University of Calgary

PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2019-12

Performance of Steel Braced Frame Reinforced with

Shape Memory Alloy Wires

Varughese, Kezia Ann

Varughese, K. A. (2019). Performance of Steel Braced Frame Reinforced with Shape Memory
Alloy Wires (Unpublished master's thesis). University of Calgary, Calgary, AB.
http://hdl.handle.net/1880/111419
master thesis

University of Calgary graduate students retain copyright ownership and moral rights for their
thesis. You may use this material in any way that is permitted by the Copyright Act or through
licensing that has been assigned to the document. For uses that are not allowable under
copyright legislation or licensing, you are required to seek permission.
Downloaded from PRISM: https://prism.ucalgary.ca
 UNIVERSITY OF CALGARY

Performance of Steel Braced Frame Reinforced with Shape Memory Alloy Wires

by

Kezia Ann Varughese

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF MASTER OF SCIENCE

GRADUATE PROGRAM IN CIVIL ENGINEERING

CALGARY, ALBERTA

DECEMBER, 2019

© Kezia Ann Varughese 2019

 Abstract

Shape memory alloys have just recently made their way into structural design, and through studies

have proved to be quite effective in increasing a structure’s strength and performance. This thesis

focuses on the research performed on concentrically braced steel frames reinforced with nickel-

titanium (NiTi) shape memory alloy (SMA) wires for use as a seismic resistance force system. The

main objective of this research was to design an innovative self-centring system that had the ability

to dissipate energy and undergo large deformations that would occur during large earthquake

motions. Three frames were designed and tested; one control steel braced frame and two SMA-

reinforced steel braced frames. Examination of the systems under a free vibration test and a quasi-

static cyclic test was conducted to study the effect the SMA wires had on the behaviour of the

braced frames. Initial test results showed that the SMA braced frame had significant re-centring

capabilities, and a potential for moderate energy dissipation. Investigation of the anchorage

connection showed that the wires slipped out of their anchorage prematurely. With the improved

system, it was concluded that the SMA-reinforced frame attained a significantly greater energy

dissipation ability, self-centring response, and overall an enhanced performance than the

conventional steel braced frame.

ii
 Preface

This dissertation is the result of research work completed under the supervision of Dr. Raafat El-

Hacha, at the Schulich School of Engineering, Department of Civil Engineering, University of

Calgary. The following papers have been published during the

course of the research.

Peer-Reviewed Conference Papers

1. Varughese, K.A., and El-Hacha, R., “Steel Braced Frame Reinforced with Superelastic
Shape Memory Alloy”, Proceedings of the 12th Canadian Conference on Earthquake
Engineering, Quebec , Canada, June 17-19, 2019,5p.

2. Varughese, K.A and El-Hacha, R., “Seismic Performance of Steel Braced Frames
Reinforced with Shape Memory Alloy Wires”, CSCE 2019 General Conference, Montreal,
Canada, June 12-15, 2019, 10p.

Varughese, K.A., and El-Hacha, R., “NiTi SMA Steel Braced Frame Design
Considerations and Construction Challenges”, Proceedings of the 7th International
Colloquium on Performance, Protection & Strengthening of Structures Under Extreme
Loading & Events (PROTECT 2019), Whistler, BC ,Canada, September 16-17, 2019, 10p

Peer-Reviewed Conference Abstract

1. El-Hacha, R., Varughese, K.A., “Rocking Frame Reinforced with Superelastic Nickle-
Titanium Shape Memory Alloy”, Proceedings of the 2nd International Conference on
Materials Science and Engineering, Paris, France, February 25-26, 2019, 2p.

iii
 Acknowledgements

First and foremost, I would like to extend my gratitude to my supervisor Dr. Raafat El-Hacha for

his support and guidance throughout this research.

I gratefully acknowledge Dr. Madhar Haddad (UAE University), Dr. Guilherme Parsekian

(Universidade Federal de São Carlos), Dr. Samer Adeeb (University of Alberta), Dr. Mark Green,

Demewoz Meena and Jamie Valeria (Queens University) for their contribution to this research. I

would also like to thank the technical staff in the Department of Civil and Mechanical Engineering

at the University of Calgary; Dan Tilleman, Mirsad Berbic, Terry Quinn, Don Anson, Daniel

Larson, Timothy Williams, Larry Trudeau, and Brandon Ferguson all of whom have gone above

and beyond when offering their assistance and expertise throughout my research. They are highly

appreciated for their help in the laboratory work.

The financial support of the University of Calgary as well as the Natural Sciences and Engineering

Research Council of Canada (NSERC) for this research project is greatly appreciated.

A big thank you to my friends and colleagues Dr. Fadi Oudah, Dr. Khaled Abdelrahman, Dr.

Hothifa Rojob, and Mr. Benjamin Forrest for all of their input and expertise during the course of

this research. I would also like to extend a special thank you to Dr. Markus Dann for his

mentorship, support and encouragement throughout my program.

Finally, I would like to thank my parents, brothers and my precious Leo for their love, support and

prayers throughout this journey.

iv
 Table of Contents

Abstract ........................................................................................................................................... ii
Preface............................................................................................................................................ iii
Acknowledgements ........................................................................................................................ iv
Table of Contents .............................................................................................................................v
List of Tables ............................................................................................................................... viii
List of Figures and Illustrations ..................................................................................................... ix
List of Symbols, Abbreviations and Nomenclature ..................................................................... xiv

Chapter 1: Introduction ................................................................................................................1

1.1 General .................................................................................................................................1
1.2 Idea and Motivation .............................................................................................................5
1.3 Research Significance ..........................................................................................................5
1.4 Objectives ............................................................................................................................7
1.5 Methodology ........................................................................................................................7
1.6 Thesis Content Overview .....................................................................................................7
Chapter 2: Literature Review .......................................................................................................9
2.1 Introduction ..........................................................................................................................9
2.2 Shape Memory Alloy Overview ........................................................................................11
2.2.1 Factors Affecting the Behaviour of Pseudoelastic SMA ...............................................13
2.2.1.1 Effect of Temperature ............................................................................................13
2.2.1.2 Effect of Loading ...................................................................................................16
2.2.1.3 Effect of Size..........................................................................................................19
2.2.1.4 Fatigue Properties ..................................................................................................21
2.3 Seismic Applications .........................................................................................................24
2.3.1 SMA Based Connections ...............................................................................................24
2.3.2 SMA-Based Dampers ....................................................................................................28
2.3.3 SMA Braces...................................................................................................................30
2.4 Further Research ................................................................................................................32
2.5 Summary ............................................................................................................................33
Chapter 3: Experimental Behaviour of PE NiTi-SMA Wires .................................................35
3.1 Introduction ........................................................................................................................35
3.2 NiTi-SMA Material Composition and Characterization....................................................35
3.3 SMA Tensile Testing .........................................................................................................36

v
 3.3.1 Experimental Analysis of Tension Tests .......................................................................39
3.4 Cyclic Behaviour of SMA Wires .......................................................................................42
3.4.1 Experimental Analysis of Cyclic Tests .........................................................................45
3.5 Training SMA Wires..........................................................................................................49
3.6 Tensile Testing of NiTi-SMA Connection Plate ...............................................................51
3.7 Summary ............................................................................................................................55
Chapter 4: Design and Behaviour of Steel Braced Frame with Yielding Fuse ......................57
4.1 Introduction ........................................................................................................................57
4.2 Testing Matrix ....................................................................................................................57
4.3 Design ................................................................................................................................58
4.3.1 Analysis of applied loads...............................................................................................58
4.3.2 Brace Design .................................................................................................................61
4.3.3 Connection Design ........................................................................................................63
4.4 Experimental Program .......................................................................................................65
4.4.1 Test Setup ......................................................................................................................65
4.4.2 Testing Procedure ..........................................................................................................67
4.5 Free Vibrations Test ...........................................................................................................67
4.5.1 Data Acquisition ............................................................................................................69
4.5.2 Results of Free Vibrations Test .....................................................................................69
4.5.2.1 Frequency analysis .................................................................................................69
4.5.2.1 Damping ratio analysis ..........................................................................................75
4.6 Quasi-Static Cyclic Loading Test ......................................................................................77
4.6.1 Test Regime ...................................................................................................................77
4.6.2 Data Acquisition ............................................................................................................78
4.6.3 Results ...........................................................................................................................80
4.6.3.1 CBF Hysteretic Response ......................................................................................80
4.6.3.2 SMA-BF Test 1 Hysteretic Response ....................................................................82
4.7 Comparison of Results .......................................................................................................84
4.7.1 Re-centring ....................................................................................................................84
4.7.2 Energy Dissipation ........................................................................................................86
4.8 Behaviour of Anchorage system ........................................................................................86
4.8.1 Development of Improved Connection System.............................................................87
4.8.2 Behaviour of Improved SMA-BF ..................................................................................89
4.8.2.1 Results of Improved SMA-BF ...............................................................................89

vi
 4.9 Stiffness Degradation .........................................................................................................93
4.10 Loading Rate Assessment ..................................................................................................97
4.11 Ductility Assessment .......................................................................................................101
4.12 Summary ..........................................................................................................................102
Chapter 5: Conclusions and Recommendations .....................................................................104
5.1 Introduction ......................................................................................................................104
5.1 Conclusions ......................................................................................................................104
5.2.1 Free Vibration Loading ...............................................................................................104
5.2.2 Quasi-Static Cyclic Loading........................................................................................105
5.3 Recommendations for Direction of Future Research.......................................................107
5.3.1 Experimental Work .....................................................................................................107
5.3.2 Finite Element Work ...................................................................................................108
References ...................................................................................................................................110
Appendix A .................................................................................................................................125
Appendix B .................................................................................................................................158

vii
 List of Tables

Table 3-1: Chemical Analysis of NiTi-SMA wire (Memry, 2018) ............................................. 35

Table 3-2: Characteristic Properties of NiTi-SMA wire (Memry, 2018) .................................... 36

Table 3-3: Summary of the experimental tensile testing of the NiTi-SMA wires ....................... 42

Table 4-1: Specimen section details............................................................................................. 61

Table 4-2 : Steel Material Properties as per Manufacture ........................................................... 61

Table 4-3: Expected yield and failure loads for yielding fuses.................................................... 63

Table 4-4: Performance specificatoons of accelerometers(www.te.com) ................................... 63

Table 4-5: CBF estimated frequencies ......................................................................................... 70

Table 4-6: SMA-BF estimated frequencies ................................................................................. 70

Table 4-7: Experimental frequencies and damping ratios from free vibration test ..................... 72

Table 4-8:Theoretical and experimental stiffnesses of CBF and SMA-BF systems ................... 73

Table 4-9: Equivalent Viscous Damping results for the CBF and SMA-BF-2 systems .............. 93

Table 4-10: Ductility assessment of the CBF and SMA-BF-2 systems ..................................... 102

viii
 List of Figures and Illustrations

Figure 1-1: Schematic representation of the FBD method for seismically isolated buildings ...... 2

Figure 1-2: Schematic representation of the performance-based method for seismically

isolated building ...................................................................................................................... 2

Figure 2-1: Typical phase diagram of SMA material (Abdelrahman, 2017)............................... 12

Figure 2-2: Stress-strain relationship for PE SMA material (Abdelrahman, 2017) .................... 13

Figure 2-3: Varying surface temperature of a PE SMA wire under uniaxial tension test (Tamai
and Kitagawa, 2002) ............................................................................................................. 15

Figure 2-4: Stress-strain relationship of PE SMA wires under varying temperatures (Pieczyska
et al., 2005) ........................................................................................................................... 15

Figure 2-5: Energy and temperature effect (Pieczyska et al., 2005) ............................................ 16

Figure 2-6: The effect of loading rate on PE SMA wires (Tobushi et al., 1998). ....................... 17

Figure 2-7: The effect of environmental conditions on energy behaviour of superelastic NiTi-
SMA wires (Pieczyska et al., 2005) ...................................................................................... 18

Figure 2-8: Stress-strain relationship of superelastic SMA material with varying diameters
(DesRoches et al., 2004) ....................................................................................................... 19

Figure 2-9: Effect of bar size on behaviour of superelastic SMA material (DesRoches et al.,
2004) ..................................................................................................................................... 20

Figure 2-10: Maximum strain under fatigue loading (Matsui et al., 2006) ................................. 22

Figure 2-11: Test results (Jaureguizahar et al., 2016).................................................................. 23

Figure 2-12: Beam-Column connection reinforced with either steel tendons or NiTi (Speicher
et al, 2011) ............................................................................................................................ 29

Figure 2-13: Reinforced concrete wall details and dimensions (Wang and Zhu, 2017) ............. 25

Figure 2-14: Schematic of rocking column with SMA, adopted from Roh and Reinhorn (2010)
............................................................................................................................................... 26

Figure 2-15: Cracking pattern at ultimate conditions (Oudah, 2014) .......................................... 27

Figure 2-16: Schematic of superelastic shape memory alloy friction damper (Qian et al., 2016)
............................................................................................................................................... 29

ix
Figure 2-17: Configuration of SCD: a) initial condition; b) lower part of wire in tension when
piston moves up; c) upper part of wire in tension when piston moves down (Zhou et al.,
2018) ..................................................................................................................................... 29

Figure 2-18: Schematic of shape memory alloy brace (Qiu and Zhu, 2017)............................... 30

Figure 2-19: Reinforced concrete shear wall with PE NiTi SMA braces (Cortes-Puentes and
Palermo, 2018) ...................................................................................................................... 31

Figure 2-20: Schematic of SMA device (Dolce et al., 2000) ...................................................... 32

Figure 3-1: Typical braced frame configuration .......................................................................... 37

Figure 3-2: Instrumentation and test set-up of the NiTi-SMA tension test ................................. 38

Figure 3-3: DIC and strain gauge application on 2mm diameter SMA wire ............................... 39

Figure 3-4: Strain gauge analysis of the SMA wires subject to tensile loading .......................... 40

Figure 3-5: DIC technique analysis of the SMA wires subject to tensile loading ....................... 40

Figure 3-6: Stress-strain diagram of the strain gauge and DIC technique of the NiTi-SMA
wires subject to tensile loading ............................................................................................. 42

Figure 3-7: Cyclic load test setup ................................................................................................ 43

Figure 3-8: Aluminium sleeves used for gripping PE NiTi SMA wire ....................................... 44

Figure 3-9: Strain-controlled loading history .............................................................................. 44

Figure 3-10: Stress-strain response of PE NiTi SMA wire subject to cyclic loading ................. 46

Figure 3-11: Definition of the parameters used to evaluate the cyclic response of the PE NiTi
SMA wire .............................................................................................................................. 47

Figure 3-12: PE NiTi subject to quasi-static cyclic loading showing: (a) residual strain (εr), (b)
equivalent viscous damping ratio (ζeq), (c) loading plateau stress (σL), and (d) unloading
plateau stress (σUL) ................................................................................................................ 49

Figure 3-13: Stress-strain curve for untrained NiTi SMA wire (McCormick et al., 2005) ......... 50

Figure 3-14: Initial anchorage system ......................................................................................... 51

Figure 3-15: Experimental set-up of anchorage system in MTS machine .................................. 53

Figure 3-16: DICT Load-Strain results of Left and Right Wire groups ...................................... 54

Figure 3-17: LSC Load-Strain Results of anchorage system ...................................................... 54

x
Figure 4-1: Specimen details ....................................................................................................... 58

Figure 4-2: Typical steel stress-strain curve ................................................................................ 59

Figure 4-3: Schematic of partial fixity at column base ................................................................ 64

Figure 4-4: Test set-up for free vibration and quasi-static cyclic tests ....................................... 66

Figure 4-5: Placement of rollers beneath vertical actuators on loading beam ............................. 66

Figure 4-6: Accelerometer set-up for free vibrations test ............................................................ 68

Figure 4-7: Accelerometer response for CBF and SMA-BF systems ........................................ 71

Figure 4-8: In-plane frequency domains for CBF and SMA-BF systems ................................... 72

Figure 4-9: Out-of-plane frequency domains of CBF and SMA-BF systems ............................ 72

Figure 4-10: Response acceleration of CBF under free vibration .............................................. 76

Figure 4-11: Displacement-controlled loading history ................................................................ 78

Figure 4-12: Instrumentation set-up for quasi-static cyclic test for CBF .................................... 79

Figure 4-13:Instrumentation set-up for quasi-static cyclic test for SMA-BF .............................. 79

Figure 4-14: Hysteresis Loop - CBF............................................................................................ 80

Figure 4-15: Fracture surface of steel yielding fuse in CBF........................................................ 81

Figure 4-16: Fracture surface from coupon tension test .............................................................. 82

Figure 4-17: Hysteresis Loop - SMA-BF-1 ................................................................................ 83

Figure 4-18: Residual strain in the brace after the last cycle at each drift level .......................... 85

Figure 4-19: Slipped wires in anchorage system ......................................................................... 85

Figure 4-20: Energy dissipated over a range of drift levels for CBF and SMA-BF-1 ................ 86

Figure 4-21: Initial behaviour of SMA-BF .................................................................................. 87

Figure 4-22: Newly developed anchorage system ....................................................................... 88

Figure 4-23: Installed anchorage system and slotted connection on SMA-BF ........................... 89

Figure 4-24: Hysteresis response of SMA-BF-2 ......................................................................... 90

Figure 4-25: Failed gusset plate at approximately 3.5% drift...................................................... 91

xi
Figure 4-26: Uneven load distribution due to failed gusset plate ................................................ 91

Figure 4-27: Energy dissipated over a range of drift levels ......................................................... 92

Figure 4-28: Stiffness degradation of CBF and SMA-BF systems ............................................. 94

Figure 4-29: Different stiffness degradation experienced by SMA-BF systems ......................... 95

Figure 4-30: Part 1 of the stiffness degradation in the SMA-BF systems ................................... 96

Figure 4-31: Placement of thermocouples on NiTi SMA wire .................................................... 97

Figure 4-32: Displacement-Time relationship at 0.5% strain ...................................................... 98

Figure 4-33: Relationship between Displacement, Load and Time at 0.5% strain ..................... 99

Figure 4-34: NiTi SMA wire returning to original shape ............................................................ 99

Figure 4-35: Rate of change of NiTi wire vs. strain relationship .............................................. 100

Figure B.1: Coupon test specimen details ................................................................................. 158

Figure B.2: Coupon test fracture pattern ................................................................................... 158

Figure B.3: Tensile stress-strain behaviour of steel coupon specimens .................................... 159

Figure B.4: FE model of steel coupon test................................................................................. 160

Figure B.5: Comparison of experimental and FE model stress-strain behaviour ...................... 161

Figure B.6: Symmetric coupon test ........................................................................................... 162

Figure B.7: FE model of CBF .................................................................................................... 165

Figure B.8: Mesh of CBF .......................................................................................................... 165

Figure B.9: Load-Displacement Hysteresis Behaviour of CBF ................................................ 166

Figure B.10: Comparison between experimental and FE model hysteresis response of CBF .. 166

Figure B.11: FE model of partial CBF system .......................................................................... 167

Figure B.12: Mesh of partial CBF system ................................................................................. 167

Figure B.13: Hysteresis behaviour of partial CBF system ........................................................ 168

Figure B.14: Only braces modelled ........................................................................................... 169

Figure B.15: Mesh of just braces ............................................................................................... 169

xii
Figure B.16: Deformed shape of just braces model................................................................... 170

Figure B.17: Force-Displacement result of just braces model................................................... 171

xiii
 List of Symbols, Abbreviations and Nomenclature

Symbol Definition
A Cross sectional area
Af Austenite finish temperature
As Austenite start temperature
c Coefficient of friction of bolt
C Carbon
CBF Control Braced frame
Co Cobalt
Cr Chromium
Cu Copper
D Diameter of bolt
Di Target displacement at level i
E Elastic modulus
EA Elastic modulus of Austenite phase
ED Energy dissipated in a cycle
ED Energy dissipated in a cycle
EM Elastic modulus of Martensite phase
Er Recoverable energy
ES0 Elastic energy in a cycle
ESo Elastic energy in a cycle
F Axial force
FA Finish Austenite
Fd Design force
Fe Iron
Fi Force at level i
FM Finish Martensite
Fr Frictional force
H Hydrogen
He Height of equivalent system
IE Importance factor for earthquakes
k Stiffness
Ke Equivalent lateral stiffness
LLRS Lateral load resisting system
m Mass
Md Detwinned Martensite
Me Mass of equivalent system
Mf Martensite finish temperature
Ms Martensite start temperature
Mt Twinned Martensite
Mv Factor to account for higher mode effects on base
shear
N Total clamping force
N Nitrogen
Nb Niobium

xiv
Nf Fatigue life
Ni Nickel
O Oxygen
Rd Ductility modification factor
Ro Overstrength modification factor
S(Ta) Design spectral response acceleration for a
period T
SA Start Austenite
SFRS Seismic force resisting system
SM Start Martensite
SMA-BF-1 Shape memory alloy braced frame test 1
SMA-BF-2 Shape memory alloy braced frame test 2
Sr Stress ratio
T Torque; Temperature
Ta Fundamental lateral period of vibration of the
structure; Applied Temperature.
TD Damped period
Ti Titanium
Tn Natural period
üi Peaks of acceleration-time graph
V0 Base Shear
Vi Shear at level i
W Dead load of the structure
w Loading frequency
Wd Dissipated/absorbed energy
Δd Design displacement
Δi Displacement at level i
Δu Displacement at ultimate
Δy Displacement at yield
εMs Martensite start strain
εr Residual strain
εu Ultimate tensile strain
ζ Viscous damping ratio
ζeq Equivalent viscous damping ratio
θ Angle
μ Coefficient of friction of steel
σAf Austenite finish stress
σAs Austenite start stress
σf Finish stress transformation
σL Loading plateau stress
σMf Martensite finish stress
σMs Martensite start stress
σs Start stress transformation
σUL Unloading plateau stress
ωD Damped frequency
ωn Natural frequency

xv
 1

Chapter 1: Introduction

1.1 General

Numerous earthquakes occur around the world every year, impacting thousands of lives.

Infrastructure is significantly impacted, cities lose their ability to function as bridges and important

buildings, such as hospitals, are destroyed; costing the city billions of dollars in damage. Currently,

seismic design allows for structures to experience plastic deformations to dissipate the seismic

energy. However, many buildings may still be severely damaged, beyond repair, and need to be

demolished. Development of seismic engineering techniques is constantly changing, and these

techniques will never eliminate the impacts of an earthquake event. However, it is possible to

minimize the damage done so that little to no damage is done on the main structural members.

The seismic design philosophy is to ensure that the structure undergoes controlled ductile

behaviour to avoid collapse during a design-level-earthquake. Designers allow this ductile failure

at certain locations; these locations are called plastic hinges. These plastic hinges undergo inelastic

deformations that allow for the earthquake energy to be dissipated. When the designer is choosing

the locations of the plastic hinges, careful consideration of undesirable inelastic deformation

mechanisms, such as shear failure, should be avoided with proper strengthening to ensure a

flexural failure (Priestley et al., 1996).

