Trends in Mechanical and Biomedical Design: Esther Titilayo Akinlabi P. Ramkumar M. Selvaraj Editors

Lecture Notes in Mechanical Engineering
Esther Titilayo Akinlabi

P. Ramkumar
M. Selvaraj Editors
Trends in
Mechanical
and Biomedical
Design
Select Proceedings of ICMechD 2019
Series Editors
Francisco Cavas-Martínez, Departamento de Estructuras, Universidad Politécnica
de Cartagena, Cartagena, Murcia, Spain
Fakher Chaari, National School of Engineers, University of Sfax, Sfax, Tunisia
Francesco Gherardini, Dipartimento di Ingegneria, Università di Modena e Reggio
Emilia, Modena, Italy
Mohamed Haddar, National School of Engineers of Sfax (ENIS), Sfax, Tunisia
Vitalii Ivanov, Department of Manufacturing Engineering Machine and Tools,
Sumy State University, Sumy, Ukraine
Young W. Kwon, Department of Manufacturing Engineering and Aerospace
Engineering, Graduate School of Engineering and Applied Science, Monterey,
CA, USA
Justyna Trojanowska, Poznan University of Technology, Poznan, Poland
Lecture Notes in Mechanical Engineering (LNME) publishes the latest develop-
ments in Mechanical Engineering—quickly, informally and with high quality.
Original research reported in proceedings and post-proceedings represents the core
of LNME. Volumes published in LNME embrace all aspects, subfields and new
challenges of mechanical engineering. Topics in the series include:
• Engineering Design
• Machinery and Machine Elements
• Mechanical Structures and Stress Analysis
• Automotive Engineering
• Engine Technology
• Aerospace Technology and Astronautics
• Nanotechnology and Microengineering
• Control, Robotics, Mechatronics
• MEMS
• Theoretical and Applied Mechanics
• Dynamical Systems, Control
• Fluid Mechanics
• Engineering Thermodynamics, Heat and Mass Transfer
• Manufacturing
• Precision Engineering, Instrumentation, Measurement
• Materials Engineering
• Tribology and Surface Technology
To submit a proposal or request further information, please contact the Springer
Editor of your location:
China: Dr. Mengchu Huang at [email protected]
India: Priya Vyas at [email protected]
Rest of Asia, Australia, New Zealand: Swati Meherishi at
[email protected]
All other countries: Dr. Leontina Di Cecco at [email protected]
To submit a proposal for a monograph, please check our Springer Tracts in
Mechanical Engineering at http://www.springer.com/series/11693 or contact
[email protected]
Indexed by SCOPUS. The books of the series are submitted for indexing to
Web of Science.
More information about this series at http://www.springer.com/series/11236

Esther Titilayo Akinlabi P. Ramkumar
• •
M. Selvaraj
Editors
Trends in Mechanical
and Biomedical Design
Select Proceedings of ICMechD 2019
123
Editors
Esther Titilayo Akinlabi P. Ramkumar
University of Johannesburg Indian Institute of Technology Madras
Johannesburg, South Africa Chennai, Tamil Nadu, India
M. Selvaraj
Sri Sivasubramania Nadar (SSN) College
of Engineering
Chennai, Tamil Nadu, India
ISSN 2195-4356 ISSN 2195-4364 (electronic)

ISBN 978-981-15-4487-3 ISBN 978-981-15-4488-0 (eBook)
https://doi.org/10.1007/978-981-15-4488-0
© Springer Nature Singapore Pte Ltd. 2021
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Preface
We are delighted to publish the selected papers presented in the second interna-
tional conference on Mechanical Engineering Design 2019 (ICMechD 2019) held at
Sri Sivasubramania Nadar (SSN) College of Engineering, Chennai, India during
25–26 April 2019.
In this book, around 100 quality papers in the areas of design of machines and
mechanisms, fatigue and fracture mechanics, finite element analysis, fluid
mechanics and heat transfer, robotics, biomedical and control engineering, thermal
and engine engineering, tribology, wear and surface engineering and vibration
engineering were selected for publication.
We are happy to put together this collection of thoughtful papers on the theme of
“Mechanical Engineering and Biomedical Design”. We are sure that this book will
help to nurture knowledge among the research society.
We are grateful to the reviewers for their valuable suggestions for improving the
quality of the papers. Also, we thank the session chairs and organizing committee
members for their steadfast support and suggestions.
Johannesburg, South Africa Esther Titilayo Akinlabi

Chennai, India P. Ramkumar
Chennai, India M. Selvaraj
v
Contents
Design of Machines and Mechanism

Rigid Logged Coupling: An Improved Coupling with Less Material
and Low-Stress Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Aniket Modak
An Efficient Grey Water Stagnancy Handler . . . . . . . . . . . . . . . . . . . . . 15
K. Banumalar, M. Aiswarya, M. Asvanthini Eswari, and M. Karthika
Experimental Design and Testing of Scale Development Reduction
System in Domestic Water Geysers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Kabilan Sankar, Karthick Selvam, and K. Joy Ashwin
Transient Analysis of Rotor System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
D. S. Megharaj and Amit Malgol
Investigating the Contact Load Capacity of Asymmetric
Helical Gears . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
P. Dinesh Babu, Swathi Balaji, and P. Marimuthu
Close-Proximity Dynamic Operations of Spacecraft with Angles-Only
Rendezvous in a Circular Phasing Orbit . . . . . . . . . . . . . . . . . . . . . . . . 55
Thangavel Sanjeeviraja, R. Santhanakrishnan, S. Lakshmi, and R. Asokan
Optimal Dimensional Synthesis of Rhombus Path Generating
Adjustable Four-Bar Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Ganesan Govindasamy, Rajakumar S. Rai, and Gadudasu Babu Rao
Sensor Fusion for Automotive Dead Reckoning Using GPS
and IMU for Accurate Position and Velocity Estimation . . . . . . . . . . . . 83
Lalith Prabu Nalla Perumal and Arockia Selvakumar Arockia Doss
vii
viii Contents
Optimal Dimensional Synthesis of Double-Semi-circle

and Double-Semi-ellipse Path Generating Adjustable Four-Bar
Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Ganesan Govindasamy, Gadudasu Babu Rao, and Rajakumar S. Rai
Effect of Interference on a Floating Axis Epitrochoidal
Hydrostatic Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Harish Ramachandran, Aditya Shriwastava, and Abhijit Nag
Design and Analysis of Rocker Arm Shaft in IC Engine Hino
Series for Reduction of Assembling Dismantling Time . . . . . . . . . . . . . . 127
R. Iruthayaraj, S. Palani, R. Ajay Vishal, T. Living Rockson, and K. Anand
Kinematic Analysis and Dimensional Synthesis of Filleted-Rhombus
Path Generating Adjustable Four-Bar Mechanism . . . . . . . . . . . . . . . . . 139
Ganesan Govindasamy, Clinton Wilson, and K. Neeraj
Experimental Investigations of Initial Push Forces on an Industrial
Trolley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Masepogu Wilson Kumar, Gadudasu Babu Rao, K. Balasubramanian,
and I. Kantharaj
Mathematical Analysis of Stiffness of Orthotropic Beam
with Hollow Circular and Rectangular Cross-sections . . . . . . . . . . . . . . 159
Michael C. Agarana, Esther T. Akinlabi, and Anuoluwapo M. Olanrewaju
Analytical Investigation of the Stiffness of Homogenous Isotropic
Mechanical Materials with Different Cross-sections . . . . . . . . . . . . . . . . 169
Michael C. Agarana, Esther T. Akinlabi, and Okwudili S. Ogbonna
Fatigue and Fracture Analysis

Fatigue Analysis of Vocal-Folds Using Discretized Aeroelastic
Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Veda Samhitha Kanduri, James Emilian, and Venkatramani Jagadish
Response of a Layered Composite Beam Subjected to Static
Loading Using Point Interpolation Meshless Technique . . . . . . . . . . . . . 189
Kunal S. Shah, Appaso M. Gadade, and Sanjiv M. Sansgiri
Dynamic Behaviour of Laminated Composite Beam Undergoing
Moving Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Lalit Babu Saxena, Appaso M. Gadade, and Sanjiv M. Sansgiri
Demystifying Fractal Analysis of Thin Films: A Reference
for Thin Film Deposition Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
F. M. Mwema, Esther T. Akinlabi, and O. P. Oladijo
Contents ix
Dynamic Analysis of Rectangular Aluminum Plate Under Transverse

Loading Using Finite Difference Algorithm . . . . . . . . . . . . . . . . . . . . . . 223
Michael C. Agarana, Esther T. Akinlabi, and Michael O. Ikumapayi
Ballistic Performance of Light Weight Magnesium (AZ31B)
and Aluminium (AL 6061) Plates Using Numerical Method . . . . . . . . . . 231
M. Selvaraj, S. Suresh Kumar, and Ankit Kumar
Comparison of Energy Absorption Characteristics of the Plain Fold
and Spot-Welded Fold Tubes Under Three-Point Bending . . . . . . . . . . . 239
M. Nalla Mohamed and R. Sivaprasad
An Efficient Energy Absorber Based on Welded Fold Tubes for
Automotive Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
Finite Element Analysis

Finite Element Modelling of a Compression Test on AISI 1016
Cylindrical Steel: A Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
V. Musonda and E. T. Akinlabi
Evaluation of Tie Wing Deformation in 0.022 Inch Stainless Steel
Orthodontic Bracket—A Finite Element Analysis . . . . . . . . . . . . . . . . . 277
Akhil Minu Ajayan, V. Magesh, P. Harikrishnan,
and D. Kingsly Jeba Singh
A Finite Element Analysis to Study the Effect of Various Loading
Conditions on the Intervertebral Disc in L4–L5 Section
of Lumbar Spine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
J. Daniel Glad Stephen, M. Prakash, V. K. Nevedha, and Manu Pandey
Numerical Simulation of a Small-Scale Shock Tube Using
OpenFOAM® . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
Satheesh Kumar Aravind, Anugya Singh, and B. T. Kannan
Contact Stress Analysis on a Functionally Graded Spur Gear Using
Finite Element Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
V. Aravind, S. Adharsh, D. Prakash, S. Martin Vinoth, and P. Marimuthu
Finite Element Analysis of Knee Joint with Special Emphasis on
Patellar Implant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
M. A. Kumbhalkar, D. T. Rangari, R. D. Pawar, R. A. Phadtare,
K. R. Raut, and A. N. Nagre
Simulation and Hardware Implementation of Interleaved SEPIC
Converter with Valley-Fill Circuit for HBLED System . . . . . . . . . . . . . 335
B. Lakshmi Praba and R. Seyezhai
x Contents
Finite Element Analysis of a Two-Post Rollover Protective Structure

of an off-Highway Motor Grader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
V. Kumar, G. Mallesh, and K. R. Radhakrishna
Fluid Mechanics and Heat Transfer

Investigating the Route to Flutter in a Pitch–Plunge Airfoil Subjected
to Combined Flow Fluctuations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
Nivedhithaa Santhakumar, Divyangi Singh, and Venkatramani Jagadish
A Computational Study on the Effect of Supercritical CO2
in a Combustor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
Vishnumaya Nair and Selvaraj Prabhu
Design of Shell-and-Tube Heat Exchanger with CFD Analysis . . . . . . . 393
B. Jayachandraiah and C. Dinesh Kumar Patel
Characterization of Rayleigh–Taylor Instability at the Fluid–Fluid
Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
Nihal Raj and S. Karthick
Heat and Mass Transfer Analysis of Al2O3-Water and Cu-Water
Nanofluids Over a Stretching Surface with Thermo-diffusion and
Diffusion-Thermo Effects Using Artificial Neural Network . . . . . . . . . . . 417
D. Seenivasan, M. Elayarani, and M. Shanmugapriya
Design of Macro-rough Surface and Its Influence on Side Wall
Heated Square Enclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
Ashwin Mahendra and Rajendran Senthil kumar
Performance Enhancement of a Savonius Vertical Axis Wind
Turbine with Bio-Inspired Design Modifications . . . . . . . . . . . . . . . . . . 449
S. M. Hasan Fayaz, Uditya Tyagi, Apurva Jain, and Nishant Mishra
Experimental Investigation of Heat Treatment Processes on Dissimilar
IS2602-EN9 MIG Welded Joint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
Mallapuram Bala Chennaiah, Podaralla Nanda Kumar,
and Karanam Prahlada Rao
Feasibility of Al2O3/Water Nanofluid in a Compact Loop
Heat Pipe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
Emerald Ninolin Stephen, Lazarus Godson Asirvatham,
Kandasamy Ramachandran, Arulanatham Brusly Solomon,
and Pitchaimuthu RamKumar
Analysis of Externally Pressurized Thrust Bearing with Inclination
Angle Using Yield Stress Fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485
G. Alexander Raymand and I. Jayakaran Amalraj
Contents xi
Numerical Investigation of Bingham Fluid Flow in the Entrance

Region of Rotating Annuli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
S. Mullai Venthan and I. Jayakaran Amalraj
Nanofluids in Improving Heat Transfer Characteristics of Shell
and Tube Heat Exchanger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519
S. Seralathan, R. Vijay, S. Aravind, S. Sivakumar, G. Devaraj, V. Hariram,
P. S. Raghavan, and T. Micha Premkumar
Design and Numerical Analysis of Gearless Transmission Used
in Small Wind Turbine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
Micha T. Premkumar, V. Hariram, S. Seralathan, Pinku Kumar,
and Godwin John
Experimental Investigation of Unsteady State Heat Transfer
Behaviour of Nanofluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 543
Yogiraj Bhumkar, A. R. Acharya, and A. T. Pise
Robotics, Biomedical, and Control Engineering

Switched Staircase-Type Multilevel Inverter Structure with Reduced
Number of Switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557
V. Thiyagarajan
Hydro Pneumatic Parking Brake Actuation System for Motor Grader
Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 569
K. Rajasekar, S. Karthikeyan, V. Kumar, and H. S. Satish Chandra
Design and Analysis of Feature Primitive Scaffold Manufactured
Using 3D-Printer—Fused Deposition Modelling (FDM) . . . . . . . . . . . . . 577
P. Vishnurajan, S. Karuppudaiyan, and D. KingslyJeba Singh
Design and Development of Modular Parallel XY Manipulator . . . . . . . 589
Santosh R. Thorat and R. B. Patil
Design of hyper-redundant In-Vivo Robot: A Review . . . . . . . . . . . . . . 603
Sushil Dev Rout, Mihir Kumar Sutar, and Sarojrani Pattnaik
Effect of Insertion Force for Successful Penetration of a Conical
Shaped Microneedle into the Skin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613
Gera Aswani Kumar, Burra Rajesh Kumar, Vankara Jyothi,
and Sowmya Injeti
Lumbar Discography: Study of Biomechanical Changes in the L1-L2
Intervertebral Disc of the Human Lumbar Spine Using Finite Element
Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 623
S. Balamurugan, P. Susai Manickam, and Sachit Chawla
xii Contents
C4–C5 Segment Finite Element Model Development and Investigation

of Intervertebral Disc Behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 635
P. Susai Manickam
Simulation and Analysis of Integrated SEPIC-Flyback AC-DC PFC
Converter for LED Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645
Sridhar Makkapati and Seyezhai Ramalingam
Simulation Study of Shading Effects in PV Array . . . . . . . . . . . . . . . . . 655
S. Harika, R. Seyezhai, and A. Jawahar
Speech Recognition Using Neural Network for Mobile Robot
Navigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 665
Prashant Patel, Arockia Selvakumar Arockia Doss, L. PavanKalyan,
and Parag J. Tarwadi
Range Sensor-Based Obstacle Avoidance of a Hyper-Redundant
Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 677
G. Mohammed Nawaz and N. Karthikeyan
Literature Survey on Four-Legged Robots . . . . . . . . . . . . . . . . . . . . . . . 691
G. Satheesh Kumar, S. Aravind, R. Subramanian,
Shashank Kannan Bharadwaj, Ronak R. Muthuraman, R. Steve Mitchell,
Aditya Bucha, M. Sriram, K. J. Shri Hari, and Nikhil Jerome Robin
In-Pipe Robot Mechanisms—State-of-the-Art Review . . . . . . . . . . . . . . 703
G. Satheesh Kumar and D. Arun
Thermal and Engine Engineering

Design and Thermal Performance Analysis on Solar Still . . . . . . . . . . . 717
Shaik Subhani and Rajendran Senthil Kumar
The Use of Hydrocarbon Refrigerants in Combating Ozone Depletion
and Global Warming: A Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 731
T. O. Babarinde, S. A. Akinlabi, and D. M. Madyira
Experimental Study of Performance of R600a/CNT-Lubricant
in Domestic Refrigerator System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 741
Assessing the Predicting Capability of RSM and ANN on
Transesterification Process for Yielding Biodiesel
from Vitis vinifera Seed Oil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753
V. Hariram, A. Bose, S. Seralathan, J. Godwin John,
and T. Micha Premkumar
Waste Heat Recovery (WHR) of Diesel Engine Using Closed-Loop
Pulsating Heat Pipe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 765
Saurabh B. Dhone and A. T. Pise
Contents xiii
An Experimental Investigation of Heat Transfer in Heat Exchangers

Using Al2O3 Nanofluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 775
S. Bhuvaneswari and G. Elatharasan
Numerical Study of Heat Transfer and Pressure Drop
in a Helically Coiled Tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 785
Tribology, Wear and Surface Engineering

Comparative Study on the Mechanical Performance of Solid
Lubricants Over Peek Polymer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 799
N. Venkatesh and P. K. Dinesh Kumar
ZrB2 Influences on the Dry Sliding Wear Resistance
of AA7075 Alloy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 809
P. Loganathan, A. Gnanavelbabu, K. Rajkumar, and S. Ayyanar
Investigation on Wear Properties of Nickel-Coated Al2O3P-Reinforced
AA-7075 Metal Matrix Composites Using Grey Relational
Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 819
D. Vijay Praveen, D. Ranga Raju, and M. V. Jagannadha Raju
Optimizing the Tribological Properties of UHMWPE Nanocomposites
—An Artificial Intelligence based approach . . . . . . . . . . . . . . . . . . . . . . 831
A. Vinoth, K. N. Nirmal, Rohit Khedar, and Shubhabrata Datta
Friction and Wear Study of Laser Surface Textured Ti-6Al-4V
Against Cast Iron and Stainless Steel Using Pin-on-Disc
Tribometer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 845
N. Sankara Manikandan, M. Prem Ananth, L. Poovazhagan, B. Sudarsan,
and A. Vishnuvarthan
Tribological Study on Sliding Contact Between Laser Surface
Textured Titanium and Aluminium Alloy Under Lubrication . . . . . . . . 855
Ramesh Rajesh, M. Prem Ananth, Sangam Harish, Sakthivel Balaji,
and R. Sivaguru
Some Studies on Surface Roughness of AISI 304 Austenitic Stainless
Steel in Dry Turning Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 869
D. Philip Selvaraj and P. Richard Philip
Vibration and Control Engineering

Influence of Fiber Orientation on Mechanical Properties and Free
Vibration Characteristics of Glass/Hemp Hybrid Composite
Laminates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 881
R. Murugan, V. Muthukumar, K. Pradeepkumar, A. Muthukumaran,
and V. Rajesh
xiv Contents
Free Vibration Analysis of Functionally Graded Beam with Linearly

Varying Thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 891
Rajat Jain and Mihir Chandra Manna
Parametric Study of Composite Plate Using Free Vibration
Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905
Rajendra Kumar Sadangi, Mihir Kumar Sutar, and Sarojrani Pattnaik
Free Vibration Analysis of Hybrid Composite Beam Under Different
Boundary Conditions and Thermal Gradient Loading . . . . . . . . . . . . . . 915
Prabhat Pradhan, Mihir Kumar Sutar, and Sarojrani Pattnaik
Computational Analysis of Boring Tool Holder
with Damping Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 925
G. Lawrance, P. Sam Paul, and A. S. Varadarajan
Comprehensive Review of the Effects of Vibrations on Wind Turbine
During Energy Generation Operation, Its Structural Challenges
and Way Forward . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 935
I. P. Okokpujie, E. T. Akinlabi, N. E. Udoye, and K. Okokpujie
Design and Analysis of a Novel Hybrid Car Bumper Using
Non-Newtonian Fluid and High-Density Polyethylene . . . . . . . . . . . . . . 949
K. Praveen Jerish, J. Rakesh Kumar,
R. Ramakrishnan, and K. S. Vijay Sekar
Vibration Signature-Based Monitoring on FSW Process
and Verification by FEA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 959
Akshay Todakar, R. S. Nakandhrakumar, M. Ramakrishnan,
and V. Meenakshisundaram
Power Generation from Hydraulic Shock Absorber Using
Piezoelectric Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 971
B. Jain A. R. Tony, M. S. Alphin, and V. Yeshwant
Static Deformation Analysis with and Without of Piezo-electric
Material Attachment in Hydraulic Suspension System . . . . . . . . . . . . . . 979
B. Jain A. R. Tony, M. S. Alphin, and Nishanth P. Shah
Study of Vibration and Tensile Characteristics of Multilayer
Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 987
D. Vishal, M. Selvaraj, and S. Vijayan
Influence of Vibro-isolator Attachment for a Jackhammer
to Reduce Vibration Discomfort . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 995
B. Jain A. R. Tony, M. S. Alphin, and Vishal Venkatesh
About the Editors
Professor Esther Titilayo Akinlabi is a Full Professor at the Department of

Mechanical Engineering Science, Faculty of Engineering and the Built
Environment, University of Johannesburg, South Africa. She has had the privilege
to serve as a Head of Department of the Department of Mechanical Engineering
Science and as the Vice Dean for Teaching and Learning at the University of
Johannesburg. Her research interest is in the field of modern and advanced man-
ufacturing processes – Friction stir welding and additive manufacturing. Her
research in the field of laser based additive manufacturing includes laser material
processing and surface engineering. She also conducts research in the field of
renewable energy, and biogas production from waste. She is a rated National
Research Foundation (NRF) researcher and has demonstrated excellence in all
fields of endeavors. Her leadership, mentorship and research experience is enviable
as she guides her team of postgraduate students through the research journey. She is
a recipient of several research grants and has received many awards of recognition
to her credit. She is a member of the prestigious South African Young Academy of
Science and registered with the Engineering Council of South Africa. Prof Akinlabi
has filed two patents, edited two books, published five books and authored/
co-authored over 400 peer-reviewed publications.
Dr. P. Ramkumar is currently working as an Associate Professor in the Depart-

ment of Mechanical Engineering in IIT Madras. Prior to coming to IIT Madras,
he was an Associate Professor at SSN College of Engineering, Chennai.
Dr. Ramkumar received his B.E in Mechanical Engineering from College of
Engineering, Guindy, Anna University and his M.Tech (Industrial Tribology) from
IIT Madras. He obtained his Ph.D. (2003-07) from University of Southampton with
specialisation of Engine Tribology, UK and did his post-doctoral fellowship in
University of Leicester (2009-10), UK for developing high speed Cam-Tappet
tribometer Rig. He has 14 years research experience in the field of tribology and
engine tribology. He won the best presentation award for his PhD from IMechE,
London, UK in 2005. His primary research interests are in the field of tribology,
xv
xvi About the Editors
engine tribology, gearbox design, wind turbine gearbox bearing failures, biotri-
bology, multilayer coatings, ionic fluids, wear modelling and developing new
composite materials for brake applications. He has published more than 20 papers
in peer-reviewed journals and 40 international conference proceedings including
1 patent as well. As sole PI, he has received two research grants with total worth of
US $1.4M, while as Co-PI US $ 14M from DST. Currently, he is handling a lot of
consultancy projects from GE, Caterpillar, Rane, Ford, Kone Elevator and L&T.
In his supervision 3 MS scholars and 14 M. Tech. students were graduated and
currently he is supervising 8 Ph.D., 5 MS research scholars, 2 project associates,
1 JRF and 2 M.Tech. degrees students. He is a member in IMechE, IoP, STLE, and
life time member in TSI. He is reviewer to various top journals including Applied
Surface Science, Mechanism and Machine Theory, Part J : Journal of Engineering
Tribology, Surface Coatings and Technology, Tribology International, Tribology
Letters, Wear, etc.
Dr. M. Selvaraj, Associate Professor in the Department of Mechanical Engi-

neering at Sri Sivasubramania Nadar (SSN) College of Engineering has over 20
years of teaching experience. He did his B.E. in Mechanical Engineering from
Madurai Kamaraj University, M.E. in Engineering Design from Government
College Technology, Coimbatore and Ph.D. from Anna University, Chennai. His
research areas are mechanical vibrations, composites and modeling of friction stir
welding. He has authored 14 papers in journals of national and international repute
and presented over 20 papers in various conferences. He is a member of Indian
Society of Technical Education (ISTE) and American Society of Mechanical
Engineers (ASME).
Design of Machines and Mechanism
Rigid Logged Coupling: An Improved
Coupling with Less Material
and Low-Stress Concentration
Aniket Modak
Abstract The paper is about a new coupling design that can be used to connect two
shafts in the same line without any axial or angular misalignment. The goals of the
new design are to reduce the amount of material required to produce other couplings
and reduce the chances of failure due to different stress distribution patterns. Rigid
flanged coupling is used as a reference for comparison. The design consists of mainly
three parts body, pins and logs (strips of metal used to connect the body). The holes
are done on outer surfaces of the hexagonal or a square prism of each body while
one surface of each body faces each other. The logs are used to join the two bodies
with the help of pins in the holes. The coupling is designed using Autodesk Fusion
360. Proper mathematical analysis of the couplings is done to get the difference in
the material used. This design can be used to connect both, shafts with the same
diameter or shafts with different diameters. The design has less stress concentration
around holes and in whole design in general and is similar in the manufacturing
process as the presently available couplings. For the safety of the operator or user,
the coupling is covered with a protective case to avoid the failed parts (if any) to
hit the operator or the user. The dimensions are found using stress analysis and
considering the modes of failure for each section of the part. Shear and tensile failure
have been considered at the parts of the assembly. The bending effect has also been
considered for parts wherever adequate. The stress analysis results have also been
found out using Autodesk Fusion 360. To observe the effects and results of stress
analysis, similar working conditions have been simulated with the coupling designed
in Autodesk Fusion 360.
Keywords Coupling · Rigid · Logged · Design
A. Modak (B)
Jalpaiguri Government Engineering College, Jalpaiguri, West Bengal, India
e-mail: [email protected]
© Springer Nature Singapore Pte Ltd. 2021 3

E. T. Akinlabi et al. (eds.), Trends in Mechanical and Biomedical Design,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4488-0_1
4 A. Modak
1 Introduction
Most of the rigid couplings popularly in use today comprise two flanges attached
to each other and transmitting motion, power and torque. This particular method of
coupling is required to join two shafts that are one from the power receiving end and
the other power generation end. The flanges, for manufacturing, require a lot of metal
and due to the various methods by which they are joined, stress is also concentrated
to certain parts which lead to failure of the flanges or a high factor of safety needs to
be considered.
The use of the designed coupling is to use less material considering the same
factor of safety and even more reduction of material when less factor of safety is
considered due to the reduction in stress concentration in power transmission areas.
2 Design Process
2.1 Logged Coupling
See Figs. 1 and 2.
2.2 Empirical Relations [1]
Empirical relations are mostly the pre derived, assumed and universally accepted
relations. In this derivation, the empirical relations are kept the same as rigid flanged
coupling.
Fig. 1 Full assembly

Rigid Logged Coupling: An Improved … 5
Fig. 2 Section of full

assembly

16M
d=
3
(1)
πτ
dh = 2d (2)
ll = d (3)
lh = 1.5d (4)
t =d (5)
Din = 2.5d (6)
Din
b1 = 1.25d = (7)
2
0.5d
d1 = √ (8)
N
In the above relations, the terms used have the following meanings (Table 1):
d Diameter of the shaft
M Moment
τ Torque to be transmitted
dh Diameter of the hub of flange
ll Length of the flange
lh Length of the hub
t Thickness of the flanges
Din Inner diameter of the flange (Flanged coupling)
6 A. Modak
Table 1 Values for no. of

Value of N d start d end
side of polygon with respect
to diameter in mm 3, 4 0 20
6 20 80
8 80 120
10 120 –
All the calculations for rigid logged coupling have been done for
diameter size 20–80, that is, N = 6
b1 Size of outer edge

d1 Diameter of pins.
2.3 Shear Failure of Log
See Fig. 3.
b
Shear area = ∗h
2
P
τ= (9)
b
2
∗h
√
Din 3
P = M/ ∗ (10)
2 2
b ≤ b1 (11)
Fig. 3 Maximum shear area

of log
√
2
Din 3 b
τ = M/ ∗ ∗ ∗h (12)
2 2 2
√
b ∗ h = 8M/ 2.5 ∗ d ∗ 3 ∗ τ (13)
π
τ∗ ∗ d12 = P (14)
4
√
Din 2
3 π
d1 = M/τ ∗
2
∗ ∗ (15)
2 2 4

√
d1 = 4 ∗ M/ τ ∗ 2.5 ∗ d ∗ 3 ∗ π (16)
In the above relations the new terms used has following meanings:
b: Breadth of the log.
2.4 If Pins Are Loosely Kept [Bending Consideration]
See Fig. 4.
lb
M=P∗ (17)
2
M 4M
P= √ = √ (18)
Din
2
∗ 2
3 2.5 ∗ d ∗ 2 ∗ 3
4Mlb
Mb = √ (19)
2.5 ∗ d ∗ 2 ∗ 3
Fig. 4 Bending of pin

8 A. Modak
32Mb
σb = (20)
π d13
32 4M
σb = ∗ √ (21)
π d13 2.5 ∗ d ∗ 2 ∗ 3
64 ∗ M ∗ lb
d1 = 3
√ (22)
2.5 ∗ 3 ∗ d ∗ π ∗ σb
2.5 Shear Stress in Connecting Pin
See Fig. 5.
π
τ = P/ ∗ d22 (23)
4

dh π
τ = M/ ∗ ∗ d22
(24)
2 4

∴ d2 = 4M/{π ∗ d ∗ τ } (25)
Fig. 5 Shear area of

connecting pin
3 Simulation Data and Results
The rigid logged coupling and the rigid flanged coupling were simulated in Autodesk
Fusion 360 under different conditions; the conditions of rigid logged coupling having
more tough testing conditions.
The conditions and results are listed below for both rigid flanged and rigid logged
coupling.
3.1 General Settings
See Tables 2, 3, 4 and 5.
Table 2 General
Contact tolerance 0.1 mm
Remove rigid body modes No
Table 3 Mesh
Average element size (% of model size)
Solids 10
Scale mesh size per part No
Average element size (absolute value) –
Element order Parabolic
Create curved mesh elements Yes
Max. turn angle on curves (deg.) 60
Max. adjacent mesh size ratio 1.5
Max. aspect ratio 10
Minimum element size (% of average size) 20
Table 4 Adaptive mesh

Number of refinement steps 0
refinement
Results convergence tolerance (%) 20
Portion of elements to refine (%) 10
Results for baseline accuracy Von Mises stress
Table 5 Materials
Component Material Safety factor
Component 1:1 Steel Yield strength
10 A. Modak
Table 6 Mesh
Type Nodes Elements
Solids 4800 2682
Table 7 Pressure1
Type Pressure
Magnitude 4.8 MPa
Table 8 Pressure2
Type Pressure
Magnitude 1.896 MPa
Fig. 6 Von Mises

[MPa] 0.17 15.5
3.2 Rigid Flanged Coupling
See Tables 6, 7, and 8.
3.2.1 Results
See Fig. 6.
3.3 Rigid Logged Coupling
Loads (Tables 9, 10, and 11).

Results (Fig. 7).
Table 9 Mesh
Type Nodes Elements
Solids 5948 3383
Table 10 Pressure1
Type Pressure
Magnitude 4.8 MPa
Table 11 Pressure2
Type Pressure
Magnitude 7.8 MPa
Fig. 7 Von Mises

[MPa] 0.48 13.65
4 Dimensions for Construction
4.1 Shaft

16M
d=
3
πτ
4.2 Flange
dh = 2d
lh = 1.5d
t =d
12 A. Modak
Din = 2.5d
Din
b1 = 1.25d =
2
4.3 Log
8M
b∗h = √
2.5 ∗ d ∗ 3 ∗ τ
ll = 2d
b < b1
h is the height of the log.
4.4 Pin
64 ∗ M ∗ lb
d1 = 3
√
2.5 ∗ 3 ∗ d ∗ π ∗ σb

M
d1 = 4 ∗ √
τ ∗ 2.5 ∗ d ∗ 3 ∗ π
4.5 Nut
dinner = d1
ln = 0.8 ∗ d1
4.6 Connector Pin

4M
∴ d2 =
π ∗d ∗τ
4.7 Protective Case
√
2
3
d pc = Din ∗ + 2h + 10
2
4.8 Material Saved as Compared to Rigid Flanged Coupling
π
5.14488467d 3 − 6bhd − N dd1 + 15π d12
8
5 Conclusions
The following comparisons can be drawn by the analysis of the two couplings namely
the rigid flanged and the rigid logged one (Table 12).
With respect to the results of simulations and derived equations, it can be
stated that logged coupling gives a better performance with less material and stress
concentration.
14 A. Modak
Table 12 Comparison
Serial Rigid flange coupling Basis Rigid logged coupling
No.
1. Stress
Maximum: 15.5 MPa Maximum: 13.65 MPa

2. Excess material required to be built Material π
5.14488467d 3 − 6bhd − 8 N dd1 +
15π d12
Volume material saved
Reference
1. Design of Machine elements by Bhandari

An Efficient Grey Water Stagnancy
Handler
K. Banumalar, M. Aiswarya, M. Asvanthini Eswari, and M. Karthika
Abstract This work proposes the idea of “Human hair boom skimmer” in addition
to the conventional flow grey water treatment. The conventional method of treating
grey water doesn’t pose a promising solution in reducing the amount of oil content
from it. Rather than discarding the human hair, it can be used to absorb the oil from
the grey water. The oil adsorption capability of wasted hair fibers could produce
valuable solutions for our prevalent and modern society. Our concept “An Efficient
Grey Water Stagnancy Handler” concentrates on three major phases: Phase-1, deals
with the drastic problem being faced by the surrounding environment with slip and
falls due to the improper collection of grey water from the bathroom floors, Phase-
2 address the eco-friendly and cost-efficient methodology to treat the grey water,
Phase-3 tackles the dwindling water resources for crops cultivation by the diversion
of processed water to the required place wherein immediate need of it. Thereby,
the grey water from the bathroom sink is no doubt a valuable resource that can be
used to alleviate the water shortage problems and increase water conservation in
individual households. Therefore, an alternative eco-friendly method rather than the
conventional method as mentioned above could pose a firm solution.
Keywords Grey water · Skimmer · Irrigation control
1 Introduction
This work on “An efficient grey water stagnancy handler” highlights the solutions
for the slip and falls in the bathroom, water scarcity for crops cultivation, dwindling
of ground water resources. Every year about 235,000 people over age 15–65 visit
emergency rooms because of injuries suffered in the bathroom, and almost 14% are
being hospitalized. In this case, about one-third of the injury happens while bathing
or showering. Injuries increase with the age and get peaking after 85, according to
K. Banumalar (B) · M. Aiswarya · M. Asvanthini Eswari · M. Karthika

Department of Electrical and Electronics Engineering, Mepco Schlenk Engineering College,
Sivakasi, Tamil Nadu, India

https://doi.org/10.1007/978-981-15-4488-0_2
16 K. Banumalar et al.
the research results. But injuries around the tub or shower are proportionately most
common among those ages 15–24 and least common among those over 85. Injuries
in or near the bathtub or shower account for more than two-thirds of emergency
room visits. In addition to this, about 85% of water resources are deployed in the
agriculture sector in India. Every year, hundreds of farmers committing suicides and
unable to deal with the devastating effects of water shortages on their crop production.
Small farmers being poor and can’t afford more advanced agriculture tools to extract
water and to conserve it. The adversity for small farmers is poverty and illiteracy.
As the population increases, the utilization of groundwater and surface water for
the domestic, industrial sectors and agriculture exaggerate, leading to tensions and
extreme pressure on the environment. Groundwater is a critical resource for farmers.
Over half the irrigated agricultural areas is dependent on groundwater. According
to 2016 World Bank report made it clear that our groundwater levels are sharply
falling and showing a downward trend. Over 50% of the country is already facing
extremely high-water stress. This document explains the ways and means to avoid
the stagnation of water in the bathroom and also poses the solution to reduce the
oil content in the grey water than by the conventional treatment method [1, 3] and
to meet the dwindling water resources for the crop cultivation. The objective of our
work:
1. To reduce the oil content in the grey water with an idea of human hair boom and
coconut husk skimmer.
2. To improve the oil content reduction than the conventional grey water treatment
method.
3. To clear the blockage for plants to produce food through photosynthesis with no
effect of oil content in recycled grey water.
2 Proposed System
The overall system has been divided into three major parts, which include stagnancy
prevention, grey water treatment, and garden irrigation as shown in Fig. 1.
2.1 Phase-1: Stagnancy Handler—Effective Collection
Figure 2 shows the implementation of stagnancy handler with the help of pervious
concrete. The main objective of this part is to collect the water that falls on the
bathroom floors with the implementation of pervious concrete tiles over the floors
[2]. As of now, the grey water has been diverted through small opening from the
bathrooms. It is clear that not fully the entire water that falls on the floor gets diverted.
Little or some water may stagnant in the floors which makes the floor prone to slip.
Therefore, it’s very important to make the floor free from water so that contact
An Efficient Grey Water Stagnancy Handler 17
Fig. 1 Overall flow of an efficient grey water stagnancy handler
Fig. 2 Implementation of
pervious concrete
between the feet and ground is never lost, thereby the possibility of loss balance can
be reduced. Pervious concrete can be inserted as tiles, over the bathroom ground
area so that water that falls on it would get percolate through them and made to get
collected in a storage tank.
2.2 Phase-2: Water Retreatment
Grey water denotes the waste water coming from the bathroom floors. This grey
water collected in a storage tank can be reused for non-potable purposes, with mild
filtration. According to the study about the characteristic of this grey water, it is clear
that grey water can be reused in gardens with little or no treatment.
Besides, the untreated grey water can’t be stored for more than one day. The
organic matter present in it gets multiplied over time results in bad odour and
makes the water unhealthy. It may also pose serious issues when this water (with
pathogens from sick people) comes into contact with humans. However, if the water
left untreated, this water can’t be extended for our future use. After taken into consid-
eration, its effect and nature it is necessary to treat the water to remove the bacteria
and viruses. Figure 3 depicts the proposed grey water treatment.
This treatment includes three stages:
Stage 1: Screening: water collected in the storage tank is made to pass through
lint filter, thereby small suspended particles in grey water would be eliminated.
Stage 2: Human Hair boom skimmer [3], the grey water contains oils in it which
may remove efficiently by using hair boom. As these hair booms are made of nylon
cloth filled with human hair, thereby making it cost-efficient when compared to other
skimming polymer materials. The manual operation of this hair boom skimming
process has been undertaken in an environmental laboratory and positive testing
results are obtained from the final skimmed water sample.
Stage 3: Filtration
Fig. 3 Proposed water treatment method

• Sediment filter: Sediment filter acts as a sieve to remove any particulate matter
that can be transported by fluid flow and which eventually is deposited as a layer
of solid particles on the bed or bottom of the grey water.
• Carbon filter: Carbon filtering is a method of filtering that uses a bed of activated
carbon to remove contaminants and impurities using chemical adsorption.
After the completion of the entire treatment, the water should be suitably stored
in a storage tank.
2.3 Phase-3: Auto Irrigation Control
The treated water is now suitable for watering gardens. In order to make the irrigation
to plants automatic and efficient, by sensing the moisture content of the soil the water
is diverted to the plant’s usage. As a result, plants would get sufficient supply of water
depending upon its requirement without overflow and prevents the spoilage of root
system [4].
Figure 4 depicts the auto irrigation control, which is achieved with the help of
interfacing soil moisture sensor, pumping motor, and display with Arduino Uno.
With the sensing of soil moisture content, the motor turns on and off to pump the
water.
Fig. 4 Auto irrigation control

Fig. 5 Conventional treatment method
Fig. 6 Grey water treatment employed with human hair and coconut husk
3 Results and Analysis
3.1 Conventional Grey Water Treatment
Figure 5 shows conventional treatment method. Initially the grey water is screened
to remove the small and medium particles present in it. After the screening process,
the treatment of grey water is followed by passing the outcome of screened grey
water into the sediment filter. In this sediment filtration, very small particulates get
removed and it is sent to the carbon filter to reduce the hardness content in the grey
water. Finally, the water is being collected into a recycled storage tank for future
usage. The treated grey water sample is found with oil content left in it. When this
water is being used for plantation purposes, the plant growth is diminishing day by
day.
3.2 Novel Idea of Grey Water Treatment
In order to reduce the amount of oil content in the resultant treated sample, a novel
idea is proposed to introduce the human hair and coconut husk skimmer between the
screening and sediment filtration process as shown in Fig. 6. The resultant grey water
after the skimming with human hair and coconut husk and with the conventional flow,
the amount of oil content is greatly reduced [5–10].
3.3 Hardware Arrangement
Figure 7 depicts the complete flow of the project from left to right. The process
starts with the effective collection of grey water from the bathroom floors with an
arrangement of pervious concrete over the floors. The collected water from the floors
is collected in a pink storage tank. The water from the pink tank is pumped into the
Fig. 7 Hardware setup
orange storage tank with the help of mini submersible water pump(motorpump1). The
water pumping is controlled with the help of two float switch(waterlevelsensor1and
waterlevelsensor2) inside the skimming tank. The grey water once reached the top
float switch the motor 1 in the storage tank turns off. The wheel and clamp arrange-
ment with hair and coconut husk enters into the up and down motion for a period
of 10 s. After this, the wheel setup moves forward by giving pulse to 2 dc motor in
the wheel to reach the compressing tank with sieve arrangement to dispose of the
adsorbed oil content in the boom through to and fro motion into the sieve. After
compression, the wheel is given with the pulse to move backward and to reach the
skimmer tank for further filtration. This process continues and the entire flow is being
controlled by the Arduino Mega kit. The water from the skimming tank is pumped
into another storage tank with the help of motorpump2 after the skimming process.
The treated grey water is now allowed to pass through two filters sediment and carbon
filter to remove the hardness content in the grey water. The treated water from the
filters is now stored in a tank for the auto irrigation purpose. This control is achieved
with the sensing of soil moisture content and Arduino control. Depending upon the
moisture content the water pumping turns on and off.
3.4 Experimental Result
Our experimental results are shown in Table 1.

Our experimental results are shown in Table 1. The quantity of input grey water
sample is about 3.5 L and the quantity of output grey water sample is about 2.5 L.
In conventional grey water treatment the amount of oil content reduction is about
11.27% whereas, with human hair and coconut husk, it is found that the reduction of
Table 1 Comparative analysis of grey water samples

S. No Method of treatment Amount of oil content Percentage of decrease (%)
(mg/L)
1 Conventional method [5] 282.3 11.27
2 Coconut husk skimming 300.18 5.6
method
3 Oil skimming method 250.48 21.25
(human hair boom +
coconut husk)
4 Combination of three 166.9 47.51
methods (human hair
boom skimming +
coconut skimming +
conventional method) [6]
Quantity of input water sample = 3.5 L
Quantity of output water sample = 2.5 L
Amount of oil content in untreated grey water sample = 318.08 mg/L
oil content is 21.25%. The entire treatment process of our project results in 47.51%
reduction of oil content in the grey water sample.
4 Conclusion
This project thereby highlights the socio-economic importance of the grey water
handling technique to meet the following situations, with the reuse of discarded
human hair, the additional stress on the environment by human hair waste disposal
can be drastically reduced. The effective collection of grey water from bathroom
floors with the use of pervious concrete can reduce the slip and fall accidents faced
by the seniors inside the bathrooms. Water scarce problems for the farming lands
can be met out for this proposed method.
References
1. Jadhav AS, Naniwadeka MY, ShindeN HS, Anekar V (2011) Study of adsorbtion of oil from
oily water using human hair. Int J Adv Eng Technol II:37–51
2. Obla KH (2010) Pervious concrete—an overview. Indian Concr J 84(8)
3. Gupta A (2014) Human hair waste and its utilization: gaps and possibilities, Hindawi Publishing
Corporation. J Waste Manage
4. Capral A (2001) Scicolone: wastewater reuse by drip irrigation. Trans Biomed Health 5
5. Parameshwara Murthy PM, Sadashiva Murthy BM, Kavya S (2016) Greywater treatment and
reuse: a technological review. Crit Rev Environ Sci Technol 41(7):657–698
6. Singh S, Pradhan N, Ojha n, Roy b, Bose S (2016) Grey water treatment and its application in
cultivation of plants. Asian J Asian J Microbiol Biotechnol Environ Sci 18(4)
7. Ingole NW (2013) Adsorption of oil from waste water by using human hair. J Environ Sci
Comput Sci Eng Technol 3
8. Capra A (2011) Wastewater reuse by drip irrigation. Trans Biomed Health 5
9. Libutti A, Gatta G, Gagliardi A, Vergine P, Pollice A, Beneduce L, Disciglio G, Tarantino E
(2018) Agro-industrial wastewater reuse for irrigation of a vegetable crop succession under
Mediterranean conditions. Agric Water Manage 196:1–14
10. Valipour M, Singh VP (2016) Global experiences on wastewater irrigation: challenges and
prospects. In: Balanced urban development: options and strategies for liveable cities
Experimental Design and Testing of Scale
Development Reduction System
in Domestic Water Geysers
Kabilan Sankar, Karthick Selvam, and K. Joy Ashwin
Abstract The aim of this project is to reduce the extent of scale formation in domes-
tic water geysers with minimal investment that is efficient in cleaning and can be
commercialized for consumer use. Modern water geysers have the disadvantage of
a life of maximum of one year due to formation of scales on the heating coil and
the walls of the geyser thereby reducing the efficiency of boiler and decreasing the
life of geyser. The geyser is to be fitted with a 3D printed polymer float bounded
by a soft brush. The variation in level of water inside the geyser will cause up and
down movement of the float, brushing the internal wall of the geyser removing any
salt presence. A nozzle system will be activated during the manual cleaning phase
to provide a more efficient cleaning for the internal parts of the geyser and to wash
away the salts from the brushes while drain them through a separate pipe. A valve
system controls the direction of flow of fluid during the cleaning. The testing was
done for a period of 240 days. The results showed a significant improvement in the
performance of the geyser and reduced the extent of scale formation.
Keywords Domestic water geyser · Scale formation · 3D printed polymer float ·

Nozzle system · Valve system
1 Introduction
Scale formation occurs due to the presence of salts in water. Generally, water is
classified into two types based on presence of salt. One type is hard water, and the
other type is soft water. The main reason for scale formation in the water geyser
is hard salts that present in the water. The presence of hard salt in water leads to
clogging of the geyser outlet pipe and malfunction of electrical units due to leakage
K. Sankar (B) · K. Selvam · K. J. Ashwin

Department of Mechanical Engineering, KCG College of Technology, Chennai, India
K. Selvam

https://doi.org/10.1007/978-981-15-4488-0_3
26 K. Sankar et al.
of water leading to formation of scale formation of salt deposition and rust forma-
tion which in turns to corrosion. Scale formation also leads to gradual occurrence
of pitting on the inner metal surface forming minute cracks or craters. McAdam
and Parsons [1] performed experiments on calcium carbonate scale formation and
control. Muryanto et al. [2] demonstrated the effects of calcium carbonate and salt
deposits in clogging of pipes and major pipelines. Gal et al. [3] have performed
experiments on calcium carbonate salt solubility in pipe clogging and their forma-
tion on the circumference of pipes. Kourtidou et al. [4] have worked on the impact of
scale deposition in geysers and their connection to corrosion of metal. Tyler et al. [5]
worked on experiments that studied the growth of scale deposition with increase in
temperature. The scale formation in water geyser will reduce the life expectancy of
the geyser and damages the electrical wiring in the form of leaks [6]. There are two
types of hardness in water, temporary hardness and permanent hardness. Temporary
hardness is a type of water hardness caused by the presence of dissolved bicarbon-
ate minerals (calcium bicarbonate and magnesium bicarbonate). When dissolved,
these minerals yield calcium and magnesium cations and carbonate and bicarbon-
ate anions (CO3 2− , HCO3 − ) [7]. Permanent hardness is hardness that cannot be
removed by boiling. When this is the case, it is usually caused by the presence of cal-
cium sulfate/calcium chloride and/or magnesium sulfate/magnesium chloride in the
water, which do not precipitate out as the temperature increases. The presence of the
metal cations makes the water hard [8] and increases the rate of salt deposition on
the surface of the vessel in use [9, 10]. Hence, it will increase the maintenance cost
and decrease the life expectancy of the geysers.
2 Methodology
Initially, two identical one-liter water geysers were selected for performing this
experiment. Water geyser 1 did not have the scale reduction system installed inside,
whereas water geyser 2 had the scale reduction system installed inside it. Water
geyser 1 was used at a frequency of twice a day for a 240-day period. Water geyser
2 had its tank drilled at two locations, nozzles were installed on the walls of the
geyser and the float is fitted inside the geyser. The float has some clearance between
the geyser wall and itself which is filled by the brush. This brush bounds the float.
The buoyancy force by the inlet water pushes the float up, brushing against the wall
scrapping any salt presence. The cleaning phase was done once every month, thus
totaling for eight times during this experiment. The properties of water such as hard-
ness and salt deposits in it were periodically tested every month. The growth of scale
development inside the geyser was also noted. Results were tabulated and plotted in
the form of graphs.
Experimental Design and Testing of Scale … 27
3 Experiment Details
3.1 Construction
The water geyser tank was made up of stainless steel. The two side of the cylindrical
tank was drilled for attaching nozzles. The nozzles were attached to the cylindrical
wall, with the help of adhesives, to spray the water to the internal surface of tank to
reduce the initial scale formation. The two valves are connected to the geyser system
to control water flow as shown in Fig. 1. The one pair of valves was connected in
the inlet to control inlet flow and another pair was connected in the outlet to control
outlet flow. The ABS polymer float was placed inside the water geyser. The brush
bounds the ABS polymer float to wipe out the scale formation vertically inside the
geyser. The connecting pipes were used to connect the nozzle and valves to allow
the fluid to the inlet flow. The gray water that is present inside the geyser after the
Fig. 1 Construction design of water geyser

28 K. Sankar et al.
1190
phase 1 phase 2
1180
1170
1160
Dissolved Solids (mg/l)
1150
1140
1130
1120
1110
1100
DAY 30 DAY 60 DAY 90 DAY 120 DAY 150 DAY 180 DAY 210 DAY 240
Fig. 2 Comparison of total dissolved solids in mg/l
700
phase 1 phase 2
600
500
Total Hardness (mg/l)
400
300
200
100
0
Fig. 3 Comparison of total hardness in mg/l
cleaning phase is exited through a separate outlet drain where the water is diverted
using the valve system.
3.2 Working
The primary aim of the project is to study the formation of scales in the geyser and
its effects on the efficiency of boiler, water output, and clogging of pipes. Testing of
water was initiated to get the result for the two phases. Water from both the geysers
were provided for testing at an external water testing laboratory in Chennai, India.
The results of water hardness and dissolved solids were obtained directly from the
laboratory upon submission of water samples. From the results of these two phases,
we performed an analysis to compare both the set of results to find the conclusion.
Phase 1
The geyser was setup without our cleaning system to test the scale deposition rate
and water quality for 34 weeks. The geyser was operated at a frequency of two times
per day for 240 days where the water was heated to a temperature of 85 °C. The
input water sample was tested initially. After a period of 240 days, we opened up
the geyser to find out the thickness of scale deposited in the geyser. The rate of scale
development was studied every month. The hardness and salt deposits value of the
water were also noted. The results were tabulated.
Phase 2
Now, the geyser was setup with our cleaning system, and the same frequency of
operation was continued for a period of 34 weeks. The cleaning was done once
per month. The geyser operated well for 240 days heating water to a maximum
temperature of 85 °C. The rate of scale development was studied every month. The
hardness and salt deposits value of the water were also noted. The results were
tabulated.
4 Results
4.1 Sample Collection and Testing
The output of the experiment is based on three parameters, total dissolved salts, total
hardness content, and scale formation layer. Our aim is to reduce the value of these
parameters by implementing our cleaning system. At the end of every 30-day period,
we collected water sample from the output of the geyser. As phase 1 does not have
any cleaning system, we could directly collect the water from the outlet of the geyser.
However, in phase 2, the cleaning system is activated and only after that the water
sample is collected from the geyser. The water samples are collected and submitted
for testing at an external water testing laboratory in Chennai, India. The laboratory
tests the water for total dissolved salts and total hardness, and we got the results
directly from them. The results of each period are tabulated and plotted in the form
of a graph.
Phase 1: without scale reduction system
Phase 2: with scale reduction system.
30 K. Sankar et al.
Table 1 Total dissolved solids in mg/l relative to the number of days

Day 30 Day 60 Day 90 Day Day Day Day Day
120 150 180 210 240
Phase 1 1132 1137 1140 1149 1158 1165 1174 1182
Phase 2 1129 1131 1133 1133 1135 1138 1139 1141
4.2 Total Dissolved Solids
The water sample is collected from the geyser at the end of every 30-day period to
test for content of dissolved solids present in the water. Both the phases are started
on the same day. At the end of the 30-day period, water sample is collected from the
outlet of the geyser and sent for testing to an external water testing laboratory. It can
be seen from the results table that the value of total dissolved solids increases rapidly
in phase 1 geyser compared to phase 2 geyser. In phase 1, the scale development is
more, and hence, there is a possibility of those salts on the walls of the geyser to
be dissolved into the new water that has entered the geyser. In phase 2 geyser, the
cleaning system removes the salt clearing the internal walls of the geyser of any salt.
Thus, the rise in value is minimal for phase 2 (Table 1).
4.3 Total Hardness Content
The water sample is collected from the geyser at the end of every 30-day period to
test for total hardness content present in the water. Both the phases are started on the
same day. At the end of the 30-day period, water sample is collected from the outlet
of the geyser and sent for testing to an external water testing laboratory. Hardness of
water generally depends on the area of source of water and increases with frequency
of usage of geyser since more salts get deposited over a period of time as evident in
the results of phase 1. However, in phase 2, the geyser with the cleaning system is
able to reduce the rate of increase of hardness content because the cleaning system
removes the salts deposited from the internal wall of the geyser. Thus, we are able
to control the rate of increase of hardness to a considerable extent (Table 2).
Table 2 Total hardness content in mg/l relative to the number of days

120 150 180 210 240
Phase 1 379 409 439 478 498 553 581 603
Phase 2 382 385 391 402 462 449 465 478
Table 3 Scale formation in mm relative to the number of days

120 150 180 210 240
Phase 1 0 0.16 0.39 0.53 0.67 0.89 1.0 1.15
Phase 2 0 0 0.1 0.12 0.15 0.18 0.22 0.24
4.4 Scale Formation
Scale formation is the layer of salt deposited on the internal wall of the geyser due to
the salts present in the water and frequent heating and cooling of the water present
inside the geyser. This layer of scale’s thickness varies at different points on the
geyser. Thus, we take the average of five readings. The task of measuring the scale
thickness is a tedious task as the thickness value is in millimeters. Thus, in order to
obtain accurate result, we peeled a small layer of the scale from the geyser wall and
measured using a digital gauge. A small layer of scale can be peeled or scraped off
the wall. This layer is highly delicate and fragile, and so the utmost care was taken to
take the reading as soon as it is peeled within the geyser. Initial readings were tough
as the scale layer was of negligible thickness. However, with increase in time, it was
possible to obtain the readings by repeated trial and errors (Table 3).
5 Conclusion
It is clear from Fig. 4 that the growth of scale inside the geyser has been brought down
to a considerable extent in phase 2 compared to phase 1. The reduced scale film also
helped with maintaining the hardness value of the water as shown in Fig. 3, since the
1.4
phase 1 phase 2
1.2
Scale formation in mm
0.8
0.6
0.4
0.2
0
Fig. 4 Comparison of scale formation in mm

32 K. Sankar et al.
presence of salts is lesser inside the geyser. The total dissolved solids in the water are
also maintained as shown in Fig. 2 as there is very limited amount of scale film for it to
dissolve into. From the overall experiments and analysis performed, we can conclude
that the cleaning system proved successful in its task by efficiently removing the scale
deposition on geyser with minimal labor. This has had a significant effect on the life
of the geyser and improves the durability as such. The installation of this setup is also
simplified and easier, thus reducing the maintenance cost of the geyser. Servicing a
geyser is a tedious task. It either requires an expert or had to be taken to a service
center for servicing. The geyser has to be opened every time, and scales have to be
scrapped using an abrasive sheet or any other scale removal method every time. The
float system substitutes this process, making the task easier and reducing the scale
formation to a considerable extent. Furthermore, the efficiency of the geyser is also
increased by consuming less energy when compared to energy consumption during
presence of scales in geyser. This can be justified the presence of less salts in the
water. Overall, the experiment was successful, and the results are in favor.
References
1. McAdam J, Parsons SA (2004) Calcium carbonate scale formation and control. Published
Views Environ Sci Technol 3(2)
2. Muryanto S, Bayuseno AP, Usamah M, Mamun H (2014) Calcium carbonate scale formation in
pipes: effect of flow rates, temperature, and malic acid as additives on the mass and morphology
of the scale. Procedia Chem 9:69–76
3. Gal JY, Bollinger JC, Tolosa H, Gache N (1996) Calcium carbonate solubility: a reappraisal
of scale formation and inhibition. Published Talanta 43(9):1497–1509
4. Kourtidou D, Chaliampalias D, Tarani E, Pavlidou E (2019) Deposition of Ni-Al coatings
by pack cementation and corrosion resistance in high temperature and marine environments.
Elsevier Corros Sci 148:12–23
5. Tyler J, Taler D (2007) Tubular type heat flux meter for monitoring internal scale deposits in
large steam boilers. J Heat Transf Eng 28(3):230–239
6. Jamialahmadi M, Muller-Steinhagen H (2011) Scale formation during nucleate boiling—a
review. Corros Rev 11(2)
7. The National Board of Boiler and Water Pressure Vessel Inspectors. https://www.nationalboard.
org/index.aspx?pageID=164&ID=224. Last accessed on 23 Jan 2019
8. Shen HJ, Xie ZB (2010) The causes and solutions of boiler scale in steam appliances. Published
Appl Mech Mater 37–38:1421–1424
9. Suez Water Technologies Webpage. https://www.suezwatertechnologies.com/handbook/
chapter-12-boiler-deposits-occurence-and-control. Last accessed 02 Oct 2018
10. Aries Chemicals Webpage. https://www.arieschem.com/boiler-deposit-scale-control/. Last
accessed 10 Sept 2018
Transient Analysis of Rotor System
D. S. Megharaj and Amit Malgol
Abstract This paper introduces a transient vibration investigation of horizontal rotor

system with three distinctive disk positions mounted on a shaft, for the simply sup-
ported case. The material properties of the shaft and disk are the same. A transient
analysis is performed to obtain the amplitudes for three distinctive disk positions. A
force is applied on the disk for a small time frame. Design a dynamic rotor struc-
ture; it is essential to decide the vibration parameter, i.e., natural frequency, critical
speeds, and amplitudes. The transient analysis was performed by “ANSYS” para-
metric design tool. The results obtained from the analysis are helpful for the design
of the rotor system.
Keywords Rotor dynamics · Transient · ANSYS · Amplitudes
1 Introduction
Transient vibration is non-occasional motion occurs in a rotating system. Transient

vibration is non-periodic motion in actual circumstances. When a vehicle experiences
a pot opening, the amplitudes diminish gradually because of the stiffness, damping,
and force acting on the system. The amplitude of the system decreases depending on
the system parameters and time. In the transient system, a load is applied for a small
period, and hence, due to this time-dependent loading, the vibration decay gradually.
The vibration energy must be less than the rotational energy of a system. The param-
eters of rotating systems play a vital role in maintaining the stability of a system.
The systems are classified as a linear system and nonlinear system. All systems, in
reality, are a nonlinear system, but a system with small oscillations or amplitudes
D. S. Megharaj
Department of Mechanical Engineering, Islamiah Institute of Technology, Bangalore, Karnataka,
India
A. Malgol (B)
Department of Mechanical Engineering, National Institute of Technology, Calicut, Kerala, India

https://doi.org/10.1007/978-981-15-4488-0_4
34 D. S. Megharaj and A. Malgol
are considered as a linear system. It is essential to decide the dynamic parameters of

a rotor, i.e., natural frequency, whirling speeds, and amplitudes. In a system, there
can be nonlinearity due to geometry, misalignments, loose supports, unbalance mass,
etc. In the design of a rotating system, the disk mass, bearing stiffness, and damping
are important parameters. For a dynamic system, the oscillation gradually reduces
and ultimately vanishes, and such a phenomenon of a dynamic system is known as
transient vibration. Transient vibration may involve free or forced vibrations or both.
The time-dependent transient solution briefs about a variety of sufficiency regarding
time under an action a force for a small period.
The investigation of the transient system is progressively imperative in spinning
machinery generally for high-speed rotors—the two important elements of study in a
dynamic rotor system. The first component is the time-based analysis. The physical
components are crucial, but that cannot be particularly legitimated as an element of
recurrence precedents that are conducted of bearings, seals, and dampers. The second
component characterizes the time-subordinate reaction instances of transient reaction
framework comprise of synchronous engines and blowers and motors. Generally,
a disk, coupling or a disk with blades attached to a shaft such a rotating system
is called as a rotor. Applications of a rotating system are pumps, generators, auto
engines, blowers, steam turbines, etc. Instability in a rotating system is the primary
cause of failure. In a rotating system, changing some parameters, i.e., disk mass,
stiffness, disk position, and different parameter of the rotor system, the amplitudes
of the rotating system can be diminished, and such design modification helps us to
operate in suitable and safe conditions.
The modal analysis and harmonic analysis of the rotor for three distinct positions
of the rotor disk system are performed. In the analysis, natural frequency and critical
speeds of the rotating system increase as the disk position moves toward the support,
as well the amplitudes decrease presented by Malgol and Potdar [1]. Sharama [2]
considered the random form of vibration due to loss of blade and the vibration fre-
quency and demonstrated that nearly at all the working velocities, three-component
control law gives more negative eigenvalues signifying better transient properties.
The stationary and transient motions in rotor systems with higher critical speeds
exceed by the operating speeds. The rotor of a turbo-driven pump assembly is used
in rocket engines where the rated speed of which exceeds the second basic speed
and in which rings serve as sealing. Babukanth and Vimal [3] constructed an FE
model of a rotor in which the shaft segments are considered as beam and considering
the shear deformation. The blade wheels and the seal rings are considered as rigid
disks and elastically attached on a solid bodies. Malcolm [4] described the traditional
analysis approach which sights the dynamic performance of a rotor framework as an
element of recurrence and investigates rotor elements in the time area and drawing
nearer into the non-direct routine. Volokhovskaya and Barmina [6] delineated the
underlying redirection in twisting and leftover unbalances of rotor framework to the
size of the amplitudes of its transient vibrations at normal frequencies underneath its
working turn recurrence on the summary. Szolc et al. [5] study shows a difference
between various dynamic and static characteristics of dynamic, asynchronous motor
and properties of rotor-shaft system. Balakh and Nikiforov [7] considered the effects
Transient Analysis of Rotor System 35
Table 1 Material properties

Sl. No Parameters Value Units
of the rotor
1 Shaft and disk material Mild steel –
2 Young’s modulus (E) 2 × 1011 N/m2
3 Density (ρ) 7800 kg/m3
4 Poisson’s ration (μ) 0.33 –
Table 2 Data of the disk and

Sl. No Parameters Value Units
shaft
1 Diameter of shaft (d) 0.01 m
2 Diameter of disk (D) 0.15 m
3 Thickness of disk (t) 0.01 m
4 Length of shaft (L) 0.4 m
5 Mass of disk (M) 1.3783 kg
of hydrodynamic force due to interaction of rotor and seal lead to disappearance of

critical speed with increasing stiffness. Jangde et al. [8] described the dynamics of
the single rotor system. Fleming et al. [9] investigated the transient response of rotor
system supported with rolling element bearing with internal clearance. The dead band
clearance shows a significant effect on the synchronous rotor response. Annentrout
and Gunter [10] described the characteristics of rotor system with consideration of
squeeze film damper and hydrodynamic effects.
2 Methodology
2.1 ANSYS Modeling
The rotor model consists of a shaft and disk. To develop a model of rotor, Table 1
shows the properties of the rotor, and Table 2 shows the data of disk and rotor which
are considered in ANSYS. Figure 1 shows the model of the rotor.
2.2 Transient Analysis in ANSYS
The transient response of the rotor system is performed, and a 100 N time-dependent
load is applied on the disk for 1 s. Three different load step files are written, i.e., when
there is no load on the rotor system, a load applied for 1 s, and load removed from
the rotor system, performed for cases of different disk position. The time history
post-processor in ANSYS is utilized to draw amplitude versus time plot for rotor
Fig. 1 Rotor model in ANSYS
system by selecting a node on mid of disk, and maximum and minimum amplitudes
are obtained at the maximum time (20 s) and minimum time (0.01 s) for the selected
node. Figures 2a, 3a, and, 4a show the deviation of amplitude concerning the time
of the rotor system for three different positions of a disk mounted on a horizontal
shaft for a simply supported case. Figures 2b, 3b, and, 4b show that the amplitudes of
vibration increase with the natural frequency, and after the first critical speed, again
amplitudes of vibration decrease for three different disk positions for damping ratio
= 0.2.
2.2.1 For the Position of Disk a = 0.2 m and b = 0.2 m
See Fig. 2.
See Fig. 3.
Fig. 2 a Amplitude versus time plot for a = 0.2 m and b = 0.2 m. b Variation of amplitudes of
vibration with operating frequency of rotor system at damping ratio = 0.2
Fig. 3 a Amplitude versus time plot for a = 0.133 m and b = 0.267 m. b Variation of amplitudes
of vibration with operating frequency of rotor system at damping ratio = 0.2
Fig. 4 a Amplitude versus time plot for a = 0.066 m and b = 0.334 m. b Variation of amplitudes
of vibration with operating frequency of rotor system at damping ratio = 0.2
See Fig. 4.
3 Analytical Method
The equation of motion for free undamped vibration is given
mx + kx = 0 (1)
where,
k Stiffness of shaft
m Mass of the disk.
The natural frequency of undamped free vibration

1 k
fn = (2)
2π m
The equivalent stiffness for a rotor with the disk at middle position simply
supported case
48E I
K eq = (3)
L3
where,
E Young’s modulus (MPa)
I Area moment of inertia (m4 )
L Shaft length (m).
The stiffness for a disk with offset or different position is given by
3E I L
K = (4)
a 2 b2
The equation of motion for damped forced vibration
mx + cx + kx = Fo sin(wt) (5)
The amplitudes of vibration (X) are obtained on solving Eq. (5) as

Fo
X = K
2
1 − r 2 + (2ξr)2
where,
ω
r= ωn
Frequency ratio,
fn Natural frequency of the system (Hz),
Fo Applied force on the disk (N),
ω Excitation frequency (Hz),
ξ Damping ratio, and
X Amplitude of forced vibration (m).
Deflection for a rotor with disk offset position for the simply supported case is
given by
W a 2 b2
δ= (6)
3E I L
where,
W Weight of the disk (kg).
Natural frequency for a rotor with disk offset position for the simply supported
case is given by

1 g
fn = (7)
2π δ
0.4987
fn = √ (8)
δ
4 Results
See Table 3.
5 Conclusion
The academic finite element ANSYS tool is good to know the dynamic behavior of
the rotor. Transient investigation of rotor system for three distinct disk positions is
achieved. A load of 100 N is applied on the disk for 1 s. In all the different cases,
we can observe that amplitudes are gradually reduced due to the application of a
load dependent on time, and plots are obtained for variation amplitudes concerning
time, for simply supported rotor system under the action of time-dependent load.
For mid disk position, the stiffness of rotor is less compared to disk location near to
the support, and hence, comparing mid disk position and disk close to the support,
Table 3 Analytical and ANSYS results

Sl. Position of disk Node Result Results
No. ANSYS results Analytical
results
a (m) b (m) Minimum Maximum Maximum
value value value
1 0.2 0.2 Mid Time 0.001 2 2
disk Y 2.080325× 3.98376 × 3.9582 ×
component 10 − 8 10 − 8 10 − 8
of
displacement
2 0.133 0.267 Mid Time 0.001 2 2
disk Y 2.30689 × 3.1676 × 3.1196 ×
component 10 − 8 10 − 8 10 − 8
of
displacement
3 0.066 0.332 Mid Time 0.001 2 2
disk Y 1.07849 × 1.37206 × 1.2022 ×
component 10 − 8 10 − 8 10 − 8
of
displacement
the amplitude of forced vibration is lower for disk closer to the support. Therefore,
the higher the stiffness value lowers the vibration amplitudes. As the disk, location
closer to the support increases stiffness with decreasing amplitudes of vibration.
Further, we can conclude that using a suitable damping ratio, the vibration amplitude
of system reduces, and changing disk position the amplitude further reduces to a
minimum value. These parameters are useful in the design of the rotor system and
finally conclude that the vibration amplitudes depend on the effect of position of the
disk, properties, and geometry of the rotor system. This investigation helps us in a
steady-state and safe operation of the rotating system. Further, these investigations
can be carried out for different materials, multi-disk rotor system, and experimental
verification.
The transient analysis with different disk positions can be utilized for design
modification of a real rotor system such as textile rotor. The textile rotor is used to
wind the band. The gyroscopic effect is considered in the system. The analysis shows
that the amplitude of vibration decreases as the disk position nearer to the support
compared to the mid position of the disk on the shaft.
References
1. Malgol A, Potdar Y (2017) Modelling and simulation of single rotor system. Int J Latest Technol
Eng Manage Appl Sci 6(4):163–169
2. Sharma N, Shekhar S, Dutt JK (2018) Transient excitation suppression capabilities of elec-

tromagnetic actuators in rotor-shaft systems. Struct Health Monit Photogrammetry DIC
6:185–194
3. Babukanth G, Vimal Teja M (2012) Transient analysis of disk brake by using ANSYS software.
Int J Mech Ind Eng 2(1):21–25
4. Malcolm E Time transient analysis and non-linear Rotordynamics, Applied Machinery
Dynamics Co.
5. Szolc T, Konowrocki R, Pisarski D, Pochanke A (2018) Influence of various control strategies on
transient torsional vibrations of rotor-machines driven by asynchronous motors. In: Proceedings
of the 10th international conference on rotor dynamics, IFToMM 2018, 205–220
6. Volokhovskaya OA, Barmina OV (2016) Effect of initial bending and residual eccentricity of
a turbine rotor on its transient vibration amplitudes. J Mach Manuf Reliab 45(2):113–121
7. Balakh LY, Nikiforov AN (2013) The reduction of the vibration level in high-speed rotor
systems by means of floating seal rings. J Mach Manuf Reliab 42(4):276–280
8. Jangde V, Chouksey M, Patil J, Jain V (2015) Modal and Transient analysis of single disc rotor
system. Int J Res Mech Eng Technol 5(2):30–33
9. Fleming DP, Murphy BT, Sawicki JT, Poplawski JV (2006) Transient response of rotor on
rolling element bearing with clearance. In: Seventh international conference on rotor dynamics
sponsored by the international federation for the promotion of mechanism and machine science
(IFTOMM), Vienna, Austria, 25–28 Sept 2006
10. Annentrout RW, Gunter EJ (1999) Transient modal analysis of nonlinear rotor-bearing systems,
RODYN Vibration Analysis, Inc. 1932 Arlington Boulevard Charlottesville, VA 22903-1560
Investigating the Contact Load Capacity
of Asymmetric Helical Gears
P. Dinesh Babu , Swathi Balaji, and P. Marimuthu
Abstract Gears are one of the machine elements used to transmit a torque from one
point of application to another point of application of a mechanical system through the
shaft. In many applications, the power flow is unidirectional. In this case, asymmetric
gears preferred over the symmetric gears. The present work focuses on evaluating
the contact stress through finite element analysis for the symmetric and asymmetric
helical gears. To calculate the stress between the contact points of the gears the
Lagrange multiplier algorithm is applied. Probable increase in the percentage of
contact load capacity of asymmetric helical gear is measured and is duly compared
with values obtained for symmetric helical gear.
Keywords Asymmetric helical gear · Contact stress · Finite element analysis ·

Symmetric helical gear
1 Introduction
Gear is a machine element that transmits movement and power from in between
connecting shafts through progressive engagement teeth and finds its applications
in automobile gearbox and places where the power transmitting distance is shorter.
Gears are also used to vary speed, magnitude and direction and are normally pre-
ferred because of its high transmission efficiency. Originally, symmetric tooth profiles
are proposed and designated for involute gears [1]. But the condition on drive side
is different from that of the coast side during meshing is a well-agreed fact while
designing. The use of asymmetric gears in unidirectional power transmission reduces
contact stress and increases load-carrying capacity per gear size [2, 3] and durability
to the tooth profile. Increasing the contact ratio along with the operating pressure
angles is possible with asymmetric tooth profiles and is beyond the scope of lim-
its offered by conventional gears. When torque is transmitted along any particular
P. Dinesh Babu · S. Balaji · P. Marimuthu (B)

School of Mechanical Engineering, SASTRA Deemed to be University, Thanjavur, India

https://doi.org/10.1007/978-981-15-4488-0_5
46 P. Dinesh Babu et al.
direction, asymmetric gears are more appropriate for those applications. Asymmet-
ric tooth profiled gears also impart tooth stiffness and higher load sharing by also
simultaneously maintaining the preferred pressure angle and contact ratio.
When load is applied on two mating gears the contact stresses arising at the
point of contact of coupling gears change. This leads to failure of gear tooth such
as pitting, surface wear and scuffing on the surface of the tooth [4]. Pitting occurs
due to continuous application load on contact surfaces which happens at the time of
meshing. Stress occurring at the contact point is considered as a limiting factor for
engineers. Francesco and Marini [5] considered the drive side assigned with a low-
pressure angle tooth profile and a high-pressure angle was given for coast side tooth
profile called buttress teeth profile. For symmetric gears with equal pressure angles,
there is decreased bending stress but, no changes in the contact stresses. Yi-Cheng
along with Chung-Biau [6] studied stresses at the contact point for helical gears with
localized bearing contact. Their results give information about the influence of gear
parameters on the distribution of stress. Hwang [4] investigated the contact stresses
developed on mating gear pairs. Their results were calculated at different points of
contact and they showed the variation in contact stress.
The gears design presented in this is different from the ones presented in [5].
Here large pressure angle is provided to drive side profile and low-pressure angle
for coast side involute profile. The load-carrying capacity largely depends on the
pressure angle. Increase in pressure angle increases radii curvature and decreases
Hertzian contact stress [7, 8]. In this paper, contact stresses for asymmetric helical
gear, pair is identified and measured using finite element analysis (FEA) to study
the effect of different pressure angles on drive side and coast side and these values
are compared with symmetric helical gears for same gear parameters. SolidWorks
13 is used to design 3D models gear pairs and FEA software Ansys 14.0 is for the
analysis.
2 Materials and Methods
2.1 Modelling of Gears
With respect to the Cartesian coordinates along x and y directions, the involute of a
circle is defined by the parametric equation
x = r (cost + tsint) (1)
y = r (sint − tcost) (2)
In symmetric gears, tooth profiles are generated form single base circles whereas
in asymmetric gear two base circles are present. One is for coast side involute and
another one is for drive side involute. SolidWorks is utilized for designing 3D gear
Investigating the Contact Load Capacity of Asymmetric … 47
models. Equation driven curves are used for accurate modelling. Coast side involute
curve is drawn from first base circle with pressure angle 20° and drive side involute
from second base circle with pressure angle 35° (Figs. 1 and 2).
Parametric equation for coast side involute is
x = bd1 × 0.5 × (cos(t + π/2 + 2.9 × π/180) + t × sin(t + π/2 + 2.9 × π/180))
(3)
Parametric equation for drive side involute is
y = db2 × 0.5 × (sin(t + π/2 − 1.37 × π/180) − t × cos(t + π/2 + 1.37 × π/180))
(4)
Fig. 1 Asymmetric involute
Fig. 2 Involutes of pressure angle 20◯ and 35◯

Table 1 Asymmetric pinion

Symbol Pinion Gear
and helical gear pair
parameters mn 2.5 mm 2.5 mm
z1 , z2 24 32
αc 20◯ 20◯
αd 35◯ 35◯
B 20◯ 20◯
gw 20 mm 20 mm
mt 2.66044 mm 2.66044 mm
d a1 68.8507 mm 90.1342 mm
d p1 63.8507 mm 85.1342 mm
d b11 60 mm 80 mm
d b22 52.3034 mm 69.7379 mm
dr 1 57.6007 mm 78.8842 mm
pa1 15◯ 11.25◯
2.2 Finite Element Analysis
The models have been developed through SolidWorks modelling software and
imported in finite element solver Ansys 14.0 for the static structural analysis. To
compare the results, same finite element mesh tetrahedral has been used for each
pair of gears. Torque of 59,000 N-mm is applied at the hub of pinion. The material
properties of the pinion and gear material are listed in Tables 1 and 2.
To reduce the computational time and save computer memory, fine mesh is not
recommended for the entire gear model. Only teeth in contact have been meshed with
lesser element sizes to favour obtaining a reasonable accuracy of results (Fig. 3). Face
refinement is done to teeth that are in contact by splitting the faces in Design Modeller.
Mechanical contact between pinion and gear pair is modelled as flexible–flexible and
interference treatment is set to adjust to touch. Friction is neglected at the contact
surfaces and Augmented Lagrange formulation has been chosen for the calculation
(Fig. 4).
The following assumptions have been accepted:
Table 2 Data on properties

Properties Values
of steel
Density 7850 kg/m3
Modulus of elasticity 210 GPa
Poisson’s ratio 0.3
Fatigue 275 MPa
Ultimate tensile strength 690 MPa
Yield strength 410 MPa
Pinion
Gear
Fig. 3 Asymmetric helical gear pair and symmetric helical gear pair
Fig. 4 Finite element model of asymmetric and symmetric helical gear
1. Frictional is neglected in the contact region.

2. Both pinion and gear are made of isotropic and homogeneous material.
The gear assembly is operated under the boundary conditions of the following:
• Gear hub is fixed.
• Remote boundary condition is applied to pinion.
• 59,000 N-mm torque is applied at the hub of pinion.
According to Lewis beam strength equation gear is like a simple cantilever beam
where the root is considered as fixed support and tooth is considered as base. The load
acting at pitch point along the pressure line can be resolved into radial component
and tangential component using a drive side pressure angle. The effect of radial force
is neglected and only the tangential component of force is taken into account.
Solution is done for von Mises stresses at the contact surfaces and fillets.
3 Results and Discussion
The von Mises stress for the asymmetric and symmetric helical gear pair is deter-
mined through finite element analysis. The von Mises stress plot at the contact region
of asymmetric and symmetric helical pair is shown in Figs. 5 and 6. It is observed
that the von Mises stress of the asymmetric helical gear is significantly lower than
that of the symmetric helical gear.
Figures 7 and 8 shows the von Mises plot at the root region of the asymmetric and
symmetric helical pinion. It is noticed that the von Mises of the asymmetric helical
pinion is higher than that of the symmetric helical pinion.
The von Mises stress of asymmetric and symmetric helical gear at the contact
region is compared and it is shown in Fig. 9. It is inferred that the von Mises stress
at the contact region of asymmetric helical gear is 34.92% lower than that of the
symmetric helical gear. An increase in the pressure angle and the load sharing results
to a significant reduction of von Mises stresses at the contact region. For the through-
hardened gears, module is determined based on the contact stresses at the contact
region. Since the von Mises is maximum at the contact region, it is considered for
design criteria to calculate the module. Hence, the contact load capacity can be
improved using asymmetric helical gear drives.
Fig. 5 von Misses stress of asymmetric helical gear at the contact region
Fig. 6 von Misses stress of symmetric helical gear at the contact region
Fig. 7 von Mises stress of asymmetric helical pinion at the root region
Similarly, the von Mises at the root region of the asymmetric and symmetric
helical pinion is compared and it is shown in Fig. 10. It is noticed that the von Mises
stress at the root of asymmetric helical pinion is 33.83% is higher than the symmetric
helical gear pair due to increase in the moment arm. For the case hardened gears the
module is determined based on the stress at the root region.
Fig. 8 von Mises stress of symmetric helical pinion at the root region
Fig. 9 Comparison of von Mises stress of asymmetric and symmetric helical gear pair at the contact
region
4 Conclusion
The following inferences are made based on the quasi-static structural analysis:
1. Asymmetric involute helical gear has been developed using parametric equations.
Fig. 10 Comparison of von Mises stress of asymmetric and symmetric helical pinion at the root
region
2. Larger pressure angle should be considered as a drive side to reduce the contact
stresses.
3. von Mises Stresses are determined through finite element analysis for the
asymmetric and symmetric helical pinion and gear pairs.
4. The maximum von Mises for symmetric and asymmetric helical gear is
544.76 MPa and 354.51 MPa respectively. There is a 34.92% reduction in the
von Misses stress for an asymmetric helical gear compared to symmetric helical
gear. Hence, the contact load capacity can be improved the through hardened
gears. At the same time, the bending load capacity decreases for the asymmetric
helical pinion.
Acknowledgements The authors are grateful to Science and Engineering Research Board
(SERB), Department of Science and Technology, India for providing financial grant (File No.
ECR/2017/000425) to do this work.
References
1. Buckingham E (1963) Analytical mechanics of gears. Dover, New York

2. Kapelevich A (September, 2011 ) Designing asymmetric gears 37–51
3. Kapelevich A (2000) Geometry and design of involute spur gears with asymmetric teeth. Mech
Mach Theory 35:117–130
4. Hwang S-C, Lee J-H, Lee D-H, Han S-H, Lee K-H (2013) Contact stress analysis for a pair of
mating gears. Math and Comp Modelling 57:40–49
5. Di Francesco G, Marini S (1997) Structural analysis of asymmetrical teeth: reduction of size

and weight. Gear Technol 14(5):47–51
6. Yi-Cheng CB (2012) Stress analysis of a helical gear set with localized bearing contact. Finite
Elem Anal Des 38:707–723
7. Marimuthu P, Muthuveerappan G (2016) Design of asymmetric normal contact ratio spur gear
drive through direct design to enhance the load carrying capacity. Mech Mach Theory 95:22–34
8. Marimuthu P, Muthuveerappan G (2016) Investigation of load carrying capacity of asymmetric
high contact ratio spur gear based on load sharing using direct gear design approach. Mech Mach
Theory 96:52–74
Close-Proximity Dynamic Operations
of Spacecraft with Angles-Only
Rendezvous in a Circular Phasing Orbit
Thangavel Sanjeeviraja, R. Santhanakrishnan, S. Lakshmi, and R. Asokan
Abstract Through this paper, it is to assess a closed-loop guidance algorithm for

the range of observability of angles-only by rendezvous position and proximity oper-
ation. The prominence and influence of Clohessy–Wiltshire dynamics (CWD) is an
emerging area which deals with angles-only guidance coupling algorithm and rela-
tive position. Observability analysis of the of rendezvous at a low earth circular orbit,
as opposed to one spacecraft remaining fixed in its orbit, and the implications for
total times are expended. A closed-loop guidance design scheme is generally based
on unscented Kalman filter (UKF) and coupling relationship. The proposed method
is used in the analysis and the following results are obtained for an initial separa-
tion, initial state uncertainties, line-of-sight angles correctness and V (change in
velocity) from the navigation and guidance accuracy. The novelty of the approach is
to minimize the delta-V for close-proximity operation in phasing spacecraft to the
other.
Keywords Rendezvous docking · Proximity operation · Circular phasing ·

Guidance control and Clohessy–Wiltshire dynamics
1 Introduction
Autonomous spacecraft rendezvous plays an important role in many space missions

to enabling technology. In 1967, a successful autonomous spacecraft rendezvous [1]
was tested; since then, bounteous mission proposals have been lunched and attempted
T. Sanjeeviraja (B) · R. Santhanakrishnan · R. Asokan

Department of Aeronautical Engineering, Hindustan University, OMR Padur, Chennai 603103,
India
S. Lakshmi
Department of Electronics and Communication Engineering, Tagore Engineering College,
Chennai 600048, India

https://doi.org/10.1007/978-981-15-4488-0_6
56 T. Sanjeeviraja et al.
by NASA, Soviet Russia, European Space Agency (ESA), Japan Aerospace Explo-
ration Agency (JAXA) and China National Space Association (CNSA), and meth-
ods were assorted for exploring an autonomous rendezvous and docking operation.
In addition, Demonstration of Autonomous Rendezvous Technology (DART) [2],
Orbital Express (OE), [3] and Experimental Small System (XSS) [4] have been
developed by USA to demonstrate the possibility of autonomous rendezvous and
docking technologies. Due to space race, Russian Federal Space Agency’s Soyuz
progress a spacecraft, ESA has designed an Automated Transfer Vehicle (ATV) and
JAXA’s H-II Transfer Vehicle (HTV) [5] and were considered to rendezvous with
the International Space Station (ISS). Recently, CNSA also has conducted several
rendezvous and docking tests with Tiangong-1 (TG-1) target spacecraft since 2011
[6].
Many researchers have shown a keen interest in flight safety during the process of
spacecraft rendezvous and docking, on-orbit tasks, exclusively in the close-proximity
operation to avoid the collision into the target [7]. In the aforementioned research [8],
the authors are strenuous to safety in the rendezvous trajectory planning. Another
practical approach is proposed in Ref. [8], the strategy for initial separation and con-
sidering the measurement uncertainties in a planned trajectory, for optimized colli-
sion avoidance. This case aimed to send separate spacecraft to rendezvous with the
ISS. It allows the station to remain fixed in orbit, incessantly conducting experiments,
while fresh crews and supplies are brought to it, from the ground.
2 Problem of Interest
This investigation will focus on close-range rendezvous in the Hill frame using the
Clohessy–Wiltshire equations and phasing orbits for rendezvous of spacecraft on the
same orbit. Rendezvous is important to crucial objectives such as resupply, in-orbit
spacecraft servicing and repair, contingency planning (e.g. proposed Space Shuttle
“rescue” missions) and in-orbit construction. One of the main literature gaps to which
this research will attempt to contribute is the topic of planning of spacecraft guidance
and control over multiple stages in the presence of uncertainties.
2.1 Solution Method: Relative Motion
This section describes the solution method of relative motion. Figure 1 shows the
RSW (Radial, Along Track and Cross Track) coordinate frame, where R is in line
with the position vector, S is in the direction of the velocity vector but aligned with
the horizontal (perpendicular to the position vector), and W is normal to R and W
(forming a right-handed system); Hill’s equations are a convenient way to express
relative spacecraft dynamics in the RSW frame (see Fig. 1) when two spacecraft are
close together. They are derived, assuming that the satellites are only a few km apart,
Close-Proximity Dynamic Operations of Spacecraft … 57
Fig. 1 RSW coordinate

frame
the frame (origin of the RSW frame) is in a circular orbit and there are no external
forces, such as solar radiation pressure or drag. Figure 1 shows RSW coordinate
frame where x, y and z correspond to r, s and w, respectively, and the equation is as
follows:

x0 2y0 2y0
x(t) = sinωt − 3x0 + cosωt − 4x0 +
ω ω ω

4y0 4x0 2x0
y(t) = 6x0 + sinωt + cos ωt − (6x0 + 3y0 )t + y0 +
ω ω ω
z0
z(t) = z 0 cos ωt + sinωt (1)
ω
y(t) = (6ωx0 + 4y0 )cosωt − 2x0 sinωt − (6x0 + 3y0 )

z(t) = −z 0 ωsinωt + z 0 cosωt (2)
Setting the first three equations equal to zero (x = y=z = 0) and solving the
velocities give the necessary velocities to set the spacecraft, on a trajectory to reach
the origin of the RSW frame (usually centred on a target spacecraft)
for a given time
to rendezvous. Those equations are shown below, where ω = a2μtgt with atgt being
the radius of the circular orbit of the target/origin of the RSW frame. Collectively this
will be called v0 because it is the initial velocity in the RSW frame of the intercept
trajectory. Figure 2 shows that relative motion coordinates for better understanding.
Fig. 2 Represent the coordinates of relative motion
Note that initial velocity in the y-direction must be solved first because initial velocity
in the x-direction is a function of dydt
0
.
−ωx0 (4 − 3cosωt) + 2(1 − cosωt)y0

x0 =
sinωt
(6x0 (ωt − sinωt) − y0 )ωsinωt − 2ωx0 (4 − 3cosωt)(1 − cosωt)
y0 =
(4sinωt − 3ωt)sinωt + 4(1 − cosωt)2
z(t) = −z 0 ωcotωt (3)
2.2 Clohessy–Wiltshire Dynamics
The well-known method of Clohessy–Wiltshire equation, for the relative motion

dynamics, is to the near-circular orbit. The two-body problem assumption made a
space between the target and the chaser. The distance, compared from the target to
the centre of Earth, is very small. The following equations are appropriate in the
relative position of chaser.
f x = ẍ − 2ωz
f y = ÿ + ω2 y
f z = z̈ + 2ωx − 3ω2 z (4)
where ω represents the angular rate of RSW frame. The axis of x, y and z describes
the elements of r(t), f x , f y and the outward forces on the chaser element is f z . The
discrete propagating equation for the system state is X = [r; v]. Equation 4 is utilized
for ignoring the disturbing forces and the pulse manoeuvre scheme, the expression
of propagating equation, for X can be described below
⎡ ⎤
1 0 6{ω(tk − t0 )} 4sin[ω(tk − t0 )] 0 2{1 − cos[ω(tk − t0 )]}
⎢ sin[ω(tk −t0 )] ⎥
⎢ 0 cos[ω(tk − t0 )] 0 0 0 ⎥
⎢ ω ⎥
⎢ −3cos ω(tk − t0 ) 2cos[ω(tk −t0 )]−1 sin[ω(tk −t0 )] ⎥
⎢ 0 0 4 0 ⎥
ϕ(k, 0)⎢
⎢
ω ω
⎥
⎥
(5)
⎢ 0 0 6{ω(tk − t0 )} cos[ω(tk − t0 ) − 3] 0 2sin[ω(tk − t0 )] ⎥
⎢ ⎥
⎣ 0 −ωsin[ω(tk − t0 )] 0 0 cos[ω(tk − t0 )] 0 ⎦
0 0 3ωsin[ω(tk − t0 )] −2sin[ω(tk − t0 )] 0 cos[ω(tk − t0 )]
And, Eq. 2 is re-expressed as,
X k = Φ(k, 0)X 0 + Gv (6)
where G represents the control-driven matrix:

G = Φ(k, 0) 06×3 , I3×3 (7)
3 Observation Equations
In the observation equation, two reasonable assumptions are to be made to establish

the equation. It will make the analysis, easier in the following section. Initially, the
RSW reference frame, of the chaser and target, can be observed, as being parallel in
the setting up of the proximity operation.
In this case, an orbital height 400 km is considered (Altitude of International Space
Station) and the primary distance mark, between the chaser and target, is 10 km, and
the accuracy of orbital phase angle in a deviation will be about 1.472 × 10−3 rad.
Even the poor camera setup. In addition, make the angle deviation will be smaller,
even when the equivalent, when, the chaser should not be far away from the target.
The obtained measurement transformation equation is from chaser to target RSW
frame. Figure 3 shows the relative measurement geometry.

ε −10 ε̇ 0
= + (8)
θ 0 − 1 θ̇ π
The second assumption can be realized by attitude control, when the camera mea-
suring frame is coinciding with the chaser’s RSW frame which has been demonstrated
Fig. 3 Relative
measurement geometry
in PRISMA mission. Then, an observation equation in the target’s RSW frame is to

be obtained.
z
z 1 = e + ve = arctan + ve ;
x2 + y2
z
z 2 = θ + ve = arctan + ve (9)
x 2 + y2
The phase of mid-range proximity for a cooperative and non-cooperative target can
be used in both Eq. 7 (observation equation) and Eq. 8 (measurements transformation
equation), where x, y and z unit vectors are representing the element of the line of
sight, as well as θ and ε are the azimuth angle and the elevation, respectively, vθ
and vε are the measurement noises. It is commonly modelled as zero-mean Gaussian
noise, i.e. v ∼ N(0, σ 2 ) and v ∼ N(0, σ 2 ).
3.1 Observability Analysis
For the circular phasing rendezvous case, we have two spacecraft initially on the same
circular orbit, as shown in Fig. 4. Conceptually, this method transfers the intercepting
spacecraft onto a phasing orbit, with a differing period, than the initial orbit. The
period, of this phasing orbit, is chosen, as such, that alerts the desired number of full
orbits, when the intercepting spacecraft returns to the initial orbit, at the same time,
as the target spacecraft [5], when the target leads the interceptor, the phase angle, is
negative.
Fig. 4 Phasing orbit
When the target spacecraft leads the intercepting spacecraft, the intercepting
spacecraft must “speed up”, and thus must enter an orbit, with the smaller semi-
major axis (faster orbital period) to “catch up” to the target spacecraft. Similarly,
if the target spacecraft is trailing, the intercepting spacecraft must “slow down”
and enter a phasing orbit with the larger semi-major axis (slower orbital period) to
allow the target spacecraft to “catch up”. Mathematically, this is performed using
the equations given below [5]. Note that, this assumes two-impulse transfers, with
instantaneous application of delta-V.
Figure 2 shows the phase angle θ is negative in this case where the target leads
the interceptor, θ is positive, when the interceptor leads the target.
First, we must find the angular velocity of the target spacecraft ωtgt = a2μtgt .
Then, the time of the phasing orbit is given by
ktgt (2π ) + θ
τphase = (10)
ωtgt
where ktgt is the desired number of orbits, the target spacecraft completes before
rendezvous? Using that, the semi-major axis, of the phasing orbit, can be found with
the expression
13
τphase 2
aphase = μ (11)
2π kint
where kint is the desired number of orbits, the intercepting spacecraft completes
before rendezvous? For the analyses in this project, kint = ktgt for simplicity, though,
this certainly needs not to be true. Another important consideration is the radius of
perigee of the phasing orbit when the phasing orbit is smaller than the initial orbit (θ is
Fig. 5 Spacecraft orbits for

circular phasing rendezvous
in the PQ frame
negative). The phasing orbit must not intersect the central body (Earth for this project)
or go too far into the atmosphere. rp = 2aphase − ra with ra = atgt (atgt = ainitial ) for
the intercepting spacecraft. For this project, r p is not allowed to drop lower than
160 km perigee altitude above Earth. Finally, the V required, for the intercepting
satellite, is found by taking the difference of the velocity on the initial orbit and the
velocity on the transfer orbit where it intersects the initial orbit. There will be two
burns of equal magnitude and opposite direction, so the total

2μ μ μ

v = 2 − − (12)
atgt aphase atgt
An example solution using this method is illustrated in Fig. 5. Here, the initial
orbit has a radius of 8.378 * 103 and the target spacecraft is the red spacecraft (solid
orbit line). The target spacecraft has initial true anomaly of 0°, while the intercepting
spacecraft (green point, dotted line) has true anomaly of 315°, giving a phase angle
of −45°. Working through the process above with k = 1 gives the result below,
with total delta-V = 0.6579 km/s and a time to rendezvous of 1.86 h. The designed
closed-loop guidance system is represented in Fig. 6.
4 Rendezvous in Circular Phasing Orbit
Given the orbital radius of the target spacecraft and the coordinates of the intercepting
spacecraft in the RSW frame of the target, the V, required to rendezvous in a desired
amount of time, can be calculated.
Fig. 6 Designed closed-loop guidance model
4.1 Extension
The rendezvous problem in this context is typically posed, as having one spacecraft
rendezvous, with a second passive spacecraft, that does not manoeuvre. In this exten-
sion, the idea, of rendezvous at an intermediate point, will be explored. This also
includes investigating rendezvous locations for more than two spacecraft as well.
These types of investigations are becoming more important, as mission designers
increasingly utilize multi-spacecraft systems, to accomplish mission goals. For exam-
ple, NASA’s Asteroid Redirect Mission requires a manned spacecraft to rendezvous,
with an asteroid sample, that will be placed in lunar orbit [9].
A possible Mars sample return mission would, likely, require some type of Mars
ascent vehicle to rendezvous, with a spacecraft in Mars orbit, before returning to Earth
[10]. Finally, future manned Mars missions may require the coordination/rendezvous
of multiple supply spacecraft, which has been sent to the planet, ahead of time
[11]. There is a growing interest in future space mission concepts which involve
the interaction of multiple spacecraft/satellites [12–16]. In each of these cases, the
multiple spacecraft involved would have the ability to manoeuvre and the finding of
an intermediate rendezvous point may help reduce the demand on a single spacecraft,
that may not have the required V or time available. It may even reduce the total
V or time to rendezvous.
4.2 Circular Phasing Extension
To explore intermediate rendezvous points for circular phasing rendezvous, the same
process, as above, is followed with only minor modification. Here, the “target satel-
lite” is, instead, a point in the initial orbit that, the other spacecraft’s aim to rendezvous
with. This point defined, by its initial true anomaly, is at the start of the simulation
thus evolves with time. The circular phasing method, as described above, is per-
formed for each satellite in the system, across the full range of possible rendezvous
locations, to examine the trends. For the two-satellite case, the minimum time case
(initial true anomaly of target orbit = 240° for k = 4) is animated below, to illustrate
the method.
Setting up the initial true anomaly, of the target orbit equal to the initial true
anomaly of satellite 1, is equivalent to the setting up of the rendezvous point at
satellite 1. Satellite 2 is leading the target orbit (θ is positive), and thus satellite 2
must enter a phasing orbit with greater semi-major axis to rendezvous with satellite
1. Recall that

2μ μ μ

v = 2 − − (13)
atgt aphase atgt
This expression shows, that transfer orbits with smaller required, because, all
the other parameters in the equation are constant. Thus, the leading satellite should
rendezvous with the trailing satellite when using phasing orbits to achieve rendezvous
with two spacecraft at a minimum V cost. Another interesting relation, to note, is
the drop off in V part way, through the plot (at around 200 degree) (Fig. 7).
This occurs because the phase angle switches from negative (trailing the target)
to positive (leading the target). This switch is a result of the phase angle, being the
Fig. 7 Delta versus true anomaly of target location

CW Analytical
Model
Reference Relative
Orbit Dynamics
Guidance Control
Algorithms Algorithms
Plant Model
Fig. 8 Schematic approach of control system
smaller of the two angles between the two position vectors, and changes the phasing
orbit from being smaller than the initial orbit to being larger than the initial orbit. This
larger orbit, however, is not as extreme, a change, as was required for the previous
orbit, so V is required drops (Fig. 8).
Effect of varying target location identifies V which is required for a continuous
rendezvous. There are also trends observed in rendezvous time, plotted below. Note
that, while using one of the satellites as a rendezvous point can minimize V; this
does not save total time, because the second satellite must still complete its phasing
orbits. Figures 9, 10, 11, 12, 13 and 14 represent that, an intermediate rendezvous
orbit is better if minimizing this time is more important than minimizing V. This
time minimizing point occurs, where both satellites are still trailing the target initially
and thus enter phasing orbits with periods, faster than the phasing orbits needed, when
leading the target. Once, one of the satellite transitions leads the target, its rendezvous
time jumps up and leaves behind a minimum.
Note that, this minimum is highly expensive in terms of V, because it is the
scenario with the greatest changes in orbit needed, to achieve rendezvous (the smaller
the phasing orbit, the more V needed to transfer to it).
4.3 Relative Motion Extension
The relative motion rendezvous problem was extended in a similar manner, with the
target, being a selected arbitrary orbit rather than on a specific spacecraft. However,
for this problem, the initial orbits, of the spacecraft, were defined by using the classical
orbital elements and then transformed into the RSW frame [4].
x = δr
Fig. 9 Rendezvous versus true anomaly
Fig. 10 Delta-V versus altitude

Fig. 11 Delta-V versus inclination
Fig. 12 Delta-V versus RAAN

Fig. 13 Delta-V versus true anomaly
Fig. 14 Delta-V versus argument of perigee

y = r (δθ + cosiδΩ)
z = r (sinθi − cosθ sin iδΩ (14)

vr 1 1
x = − δa + − hδΩ . . .
2a r p
δq1 δq2
+ (vr aq1 + hsinθ ) + (vr aq2 − hcosθ )
p p
3vt δq1
y=− δa − vr δθ + (3vt aq1 + 2hcosθ ) ...
2a p
δq2
+ (3vt aq2 + 2hsinθ ) + vr cosiδΩ
p
z = (vt cosθ + vr sinθ )δi + (vt sinθ − vr cosθ ) sin iδΩ (15)
Equations from [5] have been used to convert from orbital elements to RSW
frame. This gives the ability to see the ideal target location in terms of different
orbital elements. In cases with an odd number of satellites, the optimal location,
instead, appears to be on the same orbit, as the middle satellite (middle meaning the
satellite in between the other satellites with respect to whichever orbit element is
being varied). Thus, with an even number of satellites, it may be beneficial to look at
intermediate orbits; depending on the distribution of satellites, there may not be one
on an orbit that would be optimal to target for rendezvous. Of course, this analysis is
only confined to these specific parameters. A more generalized investigation would
give more concrete recommendations. Further, investigation, of the coupling between
varying different parameters, may also be of interest.
5 Conclusion
In this project, some more conventional rendezvous problems, with goals of reaching
a non-manoeuvring spacecraft, have been described. These methods used two-body
dynamics for the circular phasing orbits and dynamics transformed into a relative
motion frame to calculate delta-V, required for rendezvous. We then extended these
same methods to look at rendezvous in intermediate orbits. We showed that for
circular phasing orbits, to minimize delta-V, the leading satellite should rendezvous
with the trailing satellite, if V minimization is a priority and the leading satellite
has sufficient V. However, when minimizing time while still using circular phasing
orbits, there may be a more desirable intermediate rendezvous orbit, depending on
the mission and how important time versus V is. In this case, for t f = 5 min, V 1
= 1.2918 m/s, V 2 = 1.2054 m/s and V total = V 1 + V 2 = 2.4972 m/s.
The extension in the relative frame also gave insight into intermediate rendezvous
locations and indicates that rendezvous orbits that do not coincide with a spacecraft
should be considered in some cases (even number of spacecraft with certain initial
conditions). Analyses, such as these, can be useful for the understanding of how to
balance resource usage across multiple spacecraft, considering the different needs
for each, as opposed to globally optimizing a solution, that may not be optimal for a
single spacecraft. Further research would ideally extend the relative motion analysis
to investigate more general conditions and to ensure that, the initial conditions of
the spacecraft do not exceed the distances allowable by Hill’s equations. An even a
better investigation would be the research optimal control solutions for two or more
spacecraft attempting to rendezvous where the different parameters could be varied,
based on needs of the mission.
References
1. Fehse W (2003) Automated rendezvous and docking of spacecraft. Cambridge University

Press (2003)
2. Rumford TE (2003) Demonstration of autonomous rendezvous technology (DART) project

summary. In: Proceedings of SPIE—The international society for optical engineering, Orlando,
Florida (2003)
3. Stamm S, Motaghedi P (2004) Orbital express capture system: concept to reality. In:
Proceedings of SPIE—The international society for optical engineering 54(19):78–91
4. Davis TM, Melanson D (2004) XSS-10 microsatellite flight demonstration program results.
In: Proceedings of SPIE—The international society for optical engineering, vol 54, no 19, pp
16–25
5. Chullen C, Blome E, Tetsuya S (2011) H-II transfer vehicle (HTV) and the operations concept
for extravehicular activity (EVA) hardware
6. Stone R (2012) A new dawn for China’s space scientists. Science 36(60):1630–1635
7. Luo YZ, Sun ZJ (2017) Safe rendezvous scenario design for geostationary satellites with
collocation constraints. Astronautics 1(1):71–83 (2017)
8. Chu X, Zhang J, Lu S, Zhang Y, Sun Y (2016) Optimised collision avoidance for an ultra-close
rendezvous with a failed satellite based on the Gauss pseudo spectral method. Acta Astronaut
128(1):363–376
9. Shi S, Yang1 L, Lin J, Ren Y, Guo S, Jigui (2018) Measur Sci Technol 29(4)
10. Li JR, Li HY, Tang GJ (2011) Research on the strategy of angles-only relative navigation for
autonomous rendezvous. Sci China Technol Sci 54(7):1865–1872
11. Grzymisch J, Ficher W (2014) Observability criteria and unobservable manoeuvres for in-orbit
bearings-only navigation. J Guidance Control Dyn 37(4):1250–1259
12. Klein I, Geller D (2012) Zero delta-v solution to the angles-only range observability problem
during orbital proximity operations. In: Itzhack Y (ed) Bar-Itzhack memorial symposium,
Haifa, Israel
13. Zhang Y, Huang P, Song K, Meng Z (2019) An angles-only navigation and control scheme for
non-cooperative rendezvous operations. IEEE Trans Ind Electron
14. Geller D, Klein I (2014) Angles-only navigation state observability during orbital proximity
operations. J Guidance Control Dyn 37(6):1976–1983
15. Gasbarri P, Sabatini M, Palmerini G (2014) Ground tests for vision based determination and
control of formation flying spacecraft trajectories. Acta Astronaut 102(1):378–391
16. Gong B, Geller D, Luo J (2015) Initial relative orbit determination analytical error covari-
ance and performance analysis for proximity operations. In: AAS 15-723, 2015 AAS/AIAA
astrodynamics specialist conference, Vail, Colorado USA
Optimal Dimensional Synthesis
of Rhombus Path Generating Adjustable
Four-Bar Mechanism
Ganesan Govindasamy, Rajakumar S. Rai, and Gadudasu Babu Rao
Abstract Precise and continuous generation of rhombus path by adjustable four-bar

mechanism is investigated and confirmed using Reconstructed Adjustable Parameter
Curve (RAPC) technique. RAPC method is able to extract adjustable parameter
value of intermediary path points along the path generation profile corresponding
to uniformly incremented crank angle values. Harmonic Spacing (HS) algorithm is
used to improve the precision points spacing along the perimeter of intended rhombus
path. Optimal dimensional synthesis of adjustable four-bar mechanism makes use
of these precision points. Simulation of motion of adjustable four-bar mechanism to
generate profile path of rhombus is carried out to verify the dimensional synthesis
of the linkage. Optimization combining Genetic Algorithm (GA) and Pattern Search
(PS) is implemented to synthesize rhombus path generating mechanism. Results
show the achievement of precise and continuous path generation with high accuracy
for the selected rhombus path profile.
Keywords Rhombus path · Adjustable Four-bar linkage · Reconstructed

adjustable parameter curve
1 Introduction
Path generation is the motion of coupler point to trace a path profile. Point-to-
point path generation is to follow the computed precision points while the trajectory
between the computed points is approximated. The other type of path generation is of
continuous type which specifies whole path with many points which approximately
pass through the specified curve. Adjustable mechanism is neither equal to sophisti-
cated flexible robot nor equal to non-adjustable mechanism. It takes the mid-position
between non-adjustable mechanism and robot [1] in terms of speed, precision and
G. Govindasamy (B) · R. S. Rai · G. B. Rao

Department of Mechanical Engineering, Karunya Institute of Technology and Sciences, Karunya
Nagar, Coimbatore, Tamil Nadu 641114, India

https://doi.org/10.1007/978-981-15-4488-0_7
72 G. Govindasamy et al.
simplicity. In contrast to analytical methods that satisfy fewer precision points pre-
cisely, optimization methods such as Genetic Algorithm, Pattern Search, and Neural
Networks etc. satisfy many precision points. Haws and Kay [2] used an adjustable
mechanism employing a cam-link to generate one-fourth of a square and a cardioid
path. Kohli and Singh [3] made use of cam-link linkages to generate exact path.
Ogawa et al. [4] synthesized adjustable multi-link linkages to produce L-shaped
and straight lines paths connecting several precision points. Mundo and Gatti [5]
employing non-circular gears synthesized five-bar linkages for accurate path gener-
ation. Mundo et al. [6] synthesized adjustable cam linkage mechanism to adjust the
position of the fixed pivot of link using a cam for exact path generation. Ghoshal
and Chanekar [7] synthesized adjustable linkages in two stages in order to reduce
the variables required to design a linkage.
Synthesis of continuously adjustable crank-rocker linkage to produce rectangular
path was tried by Zhou [8]. Attempt has also been made by Hrishikesh et al. [9]
synthesized adjustable four-bar linkage to generate a rectangle path by improving
the slope continuity of the adjustable link length curve. Ganesan and Sekar [10] used
graphical method to obtain adjustable parameter curve (APC) from the dimensions
of the mechanism synthesized by Zhou [8] to produce rectangular path. Ganesan
and Sekar [11] optimally synthesized continuously adjustable linkages to generate
rectangle with filleted corners. Ajith Kumar et al. [12] synthesized continuously
adjustable linkages to generate rectangle with chamfered corners, rectangle with full
chamfering and one DOF cam-link adjustable mechanism to generate oblong-hole
path.
It has been found from the above literature that the rhombus path generation
mechanism is not attempted. Section 2 presents adjustable linkage model, precision
points on rhombus path, dimensional synthesis using hybrid optimization. Section 3
presents RAPC method to extract path points on the rhombus path for one degree
incremental value of crank for one revolution. Section 4 presents the results of precise
rhombus path generation for various path points. Conclusions and scope for future
work are presented in Sect. 5.
2 Adjustable Linkage Model to Generate Rhombus Path
Zhou’s [8] four-bar adjustable mechanism model is used in the synthesize to generate
precise and continuous rhombus path, as shown in Fig. 1, in which a rocker (L 3 ) is
carried by a slider (D) which forms the adjustable link. This adjustable link length (S)
from the fixed point (A) in the direction of x’-axis of a local co-ordinate frame to the
pin joint between the rocker and slider is called ‘adjustable link length or parameter’
of the mechanism. Link AB is crank (L 1 ). In triad link BCM, side BC forms coupler
link (L 2 ), side BM (L 6 ) carries coupler point (M). In this mechanism, the required
adjustable parameter values (S) are determined so that the coupler point can trace
closed rhombus path precisely when the crank makes one revolution.
Optimal Dimensional Synthesis of Rhombus Path … 73
Fig. 1 Adjustable linkage model
2.1 Precision Points of Rhombus Path
Coupler point has to travel along the four sides of the rhombus path which comprises
two obtuse angles and two acute angles between the sides. Difficulty arises when
the coupler point negotiating the vertices where it changes direction. This demands
closer precision points near the vertices. The following harmonic spacing algorithm
satisfies the closer precision points at the ends of a straight line or an arc.

xi + x f (xi − x f ) ( j − 1)
xj = + cos (1)
2 2 N −1
where x is variable with an interval [xi , x f ] and j = 1, 2, … N, x j -intermediary point

coordinate value, xi -initial point coordinate value, x f -end point coordinate value. N
(number of precision points inclusive of initial point and end point) (Fig. 2).
Fig. 2 Precision points of rhombus path
2.2 Synthesis of Mechanism to Generate Rhombus Path
The coordinates of the precision points are extracted from Eq. (1) and used in the
synthesis as the known values of the coupler points. The mechanism is synthesized
using Eqs. (10)–(21) which govern the geometry of the intended mechanism to satisfy
all the 360 precision points. Synthesized mechanism’s unknown independent param-
eters (φ, x A , y A , β, L e , L 2 , L 3 ) and dependent parameters (L 6 and L 1 ) are given in
Table 1.
The objective is to minimize link length value (S) for a set of independent variables
(φ, x A , y A , β, L e , L 2 , and L 3 ) of the linkage whose lower bound (LB) values are −
3, −2, 2.5, 2.5, 1, 50, −90 and upper bound (UB) values are −2, −1, 4, 4, 2, 80, −
60 respectively for these variables. The difference between the maximum adjustable
link length and minimum adjustable link length is the error function i.e. minimization
Table 1 Parameters of the synthesized adjustable mechanism

Optimized rhombus path generation parameters Values of the variables and other parameters
and other variables
Fixed pivot x-coordinate value (x A ) −2.3554
Fixed pivot y-coordinate value (y A ) −1.4542
Coupler length (L 2 ) 3.5747
Rocker length (L 3 ) 3.8811
Slider offset length (L e ) 1.4641
Coupler point arm angle (β) 62.9637°
Frame orientation angle (θ) −78.6008°
Crank length (L 1 ) 0.8345
Coupler pointer length (L 6 ) 2.8225
Starting crank angle value when pointer is at P1 23.4315
(θ1 )
Maximum value of adjustable link length (Sm ) 6.1869
Minimum value of adjustable link length (Sn ) 5.7168
Error function (E a ) 0.4701
of this error is the objective of the synthesis of the mechanism.
E a = Sm − Sn (2)
Initial value of L 1 and L 6 are 1.0 and 2.5 unit respectively. Penalty factor and
population size are given 400 and 20 respectively and tournament selection method
is used. Objective function must satisfy the following constraint functions (3)–(8) to
satisfy link length criterion of Grashof and also to avoid linkage dead points.
|L ce |max − L 3 ≤ 0. Where L ce = yC − L e (3)

L1 + Sm2 + L 2e − L 2 − L 3 ≤ 0 (4)

L2 − L3 + L1 − Sn2 + L 2e ≤ 0 (5)

L1 − L2 + L3 − sn2 + L 2e ≤ 0 (6)
2
L 22 + L 23 − Sn2 + L 2e − L1 − 2L 2 L 3 cosμ1 ≤ 0 (7)
where μ1 = 20◦ and S n , minimum value.

2
L 22 − L 23 + Sm2 + L 2e + L1 + 2L 2 L 3 cosμ2 ≤ 0 (8)
where μ2 = 160◦ and Sm , maximum value.

In order to trace the path points on the path profile by the pointer in the right order
and overcome ‘order defect’ during full rotation of the crank in the anti-clockwise
direction at uniform angular velocity a constraint Eq. (9) is forced.
(θ1 )i < (θ1 )i+1 (9)
2.3 Optimization Using Genetic Algorithm Followed

by Pattern Search
In Genetic Algorithm (GA) optimum value of the objective function may not be
always reached due to the stopping criteria imposed on GA. This requires further
exploration of search space. By changing the stopping criteria, it is possible to find
the optimum solution. However, it may require many more function evaluations to
reach the optimum value. A local search can be a more efficient in such a scenario.
Pattern search is used for local search that starts from the point where GA reached
[12]. Table 1 shows the parameters of the synthesized adjustable mechanism after
the optimization process.
3 Rhombus Path Generation Using RAPC Method
RAPC method is used to generate rhombus path. The sides of the rhombus path
are inclined to the global x-axis and angles between the sides are acute and obtuse.
The equations required to find the path points for uniform incremental values of the
crank rotation along the four sides of the rhombus are explained in the following
paragraphs. The coordinates of the four vertices of the rhombus P1 , P2 , P3 and P4
are shown in Fig. 1. The closed rhombus path is represented by P1 P2 P3 P4 P1 .
It is anticipated that the coupler point (M) of coupler link in the mechanism
travel through the complete profile path, shown in Fig. 1, through endless path points
for uninterrupted generation of intended path. Path point refers to any point on the
boundary of the rhombus. RAPC method [11] verifies whether the intended rhombus
path generation is achieved for N number of points. Path points on the boundary of
rhombus are obtained from synthesized values (x A , y A , φ, β, L 2 , L 1 , L 3 , L e and L 6 )
of the mechanism. Number of precision points N is taken as 360 for better result.
Precision points used and their corresponding crank angles obtained in the synthesis
do not match to uniform crank arm angular displacements.
However, the dependent and independent variables satisfying the precision points
along the designated profile must also satisfy infinite number of path points for precise
and continuous generation of desired path. The coordinate values of the first precision
point M1 at P1 is (1, 0) and corresponding angular position (θ1 )1 is obtained from
Table 1. Precision point at P1 and initial path point are one and the same. The length
between first path point M1 and fixed point A is given by Eq. (10). Equation (11)
gives the angle between link L 1 and line AM1 .

2 2
L AM(1) = x M(1) − x A + y M(1) − y A (10)
L 21 + L 2AM(1) − L 26
β1(1) = cos−1 (11)
2L 1 L AM(1)
Equation (12) also can be used to get (θ1 )1 . In Eq. (12) the ± sign specifies the
two alignments of dyad ABM as presented in Fig. 1 as continuous and dashed lines.

(θ1 )1 = tan−1 y M(1) − y A , x M(1) − x A ± β1(1) (12)
The Eqs. (13) and (15) give values of coordinates of joint B for first path point.
Equation (14) gives angle (θ6 )1 .
x B(1) = x A + L 1 cos θ1(1) (13)

θ6(1) = tan−1 y M(1) − y B(1) , x M(1) − x B(1) (14)
y B(1) = L 1 sin θ1(1) + y A (15)
The value of β is got from synthesized values (Table 1). The coordinates of joint
C1 are attained from Eqs. (16)–(18)
θ2(1) = θ6(1) − β (16)
xC(1) = L 2 cos θ2(1) + x B(1) (17)
yC(1) = L 2 sin θ2(1) + y B(1) (18)
For ease of analysis of slider dyad a local coordinate frame is used. Joint A is
taken as the local coordinate origin. Coordinates of joint C in local coordinate are
obtained from Eqs. (19) and (20) from the global coordinates.

xC(1) = xC(1) − x A cosϕ + yC(1) − y A sinϕ (19)

yC(1) = − xC(1) − x A sinϕ + yC(1) − y A cosϕ (20)
Location of the slider from fixed the pivot A is given by, adjustable link length,
S(1) when the coupler is indicating the first path point, is found from Eq. (21).

2
S(1) = xC(1) ± L 23 − yC(1) − L e (21)
For the first path point, using Eqs. (10)–(21), adjustable parameter value S(1) is
obtained. The second path point is obtained by intersection of line P1 P2 at M(2) and
an arc of length L 6 from joint B. This gives the coordinates X M(2) and Y M(2) of the
path point M(2) from Eqs. (22)–(24) after incrementing the crank by 1°.
2 2
y M(2) − y B(2) + x M(2) − x B(2) = L 26 (22)
y M(2) = mx M(2) + d, (23)
where ‘m’ is slope, and ‘d’ is intercept.

Substituting y M(2) in Eq. (22) and simplifying gives a quadratic Eq. (24) in x M(2) .
2
ax M(2) + bx M(2) + C = 0 (24)
Fig. 3 Synthesized mechanism to generate rhombus path

where a = (1 + m)2 , b = 2dm − 2my B(2) − 2x B(2) , and C = y B(2) 2
+ x B(2)
2
+ d2 −
2dy B(2) − L 6 )
2
Solving Eq. (24) for x M(2) and substituting in Eq. (22) gives y M(2) in Eq. (25).

y M(2) = y B(2) ± L 26 − (x M(2) − x B(2) )2 (25)
Incrementing crank angle by θ1 and substituting x M(2) , y M(2) through Eqs. (10)–
(21) in the place of x M1 , y M(1) to get the adjustable parameter value S(2) .
Same procedure is followed till all the path points along P1 P2 until the value
of x M(1) value is less than or equal to 0 and y M(1) value is less than or equal to
0.25. Similarly path points for the other sides of the rhombus are obtained using the
required slope and intercept values in the Eqs. (23). Thus the adjustable parameter
values S(1) and corresponding θ1 values of all the path points constitute the RAPC
method. These path points are used to generate the required rhombus path. Figure 3
shows synthesized adjustable mechanism to generate rhombus path.
4 Results and Discussions of Rhombus Path Generation
Figure 4 shows the desired and generated rhombus path for different number of path
points. As mentioned in the previous section, the rhombus path generation displays
deviation from the desired path only at the corner points. Figures 5, and 6 show closer
view of corner point at P2 and P3 respectively.
The deviation from the desired path at P2 is more for 360 path points which
correspond to 1° crank angle increment compared to 720 and 1440 path points which
Fig. 4 Desired and generated rhombus paths
Fig. 5 Closer view of desired and generated rhombus paths at P2
correspond to half and quarter degree crank rotation respectively. The deviation at P3
is the lowest. The decimal values of Fig. 6 validates the accuracy of the RAPC method
proposed in the case of rhombus path generation. Validity of this path generation is
compared and verified with the path generation of corners-chamfered-rectangle [12]
and found correct for precise and continuous rhombus path generation as shown in
Fig. 6. Deviation error from the desired path in terms of area is the lowest at P3
corner compare to P2 and P4 corners of rhombus path.
Fig. 6 Closer view of desired and generated rhombus paths at P3
It is to be noted that the RAPC method comfortably handles both obtuse and acute
angle of the rhombus path at P2 and P3 corners and predicts the deviation with high
accuracy as shown in Figs. 5 and 6 respectively. Figure 7 shows synthesized and
RAP curves for different path points.
5 Conclusions
Precise rhombus-path generation mechanism is synthesized and its precision is veri-

fied using reconstructed adjustable parameter curve method using 360,720 and 1440
path points. Increasing path points further, the desired curve can be attained. RAPC
method is used to verify the results. The results show that the adjustable parameter
curve required to generate rhombus path has sharp turning points which can further
be studied by filleting the two obtuse vertices.
Fig. 7 Synthesized and RAP curves of rhombus path generation
References
1. Chong P, Sodhi RS (2010) Optimal synthesis of adjustable mechanisms generating multiphase

approximate paths. Mech Mach Theory 45:989–996
2. Kay F, Haws R (1975) Adjustable mechanisms for exact path generation. ASME J Eng Ind
97(2):702–707
3. Singh Y, Kohli D (1981) Synthesis of cam-link mechanism for exact path generation. Mech
Mach Theory 16(4):447–457
4. Shimojima H, Ogawa K, Fujiwara A, Sato O (1983) Kinematic synthesis of adjustable
mechanisms. Bull JSME 14:621–632
5. Mundo D, Gatti G (2008) A graphical-analytical technique for the synthesis of noncircular
gears in path-generating geared five-bar mechanisms. Trans CSME 32(3–4):488–497
6. Mundo D, Liu JY, Yan H (2006) Optimal Synthesis of cam-linkage mechanisms for precise
path generations. J Mach Des 128:1253–1260
7. Chanekar PV, Ghoshal A (2013) Optimal synthesis of adjustable planar four-bar crank-rocker
type mechanisms for approximate multi-path generations. Mech Mach Theory 69:263–277
8. Zhou H (2009) Dimensional synthesis of adjustable path generation linkages using the optimal
slider adjustment. Mech Mach Theory 44:1866–1876
9. Hrishikesh Y, Ameya P, Gunesh R (2013) Optimum synthesis of path generating adjustable
mechanisms with improved slope continuity. Trans Can Soc Mech Eng 37(2):197–214
10. Ganesan G, Sekar M (2014) Kinematic analysis of rectangular path generating adjustable
four-bar linkage. Appl Mech Mater Trans Tech Publ 592–594:1094–1098
11. Ganesan G, Sekar M (2017) Optimial synthesis and kinematic analysis of adjustable four-bar
linkages to generate filleted rectangular paths. Mech Based Des Struct Mach 45(3):363–379
12. Ajith Kumar G, Ganesan G, Sekar M (2018) Near perfect path generation of corners-chamfered-
rectangle and single synthesis cam-link mechanism to generate special-slot path. Mech Based
Des Struct Mach 46(4):483–498
Sensor Fusion for Automotive
Dead Reckoning Using GPS and IMU
for Accurate Position and Velocity
Estimation
Lalith Prabu Nalla Perumal and Arockia Selvakumar Arockia Doss
Abstract In recent years, detecting the position of the vehicle is an important task.
The goal of detecting the position of the vehicle can be achieved using Global Posi-
tioning System (GPS) technology by getting approximate position data of the ego-
vehicle. For better and accurate positioning, automotive dead-reckoning can be used.
Automotive dead-reckoning (ADR) technology is a high-end navigation system. In
this paper, an ADR is used to calculate the position based on the distance and direc-
tion of the vehicle traveled from the last known location. This gives an accurate
estimation of the 3-axis position and velocity components based on vehicle data
retrieved from GPS and Inertial Measurement Unit (IMU) sensor. ADR not only
allows full coverage in indoor car parking, tunnels, and underpasses but also effec-
tively eliminates the impact of multipath effects in urban canyons. The vehicle is
incorporated with sensors that record wheel rotation along with the steering direc-
tion and the sensors are allowed to take continuous measurements with the help of
the last known location. The position and velocity shall be corrected even if quality
GPS data is available. In case of partial GPS blockage, the estimation shall be done
with different strategy/weightage factors. In case of complete GPS blockage, the
prediction shall continue for some approximation time, with the help of vehicle and
IMU sensor data. Finally, various experiments are conducted using Kalman filter and
extended Kalman filter, and the prediction results are found to be satisfactory.
Keywords Navigation system · Kalman filter · ADR · PS · IMU · Prediction
1 Introduction
Across the globe, advanced driver assistance system (ADAS) technology is one of
the most and fastest growing technologies. The main purpose for ADAS technol-
ogy is used to make comfort to the driver and reduce the accidents. Corresponding
L. P. Nalla Perumal · A. S. Arockia Doss (B)

School of Mechanical Engineering, Vellore Institute of Technology, Chennai,
Tamil Nadu 600127, India

https://doi.org/10.1007/978-981-15-4488-0_8
84 L. P. Nalla Perumal and A. S. Arockia Doss
to ADAS technology, a high integrity navigation system for vehicle is developed.

Mainly, two types of sensors are required to compute the navigation system. One
is dead-reckoning sensor and another is external sensor. Dead-reckoning sensors
(IMU Sensor) will accumulate more error with respect to time. External sensor will
give an accurate position of ego-vehicle with the help Global Positioning System
(GPS) [1, 2]. In actual time, GPS technology is mostly used for locating and point-
ing out the position and assists for the vehicle. The signals from GPS are so wobbly
because of huge buildings and constructions which will give inaccuracy to indicate
the position. To get the better of these conditions, the GPS system and dead-reckoning
system are integrated [3]. This trustworthy system is used to reduce GPS errors and
guide to evaluate the dead-reckoning system [4]. Integrating with GPS is two differ-
ent types of approaches: one is tightly coupled and another one is loosely coupled.
Loosely coupled integration method is mainly used when any one of the sensors get
failed with lower processing time, and it will get results from another sensor by using
less state vector. The main drawback in tightly couple is that the state vector size and
will increase with large processing time. Similarly, the disadvantage in loose couple
is that it is unworkable to give measurement update from GPS filter where GPS gives
inaccurate signal [5]. Usually, dead-reckoning sensors are not easy to compute, and it
generates errors with respect to time [5]. Basically, IMU sensors are the combination
of accelerometer, gyroscope, and magnetometer and are implemented as the sensor
fusion [6] with Kalman filter (KF) and extended Kalman filter(EKF) of GPS and
IMU [7].
1.1 Kalman Filter
Kalman Filter is an optimal state estimation algorithm and iterative mathematical

process that uses a set of equation and consecutive data input. The KF can estimate
the ego-vehicle position. GPS unit can be equipped with an ego-vehicle. For mini-
mum distance, Kalman will estimate and give high accuracy position, and it follows
the law of physics, by integrating its ego-vehicle velocity by continuously following
the steering wheel angle and wheel revolutions. This type of smart approach is called
dead reckoning, but sometimes it will accumulate small errors due to drift over time.
Basically, KF has two major steps to predict and update. This is the famous technique
for filtering and prediction in linear Gaussian systems. By physics’ law of motion, it
will change the ego-vehicle position with respect to time in prediction step. Similarly,
new covariance and new estimated position will be calculated. As it may be the veloc-
ity of ego-vehicle which is proportional to the covariance, it is found to be doubted
about the precise value of the dead-reckoning estimating position with high velocity
of the ego-vehicle but undoubted about the estimating position of the ego-vehicle
when moving with low velocity. The state transition probability and measurement
probability must be linear and with added Gaussian noise. This probability therefore
assumes linear dynamics and in this form is called linear Gaussian. In update steps,
the measurement of ego-vehicle is observed from GPS unit. The relative covariance
Sensor Fusion for Automotive Dead Reckoning … 85
from prediction step will tell that the new measurement is doubtful or un-doubtful
for updated prediction. Discrete KF is proposed in this work; generally, it is used for
stochastic system.
1.2 Kalman Filter Algorithm with Equations
Figure 1 shows the workflow of each step of the KF algorithm with equations. Figure 2
shows the effects of Kalman gain between measurements and estimation, where
X—State matrix, P—State covariance matrix (represents errors in the estimate),
U—Input matrix, W —Process noise matrix, Q—Process noise covariance matrix,
R—Sensor noise covariance matrix (measurement error), H—Conversion matrix (C
= H), K—Kalman gain.
Prediction Process
Fig. 1 Algorithm flowchart of KF

Fig. 2 Effects of Kalman

gain
X = AX Prev + Bu + Wk (1)
It predicts the estimate of the next state based on the previous state X k−1 and
control input u k . A and B matrices map X k−1 and u k , respectively. It predicts the
covariance matrix of the next estimate. Prediction process of any system is never
perfect. For this reason, some noise is added into the prediction of covariance matrix.
This equation is derived from Riccati equation.
P = A PPrev A T + Q k (2)
Update Process

P × HT
K = (3)
H × P × HT + R
It calculates Kalman gain, a factor which weights the measurement against predic-
tion. A higher Kalman gain will bring the final estimate closer to the measurements
while a lower Kalman gain will make a good estimate for the prediction.
X Prev = X K = X + K [Y − (H × X )] (4)
It updates the state vector. This is the final state estimate that forms the filter.
Using the weightage factor K, the equation modifies the prediction by some portion
of the innovation Y. If Y and HX are equaling system state, and the measurements
are perfectly accurate.
PPrev = Pk = (I − K H )P (5)
It updates the covariance matrix. The update step can give only result in a reduction
of covariance because the update step adds sensor information to the estimate. This
is the reason to subtract the covariance and product of Kalman gain and covariance.
1.3 Extended Kalman Filter Algorithm with Equations
Usually, linear system cannot give accurate prediction results in practically.

Hence using EKF can able to reach the accurate prediction. Figure 3 explains the
workflow of each step about EKF algorithm with equations.
The EKF tranquilizes the linearity assumption by assuming that the state transition
and measurement probabilities are controlled nonlinear functions.
X k = f (X k−1 , u k , wk )
yk = h(X k , vk )
where f predicted state from the previous estimate, h predicted measurement

from the predicted state. The functions of f and h cannot be applied to the covari-
ance directly. Instead a matrix of partial derivatives (the Jacobian) is computed. At
each timestamp, the Jacobian is evaluated with current predicted states. This process
Fig. 3 Mathematical equations of algorithm flowchart of EKF

essentially linearizes the nonlinear function around the current estimate. This is the
major differences between Kalman and EKF. The EKF is additionally using Jacobian
in the predicted states but not in KF, where X—State matrix, F(X, U)—Nonlinear
state transition matrix, P—State covariance matrix (represents errors in the esti-
mate), U—Input matrix, W —Process noise matrix, Q—Process noise covariance
matrix, R—Sensor noise covariance matrix (measurement error), F, H—Jacobian
matrix, K—Kalman gain, Y —Measurement state, Z—Measurement noise low qual-
ity of IMU devices are highly decrease estimating the position when bad GPS signal
occurs.
In autonomous vehicle, unnecessary noises are filtering out for effective position-
ing method [8]. By Kalman and EKF algorithm, performance shows that the com-
parison of unfiltered (measurement) and filtered (Kalman output) data [9]. Using
CarMaker software will create the environment. EKF is used in the filtering process
with vehicle dynamics models. By different scenarios, the simulation environment
will give the different data collected and used to estimate the position of the ego-
vehicle [10]. There are few techniques to find the position of the ego-vehicle when
GPS is not available. Basically, GPS will give in the form of latitude and longitude.
There are few techniques to convert lateral and longitude into position of X and Y
like flat earth and Gauss–Kruger, and it will give good accuracy. Experimental results
explained that the KF and EKF estimation-based algorithm was observed for signal
measurement and correlated with raw data [8] to check whether accuracy is improved
or its again robust to retain positioning output during short term of GPS signal dropout
[11]. From the literature survey, KF and EKF are one of the best methods for estimat-
ing the prediction values for linear system and nonlinear system, respectively. This
algorithm is mainly used for GPS available and unavailable. The proposed system
software implementation is made clear in Sects. 3 and 4, respectively.
2 Proposed System
In this work, KF and EKF algorithms are proposed to estimate and predicting the
positions (Px and Py ), velocity (V ), yaw (ψ). Figure 4 shows the visionary of the pro-
posed work. Receiving the data from IMU and vehicle sensors with the combination
of static and dynamic data, and it will be accumulated to the KF and EKF which gives
static data. Usually, KF can able to predict for linear system, and especially, EKF
can able to predict for nonlinear system. These filters are mainly used to eliminate
noise and predict the states.
3 Implementation
In this research work, the data is developed from CarMaker software and collected
data consign to the model. Figure 5 shows the kinematic model to evaluate the process
Fig. 4 Overview of
methodology of the project
Fig. 5 Process of KFs

Fig. 6 Calculation of
discrete KF
in KF. Figure 6 shows the process flow of EKF, and both the processes are similar;
but in EKF before kinematic model, Jacobian matrix is used to nonlinear to linearize
system.
3.1 Discrete Kalman Filter (DKF) Without Using CarMaker

Data
Figure 7 shows the conversion mathematical equation flowchart of DKF. It has several
steps to follows initially to generate the raw acceleration signal.
3.2 Extended Kalman Filter (EKF) Using CarMaker Data
In EKF, there are two steps: prediction and update. The flowchart of EKF when GPS
available and unavailable is shown in Fig. 8. When GPS is not available, EKF should
take past data values to estimate the position for few seconds. There are some few
techniques available for prediction. One technique is disable the updating part and
estimate the position, and another technique is without disable the updating part and
take the GPS available past data of Px , Py , yaw and measurement velocity. In this
case, it is not require to update velocity because velocity will get vary continuously.
Velocity will entirely depend upon the ego-vehicle drive. Velocity details will be
received from vehicle sensor. Basically, GPS will give lateral and longitude data
of Px , Py, and yaw, not velocity. Hence, when GPS is unavailable, it is possible
Fig. 7 Conversion
mathematical equations
flowchart of DKF
to estimate the position by this technique. This technique will give more accuracy
results than disable update technique.
4 Experimental Results and Discussion
Figures 9 and 10 show the discrete KF output of linear acceleration and static angle
in OCTAVE.
Figures 11, 12 and 13 show the output of EKF in OCTAVE, and here there are four
states and two measurements. The states are position (x, y), velocity, yaw angle, and
the input as acceleration and yaw rate. Hence the Jacobian matrix should be (4 × 4).
In the plots, it has clearly shown that all the three states are predicted properly when
GPS is unavailable. From the experimental results, it is found that the EKF prediction
works satisfactory. X-axis is time for the above graphs. Y-axis for position (x, y) is in
meters, and yaw is in degrees. Here velocity is measurement hence there is no point
to predict the velocity because from the ego-vehicle will get data continuously from
the vehicle sensor. GPS data does not give velocity of the vehicle, and it will give
only position and yaw rate.
Fig. 8 Flowchart of EKF

when GPS available and
unavailable
Fig. 9 Linear acceleration

Fig. 10 Static angle
Fig. 11 Px (Original and

prediction)
5 Conclusion
Based on the study of Automotive dead-reckoning fusion sensor system, the predic-
tion of ego-vehicle position and velocity are obtained irrespective of the availability
of the GPS signal using KF and EKF. Kalman Filter is found to be producing accurate
results for linear system, but the system is considered to be as nonlinear system; due
to this nonlinearity issue, an Extended Kalman Filter is incorporated to predict the
Fig. 12 Py (Original and

prediction)
Fig. 13 Py (Original and

prediction)
position and velocity for nonlinear system. EKF is found to be suitable for nonlinear
system. The experimental results are taken based on three sets of time varying with
8.13 seconds; in the first set of time (0–8.13) seconds, the GPS is available. But in
the second set of time (8.13–16.27) seconds, have unavailability of GPS, and finally
in the third set of time (16.27 to 24.40) seconds, the GPS is available. Based on the
experimental results, it is found that in the first and third sets of time the GPS data
received accurately; but in the second set of time, the EKF will start to predict current
state with the help of past state. In this condition, the EKF is found to be predicted
up to the accuracy level of 95–98% when the GPS is in unavailable time, i.e., when
the sensor data are not received accurately. In the future, the researcher can focus
with the fusion of EKF with flat earth/Gauss–Kruger technique. This technique is
used for conversion process which is used to convert the three-dimensional points to
two-dimensional planes. It is used to calculate the position with the help of latitude
and longitude.
References
1. Li Q, Fang Z, Li H (2004) The application of integrated GPS and dead reckoning positioning
in automotive intelligent navigation system. Positioning 1.08(2004):0
2. Tarfe VS, Arockia Selvakumar A (2016) Path tracing for ground wheeled robot in partially
known environment. Int J Inf Commun Technol 9(3):356–365
3. Dicu N, Andreescu G-D, Gurban EH (2018) Automotive dead-reckoning navigation system
based on vehicle speed and YAW rate. In: 2018 IEEE 12th international symposium on applied
computational intelligence and informatics (SACI). IEEE
4. Zhao L, Ochieng WY, Quddus MA, Noland RB (2003) An extended Kalman filter algorithm
for integrating GPS and low cost dead reckoning system data for vehicle performance and
emissions monitoring. J Navig 56(2):257–275
5. Fakharian A Gustafsson T, Mehrfam M (2011) Adaptive Kalman filtering based navigation:
An IMU/GPS integration approach. In: 2011 International conference on networking, sensing
and control. IEEE
6. Groves PD (2013) Principles of GNSS, inertial, and multisensor integrated navigation systems.
Artech House
7. Kasargod S et al (2014) Vehicle navigation system with dead reckoning. U.S. Patent No.
8,825,397, 2 Sept 2014
8. Jang S et al (2015) A study on integration of particle filter and dead reckoning for efficient
localization of automated guided vehicles. In: 2015 IEEE international symposium on robotics
and intelligent sensors (IRIS)
9. Romaniuk S, Gosiewski Z (2014) Kalman filter realization for orientation and position
estimation on dedicated processor. Acta Mechanica et Automatica 8.2(2014):88–94
10. Lanfelt C, Rogenfelt A. Dead Reckoning during safe stop of autonomous vehicles.
11. Zhang M, Liu K, Li C (2016) Unmanned ground vehicle positioning system by GPS/dead-
reckoning/IMU sensor fusion. In: 2nd Annual international conference on electronics, electrical
engineering and information science (EEEIS 2016). Atlantis Press
Optimal Dimensional Synthesis
of Double-Semi-circle
and Double-Semi-ellipse Path Generating
Adjustable Four-Bar Mechanisms
Ganesan Govindasamy, Gadudasu Babu Rao, and Rajakumar S. Rai
Abstract Synthesis of adjustable four-bar mechanisms to generate continuous and

precise path of double-semi-circle-straight-line (DSCSL) and double-semi-ellipse-
straight-line (DSESL) path generation is investigated and verified. Harmonic spacing
(HS) algorithm is used to improve precision points spacing along the intended profile
paths. Bi-cubic interpolation is used to extract adjustable parameter value of inter-
mediary path points along the profile of path generation conforming to uniformly
incremented crank rotation value. Precision points are used in the synthesis to get
dimensional values of link lengths and other parameters of adjustable mechanisms.
Motion simulation of adjustable mechanism to generate intended coupler path pro-
files is carried out to verify the dimensional synthesis of the mechanisms. Pattern
search (PS) is implemented in the synthesis of mechanism from the results of genetic
algorithm (GA) as a hybrid method. Results demonstrate continuous and precise path
generation of selected path profiles.
Keywords Double-semi-circle-straight-line path generation ·

Double-semi-ellipse-straight-line path generation · Adjustable path generation
mechanism
1 Introduction
Path generation is the motion of coupler point to trace a path profile. Point-to-point
path generation is to follow the computed precision points while the trajectory
between the computed points is approximated. The other type of path generation
is of continuous type which specifies whole path with many points which approx-
imately pass through the specified curve. Adjustable mechanism is neither equal
to sophisticated flexible robot nor equal to non-adjustable mechanism. Adjustable
mechanism takes the middle ground between robot and non-adjustable mechanism
G. Govindasamy (B) · G. B. Rao · R. S. Rai

Nagar, Coimbatore 641114, India

https://doi.org/10.1007/978-981-15-4488-0_9
[1]. The characteristics of non-adjustable mechanism such as precision, high speed,

and simplicity and some of the abilities of flexible multi-axis robots are similarly
attained in adjustable mechanisms. In contrast to analytical methods that satisfy
fewer precision points precisely, optimization methods like genetic algorithm, pat-
tern search, neural networks, etc., satisfy many precision points. Haws and Kay [2]
used an adjustable mechanism employing a cam-link to generate one-fourth of a
square and a cardioid path. Kohli and Singh [3] made use of cam-link linkages to
generate exact path. Ogawa et al. [4] synthesized adjustable four-bar multi-link link-
ages to generate L-shaped and straight lines paths passing through several precision
points. Mundo and Gatti [5] employing non-circular gears in between two input links
synthesized optimally five-bar linkages for precise path generation. Mundo et al. [6]
synthesized adjustable mechanism with a cam-link to adjust the location of fixed
pivot of the driven side link for exact path generation. Ghoshal and Chanekar [7]
synthesized adjustable four-bar linkages in two stages in order to reduce number of
variables in the linkage design.
Synthesis of continuously adjustable four-bar crank-rocker linkage to generate
rectangle path was attempted by Zhou [8]. Attempt has also been made by Hrishikesh
et al. [9] synthesized adjustable four-bar linkage to generate a rectangle path by
improving the slope continuity of the adjustable link length curve. Ganesan and
Sekar [10] used graphical method to obtain adjustable parameter curve (APC) from
the dimensions of the mechanism synthesized by Zhou [8] to produce rectangular
path. Ganesan and Sekar [11] optimally synthesized adjustable four-bar linkages
to generate rectangular path with filleted corners. Ajith Kumar et al. [12] synthe-
sized continuously adjustable four-bar mechanisms to generate corners-chamfered-
rectangle, fully-chamfered-rectangle, and rectangle employing RAPC method to
design an adjustable four-bar cam-linkage mechanism with one DOF to generate
oblong-hole path.
It has been found from the literature that the DSCSL and DSESL path generation
mechanisms are not investigated. Section 2 presents adjustable linkage model, pre-
cision points on DSCSL and DSESL path, and dimensional synthesis using hybrid
optimization. Section 3 presents bi-cubic interpolation of adjustable parameter val-
ues of precision points and their corresponding crank angles to extract path points
along the DSCSL and DSESL paths for one degree increment of crank rotation for
one revolution. Section 4 presents the results of precise DSCSL and DSESL path gen-
eration for various path points. Conclusions and scope for future work are presented
in Sect. 5.
2 Adjustable Mechanism Model to Generate DSCSL

and DSESL Paths
Zhou’s [8] four-bar adjustable mechanism model is used in the synthesis to generate
precise and continuous DSCSL and DSESL paths, as shown in Fig. 1 for DSCSL
Optimal Dimensional Synthesis of Double-Semi-circle … 99
Fig. 1 Adjustable linkage model
path, in which a rocker (L 3 ) is carried by a slider (D) which forms the adjustable link.
This adjustable link length (S) from the fixed point (A) in the direction of x’-axis
of a local co-ordinate frame to the pin joint between the rocker and slider is called
‘adjustable link length or parameter’ of the mechanism. Link AB is crank (L 1 ). In
triad link BCM, side BC forms coupler link (L 2 ), and side BM (L 6 ) carries coupler
point (M). In this adjustable four-bar mechanism, the required adjustable parameter
values (S) are determined so that the coupler point can trace DSESL and DSCSL
paths precisely when the crank makes one revolution.
2.1 Precision Points Algorithm for DSCSL and DSESL Paths
Coupler point has to travel along the P1 P2 P3 P4 P5 P1 path which comprises two
inverted semi-circles joined by a straight line as shown in Fig. 1 and DSESL path,
which comprises two inverted semi-ellipses, with semi-major axis is twice the semi-
minor axis, joined by a straight line.
Difficulty arises when the coupler point negotiating P1 , P3, and P5 positions where
it changes direction drastically. This demands closer precision points near these
points. The following harmonic spacing algorithm (1) satisfies the closer precision
points at the ends of a straight line or an arc [11] as shown in Fig. 2.
0.5
0.2
0
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
-0.5 -0.3
Fig. 2 Precision points of DSCST and DSEST paths

xi + x f (xi − x f ) ( j − 1F)
xj = + cos (1)
2 2 N −1
where x is variable with an interval [xi , x f ] and j = 1, 2, … N, x j —intermediary point

coordinate value, xi —initial point coordinate value, and x f —end point coordinate
value. N—number of precision points inclusive of initial point and end point.
2.2 Mechanism Synthesis of to Generate DSCSL and DSESL

Path
synthesis as known values of the coupler points. The mechanism is synthesized
using Eqs. (10)–(21) which govern the geometry of the intended mechanism to sat-
isfy all the 360 precision points. The unknown parameters (x A , y A , φ, β, L 2 , L 3 , L e )
and dependent parameters (L 1 and L 6 ) of the synthesized mechanisms are given in
Table 1.
The objective is to minimize adjustable link length value ‘S’ for a set of indepen-
dent variables (x A , y A , L 2 , L 3 , L e , β and φ) of the adjustable linkage whose lower
bound (LB) values are −2.5, −2, 2.2, 3, 1.1, 50, −90 and upper bound (UB) values
are −1.5, −1.2, 3.5, 4, 70, −70, respectively, for these variables for DSCSL path.
The adjustable link length range forms the error function, i.e., minimization of this
error is the objective of the synthesis of the mechanism.
E = Sm − Sn (2)
Initial values of L 1 and L 6 are 1.0 and 2.5 units, respectively. Penalty factor,
population size, and selection method used are 400, 20, and tournament, respectively.
Objective function must satisfy the following constraint functions (3)–(8) to satisfy
link length criterion of Grashof and to evade dead points of the linkage.
|L ce |max − L 3 ≤ 0 Where L ce = yC − L e (3)

L1 + Sm2 + L 2e − L 2 − L 3 ≤ 0 (4)
Table 1 Parameters of the synthesized adjustable mechanism

Optimized parameters and other variables of the mechanisms Values of the variables
DSCSL path DSESL path
Fixed pivot x-coordinate value (x A ) −2.4989 −2.3567
Fixed pivot y-coordinate value (y A ) −1.6616 −1.1997
Coupler length (L 2 ) 3.4999 3.9949
Rocker length (L 3 ) 3.7406 4.9681
Slider offset length (L e ) 1.7253 1.3351
Coupler point arm angle (β) 65.1044° 54.6772°
Frame orientation angle (θ) −70.6490° −64.5652°
Crank length (L 1 ) 0.9145 0.8968
Coupler pointer length (L 6 ) 3.0139 2.6756
Starting crank angle value when pointer is at P1 (θ1 ) 7.8913 13.1390
Maximum value of adjustable link length (Sm ) 6.5365 8.2772
Minimum value of adjustable link length (Sn ) 5.9345 7.3500
Error function (E a ) 0.6020 0.9272

L1 + L2 − L3 − Sn2 + L 2e ≤ 0 (5)

L1 + L3 − L2 − sn2 + L 2e ≤ 0 (6)
2
L 22 + L 23 − Sn2 + L 2e − L 1 − 2L 2 L 3 cosμ1 ≤ 0 (7)
where μ1 = 20◦ and Sn (minimum value)

2
L 22 − L 23 + Sm2 + L 2e + L1 + 2L 2 L 3 cosμ2 ≤ 0 (8)
where μ2 = 160◦ and Sm (maximum value).

In order to trace the path points on the path profile by the coupler point in the
correct order and overcome ‘order defect’ when the crank is rotated in the anti-
clockwise direction at uniform angular velocity, an additional constraint Eq. (9) is
forced.
(θ1 )i < (θ1 )i+1 (9)

Apply GA Apply PS to fine-

Define objective
and, constraint for exploring tune the result of GA to
Start functions. Define the solution explore the End
variables’ bounds space Solution space
Fig. 3 Hybrid optimization process flowchart
2.3 Optimization Using Genetic Algorithm and Pattern

Search
In genetic algorithm (GA), optimum value of the objective function may not be
always reached due to the stopping criteria imposed on GA [13]. This requires further
exploration of search space. By changing the stopping criteria, it is possible to find the
optimum solution. However, it may require many more function evaluations to reach
the optimum value. A local search can be a more efficient in such a scenario. Pattern
search [14] is used for local search that starts from the point where GA reached thus
making a hybrid optimization process (Fig. 3).
3 Bi-cubic Interpolation Method to Generate Path Points
Bi-cubic interpolation method is used to generate intermediate path points of DSCSL

and DSESL paths. The coordinates of P1 , P2 , P3 P4 and P5 of DSCSL geometry are
shown in Fig. 1. Coupler path is represented by P1 P2 P3 P4 P5 P1 .
It is anticipated that the coupler point (M) of a four-bar adjustable mechanism
travels through the complete profile path, shown in Fig. 1, through endless number
of path points for uninterrupted path generation. Here, precision point refers to the
points obtained by harmonic spacing on the perimeter of the DSCSL. Precision points
on the DSCSL path are obtained from the values of x A , y A , φ, β, L 2 , L 3 , L e , L 1 and
L 6 of the mechanism synthesized. The number of precision points (N) is taken as
360 for better results.
Precision points used and their corresponding crank angles obtained in the syn-
thesis do not match to uniform crank arm angular intervals. Coordinate value of the
first precision point M1 at P1 is (1, 0) and corresponding angular position (θ1 )1 is
obtained from Table 1. The length between fixed point A and first path point M1 is
obtained from Eq. (10). From Eq. (11), the angle between line AM1 and link length
L 1 is obtained.
2 2
L AM(1) = x M(1) − x A + y M(1) − y A (10)
L 21 + L 2AM(1) − L 26
β1(1) = cos−1 (11)
2L 1 L AM(1)
Equation (12) also can be used to get (θ1 )1 . In Eq. (12), the ± sign specifies the

(θ1 )1 = tan−1 y M(1) − y A , x M(1) − x A ± β1(1) (12)
Equations (13) and (15) give values of coordinates of joint B for first path point.
Angle (θ6 )1 . is given by Eq. (14).
x B(1) = x A + L 1 cos θ1(1) (13)

θ6(1) = tan−1 y M(1) − y B(1) , x M(1) − x B(1) (14)
y B(1) = y A + L 1 sin θ1(1) (15)
The value of β is obtained from synthesized values (Table 1). The coordinates of
joint C1 are attained from Eqs. (16)–(18)
θ2(1) = θ6(1) − β (16)
xC(1) = x B(1) + L 2 cosθ2(1) (17)
yC(1) = y B(1) + L 2 sin θ2(1) (18)
For ease of analysis of slider dyad, a local coordinate frame is used. Joint A is
taken as the origin of the local coordinate. Coordinates of joint C in local coordinate
are obtained from Eqs. (19) and (20) from the global coordinates.


Location of the slider from fixed the pivot A is given by, adjustable link length,S(1)
when the coupler is indicating the first path point, is found from Eq. (21).

2
S(1) = xC(1) ± L 23 − yC(1) − L e (21)
Equations (10)–(21) give adjustable parameter S(1) value. Similarly, APV for all
the precision points and corresponding angles is obtained from optimal synthesis
process. Thus, S(i) where i = 1, 2, …, 360, adjustable parameter values of precision
Position of Slider pin from fixed pivot vs. Crank Angle
6.5
6.4
6.3
Adjustable Link Length
6.2
6.1
5.9
5.8
Double Semi-Circle Straight Line path generation
5.7
5.6
100 150 200 250 300 350 400
Crank Angle(degree)
Fig. 4 Adjustable link length (S) versus crank angle for DSCSL path
points are obtained. But the difference between θ 2 values of successive precision
points is not uniform. A bi-cubic interpolation method is used to get path point for
incremental crank angle (1°) values from 360 precision points and corresponding
crank arm positions (θ 2 ) using MATLAB bi-cubic interpolation. Figure 4 shows
adjustable link length (S) versus crank angle incremented by 1° for DSCSL path
generation.
Synthesized adjustable mechanism to generate DSCSL path, employing bi-cubic
interpolated path points in the animation, is shown in Fig. 5. Similar approach is used
to synthesize adjustable mechanism shown in Fig. 6 to generate DSESL path.
4 Results and Discussion of DSCSL and DSESL Path

Generation
Figure 7 shows the desired and generated DSCSL path for different number of path
points. In the previous section, it is mentioned, the DSCSL path generation also
displays desired path deviation at the corner point of P5 . Figure 8 show the closer
view at transition point from curve to straight line at P5 .
The deviation from the desired path at P5 is more for 360 path points which
correspond to 1° crank increment compared to 720 (half degree crank increment)
path points. Similar deviation is also observed at curved portions of the generated
Adjustable Four-bar Mechanism to Generate Double Semi-Circle Straight Line Path
2.5
1.5
1
Length Units
0.5
-0.5
-1
-1.5
-2
-3 -2 -1 0 1 2 3 4
Length Units
Fig. 5 Synthesized mechanism to generate DSCSL path
Adjustable Four-bar Mechanism to generate Double Semi-Ellipse Straight Line Path
1
Length Units
-1
-2
-3 -2 -1 0 1 2 3 4 5 6
Length Units
Fig. 6 Synthesized mechanism to generate DSESL path
path. It is clear from Fig. 8 that when the number of path point increases, generated
path approaches desired DSCSL path. The decimal values of Fig. 8 validate the
accuracy of the bi-cubic interpolation method proposed in the case of DSCSL path
generation. Validity of this path generation is compared and verified with the path
generation of filleted rectangular path [11] and corners-chamfered-rectangle [12] and
found correct for continuous and precise path generation of DSCSL as depicted in
Double Semi-Circle Straight Line Path Generation
0.6 Desired DSCSL Path

Interpolated: 360 Path points
0.4 Interpolated: 720 Path points
Y - axis length units
0.2
-0.2
-0.4
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

X - axis length units
Fig. 7 Desired and generated DSCSL path
-3
x 10 Double Semi-Circle Straight Line Path Generation
3
Desired DSCSL Path
2 Interpolated: 360 Path points
Interpolated: 720 Path points
1
0
-1
-2
-3
-4
-5
-6
-7
-1.002 -1.001 -1 -0.999 -0.998 -0.997 -0.996 -0.995
X - axis length units
Fig. 8 Closer view of generated DSCSL path at P5 point

Figs. 7 and 8. In Fig. 8, a closer view of P5 point is given. Generated path approaches
perfect DSCSL path when the number of path points is increased. Similarly DSESL
path is verified but not presented for brevity of the paper.
5 Conclusions
Precise DSCSL and DSESL path generation mechanisms are synthesized and their
precise path generation is verified using bi-cubic interpolation method using 360 and
720 path points. Increasing path points further, the desired curve can be approached.
Bi-cubic interpolation method requires less computational time than RAPC method.
However, RAPC method is superior to bi-cubic interpolation method since former
uses synthesized trajectory on which path points are obtained. Verification of precise
path generation for DSCSL and DSESL using RAPC method is suggested as future
work.
References
1. Chong P, Sodhi RS (2010) Optimal synthesis of adjustable mechanisms generating multiphase

2. Kay F, Haws R (1975) Adjustable mechansims for exact path generation. ASME J Eng Ind
97(2):702–707
6. Mundo D, Liu JY, Yan H (2006) Optimal synthesis of cam-linkage mechanisms for precise
9. Hrishikesh Y, Ameya P, Gunesh R (2013) Optimum synthesis of path generating adjustable
mechanisms with improved slope continuity. Trans Can Soc Mech Eng 37(2):197–214
13. Goldberg DE (2009) Genetic algorithms in search, optimization, and machine learning. 4th
impression. Pearson Education, New Delhi
14. Conn AR, Goiuld NI, Toint PL (1997) A globally convergent augmented Lagrangian Barrier
algorithm for optimization with general inequality constraints and simple bounds. Math Comput
66:261–288
Effect of Interference on a Floating Axis
Epitrochoidal Hydrostatic Unit
Harish Ramachandran, Aditya Shriwastava, and Abhijit Nag
Abstract In the present work, the effect of inclusion of interference on the perfor-
mance of an Epitrochoidal Hydrostatic Rotary Piston Machine, namely Orbit motor
is studied by building its mathematical model and implementing the algorithm in
MATLAB. In this approach, the rotor is rigid, rollers are elastic and interference
between the rotor and the stator is provided by changing the roller radius, chordal
thickness and the pitch circle radius. In an Orbit motor of interference-fit type, the
contact points and the rotor center deviate from their original position as found in an
Orbit motor of perfect-fit type. A corrective technique based on minimum potential
energy of the system is used to obtain the rotor center of an interference-fit motor,
starting from the geometrically obtained rotor center of a perfect-fit motor, correcting
itself to its final position. For various positions of the output shaft, the forces and
torque acting on the rotor are calculated. For positions other than those corresponding
to that of maximum compression and maximum expansion of chambers, it is found
that a net unbalanced torque acts on the rotor. This torque tends to bring the rotor
back to the nearest position of maximum compression. Apart from the determina-
tion of the unbalanced torque, the variation of deformation and maximum contact
pressure at the contact points between the rotor and the stator due to sole inclusion
of interference, for the first two phases of the rotor rotation is also investigated.
Keywords Orbit motor · Interference-fit · Unbalanced torque · Contact

deformation · Maximum contact pressure
Nomenclature
[ei ] Position vector of the ith point on the periphery of the rotor in frame
[ei ] Position vector of the ith point on the periphery of the rotor in frame
F Resultant force acting on the rotor
H. Ramachandran (B) · A. Shriwastava · A. Nag

Birla Institute of Technology, Mesra, Ranchi, Jharkhand 835215, India

https://doi.org/10.1007/978-981-15-4488-0_10
110 H. Ramachandran et al.
Potential energy of the system due to deformation at the contacts

δj Magnitude of deformation of the jth roller
fj Reaction force acting on the rotor due to deformation of the jth roller
ξ Angular displacement of the output shaft in counter-clockwise direction
ξ0 Phase angle, ξ0 = Z (Zπ−1)
Ki Center of the ith roller for 1 ≤ i ≤ Z . Also known as the crunode
Ni Contact point of the rotor with the ith roller for 1 ≤ i ≤ Z for a perfect-fit
motor
Ni Contact point of the rotor with the ith roller for 1 ≤ i ≤ Z , for an interference-
fit motor
Ni Estimated point of contact of the rotor with the ith roller for 1 ≤ i ≤ Z , for
an interference-fit motor
S Stator center
O Rotor center of an interference-fit motor
O Rotor center of a perfect-fit motor
O Estimated rotor center of an interference-fit motor
r Radius of the inner centrode
R Radius of the outer centrode
A Distance of epitrochoid generating point from the center of the outer centrode
or radius of a circle passing through the roller centers
rm Amount of constant difference modification (i.e. the parallel shifting of the
profile), also equal to the roller radius
k Stiffness coefficient of the roller
c Center to center distance between the rotor and the stator
Ci ith Chamber
Z Number of rollers
r m Amount of change in the roller radius
ct Amount of change in the chordal thickness
pcr Amount of change in the pitch circle radius
n Number of points in which periphery of the rotor is divided
Amount of shift of O in each computational step
T Net torque acting on the rotor
Coordinate frame {X Y Z} attached to the stator with center at point ‘S’
Coordinate frame {x y z} attached to the rotor with center at point ‘O ’
i Initial potential energy of the system
f Final potential energy of the system
ξi Initial position of the output shaft
ξf Final position of the output shaft
ξ Increment in the output shaft rotation
w Width of the rotor
Effect of Interference on a Floating Axis … 111
1 Introduction
Orbit motor is a Low Speed High Torque (LSHT) hydraulic motor capable of pro-
viding high torque to inertia ratio. The heart of this motor comprises of a rotor and
a stator, where the stator consists of ‘Z’ equispaced rollers. The difference between
the number of rollers and the number of lobes on the rotor is always unity. The geo-
metric design technique for the epitrochoid generated rotor and stator profile was
first developed by Ansdale and Lockley [1] while working on the design of Wankel
engine and similar Rotary Piston Machines (ROPIMAs). Investigations on epitro-
choidal profile useful in ROPIMAs were carried out in more detail by Colbourne
[2, 3] where he described a unique method for finding the envelope of the trochoids
which perform the planetary motion. Robinson and Lyon [4] developed the actual
useful profile in ROPIMAs, which is a constant difference shifted profile known as
the constant difference modified epitrochoid.
In the geometrically form-closed rotor-stator set, all the rollers are in contact
with the rotor creating closed volumes known as the chambers. These chambers
vary in size as the rotor rotates about its orbiting center (i.e. the rotor is floating and
performs epicyclic motion about the stator center). When pressurized fluid is supplied
to certain predetermined chambers [5], some rollers experience deformation, while in
others gaps are generated. If gaps are generated at the transition contacts (i.e. contacts
separating the High-Pressure Zone (HPZ) from the Low-Pressure Zone (LPZ)), there
is a reduction in pressure differential of fluid due to leakage from HPZ to LPZ [6,
7]. This leakage will reduce the efficiency of the motor. In an Orbit motor, profile
of both the rotor and the stator are the working profile, thus external sealing is not
possible at the contacts. A feasible solution to eliminate this inter-chamber leakage is
by introducing interference between the rotor and the stator. Three popular methods
of introducing interference [8] are by modifying the following:
1. Roller radius (Fig. 1a).
2. Chordal thickness (Fig. 1b).
3. Pitch circle radius (Fig. 1c).
Fig. 1 a Interference provided by an increase in roller radius. b Interference provided by increase

in chordal thickness. c Interference provided by decrease in pitch circle diameter
In this study, interference is provided between the rotor and the stator by a com-
bination of all the three methods mentioned above and all the chambers of the motor
are at zero gauge pressure to investigate the effect of sole inclusion of interference.
Experimentally, it has been observed that in an Orbit motor of interference-fit type,
the rotor is stable only at certain discrete positions. When the rotor is disturbed from
these positions, it returns to the nearest stable position. But, there is no mathematical
justification available in open literature to explain this experimental observation.
This paper provides a mathematical model to determine the effect of interfer-
ence on the performance of Orbit motor. The model justifies the above-mentioned
experimental observation by calculating the unbalanced torque acting on the rotor
for various positions of output shaft. Apart from the unbalanced torque variation,
the variation of roller deformation and maximum contact pressure with output shaft
rotation due to the sole presence of interference is also determined.
The remainder of the paper is subdivided as follows: The next section presents
the underlying assumptions considered while formulating the model and a general
overview of the algorithm. Section 3 delineates each key aspect of the algorithm.
Section 4 presents the results on the effect of introducing interference and finally,
the conclusions of the investigation are drawn in Sect. 5.
2 Perfect-Fit Versus Interference-Fit
For an Orbit motor of perfect-fit type, when all the chambers are at zero gauge
pressure, there is neither deformation nor gap generated at the contacts. This type of
motor is purely hypothetical as there will always be some amount of manufacturing
or assembly error. But, in an Orbit motor of interference-fit type, even when all the
chambers are at zero gauge pressure, there is some deformation at all the contacts. As
a consequence, there will be a reaction force exerted by each roller on the rotor. Due to
these reaction forces, in general, there also will be an unbalanced torque acting on the
rotor. The present investigation aims at finding the nature of this unbalanced torque
acting on the rotor, deformation of the rollers and the maximum contact pressure for
various positions of the rotor.
The initial position and position of the rotor when the output shaft is rotated by ξ
is shown in Fig. 2a and b respectively.
Let ξ be the angle by which the rotor rotates from its initial position in counter-
clockwise direction. To calculate the unbalanced torque acting on the rotor, for a
particular ξ , an external torque is applied on the rotor, such that rotor is in static
equilibrium. This externally applied torque has the same magnitude as that of the
unbalanced torque but acts in the opposite direction.
In this investigation, the following assumptions are taken into consideration while
formulating the approach mentioned in the upcoming section:
Fig. 2 a Initial position of the rotor. b Position of the rotor when output shaft is rotated by an angle
ξ
1. Rotor is a rigid body.

2. Rollers are elastic and exert reaction force on the rotor which is linearly
proportional to deformation at the contacts [14].
3. Coefficient of friction is zero at all the contacts.
3 Contact Analysis of an Interference-Fit Motor
3.1 Determining the Initial Rotor Center
When the rotor rotates about its own axis by an angle ξ , its orbiting center revolves
around the stator center by an angle ξ (Z − 1) in the opposite direction. Thus, the
position vector of the point O (rotor center of a perfect-fit motor) in the stator frame,
is given by Eq. 1.
⎡ ⎤
cos(−ξ (Z − 1))
[O ] = c⎣ sin(−ξ (Z − 1)) ⎦ (1)
0
To determine the position vector of point O (rotor center of an interference-fit

motor), a new point O representing estimated rotor center is introduced. If there is
no error in estimation, O ≡ O. A corrective method based on minimization of the
potential energy of the system is adopted to find O closest to O. In this study, the
system comprises of the rigid rotor and the elastic rollers. Initially, let the estimated
rotor center, O coincide with O (rotor center of a perfect-fit motor).
3.2 Profile Generation of the Rotor
Let the periphery of the rotor be divided into ‘n’ points as shown in Fig. 3. Position
vector of the ith point on the periphery of the rotor with respect to its own frame,
([ei ] ) is given by Eq. 2.
⎡ ⎤
Acos(θi ) + ccos(Z θi ) − rm cos(θi + φi ) + ct cos(θi )
[e i ] = ⎣ Asin(θi ) + csin(Z θi ) − rm sin(θi + φi ) + ct sin(θi ) ⎦ (2)
0
where
2π
θi = ×i (3)
n
and

−1 Rsin(θi (Z − 1))
φi = tan (4)
A + Rcos(θi (Z − 1))
Fig. 3 Division of periphery

of the rotor into ‘n’ points
3.3 Transforming Profile of the Rotor from the Rotor’s

Frame to the Stator’s Frame
Frame (Rotor’s frame) is related to frame (Stator’s frame) by a rotation tensor

Q and the position vector of point O in frame is given by [O ] then,
[e i ] = [O ] + [ Q] [e i ] (5)
where
⎡ ⎤
cos(ξ ) −sin(ξ ) 0
[ Q] = ⎣ sin(ξ ) cos(ξ ) 0 ⎦ (6)
0 0 1
and initially, let the rotor center of an interference-fit motor coincide with the rotor
center of a perfect-fit motor (Refer Eq. 7).
⎡ ⎤
cos(−ξ (Z − 1))
[O ] = [O ] = c⎣ sin(−ξ (Z − 1)) ⎦ (7)
0
3.4 Determining Center of the Rollers
The center of all the rollers lies on a circle of radius (A + pcr ), centered at S and
occurs at an angular interval of 2π/Z . Thus, the position vector of point Kj (i.e.
center of the jth roller), in frame is given by Eq. 8.
⎡ ⎤
π(2 j−1)
cos
⎢ Z ⎥
[K j ] = (A + pcr )⎢
⎣ sin
π(2 j−1) ⎥ ∀1 ≤ j ≤ Z
⎦ (8)
Z
0
3.5 Contact Point Determination
The estimated contact point, N j , can be considered to be the point on the periphery
of the rotor closest to the point K j (Refer Appendix). Thus, the position vector of
the point N j (∀1 ≤ j ≤ Z ) in frame will be equal to [el ] , where l is a point on
the periphery of the rotor which is closest to point K j . It is mathematically shown
in Eq. 9. A similar approach was adopted by Lin et al. [9] to determine the contact
points in a cycloidal speed reducer.
[N j ] = [el ] (9)
iff,
el − K j ≤ e i − K j ∀1 ≤ i ≤ n
3.6 Deformation and Reaction Force at the Contacts
The contacts in this analysis are modelled as linear springs with stiffness coefficient
‘k’ as shown in Fig. 4. Shung and Pennock [10] also used a similar model to
estimate the contact forces.
The magnitude of deformation in jth roller and the reaction force exerted by the
jth roller (due to its deformation) on the rotor, in frame ([f j ] ), is determined using
Eq. 10 and Eq. 11.
δ j = (rm + rm ) − [N j ] − [K j ] (10)
Fig. 4 Contact force model

[N j ] − [K j ]
[ f j ] = kδ j ∀1 ≤ j ≤ Z (11)
[N j ] − [K j ]
The stiffness coefficient (Refer [11]) of the roller is obtained using Eq. 12.
πw
k= (12)
1−ν 2 π
4 E
1+ 4
3.7 Unbalanced Torque Calculation
Potential energy of the system and the net reaction force acting on the rotor in
frame is obtained from Eqs. 13 and 14 respectively. Here, the potential energy of the
system is simply equal to the elastic strain energy stored in the rollers as the rotor is
rigid.

Z
1
= kδ 2j (13)
j=1
2

Z
[F ] = [ f j ] (14)
j=1
Now, point O is shifted along the direction of F by an amount . The poten-

tial energy is calculated in this new state. If potential energy decreases, the above-
mentioned procedure is repeated all over until it starts to increase, i.e. the procedure
is iterated until the local minima of potential energy is obtained. At this state of local
minima, point O will be within a circle of radius centerd at O (Fig. 5). Smaller
the value of , closer is the point O to the point O and more accurate result is
obtained. At this state of minimum potential energy, the torque acting on the rotor
with respect to the stator frame ‘’ is calculated using Eq. 15.

Z
T = [001] [N j ] − [O ] × [ f j ] (15)
j=1
The whole process is repeated for different values of ξ and the results are presented
in the next section. The algorithm of the proposed approach is presented in the form
of a flowchart depicted in Fig. 6.
Fig. 5 Uncertainty in the

estimated rotor center
3.8 Flowchart of the Proposed Approach
See Fig. 6.
The results are obtained for the numerical data of the design and computational
parameters as shown in Table 1.
4.1 Variation of the Unbalanced Torque and Potential

Energy of the System
The potential energy plot (Fig. 7b) has a local minimum in the positions corre-
sponding to maximum compression of the chambers (ξ = 2nξ0 where, n ∈ Z).
The positions of local minima correspond to the state of stable equilibrium. There
is no unbalanced torque (Fig. 6a) acting on the rotor in these positions. When the
rotor is disturbed from these positions, an unbalanced torque acts on the rotor. This
unbalanced torque brings the rotor to the nearest state of stable equilibrium.
Local maxima of the potential energy (Fig. 7b) is observed in the positions cor-
responding to the maximum expansion of the chambers (ξ = (2n + 1)ξ0 where, n ∈
Z). The unbalanced torque (Fig. 7a) is also zero at these positions. When the rotor
Fig. 6 Flowchart of the proposed approach
is disturbed from these positions, the unbalanced torque acts on the rotor in such a
way that it brings the rotor to the nearest state of maximum compression. Thus, the
positions of local maxima correspond to the state of unstable equilibrium.
For any other position of the rotor in between the states of local maxima and local
minima, the unbalanced torque acts on the rotor in such a way that it brings the rotor
to the nearest state of stable equilibrium.
Table 1 Numerical data of design and computational parameters

Parameter category Parameter Values
Geometry [12] R 19.67 mm
A 31.98 mm
rm 7.97 mm
c 2.81 mm
w 10 mm
Z 7
Material k 9.87 × 108 N/m
Interference r m 71 µm
ct −51 µm
pcr −41 µm
Computational n 10,000
10−3 µm
Fig. 7 Variation of a unbalanced torque and b potential energy of the system with output shaft
rotation
4.2 Deformation at the Contacts
Figure 8 shows the variation of deformation (δ) of all the rollers with the output
shaft rotation (ξ ) for an Orbit motor of interference-fit type (Refer Table 1). It can
be observed from the plot that all the rollers experience positive deformation. This
plot has a similar nature compared to a plot obtained by Maiti and Nagao [13] for an
interference-fit motor using trial and error method.
Fig. 8 Variation of deformation in the rollers, due to the presence of interference with output shaft
rotation (ξ )
4.3 Estimation of Maximum Contact Pressure
Due to the inclusion of interference between the rotor and the stator, the contact points
are prestressed. Contact force exerted by each roller on the rotor is determined using
Eq. 11. The maximum contact pressure evaluated at each contact is obtained using
the Hertzian contact theory [14], assuming a cylinder on cylinder contact (Refer
Eq. 16) (Fig. 9).
The following assumptions are taken into consideration while implementing the
Hertzian contact stress model: the surfaces in contact are smooth, the contact are
non-conformal, no lubrication exists at the contacts, deformation at the contact is
small compared to the radius of curvature of the contacting bodies and only radial
loading exists at the contacts.

2 f j
σi = (16)
π bw
‘b’ is known as the half-contact bandwidth, given by Eq. 17

8[ f j ] 1−ν 2
E
b= (17)
π w Rroller + Rrotor
1 1
Fig. 9 Variation of contact pressure on the roller with output shaft rotation
Rrotor and Rroller are the radius of curvature of the rotor and the roller respectively.
((x )2 + (y )2 ) 2
3
Rrotor =
|x y − y x |
x = Acos(θ ) + ccos(Z θ ) − rm cos(θ + φ) + ct cos(θ )
y = Asin(θ ) + csin(Z θ ) − rm sin(θ + φ) + ct sin(θ )
x = ∂ x/∂θ y = ∂ y/∂θ
Rroller = rm + rm (18)
5 Conclusion
Leakage problem through the transition contacts can be mitigated or made zero by
providing interference between rotor and stator. But, as a result of that, an unbalanced
torque ripple is introduced. It is obvious that this unbalanced torque will increase
the load holding and the braking torque capacity, but will decrease the overall torque
efficiency. The aim of the study was to investigate the nature of this unbalanced
torque generated due to the sole inclusion of interference. The unbalanced torque has
a sinusoidal nature and at zero gauge pressure in all the chambers, it was observed
that the unbalanced torque tends to bring the rotor to the nearest state of stable
equilibrium, i.e. it orients the rotor in the nearest position corresponding to maximum
compression of the chambers. Future researchers can use this model to optimize the
design parameters of the motor by performing a trade-off between the magnitude of
unbalanced torque and the amount of leakage at the contacts.
Appendix
Figure 10a–e shows the forces acting on the rotor for different positions of the output
shaft. It can be clearly seen that for the positions of the rotor corresponding to maxi-
mum compression (Chamber C7 in Fig. 10b) and maximum expansion (Chamber C3
in Fig. 10d) of the chambers, the forces acting on the rotor are symmetrical. Thus,
the net torque about the rotor center is zero in these positions as observed in Fig. 7a.
Figure 11 shows the enlarged view of Roller-1 in contact with the rotor. Let the
periphery of the rotor be divided into five points. To determine which point out of
these five is the contact point for roller 1, the algorithm calculates the distance of
each point from the roller center K1 and selects the point which is closest to K1.
From Fig. 11, it is clear, that d3 has the smallest value. Therefore, point 3 is the
contact point.
Fig. 10 Forces acting on the rotor for a ξ = −0.5ξ0 , b ξ = 0, c ξ = 0.5ξ0 , d ξ = ξ0 , e ξ = 1.5ξ0

Fig. 11 Contact point

estimation
References
1. Ansdale R (1968) The Wankel RC engine, 1st edn. Iliffe, London

2. Colbourne J (1974) The geometry of trochoid envelopes and their application in rotary pumps.
Mech Mach Theory 9(3–4):421–435
3. Colbourne J (1975) Gear shape and theoretical flow rate in internal gear pumps. Trans Can Soc
Mech Eng 3(4):215–223
4. Robinson F, Lyon J (1976) An analysis of epitrochoidal profiles with constant difference
modification suitable for rotary expanders and pumps. J Eng Ind 98(1):161–165
5. Maiti R (1992) Distributor valve port sequences in epitrochoid generated rotary piston machine
‘ROPIMA’ type hydrostatic units. Arch Appl Mech 62(4):223–229
6. Maiti R (1991) Active contact stresses at epitrochoid generated rotor-stator set of fixed axis or
equivalent system type hydrostatic units. J Eng ind 113(4):465–473
7. Maiti R (1993) Active contact problems in epitrochoid generated floating axis orbital rotary
piston machines. J Eng Ind 115(3):337–340
8. Nag A (2015) Analysis of active contacts in epitrochoid generated form closed rotor-stator set
with specific reference to hydrostatic motor performance. Ph.D. thesis, IIT Kharagpur
9. Lin K-S, Chan KY, Lee J-J (2018) Kinematic error analysis and tolerance allocation of cycloidal
gear reducers. Mech Mach Theory124:73–91
10. Shung J, Pennock G (1994) The direct contact problem in a trochoidal-type machine. Mech
11. Cornell R (1981) Compliance and stress sensitivity of spur gear teeth. J Mech Des 103(2):447–
459
12. Nag A, Maiti R (2013) Unification of epitrochoid origin profile design approaches for external
lobed star member used in hydrostatic and gear units. Proc Inst Mech Eng Part C: J Mech Eng
Sci 227(2):299–310
13. Maiti R, Nagao M (1999) Prediction of starting torque charecterstics of epitrochoid generated
orbital rotary piston hydraulic motors. JSME Int J Ser C 42(2):416–426
14. Jhonson K (1985) Contact mechanics, 1st edn. Cambridge University Press, Cambridge, UK
Design and Analysis of Rocker Arm Shaft
in IC Engine Hino Series for Reduction
of Assembling Dismantling Time
R. Iruthayaraj, S. Palani, R. Ajay Vishal, T. Living Rockson, and K. Anand
Abstract Nowadays, in automobile industries the major competition exists in the

field of design and cost of any product. Automobile plays a vital role in our daily life.
One of the important components of the automobile is the engine of the automobile.
Assembling and dismantling of the engine are difficult and tedious process. It is
considered as time-consuming process also. If it is in the case of any valve leak,
valve bend or due to any abnormal combustion in the engine, then the total set-up
of the engine is to be dismantled part by parts for reworking purpose. The engine
involves the large number of parts, and it takes too much of time for removing them
completely into individual part. For reducing the time consumption, “the replica of
rocker arm shaft” is designed. By using this, whole cylinder head is dismantled from
the engine without any major damage. The time consumption would be very less,
comparing with the existing method. The ultimate aim of our project is to reduce the
time consumption for dismantling and reassembling of engine, without any major
damage to engine parts and other surrounding parts.
Keywords IC engine · Rocker arm · Assembling · Dismantling · Time-consuming
1 Introduction
Ashok Leyland is an Indian automobile company headquarters in Chennai, India. It

is owned by the “Hindu group”, it is the second-largest commercial vehicle manu-
facturer of buses in the world and the tenth-largest manufacturer of trucks globally.
It is operating nine plants; Ashok Leyland also makes spare parts and the engines for
industrial and marine applications. Engines are the heart of the automobiles. Engines
are used to move up the whole automobile body. Value analysis and value engineer-
ing should be employed in the engine design. In most of the automobiles, different
varieties of engines are used. Engine varies depending upon the size, variety, type,
R. Iruthayaraj · S. Palani · R. Ajay Vishal (B) · T. Living Rockson · K. Anand

Department of Mechanical Engineering, Vel Tech Multitech, Avadi, Chennai, Tamilnadu 600062,
India

https://doi.org/10.1007/978-981-15-4488-0_11
128 R. Iruthayaraj et al.
applications, specialized purposes. Engine’s capacity also varies depending upon the
application for light-duty vehicles; four cylinder engines are used; for heavy-duty
vehicles, such as trucks and lorries, six cylinder engines are used such as H-series
engines. H stands for “Hino”. H-series engine is used for heavy-duty purpose; H-
series engine is huge in construction. It comprises of many major and minor com-
ponents. In H-series engine, many parts are available. Several important parts of the
H-engines are intake and exhaust values, intake and exhaust manifold, rocker arm
assembly, cylinder head, cylinder top, fuel injector pipe, fuel injector, bolts, nuts,
studs, common rails, injector pipes, turbocharger, exhaust gas recirculation, poppet
valves, tappet valves, pushrod, camshaft, flywheel, piston head, rocker arm shaft,
fuel filter, fuel sump and thermostat.
Nassef and Elkhabit [1] investigated the failure of a rocker arm shaft of a passen-
ger car. The shaft failed by brittle fracture across one of the four holes supporting the
shaft into the cylinder head. Microscopic observation of the failed shaft revealed that
the four dark etching areas are surface hardened zone of martens tic microstructure.
Microscopic investigation of the failed shaft reveals the presence of microcracks close
to the supporting holes. This premature failure has occurred by the rapid crack prop-
agation because of the lower toughness of martensite that excluding design reasons,
failures such as consequential failure, improper lubrication, faulty manufacturing.
For all the above reasons and given the results of the above investigation, such fail-
ure is attributed to improper heat treatment of the shaft during manufacturing. It is
recommended to conduct a proper heat treatment to the whole body of the arm in
order to prevent recurrent failures in the future [1].
Chirag et al. [2] focused to reduce the bottleneck areas by applying necessary tech-
niques for improving the productivity. It is found that the main areas where bottleneck
happens are head assembly key up station, tappet line and hot testing of engines. It
is notified that various non-value added (NVA) activities and with the help of pare to
principle; they sorted out most probable causes, i.e. traffic problem of cart, unwanted
movements and handling of equipments, etc. By applying all these techniques, then
the productivity improvement in one assembly system causes improvement in other
subsystem too, NVA reduction by proper sequences of arrangement in activities by
the installation of new tool and trouble shooting for head sub assembly, tappet line
sub assembly and hot test cells etc [2]. Mujahid and Sheikh [4] reviewed the various
types of rocker arms based on the sources from the last 40 years, in order to under-
stand the rocker arm, for its problem identification and further optimization. In their
work, the various types of rocker arms and their materials are studied. Rocker arm
is an important component of engine; failure of rocker arm makes an engine useless
and also requires costly improvement. The most suitable material for the rocker arm
is steel because it has better fatigue strength than the aluminium [3].
Bache and Swaminadhan [4] discussed the failure of rocker arm in valve actuating
mechanism to study and analyse the stresses generated, and they are modelling the
rocker arm using CATIA CAD software and various regions of stresses and deforma-
tion found out using OPTISTRUCT-RADIOSS structural analysis solver which gives
solution for structural optimization using finite element analysis (FEA). The main
purpose of their study is to determine the value of stresses in rocker arm at extreme
Design and Analysis of Rocker Arm Shaft in IC … 129
conditions. The design can be optimized by reducing the amount of material used.
Stresses within the rocker body are calculated by using finite element analysis and
validated with type help of result obtained by using analytical mathematical formu-
lae [4]. Sonwane and Debase [5] started the discussion as currently manufacturing
industries are facing a greater competition in the market. Due to this, competition
among industries is trying to improve and increase both quality productivity contin-
uously. So the main focus was given on the review of recent research related to the
continuous improvement of automobile engines assembly line and a case study of
automotive industry. The present work gives an overview of previously done research
work and for new research; it gives a direction for continuous improvement of auto-
mobile engine assembly line. Also with the help of case study, it is explained how to
reduce major breakdown causing production loss at engine assembly line [5].
Gide et al. [6] proposed that the purpose of valve train is to operate the inlet and
outlet valves of engine. The valve train mainly consists of rocker arm, pushrod, cam,
poppet valve and spring for keep the valve in closed position. The valve spring param-
eters are optimized based on the space availability, buckling of pushrod and natural
frequency of system. It is observed that the valve jump engine speed with respect
to optimized valve spring is enhanced. In order to capture the dynamic behaviour
of the valve train system closely, each coil of the nested valve springs is modelled
as separate flexible body and the contacts between coils of these flexible bodies are
established [6]. Vinay Kumar and Sudeendra Srinivas [7] used ANSYS Software
to design and perform structural analysis of the rocker arm. The geometric model
of the rocker arm is developed and created by using CAD Software CATIA. The
present work deals with the three-dimensional solid modelling, design and analysis
of rocker arm of an IC engine. The rocker arm is designed, and the forces on the
same are calculated. The structural analysis is carried out using different materials
in order to arrive at optimized design of rocker arm. The finite element method is the
most popular approach and found commonly used for analysing fracture mechanics
problem [7].
Jadhav et al. [8] found that the rocker arms oscillate about rocker arm shaft because
of action of push rod on one side and spring action on other side which causes
bending of rocker arm; as a result, the bending stresses have induced the failure of
rockers which takes place due to these stresses. Their research work deals with the
theoretical and finite element analysis of rocker arm. The bending stress developed
by the theoretical and finite element analysis of rocker arm is under static condition.
Among the above two methods, the most static condition is considered for better
results of rocker arm [8]. Bansod and Neigh [9] stated that the internal combustion
engine, the tappet noise, is a major function. The tappet noise is a noise made by
the lash or clearance between rocker arm and valve stem in an engine tappet noise
that is characterized by its characteristic “Take Take” sound which is clearly audible
when the engine dynamometer runs at 1400 rpm. The noise of the tappet generally
occurred by the improper setting of tappets at the subassembly of tappet section and
at the main section of assembly stations. By various corrective actions are counter
measures taken as per the requirements of the tappet sections, provides increased
tension of the timing gear chain for better and noiseless performance [9]. Shyam
and Thacker [10] shown the main objective of any organisation is to increasing the
productivity and profitability in order to increase the productivity and profitability.
They reduce the defects and rejection of the product by the seven quality control
tools and techniques. In today’s competitive market, the major concern is to satisfy
the customers for their requirements [10].
In our research work for reducing the time consumption, “the replica of rocker
arm shaft” is designed. By using this, whole cylinder head is dismantled from the
engine without any major damage. The time consumption would be very less, com-
paring with existing method. Our ultimate aim of this project is to reduce the time
consumption for dismantling and reassembling of engine, without any major damage
to engine parts and other surrounding parts.
2 Designing the Replica of Rocker Arm Shaft Identification

of the Problem
2.1 Identification of the Problem
In engine assembly shop, the major problem exists in the cylinder head, due to
valve leak, valve bend, etc. If so, then it is subjected to rework cylinder head in the
rectification area.
2.2 Selection of Suitable Solution
As the engine comprises of many components, and due to its complex structure, “the
replica of rocker arm shaft” is designed to remove the cylinder head together with
major parts, without causing any damage to other parts.
2.3 Design Calculation of the Shaft
Known details: Diameter of shaft = 19 mm; Length of the shaft = 723 mm; Maximum
equivalent stress = 115.91 Mpa; Deformation (or) deflection = 0.6 mm 5. Load =
183 kg = 1765.8 N.
Finding the tensile stress,
Tensile stress, σt = F/A = 1739.23/(π/4 × d 2 ).
σt = F/A = 1739.23/(π/4 × d 2 )
Finding the torque (M t ):

π
Mt = τ × D3
4
π
Mt = 115.91 × 183
4
Mt = 1.33 × 1016 N m
To find the maximum shear stress (N/mm2 ), Bending stress is:

π 4
σb = M y /I = M y / × d
32
Mb = (W × L)/4
= 1789.23 × 723 mm
Mb = 323.40 N m.
π
σb = (323.40 × 0.6e−3 )/ × 184
32
= 1.508 N/mm2

τmax = 16/(π × d 3 ) (Mb × K b )2 +(Mt × K t )2
K b and K t for suddenly applying load with minor shock K b = 1.5 to 2 take K b =
1.5.
K b = 1.0 to 1.5 take K t = 1.0 Symbols and Names.
Here, σb = Bending stress, M t = Maximum torque, σt = Tensile stress.
K b , K t = Minor shocks applied.
To find maximum torque,

√
τmax = 16/(π × d 3 ) (Mb × K b )2 + (Mt × K t )2
√ 2
τmax = (16)/(π × d 3 ) (323.40 × 1.5)2 + 1.56 × e16 × 1
√
= (16)/(π × 183 ) (323.40 × 1.5)2 × 2.43 × 103
= (16)/(π × 183 ) × 1.55 × 1016
τmax = 1.15 × 1013 N/mm2
After designing the replica of the rocker arm shaft, the drafting for the particular
shaft is performed by using the SolidWorks software. This drafting action can accom-
plish by the use of SolidWorks. This drafting helps us to understand clear structure
and dimensions of the rocker arm shaft. The design of the replica of rocker arm shaft
is designed first with proper designation. After the designing process, drafting action
has been performed for previewing the rocker arm shaft.
The drafting view of the replica of rocker arm shaft is shown in Fig. 1.
Fig. 1 Drafting view of the replica of rocker arm shaft
In the real rocker arm shaft, rocker arms are available in the rocker arm shaft. The
purpose of existing rocker arm shaft is to transmit the radial motion of the cam lobe
into the linear movement to operate the poppet valves for opening and closing the
flow. In the replica of rocker arm shaft, there a long shaft, 2 movable (adjustable)
support rocker valves and 2 both side fixed rockers. In this replica of rocker arm
shaft, rocker arms are not involved. The main purpose of this shaft is to hold and
withstand the weight of the cylinder head. Figure 2 shows the design of the shaft
using CATIA software.
Fig. 2 Design of shaft using CATIA software

Fig. 3 Design of part. a Left side, b right side using CATIA software
Fig. 4 3D view of rocker arm shaft assembly using SolidWorks
The designs of parts (a) left side (b) right side using CATIA software are shown
in Fig. 3.
After designing the rocker arm shaft, the shaft is to be manufactured and fitted into
the cylinder head. The rocker arm shaft is to be fitted into the holes of the cylinder
head, bolted up and lifted by using the crane hook. The 3D view of rocker arm shaft
assembly using SolidWorks is shown in Fig. 4.
2.4 Principles of Operation
Engine varies according to the size, variety, type, applications and other specialized
purposes. Engine’s capacity also varies depending upon the application such as light-
duty vehicles and heavy-duty vehicles. For light-duty vehicles, 4 cylinder engines
are used, and for heavy-duty vehicles, such as trucks, Lorries, 6 cylinder engines
such as H-series engines are used (H stands for “Hino” which is used). In an internal
combustion engine, the cylinder head (often informally abbreviated to just head) sits
above the cylinders on top of the cylinder block. Due to these complex structure that
involving many parts, in the engine, identification of error in the engine is difficult,
if the engine is not functioning. After the identification of the error in particular part,
i.e. fuel injectors in the cylinder head, and then only the rework or the rectification
work of cylinder head is performed.
In general, same as the assembling process, disassembling process is very diffi-
cult and time-consuming one. Due to its complex structure, each and every part is
to be removed carefully for the reworking purpose. Engine is absorbed abnormal
combustion, and then we proceed to check out the fuel injector problems. In early
process to remove the cylinder head, the following things are to done. The following
parts to be removed are rocker arm shaft, exhaust manifold, fuel injector, remove the
rocker arm shaft from the head, remove inlet and exhaust manifold, and then remove
exhaust gas recirculation (EGR), turbocharger and TG Elbow, Thermostat, common
rail injector, cylinder head bolts, valve cap, leak pipes and gaskets. Disassembling of
parts is time-consuming sometimes disassembling causes damage to the parts. For
this, we are designing the special shaft which looks similar to the rocker arm shaft,
i.e. “replica of rocker arm shaft”. Electrically operating hook is used to hold and lift
the whole cylinder head set-up with the replica of rocker arm shaft, from the top of
the engine. Then, we can easily rework with the cylinder head, which was lifted and
hung over the engine through crane hook.
Time Calculation by Existing Method

Dismantling:
Step 1:
Inlet and Exhaust manifold; Rocker Arm Shaft; EGR, Thermostat, Common Rail
and injector = 35 min
Step 2:
Turbocharger and Elbow; Cylinder Head bolts, Manifold Gaskets; Valve cap pins,
Lube oil Pipes = 30 min
Step 3:
Banjo Bolt, Tappet; Oil Supply Line and Liners; Other Exterior Parts = 15 min.
In this method for dismantling the cylinder head, time consumption is 1 h 20 min.
Assembling
After finished, they dismantle the parts to check and analyse the damages in piston.
For reassemble, they consume the same time. In this method, we found the drawbacks,
which they are more time consumption and material damaging for the removing
process. So, we use the new technique in the removal of cylinder head to avoid
unwanted things.
Total time taken for assembling = 10 + 25 + 25 = 60 min.

