Investigation of Braid Geometry for

Complex Braided Composite Structures

A thesis submitted to the University of Manchester for the degree of

Master of Philosophy in the Faculty of Science and Engineering

2018

Danjie Yang

SCHOOL OF MATERIALS

1
 Table of contents
1. Introduction....................................................................................................... 17

1.1. Introduction to the research topic ................................................................... 17

1.2. Research objective ............................................................................................ 19

1.3. Outline of thesis................................................................................................. 20

2. Literature Review ............................................................................................. 21

2.1. Introduction to braiding................................................................................... 21

2.2. Braid machine components .............................................................................. 22

2.3. Braid parameters .............................................................................................. 22

2.3.1. Braid angle ......................................................................................................... 23

2.3.2. Unit cell .............................................................................................................. 24

2.3.3. Lay length .......................................................................................................... 25

2.3.4. Cover factor ....................................................................................................... 26

2.3.5. Crimp ................................................................................................................. 26

2.4. Composite materials ......................................................................................... 28

2.4.1. Application of composites ................................................................................. 28

2.4.2. Braided composites ............................................................................................ 29

2.4.3. Application of braids and braided composites ................................................... 30

2.5. Relation between mandrel position and compensation length during

braiding ..................................................................................................................... 31

2.5.1. Introduction of compensation length ................................................................. 31

2.5.2. Length from eyelet to braid ring ........................................................................ 32

2.6. Braiding using non-circular braid ring .......................................................... 33

2.7. Wrinkle formation on braided preform ......................................................... 36

2.7.1. Causes of wrinkle formation .............................................................................. 37

2.7.2. Wrinkle caused by bending ................................................................................ 38

2.7.3. Minimising wrinkles .......................................................................................... 41

2.8. Conclusions ........................................................................................................ 42

2
3. Experimental Methodology.............................................................................. 44

3.1. Introduction....................................................................................................... 44

3.2. Braiding preparation ........................................................................................ 45

3.2.1. Mandrel selection and preparation ..................................................................... 45

3.2.2. Bobbin winding .................................................................................................. 47

3.2.3. Braiding.............................................................................................................. 47

3.3. Composite fabrication ...................................................................................... 49

3.3.1. Introduction ........................................................................................................ 49

3.3.2. Resin formulation and degassing ....................................................................... 50

3.3.3. Resin infusion .................................................................................................... 50

3.3.4. Curing and demoulding...................................................................................... 51

3.4. Conclusions ........................................................................................................ 53

4. Effect of Linear Compression on Braided Preforms ..................................... 54

4.1. Introduction....................................................................................................... 54

4.2. Prediction of biaxial braid angle under compression/tension loading ......... 54

4.3. Linear tension and compression loaded on sleeves ........................................ 58

4.4. Measurement ..................................................................................................... 62

4.4.1. Braid angle measurement ................................................................................... 62

4.4.2. Braid diameter measurement ............................................................................. 64

4.5. Conclusions ........................................................................................................ 69

5. Effect of Bending on Braid Parameters .......................................................... 71

5.1. Introduction....................................................................................................... 71

5.2. Relationship among braid angle, diameter and length of braids ................. 71

5.3. Bending machine ............................................................................................... 73

5.4. Bending process................................................................................................. 74

5.5. Measurement ..................................................................................................... 76

5.6. Braid parameter formation.............................................................................. 76

5.7. Effect of bending on braided parameters ....................................................... 78

3
5.8. Conclusions ........................................................................................................ 85

6. Methods for Local Braid Angle Change ......................................................... 86

6.1. Introduction....................................................................................................... 86

6.2. Change in shape of braid ring ......................................................................... 86

6.2.1. Braiding.............................................................................................................. 88

6.2.2. Limitation of elliptical braid ring in maypole braider ........................................ 91

6.3. Change in the position of the core ................................................................... 91

6.3.1. Relation between carrier compensation length and off-centre braiding............. 93

6.3.2. Limitation of off-centre braiding ....................................................................... 94

6.3.3. Braiding with off-centre setting ......................................................................... 96

6.3.4. Effect of core position on braid parameters ....................................................... 98

6.4. Conclusions ...................................................................................................... 102

7. Analysis and Discussions ................................................................................ 104

7.1. Effect of linear compression and linear tension on braided preforms....... 104

7.2. Effect of bending on braided preforms ......................................................... 105

7.3. Effect of shape of braid ring (elliptical braid ring)...................................... 107

7.4. Effect of off-centre setting of the braid ring ................................................. 108

7.5. Wrinkle formation after resin infusion......................................................... 108

7.6. Conclusions ...................................................................................................... 111

8. Conclusions and Future Work ...................................................................... 113

9. Bibliography .................................................................................................... 116

4
 List of Figures
Figure 1 A composite sample with wrinkle analysed to observe the extent of out

of plane deformation (a) Inner side of a bent braided composite tube with

wall wrinkle generated during composite manufacturing. (b) Bent

composite tube was cut along its length. (c) Details of wrinkle formation 18

Figure 2 The angle α between yarn axis and braid axis is braid angle .............. 23

Figure 3 The braiding unit cell [15] ................................................................... 25

Figure 4 Tow path of lay length ......................................................................... 25

Figure 5 Side view of a braided textile illustrating measurements for crimp

calculation notation [19] ............................................................................ 27

Figure 6 Composite materials are used (a) throughout the body of Boeing 787

[24] and (b) Lamborghini car [25] ............................................................. 29

Figure 7 (a) Complex shapes are manufactured (b) RTM (Resin Transfer

Molding) technology used to manufacture complex shaped braided

composite [33]............................................................................................ 31

Figure 8 Schematic of front view of bobbin carrier track (left) and side view of

carriers (right) [34] ..................................................................................... 32

Figure 9 (a) Maypole braiding machine with an elliptical braiding ring. (b)

Braiding angle changes with the changes on the radius of the creating

ellipses [14] ................................................................................................ 33

5
Figure 10 Schematic diagram of the convergence zone between the elliptical

ring and the fell point showing change in convergence length and fell point

position due to the ellipticity of the ring .................................................... 34

Figure 11 Braiding using different braid ring [35]............................................. 35

Figure 12 Optical microscopic image showing a circumferential cross-section of

the tube. The wrinkle amplitude is twice as much as the wall. [45] .......... 36

Figure 13 Wrinkling due to shearing. Wrinkle appears after the locking angle

achieved [49] .............................................................................................. 37

Figure 14 Microphotographs of the cross section of samples with defect at the

corner of the laminates (a) (b) wrinkles appeared on different materials

(convex tool). (c) (d) wrinkles appeared on different materials (concave

tool) ............................................................................................................ 39

Figure 15 Photographs of cross section of samples showing external wrinkle.. 39

Figure 16 Mechanics of the bending: analogy with a solid rod. The neutral axis

moves toward outer surface of bend [52] ................................................... 40

Figure 17 Weave Modifications to Remove Wrinkles (a) Normal fit. (b) Re-

engineered fabric [56] ................................................................................ 42

Figure 18 Wrinkle appears on the metal after bending (a) side view (b) top view

.................................................................................................................... 45

Figure 19 PVC tube used in this study............................................................... 46

6
Figure 20 Herzog USP 300. (a)Winding machine (b) fibre guide (c) bobbin (d)

back view of the winding machine............................................................. 47

Figure 21 The Cobra braiding machine during manufacture of a specimen ...... 48

Figure 22 Overbraided 180° bent tubular mandrel under vacuum along with

additional vacuum bagging assembly ........................................................ 51

Figure 23 Cure cycle of epoxy resin (Araldite LY 564 and Aradur 2954) for

composite manufacturing. The dotted line represents the uncontrolled

cooling within the oven chamber ............................................................... 52

Figure 24 (a) Regular braid structure (b) Considering the braid unit cell acts as a

trellis in which the intersecting four sides are pivoted [60] (c) unit cell

schematic (d) unit cell schematic under compression assuming the sides

are pivoted [61] .......................................................................................... 55

Figure 25 Device used for linear compressing of braid. Blue arrows are

compression forces ..................................................................................... 58

Figure 26 Images show the linear compression of a braid. The two ends of

sleeve were folded instead of securing with cello tape which could apply a

boundary condition and prevent free movement of fibre duing compression

.................................................................................................................... 59

Figure 27 Linear compression force applied on a braided sleeve. Initially braid

angle is ±45°. The decrease in length leads to increase in braid angle and

braid diameter after compression ............................................................... 61

7
Figure 28 Linear tension force applied on a braided sleeve. Initially braid angle

is ±45°. The increase in length leads to braid angle and braid diameter

decrease after compression. ....................................................................... 61

Figure 29 (a) Braid angle measurement using digital protractor (b) Braid

diameter measurement using digital calliper.............................................. 62

Figure 30 Relation between displacement and angle. Braid angle is ±45° when

displacement is 0. Braid angle increases under compression and decreased

under tension until it reaches to locking angle. The dotted red line when

axial displacement is beyond 40 mm means the sleeve cannot be stretched

any more ..................................................................................................... 63

Figure 31 Gap between interlaced fibre tow. Gap has an impact on diameter

measurement using callipers. ..................................................................... 65

Figure 32 Relation between displacement and diameter. Braid diameter

increases or decreases under compression or tension respectively until

braid angle reaches to locking angle. ......................................................... 66

Figure 33 Areal density increased with the increase of displacement. .............. 67

Figure 34 Way of yarn spacing measurement .................................................... 68

Figure 35 Yarn spacing of the braid. Under linear compression and tension,

yarn spacing decreased until braid angle reached to locking angle. .......... 69

Figure 36 Schematic of a 90° bent tube [61] ..................................................... 72

Figure 37 (a) Top view of a carbon fibre braided core on bending equipment (b)

Hydraulic bending equipment used for this study...................................... 73

8
Figure 38 Cross-section of braided tube changed after bending. ‘Brazier effects’

could not be eliminated by filling the tube with sand but it was improved.

Also it was verified that wet sand had better effect than dry sand. ............ 74

Figure 39 Bent tube. Tubes were bent to 90° and 180° respectively ................. 75

Figure 40 Schematic of a 180° bent tube indicating different bending

parameters [61]........................................................................................... 76

Figure 41 Compression side, tension side and neutral sides at the curvature

section after bending .................................................................................. 77

Figure 42 Braid angle formation after bending .................................................. 79

Figure 43 Schematic shows braid length change under compression of a

multilayer braid .......................................................................................... 80

Figure 44 After bending a preform braided on a 25.4 mm diameter mandrel, at

compression side, braid angle increased and reached to the highest angle at

the centre of the curvature, and vice versa. The braid angle of double

layers braids increased more rapidly than single layer braids at

compression side but decreased slower than single layer braids. .............. 82

Figure 45 When a preform braided on a 12.7 mm diameter mandrel was bent,

the change of braid angle was similar to 25.4 mm diameter mandrel bent.

But braid angle changed more gradually from edges to the centre. ........... 83

Figure 46 Figure showing the maximum braid angle at compression side and

minimum braid angle at tension side for a single layer braid after bending

at 90° and 180°. The braid angles are collected from the centre of the

9
 curvature. Predicted braid angle were predicted by using equation 21

illustrated in Chapter 4.2 ............................................................................ 84

Figure 47 Image of elliptical braid ring ............................................................. 87

Figure 48 (a) Braid machine with elliptical braid ring (b) Schematic of a front

view of the braiding machine (c) Two sets of braid angle were produced on

four sides of the mandrel with the same angle in opposing sides .............. 89

Figure 49 Image of plate used for mandrel off-centre setting ............................ 92

Figure 50 Schematic of the front view of a braiding machine. Horizontal plane

off-centre location of the mandrel is shown with respect to braid ring. Four

measurement positions on the mandrel are shown in anti-clockwise

direction...................................................................................................... 92

Figure 51 Schematic of circumferential braiding compensation length variation

between braid ring and braid point. (left) side view. (right) top view ....... 93

Figure 52 Factors affecting the required compensation length during braiding.

(a) Original setting is setting the distance between a braid ring and the

eyelet to 100 mm with braid ring radius 100 mm and convergence length

110 mm. (b) (c) (d) When the mandrel was set 50 mm off-centre, the

increase in the distance between braid ring and carrier eyelet, braid ring

radius and convergence length (lower braid angle) resulted in more

compensation length was required. ............................................................ 95

Figure 53 Fibre loosed when mandrel was set 62.5 mm away from the machine

centre .......................................................................................................... 97

10
Figure 54 Effect of compensation limitation in braiding process. Mandrel was

set 100 mm off-centre (a) front view of fibre slack (b) fibre slack during

braiding (c) fibre hoop was created because of slack fibre tow (d)

convergence length various due to fell point moving (e) fibre tow out of

guide causing fibre damage ........................................................................ 98

Figure 55 Comparing braid angle (maximum and minimum) before bending and

after bending to 180° (12.5 mm offset). The dotted line indicates braid

angle at the neutral axis. ............................................................................. 99

Figure 56 Change in braid angle (maximum and minimum) after bending

(elliptical braid ring) ................................................................................ 106

Figure 57 Half piece of composite tubes for observation of wall thickness (a)

Wall thickness variation with small wrinkles are observed for composite,

the preform of this composite was braided by using circular braid ring

where the braid angle changed at the compression side to ±54° from ±45°

at neutral axis. (b) Uniform wall thickness at compression side of the tube

and the preform of this tube was braided by using elliptical braid ring

where the angle changed to ±45° at compression side after bending from

±41° before bending ................................................................................. 109

Figure 58 Cross section of inner bent section. (a) Preform braided by using

circular braid ring. (b) Preform braided by using elliptical braid ring ..... 111

11
 List of Tables
Table 1 Basic data for TORAYCA: T700SC 12000 60E [57] .......................... 44

Table 2 Outlines the details of tube braiding ..................................................... 80

Table 3 Length change under compression on a multilayer braid ..................... 81

Table 4 Braiding process parameters ................................................................. 88

Table 5 Braid angle are the same at opposing quadrants around the mandrel ... 90

Table 6 Effect of convergence length on braid feasibility ................................. 91

Table 7 Data setting for off-centre braiding ....................................................... 97

Table 8 Change of braid angle (°) after bending to 180................................... 101

Table 9 The maximum braid angle (°) at compression side after bending ...... 107

Table 10 The minimum braid angle at tension and the maximum braid angle at

compression side after bending to 180° ................................................... 110

Table 11 Mean average thickness (mm) at tension and compression side after

resin infusion ............................................................................................ 110

12
 Abstract

The University of Manchester

Danjie Yang
Master of Philosophy
Investigation of Braid Geometry for Complex Braided Composite Structures
February 2018

The aims of this study are to investigate the mechanism of wrinkle formation on
cylindrical preform during bending and minimise the wrinkle formation to improve
the properties of the final products. It was found that changes in braid parameters
due to the scissoring of fibre tows was one of the major influencing factors for
causing out-of-plane deformation. In order to minimise or potentially eliminate this
out of plane deformation that leads to wrinkle formation, the changes in braid angle
and braid diameter under uniaxial tension and compression loading have been
studied.
In this study, biaxial braid with ±45° braid angle was produced using T700 Toray
carbon fibre tows on a circular cross-section cylindrical mandrel. Bending of the
over-braided preform produced a significant change in braid angle around the
circumference as well as along the length as a result of opposing forces (tensile and
compressive) at the outer and inner of the bend. Optimisation of braid parameters
was achieved by changing the braid angle locally around the circumference of the
braid to compensate for distortion during the bending process.
Reducing the braid angle during braiding at the compression side of the mandrel can
minimise the wrinkle formation during preforming using bending. Two methods
were used to accomplish the change in braid angle in this study. One method was
braiding using an elliptical braid ring instead of a circular braid ring. Two diferent
braid angles were produced on the mandrel major and minor diameter of braid ring.
The secondary method was braiding the mandrel/core off-centre of the braid ring.
Different braid angles were achieved on the plane of off-set because of the difference
in distance from braid ring to braid fell point. The braid angle measured after
bending these braids and the change in braid angle were optimised due to changes in
braid angle in the tension and compression sides as opposed to the angle along the
neutral axis. This change eliminated one of the major factors that give out-of-plane
deformation, resulting in wrinkles.
Braided structures that have been pre-formed to complex geometrical shapes were
then resin infused to produce composites. Two preforms were braided using circular
braid ring or an elliptical braid ring respectively. After bending and under pressure
during resin infusion, optimised structures showed an improvement in wall thickness
as well as wrinkle formation. Wrinkles were formed during resin infusion for the
braid produced with a circular braid ring. The braid angle reduction in the preform
stage resulted in the elimination of wrinkle formation. The optimised braid structure
improved wall thickness of the composite.