Over the past couple of decades, design for seismic resistance has changed from a Strength-Based

Design approach to a Performance-Based Design. In traditional Force-Based Design (FBD)

approach in terms of seismic design, safety and collapse prevention were the number one priority.

However, Park and Paulay realized that a structure would perform better if the strength was
 2

distributed throughout the building rather than designing it based on the absolute base shear (Park

and Paulay, 1975). Furthermore, it was recognized that if plastic hinges were to form in beams

rather than columns, the building would perform much better. These parameters can be identified

as the basis for performance-based design (Priestley, 2000).

(a) (b)

Figure 1-1: Schematic representation of the FBD method for seismically isolated buildings

(a) (b) (c) (d)

Figure 1-2: Schematic representation of the performance-based method for seismically isolated
building

Figures 1-1 and 1-2 highlight the important differences between the FBD and performance-based

methods. The FBD method, Figure 1-1(a), is modelled as an equivalent linear single-degree-of-
 3

freedom (SDOF) system with an effective period of vibration and equivalent damping ratio; in this

method the initial stiffness of the structure is assumed. With the period of vibration and equivalent

damping ratio, a design acceleration is found; with this acceleration a design base shear (V0) at

ultimate state can be determined (Figure 1-1(b)) and the structure can then be designed

accordingly. The interstorey drifts are checked at the end of the design. If the drifts exceed the

code limits, the process is repeated until a suitable design can be found resulting in an iterative

process. Problems with the conventional force-based seismic design are the interdependency of

strength and stiffness that lead to a long iterative design process, the distribution of lateral forces

based on elastic estimates of stiffness leads to concentrated strength in inappropriate members,

inaccuracy of determining the fundamental period based on initial elastic periods, and the

inconsistency of determining the ductility capacity factor for different types of structures (Priestley

et al., 2007).

The known performance-based design approach is called Direct Displacement Based Design

(DDBD), Figure 1-2. The philosophy behind DDBD is that structures can be designed to achieve

a specified performance level, defined by either strain or drift limits, under a specified seismic

intensity. The main objective is to allow for a structure to experience a stable and reliable hysteresis

response with high levels of energy dissipation, whilst having a lower stiffness. The multiple

degree of freedom (MDOF) system (Figure 1-2(a) and (b)) is modelled as an equivalent linear

SDOF system with an equivalent lateral stiffness and the equivalent viscous damping (Figure 1-

2(c) and (d)) of the real structure at the peak displacement response. Here, the flexibility of the

structure is taken into consideration. The displacement profile of the structure is specified by the

engineer at the beginning of the design, to ensure a specified performance level (Cardone et al.,
 4

2008). It has been found that damage of a structure is better correlated to the displacements rather

than the forces it experiences; this is why engineers have moved away from a strength-based

approach (Priestley et al., 2007; Priestley, 2000). Furthermore, fundamental problems with the

FBD method have been taken into consideration when developing the DDBD method (Priestley et

al., 2007). When designing systems using the DDBD approach it is preferable to use smart

materials with low modulus of elasticity, such as certain fibre reinforced polymers (FRP) and shape

memory alloys (SMA), not only to allow for such displacements but to act as part of the

reinforcement at the plastic hinges.

Currently, most standards, including the National Building Code of Canada (NBCC 2015), specify

that the design of buildings follow the FBD approach. As stated in NBCC Commentary J Design

for Seismic Effects, the main objectives of seismic design are as follows:

1. To protect the life and safety of building occupants and the general public as the building

responds to strong ground shaking,

2. To limit building damage during low to moderate levels of ground shaking, and

3. To ensure that post-disaster buildings can continue to be occupied and functional following

strong ground shaking, though minimal damage can be expected in such buildings.

There is a minimum performance level that buildings must meet, however, the NBCC deems the

performance level as “extensive damage”, meaning that although the structure may be very heavily

damaged and may have lost a substantial amount of its initial strength and stiffness, it still retains

some margin of resistance against collapse. The primary objective is to minimize the loss of life.

However, designing to this bare minimum requirement leads to enormous economic loss when a

large earthquake strikes. Having a structure that is able to re-centre itself once the earthquake
 5

excitation ceases would be very beneficial. This leads to using innovative materials that have the

ability to dissipate large amounts of energy as well as the ability to experience zero residual strain,

such as SMA.

1.2 Idea and Motivation

In structural engineering, the number one priority is always the human life factor. The goal is to

design and create a safe and reliable structure; this same mindset especially applies for structures

in seismic areas. Recently, seismic design has shifted focus from solely ensuring a structure can

remain standing until the earthquake motions stop, regardless of the damage done, to focussing on

minimizing the amount of damage done using advanced materials (Bradley et al., 2008; Mander

and Cheng, 1997; Cimellaro et al., 2007).

SMA is a unique material with the ability to undergo large strains and revert to its original shape

with zero residual strain with either stress removal or through heating. In the research described in

this thesis, I propose to reinforce a steel braced frame with pseudoelastic (PE) NiTi SMA. The PE

NiTi wires are intended to relocate the plastic hinge to the location of the SMA material so that

they can dissipate the demand energy, and when the load is removed, re-centre the frame.

1.3 Research Significance

When an earthquake strikes, significant damage can be done to a structure, whether it be concrete,

steel, masonry, or timber. In seismic design, there have been numerous methods used to combat

the effects of an earthquake: either a passive or active system is used. Passive systems are those

that do not require any additional input energy to dissipate the seismic energy. Devices that have

been commonly used are viscous fluid dampers, viscoelastic solid dampers, friction dampers, and
 6

metallic dampers (Symans et al., 2008). Active systems are those that impose an additional force

on the structure to counteract the earthquake forces. Active systems, such as prestressing and

computer-based sensors are relatively new and complex, and have not been widely used. The

drawback of passive dissipating systems is either the device would need replacing after an

earthquake event, or the base shear increases and the design of the structure would need to

accommodate this increase (Symans et al., 2008; Trindade et al., 2014) which leads to a larger

stiffness and increased material costs.

Fibre reinforced polymers (FRP) have been studied over the past couple of decades for their use

in seismic retrofitting. FRPs have a high strength-to-weight ratio, are durable and resistant to

corrosion, making them ideal for retrofitting any type of structure. It is not possible to design a

concrete structure to experience no damage at all when subjected to earthquake loads. However,

with proper retrofitting/design it is possible to improve the seismic performance to minimize

damage done to the structure. Though FRPs are a viable option for seismic strengthening of

structures, there are other drawbacks such as: cost, the structure will still experience moderate

damage, and efficiency of the material.

Though there are options for seismic designs, there is still a pressing need for more effective

techniques. The use of SMA in combination with a steel brace as proposed in this research study

will not only result in the frame remaining functional after an earthquake strikes but will also re-

centre the structure and minimize the structural damage so that no structural members need

replacing after the earthquake motions cease.

 7

1.4 Objectives

The main objective of this research is to investigate the effectiveness and feasibility of active

techniques for seismic retrofitting of steel frames using PE NiTi based SMA reinforcement. In the

course of this study, the main objective was sub-divided into the following sub-objectives:

• Review the up-to-date research on the rocking structures method and the factors that affect

the behaviour of SMA material.

• Characterize the monotonic and cyclic properties of SMA wires for their use in seismic

design.

• Design and manufacture a suitable connection for the wires to the steel brace.

• Conduct a small-scale test of the SMA to steel brace connection to determine the success

of the connection.

• Design and experimentally validate a model steel frame to undergo different loading tests

including: Free Vibrations test and Quasi-Static Cyclic loading to investigate the behaviour

of the frame reinforced with the SMA braces.

1.5 Methodology

The experimental approach used in this research involved two types of tests, in-plane free vibration

and quasi-static cyclic loading to achieve the objectives of this research. Three steel braced frames

were constructed; one with a steel yielding fuse and two with pseudoelastic SMA wires as the

yielding fuse. Design calculations for the braced frames can be found in Appendix A.

1.6 Thesis Content Overview

Following the research significance and objectives, the contents of this thesis consists of 5

chapters.
 8

• Chapter 2. Background information on SMA’s and the literature review on the research

conducted using SMA material in seismic applications is presented in this chapter.

• Chapter 3. Experimental testing of the SMA wires to obtain the material and physical

properties is included in this chapter. The development of the SMA connection plate to the

steel brace is also explained in this chapter.

• Chapter 4. The design and experimental validation of the steel and SMA braced frames are

included in this chapter.

• Chapter 5. The main conclusions and recommendations for future work are included in

this chapter.

• Appendices. There are two appendices provided. Appendix A provides some information

on the detailed design of the specimens. Appendix B provides information of the finite

element modelling that was started in this phase to be used in future research.
 9

Chapter 2: Literature Review

2.1 Introduction

Every year there are multiple earthquakes that occur around the world: thousands of lives are

impacted; entire societies destroyed in a matter of minutes. Currently, new buildings in areas of

high seismic activity are designed to resist these loads. However, there are plenty of existing

structures that are not designed to resist seismic loads. For structural engineers, the human life

factor plays a huge role in how structures are designed. It is our responsibility to design and/or

rehabilitate structures to be as safe as possible whether it be at home or abroad.

Conventional seismic design systems rely on the inelastic behaviour of certain members to

dissipate the seismic energy. Engineers choose and design these members to prevent collapse of a

building; however, the structure will still undergo severe damage and these members will need to

be replaced after each event. Several techniques/materials have been used to retrofit RC columns,

with the commonly used one being concrete and steel jacketing. Concrete jacketing requires

intensive preparation of formwork and increases the weight and size of the retrofitted member.

Steel jacketing is a labour-intensive technique, costly, heavy, and prone to corrosion. However,

these techniques are deficient in the sense that the materials themselves are subjected to the same

sort of deterioration that affects the original structure (Karbhari and Eckel II, 1994). As an

alternative, wrapping columns using epoxy-bonded Fiber Reinforced Polymers (FRP) sheets was

found to be more efficient due to their higher strength-to-weight ratio, ease of application,

durability, and corrosion resistivity. The FRP system is unstressed at the time of installation and

functions through a passive confinement mechanism that is only effective once the concrete

experiences significant damage. Thus, for the system to be used more efficiently it must be an
 10

active system by prestressing the FRP sheet but that is considered impractical, problematic, labour

and cost intensive. There is also the challenge in developing an anchorage system for the

prestressing operation; few attempts have been made (Nesheli and Meguro, 2006; Tam, 2007).

Nesheli and Meguro (2006), wrapped earthquake-damaged columns with pre-tensioned FRP

sheets to restore the lateral strength in the column. They found that the lateral capacity dropped

20% and the efficiency of the wrapping was dependent on the severity of the damage done to the

column. In other words, if the column was severely damaged, in order for the wrapping to be

effective it would need to be combined with another type of repair operation. Over the past couple

of decades, techniques such as damage avoidance design (DAD) (Mander and Cheng, 1997;

Bradley et al., 2008) and resilience-based design (Cimellaro et al., 2007)has focused on

minimizing damage using advanced materials. Park and Paulay realized that a structure would

perform better if the strength was distributed throughout the building rather than designing it based

on the absolute base shear (Park and Paulay, 1975). They found that for moment frames, if the

plastic hinges were formed in beams rather than columns, the building would perform much better.

These parameters can be identified as the basis for performance-based design, which focuses on

displacement rather than forces; the approach is known as the Direct Displacement Based Design

(DDBD) (Priestley, 2000). The DDBD allows for a structure to experience a stable and reliable

hysteresis response with high levels of energy dissipation, whilst having a lower stiffness. Studies

have shown that the damage a structure experiences during a seismic event can be better correlated

to the displacements rather than the forces it experiences (Priestley et al., 2007; Cardone et al.,

2008). Therefore, when designing systems using the DDBD method, a more flexible system is

desired and using smart materials with a low modulus of elasticity will allow for such

displacements.
 11

Recently, a new class of smart materials called nickel-titanium Shape Memory Alloys (NiTi-

SMA), with considerably lower cost, has been attracting researchers from different fields (Dong

et al., 2009; Li et al., 2013; Tarzaz and Saiidii, 2015; Rojob and El-Hacha, 2015). The SMA is a

unique class of alloy with the ability to undergo large deformations (up to 6%) and return to its

original shape through stress removal (DesRoche et al., 2004). Limited studies have shown that

using PE NiTi-SMA wires provide active confinement to RC columns (Andrawes, 2010; Choi et

al., 2008; Park et al., 2011; Shin et al., 2010, Abdelrahman, 2017) and produced significantly

ductile structural members. This highly improves the performance of structures, specifically those

located in seismically active regions and makes SMA attractive materials as braces, base isolators,

damage repair devices and bridge restrainers (Cardon et al., 2004; Ocel et al., 2004; Soroushian et

al., 2001). The addition of the active control system that dissipates the energy from the earthquake

excitation is highly dependent on the structure’s ability to deform: in other words, if a structure is

very stiff under seismic loading the dampers would not be very effective (Ozkaynaka et al., 2011).

The goal of this chapter is to look at how SMAs have been used in seismic design and assess the

possibility of creating a design that could potentially eliminate structural damage without having

to replace the damper after each seismic event.

2.2 Shape Memory Alloy Overview

Shape memory alloys are a unique material that can be deformed from either temperature induced

(shape memory effect) or stress induced (pseudoelasticity (PE) effect) loading (Ozbulut et al.,

2011). The pseudoelastic effect is defined as the ability of the material to undergo hysteretic

response when mechanically loaded under constant temperature. The shape memory effect is

defined as the ability of the material to recover its original shape after being deformed in a
 12

temperature cycle, however after unloading there is residual strain. Therefore, in order to take full

advantage of the SMA’s properties, the behavioural characteristics of the PE will be studied.

The unique characteristic of SMA’s is that they have different phases based on the induced

temperature and mechanical loading. The material can be in two phases, at high temperatures the

material is in the austenite phase (A), and at lower temperatures the martensite phase (M). Within

the martensite phase, the material is either twinned or detwinned; where twins are typically defined

as two or more crystalline domains which are related by an orientation relation that occur

frequently (Christian, 2001). These different phases are clearly shown in Figure 2-1,

(Abdelrahman, 2017).

Figure 2-1: Typical phase diagram of SMA material (Abdelrahman, 2017)

In the stress-free state, the four characteristic phases are as follows: the martensite start temperature

(Ms), the martensite finish temperature (Mf), the austenite start temperature (As), and the austenite

finish temperature (Af). The PE effect is exhibited when the material is stressed at a temperature

greater than Af. Then the material starts to transform into martensite when it is stressed above Ms
 13

and then fully martensite at Mf. When the material is unloaded, it starts to transform back into

austenite at As and is then in fully austenite when the specimen is unloaded to Af.

The stress-strain hysteresis of a SMA specimen is shown in Figure 2-2. This figure shows the PE

of SMAs and clearly demonstrates that the material makes a full recovery with zero residual strain,

making it ideal for cyclic loading conditions.

Figure 2-2: Stress-strain relationship for PE SMA material (Abdelrahman, 2017)

2.2.1. Factors Affecting the Behaviour of Pseudoelastic SMA

2.2.1.1 Effect of Temperature

The behaviour of SMA’s is dependent on both the thermal and mechanical properties of the

material; this implies that whether the material is heated or cooled it will change the response of

the pseudoelastic characteristics in terms of the transformation stresses: σMs, σMf, σAs, and σAf. In-

depth analysis of the thermal effect on the strength properties is crucial in determining their use in

seismic design as knowing the actual strength of the member is considered to be a very important

parameter in addition to the stiffness and ductility properties (Paulay and Priestley, 1992; Elnashai

and Di Sarno, 2008).

 14

As the phase transformations take place, energy is either released or absorbed by the material. The

forward and reverse transformation of martensite takes place through fronts moving along the

material. These fronts act as heat sources which causes the material to undergo exothermic and

endothermic reactions during the forward and reverse transformations respectively. This energy is

either released into the surrounding environment or stored in the material (Lin et al., 1994; Soul et

al., 2010). When the SMA is loaded above Af (superelastic behaviour) the temperature of the

material will increase, whilst unloading will reduce it depending on the applied loading rate.

Consequently, two identical pieces of material may experience different hysteretic responses

depending on the surrounding environmental conditions. Careful consideration of whether the

testing environment is controlled or uncontrolled should be taken into account. The uncontrolled

temperature condition relates to applications where the SMA is used in air without airflow while

the controlled temperature condition corresponds to the SMA being exposed to high velocity air

flow or used in water (Pieczyska et al., 2005). The behaviour of SMA in concrete structures is in

between controlled and uncontrolled conditions as the thermal conductivity of concrete is around

150 times that of dry air (Morabito, 1989; Hilsenrath, 1955).

Tamai and Kitagawa (2002), studied the effect that varying surface temperature has on an SMA

wire while undergoing uniaxial testing, the results are shown in Figure 2-3. It was found that during

the martensitic forward transformation and reverse transformation, the pattern is almost identical

as well as the net temperature change is almost zero (Tamai and Kitagawa, 2002).
 15

Figure 2-3: Varying surface temperature of a PE SMA wire under uniaxial tension test (Tamai

and Kitagawa, 2002)

Pieczyska et al., (2005), studied the effect temperature has on the stress-strain response of SMA

wires in a controlled environment. As seen in Figure 2-4, with increasing temperature the

transformation stresses increase. This can be attributed to the positive slope of the transformation

lines (refer to Figure 2-1), the increase in the applied temperature results in higher characteristic

stresses in relation to the slopes of these lines.

Figure 2-4: Stress-strain relationship of PE SMA wires under varying temperatures (Pieczyska
et al., 2005)
 16

To use PE SMAs as a damper it is important to understand the relationship between temperature

and energy dissipation of the material (Pieczyska et al., 2005). The area within the hysteresis loop

of a stress-strain curve represents the dissipated/absorbed energy, Wd, while the area under the

unloading curve represents the recoverable energy, Er. This study showed that Wd increases

linearly with increasing temperature; where Er is almost constant for temperatures below Af and

then increases linearly for temperatures above Af.

a) Definition of energies b) Energy relationships with temperature

Figure 2-5: Energy and temperature effect (Pieczyska et al., 2005)

2.2.1.2 Effect of Loading

In the seismic design of structures, depending on the location of the structure, the structure will

either be subjected to low or high frequency dynamic vibrations. Therefore, it is important to study

the effect the rate of applied mechanical loading has on PE NiTi SMAs. The working seismic

loading results in strain rates in the range of 5% s-1 to 8% s-1 (Tamai and Kitagawa, 2002), and

thus, dynamic effects should also be considered in relation with thermal effects (Tobushi et al.,

1999). The effect of the strain rate on the PE behaviour of SMA wires before and after training

under controlled temperatures was investigated by Tobushi et al., (1998). Figure 2-6 shows that
 17

the slopes of the stress-strain curves in the martensitic forward and reverse transformations

increase with increasing strain rate,while the austenitic transformation stresses are decreased. From

these two curves, it can be determined that training does not alter the trends of the transformation

stresses.

a) Untrained b) Trained
Figure 2-6: The effect of loading rate on PE SMA wires (Tobushi et al., 1998).

Under uncontrolled conditions, the properties of the PE NiTi SMA differ than those in the

controlled conditions, where uncontrolled conditions were where the ambient temperature of the

room was not controlled (Pieczyska et al., 2005). Pieczyska et al., (2005), investigated this effect

and found that under low strain rates, the hysteresis loop exhibits the typical flag shape experienced

in PE behaviour. However, when the same wire is subjected to high strain rates, the behaviour

changes significantly when compared with the same wire tested under controlled conditions as

shown in Figure 2-7. When looking at the effect the increasing strain rate has on the energy

dissipated Wd and the recoverable energy Er, it was found that in an uncontrolled testing

environment, Er increases with increasing strain rate, while Wd decreases. This is because of the

heat generated in the martensitic transformation; there is not enough time for the heat to be

transferred into the air, therefore there is a large temperature variance within the specimen which
 18

leads to a large stress variance. It is this variance that results in the Er increase and the Wd decrease

under high strain rates.

Figure 2-7: The effect of environmental conditions on energy behaviour of superelastic NiTi-
SMA wires (Pieczyska et al., 2005)
In general, for SMA material, the temperature is coupled with the transformation stress, and thus

the dynamic effects are more pronounced in SMAs than other metals. Under a controlled testing

environment (isothermal case), the thermal effect is isolated, and the behaviour change is solely

due to the effect of the dynamic loading. However, in an uncontrolled testing environment, the

material is subject to both thermal and dynamic effects. The change in the behaviour is more

significant in the case of high rates of loading since the time is very short for the heat generated

by the martensitic transformation to be transferred into the air (Pieczyska et al., 2005).

It should be noted that under very high strain rates, the dissipated energy per cycle does not

experience significant variation, regardless of the temperature increase due to the dynamic effects,

and thus, the damping properties of the PE NiTi SMA should not deteriorate under dynamic

loading (Tamai and Kitagawa, 2002). DesRoches et al. (2004), states that though the dynamic

loading has a significant effect on the behaviour of SMA, its ability to return to its original form

does not depend on the loading rate.

 19

2.2.1.3 Effect of Size

Many of the early applications of SMA were for biomedical devices and electrical and mechanical

engineering components; these only required wires with diameters that ranged from 0.026 mm to

2.7 mm (Soul et al., 2010; Chen and Schuh, 2011). Up until recently, there was no motivation to

investigate large diameter bars. Additionally, the high cost of SMAs hinders their ability to become

a common construction material.

Limited studies have been conducted to study the effect large diameter bars have on the behaviour

of PE SMA (DesRoches et al., 2004; Soul et al., 2010). DesRoches et al., (2004) examined the

behaviour of PE SMA bars with diameters that ranged from 1.8 mm to 25.4 mm, loading each up

to 6% strain. The stress-strain relationship is shown in Figure 2-8. It should be noted that the

samples were all heat treated and tested under identical conditions. It was found that the hysteresis

loop is largest for the 1.8 mm diameter and lowest for the 12.7 mm bar. This suggests that the

energy dissipated is greater for a smaller diameter, attributing to the fatigue principle that larger

surface areas have more material discontinuities and thus have limited abilities.

Figure 2-8: Stress-strain relationship of superelastic SMA material with varying diameters
(DesRoches et al., 2004)
 20

Figure 2-9 shows the loading plateau stress (σMs), unloading plateau stress (σAf), residual strain,

and the equivalent viscous damping versus the maximum cyclic strain. The residual strain for all

the bars was less than 0.75% after being subjected to 6% cyclic strain, which suggests that the

ability of the PE bars to return to their original form is independent of the bar diameter.

Figure 2-9: Effect of bar size on behaviour of superelastic SMA material (DesRoches et al., 2004)

Soul et al. (2010) suggest that the response of the NiTi SMA due to the size effect can be

understood by looking at the surface-to-volume ratio. They indicated that the surface-to-volume

ratio was proportional to 4/db (where db is the bar diameter). Small diameter wires have a high

surface-to-volume ratio, therefore the heat generated during the martensitic forward and reverse

transformation is transferred more effectively resulting in a larger hysteresis loop. Larger

diameters have a smaller surface-to-volume ratio, and consequently the heat transfer is not as

effective as small wires. The results showed the 1.8 mm diameter wire has a steep loading and
 21

unloading transformation plateau because the increase in the heat leads to the increase in the

transformation stresses.