Time Calculation by proposed Method
Dismantling
Step 1:
Removing the Rocker Arm Shaft; Further, Bolts nuts of this component = 15 min
Step 2:
Exhaust and Inlet pipes Springs; Valve cap, Fuel Liner = 10 min
Step 3:
EOT is placed at in the Dummy Rocker Arm shaft = 5 min
It can be pulled upside to rigid condition.
Assembling
Due to its more complexity structure, the assembling of the single engine is consid-
ered as the time-consuming process. It requires more manpower for fixing it with
proper proportions and alignments. As in case of any problem or any improper
assembling, the whole set-up is to be removed or dismantled one by one. Same as
the assembling process, disassembling process is very difficult and time-consuming
one. Due to its complex structure, each and every part is to be removed carefully for
the reworking purpose.
Total time taken for assembling = 5 + 10 + 10 = 25 min.
In this method, the time consumption for assembling is 25 min, which is very less
compared to the existing method. Figure 5 shows the pie chart for time utilized (a)
exiting and (b) proposed methods. Figure 5 shows the pie chart for time utilized (a)
exiting and (b) proposed methods (Table 1).
Total Time Saved by Overcoming the Old Technique
Total time saved = (Total time taken by normal method)
Fig. 5 Pie chart for time utilized. a Exiting, b proposed methods

Table 1 Comparison of total time taken by both existing method and new method
Sl. no. Technique used Dismantling time Assembling time Total time
1 Normal method 1 h: 20 min 1h 2 h: 20 min
2 Proposed method 30 min 25 min 55 min
−(Total time taken by proposed method)

= (2 h: 20 min)−(55 min)
Total time saved = 1 h: 25 min.

By comparing those two methods, proposed method is considered as the most promis-
ing method. In existing method, the total time used is 2 h: 20 min. And in the proposed
method, the total time used is 55 min. The total time saved by proposed method is
1 h: 25 min.
4 Conclusions
In the proposed technique, the dismantling time of cylinder head may be reduced
in a greater extent compared to the traditional technique. By overcoming the time
consumption of dismantling of cylinder head, rate of production may be increased.
Here, also avoid the damages of the surrounding parts of the cylinder heads. The
replication of rocker arm shaft may be used to remove the cylinder head in effec-
tive manner without removing the major parts such as inlet and exhaust manifold,
and EGR. Also In this research work designed the rocker arm shaft and given for
manufacturing, after receiving the product, able to calculate the time consumption
by using this method. By using this replica of rocker arm shaft, easily remove the
cylinder head for the reworking purposes. Furthermore, improve the production rate
by reducing the time consumption for rectifying the defective engines.
References
1. Nassef GA, Elkhabit A (2016) Analysis of a failed rocker arm shaft of passenger car engines.
Sci Direct 5(6):10–14
2. Deotare CK, Dabade UA (2015) Productivity improvement in engine assembly line at GM
India Ltd. Pune, India. Int J Latest Trends Eng Technol 6(1):9–16
3. Mujahid S, Sheikh S (2013) Design and analysis of rocker arm. Int J Mech Eng Rob Res
2(3):191–197
4. Bache S, Swaminadhan P (2018) Design and static structural analysis of rocker arm in IC
engine. Int J Eng Sci Res Technol 7(4):755–763(2018)
5. Sonwane AB, Debase DA (2017) Continuous improvement of automobile engine assembly
line: A review and case study. J Adv Sci Technol 13(1):245–252
6. Gide S, Shrives J, Kharde RR. Review on critical speed improvement in single cylinder engine
valve train. Int J Innov Eng Res Technol. ICITDCEME’15 conference proceedings ISSN No
2394-3696
7. Vinay Kumar D, Sudeendra Srinivas B (2017) Design and fabrication of rocker arm. Int J Core
Eng Manage NCETME 193–200. ISSN 2348-9510
8. Jadhav S, Patil PJ, Mulik, PV (2018) Experimental and finite element analysis of rocker arm
for bending failure. Int J Adv Eng Res Dev 5(7):1–9(2018)
9. Bansod N, Neigh HS (2015) Tappet noise reduction in motor cycle engine. Int J Eng Dev Res
3(2):461–474
10. Bambharoliya SH, Thacker HR (2015) Reducing reduction rate in small scale machining unit
using seven quality control tools. Int J Eng Dev Res 3(4):582–586
Kinematic Analysis and Dimensional
Synthesis of Filleted-Rhombus Path
Generating Adjustable Four-Bar
Mechanism
Ganesan Govindasamy, Clinton Wilson, and K. Neeraj
Abstract Dimensional synthesis and kinematic analysis of adjustable four-bar

mechanism to generate accurate and uninterrupted path of Filleted-Rhombus are
investigated. Harmonic Spacing (HS) algorithm is used to improve the positioning
of precision points along the intended profile path. Reconstructed adjustable parame-
ter curve (RAPC) is employed to extract adjustable link length value of intermediary
path points along the intended path conforming to an equally incremented crank rota-
tion value. Simulation of motion of adjustable mechanism to generate the intended
coupler path is carried out to verify the dimensions of the links and other parameters
of the mechanism. Optimization employing Genetic Algorithm (GA) followed by
Pattern Search (PS) is implemented using MATLAB codes in the synthesis of mecha-
nism. Results show precise and continuous coupler path to generate filleted-rhombus
profile and variation of kinematic parameters versus crank angle.
Keywords Kinematic analysis of adjustable mechanism · RAPC · Adjustable

four-bar linkage · Filleted-Rhombus path
1 Introduction
Kinematic synthesis is applied to obtain motion generation, function generation and

path generation of mechanisms. Path generation is the motion of the coupler point to
trace a path profile. Point-to-point path generation is to follow the computed precision
points while the trajectory between the computed points is approximated. The other
type of path generation is of the continuous type which specifies the whole path
with many points that approximately pass through the specified curve. Adjustable
mechanism is neither equal to sophisticated flexible robot nor equal to non-adjustable
mechanism. It takes the mid-position between non-adjustable mechanism and robot
[1] in terms of speed, precision and simplicity. In contrast to analytical methods
G. Govindasamy (B) · C. Wilson · K. Neeraj

Nagar, Coimbatore, Tamil Nadu 641114, India

https://doi.org/10.1007/978-981-15-4488-0_12
that satisfy fewer precision points precisely optimization methods such as Genetic
Algorithm, Pattern Search, and Neural Networks, etc. satisfy many precision points.
Haws and Kay [2] used an adjustable mechanism employing a cam-link to generate
one-fourth of a square and a cardioid path. Kohli and Singh [3] made use of cam-
link linkages to generate an exact path. Ogawa et al. [4] synthesized adjustable
multi-link linkages to produce L-shaped and straight lines paths connecting several
precision points. Mundo and Gatti [5] employing non-circular gears synthesized five-
bar linkages for accurate path generation. Sodhi and Russel [6] applied a new design
method for adjustable slider-crank mechanisms to realize two-phase path generation.
Mundo et al. [7] synthesized adjustable cam linkage mechanism to adjust the position
of the fixed pivot of the link using a cam for exact path generation. Ghoshal and
Chanekar [8] synthesized adjustable linkages in two stages in order to reduce the
variables required to design a linkage.
Synthesis of continuously adjustable crank-rocker linkage to produce rectangular
path was tried by Zhou [9]. Ganesan and Sekar [10] used a graphical method to obtain
an adjustable parameter curve (APC) from the dimensions of the mechanism synthe-
sized by Zhou [9] to produce a rectangular path. Ganesan and Sekar [11] optimally
synthesized adjustable linkages to generate a rectangle with filleted corners. Ajith
Kumar et al. [12] synthesized continuously adjustable linkages to generate rectan-
gle with chamfered corners, rectangle with full chamfering and one DOF cam-link
adjustable mechanism to generate oblong-hole path.
It has been found from the above literature that the precise filleted-rhombus path
generation mechanism and kinematic analysis of such a mechanism are not attempted.
Section 2 presents an adjustable linkage model, precision points on filleted-rhombus
path, dimensional synthesis using hybrid optimization. Section 3 presents RAPC
method to extract path points on the filleted-rhombus path for one degree incremental
value of crank for one revolution. Velocity and acceleration analysis of various links
of mechanism to produce filleted-rhombus path is given in Sect. 4. Section 5 presents
results and discussions. Conclusions and suggestions are presented in Sect. 6.
2 Filleted-Rhombus Path Generating Adjustable

Mechanism
Four-bar adjustable mechanism model of Zhou’s [9] is used in the synthesis to gen-
erate a precise and continuous filleted-rhombus path, as shown in Fig. 1, in which a
rocker (L 3 ) is carried by a slider (D) which forms the adjustable link. This adjustable
link length (S) from the pivot (A) along the x’-axis of a local coordinate system to
the pin joint between the rocker and slider is called ‘adjustable parameter’ of the
mechanism. Link AB is crank (L 1 ). In triad link BCM, side BC forms coupler link
(L 2 ), side BM (L 6 ) carries coupler point (M). In this adjustable four-bar mechanism,
the required adjustable parameter values (S) are determined so that the coupler point
can trace filleted-rhombus path precisely when the crank makes one revolution.
Kinematic Analysis and Dimensional Synthesis … 141
Fig. 1 Adjustable mechanism model
Coupler point has to travel along the P1 P2 P3 P4 P1 path which comprises four
straight lines and two arcs as shown in Fig. 1. The difficulty arises when the coupler
point negotiates a change in direction. This demands closer precision points near
the points where changes in direction take place. The following harmonic spacing
algorithm satisfies the closer precision points at the ends of a straight line or an arc.
2.1 Precision Points Algorithm for Filleted-Rhombus Path
Difficulty arises when the coupler point negotiating positions where it changes direc-
tion drastically. This demands closer precision points near these points. The following
harmonic spacing algorithm (1) satisfies the closer precision points at the ends of a
straight line or an arc as shown in Fig. 2.

xi + x f (xi − x f ) ( j − 1)
xj = + cos (1)
2 2 N −1
where x is variable with an interval [xi , x f ] and j = 1, 2, …, N, x j -intermediary point

coordinate value, xi -initial point coordinate value, x f -endpoint coordinate value, N
(number of precision points inclusive of initial point and end point).
Fig. 2 Precision points of 0.5

filleted-rhombus path
0.25
0
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
-0.25
-0.5
2.2 Synthesis of Mechanism to Generate Filleted-Rhombus

Path
synthesis as the known values of the coupler points. The mechanism is synthesized
using Eqs. (10)–(21) which govern the geometry of the intended mechanism to satisfy
all the 360 precision points. Synthesized mechanism’s unknown independent param-
eters (φ, x A , y A , β, L e , L 2 , L 3 ) and dependent parameters (L 6 and L 1 ) are given in
Table 1.
The objective is to minimize link length value (S) for a set of independent variables
(φ, x A , y A , β, L e , L 2 , and L 3 ) of the linkage whose lower bound (LB) values are −
3, −2, 2.5, 2.5, 1, 50, −90 and upper bound (UB) values are −1.5, −1.0, 4, 4, 2,
80, −50, respectively, for these variables. The difference between the maximum
Table 1 Parameters of the synthesized mechanism

Optimized filleted-rhombus path generation Values of the variables and other parameters
parameters and other variables
Fixed pivot x-coordinate value (x A ) −2.4952
Fixed pivot y-coordinate value (y A ) −1.3322
Coupler length (L 2 ) 3.9999
Rocker length (L 3 ) 3.9713
Slider offset length (L e ) 1.9776
Coupler point arm angle (β) 53.8528°
Frame orientation angle (θ) −78.2318°
Crank length (L 1 ) 0.8689
Coupler pointer length (L 6 ) 2.8715
Starting crank angle value when pointer is at P1 20.8642
(θ1 )
Maximum value of adjustable link length (Sm ) 6.5327
Minimum value of adjustable link length (Sn ) 5.8469
Error function (E a ) 0.6858
adjustable link length and minimum adjustable link length is the error function, i.e.
minimization of this error is the objective of the synthesis of the mechanism.
E a = Sm − Sn (2)
Initial value of L 1 and L 6 are 1.0 and 2.5 units, respectively. Penalty factor and
population size are given 500 and 20 respectively and tournament selection method
is used. Objective function must satisfy the following constraint functions (3)–(8) to
satisfy link length criterion of Grashof and also to avoid linkage dead points.
|L ce |max − L 3 ≤ 0 Where L ce = yC − L e (3)

L1 + Sm2 + L 2e − L 2 − L 3 ≤ 0 (4)

L2 − L3 + L1 − Sn2 + L 2e ≤ 0 (5)

L1 − L2 + L3 − sn2 + L 2e ≤ 0 (6)
2
L 22 + L 23 − Sn2 + L 2e − L1 − 2L 2 L 3 cosμ1 ≤ 0 (7)
where μ1 = 20◦ and Sn = minimum value

2
−L 22 − L 23 + Sm2 + L 2e + L1 + 2L 2 L 3 cosμ2 ≤ 0 (8)
where μ2 = 160◦ and Sm = maximum value

In order to trace the path points on the path profile by the pointer in the right order
and overcome ‘order defect’ during full rotation of the crank in the anti-clockwise
direction at uniform angular velocity, a constraint Eq. (9) is forced.
(θ1 )i < (θ1 )i+1 . (9)
2.3 Optimization Employing Genetic Algorithm Followed

by Pattern Search
In the Genetic Algorithm (GA) optimum value of the objective function may not be
always reached due to the stopping criteria imposed on GA.
This requires further exploration of the search space. By changing the stopping
criteria, it is possible to find the optimum solution. However, it may require many
more function evaluations to reach the optimum value. A local search can be more
efficient in such a scenario. Pattern search is used for local search that starts from
the point where GA reached thus making a hybrid optimization process.
3 Filleted-Rhombus Path Generation Using RAPC Method
Reconstructed adjustable parameter curve method is employed to extract path point

of filleted-rhombus path. The required equations to find path points for uniform
incremental values of the crank rotation along the four straight lines and two circular-
arcs of radius 0.5 units of filleted-rhombus are described in the remaining section.
Closed filleted-rhombus path is given by P1 P2 P3 P4 P1 .
It is anticipated that the point (M) of coupler link in the mechanism travel through
the complete profile path, shown in Fig. 1, through endless path points for unin-
terrupted generation of intended path. Path point refers to any point on the filleted-
rhombus path. Path points on the filleted-rhombus path are obtained from synthesized
values (x A , y A , φ, β, L 2 , L 1 , L 3 , L e and L 6 ) of the mechanism. Number of precision
points N is taken as 360 for better results. Precision points used and their correspond-
ing crank angles obtained in the synthesis do not match to uniform crank arm angular
displacements.
The coordinate values of the first precision point M1 at P1 is (1, 0) and corre-
sponding angular position (θ1 )1 is obtained from Table 1. The length between first
path point M1 and fixed point A is given by Eq. (10). Equation (11) gives the angle
between link L 1 and line AM1 .

2 2
L AM(1) = x M(1) − x A + y M(1) − y A (10)
L 21 + L 2AM(1) − L 26
β1(1) = cos−1 (11)
2L 1 L AM(1)
Equation (12) also can be used to get (θ1 )1 . In Eq. (12) the ±sign specifies the

(θ1 )1 = tan−1 y M(1) − y A , x M(1) − x A ± β1(1) (12)
The Eqs. (13) and (15) give values of coordinates of joint B for first path point.
Equation (14) gives angle (θ6 )1 .
x B(1) = x A + L 1 cos θ1(1) (13)

θ6(1) = tan−1 y M(1) − y B(1) , x M(1) − x B(1) (14)
y B(1) = L 1 sin θ1(1) + y A (15)
The value of β is got from synthesized values (Table 1). The coordinates of joint
C1 are attained from Eqs. (16)–(18)
θ2(1) = θ6(1) − β (16)
xC(1) = L 2 cosθ2(1) + x B(1) (17)
yC(1) = L 2 sin θ2(1) + y(1) (18)
For ease of analysis of slider dyada local coordinate frame is used. Joint A is
taken as the local coordinate origin. Coordinates of joint C in local coordinates are
obtained from Eqs. (19) and (20) from the global coordinates.


Location of the slider from fixed the pivot A is given by, adjustable link length,S(1)
when the coupler is indicating the first path point, is found from Eq. (21).

2
S(1) = xC(1) ± L 23 − yC(1) − L e (21)
For the first path point, using Eqs. (10)–(21), adjustable length S(1) is attained.
The second path point is obtained by the intersection of line P1 P2 at M(2) and an arc
of length L 6 from joint B. This gives the coordinates X M(2) and Y M(2) of the path
point M(2) from equations by solving an arc and straight line after incrementing the
crank by 1°. Substituting x M(2) , y M(2) through Eqs. (10)–(21) in the place of x M1 ,
y M(1) to get the adjustable parameter value S(2) .
Intersection of two arcs along the filleted part of the profile is used to find path
points. Using the known slope and intercept values of the straight lines path points
along the straight line are obtained. Thus, S(i) where i = 1, 2, …, 360, adjustable
length values of path points are obtained. These path points are used to construct the
required filleted-rhombus path. Figure 3 shows a synthesized mechanism to generate
filleted-rhombus path. Figure 4 shows the adjustable link length (S) versus crank
angle incremented by 1° for filleted-rhombus path generation. It is related to a non-
adjustable mechanism by fixing the slider at a mean position.
Adjustable Four-bar Mechanism to generate Filleted-Rhombus Path

2.5
1.5
1
Length Units
0.5
-0.5
-1
-1.5
-2
-2.5
-4 -3 -2 -1 0 1 2 3 4
Length Units
Fig. 3 Synthesized adjustable mechanism
Position of Slider pin from fixed pivot

vs. Crank Angle
6.6 Filleted-Rhombus Path
Coupler Curves of Adjustable and Non- fixedpivot at S = 6.1899units
6.5
adjustable Mechanisms
Adjustable Link Length
6.4
Filleted-Rhombus path motion
0.6
fixedpivot at S =6.1899units 6.3

0.4 6.2
0.2 6.1
0 6
-0.2 5.9
-0.4 5.8
-1 -0.5 0 0.5 1 1.5 100 150 200 250 300 350 400 450
X - axis length units Crank Angle
Fig. 4 Generated coupler path and adjustable parameter curve (APC) required
4 Kinematic Analysis of the Mechanism
For every incremental value of crank rotation, positions of the links in the mechanism
are found from known dimensions of the links and alignment angles of coupler (θ2 )i
and rocker (θ3 )i . From known velocity of crank, coupler and rocker velocities are
found from Eq. (22) and coupler and rocker angular accelerations are obtained from
Eq. (23).

−L 2 sinθ2 L 3 sinθ3 θ̇2 L 1 θ̇1 sinθ1
= (22)
−L 2 cosθ2 L 3 cosθ3 θ̇3 L 1 θ̇1 cosθ1
Kinematic analysis is of the mechanism is related to a non-adjustable mechanism

by fixing the moving pivot at a mean value of Sm and Sn .

−L 2 sinθ2 L 3 sinθ3 θ̇2
−L 2 cosθ2 L 3 cosθ3 θ̇3

L 1 θ̈1 sinθ1 + L 2 θ̇22 cosθ2 + L 1 θ̇12 cosθ1 − cosθ3 L 3 θ̇32
= (23)
L 1 θ̈1 cosθ1 − L 2 θ̇22 sinθ2 − L 1 θ̇12 sinθ1 − sinθ3 L 3 θ̇32
Figure 5a–d shows the variation coupler angle, velocity, acceleration and rocker
acceleration, respectively versus crank angle. Figure 5a shows that the coupler angle
varies non-uniformly with crank angle. Figure 5b shows coupler angular velocities
Coupler Angle vs. Crank Angle Coupler Angular Velocity vs. Crank Angle
70
Coupler Angular Velocity rad/s
1.5
65
1
Coupler Angle
60
0.5
55
0
50
-0.5
45
-1
Filleted-Rhombus Path Filleted-Rhombus Path
40
fixedpivot at S =6.1899units -1.5 fixedpivot at S =6.1899units
100 150 200 250 300 350 400 450 100 150 200 250 300 350 400 450
Crank Angle Crank Angle
a b
Coupler Angular Acceleration vs. Crank Angle Rocker Angular Acceleration vs. Crank Angle
Rocker Angular Acceleration rad/s2
Coupler Angular Acceleration rad/s2
10 Filleted-Rhombus Path
10
8
Filleted-Rhombus Path 8 fixedpivot at S =6.1899units
fixedpivot at S =6.1899units
6 6
4 4
2 2
0 0
-2 -2
-4 -4
-6 -6
-8 -8
-10 -10
100 150 200 250 300 350 400 450 100 150 200 250 300 350 400 450
Crank Angle Crank Angle
c d
Fig. 5 Kinematic parameters versus crank angle

are almost the same. Figure 5c, d shows that the maximum magnitude of coupler
acceleration and rocker acceleration are less than non-adjustable mechanism.
6 Conclusion
Precise filleted-rhombus path generation mechanism is synthesized. Its path points

for one degree crank rotation are extracted. Kinematic analysis performed for the
mechanism to generate filleted-rhombus path. Adjustable link length curve required
to generate filleted-rhombus path does not show any sharp turning points. This obser-
vation motivates us to explore the possibility of converting this mechanism to achieve
one degree of freedom with the help of a cam-link.
References
1. Chong P, Sodhi R (2010) Optimal synthesis of adjustable mechanisms generating multiphase

2. Kay F, Haws R (1975) Adjustable mechanisms for exact path generation. ASME J Eng Ind
97(2):702–707
6. Russell K, Sodhi RS (2005) On design of slider-crank mechanisms Part II: multi-phase path
and function generation. Mech Mach Theory 40:301–317
7. Mundo D, Liu JY, Yan H (2006) Optimal synthesis of cam-linkage mechanisms for precise
Experimental Investigations of Initial
Push Forces on an Industrial Trolley
Masepogu Wilson Kumar, Gadudasu Babu Rao, K. Balasubramanian,

and I. Kantharaj
Abstract Manual material handling using push-pull trolleys in small- and medium-
scale enterprises is an inevitable activity. A study carried out on a push-pull trol-
ley using push-pull force gauge, surface electromyography, electro-goniometer, and
heart rate monitor helped the researchers to find an appropriate handle height of a
trolley for industrial use. Central composite design, a response surface methodol-
ogy, was used for designing experiments. These experiments were carried out in a
laboratory using five subjects with stature 1650, 1730, 1740, 1750 and 1820 mm,
at 125 kg load under five different heights of handle, viz. 900, 950, 1000, 1050 and
1100 mm. A 35% of reduction was found in the initial forces after using 1100 mm
handle height by the subjects during the pushing activity.
Keywords Manual material handling · Push-pull forces · Handle height · Trolley ·

Central composite design (CCD)
1 Introduction
The use of poorly designed manual material handling (MMH) equipments like push-
pull trolleys may consume extreme energy from the industrial workers. The efforts of
the industrial workers can be reduced in the tasks, especially in pushing and pulling
the heavily loaded trolleys by providing ergonomically designed manual material
handling assisting devices. Height adjustable handles for the MMH carts or trolleys
can be of great use to all the industrial workers.
MMH tasks such as lifting or lowering, carrying and pushing or pulling are physi-
cal work activities that involve the exertion of considerable force because a particular
load is heavy or the cumulative loads during working days are heavy [1]. Pushing and
pulling tasks can be ergonomically evaluated at the workplace through the assess-
ment of the magnitude of forces exerted by the subject/worker on the trolley. The
push or pull forces are often distinguished as initial forces which are required to
M. Wilson Kumar (B) · G. B. Rao · K. Balasubramanian · I. Kantharaj

Karunya Institute of Technology and Sciences, Coimbatore, Tamil Nadu 641114, India

https://doi.org/10.1007/978-981-15-4488-0_13
150 M. Wilson Kumar et al.
move the object from the initial position, whereas the sustained forces are required
to keep the object in constant or varied speed or velocity. The ending forces may be
equal or more like initial forces to stop or bring the trolley to halt [2]. The laboratory
experiments investigated the maximum acceptable forces for a particular population
using handle heights, distance and frequencies matching to different pushing tasks
from a psychosocial point of view [3]. Laboratory experiments conducted on a cart
with two different moving speeds 1.8 and 3.6 km/h and three different heights of
exertion 660, 1090 and 1520 mm found that pulling a cart ended up in more back
loading than pushing. Body weight of subject ranging from 50 to 80 kg affected the
lower-back loadings more significantly in pulling for forces 98, 198 and 294 N (50%
increase as body weight increase from 50 to 80 kg) than in pushing (25% increase)
[4].
Pushing and pulling should be carried out on an ergonomically well-designed
cart/trolley for transferring heavy loads. These redesign strategies including the var-
ious MMH tasks and equipment will positively reduce the costs involved in com-
pensation claims [5]. The laboratory experiments conducted on the trolley with five
female and five male university students, two trolley loads 73 and 181 kg revealed
that the initial forces were upward for the shoulder height and downward for knuckle
height, whereas the least vertical forces were found to be at elbow height during
pushing and pulling on the carpet surface. Here, the most direct horizontal push
occurred at the elbow height after using three different handle heights [6].
Maximum acceptable forces tables were developed after conducting investigations
on various types and frequencies of repetitive wrist motion including wrist flexion
with a power grip, wrist flexion with a pinch grip and wrist extension with a power grip
[7]. Force exertions and muscle activities were analysed while operating a manual-
guided vehicle using two 3D load cells for measuring push-pull forces and a wireless
EMG system to measure the eight muscle groups involved in cart pushing and pulling
activities. It was found that the force direction and handle height variations are
significant in most cases except for ending force and for the EMG of the anterior
deltoid [8].
Peak push forces reached 200 N for females and 500 N for male subjects when
they pushed at different heights for the loads ranging from 45 to 450 kg using an
adjustable handle height with a load cell attached to a typical industrial cart [9].
Modified EMG-assisted biomechanical model was designed for evaluating lifting
tasks to evaluate push-pull tasks, and person-specific biologically assisted models
were developed to understand the spine loading during various push-pull conditions
[10].
From the information given above, it is required to find the suitable handle height
of the trolley based on the anthropometry of the local population. The work reported
in this paper is an experimental study using different handle heights. Five university
students were selected as subjects for conducting the experiments in the laboratory on
a four-wheeled trolley, and the best suitable handle height is reported here. This study
used heart rate monitor, surface electromyography (sEMG) and electro-goniometer
for validating the results.
Experimental Investigations of Initial Push Forces … 151
2 Materials and Methodology
2.1 Subjects
The age, height and weight of five male subjects are given in Table 1. They have
participated in the laboratory experiments on the push-pull trolley. These subjects
were healthy and reported to have no history of musculoskeletal disorders.
2.2 Trolley
A four-wheeled push-pull trolley was used to conduct the experiments as shown

in Fig. 1. It has an adjustable handle height ranging from 900 to 1100 mm. The
trolley was fitted with four polyurethane caster wheels of 100 mm diameter and
Table 1 Age, height and

Subject number Age (yrs) Stature (mm) Weight (kg)
weight of the subjects
1 19 1650 68
2 19 1730 59
3 18 1740 61
4 19 1750 85
5 19 1820 58
Fig. 1 sEMG and

goniometer sensors attached
to the subject; push-pull
gauge attached to the
trolley’s horizontal handle
bar
38 mm width. The front wheels were non-swivelling, but the two hind wheels near
the handle side are 360° swivelling.
All the four wheels were oriented in the forward direction before the starting of
each trial. The handle was firmly welded to the horizontal base structure at an angle
of 110°. The push-pull force gauge was fixed to the trolley handle, whereas the EMG
sensors, twin-axis flexible goniometer and polar heart monitor were fixed on the
subject.
2.3 Force Measurement
A SHIMPO FGV-1000 HX push-pull gauge was used to measure the pushing force
with an accuracy of ±0.2% full-scale measures up to 5000 N with a sample rate
of 1000 times per second. It has a four-digit LCD display which is 180° reversible.
The peak values can be viewed by selecting a peak button. An M10 hexagonal nut
(internal threaded) was welded at the centre of the horizontal handle bar of the trolley
to which the force gauge was fastened.
2.4 Heart Rate Measurement
Polar RS100 heart rate monitor which measures, displays and also records the heart
rate was used for continuous recording during the trials. The elastic strap holds the
transmitter (detects the heart rate) around the subject’s chest. The wrist unit and
chest strap were worn by every subject before the start of each trial. The wrist unit
needs to be switched on just before the trial run and to be stopped when the subject
reaches the finishing point of the 15,000 mm distance and the average heart rate for
that trial to be recorded. This device works with an accuracy of ±0.5 s at 25 °C/77°F
temperature and also an accuracy of ±1% or 1 bpm, during steady-state condition
having a measurement range between 15 and 240 beats/minute (bpm).
2.5 Surface Electromyography (SEMG)
Precision bipolar EMG sensors SX230, Biometrics Ltd. London, UK, were used in
this study. Due to the high quality of the SX230 electronics, with this little skin
preparation, an ultra-high-quality signal was obtained without the use of conductive
gels or creams. Highly flexible grade, 1.25-m-long ground reference R306 was used
for grounding purpose. These sensors were placed over the bulk muscles, viz. flexor
digitorum, deltoid and upper trapezius parallel to the longitudinal axis of the muscle
fibres on each subject. The output was square-rooted for each average value. The
EMG signals were acquired real time and were saved in the micro SD of the DataLOG
unit and were simultaneously transmitted via Bluetooth to the laptop.
2.6 Flexible Goniometer
The twin-axis flexible electro-goniometer SG65 was attached on the subject’s dorsal
surface. The telescopic block was fixed over the third metacarpal and the distal
block over the midline of the forearm, with the wrist in the neutral position. This
sensor can simultaneously measure the angles up to two planes of movement, viz.
wrist flexion/extension and radial/ulnar deviation and works with an accuracy of ±2°
measured over 90° from the neutral position and has better than ±1° repeatability.
2.7 DataLOG
The portable DataLOG MWX8 unit (small, light weighted and battery operated) was
used to acquire experimental data. This device has a colour graphics LCD, joystick,
micro-SD card interface, accommodates eight analogue lead wires and two digital
wires simultaneously with a real-time wireless data transfer via Bluetooth linked to
a laptop facilitating an automatic backup to the micro-SD card.
2.8 Biometrics Analysis Software v8.51
Biometrics software v8.51 was used to save and analyse the raw data that was acquired
in real time through DataLOG. Prior to the starting of trials, the connectivity of the
DataLOG and sensors were inspected for its readiness. The channels used were
zeroed, and sensitivity was set for the goniometer and EMG. At the end of each trial,
the acquired raw data was saved in the laptop computer for further analysis. The
SX230 EMG pre-amplifier has the capacity in detecting low-level signals in a noisy
environment. A high-pass filter was used to remove the DC offsets due to membrane
potentials, and a low-pass filter was used to filter the unwanted frequencies above
450 Hz. The EMG raw data was filtered using an RMS filter for a better amplitude
values, and using the vertical markers, the first four seconds of time was selected
for measuring the muscle activity generated during the initial pushing in millivolts
(mV).
2.9 Experimental Design
Design-Expert DX9 software was used to formulate the design of experiments. Cen-
tral composite design (CCD) of RSM method was used to study the effect of several
variables influencing the output responses by varying them simultaneously and car-
rying out a limited number of twenty experimental combinations as shown in Table 1.
A three factorial with five-level CCD matrix was used to design the experiments to
predict the effects of the handle height and load in the trolley on push force exerted by
the subjects. Each trial was conducted thrice, and an average was taken for prediction.
2.10 Task and Procedure
Electrical activity generated within the muscles was measured using surface EMG
electrodes placed on the left-arm flexor digitorum of the subjects. The experimental
procedures were explained to the subjects, and written consent was also obtained.
Three trials were performed for each set of the run (combination), and the average
of the three trials was taken as the input data in the CCD matrix of the DOE software
as given in Table 2.
Each subject has to push the trolley for three times at his own walking pace and
at every handle height for fifteen metres. They were alerted by a human voice for
each trial. The subject has to stop on his own when he reaches the finish line. All
the initial pushing forces were recorded and saved in the system. An average of the
three trials was calculated and analysed further in the DoE software.
3 Results and Discussions
3.1 Initial Pushing Force Analysis
Figure 2 shows the initial pushing force exerted by five male subjects at five different
handle heights when the total weight of the trolley was 125 kg (weight of the empty
trolley 50 kg and load in the trolley is 75 kg). The maximum force recorded by all
the subjects was in the range from 110 to 125 N when the height of the handle was at
900 mm. The exerted pushing force gradually decreased as the height of the handle
increased from 900 to 1100 mm. It was found that the pushing force exerted by all
the subjects was ranging from 80 to 95 N at the handle height of 1100 mm resulting
in a reduction of 25.5% of the effort required to push the trolley as the handle height
has to be below the elbow height. It also shows that Subject 1 exerted a force of
118 N when compared to Subject 5 who exerted a force of 125 N at the handle height
900 mm. The force exertion gradually decreased as the handle height increase from
900 to 1100 mm.
Table 2 Design matrix coded factors with corresponding values

Run number Design matrix coded factors Subject Handle Trolley load
Factor 1 Factor 2 Factor 3 height (mm) height (mm) (kg)
1 1 −1 1 1780 950 169

2 1 −1 −1 1780 950 106
3 −1 1 1 1690 1050 169
4 0 2 0 1740 1100 138
5 −1 1 −1 1690 1050 106
6 0 0 0 1740 1000 138
7 0 0 0 1740 1000 138
8 −1 −1 1 1690 950 169
9 0 0 0 1740 1000 138
10 1 1 −1 1780 1050 106
11 2 0 0 1820 1000 138
12 0 0 0 1740 1000 138
13 0 0 −2 1740 1000 75
14 −2 0 0 1650 1000 138
15 −1 −1 −1 1690 950 106
16 0 −2 0 1740 900 138
17 0 0 0 1740 1000 138
18 1 1 1 1780 1050 169
19 0 0 2 1740 1000 200
20 0 0 0 1740 1000 138
Fig. 2 Initial push force (N)

on 100 mm wheel at load
125 kg
3.2 Heart Rate Analysis
Figure 3 shows the heart rate of all the five subjects experienced at five different
handle heights when the total weight of the trolley was 125 kg. It can be observed
that all the subjects experienced heart rate ranging between 70 and 83 beats per
Fig. 3 Heart rate (BPM)

during initial pushing of the
trolley
Fig. 4 %MVC for the

left-arm flexor digitorum
muscle during initial pushing
of the trolley on the 100 mm
wheel at load 125 kg
minute (bpm) when the handle height was set at 900 mm, and the heart rate ranges
from 60 to 78 bpm when the handle height was set at 1100 mm. Subjects 1 and 5
(the shortest and the tallest) experience a lesser heart rate compared to Subjects 2, 3
and 4 (with medium heights).
3.3 EMG Analysis
Figure 4 shows the recorded muscle activity measured in microvolts during initial
pushing for the five subjects at five different handle heights. It was found that the
muscle force was higher for the Subjects 4 and 5 when the handle height is 900 mm.
For Subjects 1 and 2, the maximum occurs at the handle height 950 mm, and for
Subject 3, the maximum occurs at handle height 1000 mm. It was also found that
the muscle activity for all the subjects decreased when the height of the handle is
1100 mm.
3.4 Goniometer Analysis
Figure 5 shows the wrist extension measured using the flexible goniometer for all
the five subjects at five different handle heights. The wrist extensions for all the five
Fig. 5 Wrist extension (°) of

left-arm wrist and forearm
during initial pushing of the
trolley on 100 mm wheel at
load 125 kg
subjects were found to be about 16 degrees when the handle height was 900 mm. It
was observed that the wrist extensions have gradually decreased (8–12) degrees with
the increase of the handle height to 1100 mm. The optimal handle height should be
in between 900 and 1150 mm. Generally, the handle height should be slightly below
the elbow height. There should be at least 200 mm distance from the back edge of
the trolley to the handle in order to provide a good room for normal walking [11,
12].
It can be observed that the results of push-pull gauge, heart rate analysis, sEMG
analysis and goniometer analysis show that the handle height of 1100 mm tends to
give reduced push forces, heart rate and muscle forces for the subjects while pushing
a loaded trolley.
4 Conclusions
It was found that the usage of handle height of 1100 mm reduced 25–35% of the
pushing efforts (including muscle force and heart rate) exerted by the subjects when
compared to 900 mm handle height. It can be taken as the most comfortable height
for the selected population whose anthropometric data provides the range of height
from 1650 to 1820 mm. The forces exerted, the heart rate experienced, the muscle
activity generated and the wrist extensions were considerably low at a handle height
of 1100 mm when compared to the other handle heights.
References
1. Darcor and Ergoweb® (2001) The ergonomics of manual material handling-pushing and pulling
tasks. Darcor Casters, 7 Staffordshire Place, Toronto, Ontario, Canada M8W 1T1, www.darcor.
com Ergoweb, 93 West Main Street, P.O. Box 1089, Midway, Utah 84049 USA, www.ergoweb.
com
2. Jansen Jorrit P, Marco JM, Hoozemans AJ, van der Beek MHW, Frings-Dresen MH (2002)
Evaluation of ergonomic adjustments of catering carts to reduce external pushing forces. App
Ergon 33:117–127
3. Snook SH, Ciriello VM (1991) The design of manual tasks: revised tables of maximum
acceptable weights and forces. Ergonomics 34:1197–1213
4. Lee KS, Chaffin DB, Herrin GD, Waikar AM (1991) Effect of handle height on lower-
backloading in cart pushing and pulling. Appl Ergon 22(2):117–123
5. Ciriello VM, Snook SH (1999) Survey of manual handling tasks. Int J Ind Ergon 23:149–156
6. Al-Eisawi KW, Kerk Congleton J, Amendola A, Jenkins OC, Gaines WG (1999) The effect
of handle height and cart load in the initial forces in cart pushing and pulling. Ergonomics
42:1099–1113
7. Snook SH, Vaillancourt DR, Ciriello VM, Webster BS (1995) Psychophysical studies of
repetitive wrist flexion and extension. Ergonomics 38(7):1488–1507
8. Lin C-L, Chen M-S, Wei Y-L, Wang M-JJ (2010) The evaluation of force exertions and muscle
activities when operating a manual-guided vehicle. Appl Ergon 41:313–318
9. Resnick ML, Chaffin DB (1995) An ergonomic evaluation of handle height and load in maximal
and submaximal cart pushing. Appl Ergon 26(3):173–178
10. Theado EW, Knapik GG, Marras WS (2007) Modification of an EMG-assisted biomechanical
model for pushing and pulling. Int J Ind Ergon 37:825–831
11. Wilson Kumar M, Vijay SJ, Vasudev KL, Darius Gnanaraj S (2017) Reduction of discomfort
in pushing an industrial trolley using ergonomics. IOP Conf Ser Mater Sci Eng 263:062042
12. Masepogu WK, Sahu M, Vijay SJ (2020) Experimental investigations on the effect of wheel
size on an industrial trolley. In: Advances in materials and manufacturing engineering. Lecture
notes in mechanical engineering, Springer, pp 557–564
Mathematical Analysis of Stiffness
of Orthotropic Beam with Hollow
Circular and Rectangular Cross-sections
Michael C. Agarana, Esther T. Akinlabi, and Anuoluwapo M. Olanrewaju
Abstract Stiffness and strength of mechanical materials have to be thoroughly

investigated to avoid functional and physical failures. This study investigates, analyt-
ically, the stiffness of wood beam, which is an example of orthotropic material, with
different cross-sectional areas. The moment of inertia for each type of wood, consid-
ered, was analysed. The axial stiffness and bending stiffness were also computed for
different cross-sectional areas. Comparisons were made and results are consistent
with the ones in the literature. Specifically, it was observed that the cross-sectional
areas of wood have a lot of influence on its mechanical properties. The axial stiff-
ness, bending stiffness, shear stresses and moment of inertia of the orthotropic solid
materials, considered, depend on their cross-sectional areas, to a large extent.
Keywords Stiffness · Orthotropic beam · Mathematical analysis · Hollow

cross-sections
1 Introduction
In material science and solid mechanics, orthotropic materials have material proper-
ties that differ along three mutually-orthogonal twofold axes of rotational symmetry.
They are a subset of anisotropic materials because their properties change when
M. C. Agarana (B) · A. M. Olanrewaju

Department of Mathematics, Covenant University, Ota, Ogun, Nigeria
A. M. Olanrewaju
M. C. Agarana · E. T. Akinlabi
Department of Mechanical Engineering Science, University of Johannesburg, Auckland Park
Kingsway Campus, P.O. Box 524, Johannesburg 2006, South Africa
E. T. Akinlabi
Department of Mechanical Engineering, Covenant University, Ota, Ogun, Nigeria

https://doi.org/10.1007/978-981-15-4488-0_14
160 M. C. Agarana et al.
measured from different directions. So every orthotropic material is anisotropic. A

familiar example of an orthotropic material is wood [1, 2].
Stiffness is the ability of the material to resist deformation or deflection (functional
failure). Understanding the roles that stiffness play is essential to the decision process
when choosing foundational support to minimise risk [2, 3]. Stiffness is different from
strength. Strength is a measure of the anxiety that can be connected to a material
before it for all time disfigures (yield quality) or breaks (rigidity). In the event that
the forced stress is not as much as the yield strength, the material comes back to
its unique shape when the stress is expelled. On the off chance that the connected
anxiety surpasses the yield quality, plastic or perpetual twisting happens, and the
material can never again come back to its unique shape once the load is evacuated
[3–5].
A polygon is a two-dimensional shape with no less than three sides that are straight
and create interior angles where they meet. Polygons can be regular or irregular. A
regular polygon is a polygon with all sides of equivalent length and every interior
angle of equivalent measure. An irregular polygon, on the other hand, is a polygon
that is not regular. That is, an irregular polygon is a polygon that does not have all sides
of equivalent lengths and all interior angles equal. Polygons can have any number of
sides insofar as there are at least three of them. A lot of n-sided polygons have special
names [5–8]. A polygon with three sides is a triangle. A polygon with four sides is
known as a quadrilateral and one with five sides is a pentagon. A polygon with six
sides is a hexagon, one with seven sides is known as a heptagon, and a polygon with
eight sides is an octagon, and so on [8, 9].
The difference between bending moment and moment of inertial is as follows: the
reaction induced in a structural element when an external force or moment is applied
to the element causing the element to bend is referred to as the bending moment,
while moment of inertia is the capacity of a cross-section to resist bending [9–12].
Rigidity of an object, otherwise known as bending stiffness, is the measurement of
the extent to which the member resists bending deformation in response to an applied
load. This paper analyses, mathematically, the stiffness and strength of wood, an
orthotropic material, with a regular diagonal cross-sectional area of different number
of vertices.
2 Formulation of Problem
σ (ε) F/A F L0
E ≡ σ (ε)ε = F/AL/L 0 = F L 0 AL E ≡ = =
ε L/L 0 AL
Strain can be expressed as
ξ = L/L (1)
Mathematical Analysis of Stiffness of Orthotropic Beam … 161
where
ξ strain (m/m) (in/in)
L elongation or compression (offset) of the object (m) (in)
L length of the object (m) (in).
Stress can be expressed as
σ (ξ ) = F/A (2)
where
σ (ξ ) stress (N/m2 , lb/in2 , psi)
F force (N, lb)
A area of object (m2 , in2 ).
Young’s modulus (E) can be expressed as
E stress/strain.
σ (ξ ) F L0 F L0
E= = . = (3)
ξ A L AL
where
σ (ξ ) is the tensile stress;
ξ is the engineering extentional strain;
E is the Young’s modulus (modulus of elasticity);
F is the force exerted on an object under tension;
A is the actual cross-sectional area, which equals the area of the cross-section
perpendicular to the applied force;
L is the amount by which the length of the object changes (L is positive if the
material is stretched and negative when the material is compressed);
L0 is the original length of the object.
The Young’s modulus of a material can be used to calculate the force it exerts
under specific strain
E AL
F = E AL L 0 F =
L0
F is the force exerted by the material when contracted or stretched by L. The
force F is given as
E AL
F= (4)
L0
The axial stiffness is given as

AE
k= (5)
L
The bending stiffness is given as
F
K = (6)
w
where
F is the applied force
w is the deflection.
Substituting Eq. (6) into (8) gives
E AL
L0 E AL
K = = (7)
w wL 0
The moment of inertia of the wood beam, along the x-axis, with different cross-
sections are given as follows:
π(d)4
Ixc = (8)
64
1 3
Ixr = bh (9)
2
π
Ixch = (d2 )4 − (d1 )4 (10)
64
1 1
Ixr h = B H 3 − bh 3 (11)
12 12
where
d is thediameter of the circular cross-section
b is the horizontal distance of the cross-section
h is the height of the cross-section
d2 is the diameter of the outer circle of the cross-section
d1 is the diameter of the inner circle of the cross-section
B and H are the horizontal distance and height of the outer rectangle of the cross-
section
B and h are the horizontal distance and height of the inner rectangle of the cross-
section
I xc is the moment of inertia of wood beam with a circular cross-section.
I xr is the moment of inertia of wood beam with a rectangular cross-section.
I xch is the moment of inertia of wood beam with a circular hollow cross-section.
I xrh moment of inertia of wood beam with a rectangular hollow cross-section
3 Numerical Example
Considering uniform blocks of wood beam with the following cross-sectional areas:
circular, rectangular, hollow circular and hollow rectangular, stiffness of such struc-
tures of wood is calculated and compared. Different values length (L), cross-sectional
area (A) and value of Young’s modulus of wood, were used to carry out the numer-
ical analysis of the stiffness of block of wood beam with different cross-sectional
areas and lengths. For brevity, only the cases of Oak wood beam with four different
cross-sections were considered in this paper.
The moment of inertia of the wood hollow circular beam is obtained using Eq. (10),
converting inches to meters, as follows:
π
Ixch = (d2 )4 − (d1 )4 = ≈ 0.0000008549 kg m2 (12)
64
The axial stiffness of the wood hollow circular beam is obtained by adopting
Eq. (5) and taken the Young’s modulus of Oak wood beam as 12,300 MPa:
A2 E − A1 E
k=
L
E E
= πr22 − πr12
L L
22 E 2
= r2 − r12
7 L
2 2
22 12,300 5.0 3.6
= × 0.03 − × 0.03
7 40 × 0.03 2 2
≈ 87.27 N/m (13)
The bending stiffness of the wood hollow circular beam is obtained by adopting
Eq. (6) and taken the Young’s modulus of Oak wood beam as 12,300 MPa:
E A2 L E A1 L L E
K = − = (A2 − A1 )
wL 0 wL 0 wL 0
(14)
The deflection w is given as
L3 F
w= (15)
3E Ix
Substituting Eq. (4) into (15) gives
L 3 E AL
L0
w= (16)
3E Ix
Substituting Eq. (16) into Eq. (14) and simplifying gives
3E Ix
K = (17)
L3
Substituting the value of the parameters gives
3(12,300)(0.0000008549)
K =
(40 × 0.03)3
≈ 0.02 Nm2 (18)
The moment of inertia of the wood hollow rectangular beam cross-section is

obtained using Eq. (11):
1 1
Ixr h = B H 3 − bh 3
12 12
1 1
= 6(0.0254).{12(0.0254)}3 − 4(0.0254).{10(0.0254)}3
12 12
≈ 0.0002222 kg m2 (19)
The axial stiffness of the wood hollow rectangular beam, whose cross-section can
be obtained by adopting Eq. (5) and taken the Young’s modulus of Oak wood beam
as 12,300 MPa:
The area of the outer rectangle A2 is
12(0.0254) × 6(0.0254)
= 1.83 m2 (20)
The area of the inner rectangle A1 is
10(0.0254) × 4(0.0254)
= 1.02 m2 (21)
Therefore,
A2 E − A1 E
k=
L
E(1.83 − 1.02)
=
0.03(40)
12,300(0.81)
=
1.2
≈ 8.3 N/m (22)
The bending stiffness of the wood hollow rectangular beam with length 40 inches,
whose cross-section can be obtained by adopting Eq. (17) and taken the Young’s
modulus of Oak wood beam as 12,300 MPa:
3E Ix
K =
L3
3(12,300)(0.0002222)
=
(40 × 0.03)3
≈ 4.74 Nm2 (23)
Now, in order to compare with the hollow solid shapes, the moment of inertia,
axial stiffness and bending stiffness of both wood beam with a circular cross-section
and wood beam with rectangular cross-section are considered in this section. The
moment of inertia of the wood circular beam, assuming the hollow is not there, is
obtained using Eq. (8), and converting inches to meters, as follows:
π(d)4
Ixc =
64
22 (5.0 × 0.03)4
=
7 64
≈ 0.00002486 kg m2 (24)
The moment of inertia of the wood rectangular beam, assuming the hollow is not
there, is obtained using Eq. (9), and converting inches to meters, as follows:
1 3
Ixr = bh
2
1
= (6 × 0.03)(12 × 0.03)3
2
≈ 0.004199 kg m2 (25)
The axial stiffness of the wood circular and wood rectangular beam, assuming
there is no hollow, are obtained by adopting Eq. (5) and taken the Young’s modulus
of Oak wood beam as 12,300 MPa:
For circular cross-section
AE
k=
L
πr 2 (12,300)
=
(40 × 0.03)

22 5 2 12,300
=
7 2 1.2
≈ 201339.29 N/m (26)
For rectangular cross-section
AE
k=
L
(12 × 6 × 0.03)(12,300)
=
(40 × 0.03)
≈ 22.14 N/m (27)
The bending stiffness of the wood circular and rectangular beam with length 40
inches, whose cross-section, assuming the hollows are not there, can be obtained by
adopting Eq. (20) and taken the Young’s modulus of Oak wood beam as 12,300 MPa:
For Circular cross-section
3E Ix
K =
L3
3(12,300)(0.00002486)
=
(40 × 0.03)3
≈ 0.53 Nm2 (28)
For Rectangular cross-section
3E Ix
K =
L3
3(12,300)(0.004199)
=
(40 × 0.03)3
≈ 89.67 Nm2 (29)
From the analysis result values, depicted in Table 1, it can be seen that the value of the
Table 1 The stiffness and moment of inertia for Oak wood beam with different cross-sections
S/N Cross-section I xc (kg m2 ) I xr I xch (kg m2 ) I xrh k (N/m) K
(kg m2 ) (kg m2 ) (Nm2 )
1 Circular – – 0.0000008549 – 87.27 0.02
hollow
2 Rectangular – – – 0.0002222 8.3 4.74
hollow
3 Circular 0.00002486 – – – 201,339.29 0.53
4 Rectangular – 0.004199 – – 22.14 89.67
moment of inertia for both the rectangular cross-section (I xr ) and rectangular hollow
cross-section (I xrh ) are greater than that of the circular cross-section (Ixc ) and circular
hollow cross-section (I xch ). This implies that the wood beam with a rectangular cross-
section can resist rotational force than the wood beam with a circular cross-section.
This also suggests that the stress in the wood circular beam is less than that in the
wood rectangular beam, going by the formula for stress of a beam: The axial stiffness
(k) value for both circular and circular hollow cross-sections is higher than that of
rectangular and rectangular hollow cross-sections. This implies that more force is
required to produce unit axial deformation in the wood circular beam than in the
wood rectangular beam. However, the value of the bending stiffness (K) for both
rectangular cross-section and rectangular hollow cross-section is higher than that of
circular and circular hollow cross-sections. This implies that the former is more rigid
than the later. It shows that the resistance of a member against bending deformation is
higher in the wood rectangular beam than in wood circular beam. It is also observed
that the moment of inertia of all the cross-sections considered in this paper is less
than one (I xch < I xrh < I xc < I xr < 1). The wood beams without hollow cross-section
seem to resist rotation than the ones with hollow cross-section.
5 Conclusion
This study set out to analyse mathematically the stiffness of orthotropic beam of
different cross-sections. Wood, an example of orthotropic material, was considered.
Also, Oak, a type of wood, with Young’s modulus of 12,300 MPa, was considered.
The moment of inertia, axial stiffness and bending stiffness for an Oak wood beam
with different cross-sections were analysed. The special cross-sections considered
are circular cross-section, rectangular cross-section, hollow circular cross-section
and hollow rectangular cross-section. This study revealed that the type of cross-
section of an Oak wood beam affects its stiffness. It was noticed among the types
of cross-sections considered, that an Oak wood beam with rectangular cross-section
has the highest bending stiffness, and the one with hollow circular cross-section has
the least bending moment. Also, the highest moment of inertia goes to the Oak wood
beam with rectangular cross-section, and the lowest goes to the beam with the hollow
circular cross-section. Therefore this would help a structural engineer to make an
informed decision on the type of wood beam to use in construction, as regards the
cross-section, in order to minimise the risk of bending due to eternal forces.
Acknowledgements The authors hereby acknowledge the support of Covenant University and the
University of Johannesburg towards the successful completion of this study.
References
1. Agarana MC, Ede AN (2016) Free vibration analysis of elastic orthotropic rectangular inclined
damped highway supported by Pasternak Foundation under moving aerodynamic automobile.
In: Proceedings of the world congress on engineering 2016 June 29–July 1, London, UK, pp
9708–981
2. Agarana MC, Agboola OO (2015) Dynamic analysis of damped driven pendulum using laplace
transform method. Int J Math 26(3). ISSN 0974-570X (online) 0974-5718(Print)
3. Burnett C (2016) Mental strength and stiffness: what’s the difference? ThermoFisher
Scientific. https://www.thermofisher.com/blog/metals/metal-strength-and-stiffness-whats-the-
difference/
4. Agarana MC, Gbadeyan JA, Ajayi OO (2016) Dynamic response of inclined isotropic elastic
damped rectangular Mindlin Plate resting on Pasternak Foundation under a moving load. In:
Proceedings of the international multi conference of engineers and computer scientists, March
16–18, 2016, Hong Kong, pp 713–716
5. Agarana MC, Gbadeyan JA (2015) Finite difference dynamic analysis of railway bridges sup-
ported by Pasternak foundation under uniform partially distributed moving Railway vehicle. In:
Proceedings of the world congress on engineering and computer science 2015, vol II WCECS
2015, October 21–23, 2015, San Francisco, USA
6. Agarana MC, Ede AN (2016) Free Vibration analysis of elastic orthotropic rectangular inclined
978–981
7. https://en.wikipedia.org/wiki/Young%27s_modulus
8. https://physics.stackexchange.com/questions/153008/what-is-difference-between-
homogeneous-and-isotropic-material
9. http://www.engineeringtoolbox.com/young-modulus-d_417.html
10. https://en.wikipedia.org/wiki/Regular_polygon
11. https://www.mathsisfun.com/geometry/regular-polygons.html
12. Friswell MI, Adhikan S, Lei, Y (2007) Vibration analysis of Beams with non-local foundations
using the finite element method. Int J Numer Methods Eng 71:1365–1386
Analytical Investigation of the Stiffness
of Homogenous Isotropic Mechanical
Materials with Different Cross-sections
Michael C. Agarana, Esther T. Akinlabi, and Okwudili S. Ogbonna
Abstract In order to achieve the required high level of mechanical material com-
ponents as regards sustainability and safety, the stiffness of such materials has to
be thoroughly investigated to avoid failure. This study investigates, analytically, the
stiffness of some mechanical materials with different cross-sectional areas, whose
mechanical properties are homogeneous and isotropic. The exerted force for steel
material solids with different cross-sections was investigated. Comparisons were
made and results are consistent with the ones in literature. A computer software,
Maple, was used to plot the three-dimensional graphs of the relationship between
the parameters. Specifically, it was observed that the steel hollow rectangular beam
cross-section has a high axial stiffness compared to that of steel hollow circular beam
cross-section. Also, the moment of inertia of the steel beams, considered, depend on
their cross-sectional areas.
Keywords Stiffness · Homogeneous isotropic materials · Analytical

investigation · Cross sections
M. C. Agarana (B)
Department of Mathematics, Covenant University, Ota, Ogun, Nigeria
M. C. Agarana · E. T. Akinlabi · O. S. Ogbonna
O. S. Ogbonna
E. T. Akinlabi

https://doi.org/10.1007/978-981-15-4488-0_15
1 Introduction
Isotropic materials are materials with indistinguishable estimations of a property

every direction. Glass and metals are good examples of isotropic materials [1]. A
homogeneous material, on the other hand, is a material with similar properties at
each point; it is uniform without inconsistencies. Materials can be both homogeneous
and isotropic, in terms of properties; however, these two have diverse implications
as regards their property of body and heading of the position [2]. The distinction
between the two is that homogeneous material has a similar group of properties
at each place, but an isotropic material has a similar-looking in the majority of
the bearings at various purposes of the property. For example, steel demonstrates
isotropic behaviour, although its microscopic structure is non-homogeneous [3–5].
Strain is “deformation of a solid due to stress”—change in dimension divided by
the original value of the dimension. Stress is force per unit area. There are different
types of stress [5–9]: Tensile stress, Compressive stress and Shearing stress. Stiffness
is the rigidity of an object. Young’s modulus, otherwise called the flexible modulus,
is a measure of the solidness of strong material. It is a mechanical property of direct
flexible strong materials. [10, 11]. The measure of the stiffness of solid material is
known as elastic modulus or Young’s modulus. A material with a very high Young’s
modulus can be approximated as rigid. Young’s modulus is the ratio of stress to strain
[12, 13]. In beams, the area moment of inertia can be used to predict deflection.
In this study, the stiffness of such block of steel in the form of beam with circular
hollow and rectangular hollow cross-sections was investigated. In particular, the axial
stiffness, Bending stiffness and moment of inertial, for beam with circular hollow and
rectangular hollow cross-sectional areas, were analyzed. Also, the second moment
of area of these solids was investigated.
The force F, exerted by the material when displaced by L can be written as follows:
E AL
F= (1)
L0
This is achieved by using Young’s modulus.

The axial stiffness is given as
AE
k= (2)
L
where
A the cross-sectional area,
Analytical Investigation of the Stiffness of Homogenous … 171
E Young’s modulus,
L the length of the element.
The bending stiffness is given as:
F
K = (3)
w
where
F the applied force,
w the deflection.
Substituting Eqs. (1) into (3) gives
E AL E AL
K = L0
= (4)
w
wL 0
3 Numerical Example
The moment of inertia of the steel hollow circular beam is obtained, converting inches
to meters, as follows:
π
Ix = (d2 )4 − (d1 )4 ≈ 0.0000008549 kgm2 (5)
64
where
d 2 = 0.20055 (the longer diameter)
d 1 = 0.20000 (the shorter diameter).
The axial stiffness of the steel hollow circular beam is obtained by adopting Eq. (5)
and taken the young’s modulus of the beam as 209 Mpa:

A2 E − A1 E 22 0.127 2 22 0.091 2
k= = 0.98 E− E ≈ 1.27 N/m (6)
L 7 2 7 2
The bending stiffness of the steel hollow circular beam is obtained by adopting
Eq. (6) as follows:
E A2 L E A1 L L E
K = − = (A2 − A1 ) (7)
wL 0 wL 0 wL 0
Substituting the deflection, w, and some manipulation gives:

3E Ix
K = (8)
L3
Substituting the value of the parameters gives:
3(209)(0.000001)
K = ≈ 0.0006 Nm2 (9)
1.0163
The moment of inertia of the steel hollow rectanguar beam cross-section is
obtained as follows:
1 1
Ix = B H 3 − bh 3 ≈ 0.0002222 kg m2 (10)
12 12
where
B the breadth of the outer part of the hollow rectangle
B the breadth of the inner part of the hollow rectangle
H the height of the outer part of the hollow rectangle
h the height of the inner part of the hollow rectangle.
The axial stiffness of the steel hollow rectangle beam was calculated to obtain:
A2 E − A1 E E(1.83 − 1.02)
k= = ≈ 166.62 N/m (11)
L 0.0254(40)
The bending stiffness of the steel hollow rectangular beam was obtained:
3E Ix 3(209)(0.0002222)
K = = ≈ 0.13284 Nm2 (12)
L3 1.0163
From the analysis, it can be gathered that the value of the moment of inertia and
the bending stiffness of both steel hollow circular beam and steel hollow rectangular
beam, under consideration in this paper, is less than one. However, the moment of
inertia of the steel hollow rectangular beam is greater than that of steel hollow circular
beam. This suggests that the stress in the steel hollow circular beam is less than that
in the steel hollow rectangular beam, going by the formula for stress of a beam:
MY
∂= (13)
I
where σ is the stress, M is the internal moment, y is the distance from the neutral axis
and I is the area moment of inertia. Both the axial stiffness and bending stiffness of
the steel hollow rectangular beam are greater than that of steel hollow circular beam.
This implies that the former is more rigid than the later. It shows that the resistance
of a member against bending deformation in higher in the steel hollow rectangular
beam than in steel hollow circular beam. Also, more force is required to produce
unit axial deformation in of steel rectangular circular beam than in of steel hollow
circular beam. From Fig. 3, the moment of inertia of the steel hollow circular beam
increases as the inner diameter decreases and outer diameters increase. This implies
the smaller the hollow in the beam the higher the moment of inertia which suggests
less stress on the beam (Figs. 1, 2, 4, 5 and 6).
Fig. 1 Show 3D plotting of

the relationship between Ix,
d2 and d1 for the circular
hollow beam
Fig. 2 Show 3D plotting of

the relationship between k, A
and L for the circular hollow
beam
Fig. 3 Showing 3D plotting

of the relationship between
K, Ix and L for the circular
hollow beam
Fig. 4 Shoing 3D plotting

I x , H, and h for the
rectanglar hollow beam
5 Conclusion
The study set out to investigate analytically the stiffness of homogenous isotropic
mechanical materials with different cross-sections. Steel circular hollow beam and
steel rectangular hollow beam were used as case studies. The moment of inertia, the
strain, the stress, the stiffness, the deflection, the Young modulus and the relationship
between these were investigated and analyzed mathematically. It was observed that
the steel hollow rectangular beam is more rigid than the steel hollow circular beam,

k, A, and I for the
rectangular hollow beam

K, I x , and L for the
rectangular hollow beam
going by the numerical values used to calculate their axial stiffness and bending
stiffness.
Acknowledgments The authors acknowledge the supports received from Covenant University and
the University of Johannesburg, where the first author is presently a post-doctoral fellow.
References
1. Agarana MC, Gbadeyan JA, Ajayi OO (2016) Dynamic response of inclined isotropic elastic
damped rectangular mindlin plate resting on Pasternak foundation under a moving load. In:
Proceedings of the international multi conference of engineers and computer scientists 2016,
Hong Kong, pp 713–716
2. Gbadeyan JA, Dada MS (2006) Dynamic response of a mindlin elastic rectangular plate under
a distributed moving mass. Int J Mech Sci 48:323–340
3. Zhang T, Zheng G (2010). Vibration analysis of an elastic beam subjected to a moving beam
with flexible connections. Am Assoc Civil Eng 136(1):120–130
4. Agarana MC, Gbadeyan JA (2015) Finite difference dynamic analysis of railway bridges sup-
ported by Pasternak foundation under uniform partially distributed moving Railway vehicle.
In: Proceedings of the world congress on engineering and computer science, vol II, October
21–23, 2015, San Francisco, USA
978–981
6. Agarana MC, Gbadeyan JA, Emetere M (2016) On response of elastic isotropic damped shear
highway bridge supported by sub-grade to uniform partially distributed moving vehicle. Int J
Appl Eng Res 11(1):244–258
7. Kumar Y (2013) Differential transform method to study free transverse vibration of monoclinic
rectangular plates resting on Winkler foundation. Appl Comput Mech 7:145–154
8. Rasian KR, Zain F, Sheer A (2013) Differential transform method for solving non-linear systems
of partial differential equations. Int J Phys Sci 8(38):1880–1884
9. Agarana MC (2010) Torsional rigidity of beams of given areas with different cross sections. J
Math Assoc Nigeria (Abacus) 37(2):117–123
10. Li R, Zhang Y, Tian B, Liu Y (2009) On the finite integral transform method for exact bending
solutions of fully clamped orthotropic rectangular thin plates. Appl Math Lett 22(12):1821–
1827
11. Agarana MC, Agboola OO (2015) Analysis of torsional rigidity of circular beams with different
engineering materials subjected to St. Venant Torsion. IMPACT: Int J Res Eng Technol 3(2):33–
46
transform method. Int J Math Comput 26(3):99–109
13. Agarana MC, Iyase SA (2015) Analysis of Hermite’s equation governing the motion of damped
pendulum with small displacement. Int J Phys Sci 10(12):364–370
Fatigue and Fracture Analysis
Fatigue Analysis of Vocal-Folds Using
Discretized Aeroelastic Model
Veda Samhitha Kanduri, James Emilian, and Venkatramani Jagadish
Abstract Voice disorders are common physiological challenges that arise largely
due to vocal fatigue, i.e. fatigue of vocal folds and associated phonatory structures.
Since the incurrence of fatigue in vocal folds directly depends on the qualitative
and quantitative nature of the dynamical signature of vocal fold oscillations, the
hitherto literature has resorted to estimating the time histories and stress histories of
the vocal fold oscillations. Often, vocal fold dynamics are rid with different types of
complexities. Therefore, the literature often resorts to the use of a mathematical model
that closely resembles the human phonatory system. Modelling the mechanism of
vocal folds as a two-mass model with 2DOF provides a reasonably accurate depiction
of the same. In order to obtain the real-time histories of the vocal fold vibrations,
a continuously changing parameter called phonation threshold pressure (PTP) is
defined so as to describe the extent of accumulated fatigue. PTP can be determined
numerically for a given set of physical and physiological parameters of the system.
Since the vocal fold vibrations are of varying amplitude as time progresses, it is
necessary to use a standard counting algorithm for fatigue analysis, like the Rainflow-
Counting algorithm. Then, the incurred damage is computed with the help of a S-N
curve and the Miner’s rule. With the time span of phonation as the control parameter
and PTP as a continuously changing parameter, the responses of the system are
systematically obtained. Later, comparisons on the fatigue damage are made.
Keywords Vocal fatigue · PTP · Rainflow-counting algorithm (RFC)
1 Introduction
Voice production involves the self-sustained, flow-induced oscillations of the vocal

folds. Vocal folds are a layered structure of three different tissues, namely, epithelium,
lamina propria and thyroarytenoid muscles. Prolonged phonation can damage these
layers due to mechanical fatigue induced. Vocal fatigue is the root cause of several
V. S. Kanduri · J. Emilian · V. Jagadish (B)

Department of Mechanical Engineering, Shiv Nadar University, Greater Noida, India

https://doi.org/10.1007/978-981-15-4488-0_16
180 V. S. Kanduri et al.
conditions, from vocal fold lesions, up to sleep apnea. After the fatigue of vocal folds
has taken place surrounding structures compensate for the reduced vocal-fold activity,
in turn, fatiguing and damaging themselves with prolonged phonation. Hence, it
is of prime importance, clinically, and economically, to study fatigue and identify
precursors to irreversible damage so that the same can be avoided.
Existing literature in the field of vocal fatigue extensively covers objective and
subjective parameters to detect fatigue, mechanisms of fatigue, etc. [1]. So far no
attempt has been made to quantify fatigue and develop a framework to predict pre-
cursors to fatigue. Here, we use the simplified two-mass model, as formulated in [2]
to extract the vibration characteristics of the vocal folds. Then, taking into consider-
ation the material properties of the multilayered vocal fold, the stress time histories
are computed. The RainFlow counting algorithm (RFC) along with the Miner’s rule
is used to compute the accumulated damage.
The organization of the paper is as follows. A brief overview of the mathematical
model is provided in Sect. 2. Section 3 presents a brief overview of vocal fatigue and
its quantification using the RainFlow-Counting algorithm and Miner’s rule. In Sect. 4
the system response over different lengths of phonation, along with the corresponding
stress histories and damage levels are presented.
2 Mathematical Model
The model used in our study is a two-mass model developed by Herzel et al. in [2],
which is a simplification of the Ishizaka and Flanagan model [3]. Herzel’s model
takes into account the basic principle of the possibility of a phase difference between
the lower and upper edge of the vocal fold which has been proven to be a neces-
sary condition for phonation. For the sake of simplicity, we have ignored the cubic
nonlinearities introduced in [3] to account for the non-linear nature of vocal fold
tissue. We also have ignored the role of subglottal and supraglottal structures and
their resonances in phonation. Despite this being a gross simplification of the actual
dynamics of phonation, we still obtain the bifurcations that are observed to occur in
excised larynges [4], and hence this approach of solving simplified equations for the
purpose of analyzing fatigue is validated.
The motions of the masses are described by the following equations:

m i ẍi + ri ẋi + ki xi + (−ai )ci (ai /2l) + kc xi − x j = Fi (x1 , x2 ) (1)
where

1, x > 0
(x) = (2)
0, x ≤ 0
Fatigue Analysis of Vocal-Folds Using Discretized Aeroelastic … 181
Fig. 1 Schematic
representation of the
two-mass model
where m i , ki , kc , ci , ri , Pi are masses, spring constants, coupling constant, additional

spring constants for collision, damping constants and pressure inside the glottis,
respectively. Here i = 1, 2 represents the lower and upper masses, respectively.
The forces Fi acting on the masses m i are given by
Fi = ldi Pi (3)
The following parameters correspond to our standard, symmetric vocal fold:

m1 = 0.125, m2 = 0.025, r 1 = 0.02, r 2 = 0.02, k 1 = 0.08, k 2 = 0.008, c1 = 3k 1 ,
c2 = 3k 2 , k c = 0.025, d 1 = 0.25, d 2 = 0.05, a01 = a02 = 0.05 and ρ = 0.0013.
All the above values are in centimeters, grams, milliseconds and their combina-
tions.
For further details on the mathematical model, refer [2].
2.1 Phonation Threshold Pressure (PTP)
PTP is defined as the minimum lung pressure to be exerted, all other conditions
remaining constant, for phonation onset (defined as loss of stability of system via
Hopf bifurcation and onset of oscillations).
Application of the Bernoulli’s equation to our model, including the assumption
of buildup of jet gives:
ρ U 2 ρ U 2
Ps = P1 + = P0 + (4)
2 a1 2 amin
Here,
amin = min(a1l , a2l ) + min(a1r , a2r ) (5)
where Ps is subglottal pressure, P0 is supraglottal pressure, U is volume flow velocity

and ρ is air density.
Accumulation of fatigue in vocal folds due to prolonged phonation changes sev-
eral characteristics of the vocal folds, and all such changes have been observed to
ultimately change PTP [5]. Since PTP has been proven to be a good indicator for
fatigue as it closely correlates with perceived exertion of phonation (PPE) [6], it is
used as the parameter that varies with time and accounts for the fatigue accumulated
in the system. Data from [6] was used to find the variation of PTP as time progresses,
during a continuous vocal loading task. Since the variation of PTP with time is not
available in the literature, we have approximated as a linear function.

t
Ps = 0.2030 + 2919 (6)
1800
where Ps is in cm of H2 O and t is in s.
The above relation is useful in our objective of setting benchmarks of PTP values
for different levels of fatigue, as presented later in this paper.
3 Vocal Fatigue and Its Quantification
The hallmark of vocal fatigue is considered to be the self-report of an increased sense

of effort with prolonged phonation. Vocal fatigue occurs due to several biomechanical
and neuromuscular factors. These include fatigue of the respiratory and laryngeal
muscles, fatigue of the non-muscular vocal-fold tissues, and changes in the viscous
properties of the vocal folds.
The potential mechanisms that contribute to vocal fatigue are
1. Neuromuscular fatigue
2. Non-muscular tissue fatigue and viscosity.
Mechanical fatigue comes under non-muscular tissue fatigue. This sort of fatigue
reflects the amount of cyclic stresses that a material can tolerate before breaking
down. Fatigue is the progressive structural damage that results from stress imposed
by strain on the material. Because of the deformation of the three layers of the
vocal folds, they are subjected to mechanical stresses with every cycle of vocal-fold
oscillation during phonation. These extended cyclic stresses result in fatigue damage
of the vocal fold tissue. Many fatigue damage theories have been used over the past
years and they indicate that fatigue damage is strongly associated with the cycle ratio,
(n i /Ni ) where n i and Ni are number of stress cycles of a specific stress amplitude
exerted on the specimen and the number of cycles to failure for that stress amplitude,
respectively.
In this model, the vocal folds are subjected to cyclic loads. Since the amplitude of
the loading is changing with time, it is difficult to determine which cycles contribute to
fatigue and their corresponding amplitudes. Hence, several cycle counting techniques
have been introduced to reduce a complicated variable amplitude loading history into
a number of discrete simple constant amplitude loading events, which are associated
with fatigue damage. And out of all these methods, Rainflow-cycle (RFC) counting
method is generally regarded as the method leading to the best estimators of ‘fatigue
life’. This is a method to determine the number of cycles present in a stress history and
along with the SN-curve, the fatigue damage is evaluated using a Palmgren-Miner
linear damage accumulation theory. Hence, RFC method breaks down any load-time
history into its constituent fatigue cycles so as to estimate the fatigue life.
The governing equations of motion are written in state-space form and solved in
MATLAB using an adaptive time-step Runge–Kutta ODE solver, to obtain time-
responses of the system for different values of the control parameter.
The governing equations can be written as:
ẋ1 = v1 (7)
1 a1
v̇1 = P1ld1 − r1 v1 − k1 x1 − (−a1 )c1 − kc (x1 − x2 ) (8)
m1 2l
ẋ2 = v2 (9)
1 a2
v̇2 = −r2 v2 − k2 x2 − (−a2 )c2 − kc (x1 − x2 ) (10)
m2 2l
2
amin
P1 = Ps 1 − (amin ) (a1 ) (11)
a1
a1 = a01 + 2lx1 (12)
a2 = a02 + 2lx2 (13)

⎧
⎨ a1 , if 0 < x1 < x2
amin = a2 , if 0 < x2 ≤ x1 (14)
⎩
0, if otherwise
The function (x) is approximated as:


tanh[50(x/x0 )], x > 0
(x) = (15)
0, x ≤0
The differential equations mentioned above are solved using ODE solver over a
time span of 1 h and the time responses are plotted. The amplitudes of the masses are
observed to be within 0.15 mm (see Fig. 2). Similarly, the time responses are plotted
for the 2nd and 3rd hour of phonation and they are observed to be within 0.2 mm and
0.3 mm respectively (see Figs. 3 and 4). As expected, a phase difference is observed
between the motion of the two masses which, as mentioned earlier, is a necessary
Fig. 2 A section of the time response obtained for 1 h of phonation

condition for phonation. In this way, the time responses can be obtained for different
time durations of phonation so as to understand the dynamics of vocal folds.
From these time responses, the corresponding stress histories are obtained using
tensile and compressive stress–strain equations. A linear relationship between stress
and strain is assumed to obtain the stress–time history:
σ = Eε (16)
where E is Young’s modulus of vocal folds in the transverse direction, for low moduli
of strains [7].
From the stress history plots above, we can see that, for 1 h of phonation, the
maximum stress amplitude is observed to be 0.0086 g/(cm) (ms)2 (see Fig. 5). In the
same way, the maximum stress amplitudes for 2 and 3 h of phonation are observed
to be 0.012 and 0.0139 g/(cm) (ms)2 (Figs. 6 and 7). Hence, it can be inferred that,
as time progresses and fatigue accumulates, the vocal folds experience a higher
magnitude of stress.
Now, our objective is to find the amount of fatigue incurred for different lengths of
continuous phonation using RFC algorithm. To apply Rainflow-Counting Algorithm,
stress-time history and S-N curve of the vocal folds are required.
Fig. 5 A section of stress history obtained for 1 h of phonation

Since existing literature does not detail the S-N curve of this multilayered,
anisotropic structure, and because calculating the same is beyond the scope of this
paper, we have used the S-N curve of the Achilles tendon [1], since it is similar in
characteristics to the vocal fold, and especially to the lamina propria, which is the
most prone to fatigue damage of the three layers that compose the vocal fold (see
Fig. 8).
Now, using the WAFO toolbox in MATLAB, the amount of damage (D) is
calculated for different lengths of phonation and is tabulated in Table 1.
As seen from the data, the damage value is higher for a longer duration of phona-
tion. The PTP value at the end of phonation is also found to be higher for a longer
Fig. 8 S-N curve of

Achilles tendon
Table 1 Amount of damage for different lengths of phonation

S. No. Duration of phonation (h) Damage index PTP (at end of phonation
duration) (cm of H2 O)
1 1 2.8501 × 10−25 3.325
2 2 4.4930 × 10−24 3.731
3 3 4.0110 × 10−23 4.137
phonatory duration. Hence, these PTP values can be linked to their respective damage
indices.
5 Concluding Remarks
This study is thus a preliminary attempt at quantifying the amount of damage incurred
in vocal folds and linking the same to a measurable, objective quantity (PTP). This
has been done as a step towards recognizing precursors to permanent, serious damage
to the vocal folds. This study hence provides a framework for quantifying fatigue for
preventive applications. To make this study directly useful in reality, a few improve-
ments are in order—a separate study is required to estimate the S-N curve of a
realistic model of the vocal folds; tissue nonlinearities have to be taken into account;
perturbations in the aerodynamic forcing have to be considered, and the role of sub-
glottal and supraglottal structures in aiding phonation have to be addressed. Future
work will focus on improving accuracy of the damage estimates by implementing
the above improvements. Further, the results of this study will prove useful when
corroborated with in vivo and ex vivo experimental studies on vocal fold fatigue, and
the authors intend to take up the same in future projects.
References
1. Edwards WB (2018) Modeling overuse injuries in sport as a mechanical fatigue phenomenon.

Exerc Sport Sci Rev 46(4):224–231
2. Steinecke I, Herzel H (1995) Bifurcations in an asymmetric vocal-fold model. J Acoust Soc Am
97(3):1874–1884
3. Ishizaka K, Flanagan JL (1972) Synthesis of voiced sounds from a two-mass model of the vocal
cords. Bell Syst Tech J 51(6):1233–1268
4. Herzel H, Berry D, Titze I, Steinecke I (1995) Nonlinear dynamics of the voice: signal analysis
and biomechanical modeling. Chaos 5(1):30–34
5. Solomon NP (2008) Vocal fatigue and its relation to vocal hyperfunction. Int J Speech-Lang
Pathol 10(4):254–266
6. Chang A, Karnell MP (2004) Perceived phonatory effort and phonation threshold pressure across
a prolonged voice loading task: a study of vocal fatigue. J Voice 18(4):454–466
7. Alipour F, Vigmostad S (2012) Measurement of vocal folds elastic properties for continuum
modeling. J Voice 26(6):816–821
Response of a Layered Composite Beam
Subjected to Static Loading Using Point
Interpolation Meshless Technique
Kunal S. Shah, Appaso M. Gadade, and Sanjiv M. Sansgiri
Abstract The present work proposes a meshless model to analyze the response of
laminated composite beam under different types of static loading. The point interpo-
lation method based on polynomial basis function is used for solving 1D higher-order
beam theory equations. The distribution of transverse shear strain and stresses along
thickness are obtained using Reddy’s shape function. A displacement response of a
laminated beam is obtained for several lamination schemes, aspect ratio, and bound-
ary conditions. The authenticity of the present algorithm is confirmed by comparing
results with different cases from literature.
Keywords Laminated composite beam · Higher-order beam theory · Meshless

method · Point interpolation method
1 Introduction
The requirement of composite materials is growing continuously in most of the

industries. Composite materials can be formulated mathematically using different
types of shear deformation theories. Computer-based analysis for solving problems
in solid mechanics, fluid dynamics or thermodynamics, etc. has increased due to
ease, accuracy, and time efficiency. In the last two decades, many authors have shown
keen interest in the numerical approach based on the meshless method, which differs
from finite element methods on the basis of approximate techniques. The domain
representation in the finite element method depends on the mesh of elements. The
K. S. Shah · S. M. Sansgiri
Mechanical Engineering, Army Institute of Technology, Pune 411015, India
S. M. Sansgiri
A. M. Gadade (B)
Mechanical Engineering Department, Thapar Institute of Engineering and Technology, Patiala
147004, India

https://doi.org/10.1007/978-981-15-4488-0_17
190 K. S. Shah et al.
shape of the element also needs to be as regular as possible to have valid results,
which is not possible in complex geometries, so the analyst has to spend a lot of
time in creating an appropriate mesh. Also, remeshing at each step is required in
case of an adaptive analysis. The accuracy of secondary or derivative variables such
as strains and stresses in solid mechanics is not adequate. The meshless methods
give an advantage over these issues. The various meshless methods with different
types of approaches and approximation techniques were reviewed [1, 2]. The point
interpolation method is one of the approximation techniques used for stress analysis
of two-dimensional solids [3, 4]. The meshless methods based on global weak forms
were described practically with MATLAB code [5].
A mathematical approach for analyzing laminated composite structures has been
developed from classical laminate plate theories to higher-order shear deforma-
tion theory [6]. The higher-order shear deformation theory present displacement
and stresses more accurately than classical or first-order theory. It accounts for the
parabolic distribution of transverse shear strain through thickness [7]. The exact
solutions of these different types of shear deformation theories were analyzed for
the bending of symmetrical and asymmetrical laminated beams [8]. The meshless
method based on multi-quadric radial basis function was used for higher-order beams
and plates analysis [9]. The radial point interpolation method was used for the analysis
of 2D laminated composite beam [10].
The present paper focuses on the behavior of static bending of the laminated beam
with respect to change in lamination scheme, boundary condition, and aspect ratio
subjected to uniformly distributed load as well as a point load.
2 General Formulation
The displacements (u, v, w) are considered along with Cartesian coordinates (x, y,
z) respectively. The displacement v is assumed to be zero. Thus (u, v, w) are the
function of x and z. In higher-order Beam Theory (HBT), the displacement field
considers a quadratic variation of transverse strain vanishing at the free surface.
Hence, shear correction factor is not required. Field variables can be represented as
axial displacement (u0 ), transverse
displacement (w0 ), rotation about normal to the
mid-surface (θx ) and slope (∂w0 ∂ x). Thus, the displacement field can be written
as [9],

∂w0
u = u 0 + zθx + αz 3 θx + ; w = w0 (1)
∂x
where
α = −4/3h 2
Response of a Layered Composite Beam Subjected to Static … 191
The linear strain vectors corresponding to displacement fields in (1) can be written
as

{ε} = ε1 ε5 (2)
where

∂u
ε1 = = ε10 + zε11 + z 3 ε13 ;
∂x

∂u ∂w
ε5 = + = ε10 + z 2 ε52 (3)
∂z ∂x
The strain–displacement relations of the HBT are

∂u 0 ∂w0
ε10 = ; ε5 = θx +
0
;
∂x ∂x

∂θx ∂θx ∂ 2 w0
ε1 =
1
; ε1 = α
3
+ ;
∂x ∂x ∂x2

∂w0
ε52 = 3α θx + (4)
∂x
Thus, vector of strain components is

{ε̄}T = ε10 ε11 ε13 ε50 ε52 (5)
and vector of field variable can be represented as

∂w0

{U } = u 0 w0 ∂x
θx (6)
The constitutive law gives a relation between stress and strain of laminated beam
as

σ1 Q 11 0 ε1
= (7)
σ5 0 Q 55 ε5
where Q i j are transformed material stiffness components of kth lamina at angle of

fiber orientation α k as described by Reddy [11].
3 Point Interpolation Method
The Û (x) is a continuous function in a domain . It can be represented by a set of

field nodes and approximated at point of interest using polynomial basis function
[12] as
n
Û T (x) = P T (x)Pm−1 U = φi u i = T
(x)U (8)
i=1
where (x) is a vector of shape functions defined by

T
(x) = P T (x)Pm−1 = φ1 (x) φ2 (x) · · · φn (x) (9)
The variation of Eq. (8), can be written as,
δÛ4×1 = 4n×4n δU4n×1 (10)
Consider a one-dimensional laminated composite beam on a domain bounded

by . The weak form of equilibrium equation can be obtained by using the Galerkin
method.

(LδU ) D(LU )d − δU bd − δU T t¯d = 0
T T
(11)
t
where L is a differential operator, b is body force vector, t¯ is traction force at the

natural boundaries and ū is displacement at the essential boundaries.
Replacing the trial functions, test functions into the weak from Eq. (11), which
provides the final discretized equation,
[K ](4n×4n) {U }(4n×1) = {F}(4n×1) (12)
where ‘n’ is number of nodes in domain around point of interest,

[K ]4n×4n = [B]T4n×5 [D]5×5 [B]5×4n d (13)

{F}(4n×1) = (4n×1) t¯d + (4n×1) bd (14)

[B]5×4n = [L]5×4 [ ]4×4n (15)
[D] is material property matrix represented as,

⎡ ⎤
[A1 ] [B1 ] [E 1 ] 0 0
N L z k
⎢ [B ] 0 ⎥
⎢ 1 [C1 ] [F1 ] 0 ⎥
⎢ ⎥
[D] = [T ]T Q [T ]dz = ⎢ [E 1 ] [F1 ] [H1 ] 0 0 ⎥ (16)
⎢ ⎥
k=1 z
k−1
⎣ 0 0 0 [A5 ] [C5 ] ⎦
0 0 0 [C5 ] [F5 ]
where
n zk
(A1 , B1 , C1 , E 1 , F1 , H1 ) = Q 11 1, z, z 2 , z 3 , z 4 , z 6
k=1 z −1
k
n zk
(A5 , C5 , F5 ) = Q 55 1, z 2 , z 4
k=1 z −1
k
and [T] is a matrix of coefficient of strain vector
1 z z3 0 0
[T] = (17)
0 0 0 1 z2
The constraints are applied by a direct method, which modifies the global force
vector,

ū i i= j
Fi j ⇒ (18)
F j − K ji Ui i = j
The present study is confirmed with the help of numerical results for different cases
of the laminated beam. The relation of material properties considered in order to
obtain nondimensional results is
E 11 = 0.04E 22 , G 12 = G 13 = 0.5E 22 , G 23 = 0.2E 22 , ν12 = 0.25
where E 11 and E 22 are Young’s modulus in longitudinal and transverse direction of

lamina, respectively. G12 , G23 , and G13 are the rigidity modulus along longitudinal
and transverse plane and ν12 is Poisson’s ratio. All laminae are considered to be of
equal length and thickness made up of the same orthotropic material. Thus, maximum
nondimensional transverse displacement can be obtained by,
100E 22 h 3
w̄ = wmax (19)
Pl 3
where P is point load, l and h are length and total height of the laminated beam
respectively. For uniformly distributed load P = q0 l with q0 as load per unit length.
Constraints are applied as hinged-hinged (H-H), clamped-hinged (C-H),
clamped-clamped (C-C), and clamped-free (C-F) by restricting the degree of free-
dom as explained in Table 1. The present work is validated by nondimensional
transverse displacement obtained using MATLAB program with previously available
literature. Tables 2, 3, and 4 represents a comparison of maximum nondimensional
transverse displacement for different boundary conditions such as hinged-hinged,
clamp-clamp, and clamp-free respectively. The results are obtained for three differ-
ent aspect ratios 10, 20, 100, etc. The lamination schemes considered for results
are [0], [90], and [0/90]s respectively. Two types of mechanical loads are applied
on a laminated beam, which is point load and Uniformly Distributed Load (UDL).
Results satisfy the requirement of validation. Table 5 shows results for laminated
beam subjected to clamped-hinged condition continuing the pattern of lamination
scheme and aspect ratio as explained before.
Figure 1 shows the effect of aspect ratio over the maximum nondimensional dis-
placement of the beam with lamination scheme [0/45/−45/90]s for different boundary
Table 1 Restricted degree of freedom for restricted node

Boundary condition Restricted degree of freedom

Clamped u0 , w0 , θx , ∂w0 ∂ x
Hinged u0 , w 0
Free –
Table 2 Non-dimensional maximum transverse displacement (w̄) of simply supported laminated

beam
l/h Author Lamination scheme
0 90 [0/90]s 0 90 [0/90]s
UDL Point load
10 Present 0.9220 16.3356 1.2609 1.5847 26.4259 2.2025
Reddy [11] 0.9250 16.3750 1.1370 1.6000 26.5000 1.9910
Sadek [10] 1.0130 16.5190 1.4243 1.7740 26.7760 2.5308
20 Present 0.6983 15.7734 0.8475 1.1458 25.3108 1.4063
Reddy [11] 0.7000 15.8130 0.8160 1.1500 25.3750 1.3480
Sadek [10] 0.7130 15.8330 0.8785 1.1750 25.4140 1.4665
100 Present 0.6264 15.5934 0.7140 1.0035 24.9524 1.1446
Reddy [11] 0.6280 15.6330 0.7130 1.0010 25.0150 1.1430
Sadek [10] 0.6270 15.5880 0.7445 1.0040 24.9440 1.1938
Table 3 Nondimensional maximum transverse displacement (w̄) of clamped-clamped laminated

beam
0 90 [0/90]s 0 90 [0/90]s
UDL Point load
10 Present 0.4127 3.8577 0.6545 0.8246 7.7133 1.3077
Reddy [11] 0.4250 3.8750 0.5700 0.8500 7.7500 1.1410
Sadek [10] 0.4160 3.8600 0.6820 0.9180 7.8520 1.4919
20 Present 0.1982 3.3035 0.2755 0.3962 6.6065 0.5506
Reddy [11] 0.2000 3.3120 0.2490 0.4000 6.6250 0.4980
Sadek [10] 0.1990 3.3070 0.2841 0.4080 6.6340 0.5842
100 Present 0.1277 3.1246 0.1472 0.2553 6.2493 0.2944
Reddy [11] 0.1280 3.1320 0.1460 1.7930 6.2650 0.2920
Sadek [10] 0.1280 3.1230 0.1536 0.2250 6.2460 0.3074
Table 4 Nondimensional maximum transverse displacement (w̄) of clamped-free laminated beam

0 90 [0/90]s 0 90 [0/90]s
UDL Point load
10 Present 7.1625 152.6064 8.9464 18.3361 404.9809 22.5062
Reddy [11] 7.2000 153.0000 8.5200 18.4000 406.0000 21.5800
Sadek [10] 7.1580 152.7160 9.0838 18.4790 405.7620 23.0457
20 Present 6.2821 150.3726 7.3477 16.5570 400.4974 19.2400
Reddy [11] 6.3000 150.7500 7.2300 16.6000 401.5000 19.0100
Sadek [10] 6.3120 150.5300 7.6924 16.6620 401.0490 20.1650
100 Present 5.9970 149.6549 6.8233 15.9840 399.0597 18.1809
Reddy [11] 6.0100 150.0000 6.8200 16.0200 400.0000 18.1800
Sadek [10] 5.9970 149.2020 6.7287 16.0240 397.8760 17.9304
Table 5 Nondimensional maximum transverse displacement (w̄) of clamped-hinged laminated

beam
l/h Lamination scheme
0 90 [0/90]s 0 90 [0/90]s
UDL Point load
10 0.5927 7.3377 0.8925 1.0864 12.7107 1.6528
20 0.3443 6.6971 0.4504 0.6044 11.5385 0.8005
100 0.2628 6.4916 0.3010 0.4523 11.1672 0.5184
Fig. 1 Graph of
15 C-C
displacement versus aspect
ratio with lamination scheme H-H
[0/45/−45/90]s and different C-H
boundary conditions C-F
subjected to UDL 10
Displacement
5
0
0 20 40 60 80 100
Aspect ratio( l/h)
conditions obtained using the present approach. The results state that the clamped-
free boundary condition shows more maximum nondimensional displacement for
all aspect ratios. Maximum nondimensional displacement decreases with increas-
ing aspect ratio initially and then remains constant for all boundary conditions.
Figure 2 shows nondimensional displacement of layered composite beam for lami-
nation scheme [θ 4 /04 /θ 4 ] and different boundary conditions at aspect ratio 10. The
results show that displacement increases with increasing lamination angle. Clamped-
free boundary conditions show more response as compared to others for all lamination
angles. Figures 3, 4, 5 and 6 shows the transverse displacement of laminated beam
with lamination scheme (0/45/−45/90)s and aspect ratio 10. The beam is subjected
to point load and UDL with boundary conditions hinged-hinged, clamped-hinged,
clamped-free, and clamped-clamped respectively. Behavior of beam deflection due
to boundary conditions is captured. The laminated beam shows maximum deflection
Fig. 2 Graph of 90
displacement versus C-C
80
lamination scheme H-H
[θ 4 /04 /θ 4 ] with aspect ratio 70 C-H
10 and different boundary
60 C-F
Displacement
conditions subjected to UDL

50
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90
Fig. 3 Graph of 2.5

displacement of midplane UDL
versus nodes with lamination Point load
scheme [0/45/−45/90]s , 2
aspect ratio 10 and

hinged-hinged beam
1.5
Displacement
1
0.5
0
0 2 4 6 8 10
Beam length
Fig. 4 Graph of
versus nodes with lamination 1.5 Point load
scheme [0/45/−45/90]s ,
aspect ratio 10 at clamped
hinged beam
1
Displacement
0.5
0
0 2 4 6 8 10
Beam length
Fig. 5 Graph of 30
displacement of midplane
UDL
versus nodes with lamination 25
scheme [0/45/−45/90]s , Point load
aspect ratio 10 and

20
clamped-free beam
Displacement
15
10
0
0 2 4 6 8 10
Beam length
Fig. 6 Graph of
1.2
versus nodes with lamination Point load
scheme [0/45/−45/90]s ,
1
aspect ratio 10 and
clamped-clamped beam
0.8
Displacement
0.6
0.4
0.2
0
0 2 4 6 8 10
Beam length
at its midpoint for (C-C, C-H, and H-H) boundary conditions while C-F shows max-
imum deflection at its edge with no constraints. It can be predicted that the effect of
point load on deflection is more as compared to UDL.
5 Conclusion
The present study infers that the higher-order beam theory using the meshless method
shows accurate results as correlated to classical or first-order shear deformation
theory. The application of constraints in point interpolation method is done by direct
method similar to FEM, thus easier as compared to other meshless methods. The
behavior of laminated beam is observed by plotting various graphs of displacement
with respect to lamination scheme, aspect ratio, loading condition, and boundary
conditions. As aspect ratio (l/h) increases the displacement reduces and becomes
consistent after crossing 20. Minimum displacement is observed for the lamination
scheme with angle [θ = 0°] and it increases and approaches the maximum at an angle
[θ = 90°]. The nondimensional displacement obtained for concentrated point load
is more as compared to a uniformly distributed load.
References
1. Belytschko T, Krongauz Y, Organ D, Fleming M, Krysl P (1996) Meshless methods: an

overview and recent developments. Comput Methods Appl Mech Eng 139:3–47
2. Liu GR (2016) An overview on meshfree methods: for computational solid mechanics. Int J
Comput Methods 13(5):1630001–1630042
3. Liu GR, Gu YG (2001) A point interpolation method for two-dimensional solids. Int J Numer
Methods Eng 50:937–951
4. Liu GR, Zhang GY, Dai KY (2005) A linearly confirming point interpolation method (LC-PIM)
for 2D solid mechanics problems. Int J Comput Methods 2(4):645–665
5. Nguyen VP, Rabczuk T, Bordas S, Duflot M (2008) Meshless methods: a review and computer
implementation aspects. Math Comput Simul 79:763–813
6. Abrate S, Sciuva MD (2017) Equivalent single layer theories for composite and sandwich
structures: a review. Compos Struct 179:482–494
7. Reddy JN (1984) A simple higher order theory for laminated composite plates. J Appl Mech
51:745–752
8. Khdeir AA, Reddy JN (1997) An exact solution for the bending of thin and thick cross-ply
laminated beams. Compos Struct 37:195–203
9. Ferreira AJM, Roque CMC, Martins PALS (2004) Radial basis functions and higher-order
shear deformation theories in the analysis of laminated composite beams and plates. Compos
Struct 66:287–293
10. Sadek SHM, Belinha J, Parente MPL, Natal Jorge RM, César JMA, Ferreira AJM (2017) The
analysis of composite laminated beams using a 2D interpolating meshless technique. Acta
Mech Sin 1–18
11. Reddy JN (2004) Mechanics of laminated composite plates and shells, 2nd edn. CRC Press
LLC
12. Liu GR (2003) Mesh free methods: moving beyond the finite element method, 1st edn. CRC
Press LLC
Dynamic Behaviour of Laminated
Composite Beam Undergoing Moving
Loads
Lalit Babu Saxena, Appaso M. Gadade, and Sanjiv M. Sansgiri
Abstract This study is about dynamic response of laminated composite beam under-
going moving loads. Rotary inertia and shear deformation effects are considered by
using Timoshenko Beam Theory (TBT). Finite Element Method (FEM) is applied
to discretize the structural element into space and for time discretization Classical
Midpoint Rule with Midpoint Acceleration (MPR-MPA) is employed which is a part
of Generalized Single Step Single Solve (GSSSS) family of algorithms. A MATLAB
code is developed to obtain dynamic responses such as dynamic magnification fac-
tor and maximum dynamic deflection at the mid-span of simply supported isotropic
and laminated composite Timoshenko beam. Numerical results are obtained and
validated with the literature available and it shows good agreement.
Keywords Timoshenko beam theory · Midpoint rule with midpoint acceleration ·

Moving loads · Laminated composite beam · Dynamic magnification factor ·
Finite element method
1 Introduction
Moving load problems is of great interest to engineers. Dynamic deflection, vibra-

tions and stresses effects vary greatly in structures as compared to effects due to
static deflection, vibration and stresses. Bridges subjected to moving vehicle loads,
railway tracks subjected to moving train loads, landing aircraft on carrier are some of
the common moving load cases. In 1905, Krylov [1] examined the simply supported
L. B. Saxena · S. M. Sansgiri
Mechanical Engineering, Army Institute of Technology, Pune 411015, India
S. M. Sansgiri
A. M. Gadade (B)
Mechanical Engineering Department, Thapar Institute of Engineering and Technology, Patiala
147004, India

https://doi.org/10.1007/978-981-15-4488-0_18
202 L. B. Saxena et al.
beam case under a point load where the mass of beam was taken more as compared
to the mass of the load. In 1908, Timoshenko [2] solved the same case using the
eigenfunctions method. In 1972, Fryba [3] proposed many studies related to struc-
tures under moving loads. However, few studies have been done on the dynamic
behaviour of layered composite beams undergoing moving loads. In 1998, Kadi-
var and Mohebpour [4] analyzed induced vibrations in asymmetric layered beam
subjected to moving loads. In 2003, Zibdeh and Abu Hilal [5] studied stochastic
response of layered composite beams under random moving loads. In 2005, Kavipu-
rapu [6] studied dynamic behaviour of glass/epoxy simply supported beams under-
going moving loads in hygro-thermal environment. In 2008, Kiral and Kiral [7]
presented dynamic behaviour of symmetric layered beams undergoing moving loads
using a 3-D model based on classical lamination theory. In 2011, Mohebpour et al. [8]
studied the dynamic behaviour of layered beams under moving oscillators. In 2012,
Kahya [9] investigated dynamic response of layered beams undergoing moving loads
using multi-layered beam element. In 2016, Tao et al. [10] studied nonlinear dynamic
behaviours in thermal environment of fiber metal layered beams under moving loads.
In 2017, Chen et al. [11] investigated nonlinear dynamic behaviours of fiber metal
layered beam resting on elastic foundation undergoing moving harmonic load and
thermal load. It is clear from the literature that no one has used MPR-MPA technique
for solving moving load problems yet.
2 General Formulation
Consider a layered composite Timoshenko beam consisting of n orthotropic layers

with principle material coordinates (x1k , x2k , x3k ) of kth lamina oriented at an angle
θ k to laminate coordinate, x. The thickness, width and length of beam are h, b and l,
respectively (see Fig. 1). Origin of the material coordinates at the middle of laminate.
Displacement field for Timoshenko Beam Theory is
u(x, z, t) = u 0 (x, t) + zφ(x, t) and w(x, z, t) = w0 (x, t) (1)
where (u, w) are displacements along (x, z) coordinates, respectively, u 0 and w0

are the displacements of a point on the midplane (x, 0) and φ is the rotation of the
transverse normal.
Linear strain vectors corresponding to displacement fields are
∂u ∂w ∂u
εx x = ε1 = and εx z = ε5 = + (2)
∂x ∂x ∂z
that can be written as

T
{εl } = ε1 ε5 l (3)
Dynamic Behaviour of Laminated Composite Beam Undergoing Moving … 203
Fig. 1 Displacement field model and layers arrangement
The associated strains for the assumed displacement field can be written as

εi = εi0 + zψi0 (i = 1, 5)
(4)
ε1 = ε10 + zψ10 and ε5 = ε50 + zψ50
The strain–displacement relations are
∂u 0 ∂φ ∂w0
ε10 = ∂x
, ψ10 = ∂x
, ε50 = φ + ∂x
, ψ50 = 0 (5)
from Eqs. (3) and (4)
{εl } = [T ]{ε̄l } (6)
where [T ] is the thickness matrix expressed as

1z0
[T ] = and {ε̄l }T = ε10 ψ10 ε50 (7)
001
Using (5), Eq. (7) can be rewritten in matrix form as
{ε̄l } = [L]{} (8)

where [L] is the differential operator and is given by

⎡ ⎤
∂/∂ x 0 0 T
[L] = ⎣ 0 0 ∂/∂ x ⎦ and {} = u 0 w0 φ (9)
0 ∂/∂ x 1
where {} is the mid-plane displacement vector for the C0 continuous model.
The Stress–Strain relation for a lamina with respect to the fiber-matrix coordinate
axis can be given as

k σx x Q 11 0 εx x
σp = Q ε p it can be written as, = (10)
τx z 0 Q 55 γx z
where

Q 11 = Q 11 cos4 θk + Q 22 sin4 θk + 2(Q 12 + 2Q 66 ) sin2 θk cos2 θk
(11)
Q 55 = Q 55 cos2 θk + Q 44 sin2 θk
and

μ21 E 11
Q 11 = (1−μE1211μ21 ) , Q 22 = (1−μE1222μ21 ) , Q 12 = (1−μ 12 μ21 )
,
(12)
Q 44 = G 23 , Q 55 = G 13 , Q 66 = G 12
where Q i j are material constants, Q i j are the transformed material constants. Mate-
rial properties E 11 and E 22 are Young’s modulus in longitudinal and transverse direc-
tions. G12 is Rigidity modulus in longitudinal plane whereas, G13 and G23 are Rigidity
modulus in transverse plane. Poisson’s ratios are μ12 and μ21 . θ k is the angle of fiber
orientation of lamina.
Material property matrix can be expressed as
⎡ ⎤
Q 11 z Q 11 0
it can be written as, [D] = ⎣ z Q 11 z 2 Q 11 0 ⎦
k
[D]3×3 = [T ]T3×2 Q [T ]2×3
2×2
0 0 Q 55
(13)
It can further be expressed as

⎡ ⎤
NL z k
A1 B1 0
Q [T ]dz = ⎣ B1 E 1 0 ⎦
k
[D] = b × [T ]T (14)
k=1 z
k−1 0 0 S1 3×3
where
NL z k
NL z k

k k
(A1 , B1 , E 1 ) = b × Q 11 1, z, z 2 dz and S1 = b × K Q 55 dz (15)
k=1 z k=1 z
k−1 k−1
where (A1 , B1 , E 1 , S1 ) are stiffness coefficients, NL is number of layer. K is shear

correction factor and its value is 0.833 for rectangular cross-section [12].
3 Finite Element Formulation
In this study, a C0 higher-order cubic finite element is used. Every element has
hence four nodes which are equally spaced with three DOF per node and are used to
interpolate both, displacement field as well as geometric coordinates.
The trial function can be written as

Z
T
{ }= Ni { i} and { } = u 0 w0 φ (16)
i=1
where Z = total number of nodes per element, Ni = shape functions for ith node.
The shape functions of higher-order cubic element can be written as

N1 = − 16
9
(ξ − 1) ξ − 13 ξ + 13 , N2 = 16 (ξ
27
− 1)(ξ
+ 1) ξ − 13 ,
(17)
N3 = − 27
16 (ξ
− 1)(ξ + 1) ξ + 13 , N4 = 16 (ξ
9
+ 1) ξ − 13 ξ + 13
The strains are related to displacement by strain–displacement matrix [B] as,
[B] = [L][N ] (18)
where [N ] is given by,

⎡ ⎤ ⎡ ⎤
Nu 0 N1 0 0 N2 0 0 N3 0 0 N4 0 0
[N ] = ⎣ Nw0 ⎦ = ⎣ 0 N1 0 0 N2 0 0 N3 0 0 N4 0 ⎦ (19)
Nφ 0 0 N1 0 0 N2 0 0 N3 0 0 N4
Elemental stiffness matrix and mass matrix is computed by transforming the

physical coordinate system (x, 0) into natural coordinate system (ξ, 0) and can be
written as,
1 1
[K ] = [B] [D][B] det[J ]dξ and [M] =
T
[N ]T ρ[I ][N ] det[J ]dξ (20)
−1 −1
where [J ] is the Jacobian matrix. ρ is the density of the material and [I] is mass
coefficients matrix and can be expressed as,
⎡ ⎤
I0 0 I1 NL z k

[I ] = ⎣ 0 I0 0 ⎦ where (I0 , I1 , I2 ) = b 1, z, z 2 dz (21)
I1 0 I2 k=1 z
k−1
The moving load is a concentrated mass acting at a point on layered composite

Timoshenko beam and the load (f m ) is moving from left to right (see Fig. 1). The
point load due to concentrated mass can be represented as,
f m (x, t) = Wm δ[x − xm (t)] and Wm = −mg (22)
where δ is Dirac delta function and W m is the weight of moving mass, m is mass
and g is the gravitational acceleration and its value is 9.81 m/s2 . vm is moving load
velocity in m/s. x m is mass location at time t and is given by
xm (t) = vm t (23)
At any given time t mass is located at lm on element e. Mass location relative to

element length L e can be represented as
η(t) = lm /L e (24)
The discretized force vector can be written as
f m (t) = Wm Nw0 (η(t)) (25)
4 Solution Technique
MPR-MPA is an explicit second-order accurate Linear Multistep method (LMS).

MPR-MPA is a time-stepping algorithm that defines the derivatives with respect to
time. All LMS algorithms including MPR-MPA and Newmark method are found
within GSSSS family of algorithms.
Equation of motion can be expressed as
M ü + C u̇ + K u = f m (t) (26)
Initial conditions are
u̇(0) = u̇ 0 and u(0) = u 0 (27)

MPR-MPA algorithm is defined by
1 1 1 ⎫
(M + C t + K t 2 ) ü = −M ü n − C(u̇ n + ü n t)⎪ ⎬
2 4 2 (28)
1 1 1 ⎪
⎭
− K (u n + u̇ n t + ü n t 2 ) + ( f n + f n+1 )
2 4 2
1 1 ⎫
u n+1 = u n + u̇ n t + ü n t 2 + ü t 2 ⎪
⎬
2 2 (29)
1 ⎪
u̇ n+1 = u̇ n + ü n t + ü t and ü n+1 = ü n + ü ⎭
2
Ratio of maximum dynamic deflection to maximum static deflection is known as

Dynamic Magnification Factor (DMF). SP is the Speed Parameter, which is ratio of
fundamental time period to actual time taken by the moving load to travel the entire
beam length (T f /T ).
The present results using FEA and MPR-MPA algorithm shows good agreement
with the published results. Data 1 of Table 1 is used in Table 2. At critical velocity vcr ,
T f = T so SP will be 1. Maximum DMF of simply supported beam is at SP = 1.25.
At low speeds, maximum dynamic displacement is nearly equal to maximum static
displacement but at higher speeds, it is 170% of the maximum static displacement.
In case of extremely higher speeds when SP > 1.25, the DMF starts decreasing so
also the maximum dynamic displacement. The critical velocity and fundamental time
period are calculated by
vcr = (ω1l)/(iπ ), i = 1, 2, 3 . . . and Tf = l/vcr (30)
where ω1 is the first fundamental frequency of the beam.

Data 2 of Table 1 is used in Table 3. The maximum DMF of simply supported
layered composite beam occurs between SP = 1.2 and 1.3. The vcr = 311.2 m/s
Table 1 Required input data

No. Description References
Data 1 f m = 4.45 N, l = 0.1016 m, b = h = 0.00635 m, [4]
E = 2.068 × 1011 N/m2 , ρ = 10,686.9 kg/m3 , ν = 0.3
Data 2 f m = 4.45 N, l = 0.1016 m, b = 0.00635 m, h = 0.00745 m, [4]
E 11 = 1.448 × 1011 Pa, E 22 = 9.65 × 109 Pa,
G 12 = G 13 = 4.136 × 109 Pa, G 23 = 3.447 × 109 Pa,
ρ = 1389.297 kg/m3 , ν12 = 0.25
208
Table 2 DMF at mid-span of simply supported isotropic beam (20 time steps are used)
SP vm (m/s) Present FEM [8] FEM [4] FEM [13]a SIM [14] FEM [15] Analytical [16] Analytical [17]
0.125 15.6 1.054 1.053 1.063 1.049 1.042 1.055 – 1.025
0.250 31.2 1.141 1.139 1.151 1.121 1.082 1.112 1.110 1.121
0.500 62.4 1.270 1.267 1.281 1.266 1.266 1.252 1.240 1.258
0.750 93.6 1.571 1.569 1.586 – – – – 1.572
1.000 124.8 1.687 1.687 1.704 1.703 1.662 1.700 1.680 1.701
1.250 156.0 1.710 1.711 1.727 – – – – 1.719
1.500 187.2 1.679 1.681 – – – – – –
2.000 250.0 1.524 1.528 1.542 1.540 1.518 1.540 1.540 1.548
SIM: Structural Impedence Method
a Using nonconforming plate element with 25 time steps
L. B. Saxena et al.
Table 3 DMF at mid-span of symmetric [θ/ − θ/ − θ/θ] angle-ply simply supported AS/3501-6
graphite-epoxy beam (50 time steps are used for better convergence)
SP 0.13 0.26 0.52 0.78 1 1.3 1.56 1.82
vm 40.9 81.8 163.6 245.4 311.2 409 490.8 572.6
0◦ 1.054 1.132 1.270 1.565 1.636 1.655 1.627 1.561
SP 0.125 0.25 0.50 0.75 1 1.25 1.50 2
vm 15 30 60 90 120 150 180 240
30◦ 1.058 1.129 1.263 1.575 1.697 1.715 1.683 1.532
SP 0.12 0.24 0.48 0.72 1 1.2 1.44 1.68
vm 10.2 20.4 40.8 61.2 87 102 122.4 142.8
60◦ 1.058 1.093 1.202 1.531 1.700 1.724 1.704 1.658
SP 0.12 0.24 0.48 0.72 1 1.2 1.44 1.68
vm 10.2 20.4 40.8 61.2 86.8 102 122.4 142.8
90◦ 1.059 1.094 1.203 1.532 1.700 1.723 1.704 1.657
is maximum for 0◦ angle-ply beams while it is minimum, vcr = 86.8 m/s at 90◦
angle-ply beams. As the ply orientation increases, the critical velocity decreases.
The maximum DMF is between 1.655 and 1.724 while 1.724 is for the 60◦ angle-ply
beams. Stiffness response of 0◦ angle-ply beams is maximum in comparison to any
other ply orientation.
Data 2 of Table 1 is used for the graphs (see Fig. 2). It can be seen from the
graphs that maximum static displacement occurs at center of beam, but the maximum
dynamic displacement occurs at nearly three quarters from the left end of the beam.
This is due to the delay in time and this time delay increases as vm increases. This
time delay is more beyond the critical nonlinear region as compared to below the
critical linear region. The pattern of graphs is same for all simply supported isotropic
and layered composite beam at any velocity and ply scheme.
4 10-5 10-5
4 v m =120
3 v m =150
Displacement w0 (m)
Displacement w0 (m)
v m =180
2 3
v m =240
Static
1 2
v m =15
v m =30
0
v m =60
1
-1 v m =90
Static
0
-2
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Normalized Time t (s) Normalized Time t (s)
Fig. 2 Maximum transverse displacement w0 (m) versus Normalized time t (s) at mid-span of
symmetric [30/−30/−30/30] angle-ply simply supported beams
6 Conclusions
FEM along with MPR-MPA explicit numerical time discretization technique is robust
enough to handle moving load structural dynamic problems. At very low speed, max-
imum dynamic deflection is almost equal to static deflection but at higher speeds,
maximum dynamic deflection is greater than 170% of static deflection. Initially as
speed increases, the dynamic magnification factor and maximum deflection increase
but at very high speed, the dynamic magnification factor and the maximum deflection
decrease. Layered composite beam with symmetric [0/−0/−0/0] angle-ply lamina-
tion scheme shows a much stiffer response as compared to any other angle-ply lami-
nation scheme. It hence contributes to minimum transverse deflection. The maximum
DMF of simply supported beams occurs in the range SP = 1.2–1.3. The maximum
dynamic displacement occurs at nearly three-quarter of the beam due to delay in
time and this time delay increases as the vm increases. Due to high strength, low
density, long term durability, greater corrosion and fatigue resistance, the laminated
composite material specially graphite-epoxy beams can be used in bridge structures
instead of conventional ones.
References
1. Krylov AN (1905) Mathematical collection of papers of the Academy of Sciences, vol 61.
Saint Petersburg, Russia
2. Timoshenko SP (1908) Forced vibration of prismatic bars. Izvestiya Kievskogo Politekhnich-
eskogo Instituta, Russia
3. Fryba L (1972) Vibration of solids and structures under moving loads. Noordhoff International
Publishing, Groningen, Netherland
4. Kadivar MH, Mohebpour SR (1998) Finite element dynamic analysis of unsymmetric com-
posite laminated beams with shear effect and rotary inertia under the action of moving loads.
Finite Elem Anal Des 29:259–273
5. Zibdeh HS, Abu Hilal M (2003) Stochastic vibration of laminated composite coated beam
traversed by a random moving load. Eng Struct 25:397–404
6. Kavipurapu PK (2005) Forced vibration and hygrothermal analysis of composite laminated
beams under the action of moving loads. West Virginia University, Morgantown, West Virginia
7. Kiral Z, Kiral BG (2008) Dynamic analysis of a symmetric laminated composite beam subjected
to a moving load with constant velocity. J Reinf Plast Compos 27:19–32
8. Mohebpour SR, Fiouz AR, Ahmadzadeh AA (2011) Dynamic investigation of laminated com-
posite beams with shear and rotary inertia effect subjected to the moving oscillators using finite
element method. Compos Struct 93:1118–1126
9. Kahya V (2012) Dynamic analysis of laminated composite beams under moving loads using
finite element method. Nucl Eng Des 243:41–48
10. Tao C, Fu Y, Dai HL (2016) Nonlinear dynamic analysis of fiber metal laminated beams
subjected to moving loads in thermal environment. Compos Struct 140:410–416
11. Chen Y, Fu Y, Zhong J, Tao C (2017) Nonlinear dynamic responses of fiber metal laminated
beam subjected to moving harmonic loads resting on tensionless elastic foundation. Compos
B 131:253–259
12. Cowper GR (1966) The shear coefficient in Timoshenko’s beam theory. ASME 335–340
13. Taheri MR, Ting EC (1990) Dynamic response of plates to moving loads finite element method.
Comput Struct 34:509–521
14. Taheri MR, Ting EC (1989) Dynamic response of plate to moving loads structural impedence
method. Comput Struct 33:1379–1393
15. Yoshida DM, Weaver W (1971) Finite element analysis of beams and plates with moving loads.
IABSE 31:179–195
16. Filho FV (1966) Dynamic influence lines of beams and frames. J Struct Div 371–386
17. Meirovitch L (1967) Analytical methods in vibrations. The Macmillan Company, New York
Demystifying Fractal Analysis of Thin
Films: A Reference for Thin Film
Deposition Processes
F. M. Mwema , Esther T. Akinlabi , and O. P. Oladijo
Abstract In this article, a variety of synthetic (or simulated) surfaces of various

morphologies of thin films and their fractal analyses are presented. Similar scaling
factors have been used to generate the synthetic images in GwydionTM software.
The surfaces are based on the actual morphologies arising from various thin film
deposition techniques. Using actual thin films of CdTe deposited by radio-frequency
(RF) sputtering technique, we have successfully shown that the fractal analyses on
the synthetic surfaces can be used to explain, theoretically, the development and self-
affinity of various thin films. Based on this validation, the results of fractal analyses on
different morphologies of thin films were generated using different fractal methods in
Gwydion software. The methods used here include Minkowski functionals, height-
to-height correlation, areal autocorrelation, and power spectral density functions. The
article will be a good resource for explaining the fractal behavior and morphology
of thin films arising from different deposition methods.
Keywords Correlation · Fractal · Minkowski · Surfaces · Thin films
1 Introduction
During the deposition process of thin films, there are different morphologies of
structures formed depending on the deposition type, process parameters, films, and
substrate types [1]. Scanning probe microscopy (SPM) techniques such as atomic
force microscope (AFM) are used to study the surface morphology of various thin
films and coatings [2–5]. The micrographs obtained from the SPM techniques are
used to undertake roughness analyses such as statistical [6, 7] and fractal measure-
ments [8–12]. Fractal methods offer a detailed description of lateral roughness [13]
and the nature of the surface morphology can be captured [14]. Although fractal
F. M. Mwema (B) · E. T. Akinlabi · O. P. Oladijo

University of Johannesburg, Johannesburg, South Africa
O. P. Oladijo
Botswana International University of Science and Technology, Palapye, Botswana
https://doi.org/10.1007/978-981-15-4488-0_19
214 F. M. Mwema et al.
characterization is widely reported in the literature [6, 12, 15–17], very little is
reported on the relationship between the fractal measurements and the structure
type/morphologies of the films. Therefore, the purpose of this work is to generate
fractal profiles (using Minkowski functionals, autocorrelation, height-height corre-
lation, and power spectral density functions) based on theoretical/synthetic surfaces
of different morphologies.
2 Methods
Various synthetic morphologies of thin films were produced using scanning probe
microscopy (SPM) software Gwydion (Fig. 1). These films depict different struc-
tural types that are obtained through various deposition processes such as sputtering
and thermal spray. These structures are columnar, ballistic, fibrous, and pile-up struc-
tures (Fig. 1) and represent some of the most common morphologies observed in thin
films. The process of creating synthetic (simulated) surfaces in Gwydion software
are described elsewhere [18, 19]. All the images were single-layer, with a maximum
Fig. 1 Illustrating simulated surfaces of thin films consisting of various structural morphologies
a columnar b ballistic c fibrous, and d pile-up particles. Corresponding 3D images are shown as
insets on each image
Demystifying Fractal Analysis of Thin Films: A Reference … 215
Fig. 2 Flowchart illustrating the image and fractal analyses procedures
height of 1000 nm and a scan area of 3 × 3 µm2 . The fractal analyses of the simu-
lated AFM images were undertaken according to the flowchart in Fig. 2. To validate
the simulated fractal analyses, fractal values of a typical columnar AFM of CdTe
thin films sputtered on glass substrates (Fig. 3) were computed and compared to
the simulations. This process was iterative until comparable results were obtained
(e.g., Fig. 4). Subsequently, all computations were conducted for the other simulated
structure and results presented in Table 1.
The results of the fractal analyses of the simulated AFM surfaces of thin films are
presented in Table 1. A short description of the results in Table 1 is as follows:
• Minkowski connectivity (X): Negative values dominate the X for columnar, bal-
listic, and fibrous structures whereas positive dominates for pile-up particles. The
profiles vary with the type of structures.
• Minkowski boundary: There are significant differences; while columnar and bal-
listic tend to nearly Gaussian profiles, the maximum values of boundary lengths
for fibrous, and pile-up are skewed right and left, respectively.
• Minkowski volume: The profiles for columnar, ballistic and pile-up particles are
symmetrical about V = 0.5, and exhibit S-shape [11, 20, 23]. The fibrous structures
are asymmetrical and exhibit quarter-circle shaped Minkowski volume.
• Power spectral density: For columnar surface structures, the profile has a flat
region at low frequencies and linearly decreasing PSD at high frequency with
Fig. 3 a SEM micrograph along the cross-section of CdTe thin films deposited by RF magnetron
sputtering. The white arrows show columnar structures of the films perpendicular to the substrate.
b Showing the AFM image (scan area of 0.5 × 0.5 µm2 ) at the top surface of the films. Recalibrated
3D AFM image of the CdTe films. Obtained from Camacho-Espinosa et al. [22] under open access
creative commons
Fig. 4 Bi-logarithmic plots for power spectral density (PSDF) against the spatial frequency (k) of
(a) typical columnar CdTe films deposited on glass substrates and b the corresponding simulated
profile plot. The shapes of the profiles are comparable and are characterized by withers at the
transition region between the flat and the linear areas of the PSDF profile
Table 1 Illustrating various fractal analyses results from different simulated (synthetic) structures of thin films
Analysis Type of structure for thin films
Columnar Ballistic Fibrous Pile-up particles
Minkowski connectivity
Minkowski boundary
Demystifying Fractal Analysis of Thin Films: A Reference …
(continued)
217
Table 1 (continued)
218

Minkowski volume
Power spectral density function
(continued)
F. M. Mwema et al.
Table 1 (continued)
Areal autocorrelation function
Height-height
correlation/structure function
Demystifying Fractal Analysis of Thin Films: A Reference …
219
withers at the transition point [21, 24, 25]. For ballistic surfaces, the 1-d PSD
profile consists of flat region and nonlinearly decreasing PSD. The flat region is
not clear in fibrous surfaces whereas the pile-up surfaces have distinct flat and
linear regions at low and high spatial frequencies respectively.
• Areal autocorrelation (ACF): For columnar surfaces, the profile exhibit oscillatory
behavior with decreasing and increasing values at low and high shifts respectively.
The ACF decreases sharply to nearly r = 1.0 and then nearly remains constant
for ballistic and fibrous. For pile-up surfaces, the ACF profile exhibit U-shape.
• Height-height correlation (HCF): The HCF increases with r for all surfaces up
to certain values. At very large r (mounded surface characteristics) oscillatory
behavior of the profile was observed for columnar and ballistic surface structures
[12, 23, 25]. The flat region (at large r) is not distinct for ballistic surfaces. The
HCF decreases at nearly constant r at the end of the flat region for columnar and
pile-up surfaces.
4 Conclusion
The profile plots of the most common fractal analyses of thin film surfaces of different
synthetic morphologies have been presented. The surfaces were generated using
Gwydion software and a typical validation of the columnar structure showed that the
software provides a good approximation of deposited films. Profiles of Minkowski
functionals, autocorrelation, height-height correlation and power spectral density
functions of the synthetic morphologies (columnar, ballistic, fibrous and pile-up
particles) presented in Table 1 will be a useful reference in relating the fractal results
to the films’ deposition techniques and conditions.
References
1. Mwema FM, Oladijo OP, Akinlabi SA, Akinlabi ET (2018) Properties of physically deposited
thin aluminium film coatings: a review. J Alloy Compd 747:306–323. https://doi.org/10.1016/
j.jallcom.2018.03.006
2. Eaton P, West P (2010) Atomic force microscopy. Oxford University Press Inc., New York
3. Mwema FM, Akinlabi ET, Oladijo OP (2019) Correction of artifacts and optimization of atomic
force microscopy imaging: a case of thin aluminum films for prosthetic applications. In: Kumar
K, Paulo Davim J (eds) Design, development, and optimization of bio-mechatronic engineering
products, IGI Global, pp 158–179. https://doi.org/10.4018/978-1-5225-8235-9.ch007
4. Kwoka M, Ottaviano L, Szuber J (2007) AFM study of the surface morphology of L-CVD
SnO2 thin films. Thin solid films 515:8328–8331. https://doi.org/10.1016/j.tsf.2007.03.035
5. Sobola D, Ţălu S, Solaymani S, Grmela L (2017) Influence of scanning rate on quality of AFM
image: study of surface statistical metrics. Microsc Res Tech 80(12):1328–1336. https://doi.
org/10.1002/jemt.22945
Demystifying Fractal Analysis of Thin Films: A Reference … 221
6. Erinosho MF, Akinlabi ET, Johnson OT (2018) Characterization of surface roughness of laser
deposited titanium alloy and copper using AFM. Appl Surf Sci 435:393–397. https://doi.org/
10.1016/j.apsusc.2017.11.131
7. Mwema FM, Oladijo OP, Sathiaraj TS, Akinlabi ET (2018) Atomic force microscopy analysis
of surface topography of pure thin aluminium films. Mater Res Express 5(4):1–15. https://doi.
org/10.1088/2053-1591/aabe1b
8. Stach S, Sapota W, Ţălu S, Ahmadpourian A, Luna C, Ghobadi N, Arman A, Ganji A (2017)
3-D surface stereometry studies of sputtered TiN thin films obtained at different substrate
temperatures. J Mater Sci Mater Electron 28(2):2113–2122. https://doi.org/10.1007/s10854-
016-5774-9
9. Ţălu S, Stach S, Raoufi D, Hosseinpanahi F (2015) Film thickness effect on fractality of tin-
doped In2O3 thin films. Electron Mater Lett 11(5):749–757. https://doi.org/10.1007/s13391-
015-4280-1
10. Ţălu S, Bramowicz M, Kulesza S, Solaymani S (2018) Topographic characterization of thin
film field-effect transistors of 2,6-diphenyl anthracene (DPA) by fractal and AFM analysis.
Mater Sci Semicond Process 79:144–152. https://doi.org/10.1016/j.mssp.2018.02.008
11. Salerno M, Banzato M (2005) Minkowski measures for image analysis in scanning probe
microscopy. Microsc Anal 19(4):13–15
12. Yadav RP, Kumar M, Mittal AK, Dwivedi S, Pandey AC (2014) On the scaling law analysis
of nanodimensional LiF thin film surfaces. Mater Lett 126:123–125. https://doi.org/10.1016/j.
matlet.2014.04.046
13. Yadav RP, Dwivedi S, Mittal AK, Kumar M, Pandey AC (2012) Fractal and multifractal analysis
of LiF thin film surface. Appl Surf Sci 261:547–553. https://doi.org/10.1016/j.apsusc.2012.
08.053
14. Senthilkumar M, Sahoo NK, Thakur S, Tokas RB (2005) Characterization of microroughness
parameters in gadolinium oxide thin films: A study based on extended power spectral density
analyses. Appl Surf Sci 252(5):1608–1619. https://doi.org/10.1016/j.apsusc.2005.02.122
15. Dallaeva D, Talu S, Stach S, Skarvada P, Tomanek P, Grmela L (2014) AFM imaging and
fractal analysis of surface roughness of AlN epilayers on sapphire substrates. Appl Surf Sci
312:81–86. https://doi.org/10.1016/j.apsusc.2014.05.086
16. Starodubtseva MN, Starodubtsev IE, Starodubtsev EG (2017) Novel fractal characteristic of
atomic force microscopy images. Micron 96:96–102. https://doi.org/10.1016/j.micron.2017.
02.009
17. Raoufi D (2010) Fractal analyses of ITO thin films: a study based on power spectral density.
Phys B Condens Matter 405(1):451–455. https://doi.org/10.1016/j.physb.2009.09.005
18. Nečas D, Klapetek P (2012) Gwyddion: an open-source software for SPM data analysis. Cent
Eur J Phys 10(1):181–188. https://doi.org/10.2478/s11534-011-0096-2
19. Klapetek P, Necas D, Anderson C (2013) Gwyddion user guide 1–122. http://gwyddion.net/
download/user-guide/gwyddion-user-guide-en.pdf
20. Mwema FM, Akinlabi ET, Oladijo OP, Majumdar JD (2019) Effect of varying low substrate
temperature on sputtered aluminium films. Mater Res Express 6(5):056404. https://doi.org/10.
1088/2053-1591/ab014a
21. Mwema FM, Oladijo OP, Akinlabi ET (2018) Effect of substrate temperature on aluminium thin
films prepared by RF-magnetron sputtering. Mater Today Proc 5(9/3):20464–20473. https://
doi.org/10.1016/j.matpr.2018.06.423
22. Camach-Espinosa E, Rosendo E, Díaz T, Oliva I, Rejon V, Peria J (2014) Effects of temperature
and deposition time on the RF- sputtered CdTe films preparation. Superf. y vacío 27(1):15–19.
http://smctsm.org.mx/ojs/index.php/SyV/article/view/148
23. Mwema FM, Akinlabi ET, Oladijo OP (2019) Fractal analysis of hillocks: a case of RF sputtered
aluminum thin films. Appl Surf Sci 489:614–623. https://doi.org/10.1016/J.APSUSC.2019.
05.340
24. Shakoury R, Rezaee S, Mwema FM, Luna C, Ghosh K, Jurečka S, Ţălu S, Arman A, Korpi
AG (2020) Multifractal and optical bandgap characterization of Ta2 O5 thin films deposited
by electron gun method. Opt Quantum Electron 52(2). https://doi.org/10.1007/S11082-019-
2173-5
25. Mwema FM, Akinlabi ET, Oladijo OP (2020) Fractal analysis of thin films surfaces: a brief
overview. In: Awang M, Emamian S, Yusof F (eds) Advances in material sciences and engi-
neering. Lecture notes in mechanical engineering. Springer, Singapore. https://doi.org/10.1007/
978-981-13-8297-0_28
Dynamic Analysis of Rectangular
Aluminum Plate Under Transverse
Loading Using Finite Difference
Algorithm
Michael C. Agarana, Esther T. Akinlabi, and Michael O. Ikumapayi
Abstract Aluminum is one of the most used mechanical elements in Manufacturing.

This paper focuses on the analysis of its dynamic response of an Aluminum plate,
under a moving load. The moving load, in this case, is assumed to be partially
distributed. Rotatory inertia and damping effects were neglected, while the effect
of shear deformation was put into consideration. Also, the rectangular Aluminum
plate was supported by a simple form of foundation. A numerical algorithm–Finite
difference was adopted in solving the mathematical model governing the deflection of
plates under moving load under consideration. It was observed, among other results,
that the maximum amplitude of the deflection of the Aluminum plate is a function
of the contact area of the load and velocity of the moving load, which is in line with
the results in the literature.
Keywords Stiffness · Homogeneous isotropic materials · Analytical

investigation · Cross sections
1 Introduction
Loads moving on flat materials are a common occurrence in Mechanical Engineering,

Civil Engineering and related disciplines. In building and geotechnical structures,
also, plate analysis is often carried out, especially those resting on a subgrade [1].
M. C. Agarana (B)
Department of Mathematics, Covenant University, Ota, Nigeria
E. T. Akinlabi · M. O. Ikumapayi
Department of Mechanical Engineering Science, University of Johannesburg, Johannesburg,
South Africa
M. O. Ikumapayi
E. T. Akinlabi
Department of Mechanical Engineering, Covenant University, Ota, Nigeria
https://doi.org/10.1007/978-981-15-4488-0_20
In this study, the situation is modelled as dynamic response of a rectangular plate

supported by an elastic foundation. The mechanical behaviour of the foundation
and the form of interaction between it and the plate was analyzed. In Winkler model,
unlike in Pasternak, it is assumed that the foundation consists of linear elastic springs
that are closely spaced and independent of each other [2–4].
Aluminum is a decent conduit of power and heat. It is light and strong. It can
be pounded into sheets (malleable) or maneuverer out into wires (ductile). It is an
exceptionally receptive metal, in spite of the fact that it is not corrosive. Aluminum
anticipates corrosion by framing a little, thin layer of aluminum oxide on its surface.
This layer shields the metal by keeping oxygen from achieving it. Corrosion cannot
happen without oxygen. In light of this thin layer, the reactivity of aluminum is not
seen. Numerous things are made of aluminum. Quite a bit of it is utilized as a part
of overhead electrical cables. It is additionally broadly utilized as a part of window
casings and airplane and ship bodies. It is found at home as pans, soda pop jars and
cooking foil. Aluminum is likewise used to coat auto headlamps and smaller plates
[5].
Transverse loading or force is a force applied vertically to the plane of the longi-
tudinal axis of a configuration, an example is a wind load. In an elastic material, it
causes it to bend and rebound from its original position, with inner tensile and com-
pressive straining associated with the change in curvature of the material. Transverse
force increases the slanting deflection [6].
Finite difference methods (FDM) are numerical methods for solving differential
equations by approximating them with difference equations, in which finite differ-
ences approximate the derivatives. FDMs are thus discretization methods [7, 8]. This
study adopted the finite difference algorithm to analyze the effect of the moving load
on an aluminum rectangular plate resting on a elastic subgrade.
An Aluminum plate, with a moving load and different boundary conditions, was con-
sidered. The load is relatively small, so its inertia can be neglected, and is traversing
along the mid-space on the surface of the rectangular aluminum plate, supported by
a Winkler foundation. The following assumptions were made: The rectangular Alu-
minum plate is of constant cross-section, the moving load traverses with a constant
speed, the moving load is guided in such a way that it keeps contact with the plate
throughout the motion, the plate is continuously supported by a Winkler founda-
tion, the moving load is a partially distributed moving load, and that the rectangular
Aluminum plate is elastic [8–10].
The governing equation is as follows [10–12]:
Bρh 3 ∂ 3 ψx ∂ 2m x ∂ 2m x y
+ +
12 ∂ x∂t 2 ∂x 2 ∂ x∂ y
Dynamic Analysis of Rectangular Aluminum Plate … 225
Bρh 3 ∂ 3 ψ y ∂ 2m y ∂ 2m x y ∂ 2w
+ + + − ρh + kw
12 ∂ y∂t 2 ∂ y2 ∂ x∂ y ∂ y2
∂ 2w
+ m f 2 = p(x, y, t) (1)
∂t
where ψx and ψ y are local rotations in the x and y directions respectively, mx and
my are bending moments in the x and y directions, respectively, mxy is the twisting
moments, h and h1 are thickness of the plate and load, respectively, ρ and ρL are
the densities of the plate and the load per unit volume, respectively. w(x, y, t) is the
traverse displacement of the plate at time t, g is acceleration due to gravity. K is the
foundation stiffness. Also, B = Bx By , where Bx and By are well-defined in papers
[1, 2, 4]. The right-hand side of Eq. (1) which represents the applied force can be
expressed as follows [12–14]:

1 ∂ 2w
p(x, y, t) = −M L g − M L 2 B (2)
με ∂t
Substituting Eq. (2) into Eq. (1) and rearranging gives:

3
Bρ 3 ∂ 3 ψx ∂ 3ψy 3 ∂ ψx ∂ 3ψy ∂ 2m x ∂ 2m x y ∂ 2m y
h + + Lh + + + 2s +
12 ∂ x∂t 2 ∂ y ∂t 2 1
∂ x∂t 2 ∂ y ∂t 2 ∂x2 ∂ x∂ y ∂ y2
2
∂ w
2
∂ w
2
−M L ∂ w
− ρh 2 + K w + m f 2 = g+ 2 B (3)
∂y ∂t μξ ∂t
Application of the boundary conditions to the non-dimensional form of Eq. (3)

yields the equations that were solved. A simply supported rectangular Aluminum
plate has been taken as an illustrative example. If the edge y = 0 of the Aluminum
plate, the deflection w along this edge must be zero. Also, there are no bending
moments along this edge, this makes it rotate freely with respect to the x-axis [15–17].
3 Solution Method
Equation (3) was solved using a numerical method based on the finite difference
algorithm. This third-order partial differential equation was converted to first-order
partial differential equations, then transformed into its equivalent algebraic form
using finite difference method.
The finite difference definition of first-order partial derivative of a function F(x,
y, t), say, with respect to x, y and t, respectively, can be written as follows [18–20]:
∂F 1 K +1 K +1 K +1 K +1

= ∗ Fi+1, + F + F + F − F K
− F K
− F K
− F K
∂t 4r j+1 i+1, j i, j+1 i, j i+1, j+1 i+1, j i, j+1 i, j
(4)
∂F 1 K +1 K +1 +1

= ∗ Fi+1, + Fi+1, − Fi,Kj+1 − Fi,Kj+1 + Fi+1,
K
j+1
K
+ Fi+1, j − Fi,Kj+1 − Fi,Kj (5)
∂x 4h j+1 j
∂F 1 K +1 K +1 K +1 K +1 K K K K

= ∗ Fi+1, j+1 + Fi+1, j − Fi, j+1 − Fi, j + Fi+1, j+1 + Fi+1, j − Fi, j+1 − Fi, j (6)
∂y 4k
where, F is the function value of the centre of a grid, which is well approximated by
the average of its values at the grid nodes [20].

h∗ k∗ r∗
F x + , y + ,t +
2 2 2
1 K +1 K +1 K +1 K +1

= Fi+1, j+1 + Fi+1, j + Fi, j+1 + Fi, j + F K
i+1, j+1 + F K
i+1, j + F K
i, j+1 + F K
i, j
8
(7)
The above finite difference definition was used on Eq. (3) and a set of algebraic
equations to be solved for the dependent variables emerged. These sets of alge-
braic equations were written in matrix form and solved using computer programs in
conjunction with computer software—Octave.
4 Results Discussion
The numerical calculations were carried out for a simply supported rectangular Alu-
minum plate supported by a simple subgrade subjected to a moving load. Both
rotatory and damping effects were neglected. The dynamic response of the Alu-
minum plate at various times, for a specific value of foundation stiffness, K, and
different values of velocity, was evaluated and represented in Fig. 1. It can be seen
that the maximum amplitude is highest at the highest value of velocity considered.
This implies the higher the velocity the higher the deflection of the aluminum plate.
Figure 2, on the other hand, shows the Comparison of the effect of different values
of foundation stiffness, K, on the deflection of Aluminum plate resting on a sub-
grade when velocity u takes a particular value. It was observed that the higher the
foundation stiffness the less the deflection of the aluminum plate. The effect of the
contact area of the load with the plate was also considered; Fig. 3 shows the plotting
for different values of contact area, A, as a function of time. It was observed that the
Fig. 1 Deflection of
aluminum plate at K =
200 Nm2 and different
values of velocity and time
Fig. 2 Comparing the effect

of K = 100 Nm2 with K =
200 Nm2 on the deflection of
aluminum plate resting on
Winkler foundation when
velocity u = 5.5 (m/s)
Fig. 3 Deflection of the

plate at K = 100 Nm2 and
various value of contact area
A (m2 ) and time t (s)
response maximum amplitude of the rectangular Aluminum plate decreases with an

increase in the contact area, A, for a fixed value of velocity U.
5 Conclusion
The structure of interest was Aluminum rectangular plate on an elastic subgrade,

under the influence of a moving load. A numerical algorithm was adopted in solv-
ing the resulting first-order coupled partial differential equations obtained from the
equations governing such simply supported Aluminum plate. The effect of rotating
inertia and damping were neglected but the effect of shear deformation and gravi-
tational force were considered. The elastic plate is the Aluminum rectangular plate.
The study of such a problem is of practical importance because of its applications
in Engineering. Numerical discussion of the deflection of the Aluminum plate on a
subgrade was presented. Among other results, it was observed that the velocity, con-
tact area of the load with the plate, foundation stiffness and the material properties
of Aluminum, affect the deflection of the plate under consideration, as a partially
distributed load moves on it. In Fig. 1, it can be seen that the maximum amplitude of
the deflection occurred when the highest value of velocity; as the velocity increases
the maximum amplitude also increases. Figure 2 shows that the maximum amplitude
of the deflection increases as K decreases in value. In the figure, it can clearly be seen
that the higher the value of the contact area the lower the maximum amplitude of the
deflection. Some assumptions were taken for brevity sake, The study has contributed
to scientific knowledge by showing that velocity, contact area of the load with the
Aluminum plate, the material properties of Aluminum and the subgrade, on which
the Aluminum plate rests, have significant effect on the deflection of the Aluminum
plate to a uniform partially distributed moving load.
Acknowledgements The authors hereby acknowledge the supports of Covenant University and
the University of Johannesburg towards the successful completion of this study.
References
1. Gbadeyan JA, Agarana MC (2014) Dynamic analysis of railway bridges supported by winkler
foundation under uniform partially distributed moving railway vehicle. Int J sustain Develop
plann 135(of WIT Transactions on built environment):873–883
2. Agarana MC, Gbadeyan JA (2015) Dynamic response of mindlin elastic plate resting on
pasternak foundation under uniform partially distributed moving load. Aust J Appl Math
2(3):17–26
3. Agarana MC, Ede AN (2016) Application of differential transform method to vibration analysis
of damped railway bridge on pasternak foundation under moving train. In: Proceedings of the
world congress on engineering, June 29–July 1, London, UK, pp 978–981
4. Dowell EH (1974) Dynamic analysis of an elastic plate on a thin elastic foundation. J Sound
Vib 35:343–360
5. https://simple.wikipedia.org/wiki/Aluminium
6. https://www.corrosionpedia.com/definition/1599/transverse-loading
7. https://en.wikipedia.org/wiki/Finite_difference_method
8. Agarana MC, Gbadeyan JA, Emetere M (2016) On response of elastic isotropic damped shear
highway bridge supported by sub-grade to uniform partially distributed moving vehicle. Int J
Appl Eng Res 11(1):pp 244–258, ISSN: 0973–4562
9. Law RTDCS (1982) 1982 for vibrating plates. Am J Phys 50(3)
10. Fryba L (1972) Vibration of solids and structures under moving loads. Noordloff International
Publishing Groningen, The Netherlands
11. Civalek O (2005) Large deflection static and dynamic analysis of thin circular plates resting
on two parameter elastic foundation HDQ/FO couple methodology approaches. Int J Comput
Mech 2(2):271–29
12. Agarana MC, Gbadeyan JA, Agboola OO, Anake TA, Adeleke OJ (2015) Dynamic response
of an inclined railway bridge supported by winkler foundation under a moving railway vehicle.
Aust J Basic Appl Sci 9(11):355–361
damped highway supported by pasternak foundation under moving aerodynamic automobile.
978–981
transform method. Int J Math 26(3):0974–5718
15. Gbadeyan JA, Dada MS (2006) Dynamic response of a Mindlin elastic rectangular plate under
a distributed moving mass. Int J Mech Sci 48:323–340
16. Zhang T, Zheng G (2010) Vibration analysis of an elastic beam subjected to a moving beam
with flexible connections. Am Assoc Civ Eng 136(1):120–130
17. Gbadeyan JA, Dada MS (2007) The effects of linearly varying distributed moving loads on
beams. J Eng Appl Sci 2(6):1006–1011
18. Friswell MI, Adhikan S, Lei Y (2007) Vibration analysis of beams with non-local foundations
using the finite element method. Int J Numer Methods Eng 71:1365–1386
19. Agarana MC, Gbadeyan JA, Ajayi OO (March, 2016) Dynamic response of inclined isotropic
elastic damped rectangular Mindlin plate resting on Pasternak foundation under a moving load.
In: Proceedings of the international multi conference of engineers and computer scientists 2016.
Hong Kong, pp 713–716
20. Agarana MC, Gbadeyan JA (October, 2015) Finite difference dynamic analysis of railway
bridges supported by Pasternak foundation under uniform partially distributed moving Railway
vehicle. In: Proceedings of the world congress on engineering and computer science 2015, vol.
II. WCECS 2015, San Francisco, USA, pp 21–23
Ballistic Performance of Light Weight
Magnesium (AZ31B) and Aluminium
(AL 6061) Plates Using Numerical
Method
M. Selvaraj, S. Suresh Kumar, and Ankit Kumar
Abstract Ballistic performance of lightweight Aluminium (Al6061) and Magne-

sium (AZ31B) plates has been determined using numerical method. In recent days,
Aluminium and Magnesium plates are widely used in various structures because of
their light weight and low density. Magnesium is 33% lighter than Aluminium and
77% lighter than Steel whereas Aluminium is one-third the weight of Steel. In the
present work, the numerical ballistic performance has been investigated for differ-
ent projectile velocity using ABAQUS finite element code to understand the impact
behaviour, energy-absorbing capability and failure modes. The conical-shaped pro-
jectile was used and the plate thickness for both alloys was kept constant. The inves-
tigation has been done with three different velocities such as 100, 400 and 800 m/s.
For the constant volume of Magnesium and Aluminium targets considered, lower
depth of penetration and higher ballistic performance was observed for Magnesium
target. The simulation result was verified by conducting a low-velocity impact test
on the Gas Gun Test Setup and a good correlation was observed.
Keywords Ballistic performance · AZ31B magnesium plate · Al6061 aluminium

plate · Depth of penetration
1 Introduction to Light Weight Targets
Magnesium and Aluminium alloys are gaining popularity in almost all the field of
engineering because of their novel feature of light weight to high strength charac-
teristic. Recent research works indicate that the use of Aluminium and Magnesium
alloys is increasing day to day especially in the field of ballistic protection and armour
applications. Both these alloys are ductile, highly formable, light weight and possess
good strength. But still, use of these alloys in ballistic protection is limited due to
M. Selvaraj (B) · S. Suresh Kumar · A. Kumar

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar (SSN) College of
Engineering, Kalavakkam, Chennai 603110, India

https://doi.org/10.1007/978-981-15-4488-0_21
232 M. Selvaraj et al.
the cost and time involved in developing material accordance with ballistic charac-
teristics. The preparation of light weight alloys for possible ballistic protection is
limited due to the lack of exposure in metallurgical behaviour of lightweight mate-
rials. A large number of experiments are done to improve the light weight material
and rework it to increase its strength as well as stiffness for possible ballistic appli-
cations. Friction stir processing of alloys using nanoparticles (Carbon Nano Tubes,
Zirconium Oxide and Graphene) to improve surface stiffness is also done. The exper-
imental ballistic impact is a highly non-linear dynamic phenomenon exhibiting high
strain rates, deformation and fracture with constantly varying boundary conditions
enabling researchers to opt for numerical simulation. Numerous experimental inves-
tigations have been conducted so far for a normal or an oblique projectile impact on
the homogeneous or sandwiched metallic or composite plate. But the cost and time
involved in ballistic experiments restrict the researcher’s dependency on experiments
for all impact-related studies. The complexity of the ballistic impact and penetra-
tion events often limits the general use of closed-form analytical solutions. Thus,
a numerical simulation analysis is often preferred and resorted as a supplement to
ballistic experiments. Simulations help in developing new projectiles and armours
in a shorter period and permit easy design modifications and improvements. They
reduce the experimental needs to a minimum extent of limiting to acceptance or a
qualification test. However, only a few numerical studies at ordnance velocities are
available because it is largely dependent on numerical inputs (material and fracture
models) and numerical formulations.
Recent developments in commercial finite element codes are capable of simu-
lating the complex ballistic impact events. The numerical simulation using finite
element (FE) codes requires a sophisticated constitutive model, equation of state
and in-built failure criteria as numerical inputs either individually or in combination
to appropriately capture the complete behaviour of ballistic events. A number of
constitutive models are available in the literature with varying capabilities in charac-
terizing the material behaviour during impact. Many of them are material dependent
and developed empirically or semi-empirically based on plasticity approach. They
offer different levels of difficulty in finding a number of material constants presented
in material models from different physical tests that are mostly and completely not
available in the public domain.
1.1 Concept of Ballistic Impact Mechanism
At low impact velocities (about few hundred meters per second), the penetration resis-
tance of a material is governed by the dynamic deformation mechanisms within the
projectile and target. However, as the impact velocity increases into the hypervelocity
regime (several thousand meters per second), hydrodynamic effects dominate and
the penetration behaviour becomes controlled by only the density of the impacted
material and projectile. Since the resistance of a material against penetration by
low-velocity projectiles is controlled by dynamic deformation and impact fracture
Ballistic Performance of Light Weight Magnesium … 233
mechanisms, many researchers have involved themselves in the study of ballistics.

In general, the ballistic performance depends upon the thickness, strength, ductility,
toughness, acoustic impedance and density of both the target material and projectile
and the velocity of the projectile. Figure 1 shows the different failure mechanisms
caused by ballistic impact.
For thin targets that are much softer than the projectile, the panel bends creating
tensile hoop stress in the plate. This leads to a failure by the growth of several
radial cracks which leads to a “petalling mode” of failure. Whereas thicker materials
projectile/material combinations lead to crater formation on the impact face and
a “plug” that is pushed out of the plate. The crater forms because of the plastic
flow needed to accommodate the volume of the projectile. The “plug” results from
high shear stresses and large (adiabatic) shear band formation in the material near
the periphery and just ahead of the projectile. In this case, increasing the material’s
shear strength and strain and strain rate hardening are important. For shear plugging to
occur in thicker targets, very high stress is required. Instead, the projectile penetrates
Fig. 1 Projectile penetration

mechanism for a soft
material impacted by a hard
projectile [1]
by “ductile hole enlargement” which is governed by the dynamic flow stress of the
target material. Efforts to increase the resistance to penetrate have focused upon the
development and use of materials of very high yield strength. However, other failure
modes can then begin to become important (for example, spallation and fracture).
Visco-plasticity is a theory in continuum mechanics that describes the rate-dependent
inelastic behaviour of solids. Rate-dependence indicates that the deformation of the
material depends on the rate at which loads are applied. The inelastic behaviour
causes the material to undergo unrecoverable deformations when a load level is
reached. Rate-dependent plasticity is important for transient plasticity calculations.
2 Literature Review
Many researchers have studied the ballistic performance of aluminium-based targets.