13
 Declaration
I declare that no portion of the work in the thesis has been submitted in support of
another degree or qualification in this or any other university or another institute of
learning.

14
 Copyright Statement
i. The author of this thesis (including any appendices and/or schedules to this thesis)

owns certain copyright or related rights in it (the “Copyright”) and s/he has given

The University of Manchester certain rights to use such Copyright, including for

administrative purposes.

ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic

copy, may be made only in accordance with the Copyright, Designs and Patents Act

1988 (as amended) and regulations issued under it or, where appropriate, in

accordance with licensing agreements which the University has from time to time.

This page must form part of any such copies made.

iii. The ownership of certain Copyright, patents, designs, trademarks and other

intellectual property (the “Intellectual Property”) and any reproductions of copyright

works in the thesis, for example graphs and tables (“Reproductions”), which may be

described in this thesis, may not be owned by the author and may be owned by third

parties. Such Intellectual Property and Reproductions cannot and must not be made

available for use without the prior written permission of the owner(s) of the relevant

Intellectual Property and/or Reproductions.

iv. Further information on the conditions under which disclosure, publication and

commercialisation of this thesis, the Copyright and any Intellectual Property and/or

Reproductions described in it may take place is available in the University IP Policy

(see http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=487), in any relevant

Thesis restriction declarations deposited in the University Library, The University

Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/regulations)

and in The University’s policy on Presentation of Theses.

15
 Acknowledgement

I would like to take this opportunity to extend my sincere thanks to all the

individuals and organisations for their kind support and help in the completion of

this project. First and foremost, I would like to show my greatest appreciation to Prof

Prasad Polturi who provided me with this opportunity to carry out this study. Also

thanks for his valuable feedback, guidance and encouragement.

I would like to thank my academic advisor Dr Sree Shankhachur Roy who offered

me invaluable assistance and support through the whole project duration. He

explained knowledge in this subject area and analysis to me. Under his guidance, I

managed to conduct the experimental work successfully. I also got a better

understanding of the composite mechanism under his instructions.

I thank colleagues Dr Mayank Gautam, Kinjalkumar Patel and Lamis Jfairi for

maintaining a friendly environment in the office. I thank Dr Vivek Koncherry, Dr

Mohammad Islam, Hussein Dalfi and Roy Conway for everyday support that

supports me spiritually. Also I thank Mr Thomas Kerr from weaving laboratory for

his help.

I also wish to express my gratitude to the School of Materials at the University of

Manchester for providing all the instruments and materials I needed to carry on my

project research. The project went on well with their support.

Finally, I want to thank my parents and friends for their constant encouragement and

support.

16
1. Introduction
1.1. Introduction to the research topic

Composites structures can sometimes have complex cross sections, rather than linear,

with bending along the length. However, fibre orientation which is a significant

parameter in determining the composite mechanical properties, can also change due

to the change in shape [1]. Wrinkling is a serious defect that could potentially form

during composite manufacturing due to such change in fibre orientation in bent

sections (Figure 1) and it can lead to lower mechanical properties. In addition, if an

over-braided preform is bent to a complex shape, fibre orientation can change as the

fibres shear. This change in fibre orientation can influence wrinkle formation. This

study will find out the relationship between fibre orientation and post-bending

wrinkle formation. Two methods will be presented to minimize or even eliminate

wrinkle formation.

One of the methods that will be attempted in order to minimise wrinkles is a change

in fibre orientation (in case of a braid, the braid angle) circumferentially where

bending forces a change braid angle. The rotational speed of carriers on the braiding

machine and take-up speed of the core (mandrel) are the two main factors that can

change the braid angle along the length of the mandrel (keeping the core diameter

constant) [2].

17
 (a)
(c)

(b)

cm

Figure 1 A composite sample with wrinkle analysed to observe the extent of out of
plane deformation (a) Inner side of a bent braided composite tube with wall wrinkle
generated during composite manufacturing. (b) Bent composite tube was cut along
its length. (c) Details of wrinkle formation

This change in braid angle is the same around the circumference of the mandrel as a

braiding machine uses a circular braid ring or former which helps in braid formation.

But braid angle cannot be changed around the circumferential directions with help of

a circular braid ring. In this context, an elliptical braid ring can potentially achieve

different braid angles circumferentially. Another method used in this study was

mandrel off-centre setting to change the braid angle locally around the circumference

direction.

CFRP (carbon fibre reinforced polymer) is formed of two components: polymer

(often epoxy resin) as the matrix and carbon fibre as the reinforcement. CFRP can be

expensive to produce [3] but is widely used in applications requiring high strength-

18
to-weight ratio and high rigidities, such as aerospace, automotive, civil engineering,

sports goods and other technical applications. In structural composite applications

performance can be affected by wrinkle formation. Vacuum bagging is a widely used

technique for manufacturing CFRP composites. However, this technique can

generate wrinkles when manufacturing braided composites. Vacuum bagging was

used to apply mechanical pressure on the laminate in a consolidation process. When

pressure is applied on the fibre preform, the thickness of the preform will reduce

during a debulking process. In this case, with the total amount of fibres unchanged,

the excess fibres are stacked to form wrinkles. This study used an elliptical braid ring

and set the mandrel off-centre to achieve local braid angle changes and eliminate the

unexpected wrinkle defects.

1.2. Research objective

The aim of this study is to optimise braided preforms for bent tubes containing

region double curvature. The objectives of this research were categorized as follows.

• Study the change in the braid angle and minimise the angle distortion during

bending.

• Investigate the mechanism of wrinkle formation during bending.

• Investigate methods to minimise wrinkle formation to improve the properties

of final products.

19
1.3. Outline of thesis

Chapter 1 briefly introduces braiding and braided composites. Methods to mitigate

wrinkle formation in bent braided composites were put forward.

In Chapter 2, a detailed review of literature in reference to braiding, braided

composites, compensation length, braiding using non-circular braid ring and wrinkle

formation is presented.

In Chapter 3 describes bobbin winding and braiding process. The process of

manufacturing of composite by vacuum bagging was also covered in this chapter.

In Chapter 4, a detailed study of braid angle change under linear tension and

compression was conducted due to its relevance with bending of braided preforms.

Chapter 5 describes the observations as the bending process was carried out. The

braid angle increased under compression loading at the compression side and

decreased at the tension side.

The Chapter 6 describes two methods to optimise the braid angle in the preform. One

is using an elliptical braid ring instead of a circular braid ring for braiding. The other

one is setting the mandrel off-centre to achieve local changes in the braid angle.

In Chapter 7, the results are shown with analysis. At the end of this chapter, the

conclusions from the study are presented and recommendations for future research

work are proposed.

20
2. Literature Review

A comprehensive review of literature with reference to the subjects of research is

presented in this chapter. Firstly, some important braid parameters relevant to the

study are introduced. Then an introduction to composite materials, composite

manufacturing and the application of composites is given. This is followed by

literature related to a technology of non-circular braid ring braiding. Finally, wrinkle

formation in braided composites is investigated.

2.1. Introduction to braiding

Braiding is a textile method of producing a braid structure that was defined as “a

process of interlacing three or more threads diagonally to the product axis (parallel to

the longest dimension of the resulting product) in order to obtain a thicker, wider or

strong product or in order to cover (over-braid) some profile” [4]. One of the earliest

documented braid textile structures was produced in China over 4000 years ago [5].

The maypole braiding machine was inspired by the maypole dance, which was a

conventional braiding machine that can produce tubular braided structures. Guyader

et al. [6] illustrated the whole braiding process. In order to produce a braid, the horn

gears are placed on the machine and the gears rotate with a constant angular velocity

and the carriers move along the track path in two opposite directions. Thus, the

interlacement of the yarns is deposited on the mandrel. Yarns are wound on bobbins

placed on to the carriers. The yarn ends are fixed on the mandrel which is pulled

through the braid ring by take-up mechanism in the centre of the whole machine.

21
The maypole braiding machine is inexpensive and can be run at high production

speeds that makes it suitable for producing low-cost, high volume composites. Over-

braiding a core is becoming popular in composite manufacturing. In this project,

braiding of high performance carbon fibres will be carried out with the aim of

developing composite.

2.2. Braid machine components

The maypole braiding machine consists of three primary components: a track plate,

carriers, a braid ring and take-up device [7]. The yarns are wound on bobbins, which

are mounted on carriers. The carriers drive horn gears move follow a track, which

determines the carrier path and provides the resistance to the carrier to keep the track

following the carrier in the horn dogs [2, 8]. Carriers are placed on horn gears, which

usually have 4 slots. When a carrier reaches a track intersection, the driving forced

combined with shape of the track transfer the carriers from one horn gear to the next

[9]. The take-up mechanism pulls the fibre tow out from a bobbin when a carrier

rotates during braiding.

2.3. Braid parameters

The structural parameters can affect the braid geometry and performance. Braiding

provides the primary features of composite reinforcement by laying fibres in many

different ways. The process can vary the microgeometry of braid structure and

produce varied parts [10]. The structural geometry of the braid is determined by

many parameters, such as braid angle, convergence length and cover factor.

22
2.3.1. Braid angle

Braided composite has a feature that its mechanical properties can be changed by

changing the fibre orientation, which is also called ‘braid angle’. The braid angle,

which is the most important parameter in the braiding process, is made by the yarn

with respect to the braid axis. To change the braid angle, either or both the take-off

speed of the mandrel as well as the rotational speed should be changed. The braid

angle (α) can be calculated using the equation 1[2].

ωR (1)
α = tan−1 ( )
v

Where α is braid angle (between yarn axis and braid axis); R is radius of cylindrical

mandrel (cm); v is take-up speed (cm/s); ω is average angular velocity of bobbins

around the machine centre (rad/s).

Yarn axis

α
Braid axis

Figure 2 The angle α between yarn axis and braid axis is braid angle

Braid angle is related to rotation speed of horn gears around their own centre, the

number of horn gears as well as take-up speed. The equation [2] is shown below.

23
 2ωh R (2)
α = tan−1 ( )
Nh v

Where ωh is average angular velocity of the horn gears around their own centre

(rad/s); Nh is number of horn gears

The longitudinal velocity of the mandrel and the angular rate of the spindle can be

changed to change the braid angle on a cylindrical braided fabric. Fibre orientation is

one significant parameter as it has an effect on mechanical properties [11, 12].

Lower braid angle leads to higher stiffness longitudinally, along the braid cylinder

axis. Higher angle improves the radial strength [13]. However, change the

longitudinal velocity of mandrel and rotation speed of the machine can only change

the braid angle in axial directions. Nishimoto et al. [14] experimented to change the

braid angle in circumference direction is that changing the form of braid ring to

ellipse shape rather than a circle. It will be discussed in section 2.6.

2.3.2. Unit cell

The braid unit cell for the 2D braided architecture is shown in Figure 3. The overall

volume of the braid unit cell is [15]:

4Ha2
V= (3)
sin2α

Where H= braid thickness; a= yarn spacing.

24
 Figure 3 The braiding unit cell [15]

2.3.3. Lay length

Each yarn forms a helical path around a mandrel. The lay length, which is known as

pitch length, is the longitudinal distance along the mandrel required for one complete

helical wrap. Lay length is calculated using equation 4. Variation of lay length

influences the final product performance due to its fibre orientation.

πD (4)
L=
tanα

Ll

Figure 4 Tow path of lay length

25
2.3.4. Cover factor

In a review by Potluri et al. [2], it has been described that cover factor is the

percentage of yarn covering the mandrel. It is the ratio of covered area over total area

of the mandrel. Cover factor for a biaxial braid can be mathematically represented as:

Wy Nc 2 (5)
CF = 1 − (1 − )
4πRcosα

In this equation, Wy = yarn width; NC = number of yarn carriers; R= effective

mandrel radius; α= braid angle

According to Zhang Q et al. [16] analysis, cover factor is related to braid angle and

mandrel diameter. Increasing the braid angle or decrease the mandrel diameter can

result in an increase in cover factor. Higher coverage occurs at particularly high or

low angles. The minimum cover factor always occurs at a braid angle 45°.

2.3.5. Crimp

Crimp is defined [17] as "the waviness of a fibre" and is normally expressed

numerically as either "the number of waves or crimps per unit length" or "the

difference in distance between points on the fibre as it lies in a crimped condition

and the same two points when the fibre is straightened under suitable tension". As

the crimp reduces, the mechanical properties of composites reinforced with braid

increased. But drape and in-plane permeability are generally decreased as the crimp

decreases.

The Crimp (C), which is referred as the wave of a yarn in an interlaced structure, is a

measure of the amplitude of the yarn in the 2D braided structure. The crimp is

26
calculated by the change in length of straightened yarns and the length of fabric. The

equation in shown as [18]

𝐿𝐿 − 𝑙𝑙 (6)
C= × 100%
𝑙𝑙

Where C is crimp percentage; L is the length of straightened yarn before crimping; l

is the length of fabric after crimping.

Figure 5 Side view of a braided textile illustrating measurements for crimp

calculation notation [19]

This parameter may change due to factors that increase or decrease the length of yarn

path. For example, change the diameter of the yarns or the float length of the

interlacement repeat. The increase in yarn's unit length or yarn thickness can result in

a higher crimp percentage. Thus, it is believed that the crimp in a diamond braid is

higher than any other structures. The crimp can be affected by variables braid angle

and the number of yarns as well.

27
2.4. Application of composites and braided composites

A composite material is an engineering material that is made up with two or more

materials, which consists of a core reinforcement reinforced by a matrix [20]. The

most efficient method that has been found to use fibres as engineering materials is to

‘combine a fibrous material of high tensile strength and high modulus of elasticity

with a lightweight bulk material of lower strength and lower modulus of elasticity’

[21].

The fibre-reinforced composite is generally composed of three elements: the fibres,

the matrix and a fine interphase region which responsible for assuring the connection

between the matrix and fibre as the interface [22]. The resultant material has superior

performance to that of individual materials [23].

2.4.1. Application of composites

One of the key factors that leads to increasing global temperatures is transportation

which contributes to the output of greenhouse gases. Lightweight composite

materials help to reduce the mass of vehicles, is one solution that can have a

dramatic effect on the production of these emissions as with reducing vehicle weight,

the fuel consumption reduces. The use of lightweight materials such as carbon fibre

can trigger dramatic weight savings and thus reduce fuel consumption.

28
 (a) (b)

Figure 6 Composite materials are used (a) throughout the body of Boeing 787 [24]

and (b) Lamborghini car [25]

Composite materials have a wide range of applications in various industries. These

are widely used in automobile industries (e.g. substitute of Steel & Aluminium body);

Marine applications like shafts, hulls, spars (for racing boats); Aeronautical

application like components of rockets, aircrafts (business and military), missiles etc.;

Communication antennae, electronic circuit boards (e.g. PCB, breadboard); Safety

equipment like helmets for ballistic protection.

