investigation of deformation behaviour of dry textiles

Investigation of deformation behaviour of

dry textiles under forming forces by

computed tomography

A thesis submitted to The University of Manchester for the Degree of

Doctor of Philosophy

In the Faculty of Engineering and Physical Sciences

2014

Zeshan Yousaf

School of Materials

Department of Textile and Paper

The University of Manchester

Contents

2

Contents

List of Tables 6

List of Figures 7

Abstract 11

Declaration 12

Copyright Statement 13

Dedication 15

Acknowledgment 16

Chapter 1. Introduction 17

1.1. Introduction 17

1.2. Problem definition 17

1.3. Research aim 18

1.4. Research objectives 18

1.5. Thesis lay out 19

Chapter 2. Literature Review 22

2.1. Textile fabrics 22

2.1.1. Woven fabrics 22

2.1.1.1. 2D woven fabrics 22

2.1.1.1.1. Plain fabrics 23

2.1.1.1.2. Twill fabrics 23

2.1.1.1.3. Satin and sateen fabrics 24

2.1.1.2. 3D woven fabrics 25

2.1.1.2.1. Orthogonal fabrics 25

2.1.1.2.2. Angle-interlock fabrics 26

2.1.2. Non-crimp stitched fabrics 27

2.1.3. Braided fabrics 28

2.2. Composite manufacturing processes 28

2.2.1. Resin transfer moulding 28

2.2.2. Vacuum bagging 29

2.2.3. Autoclave processing 30

2.3. Compaction of textile preforms 31

Contents

3

Chapter 3. Deformation of single layer dry preforms under

compaction

64


3.2. Material and mechanical testing 68

3.2.1. Mechanical testing 69

3.3. Tomography and in-situ compression rig 69

3.4. Results and discussion 72

3.4.1. Macroscopic deformation 72

3.4.2. Meso-structural analysis by computed tomography 77

3.4.3. Resin channels in dry preform 87

3.4.4. Yarn packing fraction 93

3.4.5. Yarn and fibre volume fraction of the preform 95

3.5. Conclusion 98

Chapter 4. Deformation of multilayer dry preforms under

compaction

105


4.2. Nesting factor and tow geometry parameters 109


4.3.1. Material 110


4.4. Tomography and in-situ compression set up 112



4.5.2. Meso-structural by computed tomography (CT) 114

4.5.2.1. Nesting factor analysis 122

4.5.2.2. Quantification of inter-tow voids using image analysis 123

4.5.2.3. Yarn packing fraction 126

4.6. Conclusions 127

Chapter 5. Compaction and nesting in textile preforms influenced

by tow architecture

134


5.2. Material & Experimental details 136

Contents

4

5.2.1. Material 136

5.2.2. Experimental methodology 138

5.2.2.1. Mechanical testing 138

5.2.2.2. Computed tomography (CT) analysis 139

5.3. Analysis of macro-scale deformations 141

5.3.1. Woven fabrics 141

5.3.2. Non-crimp fabrics 146

5.3.3. Nesting of layers 148

5.4. Analysis of meso-scale deformations 151

5.4.1. Tow waviness 152

5.4.2. Inter-tow voids 155


Chapter 6. Deformation of dry and wet preforms under compaction 164



6.2.1. Material 166




6.3.2. Tomography and in-situ compression set up 173

6.3.3. Meso-structural analysis by computed tomography (CT) 175

6.3.3.1. Tow waviness 177

6.3.3.2. Inter-tow voids 179


Chapter 7. Development of an in-situ technique to analyse the

meso-structure of dry fabrics under biaxial loadings

184


7.2. Development of biaxial testing machine 187

7.2.1. Load cell calibration 189

7.3. Tomography and in-situ loading rig 190


Chapter 8. Conclusions and future work 197


Contents

5

8.2. Future work 200

Publications 202

List of tables

6

List of Tables

Table 3.1. Tow geometry parameters (warp) 79

Table 3.2. Tow geometry parameters (weft) 79

Table 3.3. Tow area calculated by CT and Texgen 87

Table 3.4. Yarn packing fractions of warp cross-sections 95

Table 3.5. Yarn packing fractions of weft cross-sections 95

Table 4.1. Parameters used to describe the fabric geometry 115

Table 4.2. Tow geometry parameters measured by X-ray tomography 118

Table 4.3. Nesting measurements for the 6 layer preform 120

Table 4.4. Yarn packing fraction 127

Table 5.1. Woven fabric specifications 137

Table 5.2. Stitched non-crimp fabric (NCF) specifications 137

Table 6.1. Material used for mechanical testing 166

Table 6.2. Material used for meso-structure analysis 166

List of figures

7

List of Figures

Figure 2.1. Plain weave 23

Figure 2.2. Twill weave 24

Figure 2.3. Satin weave 24

Figure 2.4. Orthogonal weave 26

Figure 2.5. Angle-interlock 27

Figure 2.6. NCF fabrics 27

Figure 2.7. Braided fabric 28

Figure 2.8. Resin transfer moulding 29

Figure 2.9. Vacuum bagging 30

Figure 2.10. Autoclave processing 31

Figure 2.11. Typical pressure thickness cure 33

Figure 2.12. Stages of fabric compression 34

Figure 2.13. Schematic representation of time dependent behaviour of fibrous

materials

36

Figure 2.14. Main factors affecting the compaction behaviour of preforms 47

Figure 3.1. Pressure thickness curve for a woven fabric under compaction 66

Figure 3.2. Scanned image of plain weave (1/1) fabric 68

Figure 3.3. Schematic of the compression rig 70

Figure 3.4. (a) Compression rig accommodated on the tomography stage, and

(b) close-up of the rig

71

Figure 3.5. Pressure-thickness response for a single layer under dry and wet

conditions

73

Figure 3.6. Effect of loading time on the thickness of the single layer dry fabric 73

Figure 3.7. t/t0 plotted against time for different pressures 74

Figure 3.8. The constants plotted against pressure 75

Figure 3.9. Experimental and predicted thicknesses with time at different

pressures, (a) dry fabric and (b) wet fabric

76

Figure 3.10. Definition of the yarn geometry parameters 77

Figure 3.11. Segmented virtual slices of warp and weft yarns 78

Figure 3.12. 3D reconstruction of single layer fabric 78

Figure 3.13. Cross-sectional view of the preform at the centre of the tow 81

Figure 3.14. 3D reconstructions of yarns, (a) warp yarn, (b) weft yarn 81

List of figures

8

Figure 3.15. The crimp amplitude of the warp and the weft yarns 82

Figure 3.16. The crimp percentage of warp and weft yarns 82

Figure 3.17. The crimp angle of warp and weft yarns 83

Figure 3.18. Cross-sectional images of the warp and the weft yarns in between

two tows

85

Figure 3.19. Tow cross-section at the centre due to crimp angle 86

Figure 3.20. Inter-tow voids in 3D structure 89

Figure 3.21. Inter-tow voids at the centre of the tow intersections 89

Figure 3.22. Inter-tow voids in between two tows 90

Figure 3.23. Average void thickness with pressure 91

Figure 3.24. The frequency distribution of the void thickness 91

Figure 3.25. Resin channels in dry preform during compression 93

Figure 3.26. Yarn packing fractions of dry preform against different loads 94

Figure 3.27. Yarn packing fraction on application of pressure 94

Figure 3.28. Yarn volume fractions calculated at the centre of the tow

intersections

96

Figure 3.29. Yarn volume fractions calculated in between two tows 96

Figure 3.30. FVF calculated by mechanical testing and tomographic analysis 98

Figure 4.1. Typical pressure thickness curve for a woven fabric under

compaction

107

Figure 4.2. Layer thickness (a) Single layer, (b) 2 layer without nesting, (c) 2

layer with shifting and nesting

110

Figure 4.3. Compression Rig (a) Compression rig fixed on the tomography

stage, (b) close-up of the rig

113

Figure 4.4. Pressure thickness curves of different layers 114

Figure 4.5. Yarn geometry parameters 115

Figure 4.6. 3D view of the six layer dry preform 116

Figure 4.7. Tomographic sections through the stack of six layers at different

pressures

119

Figure 4.8. Yarn isolated from the tomograph showing crimp in crossing yarns 120

Figure 4.9. Tomographic weft cross-sections showing layers with and without

shift

121

Figure 4.10. Nesting factors from mechanical testing and tomographic analysis 123

List of figures

9

Figure 4.11. Flow channels at (a) 4 kPa, (b) 45 kPa, and (c) 100 kPa 124

Figure 4.12. 3D representations of inter-tow voids at (a) 4kPa, (b) 45kPa, and

(c) 100kPa

125

Figure 4.13. Inter-tow voids at different slices and average inter-tow voids 126

Figure 5.1. Woven fabrics (a) Plain, (b) 3/1 twill, (c) 5H sateen 137

Figure 5.2. Scanned images of woven fabrics and non-crimp stitched fabrics 138

Figure 5.3. Compression set up 139

Figure 5.4. Compression rig used for set up 140

Figure 5.5. Single layer thickness results against pressure 142

Figure 5.6. Compression behaviour of multi-layer fabrics, (a) plain, (b) twill,

(c) sateen

144

Figure 5.7. Different layers thickness results against pressure, (a) two layers (b)

six layers

145

Figure 5.8. Non-crimp fabric thickness results as a function of pressure, (a)

+45о/-45

о, (b) 0

о/90

о

147

Figure 5.9. Nesting factors (a) woven fabrics (b) non-crimp stitched fabrics 150

Figure 5.10. 3-D reconstruction of single layer (right) and 10 layers (left) 152

Figure 5.11. CT images of single layer 152

Figure 5.12. Tow showing crimp 153

Figure 5.13. Yarn crimp percentage (a) warp yarn, (b) weft yarn 153

Figure 5.14. CT images of ten layer stack 155

Figure 5.15. Inter-tow voids in single layer and multilayer stack 157

Figure 5.16. Inter-tow voids in single layer along slices (a) 100 kPa, (b) 300

kPa

157

Figure 5.17. Inter-tow voids in multilayer stack along slices (a) 100 kPa, (b)

300 kPa

157

Figure 5.18. Plain-weave repeat presenting, between two tows and centre of the

tow

158

Figure 6.1. Thickness vs pressure of single layer and two layers of P1 fabric 168

Figure 6.2. Thickness vs pressure of single layer and two layers of T1 fabric 168

Figure 6.3. Thickness vs pressure of single layer and two layers of S1 fabric 169

Figure 6.4. Thickness vs pressure of single layer and two layers of NCF 169

Figure 6.5. Thickness vs pressure of single layer and two layers of plain fabric 169

List of figures

10

P2

Figure 6.6. Td-Tw/T0 plotted against pressure for plain (P2) woven glass fabric 171

Figure 6.7. Experimental and predicted wet thickness of plain fabric (P1) 171

Figure 6.8. Experimental and predicted wet thickness of twill fabric (T1) 172

Figure 6.9. Experimental and predicted wet thickness of sateen fabric (S1) 172

Figure 6.10. Experimental and predicted wet thickness of NCF 172

Figure 6.11. Experimental and predicted wet thickness of plain fabric (P2) 173

Figure 6.12. Compression Rig (a) Compression rig fixed on the tomography

stage, (b) close-up of the rig

174

Figure 6.13. CT images showing slice of (a) warp cross-section, (b) segmented

inter-tow voids and (c) segmented yarns

175

Figure 6.14. Reconstructed image of single layer fabric compacted in

polycarbonate plate

175

Figure 6.15. Yarn geometry parameters 176

Figure 6.16. Cross- sectional images of single layer dry and wet fabrics 177

Figure 6.17. Tow waviness of dry and wet fabrics (a) warp yarns, b) weft yarns 178

Figure 6.18. Inter-tow voids in dry and wet fabric 180

Figure 6.19. Inter-tow voids in (a) dry fabric and (b) wet fabric at 10 kPa 181

Figure 7.1. Rig for biaxial shear and tensile testing 188

Figure 7.2. Biaxial rig with cruciform specimen 189

Figure 7.3. Rig to grip the biaxially loaded fabric 190

Figure 7.4. CT images of the biaxially loaded dry fabric, a) side view, b) top

view and crossing yarns in two slices

192

Abstract

11

Abstract

Textile preforms undergo different deformations during preforming including transverse

compression, biaxial tensile, in-plane shear and out-of-plane bending deformation. All

these deformations change the tow and resin channel geometry of the preform and

consequently affect the mechanical properties of the final product. To make simulation

tools and to execute structural analysis, accurate experimental data during these

deformations is required.

In the present work, an in-situ measurement technique has been developed to study the

deformation behaviour of textile preforms at meso-scale during the compression mode.

This technique for the first time, enables measurement of the meso-structure of dry and

wet fabrics under in-situ loadings. Tow geometry changes of single layer and multilayer

preforms have been captured under in-situ loadings using computed tomography (CT).

Resin channels in single layer and multilayer preforms have been investigated in detail.

In multilayer fabrics, it has been observed that nesting of the layers strongly influences

the inter-tow voids. It has been observed that crimp reduction in single layer fabric and

nesting of layers in multilayer fabric is responsible for the fabric thickness reduction in

a low load regime (~100 kPa).

During study of different fabric architectures, it has been observed that the float length

of the weave plays an important role in compression behaviour.

The deformation of wet fabrics has also been studied at macro and meso-levels using

mechanical test results and CT. It has also been observed that at low loads (~300 kPa)

there is higher crimp reduction in wet fabrics compared to dry fabrics.

Additionally, an in-situ measurement technique has also been developed to study the

meso-structure of dry fabrics under biaxial tensile and shear loading using CT.

Declaration

12

Declaration

No portion of the work referred to in the thesis has been submitted in support of an

application for another degree or qualification of this or any other university or other

institute of learning.

Copyright statement

13

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 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

Copyright statement

14

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.