Gupta et al. [2] conducted experiments on aluminium plates of one mm thickness by
using a gas gun and projectiles with blunt and hemispherical noses. Target plate was
impacted with varying impact velocity. Impact and residual velocities of the projec-
tile were measured. It was noticed that ballistic limit was higher for hemispherical
projectiles than that for blunt projectiles. The effect of nose shape on the deformation
of the plate was also studied. Numerical simulations of the impact were conducted by
using an explicit finite element code. “Johnson–Cook elasto-viscoplastic model” was
used to carry out the analysis. Results obtained from finite element simulations were
compared with those of experiments. Good correlation was found between the two.
Iqbal et al. [3] conducted an experimental and finite element investigation to explore
the influence of target to projectile diameter ratio (D/d) on the ballistic performance
and failure mechanism of thin aluminium plates. 1100-H12 aluminium target plates
of 1 mm thickness were impacted by 19 mm diameter ogive and blunt-nosed pro-
jectiles. The D/d ratio was varied by varying the span diameter of the target keeping
the projectile diameter constant, 19 mm. The finite element simulations were carried
out at D/d ratios 3.6, 5, 7.9, 10, 15, 20, 25, 30, 35 and 40. The validation of the
simulation results was carried out by performing experiments at D/d ratio 3.6, 5, 7.9,
10 and 15. The total work done in plastic deformation of the target was disintegrated
into circumferential, radial, axial and tangential stretching. An initial increment in
D/d ratio from 3.6 to 10 has been found to have a prominent effect on the ballistic
limit, particularly against blunt-nosed projectile. It was observed that subsequent
increase in the D/d ratio, however, could not influence the ballistic resistance signif-
icantly. The energy absorption in plastic deformation was found maximum for D/d
ratio 10. The maximum energy dissipation occurred in circumferential stretching
against bunt and in tangential stretching against ogive nosed projectile. The mini-
mum energy dissipated in axial stretching against both the projectile. Madhu et al.
[4] studied the phenomenon of ordnance velocity impact of a projectile on mono-
lithic plates. In their work, an experimental study of normal and oblique impacts
of all ogive shaped, hard steel projectile on single and layered plates of mild steel
and aluminium were investigated. The projectiles were fired at an impact velocity of
about 820 m/s. The plate thickness was varied in the range 10–40 mm and the ratio of
plate thickness to the diameter of the projectile varied in the range 1.5–13.0. Obser-
vations on target damage and measurements of the incident and residual velocities
for different angles of impact were presented. Plate thickness for which the incident
velocity is the ballistic limit was determined. Computer simulations were carried out
using a hydrodynamic code to simulate the normal impact of a projectile and they
compared these with the experimental results. They performed the experiments to
evaluate the response of these plates of intermediate thickness when layered, and
the results were compared to the results of a single plate of the same total thickness.
The aforementioned attempts are limited to a single steel target subjected to various
projectile impact. Lack of work was observed to determine the ballistic performance
of lightweight targets using numerical method. In the present work, effort has been
made to determine the ballistic performance of light weight targets using numerical
method.
3 Numerical Determination of Ballistic Performance
The numerical simulation was carried out using finite element code ABAQUS 6.14,
which is well known for simulation of non-linear dynamic explicit impact testing
problems which gives an accurate result. The software has good flexibility to manipu-
late the target material properties, testing procedure, controlling environment, taking
most of the inputs into account for getting almost accurate result moreover it is
very user friendly and flexible with varying requirements. The target used for Mg
AZ31B was 115 mm × 115 mm × 4 mm and Al6061 was 115 mm × 115 mm ×
3 mm and the conical steel projectile was used with 9 mm diameter. The projectile
hits on the target at 0° angle at three different velocities such as, low (100 m/s),
medium (400 m/s) and high (800 m/s). Both the Magnesium and Aluminium targets
were meshed with 10,894 elements and 16,692 nodes whereas projectile was meshed
with 169 elements and 168 nodes. The field outputs required for these simulations
are Principle stress, Principle strain, Johnson–Cook damage initiation criteria at the
integration point (JCCRT) and Kinetic energy absorbed by the target. Material prop-
erties of AZ31B and Al6061 with Johnson–Cook damage model is assigned to the
target plate. Boundary conditions were applied and step time was given for solving
the simulation. The edges of the plate were completely constrained for all degrees of
freedom and the projectile was provided with only Z-axis displacement at different
velocities as mentioned above according to the requirement. After had assigned the
boundary conditions, the simulation was run to determine the output parameters.
Mesh convergence study was conducted in order to determine the optimum num-
ber of elements. Figure 2a–c shows the meshed target and projectile. The tensile
yield strength of the Aluminium and Magnesium targets are 276 MPa and 200 MPa
respectively.
Fig. 2 a Target plate with

projectile, b meshed target
plate, c contact region of
target and projectile
The numerical simulation was mainly focused to determine the maximum princi-
ple stress (S) at the integration point, equivalent plastic strain (PEEQ) at the integra-
tion point, Johnson–Cook damage initiation criteria at the integration point (JCCRT)
and Kinetic energy absorbed by the plate. The simulation gives a closer observation
of how the material is about to behave under different impact conditions. The test
specimen was modelled in such a way that, the weight of both the alloys remains
same for various target thickness. Since Magnesium alloy is lighter than aluminium
alloy, the plate thickness of aluminium specimen will be less than the Magnesium
specimen to keep their weight same. The dimensions of Mg AZ31B is 115 mm ×
115 mm × 4 mm. The optimum thickness for aluminium is calculated as follows:
(Density × Volume) Al = (Density × Volume) Mg
2700 × 0.115 × 0.115 × t = 1780 × 0.115 × 0.115 × 0.004
Thickness, t = 0.00263 m ≈ 0.003 m.
Standard thickness of 3 mm was considered to facilitate easy availability of Alu-
minium plate. The projectile used in the present work is of standard steel material
which is hardened and is conical in shape. The dimensions are of 9 mm diameter and
15 mm length.
4 Result and Discussion
Ballistic performance of lightweight Aluminium and Magnesium targets subjected

to different velocities of the projectile was carried out using ABAQUS numerical
code. The following observations were made from the numerical simulation. It was
noted that maximum induced impact stress in AZ31B is more than in Al6061 due to
which a crack is formed on the rear portion of AZ31B plate. Both targets showed the
formation of a dent and no crack was observed in Aluminium target. Thus one can
expect Magnesium target can withstand more impact stress compared to Aluminium
target. Figure 3 shows the impact stress distribution of light weight targets for a
projectile velocity of 100 m/s.
It was also noticed that plastic Strain in AZ31B is less compared to Al6061. This
is due to higher elongation of, Al6061 compared to AZ31B before failure, which
also proves aluminium is more ductile than Magnesium. JC damage initiation in
AZ31B is less than in Al6061 which signifies that aluminium is more ductile it starts
to deform earlier when compared to Magnesium. Higher depth of penetration was
observed for Al6061 than AZ31B plate due to higher plastic strain.
Figure 4 shows the impact stress distribution of light weight targets at higher
velocity. At higher projectile velocity (800 m/s) higher impact stress was observed
for Magnesium target compared to Aluminium. Thus one can expect higher impact
Fig. 3 Projectile velocity of

100 m/s
a Magnesium Target b Aluminium Target

(AZ31B) (Al6061)
c Depth of Penetration - d Depth of Penetration -

Magnesium Target (AZ31B) Aluminium Target (Al6061)
Fig. 4 Projectile velocity of

800 m/s
Magnesium Target (AZ31B) Aluminium Target

(Al6061)
Depth of Penetration - Magnesium Depth of Penetration -

Target (AZ31B) Aluminium Target
(Al6061)
stress for Magnesium target irrespective of projectile velocity. This may be due to
higher thickness of Magnesium target compared to Aluminium for the same volume
considered. In this condition also lesser plastic strain and JC damage initiation was
observed for Magnesium target compared to Aluminium due to higher ductility of
Aluminium target compared to Magnesium.
5 Summary
Numerical and Experimental estimation of ballistic performance was performed on

Magnesium AZ31B and Aluminium 6061 targets to compare their failure behaviour
under various projectile impact velocity. The impact velocity ranging between 100
and 800 m/s was considered for numerical simulation and 100 m/s was considered
for experimental work. Good correlation of depth of penetration was observed for
the numerical and experimental results. One can expect better ballistic performance
for Magnesium target compared to Aluminium target of the same weight. Hence,
AZ31B can be recommended to replace Al6061 wherever the application is suitable,
since it is lighter in weight and better in impact strength compared to Al6061.
References
1. Arias A, Rodríguez-Martínez JA, Rusinek A (2008) Numerical simulations of impact behaviour

of thin steel plates subjected to cylindrical, conical and hemispherical non-deformable
projectiles. Eng Fract Mech 1635–1656
2. Gupta NK, Iqbal MA, Sekhon GS (2006) Experimental and numerical studies on the behavior
of thin aluminum plates subjected to impact by blunt-and hemispherical-nosed projectiles. Int J
Impact Eng 32(12):921–1944
3. Iqbal MA, Tiwari G, Gupta PK, Bhargava P (2015) Ballistic performance and energy absorption
characteristics of thin aluminium plates. Int J Impact Eng 77:1–15
4. MadhuV, Bhat TB, Gupta NK (2003) Normal and oblique impacts of hard projectile on single
and layered plates—an experimental study. Defence Sci J 53(2):147–156
Comparison of Energy Absorption
Characteristics of the Plain Fold
and Spot-Welded Fold Tubes Under
Three-Point Bending
Abstract Thin-walled square tubes are used as automotive vehicular crash struc-
tures since they can be easily constrained at the endpoints. These tubes are generally
produced by costly extrusion process and special mold designs are required for mak-
ing the same. This research work proposes a novel idea of making by folding thin
metal sheets into square tubes that can be easily prepared, cost-effective and flexible
in sectional shape. However, spot welding can be used to enhance the bending resis-
tance and energy absorption. Bending collapse of tubes is one of the most important
deformation mechanisms to be ascertained to ensure the safety of people or cargo
under accidental crash events. Three-point bending tests are generally employed
to investigate the bending characteristics and hence performed on plain folded and
spot-welded fold tubes. Results revealed that spot-weld and number of turns have an
important influence on the bending resistance of the folded tubes. The comparison
between traditional tubes and welded tubes showed that the welded tubes outperform
the traditional tubes in some aspects like minimizing the peak force and maximiz-
ing energy absorption. Adopting a number of turns further increased the bending
resistance of the tubes compared with traditional square tubes.
Keywords Crush force · Fold tubes · Spot-weld
1 Introduction
In the past decades, thin-walled tubes are widely used as energy absorbers as they can
absorb a large amount of energy during a vehicle crash event [1–5]. Among them,
square tubes are mostly preferred since they can be easily constrained at end points
[6]. Particularly, square tubes made of aluminium are preferred due to its lightness
M. Nalla Mohamed (B) · R. Sivaprasad

Engineering, Kalavakkam, Kancheepuram District, Tamilnadu, India
R. Sivaprasad

https://doi.org/10.1007/978-981-15-4488-0_22
240 M. Nalla Mohamed and R. Sivaprasad
and progressive deformation behaviour during crash [7]. Hence, investigation of

energy absorption and folding mechanisms of aluminium square tubes has been
done in the last decade using analytical, experimental and numerical methods [8–
10]. In recent years, multi-wall thickness square tubes have been used to improve the
crashworthiness performance of tubes. A common obstacle for the manufacturing
of such tubes is that (i) costly extrusion process and (ii) special mold is required
to fabricate such structures. Of particular interest, this research work proposes a
novel idea of making by folding thin metal sheets into square tubes that can be
easily prepared, cost-effective and flexible in sectional shape. Multi-wall thickness
and multi-cell sections can be easily obtained by the number of folding layers. This
method of preparation is cost-effective as well as obtained with less effort when
compared to extrusion method [11]. At the same time, the bending resistance of the
folded tubes is not as good as heir axial crushing resistance. However, the loading
directions have an important influence on the bending resistance of folded square
tubes [12]. Folded tubes with weaker configurations in the compression region show
less bending resistance and energy absorption. Hence upward closing is needed
for folded tubes to improve bending resistance of the structure. Various methods
are available to improve the crashworthiness performance and increase the energy
absorption efficiency of them. Upward closing through welding is one of the options
to change the deformation behaviour and increase the energy absorption capacity and
efficiency of them [13]. Several investigations have been performed experimentally,
theoretically and numerically to study the bending behaviour of the welded section.
For instance, experimental studies have been performed to test the bending of the
simple structure joined by spot-weld [14, 15]. Some researchers examined the static
collapse and the bend behaviour of hybrid hat section stub columns [16, 17]. The
above-mentioned literature review shows that more information is available on the
bending behaviour of square tubes but that relating to folded tubes with welded
section was scanty. This shortcoming motivated the authors to suggest a new design
of folded tubes for enhancing the safety of the vehicles. In this paper, bending collapse
responses of plain fold and weld fold tubes under three-point bending are investigated.
The bending resistance and energy absorption characteristics of suggested tubes
with various configurations are analyzed, and the outcomes are compared with that
of traditional extruded tubes. The relative merits of tubes are analyzed, and some
conclusions are drawn for such folded tubes subjected to bending loads.
2 Experimental Methodology
2.1 Material Properties
The structural material used for the folded tubes in this experiment is aluminium
alloy AA6061-O with a chemical composition of 98.05% Al, 0.62% Fe, and 0.41%
Mg. Mechanical properties of AA6061-O were determined using standard tensile
Comparison of Energy Absorption Characteristics … 241
Fig. 1 Stress–strain curves of AA6061-O
specimens as defined in ASTM standard E8M. The aluminium alloy 6061-O was
selected due to its common usage in automotive parts and aircraft structures, such as
wings and fuselages for crash energy management. Figure 1 presents the engineering
stress–strain curves for AA6061-O.
2.2 Fabrication of Plain Fold and Weld Fold Square Tubes
Aluminium alloy AA6061-O sheets of 1 mm thickness were folded into square tubes.
During the folding of sheets, fillets were created along the lateral edges which
reduces the stress concentration during bending, unlike fillets are rare in commer-
cially available extruded square tubes. The effective thickness of the fold tube was
increased by increasing the number of turns as shown in Fig. 2.
Four types of both plain fold and weld fold tubes were fabricated with AA6061-O
sheets with 1 mm thickness as shown in Figs. 3 and 4, respectively. The length of each
specimen is 180 mm and average cross-sectional width is 30 mm and corner radius
is measured to be 2 mm. During bending the sheet, it was ensured to give bending
allowances to compensate spring back and distortion errors and ensured to achieve
structural accuracy. In the case of weld folded tubes, spot welds of 6 mm diameter
were provided longitudinally along the loose end of the folded tube at 30 mm of
interval.
Fig. 2 Dimensions for the specimens
Fig. 3 Fabricated plain fold

tubes
2.3 Experimental Setup
Quasi-static three-point bending tests were performed in a computerized UTM of 60

t capacity with computer control and data acquisition systems, which is shown in
Figs. 5 and 6. The diameter of the cylindrical punch is 25.4 mm and the span between
the supports 140 mm. The displacement of the punch head was moved at 2 mm/min
during quasi-static loading. The punch force response versus punch displacement
data was recorded by a digital data acquisition system for each specimen.
Fig. 4 Fabricated weld fold

tubes
Fig. 5 Three-point bending

setup
Fig. 6 Geometry for

three-point bending test
The deformation patterns and punch force–displacement responses of the plain fold
and welded fold tubes are analyzed here. The closing side of the tube is kept at the
top and in contact with the punch. Figure 7 shows the bending deformation history
of the “P1 ” tube at the difference punch displacement level. It was observed that
the deformation in the corner regions is the major energy dissipation mechanism
for bending collapse. It was also observed that, in P1 and P2 types, the overlapping
walls of the tubes separated during the loading which leads to the weakening of
Fig. 7 Bending deformation

history of P1 type fold tube
corner regions. Hence, it leads to the drop of force response and consequently, the
energy dissipation of P1 and P2 is less than that of other types. But P3 and P4
exhibit the highest bending resistance. This is due to the fact that there are three
vertical sides to deform and dissipate energy when bending in this direction. The
dissymmetry resistance of the three vertical sides also brings some extent of twist
during the deformation of specimen, as shown in Fig. 8. Figure 9 shows the punch
force versus displacement curves for all the plain folded tubes. Figure 10 shows
energy versus displacement curves for all the plain folded tubes. When the number
of turns increased, the bending resistance was also increased. Maximum peak force
is obtained for P4 type and minimum peak force is obtained for P1 type. As the
bending resistance increased consequently the energy absorption also increased.
It should be mentioned that the welding had some influence on the bending
responses of the folded tubes. It is interesting to find that, highest energy absorption
and mean bending resistance were observed for all the four types of welded fold tubes
Fig. 8 Deformation pattern

of all the plain folded tubes
Fig. 9 Punch force versus

displacement curves for all
the plain folded tubes
Fig. 10 Energy versus

displacement curves
compared to their corresponding plain folded tubes. For type I, the difference is about
10%, while for all other types, the difference is 25% compared to their counterpart.
The increase of energy absorption for a corresponding increase of turns is relatively
regular and the increase in percentage is generally around 10–25% (Figs. 11 and 12).
Figure 11 presents deformed shape of welded tubes (W 1 ) under 3-point loads with
different span. Final deformed shape of all the welded tubes under 3-point bending
test is presented in Fig. 12.
The performances of plain fold and weld fold tubes under three-point bending
loading were discussed in the previous sections. However, it should be proved that
whether they outperform traditional single tube. The force responses of the welded
tubes are compared with traditional square tubes (SQ) having the same width of
30 mm and thickness 1 mm in Fig. 14. As shown in Fig. 13, the bending resistance
Fig. 11 Deformed shapes of

welded tubes (W 1 ) under
3-point loads with different
span
Fig. 12 Final deformed

shape of welded tubes in
3-point bending test
of W 4 is superior to the traditional extruded square tubes. Similarly, the energy

absorption capacity of W 4 tube is 2 times than that of traditional square tubes which
is evidenced from Fig. 15.
4 Conclusion
Quasi-static three-point bending characteristic of plain fold and weld fold tubes is
analyzed experimentally in the present work. The bending resistance and energy
absorption characteristics are compared, and the conclusions are summarized as
follows:
• The P1 type plain folded tubes failed easily due to separation of fold side and
hence the energy absorption is less. Hence use of single side fold tubes is not
advisable for crashworthy applications.
P1 &W1 P2 &W2
P3 &W3 P4 &W4
Fig. 13 Comparison of punch force versus displacement curves between plain fold and weld fold
tubes
Fig. 14 Comparison of
punch force versus
displacement curves between
extruded square tube and
weld fold tube
energy dissipation capacity
of proposed tubes with
extruded square tube
• The energy absorption and mean crushing force of folded tubes under three-point
bending are found to be directly proportional to the number of vertical side panels.
Hence P3 and P4 types exhibited more bending resistance compared to P1 and
P2 types. The application of folded tubes with more vertical side panels is hence
suggested for transverse loading.
• Experimental results proved that the weld folded tubes had more stable response
due to the non-separation of the overlapping side compared with plain folded
tubes. Hence weld tubes with more vertical sides had more bending resistance
and energy absorption than the plain folded one.
• Bending resistance and energy absorption of the weld fold tube is superior to
the extruded square tube. Hence the cost-effective and easily prepared approach
of manufacturing tubes through metal sheet bending is suggested for safety
protection in automotive applications.
References
1. Abramowicz W, Jones N (1986) Dynamic progressive buckling of circular and square tubes.
Int J Impact Eng 4:243–270
2. Alghamdi AAA (2001) Collapsible impact energy absorbers, an overview. Thin-Wall Struct
39:189–213
3. Alavi NAJ, Hamedani H (2010) Comparative analysis of energy a sorption and deformations
of thin walled tubes with various section geometries. Thin-Wall Struct 48:946–954
4. Nalla Mohamed M, Yuvarajan P, Umasankar M (2015) Energy enhancement of long cylindrical
tubes with grooves subjected to axial impact. App Mech Mater 787:345–349
5. Nalla Mohamed M (2017) New insight to energy absorption characteristics of long circular
tubes with stiffeners as controllable energy-dissipating devices. Procedia Eng 173:1300–1406
6. Chen S, Haiyan Y, Fang J (2018) A novel multi-cell tubal structure with circular corners for
crashworthiness. Thin-Wall Struct 122:329–343
7. Zhang X, Zhang H, Ren W (2018) Axial crushing of tubes fabricated by metal sheet bending.
Thin-Wall Struct 122:252–263
8. Jensen, Langseth M, Hopperstad OS (2004) Experimental investigations on the behaviour of
short to long square aluminium tubes subjected to axial loading. Int J Impact Eng 30:973–1003
9. Galib DA, Limam A (2004) Experimental and numerical investigation of static and dynamic
axial crushing of circular aluminum tubes. Thin-Wall Struct 42:1103–1137
10. El-Hage H, Mallick PK, Zamani N (2006) A numerical study on the quasi-static axial crush
characteristics of square aluminum-composite hybrid tubes. Compos Struct 73:505–514
11. Zhang X, Zhang H, Ren W (2016) Bending collapse of folded tubes. Int J Mech Sci 117:67–78
12. Bai J, Meng G, Wu H, Zuo W (2019) Bending collapse of dual rectangle thin-walled tubes for
conceptual design. Thin-Wall Struct 135:185–195
13. Zhang X, Zhang H, Leng K (2019) Experimental and numerical investigation on bending
collapse of embedded multi-cell tubes. Thin-Wall Struct 127(2018):728–740
14. Zhang X, Zhang H, Wang Z (2016) Bending collapse of square tubes with variable thickness.
Int J Mech Sci 106:107–116
15. Bilston D, Ruan D, Candido A, Durandet Y (2019) Parametric study of the cross-section shape
of aluminum tubes in dynamic three-point bending. Thin-Wall Struct 136:315–322
16. Lee S, Hahn C, Rhee M, Oh JE (1999) Effect of triggering on the energy absorption capacity
of axially compressed aluminum tubes. Mater Des 20:31–40
17. White MD, Jones N (1999) Experimental quasi-static axial crushing of top-hat and double-hat
thin-walled sections. Int J Mech Sci 41:179–208
An Efficient Energy Absorber Based
on Welded Fold Tubes for Automotive
Applications
Abstract Thin-walled square tubes have proven their energy-absorbing capabil-

ity in the automotive industries for safety applications. Unfortunately, special mold
designs are required while making these tubes through extrusion process. This fab-
rication technique is in lack of flexibility or leads to low cost-effectiveness. In place
of the existing extruded square tubes, this paper presents a novel energy absorber by
folding thin metal sheets into tubes that are easily fabricated and cost-effective. How-
ever, it is unsafe to use these folded tubes directly due to its global buckling failure
with low-energy absorption. Therefore, methods to improve the energy absorption
characteristics of thin-walled structures have become essential. Spot and continuous
welding are the effective joining techniques which can be used to enhance the energy
absorption. In order to test the suitability of these tubes for crashworthiness, quasi-
static tests were performed using a universal testing machine. The energy absorption
capacity of these welded folded tubes was analyzed and compared with traditional
square tubes with the same mass to quantify the relative merits. It was found that
spot welding leads to a reduction in peak force and continuous welding leads to
a reduction in performance due to its global bending. The results showed that the
initial peak force of spot-welded tube is 10–15% lower than the continuous weld
and 20–25% lower compared to the extruded square tube. Finally, it was found that
spot-welded tubes had more crush length than that of continuous and plain folded
one which leads to an increase in energy absorption. The outcomes of the present
study would facilitate the design of welded tubes with better energy absorption.
Keywords Continuous weld · Folded weld tubes · Spot weld
M. Nalla Mohamed (B) · R. Sivaprasad

Engineering, Kalavakkam, Kancheepuram District, Tamilnadu, India
R. Sivaprasad

https://doi.org/10.1007/978-981-15-4488-0_23
1 Introduction
Thin-walled square tubes are being used in the vehicle body for increasing the crash-
worthiness and consequently decreasing injuries [1]. Over the past few decades, a lot
of effort has been focused on investigating the energy absorption capability of tra-
ditional square tubes using theoretical, numerical, and experimental methods [2–4].
Although the conventional square tubes demonstrated good performance for energy
absorption, the high initial peak force is the crucial problem that has the potential
to cause serious injury to the occupants [5]. Also, special mold designs are required
while making these tubes through extrusion process. The present extrusion process
is in lack of flexibility or leads to low cost-effectiveness [6]. As an alternative, folded
tube by bending metal sheets is put forward in this situation. Moreover, their easy
fabrication, cost-effective and flexible in both sectional shape and geometric param-
eters attracted the researcher. At the same time, these folded tubes cannot be used
directly as energy absorber due to its irregular deformation or switch to global buck-
ling failure mode which leads to the poor energy absorption [7]. If the folded tubes
are in closed-form, the deformation may be progressive again and also much stable
and regular. Hence, they may compete with traditional tubes [8]. The closed-form
of the folded tubes can be achieved by joining techniques like welding. It is a very
cheap and affordable technology to join metals sheets. Several investigations have
been performed experimentally, theoretically and numerically to study the crushing
behaviour of welded section. For instance, experimental studies have been performed
to test the crashworthiness of quasi-static of the simple structure joined by spot weld
[9, 10]. Some researchers examined the static collapse and the bend behaviour of
hybrid hat section stub columns [11, 12].
The above-mentioned literature review shows that more information is available
on the crushing behaviour of square tubes but that relating to folded tubes with welded
section was scanty. This shortcoming motivated the authors to suggest a new design
of folded tubes for enhancing the safety of the vehicles.
This paper aims at addressing the crushing characteristics of spot and continuous
welded fold tubes under axial crushing experimentally. Folded tubes were tested
for comparison purposes. The performance indices such as peak crush force (PCF)
and the energy absorption capacity (EA) of the proposed tubes were compared with
traditional square tubes for better understanding the merits.
2 Experimental Methodology
The material used for the folded tubes in this experiment is aluminium alloy AA6061-
O with a chemical composition of 98.05% Al, 0.62% Fe, 0.41% Mg. Mechanical
properties of AA6061-O were determined using standard tensile specimens as defined
An Efficient Energy Absorber Based on Welded … 253
Fig. 1 Geometric details of tensile specimen
Fig. 2 Stress–strain curves

of AA6061-O
in ASTM standardE8M. The geometric details and the fractured specimen after the
test are shown in Fig. 1. Figure 2 presents the tensile engineering stress–strain curves.
2.2 Fabrication of Spot and Continuous Folded Tubes
First, two sets of four different types of plain folded square sections were fabricated
using Aluminium alloy AA6061-O sheets with 1 mm thickness as shown in Fig. 3.
The length of the specimens is 180 mm and the width of the inside square is 30 mm.
The aluminum sheets are bent by using a square steel rod and fillets are formed
‘1’ type ‘2’ type
‘3’ type ‘4’ type
Fig. 3 Dimensions for the specimens
in the corners. The radius of the fillets in the specimens is measured to be 2 mm.
Special care was taken to achieve a very high structural accuracy with reduced spring
back and artificial errors. Then, spot welds were done with 6 mm diameter spots and
30 mm pitch in the first set of plain folded tubes [13]. Continuous welds were also
done with the second set of plain folded tubes. Representative specimens are shown
in Fig. 4.
2.3 Equipment and Procedure
In order to better understand the deformation characteristics of proposed tubes, the

quasi-static axial compress tests were performed. Universal Testing Machine (UTM)
of model number TUE CN-600 with 60 t capacity manufactured by Fine group’s
company shown in Fig. 5 is chosen to conduct the test. The test tubes are placed
between two parallel rigid platens. The tests were conducted by a cross-head moving
Fig. 4 Fabricated tubes specimen spot and continuous welded tubes
Fig. 5 UTM machine
at preselected speeds of 2 mm/min. The compressive load was gradually applied up

to the deformation was 80% of the tube length. The variation of reacting force and
axial displacement was recorded automatically by a digital data acquisition system.
Typically, these data can be used to derive information on the energy-absorbing
characteristics.
3.1 Quasi-static Test on Plain Folded Welded Square Tubes
To investigate the influence of the welding, we compare the crash behaviour of the
folded tubes with spot and continuous welding in this section. The comparison of
the deformation history of the spot-welded (SW1-type specimen) and continuous
welded fold tubes (CW1 type) at different values of axial compression are shown
in Fig. 6. It is observed that all types of spot-welded tubes deformed progressively
which leads to more energy absorption. But, in con welded tubes, the fold started
with progressive deformation followed by global buckling failure. Hence, during
experiments, tests were stopped when the tubes lost structural resistance. This type
of global buckling failure is recommended to be avoided in crashworthy applications
due to their unpredictable results in energy absorption performance. The comparison
of the final deformation mode of all the welded fold tubes is given in Fig. 7.
a) Spot weld specimen b) Continuous weld specimen
Fig. 6 Comparison of the deformation history of spot weld and continuous weld specimen
Fig. 7 Crushing pattern of all spot welded fold tubes

The comparative results of axial crush force versus deformation behaviour of

the spot and continuous welded tubes tested under static loading conditions are
depicted in Fig. 8. The results showed that the continuous welded tubes showed
a fluctuating mean crush force, whereas the spot-welded tubes showed a gradual
mean force due to its progressive deformation. Initial peak force increased when
the number of fold increased. The total energy absorbed by the spot and continuous
welded tubes obtained through experiments was compared and the comparison is
depicted in Fig. 9. From the comparative plot, it was observed that the spot-welded
tube absorbed more energy than the continuous one. Specimens having higher wall
thickness gained comparatively higher crushing energy absorption capacity than the
other specimens considered. The energy absorption capacity of the spot weld tubes
was found in the range of 405–1144 J and the increase in energy absorption capacity
was estimated as 30–40% compared with continuous weld tubes.
Type ‘1’ Type ‘2’
Type ‘3’ Type ‘4’
Fig. 8 Comparison of axial crush force versus deformation of spot and continuous welded tubes
Fig. 9 Comparison of peak 20

force and energy absorption Spot
of fold weld tubes
15
Peak force(kN)
10
0
I II III IV
Type of Specimen
25
20
Peak force(kN)
Continuous weld
15
10
0
I II III IV
Types of specimen
4 Conclusion
Axial crushing of spot and continuous weld folded tubes was experimentally
investigated in this paper. The major results are summarized as follows:
• Continuous weld folded tubes failed in local buckling under axial crushing. In
addition, the local buckling failure will greatly reduce the crush resistance of
welded tubes and should be avoided by considering the stability of the structure
in design. Instead, spot weld in the overlapped side helped to avoid this instability
during deformation.
• The initial peak force of spot and continuous weld tube is 10–15% significantly
lower than the extruded square tube and can be used as a good alternative to the
conventional energy-absorbing structures that are sensitive to deceleration level.
• The energy absorption efficiency (SEA) of spot-welded tubes approaches that of
the traditional square tube with the increase of the number of overlapped side.
• In summary, in order to meet the mass requirements of energy-absorbing com-
ponents in automobile industries, the spot-welded fold tubes are most suitable as
energy absorbing members due to their cost-effective for small-scale production,
easily prepared, and flexible in both sectional shape and geometric parameters.
References
1. Langseth M, Hopperstad OS (1996) Static and dynamic axial crushing of square thin—walled
aluminum extrusions. Int J Impact Eng 18:949–968
2. Jensen M, Langseth OS (2004) Hopperstad, Experimental investigations on the behaviour of
short to long square aluminium tubes subjected to axial loading. Int J Impact Eng 30:973–1003
3. Galib DA, Limam A (2004) Experimental and numerical investigation of static and dynamic
axial crushing of circular aluminum tubes. Thin-Wall Struct 42:1103–1137
4. El-Hage H, Mallick PK, Zamani N (2006) A numerical study on the quasi-static axial crush
characteristics of square aluminum-composite hybrid tubes. Compos Struct 73:505–514
5. Bodlani SB, Yuen SCK, Nurick GN (2009) The energy absorption characteristics of square
mild steel tubes with multiple induced circular hole discontinuities-part 1: experiments. J Appl
Mech 76:041012
6. Zhang X, Zhang H, Ren W (2018) Axial crushing of tubes fabricated by metal sheet bending.
Thin-Wall Struct 122:252–263
7. Zhang X, Zhang H, Ren W (2016) Bending collapse of folded tubes. Int J Mech Sci 117:67–78
8. Ohkubo Y, Akamatsu T, Shirasawa K (1974) Mean crushing strength of closed-hat section
members. SAE Paper 740040
9. Yoon J, Cazacu O, Mishra RK (2013) Constitutive modeling of AZ31 sheet alloy with
application to axial crushing. Mat Sci Eng A565:203–212
10. Song J, Zhou YF, Guo FL (2013) A relationship between progressive collapse and initial
buckling for tubular structures under axial loading. Int J Mech Sci 75:200–211
11. Lee S, Hahn C, Rhee M, Oh JE (1999) Effect of triggering on the energy absorption capacity
of axially compressed aluminum tubes. Mater Des 20:31–40
12. White MD, Jones N (1999) Experimental quasi-static axial crushing of top-hat and double-hat
thin-walled sections. Int J Mech Sci 41:179–208
13. Song HW, Fan ZJ, Yu G, Wang QC, Tobota A (2005) Partition energy absorption of axially
crushed aluminum foam-filled hat sections. Int J Solids Struct 42:2575–2600
Finite Element Analysis
Finite Element Modelling
of a Compression Test on AISI 1016
Cylindrical Steel: A Review
V. Musonda and E. T. Akinlabi
Abstract Predicting the material flow behaviour of steel is always an important

undertaking especially when there is a need to decide on a suitable material required
for a particular application. Compression test is normally considered as a standard
bulk workability test and a common quality control test, which can be applied to
hot forging operations, or cold upset forging. The test is useful when the friction
conditions, especially in hot working, require an evaluation. The flow stress data for
metals at various temperatures and strain rates could also be predicted. The aim of
this study was to conduct a hot compression test on AISI 1016 carbon steel using
finite element modelling (FEM) in order to predict the flow stress behaviour of the
material including the damage prediction due to shear cracking. Three lubricating
conditions namely: Coulomb μ = 0.3(lubricated), shear μ = 0.3 lubricated and
dry μ = 0.7 were used in the modelling. The results indicate that the largest defor-
mation appears as a shear cross in form of an hourglass in the plastic zone of the
specimen at the centre of the cylinder and varies according to the lubricating condi-
tion, while the less deformed regions are stagnant as dead-metal zones (DMZ). The
highest damaging parameter, though moderate appears at the bulge of the cylinder
and this can be the source of shear cracking in the material. It was also observed
that there is inhomogeneous deformation inside the workpiece, which could result in
substantial damage to the forging process and forging quality respectively. Besides,
inhomogeneous deformation could also worsen the finish-forging process.
Keywords Material flow behaviour · Compression test · Flow stress · FEM ·

DMZ · Shear cross · Damage prediction
V. Musonda (B) · E. T. Akinlabi

V. Musonda
Department of Mechanical Engineering, School of Engineering, The University of Zambia, Great
East Road Campus, P.O. Box 32379, Lusaka, Zambia
E. T. Akinlabi

https://doi.org/10.1007/978-981-15-4488-0_24
264 V. Musonda and E. Akinlabi
1 Introduction
The compression test is an extensively employed method for attaining the flow curve
of metallic materials. Principally, the method is simple to use and that a cylindrical
specimen is commonly used and placed between reasonably rigid platens or dies in
order to provide a stress state of simple compression. This, however, is on condi-
tion that the movement is not constrained at the loaded interfaces [1]. From a metal
forming perspective, the description of strain hardening during plastic deformation
becomes realistic when flow curves are used. Moreover, the mechanical behaviour
of metals can be modelled and the constitutive equations of plasticity, which may
be nonlinear, can be set up. Flow curves obtained from a compression test are also
key in the estimation of the forces and pressures applied on the workpiece and the
dies. Furthermore, these curves can be used as a guide in controlling the process
operating conditions, which apparently is fundamental to the realisation of practi-
cal engineering solutions. The biggest challenge in the cylindrical compression test,
however, is how to minimise friction at the workpiece–die interface so that homoge-
nous deformation can be achieved. Homogeneous deformations in a compression test
are only possible under frictionless conditions against the compression platens. Such
conditions guarantee a flawless cylindrical shape during the compression test. This,
however, is on condition that the material of the cylinder is isotropic [2]. Numerous
reasons, nevertheless, have been reported [3] for not attaining homogeneous plas-
tic deformation conditions practically. First, the quality of lubricants used on the
workpiece-die interface can affect the result upon which the experimental procedure
depends. Moreover, the “barrelling” effect which is evident as a flaw will always
be present and hence it is not possible to guarantee frictionless conditions, even if
the most effective lubricants are used [3, 4]. This is also an indication that there is
residue friction related with the compression test, which is always present regardless
of the presence of the lubricant at the material-die interface.
Inhomogeneous deformation, nonetheless, is characterised by a bulging, or bar-
relling effect when friction is present at the interface [2]. During the sliding motion
of the cylinder in the radial and outward direction, the frictional stresses appear to
come against this motion over the die platens so that the friction shear stresses can
reduce the radial outward flow of the top and bottom faces of the cylinder next to the
die platens. There is, however, a need to compensate for this reduction of the radial
flow, and this is achieved by a corresponding increase in the metal flow outwards in
the middle of the cylinder. Therefore, the frictional effect between the die and the
tooling instantaneously transforms into a flow pattern called barrelling. Barrelling
appears on the outer circumferential surface at the midheight of the cylinder, and the
material develops a shape similar to a barrel. The bulging effect also increases with
an increased amount of friction and the shape becomes more pronounced throughout
the compression stroke. Conical metal-dead zones then appear as stagnant regions
inside and on both sides of the horizontal midplane of the cylinder, at each end of
the specimen [5]. The plastic deformation of the material inside the DMZ, is not as
much, compared with the rest of the material in the cylinder. This variation in the
Finite Element Modelling of a Compression Test … 265
Fig. 1 Conical metal-dead zones owing to the friction at the interface during the compression test.
a Has the larger length to diameter ratio, b smaller length to diameter ratio [5]
deformation patterns within the volume of the material is what creates the inhomo-
geneity in the cylinder compression with friction. Figure 1, shows conical metal-dead
zones formed owing to the friction at the interface during the compression test.
Friction plays a key role in metal forming process due to its influence on the energy
required for the compressive force. Over and above, friction affects the flow behaviour
in the material inside the die including the quality of the product and tool life [6].
Theoretical and numerical investigations have been used in the barrel compression
test analysis in order to evaluate the frictional effects on strain-hardening behaviour of
the material [7]. According to the theoretical analysis, the strain-hardening exponent
of the material affects the profiles of the barrel [5]. However, analytical evaluation of
the cylinder compression test is a very difficult process to comprehend. This has been
attributed to the complex nature of the stress–strain distribution and the tribological
conditions within the contact zone. Therefore, researchers have been compelled to
use Finite Element (FE) analysis as a pragmatic and effective tool to resolve the
problem [7].
During compressive deformation of a specimen, there is a change in the area of its
cross-section as the material spreads over the die. During this increase in the area, the
frictional stresses between the specimen and the die act in the opposite direction to
the flow of the material and hence providing a negative compressive force inwards.
A negative sign to denote the direction of frictional stresses is usually indicated in
the flow curves. Therefore, the regions of the specimen, which are in contact with
the die including the regions just near the die, are subjected to triaxial compressive
stresses [2, 5]. A cutting line is usually visible around these highly stressed regions
at the edge of the cylindrical specimen. On the other hand, the compressive forces
in the transverse direction, restrain the outer flow of the specimen material and these
forces are at maximum on the surfaces at the ends, which are in contact with the
dies. A gradual decrease by these forces is also noticed near the midlength of the
specimen [5]. The flow curve during the reduction of the specimen in length at any
given value is characterized by an upward shift. This is on condition that the ratio
of initial diameter D0c to initial length L 0c , i.e. D0c /L 0c of the specimen increases.
Some of the factors that influence the forming load are die geometry and the forming
speed especially in hot forming [2].
The Upper Bound Method (UBM) [8] has been reported to be useful in analysing
the pressure applied against the die and the load required during the forming process.
The method is based on the computation of the energy consumption in the process,
obtained from the flow field in the actual forming process although the velocity field
may be discontinuous on a finite number of imaginary internal surfaces [9]. The
UBM is known for providing quick and precise solution to many 2D or 3D forming
problems. Using UBM, the total power consumption for a particular metal forming
process can be calculated according to [2, 10, 11] such that:

ẆT = F · vs = ẆD + ẆS + ẆF = σ ε̇dV + k|v|dA + τi vi dA (1)
V A A
In Eq. (1), ẆT , is the total power, F, is the forming load and vs the velocity of
forming die or punch velocity, ẆD is the power required to deform the workpiece
homogeneously, ẆS is the power required to shear-deform the workpiece. Shear
deformation occurs if some or all of the workpiece material flows through a veloc-
ity discontinuity during the course of forming, ẆF is the power consumed due to
frictional sliding of workpiece material over the interface between the die and the
workpiece, σ and ε̇ are the equivalent stress and strain rate respectively, v is the
magnitude of the velocity discontinuity surfaces SD .
Applying the UBM to the axisymmetric cylinder, the compressive force during
the forming process when the friction factor m is applied, can be calculated from a
series of equations as articulated [2]:
vD
ẆD = ∫ σ ε̇dV = π R 2 h · σ · h
= π R 2 σ vD
V (2)
Ws = 0 (shows that the velocity discontinuity is not there)
R
mσ vsr mσ vs 2π
ẆF = 2 τi vi dA = 2 √ · · 2πr dr = 2 √ r 2 dr
3 2h 2 3·h
A A 0
2mσ vs π R 3
= √
3 3·h
Summation of the three parts: ẆD , Ẇs and ẆF leads to Eq. (3)
ẆT = F · vs = ẆD + ẆS + ẆF

2π m σ vs R 3
= σ ε̇dV + 0 + τi vi dA = π R 2 σ vs + √ (3)
3 3 h
V A
Hence, the compressive force is finally calculated as shown in Eq. (4)

ẆT 2μR
F= = πR σ 1+ √
2
(4)
vs 3 3·h
Evaluation of the material performance during the production process at an indus-

trial level for components such as turbine discs and turbine shafts require an under-
standing of the microstructure evolution during their forming process [12]. Such
components are considered as safety components; and therefore, recrystallization and
grain growth during the hot forging process of these components becomes important.
FEM is so far capable of analysing these events.
Aero-engine turbine discs, for example, are IN718 Nickel (Ni)-based superal-
loys, and obtaining a uniform and fine microstructure for these alloys is important
to ensure that the required mechanical properties are obtained for these safety com-
ponents. However, various microstructural defects such as coarse and duplex grain,
respectively, can occur during the forging of IN718 turbine discs [13]. Therefore,
obtaining the optimum parameters for designing the forging process becomes impor-
tant [13]. Hence, a justification to conduct a quality control test such as compression
can be made. Compression test studies have been reported [14–20] in which the
characteristics of deformation as well as the evolution of the microstructure mecha-
nism of superalloy IN718 have been described. These studies confirm some typical
industrial applications of the test as applied to aero-engine turbine discs and shafts
[13]. AISI 1016 steel was used in this study because it is one of the commonly used
material in upsetting operations. The compression test results of AISI 1016 shows
the deformation behaviour typical of an hourglass in “X” configuration at the centre
of the cylinder and dead-metal zones at the top and bottom of the cylinder. This defor-
mation pattern is similar to the one depicted in a 3D simulated model of a Forged
final shape of a turbine rotor shown in [12]. Therefore this test provided an insight
into what could be expected when other important components such as aero-engine
turbine discs or turbine shafts or rotors are to be manufactured.
2 Material Specifications and FEM Modelling
In this study, a hot compression test simulation was conducted on AISI 1016 carbon
steel and DEFORM-3D database was used to obtain the mechanical and thermo-
physical properties of this material. The 2D axisymmetric model was used in the
simulation and the deformation patterns were obtained in 2D and 3D views after the
post-processing. The ASTM medium length standards for the specimen as articulated
[21] was adopted in this study. This standard considers a cylindrical specimen used
for general purpose, to have the initial length L 0c to initial diameter D0c ratio to be
equal to 3, i.e. L 0c /D0c = 3. Table 1 shows the medium length cylindrical specimen
Table 1 Medium length

Initial length (L 0c ) Initial diameter (D0c ) L 0c /D0c
cylindrical specimen
(mm) (mm)
dimensions [21]
1 38.1 12.7 3
2 60.325 20.2692 2.976
3 76.2 25.4 3
4 85.725 28.575 3
dimensions. The initial length (L 0c ) and initial diameter (D0c ) used in the simulation
were 76.2 mm and 25. 4 mm respectively which gives a ratio of 3 as per the standard.
A cylindrical specimen was used in the simulation and compressed isothermally
between two flat polished dies (or platens). The initial dimensions of the specimen
were D0c = 25.4 mm and L 0c = 76.2 mm respectively. The compression platens
can be either tool steel, tungsten carbide, or ceramic composite, depending on the
temperature, and these should be flat and parallel according to ASTM E209 [22].
The maximum reduction in height was 75% at a constant speed of 2 mm/s and a
maximum displacement of 57.2 mm. This speed was related to the crosshead at
a constant temperature of 1200 °C. Tables 2 and 3 shows the type of lubricating
Table 2 Friction types and state variables during a hot compression test of AISI 1016 carbon steel
Friction Coefficient of Strain- Stress-effective Total XY
type friction (µ) rate (MPa) (step displacement strain-effective
effective 450) (mm) (step (step 450)
(s−1 ) 450)
(step 450)
Coulomb 0.3 0.786 62.9 57.2 2.36
(Lubricated)
Shear 0.3 0.718 61.9 57.2 1.92
(Lubricated)
Dry 0.7 0.928 64.6 57.2 2.48
Table 3 Friction types and state variables during a hot compression Test of AISI 1016 Carbon
Steel
Friction type Coefficient Damage Load (N) Force (N) Normal
of friction factor (Max) (step 450) (step 450) Pressure
(µ) (step 450) (Max.)
(MPa) (step
450)
Coulomb 0.3 0.333 1e+05 768 223
(Lubricated)
Shear 0.3 0.250 9.52e+04 923 253
(Lubricated)
Dry 0.7 0.342 1.05e+05 721 194
conditions used and the results of the state variables obtained after a post-processing
in DEFORM-3D. The final step 450 of the simulation results was used as the basis for
comparing the deformation patterns in the compressed material given the variations
in the three lubricating conditions at the interface.
3.1 Stress and Strain Rate Effective
The largest deformation patterns in all the friction types indicate that the highest
stress appears inside the cylinder at the centre, and these patterns are characterised
by a typical shear band. The degree of deformation, however, depends on the value
of μ. Figure 2a, b shows the three-dimensional FEM-predicted distribution of the
maximum stresses and deformation bands when the coulomb and shear conditions
are used. Figure 2c, d shows the respective 2D stress and strain rate patterns for the
shear condition. Figure 2e shows the highest stress at 64.6 MPa, which occurred
when a dry interface (μ = 0.7) was used. In all the lubricating conditions used in the
study, the deformation pattern showed a shear cross or hourglass with less deformed
stagnant regions appearing at the top and bottom cylinder as dead-metal zones. The
bulging or “barreling” effect appeared in the middle of the cylinder with moderate
deformation.
A visible cutting line (Fig. 2a) at the top and around the cylinder shows the
effect of high tensile stresses acting in the direction around the circumference of
the cylinder. Inhomogeneous deformation was largest in dry friction conditions and
this was characterised by jerky or serrated yielding as shown on the XY stress graph
in Fig. 2e. The XY stress was 30.1 MPa at maximum stress-effective of 64.6 MPa.
Figure 3a, b shows the FEM-predicted distribution in 2D axisymmetric for stress-
effective and strain rate-effective when μ = 0.3 (coulomb lubricated).
3.2 Strain Rate–Effective and Strain Distribution
The distribution of strain rate and strain was homogenoues during the initial stages of
compression in shear and dry conditions and this was characterised by an insignif-
icant effect of barelling in the middle of the cylinder. The distribution, however,
was inhomogenous towards the end of the compression test and the barelling effect
was significant though moderate. During the last stage, the deformation pattern trans-
formed into a shear band in a clearly visible “X” configuration and the corresponding
XY graph was typical of a jerky or serrated yielding, indicating that, the deformation
was not homogenous. Figure 4a shows the FEM-predicted distribution of effective
strain rate with a jerky flow in shear (μ = 0.3) beginning after 18 s of compression.
Cutting line
Dead -Metal Zone
Barrelling Shear cross
(a) (b)
(c) (d)
(e)
Fig. 2 FEM-predicted distribution of maximum stresses and deformation bands in hot compression
of AISI 1016 cylinder a 62.9 MPa when Coulomb μ = 0.3 lubricated, b 61.9 MPa when Shear
μ = 0.3 lubricated, c, d 2D axisymmetric patterns, e Jerky flow in XY stress (μ = 0.7)
Fig. 3 FEM-predicted distribution in 2D axisymmetric. a Stress-effective (62.9 MPa), b strain

rate-effective (0.786 s−1 ) when μ = 0.3 (coulomb lubricated)
This behaviour continued until the last stage of compression when strain rate-effective
was 0.718 s−1 . Figure 4b shows FEM-predicted distribution of effective strain rate
with an uneven flow curve at the beginning of compression followed by a pronounced
jerky flow occurring between 18 and 24 s of the compression stage when μ = 0.7.
After 24 s of compression, a near steady flow curve was noticed until the last stage
of compression at 0.928 s−1 . This behaviour is different from that depicted in a shear
flow curve where the curve was steady during the initial stages of compression until
after a jerky flow began after 18 s of compression.
The frictional resistance values (red curve on the graph) in the two graphs alluded
to are very close to the true values (green curve on the graph) with the absolute error
being 0.016 in shear and 0.014 for dry condition. This translates into 4 and 1.6%
errors respectively. The frictional resistance values were assumed to be theoretical
in all the cases and they were very close to the true values. The strain rate, however,
was not uniform in a dry condition (μ = 0.7) as can be seen from the irregular flow
curve (Fig. 4b) during the initial 75% maximum reduction. This was followed by a
jerky flow curve before a near steady flow was assumed until the final compression
stage.
The FEM-predicted 2D axisymmetric distribution patterns for stress and strain
rate for dry condition presented in Fig. 5a, b showed the highest stress and strain
rate-effective values of 64.6 MPa and 0.928 s−1 , respectively. Figure 6, shows the
FEM prediction for the largest deformation in “X” configuration at the center of the
cylinder when strain-effective is 2.48 in dry condition. It should be noted that the
frictional resistance values (in red) and the true values (in green) in the XY strain–
effective graph (Fig. 6) are very close to each other despite some minor deviations
in the flow curve. The flow curve also shows a slight deviation from the typical load-
stroke compression curve. It can, therefore, be deduced that a dry workpiece-die
interface promotes inhomogeneous deformation in a hot compressed workpiece.
Beginning of
Jerky flow
(a) shear
Jerky flow
(b)
Fig. 4 FEM-predicted distribution of effective strain rate in the plastic zone of the specimen at the
centre in hot compression of AISI 1016 cylinder. a XY strain rate at 0.718 s−1 in shear, b XY strain
rate at 0.928 s−1
3.3 Damage Parameter During Loading
The most probable site for the crack initiation and consequently fracture in a ductile
material during a hot compression test could occur in the bulged portion of a cylin-
der. Figure 7 confirms that the highest damage parameter can be predicted to occur
in the middle of the bulge. The high tensile stresses acting in the direction around
(a) (b)
Fig. 5 FEM-predicted 2D axisymmetric distribution patterns for stress and strain rate. a Stress-
effective(64.6 MPa), b strain rate-effective (0.928 s−1 ) when μ = 0.7 (dry)
Fig. 6 FEM-predicted distribution of strain-effective in the plastic zone with the largest deformation
appearing in “X” configuration at the center
the circumference of the cylinder are responsible for this damage. Strain softening
can affect the rate of decrease in the strength of the material, and when this amount
exceeds the permissible proportion of the increase in the area of the specimen, a
“kink” develops in the specimen to indicate the unstable mode of deformation [23].
This mode of deformation was predicted by FEM in the damage parameter evaluation
for the dry condition. However, the damage parameters in all the lubricating condi-
tions in this study exhibited a steady exponential curve except for the dry condition,
which showed a “kink” (in black dashed circle) in the early stages of compression.
After this point, the curve was steady until the last stage. The highest damaging factor
Kink
Fig. 7 FEM-predicted damage in a hot compression of AISI 1016 steel cylinder with maximum
damage factor of 0.342 occurring in the middle of the bulge when μ = 0.7, with a kink in the flow
curve
occurred in dry condition with a maximum value of 0.342. Figure 8, also shows the
load-stroke prediction curve with the maximum load of 1.05e+05 N at Strain-Total-
Von misses of 2.48 maximum for dry condition (μ = 0.7). This value corresponds
to the highest deformation pattern inside the cylinder at the center.
4 Concluding Remarks
In this study, it has been shown that the deformation conditions in a hot cylindrical
compression test can be characterized by FEM. The input to the FEM simulation was
the workpiece material AISI 1016 carbon steel and the coefficient of friction (COF)
data for coulomb μ = 0.3(lubricated), shear μ = 0.3(lubricated) and μ = 0.7(dry)
conditions were used. These conditions were used to compare the deformation pat-
terns of homogenous and inhomogenous conditions. High stress values were noted in
dry and coulomb conditions with dry having the highest at 64.6 MPa while coulomb
had 62.9 MPa and shear 61.9 MPa. In all the conditions, the damage factor was
characterised by an exponential curve except for the dry condition, which showed a
“kink” during the initial stages of the compression. The least damage factor of 0.250
occurred under shear conditions. The 3D deformation patterns at the centre and
inside of the cylinder in all the conditions were characterised by a typical hourglass
or shear cross in the plastic zone of the specimen. Areas of less deformed material
appeared as stagnant zones popularly known as dead-metal zones. The maximum
Fig. 8 FEM-load-stroke prediction and largest deformation in a hot compressed steel with strain-
total Von Misses of 2.48 when μ = 0.7 at maximum load of 1.05e+05 N
strain occurs at the center inside the cylinder and on the edge of the cylinder. It has
also been demonstrated that homogenous deformation is not guaranteed with par-
ticular use of lubricant at the interface. However, there is confirmation in this study,
that dry conditions lead to non-uniformity of strain rate which results in inhomoge-
neous deformation patterns. It can, therefore, be deduced that barrelling makes the
determination of the true axial compression stress complicated when the compres-
sion test is used to measure the flow properties of the material. The remedy to the
“barrelling” effect is to evaluate the properties of the material, die geometry and the
kind of lubricant used in the compression test if homogenous deformation has to be
achieved.
References
1. Williams JG, Chaiwut G (2008) Using the simple compression test to determine Young’s
modulus, Poisson’s ratio and the Coulomb friction coefficient. Int J Solids Struct 45:4448–4459
2. Valberg HS (2010) Applied metal forming. Including FEM analysis. Cambridge University
Press
3. Christiansen P, Martins PAF, Bay NO (2016) Friction compensation in the upsetting of cylin-
drical test specimens. Exp Mech 56(7):1271–1279. https://doi.org/10.1007/s11340-016-0164-
z(2016)
4. Shahriari D, Sadeghi MH, Ebrahimi GR, Kim KT (2011) Effects of lubricant and temperature on
friction coefficient during hot forging of Nimonic 115 superalloy. Kovove Mater 49:375–383.
https://doi.org/10.4149/km_2011_5_375
5. Bhaduri A (2018) Mechanical properties and working of metals and alloys. Springer Series in
Materials science, vol 264. https://doi.org/10.1007/978-981-10-7209-3_2
6. Erinosho MF, Akinlabi ET (2016) Study of friction during forging operation. Proceedings of
the World Congress on Engineering, vol II, WCE, 29 June–1 July, London, UK
7. Yao Z, Mei D, Shen H, Chen Z (2013) A friction evaluation method based on barrel compression
test. Tribol Lett 51:525–535. https://doi.org/10.1007/s11249-013-0164-4(2013)
8. Prager W, Hodge PG (1941) Theory of perfectly plastic solids. Wiley, New York
9. Wu WT, Jinn JT, Fischer CE (2003) Modeling techniques in forming processes (Chap. 15).
In: Dieter GE, Kuhn HA, Semiatin SL (eds) Handbook of workability and process design, pp
220–231. https://doi.org/10.1361/hwpd2003p220
10. Altan T, Oh S-I, Gegel HL (1983) Metal forming: fundamentals and applications. ASM Int
131–142
11. Avitzur B (1968) Metal forming: processes and analysis. McGraw-Hill, New York
12. Li X, Kopp R, Wolske M (2005) Development of forgeable Ni-base alloys for USC steam
turbine applications by use of microstructure simulation and formability testing. In: Loria EA
(eds) Superalloys 718, 625, 706 and derivatives, TMS (The Minerals, Metals & Materials
Society)
13. Zhang HY, Zhang SH, Li ZX, Cheng M (2009) Hot die forging process optimization of super-
alloyIN718 turbine disc using processing map and finite element method. In: Proceedings of
IMechE, vol 224: Part B: J Eng Manuf. https://doi.org/10.1243/09544054jem1571 (2009)
14. Srinivasan N, Prasad YVRK (1994) Microstructural control in hot working of IN-718 superalloy
using processing map. Metall Mater Trans A 25:2275–2284
15. Medeiros SC, Prasad YVRK, Frazier WG, Srinivasan R (2000) Microstructural modelling of
metadynamic recrystallization in hot working of IN718 super-alloy. Mater Sci Eng A 293:198–
207
16. Zhou LX, Baker TN (1994) Effects of strain rate and temperature on deformation behaviour
of IN718 during high temperature deformation. Mater Sci Eng A 177:1–9
17. Zhou LX, Baker TN (1995) Effects of dynamic and metadynamic recrystallization on
microstructures of wrought IN-718 due to hot deformation. Mater Sci Eng A 196:89–95
18. Thomas A, El-Wahabi M, Cabrera JM, Prado JM (2006) High temperature deformation of
Inconel 718. J Mater Process Technol 177:469–472
19. Yuan H, Liu WC (2005) Effect of the delta phase on the hot deformation behaviour of Inconel
718. Mater Sci Eng A 408:281–289
20. Wang Y, Shao WZ, Zhen L, Yang L, Zhang XM (2008) Flow behaviour and microstructures
of superalloy718 during high temperature deformation. Mater Sci Eng A 497:479–486
21. Davis HE, Troxell GE, Wiskocil CT (1964) The testing and inspection of engineering materials,
3rd edn. McGraw-Hill Book Company, New York
22. ASTM E 209-65 (1993) Standard practice for the compression tests of metallic materials at
elevated temperature. ASTM International
23. Dieter GE (2003) Hot compression test (Chap. 6). In: Dieter GE, Kuhn HA, Semiatin
SL (eds) Handbook of workability and process design, pp 61–67. https://doi.org/10.1361/
hwpd2003p061, (2003)
Evaluation of Tie Wing Deformation
in 0.022 Inch Stainless Steel Orthodontic
Bracket—A Finite Element Analysis
Akhil Minu Ajayan, V. Magesh, P. Harikrishnan, and D. Kingsly Jeba Singh
Abstract Orthodontics is the field of dentistry, which deals with the misalignment
(malocclusion) of teeth and its treatment for correcting the improper bite in the jaw.
Orthodontic brackets are a component of fixed appliances that are used in orthodon-
tics that aligns and rectifies the misalignment. Archwires are used to control the
movement of teeth by giving load on the bracket, which transfers the force to the teeth.
Since the nineteenth century, with the development of orthodontics, various models
of brackets that varies in shape and size were invented. The response behaviour and
the effect on the bracket for the load applied by the archwire varies for different
shape and slot size. As a result, the deformation and stress evolved on the bracket
varies. Thus, our study aims to find the tie wings deformation in a conventional
orthodontic bracket. Archwire is made to rotate with different angles inside the slot
of the bracket. In our methodology, the bracket and archwire assembly were mod-
elled, meshed and finite element analysis (FEA) was done on the brackets to evaluate
the tie wings deformation for the various angle of twist on the archwire as done in a
clinical situation.
Keywords Tie wing deformation · Conventional bracket · Angle of twist · Finite

element analysis
A. M. Ajayan (B) · V. Magesh

Department of Mechanical Engineering, SRM Institute of Science and Technology,
Kattankulathur, Kancheepuram, Tamil Nadu, India
V. Magesh
P. Harikrishnan
Teeth “N” Jaws Centre, Lake Area, Nungambakkam, Chennai, Tamil Nadu, India
D. Kingsly Jeba Singh
School of Mechanical Engineering, SRM Institute of Science and Technology, Kattankulathur,
Kancheepuram, Tamil Nadu, India

https://doi.org/10.1007/978-981-15-4488-0_25
278 A. M. Ajayan et al.
1 Introduction
For correcting the malocclusion or improper alignment in teeth, orthodontic brackets

are used. The improper alignments include under bite, overbite, crooked and crowded
teeth. A force is transferred from the archwire to the teeth; through the brackets used
in the fixed appliance therapy. When there is force transfer mechanics, possibilities of
bracket and wire deformation are more. By applying a force in the bracket slot, force
moments are received during the treatment of malocclusion [1]. The stability of a
tooth root in a location is related to the rectangular archwire and bracket relationship.
The variables like interincisal angle, overjet and optimal teeth positioning are rectified
using incisor torque. Torque expression applied by the archwire is affected by torsion
magnitude, wire dimension, play between wire and bracket, bracket deformation
and length of the bracket and wire [2–4]. Though experimental studies used for
finding bracket deformation is built upon complicated design principles, it fails to
graphically display the deformation and stress distribution in both bracket and wire
[5–7]. Clinically, the archwire is twisted in various angles called the angle of twist
for the correction of malocclusion in teeth.
The Finite element analysis (FEA) is one of the appropriate tools for finding
the stress distribution and deformation in orthodontic applications [8, 9]. Few FEA
findings were done, showing the slot wall deformation of the orthodontic brackets
[10, 11]. Studies on tie wing deformation using FEA in a bracket was carried out
by applying a couple on the bracket slot [12]. Our hypothesis in this study includes
rotating the archwire in the bracket with various angles of twist in the bracket. Thus,
we evaluated the tie wing deformation by twisting the archwire in various angles in
0.022-inch stainless steel (SS) bracket.
Standard Edgewise maxillary right central incisor SS Bracket with 0.022 inch ×
0.028 inch slot (Leone, Italy) and a rectangular SS archwire size of 0.019 inch ×
0.025 inch (G&H Orthodontics, USA).
The bracket dimensions were measured using an Optive LITE OLM vision mea-
suring system (Hexagon Manufacturing Intelligence, Great Britain). The 2D profile
of the orthodontic bracket was exported as an output file from the machine. This 2D
profile was converted into a 3D model using AutoCAD (Autodesk, USA), and was
exported to Solidworks (Dassault Systemes, France) for assembling the archwire
with the orthodontic bracket. Finite element (FE) model for the assembly was gen-
erated using Hypermesh (Altair Engineering, USA) as pre-processor. The assembly
model was meshed using hexahedral and pentahedral elements for maintaining the
mesh flow as shown in Fig. 1. The meshed model was analysed using Ansys Work-
bench (Ansys, Inc, USA) as the processor. Linear, elastic and isotropic materials
were used (Table 1).
Evaluation of Tie Wing Deformation in 0.022 Inch … 279
Fig. 1 Finite element model of bracket and wire assembly

Property Value
of stainless steel
Density 7.750 kg m−3
Young’s modulus 1.93 × 105 MPa
Bulk modulus 1.693 × 105 MPa
Shear modulus 0.7355 × 105 MPa
Yield strength 207 MPa
Ultimate Strength 586 MPa
3 Finite Element Analysis of Bracket
The archwire which is placed inside the slot is twisted in different angles for simu-
lating the palatal root torque condition. The angle of twist varied from 5° to 25° with
an interval of 5°. The bracket base was fixed to arrest the translation in x, y and z
axis. The deformed bracket and archwire are shown in Fig. 3 with an angle of twist
of 25°. In the FEA, the nodal deformation on the top surface of the bracket for the
four wings was measured for evaluating the tie wing deformation of the bracket wire
assembly as shown in Fig. 2.
The archwire is made to rotate in the clockwise direction known to be as Labial
crown torque by giving angle of twist on the wire. The wings on the right of the
bracket are known to be Gingival Wings and the wings on the left to be occlusal
wings.
Fig. 2 Nodal points for tie wing deformation (OW—Occlusal Wing, GW—Ginigival Wing)
Fig. 3 Deformation of bracket and wire for an angle of twist with 25°
The FEA of orthodontic bracket with archwire was done for evaluating the tie wing
deformation with an angle of twist varying from 5° to 25° on the archwire. The
archwire was twisted at both ends with similar to the clinical conditions. The analysis
results are shown below.
Figure 4a, b depict the orthodontic bracket tie wing deformation by applying an
angle of twist on the archwire. For the applied angle of twist, bracket wire assembly
showcased varying tie wing deformation for different angles of twist and which
varied between the four wings.
From Tables 2 and 3 and Figs. 5 and 6, it is proved that, when palatal root torque
is applied on the wire, Gingival wing undergoes more deformation than the occlusal
wing in both x- and y-axis, and the outer edges of the wing undergoes more deforma-
tion than the inner edges as the wire is twisted at its ends. As the wire is twisted only
at the ends, the load acting on wire throughout its length also varies. Therefore, the
Fig. 4 a Bracket with 0° angle of twist on the archwire. b Bracket with 25° angle of twist on the
archwire
Table 2 Nodal deformations in gingival and occlusal tie wings in x-axis

Angle of twist Tie wing deformation in x-axis (µm)
(degree) GW1a GW1b GW2a GW2b OW1a OW1b OW2a OW2b
5 0 0 0 0 0 0 0 0
10 4.412 2.268 2.271 4.406 1.043 0.432 0.432 1.042
15 13.191 6.854 6.855 13.206 3.346 1.396 1.396 3.345
20 22.09 11.493 11.502 22.095 5.973 2.508 2.508 5.981
25 31.272 16.105 16.157 31.115 9.003 3.688 3.739 9.054
Table 3 Nodal deformations in gingival and occlusal tie wings in y-axis

Angle of twist Tie wing deformation in y-axis (µm)
5 0 0 0 0 0 0 0 0
10 0.995 0.685 0.686 0.994 0.129 0.156 0.157 0.129
15 2.989 2.104 2.097 3.007 0.431 0.507 0.508 0.428
20 4.965 3.482 3.512 4.907 0.774 0.883 0.893 0.778
25 7.009 5.031 4.992 7.024 1.178 1.333 1.343 1.175
Tie wing deforma on of Wing (μm) (x-axis)

35 GW1a
30
Deforma on (μm)
GW1b
25 GW2a
20
GW2b
15
0W1a
10
OW1b
5
0 OW2a
5 10 15 20 25
OW2b
Angle of Twist (degree)
Fig. 5 Graphical representation for nodal deformations in gingival and occlusal tie wings in x-axis
deformation of the inner edges of wing is comparatively lower than the outer edges.
The deformation in occlusal wing is found to be 25% of gingival wing deformation.
From the analysis, it is observed that the angular deformation also occurs on the
tie wings when the archwire is twisted. Table 4 and Fig. 7 presents the tie wing
angular deformation in gingival and occlusal wings. Thus, the angular deformation
of tie wing is more in the location where bracket and wire are in contact.
Figure 8 shows the stress distribution in the bracket for the angle of twist of
25°. The point on which the bracket and the wire comes in contact with each other,
Tie wing deformation of Wing (μm) (y-axis)

8 GW1a
7
Deformation (μm)
GW1b
6
5 GW2a
4 GW2b
3
0W1a
2
1 OW1b
0 OW2a
5 10 15 20 25
OW2b
Angle of Twist (degree)
Fig. 6 Graphical representation for nodal deformations in gingival and occlusal tie wings in y-axis
Table 4 Tie wing angular deformation in gingival and occlusal wings

Angle of twist Tie wing angular deformation (degree)
5 0 0 0 0 0 0 0 0
10 12.92 17.31 17.31 12.92 7.09 20.69 20.82 7.09
15 12.98 17.59 17.53 13.05 7.38 20.81 20.85 7.33
20 12.88 17.36 17.49 12.72 7.42 20.17 20.40 7.45
25 12.63 17.34 17.17 12.72 7.45 19.87 19.76 7.39
Tie wing angular deformation (degree)

25 GW1a
Deformation (degree)
GW1b
20
GW2a
15
GW2b
10
OW1a
5 OW1b
0 OW2a
5 10 15 20 25
Angle of twist (degree) OW2b
Fig. 7 Graphical representation for tie wing angular deformation
experiences maximum stress. Also, the point at which the stress concentration occurs
is where the maximum load acts.
Figure 9 illustrates that the stress distribution in the wire throughout its length is
not uniform. The point where bracket and wire are in contact undergoes maximum
Fig. 8 Stress distribution in the bracket for 25°
Fig. 9 Stress distribution in

the wire for 25°
stress, while the middle portion of the wire undergoes minimum stress. As both the
ends of the wire are twisted, they undergo an average stress.
5 Clinical Significance
Clinically, various angles of twist are applied for finer tooth movements. In this study,
for the applied palatal root torque more deformation was recorded in the gingival
tie wings than the occlusal tie wings. These deformations might indirectly indicate
not only changes in the bracket tie wings but also in the bracket slot as well. Hence,
clinicians should be aware of these bracket changes which will affect the final tooth
positioning.
6 Conclusion
This Insilico study evaluated the tie wing deformation in 0.022-inch. SS orthodontic
bracket by varying the angle of twist on the archwire. It is concluded that the tie wing
deformation varies for the applied angle of twist by the archwire on the bracket. Thus,
tie wing deformations on bracket have greater clinical significance with the angle of
twist on the archwire.
Funding This work was funded by the SRM Institute of Science and Technology under selective
excellence initiative program.
References
1. Wagner JA, Nikolai RJ (1985) Stiffness of incisor segments of edgewise arches in torsion and
bending. Angle Orthod 55(1):37–50
2. Odegaard J, Meling E, Meling T.: An evaluation of the torsional moments developed in
orthodontic applications, an in vitro study. Am J Orthod Dentofac Orthoped 105(4):392–400
3. Siatkowski RE (1999) Loss of anterior torque control due to variations in bracket slot and
archwire dimensions. J Clin Orthod 33(9):508–510
4. Fischer H, Orthuber W, Es-Souni M, Meyer S (2000) Torque transmission between square wire
and bracket as a function of measurement, form and hardness parameters. J Orofac Orthoped
61(4):258–265
5. Lacoursiere RA, Nobes DS, Homeniuk DL, Carey JP, Badawi HH, Major PW (2010) Mea-
surement of orthodontic bracket tie wing elastic and plastic deformation by arch wire torque
expression utilizing an optical image correlation technique. J Dent Biomech
6. Major TW, Carey JP, Nobes DS, Heo G, Major PW (2011) Measurement of plastic and elas-
tic deformation due to third-order torque in self-ligated orthodontic brackets. Am J Orthod
Dentofac Orthop 140(3):326–39
7. Melenka GW, Nobes DS, Major PW, Carey JP (2013) Three-dimensional deformation of
orthodontic brackets. J Dent Biomech 4(1758736013492529)
8. Wakabayashi N, Ona M, Suzuki T, Igarashi Y (2008) Nonlinear finite element analyses:
advances and challenges in dental applications. J Dent 36(7):463–471
9. Papageorgiou SN, Keilig L, Hasan I, Jäger A, Bourauel C (2016) Effect of material variation
on the biomechanical behaviour of orthodontic fixed appliances: a finite element analysis. Eur
J Orthod 38(3):300–7
10. Harikrishnan P, Magesh V, Kingsly J, Singh D (2017) Comparison of slot deformation in

stainless steel and ceramic brackets during torque—a finite element analysis. South Eur J
Orthod Dentofac Res. https://doi.org/10.5937/sejodr4-1284, May 2017
11. Magesh V, Harikrishnan P, Kingsly J, Singh D (2018) Finite element analysis of slot wall
deformation in stainless steel and titanium orthodontic brackets during simulated palatal root
torque. Am J Orthod Dentofac Orthoped 153(4)
12. Magesh V, Harikrishnan P, Sekhar TC (2018) Finite element analysis of orthodontic bracket
tie wing deformation due to labial crown torque. IOP Conf Ser: Mater Sci Eng 402(012085)
A Finite Element Analysis to Study
the Effect of Various Loading Conditions
on the Intervertebral Disc in L4–L5
Section of Lumbar Spine
J. Daniel Glad Stephen, M. Prakash, V. K. Nevedha, and Manu Pandey
Abstract Lower back pain has been one among the difficulties faced in general,
occurring in prevalent due to natural intervertebral disc failures in the lumbar spine
region. Studies are required to examine natural disc failures due to day to day activ-
ities. A model of human lumbar spine section L4-L5 was generated to check for
stability of natural disc under various loading conditions using finite element analy-
sis. Computer tomography scan images were compiled together to generate a three-
dimensional model of lumbar spine section L4-L5 by segmentation technique. Fur-
ther, the model was prepared along with the generation of cortical bone, annulus
pulposus, nucleus pulpous, vertebral endplates and corresponding ligaments. The
material properties for each of the components were incorporated into the model
from the literature available and meshed for analysis. The L4–L5 lumbar spine section
model created, upon being checked for different motions of spine showed satisfy-
ing results yielding normal human motion preservation. Natural disc deformation
parameters were observed under different loading conditions for daily activities.
Keywords Lumbar spine · Intervertebral disc failure · Finite element analysis
1 Introduction
There are basically four regions of the spinal column—cervical (C1–C7), thoracic
(T1–T12), lumbar (L1–L5), sacral (S1–S5). Since the lumbar vertebrae are the
strongest and largest among all, they are optimized for structural support rather
than flexibility. There are intervertebral discs present in between the vertebral bodies
that provide a cushion to distribute and transmit the forces.
These discs are responsible for linking the surfaces of the vertebrae and is sand-
wiched between upper and lower endplates. Intervertebral discs provide the required
flexibility in the spine and are a medium to transmit and distribute loads.
J. Daniel Glad Stephen · M. Prakash (B) · V. K. Nevedha · M. Pandey

Kattankulathur 603203, India

https://doi.org/10.1007/978-981-15-4488-0_26
288 J. Daniel Glad Stephen et al.
Natural discs consist of gelatinous nucleus pulposus center surrounded by annu-

lus fibrosus structure (Fig. 2). The highly layered annulus is well oriented by
fibrous collagen in the proteoglycan matrix. Nucleus comprises mainly of collagen,
proteoglycan, and water [1].
There are various intervertebral disc diseases which include degenerative disc,
bulging disc, herniated disc, thinning disc, and disc degeneration with osteophyte
formation. Degeneration, basically the deterioration of the tissue, affects the material
property of the tissues in the disc and thereby changing the behavior and function of
the disc. The major cause of lower back pain and disability observed in the adults is
due to the degenerative disc disease in the lumbar region of the spine.
With intervertebral disc failures, arises normal motion discomfort and difficulties
in pursuing daily activities. Such cases can be cured by medication as well as surgical
procedures consisting of either spinal fusion or total disc replacement using artificial
implants. Pain relief occurs by stopping the motion of the painful disc by the direct
bone connection between the vertebrae surrounding the painful disc in the spinal
fusion procedure. Artificial disc replacement aims at removing the worn or damaged
disc material and replacing it with a prosthetic implant [1].
The implant loads for different activities show a large inter- and intra-individual
variation depending on the way the activities were performed [2]. To start with the
study, the primary need is to understand the various activities performed by the
body, the physiologic range and pattern of motion, and the method to be undertaken
to perform the study.
In an earlier study, it is found out that most of the individuals do not use the
full range of motion (ROM) and employ only a small part of their full active ROM
when performing the daily activities [3]. Xia et al. [4] have observed that each of the
vertebral levels reacts differently to external loads applied at different vertebral levels.
A large variation in the pressure at the L4–L5 level was observed by Dreischarf et al.
for different body positions [5]. The technique for the determination of the various
loads acting on the lumbar spine is found to be well correlated with body weight [6].
The loading within the spine nearly shows a linear relationship with the body weight
and height though the height does not have much influence [7].
The body posture and activities that necessitate flexion in the spine results in higher
forces when compared to the spine in a neutral position or extension indicating that
both influence the loads in lumbar spine [8]. Higher compressive load is experienced
by the intervertebral disc in an upright position [9]. Miller et al. [10] have observed
that the lumbar motion segments could resist considerably larger compression loads
in bending without failure. Studies have indicated that the application of follower
load can deliver realistic results for motion simulation [11].
There are generally two methods to conduct the biomechanical study: finite ele-
ment analysis (FEA) and in vitro cadaveric test. Finite element analysis involves
creating a computational model and is used for the simulation of biomechanical
behavior of the spinal segment, whereas in vitro cadaveric tests are generally used to
observe the normal motion responses with varying loading conditions in the cadav-
eric specimen. Finite element (FE) is a very efficient and powerful technology in the
biomechanics of the spine segment in the lumbar region. A 3-D geometric model was
A Finite Element Analysis to Study the Effect … 289
created which includes the anatomical structure based on the Computer Tomography
(CT) and Magnetic Resonance Imaging (MRI) data [12].
The present FEM method applies segment-wise compressive preload and the
modifications can easily be made to apply the varying compression loads at different
levels of the lumbar spine [13]. Once the finite element model (FEM) is validated, the
further evaluation of range and pattern of motion of the L4–L5 spine segment under
different loading conditions during daily activities can be done. The main focus lies
under evaluating the condition of intervertebral disc subjected to various loading
conditions under different daily task categories.
CT scan images of the vertebral section were obtained and compiled together to
generate a three-dimensional model of lumbar spine section L4–L5 by segmentation
technique. The surfaces were smoothened further by editing the three-dimensional
data. A finite element L4–L5 lumbar spine section was modeled comprising of 95,457
solid elements and nine cable elements (Fig. 1a).
Structural eight-noded solid elements are used for cancellous bone, cortical bone,
vertebral endplates, annulus pulposus and nucleus pulposus (Fig. 2), and the isotropic
linear material properties such as Young’s modulus and Poisson ratio values are
chosen from literature [14]. During the application of compression load, annulus has
been found to experience large tensile strains in a radial direction with time [15]. So
Fig. 1 a FE model of the L4–L5 lumbar spine section. b Enlarged view of L4–L5 intervertebral
disc showing anterior and posterior positions
Fig. 2 Sectional cut view of L4–L5 finite element model created detailing various components
that annulus pulposus was thus assumed to have non-linear, hyperelastic behavior in
this study by using the constant as per the Ref. [16].
Ligaments were integrated into the L4–L5 section created and represented by two-
noded, three-dimensional spar elements having three degrees of freedom at each
node. Linear isotropic behavior was incorporated into all seven ligaments of the
lumbar spinal unit (anterior longitudinal ligament, posterior longitudinal ligament,
ligament flavum, facet capsulary ligament, intertransverse ligament, interspinous
ligament, and supraspinous ligament) created. Properties comprising of Young’s
modulus and cross-sectional area values were specified for all ligaments [14]. The
L4–L5 curved facet joint was created as surface contact elements. The material
characteristics of each of the components used in FE model have been mentioned in
Tables 1 and 2.

Components Young’s Poisson’s References
of different components
modulus ratio
present in the L4–L5 FE
(MPa)
model
Cancellous 100 0.2 [14]
bone
Cortical bone 12,000 0.3 [14]
Endplate 24 0.4 [14]
Nucleus 1 0.499 [14]
Annulus (Hyperelastic, Mooney-Rivlin) [16]
c1 = 0.42, c2 = 0.105
Table 2 Material properties of various ligaments in the lumbar spine

Components Young’s modulus (MPa) Cross-section (mm2 ) References
Anterior longitudinal 7.8 63.7 [14]
ligament
Posterior longitudinal 1 20 [14]
ligament
Ligamentumflavum 1.5 40 [14]
Facet capsulary ligament 7.5 30 [14]
Intertransverse ligament 10 1.8 [14]
Interspinous ligament 1 40 [14]
Supraspinous ligament 3 30 [14]
The inferior surface of L5 vertebra was rigidly constrained for all degrees of free-
dom. Moments of 10 Nm were applied on the superior surface of L4 vertebra to
measure an angular range of motion for all spine motions (flexion, extension, left
lateral bending, right lateral bending, left axial rotation and right axial rotation).
Further, to examine the deformation pattern of intervertebral discs subject to vari-
ous loading conditions during daily activities, compression load was applied on the
superior surface of the L4 vertebra. The load administered for various daily activities
was obtained from the literature.
For the L4–L5 lumbar section, loading condition of 10Nm was utilized as provided
by the study of Yamamoto et al. [17] and angular range of motion for each of the
spinal motions (flexion, extension, lateral bending) was observed as shown in Fig. 3.
Each motion of spine has been examined and compared to the in vitro experimental
Angular range of motion
8.9
(in degrees)
7.56
5.8 6.29 5.9
5.5 5.238 5.388
FLEXION EXTENSION LEFT LATERAL RIGHT LATERAL

BENDING BENDING
Motions of spine
Yamamoto et al. Present study
Fig. 3 Comparison of ROM data between in vitro experimental data [17] and the present study
Table 3 ROM data of the FE

Motions of spine Yamamoto et al. Present study
model compared to
Yamamoto et al. [17] Flexion 8.9 ± 0.7 7.6
Extension 5.8 ± 0.4 6.3
Left lateral bending 5.5 ± 0.5 5.2
Right lateral bending 5.9 ± 0.5 5.4
data obtained by Yamamoto et al. [17]. The validation of axial rotation has been
excluded since further focus lies distinctively on studying the natural disc deforma-
tion pattern under various loading conditions. The values of ROM under the same
loading condition as acquired in this study are listed down as in Table 3.
In order to examine the condition of intervertebral disc during daily activities
carried out, a further study was undertaken to make use of the validated finite element
model generated. Activities such as standing, walking, twisting, jumping, coughing
and laughing were chosen for analysis. Loading details for various activities have
been acquired from the study [18]. A compression load was applied under each
case on the superior surface of the L4 vertebra in a vertically downward direction
distributed uniformly over its surface. Disc deformation and Von Mises stress values
were procured to understand their correlation with varying loads. The acquired results
have been mentioned in Fig. 5a–l.
In Fig. 4, an increase in the amount of compression load shows a consistent
increase in the amount of disc deformation. The least value of displacement has been
accounted for standing (Fig. 5a) whereas the highest value for laughing condition
(Fig. 5f) in this study. The displacement values vary from the least value of 2.4 mm
to the highest of nearly twice this amount. The deformation plot of the natural disc
suggests that a greater part of displacement can be sighted by the anterior region of
the disc under the compression load. Although the deformation values are higher in
cases of coughing, laughing and jumping (Fig. 5c–f) only a less portion of anterior
region experiences such change along with an expeditious decrease in deformation
traveling towards the posterior area. Whereas standing and walking conditions with
lesser compression load in comparison (Fig. 5a, b), still has a stronger influence over
DeformaƟon values
4.60
4.06 4.06
(mm)
2.98 3.16
2.42
standing walking jumping twisƟng coughing laughing

Types of daily acƟviƟes
Fig. 4 Natural disc displacement variation for different activities

Loading
Displacement Von mises stress
condition
Standing
Load
applied
700 N
(a) Maximum displacement - 2.42 mm (g) Maximum stress – 1.24 N/ mm2
Walking
Load
applied
850 N
(b) Maximum displacement - 2.98 mm (h) Maximum stress – 1.39 N/ mm2
Jumping
Load
applied
1100 N
(c) Maximum displacement - 4.06 mm (i) Maximum stress – 1.71 N/ mm2
Twisting
Load
applied
900 N
(d) Maximum displacement - 3.16 mm (j) Maximum stress – 1.44 N/ mm2
Cough
ing
Load
applied
1100 N
(e) Maximum displacement - 4.06 mm (k) Maximum stress – 1.71 N/ mm2
Laugh
ing
Load
applied
1200 N
(f) Maximum displacement - 4.60 mm (l) Maximum stress – 1.93N/ mm2
Fig. 5 Contour plots of the natural disc under various loading conditions a deformation plot b stress
plot (the anterior and posterior positions of the disc have been depicted in Fig. 1b)
a slightly larger portion of the anterior region of the intervertebral disc followed by
a gradual decrease in its value bringing out the uniform distribution of deformation.
With the increasing value of load under varying daily activities, the maximum
stress has been observed to considerably increase in magnitude. The range of max-
imum stress observed in the disc model varies from 1.24 to 1.93 N/mm2 . For all
mentioned activities under the study, the stress induced in natural disc noticeably
increases with depth as it reaches the L5 section. Stress concentration is more preva-
lent in the posterior area of disc consisting of annulus pulposus for all activities
undertaken in our study.
In cases of walking, standing and twisting maximum stress is observed to be
concentrated at the posterior disc rim (Fig. 5g, h, j). In contrast to this, rest activities
show a lesser concentration of maximum stress in the same position. Incremental rise
in the magnitude of stress in the lateral direction of the disc has been observed under
jumping, coughing and laughing activities as shown in Fig. 5i, k, l. Furthermore, in
activities of coughing, jumping and laughing maximum stress is found to have more
concentration at depths closer to L5 surface compared to the rest.
4 Conclusion
In conclusion, the FE model generated was useful in carrying out a deformation

analysis of intervertebral discs under varying loads that act on the lumbar spine region
in accordance with the activity performed or posture attained. The loading value is
influenced if a complex posture or activity is performed. It is seen that high forces on
the spinal segment increase the amount of displacement and the stress induced in the
natural disc and which eventually causes degeneration. Disc deformation and stress
plots suggested activities like coughing, jumping and laughing have a greater impact
on natural disc functioning chosen in this study. The results obtained can be used
further in postoperative motion analysis of the lumbar spine region in performing a
variety of daily tasks.
Acknowledgements This work was supported by the Biomechanics lab, Department of Mechan-
ical Engineering, SRMIST, Kattankulathur–603203. We thank Mr. Raja. D, Asst Professor
(S.G) and Research Head, Biomechanics lab, Department of Mechanical Engineering, SRMIST,
Kattankulathur–603203. For assistance to improve the manuscript.
References
1. Kurtz SM, Edidin A (2006) Spine Technology handbook, 2nd edn

2. Rohlmann A, Pohl D, Bender A, Graichen F, Dymke J, Schmidt H, Bergmann G (2014)
Activities of everyday life with high spinal loads. PLoS One 9(5):e98510
3. Bible JE, Biswas D, Miller CP, Whang PG, Grauer JN (2010) Normal functional range of motion
of the lumbar spine during 15 activities of daily living. J Spinal Disord Tech 23(2):106–112
4. Li G, Wang S, Passias P, Xia Q, Li G, Wood K (2009) Segmental in vivo vertebral motion

during functional human lumbar spine activities. Eur Spine J 18(7):1013–1021
5. Dreischarf M, Schmidt H, Shirazi-Adl A, Arjmand N, Rohlmann A (2015) Estimation of loads
on human lumbar spine: a review of in vivo and computational model studies. J Biomech
49(6):833–845
6. Ledet EH, Sachs BL, Brunski JB, Gatto CE, Donzelli PS (2000) Real-time in vivo loading in
the lumbar spine: part 1. Interbody implant: load cell design and preliminary results. Spine
(Phila. Pa. 1976) 25(20):2595–2600
7. Han KS, Rohlmann A, Zander T, Taylor WR (2013) Lumbar spinal loads vary with body height
and weight. Med Eng Phys 35(7):969–977
8. Ledet EH, Tymeson MP, DiRisio DJ, Cohen B, Uhl RL (2005) Direct real-time measurement
of in vivo forces in the lumbar spine. Spine J 5(1):85–94
9. Wang S, Park WM, Kim YH, Cha T, Wood K, Li G (2014) In vivo loads in the lumbar L3-4
disc during a weight lifting extension. Clin Biomech 29(2):155–160
10. Miller JAA, Schultz AB, Warwick DN, Spencer DL (1986) Mechanical properties of lumbar
spine motion segments under large loads. J Biomech 19(1):79–84
11. Rohlmann A, Zander T, Rao M, Bergmann G (2009) Applying a follower load delivers realistic
results for simulating standing. J Biomech 42(10):1520–1526
12. Li H, Wang Z (2006) Intervertebral disc biomechanical analysis using the finite element
modeling based on medical images. Comput Med Imaging Graph 30(6–7):363–370
13. Renner SM, Natarajan RN, Patwardhan AG, Havey RM, Voronov LI, Guo BY, Andersson GBJ,
An HS (2006) Novel model to analyze the effect of a large compressive follower pre-load on
range of motions in a lumbar spine. J Biomech 40(6):1326–1332
14. Shin DS, Lee K, Kim D (2007) Biomechanical study of lumbar spine with dynamic stabilization
device using finite element method. CAD Comput Aided Des 39(7):559–567
15. Schmidt H, Shirazi-Adl A, Galbusera F, Wilke HJ (2010) Response analysis of the lumbar
spine during regular daily activities-A finite element analysis. J Biomech 43(10):1849–1856
16. Dreischarf M, Zander T, Shirazi-Adl A, Puttlitz CM, Adam CJ, Chen CS, Goel VK, Kiapour
A, Kim YH, Labus KM, Little JP, Park WM, Wang YH, Wilke HJ, Rohlmann A, Schmidt H
(2014) Comparison of eight published static finite element models of the intact lumbar spine:
predictive power of models improves when combined together. J Biomech 47(8):1757–1766
17. Yamamoto I, Panjabi MM, Crisco T, Oxland T (1989) Three-dimensional movements of the
whole lumbar spine and lumbosacral joint. Spine J. 14(11):1256–1260
18. White III AA, Panjabi MM (1990) Clinical biomechanics of the spine, 2nd edn
Numerical Simulation of a Small-Scale
Shock Tube Using OpenFOAM®
Satheesh Kumar Aravind, Anugya Singh , and B. T. Kannan
Abstract Shock waves are a phenomenon that can be produced under constrained
conditions inside a shock tube. These waves have a wide scope of studies due to
their multi-disciplinary applications. Therefore, in the present work, the main focus
is to conduct transient numerical simulations of a small-scale shock tube which
has been designed and fabricated. Comparison of 1D, 2D, and 3D cases, and grid
independence studies have been carried out successfully along with parametric study
by variation of driver section pressures. Simulations are conducted using the open-
source Computational Continuum Mechanics toolbox, OpenFOAM®. Processing
of the simulated data is carried out using Paraview via quantitative and qualitative
methods.
Keywords Shock tube · Shock wave · Transient · OpenFOAM® · Visualization
1 Introduction
The shock waves can be created using various experimental setups for studying the
behavior and physics of different gas dynamics phenomena. It is multi-disciplinary
in nature and thus there are many ways to generate a shock wave, one such apparatus
which produces shock wave is a shock tube. It can generate a shock wave with the
sudden rupture of a diaphragm which separates the driver (high-pressure) and driven
(low-pressure) sections.
The initial idea and the field of research in shock tube were developed a century
ago and since then many experimental advancements have been achieved in this
S. K. Aravind (B) · A. Singh · B. T. Kannan (B)

Department of Aerospace Engineering, SRM Institute of Science and Technology, Kattankulathur
603203, India
B. T. Kannan
A. Singh
https://doi.org/10.1007/978-981-15-4488-0_27
298 S. K. Aravind et al.
field by different pioneering scientists and engineers throughout the last century. The
introduction of Computational Fluid Dynamics (CFD) brought in a major change in
the field of Fluid Mechanics. Numerical simulations of normal shock wave reflections
were carried out by Weber et al. [1] in a two-dimensional channel for the investigation
of the unsteady viscous interaction aspects of bifurcation of the shock. It indicated
that the shear layer in the bifurcation being unsteady in nature and the vortices of
large and small scales lead to the creation of complex flow patterns and heat transfer
phenomena. Amir et al. [2] developed a two-dimensional Euler solver for shock tube
applications which validated the performance of a short-duration hypersonic shock
tube facility based on shock speed and pressure. An experimental and numerical study
on shock was made by Moradi et al. [3] in which both driver and driven gas used
was air. The CFD was conducted in a two-dimensional geometry with a transient
form of laminar viscous flow regime and the results were in good compatibility
with the experimental results. Further numerical simulations on the flow evolution
mechanism inside the shock tube were done by Kiverin and Yakovenko [4] in which
the flow patterns behind the shock waves and temperature non-uniformities in shock
tube with ignition were investigated. Another detailed parametric CFD study on the
characteristics of the shock train inside the shock tube was done by Kim et al. [5]
which on the complex flow phenomena due to the interaction caused between the
reflected shock wave and boundary layer leading to the generation of shock train.
Since it has multi-disciplinary applications Luan et al. [7] conducted a shock tube
simulation attached with a small exit nozzle to support the analysis of chemical
mixtures and its reactions. The flow physics inside the tube and the nozzle were
investigated and it was concluded that for most chemical kinetic applications, the
influence of flow field inside the shock tube can be neglected.
Thus, the main focus of the present work is to conduct numerical simulations
using the open-source Computational Continuum Mechanics toolbox, OpenFOAM®
on a shock tube using two different solvers viz. sonicFoam and rhoCentralFoam, to
conduct qualitative (visualization) and quantitative (pressure, density, and veloc-
ity) post-processing and compare the results obtained with the theoretical and the
experimental results which have been carried out [6].
2 Computational Methodology
2.1 Geometry and Meshing
The geometry was created along x, y, and z-axes using OpenFOAM® . Geometries
were created for one-dimensional (1D), two-dimensional (2D) and three-dimensional
(3D) cases with a length of driver section of 1 m and driven section of 1.5 m, as can
be seen in Figs. 1 and 2. The 1D and 2D geometry were created the same with unit
thickness and the boundary conditions determine the solution of the equations in the
Numerical Simulation of a Small-Scale Shock … 299
Fig. 1 Geometry for 1D and

2D case
Fig. 2 Geometry for 3D

case
required direction. The geometry was created with respect to the experimental setup
which had been fabricated [6].
For 1D and 2D geometries, a single block and in the 3D case, a four-blocks were
used with following edge/face names, i.e., Ends, Walls, Front, and Back, to obtain
uniform structured grid and the ease of meshing. Many iterations of meshing were
considered following from 1D to 3D for capturing of the shock wave. Figures 3, 4,
and 5 are provided as an example of the meshes that were created, and Table 1 shows
the number of cells included in each case.
Fig. 3 1D mesh with 5000 cells
Fig. 4 2D mesh with 25,000 cells

Fig. 5 3D mesh (front and side view) with 1,000,000 cells
Table 1 Number of cells and its distribution in each case of mesh

Case Case number Number of divisions in each Total number of cells
direction
X Y Z
1D 1 2000 – – 2000
2 2500 – – 2500
3 3000 – – 3000
4 5000 – – 5000
5 10,000 – – 10,000
2D 6 5000 5 – 25,000
7 5000 10 – 50,000
8 5000 40 – 200,000
3D 9 5000 5 5 500,000
10 5000 5 10 1,000,000
2.2 Governing Equations
The flow inside the shock tube in “sonicFoam” which is a transient pressure-based
solver for trans-sonic/supersonic, laminar or turbulent flow of a compressible gas is
governed by the Eqs. 1–6 [8]. The mass and momentum conservation equations are
given as follows:
∂ρ
+ ∇ · (ρ u) = 0 (1)
∂t
∂ρ u
+ (
u · ∇)ρ u − ∇ · Reff = −∇ p (2)
∂t
The energy equation is given as follows:
∂ρe ∂ρ K
+ ∇ · (ρ ue) + + ∇ · (ρ u K ) + ∇ · (
u p) − ∇ · (αeff ∇e) = 0 (3)
∂t ∂t
where, ρ: Density, e: Internal Energy, K: Kinetic Energy, p: Pressure, u :

Velocity Vector, Reff : Stress Tensor notation in OpenFOAM® , αeff is the effective
thermal diffusivity.
The governing equation of “rhoCentralFoam” which is a density-based compress-
ible flow solver based on central-upwind schemes of Kurganov and Tadmorare [9].
The continuity equation is given as follows:
∂p ∂
+ (u i ρ) = 0 (4)
∂t ∂ xi
The momentum equation is given by:

∂ ∂p

∂u i
+ ui u j + =0 (5)
∂t I ∂x j ∂ xi
The energy equation is given as follows:

∂E ∂
∂ ∂u j ∂u i 2 ∂u k
+ uk E + p − μu j + − δi j = 0 (6)
∂t ∂ xk ∂ xi ∂ xi ∂x j 3 ∂ xk
I

From E , the temperature is calculated through,

1 E uk uk
T = − (7)
Cv ρ 2
where p: Pressure, T: Temperature, t: time, ρ: Density, μ: Dynamic Viscosity, δ:

Kronecker delta, u i : velocity vector, k: Turbulent Kinetic Energy, x: Space Variable,
i, j, k: Tensor Indices, E: Energy.
The algorithm used by both the solvers is PIMPLE which is a combination of
SIMPLE (Semi-Implicit Method for Pressure-Linked Equations) and PISO (Pressure
Implicit with Splitting of Operators).
2.3 Boundary Conditions
For all the cases, “patch” was assigned to both the ends. In 1D case, all the rest of
the faces were assigned “empty” whereas in 2D case, only the front and back faces
were assigned “empty” while the top and bottom were assigned to ‘wall (no slip)’.
However, in the 3D case, all the faces except the ends were assigned to “wall (no
slip)”.
The pressure in the driver and driven sections were 199,948 Pa and 101,325 Pa
respectively, and temperatures in the driver and driven sections were 300.15 K for
the initial case validation with the experimental results. Four more cases were carried
out with all other parameters as constant and increase in driver section pressures as
follows: 225,000 Pa, 250,000 Pa, 275,000 Pa and 300,000 Pa.
A region was defined using coordinates from the position of the diaphragm to the
end of the driven section of the tube and the gas in both the sections was given as air.
The simulation was made to run from t = 0 s to t = 0.0035 s, to draw similarities
from the experimental data [6].
Paraview was used to conduct post-processing of the data obtained from the numerical
simulations. Initially, grid independence study was carried out in each direction, as
can be seen from Fig. 6, curves of Case 1 to 3 and Case 4 to 5 overlap over each
other completely. The change in curvature from Case 3 to 4 can be observed, due to
the increment in the number of cells. Since Case 4 was most efficient and accurate,
the number of cells in the X direction was fixed to 5000 for further analysis.
Similarly, grid independence study was carried out for 2D and 3D cases separately
and it was found that all of them overlapped successfully. Hence, the number of cells
in Y and Z direction was fixed at 5 and 10, respectively, in order to capture the
properties of the shock wave with increasing time. Another comparison was drawn
between the selected cases from 1D, 2D, and 3D as shown in Fig. 7. It can be seen
that all the curves overlap over each other with a very slight difference. Therefore,
for saving computational time and resources, Case 4 was selected for conducting
parametric studies by varying the driving section pressures.
225000
Pressure (Pa)
Case 1
175000
Case 2
Case 3
125000
Case 4
Case 5
75000
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 1.25 1.5
Position along the length of Shock Tube (m)
Fig. 6 1D grid independence study at time t = 0.0015 s
225000
Pressure (Pa)
1D
175000 2D
125000 3D
75000
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 1.25 1.5
Fig. 7 1D, 2D, and 3D case comparison at time t = 0.0015 s

225000
sonicFoam
Pressure (Pa)
175000 rhoCentralFoam
125000
75000
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 1.25 1.5
Fig. 8 Comparison of solvers at time t = 0.0015 s
Once the meshing was finalized, the comparison of both solvers was carried out.
As shown in Fig. 8, only a slight difference is present between both the curves which
can be observed at the pressure drop that occurs in the driver section region across the
expansion fan. Thus, for obtaining data on the velocity of the particles in between the
shock wave and the expansion fan and the density inside the shock tube sonicFoam
and rhoCentralFoam were used, respectively.
The difference in pressure at different time steps can be seen in Figs. 9 and 10. At
t = 0 s, the change of pressure is visible as a sudden single drop in the curve, after
the rupture of the diaphragm as the time increases there are two pressure drops in
the in curvature, due to the creation of expansion fan and shock wave. The constant
225000 t=0 secs

Pressure (Pa)
175000 t=0.0005 secs

125000 t=0.0015 secs
75000 t=0.0025 secs
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 1.25 1.5
Position along the length of Shock Tube (m) t=0.0035 secs
Fig. 9 Pressure variation inside the shock tube at different time steps using “sonicFoam”
Fig. 10 Pressure contour plot inside the shock tube at different time steps using “sonicFoam”
Par cle Velocity 100

(m/s) t=0.0005 secs
50 t=0.0015 secs
t=0.0025 secs
0
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 1.25 1.5 t=0.0035 secs
Posi on along the length of Shock Tube (m)
Fig. 11 Velocity of the particles in between the shock wave and the expansion fan inside the shock
tube at different time steps using ‘sonicFoam’
2.5
Density (kg/m3)
t=0.0005 secs
2.25
t=0.0015 secs
2
t=0.0025 secs
1.75
t=0.0035 secs
1.5
1.25
1
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 1.25 1.5
Fig. 12 Density variation inside the shock tube at different time steps using “rhoCentralFoam”
line after the first pressure drop depicts that the pressure does not change in between
the expansion fan and shock wave.
Figure 11 shows the particle velocity in between the expansion fan and shock wave
at different time steps. Since there is no pressure difference between the particles
behind the shock tube and behind the shock wave, therefore, the Velocity of Particles
remains constant in this region as can be seen from the figure.
The variation of density inside the shock tube at different time steps is shown
in Figs. 12 and 13 shows the gradient contour plot of density. The presence of the
contact surface which is an imaginary separation line between the gases in the driver
and driven section can be seen. There are three drops in the curves in each time step,
due to the sudden change in the density.
The temperature variation inside the shock tube at different time steps is captured
in Figs. 14 and 15. The temperature is uniform throughout the tube at time t = 0 s as
given in the input boundary condition. But, once the shock wave is produced, it can
be seen that there is a sudden increase in temperature behind the shock tube till the
contact surface where the temperature drops and remains constant till the expansion
fan. This sudden increase in temperature is caused due to abrupt pressure differences
created inside the shock tube.
Figure 16 shows the pressure variation inside the shock tube with different driver
section pressures. The distances between the curves increase with the increase in
pressure values, but due to the same driven section pressure given in all the cases,
curves overlap towards the end which is the portion in front of the shock wave.
The Mach number and the velocity of the particles in between the expansion fan
and shock wave were calculated using the following formulae [10, 11]:
Fig. 13 Density gradient plot inside the shock tube at different time steps using ‘rhoCentralFoam’
350 t=0 secs

Temperature (K)
325
t=0.0005 secs
300
275
t=0.0015 secs
250 t=0.0025 secs
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 1.25 1.5
t=0.0035 secs
Fig. 14 Temperature contour plot inside the shock tube at different time steps using “sonicFoam”
Fig. 15 Temperature contour plot inside the shock tube at different time steps using “sonicFoam”
Pressure (Pa) 325000 199948 Pa

275000
225000 225000 Pa
175000 250000 Pa
125000
75000 275000 Pa
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 1.25 1.5
300000 Pa
Fig. 16 Comparison of the pressure variation plots with different driver section pressures inside
the shock tube at t = 0.0015 s
P2 2γ1 2
=1+ M1 − 1 (8)
P1 γ1 − 1

2C1 1
V2 = M1 − + V1 (9)
γ1 + 1 M1
where P1 : Pressure in front of the Shock Wave (Pa), P2 : Pressure behind the Shock
Wave (Pa), γ1 : Heat Capacity Ratio (for air = 1.4), M1 : Shock Mach Number, V1 :
Velocity of particles in front of the Shock Wave (m/s) (i.e., 0), V2 : Velocity of particle
behind the Shock Wave (m/s), C1 : Speed of Sound in the medium (346 m/s for air
in 300.15 K).
Theoretical calculations were done using gas dynamic isentropic relations inside
a shock tube and the comparison of values is shown in Table 2.
The difference between the theoretical, experimental (in the case of 199,948 Pa
driver section pressure: Shock Mach Number = 1.155; Velocity of Particles =
83.39 m/s [6]) and computational Shock Mach Number is under 10.52% and for
the Velocity of Particles, it is under 1.63% for all the cases. This is due to the lack of
following boundary conditions: a The viscous effects of the fluid and friction effects
added by the walls. b As the shock wave moves forward, a boundary layer is formed
behind the shock wave which causes dissipation in the kinetic energy as heat which
is conveyed to the walls via heat transfer. This causes deceleration of shock wave
and acceleration of the contact surface as the time increases.
Table 2 Comparison of the values obtained with different driver section pressures
Driver section pressure (Pa) Shock mach number Velocity of particles (m/s)
Computational Theoretical Computational Theoretical
199,948 1.282 1.16 84.5 85.9
225,000 1.296 1.19 99 100.82
250,000 1.309 1.21 112.08 112
275,000 1.321 1.24 123.91 125
300,000 1.333 1.26 134.7 134.5
4 Conclusion
Numerical simulation of a Shock Tube was conducted using the open-source Com-
putational Continuum Mechanics toolbox, OpenFOAM® . The geometry used was
inspired by the fabricated Shock Tube [6]. Grid independence study was carried out
in 1D, 2D, and 3D to fix the number of divisions in each direction using ten different
mesh cases. The pressure curves obtained were compared to finalize the mesh for
further analyses. The selected mesh was used to compare two different solvers viz.
sonicFoam and rhoCentralFoam. The sonicFoamsolver was used to study the differ-
ences in Pressure and Temperature and also the Velocity of Particles in between the
Shock Wave and Expansion Fan, and rhoCentralFoam solver was used to study the
density differences at various time steps. The driver section pressures were varied
to conduct a parametric study keeping all other parameters as constant. The Shock
Mach Number and Velocity of Particles were calculated using the data obtained and
compared for the cases of driver section pressure values. The difference in the simu-
lated and theoretical data was found to be due to the lack of comprehensive boundary
conditions. Pressure and Temperature contour and Density gradient plots revealed
the magnitude of the pressure, temperature, and density gradient, respectively, at dif-
ferent time steps along with the direction of the motion of Shock Wave, Expansion
Fan, and the contact surface inside the shock tube.
Acknowledgements The authors acknowledge the usage of Advanced Computing Laboratory

(ACL), Department of Aerospace Engineering at SRM Institute of Science and Technology
(SRMIST) for the simulations performed towards the present research work.
References
1. Weber YS, Oran ES, Boris JP, Anderson JD Jr (1995) The numerical simulation of shock
bifurcation near the end wall of a shock tube. Phys Fluids 7:2475–2488
2. Amir A-F, Yusoff M, Yusaf T (2008) Numerical simulation of inviscid transient flows in shock
tube and its validations. Int J Math Phys Eng Sci 2
3. Moradi A, Fard FM, Mani M, Ranjbar M, Mottaghipour M (2012) Experimental and numerical
investigation of AUT shock tube. Eur J Exp Biol 2(5):1820–1826
4. Kiverin A, Yakovenko I (2018) On the mechanism of flow evolution in shock-tube experiments.
Phys Lett A 382(5):309–314. ISSN 0375-9601
5. Kim DW, Kim TH, Kim HD (2017) A study on characteristics of shock train inside a shock
tube. Theor Appl Mech Lett 7(6):366–371. ISSN 2095-0349
6. Singh A, Aravind S, Kannan BT (2020) Studies on blasting effects of shock waves from a small
– Scale shock tube. AIP Conference Proceedings, vol 2204, p 030003
7. Luan Y, Olzmann M, Magagnato F (2018) Simulation of a shock tube with a small exit nozzle.
J Thermal Sci 27
8. Petrescu M (2015) Computational fluid dynamics in mesoscopic nozzles. M.S. thesis, McGill
University
9. Winter M (2013) Benchmark and validation of Open Source CFD codes, with focus on com-
pressible and rotating capabilities, for integration on the SimScale platform. M.S. thesis,
Chalmers University of Technology
10. Bokil M (2010) Characterization of the pressure wave from a shock tube using numerical
simulations. M.S. thesis, The University of Utah
11. Information on https://nptel.ac.in/courses/101106040/chapter%203.pdf. (Module 13, Experi-
mental Aero (Gas) Dynamics, NPTEL Course by Dr. Job Kurian)
Contact Stress Analysis on a Functionally
Graded Spur Gear Using Finite Element
Analysis
V. Aravind, S. Adharsh, D. Prakash, S. Martin Vinoth, and P. Marimuthu
Abstract Gears are one of the most essential and most significantly used power
transmitting parts. Spur gears are more commonly used in practice as it is easy
to manufacture and it is simple in design. The failure of gear is mainly attributed
to stresses developed in the gears. Contact and bending stresses developed in the
gears significantly reduce the lifetime of the gears being used. Materials which are
tailored for specific applications can be put under test to minimise the stresses devel-
oped and increase the lifetime. Functionally graded materials are advanced mate-
rials with spatial gradation in composition to achieve specifically tailored proper-
ties. This work attempts to implement functionally graded materials (aluminium-
steel, steel-zirconium) in the gears drives. Further, finite element analysis is used
to evaluate its contact stresses for various distributive laws (exponential, linear, and
power). In addition to that, the contact stress of the functionally graded spur gear is
compared to a conventional gear material (EN8 Steel) to evaluate its contact load
capacity. The contact stresses were assessed for various torque values. It was found
that the change in the distributive laws influenced the contact stresses induced in
gear. Also, aluminium-steel showed comparatively less contact stress compared to
steel-zirconium and homogeneous steel.
Keywords Functionally graded material · Finite element analysis · Spur gear ·

Contact analysis
1 Introduction
Gears are one of the most widely used power transmitting parts. This is because of its
ability to change the magnitude, speed, and direction of the power source. Contact and
bending stresses become an important parameter in determining the life of the spur
gear. Contact stress in the dominant stress comparatively. Lewis equation and Hertz
contact equations are analytically used to calculate the bending and contact stress
V. Aravind · S. Adharsh · D. Prakash · S. Martin Vinoth · P. Marimuthu (B)

School of Mechanical Engineering, SASTRA Deemed to Be University, Thanjavur, India

https://doi.org/10.1007/978-981-15-4488-0_28
310 V. Aravind et al.
of the spur gear widely [1]. Bending stresses do not influence the materials being
considered [2]. Finite element analysis proves out to be a useful technique to solve the
contact stress problem [3, 4]. Bharat Gupta et al. suggested that the change in module
had changed the contact behaviour of the gear [5]. Seouk-chuk Hwang et al. carried
out contact stress analysis on a 2D spur gear on the various changes in positions
using AGMA standards and found that the design considered was stricter than the
AGMA Standards [6]. Gaurav Mehta et al. conducted contact stress analysis on spur
gear using two different materials and suggested that a wide range of materials could
provide a variety of contact stresses [7]. This was also confirmed by the results of K.
Sivakumar et al. who conducted contact stress analysis for different materials [8]. S.
Rajesh Kumar et al. used a composite material in spur gear analysis using FEM with
different rotational speed and found out that it influenced the behaviours like stress,
strain, deflection, etc. [9]. The above literature shows the influence material has on the
contact stress of a spur gear. Experimental procedures tend to be not only expensive,
but any design error is very laborious to rectify, whereas finite element technique
is relatively simple and is comparatively less arduous. Hence, finite element-based
contact stress analysis is perceived as an alternative to experimental methods.
Functionally graded materials are advanced composite materials with spatial gra-
dation in its structure and composition to achieve specifically tailored properties
[10]. Functionally graded materials are characterised by changing the property gra-
dient concerning its spatial position. Tohid Mahmoudi et al. conducted a thermo-
mechanical analysis on a functionally graded wheel-mounted brake disc and con-
cluded that the stresses and factor of safety considerably improved compared to the
conventional cast iron [11]. Haidar F. Al-Qrimli et al. reviewed and suggested that
FGM could solve the contact tooth surface problem of the gears [12]. M. M. Shahza-
manianet et al. conducted thermal and mechanical analysis on a functionally graded
brake disc and found mechanical and thermal properties could be influenced by the
distribution of FGM with variable parametric constant distributed under power law
[13]. Aravind et al. conducted the bending stress analysis for various distributive laws
and found that the distributive laws had considerable influence on the deformation
of the spur gear [14]. Soufi Mohammadi et al. concluded from his stress and strain
analysis of a functionally graded annular plate and found that inhomogeneity of the
material influenced the mechanical behaviour of the article [15].
From the above-reviewed literature, it was found that functionally graded materi-
als can be employed in spur gears for enhancement of contact properties. This work
tries to evaluate the contact stress behaviour of two pairs of the functionally graded
spur gear. The materials are distributed radially along the involute profile of the gear.
Various laws such as linear, exponential, and power laws were used. Power laws
were varied with different parametric constant, and its behaviour was studied. Other
behaviours such as equivalent strain and total deformations were also observed for
different cases. The problem was using a commercially used FE software ANSYS.
The problem was solved for two pairs of FGM using various distributive laws.
Contact Stress Analysis on a Functionally Graded … 311
2 Spur Gear Design and Model
The 2D spur gear was designed and simulated in ANSYS 14.5. The design of the 2D
spur gear is given in Table 1. The geometry of the spur gear is shown in Fig. 1. The
spur gear design was divided into 20 equal segments radially, and material properties
were distributed across its involute profile based on the governed law.
2.1 Numerical Simulation and Procedure
Finite element analysis of the given spur gear was performed in ANSYS 14.5. The
simulation was 2D plane strain [16]. The driving end of the gear was given a torque
and frictionless support, and the driven end was given fixed support on the inner
rim. The contact was frictional with a frictional coefficient of 0.15 for the homo-
geneous case [4]. Since the coefficient of friction of functionally graded material
is unknown, no separation contact pair was assumed. The contact formulation of
augmented Lagrange was chosen. Quad mesh with an element size of 1 mm was
used to solve the problem. A fine mesh was provided in the edges to get an accurate
result. The element size was fixed with convergence study and is shown in Fig. 2.
Table 1 Spur gear parameter

Parameter Value
Pitch circle diameter 65 mm
Internal diameter 31.75 mm
Outer diameter 68 mm
Pressure angle 20°
Face width 10 mm
No. of teeth 20
Fig. 1 Geometry of the spur gear

Fig. 2 Mesh convergence

study
The boundary conditions and the meshed geometry are shown in Figs. 3 and 4.
Fig. 3 Boundary condition

Fig. 4 Meshed geometry
2.2 Material Property and Distributive Law
The pairs of functionally graded materials chosen were Al-steel and steel-zirconium.
Homogeneous pairs of steel and zirconium were selected for comparison [14]. The
materials were distributed across the gear involute profile. The material was dis-
tributed from the inner radius to the outer radius. The inner radius of the gear had the
former FGM material dominant (aluminium and steel). The outer radius of the gear
had later FGM material dominant (steel and zirconium). Using the inner and outer
radius as the boundary condition and the distributive laws as the equation, an equation
for distribution was generated. The equation was used to generate discrete material
property for the radially divided 20 segments of the gear. The material property was
designated to the 20 equally divided segments based on the equation. The distributive
laws used were linear, exponential, and power law with four parametric constants.
The properties of the materials are shown in Table 2. The governing equations are
shown in Table 3.
Where P(r) is the property chosen across the radial profile, β = (1/(a − b)) * ln
(PA /PB ), a and b are the inner and outer radius of the gear, m is the slope, and c
is the y-intercept. PA is the property of the material in the inner region, and PB is

S. No. Material Young’s modulus (GPa) Density (kg/m3 ) Poisson’s ratio
1 Aluminium-steel E A = 70 ρ A = 2700 μA = 0.3
E B = 200 ρ B = 8166 μB = 0.33
2 Steel-zirconium E A = 200 ρ A = 8166 μA = 0.33
E B = 244 ρ B = 5700 μB = 0.288
Table 3 Distribution laws

Law Representation Equation
Exponential EX P(r) = Po eβr
Linear LI P(r) = mr + c
Power law P(k) P(r) = PA + (PB − PA ) * (x − a/a − b)k
the property of the material in the outer region. k is the parametric constant. Hertz
equation was used to calculate the maximum contact stress and for validation of the
case.

F 1 + R1
R2
σC = (1)
R1 Bπ ( 1−ϑ12 )
+ ( E2 2 ) sin φ
1−ϑ 2
E1
where σc is the maximum contact stress, F is the force, R1 and R2 is the pitch circle
radii, B is the face width of the spur gear, E 1 and E 2 are the Young’s modulus of the
gear, ϑ is the Poisson’s ratio, and Φ is the pressure angle.
The test was conducted for two material pairs (Al-Steel and Steel-Zr). The torque
was varied from 100 to 250 Nm with an increment of 50 Nm. The materials were
distributed using linear, exponential, and power law. Four values of the parametric
index were used in power law (k = 1, 2, 3, 4). As a sample case, the von Mises
stress distribution of homogeneous steel is shown in the figure. The maximum stress
developed is 529.98 MPa, under a torque of 100 Nm.
3.1 Influence of Torque
In all cases, the maximum von Mises stress increases with an increase in the torque
value. The variation of contact stress with torque is shown in Fig. 5. Similar trends are
observed in both materials. The maximum stress in case of Steel-Zr is 1021 MPa and
for Al-Steel is 911 MPa. The value of contact stress increases by 82.6% for an increase
of contact stress from 100 to 250 for Steel-Zr distributed exponentially. Similarly,
an increase of 85.5% was seen in Al-Steel sample distributed exponentially. Al-
Steel showed comparatively less stress for all the laws under the given torque value.
The distribution of Young’s modulus and Poisson’s ratio could in the contact region
account for the difference in the contact stresses of the material.
Fig. 5 von Mises stress plot for a sample case of homogeneous steel
The variation of equivalent strain with torque is shown in Fig. 5. As depicted

in the figure, the equivalent strain is found out to be less in the case of Steel-Zr
compared to Al-Steel. The highest value of strain is obtained for Al-Steel pair under
a torque of 250 Nm as 0.0055. The equivalent strain increases with an increase in the
torque value. The Steel-Zr shows a percentage increase of 80.8% increase in its strain
value when the torque is varied from 100–250 Nm. Similarly, the Al-steel shows an
increase of 90.5%.
Similar to stress and strain, the total deformation also varies linearly concerning
torque. The variation of total deformation under torque is shown in Fig. 5. It can be
seen from that equivalent strain, and total deformation is less in the case of Steel-Zr
and is high for Al-Steel sample. This can be explained by the distribution of Young’s
modulus in the contact region. Increase in Young’s modulus increases the stress,
whereas it has an indirect influence on the strain and total deformation values. The
highest deformation was found for the Al-Steel sample (7.28 × 10−5 m). The least
was 1.72 × 10−5 m for the Steel-Zr under 100 Nm. The percentage increase when
the torque varies between 100 and 250 Nm was 141% for Steel-Zr and 144% for
Al-Steel.
3.2 Influence of Material Distribution
The materials were distributed along the radial profile of the spur gear in 6 different
fashion [linear, exponential, power law (k = 1, 2, 3, 4)]. The influence of material
distribution under a constant torque is shown in Fig. 6. The contact stress is found
out to be highest for the Steel-Zr under linear distribution. It can also be seen that
contact stresses developed are almost similar for power law k = 1 and linear. This
is because of the similarity in distributional behaviour. The highest value of contact
stress is 575 MPa for Steel-Zr under linear distribution; as the value of k increases,
Fig. 6 Variation of contact stress, equivalent strain, total deformation varying torque
it can be seen that the values of contact stress decrease. The value of k can further be
increased beyond 4 to increase the value of contact stress. In all the cases, Steel-Zr
showed comparatively more stress than Al-Steel.
The variation of equivalent strain under varying material distribution is shown in
Fig. 6. The equivalent strain is less in the case of Steel-Zr and more comparatively
for Al-Steel. The value of equivalent strain is least for the linear distribution law. A
power law with k = 1 and linear law show similar strain behaviour. As the k value
increases, it can be seen that the value of strain increases. The highest value observed
for Al-Steel under power law k = 4 as 0.00329.
Total deformation shows similar behaviour as strain under various material dis-
tributions. The variation of total deformation under different laws is shown in the
figure. The total deformation increases with increase in the value of k. Total defor-
mation is higher for Al-Steel comparatively. The highest value of total deformation
is observed for Al-Steel as 4.06 × 10−5 m under power law k = 4 and the lowest for
Steel-Zr as 1.5728 × 10−5 m under linear distribution (Fig. 7).
4 Conclusion
Functionally graded materials showed significantly fewer stresses when compared to

conventional materials. The amount of stresses induced in Al-Steel pair is about 15%
less than that of traditional steel material (power law k = 4). The distributive laws
played a significant role in deciding the stresses, strains, and the total deformation
induced in the materials. A power law with k = 4 showed least stresses in both the pairs
of FGMs. The total deformation is less as in the case of linear law. This is because of
the distribution of material with higher Young’s modulus in the contact region. This
also explains why the contact stresses are high compared to other distributive laws as
Fig. 7 Variation of contact stress, equivalent strain and total deformation under various law
contact stress is directly proportional to the square root of Young’s modulus. Also,
varying the material distribution can significantly help us in designing the properties
of the materials in the contact region. Hence, functionally graded materials can be
used as a substitute in place of conventional materials in gears as it helps us in
tailoring the desired properties and can be used to increase the lifetime of the gears
significantly.
Acknowledgements The authors are grateful towards Science & Engineering Research Board
(SERB), Department of Science and Technology, India, for providing financial grant (File No.
ECR/2017/000425) to do this work.
References
1. Pawar P, Utpat A (2015) Analysis of composite material spur gear under static loading condition.
Mater Today Proc 2:2968–2974. https://doi.org/10.1016/j.matpr.2015.07.278
2. Prabhakaran S, Ramachandran S (2014) Bending stress and contact stress analysis of a spur
gear for material steel C45. Inf (Jpn) 17:4599–4611
3. Hassan A (2009) Contact stress analysis of spur gear teeth pair. World Acad Sci Eng Techn
58:611–616
4. Farhan M, Karuppanan S, Patil S (2015) Frictional contact stress analysis of spur gear by
using finite element method. Appl Mech Mater 772:159–163. https://doi.org/10.4028/www.
scientific.net/AMM.772.159
5. Balaji D, Prabhakaran S, Harish Kumar J (2017) Analysis of surface contact stress for a spur
gear of material steel 15Ni2Cr1Mo28. ARPN J Eng Appl Sci 12
6. Hwang S-C, Lee J-H, Lee D-H, Han S-H, Lee K-H (2013) Contact stress analysis for a pair of
mating gears. Math Comput Model 57. 10.1016/j.mcm.2011.06.055
7. Mehta G, Somani M, Narendiranath Babu T, Watts T (2018) Contact stress analysis on com-
posite spur gear using finite element method. Mater Today Proc 5:13585–13592. https://doi.
org/10.1016/j.matpr.2018.02.354
8. Sivakumar K, Saravana Manikandan MT, Sethuraman B, Rajagopalan S, Hariharan R (2015)
Contact stress analysis of a spur gear tooth pair for two different materials. Int J Appl Eng Res
10
9. Rajeshkumar S, Manoharan R (2017) Design and analysis of composite spur gears using the
finite element method. IOP Conf Ser Mater Sci Eng 263:062048. https://doi.org/10.1088/1757-
899X/263/6/062048
10. Shirvanimoghaddam K, Naebe M (2016) Functionally graded materials: A review of fabrication
and properties. Appl Mater Today 5:223–246. https://doi.org/10.1016/j.apmt.2016.10.001
11. Mahmoudi Tohid, Parvizi Ali, Poursaeidi E, Rahi Abbas (2015) Thermo-mechanical analysis
of functionally graded wheel-mounted brake disk. J Mech Sci Technol 29:4197–4204. https://
doi.org/10.1007/s12206-015-0914-3
12. Al-Qrimli H, Ataollahi Oshkour A, Alnaimi F, Mahdi, FA (2015) Material design consideration
for gear component using functionally graded materials. Int J Mater Eng Innov 6:243–256.
10.1504/IJMATEI.2015.072849
13. Shahzamanian M, Sahari B, Bayat M, Mustapha F, Ismarrubie ZN (2013) Elastic contact
analysis of functionally graded brake disks subjected to thermal and mechanical loads. Int J
Comput Methods Eng Sci Mech 14:10–23. 10.1080/15502287.2012.698701
14. Aravind V, Adharsh S, Prakash D, Babu K (2019) Stress analysis on functionally graded spur
gear: select proceedings of ICEMMM 2018. https://doi.org/10.1007/978-981-13-1724-8_39
15. Soufi Mohammadi A, Sabokroohieh P, Falakian A, Hadi A, Hosseini M, Ehsani F (2013) Stress
and strain analysis of functionally graded annular plate subjected to transverse loading
16. Cornell RW (1983) Compliance and stress sensitivity of spur gear teeth. J Mech Des 103.
10.1115/1.3254939
Finite Element Analysis of Knee Joint
with Special Emphasis on Patellar
Implant
M. A. Kumbhalkar, D. T. Rangari, R. D. Pawar, R. A. Phadtare, K. R. Raut,

and A. N. Nagre
Abstract Patella is supporting part of knee and also guide for quadriceps or patellar
tendon. Patellar implant is used for proper functioning of patella after injury. For
implant, it is required to cut injured portion of host patella and keep remaining part
minimum up to 12–14 mm and overstuffing of 2 mm to prevent patellar fractures.
The patellar implant makers provide only 8-mm-thick implant to maintain original
patellar thickness which is difficult to achieve especially in patients with host bone
thickness less than 20 mm. Hence, there is need to analyze for reduction in thickness
of patellar implant from 8 to 6 mm for perfect engagement with adequate residual
bone. The critical force analysis on host patella with tendon is carried out for quadri-
ceps force, patellofemoral force, and patellar tendon force using analytical and finite
element method. Two cases are considered for the force and stress analysis of patella,
and the comparison of various implant thicknesses is discussed.
Keywords Patella · Knee joint · Finite element analysis · Patellar implant

(button) · Patellofemoral force
M. A. Kumbhalkar
Assistant Professor, Department of Mechanical Engineering, JSPM Narhe Technical Campus,
Pune, Maharashtra, India
D. T. Rangari
Lecturer, Department of Pharmacy, Government Polytechnic, Amravati, Maharashtra, India
R. D. Pawar (B) · R. A. Phadtare · K. R. Raut · A. N. Nagre
Student, Department of Mechanical Engineering, JSPM Narhe Technical Campus, Pune,
Maharashtra, India

https://doi.org/10.1007/978-981-15-4488-0_29
320 M. A. Kumbhalkar et al.
Fig. 1 Anatomy of knee and biomechanical model of knee joint with patellar implant [1]
1 Introduction
The rotary motion of knee joint is based on the successful movement between femur-
tibia and patella. The tendons attached at femur and tibia cover and compress patella
toward femur while rotary motion of knee joint. Due to compressive force or frictional
force between femur and patella, the patella may damage and therefore it is required
to provide implant over it. Figure 1 shows biomechanical model of knee joint with
patella implant.
For the possible movement of knee joint, it is necessary to maintain patellar
thickness after knee arthroplasty [1]. The host patella thickness is required to maintain
up to 12–14 mm during knee arthroplasty for the patellar thickness less than 20 mm,
and standard implant of 8 mm is used to recover original patellar thickness [1–3].
Sometimes intraoperative patellar thickness increases due to 8 mm standard patella
implant for the person having original patellar thickness less than 20 mm. If the
original thickness of patella is not maintained by applying patellar implant, then the
knee cannot bend at that extent and pain will occur during walking, climbing, or
running. Hence, for perfect meeting of original patella thickness after operation, it is
required to use patella implant having thickness less than 8 mm. This research aim
is to check effect on patellar implant by reducing its thickness from 8 mm to 6 mm
at various flexion angles.
2 Mathematical Formulation and Force Analysis
As per the experimental study done in various literature papers, it is observed that
the tendons attached with tibia and femur will stretch while motion due to which a
compressive force is acted on patella. The magnitude of forces and direction acting
on leg vary with the knee flexion angle and weight. Two lateral forces i.e., patellar
tendon force and quadriceps force are acting on patella which acts patellofemoral
Finite Element Analysis of Knee Joint … 321
reactive force toward contact point of patella and femur. Patellar tendon force acting
between the patella and tibia, patella quadriceps force acting between patella and
femur, and patella femoral force acting on patella as a compressive because of the
forward moment are locked due to the quadriceps and ligament. A free body diagram
for forces acting on patella is shown in Fig. 2.
The patellofemoral force is acting on patella at various knee flexion angles at
different conditions of human movement. The full body weight of human is acting
on knee joint at straight condition which is fixed at end of tibia, but while walking,
climbing, or running, the moment acts at joint which distributes body weight into
different forces. The free body diagram of patellofemoral joint with force distribution
and joint flexion moment is shown in Fig. 3 [4, 5]. As patella changes its own
position according to flexion angle, forces acting on patella change according to the
flexion angle. The main three forces acting on patella, i.e., quadriceps force, patella
tendon force, and patellofemoral forces are derived by considering various angular
movements at joint.
Also other two elements are plotted as free body diagrams, i.e., for patellar tendon
and for quadriceps tendon where the forces, angles, and the different lengths are
shown. For force analysis on patella, specific geometric form is not considered but
the specific thickness of patella is considered during finite element analysis. Fekete
et al. [4] considered dimensionless parameters to simplify the results of relationship
between patella and tendon and are given in Table 1.
Fig. 2 Forces acting on patella [13]

Fig. 3 Free body diagram of patellofemoral joint and joint flexion moment by Fekete et al. [4] and
Mason et al. [5]
Table 1 Dimensionless parameter [4]

Description Formulas
Dimensionless, intersected tibia length function λ1 = l1/l10
Dimensionless, intersected femur length function λ3 = l3/l30
Dimensionless length patella tendon λp = lp/l10
Dimensionless thickness of shin λt = lt/l10
Dimensionless thickness of thigh λf = lf/l30
2.1 Patellar Tendon Force
From Fig. 3, free body diagram of tibia and femur gets separated to find its force
distribution. Figure 4 illustrates the forces acting on tibia due to patellar tendon and
body weight. The lengths denoted in free body diagram as length of tibia (l10 ), length
between line of action of tibia and body weight (l1 ), length of patellar tendon (lp ),
length from axis of tibia, and tibia tuberosity (lt ). Taking the moment at point B, by
fixing the point B.
Fig. 4 Forces acting on tibia and patellar tendon

MB = 0
0 = −lp · f pt · sin β − lt · f pt · cos β + l1 BW · sin γ
l1 BW · sin γ = lp · f pt · sin β + lt · f pt · cos β

l1 BW · sin γ = f pt lp · sin β + lt · cos β
lp · sin γ
f pt = BW · (1)
lp · sin β + lt · cos β
Rearranging Eq. (1)
f pt lp · sin γ
= (2)
BW lp · sin β + lt · cos β
f pt λ1 · sin γ
= (3)
BW λp · sin β + λt · cos β
2.2 Quadriceps Force
Figure 5 illustrates the forces acting on femur due to quadriceps tendon and body
weight. The parameters denoted in free body diagram are length of femur (l30 ), length
between line of action of tibia and body weight (l3 ), angle between the axis of femur
and quadriceps force (ψ), and angle between the axis of femur and line of action of
body weight (δ). Taking moment at point B, by fixing the point B.
Fig. 5 Forces acting on femur

Mb = 0
0 = lf · f q · cos ψ + l30 f q · sin ψ − l3 BW · sin δ
δ =α−γ
ψ = 0 Assumption
l3 BW · sin δ = lf · f q · cos ψ + l30 f q · sin ψ
f q (lf · cos ψ + l30 · sin ψ)
BW =
l3 · sin δ
fq l3 · sin δ
=
BW l f · cos ψ + l30 · sin ψ
fq λ3 sin(α − γ )
= (4)
BW λf
2.3 Patellofemoral Force (Fpf )
It is reaction force acting on femur through patella and ligament forces that is quadri-
ceps and tendon forces which is obtained from F q and F pt forces by parallelogram
theorem of resultant forces.
By using equilibrium equation,

fy = 0

fx =0
0 = f q sin δ − f pt sin(γ + β) + Fpfx
Fpfx = − f q sin δ − f pt sin(γ + β)
0 = f q cos δ − f pt cos(γ − β) + Fpfy
Fpfy = −Fq cos δ + Fpt cos(γ + β)
Resultant of two concurrent forces (Law of Parallelogram)

f pt
2
f pfx + f pfy
2
f q2 + f pt2 − 2 f q f pt cos(β + δ + γ )
= =
BW BW BW
f pt f q + f pt − 2 f q · f pt cos(β + δ + γ )
2 2
= (5)
BW BW
2.4 Case Study
Two cases having different body weight and height are considered to find patellar
tendon force, quadriceps force, and patellofemoral force as per the formulae derived
in Sects. 2.1, 2.2, and 2.3 for various flexion angles. The manual drawings of both
cases having weight 74 kg and 61 kg and height 5.9 ft. and 5.5 ft., respectively, are
drawn to plot various lengths and angle for calculation of forces. As the tibia and
femur length vary with respect to height and weight, these two different cases are
considered for force calculation. The manual drawings of two cases are shown in
Fig. 6, and all parameters like length, angle, and dimensionless values are shown in
Table 2.
The mathematical formulation represents forces acted on knee joint at quasi-static
condition, and it is exposed to high magnitude of compression force during walking,
climbing, and running condition [6, 7]. As per literature [8–10], the force acting on
knee joint during walking is 1.3 times body weight, during climbing stair is 3.3 times
body weight, and during knee bends is 7.8 times body weight. The compression force
and area of articular contact increase with knee flexion and are max between 60° and
90° [11, 12].
3 Finite Element Analysis of Patella
In this paper, finite element analysis technique is used for stress analysis of knee
joint with special emphasis on patellar implant with varying thickness from 6 to
8 mm. For finite element analysis of patella, a critical CAD model of knee joint
(a) CASE-I: Person having weight 73 kg (b) CASE-I: Person having weight 61 kg
and height 5.9 feet and height 5.5 feet
Fig. 6 Manual drawing of two different case study for force calculation
Table 2 Parameters of two

Sr. No. Parameter CASE-I CASE-II
different cases
1. BW 73 kg 61 kg
2. L1 29.5 cm 27.5 cm
3. L3 18.5 cm 21.2 cm
4. Lf 5.5 cm 4.5 cm
5. Lp 12 cm 11 cm
6. Lt 2.7 cm 2.3 cm
7. L 10 42 cm 38.5 cm
8. L 30 43 cm 41 cm
9. λ3 0.4302 0.517
10. λ1 0.7023 0.714
11. λp 0.2857 0.285
12. λt 0.0642 0.059
13. λf 0.1279 0.109
14. γ 30 º 35º
15. δ 55º 47º
16. A 85º 82º
17. B 10º 8º
with patella and tendon is prepared in solid modeling software CATIA and its stp
file is imported in FEA tool, ANSYS. At first, the analytical patellofemoral force is
verified with ANSYS result by applying boundary condition as patella tendon force
and quadriceps force at the end of tendon and is fixed and at femur end. This reaction
force is exactly matched by analytical and finite element method, and the results are
shown in Table 3. All three forces are also calculated for different conditions such as
walking, climbing, and running as per the ratio discussed in 2.4, and patellofemoral
force is verified in ANSYS at different knee flexion angles.
The research aim is to replace standard available patellar implant of 8 mm thick-
ness to minimum 6 mm thickness for the host patella size of less than 20 mm thick-
ness. Hence, the stress analysis is carried out on host patella first and checks stresses
induced at point of contact of patella and femur. Then, the stress analysis is car-
ried out on patella with implant for different thicknesses. For stress analysis, direct
patellofemoral force is applied on upper surface of patella and femur is fixed at end.
The patellar implant with varying thickness is shown in Fig. 7, stress analysis result
on contact point of patella and femur is shown in Fig. 8, and stresses on patella are
Table 3 Force acting on patella for two different cases and at different flexion angles using analytical
and finite element method
Condition Case Case-I Case-II
Flexion 65° 85° 110° 65° 85° 110°
angle
Static F pt 2427.42 2289.7 4460.6 2894.90 2539.36 3976.63
Fq 1714.39 2026.1 3022.5 2207.16 2109.85 3021.42
F pf 2390.31 3186.9 6358.2 3015.2 3301.48 5545.8
F pf 2390.3 3186.9 6358.3 3015.2 3301.5 5545.8
(FEA)
Walking F pt 3186 2983.7 5850.5 3763.3 3301.1 5169.5
Fq 2242.9 2652.8 3958.5 2869.3 2742.8 3927.8
F pf 3134.8 4161.6 8335.1 3919.5 4291.8 7209.5
F pf 3134.9 4161.7 8335 3919.7 4291.9 7209.5
(FEA)
Climbing F pt 8087.6 7574.1 14851.3 9553.2 8379.9 13122.9
Fq 5693.6 6733.9 10048.5 7283.6 6962.5 9970.60
F pf 7957.65 10564 21158.2 9950.3 10894.9 18301.1
F pf 7957.8 10564 21158 9950.2 10895 18301
(FEA)
Running F pt 19116.2 17902.4 35103.2 22580.2 19807 31017.71
Fq 13457.6 15916.8 23750.9 17215.8 16456.8 23567.12
F pf 18808.9 24970.2 50010.5 23518.7 25751.5 43257.32
F pf 18,809 24,970 50,010 23,519 25,752 43,257
(FEA)
Fig. 7 Patellar implant (button) with varying thickness [3]
Fig. 8 Equivalent stress and maximum principal stress acting on contact of patella and femur at
knee joint
shown in Fig. 9.
Stress analysis on host patella with patellar implant is also carried out using finite
element analysis. The host patella is cut from interior side and keeps 12 mm thickness
of posterior part of patella where 8 mm implant is fitted. The maximum principal
stress and equivalent stress on patellar implant are shown in Fig. 10, and the results
are given in Table 4 for flexion angle of 110° and 65°.
Fig. 9 Equivalent stress and maximum principal stress acting on patella

Fig. 10 Equivalent stress and maximum principal stress on patellar implant using finite element
analysis
Table 4 Maximum principal stress and equivalent stress result for host patella and patellar implant
for various condition at 110° and 65° flexion angle
Angles 65° 110°
Condition Stresses Host 8 mm 6 mm Host 8 mm 6 mm
patella implant implant patella implant implant
Static Max 4.146 14.115 8.109 12.926 56.241 53.608
principle
stress
Equivalent 8.8343 24.579 22.4 40.827 70.487 69.158
stress (von
misses)
Walking Max 5.438 18.515 10.637 16.946 73.729 70.277
principle
stress
Equivalent 11.58 32.24 29.38 53.522 92.405 90.662
stress (von
misses)
Climbing Max 13.804 46.998 27.001 43.016 187.16 178.39
principle
stress
Equivalent 29.412 81.84 74.58 135.86 234.56 230.14
stress (von
misses)
Running Max 32.629 111.08 63.817 101.67 442.38 421.66
principle
stress
Equivalent 69.521 193.43 176.29 321.13 554.43 543.97
stress (von
misses)
4 Results Validation
Dr. Anoop Jhurani et al. [2] carried out an experimentation on 54 female patient
knees to restore the native patellar thickness less than 20 mm. This experimentation
has been observed for the period of two year which proved that the 6.2 mm thickness
plastic patellar button can be used instead of 8 mm thickness to restore native patel-
lar thickness less than 20 mm. A vernier caliper was used to measure and restore
intraoperative patellar thickness during total knee arthroplasty as shown in Fig. 11.
Shelburne et al. [8] and van Eijden et al. [13] reported critical force analysis
to estimate the quadriceps force, patello tendon force, and patellofemoral force at
different gait cycle and flexion angles shown in Fig. 12. At maximum extension of
patellofemoral joint quadriceps force is obtained as 2000 N, whereas the maximum
force is obtained as 8000 N at 75° flexion angle. Patello tendon force becomes
progressively smaller than quadriceps force at large flexion angles and reaches a
maximum of 5000 N at 60° flexion angle. At all flexion angles, a patellofemoral
force is smaller than quadriceps force [13].
Also, the forces acting on patella for two different cases during walking (Fig. 13)
are revealed that the quadriceps force produced for maximal extension is 2243 N
at 65° flexion angle and maximum patellofemoral force is 8335 N at 110° flexion
angle. The forces acting in the literature paper nearly equal the forces acting on the
patella for the cases considered in research.
A finite element analysis has been carried out to find stresses in patellofemoral
cartilage due to internal and external rotations of the femur influencing contact areas,
pressures, and cartilage stress distributions [14]. Besier et al. [14] analyzed a finite
element mesh model of patellofemoral joint as shown in Fig. 14 which illustrated the
element contact between quadriceps and patellar tendon and also within the femoral
and patellar cartilage. The simple joint stresses due to contact area and stresses in
patellar cartilage shown in Fig. 14 by finite element method revealed that the peak
Fig. 11 Intraoperative measurement of patellar thickness using vernier caliper [2]

Fig. 12 Forces acting on patella versus flexion angle [8, 13]
Fig. 13 Force acting on patella during walking for two different cases and at various flexion angles
stress values is obtained as 3.5 MPa at static or neutral condition for 15° internal and
external femoral rotation which is lower than mean stress values.
A finite element analysis of patellar knee joint also revealed that the maximum
principal stress is obtained as 4 MPa at static condition for host patella which is
increased to 8 MPa for 6 mm patellar implant at 65° flexion angle. The stresses in
patella are increased for different conditions, i.e., walking, climbing, and running.
Fig. 14 Finite element mesh of the patellofemoral joint and correlations between simple measures
of joint stress and stresses estimated in the patellar cartilage by the finite element method [14]
From finite element analysis, it is observed that the stress in 6.2 mm patellar implant
is less than 8 mm patellar implant for every condition as shown in Fig. 15.
Fig. 15 Maximum principal stress and equivalent stress for host patella and patellar implant during
walking
5 Conclusion
The mathematical formulation for biomechanical model of patellar knee joint is

carried out to find forces acting on patella using free body diagram. The patellar
tendon force, quadriceps force, and patellofemoral forces are calculated for two
different cases of 61 and 73 kg person at various flexion angles, i.e., at 65°, 85°, and
110°. The analytical patellofemoral reaction forces are verified using finite element
analysis which is obtained by applying patellar tendon force and quadriceps force
at the end. The stress analysis on host patella and patellar implant with 8 mm and
6 mm thickness is carried out which reveals that the stress increases for 8 mm patellar
implant than host patella and the stress reduces for 6 mm patellar implant than 8 mm.
From the results, it can be concluded that the 6 mm patellar implant can be used
instead of 8 mm implant during intraoperative arthroplasty for patella thickness less
than 20 mm.
References
1. Abolghasemian M, Samiezadeh S, Sternheim A, Bougherara H, Lowry Barnes C, Backstein

DJ (2014) Effect of patellar thickness on knee flexion in total knee arthroplasty. J Arthroplasty
2(1):80–84
2. Jhurani A, Agarwal P, Aswal M, Saxena P, Singh N (2018) Safety and efficacy of 6.2 mm
patellar button in resurfacing less than 20 mm thin patella: a matched pair analysis. Knee Surg
Rel Res 30(2):153–160
3. Bengs BC, Scott RD (2006) The effect of patellar thickness on intraoperative knee flexion and
patellar tracking in total knee arthroplasty. J Arthroplasty 21(5):650–655
4. Fekete G, Csizmadia BM, Wahab MA, De Baets P, Vanegas-Useche LV, Bíró I (2014)
Patellofemoral model of the knee joint under nonstandard squatting. Dyna 81(183):60–67
5. Mason JJ, Leszko F, Johnson T, Komistek RD (2008) Patellofemoral joint forces. J Biomech
41:2337–2348
6. Kainz H, Reng W, Augut P, Wurm Simone (2012) Influence of total knee arthroplasty on
patellar kinematics and contact characteristics. Int Orthop 36(1):73–78
7. Nareliya R, Kumar V (2011) Biomechanical analysis of human femur bone. Int J Eng Sci
Technol 3(4)
8. Shelburne KB, Torry MR, Pandy MG (2005) Muscle, ligament, and joint-contact forces at the
knee during walking. Med Sci Sports Exerc 37(11)
9. Heegard J, Leyvraz PF, Curnier A, Rakotomanana L, Huiskes R (1995) The biomechanics of
the human patella during passive knee flexion. J Biomech 28(11):1265–1279
10. Pokorny J, Kren J (2007) The patellofemoral joint and the total knee replacement. Appl
Computat Mech 595–602
11. Magister S, Yarboro S (2018) Biomechanical evaluation of a novel suture augment in patella
fixation. Am J Orthop 46(6):E468–E473
12. Hsu HC, Luo ZB, Rand JA, NauAn K (1996) Influence of patellar thickness on tracking and
patellar femoral contact characteristics after total knee arthroplasty. J Arthroplasty 11(1):69–80
13. van Eijden TMGJ, Weijs WA, Kouwenhoven E, Verburg J (1987) Forces acting on the patella
during maximum voluntary contraction of the quadriceps femoris muscle at different knee
flexion/extension angles. Acta Anat 129:310–314
14. Besier TF, Gold GE, Delp SL, Fredericson M, Beaupre´ GS (2008) The influence of femoral
internal and external rotation on cartilage stresses within the patellofemoral joint. J Orthopaedic
Res
Simulation and Hardware
Implementation of Interleaved SEPIC
Converter with Valley-Fill Circuit
for HBLED System
B. Lakshmi Praba and R. Seyezhai
Abstract LED lamps are nowadays preferred for lighting compared to the fluores-
cent lamps due to its longer lifetime and low power consumption. But the design of
driver circuits designed for high brightness (HB) LED demands high power factor,
more reliability, high efficiency and precise control. Therefore, to achieve this, a
single-stage power factor correction model (PFC) and valley-fill is proposed in this
paper. Valley-fill incorporated in SEPIC reduces the strain on output diode and middle
capacitor, thus achieving great power factor and competence. Further to moderate
the ripple at input and output stages, an interleaving concept is applied to SEPIC
with valley-fill circuit. The combination of interleaving and valley-fill circuit leads
to reduced voltage ripple, improved power factor and efficiency. A simulation study
of the proposed SEPIC converter is carried out by MATLAB/SIMULINK. The per-
formance factors such as source power factor, supply THD, supply distortion fac-
tor, supply displacement factor, power loss, output voltage ripple and efficiency are
computed, and the results are compared with the classical SEPIC converter without
valley-fill circuit. A prototype of an interleaved SEPIC converter is assembled to
confirm the simulation results.
Keywords Light-emitting diode (LED) · Valley-fill · Power factor correction

(PFC) · Source displacement factor · Total harmonic distortion (THD) ·
Interleaved · High brightness (HB)
B. Lakshmi Praba (B)

Department of Electrical and Electronics Engineering, Sri Sivasubramaniya Nadar (SSN) College
of Engineering, Kalavakkam, Chennai, India
R. Seyezhai
Department of Electrical and Electronics Engineering, REC Lab, Sri Sivasubramaniya Nadar
(SSN) College of Engineering, Kalavakkam, Chennai, India

https://doi.org/10.1007/978-981-15-4488-0_30
336 B. Lakshmi Praba and R. Seyezhai
1 Introduction
Electrical lighting sources are important in day-to-day life. The first incandescent
lamp was invented in 1879, followed by the invention of fluorescent lamps in 1938,
which achieved a milestone in the history of electrical engineering. Greatly new
types of electrical lamps were devised, namely mercury lamps (HPM), sodium lamps
(HPS), metal halide lamps (MH), and flash lamps filled with inert gas. Due to the
disadvantages of electrical lamps, this paper deals with LED. On comparing with the
electrical lamps, LEDs are more efficient, very small in size, light up very quickly,
and it will achieve full brightness in under a microseconds. A LED lamp is a type of
solid-state lighting that contains light-emitting diodes as the source of light.
For this LED lighting applications, we can choose any type of converters like
boost, buck, Cuk and SEPIC. But, this paper fully deals with SEPIC converter because
the SEPIC converter has improved supply power factor and reduced ripple related
to other types of converters. A single-stage interleaved SEPIC converter has 50%
reduced ripple compared to classic SEPIC converter.
A single-stage interleaved SEPIC converter for LED lighting is discussed in
project where the model is functioned in intermittent conduction mode (DCM) with
buck operation [1, 2]. The proposed topology is a modified SEPIC converter by com-
bining the classical SEPIC and interleaved SEPIC with the middle capacitor substi-
tuted by a valley-fill circuit. This additional circuit is employed to reduce stress on
the freewheeling diode and the middle capacitor, thus achieving high power factor
and efficiency. Furthermore, interleaving concept is applied to the proposed SEPIC
which reduces the ripple at input and output stage. The combination of interleaving
and valley-fill circuit results in reduced output voltage ripple, improved power factor
and efficiency.
Simulation studies of the SEPIC converter are done using MATLAB/SIMULINK.
The performance factors such as input power factor, supply THD, supply distortion
factor, supply displacement factor, power loss, output voltage ripple and efficiency are
computed, and the results are compared with the classical SEPIC. To normalize the
output and improvement in source power factor, closed-loop current control methods
are implemented.
The broadsheet is arranged as follows: Division 2 discusses about different topolo-
gies of AC–DC SEPIC converter. Division 3 focuses on the design equation of pro-
posed model. Division 4 grants the simulation results. Division 5 is pact with current
control techniques, and finally Division 6 grants the conclusion and the results.
2 Proposed Topologies
A DC–DC single-ended primary inductor is an individual kind of converter. It may

operate as buck, boost and also buck–boost. The SEPIC converter output has been
Simulation and Hardware Implementation of Interleaved … 337
Fig. 1 Circuit diagram for AC–DC SEPIC converter
varied by changing the duty cycle of the devices. There is a mutual exchange between
the capacitor and the inductor in the SEPIC converter.
The topologies of AC–DC power factor modification converter analysed in this
project are as follows:
A. Conventional AC–DC SEPIC
B. Conventional AC–DC SEPIC with valley-fill circuit
C. Interleaved AC–DC SEPIC converter
D. Interleaved AC–DC SEPIC converter with valley-fill
A. Conventional AC–DC SEPIC Converter
From Fig. 1, when the main switch is conducting, current I L1 increases and the current
I L2 goes negative. When the device brought to off, the capacitor current I C1 is equal
to the current through the inductor I L1 , as inductor does not permit the instantaneous
change in the current [3]. When current through the inductor L 2 falls to zero, the
converter goes to DCM.
B. Conventional AC–DC SEPIC Converter with Valley-Fill
The circuit of SEPIC converter with valley-fill is exposed in Fig. 2. The valley-fill
comprises of two capacitors and three diodes. The two capacitors are charged in
series and parallel to feed the load. The converter operates in DCM. The switches
and passive elements are assumed to be ideal. The significance of valley-fill circuit
is that it leads to the output diode as well as the middle capacitor, maintaining the
power factor, resulting in high efficiency [4].
C. Interleaved AC–DC SEPIC Converter
Interleaving topology has been used in electronics field, predominantly in great power
applications. Since in great power applications, strain across the device can go outside
the limit, the power device cannot handle. In this case, interconnecting the devices
Fig. 2 Circuit diagram for AC–DC SEPIC with valley-fill
can provide the solution, but one major concern is the current sharing and voltage
sharing [5]. There is another solution, that is paralleling the converters instead of the
power devices. The concept of interleaving provides ripple cancellation and better
thermal performance. Figure 3 illustrates the circuit of AC–DC interleaved SEPIC.
The process of an interleaved SEPIC is like conventional SEPIC, but this is
operated in buck mode. In buck operation, there is no overlap between the device
conductions. Therefore, at a time one switch will conduct.
D. Interleaved AC–DC SEPIC with Valley-Fill
Fig. 3 Circuit of interleaved AC–DC SEPIC converter

Fig. 4 Course diagram for interleaved AC–DC SEPIC with valley-fill
Interleaved AC–DC SEPIC is displayed in Fig. 4. The converter works in discontin-

uous current mode. This topology can result in extraordinary power factor also able
to withstand great power using compact circuit.
When the device S 1 is turned on, S 2 is in off state. The capacitors C 1 and C 2
are charging towards input voltage with leftward positive plate. An inductor L 3 will
remain charging, capacitor C 2 will be discharging via S 2 , L 4 . It creates the diode D4
to be reversed bias. Subsequently, S 1 is turned off, D1 will be forward biased, and an
inductor L 1 , L 2 will be discharged via the load.
3 Design Equations
The SEPIC is operated in DCM mode, and the design equations are as follows:
The conversion gain is given by
Vo D
• Conversion gain = (1)
Vin 1− D
where V0 is the output voltage, Vin is the input voltage, and D D is the duty cycle.
The inductor value is given by
DVin
• Inductor L c = (2)
I0 2 f s
where f s is the switching frequency, I0 is 30 to 50% of I0 .

The filter capacitor value is given by
D I0
• Capacitor CC = (3)
V0 2 f s
where V0 is 1 to 5% of V0 .
4 Simulation and Results
All the topologies of SEPIC discussed in Sect. 4 are simulated in MAT-

LAB/SIMULINK. Constructed based on design equations, and also model parame-
ters are calculated and are exposed in Table 1.
The MATLAB simulation for conventional AC–DC SEPIC is exposed in Fig. 5.
The output of conventional AC–DC SEPIC is exposed Fig. 6, where the output
voltage obtained is about 15.50 V.
The supply current THD of conventional AC–DC SEPIC exposed in Fig. 7, where
the supply current THD obtained is about 18.38%.
The AC–DC conventional SEPIC with valley-fill MATLAB simulation diagram
is exposed in Fig. 8.
The output waveform of conventional AC–DC SEPIC with valley-fill is depicted
in Fig. 9, where an output voltage obtained is about 14.40 V.
The supply current THD of the conventional AC–DC SEPIC with valley-fill is
exposed in Fig. 10, where the supply current THD obtained is about 12.96%.
The simulation circuit of an interleaved AC–DC SEPIC is exposed in Fig. 11.
Output waveform of an interleaved AC–DC SEPIC is exposed in Fig. 12, where
the output voltage obtained is about 20.07 V.
Supply current THD of interleaved AC–DC SEPIC is exposed in Fig. 13, where
supply current THD obtained is about 14.28%.
Simulation diagram of an interleaved AC–DC SEPIC with valley-fill is exposed
in Fig. 14.
The output voltage of interleaved AC–DC buck SEPIC power factor circuit incor-
porating valley-fill is exposed in Fig. 15, where output voltage obtained is about
Table 1 Simulation parameters

Parameters Values
Input voltage 22 V
Inductors 110 µH
Capacitors 225 nF
Switching frequency 50 kHz
Duty ratio 0.45
Filter capacitor 497 µF
Load resistor 40
Fig. 5 MATLAB simulation diagram of conventional AC–DC SEPIC
Fig. 6 Conventional AC–DC SEPIC output voltage
18.92 V.
The supply current THD of the interleaved AC–DC SEPIC with valley-fill is
exposed in Fig. 16, where the supply current THD obtained is about 5.42%.
The performance parameters like output voltage ripple, supply THD, supply
power factor, supply distortion factor, supply displacement factor, power losses, and
efficiency are computed and that parameters are exposed in Table 2 [6].
From Table 2, the interleaved AC–DC SEPIC converter with valley-fill has the bet-
ter performance compared to the other three topologies. This topology has improved
power factor, supply current THD, displacement factor, reduced losses, and also
improved efficiency.
Fig. 7 Supply THD for conventional AC–DC SEPIC
Fig. 8 MATLAB simulation circuit of conventional AC–DC SEPIC with valley-fill
5 Closed-Loop Current Control Techniques
To further reduce the supply current THD, current shaping techniques are employed.
For this work, the methodologies such as peak current and average current control
methods are executed in MATLAB/SIMULINK. By these techniques, the power
factor is enhanced with lower harmonic profile of supply current [7].
• Peak current control technique
Fig. 9 Output voltage waveform of conventional AC–DC SEPIC with valley-fill
Fig. 10 Supply THD of conventional AC–DC SEPIC with valley-fill
In voltage loop, Verror is created and the output is matched with reference voltage.
A fault voltage, the sinusoidal reference current is multiplied to produce the reference
current. Whenever an inductor current reaches zero, switch is in on condition and it
reaches the reference current, reset the flip-flop and the switch tends to off state.
• Average current control technique
Average current control scheme is a two-loop control technique that has inner
current and the outer voltage control loop. The average inductor current is taken as
the reference and the inductor current is mandatory to follow it. The switch becomes
on; whenever the inductor current touches zero, the switch is turned off.
Fig. 11 MATLAB simulation circuit of interleaved AC–DC SEPIC
Fig. 12 Output voltage waveform of interleaved AC–DC SEPIC
The MATLAB simulation diagram for peak current control method of interleaved
AC–DC SEPIC is exposed in Fig. 17.
The output voltage waveform for the interleaved AC–DC SEPIC with valley-fill
in peak current control mode is exposed in Fig. 18, where output voltage obtained is
about 18.65 V.
The supply current THD for an interleaved AC–DC SEPIC with valley-fill in peak
current control is shown in Fig. 19, where the supply current THD obtained is about
4.66%.
The MATLAB simulation circuit for interleaved AC–DC SEPIC with valley-fill
in average current control mode is shown Fig. 20,
Fig. 13 Supply current THD of interleaved AC–DC SEPIC converter
An output voltage for interleaved AC–DC SEPIC with valley-fill in average current
control is exposed in Fig. 21, where output voltage obtained is about 18.53 V.
The supply current THD of interleaved AC–DC SEPIC with valley-fill in average
current control scheme is exposed in Fig. 22, where supply current THD obtained is
about 3.78%.
From Table 3, the interleaved AC–DC SEPIC converter with average current
control method has the better performance compared to other topologies. This has
the better power factor, improved efficiency, reduced supply current THD, and
reduced losses. Moreover an interleaved AC–DC SEPIC is most preferable for LED
applications.
6 Hardware Implementation
Hardware prototype of interleaved AC–DC SEPIC with valley-fill is developed

employing MOSFET switches, gate driver circuit, and power circuit. Pulses to the
switches are generated using ARDUINO software. These pulses are then given to the
ARDUINO board. The output of proposed converter and the performance parame-
ters are calculated and verified. Experimental setup of an interleaved AC–DC SEPIC
converter with valley-fill circuit is constructed based on the specifications that are
shown in Table 4.
Fig. 14 MATLAB simulation circuit of interleaved AC–DC SEPIC converter with valley-fill
Fig. 15 Output voltage waveform of an interleaved AC–DC SEPIC with valley-fill

Fig. 16 Supply current THD of interleaved AC–DC SEPIC converter with valley-fill
Table 2 Evaluation of performance parameters

Topology O/P O/P Supply Supply Supply Supply Total Efficiency
(V) ripple current power distortion displacement power (ï) (%)
(%) THD factor factor factor losses
(%) (W)
Con-SEPIC 15.5 10.2 18.38 0.779 0.034 0.78 12.5 68.75
Con-SEPIC 14.4 2.8 12.96 0.845 0.054 0.841 10.5 73.75
valley-Fill
Interleaved 20.0 6.85 14.28 0.873 0.0698 0.871 8.52 78.75
SEPIC
Interleaved 18.9 3.4 5.42 0.891 0.184 0.902 7.54 80.75
SEPIC
valley-fill
Pulses generated for the switches present in interleaved AC–DC SEPIC converter
using ARDUINO UNO board are shown in Fig. 23.
Experimental setup of the proposed converter contains a rectifier for AC–DC
conversion, and this rectifier is integrated with an interleaved SEPIC with valley-fill
is exposed in Fig. 24. An output voltage obtained is about 18.1 V.
Output voltage ripple is calculated using DSO and that is shown in Fig. 25. The
output voltage ripple obtained is about 3.11%.
Fig. 17 Simulation circuit of interleaved AC–DC SEPIC with valley-fill in peak current control
mode
Fig. 18 Output voltage waveform of interleaved AC–DC SEPIC in peak current control mode
Fig. 19 Supply current THD of interleaved SEPIC peak current control technique
7 Conclusion
In this paper, conventional AC–DC SEPIC is used for power factor correction in
supply side. Interleaved buck SEPIC power factor correction circuit incorporating
valley-fill circuit is investigated. From the results, it is found that the valley-fill
circuit results in reduced output voltage ripple, improved power factor, and better
spectral quality of the supply current waveform compared to the conventional SEPIC
converter. Moreover, the presented topology provides a high efficiency with reduced
losses, and hence, this converter will be a suitable one for LED applications. The
simulation has been carried out using MATLAB/SIMULINK. Hardware prototype
has been developed, and the results are validated.
Fig. 20 Simulation circuit of interleaved AC–DC SEPIC with valley-fill in average current
mechanism
Fig. 21 Output voltage waveform for average current control

Fig. 22 Supply current THD of interleaved SEPIC in average current control method
Table 3 Comparison between current control techniques and open loop

Techniques O/P O/P THD Power Distortion Displacement
voltage voltage (%) factor factor factor
ripple
(%)
Open loop 18.92 3.4 5.42 0.891 0.184 0.902
PCC 18.65 3.8 4.66 0.901 0.226 0.906
ACC 18.53 3.2 3.78 0.915 0.259 0.912
Table 4 Hardware
Parameters Values
specifications
Input voltage 22 V
Inductors 110 µH
Capacitors 225 nF
MOSFET switch IRF 840
Switching frequency 50 kHz
Filter capacitor 497 µF
Load 40
Diode MUR50
Transformer (0–24) V
Fig. 23 Pulse pattern for the switches
Fig. 24 Experimental setup
Interleaved SEPIC
Rectifier
Output Voltage
Arduino UNO
AC Supply
Fig. 25 Output voltage ripple. Maximum voltage is 18.1 V, minimum voltage is 17.7 V, output
ripple voltage is 3.11%
References
1. Hu Q, Zane R (2010) LED driver circuit with series-input connected converter cells operating
in continuous conduction mode. IEEE Trans Power Electron 25(3):574–582
2. Saravanan S, UshaRani P (2014) Review of bridgeless discontinuous conduction mode SEPIC
power factor correction rectifier. IEEE international conference (ICGCE), pp 594–599, 02 June
2014
3. Lodh T, Majumder T (2016) High gain and efficient integrated flyback SEPIC converter DC-
DC with leakage energy recovery mechanism. International conference on signal processing,
communication, power and embedded system, SCO PES, 7955689, 2016-10, 1109
4. Costa MAD, do Proda RN, Campos A, Seidal AR (2003) An analysis about valley-fill filters
applied to electronic ballasts. in Conference Record of IAS Annual Meeting, pp 779–784
5. Duran E, Galan J, Sidrach-de-Cardona M, Segura F (2008) An application of interleaved DC-DC
converters to obtain I-V characteristic curves of photovoltaic modules. In: 34th annual conference
of IEEE industrial electronics, April 2008, pp 2284–2289. http://dx.doi.org/10.1109/iecon.2008.
4758313
6. Calleja AJ, Alonso JM, Ribas J, Lopez E, Cardesin J, Garcia J, Rico-Secades M (2002) Electronic
ballast based on single stage high power factor topologies: a comparative study. In: Proceedings
of IEEE IECON Conference, pp 1196–1201
7. Azcondo FJ, Branas C, Casanueva R, Bracho S (2005) Power-mode controlled power-factor
corrector for electronic ballast. IEEE Trans Power Electron 52(1):56–65
Finite Element Analysis of a Two-Post
Rollover Protective Structure
of an off-Highway Motor Grader
V. Kumar, G. Mallesh, and K. R. Radhakrishna
Abstract Rollover protective structure (ROPS) is a critical safety system mounted

on off-highway equipments, which serves as a safety structure for an operator survival
during rollover accidents. ROPS is classified as single post, two post, four post, and
multi-post; selection of this ROPS depends on the type of the application. Commonly
used off-highway equipments are dump truck, articulated scrapers, dozers, water
sprinklers, excavators, loaders, motor grader, etc.; among these equipments, motor
graders were extensively used in off-highway and mining works. It is noticed from
the accidents summary reported by the Directorate General of Mine and Safety, India,
and Occupational Safety and Health Administration (OSHA), USA, that a significant
failure was noticed by the motor grader ROPS structure. Hence, it is necessary to
study the failure behavior and strength criteria of rollover protective structure. In the
present research work, attempts are made to study the failure behavior and strength of
a two-post motor grader (25 tons GVW) ROPS using finite element analysis (FEA)
software ANSYS. Further, an interlock section was introduced in the ROPS column
to minimize the overall size and to enhance the performance of the ROPS. Later,
comparative analysis has been done on this ROPS in terms of maximum deformation,
rate of energy absorption, and C.G height. It is evident from the results that proposed
ROPS has 30% reduction in overall size, 5% reduction in weight, and enhanced
C.G of height. Further, the proposed ROPS meets the performance requirement of
existing standards.
V. Kumar (B)
Department of Mechanical Engineering, Sri Jayachamarajendra College of Engineering,
Mysore, India
G. Mallesh
Department of Mechanical Engineering, Sri Jayachamarajendra College of Engineering,
JSS Science and Technology University, Mysore, India
K. R. Radhakrishna
BEML Limited, Bangalore, Karnataka, India

https://doi.org/10.1007/978-981-15-4488-0_31
356 V. Kumar et al.
Keywords Rollover Protective Structure · Deflection-Limiting Volume · Energy

Absorption · Motor Grader · FEA
1 Introduction
Rollover protective structure (ROPS) is a critical safety system mounted on off-

highway equipments, which serves as a safety structure for an operator survival
during rollover accidents (Fig. 1).
A survey of rollover accidents was conducted for off-highway vehicle operators at
the different parts of the country [1] and found that slope, uneven terrain, loose soil,
and harsh environment are related to the working situations (Fig. 2). Similarly, break
failure, over turn, high center of gravity, and visibility lead to the failure of ROPS.
Further, it is observed that dump truck, dozer, excavator, scrapers, and motor graders
were commonly used off-highway vehicles. It is evident from the Occupational Safety
and Health Administration (OSHA), USA, that around 16% of failure is noticed only
Fig. 1 Cause–effect diagram for ROPS failure
Fig. 2 Motor grader with ROP structure [BEML]

Finite Element Analysis of a Two Post … 357
Fig. 3 ROPS with variable

rectangular cross section
by the motor grader [2]. Hence, in this paper, attempts are made to design and develop
new ROPS structure for motor grader by considering the safety of an operator.
Figure 2 shows the motor grader which is commonly used in off-highway for
grading and is classified as light, medium, and heavy equipment based on gross
vehicle weight (GVW) of the machine. Companies such as BEML, Mahindra &
Mahindra, Komatsu, Volvo, Case, LeeBoy, Caterpillar, and Tata Hitachi are the major
manufacturer of motor grader in India. Based on GVW, various types of cross sections
were used to design and develop ROPS by various manufacturers. C-section, square,
rectangular, and variable sections are widely used along with circular sections in
ROPS to resist impact, bending, and torsion load during rollover. It is evident from
the literature and manufacturer catalogue that variable rectangular cross sections
shown in Fig. 3 were preferred in ROPS design due to its higher resistance to impact
load and energy absorption capability [3].
As per ISO: 3164 standard [4], ROPS is required to satisfy the safety requirements;
during rollover accident, any structural part should not enter deflection-limiting
volume (DLV) shown in Fig. 4 [5–7].
Many researchers are attempted to study the performance of ROPS using expen-
sive physical testing [5] and simulation modeling technique. Physical testing is a
tedious and cumbersome procedure; the simplicity and power of analytical tools,
like finite element method (FEM), have overshadowed intricate methods like physi-
cal testing. Henceforth, finite element analyses were carried out on existing two-post
ROPS structure and proposed ROPS model to evaluate the performance as per ISO:
3471 standards [8].
The detailed dimensions of the ROPS were taken physically on the existing ROPS
placed on the motor grader at manufacturer’s site. Based on the dimensions, 3D
models were developed using CAD modeling software.
358 V. Kumar et al.
Fig. 4 Size and

deflection-limiting volume
for ROPS
2 Finite Element Analysis of ROPS
The finite element method has evolved as a specialized technique for the general
numerical solution method applicable to a broad range of physical problems. In gen-
eral, the finite element method is based on a theory, whereby an original object is
viewed as an assembly of discrete building blocks called elements. The application
of the method involves dividing the body into an optimum number of blocks. These
blocks are connected to each other at specified points known as nodes, forming a
network called mesh. The number of elements is determined by two factors—the
capability of the computer being used and the accuracy of the results. The realistic
use of finite element method for solving the problem involves preprocessing, analy-
sis, and post-processing. The preprocessing involves the preparation of data such as
nodal coordinates, element connectivity, boundary conditions, loading, and material
properties. The analysis stage involves stiffness generation, stiffness modification,
and solution of the equations resulting in the evaluation of nodal displacements and
derived quantities such as stresses. The post-processing deals with the presentation
of the results. In the present research, 3D models were developed using CAD soft-
ware and are converted into *.iges file formats and are imported into the analyses
environment to carryout static and dynamic analyses of existing and proposed ROPS
(Figs. 5, 6 and 7).
Fig. 5 Existing ROPS drawings
Fig. 6 Proposed ROPS drawings
2.1 Physical and Mechanical Properties of ROPS
ASTM A 500 Gr C is commonly used material to manufacture ROPS by various

companies. Hence, in this research, ASTM A 500 Gr C material is used for existing
and proposed ROPS. The properties of ASTM A 500 Gr C are yield strength 325 MPa,
ultimate strength is 490 MPa, Young’s modulus 200 GPa, Poisson’s ratio 0.3, and
density is 7850 kg/m3 ; these properties and 10-node tetrahedral elements shown in
Fig. 8 are used to discretize the domain into number of finite elements. Finite element
model of the proposed ROPS structure is shown in Fig. 9.
360 V. Kumar et al.
Fig. 7 3-D models of existing and proposed ROPS
Fig. 8 3-D 10-node tetrahedral element
Fig. 9 Finite element model of a ROPS

2.2 Loads and Boundary Conditions for Static Analysis

of ROPS
It is noticed that as per the failure summary during rollover the ROPS is subjected to
various types of load such as lateral, vertical, and longitudinal loads. Magnitude of the
loads is calculated based on gross vehicle weight (GVW) as per ISO 3471 standard.
For a motor grader with the mass ranging from 2140 to 38,010 kg, the following
formulas are used to determine lateral, vertical, and longitudinal loads. These loads
are applied on two-post ROPS as shown in Fig. 10 to determine deformation and
stresses induced in the ROPS (Table 1).
1.1
m
Lateral Load = 70,000 × (1)
10,000
Vertical Load = 9.6 m (2)

1.1
m
Longitudinal Load = 56,000 × (3)
10,000
As the ROPS is mounted on the main frame of the motor grader using fasteners,
bottom of the posts is constrained in all the directions and different loading conditions.
Lateral load Vertical load Longitudinal load
Fig. 10 Position of loads on ROPS [1]
Table 1 Loads on ROPS

Machine Lateral load Vertical load Longitudinal
mass (kg) (N) (N) load (N)
25,750 198,131 504,958 158,505
362 V. Kumar et al.
3 Dynamic Analysis of ROPS
Same load and boundary conditions are used to determine the dynamic behavior of
the ROPS during rollover. During rollover as per the failure summary and literatures,
it is found that ROPS is subjected to impact load, and many researchers are attempted
to conduct experiments to know the dynamic behavior of the ROPS and noticed that
conducting the experiments is costlier, time-consuming, tedious, and cumbersome
procedures. Hence, attempts are made to analyze the dynamic characteristics of the
ROPS using FEA. Further, it is noticed that during rollover, ROPS are subjected to
impact load between the ground surface and the ROPS results in energy absorption
by both ROPS and the ground. To simplify the model, the ground surface was ide-
alized as a rigid body that is able to transfer kinetic energy to the ROPS (energy
absorbed by the ROPS). During 1971, Klose and Chou et al. introduced the concept
of FEA to estimate the energy absorption using MADYMO software. Therefore, it
is essential to determine the energy absorption capacity of ROPS to estimate the
failure criteria. Hence, in this research work, attempts are made to determine energy
absorption capacity using theoretical Eq. (4) as per ISO: 3471 standard and results
were compared with advanced finite element software.
1F1 F1 + F2 FN −1 + FN
U= + (2 − 1 ) + · · · + ( N − N −1 ) (4)
2 2 2
where
F = Force,
= Deflection
U = Energy absorption by ROPS.
4.1 Static Analyses
A critical analysis was carried out using finite element analyses software ANSYS
15.0 to estimate deformation due to lateral, vertical, and longitudinal load for the
existing ROPS. It is noticed from the Fig. 11 that maximum deformation of 47.1,
18.59, and 16.97 mm is due to lateral, vertical, and longitudinal loads, respectively.
Further, similar analyses were conducted for the proposed ROPS with same load and
boundary conditions. Figure 12 shows the deformation due to different loads; it is
evident from the results that deformation in the proposed ROPS may increase but the
proposed model exhibits safe deformation limits. It is also observed from the Table 2
that existing model is robust and consumes lots of material; further, it is observed
that center of gravity of the proposed ROPS is minimum compared to the existing
and hence the stability of the vehicle can be improved with the new design.
Lateral Load Vertical Load Longitudinal Load
Fig. 11 Deformation of existing ROPS
Lateral Load Vertical load Longitudinal Load

Fig. 12 Deformation of proposed ROPS
Table 2 Design
Description Existing Proposed
characteristics of existing and
ROPS ROPS
proposed ROPS
Mass, kg 565 542
Center of gravity (CG) of 1489 1422
height, mm
Deformation, Lateral 18.5 38.6
mm Vertical 47.1 57.2
Longitudinal 16.9 36.7
Von Mises Lateral 514 526
Stress, MPa Vertical 578 520
Longitudinal 195 324
364 V. Kumar et al.
The study revealed that results obtained from the FEA are in good correlation with
physical testing and results of other researchers [5, 7, 9–11]. Hence, the proposed
ROPS model achieved 30% reduction in size and 5% reduction in overall weight
with respect to the existing ROPS.
4.2 Dynamic Analysis
It is observed from the failure summary that ROPS is subjected to impact loads
which are dynamic in nature with time; therefore, it is necessary to study the dynamic
behavior before finalizing the design. It is evident from the literature that most of the
ROPS failure is due to lateral load. Hence, in this research work, dynamic behavior
of the proposed ROPS was carried out for lateral load to understand the failure.
It is evident from the literature that energy absorption capacity of the ROPS plays
a vital role in the dynamic response of a structure, and dynamic behavior of the
proposed ROPS is given in Table 3. Further, it is noticed that as per ISO:3471 energy
absorption capacity is dependent on the gross vehicle weight of the motor grader;
for the present study, energy absorption for the ROPS is 48,973 J and time taken
for the deformation of 38.68 mm is 28.89 ms. Similarly, for a load of 198 kN, the
energy absorbed by the structure for a displacement of 204.5 mm was found to be
658,845.3 J and the time taken to deform is 90 ms. Hence, as per the results, the
proposed ROPS is safe and meets standard performance requirements.
5 Conclusions
Finite element analyses of existing and proposed ROPS models for a motor grader
were conducted to know the static and dynamic behavior during rollover accidents
and found that the proposed ROPS enhances the safety of the operator; based on the
results, the following conclusions were drawn;
• Proposed Interlock cross section ROPS achieves 30% reduction in overall size.
• Lower the center of gravity height by 67 mm and significant weight reduction
• Optimization of size ensures better operator visibility.
• Due to multiple impact loads during rollover accidents energy absorption capacity
of proposed ROPS increases compared to the existing ROPS.
• Proposed ROPS structure meets the performance criteria as per ISO 3471 standard.
Table 3 Dynamic response

Time, ms Displacement, mm Load, KN Energy, J
of ROPS
0 0 0 −0.1825
5 10.0437 11 292.6258
10 22.90393 22 2782.431
15 37.00793 33 10155.37
20 39.43152 44 27282.38
25 57.86955 55 43847.95
28.8997 65.5 65 48973.18
30 67.03812 66 62680.51
35 80.79097 77 100824.5
40 89.95954 88 164361.1
45 103.7124 99 242545.6
50 112.881 110 330949.7
55 126.6338 121 429340.6
60 135.8024 132 537958
65 149.5553 143 446923.2
70 158.7238 154 489307.6
75 172.4767 165 531692
80 181.6452 176 574076.4
85 195.3981 187 616460.9
90 204.5667 198 658845.3
95 218.3195 209 701229.7
100 227.4881 220 743614.1
105 241.241 231 785998.5
110 250.4095 242 828382.9
115 264.1624 253 870767.3
120 287.0838 264 913151.7
References
1. International standard-3471 (2008) Earth-moving machinery—roll-over protective struc-

tures—laboratory tests and performance requirements
2. Hinze JW, Teizer J (2011) Visibility-related fatalities related to construction equipment. Saf
Sci 49:709–718
3. Nagel GM, Thambiratnam DP (2004) A numerical study on the impact response and energy
absorption of tapered thin-walled tubes. Int J Mech Sci 46(2004):201–216
4. International standard-3164 (2013) Earth-moving machinery–Specification for deflection
limiting volume
5. Thambiratnam DP, Clark BJ, Perera NJ (2009) Performance of a rollover protective structure
for a bulldozer. J Eng Mech 135:31–40
6. Stockton AD, O’Neill DH, Hampson CJ (2002) Methods for optimising the effectiveness of
roll-over protective systems. Silsoe Research Institute for the Health and Safety Executive,
366 V. Kumar et al.
United Kingdom, 425/2002

7. Karlinski J, Dzialak P, Ptak M (2013) Simulation tests of roll-over protection structure. Arch
Civil Mech Eng 13:57–63
8. Mandal A, Sengupta D (2000) The analysis of fatal accident in Indian coal mines. Indian
statistical institute, 203 BT Road Kolkata -700035, AMS classification number, primary 6207,
secondary 62P99, 62N05
9. Clark BJ, Thambiratnam DP, Perera NJ (2006) Analytical and experimental investigation of
the behavior of a rollover protective structure. Struct Eng 84(1):29–34
10. Karlinski J, Rusinski E, Smolnicki T (2008) Protective structures for construction and mining
machine operator. Automat Construct 17:232–24
11. Thambiratnam DP, Clark BJ, Perera NJ (2008) Dynamic response of a rollover protective
structure. Comput Aided Civil Infrastruct Eng 23(6):448–464
Fluid Mechanics and Heat Transfer
Investigating the Route to Flutter
in a Pitch–Plunge Airfoil Subjected
to Combined Flow Fluctuations
Nivedhithaa Santhakumar, Divyangi Singh, and Venkatramani Jagadish
Abstract The dynamics and response a pitch–plunge airfoil with cubic hardening
nonlinearity in the pitch degree of freedom are investigated numerically. The aero-
dynamics is assumed to be linear and modeled using the unsteady aero-dynamical
formulation. The flow is fluctuating in both the longitudinal and vertical directions.
The fluctuating flow is mathematically modeled as a long time-scale random process.
The mean flow speed is used as the bifurcation parameter, and response analysis is
carried out by systematically varying the bifurcation parameter. The route to flut-
ter is presented, and the role of noisy flow fluctuations in the same is analyzed by
examining the dependency on noise intensity.
Keywords Flutter · Intermittency · Vertical turbulence · P-bifurcation
1 Introduction
1.1 A Subsection Sample
Aeroelastic phenomena involve the dynamic interaction between the inertial, struc-
tural and aerodynamic forces of a fluid-structural system. In the case of a wing or any
control surface subjected to an incoming airstream, under certain flow conditions,
the structure experiences a dynamic instability known as flutter. Flutter is an unstable
self-feeding behavior in which the aerodynamic forces couple with the elastic and
inertial nature of the structure, resulting in sustained limit cycle oscillations (‘LCOs’)
beyond a particular critical value of mean flow speed of wind. The appearance of
LCOs in aeroelastic systems is considered undesirable, as it can induce fatigue dam-
age in the structural components and affect the structural integrity, posing a threat to
the structural safety. Consequently, investigating the onset and route to flutter attracts
considerable research attention. Methods for prediction and characterization of flutter
N. Santhakumar · D. Singh · V. Jagadish (B)

Department of Mechanical Engineering, Shiv Nadar University, Greater Noida, India

https://doi.org/10.1007/978-981-15-4488-0_32
370 N. Santhakumar et al.
behavior have existed for decades; however, the nonlinear nature of both the airfoil
structure and the aerodynamic environment continue to make this phenomenon dif-
ficult to model precisely. Worn hinges of surfaces and loose control linkages may be
the sources structural nonlinearities—which can be approximated by cubic, hystere-
sis or a bilinear nonlinearity. This paper focusses on the cubic structural nonlinearity
[1].
In case of uniform deterministic flow, onset of flutter is identified by birth of
LCOs. At a certain critical velocity, the system will undergo a change from damped
oscillations to LCOs, and the critical point becomes a supercritical Hopf bifurca-
tion point. In reality, the input flow comprises many irregularities which have the
capability to change the dynamical nature of the aeroelastic responses, and a Hopf
bifurcation analysis proves to be insufficient in this case. In-field situations involve
flow fluctuations both in the longitudinal and vertical direction. Further, vertical fluc-
tuations being more ubiquitous [2] act as an external forcing and can bring about
radical changes in the response dynamics [3].
The open problems this paper aims to tackle include: flutter prediction under tur-
bulent flow conditions and resolution cum interpretation of the dynamical response
signatures obtained, with emphasis on ascertaining the role of fluctuations in the verti-
cal direction with respect to the airfoil. Vertical turbulence is a pervasive phenomenon
[3] whose role in airfoil response dynamics and flutter prediction is unresolved, with
studies [4, 5] on combined turbulence providing only transitional changes in the
topology of the joint probability density function of the state variables.
2 Problem Description
2.1 Aeroelastic Model
A two-degree-of-freedom airfoil oscillating in pitch ‘α’ and heave ‘y’ is considered—

which are modeled by rotational and translational springs that are attached to the
elastic axis at a point. The choice of a bending–torsion model comes from experiments
conducted on a cantilever wing that shows flutter oscillations primarily along the
flexural and torsional modes of the structure. The structural equations are derived
from Fung [1] (Fig. 1).
mxα bα̈ + m ÿ + K h y = −L(t) (1)
I E A α̈ + mxα b ÿ + K α α + K 3 α 3 = M E A (t) (2)
In the above equations, U is the speed of the oncoming wind; bxα is the distance
from the elastic axis to the airfoil center of mass; bah is the semi-chord of the airfoil;
IEA is the moment of inertia about the elastic axis; K h and K α are the linear plunge and
pitch spring stiffnesses, respectively, and K 3 is the coefficient of nonlinear plunge
Investigating the Route to Flutter in a Pitch-Plunge … 371
Fig. 1 Schematic of airfoil
stiffness. L(t) and MEA (t) are the unsteady lift and moment generated due to the
aerodynamic effects, with τ representing non-dimensional time.
The unsteady aerodynamic load, owing to the longitudinal component, assum-
ing incompressible, inviscid flow, is estimated via Duhamel’s integral and the two-
state representation of Wagner’s function, which accounts for the wake structure
developing behind the airfoil.
L ϕ (t) = πρb2 [ ÿ + U α̇ − bah α̈]

t dϕ(t − s)
+ 2πρbU ω 43 ϕ(0) − ∫ ω 43 ϕ(s) ds (3)
0 ds

1
Meaϕ (t) = πρb2 bah ÿ − b[0.5 − ah ]U α̇ − b2 ah2 + α̈
8

t dϕ(t − s)
+ 2πρb2 [0.5 + ah ]U ω 34 ϕ(0) − ∫ ω 34 ϕ(s) ds (4)
0 ds
where
ω 34 = ẏ + U α + b[0.5 − ah ]α̇ (5)
ϕ(τ ) = 1 − 1 e− 1τ
− 2 e− 2τ
(6)
The constants 1 = 0.165, 2 = 0.335, ε1 = 0.0455 and ε2 = 0.3 are obtained

from Jones [7].
Vertical turbulence also contributes to the aerodynamic force and moment. The
unsteady effects are expressed in terms of Kussner’s function which accounts for
compounded circulation as a gust hails the leading edge of the airfoil.

t dψ(t − s)
L ψ (t) = 2πρbU ωt ψ(0) − ∫ ωt ψ(s) ds (7)
0 ds

t dψ(t − s)
Meaψ (t) = 2πρb2 [0.5 + ah ]U ωt ψ(0) − ∫ ωt ψ(s) ds (8)
0 ds
ψ(τ ) = 1 − 0.5792e−0.1393τ − 0.4208e−1.802τ (9)
The equations of motion are obtained upon the combination of Eqs. 7 and 8 with
Eqs. 3 and 4, while the pitch angle and the angle of attack are taken to be the same.
Four additional states, two lag terms in Wagner’s function and two lag terms in Kuss-
ner’s function, are introduced for the transformation of the above integro-differential
equations into differential form. The transformation facilitates the solution proce-
dure while allowing the delineation of the model into components of a traditional
mechanical system consisting of inertia, damping, stiffness and external forcing
terms. The non-dimensional aeroelastic equations of motion in the differential form
can be obtained from [3].
2.2 Flow Model
The flow speed U is assumed to comprise of a mean flow Um superimposed with

a randomly longitudinally fluctuating flow component f (τ ), where τ is the non-
dimensional time. A simple model [6] for the input flow fluctuation is utilized:
U = Um + f (τ ) (10)
f (τ ) = Um σ1 sin(ωr τ ) (11)
ωr (τ ) = ωo + κ R(τ ) (12)
The uniform flow is superimposed with a sinusoidal component whose frequency

of oscillation varies with time about a dominant frequency. Um is the mean wind
speed, σ indicates the amplitude of the fluctuating component and ωr (t) is the fre-
quency of the sinusoid. Here, κ is a constant and R is a uniformly distributed random
variable that takes values in [0, 1] at each instant of time.
The function g(τ ) accounts for fluctuation in the vertical direction. The fluctuating
flow is mathematically modeled as a long time-scale random process.
g(τ ) = Um σ2 sin(ωr τ ) (13)

The eighth-order aeroelastic system is then expressed in state-space form and

solved numerically. A fourth-order Runge–Kutta scheme is used to obtain the time
responses.
The cases presented are for an airfoil with the following non-dimensional airfoil
parameters:
ωy
k3 = 400; = 0.6325; xα = 0.25; rα = 0.5; ah = −0.5; μ = 100
ωα
For the given set of parameters, the deterministic flutter speed Um is 4.3 [3].
3.1 Vertical Turbulence
A three-stage analysis is carried out for analyzing the effect of vertical turbulence
on the response of the airfoil.
Firstly, in addition to longitudinal turbulence, an impulse forcing in the vertical
direction is introduced, to isolate the response corresponding to the impulse. Further,
the airfoil response is interpreted by means of a stochastic bifurcation analysis in
case of fluctuating flows. The P-bifurcation (J-PDF) is concerned with a qualitative
topological change in the probability structure of the dynamic behavior as a control
parameter is varied. Shown below are some of the pitch time histories and J-PDFs.
The plunge response time histories are qualitatively similar to the pitch responses
(Figs. 2 and 3).
Low-amplitude noisy response is observed to grow in amplitude (intermittency)
and give rise to random LCOs. The shape of the J-PDF changes from a bell-shaped
bi-dimensional one to a crater shape. These results are in accordance with those
obtained for a longitudinal turbulence analysis [6].
For the next step, longitudinal noise intensity was varied for a fixed vertical noise
intensity, to gauge the effect of variation in longitudinal turbulence in the presence
of a constant vertical fluctuation (Figs. 4, 5 and 6).
As mean flow speed is increased, a response typical of LCOs is not observed.
However, a high-frequency noisy response is noticed, with greater prominence of
high amplitude oscillations alongside bursts of lower amplitude oscillations. The
corresponding J-PDF shows a crater with two peaks, with turbulence distorting the
structure of the LCOs. However, upon variation of intensity, demarcation of responses
into separate regimes is not apparent through qualitative analysis involving PDFs.
Further, intensities of fluctuations in both directions are varied, to capture
responses and simulate real-time in-field conditions (Figs. 7 and 8).
A decaying response with noisy intermittent bursts is observed in the response for
Um = 3.5 and magnitude of noise intensity in either direction being 0.1. The J-PDF
for this case shows a collection of peaks. An LCO-like response is obtained for
Fig. 2 a Pitch response and b J-PDF for Um = 4.5 at σ1 = 0.5
Um = 8.5 with longitudinal and vertical noise intensities being 0.3 and 0.1,
respectively. The corresponding J-PDF shows a crater with two peaks.
4 Conclusion
A study of airfoil response under combined fluctuations has been carried out for a two-
degree-of-freedom pitch–plunge airfoil with respect to the mean speed of the flow and
the noise intensities. The mean flow speed of the longitudinal and vertical components
is used as the bifurcation parameter. A three-step analysis has been undertaken,
consisting of time histories and J-PDFs. It is observed that the pre-flutter response
displays a distinct noisy signature akin to an intermittent behavior. This observation
is substantiated from the J-PDFs as well. Further, with increase in the noise intensity,
the responses display an uncertain transition from one basin of attraction to another.
However, the specific role of longitudinal or vertical component’s contribution toward
this behavior could not be identified. The authors will take up this interesting problem
in a subsequent study.
Fig. 3 a Pitch response and b J-PDF for Um = 6 at σ1 = 0.1
Fig. 4 a Pitch response and b J-PDF for Um = 4.5 at σ1 = 0.6 and σ2 = 0.2
References
1. Lee B, Jiang L, Wong Y (1999) Flutter of an airfoil with a cubic restoring force. J Fluids Struct
13:75–101
2. Fung YC (1993) An introduction to the theory of aeroelasticity. Dover
3. Hoblit FM (1988) Gust loads on aircraft: concepts and applications. Am Inst Aeronaut Astronaut
4. Poirel D, Price SJ (2007) Bifurcation characteristics of a two-dimensional structurally non-linear
airfoil in turbulent flow. Nonlin Dyn 48(4):423–435
5. Zhao D, Zhang Q, Tan Y (2009) Random flutter of a 2-DOF nonlinear airfoil in pitch and plunge
with freeplay in pitch. Nonlin Dyn 58(4):643
6. Venkatramani J (2017) Intermittency in pitch-plunge aeroelastic system. Doctoral dissertation
7. Jones RT (1940) The unsteady lift of a wing of finite aspect ratio. NACA Reports 681
A Computational Study on the Effect
of Supercritical CO2 in a Combustor
Vishnumaya Nair and Selvaraj Prabhu
Abstract The ability of supercritical carbon dioxide (sCO2 ) oxy-combustion power

cycles like Allam cycle to capture 99% of the carbon produced in combustion started
gaining attraction nowadays. The Allam cycle works at extreme pressure conditions
(200–300 bar), and it substantially increases the efficiency. To promote and accelerate
the growth and development of combustors for such cycles, a study on kinetic models
using sCO2 as diluent has to be conducted by exposing it to extreme operating
conditions such as higher pressures (300 bar) and high inlet temperatures (1000 K)
along with different levels of CO2 dilution. The paper focuses on identifying the effect
of supercritical CO2 in a combustor where the fuel used is methane by analyzing
its impact on parameters such as temperature and ignition delay time (IDT) for
the CH4 –O2 –CO2 mixture with variation in CO2 dilution. The software ANSYS
Chemkin-Pro is used for the initial simulations where two zero-dimensional models
are analyzed with the above said conditions of pressure and temperature for various
mixture ratios of CH4 –O2 –CO2 . Saudi Aramco 2.0 mechanism is used for Chemkin
simulations. Finally, computational fluid dynamics (CFD) simulation of combustor
is done in ANSYS Fluent by importing GRI-Mech 3.0 and the reduced Aramco 2.0
mechanisms and compared the results for different CO2 dilutions.
Keywords Allam cycle · CFD · GRI-Mech 3.0 · Saudi Aramco 2.0
1 Introduction
Supercritical CO2 (sCO2 ) cycles make use of a combustor which burns fuel and
pure oxidizer with the CO2 acting as a diluent and temperature moderator. sCO2
is used as the working fluid medium in supercritical carbon dioxide power cycle.
V. Nair · S. Prabhu (B)

Department of Mechanical Engineering, Amrita School of Engineering, Amrita Vishwa
Vidyapeetham, Coimbatore, India
V. Nair

https://doi.org/10.1007/978-981-15-4488-0_33
380 V. Nair and S. Prabhu
Delimont et al. [1] have discussed the geometry and boundary conditions likely to
be seen in an oxy-fuel combustor using sCO2 which are specifically relevant to the
upcoming 1 MWth combustor design and also the behavior of sCO2 swirl-stabilized
fuel injector is studied by analyzing the effect of various design variables such as
geometry of the chamber, injection, cooling schemes and mainstream flow angle. As
the growing demand for energy and the usage of fossil fuels alarm the world about
the irreversible climatic change due to global warming, there must be an action taken
in order to reduce and completely eliminate such critical situations, and one such
initiative can be in the form of reducing emissions which can be accomplished with
the support of supercritical CO2 combustors as a result of which Manikantachari
et al. [2] have performed detailed study on the reductions of mechanisms to be used
for sCO2 combustor simulations. Only very few data are available for combustors
operating close to the Allam cycle conditions, which is currently the precursor for
direct sCO2 power cycles, and hence, Strakey [3] has attempted to study the turbulent
length and time scales along with the turbulent flame speed with the help of CFD
and simplified 0D and 1D flame calculations. Liu et al. [4] have done a comparison
of auto-ignition delay measurement of CH4 –O2 –CO2 from the shock tube with the
numerical simulation. Full carbon capture together with the higher efficiencies is
inevitable for the successful operation of the plant. Zhu [5] has described the inno-
vative power generation systems using novel cycles like sCO2 power cycles which
have the potential for achieving the above said characteristics at lower costs and also
help in providing alternative power generation systems. Zhu [5] has also discussed
the turbomachinery, heat exchangers, combustors and materials to be used with the
sCO2 cycles.
The Allam cycle (high pressure, low pressure ratios, combustion of oxygen-fuel)
is one such new sCO2 cycle design including carbon capture and storage. It is a
highly recuperated sCO2 cycle with just a modest rise in the combustion engine
temperature as described by Zhu [5]. In this process, two combustion products formed
are, carbon dioxide and water, respectively, where an oxy-methane mixture is burned
using carbon dioxide as a diluent, in a direct-fired combustion plant. A combustion
operating pressure of 300 bar, an inlet temperature of 750 °C and an outlet temperature
of 1150 °C are the nominal conditions as defined by the Allam cycle. Pryor et al. [6]
have done the measurements of different ignition delay times and time histories of
methane species for methane-O2 mixtures at a high CO2 diluted environment using
shock tube and laser absorption spectroscopy at pressures between 6 and 31 atm. A
supercritical fluid has some liquid as well as some gas properties. The validation of the
chemical kinetic mechanism for supercritical CO2 (sCO2 ) oxy-methane combustion
that can be used for computational fluid dynamic code (CFD) simulations in oxy-
combustion development has been described by Vasuet al. [7]. Iwai et al. [8] have
carried out a 3-D CFD simulation of the combustor internal flow field in order to
obtain information on mixing of fuel oxidizer, position of the flame and heat transfer
near the wall. Ratna Kishore et al. [9, 10] have done different studies on laminar
burning velocities using different diluent concentrations. The CO2 peculiarities are
a relatively low critical pressure of 7.4 MPa and a critical temperature of 31 °C. As
a result, CO2 can be directly compressed to supercritical pressures and heated to
A Computational Study on the Effect of Supercritical … 381
supercritical conditions under moderate conditions. In its supercritical state, CO2 is

also approximately twice as dense as steam. The high density and volumetric heat
capacity of sCO2 make it more energy dense compared to other working fluids. A
smaller plant footprint can therefore be achieved with the reduction in size of all
components of the system such as the turbine and heat exchangers.
This research contributes to the growth and expansion of new technologies and
methods that are cleaner and also more cost effective. The work is done with CO2
as the primary diluent at higher pressures and temperatures which is not a much
explored area. This research will validate the Aramco mechanism for supercritical
conditions and extend the work for the combustor analysis.
1.1 Chemical Kinetic Mechanisms
The precise modeling of combustion processes requires a chemical kinetic mecha-

nism that is valid for the range of conditions under study. The mechanism considered
for this simulation is GRI-Mech Version 3.0 with 53 chemical species and 325 reac-
tions [11], which is used for modeling the natural gas combustion. A more recently
developed mechanism, referred to as Saudi Aramco 2.0, has also been considered for
the methane–air combustion mechanism. Saudi Aramco 2.0 mechanism consists of
493 species and 2714 reactions that can be used for fuel up to six carbon components.
The above mentioned mechanism can be used for natural gas combustion.
ANSYS Chemkin-Pro. Chemkin-Pro is used in modeling as well as simulating
complex gas phase and surface chemistry reactions and has been used for doing the
constant volume combustor and perfectly stirred reactor (PSR) simulations in order to
understand the changes in ignition delay time and combustion temperature with CO2
dilution in CH4 –O2 mixture at elevated temperatures and pressures. The variations in
IDTs and combustion temperatures are also noted with N2 as the diluent. The Aramco
2.0 mechanism is one of the most accurate mechanisms, and the validation against
a large array of experimental analysis has been performed for this mechanism, and
hence, it is employed for simulating the combustor with CH4 –O2 –CO2 mixture at
different mixture ratios.
2 Constant Volume Reactor
The mixture of all gas particles are assumed to be uniform and homogeneously mixed
inside the constant volume reactors. A closed 0-D constant volume batch reactor
and perfectly stirred reactor are used for the analysis. The ignition delay times of
a CH4 –O2 fuel mixture at high initial pressure for a series of initial temperatures
can be predicted using the 0-D closed reactor model. At the time of ignition, a
pronounced spike can be observed in the OH mole fraction. As a result, OH spikes
and the maximum temperature change rate (inflection point) determine the time for
Table 1 Effects of EOS on the IDTs at T = 1000 K and P = 300 atm

Equation of state (EOS) Ignition delay time, IDT (µs)
CH4 –O2 –CO2 = 7.5–15–77.5 H2 –O2 –CO2 = 10–5–85
Redlich–Kwong (RK) 130 155
Soave–Redlich–Kwong (SRK) 136 216
Peng–Robinson (PR) 132 183
the ignition. The equation of state (EOS), EOS_SRK, that is, Soave–Redlich–Kwong
EOS incorporated in the mechanism will be used in the simulations to determine the
P-T-V relationship of the gas mixture.
2.1 Chemkin Results
The simulations are carried out to understand the influence of EOS, Redlich–Kwong
(RK), Soave–Redlich–Kwong (SRK), Peng–Robinson (PR) in the methane and
hydrogen mixtures. Table 1 makes it clear that the impact of EOS on CH4 –O2 –CO2
is insignificant while affecting the IDTs of the H2 –O2 –CO2 mixture. A significant
effect of the EOS can be observed on the H2 mixture but not on the CH4 mixture.
H2 O is the main combustion product of the H2 mixture, while it is both CO2 and
H2 O for the CH4 mixture.
The critical CO2 pressure is approximately 73 atm, while the H2 O pressure is
approximately 218 atm, which is about three times higher than the CO2 pressure.
The higher amount of H2 O in the H2 mixture combustion product therefore increases
the resulting critical point of the mixture. Near the critical point, the EOSs largely
differ. The overall critical point is further reduced with the formation of CO2 in the
case of the CH4 mixture, which results in a smaller deviation of the EOSs. An ideal
gas behavior is exhibited by the real gases at elevated temperatures and pressures far
beyond the critical point since the compressibility factor is closer to unity at high
operating conditions.
A clear indication of the reduction in the ignition delay time with temperature can
be observed. Figure 1 shows the variation of ignition delay time with the inverse of the
temperature, hence showing an increment in ignition delay time. Two main kinetic
mechanisms used for the combustion of natural gas are considered here, the GRI-
Mech 3.0 and the recent Aramco-2.0. Figure 1 shows the difference in the predictions
of ignition delay times between the GRI-3.0 mechanism and the Aramco-2.0 at a
higher pressure level (300 atm). The GRI-Mech forecasts a higher IDT value than
the Aramco 2.0 mechanism for the CH4 –O2 –CO2 = 7.5–15–77.5 mixture at P =
300 atm and temperature T between 1000 and 1080 K.
From Fig. 2, it is understood that the mixture with the higher dilution of CO2 is
having a larger value for ignition delay time. Ignition delay generally decreases with
increasing preheat temperature as shown in Figs. 2 and 3 which show a comparison
0.01
IDT(ARAMCO 2.0)
IDT(GRI 3.0)
IDT (S)
0.001
0.0001
0.92 0.93 0.94 0.95 0.96 0.97 0.98 0.99 1
1000/K (1/K)
Fig. 1 Comparison of GRI-Mech 3.0 and Aramco 2.0 mechanisms for the mixture CH4 –O2 –CO2
= 7.5–15–77.5
0.01
IDT(86.17% CO2)
IDT(77.5% CO2)
IDT (S)
0.001
0.0001
0.82 0.84 0.86 0.88 0.9 0.92 0.94 0.96 0.98 1 1.02
1000/K (1/K)
Fig. 2 Comparison of IDTs for the mixtures CH4 –O2 –CO2 = 3.91–9.92–86.17 and CH4 –O2 –CO2
= 7.5–15–77.5
3500
TEMPERATURE (K)
3000
2500
2000
1500 71.5%N2
71.5%CO2
1000
500
0 0.001 0.002 0.003 0.004 0.005
IDT (s)
Fig. 3 Temperature rise for 71.5% N2 and CO2 diluted mixtures
of ignition delay times for stoichiometric methane–oxygen combustion at different

concentrations of CO2 . These calculations are carried out using the Saudi Aramco
2.0 mechanism, which has been validated at a much higher pressure than the GRI
3.0.
Figure 3 represents the comparison of the temperature rise in the stoichiometric

CH4 –O2 mixture at 71.5% CO2 and N2 dilution levels at 1000 K for the mixture 3 as
indicated in Table 2. The figure clearly demonstrates the difference in temperatures
when diluted with CO2 and N2 . The temperature increase due to combustion process
with N2 as the diluent is more when compared to CO2 diluted combustion. The
maximum temperature attained for those mixtures diluted with CO2 is around 2500
K and the mixtures diluted with N2 is around 3100 K. Hence, it shows that CO2 has
the ability to withstand more temperature than N2 . The temperature inflection point
as seen in the figure is somewhat delayed in the case of CO2 diluted combustion as
a result of which ignition is delayed in CO2 diluted combustion as compared to N2
diluted combustion process.
Figure 4 shows that the increment in IDT is comparatively high with CO2 dilution
compared to that with nitrogen as the diluent. Carbon dioxide slows the ignition
process slightly more.
From Table 3, it is evident that as the temperature increases, the ignition delay
time reduces as a result of the increased reactivity of the mixtures. The IDTs of the
mixture for various N2 and CO2 dilutions have been calculated. The mole fractions
of CH4 and O2 in the simulations are taken to be XCH4 = 0.095 and XO2 = 0.19.
The EOS SRK is used for the simulations.
Table 2 Stoichiometric CH4 –O2 mixture at various CO2 dilution levels

Mixtures Mole fractions
X CH4 X O2 X CO2 / X N2
1 0.161 0.322 0.517
2 0.134 0.268 0.598
3 0.095 0.19 0.715
4 0.056 0.112 0.832
0.0025
IDT_CO2
0.002 IDT_N2
ɸ=1
IDT (S)
0.0015
P=300ATM
0.001
0.0005
0
50% 55% 60% 65% 70% 75% 80% 85%
% OF CO2 (OR N2)
Fig. 4 Comparison of IDTs at different CO2 /N2 dilution levels for the stoichiometric mixtures
Table 3 Ignition delay time simulation for various N2 and CO2 dilutions at P = 300 atm
T (K) X N2 XCO2 IDT (µs)
1000 0.715 0 878
1050 407
1000 0.465 0.25 913
1050 434
1000 0.215 0.5 979
1050 459
PSR. Continuously stirred tank reactor (CSTR) is often referred to as a PSR or

perfectly stirred reactor in the literature on chemical engineering. This type of sim-
ulation, a gas turbine combustion tool since the 1950s, is a zero-dimensional one.
Primary combustion zone can be simulated using this. The reactor inlet temperature
chosen for simulations is 1000 K, while the pressure is 300 atm. Using a PSR model
in Chemkin-Pro, various possible CO2 dilution levels for stoichiometric methane–
oxygen mixture in the primary zone are simulated. In the mechanism, SRK EOS
is considered for simulation. 1000 K and 300 atm are the inlet temperature of the
reactor and the pressure taken into consideration. Figure 5 indicates the reduction in
the exit temperature with increase in CO2 dilution levels.
Mechanism reduction. The powerful tools for the numerical simulation in order to
study complex turbulent reacting flows are the detailed mechanisms. The reduction
of such mechanisms is a necessary step because of the huge computational cost and
time associated with the detailed one. Here, direct relation graph (DRG) with error
propagation (DRGEP) method is used to reduce detailed Aramco mechanism for
importing it in ANSYS Fluent. Using this method, a skeletal mechanism with 56
species is generated for 1000 K and 300 bar from the detailed one consisting of 493
species.
The reduced 56 species Aramco mechanism is compared with the detailed mech-
anism, and the reduced mechanism is in good agreement with the detailed as shown
PSR EXIT TEMPERATURE
3000
51.7%CO2 ɸ=1,
2500 59.8%CO2 1 0 0 0 K , 3 0 0 A T M
71.5%CO2
2000 77.5%CO2
(K)
1500
1000
500
1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3
PSR NO.
Fig. 5 Comparison of the exit temperature of a PSR at various CO2 dilution levels
0.0016
0.0014 IDT_REDUCED
0.0012 IDT_DETAILED
0.001
IDT (S)
0.0008
0.0006
0.0004
0.0002
0
0.8 0.85 0.9 0.95 1 1.05
1000/K (1/K)
Fig. 6 Comparison of IDTs obtained by detailed Aramco 2.0 with the reduced mechanism
in Fig. 6. Hence, the reduced mechanism can be imported to fluent for carrying out
the simulations without much computational time and cost.
3 CFD Solver Setting
A simple 2-D combustor is modeled in modeled and mesh independent study is

completed. Boundary conditions include CH4 and O2 /CO2 at the inlet and an outlet
pressure condition as shown in Fig. 7.
The meshing operation is performed by means of edge sizing by choosing the
uniform size function, and face sizing is done with quadrilaterals as the geometry is
not a complicated one. The total number of nodes obtained with this mesh setting is
12,261 while that of elements is 12,000, and default value has been chosen for the
growth rate and is kept as 1.2, and the growth rate represents how the cell volumes
grow gradually for capturing the flow physics. The standard k-epsilon model, one of
the most efficient models, has been used for the turbulent flow. The eddy dissipation
Fig. 7 A 2-D combustor model

concept is chosen for the turbulence chemistry interaction along with the Chemkin
CFD solver in the species transport model. The SIMPLE scheme is used along with
standard hybridization. The simulations are carried out at the boundaries conditions
as follows: pressure = 300 bar, O2 /CO2 inlet temperature = 1000 K, CH4 inlet
temperature = 320 K, respectively. The equations for the viscous fluid flow are as
follows:
∂ρ
+ ∇ · (ρV ) = 0 (1)
∂t
∂(ρV )
+ ∇ · (ρV V ) = ρ f + ∇ · τi j − ∇ p (2)
∂t

∂ V2 V2
ρ e+ +∇ · ρ e+ V = ρ q̇ + ∇(K ∇T )
∂t 2 2
− p∇V + ρ f · V + ϕ (3)
∂
(ρYi ) + ∇ · (ρ vYi ) = −∇ · Ji + Ri + Si (4)
∂t
The above Eqs. (1)–(4) are known as the continuity, momentum, energy and
species transport equations, while the equations corresponding to the standard k-ε
model are as follows;

∂(ρk) ∂(ρku i ) ∂ μt ∂k
+ = + 2μt E i j E i j − ρε (5)
∂t ∂ xi ∂ x j ρk ∂ x j

∂(ρε) ∂(ρεu i ) ∂ μt ∂ε ε ε2
+ = + C1ε 2μt E i j E i j − C2ε ρ (6)
∂t ∂ xi ∂ x j ρk ∂ x j k k
where (5) and (6) are the equations for turbulent kinetic energy k and dissipation ε.
4 Results
By importing the GRI-Mech 3.0 and reduced 56 species Aramco 2.0 mechanisms, the
temperatures for the cases with 50%, 60%, 70% and 80% CO2 dilution are compared.
Both mechanisms predicted similar results, and the temperature values for 50%, 60%,
70% and 80% CO2 dilutions are 2680 K, 2540 K, 2320 K and 1910 K, respectively,
as shown in Figs. 8, 9, 10 and 11. The observations clearly indicate a temperature
reduction of 5.22%, 13.4%, 28.73% for 60%, 70% and 80% CO2 dilutions with
respect to 50% dilution. This reduction in temperatures is mainly due to the fact that
CO2 has the ability to withstand more temperature, and it facilitates combustion in
such a way that it takes place in a mild manner. The increment in CO2 concentration
reduces the concentration of the reactants and has the power to lower the net reaction
Fig. 8 Temperature contours for 50% CO2 dilution using GRI 3.0 and reduced Aramco 2.0
mechanisms
mechanisms
rate. This ability of CO2 dilution to reduce temperature is more as compared to other
diluents such as nitrogen and has been proved from the Chemkin results as shown
in Fig. 3. Moreover, CO2 acts as a heat sink which gives rise to reduction in reaction
temperature.
mechanisms
mechanisms
5 Conclusion and Future Work
In this paper, the effect of CO2 dilution on methane combustion is found out by means
of Chemkin simulations where various equations of states are used to determine the
ignition delay times including comparisons conducted between the detailed Saudi
Aramco 2.0 and GRI-Mech 3.0 mechanisms. The GRI-Mech 3.0 overpredicts the
ignition delay time in comparison with the Saudi Aramco 2.0 mechanism at higher
pressures. Saudi Aramco 2.0 mechanism is found be a more accurate one for the
combustion of sCO2 . The ignition delay times of supercritical oxy-methane-CO2
mixtures is less influenced by the equation of state (EOS) in contrast to that of the
supercritical oxy-hydrogen-CO2 mixture.
The ignition delay time data have been found out at higher pressures and tem-
peratures (1000–1200 K) for the oxy-fuel combustion at different carbon dioxide
dilution levels. Chemkin simulations show that as temperature is made to increase,
the ignition delay times are decreased, whereas ignition is delayed as CO2 dilution
is increased. Compared to the IDTs with N2 dilution, CO2 dilution has been found
to result in more increment of ignition delay times. This delayed ignition is mainly
caused by the chemical effect of CO2 . A substantial reduction in the temperature can
be observed with CO2 dilution. It retards the ignition and slows the overall response
rate, making it easier for moderate or intense low-oxygen dilution combustion to
occur. The reduction in temperature is due to the physical effect of the CO2 dilution.
In addition to the kinetic effects, flow effects are also considered in CFD analy-
sis. The reduced Aramco 2.0 mechanism is used in the simulations, and the results
are compared with GRI 3.0 mechanism. The change in the peak temperature with
different dilution limits of CO2 is observed at high pressure conditions.
Further to understand the effect of supercritical conditions, a detailed investigation
of combustion of methane with large amounts of CO2 dilution has to be conducted.
Future work would include the 3-D simulation of a more refined model of combustor
by incorporating the real gas models at supercritical conditions.
References
1. Delimont J, Andrews N, Chordia L (2018) Exploration of combustor design for direct fired
oxy-fuel application in a sCO2 power cycle. In: Proceedings of Montreal, Global Power and
Propulsion Forum, GPPS-NA-2018-0157
2. Manikantachari KRV, Vesely L, Martin S, Bobren-Diaz JO, Vasu S (2018) Reduced chemi-
cal mechanisms for oxy/methane supercritical CO2 combustor simulations. J Energy Resour
Technol 140/092202-1
3. Strakey PA (2018) Oxy-combustion flame fundamentals for supercritical CO2 power cycles.
In: The 6th international supercritical CO2 power cycles symposium, Pittsburgh, Pennsylvania
4. Liu Z, Gao X, Karimi M, Ochs B, Acharya V, Sun W, Delimont J, Andrews N (2018) Simula-
tion of sCO2 oxy-combustion using reduced chemical kinetic mechanism: effects of reduced
mechanism and sensitivity to mechanism parameters. In: The 6th international supercritical
CO2 power cycles symposium, Pittsburgh, Pennsylvania
5. Zhu Q (2017) Innovative power generation systems using supercritical CO2 cycles. Clean
Energy 1(1):68–79
6. Pryor O, Barak S, Lopez J, Ninnemann E, Koroglu B, Nash L, Vasu S (2017) High pressure
shock tube ignition delay time measurements during oxy-methane combustion with high levels
of CO2 dilution. J Energy Resour Technol
7. Vasu SS, Pryor OM, Kapat JS, Masunov A, Martin SM (2016) Developing a validated
chemical kinetics model for sCO2 combustion and implementation. In: The 5th international
symposium—supercritical CO2 power cycles, San Antonio, Texas
8. Iwani Y, Itoh M, Morisawa Y, Suzuki S, Cusano D, Harris M (2015) Development approach to
the combustor of gas turbine for oxyfuel, supercritical cycle. In: Proceedings of ASME Turbo
Expo, turbine technical conference and exposition, Montréal, Canada, GT2015
9. Ratna Kishore V, Ravi MR, Ray A (2011) Adiabatic burning velocity and cellular flame
characteristics of H2 -O2 -CO2 -air mixtures. Combust Flame 158
10. Ratna Kishore V, Muchahary R, Ray A, Ravi MR (2009) Adiabatic burning velocity of H2 -O2
mixtures diluted with CO2 /N2 /Ar. Int J Hydrogen Energy 34
11. GRI-Mech homepage, Gas Research Institute, Chicago (1999). www.me.berkeley.edu/gri_
mech/
Design of Shell-and-Tube Heat
Exchanger with CFD Analysis
B. Jayachandraiah and C. Dinesh Kumar Patel
Abstract Heat exchanger is a device used to transfer heat between two fluids which
are at different temperatures. An attempt is made in this paper. A shell-and-tube
heat exchanger (STHE) is modeled by using CATIA V5 R20 software with TEMA
standards. Meshing and analysis are done by using Autodesk CFD Simulation com-
mercial software. The STHE is considered for two different fluids of kerosene and
crude oil with inlet temperature and inlet velocity of fluids as boundary conditions.
The results obtained are contours of velocity magnitude, temperature distribution and
pressure drop of the fluid. Then, results are plotted for the temperature distribution,
pressure drop and fluid velocity profiles. The simulation results are compared with
the analytical results and find good agreement with the analytical results.
Keywords STHE · CFD · Simulation
1 Introduction
Heat exchanger is a device specially designed for efficient transformation of heat from
one fluid to another fluid over solid surface. In heat exchangers, the temperature of
each fluid changes as it passes through the exchangers, and hence, the temperature
of dividing wall between the fluids also changes along the length of Exchanger.
The most widely applied heat exchangers are those constructed of a “shell” which
contains one of the fluids as well as the “tubes” which contain the other fluid.
The shell-and-tube-type heat exchanger allows a great deal of flexibility in design
applications and as a result is frequently found in heavy-duty applications (Fig. 1).
B. Jayachandraiah (B) · C. Dinesh Kumar Patel

SKIT, Srikalahasti, India

https://doi.org/10.1007/978-981-15-4488-0_34
394 B. Jayachandraiah and C. Dinesh Kumar Patel
Fig. 1 Double-pass
shell-and-tube heat
exchanger
2 Literature Survey
Caputo et al. [1] have done a procedure for optimal design of shell-and-tube heat
exchangers, which utilizes a genetic algorithm to minimize the total cost of the
equipment including capital investment and the sum of discounted annual energy
expenditures related to pumping. Ebieto and Eke [2] have done the performance of
shell-and-tube heat exchangers by analytical method, and the program was written in
MATLAB to check for the thermal and hydraulics suitability of the heat exchangers.
Chit et al. [3] have done the flow analysis in shell side on the effect of baffle spac-
ing of shell-and-tube heat exchanger which has been studied using theoretical and
numerical methods. The shell-side pressure drop for acceptable limits is 0.3 bars for
shell-and-tube heat exchanger of oil cooler of locomotive. Pavani Nookaratnam and
Dharma Raju [4] have studied material properties of various nanofluids that are cal-
culated at different volume fractions and compared. Performance of these nanofluids
is evaluated using computational fluid dynamics with the help of ANSYS Fluent
module. Irshad et al. [5] have done analyses on the performance of shell-and-tube
heat exchanger by comparing several shell-and-tube heat exchangers with segmen-
tal baffles. Mukadam et al. [6] have studied the performance of the shell-and-tube
heat exchanger and the effect of various parameters such as pitch layout and baffle
spacing.
2.1 Objective
An attempt is made in this paper to do the design of shell-and-tube heat exchanger by

modeling in CATIA V5 R20 and meshed using Autodesk meshing software. The anal-
ysis is done by Autodesk CFD 2015 software by using computational fluid dynamics
(CFD) and to find the various contours of temperature, pressure and velocity across
the double-pass shell-and-tube heat exchanger by evaluating the performance of heat
exchanger.
Design of Shell-and-Tube Heat Exchanger with CFD Analysis 395
Table 1 Design parameters

S. No. Description Unit Value
for heat exchanger
1 Heat exchanger length (L) mm 1000
2 Shell inner diameter (Di ) mm 100
3 Shell outer diameter (d o ) mm 106
4 Tube length (l) mm 1000
6 Tube outer diameter (d o ) mm 19.06
7 Number of tubes (N t ) – 5
8 Tube pitch triangular (Pt ) mm 25
9 Baffle cut portion % 25
10 Baffle spacing (B) mm 85
11 Side plate diameter (Dsp ) mm 100
12 Side plate thickness (T sp ) mm 3
13 Baffle thickness (BT ) mm 3
3.1 Materials
The material used for shell is stainless steel (304), for tubes is copper and for baffles
and tube sheet is copper.
3.2 Design Parameters
Design parameters considered for design of heat exchanger are shown in Table 1.
3.3 Fluid Properties for Kerosene and Crude Oil
The properties of kerosene and crude oil are shown in Table 2.
Inlet temperature of the hot fluid at shell side: 950 °C

Inlet temperature of the cold fluid at the tube side: 250 °C
Mass flow rate at shell side: 10 kg/s
Mass flow rate at tube side: 18.8 kg/s
Table 2 Properties of kerosene and crude oil

Mass T ρ Cp μ (Pa s) k Rf (m2
flow input (kg/m3 ) (kJ/kg K) (W/mK) K/W)
(kg/s) (°C)
Shell side: 10 95 850 2.47 0.00040 0.13 0.00061
kerosene
Tube side: 18.8 25 995 2.05 0.00358 0.13 0.00061
crude oil
Outlet pressure at the shell side: zero gauge pressure

Outlet pressure at the tube side: zero gauge pressure.
3.5 Mathematical Analysis
Maximum heat transfer Qmax = C min (T hi − T ci ) = 17,29,000 W

Heat transfer area (A) = L*P*do*Nt = 0.598473 m2 .
3.5.1 Tube-Side Calculations
Tube velocity (vt ) = 19.401 m/s

Reynolds number (Ret ) = 84,911
Prandtl number (Pr) = 56.4538
Convective heat transfer coefficient (ht ) = 7528 W/m2 K.
3.5.2 Shell-Side Calculations
Equivalent diameter (d e ) = 0.0169 m

Area of shell (As ) = 0.0020 m2
Shell velocity (vs ) = 5.882 m/s
Reynolds number (Res) = 21,123.7
Prandtl number Prs = 7.6
Convective heat transfer coefficient (hs ) = 4687.549 w/m2 K
Overall heat transfer coefficient U = 581.621 W/m2 K
NTU = (UA)/C min = 0.01409
Effectiveness ε = 0.5497 = 54.97%
Actual heat transfer = Qact = ε * Qmax = 949,221 W
Q = 24,700(95 − T ho )
T ho = 56.57 °C
Fig. 2 Design and meshing
Q = 38,540(T co − 25)
T co = 49.62 °C.
4 Modeling and Meshing
Computer-Aided Three-Dimensional Interactive Application (CATIA V5) was

developed by Dassault systems from France which is used for the modeling of the
heat exchanger. The model is imported to the Autodesk Simulation CFD and filled
the fluid volume. Prior to running an Autodesk Simulation CFD analysis, the geom-
etry is broken up into small pieces called elements. The corner of each element is a
node. The calculation is performed at the nodes. These elements and nodes make up
the mesh. The model and meshing of shell, tubes and baffles are shown in Fig. 2.
Temperature distribution of shell, tube and baffles is shown in Fig. 3. The temperature
at the inlet of shell-side hot fluid is higher at the entrance and up to 5 and gradually
losses its heat to the cold fluid.
The movement of fluid in the form of traces is shown in Fig. 4. The fluid movement
trace for the shell side is one pass and for the tube is two passes. The wireframe for
the temperature distribution is maximum at shell inlet is 95 °C.
Figure 5 shows the temperature distribution at each baffle. The baffle near to the
Fig. 3 Temperature distribution of fluids
Fig. 4 Trace lines for temperature distribution
Fig. 5 Temperature distribution in baffles
shell-side fluid entrance got more heat and next to the first baffle also got some high
temperatures.
Figure 6 shows velocity and pressure distribution in shell and tube. Throughout
the shell, the velocity distribution is constant as 5.882 m/s and in the tube is also
constant as 19.401 m/s. The pressure drop in the shell is 818,462.0 N/m2 and at the
tube is 1,417,310.0 N/m2 .
Graph 1 shows there is a continuous decrease in the shell-side fluid temperature.
The cold water temperature gradually increased up to 76 °C in the first tube pass and
decreased to the 47.45 °C in the second pass.
Fig. 6 Velocity and pressure distribution in shell and tube
Graph 1 Temperature 100 Kerosene-Crude oil

temperature
distribution of kerosene and

crude oil
50
0
0 500 1000 1500
distance
shell temperature tube in tube out
6 Conclusion
A novel shell-and-tube heat exchanger is designed in CATIA V5 and analyzed using

Autodesk Simulation CFD 2015. The elements are 380,540, and the nodes obtained
are 152,444. Standard heat exchanger design procedure is considered for the mathe-
matical calculations. The outer temperature of the shell-side fluid is 56.57 °C which
is approximately equal to the simulation results. The outer temperature of the tube-
side fluid is 49.62 °C which is also nearly equal to the simulation result. Graphs
show the clear temperature distribution in the shell-and-tube heat exchanger for the
given geometry, and the compared theoretical results show good agreement with
each other.
References
1. Caputo AC, Pelagagge PM, Salini P (2008) Heat exchanger design based on economic
optimization. Appl Therm Eng 28:1151–1159
2. Ebieto CE, Eke GB (2012) Performance analysis of shell and tube heat exchangers: a case study.
J Emerg Trends Eng Appl Sci 3(5):899–903
3. Chit SP, San NA, Soe MM (2015) Flow analysis in shell side on the effect of baffle spacing of
shell and tube heat exchanger. Int J Sci Technol Soc 3(5):254–259
4. Pavani Nookaratnam D, Dharma Raju T (2017) Analysis of shell and tube heat exchanger using
different nano fluids. Int J Res Appl Sci Eng Technol (IJRASET), 2321–9653
5. Irshad M, Kaushar M, Rajmohan G (2017) Design and CFD analysis of shell and tube heat
exchanger. Int J Eng Sci Comput, 6453–6457
6. Mukadam SS, Naik OV, Naik PS, Naik PP, Yekane AS (2018) Temperature analysis of shell and
tube heat exchanger. Int Res J Eng Technol (IRJET), 3264–3269
Characterization of Rayleigh–Taylor
Instability at the Fluid–Fluid Interface
Nihal Raj and S. Karthick
Abstract Two fluids of different densities superposed one over the other or accel-
erated toward each other develop instability at the plane interface between the two
fluids. This instability known as Rayleigh–Taylor instability (R-T) plays an integral
role in fluid atomization, metal liner electromagnetic implosion, inertial confinement
fusion, plasma fusion reactors and deuterium–tritium fusion target laser implosion.
In the secondary stages of atomization, R-T instability dictates the quality of spray
influencing the performance of engines. The objective of this study is to character-
ize the interfacial instability at the fluid–fluid interface based on growth rate and
the wavenumber of the disturbances generated at the interface. The stability of the
fluid–fluid interface under the influence of inertial, surface tensional and rotational
forces has been investigated. The test fluids taken for this study include standard flu-
ids, namely water and air, forming interfaces with commercially used fluids, namely
nitromethane, ethylene glycol, gasoline and diesel. An attempt has been made to
understand the factors contributing to instability of the fluid–fluid interface assumed
to be inviscid based on the growth rate and critical wavenumber of the generated
disturbances.
Keywords Rayleigh–Taylor instability · Interfacial instability · Critical

wavenumber
1 Introduction
Instability of fluids indicates the property of not being stable, balanced or predictable.
A simple example can be a ball on top of the hill; if given a small push, the ball will
N. Raj
Student, Department of Mechanical Engineering, Amrita School of Engineering, Coimbatore,
Amrita Vishwa Vidyapeetham, Coimbatore 641112, India
S. Karthick (B)
Faculty, Department of Mechanical Engineering, Amrita School of Engineering, Coimbatore,
Amrita Vishwa Vidyapeetham, Coimbatore 641112, India

https://doi.org/10.1007/978-981-15-4488-0_35
402 N. Raj and S. Karthick
go and continue to go down the hill while gaining speed in the process. The ball
that finally reaches downhill oscillates about its initial position until friction brings
it to rest. In addition, when a spherical balloon is filled with gas which is heavier
than air normally carbon dioxide, the balloon will tend to go down; on the contrary,
the balloon filled with helium will rise. Similarly, fluid instability in the form of
Rayleigh–Taylor occurs in liquids, gases and plasma. Rayleigh–Taylor instability
(RTI) results when two fluids of different densities are superposed or accelerated
toward each other. For example, water over oil in a container, denser water pushes the
oil up and develops an unstable condition resulting in RTI. This instability has been
observed during secondary atomization, a critical event that decides the atomization
process in combustion, spray drying and coating, drug delivery and evaporation-based
heat exchangers.
When the heavier fluid is displaced downward with an equal volume of lighter
fluid displacing upward, the potential energy of the fluid system is lower than the
initial state [1]. This results in the development of the perturbation at the interface
due to the established instability and subsequently results in the further release of
potential energy. In addition, under the influence of gravitational field, the heavier
fluid gets displaced downwards and the lighter fluid displaced upwards as observed
by Lord Rayleigh [1]. Taylor [2] realized that the fluid system of Lord Rayleigh cor-
responding to the state of fluids accelerating to each other with the less dense fluid
accelerating into the denser fluid [2, 3]. Rayleigh–Taylor instability under different
effects has been explained in detail, Chandrasekhar [4]. William et al. investigated
different ways to measure surface tension at the interface between the fluids [5].
Bhatia [6] observed the behavior of R-T instability among conducting fluids [6].
Rayleigh–Taylor instability of conducting, rotating, a stratified field in presence of
horizontal magnetic field is studied by Chakraborty [7]. Sharp [8] reviewed R-T
instability with different influencing effects [8]. Sharma and Chhajlani [9], Hoshoudy
[10] investigated R-T instability under the effect of rotation for magnetized conduct-
ing plasma [9, 10]. Young and Ham [11] investigated the effect of surface tension
intended for inviscid fluids [11]. Guildenbecher et al [12] reported the secondary
atomization and breaking up of fluid [12]. R-T instability growth during deflagration
in engine combustion was observed by Keenan et al [13]. Baldwin et.al [14] investi-
gated experimentally the effect of rotation and magnetic field on interfacial instability
[14]. Scase and Hill [15] studied the effect of high rotation on two liquid layers [16].
For brevity, only few contributions of importance have been included in this paper.
Even though considerable research has been done in this topic theoretically [2–4,
8] and also experimentally [13, 14], the complexity of the problem demands com-
prehensive theoretical investigations. To demonstrate this, the influence of dynamic
forces on the fluid–fluid interfaces, magnetic fields and rotation on R-T instability
has been investigated theoretically in this paper. The objective of this research is to
characterize the Rayleigh–Taylor instability at the interface based on the stability
parameters, namely growth rate and the dominant wavenumber of the unstable fluid
system. This is attempted for interfaces of different fluid combinations of practical
relevance using the well-established governing equations [4].
Characterization of Rayleigh–Taylor Instability … 403
2 Mathematical Formulation
For depicting Rayleigh–Taylor interface, two fluids of different densities are placed
one above the other with heavier fluid on top and lighter fluid at the bottom resulting
in a fluid–fluid interface as shown in Fig. 1a and has been used. The coordinate
system chosen for this study is presented in the flow domain (Fig. 1b).
Continuity equation based on volume basis with V velocity vector and ∇ vector
differential operator is shown below:
∂ρ
+ ∇. ρ V = 0 (1)
∂t

∂ρ ∂ ∂ ∂
+ î + ĵ + k̂ . ρU î + ρV ĵ + ρW k̂ = 0 (2)
∂t ∂x ∂y ∂z
∂ρ ∂ ∂ ∂
+ (ρU ) + (ρV ) + (ρW ) = 0 (3)
∂t ∂x ∂y ∂z
The system assumed as two dimensional and incompressible and neglecting vari-
ations in the y-direction for the given layer of fluid has to satisfy the continuity
equation below:
∂U ∂W
+ =0 (4)
∂x ∂z
The components of the perturbed flow based on the initial value of the velocity
and the perturbation quantities are:
U = Uo + u (5)
V = Vo + v (6)
W = Wo + w (7)
Fig. 1 a Fluid–fluid interface, b coordinate system

where subscript ‘o’ specifies that constant initial values and u, v, w are perturbation
velocities. Substituting the components of the perturbed flow in Eq. (4) results in
∂(Uo + u) ∂(Wo + w)
+ =0 (8)
∂x ∂z
∂Uo ∂u ∂ Wo ∂w
+ + + =0 (9)
∂x ∂x ∂z ∂z
Utilizing the continuity equation, Eq. (9) reduces to the form below:
∂Uo ∂ Wo
+ =0 (10)
∂x ∂z
Utilizing Newton’s second law of motion for a unit volume of the selected fluid
interface

∂U ∂U ∂U ∂p
ρ +U +W =− (11)
∂t ∂x ∂z ∂x
Being an incompressible flow, perturbation density is constantly resulting in a

reduction of Eq. (11) as shown below:

∂(Uo + u) ∂(Uo + u) ∂(Uo + u)
(ρ + (δρ)) + (Uo + u) + (Wo + w)
∂t ∂x ∂z
∂( p + (δp))
=− (12)
∂x
The modified form of x-momentum equation is obtained by utilizing the initial
conditions (U0 = 0) and linearizing Eq. (12),

∂u ∂(δp)
ρ =− (13)
∂t ∂x
Similarly, the reduced form of z-momentum equation is:

∂w ∂(δp)
ρ = −(δρ)g − (14)
∂t ∂z
In order to close the problem, one more equation is required. An equation in terms
of density perturbation:
∂(ρ + (δρ)) ∂ ∂
+ ((ρ + (δρ))u) + ((ρ + (δρ))w) = 0 (15)
∂t ∂x ∂z
Utilizing linear assumption, Eq. (15) becomes:


∂ρ ∂(δρ) ∂ρ ∂u ∂w ∂ρ
+ +u +ρ + +w =0 (16)
∂t ∂t ∂x ∂x ∂z ∂z
Applying the continuity equation for incompressible flow and density variations
in z-direction, Eq. (16) reduces to:
∂(δρ) ∂ρ
= −w (17)
∂t ∂z
The list of the assumptions made is as follows:

1. Two-dimensional, i.e. ∂/∂y = 0
2. Incompressible (density constant)
3. Velocity restated in terms of initial constant values with perturbed velocities
4. Inviscid interface
5. Flow is static initially (velocities are zero)
6. Linearized the momentum equation
7. The energy equation is neglected (isothermal)
Linear stability analysis involves determining the stability of the fluid flow by
introducing small-scale disturbances and tracking its temporal progression. Consid-
ering a small perturbation in terms of the periodic wave along z = 0 for single-mode
disturbance with the amplitude A is:
A(x, z, t) = Ak (z, t)eikx = Ak (z)ent eikx = Ak (z)eikx+nt (18)
where k is the wavenumber.