2.4.2. Textile composites

Textile composites, which are also defined as textile-reinforced composite materials

(TRCM), become more welcome than conventional materials because of its high

weight to stiffness ratio (specific stiffness), as well as improved fatigue life,

corrosion resistance, and with good design practice, reduced assembly costs due to

fewer detail parts and fasteners.

Variety shapes of composite structure can be produced because of flexibile fibrous

structures that can be shaped easily due to their drapability. Tubular composites are

29
usually made of long continuous fibre rather than short fibre because the continuous

fibre composites can reach the specific requirements to stiffness and weight in the

products [23]. Long continuous fibres are widely used to produce composite

structure by weaving, filament winding and braiding. In this project, preforms were

produced by 2D (two dimensional) braiding.

2.4.3. Application of braids and braided composites

Braiding can be used for many critical applications ranging from ancient decorative

belts to a variety of applications on air and spacecraft [7]. Braiding has been chosen

for producing cables, ropes, laces and so on. On the other hand, braids are also used

in many industrial applications such as fan blade containment in commercial

aircrafts; for energy-absorbing crash structures in Formula One racing cars; as

reinforcement for aircraft propellers and stator vanes in jet engines; for building

lightweight frames and structures such as trusses; for use in precise manufacture of

composite parts; ideal reinforcement for drive shafts and torque transfer components,

such as flanged hubs; for products with changing geometries like prosthetics and

hockey sticks etc. [26]. Brookstein[27] listed some of structural columns, rods, shafts,

pressure vessels and plates used in traditional applications where braid enhancement

has replaced conventional materials.

Braids are not limited within cylindrical structure anymore. As Branscomb et al.[28]

pointed that ‘braids can be, cylindrical (round), flat, or a variety of other cross-

sections’. For example, square, hollow square, I-beam, U-beam, T-beam, X-beam

[29], L-beam [30] and etc. It is possible to produce more complex-shaped parts by

using braiding technology [31, 32].

30
 (a) (b)

Figure 7 (a) Complex shapes are manufactured (b) RTM (Resin Transfer Molding)
technology used to manufacture complex shaped braided composite [33]

2.5. Relation between mandrel position and compensation length

during braiding

During braiding process, the lengths of fibre tow from eyelet to fell point are not

constent. This length variation is exactly the length that needs to be compensated by

the compensator roller. The braiding machine has its own ability to compensate for

length difference. So compensation length will be studied in this section.

2.5.1. Introduction of compensation length

As carriers move along the track, which is in a wave shape but not a circle, there is a

distance variation between its nearest and furthest point to the centre of the braiding

machine or from the fell point (Figure 8 left) [8]. Due to the fact that the take-up

speed of mandrel is constant during braiding process, which is set manually, fibre

tow will be pulled out more when a carrier moves from the nearest point to the

furthest point and that will result in an extra length of fibre tow. Compensators, play

its role to compensate the fibre tow during braiding and make the braiding process

go smoothly.

31
 l3
l1
l2
Contact point
Contact point
l4

Braid
Braid ring
ring

Figure 8 Schematic of front view of bobbin carrier track (left) and side view of
carriers (right) [34]

Usually, mandrel is placed at the centre of the whole braiding machine. The

compensation length will only be caused by variation distance from eyelets to centre

of the machine. Hence the carriers are designed to compensate the amount of fibre

pullback that is required due to the dancing motion of the carrier. However, in this

study, the mandrel was taken off-centre of the braiding machine. This caused

increase in fibre length difference in addition to the length difference caused by the

carrier motion. For a successful braiding keeping the mandrel off-centre more

compensation length will be needed.

2.5.2. Length from eyelet to braid ring

l1 is the nearest distance from eyelet to braid ring; l2 is the furthest distance from

eyelet to braid ring; l3 and l4 is the distance from the intersection point of two

bobbins to braid ring.

In order to find out the maximum compensation length between eyelets and braid

ring, the difference between l1 and l2 could be calculated,

32
 (7) [34]
∆lc = l2 − l1 = �(h1 + h2 )2 + d2 − �h1 2 + d2

Where Δlc is the difference between l1 and l2; h1 is the nearest vertical distance from

guide eyelet to braid ring contact point; h2 is the distance between two guide eyelets;

d is the horizontal distance from guide eyelet to braid ring contact point.

2.6. Braiding using non-circular braid ring

The braided composite has an advantage that their stiffness distribution can be

changed by changing the braiding angle. In order to change the braiding angle on a

cylindrical braided fabric, the longitudinal velocity of the mandrel and the rotation

speed of the carriers in a braider should be changed. However, these changes are

only in the longitudinal direction but not in the circumferential orientation. In this

case, Nishimoto et al [14] presented a method (using an elliptical braiding ring

instead of circular braiding ring) to change the circumferential distribution of

braiding angles, which made the braiding angle in the circumferential orientation

changed (Figure 9).

Figure 9 (a) Maypole braiding machine with an elliptical braiding ring. (b) Braiding
angle changes with the changes on the radius of the creating ellipses [14]

33
Also, the convergence length was not constant by using elliptical braid ring (Figure

10) because if the longitudinal velocity of mandrel changes, there is some delay for

actual braid angle reach the designated braid angle. Simulation the convergence

length was constant; the braiding angle θ is determined only by the changes in the

radius of the creating ellipse. Therefore, changing the shape of braid ring can change

the braid angle distribution along the circumferential direction.

Convergence length

Mandrel

Braid

Figure 10 Schematic diagram of the convergence zone between the elliptical ring and
the fell point showing change in convergence length and fell point position due to the
ellipticity of the ring

Ebel and Hans [35] braided on different cross-section mandrel and found that it

affect yarn path and thus braid angle. But their experimental braid angle had

deviation compared to theoretical yarn path. They minimised deviations by changing

shape of braiding ring. Circular and three elliptical braid ring with different aspect

ratio were used to achieve it. Also, three different cross-section of mandrels (circular,

34
square and rectangular) were braided on. They presented that braid angle could be

controlled by changing the braiding ring shape. The four different shapes of braid

rings used in their study are shown in Figure 11. It could be seen that the increase in

one diameter of the braid ring led to braid angle on right and left sides decrease, but

braid angle on top and bottom sides increased. It was the effect of braid ring ratio on

braid angle deviation. Their study showed that different cross-section mandrels all fit

this result.

Figure 11 Braiding using different braid ring [35]

Moreover, ±30° and ±60° structures were braided to find that the influence of ring

shape decreases with increasing braid angle. However, it was not suitable for small

angle structure. Braid angle formation was irregular on ±30° braid.

In both the studies [14] [35] with elliptical braid ring, braid angle were same at

opposite sides of mandrel. With the increase in major diameter of the elliptical braid

ring, the angle deviation increased as well. In Ebel and Hans’ [35] study, the angle

changed by using circular braid ring (Figure 11) was because of the square mandrel

35
as the flat surfaces will affect the fibre deposition. [36]

2.7. Wrinkle formation on braided preform

Defects in composite materials that may be created during manufacturing tend to

degrade the performance characteristics of the final product [37, 38]. There are many

types of defects, such as voids, delamination, incomplete resin cure, fibre crimp,

wrinkle, etc. may appear during textile composite reinforcement [39]. One of the

more serious flaws that arise during forming is wrinkling [40, 41], which tend to

significantly degrade the performance characteristics of the final product [42, 43].

Wrinkling would lead to a serious loss of mechanical properties [44]. Hence,

reducing wrinkles in the final product in order to improve the forming of the material

is one of the main aims of optimizing textile composite structures.

Figure 12 Optical microscopic image showing a circumferential cross-section of the

tube. The wrinkle amplitude is twice as much as the wall. [45]

36
A lot of parameters can cause wrinkle, such as the number of plies, ply placement,

fibre architecture, etc. during forming. Figure 12 shows a wrinkle in a braided

composite tube manufactured using vacuum bagging method.

2.7.1. Causes of wrinkle formation

When the fabric is deformed in the compression direction, buckling will occur and

wrinkles will be formed. Likewise, it is also a problem in fabric reinforced

composite forming. Wrinkling will occur when the preform is required to shear too

high to form a particular geometry [37, 46]. The locking angle can be used to

describe the potential of a fabric to produce a particular surface. Prodromou and

Chen [37] defined the locking angle as ‘the angle achieved just before the onset of

buckling’. The formation process of wrinkle is illustrated in

Figure 13. They observed that the deformation behaviour of fabric can be divided

into two parts: when a low load is applied, trellising will occur before fabrics lock-up;

while applying a sudden increase in load, out-of-plane bending (wrinkling) occurs

after lock-up. Some other in-plane shear studied also analysed wrinkling focusing on

locking angle as well [47, 48].

Figure 13 Wrinkling due to shearing. Wrinkle appears after the locking angle
achieved [49]

37
Lebrun et al. [50] used a bias-extension test to measure the shear properties. During

testing, the specimen was applied a pure intra-ply shear deformation, where the angle

between tows decreased gradually until it reaches locking angle and formed a

wrinkle. Zhu et al. [40] proved Prodromou and Chen’s opinion. They figured out that

two deformation types, in-plane shear and out-of-plane wrinkling, may take place

during the textile fabric formation. They built up a theoretical model based on

previous specimen to see how the fibre sheared under stretching. Boisse et al. [49]

applied load on a unit woven cell and analysed the role of the three rigidities of

woven fabric in wrinkling simulations. The three rigidities are tensile, in-plane shear

and bending. They found that tensile rigidity was usually avoided in forming process.

The in-plane shear stiffness increased when the shear angle became larger. Bending

stiffness played a significant role in wrinkle size.

2.7.2. Wrinkle caused by bending

Different cross-section of mandrels can affect the braiding angle contribution along

circumferential direction. Also different shape of the mandrel (curving mandrel) can

influence the braiding angle. However, when braiding on these curving mandrel,

Hubert, Poursartip[51] found that angles processed on convex tools in general

exhibited thinning at the corner, whereas thickening at the corner was observed for

angles processed on concave tooling. Fibre movement during the curve process

which causing the changes in geometry and properties, and leading to stresses and

distortion. As shown in Figure 14 (a) (b), the presence of wrinkle resulted in the

higher thickness at the corner. The resin that is shown in Figure 14 (c) was created

because of the bag bridging at the corner. Resin filled the gap that created by bag

bridges. Figure 14 (d) shown bag indentation at the corner of concave tool.

38
 1 mm

Figure 14 Microphotographs of the cross section of samples with defect at the corner
of the laminates (a) (b) wrinkles appeared on different materials (convex tool). (c) (d)
wrinkles appeared on different materials (concave tool)

1 mm

Figure 15 Photographs of cross section of samples showing external wrinkle

39
The materials layers between preform surface and the bag can form wrinkles due to

the curvature of the mould surface. Typically it was more difficult to eliminate the

wrinkle with a concave tool.

Generally it is assumed that the fibres on longitudinal direction are under an amount

proportional tension to their distance from the neutral axis, which is also, happens to

the centroidal axis. However, in the experiment, it was found that the neutral axis

moved a little bit forward to tension side.

Figure 16 Mechanics of the bending: analogy with a solid rod. The neutral axis
moves toward outer surface of bend [52]

According to Sharma et al. work, it was no longer linearly proportional because of

the limit to the fabric extension [52]. It would cause the neutral axis not coincide to

the centroidal axis (Figure 16). If the maximum tension was same as the maximum

compression, the neutral axis would be same as centroidal axis. But for most fabric,

40
the extension was limited. The neutral axis was usually above the centroidal axis.

They also indicated that the extension on outer surface of a circular bent tube applied

along both the length direction and the cross-section direction of the preform. Here

‘Brazier effects’ were come up with which cause a circular cross section deformed to

be an oval shape.

Karman [53] indicated that when the tube is subjected to bending, for example, in

order to increase the curvature, the stress will have a tendency to push the

components on both sides of the transverse section towards the neutral surface. Thus,

the farthest material from the neutral axis is enabled to shirk its effect and greatly

reduce the stress, and the maximum stress will occur at point closer to the neutral

axis.

2.7.3. Minimising wrinkles

It was assumed that fabrics having higher friction coefficients will have larger

locking angles, while those having lower friction will have smaller locking angles

[37]. According to Boisse et al.’s study, it shows that the bending stiffness only

determines the shape of wrinkles. Hence the increase of bending stiffness can lead to

an increase of wrinkle size.

According to Yu et al [54], the blank holder force can be used to prevent wrinkling

as an independent optimization parameter. Lin et al [55] found two ways to prevent

wrinkling during composite forming. One is optimizing processing parameters and

the other one is by material selection. They focused on the effect of various blank

holding forces on wrinkling and they demonstrated that using the blank holder

profile can be used to minimize winkling because of the orthotropic anisotropic

behaviour of textile composites. In another study [56], a tubular woven fabric

41
preform was draped over a bent shape (Figure 17). In the study it was observed that

the fabric at compression side produce ‘wrinkles’. The authors proposed an

algorithm to re-engineer the fabric by changing weft density. In this study, changing

braid angle around the circumference will provide similar opportunity to decide

bending side to prevent wrinkle formation.

Figure 17 Weave Modifications to Remove Wrinkles (a) Normal fit. (b) Re-
engineered fabric [56]

2.8. Conclusions

The braid parameters described in this chapter, play a sufficient role in determining

braid geometry and structure. These factors influence each other and collectively

affect the braid density. From the studies conducted in literature, it has been found

that the braid angle distributions along circumferential direction can be changed by

using an elliptical braid ring instead of a circular braid ring. If an elliptical ring rather

than the circular ring is used, the braid angle in the axial direction remains constant,

but in the circumferential direction, the braid angle will change.

42
With the wide use of carbon fibre reinforced composites in the industries, there are

also demand for non-linear and complex shape products. During complex shape

manufacturing, the literature shows that unbalanced force leads to fibre shearing and

result in wrinkle formation. Although literature reviews were carried out on wrinkle

for textile preforms in composites, no studies on wrinkle for braided composite

manufacturing were found. In this study, based on the findings of the literature,

methods to rearrange the braid parameters will be investigated to minimise wrinkle

formation in final products.

43
3. Experimental Methodology
3.1. Introduction

A brief introduction to experimental process steps of this project has been presented

at the early part of this chapter. The preparation including bobbin winding, braiding

and vacuum bagging.

The carbon fibre used in this project was T700SC 12000 60E, and the properties of

this fibre is shown in Table 1.

Table 1 Basic data for TORAYCA: T700SC 12000 60E [57]

Tensile Strength 4900 MPa

Elastic Modulus 230 GPa

Strain 2.1 %

Density 1.80 g/cm3

Filament Diameter 7 µm

Number of filaments in a tow 12000

Fibre tow Count 800 tex

44
3.2. Braiding preparation

3.2.1. Mandrel selection and preparation

In this project, metal mandrel was not chosen because the over-braided mandrel

should be able to be bent flexibly by using bending machine, which will be

mentioned in Chapter 5.

(a)

(b)

Figure 18 Wrinkle appears on the metal after bending (a) side view (b) top view

However, wrinkle formation after bending is a serious issue which cannot be ignored.

As shown in Figure 18, the wrinkle that appears on the mandrel will definitely

influence the braids structures as well as the properties of braids. Hence, flexible

tube was used as mandrel rather than metal. Considering that mandrel will be placed

in the oven with braids for composite manufacturing, the operating temperature,

Glass transition temperature (Tg) as well as melt temperature should be suitable for

composite manufacturing procedure. The glass transition temperature (Tg) of Poly

vinylchloride (PVC) are about 82°C and the resin curing temperature was 80°C.