2.2.1.4 Fatigue Properties

It is important that when designing a structure to undergo seismic loading, careful consideration

of the fatigue properties of the design elements should be taken. In concrete design, reinforcing

bars should be designed to resist high amplitude cyclic loading. Therefore, their behaviour under

low-cyclic fatigue loading should be examined. For SMA's there are two types of fatigue, structural

fatigue and functional fatigue. Structural fatigue is defined as the fracture of the SMA under cyclic

loads. Functional fatigue refers to the degradation of the properties of the material such as loss in

energy dissipation and ability to return to undeformed length. Unfortunately, fatigue testing of

SMAs under seismic loading is very limited (Maletta et al., 2012).

The fatigue testing of shape memory alloys has been conducted through various methods such as,

alternating-plane bending (Furuichi et all, 2003; Matsui et al, 2006), rotating-bending (Tobushi et

al., 2000; Matsui et al., 2006; Figueiredo et al., 2009), pulsating-plane bending (Matsui et al.,

2006), and uniaxial fatigue tests (Tabanli et al., 1999; Moumni et al., 2005; Maletta et al., 2012).

In Matsui’s study, PE SMA wires were subject to pulsating-plane bending fatigue loading at a

frequency of 500 cpm at room temperature with varying stress ratios R (minimum stress/maximum

stress). Figure 2-10 shows that the fatigue life is higher when R is higher. It is interesting to note

that the endurance limit is 0.65%, which is close to the strain initiation of the martensitic

transformation. This suggests that if the PE wire is strained below this limit, theoretically, it can

be cycled for an infinite amount of cycles, assuming no flaws exist in the specimen.
 22

Figure 2-10: Maximum strain under fatigue loading (Matsui et al., 2006)

Another study was done to examine the structural fatigue life of SMA wires. Here, Jaureguizahar

et al. (2016) had two main steps: first, stabilize the PE behaviour by performing low speed

isothermal cycles in a limited region. This creates a “virtual” dog-bone shape specimen due to the

reduced transformation stresses in the pre-cycled region. Second, apply the cyclic loading to

determine the fatigue life of the SMA in a purely martensite state and in a purely austenite state.

Results from this study, as seen in Figures 2-11(a) and 2-11(b), showed that once the strain was

above the initial initiation of the martensitic transformation strain, fatigue life was not dependent

on the applied strain range or the mean strain, but on the number of cycles of transformations the

wire underwent. Fatigue lives of the 0.5 mm diameter wire was shown to be around 10,000 cycles,

failure due to fatigue at this number are typically known as low-cyclic fatigue.
 23

(a) Stress vs Strain results (b) Fatigue life resultsFigure 2-11:

Cyclic test results of SMA wires (Jaureguizahar et al., 2016)

Moumni et al., (2005) applied the energy approach in examining the fatigue life of PE SMA’s

subjected to uniaxial loading. They found that the fatigue life was dependent on the R (or S) value.

The stress amplitude versus the fatigue life was plotted and it was clear that for the same stress

amplitude, the fatigue life is greater when R=-1 (mean stress is zero). This is because of the

beneficial effect compressive loading has on fatigue behaviour because it tends to close

microcracks. A linear trend was found for the fatigue life and the dissipated energy for all tested

R values.

DesRoches et al. (2004) looked at the cyclic properties of shape memory alloy wires and bars to

evaluate their potential use for seismic resistant design applications. They found that cyclic strains

greater than 6% led to degradation in the damping and self-centring properties; in other words the

SMA started to experience functional fatigue when the strain was greater than 6%.

More recently, Sherif and Ozbulut (2019) investigated the functional fatigue characteristics of a

PE NiTi cable. The cable was composed of multiple strands with an outer diameter of 5.5 mm;

they subjected the cable to cyclic tension loading under various strain amplitudes. They found that
 24

after the first 100 cycles the superelastic properites of the NiTi cable stabilized and the cable was

able to sustain a very high functional fatigue. The energy dissipation capabilities was highest under

5% loading strain with little degradation throughout the cycles.

2.3 Seismic Applications

Over the last couple of years, a significant amount of research has been done in regard to the use

of SMAs for seismic applications. Researchers have investigated SMA-based dampers (Dolce et

al., 2007; Ma and Cho, 2008; Shook et al., 2008; Casciati and Favavelli, 2008, 2009; Ozbulut and

Silwal, 2016; Qian et al., 2016), SMA-based structural connections (Speicher et al., 2011; Fang et

al., 2014; Oudah, 2014; Wang et al., 2017) and SMA braces (Asgarian and Moradi, 2011; Araki

et al., 2016; Qiu and Zhu, 2017; Cortes-Puentes and Palermo, 2018; Sultana and Youssef, 2018).

Recent studies have shown that the use of SMA in RC members significantly reduces

displacements and improves the overall seismic performance of the structure (Johnson et al., 2008;

Shin and Andrawes, 2011; Billah and Alam, 2015; Shrestha and Hao, 2016). The following are

some cases that demonstrate the benefits of using SMA in seismic designs.

2.3.1. SMA-Based Connections

As stated previously, there is a need to design and rehabilitate structures to be more robust under

seismic loads. One way researchers have found to initiate such structures, is by designing more

ductile connections. Speicher et al. (2011) had the goal of creating a simple ductile recentring

system so they tested a steel beam-column connection with steel tendons and NiTi tendons

embedded at the connection point, refer to Figure 2-12. They found that the NiTi SMA tendons

were able to recover 85% of their strain when cycled to 5% drift; whereas the steel tendons lost
 25

their recentring abilities at a 1.5% drift. The SMA system proved to be very effective as the

connection was able to reach a 5% drift with little residual deformations.

Figure 2-12: Beam-Column connection reinforced with either steel tendons or NiTi (Speicher et
al., 2011)
SMA connections are not only for beam-column joints but they have been applied to shear walls.

Wang and Zhu (2017), placed SMA bars at the base and at the top of the wall, where the plastic

hinge would typically develop as seen in Figure 2-13. When the wall was subjected to the loading,

it was able to reach a peak drift of 2.5% and a maximum strain of 3.3% in the SMA bar with almost

zero residual deformation.

Figure 2-13: Reinforced concrete wall details and dimensions (Wang and Zhu, 2017)
 26

In 2010, Roh and Reinhorn (2010) studied the hysteretic behaviour of rocking segmental bridge

piers reinforced with SMA bars. A rocking column is a type of double hinged column where it is

only connected through compression at the contact surfaces, it has been shown to reduce the

maximum acceleration and thus the associated forces of the earthquake. Twelve unbonded NiTi

SMA bars were placed around the square precast concrete column and anchored to the footing of

the column to prevent premature failure, this is shown in Figure 2-14. The SMA bars removed the

plastic hinge that would form when the column rocked. A bonded post-tensioned (PT) cable was

placed down the centre of the segmental pieces to hold the segments together as well as it provided

additional re-centring abilities. By providing the SMA, they found that not only was complete

damage to the column avoided, but the self-centring ability allowed for the structure to undergo

subsequent shocks.

Figure 2-14: Schematic of rocking column with SMA, adopted from Roh and Reinhorn (2010)
Oudah and El-Hacha (2016, 2017) studied the seismic performance of single-slotted and double

slotted self-centring concrete beam-column connections that were strengthened using PE NiTi

SMA bars as seen in Figure 2-15. By doing so, he suggested that it would relocate the location of
 27

the plastic hinge away from the face of the column. At the connections, he placed the SMA bars

to re-centre the structure as well as to dissipate some of the seismic energy. He found that using

PE SMA bars significantly improved the seismic response of the proposed system by achieving a

self-centring behaviour and thus minimizing permanent damage done on the structure. However,

he found that the use of SMA bars led to a reduction in the damping capabilities. Another drawback

of the system is providing proper anchorage for the SMA bar so that failure does not occur before

the full capacity of the material is met. Another careful consideration that should be taken into

account by a designer is that because NiTi SMA has a much lower stiffness than conventional

steel, the amount of SMA used needs to be taken into account so that the system does not exceed

the maximum drift limits.

Figure 2-15: Cracking pattern at ultimate conditions (Oudah, 2014)

Alam et al. (2012), combined the use of typical steel rebar with NiTi SMA for the main

reinforcement for concrete beams in an analytical model. They did a comparative study, where
 28

one frame specimen had only steel reinforcement, one had SMA’s only in the plastic hinge region

and steel elsewhere, and a specimen with SMA in the beams and steel in other regions. The SMA’s

were placed at the beam-column connection where the plastic hinge region would develop. It was

found that by placing the SMA’s at the plastic hinge locations the members were able to dissipate

adequate energy as well as re-centre the structure. Therefore, members experience little to no

damage and the repair cost is reduced after a seismic event. For a 6-storey frame, the average inter-

storey drift demands were 2.73%, 2.67% and 2.3% for SMA, Steel-SMA and Steel frames

respectively. This result suggests that due to the lower stiffness of the SMA the system will be

more flexible and prone to higher drifts.

2.3.2. SMA-based Dampers

Devices that are most commonly used for seismic resistance of structures are dampers, which

include viscous fluid dampers, viscoelastic solid dampers, friction dampers and metallic dampers.

The objective of these devices is to reduce the inelastic energy demand on the system in order to

reduce damage done to the structure. Though seismic these isolators have proven to enhance the

seismic performance of a structure they are susceptible to large deformations under strong events

resulting in the need for repair or replacement. Ozbulut and Silwal (2016), examined the

performance of Superelasic-Friction Bearing Isolator (S-FBI) composed of flat steel sliding

bearing and superelastic SMA cables. A three-story building was modelled and assessed using

numerical methods. They determined that when the structure was subjected to seismic loading the

control fixed-base model experienced a higher drift and larger floor accelerations when compared

to the S-FBI system. In particular, the interstory drift decreased from 3.2% to 1.5% and the peak

story acceleration was reduced by 64% relative to the fixed-base model. Similarly, Qian et al.,
 29

(2016), combined SMA wires and friction devices to create a damper that was more efficient in

dissipating the energy and recentring the system; a schematic of the design is shown in Figure 2-

16. The research findings showed the success of the developed SMA friction damper with

enhanced structural performance compared to the system without the device.

Figure 2-16: Schematic of superelastic shape memory alloy friction damper (Qian et al., 2016)

Recently, a PE SMA self-centring device (SCD) was created and implemented in a scaled cable-

stayed bridge for controlling the seismic response (Zhou et al., 2018). Experimental tests on the

device and the model showed a dramatic reduction of the tower accelerations and relative

displacements of the deck. Figure 2-12 shows the basic configuration of the SCD.

a) b) c)
Figure 2-17: Configuration of SCD: a) initial condition; b) lower part of wire in tension when

piston moves up; c) upper part of wire in tension when piston moves down (Zhou et al., 2018)

Here Zhou et al. (2018), developed a device that fully utilizes the PE and energy dissipation

capacity of the SMA wires. From Figure 2-12, it can be seen that upon loading and reverse loading,
 30

part of the SMA wire will always be in tension allowing for a damping force to be produced under

both conditions, ideal for seismic loading conditions.

2.3.3. SMA Braces

Braced frames are another common system used to improve the lateral response of a structure.

Results from experimental and numerical studies have shown that braces that incorporate PE SMA

material are effective in recentring the system as well as dissipating some of the energy. Qiu and

Zhu (2017) investigated the seismic performance of a six-story concentrically braced SMA frame

through a numerical analysis. Figure 2-18 shows a schematic design of the SMA brace. Results

showed that the system can successfully mitigate the high-mode effect which results in a uniform

distribution of peak inter-story drift ratios, as well as the frames exhibit limited structural and

permanent damage even after a very strong earthquake.

Figure 2-18: Schematic of shape memory alloy brace (Qiu and Zhu, 2017)

SMA bracing systems have also been applied to concrete frames as a retrofit. Cortes-Puentes and

Palermo (2018) developed a bracing system composed of tension-only PE NiTi SMA and applied

it to a seismically deficient squat RC shear wall, as shown in Figure 2-19. They found that the

energy dissipation was substantially increased by the SMA braces and the wall was able to recover

over 50% of the imposed lateral displacements.

 31

Figure 2-19: Reinforced concrete shear wall with PE NiTi SMA braces (Cortes-Puentes and

Palermo, 2018)

In the 1990’s, a European project called ‘The MANSIDE Project’ (Memory Alloys for New

Seismic Isolation and Energy Dissipation Devices) studied the use of SMA’s as a passive device

for seismic resistance. The device was initially designed to serve as a bracing system for buildings

or an isolation system for bridges. It was composed of pre-tensioned PE wires that were only

strained in tension to provide the re-centring capability and then additional martensitic bars or

austenitic wires for energy dissipation (Dolce et al., 2000). A schematic of the device is shown in

Figure 2-20. When implemented in a 1500-ton masonry building, they moved the building 140

mm laterally and then released it suddenly. After only a few damped oscillations the structure fully

re-centred itself with zero residual displacement.

 32

Figure 2-20: Schematic of SMA device (Dolce et al., 2000)

Recently, Abou-Elfath (2017) investigated the ductility characteristics of a steel buckling

restrained brace composed of PE SMA through a numerical analysis. The SMA bars were

unbonded and placed in grouted tubes so that they would not buckle under the compressive

loads. They found that the SMA frame exhibited undesirable strength and stiffness increase after

it reached its yield level as the SMA exceeded the maximum allowable drift limit of 5.5%. Their

concern, by allowing the frame to drift, is the potential danger it poses on the other structural

members, such as the columns, as they may not be able to recover from such a drift.

2.4 Further Research

As mentioned in the introduction, earthquakes occur all over the world and sometimes one place

can be hit with multiple earthquakes in a year. Buildings and bridges are frequently damaged and

are typically rendered structurally deficient and need to be demolished. This research proposes to
 33

investigate an innovative and efficient method that aims to reduces the permanent damage done to

a structure so that they can undergo subsequent events.

So far, the behaviour of SMAs and their use in seismic design have been discussed. SMA is an

excellent material for structural engineering applications. Over the past decade, a significant

amount of research has been done on the use of SMA and its need in the industry. However, SMA

is still not being utilized, especially in the seismic engineering field. From the studies looked at in

this chapter, using SMA as the main reinforcement in the plastic zone of a structure can fully re-

centre the structure and dissipate a significant amount of energy with almost zero residual

deformations. From Abou-Elfath’s (2017) study, an issue with using SMA as the mechanism to

dissipate and re-centre the structure is that it is too flexible and may exceed the drift limits, thus

putting too much stress on other structural members. Since the MANSIDE project, research on

NiTi SMA has been focused on the behaviours of the material under different loading conditions

as well as implementing it in numerical models. One of the main issues with designing with

composite materials is determining the connection or anchorage points. This research proposes to

investigate the experimental behaviour of a steel braced frame reinforced with PE NiTi SMA wires

using a simple anchorage design. Furthermore, this system can be applied to both new and existing

framed structures.

2.5 Summary

This chapter has reported the necessary literature review of SMA for conducting research in the

area of seismic design, with an emphasis on the response of the PE NiTi SMA. The different factors

that affect the behaviour of SMA were discussed, factors such as temperature, loading, size, and

fatigue in the context of seismic applications. Finally, the different applications of PE SMA in
 34

seismic design were discussed. SMA has been used as dampers, connections, and braces. The

studies done for RC members showed that the plastic hinge would develop at the location of the

SMA thus effectively dissipating the energy as well as reducing the residual displacement. Since

PE SMA has the ability to sustain almost zero residual strains, the RC structures experienced little

to no damage after being subject to cyclic loading. Overall, for this research, it was proposed to

design and test a steel braced frame reinforced with NiTi SMA wires and assess its feasibility for

seismic applications.
 35

Chapter 3: Experimental Behaviour of PE NiTi-SMA Wires

3.1 Introduction

Numerous testing procedures on determining the material characteristics of SMA wires have been

reported in the literature. Since the use of SMA for structural applications is fairly new, standards

have not yet been developed for tension testing of SMA wires specifically. The procedures adopted

here were deemed sufficient to characterize the material properties of the SMA wires utilized for

seismic applications by following the ASTM E8 standard. The aim of this study was to investigate

the monotonic and cyclic behaviour of the PE NiTi-SMA wires with emphasis on the damping and

energy dissipation properties. This chapter includes a preliminary experimental study that was

conducted to assess the characteristics of the PE NiTi-SMA wires before applying them to the

braced steel frame.

3.2 NiTi-SMA Material Composition and Characterization

The elemental composition of the 2.0 mm diameter NiTi-SMA wires is included in Table 3-1 and

Table 3-2 as certified by the manufacturer (Memry, 2018). The fully annealed austentic start (As)

temperature ranges from -15°C to -10°C; therefore, the NiTi-SMA material exhibits a pseudo-

elastic (PE) response at room temperature. According to the manufacturer, the ultimate tensile

strength and elastic modulus of the SMA wires were 1386MPa and 75GPa, respectively (Memry,

2018). However, these values were verified through the traditional tensile test. Tables 3-2 and 3-3

show the mechanical properties of the SMA wires.

Table 3-1: Chemical Analysis of NiTi-SMA wire (Memry, 2018)

Element Requirement (%) Test Results (%)
Ni 54.5-57.0 56.04
Ti - -
C 0.05 0.0328
 36

Co 0.05 0.0001
Cr 0.01 0.0007
Cu 0.01 0.0008
Fe 0.05 0.0130
H 0.005 <0.0050
Nb 0.025 0.0001
N - 0.0005
O - 0.0220
N+O 0.05 0.0225
Other single trace elements - <0.01
Total Measured Impurities - <0.2000
“-”: Not reported by the Manufacturer

Table 3-2: Characteristic Properties of NiTi-SMA wire (Memry, 2018)

Property Requirement Test Results Standard
Wire Diameter (mm) 2.025 max 1.998 max
1.975 min 1.995 min
Residual Strain - 0.1% ASTM F2516
Young’s Modulus 40(M)/75(A) GPa 75 GPa
Ultimate Tensile Stress 1100-1500 MPa 1386.5 MPa ASTM F2516
Total Elongation - 10.6% ASTM F2516
Austenite Start Temperature -15°C ±10°C -16/-14°C
“M”: Martensite phase
“A”: Austenite phase
“-”: Not reported by Manufacturer

3.3 SMA Tensile Testing

The NiTi-SMA braced frame is composed of steel columns, a cross beam and steel diagonal

members that are disconnected close to the midpoint. These disconnected steel braces are then

connected with the PE NiTi-SMA wires as shown in Figure 3-1; this will be referred to as a fuse.

The wires are then activated once the stress in the wires reaches the yielding stress (Ms)

transforming the wires from their austenite state into their martensitic state. The unique properties

of PE SMA, as discussed in Chapter 2, will allow for some energy dissipation with each cycle. It

should be noted that since PE NiTi-SMA is being used, when the material is under tension it will

start in the austentic phase and transform to the martensite phase and during unloading it reverts
 37

to its parent austenitic phase. Therefore, the tensile characteristics of the NiTi-SMA wires were

determined in their parent state (austenitic phase) to understand the behaviour of the material

better.

(a) Control Braced Frame (CBF) (b) SMA Braced Frame (SMA-BF)

Figure 3-1: Typical steel braced frame configuration

The setup of the monotonic tests of the PE NiTi-SMA wires is shown in Figure 3-2. Three

specimens were tested using a 20 kN Tinius Olsen testing apparatus fitted with friction grips. Each

wire specimen had a length of 300 mm, with a gripping length of 57 mm on either side of the

specimen, leaving a 186 mm nominal gauge length (According to ASTM E8, the minimum gauge

length is 4diameter). The monotonic test consisted of the application of ramp displacement-

controlled loading at a rate of 10.0 mm/min until failure. The average stress and strains were

determined from the results and then graphed, as shown in Figure 3-2.
 38

Tensile Testing
Machine Light

PE NiTi-SMA
Wire
Data
Acquisition
High Resolution
System
Camera

Figure 3-2: Instrumentation and test set-up of the NiTi-SMA tension test

The instrumentation consisted of a strain gauge (SG) mounted at the mid-length of the wire to

monitor the localized strains and the Digital Image Correlation (DIC) technique to measure the

axial strain across the gauge length of the wire. The maximum elongation of the strain gauges is

2-3% and it was expected for the PE NiTi-SMA wires to elongate up to 10% according to the

manufacturer, as stated in Table 3-2. Therefore, the DIC technique was used not only to monitor

the strain throughout the wire but to measure the strain at the failure load. Typically, the

application of the DICT requires some surface preparation such as painting the surface black and

applying white spots randomly across the surface. However, since the SMA wire is already black

and is quite small in diameter, only white dots spaced at about 25 mm along the wire were placed

as seen in Figure 3-3. As the experiment is running, a 15.1-megapixel high resolution Canon
 39

Digital Rebel camera is used to capture images at a rate of one frame per 5 seconds until fracture.

The DIC technique works by tracking these predefined points in a Matlab program in a series of

images captured by the camera.

End Clamps

Strain Gauge

PE NiTi-
SMA Wire

Strain Gauge

End Clamps

Figure 3-3: DIC and strain gauge application on 2 mm diameter SMA wire

3.3.1 Experimental Analysis of Tension Test

The monotonic stress-strain relationship of the PE NiTi-SMA wires is shown in Figure 3-4. Here,

results of all three wires are graphed to show the consistency of the results. As mentioned before,

the strain gauge has a maximum elongation of 2-3% and for 2 of the 3 wires, stopped producing

results after reaching the yielding plateau. However, as can be seen, the stress-strain response is

similar in all three specimens.

 40

Figure 3-4: Strain gauge analysis of the SMA wires subject to tensile loading

Figure 3-5: DIC technique analysis of the SMA wires subject to tensile loading

The strain throughout the NiTi-SMA wire at various stresses is shown in Figure 3-5. The

continuous lines were obtained from the DIC technique method where the star-shaped points was

obtained from the SG reading at the mid-length of the wire. From the results it can be seen that the

strain is almost consistent throughout the wire while it is still in the austenite phase. As the material
 41

starts to transform from austenite to martensite, σMs, the strain does not change much as the stress

increases. When the stress is equal to the martensite start stress, yielding occurs and the strain

increases as the stress increases. This effect can be seen when looking at the strain levels at 503

MPa and then at 533 MPa. At the fracture stress, 1248 MPa, the DIC technique was not able to

track all the points on the wire, this may have been due to the strains being too small to track at

that point, hence the discontinuity. The SG readings are close to the DIC technique measurements

up to 533 MPa which suggests the DIC technique is functioning properly.

To ensure that the DIC technique was functioning throughout the test, the SG graph was compared

to the DIC technique, as shown in Figure 3-6. Here, an average of the three specimens was taken

and the results are almost identical up to approximately 6% strain. As the NiTi-SMA wire reached

the full martensite phase, from 6% to about 9% strain, the measurements between points increase

at larger increments. This could be because the program was not able to track the points at those

strain values due to the camera not being close enough to the specimens. Overall, the correlation

between the two methods is very similar and thus the DIC technique is accepted as an effective

method for measuring strain.