Dedication

15

Dedication

Dedicated to the soul of my beloved mother

Acknowledgements

16

Acknowledgements

I would sincerely like to acknowledge my supervisor Prof. Prasad Potluri for his

guidance, help and encouragement for my research work, especially his motivation

towards my goals which I can never forget throughout my life. In fact, he has been a

source of motivation for me in all aspects.

Special thanks to my co-supervisor Prof. Philip Withers to guide me towards my

research work and providing me an excellent opportunity to perform my experimental

work in X-ray Imaging Facility which enabled me to successfully achieve my research

goals.

I also wish to thank X-ray imaging Facility team, especially Mr Fabien Leonard to help

me during the experimental work. Thanks to Stuart Morse in Material Science Lab and

Thomas Kerr in weaving Lab for their help.

My sincere thanks to Dr. Richard Kannon for his endeavouring support during my

write-up. Special thanks to Sabahat Nawaz, Khayale Jan and Ibraheem for their help.

I would also like to thank all my colleagues and friends for their help and support

during my research work especially Vivek, Haseeb, Alvaro, Dhaval, Mubeen,

M.Peerzada and Rishad.

I would like to thank Bhauuddin Zakria University Multan for funding my research.

Finally, I would like to express my thanks to my family, especially my father and sister

who continuously encouraged me throughout my work.

Chapter 1 Introduction

17

Chapter 1

Introduction

1.1 Introduction

Low-cost composites manufacturing techniques based on resin infusion of dry textile

preforms are becoming popular in the aerospace, automotive and energy sectors. In

these processes, several fabric plies are draped on the tool surface and are subjected to

transverse compaction and biaxial tensile and shear forces either by vacuum (Vacuum

Infusion) or external pressure (Resin Transfer Moulding), and at the same time infused

with liquid resin. These forces change the tow and resin channel geometry, resulting in

change in the mechanical properties of the final product. To make simulation tools and

to perform structural analysis, accurate experimental data for these geometrical changes

during forming is required.

1.2 Problem definition

Deformation of textile preforms under different loading conditions (transverse

compaction, biaxial tensile and shear) during forming plays an important role on tow

and resin channel geometry and on the mechanical properties of the composites.


18

Extensive research work has been performed on the deformation behaviour of textile

preforms during composite manufacturing. The limitation of the existing research work

is that structural analysis at meso-scale during forming is only available on laminated

composites. No research work is available on the deformation of dry and wet textile

preforms prior to curing during forming at meso-scale. Additionally, the effect of weave

structure of textile fabrics has been studied by many researchers but no attempt has been

taken to keep the tow spacing and tow count similar during the weave structure

comparisons. In order to make structural analysis and simulation tools accurate,

knowledge of deformations in dry textile fabrics under different loadings prior to curing

is required.

1.3 Research Aim

The aim of the present research work is to study deformation behaviour of dry and wet

textile fabrics under transverse compaction forces at macro and meso-scale and to

develop a technique to measure the dry fabric behaviour under in-situ biaxial tensile and

shear forces.

1.4 Research Objectives

Mechanical testing of single layer and multilayer dry and wet glass fabric

samples using an Instron testing machine.

Development of a compression rig calibrated by the mechanical test results to

measure the meso-structure of glass fabrics under in-situ compression loading.

Structural analysis of glass fabric samples under in-situ compression loading

using computed tomography (CT).


19

Development of different woven fabric architectures by keeping similar thread

densities and thread counts.

Macro-scale deformation study of developed woven fabrics and non-crimp

fabrics by testing of the fabric samples using an Instron testing machine.

Development of the biaxial testing machine and an in-situ rig to study the meso-

structure of dry textile preforms under biaxial tensile and shear loadings using

CT.

1.5 Thesis layout

Chapter-2: Literature review

This chapter gives an introduction of the basic textile architectures, different composite

manufacturing techniques and detail of work done on the deformation of textile

preforms during forming.

Chapter-3: Deformation of single layer dry preforms under compaction

This chapter presents an introduction to compaction; the methodology to investigate

mechanical testing and the computed tomography (CT) scanning and deformation

behaviour of single layer preforms at meso-level using computed tomography (CT).

Different tow geometry parameters of tow changes, the resin channels in the preform

during compaction, comparison of computed tow area calculated from Texgen software

and CT images and comparison of fibre volume fraction calculated from mechanical test

results and from computed tomography data are examined.


20

Chapter-4: Deformation of multilayer dry preforms under compaction

In this chapter the deformation behaviour of six layer plain woven preforms under

compaction has been presented at meso-scale.

A brief introduction of the deformation behaviour of multilayer preforms and the

nesting of layers has been discussed. The methodology used for the mechanical testing

and computed tomography has been presented.

Macro and meso-level deformation of six layer preforms have been presented and

different tow geometry parameters calculated for six layer preforms have been

discussed at low loads of 100 kPa.

Chapter-5: Compaction and nesting in textile preforms influenced by tow

architectures

In this chapter, the deformation behaviour of different textile preforms has been studied

at macro-level. The comparison of single layer and multilayer preforms of each

architecture under compaction has been presented and nesting factors have been

calculated for all these using mechanical test results. The tow waviness and resin

channels of single and ten layer specimens have been investigated using CT images.

Chapter-6: Deformation of dry and wet preforms under compaction

Here the deformation of dry and wet fabrics has been presented at macro and meso-

levels. The deformation behaviour of different fabrics has been studied using the

mechanical test results at macro level. Power law relations have been derived to predict


21

the wet thickness results from dry fabric thickness. Tow waviness and resin channels

have been investigated and compared for single layer dry and wet preforms.

Chapter-7: Development of an in-situ technique to analyse the meso-

structure of dry fabrics under biaxial loading

In this chapter, a method to measure the deformation of dry preforms under in-situ shear

and tensile loading has been presented. Plain woven fabric has been investigated under

biaxial tensile loading to verify the accuracy of the technique.

Chapter-8: Conclusions and future work

This chapter includes the conclusions and suggestions for future work.

Chapter 2 Literature review

22

Chapter 2

LITERATURE REVIEW

2.1 Textile fabrics

A textile fabric is defined as a manufactured assembly of fibres and/or yarns, which has

a substantial surface area in relation to its thickness and sufficient inherent cohesion to

give mechanical strength to the assembly [1].

Based on the manufacturing techniques, conventional textile fabrics can be divided into

woven, non-woven, knitted and braided formations. A brief introduction of these

structures is presented here.

2.1.1 Woven fabrics

Woven fabrics that are used in composites can be classified as two dimensional (2D)

and three dimensional (3D) structures [2].

2.1.1.1 2D woven fabrics

2D woven fabrics are made by the interlacement of two groups of threads, warp and

weft, which are placed at right angles to each other in the plane of the fabric. The warp

and the weft yarns are oriented along the length and the width of the fabric respectively.

Warp yarns are also known as ends and weft yarns as picks. Depending on the repeat


23

pattern of the interlacement, 2D woven fabrics can be further classified; some examples

are plain, twill and satin fabrics [3].

2.1.1.1.1 Plain fabrics

The plain weave is the simplest of all the available weaving patterns. This weave has the

highest possible interlacing frequency. Every weft yarn passes over and under the

successive warp yarn and repeats the same pattern with alternate yarns in the following

row of weft yarn. A single repeat of the plain weave is depicted in Figure 2.1.

Further derivatives of the plain weave fabric can be made by passing two or more

adjacent warp threads and/or two or more adjacent weft threads at the same time. In this

way it is possible to obtain larger warp and weft covered areas than in plain weave

fabrics [4].

Figure 2.1. Plain weave [5]

2.1.1.1.2 Twill fabrics

Twill weave fabrics can be identified by the diagonal lines on the face of the fabric due

to warp and weft floats [3, 4]. This weave can further be categorized by “S” or “Z” twill

depending on the direction of the diagonal lines. The twill weave can be produced on a


24

minimum of 3 ends and picks (2/1-twill) with no theoretical upper limit [3]. A single

repeat of 3/1-twill is presented in Figure 2.2.

Figure 2.2. Twill weave [5]

2.1.1.1.3 Satin and sateen fabrics

Satin weave is a warp faced weave whereas the sateen weave is a weft faced weave. In

both satin and sateen weaves, the weave interlacements are arranged in such a way that

a smooth surface is achieved free from twill lines [3]. The smallest repeat of satin and

sateen is 5, a repeat of 5 end satin is shown in Figure 2.3. The most popular weave

repeats are 5 and 8 [4].

Figure 2.3. Satin weave [5]


25

2.1.1.2 3D woven fabrics

3D fabrics are recognised by the presence of thickness of the fabric in the Z-direction in

addition to the X and Y directions. These fabrics are woven with multiple warp or weft

layers. The thickness of a 3D fabrics is considerable in comparison to 2D fabrics [2].

Orthogonal and angle interlock fabrics are the most renowned classes of 3D woven

fabrics.

2.1.1.2.1 Orthogonal fabrics

In orthogonal fabrics, the straight yarns are arranged perpendicular to each other in X, Y

and Z directions as shown in Figure 2.4. The absence of crimp in the warp and weft

yarns makes this structure ideal for applications where non-crimp features are required.

Both isotropic and anisotropic preforms can be achieved by arranging the number of

yarns in each dimension [2, 6]. The orthogonal fabrics can further be classified into two

types, through-the-thickness orthogonal (Figure 2.4a) and layer-to layer-orthogonal

(Figure 2.4b). In orthogonal through-the-thickness fabrics, binding warp travels from

one surface of the preform to the other, holding together all the layers of the preform.

The orthogonal layer-to-layer is a multilayer woven fabric in which binding warps

travel from one layer to the adjacent layer and back.


26

Figure 2.4. Orthogonal weave, (a) Through-the-thickness, (b) Layer-to-layer

2.1.1.2.2 Angle-interlock fabrics

In angle-interlock structures, the warp (or weft) yarns are used to bind many layers of

weft (or warp) yarns with weft/warp yarns being straight. A third set of yarns (stuffer

yarns) can also be added in angle-interlock fabrics to increase fibre volume fraction and

in-plane strength [2]. Figure 2.5 presents the angle-interlock structure with warp, weft

and binder yarns. Like orthogonal fabrics, angle interlock structure can also be divided

into through-the-thickness (Figure 2.5a) and layer-to-layer (Figure 2.5b) angle interlock

depending on the passage of binder yarn

(a) (b)


27

Figure 2.5. Angle-interlock, (a) Through-the-thickness, (b) Layer-to-layer

2.1.2 Non-crimp stitched fabrics (NCF)

In non-crimp stitched fabrics, a light weight fibre/yarn is used as a loop which is sewn

or knitted around the reinforcement tow to create the fabric as can be seen in Figure 2.6.

These fabrics are commonly known as non-crimp fabrics (NCF) as the reinforcement

tow remains straight without crimp [7].

Composites made of NCF are used in aerospace, automotive, civil engineering and the

wind turbine industry due to their high strength properties [8].

Figure 2.6. NCF fabric [5]

(a) (b)


28

2.1.3 Braided fabrics

In braided fabrics, three or more threads are interlaced in such a way that they cross one

another in diagonal formation to form the braided structure. A braided structure is

presented in Figure 2.7. Braiding is a simple form of narrow fabric construction and the

limitations of knitting and weaving make braiding an important method of fabric

construction in the textile composites industry [1, 9].

Figure 2.7. Braided fabric [5]

2.2 Composite manufacturing processes

There are a number of different techniques for fibre reinforced polymer (FRP)

composite manufacturing. A brief introduction of some main manufacturing processes

is presented here.

2.2.1 Resin transfer moulding process (RTM)

Resin transfer moulding is the most common manufacturing process for structurally

capable composites. RTM is capable of producing large, complex and highly integrated

components. Additionally it has advantages of low capital cost, low mould cost and a

good work environment. In this process, the dry reinforcement is placed into a closed

mould and liquid resin is injected into the closed mould to impregnate the reinforcement.


29

A vacuum can also be applied in the mould cavity to infiltrate resin into the fabric,

which is known as vacuum assisted resin transfer moulding. During RTM, the injection

pressures range from 0.1 MPa to over 1 MPa depending on fibre content and resin type

[10, 11]. Figure 2.8 presents a schematic representation of RTM.

Figure 2.8. Resin transfer moulding [10]

2.2.2 Vacuum bagging

In the vacuum bagging process pressure is applied to the laminate once it is laid-up in

order to improve its consolidation. This is achieved by sealing a plastic film over the

dry laid-up laminate and onto the tool. To remove the air under the bag, a vacuum pump

is used and a pressure of one atmosphere is applied to achieve the consolidation of the

laminate. Figure 2.9 represent a typical vacuum bagging set up.

Mould tool

Mould tool

Resin

injection

under

pressure

Compression to secure

the mould

Optional

vacuum

assistance

Dry reinforcement preform


30

Figure 2.9. Vacuum bagging [12]

2.2.3 Autoclave processing

To achieve the maximum performance of thermoset composite materials, an increase in

the fibre to resin ratio and removal of all air voids is required. This can be obtained by

subjecting the material to elevated pressures and temperatures. Vacuum bagging can

apply atmospheric pressure equal to one bar. To increase pressure to more than one bar,

additional external pressure is required. This is achieved by autoclave processing in

which pressures up to 5~7 bar can be applied [13]. In the autoclave, a composite is laid

up and enclosed in a vacuum bag. A full or partial vacuum is drawn within the bag and

the bag is kept inside the autoclave chamber where the pressure is kept at more than one

bar. This extra pressure is added on the exterior of the bag. Simultaneously the

temperature is raised which results in reduction in viscosity of the polymer. Thus the

wetting of the reinforcing fibres and the consolidation becomes better. A schematic

representation of the autoclave process is presented in Figure 2.10.