45
Since PVC is an amorphous thermoplastic, it softens gradually as temperature rises.

The melt point of PVC is between 100-260°C. PVC tube is flexible enough to be

bent to a certain curvature without cross section shape change. Due to these factors,

PVC tube was selected as suitable mandrel material for this project. Also wrinkle

will not be created after bending. In this study, 25.4 mm outer diameter PVC tube

and 12.7 mm outer diameter PVC tube were selected as mandrels.

In general, the mandrel used for braiding is required to be straight. The PVC tube

was received in the roll form and hence it needed straigtening. In this study, an

aluminium tube was inserted into the PVC tube to hold it and keep it straight during

braiding process. Before that, the PVC tube was put in oven and temperature was

raised to 82°C. Once the tube reached the Tg, it was cooled down to room

temperature. The process made the PVC tube mandrel straight.

Figure 19 PVC tube used in this study

46
3.2.2. Bobbin winding

(a)

(b) (d)

(c)

Figure 20 Herzog USP 300. (a)Winding machine (b) fibre guide (c) bobbin (d) back
view of the winding machine

In this project, for producing a regular braid structure (2/2) 48 bobbins were wrapped

with carbon fibre. The machine used for winding is Herzog rewinding unit USP 300

Eltra as shown in Figure 20.

3.2.3. Braiding

Each bobbin is mounted on a carrier and the fibre tow was passed around the

compensator roller and other guides. During braiding as the carrier moves in a

serpentine path, the fibre length between the carrier and braid fell point varies. As a

result the fibres lose tension and become slack. To prevent this, the carrier has spring

compensation method. The compensation length of the carriers used on the machine

is 210 mm. This means a length of up to 210 mm slack can be pulled back by the

compensation roller to under tension. The braiding machine has 48 carriers and the

47
machine was full of bobbins to produce a 2/2 (regular braid) structure. The roller

tensions were set by adjusting the torsion spring to a force of 1.5N. The tows were

passed through the braid ring and all were placed onto a mandrel. The ends of the

tows were secured onto the mandrel and braiding was carried out manually to set a

starting point.

Bobbins

Braid ring
(circular)

Overbraided
core (mandrel)

Clamp on the
take-up
mechanism

Figure 21 The Cobra braiding machine during manufacture of a specimen

The mandrel was clamped at one end by thread clamping system as tightly as

possible to prevent the slippage of the mandrel during braiding. The mandrel

slippage during braiding can lead to angle deviation from the setup. Braiding take-up

was controlled by an integrated computer operating CAM software developed by

Trio motion Ltd.

This automatic motion was determined by the input parameter, including take-up

speed and carriers rotation speed. The carriers were driven at a constant rate of 0.63

48
rad/s. The take-up speed of mandrel was driven at a specific speed, which is

depending on the desired braid angle. It was calculated by using equation 1.

The machine was kept at these speeds until the fell point approached the free end of

the mandrel, at which point the machine was stopped manually. As the braiding was

completed, the braid was tightly wrapped with a transparent adhesive tape to avoid

slippage of the braids, which is a key step in preventing the disruption of the braid

angle.

At this point the mandrel could be removed for a single layer braid. However, two

layers were needed to compare with single layer. The process was same as multiple

layers. Outer diameter was measured by using digital calliper. The second layer was

then braided in the same manner with the take-up speed adjusted to produce a

constant braid angle.

In this study, one 25.4 mm diameter tubes and one 12.7 mm diameter tubes were

braided to produce ±45° structures. Another two different sizes of tubes had two

layers of braids to identify the difference between one layer braid and two layer

braids.

3.3. Composite fabrication

3.3.1. Introduction

All preforms were consolidated by using vacuum bagging method. Composites

manufacturing can be summarised as two stages: preform manufacturing and resin

infusion. The preform is manufactured in braiding process in this study as explained

in previous section. The steps of vacuum assisted liquid resin infusion and curing is

described in the following sections.

49
3.3.2. Resin formulation and degassing

High curing temperature epoxies are widely used in aerospace composite

manufacturing. The epoxy resin was formulated by mixing the resin (Araldite LY

564) and hardener (Aradur 2954) at ratio 100:35 by weight in a container.

After mixing of two parts, the container was put in a vacuum chamber. As the

vacuum pump draws the air out of the chamber, the degassing process removes all

the air bubbles added during mixing.

3.3.3. Resin infusion

The overbraided mandrel was covered using release fabric, also known as ‘peel ply’,

is the first layer. It is used for easy separation of the composite from other infusion

aid materials after resin infusion. The infusion mesh which is usually highly

permeable warp knitted fabrics overwrapped on top of the release film wrapped braid.

It acts as a medium to transport the resin from inlet to outlet, also exhausts the

volatiles, air and moisture. The outermost layer is a nylon bag film that acts as

vacuum membrane. The bagging film is sealed using sealant tape on top of the

mandrel keeping the preform and other infusion aids inside. This acts as an envelope.

Inlet and outlet tube is connected to this envelope to draw resin into the braided

preform and remove air, moisture, solvents and volatiles respectively.

Two resin traps were made by sleeve bagging film enclosed an absorbent non-woven

fabric. One was connected to the outlet infusion tube to absorb excess resin. The

other one had a valve pierced through the bag to connect with vacuum pump.

50
 Valve
(connected
Resin to vacuum
traps pump)

Sealant tape

Preform
covered by
release fabric
and mesh
Outlet
resin Vacuum
tube bagging

Inlet resin
tube

Figure 22 Overbraided 180° bent tubular mandrel under vacuum along with
additional vacuum bagging assembly

Followed by closing all the open tubes for air to the bag film, the valve was

connected to the vacuum pump. The vacuum environment was achieved after

dropping the air out. The process was completed by delivering the resin to the

preform from inlet tube and closing the vacuum trap and inlet tube once the preform

was full permeated in the resin.

Upon completed infusion, the whole system was then put into the oven for curing

cycle duration. Once fully cured all bagging materials were discarded.

3.3.4. Curing and demoulding

Once the resin infusion was completed, the resin infused mandrel was kept in an

oven. The curing and post-curing was carried out at 80°C and 140°C respectively. After

curing cycle had completed, the oven was switched off and temperature was decreased

51
slowly to bring back to the room temperature. Once the part was taken out of the oven,

all the materials added on the preform was removed to separate the composite.

Curing cycle of epoxy resin

160
Temperature (°C)

140
120
100
80
60
40
20
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time (hr)

Temperature Vs Time

Figure 23 Cure cycle of epoxy resin (Araldite LY 564 and Aradur 2954) for
composite manufacturing. The dotted line represents the uncontrolled cooling within
the oven chamber

After all materials and PVC mandrel was removed, the composite tube was cut along

its length. In this study, composite tube braided by using circular braid ring and that

braided by using elliptical braid ring were compared.

52
3.4. Conclusions

Fibre tows were interlaced on mandrel, the shape of the braid was determined on the

shape of mandrel. Flexible hollow tubes were used in this study as this kind of tubes

were easily to be bent in the next experiment. The winding machine could be used to

transfer the fibre from fibre package to 48 bobbins. Braiding process was controlled

by an automatic motion, which was determined by input take-up speed and carriers

rotation speed. Then the desired braid angle could be produced. Afterwards, bending

process that will be discussed in Chapter 5 was followed. After preforms were

produced, all of them were consolidated by using vacuum bagging. Preforms were

sealed in a vacuum bag with release fabric and mesh covered. Resin infusion

happened after resin degassing. This process removed all air bubbles. After curing

and demoulding, composite manufacturing process was completed. In the end, the

tubular composites were cut along its length to see whether wrinkle appeared or not.

53
4. Effect of Linear Compression
on Braided Preforms
4.1. Introduction

In this study, the over-braided mandrels were subjected to bending which resulted in

a tension and compression on the inner and outer bent of the braided sleeve. The

biaxial braid architecture has a phenomenon of fibre shearing while suffering tension

or compression force. When the braid is axially stretched, the mechanism of

deformation is mainly sheared and parallel to each tow [58]. The “Chinese finger

trap” principle was demonstrated in a recent study [59]. When a tension force

applied on the braid, the length increase can result in decreasing of fibre orientation

and subsequently braid diameter increases, vice versa. A ‘scissor effect’ [58] was

presented, which is that as the braids are stretched, the tows scissor like the shape of

a trellis. In this chapter the braid tow scissoring in relation to tension and

compression is presented.

4.2. Prediction of biaxial braid angle under compression/tension

loading

Owing to the fact that compressions at the inner bend causes braid angle increases

while compressing, this also leads to diameter change. In order to find out how the

braid angle and diameter change with compression, relationship among braid angle,

diameter and length could be developed.

54
 (a) (b)

Unit
cell

(c)

(d)

Figure 24 (a) Regular braid structure (b) Considering the braid unit cell acts as a
trellis in which the intersecting four sides are pivoted [60] (c) unit cell schematic (d)
unit cell schematic under compression assuming the sides are pivoted [61]

In this study, a unit cell was regarded as a trellis with a diamond shape.

𝑝𝑝𝑝𝑝 𝐶𝐶 (8)
𝛼𝛼 = tan−1 = tan−1
𝑞𝑞𝑞𝑞 𝐿𝐿

Where: L= Length of a unit cell; C= Width of a unit cell; α =Braid angle

55
In case of a uniform cylindrical braid structure, when α = 45°, the unit cell widths on

all four sides are equal. With equal sides, the diagonals are equal too. Therefore,

pq = qr = rs = sp = a (9)

and pr = qs

So, L=C (10)

Hence, L2 + C2 = 2a2 (11)

When the unit cell suffers a compression force, the diagonal distances pr and qs

changes. For the fibre structure, fibre slippage will happen at the cross over points

hence the position of the intersection will not remain fixed. The slippage can result

in change in fibre structure parameters such as fibre tow width, spacing, etc., which

can influence the unit cell dimension. In addition, due to the interlacement of braid

tow, there will be frictional resistance to fibre shearing as well. In this study it is

assumed that the four sides of the unit cell remain equal during compression or

tension. Also, the four points at the boundary of the unit cell acts as pivots to form a

trellis changing the position under compression or tension without any resistance or

slippage.

Considering the above assumption of pivoted sides, the diagonal distances will

change under compression (Figure 24d), so,

L1 2 + C1 2 = 2a2 (12)

So, from equation 13 and 14,

L2 + C2 = L1 2 + C1 2 (13)

If the compressive strain is ε, the unit cell length after compression L1 can be written

as,

56
 L1 = L (1+ ε) (14)

In case of a regular braid with Nc number of carriers, the number of unit cells in the

circumference,

𝑁𝑁𝑐𝑐 (15)
𝑛𝑛 =
8

So the width of a unit cell,

𝜋𝜋𝜋𝜋 (16)
𝐶𝐶 =
𝑛𝑛

In the above equation, D is the braid diameter.

Similarly, the width of the compressed unit cell,

𝜋𝜋𝐷𝐷1 (17)
𝐶𝐶1 =
𝑛𝑛

The compressed braid diameter can be calculated from equation 12 as follows,

(18)
𝐶𝐶1= ��(L2 + 𝐶𝐶 2 ) − 𝐿𝐿1 2 �

Substituting the L1, C and C1 from equation 16, 17 and 18 into equation 15,

(19)
𝑛𝑛2 𝐿𝐿2
𝐷𝐷1 = � 2 (1 − (1 + 𝜀𝜀)2 ) + 𝐷𝐷2
𝜋𝜋

Under compression, alongside the change in diameter, the braid angle will also

change. As the angle changes to α1 when the length and width changes to L1 and C1

respectively, from equation 10,

𝐶𝐶1 (20)
𝛼𝛼1 = tan−1
𝐿𝐿1

Substituting the L1 and C1 from equation 16 and 18 into equation 22,

57
 𝜋𝜋𝐷𝐷1 (21)
𝛼𝛼1 = tan−1
nL (1 + ε)

In the above equations, D is the initial braid diameter and D1 is the braid diameter

under loading, L is the initial length of the unit cell (Figure 24c) and ε is the

longitudinal strain under loading. The notations, α, α1 and n indicate initial braid

angle, braid angle after deformation and the number of unit cells around the

circumference respectively.

4.3. Linear tension and compression loaded on sleeves

Support base
Ring
Shaft Braids

Compression

force

Figure 25 Device used for linear compressing of braid. Blue arrows are compression
forces

58
In order to find out how the braid parameters change under linear compression, a

braided sleeve was compressed on a mandrel to simulate the situation that happened

on compression curvature during bending.

The device shown in Figure 25 was used in this experiment. The whole device

consists of three parts: supporting base, shaft and a ring that slides onto the shaft in

order to provide necessary displacement. The support works to stop sleeve to move.

The shaft plays a role as a mandrel. Pressing the ring by hand, a linear compressive

force could be applied on the sleeve. During compression loading, braid angle and

diameter were recorded after every 10 mm change.

Figure 26 Images show the linear compression of a braid. The two ends of sleeve
were folded instead of securing with cello tape which could apply a boundary
condition and prevent free movement of fibre duing compression

59
A 200 mm long braided sleeve was placed on the 25.4 mm diameter shaft and

connecting with supporting base and compressing ring Figure 26. Two ends of the

sleeve were folded to prevent fibre tows from getting released. The initial braid angle

was ±45° and outer diameter was 26.67 mm. The sleeve was pressed gradually, a

compressing force was applied on sleeve along its axial direction, which caused

sleeve shorter and increased the braid angle as well as outer diameter. When the

sleeve was compressed to 190 mm, the braid angle and outer diameter could be

recorded. Continue to compress the sleeve to 180 mm, 170 mm, 160 mm and so on

until the braid angle did not increase. The corresponding braid angles and outer

diameters were recorded after each step.

As for linear tension, there is no shaft used for holding. The sleeve was a hollow

structure and the tension force was applied manually. A sleeve of 200 mm length

was placed on the table and taped two ends of the sleeves. Holding two ends of the

sleeve and applying a tension force manually along its axial direction. The braid

angle and diameter data were recorded every 10 mm change in length as well.

Recording the braid angle and outer diameter value after the length of sleeve was

increased to 210 mm. Continue to stretch until the product was stretched to 220 mm,

230 mm and 240 mm, respectively, to record the corresponding braid angle and outer

diameter.

The diameter of 12.7 mm sleeve was measured by using the same way as 25.7 mm

diameter sleeve. The sleeve was compressed and stretched respectively. The braid

angle and outer diameter were recorded after each step.

60
 1 cm

（a)
(a) (b)（b) (c)

Figure 27 Linear compression force applied on a braided sleeve. Initially braid angle
is ±45°. The decrease in length leads to increase in braid angle and braid diameter
after compression

1 cm

(a) (b) (c)

Figure 28 Linear tension force applied on a braided sleeve. Initially braid angle is
±45°. The increase in length leads to braid angle and braid diameter decrease after
compression.

As shown in Figure 27(a) and Figure 28(a), braid angle was ±45° without any

loading. After compression, the decrease in braid length resulted in increasing of

braid angle and subsequently led to braid diameter increase [Figure 27 (b) (c)]. On

the contrary, when a tension force loaded on a braided sleeve [Figure 28(b) (c)],

braid length increase led to braid angle as well as braid diameter decreased.

61
4.4. Measurement

4.4.1. Braid angle measurement

Braid angle was measured by using a digital protractor. This Wixey Digital

Protractor shown in

Figure 29(a) is precise and convenient for measuring braid angle. It measures angles

with ±0.1° resolution accuracy. As shown in the Figure 29(a), the edges of the digital

protractor were aligned to the two adjacent tows helically running along clock and

anti-clockwise.