 42

Figure 3-6: Stress-strain diagram of the strain gauge and DIC technique of the NiTi-SMA wires
subject to tensile loading

The average stress and strain values of the martensite starting stress (σMs, εMs), the ultimate tensile

stress and strain (σu, εu), and the elastic modulus of the austenite and martensite phase (EA, EM) are

included in Table 3-3. It should be noted that EA and EM were determined from the slopes of the

linear portions of the obtained graphs.

Table 3-3: Summary of the experimental tensile testing of the NiTi-SMA wires
σMs (MPa) εMs σu (MPa) εu EA (GPa) EM (GPa)
500 0.012 1247 0.09 41.7 26.5

3.4 Cyclic Behaviour of SMA wires

The PE NiTi-SMA wires are to be used as a damper as well as a re-centring mechanism for the

steel braced frame, therefore, the wires will be subject to cyclic loading. The setup of the cyclic

loading test is shown in Figure 3-7. Three specimens were tested using an MTS hydraulic testing

apparatus fitted with friction grips. Each 2 mm diameter SMA wire specimen had a length of 300

mm, with a 75 mm long 12 mm outer diameter aluminium sleeve at each end of the wire as shown
 43

in Figure 3-8. The aluminium sleeves were used to avoid the biting action of the wedge tooth on

the PE NiTi wire and developing stress concentrations by applying a transverse stress on the wire

preventing this premature failure. Previous research studies have used similar methods for testing

Fiber Reinforced Polymer (FRP) bars as well as anchoring FRP bars for prestressed applications

(Mohee and Al-Mayah, 2018; Reda Taha and Shrive, 2002).

MTS Testing Machine

PE NiTi-SMA wire Data Acquisition System

Extensometer

Figure 3-7: Cyclic load test setup

 44

Figure 3-8: Aluminium sleeves used for gripping PE NiTi SMA wire

The cyclic test consisted of the application of tension-tension cyclic loading, adopted from

DesRoches et al., (2004), as illustrated in Figure 3-9. The loading protocol used consisted of

increasing strain cycles of 0.5%, 1.0% to fracture by increments of 1%. The specimen was cycled

once at 0.5% strain, twice from 1-2% and then three times from 3% strain to when fracture

occurred. For the first test specimen the loading was performed at a frequency of 0.05 Hz, for the

second and third tests, the frequency was reduced to 0.025 Hz. The change in frequency for the

last two tests was done to reduce the amount of residual strain in the wire after each cycle.

Figure 3-9: Strain-controlled loading history

 45

The instrumentation consisted of a 25 mm extensometer mounted at the mid-length of the wire to

monitor the localized strains and the DIC technique to measure the axial strain across the gauge

length of the wire. The extensometer had to be removed at 8% strain as to not damage the device

when the specimen fractured.

3.4.1 Experimental Analysis of Cyclic Tests

The cyclic response is plotted in Figure 3-10. The DIC technique was unsuccessful at providing

accurate results for the cyclic tests. Therefore, only the strain readings obtained from the

extensometer were used to plot the cyclic response of the PE NiTi wire up to 8% strain, however

the cyclic test was carried through until fracture in the wire occurred, which was approximately at

12% strain. The cyclic response was evaluated by examining the effect of the cyclic loading on

the following parameters: residual strain (εr), the equivalent viscous damping ratio (ζeq), the

loading stress plateau (σL), and the unloading stress plateau (σUL). The residual strain was

determined by examining the strain at the end of maximum cyclic strain cycle. The equivalent

viscous damping ratio was determined using Equation 3-1.

1 𝐸𝐷
𝜁𝑒𝑞 = 𝜔 𝐸
4𝜋(𝜔 ) 𝑆𝑜 Equation 3-1
𝑛

where

- ED is the energy dissipated in a cycle,

- ESo is the elastic energy in a cycle,

- w is the loading frequency, and

- ωn is the natural frequency.

 46

The equivalent viscous damping ratio is a property of the system and depends on its mass and

stiffness (Chopra, 2012). It is the simplest form of damping since the governing differential

equation of motion is linear and can be evaluated through simple experimental testing.

According to Equation 3-1, the ζeq depends on the frequency at which the test is conducted. It is

recommended to conduct the test at a frequency near the resonance frequency since the response

of the system is most sensitive to damping at that frequency (Chopra, 2012). Due to the difficulty

of conducting the cyclic testing at high frequencies, it has been a common practice to take the

testing frequency as the natural frequency of the system. Therefore, the equivalent viscous

damping ratio taken is that at the natural frequency. Equation 3-1 now becomes:

1 𝐸𝐷 Equation 3-2
𝜁𝑒𝑞 =
4𝜋 𝐸𝑆𝑜

The loading and unloading stress plateau parameters are defined in Figure 3-11.

Figure 3-10: Stress-strain response of PE NiTi SMA wire subject to cyclic loading
 47

Figure 3-11: Definition of the parameters used to evaluate the cyclic response of the PE NiTi
SMA wire

The four parameters are plotted against the maximum cyclic strain as shown in Figure 3-12. The

residual strains after unloading are more dominate in the specimen cycled at a higher loading rate

(Figure 3-12(a)). Though the residual strain is quite small, less than 0.7%, it can be suggested that

at the higher loading rate the PE NiTi wire is in its plastic state when it is reloaded, thus maintaining

a very small residual strain. Previous research has shown that the increase in residual strain with

cycling is due to small levels of localized slip and build-up of dislocations and may stabilize with

increased cycling, otherwise known as training (DesRoches et al., 2004; McCormick et al., 2005).

Though a small residual strain is seen at a slightly higher loading rate, it can still be shown that

their re-centring capability of the material is intact after being cycled to 8% strain.

The equivalent viscous damping ratio (ζeq) increases with increasing strain until it reaches a peak

at a strain of approximately 5% as shown in Figure 3-12(b). At a loading rate of 0.025 Hz, the
 48

material is able to dissipate more energy than at the slightly higher rate. A similar trend is shown

in DesRoches et al., (2004).

Results show that the trend of the loading and unloading stresses are not significantly affected by

the loading rate. For the first couple of cycles, the forward transformation stress (loading plateau

stress σL) occured at approximately 573 MPa and 629 MPa for the 0.05 Hz and 0.025 Hz rates

respectively (Figure 3-12(c)). At 8% strain, σL decreases to approximately 382 MPa and

414 MPa. As mentioned in Chapter 2, Section 2.3.1.4, the effect of decreasing loading plateau

stressess with increasing cycles is functional fatigue. For the reverse transformation stress

(unloading plateau stress σUL), it starts at approximately 200 MPa (0.05 Hz) and 245 MPa

(0.025 Hz) and decreases to 137 MPa and 150 MPa (Figure 3-12(d)). It has been noted in previous

research, that σUL generally decreases less over the cycles when compared to the loading plateau

(Sherif and Ozbulut, 2018).

 49

Figure 3-12: PE NiTi subject to quasi-static cyclic loading showing: (a) residual strain (εr), (b)

equivalent viscous damping ratio (ζeq), (c) loading plateau stress (σL), and (d) unloading plateau

stress (σUL)

3.5 Training SMA Wires

As stated previously, SMA material has the unique ability to undergo large deformations and return

to its original undeformed shape through either the shape memory effect or the pseudoelastic

effect. Furthermore, the properties that make them desirable for seismic applications is their

hysteresis response, re-centring ability, and its high resistance to low and high cycle fatigue.

Though SMA material has proven to provide adequate dissipation and re-centring capabilities, it

has also been shown that cyclic loading results in degradation of both of these properties. In a

study done by McCormick et al., (2005), they investigated the effect training NiTi-SMA wires had

on their response to cyclic loads. Training refers to the procedure of cycling the material to a
 50

certain strain before using it for its main application. Figure 3-13 shows a 2.16 mm NiTi-SMA

wire under quasi-static loading, untrained.

Figure 3-13: Stress-strain curve for untrained NiTi SMA wire (McCormick et al., 2005)

From Figure 3-13, it can be seen that the NiTi-SMA wire does not reach a stable response until it

has been subjected to a number of cycles. In order to utilize the full capabilities of the SMA in

seismic resistant designs, a stable response is required. Furthermore, a stable response will produce

more a predictable response of the structure during any seismic event. In 1986, Miyazaki et al.,

suggested that by training the NiTi-SMA material, the fatigue effects can be limited. By training

the SMA before use, it will reach a stable hysteresis which can then be extracted and used in

analytical models to predict future behaviour of the structure under different loading conditions.

Following this study, McCormick et al., (2005), the optimal number of cycles and strain level are

60 and 5%, respectively. It should be noted that the 5% was based on the specimen being fully

transformed to its detwinned martensite phase. Unfortunately, for this project the manufacturer

was not able to train the wires, and due to time and equipment constraints the wires were unable
 51

to be trained. It should be addressed that for future studies with this material, training should be

done to produce more predictable results.

3.6 Tensile Testing of NiTi-SMA Connection Plate

In order to utilize the full capacity of the PE NiTi-SMA wires, a custom-made gripping device had

to be designed and manufactured. As shown in Figure 3-14, each slot held four NiTi-SMA wires

and was then clamped down with 6 mm diameter high strength bolts; each bolt was torqued to 13.6

N·m. The resisting force of the wires when a tension load is applied is the frictional force between

the grooved steel and the wire.

Figure 3-14: Initial anchorage system

Using the typical mechanics torque-force relationship, the frictional force was calculated using

Equation 3-2:

𝑇 = 𝑐𝐷𝐹 Equation 3-2

 52

where T is torque, c is the coefficient of friction of the bolt (taken as 0.2 to be conservative,

(Engineers Handbook, 2004)), D is the diameter of the bolt, and F is the axial force.

By rearranging Equation 3-2, the axial clamping force of each bolt was determined to be 10.7 kN.

Then Equation 3-3 was used to determine the capacity of frictional force of each bundle of NiTi-

SMA wires.

𝐹𝑟 = 𝜇𝑁 Equation 3-3

where μ is the coefficient of friction of the steel (taken as 0.2 to be conservative, (Engineers

Handbook, 2004)) and N is the total clamping force. Each bundle was straddled by two bolts;

therefore, the total clamping force was 21 kN. By plugging these values into Equation 3-3, it was

determined that the frictional force each bundle had was 4.3 kN at each connection point and

8.6 kN of resistance at each bundle. With this, the brace could theoretically undergo 189 kN of

tensile load.

Before placing the NiTi-SMA wires in the braced frame, the anchorage system was tested in a

1MN 10” stroke MTS machine to verify the load at which the NiTi-SMA wires slip. In the testing

apparatus, the gripping strength was tested by connecting 8 wires to the apparatus bundled in two

sets of 4 wires each. Two sets of wires (to be referred to “Left Wires” and “Right Wires”) were

tested to examine the gripping strength before slipping occurred between the grips as well as to

observe the load distribution to the two sets. It should be noted that in this phase of testing, it was

not feasible to test the anchorage system with the full set of wires (i.e. 44 wires as the frame would

have). The goal was to test the experimental load at which slip would occur with the theoretical

load, assuming equal load distribution. In order to place the connection system in the MTS machine

temporary round bars were cut flush to the back of the connection plate and welded. This allowed

for a more evenly distribution of load across the wires. It should be noted that in order to eliminate
 53

the misalignment of the two connection plates when the wires were anchored in, a square 100 mm

wide steel plate was placed in between the plates to ensure that a constant 100 mm gap was kept

between the plates. The setup of the test is shown in Figure 3-15. A ramp displacement of 2

mm/min was applied. A Linear Strain Conversion (LSC) device and the DIC technique were used

to track the displacements in the wires.

Figure 3-15: Experimental set-up of anchorage system in MTS machine

Based on the above equations, the wires should have been able to sustain at least a load of

approximately 17 kN. Figures 3-16 and 3-17 show the results from the DIC technique and LSC,

respectively. From the LSC and DIC data, slip occurred at roughly 14.7 kN.
 54

Figure 3-16: DIC Load-Strain results of Left and Right Wire groups

Figure 3-17: LSC Load-Strain results of anchorage system

 55

The data collected by the LSC shows the slip more clearly than the DIC technique. The slip

between the steel connection plate and NiTi-SMA wires suggests that the coefficient of static

friction initially assumed was incorrect and is lower than 0.2. This could be due to the black oxide

surface finish of the NiTi-SMA wire. In order to increase the resistance of the wires so that slip

does not occur at a low load, roughening the surface to eliminate the black oxide finish as well as

increasing the amount of torque applied to each bolt was done. It should be noted that both the

Left wire and Right wire groups experienced almost the same loads suggesting that the load was

distributed evenly to both groups, so it can be assumed that slip at each bundle would occur at a

tension load of 7.35 kN.

For the main experiment the expected yielding load in the brace was to be 125 kN. In each brace

there are 22 bundles of wire, it was assumed that the load would be evenly distributed to each

bundle, therefore, each bundle would take approximately 5.68 kN at yielding which is less than

the expected slipping load. For the main experiment, the bolts were torqued to 20 N·m and the

wires were roughened using sand paper to increase the coefficient of friction between the steel and

NiTi-SMA wire.

3.7 Summary

This chapter included the elemental composition and the characterization of the pseudoelastic

properties of the NiTi SMA wires. Tension tests were conducted to determine the tensile

characteristics of the SMA wires in the martensite state. Quasi-static cyclic tests were conducted

on the wires to determine the cyclic properties and potential of PE NiTi SMA in seismic resistant

design. Properties such as recentring, damping, and loading and unloading plateau stresses were

investigated. This chapter also included the design and testing procedure of the unique SMA
 56

anchorage system that was implemented in the main frame experiment. The importance of training

NiTi SMA wires before being subject to cyclic loading is discussed, and though training of the

wires was not done in this research phase, it is recommended that it be done for future studies. The

outcome of the study presented in this chapter validated the success of the proposed method, and

the research study can further progress towards investigating the behaviour of steel braced frames

combined with NiTi-SMA wires.

 57

Chapter 4: Design and Behaviour of Steel Braced Frame with Yielding Fuse

4.1 Introduction

In the previous chapter, an in depth look at the NiTi SMA material properties was examined. In

this chapter, the rationale behind the large-scale experimental program is explained in depth. First,

the motivation behind this research is presented. Next, the methodology of this research is

described; the main aspects of the design process are included while the detailed calculations are

presented in Appendix A. The goal of this research was to do a simple comparison between steel

and SMA braced frames to demonstrate the benefits of using SMA in a seismic structural design.

Three specimens were tested under free vibration and quasi-static cyclic loading. In each specimen

the yielding portion was placed at approximately two-thirds up the cross brace. The yielding

portions were analyzed and compared in terms of hysteric response, residual strains and

deformations. Finally, a comparison between the steel frame and SMA frame is drawn, outlining

the important advantages and disadvantages of each system.

4.2 Testing Matrix

The objective of this phase of the research is to investigate the performance of PE NiTi SMA steel

braced frame. Three large-scale frames were examined experimentally; a control steel braced

frame (CBF), PE SMA-reinforced braced frame (SMA-BF-1), and a SMA-BF with an improved

anchorage system and slotted connections (SMA-BF-2). In the latter two systems, their

performance was evaluated against three performance objectives:

a) achieving a self-centring response;

b) achieving a target capacity of 4% drift; and

c) achieving adequate energy dissipation.

 58

The design considerations of the proposed systems are discussed in the following sections.

4.3 Design

4.3.1 Analysis of applied loads

Lateral load resisting systems can be composed of concrete, steel, masonry, timber or a

combination of different materials. In this experiment, the main objective was to extract the

hysteresis response of the SMA material so that it could be implemented into a numerical analysis

where different parameters could then be changed, such as material. Therefore, for simplicity and

practicality, a general steel braced 2m  2m frame was considered for this study.

(a) Control Braced Frame (CBF) (b) SMA Braced Frame (SMA-BF)

Figure 4-1: Specimen details

Figure 4-1 shows the elevation profile of the two frame setups, Control Braced Frame (CBF) and

SMA Braced Frame (SMA-BF). One of the objectives of this study was to ensure a ductile failure

of the structure, at the location of the plastic hinge in the brace. The plastic hinge was designed
 59

based on two main criteria: the maximum load the actuator could apply and the strength of

structural steel.

Figure 4-2: Typical steel stress-strain curve

The maximum load that the available lateral actuator at our structural lab could apply was 250 kN

with a 254 mm stroke. In order to ensure a ductile failure, the specimens had to be designed so that

failure would occur within the 250 kN load. From Figure 4-2, a typical stress-strain diagram is

shown. For most ductile structural steels, an elastic portion exists to about 0.2% strain (point A).

However, after a critical stress is reached, Fy, the specimen will undergo some deformation with

little change in stress: this is known as the yield plateau. Then it will experience a large

deformation, sometimes up to 200 times as its deformation before yield, to a maximum stress

value: this is the strain hardening. At this point the specimen will start to neck, the cross-sectional
 60

area decreases, with some load decrease, then rupture occurs. In structural design, most elements

are designed based on the yield strength of the member, in other words an elastic limit design is

always used. However, there is still a significant amount of reserve strength that exists before the

member failures (between points B and C). In accordance with the 2015 National Building Code

of Canada (NBCC, 2015), the minimum lateral earthquake design force using the equivalent static

force procedure is given by Equation 4-1:

𝑀𝑣 𝑆(𝑇𝑎 )𝐼𝐸 𝑊
𝑉𝑠 = Equation 4-1
𝑅𝑑 𝑅𝑜

where

- Mv is the factor to account for higher mode effects on base shear

- S(Ta) is the design spectral response acceleration for a period T

- Ta is the fundamental lateral period of vibration of the structure

- IE is the importance factor for earthquakes

- W is the dead load of the structure

- Rd is the ductility modification factor

- Ro is the overstrength modification factor

- Rs = R0Rd is the reserve strength factor

The Rs factor takes into account the dependable portion of reserve strength in the Seismic Force

Resisting System (SFRS) as well as the ability of the system to dissipate energy through reversed

cyclic inelastic behaviour. According to the NBCC (2015), the Rs factor that should be applied

when designing for seismic applications is 1.5. This brings the load from the ultimate, 250 kN, to

the elastic limit, 166 kN load. The columns and cross beam were designed to take this lateral load
 61

of 166 kN. However, to ensure full utilization of the braced members, an additional 1.25 load

factor was applied (Rahgozar and Humar, 1997). Furthermore, to simulate real world situations,

vertical loads were applied to the columns to act as gravity loads. Roh and Reinhorn (2010),

suggested gravity loads between 5% and 20% of the column’s axial strength; based on this a

65 kN load was placed on each column, roughly 7% of the columns axial strength. Table 4-1 has

the list of member sizes. Table 4-2 shows the yield strength, tensile strength and grade of steel as

per the manufacturers testing in accordance with ASTM A370.

Table 4-1: Specimen section details

Columns Cross beam CBF brace Steel yielding SMA-BF SMA fuse
fuse brace
W15030 W25022 L1021026.4 PL300456 PL2006 88-2mmϕ wires

Table 4-2: Steel Material Properties as per Manufacturer

Member σy (MPa) σu (MPa) Grade

Columns 355 468 350W
Cross Beam 391 474 350W
CBF Brace 363 529 44W
Steel yielding fuse 413 501 44W
SMA-BF brace 438 504 44W

4.3.2 Brace Design

An idealized system consists of a mass concentrated at the roof level, a massless frame that

provides stiffness to the system, and a viscous damper that dissipates vibrational energy of the

system (Chopra, 2012). The properties of each structural member are concentrated into three

separate components: mass (cross beam), stiffness (columns), and damping component (cross

braces). The system was constrained to only move in the direction of the excitation, making it a

single-degree-of-freedom (SDOF) system.

 62

In current practice, steel braced frames are designed on the assumption that the compression

diagonal has negligible capacity and the tension diagonal resists the total lateral load. This

approach is a conservative design, as the actual lateral load on the system is much larger than the

design load because of the load sharing that occurs between the diagonals. This approach was

adopted for this research. In this research, three tension-only braces (i.e. braces were designed to

resist tension force only and the members were not connected at the intersection point) systems

were considered. To understand the potential of NiTi-SMA wires fully, the specimens were

designed to an equivalent load capacity. Since, steel and SMA have a different modulus of

elasticity 200 MPa and 42 MPa, respectively, the materials may yield at the same load, but their

yielding strains will be different. It is the yielding of the material that dissipates the seismic energy;

if the specimen is able to yield at a relatively small drift, more energy can be dissipated throughout

the excitation. Using Equation 4-2, the area of each material was calculated based on the

assumption that the yield strength of both materials was 350 MPa and a material resistance factor

of 0.9 was taken as per CSA S16-14.

𝑇𝑟 = ∅𝐴𝐹𝑦 Equation 4-2

The steel yielding fuse adopted in this research is not covered in CSA S16-14 or the NBCC (2015),

as it is expected that braces with short lengths fail at limited drifts, especially if the fuse is subject

to flexural yielding in compression due to buckling. So rather, bracing members having the same

cross-section throughout their length is what is typically done. However, the CBF was not designed

to represent current industry practice, but rather, for a comparative means for the behaviour of the

SMA-BF. Furthermore, being limited by the maximum lateral load that could be applied by the

actuator and a realistic amount of NiTi-SMA wires that could be used, a small yielding fuse was
 63

developed and designed. It should be noted that the term “fuse” refers to the part of the brace that

is acting as the plastic hinge. For the CBF, it is the steel plate, for the SMA-BF it is the NiTi SMA

wires.

The SMA-BF was composed of 88-2 mm diameter PE NiTi-SMA wires, placed in the anchor

developed as described in Chapter 3. The CBF system was composed of conventional 300456

mm steel plate. The other steel members were designed in accordance with CSA S16-14 to remain

in the elastic limit when loaded. Table 4-3 presents the expected yield and failure loads for the

fuses.

Table 4-3: Expected yield and failure loads for yielding fuses
Yield Area Expected Rupture Expected Expected
strength (mm2) yield strength rupture lateral load
(MPa) force (MPa) force at rupture
(kN) (kN) (kN)
456 plate 390 270 95 496 134 95
SMA (88-2 mm) 495 276.5 123 1248 345 244

4.3.3 Connection Design

Lateral load resisting systems (LLRS) are provided in the vertical plane of structures to resist

horizonal loads and to ensure lateral stability. There are three primary LLRS: vertical braced

frames, rigid frames, and shear walls. For this research, a vertical braced frame was examined.

Vertical braced frames have all pinned connections and the columns are only subjected to axial

loads; these factors help contribute to the rocking aspect of the system. As stated before, the

objective of this experiment is to ensure a ductile failure, and typically braced frames are more

stiff than rigid frames thus limiting their ductility. However, by designing for a pin connection at

the foundation level, the system will be able to rotate, and reduce the overall stiffness of the system;
 64

furthermore, by implementing a yielding fuse within the cross brace the flexibility of the system

is increased. With the pin-pin connection of the frame to the foundation beam, though in design

we consider no moment transfer, but because the frame is bolted using a plate, there is a partial

fixity that causes a moment transfer. Figure 4-3 demonstrates the effect when a lateral load is

applied to the frame.