To vacuum pump To vacuum gauge

Vacuum bagging film

Sealant tape

Release film Release coated

mould Laminate

Peel ply

Breather/absorption

fabric


31

Figure 2.10. Autoclave processing [14]

2.3 Compaction of textile preforms

During the manufacturing of composites, the dry preform is compacted under a certain

level of pressure, which changes the tow geometry of the dry preform, the resin

permeability and the final mechanical properties of the composites [15]. Hence, this

compaction process is considered an important parameter of the manufacturing

processes [16]. An intensive study is important to keep a record of the changes of tow

geometry at each and every stage of composite manufacturing so that reliable data can

be extracted for making accurate simulation tools [15].

Extensive research has been carried out on the compaction behaviour of different fabrics

using mechanical test methods and structural analysis [15-49].

Van Wyk [27] is considered the first researcher who treated the fibres under the

compression as a system of bending units. He derived a relationship between the

pressure and volume of a mass of wool fibres by assuming that compression of the wool

Bleeder

Vacuum connection Autoclave

Prepreg

Sealant tape Peel ply

Release film

perforated

Mould


32

fibres is due to bending of the wool fibres. He also suggested that compressibility

measurements constitute a convenient method of comparing the flexural elastic

properties of the fibres composing different wool samples.

Gutowski et al. [50] studied the compression behaviour of aligned fibre bundles. They

proposed a simple elastic deformation model for the transverse compression of aligned

fibre bundles by using Van Wyk’s idea of deformation. Later on, the deformation of

lubricated fibre bundles was also studied by Gutowski et al. [51] and they proposed a

method to measure the fibre volume fraction during consolidation.

De Jong et al. [21] developed a mechanical model for the lateral compression of wool

woven fabrics based on Van Wyk’s compression law of fibre assemblies. They

observed that the compressible surface layers were following Van Wyk’s law.

Compression behaviour of cotton woven fabrics in low load regions was studied by Hu

et al. [52]. They expressed the pressure thickness relationship of woven cotton fabrics

by a mathematical expression in the low load regime. In their research, they observed

that the equation proposed by Van Wyk [27] can be applied to cotton woven fabrics,

which was previously applied to the compression of wool fabrics by De Jong et al. [21].

A detailed study of the fabric compression behaviour by using a KES compression

tester and microscope images of fabric samples was carried out by Matsurdaira et al.

[53]. On the basis of their research, they divided the fabric pressure thickness curve into

three parts. The first and third steps of the fabric compressional curve were seen to obey

a linear relationship, whereas the second part of the curve was following an exponential

relationship. They attributed the first step of the curve to the bending of the fibres on the


33

fabric surface, due to which the curve behaved linearly. This typical behaviour is shown

in Figure 2.1.

Figure 2.11. Typical pressure thickness curve

The second step of the compression curve, which has an exponential behaviour, was

related with the hardness in compression due to friction between fibres. The third linear

step of the compressional curve was attributed to the fibre material and was explained

by the initial lateral compressional modulus of the fibres.

Pearce et al. [24] conducted experimental work on plain woven glass fibre

reinforcement to study the compressibility of plain fabric. They fitted the loading cycle

response of the fabric to the power law relationship and the relaxation cycles were fitted

to an exponential decay function. They also observed that there was more reduction in

thickness of single layers compared to the average layer thickness of a 5 layer stack.

They concluded that the reason for higher single layer thickness reduction was due to

fabric-to-fabric interaction rather than fabric-to-platen interactions.


34

Potluri et al. [15] characterised the different structural changes, which take place during

compression of fabric as shown in Figure 2.1. According to this, during the initial stage

of compression the gap between the plates and the fibre surface gradually reduces to

zero, which results in a decrease in crimp in one set of yarns and an increase in the

second set of yarns. When the gap between the plates and the fabric surface is zero, any

additional compression results in yarn flattening accompanied by crimp reduction in the

crossing yarns. This process of yarn flattening continues until the fibres are maximally

packed into each other.

Figure 2.12. Stages of fabric compression [15]

The compaction of fabrics is affected by different processing parameters. After resin

infusion during composite manufacturing, the dry fabric changes to a wet state. The

compaction behaviour of wet fabrics is different from that of dry fabrics. There is more

thickness reduction of the wet fabric compared to dry fabric due to compaction as

studied by various researchers [18, 25, 29, 38, 39, 54-60].


35

Robitaille et al. [18] performed experimental work on woven fabric to study the

compaction and relaxation behaviour of dry and wet fabrics. Distilled water was used to

saturate the fabric samples to study the wet compaction. They concluded that saturation

of fabric samples with water plays an important role on the compaction behaviour.

Kelly et al. [38] studied the response of dry and wet fibrous materials during

compaction. In their research, they performed experimental work on glass fibre

continuous filament mats. They observed the different compression responses of dry

and wet reinforcements.

Saunders et al. [25] studied the compaction response of glass woven fabric samples

after application of different resins. They applied three types of resin to the fabric

surface with higher to lower viscosity and observed the change in the thickness of the

layers. They concluded that there was no significant effect of the difference of the resin

viscosity on the compaction behaviour of the fabrics.

The compaction behaviour of fibrous materials is also affected by loading time as

observed by many researchers that pressure drops to maintain a specific preform

thickness under compression loading with time or thickness drops to maintain a

constant load with time [18, 22, 26, 29, 38, 39, 61].

This behaviour is depicted in Figure 2.13. As can be seen from this figure there is a

drop in thickness or stress with time to maintain constant pressure or constant thickness.


36

Figure. 2.13 Schematic representation of time dependent behaviour of fibrous

materials

Gutowski et al. [51] studied the elastic deformation of lubricated carbon bundles. In

their work they plotted the experimental results of fibre volume fractions against the

time scale of the experiment and observed that there was strong correlation between

fibre volume fraction and time. The experimental results showed that higher amounts of

fibre volume fraction can be achieved by longer processing times.

Kim et al. [22] performed experimental work on fibrous reinforcements to study the

compression and relaxation behaviour of reinforcement material. In their research, they

studied E-glass, graphite cloth, mat, unidirectional material and combinations of two

different fibre orientations. They observed that there was stress relaxation during

compression. They related the amount of stress relaxation to fibre orientation and

observed that relaxation behaviour decreased with fibre alignment. They explained the

relaxation behaviour of fibre reinforcements during compression with the Maxwell-

Wiechert viscoelastic model and fitted their data using this model and observed good

conformity between the model and experimental results.

Time Time

Th

ick

nes

s

Str

ess


37

Francois et al. [16] collected the experimental results of compression and relaxation

from several previous researchers and fitted the relaxation data using a power law

equation in which pressure was taken as a function of time as in equation 2.1.

(2.1)

In this equation P is the final compaction pressure, P0 is the initial compaction pressure,

C is the pressure decay after 1 second and D is the relaxation index.

Luo et al. [23] studied the compression and relaxation response of a new sandwich

textile preform for liquid composite moulding and fitted the experimental data with the

power law equation in the following form as in equation 2.2.

(2.2)

Where, P and P0 are the compression pressure before relaxation and 1000 seconds after

relaxation respectively. is the relaxation index and is defined by the following

equation 2.3.

(2.3)

During the comparison study of the dry and wet fabric samples different wetting agents

were used to wet out the fabric samples. Glucose syrup was used by Kelly et al. [38] to

wet the glass fibre filament mat samples. Bickerton et al. [29] used a non-reactive corn

syrup solution as the test fluid for the wet test experiments. Robitaille et al. [18]

performed experimental work on different fabrics in dry and wet conditions to study

their compaction response. They used distilled water to saturate the fabric samples.


38

Lawrence et al. [54] studied the compression of the fibrous preform materials by using

experimental and simulation tools. In the research they studied the wet preform samples

saturated with water and oil for the compression behaviour. They observed that in water

saturated samples, there were no viscous forces present. They suggested that by using

water as a lubricating fluid, the force exerted on the fabric can be directly measured.

The fibre volume fraction of the preform during compaction loading can be varied by

changing the cross-head movement of the compression machine. It was noticed by

different researchers that cyclic compaction increases the fibre volume fraction of the

preform by decreasing the fabric thickness [19, 22, 39].

Kruckenberg et al. [39] studied the effect of vibration compaction on plain woven

fabrics. They observed that vibration compaction plays a significant role in the

compaction response of plain woven fabrics.

It was noticed by Kim et al. [22] that there was a change in the thickness of the fibre

reinforcement with change in the tensile machine cross-head speed. The thickness of the

preform was seen to decrease with slower speed. They discussed this behaviour with the

rearrangement of the fibres with a prolonged period at a slow speed. Saunders et al. [25]

performed compression testing on different woven fabrics and observed that there was

no significant effect of machine speed on the thickness of the fabrics in the dry state but

they observed significant effects of machine speed on the compression of wet samples.

Compaction behaviour of single layer and multilayer preforms was also studied by

using theoretical models. Chen et al. [30] developed a 3D model of plain woven fabric

to predict the compressive behaviour of single layer preforms. For this model, the yarns

were treated as transversely isotropic solids. It was assumed that there were no voids


39

and gaps between the yarns and there is deformation only in the yarn shape without

change in the yarn cross-sectional area. Using this model, they established analytical

expressions for the fibre volume fraction, the applied compressive force and the preform

thickness.

The micromechanical model of Chen et al. [34] describes the compaction behaviour of

single layer fabrics with plain, twill and satin weaves. The deformation mechanism was

studied at two different levels including micro deformation of yarn cross-section and

macro deformation of yarn bending along with yarn waveform flattening. The

relationship between thickness and pressure for plain, twill and satin fabrics was also

investigated. They observed that single layer plain weave fabric is most difficult to

compact whereas the single layer satin fabric is the easiest to compact. The least

deformation of single layer plain fabric was attributed to the presence of its highest ratio

of curved parts to the straight parts due to which there was more contribution of macro-

bending deformation in resisting compaction of single layer plain weave fabric. In

addition it was shown that compression behaviour of single layer fabric was affected by

the initial fibre packing ratio of the yarns.

An analytical model for compaction of multilayer woven preforms was presented by

Chen et al. [31] in which they established a relationship between fibre volume fraction,

the applied compressive force and the preform thickness for nested and non-nested

cases.

A micromechanical model was developed by Chen et al. [35] to investigate the

compaction behaviour of multilayer plain woven fabric preforms. The deformation

mechanism of fabrics at different hierarchical levels was studied. The deformation and


40

compaction of yarn cross-section, the flattening of the yarn waveform, nesting between

adjacent layers and inter-layer packing were taken into consideration to develop this

model. They showed in their model that the nesting of the layers is affected by the

number of layers. Also the shifting of the layers was shown as an important factor in

defining the compaction behaviour of the multilayer preforms.

The compaction behaviour of multilayer preforms differs from single layer preforms

due to the presence of nesting in the multilayer preforms. Nesting of the layers in a

multilayer preform is a geometrical and mechanical phenomenon which influences the

thickness of the preform during the composite manufacturing process and the

mechanical properties of the composite [44]. The study of the compaction of multilayer

preforms has been investigated by a number of researchers [15, 20, 24, 25, 31-33, 35, 39,

40, 44, 48, 57, 62, 63].

Nesting of the layers reduces the gaps between neighbouring layers in the stack and

increases the gap between the mould and the topmost layer during composite

manufacturing [64]. The effect of the nesting on the permeability of the layers was

studied by various researchers [62, 64-66].

Karahan [67] performed a damage study on 2/2 twill woven carbon multilayer

composites and observed that the nesting of the layers reduces the delamination of the

layers in the multilayer stack.

Tekalur et al. [68] studied the mechanical behaviour of glass and carbon composites.

The glass and carbon composites were subjected to quasi-static and high strain loading

to study the tensile, compressive and shear properties of the composites. They observed

that the inter-laminar shear properties of the glass composites were better than for the


41

carbon composites. This was due to higher levels of nesting in glass composites

compared with carbon composites.

Ebraheem et al. [69] developed a geometrical model for plain woven fabric to analyse

the inter-laminar shear distribution between nested layers of composites. By using this

geometrical model they concluded that nesting caused a reduction in the inter-laminar

shear stress and compaction of nested layers reduced the variation in the stress

distribution within the layers.

Hoes et al. [66] performed experimental work on plain woven glass fabric to study the

effect of nesting on the permeability scatter using electrical resistance sensors. In their

research, they studied the scatter in permeability with minimum, random and maximum

nested cases. They came to the conclusion that the nesting of the individual layers was

the major source of the scatter of the permeability measurement data.

Nesting of the layers in a multilayer stack depends upon different parameters. It was

observed by Kruckenberg et al.[39] that nesting of the layers depends upon the shifting

of the layers.

Potluri et al. [15] described cross-sectional shape of yarn, yarn spacing, applied pressure

and surface waviness as parameters which effect the nesting of the layers.

Lomov et al. [44] established a geometrical model for the nesting of the layers in textile

laminates. They studied the layer nesting in multilayer preform by varying the different

parameters. The effects of flatness of the yarns, tightness and balance of the fabric, the

number of layers in the multilayer stack and the effect of weaving/braiding and knitted

pattern was studied for the nesting phenomenon. They observed that higher thread

spacing in the fabric enhanced the nesting of the layers in the stack and vice versa.


42

Higher nesting was seen with a decrease in the float length of the weave. In the case of

non-crimp stitched fabric, they defined the nesting by the stitching pattern.

Nesting of the layers in the multilayer stack affect the stack layer thickness [20, 25, 31,

33, 41, 44, 57] and the permeability of the preform [62, 66].

Nesting of the layers was calculated by different researchers in terms of nesting factor

or nesting coefficients [15, 41, 70, 71]. Potluri et al.[15, 70] defined the nesting factor

by equation 2.4.