(a) (b)

Figure 29 (a) Braid angle measurement using digital protractor (b) Braid diameter
measurement using digital calliper

According to the definition of braid angle, the braid angle is half of the angle shown

on screen. In order to improve accuracy in the measurement ten data values were

collected along its length for the same angle measurement.

62
According to the equation 21, braid angle under compression and tension can be

predicted as shown in Figure 30. When a uniaxial compression and tension force

applied on the sleeve, braid angle could be calculated from the relationship between

angle and axial displacement.

Figure 30 Relation between displacement and angle. Braid angle is ±45° when
displacement is 0. Braid angle increases under compression and decreased under
tension until it reaches to locking angle. The dotted red line when axial displacement
is beyond 40 mm means the sleeve cannot be stretched any more

In order to prove this equation, a braid sleeve with 200 mm in length, ±45° of braid

angle was prepared. At this stage, no force is applied on the braids. Then braid angle

63
increased under linear compression and decrease under linear tension. As shown in

Figure 30, axial displacement is 0 while no loading applied on the sleeve. Under

compression, change in length led to displacement change, which result in braid

angle increase, vice versa. However, braid angle rose while the braids reaching to

jammed state. At this stage, braid angle reaching to locking angle. Then further

loading would result in wrinkle formation and the braid angle maintained steady.

4.4.2. Braid diameter measurement

Braid diameter was measured by using digital calliper (Figure 29b). The digital

calliper gives a direct reading of the distance measured with ±0.01mm accuracy.

When measuring the braid diameter, one important thing should be focused on is

measuring without any pressure. Care was taken during the diameter measurement to

prevent applying extra compression by the callipers jaw on the surface of braids.

Also, the braids consist of a lot of interlaced filament fibre tow, which results in an

uneven surface as shown in Figure 31. But the width of the calliper jaw is much

smaller than the width of fibre tow. It made that the jaws of the calliper were easily

touch the gap that between two tows interlaced. In this case, the measured diameter

would be smaller than the real value. It was ensured that the calliper jaws did not

touch on the gap position between fibre tows. The braid diameter measurement was

carried out at the same points as braid angle measurement.

64
 Calliper jaw
2 mm

Figure 31 Gap between interlaced fibre tow. Gap has an impact on diameter
measurement using callipers.

Diameter could be predicted as well. By using equation 19 that came out in Section

4.2, the relation between diameter and axial displacement is shown in Figure 32.

Diameters change under compression and tension were also recorded to prove

diameter equation. Braid diameter increase with the increase of the braid angle. As

soon as the braid angle reaching to the locking angle, braid diameter would not

increase anymore.

The original length was 200 mm, at which time the sleeve was not deformed. When

the compression force applied on the sleeve, the axial displacement increased

nonlinearly. The greater the deformation of the braids, the smaller increase rate of

displacement. Compare the Figure 30 and Figure 32, when the displacement increase

to about 80 mm, the angle increased to locking angle and stopped increasing further.

As for diameter, it stopped increasing at the point that displacement increase to about

80 mm as well. Hence, diameter increased with the increase of braid angle under

compression.

As for tension, diameter decreased with the increase of displacement. According to

the prediction, in the case of stretching, the decrease rate of diameter went faster.

However, it was found in the experiment that when the strain increased to 40, the

65
diameter was not reduced because the density of the fibre tow got to maximum. Also

the increase in braid angle led to cover factor (CF) increase. The maximum CF is 1.

Once the CF increased to the maximum, it could lead to braid being unable to be

stretched.

Figure 32 Relation between displacement and diameter. Braid diameter increases or

decreases under compression or tension respectively until braid angle reaches to
locking angle.

Comparing the change of braid angle and diameter with compressing length, it was

obvious that the simulation curve fits the experimental results very well. So

66
equations are effective to be used to predict the braid angle and diameter under

compression or tension.

4.4.3. Areal density and yarn spacing

A 200 mm long braid sleeve weights about 11.1 g. The initial areal density was

about 663 g/m2. The mass of braid did not change under tension or compression. So

the change in surface area changed the areal density.

Figure 33 Areal density increased with the increase of displacement.

When the braid angle was ±45°, the braids had the lowest areal density. With the

tension or compression force applied on the sleeve, the increase or decrease of the

braid length led to decrease in surface area. Also the tow width changed to smaller

value as the angle increased or decreased. As a result fibre packing increased in a

given surface area. So the areal density increased. However, when the braid angle

reached to locking angle (displacement is -80 mm), further loading could result in

67
wrinkle formation. In that case, the surface area could not be measured or calculated.

Thus areal density could not be calculated.

Figure 34 Way of yarn spacing measurement

Yarn spacing is an important quality characteristic in textiles. It refers to the distance

between two adjacent fibre tows in braids. In this study, yarn spacing was measured

followed the way shown in Figure 34. As 2/2 structures were produced in this study,

the initial cover factor was nearly 1, which means the mandrel was full covered.

When tension or compression load applied on a braided sleeve, cover factor could

not increase anymore. In this case, yarn spacing is nearly the same as tow width. As

tow width decreased under compression, yarn spacing also decreased. However, as

soon as the braid angle reached to locking angle, further loading could not lead to

change in tow width. Thus, yarn spacing did not change as well. On the contrary, the

number of fibre tows remained same. The tow width decreased under tension

because braid diameter reduced. So yarn spacing decreased as well.

68
 Figure 35 Yarn spacing of the braid. Under linear compression and tension, yarn
spacing decreased until braid angle reached to locking angle.

4.5. Conclusions

In this study it is assumed that the unit cell keeps its four sides equal and rotates at

four points through acting as a trellis without major resistance or slippage during

compression or tension until it reaches the lock limit. Then braid angle and braid

diameter under loading can be calculated. In this case, the linear compression

process as well as linear tension process can be used to simulate the braid parameter

(especially braid angle and braid diameter) change happened at compression side and

tension side while bending because braids suffered compression and tension as well

during bending. The braid angle could increase when a compression force loading on

a braided preform. When it increased to a certain extent (braid angle reached to

69
locking angle), it stopped increasing. Meanwhile, braid diameter did not increase any

more until the braid angle reached to locking angle. Then further compression could

result in wrinkle created on braids. On the contrary, during tension loading process,

braid angle and diameter decreased but that reduced until the sleeve could not be

stretched more. The change in braid angle and diameter during compression were

higher than that during tension because when compression load applied on the

preform, more space were created between interlaced fibre tow and the mandrel, but

during tension process, more friction were created between interlaced fibre tow and

the mandrel. The frictional resistance resulted in smaller changes of braid angle

when compared to compression side. Similarly during bending, compression and

tension force applied at inner and outer sides of the curvature. The change of braid

angle observed during linear compression and tension will be compared with the

changes after bending as studied in Chapter 5.

70
5. Effect of Bending on Braid
Parameters
5.1. Introduction

In this chapter, braided tube were bent to 90° and 180° to studied the effect of

bending on braid angle and braid diameter formation. Braid angle and diameter data

were collected from tension side, compression side and neutral axis. Two sizes of

tubes (25.4 mm diameter and 12.7 mm diameter) were bent to illustrate the braid

angle and diameter formation on different size of preforms. Meanwhile, 1 layer

braids and 2 layers braids were bent to find out the effect on braid parameter

formation.

5.2. Relationship among braid angle, diameter and length of

braids

During bending, as the inside radius (Ri), which is also known as bending diameter

radius, (Figure 36) is smaller than that of the outer (Ro) and neutral axis radius (R),

the reduction in length at inner bend leads to compression (Figure 36). The outer

bend is similarly subjected to tension due to the increase in length from the centre

line.

If the degree of bend is θ, braid radius is rb, total braid thickness is tb, Ri is the inner

diameter of the braid (or mandrel dia) and the bending die radius is Rd, the length of

the braided tube at the inner curve (Li), neural axis (L) and outer curve (Lo) can be

calculated as follows.

71
 𝐿𝐿𝑖𝑖 = (𝑅𝑅𝑑𝑑 + 𝑡𝑡𝑏𝑏 ) 𝜃𝜃 (22)

𝐿𝐿 = 𝑅𝑅𝑅𝑅 = (𝑅𝑅𝑑𝑑 + 𝑟𝑟𝑏𝑏 )𝜃𝜃 = (𝑅𝑅𝑑𝑑 + 𝑅𝑅𝑖𝑖 + 𝑡𝑡𝑏𝑏 )𝜃𝜃 (23)

𝐿𝐿𝑜𝑜 = 𝑅𝑅𝑜𝑜 𝜃𝜃 = (𝑅𝑅𝑑𝑑 + 2𝑟𝑟𝑏𝑏 )𝜃𝜃 = (𝑅𝑅𝑑𝑑 + 2(𝑅𝑅𝑖𝑖 + 𝑡𝑡𝑏𝑏 ))𝜃𝜃 (24)

As Ri< R the reduction in length (Li<L<Lo) at inner bend leads to compression on

the braid. The outer bend is similarly subjected to tension due to the increase in

length (Lo>L) from the centre line.

Tension
side

Rd
Ro R
Compression
side

Neutral axis

Figure 36 Schematic of a 90° bent tube [61]

In order to study the effect of bending, initially braiding was carried out using a

circular braid ring and keeping the core at the centre of the ring. This is a standard

practice for preforming using braiding process. The braiding process was executed

following Section 3.2.3. For the ease of bending, PVC tube was used for the over-

braiding purpose. After single layer braiding, the over-braided PVC tubes were bent

72
to 90° and 180° respectively. Then braid angle and diameter measurement were done

around the bending radius of tube.

5.3. Bending machine

A hydraulic pipe bender was used to ensure consistent bending process between

samples. The machine used for bending is Clark hydraulic 12 Ton pipe bender as

shown in Figure 37. This bender consists of a frame, 12-ton bottle jack, hex bolt,

guide roller, hitch pin, bending dies and jack handle. It is a powerfully built, easy to

use machine for bending all manner of steel pipe and tubing up to 180°. This bender

can be used on a variety of pipework installations and for general engineering

applications. A hydraulic pump is fitted within the heavy duty steel frame. Pipes

could be bent into different angle by changing different bending dies. In this project,

only 90° and 180° bending were studied.

Bending dies Guide

ll Hex bolt
for guide

Hydraulic
value
Operating
(a) (b) lever

Figure 37 (a) Top view of a carbon fibre braided core on bending equipment (b)
Hydraulic bending equipment used for this study

73
5.4. Bending process

A 25.4 mm outer diameter tubes and 12.7 mm outer diameter tubes were chosen for

braiding and corresponding dies were used on the bending machine. Before bending,

the original tube was in circular cross-section shape. But after bending, the cross-

section shape deformed from circular shape to an oval shape.

During braiding process, an aluminium tube was inserted into the hollow tube to

ensure the cross section of outer PVC tube was circular. The aluminium tube was

taken off before bending. In order to prevent the circular tube cross-section flattening

to an oval shape during bending, the over-braided PVC tubes were filled with sand.

Before bending

24.3 mm
26.6 mm

27.2 mm

After bending

Figure 38 Cross-section of braided tube changed after bending. ‘Brazier effects’

could not be eliminated by filling the tube with sand but it was improved. Also it was
verified that wet sand had better effect than dry sand.

Once finishing the braiding process, the braided tube was removed from braiding

machine and filled with wet sand. Two ends of the tube were sealed with rubber

bungs to keep the sand inside and stainless steel hose clamp was used to keep the

bungs in place. The clamps were also used to tie a string to keep the tube in bent

74
shape for measurement. Before bending, braid angle and diameter was measured and

recorded in order to compare with the data after bending. The first step of bending

process is opening the hydraulic value by turning it anticlockwise to secure the guide

rollers in place with the ram in its lowest position. When the tube was ready to be

bent, placed the tube on the die. Then the hydraulic value was closed by turning it

clockwise fully and pumps the operating lever gently by using hitch pin. After

bending the tube to 90°, the two ends of the tube were connected with a string, which

will keep the tube with this shape. The hydraulic value could be opened when

completed, thus allowing the product to be removed. The braid angle and outer

diameter of the curve section were measured. The measurement methods were

discussed in Section 4.5.

After the measurements were taken from a 90° bent tube, the tube was placed back

on the die the bending operation was repeated to bend the tube to 180°. The string

length was adjusted to ensure that the tube could be kept in this shape after removing

it from the dia. Then measuring the parameters and recording the values.

Figure 39 Bent tube. Tubes were bent to 90° and 180° respectively

75
5.5. Measurement

All samples were measured in ten points with even distance along the longitudinal of

mandrel. As shown in Figure 40, the inner side is compression side; the outer side is

tension side. The braid angle measurement was carried out at both tension and

compression side and along the neutral axis on both sides. The diameter was

measured on two opposite sides. All data were all presented in Section 4.5.

Tension side

Neutral axis

Compression side

Figure 40 Schematic of a 180° bent tube indicating different bending parameters [61]

The braid angle and diameter were all measured before bending. However, on

curvature section, these values were changed after bending. So data was collected on

the same point at curvature section to find out the braid angle and diameter change

after bending.

5.6. Braid parameter formation

Generally, using a circular braid ring braiding on a cylindrical mandrel, the braid

angles will keep constant on both along the length and circumference direction.

However, when the tube is bent to 90° or 180°, forces were applied to different side

76
of the tube. During bending, one side of the tube goes in compression, which were

called ‘compression side’, the opposite side in tension was called ‘tension side’ and

other two sides were applied neither compression nor tension, so they were called

neutral axis.

Lower braid angle

Tension side

Compression side Higher braid angle

Figure 41 Compression side, tension side and neutral sides at the curvature
section after bending

Because the tube was straight before bending, the lengths of tube along any axis

were same. After bending, the length of braids along neutral axis did not change. As

for compression side, the braid was compressed and become shorter, the braid angle

would increase as observed during the linear compression of braid in the previous

chapter. But for tension side, the braid was stretched and that force made the length

of braid become longer. This resulted in smaller braid angle similar to the

observation of linear tension of braided sleeve.

The bent tube was presented in a curve shape, which indicated that the force applied

on the tube was not evenly distributed along its length. Thus stress irregularity would

eventually cause the braid angle changed gradually from two edges to the centre of

the curve. So the braid angle at the centre of the whole braids has the largest

77
deviation. On the other hand, the length of braids along neutral axis did not change,

so that the braid angle and braid thickness did not change as well.

In general, the braid thickness is constant at any position of the braids. However, it

will only be achieved under standard conditions. Once the tube was bent, horizontal

force was applied to the tube at tension and compression side, which would cause the

braid thickness to be changed. Different braid angle as well as tension that applied

on the fibre tows would result in different thickness. Thus, at compression side,

decreased length of braids caused higher braid angle and increased thickness. But at

tension side, increased length of braids resulted in lower braid angle but reduced

thickness.

5.7. Effect of bending on braided parameters

In order to find out how the braid parameters change before bending and after

bending, braid angle and outer diameter were measured. These parameters were

measured for all the braids produced in the whole experiment at compression side,

tension side and along the two neutral axes before bending and after bending. Due to

the fact that the braid angle produced by braiding machine was controlled by

machine setting, all these samples had same braid angle (±45°) before bending.

Figure 42 shows that along the bending curve a gradual change in braid angle was

observed from the edge of the tangent to the centre of the bending curve. On the

compression side, the maximum angle was observed at the centre of the bending arc

and the decrease in angle on the tension side was observed with the lowest angle

located at the centre of the bending curve. At compression side, the braid angle

78
increased to as high as ±54°. This angle is below the locking angle observed during

linear compression at which the wrinkle appeared.