Figure 4-3: Schematic of partial fixity at column base

With that said, the connections were mainly designed to withstand the respective shear forces. The

beam-column connection was designed to take the self-weight of the beam and the column-

foundation connections were designed to take the maximum lateral load the actuator could apply.

The braces were welded to 9.525 mm (3/8”) thick gusset steel plates designed to take the maximum

lateral load of the actuator. The gusset plates were welded on their vertical edge to the web of the

column and were designed for the maximum tension, shear and moment loads that would occur

when the brace was loaded to maximum capacity. It was very important in the design phase of the
 65

project that the connections be designed with large safety factors to ensure that premature failure

does not occur and the intended failure occurs at the fuse location.

4.4 Experimental Program

4.4.1 Test Setup

The test frame used in this experiment consisted of a steel frame and three actuators, two vertical

and one lateral, as shown in Figure 4-4. The capacity of each vertical actuator was 1.5 MN and the

lateral one was 250 kN with a maximum stroke of 254 mm.

A steel W-section was placed both on top and on the bottom of the specimens to act as a loading

beam and a foundation beam, elements #4 and #7 in Figure 4-4, respectively. The loading beam

was connected to the columns of the specimens by welding a steel plate to the top of the columns

and to the bottom of the loading beam (element #5 in Figure 4-4) and then using 25 mm grade 8

bolts to connect the two members. A similar method was used to connect the frames to the

foundation beam. By using a separate steel I-beam as the loading beam, it ensured that only the

columns would experience an axial load and the cross beam would experience no bending stresses.

A cap plate (element #8 in Figure 4-4) was welded to the end of the loading beam to connect the

lateral actuator to the system as shown in Figure 4-4. As the frame moves laterally with the load,

it was intended to maintain the axial loads at 65 kN, in order to achieve this, rollers were placed

underneath the vertical actuators as shown in Figure 4-5.

 66

Figure 4-4: Test set-up for free vibration and quasi-static cyclic tests

Vertical Actuators

Rollers

Figure 4-5: Placement of rollers beneath vertical actuators on loading beam

 67

4.4.2 Testing Procedure

Two types of test were carried out on each specimen: in-plane free vibration and in-plane quasi-

static cyclic loading. First, the frames were tested in the undamaged condition for the in-plane free

vibration test. The in-plane free vibration test was carried out with no vertical loads applied to the

specimens. For the free vibration test, a cable was fastened to the plate on the loading beam and

then attached to the lateral actuator. The actuator was pulled back horizontally in the in-plane

direction to create a small displacement, but small enough not to cause yielding in the system. The

cable was then cut, releasing the frame to vibrate freely. Then an in-plane quasi-static cyclic test

was performed on the specimens until failure was reached. For this research, failure was defined

as either fracture in the fuse or failure in the connections.

4.5 Free Vibration Test

A free vibration test on a structure occurs when it is disturbed from its static equilibrium state and

is then allowed to vibrate without any external dynamic excitations. From this test, the natural

frequency of vibration and the damping ratio are determined. Figure 4-6 shows the basic set-up for

the free vibrations test for the specimens.

 68

Accelerometers

Figure 4-6: Accelerometer set-up for free vibrations test

Using 6 mm diameter aircraft cable, the maximum tensile load that the cable was able to carry was

9 kN. The specimen was pulled till the actuator reached about 8.8 kN, which resulted in a lateral

displacement at the top of the frame of 0.5 mm for the CBF and 0.57 mm for the SMA-BF. From

structural analysis, the theoretical lateral displacement to cause yielding in the steel fuse is

approximately 7 mm (refer to Appendix A for detailed calculation). Therefore, it can be assumed

that the frame remained in the elastic state during the free vibration test.
 69

4.5.1 Data Acquisition

During the free vibrations test, the displacement of the top of the frame was measured using 2

laser-based displacement transducers, both aligned in the in-plane direction. Three high accuracy

accelerometers were mounted on the frame. Two were placed on each column to measure the in-

plane vibration, and the other was placed on the cross beam to measure the out-of-plane vibration.

The National Instrument (NI) SCXI DAQ system was used to extract data from the accelerometers.

This DAQ was a high performance, multi-channel signal conditioning and switching platform for

automation and measurement systems. The recommended software to connect to the DAQ system

was LabVIEW, which was a graphical development platform for design, control and testing. Data

were acquired at a frequency of 2000 Hz during the free vibrations test. Table 4-5 lists the

performance specifications of the accelerometers.

Table 4-4: Performance specifications of accelerometers (www.te.com)

Model MEAS 4604
Range (g) ±2
Sensitivity (mV/g) 1000
Frequency Response (Hz) 0-300
Natural Frequency (Hz) 700
Transverse sensitivity (%) <2
Damping Ratio 0.7
Shock limit (g) 5000
Residual noise (μV RMS) 600
Spectral noise (μg/√Hz) 38

4.5.2 Results of Free Vibrations Test

4.5.2.1 Frequency analysis

The NBCC (2015), estimates the natural period of braced steel frames to be 0.025 H; where H is

the height of the frame in meters. The height of the braced frame is 2 m, this results in a natural

period of 0.05s/cycle and a natural frequency of 20 Hz. The following Tables 4-5 and 4-6
 70

demonstrate the estimated frequencies obtained from SAP2000, a structural analysis software

program and the NBCC (2015) and compares them to the experimental results in Table 4-7.

Table 4-5: CBF estimated frequencies

Estimated frequency (Hz)
NBCC (2015) 20
SAP 2000 – Pinned Connection 118.5
SAP 2000 – Fixed Connection 122.8

Table 4-6: SMA-BF estimated frequencies

Estimated frequency (Hz)
NBCC (2015) 20
SAP 2000 – Pinned Connection 5.70e-06
SAP 2000 – Fixed Connection 3.77

From Tables 4-5 and 4-6, it can be seen that the predicted values from SAP2000 for the CBF and

SMA-BF vary significantly; suggesting that the difference in elastic modulus between the two

materials impacts the stiffness of the system, as expected. Further, the predicted values between

SAP2000 and the value given by NBCC (2015) vary greatly. The NBCC predicts the natural

frequency of a structure based on the height of the structure multiplied by a factor of 0.025: it does

not take into account the stiffness or mass of the structure. For simplicity, the braces were modelled

as one full length member; in other words, the fuse was not modelled in SAP. For the CBF system,

the L1021026.4 was modelled and for the SMA-BF, the 2006 plate was modelled with the

SMA properties determined in Chapter 3.

For the experiment, a viscous-based free vibration analysis was used to determine the natural

frequency of the system as the yielding of the steel fuse acted as a damper, shortening the natural

frequency and lengthening the natural period from the undamped period to a damped period,
 71

Equation 4-3. However, when the damping ratio is less than 20%, as most structures are, these

effects can be neglected (Chopra, 2012), therefore, TD = Tn and ωn = ωD.

𝑇𝑛 Equation 4-3
𝑇𝐷 =
√1 − 𝜁 2

where

- ζ is the viscous damping ratio;

- TD is the damped period;

- Tn is the natural period;

- ωn is the natural frequency; and

- ωD is the damped frequency.

From the free vibration test, the following results were obtained.

a) CBF b) SMA-BF
Figure 4-7: Accelerometer response for CBF and SMA-BF systemsTHE FONT TYPE AND

SIZE IN THE LOWER TWO FIGURES IS DIFFERENT THAN THE ABOVE ONES,
 72

a) CBF b) SMA-BF

Figure 4-8: In-plane frequency domains for CBF and SMA-BF systems

b) CBF b) SMA-BF

Figure 4-9: Out-of-plane frequency domains of CBF and SMA-BF systems

Table 4-7: Experimental frequencies and damping ratios from free vibration test
In plane frequency Out of plane frequency Damping ratio
CBF 35 Hz 39 Hz 1.3%
SMA-BF 35 Hz 46.8 Hz 1.0%
 73

The in-plane free vibration tests revealed some important information regarding the design of the

braced systems. Typically, braced systems are a superposition of two different systems: one being

a rigid frame that supports the vertical loads, and the second is a vertical bracing system that is

regarded as a pin-connected truss that resists the lateral forces. Therefore, the lateral stiffness of

the braced frame can be estimated as the sum of the lateral stiffness of the individual braces as

seen in Equation 4-4 (Chopra 2012).

𝐴𝐸
𝑘𝑏𝑟𝑎𝑐𝑒 = 𝑐𝑜𝑠 2 (𝜃) Equation 4-4
𝐿

where

- A is the cross sectional area of the brace

- E is the elastic modulus

- L is the length of the brace

- θ is the angle of the brace

Table 4-8: Theoretical and experimental stiffnesses of CBF and SMA-BF systems
Theoretical Stiffness* Experimental Stiffness % Difference
(kN/mm) (kN/mm)
CBF 9.5 17.6 -7.81
SMA-BF 2.05 15.4 651.2
*Note: With the assumption that the entire cross brace length is composed of the yielding fuse material.

However, from the test results, both frames have the same in-plane natural frequency (Table 4-6).

This suggests that the fixation at the columns dominates the rigidity of the system rather than the

stiffness of the braced members. The NiTi-SMA brace is approximately 22% the stiffness of the

steel brace, it would be expected for the frequency of the NiTi-SMA brace to be higher than the

steel brace. However, because of the partial fixity, the frames behave more like a rigid frame than

a pin-connected frame. Furthermore, the experimental stiffnesses of both systems are relatively
 74

close to each other, Table 4-8, which again suggests that the frames behave similar to a rigid frame.

Therefore, the stiffness of the braces does not affect the in-plane natural behaviour of the frame.

From the SAP2000 results in Tables 4-5 and 4-6, it can be seen that the predicted results are greatly

different than the experimental. There are two main components of natural frequency: the stiffness

of the system and its mass. Equation 4-5 shows the relationship between frequency, stiffness and

mass for a multiple degree of freedom system (Chopra, 2012).

𝑴𝒒̈ + 𝑲𝒒 = 𝑷(𝑡 Equation 4-5

where

- M is a diagonal matrix of the generalized modal masses Mn

- q is a vector of modal coordinates

- K is a diagonal matrix of the generalized modal stiffnesses Kn

- P(t) is a column vector of the generalized modal forces Pn(t)

From Equation 4-5, the frequency of the system at different modes is found by determining the

eigenvalues. From an analysis perspective, the parameters that are to be manipulated to achieve a

certain frequency are the mass and the stiffness of the system. However, since the members and

elastic modulus of the system are already known, only the mass needs to change. By adding an

additional 14 kN of gravity load to the CBF SAP system, a frequency of 34.8 Hz was found.

Further, for the SMA-BF system, the predicted frequency was far less than the experimental,

suggesting that within the SAP model, the mass of the system had to be reduced or the stiffness

had to be increased. The SMA-BF system had an elastic modulus of 42 GPa in the austenite phase,

as determined in Chapter 3, Section 3.3.1; from the SAP model this reduction in stiffness had a
 75

severe impact on the predicted frequency. However, as seen in the experimental results, the make-

up of the brace did not affect the natural behaviour of the frame. Overall, from the additional load

in the CBF system and the reduced stiffness of the SMA-BF system, it is postulated that the rigidity

of the connection points in the system controlled the way the braced systems behaved under the

free vibration test.

It should be noted that the out-of-plane frequencies are different for the different systems. In the

out-of-plane direction the stiffness of both systems is less as it is the weak axis, thus giving a higher

frequency. Reasons for the systems having different out-of-plane frequencies is due to the

difference in plastic hinge, one being steel, the other NiTi SMA. Furthermore, the steel bracing

member for each system is different. For the steel braced frame, a L1021026.4 angle is used

and for the NiTi-SMA brace, a 2006 mm plate is used. The angle in the steel frame, increases the

out-of-plane stiffness of the system compared to using a 6 mm thick plate, thus having a lower

frequency than the SMA frame.

4.5.2.2 Damping ratio analysis

In structures, there can be several energy dissipating mechanisms acting simultaneously, and

typically they are idealized by equivalent viscous damping. There are some cases where it may be

more appropriate to use the Coulomb damping method of analysis when Coulomb frictional forces

are utilized. However, in the case of a steel braced frame, due to its simplicity using the equivalent

viscous damping method is sufficient (Chopra, 2012). For an underdamped single-degree-of-

freedom (SDOF) system, the equation of motion is of the form:

𝑢(𝑡) = 𝑒 −𝜁𝜔𝑛𝑡 (𝐴𝑐𝑜𝑠𝜔𝐷 𝑡 + 𝐵𝑠𝑖𝑛𝜔𝐷 𝑡) Equation 4-6

 76

where A and B are real-valued constants determined from the initial conditions of the system.

The important effect of damping is the rate at which the free vibration decays. For lightly damped

systems, such as the steel braced frame, the damping ratio can be determined from the following

equation:

1 𝑢̈ 𝑖 Equation 4-7
𝜁= ln( )
2𝜋𝑗 𝑢̈ 𝑖+𝑗

where üi and üi+j are peaks of the acceleration-time graphs.

A sample of the experimental measurement from the accelerometer #1 of acceleration versus time

is shown in Figure 4-10.

Figure 4-10: Response acceleration of CBF under free vibration

In this study for damping analysis, the acceleration-time relationships were used to determine the

damping ratio of the systems. It was found that the steel braced frame had a slightly larger damping
 77

ratio than the NiTi-SMA frame, with it being 1.3% and the NiTi-SMA frame being 1.0%. The

slight variation in damping ratio suggests that like the natural frequency, the stiffness of the system

as a whole is dominating the behaviour of the natural system as opposed to the fuse affecting the

behaviour.

4.6 Quasi-Static Cyclic Loading Test

For the quasi-static cyclic test, the structural parameters which were investigated included the

maximum lateral displacement, the hysteresis response, and the dissipated energy.

4.6.1 Test Regime

The columns of each specimen were subject to an axial load of 65 kN to simulate real world

situations. Roh and Reinhorn (2010), suggested gravity loads between 5% and 20% of the

column’s axial strength; the 65 kN was roughly 7% of the columns axial strength. Then the

specimens were subjected to a quasi-static loading in a displacement control mode. The quasi-

static testing program started at a displacement of 2 mm in both positive and negative directions,

then increased in increments of 2 mm at a rate of 0.5 mm/s. Once drift levels reached 1%, the

displacement increased to increments of 4mm at a rate of 1 mm/s up to a drift of 3% in both positive

and negative directions and then increased to 2 mm/s. From 0-1% drift, one cycle was conducted

at each increment; after 1% drift was achieved, 3 cycles were conducted at each displacement

increment, as shown in Figure 4-11. It can be noted that the loading protocol from this research

was adopted from the FEMA recommendation for steel structures (FEMA, 2009). Investigating

the behaviour of the SMA-BF system was the primary objective of this research, and due to the

unique characteristics of NiTi it was decided to use a loading protocol that was slower with more

cycles.
 78

Figure 4-11: Displacement-controlled loading history

4.6.2 Data Acquisition

The specimens were instrumented with laser-based displacement transducers (LDT) with a ±1 mm

accuracy, linear strain converters (LSC) with a 0.05% accuracy, and strain gauges (SG) with a

0.2% accuracy, to monitor the behaviour during testing. The LDTs were aligned in both the in-

plane and out-of-plane direction to measure displacement. An LDT was placed on the North side

of the specimen, to measure the in-plane displacement. One LDT was placed on each fuse to

measure the out-of-plane displacement when the respective brace went into compression. An LSC

was placed at each column to measure any uplift. It was decided to use LDT’s to monitor and

measure the displacements of the specimens as large displacements were expected and the LDT’s

had a larger range. Furthermore, from past experiments, LDT’s are more reliable and accurate than

LSC’s for measuring displacements greater than 50 mm. For the CBF, PI displacement transducers

were installed to measure the axial strains in the yielding fuse. For the SMA-BF, six strain gauges
 79

were installed on individual wires, per brace. Figures 4-12 and 4-13 shows the instrumentation set-

up for the CBF and SMA-BF systems, respectively.

1 – PI displacement 2 – West out-of-plane 3 – East out-of-plane LDT 7 – Out-of-

transducer LDT plane LDT
4 – South LSC 5 – North LSC 6 – In-Plane LDT

Figure 4-12: Instrumentation set-up for quasi-static cyclic test for CBF

1 – Strain Gauges placed 2 – West out-of-plane 3 – East out-of- 7 – Out-of-plane

on SMA wires LDT plane LDT LDT
4 – South LSC 5 – North LSC 6 – In-Plane LDT
Note: Not all PE NiTi SMA wires are shown for clarity purposes

Figure 4-13: Instrumentation set-up for quasi-static cyclic test for SMA-BF
 80

4.6.3 Results

4.6.3.1 CBF Hysteretic Response

The hysteretic load-displacement relationships of the CBF specimen is shown in Figure 4-14. The

failure mode of the specimen was fracture in the steel brace at approximately 2% lateral drift. The

fuse remained elastic and stable for small drift levels (below 0.7%). At 0.7% drift, the fuse began

to yield, which resulted in a decrease in stiffness in the subsequent cycles. This trend continued as

the steel fuse accumulated residual strain, and eventually fractured. The failure mode of the fuse

was in tension at a lateral load of approximately 127 kN. The expected fracture load was 95 kN,

this difference can be attributed to the design assumption that the compressed brace does not take

any load.

Plate Fracture

Figure 4-14: Hysteresis Loop - CBF

 81

The CBF failed in tension during the third cycle of the 36 mm displacement amplitude cycle. It is

observed that the hysteretic response shows a pinching behaviour due to the accumulation of

residual deformations formed through inelastic buckling during subsequent compression cycles.

This pinching behaviour leads to a reduction in stiffness and the amount of energy dissipated

during the cyclic loading (FEMA, 2009). Figure 4-15 shows the fracture surface of the steel fuse;

as expected, due to the additional stress concentrations at the connection point between the brace

angle and the steel fuse, the failure of the yielding fuse occurred along this interface.

Figure 4-15: Fracture surface of steel yielding fuse in CBF

In a typical tension test of a ductile metal, such as mild steel, it would be expected to see some

necking and then a cup and cone deformed surface as seen in Figure 4-16.
 82

Figure 4-16: Fracture surface from coupon tension test

However, the fracture surface of the steel yielding fuse had minor necking and a smooth surface,

similar to a brittle failure. This type of failure is classified as low cyclic fatigue failure.

4.6.3.2 SMA-BF Test 1 Hysteretic Response

The hysteretic load-displacement response of the first SMA braced frame (SMA-BF-1) is shown

in Figure 4-17. The response of the system achieved a self-centring response and a target capacity

of 4% drift of the performance objectives (Section 4.2), as evident from the hysteretic response.

The system, however, experienced insignificant energy dissipation capability. This was due to the

slippage of the SMA wires inside the anchor upon load-reversal. As the tensioned brace was

undergoing the reverse loading, the wires started bending out-of-plane. However, as it was being

loaded in tension, the wires started tensioning. This behaviour led to significant reduction in the
 83

structural stiffness and the energy dissipation capability of the system. The slippage did not allow

for the full utilization of the material.

The SMA-BF-1 reached a maximum load of 122 kN on the first cycle of the 2% lateral drift

increment, before dropping off. At this point in the experiment there was no visible indication of

failure, all the connections were still structurally viable. Therefore, it was concluded that at this

point, the wires started slipping.

Max Load

Figure 4-17: Hysteresis Loop - SMA-BF-1

 84

4.7 Comparison of Results

4.7.1 Re-centring

An advantageous property of PE NiTi SMA is the ability to be strained up to 8-10% and return to

its original shape without residual deformations (DesRoche et al., 2004; Memry, 2018). The re-

centring ability of the proposed system was evaluated by examining the residual strains recorded

after the last cycle at each drift cycle. The residual strains were measured using strain gauges

mounted on the SMA wires in the SMA-BF system and mounted on the steel fuse plate in the CBF

system. The residual brace strain at the end of the last cycle at each drift cycle is shown in Figure

4-18. The amount of residual strain developed in the CBF fuse was 0.41% at a 1% drift compared

to the SMA wires which was 0.04% at a 1% drift. It is, thus, concluded that the SMA-BF specimen

experienced a self-centring behaviour. It is interesting to note that though visible slipping of the

wires was observed, it did not have an effect on the re-centring properties of the system. With the

slipping, a gradual loss of the re-centring capacity would be expected, however that is not the case

as seen in Figure 4-18. An explanation for this is that the wires never slipped out of their connection

completely, refer to Figure 4-19; a repetitive sequence of minor slipping then the clamps re-

grabbing hold of the wire continued for the duration on the test thus still providing re-centring

abilities of PE NiTi-SMA wires.

 85

Figure 4-18: Residual strain in the brace after the last cycle at each drift level

(a) Before quasi-static cyclic test (b) After quasi-static cyclic test

Figure 4-19: Slipped wires in anchorage system

 86

4.7.2 Energy dissipation

As seen in Figure 4-17, the amount of energy dissipated in the SMA-BF specimen is very small

compared to that of the CBF specimen shown in Figure 4-14. This is further detailed in

Figure 4-20 as the value of energy dissipated was calculated for the first cycle at each drift level.

At approximately 1% drift, both systems had reached their yield; the CBF exhibited a larger energy

dissipated than the SMA-BF. Then, as the drift level increased, the CBF energy dissipation

increased at a faster rate than the SMA-BF. As the drift increased past 2%, the energy dissipated

for the SMA-BF started to plateau. An explanation for this, is that the wires were slipping within

their anchorage system before they were able to reach their full capacity.

Figure 4-20: Energy dissipated over a range of drift levels for CBF and SMA-BF-1

4.8 Behaviour of Anchorage System

The SMA-BF was designed to be a tension-only braced system, therefore, when one of the braces

was in tension the other was in compression and was assumed to have buckled. The NiTi-SMA

wires used in the experiment had approximately 20% of the stiffness of steel, leading to a greater
 87

deformation when loaded in compression, thus buckling was considered to have occurred when

the brace experienced any out-of-plane displacement. The amount of out-of-plane displacement

was unforeseen and caused an additional tension load on the anchorage system when the brace was

in compression. This additional load was greater than the design and caused the wires to slip out

of their anchorage. Figure 4-21 shows one of the braces in tension and the other in compression.

Tension Brace Buckled Brace

Figure 4-21: Initial behaviour of SMA-BF

Slippage occurred at approximately 2% drift, which corresponded to a 100 mm out-of-plane

displacement of the buckled brace. The slippage did not allow for full utilization of the material

and therefore, the system had to be modified in order to assess the feasibility of using NiTi-SMA

in a braced frame.

4.8.1 Development of Improved Connection System

Due to the occurrence of slippage experienced with the initial design of the SMA-BF, it was

necessary to develop a new connection system that would provide a greater resistance. Two
 88

modifications were made to the system, one was made to improve the frictional resistance of the

anchorage system, the second was to eliminate the out-of-plane displacement. To increase the

frictional resistance, the wires were glass beaded to remove the black oxide finish and increase the

coefficient of friction. Additionally, a third row of clamps was installed; the new connection plate

is shown in Figure 4-22.