Positive eigenvalue gives the positive exponential and unstable state; on the
other hand, a negative eigenvalue will result in a stable condition. For the selected
velocities, density and pressure, the corresponding normal modes are:
u = u(x, z, t) = u k (z)eikx+nt (19)
w = w(x, z, t) = wk (z)eikx+nt (20)
δρ = δρ(x, z, t) = δρk (z)eikx+nt (21)
δp = δp(x, z, t) = δpk (z)eikx+nt (22)
On substituting Eqs. (19), (20), (21) and (22), respectively, in the following forms
of Navier–Stokes, Eqs. (10), (13), (14), (17) result in:
∂w
iku + =0 (23)
∂z
ρnu = −ik(δp) (24)
∂(δp)
ρnw = −(δρ)g − (25)
∂z
∂ρ
n(δρ) = −w (26)
∂z
On solving the above four Eqs. (23–26)

2
d dw k dρ
ρ − ρk 2 w = −wg 2 (27)
dz dz n dz
Equation (27) is the governing equation for predicting the instability of fluid
interface(R-T) generated by the heavier fluid on top of the lighter fluid.
2.1 Rayleigh–Taylor Instability for Two Incompressible

Fluids with Inertial Effects
For fluids of constant density ρ1 and ρ2 forming interface, the governing Eq. (27) is
reduced to:
d 2w
− k2w = 0 (28)
dz 2
Using the boundary conditions that the velocity is zero at large distance above
and below the interface and the fact that the velocity at the interface matched for the
two solutions becomes,
w1 = w0 e+kz (z < 0) (29)
w2 = w0 e−kz (z > 0) (30)
Multiplying dz and integrating

dw k2
∫d ρ − ∫ ρk 2 wdz = − ∫ wg 2 dρ (31)
dz n
Integrating across the interface across an infinitesimal distance dz = 0, the second

term becomes zero:

dw k2
ρ = −wg 2 ρ (32)
dz n
k2
ρ2 (−kw) − ρ1 (kw) = −wg (ρ2 − ρ1 ) (33)
n2
k
−(ρ2 + ρ1 ) = −g (ρ2 − ρ1 ) (34)
n2
Equation (34) on rearrangement yields an expression for determining growth rate
(n):

ρ2 − ρ1
n = gk
2
(35)
ρ2 + ρ1
It is to be noted that the fluid system is unstable if the heavy fluid is above the
lighter fluid (ρ2 > ρ1 ), because the eigenvalues are real and system unstable. On the
contrary, if the lighter fluid is above the heavier fluid (ρ2 < ρ1 ), the eigenvalues are
the imaginary and the system is stable.
To find the mode of maximum perturbation or growth, setting the first derivative
of Eq. (35) (dn/dk = 0) to zero,

ρ2 − ρ1 1
g √ =0 (36)
ρ2 + ρ1 2 k
On squaring both sides of the resulting equation,

ρ2 − ρ1 1
g =0 (37)
ρ2 + ρ1 4k
Equation (37) enables determination of the dominant wavenumber:
kD = ∞ (38)

Fluids with Surface Tension Effect
On introducing the effect of surface tension in Eq. (27), the fundamental equation
reduces to:
D2 − k 2 w = 0 (39)
Utilizing the vanishing boundary conditions, namely z→−∞ and z→+∞,

Eq. (32) results in:
k2
0 (ρ Dw) = − g(ρ2 − ρ1 ) − k 2 σ w0 (40)
n2
where w0 is the common value of w at z = 0. The resulting solution helps in
determining the growth rate:
k2
−k(ρ2 + ρ1 ) = − g(ρ2 − ρ1 ) − k 2 σ (41)
n2
The corresponding growth rate and wavenumber involving inertial and surface
tension effects [4]:

ρ2 − ρ1 k2σ
n = gk
2
− (42)
ρ2 + ρ1 g(ρ2 + ρ1 )
Equation (42) reveals the stability of the interface when ρ2 < ρ1 , otherwise
unstable. The corresponding critical wavenumber was determined to be:
21
(ρ2 − ρ1 )g
kc = (43)
σ
Mode of maximum growth has been obtained using the first derivative test
(dn/dk = 0), based on Eq. (42):

1 1 1 − 21 1 1 1
k 2 g(ρ2 − ρ1 ) − k 2 σ (−2kσ ) + k − 2 g(ρ2 − ρ1 ) − k 2 σ 2
=0
ρ2 + ρ1 2 2
(44)
21
1 (ρ1 − ρ2 )g
kD = √ (45)
3 σ
Equation (45) indicates the relationship between critical and dominant wavenum-
ber:
kc
kD = √ (46)
3
On substituting Eq. (46) in Eq. (42), expression for dominant growth rate has been
obtained:
⎛ ⎞1/ 2
3/ 3/
2 (ρ − ρ ) 2 g 2
nD = ⎝ √ ⎠
2 1
(47)
3 3 (ρ + ρ )1/ 2
2 1

Fluids with of Rotational Effect
To study the effect of rotation in the stability of the inviscid fluid, the fluid is subjected
to a uniform rotation with angular velocity . The momentum equation involving
the effect of rotation and pressure is utilized to arrive at the governing equations:

du ∂
ρ − 2ρv = − (δp) (48)
dt ∂x

dv ∂
ρ − 2ρu = − (δp) (49)
dt ∂y
∂2
dw ∂ ∂2
ρ =− (δp) − g(δρ) + σs + (δz s )δ(z − z s ) (50)
dt ∂z s
∂x2 ∂ y2
∂
(δρ) = −w Dρ (51)
∂t
∂u ∂v ∂w
+ + =0 (52)
∂x ∂y ∂z
Density may change discontinuously at pre-assigned level (z s ), a known surface

(s) and surface tension effect that is included in Eq. (50).
On linearizing above equations, namely (48), (49), (50), (51) and (52), and subse-
quently introducing the respective perturbations in the form of normal modes, results
in equations mentioned below:
ρnu − 2ρv = −ik x (δp) (53)
ρnv − 2ρu = −ik y (δp) (54)

ρnw = −D(δp) − g(δρ) − k 2 σs (δz s )δ(z − z s ) (55)
s
n(δρ) = −w D (56)
ik x u + ik y v = −Dw (57)
z-component of vorticity is defined below:
ζ = ik x v − ik y u (58)
Effective utilization of Eqs. (53), (54), (57) and (58) helps in determination of
vorticity: ζ = 2Dw/n and reduction of Eq. (55)
g k2
D(δp) = −ρnw + (Dρ)w − σs ws δ(z − z s ) (59)
n n s
On eliminating δp in Eq. (59),

2 2
42 gk k
1 + 2 D(ρ Dw) − k ρw = −
2
(Dρ) − σs δ(z − z s )
n n2 g s
(60)
Expressing,
k2
θ2 = (61)
1 + 42 /n 2
On rewriting Eq. (60) by utilizing Eq. (61):

2 2
gθ k
D(ρ Dw) − θ ρw = −
2
(Dρ) − σs δ(z − z s ) w (62)
n2 g s
It has to be noted that θ 2 is always positive for n 2 > 0 or n 2 < −42 resulting
in alteration of Eq. (62):
2
θ
s (ρ Dw) = − 2 gs (ρ) − k 2 σs ws (63)
n
while
2
gθ
D(ρ Dw) − θ ρw = −
2
(Dρ)w (z = z s ) (64)
n2
Equation (42) transforms into the form below on solving Eq. (64) [4]:

ρ2 − ρ1 k2σ
n = gθ
2
− (65)
ρ2 + ρ1 g(ρ2 + ρ1 )
On substitution Eq. (61) in Eq. (65), the expression for the growth rate is obtained:
1
42 2 ρ2 − ρ1 k2σ
n 1+ 2
2
= gk − (66)
n ρ2 + ρ1 g(ρ2 + ρ1 )
Expressing the left-hand side of Eq. (66) in terms of n 0


42
n 2
1+ = n 20 (67)
n2
The possible solutions are [4],

n = −2 +
2 2
44 + n 40 i f n 20 > 0 (68)

n 2 = −22 − 44 + n 40 i f n 20 < 0 (69)
Utilizing Eqs. (68) and (69) behavior of fluid–fluid interface under rotation can
be predicted.
The theoretical analysis for determining the stability of Rayleigh–Taylor interfaces

has been carried out to investigate the influence of inertial and surface tension and
rotational effect in promoting instability at the interface utilizing Eqs. (35), (42),
(65), respectively. The results are presented in the following order: inertial alone
followed by combined influence of inertial and surface tension and finally including
rotational effect to combined influence of inertia and surface tension.
Figure 2 depicts the growth rate in a pure inertial situation based on Eq. (35) for
wavenumbers in the range of 0 to 1000. The results indicate an increase in growth rate
with an increase in wavenumber for the range of condition tested. This confirms the
possibility of the mode of maximum instability only at infinity. This may be due to the
dominance of inertial forces over the resistive force suppressing the disturbances to
put back the system into stable condition. This leads to rapid growth of perturbation
resulting in exertion of extraction of dominant wavenumber.
Fig. 2 Growth rate of perturbation under the influence of inertial effect

Fig. 3 Growth rate of perturbation under the combined influence of inertial and surface tensional
effect: a air-based interface, b water-based interface
The effect of inertial and surface tension forces on the stability of the fluid–
fluid interface has been studied using Eq. (43). Figure 3 includes the growth rate
predictions of the disturbances generated at the interface involving, air (Fig. 3a)
or water (Fig. 3b) with the tested fluids. The results based on air-based interfaces
delineate the existence of a distinct mode of maximum instability that can be extracted
in contrast to the corresponding pure inertial situation. In addition, the dual nature
of surface tension in promoting instability when water is involved and suppressing
instability for air has been noted. It stabilizes the air-based fluid–fluid interfaces and
on the contrary, destabilizes for water-based fluid–fluid interfaces, in comparison
with the inertial situation. Correspondingly, the wavenumber at which maximum
growth rate happens appears to be depended on the fluid.
Figure 4 reveals the comparison of growth rates for different fluids for determining
dominant wavenumber under the influence of inertial with and without surface ten-
sion. Influence of surface tension is to stabilize the generated interfacial waves selec-
tively enabling extraction of the dominant wave responsible for maximum growth
rate among the spectrum. It is evident for all the fluids tested irrespective of air- or
water-based interfaces.
The effect of rotation on a disturbed Rayleigh–Taylor interface should be stabi-
lizing the fluid interface bringing chunks of fluid closer to each other that has been
separated otherwise. Figure 5 predicts the growth rate as a function of wavenum-
ber for water and air-based interfaces that are already under combined influence of
inertial and surface tension. The effect of rotational field is to suppress the growth
of the introduced perturbation. Higher rotational speed results in mixing of fluid
particles ensuring the stability of the flow field. From the figures, it is evident that
the stability of the fluid can be increased by maintaining higher rotational speeds.
Rotational speed was varied from 0.002 to 13 rad/s, and the corresponding effect on
the growth rate was determined. Individual evolution reveals the existence of unique
Fig. 4 Growth curve for inertial and surface tension a water–air, b diesel–air, c gasoline–air,
d nitromethane–water, e ethylene glycol–water, f ethanol–water, g glycerol–water
Fig. 5 Rayleigh–taylor instability for an interface under the combined effects of inertial, surface
tension and rotation for different rotational speeds: a water–nitromethane, b water–ethylene glycol,
c water–ethanol, d water–glycerol, e water–air, f air–gasoline, g air–diesel
wavenumber having maximum growth rate. The growth rate curves are identical irre-
spective of the liquid used for air-based interfaces suggesting that rotational effect is
autonomous of fluids. However, significant variation is observed in the water-based
interfaces. Correspondingly, an increase in rotational speed increases the stability of
water-based fluid interfaces.
Fig. 6 Rayleigh–taylor instability of interfaces under the influence of specific rotational speeds:
a = 0.002, b = 0.02, c = 1, d = 5, e = 10, f = 13
Figure 6 suggests the behavior of fluid interfaces subjected to different rotational

speeds ranging from 0.002 to 13 rad/sec. The influence of an increase in rotational
speed on the stability for water-based interfaces is more prominent than the air-based
interfaces. Also, the increase in rotational speed suppresses the developed instability
significantly overcoming the influence of inertia resisting the attainment of interfacial
stability.
4 Conclusion
In the present study, the behavior of fluid interfaces at rest subjected to Rayleigh–
Taylor type of instabilities has been attempted. The objective is to predict stability of
interface under the influence of combined inertial, surface tensional and rotational
effects for different fluid combinations. It is evident that under the influence of
inertial forces only, the interface is always unstable. On the contrary, including surface
tension effects to pure inertial condition, there is a possibility for the fluid interface to
attain stable condition. Also, unimodal behavior observed supports in determination
of maximum growth rate and the associated dominant wavenumber. Implicit analysis
reveals the dual nature of surface tension; the air-based interfaces are stabilized and
on the contrary, destabilizing the water-based interfaces. In addition, for air-based
systems, the sensitiveness of dominant wavenumber resulting in maximum growth
rate on fluid properties has been observed. The effect of rotation to an interface
involving inertial and surface tension effects has been to suppress the growth of the
instabilities with increase in rotational speed. Besides, at higher speeds, the effect
of rotation in attainment of stability has been prominent in water-based systems in
comparison with air-based interfaces. By considering the viscous influence of fluids

forming interfaces in the future studies, a better understanding of R-T instability
can be realized. In the case of nuclear reactors, this kind of fluid–fluid interface is
predominant as mentioned in Siddharth et al. [17]. Karthick et al. [16] reported the
Kelvin–Helmholtz (K-H) instability at the fluid–fluid interface due to accelerating
fluids. This is different from the Rayleigh–Taylor (R-T) instability mentioned in this
paper. Nevertheless a fluid–fluid interface can experience K-H or R-T instability
depending upon the relative positions of heavier and lighter fluids.
References
1. Rayleigh L (1883) Investigation of the character of the equilibrium of an incompressible heavy

fluid of variable density. Proc Lond Math Soc 14:170–177
2. Taylor GI (1950) The instability of liquid surfaces when accelerated in a direction perpendicular
to their planes. Proc Roy Soc Lon Series A, Math Phys Sci 201(1065):192–196
3. Lewis DJ (1950) The instability of liquid surface when accelerated in perpendicular to their
planes (II). Proc R Soc Lond Ser A 202:81
4. Chandrasekhar S (1961) Hydrodynamic and hydromagnetic stability. Dover publications
chapter X 428–477
5. Harkins WD, Humphrey EC (1915) The surface tension at the interface between two liquids and
the effect of acid, salts, and base upon the interfacial tension. J Am Chem Soc 38(2):242–246
6. Bhatia PK (1973) Rayleigh-Taylor instability of two viscous superposed conducting fluids.
Astrophys Space Sci 19B(2)
7. Chakraborty BB (1982) Hydromagnetic Rayleigh-Taylor instability of a rotating stratified fluid.
Phys Fluids 25:743
8. Sharp DH (1983) An overview of Rayleigh-Taylor Instability. Los Alamos Univ New Mex
12:3–10
9. Sharma PK, Chhajlani RK (1998) Effect of rotation on the Rayleigh-Taylor instability of two
superposed magnetized conducting plasma. Phys Plasmas 5:2203
10. Hoshoudy GA (2012) External magnetic field effects on the rayleigh-taylor instability in an
inhomogeneous rotating quantum plasma. J Mod Phys 3:1792–1801
11. Young YN, Ham FE (2006) Surface tension in incompressible rayleigh-taylor mixing flow. J
Turbul 7(71)
12. Guildenbecher DR, López-Rivera C, Sojka PE (2009) Secondary Atomization. Exp Fluids
46:371–402
13. Keenan JJ, Makarov DV, Molkov VV (2014) Rayleigh-taylor instability: modeling and effect
on coherent deflagrations. Int J Hydrogen Energy 39(35)
14. Baldwin KA, Scase MM, Hill RJA (2015) The inhibition of rayleigh-taylor instability by
rotation. Sci Rep 5:11706
15. Scase MM, Hill RJA (2018) Centrifugally forced rayleigh-taylor instability. J Fluid Mech
852:543–577
16. Karthick S, Srivatsan M, Raghav R, Vishnu Supreet, K (2015) Theoretical investigation of
primary breakup of liquid jet using linear stability analysis in the first wind induced regime
based on kelvin-helmholtz instability. 1901–3905
17. Siddharth SS, Vikram G, Natesan K, Karthick S (2018) Modeling, simulation and analysis of
primary pipe rupture in sodium cooled fast breeder reactor. 045 ICMMC December
Heat and Mass Transfer Analysis
of Al2 O3 -Water and Cu-Water
Nanofluids Over a Stretching Surface
with Thermo-diffusion
and Diffusion-Thermo Effects Using
Artificial Neural Network
D. Seenivasan, M. Elayarani, and M. Shanmugapriya
Abstract The present paper deals with the artificial neural network (ANN) model-
ing of heat and mass transfer on magnetohydrodynamic (MHD) convective flow of
Al2 O3 -water and Cu-water nanofluids past a stretching sheet through porous media
with thermo-diffusion and diffusion-thermo effects. The set of suitable similarity
transformations are employed to alter the nonlinear partial differential equations
into ordinary differential equations. The solutions of the resulting nonlinear dif-
ferential equations are solved numerically with the help of Runge–Kutta Fehlberg
fourth–fifth order method accompanied by shooting technique, and then, the ANN
is applied to them. The multilayer feedforward neural network with backpropaga-
tion training algorithm is used for predicting the desired outputs. The influence of
various physical parameters on velocity, temperature and concentration profiles is
explored and discussed in detail. The friction factors, heat and mass transfer rates
are predicted using ANN. The numerical results and the results of the ANN are in
good agreement with errors less than 5%. According to the findings of this paper, the
ANN approach is reliable and more effective for simulating heat and mass transfer
in MHD nanofluid flow over a stretched sheet.
Keywords Magnetohydrodynamics · Nanofluid · Heat transfer · Stretching sheet ·

Artificial neural network
D. Seenivasan
Department of Mathematics, Jeppiaar SRR Engineering College, Padur, Chennai, Tamil Nadu
603103, India
M. Elayarani · M. Shanmugapriya (B)
Department of Mathematics, SSN College of Engineering, Kalavakkam, Chennai, Tamil Nadu
603110, India

https://doi.org/10.1007/978-981-15-4488-0_36
418 D. Seenivasan et al.
1 Introduction
Heat and mass transfer flows with chemical reactions have received great attention
among the researchers due to its significant applications in processes such as evap-
oration, drying, energy transfer, flow in a desert cooler, distribution of temperature
and moisture over agricultural fields and trees. Combined effects of heat and mass
transfer on MHD flow over a stretching sheet in the presence of chemical reac-
tion under various physical aspects have been analyzed by the researchers [1–4].
The effects of heat generation or absorption play a vital role in cooling processes,
exothermic chemical reactions, storage of foodstuffs, disposal of radioactive waste
materials and separating fluids in packed-bed reactors. Studies on MHD nanofluid
flow with heat generation or absorption over different surfaces were addressed in
the publications [5–8]. Magnetohydrodynamic convective flow of nanofluids over
various stretching velocities, namely linear, radial, non-isothermal and exponential,
have been investigated by the authors [9–12]. The combined effects of Soret and
Dufour are significant when there exists density difference in the flow regime and
finds potential applications in the areas of chemical engineering and geosciences.
Many researchers [13–18] interpreted the Soret and Dufour effects of MHD flow
with chemical reaction under varied geometries.
In the recent years, ANNs have been extensively used in the fluid flow problems
to reduce the time, expenses and complicated computation works. Ziaei-Rad et al.
[19] applied a multilayer neural network model for finding the skin friction factor
and Nusselt number of dissipative nanofluid flow under a variable MHD field over a
permeable horizontal stretching surface. Aminian [20] used a cascade-forward neural
network model for predicting the effective thermal conductivities of 26 different
types of nanofluids. The MHD convective boundary layer slip flow on a permeable
stretched cylinder with chemical reaction was modeled using artificial neural network
technique by Reddy et al. [21]. Tafarroj et al. [22] proposed an ANN model for
calculating the heat transfer coefficient and Nusselt number on the flow of TiO2 -
water nanofluid in a microchannel heat sink. Tripathy et al. [23] employed ANN
approach for forecasting the free convective flow and heat transfer between two
coaxial cylinders of an elastic-viscous liquid. Maddah et al. [24] estimated the heat
transfer efficiency of water–iron oxide nanofluid in a double pipe heat exchanger
equipped with twisted tape using experiment and artificial neural networks. The
current paper extends the artificial neural network technique to predict the friction
factor, heat and mass transfer rates due to the MHD convective flow of Al2 O3 -water
and Cu-water nanofluids on a linear porous stretching surface in the presence of Soret
and Dufour effects.
Heat and Mass Transfer Analysis of Al2 O3 -Water and Cu-Water … 419
2 Mathematical Analysis
We consider a two-dimensional, steady, laminar, MHD boundary layer flow of an

incompressible, viscous nanofluid past a stretching sheet through porous medium in
the plane y = 0, and the flow is assumed to be confined in a region y > 0. The
stretching sheet coincides with the x-axis in the direction of the flow, and y-axis
is perpendicular to the surface. A constant magnetic field B0 is applied along the
y-direction, and the stretching sheet is extended with velocity u w (x) = ax, where
a > 0 is a constant. The uniform temperature Tw and nanoparticle fraction Cw are
assumed to be greater than the ambient temperature T∞ and nanoparticle fraction
C∞ , respectively. In this problem, water-based nanofluid is considered with two
different types of nanoparticles, namely Al2 O3 and Cu. The fluid and nanoparticles
are assumed to be in thermal equilibrium, and no slip occurs between them, and the
thermophysical properties are given in Table 1. The boundary layer equations that
govern the problem are given by
∂u ∂v
+ =0 (1)
∂x ∂y
∂u ∂u ∂ 2u νn f σ B02 u
u +v = νn f 2 − u− (2)
∂x ∂y ∂y K ρn f
2
∂T ∂T ∂2T 1 ∂qr μn f ∂u
u +v = αn f 2 − +
∂x ∂y ∂y ρc p n f ∂ y ρc p n f ∂y
1 D2 ∂ 2 C
+ q + (3)
ρc p n f ρc p n f ∂ y 2
∂C ∂C ∂ 2C ∂2T
u +v = Dm 2 + D1 2 − K 0 (C − C∞ ) (4)
∂x ∂y ∂y ∂y
with the associated boundary conditions
u = u w (x) = ax, v = vw , T = Tw , C = Cw at y = 0 (5)
u = 0, T → T∞ , C → C∞ as y → ∞ (6)
Table 1 Thermophysical properties of water and nanoparticles [9]

Fluid ρ (kg/m3 ) c p (J/kg K) κ (W/m K) β × 105 K−1
Water 997.1 4179 0.613 21
Alumina (Al2 O3 ) 3970 765 40 0.85
Copper (Cu) 8933 385 401 1.67
where u and v are the velocity components along x- and y axes, respectively, T is
the temperature, C is the concentration, vw is the suction or injection velocity, K is
the permeability of porous medium, σ is the electrical conductivity, D1 and D2 are
the mass and heat fluxes through temperature and concentration gradient, Dm is the
species diffusivity and K 0 is the chemical reaction parameter.
The effective density ρn f , dynamic viscosity μn f , thermal conductivity
κn f , kine-
matic viscosity νn f , thermal diffusivity αn f and heat capacitance ρc p n f of the
nanofluid are defined as follows:
μ
ρn f = (1 − ϕ)ρ f + ϕρs , μn f = (1−ϕ)f 2.5 ,

κn f = κ f κs +2κ f +2ϕ (κ f −κs ) , νn f = ρnnff ,
κ +2κ −2ϕ κ −κ μ
( f s) (7)
κ s
f

αn f = ρcn f , ρc p n f = (1 − ϕ) ρc p f + ϕ ρc p s
( p )n f
where ϕ represents the solid volume fraction of nanoparticles in the fluid.

The non-uniform heat source/sink q is defined by
q = A(Tw − T∞ ) f + B(T − T∞ ) (8)
where A and B are the coefficients of space and temperature-dependent heat

source/sink parameters, in which the case A > 0, B > 0 represents internal heat
source and the case A < 0, B < 0 represents internal heat sink.
Under Rosseland approximation, the radiative heat flux qr takes the form
4σ ∗ ∂ T 4
qr = − (9)
3k ∗ ∂ y
where σ ∗ and k ∗ are the Stephan–Boltzmann constant and mean absorption coeffi-
cient. Consider the temperature difference within the flow is sufficiently small such
that T 4 can be expressed as the linear function of temperature T∞ . Expanding T 4 in
a Taylor series about T∞ and neglecting higher order terms yield
T 4 ≡ 4T∞
3
T − 3T∞
4
(10)
3 ∗
16T∞ σ ∂T
qr = − (11)
3k ∗ ∂ y
The stream function ψ is defined as
∂ψ ∂ψ
u= , v=− (12)
∂y ∂x
We introduce the following similarity transformations

√
ψ= aν f x f (η), η = νaf y
√
u = ax f (η), v = − aν f f (η) (13)
θ (η) = TTw−T ∞
−T∞
, φ(η) = CC−C ∞
w −C ∞
By using Eqs. (8), (11) and (13), the Eqs. (1)–(4) are reduced to
2 M
f + A1 A2 f f − f − f − K1 f = 0 (14)
A2

4
1 + R θ
3
A3 Ec 2 1
+ Pr f θ − 2 f θ + f + A1 f + B1θ + Duφ = 0 (15)
A5 A4 A3
φ + Sc f φ + Srθ − C1φ = 0 (16)
The corresponding boundary conditions are
f = S, f = 1, θ = 1, φ = 1 at η = 0 (17)
f = 0, θ = 0, φ = 0 at η = ∞ (18)
σ B02
where prime denotes the differentiation with respect to η, M = aρ f
the magnetic
νf 4σ ∗ T∞ 3
parameter, K 1 = Ka
the porosity parameter, R = κn f k ∗
the radiation parameter,
νf u 2w
Pr = the Prandtl number, Ec = the Eckert number, A1 = a ρcA
αf (c p ) f (Tw −T∞ ) ( p) f
the space-dependent heat source/sink, B1 = a ρcB the temperature-dependent heat
( p) f
(Cw −C∞ ) ν
source/sink, Du = ν ρc (Tw −T∞ ) the Dufour number, Sc = Dmf the Schmidt
D2
f( p) f
(Tw −T∞ )
number, Sr = Dν f1 (C w −C ∞ )
the Soret number, C1 = Ka0 the chemical reaction
parameter.

ρs ρc p s
A1 = (1 − ϕ) , A2 = (1 − ϕ) + ϕ , A3 = (1 − ϕ) + ϕ
2.5
,
ρf ρc p f

ρc p s κn f
A4 = (1 − ϕ)2.5
(1 − ϕ) + ϕ , A5 =
ρc p f κf
The special interest in this problem is the local skin friction coefficient C f x , the
local Nusselt number Nux and the local Sherwood number Shx which are defined as
τw xqw x Jw
Cfx = , Nux = , Shx = (19)
ρ f u w /2
2 κ(Tw − T∞ ) Dm (Cw − C∞ )
where τw is the wall shear stress, qw and Jw are the heat and mass fluxes at the
surface.
Substituting Eq. (13) into Eq. (19), we get the following expressions in
dimensionless form
1
−2 f (0)
C f x Rex2 = (1−ϕ)2.5
− 21 κn f
Nux Rex = − κ f 1 + 43 R θ (0) (20)
−1
Shx Rex 2 = −φ (0)
3 Numerical Procedure
In this paper, Runge–Kutta Fehlberg fourth–fifth order method has been used with
shooting technique to solve the system of coupled nonlinear ordinary differential
Eqs. (14)–(16) with the boundary conditions (17) and (18). First, we define the
variables f = Y1 , f = Y2 , f = Y3 , θ = Y4 , θ = Y5 , φ = Y6 , φ = Y7 .
Equations (14)–(16) are reduced to system of first-order differential equations as
follows

Y1 = Y2

Y2 = Y3
M
Y3 = −A1 A2 Y1 Y3 − (Y2 )2 − Y2 + K 1 Y2
A2

Y4 = Y5
1
Y5 = −
1 + 43 R
A3 Ec 1

Pr Y1 Y5 − 2Y2 Y4 + (Y3 )2 + A1Y2 + B1 Y4 + DuY7
A5 A4 A3

Y6 = Y7

Y7 = −Sc Y1 Y7 + SrY5 − C1Y6 (21)
with the corresponding boundary conditions
Y1 (0) = S, Y2 (0) = 1, Y3 (0) = c1 , Y4 (0) = 1

Y5 (0) = c2 , Y6 (0) = 1, Y7 (0) = c3 (22)
In order to get the solution of the above initial value problem, the unknown initial
values f (0), θ (0) and φ (0) are to be determined by making initial guesses, and then,
integration is carried out with the aid of Runge–Kutta Fehlberg fourth–fifth order
scheme. The initial values are chosen in such a way that it satisfies the condition
f (η∞ ) → 0, θ (η∞ ) → 0 and φ(η∞ ) → 0 for finite domain length ηmax . The
above procedure is repeated until the required accuracy is achieved with the step size
of η = 0.001.
4 Artificial Neural Networks
Artificial neural network is an advanced computing tool consists of an extremely

interconnected neurons and processes information using neurocomputing technique.
A multilayer neural network model is shown in Fig. 1, which comprises an input layer,
output layer and a hidden layer of nine neurons. Back error propagation (BEP) train-
ing algorithm has been used with sigmoid hidden neurons and linear output neurons.
The neural network adjusts weights at every node in all layers during training process
in order to minimize the error between the predicted and actual result. In this study,
Fig. 1 A multi-layer neural network model [25]

the network uses twelve parameters (ϕ, M, K 1 , R, Pr, Ec, A1, B1, Du, Sc, Sr, C1)
as input nodes, and output parameters are f (0), θ (0), φ (0) for alumina–water and
copper–water nanofluid. The numerical data set was divided into training (70%),
testing (15%) and validation (15%) sets of the neural network.
The accuracy of the current numerical method is assessed by obtaining the values
of skin friction coefficient for the various values of M, and ϕ in the absence of
K 1 , R, Ec, A1, B1, Du, Sc, Sr, C1 when Pr = 6.2 is shown in Table 2. It is clear
that the results are in good agreement with Hamad [9]. Figures 2, 3 and 4 display
the velocity ( f ), temperature (θ ) and concentration (φ) profiles plotted against η for
varying ϕ. From Fig. 2, it is observed that enhancing ϕ leads to increase the velocity
field for Al2 O3 -water and decreases for Cu-water nanofluid. From Figs. 3 and 4, it is
seen that the dimensionless temperature elevates with the growing values of ϕ, and
a reverse trend is occurred in the concentration profile for both nanofluids.
Figures 5, 6 and 7 represent the impact of M on velocity, temperature and con-
centration distributions. This is due to the fact that the transverse magnetic field
produces a resistive type of force (Lorentz force) when applied to an electrically
conducting fluid. The Lorentz force has the tendency to diminish the motion of the
fluid and accelerate the temperature and concentration profiles. As a consequence,
it is perceived from Fig. 5 that the magnetic parameter M weakens the velocity for
alumina–water and copper–water nanofluids and thereby strengthen the temperature
and concentration inside the boundary layer for both fluids which are shown in Figs. 6
and 7.
Figures 8 and 9 show the influence of temperature-dependent heat source/sink
coefficient B1 on dimensionless temperature and concentration for both Al2 O3 -water
and Cu-water nanofluids. From Fig. 8, it is noticed that the temperature profile rises
for the heat source B1 > 0 and decreases for the heat absorption B1 < 0. This
Table 2 Comparison of − f (0) with previously published results

M ϕ Hamad [9] Al2 O3 Cu Present results Al2 O3 Cu
0 0.05 1.00538 1.10892 1.00554 1.10901
0 0.1 0.99877 1.17475 0.99895 1.17481
0 0.15 0.98185 1.20886 0.98204 1.20891
0 0.2 0.95592 1.21804 0.95615 1.21809
0.5 0.05 1.20441 1.29210 1.20442 1.29211
0.5 0.1 1.17548 1.32825 1.17550 1.32825
0.5 0.15 1.13889 1.33955 1.13891 1.33956
0.5 0.2 1.09544 1.33036 1.09547 1.33036
Fig. 2 Velocity profiles for varying ϕ
Fig. 3 Temperature profiles for varying ϕ
is because of the fact that the boundary layer creates energy when B1 > 0 and
absorbs energy when B1 < 0. In Fig. 9, an opposite trend has observed in the
concentration profile for both heat generation and absorption cases of temperature-
dependent coefficient B1 in both nanofluids.
The mass and energy fluxes caused by temperature gradient and concentration
difference are called thermo-diffusion (Soret) and diffusion-thermo (Dufour) effects,
respectively. Figures 10 and 11 illustrate the combined effect of Soret and Dufour
Fig. 4 Concentration profiles for varying ϕ
Fig. 5 Velocity profiles for varying M
numbers with increasing Sr and decreasing Du. The values of Sr and Du are taken
in pairs such as (0.1, 0.5), (0.3, 0.4), (0.5, 0.2) and (1, 0.1). It is interesting to notice
that the dimensionless temperature and concentration enhance significantly with a
rise in Soret number Sr (or decrease in Dufour number Du) for both Al2 O3 -water and
Cu-water nanofluids.
The proposed ANN model contains an input layer of 12 neurons, output layer of
six neurons and a hidden layer. To avoid the overfit or underfit of neurons, the network
was trained with the increasing number of hidden neurons. Figures 12 and 13 exhibit
Fig. 6 Temperature profiles for varying M
Fig. 7 Concentration profiles for varying M
the coefficient of determination (R2 ) and test errors (MSE and MAE) for varying
neurons in the hidden layer. It is perceived that the optimum R2 was achieved for
nine hidden neurons, and the corresponding R2 , MSE and MAE values are 0.973729,
0.002785 and 0.016838. Figures 14, 15 and 16 illustrate the regression plot between
the numerical and the predicted ANN values of the skin friction coefficient, Nusselt
number and Sherwood number for the overall data set. The correlation coefficient
Fig. 8 Temperature profiles for varying B1
Fig. 9 Concentration profiles for varying B1
R of f (0), θ (0) and φ (0) for Al2 O3 -water and Cu-water nanofluids is 0.98443,
0.97102 and 0.99619, which shows the accuracy of the neural network.
Fig. 10 Temperature profiles for varying Sr and Du
Fig. 11 Concentration profiles for varying Sr and Du
6 Conclusion
In this paper, we have investigated the numerical and ANN modeling of two-
dimensional MHD convective boundary layer flow, heat and mass transfer character-
istics of alumina–water and copper–water nanofluids over a porous linear stretching
surface in the presence of Soret and Dufour effects. The results are summarized as
follows:
Fig. 12 Coefficient of determination R2 for varying hidden neurons
Fig. 13 MSE and MAE test errors for varying hidden neurons
• As the solid volume fraction ϕ increases, the momentum boundary layer becomes
thicker for Al2 O3 -water and thinner for Cu-water nanofluids, whereas the thick-
ness of the thermal boundary layer upsurges and concentration boundary layer
decelerates for both nanofluids.
• The increase of magnetic parameter M leads to decrease the momentum boundary
layer thickness due to the effect of Lorentz force and improves the thickness of
the thermal and solutal boundary layer in the flow regime.
Fig. 14 Comparison between numerical and ANN predictions of f (0)
Fig. 15 Comparison between numerical and ANN predictions of θ (0)

Fig. 16 Comparison between numerical and ANN predictions of φ (0)
• The presence of temperature-dependent coefficient B1 enhances the thermal

boundary layer thickness for the heat generation B1 > 0 and reduces for the
heat absorption B1 < 0. The exact opposite trend appears in the thickness of the
concentration boundary layer for both heat source and heat sink cases.
• The diffusive species with high Soret number and low Dufour number tends to
increase the thickness of the thermal and solutal boundary layer.
• The correlation coefficient R shows that the proposed ANN model provides better
accuracy in the prediction of MHD convective nanofluid flow problems.
• A well-trained neuromorphic model could be an alternative and affordable way for
numerical and experimental studies which may be time consuming and expensive.
Acknowledgements The authors thank the management of SSN College of Engineering for
providing the necessary facilities to carry out the present work.
References
1. Sandeep N, Sulochana C (2016) MHD flow over a permeable stretching/shrinking sheet of a

nanofluid with suction/injection. Alexandria Eng J 55:819–827
2. Prasannakumara BC, Gireesha BJ, Krishnamurthy MR, Ganesh Kumar K (2017) MHD flow
and nonlinear radiative heat transfer of Sisko nanofluid over a nonlinear stretching sheet. Inform
Med Unlocked 9:123–132
3. Sreekala B, Janardhan K, Ramya D, Shravani I (2017) MHD boundary layer nanofluid flow of
heat transfer over a nonlinear stretching sheet presence of thermal Radiation and partial slip
with suction. Global J Pure Appl Math 13(9):4927–4941
4. Shanmugapriya M (2018) Heat and mass transfer analysis of MHD stagnation point flow past
an unsteady stretching sheet. Int J Mech Prod Eng Res Dev 8(7):577–584
5. Nandy SK, Mahapatra TR (2013) Effects of slip and heat generation/absorption on MHD
stagnation flow of nanofluid past a stretching/shrinking surface with convective boundary
conditions. Int J Heat Mass Transf 64:1091–1100
6. Ganga B, Mohamed Yusuff Ansari S, Vishnu Ganesh N, Abdul Hakeem AK (2016) MHD flow
of Boungiorno model nanofluid over a vertical plate with internal heat generation/absorption.
Propulsion Power Res 5(3):211–222
7. Sandeep N, Sulochana C (2016) Dual solutions of radiative MHD nanofluid flow over an
exponentially stretching sheet with heat generation/absorption. Appl Nanosci 6(1):131–139
8. Sudarsana Reddy P, Chamkha AJ, Al-Mudhaf A (2017) MHD heat and mass transfer flow of a
nanofluid over an inclined vertical porous plate with radiation and heat generation/absorption.
Adv Powder Technol 28(3):1008–1017
9. Hamad MAA (2011) Analytical solution of natural convection flow of a nanofluid over a linearly
stretching sheet in the presence of magnetic field. Int Commun Heat Mass Transf 38:487–492
10. Bilal Ashraf M, Hayat T, Alsaedi A, Shehzad SA (2015) Convective heat and mass transfer
in MHD mixed convection flow of Jeffrey nanofluid over a radially stretching surface with
thermal radiation. J Central South Univ 22:1114–1123
11. Pal D, Mandal G (2015) MHD convective stagnation-point flow of nanofluids over a non-
isothermal stretching sheet with induced magnetic field. Meccanica 50:2023–2035
12. Nayak MK, Akbar NS, Tripathi D, Khan ZH, Pandey VS (2017) MHD 3D free convective
flow of nanofluid over an exponentially stretching sheet with chemical reaction. Adv Powder
Technol 28(9):2159–2166
13. Pal D, Mondal H (2011) MHD non-Darcian mixed convection heat and mass transfer over a
non-linear stretching sheet with Soret-Dufour effects and chemical reaction. Int Commun Heat
Mass Transf 38:463–467
14. Chamkha AJ, Rashad AM (2013) Unsteady heat and mass transfer by MHD mixed convection
flow from a rotating vertical cone with chemical reaction and Soret and Dufour effects. Can J
Chem Eng 92(4):758–767
15. Mohan Krishna P, Vidyasagar G, Sandeep N, Sugunamma V (2015) Soret and Dufour effects
on MHD free convective flow over a permeable stretching surface with chemical reaction and
heat source. Int J Sci Eng Res 6(9):143–147
16. Karthikeyan S, Bhuvaneswari M, Sivasankaran S, Rajan S (2016) Soret and Dufour effects on
MHD mixed convection heat and mass transfer of a stagnation point flow towards a vertical
plate in a porous medium with chemical reaction, radiation and heat generation. J Appl Fluid
Mech 9(3):1447–1455
17. Sudarsana Reddy P, Chamkha AJ (2016) Soret and Dufour effects on MHD convective flow
of Al2 O3 -water and TiO2 -water nanofluids past a stretching sheet in porous media with heat
generation/absorption. Adv Powder Technol 27(4):1207–1218
18. Sreedevi G, Prasada Rao DRV, Makinde OD, Venkata Ramana Reddy G (2017) Soret and
Dufour effects on MHD flow with heat and mass transfer past a permeable stretching sheet in
presence of thermal radiation. Ind J Pure Appl Phys 55:551–563
19. Ziaei-Rad M, Saeedan M, Afshari E (2016) Simulation and prediction of MHD dissipative
nanofluid flow on a permeable stretching surface using artificial neural network. Appl Therm
Eng 99:373–382
20. Aminian A (2016) Predicting the effective thermal conductivity of nanofluids for intensification
of heat transfer using artificial neural network. Powder Technol 301:288–309
21. Reddy PBA, Das R (2016) Estimation of MHD boundary layer slip flow over a permeable
stretching cylinder in the presence of chemical reaction through numerical and ANN modeling.
Eng Sci Technol Int J 19:1108–1116
22. Tafarroj MM, Mahian O, Kasaeian A, Sakamatapan K, Dalkilic AS, Wongwises S (2017) Artifi-
cial neural network modeling of nanofluid flow in a microchannel heat sink using experimental
data. Int Commun Heat Mass Transf 86:25–31
23. Tripathy UK, Patel SM (2017) Solution of flow and heat transfer of an elastic-viscous liquid
between two co-axial cylinders by artificial neural network. Int J Res Sci Innov 4(11):20–27
24. Maddah H, Ghasemi N (2017) Experimental evaluation of heat transfer efficiency of nanofluid
in a double pipe heat exchanger and prediction of experimental results using artificial neural
networks. Heat Mass Transf 53:3459–3472
25. Haykin S (2004) Neural networks: a comprehensive foundation, 2nd edn. MacMillan
Design of Macro-rough Surface and Its
Influence on Side Wall Heated Square
Enclosure
Ashwin Mahendra and Rajendran Senthil kumar
Abstract In the power electronics sector, silicon carbide devices are operational
at high junction and case temperatures. Drastic reduction in the size of electronic
components including heat sink has been achieved by using silicon carbide rather than
silicon. Depending upon the application, the challenges in thermal management can
be addressed, preferably passive cooling. If electronic components and its casing
have higher operating temperatures, the analysis of natural convection in power
electronics is more important. Hence, the present study elucidates the influence of
macro-rough surface on natural convection. In this study, a square enclosure with
and without macro-roughness have been modelled by maintaining top and bottom
walls as adiabatic. The parameters varied are Rayleigh number, number of roughness
elements and its thickness. The fluid flow and heat transfer characteristics have
been explained with the help of velocity and temperature contours. From the results
obtained, we have concluded that the increase in roughness elements and its thickness
have reduced convection near the roughness elements.
Keywords Natural convection · Laminar flow · Enclosures · Macro-roughness

elements
1 Introduction
Natural convection inside cavities or enclosures has received immense attention by

researchers like Ostarch et al. [1] owing to its plethora of applications in different
areas like electronic equipment, solar collectors, nuclear reactors, energy storage, fire
control, etc. The buoyancy-driven flows and heat transfer taking place inside these
enclosures were the basis of most studies. The fluid layer which is closest to the wall
is stationary compared to the successive layers, and hence, heat transfer here takes
place by conduction. For the layers in motion, convection is responsible for heat
transfer. The heated fluid experiences a decrease in its density. The lighter fluid then
A. Mahendra · R. S. kumar (B)

Department of Mechanical Engineering, SRMIST, Chennai, India

https://doi.org/10.1007/978-981-15-4488-0_37
436 A. Mahendra and R. S. kumar
moves upwards, and as its temperature decreases, it moves downwards resulting in

the recirculation of fluid inside the enclosure. Advancements in the study involved the
attachment of projections or roughness elements on either wall to modify the average
Nusselt number variation in the enclosure [2–8]. The flow physics and transfer of
heat within the enclosure are interesting and also need to be well understood. Hence,
the present study computes the influence of macro-rough surface in heat dissipation.
2 Literature Survey
Buoyancy-driven convection inside enclosures with heated and cooled walls has
been the basic model in various industrial applications as explained by Ostarch et al.
[1]. In many studies, the phenomenon was analysed inside square enclosures by De
Vahl Davis [9], Nag et al. [2], Shi and Khodadadi [3] and Elatar et al. [4]. De Vahl
Davis has done a benchmark study for natural convection inside square cavities by
comparing the results obtained by various researchers. Later, researchers introduced
roughness elements into the geometry. These roughness elements increase the surface
area available for transfer of heat. Nag et al. [2] studied square cavity with a single
plate on the hot vertical wall to study natural convection inside and concluded that
the colder wall has Nusselt number greater than no fin model. Bilgen [5] studied
a similar model in which the partition was fixed on the horizontal adiabatic wall.
He claimed that heat transfer reduced if double partitions were introduced. Also,
decreasing the aspect ratio or moving the partition farther away from the hot wall
reduced heat transfer. Bilgen [6] studied the influence of a thin fin on hot wall on
natural convection in square cavity. Based on his findings, as the fin length increases,
the Nusselt number decreases. Elatar et al. [4] has done a similar study with a thick
wall instead of a thin wall and concluded that the thickness of fin has minimal to no
effect on the rate of heat transfer for Rk = 10–1000. The increasing conductivity ratio
improves heat transfer (10 ≤ Rk ≤ 1000). Dindarloo et al. [7] carried out a study
of the effect of fin thickness and groove depth on the fin attached to the hot wall to
reduce the heat transfer. His results conclude that the optimal fin length decreases with
increasing conductivity ratio. Also, a thick fin with grooves reduces transfer of heat
than a straight fin. Hasnaoui et al. [8] studied the natural convection inside rectangular
enclosures with adiabatic roughness elements on the heated wall. His findings show
that for tall cavities having roughness elements with a dimensionless fin length of
25 and 50%, the heat transfer takes place by conduction alone. The present survey
has demonstrated the scope for a parametric study based on an increasing number of
roughness elements and fin thickness.
Design of Macro-rough Surface and Its Influence … 437
Fig. 1 Schematic Adiabatic Wall

representation of the given
problem
H=5
g
Adiabatic Wall
3 Problem Statement
A two-dimensional square enclosure is modelled for simulation. The left wall is kept
at a constant temperature (Th ) greater than that at the right wall (Tc ), and both the
horizontal (top and bottom) walls are adiabatic (insulated). The length of a side of
the enclosure is (H). Roughness elements of different thicknesses are introduced
on the left wall. The number of roughness elements ranges from 1 to 9, and the
length (L f ) of the roughness element is fixed at 20% of wall length. The boundary
conditions of the roughness elements are fixed to be the same as that of the left wall.
The Prandtl number is 0.71. The fluid inside the enclosure is incompressible and
obeys Boussinesq approximation. The simulation was done for Rayleigh numbers
ranging from 103 to 106 (Fig. 1).
4 Governing Equations, Boundary Conditions and Solution

Procedure
The dimensionless form of governing equations is obtained by introducing the

following terms into the traditional equation:
x y u Ll vL l
X= ,Y = , U = ,V =
Ll Ll α α
pL l2 (T − Tc )
P= , θ= (1)
ρα 2 (Th − Tc )
where the velocities in the x and y directions are denoted by ‘u’ and ‘v’, respectively,
and temperature by ‘T ’. ‘P’ stands for reduced pressure. ‘ρ’ denotes the density, and
α denotes the thermal diffusivity of the fluid.
On introducing these terms into the continuity and momentum equations and also
incorporating Boussinesq approximation, we obtain the final equations as seen in the
study by Elatar et al. [4]:
∂U ∂V
+ = 0, (2)
∂X ∂Y
2
∂U ∂U ∂ U ∂ 2U ∂P
U +V = Pr + − (3)
∂X ∂Y ∂ X2 ∂Y 2 ∂X
2
∂V ∂V ∂ V ∂2V ∂P
V +V = Pr + − + Ra ∗ Pr ∗θ, (4)
∂X ∂Y ∂X 2 ∂Y 2 ∂Y
∂θ ∂θ ∂ 2θ ∂ 2θ
U +V = + (5)
∂X ∂Y ∂ X2 ∂Y 2
gβ(Th − Tc )L 3 v
Ra = , Pr = (6)
vα α
Ra = Rayleigh number and Pr = Prandtl number.
where kinematic viscosity and thermal expansion coefficient of the fluid are
denoted by v and β, respectively.
The boundary conditions for the given model are as follows
On the left wall, U = 0, V = 0 and º = 1

On the right wall, U = 0, V = 0 and º = 0
On the top and bottom wall, U = 0, V = 0 and dY
dθ
= 0.
On the roughness elements, U = 0, V = 0 and º = θw
where θw is the temperature of the respective wall.

The problem was modelled and meshed using ANSYS Workbench, and the sim-
ulations were carried out using ANSYS Fluent-18. The convergence criteria for the
continuity residual are set as 1.E−06. The Boussinesq approximation is used for
fluid density variation at various temperatures.
5 Grid Independence
The optimum grid size was determined in the grid independence study carried out
by Shi and Khodadadi [3]. A grid size of 120 × 120 was seen to provide the most
(a) Plain Enclosure (b) Enclosure with projections
Fig. 2 Computational domain with meshing
accurate results. The same has been adopted in this paper, and the results of the
simulation have been validated with their results as can be seen in further sections
of this paper (Fig. 2).
6.1 Validation of Computation
A total of 84 cases were simulated to analyse the heat transfer and flow characteristics
of the problem. This involved studying the effect of macro-roughness on the hot wall
(0, 1, 3, 5, 7 and 9 roughness elements) at different thicknesses (2, 4 and 10%). The
Rayleigh number ranges from 103 to 106 . The element’s length was fixed at 20%
of the wall’s length. The present computational results have been validated with the
literature by De Vahl Davis [9], Nag et al. [2], Shi and Khodadadi [3] and Elatar et al.
[4].
Table 1 explains the accuracy of the present study by comparing the average
Nusselt number obtained on the cold wall at different Rayleigh numbers with previous
studies. It shows that the maximum variation is 3.2% for Ra = 103 1.6% for Ra =
104 , 3.9% for Ra = 105 and 3% for Ra = 106 which are all within the acceptable
limits.
Furthermore, the average Nusselts number was validated with that of Nag et al.
[2] for square enclosure with a thick fin attached to the vertical hot wall as tabulated
in Table 2. The fin length (L l ) is fixed at 0.2, and the dimensionless conductivity ratio
Table 1 Nusselt number validation on hot or cold wall

Ra 103 104 105 106
De Vahl Davis [9] 1.118 2.243 4.519 8.800
Nag et al. [2] – 2.240 4.510 8.820
Shi and Khodadadi [3] – 2.247 4.532 8.893
Elatar et al. [4] – 2.234 4.517 8.948
Present 1.154 2.210 4.686 8.679
Table 2 Validation of
Non-dimensional thickness 0.02 0.04 0.1
Nusselt number obtained for
different non-dimensional Nag et al. [2] 8.861 8.888 9.033
thicknesses at Ra = 106 Present 9.110 9.160 9.394
(Rk ) is 7750. The table shows that the maximum variation is 3.9% for a thickness of
10%, and hence, the accuracy of the program holds well for further studies.
6.2 Velocity and Temperature Contours
The velocity profiles and isotherms have been presented for increasing Rayleigh
numbers in enclosures without any roughness element in Fig. 3. The figure explains
the improvement in motion due to the increase in Rayleigh number. The velocity
magnitude was seen to increase from 0.00182 ms−1 for Ra = 103 to 0.285 ms−1
for Ra = 106 . As density decreases, the fluid along hot wall rises, and as density
increases, the fluid along cold wall falls. This results in the formation of a clockwise
vortex in the enclosure. For lower Rayleigh number, the boundary layer influence is
more leading to the formation of eddies. This is due to the interaction of the fluid
subjected to vortex flow and the fluid under the influence of boundary flow. As the
Rayleigh number increases, this interaction is seen to decrease, and the eddies are
suppressed and confined near the thermal walls. With the increase in temperature,
the viscosity decreases, thereby reducing the boundary layer influence. At Ra =
106 , the velocity is maximum, and the effect of the boundary layer is overshadowed
by the vorticity. All the cases are seen to be unicellular, but for lower Ra values,
the vorticity is not uniform and the shape is distorted. But, at Ra = 106 , uniform
unicellular vorticity is generated. The temperature contours capture the flow guided
heat transfer data as shown with the isotherms being densely packed near the walls
at Ra = 105 . As the vorticity increases, the bulk fluid mixing increases which aid in
improving convection.
Figure 4 shows a comparison of the enclosure with nine roughness elements for
all four Ra values. For low Ra values, the velocity contours follow the same pattern
as that of the plain enclosure. As the Rayleigh number increases, the eddies are
300.3939K
300K
Ra= Vorticity = 0.430/s
(a) Velocity = 0.00182m/s (a’)
303.939K
300K
(b) Velocity = 0.00770m/s (b’)

339.39K
300K
(c) Velocity = 0.0187m/s (c’)

300K
.693.9K
(d) Velocity = 0.285m/s (d’)
Fig. 3 Velocity and temperature contours for plain enclosures at different Ra values
300.3939K
300K
(a) Velocity = 0.00132m/s (a’)
303.939K
300K
(b) Velocity = 0.00683m/s (b’)

339.39K
300K
(c) Velocity = 0.0163m/s (c’)

300K

693.9K
(d) Velocity = 0.0337m/s (d’)
Fig. 4 Velocity and temperature contours for nine-finned enclosure at different Ra values
suppressed. But, for Ra = 106 a uniform vortex as seen in the plain enclosure cannot
be obtained. Due to the resistance offered by the roughness elements to fluid flow,
the vorticity is still in the transition phase, and the effect of eddies are not suppressed
completely. Also, the fluid starts to penetrate between the lower roughness elements
as the Rayleigh number increases, due to momentum which affects the flow. The
related heat transfer data has been captured in the temperature contours as shown.
It is inferred that the isotherms become more packed as the Ra value increases,
especially near the cold wall where the thermal boundary layer becomes denser and
also below the lowest roughness element. Therefore, increasing the Rayleigh number
increases the isotherm density in these regions.
Figure 5 clearly visualizes the fluid flow at Ra = 106 . The velocity contours
have been captured for enclosures with an increasing macro-roughness elements.
Here, the flow physics have been captured between the roughness elements unlike
the cases with lower Ra values. The case shows the circulation of fluid between the
roughness elements which is seen to decrease with increasing number of roughness
elements. Therefore, as the roughness increases, the stagnation of fluid within the
roughness elements increases resulting in the reduction of convection in the region.
The roughness elements also resist the flow inside the enclosure. The sharp edges
of the roughness elements lead to the local recirculation of the fluid in the region.
Hence, convection is enhanced at the roughness element’s tip. This is clearly captured
in the temperature contours as well. The isotherms show that the isotherm density
between the roughness elements decreases with increasing roughness. This explains
the prevalence of conduction in these regions over convection. The fluid gets stagnant
here due to the flow restriction offered by the roughness elements resulting in reduced
convection here.
Table 3 illustrates the average Nusselt number in a square enclosure at different
walls with increasing macro-roughness on the hotter wall. It shows that the increase
in macro-roughness reduces the hot wall Nusselt number, and hence, convection
is reduced. This can be explained by the stagnation of fluid in the region between
the roughness elements. Hence, heat transfer here takes place predominantly by
conduction.
Figure 6 shows the velocity contours for an enclosure with five roughness elements
for different thicknesses at different Rayleigh numbers. The figure shows that for a
given Ra value, the penetration of the fluid between the roughness elements reduces
as the thickness increases. For minimum thickness, the sharp tip induces local recir-
culation of fluid, thus enabling it to partially enter the region between the roughness
elements. For higher Ra values, as the velocity of the fluid increases, the high momen-
tum flow allows recirculation in the regions between lower roughness elements. In
the successive roughness elements, local recirculation prevails. As the thickness
increases, the local recirculation decreases. Thus, convection is adversely affected.
But, the comparative increase in surface area available for convection compensates
this decrease.
Figure 7 plots the Nusselt number for different thicknesses in a square enclosure
with a single roughness element with respect to the Rayleigh number. It shows that
though the variations are minor, the Nusselt number increases with thickness owing
693.9K
300K
Vorticity = 26.153/s
(a) Velocity = 0.0413m/s (a) 1 Roughness Element
693.9K
300K
(b) Velocity = 0.0390m/s (b) 3 Roughness Elements

693.9K
300K
(c) Velocity = 0.0357m/s (c) 5 Roughness Elements

300K
693.9K
(d) Velocity = 0.0352m/s (d) 7 Roughness Elements
Fig. 5 Velocity and temperature contours for an increasing number of roughness elements at Ra
= 106
Table 3 Nusselt number

Present Cold wall Hot wall Fin wall
values for different walls in
the enclosure with different No roughness element 8.678 8.678 –
number of roughness 1 roughness element 9.110 7.376 4.496
elements at Ra = 106
3 roughness elements 8.716 3.899 4.008
to an increase in the overall surface area available for heat transfer, thereby improving
convection.
Figure 8 shows the variations in Nu with respect to Ra. The graph on the left
denotes the Nu on the hot wall and the one on the left denotes the Nu on the cold wall.
The Nusselt number decreases on both the walls as the macro-roughness increases,
but the decrease is more significant on the hot wall. As the number of roughness
elements increases, the stagnation of fluid between the roughness elements (Fig. 5)
and along the hot wall also increases resulting in transfer of heat by conduction.
The introduction of roughness elements has seen to disrupt the vortex circulation
strength.
7 Conclusion
The flow and heat transfer within the square enclosure by keeping adiabatic horizontal
walls with and without macro-roughness elements have been studied for different
Rayleigh numbers (103 , 104 , 105 , 106 ) under laminar flow conditions. The velocity
and temperature contours expose fluid flow behaviour and transfer of heat within the
cavity and also the following have been concluded.
1. As the Rayleigh number increases, the average Nusselt number on the wall also
increases. This is because, the Nusselt number increase improves the movement
of fluid inside the enclosure, thereby enhancing convection.
2. As the thickness of roughness elements increases, the average Nusselt number
on the cold wall increases. This can be understood by the increase in surface area
of the roughness elements as the thickness increases. The fluid, therefore, has
a greater area to interact with the roughness elements, and hence, convection is
improved.
3. The convection near macro-roughness elements improves with increase in its
thickness and weakens with increasing roughness elements.
4. Increasing the roughness elements for the given Ra, increases the Nusselt number
owing to reasons similar as above.
Ra= Vorticity = 0.352/s Vorticity = 0.360/s Vorticity = 0.383/s
(a) Velocity = 0.00132m/s (a’) Velocity = 0.00134m/s (a’’) Velocity = 0.0014m/s
(b) Velocity = 0.00671m/s (b’) Velocity = 0.00688m/s (b’’) Velocity = 0.00698m/s
(c) Velocity = 0.0159m/s (c’) Velocity = 0.0164m/s (c’’) Velocity = 0.0176m/s
(d) Velocity = 0.0357m/s (d’) Velocity = 0.0356m/s (d’’) Velocity = 0.0369m/s
Fig. 6 Velocity contours of an enclosure with five roughness elements for different thicknesses at
different Ra values
Fig. 7 Variation of Nusselt number in one roughness element enclosure with different thicknesses
with Ra
Fig. 8 Variation of Nusselt number in a cavity with different number of roughness elements with
Ra
Acknowledgements We would like to thank the product development laboratory authorities and
SRM Institute of Science and Technology for providing us with the resources to make this possible.
We would also take this opportunity to thank SSN College of Engineering for providing us with a
platform to present our ideas and findings.
References
1. Ostrach S (1988) Natural convection in enclosures. ASME J Heat Transf 110:1175–1190

2. Nag A, Sarkar A, Sastri VMK (1994) Effect of thick horizontal partial partition attached to one
of the active walls of a differentially heated square cavity. Numer Heat Transf Part A:611–625
3. Shi X, Khodadadi JM (2002) Laminar fluid flow and heat transfer in a lid-driven cavity due to
a thin fin. ASME J Heat Transf 124:1056–1063
4. Elatar A, Teamah MA, Hassab MA (2016) Numerical study of laminar natural convection inside
square enclosure with single horizontal fin. Int J Therm Sci 99:41–51
5. Bilgen E (2002) Natural convection in enclosures with partial partitions. Renew Energy 26:257–
270
6. Bilgen E (2005) Natural convection in cavities with a thin fin on the hot wall. Int J Heat Mass
Transf 48:3493–3505
7. Rafie Dindarloo M, Payan S (2019) Effect of fin thickness, grooves depth, and fin attachment
angle to the hot wall on maximum heat transfer reduction in a square enclosure. Int J Thermal
Sci 136:473–490
8. Hasnaoui M, Vasseur P, Bilgen E (1992) Natural convection in rectangular enclosures with
adiabatic fins attached on the heated wall. Warme- und Stoffubertragung 27:357–368
9. De Vahl Davis G (1983) Natural convection of air in a square cavity: a bench mark numerical
solution. Int J Numer Methods Fluids 3:249–264
Performance Enhancement of a Savonius
Vertical Axis Wind Turbine
with Bio-Inspired Design Modifications
S. M. Hasan Fayaz, Uditya Tyagi, Apurva Jain, and Nishant Mishra
Abstract One of the prime commodities in modern civilization is energy. The

amount of energy consumption has become the indicator for the standard of liv-
ing and the degree of industrialization. People use fossil fuels to meet nearly all of
their energy needs, such as powering vehicles, producing electricity for light and
heat and running factories, thus greatly exhausting the fossil fuel reserves along with
polluting the environment with greenhouse gases. Renewable energy sources are
viable alternatives, and among the various types of renewable energy sources avail-
able, wind energy is the sector which has a lot of untapped potential. Our objective
is to improve the efficiency of a Savonius-type vertical-axis wind turbine (VAWT)
which currently has the least efficiency among existing wind turbine designs. Savo-
nius turbines have a very compact structure and can run at low wind speeds which
are desirable characteristics for commercial-scale power production. This research
paper focuses on improving the efficiency of Savonius wind turbine. Since Savonius-
type wind turbines are drag-based wind turbines, we need to reduce the impulsive
force acting on the negative face in order to increase the drag difference between
the positive and negative side of the rotor blades thereby improving the efficiency
of the turbine. Our proposition for attaining higher efficiency is by incorporating the
concept of dimples (inspired from golf ball) on the negative side of the rotor blade
and tubercles (inspired from whales) on the leading edge of the rotor blade. Dimples
reduce the amount of wake region in the case of golf ball by increasing turbulence
which is one of the desirable characteristics in the case of Savonius wind turbine.
Tubercles help in reducing the wake region behind the rotor blades by increasing the
turbulence of air near the surface of the rotor thereby improving the efficiency of the
turbine.
Keywords VAWT · Savonius · Drag based · Dimples · Tubercles · Wake region
S. M. Hasan Fayaz · U. Tyagi · A. Jain · N. Mishra (B)

Shiv Nadar University, Greater Noida, India

https://doi.org/10.1007/978-981-15-4488-0_38
450 S. M. Hasan Fayaz et al.
Nomenclature
VAWT Vertical-axis wind turbines

Pω Total wind power available
Po Power output by the wind turbine
Cp Coefficient of power
A Swept area of the turbine
ω Angular velocity of turbine in rad/sec
ρ Density of the air
λ Tip-speed ratio
R Radius of the turbine
H Height of turbine
D Diameter of the turbine
rpm Turbine speed in revolutions per minute
d Diameter of turbine blades
1 Introduction
Power production in an efficient way without affecting the environment has been a
growing concern around the globe. Various means of eco-friendly methods of power
production have been devised. Some of the eco-friendly means of power production
are harnessing solar energy, wind energy, hydro power, biomass, etc. Wind energy is
a field having a lot of scope for improvement of efficiency. Wind power is also one
of the most economical sources of renewable energy, land-based utility-scale wind
power is the most economical source of renewable energy as on date [1, 2]. Wind
is an environment-friendly source of energy that has got huge potential to satisfy
energy needs for people and also to mitigate the climate change from greenhouse
gases emitted by the burning of fossil fuels.
Wind turbines can be classified into two types based on the orientation of the
shaft connecting the turbine with the generator, namely horizontal-axis wind turbine
(HAWT) and vertical-axis wind turbine (VAWT). Most of the commercial-scale
power produced by tapping wind energy is based on horizontal-axis wind turbine
technology. The horizontal-axis wind turbines are more efficient as compared to
that of vertical-axis wind turbines [3]. HAWT has an efficiency range of 35–45%,
whereas VAWT has a comparatively lower efficiency. VAWT has a greater scope
of improvement with respect to its design and material in order to increase its effi-
ciency, so that it can be used for commercial-scale power production. VAWT rotors
have different types, such as Savonius rotor and (eggbeater) Darrieus, or H-Darrieus
rotor, out of which Savonius is the turbine with least efficiency. Savonius turbines
have a very compact structure and can run at low wind speeds which are desirable
characteristics for commercial-scale power production. As a main target application,
Performance Enhancement of a Savonius Vertical Axis Wind Turbine … 451
optimized Savonius wind turbines would be perfectly suited for electricity produc-
tion at the level of individual buildings with a flat roof or for commercial-scale power
productions on both onshore and offshore platforms.
The objective is to increase the efficiency of the Savonius-type vertical-axis wind
turbine by introducing and assessing the effect of arrays of dimples and tubercles.
It has been proven that dimple structures on a golf ball reduce drag on the surface
enabling it to cover longer flight distance [4–6]. The magnitude of drag force on a
moving ball is largely affected by the size of wake region behind the moving ball.
Decreasing the wake region by increasing turbulence on the surface of the ball which
in turn delays the separation of air from the ball leads to an increase in total pressure
behind the ball. Therefore, the pressure difference between the front of the ball and
behind the ball will decrease resulting in decrease of drag force constituted by air.
These researches provided us with key insights into the analysis of the dimples and
also led to contemplate that if the same concept can be applied to reduce the drag
on the negative blade of the Savonius turbine. Spherical dimples were created on the
negative surface and the turbine’s performance was compared with and without the
use of dimples. Tubercles were introduced as a design modification in order to reduce
the negative torque acting on the wind turbine. Concept of tubercles was inspired
from the whales wherein tubercles are used to increase lift thereby helping them to
manoeuvre through steep climbs.
Fig. 1 a Comparing golf ball and normal ball [6], b blades of Savonius turbine
2 Methodology
2.1 CAD Model
Once the design parameters were decided, CAD models of the two configurations
of the Savonius wind turbine were made using SolidWorks. Individual components
were made and were assembled in order to design the two-stage Savonius turbine. The
dimensions of rotor blades were determined by the aspect ratio chosen according to
the literature survey [7–10]. Rotor blades were made (using extrude option), and dim-
pling was done according to requirement (using the revolved cut tool). Continuous
triangular tubercles of base 10 mm were made using linear pattern.
Dimensions of the two-stage turbine are:
End plate thickness = 4 mm
End plate diameter = 152 mm
Rotor inner diameter = 75 mm
Rotor outer diameter = 85 mm
Rotor length = 120 mm
Overlap length, e = 13.5 mm
Dimple diameter = 4 mm
Dimple depth = 1.5 mm
Number of dimple columns = 13
Number of dimple rows = 11
Base length of tubercles (isosceles triangle) = 10 mm
Number of tubercles = 11 (Fig. 2).
Fig. 2 CAD models of Savonius wind turbine with dimple and tubercle
2.2 3D Printing and Assembling the Rotor Blades
The final models of two different types of rotor blades (with and without dimples and
tubercles) were printed using a 3D printer. The specifications of the printing process
are:
Nozzle diameter = 0.4 mm
Fill density = 100%
Fill rate = 100
Platform adhesion type = Brim
Print temperature = 210 °C.
The printed blades were assembled according to the overlap ratio chosen from the
literature survey [7–10]. The rotor blades were assembled with the end plates using
bondtite. Both types of turbines were assembled as a two-stage turbine with a phase
difference of 90° which was observed from the literature review [11, 12].
2.3 Experimental Set-Up
The experiment comprises of two different models, namely:

1. Double stage plain turbine (smooth)
2. Double stage turbine with dimples and tubercles.
The experimental study basically comprises of observation and comparison of the
exit wind speed, rpm of the turbine and the power generated by the two different
turbines. The wind speed is measured using an anemometer by placing it at various
points on the path traced by the wind entering the turbine. Tachometer is used to
measure the rpm of the turbine and the power generated by the turbine can be observed
from the multimeter connected to the generator terminals.
The schematic diagram of the test rig used for the experimental study is shown in
Fig 3.
3 Mathematical Formulations
Maximum amount of power, P, that can be harnessed from wind in an ideal case is
ρ × A × V3
Pω = (1)
2
where
ρ = density of the air
Fig. 3 a Schematic diagram of test rig (above), b wind tunnel (below)
A = swept area of the turbine

V = wind speed.
According to Betz law, the amount of kinetic energy of the wind that can be
harnessed for power production can never exceed 59.3%. Power coefficient C p is
defined as the ratio of the cumulative power output Po by a wind turbine to the total
power Pω available to the wind turbine
Po
Cp = (2)
Pω
So, no more than 59% of the energy carried by wind can be extracted by a wind
turbine. The value of C p is unique for each turbine. Value of C p varies based on
factors like strength of the turbine, weight of the turbine, friction involved due to
generator and gearbox, surrounding air conditions, etc. Therefore, maximum power
that can be extracted from the wind turbine considering the aerodynamic constraints
can be found by the equation
ρ × Cp × A × V 3
Pω = (3)
2
C p also varies with respect to the rotational speeds of the turbine. Hence, C p is a
function of tip-speed ratio (TSR) λ as well, which is defined as
Tip Speed of the Blade

λ= (4)
Wind Speed
ω×R
λ= (5)
V
Power output Po can be found experimentally using,
Po = T × ω (6)
Here,
T = mechanical torque
ω = angular velocity of the rotor blades.
2π N
ω= (7)
60
Here,
ω = rotational velocity of the turbine
R = turbine radius
V = wind Speed
N = RPM of turbine.
4 Comparison Between Savonius Wind Turbines

with and Without Modifications (Dimples and Tubercles)
The following data were recorded during the experiment performed on the plain and
modified wind turbine: (Figs. 4 and 5).
5 Conclusion
In this experimental publication, the concepts of dimples and tubercles have been
introduced on an existing design of Savonius wind turbine. Plain turbine and the
modified turbines were tested under the same wind speed in a wind tunnel, and the
results were compared.
Following can be concluded from the experimental results:
Fig. 4 Variation in efficiency with wind speed for turbines with and without design modifications
Fig. 5 Variation in efficiency with rpm for turbines with and without design modifications
It can be observed that the modified turbine (with dimples and tubercles) generates
more power at the same wind speed as compared to the plain turbine.
Acknowledgements The authors would like to thank the laboratory assistants, Sanjay Singh, Amit
Kumar, Sunil Gupta and Vishal Gaur for their assistance in establishing the experimental rig. Authors
would also thank Shiv Nadar University for their support in carrying out the experiment.
References
1. Power Technology, https://www.power-technology.com/features/cheapest-renewable-energy-

sources/. Last accessed 13 Feb 2019
2. OIL PRICE, https://oilprice.com/Latest-Energy-News/World-News/Wind-Solar-Are-Now-

The-Cheapest-Sources-Of-Power-Generation.html. Last accessed 14 Feb 2019
3. Johari MK, Jalil MAA, Shariff MF (2018) Comparison of horizontal axis wind turbine (HAWT)
and vertical axis wind turbine (VAWT). 7:74–80. https://doi.org/10.14419/ijet.v7i4.13.21333
4. Alam F, Steiner T, Chowdhury H, Subic A (2011) A study of golf ball aerodynamic drag. In:
5th Asia-Pacific congress on sports technology (APCST) Procedia Engineering, vol 13, pp
226–231, https://doi.org/10.1016/j.proeng.2011.05.077
5. Alam F, Chowdhury H, Moria A, Subic A (2010) A comparative study of golf ball aerody-
namics. In: Conference: 17th Australasian fluid mechanics conference 2010, Auckland, New
Zealand, pp 599–602, ISBN: 978-1-61782-913
6. Aerospace Web, http://www.aerospaceweb.org/question/aerodynamics/sports/sphere-flow-
comparison.jpg. Last accessed 05 May 2019
7. Gupta AK (2015) Efficient wind energy conversion: evolution to modern design. ASME J
Energy Resourc Technol 137(5):051201–051201-10, JERT-14-1427. https://doi.org/10.1115/
1.4030109
8. Alom N, Saha UK (2018) Four decades of research into the augmentation techniques of Savo-
nius wind turbine rotor. ASME J Energy Resourc Technol 140(5), 050801, 14 pages. Paper
No: JERT-17-1620. https://doi.org/10.1115/1.4038785
9. Mari M, Venturini M, Beyene A (2017) A novel geometry for vertical axis wind turbines based
on the Savonius concept. ASME J Energy Resourc Technol 139(6):061202–061202-9. https://
doi.org/10.1115/1.4036964
10. Mahmoud NH, El-Haroun AA, Wahba E, Nasef MH (2012) An experimental study on
improvement of Savonius rotor performance. Alexandria Eng J 51(1):19–25
11. Rosmina N, Jauharib AS, Mustaamalc AH, Husind F, Hassane MY (2015) Experimental study
for the single-stage and double-stage two-bladed Savonius Micro-sized turbine for rain water
harvesting (RWH) System. In: 2nd international conference on sustainable energy engineering
and application, ICSEEA Energy Procedia, 68, pp 274–281
12. Kamoji MA, Kedare SB, Prabhu SV (2008) Experimental investigations on single stage, two
stage and three stage conventional Savonius rotor. Int J Energy Res 32(10):877–895. https://
doi.org/10.1002/er.1399
Experimental Investigation of Heat
Treatment Processes on Dissimilar
IS2602-EN9 MIG Welded Joint
Mallapuram Bala Chennaiah , Podaralla Nanda Kumar ,

and Karanam Prahlada Rao
Abstract The present investigation, influence of different heat treatment

processes which will yield better tensile strength and microstructures of 8-mm-thick
plates metal inert gas welded IS2062-EN9 joint. Mechanical properties of the joint
increased with annealing heat treatment processes, higher grain size was measured at
welded zone, and it decreased. It is found that the joint fabricated at a low heat input
condition showing excellent mechanical and metallurgical properties. After anneal-
ing heat treatment process, tensile strength improved around 14.52%, the pearlite
phase increased from 36.75 to 47.15%, and ferrite phase has decreased from 59.36
to 50.61% and other decreased from 3.88 to 2.24%. The tensile strength is improved
due to grain size that was observed to be 7.5 before heat treatment and 9 after
annealing heat treatment, and ferrite phase is transformed into pearlite phase.
Keywords Metal inert gas welding · Low heat input · High heat input ·
Annealing · Normalizing · Tempering · Mechanical and metallurgical properties
1 Introduction
Steel is mostly used material; reason is steels dominance is usually considered to be

the abundance of iron ore and the ease by which it can be refined from ore, neither
of these is necessarily correct. Steel is such an important material because of its
tremendous flexibility in fabrication processes and metal working processes. Two
dissimilar plates were selected (IS2602–EN9) for MIG welding and their industry
importance and wide application, IS2062, ISC55/EN9 Steel plates having 8 mm
thickness were selected as base material in our work. Chemical composition and
M. Bala Chennaiah (B)

V.R. Siddhartha Engineering College, Vijayawada 520007, India
P. Nanda Kumar
N.B.K.R Institute of Science & Technology, Nellore, India
K. Prahlada Rao
Jawaharlal Nehru Technological University, Anantapur, Ananthapuram, India
https://doi.org/10.1007/978-981-15-4488-0_39
460 M. Bala Chennaiah et al.
mechanical properties of base material are given in Table 1. The carbon equivalents
of each plate IS2602 and EN9 were 0.54 and 0.96, respectively [1].
One of the steel plates is hypo-eutectoid while other is hyper-eutectoid with
respect to carbon equivalent. With each heat input condition, the four weldments
were prepared. One of the finished and defect-free joints is shown in Figs. 1 and 2.
Al-Mazrouee et al. [2], the main role of post-weld heat treatment is to relive the
residual stress by allowing plastic flow at the stress-relieving temperature, as a result
of improving the properties of the element. Post-weld heat treatment is mandatory
Table 1 Chemical compositions of various materials

S. Name of Composition in % of weight
No. material Carbon Manganese Silicon Chromium Sulphur Phosphorus
1 Mild steel (IS 0.2 – 0.25 – 0.05 0.05
2602)
2 Medium-carbon 0.55 0.15 0.65 – 0.06 0.06
alloy steel
(EN-9)
Fig. 1 Finished joint
Fig. 2 Defect-free joint

Experimental Investigation of Heat Treatment Processes … 461
after weld fabrication specially used in various industries for manufacturing the
components like pressure vessel, boiler, etc. The weldments were undergone with die
penetrant and X-ray tests to find the surface and subsurface defects before conducting
destructive tests. One of the four weldments was used to investigate the mechanical
and metallurgical properties without heat treatment. The remaining three weldments
were used to investigate the mechanical and metallurgical properties after the heat
treatment processes, viz. annealing, normalizing and tempering. The results were
tabulated and compared with the results of weldment without heat treatment.
2 Experimental Procedure
2.1 Base Material Selection
Firstly, the materials which are of the size (250 × 100 × 8) mm are to be taken.
Select the universal milling machine for weld groove bevelling and deburr the plates.
For bevelling, we have to turn the facing head up to 300–450, and feed is given
slowly. So, the edge preparation is done to do welding as single V butt joint. After
edge preparation, we have to select the combinations of material pair for further
purpose. Here, the combination of the weld plates is selected in such a way that for
the first joint low- and medium-carbon steel plates. Here, we consider the metal inert
gas (MIG) welding, where as MIG wire (copper-coated mild steel) with diameter
1.2 mm is taken. The importance of copper coating on mild steel is used to prevent
rust and also current is passed easily. Anti-spatter spray is sprayed on the wire for
easy clean up after a flux core MIG welder. Then, welding is done for five passes
with respect to voltage, current and welding speed which is considered.
2.2 Importance of Post-weld Heat Treatment
After having completed with the welding for the similar/dissimilar metal plates,
the post-weld heat treatment was carried out to evaluate and improve the mechan-
ical properties and microstructure of the metal plate. annealing, normalizing and
tempering post-weld heat treatment process were carried out in the experiment.
2.3 Tensile Properties
Welding process is subjected to heating and cooling leading to development of dif-

ferent microstructures in different zones, with respect to transverse direction of the
weld axis. Eight dissimilar weldments of IS2602-EN9 were prepared by using MIG
Table 2 Tensile strength results of IS2602-EN9 low heat input weldments

Type of heat Tensile strength low heat input weldment (MPa)
treatment Before heat PWHT Improvement Percent of
treatment improvement
Annealing 482.79 543.41 60.62 14.52
Normalizing 520.94 39.15 10.16
Tempering 530.96 48.17 9.05
Table 3 Tensile strength results of IS2602-EN9 high heat input weldments

Type of heat Tensile strength high heat input weldment (MPa)
treatment Before heat treatment PWHT Improvement % of improvement
Annealing 483.91 543.97 70.87 12.55
Normalizing 533.10 49.19 7.90
Tempering 527.71 43.8 9.97
welding process. Three testing samples were prepared according to ASTM E8-04
from dissimilar weldment. The tensile strengths of these three samples of each weld-
ment were measured, and their average values were tabulated (Tables 2 and 3). One
of the most significant observations with different heat treatment conditions is that
the tensile strength gives higher values with heat treatment than that without heat
treatment for both heat inputs.
2.4 Microstructural Observations
For metallographic observations before post-weld heat treatment and after post-weld
heat treatment, specimens were etched with 4% Nital for 20 s, and consequently, the
microstructures of the base, weld and the heat-affected zone were observed under
100X magnification. Heat input seriously affects the size and shape of the grains of
those metals composed with filler materials. The grains of the welded zone grow to
larger size compared with that of other zones. Ili et al. [3] studied the variations in the
structures, and grain sizes were mainly due to directional heat flow from the fusion
line. The microstructures at different zones of the weldment were taken before and
after heat treatment for further comparison.
3.1 Tensile Strength of IS2602—EN9 Dissimilar Welded

Joint
Eight dissimilar weldments of IS2602-EN9 were prepared by using MIG welding

process with low and high heat input condition, each heat input condition for four
dissimilar weldments was prepared. Three testing samples were prepared according
to ASTM E8-04 from each weldment. The tensile strengths of those three samples
of each weldment were measured, and their average values were tabulated. One
of the most significant observations with different heat treatment conditions is that
the tensile strength gives higher values with heat treatment than that without heat
treatment for both heat inputs [4].
To compare with the results, a histogram was presented as Fig. 3 which shows
that the tensile strength obtained in annealing process was effective in improvement,
compared with the other heat treatment processes. This is true for both heat input
conditions. It is also observed that high heat input conditions weldments show higher
improvements in tensile strength values in comparison with low heat input conditions.
The exception is that the tensile strength of samples subjected to tempering heat
treatment process showed betterment for low heat input.
550 550
Tensile Strength(MPa)
Tensile Strength(MPa)
540 540
530 530
520 520
510 510
500
500 490
490 480
480 470
470 460
460 450
450
Heat Treatment Processes Heat Treatment Processes
a)Low heat input b)High heat input
Fig. 3 Tensile strength of IS2602–EN9 weldment

3.2 Microstructure Analysis of IS2602—EN9 Dissimilar

Welded Joint
Microstructure observations was performed on specimens without heat treatment

and those heat treated by annealing process [5]. The metallographic observation
before post-weld heat treatment is shown in Fig. 4. This showed the microstructure
of base metal (IS2602) with equiaxed structure and uniformly distributed with very
fine strengthening precipitates. ASTM grain size of 7.5 is observed. Light etched
structures which were observed in white colour are the ferrite phase, and dark etched
structure which was pink in colour represents pearlite phase. From Fig. 4b which
indicates the microstructure of base metal IS2062 after heat treatment, the grain size
increased to 9 after heat treatment. Figure 4c, d shows that HAZ(IS2602) shows a
grain size of 7.5 without heat treatment and 9 with annealing heat treatment. Thus,
it is clearly evident that the grain size increased from 7.5 to 9 with heat treatment.
Figure 4e, f shows the microstructures in the weld zone without heat treatment
Before Heat Treatment After Heat Treatment

Grain Size Volume Fraction Grain Size Volume Fraction
IS2602-BM
(a & b)
IS2602-HAZ
(c & d)
WZ
(e&f)
EN9-HAZ
(g&h)
EN9-B(i&j)
Fig. 4 Microstructural images of EN9

and with heat treatment, respectively. In case, grain size was observed to be 7.5
before heat treatment and 9 after heat treatment. Figure 4g shows the microstruc-
ture observed in the heat-affected zone (EN9) without heat treatment. From Fig. 4h
indicates the microstructure of heat-affected zone (EN9) with heat treatment; the
microstructure observation also shows that the grain size increases to 7.5 after heat
treatment compared with a value of 8 before the heat treatment. Figure 4i shows the
microstructure of base metal (EN9) which has a uniform structure with uniformly
distributed very fine strengthening precipitates, with ASTM grain size of 7. The
light etched structures which were observed in white colour are the ferrite phase,
and dark etched structure which was grey in colour represents pearlite phase. From
Fig. 4j which indicates the microstructure of base metal (EN8) after heat treatment,
the microstructure observation also shows that the grain size increases from 7 to 7.5.
After completion of grain size measurements, the same specimens were con-
sidered for phase/volume fraction analysis. Light optical microscope (LOM) with
magnification of 100X was used for the study of phase/volume fraction analysis.
In order to study the different phases of the microstructures, colour grading was
selected. The ferrite phases were represented with dark blue colour, and the pearlite
phases were represented with grey colour. Figure 4a–j shows the microstructure
images and the corresponding phase observed along with their volume fractions of
the three principal zones (base metal, heat-affected zone and weld zone).
4 Conclusions
1. Tensile strength of low heat input dissimilar weldment IS2602-EN9 was observed
as 482.79 MPa before heat treatment. For improving this strength, we adopted
the popular heat treatment processes. With annealing heat treatment process, the
tensile strength was observed as 543.41 MPa. Increase in tensile strength was
observed to be 12.55%. After normalizing heat treatment process, tensile strength
was observed as 520.94 MPa. Thus, 7.90% improvement in tensile strength has
been observed. Even after tempering heat treatment process, the tensile strength
was increased by 9.97% i.e., to 530.96 MPa. Further among the three heat treat-
ment processes, after annealing process offered the best improvement in tensile
strength compared with remaining two heat treatment processes.
2. The microstructure of the IS2602-EN9 welded joint after heat treatment con-
sisted mainly of three phases 47.15% pearlite phase, 50.61% ferrite phase and
others 2.24%. After the heat treatment, the pearlite phase increased from 36.75
to 47.15%, ferrite phase has decreased 59.36–50.61% and other decreased from
3.88 to 2.24%. Therefore, it is understood that when the ferrite phase is reducing,
the pearlite phase is increasing, which means an account of heat treating shows
that ferrite phase is transformed into pearlite phase.
Acknowledgements I express my sincere thanks to UGC for assisting in my work through funding
Minor Research Project, MRP-6188/15 (SERO/UGC) for the purchasing the optical microscope
which helped me to carrying out my work successfully.
References
1. Ali MS (2011) Study of mechanical properties in a weld bead of dissimilar metals. Int J Eng Sci
Technol (IJEST) 3(5). ISSN: 0975-5462
2. Al-Mazrouee A, Singh Raman RK, Ibrahim RN (2005) Effect of post weld heat treatment on
the oxide scaling of Cr–Mo steel weldments. J Mater Process Technol 164–165:964–970
3. Ili N, Jovanovich MT, Todorovi M, Trtanj M, Aponji P (1999) Micro structural and mechanical
characterization of post weld heat-treated termite weld in rails. Mater Charact 43(4):243–250
4. Standard test methods for tension testing of metallic materials. Designation: (E8/E8M-09)
5. Sireesha M, Albert SK, Sundaresan S (2001) Microstructure and mechanical properties of weld
fusion zones in modified 9Cr-1 Mo Steel. J Mater Eng Perform 10:320–330
Feasibility of Al2 O3 /Water Nanofluid
in a Compact Loop Heat Pipe
Emerald Ninolin Stephen, Lazarus Godson Asirvatham,

Kandasamy Ramachandran, Arulanatham Brusly Solomon,
and Pitchaimuthu RamKumar
Abstract Effect of low volume concentration of water-based aluminium oxide

(Al2 O3 ) nanofluids on the thermal performance and entropy generation of a com-
pact loop heat pipe (CLHP) with the flat square evaporator of dimensions (34 mm ×
34 mm × 19 mm) is experimentally investigated. The heat input and volume concen-
tration are varied from 30 W to 500 W and 0.03%, 0.09% and 0.12%, respectively.
The effects of performance parameters such as heat supplied and nanoparticle con-
centration on the entropy generation, thermal resistance, evaporator and condenser
heat transfer coefficients (HTC) are analysed. Results showed that the formation of
thin porous deposition on the wick and wall plays a key role in the enhancement of
heat transfer of the CLHP. The heat transfer coefficient was found to be enhanced by
24.42% in the evaporator, and the thermal resistance was reduced by about 21.29%.
The entropy generation is also reduced by 12.36% when Al2 O3 nanofluid is used
as the working fluid. Adding a little quantity of Al2 O3 nanoparticles enhanced the
operating range of CLHP by 12% when compared with that of the base fluid.
Keywords Loop heat pipe · Heat transfer · Aluminium oxide nanoparticles
1 Introduction
Modern electronics industries are facing considerable challenges in the removal of

heat from electronic devices. Due to the reduced size, increasing processor speed and
high heat loads, these devices require efficient cooling systems. The conventional
method of cooling like simple conductive and convective cooling is not appropriate
for microelectronic devices with high heat generation. In recent years, the heat pipes
have got more attention in dissipating the heat from high heat flux devices [1]. Heat
E. N. Stephen · L. G. Asirvatham (B) · A. B. Solomon · P. RamKumar

Department of Mechanical Engineering, Karunya Institute of Technology and Sciences,
Coimbatore 641114, India
K. Ramachandran
Department of Physics, Bharathiar University, Coimbatore 641046, India

https://doi.org/10.1007/978-981-15-4488-0_40
468 E. N. Stephen et al.
pipes are also used in spacecraft’s thermal control, electronic boards, air conditioning
systems, energy collectors, chemical engineering, engine cooling, power generation,
etc. The heat pipe is capable of transmitting heat at high rates over vast distances with
very small temperature drops [2]. Different types of heat pipes like planar heat pipes,
variable conductance heat pipes, pressure-controlled heat pipes, diode heat pipes,
thermosyphons, rotating heat pipes, loop heat pipe, oscillating or pulsating heat pipe
are used for heat removal applications. Among the different types of heat pipe, loop
heat pipes (LHPs) are acknowledged as a superior heat transfer device in the domain
of thermal management as well as electronic cooling and cooling of computers [3].
The LHP is an effective, self-circulating heat transfer device based on capillarity.
Generally, LHPs offers low thermal resistance, and it has the capacity to transport
heat over a long distance with a low-temperature gradient and requires less space.
The analysis done by Maydanik et al. [4] in a 21 m long stainless steel LHP with
ammonia and nickel wick achieved the highest heat load of 1700 W (12 W/cm2 )
with the lowest resistance of 0.034 °C/W. A long LHP of 10 m long was tested
in the horizontal position using ethanol as cooling fluid. It was identified that the
maximum power could be transferred by the LHP is 340 W with a thermal resistance
of 0.11 °C/W [5]. The flat evaporator has many advantages over cylindrical, disc-
shaped evaporators, especially easy mounting, elimination of additional interface
material, increased active area, reduction in the resistance and weight of the LHPs.
Therefore, flat shape LHPs are favoured in cooling electronic equipment in ground
as well as in space applications [6–8].
Most LHPs are effectively employed in space applications, and when coming to
electronic systems, nowadays the size is minimized, power is increased, and the heat
removal becomes a major challenge. The use of miniaturized LHP with flat evapora-
tor can be one of the options to satisfy the compact enclosure of electronic component
without compromising the performance of the device. Tang et al. [9] tested experi-
mentally the start-up and thermal performance of a specially made miniature LHP
(mLHP). The experiment is conducted with water as well as with ethanol as working
fluids. Water and ethanol are used as working fluids. The highest temperature in
the evaporator and the lowest thermal resistance is recorded at 150 W as 93.7 °C
and 0.0315 K/W. Wang et al. [10] fabricated a LHP with an evaporator of flat disc
shape. The evaporator diameter and height are 40 and 19 mm. The wick used in
this study is a stainless steel wire mesh, which contains about 630 meshes, and the
working fluid used is methanol. The maximum heat capacity reached was 240 W
and heat flux of 19.1 W/cm2 . Tests were conducted by Anand et al. [11] with four
working fluids such as acetone, ethanol, n-pentane and methanol for different heat
inputs in the miniature LHP (mLHP) containing flat evaporator. The experimental
results revealed that methanol had the highest range in heat transport and the lowest
operating temperature is recorded by n-pentane.
The significant features that distinguish compact loop heat pipe (CLHP) from
conventional LHPs are the dimensions of the connecting lines. Based on the dimen-
sions of the connecting lines [12], the LHPs are categorized as CLHP if the diameter
ranges from 3 to 5 mm. The working fluid used by Gunasegaram et al. [13] in the LHP
Feasibility of Al2 O3 /Water Nanofluid in a Compact … 469
is distilled water and silica nanoparticles. The mass concentration of silica nanopar-
ticles was ranged from 0 to 3%. Based on the results, it is observed that up to a certain
mass concentration (0.5%) the total thermal resistance is found to be decreasing in
the LHP, and beyond this mass concentration, it started increasing again. Tharayil
et al. [14] used graphene–water nanofluids with different concentration and tested
the performance of the mLHP. When compared to water, the thermal resistance at the
evaporator was found to reduce by 25%. The thermal performance of the mLHP was
observed to be improved when the nanoparticles concentration was increased. Wan
et al. [15] fabricated the mLHP and compared the performance with distilled water
and copper–water nanofluid for different mass concentrations (1.0, 1.5 and 2.0%).
The nanofluid-filled mLHP showed an enhanced performance once compared with
distilled water, and 1.5% was found to be the optimum mass concentration. Riehl
[16] studied the heat transport behaviour of a LHP. Water–nickel nanofluid was used
as the working fluid. In this study, the use of a nanofluid increased the pressure drop,
and higher operating temperatures were observed. The tendency of nanofluid is to
improve the heat transfer performance of the mLHP. Use of nanofluid in CLHP is
an encouraging research area which is quiet in its early stage. The purposes of this
experimental study are to examine and analyse the performance of CLHP and the use
of the Al2 O3 /water nanofluid. In the current study, distilled water and Al2 O3 /water
nanofluid with three different volume concentrations are used to study the heat trans-
fer performance and entropy generation in a CLHP. The effect of heat input and
volume concentration on entropy generation, thermal resistance and evaporator and
condenser heat transfer coefficients are investigated.
2 Experimentation
2.1 Preparation of Nanofluid
The nanofluid is prepared by suspending the Al2 O3 nanoparticles in water with three
volume concentrations. The Al2 O3 nanoparticles are bought from Alfa Acer with
product number 44,931 having the size of 40–50 nm and density 3950 kg/m3 . By
the two steps method, nanofluids are synthesized. The volume concentrations used
are 0.03, 0.09 and 0.12%. Surfactants are not used in the preparation of nanofluid.
Adequate quantity of nanoparticles corresponding to the desired volume fraction is
measured and dispersed straight in the distilled water and stirred. Then, the nanofluid
is sonicated in the ultrasonicator at 50–60 kHz for one hour so as to get good disper-
sion. By conducting visual observation test, the stability of the prepared nanofluids
is checked. It is found to be stable even after two weeks.
2.2 Characterization of Nanofluid
The characterization of synthesized nanoparticles was carried out by scanning elec-

tron microscope (SEM) analysis techniques and zeta potential distribution. Figure 1
shows the SEM image taken at 30,000 magnifications. SEM image gives the distri-
bution pattern and size of the Al2 O3 nanoparticles. There is crystal shape exist in this
nanoparticles. Figure 2 shows the zeta potential distribution of the Al2 O3 nanopar-
ticles. Al2 O3 nanoparticles with zeta potential values greater than +25 mV or less
than −25 mV typically have high degrees of stability.
Fig. 1 SEM image of Al2 O3

nanoparticles
Fig. 2 Zeta potential of

Al2 O3 nanoparticles
2.3 Description of the CLHP
The fabricated CLHP comprises of a square flat evaporator, condenser, smooth liq-
uid transport line, smooth vapour transport line and copper screen mesh wick. The
assembly of the evaporator section is shown in Fig. 3. The entire parts of the CLHP
are made up of copper. The size of the evaporator is 34 mm × 34 mm × 9.5 mm. The
wall thickness of the evaporator is 2 mm. The evaporator has two parts. The upper
cover is the compensation chamber (CC), and the bottom cover is the boiling cham-
ber. The boiling chamber comprises of five fin structures inside in order to enhance
the heat transfer area. The size of the fin is 25 mm × 5.5 mm × 1 mm. The capillary
wick used comprises of a copper screen mesh wick with 63% porosity. The size of
the wick is 300 mm × 25 mm × 0.25 mm. This copper screen mesh wick is wrapped
with eight turns over the fins, and ten layers of size 30 mm × 25 mm × 0.25 mm
is placed above the fins which act as a separator between CC and boiling chamber.
With the help of three projections in the CC, the copper screen mesh wick is held
tightly in place. In the CC, a partition is provided to avoid the mixing of vapour and
the liquid. Copper tubes having a 5 mm inner diameter are used as the liquid and
vapour transport lines. A water-cooled condenser having a single-pass cross-flow
forced flow with 21 °C as constant inlet temperature which is used to condense the
vapour. A cooling water flow rate as 25 LHP is used in the condenser. The heat loss
to the ambient is prevented by insulating the entire CLHP, excluding the condenser
area with thick layers of glass wool. Very low pressure of 10−4 millibar is maintained
inside the CLHP.
Fig. 3 Assembly of
evaporator
2.4 Experimental Layout
The heat transfer performance of the CLHP is analysed for the different working
fluids. The experimental test set-up is shown in Fig. 4. It consists of CLHP, heater
block, thermocouples, voltmeter and ammeter connected to a dimmerstat, computer,
data logger, cooling water pipes, chilling element, rotameter and pump. The heater
assembly comprises of a copper block and cartridge heaters. Three cartridge heaters
of each 200 W are fitted in the copper block. In order to provide the varying heat
inputs, a dimmerstat is connected to cartridge heaters, and the varying heat inputs
are measured using voltmeter and ammeter. The data logger DARWIN DAQ-100
records the temperatures for every 3 s. The cooling water supplied is maintained
at constant temperature by the chilling unit. The mass flow rate of cooling water
supplied to the condenser is measured with a rotameter and is maintained constant
by using a flow-regulating valve attached to the pump.
Seventeen thermocouples of T-Type with ±0.5 °C accuracy were positioned at
different locations in the CLHP assembly to record the wall and vapour temperatures.
The thermocouples positions are clearly mentioned in the schematic diagram of
the experimental testing set-up in Fig. 5. Instrument welder is used to fixing the
thermocouples to the outer surface of CLHP. The thermocouple joint is strengthened
with fast-curing epoxy compounds like M-seal and Araldite.
T 1 and T 2 gives the heater temperature and interface temperature between the
heater and evaporator. T 3 and T 4 measure the surface and vapour temperature at
evaporator. T 5 and T 6 measure the wall temperatures at vapour line, and T 7 gives
the vapour temperatures at the vapour line. T 8 , T 10 and T 12 provide the surface
temperature inside the condenser. T 9 and T 11 provide the vapour temperature inside
the condenser. T 13 gives the vapour temperatures at liquid line. T 14 and T 15 provide
the wall temperatures at the liquid line. T 16 and T 17 measure the inlet and outlet
temperatures of the cooling water. For every 3 s, data from the thermocouples are
recorded at the DARWIN DAQ-100 data logger. A computer connected with the
DARWIN DAQ-100 data logger records all the necessary data for analysis.
Fig. 4 A photographic view of the experimental set-up

Fig. 5 Schematic diagram of the experimental testing set-up of CLHP
2.5 Experimental Procedure
The test is performed on the CLHP with the distilled water and with Al2 O3 /water
nanofluid for various concentrations. Based on the preliminary studies, the working
fluid is charged for 60% of the total volume of the CLHP. Literature has proposed
that 30–80% as the optimum fill ratio for the CLHP to operate effectively. The tests
are conducted for the heat load range of 30–500 W. The heat loss to the surroundings
is avoided by insulating the entire CLHP assembly with the glass wool of sufficient
thickness. All the tests are conducted in vertical position, i.e., the evaporator is
positioned below the condenser. As the power input is switched initially, 30 W of the
heat load is applied to the heater through dimmerstat, and after 30 min, the steady-
state condition is reached in the CLHP. Successively, heat input of 50 W is used for a
duration of 20 min to attain the steady state. Progressively, the heat input is increased
by 50 W until the heat load becomes 500 W. For every 3 s, the data loggers record
all the temperature readings of the thermocouples and store in a computer.
2.6 Data Reduction
The thermal resistance at the evaporator (Re ) is defined as

(Te − Tv )
Re = (1)
Q
where Re is evaporator thermal resistance, T e is the surface average evaporator tem-

perature, T v is the vapour average evaporator temperature and Q is the applied heat
input.
The loop thermal resistance is defined by Eq. (2).
(Te − Tc )
Rlhp = (2)
Q
where Rlhp is loop heat pipe thermal resistance and T c denotes the average temperature
at the condenser.
The total thermal resistance is defined as
(T 2 − T a)
Rtot = (3)
Q
where Rtot is the total thermal resistance of the heat pipe, T 2 is the temperature at the
interface and T a is the ambient temperature.
The thermal resistance at the condenser (Rc ) is defined as
(Tvc − Tc )
Rc = (4)
Q
where Rc is condenser thermal resistance, T vc is the average vapour condenser

temperature and T c is the average wall condenser temperature.
The evaporator heat transfer coefficients are defined as
Q
he = (5)
Ae (Te − Tve )
where he denotes the evaporator heat transfer coefficient, Ae denotes surface area at
the evaporator and T e and T ve indicate the average wall evaporator temperature and
average vapour evaporator temperature.
The condenser heat transfer coefficients are defined as
Q
hc = (6)
Ac (Tvc − Tc )
where hc denotes the condenser heat transfer coefficient, Ac is the surface area at the
condenser and T vc and T c indicate the average vapour condenser temperature and
average wall condenser temperature.
At the condenser, heat rejection is defined using the Eq. (7).
Q c = m w Cp (T17 − T16 ) (7)

where mw is the mass flow rate of the cooling water supplied, C p is the specific heat
of the cooling water and T 16 and T 17 are the temperature of cooling water at the inlet
and outlet of the condenser.
The thermal efficiency of the CLHP is defined as
Q
η= × 100 (8)
Qc
The entropy generation due to heat transfer from the evaporator to the condenser
is given by
Q Q
Sht = − (9)
Tc Te
where Q indicates the applied heat load, T c denotes the average wall condenser
temperature and T e denotes the average wall evaporator temperature.
The entropy generated at the vapour transport line is calculated as
m × p
Svl = (10)
ρv Tv
where m denotes the mass flow rate, p indicates the pressure difference, ρ v is
average density and T v is the average vapour temperature at the vapour transport
line. The entropy generated at the liquid transport line is calculated as
m × p
Sll = (11)
ρl Tl
where m denotes the mass flow rate, p indicates the pressure difference, ρ l is
average density, T l is the average liquid temperature at the liquid transport line. The
total entropy generated in CLHP is calculated as
Stot = Sht + Svl + Sll (12)
2.7 Uncertainty Analysis
To estimate the percentage of uncertainties involved in the experiments, the flawless

uncertainty study is done on the different elements used.
The uncertainty allied with the heat input consists of current (I) and voltage (V )
uncertainties.
2 2
Q V I
= + (13)
Q V I
The uncertainty associated with the evaporator heat transfer coefficient includes
the heat flux (q) and variation of temperature between vapours and surface (T vs )
uncertainties.

h q 2 (T ) 2
= + (14)
h q TVS
The uncertainty associated with the thermal resistance is calculated as

2 2
Rt Q (T )
= + (15)
Rt Q Tec
T ec denotes the temperature difference between evaporator and condenser. The

thermal efficiency is calculated from the applied heat input, mass flow rate (m) and the
outlet and inlet cooling water temperature difference. The uncertainty is calculated
as

η Q 2 m 2 (T )W 2
= + + (16)
η Q m TW
The uncertainty associated with entropy generated in heat transfer is calculated

as
2 2 2
Sgen.ht Q T H TL
= + + (17)
Sgen.ht Q TH TL
where T H is the heat source temperature and T L is the temperature of the heat sink.
The uncertainty associated with entropy generation due to friction in the transport
line is calculated as

Sgen.fl m 2 (P) 2 ρ 2 T 2
= + + + (18)
Sgen.fl m Pec ρ T
where m denotes the mass flow rate, Pec represents the transport line pressure
difference, ρ is the density of fluid and T is the temperature for working fluid. The
uncertainties obtained are 3.9, 4.16, 4.9, 6.8, 3.7 and 3.2% for applied heat input,
heat transfer coefficient at the evaporator, thermal resistance, thermal efficiency and
entropy generation due to heat transfer and flow friction.
The functioning of the CLHP is established by the vapour and wall temperatures
shown by the thermocouples positioned at various parts of CLHP like evaporator
region, vapour transport line, condenser region and liquid transport line. Once the
heat input applied is increased gradually, the temperature at the different location
of the CLHP is also correspondingly increased. This strongly confirms the perfect
functioning of the CLHP. At the evaporator region, the temperature reports higher
than the temperature at the other parts of CLHP confirms that CLHP functioning is
perfect. It is found that once the heat load is increased, the temperature difference
(T ) of the CLHP is also increased gradually. Figure 6 shows the temperature dif-
ference (T e − T c ) of the evaporator section and the condenser section of the CLHP.
It is clearly seen from Fig. 6 that the difference in temperature (T e − T c ) increases
with the increase in heat load for all the concentration of Al2 O3 /water nanofluids and
for distilled water. The distilled water temperature difference is found to be higher
when compared with nanofluids. From the evaporator region, more amount of heat is
transferred to the condenser region when Al2 O3 nanofluids are used as working fluid.
This happens due to the high thermal conductivity of Al2 O3 nanofluid. The tempera-
ture difference obtained for 0.12 vol% Al2 O3 nanofluids is reduced by 13.94% while
compared with distilled water, and the same is reduced by 11.12% for 0.09 vol%
Al2 O3 nanofluids while compared with distilled water.
Figure 7 illustrates the influence of thermal resistance of evaporator with the heat
input range from 30 to 500 W. The thermal resistance value of evaporator for all cases
declines as heat load increases. The thermal resistance value of evaporator declines
with the increases in volume concentration of Al2 O3 nanofluids. In general, large size
Fig. 6 Temperature
difference with heat load
Fig. 7 Thermal resistance

with the heat load
bubbles are formed in the evaporator, and due to the presence of nanoparticles, these
larger bubbles are split down into smaller vapour bubbles which offer less resistance
to heat transfer. The evaporator thermal resistance is found to be reduced by 21.29%
for 0.12 vol% of Al2 O3 nanofluids and 17.71% for 0.09 vol% of Al2 O3 nanofluids
when compared with distilled water.
The loop thermal resistance variation with heat load for four CLHPs is shown
in Fig. 8. It is observed from Fig. 8 that the thermal resistance decreases with the
increase in heat load as well as the volume concentration of Al2 O3 nanofluids.
When the heat load is increased, more vapour are generated, and a large amount
of vapour is passed to the condenser, which yields in low thermal resistance. The
nanoparticles stick to the copper screen mesh wick and form a thin layer with a
large number of small nucleation site, which increases the capillary force and the
solid–liquid wettability [17]. The loop thermal resistance is reduced by 14.3% for
0.12 vol% of Al2 O3 nanofluids when compared with distilled water, and the same is
reduced by 13.5% for 0.09 vol% Al2 O3 of nanofluids. The evaporator heat transfer
coefficient variation with the heat load is shown in Fig. 9 for distilled water as
well as for three volume concentrations of Al2 O3 nanofluids. In the evaporator, the
convective heat transfer coefficient is a picture of convective boiling. As shown in
Fig. 9, the heat transfer coefficient at the evaporator increases with the increase in
heat load and volume concentration. At low heat loads, the boiling rate is low in the
evaporator region and is high for higher heat loads. As the heat input increases, more
amount of nucleation sites may be activated that leads to a rise in trend in the heat
transfer coefficient. For 0.12 vol% of Al2 O3 nanofluids, the convective heat transfer
coefficient is enhanced by 24.42% at evaporator when compared with distilled water.
The heat transfer coefficient is increased with the increase of Al2 O3 nanoparticles
Fig. 8 Loop thermal

resistance with heat load
Fig. 9 Evaporator HTC with

the heat load
concentration in the condenser region. In the condenser at high heat loads, vapour
requires a larger area to condense. Inadequate surface area and constant cooling water
flow rate resist the heat transfer coefficient in the condenser region when comparing
with the evaporator. The condenser heat transfer coefficient variation related with
heat load is shown in Fig. 10.
Fig. 10 Condenser HTC

with the heat load
From Fig. 10, it is observed that the heat transfer coefficient at condenser is
also increased with the increase in heat load. The heat transfer coefficient at the
condenser is enhanced by 12.9% for 0.12 vol% Al2 O3 nanofluids when compared
with that of distilled water. Figure 11 shows the variation of thermal resistance of
condenser with the heat load. As the heat load increases, the thermal resistance of
Fig. 11 Thermal resistance

with the heat load
condenser decreases. When compared with distilled water, the condenser thermal
resistance with 0.12 vol% Al2 O3 nanoparticles is observed to be reduced by 26.75%
and for 0.09 vol%, and the same is reduced by 20.12%. The generation of entropy
is evaluated as the outcome of irreversibilities occurred. Entropy generation occurs
in the CLHP that is due to heat transfer with evaporator and condenser, and due to
the frictional effect in the transport line, the entropy generated in the copper screen
mesh wick and at the condenser is not accounted because of the negligible pressure
drops in them. The entropy generation because of pressure drops is extremely low
once evaluated with entropy generation due to heat transfer. This is due to the heat
transfer irreversibility caused by the latent heat of vapourization that dominates over
the frictional irreversibility caused by viscosity.
It was observed that with the rise in heat load, the entropy generation is increased
alongside, and with the increase of volume concentration of Al2 O3 nanoparticles,
the entropy generation decreases. The average fall in entropy generation is 6.47%
for 0.09 vol% and 11.11% for 0.12 vol% of Al2 O3 /water nanoparticles. The entropy
generation is negligible for the liquid flow in comparison with the vapour flow in the
CLHP. In the liquid transport line, the outcome of heat transfer is very low because of
the tremendous low-temperature drop of liquid in the liquid transport line. The total
entropy generation variation of the CLHP is shown in Fig. 12. This happens due to
heat transfer and pressure drop caused by friction. With an increase in the heat load,
the total entropy generation also increases, but with increase in nanofluid concentra-
tion, the total entropy generation decreases. The entropy generation is reduced in the
CLHP due to the Al2 O3 nanofluids. The presence of nanoparticles with high thermal
conductivity in the base fluid enhances the heat transfer. The entropy generation due
to the heat transfer was dominated in comparison with the entropy generation due to
Fig. 12 Total entropy

generation in CLHP
Fig. 13 Heat input versus heat output
friction. In the CLHP, energy balance is done for the heat applied at the evaporator
and for the heat removed at the condenser.
The energy balance of the CLHP for distilled water and Al2 O3 /water nanofluid at
different heat loads is shown in Fig. 13. It is observed that the average output gained is
82% for distilled water, and when matched with distilled water, enhancement of 11%
additional is achieved for Al2 O3 /water nanofluid. Hence, it is proved that the CLHP
thermal performance is improved with the use of nanoparticles with small volume
concentrations. The increase in the performance of the CLHP with nanofluids makes
it significant in cooling modern electronic devices.
4 Conclusion
CLHP with flat square evaporator is fabricated, and the thermal performance is
experimentally tested for distilled water and Al2 O3 /water nanofluids with 0.03, 0.09
and 0.12% volume concentration. The CLHP was capable of operating up to 500 W
without the occurrence of dry out condition. When compared with distilled water,
the use of Al2 O3 /water nanofluids in CLHP as working fluid showed a better thermal
performance. It is revealed that the thermal resistance decreases as the heat load
and concentration of the Al2 O3 nanofluid increases. The heat transfer coefficients
at the evaporator and condenser and thermal efficiency of CLHP increased with the
increase in Al2 O3 nanofluid concentration. In the CLHP, the total entropy generated
is increased with the increase in the heat load and decreased with the increase in
the concentration of Al2 O3 nanofluid. It is concluded that the use of small quantity
Al2 O3 nanoparticles in CLHP enhances the heat transfer performance significantly.
Retrofitting of conventional heat sinks with nanofluid filled CLHP could be one of
the possible solutions for the cooling of modern electronic devices.
References
1. Maydanik YF (2005) Loop heat pipes. Appl Therm Eng 25:635–657

2. Shah B (2016) Recent trends in heat pipe applications: a review. Int J Sci Eng Technol Res
(IJSETR) 5(7)
3. Vasiliev L, Lossouarn D, Romestant C, Alexandre A, Bertin Y, Piatsiushyk Y, Romanenkov
V (2009) Loop heat pipe for cooling of high-power electronic components. Int J Heat Mass
Transf 52:301–308
4. Maydanik Y, Pastukhov V, Chernysheva M (2018) Development and investigation of a loop
heat pipe with a high heat-transfer capacity. Appl Therm Eng 130:1052–1061
5. Nakamura K, Odagri K, Nagano H (2016) Operating characteristics of 10 m-class long-distance
loop heat pipe under anti-gravity condition, In: Proceedings of the joint 18th international heat
pipe conference and 12th international heat pipe symposium, pp 384–39
6. Wang D, Liu Z, Jiang C, Chen B, Yang J, Tu Z, Liu W (2014) Experimental study of the loop
heat pipe with a flat disk-shaped evaporator. Exp Thermal Fluid Sci 57:157–164
7. Yang Y, Zhu K, Wang Y (2016) Experimental investigation and visual observation of a vapor–
liquid separated flat loop heat pipe evaporator. Appl Therm Eng 101:71–78
8. He S, Zhao J, Liu Z, Tian W, Yang J, Liu W (2018) Experimental investigation of loop heat
pipe with a large squared evaporator for cooling electronics. Appl Thermal Eng 144:383–391
9. Tang Y, Xiang J, Wan Z, Zhou W, Wu L (2010) A novel miniaturized loop heat pipe. Appl
Therm Eng 30:1152–1158
10. Wang Dongdong, Liu Zhichun, Shen Jun, Jiang Chi, Chen Binbin, Yang Jinguo, Tu Z, Liu
W (2014) Experimental study of the loop heat pipe with a flat disk-Shaped evaporator. Exp
Thermal Fluid Sci 57:157–164
11. Anand AR, Akhil J, Amrit A, Pradip D (2018) Experimental studies on a miniature loop heat
pipe with flat evaporator with various working fluids. Appl Therm Eng 144:495–503
12. Ji L, Daming W, Peterson GP (2010) Experimental studies on a high performance compact
loop heat pipe with a square flat evaporator. Appl Therm Eng, 30:741–752
13. Prem G, Abdullah MZ, Yusoff MZ, Abdullah SF (2015) Optimization of silica nanoparticle
mass concentration and heat input on a loop heat pipe. Case Stud Therm Eng 6:238–250
14. Tharayil T, Asirvatham LG, Ravindran V, Wongwises S (2016) Thermal performance of
miniature loop heat pipe with graphene-water nanofluid. Int J Heat Mass Transf 93:957–968
15. Wan Z, Deng J, Li B, Xu Y, Wang X, Tang Y (2015) Thermal performance of a miniature loop
heat pipe using water-copper nanofluid. Appl Therm Eng 78:712–719
16. Riehl RR (2012) Heat transport behavior of a miniature loop heat pipe using water-nickel
nanofluid. Heat Pipe Sci Technol Int J 3:83–96
17. Kim HD, Kim J, Kim MH (2007) Experimental studies on CHF characteristics of nanofluids
at pool boiling. Int J Multiphase Flow 33:691–706
Analysis of Externally Pressurized
Thrust Bearing with Inclination Angle
Using Yield Stress Fluids
G. Alexander Raymand and I. Jayakaran Amalraj
Abstract In this investigation, the effects of the non-Newtonian characteristics and

angle of inclined plane (F) of the externally pressurized bearing lubricated with yield
stress fluid is studied. We have obtained mathematical formulation for varying film
thickness. Applying appropriate boundary conditions and numerical procedure, the
governing equations are solved iteratively. There is a formation of an unyielding
core in the flow region for yield stress fluids, and the shape and size of this core
have been calculated numerically. Further, the effects of the angle of inclined plane
(F) and non-Newtonian characteristic on pressure distribution and load capacity are
discussed.
Keywords Externally pressurized bearing · Yield stress fluid · Inclination angle ·

Load capacity
1 Introduction
In the modern world, machinery industries focus on the bearing model and the innate
nature of the lubricant as it increases the bearing’s performance. Thrust bearings are
innately developed to withstand heavy axial load. Friction is developed between the
plates, while the bearing operates as it is subjected to high-speed operations, heavy
loads, high stiffness, etc. To reduce this friction, the bearing is lubricated. Generally,
lubricants are classified into two types, i.e. Newtonian and non-Newtonian. Any fluid
that does not follow Newton’s law is termed as non-Newtonian fluids, and they are
characterized with high viscosity. Time-independent non-Newtonian fluids are the
fluids in which the rate of shear at any point is a function of shearing stress only. Eg:
Grease, mineral oil, etc.
G. Alexander Raymand (B) · I. Jayakaran Amalraj

Department of Mathematics, Sri Sivasubramaniya Nadar (SSN) College of Engineering,
Kalavakkam, Chennai 603110, India

https://doi.org/10.1007/978-981-15-4488-0_41
486 G. Alexander Raymand and I. JayakaranAmalraj
Non-Newtonian fluids have varied properties compared to Newtonian fluids as

they resist flow and thereby significantly increase load-carrying capacity in bear-
ings. The shear-thinning and shear-thickening characteristics are the key factors that
influence them. However, non-Newtonian fluids characterized by a yield value are
keenly studied by Tribologists, and they include fluids such as Bingham, Casson
and Herschel Bulkley. Therefore, non-Newtonian fluids in the externally pressurized
converging circular thrust bearings have been a necessity and developing interest.
By considering the above principles of the fluid with yield value, the bearing
surfaces in this study are separated by a fluid film that is formed and sustained by
external means. The pioneering researchers in this field have laid the foundation of
bearings lubricated with Newtonian fluids. Nowadays, Tribologists concentrate on
non-Newtonian fluids characterized by a yield value due to its varying viscosities.
Tribologists have most often researched the rheodynamic lubrication in an exter-
nally pressurized thrust bearing. Kalil et al. [1] investigated numerically the effects
of convective and centrifugal inertial forces, on the performance of externally pres-
surized conical thrust bearings under turbulent flow and observed that the circular
bearing had a slightly higher dimensionless pressure, load and torque than the conical
bearing. It was observed that the turbulent flow solution gave higher dimensionless
pressure, load and torque compared to the laminar flow solution. Elsharkawy and
Kassab [2] have analysed the effect of a porous layer on the hydrodynamic lubrica-
tion performance of an externally pressurized circular recessed bearing and observed
that dimensionless viscosity parameter has significant effects on the pressure profiles
and the load-carrying capacity of the externally pressurized porous bearings. Lin [3]
theoretically studied the combined effects of couple stress, fluid inertia and recess
volume fluid compressibility on the steady-state and the dynamic characteristics of
hydrostatic circular step thrust bearings and concluded that steady-state load-carrying
capacity decreases with fluid inertia but remains the same even for higher values of
the couple stress parameter. The problem of an externally pressurized thrust bearing
lubricated with Herschel–Bulkley fluids has been considered by Kandasamy et al.
[4], and numerical solutions have been obtained for the pressure distribution and the
load-carrying capacity.
Using yield stress fluids as lubricants, the problem of different types of bearings
has been investigated. H-B fluids are a class of non-Newtonian fluids that require finite
stress known as yield stress, to deform. Therefore, when the local shear is below the
yield stress, these materials behave as a rigid body. Once the yield stress exceeds, the
material flows with a nonlinear stress–strain relationship which exhibits itself either
as a shear-thinning or shear-thickening fluid. Alexandrou et al. [5] studied a steady
flow of Herschel–Bulkley fluids in a canonical three-dimensional expansion and
examined the topology of the yielded and unyielded surfaces. The results revealed a
strong interplay between the Bingham and Reynolds numbers. Using the augmented
Lagrangian method, Huilgol et al. [6] investigated the problems with the steady flow
of visco-plastic fluids in pipes of circular and square cross-sections. The solutions
for plug flow velocity, the flow rate and its pattern, the velocity profile, the region
of yielded/unyielded surfaces, the stopping criteria and the friction factor have been
obtained. Vishwanath and Kandasamy [7] analysed the effects of fluid inertia forces in
Analysis of Externally Pressurized Thrust Bearing … 487
circular squeeze film bearing lubricated with Herschel–Bulkley fluids with constant
squeeze motion. They observed that the pressure and the load capacity of the bearing
proportionately increase with the increase of Herschel–Bulkley number and power
law index of the fluids. Amalraj et al. [8] analysed the problem of an externally
pressurized thrust bearing lubricated with yield stress fluid under the sinusoidal
flow rate concluding that the effects of non-Newtonian characteristics on the load
capacity are more significant when the amplitude of the sinusoidal motion is very
high. Walicki and Jurczak [9] investigated the influence of a wall porosity on the
pressure distribution in a curvilinear squeeze film bearing lubricated with a yield
stress type lubricant and obtained a formula expressing the pressure distribution.
The geometric shape of the bearing plays a vital role as it wholly increases the
performance of the bearing. This factor drives the following studies to research on the
converging film thickness using non-Newtonian fluids with its numerous potential
advantages. Most of the studies focus on externally pressurized thrust bearings with
uniform film thickness and not with variable film thickness. A few researchers studied
the performance of converging and diverging bearings. Roy et al. [10] discussed the
effect of inertia forces in an externally pressurized bearing with both converging
and diverging film thickness using a visco-elastic lubricant. Vishwanath et al. [11]
analysed the problem of a squeeze film bearing with the converging and diverging
film thickness. Amalraj and Raymand [12] studied the effects of the varying film
in rheodynamic lubrication using externally pressurized bearing and have shown
theoretically that as the inclination angle increases, the load-carrying capacity also
increases.
The purpose of this research is to investigate the impact of the inclination angle and
non-Newtonian characteristics on the performance of the circular thrust bearing using
yield stress fluids. Numerical solutions have been obtained for bearing performances
such as the core thickness, velocity, pressure distribution and load-carrying capacity
for various values of yield stress fluid number (N), power law index (n) and inclination
angle (F).
2 Formulation of the Problem
The region between the circular plates of the bearing is symmetric, so the upper
portion of the bearing is considered for analysis. The geometry of the problem is
shown schematically in Fig. 1.
The 3-D constitutive equation of a yield stress fluid is given by (Alexandrou et al.)
(n−1) √
DII 2 η2 1 − exp −m DII /2
τ = η1 + √ ∗D (1)
2 DII /2
Whorlow gave 1 − D form of Eq. (1) in 1980.

Fig. 1 Schematic diagram

of externally pressurized
inclined circular thrust
bearing
τ = η2 + η1 γ̇ n , where γ̇ represents shear rate. (2)
There will be a core formation when the shear stress is less than the yield stress
of the material. The core will move with the constant velocity, vc . Let the boundaries
of the core be z = − δ(r2)h and z = δ(r2)h as shown in Fig. 2.
Using the assumption of lubrication theory, the governing equations are reduced
and expressed as follows
Continuity Equation:
1 ∂ ∂vz
(r vr ) + =0 (3)
r ∂r ∂z
Momentum Equation:
∂τr z ∂p
=− (4)
∂z ∂r
∂p
=0 (5)
∂z
Fig. 2 Formation of core in

externally pressurized thrust
bearing
n
∂vr
τr z = η2 + η1 (6)
∂z
The considered governing equations are to be solved by following boundary

conditions
h
vr = 0 at z = ± (7)
2
δh
vr = vc at z = ± (8)
2
∂vr
is continuous, at τ = η2 (9)
∂z
p = pa at r = R2 (10)
3 The Solution to the Problem
Solving Eq. (4) by integrating and using Eq. (6) and applying boundary condition
(7) and (8), we obtain velocity distributions in the flow region as
1 1
n 1 n −d p n
vr =
n + 1 η1 dr
n1 +1 1
δh h δh n +1 δh h
∗ z− − − , where ≤z≤ (11)
2 2 2 2 2
velocity in the core region as
n +1
1 1 1
n 1 n −d p n h δh δh
vc = − − , where 0 ≤ z ≤ (12)
n+1 η1 dr 2 2 2
The equation of conservation of mass which depends on the bearing shape in this
case is
h
2
Q = 4πr ∫ vr dz (13)
0
where Q is flow rate per unit width

Using velocity distributions vr and vc from Eqs. (11) and (12) in Eq. (13) and
integrating, we obtain
⎧ 1 ⎫
1 ⎪
⎨ πr h n +2 −d
1 p n ⎪
⎬
n 1 n dr
Q=−
(n + 1)(2n + 1) η1 ⎪
⎩
1
2n ⎪
⎭
∗ (1 − δ) n +1 (nδ + n + 1)
1
(14)
Now, consider the equilibrium of an element in the yield surface − δh2 ≤ z ≤ δh

2
,
we get
dp 2η2
= (15)
dr δ(r )h
By eliminating the pressure gradient from Eqs. (14) and (15), we get an algebraic
equation to find the thickness of the yield surface δ(r)
1
(n+1)(2n+1)
n
Q η1n (1 − δ) n +1 (nδ + n + 1)
1
1 = 1 (16)
π r h 2 η2n δn
From Eq. (14), we get,

n
dp 2 (n+1)(2n+1)
n
Q n η1
= n n 2n+1 (17)
dr π r h (1 − δ)n+1 (nδ + n + 1)n
From Fig. 1, the variation of the film thickness of the lubricant due to the angle
of inclination of the bearing can be defined as

r
h(r ) = h 0 − h 0 tan φ (18)
R2
Non-dimensional parameters are introduced as follows.

r
r∗ = ; δ∗ = δ r ∗ ;
R2
∗ p h z
p = ; h∗ = ; z∗ = ;
Q n η1 h0 h
π n h 02n+1 R2n−1
1
π R2 h 20 η2 n
N= (19)
Q η1
The non-dimensional form of the velocity profile, core thickness as

1
(2n + 1)(2) n δ∗ h ∗ h∗
vr∗ = 1 , where − ≤ z∗ ≤ (20)
r ∗ (1−r ∗ tan φ)(nδ ∗ +n+1)(1−δ ∗ ) n +1
2 2
1
1 +1 ∗ n +1
( 21 ) n (1−δ∗ ) n +1 − z ∗ − δ2
1
(2n + 1) δ∗ h ∗
vc∗ = , where 0 ≤ z ∗
≤ (21)
2r ∗ (1 − r ∗ tan φ)(nδ ∗ + n + 1) 2
n
(n+1)(2n+1)
(1 − δ ∗ )1+n (1 + n + nδ ∗ )n n
− =0 (22)
δ∗ (N ) (r ) (1 − r ∗ tan φ)2n
n ∗ n

d p∗ 2
=N n
(23)
dr ∗ δ ∗ (1 − r ∗ tan φ)
From the nonlinear algebraic Eq. (22), the core thickness can be calculated.
The roots δ(r ∗, N , n, ϕ) of this equation define the shape of the plug core region
is which is positive and smaller than unity. The values of δ have been determined for
different values of angle of inclination, yield stress fluid and power law index (φ, N
& n) by iterative techniques.
Knowing (δr*) and integrating Eq. (23) numerically and using boundary condition
p = pa , the pressure distribution can be found and is given by,
1 d p∗
P ∗ − Pa∗ = ∫ dr ∗ (24)
r∗ dr ∗
Similarly, the load-carrying capacity of the externally pressurized bearing can be

obtained by integrating the pressure along the thrust bearing and is given by
1
W = (P ∗ − Pa∗ )r ∗ dr ∗ (25)
R∗
where R ∗ = RR21 is the ratio of inside to outside radius of the bearing. This integration
is performed numerically for various values of inclination angle, yield stress fluid
number and power law index (φ, N &n).
The shape and thickness of the core for various inclination angle, yield stress fluid
number and the power law index (F, N & n) at every point in the radial direction
are computed, and the results are given in Figs. 3, 4 and 5. In the case of varying
film thickness, for a given yield stress fluid and the power law index (N & n) at
each point in the radial direction, its thickness decreases as inclination angle (F)
Fig. 3 Core thickness for different inclination angle when n = 0.7
Fig. 4 Core thickness for different inclination angle when n = 1
increases. Further, the thickness of the core increases with an increase in a yield
stress fluid (N) for a constant power law index (n). Similarly, the thickness of the
core marginally decreases with an increase in power law index (n), for a constant
yield stress fluid number (N). The velocity profile for various inclination angle, yield
stress fluid number and power law index (F, N & n) along the axial direction (z*)
Fig. 5 Core thickness for different inclination angle when n = 1.3
for various values of the radius(r*) is depicted in Fig. 6. The thickness of the core
as observed earlier is reflected in the velocity profile. The pressure distribution for
various values of inclination angle, yield stress fluid number and power law index
(F, N & n) in the radial direction is as shown in Figs. 7, 8, 9 and 10. The pressure
profile is found to be maximum at the centre and gradually decreases as it moves
towards the periphery in the radial direction. The load-carrying capacity for various
Fig. 6 Velocity profile

Fig. 7 Pressure distribution for zero inclination angle
Fig. 8 Pressure distribution for different inclination angle (F) and yield stress number (N) when
n = 0.7
values of inclination angle, yield stress fluid number and power law index (φ, N &
n) in the radial direction is tabulated in Tables 1, 2 and 3.
Case 1: Increasing the inclination angle (F), the pressure is maximum at the centre
and decreases as it moves towards its periphery, and increment in load-carrying
capacity is gradual and marginal for a constant yield stress fluid number and power
law index (N & n). Case 2: For dilatant fluids (n > 1), the pressure distribution and
load-carrying capacity are high, whereas for pseudoplastic fluids (n < 1), pressure
distribution and load-carrying capacity are low for constant inclination angle and
Fig. 9 Pressure distribution for different inclination angle and yield stress number (N) when n = 1
Fig. 10 Pressure distribution for different inclination angle and yield stress number (N) when n =
1.3
Table 1 Load capacity for different value of N and φ for n = 0.7

n = 0.7 N =5 N = 10 N = 15 N = 20 N = 25
φ =0 2.48 3.06 3.58 4.06 4.51
φ = 10 3.02 3.66 4.25 4.80 5.31
φ = 20 4.02 4.72 5.40 6.04 6.65
φ = 30 6.69 7.35 8.13 8.91 9.67
Table 2 Load capacity for different value of N and φ for n = 1

n =1 N =5 N = 10 N = 15 N = 20 N = 25
φ =0 4.25 6.32 8.30 10.24 12.13
φ = 10 5.39 7.75 10.08 12.35 14.57
φ = 20 7.30 10.30 13.16 15.95 18.68
φ = 30 12.63 16.68 20.59 24.38 28.10
Table 3 Load capacity for different value of N and φ for n = 1.3

n = 1.3 N =5 N = 10 N = 15 N = 20 N = 25
φ =0 7.50 13.61 20.13 26.99 34.15
φ = 10 9.53 16.94 24.75 32.94 41.45
φ = 20 13.35 22.90 32.85 43.19 53.88
φ = 30 23.63 37.98 52.51 67.40 82.67
yield stress fluid number (F & N). Case 3: Increasing yield stress fluid number (N),
the pressure is maximum at the centre and decreases as it moves towards its periphery,
and the increase in load-carrying capacity is gradual and marginal for a constant (F
& n). However, pressure distribution and load-carrying capacity are more significant
for Case 2 when compared with Case 1 and Case 3.
We compute the relative percentage of increase for pressure distribution and load-
carrying capacity for inclination angle φ > 0 by keeping φ = 0 as a reference angle and
depicted in Figs. 11 and 12. It is evident that the increase in the relative percentage
of pressure and load-carrying capacity is quite significant with respect to power law
Fig. 11 Relative percentage of pressure difference for N = 5

Fig. 12 Relative percentage of load-carrying capacity for various φ
index (n). This will help in designing the bearing to have the optimal load-carrying
capacity in the case of an externally pressurized thrust bearing.
4.1 Validation
The shape and extent of the core along the radial direction, which we determined,
are found to be in good agreement with the results published earlier by Jayakaran
et al. [8] for uniform film bearing (F = 0).
5 Conclusion
The above investigation reveals that the Herschel−Bulkley number (N), power law
index (n) and the inclination angle (F) have significant effects in the bearing per-
formance such as pressure distribution and load-carrying capacity of the bearing.
Further, it has been observed that the shear-thickening fluids (n > 1) improve the
bearing performance, whereas the shear-thinning fluids (n < 1) decrease for any
given N and F.
Acknowledgements The authors thank the Principal and management of SSN institutions for their
support in carrying out this work.
References
1. Khalil MF, Kassab SZ, Ismail AS (1993) Performance of externally pressurized conical thrust
bearing under laminar and turbulent flow conditions. Wear 166:147–154
2. Elsharkawy AA, Kassab SZ (1996) Lubrication analysis of externally pressurized porous
bearings. App Math Modell 20:870–876
3. Lin J-R (1999) Static and dynamic characteristics of externally pressurized circular step thrust
bearings lubricated with couple stress fluids. Tribol Int 32:207–216
4. Kandasamy A, Narasimman S, Jayakaran Amalraj I (2006) Rheodynamic lubrication of an
externally pressurized thrust bearing with Herschel-Bulkley fluids. In: Proceedings of the
international conference on Industrial Tribology, IISc, Bangalore, India
5. Alexandrou AN, McGilvreay TM, Burgos G (2001) Steady Herschel- Bulkley fluid flow in
three-dimensional expansions. J Non-Newton Fluid Mech 100:77–96
6. Huilgol RR, You Z (2005) Application of the augmented lagrangian method to steady pipe flows
of Bingham, Casson, and Herschel- Bulkley fluids. J Non-Newton Fluid Mech 128:126–143
7. Vishwanath KP, Kandasamy A (2010) Inertia effects in circular squeeze film bearing using
Herschel-Bulkley lubricants. Appl Math Model 34:219–227
8. Amalraj IJ, Narasimman S, Kandasamy A (2012) Rheodynamic lubrication of an externally
pressurized thrust bearing using Herschel-Bulkley fluid with sinusoidal injection. J Appl Fluid
Mech 5(4):71–79
9. Walicka A, Jurczak P (2016) Pressure distribution in a porous squeeze film bearing lubricated
with a Herschel-Bulkley fluid. Int J Appl Mech Eng 24(4):951–965
10. Sinha Roy J, Padhy S, Bhopa LK (1993) Inertia effect in the externally pressurized thrust bearing
with converging and diverging film using visco-elastic lubricant. Actamechanica 96:1–12
11. Vishwanath KP, Kandasamy A (2008) Rheodynamic lubrication of a rectangular squeeze film
bearing with an exponential curvature variation using Bingham lubricants. Int Sci Index Phys
Math Sci 02(07):380–384
12. Jayakaran Amalraj I, Alexander Raymand G (2016) Core variation in an externally pressurized
converging thrust bearing with Bingham lubricant. Appl Mech Mater 852:428–434
Numerical Investigation of Bingham
Fluid Flow in the Entrance Region
of Rotating Annuli
S. Mullai Venthan and I. Jayakaran Amalraj
Abstract The study of the entrance region flow is, nowaday, assuming consider-
able technical importance due to its immediate applications in various designs of
chemical and biomedical equipments in which the flows of non-Newtonian fluids
are encountered. The flow characteristic of non-Newtonian fluid that is independent
of time, and which obeys Bingham’s stress–strain relations, has been investigated
at the entrance region of the annular space between two rotating coaxial cylinders.
This analysis has been carried out for both the cases when each rotates with different
speed in the same direction and as well as in the opposite direction with constant
angular velocity. Discussions have been presented for a steady, laminar, isothermal
flow condition of the Bingham fluid. The continuity and momentum equations are
solved iteratively with finite difference method by using the Prandtl’s boundary layer
assumptions. The velocity components and pressure distributions have been obtained
numerically for various values of Bingham number and aspect ratio.
Keywords Entrance region · Rotating cylinders · Bingham fluid · Finite difference

technique · Speed variation
Nomenclature
η, ζ Mesh sizes in the radial and axial directions respectively

ρ Density of the fluid
μ Apparent viscosity of the model
μr Reference viscosity
ω Angular velocity
η, ζ Dimensionless coordinates in the radial and axial directions respec-
tively
S. Mullai Venthan (B) · I. Jayakaran Amalraj

Department of Mathematics, Sri Sivasubramania Nadar (SSN) College of Engineering, Chennai,

https://doi.org/10.1007/978-981-15-4488-0_42
500 S. Mullai Venthan and I. Jayakaran Amalraj
B Bingham number
r 1 /r 2 (Flow region of an annular space)
r1, r2 Radius of the inner and outer cylinders respectively
k Coefficient of fluidity
m Number of radial increments in the numerical mesh network
P Dimensionless pressure
p, p0 Pressure and initial pressure respectively
R, θ & z Cylindrical coordinates
Re, Ta Modified Reynolds number and Taylor number respectively
Rt = Re2 /Ta Ratio of Reynolds number with Taylor number
R Aspect ratio of the annulus
s1 , s2 Speed coordinates for inner and outer cylinders respectively
v0 Uniform inlet velocity
vz , vr , vθ Velocity components in z-, r-θ -directions respectively
U, V, W Dimensionless velocity components
1 Introduction
The study of entrance region flow in rotating cylinders has been widely employed
in various industries, ranging from the pharmaceutical, food and agricultural indus-
tries to the chemical industries, wherein Newtonian and non-Newtonian fluids are
frequently being encountered. Lately, there has been an increasing concentration of
problems involving fluids with variable viscosity such as Bingham plastic, Casson
and Herschel–Bulkley fluids which are characterized by a yield value.
The development of fluid flow for non-Newtonian and power-law fluids in the
entrance region of concentric annular cylinders has been studied by various authors,
using various approximation methods, and the corresponding literature reviews have
been mentioned below.
Schlichting and Gersten studied the flow of typical materials, particularly non-
Newtonian fluids and the flow of any material under the influence of an applied force
or stress in the preview of the boundary layer theory [1]. Escudier et al. numerically
analyzed the fully developed laminar flow of an inelastic shear thinning power-law
fluid through an eccentric annulus with inner cylinder rotation. Further, they also had
analyzed more complex rheological models [2]. Ramesh Gupta has obtained velocity
distribution by hydrodynamic study, the development of laminar power-law fluid flow
in pipes and straight channels for equivalent Newtonian models [3]. Radka and Karel
investigated the steady and unsteady incompressible flows for Newtonian and non-
Newtonian shear thickening fluid flow through a branching channel by numerical
modeling [4]. Poole and Chhabra numerically analyzed the development of laminar
pipe flow of yield stress fluids which obey the Bingham-type models. And they have
shown the development lengths for various values of Reynolds number [5].
Numerical Investigation of Bingham Fluid Flow … 501
Huilgol studied the minimum pressure drop required to initiate the flow of vis-
coplastic fluids with a constant yield stress, such as Casson and Herschel–Bulkley
fluids in the pipes of the following symmetric cross sections: circular, annular, square,
rectangle, equilateral triangle, elliptical and L-shape [6]. Khali et al. numerically
investigated the effect of a porous layer on non-Newtonian fluids between rotat-
ing concentric annular ducts using the Lattice Boltzmann method [7]. In concentric
annular cylinders with known diameter ratio and inner cylinder rotational speed, the
mean velocity and the Reynolds shear stress of both Newtonian and non-Newtonian
fluids have been calculated by Nouri and Whitelaw [8].
Rachid Chebbi analyzed flow in both the inlet and filled regions and obtained
the lengths of the entrance, inlet and filled regions as a function of the general-
ized Reynolds number and the power-law fluid index by using an integral boundary
layer method [9]. Round and Yu analyzed Bingham fluid in a general perspective in
concentric annular cylinders for the development of flows using up-winding finite
difference technique in order to solve the equation of motion. They obtained the
results of velocity and pressure drop for different radius ratios, power-law indices
and Bingham numbers [10]. Schulz and Pfister experimented on concentric annular
cylinders using Taylor–Couette flow between rotating and counter-rotating cylin-
ders. The numerical results, derived using finite element method, have an excellent
quantitative agreement with experimental measurements for a wide range of aspect
ratios and rotation rates [11].
Nallapu Santhosh et al. studied Herschel–Bulkley fluid obeying a two-fluid model
flow through tubes of small diameters and a slip on the wall [12]. Guang Lu et al.
investigated the cross-sectional flow of non-spherical particles in horizontal rotating
cylinders with and without wall roughness. Using a similar method, gravitational
acceleration and particle size were altered to investigate the effect of wall rougheners
under a range of operating conditions by using discrete element method (DEM) [13].
Rekha and Kandasamy numerically investigated the results for entrance region flow
of a Herschel–Bulkley fluid in an annular cylinder in the form of velocity profile
within the boundary layer region. Entrance flow region of an annulus and pressure
distribution has been measured for a particular Herschel–Bulkley number and various
aspect ratios [14]. Srinivasa Rao and Kandasamy numerically studied the effects of
parameters, aspect ratio and yield value on the pressure drop and velocity profiles for
the entrance region flow heat transfer in concentric annuli with rotating inner wall
for Bingham fluid [15].
The purpose of the present work is to reduce the friction, flow development at
the entrance region and to increase the performance for rotating concentric annular
cylinders by using time-independent non-Newtonian fluid obeying Bingham fluid
with excessive work. The non-Newtonian fluid has been used to quantify the possi-
ble benefits of drag-reducing fluids. The shear rate of a non-Newtonian fluid is of
importance in fixing the rheological or viscometric behavior of such materials.
In the study, the authors have analyzed numerically the velocities and pressure
distributions for the flow of the Bingham fluid at the entrance region of an annular
space between two rotating coaxial cylinders. The analysis has been carried out with
the assumption that both cylinders are rotating in the same as well as in the opposite
direction. Under Prandtl’s boundary layer assumptions, the momentum equations

and the equation of conservation of mass have been discretized and solved using
linearized implicit finite difference scheme. The system of linear equations thus
obtained has been solved by the Gauss Jordan method. The development of different
velocity profiles and pressure distributions at the entrance region of the annular space
have been computed numerically and presented for various values of Bingham num-
ber and aspect ratio. The effects of non-Newtonian characteristics and geometrical
parameters on the flow behavior at the entrance region of annular space between two
rotating coaxial cylinders have been discussed.
2 Formulation of the Problem
We have considered the Bingham fluid entering into the horizontal rotating concentric
circular cylinders, with inner and outer radius r 1 and r 2 , respectively, from a large
chamber with a uniform velocity v0 and pressure p0 initially. By fixing the origin
at the entry region on the central axis of an annulus and considering cylindrical
polar coordinates system (r, θ , z) with z-axis measured along the axial direction
perpendicular to radial direction r, with the inner and outer rotating cylinders with
angular velocity ω, under the assumptions of steady incompressible axisymmetric
flow with constant properties. The geometry of the problem is shown in Fig. 1.
The constitutive equation for Bingham fluid is given as

τ0 1 1
τi j = (μ + )εi j τ ≥ τ0 where τ = τi j τi j and ε = εi j εi j (1)
ε 2 2
where τ 0 is the yield stress, τ ij and εij are the stress tensor and the rate-of-strain
tensor, respectively, and μ is the viscosity of the fluid. The flow is governed by the
equations
∂ ∂
(r vr ) + (r vz ) = 0 (2)
∂r ∂z
Fig. 1 Geometry of the rotating cylinders

vθ2 1 ∂p
= (3)
r ρ ∂r

∂(vθ ) ∂(vθ ) vvθ 1 ∂ ∂ vθ
vr + vz + = 2 r 2 τ0 + kr (4)
∂r ∂z r ρr ∂r ∂r r
∂(vz ) ∂(vz ) 1 ∂p 1 ∂ ∂vz
vr + vz =− + (r [τ0 + k ]) (5)
∂r ∂z ρ ∂z ρr ∂r ∂r
where vz , vr , vθ are the velocity components in z-, r-, θ -directions, respectively.

The boundary conditions for inner and outer cylinders rotating with different
speed in the same direction are
For z ≥ 0 and r = r1 , vr = vz = 0 and vθ = ωr1 (s1 )

For z = 0 and r1 < r < r2 , vz = v0
p = p0 at z = 0
The boundary conditions for inner and outer cylinders rotating with different
speed in the opposite direction are

For z = 0 and r1 < r < r2 , vz = v0
p = p0 at z = 0
Using the boundary conditions, the continuity Eq. (2) can be expressed in the
following integral form:
r1
2 r vz dr = (r22 − r12 )v0 (6)

r2
Let us non-dimensionalize the parameters are

r vz ρvr r2 vθ r1
η= ,U = ,V = ,W = ,R = ,
r2 v0 μr ωr1 r2
p − p0 2vz (1 − R) τ0 r2
P= ,ζ = ,B = ,
ρv02 r2 Re kv0
2ρ(r2 − r1 )v0
Re = ,
k

2Ω ρ r1 (r2 − r1 )3
2 2 2
ωr1
Ta = , where μr = k
μr2 (r1 + r2 ) r2
By introducing the above non-dimensional parameters, in the governing Eqs. (2)–

(5) as well as in boundary conditions (6), they become
∂V V ∂U
+ + =0 (7)
∂η η ∂ζ
W2 Re2 (1 − R) ∂ P
= (8)
η 2(1 + R)T a ∂η
∂W ∂W VW ∂2W 1 ∂W W 2B
V +U + = + − 2 + (9)
∂ζ ∂ζ η ∂η2 η ∂η η η
∂U ∂U ∂P 1 ∂U ∂ 2U B
V +U =− + + + (10)
∂η ∂ζ ∂ζ η ∂η ∂η2 η
1
2 ηU dη = (1 − R 2 ) (11)
R
The boundary conditions for cylinders rotating with different speed in the same
direction, in the dimensionless form, are
For ς ≥ 0 and η = R, V = U = 0 and W = 1

For ς ≥ 0 and η = 1, V = U = 0 and W = 2/R
For ς ≥ 0 and R < η < 1, U = 1
P = 0 at ς = 0
The boundary conditions for cylinders rotating with different speed in the opposite
direction, in the dimensionless form, are
For ς ≥ 0 and η = R, V = U = 0 and W = 1

For ς ≥ 0 and η = 1, V = U = 0 and W = −2/R
For ς ≥ 0 and R < η < 1, U = 1
P = 0 atς = 0
3 Numerical Solution
The numerical investigation and the scheme of solution can be considered as a round-
about expansion work of Coney and El-Shaarawi [16]. It has been decided to solve
the above system of governing equations, using finite difference method. Considering
the mesh network of Fig. 2, the following difference representations are made.
Fig. 2 Meshwork for the

finite difference technique
The discretized solution space is shown in Fig. 2 in which η and ζ correspond

to the grid size along the radial direction and also axial direction, respectively. By
assuming appropriate finite difference equations, the non-dimensionalized system of
Eqs. (7)–(11) got reduced as

(n+1) R + iη 4η 2R + (2i + 1)η
V(i+1) = Vi(n+1) − ∗
R + (i + 1)η 4ς R + (i + 1)η

(n+1)
Ui+1 + Ui((n+1) − Ui(n+1) − Ui(n) (12)
(n+1) (n+1)
[W 2 ]i(n+1) (1 − R)Re2 Pi − Pi−1
= (13)
R + iη 2T a(1 + R) η
(n+1) (n) (n+1) (n)
Wi+1 + Wi+1 − Wi−1 − Wi−1 Wi(n+1) − Wi(n)
Vi(n) + Ui(n)
4W ς
(n+1) (n)
Vi(n) Wi(n) Wi+1 + Wi+1 − 2Wi(n+1) − 2Wi(n)
+ = Vi(n)
R + iη 2(η)2
(n) (n+1)
Wi−1 − Wi−1 Wi(n)
+ −
2(η)2 (R + iη)2
(n+1) (n) (n+1) (n)
Wi+1 + Wi+1 − Wi−1 − Wi−1 2B
+ Vi(n) + (14)
(R + iη)4η R + iη
(n+1) (n+1) (n+1)
Ui+1 − Ui−1 Ui(n+1) − Ui(n) Pi+1 − Pi(n)
Vi(n) + Ui(n) =−
2η ς ς
(n+1) (n+1) (n+1)

Ui+1 − Ui−1 Ui+1 − 2Ui(n+1) − Ui−1
(n+1)
B
+ + + (15)
(R + iη)2η 2(η)2 R + iη
where i = 0 at η = R and i = m at η = 1, the application of trapezoidal rule to Eq. (12)

gives

η m−1
1 − R2
RU0(n) + U(m)
(n)
+ η U1(n) (R + iη) = (16)
2 i=1
η
The boundary conditions give for both cases
U0(n) = U0(m) = 0
m−1
1 − R2
η U1(n) (R + iη) = (17)
i=1
η
For both the cases of rotation, i.e., rotating in the same direction and opposite
direction of the annular cylinder, the above set of Eqs. (12)–(15) and (17) have been
solved using an iterative process. At first, Eq. (14) has been solved at the n = 0
columns, and i varied from 1 to m − 1 to obtain the system of linear equations
which is solved using the Gauss Jordan method to obtain the tangential velocity.
Equations (13), (15) and (17) have been solved under similar conditions to determine
a system of linear equations. This system has been solved by the Gauss-Jordan method
to obtain the axial velocity and pressure at the adjacent column n = 1. Finally, the
radial velocity is obtained from Eq. (12) by Gauss-Jordan method using the known
values of U. Repeating the above procedure, we obtain the result column by column,
along the axial direction of the annulus until the flow becomes axially and tangentially
fully developed. The above equations, to computationally solve, lead to the following
results.
Numerical solutions have been obtained for some permissible values of Bingham
number B and aspect ratio R. For the aspect ratio R = 0.3–0.8, the ratio of Reynolds
number to Taylor number as Rt = 20, 30; ζ = 0.02, 0.03 and η = 0.007, 0.002 was
fixed, respectively. The velocity profiles and pressure distribution along the radial
direction have been computed for varying Bingham numbers, B = 0, 10, 20, 30 for
two different cases namely, case 1: Rotating cylinders in the same direction and case
2: Rotating cylinders in opposite direction. The cylinders are rotating with different
speed when the outer cylinder rotating speed increases in both cases. Numerical
results for case 1 and case 2 are depicted in Figs. 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13,
Fig. 3 Tangential velocity for different Bingham number when R = 0.3
14, 15, 16, 17, 18, 19, 20, 21 and 22, respectively.
The tangential velocity profile of cylinders that are rotating with different speeds
in same and opposite directions is shown in Figs. 3, 4, 13 and 14, respectively, for
various values of Bingham number and aspect ratio. As the outer cylinders rotating
speed increases, in the case of cylinders rotating in the same direction, the tangential
velocity increases with an initial downturn of the increasing values from the inner
wall, and then, it continues increasing toward the outer wall of the annulus when
Fig. 5 Axial velocity for different Bingham number when R = 0.3
the annular space is small. However, in the case of cylinders rotating in opposite
direction, the tangential velocity decreases with an initial rise from the inner wall
and reaches its minimum at different radial positions for different Bingham numbers,
and then, it steadily increases as it approaches the outer wall. It is also observed that in
both the cases, the maximum tangential velocity value near the outer wall decreases
for different Bingham numbers when we increase the aspect ratio from R = 0.3 to
0.8, i.e., as the annular space decreases.
Fig. 7 Radial velocity for different sections when R = 0.3
Figures 5, 6, 15 and 16 show the axial velocity distribution for different values of
Bingham number and aspect ratio for both the cases. Increasing the outer cylinders
rotating speed, irrespective of whether the cylinders are rotating in the same or
opposite directions, for a fixed Bingham number and aspect ratio, the axial velocity
component increases from the inner wall to reach its maximum at a particular position
and then decreases as it moves toward the outer wall of the annulus. For large annular
space at R = 0.3, when Bingham number increases, the increase in axial velocity is
found to be significant. For smaller annular space at R = 0.8, the increase in axial
Fig. 9 Radial velocity for different Bingham number when R = 0.3
velocity is found to be negligible as Bingham number increases. Moreover, it could

also be observed that the increase in the maximum value of the axial velocity at
a particular radial position decreases as there is an increase in the outer cylinders
rotating speed.
The radial velocity V for the aspect ratios R = 0.3 and 0.8 at different points in the
axial direction for both the cases is shown in Figs. 7, 8, 17 and 18, respectively. It has
Fig. 11 Pressure for different Bingham number when R = 0.3
been found that for any particular Bingham number, the radial velocity is positive
near the inner wall and negative in the region near the outer wall. Also, the values
of radial velocity decrease as we progress in the axial direction. Furthermore, at any
particular cross-section in the axial direction, the radial velocity for a larger annular
space is found to be higher than that of the smaller annular space. From Figures
depicted in 9, 10, 19 and 20, it could be observed that at any particular cross-section
in the axial direction, the radial velocity increases with an increase in the Bingham
number near the inner wall of the annulus when the aspect ratio is small. Moreover,
the radial velocity decreases toward the outer wall for higher values of Bingham
number when the outer cylinders rotating speed increases.
For cylinders rotating in the same direction or in the opposite direction with
different speeds, the pressure distributions computed for various values of Bingham
number and aspect ratio are depicted in Figs. 11, 12, 21 and 22, respectively. The
pressure is found to increase from the inner wall to the outer wall of the annulus, not
depending on the Bingham number or the aspect ratio. It has been observed that the
pressure increases as the Bingham number increases for any aspect ratio. Moreover,
for large annular space when R = 0.3, the increase in pressure for higher Bingham
number is found to be significant than that of the smaller annular space. Also, it has
been observed that the pressure is higher for larger aspect ratios, i.e., when space
is small. When increasing the speed of the outer cylinders, the values, which are
increasing, decrease as the pressure increases.
For the particular case of the inner cylinder rotating and the outer cylinder at
rest, the results were found to match the results of Srinivasa Rao [15]. In the case of
stationary concentric annular cylinders, the results of axial velocity components in
the analysis matched well with the results of Kandasamy [17].
5 Conclusion
In this study, the development of different velocity profiles and pressure distributions
are numerically computed. The following conclusions are observed when Bingham
number and aspect ratios are increased in the entrance region of the annular space
when the concentric cylinders are rotating in the same as well as opposite directions
with different speeds.
• By comparing the results obtained, a significant change in the tangential velocity
profile and pressure profile could be noticed, and it increases rapidly when cylin-
ders are rotating with different speeds rather than the same speed in both cases.
However, the profile corresponding to radial and axial velocity remains almost
similar in both cases.
• The entrance length distance decreases as velocities increase, and the flow
becomes fully developed before the usual or expected time.
• As Bingham number increases, time taken for the Bingham fluid to flow between
the rotating annulus is reduced, resulting in enhanced surface area contact which
is ascertained so as to reduce the friction in the flow region of the annulus.
• Bingham fluid shall be used as working fluid in transmission systems and flow
characteristics which rapidly dissipate the heat generated due to friction.
• The performance level increases when annular cylinders rotate in the same as
well as opposite directions with different speeds by using time-independent non-
Newtonian fluid obeying Bingham fluid.
Acknowledgements The authors thank the Principal and management of SSN institutions for their
support in carrying out this work.
References
1. Schlichting H, Gersten K (2000) Boundary layer theory, 8th edn. Springer

2. Escudier MP, Oliveira PJ, Pinho FT (2002) Fully developed laminar flow of purely viscous
non-Newtonian liquids through annuli, including the effects of eccentricity and inner-cylinder
rotation. Int J Heat Fluid Flow 23:52–73
3. Ramesh Gupta C (2001) On developing laminar non-Newtonian flow in pipes and channels.
Nonlinear Anal Real World Appl 2:171–193
4. Keslerova R, Kozel K (2010) Numerical modelling of incompressible flows for Newtonian and
non-Newtonian fluids. Math Comput Simul 80:1783–1794
5. Poole RJ, Chhabra RP (2010) Development length requirements for fully developed laminar
pipe flow of yield stress fluids. J Fluid Eng 132:1–4
6. Huilgol RR (2006) A systematic procedure to determine the minimum pressure gradient
required for the flow of viscoplastic fluids in pipes of symmetric cross section. J Non-Newton
Fluid Mech 136:140–146
7. Khali S, Nebbali R, Bouhadef K (2017) Effect of a porous layer on Newtonian and power-law
fluids flows between rotating cylinders using lattice Boltzmann method. J Braz Soc Mech Sci
Eng 39:3881–3895
8. Nouri JM, Whitelaw JH (1994) Flow of Newtonian and Non-Newtonian flu ids in a concentric
annulus with rotation of the inner cylinder. J Fluid Eng 116:821–827
9. Chebbi R (2002) Laminar flow of power-law fluids in the entrance region of a pipe. Chem Eng
Sci 57:4435–4443
10. Round GF, Yu S (1993) Entrance laminar flows of viscoplastic fluids in concentric annuli. Can
J Chem Eng 71:642–645
11. Schulz A, Pfister G (2003) The effects of outer cylinder rotation on Taylor-Couette flow at
small aspect ratio. Phys Fluids 15:417–425
12. Santhosh N, Radhakrishnamacharya G, Chamkha AJ (2015) Effect of slip on Herschel Bulkley
fluid flow through narrow tubes. Alexandria Eng J 54:889–896
13. Lu G, James Third R, Christoph Mller R (2014) Effect of wall rougheners on cross-sectional
flow characteristics for non-spherical particles in a horizontal rotating cylinder. Particuology
12:4453
14. Pai RG, Kandasamy A (2014) Entrance region flow of Herschel-Bulkley fluid in an annular
cylinder. Int J Appl Eng Res 9:7083–7101
15. Nadiminti SR, Kandasamy A (2016) Entrance region flow heat transfer in concentric annuli
with rotating inner wall for bingham fluid. J Comput Appl Mech 11(2):137–157
16. Coney JER, El-Shaarawi MAI (1974) A contribution to the numerical solution of developing
laminar flow in the entrance region of concentric annuli with rotating inner walls. ASME Trans
J Fluid Eng 96:333–340
17. Kandasamy A (1996) Entrance region flow heat transfer in concentric annuli for a Bingham
fluid. In: Third Asian-Pacific conference on computational mechanics, pp 1697–1702
Nanofluids in Improving Heat Transfer
Characteristics of Shell and Tube Heat
Exchanger
S. Seralathan, R. Vijay, S. Aravind, S. Sivakumar, G. Devaraj, V. Hariram,

P. S. Raghavan, and T. Micha Premkumar
Abstract The thermal performances of shell and tube heat exchanger [STHE] are
investigated using Ag–W and CuO–W nanofluids with suspended particle volume
concentrations between 0.02% and 0.06% of Ag and CuO. The comparison is made
with reference to base fluid. The result revealed that thermal conductivity of the
nanofluids, which is dependent on the particle volume concentration, influenced the
heat transfer ability. Highest overall as well as convective heat transfer coefficients
and highest actual heat transfer are obtained for 0.06% volume concentration CuO–W
nanofluid. An improvement of about 19% in heat transfer coefficient is recorded for
0.06 vol% of CuO–W nanofluids with respect to water. Also, the overall heat transfer
coefficient enhanced between 64 and 79% for 0.06 vol% of CuO–W nanofluids.
However, in the tube-side, pressure drop increases with increase in nanofluid volume
concentrations. Actual heat transfer improved by 39–56% with reference to water. It
can be concluded that better heat transfer characteristic for the STHE is obtained by
keeping the shell-side mass flow as fixed and varying the tube-side mass flow rates.
Keywords Heat exchanger · Shell and tube · Nanofluid · Ag–W · CuO–W
Nomenclature
α Thermal diffusivity (m2 /s)

ρ Density (kg/m3 )
S. Seralathan (B)
Department of Aeronautical Engineering, Hindustan Institute of Technology and Science,
Chennai, India
R. Vijay · S. Aravind · S. Sivakumar · G. Devaraj · V. Hariram · T. Micha Premkumar
Department of Mechanical Engineering, Hindustan Institute of Technology and Science, Chennai,
India
P. S. Raghavan
Department of Chemistry, Hindustan Institute of Technology and Science, Chennai, India

https://doi.org/10.1007/978-981-15-4488-0_43
520 S. Seralathan et al.
μ Dynamic viscosity (N s/m2 )

γ Kinematic viscosity (m2 /s)
T LMTD Logarithmic mean temperature difference (°C)
A Heat transfer area [A = π Di L] (m2 )
Cp Specific heat capacity (J/kgK)
k Thermal conductivity (W/mK)
T Temperature (K)
V Velocity of the nanofluid (m/s)
Subscripts
f Base fluid (water)

wall Tube wall
out Outlet
in Inlet
c Cold side
nf Nanofluid
h Hot side
w Water
1 Introduction
Designing energy-efficient thermal energy system along with enhanced heat transfer
is one of the vital requirements in present energy conservation scenario. Shell and tube
heat exchanger [STHE] are predominately deployed in variety of process industries,
oil refineries, etc. The performance of STHE could be enhanced by augmenting
the heat transfer characteristic of the working fluids used by suspending it with
nanoparticles. This colloidal solution, also called ‘nanofluid,’ contains nanoparticles
suspended uniformly in a base fluid. Nanofluid is an innovative class of fluid with
enhanced thermal properties to meet the heat transfer requirements. The thermal
conductivity and viscosity of nanofluid are higher compared to base fluid, and these
properties rely on various parameters, namely nanoparticles mass fraction, particle
size, and temperature. The widely used nano powders are Fe, Ag, Ti, Al, and Cu in
its pure form and CuO, Al2 O3 , TiO2 , Fe2 O3 in the oxide form. Consequently, with
the use of nanofluid, compact heat transfer equipments may find its place in many
industrial applications.
Lee et al. [1] used 4.0% vol. CuO–ethylene glycol nanofluid and found that the
thermal conductivity improved by 20%. On the other hand, Das et al. [2] obtained
improved thermal conductivity using CuO nanoparticles suspended in water. Li and
Xuan [3] reported improved heat transfer by using smaller volume fraction of CuO
Nanofluids in Improving Heat Transfer Characteristics … 521
nanoparticles. Godson et al. [4] used different Ag particle volume concentrations

in Ag–W nanofluid and reported 12.41% improvement in convective heat transfer
coefficient. The review of literatures [5–8] revealed a variety of metal oxide-based
nanoparticles suspended in base fluid, namely ethylene glycol, mineral oil, and water
which were used to augment the heat transfer characteristics of STHE. A detailed
review by Das et al. [9] on nanoparticles suspended in base fluid revealed higher heat
conduction, enhancement of thermal conductivity as well as better stability. Besides
improvement in thermal conductivity by dispersed nanoparticles, it is understood
that boundary layer suppression, intensification of turbulence also aid in improving
the heat transfer ability. All the existing studies by researchers focused on using
metal oxide-based nanoparticles with volume fractions varying from 0.01 to 6% in
the base fluid [5–11]. The use of high concentrations led to drop in pressure which
required additional pumping resource. Moreover, a limited research was carried out
in using pure metal oxide like Ag at lower concentrations levels in a STHE.
Therefore, the focus of this present study is to use a pure metal nanoparticle,
namely silver nanoparticles dispersed in water (Ag–W) with volume concentra-
tions of nanoparticles varying from 0.02 to 0.06% in a STHE. A metal oxide-based
nanoparticle namely copper oxide dispersed in water (CuO–W) with volume con-
centrations of nanoparticles varying from 0.02% to 0.06% is also used. All the
comparison in this study is carried out with respect to base fluid (water).
2 Preparation of Ag–Water and CuO–Water Nanofluids
Silver nitrate (AR grade) is dissolved in a minimum quantity of distilled water. Gum
Acacia is used as the dispersing matrix. 0.25% of Gum Acacia is dissolved in another
beaker. Both are mixed together and stirred for 30 min. 0.25% of di-ethanol amine is
dissolved in water and added in drops to the above content under vigorous stirring.
A silver–amine complex is formed with light brown color. This is followed by the
addition of glucose solution in drops which leads to the formation of nanosilver
particles dispersed in water. The final nanosilver suspension contains 5000 ppm
of silver particles. The transmission electron microscopy (TEM) photograph of the
nanosilver is displayed in Fig. 1a, and the energy dispersive X-ray (EDAX) analysis
(chemical analysis) result is shown in Fig. 2a. The nanoparticles are found to be
dispersed well in certain agglomeration.
Similarly, calculated quantity of copper nitrate is dissolved in a minimum quantity
of distilled water. 0.25% of Gum Acacia is dissolved in distilled water and added to
the above solution under vigorous stirring. After 30 min, 25% of liquor ammonia is
added in drops to form copper–amine complex followed by the addition of reducing
agent (hydrazine) to form nano CuO dispersed in water. The concentration of CuO
in the final suspension is 5000 ppm. The TEM photograph of the nanocopper is
shown in Fig. 1b reveals that nanoparticles are found to be well dispersed in certain
aggregation. The EDAX analysis (chemical analysis) result is shown in Fig. 2b. The
Fig. 1 TEM photograph a nanosilver and b nanocopper
Fig. 2 EDAX analysis a nanosilver and b nanocopper
Ag–W and CuO–W nanofluids prepared are treated using ultrasonicator to ensure
continuous suspended, uniform and stable colloidal solution.
3 Experimentation
The graphic arrangement of experimental setup is depicted in Fig. 3, and its specifi-
cations are listed in Table 1. It consists of two loops, one for hot fluid (water) which
flows through the tube and another for cold fluid (Ag–W/CuO–W/W), which flows
through the shell. Each loop has a reservoir, pump, flow meter, and a valve to regu-
late the flow rate in the system. Four ‘k’ type thermocouples with error measurement
Fig. 3 Schematic layout of STHE
Table 1 STHE specification

Specification Tube Shell
Diameter inner (Di ) (mm) 8 145
Diameter outer (Do ) (mm) 12 200
Length (L) (mm) 475 500
Number of tubes 9 –
Material Mild steel Copper
(±1 °C) are used to record the temperatures at outlet as well as inlet of the tube
and shell, respectively. Digital flow meters determine the flow rate of the working
fluids. STHE is shielded well to prevent any heat rejections to the surroundings.
All the electrical signals received from thermocouples and digital flow meters are
sent to data logger as readable information using LABVIEW software. The raw data
obtained during experimentation is used to estimate the heat transfer rate, Nusselt
number, Peclet number, and convective as well as overall heat transfer coefficients of
nanofluids with varied volume concentrations of Ag nanoparticles and CuO nanopar-
ticles. Uncertainty analysis for the measurements recorded using the instruments is
carried out, and the maximum error is 3.89% for overall heat transfer coefficient.
The influence of varying the mass flow rates of shell-side fluid and tube-side fluid
on the heat transfer characteristics of STHE is studied and analyzed. Initially, the
tube-side fluid mass flow rate (mTube_Side_Fluid ) is varied by keeping the shell-side fluid
mass flow rate (mShell_Side_Fluid ) as constant. Later, mShell_Side_Fluid is changed keeping
the mTube_Side_Fluid as constant. Experiments are carried out initially using water in
the shell-side and later with Ag–W as well as CuO–W nanofluids with different
particle vol.% concentrations. The data are extracted after the fluid flow attains fully
developed conditions. The heat transfer rate convective as well as overall heat transfer
coefficients and Nusselt number are found using the formulae given below.
Heat transfer rate of nanofluid (Qnf ) is expressed as

Q nf = mCpnf Tnf_out − Tnf_in (1)
Convective heat transfer coefficient of nanofluid (hnf ) is estimated from the overall
heat transfer coefficient (U) of STHE which includes hot fluid namely nanofluid and
cold fluid (water) thermal resistances along with tube wall.