∑

(2.4)

In equation 2.4, higher the nesting factor, the lower will be the nesting of the layers and

vice versa. They described that the nesting factor as 1 if the layers sit exactly on top of

each other without nesting. But in actual cases, the layers do not sit exactly upon each

other due to shifting of the layers so the nesting factor will be less than 1.

By using equation 2.4, they calculated the nesting factor for two layer stacks of glass

plain woven fabric. They observed that the nesting factor at the start of the compression

without any pressure was 0.83, which decreased to 0.77 with the application of 5 kPa

pressure [15] which shows nesting increases with increase in pressure.

Karin et al.[71] studied the nesting of the layers in biaxial and triaxial braids. They used

the thickness of a single layer and the laminate thickness for the calculation of the


43

nesting factor. The nesting factors for biaxial and triaxial braids were calculated by

using the equations 2.5 and 2.6.

(2.5)

(2.6)

In these equations, nl is the number of layers in the laminate, tSLb is the single layer

thickness of biaxial braid, tSLt is the single layer thickness of triaxial braid and tlam is the

laminate thickness. The single layer thickness of biaxial and triaxial braids was

described by the thickness of the crossing yarns.

Lomov et al. [41] calculated the nesting of the multilayer stack in terms of the nesting

coefficient by using the following equation 2.7.

(2.7)

In the above equation, N is the number of plies, t is the ply thickness and p is the

pressure applied. In this equation, the higher the nesting coefficient, the higher will be

the nesting of the plies and vice versa. In their experimental work, they calculated the

nesting coefficients for 4 layer fabrics. The observed higher nesting coefficient for the

plain woven fabric (16.2 % ), whereas the nesting coefficient for the satin fabric was

5.2 %, which shows that plain fabrics have better nesting compared to satin fabrics.

Yurgartis et al. [63] developed a method to quantify yarn shape and nesting in plain

woven composites. In their method they used the angle match data to characterize the

nesting of the layers.


44

Kruckenberg et al. [39] performed work on fibreglass plain woven fabrics and studied

the nesting of the layers. In their work, they studied the nesting of the layers using the

microstructure of the composite laminates and nesting shift was calculated using the

following equation 2.8.

(2.8)

In equation 2.8, ‘a’ is half of the yarn width, ‘d’ is the tow shift and ‘s’ is the spacing

between the yarns. They calculated these values from the microstructure images of the

laminates.

As already discussed the compaction and nesting of the layers affect the resin

permeability during composite manufacturing. The resin permeability has been studied

extensively by a number of researchers using experimental and simulation tools [36, 40,

64, 65, 72-88]. The resin flow during composite manufacturing depends on different

parameters. Grujicic et al. [62] studied the effect of shear, compaction and nesting of the

layers on the permeability of orthogonal plain woven fabrics. According to them, the

resin flows mainly through the pores and channels or inter-tow voids in the

neighbouring layers. Yu and Lee [83] developed a simple in-plane permeability model

for textile fabrics. They calculated the flow in the channels or gaps between fibre tows

by using the one-dimensional Stokes equation and used the one-dimensional Brinkman

equation for flow in the fibre tows. From the analysis of their research results, they

concluded that the permeability of fibre preforms is mainly determined by gaps or

channels between the fibre tows whereas the effect of intra tow voids on the

permeability of the preform is negligible. It was emphasised by Chen et al. [33] that


45

the understanding the existence of resin channels is essential for making resin

simulation tools.

Hoes et al. [66] studied the effect of layer nesting on the resin permeability of the

woven glass fabrics and concluded that layer shifting and nesting in multilayer preforms

are the major sources of variations in permeability values.

The effect of compaction on tow geometry and voids during composite manufacturing

has been studied by different researchers [17, 20, 25, 33, 39, 89, 90]. Potluri et al. [17]

developed a novel stress freezing technique for studying the compressional behaviour of

woven fabrics. In their method, fabric impregnated with clear resin was subjected to

compressive stress and held under load allowing the resin to cure. The cured fabric

samples were analysed using a scanning electron microscope (SEM) and different tow

geometry parameters were investigated. They observed a decrease in crimp percentage

and crimp amplitude of the tows while an increase was observed in yarn spacing on

application of pressure.

Saunders et al. [20] studied the microstructure of plain woven fabrics under

compression loading. The resin impregnated fabric samples were cured and studied

using the SEM. The study contains 20 ply laminate samples compressed at different

pressure levels (4.4 kPa to 1768 kPa). On the basis of their analysis, they concluded that

the compression of woven cloths can be modelled as three different modes. The first

compression mode was dominated by the nesting of the layers in which layers came

closer to each other. In the second mode of compression, the yarns were deformed by

yarn amplitude reduction and the thickness of the individual plies reduced. In the third

mode of compression, they observed that fibrous yarns were deformed and compressed


46

individually. According to this study, the first mode of compression was dominant over

a wide range of pressures in the low and intermediate pressure regions while the second

and third modes were significant at higher pressures .

Kruckenberg et al. [39] studied the microstructure of plain woven composite laminates.

In this research, they applied vibration compaction to four ply plain woven fabric

samples at different loads and cured the resin impregnated samples in order to study the

microstructure of the fabric. They observed that nesting of the layers is dominant at low

loads. The yarn width remained constant throughout the compaction and there was

decrease in yarn height.

Saunders et al. [25] performed microstructural analysis on laminates of plain, twill and

satin fabrics by using SEM. They studied average area porosity (resin rich areas), areas

pore structure and voids for all the fabrics.

Chen et al. [33] conducted experimental and theoretical studies on the compaction

behaviour of fabric preforms in the resin transfer moulding process. The research was

conducted on continuous strand mats, plain woven fabrics and unidirectional knitted

materials. From analysing the results, several factors were identified which were the

main contributors to the preform compaction. These factors are presented in Figure 2.14.


47

Figure 2.14. Main factors affecting the compaction behaviour of preform [33]

Different image analysis techniques are being used for microstructural study of fibre

composite materials. A review of different computer based techniques for image

analysis of fibre reinforced composites can be seen in literature [91].

Schell et al. [92] conducted a study of the meso-structure analysis of glass fibre

reinforced polymers by using computed tomography (CT). They studied the resin

channels in fibre reinforced composites and analysed the voids in the composite

samples by selecting the voids by a threshold method.

Computed tomography is a non-destructive technique, which can provide geometrical

features of the internal structure of an object. CT was initially used in the biomedical

industry [93]. The use of CT became popular in the composites industry due to its

ability to study the objective in 3D. Several researchers have used CT for their research

work in the composites field [45, 89, 92, 94-122].

Desplentere et al. [106] used CT to characterise the micro-structural variation of 3D

fabrics. They determined the accuracy of the CT technique by comparing the measured

values of yarn thickness, width and spacing between the yarns obtained from CT with


48

measured values of the same parameters obtained from optical micrographs of the same

textiles. They observed that the difference between the data obtained by these two

techniques was not significant. They concluded that CT is a reliable technique to obtain

the input for modelling textiles or textile composites.

Schell et al. [92] studied glass fibre reinforced polymer samples to investigate the

geometry of the fibre bundles and voids in the samples using micro CT. They described

CT as an appropriate technique to investigate the fibre bundle structure.

Little et al. [123] conducted a study to compare the accuracy and reliability of various

void characterisation techniques in composite materials. They studied voids in the

samples of carbon fibre reinforced composites using microscopy, Archimedes density

measurements and micro CT. After their study they concluded that micro CT analysis is

the most accurate and reliable technique for characterising voids in composite materials.

Kastner et al. [124] used CT for the characterisation of carbon fibre reinforced

composites and compared the results with other different segmentation methods

including ultrasonic testing and acid digestion. They observed that X-ray CT is a

powerful non-destructive technique for characterisation and especially an excellent tool

for measuring the volumetric porosity of carbon fibre reinforced composites. They also

mentioned that CT is not only limited to measuring the porosity but additionally the size,

shape and position of the pores can also be studied using this technique. Finally they

concluded that CT can be used as a substitute method for non-destructive measurement

of porosity in composites.

Computed tomography has limitation that it requires a rigid sample that does not move

during the CT scan. In some cases, particularly when going to very high resolution,


49

special care is needed to achieve this, e.g. by consolidation of fixation of the sample.

This can be considered as destruction of the original sample. Due to variety in sample

size, shape and composition, no fixed and generally accepted protocols exist for CT

scanning. A large number of free parameters can be chosen arbitrary, such as tube

voltage , exposure time, etc., all influencing the final results and making this technique

more manual dependent [125]. Additionally the operational cost of this technique is

high.

References

[1] Textile Terms and Definitions. The Textile Institute.

[2] Badawi SS. Development of the weaving machine and 3D woven spacer fabric

structures for lightweight composites materials. Dr-Ing Dissertation. 2007.

[3] Robinson A.T.C MR. Woven cloth construction. The Textile Institute. 1973.

[4] Horrocks A R ASC. Hand book of technical textile. The Textile Institute.

[5] http://texgen.sourceforge.net/index.php/Main_Page.

[6] http://www.texeng.co.uk/orthogonal.html.

[7] Plymouth Uo. Composites design and manufacture.

http://wwwtechplymacuk/sme/mats324/.

[8] Lomov SV. Non-crimp fabric composites: manufacturing, properties and

applications. Woodhead publishing.

[9] Gopalakrishnan D B, Ramakrishnan V , Sabarinath T. Braiding & narrow width

fabrics. The Indian Textile Journal. 2007.

http://texgen.sourceforge.net/index.php/Main_Page

http://www.texeng.co.uk/orthogonal.html

http://wwwtechplymacuk/sme/mats324/


50

[10] <http://www.carbonfiberguru.com/> carbon fibre processing part 3 of 12-Resin

transfer molding ( accessed: 12 February 2014).

[11] Astrom BT. Manufacturing of polymer composites. 1997.

[12] <http://www.netcomposites.com/guide/vacuum-bagging/53>, Vaccum bagging,

( accessed 12 Feburary 2014).

[13] <http://www.netcomposites.com/guide/autoclave-design/59> Autoclave design,

(accessed: 12 Feburary 2014).

[14] <http://www.rampf-tooling.de/en/products-solutions/composites-a-fiber-

composite-material-with-a-promising-future/raku-tool-composites-production-

processes/>, Autoclav process, ( accessed 12 Feburary 2014)

[15] Potluri P, Sagar TV. Compaction modelling of textile preforms for composite

structures. Composite Structures. 2008;86(1-3):177-85.

[16] Robitaille F, Gauvin R. Compaction of textile reinforcements for composites

manufacturing. I: Review of experimental results. Polymer Composites.

1998;19(2):198-216.

[17] Potluri P, Wilding MA, Memon A. A novel stress-freezing technique for studying

the compressional behavior of woven fabrics. Textile Research Journal.

2002;72(12):1073-8.


manufacturing. II: Compaction and relaxation of dry and H2O-saturated woven

reinforcements. Polymer Composites. 1998;19(5):543.

http://www.carbonfiberguru.com/

http://www.netcomposites.com/guide/vacuum-bagging/53%3e

http://www.netcomposites.com/guide/autoclave-design/59

http://www.rampf-tooling.de/en/products-solutions/composites-a-fiber-composite-material-with-a-promising-future/raku-tool-composites-production-processes/%3e




51


manufacturing. III: Reorganization of the fiber network. Polymer Composites.

1999;20(1):48.

[20] Saunders RA, Lekakou C, Bader MG. Compression and microstructure of fibre

plain woven cloths in the processing of polymer composites. Composites Part A:

Applied Science and Manufacturing. 1998;29(4):443-54.

[21] De Jong S, Snaith JW, Michie NA. A mechanical model for the lateral compression

of woven fabrics. Textile Research Journal. 1986;56(12):759-67.

[22] Kim YR, McCarthy SP, Fanucci JP. Compressibility and relaxation of fiber

reinforcements during composite processing. Polymer Composites. 1991;12(1):13-9.

[23] Luo Y, Verpoest I. Compressibility and relaxation of a new sandwich textile

preform for liquid composite molding. Polymer Composites. 1999;20(2):179-91.

[24] Pearce N, Summerscales J. The compressibility of a reinforcement fabric.

Composites Manufacturing. 1995;6(1):15-21.

[25] Saunders RA, Lekakou C, Bader MG. Compression in the processing of polymer

composites 1. A mechanical and microstructural study for different glass fabrics and

resins. Composites Science and Technology. 1999;59(7):983-93.

[26] Seung Hwan L, Jeong Hwa H, Seong Yun K, Jae Ryoun Y, Young Seok S.

Compression and relaxation behavior of dry fiber preforms for resin transfer molding.

Journal of Composite Materials.44(15):1801-20.

[27] Van Wyk CM. Note on the compressibility of wool. Journal of the Textile Institute

1946;37(12):T285-92.


52

[28] Acheson JA, Simacek P, Advani SG. The implications of fiber compaction and

saturation on fully coupled VARTM simulation. Composites Part A: Applied Science

and Manufacturing. 2004;35(2):159-69.

[29] Bickerton S, Buntain MJ, Somashekar AA. The viscoelastic compression behavior

of liquid composite molding preforms. Composites Part A: Applied Science and

Manufacturing. 2003;34(5):431-44.

[30] Chen B, Chou T-W. Compaction of woven-fabric preforms in liquid composite

molding processes: single-layer deformation. Composites Science and Technology.

1999;59(10):1519-26.

[31] Chen B, Chou T-W. Compaction of woven-fabric preforms: nesting and multi-

layer deformation. Composites Science and Technology. 2000;60(12-13):2223-31.

[32] Chen B, Cheng AHD, Chou TW. A nonlinear compaction model for fibrous

preforms. Composites Part A: Applied Science and Manufacturing. 2001;32(5):701-7.

[33] Chen B, Lang EJ, Chou T-W. Experimental and theoretical studies of fabric

compaction behavior in resin transfer molding. Materials Science and Engineering A.

2001;317(1-2):188-96.