Tension side

Compression side

Neutral axis

55
53
51
Braid angle (°)

49
47
45 Braid angle before
43 bending (±45°)
41
39
37
35
0 45 90 135 180
Measurement position around the bend (°)
Braid angle at compression side
Braid angle at tension side

Figure 42 Braid angle formation after bending

Two mandrels of 25.4 mm diameter and two mandrels of 12.7 mm diameter were

chosen to be braided. The mandrel details are shown in Table 2.

79
 Table 2 Outlines the details of tube braiding

Tube number 1 2 3 4

Mandrel Diameter 25.4 mm 25.4 mm 12.7 mm 12.7 mm

Number of layers 1 2 1 2

As shown in Table 2, #1 and #3 tubes had single layer, #2 and #4 tubes has double

layers. When compared braid angle formation on #1 and #2 tubes or compared the

braid angle formation on #3 and #4 tubes, the effect of multiple layers on braided

parameters could be found out.

Figure 43 Schematic shows braid length change under compression of a multilayer

braid

As shown in Figure 44 and Figure 45, the change of braid angle on different

diameter tubes followed the same trend for single layer. In both cases, two layers

were braided to observe the effect of bending on a case of multilayer preform. The

change in braid angle was significantly higher for a two-layer braid than a single

layer braid. In a two layer braid, in compression side, the outer (i.e., second) layer

80
bending radius is lower than the first layer. As a result, the braid length under

compression (Li) is also lower than that of the first layer braid. This can be

calculated by using equation 22.

For a 2 layer braid, the length of 2nd layer braid under compression will be as follows

where tn indicates the thickness of the nth layer.

𝐿𝐿𝑖𝑖 = (𝑅𝑅𝑑𝑑 + 𝑡𝑡𝑛𝑛 ) 𝜃𝜃 = (𝑅𝑅𝑑𝑑 + 𝑡𝑡2 ) 𝜃𝜃 (25)

Similarly, for the 2 layer braid, the length of 1st layer braid under compression will

be as follows where tb indicates the total braid thickness as shown in the Figure 43.

𝐿𝐿𝑖𝑖 = (𝑅𝑅𝑑𝑑 + 𝑡𝑡𝑏𝑏 ) 𝜃𝜃 = (𝑅𝑅𝑑𝑑 + 𝑡𝑡1+2 ) 𝜃𝜃 (26)

Table 3 Length change under compression on a multilayer braid

Braid Braid at Braid length Compressed

Bending Degree of
radius neutral under length from
die radius bending
(rb) axis (L) compressio neutral axis
(Rd) (mm) (θ)
(mm) (mm) n (Li) (mm) (mm)
Layer
60 13.34 180° (π) 232.7 192.9 39.8
1
Layer
60 14.12 180° (π) 232.7 190.8 41.9
2

The linear compression experiment and prediction shows that higher the braid length

under compression, higher the change in angle. Since the above table shows that the

braid length change under compression for second layer is higher than that of the

first layer, the change in angle should be higher. This was also observed in Figure 44

and Figure 45.

81
The change in angle was only recorded around the bending curve and the effect was

negligible on the tangent sides. This observation indicates that the fibre shearing

under tension and compression mostly accommodates the change in parameters

within the bending curve region.

60

55

50
Braid angle (°)

45

40

35
0 45 90 135 180
Braid angle along compression side (single layer)
Braid angle along tension side (single layer)
Braid angle along compression side (double layers)
Braid angle along tension side (double layers)

Figure 44 After bending a preform braided on a 25.4 mm diameter mandrel, at

compression side, braid angle increased and reached to the highest angle at the centre
of the curvature, and vice versa. The braid angle of double layers braids increased
more rapidly than single layer braids at compression side but decreased slower than
single layer braids.

82
 60

55

50
Braid angle (°)

45

40

35
0 45 90 135 180
Braid angle along compression side (single layer)
Braid angle along tension side (single layer)
Braid angle along compression side (double layers)
Braid angle along tension side (double layers)

Figure 45 When a preform braided on a 12.7 mm diameter mandrel was bent, the
change of braid angle was similar to 25.4 mm diameter mandrel bent. But braid
angle changed more gradually from edges to the centre.

The data collected from one 25.4 mm diameter single layer braided mandrel (#1) was

shown in Figure 44. Before bending, braid angle was ±45° at any point on the tube.

After bending to 90°, the average braid angle at compression increased and that at

tension side decreased. After bending to 180°, the greater degree of bending made

83
the angle change more obvious. The minimum braid angle referred to braid angle at

the centre of the tension side. Similarly, the maximum braid angle means the braid

angle at the centre of compression side.

53

51

49
Braid angle (°)

47

45

43

41

39
Tension side (°) Compression side (°)

Braid angle (Bending to 90°) Braid angle (Bending to 180°)

Predicted braid angle after bending to 90°

Predicted braid angle after bending to 180°

Braid angle before bending

Figure 46 Figure showing the maximum braid angle at compression side and
minimum braid angle at tension side for a single layer braid after bending at 90° and
180°. The braid angles are collected from the centre of the curvature. Predicted braid
angle were predicted by using equation 21 illustrated in Chapter 4.2

84
5.8. Conclusions

Preforms could be bent to desired shape before composite manufacturing. During the

bending process two types of forces are applied on the braided tube- tension and

compression. Braid angle was decreased after bending on tension side of the braided

tube. On the other hand, the compression load led to braid angle increasing. The

change in braid angle and braid diameter after bending was similarly as linear

compression and tension applied on sleeves. So the braid angle and braid diameter

could be predicted by equations in Chapter 4. The predicted angle and diameter did

not exactly match the experimental results because the equation was built based on

the assumption that the braid unit cell acts as a trellis. In fact, fibre tows slippage

happened during fibre shearing as well as friction resistance played a role in the

change.

After bending, at tension side, braid angle and braid diameter decreased gradually

from two edges to the centre. The lowest value was collected at the centre of the bent

arc. As for compression side, the highest braid angle and diameter were recorded at

the centre of the curvature. The change in braid angle changed other parameters such

as tow width, length of braid unit cell which eventually changed the local fibre

density. that the increase in local fibre density and bulk as well as braid diameter can

result in wrinkle formation at compression side as noticed in the microscopic

analysis of the composite tube cross-section. To prevent such increase in local fibre

density due to increase in braid angle, methods that could generate a suitable braid

angle will be discussed in Chapter 6.

85
6. Methods for Local Braid Angle
Change
6.1. Introduction

As braid angle change was observed after bending as analysed in Chapter 5, in order

to accommodate the change of braid angle that occurs during bending, a local change

in braid angle is required. In the compression side the angle was increased from that

of the neutral axis. To prevent increase in angle thus increase in thickness, before

bending a lower angle than neutral axis is required at compression side. By changing

the take-up speed or rotational speed of the process, change in braid angle is possible

to achieve. However, this change in angle only appears along the length of the

mandrel and not localised to compression side. The effect of bending on braid

parameters showed that the change in braid angle occurs in the circumferential

direction of the mandrel. Hence, an alternative approach of using elliptical ring and

off-centre braiding were taken to achieve local change in braid angle in the

circumferential direction.

6.2. Change in shape of braid ring

An Elliptical braid ring was used in this study to investigate the effect of elliptical

braid ring on braid parameters as well as change the braid angle at compression side

to prevent the increase in braid thickness and angle after bending. By using both a

circular and an elliptical braid ring ±45° braid structures were produced.

86
 150 mm

300 mm

Figure 47 Image of elliptical braid ring

Use of circular braid ring is a universally used approach for braiding process.

Generally, the mandrel was mounted at the centre of the braiding machine. The

standard circular braid ring can ensure that the distance from bobbins to the point on

the braid ring and from braid ring point to fell point are all constant, so that the braid

angle could keep constant in all directions on the mandrel. In this case, the changes

in rotation speed as well as take-up speed could only change the fibre orientation

along longitudinal direction. But the fibre orientation remains the same around

circumferential direction. Hence, if the braid ring is not circular shape, then the

distance from the point on the braid ring to fell point will be changed, which will

result in different braid angle around the circumferential direction. In order to

achieve that, non-circular braid ring- elliptical braid ring was used in this study.

Elliptical braid ring has two minor diameters and two major diameters on opposite

sides. The gradual change in diameter resulted in gradual change in braid angle

around the circumference direction. Elliptical braid ring that was used in this study

87
has 300 mm in major diameter and 150 mm in minor diameter as shown in Figure 47.

The circular braid ring with 300 mm diameter was also used for comparison.

6.2.1. Braiding

The circular braid ring was mounted at the centre of the machine and the

convergence length, rotation speed and take-up speed are shown in the Table 4. Once

braiding was finished, the elliptical braid ring was mounted on the braiding machine

(Figure 48).

Table 4 Braiding process parameters

Circular braid Elliptical braid

ring ring

Convergence length (mm) 100 75

Take up speed, v (machine setting) (mm/s) 8 9.8

Distance between ring and eyelet (mm) 65 100

Braid angle located along the vertical plane

of the ring (along minor dia of the elliptical ±45 ±45
ring) (°)

Braid angle located along the horizontal

plane of the ring (along major dia of the ±45 ±41
elliptical ring) (°)

The elliptical braid ring was set at the same position as the circular braid ring.

According to the braid angle calculation equation, in order to create ±45° structure

braid, bobbins rotation speed and take-up speed could be input into the system as

following (Table 4).

88
 (a)

Circular
Elliptical
shape core
braid ring

(b)

α>β
(c)
α

β β
α

Figure 48 (a) Braid machine with elliptical braid ring (b) Schematic of a front view
of the braiding machine (c) Two sets of braid angle were produced on four sides of
the mandrel with the same angle in opposing sides

The braid angle of ±45° could be located along two minor diameters sides. The braid

angle at two major diameter sides was less than ±45°. Therefore, major diameter

corresponds to small braid angle, and vice versa. By using an elliptical braid ring

(Figure 48) different braid angles at its major (β) and minor diameter (α) was

89
produced where β< α. The braid angle appears to be the same at opposing quadrants

around the core (Table 5).

Table 5 Braid angle are the same at opposing quadrants around the mandrel

Position around mandrel (°) Braid angle (°)

0 ±41 (β)

90 ±45 (α)

180 ±41 (β)

270 ±45 (α)

By using elliptical braid ring rather than standard ring in braiding process, braid

angle formation before bending were shown in Table 5. Braid angle at two opposite

sides of the mandrel kept ±45°. But at the other two sides, braid angle became lower,

only ±41°. It is caused by change of radius of braid ring.

After braiding, take-off the mandrels and put them on the die of the bending machine

for bending. The same bending process was used here to bend the tube to 90° and

180°. Owing to the fact that the elliptical braid ring have two major diameters and

two minor diameters, which resulted in different braid angle around the

circumference. Besides, after bending process, the braid angle at compression side

increases and angle decreases at the tension side, the angle at the other two sides

keep constant. So one of small braid angle side was placed at the compression side

and the other small braid angle side was tension side during bending. Hence, braid

angle of ±45° was kept constant after bending. One side of large braid angle could

increase to nearly ±45°. Thus, braid angle at three sides of the tube could be same.

90
6.2.2. Limitation of elliptical braid ring in maypole braider

Braid angle has a wide range while braiding using circular braid ring. However,

when braided using elliptical braid ring, high angle braid could not be braided

because the convergence length for a high angle braid was too small to be in contact

with the braid ring. This is a limitation in maypole braider. So during braiding

process, elliptical braid ring will play a role in locally braid angle change only when

all of the fibre tows touch the braid ring.

Table 6 Effect of convergence length on braid feasibility

Distance from fell point to

Braid angle (°) Feasibility
eyelet (mm)
±30° 225 Yes

±45° 200 Yes

±60° 135 No

6.3. Change in the position of the core

In standard braiding machine setting, the mandrel is set at the centre of the machine

to ensure the position of mandrel would not affect the fibre orientation. However,

this study aims to change the fibre orientation. So changing the position of the

mandrel would definitely change the braid angle formation. In order to set the

mandrel away from the centre, an aluminium plate with some holes (Figure 49) was

used in this study. The mandrel could be set even distance from the centre.

91
 12.5 mm
25 mm

Centre

Figure 49 Image of plate used for mandrel off-centre setting

90° Braid
180° ring
0°

270°
Mandrel

Figure 50 Schematic of the front view of a braiding machine. Horizontal plane off-
centre location of the mandrel is shown with respect to braid ring. Four measurement
positions on the mandrel are shown in anti-clockwise direction.

In this part of the study, mandrels were mounted on the take-up device at off-centre

position of the whole machine and four points of mandrel are pointed out in Figure

92
50. The braiding process was following the process that was discussed in section

3.2.3.

6.3.1. Relation between carrier compensation length and off-centre

braiding

As it was illustrated in Section 2.5 that when the carrier moves from the furthest

point to the nearest point, an extra length was created during braiding which led to

extra fibre tow on uneven braid surface. In the literature [34], the compensation

length was studied and the length from eyelet to braid ring could be calculated by

equation 7. In this study, more compensation length was needed because an extra

length was created by braid ring off-centre setting. The difference in length from

braid ring to fell point could also be calculated as follow.

Bobbins Centre line

Braid ring

P1 c
P2

f
Mandrel

Figure 51 Schematic of circumferential braiding compensation length variation

between braid ring and braid point. (left) side view. (right) top view

ΔP = 𝑃𝑃2 − 𝑃𝑃1 = �(𝑅𝑅 + 𝑓𝑓)2 + 𝑐𝑐 2 − �(𝑅𝑅 − 𝑓𝑓)2 + 𝑐𝑐 2 (27)

93
Where ΔP is the difference between P1 and P2; P1 is the nearest distance from braid

ring contact point to fell point; P2 is the furthest distance from braid ring contact

point to fell point; R is the radius of braid ring; f is the horizontal distance from

mandrel centre line to machine centre line; c is convergence length.

Owing to the fact that the total compensation length required during braiding is the

maximum differences from eyelet to fell point, so

ΔC = Δlc + ΔP (28)

Where ΔC is the compensation length that required during braiding..

Substituting equation (7) and (27) into equation (28) yields the maximum

compensation length.

∆C = �(h1 + h2 )2 + d2 − �h1 2 + d2 + �(R + f)2 + c 2 − �(R − f)2 + c 2 (29)

6.3.2. Limitation of off-centre braiding

The parameters h1, h2, d, R, f and c was measured by using measuring tape.

Substituting h1, h2, d, R, f and c into equation (29), the compensation length from

eyelet to braid ring (Δlc) and distance from braid ring to braid point (ΔP ) can be

calculated. Also, the total compensation length that is required during braiding (ΔC)

can be calculated (shown in Figure 52).

94
Figure 52 Factors affecting the required compensation length during braiding. (a)
Original setting is setting the distance between a braid ring and the eyelet to 100 mm
with braid ring radius 100 mm and convergence length 110 mm. (b) (c) (d) When the
mandrel was set 50 mm off-centre, the increase in the distance between braid ring
and carrier eyelet, braid ring radius and convergence length (lower braid angle)
resulted in more compensation length was required.

As it was shown in Figure 52, the distance from mandrel centre line to machine

centre line (f), the horizontal distance from braid ring to eyelet (d), braid ring radius

(R) and convergence length were all played an important role on compensation

length required during braiding process. Equation 29 and Figure 52 showed the

relation between compensation length and these factors. As h1 and h2 were

unchangeable, the compensation length (ΔC) was proportional to the distance from

mandrel centre line to machine centre line (f) and braid ring radius (R) but was

95
inversely proportional to the distance from braid ring to eyelet (d) and convergence

length (c).