Do
1 1 Di ln Di Di 1
= + + × (2)
U h nf 2K wall Do hf
Overall heat transfer coefficient (U) is found out from the expression
Q = AU TLMTD (3)
The LMTD (TLMTD ) is found from the expression
(Thnf,in − Tcw,out ) − (Thnf,out − Tcw,in )

TLMTD = (4)
−Tcw,out
ln TThnf,in
hnf,out −Tcw,in
Nusselt number of the nanofluids (Nunf ) is expressed as
h nf Di
Nunf = (5)
K nf
Nunf is found using Dittus and Boelter’s [12] single-phase liquid relation which
is also suggested by Azmi et al. [13]
0.40
Nu = 0.023Re0.80 Pr (6)
Prandtl number (Prnf ) and Reynolds number (Renf ) for the nanofluid inside the
tube are expressed as follows
V Di
Renf = (7)
γnf
γnf C pnf μnf
Pr = = (8)
nf αnf K nf
Thermal diffusivity (α nf ) of nanofluid is estimated using the expression

K nf
αnf = (9)
ρnf C pnf
4.1 Tube Side Fixed Shell Side Varied Flow (Case: TSFSSV
Flow)
The shell-side fluid mass flow rates are changed between 110 and 150 kg/h, and at
the tube-side, the fluid is kept constant as 150 kg/h. Figure 4 illustrates the varia-
tions of actual heat transfer for different mass flow rates for varied particle volume
concentrations of Ag–W and CuO–W nanofluids. Actual heat transfer (Q) enhances
with rise in mShell_Side_Fluid . Adding up nanoparticle to the base fluid enhances ‘Q’,
and it augments with enhancement in volume concentrations of the nanoparticles.
Heat transfer rate improved by 36–50% for Ag–W nanofluid, whereas for CuO–W
nanofluid, it improved by 84–91% with reference to water. As can be noticed in
Fig. 4, CuO–W nanofluid with 0.06% vol. concentration gives the maximum ‘Q’ for
mShell_Side_Fluid and mTube_Side_Fluid of 150 kg/h, respectively.
The change of overall heat transfer coefficients (U) against mShell_Side_Fluid is plot-
ted in Fig. 5. The highest ‘U’ is observed for an mShell_Side_Fluid of 150 kg/h and
mShell_Side_Fluid of 150 kg/h. The CuO–W nanofluid with 0.06% particle volume con-
centration has a better thermo physical properties, namely specific heat, density,
and thermal conductivity, which is reflected in terms of highest ‘U’. Overall heat
transfer coefficient enhanced by 27–47% for Ag–W nanofluid, whereas for CuO–W
nanofluid, it is found to be 70–93%.
Fig. 4 Case: TSFSSV flow-variations of actual heat transfer

Fig. 5 Case: TSFSSV flow-variations of overall heat transfer coefficient
It is known that convective heat transfer coefficient (h) improves with diminution
in thermal boundary layer thickness. Moreover, ‘h’ is influenced by nanofluid thermal
conductivity. Figure 6 depicts the changes in convective heat transfer coefficient and
mShell_Side_Fluid keeping tube-side fluid as constant. Around 8–16% rise in ‘h’ value is
observed for Ag–W nanofluid. On the other hand, it is 19–28% for CuO–W nanofluid.
Highest ‘h’ value is realized for 0.06% vol. concentration of CuO–W nanofluid as
its thermal conductivity is high for this particle volume concentration.
Figure 7 depicts the variation between overall heat transfer coefficient and Peclet
number (Pe) for the varying mShell_Side_Fluid keeping mTube_Side_Fluid as constant. Over-
all heat transfer coefficient improves with rise in ‘Pe’ and also with enhancement in
particle volume concentrations. Around 5–10% improvement of ‘U’ is observed for
Fig. 6 Case: TSFSSV flow-variations of convective heat transfer coefficient

Fig. 7 Overall heat transfer coefficient versus Pe
Ag–W nanofluid with respect to water. It improved by 8–12% for CuO–W nanofluid.
0.06% vol. concentration CuO–W nanofluid shows the maximum overall heat trans-
fer coefficient among all nanofluids and base fluid. Figure 8 depicts the variations
between Nusselt number (Nu) and Pe. Nusselt number of all nanofluids is higher
than water. Nusselt number improves with enhancement in particle volume con-
centrations. Nusselt number is found to improve by 15–27% for Ag–W nanofluid.
Similarly, for CuO–W nanofluid, it is improved by 36–42%. Highest ‘Nu’ is realized
for 0.06% vol. concentration of CuO–W nanofluid.
Fig. 8 Nu versus Pe
Fig. 9 Case: SSFTSV flow-variations of actual heat transfer
4.2 Shell Side Fixed Tube Side Varied Flow (Case: SSFTSV
Flow)
The tube-side fluid mass flow rates are altered between 110 and 150 kg/h, and at
the shell-side, the fluid is kept constant as 150 kg/h. Figure 9 depicts the variations
between actual heat transfer with varying mTube_Side_Fluid . Highest actual heat transfer
is realized for 0.06% vol. concentration CuO–W nanofluid at mShell_Side_Fluid as well as
mTube_Side_Fluid of 150 kg/h, respectively. Actual heat transfer is found to be improved
by 26–34% for Ag–W nanofluid. But, it is improved by 39%–56% with reference to
base fluid for CuO–W nanofluid.
Figure 10 highlights the variations between overall heat transfer coefficients
against varied mTube_Side_Fluid . The ‘U’ value improved by 51–57% for Ag–W
nanofluid, whereas for CuO–W nanofluid, it is improved significantly by 64–79%.
Again, the highest ‘U’ value is realized for 0.06% vol. concentration CuO–W
nanofluid at mShell_Side_Fluid as well as mTube_Side_Fluid of 150 kg/h. This is due to
improved thermal conductivity of the 0.06% vol. concentration CuO–W nanofluid.
Figure 11 depicts the variations in convective heat transfer coefficient with var-
ied mTube_Side_Fluid . The ‘h’ is improved by 8% for Ag–W nanofluid, whereas it is
improved by 19% with respect to water. As the thermo physical property of 0.06%
vol. concentration CuO–W is better, maximum ‘h’ value is realized for 0.06% vol.
concentration CuO–W nanofluid.
The consequence of nanofluid concentrations on the pressure drop is studied for a
constant mass flow rate (i.e., m = 130 kg/h) is plotted in Fig. 12. The results revealed
that pressure drop increases with enhancement in nanofluid volume concentrations.
This pattern observed correlates well with earlier findings by Kabeel and Mohamed
Abdelgaied [10].
Fig. 10 Case: SSFTSV flow-variation of overall heat transfer coefficient
Fig. 11 Case: SSFTSV flow-variation of convective heat transfer coefficient
5 Conclusions
This experimental investigation deals with thermal performance of a STHE using

Ag–W and CuO–W nanofluids with particle volume concentrations varied from 0.02
to 0.06%. The comparison is made with reference to base fluid (water). The result
highlighted that thermal conductivity of the nanofluids which rely on the particle
volume concentration influenced the heat transfer ability. Highest overall as well as
convective heat transfer coefficients and highest actual heat transfer are obtained for
0.06% vol. concentration CuO–W nanofluid. A better heat transfer characteristic for
Fig. 12 Variations of pressure drop with respect to nanofluid volume fraction
the STHE is realized by ensuring the mass flow at shell-side as fixed and varying the
tube-side mass flow rates [14]. An improvement of about 19% in heat transfer coef-
ficient is recorded for 0.06% vol. concentration of CuO–W nanofluids with respect
to water, i.e., base fluid. Also, the overall heat transfer coefficient improved between
64 and 79% for 0.06% vol. concentration of CuO–W nanofluids. However, in the
tube-side, pressure drop increases with increase in nanofluid volume concentrations.
Actual heat transfer also improved by 39–56% with respect to water. Hence, use
of nanofluids as a working fluids in STHE results in an improved energy-efficient
thermal system.
References
1. Lee S, Choi SUS, Li S, Eastman JA (1999) Measuring thermal conductivity of fluids containing
oxide nanoparticles. J Heat Transf 121(2):280–289
2. Das SK, Putra N, Thiesen P, Roetzel W (2003) Temperature dependence of thermal conductivity
enhancement for nanofluids. J Heat Transf 125 (4):567–574
3. Li Q, Xuan Y (2000) Experimental investigation of transport properties of nanofluids. Heat
transfer science and technology. Buxan Wang edn. Higher Education Press, pp 757–762
4. Godson L, Deepak K, Enoch C, Jefferson B, Raja B (2014) Heat transfer characteristics of
silver/water nanofluids in a shell and tube heat exchanger. Arch Civil Mech Eng 14:489–496
5. Shahrul IM, Mahbubul IM, Saidur R, Khaleduzzaman SS (2016) Performance evaluation of a
shell and tube heat exchanger operated with oxide based nanofluids. Heat Mass Transf 52:1425–
1433
6. Leong KY, Saidur R, Mahlia TMI, Yau YH (2012) Modeling of shell and tube heat recovery
exchanger operated with nanofluid based coolants. Int J Heat Mass Transf 55:808–816
7. Zeinali Heris S, Etemad SG, Nasr Esfahany M (2006) Experimental investigation of oxide
nanofluids laminar flow convective heat transfer. Int Commun Heat Mass Transf 33:529–535
8. Nishant K, Sonawane SS (2016) Experimental study of thermal conductivity and convective

heat transfer enhancement using CuO and TiO2 nanoparticles. Int Commun Heat Mass Transf
76:98–107
9. Das SK, Choi SUS, Patel HE (2008) Heat transfer in nanofluids- a review. Heat Transf Eng
27(10):3–19
10. Kabeel AE, Abdelgaied M (2016) Overall heat transfer coefficient and pressure drop in a typical
tubular exchanger employing alumina nano-fluid as the tube side hot fluid. Heat Mass Transf
52:1417–1424
11. Saravanna JY, Kantamnen R, Fasil N, Sivamani S, Hariram V, Micha Premkumar T, Mohan T
(2017) Modelling and analysis of water heating using recovered waste heat from hot flue gases
of chulha. ARPN J Eng Appl Sci 12(21):6164–6171
12. Dittus FW, Boelter LMK (1930) Heat transfer in automobile radiators of the tubular type, vol
2. University of California Publications on Engineering, California
13. Azmi WH, Usri NA, Rizalman M, Sharma KV, Noor MM (2017) Force convection heat transfer
of Al2 O3 nanofluids for different based ratio of water ethylene glycol mixture. Appl Therm
Eng 112:707–719
14. Shahrul IM, Mahbubul IM, Saidur R, Khaleduzzaman SS, Sabri MFM, Rahman MM (2014)
Effectiveness study of a shell and tube heat exchanger operated with nanofluids at different
mass flow rates. Numer Heat Transf Part A Appl: Int J Comput Methodol 65:699–713
Design and Numerical Analysis
of Gearless Transmission Used in Small
Wind Turbine
Micha T. Premkumar, V. Hariram, S. Seralathan, Pinku Kumar,

and Godwin John
Abstract The main function of the gearless transmission system is to transmit power
between two shafts through the sliding links that develop revolute pair between the
hubs of two shaft. Normally the output link is bent at any angle between 0 and 180°.
Moreover, as the hub in the driving shaft rotates, holes in this hub rotate, which in turn
pushes the links and it drive the output. Moreover, torque developed in the small wind
turbine is small so it is effectively used in this low torque applications. The objective
of this project is to analysis gearless drive that is being simple and transmitting
power preciously at almost right angle without any bevel gears. This mechanism is
completely analysis using Ansys© software to lookout the elbow rod the hub under
different working condition. Numerical analysis is made by rotating the mechanism
at different rpm, viz. 0–150 rpm. Reaction forces and reaction moment are analyzed
for 5 s simulation and compared with allowable stress. In this investigation, the
effect of number of links and size of link is studied to find the permissible operating
speed of mechanism. The major outcome of this work is to get the permissible limit
of the stress, strain, speed, torque transmitting capacity of this mechanism as it is
very much attracting the research community in replacing the conventional bevel
gears. It is observed that the optimum number of link for better transmission is 8
and its diameter is 8 mm, and optimum speed of the mechanism is 100 rpm, so this
mechanism is suitable for the small wind turbine.
Keywords Gear less transmission · Vertical axis wind turbine · Numerical
M. T. Premkumar (B) · V. Hariram · P. Kumar

Department of Mechanical Engineering, Hindustan Institute of Technology and Science, Chennai
603103, India
S. Seralathan
Chennai 603103, India
G. John
Department of Automobile Engineering, Hindustan Institute of Technology and Science, Chennai
603103, India

https://doi.org/10.1007/978-981-15-4488-0_44
534 M. T. Premkumar et al.
1 Introduction
Gearless mechanism or also known as Elbow mechanism which transmits power

between driving and driven shaft at any fixed angle. It has the sliding link and rotat-
ing pair mechanism, which allows to transmit torque between two shafts. The slid-
ing link slides inside the holes in the hub of both driving and driven shaft which
forms mechanism of power transmission. This mechanism is the best alternate for
replacement of the bevel gear, so lot of researchers have an eye toward the gearless
transmission mechanism. Chung [1] presented a new type of gearless traction mech-
anism for high-speed elevators with a permanent magnet motor drive system and
explained gearless transmission mechanism which includes a multiplicity of similar
linkages, which includes off-center center of rotation between an input and an output
shaft. However, many reasons supporting that this type of bevel gear drives is not
favoring large mills due to not having sufficient VFD option [2], unlike universal
joints, performance losses are low even there is offset in shaft and smooth running
of operation [3]. Janakiraman and Paramasivam [4] and Bohra et al. [5] performed
comparative analysis using the gearless wind turbines which related to some control
techniques to address the offset issue for smooth operation. Moreover, Ghagare et al.
[6] tested the power transmission between the two with angular offset. The outcome
of this work is that the number of pin contributes more for the smoother operation
and unlike bevel gear, the force distribution is equal in this mechanism. Because of
uneven force distribution, bevel gear transmission system has high wear and tear,
and hence, the gearless mechanism has smooth and controlled acceleration without
having jerk in the transmission gearless mechanism or also known as Elbow mech-
anism which transmits power between driving and driven shaft at any fixed angle.
It has the sliding link and rotating pair mechanism, which allows to transmit torque
between two shafts. The sliding link slides inside the holes in the hub of both driving
and driven shaft which forms mechanism of power transmission. This mechanism is
the best alternate for replacement of the bevel gear, so lot of researchers have an eye
toward the gearless transmission mechanism [7].
The development of this type of mechanism has been explored the relatively
slackness between both gear drives and simple gearless drives. Moreover, recent
advancement in technologies addresses the possible disadvantage of the gear drive
[8]. The failure analysis under on the basis of bending stress for smooth running of
the gearless mechanism under heavy load condition reveals the failure stress of the
mechanism. The analysis has been done using ANSYS environment [9]. Moreover,
real-time study of gearless transmission mechanism is analyzed using solid works
software to study reaction stress in the elbow pin and hub, and then, the mechanism
is fabricated to permissible speed of the elbow rod and hub of the mechanism [10].
The application of this gearless transmission is inclined toward the wind turbine.
These are becoming promising solution for smooth and effective operation for the
wind turbine application [11]. Kiran et al. [12] and Yasarkhan et al. [13] analyzed the
effect of the elbow pin and hub interaction during operation of the mechanics. The
Design and Numerical Analysis of Gear Less Transmission … 535
Fig. 1 Schematic of vertical

axis wind turbine with
gearless transmission
mechanism
peak value of force and moments was analyzed and compared with the allowable
and designed stress.
Figure 1 shows the schematic of the vertical axis wind turbine system employed
with gearless transmission mechanism. The main objective of this paper is to design
and analyze the gearless mechanism of wind turbine which performs at low height
effectively. Prior to the implementation of vertical axis wind mills with gearless
technology, extensive analysis has been done using Ansys© software by changing
the parameter like number of pin and diameter of pin for various speed of the rotation.
In the next section, the methodology to do the analysis and validation of the model
is discussed. In third section, result and discussion of this work are presented.
2 Methodology
The commercially available FEA tool like Ansys© structural analysis is used in this
simulation to enable us to solve complex structural engineering mechanism. This
gearless transmission mechanism can easily model and analyze for better and faster
design optimization using this software. Ansys© mechanical has smart meshing tech-
nology, enabling to get optimum meshing quickly. Figure 2 shows the 3D meshing of
the gearless transmission mechanism. The boundary condition of rotation for driving
shaft and only one degree of freedom of sliding is allowed for the link. The mesh sen-
sitivity study is carried out using the automatic and intelligent algorithms to ensure
that high quality meshes were generated for this simulation. The validation of the
simulation is shown is Table 1. The validation has been done analytically, and the
error in the simulation is maximum 3.8%, which is in allowable limit. With this con-
fidence, further simulation is continued for different case in the gearless transmission
mechanism.
Fig. 2 3D mesh with boundary conditions
Table 1 Validation of
Software Numerical
numerical studies
Bending Bending Percentage
stress (MPa) stress (MPa) error
Link 11,049 11,337 2.5
Hub and 438 456 3.8
Shaft
In this work, three different cases of simulation have been done. In the first case,
the diameter of the link is 4 mm, and in the second case, the diameter of the link is
6 mm, similarly in the third case, the diameter of the link is 8 mm, and number of
link varied from 3, 5 to 9 for all three cases is shown in Table 2. In next section, the
Table 2 Consolidation of
CASE 1: No. of link 3
various cases studied in the
Diameter of Link is 4 mm fixed No. of link 7
analysis
No. of link 9
No. of link 9
No. of link 9
effect of changing the diameter and number of the link for different speed of driving
shaft is discussed.
In the gearless transmission system, number of links used must be odd 3,5,7,9, etc.,
and these links are fixed in the drilled holes of the hub at the both driving and driven
shaft ends because of which rotation is transferred. In this study, three cases with 4, 5
and 6 mm diameter of link are analyzed for different number of link, viz. 3, 7 and 9.
The minimum and maximum value of the total deformation, equivalent elastic strain
and equivalent stress of the 5 link and 7 link are almost same, so result of 5 link is not
discussed in this paper. Moreover, when the number of link is 11, the pitch distance
between the two links is very small which lead to more stress concentration, and
possibility of failure of mechanism is more. In this section, illustration of the 9 link
alone is presented, and consolidation results of all there link are presented in Table 3.
In Fig. 3, the von Mises yield principle put forward that the yielding of materials
starts when the second derivate stress tensor reaches a critical value and its value is
5.88 × 10−5 which is less than the other two cases. It is observed that the part of a
plasticity theory applies best to the materials, such as mild steel used in this study.
Moreover, prior to yield, material reaction can be presumed to be nonlinear elastic
or viscoelastic. The subset of in Fig. 3 is the enlarged view of enclosed area where
maximum strain is predicted because of yielding of materials under any loading
condition.
Figure 4 shows the deformation in the link and hub of the mechanism which is
Table 3 Consolidated list of total deformation, equivalent elastic strain and equivalent stress
Total Equivalent elastic Equivalent stress
deformation strain (von Mises) (Mpa)
(mm)
Min Max Min Max Min Max
CASE 1: No. of 3 link 0 1.5963 6.28e-9 3.57e-4 9.17e-4 54.925
Diameter of No. of 7 link 0 0.9731 6.28e-9 2.02e-4 4.86e-4 31.275
link is 4 mm
fixed No. of 9 link 0 0.8341 5.03e-9 1.85e-4 3.61e-4 28.665

link is 5 mm

link is 6 mm
Fig. 3 Equivalent elastic strain for case 3 of 6 mm diameter link and 9 number of link
Fig. 4 Total deformation for case 3 of 6 mm diameter link and 9 number of link
the conversion of a body from a reference configuration to an existing configuration

due to the sliding mechanism of the link inside the hub. An existing configuration is
a set that contains the positions due to the deformation of the body. This deformation
is caused by external loads like friction force of the sliding motion, body forces due
to the self-weight of the mechanism and the connecting load at the driven end. This
strain is an explanation of deformation in terms of relative position of particles in the
shaft that ignores rigid-body motions of the mechanism. In a continuous mechanism,
Fig. 5 Equivalent (von Mises) stress for case 3 of 6 mm diameter link and 9 number of link
a deformation represents the results from a stress field which is promoted by applied
forces inside the component of the sliding and rotating mechanism.
Figure 5 shows the von Mises stress developed in the link, and it is used to predict
yielding of materials under the given condition due to the both bending and twisting
momentum. It fulfills the property that two states of stress with same distortion
energy have similar von Mises stress. Moreover, the von Mises yielding principle is
not dependents of the first stress invariant, and it is applicable for the simulation of
plastic deformation of the material that chosen for this analysis. The von Mises yield
criterion in this study can be also expressed in terms of the von Mises stress, a scalar
stress value that can be calculated from the Cauchy stress tensor. The subset in Fig. 5
shows the enlarged view in the region where von Mises stress is more and no net
mass addition through the inlet or exit is permitted in this transmitting mechanism.
In Table 3, the consolidated list of total deformation, equivalent elastic strain
and equivalent stress for the three different cases is presented. It is observed in the
simulation that case of 6 mm diameter link with 9 number of link has minimum
maximum deformation, equivalent elastic strain and equivalent stress. So, this case
is selected for further analysis to optimize the speed of the driving shaft is shown
in Table 4. Rotational analysis is performed by operating the mechanism at various
speed from 0 to 150 rpm, reaction forces and reaction moment are measured for
5 s, and peak value of the reaction force and moments was measured and compared
with the allowable. It is found that the optimum permissible speed for smooth and
efficient operation is 100 rpm and reaches the allowable stress at this speed for the
factor of safety of 2.5. Hence, it is becoming a stimulating solution for small vertical
axis wind turbine applications.
Table 4 Variation of
S. No Speed (rpm) Equivalent stress (MPa)
equivalent stress with
rotational speed 1 0 0
2 25 10.85
3 50 30.2
4 75 60.72
5 100 96.12
6 125 110.13
7 150 120.32
4 Conclusion
Gearless transmission mechanism that is suitable for vertical axis wind turbine has
been analyzed using commercially available Ansys© software. Three different cases
were analyzed, and the effect changing the diameter of elbow rod and number of
elbow rod is being studied with time by using Ansys© post processor. It is found
that the case 3 with 6 mm diameter of elbow rod and 9 number of elbow rod has the
minimum value of maximum deformation, von Mises stress and equivalent strain. So,
the dimensioned is optimized with various case study. Using the optimum dimension,
the simulation has been done with various motor speed. The various motor speed was
defined with varied torque in the simulation. It is observed that deformation in hub is
very small at almost all values of rpm, whereas the link reaches its allowable stress
value at 100 rpm for factor of safety of 2.5. It shows that the smooth and safe running
of mechanism will occur when speed is below 100 rpm. Hence, it is concluded that
gearless transmission mechanism is efficiently running up to 100 rpm under normal
working condition.
Acknowledgements This research work is supported, in part by the Ministry of New and Renew-
able Energy, Government of India, New Delhi, India, vide Grant No. 24/6/2014-SWES(R&D)
dated: 30/12/2014, and the authors also gratefully acknowledge the support of Dr. Rajesh Katyal,
Mr, Haribhaskaran, Mr. David Solomon and Ms. Deepa Kurup. The authors are also grateful to Dr.
Uppu Ramachandraiah, HoD (EIE), HITS, for providing valuable suggestions.
References
1. Chung DW, Ryu HM, Lee YM, Kang LW, Sul SK, Kang SJ, Song JH, Yoon JS, Lee KH, Suh
JH (2001) Drive systems for high-speed gearless elevators. IEEE Ind Appl Soc 7(5):52–56
2. Jagushte GS, Mayur L, Rahul H, Prasad P, Tejas S (2017) Design and automation of bar cutting
machine. Int J Eng Sci Res Tech 3(2):658–661
3. Mahantesh T, Yapalaparvi SB, Inamdar DN, Chiniwalar GV (2014) Gearless power transmis-
sion offset parallel shaft coupling. Int J Res Eng Technol 3(3):52–56
4. Janakiraman R, Paramasivam S (2012) Model and design analysis of gearless PM stator-less
contra-rotation wind power generator. Int J Eng Res Technol 1(4):1–8
5. Bohra D, Bhaskar MK, Jain D, Manoj C, Vishvendra PG (2018) Performance analysis of

permanent magnet synchronous generator wind turbine for variable wind velocity. Int J Res
Appl Sci Eng Technol 6(1):2050–2059
6. Ghagare SD, Abhijeet SS, Anant DA, Abhijit MZ, Hanmant MK, Vahid MJ (2017) Hybrid
power generation using Maglev turbine. Int J Recent Res Asp 4(4):409–412
7. Danish SM, Moinoddin M, Nawazish SM (2015) Investigations on the fluid coupling of gearless
two wheeler. Int J Sci Eng Appl Sci 1(7):220–234
8. Kumar A, P P, R Sagar, Bist S (2015) Multi-angular gearless drive. Int J Sci Eng Res 6(7):974–
977
9. Gupta S, Sheikh AG, Adnan ZB, Pawan A, Kushal K, Mohsin AW (2018) Design and analysis
of gearless power transmission by rods using software simulation. Int J Innov Res Sci Eng
Technol 7(5):5215–5218
10. Somraj R, Sailesh B (2017) Design and fabrication of gearless power transmission for skew
shafts. Int Res J Eng Technol 4(4):1018–1023
11. Jagushte GS, Hrishikesh K, Vikas P, Vishal V (2016) Design, analysis and fabrication of gearless
transmission by elbow mechanism. Int J Sci Res Dev 4(2):688–689
12. Kiran S, Yadav SPS, Sandeep GM, Naik R (2017) Design, analysis and fabrication of gearless
power transmission by using elbow mechanism. Int J Eng Res Technol 6(4):739–741
13. Yasarkhan RS, Pankil P, Hiren P, Patel D (2017) A review paper on design and analysis of
gearless transmission mechanism using elbow mechanism. Int J Adv Mech Eng 7(1):31–40
Experimental Investigation of Unsteady
State Heat Transfer Behaviour
of Nanofluid
Yogiraj Bhumkar, A. R. Acharya, and A. T. Pise
Abstract This paper provides the method for preparation of nanofluids and
measuring thermal properties of nanofluids by various methods. In this paper, we
have focused on effect of variation of temperature, particle size and volume fraction
of nanofluid on properties of nanofluid such as density, viscosity, thermal conduc-
tivity and heat transfer coefficient. We have made comparison between CuO, ZnO
and SiO2 nanoparticles with particle size 30–50 nm and 50–70 nm and with three
volume fractions 0.5, 1 and 1.5%. The heat transfer coefficient (h) of all nanofluids is
determined from unsteady state heat transfer apparatus. The results indicate that the
thermal conductivity and h of nanofluid enhance with increase in %vol. fraction and
decrease in particle size. CuO–DI water nanofluid with 1.5% volume fraction and
30–50 nm particle size has higher heat transfer coefficient, and thermal conductiv-
ity is enhanced by 59.59% than de-ionized water. Hence, CuO–DI water nanofluid
(1.5% vol. fraction and 30–50 nm particle size) reaches steady state faster than other
nanofluids.
Keywords Preparation of nanofluid · Properties of nanofluid: Heat transfer

coefficient · Thermal conductivity · Viscosity · Density and specific heat ·
Unsteady state heat transfer behaviour (Unsteady state heating curves) of nanofluid
1 Introduction
Nanofluids have higher thermal properties than current heat transfer fluids. Hence,
nanofluid has future scope in heat transfer fluids. Metal nanoparticles have more
thermal conductivity than other nanoparticles but have less dispersion properties
and poor stability. Oxide nanoparticles have better dispersion properties and stable
suspension. Particle size and volume fraction of particles affect the properties of
nanofluid. Researchers have given different correlations for properties of nanofluids.
We have to observe important conclusions of some researchers.
Y. Bhumkar (B) · A. R. Acharya · A. T. Pise

Department of Mechanical Engineering, GCE, Karad, Maharashtra, India

https://doi.org/10.1007/978-981-15-4488-0_45
544 Y. Bhumkar et al.
Bashirnezhad et al. [1] concluded that viscosity increases with increase in volume
fraction of nanoparticles and decrease in temperature. Author reported the future
scope for doing experimental studies by acknowledging temperature, nanoparticle
size, volume fraction, sonication time and base fluid.
Liu et al. [2] observed that thermal conductivity of Cu–water nanofluid at 0.1%
volume fraction is increased by 23.8%. Researcher performed the experiment on
nanofluids without use of surfactant. Due to poor dispersion properties of nanofluid,
thermal conductivity depends on time interval.
Gupta et al. [3] provided the correlations for various properties of nanofluid.
Researcher reviewed the factors affecting the thermophysical properties of nanofluid.
Researcher has mentioned the future scope for cost effective and efficient nanofluid.
Albadr et al. [4] used the Al2 O3 nanoparticles of 30 nm particle size. The researcher
concluded that heat transfer coefficient (h) increases with increase in mass flow rate
and volume fraction. But friction factor increases with increase in vol. concentration
of nanoparticles, which results in pressure drop.
Li et al. [5] summarized the recent improvement in research of stationary
nanofluid. The researcher mentioned that long-term stability of nanofluid is a key
issue.
2 Nanofluid
Nanofluid contains nanometre-sized particles made of metal, oxides, carbides,

borides, nitrides, etc. There are two ways for the preparation of nanofluids.
2.1 Preparation of Nanofluid
2.1.1 One Step Method
The manufacturing of nanoparticles and dispersion in a base fluid are done

simultaneously. Some general one step methods are as follows:
(a) Direct evaporation
(b) Chemical vapour condensation
(c) Chemical precipitation.
2.1.2 Two Step Method
The nanoparticles are first manufactured and then dispersed into the base fluid. Exam-
ple of two step method is gas condensation. We have preferred the two step method
for the preparation of nanofluid. In this method, first of all, we have calculated the
Experimental Investigation of Unsteady … 545
Fig. 1 a Weighing machine. b Magnetic stirrer. c Ultrasonic cleaner
mass of nanoparticles for 0.5, 1 and 1.5% vol. fraction of nanoparticles into the base
fluid which is de-ionized water.
Mass of Nanoparticles = ρ p ∗ %Vol. Fraction ∗ Volume of base fluid
where ρ p = Density of Nanoparticles (gm/cm3 ).

Surfactant used: Sodium Dodecyl Sulphate (SDS).
Mass of Surfactant = 10% of mass of nanoparticle
We can use different surfactants such as Sodium Dodecyl Sulphate (SDS), Ammo-
nium Cetyl Cetyl (CTAB) and Dodecyl Benzen Sulphonate (SDBS). But SDS is more
effective for stability of oxide nanoparticles in base fluid.
2.2 Process of Nanofluid Reparation
This measured quantity of nanoparticles and surfactant are mixed with base fluid
in magnetic stirrer for 30 min. After magnetic stirring, the fluid is transferred to
ultrasonic bath for 2 h. Ultrasonic waves disperse the nanoparticles into the base
fluid and form uniform mixture. Use of ultrasonic bath more than 2 h does not show
more effect of dispersion of nanoparticles (Fig. 1).
2.3 Stability of Nanofluid
The long-term stability of nanofluid is the key issue. Metal oxide nanofluids are more
stable than metallic nanofluids. Nanofluid becomes stable after the proper dispersion
Fig. 2 SiO2 –DI water nanofluid a without SDS b with SDS after 2 h c with SDS after 48 h d with
SDS after 72 h
of nanoparticles due to existence of Brownian motion. Due to Brownian motion,

nanoparticles are in motion and get dispersed properly in base fluid stability can be
analysed by the following methods (Fig. 2):
(a) Visual Inspection
(b) Zeta Potential (pH Value).
3 Properties of Nanofluids
After the preparation of nanofluid, it is necessary to find out the following properties
of nanofluid. Some important properties are as follows:
1. Density
2. Viscosity
3. Thermal Conductivity
4. Specific Heat.
3.1 Density
Density of nanofluid can be determined from following two methods:
Method 1 Pak and Cho correlation
ρnf = [(1 − Φ)ρnf ] + (Φρbf ) (1)
where ρ nf = Density of nanofluid in gm/cm3

Fig. 3 Variation of density with respect to temperature of different nanofluids
ρ bf = Density of base fluid

F = Volume fraction.
Method 2 Experimental Method

W2 − W1
ρnf = × 103 (2)
50
where W 2 = Weight of jar with 50 cc of nanofluid

W 1 = Weight of empty jar
ρ nf = Density of nanofluid in kg/m3 (Fig. 3).
3.2 Viscosity
Method 1 Einstein correlation
μnf = [1 + 2.5Φ]μbf (3)
where μnf = Density of nanofluid in kg/ms

μbf = Density of base fluid in kg/ms
Φ = Volume fraction.
Method 2 Redwood Viscometer


B
γ = A×t − × 10−6 (4)
t
where γ = Kinematic viscosity of nanofluid m2 /s

t = Time for collecting 50 cc of nanofluid in sec
A and B = Redwood constants
A = 0.264 and B = 190.
μ=ρ×γ (5)
where µ = Density of nanofluid in kg/ms (Fig. 4).
Fig. 4 Variation of dynamic viscosity with respect to temperature of different nanofluids

3.3 Thermal Conductivity
Method 1 Wasp Correlation
2K bf + K np − 2Φ(K bf − K np )
K nf = K bf × (6)
2K bf + K np + Φ(K bf − K np )
where K nf = Thermal conductivity of nanofluid in W/mK

K bf = Thermal conductivity of base fluid in W/mK
F = Volume fraction.
Method 2 Ultrasonic Interferometer
2/3
N
K nf = 3 × × K B × vs (7)
V
where N = Avogadro’s Number = 6.02 ∗ 1023

K B = Boltzmann’s Constant = 1.3807 × 10−23 × 10−23 J/K
vs = Ultrasound Velocity in m/s
V = Molar Volume = mρ (Fig. 5).
3.4 Specific Heat
Pak and Choi have given the following correlation for specific heat:
CPnf = Φ(CP )np + (1 − Φ)(CP )bf (8)
where CPnf = Specific Heat of Nanofluid.
4 Experimental Apparatus
See (Fig. 6).

Specification:
1. Electric Heater: 500 W (1.5 Lit)
2. Digital Temperature Indicator: 0 °C–199.9 °C
3. Thermocouples: Cr–Al type (K-Type)
4. Specimens Material: Copper (Dia—30 mm, Length—30 mm) (K = 386 W/m K).
Thermal Conductivity of CuO-DI Water (50-70nm)
Thermal Conductivity
1.2
1
0.8 0.5% Vol.
W/mK
0.6 FracƟon
0.4
0.2
0
30 50 70
Temperature ( 0C)
Thermal Conductivity of CuO-DI Water (30-50nm)

1.5
Thermal conductivity
1
W/mK
0.5% Vol. FracƟon
0.5
1% Vol. FracƟon
1.5% Vol. FracƟon
0
30 50 70
Temperature ( 0C)
Thermal Conductivity of ZnO-DI Water (50-70nm)

1.2
1
0.5% Vol.
0.8 FracƟon
W/mK
0.6 1% Vol.
FracƟon
0.4 1.5% Vol.
0.2 FracƟon
0
30 40 50 60 70
Temperature ( 0C)
Thermal Conductivity of ZnO-DI Water (30-50nm)

1.5
1
W/mK
0.5% Vol. FracƟon
0.5 1% Vol. FracƟon

1.5% Vol. FracƟon
0
30 40 50 60 70
Temperature ( 0C)
Fig. 5 Variation of thermal conductivity with respect to temperature for various volume fraction
of nanofluid
Thermal Conductivity of SiO2-DI Water (50-70nm)
1.2
1
W/mK 0.8
0.5% Vol. FracƟon
0.6
1% Vol. FracƟon
0.4
0.2
1.5% Vol. FracƟon
0
30 50 70
Temperature ( 0C)
Thermal Conductivity of SiO2-DI Water (30-50nm)

1.2
1
0.5% Vol. FracƟon
0.8
W/mK
0.6 1% Vol. FracƟon

0.4
0.2 1.5% Vol. FracƟon
0
30 40 50 60 70
Temperature ( 0C)
Fig. 5 (continued)
Fig. 6 Unsteady state heat

transfer apparatus
After performing the experiments for different nanofluids, we can get the heat
transfer coefficient, and also, we can compare these nanofluids by drawing unsteady
state heating curves. Following formulas are required to calculate the heat transfer
coefficient:
1. Grashof Number
g β T l 3
Gr = (8)
v2
2. Prandtl Number
μCp
Pr = (9)
K
3. Rayleigh Number
Ra = Gr × Pr (10)
4. Nusselt Number
0.518Ra1/4 −6
Nu = 0.36 + 0.559 9/16 4/9 10 < Ra < 10
9
(11)
1 + Pr
4/9
0.387 Ra1/6
Nu = 0.6 + 0.559 9/16 8/24 10 < Ra < 10
9 12
(12)
1 + Pr
5. Heat Transfer Coefficient
Nu K
h= (13)
d
See (Figs. 7 and 8).
1. Heat transfer coefficient of the nanofluid increases with increase in temperature,

increase in volume fraction and decrease in particle size.
2. Thermal conductivity increases with increase in volume fraction, increase in
temperature and decrease in particle size.
3. Thermal conductivity of CuO–DI water nanofluid (with particle size 30–50 nm,
1.5% Vol. fraction) is 59.59% more than the de-ionized water.
Fig. 7 Variation heat transfer coefficient of with respect to temperature and volume fraction for
different nanofluids
Fig. 8 Variation of temperature with respect to time for different nanofluids during unsteady state
heating
4. Viscosity of nanofluid increases with increase in volume fraction and decrease

in temperature.
5. Density of nanofluid increases with decrease in particle size, decrease in volume
fraction and decrease in temperature.
6. Specific heat increases with increase in temperature, increase in particle size and
decrease in volume fraction.
6 Conclusion and Future Scope
In this paper, we studied the thermophysical properties of nanofluids. We can con-

clude that CuO–DI water nanofluid with 30–50 nm particle size and 1.5% volume
fraction has higher heat transfer coefficient. Due to this, CuO–DI water nanofluid
reaches the steady state faster than other fluids.
At higher volume fraction, nanofluid has better thermal properties, but stability of
nanofluid is a key issue. Further work is required related to the stability of nanofluid.
New types of nanofluids having higher thermal properties should be found.
References
1. Bashirnezhad K, Bazri S, Safaei MR (2016) Viscosity of nanofluids: a review of recent

experimental studies. Int Commun Heat Mass Transfer 73
2. Liu MS, Lin MCC, Tsai CY, Wang CC (2006) Enhancement of thermal conductivity with cu for
nanofluids using chemical reduction method. Int J Heat Mass Transfer 49
3. Gupta M, Singh V, Kumar R, Said Z (2017) A review on thermophysical properties of nanofluids
and heat transfer applications. Renew Sustain Energy Rev 74
4. Albadr J, Tayal S, Alasadi M (2013) Heat transfer through heat exchanger using Al2 O3 nanofluid
at different concentrations. Case Stud Therm Eng 1
5. Li Y, Zhou J, Tung S, Schneider E, Xi S (2009) A review on development of nanofluid preparation
and characterization. Powder Technol 196
Robotics, Biomedical, and Control
Engineering
Switched Staircase-Type Multilevel
Inverter Structure with Reduced
Number of Switches
V. Thiyagarajan
Abstract Modern research has been involved in developing distinctive multilevel

inverter (MLI) circuits. The main aim of this paper is to propose a new MLI topology
based on the series–parallel connection of several dc sources with reduced number of
switching components. The basic structure of the proposed topology generates only
positive voltage levels, and thus, it requires an additional H-bridge unit to create
both negative and zero voltage levels. The generalized structure for the proposed
MLI topology has also been presented in this paper. Additionally, different strategies
are suggested to determine the magnitudes of the dc sources in order to produce more
output steps. Compared with the conventional and other recently presented improved
topologies, the proposed inverter reduces the number of switches, dc sources, and
the maximum voltage blocked by the switches. At last, simulation results for 7-level
symmetrical and 17-level asymmetrical inverters are presented in order to confirm
the performance of the proposed inverter.
Keywords Multilevel inverter · Total harmonic distortion · Symmetric ·

Asymmetric · Reduced switch
1 Introduction
The use of sustainable power sources in modern power industry applications requires
new power converters to create an improved power quality voltage waveform [1, 2].
Multilevel inverter (MLI) consists of an array of power semiconductor switches
and several input sources which provide a multi-step output ac voltage waveform
with high power quality [3, 4]. MLI reduces the total harmonic distortion (THD)
and electromagnetic interference problems. The various conventional MLI includes
cascaded H-bridge (CHB), diode clamped, and flying capacitor inverters [5].
V. Thiyagarajan (B)
SSN College of Engineering, 603110 Chennai, Tamilnadu, India

https://doi.org/10.1007/978-981-15-4488-0_46
558 V. Thiyagarajan
In [6], a new MLI structure using reverse connected voltage sources and H-bridge
inverter is presented. The presented topology shares the voltage stress across the
switches with increased higher output levels. Another MLI topology with reduced
number of switches as compared with CHB is presented in [7]. To show the feasibility,
13-level and 20-level inverter operation is presented in this paper. The new basic unit
for the MLI is presented in [8]. This basic unit consists of three dc sources and five
switches and is able to generate only positive output levels. In order to create negative
output levels, an additional H-bridge inverter is added to the presented basic unit.
The topology presented in [9] can be operated in symmetric and asymmetric modes.
The presented inverter configuration consists of multiple series connected units to
generate larger output levels. Other symmetrical and asymmetric type inverters are
presented in [10–12]. In these, the presented topologies consist of basic unit and level
changing unit. The positive output levels are achieved with the help of basic unit,
and the level changing unit helps to create the negative output levels. However, these
topologies require larger number of switches and increase the total standing voltage
(TSV) across the switches.
The main objective of this paper is to propose a switched staircase-type MLI
topology with reduced switch count and maximize the output levels. The proposed
inverter has been operated in both symmetric and asymmetric modes and shows very
good performance which considerably reduces the cost and installation area.
2 Proposed Inverter Structure
Figure 1 shows the simple circuit of the proposed MLI topology with two main parts
such as level creating part and polarity changing part. The level creating part contains
three dc sources and six main switches, and the H-bridge unit with four switches
forms the polarity changing part. This circuit produces 7-step output voltage during
symmetric operation of the inverter.
Fig. 1 Proposed inverter

Switched Staircase Type Multilevel Inverter Structure … 559
2.1 Operation
The various output levels obtained during positive mode of the proposed basic unit
is shown in Fig. 2, and the corresponding switching conditions are given in Table 1.
It is seen that only two switches in the basic unit and two switches in the H-bridge
units are turned ON to create the required output level. To get the zero output voltage
across the load terminals, the switches (T 1 , T 2 ) or (T 3 , T 4 ) are turned ON. By turning
ON the switches T 1 and T 4 , positive voltage levels are achieved, and to create the
negative output levels, the switches T 2 and T 3 are turned ON. To avoid the short
circuit, any two switches in the combination (S 1 , S 2 , S 3 ) and (S 4 , S 5 , S 6 ) should not
be turned ON simultaneously.
2.2 Generalized Inverter Topology
The generalized structure of the proposed MLI with ‘p’ units is shown in Fig. 3. The
relation between the total number of basic units ‘p’ and total number of dc sources
‘n’ is obtained as:
n = 3p (1)
The total number of switches ‘k’ in the generalized structure of the proposed MLI
is given by
k = 10 p (2)
The total number of conduction switches in the generalized structure of the

proposed MLI ‘nON-State ’ is obtained as:
n ON - State = 4n/3 (3)
For the extended structure shown in Fig. 3, the total standing voltage (TSV) is
given by:

p

TSV = 5V1 j + 7V2 j + 6V3 j (4)
j=1
The maximum output voltage is synthesized by adding the voltage magnitude of

all dc sources across the load and is given by:

p
Vo,max = (V1i + V2i + V3i ) (5)
i=1
560 V. Thiyagarajan
(a) V1 (b) V22
(c) V3 (d) V 1+V2
(e) V2+V3 (f) V1+ V3
(g) V1-V2+V3 (h) V1+V2+V3
Fig. 2 Different output levels

Table 1 Switching
S. No Conducting switches Output voltage
conditions for different output
levels 1 T 1 , T 2 or T 3 , T 4 0
2 S3 , S6 V1
3 S1 , S6 V2
4 S1 , S4 V3
5 S2 , S6 V1 + V2
6 S1 , S5 V2 + V3
7 S2 , S4 V1 + V3
8 S3 , S4 V1 − V2 + V3
9 S2 , S5 V1 + V2 + V3
3 Determination of Voltage Magnitude
In this section, two different methods are proposed to determine the voltage
magnitude of dc sources.
3.1 Algorithm-1
In algorithm-1, the dc voltage magnitude of each source is obtained as
Vi j = Vdc where, i = 1, 2, 3 and j = 1, 2, 3, . . . , p (6)
Since the voltage magnitudes are same, the inverter is called as symmetric inverter.
In this case, the maximum output voltage is given by
Vo,max = nVdc (7)
The total number of output levels ‘m’ is given by
m = 6p + 1 (8)
3.2 Algorithm-2
In algorithm-2, the magnitude of each source is determined as

562 V. Thiyagarajan
Fig. 3 Generalized structure of the proposed MLI

V1i = (17)i−1 Vdc

V2i = 2(17)i−1 Vdc
V3i = 5(17)i−1 Vdc where, i = 1, 2, 3, . . . , p (9)
Since the voltage magnitudes are different, the inverter is called as asymmetric
inverter. The total number of output levels ‘m’ is given by
m = (17) p (10)
4 Comparison Study
This section presents the comparison results of the proposed MLI with recently
presented topologies. This comparison is based on the required number of sources,
switches, and TSV value to create the required output levels. Figure 4a shows the
plot between the number of levels and sources. Figure 4b, c shows the plot between
the number of levels and switches during symmetric and asymmetric operation,
respectively. Figure 4d shows the total number of conducting switches for the given
number of dc sources. The plot between the TSV and the number of output levels
is shown in Fig. 4e. The results show that the proposed topology uses minimum
number of dc sources and switches to create higher output levels. Also, it minimizes
the total number of conducting switches and thereby minimizes switching losses.
In addition, the TSV value of the switches is reduced as compared with the other
presented topologies.
5 Simulation Results
This section presents the simulation results of symmetric 7-level and asymmetric
17-level operation of the proposed inverter. A series RL load with the values of R =
50 and L = 100 mH is considered for the simulation.
5.1 Symmetric 7 Level
During symmetric condition, all the sources have same voltage magnitude, i.e., V 1
= V 2 = V 3 = 80 V. Figure 5 shows the simulation results of 7-level inverter. It
is obtained that THD of the 7-level output voltage waveform is 12.23%. The input
power is 500 W, and the efficiency is obtained as 87% with a loss of 65 W.
564 V. Thiyagarajan
(a) Levels vs Sources (Asymmetric) (b) Level vs Switches (Symmetric)
(c) Levels vs Switches (Asymmetric) (d) Sources vs ON-State Switches
(e) Levels vs TSV
Fig. 4 Comparison results
5.2 Asymmetric 17 Level
During asymmetric condition, the voltage magnitudes are different. In this case, V 1 =
V dc = 30 V, V 2 = 2V dc = 60 V, V 3 = 5V dc = 150 V. Figure 6 shows the simulation
results of 17-level inverter. It is obtained that THD of the 17-level output voltage
waveform is 4.84%. The efficiency of the inverter is obtained as 92% with a loss of
45 W.
(a) Output Voltage
(b) THD
Fig. 5 Simulation results—7-level
6 Conclusion
This paper presents a new switched staircase-type MLI topology with reduced
switched count. The comparison study is presented in this paper between the pro-
posed MLI and other recently developed topologies. The comparison results shows
that the proposed MLI has superior features over conventional and other recently
developed topologies in view of number of switches, sources, and TSV. The pro-
posed inverter has very good performance which considerably reduces the cost and
installation area. Finally, the proposed MLI is implemented by MATLAB simulation
for 7-level and 17-level inverter operation in order to show the performance of the
inverter.
566 V. Thiyagarajan
(a) Output Voltage
(b) THD
Fig. 6 Simulation results—17-level
References
1. Khenar M, Taghvaie A, Adabi J, Rezanejad M (2018) Multilevel inverter with combined T-type
and cross connected modules. IET Power Electronics 11(8):1407–1415
2. Prayag A, Bodkhe S (2017) Novel basic block of multilevel inverter using reduced number of
on-state switches and cascaded circuit topology. Adv Electr Eng 2017:1–16
3. Odeh CI, Obe ES, Ojo O (2016) Topology for cascaded multilevel inverter. IET Power Electron
9(5):921–929
4. Prabaharan N, Palanisamy K (2017) Analysis of cascaded H-bridge multilevel inverter
configuration with double level circuit. IET Power Electron 10(9):1023–1033
5. Karthikeyan D, Vijayakumar K, Sathik MAJ (2018) Development of a switched diode
asymmetric multilevel inverter topology. J Power Electron 18(2):418–431
6. Thamizharasan S, Baskaran J, Ramkumar S, Jeevananthan S (2014) Cross-switched multilevel
inverter using auxiliary reverse connected voltage sources. IET Power Electron 7(6):1519–1526
7. Ajami A, Mokhberdoran A, Oskuee MRJ (2013) A new topology of multilevel voltage source
inverter to minimize the number of circuit devices and maximize the number of output voltage
levels. J Elect Eng Technol 8(6):1328–1336
8. Babaei E, Laali S, Bayat Z (2015) A single-phase cascaded multilevel inverter based on a new
basic unit with reduced number of power switches. IEEE Trans Indus Electro 62(2):922–929
9. Oskuee MRJ, Nikpoura S, Ravadanegha SN, Gharehpetian GB (2018) Improved configuration
for symmetric and asymmetric multilevel inverters with reduced number of circuit devices. Int
J Ambient Energy 39(4):421–431
10. Thiyagarajan V, Somasundaram P (2018) Multilevel inverter topology with modified pulse
width modulation and reduced switch count. Acta Polytech Hung 15(2):141–167
11. Khosroahahi MI (2014) Crisscross cascade multilevel inverter with reduction in number of
components. IET Power Electron 7(12):2914–2924
12. Babaei E (2008) A cascade multilevel converter topology with reduced number of switches.
IEEE Trans Power Elect 23(6):2657–2664
Hydro Pneumatic Parking Brake
Actuation System for Motor Grader
Application
K. Rajasekar, S. Karthikeyan, V. Kumar, and H. S. Satish Chandra
Abstract Motor graders are used in many applications in the Mining and Construc-
tion fields. In India, grader models running on road need to meet the Central Motor
Vehicles Rules (CMVR 1989) [1], i.e. RTO regulations formulated by the Ministry
of Road Transport and Highways (MoRTH) and in the mines has to meet the DGMS
regulations. Hydraulic or air-operated brake system is installed in graders for slow
down and stop. Air brake system is preferred due to its simple construction, low cost
and easy serviceability. Parking brake is used to hold the equipment on slopes in
stationary conditions and for parking. As per the vehicle safety standards, parking
brakes are designed to be fail safe. To meet this, spring applied, air/hydraulic release
type parking brakes are used. Motor grader having 12 ft work blade length uses the
air system for regular service brakes and internally expandable shoe type parking
brake that is actuated by mechanical ratchet pinion mechanism is taken up for study.
New disc type spring applied hydraulically released parking brake is designed to
improve the holding performance. Need arise to invent a new actuation system as
the main service brakes use air-operated brakes and hydraulic power required for
parking brake application. The new hydro pneumatic parking brake actuation sys-
tem has been designed and tested. The practical field tests reveal that new parking
brake system reduces operator fatigue, builds up more confidence in operator and
ensures safety of higher order.
Keywords Motor grader · Parking brake actuation · Hydro pneumatic
1 Introduction
Motor Grader is used in construction and mining sites for various applications.
Graders are also named as the work equipment based on nature of work carried
out. Motor Graders are categorized based on the work blade size, frame type and
K. Rajasekar (B) · S. Karthikeyan · V. Kumar · H. S. Satish Chandra

Equipment Division, Research & Development (R&D), BEML Limited,
Mysore, Karnataka, India

https://doi.org/10.1007/978-981-15-4488-0_47
570 K. Rajasekar et al.
operating mass. Motor graders are used for material spreading, moving, rolling, mix-
ing and grading operations in road making construction fields. Small and midsized
work blade graders like 10 and 12 ft are used in these applications. Larger size work
blade graders like 16 ft and above are used mainly in various surface opencast mines
for making the haul roads on which heavy trucks and dumpers are transporting coal
and other minerals extracted from the earth.
Graders are classified as Rigid and Articulated type based on the Mainframe
structure. In rigid version, graders have a single frame where the cabin, work blade
and other system aggregates are mounted whereas in articulated version, frame is
made into two parts, i.e. front and rear frames, coupled with hinge pins to achieve
the shorter turning radius by articulation and best suited for work sites with con-
stricted workspace. Graders work attachment operations are hydraulically operated.
Control levers or joysticks are used to control the work blade by the operator inside
the cabin. Though the graders are slow-moving work equipment, the safety of the
operator is ensured by essential safety standards. Cabin of the graders is designed
to meet the ROPS standards to protect the operator. The operator fatigue is reduced
by providing the power steering, power brake and other controls. Hydro shift trans-
mission ensures precise and smooth gear shift and transmission of power. Box type
heavy-duty tandem delivers the power to the rear wheels. Foot-operated, air actuated
internally expandable shoe type service brakes and parking brake ensure the stop-
ping and smooth control of the grader. It is equipped with safety measures like Auto
Fire Detection and Suppression system, ROPS cabin, etc. It also meets the DGMS
standards to run in various mines of India.
Parking brakes are primarily used to park the vehicle in Slopes. In some vehicles,
the parking brake may be used in emergency conditions. In commercial vehicles
like car and medium-duty vehicle, ratchet and pinion type parking brake actuation
system has been used. Single lever actuates the internally expandable shoe brake
at rear wheels of the vehicle through mechanical linkages. The factors influenc-
ing the performance of the mechanical type parking brake have to be evaluated to
eliminate rollaway [2]. Thermal expansion of the brake material and initial tem-
perature is affecting the clamping force of parking brakes. FMVSS 135 standard
describes the performance and functional requirements of vehicle parking brakes [3,
4]. Ergonomics of the operator has been an important factor that has to be studied to
make an effective parking brake system [5]. As the mechanical lever-actuated park-
ing brakes require more operator effort, this leads to operator fatigue. Electrical and
electronic assisted system developments in the parking brake area offer less operator
effort. Application of Sensors and Motors converts conventional mechanical system
into automatic hand-operated brake system [6, 7]. It’s significant to derive corre-
lation among gross vehicle weight, driving force, braking force, rolling and grade
resistance in hill station application [8]. In order to avoid damages due to operator
negligence (forgot) to apply the parking brake, automatic apply and release of park-
ing brake by integrating sensors, Engine control unit and other electromechanical
components have already developed and installed in on-road commercial vehicles.
Off-highway and heavy mining equipment parking brakes have been designed for
fail safe to ensure safety.
Hydro Pneumatic Parking Brake Actuation … 571
2 Problem Definition
Motor grader with a 12 ft blade is considered for this research work. Air is used
for the main brake application. Mechanical ratchet pinion type parking brake was
equipped to provide parking brake of the 16 Ton class grader. Existing parking brake
system also meets performance and the standard operating force requirement as rec-
ommended by international standards. Introduction of new power shift transmission
that is fitted with spring applied sliding caliper disc type parking brake for improved
holding performance, demands for a hydraulic release of parking brake. The main
service brakes are air-operated and hence hydraulic power is required for parking
brake application resulting in a need of a new actuation system. A new concept of
hydro pneumatic parking brake actuation system has been evolved and developed in
line with ISO 3450 standard norms.
3 Design of Hydro Pneumatic Parking Brake Actuating

System
Presently disc type spring applied hydraulically released parking brake in place of
internally expandable shoe type parking brake is being designed for improved parking
brake performance (Fig. 1).
To improve the holding performance and reduce operator fatigue on motor graders
equipped with air-operated brake system, hydro pneumatic parking brake system has
been introduced. The spring applied hydraulically released sliding caliber disc type
is used for parking brake application. When the vehicle is in OFF condition, due
to the spring force the parking brake will be applied automatically. When engine
Fig. 1 Hydro pneumatic parking brake system for motor graders

starts, the compressor (1) builds the air pressure inside the system. The compressed
air enters the parking brake solenoid valve (6) through air dryer (2), Main reservoir
(3) and system protection valve (4). Booster (7) receives pressurized air supply from
parking brake solenoid valve and boosts the hydraulic pressure required to release
the parking brake at sliding caliber disc brake (9). When the parking brake switch (5)
is operated inside the cabin, the parking brake solenoid valve is energized and cuts
the air supply to the booster. Due to the lack of air pressure, spring force comes into
action which leads to the application of parking brake. Further, the oil will return to
the brake oil tank (8).
4 Development and Installation
Based on the gross vehicle weight and parking brake requirements as per international
vehicle safety, calculations were made to select and evaluate disc type sliding caliber
parking brake. Static and dynamic coefficients of friction assumed as 0.34 and 0.28
respectively. As the calculation concludes the pressure required to release the parking
brake is 98 bar and Max. Permissible release pressure is 206 bar which is not possible
to achieve by air-operated system alone. The booster is selected to meet the pressure
limits and sliding velocity of the caliber to ensure smooth application and release of
parking brake. Magnification factor achieved in the selected booster is 24.
Performance characteristics of the booster are critically investigated to confirm
the adaptability for the present requirement. Brake oil tank is sized based on the oil
flow and stroke length of the booster cylinder. Low-pressure switch is employed in
the parking brake line to sense and send the signal to the transmission control unit.
ON/OFF switch is provided inside the cabin to control the parking brake by operator
(Fig. 2).
5 Performance Evaluation and Testing
Parking brake performance and evaluation have been done as per ISO 3450-standard
norms [9]. Before carrying out the parking brake performance test, the overall brake
system has been ensured for zero leak condition by soap water test. Also, the equip-
ment is charged to its maximum working pressure and switch-off for 30 min. Then
the pressure values are checked and ensured that the pressure drop is not beyond
0.5 bar. The stopping distance test has been measured to conform the normal work-
ing of the air brake system. As per the operating mass of the grader, 32 km/hr speed
has been chosen for stopping distance tests. After application of the service brake,
the stopping distance and time have been measured and confirmed that the results
meet ISO 3450 standard requirement.
Fig. 2 Parking brake system aggregates mounting arrangement
5.1 Holding Performance Test
The parking brake has been tested with a 20% slope gradient to reflect the worst
condition in mines and hills station. The holding performance of the parking brake
is validated in 20% Gradient and also checked with 36% Gradient. Parking brake
apply and release actuation time has been examined and compared with the standard
values. Test results ensure the parking brake meets the standard requirements (Fig. 3).
5.2 Parking Brake Functional Run Test
Run test has been carried out to check the release characteristic of parking brake.
Operational run test for continuous of 2 hours at actual working environment has been
carried out. During trials, release and apply of parking brake have been examined by
switching the parking brake ON/OFF using the switch provided inside the cabin to
ensure the function. Further, it is also ensured that the parking brake is not engaged
accidentally during operation since the air pressure in the system is above the low-
pressure warning limit. Functional run test confirms the parking brake requirements
as per ISO 3450 standard (Fig. 4).
Fig. 3 Parking brake holding test
Considering equipment mass, parking brake arrangement and grade holding require-
ments as per ISO 3450 standard, minimum hydraulic pressure required to release
parking brake is around 98 bar. With the air to hydraulic booster magnification factor
of 24 times, nearly half of the working pressure of air system is enough to release the
parking brake, i.e. about 4.2 bar. Electro Pneumatic solenoid offers a quick release
of parking brake as soon as the air pressure reaches the half of the normal working
pressure (i.e. 8 bar).
Input air line to release the parking brake has been directly taken from the system
protection valve port no. 23. The dynamic opening pressure of the system protection
valve’s port no. 23 is about 6.9 bar. Uninterrupted release of parking brake is ensured
with the minimum of 6.9 bar pressure available in parking brake line continuously
against the minimum required pressure of 4.2 bar to release the parking brake with
said booster magnification. SAE 10 W oil is used in the booster hydraulic line to
deliver the brake release pressure at parking brake end. Smooth apply and release
of parking brake is achieved by practical test with Electro pneumatic solenoid and
ON/OFF switch setup.
The holding performance test reveals the present parking brake system offers
holding in even 36% slope against the requirement of 20% slope as per the ISO 3450
standard. Continuous 2 hours run test confirms there were no malfunctions in hydro
pneumatic parking brake actuation system.
7 Conclusion
Most of the present-day motor graders are equipped with a hydraulically operated
accumulator-assisted parking brake system. Hydraulically operated brake system
contributes approximately 8–10% of excess heat load compared to air-operated brake
Fig. 4 Functional run test
system. The present motor grader equipment considered for these trials produces
nearly 15 kw of heat load in excess when operated with hydraulic operated brake
system. Hence an additional Cooling system needs to be introduced in hydraulic
brake system due to its higher heat generation. The complexity of the system is high
resulting in need of higher skills for maintenance and service.
Pneumatic or air-operated brake system offers ease of maintenance and low cost
over hydraulic brake system due to its simplicity. In motor graders operated with air
brake system, to achieve the pressure required to release the parking brake (about
98 bar) is not possible. To overcome this, air actuated & hydraulically released type
parking brake system has been designed for 12 ft, 16 Ton class motor graders.
Major aggregates of the system were critically evaluated and selected to suit
the present requirement. Air pressure from system protection valve is magnified by
24 times in the booster hydraulic line. With the combination of system protection
valve and brake booster, hydraulic pressure of 98 bar is delivered continuously when
the equipment is On and is under operation, to release the parking brake. Operator
fatigue has been eliminated by using soft-touch ON/OFF switch inside the cabin
which controls the apply and release of the parking brake through parking brake
solenoid valve based on the operator input.
Prototype has been developed with full-scale model and various tests have been
conducted in actual working conditions as per ISO 3450 standard. More than 75%
increase in holding performance makes the motor grader best suits in Hill Stop
parking brake holding. The 2 hours practical field run test revealed that new Hydro
Pneumatic parking brake system reduces the operator fatigue, builds more confidence
in operators and ensures safety at the higher order.
References
1. The Central Motor Vehicles Rules, India (1989)

2. Mckinlay AJ (2007) The phenomenon of vehicle park brake rollaway. Ph.D. Thesis-School of
mechanical Engineering, University of Leeds, August
3. Mohd-Razmi I, Jamaludin-Mohd T, Abd-Rahim AB, Wan-Zaidi WO, Ali B (2018) Brake torque
analysis of fully mechanical parking brake system: theoretical and experimental approach.
IngenIería InvestIgacIón y tecnología 19(1): 37–49
4. Rozainia AH, Ishak MR, Abu Bakar AR, Mohd Zain MZ (2013) Performance of a fully mechan-
ical parking brake system for passenger cars. In: 2nd international conference on mechanical
engineering research, pp 1–10. IOP Publishing, Malaysia
5. Noble V, Frampton R, Richardson J (2014) Ergonomics of parking brake application: factors
to fail. In: Proceedings of 2014 international conference on ergonomics and human factors:
contemporary ergonomics and human factors, pp 481–488. Loughborough University, Great
Britain
6. Thivagar S, Kumar CN (2016) Automatic hand brake system. Int J Eng Res Gen Sci 4(1):214–
221
7. Maske A, Tuljapure SB, Satav PK (2016) Design & analysis of parking brake system of car.
Int J Innov Res Sci, Eng Technol 5(7):12578–12590
8. Wang H, Li N, Gao F, Dai X (2014) Research on the hill-start control of vehicles based on the
pneumatic auto parking brake system. International conference on mechatronics. Controls and
electronic engineering. Atlantis Press, China, pp 42–49
9. ISO 3450—earth-moving machinery—braking systems of rubber-tyred machines—systems
and performance requirements and test procedures
10. Scales W, McCulloch DM (2007) Best practices for compressed air systems, compressed air
challenge
Design and Analysis of Feature Primitive
Scaffold Manufactured Using
3D-Printer—Fused Deposition Modelling
(FDM)
P. Vishnurajan, S. Karuppudaiyan, and D. KingslyJeba Singh
Abstract The scaffold guided tissue is a branch of tissue regenerative medicine

approach. The scaffold acts as a temporary structure for bone growth and provides a
better environment and architecture for the regeneration of bone. In this work, five
different features were defined in parametric form using CATIA. The relationship
between geometric parameters and porosity, surface area-to-volume ratio was devel-
oped with constant inter-pore distance. The developed relationship of scaffolds was
evaluated by designing of 40% porosity. All these scaffolds were manufactured with
biocompatible Poly Lactic Acid (PLA) material using Ultimaker3 –FDM system.
The mechanical properties of a scaffold were evaluated by the compression test, and
finite element analysis was performed using ANSYS software. The experimental and
finite element analysis (FEA) results were compared to understand the microscale
level modelling and variability of mechanical properties of scaffold manufacturing
with 3D printing technology. In this study, we have obtained experimental effec-
tive elastic modulus (or) Young’s modulus is 858.69 MPa and compressive strength
(27.693 MPa) of a scaffold (P1), maximum porosity was (38.06%) for scaffold (T1).
In FEA, the maximum effective Young’s Modulus is 1086 MPa for scaffold (P1),
maximum stress is 29.54 MPa for scaffold (C1) and maximum strain is 0.331 for
scaffold (T1).
Keywords Scaffold · Additive manufacturing · 3D printer · Ultimaker3 –FDM
P. Vishnurajan · S. Karuppudaiyan (B)

Kattankulathur, Chennai, India
P. Vishnurajan
D. K. Singh
School of Mechanical Engineering, SRM Institute of Science and Technology, Kattankulathur,
Chennai, India

https://doi.org/10.1007/978-981-15-4488-0_48
578 P. Vishnurajan et al.
1 Introduction
Bone is a natural heterogeneous structure; it has, in macroscale level, two-layer

architecture as cortical and cancellous. When the bone experiences damage due to
accident, disease and aging, it is a challenging task for orthopedic surgery to repair
and restore its functions [1, 2]. In general, bone possesses a good healing capacity of
defects under suitable physiological, environmental conditions. If the defects are a
larger size, particularly in the load-bearing bone-like femur, vertebrae and tibia have
a problem clinically to resolve.
Tissue engineering (TE) is a recent growing approach to repair of injured or
fracture bone [3]. The success of bone tissue engineering depends on a scaffold,
which acts as a template for cells to interactions and formation of bone to provide
structural support to the newly formed tissue [4]. A desirable scaffold should provide
preliminary mechanical support at the site of the defect until the new bone tissue
is regenerated [5–7]. The scaffolds should have enough strength to resist load at
the site of implantation also have high porosity and integrity to completion of the
regenerative process. The minimum parameters of any scaffold are porosity, pore-
interconnectivity and surface area-to-volume ratio and biocompatible material and
manufacturing technology.
1.1 Additive Manufacturing (or) Rapid Prototyping Method
The additive manufacturing process is commonly known as 3D printing. It is mainly

used to manufacturing low-volume parts with complex shapes also often to multi-
ple functions. Additive manufacturing process successfully used for the fabrication
of patient-specific implants and porous scaffolds with custom-tailored internal and
external architectures [8–11]. The rapid prototyping techniques build apart directly
from the CAD model using layer by layer approach.
The development of additive manufacturing techniques are widely used for tissue
engineering applications. They are used mainly to make 3D porous scaffolds with
custom-tailored of internal and external architecture. The following rapid prototyp-
ing systems are commercially available to fabricate scaffolds for tissue engineering
application; they are SLS (Selective Laser Sintering), FDM (Fused deposition mod-
elling), STL (Stereo lithography) Precision extrusion deposition, (SLM) Selective
Laser Melting, Electron Beam Melting (EBM) and bio-plotter [12, 13]. The limi-
tations this technology is accuracy, resolution, material used to control appropriate
internal and external architecture of scaffold structure. However, recent technologies
in rapid prototyping combined together with composite (HA-PCL/HA-TCP) bioma-
terials [14–17]. The rapid prototyping process yet to recognize as a standard to design
and optimization of the architecture of scaffold till now, not even for manufacture
specific target of tissue [18].
Design and Analysis of Feature Primitive Scaffold … 579
The objective of this work, to design the feature-based internal architecture of

scaffold and manufacture using the 3D printer with PLA biocompatible material,
mechanical characterization of PLA scaffold by both experimental and numerical
methods.
2 Methods and Methodology
2.1 Design of Feature Primitive Scaffold
In general, scaffold design is difficult to reproduce a microstructure of exact internal

bone in the scaffold. We created simplified models with functionally equivalent to
mechanical properties that help tissue to repair apart from the biological requirement.
The goal of the design is to produce a porous structure which effectively reduces the
strength and stiffness of scaffold and provide sufficient space for regeneration of new
bone tissue at the implanted site.
In this work, scaffold primitives are designed using CATIA software with the
following features (Circle, Triangle, Square, Hexagonal, and Pentagonal) size varies
from 0.4 to 1 mm with a constant inter-pore distance of 1.25 mm and dimension of 10
× 10 × 20 mm. The edge size/diameter is controlled parameters with desired porosity.
The designer can interact with CATIA design table and select the appropriate feature
and design scaffold. A typical scaffold library was designed with 40% porosity is
shown in Fig. 1.
2.2 Design and Optimization of Geometric Parameters

of Feature Primitive Scaffold
To understand the relationship between design geometric parameters with surface

area-to-volume ratio (K) and porosity (P). In a circular feature, scaffold primitives
are modelled with a diameter of 0.4–1 mm in steps of 0.1. The obtained parameters
are listed in Table 1 (Fig. 2).
The design porosity for each model was calculated using Eq. (1)

Vs
Porosity(P) = 1 − × 100 (1)
Vapp
where V s —design volume of the porous scaffold, and

V app —apparent volume of scaffold model.
The surface area-to-volume ratio (K = SA/Vs) is an essential parameter for tissue
generation, cells attachment, and proliferation. The scaffold area is a primary site
for the interaction of cells in the surrounded region of the implanted site. The cells
Fig. 1 Representative
feature primitive scaffold of
CAD model designed with
40% porosity
Table 1 Design parameters of the scaffold-circular feature

Pore diameter, Scaffold volume, Porosity, (P) (%) Surface area, K = (SA/Vs)
(d) (mm) (Vs) (mm3 ) (SA) (mm2 )
0.4 1679.5 16.0 3807.5 2.3
0.5 1525.2 23.7 4177.2 2.7
0.6 1353.7 32.3 4414.0 3.3
0.7 1171.2 41.4 4518.0 3.9
0.8 984.0 50.8 4489.0 4.6
0.9 798.3 60.1 4327.3 5.4
1 620.4 69.0 4032.6 6.5
used in tissue engineering are mostly anchorage-dependent, so the scaffold surfaces

should have enough surface roughness to facilitate cell proliferation and attachment.
The scaffold designing depends on the requirement of geometric of the damaged site.
The surface area-to-volume ratio and porosity in the designed scaffolds are linearly
variable with the pore size (Fig. 3).
Fig. 2 Relationship between porosity and pore size
Fig. 3 Relationship between surface area-to-volume ratio and pore size

Table 2 Design parameter of five pore shape scaffold for porosity (P) with 40%
Scaffold type Circular pore Triangular Square pore Pentagonal Hexagonal
shape (C1) pore shape shape (S1) pore shape pore shape
(T1) (P1) (H1)
Design d t t t t
parameter
Design 0.678 0.93 0.608 0.465 0.38
parameter
value (mm)
3 Additive Manufacturing of Porous Scaffold Using the 3D

Printer
In the past three decades, 3D printer is potentially used for developing complex struc-
tures without tools in the area of engineering, biomedical, aerospace and automotive
applications. In this work, scaffold was manufactured using 3D printerUltimaker3 .
Ultimaker3 is a user-friendly additive manufacturing machine based on the principle
of fused filament fabrication, which is similar to the Fused Deposition Modelling
process. This printer can print a 3D part with ABS (Acrylonitrile butadiene styrene),
Nylon and PLA (Polylactic acid) materials. Ultimaker3 produces the fastest build
time for fabricating a model. Similar to all other rapid prototype systems, the scaffold
model was converted into STL file using CATIA software. Then, the pre-processing
was done with CURA software. STL file is used in the software to convert the model
into the number of the slice for manufacturing; then the final file is sent to Ultimaker3
to fabricate layer by layer process.
3.1 Modelling and Manufacturing of Scaffold
The linear relationship were evaluated to design scaffold with 40% of porosity. The
model designed to have a lesser wall thickness, which cannot be printed in ultimaker3 ,
so the model was scaled into 2.5%, so the dimension changed to 25 × 25 × 50 mm.
For manufacturing of scaffold, the layer height of 0.06 mm and no infill option was
used to avoid the removal of support material. The calculated design parameters of
the features are shown in Table 2, and manufactured scaffolds are shown in Fig. 4
(Table 3).
3.2 Porosity Calculation for Manufactured Scaffold
The porosity of all manufactured scaffold is calculated using Eq. (2) and listed in
Table 4.
Fig. 4 Manufactured scaffold-3D printer
Table 3 Supplier’s information for PLA material

Polymer Modulus of elasticity Poisson’s ratio Density
Polylactic acid 2745.68 MPa 0.36 1.24 g/cm3
Table 4 Experimental porosity calculated from the manufactured scaffold model

Scaffold type Volume of Mass of Scaffold Designed Porosity
the scaffold scaffold (g) porosity (%) porosity (%) difference
(mm3 ) (%)
C1 32,164.72 25.32 36.51 40.5 9.83
T1 32,170.94 24.72 38.01 40.2 5.44
S1 32,176.06 25.12 37.03 40.4 8.31
P1 32,147.95 24.97 37.35 40.3 7.29
H1 32,317.37 25.43 36.54 40.6 9.99

ρc
Porosity(%) = 1 − × 100 (2)
ρm
where ρc —Density of manufactured scaffold,

ρm —Density of bulk material of PLA (as per supplier data in Table 3).
The density of the manufactured scaffold is predicted by measuring mass and
volume. The obtained results were listed in Table 4.
3.3 Mechanical Testing
The compression test was performed using TUE CN 200 machine for all the samples.
The ultimate strength, load and displacement were plotted as shown in Fig. 5. The
test was carried out on all the samples with 5 mm displacement approximately 10%
Fig. 5 The load–displacement curve for C1 Scaffold
of its length. The compressive strength, compression at peak, and effective Young’s
modulus of the scaffolds calculated. The stress and strain were calculated and plotted
as shown in Fig. 6 (Table 5).
3.4 Finite Element Analysis of Scaffold
Finite element analysis was carried out to evaluate strength, stress and strain of
scaffold with 40% porosity. This approach can decrease physical iteration numbers,
which is used to optimize scaffold geometry with the best structural integrity and
porosity. In this work, static structural analysis was performed using ANSYS. The
mesh geometry was created in ANSYS. The tetrahedron mesh-type was used to
create the mesh. As a representative scaffold (C1) type, nodes and elements are 10,
75,130 and 6, 14,889, respectively. A similar approach is followed for all other types
of scaffolds. The boundary conditions are fixed support with zero degrees of freedom
was applied on the lower end surface, on the top surface was allowed to displacement
value obtained by experimentally as listed in Table 6 to use on the opposite top surface
as shown in Fig. 7a. The material properties assigned to the model which is obtained
Fig. 6 Stress-strain curve for C1 scaffold
Table 5 Compression test results of five different types of scaffold and its effective Young’s
modulus
Scaffold type Load at peak Compression at Compression Effective
(KN) peak (mm) strength Young’s modulus
(N/mm2 ) (MPa)
C1 15.710 3.450 24.460 757.65
T1 14.260 4.910 22.226 639.88
S1 14.820 2.320 23.077 827.89
P1 17.790 4.320 27.693 858.69
H1 15.890 4.580 24.632 810.38
from the supplier as given in Table 5 and analysis was carried out and found the
reaction force at the fixed face and stress–strain was obtained, and results are in
given Table 6.
Table 6 Finite element analysis of five types of the scaffold with displacement value taken from
compression test on the manufactured scaffold
Scaffold Displacement Von Directional Strain Reaction Effective
type (mm) misses stress force Young’s
stress (z-axis) (KN) modulus
(MPa) (MPa) (MPa)
C1 3.450 693.53 29.54 0.25327 36.542 847.74
T1 4.910 856.8 28.12 0.33056 47.581 775.3
S1 2.320 332.46 23.077 0.12109 28.563 984.9
P1 4.320 822.26 27.693 0.32515 58.645 1086.0
H1 4.580 677.25 24.632 0.24665 54.152 945.9
Fig. 7 a Pre-processing model. b Stress distribution. c Strain distribution
The results obtained experimentally and finite element analysis was listed in Tables 5
and 6. In this study, we compared results of maximum averages stress obtained by
the compression test, which less than that of average stress obtained by finite element
analysis. The scaffold type C1 and T1 has a difference in effective Young’s modu-
lus of 10.59% and 17.4%, and compressive is 17.19% and 20.96%, the porosity is
9.83% and 5.445%, respectively, which is lower than the design porosity. The reason
for lower experimental results may be manufacturing artefacts such as interlayer’s
failure, material flow characteristics in the nozzle, and reduced printing resolutions
at microscale level. It confirms that the theoretical porosity of scaffold lower than the
designed porosity due to the filling of material inside the pores. The cross-sectional
area of scaffold is higher than the designed area due to the inaccuracy of printer
resolutions.
5 Conclusion
In this work, the scaffold library was created in CATIA consist of five different
basic features. It needs to be developed with more combined features to obtain a
heterogeneous structure. The developed library investigated porosity (P), Surface
area-to-volume ratio (K) of the scaffold with constant inter-pore distance for tissue
engineering applications. We conclude that to confirm the finding of in this study
we need to do further experiments with large samples manufactured with system
controlled parameters and finite element analysis by varying other parameters like a
change of materials, printing nozzle with different combined features.
References
1. Duda GN, Heller M, Albinger J, Schulz O, Schneider E, Claes L (1998) Influence of muscle
forces on femoral strain distribution. J Biomech 31(9):841–846
2. Edwards WB, Gillette JC, Thomas JM, Derrick TR (2008) Internal femoral forces and moments
during running: implications for stress fracture development. Clin Biomech 23(10):1269–1278
3. Langer R, Vacanti J (1993) Tissue engineering. Science 260:920–926
4. Yeong WY, Chua CK, Leong KF, Chandrasekaran M (2004) Rapid prototyping in tissue
engineering challenges and potential. TRENDS Biotechnol 22(12)
5. Hutmacher DW, Sittinger M, Risbud MV (2004) Scaffold-based tissue engineering: rationale
for computer-aided design and solid free-form fabrication systems. TRENDS Biotechnol 22(7)
6. Sun W, Starly B, Nam J, Darling A (2005) Bio-CAD modeling and its applications in computer-
aided tissue engineering. Comput-Aided Des 37:1097–1114
7. Habib FN, Nikzad M, Masood SH, Saifullah ABM (2016) Design and development of Scaffolds
for tissue engineering using three-dimensional printing for bio-based applications. 3D Print
Addit Manuf 3(2)
8. Lal P, Sun W (2004) Computer modeling approach for microsphere-packed bone scaffold.
Comput Aided Des 36:487–497
9. Zhang XY, Fang G, Zhou J (2017) Additively manufactured scaffolds for bone tissue
engineering and the prediction of their mechanical behavior: a review. Materials 10(50)
10. Starly B, Lau W, Bradbury T, Sun W (2005) Internal architecture design and freeform
fabrication of tissue replacement structures. Comput-Aided Des 38:115–124
11. Chua CK, Leong KF, Cheah CM, Chua SW (2003) Development of a tissue engineering scaffold
structure library for rapid prototyping. Part 2: parametric library and assembly program. Int J
Adv Manuf Technol (21): 302–312
12. Zein I, Hutmacher DW, Tan KC, Teoh SH (2002) Fused deposition modeling of novel scaffold
architectures for tissue engineering applications. Biomaterials 23:1169–1185
13. JyotiKalita Samar et al (2003) Development of controlled porosity polymer-ceramic composite
scaffolds via fused deposition modeling. Mater Sci Eng, C 23:611–620
14. Lin CY, Kikuchi N, Hollister SJ (2004) A novel method for biomaterial scaffold internal
architecture design to match bone elastic properties with desired porosity. J Biomech 37:623–
636
15. Eshraghi S, Das S (2010) Mechanical and microstructural properties of polycaprolactone scaf-
folds with one-dimensional, two-dimensional, and three-dimensional orthogonally oriented
porous architectures produced by selective laser sintering. J Acta Biomater 6:2467–2476
16. Cahill S, Lohfeld S, McHugh PE (2009) Finite element predictions compared to experimental
results for the effective modulus of bone tissue engineering scaffolds fabricated by selective
laser sintering. J Mater Sci Mater Med 20:1255–1262
17. Wettergreen MA, Bucklen BS, Starly B, Yuksel E, Sun W, Liebschner MAK (2005) Creation of
a unit block library of architectures for use in assembled scaffold engineering. Comput-Aided
Des 37:1141–1149
18. Giannitelli SM, Accoto D, Trombetta M, Rainer A (2014) Review on current trends in the
design of scaffolds for computer-aided tissue engineering. J Acta Biomater 10:580–594
Design and Development of Modular
Parallel XY Manipulator
Santosh R. Thorat and R. B. Patil
Abstract Flexure mechanism provides precise and repeated motion over a small
range. Many monolithic designs have been discussed in the literature however
they are costly to manufacture and give no flexibility for change in the parame-
ter. This paper represents flexure micromotion stages with modular design. Compli-
ance matrix method has been used for designing the flexure mechanism. Nonlinear
2-DOF model is used to characterize the stiffness of XY stage, maximum stress-
induced. Proposed XY motion stage has a travel range of ±3.2 mm2 with 0.12 mm
parasitic error. Dynamic analysis is performed to determine modal frequency of the
stage. Maximum error estimated in analytical and FEA model is 26.38%. Linear and
nonlinear analytical results are compared with FEA and are in agreement.
Keywords Compliant mechanism · Cross-coupling · Stiffness matrix · Nonlinear

analytical modeling
1 Introduction
Compliant mechanism transmits motion by deformation of flexure elements. Mostly

used flexure elements are blade flexure and wire flexure or slender rod [1]. Flex-
ure elements are arranged in series or parallel. In series arrangement two or more
1-DOF are connected to obtain desired motion and motion can be controlled at the
expense of high inertia, low natural frequency, and cumulative errors [2, 3]. Parallel
mechanism has advantages of high payload capacity, lower inertia, and high natu-
ral frequency [4]. However, the parallel mechanism has disadvantages such as low
workspace and parasitic error. Flexure parallel stage is used in many applications in
precise machines and instruments. Compact XY Flexure parallel stage with a large
motion range is desirable in many applications, atomic force microscopy, MEMS,
S. R. Thorat (B) · R. B. Patil

Army Institute of Technology, 411015 Pune, India
R. B. Patil
https://doi.org/10.1007/978-981-15-4488-0_49
590 S. R. Thorat and R. B. Patil
biomedical implant and space application [3, 5, 6]. They are used in various appli-
cations because of no backlash, free of friction, no noise emission and no need for
lubrication [4]. Designing the flexure mechanism has some challenges because of its
nonlinear behavior, lack of motion range and cross-axis coupling [7]. Flexure parallel
stages are mostly driven by piezoelectric or electromagnetic actuators. Stages driven
by piezoelectric actuator has small motion range, high stiffness and compactness and
those driven by electromagnetic actuator have large workspace, low stiffness, and
ease to vibrate [2]. Compliance matrix method as reported in [1–4, 8–13] has been
used for designing the micromotion stage.
In the literature, many parallel flexure mechanisms have been proposed and ana-
lyzed. Herpe et al. [14] presented a model used to characterize micromotion stage.
Two-DOF nonlinear model is used to characterize force-displacement and stress
analysis. Dynamic analysis is carried to find natural frequency of stage. The pro-
posed mechanism has motion range of ±2.3 with parasitic error of 60 micron. Awtar
[10] in his Ph.D. thesis presented analytical formulation including geometric non-
linearity for family of symmetric XY Flexure mechanism with a large workspace
and small parasitic error. Pham et al. [9] address the stiffness models based on the
way flexure elements are connected together. Awtar et al. [11] presented a nonlinear
force-displacement model for 2-D beam flexure. Su et al. [15] presented a screw the-
ory approach for synthesis and analysis of compliant joint. Li et al. [3] presented the
idea of totally decoupling and analyzed double parallelogram flexure using matrix
method for modeling compliance and stiffness. Wan et al. [12] presented a survey of
recently developed flexure mechanisms with large motion range and greater accu-
racy. Su et al. [1] presented symbolic formulation for compliance and synthesis of
mechanism with serial, parallel and hybrid topologies based on screw theory. Jia
et al. [16] presented a parameterized compliance approach for synthesis and analysis
of flexure. Xu et al. [2] proposed new multistage compound parallelogram flexure.
The motion range of the mechanism is greater than 10 mm.
In this paper, modular design for a flexure parallel stage is presented. Many mono-
lithic micromotion stages have been discussed in the literature. Monolithic design
does not require assembly of component however when flexures are deformation
beyond yield strength they get fractured. This, in turn, led to remanufacturing of
whole mechanism due to its monolithic nature. Clearly, there is the motivation for
design of modular micromotion stage. Modular design helps in manufacturing and
assembly all the components separately. Modular design has an advantage that param-
eters of flexure element can be varied to some extent and fabrication can be done
at relatively low cost. The flow of this paper is as follows. Section 2 determines
stiffness of motion stage. Sections 3 and 4 consider nonlinear force-displacement
and maximum stress analysis. Section 5 determines the resonant frequency of the
stage. The coupling analysis of stage is carried out in Sect. 6. Section 7 compares
the analytical solution with FEA and finally conclusions are drawn.
Design and Development of Modular Parallel XY Manipulator 591
2 Stiffness of Mechanism

Let us consider a general twist X− = δx , δ y , δz , θx , θ y , θz and general wrench

F− = Fx , Fy , Fz , Mx , M y , Mz [8] with 6 DOF (degree of freedom) at reference
point OXYZ. The parameters of a beam are shown in Fig. 1. The dimensions and
configuration of the XY motion stage are shown in Fig. 2 and CAD model is shown
in Fig. 3. Mechanism consist of four rigid blocks located at corners are fixed. There
are four intermediate blocks (P, Q, R and S) where input force is applied and a motion
stage (C) at the center.
All flexure elements are numbered 1–16 have the same dimension and are given
in meter.
According to linear elastic theory, the relation between twist and wrench can be
written as in
Fig. 1 Beam parameters [2]
Fig. 2 Spring equivalent

model [14]
Fig. 3 CAD model
F− = Sfix · X− (1)
where Sfix is 6 × 6 stiffness matrix of fix-guided beam which is given as

⎛ ⎞ ⎡ EA ⎤⎛ ⎞
Fx L
0 0 0 0 0 δx
⎜F ⎟ ⎢ 0 12EIz 6EIz ⎥⎜ ⎟
⎜ y⎟ ⎢ L 2 ⎥⎜ δ y ⎟
L3
0 0 0
⎜ ⎟ ⎢ 12EI 6E I ⎥⎜ ⎟
⎜ Fz ⎟ ⎢ 0 0 y
0 − L2 y 0 ⎥⎜ δz ⎟
⎜ ⎟=⎢ L3 ⎥⎜ ⎟ (2)
⎜ Mx ⎟ ⎢ 0 0 0 GJ
0 0 ⎥⎜ θx ⎟
⎜ ⎟ ⎢ L ⎥⎜ ⎟
⎝ My ⎠ ⎣ 0 6EI
0 − L 2y 0
4EI y
0 ⎦⎝ θ y ⎠
L
Mz 0 6EIz
L2
0 0 0 4EIz
L
θz
where E and G is Young’s modulus and shear modulus, respectively, A = t × w is

3
beam’s cross-section area, L is length of beam, J is torsion constant and Iz = w×t 12
3
and I y = t×w
12
are the area moments [1]. This matrix is at local frame which needs to
be shifted to global frame by using the shifting law from screw theory and presented
in [5, 6, 13, 14, 17, 18]. This is implemented by pre multiplying by inverse of (Tr)
transpose and post multiplying by inverse of (Tr) to Sfix as follows.
j j −T j −1
Si = (Tr)i · Sfix · (Tr)i (3)
where ‘i’ is local reference of beam and ‘j’ is global reference at the center of motion
j
stage. T is the transpose of matrix and (Tr)i is 6 × 6 adjoint transformation matrix
given as

j j j
j Ri S ti Ri
(Tr)i = j
(4)
0 Ri
Owing to symmetry stiffness of mechanism can be evaluated from the stiffness of

one-quarter part of parallelogram, i.e. parallelogram 1–2–3–4. The stiffness of beam
1 at center C is given by S1c It can be obtained by translating along X by 0.03 and
Y direction by (L + 0.03) and zero in Z direction. Therefore, t1C = |0.03, −(L +
0.03), 0|.
−T −1
S1c = (Tr)c1 · Sfix · (Tr)c1 (5)
The stiffness of flexure 2 at center C is obtained by revolving flexure 1 about

Y-axis by π radians with no translation as given below.
−T −1
S2C = Ry(π )C2 · S1C · Ry(π )C2 (6)
Similarly, stiffness of flexure 3 at point C can be obtained by revolving the flexure

1 by −π /2 rad about Z-axis and translating 0.02 in X direction 0.03 in Y direction
and zero in Z direction. Hence t3C = |0.02, −0.03, 0|.
−T −1
S3C = (Tr)C3 · Sfix · (Tr)C3 (7)
Stiffness of flexure 4 is obtained by simply revolving flexure 3 by π radians about

Y-axis.
−T −1
S4C = Ry(π )C3 · K 3C · Ry(π )C3 (8)
The stiffness of the parallelogram 1–2–3–4 at center C is given as
1
SC1 = −1 −1 (9)
S1C + S2C + S3C + S4C
The stiffness of mechanism can be obtained by rotating the parallelogram P by

−π /2 around the Z-axis as follows.
S = SC1 + SC2 + SC3 + SC4 (10)
where SC1 , SC2 , SC3 , SC4 are stiffness of three-quarter of mechanism.
3 Nonlinear Modeling
Owing to large deformation, the length of beam flexure is not constant thus nonlinear
analysis is considered. This nonlinearity is due to tension loading terms which cause
the change in length of beam flexure. Each beam flexure is considered as spring
connected to rigid body as shown in Fig. 2. Further it is assumed that parallelogram
P and R can travel in only Y direction and parallelogram Q and S can travel only
in X direction. Let us consider outer parallelogram i.e. beam 1–2, 5–6, 9–10, and
13–14 and inner parallelogram i.e. beam 3–4, 7–8, 11–12, and 15–16. From Fig. 2,
the deformation at the intermediate stage is written as
⎧
⎪
⎪ δPy = δ1y = δ2y
⎨
δRy = δ9y = δ10y
(11)
⎪ δQx = δ5x = δ6x
⎪
⎩
δSx = δ13x = δ14x
The reaction forces at the intermediate stage are written as follows.

⎧
⎪
⎪ FPy = F1y + F2y
⎨
FRy = F9y + F10y
(12)
⎪
⎪ F = F5x + F6x
⎩ Qx
FSx = F13x + F14x
where FPy represent reaction force at intermediate stage p in y direction, F1y is

reaction force in beam 1 and so on. As in [14] the total stiffness of parallelogram
1–2, 5–6, 9–10 and 13–14 can be derived by considering a single beam as shown in
Fig. 3.
The downward bending force as given in on beam 1 is given by
Fig. 4 Deformed condition

of beam 1 [1]
Fig. 5 Spring model of

perpendicular elements
12EIz δ1y
F1ybend = (13)
L 31
where δ1y is a deflection of beam 1. The component of the tensile force in Y direction
is given by
F1ytens = EA ∈1 sin ∝1 (14)
where ∈1 = (L tens − L)/L is linear strain and ∝1 = tan−1 (δ1 /L) is the angle made
by beam 1 with horizontal. Therefore, total force acting on beam 1 along the Y-axis
is given by
F1y = F1ybend + F1ytens (15)
Now taking into consideration inner structure the stiffness can be obtained by
considering one outer beam connected with single inner one as shown in Fig. 5. For
small deflection of beam 16 there is a very small deflection of beam 14, therefore,
stiffness of beam 14 is negligible as compared to stiffness of beam 16. The force
acting along the X-axis on beam 16 can be obtained as
12EIz δ14x EA · L 14 sin α14

F16x T = 3
+ (16)
L 14 L 14
The force acting along the Y-direction is, therefore, a combined effect of the forces
applied by beams 14 and 16 as given.
F16x T δ16y 12EIz δ16y

F16y = + (17)
L 16x L 316
Therefore, the reaction forces at the center C motion stage due to δ Cx and δ Cy are
FCx = FQx + FSx + F3x + F4x + F11x + F12x (18)

FCy = FPy + FRy + F7y + F8y + F15y + F16y (19)
The deflection at the intermediate stage is related to the deflection of the center of
the motion stage. The following equation can be deducted by applying Pythagoras
theorem to deformed and undeformed conditions between intermediate and motion
stages.
⎧
⎪
⎪
⎪
⎪ δPy = δC y − L − L − δCx2 2
⎪
⎪
⎪
⎪
⎨
δRy = δC y + L − L − δCx2 2
(20)
⎪
⎪
⎪
⎪ δ = δ − L − L 2 − δ2
⎪
⎪ Qx C x
Cy

⎪
⎪
⎩ δSx = δC x + L − L 2 − δ 2
Cy
4 Stress Analysis
Stresses in the XY mechanism are determined to know allowable displacement the

stage can undergo within the elastic limit of the material. The maximum stress occurs
at one of the corner of flexure beam where it is attached to rigid body. Maximum
Bending stress in beam is given by σ max = M b Y /I. Where Y is the farthest point from
neutral axis (half of beam thickness) and M b = F 1y_tens × 0.5 × L. From Eq. (13),
maximum bending stress can be given by.
3Ewδ1y
σ1bend = (21)
L 21
Also, stress induced due to tensile loading in beam 1 is given by
σ1tens = E ∈1 (22)
From Eqs. (23) and (24) we can write maximum stress is
σ1 = K 1 σ1bend + K 2 σ1tens (23)
where K 1 is a stress concentration factor for bending loading and K 2 is the stress
concentration factor for tensile loading.
5 Dynamic Analysis
Dynamic analysis helps in finding the resonance frequency of mechanism. The

equation of motion for un-damped free vibration is given.
M ẍ + S ẋ = 0 (24)
where M is mass matrix and K is a stiffness matrix of mechanism. The mass matrix
is given by
⎡ ⎤
Mxx 0 0 0 0 0
⎢ 0 0 ⎥
⎢ Myy 0 0 0 ⎥
⎢ ⎥
⎢ 0 0 Mzz 0 0 0 ⎥
⎢ ⎥ (25)
⎢ 0 0 0 Ixx 0 0 ⎥
⎢ ⎥
⎣ 0 0 0 0 Iyy 0 ⎦
0 0 0 0 0 Izz
where Mxx , Myy , and Mzz are moving mass in X, Y, and Z direction respectively
Ixx , Iyy and Izz are the moment of inertia about X, Y and Z direction, respectively.

33
Mxx = Myy = m 0 + 2m p + 8 × m beam + (4 × m beam ) (26)
140
where mass of motion stage m 0 = 0.062 × W × ρ, mass of parallelogram m p =

0.06 × 0.02 × ρ, mass of beam m beam = W × t × L × ρ

33
Mzz = m 0 + 8 × m beam (27)
140

m 0 0.062 + w2 0.022 + w2 0.062 + w2
Ixx = Iyy = + 2m p + (L + 0.03)2 +
12 12 12
(28)
2 2
0.06 + 0.02
2
0.03 + 0.03
2
Izz = 4m p + m0 (29)
12 12
The resonance frequency of stage can be determined by

1 S
f = (30)
2π M
6 Coupling Analysis
The parasitic error in flexure mechanism is due deformation of intermediate stage P

and R when input displacement is applied to intermediate stage Q and S. This error is
also called cross-coupling. The parasitic error in the X direction can be determined
by subtracting the resulting displacement δ Sx of parallelogram S from desired output
displacement δ Cx as follows [14].
δx_par = δCx − δSx (31)
Similarly, the parasitic error in Y direction can be determined by subtracting

resulting displacement δRy of parallelogram R from desired output displacement δCy
as follows.
δ y_par = δCy − δRy (32)
7 FEA Validation
The analytical model is validated using ABAQUS 6.14 software. The parameters of
the beam and material properties are given in Table 1.
7.1 Force-Displacement Analysis
The force-displacement relation is studied by applying gradual displacement of

3.25 mm at one of the intermediate stage and reaction forces are noted. Nonlin-
ear behavior is due to the load stiffening phenomenon at lager displacement. The
force required for 3.25 mm displacement is 166.7 N analytically and 190.5 N from
FEA. Thus the nonlinear analytical model is validated. Comparing the result of ana-
lytical model for linear (Eq. 10) and nonlinear (Eq. 21) behavior and FEA shows
some linearity for small deflection and deviates for large displacement. These results
are shown in Fig. 6.
Table 1 Parameters of beam

Parameters t (m) w (m) L (m) E (N/m2 ) G (N/m2 ) ρ (kg/m3 )
Values 0.0008 0.015 0.075 68.9e9 26.9e9 2810
Fig. 6 Force-displacement 250

NONLINEAR FEA
plot
LINEAR ANALYSIS
200
ANALYTICAL
Reaction force (N)

150
100
50
0
0 0.5 1 1.5 2.25 3.25
Displacement (mm)
Fig. 7 Stress displacement 500 NONLINEAR FEA

plot 450 LINEAR ANALYSIS
400 ANALYTICAL
350
Stress (MPa)
300
250
200
150
100
50
0
0 0.5 1 1.5 2.25 3.25 4.25 5
Displacement (mm)
7.2 Stress Analysis
In order to study the stress variation in mechanism, the stress concentration factor K1
is taken as 1 and K2 is taken as 2 as in [14]. The maximum displacement evaluated
from the analytical model is 4.3% smaller than FEA analysis. The error in the FEA
and analytical model is 26.38%. Figure 7 show that yield strength of material has
reached 5 mm displacement.
7.3 Coupling Analysis
Coupling analysis is carried to determine maximum positioning error in model. In

FEA displacement is applied in Y direction at center of motion stage in steps and
Fig. 8 Parasitic error
Output displacement in X-direction

0.12 ANALYTICAL
FEA
0.1
0.08
0.06
0.04
0.02
0
0 0.5 1 1.5 2 2.5 3
Input displacement in Y-direction
output displacement is recorded in at intermediate stage S in X direction. The parasitic

displacement plot is shown in Fig. 8. Maximum parasitic error is 0.12 mm from FEA
and error in analytical and FEA is 2%.
7.4 Modal Analysis
Frequency analysis is done in ABAQUS 6.14 using Lanczos Eigen solver. The first
four mode shapes are shown in Fig. 9. The first two resonant frequencies in X and Y
directions are the same and occur at 27.9 Hz form FEA and 26.7 Hz analytically. The
Fig. 9 Mode shapes of

mechanism
third resonant frequency occurs at 125.15 Hz from FEA and 134.4 Hz analytically.
Fourth frequency from FEA is 261.8 and 285.2 Hz analytically. The error in analytical
and FEA model is 9.24%.
8 Conclusion
An analytical model incorporating linear and nonlinear behavior was presented.

MATLAB was used to characterize the compliant XY stage. Analytical model was
successfully applied to predict stiffness, motion range considering the limitation
such as maximum stress. The model proposed can predict the output displacement
as input displacement is applied. Its micromotion stage is used for position control
accurately. The results from FEA and analytical model are within 26.38%. The
micromotion stage has a travel range of ±3.2 mm with cross-coupling of 0.12 mm.
The yielding occurs at the displacement of 5 mm. The ratio between the first two
frequencies and third is greater than 0.22. Modular design can be successfully applied
to micromotion stage. Modular design will greatly reduce the manufacturing cost for
the same characterization of a monolithic structure.
References
1. Su H, Shi H, Yu J (2012) A symbolic formulation for analytical compliance analysis and

synthesis of flexure mechanisms. J Mech Des 134:1–9
2. Xu Q (2012) New flexure parallel-kinematic micropositioning system with large workspace.
IEEE Trans Robot 28:478–491
3. Li Y, Xu Q (2009) Design and analysis of a totally decoupled flexure-based XY parallel
micromanipulator. IEEE Trans Robot 25:645–647
4. Xiao S, Li Y, Zhao X (2011) Design and analysis of a novel flexure-based XY micro-positioning
stage driven by electromagnetic actuators. In: Proceedings of 2011 international conference on
fluid power and mechatronics, pp 953–958
5. Shang J, Tian Y, Li Z, Wang F, Cai K (2015) A novel voice coil motor-driven compliant
micropositioning stage based on flexure mechanism. Rev Sci Instrum 86:1–10
6. Li Y, Xu Q (2010) Development and assessment of a novel decoupled XY parallel microposi-
tioning platform. IEEE Trans Mechatronics 15:125–135
7. Howell LL, Magleby SP, Olsen BM (2013) Handbook of compliant Mechanism. Compliant
mechanism. 1st edn. Wiley Publication, New Delhi
8. Yu J, Li S, Su HJ, Culpepper ML (2011) Screw theory based methodology for the deterministic
type synthesis of flexure mechanisms. J Mech Robot 3: 1–14
9. Pham HH, Chen IM (2005) Stiffness modeling of flexure parallel mechanism. Precis Eng
29:467–478
10. Awtar S (2004) Synthesis and analysis of parallel kinematic XY flexure mechanisms. Ph.D.
Thesis, Massachusetts Institute of Technology
11. Awtar S, Sen S (2010) A generalized constraint model for two-dimensional beam flexures:
nonlinear load-displacement formulation. J Mech Des 132:1–11
12. Wan S, Xu Q (2014) A survey on recent development of large-stroke compliant microposition-
ing stage. Int J Robot Autom 1:19–35
13. Zhang Y, Su H, Liao Q (2014) Mobility criteria of compliant mechanisms based on

decomposition compliance matrices. Mech Mach Theory 79:80–93
14. Herpe X, Walker R, Dunnigan M, Kong X (2018) On a simplified nonlinear analytical model
for the characterization and design optimization of a compliant XY micromotion stage. Robot
Computer-Integrated Manuf 49:66–76
15. Su HJ, Dorozhkin DV, Vance JM (2009) A screw theory approach for the conceptual design of
flexible joints for compliant mechanisms. J Mech Robot 1:1–8
16. Jia M, Jia RP, Yu JJ (2015) A compliance-based parameterization approach for type synthesis
of flexure mechanisms. J Mech Robot 7:1–12
17. Lobontiu N (2014) Compliance-based matrix method for modeling the quasi-static response
of planar serial flexure-hinge mechanisms. Precis Engg 38:639–650
18. Tang H, Li Y (2013) Design, analysis, and test of a novel 2-DOF nanopositioning system driven
by dual mode. IEEE Trans Robot 29:650–662
19. Yu J, Li Z, Lu D, Zong G, Hao G (2014) Design and analysis of new large-range XY compliant
parallel micromanipulators. In: Proceedings of the ASME 2014 International Design Engi-
neering Technical Conferences & Computers and Information in Engineering Conference, pp
1–10
Design of hyper-redundant In-Vivo
Robot: A Review
Sushil Dev Rout, Mihir Kumar Sutar, and Sarojrani Pattnaik
Abstract Conventional clinical process involves a large incision in the stomach to

perform the surgery or biopsy to collect the infected tissue from the abdominal cavity.
However, the process requires expertise and often leads to a surgeon’s tremor and
patient’s discomfort. The inclusion of a robot to assist in biopsy and surgery is a
boon for both surgeon and patient as it reduces the surgeon’s error in pinpointing the
infected area and also results in a minimum incision requirement to do so, which is
called as Minimally Invasive Surgery (MIS). In some cases, the entire robot is placed
inside to do the task and called an in vivo robot. The current study focuses on the
various design strategies adopted by researchers around the globe and the outcomes.
This state of the art will be useful to enhance the design strategies further to refine
the most cutting-edge research outcomes further.
Keywords Minimally invasive surgery · In vivo robot · Biopsy · Design of

surgical robot
1 Introduction
Conventional biopsy and surgical procedure bring tremendous pain and consume
lots of time. Moreover, expensive medical bill adds trauma to the patient. Inclusion
of robot to assist in surgery and biopsy was remarkable progress in the medical
field. However, robotic-assisted surgery has some challenges with its advantages.
The constraints present in designing and controlling the robot includes the trocar
complicacy in case of minimally invasive surgery through a small opening port in
the stomach, dimensions available in the natural opening such as the esophagus,
constricted degrees of freedom (dof) for the complete insertion of robot in case of
in vivo robot-assisted surgery or biopsy. Robot-assisted surgery for stomach biopsy
and surgery is categorized into two types one is laparoscopy, which is a minimally
S. D. Rout · M. K. Sutar (B) · S. Pattnaik

Mechanical Engineering Department, Veer Surendra Sai University of Technology, Burla, Odisha,
India

https://doi.org/10.1007/978-981-15-4488-0_50
604 S. D. Rout et al.
invasive surgery (MIS), and in this case, the robot is kept outside during the clinical
procedure. The second one is one in which the entire robot is placed inside the
stomach through natural openings such as the esophagus, and this type of robot is
called in vivo robot. The issues of concerns in designing and controlling such type of
robots are lost dof at the insertion position of trocar incision, surgeon’s expertise in
performing the clinical task, complicated forward and inverse kinematic trajectories
which lead to an unsuccessful attempt to reach at the specified goal point by the
robotic tip, absence of a proper and better control strategy to control the tip from
damaging any healthy tissue during the clinical procedure.
This study briefs about various design principles adopted by researchers to impart
an optimum solution to the described constraints. The state of the art consists of
different strategies that have been adopted by researchers to impart an efficient design
for smooth clinical procedures and constraints involved in them.
2 Methodology
Design of a surgical robot decides the effectiveness of the instrument for performing
the clinical tasks such as collecting biopsy samples from the infected area, performing
surgical tasks with ease because of its simplicity in design and with available dofs.
Dexterity is one of the most important parameters that set the design implacability.
Proper dexterity allows the robotic tip to attain any point in the abdominal cavity
with easiness. Minor and Mukherjee [1] proposed a dexterous manipulator design
for the MIS. The end effecter was provided with bi-directional 180° articulation and
rotation was provided by a compact multilink structure comprised of gears and gear-
link as shown in Fig. 1. A 2-dof robotic manipulator shown in Fig. 2 was proposed
by Yamashita and Kim [2] for performing endoscopy and tissue sampling tasks.
The robotic forceps were designed and fabricated for holding, suturing the affected
tissues in the liver abdomen. The two-way bending mechanism suggested in the
◦
design attained a workspace ±80 for each bending mechanism.
Fig. 1 Robot structure [1]

Design of hyper-redundant In-Vivo Robot: A Review … 605
Fig. 2 A schematic of 2-dof

robotic manipulator design
[2]
Peirs et al. [3] proposed a self-propelling endoscope for inspection of the human
colon. The robotic manipulator has 2 dof as it can tilt and bend both sides by 40° .
The body of the robot consists of two modules separated by a pneumatic expansion
system and actuated by an electromagnetic motor. The front module consists of the
inspection tool, illumination, camera and channel for air and water injection. The rear
module is of larger size and is hermetically sealed with the front one. The advantage
of this type of design is that it allows to propel the whole system with proper control
on the bellows and it can be steered around as and when required. However, the
design has an issue that the bellow wall may get damaged because of the channels
in it and it may endanger the patient’s condition. The design and all the components
in it are illustrated in Fig. 3.
Fig. 3 CAD model of the

self-propelling robotic
endoscope [3]
Lehman et al. [4] developed a robotic platform for Natural Orifice Transluminal
Endoscopic Surgery (NOTES) and tested its laparoscopic capabilities and control.
The retraction and position of the robotic manipulator with disconnected linkage
are shown in Fig. 4a–b, respectively. The design basically consists of a grasper and
cautery arms connected to a central module having the camera unit. The two arms
have individual rotational and translational dof coupled with the left and right yaw
of the central body.
The designed robot was able to apply up to 10 N of force axially and 5 N of force
transversely. However, the design reported unstable image quality and complicacy
during simultaneous clinical procedures.
Lehman et al. [5] performed the design and control of a fixed-based in vivo robot
for abdominal exploration to perform the necessary clinical tasks. The arrangement
of the robotic platform along with its accessories is shown in Fig. 5. The robot had
Fig. 4 The NOTES robot on both articulation a retraction and, b position [4]
Fig. 5 Pan and tilt fixed

focus camera robot [5]
Fig. 6 Pan and tilt

adjustable focus camera
robot [6]
a limitation of focusing on the intricated portions inside the abdominal cavity at

a different depth, leading to poor viewing feedbacks. This led to a variable focus
camera robot by Rentschler et al. [6] shown in Fig. 6.
Further, the designed was enhanced to develop a wheeled mobile robot with an
ability to advance inside the abdominal organs [6]. The wheel design was finalized
after several tests and later improvised through visco-elastic modelling. The design
is shown in Fig. 7.
Xu and Simaan [7] proposed a multi-backbone snake-like robot with increased
dexterity. The system has two sections along with a detachable gripper as shown in
Fig. 8. It has 1° accuracy in path tracking.
Garg et al. [8] presented a design of a 4-dof hyper-redundant robotic manipulator
for stomach biopsy. Hyper-redundancy increases the dexterity for reaching the target
points inside the abdominal cavity. The in vivo robot has 4 dof apart from the available
4 dof of the conventional endoscope in which it is connected. The CAD model of
the design proposed is shown in Fig. 9.
Workspace analysis was also performed for the four-scaled design and it was
observed that a total of 0.00418 m3 of workspace is covered with the proposed design
whereas the total workspace for a four-scaled stomach model is 0.00376 m3 . So, the
proposed design can easily reach all the parts of the stomach for biopsy even up to
Fig. 7 Mobile camera and

biopsy robot with helical
screw design [6]
Fig. 8 Snake-like robot with

detachable gripper
Fig. 9 CAD model of the

4-dof hyper-redundant
manipulator with tool
channel [7]
the duodenum. The torque required for actuating the joints were also measured to be
0.625mN-m. The manipulator is wire actuated. The relation between the length of
wire to be pulled to the angle of rotation of each joint in Cartesian space is calculated
in the analysis and also the biopsy force involved for collecting the tissue has been
calculated by modelling a four-scaled size of the actual robot. The articulated links
are interconnected. The first link is connected to the coupler and the last is connected
to the gripper. The coupler consists of 1500 of conical slots on either side to facilitate
the titanium wire for better actuation. Total 360° grooves are avoided as it may lead
to wire entanglement. The slots are shown in Fig. 10.
Further capsule endoscopy gained momentum over a decade and various work
advancement is noticed in this segment as well. The capsule endoscopy is only
Fig. 10 Rotary joint with

wire guide for actuation [8]
Fig. 11 M2A capsule a the actual model, b cross-sectional details [9]
meant to gather clinical information about the tissue health inside the abdominal
and colon. Glukhovsky [9] developed a capsule endoscope that can steer inside the
gastrointestinal tract to collect the information about the infected tissue inside. The
collected information subsequently is utilized in curing diseases like cancer, polyps
and other medical concerns. The integrated battery inside the capsule gives a back-
up for around 8 h. The advantage of this design is its low-cost effectiveness and
battery back with a pin point detection of data regarding infectious tissues. The
flaw-side of the design is that the instrument is meant for only one-time use. The
design of the capsule endoscope is shown in Fig. 11. The capsule is called a M2A
capsule. Liu and Gao [10] reported the design and fabrication of magnetic energy
propelled capsule endoscope to collect pictorial information of the GI tract. This
information collected is further utilized for the clinical diagnosis of the patient.
The capsule contains four modules namely transmission, illumination, vision and
control. All the modules are assembled and kept inside a shell with optical dome.
A ball screw arrangement is made which converts the rotary motion of the motor
into linear motion. And the position of the capsule inside the patient is controlled
by the interaction of permanent magnet assembly and magnet present inside it. The
arrangement is illustrated in Fig. 12.
Simi and Valdastri [11] proposed hybrid locomotion for wireless-capsule endo-
scope which generates 3.8 N at the tip of the legged mechanism and a magnetic link
force up to 135 mN.
Miyashita et al. [12] presented an ingestible origami robotic capsule that can be
used to perform all such clinical tasks that are required to get medical data about
the infection and also to patch the same after pin point location. The robotic capsule
is reconfigurable to match the space requirement and wrapped around with a bio-
compatible cover. Inside the cover, it consists of a miniature robot and the deliverer
to enable the robot to reach the pin point location (Fig. 13).
3 Conclusions
Still better designs yet to come with an ability to perform flaw-less medical tasks
assigned to them. Cost-effective, more agile, and dexterous miniature robots are need
Fig. 12 Different parts of the magnetic propelled endoscopic system [10]
Fig. 13 Ice capsule with the

deliverer [12]
of the hour for better medical assistance to the surgeon performing clinical diagnosis,
biopsy and surgery. Followings were observed in the brief literature review of the
design and analysis of surgical robot:
• The design of surgical robot mainly the in vivo robot has several constraints that
have to be addressed during the process, e.g. the maximum possible tool channel
diameter of available videoscope is 6 mm under Indian condition.
• Simpler and effective designs are need of the hour for improvised dexterity.
• Some of the proposed designs may have limitations, e.g. the self-propelling
endoscope [3] has a tilting angle limited to 40° . However, these designs are the
foundation to proceed further for new and effective designs.
• The capsule endoscopes are only meant for capturing data inside the abdominal
cavity. They are unable to perform any clinical tasks inside.
Acknowledgements The authors are grateful to the Department of Mechanical Engineering,

VSSUT Burla for rendering all necessary co-operation to carry out our research work.
References
1. Minor M, Mukherjee R (1999) A dexterous manipulator for minimally invasive surgery. In:
Proceedings of the IEEE international conference on robotics and automation, pp 2057–2064
2. Yamashita H, Kim D, Hata N, Dohi T (2003) Multi-slider linkage mechanism for endoscopic
forceps manipulator. In: Proceedings of the IEEE/RSJ international conference on intelligent
robots and systems, pp 2577–2582
3. Peirs J, Reynaerts D, Van Brussel HA (2001) Miniature manipulator for integration in a self-
propelling endoscope. Sens Actuators 92:343–349
4. Lehman AC, Wood Nathan A, Dumpert J, Oleynikov D, Farritor Shane M (2008) Dexter-
ous miniature in vivo robot for notes. In: Proceedings of the 2nd biennial IEEE/RAS-EMBS
international conference on biomedical robotics and biomechatronics, USA, pp 244–249
5. Lehman AC, Rentschler ME, Farritor SM, Oleynikov D (2007) The current state of miniature
in vivo laparoscopic robotics. J Robot Surg 1:45–49
6. Rentschler ME, Platt SR, Dumpert J, Farritor SM, Oleynikov D (2006) In vivo laparoscopic
robotics. Int J Surg 4(3):167–171
7. Xu K, Simaan N (2006) Actuation compensation for flexible surgical snake-like robots with
redundant remote actuation. In: Proceedings of the IEEE international conference on robotics
and automation, Florida, pp 4148–4154
8. Garg A, Vikram CS, Gupta S, Sutar MK, Pathak PM, Sharma AK, Mehta NK, Gupta VK
(2014) Design and development of in-vivo robot for Biopsy. J Mech Based Des Struct Mach
42:278–295
9. Glukhovsky A (2003) Wireless capsule endoscopy. Sens Rev 23:128–133
10. Gao M, Liu S (2010) Design and fabrication of a magnetic propulsion system for self-propelled
capsule endoscope. IEEE Trans Biomed Eng 57:2891–2902
11. Simi M, Valdastri P (2010) Design, fabrication, and testing of a capsule with hybrid locomotion
for gastrointestinal tract exploration. IEEE/ASME Trans Mechatron 15:170–180
12. Miyashita S, Guitron S, Yoshida K et al (2016) Ingestible, controllable, and degradable
origami robot for patching stomach wounds. In: IEEE international conference on robotics
and automation (ICRA), 16–21 May 2016, Stockholm, Sweden, pp 909–916
Effect of Insertion Force for Successful
Penetration of a Conical Shaped
Microneedle into the Skin
Gera Aswani Kumar , Burra Rajesh Kumar, Vankara Jyothi,

and Sowmya Injeti
Abstract Microneedles and microneedle-related technologies are the present trends

in medical science which are voraciously being investigated by the research com-
munity towards various therapeutic and medical diagnostic applications, since they
are treated and also to be considered as painless methodology of administering vac-
cines. Apart from that these microneedles clearly aids in mitigating the potential risk
of infections and ailments related when needle transfusions when the medicine is
transmitted into the body. Microneedles have gained much upfront and considered
to be a novel way for administering drugs and vaccines through transdermal drug
delivery. This particular area primarily mends upon the design and analysis of silicon
and stainless steel based hollow conical microneedles for transdermal drug delivery
(TDD). The effect of axial and transverse load on a microneedle has been investigated
along with the mechanical properties. The analysis predicts that the resultant stresses
due to applied bending and axial loads are in the safe & comfort desired range. The
primary work was focused on the conical-shaped hollow needles of micron-sized
dimensions. In this paper simulation studies are carried out on the stress analysis
along with displacement parameters.
Keywords Drug delivery · Microneedle · Skin · Insertion force · Stress
1 Introduction
Microneedles are usually being fabricated with various kinds of materials, such as
metal, silicon and polymeric [1], are micron-sized needles that can easily pierce
into the human skin. They are fabricated as either solid (Hashmi et al. 1995) or
hollow needles [2] and have recently been applied in the biomedical field for such
applications as the blood glucose measurement, transdermal delivery [2–5], skin
therapy and so on. Although microneedles may not totally replace the traditional
G. Aswani Kumar (B) · B. Rajesh Kumar · V. Jyothi · S. Injeti

GITAM Institute of Technology, GITAM (Deemed to Be University), Visakhapatnam
Rushikonda, 530045, India

https://doi.org/10.1007/978-981-15-4488-0_51
614 G. Aswani Kumar et al.
needles, they do possess some capabilities that the traditional needles do not and
show a great advantage in the biomedical field.
Most of the recent studies of microneedles have focused on their fabrication meth-
ods and drug delivery capabilities. In the year 1998, researchers created micronee-
dles based on the fundamental principles of needle injection which is the conven-
tional injection towards transdermal drug delivery. Researchers demonstrated that
biodegradable polymer microneedles possess the mechanical robustness and sharp-
ness to penetrate into human skin. In the last years of the nineteenth century, research
community created hollow metal microneedles for localized drug delivery. In early
2000, the usage of a single glass microneedle to pierce into the skin of diabetic
hairless rats through invasive technique to deliver insulin was investigated. Sivamani
et al. [6] tested hollow microneedles [6] by surmounting them onto a conventional
hypodermic needle syringe for the delivery of methyl nicotinate on human partici-
pants which yielded them good results. Further extension of the microneedle design
and fabrication towards drug delivery was demonstrated by Prausnitz [7] and his
team in the year 2004. At the outset in spite of a wide range and variety of micronee-
dle designs towards different applications that have been reported in the last few
decades, the fundamental mechanics and the techniques towards the insertion of
microneedles into the skin of human subjects, have received a very limited notice
and attention. The core design objective of microneedles is that they should be care-
fully administered into the skin by following precautionary safety measures and see
to it that the microneedles do not break or bend to achieve a painless and minimal
invasiveness based treatment. The insertion force is considered to be one of the key
indicating parameters which are generally estimated for the safe use of micronee-
dles. When the insertion force is too large or more, the microneedles will have a
greater tendency to break or buckle before the microneedle is totally inserted into the
skin. A better understanding of the insertion process and the relationship between
the microneedle geometry and the insertion force is helpful to the microneedle opti-
mization design. Davis et al. (2004) measured and predicted that the force necessary
to insert a tapered, hollow metal kind of microneedle into human skin with the help
of a force-displacement measurement kind of device which gives an approximately
linear dependency of the insertion force on the cross area of the needle tip. In their
study, tapered hollow metal microneedles were used having a tip radius of 30-80 mm,
wall thickness of 5-58 mm and having a constant length of 720 mm.
The insertion force for a single microneedle was found to vary from 0.08 to 3.04 N.
Recently, Roxhed et al. [8] have developed [8] and fabricated a type of ultra-sharp hol-
low microneedles whichgaveaninsertionforcelessthan10mN. All of the experiments
mentioned above have studied the insertion properties of microneedles into human
skin. However,noefficientnumericalmodelhasbeensetupfor predicting the insertion
process of a microneedle into human skin.
In this paper, a conically shaped microneedle was considered which will have a
primary force being applied in the vertical direction and a lesser secondary force
to be applied in the horizontal direction. The von-mises stresses in the needle as
well as the deformation of the needle will be analyzed in a 3-d plot. A multilayered
skin model consisting of three layers, stratum corneum, dermis and hypodermis, is
Effect of Insertion Force for Successful Penetration … 615
proposed. With the multilayered microneedle-skin model, the skin deformation and
failure involved in the insertion process [9, 10], as well as the influences of several
parameters (e.g. mechanical properties of skin, wall angle, tip are and wall thickness
of the microneedle) on the insertion force of the microneedle, have been studied. The
finite element analysis [11, 12] provides some new insights into the insertion process
of the microneedle into human skin.
1.1 SkinPhysiology
Human skin, being the largest organ of the human body, is a highly systematically
organized and its structure chiefly consisting mainly of three layers: epidermis [13]
mainly serves as a front line barrier [14] towards infection; the dermis layer which
serves as a subcutaneous layer with adipose being called also as a basement mem-
brane (Fig. 1). The outermost layer of the skin or epidermis is called the stratum
corneum [15] with an ideal thickness of about 10-20 mm. The second layer of the
skin is called a dermis layer, with a dense fibro-elastic [16] connective tissue layer
of thickness 1-4 mm (Odland [16]). The third layer also called the subcutaneous fat
Fig. 1 Schematic view of the cross-section of human skin

layer or the hypodermis layers is almost composed of loose fatty connective tissue
connecting to the dermis.
1.2 Mechanical Properties
The mechanical behavior of human skin was under investigation [17] and research
is focused on the skin behavior (Payne [17]; Holzapfel 2001; Hendrikset al. [18]).
As a matter of fact, the human skin is generally considered to be a heterogeneous
substance with a property of viscoelasticity [19] with an irregular pattern being under
continual stress (Langer [19]). Skin as a homogeneous material without consider-
ing the individual contributions of the epidermis, dermis and hypodermis. Actually,
different layers have rather different mechanical properties.
1.3 Basic Assumptions and Mechanical Properties

of the Skin
The geometrical parameters of the microneedles are analyzed in the present study
which is partially dependent upon the experimental design proposed by Davis et al.
(2004). All the microneedles used in this numerical modeling [12] are tapered with
respective to their geometrical shape. Here, the length which is a primary parameter
of the tapered microneedles is also considered as the same order as of the thickness
of skin layer. In order to get a better insight of overall insertion process, the mechan-
ical behavior of different skin layers should be studied on an initial basis. In this
simulation, a multilayer finite element model of skin including stratum corneum [15,
20], dermis and hypodermis is developed. It was taken into consideration that the
skin is to be as a simplistic model primarily consisting of stratum corneum, dermis
and the hypodermic layer.
Table 1 lists the mechanical properties of different layers of the skin used in
the numerical [11, 20] simulations. Although experiments need to be conducted to
determine the exact mechanical failure properties [17] of the lower layers of the
epidermis. Initially, the lower layers of the skin which is the epidermis will tend to
have a lesser impact as compared to that of the whole deformation of the skin model as
Table 1 Mechanical
Density r Failure stress Friction
properties of human skin
(kg/m3) s (MPa) coefficient m
Stratum 1300a 13–44b 0.42c
corneum
Dermis 1200d 7.3e 0.42c
Hypodermis 971e – –
Table 2 Parameters required

Mesh quality Normal
for processing
Number of node points Not applicable
Total number of elements 6805
Degree of freedom 31,755
a matter of fact due to smaller thickness. The force which is acting horizontally on the
needle is assumed to be significantly less than the force required to penetrate into the
skin, which is called the force of insertion or the insertion force, which can be easily
justified by the elastic property [2, 21] of the skin. Any force which when applied
in the horizontal direction would be unintentional and therefore considered to be
significantly smaller in magnitude, than that of the force which when punched in the
vertical direction. When considering each needle on the patch, they are assumed to be
identical and by adding more needles to the model, it would not significantly change
the forces acting on each individual needle, the forces here gets distributed which
will decrease the stress on a single needle, but will distribute equally throughout
the patch. Since the focus is on estimating the maximum stress there gives a scope
to neglect the decreasing stress factor. The interface between the top of the needle
and adjoining surface is always maintained to be at a fixed constraint which can be
justified by interfacing between the top of the needle and base of the needle surface
which is by default considered to be a continuous geometry. In simple it means that
there is so physical connection such as a weld that would fail before the rest of the
structure.
The simulation was a stationary model using a fully coupled stationary direct
solver. The advanced parameters were a MUMPS solver with 1.2 memory allocation
factors and automatic pre-ordering algorithm (Table 2).
2 Geometry and Force of Insertion
Height of the needle to be considered as 400 µm, Top outer radius 75 µm, Top
inner radius 25 µm, Bottom outer radius 50 µm, Bottom inner radius 10 µm. The
microneedle needed to be long enough to penetrate the epidermis to a thickness
level = 130 µm. The microneedle needs to be sturdy enough to withstand the force
required to penetrate human skin Force to pierce skin = 2 N. The microneedle needs
to be sturdy enough to withstand any lateral shearing force caused by motion of the
needle after insertion. Lateral shearing force = 0.5 N. The microneedle tip opening
needs to large enough to provide a sufficient flow of vaccine to the skin Inner radius
= 20 µm. The thickness of each layer of skin is considered as the average value
of human body, for example, the stratum corneum is 20 mm, dermis is 1.5 mm and
hypodermis is 1 mm.
2.1 Governing Equations
Compressive Force
The maximum compressive force the microneedle can withstand without breaking
is given by
FCompressive = бy. A (1)
where by is the fracture strength of the material. Where A is the cross-sectional area
of the microneedle.
Skin Puncturing Force
As a microneedle is inserted, it experiences resistive forces by human skin. In
order to successfully penetrate human skin, the applied force should be greater than
this opposing force (Fig. 2 and 3, Table 3).
FSkin = Ppuncturing A (2)
Fig. 2 Principle stress exerted by the needle

Fig. 3 Total displacement of the needle
Table 3 Parameters
Parameter Value
considered with respective
values Young’s modulus 17.0 GPa
Density 1500 kg/m3
Yield strain 0.05
Yield stress 80 Pa
3 Results
(Table 4, 5, 6 and 7)
Table 4 Resultant outputs

Parameter Value
for various parameters
Length of the needle 400 µm
Diameter of the needle 20 µm
Resultant stress 0.541 GPa
Total displacement 28.2 µm
Table 5 Deflection versus

bmax 0.541 GPa
max stress
Total deflection 28.2 µm
Table 6 Increase of needle

Top outer radius 92.5
diameters by 25%
Bottom outer radius 62.5
Table 7 Change in stress

Max stress bmax of increase in dimensions 0.3034 GPa
level
Percentage 23%
4 Conclusion
A conically shaped microneedle will have a primary force applied in the vertical
direction and a lesser secondary force will be applied in the horizontal direction. The
von-mises stresses and the displacement of the needle are analyzed in a 3-d plot.
We could able to meet the physical design criteria by constructing a needle to fit the
specified requirements. The length of 400 µm is long enough to pierce the epidermis
and the diameter of 20 µm is large enough to achieve sufficient flow through the tip
of the needle. From our first plot, we see that the maximum stress reached any point
in the needle is 0.541 GPa. From this, we can be assured that the needle will not
break. 0.541 GPa is also less than the yield strength of the needle which is 17.0 GPa.
It underwent a displacement of 28.2 µm from its center. By altering the needle radius
parameters, we can see there is a change in the max stress being exerted down to
0.3034 GPa which has a decrease in percentage level. By increasing the needle radii
there is a significant decrease in the percentage of stress exerted. This gives an insight
into the scientific community to look into needle diameters which has an impact on
the stress level. From the stress analysis, we see that the needle can withstand the
applied normal forces and applied shearing forces. Simulations were carried out on
COMSOL 5.0 version.
Future work
From the present research analysis, it can be concluded that the work can be
extended to complete structural analysis of design which includes bending stress,
buckling stress and thereby to fabricate and test the microneedle.
Acknowledgements This work was executed with the support of the Department of Science and
Technology-Science and Engineering Research Board (DST-SERB) New Delhi, India for providing
[Project No: ECR/2017/001183] under ECRA-SERB scheme.
Conflict of Interest: The contributing authors declare that there is no conflict of interest
regarding the article.
References
1. Moon SJ, Seung SL (2003) Fabrication of microneedle array using inclined LIGA process.
Paper presented at: proceedings of transducers’03. Boston, USA
2. Lin L, Pisano AP (1999) Silicon-processed microneedles. J Microelectromech Syst 8:78–84
3. Henry S, McAllister DV, Allen MG, Prausnitz MR (1998) Micro fabricated microneedles: a
novel approach to transdermal drug delivery. J Pharm Sci 87:922–925
4. Ashraf MW, Tayyaba S, Afzulpurkar N, Nisar A, Bohez ELJ, Lomas T, Tuantranont A (2010)
Design, simulation and fabrication of silicon microneedles for bio-medical applications, in
ECTI-CON 9(1)
5. McAllister DV (2000) Microfabricated needles for transdermal drug delivery [PhD thesis].
Georgia institute of technology
6. Sivamani R, Stoeber B, Wu G, Zhai H, Liepmann D, Maibach H (2005) Clinical microneedle
injection of methyl nictinate: stratum corneum penetration. Skin Res Technol 11:152–156
7. Lee JW, Park J-H, Prausnitz MR (2008) Dissolving microneedles for transdermal drug delivery.
Biomaterials 2113–2124
8. Roxhed N, Gasser TC, Griss P (2007) Penetration-enhanced ultra-sharp microneedles and
prediction on skin interaction for efficient transdermal drug delivery. J Microelectromech Syst
16:1429–1440
9. Shergold OA, Fleck NA (2004) Mechanisms of deep penetration of soft solids with application
to the injection and wounding of skin. Proc R Soc London Ser A 460:3037–3058
10. Shergold OA, Fleck NA (2005) Experimental investigation into the deep penetration of soft
solids by sharp and blunt punches, with application to the piercing of skin. J Biomech Eng
127:838–848
11. Louzidou EZ, Williams NA, Barrow DA, Eaton MJ, McCrory J, Evans SL, Allender CJ
(2015) Structural characterization and transdermal delivery studies on sugar microneedles:
experimental and finite element modelling analyses. Euro J Pharm Biopharm 89:224–231
12. Olatunji O, Das DB, Garland MJ, Belaid L, Donnelly RF (2013) Influence of array interspacing
on the force required for successful microneedles skin penetration: theoretical and practical
approaches. J Pharma Sci 102:1209–1221
13. Geerligs M, van Breemen L, Peters G, Ackermans P, Baaijens F, Oomens C (2011) In vitro
indentation to determine the mechanical properties of epidermis. J Biomech 44:1176–1181
14. Norlén L (1999) The skin barrier: structure and physical function [PhD thesis]. Karolinska
Institute, Stockholm, Sweden
15. Menon GK, Cleary GW, Lane ME (2012) Structure and function of stratum corneum. Int J
Pharma 435:3–9
16. Odland GF (1991). Structure of the skin. In: Physiology biochemistry, and molecular biology
of the skin. Oxford University Press, New York
17. Payne PA (1991) Measurement of properties and function of skin. Clinical Phys Physiol Meas
12:105–129
18. Hendriks FM, Brokken D, Oomens CWJ, Baaijens FPT (2004) Influence of hydration and
experimental length scale on the mechanical response of human skin in vivo, using optical
coherence tomography. Skin Res Technol 10:231–241
19. Langer K (1978) On the a natomy and physiology of the skin. Br J Plast Surg 31:3–8
20. Mane N, Gaikwad A (2014) MEMS based structural and fluid flow mechanics and simulation
analysis of microneedle for drug delivery. Int J Mech Prod Eng 2 ISSN:2320-2092
21. Magnenat-Thalmann N, Kalra P, Leveque JL, Bazin R, Batisse D, Querleux B (2002) A
computational skin model: foldand wrinkle formation. IEEE Trans Info Technol Biomed 6:317
22. Martanto W, Baisch SM, Costner EA, Prausnitz MR (2005) Fluid dynamics in conically tapered
microneedles. AIChE J 51:1599–1607
Lumbar Discography: Study
of Biomechanical Changes in the L1-L2
Intervertebral Disc of the Human
Lumbar Spine Using Finite Element
Methods
S. Balamurugan, P. Susai Manickam, and Sachit Chawla
Abstract Lumbar discography is often used as a pain recognition technique for

evaluating patients who have otherwise not responded well to non-surgical methods
(radiology). While the injection technique provides unique information about the
disc which is not obtainable through other methods, the procedure is suspected to
cause prolonged damage to the disc. A study concentrating on equivalent stress in
different regions of the disc shall provide useful insight on the prolonged effects of
the lumbar discography procedure. A computational model of the functional spinal
unit that accurately mimics its material properties and motion is required to simulate
the medical procedure performed on the intervertebral disc. This study uses a hybrid
finite element model of the L1-L2 motion segment of the lumbar spine of a 35-year-
old male subject. In order to mimic spinal motion during flexion and extension, all
supporting ligaments were added to the finite element model. Hyperelastic material
properties were altered to simulate the discography procedure and equivalent von
Misses stress values across the Annulus Fibrosus were recorded. A considerable
increase in the equivalent stress across the annulus was observed in the punctured
controls. Thus, the injection technique alters the stress concentrations across the disc
and can have prolonged degenerative effects on the IVD.
Keywords Intervertebral disc · Lumbar discography · Lumbar spine
1 Introduction
Degeneration of the intervertebral is characterized by the decrease in disc height and

desiccation. [1, 2] A degenerated IVD can lead to lower back pain. The most common
method used for imaging lumbar spine pathologies is magnetic resonance imaging.
MRIs provide a detailed image of the degenerated discs and other abnormalities in
the region of the lumbar spine. While the modality is useful for identifying abnormal
S. Balamurugan (B) · P. S. Manickam · S. Chawla

Department of Mechanical Engineering, SRM Institute of Science and Technology, SRM Nagar,
Kattankulathur, Chengalpattu 603 203, Tamil Nadu, India
e-mail: [email protected]; [email protected]

https://doi.org/10.1007/978-981-15-4488-0_52
624 S. Balamurugan et al.
disc pathology in the spine, it cannot prove if the abnormality is the cause of pain. [3,
4] To overcome this problem an invasive medical procedure, lumbar discography, is
used by physicians to record the patient’s response to the pain.
1.1 Lumbar Discography
This invasive medical procedure has been used by physicians for over 60 years and
involves an injection technique to insert fluoroscopy-guided needle in order 22- or
25-gauge needle from a posterolateral approach into the IVD [5]. Upon injection
of contrast, the patients’ response to the pain sensation is recorded. The response is
recorded on a Likert scale where patients rate their pain between 0 and 10, where 0
is no pain and 10 is unbearable pain [6]. The contrast is then viewed in CT images to
check for abnormalities in the disc as well as any ruptures present in the annulus region
which may be reflected by contrast. The procedure is conducted on both abnormal
and normal controls. Proponents of the procedure believe that information extracted
from discograms is unobtainable via any other non-surgical methods. However, it
has been strongly argued that the information obtained is extremely sensitive and
specific to patient response [7]. A considerable number of cases have been registered
wherein the patients experience similar or in some cases, an increase amount of
pain upon the removal of pain generator based on the discography results. Further, a
puncturing the disc during the process of injecting contrast can also lead to ruptures
in the annulus and can lead to prolonged degeneration of the IVD [8]. The aim of this
study is to characterize changes in the equivalent stress across the annulus fibrosus
before and after the lumbar discography procedure utilizing a computational model
of L1-L2 motion segment of the lumbar spine. The FE model was validated using
data from in vitro studies [9] and adjusted to achieve the desired accuracy.
1.2 Intervertebral Disc
The intervertebral discs lie between the vertebral bodies and link them together. They
are the main joints of the spinal column and occupy one-third of its height. Their major
role is mechanical, as they constantly transmit loads arising from body weight and
muscle activity through the spinal column. They provide flexibility allowing bending,
flexion and torsion. The intervertebral disc constitutes of mainly three components,
namely the annulus fibrosus, nucleus pulposus and the cartilage end-plates.
1.2.1 Nucleus Pulposus
The central nucleus pulposus contains collagen fibers, which are organized randomly,
and elastin fibers (sometimes up to 150 mm in length), which are arranged radially;
Lumbar Discography: Study of Biomechanical … 625
these fibers are embedded in a highly hydrated aggrecan-containing gel. The nucleus
acts as an incompressible unit that transfers loads evenly across the annulus and also
aids the transfer of nutrients to and from the intervertebral disc.
1.2.2 Annulus Fibrosus
This is made up of a series of 15 to 25 concentric rings, or lamellae, 11 with the

collagen fibers lying parallel within each lamella. The fibers are oriented at approxi-
mately 60 to the vertical axis, alternating to the left and right of it in adjacent lamellae.
Elastin fibers lie between the lamellae, possibly helping the disc to return to its orig-
inal arrangement following bending, whether it is flexion or extension. They may
also bind the lamellae together as elastin fibers pass radially from one lamella to the
next.
1.2.3 Cartilage End-Plate
The third distinct feature of the IVD is the cartilage end-plates. These are a thin
horizontal layer, usually less than 1 mm thick, of hyaline cartilage. The college fibers
within the end-plates are aligned parallel and horizontal to the vertebral bodies.
2 Methodology
A computational model of the L1-L2 motion segment of the lumbar spine was used
to investigate biomechanical changes. A hybrid finite element model that used both
quad and tetrahedral elements was developed where the main vertebral body and the
intervertebral disc components were meshed using quad elements while the spinous
process utilized tetrahedral elements. The use of the same type of elements across
the functional spinal unit provides better load transfer across biological components
compared to using different types of elements. The generation of the finite element
model from medical imaging techniques involved mainly three stages:
2.1 Geometric Modelling
CT images of the lumbar spine obtained from a 35-year-old male from SRM Medical
College were processed using Mimics Medical by Materialize (Fig. 1). Thresholding
values were adjusted to mask the vertebral column separately. The intervertebral
disc tissue shares its thresholding values with the muscular tissue surrounding it thus
making it difficult to extract the exact geometry of the IVD using slices obtained
from CT method. This meant that the disc needed to be reserve engineered using the
Fig. 1 Stages of obtaining

CAD model from CT images
(a) Masked vertebral bodies
(b) Unprocessed CAD model
(c) CAD model after topological refinement

inferior and superior faces of the L1 = L2 vertebra respectively. The CT images were
edited slice by slice along the transverse plane in order to separate the two vertebral
bodies and allow for their unrestricted motion in the finite element model. Topological
refinements of the now obtained geometric model were carried out using 3-Matics
by Materialize. The surfaces of the vertebral bodies consisted of planar holes which
were needed to be filled in order to maintain the continuity of the outer cortical shell.
The internal trabecular structure of the vertebrae was also removed so that the inner
portion of the vertebrae could be treated as the cancellous core.
2.2 Finite Element Modelling
Finite element model was generated using Hypermesh by Altair Engineering (Figs. 2,
3). Morphing options available in Hypermesh were used to morph geometrically
symmetric 2D meshed surfaces to oblong vertebral bodies. A concentric circular
pattern was developed on the inferior and superior faces of the vertebral bodies and
annulus fibers were created along these circular paths. Tension-only 1D elements
(LINK 10) were used to create the annulus fibers. These 1D elements connected
the diagonally opposite nodes of the quad elements and mimicked the non-linear
behaviour of the disc under extension and flexion. To simulate cartilage end-plates a
solid layer of quad elements 0.3 mm thick was created on both inferior and superior
faces of the L1 and L2 vertebrae respectively. All ligaments connecting the two
vertebral bodies were also incorporated in the finite element model to facilitate true
motion of the motion segment under load. The seven ligaments included in the
model were the anterior longitudinal (ALL), posterior longitudinal (PLL), flavum
(a) Finite Element Model with (b) Cross-section of the FEM along
all cable elements the mid-Sagittal plane
Fig. 2 a Cortical shell b Cancellous bone c Cartilaginous end-plates d Nucleus pulposes e Annulus
ground substance f Spinous process
(a) FEM of IVD complete with Annulus Ground Substance and

Nu-cleus
(b) Annulus Fiber generated using LINK10 elements from

Layer 1(in-nermost) to Layer 6(outermost)
Fig. 3 Intervertebral disc
(LF), intertransverse (ITL), interspinous (ISL), and supraspinous (SSL) ligaments.

These ligaments were modelled using 1D elements which displayed zero stiffness
value under compression. These elements were attached to appropriate landmarks
on the vertebral bodies. Node to node contact was defined between the intervertebral
disc and the vertebral bodies to achieve complete load transfer along the motion
segment.
Fig. 4 Material and mechanical properties of the cortical and cancellous bone as well as spinous
process
Fig. 5 Mechanical properties of annulus fibers and ligaments
2.2.1 Material Models
Material models that accurately depicted the biological properties of the vertebral
bodies as well as the intervertebral disc were assigned to different components in
the finite element model. The anisotropy in the material properties of the vertebral
bodies was considered [10, 11] (Fig. 4). A Mooney-Rivlin model was used to define
the material properties of the Annulus and the Nucleus. Mooney-Rivlin constants
were altered to simulate the discography procedure [12] (Fig. 7). The mechanical
properties and cross-sectional area of all 1D elements were assigned to best mimic
the biomechanical behaviour of the motion segment [13, 14] (Fig. 5).
The coordinate system used by Panjabi et al. [9] was followed in order to validate the
model. A pilot node was defined at the center of the superior face of the L1 vertebral
Fig. 6 Range of motion of functional spine unit for Flexion and Extension
Fig. 7 Material and mechanical properties of nucleus pulposes before and after the lumbar
discography procedure
body and moment loads were applied on this point about the X-axis. The pilot node
was connected to the superior face using rigid elements. The inferior face of the L2
vertebra was constrained with 0 DOF in all directions.
3 Validation of Finite Element Model
The mechanical behaviour of generated FE model was validated by comparing exper-

imental and finite element values (Fig. 6) for the range of motion of the L1-L2 motion
segment.
3.1 Flexion-Extension-Moment Loading
The values for range of motion segment under pure flexion and extension loading
obtained from the FE model were recorded and compared to in vitro studies [9].
ROM noted from the FE model lie within the standard deviation of the experimental
mean values for most of the loading cases. The average error between ROM values
predicted by the FE model and those obtained from literature for extension and
flexion was 42.1% and 26%, respectively.
In order to simulate disc puncture the material properties of the nucleus were changed.
Due to water loss during disc puncture there is thought to be a stiffening of the NP
region. This concept is based on the results of a study by Iadritis et al. stating that
even a small disruption to the annulus fibrosis of the disc due to disc puncture can
lead to depressurization of the nucleus pulposus [15]. Flexion-extension-moment
loads were applied on the L1-L2 motion segment and the maximum von Misses
stress values were recorded. This simulation was performed for both normal and
punctured controls. The results predicted by the FE model in both cases were then
compared.
4.1 Flexion-Moment Loading
A comparison of equivalent von Misses stress during flexion-moment loading showed

that the discography procedure causes an increase in equivalent stress across the AF
region (Fig. 8). The stress values recorded after simulating the procedure displayed
a non-linear trend and a decrease in stress were observed for high moment loads
beyond 9.5 Nm. This non-linearity in results can be explained by the non-linear
nature of the annulus fibers. The maximum percentage increase in stress observed
for flexion loading was 3.11%.
Fig. 8 Comparison of von misses stress produced across the AF region as predicted by the FE
model during flexion-moment loading before and after lumbar discography procedure
Fig. 9 Comparison of von misses stress produced across the AF region as predicted by the FE
model during extension-moment loading before and after lumbar discography procedure
4.2 Extension-Moment Loading
A similar trend of increased equivalent stress was observed for extension-moment

loading in the punctured disc. A decrease in equivalent stress was observed for lower
values of moments between 3 Nm and 4.5 Nm. The maximum percentage increase
in stress observed for extension loading was 6.80% (Fig. 9).
5 Conclusion
The current study on L1-L2 motion segment of the lumbar spine examines the biome-
chanical changes in the intervertebral disc brought about by the lumbar dicscography
procedure. The simulations carried out on the validated FE model predict an overall
increase in the von Misses stress across the Annulus Fibrosus upon disc puncture
caused by the injection technique. A considerable increase of 6.80% in equivalent
stress during extension provides proof that the medical procedure causes changes
in the biomechanics of the disc even under simple loading conditions. Complex or
impact loading of the motion segment may cause even more significant changes in
the IVD biomechanics. Since the healing of the IVD after puncture is frustrated by a
small population of cells that have limited availability to nutrients and is repeatedly
exposed to mechanical motions, [16] these changes in the stresses induced in the
disc can cause prolonged degeneration of the Annulus Fibers. Thus, a change in the
biomechanics of the disc due to the puncture was confirmed but if these changes can
onset early degeneration of the disc is still being studied.
Acknowledgements This work was supported by the Bio mechanics lab, Department of Mechan-
ical Engineering, SRMIST, Kattankulathur—603203. We thank Mr. Raja D, Asst Professor
(S.G) and Research Head, Bio mechanics lab, Department of Mechanical Engineering, SRMIST,
Kattankulathur—603203 for assistance to improve the manuscript.
References
1. An HS, Inoue N, Masuda K (2006) Intervertebral disc degeneration: biological and biome-
chanical factors. J Orthop Sci 11(5):541–552
2. Urban JP, Roberts S (2003) Degeneration of the intervertebral disc. Arthritis Res Ther 5(3):120–
130
3. Guyer RD, Ohnmeiss DD (2003) Lumbar discography. Spine J 3(3 Suppl):11S–27S
4. Kim YH, Kang CH, Lee S et al (2009) Can magnetic resonance imaging accurately predict
concordant pain provocation during provocative disc injection? Skeletal Radiol 38(9):877–885
5. Hirsch C (1948) An attempt to diagnose level of disc lesion clinically by disc puncture. Acta
Orthop Scand 18:131–140
6. Moneta GB, Videman T, Kaivanto K et al (1994) Reported pain during lumbar discography
as a function of anular ruptures and disc degeneration. a re-analysis of 833 discograms. Spine
(Phila Pa 1976) 19(17):1968–1974
7. Carragee EJ, Alamin T (2001) Discography. a review. Spine J 1(5):364–372
8. Carragee EJ, Don AS, Hurwitz E et al (2009) Does discography cause acceler-ated progression
of degeneration changes in the lumbar disc: a ten-year matched cohort study. Spine (Phila Pa
1976) 34(21):2338–2345
9. Panjabi M, Oxland T, Yamamoto I, Crisco J (1994) Mechanical behavior of the hu-man lumbar
and lumbosacral spine as shown by three-dimensional load-displacement curves. J Bone Joint
Surg 76(3):413–424
10. Ueno K (1984) A three dimensional nonlinear finite element model of the lumbar intervertebral
joint. [phd thesis]. The University of Iowa, Iowa
11. Lang SB (1970) Ultrasonic method for measuring elastic coe cients of bone and results of fresh
and dried bovine bones. IEER Trans Biomed Eng 17:101–105
12. Lipscomb KE, Sarigul-Klijn N, Klineberg E, Mohan V (2017) Biome-chanical e ects of human
lumbar discography: in-vitro experiments and their finite element validation. Clin Spine Surg
30(3):E219–E225
13. Shirazi-Adl S.A., Shrivastava, S. C., Ahmed, A. M (1984). Stress analysis of the lumbar disc-
body unit in compression. a three-dimensional nonlinear finite element study. Spine 9, 120134
14. Smit TH, Odgaard A, Schneider E (1997) Structure and function of vertebral trabec-ular bone.
Spine 22:2823–2833
15. Iatridis JC, Michalek AJ, Purmessur D et al (2009) Localized intervertebral disc injury leads to
organ level changes in structure, cellularity, and biosynthe-sis. Cell Mol Bioeng 2(3):437–447
16. Adams MA, Stefanakis M, Patricia D (2010) Healing of a painful interverte-bral disc should
not be confused with reversing disc degeneration: im-plications for physical therapies for
discogenic back pain. Clin Biomech (Bristol, Avon) 25(10):961–971
C4–C5 Segment Finite Element Model
Development and Investigation
of Intervertebral Disc Behaviour
P. Susai Manickam
Abstract Clinical problems of the human cervical spine continue to be widespread

in our society. In the past few decades, many experimental and simulation tech-
niques were adopted to study the motion of the cervical spine. The functional spine
unit is made of vertebrae, end plates, intervertebral disc and ligaments. To under-
stand the biomechanical behaviour of cervical spine a detailed finite element model
is needed. The focus of the study is to develop the cervical spine from (C4–C5)
finite element model with accurate dimensions and representation of the material
properties. A three-dimensional finite element (FE) model of cervical spine seg-
ment (C4–C5) was developed using computed tomography CT scan and applied to
study the stresses distribution of the intervertebral discs under quasi-static loading
conditions. A pure moment loading of 1.0 Nm was applied to study the physiolog-
ical motion. The model accuracy was validated by comparing the results with the
previously published experimental and numerical results for various physiological
motions. Several movements were analysed: flexion, extension, lateral bending and
axial rotation. By using elastic and hyperelastic model behaviour of the intervertebral
disc was accurately simulated. The results obtained by the elastic and the hyperelastic
models showed good agreement with experimental and numerical data. The current
model which reflects the behaviour of human cervical spine can be effectively used
to study further biomechanics and traumatic studies.
Keywords Spine · Cervical spine · Finite element model
1 Introduction
Injury in the cervical spine can be analysed according to the physiological motion
of the neck and the mechanical loads acting in the spine. Generally in the traffic
accidents, the damages are caused in the intervertebral disc soft tissues, ligaments
P. Susai Manickam (B)

Research Scholar, Department of Mechanical Engineering, SRM Institute of Science and
Technology, SRM Nagar, Kattankulathur, Chengalpattu 603203, Tamil Nadu, India

https://doi.org/10.1007/978-981-15-4488-0_53
636 P. Susai Manickam
and muscles. The load transmission and the shock absorber effect in the compression
is achieved by intervertebral disc. Clinician identified that the biomechanical study
in the disc helps to understand the loads and the physiological motion of the cervical
spine. Biomechanical models, such as in finite element (FE) models and vitro, helps
us to understand the mechanisms of injury and diagnosis and effective treatment of
cervical spine problems [1, 2]. The objective of the work is to model C4–C5 level
of the cervical spine by using finite element analysis. The intervertebral disc is the
major load transmitting components so the material models need to be assigned.
For the basic material properties, the validation was done using the experimental
results. The material models were assigned as linear elastic properties and non-linear
material models for the functional spine unit. The models can be used to predict the
physiological motion and the stress distribution across the intervertebral disc.
2.1 Model Construction
The vertebrae were classified into cortical, cancellous and the posterior side. The
thickness of the cortical layer is 0.5 mm. The intervertebral disc was classified as
disc-annulus fibrosis, nucleus pulposus, with inferior and superior endplates. The
thickness of the endplate is 0.6 mm. The ligaments are modelled for the functional
spine unit anterior longitudinal ligament (ALL), capsular ligament (CL), and inter-
spinous ligament (ISL), ligamentum flavum (LF), posterior longitudinal ligament
(PLL) (Figs. 1 and 2) .
Fig. 1 C4–C5 meshed

model
C4–C5 Segment Finite Element Model Development … 637
Fig. 2 C4–C5 model in

extension for a load of 1 Nm
2.2 FEA Modelling
The geometry of the C4–C5 model was obtained from computed tomography (CT)
data of a healthy 28-year old man from SRM medical college. The CT data were
processed with medical image processing software MIMICS (MIMICS_ Version
10.0; Materialise, Inc., Leuven, Belgium). The meshing of the vertebrae cortical
and cancellous bone, Intervetebral disc and endplates was performed using Hyper
mesh 11.0 (Altair Engineering, Inc., Executive Park, CA, USA). In the finite element
model of cortical and cancellous on the anterior side is modelled using hexahedral
elements and it consists of 6651 elements and 7920 nodes were used to build the
model. The posterior side of the model is having a high level of irregular surface so
it was meshed using tetrahedral elements and it consists of 109,269 elements and
23,321 nodes. According to the anatomy, the intervertebral disc and the ligaments
were modelled. In the intervertebral disc cross-sectional area approximately 50% was
comprised as nucleus pulposus [3]. The hexahedral elements were used to construct
the inferior and superior endplates. All the ligaments were modelled using non-linear
spring elements. The finite element model was developed for C4–C5 segment based
on the available material properties as shown in Table 1. After the mesh sensitivity
effect size of the elements are finalized. The material is assigned as shown in Table 1
for the bones and Table 2 for the ligaments.
Table 1 Finite element model details

S. No. Description Young’s modulus (Mpa) Poisson ratio
1 Cortical bone 12,000 0.29
2 Cancellous bone 450 0.29
3 End plate 500 0.4
4 Disc-annulus 3.4 0.4
4.1 Disc-annulus hyper elastic C1 = 0.56, C2 = 0.14 –
Mooney–Rivlin
5 Disc-nucleus 1 0.49
Ligament
6 Posterior longitudinal ligament 20 0.3
7 Anterior longitudinal ligament 54.5 0.3
8 Interspinous ligament 1.5 0.3
9 Ligamentum flavum 1.5 0.3
10 Capsular ligament 1.5 0.3
3 FEA Validation
In order to validate the FE model for the predicted physiological motion various
boundary condition was compared against the values reported in the following lit-
erature [2]. In the lower surface of C5, all the degree of freedom was constrained.
To simulate the range of motion a pure moment of 1 Nm is applied at the top of the
C4 vertebrae in the three planes. The validation of the FE model was performed in
ANSYS-19.0.
4 FE Analysis
In the simulation, the physiological motion was compared with the literature and
validated. The elastic property model was validated with the experimental results as
shown in Fig. 3. In this study, the elastic model material property and the hyperelastic
material properties were compared. The intervertebral disc was considered as elas-
tic and hyperelastic model was modelled. The material properties for hyperelastic
Mooney–Rivlin was assigned from [4]. After solving with the assigned boundary
condition the stress distribution across the disc needs to be studied.
Table 2 Ligament details of the finite element model
ALL PLL LF ISL CL
D F D F D F D F D F
0 0 0 0 0 0 0 0 0 0
1 35.5 0.9 1.33 1.7 2.2 1.2 0.75 1.7 2.452
2 64.9 2 29 3.74 45.9 2.7 16.9 3.9 53.6
4 89.7 3 51.4 5.61 82.9 4 24.4 5.8 87.9
C4–C5 Segment Finite Element Model Development …
5.00 108.60 4 71.38 7.48 119.6 5.4 29.5 7.7 109.4

6.00 119.60 5 85.8 9.35 133.7 6.7 32.9 9.7 125.8
6 94.7 11.3 147.2 8.1 34.9 11.5 134.8
D Displacement in mm and F force in Newton
639
Fig. 3 Comparison of the various physiological motion of cervical spine C4–C5 level
5.1 Flexion and Extension
The response of the intact model shows a good response with the existing experi-
mental results from [5]. In the flexion the model was more flexible and the angle is
little higher than the experimental values but in the extension, while comparing with
[6] the values are closer to it because it’s also an FEM model as shown in Table 3.
While comparing with the disc we modelled with the elastic model and hyperelastic
model. In the elastic model, we measured the stresses in the anterior side and the
posterior side of the annulus fibre was same. But while modelling it as a hyperelastic
model, the stresses are more in the anterior side than the posterior side during flexion
Table 3 Physiological motion of C4–C5 level

Author Panjabi Panzer Kallemeyn Haghpanahi Erbuluta Current
name [5] [7] [6] [8] [9] study
Year 2001 2009 2009 2012 2014 2019
Method Expt FEM FEM FEM FEM FEM
Load 1 Nm 1 Nm 1 Nm 1 Nm 1 Nm 1 Nm
Flexion 5.2 5.17 6 3.68 4 7.24
Extension 5.2 4.51 5.1 3.04 2.5 6.1
Lateral 6 7.62 4 1.47 5 8.1
bending
Axial 6 6.62 5.1 2.64 5.5 3.5
rotation
The values are in degree
Table 4 Stress distribution in

Intervertebral disc Elastic properties Hyperelastic
the intervertebral disc of
properties
C4–C5 level
Stress in MPa Stress in MPa
Flexion extension 2.61 3.31
Lateral bending 3.07 4.48
Axial rotation 1.05 1.14
and during extension the stresses of the posterior side are more than the anterior side.
This trend replicates the hyperelastic models shows more accuracy than the elastic
models. The stresses in the disc are more in the hyperelastic model than the elastic
model. The elastic and hyperelastic disc stresses are compared in Table 4 The finite
element model with extension is shown in Fig. 2.
5.2 Lateral Bending
In the lateral bending, the model is flexible and its range of motion is high compared
to that of the experiment values [5] but while comparing with the FEM model it is
very closer to the bending results of [7]. The elastic material model exhibits the less
stresses in the sides but that may due to the facet restrictions. But in the hyperelastic
models, the stress distribution is more and equally distributed in both the sides anterior
and posterior side. There is a significant change is observed in the top and the bottom
side of the annulus fibres during lateral bending. The comparison of the elastic and
hyperelastic models are shown in Table 4.
5.3 Axial Rotation
In the axial rotation, the stress the model is stiffer while comparing with the exper-
imental results [5]. But in comparison with the [8, 9], the results are closer in the
axial rotation. This may be due to multiple factors like in the axial rotation the facet
plays a major role and the ligaments also provide a locking mechanism. In this, while
comparing the elastic and hyperelastic model, there is no much change in the stresses
because the disc will not undergo considerable tension and compression. The stress
distribution in the axial rotation is shown in Table 4.
Fig. 4 Comparison of the various physiological motion of cervical spine C4–C5 level
6 Conclusion
The results are shown above clearly indicate that model shows a good agreement
with the physiological motion of the functional spine unit Fig. 4. The intact finite
element model was validated against the experimental study [5]. The range motion
for the functional spine unit was compared. There is a significant difference was
found in Von mises stress while comparing the elastic and hyperelastic material for
intervertebral disc. However, by modifying the material models will help us in finding
the real behaviour of the functional spine unit.
Acknowledgements This work was supported by the Bio mechanics lab, Department of Mechan-
ical Engineering, SRMIST, Kattankulathur–603203. We thank Mr. Raja. D, Asst Professor (S.
G) and Research Head, Bio mechanics lab, Department of Mechanical Engineering, SRMIST,
Kattankulathur–603203 for his assistance to improve the manuscript.
References
1. Panjabi MM (1998) Cervical spine models for biomechanical research. Spine 23(24):2684–2700
2. Zhang QH, Teo E-C, Ng H-W et al (2006) Finite element analysis of moment-rotation
relationships for human cervical spine. J Biomech 39:189–193
3. Gilad I, Nissan M (1986) A study of vertebra and disk geometric relations of the human cervical
and lumbar spine. Spine 11:154–157
4. Schmidt H, Galbusera F, Rohlmann A, Zander T, Wilke HJ (2012) Effect of multilevel lumbar
disc arthroplasty on spine kinematics and facet joint loads in flexion and extension: a finite
element analysis. Eur Spine J 21 (Supp l5): S663–S674
5. Panjabi MM, Crisco JJ, Vasavada A, Oda T, Cholewicki J, Nibu K, Shin E (2001) Mechanical
properties of the human cervical spine as shown by three-dimensional load–displacement curves.
Spine 26(24):2692–2700
6. Kallemeyn N, Tadepalli SC, Shivanna K, Grosland N (2009) An interactive multiblock approach
to meshing the spine. Comput Methods Programs Biomed 95:227–235
7. Panzer MB, Cronin DS (2009) C4–C5 segment finite element model development, validation,
and load-sharing investigation. J Biomech 42:480–490
8. Haghpanahi M, Javadi M (2012) A three dimensional parametric model of whole lower cer-
vical spine (C3–C7) under flexion, extension, torsion and lateral bending. Scientia Iranica B
19(1):142–150
9. Erbulut DU, Zafarparandeh I, Lazoglu I et al (2014) Application of an asymmetric finite element
model of the C2–T1 cervical spine for evaluating the role of soft tissues in stability. Med Eng
Phys 36:915–921
Simulation and Analysis of Integrated
SEPIC-Flyback AC-DC PFC Converter
for LED Applications
Sridhar Makkapati and Seyezhai Ramalingam
Abstract Light Emitting Diode (LED) inherently acts as a diode which allows
the current in one direction and this can be achieved by AC-DC converter. It is
essential to have sinusoidal input current wave shape and constant load current to
drive the LED without any flicker. Integration of Single Ended Primary Inductor
Converter (SEPIC) and Flyback converter act as an LED drive. Both these converters
are operated in discontinuous conduction mode for supplying LED to achieve the
unity power factor. It is compared with conventional SEPIC and Flyback topology.
Analysis and simulation of 100 W LED module from 100 VP-P /50 Hz ac supply are
analyzed and their results are presented using MATLAB-Simulink.
Keywords LED · SEPIC · Flyback converter · Power factor
1 Introduction
Light emitting diode is energy efficient and predominant light source to replace the
traditional sources like fluorescent and discharge lamps. It is due to several advan-
tages such as a predominant lifetime over 1,00,000 h during appropriate operating
conditions, a high light output of more than 150 lm/w [1], pollution-free and avail-
ability of multicolor variants. LED is employed for applications such as automotive
applications, street lighting, and decorative lighting [2]. But the LED also suffers
from several issues of concern unlike conventional luminaries such as driver circuit
is essential to drive LED, in which elimination of electrolytic capacitor (E-cap) [3] in
ballast design [4], lesser in size, economical, high efficiency, and necessary to meet
the IEC standards for LED lighting [5, 6], galvanic isolation and proper current shar-
ing control. Various methodologies have been introduced in the literature to solve
the aforementioned problems.
S. Makkapati (B) · S. Ramalingam

Renewable Energy Conversion Laboratory, Department of EEE, Sri Sivasubramania Nadar (SSN)
College of Engineering, Kalavakkam, Chennai 603110, India

https://doi.org/10.1007/978-981-15-4488-0_54
646 S. Makkapati and S. Ramalingam
Passive LED driver circuits without active switches are initially introduced, with
series limiting resistor circuit [7], and valley fill circuit [8] however these circuits are
repulsive because of their bulky size, high cost and lack of output current regulation.
In order to operate in several high-frequency applications, switched-mode driver
circuits with smaller size and precise current control are suggested with single-
stage [9], two-stage [10], and three stages [11]. Single-stage involves only a single
switch with one processing stage which performs both PFC and current regulation
simultaneously. But the major hardship in this stage is the output electrolytic capacitor
(E-cap) lifetime is lesser than the lifetime of LED. Two-stage drivers consist of two
processing stage where one stage is for PFC and the other performs the output current
regulation whereas the size and cost is the issue. Three-stage topology is suitable
for multi-string usage in which the initial two act as previous while the third stage
performs the equal current sharing among the strings.
For non-isolated applications, single-stage topologies such as buck [12], buck-
boost [13], SEPIC [14] and Cuk [15] are employed. In order to achieve high conver-
sion ratio using isolated flyback [16], half-bridge [17], and push-pull converter [18]
are proposed. Several combinations of isolated [19], non-isolated [20], and resonant
converters [21] are suggested as integrated topologies. An integrated boost-buck
topology is suggested in [22] with two active switches increased the complexity,
boost-forward converter is suggested where the switch poses the voltage spikes dur-
ing its off state. In reference [23] Boost-flyback integrated topology is proposed with
the additional winding on flyback that supports the reduction in the bulk capaci-
tor at the output. In this paper single-stage single switch, PFC SEPIC with flyback
converter is suggested without E-cap at the output.
The flow of this paper is as follows. Section 2 elaborates on the operation of
SEPIC-Flyback PFC converter. Its design steps are discussed in Sect. 3 and Sect. 4
presents the comparison of non-isolated SEPIC, isolated flyback with integrated
SEPIC-Flyback converter simulation results. Finally, Sect. 5 presents the conclusion.
2 Operation of Integrated SEPIC-Flyback PFC Converter
Integrated SEPIC-Flyback converter operation is discussed in this section with its

modes of operation. In order to achieve near unity power factor during the LED
driving process, SEPIC converter is integrated with Flyback converter as shown
in Fig. 1, however, both the converters operate in discontinuous conduction mode
(DCM). Non-isolated SEPIC converter consists of L in , C B , L, C, and D1 ; while the
flyback converter includes a transformer T, D2 , Do are the diodes and the output
capacitor C o . These two converters share a common switch with SEPIC converter
that acts as a PFC circuit and DCM flyback performs the DC regulation. Transformer
model is represented using magnetizing inductance Lm and an ideal transformer.
SEPIC-Flyback circuit consists front end diode rectifier and its output is fed
to integrated SEPIC and Flyback circuit. Operating modes with six different time
instants are explained. An assumption is made that switch is ideal and the higher
Simulation and Analysis of Integrated SEPIC-Flyback … 647
Fig. 1 Circuit diagram of SEPIC-flyback LED driver
switching frequency is chosen for the semiconductor device. Transformer turns ratio
n = N p /N s . Figure 2 shows the waveforms corresponding to one switching cycle
during different time instants.
Initially, it is assumed that all the switching devices are lossless and load current is
flown from the output capacitor. Operation of above mentioned integrated topology
is explained below.
Fig. 2 Waveforms during a switching period

Mode I [t o < t < t 1 ]: Switch S turns on in this mode and the rectified output e
is applied to the input inductor L in , voltages V C and V B are applied to L and L m
respectively. This allows the current I Lin , I L , I Lm to increase linearly from zero. DTs
is the switch S on time and assure the zero current switching of the device. Diodes
D1 and Do are reverse biased and this mode terminates after the period DTs .
Mode II [t 1 < t< t 2 ]: Switch S initially turns off during this period and the diodes
D1 and Do are forward biased. Due to D1 conduction voltage across the switch is V B
+ V C and the voltage V c − nV o is inversely applied across the L lk and the leakage
energy is absorbed. The difference in leakage current and magnetizing current I Lm
forward biases the diode Do at zero current switching and diode D2 turns off.
Mode III [t 2 < t< t 3 ]: During this mode, I Lin , I L and I Lm decrease linearly and
the D2 diode is turned on. The voltage across the inductors L in and L are—(V B +
V C V rect ) and −V B respectively.
Mode IV [t 3 < t< t 4 ]: D1 and Do are conducting and leakage and magnetizing
currents decrease. Current through inductor I L becomes zero at the end of this period
and D1 turns off in this mode of operation.
Mode V [t 4 < t< t 5 ]: In this mode diode D1 is in off state and switch voltage falls
to V C . I Lm current still decreases and the remaining inductor currents fall to zero.
Diode Do get reverse bias at the lapse of this mode.
Mode VI [t 5 < t< t 6 ]: In this mode, the semiconductor switch is turned off the
current through the load is driven by the output capacitor. This operation persists
until the switch turns on again.
3 Analysis and Design Consideration of SEPIC-Flyback

Converter
Design of SEPIC-Flyback converter with a sinusoidal line voltage Vin (t) =

Vm sin ωl (t) where ωl is line frequency 2π fl Average line current that flows through
inductor L in over a switching period T s is given as
1 1
Iin avg (t) = (DTs + tfLin )ILinpeak (1)
Ts 2
The peak inductor current ILinpeak and is given as
Vm DTs
ILinpeak = sin ωl (t) (2)
L in
The fall time of inductor peak current tfLin that reaches to zero and is calculated
as
Vm sin ωl (t)
tfLin = DTs (3)
VB + VC − Vm sin ωl (t)
Average line current substituting Eqs. (3) and (2) in Eq. (1)
Vm D 2 Ts VB + VC
Iin avg (t) = sin ωl (t) (4)
2L in VB + VC − Vm sin ωl (t)
Line current depends on V B + V C and its value is more than V m

By using Eq. (4) average input power is given by
Vm2 D 2 Ts π (VB + VC )sin2 θ

Pin = ∫ dθ
2π L in 0 VB + VC − Vm sin θ
kVm2 D 2 Ts
= (5)
2π L in
where k is a function of V m & V B + V C

Neglecting converter losses
Pin = Po (6)
Input inductor is obtained as
k D 2 Vm2
L in = (7)
2π f s Po
By selecting f s and P0 choose the value of input inductor

Power delivered to load, i.e. output power is
1
Po = 2
L m ILmpeak fs (8)
2
where
VB DTs
ILmpeak = (9)
Lm
Power output is obtained as
VB2 D 2 Ts
Po = (10)
2L m
Transformer magnetizing inductance
VB2 D 2
Lm = (11)
2 f s Po
Table 1 Simulation
Parameter Notation Value
parameters of SEPIC-flyback
converter Input voltage Vs 100 V (P-P)
Switching frequency fs 50 kHz
Duty cycle D 0.1
Input inductor L in 75 µH
Inductor L 150 µH
Magnetising inductor Lm 130 µH
Turns ratio n:1 1:1
Capacitor C 4.7 µF
Capacitor CB 4.7 µF
Output capacitor Co 2*4.7 µF
Output power Po 100 W
For DCM, the turns ratio is selected based on the rate of fall of magnetizing
current. Design formulae for L, C, C B , and C o are as follows
π Vc (VB + Vc )
L= L in (12)
kVm2
QC
C= c is low-frequency voltage ripple (13)
VC
QB
CB = B is low-frequency voltage ripple (14)
VB
(nVB DTS − L m I0 )2
Co = (15)
2n 2 L m Vo V0
where o is high-frequency output capacitor voltage ripple

Based on the above design Eqs. (1–15), the simulation parameter values are
computed and presented in Table 1.
4 Simulation Results and Comparative Analysis
In order to verify the design equations, the results are simulated using MATLAB–
Simulink for 100 W power rating driven with a voltage source of 100 V P-P. The
input side parameters are majorly concentrated to obtain a high power factor and
lesser Total harmonic distortion (THD). The suggested topology is compared with
the discontinuous conduction of SEPIC and Flyback converter. Simulated circuit for
SEPIC-Flyback is represented in Fig. 3.
Input voltage and current waveform and input inductor current are shown in Figs. 4
and 5 respectively and it resembles that the converter is operating in the discontinuous
Fig. 3 Simulation circuit of SEPIC-Flyback converter
150
Input voltage
100 Input current
Input voltage
50
Input current
-50
-100
-150
0 100 200 300 400 500 600 700 800 900 1000
Time
Fig. 4 Input AC voltage and current of the integrated converter
120
100
Inductor currrent
80
60
40
20
-20
0 100 200 300 400 500 600 700 800 900 1000
Time
Fig. 5 Input inductor current of the integrated converter

mode as the current through inductor reached zero. Based on the measured harmonics
with a single-phase ac supply power factor is 0.98 and THD of the simulated circuit
is 9.36% and is shown in Fig. 6.
In order to compare the suggested SEPIC-Flyback with isolated flyback and Non-
isolated SEPIC converter under discontinuous mode to drive the LED is simulated
with the designed parameters. Input current and voltage waveforms of both fly-
back and SEPIC in DCM operation are shown in Figs. 7 and 8 respectively. Their
comparison is shown in Table 2.
Table 2 Infers that SEPIC-Flyback integrated topology produces less harmonic
distortion and better power factor in comparison with the DCM SEPIC and DCM
Flyback converter.
Fig. 6 Supply current FFT spectrum
150
Input voltage
100 Input current
Input volatge
Input current
50
-50
-100
-150
0 100 200 300 400 500 600 700 800 900 1000
Time
Fig. 7 Input AC voltage and current of DCM flyback converter

150
100
Input voltage
Input current
Input voltage
Input current
50
-50
-100
-150
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
4
Time x 10
Fig. 8 Input AC voltage and current of DCM SEPIC converter
Table 2 Comparison of input

Topology Power factor THD (%) Distortion
parameters
factor
SEPIC 0.975 65.7 0.83
converter-DCM
Flyback 0.859 121.15 0.63
converter-DCM
SEPIC-flyback 0.980 9.38 0.99
converter
5 Conclusion
In this paper single-stage, SEPIC integrated flyback converter is investigated in detail

with its design equations. This converter, when fed from AC source, achieves better
power factor with lesser total harmonic distortion compared to flyback and SEPIC
converter topologies. In the proposed topology, both the converters operate in DCM
and leakage energy is fed back by the magnetizing reactance of flyback converter
and SEPIC converter for active PFC with the lesser output capacitor.
References
1. Schubert EF (2006) Light-emitting diodes, 2nd edn. Cambridge Univ. Press, Cambridge, U.K
2. Chen W, Hui SR (2012) Elimination of an electrolytic capacitor in AC/DC light-emitting diode
(LED) driver with high input power factor and constant output current. IEEE Trans Power
Electron 27(3):1598–1607
3. Linlin G, Xinbo R, Ming X, Kai Y (2009) Means of eliminating electrolytic capacitor in AC/DC
power supplies for LED lightings. IEEE Trans Power Electron 24:1399–1408
4. Electromagnetic compatibility (EMC) (1995)—part 3: limits-section 2: limits for harmonic
current emissions (equipment input current < 16 A per phase), IEC1000-3-2
5. ENERGY STAR Program Requirements for Solid State Lighting Luminaires http://www.
energystar.gov/index.cfm?c=new_specs.ssl_luminaires
6. Hui SYR, Li SN, Tao XH, Chen W, Ng WM (2010) A novel passive offline LED driver with
long lifetime. IEEE Trans Power Electron 25(10):2665–2672
7. Gunabalan R (2016) Overview of passive light emitting diode driver circuits for street lighting.
Bull Electr Eng Inform 5(3)
8. TianFu P, HuangJen C, ShihJen C, ShihYen C (2007) An improved single-stage flyback PFC
converter for high-luminance lighting LED lamps. In: 8th international conference on electronic
measurement and instruments
9. Kim BC, Park KB, Kim CE, Lee BH, Moon GW (2010) LLC resonant converter with adap-
tive link-voltage variation for a high-power density adapter. IEEE Trans Power Electron
25(9):2248–2252
10. Almeida PS, Braga HAC, Dalla Costa MA, Alonso JM (2015) Offline soft-switched LED driver
based on an integrated bridgeless boost–asymmetrical half-bridge converter. IEEE Trans Ind
Appl 51(1):761–769
11. Milashevski I, Galkin I, Tetervenok O (2012) Assessment of buck converter powered by current
or voltage sources for LEDs luminary. In: 13th Biennial Baltic electronics conference
12. Lin RL, Lee CC, Chang YC (2011) Optimal design of LED array for single-loop CCM buck-
boost LED driver. IEEE Ind Appl Soc Annu Meet
13. Shen CL, Wu YE, Tsai CT (2011) Coupled-inductor sepic-type PFC with soft-switching fea-
ture for LED lighting applications. In: 6th IEEE conference on industrial electronics and
applications, Beijing, pp 2384–2389
14. de Britto JR, Demian AE, de Freitas LC, Farias VJ, Coelho EAA, Vieira JB (2008) A proposal
of led lamp driver for universal input using Cuk converter. In: 2008 IEEE power electronics
specialists conference, Rhodes, pp 2640–2644
15. Gacio D, Alonso JM, Calleja AJ, Garcia J, Rico-Secades M (2011) A universal-input
single-stage high power-factor power supply for HB-LEDs based on integrated buck–flyback
converter. IEEE Trans Ind Electron 58(2):589–599
16. Lamar DG, Zuniga JS, Alonso AR, Gonzalez MR, Alvarez MMH (2009) A very simple control
strategy for power factor correctors driving high-brightness LEDs. IEEE Trans Power Electron
24(8):2032–2042
17. Castro I, Martin K, Lamar DG, Arias M, Hernando MM, Sebastian J (2016) Single-stage
AC/DC dual inductor BCM current-fed push-pull for HB-LED lighting applications. IEEE
Energy Convers Congr Expo (ECCE)
18. Lu DDC, Lu HHC, Pjevalica V (2009) Single-stage AC/DC boost–forward converter with high
power factor and regulated bus and output voltages. IEEE Trans Ind Electron 56(6):2128–2132
19. Ye Z, Greenfeld F, Liang Z (2008) Design considerations of a high power factor SEPIC converter
for high brightness white LED lighting applications. In: IEEE power electronics specialists
conference
20. Wang Y, Huang J, Wang W, Xu D (2016) A single-stage single-switch LED driver based on
class-E converter. IEEE Trans Ind Appl 52(3):2618–2626
21. Cheng HL, Lin CW (2014) Design and implementation of a high-power factor LED driver with
zero-voltage switching-on characteristics. IEEE Trans Power Electron 29(9):4949–4958
22. Lin W, Chen H, Ke S (2016) Research on a single-stage flyback/boost LED driver with lower
output ripple. In: 2016 IEEE 2nd annual southern power electronics conference
23. Ding X, Yu D, Song Y, Xue B (2017) Integrated switched coupled-inductor boost-flyback
converter. In: 2017 IEEE energy conversion congress and exposition
Simulation Study of Shading Effects
in PV Array
S. Harika, R. Seyezhai, and A. Jawahar
Abstract The PV panels are interconnected in SP configuration in order to achieve

a higher voltage. The SP configuration of PV string may undergo both objective
and subjective partial shading situations which reduces the generated maximum
output power. Thus, the research work carried out the analysis of PV array rated
1500 W for different shading patterns like row-wise and column-wise partial shading.
The behavior of solar array is also studied under different irradiation levels and
different degrees of shading. The behavior of PV string characteristics for with and
without bypass diode under different shading scenarios has also been analyzed using
MATLAB/SIMULINK. The performance parameter of a solar array such as power
loss, fill factor and a maximum power of the PV string is evaluated and compared.
From the analysis, it is inferred that the generated output power gets deteriorated for
low irradiation level and low degree of shading. The results are verified.
Keywords Partial shading condition · Row-wise shading · Column-wise shading ·