[34] Chen Z-R, Ye L, Kruckenberg T. A micromechanical compaction model for woven

fabric preforms. Part I: Single layer. Composites Science and Technology.

2006;66(16):3254-62.

[35] Chen Z-R, Ye L. A micromechanical compaction model for woven fabric preforms.

Part II: Multilayer. Composites Science and Technology. 2006;66(16):3263-72.

[36] Endruweit A, Long AC. Analysis of compressibility and permeability of selected

3D woven reinforcements. Journal of Composite Materials. 2010;44(24):2833-62.


53

[37] Kelly PA. A viscoelastic model for the compaction of fibrous materials. Journal of

The Textile Institute.102(8):689-99.

[38] Kelly PA, Umer R, Bickerton S. Viscoelastic response of dry and wet fibrous

materials during infusion processes. Composites Part A: Applied Science and

Manufacturing. 2006;37(6):868-3.

[39] Kruckenberg T, Ye L, Paton R. Static and vibration compaction and microstructure

analysis on plain-woven textile fabrics. Composites Part A: Applied Science and

Manufacturing. 2008;39(3):488-502.

[40] Liu XL, Falzon PJ, Sweeting R, Paton R. Effective compressibility and

permeability of multi-layer non-crimp fiberglass reinforcements. Journal of Reinforced

Plastics and Composites. 2004;23(8):861-79.

[41] Lomov SV, Gorbatikh L, Kotanjac Z, Koissin V, Houlle M, Rochez O, et al.

Compressibility of carbon woven fabrics with carbon nanotubes/nanofibres grown on

the fibres. Composites Science and Technology. 2011;71(3):315-25.

[42] Lomov SV, Gorbatikh L, Verpoest I. Compression behaviour of a fibre bundle with

grafted carbon nanotubes. Carbon. 2011;49(13):4458-65.

[43] Lomov SV, Verpoest I. Compression of woven reinforcements: A mathematical

model. Journal of Reinforced Plastics and Composites. 2000;19(16):1329-50.

[44] Lomov SV, Verpoest I, Peeters T, Roose D, Zako M. Nesting in textile laminates:

geometrical modelling of the laminate. Composites Science and Technology.

2003;63(7):993-1007.

[45] Mahadik Y, Hallett SR. Effect of fabric compaction and yarn waviness on 3D

woven composite compressive properties. Composites Part A: Applied Science and

Manufacturing.42(11):1592-600.


54

[46] Potluri P, Parlak I, Ramgulam R, Sagar TV. Analysis of tow deformations in textile

preforms subjected to forming forces. Composites Science and Technology.

2006;66(2):297-305.

[47] Yang B, Kozey V, Adanur S, Kumar S. Bending, compression, and shear behavior

of woven glass fiber–epoxy composites. Composites Part B: Engineering.

2000;31(8):715-21.

[48] Nguyen QT, Vidal-Sallé E, Boisse P, Park CH, Saouab A, Bréard J, et al.

Mesoscopic scale analyses of textile composite reinforcement compaction. Composites

Part B: Engineering. 2013;44(1):231-41.

[49] Williams CD, Grove SM, Summerscales J. The compression response of fibre-

reinforced plastic plates during manufacture by the resin infusion under flexible tooling

method. Composites Part A: Applied Science and Manufacturing. 1998;29(1–2):111-4.

[50] Gutowski T, Kingery J, Boucher D. Experiments in composites consolidation: fiber

deformation. ANTEC 86. 1986:1316-20.

[51] Gutowski TG, Dillon G. The elastic deformation of lubricated carbon fiber bundles:

Comparison of theory and experiments. Journal of Composite Materials.

1992;26(16):2330-47.

[52] Hu J, Newton A. Low-load lateral-compression behaviour of woven fabrics.

Journal of the Textile Institute. 1997;88(3):242-54.

[53] Matsudaira M, Qin H. Features and mechanical parameters of a fabric's

compressional property. Journal of the Textile Institute. 1995;86(3):476-86.

[54] Lawrence JMS, Pavel; Frey, Pierre; Bhat, Prabhas; Gebauer, Timo; Advani, Suresh

G. The compaction behavior of fibrous preform materials during the VARTM infusion.

10th Esaform conference on material forming 2007;907:1039-45.


55

[55] Francucci G, Vázquez A, Rodríguez ES. Key differences on the compaction

response of natural and glass fiber preforms in liquid composite molding. Textile

Research Journal. 2012;82(17):1774-85.

[56] Debnath S, Madhusoothanan M. Studies on compression properties of polyester

needle-punched nonwoven fabrics under dry and wet conditions. Journal of Industrial

Textiles. 2012;41(4):292-308.

[57] Hammami A. Effect of reinforcement structure on compaction behavior in the

vacuum infusion process. Polymer Composites. 2001;22(3):337-48.

[58] Yenilmez B, Senan M, Murat Sozer E. Variation of part thickness and compaction

pressure in vacuum infusion process. Composites Science and Technology. 2009;69(11-

12):1710-9.

[59] Yuexin D, Zhaoyuan T, Yan Z, Jing S. Compression responses of preform in

vacuum infusion process. Chinese Journal of Aeronautics. 2008;21(4):370-7.

[60] Francucci G, Rodríguez ES, Vázquez A. Experimental study of the compaction

response of jute fabrics in liquid composite molding processes. Journal of Composite

Materials. 2012;46(2):155-67.


reinforcements during composite processing. Polymer Composites. 1991;12(1):13.

[62] Grujicic M, Chittajallu KM, Walsh S. Effect of shear, compaction and nesting on

permeability of the orthogonal plain-weave fabric preforms. Materials Chemistry and

Physics. 2004;86(2-3):358-69.

[63] Yurgartis SW, Morey K, Jortner J. Measurement of yarn shape and nesting in

plain-weave composites. Composites Science and Technology. 1993;46(1):39-50.


56

[64] Dungan FD, Senoguz MT, Sastry AM, Faillaci DA. Simulations and experiments

on low-pressure permeation of fabrics: Part I-3D modeling of unbalanced fabric. Journal

of Composite Materials. 2001;35(14):1250-84.

[65] Simacek P, Advani SG. Permeability model for a woven fabric. Polymer

Composites. 1996;17(6):887-99.

[66] Hoes K, Dinescu D, Sol H, Parnas RS, Lomov S. Study of nesting induced scatter

of permeability values in layered reinforcement fabrics. Composites Part A: Applied

Science and Manufacturing. 2004;35(12):1407-18.

[67] Karahan M. Investigation of damage initiation and propagation in 2 × 2 twill

woven carbon/epoxy multi-layer composites. Textile Research Journal. 2011;81(4):412-

28.

[68] Tekalur SA, Shivakumar K, Shukla A. Mechanical behavior and damage evolution

in E-glass vinyl ester and carbon composites subjected to static and blast loads.

Composites Part B: Engineering. 2008;39(1):57-65.

[69] Shady E, Gowayed Y. Interlaminar shear stress distribution between nested layers

of plain weave composites. Polymer Composites. 2010;31(11):1838-45.

[70] Potluri P, Manan A, Francke M, Day RJ. Flexural and torsional behaviour of

biaxial and triaxial braided composite structures. Composite Structures. 2006;75(1–

4):377-86.

[71] Birkefeld K, Röder M, Von Reden T, Bulat M, Drechsler K. Characterization of

biaxial and triaxial braids: Fiber architecture and mechanical properties. Applied

Composite Materials.1.

[72] Adams KL, Rebenfeld L. In-plane flow of fluids in fabrics: Structure/Flow

characterization. Textile Research Journal. 1987;57(11):647-54.


57

[73] Chen R, Dong C, Liang Z, Zhang C, Wang B. Flow modeling and simulation for

vacuum assisted resin transfer molding process with the equivalent permeability method.

Polymer Composites. 2004;25(2):146-64.

[74] Chen Z, Ye L, Liu H. Effective permeabilities of multilayer fabric preforms in

liquid composite moulding. Composite Structures. 2004;66(1-4):351-7.

[75] Gebart BR. Permeability of unidirectional reinforcements for RTM. Journal of

Composite Materials. 1992;26(8):1100-33.

[76] Griffin PR, Grove SM, Russell P, Short D, Summerscales J, Guild FJ, et al. The

effect of reinforcement architecture on the long-range flow in fibrous reinforcements.

Composites Manufacturing. 1995;6(3–4):221-35.

[77] Mogavero J, Advani SG. Experimental investigation of flow through multi-layered

preforms. Polymer Composites. 1997;18(5):649-55.

[78] Pillai KM, Advani SG. Numerical simulation of unsaturated flow in woven fiber

preforms during the resin transfer molding process. Polymer Composites.

1998;19(1):71-80.

[79] Senoguz MT, Dungan FD, Sastry AM, Klamo JT. Simulations and experiments on

low-pressure permeation of fabrics: Part II ”The variable gap model and prediction of

permeability. Journal of Composite Materials. 2001;35(14):1285-322.

[80] Simacek P, Advani SG. Modeling resin flow and fiber tow saturation induced by

distribution media collapse in VARTM. Composites Science and Technology.

2007;67(13):2757-69.

[81] Song YS, Youn JR. Flow advancement through multi-layered preform with

sandwich structure. Composites Part A: Applied Science and Manufacturing.

2007;38(4):1082-8.


58

[82] Verleye B, Lomov SV, Long A, Verpoest I, Roose D. Permeability prediction for

the meso-macro coupling in the simulation of the impregnation stage of Resin Transfer

Moulding. Composites Part A: Applied Science and Manufacturing. 2010;41(1):29-35.

[83] Yu B, James Lee L. A simplified in-plane permeability model for textile fabrics.


[84] Xiao X, Zeng X, Long A, Lin H, Clifford M, Saldaeva E. An analytical model for

through-thickness permeability of woven fabric. Textile Research Journal.

2012;82(5):492-501.

[85] Arbter R, Beraud JM, Binetruy C, Bizet L, Bréard J, Comas-Cardona S, et al.

Experimental determination of the permeability of textiles: A benchmark exercise.

Composites Part A: Applied Science and Manufacturing. 2011;42(9):1157-68.

[86] Govignon Q, Bickerton S, Kelly PA. Simulation of the reinforcement compaction

and resin flow during the complete resin infusion process. Composites Part A: Applied


[87] Ahn KJ, Seferis JC, Berg JC. Simultaneous measurements of permeability and

capillary pressure of thermosetting matrices in woven fabric reinforcements. Polymer

Composites. 1991;12(3):146-52.

[88] Amico S, Lekakou C. An experimental study of the permeability and capillary

pressure in resin-transfer moulding. Composites Science and Technology.

2001;61(13):1945-59.

[89] Mahadik Y, Brown KAR, Hallett SR. Characterisation of 3D woven composite

internal architecture and effect of compaction. Composites Part A: Applied Science and



59

[90] Jortner J. Microstructure of cloth-reinforced carbon-carbon laminates. Carbon.

1992;30(2):153-63.

[91] Guild FJ, Summerscales J. Microstructural image analysis applied to fibre

composite materials: a review. Composites. 1993;24(5):383-93.

[92] Schell JSU, Renggli M, Van Lenthe GH, Müller R, Ermanni P. Micro-computed

tomography determination of glass fibre reinforced polymer meso-structure.

Composites Science and Technology. 2006;66(13):2016-22.

[93] Muller R. The zurich experience: One decade of three-dimensional high-resolution

computed tomography. Top Magn Reson Imaging. 2002;13(5):307-22.

[94] Tan KT, Watanabe N, Iwahori Y. X-ray radiography and micro-computed

tomography examination of damage characteristics in stitched composites subjected to

impact loading. Composites Part B: Engineering. 2011;42(4):874-84.

[95] Tiseanu I, Tsitrone E, Kreter A, Craciunescu T, Loarer T, Pégourié B, et al. X-ray

micro-tomography studies on carbon based composite materials for porosity network

characterization. Fusion Engineering and Design. 2011;86(9–11):1646-51.

[96] Seltzer R, González C, Muñoz R, Llorca J, Blanco-Varela T. X-ray

microtomography analysis of the damage micromechanisms in 3D woven composites

under low-velocity impact. Composites Part A: Applied Science and Manufacturing.

2013;45(0):49-60.

[97] McCombe GP, Rouse J, Trask RS, Withers PJ, Bond IP. X-ray damage

characterisation in self-healing fibre reinforced polymers. Composites Part A: Applied



60

[98] Schilling PJ, Karedla BR, Tatiparthi AK, Verges MA, Herrington PD. X-ray

computed microtomography of internal damage in fiber reinforced polymer matrix

composites. Composites Science and Technology. 2005;65(14):2071-8.

[99] Parnas RS, Wevers M, Verpoest I. Using textile topography to analyze X-ray CT

data of composite microstructure. Polymer Composites. 2003;24(2):212-20.

[100] Latil P, Orgéas L, Geindreau C, Dumont PJJ, Rolland du Roscoat S. Towards the

3D in situ characterisation of deformation micro-mechanisms within a compressed

bundle of fibres. Composites Science and Technology. 2011;71(4):480-8.

[101] Soltani P, Johari MS, Zarrebini M. Tomography based determination of

transverse permeability in fibrous porous media. Journal of Industrial Textiles. 2013.

[102] Badel P, Vidal-Sallé E, Maire E, Boisse P. Simulation and tomography analysis of

textile composite reinforcement deformation at the mesoscopic scale. Composites

Science and Technology. 2008;68(12):2433-40.

[103] Pandita SD, Verpoest I. Prediction of the tensile stiffness of weft knitted fabric

composites based on X-ray tomography images. Composites Science and Technology.

2003;63(2):311-25.