The maximum compensation length that maypole braiding machine used in this

study could give was measured from the machine, which was 210 mm. Keeping the

fibre tow on the compensator roller, fibre tow was pulled until the bobbin started to

rotate. This was the point at the compensator roller reached to its maximum limit.

The length of the fibre tow that was pulling out is the maximum compensation

length of the maypole braiding machine. So ΔC should be no more than 210 mm.

6.3.3. Braiding with off-centre setting

In this study, a circular braid ring with 150 mm radius was set 100 mm away from

the eyelet. Convergence length was 110 mm to produce braid structure of ±45° braid

angle. According to Figure 52, the maximum distance that the mandrel could move

away from the centre was 50 mm. Because of the limitation of maypole braiding

machine’s compensation length, when mandrel was set more than 50 mm away from

the centre, fibre tows were loose and that led to braid angle distortion. On the other

hands, structure of ±60° could not be produced because fibre tow did not touch the

braid ring. In this study, the mandrel was set 6 mm, 12.5 mm, 25 mm, 37.5 mm, 50

mm, 62.5 mm and 100 mm away from the centre. The setting up data was shown in

the Table 7.

96
 Table 7 Data setting for off-centre braiding

Setup for
Mandrel axis off Actual convergence
Braid angle Feasibility
set (mm) length (mm)
(°)
6 ±45 110 Yes

12.5 ±45 110 Yes

25 ±45 110 Yes

37.5 ±45 110 Yes

50 ±45 110 Yes

62.5 ±45 110 Yes

100 ±45 N/A No

When the mandrel moved from centre to 6 mm, 12.5 mm, 25 mm, 37.5 mm and 50

mm away from the machine centre, the required compensation length was less than

210 mm. However, when the mandrel was 62.5 mm away from the machine centre,

the required compensation length was 220 mm, which was 10 mm higher than

machine’s compensation length. This small difference was not reflected in the braid.

But fibre slack was observed during braiding (shown in Figure 53)

Figure 53 Fibre loosed when mandrel was set 62.5 mm away from the machine
centre

97
 (a) (b) (c)

(d)

(e)

Figure 54 Effect of compensation limitation in braiding process. Mandrel was set

100 mm off-centre (a) front view of fibre slack (b) fibre slack during braiding (c)
fibre hoop was created because of slack fibre tow (d) convergence length various due
to fell point moving (e) fibre tow out of guide causing fibre damage

In order to see the effect of limitation of compensation length on braids, the mandrel

was set 100 mm away from the machine centre. When the required compensation

length was 64 mm, which was longer than machine’s ability, fibre hoop was created,

fell point moved and severe fibre loose was observed (Figure 54).

6.3.4. Effect of core position on braid parameters

Fibre orientation could become variety in terms of the position of the core.

Considering the mandrel was moved towards 180° direction, the distances from the

98
fell point to the contact point between fibre tow and braid ring were different at four

directions. At 180°, it was smaller than that at 90° or 270° which were equal. But for

0°, it had the longest distance.

55

50
Braid angle (°)

45

40

35
0° 180°
Position of mandrel

Before bending Bending to 90° Bending to 180°

Figure 55 Comparing braid angle (maximum and minimum) before bending and
after bending to 180° (12.5 mm offset). The dotted line indicates braid angle at the
neutral axis.

As shown in Figure 55, before bending, the smallest distance corresponds to the

maximum angle and vice versa. The braid angle at 90° and 270° were ±45°. The

braid angle at 0° was ±42°, but at 180°, it was ±47°. Bending process could be

carried out by keeping 90° and 270° as neutral axis. In this case, ±45° would keep

constant. Based on the effect of bending on braid angle distribution as discussed in

Chapter 5, 180° side should be kept under tension side so that braid angle could

99
decrease to ±42° and then reduce the chance to form wrinkle. At 0°, the lowest angle

could increase to ±49°.

Table 8 showed braid angle change after bending when mandrel moved 12.5 mm, 25

mm, 37.5 mm and 50 mm away from the machine centre respectively. Before

bending, braid angle at 0° slightly reduced but slightly increased at 180° while

mandrel moving further. But after bending, braid angle changed more at 0° but

moved less at 180° when mandrel moved further. Besides, braid angle changed more

at compression side than tension side.

100
 Table 8 Change of braid angle (°) after bending to 180

Change in angle after Change in angle after

12.5 mm off centre Before bending After bending to 90° After bending to 180°
bending bending
Braid angle at 0° 42 46 4 49 7
Braid angle at 180° 47 42 -5 40 -7
Change in angle after Change in angle after
25 mm off centre Before bending After bending to 90° After bending to 180°
bending bending
Braid angle at 0° 41 47 6 49 8
Braid angle at 180° 48 44 -4 42 -6
Change in angle after Change in angle after
37.5 mm off centre Before bending After bending to 90° After bending to 180°
bending bending
Braid angle at 0° 41 48 7 50 9
Braid angle at 180° 48 46 -2 44 -4
Change in angle after Change in angle after
50 mm off centre Before bending After bending to 90° After bending to 180°
bending bending
Braid angle at 0° 41 48 7 50 9
Braid angle at 180° 49 47 -2 45 -4

101
6.4. Conclusions

Elliptical braid ring could achieve the braid angle vary around circumferential

direction. Braid angles were the same on adjacent sides. After bending, one side with

lower braid angle increased to nearly the same as the angle on neutral axis. In that

case, braid angle at three sides became same or nearly same after bending, especially

preventing the braid angle to increase more than that of neutral axis in the

compression side. However, the braid angle on opposite side became lower than the

angle braided. Also, instead of producing a ±45°, if a larger braid angle is required

using a maypole braider, the elliptical ring concept may not be useful. This is due to

the small convergence length which will result in the fibres not touching the braid

ring and thus no effect of ellipticity will be reflected on the braid angle.

The mandrel was set off-centre only at horizontal plane of the circular braid ring and

braid angle at two opposite sides were same. The other two opposite sides had two

different braid angles unlike the elliptical ring study. The angles were lower and

higher than that of the other set of the same braid angles. During bending, the highest

braid angle was kept at tension side and the lowest braid angle was kept at

compression side. The angle at compression and tension side changed to nearly the

same as neutral axis for certain off-centre distance and degree of bending. However

in most cases the change in angle was larger than that at the neutral axis which is not

suitable to consider bending as the local fibre density will change similarly like a

concentric braided mandel leaving the possibility of wrinkle formation. Thus,

appropriate selection of off-centre distance and elliptical ring diameter can

102
effectively achieve desired braid angle locally to minimise prospect of wrinkle

formation.

In addition, during off-centre braiding, the limitation of carrier compensation length

was studied. The total compensation length could be calculated so that the braid

feasibility could be predicted. In this study, the maximum distance between mandrel

centre line and machine centre line was 50 mm. The increase in distance from

mandrel centre line to machine centre line and the braid ring radius resulted in an

increase in compensation length. On the other hand, the compensation length could

also increase when braid ring was close to the machine and reduced the convergence

length. According to the equation, it could be predicted that when the mandrel was

62.5 mm off centre setting, the required compensation length was higher than that of

the carrier capacity. If the mandrel was moved further, several slack fibres were

observed which eventually damaged some fibre tows. Thus, before selecting an off-

centre distance the total compensation length should be predicted to keep it within

the carrier’s capacity to prevent fibre damage.

103
7. Analysis and Discussions

7.1. Effect of linear compression and linear tension on braided

preforms

The experiments in this study shows that if a braided mandrel is subjected to bending,

change in braid geometry around the cross-section of the bent section was observed.

After bending, the reduction of length at inner curvature leads to increasing on braid

angle and diameter. The length increase under tension leads to braid angle and

diameter decrease. Linear compression and linear tension have similar effect as

bending. So earlier in this study, the effect of linear compression and linear tension

were studied in order to find out the relationship between the change in length and

the change in braid angle as well as diameter under loading. It was found that when

initial braid angle was ±45°, the diameter was 26.6 mm. As the compression force

was applied further, the locking braid angle was reached and it was ±66°. Beyond

this point, wrinkle formation was observed. But in standard braiding case, braid

angle could reach to ±51° maximum. So wrinkle was not observed at this stage

however, during composite manufacturing, under vacuum condition, due to the

consolidation force, the wrinkle was formed at compression side where the braid

angle was ±51°. Based on linear compression and tension equation, the decrease in

length resulted in the increase in braid angle. When the braided tube was bent to

180°, the length of inner arc and outer arc were 559 mm and 641 mm respectively

which changed from the neutral axis length of 600 mm. By using the equation,

104
the predicted braid angles were ±49° and ±41° at compression and tension side

respectively. However, the experimental value discussed in Chapter 5 shows that the

maximum braid angle at compression side was ±51° and the minimum braid angle at

tension side was ±42°. The difference between predicted braid angle and

experimental angle was because the equation was derived based on the simulation

that the four sides of the unit cell remain equal during compression or tension. Also,

the four points at the boundary of the unit cell acts as pivots to form a trellis

changing the position under compression or tension without any resistance or

slippage. The experimental results matched to predicted value quite well.

7.2. Effect of bending on braided preforms

Then standard bending (mandrels were braided by using circular braid ring) were

carried out to investigate how the braid angle change after bending at compression

side, tension side and neutral axis. In order to accommodate the change in the braid

angle occurring during the bending, a local variation of the braid angle is required.

Change in braid angle is possible to be achieved by changing the rotational speed of

carriers or linear speed of the process. However this change in angle only appears

along the length of the mandrel. According to the effect of bending on the braiding

parameters, it indicates that the change in the braid angle occurs in the

circumferential direction of the core. Hence, a non-circular braid ring is an

effectively method to change the braid angle locally around its circumference

direction. Thus, in order to change the braid angle locally, two methods were used to

achieve that. One is using a non-circular braid ring instead of the circular braid ring.

In this study, elliptical braid ring was used to change the fibre orientation around the

105
circumference direction. The other one is changing the position of the mandrel by

moving the mandrel away from the centre.

Bending was carried out by keeping the quadrants with ±45° braid angle along the

neutral axis. After bending, braid angle at neutral axis remained the same. After

mandrel bending to 90°, braid angle at tension side decreased but increased at

compression side. However, at neutral axis, braid angle kept constant because there

is no force applied along neutral axis. As for bending to 180°, braid angle changed

more than before.

47

45

43
Braid angle (°)

41

39

37

35

33
Tension side (°) Compression side (°)

Before bending Bending to 90° Bending to 180°

Braid angle at neutral axis

Figure 56 Change in braid angle (maximum and minimum) after bending (elliptical
braid ring)

The braid angle at compression side changed from ±40° to about ±45°. This change

in angle led the braid angle at compression side to be the same as that at the neutral

106
axis. This change in angle to ±45° at compression side is desirable as it will prevent

the increase in braid density by increasing braid angle and eventually prevent the

possibility wrinkle generation during vacuum bagging.

Table 9 The maximum braid angle (°) at compression side after bending

After bending After bending

Before bending
Ring type (90°) (180°)
Compression side Compression side

Circular ring 45 51 54

Elliptical ring 41 44 45

The maximum angle at the centre of the compression curvature was recorded in the

Table 9. This change in angle to ±45° at compression side is desirable as it will

prevent the increase in braid density by increasing braid angle and eventually prevent

the possibility wrinkle generation during vacuum bagging.

7.3. Effect of shape of braid ring (elliptical braid ring)

In tension side, as braid angle decreased after bending, which would not lead to

wrinkle forming, only change at compression side were taken into account. After

bending to 180°, while braiding by using circular braid ring, braid angle increase

from ±45° to nearly ±52°. Wrinkle formation was observed after resin infusion. As

for braiding using elliptical braid ring, the initial angle was ±41°. After bending, it

reached to nearly ±45°. As the larger the angle, the greater the likelihood of the

formation of wrinkles, elliptical ring reduced the chance to form wrinkle.

107
7.4. Effect of off-centre setting of the braid ring

Changing the shape of braid ring is an effective method to change the braid angle

locally. But after bending, braid angle at tension side, which was lower than braid

angle at neutral axis, became even lower. Hence, another method of changing the

position of mandrel (offset) can be used to optimise the braid angle locally as well.

In this way, the braid angle at the tension side can be developed to be higher than

that at neutral axis. As a result, under tension it can decrease to the angle similar to

that of neutral axis avoiding major change as observed with elliptical or circular ring

braid after bending.

However, in most cases, the change in angle was greater than that at the neutral axis,

which is not suitable for bending due to the local fibre density will change similarly

like a concentric braided mandel leaving the possibility of wrinkle formation. Thus,

appropriate selection of off-centre can effectively achieve desired braid angle locally,

thereby minimising prospect of wrinkle formation.

7.5. Wrinkle formation after resin infusion

The change in fibre orientation was observed when linear over-braided cylindrical

cores were subjected to be bent to 90° and 180°. Braid parameters changed due to

tension and compression force applied on the bent section during bending. Increase

in braid angle and increase in thickness can potentially lead to wrinkle generation.

As discussed above, elliptical braid ring can effectively change the braid angle

around the circumference direction. Braid angle at inner bent section is smaller than

the angle at neutral axis. Thus, after bending, braid angle increased, but it was still

108
lower than preform braided by using circular braid ring. Different braid angle had an

effect on thickness. Figure 57 shows wrinkles were eliminated by braiding using

elliptical braid ring instead of circular braid ring. Table 10 and Table 11 showed that

higher braid angle led to thicker wall thickness. Once the braid angle increased to a

certain degree (e.g. locking angle), wrinkle was created.

(a)

mm

(b)

mm

Figure 57 Half piece of composite tubes for observation of wall thickness (a) Wall
thickness variation with small wrinkles are observed for composite, the preform of
this composite was braided by using circular braid ring where the braid angle
changed at the compression side to ±54° from ±45° at neutral axis. (b) Uniform wall
thickness at compression side of the tube and the preform of this tube was braided by
using elliptical braid ring where the angle changed to ±45° at compression side after
bending from ±41° before bending

109
Table 10 The minimum braid angle at tension and the maximum braid angle at
compression side after bending to 180°

After bending
After bending (180°)
Ring type (180°) Along neutral axis
Tension side Compression side

Circular ring 38 54 44.9

Elliptical ring 36 45 45

Table 11 Mean average thickness (mm) at tension and compression side after resin
infusion

After bending
After bending (180°)
Ring type (180°) Along neutral axis
Tension side Compression side
0.86
Circular ring 0.80 1.27

Elliptical ring 0.69 0.85 0.75

The thickness showed a large variation on compression side because of wrinkle with

small amplitudes. As shown in Figure 58(a), the maximum wrinkle amplitude is 1.82

mm. But no wrinkle was shown in Figure 58(b), the maximum thickness is 0.87 mm.

Therefore, wrinkle was eliminated by using elliptical braid ring instead of circular

braid ring.

110
 (a)

(b)

1 cm

Figure 58 Cross section of inner bent section. (a) Preform braided by using circular
braid ring. (b) Preform braided by using elliptical braid ring

Braid angle could be changed locally by using these two methods. As a result, after

bending, the increase in braid angle would be no longer reach to locking angle.

Hence, wrinkle formation could be eliminated.

7.6. Conclusions

This chapter illustrated the summarised results and analysis. Change in shape

(bending) resulted in braid angle formation change around circumferential direction.