Power loss · Bypass diode
1 Introduction
Nowadays, renewable energy resources have emerged as an attractive solution for

the problem of high electricity demand and serious environmental hazards caused by
the existing conventional resources. In that photovoltaic renewable energy resources
is widely used as it is environmentally friendly and available abundant in nature.
The photovoltaic (PV) panel directly converts solar energy into electrical power. But
for high power applications, the PV panels have to be configured in SP model to
S. Harika (B) · R. Seyezhai

Department of Electrical and Electronics Engineering, REC Lab, Sri Sivasubramania Nadar
(SSN) College of Engineering, Kalavakkam, Chennai, India
A. Jawahar
Department of Electronics and Communication Engineering, Sri Sivasubramania Nadar (SSN)
College of Engineering, Chennai, India

https://doi.org/10.1007/978-981-15-4488-0_55
656 S. Harika et al.
Fig. 1 SP configuration of
1.5 W PV array
increase the output power since the output power of the single PV panel is low [1].
The solar cells or solar panel in the array receives different irradiation and mismatch
between the cells or panels due to shading effects which may affect the generated
output power. The partial shading is due to poor weather conditions, dust, building,
etc. Many researchers and scientists focus to improve the performance of PV even
under partial shading conditions [2]. Thus the paper examines the performance of
1500 W PV array under different (row-wise and column-wise partial shading) partial
shading scenarios through voltage versus current curve and power versus voltage
curve. Bypass diode plays a major role in partial shading which prevents hot spots
and reduces the thermal stress. Thus the impact of the bypass diode on PV array is also
studied using MATLAB/SIMULINK. The SP configuration of 1500 W photovoltaic
array is shown in Fig. 1.
2 Partial Shading
The interconnection of PV in the array will undergo shading due to building, trees,
birds, dust and poor weather condition. The partial shading of solar panel is con-
sidered to be a more critical issue because the shaded cell consumes power from
unshaded cells since it gets reverse bias, this leads to the reduction of output power
from PV array. It will increase the heating and thermal stress and results in hot-spot
[3, 4]. The partial shading of the solar array is characterized through the V-I and
P-V characteristics. There are different types of partial shading, this research work
focuses on row-wise and column-wise partial shading for 1.5 kW PV array.
Simulation Study of Shading Effects in PV Array 657
2.1 Row-Wise Partial Shading
In row-wise partial shading, the last row of the 2 × 3 PV array is fully shaded by
maintaining all other PV panels at full illumination (1000 W/m2 ). If an entire row
gets shaded, then the sum of photonic current is less than the array current which in
turn leads to dissipate the power generated by remaining rows.
2.2 Column-Wise Partial Shading
In column-wise partial shading, the first column of the 2 × 3 PV array is fully shaded
by maintaining all other PV panels at full illumination (1000 W/m2 ). Due to shading,
the amount of voltage required to produce output current is reduced which will further
reduce the array output voltage and increase the power loss.
3 Simulation Results
The PV panel rating of 1.5 kW (V oc = 112.2 V, I sc = 17.26 A, V mp = 92.1 V, I mp

= 16.3 A) is modeled in MATLAB/SIMULINK. The impact of bypass diode on PV
array for both row-wise and column-wise partial shading has been analyzed through
V-I characteristics and P-V characteristics and its performance is evaluated under
different shading scenarios. The performance parameter such as row voltage, open-
circuit voltage (V oc ), short-circuit current (I sc ), maximum power (Pmax ), fill factor
and power loss is evaluated for both with and without bypass diode [5–9] under
different shading level and different illumination is depicted in Tables 1, 4.
Table 1 Row-wise shading with bypass diode

Shading Irradiation Row Row Row V oc (V) I sc (A) Pmax
level (%) (W/m2 ) voltage I voltage II voltage (W)
(V) (V) III (V)
100 1000 37.4 37.4 0 74.8 17.32 975
600 36.63 36.63 0 73.26 10.4 585
300 35.59 35.59 0 71.18 5.2 290
60 1000 37.4 37.4 36.02 110.82 17.32 975
600 36.63 36.63 36.02 109.28 10.4 650
300 35.59 35.59 36.02 107.2 6.93 458
20 1000 37.4 37.4 37.06 111.86 17.32 1280
600 36.63 36.63 37.06 110.32 13.86 928
300 35.59 35.59 37.06 108.24 13.85 470
Table 2 Row-wise shading without bypass diode

Shading Irradiation Row Row Row V oc I sc Pmax
level (%) (W/m2 ) voltage I voltage II voltage (V) (A) (W)
(V) (V) III (V)
100 1000 37.4 37.4 0 74.8 11.4 450
600 36 36 0 72 6.9 175
300 25 25 0 50 3.5 40
60 1000 37.4 37.4 10.2 85 13.7 600
600 36 36 10 82 9.2 300
300 35 35 8 78 5.7 120
20 1000 37.4 37.4 14.2 89 16 750
600 36 36 11 83 11.5 450
300 35 35 10 80 8 230
Table 3 Column-wise shading with bypass diode

Shading Irradiation Row Row Row V oc (V) I sc Pmax
level (%) (W/m2 ) voltage I voltage II voltage (A) (W)
(V) (V) III (V)
100 1000 37.4 37.4 37.4 112.2 8.6 740
600 36.63 36.63 36.63 109.89 5.2 448
300 35.59 35.59 35.59 106.77 2.6 220
60 1000 37.4 37.4 37.4 112.2 12.1 1040
600 36.63 36.63 36.63 109.89 8.66 744
300 35.59 35.59 35.59 106.77 6 515
20 1000 37.4 37.4 37.4 112.2 15.6 1340
600 36.63 36.63 36.63 109.89 12.1 1040
300 35.59 35.59 35.59 106.77 9.52 820
Table 4 Column-wise shading without bypass diode

Shading Irradiation Row Row Row V oc I sc Pmax
level (%) (W/m2 ) voltage I voltage II voltage (V) (A) (W)
(V) (V) III (V)
100 1000 20 30 30 80 8.6 270
600 10 30 30 70 5.1 100
300 10 10 20 40 2.6 25
60 1000 25 30 30 85 12 500
600 22 30 30 82 8.6 270
300 15 20 30 65 6 135
20 1000 28 30 30 88 15.4 715
600 26 30 30 86 12 500
300 22 30 30 82 9.4 320
The analysis of 1500 W PV array under row-wise shading for both with and
without bypass diode model is depicted in Tables 1, 2 in which the maximum power
of 1280 W is achieved for with bypass diode model under 20% shading at 1000 W/m2
whereas for the same condition, the without bypass diode configuration achieves only
750 W. For 60 and 100% shading at 1000 W/m2 , the maximum power of 975 W is
achieved for with bypass diode model whereas for the same condition the without
bypass diode model attains only the maximum power of 600 and 450 W.
The performance of 1500 W PV array under column-wise shading for both with
and without bypass diode model is depicted in Tables 3, 4 in which the maximum
power of 1340 W is achieved for with bypass diode under 20% shading at 1000 W/m2
whereas, for the same scenario, the without bypass diode configuration achieves only
715 W. For the same irradiation level and with 60% shading, the with and without
bypass diode model achieves 1040 and 500 W whereas, with 100% shading, the
maximum power of 740 and 270 W is attained for with bypass diode and without
bypass diode configuration.
The effect of shading on the photovoltaic array is clearly studied and presented
in the table. In addition to that, the pictorial representation of fill factor and power
loss variation for the above-mentioned scenarios is also depicted. The variation in
fill factor under variable irradiation and for the different degrees of shading is shown
in Figs. 2 and 3.
Figure 2 shows, the fill factor impacts for row-wise shading. The fill factor of
0.75, 0.5 and 0.66 is achieved under 100,60 and 20% shading at 1000 W/m2 for
with bypass diode configuration whereas for the same condition the without bypass
diode achieves only 0.46, 0.51 and 0.52. Figure 3 shows the fill factor impacts for
column-wise shading. For 20% shading over the PV array, the with bypass diode
model achieves the fill factor of 0.76 under all degree of shading level at 1000 W/m2
whereas for the same condition the without bypass diode achieves only 0.39, 0.49
and 0.52.
The variation in the power loss under variable irradiation and for the different
degrees of shading is shown in Figs. 4 and 5.
Fig. 2 a Row-wise shading with bypass diode and b row-wise shading without bypass diode
Fig. 3 a Column-wise shading with bypass diode and b column-wise shading without bypass diode
Fig. 4 a Row-wise shading with bypass diode and b row-wise shading without bypass diode
Fig. 5 a Column-wise shading with bypass diode and b column-wise shading without bypass diode
From Fig. 4, it is observed that under 20% shading at 1000 W/m2 irradiation level,
the row-wise shading has a lower power loss of 0.14 W for with bypass diode model
and 0.4 W for without bypass diode configuration. And for the same irradiation level,
the with bypass diode model has a power loss of 0.35 W for 60 and 100% shading
whereas, for the same case, the without bypass diode has 0.56 and 0.6 W.
From Fig. 5, for the same irradiation level, the with bypass diode under column-
wise shading has 0.1, 0.3, and 0.5 W for 20, 60, and 100% shading but the without
bypass diode model has 0.52, 0.66, and 0.8 W for the same scenario. Thus, it is
inferred that the with bypass diode model attains the maximum peak power and
reduces the power loss which makes the system effective. And hence the voltage-
current curve and power-voltage curve of 1.5 kW PV array for with bypass diode
configuration under row-wise and column-wise partial shading for different irra-
diance levels and for the different degrees of shading are depicted in Figs. 6 and
7.
In Fig. 6, shading the last row of the 2 × 3 array to 20% leads to a maximum power
of 1280 W at 1000 W/m2 but if the degree of shading increases, the power drops to
290 W whereas the without bypass diode model reaches the power loss to 40 W as
Fig. 6 a V-I characteristics of row-wise partial shading incorporating bypass diode and b P-V
characteristics of row-wise partial shading incorporating bypass diode
Fig. 7 a V-I characteristics of column-wise partial shading incorporating bypass diode and b P-V
characteristics of column-wise partial shading incorporating bypass diode
the negative voltage is developed. In Fig. 7, 20% shading of the first column in an
array results in maximum power of 1340 W at 1000 W/m2 which is 0.6% higher than
the row-wise shading for the same case. Hence the column-wise shading results in
reduced power loss increased fill factor and reaches the maximum power compared
to row-wise shading. Thus, it is inferred that column-wise shading is better as it has
high efficiency of 89.33% than the row-wise shading.
4 Conclusion
The investigation of PV array under row-wise and column-wise partial shading con-
ditions for different irradiation levels and for the different degrees of shading is
analyzed. From the analysis, it can be observed that operating the PV array with a
bypass diode under 20% of shading at full illumination (1000 W/m2 ) has a better
performance compared to the former method. The proposed shading method gener-
ates high peak power with reduced power loss and attains a higher fill factor value of
0.7 whereas row-wise shading reaches only 1280 W peak power. Thus the proposed
column-wise shading has higher fill factor, high peak power, and reduced power loss
compared to row-wise shading.
References
1. Vemuru S, Singh P, Niamat M (2012) Modeling impact of bypass diodes on photovoltaic cell
performance under partial shading. IEEE Trans
2. Danish S, Mehta G, Yadav VK (2017) Assessing the performance of partially shaded photovoltaic
array under different radiations. In: 8th ICCCNT 2017 3–5 July
3. Alajmi BN, Ahmed KH, Fenney SJ, Williams BW (2013) A maximum power point tracking for
partially shaded photovoltaic system in micro grid. IEEE Trans Ind Electron 60(4):1596–1606
4. Batzelis E, Papathanassiou S (2016) An algorithm to detect partial shading conditions in a PV
system. Mater Sci Forum 856: 303–308
5. Hu Y, Cao W, Wu J, Ji B, Holliday D (2014) Thermography-based virtual MPPT scheme for
improving PV energy efficiency under partial shading conditions. IEEE Trans Power Electron
29
6. Veerapen S, Wen H, Du Y (2017) Design of a novel MPPT algorithm based on the two-
stage searching method for PV systems under partial shading. In: IEEE 3rd international future
energy electronics conference and ECCE Asia (IFEEC 2017–ECCE Asia), pp 1494–1498
7. Shinohara T, Uno M (2017) Micro-inverter based on quasi-Z-source inverter integrating switch-
less voltage equalizer for photovoltaic panels under partial shading. In: IEEE 3rd international
future energy electronics conference and ECCE Asia (IFEEC 2017–ECCE Asia), pp 2013–2018
8. Moballegh S, Jiang J (2014) Modeling, prediction, and experimental validations of power peaks
of PV arrays under partial shading conditions. IEEE Trans Sustain Energy 5
9. Ahmed J, Salam Z (2015) An improved method to predict the position of maximum power point
during partial shading for PV arrays. IEEE Trans Ind Inform 11
Speech Recognition Using Neural
Network for Mobile Robot Navigation
Prashant Patel, Arockia Selvakumar Arockia Doss, L. PavanKalyan,

and Parag J. Tarwadi
Abstract Automatic speech recognition (ASR) has gained a lot of popularity in the
mobile robotics, where the commands could be provided to the robot wirelessly to
maneuver. A navigation system combined with ASR is a complex system to carry
out, because the system has difficulty in recognizing the voice commands when
the environment involved already has disturbances like road noise, air conditioner,
music, and passengers. The objective of this research is to operate a mobile robot with
a single-arm manipulator, where the robot can perceive the speech and it can react
to the individual speech commands provided by the operator swiftly and precisely.
In order to recognize the speech, mel-frequency cepstral coefficient (MFCC) speech
recognition algorithm is chosen and implemented in MATLAB. Various training and
testing have been done in MFCC algorithm where it has to carry out the real-time
processing of speech data and respond to it. Based on both the training and testing
the voice commands collected from the five test subjects both male and female, the
speech recognition system achieved 89% efficiency for the test database.
Keywords Automatic speech recognition · Navigation · Mobile robot MFCC ·

MATLAB
1 Introduction
Speech is a characteristic wonder that happens each and every day. From an all-
around early point in our lives, we become familiar with the vital abilities to utilize
speech as an essential method of correspondence. Since speech comes so normally
to us, it is anything but difficult to overlook how complex it truly is. It starts with
P. Patel · A. S. Arockia Doss (B) · L. PavanKalyan

School of Mechanical Engineering, Vellore Institute of Technology, 600127 Chennai, Tamil Nadu,
India
P. J. Tarwadi
University of Auckland, Auckland, New Zealand

https://doi.org/10.1007/978-981-15-4488-0_56
666 P. Patel et al.
lungs creating wind stream and gaseous tension. The vocal tract, comprising of the
tongue, sense of taste, cheek and lips at that point well-spoken and channel the sound
[1].
In the present society, individuals lean toward alternatives in innovation that
improve exercises. Utilizing speech acknowledgment, clients could type content
or issue gadget directions through speech. Speech acknowledgment in PCs is not
even close to the speech acknowledgment capacities of the human brain. Be that
as it may, maybe utilizing frameworks that imitate human brain capacities could
prompt further progressions in the field. Cooperative robotics is a challenging area
where the above-mentioned operations need to be performed by the robot under the
supervision of an operator or user using speech as a mode of communication with
the robot over a distance. Currently, cobots are growing in-demand in various fields
like service, rehabilitation, search, rescue, and medical field robotics. Integration
and implementation of speech recognition for the cooperative robotics are complex
and the recognition of speech by the robot is difficult in the unknown environment
which is influenced by many parameters like noise and disturbances, where the per-
formance of the robot is quite unpredictable [2]. In speech recognition, following
methods were proposed to extract and match the features to find optimized results:
microcontroller with MAX-32 [3], MATLAB with Gaussian noise and Butterworth
filter [4], hidden Markov model (HMM) with customized algorithm [5], speech-and-
speaker (SAS) identification system [6], fast Fourier transform [7], hybrid artificial
neural networks (ANN) [8], short utterance speaker recognition (SUSR) [9], ANN
with back-propagation algorithm [10], and MFCC with vector quantization (VQ)
[11]. Similarly, an overview of speech recognition techniques is summarized with
merits and demerits [12].
In digital signal processing, the Mel-frequency cepstral coefficient (MFCC) and
dynamic time warping (DTW) are used to extract features and compare the complex
voice signal patterns [13]. A comparative performance study between linear dis-
criminative analyzed and Mel-frequency cepstral coefficient was carried out based
on feature extraction techniques. Based on the results obtained, it was found that
MFCC performed better in extracting features whereas LDA showed better results in
reducing the dimension of the extracted feature [14]. MFCC was found to be one of
the best feature extraction techniques for Quranic [15] and Marathi speech recogni-
tion [16]. Using MATLAB, text-dependent speaker recognition was developed and
MFCC was used to extract features from the speaker’s signal [17].
Based on the literature survey, many researchers have carried out speech recogni-
tion techniques for various applications. The aim of this research is to find the highest
speech recognition rate using neural network method and implement the speech effi-
ciently in mobile robot where the noise level is considerably high and also adapts to
the environment. Further, the work is focused to use MATLAB for an ANN model
testing, training, and validation of input data used for speech recognition command
from database to find the highest speech recognition rate.
Speech Recognition Using Neural Network for Mobile … 667
Fig. 1 ANN training and

testing process
2 Feature Extraction
Figure 1 shows the proposed methodology of the research work. The MATLAB is
utilized to record sound through PC receiver in a sound card. Once the wav documents
are recorded, a library called Auditory Toolbox is connected with MATLAB. This
library contained the quick Fourier change calculation that had chosen to use so as
to retrieve one of the kind highlights from the input sound samples.
In acoustic signals, identifying and extracting the best output signals are primary
importance because the performance and the efficiency of the robot is dependent
on the obtained signal in this phase. Two types of filters are used in this recogni-
tion process to receive low frequency and high-frequency signals above and below
1000 Hz.
3 Training Data Preparation
Hence, to execute speech recognition with neural system, the sound documents are
recorded, investigated, and controlled.
clc
clear all
close all
recObj = audiorecorder;
disp(‘Start speaking.’)
recordblocking(recObj, 5);
disp(‘End of Recording.’);
668 P. Patel et al.
play(recObj);
myRecording = getaudiodata(recObj);
plot(myRecording);
Ft = fft(myRecording);
figure,
plot(1:numel(Ft),Ft);
Utilizing MATLAB characterized a sampling frequency called Fs and a term
called duration. At that point, allot another wav document to a variable utilizing the
wav record function. After all expected to physically trim the sound records so as
to dispose of void space where no speech was spoken, so as to improve my system
error rates. Utilizing this recording technique, four distinct examples of the words
are recorded, for example, “Start, Stop, Pick, and place.” Five examples of each
word would be put aside to test the neural system after it was assembled; the other
five examples of each word were to be utilized as preparing information. Utilizing
the Auditory Toolbox work called MFCC where, ready to recover the best possible
coefficients that could be utilized for the proposed speech recognition technique.
Figures 2, 3, 4, 5, 6, 7, 8 and 9 show the feature extraction of the speech from the
test subjects. The last part of the arrangement before making the neural system was
to make the training matrix and target matrix. I shaped a framework of size 130 ×
20, which is 130 highlights for 20 words, by taking the MFCC vector of each word
and combining them into one matrix.
Fig. 2 Wav file of “start”

Fig. 3 FFT of “start”
Fig. 4 Wav file of “stop”

670 P. Patel et al.
Fig. 5 FFT of “stop”
Fig. 6 Wav file of “pick”

Fig. 7 FFT of “pick”
Fig. 8 Wav file of “place”

672 P. Patel et al.
Fig. 9 FFT of “place”
4 Neural Networks in Speech Recognition
A multilayer perceptron neural network is used for speech recognition. Feed-forward

systems [7] use many number of hidden layers of sigmoid neurons followed by an
output layer of linear neurons. Figure 10 outlines the structure of the neural system in
this project. The highlights can be the LPC coefficients or the initial three formants of
each frame. The entire database comprises of: (i) five different speakers speaking at
different rates; (ii) four words, for each speaker; and (iii) total number of utterances:
5 × 4 = 20 utterances. Table 1 shows details of the neural networks performance
with numerous hidden neurons.
About 70% of the database is used for training the multilayer neural network and
115 the rest is used for testing.
Table 1 Neural networks performance with numerous hidden neurons

S. No. ANN parameters Values
1 Learning parameter 0.22
2 Nonlinear activation function Tan-sigmoid
3 Maximum epoch 1000
4 Number of hidden layer 1
5 No. of nodes in hidden layer 10–80
6 Error goal 0.001
7 Target node 4
Fig. 10 Neural network architecture for speech recognition
5 Design of the Mobile Robot
The robot which is used for speech recognition is for “pick and place” mobile robot
with a soft catching gripper for safe handling of the sensitive objects as shown in
Fig. 11. App-based speech recognition is used for remote operation of the robot,
to follow the commands sent by the user like forward, backward, and left or right.
Bluetooth module is used for operating the robot remotely to drive motors.
Fig. 11 Mobile robot along single-arm manipulator

674 P. Patel et al.
Table 2 Speech rate using ANN recognition

Hidden nodes Start Stop Pick Place
10 84.1 84.1 88.5 81.0
20 90.3 92.9 88.5 90.3
40 93.5 92.9 93.8 93.8
80 93.8 97.3 88.5 80.5
Mean 90.4 91.8 89.82 86.4
6 Results
Speech recognition using neural network for mobile robot navigation is developed,
and based on the findings, the results are discussed in this section. When utilizing a
feed-forward multilayer perceptron, the Levenberg–Marquardt is used for preparing
calculation to characterize the words “Start,” “Stop,” “Pick,” and “Place” based on
their MFCC, and the test words are perceived without any error. These outcomes
are very positive and exhibit the capabilities of neural networks when it comes to
pattern recognition. Table 2 shows the various speech rate values obtained using
ANN recognition.
7 Summary
The speech recognition using neural network for mobile robot navigation is devel-
oped and tested, based on the methods proposed and various findings, and the
following points are summarized:
This research sets on actualizing a speech recognition framework with an artificial
neural system. MFCC algorithm is implemented for the speech recognition.
The collected voice commands from five test subjects include both male and
female in our databases. But this database is used in both training and testing purpose
after preprocessing this speech commands. Finally, we get a recognition system with
89% efficiency. The recognition can further be improved by increasing the number of
hidden neurons. Maximum of 92% recognition was obtained which can still further
be increased by using other recognition methods such as neuro-fuzzy and fuzzy
methods. The number of hidden layer can also be altered to notice the behavior of
the recognition system. While the scope of the system was reduced to an isolated
word recognition network, the outcomes were still positive.
The final system provides a user-friendly mobile robot which could be remotely
operated by the pre-programmed voice commands by the user. This approach sup-
ports the fact that the key to successfully demonstrate the cooperation between the
robot and user in a very unlike environment with a lot of disturbances and still
implement the received command swiftly and precisely.
It clarifies how neural systems can handle an issue like pattern recognition, as
well as the benefits of certain structures and training algorithms.
In future, more words can be used for speech recognition, and it can be imple-
mented for mobile robot for the execution of task where the robot can execute the
task by speech recognition from various speakers on different speech.
8 Compliance with Ethical Standards
All procedures performed in studies involving human participants were in accor-

dance with the ethical standards of the institutional research committee and with the
1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed consent was obtained from all individual participants included in the study.
References
1. Campbell JP (1997) Speaker recognition: a tutorial. Proc IEEE 85(9):1437–1462

2. Bort L, del Pobil AP (2000, October) Using speech to guide a mobile robot manipulator. In
Smc 2000 Conference Proceedings. 2000 IEEE International Conference on Systems, Man
and Cybernetics. ‘cybernetics evolving to systems, humans, organizations, and their complex
interactions’ (cat. no. 0, vol 4, pp 2356–2361). IEEE
3. Neerja SD, Rupesh SM (2013) Robotic automation using speech recognition
4. Saxena A, Sinha AK, Chakrawarti S, Charu S (2013) Speech recognition using matlab. Suresh
Gyan Vihar University, Jaipur, Rajasthan, India
5. Kim NS, Un CK (1995) On estimating robust probability distribution in HMM-based speech
recognition. IEEE Trans Speech Audio Process 3(4):279–285
6. Chandra E, Sunitha C (2009, March) A review on speech and speaker authentication sys-
tem using voice signal feature selection and extraction. In 2009 IEEE International Advance
Computing Conference IACC 2009, pp 1341–1346
7. Harris FJ (1978) On the use of windows for harmonic analysis with the discrete Fourier
transform. Proc IEEE 66(1):51–83
8. Zebulum RS, Vellasco M, Perelmuter G, Pacheco MA (1996) A comparison of different spectral
analysis models for speech recognition using neural networks. IEEE
9. Fatima N, Zheng TF (2012, May) Short utterance speaker recognition a research agenda. In
2012 International Conference on Systems and Informatics (ICSAI), pp 1746–1750
10. Shukla A, Tiwari R (2008) A novel approach of speaker authentication by fusion of speech
and image features using artificial neural networks. Int J Inf Commun Technol 1(2):159–170
11. Chakraborty P, Ahmed F, Kabir MM, Shahjahan M, Murase K (2008) An automatic speaker
recognition system. In: Ishikawa M et al. (eds) ICONIP 2007, Part I, LNCS 4984, Springer-
Verlag, Berlin Heidelberg, pp 517–526
12. Himanshu M, Kaur SS, Sharma A (2014) Real-time information system based on speech
recognition
13. Muda L, Begam M, Elamvazuthi I (2010) Voice recognition algorithms using mel frequency
cepstral coefficient (MFCC) and dynamic time warping (DTW) techniques. arXiv preprint
arXiv:1003.4083
14. Gaikwad S, Gawali B, Yannawar P, Mehrotra S (2011, December) Feature extraction using
fusion MFCC for continuous marathi speech recognition. In 2011 Annual IEEE India
Conference, Hyderabad, pp 1–5
676 P. Patel et al.
15. Razak Z, Ibrahim NJ, Tamil EM, Idris MYI (2012) Quarnic verse recitation feature extraction
using mel-frequency cepstral coefficient (MFCC). Department of Al-Quran & Al- Hadith,
Academy of Islamic Studies, University of Malaya
16. Gaikwad SK, Gawali BW, Yannawar P (2010) A review on speech recognition technique. Int
J Comput Appl 10(3):16–24
17. Kabir A, Ahsan SMM (2007, December) Vector quantization in text dependent automatic
speaker recognition using mel-frequency cepstrum coefficient. In 6th WSEAS International
Conference on Circuits, Systems, Electronics, Control & Signal Processing, Cairo, Egypt, pp
352–355
Range Sensor-Based Obstacle Avoidance
of a Hyper-Redundant Robot
G. Mohammed Nawaz and N. Karthikeyan
Abstract In recent times, the development of the bio-mimicking is increasing in

the robotics field. This paper discusses the design of the hyper-redundant robot
and manoeuvring in the environment. The robot works on the Coulomb’s friction
law where the friction between robot and surface helps to move the robot. The
robot models are fabricated by 3D printing. The robot uses passive wheels for its
locomotion. A range sensor is used to detect the obstacles in the environment. A
Bluetooth module is utilized for the wireless control of the robot. This paper describes
the design of the hyper-redundant robot to avoid the obstacles in the environment
using a range sensor and the robot follows the serpentine motion for the locomotion.
Keywords Bio-mimicking · Hyper-redundant · Passive wheels · Range sensing ·

Obstacle avoidance
1 Introduction
The bio-mimicking robots have been increasing, and the hyper-redundant robots are
used for various operations like rescue of human in the trapped region and investiga-
tion of the narrow path regions. The hyper-redundant robot is referred to as the robot
which has the greater DOF than the minimum DOF. The robot should be capable of
avoiding the obstacle in the environment and find the best path to avoid the obstacle.
The range sensor is capable of identifying the obstacle in the environment.
The robot is moved by the structure of a genuine snake. Be that as it may, there is
an extensive scale between the headway proficient of snake-like robot and genuine
snake. The snake can move in off-road, swim in the water, climb tree and even it
coast noticeable all around by a portion of the snake species. Snake robot uses the
scattered objects in condition for obstruction evasion. Manufacturing a snake robot
G. Mohammed Nawaz (B) · N. Karthikeyan

SRM Institute of Science and Technology, Kattankulathur, Chennai, Tamil Nadu 603203, India
N. Karthikeyan
https://doi.org/10.1007/978-981-15-4488-0_57
678 G. Mohammed Nawaz and N. Karthikeyan
with such action will be an alluring one. But implementing in the real time will be a
difficult task and it can be achievable. Development of such robots will motivate in
various applications like testing constant operation, hazardous areas, examining the
pipe structure, search and rescue.
The component of the snake robot has sequentially associated joints that have the
capacity to move in different planes. High level of opportunity of snake robot influ-
ences hard to control however to give productive headway aptitudes in scattered and
unpredictable condition which superior to the portability of all the more overarching
haggle robot.
There are a few sorts of headway for snake robots, where the development is
conveyed by wheels or legs on their body. A portion of the robot utilizes passive
wheels to create anisotropic grinding for headway. The passive wheels are highly
useful to the robot for faster reachability to the target.
The snake uses different types of locomotion suitable for the terrain and for the
better manoeuvrability of snake in the plane. The various methods of locomotion of
snakes are.
• Concertina
• Serpentine
• Sidewinding
• Rectilinear.
The robot moves in a particular stride, which is serpentine motion, like a lateral
undulation motion. The serpentine motion is similar to sine curve, and it has some
advantages over other locomotion (Fig. 1).
Ryo and Fumitoshi [1] analysed the dynamics of three snake locomotion gaits
where the gaits are sidewinding motion, serpentine motion and sinus-lifting motion.
They have done a kinematic model of a snake robot and done the numerical analysis
for the three walks. Furthermore, they consider the examination in just planar motion
with typical powers following up on the grounded joints. And they simulated in
MATLAB to get efficient results to amount the three gaits.
Motoyasu and Kazuo [2] proposed a shape control technique to keep up the head
and relative position of the snake robot and maintain a strategic distance from as far
as possible and self-impact. They created an inactive-wheeled snake robot that can
switch the grounded and lifted the status of its wheels. They contrasted the simulation
Fig. 1 Sine curve

Range Sensor-Based Obstacle Avoidance of a Hyper-Redundant Robot 679
with experimental results which were utilized to assess the viability of the proposed
technique.
Motoyasu and Kazuo [3] proposed control method for snake robot for ascending
and descending steps. Phase shift method with shifting conditions is used to ascend
and descend in steps. Shifting conditions derived for four process which are shift-
ing on high step, before a shift, during shifting and after a shift. Efficient shifting
conditions are derived from simulations using MATLAB.
Motoyasu and Kazuo [4] broke down the conditions for singular configura-
tions with unconstrained links and lemmas for a snake robot. Kinematic model is
determined with snake robot having passive wheels with dynamics joints. Essen-
tial and adequate conditions for the singular configurations of the snake robot with
constrained links are assessed and got viability in recreation utilizing MATLAB.
Motoyasu et al. [5] designed a snake robot and control framework in light of range
sensor information that control semi-autonomously to help the robot in staying away
from hindrances. The administrator in the control structure exhibits the coveted speed
of the first connection, and effect avoidance between coming about connections and
obstructions is consequently figured by the controller. This chooses the connections
should have been grounded and misuses excess.
Motoyasu et al. [6] proposed a control method for semi-self-ruling advance mov-
ing by a snake robot. The control system relies upon blended whole number quadratic
programming to make the reference course of the leader of the robot. It chooses the
proper positions and time term to identify nature before moving towards the move-
ment. This control technique was connected to a snake robot furnished with a laser
extend discoverer which is used for the progression area.
Motoyasu et al. [7] proposed a control method for all joints of the snake robot
by comparing simulation and experimental results. The snake robot is composed of
inactive wheels with dynamic yaw and move joints. Controller for approximate path
tracking is composed utilizing the kinematics redundancy and choosing of the proper
lifted parts.
Xiaodong and Shugen [8] concentrated on the outline of controller for a given
motion pattern. Snake robot is equipped with laser range sensor to give sensory inputs
to the control system for collision avoidance using neural controller-based central
pattern generator.
Liljeback et al. [9] proposed a control technique that joints environment adaptation
and directional control to accomplish straight line path with hindrance. Snake robot
is designed to propel using the body friction by undulating motion. Tactile sensor is
equipped in snake robot for contact type obstruction evasion.
Liljeback et al. [10] proposed a control technique for path following in straight
path. Cascade approach technique is used to control the robot to go along straight
path. Here, presumptions have been made as the forward velocity of the robot is
nonzero and positive. Path way following controller exponentially balances out the
robot to a different allotted straight path.
Liljeback et al. [11] proposed two control methods namely jam-detection scheme
and leader–follower scheme for obstacle avoidance. The snake robot is designed to
propel using the body friction. Jam-detection scheme will be undergone when the
robot stuck between obstacles. Leader–follower scheme makes the sub-links of the
robot to follow the head.
Liljeback et al. [12] investigated the properties of snake-like robot dynamics by
employing nonlinear system tools analysis. Five investigations have been done which
are partially feedback linearized model of a planar snake robot, stability analysis,
controllability analysis, mapping link velocities and Poincare map.
Liljeback et al. [13] focused on obstacle avoidance by solving linear complimen-
tary problem (LCP). Hybrid control procedure is executed to quantify contact forces
that keep up the propulsion and keep robot from being jammed between hindrances.
Contact forces were figured by formulating and solving a LCP.
Aksel et al. [14] developed a 3D kinematic model of a snake robot which propul-
sion is conveyed friction forces in light of coulomb’s law of dry friction. Thee
model is actualized for numerical treatment with numerical integrator called the time-
stepping method. Simulation is completed for serpentine movement with sensible
effectiveness.
Aksel et al. [15] developed a hybrid mathematical model for wheel-less snake
robot. Ultrasonic sensor is equipped for obstacle avoidance.
Peipei et al. [16] designed a planar snake robot with bottom wheeled modular
parts. Serpentine locomotion may be slipping problem. New angular parameters are
introduced to compensate the slipping. Simulations and experimental results for the
new angular parameter gives the effective compensation for slipping.
Chao et al. [17] designed a snake robot with universal joint. Snake robot dexter-
ity, complexities and various motion patterns are improved by the universal joint.
Inchworm locomotion gait is adopted in this robot.
Billah et al. [18] developed a smart material actuated autonomous snake robot. It
increases the flexibility of the snake robot. CPG control method helps in autonomous
navigation.
2 Introduction to Central Pattern Generator
Central pattern generators (CPG) are biological neural systems that deliver rhythmic
output without musical input. They are wellspring of the firmly coupled example of
neural action that drives musical movements like walking and breathing. Essentially,
the capacity to work without contribution from higher cerebrum regions does not
imply that CPGs do not get modulatory inputs, or that their outputs are fixed. CPGs
have been found in essentially all vertebrate species including humans (Fig. 2).
Fig. 2 Central pattern

generator [7]
3 Working Methodology
The route towards affecting the hyper-redundant robot to have particular steps of
process starts from arranging, demonstrating, produce and controlling the hyper-
redundant. The underlying advance is to recognize the issue. The second means are
to dissecting the issues and build up a model. The subsequent stage is to build up
a 3D model of parts utilizing planning programming. In the wake of demonstrating
the parts, choice of materials for manufacture happens.
The snake robot has three noteworthy area which are mechanical, electrical and
program. The program which controls the robot by programming the servo engine
with PWM pins which enables robot to move required movement. The sine wave
can be created utilizing PWM sticks by giving positive half cycle to one stick and
negative half cycle for another stick. The sine wave deals with an advanced voltage
in the microcontroller.
3.1 Mechanical Specification
See Table 1.
Table 1 Mechanical
Specification Details
specification
No. of joints 12
Link dimensions (mm) 80 × 60 × 40
Link weight (g) 200
Joint range (°) [0–180]
Actuators Dual shaft servo
Range sensor Ultrasonic sensor
3.2 Process Diagram
See Fig. 3.
4 Designing CAD Models
The following CAD module is the head, connecting frame and tail of the robot which
is a customized design need to be fabricated using 3D printing. Each link is connected
with the frame in the robot. The head and tail are the places where the controller and
power supply are placed in the robot (Figs. 4, 5 and 6).
5 Fabricated Robot
The following are fabricated by 3D printing and acrylonitrile butadiene styrene (ABS)
is the material which is used for the 3D printing of the models. The servo motor
is attached in the series manner in order to perform the serpentine motion. The
microcontroller is placed in the head and power source at the tail. The passive wheels
are attached in each frame of the robot in order to stabilize the robot (Fig. 7).
6 Control Interface
The control interface is separated into three sections: user, wireless data transfer and
the hyper-redundant. The user is used to give signal for on and off of the snake robot
using Bluetooth module. The ultrasonic sensor is used to detect the obstacle in the
environment as depicted in Fig. 8.
The hyper-redundant robot comprises of 12 servo engines and AT mega 2560 as
its microcontroller. All the servo engines are provided with a 7.4 V and 2000 mAh
battery to control up and the signs from controller.
Fig. 3 Process diagram
7 Locomotion
The robot which uses serpentine motion for its locomotion. The serpentine motion
is which is similar to the sine curve. The servo motor is made into 90° and the
connecting frame is attached between the servos motor. Likewise, 12 servo motors
are attached in the series manner to perform serpentine motion on the plain. The
serpentine motion can be achieved by the following formula
Fig. 4 Head
Fig. 5 Tail
Fig. 6 Connecting frame

Fig. 7 Fabricated robot
Fig. 8 Interfacing block diagram
Serial.write(90 + amp ∗ cos( f r eq ∗ count ∗ (π/180) − n ∗ lag)) (1)
where
count = Loop counter variable.
lag = Phase lag between segments.
freq = Oscillation frequency of segments.
amp = Amplitude of the serpentine motion of the snake.
n = Number of segment estimated from 1 to 12.
(π /180) = from degree to radiant.
Table 2 Testing of
Trails 1 m in seconds (s) Speed (kmph)
serpentine motion
1 4.50 0.8
2 4.26 0.845
3 4.40 0.818
4 4.20 0.857
Equation (1) is used for serpentine motion in the robot by dumping the codes in
the AT mega microcontroller with sensor shield for providing more digital pins.
8 Results
8.1 Serpentine Motion
The fabricated prototype of the designed hyper-redundant is tested in the real-time

environment for the serpentine motion of the robot. The 12 servos are connected to
microcontroller and the power supply is given by the lithium-ion battery placed at
the tail of the robot. Equation (1) is dumped into the microcontroller Arduino mega
using embedded c language. Equation (1) tuned to move the robot in a straight path
with serpentine motion. The experiment is done for many trails to find average speed
of the robot.
In Table 2, the robot 1 m at an average time of 4.34 s and concludes that designed
hyper-redundant Robot with passive wheels maximum speed of 0.830 (kmph) in
serpentine motion.
Figure 9 is the representation of the robot with serpentine motion in the plain
terrain. The pictures are taken per seconds of the robot motion. For each second, one
curve completes and forms a new curve.
8.2 Obstacle Avoidance
The obstacle avoidance of the hyper-redundant robot is performed with the help
of the ultrasonic sensor. The robot with the obstacle avoidance algorithm which is
explained in Fig. 10. The ultrasonic sensor is fixed in the head of the robot and the
Bluetooth module is attached in the robot. The user communicates with the robot
Fig. 9 Serpentine motion in robot
Fig. 10 Obstacle avoidance algorithm

with Bluetooth wireless transfer of data. The user only gives command to start or
stop the robot.
Algorithm
Figure 10 is the pictorial representation of the obstacle avoidance of the designed
hyper-redundant robot and the algorithm will be in the loop command so that it can
able to detect the obstacle in the environment for every time.
In Fig. 11, the box in the picture acts as obstacle in the environment once the ultra-
sonic detects the obstacle in the environment and the obstacle avoidance algorithm
will be processed to avoid the collision with the obstacle.
Fig. 11 Obstacle avoidance of the robot

9 Conclusion
This paper presents the design of the hyper-redundant and to get the maximum speed
in the serpentine motion. The obstacle avoidance of the designed robot is done with
the help of the range sensor. Experimental results demonstrate the obstacle avoidance
of the robot in the real-time environment with serpentine motion as its locomotion.
The experimental results for the obstacle avoidance in each trails shows that the robot
performs without hitting the objects.
The future works in providing the live telecast of the robot wherever it goes and
also surveillance of the environment with help of Wi-Fi camera.
References
1. Ryo A, Fumitoshi M (2017) Dynamic analysis of three snake robot gaits. IEEE Trans Robot
33(5)
2. Motoyasu T, Kazuo T (2017) Shape control of a snake robot with joint limit and self-collision
avoidance. IEEE Trans Control Syst Technol 25(4)
3. Motoyasu T, Kazuo T (2015) Control of a snake robot for ascending and descending steps.
IEEE Trans Robot 31(2)
4. Motoyasu T, Kazuo T (2016) Singularity analysis of a snake robot and an articulated mobile
robot with unconstrained links. IEEE Trans Control Syst Technol 24(6)
5. Motoyasu T, Kazuyuki K, Kazuo T (2015) Range-sensor-based semiautonomous whole-body
collision avoidance of a snake robot. IEEE Trans Control Syst Technol 23(5)
6. Kazuyuki K, Motoyasu T, Kazuo T (2016) Mixed integer programming-based semiautonomous
step climbing of a snake robot considering sensing strategy. IEEE Trans Control Syst Technol
24(1)
7. Motoyasu T, Kazuo T, Fumitoshi M (2015) Approximate path-tracking control of snake robot
joints with switching constraints. IEEE/ASME Trans Mechatron 20(4)
8. Xiaodong W, Shugen M (2013) Neurally controlled steering for collision-free behavior of a
snake robot. IEEE Trans Control Syst Technol 21(6)
9. Liljeback P, Pettersen KY, Stavdahl Ø, Gravdahl JT (2012) Snake robot locomotion in
environments with obstacles. IEEE/ASME Trans Mechatron 17(6)
10. Liljeback P, Haugstuen IU, Pettersen KY (2012) Path following control of planar snake robots
using a cascaded approach. IEEE Trans Control Syst Technol 20(1)
11. Liljeback P, Pettersen KY, Stavdahl Ø, Gravdahl JT (2011) Experimental investigation of
obstacle-aided locomotion with a snake robot. IEEE Trans Robot 27(4)
12. Liljeback P, Pettersen KY, Stavdahl Ø, Gravdahl JT (2011) Controllability and stability analysis
of planar snake robot locomotion. IEEE Trans Autom Control 56(6)
13. Liljeback P, Pettersen KY, Stavdahl Ø, Gravdahl JT (2010) Hybrid modelling and control of
obstacle-aided snake robot locomotion. IEEE Trans Robot 26(5)
14. Aksel AT, Leine RI, Glocker C, Pettersen KY (2008) 3-D snake robot motion: nonsmooth
modeling, simulations, and experiments. IEEE Trans Robot 24(2)
15. Aksel AT, Leine RI, Glocker C, Pettersen KY, Liljebäck P (2008) Snake robot obstacle-aided
locomotion: modeling, simulations, and experiments. IEEE Trans Robot 24(1)
16. Peipei S, Qianjun S, Dongtai L (2016) Design and improved serpentine curve locomotion
control of a planar modular snake robot. In: IEEE international conference on information and
automation (ICIA), Feb 2017
17. Chao W, Deng H-B, Xia F, Li Y (2016) Study on climbing locomotion mechanism of snake
robot with universal unit. In: 8th international conference on modelling, identification and
control, Nov 2016
18. Billah MM, Khan MR, Shafie AA (2015) Autonomous flexible snake robot for 3D motion. In:
IEEE international symposium on robotics and intelligent sensors (IRIS), Oct 2015
Literature Survey on Four-Legged
Robots
G. Satheesh Kumar, S. Aravind, R. Subramanian,

Shashank Kannan Bharadwaj, Ronak R. Muthuraman, R. Steve Mitchell,
Aditya Bucha, M. Sriram, K. J. Shri Hari, and Nikhil Jerome Robin
Abstract Over the last two decades, the research and development of legged loco-
motion robots has grown steadily. Legged systems present major advantages when
compared with ‘traditional’ vehicles, because they allow locomotion in inaccessible
terrain to vehicles with wheels and tracks. However, the robustness of legged robots,
and especially their energy consumption, among other aspects, still lag behind mech-
anisms that use wheels and tracks. Legs are not new to humans or animals but building
legs for a robot is a complex process. The normally noticed and ignored fact is how
a baby learns to walk and the sheer learning curve involved. If we, the intelligent
humans take years to learn to walk, imagine creating legs for a robot and teaching it
how to walk. Although there is an extensive research going on in the field of legged
robots, researchers are still in developing stage to construct a legged robot which can
replicate human walk, or for that matter any animals. And a few of their current area
of research are in maintaining dynamic stability of the robot, reflex of the robot to
sudden impacts and also interaction with the terrain to produce suitable gait. This
paper gives a brief description on a few of the existing legged robots on various
aspects of a robot and on how each one of them affects the dynamics of the robot.
This paper enhances a space for future research on understanding the various aspects
of a legged bot design.
Keywords Legged robots · Mechanism · Terrain · Stability · Dynamics
1 Introduction
Legged robots are today used in a variety of applications including areas of rescue,
excavation, research operations and places that are inaccessible by humans. This
paper shows the different mechanisms used in legged robots. These legged robot
G. Satheesh Kumar (B) · S. Aravind · R. Subramanian · S. K. Bharadwaj · R. R. Muthuraman ·

R. Steve Mitchell · A. Bucha · M. Sriram · K. J. Shri Hari · N. J. Robin
Sri Sivasubramaniya Nadar (SSN) College of Engineering, Chennai, India

https://doi.org/10.1007/978-981-15-4488-0_58
692 G. Satheesh Kumar et al.
locomotion mechanisms are generally inspired by the biological system and its envi-
ronment—the animals, insects and other legged creatures including humans give us
a basic idea of how legged robots have to be designed for its motion. These designs
have to be made keeping in mind about factors like stability, power consumption and
complexity of control of each added leg. With the increase in number of legs, the
stability of the robot increases but each leg requires at least two degrees of freedom,
one for forward movement and other for lifting, and therefore, its power consumption
increases. For a two-legged bot, there are two legs and four degrees of freedom. This
makes the robot to have a more versatile motion. But with more legs, controllability
is an issue. If there are n legs of a robot, then the number of possible motions is given
by [1]:
X = (2n − 1)!; by substituting n = 4, we get X = 5040
X is the number of ways in which a four-legged robot can be controlled. Thus, a

robot with more legs becomes more complex due to their controllability, whereas
decrease in number of legs like a bipedal robot or a one-legged robot reduces the
power consumption but decreases the stability of the robot. So generally, four-legged
robots are widely used and various locomotion, mechanisms are researched and are
discussed below. The major difference between legged and wheel robots are the legs
which compromise on stability to enhance terrain overcoming features [2]. Though
stability is compromised, different mechanisms can be used to enhance stability.
Therefore, in the present state of development, there are several aspects that need
to be improved and optimized. The factors in consideration while designing robotic
systems on how they operate include gait mechanisms, robot control system, size,
structural heterogeneity, power sources, actuators and its sensors used.
2 Mechanisms of Legged Robots
Different mechanisms can be used to enhance the stability of a robot. Some of

the mechanisms are parallel leg mechanism, pantograph mechanism, Jansen link-
age, Klann linkage, strider linkage and so on. Different robots employ different
mechanisms customized to meet their requirements [3, 4].
2.1 Big Dog
The big dog robot is a combination of hydraulic actuator and quadruped mechanism
as a whole. Each leg adopted a linkage mechanism, an elastic mechanism and a
spring shock absorber so as to improve the flexibility and terrain overcoming ability.
The hind legs use the five-bar mechanism. The robot’s body is a structure which
is cross-linked. The legs can be seen as institutions, the centre of the two links are
Literature Survey on Four-Legged Robots 693
Fig. 1 Depiction of different

segmented legs
hinged whereas the front end is fixed and the rear is connected to the sides of the body
using screw nut [5]. This makes it more stable while moving over rough terrains.
2.2 Cheetah Cub
The cheetah cub employs the spring-loaded pantograph mechanism. The pantograph
mechanism is in itself efficient for terrain crossing abilities. The springs are instilled
to make sure that the energy expended is not wasted. Both leg implementations
include a gravity-loaded leg spring. The hip actuator is mounted directly between
the robot’s body and leg [6]. The knee actuator, which is parallel to the main leg
spring, is mounted. The role of the knee actuator is only to withdraw the leg by the
pulling of the cable and not to extend the leg since springs are used for leg extensions
as shown in Fig. 1. This system ensures that the energy is not wasted and can be
utilized by the leg thereby increasing energy efficiency.
2.3 Anymal
The interaction of the forces is done by either by integrated load cells in the joints or
at the end-effector, by pressure transducers in the cylinders, or by a series elasticity
in every actuator. This ensures that environment–force interaction is in a controlled
manner thereby maintaining the stability of the system and also for the handling of
objects. The designing of the compliant actuators is so that it is analogous to natural
locomotion systems of movement such as tendons and muscles which regulate the
forces. SEAs can be described as actuators where springs are used to attach the load.
Other options include the use of elastic elements instead of a spring. This set-up
allows proper information of the force applied which helps in the control of the force
[7]. The movement of spring in motion such as expansion or contraction helps in the
calculation of force. Shock tolerance is an added benefit of using springs. Energy
storage is also done by the springs increasing the overall efficiency and increasing
the maximum power since the spring provides the additional energy. Traditional
gear mechanisms cannot be used as it provides very slow motion due to which series
elastic actuators are used.
2.4 ARL Monopod 2
In this robot [8], hip joint is actuated by the use of cable and pulleys. The long cables
make use of the extension springs. Leg actuation is done by the hip pulley to which
the cables are attached (Fig. 2).
The cables are attached to NSK ball screw by passing over idlers initially being
fixed to the other end. The hip motor makes use of the miniature HTD timing belt to
actuate the ball screw, and motor torque is converted into axial force which acts at the
ball nut which is done by the unity ratio pulleys. Hip pulley is actuated by the tensions
in the cable due to the force in the ball nut. During motion, the acceleration and
Fig. 2 ARL Monopod 2

deceleration of the motor can be aided by the use of springs which can again increase
efficiency. The mechanism is such that the use of springs can be accommodated.
3 Gait
The motion of robot is caused by switching from one mechanism set to another. We
define a gait as all operations, implemented by each mechanism set. Gait is the leg
phasing part of coordination problem, and in other words, we can say that a gait
defines the form and the characteristics of a body displacement. In a more general
approach, a gait is defined by Song and Waldron [9], as follows ‘A gait is defined
by the time and the location of the placing and lifting of each foot, coordinated with
the motion of the body in its six degrees of freedom, in order to move the body from
one place to another’ (Fig. 3).
From the observation and mathematical formulation of natural gaits, it was found
that many of the motions were periodic in nature. Hence, the motion was categorized
into two, periodic and non-periodic (free gaits). The periodic gaits were characterized
with continuous gait, in which the body moves with constant motion while all legs
move simultaneously and this is was generally used for flat and levelled terrains,
whereas for an uneven terrain, the animals generally use a special kind of continuous
gait called wave gait, at a lower speed. Mammal and insects change their gait pattern
on sharp irregular terrains, which was defined as secure gait and it is characterized
by the sequential motion of legs and body [3]. The body is pushed forward/backward
with all of the feet in proper contact with the ground and then one leg is moved with the
other three legs and body remaining stationary, and this gait is called discontinuous
gait [10–12]. Eventually, this gait causes intermittent body motion which is beneficial
to a real legged machine, where they are very easy to implement.
Gait
Periodic Non-Periodic
Discontinuous Free Gait

Continuous
Trot Bound Pace Creep
Fig. 3 Types of segmented legs

Fig. 4 Leg actuation in a

troit gait cycle
Fig. 5 Depiction of different

gait sequence
3.1 Types of Gaits
3.1.1 Troit Gait
In this type of gait, diagonally opposite legs are moved at an instant. This type of
movement is very fast and dynamic stability is used to maintain balance. Twice
during each cycle, the body is ballistic and without support, hence if the speed is
greater then there is more stability (Figs. 4 and 5).
3.1.2 Creeping Gait
There are different types of walking gait available. Among these creeping gaits, crawl
gait is most used due its stability. One leg up at each step, based on the order of lifting
legs it can have six different gaits. The 1423 creeping gait has the maximum stability
for walking along the x-direction and is called the crawl gait. If the walking direction
was −x, 1324 gives the crawl gait (Figs. 6 and 7).
Likewise, 1234 and 1432 represent the crawl gait in the −y and y directions,
respectively. On the other hand, the 1243 and 1342 creeping gaits give medium
stability and are suitable for turning. Graphical representation of gaits sequences is
shown in Fig. 5, and the minimum value of duty factor for static stable gaits is 0.75
for four-legged robots.

creep gait cycle

bounce gait cycle
3.1.3 Bounce Gait
In the bound, support alternates between pairs of legs, with the fore and hind limbs
acting in unison to thrust the body forward in a bound gait. As can be observed
from this figure, there is approximately an 1800 phase difference between the sets
of front and back legs. The bound has the shortest gait period, allowing for fre-
quent interactions of the legs with the ground. This makes the bound well suited for
obstacle avoidance and for providing propulsion. Simulation analysis first conducted
by Murphy and later validated by Neishtadt and Li showed that certain simple pla-
nar quadruped bounding models are always passively stable for dimensionless body
inertia values of less than one and duty factor (β) is always less than 0.5.
3.1.4 Wave Gait
This type of gait gives maximum stability on horizontal axis and it is defined as:
βi = β (i = 1 . . . , 4)
where 0.75 ≤ β < 1 and β is the duty factor of the wave gait. This gait is usually like
static stability where three legs on the ground and one leg is lifted which gives us
an elliptical equation which is used to find the beta and then this gait is regular and
symmetrical, so which as add is stability.
4 Stability
Stability means an equilibrium position which can be obtained by proper position of

legs and it can be measured and observed. It is very important in legged robot. There
are two types of stability: a. static and b. dynamic.
4.1 Static
It is the stability for the robot which makes it balanced so that it does not fall when
standing, i.e. the centre of gravity is within the ground contact base for understanding
we take a tripod robot it forms a triangle. So, it stands still because of its stable
structure this the robot is statically stable as also and the centre of gravity is in triangle.
This triangle is called ‘support polygon’ [13, 14] and it is a projection between all
the support points of a robot onto the surface it is standing. The minimum number
of ground contact for a robot is 3 number of legs.
4.2 Dynamic Stability
It is the stability which is to be achieved while the robot is moving. For example,
in one leg robot, it is stable as long as it is good but falls when it stops, but it is
very complex to control but moves fast and has great efficiency [15–17]. Because of
this, most four-legged robot uses this kind of stability because static is slow, that is,
because at least 3 points should be in ground contact because of this walking is slow,
whereas in dynamic stability, it can vary from 0 to n number of legs.
To get the best design, there should be both static and dynamic stability so instead
of lifting one leg two legs should be lifted so it can be less energy consumption bit
slightly less stable than static.
4.3 Stable Model
There is (2D + 1) model. In this model, there is a straight line which describes the
legs and there is an empty circle when the leg is raised and filled when it is grounded
[18]. There is a green circle that repents the centre of mass. The number indicates the
position of the robot legs, whereas the red triangle is the support polygon; by using
this, there is nearly a 10,000-robot configuration but all are not stable. To find the
stable one, we need to eliminate the unstable ones, then by using stable configuration,
gait can be obtained.
4.4 Stability of Other Robots
4.4.1 Phony Pony
It is made is such a way that it can move forward and backward. This is done by
the quadruped crawl and the diagonal gait or the trot gait. The disadvantage is that
it cannot turn. The feet are broad to make it stable. But it is not as stable as the
six-legged robot due to lack of true pitch and yaw motion at the hips.
4.4.2 KUMO1-4-Legged Walking Machine
This walking is done by an actuator and so the robot is heavier and difficult to control.
It is stable over any terrain including a rugged surface due to the legs having 3 degrees
of freedom. They can also move in any direction. So there is stability even when the
task is stopped 199 in a rugged surface.
4.4.3 Scout-1
They have dynamic stability which is achieved by leg which has 4 DOF. Another
benefit is the simplicity. Experiments conducted using treadmill found that it was
capable of walking for 32 seconds, turning and staircase climbing. It is also low cost
and efficient [19, 20]. In this, a linear actuator is in the hip controlling hoping height,
and hip only controls all the three joints.
4.4.4 Cheetah
The stability in cheetah is achieved by a flexible metal kind to provide stability and
it has elbows to lift itself [21, 22]. It also uses a computer to control each leg so it
is dynamically stable. In recent developments, it is using a bounding gait which is
between a canter and a gallop so it is the most dynamically stable gait, making it
stable in any terrain.
4.4.5 Titan Vlll
Titan VIII is an example of dynamic stability; it uses trot gait, where 2 diagonal legs
are lifted simultaneously. Its stability is based on zero moment point (ZMP). At the
time of walking, the ZMP has to be about the diagonal [17]. Thus, it makes it faster
and also efficient and making it move in all-terrain.
5 Pros and Cons of Underactuated Systems
• It is more efficient as the natural dynamics are used profitably. It saves power by
reducing the number of actuators.
• As it uses the natural dynamics of the systems, it looks more natural and smoother
whereas the fully actuated systems look artificial [23]. Moreover, having the move-
ment to be natural helps the system to tackle many real-world challenges much
easier than having to solve complex algorithms as is the case for fully actuated
systems.
• The only con is the difficulty in formulating the equations but once formulated it
is not computationally heavy.
6 Conclusion
In this paper, an overview of legged robots has been discussed and compared based
on different aspects of importance for designing a bot. The topics discussed are mech-
anism, gait and stability. Finally, the difference in underactuated and fully actuated
robots is for improving control and efficiency. The other parameters like robot con-
trol system, size and structural heterogeneity, power source, sensors and actuators
would be covered in further works.
References
1. Siegwart R, Nourbakhsh I (2004) Introduction to autonomous mobile robots: a Bradford book.

The MIT Press, Cambridge, London
2. Fu KS, Gonzalez RC, Lee CSG (1987) Robotics: control, sensing, vision and intelligence.
McGraw-Hill
3. Hirose S, Umetani Y (1980) Basic motion regulation system for a quadruped walking vehicle.
In: Design engineering technical conference, Los Angeles. Researchgate
4. Bonnier Corporation (1933) Popular science monthly, Apr 1933. Bonnier Corporation, p 46.
ISSN 0161-7370
5. Raibert M, Blankespoor K, Nelson G, Playter R (2008) BigDog, the rough-terrain quadruped
robot. In: Proceedings of the 17th IFAC world congress, COEX, South Korea, pp 10823–10825
6. Spröwitz A, Tuleu A, Vespignani M, Ajallooeian M, Badri E, Ijspeert A (2013) Towards
dynamic trot gait locomotion: design, control, and experiments with cheetah-cub, a compliant
quadruped robot. Int J Robot Res 32(8):932–950
7. Hutter M, Gehring C, Lauber A, Gunther F, Bellicoso CD, Tsounis V, Fankhauser P, Diethelm
R, Bachmann S, Bloesch M, Kolvenbach H, Bjelonic M, Isler L, Meyer K (2017) ANYmal—
toward legged robots for harsh environments. Adv Robot 31(17):918–931. https://doi.org/10.
1080/01691864.2017.1378591
8. Gregorio P (1996) Design, control and energy minimization strategies for the ARL Monopod.
Researchgate 1–54
9. Song SM, Waldron KJ (1989) Machines that walk: The adaptive suspension vehicle, The MIT
Press, Cambridge, MA, ISBN 0-262-19274-8
10. PAL Robotics. REEM-A technical specifications. https://web.archive.org/web/
20110302213317/http://www.pal-robotics.com/robots/reem-a/34-reem-a-spec. Archived
from the original on 2011-03-02
11. PAL Robotics. REEM-B technical specifications. https://web.archive.org/web/
20110720215835/http://www.pal-robotics.com/robots/reem-b. Archived from the original on
2011-03-02
12. PAL Robotics. REEM-C technical specifications. https://web.archive.org/web/
20110824004151/http://pal-robotics.com/robots/reem-a. Archived from the original (PDF) on
2013-07-17
13. Olfati-Saber R (2001) Nonlinear control of underactuated mechanical systems with application
to robotics and aerospace vehicles. Thesis (Ph.D.), Massachusetts Institute of Technology,
Department of Electrical Engineering and Computer Science, pp 1–316
14. Queiroz C, Menezes P (2000) A study on static gaits for a four-leg robot. In: Proceedings of
international conference on control, pp 1–6
15. Bottcher S (2006) Principles of robot locomotion. In: Seminar on human robot interaction, pp
3–25
16. MIT cheetah robot runs fast, and efficiently. IEEE Spectrum: Technology, Engineering, and
Science News, 2013
17. Goddard L (2012) Cheetah robot sets 28.3 mph speed record, outrunning Usain Bolt. The Verge
18. Mohamed A (2015) Assessment of the stability of a four legged robot manipulator. Int J Eng
Res 4(1):433–436. https://doi.org/10.17950/ijer/v4s8/806
19. Buehler M, Battaglia R, Cocosco A, Hawker G, Sarkis J, Yamazaki K (1998) SCOUT: a simple
quadruped that walks, climbs, and runs. Proc IEEE Int Conf Robot Autom 2(1):1707–1712.
https://doi.org/10.1109/ROBOT.1998.677408
20. Chen W, Yu Y, Zhang X (2005) A survey on the underactuated robots. Jixie Sheji yu Yanjiu
21(4):22–26
21. Li M, Wang X, Guo W, Wang P, Sun L (2014) System design of a cheetah robot toward
ultra-high speed. Int J Adv Rob Syst 11(5):1–11. https://doi.org/10.5772/58563
22. Greenemeier L (2008) DARPA pushes machine learning with legged LittleDog robot. Scientific
American, 15 Apr 2008
23. Mishra A (2014) Design, simulation, fabrication and planning of bio-inspired quadruped robot,
May 2014. Thesis. https://doi.org/10.13140/2.1.2731.6486
In-Pipe Robot
Mechanisms—State-of-the-Art Review
G. Satheesh Kumar and D. Arun
Abstract In-pipe robots are a class of robots that have been extensively studied
for a very long time and many interesting solutions have been proposed and used in
practical scenarios. There are many parameters and functional features that define the
characteristics of these robots. An extensive review of the mechanisms that allows
locomotion to the in-pipe robots is performed in this paper. A broader classification
that allows more clarity on these robots for choosing them based on their capacities
is provided. The designs brought out so far have limited itself to one type of scenario
where there is no need for the robots to work in coordination. This perspective is
different from the other review papers in that sense and is needed for the researchers
to identify new design directions for improvising the performance of their robots. A
new design perspective for in-pipe robots for swarm applications is also presented
toward the end.
Keywords In-pipe robots · Mechanism · Scavenging · Reconfigurability
1 Introduction
This paper presents a state-of-the-art review of the literature on the sewage in-pipe
robotics on the basis of the locomotion aspects. In the early 1990s, the sewage in-pipe
robot was proposed [1]. In the 2000s, there has been a steady increase in research
interest within the field of sewage in-pipe robots. It is not an uncommon sight to see
men forced to work on sewage for cleaning pipes with unreserved disregard to any
safety measures. Workers, while performing the tasks of maintenance and cleaning
sewage pipes, get affected by the toxic gases (like hydrogen sulfide or chlorine) [2],
sometimes causing instant death or at least increased the risk of serious diseases too.
Robots are designed to reduce the human elements from effortful work or danger-
ous work and additionally to act in the impassable environment. Implementation of
smart cities opens up a Pandora’s box of newer complex technical problems transfixed
G. Satheesh Kumar (B) · D. Arun

Sri Sivasubramaniya Nadar (SSN) College of Engineering, Chennai, India

https://doi.org/10.1007/978-981-15-4488-0_59
704 G. Satheesh Kumar and D. Arun
by associated social problems. Chennai, a city of 426 km2 and 4.6 million popula-
tions [3], has a sewage network of 7000 km and 77,000 plus manholes. A problem
of this scale was usually not encountered, but would have been handled by humans
manually until recently even in developed countries. Though under these conditions
it would seem appropriate to develop a robot with the equivalents of superpowers,
actually the solution would be a lot effective to have simpler robot-based solutions.
Solutions developed for this problem could be effectively expanded to provide solu-
tions for the problems of the other kind. Pipelines are used to transport drinkable
water, oil, sewage water, etc., and are prone to a lot of problems like crack, cor-
rosion, blockage, and leakages. Even for sewage pipelines continuous monitoring,
inspecting, and removing blockages are strongly recommended.
In-pipe robots and open-air robots have been exciting areas of research all these
years and their differences are quite an interesting domain for research. By open-
air, it is meant to cover all the robots which are not in-pipe robots. The technology
and components used are fundamentally alike but with vividly adapted differences.
The qualities that define the environment where the robots work also define the
characteristics of the robot. The differences would soon adopt to a scale that exists
between terrestrial and aquatic animals as research progresses. Table 1 provides a
basic analysis of the differences existing in those robots for ready reference.
2 Parameters for Comparison
In the course of development of robotics, there has been a lot of sewage cleaning in-
pipe robots developed by many researchers and some have been commercially mar-
keted successfully. The developed robots are expected to overcome these constraints
placed by the interacting environment on the in-pipe robots:
1. limits of reconfigurability
2. extreme space limitations for making turns
3. absence of light
4. presence of muddy, slushy content
5. presence of water and extremely contaminated
6. presence of stones and larger obstructing objects
7. sometimes 6 m beneath ground level—constrained communication
8. Slippery surface offered by the pipes affects traction effort
9. Should overcome aquatic pressure of up to XXX 100HP due to gravity flow
and intermediate pumps
10. Need to climb inclined pipes, sometimes up to 90°
11. Lightweight, so as not to damage the pipe but strong enough to exert load on
the obstacles to be cleared
12. Entire robot to be waterproof
13. In situ charging
14. Long interval between maintenance.
In-Pipe Robot Mechanisms—State-of-the-Art Review 705
Table 1 Comparison of in-pipe and open-air robots

S. No. Classification Open-air robot In-pipe robot
1 Mechanism Simple mechanisms alone Pipe diameter will change
would be sufficient for this frequently so need a better
type of robots mechanism to adopt a robot
through the pipe [4, 5]
2 Size Size is not determined based Should consider the size of
upon application size the robot. Because it works
determined underwater pipe, size of pipe
change continuously. Size of
the robot should adopt a pipe
[4]
3 Sensor The basic sensor needs to run High-profile sensor need.
a robot. E.g., light sensor, E.g., ultrasonic sensor,
sound sensor, temperature thermal sensor ultrasonic
sensor, contact sensor, sensor, magnetic sensor,
proximity sensor, distance infrared sensor, vision sensor
sensor, pressure sensor, tilt (camera), tactile sensor, light
sensor, positioning sensor, amplification by stimulated
acceleration, gyroscope, emission of radiation
voltage sensor, current sensor (LASER) [7]
[6]
4 Artificial intelligence High intelligence need. It Less intelligence need.
works in an outdoor, and a Because it works on
lot of problems, environment underground and there are no
change happen [6] external factors to affect a
robot. Consideration should
be in a minimum [8]
5 Traction Based on work surface High traction force need, it
traction force calculated works in the slurry, slippage
area. The wheel should be in
a grip to avoid slip [9]
6 Light It works in daylight It works in dark
7 Sealing A need for sealing is less High sealing required. It is in
water. So, sealing should
avoid water to enter a robot.
Otherwise, it affects a sensor,
board, etc. [10]
8 Weight A weight of robot considers Consideration of weight is
upon applications high. A diameter of pipe
changes so the weight of the
robot is less; otherwise, it
cannot adopt a pipe. A center
of gravity changes and it
causes failure [9]
9 Path determine No path determination robot Path of a robot does not
should move freely change [11]
10 Obstacles Cannot be determined. We can determine the
Because in the environment, obstacles. E.g., slog, tree
a lot of factors affect a robot branches, plastics, etc. [11]
In-pipe robots are broadly classified by their locomotion capabilities, sensors,

actuators used, choice of controllers, applications and purpose, functionality, cost
and reconfigurability.
Locomotion is the complement of manipulation. In locomotion, the desired envi-
ronment is mounted upon with the robot and the robot moves by transmission of force
to the environment. Manipulation, on the other hand, moves or handles the object by
way of application of force and completes a given task. Means are the same but the
ends vary entirely. In each case, the scientific basis is the study of actuators that gen-
erate interaction forces, and the mechanisms that translate it as desired. Locomotion
has the characteristics of stability, characteristics of contact, type of environment,
and vector loops among others.
Actuators and sensors are among the essential devices to drive a robot. There is a
wide range of sensors and actuators used in in-pipe robots. Sensors are used to acquire
information about the robot’s environment or perhaps to directly measure a robot’s
global position and sometimes encounter unforeseen environmental characteristics.
Similarly, the role of actuators is also inseparable from the functions of locomotion.
However, this paper deals entirely with locomotion capabilities for the purpose of
setting future research directions.
3 Mechanisms for Locomotion
The main objective of the robot is to traverse through the horizontal and vertical
cross section of the pipe. To meet this objective, there are many types of mecha-
nisms designed for locomotion and are generally classified and discussed in [12–18].
Another broad-based classification approach is discussed in this paper. They are:
1. Self-locking mechanisms
2. Climbing mechanisms
3. Twist-to-Turn mechanism
4. Reconfigurable mechanisms
5. Multiple to partially active mechanisms and
6. Simple drive-by mechanisms.
3.1 Self-locking Mechanisms
This combination of linkages makes the mechanism contract in the clockwise direc-
tion and expands in counterclockwise direction. Rotation inputs could also be
replaced by planar motion mechanisms. Most of these robots belong to this class
of mechanisms since the primary need of such robots is to clear clogs by exerting a
force on the target material. Symmetric multiple linkage mechanisms are common
in this category of in-pipe robots. Many of these robots have their maintenance unit
Fig. 1 DeWaLoP in-pipe robot mechanism [5]
together with the chassis forming a monolithic multi-module robot, which can be
easily mounted/dismounted without the need of screws. However, other robots still
have fasteners and other mods to complete the structure. The maintenance unit struc-
ture consists of usually six- to eight-wheeled legs, distributed in pairs of three/four,
on each side, separated by an angle of 120° or lesser, supporting the structure along
the center of the pipe. The advantage with these robots is its capacity to lock it
into any allowed postures. Being made of parallel loops, the mechanism inherently
brings in the load distribution, better stability, and stiffness characteristics of parallel
manipulators. This is true only when in contact with the walls of the pipes. Figure 1
shows a few robots of this type. Even the cleaning system is similar to an umbrella-
like open-and-close mechanism, which makes the robot highly adaptable to different
pipe sizes.
Contrary to the standard cylindrical robot, the location of the rotating actuator is
not in the central axis of the robot, it is located on the arm which is opposite to the
arm with the tool. Hence, this rotating actuator (drive wheel) rotates with its entire
cleaning mechanism. It is a general observation that the forward transformation for a
cylindrical robot is quite simple, because it is equivalent to the transformation from a
cylindrical to a Cartesian frame. The main challenge is to accurately place the robot
in the center of the pipe while overcoming the push back forces and vibrations caused
in the cleaning process, so that the cleaning tool is able to focus on the desired area
of the pipe surface.
3.2 Climbing Mechanisms
Wet adhesive wheels and contact points provide the ability to climb vertical walls.
But it gets quickly dried and loses adhesion. Dry adhesive medium has higher elastic
Fig. 2 THESBOT developed by Tokyo-based HiBot [1]
force and rely on Van der Waal forces. Magnetic wheel has the ability to climb vertical
wall but these wheels work only on ferromagnetic surfaces. Wall vacuum suction has
ability to attain continuous movement and high moving speed. It requires substantial
power and hence the system becomes too complex to fit into smaller piper diameters.
Figure 2 shows THESBOT with a potential to climb vertically. These robots have
the requirement to overcome gravity which it easily does by compensating with the
torque capacity of the motors. The problems faced are to ensure adequate traction
forces at the available points of contact. Most of them are also commercially available
with joystick control. A configuration that works for vertical pipelines would be able
to handle any slopes too, and however, some robots are developed to climb specific
slopes.
3.3 Twist-to-Turn Mechanisms
The main difference of this mechanism is to provide a rotary twist to the moving
unit with respect to the stationary unit of the robot in order to traverse through the
horizontal and vertical cross section of the pipe. The stationary channel acts as the
base for the mobile channel to move in the forward direction and reverse direction as
shown in Fig. 3. This skewing motion ensures the required traction forces and also
allows the robot to be positioned in any angle in relation to the axis of the pipeline.
These robots have at least two passive degrees of freedom like pitch and yaw
degrees of freedom, which allow the robot to adapt to the curved shape of pipe in
both directions. Sometimes a mechanism is provided in the robot to keep the stability
of movement and generate forces between the robot and the pipes to keep its position
in vertical and curved pipe conditions.
Fig. 3 A few twist-to-turn mechanisms D-70/2 HELI-PIPE [12] and PCV-1 [18]
A significant amount of energy is required to negotiate bumps and turns and hence
it is not preferable mechanisms under normal operating conditions. And, on the other
hand, one should provide overload protection to prevent the mechanism overloading.
3.4 Reconfigurable Mechanisms
Any understanding of the variability in the dimensions of the pipelines that the robot
would encounter would validate the needs for these robots to be reconfigurable.
The robot mechanisms are to be designed in such a way as to expand and contract
between the chosen limits. This necessitates the use of a mechanism where the input
link causes the other links to move in a uniform fashion without any crossovers.
All these characteristics bring in the need for reconfigurability in these robots [19].
Most of both the group of robots discussed in the self-locking mechanism and the
twist-to-turn types fall under this category of mechanisms, as like the one shown
in Fig. 4. During the forward motion of robot, the traction motor is in a clockwise
Fig. 4 Reconfigurable mechanisms [XX, 17, 20]

direction, but when the manipulator is stalled due to extra loading effect by extreme
blockage of sewage, the manipulator limit switch is activated.
The common umbrella-like mechanism consists of a structure with an ability
to increase its height in order to adapt to different pipe diameters. Even for the
caterpillar mechanism, turning from flat to the vertical axis inside the pipeline is
difficult. This is because of the low contact for the caterpillar wheels to have the
capacity to reach to within limits of the pipeline. At this point, when there is a less
area of contact, the robot wheels cannot have any significant bearing the usefulness
of separating the caterpillar’s pace for making a turn. The robot can make a turn
by making the wheels stationary reaching the internal corner of the T-branch and
pivoting the wheels reaching the external corner of the T-branch toward the vertical
side. The design, kinematics, and the control of these robots are relatively difficult
owing to its complexity in their vector loops. Nevertheless, they are increasingly
finding their applications in most field of robotics.
3.5 Multiple to Partially Active Mechanisms
Some scenario requires only a part of the robot to be active. In some cases, the entire
robot need not have active parts. Both the design requirements allow for some robots
to be partially active. Most of these robots would have multiple segments, some
active and some passive. Even the active segments could be a repeated segment with
same/identical functionality. Intermediate elements like universal joints, as shown
in Fig. 5 enable the robot to make a turn movement in the three-dimensional pipe
systems. The elastic sliding pairs have also adopted the synchronous mechanism to
keep the robot in stable posture and position in the pipe. Another advantage that this
mechanism offers is the possibility of more payload in more numbers.
Fig. 5 MRINSPECT robot [15]

The length of the robot places a considerable taxing on the length of the robot and
so on the cablings and makes it difficult to handle manually. Another drawback of
these robots is that the friction between the pipe and the cables for communication
and power supply makes it difficult to move a long distance. Autonomous more is
preferred in these cases.
3.6 Simple Drive-by Mechanisms
A classic example of these types of robot mechanisms is the crawler-driven mecha-

nism as shown in Fig. 6. Motion motor of the robot produces more driving force, the
adhesion force only contributed by robot weight may be insufficient, and its driv-
ing wheels may slip on the surface of pipe wall. Therefore, an additional pressure
enhancing adhesion force should be produced by the pipe diameter adaptive mech-
anism to improve the tractive capacity of the robot. Other simple wheeled units also
fall into this category. It is easier to design, build, and control these robots.
In the harsh and slippery environment of sewer, a 4-wheel drive mechanism would
be more effective and provide sufficient and stable traction in experiment. So, in the
power transmission system of the robot, the output shaft of the gearbox is coupled
with the four wheels through a right angle gearbox and a chain. Another important
factor which improves the mobility of the robot is the distance between the bottom
of the robot and the ground. Increasing the distance results in passing through bigger
obstacles and more mobility. The problem with these types of robot is the inability
to provide horizontal force to clear or perform any tasks inside the pipelines. Alone
Fig. 6 Crawler-driven mechanism [16]

these robots are only effective for surveillance and that too only where there are no
opposing forces like the presence of water or obstacles on its path.
4 Another Design Perspective
Apart from all these types of robot mechanism mentioned above, a new class would
be to have a reconfigurable in-pipe robot that allows for networking of these robots
like a transformer robot [13]. Such capabilities to configure new mechanism on the
go are needed as the robot is expected to encounter new scenarios and it cannot be
possible to have one robot with all the capabilities. Swarm robotics for in-pipe robots
calls for a complete new outlook to the usual design aspects of these mechanisms.
Many design modifications on these lines have been indicated in those conditions
[14]. A provision to hold, lift, and contact physically the other robot in order to work
together brings an entirely different scope of research, which would be covered in
another paper. The sharing of resources, shared control, and shared payload would
be the theme for this perspective. These robots are expected to use the provision
available in the environment, like a guideway or a pole to enhance its performance.
5 Conclusion
An extensive review of the mechanisms that allow locomotion to the in-pipe robots
is performed in this paper. A broader classification that allows more clarity on these
robots for choosing them based on their capacities is provided. This perspective is
different from the other review papers in that sense and is needed for the researchers
to identify new design directions for improvising the performance of their robots. A
new design perspective for in-pipe robots for swarm applications is also presented.
References
1. https://newatlas.com/hibot-thesbot-pipe-inspecting-robot/40726/
2. https://en.wikipedia.org/wiki/Sewer_gas
3. https://en.wikipedia.org/wiki/Chennai. Last accessed on 12.03.2019
4. Somashekhar D, Hiremath S, Singaperumal M, Ramakrishnan R (2013) Design and develop-
ment of a reconfigurable type autonomous sewage cleaning mobile manipulator. Procedia Eng
64:1464–1473
5. Mateos LA, Vincze M (2013) In-pipe robot with capability of self stabilization and accurate pipe
surface cleaning. Automation and Control Institute (ACIN), Vienna University of Technology
(TU WIEN), Vienna
6. Hille C, Nasir AK, Abreu AA, Piñeiro DL, Roth H (2014) Development of an outdoor mobile
robot for teleoperation as an agent for a robot network. In: Proceedings of the 19th World
Congress the International Federation of Automatic Control, Cape Town, South Africa, 24–29
Aug
7. Borenstein J, Everett HR, Feng L, Wehe D (1997) Mobile robot positioning: sensors and
techniques. J Robot Syst 14(4):231–249
8. Nassiraei AAF, Ishii K (2007) Concept of intelligent mechanical design for autonomous mobile
robots. J Bionic Eng 4:217–226
9. Yuan F, Wang L (2010) The design and study of the drainage pipelines dredging robot. In:
International Conference on Computing, Control and Industrial Engineering
10. Alireza H, Gholamhossein M (2012) Developing a sewer inspection robot through a mecha-
tronics approach. In: International Conference on Mechanical, Automotive and Materials
Engineering (ICMAME’2012), 7–8 Jan
11. Shrivastava AK, Verma A, Singh SP (2012) Development of robotic sewerage blockage detec-
tor controlled by embedded systems. In: IEEE International Conference on Advances in
Engineering, Science and Management (ICAESM-2012), 30, 31 Mar 2012, pp 290–295
12. http://scmero.ulb.ac.be/project.php?id=5&page=robotics_pipe_inspection.html
13. Kurokawa H, Tomita K, Kamimura A, Kokaji S, Hasuo T, Murata S (2008) Distributed self-
reconfiguration of M-TRAN III modular robotic system. Int J Robot Res 27(3–4):373–386
14. Ducatelle F, Di Caro GA, Pinciroli C, Gambardella LM (2001) Self-organized cooperation
between robotic swarms. Swarm Intell 5(2):73–96
15. Choi HR, Ryew SM (2002) Robotic system with active steering capability for internal
inspection of urban gas pipelines. Mechatronics 12(5):713–736
16. Yoshida Y, Ma S (2010) Design of a wall-climbing robot with passive suction cups. In: Proceed-
ings of the 2010 IEEE, International Conference on Robotics and Biomimetics, Dec 14–18,
Tianjin, China, pp 1513–1518
17. Satish PV, Dorapudi SH, Seeli H, Kumar SN (2017) Design of an In-pipe inspecion robot. Int
J Eng Res 6(03):63–68
18. Ye C, Ma S, Li H, Yang J (2010) Development of a pipe cleaning robot for air conditioning
system. In: IEEE International Conference on Robotics and Biomimetics, Tianjin, 2010, pp
1525–1529. https://doi.org/10.1109/ROBIO.2010.5723556
19. Satheesh Kumar G, Nagarajan T (2012) Reconfigurable Stewart platform for vibration isolation.
Trends Intell Robot Autom Manuf Commun Comput Inf Sci 330:37–45
20. Ajin M, Satheesh Kumar G, Karthikeyan P (2016) Design, structural and kinematic analysis
of reconfigurable sewage cleaning robot. Int J Appl Mech Mater 852(763–769):2016
Thermal and Engine Engineering
Design and Thermal Performance
Analysis on Solar Still
Shaik Subhani and Rajendran Senthil Kumar
Abstract Economically feasible and eco-friendly water treating technologies are

most expected in Asia and Africa which lead to an increase in usage of renewable
energy. Hence, solar stills are the best choice to resolve the issue related to water
scarcity. Due to the low productivity of solar stills, continuous research is being per-
formed considering different parameter changes to optimize its production. Research
had been carried out by altering and optimizing the condensing surface of solar still for
performance improvement. In this research, the modification of condensing surface
is done by multiple hemispherical slopes. The flow pattern has been analysed inside
the solar still before and after modification and the changes in performance have
been compared with single slope solar still and presented for better understanding.
Keywords Solar still · Natural convection · Rayleigh number · Nusselt number ·

Baffle · CFD
Nomenclature
g Gravitational pull (m/s2 )

H Height (m)
h Convective heat transfer coefficient KWm2
W
k Thermal conductivity K m
L Characteristic length (m)
Nu Nusselt number
Ra Rayleigh number
T Surface temperature (K)
TB Temperature at bottom (°C)
TT Temperature at top (°C)
S. Subhani · R. S. Kumar (B)

Chennai, India

https://doi.org/10.1007/978-981-15-4488-0_60
718 S. Subhani and R. S. Kumar
T∞ Bulk mean temperature (K)

v Kinetic viscosity (m2 /s)
α Thermal diffusivity (m2 /s)
β Thermal expansion coefficient (K−1 )
1 Introduction
Demand for freshwater is increasing due to an increase in population, agricultural

and industrial activities. As per sources, the water available in the world in which
97% of it is saline and 2% of it is in frozen form and only 1% of it is available
for drinking purposes. There are many desalination techniques available such as
Reverse osmosis, multistage desalination etc. but because of unavailability of such
vitalities in rural areas leads to choose the solar based desalination as the production
is free of cost and also energy is abundantly available in tropical regions. The basic
operating principle is simple as due to solar radiation the water gets heated up to
evaporating temperature, as water vapour rises and touches the condensing surface
and forms pure water removing impurities. Though it is very much available in
our region, the production rate is lower while compared to other techniques as the
availability of energy is limited and effects with factors like seasons, water mass
and other atmospheric conditions which leads researchers to work on increasing the
productivity rate with respect to energy availability.
2 Literature Review
Madhav [1] compared the performance of still made of black granite as basin mate-
rial with the still performance with basin material as iron steel basin and proved the
productivity is increased for black granite solar still around 38% than solar still with
basin iron steel where the water temperature is increased up to 87 °C in basin with
black granite as basin with iron steel is reaching around 79 °C. Badran [2] theoret-
ically compared the active single slope solar still by changing various parameters
with experimental data to get the optimized combination to increase the produc-
tivity of solar still. From the study, he proved that active solar stills have various
options to enhance the performance of the still. Abdallah et al. [3] experimentally
studied the performance of single slope solar still thermally by using different types
of absorption materials and concluded as solar still with uncoated sponge has the
highest collection rate of water during day and after considering the total production
gain at nights also they found the performance as 60% for black rocks, respectively.
Akash et al. [4] studied the effect in single-basin solar still with double slopes by
optimizing with different absorbents in the basin and further increasing the perfor-
mance of the still. Khalifa et al. [5] studied experimentally by incorporating new
Design and Thermal Performance Analysis on Solar Still 719
designs of basin-type solar stills as feedwater is preheated using solar heater and
compared with the help of performance and efficiency of the still. Ismail [6] experi-
mentally studied by designing and fabricating the hemispherical solar still and found
the efficiency as 33% with 50% as conversion ratio. Tayeb [7] experimentally stud-
ied the performance of the solar stills with different designs and by comparing the
results concluded that the highest ratio of the area of condensation to the evaporation
surface leads to the increased production of the solar still. Khalifa and Hamood [8]
studied the performance of basin-type solar still with the help of correlations and
concluded that the tilt angle of cover of about 30° gives the highest productivity
and 20% enhancement of adding soluble dyes. Akash et al. [4] studied the effect in
single-basin solar still with double slopes by optimizing with different absorbents in
the basin under same parametric conditions and found the further increase in perfor-
mance of the still. Ali et al. [9] numerically analysed by modelling the solar still with
granular activated carbon using MATLAB software and stated that with irrespective
of different heights of granular activated carbon the productive output is more up
to 2 PM and gradually decreases after that. Arunkumar et al. [10] experimentally
studied the constructed solar stills with different geometrical designs as spherical,
pyramid, hemispherical solar stills, etc., and after operating in same climatic condi-
tions and analysing the modifications on the performance concluded the solar still
with tube assistance by compound parabolic concentrator that has the maximum
production rate. Arunkumar et al. [11] experimentally studied redesigned solar still
with a hemispherical condensing surface and found that the production rate has been
increased by decreasing the temperature of condensing surface by allowing water
over it and also efficiency was increased up to 42% after cooling the condensing
surface, even though the performance has been analysed experimentally by cooling
the hemispherical condensing surface and the further research has been done by
comparing the hemispherical solar still with single-slope solar still numerically and
also increasing the number of hemispherical condensing surfaces.
3 Problem Statement
In this study, a solar still with modified condensing surface of multiple curves as
hemispherical shape is considered for numerical analysis using Ansys Fluent-18 by
considering suitable assumptions as shown in Fig. 1, and with a non-dimensionalized
design parameters, such a still walls of height H and separation between them is 3.91H
and condensing surface height is of 0.25H, 0.251H and 0.24H for single curved, dou-
ble curved and triple curved, respectively. Here, lower boundary is at the topmost
layer of water basin. Top condensing surface temperature is T T , and bottom evapo-
rating surface temperature is T B . Here, the flow is steady, 2D and laminar. Humid
air with incompressible nature is preferred as a still fluid. The fluid thermophysical
properties are referred at bulk mean temperature, i.e. [(T T + T B )/2]. The constant
operating temperature difference of 10 °C has been preferred by varying top and the
bottom wall of still ranging from 30 to 70 °C. In addition, the still performance has
Fig. 1 Solar still with single curved hemispherical condensing surface
been decided to analyse by variable operating temperature difference ranging from 5

to 25 °C to resemble the real field scenario when phase-changing material is dumped
in the still basin.
4 Governing Equations, Boundary Conditions and Solution

Procedure
The general governing equations for fluid flow and heat transfer have been reduced
to 2D form by incorporating suitable assumptions and presented in Eqs. (1)–(4)
Continuity equation
∂u ∂v
+ =0 (1)
∂x ∂y
X-Momentum equation
2
∂u ∂u ∂p ∂ u ∂ 2u
ρ u +v =− +μ + 2 (2)
∂x ∂y ∂x ∂x2 ∂y
Y-Momentum equation
2
∂v ∂v ∂p ∂ v ∂ 2v
ρ u +v =− +μ + 2 + gβ(T − T∞ ) (3)
∂x ∂y ∂y ∂x2 ∂y
Energy equation

∂T ∂T ∂2T ∂2T
u +v =α + (4)
∂x ∂y ∂x2 ∂ y2
The equations are discretized and solved using the finite volume method by impos-
ing suitable boundary conditions. No-slip boundary condition has been imposed at
walls of solar still also adiabatic boundary condition is preferred for vertical walls
of the still to resemble the insulation on the real field. The driving potential is given
on the top and bottom wall by defining constant temperatures as described in the
problem formation. The convective terms present in governing equations have been
discretized using second-order upwind scheme. SIMPLE algorithm is selected and
the residuals preferred in the order of 10−3 for the mass and momentum conservation
and 10−6 for energy.
Formulae
hL
Nu = (5)
k
g · β · (T − T∞ ) · L 3
Ra = (6)
v·α
5 Grid Generation and Grid Independence
The formulated problem consists of curved condensing surface and flat evaporation
surface leads to prefer unstructured triangular mesh elements using Ansys Work-
bench. To capture exact flow physics and heat transfer, highly intensified fine mesh
structure has adopted the region nearer to solid walls and curved surface (Fig. 2).
To ensure grid independence, a detailed study has been carried out for single-slope
solar still (Table 1) with varying mesh elements. It was found that the total grid count
(10,216 elements) has the minimum influence on final results which is also in the
acceptable range.
Fig. 2 Mesh alignment in solar still with triple curved condensing surface
Table 1 Grid independent study

No. of elements Nusselt number Percentage of variation
7984 11.95 –
9173 10.127 18.00
10,216 10.00 1.27
13,615 9.99 1.00
Table 2 Dimensions of solar still in literature

T B (°C) T T (°C) L L (m) L H (m) RH (m) θ (°)
63 48 0.438 0.075 0.187 14.35
Table 3 Validation of present

Nu (Rahbar Nu (Rashidi Nu (Present Percentage
computation with literature
and Esfahani et al. [12]) study) of variation
[13]) (%)
10.7 10.5 10.4 2.8 and 0.95
6.1 Validation of Computation
Rashidi et al. [12] and Rahbar and Esfahani [13] numerically analysed the perfor-
mance of plain solar still with dimensions 7.5 cm on the left side, 18.7 cm on the
right side with 14.35° steep angle and separation between is 43.8 cm (Table 2).
Hence, the present computation compared and validated the performance of solar
still in Table 3. Now, the present computational results closely match with literature.
6.2 Performance Analysis of Solar Still at Constant

Temperature Differences
The present study compares the velocity and temperature contours of a single-slope
solar still with modified condensing surfaces at constant temperature difference by
varying temperatures of condensing and evaporating surfaces. Firstly, the single-
slope solar still has been considered for analysis. Here, non-uniform bi-cellular flow
structure whose axis is almost normal to the condensing surface has been observed
(Fig. 3) when the temperature difference is 10 °C (T T = 30 °C, T B = 40 °C). The top
and bottom surface temperatures have been gradually increased to 40 °C and 50 °C,
respectively, but the considerable variations do not significantly influence the fluid
velocity, flow structure, vorticity magnitude (Fig. 3) and convection heat transfer
Fig. 3 Respective velocity and temperature contours in conventional still
(Table 4). Furthermore, on increasing the level of heat intensity on evaporating (60–
70 °C) and condensing (50–60 °C) surfaces, there is a considerable increase in fluid
velocity, vorticity magnitude and forms four revolving vortices at different sizes with
heat transfer enhancement of 30% (Table 4).
Now, the conventional still has been redesigned by changing the shape of con-
densing surface. Exactly when the condensing surface is single hemispherical curved
structure, the average fluid velocity increases and vorticity magnitude decreases with
appreciable increase in convective heat transfer when the top and bottom surface tem-
peratures 30 and 40 °C have been maintained in the still before and after redesign of
single curve condensing surface (Fig. 4). And further increasing temperature encour-
ages the formation of more vortices with an increase in vorticity magnitude and
decrease in average fluid velocity but the convection-based heat transfer is decreased
as the multiple revolutions of fluid segregate the heat at centre of the still and the
Table 4 Nusselt number versus Rayleigh number at constant temperature difference for before and
after redesigning of solar still
Bottom Constant Ra (103 ) Nusselt number
wall temperature Without Single Double Triple
temperature difference curves curve curve curve
(°C) with top
wall (°C)
40 10 2722.3 9.368 9.057 8.861 9.839
50 10 2350.9 9.36 9.94 9.310 9.806
60 10 2040.3 12.17 9.54 8.765 8.442
70 10 1779.3 12.17 8.06 9.994 9.132
500C
Vorticity 3.842/s
600C
Velocity Contour (0.044 m/s) Temperature Contour
600C
Vorticity 3.464/s
700C
Velocity Contour (0.0462 m/s) Temperature Contour
Fig. 4 Respective velocity and temperature contours for single curved solar still
highest top and bottom wall temperatures (Fig. 4) again form bi-cellular flow struc-
ture with heat concentrations at top and the centre of the still resulting lowest heat
transfer rate (Table 4). When the condensing surface is further modified into dou-
ble curved hemispherical shape, it forms bi-cellular flow structure symmetric to the
solar still at surface temperatures (30–40 °C) of the top and bottom walls with heat
concentration at middle and lower part of the still resulting heat transfer (Table 4)
reduction when compared to single-slope solar still (Fig. 5). Further increasing top
and bottom wall temperatures (40–50 °C) form tri-cellular flow pattern by dividing
formed vortices previously with an increase in vorticity magnitude and average fluid
velocity (Fig. 5) by encouraging heat transfer (Table 4). Exactly at higher surface
Fig. 5 Respective velocity and temperature contours for double curved solar still
Fig. 6 Respective velocity and temperature contours for triple curved solar still
temperatures, more number of revolving vortices are concentrating heat at centre of

the still with lower convective heat transfer (Table 4) when compared to single slope
solar still.
When condensing surface is again modified with triple-slope hemispherical struc-
ture, it forms more number of revolving vortices (Fig. 6) at lower surface temper-
atures (30–40 °C, 40–50 °C) of top and bottom walls with effective enhancement
of heat transfer when compared to single-slope solar still as appreciable heat is car-
ried by the fluid flow (Fig. 6). And when top wall temperature (50 °C), bottom wall
temperature (60 °C) an adverse flow pattern with non-interacting baby vortices has
been observed with decrease in natural convection all though the vorticity magnitude
increases (Fig. 6). Again at higher wall temperatures, the same favourable tri-cellular
vortices formation has been observed with effective heat transfer rate (Table 4).
6.3 Performance Analysis of Solar Still at Variable

Temperature Differences
Similarly, the performance analysis of the solar still has been done for variable
temperature difference by maintaining the top and bottom wall temperatures range
from 25 to 90 °C. Now when conventional still is taken into consideration, at lower
Fig. 7 Respective velocity and temperature contours for conventional still
temperature difference and wall temperatures, more heat transfer enhancing vortices
are formed with less velocity of the fluid.
Additionally, increase in temperature difference leads to merging of vortices
(Fig. 7) in turn decrease in heat transfer about 19% (Fig. 11), and again with increase
in temperature gives favorable number of vortices (Fig. 7) with ascension in size
where its considerable increase in amount of heat transfer, fluid velocity and vorticity
magnitude.
Again, at higher temperatures, they merge and form bi-cellular flow structure with
the axis normal to condensing surface where there will be no significant enhancement
of natural convection inside. When the condensing surface is single curved in solar
still, high-performance fluid flow structure with low vorticity magnitude and aver-
age fluid velocity has been absorbed (Fig. 8) at lower wall temperatures top (25 °C),
bottom (30 °C) with effective heat transfer (Fig. 11) but lower than single-slope solar
still. Further increasing the temperature differences, bi-cellular effective heat trans-
ferring vortices flow pattern has been absorbed with a gradual increase in vorticity
magnitude, average fluid velocity and heat transfer also with no fluctuations as in
single-slope solar still (Fig. 11). When the solar still with condensing surface is modi-
fied with double curved hemispherical shape, three revolving vortices along with one
non-interacting baby vortex are observed (Fig. 9) at lower variable temperatures and
further increase in temperature difference between top and bottom wall temperatures
tri-cellular heat transfer enhancing vortices formation is observed (Fig. 9) with merg-
ing of non-interacting baby vortex formed earlier with gradual increase in vorticity
magnitude, average fluid velocity and heat transfer also with no fluctuations as in
single-slope solar still. When condensing surface is again modified with triple slope
hemispherical structure, a tri-cellular heat transfer enhancing vortex formation with
minute flow separation has been observed (Fig. 10) at lower-to-medium temperature
Fig. 8 Respective velocity and temperature contours for single curved solar still
difference (15 °C) with gradual increase in heat transfer rate (Nu = 8.49–9.86).and by
increasing temperature difference up to maximum (i.e. 20 and 25 °C), a full surface
touching tri-cellular vortices have been formed with maximum heat transfer (Nu =
10.82 and 10.72) with no fluctuations as in single-slope solar still (Figs. 10 and 11).
Fig. 9 Respective velocity and temperature contours for double curved solar still
Fig. 10 Respective velocity and temperature contours for triple curved solar still
Fig. 11 Nu versus Ra plot

when the still is at variable
temperature difference
7 Conclusion
The flow and heat transfer analysis on single-slope and hemispherical solar still with
constant and variable temperature differences have been analysed and reported the
performance with necessary supporting contour plots and figures. Comparing single-
slope solar still with modified still of multiple curves at condensing surface, there
is an advancement in the formation of multiple vortices with an increase in circula-
tion strength and enhances the convection heat transfer considerably, by minimizing
lesser fluctuations in convective heat transfer and stable condensation. Modifying the
single-slope condensing surface into hemispherical surface gives the better revolving
structure for fluid and keeps on increasing the curves giving better results and also
multiple hemispherical surfaces can be supported easily and ensures structural sta-
bility. If now variation of Nusselt number is very small compared to other curves but
in larger-scale production, the production rate will increase. And we tried to enhance
the production rate of the still by modifying the design of single slope to curved
condensing surface meaningfully. The adverse fluid flow structure has been clearly
identified and reported in appropriate places.
Acknowledgements We sincerely acknowledge the Modelling and Computing facilities available

at Product development lab and CFD lab at Department of Mechanical Engineering, SRMIST.
References
1. Madhav J (2011) Performance of black granite basin solar still. Int J Tech Sci 2(2):161–168
2. Badran O (2011) Theoretical analysis of solar distillation using active solar still. Int J Therm
Environ Eng 3(2):113–120
3. Abdallah S, Abu-Khader MM, Badran O (2009) Effect of various absorbing materials on the
thermal performance of solar stills. Desalination 242:128–137
4. Akash BA, Mohsen MS, Osta O, Elayan Y (1998) Experimental evaluation of a single basin
solar still using different absorbing materials. Renew Energy 14(1–4):307–310
5. Khalifa AJ, Al-Jubouri AS, Abed MK (1999) An experimental study on modified simple solar
stills. Energy Convers Manage 40(17):1835–1847
6. Ismail BI (2009) Design and performance of a transportable hemispherical solar still. Renew
Energy 34(1):145–150
7. Tayeb AM (1992) Performance study of some designs of solar stills. Energy Convers Manag
33(9):889–898
8. Khalifa AJN, Hamood AM (2009) Performance correlations for basin type solar stills.
Desalination 249:24–28
9. Ali MDI, Kumar RS, Mahendran R (2011) Modelling of solar still using granular activated
carbon in matlab. Bonfring Int J Power Syst Integ Circ 1
10. Arunkumar T, Vinothkumar K, Ahsan A, Jayaprakash R, Kumar S (2012) Experimental study
on various solar still designs. ISRN Renew Energy, Article ID 569381, 10 pages
11. Arunkumar T, Jayaprakash R, Denkenberger D, Ahsan A, Okundamiya MS, Kumar S, Tanaka
H, Aybar HS (2012) An experimental study on a hemispherical solar still. Desalination 286:342–
348
12. Rashidi S, Bovand M, Esfahani JA (2016) Optimization of partitioning inside a single slope
solar still for performance improvement. Desalination 395:79–91
13. Rahbar N, Esfahani JA (2013) Productivity estimation of a single-slope solar still: theoretical
and numerical analysis. Energy 49:289–297
The Use of Hydrocarbon Refrigerants
in Combating Ozone Depletion
and Global Warming: A Review
Abstract Vapour compression systems are associated with high energy demand and
contribute to ozone depletion and global warming due to halogenated refrigerants
used in them which are unfriendly to the environment. Hydrocarbon refrigerants
have been proposed as a substitute to halogenated refrigerants due to their zero
ozone depletion potentials, very low global warming potentials, and lower energy
demand for effective operations. In this review, the terms ozone depletion and global
warming are explained and the causes. Also, the use of hydrocarbon refrigerants as
an ideal solution to energy security and environmental challenges accompanied by
vapour compression systems are discussed. Lastly, this paper revealed that hydrocar-
bon refrigerants are long term replacement for halogenated refrigerants in a vapour
compression system.
Keywords ODP · GWP · Hydrocarbon refrigerants · Energy efficiency
1 Introduction
In the world today, 17% of the world electricity demand goes into refrigeration and
air-conditioning and heat pump appliances [1]. And the trend will persist to grow
in the future to come, especially in the developing countries where we have 81% of
the world population [2]. In an effort to balance the environmental challenges such
as the depletion of the Ozone Layer (OL) and Global Warming (GW), and energy
required by Refrigeration & Air-conditioning (R&A) and heat pump systems, an
eco-solution approach must be adopted for long term sustainability.
Refrigeration and Air-conditioning (R&A) systems are designed for comforts and
preservation of foods, cooling is usually employed in order to reduce the propagation
T. O. Babarinde (B) · S. A. Akinlabi · D. M. Madyira

South Africa
T. O. Babarinde · S. A. Akinlabi

https://doi.org/10.1007/978-981-15-4488-0_61
732 T. O. Babarinde et al.
of bacteria in them. One of the advantages of cooling or freezing by preservation

is that it does not require the use of preservatives. Also, R&A systems are being
used in homes and industries for private and commercial purposes. While there is
an advancement in the technology design used today to improve comfort systems
and preservation of foods, it is leading to environmental challenges called Ozone
Depletion (OD) and Global Warming (GW) [3, 4].
The major problem facing the refrigeration industries currently are OD and GW.
Ozone molecules are made up of three atoms of oxygen, it is a poisonous gas that
if breathed in by humans or animals can lead to death. The earth stratosphere is
being surrounded by the ozone layer. The OL has its own importance, it protects
the earth from dangerous ultraviolet radiation coming directly from the sun. The
OL absorbs the ultraviolet B and the oxgen capture ultraviolet C while ultraviolet A
is being permitted but ultraviolet B is dangerous and it’s capable of destroying our
environment and threatens life [3].
It has been established that human activities are responsible for ozone layer deple-
tion due to the production of greenhouse gases called halocarbons [5]. Halocarbon
compound contains carbon and one or more halogens, this could be fluorine (Fl),
chlorine (Cl), Iodine (I) and bromine (Br). These gases are produced by humans
for industrial applications. They have wide industrial applications such as R&A,
production of foams and cleaning solvent. The three main groups of halogenated
refrigerants are chlorofluorocarbon (CFC), hydrochlorofluorocarbon (HCFC) and
hydrofluorocarbon (HFC). Refrigerants such as CFCs and HCFCs are stable and are
capable of attacking the OL that protects the earth from being exposed to danger-
ous ultraviolet radiation. When these gases are released into the atmosphere their
chemical bond is broken by UV-C radiation which leads to the release of chlorine
gas, the chlorine takes away an atom from the ozone (O3 ) and reduces it to oxygen
(O2 ). However, chlorine itself does not change but serves as a catalyst in the process.
The amount of chlorine contained in a compound determines its effect on the OL,
higher chlorine content refrigerants are capable of a longer effect than the lower
chlorine refrigerants on the OL. The chlorine content of CFCs is higher than the
HCFCs. The degree of destruction of the layer caused by the refrigerants is being
measured by Ozone Depletion Potential (ODP). In the beginning, CFC refrigerants
(Freons) were produced when there was no knowledge of its impact on the ozone
layer, the producers were only looking for a better refrigerant that would be stable
and capable of providing better cooling. Later on, HCFC refrigerants were launched
by air-conditioning manufacturers. However, today it has been discovered that they
contribute to OD and GW.
Initially, the Montreal protocol of 1987 did not include HCFC refrigerants because
they were considered to have a low ODP, an example of these refrigerants is R22
refrigerant which was the most commonly used working fluid in air-conditioning
systems, R22 has an ODP of 0.034. But few companies that saw the future started
replacing CFC refrigerants with R22 and a blend of HCFCs. But, in Copenhagen
in 1992, an amendment was made to include HCFCs among the gases that will be
phased out in the future. This was not comfortable with the companies which were
already converted from CFCs to the use of HCFCs.
The Use of Hydrocarbon Refrigerants in Combating Ozone … 733
HFC refrigerants were recommended to replace CFC refrigerants. This was not
much difficult for some companies because there is better compatibility between
HFCs and CFCs just small changes are required which is the oil type, therefore some
existing systems designed for CFCs were transformed into HFC refrigerant systems.
Global warming (GW) manifests as a result of greenhouse gases effect. Green-
house gases are not limited to CFCs and HCFCs earlier discussed, they also include
CO2 , nitrous oxide (N2 O) and methane (CH4 ). These gases have the capacity to
absorb and trap the infrared. Also, when released to the atmosphere they remain for
a long period and change the atmospheric conditions. About 30% of the solar radia-
tion that reaches the earth reflects back to space while others are passed through the
atmosphere to the earth’s surface. This infrared radiation does not have the capacity
to pass through the atmosphere due to the presence of water vapour and some other
infrared absorber present in the atmosphere, they form heat energy trapped in the
atmosphere, and the trapped heat energy causes the temperature at the earth surface
to be higher than the normal.
Although GW has a positive impact on life on earth, it is needed by humans,
animals and plant for survival, but due to human activities there is an increase in
the greenhouse gases released to the atmosphere, the quantity of infrared radia-
tion trapped in the atmosphere has increased causing increase in the atmospheric
temperature and this has led to what is being referred to as climate change today.
The causes of global warming can be categorized into two parts. The global
warming that is being caused by the release of CFCs and HCFCs in the atmosphere
and the global warming that is being caused due to the burning of fossil fuels such
as oil, natural gas and coal due to increase in energy demand.
In order to combat the challenges of OD and GW, the use of eco-friendly refrig-
erants has been adopted to replace HCFCs and CFCs in R&A systems. HCFCs will
be passed out completely by developed countries in 2020 and 2030 by developing
countries due to their higher chlorine content that causes ozone depletion and global
warming. Hydrocarbon refrigerants (HCs) have been identified as a substitute for
CFCs and HCFCs in domestic refrigeration. HCs are environment-friendly refrig-
erants due to their zero ODPs and very low GWPs compared to CFCs and HCFCs.
The environmental effect of the hydrocarbon refrigerants is shown in Table 1.
Table 1 Environmental
Refrigerant Molar mass (g/mol) ODP GWP
properties of most commonly
used hydrocarbon refrigerants R290 44.00 0 3.3
R600 58.12 0 4
R600a 58.12 0 3
R170 30.07 0 5.5
R1270 42.08 0 1.8
RC270 412.08 0 1.8
RE270 46.07 0 1
This paper discusses the conventional refrigerants used in R&A systems and
hydrocarbon refrigerants as alternatives.
2 Hydrocarbon Refrigerants
Hydrocarbon refrigerants occur naturally, the most commonly used hydrocarbons

today include R600a, R290 and R1270. Hydrocarbon refrigerants have an excellent
critical point, heat transfer and transport properties. They require less amount of mass
charge compared to CFCs, HCFCs and HFCs due to lower molar mass and density
[6, 7]. Hence, they require lower energy demand for their effective operations in
refrigeration systems. The only problem associated with hydrocarbon refrigerants
is their flammability. But, a maximum mass charge of 150 g has been endorsed for
use in domestic refrigerators and for small commercial refrigerator systems [8–10].
Due to flammability EN378-2008 [11] recommend Eq. (1) below for calculating the
allowable charge for hydrocarbon refrigerants [12].
Mmax = 2.5 × LFL1.25

m × Ha × A
0.5
(1)
where Mmax represents the maximum allowable charge in the room in kg, LFLm
denotes the lower flammability limit in kg/m3 , A represents the area of the room in
m2 and H a denotes the height of the installed appliance.
Isobutane (R600a) has excellent properties and is substitute for R12 in the house-
hold refrigerator and medium commercial refrigerator unit. In 1995 the mass produc-
tion of household refrigerators using R600a as the working fluid was first started in
Germany and spread across the whole of Europe and other parts of the world except
the United States where the safety regulations do not permit its production. R600a
is energy-efficient according to the past and recent research by some refrigeration
energy researchers [4, 7, 10, 13–15]. R600a requires between 40 and 60% of the
required mass charge amount of R12 and R134a for effective performance in refrig-
eration systems, therefore it consumes lesser energy compared to R12 and R134a in
vapour compression systems. Also, the price is cheaper compared to conventional
refrigerants.
Propane (R290) thermodynamics properties are similar to that of R22, therefore,
propane is being used as a drop-in replacement for R22 in vapour compression
system (VCRS). R290 appliances are now on increase but are restricted to the special
application due to safety. Currently, there are initiatives for the use of R290 in the
domestic air-conditioning unit in India and China.
2.1 Experimental Studies of Hydrocarbon Refrigerants

in Vapour Compression Systems
Oyedepo et al. [16] conducted an experiment on a refrigerator designed for R12

refrigerant. The R600a refrigerant was investigated in the system by varying the
capillary tube length within the range of 0.9–1.5 m, it was discovered that R600a with
0.9 m capillary tube length gave a higher cooling capacity of 9.18% and R600a with
1.5 m capillary tube length gave 24% decrease in power consumption with a higher
COP of 6.3% in the system. Righetti et al. [17] did a comparative analysis of R1234yf,
R1234ze(E) and R600a in roll bond evaporator. The result showed that R1234yf and
R1234ze(E) and R600a performed closely to R134a in the system and can be retrofit
for R134a in VCRS provided the compression displacement is adjusted or modified.
El-Morsi [18] carried out research on energy analysis of R134a, LPG and R600a in a
VCRS. The outcome revealed R600a offers highest to R134a and LPG. The COP is
10% and 5% higher respectively. However, both R600a has an advantage than R134a
in terms of affordability and GWP. Jung et al. [19] investigated the performance of
R290/R600a to substitute R12 in a household refrigerator the result showed that
R290/R600a portrayed better replacement for R12 in the system. Wongwises and
Akash and Said [20, 21] carried out a research experiment on hydrocarbon mixture
made up of R290, R600 and R600a refrigerant as a substitute for R134a in VCRS,
the result showed that R290, R600 and R600a have better performance among the
mixtures in the experiment.
Furthermore, the use of hydrocarbon mixtures had been carried out by numerous
researcher as a substitute to halogenated refrigerants in VCRS. Mohanraj et al. [22]
tested a blend of 45.2% of R290 and 54.8% of R600a as a retrofit for R134a in a VCRS.
The result showed that the R290/R600a had a higher COP, lower power consumption,
pull downtime and discharge pressure compared to the based refrigerant (R134a).
Yu and Teng [23] carried out an experiment on R290 and R600a to replace R134a
in a refrigerator system. The result showed that 40% of the mass charge of R134a
of hydrocarbon refrigerant performed better in terms of pull-down time and power
consumption with R600a offering the best performance in the system. Rasti et al.
[24] also tested R436A to replace R134a in a VCRS. The result showed 13% and
5.3% reduction in the on-time ratio and power consumption respectively, Jwo et al.
[25] conducted a study on a mixture of R600a and R290 as an alternative to R134a
in a 440L size refrigerator system. The result showed a reduction of 4.4%, 17.4%
and 40% were achieved in energy consumption, running time and refrigerant charge
respectively. Rasti et al. [26] also investigated the use of R435A and R600a as a
replacement for R134a in a household refrigerator. The outcome result revealed a
14% and 7% reduction in the power consumption at an optimum charge of R435A
and R600a respectively in the system. Babarinde et al. [10] conducted an experiment
with a mixture of R290 and R600 to retrofit R134a in vapour compression system
considering varied mass charge of mixture of R290/R600 and a modified capillary
length, the result showed a mixture of R290 and R600 performed better in the system
in comparison with R134a with the highest COP of 9.5% and a decrease in power
consumption of 12%. In a similar study Joudi et al. [27] experimented R600, R600a
and a blend of R290, R600 and R600a in a VCRS using R134a as the based line
of the experiment. The result showed that the energy consumption of the VCRS
were decreased by 2–3% using hydrocarbon refrigerants and its mixture. Oyedepo
et al. [6] investigated R600a and LPG in a domestic refrigerator by varying the mass
charge (40, 60, 80 g) of the refrigerant and also varying the capillary tube length (0.9,
1.2, 1.5 m). The result revealed that 60 g charge of LPG offered an improved COP
of 4.8% system and a better cooling capacity, however, it was reported that a 20%
decrease in compressor power consumption was achieved when compared to LPG. In
another related experiment. Babarinde et al. [28] performed an experiment on exergy
analysis of LPG and R600a in a domestic refrigerator using LPG (40 g) against R600a
(60 g). The result showed that when much lower mass charge of LPG is used the
performance is comparable with that of R600a. Li et al. [29] evaluated the energy and
exergy performance of R152a and R290 as a retrofit for R134a refrigerator system.
The result confirmed that the cooling COP was increased by 5 and 8% and the overall
exergy destruction reduced by 9.6 and 14.3% for R152a and R290, respectively.
Apart from refrigeration and air-conditioning purposes HCs are widely used for
powering internal combustion engines, heating, aerosol propellant and cooking. But
in order to prevent a fire outbreak, the safety measures associated with flammable
gases must be strictly followed.
3 Material Compatibility of Hydrocarbon Refrigerants
Vapour compression systems require oil to provide adequate lubrication to the moving
parts. When a compatible lubricant is used with refrigerant in a vapour compression
system, the compressor frictional and wear is being reduced. The reduction in the
friction of vapour compression system compressor also lead to a reduction in energy
consumed by the vapour compressor, Therefore, the amount of energy needed to
power the refrigerator for effective operations is reduced. This also translates to a
reduction in the burning of fossil fuel required to meet the energy demand by vapour
compression refrigeration systems [30–32]. The benefit of this is that the amount of
CO2 emitted into the atmosphere as a result of the burning of fossil fuel is reduced.
Though there are problems accompanying the use of lubricating oil in the vapour
compression system. Apart from supplying lubrication to the moving part of the
compressor, the oil also moves into the condenser, expansion valve and evaporator, by
this means affecting the friction and the heat transfer performance of the refrigerant
in the system but the lubricating oil and its return to the refrigerator compressor
can be minimized by using a refrigerant which is miscible with the lubricating oil.
Hydrocarbon refrigerants are miscible with mineral oil used in CFC refrigerants and
polyester oil used in HFC refrigerants in vapour compression system.
Therefore, HCs can serve as a retrofit in vapour compression systems originally
designed to work with CFCs and HCFCs. Furthermore, in refrigeration systems
various materials are being used such as copper, brass, steel, metal alloy or solder
Table 2 Selection guide for hydrocarbon refrigerants

Refrigerant Compressor Application
R290, R600a Sealed hermetic unit Fridge/freezer
R290, R600a Sealed hermetic unit Commercial equipment-medium
Accessible semi-hermetic temperature
Reciprocating open drive
R170, R290 Sealed hermetic unit Commercial equipment-medium
Accessible semi-hermetic temperature
Reciprocating open drive
R170, R290, R600a Reciprocating open drive Industrial
Screw
R290, R600a Hermetic reciprocating Mobile refrigeration and air-conditioning
Open drive
R600a Centrifugal Air-conditional
R290 Accessible semi-hermetic
R600a Screw
R600a Hermetic
with zinc, others are materials used for sealing. Metal alloy or solder with zinc is
capable of increasing corrosion rate with polyester oil and HFCs, however, HCs do
not behave in this manner, see Table 2. Generally, HCs are compatible with traditional
materials used in the refrigeration system.
4 Conclusion
The use of halogenated refrigerants in R&A systems will be phased out completely
in the near future as a result of their ozone-depleting and global warming potentials
which are a threat to the environment. Therefore, Earth the only planet that supports
life must be protected by a reduction in the release of greenhouse gases into the
atmosphere. In summary, the following conclusion can be drawn out:
• The use of hydrocarbon refrigerants in VCRS is safe, environment-friendly,
energy-efficient and a good substitute for halogenated refrigerants in R&A
systems.
• Regarding the flammability of hydrocarbon refrigerants, a maximum mass charge
of 150 g has been recommended for use in domestic refrigeration systems.
• Hydrocarbon refrigerants have shot lifetime which makes them very low in GWP.
• Hydrocarbon refrigerants require a lower mass charge and lower discharge
pressure for effective operation.
• R290 refrigerant has been commercialized as a replacement for R22 with a low
mass charge in small air-conditioners.
• Hydrocarbon refrigerants are compatible with mineral oil and polyester oil
lubricant in R&A systems.
This paper brings into awareness the advantages of using hydrocarbon refriger-
ants to replace halogenated refrigerants as a means of combating the environmental
problems and energy security associated with vapour compression systems.
Lastly, for future studies, the following studies need to be done.
• Hydrocarbon refrigerant effect on the wear of vapour compression systems
compressor.
• The possibility of using hydrocarbon refrigerants for large cooling applications
and the safety precautions required.
References
1. UNEP (2000) United Nation environmmental program. In: Hand book for international treaties
for protection of ozone layers, 5th edn. Nairobi, Kenya
2. Ciconkov R (2018) Refrigerants: there is still no vision for sustainable solutions (Frigorigènes:
toujoursaucuneapproachfavorable à des solutions durables). Int J Refrig 86:441–448
3. UNEP (2012) Handbook for the Montreal protocol on substances that deplete the ozone layer
4. Babarinde TO, Akinlabi SA, Madyira DM, Ohunakin OS, Adelekan DS, Oyedepo SO (2018)
Comparative analysis of the exergetic performance of a household refrigerator using R134a
and R600a. Int J Energy Clean Environ 19(1–2)
5. Bolaji BO, Huan Z (2013) Ozone depletion and global warming: case for the use of natural
refrigerant—a review. Renew Sustain Energy Rev 18:49–54
6. Oyedepo SO et al (2017) A comparative study on the performance of the domestic refrigerator
using R600a and LPG with varying refrigerant charge and capillary tube length. Int J Energy
Clean Environ 18(4)
7. Babarinde TO, Akinlabi SA, Madyira DM, Ohunakin OS, Adelekan DS, Oyedepo SO (2018)
Comparative analysis of the exergetic performance of a household refrigerator using R134a
and R600a. Int J Energy Clean Environ 19(1–2):37–48
8. Harby K (2017) Hydrocarbons and their mixtures as alternatives to environmental unfriendly
halogenated refrigerants: an updated overview. Renew Sustain Energy Rev 73:1247–1264
9. Gill J, Singh J (2017) Energy analysis of vapor compression refrigeration system using mixture
of R134a and LPG as refrigerant. Int J Refrig
10. Babarinde TO, Ohunakin OS, Adelekan DS, Aasa SA, Oyedepo SO (2015) Experimental study
of LPG and R134a refrigerants in vapor compression refrigeration. Int J Energy Clean Environ
16(1–4):1–10
11. Babarinde TO, Akinlabi SA, Madyira DM (2018) Enhancing the performance of vapour com-
pression refrigeration system using nano refrigerants: a review. IOP Conf Ser Mater Sci Eng
413:012068
12. Yang Z, Wu X (2013) Retrofits and options for the alternatives to HCFC-22. Energy 59:1–21
13. Kasperski J, Gil B (2014) Performance estimation of ejector cycles using heavier hydrocarbon
refrigerants. Appl Therm Eng 71(1):197–203
14. Gill J, Singh J (2017) Energetic and exergetic performance analysis of the vapor compression
refrigeration system using adaptive neuro-fuzzy inference system approach. Exp Therm Fluid
Sci 88:246–260
15. Gill J (2017) Component-wise exergy and energy analysis of vapor compression refrigeration
system using mixture of R134a and LPG as refrigerant
16. Oyedepo SO, Fagbenle RO, Babarinde T, Tunde A (2016) Effect of capillary tube length and
refrigerant charge on the performance of domestic refrigerator with R12 and R600a. Int J Adv
Thermofluid Res 2(1):2–14
17. Righetti G, Zilio C, Longo GA (2015) Comparative performance analysis of the low GWP
refrigerants HFO1234yf, HFO1234ze(E) and HC600a inside a roll-bond evaporator. Int J Refrig
54
18. El-Morsi M (2015) Energy and exergy analysis of LPG (liquefied petroleum gas) as a drop-in
replacement for R134a in domestic refrigerators. Energy 86:344–353
19. Jung D, Kim CB, Song K, Park B (2000) Testing of propane/isobutane mixture in domestic
refrigerators. Int J Refrig 23(7):517–527
20. Wongwises S (2006) Experimental investigation of hydrocarbon mixtures to replace HFC-134a
in an automotive air conditioning system. Energy Convers Manag 47:1644–1659
21. Akash BA, Said SA (2003) Assessment of LPG as a possible alternative to R-12 in domestic
refrigerators. Energy Convers Manag 44(3):381–388
22. Mohanraj M, Jayaraj S, Muraleedharan C, Chandrasekar P (2009) Experimental investigation
of R290/R600a mixture as an alternative to R134a in a domestic refrigerator. Int J Therm Sci
48(5):1036–1042
23. Yu CC, Teng TP (2014) Retrofit assessment of refrigerator using hydrocarbon refrigerants.
Appl Therm Eng 66(1–2):507–518
24. Rasti M, Hatamipour MS, Aghamiri SF, Tavakoli M (2012) Enhancement of domestic refriger-
ator’s energy efficiency index using a hydrocarbon mixture refrigerant. Meas J Int Meas Confed
45(7):1807–1813
25. Jwo C, Ting C, Wang W (2009) Efficiency analysis of home refrigerators by replacing
hydrocarbon refrigerants. Measurement 42(5):697–701
26. Rasti M, Aghamiri S, Hatamipour MS (2013) Energy efficiency enhancement of a domestic
refrigerator using R436A and R600a as alternative refrigerants to R134a. Int J Therm Sci
74:86–94
27. Joudi KA, Mohammed ASK, Aljanabi MK (2003) Experimental and computer performance
study of an automotive air conditioning system with alternative refrigerants. Energy Convers
Manag 44(18):2959–2976
28. Babarinde TO, Akinlabi SA, Madyira DM (2018) Exergetic performance of household
refrigerator using R600a and LPG. IOP Conf Ser Mater Sci Eng 413:012069
29. Li G, Eisele M, Lee H, Hwang Y, Radermacher R (2014) Experimental investigation of energy
and exergy performance of secondary loop automotive air-conditioning systems using low-
GWP (global warming potential) refrigerants. Energy 68:819–831
30. Fatouh M, El Kafafy M (2006) Experimental evaluation of a domestic refrigerator working
with LPG. Appl Therm Eng 26(14–15):1593–1603
31. Ohunakin OS et al (2018) Performance of a hydrocarbon driven domestic refrigerator based
on varying concentration of SiO2 nano-lubricant. Int J Refrig
32. Wongwises S, Chimres N (2005) Experimental study of hydrocarbon mixtures to replace HFC-
134a in a domestic refrigerator. Energy Convers Manag 46(1):85–100
Experimental Study of Performance
of R600a/CNT-Lubricant in Domestic
Refrigerator System
Abstract Energy consumption and environmental problems have been the major
consideration for comfort systems manufacturers. This research work investigated a
varied mass charge of R600a, an eco-friendly refrigerant with a low concentration
of 0.4 and 0.6 g/L of CNT nanolubricant concentration in a domestic vapor com-
pression refrigerating system working with pure mineral oil as the base lubricant.
The experimental test performance of the system was studied considering pull-down
time, COP, power consumption, and cooling capacity. The result showed that CNT
nanolubricant had lower evaporator air temperature with higher COP and cooling
capacity with a reduction in power consumption compared to R600a in the base
lubricant in the system.
Keywords R600a · CNT · COP · Power consumption
Nomenclature
COP Coefficient of performance

EG Ethylene glycol
h Enthalpy (kJ/kg)
m Mass flow rate (kg/s)
MWCNT Multiwall carbon nanotube
PAG Polyalkylene glycol
Qevap Cooling capacity (W)
SG Solid grinding
SWCNT Single wall carbon nanotube
T. O. Babarinde (B) · S. A. Akinlabi · D. M. Madyira

South Africa
T. O. Babarinde

https://doi.org/10.1007/978-981-15-4488-0_62
T Temperature (°C)
TEM Transmission electron microscope
VCRS Vapor compression refrigeration system
WC Compressor power in put (W)
Subscripts
1 Compressor inlet
2 Compressor outlet
3 Condenser outlet
4 Evaporator outlet
1 Introduction
Most energy demanded in homes and commercial building is from refrigeration

and air-conditioning system, heating, and ventilation. Hence, energy researchers
are looking into improving the energy consumed by comfort systems such as the
domestic refrigerators and air-conditioning systems [1, 2]. Introducing nanofluids
into vapor compression systems is capable of giving that solution to the high energy
consumption associated with refrigeration systems.
Choi et al. found out that the use of nanoparticles as additives in refrigerator
lubricants and refrigerants has the potential to enhance its thermal properties [3].
When nanoparticles are being dispersed in lubricant or refrigerant, they are called
nanolubricant or nanorefrigerants. The major advantages of nanoparticles in the
compressor lubricant are to enhance the thermal conductivity and viscosity which in
turn cause an improvement in the energy efficiency of the system [2]. However, the
performance of nanoparticles in the base fluid largely depends on its size, shape, and
concentration [4].
Some researchers have studied the use of nanolubricants and in cooling systems
such as VCRS in order to enhance the performances [5–8]. Nabil et al. [9] inves-
tigated the heat transfer performance of water-ethylene glycol (EG) mixture using
TiO2 nanoparticles within a temperature range of 30–80 °C. The result showed 15.4%
improvement at 60 °C. Wang et al. [10] compared fullerene (170) and NiFe2 O4 nanol-
ubricant with base lubricant in an automobile air-conditioning system. The result
showed that coefficient of performance (COP) of the air-conditioning system was
increased by 23%. Also, a reduction in the coefficient of friction was achieved which
means a reduction in wear. Melnyk [11] tested Al2 O3 and TiO2 nanoparticles with
R600a refrigerant using mineral oil in vapor compression refrigeration system. The
outcome result indicated that R600a/Al2 O3 and R600a/TiO2 performances improved
Experimental Study of Performance of R600a/CNT-Lubricant … 743
in terms of thermal conductivity and viscosity in the system. Akilu et al. [12] exper-
imented TiO2 –CuO nanocomposite in ethylene glycol by varying the TiO2 –CuO
nanocomposite volume concentration. The result revealed the thermal conductivity
and viscosity are functions of nanoparticles volume concentration and temperature.
The thermal conductivity and viscosity increased at 20% volume concentration in
comparison with the pure ethylene glycol at 40.4 °C.
Sun et al. [13] evaluated the performance of the heat transfer of MWCNT-
OH/R141b and MWCNT-COOH/R141B in a cooling system by varying the volume
concentration (0.059, 0.117, and 0.17655% vol.) at vapor quality range of 0.2–0.7.
The result indicated the heat transfer coefficient of MWCNT-COOH/R141B was
higher compared to MWCNT-OH/R141b. Mahmoud [14] also investigated the per-
formance of SWCNT in refrigeration system using R134a as working fluid. The
nanoconcentration was varied from 1 to 5 vol.% within the temperature range of
300–320 K. The result showed increase in thermal conductivity, viscosity and spe-
cific heat transfer of SWCNT/R134a in the system compared to R134a refrigerant.
Hung and Gu [15] also carried out research on MWCNT nanofluid in a hybrid energy
system. The MWCNT was added as additives in water with a weight fraction of 0.125,
0.25, and 0.5% to produce a nanofluid. The 0.125% nanofluid showed 5% increase
in heat transfer behavior compared to the water.
From the existing literature, CNT nanoparticles are capable of increasing the
thermal conductivity, viscosity, and heat transfer rate when used as additives in
the base fluid. Currently, there is scant literature on the use of CNT in the vapor
compression system. Furthermore, CNT nanoparticles have not been investigated in
hydrocarbon refrigerants in which R600a is one of such. Today, R600a is the most
widely used hydrocarbon refrigerant used as a substitute for R134a in a domestic
refrigerator. Therefore, there is a need for its energy improvement in the domestic
refrigeration system. This work presents an application of CNT nanoparticles in
domestic refrigerator system using R600a refrigerant, which is clean and cheaper.
2 Experimental Procedures
The experimental test rig for this experiment was a household vapor compression
refrigerator system initially designed to work with R134a refrigerant. The refrigera-
tor system was evacuated with the aid of vacuum flasher before use and after each use.
A multi-wall CNT was considered for this experiment. The CNT nanoparticles were
measurement with a digital charging scale, the CNT nanoparticles were dispersed in
a mineral oil and agitated together with the aid of digital ultrasonic machine, and two
samples (0.4 and 0.6 g/L) of CNT nanolubricant concentration were prepared for 50,
60, and 70 g mass charge of R600a refrigerant each. The refrigerants were charged
into the system with the aid of digital charging scale. The system was evacuated and
flushed after each experiment to ensure better accuracy. The temperature readings
were taken at the inlet and outlet of each refrigerator components (compressor, con-
denser, expansion valve, and evaporator) with K-type thermocouples connected to
the inlet and outlet of each component, and also the evaporator air temperature of
the refrigerator was monitored with K-type thermocouple. Two pressure gauges were
connected to the inlet and out of the compressor to measure the suction and discharge
pressure of the compressor. A digital wattmeter was used to measure the power con-
sumed by the refrigerator compressor. The experiment was carried out at an average
environmental temperature of 27 °C, and the experiment was repeated five times
at an interval of 30 min for four hours. The measurement of the uncertainty of the
measuring instruments used with experimental ranges and condition are in Tables 1
and 2, and the experimental setup and the refrigeration cycle are shown in Figs. 1 and
2. The experimental output readings were used to evaluate the performances of the
system using Refprop version 9.0. Performances such as coefficient of performance
(COP), power consumption, thermal conductivity, and viscosity of the system were
evaluated.
The performance analysis of the experimental result was calculated using Eqs. (1)–
(3).
1. Cooling capacity (Qe ) is capacity given by
Q e = m(h 1 − h 4 ) (kW) (1)
2. The compressor power input is given by
Table 1 Specifications of the measuring instrument and uncertainty

S. No. Parameter Specification Range Uncertainty
1 Temperature Digital thermocouple −50 to 700 °C ±3 °C
K
2 Pressure Digital pressure 5–4000 Pa ±1%
gauge
3 Power consumption Watt/Watt-h-meter 1–2000 W ±1%
(0.0001–999.9 kWh)
Table 2 Experimental range and conditions

S. No. Parameter Experimental range
1 Refrigerant mass charge 50, 60, 70 g
2 Refrigerant R600a
3 Compressor lubricant Pure lubricant, CNT (Mult-wall) nano-lubricants
4 Carbon nanotubes (CNT) 10 nm ± 1 nm × 4.5 nm ± 0.5 nm × 3 − ~6 µm
5 Nanolubricant concentration 0.4, 0.6 g/L
6 Capillary tube length 1.5 m
7 Condenser type Air cooled
8 Evaporator size 70 L
Fig. 1 Transmission electron microscope (TEM) image of CNT nanoparticles
WC = m(h 2 − h 1 ) (kW) (2)
Q evap
3. COP = (3)
WC
The comparison of the performance of R600a in the base lubricant and CNT nanol-
ubricant concentration is discussed next in terms of evaporator air temperature,
coefficient of performance, power consumption, and cooling capacity.
3.1 Evaporator Temperature
Table 3 depicts the pull-down time of mass charge of R600a with pure lubricant,
0.4 and 0.6 g/L CNT nanolubricant. It can be observed that R600a with 0.4 g/L of
nanolubricant concentration has the shortest pull-down time and lowest evaporator
air temperature (see Table 3). The evaporator air temperature of −7, −11, and −
3 °C at a pull-down time of 180, 150, and 90 min was recorded for 50, 60, and
70 g of R600a with 0.4 g/L CNT nanolubricant concentration. Also, evaporator air
temperature of −8, −11, and −8 °C at a pull-down time of 210, 240, and 210 min was
recorded for 50, 60, and 70 g of R600a with 0.6 g/L CNT nanolubricant concentration,
respectively. This result also establishes the research work of You et al. and Malvandi
et al. that the cooling temperature is enhanced with nanofluid than the base fluid in
cooling application [16, 17]. The improvement is also as result of enhancement in
heat transfer due to increase in thermal conductivity of the system.
3.2 Coefficient of Performance
The graph in Fig. 3 shows the effect of CNT nanolubricant on the COP of the system.
The COP of R600a with CNT nanolubricant is higher than the base lubricant (see
Tables 4, 5 and 6). The COP of R600a with the 0.4 g/L of CNT nanolubricant ranges
from 2.1 to 2.8 with a 50 g charge of R600a having the highest COP, while the COP
of R600a with 0.6 g/L of nanolubricant ranges from 2.5 to 2.9 with the 60 g mass
charge of R600a offering the highest COP in the system. This can be linked influence
of the CNT nanoparticles on the system which led to reduction in the coefficient of
friction and wear in the compressor, therefore reducing the energy consumed by the
compressor in the system.
3.3 Power Consumption
Figure 4 shows the power consumption at steady state and the effect of CNT on
the power consumption of the system. The power consumption of R600a with CNT
nanolubricant concentration decreases with increase in nano-concentration until it
Table 3 Evaporator air temperature of R600a charges using pure lubricant oil and CNT lubricant concentrations
Time (min) 50 g 60 g 70 g
Pure R600a R600a/CNT R600a/CNT Pure R600a R600a/CNT R600a/CNT Pure R600a R600a/CNT R600a/CNT
(0.4 g/L) (0.6 g/L) (0.4 g/L) (0.6 g/L) (0.4 g/L) (0.6 g/L)
0 27 27 27 27 27 27 27 27 27
30 9 0 −1 1 −4 −4 5 5 1
60 2 −4 −5 −2 −7 −7 2 −1 −3
90 −1 −5 −6 −3 −8 −8 1 −3 −5
120 −3 −6 −7 −3 −10 −9 1 −3 −7
150 −4 −6 −7 −3 −11 −9 0 −3 −7
Experimental Study of Performance of R600a/CNT-Lubricant …
180 −4 −7 −7 −3 −11 −10 0 −3 −7

210 −5 −7 −8 −3 −11 −10 0 −3 −8
240 −5 −7 −8 −3 −11 −11 0 −3 −84
747
3.0
2.5
2.0
COP
1.5
1.0
0.5
0.0
re L) L) re L) L) re L) L)
Pu g/ g/ Pu g/ g/ Pu g/ g/
( 0.4 ( 0 .6 ( 0.4 ( 0. 6 ( 0.4 ( 0.6
T T T T T T
CN CN CN CN CN CN
C1 g g g
50 60 70
Fig. 3 Effect of CNT nanolubricants on COP of the R600a in the system
Table 4 Summary of the experimental result for 50 g mass charge R600a in the system
Pull downtime (min) Evaporator air temp. (°C) COP W c (W) Qe (W)
Pure 210 −5 2.1 82.1 173.6
0.4 g/L 180 −7 2.8 61.1 171.4
0.6 g/L 210 −8 2.6 71.9 187.0
Pure 90 −3 2.1 82.2 167.9
0.4 g/L 150 −11 2.6 72.8 189.3.
0.6 g/L 240 −11 2.9 63.9 185.3
Pure 210 −0 1.6 86.1 141.0
0.4 g/L 120 −3 2.1 81.4 170.4
0.6 g/L 150 −3 2.5 76.1 190.2
reaches its optimal concentration and increases with the mass charge for R600a
in the base lubricant (see Tables 4, 5 and 6). The power consumption of 0.4 and
0.6 g/L CNT nanolubricant is lower from the range of 61.2–81.4 and 63.9–76.1 W,
respectively, with 50 and 60 g of R600a in 0.4 and 0.6 g/L of CNT nanolubricant
having the lowest power consumption, respectively. The trend of the improvement
in power consumption conforms to the previous result of Wang et al., Bi et al., Jeng
et al., and Jwo et al., where a reduction in power consumption of the refrigeration
90
80
Power consumption (W)
70
60
50
40
30
20
10
0
re ) ) re ) ) re L) )
Pu /L /L Pu /L /L Pu g/ /L
. 4g (0.
6g .4g (0 .
6g 0.4 (0.
6g
T(0 T T(0 T T( T
CN CN CN CN CN CN
C1 g g g
50 60 70
Fig. 4 Effect of CNT nanolubricants on the power consumption of R600a in the system
system was achieved using nanoparticles in base lubricant for the refrigerant in the
refrigeration system [10, 18–20].
3.4 Cooling Capacity
Figure 5 shows the cooling capacity of the system. The cooling capacity of the system
increases with increasing addition of nanoparticles in the system until it reaches its
optimal charge. The 60 g of R600a with 0.4 g/L CNT nanolubricant has the highest
cooling capacity of 196.5 W with an evaporator air temperature of −11 °C. This
is due to the increase in thermal conductivity of the R600a in CNT nanolubricant
compared to the base lubricant.
200
Cooling capacity (W)
150
100
50
0
re L) L) re L) L) re /L
) L)
Pu g/ g/ Pu g/ g/ Pu .4g g/
( 0.4 ( 0.6 ( 0.4 (0 .6 (0 (0 .6
T T T T NT T
CN CN CN CN C CN
C1 g g g
50 60 70
Fig. 5 Effect of CNT nanolubricants on the cooling capacity of R600a in the system
The summary of the performances of the system is expressed in Tables 4, 5 and

6.
In this research, it was noticed that for every mass charge of R600a refriger-
ant (see Tables 4, 5 and 6) CNT nanoconcentration performed better in terms of
pull-down time which established the enhancement in heat transfer rate, increase in
thermal conductivity, and most importantly the improvement in energy efficiency of
the system.
4 Conclusion
In this present study, the experimental analysis of CNT nanolubricant concentration

in R600a was compared with R600a in pure mineral oil. The different mass charge
of R600a refrigerant (50, 60, 70 g) was tested in different CNT nanolubricant con-
centrations of 0.4 and 0.6 g/L. Based on the experimental findings, and considering
the performance of the vapor compression system, the following conclusions were
made for CNT nanolubricant concentration with R600a refrigerant.
• The 60 g mass charge of R600a in 0.6 g/L CNT nanolubricant obtained the best
result in terms of COP, with evaporator temperature −11 °C at 120 min which
is found to meet the ISO 8187 standard temperature of −3 °C at 150 min. Also,
power consumption is lower compared to R600a in the base lubricant. Therefore,
it is considered as a drop-in replacement for R600a refrigerant in base lubricant
(pure mineral oil).
• The CNT nanolubricant concentrations of 0.4 and 0.6 g/L offer reduced power
consumption throughout the experiment with the 50 g charge of R600a in 0.4 g/L
CNT nanolubricant with the lowest power consumption at an evaporator air tem-
perature of −7 °C. Also, this can serve as a drop-in replacement for base lubricant
in a domestic refrigerator.
• Since CNT nanolubricant works efficiently in terms of COP, power consump-
tion, cooling capacity, and pull-down time in the domestic refrigerator, it is rec-
ommended as a substitute for base lubricant in vapor compression refrigerator
working with R600a.
References
1. Azmi WH, Sharif MZ, Yusof TM, Mamat R, Redhwan AAM (2017) Potential of nanorefrigerant
and nanolubricant on energy saving in refrigeration system—a review. Renew Sustain Energy
Rev 69:415–428
2. Babarinde TO, Akinlabi SA, Madyira DM (2018) Enhancing the performance of vapour com-
pression refrigeration system using nano refrigerants: a review. IOP Conf Ser Mater Sci Eng
413:012068
3. Choi C, Yoo HS, Oh JM (2008) Preparation and heat transfer properties of nanoparticle-in-
transformer oil dispersions as advanced energy-efficient coolants. Curr Appl Phys 8(6):710–712
4. Maheshwary PB, Handa CC, Nemade KR (2018) Effect of shape on thermophysical and heat
transfer properties of ZnO/R-134a nanorefrigerant. Mater Today Proc 5(1):1635–1639
5. Jung J, Cho C, Hyun W, Tae Y (2011) Thermal conductivity measurement and characterization
of binary nanofluids. Int J Heat Mass Transf 54(9–10):1728–1733
6. Mahbubul IM, Saidur R, Amalina MA (2013) Heat transfer and pressure drop characteristics
of Al2 O3 -R141b nanorefrigerant in horizontal smooth circular tube. Procedia Eng 56:323–329
7. Saidur R, Kazi SN, Hossain MS, Rahman MM, Mohammed HA (2011) A review on the
performance of nanoparticles suspended with refrigerants and lubricating oils in refrigeration
systems. Renew Sustain Energy Rev 15(1):310–323
8. Sun B, Yang D (2013) Flow boiling heat transfer characteristics of nano-refrigerants in a
horizontal tube (Caractéristiques du transfert de chaleur d’écoulement en ébullition de nano-
frigorigénes dans un tube horizontal). Int J Refrig 38:206–214
9. Nabil MF, Azmi WH, Hamid KA, Mamat R, Hagos FY (2017) An experimental study on the
thermal conductivity and dynamic viscosity of TiO2 -SiO2 nanofluids in water: ethylene glycol
mixture. Int Commun Heat Mass Transf 86:181–189
10. Wang R, Zhang Y, Liao Y (2017) Performance of rolling piston type rotary compressor using
fullerenes (C70) and NiFe2 O4 nanocomposites as lubricants additives
11. Melnyk AV (2017) A complex investigation of the nanofluids R600a-mineral oil-Al2 O3
and R600a-mineral oil-TiO2 (Thermophysical properties Une étude complexe des propriétés
thermophysiques des nanofluides R600a-huile minérale-Al2 O3 et R600a-huile). Int J Refrig
74:488–504
12. Akilu S, Baheta AT, Sharma KV (2017) Experimental measurements of thermal conductivity
and viscosity of ethylene glycol-based hybrid nanofluid with TiO2 -CuO/C inclusions. J Mol
Liq 246:396–405
13. Sun B, Wang H, Yang D (2018) Effects of surface functionalization on the flow boiling heat
transfer characteristics of MWCNT/R141b nanorefrigerants in smooth tube. Exp Therm Fluid
Sci 92:162–173
14. Mahmoud O (2017) OMAE2016-54071, pp 1–10
15. Hung Y, Gu H (2014) Multiwalled carbon nanotube nanofluids used for heat dissipation in
hybrid green energy systems. J Nanomater 2014(162):1–32
16. Malvandi A, Heysiattalab S, Ganji DD (2016) Thermophoresis and Brownian motion effects
on heat transfer enhancement at film boiling of nanofluids over a vertical cylinder. J Mol Liq
216:503–509
17. You SM, Kim JH, Kim KH (2003) Effect of nanoparticles on critical heat flux of water in pool
boiling heat transfer. Appl Phys Lett 83(16):3374–3376
18. Bi SS, Shi L, Zhang LL (2008) Application of nanoparticles in domestic refrigerators. Appl
Therm Eng 28(14–15):1834–1843
19. Jwo C-S, Jeng L-Y, Teng T-P, Chang H (2009) Effects of nanolubricant on performance of
hydrocarbon refrigerant system. J Vac Sci Technol B Microelectron Nanomet Struct 27(3):1473
20. Tiwari AK, Ghosh P, Sarkar J (2013) Performance comparison of the plate heat exchanger
using different nanofluids. Exp Therm Fluid Sci 49:141–151
Assessing the Predicting Capability
of RSM and ANN on Transesterification
Process for Yielding Biodiesel from Vitis
vinifera Seed Oil
V. Hariram, A. Bose, S. Seralathan, J. Godwin John,

and T. Micha Premkumar
Abstract In this present investigation, the process parameters to obtain maximum

fatty acid methyl ester yield from Vitis vinifera seed bio-oil by transesterification
were explored using the central composite design with variable input parameters
like catalyst concentration (0.5–1.5% of KOH), reaction duration (30–60 min), and
molar ratio (3:1–7:1). Response surface methodology (RSM) and artificial neural
network (ANN) were employed to predict the optimized biodiesel yield and model
the transesterification process. The experimental outputs were simulated using a
quadratic model generated by RSM. The maximum biodiesel yield parameters were
determined by RSM, and it was found to be 6.4246:1 molar ratio, 66.8205 min
reaction time, and 1.1719% of catalyst concentration. The transesterification pro-
cess performed with this experimental combination resulted in methyl ester yield of
around 97.53% which correlated well with the yield predicted by RSM. The statisti-
cal analysis was carried out to determine the model validity, accuracy, and predictive
capability of both ANN and RSM models. The biodiesel obtained by this process was
subjected to analysis for estimating the physiochemical properties like cetane num-
ber, calorific value, density, acid value, flash and fire point, and kinematic viscosity,
and it was found to be within ASTM limits.
Keywords Biodiesel · RSM · Molar ratio · ANN · Transesterification
V. Hariram (B) · A. Bose · T. Micha Premkumar

Department of Mechanical Engineering, Hindustan Institute of Technology and Science, Chennai,
S. Seralathan
Chennai, Tamil Nadu 603103, India
J. Godwin John
Department of Automobile Engineering, Hindustan Institute of Technology and Science, Chennai,

https://doi.org/10.1007/978-981-15-4488-0_63
754 V. Hariram et al.
1 Introduction
The continuous decrease in availability of fossil fuels like crude oil and natural gas
has resulted in the search for new, renewable as well as ecology-friendly source of
fuel. Based on the depletion rate of fossil fuels, the petroleum-based fuels might not
last more than a few decades. Biofuels are one of the promising resources which
can supplement the use of fossil fuels. Biodiesel are more environmentally friendly,
free from pollution, and are renewable [1]. The biodiesels are mostly synthesized
from edible oils which lead to a higher price in production and affect the avail-
ability in food market. This led to focus on synthesizing biofuels from non-edible
vegetable-based oils. Few examples of similar non-edible oils that are mostly used
to produce biodiesel are jatropha, neem, karanja, cotton seed oil, grape seed, etc. [2].
The biodiesel is generally produced from transesterification process. Transesterifi-
cation is the method of alcoholysis of triglyceride present in the bio-oil into glycerol
and alkyl esters in existence of catalyst. The biodiesel produced from the transester-
ification depends on major parameters like mixing intensity, catalyst concentration,
molar ratio of oil to alcohol, reaction duration, and reaction temperature. In order to
achieve maximum quantity and quality biodiesel yield, it is essential to improve the
process parameters [3, 4]. Soft computing tools like ANN, RSM, and ANFIS can
be used to optimize the operational parameters. RSM is a statistical tool which gen-
erates quadratic polynomial equation based on the investigational output. It is used
to derive the inter-relationship between the dependent and independent parameters.
ANN is a soft computing tool which is based on the human brain neurons ability
to learn and adapt on different conditions. The ANN consists of nodes and neurons
which act like human neurons using algorithms which enable this adaptability. Many
studies were conducted earlier using RSM and ANN for predicting the responses and
optimizing the experimental process [5, 6].
Selvaraj et al. examined the synthesis of biodiesel from waste cooking oil and
optimized its operational parameters using ANN and RSM. Optimal biodiesel yield
of 95% was achieved at optimum process conditions of 6:1 molar ratio, 1% catalytic
concentration, 75 °C, and a reaction time of 60 s with a high coefficient of determi-
nation of 0.98 for RSM and 0.99 for RSM [7]. Prakash Maran and Priya studied the
production of biodiesel from neem bio-oil and optimized it using ANN and RSM. Its
predictive capability was also compared. The optimum selection of neurons in hid-
den layer and algorithms was discussed by the author [8–10]. Chizoo et al. optimized
the production of biodiesel from sweet almond seed oil through transesterification.
The ANN and RSM were used for optimization and the CCD design. The authors
achieved an optimized yield of 95.45% from ANN and 94.36% from RSM based on
the optimized parameter, namely 1.5 w/w % catalyst concentration, 5:1 molar ratio
(methanol to oil), and 65 min reaction duration at reaction temperature of 50 °C.
Many researches were conducted using RSM and ANN to predict and optimize the
biodiesel production from various oils [4, 5, 11–13]. But, to the authors’ knowledge,
no research was identified on the optimized biodiesel production from Vitis vinifera
oil using RSM and ANN. Therefore, the objective of this investigational study is
Assessing the Predicting Capability of RSM and ANN … 755
to optimize the production of biodiesel from Vitis vinifera oil and assessing the
predictive capability of both RSM and ANN models. The physiochemical properties
of this derived Vitis vinifera biodiesel are also analyzed.
Grape seed bio-oil used in this investigational study was extracted by traditional
cold pressing. The acid value of the grape seed bio-oil was estimated to be 1.4 mg
of KOH/g. The seeds used in this extraction were supplied by Noyer Overseas India
Pvt. Ltd. The seeds were manually cleaned from dirt and stones. The seeds were
dried in sun for two days, and then, the bio-oil extraction process was initiated.
The chemicals used in the reaction—methanol (CH3 OH) and sodium hydroxide
(NaOH)—were supplied by Vetri Chemicals, Chennai.
2.1 Transesterification of Grape Seed (Vitis vinifera) Oil
50 ml of grape seed oil was taken in a beaker and heated up to 100 °C to eliminate
the existence of water molecules in the bio-oil. Then, the oil was cooled down to
reaction temperature of 60 °C which was less than the boiling point of methanol. An
appropriate amount of sodium hydroxide catalyst was mixed with the methanol to
form a homogenous sodium hydroxide solution. The catalyst solution was mixed with
the oil and agitated at an intensity of 650 rpm for about 60 min. The glycerol settles
down in the separating flask after the 24 h settling period. The yield percentage
of biodiesel was determined using Eq. (1). The biodiesel yield was influenced by
catalyst loading percentage (0.5–1.5%), reaction temperature (30–90), and methanol
to oil molar ratio (3:1–5:1) [14].
weight of biodiesel produced (g)

Yield (%) = ∗ 100 (1)
weight of oil used (g)
2.2 Experimental Design of RSM Model
The optimization of biodiesel production was carried out using the Box–Behnken
design of RSM. Three variable parameters, namely catalyst loading percentage, reac-
tion duration, and molar ratio as three factors at three levels which had 17 exper-
imental data were designed using the Box–Behnken approach. The variables were
coded in the levels of −1, 0, and +1. The experiments were carried out as per the
Table 1 Experimental and predicted values using ANN and RSM

Molar ratio Catalyst Reaction Experimental ANN RSM
(mol/mol) loading time (min) result predicted predicted
percentage value value
5 (0) 1 (0) 60 (0) 96.20 96.525 96.520
7 (+1) 1.5 (+1) 60 (0) 96.45 96.450 96.300
5 (0) 1 (0) 60 (0) 96.50 96.525 96.500
3 (−1) 1.5 (+1) 60 (0) 92.90 92.894 93.015
5 (0) 0.5 (−1) 30 (−1) 90.60 90.600 90.543
3 (−1) 1 (0) 90 (+1) 91.40 91.400 91.231
7 (+1) 0.5 (−1) 60 (0) 94.20 93.779 94.088
3 (−1) 1 (0) 30 (−1) 90.10 90.098 90.006
5 (0) 1 (0) 60 (0) 96.51 96.525 96.522
5 (0) 1.5 (+1) 30 (−1) 92.30 92.300 92.281
5 (0) 1.5 (+1) 90 (+1) 95.20 95.200 95.256
7 (+1) 1 (0) 30 (−1) 94.00 94.000 94.169
3 (−1) 0.5 (−1) 60 (0) 88.50 88.632 88.650
5 (0) 1 (0) 60 (0) 96.70 96.525 96.522
5 (0) 1 (0) 60 (0) 96.70 96.522 96.522
5 (0) 0.5 (−1) 90 (+1) 90.40 90.419 90.419
7 (+1) 1 (0) 90 (+1) 95.70 95.794 95.794
experimental design, and the values are given in Table 1. A second-order quadratic
polynomial equation which defined the yield was fitted as shown in Eq. (5)
2.3 ANN Modeling
Box–Behnken design was employed to generate seventeen test data point, and those
were given as the experimental input to the ANN model. The back propagation
algorithm of Levenberg–Marquardt with tangent sigmoidal function as the activation
function was used to train the input variables. The neural network was divided into
three layers, namely input layer, hidden layer which consisted of hidden neurons,
and output layer. The independent parameters were in the input layers, and the output
layer was biodiesel yield. The performance of the prediction was verified using mean
square error (MSE). The MSE, SEP, and AAD can be calculated using Eqs. (2), (3),
and (4).
1 2
MSE = yei − ypi (2)
n
(RMSE)
SEP = × 100 (3)
ye
n
1 yia − yip
AAD = × 100 (4)
n i=1 yia
where yei was the experimental output at ith run, ypi was the predicted output at the
ith run, and n was number of experimental data’s.
3.1 Transesterification Reaction—RSM Model
The quantity of biodiesel yield depended on the reaction parameters and to obtain a
maximum yield; it was vital for optimizing the process conditions. The transesterifi-
cation reaction was carried out using seventeen runs generated by the Box–Behnken
design. The design consisted of 17 experimental runs which were defined by 3-factor
3-level design with 5 center points. The experimental data and RSM predicted data
with the coded values were given in Table 1. As listed in Table 2, it can be observed
that the quadratic model was chosen since the p-value was less than 0.05 which
defined the significance of the model, and the adjusted coefficient of determination
(adj R2 ) and the coefficient of determination (R2 ) were closer to one [15]. The exper-
imental data was fitted into a second-order polynomial equation which defined the
yield obtained by the grape seed oil at a given parameter.
Y = 60.82306 + 5.07125x1 + 19.6130x2 + 0.27105x3 − 0.5375x1 x2

+ 0.001667x1 x3 + 0.051667x2 x3 − 0.354313x12 −8.369x22 − 0.002561x32 (5)
where y was the yield, x 1 , x 2 , and x 3 were the input parameters molar ratio, reaction
duration, and catalyst loading percentage, while x 1 x 2 , x 2 x 3 , x 1 x 3 , and x 21 , x 22 , x 23
were the interactive and quadratic terms of the input parameters. ANOVA was used
to statistically analyze the fitness of the model with the experimental data. The
parameters of the ANOVA table are given in Table 2. The statistical significance of
Table 2 Model selection

Source Sequential p-value Lack of fit p-value Adjusted R2 Predicted R2 R2
2FI 0.8818 <0.0001 0.2802 −0.0799 0.5501
Linear 0.0195 0.0001 0.4101 0.2930 0.5207
Quadratic <0.0001 0.4145 0.9940 0.9780 0.9974
Cubic 0.4145 – 0.9945 – 0.9945
Table 3 ANOVA for quadratic model

Source df Sum of p-value F-value Mean square Significance
squares
Model 9 122.08 <0.0001 297.14 13.56 Significant
A-molar ratio 1 38.06 <0.0001 833.78 38.06
B-catalyst 1 21.62 <0.0001 473.49 21.6
loading
C-time 1 4.06 <0.0001 88.96 4.06
AC 1 0.04 0.3804 0.8762 0.04 Not significant
BC 1 2.40 0.0002 52.63 2.40 Significant
AB 1 1.16 0.0015 25.31 1.16
A2 1 8.46 <0.0001 185.26 8.46
B2 1 18.43 <0.0001 403.75 18.43
C2 1 22.37 <0.0001 489.93 22.37
Pure error 4 0.1677 – – 0.0419
Lack of fit 3 0.1519 0.4145 1.21 0.0506 Not significant
Residual 7 0.3196 – – 0.0457
Core total 16 122.40 – – – –
the parameters and the model can be determined by observing the f-value and p-value
[7].
Based the ANOVA listed in Table 3, the importance of the developed model can
be seen by the p-value which was lesser than 0.0001. The linear parameters and
quadratic squares of catalyst loading percentage, reaction duration, and molar ratio
were statistically found to be significant (p-value < 0.05). The interactions between
molar ratio with catalyst loading percentage and catalyst loading percentage with
reaction time were found significant, while the interaction of catalyst concentration
with molar ratio was found insignificant. The ‘F’ value of 297.14 was observed with
low probability value (p < 0.05) which defined the acceptability of the model. Higher
p-value which was greater than 0.05 made the lack of fit value to be insignificant.
This signifies that there was a good fit between the predicted and the experimental
variables. The coefficient of variation (C.V %) and standard deviation of the devel-
oped RSM model were found to be 0.2278% and 0.2137, respectively. The effect of
the parameters and its synergetic effect on the biodiesel yield can be calculated from
the sum of squares of deviation by using Eq. (6).
sum of squares of individual terms

The contribution of terms % = (6)
total sum of square of all terms
Based on Eq. (6), the molar ratio had the highest effect of 32.6415% on biodiesel
yield which was greater than the catalyst loading percentage and reaction time whose
values were 18.542% and 3.819%, respectively. The individual parameters had more
effect on the yield of biodiesel which was greater than the interaction between param-
eters and the quadratic terms. The fitness of the model was assessed by using the R,
R2 , and adj R2 values. The high value of R2 showed that the model was about 99.74%
of variation with the experimental data. The lesser difference between R2 and pre-
dicted R2 shows that the model was highly reliable; the R defined the exceptional
correlation between the predicted response and experimental result. The low C.V %
(0.2278%) demonstrated the reliability of the model and the experiment conducted.
As plotted in Fig. 1a, the effect of catalyst loading and reaction duration on the
Vitis vinifera biodiesel yield at optimum molar ratio was observed. The synergetic
effect between reaction time and the catalyst concentration can be evidenced during
the transesterification process. With the upsurge in catalyst concentration, increased
biodiesel yield was evidenced up to a certain limit. As the catalyst concentration
surpassed the optimal value, there was a decrease in yield observed. The same can
be observed with the reaction time when raised above optimal value (66.8205 min).
This was due to the reverse reaction that takes place which caused the decrease in the
yield. A considerable amount of time was vital for the reaction to complete. Hence, a
low yield of biodiesel can be observed at low reaction time, while at surplus reaction
Fig. 1 RSM predicated optimized values of 66.8205 min reaction time, 6.4246:1 molar ratio, and
1.1719% catalyst loading to obtain maximum biodiesel yield
time, alkyl ester will start to react with the glycerol to glycerides. This led to lesser
yield at both high molar ratio and reaction time.
As shown in Fig. 1b, the synergy between the catalyst concentration and the molar
ratio at optimum reaction time can be witnessed. There was an escalation in yield as
the reaction time was increased from 30 to 60 min. The increase in catalyst more than
the optimum value prevented the reaction of methanol and oil and led to formation of
soap solution leading to decrease in yield. The same happened when the molar ratio
was amplified more than the optimum value. As the transesterification reaction was
a reverse reaction, the excess supply of methanol led to reduction in the biodiesel
yield. Figure 1c depicts the effect of reaction duration and molar ratio on the yield
of biodiesel at optimum catalyst concentration (1.1719%). There was an overall
increment in the biodiesel yield observed when the concentration of catalyst was
augmented from 0.5 to 1%. But, when it was increased from 1 to 1.5%, the increment
was only observed on lower levels. With the upsurge in the molar ratio, an rise in the
biodiesel yield was witnessed between the ranges of 6:1 and 6.5:1. When the molar
ratio reached more than that, there was a decrement in yield. As transesterification
was a reverse reaction with excess methanol and reaction time, reverse reaction tends
to take place. It led to decrease in yield. Hence, yield decrement can be observed at
higher levels of reaction duration and molar ratio. The predicted yield by RSM fits
well with the actual experimental yield. The desirability function was used by RSM
to find the maximum yield of biodiesel. The maximum yield of biodiesel predicted by
RSM was 97.6625% which was predicted to be attained at catalyst loading percentage
1.1719 wt% of oil, molar ratio of 6.4246:1, and 66.8205 min of reaction time with
desirability value of 0.964. The optimum experimental condition was used to carry
out the experiment in triplicate, and a yield of 97.53% was obtained. An error of
about 0.1325% was found between the experimental biodiesel yield and predicted
response yield. Based on this, it was possible to state that experimental condition
had less error and the experimental condition was valid.
3.2 ANN Modeling
The nonlinear relationship between the output response and input individual param-
eters was established using ANN. The data (experimental value) from Box–Behnken
design was used in artificial neural networks. In this study, the ANN was modeled
using Levenberg–Marquardt back propagation algorithm which mostly used one,
and it takes less time to process the information. It was selected through random-
ized trial and error technique. The number of neurons was increased in the order of
two, and the optimum number of neurons was found by randomized trial and error
method. 3–6–1 neuron topology for input-hidden and output layer was selected for
ANN. The neurons were selected based on the correlation coefficient of the model
topology, and there were three layers in ANN structure—input, hidden, and output
layers. The input layer consisted of three neurons which represented catalyst loading
percentage, molar ratio, and reaction time, while hidden layer contained six neurons
Fig. 2 Model topology for ANN
and output layer comprised of 1 neuron which represented the yield of biodiesel as
shown in Fig. 2. 17 experimental data which was given as an input were randomly
divided into 70% data, 15% data, and 15% data which were used for training data,
testing data, and validation data, respectively. The training data consisted of 11 data
in training, 3 data in testing, and 3 data in validation. The predicted and the experi-
mental value for the particular data and the linear equation can be observed in Fig. 3.
The coefficient of variation (R) for training, testing, validation, and whole data were
0.9981, 1, 1, and 0.99855. It can be observed in the graph there were two lines which
represented the best fit and perfect for each training, testing, validation, and all data.
MSE of the model was used to validate the performance, and high coefficient of
correlation depicted that the value determined by the ANN had high accuracy. From
Fig. 3, it was witnessed that best performance of the artificial neural network model
was found at epoch 0 with an MSE of 0.00020833. From the regression graph of
the predicted and experimental data, it can be observed that there was less deviation
from the experimental data.
3.3 Comparison of ANN and RSM
The statistical parameters like RMSE, MSE, SEP, R, R2 , Adj R2 , MAE, and AAD
were used to compare and analyze the validity and accuracy of the results given by
RSM and ANN models. The statistical parameters were calculated for both ANN
and RSM and given in Table 4. The value of R, R2 , and Adj R2 should be near to
one which defined better correlation between the predicted and experimental data.
MSE, RMSE, MAE, SEP, and AAD should be closer to zero for better predictive
ability. The R, R2 , MSE, RMSE, and SEP values for RSM was less than that of ANN,
while MAE and AAD values for the RSM was higher than that of the ANN. The
values predicted by RSM had a better correlation with the experimental values and
the interactive effect between parameters and the response can be observed more
in RSM by applying the second-order polynomial equation to predict the yield. The
relation between the nonlinear equations can be better inter-related by ANN, while its
synergic effect was showcased by RSM having higher values of R, R2 , MSE, RMSE,
and standard error of prediction, while ANN depicted lower value of AAD and MAE
which defined that ANN had higher generalized predictive ability. Moreover, RSM
Fig. 3 Regression graph for predicted versus experimental value for ANN
Table 4 Comparison of
Parameters ANN RSM
statistical parameters for
predicted values of RSM and R 0.99855 0.9986
ANN R2 0.997001 0.997382
Adj R2 0.996309 0.996778
SEP 0.143494 0.134775
MAE 0.075376 0.11246
AAD 0.079506 0.11952
RMSE 0.134577 0.1264
MSE 0.018111 0.015978
Table 5 Grape seed biodiesel properties

Fuel properties Units Standards Bio-oil Biodiesel
Fire point °C ASTM D93 262 182
Density @ 15 °C kg/m3 ASTM D127 905 880
Kinematic viscosity @ 40 °C cSt ASTM D445 26.62 4.04
Calorific value kJ/kg ASTM D240 37.7 39.52
Acid value mg of KOH/g ASTM D6751 1.4 0.45
Flash point °C ASTM D93 250 174
predicted data had better correlation to the experimental data and synergy between
operating parameters.
The kinematic viscosity of Vitis vinifera biodiesel was 4.04 cSt which was well
inside the ASTM limits. The density of fuel was around 880.2 kg/m3 . The flash
point of the biodiesel was determined as 174 °C which too had met the ASTM D93
standards. The Cetane number should have a minimum value of 47 as per the ASTM
D613. The acid value of the biodiesel was around 0.36 mg of KOH/g which met the
ASTM D6751 standards which is less than 0.5 mg of KOH/g. As listed in Table 5, it
can be observed that the biodiesel had good fuel properties compared to the mineral
diesel, and it was with the stipulated ASTM standards.
4 Conclusion
In this investigational study, optimization of Vitis vinifera oil biodiesel production

was done using methanol, sodium hydroxide as catalyst, and reaction time at three
different levels to find the optimum process parameters. Box–Behnken design of
experimental values with 17 runs at three different levels was used to determine the
optimal process conditions. A yield of about 97.665% was predicted to be attained
at a catalyst loading percentage 1.1719 wt% of oil, 66.8205 min of reaction time
and molar ratio of 6.4246:1. The value was experimentally validated, and a yield
of 97.53% was obtained with an error of 0.1325%. The modeling and predictive
ability of ANN and RSM were assessed using statistical parameters. It was found
that RSM had a better correlation with the experimental value, while ANN had
better generalized predictive ability. The physiochemical properties of Vitis vinifera
biodiesel were estimated, and it was compared with the ASTM standards. Hence,
Vitis vinifera bio-oil can be used as a future potential feedstock for the production of
biodiesel.
References
1. Ong HC, Milano J, Silitonga AS, Hassan MH, Shamsuddin AH, Wang CT, Indra Mahlia TM,
Siswantoro J, Kusumo F, Sutrisno J (2019) Biodiesel production from Calophyllum inophyllum-
Ceiba pentandra oil mixture: optimization and characterization. J Clean Prod
2. Rocabruno-Valdés CI, Ramírez-Verduzco LF, Hernández JA (2015) Artificial neural network
models to predict density, dynamic viscosity, and cetane number of biodiesels. Fuel 147:9–17
3. Meng X, Jia M, Wang T (2014) Neural network prediction of biodiesel kinematic viscosity at
313K. Fuel 121:133–140
4. Venkatesan H, John G, Sivamani S (2017) Cotton seed biodiesel as alternative fuel: produc-
tion and its characterization analysis using spectroscopic studies. Int J Renew Energy Res
7(3):1333–1339
5. Ameer K, Bae S-W, Jo Y, Lee H-G, Ameer A, Kwon J-H (2017) Optimization of microwave-
assisted extraction of total extract, stevioside and rebaudioside-A from Stevia rebaudiana
(Bertoni) leaves, using response surface methodology (RSM) and artificial neural network
(ANN) modelling. Food Chem 229:198–207
6. Betiku E, Omilakin OR, Ajala SO, Okeleye AA, Taiwo AE, Solomon BO (2014) Mathe-
matical modeling and process parameters optimization studies by artificial neural network
and response surface methodology: a case of non-edible neem (Azadirachta indica) seed oil
biodiesel synthesis. Energy 72:266–273
7. Maran JP, Priya B (2015) Comparison of response surface methodology and artificial neural
network approach towards efficient ultrasound-assisted biodiesel production from muskmelon
oil. Ultrason Sonochem 23:192–200
8. Selvaraj R, Moorthy IG, Kumar RV, Sivasubramanian V (2019) Microwave mediated produc-
tion of FAME from waste cooking oil: modelling and optimization of process parameters by
RSM and ANN approach. Fuel 237:40–49
9. Prakash Maran J, Priya B (2015) Modeling of ultrasound assisted intensification of biodiesel
production from neem (Azadirachta indica) oil using response surface methodology and
artificial neural network. Fuel 143:262–267
10. Chizoo E, Dominic OO, Uwaoma OA (2018) Optimization of methyl ester production from
Prunus amygdalus seed oil using response surface methodology and artificial neural networks.
Renew Energy
11. Dharma S, Masjuki HH, Ong HC, Sebayang AH, Silitonga AS, Kusumo F, Mahlia TMI (2016)
Optimization of biodiesel production process for mixed Jatropha curcas–Ceiba pentandra
biodiesel using response surface methodology. Energy Convers Manag 115:178–190
12. Betiku E, Ajala SO (2014) Modeling and optimization of Thevetia peruviana (yellow oleander)
oil biodiesel synthesis via Musa paradisiacal (plantain) peels as heterogeneous base catalyst: a
case of artificial neural network vs. response surface methodology. Ind Crops Prod 53:314–322
13. Avramović JM, Veličković AV, Stamenković OS, Rajković KM, Milić PS, Veljković VB (2015)
Optimization of sunflower oil ethanolysis catalyzed by calcium oxide: RSM versus ANN-GA.
Energy Convers Manag 105:1149–1156
14. Hariram V, John JG, Seralathan S (2019) Spectrometric analysis of algal biodiesel as a fuel
derived through base-catalyzed transesterification. Int J Ambient Energy 40(2):195–202
15. Venkatesan H, Godwin JJ, Sivamani S (2017) Data set for extraction of bio-oil from
Stoechospermum marginatum, a brown marine algae. Data Brief 14:623–628
Waste Heat Recovery (WHR) of Diesel
Engine Using Closed-Loop Pulsating
Heat Pipe
Saurabh B. Dhone and A. T. Pise
Abstract In internal combustion engine, around 25–40% of energy is expelled

through the exhaust gases into environment which causes pollution. So this energy
should be recovered for other applications. This present work represents the study of
the pulsating heat pipe (PHP) technique used for heating the air with the help of recov-
ered heat from exhaust of the engine. The CLPHP is most effective, promising heat
transfer device than the others, which uses the pulsating or oscillating actions of the
working fluid for transferring heat. The performance of CLPHP basically depends
on different parameters like filling ratio (F.R), internal diameter of the tube(Di ),
cross section area of tube (Ac ), heat input(Qin ), angle of inclination (θ ), no of turns
(N), working fluid, length of tube in evaporator and condenser section, etc. In this
work, eight-loop PHP is made of copper material having ID = 3.2512 mm and OD
= 4.7752 mm with water as working fluid. The experimental investigation is per-
formed for different filling ratios (50, 60 and 70%) and at various heat inputs (715.98,
861.48, 935.41, 1004.96 and 1070.5 W) by varying the loads on engine. The result
shows that for 50% filling ratio, overall thermal resistance is less and heat transfer
coefficient value is more as compared to other filling ratios.
Keywords CLPHP · Filling ratio · Thermal resistance
1 Introduction
In internal combustion engine, around 25–40% of energy is expelled through the

exhaust gases into environment which causes pollution. So this energy should be
recovered for other applications. There are a number of devices used for waste heat
recovery of engine like heat exchangers, thermoelectric generator, heat pipe, etc. The
heat pipe is also having more efficiency for transferring the heat.
Heat pipe is a device generally used to transfer the heat from one location to another
location with minimum heat loss. The heat pipe consists of 3 sections, i.e. heating
S. B. Dhone (B) · A. T. Pise

Department of Mechanical Engineering, GCE, Karad, Maharashtra 415124, India

https://doi.org/10.1007/978-981-15-4488-0_64
766 S. B. Dhone and A. T. Pise
Fig. 1 Classification of pulsating heat pipe [PHP] [2]
section (evaporator), adiabatic section and cooling section (condenser). It absorbs the
heat from hot source at evaporator and releases it at sink, i.e. at the condenser. While
heat transfer the phase change occurs in both evaporator and condenser sections, huge
amount of heat transfer takes place. It has wide range of applications in electronic
cooling, aerospace cooling, waste heat energy recovery of IC engine, gas turbine,
HVAC, VARS, and VCRS and is also used for high temperature range of applications
like in boiler or nuclear reactor etc.
In 1990, Akachi introduced a “wickless heat pipe”. The pulsating heat pipe (PHP)
and thermosyphon are the types of wickless loop heat pipe. The pulsating heat pipe
is different than the other types: In conventional heat pipe, the liquid returns from
condenser to evaporator by the capillary action; as in thermosyphon, circulation takes
place due to density difference, whereas in pulsating heat pipe, the vapour pressure
difference acts as a main driving force for the circulation of fluid [1].
2 Classification of Pulsating Heat Pipe [PHP]
The classification of PHP is as shown in Fig. 1. It is classified into three types

such as open loop, closed loop and closed loop with check valve. In OLPHPs, the
two ends of tube are not welded so that there is no possibility of circulation of
working fluid; hence, annular flow cannot be achieved. The flow will be as a pulsating
capillary slug flow regime with long vapour bubbles forming at higher heat fluxes.
Local internal heat transfer is usually more in the convective annular flow regime.
Therefore, CLPHP shows better performance than OLPHP [2].
3 Principle of Heat Transfer in CLPHP
The set-up consists of mainly three sections, i.e. condenser, adiabatic and evaporator
sections. Due to input of heat flux at evaporator, there is variation in temperature and
pressure gradient between evaporator and condenser regions. Due to continuously
heating at the evaporator, vapour pressure of fluid is increased; hence, small growth of
bubbles takes place. These bubbles are combined together and form a large bubble,
Waste Heat Recovery (WHR) of Diesel Engine Using … 767
and then it turns into the vapours. The vapour pressure difference acts as a main
driving force for the circulations of fluid. The liquid vapour slug-plug moves towards
condenser region, whereas the vapour pressure is less; therefore, condensation of the
liquid takes place. While condensation vapour releases the latent heat and the liquid
releases the sensible heat, the heat transfer performance increases.
4 Parameters Affecting the Performance of CLPHP
The following main factors which affecting the performance of CLPHP are:
• Internal diameter of tube
• Cross section
• Heat input
• No. of. loops/turns
• Length of evaporator and condenser section
• Working fluid
• Filling ratio
• Angle of inclination.
Internal Diameter (D)

The internal tube diameter is the most essential factor which defines a PHP. The
occurrence of pulsating motion inside a CLPHP is limited by the internal diameter
of tube. The critical diameter of heat pipe is given by,

σ
Dmax = Dcrit = 2 ×
g × (ρl − ρv )
where,
σ surface tension working fluid,
g gravitational acceleration,
ρl liquid density,
ρv vapour density.
If the tube internal diameter is increased beyond the critical diameter (i.e. D
>>> Dcrit ), then the surface tension gets reduced which causes the stratification of
liquid by the gravity, and therefore, heat pipe acts like a interconnected array of two-
phase thermosyphons as shown in Fig. 2 (case A and case B). If internal diameter
of the tube is very very less (i.e. D <<< Dcrit ), then the surface tension of working
fluid is more dominant and therefore a stable liquid slugs are formed inside a tube
as shown in Fig. 2 (case C) [3]. So that for the better achievement of performance
of CLPHP, the internal diameter of tube should be required slightly less than that of
critical diameter of the tube.
Fig. 2 Effect of diameter on fluid distribution inside the circular tube of CLPHP under adiabatic
and operating conditions [3]
Cross Section the Tube (Ac )

Fourgeaud et al. [4] used circular area of tube, and they conclude that with circular
section the liquid film is attached to the inner side wall of tube so that the maximum
heat is absorbed at the evaporator and released at the condenser section, and also
there is decrease in the drag force resistance offered by the fluid, hence improving
the oscillation characteristics inside the tube.
Input Heat Flux (Qin )

There is necessity of higher heat input flux for the continuous operation of the CLPHP.
With increasing the heat input, the temperature and vapour pressure of liquid increase
at the evaporator, therefore improving the bubble growth and decreasing the dynamic
viscosity of the fluid. Hence, total heat transfer rate increases.
No. of Turns (N)

The no. of turns is also another most important parameter which affects the perfor-
mance of PHP. If the no of turns or loops are increased, then the availability of area
for accepting the heat increases, therefore increasing the vapour pressure difference
in each loop of the heat pipe, which will help to form more dynamic instability or
the pulsating motion inside a tube; hence, the performance of PHP is increased.
Length of Evaporator and Condenser

Generally, the condenser area should be taken greater than the evaporator area in
order to avoid the dryout condition. As the evaporator length increased, the thermal
resistance offered by the fluid also increases and therefore the performance of heat
pipe is decreased. If the condenser length increased, then the area for dissipation of
heat also increases; hence, the performance of PHP is increased. Hence, the selections
of lengths of evaporator and condenser are the most important factor in the design
of heat pipe.
Working Fluid
The working fluid should have the following characteristics:
1. Low latent heat
2. Higher thermal conductivity
3. Lower surface tension
4. Low dynamic viscosity
5. Higher specific heat.
Filling Ratio (F.R.)