[104] Cosmi F, Bernasconi A, Sodini N. Phase contrast micro-tomography and

morphological analysis of a short carbon fibre reinforced polyamide. Composites


[105] Feng Y, Feng Z, Li S, Zhang W, Luan X, Liu Y, et al. Micro-CT characterization

on porosity structure of 3D Cf/SiCm composite. Composites Part A: Applied Science



61

[106] Desplentere F, Lomov SV, Woerdeman DL, Verpoest I, Wevers M, Bogdanovich

A. Micro-CT characterization of variability in 3D textile architecture. Composites


[107] Pazmino J, Carvelli V, Lomov SV. Micro-CT analysis of the internal deformed

geometry of a non-crimp 3D orthogonal weave E-glass composite reinforcement.

Composites Part B: Engineering. (0).

[108] Chateau C, Gélébart L, Bornert M, Crépin J, Boller E, Sauder C, et al. In situ X-

ray microtomography characterization of damage in SiCf/SiC minicomposites.


[109] Scott AE, Mavrogordato M, Wright P, Sinclair I, Spearing SM. In situ fibre

fracture measurement in carbon–epoxy laminates using high resolution computed

tomography. Composites Science and Technology. 2011;71(12):1471-7.

[110] Pandita SD, Falconet D, Verpoest I. Impact properties of weft knitted fabric

reinforced composites. Composites Science and Technology. 2002;62(7–8):1113-23.

[111] Wright P, Moffat A, Sinclair I, Spearing SM. High resolution tomographic

imaging and modelling of notch tip damage in a laminated composite. Composites


[112] Rinaldi RG, Blacklock M, Bale H, Begley MR, Cox BN. Generating virtual

textile composite specimens using statistical data from micro-computed tomography:

3D tow representations. Journal of the Mechanics and Physics of Solids.

2012;60(8):1561-81.

[113] Blacklock M, Bale H, Begley M, Cox B. Generating virtual textile composite

specimens using statistical data from micro-computed tomography: 1D tow


62

representations for the Binary Model. Journal of the Mechanics and Physics of Solids.

2012;60(3):451-70.

[114] Ullah H, Harland A, Silberschmidt V. Experimental and Numerical Analysis of

Damage in Woven GFRP Composites Under Large-deflection Bending. Applied

Composite Materials. 2012;19(5):769-83.

[115] Sket F, Seltzer R, Molina-Aldareguía JM, Gonzalez C, Llorca J. Determination of

damage micromechanisms and fracture resistance of glass fiber/epoxy cross-ply

laminate by means of X-ray computed microtomography. Composites Science and

Technology. 2012;72(2):350-9.

[116] Kim KJ, Yu W-R, Lee JS, Gao L, Thostenson ET, Chou T-W, et al. Damage

characterization of 3D braided composites using carbon nanotube-based in situ sensing.


[117] Scott AE, Sinclair I, Spearing SM, Thionnet A, Bunsell AR. Damage

accumulation in a carbon/epoxy composite: Comparison between a multiscale model

and computed tomography experimental results. Composites Part A: Applied Science


[118] Djukic LP, Herszberg I, Walsh WR, Schoeppner GA, Gangadhara Prusty B, Kelly

DW. Contrast enhancement in visualisation of woven composite tow architecture using

a MicroCT Scanner. Part 1: Fabric coating and resin additives. Composites Part A:


[119] Amenabar I, Mendikute A, López-Arraiza A, Lizaranzu M, Aurrekoetxea J.

Comparison and analysis of non-destructive testing techniques suitable for delamination

inspection in wind turbine blades. Composites Part B: Engineering. 2011;42(5):1298-

305.


63

[120] Hufenbach W, Böhm R, Gude M, Berthel M, Hornig A, Ručevskis S, et al. A test

device for damage characterisation of composites based on in situ computed

tomography. Composites Science and Technology. 2012;72(12):1361-7.

[121] Bull DJ, Helfen L, Sinclair I, Spearing SM, Baumbach T. A comparison of multi-

scale 3D X-ray tomographic inspection techniques for assessing carbon fibre composite

impact damage. Composites Science and Technology. 2013;75(0):55-61.

[122] Faessel M, Delisée C, Bos F, Castéra P. 3D modelling of random cellulosic

fibrous networks based on X-ray tomography and image analysis. Composites Science

and Technology. 2005;65(13):1931-40.

[123] Little JE, Yuan X, Jones MI. Characterisation of voids in fibre reinforced

composite materials. NDT & E International. 2012;46(0):122-7.

[124] Kastner J PB, Salaberger D, Sekelja J. Defect and porosity determination of fibre

reinforced polymers by X-ray computed tomography. 2nd International Symposium on

NDT in Aerospace. 2010.

[125] Cnudde V, Boone MN. High-resolution X-ray computed tomography in

geosciences: A review of the current technology and applications. Earth-Science

Reviews. 2013;123(0):1-17.

Chapter 3 Deformation of single layer dry preforms under compaction

64

Chapter 3

DEFORMATION OF SINGLE LAYER DRY

PREFORMS UNDER COMPACTION

Z.Yousaf, P.Potluri, P.Withers

Abstract

Meso-scale geometrical changes during transverse compression of a dry textile preform

have a significant influence on resin permeability during infusion as well as on the

mechanical properties of the resulting polymer composite. Here tow deformations have

been analysed in detail by x-ray computed tomography (CT) for dry fabric preforms as a

function of compressive loading. With the aid of an in-situ loading device, 3D images

have been captured at each pressure increment up to 600 kPa, to simulate the forces

applied by a range of manufacturing processes including vacuum infusion and autoclave

curing. The evolutions of meso-scale geometrical features such as tow cross-section,

waviness, and inter-tow channels have been quantified. Good contrast has allowed the

geometry of the individual tows, as well as inter-tow channels to be described by 3D

solid objects. This technique, for the first time, facilitates the capture of detailed

preform geometry under compressive loading that can be used for resin flow

simulations as well as for structural analysis.

Keywords: compaction, tow geometry, woven textile, resin infiltration, inter-tow voids


65

3.1 Introduction

Recent accelerated growth in the use of advanced composites is sustainable only

through continuous development of advanced manufacturing techniques matched by

improved design and simulation tools. Prepreg hand lay-up techniques developed

primarily for high-value markets such as military airframes and Formula 1 cars are

being gradually replaced by automated tape-laying and tow-placement techniques.

Further expansion to cost-sensitive high-volume markets, such as the passenger car

industry, will only be feasible through a significant reduction in manufacturing and

product development costs. Liquid resin infusion technologies in conjunction with dry

fibre preforming have been widely recognised as a way forward for achieving the next

level of affordable composites manufacturing. Dry textile preforms have meso-scale

features due to the interlacement of tows that affect the ease of resin infiltration and the

final properties of the manufactured composite. These meso-scale tow-level geometric

features need to be captured as a function of external loading so that they may be used

further for structural analysis and for making resin simulation tools.

For composite manufacturing, a number of different techniques are employed. These

manufacturing processes involve compaction of dry preform to a certain level of

pressure which changes the tow and the resin channel geometry of the preform affecting

the fibre volume fraction and the resin permeability. Based upon these facts, the

compaction process is taken as an important parameter of the manufacturing process

[1]. Extensive research has been performed on the compaction behaviour of different

fabrics [2-15]. It is considered that Van Wyk [3] was probably the first to treat the fibres

under compression as a system of bending units. Pearce et al. [11] performed


66

experimental work on the compressibility of the reinforcement fabric under loading and

relaxing cycles, and attributed the loading cycle response to a power-law relationship. A

typical pressure - thickness curve was proposed by Hu et al. and Mastudaira et al. [16,

17], as shown in Figure 3.1. The curve can be divided into three regimes; namely two

linear stages separated by an exponential stage. Mastudaira et al. attributed the first

stage to the bending of the fibres, the second to friction between the fibres, and the third

part to the lateral compressional modulus of the fibres themselves. According to Chen et

al. [8], the third stage of the compression curve was due to bending deformation of the

yarn. A compaction model for a single layer woven fabric preform was developed by

Chen et al. [10]. In the compaction model, micro-deformation of the yarn cross-section

and macro-deformation by yarn bending accompanied by yarn waveform flattening

were described. In addition, the effects of single layer woven fabric microstructure on

the compaction behaviour were evaluated. It was shown that the macro-bending

stiffness of the fibre and the initial fibre packing ratio of the yarn affect the compaction

behaviour of single layer fabric.

Figure 3.1 Pressure - thickness curve for a woven fabric under compaction


67

Saunders et al. [18] studied the compaction of multilayer plain woven cloths using

mechanical testing and microstructural images of the laminates. They divided the fabric

compression into three modes of deformation: the first mode being the nesting of cloths,

the second mode the deformation of the yarn waveform and the third mode as the

deformation of the yarn cross-section.

Microstructural analyses on multilayer woven preforms for different tow geometry

parameters under compaction were conducted by Saunders et al. and Kruckenberg et al.

[18-20]. They observed the nesting of the layers as the dominant factor in the initial

mode of compression for multilayer fabrics. Mahadik et al. [21-23] investigated the

effect of compaction on 3D woven fabric’s internal architecture and composite

compressive properties. They found the resin channels are highly influenced by the

weave style of the fabrics.

As already mentioned, for calculating the permeability of the reinforcements during

composite manufacturing, the study of the meso-structure of the reinforcement is

necessary. The meso-structure of the glass laminate for resin channels and inter-tow

voids was studied by Schell et al. [24] using computed tomography (CT). Saunders et

al. [19] studied the different glass fabrics under compression and provided data for

average area porosity, area pore structure and voids for different type of fabrics using

microstructure images of composites. The area pore distribution in cross-sections of

laminates was measured by Griffin et al. [25]

The deformation behaviour of single layer fabrics has been studied extensively using

mechanical methods while the microstructure of single layer fabrics has been analysed

only on composite laminates [12]. In the previous literature, no study has been reported


68

on the meso-structure of single layer dry fabrics. The laminate structure may not

represent the meso-level features of dry fabric. This limitation has been pointed out by

Schell et al. [24]. Thus in order to validate tow geometry and permeability simulations,

the meso-structure analysis of dry fabric is necessary. In the present work, the

deformation behaviour of single layer dry preforms was investigated in-situ under

increasing compressive loading using CT. The changes in tow cross-section, yarn

waviness, yarn packing fraction, yarn volume fraction and resin channels were analysed

at different pressure levels. To carry out this study, the fabric preform was compressed

in a rig comprising two clear polycarbonate plates, and then scanned by CT for meso-

structural analysis.

3.2 Material and mechanical testing

The material under observations was E-glass plain woven fabric having a warp density

of 4.8 /cm and a weft density of 4.4 /cm. The warp and the weft counts were 650 Tex,

and the density of the fibres was 2.60 g/cm3. The initial unloaded thickness of the fabric

was 0.92 mm. The scanned image of the glass fabric used is presented in Figure 3.2.

Figure 3.2 Scanned image of plain weave (1/1) fabric


69

3.2.1 Mechanical testing

An Instron 5569 machine with a 5 kN load cell was used for the mechanical testing of

the single layer fabric samples. The sample area was 15 cm2 and the cross-head speed

during testing was 1 mm/min. In glass yarn fibres tend to rearrange themselves under

load with the passage of time. Consequently, either the pressure decreases at constant

thickness or the thickness decreases at constant pressure when the fabric is kept under

load for long periods [15, 26-29]. For the mechanical testing of the fabric samples, a

static test method was used to maintain constant thickness at constant pressure. In this

method, the machine crosshead was moved to apply the desired constant pressure at a

constant speed of 1 mm/min and was held for 5 minutes at that point before moving to

the next position. The holding time of 5 minutes is considered adequate for these types

of loadings [20, 29]. At each target load, the initial and final thickness readings were

recorded. Prior to testing, the compliance curve was taken to adjust the minor

compliances during the compression. The compliance values were then removed from

the thickness readings, so that the compliance effect due to the machine was balanced.

3.3 Tomography and in-situ compression rig

A compression rig, shown in Figure 3.3, was developed to compress dry preforms in-

situ within the tomography machine. The rig comprises two clear polycarbonate plates

60 x 35 x 12 mm in length, width and thickness, respectively. Two side screws were used

to compress the plates from both sides, and two thickness gauges were placed on either

side to maintain uniformity on both sides. The edge-to-edge distance of the two side

screws was 35 mm, and the sample size of dry fabric compressed between the two plates

was 40 x 35 mm. At each pressure level, a slip gauge of known thickness was placed


70

between the two plates on each side as illustrated in Figure 3.3. The bending strength of

the plates was estimated by applying beam bending theory to ensure that there was no

significant bending of the plates under load. Once mounted on the x-ray scanner, the

side screws were tightened to compress the fabric preform to the desired pressure. From

the pressure - thickness curve, the thickness value against desired pressure was taken

and the slip gauge of thickness corresponding to that pressure was put on both sides of

the plates. Figure 3.4 presents the rig mounted on the x-ray scanner.

Figure 3.3. Schematic of the compression rig

Tightening screw Top plate

Thickness gauge

Bottom plate Dry fabric Nut


71

Figure 3.4. (a) Compression rig accommodated on the tomography stage, and (b)

close-up of the rig

The CT method involves collecting a large set of radiographs (projections) of the

sample as it is rotated through 180°. Together with a small number of calibration

images, these images are reconstructed into a 3D volume, which represents the

attenuation through the sample. A Nikon Metris 225/320 kV Custom Bay system was

used for scanning. The current and voltage were adjusted to 115 µA and 85 kV,

respectively. For each scan, a total of 3142 projections were taken. These images were

then reconstructed using Metris X-Tek CT Pro software, and the data processed using

VSG Avizo Fire 6.3 software.


72

3.4 Results and discussion

3.4.1 Macroscopic deformation

The compaction of the single layer fabric was performed in both dry and wet conditions.