Three rows of clamps

to increase frictional
resistance.

Figure 4-22: New developed anchorage system

To eliminate the out-of-plane displacement, a slotted connection was designed in accordance to

CSA S16-14 and installed on the SMA-BF as shown in Figure 4-23. The slotted connection plates

were oiled well to ensure that the steel-on-steel interfacial friction was minimized.
 89

Slotted
connection

Figure 4-23: Installed anchorage system and slotted connection on SMA-BF

4.8.2 Behaviour of Improved SMA-BF

Testing of the SMA-BF, with the improved connection system used in conjunction with the slotted

connections, significantly reduced the out-of-plane displacement and slippage of the SMA wires.

Furthermore, the SMA-BF achieved higher loads and higher drift levels prior to fracturing of the

SMA wires.

4.8.2.1 Results of Improved SMA-BF

The hysteretic load-displacement response of the second SMA braced frame (SMA-BF-2). with

the improved system is shown in Figure 4-24. Similar to the first test, the system was evaluated

against three performance objectives and all three objectives were achieved. From the hysteretic

response, the permanent deformation upon unloading was insignificant compared to the CBF and

similar to the SMA-BF-1 performance.

 90

Twelve strain gauges were placed on the SMA-BF-2 wires: however, only one strain gauge

collected useable data. During the improvement of the frame, the wires had to be removed from

the original anchorage and then put back in place when the third row of clamps was welded into

place. It is thought that the strain gauges were damaged during this process. The SMA-BF-2

reached a maximum load of 160 kN on the first cycle of the 4% lateral drift increment. At this

cycle a couple of SMA wires fractured.

Figure 4-24: Hysteresis response of SMA-BF-2

With the slotted connection plates added to the braces the out-of-plane displacement at the fuse

location was significantly reduced. At 2% drift, the out-of-plane displacement at the fuse location

of the original SMA-BF was 108 mm, for the SMA-BF-2 it reduced to 4.5 mm. Furthermore, with

the added row of clamps to each anchorage plate, the frictional resistance was increased and

eliminated slippage up to 3.5% lateral drift. At approximately 3.5% drift, the upper gusset plates
 91

failed at the welds, as seen in Figure 4-25. The test was still continued up to 4.5% drift: however,

due to the gusset failing at the top, the applied load to the SMA connection was not evenly

distributed and thus a moment was applied. The SMA wires at the bottom chord took more load

than the top chord of the brace, more than the resistance, and began slipping as seen in

Figure 4-26.

Figure 4-25: Failed gusset plate at approximately 3.5% drift

Figure 4-26: Uneven load distribution due to failed gusset plate

 92

Though the addition of the slotted connection plates significantly reduced the out-of-plane

displacement of the brace when reversed loading (compression loading) was applied, some out-

of-plane displacement still occurred. It was noted during the experiment that there was a lot of

friction between the slotted plates and the steel brace preventing that sliding action. It is

recommended that with slotted plates, to add an additional material, with a smaller coefficient of

friction, between the two plates to reduce this friction.

The energy dissipation ability of NiTi-SMA has been studied by researchers (Haque and Alam,

2017; Varela and Saiid, 2016; Oudah, 2014; Speicher et al., 2011; DesRoches et al., 2004) and it

has been found that NiTi-SMA wires are adequate at dissipating seismic energy. With the SMA-

BF-1, that was not the case due to the SMA wires slipping in the anchorage system. However,

Figure 4-27 shows the energy dissipated during the first cycle at different drift levels of all three

systems. Due to the improved system, the SMA-BF-2 was able to fully utilize the NiTi-SMA wires

and dissipate a signifcant amount of energy compared to the CBF. Energy is dissipated through

the inelastic response of the system, from the Figure 4-27, the increase in energy dissipation from

1% to 2% drift suggests that at approximately 1% lateral drift, the NiTi-SMA wires started to yield.

Figure 4-27: Energy dissipated over a range of drift levels

 93

Another means of measuring the energy dissipation capacity of a system is through the equivalent

viscous damping, ζeq. as described in Chapter 3 Section 3.4.1. The values of the equivalent viscous

damping ratio at the ultimate displacement cycles are included in Table 4-9 for the CBF and SMA-

BF-2 systems.

Table 4-9: Equivalent Viscous Damping results for the CBF and SMA-BF-2 systems
CBF SMA-BF-2
ζeq 0.0281 0.0460

In addition to the SMA-BF-2 system having a significant capacity to dissipate energy, it had

adequate damping abilities compared to the CBF system.

4.9 Stiffness Degradation

When steel structures undergo cyclic loading, such as those in a seismic event, they are subject to

inelastic behaviour. The inelastic behaviour is a result of certain elements, such as a plastic hinge,

yielding and dissipating energy. The result of yielding causes a stiffness degradation in the system.

Figure 4-28 shows the degradation of the CBF, SMA-BF-1, and SMA-BF-2 systems over the drift

cycles. Based on the initial design of the frames as stated in Section 4.3 the stiffness of the systems

was based on the stiffness of the braces Equation 4-4: as the area of both the steel fuse and the

NiTi SMA fuse was the same. Therefore, it would be expected that the SMA-BF systems have a

lower stiffness due to the lower modulus of elasticity. However, as seen, that is not the case. An

explanation for this, could be due to the different sections used for the steel brace. For the CBF

system an angle was used and for the SMA-BF systems, plates were used. If the system was acting

as a true braced frame, with pinned connections at all joints and the members only experience axial

loads, the stiffness of the SMA-BF system would be lower than the CBF due to the NiTi SMA
 94

wires having a lower stiffness. However, from the results it can be inferred that these systems were

not a true pinned framing system and the braced members experienced bending stresses which in

turn, affects the stiffness of the entire system. Since, the SMA-BF systems had the 200x6 mm steel

plate which had a larger stiffness than the angles used in the CBF system, during the quasi-static

cyclic loading test, the SMA-BF systems experienced a larger initial stiffness than the CBF system.

Figure 4-28: Stiffness degradation of CBF and SMA-BF systems

As expected, when the system yields there is a significant drop in stiffness for both the CBF and

SMA-BF-2 systems and then a gradual decay. For the SMA-BF-1 system, at 1% drift there is a

significant drop in stiffness which shows when the wires started slipping out of their anchorage,

then in subsequent cycles with the wires slipping at different rates the change in stiffness was not

consistent.
 95

For the SMA-BF-2 system, there is a gradual decay in stiffness which indicates that the wires are

yielding. Further, for the SMA-BF-2 system there are two parts to the stiffness degradation. From

Figure 4-24, as the system is undergoing the loading cycle there is a plateau where the stiffness of

the system is nearly zero then the system regains its stiffness, this is illustrated in Figure 4-29.

Figure 4-28 shows part 2 of the stiffness degradation over the drift levels, where a line of best fit

was taken for the loading curve. As explained in Chapter 3, the stiffness of PE SMA is different

for the loading and unloading curve due to the phase transformation that takes place.

Stiffness
Degradation Part 2

Stiffness
Degradation Part 1

Figure 4-29: Different stiffness degradation experienced by SMA-BF systems

For the loading curve, the material is in the austenite phase, which has a higher modulus of

elasticity (stiffer), and in the unloading portion, the material is in the martensite phase which has

a lower modulus of elasticity. For all drift cycles, the line of best fit was applied to the loading

portion of the hysteresis curve. Part 1 of the stiffness degradation was measured by taking the

displacement at which the line of best fit from Part 2 crossed the x-axis and plotting how that
 96

changed over the drift cycles. A plot of the stiffness degradation of the SMA-BF-1 and

SMA-BF-2 systems is shown in Figure 4-30.

Figure 4-30: Part 1 of the stiffness degradation in the SMA-BF systems

Part 1 of the stiffness degradation of the SMA-BF-1 system, shows that the system was stable up

to 1% drift, then when slip occurred, severe degradation occurred. For the SMA-BF-2 system, the

system was relatively stable up to 2% drift, at which the system would have already started

yielding. After 2%, the stiffness degradation significantly increased; this could suggest that some

minor slipping occurred. As there were no visible signs of slippage until 3.5% drift was reached,

the slipping that occurred prior could have been very small, on the micro-scale. However, a more

plausible explanation would be that the PE SMA wire underwent permanent plastic deformation,

leading to an increase in length. As described in Chapter 2, Section 2.3.1, the behaviour PE NiTi

SMA wires is dependent on both mechanical and thermal properties of the material. With the PE

behaviour, the SMA is initially in the austenite phase, when loaded it undergoes a phase

transformation to Martensite. As the phase transformation takes place, energy is released, and the
 97

material has a slightly increased temperature. Based on the hysteresis response, it is suggested that

due to the unloading rate and the test being conducted at room temperature in an uncontrolled

environment, the material did not have the time to fully return to its austenitic parent phase. To

investigate this hypothesis, separate testing of the PE NiTi SMA wire was conducted to study the

rate at which the wire undeforms: this is discussed in the following section.

4.10 Loading Rate Assessment

In the previous section, it was hypothesized that the stiffness degradation in the SMA-BF-2 system

was caused by the PE NiTi SMA wires undergoing permanent plastic deformation when being

reloaded. To investigate this further, three PE NiTi SMA wires were tested in an uncontrolled

environment to assess the rate at which the wires return to their original state when loaded to a

specific strain. Similar to the setup for the cyclic test described in Chapter 3, the specimens were

placed in an MTS testing apparatus with aluminum sleeves at the ends. An extensometer and an

LSC were used to monitor the displacements. The temperature of the specimens was monitored

using three thermocouples as shown in Figure 4-31.

Figure 4-31: Placement of thermocouples on NiTi SMA wire

 98

The loading protocol consisted of increasing the strain in the wire, starting at 0.5%, then 1.0-8.0%

at increments of 1% using the displacement-controlled function of the MTS. At each strain limit,

the MTS machine was switched to force-control and a 0 kN load was entered. Two different types

of tests were carried out; the first test consisted of the specimens being loaded and released at room

temperature. The second test consisted of the specimens being loaded at room temperature and

then when the machine was switched to 0 kN load a heat gun was used to increase the temperature

of the wire to 70°C. Figure 4-32 depicts the displacement-time behaviour of the first test at the

0.5% strain cycle.

Figure 4-32: Displacement-Time relationship at 0.5% strain

The time from when 0 kN load is entered to t0 is the MTS machine stabilizing at 0 kN. Time t0 is

when the PE NiTi SMA wire begins to return to its undeformed shape and t1 is when the wire is

fully undeformed.
 99

Figure 4-33: Relationship between Displacement, Load and Time at 0.5% strain

By overlaying the displacement-time graph with the load-time graph, Figure 4-33, it can be shown

that once the wire is undeformed the machine registers a load.

Figure 4-34: NiTi SMA wire returning to original shape

 100

The time between t0 and t1 is the time it takes the PE NiTi SMA wire to return to its undeformed

shape. The change in deformation over the amount of time it takes for the wire to undeform results

in the rate of change of the NiTi wire. Figure 4-35 shows the rate of change of the NiTi wire for

both tests as the strain increases with each pull.

Figure 4-35: Rate of change of NiTi wire vs. strain relationship

The trend of both tests is similar and as expected, when energy (heat) is added to the system, the

rate at which the wire undeforms increases. It is interesting to note that there is a sharp increase in

rate up to approximately 3% strain and then it starts to plateau. At 3% strain, the wire is halfway

through the forward transformation and at 8% is fully martensitic, this is detailed in Chapter 3,

Section 3.3.1. This suggests that the rate of change of the material is unaffected by the applied

strain when the strain is larger than 3%.

From conducting the two types of tests it can be seen that the rate at which the wire returns to its

original form is affected by temperature. From these results, it can be concluded that the SMA-BF
 101

systems were subjected to a loading protocol that was too fast for the wires to return to their parent

phase before being reloaded. Thus, this non-return to the parent phase is the likely cause of the

severe stiffness degradation.

Further research should be devoted towards the temperature effects of this PE NiTi SMA material

to better understand the effects of different temperatures, loading rates, and wire sizes.

4.11 Ductility Assessment

In Section 4.8, the energy dissipation capacity of the frames was assessed, and it was found that

the SMA-BF-2 system had a high energy dissipation capacity. Here, the ductility of the CBF and

SMA-BF-2 systems will be examined. Ductility is the ability of a structure to undergo significant

inelastic deformation without loss in structural capacity. In steel structures, a higher ductility

suggests high deformation and energy dissipation capabilities due to the plastic property of the

steel. However, for other structures such as concrete structures reinforced with elastic materials

like FRP, for instance, high ductility implies high deformation but low energy dissipation (Oudah

and El-Hacha, 2014). Therefore, it is important to identify both the deformation ability and energy

dissipation capacity of a structure.

The overall ductility factor is defined as the ratio between ultimate displacement and yield

displacement (Alagawani and Al-Qaryouti, 2016), and is expressed as follows:

∆𝑢
𝜇=
∆𝑦 Equation 4-2

where Δu and Δy are displacements at ultimate and yield points, respectively.

 102

Table 4-10 shows the ductility parameters for the CBF and SMA-BF-2 systems. Though the SMA-

BF-2 system experienced excellent re-centring capabilities and a large energy dissipation capacity,

the ductility of the system is significantly lower than the CBF system. In reinforced concrete

structures reinforced with PE NiTi-SMA at connection joints, it is expected that the ductility would

be lower due to the lower stiffness of the PE NiTi-SMA. The same can be said here as the SMA is

acting as a type of connection. The reduced stiffness of the PE NiTi-SMA fuse results in a reduced

stiffness of the overall system which leads to a high displacement at yielding compared to the

control. Future research should be devoted towards enhancing the performance of the anchorage

system so that the PE NiTi-SMA wires yield at a lower displacement, and consequently, increase

the ductility capacity of the bracing system.

Table 4-10: Ductility assessment of the CBF and SMA-BF-2 systems

Δy (mm) Δu (mm) μ
CBF 14 36 2.57
SMA-BF-2 57 80 1.40

4.12 Summary

The behaviour of the tested CBF and SMA-BF systems was examined in this chapter. The natural

behaviour (free vibrations test) and cyclic behaviour of the systems was examined experimentally.

A free vibration test was conducted on the CBF and SMA-BF-1 systems to assess how the

difference in fuse material affected the natural behaviour of the system. It was found that due to

the partial fixity of the system to the foundation beam, the frames behaved like a rigid frame rather

than a traditional braced frame. Therefore, the stiffness of the system depended on the stiffness of

the columns rather than the cross braces. Next, a quasi-static cyclic test was conducted on the CBF

and SMA-BF systems. Furthermore, the anchoring of the PE NiTi-SMA wires to the braced system

was modified in order to minimize slip and out-of-plane displacement, and consequently,
 103

increasing the stress and strain capacity of the PE NiTi-SMA wires at failure. The results based on

the hysteretic response, energy dissipation capacity, and damping abilities indicate superior

performance of the proposed system in achieving self-centring behaviour, and thus, reduce the

amount of permanent damage done to the rest of the structure. With the improved connection,

there was still moderate stiffness degradation in the SMA-BF-2 system, the exact cause is unknown

and further improvements to the anchorage and slotted connection plates should be done to

completely eliminate out-of-plane displacement and slip. A simple study on the temperature

effects of the PE NiTi-SMA was conducted to assess if the loading rate resulted in the severe

stiffness degradation of the system. From this study it was found that adding heat to the wire as it

returned to its original shape increased the rate of change of the wire: however, it was still smaller

than the loading rate that was applied to the SMA-BF system. Therefore, it is concluded that the

loading protocol used in this study should be modified accordingly to limit the stiffness

degradation of the system. Further, it is recommended that more in-depth temperature studies be

conducted on PE NiTi material to better understand the effects temperature has on the material.

The aim of this research was to investigate the effectiveness and feasibility of using PE NiTi SMA

wires as an active technique for seismic design and retrofitting of structures. From the results of

this study, it was found that the PE NiTi wires are sensitive to the loading rate; which may be an

issue for seismic design as the frequency and amplitude of earthquakes varies throughout the

world. Therefore, it is recommended that further testing, such as shake table tests, be conducted

on these systems to assess their effectiveness under varying frequencies.

 104

Chapter 5: Conclusions and Recommendations

5.1 Introduction

Three large-scale frames were examined experimentally; a control steel braced frame (CBF), PE

NiTi SMA-reinforced braced frame (SMA-BF-1), and a SMA-BF with an improved anchorage

system and slotted connections (SMA-BF-2). The main objective of this thesis research was to

investigate the behaviour of the proposed steel braced frame reinforced with NiTi SMA wires and

determine its performance in terms of hysteretic damping and self-centring capabilities. Through

experimental testing and analysis of large-scale steel frames, it was determined that the SMA-BF

system performed very well under cyclic loading conditions. It should be noted that the results

obtained from the experimental testing provide a representative indication regarding the

performance of the specimens and not an exact behaviour. This is due to the limited number of

specimens, the level of variability in material properties (steel and SMA), and the level of

variability in the testing procedure (accuracy of maintaining a constant vertical load to the columns

through the loading beam and the quality of construction). The major conclusions and

recommendations drawn from this research are summarized below.

5.2 Conclusions

5.2.1 Free Vibration Loading

Free vibration tests were carried out on the specimens to investigate the effect NiTi SMA wires

had on the natural behaviour of the system. It is important for designers to understand the natural

characteristics of a structure when designing for large seismic loading.

 105

• Although there were different materials at the plastic hinge location, the natural frequency

of the CBF and SMA-BF systems was the same. This suggested that the end (boundary)

conditions of the frame dictated the behaviour of the frame. In the design phase, the frames

were considered to be pin-pin, meaning that rotation was allowed to occur at the column-

foundation connection. However, the connection of the column to the foundation was done

using a base plate. Therefore, a partial fixity was created. From the results it can be

concluded that the partial fixity controlled the behaviour of the frame causing it to behave

more like a rigid frame than a braced frame.

• The damping ratios of each specimen are very close, further suggesting that the braces do

not contribute to the in-plane lateral stiffness of the system.

• The out-of-plane frequencies of the systems are different mainly due to the different

bracing members used in the design.

5.2.2 Quasi-Static Cyclic Loading

Quasi-static cyclic loading was carried out on the three braced frames to study the structural

performance of the developed NiTi-SMA system compared to the standard steel braced frame with

tension-only plates.

• The SMA-BF-1 system possess excellent self-centring behaviour. The amount of residual

strain developed in the CBF fuse was 0.41% at a 1% drift compared to the SMA wires

which was 0.04% at a 1% drift. It is, thus, concluded that the SMA-BF specimen

experienced a self-centring behaviour.

 106

• The SMA-BF-1 system sustained high drift ratios (up to 5%), as compared with the CBF

system which only experienced 1.8%. Therefore, this system can be used as a braced

system for structures that are subjected to high drift demands.

• The energy dissipation ability of the SMA-BF-1 system was low compared to the CBF

system. During the test, the SMA wires began slipping out of their anchorage system at a

lower drift than expected, therefore they were unable to reach their full potential.

• Modifications were made to the system to enhance the performance of the NiTi SMA wires

as well as to reduce the amount of out-of-plane bending due to compressive forces in the

braces under reverse loading. These modifications included, adding an additional row of

clamps to each set of wires, thus increasing the frictional resistance. Slotted connection

plates were added to each brace to allow the wires to remain in tension-only and thus

minimize the out-of-plane movement.

o The improved system minimized slip and thus increased the stress and strain

capacity of the NiTi-SMA wires at failure.

o With the improved system, the SMA-BF-2 system was able to fully utilize the NiTi-

SMA wires, where fracturing of the wires occurred, a significant amount of energy

was dissipated compared to the SMA-BF-1 and CBF systems.

o The SMA-BF-2 system had a 4.6% equivalent viscous damping ratio compared to

the CBF system which had a 2.8%. Suggesting that the SMA system had adequate

damping abilities.

o Even with the improved connection, the SMA-BF-2 system experienced moderate

stiffness degradation which resulted in a lower ductility response compared to the

CBF system.
 107

• A simple temperature study was conducted on the PE NiTi-SMA wires and it was found

that the cause of the stiffness degradation was due to the system being subjected to a

loading rate faster than the rate at which the PE NiTi-SMA wires were able to return to

their original shape, thus causing them to be partially deformed when being reloaded. This

effect resulted in the severe stiffness degradation.

5.3 Recommendations for Direction of Future Research

5.3.1 Experimental Work

• It is recommended to use trained NiTi-SMA wires in order to maintain a stabilized

hysteresis response when subjected to cyclic loading.

• A thorough cyclic response study (dynamic loading, varying loading rates, controlled vs.

uncontrolled environment etc.) of the NiTi-SMA wires should be done to understand their

cyclic behaviour and to study the fatigue life of the material.

• The NiTi SMA wires are sensitive to temperature, therefore, it is recommended to further

study the temperature effects on the performance of the NiTi-SMA material.

• From the free vibration test conducted on the braced frame systems determined that the

rigidity of the frame dominated the behaviour of the system. It is recommended that for

future work, the design of a true pin connection be implemented to study the combined

effects of allowing a system to rock with the NiTi-SMA ability to re-centre the system once

the load is removed.

• Though the addition of the slotted plates reduced the out-of-plane displacement

significantly, there was still noticeable friction between the brace and the plates during the
 108

experiment. It is recommended that a material with a smaller coefficient of friction (for

example: a plastic), to reduce the friction between the slotted plates and the steel brace.

• This experimental program has shown the potential of steel braced frames reinforced with

NiTi SMA wires. However, it is recommended that more specimens be tested to investigate

the consistency of the behaviour of the system.

• To assess the effectiveness and feasibility of using the PE NiTi SMA braced frame for

seismic design, shake table testing of these systems should be conducted to assess their

performance under varying frequency and amplitude loading rates.

5.3.2 Finite Element Work

• In addition to the experimental work, a finite element model for the CBF and SMA-BF

systems was created with the intention of conducting a parametric study on the systems.

However, due to time constraints a full analysis was unable to be conducted in this research

phase. However, a finite element analysis (FEA) was started and details on findings and

recommendations can be found in Appendix B.

• A parametric study on the design parameters, such as: aspect ratio of the braced frame,

member sizes, elastic modulus of the fuse, amount of fuse material used etc., should be

conducted using FEA in order to examine the effect of varying SMA properties (area,

elastic modulus, thermal and mechanical properties) on the self-centring abilities, the

energy dissipation ability and ductility of the braced frame reinforced with NiTi SMA

wires.

• Within the finite element model, constitutive material models should be developed to

obtain the correct cyclic behaviour for steel and PE NiTi SMA. Further, a mesh
 109

sensitivity analysis should be performed on the models to examine the effect of the mesh

size on the solution and to determine the optimum mesh density to be used in this finite

element model.
 110

References

Abdelrahman, K., (2017). “Performance of Eccentrically Loaded Reinforced Concrete Columns

Confined with Shape Memory Alloy Wires”. PhD Thesis. University of Calgary.

Abou-Elfath, H., (2017). “Evaluating the Ductility Characteristics of Self-Centring Buckling-

Restrained Shape Memory Alloy Braces,” Journal of Smart Material Structtures, vol. 26,

no. 5., 13pp.