Tension and compression force were applied on two opposite sides of tube. As

higher braid angle caused wrinkle, two methods were used to optimise the braid

angle. Elliptical braid ring and mandrel off-centre setting were used in this study to

potentially mitigate wrinkle formation. Two different braid angles were produced on

two adjacent sides of a mandrel by using elliptical braid ring. After bending, the

111
braid angle at neutral axis kept constant. On compression side, the braid angle

increased to be close to the braid angle at neutral side. However, the lower braid

angle at tension became lower than the initial angle. So another method of setting the

mandrel off-centre was studied using circular ring. In this case, three different braid

angles were produced around a manrel circumference after braiding. Two of these

angles were ±45° and these were kept at the neutral axis during bending. The other

two angles was higher and lower than ±45°. The lower braid angle and higher braid

angle were kept at compression side and tension side respectively. Thus after

bending at 90°, the lower braid angle and the higher braid angle tended to be as the

same as neutral axis for 25 mm offset. Then wrinkle could be prevented being

forming after bending. As in other off centre settings, the change in angle was larger

than that at the neutral axis, the off-centre distance needs to be selected carefully to

prevent the angle change higher than desired.

112
8. Conclusions and Future Work

Composites manufactured using braided preform can generate defects due to out of

plane deformations such as wrinkles during the vacuum assisted resin infusion

methods. Fibre orientation, which is a very important parameter for composite

mechanical properties, can change due to the shape change. This study introduced

two methods to change the braid angle on preforms to lead to an improvement of

wrinkle formation on bent composite tube. Wrinkles can get generated in composite

tubes around its circumferential direction while forming of complex shapes with a

high degree of bending. Because the curvature section is subjected to tension and

compression during bending, this influence the braid angle to change by fibre

shearing.

Considering the deformation of a braid unit cell is similar to that of a trellis, a

method for predicting the angle changes was presented. Braid angles and diameter

data were collected after braid being loaded under uniaxial tension as well as

compression. Predicted braid angle showed a good correlation with the practical

measurement until the shear locking limit was reached. Once the braid angle reached

to shear locking limit, no change in the braid angle and diameter was recorded with

additional force and deformation.

To study the effect of bending on the braided parameters, over-braided cores were

bent to 90° and 180°. The minimum and maximum braid angle values at the curve

centre were increased and decreased at the compression side (inner bend) and tension

side (outer bend) respectively. As no load applied along the neutral axis, the braid

angle remained constant. Along the bending curve a gradual change in braid angle

113
was observed from the edge of the tangent to the centre of the bending curve. On the

compression side, the maximum angle was observed at the centre of the bending arc

and the decrease in angle on the tension side was observed with the lowest angle

located at the centre of the bending curve. In compression side, the increase in braid

angle of outer layer was higher than first layer because more decrease in length result

in higher braid angle. On the contrary, in tension side, the braid angle after bending

showed a more significant increase at first layer. It is probably because the increase

in tension at outer layer eventually increases the friction resistance, which is likely to

play a role in smaller change in braid angle.

Localised change in braid angle around the circumference was achieved by using an

elliptical braid ring. The elliptical ring generated two different braid angles along

major and minor axis and the angles at the opposing quadrants of the braid

circumference were equal. Based on the change regulation in braid angle, during

bending at compression side, braiding was carried out to produce ±45° and ±41°

using the elliptical ring. After bending, braid angle at neutral axis remain same and

hence ±45° angles were kept at the neutral axis. The braid angle at compression side

changed from ±41° to about ±45°. This change in angle leads the braid angle at

compression side to be the same as that at the neutral axis. The local angle variation

improves the angular consistency around the circumference as compared to the braid

produced using a circular ring. Circular ring generated the same angle around the

core and the after bending the angle changed to about ±42° and ±51° at tension and

compression side respectively. Thus, the occurrence of wrinkle formation can be

minimised or even eliminated by changing the angles locally and using the

advantages of the deformation of parameters of the braid.

114
Localised braid angle change in circumferential direction was also completed by off-

centre setting of the braid ring. The difference in distance from eyelet to fell point

led to various distribution of braid angle. Two opposite sides with same braid angle

(±45°) were kept at neutral axis. Unlike the effect of elliptical braid ring, other two

opposite sides on the mandrel had two different braid angles. One side had a higher

and the other one side had a lower braid angle than ±45°. The highest braid angle

side was kept as tension side so that it reduced after bending. Similarly, the lowest

braid angle was kept at compression side and the angle increased under compression

load. Thus, braid angle tended to be same after bending for 25 mm offset and for 90°

bending. The change in braid angle was not significant for different off-centre

distances however the change in braid angle after bending was noticeable. Between

the specimens the braid density (tows per cm) was observed to be different and this

difference contributed to the change in angle after bending.

The compensation length is an important factor that should be taken into account for

off-centre braiding. The total required compensation length should not be more than

the compensation length that could be provided by machine. The mandrel could not

be moved further away from the centre because of the limitation of compensation

length.

Some further studies that could be done in the future are as follows:

1. Multi-layers preforms could be braided to investigate the effect of the

number of layers on wrinkle formation.

2. Composite mechanical testing could be done to study the impact of wrinkle

on mechanical properties as well as effect of change in braid angle on overall

mechanical properties as well as failure mechanism.

115
9. Bibliography
1. Dauda, B., S. Oyadiji, and P. Potluri, Characterising Mechanical Properties of
Braided and Woven Textile Composite Beams. Applied Composite Materials, 2009.
16(1): p. 15-31.
2. Potluri, P., et al., Geometrical modelling and control of a triaxial braiding machine
for producing 3D preforms. Composites Part A: Applied Science and Manufacturing,
2003. 34(6): p. 481-492.
3. Manders, P. and M. Bader, The strength of hybrid glass/carbon fibre composites.
Journal of Materials Science, 1981. 16(8): p. 2233-2245.
4. Kyosev, Y., 1 - Introduction: the main types of braided structure using maypole
braiding technology A2 - Kyosev, Y, in Braiding Technology for Textiles. 2015,
Woodhead Publishing. p. 1-25.
5. Head, A., F. Ko, and C. Pastore, Handbook of Industrial Braiding. Atkins and Pearce.
6. Guyader, G., A. Gabor, and P. Hamelin, Analysis of 2D and 3D circular braiding
processes: Modeling the interaction between the process parameters and the pre-
form architecture. Mechanism and Machine Theory, 2013. 69: p. 90-104.
7. Ko, F., C. Pastore, and A. Head, Atkins & Pearce Handbook of Industrial Braiding.
1989: Atkins & Pearce.
8. Kyosev, Y., 6 - Braiding machine components A2 - Kyosev, Y, in Braiding Technology
for Textiles. 2015, Woodhead Publishing. p. 115-151.
9. Douglass, W., Braiding and Braiding Machinery. 1964: Centrex Publishing Company;
Cleaver-Hume Press.
10. Lee, S., Handbook of Composite Reinforcements. 1992: Wiley.
11. Du, G., P. Popper, and T. Chou, Analysis and Automation of Two-Step Braiding, in
Fibre Tex'88 Conference. 1988: Greenville, S.C.
12. Omeroglu, S., The Effect of Braiding Parameters on the Mechanical Properties of
Braided Ropes. Fibres & Textiles in Eastern Europe, 2006. Nr 4 (58): p. 53-57.
13. Omeroglu, S., The effect of braiding parameters on the mechanical properties of
braided ropes. Vol. 14. 2006. 53-57.
14. Nishimoto, H., et al., Generation and prediction methods for circumferential
distribution changes in the braiding angle on a cylindrical braided fabric.
Proceedings of the Institution of Mechanical Engineers, Part L: Journal of Materials
Design and Applications, 2010. 224(2): p. 71-78.
15. Chretien, N., Numerical constitutive models of woven and braided textile structural
composites. 2002.
16. Zhang, Q., et al., Structure Analysis of a Two-dimensional Braided Fabric. The
Journal of The Textile Institute, 1997. 88:1: p. 41-52.
17. Farnfield, C. and P. Alvey, Textile Terms and Definitions. 1975: The Textile Institute,
Manchester.
18. Behera, B. and P. Hari, Woven textile structure: Theory and applications. 2010:
Elsevier.
19. Asayesh, A. and A. Jeddi, Modeling the Creep Behavior of Plain Woven Fabrics
Constructed from Textured Polyester Yarn. Textile Research Journal, 2010. 80(7): p.
642-650.
20. Ogin, S., 11 - Textile-reinforced composite materials, in Handbook of Technical
Textiles. 2000, Woodhead Publishing. p. 264-281.
21. Schwartz, M., Composite Mterials Handbook. 1984, the United States of America:
McGraw-Hill Book Company.
22. Cantwell, W. and J. Morton, The impact resistance of composite materials — a
review. Composites, 1991. 22(5): p. 347-362.

116
23. Campbell, F., Structural Composite Materials. 2010: ASM International.
24. Segui, J. Protecting Aircraft Composites from Lightning Strike Damage. 2015 [cited
2017 16/12]; Available from: https://www.comsol.com/blogs/protecting-aircraft-
composites-from-lightning-strike-damage/.
25. Zurschmeide, J. Lamborghini is rewriting the rules on carbon fibre with a crazy new
way to make it. 2016 [cited 2018 10th January]; Available from:
https://www.digitaltrends.com/cars/lamborghini-forged-carbon-fiber-
manufacturing-process/#/3.
26. Ayranci, C. and J. Carey, 2D braided composites: A review for stiffness critical
applications. Composite Structures, 2008. 85(1): p. 43-58.
27. Brookstein, D., Structural applications of advanced braided composites. 1988. 415-
424.
28. Branscomb, D., D. Beale, and R. Broughton, New directions in braiding. J. Eng.
Fibers Fabr, 2013. 8(2): p. 11-24.
29. Tolosana, N., S. Lomov, and A. Miravete, Development of a geometrical model for a
3D braiding unit cell based on braiding machine emulation. Finite Element
Modelling of Textiles and Textile Composites, St Petersbourg, 2007: p. 26-28.
30. Lebel, L. and A. Nakai, Design and manufacturing of an L-shaped thermoplastic
composite beam by braid-trusion. Composites Part A: Applied Science and
Manufacturing, 2012. 43(10): p. 1717-1729.
31. Michaeli, W., U. Rosenbaum, and M. Jehrke, Processing strategy for braiding of
complex-shaped parts based on a mathematical process description. Composites
Manufacturing, 1990. 1(4): p. 243-251.
32. Melenka, G., et al., 3 - Manufacturing processes for braided composite materials, in
Handbook of Advances in Braided Composite Materials. 2017, Woodhead
Publishing. p. 47-153.
33. [cited 2017 16/12]; Available from: http://www.xxvii.com/2017-custom-time-
fluidity-endurance-road-frameset.
34. Ken, D., Evaluation of fibre damage between 2-D radial and circumferential
braiding process. 2016, the University of Manchester. p. 67.
35. Ebel, C. and T. Hans, Interaction of braiding ring geometry and fiber lay-up in the
braiding process, in A comprehensive approach to Carbon Composites Technology.
2014, Research Campus Garching, Germany: Institute for Carbon Composites.
36. Fouladi, A. and R. Nedoushan, Prediction and optimization of yarn path in braiding
of mandrels with flat faces. Journal of Composite Materials. 0(0): p.
0021998317710812.
37. Prodromou, A. and J. Chen, On the relationship between shear angle and wrinkling
of textile composite preforms. Composites Part A: Applied Science and
Manufacturing, 1997. 28(5): p. 491-503.
38. Lin, H., et al., Predictive modelling for optimization of textile composite forming.
Composites Science and Technology, 2007. 67(15–16): p. 3242-3252.
39. Boisse, P., N. Hamila, and A. Madeo, Modelling the development of defects during
composite reinforcements and prepreg forming. Philosophical Transactions of the
Royal Society A: Mathematical, Physical and Engineering Sciences, 2016. 374(2071).
40. Zhu, B., et al., Theoretical modeling of large shear deformation and wrinkling of
plain woven composite. Journal of Composite Materials, 2009. 43(2): p. 125-138.
41. Gutowski, T., et al., Laminate wrinkling scaling laws for ideal composites.
Composites Manufacturing, 1995. 6(3): p. 123-134.
42. Arndt, R., Fabric Preforming for Structural Reaction Injection Molding. Proceedings
of the Institution of Mechanical Engineers, Part L: Journal of Materials Design and
Applications, 1991: p. 35-44.

117
43. Zhu, B., et al., Theoretical Modeling of Large Shear Deformation and Wrinkling of
Plain Woven Composite. Journal of Composite Materials, 2009. vol. 43(no. 2): p.
125-138.
44. Potter, K., Bias extension measurements on cross-plied unidirectional prepreg.
Composites Part A: Applied Science and Manufacturing, 2002. 33(1): p. 63-73.
45. Roy, S. and P. Potluri, Braid-winding for Optimum Tubular Preforming, in Texcomp-
12 Conference. 2015: Raleigh, NC, USA.
46. Tam, A. and T. Gutowski, The kinematics for forming ideal aligned fibre composites
into complex shapes. Composites Manufacturing, 1990. 1(4): p. 219-228.
47. Rozant, O., P. Bourban, and J. Månson, Drapability of dry textile fabrics for
stampable thermoplastic preforms. Composites Part A: Applied Science and
Manufacturing, 2000. 31(11): p. 1167-1177.
48. Sharma, S., M. Sutcliffe, and S. Chang, Characterisation of material properties for
draping of dry woven composite material. Composites Part A: Applied Science and
Manufacturing, 2003. 34(12): p. 1167-1175.
49. Boisse, P., et al., Simulation of wrinkling during textile composite reinforcement
forming. Influence of tensile, in-plane shear and bending stiffnesses. Composites
Science and Technology, 2011. 71(5): p. 683-692.
50. Lebrun, G., M. Bureau, and J. Denault, Evaluation of bias-extension and picture-
frame test methods for the measurement of intraply shear properties of PP/glass
commingled fabrics. Composite Structures, 2003. 61(4): p. 341-352.
51. Hubert, P. and A. Poursartip, Aspects of the compaction of composite angle
laminates: an experimental investigation. Journal of Composite Materials, 2001.
35(1): p. 2-26.
52. Sharma, S., et al., Mapping of tubular woven composite preforms on to doubly-
curved surfaces. Computer-Aided Design, 2001. 33(14): p. 1035-1048.
53. von Kármán, T., L. Dunn, and H. Tsien, The Influence of Curvature on the Buckling
Characteristics of Structures, in Collected Works of H.S. Tsien (1938–1956). 2012,
Elsevier: Oxford. p. 122-145.
54. Yu, W., P. Harrison, and A. Long, Finite element forming simulation for non-crimp
fabrics using a non-orthogonal constitutive equation. Composites Part A: Applied
Science and Manufacturing, 2005. 36(8): p. 1079-1093.
55. Lin, H., et al., Textile Wrinkling in Composite Forming. University of Nottingham and
University of Glasgow, 2006.
56. Sharma, S., et al., Manufacturing of doubly curved tubular composite structures:
Mapping and weave modifications. Journal of Thermoplastic Composite Materials,
2002. 15(3): p. 209-225.
57. Toray carbon fibres america, i., Torayca T700S data sheet.
58. Harte, A. and N. Fleck, On the mechanics of braided composites in tension.
European Journal of Mechanics - A/Solids, 2000. 19(2): p. 259-275.
59. Roy, S., W. Zou, and P. Potluri, Influence of Braid Carrier Tension on Carbon Fibre
Braided Preforms, in Recent Developments in Braiding and Narrow Weaving, Y.
Kyosev, Editor. 2016, Springer International Publishing: Cham. p. 91-102.
60. IKEA [cited 2018 9th Jan]; Available from:
http://www.ikea.com/gb/en/products/decoration/plants-pots-stands/askholmen-
trellis-grey-brown-stained-art-70258671/.
61. Roy, S., D. Yang, and P. Potluri, Influence of Bending on Wrinkle Formation and
Potential Methods of Mitigation, in 21st International conference on composite
materials. 2017: Xi'an,China.

118