For wet compaction the fabric was impregnated with water and was compacted to

measure the pressure - thickness response. The thickness results for dry and the wet

fabrics at different pressures are shown in Figure 3.5. Two curves were recorded for

each fabric, one represents the thickness immediately after the load is applied, and the

other is the thickness after 5 minutes of holding at that loading point. From Figure 3.5, it

can be seen that the compaction of the wet fabric was higher than the dry fabric at the

same pressures which is in agreement with the previous findings [30, 31]. Significant

compaction occurs over time in both cases at constant load, the change having stabilised

after 5 minutes. It is clear from Figure 3.6 that the displacement has essentially

plateaued within 5 minutes of loading for both wet and dry preforms. Unsurprisingly,

the change of thickness with time was greatest during the early stages of compaction;

this reflects the fact that the fibres relax less at the higher loads.


73

Figure 3.5. Pressure - thickness response for a single layer under dry and wet

conditions

Figure 3. 6. Effect of loading time on the thickness of single layer dry fabric

The thickness curves with time at particular pressures as illustrated in Figure 3.7 were

fitted with a power-law of the form shown in Equation 3.1:

(3.1)


74

Where t/t0 is the final thickness to initial thickness ratio, ‘T’ is the time and ‘a’ and ‘c’

are the empirical constants.

Figure 3.7. t/t0 plotted against time for different pressures

The constants ‘a’ and ‘c’ can further be varied against different pressures and it was

found that ‘a’ takes the trend of a power law with pressure while ‘c’ is almost constant.

The experimental and fitted curves of constant ‘a’ against pressure are presented in

Figure 3.8.


75

Figure 3.8. The constants plotted against pressure

The fitted power follows the form shown in Equation 3.2 for ‘a’ in the above equation

(3.2)

So the final relationship between t/to with pressure and time becomes that of Equation

3.3

(3.3)

The term ‘q’ was found to vary with the wettability of the fabric.

Figure 3.9 shows the experimental and fitted values of a single layer of dry and wet

fabrics at different pressures, which are in close agreement with each other.


76

(a)

(b)

Figure 3.9. Experimental and predicted thicknesses with time at different

pressures, (a) dry fabric and (b) wet fabric


77

3.4.2 Meso-structural analysis by computed tomography (CT)

Single layer dry fabric compressed between polycarbonate plates was scanned by CT

for its meso-structural analysis at different pressures. The parameters describing the

yarn geometry obtained from the meso-structural analysis are illustrated in Figure 3.10.

The thread spacing, P, is the distance between two consecutive yarns, the crimp

amplitude is represented by h, the crimp angle by θ, the yarn width by a, and yarn

thickness by b. The crimp percentage of the yarn was calculated by using the equation

3.4.

(3.4)

Figure 3.10. Definition of the yarn geometry parameters

The tomography technique has the advantage that the scanned structure can be

visualised over any region of interest. To investigate the behaviour of the fabric, both

the warp and the weft yarns have been extracted as virtual slices from the centre of the

tow intersections and in between two tows to study the tow at the interlacing and non-

interlacing points for both tow geometry and the void-space data as shown in Figure

3.11.


78

Figure 3.11. Segmented virtual slices of warp and weft yarns

Figure 3.12 represents the 3D structure of the fabric sample, segmented using Avizo 6.3

on the basis of simple thresholding.

Figure 3.12. 3D reconstruction of single layer fabric

In between two tows Centre of tow


79

Table 3.1. Tow geometry parameters (warp).

Pressure

(kPa)

Tow

area

(mm2)

Tow

width

(mm)

Tow

height

(mm)

Thread

spacing

(mm)

10 0.47 2.07 0.31 2.03

50 0.46 2.11 0.30 2.01

100 0.44 2.11 0.28 2.03

300 0.41 2.12 0.26 2.02

600 0.38 2.12 0.24 2.03

Table 3.2. Tow geometry parameters (weft).

Pressure

(kPa)

Tow

area

(mm2)

Tow

width

(mm)

Tow

height

(mm)

Thread

spacing

(mm)

10 0.45 1.78 0.40 2.25

50 0.44 1.77 0.39 2.26

100 0.42 1.78 0.37 2.26

300 0.39 1.79 0.34 2.25

600 0.37 1.79 0.31 2.24


80

Tomographic slices were taken at the centre of the intersections of the warp and the

weft yarns, and various tow geometry parameters of both the warp and the weft yarns

were calculated at different pressure levels as shown in Tables 3.1 and 3.2. The

corresponding cross-sectional images of the warp and the weft yarns are displayed in

Figure 3.13. The range of pressures studied varied from 10 kPa to 600 kPa to investigate

the behaviour of the geometrical changes for both vacuum infusion and autoclave

curing. During the meso-structural analysis of the compressed dry fabric preform, it was

observed that initially, the shape of the warp yarn was elliptical and the weft yarn was

lenticular. The warp yarn was highly crimped as shown in Figure 3.14a and the yarn

followed a sinusoidal path, whereas the weft yarn exhibited very low crimp as shown in

Figure 3.14b and the yarn path was nearly straight. The warp tows were in contact with

the compression plate whereas there was separation between the weft tows and the

compression plate at the initial loading. Warp tows were spreading to the available gaps

between the tows and there was a slight overlap between the warp tows at the edges of

the cross-sections .Due to the higher warp yarn crimp percentage, the crimp amplitude

and the crimp angle were also higher in the warp yarns than the weft yarns.

During the first step of compression the amplitude of the warp yarns reduced as the

pressure increased from 10 kPa to 50 kPa; by contrast, the amplitude of the weft yarns

increased as seen in Figure 3.15. Similarly, both the crimp and the crimp angle reduced

for the warp yarns and increased for the weft yarns with increasing the pressure as in

Figures 3.16 to 3.17. It should be noted that the yarn crimp reduces the compressive

strength of the laminates [32] and the magnitude of the crimp angle is important in

predicting the in-plane compressive strength of the composite. The phenomenon of

crimp behaviour may be attributed to the balancing of the warp and the weft yarns in


81

which the crimp in one yarn increases and the other yarn decreases to reach the

balancing position. This balancing of crimp was discussed in detail by Potluri et al. and

Lomov et al. [33, 34].This crimp interchange is an important phenomenon which affects

the laminate tensile moduli [35]. It is worth mentioning that the other geometrical

parameters (area, height) of dry fabric were only slightly changed by increasing the

pressure.

Weft cross-section Warp cross-section

Figure 3.13. Cross-sectional view of the preform at the centre of the tow

intersections

Figure 3.14. 3D reconstruction of yarns, (a) warp yarn, (b) weft yarn


82

Figure 3.15. The crimp amplitude of the warp and the weft yarns

Figure 3.16. The crimp percentage of warp and weft yarns


83

Figure 3.17. The crimp angle of warp and weft yarns

When the pressure was increased from 50 kPa to 100 kPa, the same phenomenon of

crimp balancing continued in which the crimp percentage, the crimp amplitude, and the

crimp angle decreased in the warp yarns and increased in the weft yarns. In all cases, the

change in the yarn height and area was not significant. Figure 3.15 indicates that for the

regime of pressure up to 100 kPa corresponding to vacuum infusion, compaction is

mainly dominated by the crimp amplitude reduction. In this region, the main factor

contributing to fabric thickness reduction was also identified, which may return to the

crimp break of warp yarns for the fabric. A similar behaviour was observed by Potluri et

al. [12] for the single layer plain woven fabric laminate in which crimp reduction was

responsible for the compression at low loads. This observation shows that the thickness

reduction in the low load regime for single layer dry fabric is different from the

multilayer laminate in which the nesting of the layers is the main factor contributing to

the thickness reduction of the laminate as observed by Saunders et al. and Kruckenberg

et al. [18, 20].


84

In cases where higher pressures (300 kPa to 600 kPa) were applied, the further reduction

in yarn amplitude and yarn crimp was insignificant, and the deformation in the tow

occurred at the current pressures as in Figures 3.15 to 3.17 and Tables 3.1& 3.2. In

contrast to the lower pressures, there was greater reduction in yarn height and yarn area.

Additionally, a slight increase was noticed in the yarn width as the pressure changed to

600 kPa. Contrary to the initial pressures, the weft yarn crimp percentage, crimp

amplitude and crimp angle were decreased at a pressure of 600 kPa as seen in Figures

3.15 to 3.17. The observation of decrease in yarn area is in contrast to the assumption of

the model made by Chen et al. [8] in which yarn shape deforms but the yarn area

remains the same. These findings indicate that after the yarn attains the balancing

condition, the crimp in both types of crossing yarns starts to decrease with the

application of further load. It should be pointed out that no significant trend was

observed for the thread spacing with the pressure increasing. The overall behaviour of

the fabric under compaction shows that applying low pressures at the start, the single

layer thickness reduction is maximised which is due to crimp interchange and at higher

pressures the thickness reduction is due to yarn deformation. This is also in agreement

with the pressure thickness curve in Figure 3.1 where a small pressure is required in the

initial mode of compression and then a higher amount of force is required even for

small amounts of thickness reduction.


85

Weft cross-section Warp cross-section

Figure 3.18. Cross-sectional images of the warp and the weft yarns in between two

tows

In addition to the centre of the interlacing points, tow cross-sections were also analysed

in between two consecutive warp and weft yarns. For this purpose, yarn slices were

taken in between two tows using Avizo software and the change in the corresponding

tow area was investigated at each pressure level. Figure 3.18 shows the images taken in

between the two tows. In the warp cross-sections only warp tows are visible whereas in

the weft cross-sections both the warp and the weft tow are visible. The reason for this

was when the tows cross-sections were taken for the warp cross-sections in between two

weft tows, due to the higher weft spacing only warp cross-sections were present at the

slices and when the weft cross-sections were taken in between the two warp yarns, the

warp yarns were overlapping on the edges and on the slices they also contributed with

the weft yarns.

As tows in between warp and weft yarns were not straight due to the crimp angle, so

when slices were taken at the centre point they gave different area values due to

additional crimp angle. As shown in Figure 3.19, the slice taken in between two tows

gave a value of bd instead of ae due to the crimp angle θ1. This crimp effect was


86

eliminated by using the trigonometric cosine function and the actual value of tow cross-

section area ( ae ) at the central point was measured.

Figure 3.19. Tow cross-section at the centre due to crimp angle

It was observed that the tow area in between the two tows was higher than the

corresponding area at the centre of the crossing point of both the warp and the weft

yarns. This may be attributed to the fact that the forming forces acting on the crossing

points are higher than those in between the two tows. This area in between the two tows

changed negligibly during the application of initial loads and decreased slightly as

pressure approached 600 kPa, giving room for more intra-tow voids and, consequently,

a lower yarn packing fraction, compared to that at the crossing point of the yarns.

The tow area calculated from the CT images was compared with the tow area calculated

by the Texgen software developed by The University of Nottingham; the figures appear

in Table 3.3. It was observed that the area calculated by Texgen and tomography differ

from each other; the reason might be that the Texgen software calculates the ideal tow


87

area whilst due to variations in the manufactured yarns, the area calculated from the CT

images is different from the Texgen area.

Table 3.3 Tow area calculated by CT and Texgen

Area (mm2)

Pressure(kPa)

By CT

By Texgen

10

0.47

0.50

50

0.46

0.50

100

0.44

0.48

300

0.41

0.44

600

0.37

0.40

3.4.3 Resin channels in dry preform

The compaction of the fabrics leads to a change in the fabric permeability [36]. This is

taken to arise because of the reduction in the inter-tow voids or the resin channels; while

the intra-tow voids are believed to exert a negligible effect [37]. Because of the

importance of the fabric permeability, several models have been developed for both

single and multi-layer fabrics [2, 37-39].

Studying the resin channels or inter-tow voids in dry fabrics is significant in developing

accurate permeability models. In the current work, the resin channels were examined in

detail at different pressure levels, and the corresponding behaviour analysed.

Volumetric as well as areal channels were analysed for both warp and weft cross-

sections, as a function of the applied pressure.


88

As seen from Figure 3.20, the inter-tow void percentage decreases exponentially with

increasing the pressure from 10 kPa to 600 kPa. The reduction is steep over the pressure

range 10-100 kPa, which corresponds to the vacuum infusion region; while in the

pressure range 300-600 kPa, the reduction was insignificant. Perhaps surprisingly, the

void percentage was markedly higher in between two tows than at the centre of the

intersections of the warp and the weft cross-sections as illustrated in Figures 3.21 &

3.22, giving larger passages for resin flow in between two tows rather than at the centre

of the intersecting point of the tows. Further, the void percentage was higher for the

weft cross-section at the centre of the tow intersections than that in the warp cross-

section. This might be due to the higher number of warp yarns than weft yarns.

However, the reverse behaviour was seen for the void percentage in between the tow

where the void percentage was larger in magnitude in the warp cross-section compared

with the weft cross-section, suggesting better resin transfer in the warp cross-section

than in weft cross-section. The reason for these larger inter-tow voids in between the

tows in the warp cross-section was that when the slices were taken in between two weft

tows in the warp cross-section, only warp yarns were present there due to the greater

yarn spacing between the weft yarns while on the other side, when the yarn slices were

taken for the weft cross-sections in between two warp yarns, the weft slices at that point

also contained warp yarns which were overlapping at the edges of each other.


89

Figure 3.20. Inter-tow voids in 3D structure

Figure 3.21. Inter-tow voids at the centre of the tow intersections


90

Figure 3.22. Inter-tow voids in between two tows

The thickness of the inter-tow voids and its full distribution were measured at different

pressure levels as shown in Figures 3.23 & 3.24. Similarly to the void percentage-

pressure relationship, the void thickness decreased exponentially with increasing

applied pressure as seen in Figure 3.23. It was found in the full distribution of the resin

channel heights that there were two peaks for each curve as shown by Figure 3.24. From

50 kPa to 600 kPa, the intensity of the peak at low thickness values increased whereas

the intensity of the peak at higher thickness values decreased. The evaluation of the

curves is consistent from 50 kPa up to 600 kPa; however, the 10 kPa curve is singular:

the peak at lower thickness values is larger than that at higher thickness values. Also,

the first peak is at the lowest thickness value. This needs to be investigated further. Also

the thickness of the resin channels was found to be different at the centre of the tow

cross-sections and in between tows.