The filling ratio is defined as the ratio of available working fluid inside the tube to
the total volume of the tube. If F.R = 0, no liquid content is present inside the tube.
Hence it increases the thermal resistance which will reduce the heat transfer rate. If
F.R = 1, there is more amount of liquid content in the tube; hence, less bubble will
form which has the low ability to transfer the latent heat, so overall heat transfer rate
decreases. The optimal filling ratio is required for the best operation of the CLPHP
is generally taking the filling ratio as in between 40 and 60%.
Angle of Inclination
As the PHP is at horizontal mode, then the gravity effect is decreased; therefore, the
circulation of the fluid stops; hence, the heat transfer rate is decreased. As an angle of
inclination increased, the temperature gradient also increases. Therefore, increasing
the dynamic instability, the heat transfer performance is increased.
5 Experimental Set-Up
The experimental set-up of CLPHP for waste heat recovery [WHR] of diesel engine
is as shown in Fig. 1. Here, the copper pipe having ID 3.252 and OD 4.7752 mm
with 8 number of loops is used and tube is filled with the water. The exhaust pipe
of the engine is connected to evaporator section of the heat pipe. This exhaust heat
is absorbed at the evaporator section, and it is transferred to the condenser section
where it condenses and releases the latent heat of vapour and sensible heat of the
liquid to the environment (air). Here, blower is used to achieve proper condensation
of the working fluid (Fig. 3).
All dimensions are in mm
Fig. 3 Experimental set-up of WHR of diesel engine using CLPHP

6 Calculation of Heat Transfer Performance of PHP
The effectiveness of pulsating heat pipe can be represented by a system of thermal

resistance. The thermal resistance “Rth ” can be represented by,
Te − Tc
Rth =
Q in
The overall heat transfer coefficient “ho” of the pulsating heat pipe can be given
by
Q in
ho =
A × Te − Tc
where,
Qin heat input (in Watt),
Te Average evaporator temperature (°C),
Tc Average condenser temperature (°C),
Ae The surface area of heat pipe (m2 ), = π d i L.
The performance PHP of pulsating heat pipe depends on various factors which we
have discussed in above section. The experimental investigation is performed for
different filling ratios (50, 60 and 70%) at various heat inputs values (715.98, 861.48,
935.41, 1004.96 and 1070.5 W) by varying the loads on engine.
Figure 4 shows the various values of heat input is achieved by varying the loads on
the engine. Figure 5 shows the heat transfer coefficient with heat input for different
filling ratios. From graph, we can see that, as the filling ratio increases, there is
increment in the thermal resistance of fluid; hence, the heat transfer coefficient is
decreased, and therefore, the value of heat transfer coefficient for 50% filling ratio
is more as compared to that of 60 and 70%.
Figure 6 shows the variation of overall thermal resistance with heat input to the
CLPHP. From fig, we can see that, as the mass concentration of the fluid increased,
then the thermal resistance of the vapour also increases. Therefore, for 50% filling
ratio, the thermal resistance value is less as compared to 60 and 70%.
Figure 7 shows the variation in average evaporator temperature with respect to
time. As the evaporator temperature is increased with increasing the heat input to
PHP, and it could be decreased by increasing the mass concentration of the fluid.
Therefore, the maximum evaporator temperature is achieved for 50% filling ratio as
compared to the 60 and 70%.
Fig. 4 Heat input versus load
Fig. 5 Heat input versus heat transfer coefficient
Fig. 6 Overall thermal resistance versus heat input

Fig. 7 Avg. evaporator temperature versus time
Figure 8 shows the variation in average condenser temperature with respect to

time. The average condenser temperature is increased with increasing the heat input,
and it could be decreased by increasing the mass concentration of the fluid. Because
as the mass concentration of fluid increases, the saturation temperature and specific
heat of the fluid also increased; therefore, the condenser temperature decreases.
Therefore, the maximum condenser temperature is achieved for 50% filling ratio as
compared to the 60 and 70%.
Figure 9 shows the variation in Avg. temperature difference between the evap-
orator and the condenser. By increasing the filling ratio, thermal resistance offered
by the fluid also increases; therefore, the temperature difference between the evap-
orator and the condenser is increased. Hence, for 50% filling ratio, the temperature
difference is less as compared to 60 and 70%.
Fig. 8 Avg. condenser temperature versus time

Fig. 9 Heat input versus average temperature difference between evaporator and condenser
8 Conclusion
The pulsating heat pipe is a simple and more efficient heat transfer device. From the
above results, we concluded that the performance of closed-loop pulsating heat pipe
for 50% filling ratio is more as compared to filling ratio of 60 and 70%.
References
1. Chavan D, Kale VM (November, 2015) On the design fundamentals of pulsating heat pipe: an
overview. IJERT 4(11)
2. https://www.google.com/search?q=heat+pipe&source=lnms&tbm=isch&sa=X&ved=
0ahUKEwifnLvHn-PhAhUN8HMBHUdLDbUQ_AUIDigB&biw=1366&bih=657
3. Haque S (December, 2011) Thermal characteristics of an aluminum closed-loop pulsating heat
pipe charged with ammonia. M. S. thesis, mechanical department, Bangladesh University of
engineering and technology Dhaka-1000, Bangladesh
4. Fourgeaud L, Nikolayev VS, Ercolani E, Duplat J, Gully P (2017) In situ investigation of liquid
films in pulsating heat pipe. Appl Therm Eng 126:1023–1028
An Experimental Investigation of Heat
Transfer in Heat Exchangers Using
Al2 O3 Nanofluid
Abstract Heat is the term distinct as the amount of energy transferred by the virtue of
the temperature difference. The heat flows from the region of maximum temperature
to minimum temperature by the different methods of heat conveyed. This paper
focuses on the study of heat transfer in the heat exchanger which is a device used for
transfer of thermal energy and consists of two or additional fluids or between surface
of fluid and solid to solid particulates and a fluid at different temperature, thermal
constant with and without outside heat and work communications. The parameters
of heat transfer are analyzed in different types of heat exchangers. This examination
explores the warmth move and the weight drop of cone helically curled cylinder heat
exchanger utilizing Al2 O3 /water nanofluids. The Al2 O3 /water nanofluids at 0.2, 0.4,
and 0.6% molecule volume fixations were set up with the expansion of surfactant
by utilizing the two advance strategies. The experimental study of heat transfer in
various heat exchangers is discussed.
Keywords Heat transfer · Heat exchangers · Mathematical equations ·

Experimental study
1 Introduction
Heat transfer is an application for engineering analysis, for example, heat exchangers,
combustors, furnaces, and certain electronic devices, which include both fluid and
solid components to perform the conjugate heat transfer analysis. The solid or liquid
domains are simultaneously solved where it behaves as interface mechanism for heat
transfer. Determination of an interface temperature is done by the energy balance of
the conductive solid and convective liquid. Heat exchanger is the device designed to
S. Bhuvaneswari
University College of Engineering, Pattukkottai, Rajamadam 614701, India
G. Elatharasan (B)
Department of Mechanical Engineering, University College of Engineering, Pattukkottai,
Rajamadam 614701, India

https://doi.org/10.1007/978-981-15-4488-0_65
776 S. Bhuvaneswari and G. Elatharasan
transfer the heat involving two or more additional fluids. The heat transfer process
differs based on the heat exchanger type like gas to gas, liquid to gas, or liquid to
liquid which occurs through the solid separator which prevents mixing of fluids or
fluid direct contact.
Design characteristics and components, heat transfer mechanism, flow config-
urations. The type of heat exchanger is primarily classified as industries used for
heating and cooling processes. This paper helps to study and analyze the transfer
of heat in different types of heat exchangers. Heat exchangers operate on the fun-
damental principles of laws of thermodynamics which explain the transfer of heat
from one fluid to another. Based on the zeroth law of thermodynamics, if the two
systems are in thermal stability with a third system, then the other two systems must
be in equilibrium with each other where all the systems are of the same tempera-
ture. If the heat flows interested in the system from the surroundings, then there is
an increase in the internal energy of the system and decrease in the energy of the
surrounding environment by the first law of thermodynamics. The entropy property
is established by the second law of thermodynamics which explains the natural and
invariable tendency of the closed thermodynamic system to increase the entropy over
time. The increased entropy moves the system to reach the state of equilibrium. The
proportion of the heat added or subtracted to the system to the absolute temperature
remains constant. Heat transfer is a fundamental concept in industries. Heat must be
added and removed or moved from one stage to another stage with the basic types
of heat transfer as conduction, convection, and radiation. Heat exchanger is a piece
of equipment to transfer heat from one medium to another medium separated by the
wall in order not to mix.
1.1 Heat Exchangers
The heat transfer and heat exchangers are widely used in the petroleum refineries,
natural gas processing and chemical or power plants. Heat exchanger has its appli-
cation in the process, petroleum industry, cryogenic, air-conditioning, refrigeration,
alternate fuels, and other industries. Based on the characteristics, we categorize it in
terms of shapes, sizes, models, and the flow direction of the fluids comparative to
each other. The categories are divided into different phases such as parallel, counter,
and cross flow. Parallel flow is applied when both types of the tube side fluid and
the shell side fluid flow in the same direction. The counter flow comes into existence
when the two fluids flow in opposite directions. The fluids enter at opposite sides of
heat exchangers. The concept of cross flow works when fluid flows are perpendic-
ular to each other. The wide variety of heat exchangers are regenerative, adiabatic
wheel, plate fin, fluid heat, waste recovery, dynamic scraped surface, plate, and shell-
tube types. The two types of heat exchangers are direct and indirect. In direct heat
exchanger, the two media are in direct contact with each other such that it is not
mixed together, whereas in the indirect type in cooperation the media are divided by
the wall through which the heat is transferred.
An Experimental Investigation of Heat Transfer … 777
1.2 Heat Exchanger Design
The heat exchanger design equation requires the entry of liquids into the heat
exchanger surface, the outlet temperature with the type, and the configurations of the
heat exchangers with parallel or opposite flow. The basic design equation is given
by Q = U ATlm where Q is clear as the rate of heat transfer between two fluids, U
is the overall heat transfer coefficient, A is the heat transfer surface area, and Tlm
is the log mean temperature difference which is calculated from the inlet and out-
let temperatures of the fluids. The three parameters are to be chosen effectively to
estimate the value of heat transfer in the heat exchanger. The log mean temperature
difference is calculated with the inlet and outlet temperature of the hot fluid and cold
fluid and is defined as
THin − TCout − THout + Ttin
Tlm = −TCout
ln TTHin
Hout −TCin
The heat transfer rate Q is calculated from the flow rate of one of the fluids
with the shell side fluid and the hot fluid as the tube side fluid with the values as
Q = m H Cρ H (THin − THout ) = m c Cρc (Tcout − Tcin ) where mass flow rates of hot and
cold fluids are given by m H , m c and heat capacities of hot and cold fluids are given
by Cρ H and Cρc , respectively. If the total heat transfer coefficient is taken as U, the
heat transfer through the wall separating the two liquids depends on the conductivity.
A heat exchanger consists of heat exchanging elements such as matrix containing
the heat transfer surface, inlet and outlet pipes, etc., whereas in the rotary regenerator
the matrix is driven to rotate at specified design speed.
The heat exchanger is classified based on the characterization of construction,
transfer process, degrees of surface compactness, flow arrangements, pass arrange-
ments, (6) fluids process, and heat transfer mechanisms. Based on the construction,
we have different types as tubular heat exchangers, double-pipe heat exchanger,
gasket, brazed, welded, spiral, and panel coil.
1.3 Heat Transfer in Heat Exchangers
The tubular heat exchangers are of double pipe, shell and tube, helically coiled
tubesplate heat exchangers are of gasket, brazed, welded, spiral, pane; coil, and
lamella. Tube fin and plate fin are of extended surface heat exchangers. Fixed and
rotary matrixes are of regenerators (Fig. 1).
Fig. 1 Helically coiled heat exchanger
2 Experimental
Figure 4 demonstrates the picture of the test segment. Figure 5 demonstrate the line
graph and photograph of the test arrangement. The exploratory arrangement has two
circles. The first is cone helically looped cylinder side which handles nanofluids. The
subsequent circle is the shell side which handles high-temp water. The subsequent
circle is associated with a capacity vessel with a size of 15 cm × 15 cm × 15 cm, with
a warmer 2 kW limit, attractive siphon, and indoor regulator. Cone wound cylinder
circle side is associated with a monobloc pump with 0.5 hp control, valve to control
the stream on cylinder side, test segment, cooling unit, and capacity vessel of six-
liter limit. The straight cylinder with fine sand is twisted to have the cone-like shape
with the assistance of wooden cone-shaped shape. The utilization of fine sand in the
cylinder is expected to maintain a strategic distance from the leveling of the copper
tube while bowing. The shell material is stainless steel, and the cone cylinder is
copper. The thermostat regulator is utilized to cut in and cut off the warmer. The inlet
and outlet temperatures are recorded by the fitted four K-type thermocouples with the
exactness of 0.1 °C. The stream passage impact is dodged by utilizing the quieting
area. U-tube mercury manometer is fitted over the cylinder with a precision of 1 mm.
The cooling unit of nanofluids handles water as a cooling medium. The shell external
surface is secured with asbestoses tape to lessen heat misfortune. Stream control is
finished with the valve, and the heated water from the cylinder is cooled by a cooling
unit (Fig. 2).
2.1 Helically Coiled Heat Exchanger
Helical coil heat exchangers are used in various processes because of its advantages
compared to other heat exchangers in view of simple configuration, easy integration
of the system, manufacturing and design, suitable for operating at high pressure and
low flow rates, cost-effectiveness. Figure 4 indicates the helical coil exchanger image
with winding of large number of small bore ductile tubes in the form of helix type
around a central core tube with many layers of tubes in radial and perpendicular axes.
With the low temperature where the heat transfer between more than two streams
desired, we use helical coiled heat exchanger than the other. Fouling deposit cleaning
is complicated because of small core tubes located on both sides. The overall heat
transfer coefficient in helical coil inlet and outlet is given by the following equations:
ki
hi = 2 ∗ Pr 0.33 ∗ 0.137 ∗ Reic0.7038 (1)
di + do

h 0 = ks ∗ Pr 0.3 ∗ 0.0175 ∗ Reic0.1 ∗ Ra 0.3276 /Deq (2)
9.81 ∗ β ∗ L 3 ∗ dt
Ra = (3)
α∗μ
where d stands for the diameter, k for thermal conductivity, μ for viscosity, β for
volumetric thermal expansion, L is the length of the coil, R for resistance, Deq as
equivalent diameter, and Pr for Prandtl number.
3 Result and Discussions
Figure 3 depicts the Dean versus the general heat transfer coefficient. It is seen that
the general warmth move coefficient increments with expanding the Dean number
and molecule volume fixation. The most extreme in general warmth move coefficient
is 63% at 0.6% nanofluid in the Dean number 4200. The general warmth move coef-
ficient is the impact of conduction and convection mode in the warmth exchanger.
While contrasting the conduction warmth move mode, the convection warmth move
is exceedingly viable than the conduction warmth move. Specifically, the inlet heat
transfer is exceptionally powerful because of more grounded convection flow along
with water. This leaves to improve progressively convective heat transfer charac-
teristics. The lower temperature distinction between the cylinder and shell is the
significant reason for improved convective heat transfer.
Figure 3 demonstrates the impact of molecule fixation on the heat transfer coef-
ficient. It is discovered that the expanding pattern in warmth move coefficient with
differing Dean number. The improved heat transfer coefficient is observed to be 14,
30, and 41% more than the water at 0.2, 0.4%, and water nanofluid separately. It
is unmistakably observed that the greatest heat transfer coefficient is acquired at
0.6% nanofluid. The expansion of more Al2 O3 expands the thermal conductivity of
nanofluids. Also, the expansion of Al2 O3 defers the arrangement of the heat limit
layer and makes the temperature profile smooth. Notwithstanding the deferring heat
limit layer, the lower relative speed of Al2 O3 with water particles along the bent
stream way is the purpose behind the higher inlet heat transfer coefficient at 0.6%
volume fixation.
Fig. 3 Dean number versus inner heat transfer coefficient

From Fig. 4, it is seen that the Nusselt number is improved by changing Dean
Number and molecule volume fixation. Increment in Nusselt numbers is observed to
be 30, 59 and 67% at 0.2, 0.4, and 0.6% Al2 O3 /water nanofluids individually when
looked at with water. The improvement is a direct result of the careful blending
of water particles and Al2 O3 ; this additionally might be the commitment Brownian
movement of the Al2 O3 . In addition, the arbitrary development of Al2 O3 exasperates
the limit layer development and the arrangements of auxiliary streams are escalated.
The Nusselt number is legitimately relative to the inward warmth move coefficient,
and along these lines Nusselt number increments with expanding the warmth move
coefficient and with expanding the volume fixation. The noteworthy Nusselt number
is the manner by which compelling the convective warmth move occurs. In this work,
the convective warmth move is profoundly compelling when expanding molecule
volume focus.
Figure 5 uncover the expanding pattern of weight drop by shifting molecule vol-
ume fixation and Dean Number. The weight drop of 0.2, 0.4, and 0.6% nanofluids are
observed to be 18, 33, and 44% higher than water separately. It is a direct result of
the higher thickness while including more Al2 O3 . It is seen that the greatest weight
drop happens at 0.6% nanofluid and at the Dean number 4200. It is clear that the
higher weight drop prompts the siphoning power punishment. The greatest weight
drop acquired is 44% when the outlet temperature of nanofluid is at 48 °C when it
leaves the cylinder. Nonetheless, the weight drop may change as for the nanofluids
outlet temperature. Higher the nanofluid temperature implies that lower the weight
drop. This is just on account of bringing down the thickness because of a higher
temperature (Fig. 5).
Fig. 4 Dean number versus Nusselt number

Fig. 5 Pressure drop
4 Conclusion
The paper focuses on the heat exchanger types and the efficiency of heat transfer
in each heat exchanger. In this paper, the fierce stream (2000 < De < 4000) heat
move attributes and weight drop of cone helically looped cylinder with Al2 O3 /water
nanofluid at 0.2, 0.4, and 0.6% molecule volume fixation have been tentatively con-
templated. It is discovered that the most extreme generally warmth move coefficient
of nanofluids is 52% higher than the water at 0.6% nanofluid in the Dean num-
ber 4200. The expansion in Nusselt numbers is observed to be 28, 52% also, 68%
at 0.2, 0.4, and 0.6% Al2 O3 /water nanofluid separately. In each type, we studied
the constructional and performance features of the heat exchangers. In this man-
ner, the conventional warmth move liquids might be supplanted with Al2 O3 /water
nanofluid at extensive weight drop in cone helically snaked cylinder heat exchangers.
Future work is required for examining the warm and stream practices of Al2 O3 /water
nanofluid with higher volume focus at various cones helically coil pitch.
References
1. Ke Y, Pei-Qi G, Yan-Cai S, Hai-Tao M (2011) Numerical simulation on heat transfer

characteristics of conical spiral tube bundle. Appl Therm Eng 31:284–292
2. Jaykumar JS (2011) Helically coiled heat exchangers. Heat exchangers—developments in heat
transfer, In Tech Publishers, pp 311–342
3. Shah RK, Sekulic DP (2003) Fundamentals of heat exchanger design. Wiley, New York
4. Minton P (1990) Process heat transfer. In: Proceedings of the 9th international heat transfer
conference, heat transfer (pp 355–362), Jerusalem, Israel, paper no. KN–2, 1
5. Taylor MA (ed) (1980) Plate-fin heat exchangers, guide to their specification and use. HTFS
(Harwell Laboratory), Oxon, UK
6. Dravid AN, Smith KA, Merrill EA, Brian PLT (1971) Effect of secondary fluid motion on
laminar flow heat transfer in helically coiled tubes. AICHE J 17(5):1114–1122
7. Varghese JA, Sreejith K, Ram TRS, Francis M, Mossas VJ, Nidhin MJ, Sushmitha S (August,
2015) Experimental investigation of a helical coil heat exchanger 5(8):01–05
8. Chinna Ankanna B, Sidda Reddy B (March, 2014) Performance analysis of fabricated helical
coil heat exchanger. Int J Eng Res 3(1):33–39, ISSN: 2319–6890
9. Bhuvaneswari S, Elatharasan G (2019) A study on the literature review on heat transfer in a
helically coiled heat exchanger. Int J Eng Res Technol 7(1):ISSN: 2278–0181, RTICCT
10. Bhuvaneswari S, Elatharasan G (March, 2019) Numerical methods of solving one dimensional
heat equation comparative study. J Appl Sci Comput VI(III):ISSN: 1076–5131
11. Indian Standard Specification for Shell and Tube Type Heat Exchangers (2007) BIS 2007, New
Delhi (IS: 4503–1967)
12. Dutta BK (2006) Heat transfer-principles and applications, 1st edn. PHI Pvt. Ltd., New Delhi
13. Couper JR, Penney WR, Fair JR, Walas SM (2005) Chemical process equipment: selection and
design, 2nd edn. Elsevier Inc.
14. Bhuvaneswari S, Elatharasan G (April, 2019) A study on the methods of solving heat equation.
J Appl Sci Comput VI(IV):ISSN: 1076–5131
Numerical Study of Heat Transfer
and Pressure Drop in a Helically Coiled
Tubes
Abstract The pressure drop and heat transfer are increased due to the secondary
flows caused by the centrifugal forces in a fluid through the curved pipes. The heat
transfer rate in a helically coil tubes is higher than that of the straight tubes. Due to its
compact structure, it has its wide application in industries. This chapter reviews on the
of heat transfer flow rates and pressure drop individuality using the parameters such
as Dean number, Reynolds number, Nusselt number, friction factor, and correlation.
Based on the data obtained experimentally from the literature, computation of heat
transfer and pressure fall is obtained by the mathematical equations. These parameters
are analyzed and tabulated for all the flows in the helical coil tube. The heat transfer
results have demonstrated that when utilizing an Al2 O3 , an expansion in warmth
move rate can be acquired when contrasted with heat transfer results got utilizing
straight heat transfer segments. It has reasoned that the expanded explicit heat of the
Al2 O3 just as the liquid elements in helical loop channels are the primary supporters
of the expanded heat move.
Keywords Heat transfer · Pressure drop · Helically coiled tubes
1 Introduction
Helically coiled heat exchanger has its wide and extensive use in all industries because
of the compact structure and high coefficient of heat transfer. The industrial appli-
cations are power generation, process plates, nuclear reactors, refrigeration, food
industries, and chemical industry process. The literature review studied by Patankar
et al. [1], Futagai and Aoyame [2], Abdulla [3], Bai et al. [4], Xin et al. [5], Jayaku-
mar and Grover [6] states that heat exchanger with helical coils has been used for the
S. Bhuvaneswari
University College of Engineering, Pattukkottai, Rajamadam 614701, India
G. Elatharasan (B)
Department of Mechanical Engineering, University College of Engineering, Pattukkottai,
Rajamadam 614701, India

https://doi.org/10.1007/978-981-15-4488-0_66
heat residual and removal systems where large nuclear reactor systems are mounted
and erected with the performance effectiveness of the parameters. They are also used
in steam generators and condensers in power plant as it has large surface area per
unit volume. The centrifugal forces produce the secondary flow which is typical to
the crucial direction of flow with the effects of circular motion and also increases
the friction factor and heat transfer. Present research work has been done to review
the intensity of the secondary flow developed within the tube that is performed of
the tube diameter and coil diameter. The performance of heat exchanger is improved
for the enhancement of heat transfer for which the size of the heat exchanger has to
be considerably decreased. The enhancement technique is classified as two groups
as dynamic and reactive. The active technique is need for some peripheral forces
like surface, fluid vibration, and field Special surface geometries, fluid additives are
thought-about the passive techniques. These two techniques have extensively used
in improving the heat transfer behavior. But in industries, helically coiled tube is
considered as the reactive heat transfer augmentation techniques due to its structure.
Due to the bend, the force ends up in the secondary flow enhancing the warmth
transfer rate that is employed in streamline flow [7].
Naphon [8] studied the most frequent uses, thermal performance, and pressure
drop of helically coiled tubes with and without the helical crimped fins. The literature
survey reveals that the heat transfer coefficient calculations are based on the geometry
features like tube diameter, coil diameter, pitch, etc.; heat transfer in the helical
coil is investigated for laminar and turbulent flow with constant wall flux boundary
conditions. Janssen and Hoogendoornh [9] studied the improvement techniques of
helical coil heat exchangers over the others. The flow pattern between the stratified
and annular transition was studied by Whalley [10]. Relaminarization was studied
by Sreenivasan and Strykowski [11]. The review of heat transfer and pressure drop
in a helical coils with two-phase flow in the horizontal helical pipes was studied by
Berger et al. [12], Awwad et al. [13], Xin RC [5], and Shah and Joshi [14]. Jayakumar
and Grover [6] and Vimalkumar et al. [15] studied the pressure drop and heat transfer
in tube-in-tube helical heat exchanger with the turbulence condition.
Rennie and Raghavan [16] studied experimentally the parallel and counter flow
configurations with the double-pipe helical heat exchanger. Annulus and inner tube
had the varied flow rates and temperature for which overall heat transfer coefficients
were determined. The experimental study done by Prabhanjan et al. [17] determined
the relative advantage of helically coil over straight tube for liquid heating using the
boundary conditions. Shokoummand et al. [18] had one investigation on the curvature
ratio and different coil pitches with parallel and counter flow configurations with
measurements of inlet and outlet temperature of tube-side, shell-side fluids, and flow
rate of the fluids. Using Wilson plots, overall heat transfer coefficients were calculated
and compared with the inner Nu number and mixed convection heat transfer in a
coil-in-shell type [19]. Experiments were conducted for the steady state by varying
Reynolds and Rayleigh numbers with different tube-to-coil diameters, pitch of the
coil, mass flow rate of shell-side and tube-side over the performance coefficient, and
effectiveness of the helical coil tube for the laminar and turbulent flows.
Numerical Study of Heat Transfer and Pressure Drop … 787
High-temp water and cold water were provided to the shell side and the cylinder
side, individually, to check for spillages and test the exactness of thermocouples
and indoor regulator. Boiling water was coursed to the shell side at a consistent
stream rate. The nanofluid with 0.2% molecule volume fixation was circled through
the cylinder side. The comparing perceptions were made. The cylinder-side stream
was fluctuated to achieve the predetermined Dean number by utilizing a valve plan
[20, 21].
1.1 Helical Coil Characteristics
The representation drawing of helical coil is given in Fig. 1. The inner pipe has the
diameter 2*r and the coil diameter is 2*R*c where c is the distance between the
centers of the pipes with H as the detachment between two neighboring turns. The
curvature ratio δ is defined as the fraction of pipe diameter to the coil diameter agreed
by r/Rc. The coil diameter is also defined as the pitch circle diameter. The ratio of
pitch to the developed length of one turn is called as non-dimensional pitch as λ =
H/2nRc. The symbol α is called helix angle which is termed as the projection of the
coil on the plane perpendicular passing through the axis of the coil. The Reynolds
number and Dean number are used to evaluate the flow characteristics.
Fig. 1 Helical coil

schematic diagram
2 Experimental Setup
Figure 2 delineates the plan of the test arrangement. The setup has shell-side circle
and helical curled cylinder-side circle. Shell-side circle handles boiling water. Helical
wound cylinder circle handles Al2 O3 /water nanofluid. The shell-side stream and the
looped cylinder-side stream are in counter stream setup. Shell-side circle comprises
a capacity vessel with a radiator of 1.75 kW limit. The cylinder-side circle comprises
of monoblock siphon, valve to control the stream on the cylinder side, test area,
cooling unit, and capacity vessel of five-liter limit. The helical cylinder is comprised
of copper, and shell is comprised of mellow steel. The temperature of the heated
water in the shell-side stockpiling vessel is kept up by an indoor regulator [22, 23].
Four ‘K’- type thermocouples of 0.1 °C exactness are utilized to quantify the channel
and outlet temperatures of shell and cylinder side. Four ‘K’- type thermocouples of
0.1 °C exactness are put on the external surface of the looped cylinder to gauge the
cylinder divider temperatures. U-tube mercury manometer is put over the helical
cylinder to quantify the weight drop. The shell is protected with fiber fleece. A valve
is given in the stream pipe associating the cooler area and the repository for stream
rate estimations and cleaning the framework between trial runs.

2.1 CFD Methodology
A geometric model was created utilizing CATIA v5, and this plan was sent out to
the CFD module in ANSYS v12 to get the limited component model to the required
examination. Cross section was done in CFD—Fluent utilizing similar programming
as appeared in Fig. 3. The complete number of nodes and components acquired for a
helically coiled heat exchanger were 359,624 and 2,175,794 individually. The admin-
istering conditions were explained utilizing limited volume technique with Compu-
tational Fluid Dynamic (CFD) programming ANSYS Fluent [24]. The arrangement
depended on weight revision strategy utilizing the SIMPLE calculation. Kumar et al.
explored the warm conduct of Al2 O3 /water nanofluid in a helically wound cylinder
heat exchanger by utilizing CFD model.
The motivation behind this paper is to display the aftereffects of an examination

concerning the general variety of heat transfer coefficient around the in-line cluster
with helical loop cylinder shapes. The numerical examination considers the impact
of nanofluid, for example, Al2 O3 on the stream and heat transfer qualities of cylin-
der banks in a physical space for various Nusselt numbers. In this study, the heat
Fig. 3 Heat transfer coefficient

transfer rate, surface temperature, Nusselt number, thermal resistance, power con-
sumption, and reliability of heat exchanger by using Al2 O3 /water nanofluids have
been numerically studied by single-phase approach.
3.1 Heat Transfer Coefficient
The fundamental goal is to increase the heat transfer coefficient in the helically coiled
heat exchanger, and Fig. 3 demonstrates the upgrade of heat transfer coefficient when
expanding the channel liquid speed and molecule volume divisions. It was watched
that heat transfer coefficient increments with expanding inlet fluid velocity. The heat
transfer coefficient was observed to be 15, 26, and 41% higher than the water at
0.2, 0.4, and 0.6% separately volume division of Al2 O3 /water nanofluids. This is
because of the higher warm conductivity and Brownian movement of nanoparticles.
The movement of particles lessens the warm limit layer and improves the warmth
move limit. A noteworthy heat transfer coefficient of 42% was accomplished at 0.6%
volume portion which brought about better drop in surface temperature in the heat
exchanger [25].
Figure 4 demonstrates the improvement of Nusselt number by fluctuating
Reynolds number and Al2 O3 nanoparticles volume parts. The Nusselt number
expanded with a comparing increment in Reynolds number, and the Nusselt number
was least at Re of 200 and most extreme at Re of 600. Further, the heat exchanger
with higher nanoparticles volume fractions of nanofluids has a higher Nusselt number
Fig. 4 Nusselt number

since when more nanoparticles take part in the nanofluid it can prompt a successful
thermal conductivity. The Nusselt number is increased by 12, 27, and 45% at 0.2, 0.4,
and 0.6% of Al2 O3 /water nanofluids individually when compared with water [26].
The expansion of Nusselt number to 32% is significant at 0.6% molecule volume
portion on the grounds that the heat coefficient of nanofluid was higher than water.
The pressure drops are crosswise over the heat exchanger from inlet to outlet
utilizing ANSYS Fluent. Here, the heat transition was connected on the base of
the inner wall surfaces [27]. The stream rate of bay water and Al2 O3 -nanofluids is
0.02–1 m/s. The pressure drops over the inlet and outlet of the direct effect were
appeared in cross-sectional area utilizing ANSYS Fluent. It was plainly seen that
there were two locales with low temperature and pressure drop (blue in shading) in
the inlet, high temperature, and pressure drop (red in shading) in the channel outlet.
The pressure drop accomplished is 3240 Pa. The weight drop emerges due to the
frictional impact of nanoparticles. The pressure drop of Al2 O3 -nanofluids expanded
with increment in volume focuses. Kumar et al. [24] researched tentatively and
numerically the warmth move and weight drop attributes of microchannel warmth
sink by utilizing half breed Al2 O3 –TiO2 nanofluids (two-stage blend model). The
outcomes demonstrated that crossover nanofluids yield no synergetic impact in heat
transfer while blending different nanoparticles with comparable shape and size.
3.2 Governing Equations for Heat Transfer and Pressure

Drop in a Helical Coil
The heat transfer rate can be calculated as follows:

T1 − T2
Q = U ATlog = U A
T1
ln T2
where Q is heat transferred, U is the overall heat transfer, A is the total heat transfer
area, and Tlog is log temperature difference or LMTD which is based on the fluid
properties. Using control volume method, the discretization is done with change in
overall enthalpy with division of equal differences h with N control volumes.
Q i = m hf chf (Thf,bulk,i − Thf,bulk,i+1 )

Q i = m hf (Thf,bulk,i − Thf,bulk,i+1 ) (1)
Q i = ḣ hf Aout,i (T̈hf,bulk,i − T̈hf,wall,i+1 )

Q i = ḧ hf Ain,i (T̈wf,wall,i − T̈wf,bulk,i+1 ) (2)
T̈hf,bulk,i = Thf,bulk,i + Thf,bulk,i+1 /2 (3)

T̈hf,wall,i = Thf,wall,i + Thf,wall,i+1 /2 (4)
T̈w f,bulk,i = Tw f,bulk,i + Tw f,bulk,i+1 /2 (5)
T̈hf,wall,i = Twf,wall,i + Twf,wall,i+1 /2 (6)

Thf,bulk,i − Twf,bulk,i+1 − Thf,bulk,i − Twf,bulk,i+1
LMTDi = Thf,bulk,i −Twf,bulk,i+1
(7)
ln Thf,bulk,i −Twf,bulk,i+1

U̇out,i = ḧ 1 + ḧ 1 dd0i + d20 ln dd0i /λtube where Q i, m, T, U, c, λ are denoted as
hf,i wf,i
heat transfer, mass flow rate, temperature of the fluid, overall heat transfer, specific
heat, and thermal conductivity.
The symbols h, A, d, and i denote the enthalpy, area of the surface, tube diameter,
and the iteration. The suffixes wf and hf denote the working and heating fluids, respec-
tively. The governing equations are obtained by applying the boundary conditions as
follows.
The continuity equation is defined as
∂(ρu) ∂(ρv) ∂(ρw)

+ + =0 (8)
∂x ∂y ∂z
Navier–Stoke equation is given for the fluid flow as

∂u ∂v ∂w ∂ρ 1 ∂ ∂u ∂v ∂w
ρ u +v +w = ρX − + μ + + + μ∇ 2 μ
∂x ∂y ∂z ∂z 3 ∂x ∂x ∂y ∂z
(9)
The energy equation is given by

∂T ∂T ∂T ∂ρ ∂ρ ∂ρ
ρc p u +v +w = u +v +w + k∇ 2 T + μρ (10)
∂x ∂y ∂z ∂x ∂y ∂z
k ∂T
The heat transfer coefficient is given by h = Tw −T
∂x
f
The Reynolds number, Dean number, and Nusselt number are given by
ρV d √
Re = , De = Re ∗ δ, Nu = 0.021 ∗ Re0.85 Pr0.6 δ 0.1
μ
0.46
The critical Reynolds number is given by Recr = 2300 1 + 8.6 Dd
where d, D, L, V, k, and μ are the tube diameter, coil diameter, length of the coil,
volume of the fluid, thermal conductivity, and viscosity, respectively. The friction
factor is given by f = ρ2∇L Vpd2 .
The graph (Fig. 2) shows the relationship with Nu, Re with the mass flow rate
(kg/s) (Figs. 5 and 6).
Relationship of Re, Nu with mass flow rate Figs. 2 and 3.
The largely heat transfer coefficient and the effectiveness with the study of Nu
and Re number are analyzed in the graph Fig. 5. The diagram reveals that Nu number
increases with the increment of mass flow rate and increase in the ratio of curvature
of the helical coil. The curvature ratio is changed from 0.05 to 0.06 at the mass flow
rate of 0.004 kg/s for the increment of Nusselt number of 3 and 5%. The Reynolds
number is increasing with the mass flow rate at different coils. The values of Re,
Nu are increases with the increases in turbulence. Figure 7 shows the contours of
temperature variations respective of the different velocities.
Fig. 5 Re versus mass flow rate
50
0
1 2 Mass flow rate(kg/sec)
3 4 5 6 7 Mass flow rate(kg/sec)
8 9 10 11 12
1 2 3 4 5 6 7 8 9 10 11 12
Mass flow rate(kg/sec) 0.002 0.0023 0.0026 0.0029 0.0032 0.0035 0.0038 0.0041 0.0044 0.0047 0.005 0.0053
Nu 5 7 9 11 13 15 17 19 21 23 25 27
Mass flow rate(kg/sec) 0.002 0.0023 0.0026 0.0029 0.0032 0.0035 0.0038 0.0041 0.0044 0.0047 0.005 0.0053
Nu 5 7 9 11 13 15 17 19 21 23 25 27
Fig. 6 Nu versus mass flow rate

Fig. 7 Inlet velocity 0.02 and mass flow rate 0.004 kg/s
4 Conclusion
Numerical simulation has been researched on heat transfer qualities and pressure
drop of Al2 O3 /water nanofluid in a minimized heat exchanger with helical loop
cylinder shapes and an in-line arrangement of cylinders under consistent state lami-
nar liquid stream. The numerical outcomes uncover the enhancement in heat transfer,
concerning the base liquid, recognized to portray nanofluid. The paper focuses on the
literature studies of the heat transfer and pressure drop in a helical coil tube with the
varying parameters. The analysis has been done for the constant heat flux boundary
conditions and constant wall temperature. The resulting heat transfer and pressure
drop are optimized by choosing the parameters appropriate. Studies were discussed
by examining the influence on heat transfer and pressure drop in the helical coil.
The centrifugal force caused by the curvature of the pipe makes the heavier fluid
to flow in the outer side with the high velocity and temperature which has greater
influence on the drop in the temperature and pressure. Temperature drop remains the
maximum for low flow rate and decreases for the increasing flow rate, whereas the
pressure drop is directly proportional to the flow rate. It has been revealed that heat
transfer coefficient is increased with expanding the inlet speed and corresponding
heat transfer coefficients were increased by 15, 26, and 41% at 0.2, 0.4, and 0.6%
of Al2 O3 /water nanofluids. It is clearly observed that the molecule volume division
individually separated when contrasted and refined water. Nanofluid containing few
nanoparticles has considerably higher warmth move coefficient than those of base liq-
uids. The Nanofluid containing few nanoparticles have considerably higher warmth
move coefficient than those of base liquids. The expansion in the volume fractions
of nanoparticles is heightened with the cooperation and impact of nanoparticles.
References
1. Patankar S, Pratap VS, Spalding DB (1974) J Fluid Mech 62:539–551

2. Futagami K, Aoyama Y (1988) Laminar heat transfer in helically coiled tubes. Int J Heat Mass
Transfer 31:387–396
3. Abdulla MA (1994) A four-region, moving-boundary model of a once through, helical coil
team generator. Ann Nucl Energy 21(9):541–562
4. Bai B, Guo L, Feng Z, Chen X (1999) Turbulent heat transfer in a horizontally coiled tube.
Heat Transfer-Asian Research 28(5):395–403
5. Xin RC, Awwad A, Dong ZF, Ebadian MA (1996) An investigation and comparative study of
the pressure drop in air-water two-phase flow in vertical helicoidal pipes. Int J Heat and Mass
Transfer 39(4):735–743
6. Jayakumar JS, Grover RB (1997) Two phase natural circulation residual heat removal. In:
Proceedings of 3rd ISHMT-ASME heat and mass transfer conference, Kanpur
7. Jayakumar JS, Mahajani SM, Mandal JC, Vijayan PK, Bhoi R (2005) Experimental and CFD
estimation of heat transfer in helically coiled heat exchangers. Chem Eng Res Des 86(3):221–
232
8. Naphon P, Wongwises S (2006) A review of flow and heat transfer characteristics in curved
tubes. Renew Sustain Energy Rev 10:463–490
9. Janssen LAM, Hoogendoorn CJ (1978) Laminar convective heat transfer in helical coiled tubes.
Int J Heat Mass Transfer 21:1197–1206
10. Whalley PB (1980) Air-water two-phase flow in a helically coiled tube. Int J Multiph Flow
6:345–356
11. Sreenivasan KR, Strykowski PJ (1983) Stabilization effects in flow through helically coiled
pipes. Exp Fluids 1:31–36
12. Berger SA, Talbot L, Yao LS (1983) Flow in curved pipes. Ann Rev Fluid Mech 15:461–512
13. Awwad A, Xin RC, Dong ZF, Ebadian MA, Soliman HM (1995) Measurement and correlation
of the pressure drop in air-water two-phase flow in horizontal helicoidal pipes. Int J Multiph
Flow 21(4):607–619
14. Shah RK, Joshi SD (1987) Convective heat transfer in curved ducts (Chapter 3). In: Kakac S,
Shah RK, Hung W (eds) Handbook of single-phase convective heat transfer. Wiley, New York
15. Kumar V, Saini S, Sharma M, Nigam KDP (2006) Pressure drop and heat transfer in tube-in-tube
helical heat exchanger. Chem Eng Sci 61:4403–4416
16. Rennie TJ, Raghavan VGS (2006) Numerical studies of a double-pipe helical heat exchanger.
Appl Therm Eng 26:1266–1273
17. Prabhanjan DG, Raghavan GS, Rennie TJ (2005) Comparison of heat transfer rates between
a straight tube heat exchanger and a helically coiled heat exchanger. Int Commun Heat Mass
Transfer 29(2):185–191
18. Shokouhmand H, Salimpour MR, Akhavan-Behabadi MA (2007) Experimental investigation
of shell and coiled tube heat exchangers using Wilson plots. Int Commun Heat Mass Transfer
35(1):84–92
19. Salimpour MR (2008) Heat transfer characteristics of a temperature-dependent-property fluid
in shell and coiled tube heat exchangers. Int Commun Heat Mass Transfer 35(9):1190–1195
20. Cioncolini A, Santini L (2006) An experimental investigation regarding the laminar to turbulent
flow transition in helically coiled pipes. Exp Thermal Fluid Sci 30:367–380
21. Seban RA, McLaughlin EF (1963) Heat transfer in tube coils with laminar and turbulent flow.
Int J HeatMass Trans 6:387–495. Int J Multiphase Flow 21(4):607–619
22. Ghorbani N, Taherian H, Gorji M, Mirgolbabaei H (2009) Experimental study of mixed con-
vection heat transfer in vertical helically coiled tube heat exchangers. Exp Thermal Fluid Sci
34(7):900–905
23. Chen CN, Han JT, Jen TC, Shao L, Chen WW (2010) Experimental study on critical heat
flux characteristics of R134a flow boiling in horizontal helically-coiled tubes. Int J Therm Sci
50(2):169–177
24. Kumar V, Sarkar J (2019) Numerical and experimental investigations on heat transfer and
pressure drop characteristics of Al2O3-TiO2hybrid nanofluid in minichannel heat sink with
different mixture ratio. Powder Technol. https://doi.org/10.1016/j.powtec.01.061(2019)
25. Bhuvaneswari S, Elatharasan G (2019) A study on the methods of solving heat equation. J
Appl Sci Comput 6(4). ISSN:1076-5131
26. Bhuvaneswari S, Elatharasan G (2019) Numerical methods of solving one dimensional heat
equation comparative study. J Appl Sci Comput 6(3). ISSN:1076-5131
27. Bhuvaneswari S, Elatharasan G (2019) A study on the literature review on heat transfer in a
helically coiled heat exchanger. Int J Eng Res Technol 7(01). ISSN:2278-0181
Tribology, Wear and Surface
Engineering
Comparative Study on the Mechanical
Performance of Solid Lubricants Over
Peek Polymer
N. Venkatesh and P. K. Dinesh Kumar
Abstract Polyetheretherketone (PEEK) and its composites have an extended range

of applications and are continuously replacing metals in aerospace, automobile, and
mechanical industries for structural and tribological applications due to its robust-
ness, good mechanical and chemical resistance properties. The mechanical proper-
ties of virgin PEEK, PEEK reinforced with 10% Polytetrafluoroethylene (PTFE) and
PEEK reinforced with 10% hexagonal Boron Nitride (h-BN) are compared and ana-
lyzed in this study. The composites were fabricated through the twin-screw extrusion
method and were then characterized by EDAX analysis. The static mechanical prop-
erties such as compressive, flexural and hardness of all the specimens were tested
as per ASTM standards, and it was identified that the virgin PEEK material had
better mechanical properties in comparison to materials with solid lubricants. Also,
the addition of h-BN to PEEK had a more drastic reduction in mechanical strength
compared to the addition of PTFE to PEEK. It was seen that by adding the solid
lubricants, the mechanical properties of virgin PEEK material deteriorated.
Keywords PEEK · PEEK-PTFE · PEEK-hBN · Mechanical properties
1 Introduction
Polymers are viscoelastic materials that can exhibit properties of solids and liquids.
Polyetheretherketone (PEEK) is a colorless organic thermoplastic polymer in the
Polyaryletherketone (PAEK) family. It is semi-crystalline with excellent mechanical
and chemical resistance properties that are retained to high temperatures. Due to its
robustness, several engineering applications in aerospace, automobile, and mechan-
ical industries are using it and are also finding its high mechanical strength, high
N. Venkatesh (B) · P. K. Dinesh Kumar

Department of Aeronautical Engineering, Rajalakshmi Engineering College, Chennai 602105,
India
P. K. Dinesh Kumar

https://doi.org/10.1007/978-981-15-4488-0_67
800 N. Venkatesh and P. K. Dinesh Kumar
thermal resistance, and high wear resistance to support its usage. However, in its
pure form, the friction coefficient is relatively high in dry sliding, which tends to
restrict its applications in terms of Tribology. To minimize the friction coefficient
of PEEK polymer particular solid lubricants are incorporated in it such as Polyte-
trafluoroethylene (PTFE) and hexagonal Boron Nitride to increase its applications
in various industries. Few studies show that PEEK reinforced with PTFE and PEEK
reinforced with h-BN tend to reduce the overall friction coefficient of the material.
Dry lubricants or solid lubricants are materials despite being in the solid-state have
outstanding lubrication properties and hence have less friction and wear rate. It can
reduce friction and thus wear between two surfaces sliding against each other with-
out the requirement of a liquid oil medium. Solid lubricants like Graphite, PTFE,
hexagonal Boron Nitride and Tungsten disulphide are mostly used [1]. These mate-
rials provide lubrication at higher temperatures than those of oil—and liquid-based.
These materials can operate to about 350 °C in case of oxidizing environments and
even further in a reducing environment. The reduced friction properties of solid
lubricants are due to a layered structure on the molecular level with weak bonding
between the layers [2, 3]. These layers slide over each other with minimal applied
force and also have low friction properties [4]. Morphological, physical, friction and
wear behavior of h-BN filled PEEK composite coatings with varying percentage of
h-BN content from 0% to 8% by flame spray coating technique for different particle
sizes were studied. It was noted that although adding h-BN increased the microhard-
ness irrespective of the size, deviation in terms of crystallinity did exist. It was noted
that by adding h-BN particles to the material reduced the specific wear rate and sim-
ilarly by adding more amount of h-BN the friction coefficient was reduced [5]. The
friction coefficients and wear rates of neat PEEK and h-BN/PEEK coatings by ther-
mal spray coating was studied and tested under varying temperature (30 °C–300 °C).
The results suggested that friction and wear rate for PEEK coatings were relying on
the testing temperature rather than h-BN content. Overall, the operating range of 100
and 200 °C was found to be the best for low wear and friction coefficients. No kind
of chemical change was observed with this work [6]. Explained about using a com-
mon parent material and changing the dry lubricant reinforced with it. The different
solid lubricants used were hexagonal boron nitride, Potassium Titanate, Boric acid,
and Mica. These dry lubricants were incorporated with a common parent material
into a composite by injection molding. All the composites were tested under similar
conditions, and the best solid lubricant based on wear rate was found to be KT (Potas-
sium Titanate). Besides, the reduction of hardness in materials due to the addition of
the fillers could be due to the surge in the plasticity of the material which tends to
deformation of the material [7, 8]. Explained how the amount of filler content to a
parent material affected the overall properties of new materials such as mechanical
and tribological properties. Based on the parent material characteristics, the filler
used, quantity of filler used, filler size and whether it is a coating or reinforcement,
the final properties of the new material will vary accordingly [9]. Indicates that the
addition of PTFE to PEEK decreases all the strength properties except impact.
Comparative Study on the Mechanical Performance … 801
2 Fabrication of Composites and Its Characterisation
Commercially available PEEK of grade 450(G) with an average diameter of 100 µm,
PTFE powder of particle size 60 µm and h-BN powder in microform was provided
by SRL chemicals. PEEK, and solid lubricants PTFE and h-BN powders of 90% by
weight and 10% by weight respectively were combined using a twin-screw extruder
at around 400 °C with a constant screw speed of 120 rpm [10–12].
The manufacturing process of the specimens occurs in two stages.
In the initial stage, ultrasonic vibration was used along with alcohol as a mixing
medium, and then the dispersed solution was dried at around 100 °C in a hot drier for
approximately 3 h to remove the excess alcohol and moisture that may be present.
Next, the blended mixture was poured into a twin-screw extruder at 400 °C and a
screw speed of 120 rpm.
Finally, the extruded pellets were cut as small granules with the help of pallet
cutter machines. The extruded granules were placed in a hot electric furnace at
around 80 °C up to 60 min to remove any moisture present. The dried granules
underwent injection molding at approximately 20 MPa and a temperature of about
380–400 °C for approximately 3 h. Composites PEEK-PTFE and PEEK-h-BN were
cooled down to room temperature. The different compositions for the composites
are given in Table 1. Figure 2a–b shows the schematic assemblage of composite S2
and S3 and the powders used in manufacturing are shown in Fig. 1a–c.
Table 1 Composite Composition

Composites PEEK (wt%) PTFE (wt%) H-BN (wt%)
S1 100 0 0
S2 90 10 0
S3 90 0 10
Source: computed
Fig. 1 a PEEK powder (100 µm). b PTFE powder (60 µm). c h-BN powder (5 µm)
Fig. 2 a Schematic fabrication of PEEK-PTFE. b Schematic fabrication of PEEK-hBN
The characterization of the two specimens with dry lubricants was done with the
help of the Energy Dispersive Spectrum Analysis (EDAX). From the EDAX test, the
characteristic crests of chemical elements are observed in S2 and S3. In composite
S2, the unique peak of F, i.e. Fluorine shows the presence of 10% of PTFE with
PEEK matrix. In composite S3, the typical peak of N, i.e. Nitride shows the presence
of 10% of h-BN with PEEK matrix.
The observation on the dispersion of the solid lubricants was noted through a
Scanning Electron Microscope (SEM) and captured. It is found that the majority of
the PTFE and h-BN powder particles were homogeneously mixed throughout the
studied samples.
From Fig. 3a, the amount of Carbon and Oxygen present in the sample indicates
the compound PEEK (S1). The presence of Fluorine is due to the addition of NaF
as a catalyst during the manufacturing of PEEK polymer. The result is similar to the
amount which was found in the literature survey [7, 13, 14]. Also, Fig. 3b indicates
the PEEK-PTFE (S2) composite. From the figure, the amount of Fluorine present in
the sample indicates the presence of PTFE in the right quantity [9, 13]. Similarly,
from Fig. 3c, the amount of Nitrogen present in the sample indicates the existence
of Boron Nitride in the appropriate quantity as reinforced.
Figure 4a–b shows the PEEK with h-BN at magnification 10 k (5 µm) and PEEK
with PTFE at magnification 15 k (3 µm) respectively. From these figures, we can
observe the distribution of h-BN particles and PTFE particles in the PEEK polymer.
The distribution of h-BN and PTFE in PEEK polymer is uniform. The small white
dots in the SEM images are the solid lubricants hBN and PTFE distributed in the
PEEK polymer, respectively.
The measure of the hardness of the different composites was compared through
Shore D hardness test with the help of Shore D Durometer as per ASTM standard
ASTM D2240 (Table 2).
The Shore D hardness test revealed that the hardness of PEEK-PTFE was 14%
more than that of the hardness of PEEK-hBN. This result could be due to the increased
plasticity of the PEEK material due to the addition of dry lubricants. The comparison
of the hardness for different specimens is given in the Figure (Fig. 5).
3 Testing Procedures and Conditions for Mechanical Study
After molding the exhibits, the test pieces of different proportions were machined
from the molded exhibits for studying the tribological and mechanical properties
as per the ASTM standards. The static mechanical properties were found out by
compression and flexural tests on the specimens S2 and S3.
The Flexural tests were carried out on cuboidal specimens of dimensions 65 mm
length(l) × 13 mm width(w) × 3 mm thickness(t) as per ASTM standard D 790 were
machined out at room temperature (25 °C) and tested in a universal testing machine
with a span length of 48 mm at a constant feed rate of 5 mm/min as flexural fixtures.
The average value for all the specimens was taken into account.
The Compression tests were carried out on cylindrical specimens of dimensions
30 mm length(l) × 10 mm diameter(d) were machined from the composite sample
as per the ASTM standard D 695 at room temperature(25 °C) and tested with the
universal testing machine at a constant feed rate of 0.5 mm/min as the compressive
fixtures. The average value for all the specimens was taken into account.
Fig. 3 a EDAX analysis of PEEK (S1). b EDAX analysis of PEEK-PTFE (S2). c EDAX analysis
of PEEK-hBN (S3)
a b
hBN particles PTFE particles
Fig. 4 a SEM image of PEEK-hBN at 5 µm. b SEM image of PEEK-PTFE at 3 µm
Table 2 Hardness
Composites Shore D hardness
Comparison
S1-PEEK 88
S2-PEEK-PTFE 86
S3-PEEK-hBN 73
Fig. 5 Comparison of Shore

D’ Hardness for different
specimens
4 Mechanical Behavior
The compression strength for the composites reduced with the addition of solid
lubricants. For pure PEEK, the compression strength was found to be the maximum
at 143 MPa, and it decreased by 9% with the addition of solid lubricant PTFE with
131 MPa and it reduced by 37.5% due to the addition of h-BN with 104 MPa. The
addition of the fillers tend to weaken the bonds present in the material thereby leading
to a reduction in the compression strength. The comparison of the compression
strength for the different specimens is shown in the figure (Fig. 6).
compression strength
flexural strength
The flexural strength too reduced with the addition of solid lubricants to PEEK
polymer. The flexural strength of the PEEK composite was found to be the most at
210 MPa, and it decreased by 43% with the addition of PTFE with 119 MPa and
reduced by 65% with the addition of h-BN with 73.4 MPa. The forces of attraction
within the materials were considerably lower than the original material thus causing
a reduction in the Flexural strength. The comparison of the flexural strength for the
different specimens is shown in the figure (Fig. 7).
5 Conclusion
The effect of reinforcement of solid lubricants PTFE and hBN at 10 wt% to PEEK
polymer were investigated in this work. When 10 wt% of either solid lubricant was
reinforced with PEEK polymer, the static mechanical properties of the composite
were considerably lower than that of neat PEEK. The materials had significantly
lower Compression margin depending on the solid lubricant which was reinforced.
The major possibility for the reduction in the mechanical properties of the composites
is due to the agglomeration of the solid lubricant particles in the PEEK polymer.
Due to the excess amount of solid lubricant particles present, proper bonds between
the molecules of the composite could not be formed. The solid lubricants added had
distressed mechanical properties than PEEK. Due to this reason, the addition of these
lubricants brought about a reduction in mechanical properties. In comparison, the

mechanical properties of h-BN are significantly lower than that of PTFE. That’s why
when comparing the two solid lubricants with PEEK polymer with the same weight
percentage, PEEK-PTFE had better hardness, Compressive and Flexural strength
compared to PEEK-hBN. A reason for this could be due to the weaker bonds among
the h-BN particles and also among the h-BN particles and PEEK material.
References
1. Wang Q, Wang Y, Wang H, Fan N, Yan F (2016) Experimental investigation on tribological

behaviour of several polymer materials under reciprocating sliding and fretting wear conditions
2. Zhang G, Wetzel B, Wang Q (2015) Tribological behaviour of PEEK-based materials under
mixed and boundary lubrication conditions. Tribol Int 153–161
3. Qu M, Yao Y, He J, Ma X, Feng J, Liu S, Hou L, Liu X (2017) Tribological study of polyte-
trafluoroethylene lubricant additives filled with Cu microparticles or SiO2 nanoparticles. Tribol
Int 57–65
4. Tharajak J, Palathai T, Sombatsompop N (2015) Morphological and physical properties and
friction/wear behaviour of h-BN filled PEEK composite coatings
5. Tharajak J, Palathai T, Sombatsompop N (2017) Recommendations for h-BN loading and
service temperature to achieve low friction coefficient and wear rate for thermal sprayed PEEK
coatings
6. Narayan Panda J, Bijwe J, Pandey RK (2017) Comparative potential assessment of solid
lubricants on the performance of poly aryl ether ketone (PAEK) composites
7. Kalin M, Zalaznik M, Novak S (2015) Wear and friction behaviour of poly-ether-ether-ketone
(PEEK) filled with graphene, WS2 and CNT nanoparticles
8. Gan D, Lu S, Song C, Wang Z (2001) Physical properties of polyether ketone ketone/mica
composites: effect of filler content
9. Bijwe J, Sen S, Ghosh A Wear—(Dec 2004) Influence of PTFE content in PEEK- PTFE blends
on mechanical properties and tribo performance in various wear modes
10. Liu L, Xiao L, Li M, Zhang X, Chang Y, Shang L, Ao Y (2016) Effect of hexagonal boron
nitride on high-performance polyether ether ketone composites
11. Kurdi A, Wang H, Chang L (2017) Study on the effect of nanosized TiO2 addition on tribological
behaviour of polyether ether ketone composite; Tribol Int
12. Rodriguez V, Sukumaran J, Schlarb AK, De Baets P (2016) Influence of solid lubricants on
tribological properties of polyetheretherketone (PEEK). Tribol Int 45–57
13. Friedrich K, Zhang Z, Schlar AK (July 2005) Effects of various fillers on the sliding wear of
polymer composites
14. Díez-Pascual AM, Naffakh M, González-Domínguez JM, Ansón A, Martínez-Rubi Y, Teresa
Martínez M, Simard B, Gómez MA, High-performance PEEK/carbon nanotube composites
compatibilized with polysulfones-II. Mechanical and electrical properties
ZrB2 Influences on the Dry Sliding Wear
Resistance of AA7075 Alloy
P. Loganathan, A. Gnanavelbabu, K. Rajkumar, and S. Ayyanar
Abstract This article presents the influence of ZrB2 particles on the mechanical
and wear behavior of AA7075 composites. Variations in AA7075 reinforced with
ZrB2 particles varied in the range of 5–15 wt% and fabricated using of the two-
step stir casting technique. Fabricated composites were subjected to T6 heat-treated
and analyzed microstructure, hardness, impact, tensile strength and dry sliding wear
behavior. Tribo test was conducted with varying applied pressure (0.39–1.59 MPa),
sliding velocities of (0.8–2.0 m/s), and a sliding distance of 3.0 km. The results
revealed the distribution of particles in the matrix materials. ZrB2 increased the
hardness of the composites from 98 to 128 Vickers hardness and reduced the impact
strength up to 44%. The addition of 15 wt% of ZrB2 particles increased the tensile
strength of AA7075 alloy by 242 MPa. Wear resistance of the composites increased
with increasing wt% of ZrB2 particles. Wear mechanism of the worn surface was
studied through use of a scanning electron microscope (SEM) and revealed the for-
mation of an oxide layer on the worn surface reducing the wear rate of the composite.
At a low applied pressure and sliding velocity, there was the formation of a plastic
deformation groove, whereas a high applied pressure and sliding velocity breakage
of tribolayer resulted in severe deformation.
Keywords ZrB2 · Worn surface · Tribo layer · Hardness
1 Introduction
Lightweight aluminum alloy is one of the major perspectives in the field of automotive
and aerospace applications due to its low density and high strength. The major appli-
cations are in brake rotors, pistons and cylinder liners, connecting rods and push rods.
P. Loganathan · A. Gnanavelbabu (B) · S. Ayyanar

Department of Industrial Engineering, CEG Campus, Anna University, Chennai, India
K. Rajkumar
Department of Mechanical Engineering, SSN College of Engineering, Kalavakkam, Chennai,
India

https://doi.org/10.1007/978-981-15-4488-0_68
810 P. Loganathan et al.
These components are subjected to abrasive wear mode along with sliding direction
against their counter surface. The wear resistance of the alloys and their mechanical
properties improved through the addition of hard ceramic reinforcement particles
on the matrix material [1]. Of late, ultra-high temperature ceramics (UHTCs) play a
major role in improving wear resistance in high-temperature environments. Among
various ceramic particles, zirconium diboride (ZrB2 ) has drawn the interest due to
its high temperature stability and oxidation protection on the surface of aerospace
structure, and it is mainly used in hypersonic flights and re-entry vehicles [2].
Recent studies on the aluminum alloy reinforced with ceramic particles are seen
in the literature [3–11]. Jerome et al. [3] carried out investigation on the mechanical
and high-temperature wear behavior of Al-TiC (5, 10 and 15 wt%) composites. There
was a decrease in the wear rate and the coefficient of friction of the composites due to
the formation of oxidative protective layer with increasing wt% of TiC. An increase
in hardness was seen due to the formation of intermetallic phases such as Al3 Ti.
Kumar et al. [4] analyzed the tensile and wear behavior of Al–7Si/TiB2 particulate
with (0, 3.4 and 6.8 vol. %) composites. Increase in hardness and tensile strength
of the composite was seen due to the presence of α-Al secondary dendritic arm
spacing and eutectic Si particle size and also the presence of fine TiB2 particles.
The effect of TiB2 particles on higher load showed a larger significant than low load
conditions due to the work hardening effect with Al-SiC composites. Liang et al. [5]
saw the tribological behavior of AZ31 Mg alloy. At low load, 5–30 N abrasion and
oxidation wear mechanisms were identified as also delamination mechanisms within
30–360 N.
Siddesh Kumar et al. [6] have stated that the wear behavior of AA2219/B4 C/MoS2
hybrid composites was fabricated through stir casting. Rise in temperature more
than 80 °C caused a rate of oxide layer removal significantly higher than the oxide
regeneration tending to increase the wear rate with increasing temperature. Johny
James and Raja Annamalai [7] made a study of the wear behavior of AA6061/ZrO2
composites with varying (5–15 wt%). Steady-state and mild wear behaviors were
observed in 5 wt% of ZrO2 , and a higher rate of delamination was obtained at 15 wt%
due to agglomeration of particles in the composite. Nishant Verma and Vettivel [8]
investigated the effect of B4 C and rice husk ash (RHA) on the mechanical behavior
of AA7075. The increment in mechanical properties was contributed mainly by B4 C
as compared to RHA.
Arivukkarasan et al. [9] analyzed the mechanical and tribological behavior of
LM13/WC composites fabricated through stir casting. The maximum tensile and
impact strength was obtained at 15 wt% of WC composites. Uniform distribution
of particles caused a decrease in mass loss of composites with increasing wt% of
WC. Jawahar Chandra et al. [10] concluded the wear behavior of LM13/TiS2 com-
posites and wear rate of composites increased with increasing applied load and slid-
ing velocity due to rise in interface temperature on contact pin and countersurface.
Ebenezer Jacob Dhas et al. [11] revealed the effect of SiC/WC/Gr on the mechanical
behavior of AA5052 hybrid composites. The formation of a brittle phase among
AA5052/WC/Gr composites caused a decrease in the impact strength more than
ZrB2 Influences on the Dry Sliding Wear … 811
AA5052/SiC/Gr composites. Both hybrid composites showed a brittle fracture due

to the presence of brittle interfaces between the reinforcements and the alloy matrix.
Work done on the mechanical and wear behavior of AA7075 alloy with ZrB2
reinforcement is not much speak of limit on the reinforcement fraction. The host
matrix was based on the reaction between the constitution chemical compounds. The
AA7075/ZrB2 combination brought the inherent mechanical properties that explain
the demand from military and aerospace industry.
This research work is an investigation of ZrB2 particles on AA7075 matrix alloy on
mechanical and wear behavior of the composites with varying weight percentages (5,
10, 15 wt%). Fabrication of the composites was carried out through the use of the stir
casting method. The mechanical properties micro Vickers hardness, impact strength
and tensile strength are analyzed. The dry sliding wear behavior of composites was
carried out on varying applied pressures 0.39–1.59 MPa, sliding velocity of 0.8–
2.0 ms−1 and constant distance 3 km. Dispersion of ZrB2 particles in the matrix was
examined through an optical microscope (OM), and analysis of wear mechanism on
worn surface was carried out through SEM.
AA7075 alloy was used as matrix material and ZrB2 particles with average size
of 5.60 µm as reinforcement particle. AA7075-ZrB2 composites were fabricated
using the two-step stir casting process. The required amount of AA7075 alloy was
placed in a graphite crucible in an electric resistance furnace at 800 °C to enable to
attainment of the molten state. The melting process carried out in argon atmosphere
at the rate of 2 CC/min to minimize high-temperature oxidation. ZrB2 particles were
preheated in the muffle furnace up to 150 °C for 2 h to remove the moisture content
and improve wettability. Alumina-coated mild steel blades were used for stirring
the aluminum melt. The initial stirring was done at 350 rpm, for the formation of
the vortex and addition of preheated particles into the vortex formation to attain
a uniform distribution of particles. After the complete addition of reinforcement
particles, the molten temperature dropped to 650 °C and a semi-solid state was
formed after10 min. The slurry temperature was raised up to the temperature of the
liquids, and stirring was done for 5 min. The molten melt was poured into a pre heated
die. The fabricated composites were designated as AZ0, AZ5, AZ10 and AZ15 based
on the concentration of ZrB2 particle varying as 0, 5, 10 and 15 wt%, respectively.
The fabricated composites were subjected to T6 heat treatment process to improve
the mechanical properties. Heat-treated composites were subjected to metallographic
examination. Specimens were etched using Keller’s solution. The microstructure of
the specimens was obtained through an optical microscope. The hardness of com-
posites was analyzed through Vickers hardness tester with an applied load of 500 gf
with a dwell time of 15 s. The tensile strength carried was as per ASTM E8 with
strain rate of 0.001/s in a computerized UTM machine (MTS Tensile Test 100 KN
capacity). Impact energy of the composite was analyzed using pendulum-type test-
ing machine with a capacity of 400 J. Wear behavior of the composite was seen out
through ASTM G-99, with applied pressure of (0.39–1.59 MPa), sliding velocities of
(0.8–2.0 m/s) and sliding distance of 3.0 km. The wear mechanism on worn surface
of composite is studied through SEM.
3.1 Microstructure Analysis
Figure 1a-b shows the microstructures of AZ0 and AZ5, respectively. It shows a good
chemical bonding among the Al particles, where Al are particles joined together to
construct a solid structure. The MgZn2 (g) phase precipitation in aluminum alloys
with a Zn content greater than 3 wt% and a Zn-to-Mg ratio greater than two represent
one of the strengthening mechanisms in these alloys, as well as the absence of voids
in the surrounding matrix.
T6 heat treatment, coarse and closely spaced precipitates along the grain bound-
aries and fine precipitates in the grains are seen relatively in the microstructures. The
distributions of reinforcements in the respective matrix are fairly uniform along the
grain boundaries as shown in Fig. 1b. The medium-sized aggregates of particles were
pushed toward final solidification regions and were eventually located in the grain
boundary. The refinement of grains could be due to the restriction in the movement
of solidification front due to the presence of ZrB2 particles and also act as nucleation
sites for matrix phase which tends to increase the number of grains.
Fig. 1 SEM image for AA7075 composite a AZ0. b AZ5

Fig. 2 Micro vickers

hardness of AA7075
composites
3.2 Micro Vickers Hardness
The details of the micro Vickers hardness of the AA7075 composites are shown in
Fig. 2. Increase in hardness with increasing wt% of ZrB2 is particles due to the high
dislocation density around the particles acting as a barrier to plastic deformation
during indentation, leading to increase in hardness of the composites [12]. Orowan
strengthening mechanisms cause increase in the hardness of composites in addition
to the interface, between the reinforcement and matrix [13]. Nearly 54.4% of max-
imum hardness was seen for AZ15 composite compared to AZ0 alloy. The large
difference in the coefficient of thermal expansion between ZrB2 (6.66 × 10–6 °C−1 )
and AA7075 alloy (2.3 × 10–6 °C−1 ) hardness and effect of T6 heat treatment
increases the hardness of the composites.
3.3 Effect of ZrB2 on Impact Strength
Impact strength of the composites is featured in Fig. 3. The addition of ZrB2 particles
caused a decrease in the impact strength of composites. Results of similar kind have
been reported by other researchers [11, 14]. The impact strength of all the composites
is less than that of the aluminum alloy by nearly 30.12%, compared to composites and
the base matrix. Aluminum alloy has a maximum fracture toughness due to plastic
deformation at stressed area. Basically, the toughness of the material is based on the
energy consumed for fracturing. The decrease in impact strength of the composite
was due to a mismatch in elastic modulus, the thermal expansion of AA7075/ZrB2
and its tendency to increase the elastic strain. The addition of ceramic particles led
to composite materials acting as brittle nature and reduces the impact strength.
Fig. 3 Impact strength of

composites
3.4 Effect of ZrB2 on Tensile Strength and % of Elongation
Figure 4 shows the tensile strength and % elongation of the composites. There was
an increase in the tensile strength of composites with increasing wt% of ZrB2 par-
ticles. Increase in tensile strength was due to strong interfacial bonding between
the aluminum alloy and ZrB2 particles. The strength of alloy mainly depends upon
deformation process on strain hardening and thermal softening. The addition of rein-
forcement caused a marked reduction in the percentage of the elongation due to the
presence of hard particles at the base material grain boundary. The stress transfer
between the base matrix and ZrB2 particles was the result of the interaction between
the dislocation and ZrB2 particles. The effect of Orowan strengthening mechanism
where a dislocation bows out considerably and leaving a dislocation loop around a
particle [15, 16].
Fig. 4 Tensile strength of

composites
3.5 Wear Behavior
The observations of wear rate and COF of the AA7075 composites are shown in
Figs. 5 and 6. The variations in wear rate with respect to applied pressure at different
wt% of ZrB2 and sliding velocity are shown in Fig. 5. The wear rate of the composites
increased with increase in sliding velocity. At the initial condition 0.8 m/s, there was
an increase in the wear rate for all compositions due to the direct contact between
the pin surface and the disk. With the addition of ZrB2 particles, wear rate of the
composite started decreasing with increasing wt% of particles, due to improvement
in hardness, strong interfacial bonding between the matrix and ceramic particles and
grain refinement in the structure. Wear rate of base alloy at 2 m/s 0.00271 mm3 /m
was more compared to AZ15 composite 0.00182 mm3 /m, and nearly 38% of wear
rate reduction in AZ15 composite was due to the addition of ZrB2 composites.
Fig. 5 Wear rate of the

composites
Fig. 6 COF of the

composite
Decrease in wear rate also caused differences in CTE between the ZrB2 particles,
and α-Al matrix was expected to develop a larger strain leading to enhancement
of the dislocation density. The effect of friction is composite material shown in
Fig. 6. The COF of the composites showed a decrease with increasing wt% of ZrB2
particles due to the hard resistance and uniform distribution of particles and formation
of mechanical mixed layer (MML) acting as a lubrication medium reducing the
COF and wear rate of composites [17]. Increase in sliding velocity caused increase
in the surface temperature due to the oxidation of the pull out of ZrB2 particles.
This is attributed to the formation of B2 O3 self-lubricating layer. The worn surface
morphology of the AA7075/ZrB2 composites is shown in Fig. 7.
Two wear mechanisms occurred during dry sliding wear of AA7075/ZrB2 com-
posites, namely abrasion and delamination. Figure 7a shows the worn surface of AZ0
alloy at 1.6 m/s and 1.194 Mpa. Formation of grooves and more surface failure and
plowing. The lower hardness of the alloy made the adhesive wear mechanism domi-
nant. Figure 7b–c shows the worn surface of AZ5 and AZ10 composites. A pull out
of particles, wear debris as white particles and thin formation of oxide layer acting
a lubrication layer lead to reduction in the COF, and dominant wear mechanism of
mild wear/oxide wear could be seen. Figure 7d shows the worn surface of AZ15
at high-load and high-speed conditions. Wear transmission for mild-to-severe wear
mode occurred due to a rise in temperature leading to an increase in the wear rate of
composites while a delamination occurred on the surface.
Fig. 7 Worn surface of AA7075/ZrB2 composites a AZ0. b AZ5. c AZ10. d AZ15

4 Conclusions
In this study, AA7075-ZrB2 composites were successfully fabricated using stir cast-
ing technique, and the mechanical and wear behavior was evaluated. The following
conclusions are drawn:
1. The distribution of the reinforcement particles obtained in matrix material was
uniform and particles lay down in the grain boundary and within the grain region.
2. The hardness and tensile strength of the composites increased with increasing
wt% of ZrB2 particles, and the impact strength of composites decreased at a
higher wt% due to the presence of hard particles which acts as brittleness.
3. Wear rate of the composites increased with increasing applied load and sliding
speed compared to base matrix. AZ15 composites show a minimum wear rate
and COF.
4. Wear mechanism of the composites has both abrasion and delamination mecha-
nisms. The mode change of the mechanism from mild-to-severe wear occurred
at 1.6 m/s and 1.19 MPa applied load.
Acknowledgements The authors would like to thank the DST-SERB, Government of India, for
supporting the research with funding under the Grant No: EEQ/2017/000382.
References
1. Kaushik NC, Rao RN (2017) High-stress abrasive wear behaviour of Al–Mg–Si hybrid
composites using regression analysis. Ind Lubr Tribol 69(2):149–157
2. Zamora V, Ortiz AL, Guiberteau F, Nygren M (2012) In situ formation of ZrB2–ZrO2 ultra-
high-temperature ceramic composites from high-energy ball-milled ZrB2 powders. J Alloy
Compd 518:38–43
3. Jerome S, Ravisankar B, Mahato PK, Natarajan S (2010) Synthesis and evaluation of mechan-
ical and high temperature tribological properties of in-situ Al–TiC composites. Tribol Int
43(11):2029–2036
4. Kumar S, Chakraborty M, Sarma VS, Murty BS (2008) Tensile and wear behaviour of in situ
Al–7Si/TiB2 particulate composites. Wear 265(1–2):134–142
5. Liang C, Li C, Lv XX, An J (2014) Correlation between friction-induced microstructural
evolution, strain hardening in subsurface and tribological properties of AZ31 magnesium alloy.
Wear 312(1–2):29–39
6. Kumar NS, Prabhu TR, Mishra RK, Eswaraprasad N, Shankar GS, Basavarajappa S Analysis
of dry sliding wear behavior of the nano composites using statistical methods with an emphasis
on temperature effects. Measurement 128:362–376
7. James J (2018) Tribological behaviour and wear fashion of processed AA6061/ZrO2 composite.
Ind Lubr Tribol 70(9):1815–1824
8. Verma N, Vettivel SC (2018) Characterization and experimental analysis of boron carbide
and rice husk ash reinforced AA7075 aluminium alloy hybrid composite. J Alloy Compd
741:981–998
9. Arivukkarasan S, Dhanalakshmi V, Stalin B, Ravichandran M (2018) Mechanical and tribo-
logical behaviour of tungsten carbide reinforced aluminum LM4 matrix composites. Part Sci
Technol 36(8):967–973
10. Jawahar Chandra C, Radhika N, Bhatija KA, Saalai Thenagan R (2018) Fabrication of light-
weight Al LM13/TiS2 metal matrix composites and investigation of its wear characteristics.
Part Sci Technol https://doi.org/10.1080/02726351.2017.1394408
11. Dhas DEJ, Velmurugan C, Wins KLD, BoopathiRaja KP (2019) Effect of tungsten carbide,
silicon carbide and graphite particulates on the mechanical and microstructural characteristics
of AA 5052 hybrid composites. Ceram Int 45(1):614–621
12. Gnanavelbabu A, Muthazhagan C, Rajkumar K, Ayyanar S, Loganathan P (2018) Experimen-
tal investigation and reliability analysis on wear performance of AA6061-B4 C-Nanographite
hybrid composites. Mater Today Proc 5(2):8436–8445
13. Sharma P, Sharma S, Khanduja D (2015) A study on microstructure of aluminium matrix
composites. J Asian Ceram Soc 3(3):240–244
14. Kumar KR, Kiran K, Sreebalaji VS (2017) Micro structural characteristics and mechanical
behaviour of aluminium matrix composites reinforced with titanium carbide. J Alloy Compd
723:795–801
15. Veeravalli RR, Nallu R, Mohiuddin SMM (2016) Mechanical and tribological properties of
AA7075–TiC metal matrix composites under heat treated (T6) and cast conditions. J Mater
Res Technol 5(4):377–383
16. Murali M, Sambathkumar M, Saravanan MS (2014) Micro structural and mechanical properties
of AA 7075/Tio 2 in situ composites. Univ J Mater Sci 2(3):49–53
17. Gnanavelbabu A, Rajkumar K, Saravanan P (2018) Investigation on the cutting quality charac-
teristics of abrasive water jet machining of AA6061-B4 C-hBN hybrid metal matrix composites.
Mater Manuf Processes 33(12):1313–1323
Investigation on Wear Properties
of Nickel-Coated Al2 O3P -Reinforced
AA-7075 Metal Matrix Composites Using
Grey Relational Analysis
D. Vijay Praveen, D. Ranga Raju, and M. V. Jagannadha Raju
Abstract The development of hard reinforcements in MMCs became an important

area of research in materials engineering. Metallic coatings on reinforced particles
can improve wettability because it plays a key role in aluminium-based metal matrix
composites. Nickel plating by electroless method having found to be an efficient
method and widely accepted in various engineering domains such as aerospace,
automobile, chemical and electrical industries. In this study, 20 micrometer-sized
(average) alumina (Al2 O3 ) particles are coated by nickel using electroless deposition
technique. Wear specimens are prepared as per ASTM-G99 standards, using coated
and uncoated reinforcements of 3, 6 and 9% weight fractions into 7075 aluminium
alloy by standard stir casting route. Wear test was conducted by utilizing the test
parameters, viz. track diameter, speed and load based on L18 Taguchi orthogonal
array, on pin-on-disc machine, and the influencing parameters of responses were
optimized using grey relational analysis. Weary surfaces were analysed using scanned
electron microscopy. The specimen with 9% coated reinforced composite at 80 mm
track diameter, 9.81 N and 1000 rpm showed improved resistance to wear.
Keywords Electroless nickel coating · Al2 o3 p/AA7075 · AMMC · Wear
1 Introduction
As a consequence of technology growth, there is enormous demand for advanced

materials which possess light in weight, economic, stronger and harder in the field of
D. V. Praveen (B)
Research Scholar, Andhra University, Vishakhapatnam 530003, India
Department of Mechanical Engineering, Bapatla Engineering College, Bapatla 522102, India
D. R. Raju
Srinivasa Institute of Engineering and Technology, Amalapuram 533216, India
M. V. J. Raju
Department of Mechanical Engineering, Andhra University, Vishakhapatnam 530003, India

https://doi.org/10.1007/978-981-15-4488-0_69
820 D. V. Praveen et al.
automotive, space and defence applications. Reinforcing the hard ceramic particles
into aluminium metal matrix can improve the mechanical and wear attributes [1].
Solid-state processing and liquid-state processing are the two major categories of
fabrication of AMCs, in which liquid-state route is preferred as it possesses major
advantages like continuous production, economical and ease of adaption [2–4]. There
are a few limitations in the above techniques like poor wettability, non-uniform
distribution of the hard reinforcements in the liquefied metal [5], partial bonding and
improper interface between reinforcement and matrix materials [6].
Nowadays, electroless composite coating technique has gained good recognition
in preparing composite coatings, discussed by Loto [7]. Moustafa et al. had investi-
gated wear studies of copper graphite composite of uncoated and coated reinforce-
ments and reported that the coated reinforcements show better wear performance
when compared with uncoated reinforced composites [8]. Zhan et al. had studied
the wear characteristics of nickel-coated SiC-reinforced composites and observed
enhanced wear resistance [9]. Davidson et al. investigated on benefits of coating and
interfacial bonding of copper(Cu)-coated silicon carbide (SiC) particle-reinforced
6061 alloy [10]. Most of the researchers had worked vigorously on the wear studies
on aluminium-based MMCs, but a few studies are available on electroless-coated on
particulate reinforced composites and reported that the mechanical and wear prop-
erties are enhanced by these coated composites when compared with the uncoated
reinforced composites.
In this present paper, experimental analysis reports the influence of nickel coating
on Al2 O3 particle on wear behaviour of prepared composite materials, which were
fabricated by stir casting routing, and the experimental responses were analysed
using grey relational approach.
2.1 Selection of Matrix Material and Reinforcement
AA7075, is Al-based zinc alloy, has capable to withstand high loads and temperature,
and it has very good wear resistance, better fatigue strength, average machinability
and considerable corrosion resistance. Because of these properties, it finds appli-
cations in transport applications, including marine, automotive, aviation and mould
tool manufacturing. The chemical composition of base alloy is shown in Table 1.
Table 1 Chemical composition of AA-7075

Si Fe Cu Zn Mg Mn Cr Ti Al
0.4 .5 1.43 4.87 2.43 0.2 0.1 0.16 Bal.
Investigation on Wear Properties of Nickel-Coated … 821
Fig. 1 Electroless nickel

coating process [12] Ultra sonic Etching
Sensitization
cleaning Process
Preparation Surface
Coating
of Bath Activation
Al2 O3 has high hardness, chemical stability and structural stability in the certain
temperature range, which is applied in the substrate surface to significantly reduce
co-efficient of friction and improve the wear resistance [11].
2.2 Preparation of Reinforcement Particles
Coating procedure on Al2 O3 microparticles was carried out by standard electroless

plating technique. Basically, the process consists of two steps. They are surface treat-
ment of the particles and, the other one is preparation of bath for coating. Pretreatment
consists of ultrasonic cleaning, etching, sensitization and surface activation. Figure 1
shows sequence of the steps that were followed for coating process. After each suc-
cessive stage, the reinforcement particles are cleansed with distilled water followed
by drying and proceeded to the next stage [12–15].
The second step is preparation of coating bath. The bath for the coat-
ing was prepared with nickel chloride (NiCl2 .6H2 O)-30 g/l, tri-sodium cit-
rate (Na3 C6 H5 NaO7 .H2 O)-25 g/l, ammonia chloride (NH4 Cl)-50 g/l and sodium
hypophosphate (NaH2 PO2 .H2 O)-25 g/l. Nickel plating had been carried out at a
temperature of 85 °C, pH value of 8 and rousing time of 20 min.
2.3 Development of Composites
Composites are fabricated by the standard liquid metallurgy route. Processed Al2 O3
particles were preheated, before being introduced into the vortex and stirring of the
molten composite were accomplished for 8 min at 350 rpm of the stirrer speed. Pour-
ing temperatures adopted were 710 °C. As the weight percentage of Al2 O3 increases
more than 10, the produced composites may exhibit brittle nature as observed in the
literature [16–18]. Within 0–10% range, as preliminary experiments, the percentage
of reinforcement has been fixed with equal intervals, to get three different compo-
sitions (i.e. 3, 6 and 9%). The cylinders of Ø15 mm × 150 mm cast composites of
were obtained for the both uncoated and coated reinforced AA7075 with weight per-
centages of 3, 6 and 9. Finally, the specimens are developed by machining the casted
Table 2 Basic properties of the material

Matrix material Reinforcement Density (gm/cc) Rockwell hardness number
@981 N, duration 10 s
AA-7075 3% wt. Al2 O3 2.759 49 ± 5.1
AA-7075 6% wt. Al2 O3 2.775 50 ± 1.9
AA-7075 9% wt. Al2 O3 2.787 48 ± 3.6
AA-7075 3% wt. Al2 O3 (Ni coated) 2.773 60 ± 7.8
cylinders into Ø12 mm × 25 mm as per ASTM-G99. Experimental density values

and Rockwell hardness number (average of three readings) of prepared composites
are shown in Table 2.
3.1 Design of Experiments
Design of experiment (DOE) is a statistical methodology for establishing the associ-

ation between different variables affecting any process. Taguchi technique had been
effectively used by many researchers in the analysis of wear properties of compos-
ites. This approach abolishes the need for repetitive experimental tests and thus saves
material, cost and time. The very important stage in the arrangement of experiments
is selection of influencing factors which have effects on the process.
Experiments were designed as multilevel problem, L18 , orthogonal array to inves-
tigate the most effecting parameters on the wear behaviour of the fabricated com-
posite material. Many researchers considered these parameters for investigation. The
level and factors considered for conducting wear study are shown in Table 3. Pin-on-
disc machine was used to conduct wear test on the specimens which were prepared
according ASTM-G99 standards.
Table 3 Control factors and levels

S. No. Factors Levels
1 Reinforcement (wt%) 3 6 9 - 3C 6C 9C
[Un-coated [Coated particles]
particles]
2 Load (N) 9.81 19.61 29.4
3 Track radius (mm) 60 80 100
4 Speed (rpm) 600 800 1000
The surfaces of the samples were cleansed with a soft paper soaked in acetone
before the test. Weight of each specimen was recorded using digital electronic balance
(Mettler Toledo: accuracy 0.0001 g). The initial and final weights of specimen were
measured to find the slide wear loss. Table 4 shows the experimental result coefficient
of friction and specific wear rate.
3.2 Grey Relational Analysis
Grey relational technique is used to solve the multi objective decision-making prob-
lems by combing the entire range of performance attribute values being considered
for every alternative into one single value. This reduces the complexity of the original
problem. The analysis is performed with the following steps:
1. Data normalization: The responses of the experimental data should be normalized
for getting uniformity of data. For the current experiment, both the responses,
i.e. wear rate and co-efficient of friction, should be lower the better.
(bi j )Max − (bi j )

ai j = (1)
(bi j )Max − (bi j )Min
where bi j is unique sequence for ith experimental result for jth experiment,
(bi j )Max is the maximum value and (bi j )Min is minimum value of the experiment
and is sequence produced of processing the data.
2. Calculation of deviation coefficient for each experimental response using the
following equation.
= 1 − ai j
3. By taking average of grey relational coefficients, the GRG can be determined for
each response using the Eq. (2).
min − ν · max
γ a0 j , ai j = (2)
i j max + ν · max
For i = 1, 2….m, j = 1, 2…..n. where i j = |bi j − ai j |
max = Max{i j , For i = 1, 2 . . . m, j = 1, 2 . . . . n}

min = Min{i j , For i = 1, 2 . . . m, j = 1, 2 . . . . n}
4. Grey relational grade (GRG): It shows the influence of process parameters. The
upper value of the grey relational grade represents the stronger relational degree
in the given sequence [19–21]. Table 5 shows the calculations of GRG of the
experimental responses.
4.1 Worn Surface Analysis
As the percentage of Al2 O3 increases from 3 to 9, it can be seen that the particulates
are distributed in the matrix and also agglomerated in some regions. Scanned electron
microscopy (SEM) images of the composites are shown in Fig. 2.
The microstructures of a few worned surfaces are shown in Fig. 3. Figure 3a, b
are related to experiment numbers 2 and 5 which were ranked as 17 and 18, show the
direction of weary surface and also show the regions of delamination. In Fig. 3b, the
specimen surface was highly damaged and craters of different depth were observed.
Due to high rise in temperature, its surface was softened and swelling was also
noticed. Figure 3c, d show the microstructure of coated composite with 3 and 9%
and of experimental set of 10 and 17. In Fig. 3c, parallel scratches, pits and surface
cracks were observed. In Fig. 3d, no considerable damage was found.
Fig. 2 SEM images of composite: a 7075 pure, b 3% Al2 O3 +AA7075, c 6% Al2 O3 +AA7075,
d 9% Al2 O3 +AA7075, e 3% Ni-coatedAl2 O3 +AA7075, f 6% Ni-coated Al2 O3 +AA7075 and
g 9% Ni-coated Al2 O3 + AA7075
Fig. 3 Scanned electron microscopy images of worned surface composite
5 Experimental Result Analysis
5.1 Analysis of the Experimental Results
Wear test results are shown in Table 4. The following Table 5 shows the normalized
values of coefficient of friction (X 1 ), normalized values of percentage wear (X 2 ),
deviation sequence (X 3 ), deviational sequence of percentage wear (X 4 ), grey rela-
tional coefficient for coefficient of friction as well as wear rate, grey relational grade
and rank [21].
ANOVA was performed using Minitab software to determine the highly influ-
encing characteristics and also the main effects plot of grey relational grade
Fig. 4.
From the grey relational analysis, it is also observed that the nickel-coated com-
posites shown better resistance to wear when compared to the uncoated reinforced
composites. From the ANOVA, the material type has highest influence (79.945%)
followed by load, (7.81%), track diameter (4.65%) and speed (4.478%). The same
were shown in Table 6 and plotted in the pi-chart in Fig. 5.
Main Effects Plot for GRG

Data Means
Material track dia load speed
0.80
0.75
Mean of Means
0.70
0.65
0.60
0.55
0.50
3% 6% 9% 3% 6% 9% 60 80 10
0 1 2 3
60
0
80
0 00
d d d d d d 10
ate ate ate at e ate ate
Co Co Co n co n co nco
U U U
Fig. 4 Main effects plot for GRG
Table 4 Experimental results of wear and friction coefficient

Exp. Reinforcement Track Weight Speed %of wear Coefficient of
No diameter (Kg) (RPM) (wt. loss) friction
(mm)
1 Uncoated 3% 60 9.81 600 0.5162 0.2268
2 Uncoated 3% 80 19.61 800 0.5358 0.2624
3 Uncoated 3% 100 29.40 1000 0.8852 0.1901
4 Uncoated 6% 60 19.61 800 0.1072 0.1765
5 Uncoated 6% 80 29.40 1000 0.8487 0.2354
6 Uncoated 6% 100 9.81 600 0.3439 0.2418
7 Uncoated 9% 60 29.40 600 0.1489 0.1730
8 Uncoated 9% 80 9.81 800 0.2569 0.2526
9 Uncoated 9% 100 19.61 1000 0.1573 0.2393
10 Coated 3% 60 29.40 1000 0.2600 0.1696
11 Coated 3% 80 9.81 600 0.2392 0.1862
12 Coated 3% 100 19.61 800 0.2910 0.1992
13 Coated 6% 60 19.61 1000 0.2442 0.1938
14 Coated 6% 80 29.40 600 0.8903 0.1539
15 Coated 6% 100 9.81 800 0.2088 0.1990
16 Coated 9% 60 29.40 800 0.3182 0.1758
17 Coated 9% 80 9.81 1000 0.1833 0.1744
18 Coated 9% 100 19.61 600 0.2436 0.2167
Table 5 Grey relational grade calculations

Exp. No X1 X2 X3 X4 X5 X6 GRG Rank
1 0.5162 0.2268 0.5080 0.5539 0.4920 0.5285 0.5162 15
2 0.5358 0.2624 0.4814 0.3360 0.5186 0.4295 0.4602 17
3 0.8852 0.1901 0.0069 0.7785 0.9931 0.6930 0.5139 16
4 0.1072 0.1765 0.2486 0.8617 0.7514 0.7833 0.5914 13
5 0.8487 0.2354 0.0565 0.5012 0.9435 0.5006 0.4235 18
6 0.3439 0.2418 0.7420 0.4621 0.2580 0.4817 0.5707 14
7 0.1489 0.3173 1.0068 0.0000 0.0000 0.3333 0.6735 10
8 0.2569 0.2526 0.8601 0.3960 0.1399 0.4529 0.6171 12
9 0.1573 0.2393 0.9954 0.4774 0.0046 0.4889 0.7399 6
10 0.2600 0.1696 0.8559 0.9039 0.1441 0.8388 0.8076 2
11 0.2392 0.1862 0.8842 0.8023 0.1158 0.7167 0.7643 3
12 0.2910 0.1992 0.8138 0.7228 0.1862 0.6433 0.6860 8
13 0.2442 0.1938 0.8774 0.7558 0.1226 0.6719 0.7375 7
14 0.8903 0.1539 0.0000 1.0000 1.0000 1.0000 0.6667 11
15 0.2088 0.199 0.9254 0.7240 0.0746 0.6443 0.7573 4
16 0.3182 0.1758 0.7769 0.8660 0.2231 0.7886 0.7400 5
17 0.1833 0.1744 0.9601 0.8745 0.0399 0.7994 0.8627 1
18 0.2436 0.2167 0.8782 0.6157 0.1218 0.5654 0.6848 9
Table 6 Main effects plot for GRG

Source DF Adj SS Adj MS F-value P-value Percentage
contribution
Material 5 0.2045 0.0409 10.49 0.006 79.945
Track diameter 2 0.0118 0.0059 1.52 0.292 4.650
Load 2 0.0199 0.0099 2.56 0.157 7.810
Speed 2 0.0114 0.0057 1.47 0.303 4.478
Error 6 0.0234 0.0038 9.152
Total 17 0.2556
Fig. 5 Percentage Percentage Contribution

contribution chart
9%
4% Material
7% track dia
4% load
76% speed
Error
6 Conclusion
In the present work, it was observed that the percentage of wear of the particle-coated
composite is improved when compared to the uncoated reinforced composite. Grey
relational analysis was successfully done, and the following conclusions were drawn
from the present investigation.
1. As the GRG value is highest at the experiment no. 17, 9% nickel-coated-
reinforced 7075 composite material has shown good result in percentage of
wear.
2. Similarly, most of the GRG values yielded good result when compared with the
uncoated reinforcement.
3. ANOVA results revealed that material type is the most influencing parameter in
the present analysis followed by load, track diameter and speed on wear studies
References
1. Garcia-Cordovilla C, Narciso J, Louis E (1996) Abrasive wear resistance of aluminium

alloy/ceramic particulate composites. Wear 192(1–2):170–177
2. Gladston JAK, Sheriff NM, Dinaharan I, Selvam JDR (2015) Production and characterization
of rich husk ash particulate reinforced AA6061 aluminum alloy composites by compocasting.
Transactions of Nonferrous Metals Society of China 25(6):686–691
3. Rittner M (2000) Metal matrix composites in the 21st century: markets and opportunities.
Report GB-108R, Business Communications Co., Inc., Norwalk, CT
4. Hashim J, Looney L, Hashmi MSJ (1999) Metal matrix composites: production by the stir
casting method. J Mater Process Technol 92:1–7
5. Asthana R (1998) Reinforced cast metals: part II evolution of the interface. J Mater Sci
33(8):1959–1980
6. Keshavamurthy R, Mageri S, Raj G, Naveenkumar B, Kadakol PM, Vasu K (2013) Microstruc-
ture and mechanical properties of Al7075-TiB2 in-situ composite. Res J Mater Sci. ISSN2320:
6055
7. Loto CA (2016) Electroless nickel plating–a review. Silicon 8(2):177–186
8. Moustafa SF, El-Badry SA, Sanad AM, Kieback B (2002) Friction and wear of copper–graphite
composites made with Cu-coated and uncoated graphite powders. Wear 253(7–8):699–710
9. Zhan Yongzhong, Zhang Guoding (2003) The effect of interfacial modifying on the mechanical
and wear properties of SiCp/Cu composites. Mater Lett 57(29):4583–4591
10. Davidson AM, Regener D (2000) A comparison of aluminium-based metal-matrix compos-
ites reinforced with coated and uncoated particulate silicon carbide. Compos Sci Technol
60(6):865–869
11. Zhang C (2014) Understanding the wear and tribological properties of ceramic matrix com-
posites. Advances in ceramic matrix composites. Wood Head Publishing, Cambridge, pp
612–669
12. Kumar DS, Suman KNS, Kumar PR (2016) A study on electroless deposition of nickel on nano
alumina powder under different sensitization conditions. Int J Adv Sci Technol 97:59–68
13. Plating E (1990) Fundamentals and applications. In: Mallory GO, Hajdu FB (eds) American
electroplaters and surface finishers society, Orlando, Forida
14. Pang JN, Jiang SW, Lin H, Wang ZQ (2016) Significance of sensitization process in electroless
deposition of Ni on nanosized Al2O3 powders. Ceram Int 42(3):4491–4497
15. Li GJ, Huang XX, Guo JK (2001) Fabrication of Ni-coated Al2O3 powders by the
heterogeneous precipitation method. Mater Res Bull 36(7–8):1307–1315
16. Javdani A, Pouyafar V, Ameli A, Volinsky AA (2016) Blended powder semisolid forming of
Al7075/Al2O3 composites: investigation of microstructure and mechanical properties. Mater
Des 109:57–67
17. Dhanalakshmi S, Mohanasundararaju N, Venkatakrishnan PG (2014) Preparation and mechan-
ical Characterization of stir cast hybrid Al7075-Al2 O3 -B4C metal matrix composites. In:
Applied mechanics and materials, vol 592. Trans Tech Publications, pp 705–710
18. Maurya Manish, Kumar Sudhir, Bajpai Vivek (2019) Assessment of the mechanical properties
of aluminium metal matrix composite: a review. J Reinf Plast Compos 38(6):267–298
19. Sharma P, Khanduja D, Sharma S (2016) Dry sliding wear investigation of Al6082/Gr metal
matrix composites by response surface methodology. J Mater Res Technol 5(1):29–36
20. Subbaya KM, Suresha B, Rajendra N, Varadarajan YS (2012) Grey-based Taguchi approach
for wear assessment of SiC filled carbon–epoxy composites. Mater Des 41:124–130
21. Siriyala R, Alluru GK, Penmetsa RM, Duraiselvam M (2012) Application of grey-taguchi
method for optimization of dry sliding wear properties of aluminum MMCs. Front Mech Eng
7(3):279–287
Optimizing the Tribological Properties
of UHMWPE
Nanocomposites—An Artificial
Intelligence based approach
A. Vinoth, K. N. Nirmal, Rohit Khedar, and Shubhabrata Datta
Abstract The longevity of the hip implants has been a major issue in recent times
due to inadequate material used for implants. Since the metal on polymer implants has
issues such as tissue degeneration and osteolysis, the focus of this study is to improve
the tribological properties of ultra-high molecular-weight polyethylene (UHMWPE)
which has been in use on acetabular cup of hip implants by considering multiple
nanoparticles like carbon fibre, carbon nanotubes and graphene as reinforcements.
It is extremely difficult and time-consuming through numerous experimental tri-
als to arrive at the optimum material composition of nanoparticles. Therefore, an
effort has been made on developing a new polymer nanocomposite by utilizing the
artificial intelligence (AI)-based design which includes the techniques, viz. artificial
neural network (ANN) and genetic algorithm (GA). The input parameters like weight
fraction and the geometry of the different nanoparticles related to the tribological
properties were collected from various published literatures, and modelling was done
through ANN for the output parameters, viz. coefficient of friction and specific wear
rate. Best ANN predictive model was chosen individually for each output parame-
ters on iterating the different hidden nodes. The fundamental correlation between the
input and output parameters was investigated through sensitivity analysis. Optimiza-
tion studies were performed using genetic algorithm (GA) with the best-chosen ANN
model as an input to get optimum input variables. Thus, the AI-based approach of
designing the UHMWPE nanocomposites shows an enhancement on the tribological
properties that pave a way for further experimental trials.
Keywords Wear · UHMWPE · Composites · Artificial neural network and genetic

algorithm
A. Vinoth (B) · K. N. Nirmal · R. Khedar · S. Datta

Kattankulathur 603203, India

https://doi.org/10.1007/978-981-15-4488-0_70
832 A. Vinoth et al.
1 Introduction
The need for a proper material in hip implant has been a major area of research in
recent years. Hence, wide variety of materials have been in use and also been studied
upon like metal implants, ceramic implants, polymer implants, etc., which could
serve its purpose as a strong, reliable and biocompatible implant. Metal implants
were the first which came into the picture but were not much successful because
of the wear of metal particles which may cause osteolysis [1]. Metal implants have
paved way for ceramic and polymers because of their uncanny characteristics and
have ability to show properties superior to metals both mechanically and chemically
[2].
Polymers have been in use since the early 1960s for making acetabular cups in
hip arthroplasty and was proposed by a British surgeon named ‘John Charnley’ [3].
Charnley made use of polytetrafluoroethylene (PTFE) polymer in acetabular compo-
nents which reduced the use of metal implants in arthroplasty [4]. But the immense
research carried out under PTFE being used in implants proved it be an inferior
component to ultra-high molecular weight polyethylene (UHWMPE) because of
its distortion and high wear rate. UHMWPE polymer has been in use since 1962
in orthopaedics because of its properties such as high toughness, self-lubricating
capability, resistance to absorb moisture and to get oxidized. Researches have been
carried out in this field to further improve the property of UHMWPE by adding
reinforcements such as carbon fibre, CNT, graphene oxide, nano-Al2 O3 , hydroxyap-
atite (HA), nano-SiO2 fibre as individual components [5]. It was seen that on adding
these reinforcements, there has been significant improvement in the properties like
wear rate and coefficient of friction (COF) when compared with pure UHMWPE [6].
This study focuses on further enhancing the tribological properties of UHMWPE on
combining different nanoparticles as reinforcement namely CNT, carbon fibre and
graphene oxide and its influence on the matrix. The experimental trial would be
possible only if it is practically tried out for various combinations of all the three
composites with regard to numerous properties and structural behaviour with the
UHMWPE matrix to find the optimal composition. It can be concluded easily that
this trial-and-error method is too complicated, tedious and consumes a lot of money
and time. In order to overcome these issues, computational data-driven models are
developed on collecting abundant data from several published literatures. The main
purpose of developing these models is to find a perfect balance between the compo-
sition, structure and the physical properties among UHMWPE and its composites.
The tribological properties to be monitored under this study include a reduction of
COF and wear rate.
Artificial intelligence techniques like ANN is employed to establish individual
relationship among input and output parameters fed into ANN as data-driven models
[7–9]. The optimal models which give proper relationship and which are closer to
the target required are chosen for further optimization. Optimization is conducted
through genetic algorithm (GA) to which the best-chosen ANN models are fed as
inputs which act as an objective function for carrying out the optimization [10].
Optimizing the Tribological Properties of UHMWPE Nanocomposites … 833
Depending on observations and pattern of output from GA, the solution can be
obtained through single objective or also if needed using multi-objective manner. The
optimal results collected from GA could be used to design a UWMWPE composite
which has a combination of reinforcements with improved tribological performance
for the use of acetabular cup in hip prosthesis.
2 Database
A sufficient set of data was collected from published literatures containing ultra-
high molecular-weight polyethylene (UHMWPE) composites with graphene oxide,
carbon fibre and Carbon nanotubes as reinforcements ranging up to 471 in num-
ber [11–24]. A total number of 16 input parameters were considered for the study
which includes the weight percentage, particle size of the above-mentioned three
reinforcements.
The inputs also contain variables like sliding distance, sliding speed, method of
test, load and hardness. These data are fed into ANN as input to develop ANN models
which give relationships between the input and output variables. The developed ANN
models are then considered for minimizing the output parameters COF and wear rate.
Table 1 depicts the minimum, maximum, mean and standard deviation of the input
and output variables.
3 Artificial Intelligence (AI) Techniques
Artificial intelligence techniques such as artificial neural network (ANN) and genetic
algorithm (GA) are used as computational methods. ANN and GA have significant
roles in developing empirical models especially in the field of materials for develop-
ing new materials. They also have the ability to capture highly complex systems from
the data. ANN is used to develop data-driven models which are fed into GA as objec-
tive function to get optimized solutions on the tribological behaviour of UHMWPE
composites. The functioning of ANN and GA is described below.
3.1 Artificial Neural Network (ANN)
Artificial neural network works similarly as our human brain. It has several neurons to
connect between layers to carry forward the stimulus similar to our neuron system. It
has huge number of nodes which act as connections or links to transfer signal between
layers to process the data. ANN is a very popular tool to develop data-driven models
which are fundamentally different from statistical or numerical approach. The initial
stage of ANN processing involves training itself from several computational trials
Table 1 Maximum and minimum ranges of each input and output parameters along with their
mean and standard deviation
Variables Min Max Mean Std. dev.
Molecular weight (million g/mol) 2.5 6 4.045 0.627
Carbon nanotube (wt%) 0 3 0.769 0.757
CNT fibre length (µm) 0 15 9.063 7.668
CNT fibre OD (nm) 0 80 32.027 25.528
Carbon fibre (wt%) 0 12 0.790 2.517
CF length (µm) 0 1000 113.890 317.300
CF OD (nm) 0 7000 736.363 2012.133
Graphene (wt%) 0 3 0.260 0.558
Graphene sheet thickness (nm) 0 1 0.045 0.201
Graphene sheet length (µm) 0 4 0.0681 0.343
Method 0 1 0.136 0.343
Sliding distance (m) 5 10,000 911.045 1176.424
Sliding speed (m/s) 0.004 1.667 0.432 0.435
Load (N) 5 140 40.047 29.174
Lubrication 0 2 0.909 0.884
Hardness (kgf/mm2 ) 67.5 1580 303.797 278.087
Coefficient of friction 0.04 0.84 0.210 0.206
Wear rate (mm3 /Nm) 0.0000002 0.00009 1.816E-05 2.109E-05
to reach desired output or target. ANN can be utilized to reach the desired target by
fine tuning its weights and biases. Specific and minute changes can be made through
connection strength or weights until the perfect match for the output is generated.
This step is generally called learning or training of ANN. ANN became more popular
and readily acceptable since 1980s because of its ability to handle nonlinear systems
[26]. ANN comprises three sets of processing units, one which receives inputs from
the outside world is called input unit which is processed in a layer called ‘input layer’.
The second unit consists of a hidden unit which receives inputs from the input node
or unit and processes the data according to the number of hidden layers required
by the user. The final set of results are extracted from the output layer receives
inputs from the hidden layer. The node number in the input layer and output layer is
taken from the application problem which is being evaluated, whereas the number
of hidden layers is to be defined by the user. The structure of ANN is represented by
16-N-1 notation in which 16 represents the number of inputs, viz. molecular weight
of UHMWPE, CNT wt%, CNT fibre length, CNT fibre OD, CF wt%, CF length,
CF OD, graphene wt%, graphene sheet thickness, graphene sheet length, method of
testing, sliding distance, sliding speed, load, lubrication and hardness. N represents
the number of hidden layers as required by the user ranging between 3 and 15. 1
represents the output variables which are computed separately. The output variables
under consideration are COF and wear rate.
The interconnection developed between each layer is through functions. Tan-
hyperbolic (tan h) function acts as link between input and hidden layers. At each
hidden unit (H j ), the weighted summation of the input (X i ) is operated on by a
nonlinear tan hyperbolic transfer function as shown in equation below, which ensures
that each input contributes to every hidden unit [26].

H j = tan h W ji X i + θ j (1)
The output node then calculates the linear weighted sum of the outputs of the
hidden layers.

Yk = Wk j H j + θk (2)
where θ denotes biased variable, W ji and W kj represent the weight assigned to the
connections between the layers, H j denotes output from hidden units, Y k denotes the
output and X i denotes the normalized inputs
3.2 Sensitivity Analysis
In order to develop any new material, it is essential to have better knowledge on

the absolute influence of the various input parameters in the final tribological prop-
erties of the composites. But, in case of ANN models, due to its tangled hidden
relationships, it is quite tedious to determine consequence of the input variables on
the output. The same can be overcome by means of sensitivity analysis that helps to
find all applicable factors from a set of possible factors. It is feasible to determine the
correlation between the input and output parameters through various methods. One
among them is the connection weight method [25] that is considered in this study. To
find the relevance importance on the parameters, it uses the connection weights of
input to hidden layers and hidden to output layers from the developed ANN model.
3.3 Genetic Algorithm (GA)
Genetic algorithm is a known tool for optimization problems which is used widely
for solving constrained and unconstrained optimization problems that is based on
natural selection that thrusts biological evolution. Three rules are used to build a
genetic algorithm optimization, namely selection, crossover and mutation. GA per-
forms multiple modifications of a selected population of individual solutions. GA
utilizes its objective function to select a random set member from the model which
is fittest among the rest and allocates them as parents so that they can be utilized
to make the next-generation children. This selection leads to the next rule where
two parents from different generations are crossed over with each other to achieve
better solutions. Further, certain characteristics of the chosen parent member may
also be varied slightly to arrive at optimal solutions which is called mutation [26].
This same process repeats itself for several generations until an optimum solution
is evolved from these generations. The last generation among the entire population
gives the optimal solution as desired. The optimization in GA is implemented by
specifying an objective function (in this case an ANN model) on a single-objective
or multi-objective function with or without constraints [27]. This study comprises
individually monitoring the performance of the optimization by minimizing each of
the output parameters. In this case, both COF and wear rate have to be minimized
to get better tribological performance of the composite. This is done by varying the
molecular weight of UHMWPE on ANN model of each output within a specified
range and checking if the output parameters are getting minimized or maximized.
Also, constraints were added on varying the sum of weight percentages of CNT and
graphene as 2, 4 and 6% for the betterment of solutions.
4.1 ANN Modelling and Sensitivity Analysis
Sixteen input variables were taken into account to develop two ANN models for two
sets of outputs separately, viz. COF and wear rate. 3–15 hidden nodes were taken into
account for the iteration to develop the best models. The model with least amount of
error or closest to the target value is taken as the best one for further studies. Figure 1
Fig. 1 Predictability of ANN model with linear fit for a wear rate and b COF
-1
-2
COF Wear rate
-3
Fig. 2 Sensitivity analysis
shows the predictability of the trained ANN models. It can be inferred from the
scattered plot that the models developed are closer to the target value. The regression
value for COF is 0.96801, and similarly that of wear rate is 0.72567.
The model is confirmed not only from its regression plot but also should have
good input-to-output relationship as well. The regression curve does not provide any
information about the influences of input variables on the output due to its complex
hidden relationships developed which is why ANN is also called a ‘Black Box’.
The sensitivity analysis is done on a trained ANN model which gives the individual
contribution of each input variable to the output. More the number of input variables
on the positive side, the better are the results. The sensitivity analysis for the above
two best models is shown in Fig. 2.
Surface plots are developed by keeping the input variables fixed and observing
their influences on the output. It is plotted by taking two input variables and one
output variable at a time to perceive the influence of input variables on the output.
Figure 3a shows that as GO wt% increases from 0 to 0.5%, there is a slight decrease
in COF and increases thereafter whereas in case of CNT wt%, there is a slight decrease
in COF value from 0.5 to 2%. In Fig. 3b, the COF increases as CF wt% increases
from 0 to 15%, whereas for CNT, COF decreases from 1 to 2 wt%. Also from the
earlier study, Naresh Kumar et al. [12] have proposed that COF decreases with an
increase of CNT wt% from 0.5 to 2%. Figure 3c shows the influence of load and
molecular weight on COF. As molecular weight increases from 2 to 6 million g/mol,
COF also increases and as load increases from 0 to 100 N, COF increases as well.
In Fig. 3d, the relationship of sliding speed and sliding distance with COF is shown.
COF increases with increase in sliding distance from 0 to 4000 m, whereas COF
decreases with increase in sliding speed up to 0.2 m/s and then increases thereafter.
Figure 4a depicts the relationship between CF and CNT wt% with respect to wear
rate. It shows that as CNT wt% increases, wear rate decreases drastically, whereas
as CF wt% increases from 0 to 15 wt%, wear rate is almost constant. Also from the
earlier study, Naresh Kumar et al. [12] have studied that with an increase of CNT
wt% from 0.5 to 2%, wear rate decreases. In Fig. 4b, it can be seen that wear rate
decreases very minutely with lower value of GO wt% and decreases drastically at
higher content of GO, whereas when CNT wt% varies wear rate decreases slightly.
Fig. 3 Surface plots indicate the variation of COF with a CNT wt% versus graphene wt%, b CNT
wt% versus CF wt%, c mol. wt. versus load and d sliding dist. versus sliding speed
Figure 4c depicts the relationship between molecular weight and load with wear rate.
It shows that as molecular weight increases, wear rate increases up to 5.5 million
g/mol and then further decreases. As load increases from 0 to 150 N, wear rate also
increases. Figure 4d shows the influence of sliding speed and sliding distance with
wear rate. It could be inferred that wear rate decreases steeply with increase in sliding
distance from 5000 to 20000 m. Also wear rate remains almost constant with increase
in sliding speed.
4.2 Single-Objective Genetic Algorithm
Single-objective genetic algorithm gives the desired output for one particular output
parameter in relation to all the input parameters. In this study, the optimization is
carried out for both COF and wear rate separately to check its influences individually
on the output parameters. The solutions were run for 1000 generations and 1000
populations for the minimization of COF and wear rate by using the best-developed
Fig. 4 Surface plot indicates the variation of wear rate with a CNT wt% versus graphene wt%,
b CNT wt% versus CF wt%, c mol. wt. versus load and d sliding dist. versus sliding speed
ANN models as objective function by varying the molecular weight of UHMWPE

composites from 2.5 to 5.5 million g/mol. Suitable constraints were also applied to
the weight percentages of the reinforcements by restricting the summation of CNT
and GO weight percentage as 4% from the simulation studies that higher amount
of nanoparticles does not cause any irregularities to get optimized solutions for the
tribological properties. Table 2 indicates the parameters which are considered for the
optimization using single-objective GA.
Table 2 Value of GA
Parameters Values
parameters
Number of generation 1000
Number of population 1000
Probability of mutation 0.1
Probability of cross-over 0.95
The solution obtained from GA suggests that there is a constant decrease in COF
value with increase in molecular weight of UHMWPE composite as shown in Fig. 5a.
This is similar to the trend required for this particular study where output has to be
minimized. But on the contrary, there is a steep increase in the wear rate with increase
in molecular weight of UHMWPE composite as shown in Fig. 5b. Wear rate does not
follow the minimization function as required or expected in this study. So a further
assessment of data is required through multi-objective genetic algorithm because of
the conflicting nature of each output.
Tables 3 and 4 depict the composition of each reinforcement with the variation of
molecular weight of UHMWPE and load which are obtained through single-objective
GA.
0.12 0.000065
0.1 0.00006
Wear rate
0.000055
COF
0.08
0.00005
0.06
0.000045
0.04 0.00004
2 4 6 2 4 6
Mol. wt (Million g/mol) Mol. wt (Million g/mol)
(a) (b)
Fig. 5 Variation of output with increasing molecular weight a mol. wt. versus COF and b mol. wt.
versus wear rate
Table 3 List of CNT wt%, CF wt%, GO wt%, load and COF with respect to mol. wt
Mol. wt. (×106 g/mol) CNT wt% CF wt% GO wt% Load (N) COF
2.5 0.430 5.303 0.7794 40.936 0.1074
3.5 0.4797 4.925 0.7247 47.057 0.0931
4.5 0.5151 4.821 0.6889 43.733 0.0803
5.5 0.5014 4.424 0.7121 45.252 0.0693
Table 4 List of CNT wt%, CF wt%, GO wt%, load and wear rate with respect to mol. wt
Mol. wt. (×106 CNT wt% CF wt% GO wt% Load (N) Wear rate (mm3 /Nm)
g/mol)
2.5 0.6176 6.7829 0.4790 43.634 5.02E-05
3.5 0.6076 6.7615 0.4708 44.729 5.44E-05
4.5 0.5997 6.0503 0.5191 46.466 5.87E-05
5.5 0.5.691 6.1093 0.5197 43.382 6.08E-05
0.12 0.000065
0.1 0.00006
Wear rate
0.08
COF
0.000055
0.06
0.00005
0.04
0.02 0.000045
0 0.00004
40 45 50 42 44 46 48
Load (N) Load (N)
(a) (b)
Fig. 6 Variation of output with increasing load a load versus COF b load versus wear rate
Figure 6 shows the relationship of the output with increasing load. It could be
concluded from Fig. 6a that COF is decreasing with increasing load till 45.2 N and
thereafter increases, whereas from Fig. 6b, it can be concluded that wear rate is
increasing with load but initially it decreases up to a load of 43.6 N and increases
thereafter. The general trend of the solutions depicts that load is also showing con-
tradictory behaviour as in the case of increasing molecular weight of UHMWPE
composite.
5 Conclusions
The conclusions from this study are as follows:

• It can be seen that ANN models could act as a perfect objective function for GA
in optimization problems.
• The two computational approaches used to conduct this study can save a huge
amount of time and money for designing a new material.
• This study shows that combined reinforcement of different nanoparticles can
improve the tribological properties when added to the UHMWPE, using artificial
intelligence techniques.
• The optimization study through single-objective GA shows that there was a
conflicting nature in output parameters, i.e. COF and wear rate with respect to
molecular weight of UHMWPE and load.
• The composition of all the three reinforcements, i.e. CNT, CF and GO, has dif-
ferent effects on the tribological properties. All three nanoparticles have a unique
contribution to the output parameters.
• Due to the confliction on each output, further study can be done using multi-
objective optimization to arrive at better results and to compare both the
tribological properties in a single frame.
References
1. Campbell P, Beaulé PE, Ebramzadeh E, Leduff M, De Smet K, Lu Z, Amstutz HC (2006) The

John Charnley Award: a study of implant failure in metal-on-metal surface arthroplasties. Clin
Orthop Relat Res 453:35–46
2. Affatato S, Ruggiero A, Merola M (2015) Advanced biomaterials in hip joint arthroplasty.
A review on polymer and ceramics composites as alternative bearings. Compos Part B Eng
83:276–283
3. Mikić Z, Lesić AR (2013) 50 years of total hip prosthesis—a tribute to Prof. Sir John Charnley.
Acta Chir Iugosl 60(1):9–13
4. Blumm J, Lindemann A, Meyer M, Strasser C (2010) Characterization of PTFE using advanced
thermal analysis techniques. International Journal of Thermophysics 31(10):1919–1927
5. Mano JF, Sousa RA, Boesel LF, Neves NM, Reis RL (2004) Bioinert, biodegradable and
injectable polymeric matrix composites for hard tissue replacement: State of the art and recent
developments. Compos Sci Technol 64(6):789–817
6. Vadivel HS, Golchin A, Emami N (2018) Tribological behaviour of carbon filled hybrid
UHMWPE composites in water. Tribol Int 124:169–177
7. Datta S, Chattopadhyay PP (2013) Soft computing techniques in advancement of structural
metals. Int Mater Rev 58(8):475–504
8. Sinha A, Sikdar Dey S, Chattopadhyay PP, Datta S (2013) Optimization of mechanical property
and shape recovery behavior of Ti-(∼49 at.%) Ni alloy using artificial neural network and
genetic algorithm. Mater Des 46:227–234
9. Bleck W, Sikdar Dey S, Bhattacharjee D, Bhattacharyya T, Brat Singh S, Bhattacharyya S
(2012) Microstructural prediction through artificial neural network (ANN) for development of
transformation induced plasticity (TRIP) aided steel. Mater Sci Eng A 565:148–157
10. Mohanty I, Bhattacharjee D, Datta S (2011) Designing cold rolled if steel sheets with optimized
tensile properties using ANN and GA. Comput Mater Sci 50(8):2331–2337
11. Bin Ali A, Abdul Samad M, Merah N (2017) UHMWPE hybrid nanocomposites for Improved
Tribological Performance Under Dry and Water-Lubricated Sliding Conditions. Tribol Lett
65(3):1–10
12. Naresh Kumar N, Yap SL, Samsudin B, Dayana FN, Khan MZ, Srinivasa P, Sreekanth R
(2016) Effect of argon plasma treatment on tribological properties of UHMWPE/MWCNT
nanocomposites. Polymers (Basel) 8(8):295
13. Manoj Kumar R, Sharma SK, Manoj Kumar BV, Lahiri D (2015) Effects of carbon nan-
otube aspect ratio on strengthening and tribological behavior of ultra high molecular weight
polyethylene composite. Compos Part A Appl Sci Manuf 76:62–72
14. Liu B, Qi Y, Yan X, An Y, Pei J, Xue Q, Tai Z (2013) Friction and wear properties of graphene
oxide/ultrahigh-molecular-weight polyethylene composites under the lubrication of deionized
water and normal saline solution. J Appl Polym Sci 131(1):1–11
15. Tai Z, Chen Y, An Y, Yan X, Xue Q (2012) Tribological behavior of UHMWPE reinforced
with graphene oxide nanosheets. Tribol Lett 46(1):55–63
16. Chukov DI, Stepashkin AA, Maksimkin AV, Tcherdyntsev VV, Kaloshkin SD, Kuskov KV,
Bugakov VI (2015) Investigation of structure, mechanical and tribological properties of short
carbon fiber reinforced UHMWPE-matrix composites. Compos Part B Eng 76:79–88
17. Tenison N, Baena JC, Yu J, Peng ZX (2017) Development of Mixing Methods of
UHMWPE/Carbon Nanotubes (CNT) Composites for Use in Artificial Joints. Key Eng Mater
739:81–86
18. Zoo YS, An JW, Lim DP, Lim DS (2004) Effect of carbon nanotube addition on tribological
behavior of UHMWPE. Tribol Lett 16(4):305–310
19. Rocha LA, Kanagaraj S, Oliveira MS, Simoes JA, Mathew MT, Fonseca A (2010) Tribological
characterisation of carbon nanotubes/ultrahigh molecular weight polyethylene composites: the
effect of sliding distance. Int J Surf Sci Eng 4(4–6):305–321
20. Baena JC, Peng Z (2018) Dispersion state of multi-walled carbon nanotubes in the UHMWPE
matrix: effects on the tribological and mechanical response. Polym Test 71:125–136
21. Golchin A, Wikner A, Emami N (2016) An investigation into tribological behaviour of multi-
walled carbon nanotube/graphene oxide reinforced UHMWPE in water lubricated contacts.
Tribol Int 95:156–161
22. Sreekanth PSR, Kanagaraj S (2015) Influence of multi walled carbon nanotubes reinforcement
and gamma irradiation on the wear behaviour of UHMWPE. Wear 334–335:82–90
23. Wang Y, Yin Z, Li H, Gao G, Zhang X (2017) Friction and wear characteristics of ultrahigh
molecular weight polyethylene (UHMWPE) composites containing glass fibers and carbon
fibers under dry and water-lubricated conditions. Wear 380–381:42–51
24. Lee JH, Kathi J, Rhee KY, Lee JH (2010) Wear properties of 3-aminopropyltriethoxysilane-
functionalized carbon nanotubes reinforced ultra high molecular weight polyethylene
nanocomposites. Polym Eng Sci 50(7):1433–1439
25. Olden JD, Joy MK, Death RG (2004) An accurate comparison of methods for quantifying
variable importance in artificial neural networks using simulated data. Ecol Model 178:389–397
26. Datta S (2016) Materials design using computational intelligence techniques. CRC Press, UK
27. Mathworks C (2016) Global optimization toolbox user’s guide R 2016 b
Friction and Wear Study of Laser
Surface Textured Ti-6Al-4V Against Cast
Iron and Stainless Steel Using
Pin-on-Disc Tribometer
N. Sankara Manikandan, M. Prem Ananth, L. Poovazhagan, B. Sudarsan,
and A. Vishnuvarthan
Abstract The reduction in the friction plays a vital role in improving the efficiency
of the working machinery. The wear is the direct effect of friction which reduces the
lifetime of the machine components. The modification of the surface of the materials
can considerably reduce the wear and friction and thus improving the efficiency
and increasing the lifetime of the machine components. The surface modification
techniques such as abrasive jet machining, ECM, etching and LST are being used to
alter the surface properties. This paper presents the laser surface texturing on titanium
alloy, cast iron and stainless steel and testing of frictional properties such as friction
and wear of titanium pins against cast iron and stainless steel discs. The texturing
patterns are designed and laser ablation method is used to texture the surfaces of the
pins and the discs. Since the friction is directly proportional to the contact area, the
reduction in the surface contact area of the interacting surfaces is obtained through
LST which plays a vital role in the reduction of friction wear. The textured patterns
on the surface of the materials also act as a reservoir for the debris particles to
get entrapped in them and reduce the three-body abrasion. The textured pins are
tested against the textured discs using pin-on-disc tribometer. The friction and wear
characteristics of the pin and the disc are determined and are compared against
the untextured disc and pins. The experimental analysis of the textured surfaces
established to reduce the friction and wear characteristic behaviour.
Keywords Laser surface texturing · Three particle abrasion · Pin-on-disc

tribometer · Friction and wear
N. Sankara Manikandan · M. Prem Ananth (B) · B. Sudarsan · A. Vishnuvarthan

Department of Mechanical Engineering, Sri Venkateswara College of Engineering, Chennai 602
117, India
L. Poovazhagan
Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai 603 110, India

https://doi.org/10.1007/978-981-15-4488-0_71
846 N. Sankara Manikandan et al.
1 Introduction
Tribology mainly focuses on the reduction of friction and wear between the contact-
ing surfaces in various applications and between various materials [1]. The tribology
is aimed for friction reduction and the reduction in friction can be achieved by var-
ious methodologies which mainly includes the lubricating the contacts which are
in relative motion [2]. The lubrication decreases the friction by not permitting the
two surfaces that interact with each other; rather it comes in contact between both
the surfaces, thus reducing the friction and wear of the materials. There are some
other potential methods which are currently being practiced by many tribologists,
one of which being the surface modification techniques. There are different surface
modification techniques which are used to reduce the real contact area between the
contacting surfaces and thereby to increase its tribological behaviour. Since the fric-
tion is directly proportional to the contact area, the reduction in the surface area
reduces the friction, which resulted in the reduction in the wear rate of the com-
ponents in contact. The various surface modification techniques [3] used in present
days are etching, pulsed arc air treatment [4], abrasive jet machining, ECM and LST.
Laser surface texturing is most widely used technology [5] because of its greater
accuracy and size control. The LST is prepared on the surfaces of the materials and
the predefined patterns are formed on the surface of the components. In this work,
the patterns that are formed on the surface are both circular and hexagonal dimples
on the disc and only circular dimples are formed on the pins. These texturing patterns
are formed on the basis of calculations done which generated a texturing density of
about 30%. The friction reduction is found to be maximum at the surface density of
30% [6]. The experiment is carried out on a pin-on-disc tribometer testing facility
which has rotating disc and the pin made to contact with the pin along with the
predefined normal load. Using the experimental apparatus, the frictional coefficient
and the wear rate on the disc and the pins were calculated and reported.
2 Experimentation
2.1 Materials and Machining
The pin and the disc materials for the experiments are titanium alloy (Grade 5), cast
iron and stainless steel. The titanium alloy is selected for the pin material and the
cast iron and stainless steel are selected as the disc materials. Table 1 shows the
constituents of the titanium alloy, cast iron and stainless steel materials used in the
experiment.
The pin and the disc materials are machined according to the dimensions required
for the pin-on-disc tribometer. The dimensions for the pin are 8 mm diameter and
Friction and Wear Study of Laser Surface Textured … 847
Table 1 Composition of
Element Percentage by weight
materials selected for the
experiment Titanium Cast iron (%) Stainless
alloy (%) steel (%)
Titanium 90 – –
Aluminium 6 – –
Vanadium 4 – –
Carbon <0.10 3.2–3.5 ≤0.15
Oxygen <0.20 – –
Nitrogen <0.05 – –
Hydrogen <0.0125 – –
Iron <0.3 93 84.3–88.5
Manganese – 0.60–0.90 ≤1.0
Phosphorous – 0.12 ≤0.040
Silicon – 2.0–2.4 ≤1.0
Sulphur – 0.15 ≤0.030
Chromium – – 11.5–13.5
28 mm in length, whereas the disc is machined to 8 mm thick and 165 mm diameter

with holes drilled into it according to the given Fig. 1a. The various machining
operations performed for the disc and pin dimensioning are cutting, turning, facing,
lapping and surface grinding and the fabricated disc is shown in Fig. 1b.
Fig. 1 a Disc dimensions as per ASTM standard b fabricated counter-body disc

3 Laser Surface Texturing
3.1 Texturing
The laser surface texturing is a laser ablation process in which the material becomes
vapourized when the laser beam comes in contact with the material surface [7]. There
are several texturing patterns that can be textured on the surface of the material that
will help in improving tribological properties of the surfaces [8]. The laser surface
texturing on the pin and disc surfaces are prepared using the laser marking system
supported by EzCAD software. The EzCAD software is used to design the required
patterns. The laser source is incorporated by a diode-pumped yttrium-doped fibre
source. The wavelength of the laser is 1064 nm and the power used for the texturing
is 20 W. The patterns that are made on the pins are circular and hexagonal dimples
whereas circular dimples are engraved on the discs. The diameter of the dimple is
calculated based on the required texturing density. The texturing density is taken as
30% which is having a greater reduction in friction. The diameter and pitch (distance
from centre of one dimple to the centre of the other dimple) for the circular dimples
are 90 µm and 140 µm, respectively, for titanium pins [9] and 50 µm and 100 µm,
respectively, for the discs. The titanium alloys are also characterized to show poor
wear resistance properties which can be improved by surface texturing methodologies
[10, 11]. The hexagonal dimples are patterned on the pins which have a side length
of 45 µm and pitch of 140 µm. Single-pulse laser beams are preferred to texture the
surfaces instead of multiple overlapped pulses. In multiple overlapped pulses, the
total energy exposure per unit area of the textured surface is higher than the single
pulse. Single-pulse laser beams are preferred to overcome the heat cumulative effect
caused by multiple overlapped pulses. Also, due to the influence of heat cumulative
effect, unique structures are created and this effect is prevented by using single-beam
pulse [12].
3.2 Sem Analysis
After the texturing is done, the samples are taken for SEM analysis where the textures
are captured. The following images show the textures done.
Figure 2a shows the SEM image of circular dimples that are formed in the disc
by laser marking system. Oxide layers are formed on the surface of the titanium
pins due to interaction of the atmospheric oxygen with the free aluminium on the
textured surface which is beneficial [13]. Figure 2b shows the formation of hexagonal
dimples on the surface of the pin. The rectangular array patterns are formed on the
pin surfaces and it is evident from Fig. 2. The textured and untextured pins are tested
against the discs using the tribometer.
Fig. 2 SEM image of a circular dimples and b hexagonal dimples
3.3 Testing
The pins are tested against the discs using the pin-on-disc tribometer. The pin-on-
disc tribometer offers uni-directional sliding motion by rotating in one direction
only. The tribometer has sensors and transducer to detect parameters like force and
displacement. The force sensor, which is a load cell type which measures the frictional
force between the contacting surfaces, linear variable differential transducer (LVDT),
to determine the wear of the pin material and a temperature sensor to acquire the
temperature at the interacting surfaces. The disc and the pin are fixed on the given
fixtures and the load of 25 N is applied on the pin through a horizontal lever [14].
The sliding velocity is taken as 2 m/s for the testing. The frictional force between
the pin and the disc was continuously monitored and measured by using the force
sensor. The wear on the disc and the pin is calculated by measuring the weight of
the specimen before and after the testing. The experiment was done under constant
sliding velocity conditions. In order to maintain the same sliding velocity, the rpm
was varied according to the track diameter. The following Table 2 shows the list of
experiments that are done and the parameters that are used for the testing.
Table 2 List of experiments

S. No. Disc Pin Track diameter (mm) Speed (rpm)
1 SS-Untextured Untextured 50 797
2 SS-Untextured Circular dimple 70 550
3 SS-Untextured Hexagonal dimple 90 425
4 CI-Untextured Untextured 50 797
5 CI-Untextured Circular dimple 70 550
6 CI-Untextured Hexagonal dimple 90 425
7 CI-Circular dimples Untextured 70 550
The tests are successfully carried out under the given parameters and the graph is
plotted for comparison of tribological properties of friction and wear. Figure 4 shows
the variation of friction coefficient with respect to sliding distance. Each test is briefly
discussed below.
Test 1: SS-Untextured Disc versus Untextured Pin

The above Figs. 3 and 4 show that there is a high friction between the surfaces
because of the direct contact between the materials. The observed wear rate is also
very high because of the three-particle abrasion mechanism. The initial wear rate is
observed to be very high due to sudden contact of both the pin and the disc.
Fig. 3 Comparison of wear versus sliding distance
Fig. 4 Comparison of coefficient of friction for different distance

Test 2: SS-Untextured Disc versus Circular Dimple Pin

The friction and wear are reduced in this test when compared with the untextured
pin which is evident from the graphs shown in Figs. 3 and 4. The textures on the pin
act as a reservoir for capturing the debris particles which cause the reduction in the
further wear. The reduction in friction is also reduced because of the reduction in the
contact area.
Test 3: SS-Untextured Disc versus Hexagonal Dimple Pin

The test result obtained for this test yielded the same result as that of the previous
test. Except for the fact that both the wear and friction coefficient for this test is less
compared with that test, which can be seen from graphs shown in Fig. 3 and 4. And
it can also be noted that the wear is gradual.
Test 4: CI-Untextured Disc versus Untextured Pin

The graphs obtained from the test are shown in Figs. 3 and 4 which indicate that the
wear is less initially and is gradually increased. This is because both the materials
are of equal in hardness. Hence, the friction as well as the wear is initially less.
After a period of some time, due to the heat generated between the disc and the pin,
the wear is gradually increased and because of the debris particles, the friction also
got increased. The wear is less on both the disc and the pin compared to the other
combinations.
Test 5: CI-Untextured Disc versus Circular Dimple Pin

The above test obtained the desired results. The wear rate is much lower compared
with the other and also the friction coefficient is good. Since both the pin and the
disc are hard materials, the friction is good. The cause for wear reduction being the
texturing which accumulated the wear particles.
Test 6: CI-Untextured Disc versus Hexagonal Pin The wear and friction coeffi-
cient values for this test and the previous test are nearly same and the graphs seem to
coincide with one other as seen in Figs. 3 and 4. This shows that the texturing pattern
seems to reduce the wear and friction values. But in hexagonal dimples, friction and
wear rates are much less when compared with circular dimples.
Test 7: CI-Circular Dimple Disc versus Untextured pin

Figures 3 and 4 show the graph between and wear and friction, respectively, against
the sliding distance obtained from this test. It can be seen that the wear rate is very
much reduced due to laser texturing. The wear particles are accumulated in the
dimples made in the disc. Since the disc has a larger area, most of the wear particles
are captured by it. So, there is a direct contact between the disc and the pin which
accounts for higher friction. The sudden drop in the friction after certain time is
because of the thermal fluctuations in the environment
All the tests listed above are experimentally carried out, and the frictional force
values and the wear at a given interval of time are continuously recorded. The specific
wear rate and the average coefficient of friction values are derived and plotted as the
graphs in Figs. 5 and 6. Figure 5 shows that there is a major reduction in the wear
rate in the textured discs and pins compared with the untextured discs and pins.
Experiment 7 shows that there is a major reduction in wear rate when the disc is
textured. This is because a major area of the disc is textured. Since the disc has a
greater area and contact area is reduced due to texturing, the wear rate is significantly
reduced.
Figure 6 shows the reduction in the coefficient of friction between the titanium
alloy pins and the discs during various untextured and textured analysis. This graph
shows that the friction coefficient is higher during untextured discs and pins. The
specific wear rate is significantly higher for the experiments 1 and 4 when compared
Fig. 5 Comparison of specific wear rate
Fig. 6 Comparison of average coefficient of friction

with the rest because in those tests both the disc and the pins were untextured, which
results in the increased surface contact area which directly implies on the increased
three-particle abrasion. The circular and hexagonal textured dimples act as a reservoir
for the debris particles generated during the contact between the disc and the pin.
These debris particles, in untextured samples, come in between the disc and the pin
and result in the three-particle abrasion. But they get encapsulated in these dimples.
In comparison with the circular and hexagonal dimples, the hexagonal dimples have a
reduced wear because of a considerable increase in the dimple area and considerably
reduced surface contact area. And also it is evident that there is a higher reduction in
friction as quoted in [6]. By combining the two graphs, it can be found that experiment
3 (SS UT—hexagonal dimple) has shown improved tribological properties compared
with the rest of the combinations.
5 Conclusions
The above tests clearly show that the texturing has reduced the friction and wear
between the interacting materials when compared with the untextured surfaces. The
specific wear rate is very less for the stainless steel (untextured disc)-hexagonal
dimple pin combination. It has also been found that the coefficient of friction is also
low between them. This is nearly 24% reduction in friction and 36% reduction in
wear when compared with the rest of the tests. Hence, the texturing has helped in
the reduction of friction and wear rate. Friction is directly proportional to the contact
area. The coefficient of friction obtained after the texturing and coating is reduced
because of the less surface contact area between the disc and the pin material. This
reduction in the surface area contributes to the major reduction in friction. The wear
is directly proportional to the friction between the contacting surfaces. The wear is
reduced since the friction between the surfaces is reduced due to the reduction in
the contact area. The above experiments are evident for reduction in friction is due
to the reduction in contact area offered by texturing. The surface texturing reduced
the area by 30%, which contributed to the reduction in friction. Since the friction
between the surfaces and the contact area is reduced, the wear is also reduced. The
textures formed on the surfaces acted as a reservoir for the debris particles. And
hence, friction and wear are reduced.
References
1. Lia L, Guob YB, Weia XT, Lib W (2013) Surface integrity characteristics in wire-EDM of
inconel 718 at different discharge energy. In: 7th international symposium on electro machining,
procedia CIRP, vol 6, pp 220–225
2. Hutchings IM (1992) Tribology- friction and wear of engineering materials. Edward Arnold,
London
3. Totten GE (1992) ASM handbook friction, wear and lubrication technology. vol 18
4. Ibatan T, Uddina MS, Chowdhury MAK (2015) Recent development on surface texturing in
enhancing tribological performance of bearing sliders. Surf Coat Technol 272:102–120
5. Moshkovitih A, Perfiiiev V, Glindin D, Parkansky N, Boxman R, Rapoport L (2007) Surface
texturing using pulsed arc air treatment. Wear 253:56–77
6. Zavestovskaya I (2010) Laser nano structuring of materials surfaces. Quantum Electron
11:942–954
7. Li K, Yao Z, Hu Y, Gu W (2014) Friction and wear performance of laser peen textured surface
under starved lubrication. Tribol Int 6(11):97–105
8. Etsion I (2005) State of the art in laser surface texturing. J Tribol Trans ASME 127:248–53
9. Galda L, Pawlus P, Sep J (2009) Dimples shape and distribution effect on characteristics of
Stribeck curve. Tribol Int 42:1505–1512
10. Laser surface texturing of Ti6Al4V alloy, stainless steel and aluminium silicon alloy Andrzej
Grabowski, Maria Sozańska, Marcin Adamiak, Mirosława K˛epińskaa, Tomasz Floriand
11. PremAnanth M, Ramesh R (2014) Tribological improvement of titanium alloysurfaces through
texturing and TiAlN coating. Surf Eng 30:758–762
12. Yerramareddy S, Bahadur S (1992) The effect of laser surface treatments on the tribological
behavior of Ti–6Al–4V. Wear 157:245–262
13. Weber R, Graf T, Berger P, Onuseit V, Wiedenmann M, Freitag Ch, Feuer A (2014) Heat
accumulation during pulsed laser materials processing. Opt Express 22:11313
14. Basiaga M, Kajzer W, Walke W, Kajzer A, Kaczmarek M (2016) Evaluation of physicochemical
properties of surface modified Ti6Al4V and Ti6Al7Nb alloys used for orthopaedic implants.
Mater Sci Eng C 68:851–860
Tribological Study on Sliding Contact
Between Laser Surface Textured
Titanium and Aluminium Alloy Under
Lubrication
Ramesh Rajesh, M. Prem Ananth, Sangam Harish, Sakthivel Balaji,

and R. Sivaguru
Abstract The friction plays a major role whenever two surfaces come into contact.
The friction is a major factor which affects the efficiency of a working component.
The life of the component depends on the wear it undergoes during working. This
wear is directly related to the friction that is produced between the surfaces. It is
impossible to completely avoid friction. But it is possible to reduce the friction
between the contact surfaces by various methods. One of the most promising meth-
ods in recent days is the laser surface texturing. There are various developments
currently being done in the LST. It is done by altering the various parameters of the
textures. Some of the parameters include the pattern, depth, diameter, pitch and so
on. In our tribological study, a pin-on-disc tribometer is used to determine the fric-
tion and the wear between the components made of Ti-6Al-4V and AA7075 alloys
under liquid lubrication condition. For improving the tribological characteristics of
the contacting pairs, laser surface texturing is performed on both pin and disc sur-
faces. The laser surface texturing concept is implemented to create micro-dimples on
the contact surface which the dimples act as a reservoir for the lubricant enhancing
lubrication to reduce the coefficient of friction, the coating material to improve its
bonding characteristics, wear debris to get accumulated. The desired dimple param-
eters are chosen for the laser surface texturing. The laser surface texturing followed
by a surface treatment has improved the wear-resisting performance of the contact
surfaces.
Keywords Micro-dimples · Ti-6Al-4 V · Pin-on-disc · Coefficient of friction and

wear debris
R. Rajesh · M. Prem Ananth (B) · S. Harish · S. Balaji · R. Sivaguru

Department of Mechanical Engineering, Sri Venkateswara College of Engineering,
Sriperumbudur, Kancheepuram, Tamil Nadu 602117, India

https://doi.org/10.1007/978-981-15-4488-0_72
856 R. Rajesh et al.
1 Introduction
The aluminium alloys and titanium alloys (Ti-6Al-4V) are extensively used in mod-
ern industries due to its high strength–weight ratio and excellent corrosion resistance
which led to a comprehensive and expanded range of applications, especially in the
aerospace and automotive industries. Since the titanium and aluminium alloys are
used for high optimum operations, the tribology characteristics such as friction and
wear play an important role in the relative sliding contact between AA7075 and
Ti-6Al-4V. As the titanium alloy is low in wear resistance and possesses high coeffi-
cient of friction, engineering components made from the same seek improvement in
wear resistance and friction coefficient. It is necessary and important to attempt for
improving wear resistance performance and to reduce friction [1]. These problems
can be overcome by changing the nature of the surface of titanium and titanium
alloys using thermo-chemical treatments such as nitriding. The surface is changed
from titanium to a hard compound of titanium. Gas and plasma nitriding are among
the most widely used thermo-chemical treatments for improving the surface prop-
erties of titanium alloys [2]. Among several surface engineering techniques, laser
surface texturing (LST) is adopted to improve the tribological characterization of the
contact. LST is used to create patterned micro-dimples on the surface of the materials
which can improve load capacity, wear rates and lubrication and reduce friction coef-
ficients. The micro-dimples on the contact surfaces act as a reservoir for the lubricant
enhancing lubrication performance to reduce the coefficient of friction, the nitriding
process to improve its bonding characteristics, the wear debris to get trapped in the
micro-reservoir during the relative motion between the two contact surfaces. The
patterned geometry permits for countless miniature lubricant reservoirs, providing
direct and immediate lubricant relief for starved areas [3]. However, it was observed
that LST may be deleterious to tribological performance under starved lubrication
conditions when the dimples are relatively deep or when the oil viscosity is relatively
high [4]. During the nitriding process, the micro-dimples act as fins creating more
exposure area to the surface for the nitriding process which permits the ammonia
particles to deposit on the micro-reservoirs.
2 Materials
2.1 Pin and Disc
The materials Ti-6Al-4V and AA7075 are selected as pin and disc, respectively,
because they are finding more applications in the industry in recent days because of
the mentioned properties. The specific composition of these components is mentioned
in Table 1. The EDS analysis of textured Ti-6Al-4V surface is shown in Fig. 1 which
Tribological Study on Sliding Contact Between Laser Surface … 857
Table 1 Composition of
Element Ti-6Al-4V AA7075
material
Titanium 90% <0.2
Aluminium 6% 87.1–91.4%
Vanadium 4% –
Carbon <0.10% –
Oxygen <0.20% –
Nitrogen <0.05% –
Hydrogen <0.0125% –
Iron <0.3% <0.5%
Chromium – 0.18–0.28%
Copper – 1.2–2%
Magnesium – 2.1–2.9%
Manganese – <0.3%
Silicon – <0.4%
Zinc – 5.1–6.1%
Fig. 1 EDS analysis of

Ti-6Al-4V textured surface
Fig. 2 Untextured lapped

AA7075 disc
confirms the major composition of Ti, Al and V present in the pin material. The
AA7075 and Ti-6Al-4V contact pair have been used extensively in aerospace and
automotive applications. For instance, in high-speed racing car engines and aerospace
engines, this AA7075 and Ti-6Al-4V contact pair is employed in cam and follower, in
which AA7075 is used as cam and Ti-6Al-4V is used as a follower. Since the follower
experiences the constant load at a point at any instance and the cam experiences a
cyclic load at a point in periods, AA7075 is chosen as disc and Ti-6Al-4V is chosen
as pin.
According to the ASTM standards for the pin-on-disc test, pin acts as the target
or the test. Friction and wear properties are studied for the pin materials extensively.
The pin is machined of 8-mm diameter and 28-mm length. The disc is machined of
165-mm diameter and 8-mm thickness, and its macro-image is shown in Fig. 2. The
weight of the pins before and after the testing is measured to calculate the specific
wear rate (SWR).
2.2 Lubricant
In most of the moving components, a lubricant is used to reduce the friction, reduce
the heat generated during the operation and also carry the debris from the interacting
surfaces. Hence to simulate a real-time working condition, a lubricant is chosen.
The oil 10 W-40 has been chosen for the lubrication due to high and maximum
performance, high viscosity index, high dirt dispensing capacity, excellent cleaning
capacity and high thermal stability. Since the lubricant has the above properties, it
is used when the test is run on the tribometer. The properties of the lubricant are
provided in Table 2.
Table 2 Properties of
Density 827 kg/m3
lubricant
Kinematic viscosity @100 °F 15.4 cSt
Viscosity index 149
Flashpoint 220 °C
Pour point 33 °C
The pin and disc which are to be tested in pin-on-disc tribometer are machined
with respect to the specifications of the tribometer. The machined pins and discs are
grinded and lapped to achieve low roughness coefficient. The laser surface texturing
process is carried out on both pin and disc to obtain designed dimple geometry on
the surface of the specimens. After laser surface texturing, the pins and discs are
washed by using acetone to remove any debris on the surface of the specimens. The
specimens were kept immersed for about 10 min to remove all dust, grease and oxides
formed on the surface as any impurities will cause effect on experiment results [5,
6]. Then, the pins and discs are placed in a clean environment and kept ready to
run experiments on the tribometer. The weight of the pins before and after testing is
measured to calculate the specific wear rate for every experiment.
4 Laser Surface Texturing
The surface of the machined and processed discs and pins is textured using the laser
marking system. There are several parameters to be determined before marking.
Some of the parameters include texture density, dimple diameters, dimple structure,
dimple pattern, dimple depth and pitch [6–8].
Type of laser used: Nd:YAG, power: 20 W.
The density of the texturing is arbitrarily chosen as 30% for study. For the study,
circular dimple having a hemispherical bottom is chosen for analysis. The diameter
and pitch are calculated. The dimpled surface topography is used for lubricant storage,
and the dimple area percentage is used as a variable parameter to express dimple
availability (1) [9–11]:
D − area = π (d/2 p)2 × 100% (1)
where d is the diameter of the dimple and p is the pitch of the dimple. The dimple
geometry for different dimple diameters of 30% D-area has been calculated and given
in Table 3.
The above texturing patterns are designed using the AutoCAD software and are
fed into the EZCAD software. Figures 3 and 4 show the pattern made using the
Table 3 Dimple geometry

D-area % D, 10−6 m p, 10−6 m Area, 10−12 m2 Closest distance between the
circumference
of adjacent circular dimples,
10−6 m
30 150 242.752078 17,678.5714 92.75207811
30 120 194.201662 11,314.2857 74.20166249
30 90 145.651247 6364.28571 55.65124687
Fig. 3 CAD diagram of

textured surface on pin
Fig. 4 CAD diagram of

dimple pattern
AutoCAD software. The patterned disc and pin are then analysed under scanning
electron microscope, and the image is captured.
Figures 5 and 6 show the SEM images obtained after the texturing. The images
show that the dimples are formed as a rectangular array. It can also be noted from the
above images that the debris formed during the laser ablation process is aggregated
near the formed dimples. The debris formed is cleaned using acetone.
Fig. 5 500× zoom SEM

image of textured pattern
Fig. 6 130× zoom SEM

image of textured pattern
5 Pin-on-Disc Tribometer
The apparatus consists of a pin that is pressured against a rotating disc. According to
the Society for Testing and Materials (ASTM), the standards that can be used for the
pin and disc are ASTM G99, 2003. It allows the wear qualification by both geometry
and mass loss methods. The coefficient of friction is calculated by measuring the
tangential force acting on the specimen.
Table 4 Specifications of
Parameter Range
pin-on-disc tribometer
Pin diameter 8–12 mm
Disc size 165 mm dia
Disc thickness 8 mm
Disc rotation Up to 2000 rpm
Normal load Up to 200 N
Wear 2000 × 10−3 mm
The pin-on-disc tribometer used for conducting the experiments is the Ducom TR
20L. The Ducom wear and friction monitor TR 20LT records friction and wear in
low-temperature conditions using the LVDT and strain gauge. The Ducom wear and
friction monitor TR 20L is particularly designed for wear and friction characterization
in low-temperature conditions. This instrument consists of a rotating disc against
which a test ball is pressed with a known force. A provision for measurement of
compound wear and frictional force is provided.
The TR 20LT series comes with WinDucom software for data acquisition and
display of results. WinDucom software is used to present data in a variety of ways
and compare view provides a powerful tool to view and compare test results. Table 4
shows the specifications of the tribometer used for the analysis.
6 Testing
The surface roughness was measured randomly in four regions on each surface for
the 4-mm stroke length using Mitutoyo SJ-430 surface roughness tester, and the
average surface roughness (Ra) was calculated.
Table 5 shows the surface roughness of the pins after various surface modifications.
Table 5 shows the experiments done using the pin-on-disc tribometer.
Table 5 Surface roughness

Material Surface roughness, Ra 10−6
test on pins and discs
m
Untextured Ti-6Al-4V pin 0.87
Textured Ti-6Al-4V pin 13.09
Untextured nitrided 1.00
Ti-6Al-4V pin
Textured nitrided Ti-6Al-4V 16.65
pin
Untextured lapped AA7075 0.09
disc
Textured AA7075 disc 8.82
Table 6 List of experiments

Exp. No. Ti-6Al-4V pin AA7075 disc Test timing Track diameter Speed (rpm)
(min) (m)
1 Untextured Untextured 30 0.105 571
2 Textured Untextured 30 0.085 706
3 Untextured Untextured 30 0.065 923
nitrided
4 Textured Untextured 30 0.045 1333
nitrided
5 Untextured Textured 30 0.105 571
6 Textured Textured 30 0.085 706
7 Untextured Textured 30 0.065 923
nitrided
8 Textured Textured 30 0.045 1333
nitrided
The gas nitriding process on Ti-6Al-4V pins has reduced the surface roughness
of the textured surface of the disc which is evident from the table. The reduction in
the surface roughness on the textured surfaces is due to accumulation of ammonia
particles from gas nitriding process in the dimples which act as micro-reservoirs.
Table 6 shows the list of experiments, combinations of a different surface engi-
neered pin and disc and its parameters. All the experiments are tested with constant
sliding velocity = 3.14 m/s, and load on pin is 50 N.
Various tests discussed above are experimentally carried out, and the results of each
test are briefly discussed below.
Figure 7 shows a graphical representation which explains the frictional behaviour
of various combinations of a different surface engineered pin and disc.
Figure 8 shows a graphical representation which explains the wear behaviour of
various combinations of a different surface engineered pin and disc.
Experiment No. 1 Untextured Ti-6Al-4V PIN Versus Untextured AA-7075

DISC
Both the pin and the disc were untextured, and hence this resulted in the high
friction between them. But there is high wear in the pin which is undesirable. The
wear rate is high during the initial stages of the testing, whereas it got stable during
the later period of the test. This is due to the total surface contact between the disc
and the pin. The major cause for the high wear is the third particle abrasion, where
the debris particle from the direct abrasion stays in between the disc and the pin and
causes more wear. The average coefficient of friction is 0.056276.
Coefficient of Friction vs Sliding Distance

0.2
0.18
Co-efficient of Friction
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
74
1026
1094
1162
1230
1298
1366
1434
1502
1570
1638
1706
1774
1842
1910
6
142
210
278
346
414
482
550
618
686
754
822
890
958 Sliding Distance , m
1 2 3 4 5 6 7 8
Fig. 7 Frictional force versus sliding distance
Wear vs Sliding distance

60 1
50 2
Wear , micron
40 3
30 4
20 5
10 6
7
0
0 603 1149 1695 2240 2786 3332 3877 4423 4969 8
Sliding distance, m
Fig. 8 Wear versus sliding distance
Experiment No. 2 Textured Ti-6Al-4V PIN Versus Untextured AA-7075 DISC

The laser surface textured Ti-6Al-4V pin is tested against the untextured lapped
AA7075 disc in lubrication. The experiment resulted in a reduced coefficient of
friction and improved wear which is evident from the graph. This is due to laser
surface texturing at the surface of the pin in which the third particle abrasion by debris
is reduced as the debris was trapped in the micro-reservoirs resulting in a weight gain
in the pin after the testing. The average coefficient of friction is 0.043726.
Experiment No. 3 Untextured Nitrided Ti-6Al-4V PIN Versus Untextured AA-

7075 DISC
The untextured nitrided Ti-6Al-4V pin is tested against the untextured lapped
AA7075 disc in lubrication. On comparing the results from experiment no. 1, this
experiment shows better results in terms of less coefficient friction and improved
wear. This is due to the thermal surface treatment of gas nitriding. The average
coefficient of friction is 0.042272.
Experiment No. 4 Textured Nitrided Ti-6Al-4V PIN Versus Untextured AA-

7075 DISC
The laser surface textured nitrided Ti-6Al-4V pin is tested against the untextured
lapped AA7075 disc in lubrication. Both coefficient of friction and wear obtained
during this were low and appreciable. This is due to the combined effect of LST
and gas nitriding. The third particle abrasion by debris is reduced as the debris was
trapped in the micro-reservoirs resulting in a weight gain in the pin after the testing.
The average coefficient of friction is 0.039753.
Experiment No. 5 Untextured Ti-6Al-4V PIN Versus Textured AA-7075 DISC

The untextured Ti-6Al-4V pin is tested against the laser surface textured AA7075
disc in lubrication. Due to the absence of nitriding, the friction was observed to
be moderate and low wear due to laser surface texturing on the disc. The average
coefficient of friction is 0.048982.
Experiment No. 6 Textured TI-6AL-4V PIN Versus Textured AA-7075 DISC

The textured Ti-6Al-4V pin is tested against laser surface textured AA7075 disc
in lubrication. From the line graph, it experiences high wear which is not appreciable.
This is due to absence of nitriding effect in the pin as it was pressured against textured
surface which has high roughness. The third particle abrasion by debris is reduced
as the debris was trapped in the micro-reservoirs resulting in a weight gain in the pin
after the testing. The average coefficient of friction is 0.047006.
Experiment No. 7 Untextured Nitrided TI-6AL-4V PIN Versus Textured AA-

7075 DISC
The untextured nitrided Ti-6Al-4V pin is made to test against textured AA7075
disc. The nitrided Ti-6Al-4V pin possesses low coefficient friction. Due to the tex-
turing on the surface of the disc, the contact area was reduced resulting in moderate
wear on the pin. The average coefficient of friction is 0.042964.
Experiment No. 8 Textured Nitrided TI-6AL-4V PIN Versus Textured AA-7075

DISC.
The combination of laser surface textured nitrided pin and textured AA7075 disc
resulted in least coefficient of friction and good improved wear. This is due to the
combined effect of nitriding and texturing done on both the surfaces. The third
particle abrasion by debris is reduced as the debris was trapped in the micro-reservoirs
resulting in a weight gain in the pin after the testing. The average coefficient of friction
is 0.034296.
The coefficient of friction at contact between pin and disc for all experiments has
been observed.
From Fig. 9, it can be inferred that the LST and gas nitriding have improved the
tribological properties. The experiment no. 8 (textured nitrided pin versus textured
disc) has the least coefficient of friction.
The specific wear rate for all the experiments has been calculated and reported
in Fig. 10. From the bar chart, it can be inferred that the bars in the negative section
correspond to the weight loss of the pin after the experiment whereas bars in the
positive sections correspond to the weight gained by the pin after the experiment.
Coefficient of Friction
0.06 0.056276
Average Coefficient of Friction
0.0489820.047006
0.05 0.0437260.042272 0.042964
0.039753
0.04 0.034296
0.03
0.02
0.01
0
1 2 3 4 5 6 7 8
Experiment No
Fig. 9 Coefficient of friction at contact for all experiments
Specific Wear Rate

0.0000012 1.11727E-06
0.000001
7.98048E-07 7.98048E-07
0.0000008
0.0000006
SWR , m3/Nm
0.0000004 3.19219E-07
0.0000002
0
1 2 3 4 5 6 7 8
-2E-07-7.98048E-08 -7.98048E-08
-4E-07 -3.19219E-07
-3.99024E-07
-6E-07
Experiment No
Fig. 10 Specific wear rate of pins for all experiments

8 Conclusion
The experiments are conducted and studied to understand the tribological character-
istics of various combinations of the selected materials with and without texturing.
The wear and frictional properties of the Ti-6Al-4V have been improved drastically
after the gas nitriding process. The dimples created on the discs enhance the effect
of lubrication and significantly reduce the coefficient of friction. Since the friction
between the surfaces is reduced due to the reduction in the contact area, the wear is
also reduced. The laser surface texturing reduces the area by 30% which contributes
to a reduction in friction. It is observed that there has been some gain in weight of
textured pins after the pin-on-disc experiment which ensures deposition of debris in
the dimples of the pin from the disc particles. This characterization prevents impu-
rification of lubricating oil by the debris enhancing the service life of the components
in their applications.
Out of 8 conducted experiments, the experiment no. 8 is chosen as the best com-
bination in which the friction has been reduced by 39% and resistance improved.
Hence, laser surface textured AA7075 and laser surface textured nitrided Ti-6Al-4V
pair is suggested for the real-time engineering applications.
Acknowledgements This work was supported by the All India Council for Technical Education
(AICTE) under the Scheme of Research Promotion Scheme (RPS) with a reference number of File
No.8-50/RFID/RPS/POLICY-1/2016-17. We would like to acknowledge them for the financial and
technical support.
References
1. Muthuvel PA, Rajagopal R (2013) Influence of surface texture on tribological performance of

AlCrN nanocomposite coated titanium alloy surfaces. J Eng Tribol 227(10):1157–1164
2. Zhecheva S, Sha WM (2006) Titanium alloys after surface gas nitriding. Surf Coat Technol
201(6):4
3. Erdemir DA (2008) Tribology of diamond-like carbon films (Chapter 3). In: Laser processing.
Springer, New York
4. Kovalchenko A, Ajayi O, Erdemir A, Fenske G, Etsion I (2005) The effect of laser surface
texturing on transitions in lubrication regimes during unidirectional sliding contact. Tribol Int
38(3):219–225
5. Prem Ananth M, Ramesh R (2014) Tribological improvement of titanium alloy surfaces through
texturing and TiAlN coating. Surf Eng 30(10):758–762
6. Ananth MP, Ramesh R (2017) Sliding wear characteristics of solid lubricant coating on titanium
alloy surface modified by laser texturing and ternary hard coatings. Trans Nonferrous Met Soc
China 27(4):839–847
7. Mishra A (2014) Analysis of friction and wear of titanium alloys. Int J od Mech Eng Robot
Res 3
8. Vilhena LM, Sedlaček M, Podgornik B, Vižintin J, Babnik A, Možina J (2009) Surface texturing
by pulsed Nd: YAG laser. Tribol Int 42(10):1496–1504
9. Voevodin AA, Zabinski JS (2006) Laser surface texturing for adaptive solid lubrication. Wear
J 261:11–12
10. Borghi A, Gualtieri E, Marchetto D, Moretti L, Valeri S (2008) Tribological effects of surface
texturing on nitriding steel for high-performance engine applications. Wear J 265(7–8):1046–
1051
11. Quazi MM, Fazal MA, Haseeb AS, Yusof F, Masjuki HH, Arslan A (2015) Laser-based surface
modifications of aluminum and its alloys. Crit Rev Solid State Mater Sci J
Some Studies on Surface Roughness
of AISI 304 Austenitic Stainless Steel
in Dry Turning Operation
D. Philip Selvaraj and P. Richard Philip
Abstract In this paper, the input variables of turning process of AISI 304 stainless
steel are optimized by employing the Taguchi method under dry machining con-
ditions. The dry turning operations are conducted at three levels of depth of cut,
cutting velocity, and feed. The test results are evaluated by employing S/N ratio and
ANOVA. It is revealed that the depth, cutting speed, and feed are the important input
variables affecting the quality of the machined surface. The optimum surface finish
is attained at the combination of lower depth, lower feed, and higher cutting speed.
Keywords Austenitic stainless steel · Dry turning · Taguchi’s method ·

Optimization · Surface finish
1 Introduction
Stainless steels (SSs) contain at least 10.5% chromium (Cr) by weight in order to have
corrosion resistance. Generally, austenitic stainless steels (ASSs) have at least 16%
chromium and small amount of nickel and/or manganese. The usage of ASS alloy
is higher compared with other stainless steel alloys. They possess face-centered-
cubic (FCC) microstructure. They have good formability, weldability, and higher
toughness. ASS is employed for making cooking utensils, architectural applications,
containers, equipment for food industries, heat exchangers, and surgical implants
[1].
Surface finish is an important factor needed by the production engineers to the
effective working of the machine elements. It is required for production engineers
to obtain required surface roughness on the machine components. Ra is widely
employed parameter for surface roughness evaluation in machine tool industries.
Various reasons for surface roughness are machine tool vibration, tool feed marks,
D. Philip Selvaraj (B)

Karunya Institute of Technology and Sciences, Coimbatore 641114, India
P. Richard Philip
Karpagam College of Engineering, Coimbatore, Tamil Nadu 641032, India
https://doi.org/10.1007/978-981-15-4488-0_73
870 D. Philip Selvaraj and P. Richard Philip
and formation of built up edge [2]. In the past few decades, CNC machines are
replacing conventional machine tools in the industries due to more productivity,
better accuracy, and reduction in rejection rate [2, 3].
Kulkarni et al. [4] analyzed the influences of input cutting variables on surface
roughness and machining force of 304 SS while turning by employing carbide tool
inserts coated with AlTiCrN. They found that optimal surface finish was attained
in the range of cutting velocity from 200 to 320 m/min and feed between 0.08 and
0.2 mm/rev. Asilturk et al. [5] found out optimum machining condition for surface
finish while turning 304 SS in CNC lathe using the Taguchi technique and RSM.
They concluded that the most dominant variable for the machined surface quality
is feed. Selvaraj et al. [6] evaluated the influence of cutting variables on the surface
quality of 304 ASS by applying the Taguchi technique in turning process in center
lathe. They employed ANOVA and S/N ratio analysis to determine the optimum
cutting variables, which improve the surface finish in turning process. Ciftci [7]
conducted turning experiments of two different alloys of ASSs using carbide tools.
The machining processes were carried out at 4 different values of cutting speeds,
constant value of depth and feed. The influences of coating of the tool inserts and
cutting speed on the surface finish were analyzed. They determined that surface
finish was increased with increasing cutting speed. The optimal cutting parameters
for turning process of 304 SS were investigated by Ihsan Korkut et al. [8]. Increase of
cutting velocity decreased the surface roughness while turning 304 SS. Xavior et al.
[9] examined the application of coolant oils and their influence on surface finish in
turning operation of ASS by employing carbide cutting tools. They concluded that
the application of coconut oil improved the surface finish of ASS. Kaladhar et al. [10]
carried turning tests to optimize surface finish of 304 SS on CNC turning center by
employing CVD-coated tool inserts. They found that cutting velocity, feed, depth,
and nose radius influencing the surface roughness by 46%, 17%, 13%, and 24%,
respectively. They found that the cutting velocity is the most important parameter
and the depth is the least important variable in controlling the surface roughness.
The Taguchi method is commonly employed for optimizing the machining opera-
tions [11]. Several researchers used the Taguchi techniques for optimizing the surface
finish of various materials in different cutting processes [11–15]. Dry turning process
is the method of turning without employing cutting fluids. Dry machining eliminates
coolant and disposal costs. It eliminates the harmful effects of cutting fluid on the
operator’s health. It improves the environmental conditions of the workplace which
leads to cleaner production and green manufacturing.
So, in the current work, the influences of turning process variables on surface
finish of 304 ASS during dry turning process in CNC lathe are analyzed by applying
the Taguchi technique.
Some Studies on Surface Roughness of AISI 304 Austenitic … 871
2 Experimental Details
The material used for the research work is 304 SS cylindrical rod of diameter 40 mm
and length 120 mm. Table 1 gives the composition of the 304 SS alloy. Batliboi
make CNC lathe was used for conducting the turning tests. The spindle speed range
available in the CNC lathe is from 40 to 4000 rpm and power rating is 16 KVA.
The specification of the carbide turning tool inserts used is WNMG 06T304 MP
TT8020. The tool holder specification is MWLNR 2020 K 06 W (ISCAR make).
Mitutoyo makes SJ-210 roughness tester which was used for measuring the surface
finish. Cutting fluid was not used during the turning tests (dry turning). Figures 1
and 2 show the experimental setup and surface finish measurement photographs.
In order to choose an appropriate orthogonal array (OA) for the tests, the number
of degree of freedom (DOF) for the experiments is to be calculated. A 3-level variable
has two DOF. The interaction effects are not considered in the current work. Hence,
the number of DOF for three machining variables is six. The selection of the OA is
based on the condition that the DOF for the OA must be greater than or equal to the
DOF of the machining parameters. In the current work, an L9 OA with 9 rows and
3 columns is employed. This OA has 8 DOF and can analyze 3 factors and 3-level
experiments. The DOF of OA is greater than the DOF of process variables. Hence,
Table 1 Chemical
Element wt%
composition of 304 SS
C 0.067
Si 0.64
Mn 1.60
S 0.023
P 0.025
Cr 19.12
Ni 9.06
Fe Balance
Fig. 1 CNC machine setup

Fig. 2 Photograph of the

surface roughness
measurement
Table 2 Levels of machining

1 Machining variables Level 1 Level 2 Level 3
variables used in turning
process Cutting speed, V (m/min) 120 160 200
Feed, F (mm/rev) 0.05 0.10 0.15
Depth, D (mm) 0.50 0.75 1.00
L9 OA is selected for the current work. Only nine tests are required to analyze the
entire machining variables using this array. Hence, experimental cost and time are
saved.
The input machining variables employed in this investigation are depth, feed, and
cutting speed. Three levels of cutting velocity (200, 160, and 120 m/min), feed rates
(0.15, 0.10, and 0.05 mm/rev), and depth of cuts (1.0, 0.75 and 0.5 mm) are used
for turning the AISI 304 SS cylindrical rods. Table 2 shows the levels of machining
variables employed in the dry turning operation. The plan of experiment using L9
OA is shown in Table 3.
The test results are examined using ANOVA and S/N ratio. Minitab software is
employed for the analysis of the results. The ratio of mean and standard deviation is
defined as S/N ratio (η). Equation (1) reported by Yang et al. [11] is used to calculate
η.
η = −10 log(M.S.D) (1)
In Eq. 1, M.S.D represents the mean square deviation.

Table 3 Plan of experiment

S. No. Levels of cutting parameters
using L9 orthogonal array
V F D
1 1 1 1
2 1 2 2
3 1 3 3
4 2 1 2
5 2 2 3
6 2 3 1
7 3 1 3
8 3 2 1
9 3 3 2
The quality category, lower-the-better should be used to attain the optimum surface
roughness. Equation (2) is used to calculate M.S.D. for surface roughness [11].
1 2
m
M.S.D = S (2)
m i=1 i
In Eq. 2, m indicates the number of experiments and S i indicates the Ra value of

ith experiment.
Table 4 gives the test results of Ra and their S/N ratio for AISI 304 ASS. The
mean S/N ratio of level 1 cutting velocity is the mean of S/N ratios of tests 1, 2, and
3. The mean of S/N ratios of tests 1, 4, and 7 gives the mean S/N ratio at level 1 feed.
The mean of S/N ratios of tests 1, 6, and 8 gives the mean S/N ratio at level 1 depth.
The mean S/N ratio for depth, feed, and cutting velocity at second and third levels
Table 4 Test results for Ra and S/N ratio

S. No. Levels of cutting parameters Ra η
V F D
1 1 1 1 0.80 1.9382
2 1 2 2 0.98 0.1754
3 1 3 3 1.26 −2.0074
4 2 1 2 0.65 3.7417
5 2 2 3 0.83 1.6184
6 2 3 1 0.78 2.1581
7 3 1 3 0.60 4.4369
8 3 2 1 0.55 5.1927
9 3 3 2 0.71 2.9748
Table 5 S/N response table for Ra

Machining parameter Cutting parameters level Max-Min
Level 1 Level 2 Level 3
V 0.0354 2.506067 4.201467 4.1660667
F 3.372267 2.328833 1.041833 2.3304333
D 3.096333 2.2973 1.3493 1.7470333
Main Effects Plot for SN ratios

Data Means
V F
4
3
2
Mean of SN ratios
0
120 160 200 0.05 0.10 0.15
D
4
3
2
1
0
0.50 0.75 1.00
Signal-to-noise: Smaller is better
Fig. 3 S/N response graph for Ra
is calculated in a similar method. The surface roughness S/N response table of AISI
304 ASS is given in Table 5.
Figure 3 illustrates the average S/N response graph for the surface finish of AISI
304 ASS. In Fig. 2, V 1 , V 2 , and V 3 represent the cutting velocity at level 1, 2, and 3,
respectively. F 1 , F 2 , and F 3 represent the feed at level 1, 2, and 3, respectively. D1 ,
D2 , and D3 represent the depth at level 1, 2, and 3, respectively. The larger S/N ratio
gives the minimum variance of the response about the desirable value. The larger
S/N ratio for surface finish of 304 ASS is achieved at level 3 cutting velocity, level
1 depth, and level 1 feed. Therefore, the optimal cutting variables for surface finish
of 304 ASS are 0.5 mm depth, 0.05 mm/rev feed, and 200 m/min cutting velocity.
ANOVA is used to find the turning process variables, which influence the output
parameter significantly. Equation (3) given by Yang et al. [11] is used to determine
total sum of squared deviations (SST ).

n
SST = (ηi − ηm )2 (3)
i=1
Table 6 Results of ANOVA for Ra

Machining DOF Sum of squares Mean square F-Ratio Percentage
variable influence
V 2 26.3338 13.1669 950.6787 67.30443
F 2 8.1758 4.0879 295.1552 20.89587
D 2 4.5891 2.29455 165.6715 11.72891
Error 2 0.0277 0.01385 0.070796
Total 8 39.1264 100
Here, n indicates a number of tests, ηi indicates an average S/N ratio for ith test,
and total average S/N ratio is indicated by ηm.
The SST consists of two sources. One is sum of the squared deviations (SSd ) cor-
responding to each input variable and the other one is sum of the squared deviations
due to error (SSe ). Table 6 gives the results of ANOVA for Ra of AISI 304 ASS.
From Table 6, it is revealed that the most influential variable for surface finish is
cutting velocity. The cutting variables influencing the surface finish are in the order
of cutting velocity followed by feed and then depth. The ANOVA table indicated that
depth, feed, and cutting velocity are influencing the surface finish of AISI 304 ASS
by approximately 12%, 21%, and 67%, respectively. It is revealed that the cutting
velocity is the most important variable and the depth of cut is the least dominant
variable in controlling the surface finish. The results are agreed with the results
reported by Kaladhar et al. [10].
Confirmation experiment is conducted at optimum cutting conditions (d =
0.5 mm, f = 0.05 mm/rev, and V = 200 m/min). The estimated S/N ratio η̂ is
employed to predict and check the output parameter at the optimum level. Value of
η̂ at the optimum level of the cutting variables can be determined by employing the
Eq. (4) [11].

o
η̂ = ηm + (η̄i − ηm ) (4)
i=1
Here, “o” is the number of cutting variables that influence the response.
The surface finish value at the optimum machining conditions was calculated by
using the Eqs. (1), (2), and (4). Predicted and experimental values of surface finish
Table 7 Confirmation test

Optimal machining variables
results for Ra
Prediction Experiment
Level V3 F1 D1 V3 F1 D1
Surface roughness 0.49 0.51
S/N ratio 6.1748 5.8486
values at the optimum cutting conditions are shown in Table 7. The test result is
closer to predicted value within an error of about 4%.
4 Conclusions
The Taguchi method was employed to find the optimum machining conditions of 304
stainless steel alloy in dry machining process. The analysis of S/N ratio and variance
were used to evaluate the machining performance. The outcomes of this work are
given as follows:
1. Higher value of cutting velocity (200 m/min), lower value of depth (0.5 mm),
and lower value of feed (0.05 mm/rev) are preferred to achieve the minimum
surface roughness.
2. From the ANOVA analysis, depth, feed, and cutting velocity were found to
influence the surface finish by around 12%, 21%, and 67%, respectively.
References
1. Davis JR (1995) ASM specialty handbook stainless steels. ASM International, Ohio
2. Radhakrishnan P (2002) Manufacturing Technology, Scitech publications (India) Pvt. Ltd.,
Chennai
3. Rao PN (2014) CAD/CAM principles and applications. McGraw Hill Education (India) Pvt.
Ltd, Delhi
4. Kulkarni AP, Joshi GG, Sangade VG (2013) Dry turning of AISI 304 austenitic stainless steel
using AlTiCrN coated insert produced by HPPMS technique. Procedia Eng 64:737–746
5. Asiltürk I, Neşeli S (2012) Multi response optimisation of CNC turning parameters via taguchi
method-based response surface analysis. Measurement 45:785–794
6. Philip Selvaraj D, Chandramohan P (2010) Optimization of surface roughness of AISI 304
austenitic stainless steel in dry turning operation using taguchi design method. J Eng Sci
Technol 5:293–301
7. Ciftci I (2006) Machining of austenitic stainless steels using CVD multilayer coated cemented
carbide tools. Tribol Int 39:565–569
8. Korkut I, Kasap M, Ciftci I, Seker U (2004) Determination of optimum cutting parameters
during machining of AISI 304 austenitic stainless steel. Mater Des 25:303–305
9. Xavior A, Adithan M (2009) Determining the influence of cutting fluids on tool wear and
surface roughness during turning of AISI 304 austenitic stainless steel. J Mater Process Technol
209:900–909
10. Kaladhar M, Venkata Subbaiah K, Srinivasa Rao C (2012) Parametric optimization during
machining of AISI 304 Austenitic Stainless Steel using CVD coated DURATOMICTM cutting
insert. Int J Ind Eng Comput 3:577–586
11. Yang WH, Tarng YS (1998) Design optimization of cutting parameters for turning operations
based on the taguchi method. J Mater Process Technol 84:122–129
12. Zhang JZ, Chen JC, Kirby ED (2007) Surface roughness optimization in an end-milling
operation using the Taguchi design method. J Mater Process Technol 184:233–239
13. Nalbant M, Gokkaya H, Sur G (2007) Application of taguchi method in the optimization of
cutting parameters for surface roughness in turning. Mater Des 28:1379–1385
14. Ghani JA, Chodhury IA, Hassan HH (2004) Application of taguchi method in the optimization
of end milling parameters. J Mater Process Technol 145:84–92
15. Philip Selvaraj D, Chandramohan P, Mohanraj M (2014) Optimization of surface roughness,
cutting force and tool wear of nitrogen alloyed duplex stainless steel in a dry turning process
using taguchi method. Measurement 49:205–215
Vibration and Control Engineering
Influence of Fiber Orientation
on Mechanical Properties and Free
Vibration Characteristics of Glass/Hemp
Hybrid Composite Laminates
R. Murugan, V. Muthukumar, K. Pradeepkumar, A. Muthukumaran,

and V. Rajesh
Abstract Fiber-reinforced polymer (FRP) composite materials made of synthetic

fibers are commonly used in aircraft and automotive structures due to their specific
property of high strength-to-weight ratio. The non-biodegradable nature of synthetic
fibers and increase in demand of environment-friendly materials now made natural
fibers as best alternative material to synthetic fibers. In this view, hybridization of
synthetic E-glass fiber with natural hemp fiber is proposed to promote the usage
of natural fibers in structural applications without much compromise in structural
integrity. In the present study, the influence of fiber orientation over effective stacking
sequence of E-glass/hemp hybrid composite laminates was investigated in terms of
mechanical and free vibration characteristics. The hybrid composite laminates are
fabricated by hand lay-up method. Two types of composite laminates were prepared
in effective stacking sequence of GHHG with two different orientations fibers such
as (0°/90°) and (45°/45°) for hemp fiber. The mechanical properties such as tensile
and flexural strength were evaluated according to ASTM standards. Free vibration
characteristics such as modal frequency and modal damping values of hybrid com-
posite laminates are analyzed by experimental modal analysis. The results of modal
frequency and mode shapes of hybrid composite beams were stated and discussed.
Keywords Natural fiber · Hybrid composites · Mechanical properties · Free

vibration characteristics · Fiber orientation
R. Murugan (B) · V. Muthukumar · K. Pradeepkumar · A. Muthukumaran · V. Rajesh

Department of Mechanical Engineering, Panimalar Institute of Technology, Chennai, India
V. Muthukumar
K. Pradeepkumar
A. Muthukumaran
V. Rajesh

https://doi.org/10.1007/978-981-15-4488-0_74
882 R. Murugan et al.
1 Introduction
Fiber-reinforced polymer (FRP) composite materials are used in a large number

of applications ranging from aerospace to automotive industries due to their cost-
effectiveness, specific properties, and ease of manufacturing. It is observed that the
structural members used in such dynamic applications experience shock and vibra-
tions. Hence, studying the vibration characteristics of FRP composite structures needs
much attention to meet the industrial requirements.
The characteristics of FRP composites depend upon many factors such as the
mechanical properties of the constituent materials, i.e. fiber and matrix, the vol-
umes of both fiber and matrix, the length and orientation of the fibers in the matrix.
Many researchers have contributed towards the effect of these factors on mechanical
properties of FRP composite laminates [1–5]. Ajith Gopinath et al. experimentally
investigated the effect of different matrices on mechanical properties of jute fiber-
reinforced composites using polyester and epoxy resin matrices. The authors revealed
that the tensile and flexural properties for jute/epoxy laminates are better than that
of the jute/polyester composites, which is preferable for automotive applications
[6]. Saira Taj et al. evaluated the mechanical properties of woven and nonwoven
jute fabric-reinforced poly-L-lactic acid-based composites. They found that woven
structure exhibited excellent mechanical behavior like tensile, flexural, and impact
loadings compared to nonwoven composite [7]. Handa et al. proved from their study
that hybrid composites reinforced with natural fibers and synthetic fibers demon-
strate good mechanical performance [8]. The comparison of mechanical properties
of natural and synthetic fiber-reinforced composite laminates was carried out by
Muthuraj et al. The authors revealed that among the preferred natural fibers like jute,
sisal, hemp, coir, banana, and palm, tensile strength and modulus of hemp/epoxy
fabric-reinforced composite is higher than the other fibers [9]. Claudio Scaponi et al.
made a comparative evaluation between E-glass and hemp fiber composites applied
in rotorcraft interiors. They replaced the existing steel electronic rack mounted on the
helicopter with eco-friendly hemp/epoxy composite material and thereby reduced the
weight of the structure considerably [10]. Girisha et al. investigated the effect of fiber
orientation on mechanical properties of jute/hemp-reinforced hybrid composites with
three different orientations 30%, 45%, and 90%. They found that orientation 90%
has higher mechanical properties when compared the to remaining two orientations
[11].
In the recent past, the influence of types of fiber and matrix material, fiber architec-
ture and stacking sequence on free vibration characteristics of natural and their hybrid
composites were studied extensively. Senthil Kumar et al. experimentally investi-
gated the effect of stacking sequence on vibration properties of coconut sheath/sisal
hybrid composites, and they found that higher natural frequency values occurred
when similar fibers are stacked together [12]. Akash et al. experimentally studied the
dynamic behavior of hybrid jute/sisal fiber-reinforced polyester composites using
free vibration characteristics and they showed that the changes in the fiber angle
yield to different dynamic behavior of the composite laminate [13]. Ömer Yavuz
Influence of Fiber Orientation on Mechanical Properties … 883
Bozkurt et al. investigated the effect of fiber orientations on damping and vibration
characteristics of basalt/epoxy composite laminates and they showed that damping
and vibration characteristics of the composite samples are strongly affected by the
fiber orientation. Also, they found that the increase in the angle of fiber orientation
resulted in a decrease in natural frequency, and the laminates having higher fiber
orientation angle had higher damping ratios [14]. Experimental investigation carried
out by J. Alexander et al. on free vibration and damping characteristics of GFRP
and BFRP laminated composites at various boundary conditions showed that fun-
damental natural frequency of woven fabric composites is better than unidirectional
composites due to its high stiffness [15]. Based on the literature review, the present
study aims toward understanding and investigating the effect of fiber orientation on
the vibration characteristics of thin-walled woven fabric hybrid composite beams
made of glass and hemp fibers for promoting good dynamic stability in structural
applications.
2 Fabrication of Glass/Hemp Hybrid Composites

Laminates
Both glass and hemp fibers are preferred in plain woven fabric form with 400 gsm
for reinforcement. Thermoset epoxy resin of grade LY556 with hardener HY951 is
used as matrix material. All glass/hemp hybrid composite laminates are fabricated by
conventional hand lay-up method with uniform fiber volume fraction of vf = 0.3. The
effective layering arrangement of four-layered glass/hemp hybrid laminate, GHHG,
is selected from literature. In an earlier investigation, Murugan et al. established the
effective stacking sequence for four-layered glass/carbon hybrid laminates based on
tensile, flexural properties, and free vibration characteristics. From the experimental
results, they concluded that the increased tensile and flexural modulus values were
obtained for the hybrid laminate fabricated with high modulus fabric in the external
layer and low modulus fabric as the core layer [16, 17]. In order to study the effect
of fiber orientation over the effective stacking sequence, in the present study, two
different fiber orientations such as (0° /90° ) and (45° /45° ) are considered. Figure 1
shows the image of the glass/hemp hybrid samples fabricated by hand lay-up tech-
nique. Table 1 shows the layering arrangement, size, and density of the glass/hemp
hybrid laminates fabricated by hand lay-up technique.
Fig. 1 Photograph showing the two types of glass\hemp hybrid laminates fabricated using hand
lay-up technique
Table 1 Layering
Specimen Layering arrangement Size (mm)
arrangement, size, and
density of the glass/hemp H1 G 0-90
250 × 250 × 2.5
hybrid laminates fabricated H 0-45
by hand lay-up technique H 0-45
G 0-90
H2 G 0-90 250 × 250 × 2.5

H 45-45
H 45-45
G 0-90
3 Testing of Glass/Hemp Hybrid Composites Laminates
3.1 Tensile Test
Tensile test was performed for the glass/hemp hybrid composite laminates as per
ASTM D3039 standard. The size of the specimen for the tensile test is 250 mm ×
25 mm × t mm where t represents the thickness of the specimen. Figure 2 shows
the image of tensile testing of glass/hemp hybrid laminates carried out in INSTRON
4204 machine. Hybrid composite specimen having a thin rectangular cross section
is mounted in the grips of a universal testing machine and monotonically loaded in
tension, and the applied load and deformation of the specimen were recorded.
Fig. 2 Image showing the tensile test conducted on glass/hemp hybrid composite laminates
using INSTRON machine
3.2 Flexural Test
Flexural properties of FRP materials were evaluated as per ASTM D790 standard
test method. The standard dimensions of the specimen for flexural test are 126 mm ×
12.5 mm × t mm where t represents the thickness of the specimen. The composite
specimens are properly cut to the standard dimensions. The flexural test was con-
ducted on the same INSTRON machine with feed rate of 1.2 mm/min. The hybrid
specimen is placed on two knife edges provided on the testing machine with equal
over-hanging on both sides. A transverse load is applied at the center point of the
specimen. Figure 3 shows the standard size of the composite specimen used for the
flexural test.
Fig. 3 Schematic diagram

showing the standard size
of flexural test specimen as
per ASTM D790 standard
(all dimensions are in mm)
3.3 Experimental Modal Analysis
Free vibration characteristics such as modal frequency and modal shapes are evalu-
ated for glass/hemp hybrid composite laminates under fixed-free end condition. The
preferred glass/hemp hybrid laminates are cut into the standard size of 250 mm ×
25 mm × t mm where t represents the thickness of the specimen for free vibration
study. For the effective transformation of excitation force into the composite beam
and to achieve fixed-free boundary condition, a fixture is used to clamp the composite
specimen. After fixing the specimen, the span length of the cantilever specimen is
kept as 200 mm as shown in Fig. 4.
For obtaining the mode shapes of the glass/hemp hybrid composite beams, rov-
ing hammer method is preferred. The composite beam is excited by an impact
hammer (9722A500) at specified points of equal interval. A triaxial accelerome-
ter (8766A500) is used to capture the vibration response of the composite beam
specimen due to the given excitation. The input excitation force and correspond-
ing output vibration response of the specimen are fed into data acquisition system
(ATA9234) as shown in Fig. 4. And these time domain signals are processed over the
fast Fourier transform (FFT) software (Dewesoft) to acquire the required frequency
response function (FRF) plots. In a similar manner, FRF plots for each excitation
point of the composite beam and thus cumulative FRF plots are obtained to evaluate
the mode shapes of the two hybrid composite beams.
Fig. 4 Image showing free vibration test setup: (1) fixture, (2) composite specimen, (3) impact
hammer, (4) accelerometer, (5) data acquisition card, and (6) computer with FFT software showing
FRF
Table 2 Tensile and flexural

Specimen Tensile strength (MPa) Flexural strength (MPa)
properties of glass/hemp
hybrid composites laminates H1 49.86 0.436
considered H2 33.19 0.434
Fig. 5 Comparison of 3.5

tensile stress/strain graphs of
3 H1
various glass/hemp hybrid
2.5 H2
Stress (MPa)
laminates
2
1.5
1
0.5
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Strain (%)
4.1 Mechanical Properties of Glass/Hemp Hybrid Laminates
The experimentally evaluated mechanical properties of preferred glass/hemp hybrid

laminates are reported in Table 2. Figure 5 shows the comparison of stress/strain
graphs of hybrid composite laminates obtained by tensile test. Table 2 showed that
there is a considerable tensile strength variation between the hybrid laminates con-
sidered due to the change in fiber orientation. From the comparison of stress/strain
graphs shown in Fig. 5, it is revealed that H1 laminate has higher tensile strength
than H2 laminate. From Table 2, it is revealed that there is minimal variation in the
flexural strength among the two hybrid laminates. Flexural strength of composite
laminates mainly governed by the stacking sequence of the fiber lamina and the fiber
orientation has the least control over it.
4.2 Free Vibration Characteristics of Glass/Hemp Hybrid

Composite Beams
Figure 6 shows the comparison of FRF plots of glass/hemp hybrid beams with two
different fiber orientations. The positive shift in FRF plot of H1 beam compared to
H2 beam confirms the higher stiffness of H1 beam than H2 beam.
The modal frequency values for the first three mode shapes of two hybrid compos-
ite beams arrived from experimentally recorded FRF function are shown in Table 3.
From Table 3, it is observed that the resonant frequency set attained for H1 hybrid
Fig. 6 Comparison of FRF 7

plots of preferred glass/hemp 6 H1
hybrid beams obtained from
Amplitude (g/N)
H2
free vibration test under 5
fixed-free boundary 4
condition
3
0
0 100 200 300 400 500 600
Frequency (Hz)
Table 3 Experimental modal frequency values of hybrid composite beams for three successive
modes
Mode no. Mode shape Modal frequency (Hz)
H1 H2
Mode I 21 20
Mode II 158 148
Mode III 457 447
beam is slightly higher side than H2 beam in all three modes. This characteristic
behavior confirms that the change in the fiber orientation of inner hemp layers in the
hybrid beam influences the stiffness of hybrid composite beam under flexural vibra-
tion. The minimal variation in flexural strength of H1 than H2 layering arrangement
caused this modification.
5 Conclusions
In the present work, the effect of fiber orientation on mechanical properties and free
vibration characteristics of glass/hemp hybrid composite beams was experimentally
investigated. The effective layering arrangement of glass/hemp hybrid laminates,
i.e. GHHG, is preferred based on the earlier investigation. In order to study the
effect of fiber orientation, two different laminates H1 with fiber orientations (0° /90° )
and H2 with fiber orientation (45° /45° ) are fabricated. Hybrid laminate H1 exhibited
increased tensile strength than the other hybrid arrangements, but the flexural strength
of hybrid laminates showed the least variation.
Experimental modal analysis of glass/hemp hybrid beams, H1 and H2, showed

minimal variation in resonant frequency set in successive modes for different fiber
orientation. The layer arrangement of hybrid beam, H1, showed increased modal
frequency values than the other laminate, H2, which is beneficial for structural
applications.
References
1. Sanjay MR, Arpitha GR, Yogesha B (2015) Study on mechanical properties of natural—glass
fiber reinforced polymer hybrid composites: a review. Mater Today Proc 2:2959–2967
2. Satish P, Kesavan R, Vijaya Ramnath B (2015) Effect of fiber orientation and stacking sequence
on mechanical and thermal characteristics of banana-kenaf hybrid epoxy composite. Silicon
9(4):577–585
3. Bhoopathi R, Rajesh M, Deepa C (2014) Fabrication and property evaluation of banana-hemp-
glass fiber reinforced composites. Procedia Eng 97:2032–2041
4. Kumar A, Singh S (2015) Analysis of mechanical properties and cost of glass/jute fiber
reinforced hybrid polyester composites. Proc Inst Mech Eng Part L: J Mat: Des Appl
229(3):202–208
5. Hareesh M, Sunil M, Naveen Kumar HN, Rohith Kumar K (2016) Mechanical characterization
of hybrid hemp/glass fiber reinforced epoxy composite. Int J Eng Res Rev 4(1):104–108
6. Gopinath A, Senthil Kumar M, Elayaperumal A (2014) Experimental investigations on mechan-
ical properties of jute fiber reinforced composites with polyester and epoxy resin matrices.
Procedia Eng 97:2052–2063
7. Taj S, Munawar MA, Khan S (2007) Natural fiber-reinforced polymer composites. Proc Pak
Acad Sci 44(2):129–144
8. Abdul Khalil HPS, Handa S, Kang CW, Nik Faud NA (2007) Agro-hybrid composite: The
effects on mechanical and physical properties of oil palm fiber (EFB)/glass hybrid reinforced
polyester composites. J Reinf Plast Comp 26:203–218
9. Muthuraj, Ravi Kumar M, Keerthiprasad KS (2016) Characterization and comparison of natural
and synthetic fiber composite laminates. Int J Eng Tech 2(5):1–8
10. Scarponi C, Messano M (2015) Comparative ealuation between E-glass and hemp fiber
composites application in rotorcraft interiors. Composites Part B 69:542–549
11. Girisha KG, Anil KC, Akash (2014) Mechanical properties of jute and hemp reinforced hybrid
composites. Int J Res Eng Techn 2(4):245–248
12. Senthil Kumar K, Siva I, Rajini N, Jeyaraj P, Winowlin Jappes JT (2014) Tensile, impact, and
vibration properties of coconut sheath/sisal hybrid composites: Effect of stacking sequence. J
Reinf Plast Compos 0(0):1–11
13. Akash DA, Thyagaraj NR, Sudev LJ (2013) Experimental study of dynamic behaviour of hybrid
jute/sisal fiber reinforced polyester composites. Int J Sci Eng Appl 2(8):170–172
14. Bozkurt ÖY, Bulut M, Özbek Ö (2016) Effect of fiber orientations on damping and vibration
characteristics of basalt epoxy composite laminates, Proceedings of the world congress on
civil, structural, and environmental engineering (CSEE’16) Prague, Czech Republic, ICSENM
pp. 115–116
15. Alexander J, Augustine BSM (2015) Free vibration and damping characteristics of GFRP and
BFRP laminated composites at various boundary conditions. Indian J Sci Technol 8(12):1–7
16. Murugan R, Ramesh R, Padmanabhan K (2014) Investigation on static and dynamic mechanical
properties of epoxy based woven fabric glass/carbon hybrid composite laminates. Procedia Eng
97:459–468
17. Murugan R, Ramesh R, Padmanabhan K (2016) Investigation of the mechanical behavior and
vibration characteristics of thin walled glass/carbon hybrid composite beams under a fixed-free
boundary condition. Mech Adv Mater Struct 23(8):909–916
Free Vibration Analysis of Functionally
Graded Beam with Linearly Varying
Thickness
Rajat Jain and Mihir Chandra Manna
Abstract The present study investigates the free vibration analysis of functionally
graded material (FGM) beam which is rectangular in cross section with linearly
varying thickness along its axis with the help of finite element formulation. This
formulation of finite element is developed based on the Timoshenko beam theory
which we called as first-order shear deformation theory. In the present analysis of the
beam with linearly varying thickness, the beam element has five nodes and thirteen
degrees of freedom. Properties of the material used in this beam element are varying
continuously according to the power law along the thickness direction. Governing
equations used for the formulation of present work are derived from Lagrange’s
equations. The natural frequency of beam is calculated using different boundary
conditions, exponents of power law, depth to span ratios, and tapered ratios. The
present beam element is accurately demonstrated by comparing the results with the
available data of publications for constant thickness and for variable thickness, and
some results are new and can be further considered for future researches.
Keywords Functionally graded material · Finite element method
1 Introduction
Functionally graded materials (FGMs) have superior composites of continuous spa-

tial variables formed with the help of two or more material constituents. These materi-
als are used to build a mixture of two or more constituents whose particles have nearly
similar form and dimensions (ceramic powder, plasma particles, etc.). These materi-
als have excellent properties such as high temperature, creep, and fatigue resistance.
Therefore, these materials are extensively used in many engineering applications,
i.e., aerospace, civil, and mechanical engineering, etc. The increasing demand for
these materials is because of their good mechanical behavior in various applications.
R. Jain (B) · M. C. Manna

Indian Institute of Engineering Science and Technology Shibpur, Howrah, West Bengal 711103,
India

https://doi.org/10.1007/978-981-15-4488-0_75
892 R. Jain and M. C. Manna
When we compare the researches of plates and shells, the work on FGM beams
are very less and very few data are available. Free vibration problems of functionally
graded beams (FGB) were solved by different analytical and numerical approaches
using different beam theories. Among many analytical works, some may be men-
tioned in this context. In the last of the twentieth century, many researchers have
done their works on free vibration analysis of FGM beams with different boundary
conditions and loading conditions [1–5]. In the area of sandwiched beams, Frostig
et al. [6] investigated the higher-order theory for sandwiched beams behavior with
transverse core and they investigated that how the shear is transferred from the core
to the skin. Later on, Sankar et al. [7] obtained the solution for elasticity problem
using FG beams with simply supported conditions exposed to transverse loading of
sinusoidal behavior and the modulus of elasticity varies along the thickness in the
exponential manner, and Sankar found when the loaded side of the beam is softer
as compared to homogeneous beam, the concentration of stresses are less, and vice
versa when the harder side is loaded.
In the field of thermal stress, Xiang and Yang [8] examined the frequencies of
thermally pre-stressed, laminated FG beam with variable thickness using the first-
order beam theory and the method of differential quadrature. This analysis shows
that if the layers of FGM are thicker with a lesser fractional index of volume in
the laminated beam, structure shows the effective increases in natural frequencies
and decrease in the amplitude of vibration. And after the analysis of Xiang and
Yang, characteristics of free vibration and the behavior of dynamic analysis of a FG
beam with simply supported ends under a concentrated moving harmonic load are
studied by Simsek and Kocaturk [9]. In this study, Lagrange’s equations are used
for deriving the equation of motion under the assumptions of beam theory given
by Euler–Bernoulli. This analysis says that the different material distribution effect,
moving harmonic load velocity, and the dynamic responses of excitation frequency on
the FG beam show very vital role in the dynamic response of the FG beam. Later on,
Li et al. [10] investigated the FG beams by taking into account the effects of shear
deformation and deriving a single governing equation for the static and dynamic
behaviors of FG beams. The results obtained from the analysis are coinciding with
the standard elasticity solutions. In the analysis of mode shape and fundamental
frequencies of FGM beams, Alshorbagy et al. [11] studied the free vibration features
and dynamic response of a functionally graded beam for various distributions of
material using FEM. The obtained results have shown that the distributed variation
of material along the axial direction and the ratio of slenderness have not executed
any effect on the fundamental frequencies or mode shapes but as the power exponent
increases the natural frequencies also increases. Alshorbagy et al. [12] also examined
the dynamic behavior of FGM thick beam by analyzing the effect of temperature.
In the area of isotropic and functionally graded (FG) sandwich beams, Nguyen
et al. [13] proposed a new higher-order shear deformation theory for buckling and
free vibration analysis. This study shows a new hyperbolic variation of shear stress
in transverse direction, and the results obtained from this theory show outstanding
promise with those derived from former studies. After that, Kahya et al. [14] studied
multi-layered shear deformable beam element for dynamic analysis of laminated
Free Vibration Analysis of Functionally Graded Beam … 893
composite beams subjected to moving load. In this investigation, results show very
significant responsive evaluation of the stacking lamina of the laminated beams. After
two years, Kahya et al. [15] also examined one more result on the model of finite
element for vibration and buckling analysis of FGM beam based on first-order shear
deformation theory using a five-noded beam element with ten degrees of freedom.
Recently, Banerjee et al. [16] investigated the free vibration of FGBs by apply-
ing the dynamic stiffness method in this analysis properties of the material alter
continuously over the thickness according to the variation of power law. After that,
Armagan et al. [17] analyzed free vibration behavior of two-directional FG beams
subjected to different boundary conditions by using the shear deformation theory of
third order where the properties of materials of the beam vary exponentially in length
as well as thickness directions. In the area of nonlinear free and forced vibrations,
Sinir and Gultekin [18] have done his analysis on non-uniform cross-sectional beam
of axially functionally graded material with Euler–Bernoulli theory. The beam has
immovable boundary conditions, which leads to mid-plane stretching because of
vibrations and the frequency-response curves that show the effect of these nonlinear
correction terms on natural frequency by the unstable regions. And very recently,
Chen et al. [19] investigated the vibration problem of axially functionally FGM beam
and parabolically varying thickness in 3D by isogeometric analysis in conjunction
with 3D theory.
From the available works of literature, it has been seen that most of the works have
been done using different beam elements for the fundamental frequency only. In the
present work, a beam element with five nodes having thirteen degrees of freedom
is used for the analysis of free vibration of the functionally graded beams with a
variable thickness along its length. In this work, the first-order shear deformation
theory has been taken into account for finite element formulation.
2 Theory and Formulation
The purpose of this study is to develop an exact finite element model with the help
of shear deformation theory of first order for free vibration analysis of functionally
graded beams (FGB) with variable thickness. The properties of materials in the beam
vary continuously through the direction of thickness according to the formulation of
power law.
The beam proposed here is an isotropic, non-homogeneous elastic beam having its
length l and cross section is b × h g which is rectangular in shape where the thickness
of the beam is varying linearly along its length as shown in Fig. 1 and h g is the
thickness at Gauss’s point. The beam is constituted with a mixture of two materials
Fig. 1 Variable thickness cross section of beam element
such as ceramic and metal, the position of these materials is at its top and bottom
surfaces, respectively. Hooke’s law is obeyed by the material. Power law rule governs
the variation of material properties along with the thickness as follows.
k
z 1
P(z) = (Pc − Pm ) + + Pm
hg 2
where k is the non-negative exponent of the power law, Pm and Pc are the equivalent
properties of the metal and ceramic ingredients, e.g., Young’s modulus E, Poisson
Ratio ν, and mass density ρ, respectively.
2.2 Finite Element Formulation
Figure 2 shows the five-node beam element with thirteen degrees of freedom with
each node having three degrees of freedom except the mid node. And only the mid
node has one degree of freedom.
Displacement field equation according to the first-order shear deformation theory
is as follows:
U (x, z, t) = u(x, t) − zφ(x, t), W (x, t) = w(x, t) (1)
where u is axial displacement, w is transverse displacement, and φ is the total bending

rotation of the cross section at any point on the neutral axis.
From Eq. (1), the relationship of strain–displacement is as follows:
∂U du dφ ∂U ∂W ∂w
εx x = = −z and γx z = + = −φ (2)
∂x dx dx ∂x ∂x ∂x
Fig. 2 Beam element with five nodes and thirteen degrees of freedom
where εx x and γx z are the normal and shear strains, respectively, and Eq. (2) can be
rewritten as:
⎧ du
⎪
⎪
⎨ dx
∂w
εX X du
−z ddxf 1 0 −z 0
= dx
∂W = ∂x
YX Z −φ 0 1 0 −1 ⎪
⎪
dφ
∂x ⎩ dx
φ
where

du ∂w dφ T
φ
dx ∂ x dx
⎡ ⎤
N1u,x 0 0 N2u,x 0 0 0 N4u,x 0 0 N5u,x 0 0
⎢ ⎥
⎢ 0 N1w,x 0 0 N2w,x 0 N3w,x 0 N4w,x 0 0 N5w,x 0 ⎥
=⎢ ⎥{δ}
⎣ 0 0 N1φ,x 0 0 N2φ,x 0 0 0 N4φ,x 0 0 N5φ,x ⎦
0 0 −N1φ 0 0 −N2φ 0 0 0 −N4φ 0 0 −N5φ
= [B]{δ} (3)
where {δ}T = {u 1 w1 φ1 u 2 w2 φ2 w3 u 4 w4 φ4 u 5 w5 φ5 }
and “x” represents the derivative with respect to x.
The shape functions in [B] are as follows:

−9 1 1
N1u = × ξ+ ξ− (ξ − 1)
16 3 3

27 1
N2u = × (ξ + 1) ξ − (ξ − 1)
16 3

−27 1
N4u = × ξ+ (ξ − 1)(ξ + 1)
16 3

9 1 1
N5u = × ξ− ξ− (ξ + 1)
16 3 3

9 1 1
N1w = × ξ+ (ξ ) ξ − (ξ − 1)
16 3 3

−81 1
N2w = × (ξ − 1)(ξ ) ξ − (ξ + 1)
16 3

1 1
N3w =9× ξ + (ξ − 1) ξ − (ξ + 1)
3 3

−81 1 1
N4w = × ξ− ξ+ (ξ )(ξ + 1)
16 3 3

9 1 1
N5w = × ξ+ (ξ ) ξ − (ξ + 1)
16 3 3

−9 1 1
N1φ = × ξ+ ξ− (ξ − 1)
16 3 3

27 1
N2φ = × (ξ + 1) ξ − (ξ − 1)
16 3

−27 1
N4φ = × ξ+ (ξ − 1)(ξ + 1)
16 3

9 1 1
N5φ = × ξ− ξ− (ξ + 1) (4)
16 3 3
where ξ = Lx and 2L is the element length.