Alagawani, B., and Al-Qaryouti, Y., (2016). “Evaluating Overall Ductility Factor of Steel Frames

with Viscoelastic Bracing System”. Journal of Engineering Science and Technology

Review, 9(4), 128–137.

Alam, S., Moni, M., Tesfamariam, S., (2012). “Seismic Overstrength and Ductility of Concrete

Buildings Reinforced with Superelastic Shape Memory Alloy Rebar”. Journal of

Engineering Structures, 34, 8-20.

American Society for Testing Materials. 2016. ASTM E8/E8M-16a: Standard Test Methods for

Tension Testing of Metallic Materials. Prepared by ASTM International, West

Conshohocken, PA, USA.

American Society for Testing Materials. 2018. ASTM F2516-18: Standard Test Method for

Tension Testing of Nickel-Titanium Superelastic Materials. Prepared by ASTM

International, West Conshohocken, PA, USA.

Standard Test Methods and Definitions for Mechanical Testing of Steel Products

American Society for Testing Materials. 2018. ASTM A370: Standard Test Methods and

Definitions for Mechanical Testing of Steel Products. Prepared by ASTM International,

West Conshohocken, PA, USA.

 111

Andrawes, B., (2010). “Active Confinement of Bridge Piers Using Shape Memory Alloys”.

Innovative Deserving Exploratory Analysis Programs; 16 pp.

Araki, Y., Shrestha, K.C., Maekawa, N., Koetaka, Y., Omori, T., Kainuma, R., (2016). “Shaking

Table Tests of Steel Frame with Superelastic Cu-Al-Mn SMA Tension Braces”. Journal of

Earthquake Engineering Structures Dynamics, 45(2):297–314.

Asgarian, B., Moradi, S., (2011). “Seismic Response of Steel Braced Frames with Shape Memory

Alloy Braces”. Journal of Construction Steel Res?????. 67:65–74.

Azuhata, T., Ishihara, T., Midorikawa, M., and Wada, A., (2006). “Seismic Response of Steel

Frames with Multi-Spans by Applying Rocking Structural System.” Proceedings of the 5th

International Conference on Behaviour of Steel Structures in Seismic Areas STESSA 2006,

Taylor & Francis Group, London, U.K.

Barthes, C.B., (2012). “Design of Earthquake Resistant Bridges Using Rocking Columns”.

ProQuest Dissertations Publishing, 2012, 3526503. University of California, Berkeley.

Billah, A., Alam, S., (2015). “Seismic Fragility Assessment of Concrete Bridge Pier Reinforced

with Superelastic Shape Memory Alloy”. Journal of Earthquake Spectra, 31 (3), pp. 1515-

1541.

Billah, A., Alam, S., (2012). “Seismic Performance of Concrete Columns Reinforced with Hybrid

Shape Memory Alloy (SMA) and Fiber Reinforced Polymer (FRP) Bars”. Journal of

Construction and Building Materials, 28.1, pp. 730-742.

Bradley, B. A., Dhakal, R. P., Mander, J. B., and Li, L.,(2008). "Experimental Multi-Level Seismic

Performance Assessment of 3D RC Frame Designed for Damage Avoidance". Journal of

Earthquake Engineering & Structural Dynamics, 37(1), 1–20.

https://doi.org/10.1002/eqe.741
 112

Canadian Standard Association – CSA (2014)., CAN/CSA-S16-14: Handbook of steel

construction. Prepared by the CSA, Toronto, ON.

Cardon, D., Dolce, M., Ponzo, F.C., and Colelho, E., (2004). “Experimental Behaviour of R/C

Frames Retrofitted with Dissipating and Re-Centring Braces”. Journal of Earthquake

Engineering, 08(03), pp. 361-396.

Cardone, D., Dolce, M., and Palermo, G., (2008). "Force-Based vs. Direct Displacement-Based

Design of Buildings with Seismic Isolation". Proceedings of the 14th World Conference

on Earthquake Engineering, Beiging, China, October 12-17, 2008. 8p.

Casciati, S., Faravelli, L. (2008). “Structural components in shape memory alloy for localized

energy dissipation”. Comput Struct 86(3):330–339.

Casciati, S., Faravelli, L. (2009). “A passive control device with SMA components: from the

prototype to the model”. Struct Control Hlth 16(7–8):751–765.

Chen, Y. and Schuh, C. A. (2011). “Size effects in shape memory alloy microwires,” Acta

Materialia, 59(2), 537-553.

Choi, E., Nam, T.H., Cho, S.C., Chung, Y.S., and Park, T. (2008). “The Behavior of Concrete

Cylinders Confined by Shape Memory Alloy Wires”. Smart Materials and Structures;

Vol.17: 10pp.

Chopra, A.K. 2012. “Dynamics of Structures Theory and Applications to Earthquake

Engineering”, 4th ed., Prentice Hall, New Jersey, NY, USA.

Chou, C. C., and Chen, Y. C. (2005). “Cyclic tests of post-tensioned precast CFT segmental bridge

columns with unbonded strands.” Earthquake Eng. Struct. Dyn.,352, 159–175.

Christian, J.W. (2001). “Twinning: Crystallography”, Encyclopedia of Materials: Science and

Technology, pp. 9423-9425.

 113

Cimellaro GP, Reinhorn AM, Bruneau M. (2010). “Seismic resilience of a hospital system”. Struct.

Infrastruct. Eng. 6:127–144.

Cimellaro, G., Reinhorn, A. M., & Filiatrault, A. (2007). "Improving Seismic Resilience of

Structural Systems through Integrated Design of Smart Structures". PhD Thesis. University

of Buffalo, Buffalo New York.

Clough, R.W., and Huckelbridge, A. A. (1977). “Preliminary experimental study of seismic uplift

of a steel frame.” Rep. No. UCB/EERC-77-22, Earthquake Engineering Research Centre

(EERC), Univ. of California, Berkeley, CA.

Cortés-Puentes, W.L., Palermo, D. (2018). “Seismic Retrofit of Concrete Shear Walls with SMA

Tension Braces”. J. Struct. 144(2), 4017200-1:13.

DesRoches, R., McCormick, J., and Delemont, M. (2004). “Cyclic properties of superelastic shape

memory alloy wires and bars,” Journal of Structural Engineering, 130(1), 38-46.

Dolce, M., Cardone, D. (2001). “Mechanical behaviour of shape memory alloys for seismic

applications. Martensite and austenite NiTi bars subject to torsion”. International Journal

of Mechanical Sciences. Vol 43, 2631-2656.

Dolce, M., Cardone, D., Marnetto, R. (2000). “Implementation and testing of passive control

devices based on shape memory alloys”. Earthquake Engineering and Structural Dynamics.

29, 945-968.

Dolce, M., Cardone, D., Ponzo, F.C. (2007). “Shaking-table tests on reinforced concrete frames

with different isolation systems”. Earthquake Engineering Structures 36(5):573–596.

Dong, Z., Klotz, U.E., Leinenbach, C., Bergamini, A., Czaderski, C., and Motavalli, M. (2009).

“A Novel Fe-Mn-Si Shape Memory Alloy with Improved Shape Recovery Properties by

VC Precipitation”. Advanced Engineering Materials, 11(1-2), 40-44.

 114

Eatherton, M.R., et al., (2014). “Design Concepts for Controlled Rocking of Self-Centring Steel-

Braced Frames,” J. Struct. Eng., vol. 140, no. 11, p. 4014082.

Elnashai, A. S. and Di Sarno, L. (2008). “Fundamentals of earthquake engineering,” John Wiley

& Sons Ltd, New York, 347 pp.

El-Tawil, S., and Kuenzli, C. (2002). “Pushover of hybrid coupled walls II: Analysis and

behavior.” J. Struct. Eng., 128.10., 1282–1289.

Fang, C., Yam, M.C.H., Lam, A.C.C., Xie, L.K. (2014). “Cyclic performance of extended end-

plate connections equipped with shape memory alloy bolts”. J Constr Steel Res 94:122–

136.

FEMA. (2009). “Effects of Strength and Stiffness Degradation on Seismic Response,” FEAM

P440A. Prepared by Applied Technology Council, Redwood City, California, USA.

Figueiredo, A. M., Modenesi, P., and Buono, V. (2009). “Low-cycle fatigue life of superelastic

NiTi wires,” International Journal of Fatigue, 31(4), 751-758.

Furuichi, Y., Tobushi, H., Ikawa, T., and Matsui, R. (2003). “Fatigue properties of a TiNi shape-

memory alloy wire subjected to bending with various strain ratios,” Proceedings of the

Institution of Mechanical Engineers, Part L: Journal of Materials Design and Applications,

217(2), 93-99.

Gajan, S., Kutter, B. L., Phalen, J., Hutchinson, T. C., and Martin, G. R. (2005). “Centrifuge

modelling of load-deformation behaviour of rocking shallow foundations.” Soil. Dyn.

Earthquake Eng., 25, 773–783

Günay, S., Korolyk, M., Mar, D., K.M. Mosalam, and Rodgers J. (2009). “Infill Walls as a Spine

to Enhance the Seismic Performance of Non-Ductile Reinforced Concrete Frames”.

Conference: ATC and SEI Conference on Improving the Seismic Performance of Existing
 115

Buildings and Other Structures. San Francisco, California USA. December 9-11, 2009.

1093-1104.

Haque, A.B.M., and Alam, M.S. (2017). “Hysteretic Behaviour of a Piston Based Self-centring (

PBSC ) Bracing System Made of Superelastic SMA Bars – A Feasibility Study”.

Structures, 12(August), 102–114.

Hewes, J. T., and Priestley, M. J. N. (2002). “Seismic design and performance of precast concrete

segmental bridge columns.” Structural Systems Research Project, Rep. No. SSRP-2001/25,

Univ. of California at San Diego

Hilsenrath, J. (1955). “Tables of thermal properties of gases; comprising tables of thermodynamic

and transport properties of air, argon, carbon dioxide, carbon monoxide, hydrogen,

nitrogen, oxygen, and steam,” U.S. Department of Commerce, National Bureau of

Standards, Washington, USA, 564 pp.

Holden, T., Restrepo, J., and Mander, J. B. (2003). “Seismic performance of precast reinforced

and prestressed concrete walls.” J. Struct. Eng.,1293, 286–296

Hosseini, F., Gencturk,B., Lahpour, S., and Gil, D.I. (2015). “An experimental investigation of

innovative bridge columns with engineered cementitious composites and Cu-Al-Mn super-

elastic alloys,” Smart Mater. Struct., vol. 24, no. 8. 16pp.

Humar, J.L. and Rahgozar, M.A. (1997). “Application of inelastic response spectra derived from

seismic hazard spectral ordinates for Canada,” Canadian Journal of Civil Engineering,

24(4):664-666.

Jaureguizahar, S.M., Chapetti, M.D., Yawny, A.A. (2016). “Fatigue of NiTi shape memory wires,”

Procedia Structural Integrity 2 (2016) 1427-1434.

 116

Johnson, R., Padgett, J.E., Maragakis, M.E., Desroches, R., Saiidi, M.S. (2008). “Large scale

testing of nitinol shape memory alloy devices for retrofitting of bridges”. Smart Mater.

Struct. 17. 10pp.

Karbhari, V.M., and Eckel II, D.A. (1994). “Effect of Cold Regions Climate on Composite

Jacketed Concrete Columns”. Journal of Cold Engineering, ASCE, Vol. 8, No. 3, pp.73-

86.

Kumar, P.K., Lagoudas, D.C. (2008). “Introduction to Shape Memory Alloys”. Modelling and

Engineering Applications. Springer Science+Business Media Deutschland GmbH,

Heidelberger Platz 3, Berlin, 14197, Germany. 1, pg. 1-51.

Kurama, Y. C., and Shen, Q. (2004). “Posttensioned hybrid coupled walls under lateral loads.” J.

Struct. Eng., 1302, 297–309.

Li, K., Dong, Z., Liu, Y., and Zhang, L. (2013). “A Newly Developed Fe-based Shape Memory

Alloy Suitable for Smart Civil Engineering”. Smart Materials and Structures, 22(4), pp.1-

7.

Lin, P. H., Tobushi, H., Tanaka, K., Hattori, T., and Makita, M. (1994). “Pseudoelastic behaviour

of TiNi shape memory alloy subjected to strain variations,” Journal of Intelligent Material

Systems and Structures, 5(5), 694-701.

Ma, H.W., Cho, C.D. (2008). “Feasibility study on a superelastic SMA damper with re-centring

capability”. Material Science Engineering A 473(1–2):290–296.

Maletta, C., Sgambitterra, E., Furgiuele, F., Casati, R., and Tuissi, A. (2012). “Fatigue of

pseudoelastic NiTi within the stress-induced transformation regime: A modified Coffin-

Manson approach,” Smart Materials and Structures, 21(11), 7 pp.

 117

Mander, J. B., and Cheng, C.-T. (1997). “Seismic resistance of bridge pier based on damage

avoidance design.” Technical Rep. No. NCEER-97-0014, Univ. at Buffalo—State Univ.

of New York, Buffalo, N.Y.

Mander, J. B., and Cheng, K. C. (1997). "Seismic resistance of bridge piers based on damage

avoidance design". Technical Report NCEER-97-0014. Buffalo.

Matsui, R., Makino, Y., Tobushi, H., Furuichi, Y., and Yoshida, F. (2006). “Influence of strain

ratio on bending fatigue life and fatigue crack growth in TiNi shape-memory alloy thin

wires,” Materials Transactions, 47(3), 759-765.

Mccormick, J., Barbero, L., Desroches, R., Mccormick, J., Barbero, L., & Desroches, R. (2005).

“Effect of mechanical training on the properties of superelastic shape memory alloys for

seismic applications”. In Proceedings of SPIE. San Diego California, USA, pp. 430-439.

Memry, SEAS Group Company, https://www.memry.com/nitinol-glossary, retrieved on 31 July

2018.

Miyazaki, S., Imai, T., IgoK, Y., & Otsuka, K. (1986). “Effect of cyclic deformation on the

pseudoelasticity characteristics of Ti-Ni alloys”. Metallurgical Transactions, 17(1), 115–

120.

Mohee, F. M., and Al-mayah, A. (2018). “Effect of barrel, wedge material and thickness on

composite plate anchor performance through analytical, finite element, experimental and

3D prototype investigations”. Engineering Structures, 175, 138-154.

Morabito, P. (1989). “Measurement of thermal properties of different concretes,” High

Temperatures. High Pressures, 21(1), 51-59.

Motaref, S., Saiidi, M.S., Sanders, D. (2014). “Shake Table Studies of Energy-Dissipating

Segmental Bridge Columns”. Journal of Bridge Engineering. 19:186–199.

 118

Moumni, Z., Herpen, A. V., and Riberty, P. (2005). “Fatigue analysis of shape memory alloys:

Energy approach,” Smart Materials and Structures, 14(5), S287-S292.

National Building Code of Canada – NBCC (2015). Canadian Commission on Building and Fire

Codes., National Research Council of Canada. Prepared by the Associate Committee on

the National Building Code, Ottawa, ON.

Nesheli, K.N., and Meguro, K. (2006). “Seismic Retrofitting of Earthquake-Damaged Concrete

Columns by Lateral Pre-Tensioning of FRP Belts”. Proceedings of the 8th US National

Conference on Earthquake Engineering, San Francisco, California, USA, Paper no. 841.

Ocel, J., DesRoches, R., Leon, R., Hess, W., Krumme, R., Hayes, J., and Sweeney, S. (2004).

“Steel Beam-Column Connections Using Shape Memory Alloys”. Journal of Structural

Engineering, 130(5), 732-740.

Ou, Y. C., Chiewanichakorn, M., Aref, A. J., and Lee, G. C. (2007).“Seismic performance of

segmental precast unbonded post-tensioned concrete bridge columns.” J. Struct. Eng.,

13311, 1636–1647.

Oudah, F. (2014). “Development of Innovative Self-Centring Concrete Beam-Column

Connections Reinforced using Shape Memory Alloys”. PhD Thesis. University of Calgary,

The Vault: http://hdl.handle.net/11023/2067.

Oudah, F. and El-Hacha, R. (2014). “Ductility of FRP reinforced RC structures: A critical review

of definition and expressions,” Proceedings of Concrete Solutions, Belfast, UK, September

1-3, 329-335.

Oudah, F. and El-Hacha, R. (2016). “Plastic hinge relocation in concrete structures using the

double-slotted-beam system”. Bulletin of Earthquake Engineering, 15 no(5), pp. 2173-

2199. 10.1007/s10518-016-0055-9
 119

Oudah, F. and El-Hacha, R. (2017). “Joint performance in concrete beam-column connections

reinforced using SMA smart material”. Engineering Structures, 151 pp. 745-760.

10.1016/j.engstruct.2017.08.054

Ozbulut, O.E., Hurlebaus, S., Desroches, R. (2011). “Seismic Response Control Using Shape

Memory Alloys: A Review”. Journal of Intelligent Material Systems and Structures. Vol

13, 1531-1549.

Ozbulut, O.E., Silwal, B. (2016). “Performance assessment of buildings isolated with S-FBI

system under near fault earthquakes”. Smart Struct Syst 17(5):709–724.

Ozkaynaka, H., Yuksela, E., Buyukozturk, O., Yalcinc, C., and Dindar, A.A. (2011). “Quasi-static

and pseudo-dynamic testing of infilled RC frames retrofitted with CFRP material”.

Composites Part B: Engineering, 42 (2), pp. 238-263.

Park, J., Choi, E., Park, K., and Kim, HT. (2011). “Comparing the Cyclic Behaviour of Concrete

Cylinders Confined by Shape Memory Alloy Wire of Steel Jackets”. Smart Materials and

Structures; 20: 11pp.

Park, R., Paulay, T. (1975). "Reinforced Concrete Structures". ed. New York: John Wiley & Sons,

p. 769.

Paulay, T. and Priestley, M. J. N. (1992). “Seismic design of reinforced concrete and masonry

buildings,” John Wiley & Sons Ltd, New York, USA, 744 pp.

Pieczyska, E., Gadaj, S., Nowacki, W. K., Hoshio, K, Makino, Y., and Tobushi, H. (2005).

“Characteristics of energy storage and dissipation in TiNi shape memory alloy,” Science

and Technology of Advanced Materials, 6(8), 889-894.

Pollino, M., Bruneau, M. (2007). “Experimental study if the controlled rocking response of steel

braced frames”. ASCE Structural Engineering Research Frontiers.

 120

Priestley, M. J. N. (2000). “Performance based seismic design”. 12th WCEE, 1(1), 1–22.

Auckland, New Zealand. January 30, 2000.https://doi.org/10.1016/j.engstruct.2003.08.003

Priestley, M. J. N., Seible, F., and Calvi, G. M. (1996). “Seismic Design and Retrofit of Bridges”.

John Wiley & Sons.

Priestley, M., Calvi, G. M., and Kowalsky, M. J. (2007). “Displacement-based seismic design of

structures”. Building, 23(33), 1453–1460. https://doi.org/10.1016/S0141-0296(01)00048-

Priestley, M.J.N. (2000). “Performance based seismic design”. 12th WCEE, vol. 1, no. 1, p. 22.

Priestley, M.J.N., Calvi, G.M., and Kowalsky, M.J. (2007). “Direct displacement-based seismic

design”. New Zealand Society for Earthquake Engineering Conference, New Zealand, vol.

23, no. 33, pp. 1453-1460.

Qian, H., Li, H.N., Song, G.B. (2016). “Experimental investigations of building structure with a

superelastic shape memory alloy friction damper subject to seismic loads”. Smart Mater

Struct 25(12).

Qiu, C.X., Zhu, S.Y. (2017). “Shake table test and numerical study of self-centring steel frame

with SMA braces”. Earthq Eng Struct D 46(1):117–137.

Roh, H. (2007). “Seismic Behavior of Structures using Rocking Columns and Viscous Dampers”.

ProQuest Dissertations Publishing, 2007, 3277759. State University of New York at

Buffalo.

Roh, H. and Reinhorn, A. M. (2010). ”Modelling and seismic response of structures with concrete

rocking columns and viscous dampers”. Engineering Structures, 32 no(8), pp. 2096-2107.

Roh, H., and Reinhorn, A.M., (2010). “Hysteretic behavior of precast segmental bridge piers with

superelastic shape memory alloy bars”. Journal of Engineering Structures, 32, 3394-3403.
 121

Rojob, H., El-Hacha, R., (2015). “Ductility behavior of RC beams strengthened in flexure with

NSM Iron-based Shape Memory Alloy bars”. 8.

Sause, R., Ricles, J. M., Roke, D. A., Chancellor, N. B., and Gonner, N. P. (2010). “Seismic

performance of a self-centring-rocking concentrically-braced frame.” Proc., 9th U.S.

National and 10th Canadian Conf. on Earthquake Engineering, Mira Digital Publishing, St.

Louis, MO.

Sherif, M. M., & Ozbulut, O. E. (2018). “Tensile and superelastic fatigue characterization of

Tensile and superelastic fatigue characterization of NiTi shape memory cables”. Smart

Materials and Structures, 27(15007), 13pp.

Shin, M., and Andrawes, B. (2010). “Experimental Investigation of Actively Confined Concrete

Using Shape Memory Alloys”. Engineering Structures; 32: 656-664.

Shin, M., and Andrawes, B. (2011). “Lateral Cyclic Behavior of Reinforced Concrete Columns

Retrofitted with Shape Memory Spirals and FRP Wraps”. J. Struct. Eng. 137:1282–1290.

Shook, D.A., Roschke, P.N., Ozbulut, O.E. (2008). “Superelastic semi-active damping of a base-

isolated structure”. Struct Control Hlth 15(5):746–768.

Shrestha, B., Hao, H. (2016). “Parametric study of seismic performance of super-elastic shape

memory alloy-reinforced bridge piers”. Struct. Infrastruct. Eng. 12:1076-1089.P

Sideris, P., Aref, A.J., Filiatrault, A. (2014). “Large-Scale Seismic Testing of a Hybrid Sliding-

Rocking Posttensioned Segmental Bridge System”. J. Struct. Eng. 140:4014025.

Soroushian, P., Ostowari, K., Nossoni, A., Chowdhury H. (2001). “Repair and Strengthening of

Concrete Structures through Application of Corrective Posttensioning Forces with Shape

Memory Alloys”. Journal of the Transportation Research Board, No. 1770, pp.20-26.
 122

Soul, H., Isalgue, A., Yawny, A., Torra, V., and Lovey, F. C. (2010). “Pseudoelastic fatigue of

NiTi wires: Frequency and size effects on damping capacity,” Smart Materials and

Structures, 19(8), 7pp.

Speicher, M.S., DesRoches, R., and Leon, R.T. (2011). “Experimental results of a NiTi shape

memory alloy (SMA)-based recentring beam-column connection”. Eng Struct 33(9):

2448–2457.