The thickness of the beam varying linearly from one end as shown in Fig. 1, the
variation of thickness at any distance “x” from one end (x = 0) is given by
t1 − t0
t x = t0 + (l − x) (5)
l

Or,tx = t0 1 + 1 − xl δ , where δ = t1 −tt0
0
(tapered ratio), t0 is the thickness at
one end (x = l) and t1 at the other end (x = 0), and x is measured from the end
where the thickness is t1 .
The element stiffness matrix can be written using the principle of virtual work as
follows:
l
[K ]e = ∫ [B]T [D][B]dx (6)
−l
]
where [B] = 1J d[N
dξ
or dN
dx
i
= dN
dξ
i
× dx
dξ
.
Again, on considering the beam element used in the present analysis we get,
⎡ ⎤
x1
∂x ∂
J= = {N1x N3x N5x } ⎣ x3 ⎦ (7)
∂ξ ∂ξ
x5

where N1x = 1/2 −ξ + ξ 2 ), N3x = (1 − ξ 2 ), and N5x = 1/2 ξ + ξ 2 . Total length
of the beam element is 2L. The abscissa of the nodes of the beam elements x1 , x3 ,
and x5 for geometric interpolation is as follows:
x1 = 0, x3 = x1 + L and x5 = x1 + 2L
On putting the above values of x1 , x3 , and x5 in Eq. (7) we get J as follows

⎧ ⎫
∂x ∂ ⎨ x1 ⎬
J= = {N1x N3x N5x } x3
∂ξ ∂ξ ⎩ ⎭
x5
⎧ ⎫
⎨ x1 ⎬
1 1
= (−1 + 2 ξ ) − 2ξ (−1 + 2ξ ) x +L
2 2 ⎩ 1 ⎭
x1 + 2L
= L or dξ/dx = 1/L
Total length of the beam element is 2L. Hence,x5 − x1 = 2L or
x5 − x1
L= (8)
2
Therefore, using Eqs. (6), (7), and (8), the elemental stiffness matrix can be given
as
1
[K ]e = ∫ [B]T [D][B] Ldξ (9)
−1
The rigidity matrix [D] is given as:

⎡ ⎤
A0 0 −A1 0
⎢ 0 B0 0 −B0 ⎥
[D] = ⎢
⎣ −A1 0 A2 0 ⎦
⎥ (10)
0 −B0 0 B0
where A0 , A1 , A2 , and B0 can be expressed as follows:

[A0 , A1 , A2 ] = ∫ E(z) 1, z, z 2 dA

= ∫ E(z) 1, z, z 2 zdz and B0

= k G(z)dA = k G(z) zdz
It has been assumed that mechanical property of Young’s modulus is varying

along with the thickness of the beam element. This Young’s modulus variation along
the thickness of the beam is governed by the power law which is as follows:
k
z 1 E(z)
E(z) = (E c − E m ) + + E m and G(z) = (11)
h 2 2(1 + ν)
where ν is taken as constant for the functionally graded material.

Therefore, putting the above relationship in the expression [A0 , A1 , A2 ] and B0
and taking the integration over the whole thickness of the beam, A0 , A1 , A2 , and B0
are obtained as:

(E c − E m ) bh 2g (E c − E m )k
A0 = bh g + E m ], A1 = ,
k+1 2 (k + 1)(k + 2)

bh 3g (E c − E m ) k 2 + k + 2 Em K
A2 = + , and B0 = A0
4 (k + 3)(k + 1)(k + 2) 3 2(1 + μ)
where h g is the total thickness of the beam at Gauss points of integration. Similarly,
the element mass matrix can be written as
l
T
[M] = e
B [ρ] B Ldξ (12)
−l

B
⎡ ⎤
N1u,x 0 0 N2u,x 0 0 0 N4u,x 0 0 N5u,x 0 0
⎢ ⎥
=⎣ 0 N1w,x 0 0 N2w,x 0 N3w,x 0 N4w,x 0 0 N5w,x 0 ⎦
0 0 N1φ,x 0 0 N2φ,x 0 0 0 N4φ,x 0 0 N5φ,x
⎡ ⎤
I0 0 −I1
⎢ ⎥
and, [ρ̄] = ⎣ 0 I0 0 ⎦
−I1 0 I2
where I0 ,I1 , and I2 can be expressed as follows:

[I0 , I1 , I2 ] = ρ(z) 1, z, z 2 dA = ρ(z) 1, z, z 2 zdz (13)
It has been assumed that densities are varying along with the thickness in the
present beam element. This variation of density along the beam thickness is governed
by power law which is as follows:
k
z 1
ρ(z) = (ρc − ρm ) + + ρm (14)
hg 2
Combining Eqs. (12), (13), and (14) and taking integral over the thickness of the
beam I0 , I1 , and I2 are expressed as follows:

(ρc − ρm )
I0 = bh g + ρm ,
k+1

(ρc − ρm )k
I1 = bh g
2
,
2(k + 1)(k + 2)

bh 3g (ρc − ρm ) k 2 + k + 2 ρm
I2 = + (15)
4 (k + 1)(k + 2)(k + 3) 3
Gauss quadrature method is used to find the element stiffness and mass matrices
from Eqs. (9) and (12) numerically, and the Gauss quadrature order used is four.
By assembling the element stiffness and mass matrices, following Eigenvalue
equation is obtained:

[K ]g − ω2 [M]g { } = {0} (16)
where [K ]g and [M]g are the global stiffness and global mass matrices, ω is the
natural frequency, and { } is the corresponding mode shape. Equation (16) is solved
by using the simultaneous iteration technique by Corr and Jennings [20] to obtain
the natural frequencies of the beam.
In the present analysis, formulation of finite element model is used based on the
shear deformation theory of the first order for the study of free vibration analysis of
functionally graded beams with linearly varying thickness along its length.
The study of convergent is approved forthe present beam element. Table 1 shows
the normalized fundamental frequencies ωn = h 0 Eρmm of FGBs with various
ωn l 2
boundary conditions for their different values, where the length of beam is l and the
thickness is h 0 at x = l. For the comparison of present work with those of Kahya
[15] and Nguyen [13], the material used in this FG beam is aluminum (Al) as metal
and alumina (Al2 O3 ) as ceramic for which Em = 70 GPa, ρm = 2702 kg/m3 ,
νm = 0.3, Ec = 380 GPa, ρc = 3960 kg/m3 , and νc = 0.3. Boundary condi-
tions for this analysis are assumed to be clamped-clamped (C–C), hinged-hinged
(H–H), and clamped-free (C–F). For calculation, the shear-correction factor is taken
as K = 5(1 + ν)/(6 + 5ν) from the work of Kahya [15] where υ is Poisson’s ratio.
In all of the following calculations, rectangular cross-sectional beam having different
length-to-thickness ratio (l/ h 0 ) ranging from 5 to 100 and different values of power
law exponent (k) varying from 0 to 10 has been considered. The normalized funda-
mental frequencies for different values of power law exponent, boundary conditions,
Table 1 Convergence study of normalized fundamental frequency for FGBs with boundary
condition l/ h 0 = 5 and k = 0
δ Number of elements C–C H–H C–F
0 4 9.9975535 5.1524837 1.8944103
8 9.9975020 5.1524793 1.8944101
12 9.9975012 5.1524783 1.8944100
15 9.9975012 5.1524783 1.8944100
0.5 4 11.3608624 6.1601690 2.8993748
8 11.3608125 6.1601620 2.8993735
12 11.3608117 6.1601619 2.8993735
15 11.3608117 6.1601619 2.8993735
1.0 4 12.3353769 6.9591461 3.8705373
8 12.3351511 6.9591034 3.8705336
12 12.3351434 6.9591027 3.8705335
15 12.3351432 6.9591027 3.8705335
and different tapered ratios (δ = 0, 0.25, 0.50, 0.75, 1.0) obtained from the present
analysis are shown in Table 3. For the value of δ = 0, it has been experiential that,
the obtained results are very close 0.8–1.2% less than the available published results
by Simsek [4], Nguyen [13] and Kahya [15] as shown in Table 2 and the results are
shown graphically in Fig. 3. From Fig. 3, it is observed that the results obtained from
the present analysis are very close to the results obtained by Kahya [15] because I
have taken 15 elements for my present analysis. From the results, it also observed
that only 15 elements (shown in Tables as P15) are sufficient for the desired accu-
racy of the obtained results compared to the other published results by Kahya [15]
and Nguyan [13]. In Table 2, the fundamental natural frequencies are presented with
available results of Simsek [4], Nguyen [13], and Kahya [15] (for δ = 0) which gives
very accurate results. And all the results with different tapered ratio except δ = 0 are
presented in Table 3 as new results and may be used for future reference for research
work in this field. In Fig. 4, it has been observed that the fundamental natural fre-
quencies of the beam increase with increasing the tapered ratio for different values of
(l/ h 0 ), keeping power law exponent (k) as constant. Again, it has been observed that
keeping (l/ h 0 ) and δ constant, the normalized natural frequencies decrease when the
power law exponent (k) increases as shown in Fig. 3. It may be further concluded that
in all cases, the normalized natural frequencies are higher for C–C beams than those
for C–F and H–H beams as shown in Fig. 5. It has been observed form Fig. 6 that
the effect of the taper ratios on the first, third, and sixth normalized frequencies are
not significant. But second, fourth, and fifth normalized frequencies are increasing
with the increase in taper ratios.
Different boundary conditions are defined as follows:
For C–C: At x = 0, u = w = φ = 0; Atx = 2l, u = w = φ = 0
For H–H: At x = 0, w = 0; At x = 2l, w = 0 and
Table 2 Comparison of normalized fundamental frequencies with different boundary conditions

for δ = 0 and l/ h 0 = 5
Conditions Element k =0 k = 0.5 k =1 k =5 k = 10
C–C P 15 9.99750 8.74257 7.89974 6.64281 6.31479
Kahya [15] 10.08647 8.75479 7.98414 6.71481 6.37413
Simsek [11] 10.0705 8.7463 7.9503 6.4934 6.1651
Nguyen [13] 10.0726 8.74674 7.9518 6.4929 6.1658
H–H P 15 5.15247 4.40816 3.97084 3.40227 3.29606
Kahya [15] 5.22193 4.46926 4.04967 3.48818 3.36434
Simsek [11] 5.1527 4.4102 3.9904 3.4012 3.2816
Nguyen [13] 5.1528 4.41108 3.9904 3.4009 3.2815
C–F P 15 1.89441 1.61867 1.46275 1.26419 1.22369
Kahya [15] 1.90722 1.62865 1.47294 1.27515 1.26363
Simsek [11] 1.8952 1.6182 1.4632 1.2591 1.2183
Nguyen [13] 1.8957 1.61817 1.4636 1.2594 1.2187
Fig. 3 Plot between 12

normalized fundamental
Normalized Fundamental Frequency
frequency v/s power law

exponent for the present 10
element with an earlier
research paper for L/h = 5, δ
8
= 0 and C–C boundary
condition
6
P15
4
Kahya
2 Simsek
Nguyen
0
k=0 k=1 k=5 k=10
Power Law Exponent
For C–F: At x = 0, u = w = φ = 0.
4 Conclusion
In the present analysis, a five-node beam element with thirteen degrees of freedom
is used to study the free vibration analysis of beam made of functionally graded
Table 3 Normalized fundamental frequencies of FGBs with various boundary conditions and
tapered ratio having different values of k at l/ h 0 = 5
Conditions k δ =0 δ = 0.25 δ = 0.50 δ = 0.75 δ = 1.0
C–C 1 7.89974 8.52200 9.04984 9.50214 9.89262
5 6.64281 7.12962 7.53598 7.87901 8.17102
10 6.31489 6.75639 7.12141 7.42668 7.68429
H–H 1 3.97084 4.38696 4.75685 5.08827 5.58646
5 3.40227 3.75003 4.05684 4.32962 4.57309
10 3.29606 3.62996 3.92342 4.18324 4.41408
C–F 1 1.46275 1.85471 2.24504 2.63015 3.00721
5 1.26419 1.59915 1.93055 2.25519 2.57063
10 1.22369 1.54622 1.86435 2.17494 2.47560

normalized fundamental
Fundamental Natural Frequency
frequency v/s tapered ratio 14

for k = 1, C–C boundary
condition and different 12
values of L/h 10
6 L/h=5
4 L/h=10
2 L/h=100
0
=0.00 =0.25 =0.50 =0.75 =1.00
Tapered RaƟo
material with linearly varying thickness under different boundary conditions for
different values of power law exponent (k) and different length to thickness ratios.
From the present analysis, it is observed that the performance of the present element
is excellent and this element can be utilized for the analysis of critical buckling of
FG beams as well as composite beams or functionally graded composite beams. It
is also observed that maximum frequencies are obtained for C–C beams as expected
than the others. It may also be seen that the power law exponent, length-to-thickness
ratio, and the tapered ratio have a significant effect on fundamental frequencies.

normalized fundamental C-C
Normalized Fundamantal Frequency

frequency v/s tapered ratio H-H
for L/h = 5, K = 1 and 10
C-F
different boundary
conditions
8
0
=0.00 =0.25 =0.50 =0.75 =1.00
Tapered RaƟo

Normalized Natural Frequency
normalized natural
frequency v/s tapered ratio 60 1st
for L/h = 5, k = 0 and Mode
different mode shapes 50 2nd
Mode
40 3rd
Mode
30 4th
Mode
20 5th
Mode
10
0
=0.00 =0.25 =0.50 =0.75 =1.00
Tapered RaƟo
References
1. Chakraborty A, Gopalakrishnan S, Reddy JN (2003) A new beam finite element for the analysis
of functionally graded materials. Int J Mech Sci 45:17–22
2. Aydogdu M, Taskin V (2007) Free vibration analysis of functionally graded beams with simply
supported edges. Mater Des 28(5):1651–1656
3. Sina SA, Navazi HM, Haddadpour H (2009) An analytical method for free vibration analysis
of functionally graded beams. Mater Des 30(3):741–747
4. Şimşek M (2010) Fundamental frequency analysis of functionally graded beams by using
different higher-order beam theories. Nucl Eng Des 240(4):697–705
5. Lee JW, Lee JY (2017) Free vibration analysis of functionally graded Bernoulli-Euler beams
using an exact transfer matrix expression. Int J Mech Sci 122:1–17
6. Frostig Y, Baruch M, Vilnay O, Sheinman I (1992) High-order theory for sandwich-beam

behaviour with transversely flexible core. J Eng Mech 118(5):1026–1043
7. Sankar BV (2001) An elasticity solution for functionally graded beam. Compos Sci Technol
61:689–696
8. Xiang HJ, Yang J (2008) Free and forced vibration of a laminated FGM Timoshenko beam of
variable thickness under heat conduction. Compos B Eng 39:292–303
9. Şimşek M, Kocatürk T (2009) Free and forced vibration of a functionally graded beam subjected
to a concentrated moving harmonic load. Compos Struct 90:465–473
10. Li XF, Wang BL, Han JC (2010) A higher-order theory for static and dynamic analyses of
functionally graded beams. Arch Appl Mech 80:1197–1212
11. Alshorbagy AE, Eltaher MA, Mahmud FF (2011) Free vibration characteristics of a functionally
graded beam by finite element method. Appl Math Model 35(1):412–425
12. Alshorbagy AE (2013) Temperature effects on the vibration characteristics of a functionally
graded thick beam. Ain Shams Eng J 4:455–464
13. Nguyen T-K et al (2015) Vibration and buckling analysis of functionally graded sandwich
beams by a new higher-order shear deformation theory. Compos B Eng 76:273–285
14. Kahya V (2012) Dynamic analysis of laminated composite beams under moving loads using
finite element method. Nucl Eng Des 243:41–48
15. Kahya V, Turan M (2017) Finite element model for vibration and buckling of functionally
graded beams based on the first-order shear deformation theory. Compos B Eng 109:108–115
16. Banerjee JR, Ananthapuvirajah A (2018) Free vibration of functionally graded beams and
frameworks using the dynamic stiffness method. J Sound Vib 422:34–47
17. Karmanli A (2018) Free vibration analysis of two directional functionally graded beams using
a third-order shear deformation theory. Compos Struct 189:127–136
18. Sinir S, Gultekin B (2018) Nonlinear free and forced vibration analyses of axially functionally
graded Euler-Bernoulli beams with non-uniform cross-section. Compos B Eng 148:123–131
19. Chen M, Jin G, Zhang Y, Niu F, Liu Z (2019) Three-dimensional vibration analysis of beam
with axial functionally graded materials and variable thickness. Compos Struct 207:304–322
20. Corr RB, Jennings A (1976) A simultaneous iteration algorithm for symmetric Eigen value
problems. Int J Numer Meth Eng 10(3):647–663
Parametric Study of Composite Plate
Using Free Vibration Analysis
Rajendra Kumar Sadangi, Mihir Kumar Sutar, and Sarojrani Pattnaik
Abstract This paper presents the free vibration analysis of laminated composite
plate to study the parametric variations on natural frequency of the plate. The param-
eters used for analysis are aspect ratio, number of layers in composite plate, support
conditions, number of layers and fibre orientation. Finite element model of the lami-
nated composite plate is prepared using ANSYS 13.0 software. The element chosen
for the analysis is shell 281, an 8-noded, six degrees of freedom. This element is used
for analysing thin to moderately thick structures. The effect of parametric variations
on the natural frequency is then studied. The results obtained from the analysis are
then compared with the experimental values obtained from the literature survey with
a satisfactory result.
Keywords Laminated plate · Orthotropic plate · ANSYS · Natural frequency and

fibre orientation angle
Nomenclature
[K] Stiffness matrix

[M] Moment matrix
ωn Angular velocity
1 Introduction
Today technology progress results in the continuous growth of structural material

types and improvement of their properties. The frequent use of composite mate-
rial as a group of metals, polymers, and ceramics have increased the demand for
R. K. Sadangi · M. K. Sutar (B) · S. Pattnaik

Veer Surendra Sai University of Technology Burla, Burla, Odisha, India
https://doi.org/10.1007/978-981-15-4488-0_76
906 R. K. Sadangi et al.
improvement in them for developing industries [1]. Since last four decades, plates
made by composite material are more used in many engineering purposes. The high
stiffness-to-weight ratio coupled with flexibility of the selected laminated material
can be turning to match the design demand and makes the laminated plate an interest
of structural component for manufacture.
In these days, generally laminated plates used in various fields to develop the
mechanical properties of material for research. The result of the above analysis in
terms of natural frequency using laminated composite plate is improved and resonant
omitted.
1.1 Illustrations
Ansar et al. [1] sequence reviews the modelling method along with their efficiency and
drawback for features of the material’s geometry, mechanical behaviour and impact
behaviour of three-dimensional woven composites. The authors have highlighted on
the recent developments in the field of modelling and design of woven composites,
so that their special mechanical character and impact behaviour can be understood.
Advantage and drawback and use of three-dimensional woven composite are also
highlighted in his paper.
Charles and Byron [2] have performed a free vibratio analysis of symmetrically
and unsymmetrical laminated anisotropic composite plates with clamped edges. Nat-
ural frequency of the laminated anisotropic plate was obtained using linear analysis.
Further an approximate solution was obtained using Rayleigh-Ritz energy method.
A good agreement was obtained between the numerical and experimental work.
Chakraborty et al. [3] found both numerical and experimental work of the free
vibration of composite fibre-reinforced plates. They used modal testing method using
impact excitation to find out the corresponding frequency response function. Finite
element methods using an isometric element determine the modal features. In his
works, they validate result obtained from commercial finite element package (NISA)
with both theoretical and experimental work.
Cawley and Adams [4] found the natural modes of vibration of laminated com-
posite plates with theoretical and experimental work. Natural frequency and mode
shapes presented different ply orientations’ comparisons between theoretical and
experimental frequencies.
First Crawly [5] found out the natural frequency and mode shape using
graphite/epoxy and aluminium plates are experimentally determined. In his observa-
tion, comparison between calculated and observed mode was excellent but discrep-
ancy in frequency result is due to the difference between dynamic flexural modulus
and static in-plane modulus.
Mishra and Sahu [6] present experimental works to find free vibration of glass
epoxy composite material with different boundary condition. Modal testing method
used for obtaining the frequency of the structure. This experimental result is compared
with the FEM numerical which is based on first-order shear deformation. In this paper,
Parametric Study of Composite Plate Using Free Vibration Analysis 907
the effect of different geometry parameters like a number of layers, aspect ratio, fibre
orientation and boundary condition of composite plate is presented.
Ju et al. [7] present the finite element formulation for free vibration analysis of
composite plate with multiple delaminations. This formulation introduces the effect
of transverse shear deformation as well as the bending extension coupling due to the
influence of delaminated plates.
Kamal and Durvasula [8] found some studies on free vibration of composite
laminates using the modification of shear deformation and composite plate theory
and applied the Rayleigh-Ritz energy method. The numerical result for clamped
boundary condition of composite plate is compared with various earlier published
results.
Qatu and Leissa [9] find free vibration of cantilever laminated composite plate
using the Ritz method and further he compared the result with experimental value.
Rikards [10] uses sandwich composite plate with viscoelastic layers for vibration
and damping analysis. In his paper, he showed the natural frequency variation of free
and damping vibration and described that efficiency of damped material.
Soares [11] use the Eigen-frequencies technique to find the mechanical proper-
ties of composite material and compared the result with experimental analysis of
composite material.
Srinivasa et al. [12] presented experimental and finite element studies on free
vibration of isotropic and laminated composite skew plates. The natural frequency
obtained taken CQUAD8 finite element of NASTRAN and compared made between
the experiment result and finite element solution.
2 Material and Methodology Composite Materials
The composite material is defined as combination of matrix and fibre reinforcement

material, the formation of properties superior to the properties of the individual
components. In the case of composite, the reinforcement is the fibres and it is used
to fortify the matrix in terms of strength and stiffness. The fibre and a base material
are known as matrix that yields best mechanical properties than its original material.
Composite materials are best for structures that require high strength/weight and
stiffness/weight.
2.1 Free Vibration of Laminated Composite Plate
The first-order shear deformation theory with finite element methods is used for
finding out natural frequency and mode of vibration of laminated composite plates.
Shear deformation theory is same as classical laminated plate theory but only includes
the transverse shear strain (interlaminar shear strain) which was not considered in
classical laminate plate theory.
Free vibration motion of dynamics structure with linear single degree of freedom
system without damping given below
d2 u
m + ku = 0 (1)
dt 2
.
For static equilibrium position, the displacement and velocity x = x(0), x =
.
x(0), respectively.
Solution of equation obtained at time period t, x = x(t), ẋ = ẋ(t)
.
x(0)
x(t) = x(0) cos ωn t + sin ωn t (2)
ωn

Natural frequency of vibration is, ωn = k
m
.
2.2 Analysis in ANSYS
The composite plate is analysed by using finite element analysis using the software
package ANSYS 13.0. Shell 281 element, which can be used for analysis of thin
to moderately thick materials, has been used for the composite late analysis. This
element has 8 nodes and each node has a 6° of freedom 3 in translation direction
and 3 in rotational direction. This material is suitable for use in linear, large rotation
and large strain nonlinear application. This element is also used in a layered applica-
tion for composite shell and sandwich plate model and accuracy of composite shell
governing by using first-order shear deformation theory (Fig. 1).
3.1 Case-1
Consider epoxy glass fibre composite material with different combination of length
(a) and breadth (b) with thickness of 20 mm (20 no. of layer lamina) and different
boundary condition and same analysis with ANSYS 13.0 using shell 281 in the
analysis (Fig. 2).
The result of ANSYS validated through analytical FEM analysis and experimental
analysis made by Srinivasa et al. [12] in his paper considers skew angle a = 0°. Here
is first 3 modes of vibration having aspect ratio (a/b = 1) of antisymmetric ply. For
laminated glass/epoxy composite plates of different ply combination given below,
these value are compared with experimental analysis by Srinivasa et al. [12] (Tables 1,
Fig. 1 Ansys shell 281
Fig. 2 Layer stacking cross angle ply composite
Table 1 Different mode of

SET Time/frequency Load step Sub-step cumulative
vibration of antisymmetric
layers 1 7.9942 1 1
2 15.205 1 2
3 18.455 1 3
Table 2 Modes of vibrations

SET Time/frequency Load step Sub-step cumulative
of cross-ply composite
1 7.7889 1 1
2 14.587 1 2
3 19.378 1 3
2 and 3).
3.2 Case-II
Again taken Graphite/epoxy composite material first three natural frequency obtained
from Finite element analysis using the Software ANSYS and same compared the
analytical work obtained by Ju et al. [7] and Experimental work by Mishra and Sahu
[6] (Table 4).
4 Conclusion
The following observations are obtained in this paper.

In case-I, comparison between different combinations of angle ply laminate with
frequency of vibration on different arrangement of the mode of vibrations. Here,
in conclusion drawn from Fig. 3, frequency of vibration in the first two modes of
vibrations is nearly same and afterwards, vibrations increase tremendously.
The chart (Fig. 4) frequency is least value when using angle ply having
(+90º /−90º /+···) and highest value whenever using angle ply laminate having ply
combination like (0º /−0º /+···). The angle ply lamination having a stacking sequence
(+90º /−90º /+···) results in reduced vibration in the system and can be used in sit-
uations where vibration is a major drawback in the system and it has to be
reduced.
In case II, consider eight-ply graphite/epoxy plate having dimension of length
152 mm and breadth 76 mm and thickness 1.04 mm with laminate slackness combi-
nation (0°/90°/45°/90°)s. In this paper, comparison among the result by three differ-
ent approaches like first in finite element analysis using ANSYS and same validated
through analytical and experimental done by authors (C. V. Srinivasa, Y. J. Suresh
and W. P. Prema Kumar) in his paper. In this ANSYS analysis, result is conver-
gence to both analytical and experimental result plot in Figs. 5, 6 and 7 by different
combinations like clamped, cantilever and simple supported plate.
It has been observed that the natural frequency has an increasing trend in all
angles of ply orientations. The present analysis found to be well in agreement with
the experimental results of Srinivas et al. [12] and with Ju et al. [7].
Table 3 Comparison of analytical and experimental values [12] of natural frequencies
M Aspect Ratio
o 1 1.5 2 2.5
d
e
[12]
[12]
[12]
[12]
FEM
FEM
FEM
FEM
Value
Value
Value
Value
Analytical
Analytical
Analytical
Analytical
Exp. Values
Exp. Values
Exp. Values
Exp. Values
o
Angle 1 15.30 15.01 15.1 15.18 14.1 14.2 14.98 13.09 12.4 14.82 14.12 13.19
ply(+00/- 2 15.99 16.1 15.08 17.19 17.01 16.01 17.53 16.21 15.89 18.61 18.1 17.19
00/+00/-) 3 19.43 17.07 18.4 27.00 27.01 26.1 36.68 25.76 34.2 40.85 41.91 39.32
Angle 1 9.45 8.09 7.99 9.54 9.421 8.99 9.039 8.23 9.01 8.85 8.09 8.09
ply(+450/- 2 12.85 12.91 15.2 16.33 16.10 15.1 19.11 18.29 19.4 22.48 22.59 22.30
450/+450/-) 3 22.68 22.1 18.4 26.46 26.41 26.3 26.11 25.19 24.2 24.54 23.08 24.76
Angle 1 7.05 7.09 6.11 7.21 7.11 7.09 7.19 7.12 7.99 6.83 6.07 6.59
ply(+900- 2 8.44 7.98 12.69 10.12 10.33 11.11 11.83 11.84 12.01 13.02 13.81 13.01
Parametric Study of Composite Plate Using Free Vibration Analysis
900+900-) 3 19.39 18.54 16.56 19.87 19.23 18.78 19.82 18.01 19.78 18.82 18.47 18.59
Cross 1 11.90 18.87 7.78 12.17 12.01 11.10 12.13 11.01 12.09 11.53 11.90 12.07
ply(00+900 2 12.77 12.56 14.58 14.10 13.89 12.89 15.4 13.89 12.83 16.09 17.90 16.92
+) 3 19.42 18.23 19.37 33.41 34.09 32.12 33.45 32.12 33.40 31.78 31.45 32.01
911
Table 4 Comparison of natural frequency

Boundary Reference Mode-1 Mode-2 Mode-3 Mode-4
condition papers frequency frequency frequency frequency
(Hz) (Hz) (Hz) (Hz)
All side Ju et al. [7] 346.59 651.51 781.06 1017.2
clamped Present FEM 365.44 720.51 839.80 2283.2
One side Ju et al. [7] 41.347 60.663 221.52 258.72
clamped and Present FEM 31.533 54.164 198.04 230.07
other 3 sides
free Experimental 41.1652 60.534 220.5114 257.7930
[12]
All sides are Ju et al. [7] 164.37 404.38 492.29 658.40
simple Present FEM 163.27 400.00 493.621 660.31
supported
boundary Experimental 163.7453 401.3175 494.5912 651.652
condition [12]
Fig. 3 Different modes of vibration
Fig. 4 Different aspect ratio variation in frequency versus angle ply chart
Fig. 5 Layer stacking cross-angle ply composite

Appendix
Material propertied used in finite element calculation for laminated composite

Graphite/Epoxy based on data book (unpublished reference) is given below.
E L = 128 GPa, E T = 11 GPa, ν L T = 0.25, G L T

= 4.48 GPa, G 13 = 1.53 GPa, density
= 1.5 × 103 kg/m3 , Plythickness = 20 mm
References
1. Ansar M, Xinwei W, Chouwei Z (2011) Modeling strategies of 3D woven composites: a review.

Compos Struct 93:1947–1963
2. Bert CW, Mayberry BL (1969) Free vibrations of unsymmetrically laminated anisotropic plates
with Mayberry clamped edges. J Compos Mater 3:282–293
3. Chakraborty S, Mukhopadhyay M, Mohanty AR (2000) Free vibrational responses of FRP
composite plates: experimental and numerical studies. J Reinf Plast Compos 19:535–551
4. Cawley P, Adams RD (1978) The predicted and experimental natural modes of free-free CFRP
plates. J Compos Mater 12:336–347
5. Crawley EF (1979) The natural modes of graphite/epoxy cantilever plates and shells. J Compos
Mater 13:195–205
6. Mishra I, Sahu SK (2015) Modal analysis of woven fiber composite plates with different
boundary conditions. Int J Struct Stab Dyn 15(1):1540001–1540017
7. Ju F, Lee HP, Lee KH (1994) Finite Element analysis of free vibration of delaminated composite
plates. Composite Engineering 5:195–209
8. Kamal K, Durvasula S (1986) some studies on free vibration of composite laminates. Compos
Struct 5:177–202
9. Qatu Mohamad S, Leissa Arthur W (1991) Natural frequencies for cantilevered doubly- curved
laminated composite shallow shells. Compos Struct 17:227–255
10. Rikards R (1993) Finite element analysis of vibration and damping of laminated composites.
Compos Struct 24(3):193–204
11. Soares MCM, de Freitas MM, Araujo AL (1993) Identification of material properties of
composite plate specimens. Compos Struct 25:277–285
12. Srinivasa CV, Suresh YJ, Prema Kumar WP (2014) Experimental and finite element studies on
free vibration of skew plates. Int J Adv Struct Eng 6:48
Free Vibration Analysis of Hybrid
Composite Beam Under Different
Boundary Conditions and Thermal
Gradient Loading
Prabhat Pradhan, Mihir Kumar Sutar, and Sarojrani Pattnaik
Abstract This paper presents a free vibration analysis of hybrid composite beam
under thermal gradient loading. Theoretical vibration analysis of the hybrid com-
posite beam is performed and variation of the (z/h) with respect to Young’s modulus
under variation of the thermal gradient is studied. The results obtained from the
analysis of hybrid composite beam were compared with the plain composite beam
and it was observed that hybrid composite beam possesses better modulus of elas-
ticity compared to that of plain composite beam. This analysis would be vital for
performing further research in this domain.
Keywords Functionally graded material · FG-CNT-Reinforced hybrid composite

beam · Power law distribution · Temperature distribution · Free vibration analysis
1 Introduction
Functionally Graded Materials (FGMs) are the innovative advance quality materials
in the field of composites with respect to their strength, mechanical and thermal prop-
erties. Nowadays, the modern requirement of the industry in the field of aerospace
and power sectors needs the rapid evolution of new components, which provides the
researchers to invent new materials in order to satisfy the functional requirements of
modern technology. Due to excellent heat-resisting property along with resistance to
corrosion, erosion and fracture of FGM materials, these are used as smart materials
in modern technologies. With the excellent thermal properties of FGM, these are also
found to have erosion and corrosion resistance and high fracture resistance; these are
used as smart materials in modern technologies.
P. Pradhan · M. K. Sutar (B) · S. Pattnaik

Mechanical Engineering Department, Veer Surendra Sai University of Technology, Burla,
Odisha, India

https://doi.org/10.1007/978-981-15-4488-0_77
916 P. Pradhan et al.
Integration of Carbon Nanotubes (CNT) with FGM, leads to a newer material

category called hybrid composite. This results in important improvements in com-
posite properties, even in very small volumes of CNT. In actual structural appli-
cations, Hybrid composite is incorporated as a type of advanced material that can
be incorporated as structural elements into beams, sheets or shells. Therefore, it is
more important to explore the mechanical results of these structures as beams. Many
researchers have imparted their effort in this particular domain.
Free vibration behaviour of CNT-reinforced functionally graded composite plates
was investigated by Selim et al. [1], using Higher Order Shear Deformation Theory
(HSDT) and rules of mixture. The frequency response and mode shape obtained
were studied in the analysis. VDUY et al. [2] have used a Finite Element Analy-
sis (FEA) to investigate the different behavioural properties of FG-CNT beam. The
results obtained were further compared for different orientations of CNTs. Kazem
et al. [3] have performed the FEA vibration analysis of functionally graded beam, sub-
jected to different thermal conditions. Armagan [4] has performed the free vibrational
analysis of two-directional functional graded beams which are subjected to differ-
ent boundary conditions like simply supported (SS) beam, clamped-clamped (CC),
clamped-simply supported (CS) and clamped-free (CF) beams which are evaluated
by applying shear deformation theory of third order. Further, a free vibrational study
of FG-CNT-reinforced nanocomposite sandwich plates resting on elastic foundation
has been performed by Dastjerdi et al. [5] and randomly oriented straight carbon
nanotubes have been reinforced in the sandwich plates. The governing equations
were derived using Hamilton’s energy principle and are solved by using Navier’s
method and various effects of CNT volume fraction and various parameters of natural
frequency have been studied.
The thermomechanical vibration analysis of sandwich beams having reinforced
with CNT composite functionally graded along thickness direction has been per-
formed by Ebrahimi et al. [6], in which the material property variation has been done
using rule of mixture and is temperature-dependent. The governing equation was
formulated using Hamilton’s principle and differential transform method (DTM) is
being used for verifying the material properties. The first-known vibration analysis
of fully functionally graded carbon nanotube-reinforced hybrid composite (FFG-
CNTRHC) laminates has been by Kuo [7]. In this case, carbon nanotubes have been
graded over graphite/epoxy laminates and were distributed by using some distri-
bution functions and rules of mixture to obtain desire stiffening effect. Then the
required stiffness and mass matrices were generated and solved for natural frequen-
cies. Liew et al. [8] have performed the mechanical analysis of FG-CNT based com-
posite where the superlative properties of carbon nanotubes along with their excellent
reinforcement properties with composites were discussed. A nonlinear free vibra-
tion of FGB embedded with nanocomposites has been performed by Liang Ke et al.
[9]. The governing equations were derived for this beam by Ritz’s method and then
nonlinear vibration frequencies are observed for FG-CNTRC beams with different
end supports. The nonlinear vibration characteristics of carbon nanotube-reinforced
composite beam subjected to thermal conditions have been analyzed by Chaudhari
et al. [10]. The linear free vibration analysis of nanocomposite beams reinforced
Free Vibration Analysis of Hybrid Composite Beam … 917
with single-walled carbon nanotubes (SWCNTs) has been performed by Lin et al.
[11]. Hamilton’s principle has been used to obtain different strain energy and kinetic
energies for such beam and p-Ritz method is used in order to solve these equations to
generate the governing equations. Finally, the vibrational frequencies for such beams
were investigated.
2 Mathematical Modelling of Hybrid Composite Beam
The carbon nanotube-reinforced functionally graded beam has been modelled math-
ematically by using power law function of FGM using Timoshenko beam theory
(First-order shear deformation theory) and further it is applied to determine the
physical properties of CNTs and matrix material, i.e. Young’s Modulus.
Consider a composite beam system having length L, width b, thickness h along
the x, y and z directions, respectively. Three types of alignment of carbon nanotubes
in CNT beams, such as uniformly distributed CNT beam (UD-CNT), functionally
graded CNT of type A (FGA-CNT), in which CNTs are more concentrated at bottom
of beam and FG beam of type X(FGX-CNT), in which CNTs are more concentrated
at the top and bottom portion of beam, have been taken into consideration for the
analysis. Figure 1 depicts the three alignment categories.
A method that is used for the constituent material distribution within FG-CNT
beam is termed as rules of mixture, in which specific laws like power/exponential
laws are used for gradation of material in a specific manner.
Fig. 1 Functionally graded carbon nanotube-reinforced composite beams: a Co-ordinate system

of beam. b Cross-section of UD-FGM, FGA-CNT and FGX-CNT beams
Here, we will use the power law to describe the material properties of the beam,
where a particular index is used for constituents. The various properties of “the
material can be evaluated by using power law, where the volume of both elements
varies in the direction of thickness. Also, the volume fraction varies in accordance
with power law along thickness direction.”
It is given according to subsequent equation [11]:
z n
Vcnt = + 0.5 (1)
h
where n is the index representing power law and h represents the thickness of the
beam and z is the co-ordinate axis which varies from (−h/2 ≤ z ≤ h/2), along
direction of thickness,
And
Vcnt + Vm = 1 (2)
where V cnt is the fractional volume of CNT in FGM and V m is the fractional volume
of matrix in FGM. This volumetric relationship gives the composition of a layer of a
beam at a particular co-ordinate in z-axis. The physical properties of FG-CNT beam
varies according to the relation [11]:
k
1 z
P(z) = Pcnt + (Pm − Pcnt ) + (3)
2 h
The variation of Young’s modulus of FGX-CNT beam with z/h, is as shown in

Fig. 2, which depicts the variation of material property of the FG-CNT (mainly for
FG-X Distribution) beam along the thickness direction obeying a power law for
different values of power law exponent. Further, it has been shown that the variation
mostly linear for the power index of k = 1 and all other variations are mostly mirror
image about this line.
3 Free Vibration Analysis of FG-CNT-Reinforced Hybrid

Composite Beam
The first-order shear deformation theory (FSDT), gives the displacement along axial
direction u and transverse displacement w at any point of beam is given by
u(x, z, t) = u 0 (x, t) + zψ(x, t) and w(x, y, z, t) = w0 (x, t) (4)
By implementing the finite element analysis on FG-CNT beam, the above govern-
ing equations are solved and after solving the governing differential equations using
interpolating functions, the stiffness and mass matrix can be obtained. The discrete
Fig. 2 Power law distribution of material of CNT over PMMA matrix of FG-CNT beam for various
power index along thickness direction
equation for the free vibration analysis of laminated composite beam is represented
by;
M d̈ + K d = 0 (5)
where
d Displacement vector
d̈ 2nd order derivative of displacement w.r.t time ‘t’
M and K are the global mass and stiffness matrix which are the assemblies of
elemental mass and stiffness matrix.
Further, the above Eq. is solved in order to get the natural frequency from equation;

[K ] − w 2 [M] φ = 0 (6)
By applying the finite element study on FG-CNT beam, the above governing
equations are solved and subsequently solving the governing differential equations
using interpolating functions, the stiffness and mass matrix can be obtained. Finally,
the derived equation of motion is given by
[M]{q̈} + [K ]{q} = 0 (7)

where [M] is the mass matrix and [k] is the stiffness matrix. By solving this equation,
we will obtain the natural frequency of vibration of FG-CNT beam.
In the present analysis of FG-CNTRC beam, the matrix material has been taken
as polymethyl methacrylate (PMMA) and single-walled armchair (10, 10) carbon
nanotubes are as the reinforcing element, hence their material properties [12] at a
room temperature of 300 K which is used in the analysis of free vibration in FEM is
tabulated as follows (Table 1).
Further, a dynamic study is performed to validate the effectiveness of the
present study for vibration analysis of FG-CNTRC beam of present formulation
and modelling in computer with the previous paperwork.
Also, to evaluate the present modelling of FG-CNTRC beam in FEM to observe
the dimensionless frequency parameter following equations have been used

ρm A L2 ρm
ω = ωL 2
. =ω . (8)
Em I h Em
An iterative program in the computer is modelled in MATLAB in order to assemble

the stiffness and mass matrices for the required number of elements in FG-CNTRC
beam.
First of all the effective material properties of FG-CNTRC were evaluated and
then the comparison study is carried out for the successful use of finite element
method for FG-CNT hybrid composite beam with all the dimensions of FG-CNTRC
beam from the literature [12].
The fundamental natural frequency for the different Slenderness ratios (aspect
ratios L/h) of a FG-X CNTRC with volume fraction of CNTs VC N T = 0.12 beam
for different boundary conditions is evaluated below. The different boundary con-
ditions taken are: clamped-free (C-F), hinged-hinged (H-H), clamped-hinged(C-H)
and clamped-clamped(C-C) are as follows:
Table 1 Material properties considered in FG-CNTRC beam (300 K) [12]

Properties(units) Matrix (PMMA) Reinforcement fiber (CNTs)
Poisson’s coefficient ν m = 0.3 ν C N T = 0.19
Mass density(kg/m3 ) ρ m = 1190kg/m 3 ρ C N T = 1400kg/m 3
Young’s modulus of Em = 2.5G Pa E C N T = 950G Pa,
elasticity(GPa) CNT
E 11 = 600G Pa, E 22
CNT
= 10G Pa
Shear modulus (GPa) – GC N T = 17.2G Pa
12
Fig. 3 Fundamental frequency versus aspect ratio of FG-X beam for different boundary conditions
The plot for variation of fundamental frequency of vibration of FG-X CNTRC

beam for various boundary conditions (C-F, C-H, H-H and C-C) are plotted for
different values of Slenderness ratios as shown in the Fig. 3.
The plot for comparison of the evaluated fundamental frequency of vibration
for different aspect ratios (L/h) for a FG-CNT beam having FG-X Distribution for
clamped-clamped(C-C) boundary conditions is as shown in Fig. 4.
Fig. 4 Comparison of evaluated fundamental frequencies with literature. [12] for FG-X C-C beam
Fig. 5 Fundamental frequency versus aspect ratio of FG-X Beam for the different volume fraction
The variation of fundamental frequency versus aspect ratio(L/h) of FG-X FG-X

hybrid composite beam for clamped-clamped boundary condition as shown in Fig. 5
for different volume fractions of CNTs (i.e. vcnt = 0.12, vcnt = 0.17 and vcnt = 0.28,
from the literature [12]) distributions over a given cross-section of the beam.
The variation of dimensionless frequency of vibration for different CNT grada-
tions over PMMA matrix with CNTs volume fraction of V CNT = 0.12, and boundary
condition has also been calculated and is listed in Table 2 (Table 3).
From this, it is observed that the FG-CNT beam having X-distributions,
comparatively larger frequency than the other distribution profile.
Table 2 The comparison of dimensionless fundamental frequency w.r.t slenderness ratio (L/h) of
FG-CNTRC beam for different boundary conditions
L/h Fundamental frequency (ω1 ) of vibration for different boundary conditions of
FG-X CNTRC beam
ω1 for C-F ω1 for H-H ω1 for C-H ω1 for C-C
15 0.213 0.562 0.850 1.214
20 0.162 0.460 0.652 1.029
25 0.139 0.363 0.550 0.794
35 0.095 0.260 0.386 0.554
40 0.078 0.247 0.347 0.486
45 0.0727 0.201 0.285 0.440
50 0.067 0.179 0.279 0.414
Table 3 The variation of dimensionless frequency of vibration for different CNT gradation and
boundary condition
Boundary cond. FG-CNT distribution profile
V
UD FG- FG-♦ FG-X
C-C 1.576 1.472 1.392 1.675
C-F 0.394 0.305 0.301 0.464
C-H 1.306 1.221 1.065 1.432
H-H 1.032 0.986 0.782 1.171
5 Conclusions
A comparative study has been carried out for the FG-CNT beam having constituent
materials of PMMA as matrix and CNTs as reinforcement in four different distri-
butions subjected to four different boundary conditions. The free vibration analysis
of FG-CNT beam through finite element method modelling concludes the following
results.
From the above analysis, it is illustrated that:
• The volume fraction distribution of CNTs having a significant effect on the natural
frequency of vibration of the beam and the beam having the highest volume
fraction of CNT is having more natural frequency of vibration.
• The Slenderness ratio (L/h) of the FG-CNT beam is also having a prominent effect
on the behaviour of natural frequency of vibration, and when this ratio increases,
then the fundamental frequency of vibration value decreases consequently.
• Carbon nanotube beams in which orientation takes place in FG-X manner having
the highest fundamental frequency as compared to other distributions.
• Increasing the volume fraction of CNT leads rise to an increase in the natural
frequency of vibration.
• This method of analysis of FG-CNT can also be applied for forced vibration,
buckling analysis and beam subjected to different higher temperature conditions.
The power law distribution plot also made in MATLAB program to analyze the
behaviour of Young’s modulus of CNT-reinforced beam with the directional variation
of z/h, which resulting in a power law index of 1.5 will be most suitable for high-
temperature application of FG-CNT beam.
References
1. Selim BA, Zhang LW, Liew KM (2015) Vibration analysis of CNT-reinforced function-
ally graded composite plate in a thermal environment based on Reddy’s higher order shear
deformation theory. Compos Struct 156:276–290
2. Vo-Duy T, Ho-Huu V, Nguyen-Thoi T (2019) Free vibration analysis of Laminated FG-CNT
reinforced composite beams using finite element method. Front Struct Civil Eng 13(2):324–336
3. Kashyzadeh KR, Asfarjani AA (2016) Finite element study on the vibration of functionally
graded beam with different temperature conditions. Adv Mater 5(6):57–65
4. Karamanli A (2018) Free vibration analysis of two directional functionally graded beams using
a third order shear deformation theory. Compos Struct 189:127–136
5. Moradi-Dastjerdi R, Payganeh GH, Malek-Mohammadi H (2015) Free vibration analyses
of functionally graded CNT reinforced nanocomposite sandwich plates resting on elastic
foundation. J Solid Mech 7(2):158–172
6. Ebrahimi F, Farazmandnia N (2018) Vibration analysis of functionally graded carbon nanotube-
reinforced composite sandwich beams in thermal environment. Adv Aircr Spacecraft Sci
5(1):107–128
7. Kuo SY (2018) Free vibration of fully functionally graded carbon nanotube reinforced
graphite/epoxy laminates. Mater Res Expr 5(3)
8. Liew KM, Lei ZX, Zhang LW (2015) Mechanical analysis of functionally graded carbon
nanotube reinforced composites: a review. Compos Struct 120:90–97
9. Ke LL, Yang J, Kitipornchai S (2010) Nonlinear free vibration of functionally graded carbon
nanotube-reinforced composite beams. Compos Struct 92(3):676–683
10. Chaudhari VK, Lal A (2016) Nonlinear free vibration analysis of elastically supported
nanotube-reinforced composite beam in thermal environment. Procedia Eng 144:928–935
11. Lin F, Xiang Y (2014) Vibration of carbon nanotube reinforced composite beams based on the
first and third order beam theories. Appl Math Model 38(15–16):3741–3754
12. Kumar P, Srinivas J (2017) Free vibration, bending and buckling of a FG-CNT reinforced
composite beam. Multidiscipline Model Mater Struct 13(4):590–611
Computational Analysis of Boring Tool
Holder with Damping Force
G. Lawrance, P. Sam Paul, and A. S. Varadarajan
Abstract In the boring process, tool vibration is an important parameter due to

its overhanging length and also leads to high cutting force, poor surface finish and
increase in tool wear. To suppress tool vibration and increase the cutting performance,
a novel idea has been developed in rheological fluid. In present work, a boring tool
holder with magnetorheological damper was analysed. Magnetorheological damper
received massive responsive due to their capability to reversely change from a linear,
free-flowing viscous fluid to semis solid within seconds when a magnetic field is
applied. In this paper, a boring tool holder was analysed using ANSYS software
with damping force applied in various direction and without MR effect. The results
prove that the use of Magnetorheological damper reduces tool vibration significantly.
Keywords Boring tool holder · Magnetorheological damper · Damping ratio ·

Harmonic analysis · Modal analysis
1 Introduction
The modern industries are more concerned about achieving a product with reduced
ecological effect during its production and achieving high standard at the end prod-
uct, increase in dimensional accuracy, good surface quality, increased production
rate, reduced tool wear and economy. In the manufacturing process, dynamic inter-
action between tool and workpiece results in tool vibration. The vibration in tool and
wavy surface on the workpiece result modified chip thickness, which also results in
invariable cutting forces to stimulate the work piece and machine [1]. Quintana and
Ciurana [2] proposed that unbalance dynamic stiffness leads to self-excited vibra-
tion which further leads to diminishing in productivity and quality. The vibration
further leads to poor surface quality, extreme tool wear and irritating noise [3]. Chen
G. Lawrance (B) · P. Sam Paul

Karunya Institute of Technology and Sciences, Coimbatore, Tamil Nadu 641114, India
A. S. Varadarajan
MES College of Engineering, Malappuram, Kerala 679573, India
https://doi.org/10.1007/978-981-15-4488-0_78
926 G. Lawrance et al.
and Tsao [4] described the vibration behaviour by using flexible workpiece to focus
on regenerative chatter rather assuming rigid workpiece. A large slender workpiece
in machining at high spindle speed leads to an increase in deformation and affects
the stable analysis whereas smaller deformation at less slender workpiece at high
spindle speed [5]. Andren et al. [6] clarify the dynamic response of a boring bar in a
non-linear and time-varying process when it is focused on three different workpieces
and is measured in varying cutting depth and cutting speed direction. Tool vibration
in boring operation is high when it is compared with turning operations due to its
larger slenderness ratio [7]. So it is essential to reduce the tool vibration in boring
operation effectively by means of damper which enhances the damping capability
when compared with self-damping of conventional tool to suppress vibration [8].
Metal cutting industries are expecting better quality, increased productivity,
decreased production cost and better technique to suppress tool vibration. To reduce
tool vibration researchers used outlines in past and amongst them, the influence of
smart material on tool has gained attention. Wang and Fei [9] proposed a boring bar
with variable-stiffness consisting of an electrorheological (ER) fluid to reduce chatter.
Magnetorheological fluid is a smart fluid and MR fluids was reported by Rabinow
[10] who is the pioneer to use magnetorheological fluid with electromagnetically
controllable clutch. Mei et al. [11] proposed Magnetorheological fluids which sup-
press vibration in the boring tool by adjusting the damping and natural frequency
of the system. A Magnetorheological damper reduces tool vibration successfully
in a turning tool when it was analysed using ANSYS software with and without
Magnetorheological damper [12]. The analysing the boring tool holder subjected to
Magnetorheological damper with different directions that reduces tool vibration by
means of computational method is done in the present investigation.
2 Selection of Workpiece and Tool Holder
In the present investigation, AISI 4340 steelwork piece which is hardened to 45

HRC is used and the workpiece consists of 80 mm outer diameter and 40 mm inner
diameter, 100 mm length were used. The AISI 4340 steel in weight percent with
chemical composition is 0.41% C, 0.87% Mn, 0.28% Si, 1.83% Ni, 0.72% Cr, 0.20%
Mo and rest Fe [13]. The tool holder is S25T PCLNR 12F3 which is a Widaxtool
with diameter 23 mm and 300 mm in length is used. The geometry of the tool holder
is shown in Fig. 1. The boring tool contains Tool shank, insert and sim. The boring
tool holder material and its properties are detailed in Table 1. Multicoated tungsten
carbide inserts CNMG 120408 MT TT5100 from M/s Taegu Tec is used for the
analysis purpose.
Computational Analysis of Boring Tool Holder with Damping … 927
Fig. 1 Line sketch of apparatus determining the damping force [1]
Table 1 Material properties of tool holder [13]

Material Properties Density (Kg/m3 ) Poisson’s ratio Young’s modulus
(N/m2 )
Tool Shank Carbon steel 7850 0.3 2.09E+11
Insert Tungsten carbide 15,800 0.28 5.55E+11
sim steel 7850 0.29 2.09E+11
3 Magnetorheological Damping Force for Boring Tool

Holder
In the present study, the computational analysis of boring tool holder was analysed
with damping force applied in various direction. A line sketch of the apparatus is
shown in figure which was used to find the damping force to be applied in various
direction. The piston (P) forces the fluid in the hydraulic chamber to move in loading
cylinder when the handle (H) is rotated and the loading piston (Q) moves up along the
loading platform. The applied voltage was controlled and the current was measured
when the coil of the damper was connected to a circuit. The loading chamber is consist
of pressure gauge and the area (Ap) of the loading piston, pressure (Pr) acting on the
piston and the force (Fp) acting on the piston is given by
Fp = Pr × Ap × 9.81 N
The MR fluid damper is made active by applying the voltage V across the coil and
the plunger touches the loading platform. When the handle (H) rotated in clockwise
there is an increase in the pressure of the chamber. When the force applied by the
Fig. 2 A model of boring tool holder
loading platform balances the resistance offered by the MR fluid damper, the plunger
just starts moving up and the limiting pressure indicated by the pressure gauge was
noted which a measure of the resisting force that is offered by the MR damper for
the applied voltage V [1] and the damping force is calculated as 102N.
4 Computational Analysis
4.1 Geometric Modelling
The boring tool holder was modelled using ANSYSsoftware and is shown in Fig. 2.
4.2 Grid Independence Study
Solid element of type Solid 185 is used in the present work. Based upon the accuracy,
number of elements is chosen and done using Grid Independent Study. The accurate
results can be obtained when there is an increase in number of element, beyond a
certain number of elements accuracy cannot be upgraded. Grid independent study is
carried out until deflection becomes constant. As in Fig. 3, beyond 0.315346 mm the
deflection does not change and the number of elements was found to be 2 × 106 . The
use of 2 × 106 elements contains degrees of freedom in a huge number and it is not
possible to store in the computer memory. Hence, the number of element required is
45 × 105 .
Grid Independence Study

0.31536
0.31534
DeflecƟon (mm)
0.31532
0.3153
0.31528
0.31526
0.31524
0.31522
0.3152
1500000 2000000 2500000 3000000 3500000 4000000 4500000
No.of.Elements
Fig. 3 Grid Independence Study
4.3 Static Analysis
The deflection values for different clamping length are calculated using static analy-
sis. The cutting force used in the insert tip is calculated based on the formula. Cutting
force = 396,000 × Depth of Cut × power consumption × feed rate. A 500 N force
is applied at the tip of the boring tool holder in a vertical direction at a distance of
85 mm from the tool tip [13]. Then damping force is applied in a vertical direction,
horizontal direction and holding direction. The performance of the static analysis for
the 100 mm, 125 mm, 150 mm overhanging length individually with the application
of damping force and the corresponding maximum deflections were noted. The same
analysis was performed again but without damping force. The deflections that were
found for various clamping lengths are provided in Table 2. It is clearly seen from
Fig. 4, 5 and 6 that when the clamping length increases the deflection decreases.
Table 2 Overhanging length versus deflection versus damping force direction

S.No. Overhanging With damper (mm) Without damper
length (mm) Axial force Vertical force Holding force (mm)
1 150 0.37541 0.36925 0.17294 0.3668

2 125 0.37977 0.36529 0.10569
3 100 0.38528 0.40801 0.063886
Fig. 4 Deflection for 150,125 and 100 mm tool overhanging length with axial force
Fig. 5 Deflection for 150,125 and 100 mm tool overhanging length with vertical force
Fig. 6 Deflection for 150,125 and 100 mm tool overhanging length with holding force
4.4 Modal Analysis
The dynamic response is excited by input on structures and fluids when it is measured
and analysed in the field is called Modal analysis. The system will collapse when
the frequency of the system reaches its natural frequency. The system frequency
should be below its natural frequency to avoid catastrophic failure. Computing the
natural frequency of the system modal analysis avoids such a failure and for different
overhanging length, the modal analysis was done. The extreme natural frequency
obtained was 4390.6 Hz at 150 mm overhanging, 5914.8 Hz at 125 mm overhanging
and the corresponding natural frequency for 100 mm overhang length was found to
be 7523 Hz for holding position are tabulated in Table 3
4.5 Harmonic Analysis
A sustained cyclic response is produced by the sustained cyclic load on a structural

system. The dynamic behaviour of forced vibrations on the structures can be pre-
dicted, which ensure whether the designs were successful in overcoming the fatigue,
resonance, and other harmful effects are known as Harmonic analysis. This technique
is used to determine steady-state response that varies sinusoidal with time form linear
structure to loads. In the present study, the value of forced frequency is attained by
applying load which is 500 N at top of the tooltip. From the Fig. 7, the procedure for
damping ratio (τ ) value was found using half-power bandwidth methods. The values
for damping ratio for different damping force direction and without damping force
is tabulated in Table 4.
Table 3 Overhanging length versus natural frequency versus damping force direction
S.No. Overhanging Axial force Vertical force Holding force Without
length (mm) with damper with damper with damper damper (Hz)
(Hz) (Hz) (Hz)
1 150 433.68 442.79 764.7 443.82
440.68 448.75 779.07 451.09
2597.2 2637.7 2446.6 2645.5
2603.6 2643.8 2465.4 2652.4
4080.8 4108.1 4390.6 4133.54
2 125 431.33 443.16 1126.3
438.82 449.99 1139.5
2631.7 2643.4 2450.5
2639 2650.3 2470.7
4067.1 4126.4 5914.8
3 100 431.33 581.3 1697.4
438.82 1670.3 1723.3
2631.7 3455.4 2448
2639 5776.3 2456.1
4067.1 6119.7 7523
Fig. 7 Harmonic analysis
In boring tool holder, finite element analysis was done in the overhanging position
of 150, 125 and 100 mm using ANSYS software. The tool deflection increases
with the increase in tool overhanging length observed from static analysis. Out of
Table 4 Overhanging length versus damping ratio versus damping force direction
S.No Overhanging Axial force Vertical forces Holding forces Without
length (mm) damping ratio damping ratio damping ratio damper
damping ratio
1 150 0.2 0.27 0.13 0.3
2 125 0.18 0.2 0.15
3 100 0.16 0.25 0.13
Table 5 Comparison of results with and without MR damper in holding direction

100 mm overhanging without MR 100 mm overhanging with MR
damper damper
Method Computational Experimental Computational Experimental
Deflection 0.3 mm 0.32 mm 0.13 mm 0.132 mm
Natural frequency 4440.54 Hz 4240 Hz 7523 Hz 7400 Hz
the different overhanging position, tool with 100 mm overhanging length was to
be the optimum, since this length has minimum tool deflection and also it satisfies
experimental criteria. Comparison of experimental and computational results for
boring tool holder with and without damping force is shown in Table 5.
As shown in Table 5, the boring tool holder with MR damper has less deflection
when compared to tool holder without damping effect. The system opposes the
movement of boring tool 50 mm from fixed end (150 mm), offers better damping
effect and suppresses tool vibration. When it is compared with damping force of
length 50 mm from fixed end, tool holder has no damping force. The natural frequency
increases when there is an increase in stiffness due to the damping force acting on the
holding direction which in turn reduces the mass acting on the tool holder. Also, it is
seen that the deflection and natural frequency obtained using computational results
matched well with the experimental results which were conducted on a Kirloskar turn
master lathe during hard boring with MR damper. During the experimental work,
30 V with direct current was used in the coil with cylinder consist of particle size
75 microns, fluid viscosity index is 40, with cone-shaped plunger and the damper
was positioned in holding direction as shown in Fig. 8. Based on the harmonic
analysis, it was found that a tool with MR damper in holding direction provide lesser
damping ratio when compared with other direction and tool without damper. From
the investigation, it has proved that a tool with magnetorheological fluid damper in
holding direction suppress the tool vibration and provide better cutting performance
which enrich the quality of the finished product.
Fig. 8 Photograph of
experimental setup
6 Conclusion
The present investigation provides a detailed explanation on the effect of tool holder
with various damping force direction and without damping force on tool vibration.
The effect of tool vibration on a boring process using Magnetorheological damper
by ANSYS software is studied. ANSYS software was used to model and analysed
the tool holder with insert and sim. The force constraints such as cutting force and
damping force were provided along with the fixed boundary conditions. The results
obtained from computational methods reasonably matched well with the experimen-
tal results. It is proven that using Magnetorheological fluid damper which is presently
holding direction reduces tool vibration and cutting performance can be improved
effectively.
References
1. Sam Paul P, Varadarajan AS, Ajay Vasanth X, Lawrance G (2014) Effect of magnetic field
on damping ability of magneto rheological damper during hard turning. Arch Civ Mech Eng
14:433–443
2. Quintana G, Ciurana J (2011) Chatter in machining processes: a review. Int J Mach Tools
Manuf 51:363–376
3. Sam Paul P, Varadarajan AS (2012) Effect of magneto rheological damper on tool vibration
during hard turning. Front Mech Eng 7:410–416
4. Chen CK, Tsao YM (2006) A stability analysis of regenerative chatter in turning process
without using tailstock. Int J Adv Manuf Technol 29:648–654
5. Chen CK, Tsao YM (2006) A stability analysis of turning a tailstock supported flexible work-
piece. Int J Mach Tools Manuf 46:18–25
6. Andren L, Hakansson L, Brandt A, Claesson I (2004) Identification of dynamic properties of
boring bar vibrations in a continuous boring operation. Mech Sys Sign Process 18:869–901
7. Moetakef-Imani B, Yussefian NZ (2009) Dynamic simulation of boring process. Int J Mach
Tools Manuf 49:1096–1103
8. Waghmare PJ, Patil RV, Waghmare GS (2015) A review on vibration mitigation of boring bar
using passive damping techniques. Int J Res Eng Technol 4:138–141
9. Wang M, Fei R (1999) Improvement of machining stability using a tunable-stiffness boring
bar containing an electrorheological fluid. Smart Mater Struct 8:511–514
10. Rabinow J (1951) Magnetic fluid torque and force transmitting device, United States Patent
US 2575360 A
11. Mei D, Kong T, Shih AJ, Chen Z (2009) Magneto rheological fluid-controlled boring bar for
chatter suppression. J Mater Process Technol 209:1861–1870
12. Paul PS, Varadarajan AS, Lawrance G, Sunil KS (2012) Analysis of turning tool holder with
Mr Fluid damper, international conference on modeling, optimization and computing. Procedia
Eng 38:2572–2578
13. Lawrance G, Paul PS, Varadarajan AS, Praveen AP, Vasanth XA (2017) Attenuation of vibration
in boring tool using spring controlled impact damper. Int J Interact Des Manuf 11:903–915
Comprehensive Review of the Effects
of Vibrations on Wind Turbine During
Energy Generation Operation, Its
Structural Challenges and Way Forward
I. P. Okokpujie, E. T. Akinlabi, N. E. Udoye, and K. Okokpujie
Abstract The effects of vibration cannot be overemphasized when it comes to gen-

erating energy via wind turbine. Vibration is one of the major challenges faced by the
wind turbine, due to the complexity of the structure and the area of installation. This
research work focuses on a compressive review of the effects of vibration occurrence
on wind turbine during energy generation operations and its economical challenges’.
Therefore, this research paper has reviewed various aspects of vibration effects in
horizontal wind turbine such as the blades region, tower structure, nacelles compart-
ment, and condition monitoring along with fault diagnosis models. The result from
this study has shown that, there are needs to develop and implement a good reliability
model, fatigue assessment process, and a well-developed monitoring model for wind
turbine during operation. When these things are properly put in place, it will help
to reduce unwanted vibration occurrence, eliminate unexpected failure of the wind
turbine in operations, and hence sustainable energy generation from wind turbine.
Keywords Vibration · Reliability for sustainability · Wind turbine blade · Wind

turbine tower
1 Introduction
Vibration will occur when a structure is dislodged from its location of stable equi-
librium. The structure will generally return to its stable position under the activity of
I. P. Okokpujie (B) · E. T. Akinlabi · N. E. Udoye

Department of Mechanical Engineering, Covenant University, Ota, Ogun State, Nigeria
E. T. Akinlabi
Kingsway Campus, Johannesburg 2006, South Africa
K. Okokpujie
Department of Electrical and Information Engineering, Covenant University, Ota, Ogun State,
Nigeria

https://doi.org/10.1007/978-981-15-4488-0_79
936 I..P. Okokpujie et al.
re-establishing powers. The towel holds the displacement of the forward and back-
ward movement of the wind turbine over the issue of its balance [1]. A structure
is combinations of element designed to act together to achieve a goal. A stationary
element is one whose operation makes its yielding time depends just on the contribu-
tion at a time while a dynamic element yield depends on previous information [2]. A
static and dynamic framework is two great types of the frameworks that consist on all
components and one powerful component, respectively [3]. The issues of vibration
in wind turbine can seriously affect the energy preservation; due to the fact that, the
vibration will help to expand the system which will lead to loss of energy generated
from the wind turbine during operations and after.
Wang et al. [4] recommended a damage identification and determination technique
for wind blades based on the unique elements property and mode shape distinction
arch data. It likewise gives a minimal effort and effective, nondestructive instrument
for wind turbine blade condition examination. Notwithstanding mix with modular
investigation strategies, the auxiliary strain field information likewise assumes an
essential job in identifying the harm existing in the breeze turbine system, including
the sharp edges and tower, on the grounds that the slight difference in basic physical
properties may prompt the rearrangement of basic strain. As a rule, the strain contrast
between nearby strain sensors can be considered as a list of the nearness of a split and
utilized as an attainable system to accomplish remote exhaustion harm recognition for
OWT towers. Nonetheless, the achievement predominantly relies upon the sensible
game plan of strain sensors around the pinnacle and the expectation of extraordinary
wind conditions [5, 6].
To discover the abilities of different tension sensors made with fiber Bragg grating
(FBG), optical backscatter reflectometer (OBR), and typical strain checks in harm
identification, an example acknowledgment system utilizing progressive nonlinear
essential segment investigation (h-NLPCA) in light of the strain field estimation was
depicted to recognize harm amid the affirmation trial of the blades [7, 8]. In light of
the trial results, the FBGs have more points of interest over the strain measures. FBGs
are reasonable for being inserted into the compound materials straightforwardly amid
the operational procedure of OWTs and observing progressively inside changes in
strain, which strain measures reinforced onto the surface cannot accomplish. What
is more, the FBGs have a more drawn out life in administration well-being observing
and harm discovery frameworks dependent on strain estimations for seaward wind
ranches. From there on, Downey contemplated a crossover thick sensor arranges
comprising of meager film sensors and FBGs and proposed an information-driven
harm recognition and limitation technique for wind turbine sharp edges. This pro-
cedure uses the blunder between the evaluated strain maps and measures strains to
characterize harm location which includes and applies a novel strategy for reviewing
vast quantities of sensors without the requirement for complex model-driven method-
ologies by intertwining sensor information into a solitary harm discovery highlight
[9]. The wind turbine also has a lot of electrical activities, which will need electrical
models to establish in the remote areas [10]. The challenge faced by researchers in
the field of wind turbine is very enormous, due to the fact that wind turbine develop-
ment and operations are very complex. There is great need to bridge the gap between
Comprehensive Review of the Effects of Vibrations … 937
researcher, teachers, and the industry in order to provide solutions to these challenges
[11–13].
However, this paper aimed at carrying out a compressive review on the effects
of vibrations occurrences during operation of wind turbine. This study will further
assist to eliminate and to reduce vibration occurrence on wind turbine.
2 Review of the Effects of Vibrations on Horizontal Wind

Turbine During Operations
Escaler and Mebarki [14] study vibration analysis on the wind turbine. The vibra-
tions were experimentally determined in precise positions applying the same type
of sensors for the period of six-month covering the whole range of working situa-
tions. The data obtained were preliminary certified to eliminate outliers based on the
hypothetical energy curves. The influential frequency points in the blade, gearbox,
and generator on vibrations were discovered and identified established on averaged
energy spectra. The amplitudes of the points induced by some environmental effects
were compared in various positions. There were observations on the broadband vibra-
tion on the wind speed reliance during this study. Finally, the authors detected a fault
case in the change of vibration enforce by a damage in the gearbox, as shown in
Fig. 1.
According to Hossain et al. [15], said that wind turbine is greatly subjected to
various fault on the mechanical system during operation which can be tired to the
issues of vibrations occurrences. Figure 2 shows the analysis of various failure rates in
wind turbine components, and Fig. 3 shows all the basic components of wind turbine
fully in operations. This analysis is in line with the observation from Okokpujie et al.
Fig. 1 Evaluation between the reference means vibration (RMS) amplitude on the gear mesh
frequency (GMF) and the harmonics (gearbox 3) showing the various levels of damage [14]
Fig. 2 Rate of failure of the various component of the wind turbine [17]
Fig. 3 Distinctive utility-scale wind turbine basic mechanisms [18]
[16] which in their design work shows that vibration is one of the major challenges
faced by a wind turbine, due to its mechanical structures.
2.1 Effects of Vibration on a Horizontal Axis Wind Turbine

Blades
A lot of researchers have carried out an experimental study on the causes of the
occurrence of sound on the wind turbine during operation. The authors concluded
that the flowing air passing through the blades (i.e., aerodynamic shape) has effects
on the turbine blades, which causes the vibration in the form of sound waves [19,
20]. One of the most significant aspects of the impact is the atmospheric turbulence
striking the blades of the wind turbine during operation (i.e., the influx of the sound
turbulence) and also the air rolling at the surface of the blades (trailing edge).
• Turbulence at the back or trailing edge of blades is created on the grounds that the
wind stream at the blades surface forms into a tempestuous layer. The recurrence
with the most noteworthy (capable of being heard) sound vitality content is gen-
erally in the scope of a couple of hundred Hz (hertz) up to around 1000–2000 Hz.
At the edge tips, conditions are to some degree diverse because of air streaming
towards the tip; however, this tip commotion is fundamentally the same as trailing
edge clamor and more often than not recognized as a significant separate source.
• Inflow disturbance is produced on the grounds that the blades slice through fierce
whirlpools that are available in the inflowing air (wind). This sound has a most
extreme sound dimension at around 10 Hz.
• Thickness sound outcomes from the relocation of air by moving blades and is
irrelevant for sound creation when the wind streams easily around the blades.
Notwithstanding, quick changes in powers on the blades result in sideways devel-
opments of the sharp edge and sound heartbeats in the infrasound district. This
prompts the normal wind turbine sound “signature” of sound dimension tops at
frequencies between around 1–10 Hz. These pinnacles cannot be heard, yet can
be found in estimations.
Tartibu et al. [19] presented a work of flap-wise, edgewise, and torsional basic
occurrences of variable length blades that have been acknowledged. Subsequently,
initiators can guarantee that characteristic vibration won’t be near the recurrence
of the fundamental excitation powers so as to maintain a strategic distance from
reverberation. The fixed part and moveable bit of the variable length of blades are
approximately a strong bar which can be slide in and out. Ten unique setups of the
variable length blades, speaking to ten distinct places of the moveable part, are exam-
ined. A MATLAB program was created to anticipate distinctive frequencies. Thus,
three-dimensional models of the variable length of blades have been created in the
limited component program Unigraphics NX5. Simultaneousness among MATLAB
and Unigraphics NX5 results has been found for the recurrence scope of intrigue.
This implies that a viable technique to figure regular frequencies of a variable length
of blades was generated.
The investigation of the impact of blades length on regular frequencies shown in
Fig. 4 has appeared with expanding blades length, the normal frequencies decrease.
This is most likely on the grounds that the blades turn out to be increasingly adaptable
as its length increases. The excitation loads are moved in the interim 0.5–30 Hz.
Kumar et al. [21] the primary goal of this examination research is to present another
material for wind turbine blades. Aluminum 2024 is chosen for the reasonableness
examination. Limited component method or finite element analysis is a strategy
utilized for the investigation of complex articles and geometries. Wind power or
wind vitality is considered as a spotless wellspring of vitality which delivers no
Fig. 4 Effects of vibration natural frequency on configurations of variable blade [19]
ecological damage amid activity. The complete wind control age capability of India
at a tallness of 50 m is 50,000 MW. As of late Indian Government has concentrated
on this inexhaustible well-spring of energy. The primary part of this exploration is to
recognize vibration frequencies and regular vibration methods of the Aluminum 2024
wind turbine blades. Wind turbine blades configuration is a perplexing technique.
For the plan of wind turbine blades, solid edge programming is utilized and the
model is imported in ANSYS 14.0 for modular investigation. For the reasonableness
examination of Aluminum 2024, the basic and modal investigation has been done.
The significances of the investigation are utilized to confirm a structure’s readiness
for use.
The examination results were confirmed with experimental result accessible in
writing. With the expectation of free vibration examination of Aluminum 2024, wind
turbine blades modular investigation utilizing ANSYS 14.0 was directed. Through
our examination, we have discovered the disfigurements, stresses, and normal fre-
quencies for initial six modes state of Aluminum 2024 wind turbine blades as pre-
sented in Fig. 5a. The examination results were checked with an experimental result
that has been done. Andrew [22] examined the on and seaward wind turbines. As
indicated by the author’s test examination, the most extreme happens at tip as shown
in Fig. 5b, and the loads are extremely less for Aluminum lightweight materials.
Taware et al. [23] composite materials have different confounded attributes as
indicated by the utilization of the constituent materials and limited conditions. In
this manner, it is hard to examine the properties of composite materials, and the
composite materials are outstanding by their superb blend of high auxiliary firmness
and low weight. Their anisotropy enables the designer to tailor the material so as
to accomplish the ideal execution necessities. Accordingly, it is important to create
devices that enable the architect to get plans, thinking about the auxiliary prereq-
uisites and utilitarian attributes, for effective utilization of composite materials in
building applications, the dynamic conduct (e.g., normal frequencies) ought to be
known. This paper centered around the conduct of little wind turbine blade fabri-
cated from composite materials. Two little wind turbine blades are produced from
the glass fiber reinforced plastic (GFRP) and GFRP with steel wire work support.
Fig. 5 a Wind turbine illustration b vibration mode v/s expected frequency of the AL 2024 alloy
blade used for the development of the wind turbine [22]
Finite element analysis (FEA) was completed by utilizing the component program-
ming ANSYS 16.0. From FEA, hypothetical normal frequencies and mode shapes of
blades made from the GFRP and GFRP with steel wire work support were acquired.
Test free vibration trial of produced blades was done to locate the useful regular
vibration frequencies and mode shapes at last, the outcomes acquired from the FEA
and exploratory test for the blade fabricated from GFRP and GFRP with steel wire
work were carried out, and the result shows that the GFRP with steel wire has great
performance when compared with the GFRP.
2.2 Effects of Vibration on a Horizontal Wind Turbine Tower
The horizontal wind turbine tower shoulders an essential role in supporting the
energy-producing part, for example, blades, hubs, and nacelles. The tower opposes
the wind loads produced by the rotational wind blades [23] and hence, to verify the
auxiliary security and dependability of the wind on the turbine tower, which can be
harmed by unanticipated solid winds, through basic wellbeing checking. For this
situation, Kim et al. [24] contemplate the dynamic attributes of a wind turbine tower
display with and without harm at different areas were examined. However, indoor
trials with the end goal of damages evaluation and auxiliary well-being observing the
test result were contrasted and those acquired utilizing limited component demon-
strate as appeared in Fig. 6 which demonstrates the variety of the recurrence and
the conceivable damages. The authors likewise proposed a fitting damage identifica-
tion system for the wind turbine tower, recurrence base damage location strategies
were connected, and the outcome was checked by utilizing the numerical model of
a 3 MW airstream turbine tower working in Jeju Island, Korea. The comparison of
the indoor experiment and the finite element (FE) model are shown in Fig. 7. The
Fig. 6 The analysis of the natural frequency variations with the damage situations of mode 1 and
2 [24]
Fig. 7 Analysis of the first three shapes of the mode a mode 1, b mode 2, c mode 3 for the FE and
FDD [24]
frequency-domain decomposition (FDD) technique was used for the identification

of the natural frequency, damping ratio, and shape mode, which are the parameters
of the modal; the result gotten from this technique were also analyzed and compared
with FE model, and Table 1 gives the analysis of the first model.
Figure 8 shows a wind turbine tower that was greatly affected by the vibration
during operation, as a result of unwanted vibration occurrences on the earth cross,
coupled with other fault that was not detected before the failure of the wind turbine
occurs.
Table 1 FE model and FDD comparisons result [24]

Wind environment Model FE FDD Percentage difference (%)
Light 1 2.7332 2.8224 3.2
676 (rpm) 2 54.669 40.103 26.6
3 170.785 100.185 41.3
Fig. 8 Failed wind turbine

tower during operation
Guo and Infield [25] carried out the examination of the early disappointment of key
segments, for example, the tower, drive train, and rotor of a substantial wind turbine.
In their work, the authors used the nonlinear state estimation technique (NSET) to
determine the turbine tower vibration impact, giving a comprehension of the tower
vibration dynamic attributes, and the fundamental elements affecting it. The created
tower vibration demonstration involves two unique aspects: (1) a sub-display utilized
for underneath evaluated wind speed and (2) for above-assessed wind speed as shown
in Fig. 9. Supervisory control and data acquisition system (SCADA) applied the data
obtained from a wind turbine from March to April 2006 to develop the model, and
the model was validated with series of experimental analysis. The NSET method has
been effectively demonstrated in this research work, which proves to be accurate and
very efficient for tower vibration analysis. The tower vibration model developed was
further used for spotting out blades angle irregularity, which is a common error that
needed to be fixed in other to hence the performance of the wind turbine and help
to eliminate or reduce fatigue damages. This study has demonstrated that condition
checking will be improved greatly, if data from vibration signals is accompanied
by the investigation of other applicable SCADA information, for example, energy
performance analysis, rotor loads, and the wind speed.
Fig. 9 NSET modeling and

prediction for tower
vibration [25]
Tibaldi et al. [26] used the aeroelastic simulation process to examined wind turbine
vibration base on both frequency and time domain. Three different aspects were
investigated such as there are need for a defined model in the operation of a close
echoing condition; the presence of resonances identification and load estimation at
low turbulence intensity during the wind turbine operation is very significant; and
the external excitation response of the wind turbine. In the first phase of the analysis,
the frequency and the damping of the aeroelastic modes were analyzed with three
wind turbine models. Two different turbulent intensities were used to investigate the
same models on fatigue loads to analyze the response from the wind turbine.
The second phase, to determine the modes that can be excited, an external force is
introduced to the wind turbine model, and therefore, the minimal excitation frequency
will be presented during operation The study shows that substantial sideways blade
vibrations can happen on modern wind turbines even if the aeroelastic damping of
the sideways modes is optimistic. When working close to echoing environments,
slight changes in the modeling can take a large nuance on the vibration level. The
sideways vibrations are less noticeable in great turbulent environments. Applying
simulations with low-level turbulence concentration will hence the performance to
avoid a redesign. Additionally, this is subjected to the external forces. The analysis
is done using aeroelastic models equivalent to a 1.5 MW class wind turbine with
insignificant differences in blade materials.
Pacheco et al. [27] work on a different amplitudes of sound time series through
simulation of noise adding band-limit to the time series acceleration measured from
the original experiment, and Fig. 10a, c presents the various wind speed, spectra
time series of the acceleration connected at the tower tip, that is already polluted
with sound. From the study at low wind speed, the adopted sound levels are higher
to measure the accelerations; however, when the wind speed gets to 2.5 m/s, the
acceleration goes beyond the first two sound levels. The acceleration signal is used
to overcome the sound levels for a high wind speed.
Fig. 10 Energy spectra of distinctive acceleration time series obtained at the tower tip for varying
wind speeds: the real signal = Acc and original signal polluted with various sound levels = Acc +
Noise1, 2, 3. For wind speeds of 2.5 and 17.5 m/s, the spectra connector with “Acc” and “Acc +
Noise1” is practically equivalent
3 Structural Challenges Face During Operation

of the Wind Turbine
The issues of reliability during wind turbine operations: The location of the 80% of
wind turbine is in a remote area, which has great effects on the rate of maintenance
that can be carried out. This will significantly affect the issue of reliability, because
reliability is very significant in every machine that is design to rotate during operations
(Fig. 11).
From this review study, it can be observed that over 80–90% of failure of the
wind turbine (WT) is caused by vibration; due to the location area that most WTs
are installed, this may be as a result from continuous contact of two metal objects
during rotation. This continuous contact will affect the life circle of the turbine in
operation [28].
The offshore wind turbine structure is always faced by various cyclic excitations,
which includes the rotor operation, wind, currents, and waves, which can lead to
high risks of damage, failure, and fatigue to the structures. This is one of the major
challenges of the wind turbine on the offshore cite; however, the study of fatigue
assessment needs to be strategically implemented in all operation of the offshore
wind turbine; from several authors, wind turbine needs a well-developed standard
model that can help control and monitor the wind turbine during operation, that will
enable operator to detect any occurrences that will lead to vibration [29–32]. These
continuous challenges of vibration occurrence in wind energy generation will affect
the economic structure of the wind turbine development and its implementations,
which will lead to high rate of manufacturing and maintainers cost.
Fig. 11 a Design of the planetary gear in the nacelles compartment. b The failure of the wind
turbine (WT) during operation
4 Way Forward for Eliminating Vibration Frequency

from a Wind Turbine During Operations
• One of the major things that must be done for a sustainable power generation from
a wind turbine is to develop a standard model that can serve as a guild for wind
turbine operations
• There is a need to carry out research on how to establish a good reliability and
fatigue assessment process for wind turbine operations
• The basic areas for installations of wind turbine should be put into consideration
during the design, in order to avoid the occurrence of vibrations which will lead
to great loss of the wind turbine
• Material selection is one of the major factors in developing wind turbine, in the
area of the blades, tower, and nacelles compartment. Most wind turbine blade is
made with glass fiber; from this study, the authors observed that there are need to
manufacture new materials which will serve as a better alternative to glass fiber,
for example, the composite material of aluminum alloy mixed with glass fiber
and a little percentage of titanium, for strength, flexibility, and firmness of the
materials.
When vibration in the wind turbine is greatly reduced to the minimum, it will help
to reduce the cost of maintainers during operation and after operations, which will
serve as economic benefit to the manufacturers, industrial company (that makes use
of the wind turbine for commercial purposes) and the nations at large.
5 Conclusions
Wind energy generation is one of the major and reliable sources of electricity gen-
eration, that is, environmental friendly. Through the application of wind turbine in
energy generation, a lot of countries and continents have employed it which has a
major source of supply in their home and official places in the world today; however,
this promising system is currently facing a lot of challenges of vibration occurrence
due to its complex nature. This paper has presented a comprehensive review on the
effects of vibration on the wind turbine blades, tower, and nacelles compartment and
also highlighted some major factors that cause the issues of the unwanted vibration
occurrences and the possible ways to eliminate or reduce vibration during operation
of a wind turbine.
Acknowledgements The authors deeply appreciate the management of Covenant University for
their sponsorship.
References
1. Li Li (2012) Vibrations analysis of vertical axis wind turbine: a thesis presented in partial
fulfilment of the requirements for the degree of master of engineering in mechatronics, school
of engineering and advanced technology, Massey University, New Zealand. Diss, Massey
University
2. Ajayi OO, Agarana MC, Akinwumi II, Okokpujie PI, Salawu EY, Abioye AA, Afolalu AS,
Leramo RO (2019) Modelling rate of traffic-induced building vibrations in Sango-Ota, Nigeria:
an assumption based analysis. WIT Trans Built Environ 182:303–313
3. Nwoke ON, Okonkwo UC, Okafor CE, Okokpujie IP (2017) Evaluation of chatter vibration
frequency InCnc turning of 4340 alloy steel material. Int J Sci Eng Res 8(2):487–495
4. Wang X, Zeng X, Li J, Yang X, Wang H (2018) A review on recent advancements of
substructures for offshore wind turbines. Energy Convers Manag 15(158):103–119
5. Lian J, Cai O, Dong X, Jiang Q, Zhao Y (2019) Health monitoring and safety evaluation of the
offshore wind turbine structure: a review and discussion of future development. Sustainability
11(2):494
6. Okokpujie K, Shobayo O, Noma-Osaghae E, Okokpujie I, Okoyeigbo O (2018) Realization
of MPLS-based VPN network for improved Qos metrics. TELKOMNIKA Telecom Comput
Electron Control 16
7. Luo H (2017) Physics-based data analysis for wind turbine condition monitoring. Clean Energy
1(1):4–22
8. Okokpujie KO, Chukwu EC, Noma-Osaghae E, Okokpujie IP (2018) Novel active queue man-
agement scheme for routers in wireless networks. Int J Commun Antenna Propag (I. Re. CAP)
8(1):53–61
9. Qiao W, Lu D (2015) A survey on wind turbine condition monitoring and fault diagnosis—part
II: signals and signal processing methods. IEEE Trans Indus Electron 62(10):6546–6557
10. Okokpujie KO, Chukwu EC, Noma-Osaghae E, Odusanmi M, Okokpujie IP (2018) A unique
mathematical queuing model for wired and wireless networks. Int J Civil Eng Technol (IJCIET)
9(9):810–831
11. Okokpujie IP, Fayomi OSI, Leramo RO (2018) The role of research in economic development.
IOP Conf Ser Mater Sci Eng 413(1):012060. IOP Publishing
12. Fayomi OSI, Okokpujie IP, Fayom GU, Okolie ST (2019) The challenge of Nigeria researcher
in meeting up with sustainable development goal in 21st century. Energy Procedia 157:393–404
13. Okokpujie IP, Fayomi OSI, Ogbonnaya SK, Fayomi GU (2019) The wide margin between the
academic and researcher in a new age university for sustainable development. Energy Procedia
157:862–870
14. Escaler X, Mebarki T (2018) Full-scale wind turbine vibration signature analysis. Machines
6(4):63. (1–16)
15. Hossain Md, Abu-Siada A, Muyeen S (2018) Methods for advanced wind turbine condition
monitoring and early diagnosis: a literature review. Energies 11(5):1309
16. Okokpujie IP, Okokpujie KO, Nwoke ON, Azeta J (2018) Development of a 0.5 KW horizontal
axis wind turbine. J Eng Appl Sci 13(8):2202–2208
17. Hahn B, Durstewitz M, Rohrig K (2007) Reliability of wind turbines. Wind energy. Springer,
Berlin, pp 329–332
18. Odia OO, Ososomi SA, Okokpujie PI (2016) Practical analysis of a small wind turbine for
domestic use on latitude 7.0670 N, longitude 6.2670 E. J Res Mech Eng 2(11):8–10
19. Okokpujie IP, Okokpujie KO, Nwoke ON, Azeta J (2018) Design and implementation of
0.5 kw horizontal axis wind turbine for domestic use. In: Proceedings of the world congress
on engineering, vol 2
20. Tartibu LK, Kilfoil M, van der Merwe AJ (2012) Vibration analysis of a variable length blade
wind turbine 630–639
21. Kumar A, Dwivedi A, Paliwal V, Patil PP (2024) Free vibration analysis of Al, wind turbine
blade designed for Uttarakhand region based on FEA. Proce Technol 14(2014):336–347
22. Andrew H (1998) Dynamic analysis of both on and offshore wind turbines in the frequency
domain. A dissertation submitted for the degree of Doctor of Philosophy in Faculty of
Engineering, University of London
23. Taware GB, Mankar SH, Ghagare VB, Bharambe GP, Kale SA (2016) Vibration analysis of a
small wind turbine blade. Int J Eng Technol (IJET). 8(5):2121–2126
24. Kim W, Yi J, Kim J, Park J (2017) Vibration-based structural health assessment of a wind
turbine tower using a wind turbine model. Proce Eng 188:333–339
25. Guo P, Infield D (2012) Wind turbine tower vibration modeling and monitoring by the nonlinear
state estimation technique (NSET). Energies 5(12):5279–5293
26. Tibaldi C, Kim T, Larsen TJ, Rasmussen F, de Rocca Serra R, Sanz F (2016) An investigation
on wind turbine resonant vibrations. Wind Energy 19(5):847–859
27. Pacheco J, Oliveira G, Magalhães F, Cunha Á, Caetano E (2017) Wind turbine vibration based
SHM system: influence of the sensors layout and noise. Proce Eng 199:2160–2165
28. Dunmade I, Udo M, Akintayo T, Oyedepo S, Okokpujie IP (2018) Lifecycle impact assessment
of an engineering project management process–a SLCA approach. In: IOP conference series:
materials science and engineering, vol 413, no 1, p 012061. IOP Publishing, 2018
29. Liu X, Lu C, Liang S, Godbole A, Chen Y (2015) Influence of the vibration of large-scale wind
turbine blade on the aerodynamic load. Energy Proced 75:873–879
30. Okokpujie IP, Akinlabi ET, Okonkwo UC, Babaremu KO, Okokpujie KO (2019) Experimental
evaluation, modeling and optimaztion of a 500 W horizontal wind turbine using definitive screen
design method for sustainable wind power generation. Int J Civil Eng Technol 10(1):2415–2431
31. Okokpujie KO, Odusami M, Okokpujie IP, Abayomi-Alli O (2017) A model for automatic
control of home appliances using DTMF technique. Int J Sci Eng Res 8(1):266–272
32. Tchakoua P, Wamkeue R, Tameghe TA, Ekemb G (2013) A review of concepts and methods
for wind turbines condition monitoring. In: 2013 world congress on computer and information
technology (WCCIT) Jun 22 2013 (pp 1–9). IEEE
Design and Analysis of a Novel Hybrid
Car Bumper Using Non-Newtonian Fluid
and High-Density Polyethylene
K. Praveen Jerish, J. Rakesh Kumar, R. Ramakrishnan,

and K. S. Vijay Sekar
Abstract The drivers of today’s cars are protected from frontal impacts with the help
of crumple zones and composite bumpers that absorb the entire impact. However, the
structural integrity is not entirely protected. In this research, a hybrid of two materials,
one a non-Newtonian fluid named Oobleck and another material HDPE, was used in
conjunction with creating a hybrid bumper system. The Oobleck fluid exhibits shear
thickening when an impact force is applied and the material changes from a liquid
to a solid state, thereby demonstrating twin characteristics. This unique property of
the fluid is exploited by housing it in a HDPE tube and fixing the HDPE tube within
the bumper beam. Such a combination has not been used in previous works and
is a novelty. This deals with the assorted mixture properties of fluid, the materials
for housing this fluid and holding it in the site and the impact strength absorbing
capabilities of the designed bumper. A comparative analysis was undertaken between
the prevailing bumper system and the newly designed bumper with non-Newtonian
fluid. The finite element results that justify deformation experienced by the modified
bumper are less compared to existing bumper, thereby resulting in less structural
damage or formation of crumpled zones.
Keywords Non-Newtonian fluid · Oobleck · Shear thickening
1 Introduction
Automotive bumper beam is an important part of the modern car, which ensures
safety to the occupants and also plays a key role in minimizing the frontal impact
on the cars [1]. The bumper is a structure integrated to the front and rear end of
a car, whose primary aim is to absorb impacts in minor collisions and if possible
self-destruct in major collisions, thereby ensuring the safety of its passengers [2].
Bumpers minimize height problems between cars and other vehicles. In the early
K. Praveen Jerish · J. Rakesh Kumar (B) · R. Ramakrishnan · K. S. Vijay Sekar

Kalavakkam 603110, India

https://doi.org/10.1007/978-981-15-4488-0_80
950 K. Praveen Jerish et al.
stages of the automotive industry, the front spring hanger bolts were replaced with
longer ones to be able to attach an ingot. Later the primary bumper manufactured
from rubber was designed to soak up impacts [3]. General Motors incorporated body-
coloured plastic front bumpers designed to absorb low-speed impacts. In fashionable
cars, the bumpers comprise of a plastic cowl over a reinforcement bar manufactured
from steel, aluminium, composite, or plastic. Today the modern cars’ bumpers are
manufactured from a mix of polycarbonate/ABS. Antecedent bumper system consists
of many components that are manufactured from typical material like steel [4]. The
integrity between components in bumper beam system is crucial in the overall bumper
beam performance, which is affected by three factors: ease of manufacturability, ease
of assembly and intrinsic value. [5]. In recent times, a solitary bumper beam was used
to scale back price. Also, hybrids using a combination of plastic and metal are in
vogue, to leverage the benefits of low weight, less corrosion, high specific stiffness,
high specific strength and high impact energy absorption capability [6].
The objective of this work is to herald the employment of non-Newtonian fluid
within tubes of plastics like HDPE and PP. Since the non-Newtonian fluid does not
follow the law of viciousness, i.e. the viciousness will amend to either an addi-
tional liquid or an additional solid phase. The non-Newtonian fluid employed is
named Oobleck, a shear thickening fluid, which is corn starch dissolved in water,
which if stirred slowly shows milk-like properties, while when stirred hardly, appears
like a viscous liquid. This combination of Oobleck and HDPE gas been structurally
analysed using modelling and finite element analysis.
2 Design of the Hybrid Bumper
The hybrid bumper design was initiated by taking an existing bumper model of a
car into consideration. The 3D model of the bumper was designed using SolidWorks
software. Two bumper designs were considered. Figure 1 shows the location of the
bumper beam in the existing vehicle. Figure 2 shows the hybrid bumper design-1.
The bumper beam was made hollow instead of the regular rigid bumper. This was
made to house the non-Newtonian fluid within the bumper beam. This provision
helps in consolidating all the components into one unit to provide more strength.
The entire rear section of the bumper was initially sealed, and the HDPE tube to the
entire length of the bumper was planned to be housed inside.
However, the weight was raised by 2.4 kg since weight plays an important role in
a car’s mileage and performance, a change in design was initiated. Instead of sealing
the entire rear portion, it was decided to seal the end portions alone. Moreover, the
length of the HDPE tube was shortened to be accommodated at the centre of the
bumper beam element. Figure 3 shows the modified bumper design.
Design and Analysis of a Novel Hybrid Car Bumper … 951
Fig. 1 Location of bumper beam in the existing vehicle
End cap
Polypropylene tube
MS plate
Bumper beam Oobleck
Fig. 2 Hybrid bumper design 1 (exploded view)
3 Selection of Housing Material
Since Oobleck is in the liquid phase before the application of any force on it, it is
necessary for a tubular structure to house the fluid. To accommodate the housing of
the non-Newtonian fluid, a thermoplastic material (because plastics do not absorb
water) was used. Some of the commonly available thermoplastics were low-density
polyethylene (LDPE), high-density polyethylene (HDPE) and polypropylene (PP).
Initially, the LDPE and PP tubes were considered for housing the fluid; but after
carrying out the comparison over the properties of the other thermoplastics, it was
found that both the materials lacked flexibility and tended to crack under an impact
force. Therefore, HDPE was selected and on analysis was found suitable for the
HDPE tube
Bumper beam
Mounting
Oobleck MS plate
Fig. 3 Hybrid bumper design 2 (exploded view)
application. It was flexible to a certain extent and was able to produce a cushioning
effect while facing an impact force. The quantity of fluid required to be filled was
made as per the dimensions of the HDPE tube: a HDPE tube of length 0.8 m and
inner diameter of 4.4 cm with a wall thickness of 0.3 cm (the tube dimensions
were taken based on the bumper beam of Maruti Suzuki Alto 800). The quantity of
Oobleck required for this dimension was obtained as 1441.34 mg which is very much
negligible that there is no significant increase in the weight of the bumper. So the
addition of shear thickening fluid does not influence the performance and mileage
characteristics of the vehicle.
4 Bumper Assembly
The end of the HDPE tubes was sealed with HDPE caps using induction heating
method. The circumference of the pipe as well as the cap was heated by placing
them on the coil and fused together. Before sealing the other end, 2 kg of corn starch
was mixed with 500 ml of water to get the required Oobleck, which was filled inside
the tube. The HDPE tube with the Oobleck fluid was housed along with the bumper
beam, by placing the tube at the centre of the bumper and then welding using plates
at each end of the bumper beam.
5 Finite Element Analysis
An explicit finite element analysis was carried out to study the dynamics of the
bumper, using Ansys software. For the analysis, a static wall made of concrete mate-
rials was considered. The concrete wall was used as a fixed support, and the moving
bumper was enabled to impact the stationary concrete slab. Figure 4 shows the fixed
concrete wall, and Fig. 5 shows the bumper impacting the wall.
Figure 6 shows the bumper before collision or impact with the concrete wall.
Figure 7 shows the bumper after collision with the concrete wall. Table 1 shows
the material properties of the concrete wall, bumper beam and the HDPE tube. The
assembly was meshed with 13,791 elements, 26,609 nodes with an element size of
0.05 mm (Table 2).
Fig. 4 Static concrete wall
Fig. 5 Bumper impacting the wall

Fig. 6 Static bumper
Fig. 7 Bumper deformation after impacting the wall


Material Density Young’s Poisson’s Bulk Shear Specific
(kg/m3 ) modulus ratio modulus modulus heat
(GPa) (GPa) (GPa) (J kg−1
C−1 )
Concrete(wall) 2300 30 0.18 15.625 12.712 780
Mild steel 7850 205 0.3 170.8 78.84 434
(Bumper
beam)
HDPE tube 950 11 0.42 22.91 3.87 2300
Table 2 Fluid properties

Material Density (kg/m3 ) Kinematic viscosity Dynamic viscosity
(m2 /s) (Ns/m2 )
Shear thickening fluid 1184.9 15.556 18,444.138
(Oobleck)
To obtain a comparable study group, the collision conditions were restricted by

choosing the speed of the vehicle as 40 kmph, which is based on the average speed
in city traffic. The explicit dynamic analysis of the conventional bumper and the
newly designed hybrid bumper were carried out. The total deformation plot of the
conventional bumper (without HDPE tube and non-Newtonian fluid) for an impact
speed of 40 kmph produced a net deformation of 20.853 × 10−3 (mm) as seen in Fig. 8.
Figure 9 shows the total deformation of the designed bumper (with HDPE tube and
non-Newtonian fluid) for an impact velocity of 40 kmph with a net deformation value
of 9.97 × 10−3 (mm).
Fig. 8 Deformation of conventional bumper

Fig. 9 Deformation of hybrid bumper
6.1 Deformation Versus Impact Force and Stress
From the deformation versus impact force graph (Fig. 10), it can be seen that the defor-
mation for a given impact force is larger in the bumper without the non-Newtonian
fluid and HDPE than the bumper with non-Newtonian fluid and HDPE tube. For
a maximum load of 20,000 N, the corresponding deformation is 2.6 mm (without
HDPE) and 1.7 mm (with HDPE). From the stress vs impact force graph (Fig. 11),
it can be seen that the stress induced for a given impact force is larger in the bumper
without the non-Newtonian fluid and HDPE than the bumper with non-Newtonian
fluid and HDPE tube. For a maximum load of 20,000 N, the corresponding stress is
795 MPa (without HDPE) and 700 MPa (with HDPE).
From the above results and discussions, it can be clearly seen that the hybrid
bumper with the HDPE tube filled with non-Newtonian fluid shows better impact
Fig. 10 Deformation versus

impact force
Fig. 11 Stress versus impact

force
resistance characteristics with lower deformation and stress than the conventional
bumper design.
7 Conclusion
The results reveal that non-Newtonian fluid is very advantageous; due to the reason
that the deformation experienced by the modified hybrid bumper is less compared
to the existing bumper, thereby resulting in less structural damage or formation of
crumpled zones. The involvement of non-Newtonian fluid in the safety segment helps
in the absorption of the impact and can be used in conjunction with existing bumpers
to augment the impact absorption strength of the bumper. The non-Newtonian fluid
acts as a damping material and when used with a material like HDPE, it enhances
the total impact strength of the entire bumper system.
References
1. Johnson W, Walton AC (1983) An experimental investigation of the energy dissipation of a

number of car bumpers under quasi-static lateral loads. Int J Impact Eng 1(3):301–308
2. Cheon SS, Choi JH, Lee DG (1995) Development of the composite bumper beam for passenger
cars. Compos Struct 32:491–499
3. Witteman WJ (1999) Improved vehicle crashworthiness design by control of the energy absorp-
tion for different collision situations Eindhoven: Technische Universiteit Eindhoven, Ph.D.
thesis
4. Hosseinzadeh R, Shokrieh MM, Lessard LB (2005) Parametric study of automotive composite
bumper beams subjected to low-velocity impacts. Compos Struct 68(4):419–427
5. Teng T-L, Ngo V-L, Nguyen T-H (2010) Design of pedestrian friendly vehicle bumper. J Mech
Sci Technol 24(10):2067–2073
6. Shyi DY (1989) Structure of multi-protection safety car bumper. US Patents US4807915A
Vibration Signature-Based Monitoring
on FSW Process and Verification by FEA
Akshay Todakar, R. S. Nakandhrakumar, M. Ramakrishnan,

and V. Meenakshisundaram
Abstract Experimental study conducted during joining butt weld in FSW process
on Al 6061 alloy of size 50 mm width × 100 mm length × 8 mm thickness of two
plates wherein the effect of the interaction between the plates, tool and the vibration
that occurs during the process is investigated and reported. In this study, joining sides
of the workpiece samples are artificially induced with air gaps of drilled holes in 2,
3, 5 mm diameter holes and 3 mm width × 4 mm depth of slots in random distances.
The vibration behaviour of the tool and workpiece joining system is characterized
by frequencies arrived in modal analysis using finite element analysis (FEA), and
each mode corresponds to tool and workpiece system. Variations in the amplitudes
of vibration signals in the particular range of frequencies from 6.0 to 7.0 kHz are
proportional to the workpiece and significant changes in a linear pattern indicate
the defective and steady joining area of workpiece. So, this method is effective
in monitoring of workpiece joining in FSW process. The fast Fourier transform
(FFT) analysis of the vibration signal shows the changes in individual frequencies
and is used for identifying the frequency range of monitoring workpiece with gap
and without gap conditions. The steady joining portions cause the vibration which
corresponds to 4th region frequencies of workpiece.
Keywords Friction stir welding process · Vibration · Frequency domain analysis ·

FEA
A. Todakar (B)
Department of Mechanical Engineering, Hindustan Institute of Technology and Science, Chennai
603103, India
R. S. Nakandhrakumar
Centre for Simulation and Engineering Design (SIMENDES), Hindustan Institute of Technology
and Science, Chennai 603103, India
M. Ramakrishnan · V. Meenakshisundaram
School of Aeronautical Sciences, Hindustan Institute of Technology and Science, Chennai
603103, India

https://doi.org/10.1007/978-981-15-4488-0_81
960 A. Todakar et al.
1 Introduction
The friction stir welding process is a solid-state combined that uses a non-expendable
tool to link two non-melting materials. This method can progress the mechanical
properties of the joint, such as the strength and hardness. The heat will be created
due to friction and plastic distortion between the tool and the workpieces. This friction
and plastic distortion result in the mixing and agitation of the materials around the
pin from the front to the rear. The heat generated by friction leads to the softening of
metals, especially near the friction welding tool. This means that mechanical energy
is converted into thermal energy in the contact areas, without the need for heat from
other sources. The main function of the friction welding tool is to heat the parts, and
then to induce the materials to flow and restrict under the shoulder and Impression
action will generate friction between two surfaces and relative motion between two
parts. People are studying to optimize process parameters for active connection of
materials [2].
2 Literature
Premature failure of the welding tool can lead to unacceptable welding joint quality
and loss of welding productivity. Friction welding is a completely machined process.
The forces and vibration generated by the process are high enough that manual
operation is not possible, except possibly for very fine materials. Therefore, online
monitoring of vibration is, therefore, in demand.
Lambiase et al. [1] have investigated force variation, temperature and torque dis-
tribution in process with an al-Si–Mg aluminium alloy that varies the tool’s rotation
speed and welding speed. Temperature measurements were made using an IR cam-
era. Prasanna et al. [4] have observed the experimental and numerical evaluation
with aluminium alloy AA6061. Temperature variation and simulation model are
tested parameters with experimental results. Buffa et al. [3] have developed the dis-
tribution of temperature, and tension in welding nugget was investigated. Projected
the relationships between the forces of the tool and the variation in the parameters.
Temperature profile almost symmetrical in the welding area was found. Reza-E-
Rabbya [5] has found pin characteristics in the flow of material and the weldability
by stirring by the friction of two aluminium alloys (AA 7050 and AA6061) with a
pin tool cylinder, including the pin smooth/without thread attached to a geometry
of shoulder displacement single invariant. Welds were made under a range of pro-
cess parameters (welding and rotation speed). Sadeesh and Kannan [6] conducted
with plates of aluminium AA2024 and AA6061 dissimilar, and obtained the optimal
parameters of the process. Different tool designs have been used to analyse the prop-
erties. Investigated the effect of welding speed on the microstructure hardness and
tensile properties of the welded joints. As the process parameters varied, seamless,
high-efficiency welded joints were produced. Shukla and Raghu [7] have observed
Vibration Signature-Based Monitoring on FSW Process … 961
mechanical and metallurgical properties by changing various parameters that FSW

can be used to study the parameters on the process in laboratory. Experiments have
been conducted to validate some of the simulation results of the ANSYS software.
Ramesh et al. [8] have investigated (FSW) the aluminium alloy 6082 to study the
tensile strength and hardness by changing the process parameter, speed of rotation
and speed of welding of tool and different weld condition. They found the effect of
tool design on mechanical properties in FSW of AA6061.
In this paper, modal analysis of workpieces and hard tool using finite element
analysis (FEA) is compared with the range of frequency occurring during experiments
in the fast Fourier transform (FFT) analysis of vibration signals. The fact of this
approach is presented from friction stir welding (FSW) experiments using hard tool
and an aluminium alloy workpiece.
3.1 Experimental Work
The workpiece materials used for experimental study are commercially available Al
6061. The size of the samples of workpiece plates is 100 mm length × 50 mm width
× 8 mm thickness. The chemical compositions and material properties are tabulated
in Tables 1 and 2 [9]. The two plates were mounted using a clamp and retrofitted on a
CNC milling machine which is shown in Fig. 1 (photograph of the experimental set-
up used to study friction stir process). Tool pin was made with D/d ratio (shoulder/pin)
Table 1 Chemical
Composition Weight %
composition of AA6061
Fe 0.244
Si 0.741
Mn 0.095
Cu 0.157
Ni 0.01
Cr 0.125
Ti 0.007
Sn 0.006
V 0.006
Co 0.002
Zn 0.020
Pb 0.005
Mg 0.901
Al 97.699
Table 2 Mechanical properties of AA6061

Yield strength Ultimate tensile Elongation % Shear strength Fracture strength
(MPa) strength (MPa) (MPa) (MPa)
276 310 12 207 94
Fig. 1 Photograph of experimental set-up with vibration monitoring system i accelerometer, ii data
acquisition card, iii connecting cable and iv Laptop
of 3 in which tool shoulder and pin are made up of diameter 16 mm × height 60 mm

and pin diameter of 6 mm × height 7.5 mm, respectively. The chemical composition
and material properties are given in Tables 3 and 4. Joining of workpiece samples
was performed on a 3-axis CNC milling machine which has 3-axis movement and
Table 3 Chemical
Composition Weight %
composition of H13
Cr 4.75–5.5
Mo 1.1–1.75
Si 0.80–1.2
V 0.80–1.2
C 0.32–0.45
Ni 0.3
Cu 0.25
Mn 0.2–0.5
P 0.03
S 0.03
Table 4 Properties of H13

Density 7800 kg/m3
Melting point 1427 °C
Tensile strength (ultimate) 1.2–1.5 GPa
Tensile strength (yield) 1–1.3 GPa
Modulus of elasticity 215 GPa
Thermal conductivity 28.6 W/mK
carried using a special CNC program which was run at speed 800 rpm and feed
33 mm/min [Ref]. Both workpieces are properly secured using a clamp. The vibration
signals were measured using a (Kistler model-8702B50) accelerometer sensor which
is positioned on the clamp and used for holding dynamometer. A data acquisition card
(NI 9133) is used to convert analog output signals into digital signals. Among four
of its channels, single analog input channel was used to collect the data by sampling
the signals at 25 kHz and interfaced with a personal computer and simultaneously
processed and recorded by LabVIEW software 8.5 (sound and vibration assistant)
(Fig. 2).
3.2 Methodology and Experimental Procedure
Figure 3 shows the flow diagram of methodology followed for comparison through
extracting the features from fast Fourier transform (FFT) of the submodule of sound
and vibration software in the LabVIEW and modal anlysis of FEA. First step, it was
planned to record vibration signals for beginning of the plate joining to ending of
the plate joining for the size of 100 mm length × 50 mm width × 8 mm thickness.
Once the vibration signals were recorded through data acquisition card of converted
analog-to-digital signal, then further signal processing was carried in LabVIEW for
feature extraction from time domain and frequency domain. In this project, vibration
characteristics such as amplitude and frequencies are extracted in the FFT and used
for comparison of frequencies arrived through modal analysis of free vibration in
ANSYS 18.1.
4.1 FEA Analysis
The workpiece with the size of 100 mm length × 50 mm width × 8 mm thickness

of two plates has been modelled together and outer perimeters are arrested to the
Fig. 2 Flow chart of describing feature extraction from FFT and comparison with FEA
dimensions equal to the dimensions of clamp of 20 mm on both sides which is used

for holding the workpiece and input parameters are given such as mass density is
2710 kg/m3 Young’s modulus is 68.9 × 109 N/m2 and Poisson’s ratio is 0.3. Entire
model is meshed and allowed to run the rectangular plate in free vibration, arrived
modal values for 30 modes. Table 5 shows that first 4 modal values of the rectangular
plate. Figure 3a shows the meshing model of two workpieces and Fig. 3b shows the
modal analysis result of rectangular plates which are performed by FEA software
(ANSYS 18.1) and shown for the first mode.
Fig. 3 a Meshing model of two workpieces b 1st mode result of rectangular plates
Table 5 Modal values of workpiece in free vibration

Component Modes of free vibration I mode II mode III mode IV mode
Workpiece By FEA, Fn (Hz) 6674.8 12,960 13,213 18,614
Tool pin By FEA, Fn (Hz) 3051.9 3052.2 13,533 15,817
Fig. 4 a Meshing model of tool pin b 1st mode result of tool pin
Tool pin is modelled for size of diameter 16 mm × height 60 mm of tool shoulder

and diameter of 6 mm X height 7.5 mm of tool pin. Top surface of tool pin is arrested
and input parameters are given such as mass density, Young’s modulus and Poisson’s
Ratio of 7800 kg/m3 , 2.15 × 1011 N/m2 and 0.3, respectively. Entire model is meshed
and is showed in Fig. 4a. Meshing model is used to run for free vibration and 1st
mode result of tool pin is shown in Fig. 4b.
4.2 Power Spectrum Analysis
Joining operations on plates have been performed using spindle speed of 800 rpm
and feed of 71.2 mm/min. Figure 5 shows that the plates are joined with defects and
Fig. 5 Plate segmented at

three stages
the joined plates are separated into three segments such as metal joining at entry,
steady joining and exit. Segmented areas are monitored through FFT and used for
analysis. During entire process of joining, sensor was positioned at the fixture of the
base plate. While joining plates, accelerometer sensor measured the vibration signal
and LabVIEW software used to store the information in the computer. Totally two
plates were used for joining up to 100 mm length. Time domain and corresponding
frequency domain signals of three segments are shown in Fig. 6. The amplitude was
measured 4.5 × 10−3 m/s2 at entry and exit of joining, while the metal in steady
joining the amplitude level was 6.6 × 10−3 m/s2 . This is captured in the frequency
range between 6000 and 7000 Hz. The variation in the amplitude level occurs because
of friction between the contacting materials, due to which temperature also raises to
solidification stage for joining metal. The results of the experiments conducted are
shown in Fig. 6, which provides time domain and corresponding power spectrum
graph. Power spectrum indicates the corresponding frequencies (Hz) arrived in the
FEA analysis. Figure 7 shows the bar graph of amplitude level against the various
stages of plates.
5 Conclusions
In this study, a relation between the workpiece, tool stiffness and the vibration signals
in joining process is reported. The stiffness of the workpiece is the most influencing
parameter which causes high frequency components in FSW process. The friction
with temperature raise also causes notable changes in high-frequency components
of power spectrum. It was found that 4th region of frequency 6400 Hz of workpiece
frequency is affected by changes in metal joining pattern. This increases the vibration
amplitude in the range of frequencies from 6.0 to 7.0 kHz. The increases in the
Fig. 6 Comparison of power spectra for different stage joining of workpiece. (i) a Time domain
signal and b corresponding FFT during entry. (ii) a Time domain signal and b corresponding FFT
during steady joining. (iii) a Time domain signal and b corresponding FFT during exit of FSW
process
Fig. 6 (continued)
vibration amplitude are observed in this frequency band and are most sensitive during
the metal process. This significant increase of amplitude by two times indicates the
friction increases between tool and workpiece. It is concluded that the monitoring of
4th region frequency of the workpiece can be utilized for effectively monitoring of
defects during FSW process.
Acknowledgements This work is done with the support provided by CNC Condition Monitoring
Lab of Centre for Automation and Robotics at Hindustan University. The authors wish to extend
Fig. 6 (continued)
their sincere thanks to Mr. Natesan, Lab Instructor, ANRO, HITS, Padur for providing necessary
support in executing this research.
For citations of references, we prefer the use of square brackets and consecutive numbers.
Citations using labels or the author/year convention are also acceptable. The following bibliography
provides a sample reference list with entries for journal articles [1], an LNCS chapter [2], a book
[3], proceedings without editors [4], as well as a URL [5].
Fig. 7 Effect of friction on

amplitude
References
1. Lambiase F, Paoletti A, Di Ilio A (2018) Forces and temperature variation during friction stir
welding of aluminum alloy AA6082-T6. Int J Adv Manuf Technol 99:337–346
2. Kumar S, Kumar P (2016) Study the effect of parameters of friction stir welding on the impact
strength of aluminium 6063. Int J Cur Eng Technol 6(3):993–998
3. Buffa G, Fratini L, Micari F, Shivpuri R (2008) Material flow in FSW of T-Joints experimental
and numerical analysis. IntJ Mater Form 1(1):1283–1286
4. Prasanna P, Subba Rao B, Krishna Mohana Rao G (2010) Experimental and numerical evaluation
of friction stir welds of AA6061-T6 aluminium alloy. ARPN J Eng Appl Sci 5(6):1–18
5. Reza-E-Rabby Md (2014) Effect of Tool pin thread forms on friction stir weldability of different
aluminium alloys. Proc Eng 90:637–642
6. Pa Sadeesh, Kannan Venkatesh (2014) Studies on friction stir welding of AA 2024 and AA 6061
dissimilar metals. Proc Eng 75:145–149
7. Shukla J, Echempati R (2016) Investigation and thermal analysis of friction stir welding process
parameters of AA6061 plates. In: Proceedings of the ASME 2016 international mechanical
engineering congress and exposition (IMECE2016), pp 11–17, Phoenix, Arizona, USA
8. Ramesh A, Indira Rani M, Prasanth A (2016) Influence of tool design on the mechanical
properties in friction stir welding of Aluminium alloy AA 6082-T6. Int J Sci Eng Res 7(6)
9. Prantik M (2012) Alloy designation processing and use of AA6XXX series aluminium alloys.
Int Sch Res Netw ISRN Metall 1–15
Power Generation from Hydraulic Shock
Absorber Using Piezoelectric Material
B. Jain A. R. Tony, M. S. Alphin, and V. Yeshwant
Abstract A lot of emphasis has been laid on harvesting energy from unconven-
tional sources these days due to environmental concerns. Harvesting energy from
vibrations is one of the most promising technologies in the present day scenario
like the vibrations of tall buildings, long bridges, vehicle systems, railroads, ocean
waves, and even human motions. For successful harnessing of this wasted energy, a
piezoelectric transducer has been used. The analysis of suspension to find stiffness
and deformation is done. Also, the force acting on the suspension over bumps is
found which is used to calculate the voltage. The power obtained is further calcu-
lated and possible applications of this energy generated are also being mentioned.
The objective is to harvest energy from the suspension system using PZT. Model of
the shock absorber and a Simulink model of quarter car suspension. Evaluation of
the changes in stiffness and deflection of the suspension would be done. A way of
efficient conversion of the vibration to electric power by finding the best position to
place the piezoelectric material. Finally, the force during a bump is found and hence
the voltage induced in PZT. Thus, the power generated and designing a circuit to
store the generated power in a battery is obtained.
Keywords Piezoelectric material · Simulink · Energy harvesting
1 Introduction
There has constantly been a lot of emphasis on the need for reduced energy consump-
tion or energy harvesting in automobiles. There is also an increase in the amount of
B. J. A. R. Tony (B)
Department of Mechanical Engineering, Malla Reddy College of Engineering and Technology,
Secunderabad, India
M. S. Alphin · V. Yeshwant

https://doi.org/10.1007/978-981-15-4488-0_82
972 B. J. A. R. Tony et al.
features and luxuries that are being used by the passengers who commute in pri-
vate four-wheelers. The problem is that most of these features are highly energy-
consuming. Thus, the need to produce or harvest energy through innovation. Few
preliminary studies on the amount of energy recovery in vehicles and at the end they
claimed the recoverable energy for a 2500 lb vehicle with an average speed of 45 mph
are about 20–70% (that amounts up to 7500 W) of the propulsion power on a typical
highway road [1]. Kawamoto et al. [2] modeled a ball-screw type electromagnetic
damper for active automobile suspension and the results of these experiments show
a 15.3 W energy recovery from one shock absorber of a vehicle on Class C road at
50 mph, mainly from vibration above 2 Hz.
A regenerative suspension with a ball-screw and a three-phase motor of a real car
on the vibration test rig and obtained 11.7 W under random excitation or 46 W for
four shock absorbers [3]. Comprehensive analysis about the mean power that can be
obtained from the vehicle and the results suggest that road roughness, tire stiffness,
and even the vehicle driving speed have a great influence on the harvesting power
potential, whenever the suspension stiffness, absorber damping, and vehicle masses
are insensitive or almost constant. At 60 mph, on good and average roads as specified
on an ISO standard, 100–400 W average power is available in the suspensions of
a middle-sized vehicle [4]. The resonance frequency of the piezoelectric transducer
is dependent upon the configuration, size, and loading conditions. They noted that
efficiency (η) of the conversion process in resonance condition is dependent upon
the coupling coefficient and mechanical quality factor of the piezoelectric [5].
Zhongjie et al. [6] have used a key component which is a unique motion mecha-
nism, which they called ‘mechanical motion rectifier (MMR)’, to convert the oscil-
latory vibration into unidirectional rotation of the generator. Vibration energy that
needs to be tapped to understand the unique motion in the cars, buildings, human
motion, and also tidal wave energy. They have also discussed how the energy gener-
ated by vibration which equals to around 300 W is actually contributing around 9%
of the energy fuel consumption regeneration [7]. But on all the above topics there has
been no work done on the stress on the piezoelectric material can be reduced without
reducing the output of the electricity. Hence, this calls the need for the purpose of
the research where it tends to solve the above-stated problems.
2.1 Piezoelectric Materials
Some of the most common forms of the Piezoelectric materials include quartz, lead
zirconatetitanate (PZT) ceramic, and polyvinylidene fluoride (PVDF) polymer. PZT
is the most widely used piezoelectric ceramics and is commercially available. It
has very high piezoelectric coupling coefficients and a relatively low maximum
operating temperature (200C). PZT has some advantages over quartz-like lower cost
Power Generation from Hydraulic Shock Absorber … 973
Table 1 Properties of coil

Parameters Properties
spring
ASTM 228
Yield strength 900–1100 MPa
Young’s modulus 210 GPa
Compressive strength 250 MPa
Ultimate strength 2840 MPa
Poisson ratio 0.3
and versatility in design by changing the composition and properties. Quartz also
has disadvantages such as the varying output according to the temperature variation
of the crystal. If the relative humidity rises above 85% or falls below 35% the output
of the material of quartz is affected. For this defect to be overcome, they have to be
coated with wax or polymer material. Thus mostly the crystal types of piezoelectric
are not preferred than ceramics. Ceramics are less sensitive to temperature changes
and hence they are used. The common ceramics are PZT and PVDF. PZT ceramic has
large electromechanical coupling factors, typically k31 = 0.34 and k33 = 0.69 (k31
is the factor for the electric field in direction 3 and longitudinal vibrations in direction
1; k33 is the factor for the electric field in direction 3 and longitudinal vibrations in
direction 3), which means it is able to convert 34% and 69% of mechanical energy
in the piezoelectric material into electric energy.
PVDF is more flexible and sensitive however, the electromechanical converting
coefficient is much smaller, k31 = 0.12, and k33 = 0.15. Single crystal piezoelectric
materials have also been used for its high energy density, high-energy converting
efficiency, and large operational temperature range. The coil spring properties are
shown in Table 1.
2.2 MATLAB Simulation
A quarter car suspension model was created in MATLAB Simulink and the deflec-
tion of the spring was analyzed. A graph was plotted between the magnitude of
displacement of the car versus time. Frequency and Phase response time was plotted
and compared with and without a thermoplastic polyurethane buffer. In Simulink, an
equation of motion for a suspension system was created using functions and blocks.
The force acting on the spring was determined using the devised equations of motion.
Considering the undulations which cause the displacement in the suspension sys-
tem, an assumption was made for the bump dimensions (Height:10 cm, during the
rise: 1 s). Fieldwork was done to find out the normal terrain which can be assumed to
produce the vibrations. In 8 kilometres, 6 speed breakers, and 10 pitholes were found
to occur. From commonly used graphs in blocks, a sine wave was used to represent
the bumps. Step functions were used to represent pitholes in the terrain for analysis.
A force versus time graph was investigated in the results caused by the pitholes and
Fig. 1 Simulink model of hydraulic suspension with piezoelectric material
bumps. Figure 1 shows Simulink model of hydraulic suspension with piezoelectric

material.
The results of the used analytic techniques mentioned below. Figure 2 shows the
results of total deflection and magnitude were found out for a period of 20 s. The
deflection was found to be 0.0135 m. Figure 3 shows a step amplitude of 2 cm which
is equal to 0.02 m is given as the input for a duration of 1 s. The total force was found
Fig. 2 Magnitude versus time

Fig. 3 Force versus time
Fig. 4 Frequency and phase response
out for a period of 20 s. The figure shows the force versus time and force was 300 N
and 1500 N.
Figure 4 shows the frequency and phase response of the system. The frequency
response shows the relation between the frequency and the magnitude. The phase
response shows the relation between the magnitude and phase.
3.1 Voltage Calculations
A bridge rectifier is an Alternating Current (AC) to Direct Current (DC) converter that
rectifies mains AC input to DC output. Bridge Rectifiers are widely used in power
supplies that provide necessary DC voltage for the electronic components or devices.
They can be constructed with four or more diodes or any other controlled solid-state
switches. Depending on the load current requirements, a proper bridge rectifier is
selected.
Components’ ratings and specifications, breakdown voltage, temperature ranges,
transient current rating, forward current rating, mounting requirements, and other
Fig. 5 Battery storage setup
Fig. 6 Relationship between

voltage and current
considerations are taken into account while selecting a rectifier power supply for an
appropriate electronic circuit’s application.
A capacitor is a passive two-terminal electrical component that stores potential
energy in an electric field. The effect of a capacitor is known as capacitance. While
some capacitance exists between any two electrical conductors in proximity in a
circuit, a capacitor is a component designed to add capacitance to a circuit. Figure 5
shows the battery storing setup. Figure 6 shows the relationship between voltage and
current.
4 Conclusions
The piezoelectric material which has been used along with the thermoplastic
polyurethane buffer in the suspension system has helped in the electric utilization
of Damper Dissipated Energy. Wasted Energy Utilization and automobiles being the
most thriving words in the present day scenario, has been given the primary impor-
tance in this project and been fruitfully researched to produce unexpected results. The
piezoelectric material which has been used along with the thermoplastic polyurethane
buffer in the suspension system has helped in the electric utilization of Damper Dis-
sipated Energy. Future scope of studies can involve the cross-section change of the
used piezoelectric material inside the TPU buffer.
References
1. Goldner RB, Zerigian P, Hull JR (2001) A preliminary study of energy recovery in vehicles by
using regenerative magnetic shock absorbers. Trans J Commercial Veh 110–2
2. Kawamoto Y, Suda Y, Inoue H, Kondo T (2007) Modelling of electromagnetic damper for
automobile suspension. J Sys Des Dyn 1(3):524–535
3. Zhang Y, Huang K, Yu F, Gu Y, Li D (2007) Experimental verification of energy-regenerative
feasibility for an automotive electrical suspension system. 2007 IEEE International conference
on vehicular electronics and safety. China
4. Zuo L, Zhang P-S (2013) Energy harvesting, ride comfort, and road handling of regenerative
vehicle suspensions. J Vib Acoust 135:011002–2
5. Priya S (2007) Advances in energy harvesting using low profile piezoelectric transducers. J
Electroceram 19(1):167–14
6. Li Z, Zuo L, Kuang J, Luhrs G (2012) Energy-harvesting shock absorber with a mechanical
motion rectifier. Smart Mater Struct 22(2):025008
7. Zuo L, Tang X (2013) Large-scale vibration energy harvesting. J Intell Mater Syst Struct
24(11):1405–1430
Static Deformation Analysis
with and Without of Piezo-electric
Material Attachment in Hydraulic
Suspension System
B. Jain A. R. Tony, M. S. Alphin, and Nishanth P. Shah
Abstract A lot of emphasis has been laid on harvesting energy from unconven-
tional sources these days due to environmental concerns. Harvesting energy from
vibrations is one of the most promising technologies in the present day scenario
like the vibrations of tall buildings, long bridges, vehicle systems, railroads, ocean
waves, and even human motions. For successful harnessing of this wasted energy a
piezo-electric transducer has been used. The analysis of suspension to find stiffness
and deformation is done. Hydraulic suspension system was modelled using SOLID-
WORKS and static analysis was performed using ANSYS. The static analysis results
were evaluated as deformation and Von-Mises stress with and without piezo-electric
material.
Keywords Hydraulic suspension system · Piezo-electric material · Static

deformation · Von-Mises stress
1 Introduction
The harvested energy from an automobile system is mainly obtained from the energy
absorbed due to the vibration of a car on irregular roads that are being damped
in order to have an ergonomic design for the passenger during his drive. Thus,
its suspension system uses a dampener or a shock absorber to absorb the energy
produced from the vibrations. But this energy harvested can be used to power the
car or any of its associated devices. Piezo-electric material can be used to absorb
these vibrations and convert it into a form of electrical energy that can be stored in
a battery. The transduction mechanism such as the piezo-electric, electrostatic and
B. Jain A. R. Tony (B)

Secunderabad, India
M. S. Alphin · N. P. Shah

https://doi.org/10.1007/978-981-15-4488-0_83
980 B. Jain A. R. Tony et al.
electromagnetism has been used to the human-based devices with the average human
of 68 kg produces energy of 67 watts but due to the current piezo-electric conversion
efficiency and other difficulties, the amount of energy that can be extracted from
the human who walks on the floor is 1.27 watts [1]. Green advocates the need for
wireless charging and also mechanisms used for generation. There are two types of
generation that take place, direct force generation, and the inertial force generation.
The direct force generation examples include piezo-electric shoes, converts the force
exerted by the feet on the shoes into electricity [2]. Targeted mechanical vibrations
lie in the range of 50–150 Hz with force amplitude in the order of 1 KN (automobile
engine vibration level). It was found that under such severe stress conditions the
metal–ceramic composite transducer “cymbal” is a promising structure. The metal
cap enhances the endurance of the ceramic to sustain high loads along with stress
amplification.
The transducer design plays a vital role in enhancing the conversion of electrical
energy [3]. However, the initial theoretical research on this energy harvesting from
the vehicle suspension began almost three decades ago [4, 5]. The possibility of using
a permanent magnetic motor as mechanical dampers, for dissipating the energy using
variable resistors [5]. Segel et al. [6] have analyzed the influence of highway pavement
roughness on vehicular resistance to motion due to tire and suspension damping and
they have indicated that approximately 200 W of power is dissipated by dampers of
a passenger car at 30 mph (miles per hour). The electrical active suspension with
LQG (Linear-Quadratic-Gaussian) control and estimated that up to 400 W of energy
is recoverable at a highway driving condition, is 5% of propulsion power to maintain
GM Impact (electrical car) at 60 mph [7]. Abouelnour et al. [8] have looked into the
concept of energy and electro-mechanic suspension and simulation based on quarter
car models that predict around 150 W of the energy was being dissipated by shock
absorbers and the energy can be converted into electrical power at 56 mph. The
primary aim of this work is to make an effective energy conversion of the vibration
arising due to the undulations on the road.
2.1 Modelling of Hydraulic Suspension System
The spring is used to coil around the cylinder part of the suspension system. The
spring was also modelled based on the standard suspension model. This includes
the cylindrical part of the suspension system. They also include a cylindrical piston
which is useful for the hydraulic motion that dampens the vibration. The buffer used
in the coil suspension system of the vehicles for various reasons. The life of the
suspension jumper increases on the usage of the buffer since the buffer takes up
some of the stress. The height of the suspension increases by 2–4 cm. The objective
is to reduce the shock felt by the passengers due to irregularities of the road.
Static Deformation Analysis with and Without of Piezo-electric … 981
Table 1 Properties of buffer

Elastomer 800 series
Density 1200
Young’s modulus 2–10 GPa
Tensile strength 40–45 MPa
Poisson ratio 0.5
As stated above since the shock is they result in the balance of the car being
maintained. They also help in maintaining the spring tension and also restoring it
back after the vibration. They also reduce the noise and vibration from the suspension
components. It also reduces the force which makes the automobile slant to its side.
It decreases the flat force which shifts the majority of the center of gravity while it’s
cornering. It enhances the flexibility if the automobile body those results in smoother
control, leaning phenomenon reduction and more constant driving. It increases the
linear force and the landing force of the tyres which results in better road breaking
and effective gripping at high speeds. Thus this reduces the braking distance of the
car. The properties of the buffer are mentioned in Table 1.
The most common type of material used in the coil spring is Carbon Steel. This
type of material is mostly used since they have high strength and also wear resistance
ability. They can also be alloyed with other metals to obtain the desired level of
properties. They are also easily available and also are at low cost compared to many
metals. The coil spring properties are shown in Table 2.
Hydraulic piston rods are a case in point where conventional steels are performing
inadequately. These instances have resulted in oil leaks or pressure loss, or else have
necessitated regular and disruptive cylinder changes. Analysis has determined that
these failures are caused by cyclic stresses, which are typical in demanding mining
and construction applications, and to which piston rods are commonly exposed over
long periods. The main driver of cost and capital in piston rods in hydraulic cylinders
is the diameter of the bar. The cost of a rod increases exponentially with the diameter
as a result.
As a general rule, the cost of a piston rod decreases by 15% by reducing its
diameter by 5 mm. Other effects are lighter weight and reduced space requirements,
Table 2 Properties of coil

spring
ASTM 228
Yield strength 900–1100 Mpa
Young’s modulus 210 GPa
Compressive strength 250 MPa
Ultimate strength 2840 MPa
Poisson ratio 0.3
Fig. 1 The assembly of

hydraulic system for
suspension
which could also translate into lower energy consumption and greater versatility
in design. But to achieve these benefits, the rod material must be strong enough
to maintain equal strength at the smaller diameter. Therefore, a key parameter in
hydraulic piston rod design is dimensioning against fatigue, buckling, and impact
failure. Thus the nickel-chromium plated steel is actually used for piston but the
thickness of that plating should be minimum as possible to reduce the weight and
improve the efficiency. Figure 1 shows the assembly of the hydraulic system for
suspension.
2.2 Peizo-Electric Material
The PZT material has been chosen as the piezo-electric material based on which the
energy can be harvested. They are actually fitted into the holes of the buffer that lies
in between the coil spring. They are placed in such a way since the direct dynamic
force acting on the PZT material can actually deter the life of the piezo material and
so they can cost a lot. The use of buffer can reduce this effect. The properties of
peizo-electric material are shown in Table 3.
Table 3 Properties of
peizo-electric material
Density 7500 GPa
Young’s modulus 63 Gpa
Voltage co-efficient 0.025 Vm/N
Ultimate tensile strength 240 MPa
Poisson ratio 0.31
2.3 Numerical Simulation
The solid model of the entire shock absorber was modelled using the solid works
software in parts. The separate parts were assembled to create the entire part mirroring
exactly the shock absorber in reality. The interactions between each part of the shock
absorber were finalized and setup as a final design. The designed solid model was
imported into Ansys using appropriate dimensions and tolerances. Static structural
analysis was carried out on the solid model without the inclusion of Thermoplastic
polyurethane buffer. The appropriate sections and materials were input in the analysis.
The model was meshed using a quad element. Loads and boundary conditions were
appropriately given on the model to be analyzed. The equivalent stress and total
deformation at each element in the model were determined and investigated.
A static structural analysis was done on the solid model with the inclusion of the
Thermoplastic polyurethane buffer using the same procedures as the model without
the TPU. Now, the Piezo-electric material was attached to the model. The Max force
that was obtained from the results in the MATLAB software was applied on the
model in the bottom surface and the top surface is fixed. The new job was analyzed
and the results were plotted into required graphs.
The static-structural analysis was performed using Ansys and results were evalu-
ated in terms of static structural deformation, and static structural Von-Mises stress.
Figure 2 shows the static structural deformation of hydraulic suspension system
with and without piezo-electric material. The numerical simulation result shows that
the static structural deformation was found 0.2078 m for suspension system with
piezo-electric material and 0.2471 m for normal suspension system.
The static deformation was controlled up to 0.0393 m. The piezo-electric mate-
rial controls the deformation up to 15.9%. The minimum deformation was found
0.02309 m for suspension system with piezo-electric material and 0.027457 m for
normal suspension system. The average of static deformation was found 0.115 m for
suspension system with piezo-electric material and 0.137 m for normal suspension
system.
(a) Hydraulic suspension system static structural deformation with piezo-electric

material
(b) Hydraulic suspension system static structural deformation without piezo-

electric material
Fig. 2 Static structural deformation with and without piezo-electrical material

Figure 3 shows the static structural Von-Mises stress of hydraulic suspension

system with and without piezo-electric material. The numerical simulation result
(a) Hydraulic suspension system Von-Mises stress with piezo-electric material
(b) Hydraulic suspension system Von-Mises stress without piezo-electric

material
Fig. 3 Static von-mises stress with and without piezo-electrical material

shows that the static structural Von-Mises stress was found 3.05 GPa for a suspension
system with piezo-electric material and 3.138 GPa for the normal suspension system.
The static Von-Mises stress was controlled up to 0.088 GPa. The piezo-electric
material controls the deformation up to 2.8%. The minimum Von-Mises stress was
found 0.0339 GPa for suspension system with piezo-electric material and 0.0348 GPa
m for normal suspension system. The average of static Von-Mises stress was found
1.695 GPa for suspension system with piezo-electric material and 1.174 GPa for
normal suspension system.
4 Conclusions
• Wasted Energy Utilization and automobiles being the most thriving words in the
present day scenario, has been given the primary importance in this project and
been fruitfully researched to produce unexpected results.
• The piezo-electric material which has been used along with the Thermoplastic
polyurethane buffer in the suspension system has helped in the Electric utilization
of Damper Dissipated Energy.
• The varied applications such as headlight of automobiles, parking rearview cam-
eras can be powered using this new harnessed energy thus reducing the extra
energy required to power luxury utilities in an automobile.
• The newly innovated model of the suspension has been found to be structurally
stable using the results from the Ansys static analysis.
References
1. Beeby SP, Tudor MJ, White NM (2006) Energy harvesting vibration sources for microsystems
applications. Meas Sci Technol 17(12):175–195
2. Mitcheson PD, Yeatman EM, Rao GK, Holmes AS, Green TC (2008) Energy harvesting from
human and machine motion for wireless electronic devices. Proc IEEE 96(9):1457–1486
3. Kim HW, Batra A, Priya S, Uchino K, Markley D, Newnham RE, Hofmann HF (2004) Energy
harvesting using a piezoelectric “Cymbal” transducer in dynamic environment. Jpn J Appl Phys
43(1): Part 1, Number 9A
4. Karnopp D (2007) Permanent magnet linear motors used as variable mechanical dampers for
vehicle suspensions. Int J Veh Mech Mobility 18(4):187–200
5. Karnopp D (1992) Power requirements for vehicle suspension systems. Int J Veh Mech Mobility
21(1):65–71
6. Segel L, Lu X-P (1982) Vehicular resistance to motion as influenced by road roughness and
highway alignment. Aust Road Res 12(4):211–222
7. Hsu P (1996) Power recovery property of electrical active suspension systems. IECEC 96.
Proceedings of the 31st intersociety energy conversion engineering conference (vol. 3, pp. 1899–
1904). Washington, DC, USA
8. Abouelnour AMA, Hammad N, Ebrahim E (2003) Electric utilization of vehicle damper
dissipated energy. Proceedings of the IASVD symposium (p. 647). Japan
Study of Vibration and Tensile
Characteristics of Multilayer Composites
D. Vishal, M. Selvaraj, and S. Vijayan
Abstract In recent years, the use of composites for engineering purposes has grown
enormously. Composites can be fabricated to the necessary requirements and can
satisfy the required needs. The epoxy resin, used in combination with a hardener, acts
as an adhesive and matrix of the composite. Glass fiber has been unanimously used
along with epoxy for better mechanical properties. Various Rubbers are incorporated
as the middle layer in the multilayer composite structures. Tensile and vibration
tests are performed on these composites. The presence of rubber in the composite
structure influences both the tensile and vibration characteristics of the multilayer
composites. Hence, the tensile strength, natural frequency, and the damping factor
shows variation due to the presence of rubber and this property can be exploited for
engineering purposes.
Keywords Glass fiber · Rubber · Tensile characteristics · Vibration characteristics
1 Introduction
Composites have been the major trend of the materials in recent times. Compos-
ites yield us the desired properties and necessary characteristics. Over the years,
Epoxy Resin has been used as the most common adhesive and matrix of the com-
posite. Epoxy has been extensively used for its low shrinkage properties. They also
exhibit brilliant mechanical characteristics. Moreover, the fiber being used in the
multilayer composites has a great influence on the strength characteristics of the
composites [1]. Different types of fiber materials [2, 3] are used along with epoxy
resin in the composite structure. Natural fibers, as well as synthetic fibers, are used
in combination with epoxy resin. Natural fibers such as flax, kenaf and synthetic
fibers such as carbon fibers, glass fibers are the most commonly used fibers. Glass
fiber shows excellent adhesion with epoxy resin [4]. The high energy absorption
D. Vishal · M. Selvaraj (B) · S. Vijayan

Kalavakam, Chennai 603110, India

https://doi.org/10.1007/978-981-15-4488-0_84
988 D. Vishal et al.
capability [5] and good tensile strength characteristics [6] of the glass fiber/epoxy
resin combination are impressive enough to overshadow the other combinations of
resin and fiber. Rubber is a viscoelastic material. It is an elastomer type of polymer
compound. Both the natural and synthetic rubbers are used in composites. The good
elastic properties of rubber [7] draws interest in their usage in composite structures.
The rubber/epoxy resin combination shows good physical and thermal properties
[8]. Rubber/epoxy combination also shows excellent toughness [9, 10], durability
[11] and other mechanical properties. Nitrile Butadiene Rubber shows high strength
and damage accumulation properties under cyclic loading conditions [12]. Styrene
Butadiene rubber shows good adhesion and viscoelastic properties [13]. The SBR
also shows good damping properties [14]. The thermoplastic polyurethane shows
good adhesion with epoxy resin [15]. Polyurethane imparts considerable strength to
the composite [16]. Silicone rubber has good tensile properties [17]. The excellent
properties of these rubbers are exploited by their usage in the composite structure
along with epoxy resin and glass fiber. Their tensile and vibration characteristics are
studied.
2.1 Epoxy Resin and Hardener
Araldite LY556 was used in the multilayer composite as the matrix. It is a two-
part resin. Hardener HY951 was employed as the curing resin. This combination of
epoxy and hardener is most commonly used for composite structures. The epoxy and
hardener were mixed in the ratio 10:1 for each layer of the multilayer composite.
2.2 Glass Fiber and Rubber
Woven glass fiber was used in the multilayer composite. The density of the glass
fiber is 1.69 gm/cm3 [18].
Nitrile Butadiene Rubber is used in the composite structure. Its market name is
NBR, Europrene and Krynac. The original names are Buna—N, and Acrylonitrile
Butadiene rubber. The Shore A hardness is 30–90 [19]. Styrene Butadiene rubber is
another rubber that is used. The market name of this rubber is SBR and Neolite. The
original name is Buna–S. The Shore A hardness is 35–95 [19]. Polyurethane rubber
is one of the rubber that is used in the composite structure. The market name of this
rubber is PU or Polyurethane rubber. The Shore A hardness is 55–98 [20]. Silicone
Rubber is also used. The market name of this rubber is Silicone Rubber. The original
name is Polysiloxane. The Shore A hardness is 40–80 [21].
Study of Vibration and Tensile Characteristics … 989
2.3 Hand Lay-up Method
Fabrication of composite laminate was done using the hand lay-up method [22]. The
first layer is the OHB sheet. A wax coating is given on the OHB sheet. Then, epoxy
hardener mixture is poured and spread over the OHB sheet. A glass fiber sheet is
placed over the epoxy hardener coating. Another layer of epoxy hardener coating
is given over the first glass fiber sheet and a second glass fiber sheet is placed over
it. Epoxy hardener coating is given on the second glass fiber sheet and rubber sheet
was placed over it. Alternatively, epoxy hardener coating and glass fiber sheets are
stacked upon one another. At last, an OHB sheet is placed on the top. After the
solidification, the two OHB sheets are removed to obtain the multilayer composite.
This procedure is repeated for different rubber sheets.
2.4 Mechanical Test
Tensile Test
Tensile test was conducted using a Universal Testing Machine called UTM. It
is also called Tensiometer or Universal Tester. The loading capacity is 53MN [23].
The fabricated composites were cut to ASTMD638 standard and tensile test was
performed.
Vibration Test
Vibration test was conducted using Dewesoft software. The fabricated composites
were cut to 25 × 250 mm2 and vibration test was performed [24].
3.1 Tensile Test
The tensile test was conducted on the composite specimens. The following results
were observed.
From Fig. 2, it can be noted that the UTS for the NBR composite plate is 92 MPa.
This UTS is 65% lesser than the glass fiber composite plate. The NBR composite
plate fails at a strain value 5.3% lesser than the glass fiber. The UTS for the SBR
composite plate is 89 MPa (Figs. 1 and 3).
This UTS is 66.3% lesser than the glass fiber composite plate. The SBR composite
plate fails at a strain value of 15.7% lesser than the glass fiber.
The UTS for the PU composite plate is 104 MPa This UTS is 60.6% lesser than the
glass fiber composite plate. The PU composite plate fails at a strain value of 20.3%
lesser than the glass fiber. The UTS for the Silicone composite plate is 118 MPa.
This UTS is 55.3% lesser than the glass fiber composite plate. This means that the
Fig. 1 Specimens post tensile test
Fig. 2 Tensile test results
Silicone composite plate fails at a strain value of 53.4% more than the glass fiber.
TheUTS for the glass fiber composite plate is 264 MPa. This UTS is more than the
other four composite plates.
3.2 Vibration Test
The vibration test was conducted on the composite specimens.

From Table 1, it can be seen that the Natural Frequency of the NBR composite
plate is 47 Hz.
This Natural Frequency is 22.9% lesser than the glass fiber composite plate. The
damping factor of NBR is 0.31. The Natural Frequency of SBR composite plate is
Fig. 3 Consolidated stress–strain graph
Table 1 Vibration test results

Composite Natural frequency (Hz) Damping factor
NBR 47 0.31
SBR 48 0.34
Polyurethane 41 0.41
Silicone 46 0.26
Glass fiber 61 0.19
48 Hz. This Natural Frequency is 21.3% lesser than the glass fiber composite plate.
The damping factor of SBR is 0.34 (Figs. 4 and 5).
It can be noticed that the Natural Frequency of Polyurethane composite plate
is 41 Hz. This Natural Frequency is 32.7% lesser than the glass fiber composite
plate. The damping factor of PU composite is 0.41. It may be noted that the Natural
Frequency of Silicone composite plate is 46 Hz. This Natural Frequency is 23.5%
lesser than the glass fiber composite plate. The damping factor of Silicone composite
is 0.26. The Natural Frequency of glass fiber composite plate is 61 Hz.
The Natural Frequency of glass fiber sandwich is highest among all the composite
plate. The damping factor of glass fiber composite is 0.19.
4 Conclusion
Various Composites with Different Rubbers were Fabricated and tested. Their
Mechanical and Dynamic Properties were analyzed, and results were discussed
below.
Fig. 4 Vibration testing
Fig. 5 Consolidated natural frequency versus acceleration graph
• It is seen that the glass fiber composite has the Highest Natural Frequency value
of 61 Hz and Lowest damping factor of 0.19.
• The Polyurethane composite has the least Natural Frequency value of 41 Hz and
Highest damping factor of 0.41.
• It is noted that the presence of rubber increases the damping property.
• We can observe that the Natural Frequency is independent of the excited

acceleration.
• It can be seen that the glass fiber ranks as the composite having highest Ultimate
Tensile Strength of value 264 MPa. This is due to good interfacial bonding between
reinforcement with a matrix.
• Out of the rubber Sandwich composites, it can be noted that the Silicone composite
plate has the highest Ultimate Tensile Strength of value 118 MPa.
• The SBR composite plate has the lowest Tensile Strength (89 MPa) out of the
four rubber Sandwich composites.
References
1. Sarikaya E, Çallioğlu H, Demirel H (2019) Production of epoxy composites reinforced by

different natural fibers and their mechanical properties. Compos Part B Eng 167(15):461–466
2. Huang JK, Bin Young W (2019) The mechanical, hygral, and interfacial strength of continuous
bamboo fiber reinforced epoxy composites. Compos Part B Eng 166(December 2018):272–283
3. Davis DC, Wilkerson JW, Zhu J, Ayewah DOO (2010) Improvements in mechanical properties
of a carbon fiber epoxy composite using nanotube science and technology. Compos Struct
92(11):2653–2662
4. Plonka R, Mäder E, Gao SL, Bellmann C, Dutschk V, Zhandarov S (2004) Adhesion of
epoxy/glass fibre composites influenced by aging effects on sizings. Compos Part A Appl
Sci Manuf 35(10):1207–1216
5. Aymerich F, Dalla Via A, Quaresimin M (2011) Energy absorption capability of nanomodified
glass/epoxy laminates. Procedia Eng 10:780–785
6. Maciel NDOR et al (2018) Comparative tensile strength analysis between epoxy composites
reinforced with curaua fiber and glass fiber. J Mater Res Technol 7(4):561–565
7. Yang H et al (2018) Highly sensitive and stretchable graphene-silicone rubber composites for
strain sensing. Compos Sci Technol 167(August):371–378
8. Irez AB, Bayraktar E, Miskioglu I (2017) Design and mechanical-physical properties of epoxy-
rubber based composites reinforced with nanoparticles. Procedia Eng 184:486–496
9. Tsai JL, Huang BH, Cheng YL (2011) Enhancing fracture toughness of glass/epoxy composites
for wind blades using silica nanoparticles and rubber particles. Procedia Eng 14:1982–1987
10. Yahyaie H, Ebrahimi M, Tahami HV, Mafi ER (2013) Toughening mechanisms of rubber
modified thin film epoxy resins. Prog Org Coat 76(1):286–292
11. Wang J, Dai Q, Guo S, Si R (2019) Mechanical and durability performance evaluation of crumb
rubber-modified epoxy polymer concrete overlays. Constr Build Mater 203:469–480
12. Higashino M, Takemura K, Fujii TJ (1995) Strength and damage accumulation of carbon
fabric composites with a cross-linked NBR modified epoxy under static and cyclic loadings.
32:357–366
13. Rodriguez N, Dorogin L, Chew KT, Persson BNJ (2018) Adhesion, friction and viscoelastic
properties for non-aged and aged styrene butadiene rubber. Tribol Int 121(January):78–83
14. Jiang Z, Easa SM, Hu C, Zheng X (2019) Understanding damping performance and mechanism
of crumb rubber and styrene-butadiene-styrene compound modified asphalts. Constr Build
Mater 206:151–159
15. Hausberger A, Major Z, Theiler G, Gradt T (2018) Observation of the adhesive- and defor-
mation—contribution to the friction and wear behaviour of thermoplastic polyurethanes. Wear
412–413(July):14–22
16. Saleh S, Zurairahetty N, Yunus M, Ahmad K, Ali N (2019) Improving the strength of weak
soil using polyurethane grouts: a review. Constr Build Mater 202:738–752
17. Ziraki S, Mojtaba S, Hadianfard MJ (2016) A study on the tensile properties of silicone rubber/
polypropylene fibers/ silica hybrid nanocomposites. J Mech Behav Biomed Mater 57:289–296
18. Gao G (2017) Mechanical properties of woven glass fiber-reinforced polymer composites.
(April)
19. Mostafa A, Bayoumi MR (2009) Insight into the effect of CB loading on tension, compression,
hardness and abrasion properties of SBR and NBR filled compounds. Mater Des 30(5):1785–
1791
20. An S, Kim B, Lee J (2016) Incomparable hardness and modulus of biomimetic porous
polyurethane films prepared by directional melt crystallization of a solvent. J Cryst Growth
21. Nowakowska-toporowska A, Malecka K, Raszewski Z Changes in hardness of addition-
polymerizing silicone-resilient denture liners after storage in artificial saliva. J Prosthet Dent
1–5
22. Avila AF, Morais DTS (2005) Science and a multiscale investigation based on variance analysis
for hand lay-up composite manufacturing. 65:827–838
23. Rajesh S, Ramnath BV, Elanchezhian C, Abhijith M, Riju RD, Kishan KK (2018) Science
direct investigation of tensile behavior of kevlar composite. Mater Today Proc 5(1):1156–1161
24. Naito H, Bolander JE (2019) Damage detection method for RC members using local vibration
testing. Eng Struct 178(September 2018):361–374
Influence of Vibro-isolator Attachment
for a Jackhammer to Reduce Vibration
Discomfort
B. Jain A. R. Tony, M. S. Alphin, and Vishal Venkatesh
Abstract Jackhammers are widely used in the construction industry to break up

rock, pavement and concrete. They are generally powered by compressed air, electric
motors, or hydraulics and can generate a large force for drilling and demolition.
However, while they are an efficient tool for this purpose, they also pose a serious
danger to the worker, due to the vibrations transmitted. Prolonged exposure to these
vibrations can cause ailments such as vibration white finger, Raynaud’s disease and
carpal tunnel syndrome (CTS). The objective of this research is to design and fabricate
a vibro-isolator attachment to damp and absorb the range of harmful vibrations
transmitted to the occupant. The attachment is clamped on top of the jackhammer and
makes use of two helical springs in parallel to reduce the higher amplitude vibrations.
Handles are provided above the springs for the user to grip the attachment and hence,
the jackhammer. The new design, setup feels worker much less hand-arm vibration
without reducing the downward drilling force of the jackhammer. Hence, the newly
designed vibration isolation, attachment reduces the hand-arm vibration, and the
jackhammer can be operated at full power without compromising on the health of
the occupants.
Keywords Vibro-isolator · Jackhammer · Carpal tunnel syndrome · Design of

springs
B. Jain A. R. Tony (B)

Secunderabad, India
M. S. Alphin · V. Venkatesh
Chennai, India

https://doi.org/10.1007/978-981-15-4488-0_85
1 Introduction
A large magnitude of vibrations is transmitted to the hand-arm system while using

hand-operated tools. The variety of occupations like construction, industry, agricul-
ture and medical applications are repeatedly using the hand-operated vibration tools.
Continues usage of vibrating tools transmits vibrations to the hand-arm system which
leads to discomfort and pain [1] and also such vibrations lead to Carpal Tunnel Syn-
drome (CTS) and Hand-Arm Vibration Syndrome (HAVS) [2]. The level of vibration
transmitted from the hand-operated tools are evaluated by various researchers such
as grass trimmer (4.5–11.3 m/s2) [3], rock drill (24–25 m/s2) [4], pneumatic hammer
(30 m/s2) [5], and orbital sander (3.9–7.3 m/s2) [6]. Improper design and continuous
exposure may lead to a combination of muscular, circulatory, neurological, bone and
joint disorders [7, 8].Vibration exposure for a single day A (8) has been categorized
by European Union (2005) based on that the exposure limit value and action value
should be 5.0 m/s2 and 2.5 m/s2, respectively [9].
In order to overcome the hand-arm vibration, the assessment methods used for
vibration measurement and its guidelines are drawn by ISO 2001 (International Stan-
dard Organization). The frequency-weighted vibration total value (ahv) was evalu-
ated by combining the vibration acceleration of all three directions. The vibration
measurement axes were suggested by ISO 5349-1 (2001); Zh-axis as the longitudi-
nal axis from the third metacarpal bone; the Yh-axis perpendicular to the Xh and
Zh axes; Xh-axis perpendicular to the Zh-axis [10]. Vibration injury was assessed
between 8–1000 Hz of the working frequency. The vibrations in the frequency range
of 6–20 Hz are more harmful to humans, but the effect of vibration decreases with
the increase of frequency [11].
The response from human for vibration was depended on four factors such as
direction of vibration, magnitude of vibration, exposure time and duration [12].
Other factors influenced hand-arm vibration transmission are forces of grip and
push, posture of hand and contact area between the hand and handle. In order to
estimate the hand-transmitted vibration objective and subjective measurements were
followed by several researchers. The perceptions of workers in the workplace were
evaluated by subjective measurements using observation and questionnaire methods
[13, 14]. Subjective measurement accuracy is based on the honesty of the subjects
[2]. Subjective measurements are very easy to estimate the hand-arm transmitted
vibration. In objective measurements, the data obtained are duration of hand-arm
vibration exposure, frequency of the vibration and magnitude of vibration. However,
objective measurements are more accurate and required expensive sensory devices
with experienced operators. The measurement accuracy of the objective method
depends on biodynamic response of the humans, working environment, working
conditions, tool condition and activity of the work [15]. In order to predict the hand-
arm vibration health effects, Poole and Mason [16] combined both objective and
subjective measurements.
Influence of Vibro-isolator Attachment for a Jackhammer … 997
Improved design of handles and accurately designed handles can provide safety
and comfort during work, increase performance and decrease disorders [17]. Anti-
vibration coatings were given to the handles in order to reduce the vibration and the
vibration reduction was evaluated in the hand-arm system [18]. Alphin et al. (2013)
considered the effect of handle diameter to overcome the vibrations transmitted to the
hand-arm system [19]. Influence of shape and size was considered by Tony et al. to
develop the human hand based handles and the vibration was reduced up to 14% than
the traditional cylindrical handles [20]. The effect of handle diameter was simulated
by Tony et al. (2019) using hybrid finger model and FE analysis [21]. The objectives
of the present study are (1) to develop a vibro-isolator attachment for jackhammer
by considering the design calculations, (2) to evaluate the effect of vibro-isolator in
vibration reduction.
2.1 Subjects and Jackhammer Specifications
Six male subjects between the age group of 21–30 years, and who were free from
musculoskeletal disorders, volunteered to participate in this experiment. Table 1
(a) and (b) shows the anthropometric descriptions of the participants and jackham-
mer specifications in detail. The detailed description of experimental procedure was
explained for every subject before the experiment was initiated.
Table 1 Description of
Characteristics Mean S.D Range
anthropometric
measurements and Age (years) 28 6.48 21–30
jackhammer specifications: a Height (cm) 164 3.92 152–170
Characteristics of subjects. b
Weight (kg) 76 4.89 61–79
Jackhammer characteristics
Middle finger 192 3.86 182–191
length (mm)
Thumb finger 110 3.66 103–116
length (mm)
Grip diameter 52 2.65 44–52
(mm)
Characteristics Specifications
Make Bosch GSH 11 E
Power rating 1500 W
Impact rate 900–1890 bpm
Impact energy 16.8 J
Weight Po 266.7119 N
2.2 Design Calculation of Springs
The presence of spring as a damper in the attachment to the handle will help reduce
the vibrations considerably and also help avoid the various health hazards caused due
to excessive vibrations. In order to design a spring as a damper, spring parameters to
be considered such as arrangement of spring, number of springs, diameters of inner
(Di ), wire (d) and mean (D), number of turns (n), spring index (C) and compressive
force (P) of each spring. The calculated spring index was used to evaluate the Wahl
stress factor (K) and found as 1.148 from the graph. The important parameters for
designing the spring is shear stress (τ ), deflection (y), and stiffness (q) were calculated
using Eq. (1), (2) & (3) as 237.77 N/mm2 , 22.415 mm, and 5.949 N/mm, respectively.
8KPC
τ= (1)
π d2
8PC 3 n
y= (2)
Gd
Gd
q= (3)
8C 3 n
2.3 Three-Dimensional Handle Attachment Design
Considering the design parameters of springs the handle with spring attachment was
designed using modelling software SOLIDWORKS 2016®. The various parts were
first modelled individually and finally assembled to create the attachment. There are
three distinct parts in the assembly. The springs are rested on the top surface of the
bottom part and enclosing the cylinders projecting from the surface. Then the top
part is added to the assembly. The axis of the solid cylinder and hollow cylinder are
made coincident and one of the side surfaces of the top and bottom part is also made
coincident. Now the top part is constrained in two directions and can move only in
one direction. The top part is then moved downwards until it is in contact with the
top of the spring. Then it is fixed in position. The damping attachment is assembled
as shown in Fig. 1.
2.4 Fabrication of Handle
The aluminium block of dimensions 210 × 130 × 20 was purchased then; the block
was measured and milled to the dimensions 200 × 120 × 20. Two holes were drilled
in the block at a distance of 55 mm from the smaller edge, a distance of 60 mm from
Fig. 1 Three-dimensional handle attachment design
the longer edge and of diameter 29.9 mm. Then, the aluminium block of 15 mm
thickness is taken and milled so that it too has a length of 200 mm and a breadth of
120 mm. On the 20 × 120 mm faces of the slab, holes of 8 mm diameter were drilled
using a vertical (manual) drill machine. These holes were of a depth of 45 mm.
Two aluminium rods of length 125 mm and diameter 34 mm are taken and were
machined (faced and turned in lathe) to a diameter of 32 mm and a length of 120 mm.
Then, the plates are drilled (centrally) with a 6 mm bit to obtain a 6 mm hole. The
hole diameter is increased to 8 mm by means of boring process. The springs are
placed over the mild steel rods centered on them.
Then, the hollow aluminium tubes are placed on either face (20 × 140) of the
aluminium slab and are screwed in using 8 mm Allen bolts of length 130 mm. Now,
the 20 mm thick slab is placed over the 15 mm thick slab so that the MS rod slides in
and out of the MS tube. The two parts are locked together when a cotter is placed in
the hole in the MS rod. The handle setup with spring attachment is shown in Fig. 2.
2.5 Experimental Procedure
The two aluminium blocks are placed on top of each other with the spring placed
over the mild steel rods centered on them. The two aluminium blocks are locked
together when a cotter is placed in the hole in the MS rod. The attachment is then
placed over the handle of the jackhammer as shown in Fig. 3.
The vibrations were recorded at the base of the handle and at the wrist of each
subject. The accelerometer was attached using the lightweight strip to avoid the
measurement errors according to the ISO 5349-2 standard. Once the subject found
the correct posture the experiment was initiated and the vibrations were measured
Fig. 2 Fabricated handle attachment setup
Fig. 3 Experimental studies with vibro-isolator attachment
up to 0-1000 Hz. IBM SPSS Statistics v.20 software® package was used to store all
the acquired data.
A tri-axial accelerometer was used to determine the amplitude of vibrations with

and without the attachment. These values at selected frequencies are tabulated below
Table 2. Time-domain and frequency domain data was stored using this experiment.
Time-domain data gives the relation between time and frequency. The below values
of the amplitude of vibration, with and without the attachment, a graph is plotted
is shown in Fig. 4. This graph is a plot of frequency domain that relates between
frequency and amplitude.
Table 2 Experimental data

S. No Frequency in Hz Amplitude before damping in Amplitude after damping in mm
mm
1 150 11.97 7.89
2 200 8.42 1.30
3 300 25.84 2.00
4 350 17.31 4.30
5 400 55.41 2.45
6 500 16.49 0.99
7 600 6.53 1.38
8 700 27.91 2.22
9 800 11.39 1.83
10 900 8.18 3.0
11 1000 11.48 0.80
60
Before Damping
After Damping
50
Amplitude (mm)
40
30
20
10
0
200 400 600 800 1000
Frequency (Hz)
Fig. 4 Amplitude and frequency distribution

The graph denotes the amplitude of vibration before and after the attachment.
Thus, a vibration-damping attachment for a jackhammer was designed and fabricated
with the necessary dimensions. The aluminium slabs were accurately machined by
milling. The solid and hollow mild steel cylinders were machined by turning and
finished by cylindrical grinding. The fabricated attachment successfully damps the
vibrations from the jackhammer as evidenced by the above graph. From the graph,
we infer that the damping effect is not the same for all the frequencies. The damping
is more for some frequencies and lesser for some other frequencies. The amplitude
of vibration is greatly reduced in the region of frequency around 400 Hz. The results
were in good agreement with experimental analysis conducted for ergonomic handle
shapes [20].
It is also very effective in the region 700 Hz and in the region around 300 Hz as
illustrated by the graph. From the graph and from the experiments, we inferred that
the attachment was not effective in the lower frequency range (frequency less than
150 Hz). A similar response was found by the various researchers in order to control
the vibrations below 150 Hz. The static gripping analysis of hybrid finger model was
in good agreement with the experimental results [21, 22].
4 Conclusions
Thus, the vibro-isolator attachment for a jackhammer was successfully designed and
fabricated. Experimental trials were conducted which signify successful damping
of the vibrations from a jackhammer. This attachment reduces the risk of ailments
that occur due to prolonged exposure to the vibrations from a jackhammer. The fact
that the attachment is relatively inexpensive serves to highlight the feasibility of
the product. Hence, using this attachment, the operator can use the jackhammer for
longer durations of time without any risk to his wellbeing.
Ethical Approval All procedures performed in studies involving human participants were in
accordance with the ethical standards of the institutional ethics and research committee (Dr. S. Sali-
vahanan, Dr. V.E Annamalai, Dr. A. Kavitha). The participants gave their concerns for experiments
and publications.
References
1. Tewari VK, Dewangan KN, Karmakar S (2004) Operator’s fatigue in field operation of hand
tractors. Biosys Eng 89:1–11
2. Mansfield NJ (2005) Human response to vibration. CRC Press LLC, United Kingdom
3. Ko YH, Ean OL, Ripin ZM (2011) The design and development of suspended handles for
reducing hand-arm vibration in petrol driven grass trimmer. Int J Ind Ergon 41:459–470
4. Oddo R, Loyau T, Boileau PE, Champous Y (2004) Design of a suspended handle to attenuate
rock drill hand-arm vibration; model development and validation. J Sound Vib 275:623–640
5. Golysheva EV, Babitsky VI, Veprik AM (2004) Vibration protection for an operator of a
hand-held percussion machine. J Sound Vib 274:351–367
6. Bovenzi M, Della Vedova A, Nataletti P, Alessandrini B, Poian T (2005) Work-related disorders
of the upper limb in female workers using orbital sanders. Int Arch Occup Environ Health
78(4):303–310
7. Stoyneva Z, Lyapina M, Tzvetkov D, Vodenicharov E (2003) Current pathophysiological views
on vibration-induced Raynaud’s phenomenon. Cardiovas Res 57:615–624
8. Griffin MJ, Bovenzi M, Nelson CM (2003) Dose-response patterns for vibration induced white
finger. Occup Environ Med 60:16–26
9. European Parliament and the Council of the European Union (2005) Directive 2005/32/EC on
the minimum health and safety requirements regarding the exposure of workers to the risks
arising from physical agents (vibration). Official J Eur OJL 191:29–58
10. ISO 5349-1 (2001) Mechanical vibration—measurement and evaluation of human exposure
to hand-transmitted vibration. Part-1: general requirements. geneva: international standard
organization
11. ISO 5349-2 (2001) Mechanical vibration-measurement and evaluation of human exposure
to hand-transmitted vibration. Part -2: practical guidance for measurement at the workplace.
Geneva: International Standard Organization
12. Griffin MJ (1997) Measurement, evaluation, and assessment of occupational exposure to hand-
transmitted vibration. Occup Environ Med 54:73–89
13. Griffin MJ (2004) Minimum health and safety requirements for workers exposed to hand-
transmitted vibration and whole-body vibration in the European Union; a review. Occup Environ
Med 61:387–397
14. Su TA, Hoe VCW (2008) Reliability of a Malay-translated questionnaire for use in a hand-arm
vibration syndrome study in Malaysia. Singap Med J 49:1038–1045
15. Dong RG, Wu JZ, Welcome DE (2005) Recent advances in biodynamics of human hand-arm
system. Ind Health 43:449–471
16. Poole K, Mason H (2005) Relationship between self-reported upper limb disability and
quantitative tests in hand-arm vibration syndrome. Disabil Rehabil 29(5):359–366
17. Eksioglu M (2004) Relative optimum grip span as a function of hand anthropometry. Int J Ind
Ergon 34:1–12
18. Tony JARB, Masilamany Santha A (2018) An analytical study—dynamic behavior of the
human hand to hold mechanical handle with and without coating. Vibro Eng Procedia 21:137–
142
19. Alphin MS, Sankaranarayanasamy K, Sivapirakasam SP (2013) Experimental investigation to
study the influence of handle diameter on low-frequency, hand–arm vertical vibration. Hum
Factor Ergon Man 23(2):140–148
20. Tony JARB, Alphin MS (2019) Assessment of ergonomically designed handle shapes for low-
frequency vibration responses. Proc Inst Mech Eng pt L-J Mater Des Appl 233(8):1564–1573
21. Tony JARB, Alphin MS (2019) Finite element analysis to assess the biomechanical behavior
of a finger model gripping handles with different diameters. Biomed Hum Kinet 11:69–79
22. Tony JARB, Alphin MS, Velmurugan D (2019) Influence of handle shape and size to reduce
the hand-arm vibration discomfort. Work. 63(3):415–426

Trends in Mechanical and Biomedical Design: Esther Titilayo Akinlabi P. Ramkumar M. Selvaraj Editors

Uploaded by

Copyright:

Available Formats

Trends in Mechanical and Biomedical Design: Esther Titilayo Akinlabi P. Ramkumar M. Selvaraj Editors

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Trends in Mechanical and Biomedical Design: Esther Titilayo Akinlabi P. Ramkumar M. Selvaraj Editors

Uploaded by

Copyright:

Available Formats

Lecture Notes in Mechanical Engineering

Esther Titilayo Akinlabi

More information about this series at http://www.springer.com/series/11236

ISSN 2195-4356 ISSN 2195-4364 (electronic)

Johannesburg, South Africa Esther Titilayo Akinlabi

Design of Machines and Mechanism

Optimal Dimensional Synthesis of Double-Semi-circle

Fatigue and Fracture Analysis

Dynamic Analysis of Rectangular Aluminum Plate Under Transverse

Finite Element Analysis

Finite Element Analysis of a Two-Post Rollover Protective Structure

Fluid Mechanics and Heat Transfer

Numerical Investigation of Bingham Fluid Flow in the Entrance

Robotics, Biomedical, and Control Engineering

C4–C5 Segment Finite Element Model Development and Investigation

Thermal and Engine Engineering

An Experimental Investigation of Heat Transfer in Heat Exchangers

Tribology, Wear and Surface Engineering

Vibration and Control Engineering

Free Vibration Analysis of Functionally Graded Beam with Linearly

Professor Esther Titilayo Akinlabi is a Full Professor at the Department of

Dr. P. Ramkumar is currently working as an Associate Professor in the Depart-

Dr. M. Selvaraj, Associate Professor in the Department of Mechanical Engi-

Keywords Coupling · Rigid · Logged · Design

© Springer Nature Singapore Pte Ltd. 2021 3

2.1 Logged Coupling

See Figs. 1 and 2.

2.2 Empirical Relations [1]

Fig. 1 Full assembly

Fig. 2 Section of full

Din = 2.5d (6)

Table 1 Values for no. of

b1 Size of outer edge

2.3 Shear Failure of Log

Fig. 3 Maximum shear area

2.4 If Pins Are Loosely Kept [Bending Consideration]

Fig. 4 Bending of pin

2.5 Shear Stress in Connecting Pin

Fig. 5 Shear area of

3 Simulation Data and Results

3.1 General Settings

See Tables 2, 3, 4 and 5.

Table 4 Adaptive mesh

Fig. 6 Von Mises

3.2 Rigid Flanged Coupling

See Tables 6, 7, and 8.

3.3 Rigid Logged Coupling

Loads (Tables 9, 10, and 11).

Fig. 7 Von Mises

4 Dimensions for Construction

h is the height of the log.

4.6 Connector Pin

4.7 Protective Case

4.8 Material Saved as Compared to Rigid Flanged Coupling

Maximum: 15.5 MPa Maximum: 13.65 MPa

1. Design of Machine elements by Bhandari

K. Banumalar, M. Aiswarya, M. Asvanthini Eswari, and M. Karthika

Keywords Grey water · Skimmer · Irrigation control

K. Banumalar (B) · M. Aiswarya · M. Asvanthini Eswari · M. Karthika