Stanton, J., and Nakaki, S. (2002). “Design guideline for precast concrete seismic structural

systems.” PRESSS Rep. Nos. 01/03-09 and UW SM02-02, Univ. of Washington, Seattle,

Wash

Sultana, P., and Youssef, M.A. (2018). “Seismic performance of modular steel frames equipped

with Shape Memory Alloy braces,” Proceedings, Annu. Conf. - Can. Soc. Civ. Eng., vol.

3.

Symans, M. D., Charney, F. A., Whittaker, A. S., Constantinou, M. C., Kircher, C. A., Johnson,

M. W., and Mcnamara, R. J. (2008). "Energy Dissipation Systems for Seismic

Applications: Current Practice and Recent Developments". Journal of Structural

Engineering, 134(1), 3–21. https://doi.org/10.1061/͑ASCE͒0733-9445͑2008͒134:1͑3͒

Tabanli, R. M., Simha, N. K., and Berg, B. T. (1999). “Mean stress effects on fatigue of NiTi,”

Materials Science and Engineering: A, 273-275(15), 644-648.

Taha, M. R., and Shrive, N. G. (2002). “New Concrete Anchors for Carbon Fiber-Reinforced

Polymer Post-Tensioning Tendons — Part 1 : State-of-the-Art Review / Design”. ACI

Structural Journal, 100(1), 86–95.

 123

Tam, S.S.F. (2007). “Durability of Fibre Reinforced Polymer (FRP) and FRP Bond Subjected to

Freeze-Thaw Cycles and Sustained Load”. MSc Thesis, University of Toronto, Ontario,

Canada.

Tamai, H. and Kitagawa, Y. (2002). “Pseudoelastic behaviour of shape memory alloy wire and its

application to seismic resistance member for building,” Computational Materials Science,

25(1-2), 218-227.

Tazarv, M., and Saiidi, M.S. (2015). “Reinforcing NiTi Superelastic SMA for Concrete

Structures,” J. Struct Eng., vol. 141, no. 8, pp. 1–10.

TE connectivity, www.te.com, Retrieved September 2018.

Tobushi, H., Nakahara, T. Shimeno, Y., and Hashimoto, T. (2000). “Low-cycle fatigue of TiNi

shape memory alloy and formulation of fatigue life,” Journal of Engineering Materials and

Technology, 122(2), 186-191.

Tobushi, H., Shimeno, Y., Hachisuka, T., and Tanaka, K. (1998). “Influence of strain rate on

superelastic properties of TiNi shape memory alloy,” Mechanics of Materials, 30(2),

141150.

Tobushi, H., Takata, K., Shimeno, Y., Nowacki, W. K., and Gadaj, S. P. (1999). “Influence of

strain rate on superelastic behaviour of TiNi shape memory alloy,” Proceedings of the

Institution of Mechanical Engineers, Part L: Journal of Materials Design and Applications,

213(2), 93-102.

Tremblay, R., Poirier, L., Bouaanani, N., Leclerc, M., Rene, V., Fronteddu, L., Rivest, S. (2008).

“Innovative Viscously Damped Rocking Braced Steel Frames”. 14th World Conf. Earthq.

Eng. (14 WCEE), Oct. 12-17.

 124

Trindade, M. A., Soong, T. T., and Costantinou, M. C. (2014). "Passive and active structural

vibration control in civil engineering". Dynamics of Smart Systems and Structures,

345(March), 65–92. https://doi.org/10.1007/978-3-319-29982-2_4

Varela, S., and Saiidi, M. S. (2016). “A bridge column with superelastic NiTi SMA and replaceable

rubber hinge for earthquake damage mitigation”. Smart Materials and Structures,

25(075012).18pp.

Viti, S.; Cimellaro, G.P.; Reinhorn, A.M. (2006). “Retrofit of a hospital through strength reduction

and enhanced damping,” Smart Struct. Syst., vol. 2, no. 4, pp. 339–355.

Wada, A., Yamada, S., Fukuta, O., and Tanigawa, M. (2001). “Passive controlled slender

structures having special devises at column connections.” 7th Int. Seminar on Seismic

Isolation, Passive Energy Dissipation and Active Control of Vibrations of Structures,

European Association of Earthquake Engineering, Istanbul, Turkey.

Wang, W., Fang, C., Liu, J. (2017) “Self-centring beam-to-column connections with combined

superelastic SMA bolts and steel angles”. J. Struct Eng, 143(2).

Wiebe L, Christopoulos C. (2015). “Performance-Based Seismic Design of Controlled Rocking

Steel Braced Frames. I: Methodological Framework and Design of Base Rocking Joint”. J.

Struct. Eng. 141:4014226.

Wiebe, L.; Christopoulos, C. (2009). “Mitigation of Higher Mode Effects in Base-Rocking

Systems by Using Multiple Rocking Sections”. Journal of Earthquake Engineering, 13 (1),

83-108.

Zhou, P., Liu, M., Li, H., and Song, G. (2018). “Experimental investigations on seismic control of

cable ‐ stayed bridges using shape memory alloy self ‐ centring dampers”. Structural

Control Health Monitoring, 25(7), pp. 1–14.

 106
125

Appendix A
A1. Introduction
Appendix A contains the design of the CBF and SMA-BF systems. The order of the calculations

are as follows:

A1.1. Design of cross beam and columns based on axial load and an accidental eccentricity as

per CSA S16.

A1.2. Gusset plate design

A1.3. Connection design for the column to foundation beam and cross beam to column.

A1.4. Design of braces for the CBF and SMA-BF systems.

A1.5. Design of yielding fuse for the CBF and SMA-BF systems.

A1.6. Connection design of the braces to gusset plates.

A1.7. Verification that foundation beam and connection to lab floor is able to undergo design

loads.

A1.8. Verification that the loading beam could undergo design loads without deforming under

applied loads.

A1.9. Required applied loads for the vertical actuators to induce 65kN into columns.
A1.1 Design of Columns based on axial load and accidental eccentricity

BEAM-COLUMN DESIGN CRITERIA FROM CSA S16-14

Criterion Class 1 Section Class 2 Section
b/t Limit 𝟏𝟒𝟓 𝟏𝟕𝟎
≤ ≤
√ 𝑭𝒚 √ 𝑭𝒚
C=0 𝟏𝟏𝟎𝟎 𝟏𝟕𝟎𝟎
Local ≤ ≤
√ 𝑭𝒚 √ 𝑭𝒚
Buckling 𝟔𝟕𝟎 𝟔𝟕𝟎
M=0
(Clause 11.3) h/t Limit ≤ ≤
√ 𝑭𝒚 √ 𝑭𝒚
C≠0 𝟏𝟏𝟎𝟎 𝑪𝒇 𝟏𝟕𝟎𝟎 𝑪𝒇
M≠0 ≤ (𝟏 − 𝟎. 𝟑𝟗 ) ≤ (𝟏 − 𝟎. 𝟔𝟏 )
√𝑭𝒚 𝝋𝑪𝒚 √𝑭𝒚 𝝋𝑪𝒚
Cross-Sectional Strength (Clause 𝑪𝒇 𝟎. 𝟖𝟓𝑼𝟏𝒙 𝑴𝒇𝒙 𝜷𝑼𝟏𝒚 𝑴𝒇𝒚
13.8.2.a and 13.8.3.a + + ≤ 𝟏, 𝜷 = 𝟎. 𝟔
𝑪𝒓 𝑴𝒓𝒙 𝑴𝒓𝒚

𝑪𝒓 = 𝝋𝑨𝒈 𝑭𝒚
𝑴𝒓𝒙 = 𝝋𝒁𝒙 𝑭𝒚
𝑴𝒓𝒚 = 𝝋𝒁𝒚 𝑭𝒚

𝒘𝟏𝒙 𝝅𝟐 𝑬𝑰𝒙
𝑼𝟏𝒙 = ≥ 𝟏, 𝑪𝒆𝒙 =
𝑪𝒇 𝑳𝟐𝒙
𝟏− 𝑪
𝒆𝒙
𝒘𝟏𝒚 𝝅𝟐 𝑬𝑰𝒚
𝑼𝟏𝒚 = ≥ 𝟏, 𝑪𝒆𝒚 =
𝑪𝒇 𝑳𝟐𝒚
𝟏− 𝑪
𝒆𝒚

Overall (In-Plane) Member Strength 𝑪𝒇 𝟎. 𝟖𝟓𝑼𝟏𝒙 𝑴𝒇𝒙 𝜷𝑼𝟏𝒚 𝑴𝒇𝒚

(Clause 13.8.2.b and 13.8.3.b) + + ≤ 𝟏, 𝜷 = 𝟎. 𝟔 + 𝟎. 𝟒𝝀𝒚
𝑪𝒓 𝑴𝒓𝒙 𝑴𝒓𝒚
≤ 𝟎. 𝟖𝟓

𝑪𝒓 = 𝝋𝑨𝒈 𝑭𝒚 (𝟏 + 𝝀𝟐𝒏 )−𝟏/𝒏

𝑴𝒓𝒙 = 𝝋𝒁𝒙 𝑭𝒚
𝑴𝒓𝒚 = 𝝋𝒁𝒚 𝑭𝒚

𝒘𝟏𝒙 𝝅𝟐 𝑬𝑰𝒙
𝑼𝟏𝒙 = ≥ 𝟏, 𝑪𝒆𝒙 =
𝑪𝒇 𝑳𝟐𝒙
𝟏− 𝑪
𝒆𝒙
𝒘𝟏𝒚 𝝅𝟐 𝑬𝑰𝒚
𝑼𝟏𝒚 = ≥ 𝟏, 𝑪𝒆𝒚 =
𝑪𝒇 𝑳𝟐𝒚
𝟏− 𝑪
𝒆𝒚
 Lateral Torsional Buckling (Out-of- 𝑪𝒇 𝟎. 𝟖𝟓𝑼𝟏𝒙 𝑴𝒇𝒙 𝜷𝑼𝟏𝒚 𝑴𝒇𝒚
plane) Strength + + ≤ 𝟏, 𝜷 = 𝟎. 𝟔 + 𝟎. 𝟒𝝀𝒚
𝑪𝒓 𝑴𝒓𝒙 𝑴𝒓𝒚
(Clause 13.8.2.c and 13.8.3.c) ≤ 𝟎. 𝟖𝟓

𝑴𝒓𝒚 = 𝝋𝒁𝒚 𝑭𝒚
𝑴𝒓𝒙 𝒇𝒓𝒐𝒎 𝑪𝒍𝒂𝒖𝒔𝒆 𝟏𝟑. 𝟔

𝑼𝟏𝒙 = 𝟏
𝑼𝟏𝒚 = 𝟏

The columns and beam were designed using the above equations. The checks are following in
the excel spreadsheets.
Section Applied Loads
W150x30
A 3790 mm2 Mfx 0 kNm
Ix 17100000 mm4 Mfy 0 kNm
Sx 218000 mm3 Cf 250 kN
rx 67.3 mm
Zx 244000 mm3
Iy 5560000 mm4
Sy 72600 mm3
ry 38.3 mm
Zy 111000 mm3
J 100000 mm4
Cw 30300000000 mm6
d 157 mm
b 153 mm
t 9.3 mm
w 6.6 mm
G 77000 Mpa
E 200000 Mpa
fy 350 Mpa
k 1
L 2000 mm
Le 2000 mm
n 1.34
ω1x 0.85
ω1y 0.6
φ 0.9
Cy 1326.5 kN
κ 0
ω2 1.75
 1) Local Buckling Class Section Class 2) Cross sectional Strength
b/t 8.225806452 2 2 Cr 1193.85 kN
Look below to
determine
h/w 20.96969697 1 secton Mrx 76.86 kNm
Mry 34.965 kNm
Class 1 7.750576015 Cex 8438.512 kN
b/t limit Class 2 9.086882225 Cey 2743.75
Class 3 10.69044968 U1x 1
U1y 1
Class 1 Class 2 Class 3 β 0.6
C=0 58.79747322 90.86882225 101.5592719
h/w limit
M=0 35.81300642 35.81300642 35.81300642 Check Interaction EQ
C>0 & M>0 53.99556876 79.26142149 87.73560757 0.209407
OK Good!
3) Overall Strength
Crx 1124.600051 kN
λx 0.395716431 4) Lateral Torsional Buckling
Cry 939.9683481 kN Cr 939.9683 kN
λy 0.69534506 Mry 34.965 kNm
Cr 939.9683481 kN U1x 1
U1y 1
Mrx 76.86 β 0.6
Mry 34.965 A 8.56E+21
Cex 8438.511763 kN B 1.66E+22
Cey 2743.750024 Mu 436.2826
U1x 0.875950992 Mp or My 85.4
U1y 0.660150375 (2/3) Mp or My
56.93333
β 0.6
Mrx 83.54453 kNm
Check interaction EQ
0.265966 Check interaction EQ
OK Good! 0.265966
OK Good!
Section Applied Loads
W250x22
A 2850 mm2 Mfx 0 kNm
Ix 28900000 mm4 Mfy 0 kNm
Sx 227000 mm3 Cf 160 kN
rx 101 mm
Zx 263000 mm3
Iy 1230000 mm4
Sy 24000 mm3
ry 20.7 mm
Zy 38100 mm3
J 43400 mm4
Cw 18700000000 mm6
d 254 mm
b 102 mm
t 6.9 mm
w 5.8 mm
G 77000 Mpa
E 200000 Mpa
fy 350 Mpa
k 1
L 2000 mm
Le 2000 mm
n 1.34
ω1x 0.85
ω1y 0.6
φ 0.9
Cy 997.5 kN
κ 0
ω2 1.75
 1) Local Buckling Class Section Class 2) Cross sectional Strength
b/t 7.391304 1 1 Cr 897.75 kN
Look below
to determine
h/w 41.41379 1 secton Mrx 82.845 kNm
Mry 12.0015 kNm
Class 1 7.750576 Cex 14261.58 kN
b/t limit Class 2 9.086882 Cey 606.9807
Class 3 10.69045 U1x 1
U1y 1
Class 1 Class 2 Class 3 β 0.6
C=0 58.79747 90.86882225 101.5592719
h/w limit
M=0 35.81301 35.81300642 35.81300642 Check Interaction EQ
C>0 & M>0 54.71063 80.98990601 89.79412096 0.178223
OK Good!
3) Overall Strength 4) Lateral Torsional Buckling
Crx 879.3834 kN Cr 398.9738 kN
λx 0.26368 Mry 12.0015 kNm
Cry 398.9738 kN U1x 1
λy 1.286556 U1y 1
Cr 398.9738 kN β 0.6
A 8.22E+20
Mrx 82.845 B 2.27E+21
Mry 12.0015 Mu 152.859
Cex 14261.58 kN Mp or My 92.05
Cey 606.9807 (2/3) Mp or My
61.36667
U1x 0.859644
U1y 0.814774 Mrx 79.2077 kNm
β 0.6
Check interaction EQ
Check interaction EQ 0.401029
0.401029 OK Good! OK Good!
A1.2 Gusset Plate design
A1.3 Connection design for the column to foundation beam and cross beam to column
A1.4 Design of braces for the CBF and SMA-BF systems
A1.5 Design of yielding fuse for the CBF and SMA-BF systems
A1.6 Connection design of braces to gusset plates
A1.7 Foundation beam to Lab floor connection
A1.8 Verification of loading beam to undergo design loads without deforming
A1.9 Required applied loads for vertical actuators to induce 65kN into columns
 158

Appendix B - FE Modelling Summary

B.1 Objective
The objective of the finite element (FE) modelling was to simulate the behaviour of the braced

frame and conduct a parametric study on the shape memory alloy braced frame (SMA-BF) and

compare it to the control braced frame (CBF). However, due to time constraints and unforeseen

circumstances the model was not able to be completed and will become future work.

B.2 Material Properties (Steel Coupon Test Simulations)

B.2.1. Experiment

The steel plate used as the steel fuse in the CBF system was tested in accordance with ASTM

E8/E8M. The specimen, the fracture pattern and the stress-strain behaviour are shown below,

Figures B.1 to B.5.

Figure B.1: Coupon test specimen details

Figure B.2: Coupon test fracture pattern

 159

Figure B.3: Tensile stress-strain behaviour of steel coupon specimens

Two different methods were used to measure the strain value of the specimens. On coupon test 1

a strain gauge was used and failed prematurely at around 5% strain. For the second specimen a

linear strain converter (LSC) was used to measure the strain. The steel properties can be found in

Chapter 4 in Table 4-2.

B.2.2. Finite Element Model of the coupon test with solid elements

In order to ensure that the larger braced frame model had the correct material properties inputted

into the steel fuses, a model of the coupon test was done. Figure B.4 shows the ABAQUS model

of the coupon test.

• Solid elements were used, more specifically C3D20R. A 20-node quadratic brick. For this

model reduced integration was used to reduce run time and size of file.

• An encastre boundary condition was used on the bottom end of the model and pinned at

the top, allowing for movement in the direction of the tension load (y-direction).
 160

• For the material input properties, the engineering stress collected from the experimental

tension test was put into the built in ABAQUS calibration tool to determine the true stress-

strain behaviour.

• Figure B.5 shows the results of the model compared with the experimental stress-strain

results.

• From this model it can be shown that the correct input material properties was determined

as the FE results are identical to the experimental results.

• A mesh sensitivity test should be done on this model to validate the results.

Figure B.4: FE model of steel coupon test

 161

Figure B.5: Comparison of experimental and FE model stress-strain behaviour

B.3 Finite Element Model of the symmetric coupon test with Shell elements

The objective of this FE model was to understand the behaviour of the steel fuses during the

experimental cyclic testing and to model their behaviour correctly. The CBF system was designed

as a tension-only braced frame, therefore, one brace will be in tension and the other brace would

be in compression. During the experiment, the fuse buckled in the first mode and therefore, it was

important that the FE model reflected that behaviour. When modeling the fuses as shown in Figure

B.6, an initial imperfection had to be implemented in order to force the fuse to buckle by initiating

a small out of plane displacement. Table B.1 shows the different hysteresis response of the model

with varying imperfection values.

 162

a) Model b) Mesh c) Boundary Conditions d) Deformed Shape

Figure B.6: Symmetric coupon test
 163

Table B.1: Behaviour of symmetric coupon test with Initial Imperfection

Imperfection Value Behaviour

0.3

0.6

1.5

• A mesh sensitivity test should be done on this model to validate the results.
 164

B.4 Finite Element Models of CBF

B.4.1. CBF FE Model with solid elements and shell elements

A complete model of the experimental CBF frame was developed in ABAQUS as shown in

Figure B.7.

• A similar cyclic displacement-controlled loading regime (Figure 4-11) was applied to the

model as in the experiment; as well as the 65 kN vertical loads on the column.

• Boundary conditions were placed on the bolts to try to simulate a true pin connection.

However, it was found that it did not make much of a difference when the pins were placed

on the underside of the base plate or to the underside of the column; the behaviour was

quite similar. This could be because it is a pin connection and the frame is still rotating

about the same axis in all cases.

• All parts were connected together using tie constraints. Due to the out-of-plane movement

of the braces a contact interaction had to be created at the intersection of the cross braces

so that the contact forces would transfer correctly between elements.

• The columns and beams were composed of linear solid elements.

• The gusset plates were quadratic wedge elements (C3D15). Due to the complex geometry,

wedge elements were the only viable elements.

• The bolts and nuts were quadratic solid elements (C3D20R).

• The angle braces were linear solid elements (C3D8R).

• The steel fuses were shell elements (S4). Shell elements were used at the fuse location

because it was a plate. Solid elements are too stiff for this application.

• A mesh sensitivity test should be done on this model to validate the results.
 165

Figure B.8 shows the mesh configuration of the CBF.

Figure B.7: FE model of CBF

Figure B.8: Mesh of CBF

Figure B.9 shows the hysteresis response of the CBF. When compared to Figure 4-14, the response

is very different than the FE model. The frame was designed so that the plastic hinge developed at

the fuse location and therefore, the behaviour of the frame was dictated by the behaviour of the
 166

braces. In the following sections, are different models that were developed to try to model the

correct behaviour of the braces under cyclic load.

Figure B.9: Load-Displacement Hysteresis Behaviour of CBF

Figure B.10 shows the comparison of the experimental and FE hysteresis behaviour.

Figure B.10: Comparison between experimental and FE model hysteresis response of CBF

B.5 CBF FE model without full columns

In the previous section, the full frame CBF model was behaving stiffer than the experimental

model. One potential cause proposed by Dr. Haddad was that the frame was dominating the
 167

behaviour rather than the braces. He proposed to remove the full height columns, Figure B.11, and

only have a portion of them and investigate how the behaviour of the system changes.

Figure B.11: FE model of partial CBF system

Figure B.12: Mesh of partial CBF system

 168

Figure B.12 shows the mesh of the CBF model. The steel fuses were composed of continuum shell

elements (SC8) and the rest were the same as the full CBF model described in the previous section.

A mesh sensitivity test should be done to assess the validity of the results.

Figure B.13: Hysteresis behaviour of partial CBF system

Figure B.13 shows the behaviour of the CBF model without the full height columns. It is interesting

to note that without the full height columns, there is a severe pinching effect, suggesting that the

braces dictated the behaviour of the system. However, the behaviour was still incorrect as it would

not be expected for steel to behave in this manner.

B.6 CBF model only braces

The objective of this model was to examine the behaviour of the braces alone to ensure that it was

deforming correctly. Shell elements (S4R) were used for the steel fuse, while SC8 elements were

used for the braces and gusset plates. A fine mesh was applied to the fuse while a coarse mesh was

used for the other members to speed up the run-time. A mesh sensitivity test should be done to

assess the validity of the results.

 169

Figure B.14: Only braces modelled

Figure B.15: Mesh of just braces

Figure B.16 shows the deformed shape of the model at the end of the 40 mm cycle. One brace is

in tension and the other has buckled out of plane, similar to the experiment. Therefore, in this

model the brace achieved the objective of matching the experimental deformed shape when cycled.
 170

Figure B.16: Deformed shape of just braces model

Figure B.17 shows the force-displacement hysteresis response. From the graph it can be seen that

the steel yielded at a lower displacement than the experimental as well as it was slightly stiffer

than the experimental CBF system.

 171

Figure B.17: Force-Displacement result of just braces model

B.7 Summary

One of the objectives of this research was to create a finite element model that demonstrated the

behaviour of the PE SMA material in a braced frame. The first step was taken in this research

phase into determining how to model the simple behaviour of the braced frame. In this research

phase the modeling of the steel yielding fuse was examined in order to model the correct properties

and behaviour of the overall system. Then, a couple of full frame models were created to model

the behaviour of the full frame with the yielding fuse. It was found that when the braces and the

gusset plates were modelled as shell continuum elements and the fuse modelled as a shell element,

the tension-compression behaviour of the braces matched the experimental results. Further, the

shape of the hysteresis response followed a similar trend to the experimental. In this phase of the

research, a starting point for the FE model was created.

B.8 References
 172

American Society for Testing Materials. 2016. ASTM E8/E8M-16a: Standard test methods for

tension testing of metallic materials. Prepared by ASTM International, West Conshohocken, PA,

USA.

M. Haddad, personal communication, March to June 2019.

S. Adeeb, personal communication, July 2019.