91

Figure 3.23. Average void thickness with pressure

Figure 3.24. The frequency distribution of the void thickness

From the above observations it is clear that the macro-voids not only differed in the

warp and weft cross-sections but also in the centre of the tow and in between two tows

because of the low cover factor.

Figure 3.25a shows the 3D reconstructed image of the compressed fabric sample with

the resin channels segmented separately in the Avizo software using the threshold


92

method. The volumetric resin channels for the compressed dry fabric preform at 10 kPa

and 600 kPa are pictorially represented in Figures 3.25b and 3.25c, respectively. The

resin channels were also observed for the connectivity of the pores using the threshold

method and it was seen that these channels were connected at each pressure level. At the

higher pressure level of 600 kPa there were thin connections of area even up to 0.01

mm2 at the centre of the tow giving the least passage to resin flow in those areas.

(a)

(b)


93

(c)

Figure 3.25. Resin channels in dry preform during compression: (a) dry preform

with resin channels, (b) channels at 300 kPa, and (c) channels at 600 kPa

3.4.4 Yarn packing fraction

The packing fraction of the yarns (YPF) was calculated by using equation 3.5

(3.5)

In equation 3.5 ‘Vf’ is the volume of the fibres in the yarn, and ‘Vy’ is the total volume

of the yarns (fibres plus air) in the stack. Herein, the yarn volume (Vy) was calculated

with the help of the Avizo software in which the yarns were selected using the threshold

method and the volume was measured using the material statistics tool in the Avizo

software. Whereas, the fibre volume (Vf) was calculated from the density of the fibre

given in the material section.


94

Figure 3.26. Yarn packing fractions of dry preform against different loads

Figure 3.26 represents the yarn packing fractions at various pressure levels. It can be

seen from Figure 3.26 that there was only a slight change in the packing fraction when

the applied pressure increased from 10 kPa to 100 kPa. However, in case where the

applied pressure increased from 100 kPa to 600 kPa, the packing fraction increased

gradually. Figure 3.27 depicts an increase in the packing fraction in the cross-sectional

area with the application of pressure.

Figure 3.27. Yarn packing fraction on application of pressure

Also when the yarn packing fractions were studied in the centre of the tows and in

between two tows, it was observed that the change in yarn packing fractions with

pressure was insignificant in between the tows for both warp and weft yarns compared

After compression


95

to the packing fraction in the centre of the tows for which data are collected in Tables

3.4 & 3.5.

Table 3.4. Yarn packing fraction of warp cross-sections

Packing fraction

Pressure(kPa) Centre of tow In between tows

10 0.53 0.53

50 0.54 0.53

100 0.57 0.53

300 0.61 0.54

600 0.66 0.56

Table 3.5. Yarn packing fractions of weft cross-sections

Packing fraction

Pressure(kPa) Centre of tow In between tows

10 0.56 0.56

50 0.57 0.56

100 0.6 0.56

300 0.64 0.57

600 0.68 0.57

3.4.5 Yarn and fibre volume fraction of the preform

The yarn volume fraction of dry preform was calculated for the warp and the weft cross-

sections at the centre of the intersections as seen in Figure 3.28 and in between two tows

as seen in Figure 3.29.


96

Figure 3.28. Yarn volume fractions calculated at the centre of the tow intersections

Figure 3.29. Yarn volume fractions calculated in between two tows

It was seen that the yarn volume fractions in the slices taken at the centre of the tow

intersections were much higher than those taken in between the two tows for both the

warp and the weft yarns and these are illustrated in Figures 3.28 & 3.29 giving more

passage to resin flow in between the two tows than at the centre of the tow cross-


97

sections. In the slices at the centre of the tows, the volume fraction for the warp yarns

was higher than that at the centre of the weft yarns and this is shown in Figure 3.28.

Whereas, in the slices in between the two tow, the opposite trend was observed and may

be seen in Figure 3.29. This phenomenon again relates to the point that the warp yarn

density was higher than the weft yarn so a higher yarn volume fraction was seen in the

warp direction than that of the weft direction in the centre of the tow while due to

overlapping of the warp yarns on the edges; the weft yarn slices contained some

portions of warp yarns at that point giving a greater yarn volume fraction in the slices

taken in between two warp tows in the weft cross-section.

Also, the fibre volume fraction was calculated and compared using both tomographic

and mechanical test results. From the tomographic results the fibre volume fraction

(FVF) was calculated using equation 3.6.

(3.6)

Where: YPF is the yarn packing fraction and YVF is the yarn volume fraction.

Whereas for the fibre volume fractions from mechanical test results, equation 3.7 was

employed.

(3.7)

Here in equation 3.7 ‘ ’ is the number of fabric layers; is the areal weight of the

fabric, the density of the fibre and ‘ ’ is the thickness of the fabric under

compression.


98

Figure 3.30. FVF calculated by mechanical testing and tomographic analysis

The fibre volume fraction results calculated both from mechanical testing and

tomographic analysis are depicted graphically in Figure 3.30. Close agreement was

observed between the FVF calculated by mechanical test result and the tomography

analysis confirming the accuracy of the imaging/segmentation process. So with good

contrast exact tow geometry can be achieved and FVF and improved geometry models

can also be exported to Abaqus and other analysis software using CT.

3.5 Conclusions

In this work, the fabric tow cross-sectional changes and the geometry of the resin

channels of a single layer dry fabric are followed in-situ by X-ray tomography during

increasing compressive loading. Pressures ranging from 10 kPa to 600 kPa were applied.

It was seen that for pressures up to 100 kPa, the reduction in fabric thickness was

mainly due to reducing crimp and crimp amplitude. In this regime, the crimp balancing


99

position was initially attained by a decrease in the crimp of the warp yarns, followed by

an increase in the crimp for the weft yarns. The reduction in the resin channels was also

prominent in this pressure range of 10-100 kPa, whereas, yarn deformation and packing

fraction were not significant contributors. At higher pressures (300 kPa to 600 kPa) yarn

deformation became prominent, and contribution from the reduction in crimp and crimp

amplitude exhausted. In addition, the macro-void percentage in between two tows was

much higher than that at the centre of the crossing of both the warp and the weft yarns.

Additionally, CT was seen as an accurate tool to find the fibre volume fractions.

Acknowledgements

The authors are grateful to the staff of the Henry Moseley X-ray Imaging Facility which

was funded by the EPSRC under EP/F007906, EP/I02249X, EP/J021229/1 with

additional support from the University of Manchester;

References



1998;19(2):198-216.

[2] Nguyen QT, Vidal-Sallé E, Boisse P, Park CH, Saouab A, Bréard J, et al.

Mesoscopic scale analyses of textile composite reinforcement compaction. Composites

Part B: Engineering. 2013;44(1):231-41.


1946;37(12):T285-92.


100



1992;26(16):2330-47.



[6] Karbhari VM, Simacek P. Notes on the modeling of preform compaction: II-Effect

of sizing on bundle level micromechanics. Journal of Reinforced Plastics and

Composites. 1996;15(8):837-61.

[7] Simacek P, Karbhari VM. Notes on the modeling of preform compaction: I -

Micromechanics at the fiber bundle level. Journal of Reinforced Plastics and

Composites. 1996;15(1):86-122.



1999;59(10):1519-26.

[9] Chen B, Chou T-W. Compaction of woven-fabric preforms: nesting and multi-layer

deformation. Composites Science and Technology. 2000;60(12-13):2223-31.



2006;66(16):3254-62.





2002;72(12):1073-8.


101





2001;317(1-2):188-96.



Manufacturing. 2003;34(5):431-44.













Manufacturing. 2008;39(3):488-502.

[21] Mahadik Y, Hallett SR. Finite element modelling of tow geometry in 3D woven

fabrics. Composites Part A: Applied Science and Manufacturing.41(9):1192-200.


102










[25] Griffin PR, Grove SM, Russell P, Short D, Summerscales J, Guild FJ, et al. The

effect of reinforcement architecture on the long-range flow in fibrous reinforcements.

Composites Manufacturing. 1995;6(3–4):221-35.

[26] Seung Hwan L, Jeong Hwa H, Seong Yun K, Jae Ryoun Y, Young Seok S.

Compression and relaxation behavior of dry fiber preforms for resin transfer molding.

Journal of Composite Materials.44(15):1801-20.



reinforcements. Polymer Composites. 1998;19(5):543-57.







103






Manufacturing. 2006;37(6):868-73.



2000;31(8):715-21.







2006;66(2):297-305.



Physics. 2004;86(2-3):358-69.



[38] Simacek P, Advani SG. Permeability model for a woven fabric. Polymer

Composites. 1996;17(6):887-99.


104

[39] Badel P, Vidal-Sallé E, Maire E, Boisse P. Simulation and tomography analysis of

textile composite reinforcement deformation at the mesoscopic scale. Composites


Chapter 4 Deformation of multilayer dry preforms under compaction

105

Chapter 4

DEFORMATION OF MULTILAYER DRY

PREFORMS UNDER COMPACTION


Abstract

Preform compression is an important mechanism during the resin infusion of the

composite manufacturing process. Here, the analysis of tow cross-section, tow

waviness, nesting of the layers and inter-tow voids in a 2D multilayer woven preform

under compressive loading in-situ was done using computed X-ray tomography (CT).

Low pressures up to 100 kPa, being representative of Vacuum Infusion (VI) were

applied to dry preform and changes in tow cross-section, yarn waviness and inter-tow

voids were observed. Nesting between individual layers was calculated and nesting

factors calculated by mechanical testing and tomographic images were compared. Yarns

packing fractions of dry preform were also calculated for this study. It was seen that the

inter-tow voids reduced drastically for this low load regime and the nesting of the layers

was observed as major contributor to the stack thickness reduction for pressures up to

100 kPa.

Keywords: dry preform, inter-tow voids, nesting factor, computed tomography (CT)


106

4.1 Introduction

The vacuum infusion process (VI) is becoming a promising manufacturing technique in

both automotive and aerospace industry for manufacturing composite parts due to its

low cost, better fibre to resin ratio and environmental benefits. A discussion about the

VI technique can be found in the literature [1]This manufacturing process involves the

compaction of the preform to a certain level of pressure. The compaction of the preform

not only affects the tow geometry of the yarn but also changes the permeability and

mechanical properties of the final product and is taken as an important parameter of the

manufacturing process [2]

Since, in most of the practical applications the use of more than one layer is needed so

the study of the compaction of multilayer stacks of preforms is significant. The

compression behaviour of different fabrics has been studied by a number of researchers

[3-15]. Van Wyk [3] was probably the first to consider fibres under compression as a

system of bending units. Gutowski at el. [4] proposed a theory of elastic deformation of

carbon fibre bundles and the compaction of the fibre bundles was attributed to the

bending of the curved filaments. De Jong et al. [5] derived a mechanical model from

Van Wyk’s compression law to describe a relationship between pressure and thickness

of woven fabrics. Pearce et al. [11] performed experimental work on the

compressibility of the reinforcement fabric under loading and relaxing cycles. They

fitted the loading cycle response to the power law relationship and the relaxation cycles

to an exponential decay functions. Saunders at el. [16] made a mechanical and micro

structural study of different fabrics under compression and provided data of the average

area porosity, area pore structure and voids of the fabric. A detailed study on the


107

internal architecture, tow geometry and effect of yarn waviness on the mechanical

properties for the 3D woven angle interlock fabric under compaction was made by

Mahadik et al. [17-19]. A typical pressure thickness curve was given by Hu et al. and

Mastudaira et al. [20, 21] as shown in Figure 4.1. The curve can be divided into three

parts, two linear parts and one exponential part. Mastudaira et al. attributed the first part

of the curve to the bending of the fibres, the second part with the hardness in

compression because of the friction between the fibres and the third part of the curve to

the lateral compressional modulus of the fibres itself. According to Chen et al. [8] the

preform compaction in the initial mode is due to the reduction of pores and gaps among

the fibres and yarns; while in the third mode it is due to the bending deformation of

yarns. Saunders et al. [22] studied the compaction of plain woven cloth using

mechanical testing and micro structural images of the compressed resin impregnated

samples. They divided the fabric compression into three modes of deformation, the first

mode as nesting of cloths, the second mode as deformation of the yarn waveform and

the third mode as the deformation of the yarn cross- section.

Figure 4.1. Typical pressure thickness curve for a woven fabric under compaction

The compaction behaviour of a multilayer preform is different from a single layer

preform due to shifting and nesting between the individual layers. The nesting between


108

individual layers plays an important role on the compressibility and permeability of the

reinforcement. It also effects the fibre volume fraction and thickness of the preform [9].

The effect of nesting was also studied with regards to the permeability of the woven

fabrics [23-25]. It was shown by Hoes et al. [26] that nesting is a major source of

variation in experimentally determined permeabilites. Keeping in view these points,

study of the nesting phenomenon in textile preforms during compaction is essential to

predict the behaviour of the composites. Yurgartis et al. [27] developed methods of

quantifying yarn shape and nesting using the inclination angle, crimp angle and angle

match. Chen et al. [9] developed an analytical methodology for predicting the elastic

deformation and nesting of multilayer woven fabric preforms under compaction using

their 3D model. Rong et al. [28] developed a micromechanical compaction model for

multilayer woven fabric preforms in which both the shifting and the number of layers

were shown as having significant effect on the compaction behaviour. Also, an increase

in nesting was seen with an increase in the number of fabric layers. Lomov et al. [29]

made a 3D geometrical model to study the nesting of the reinforcement in textile

laminates. In this model the nesting was associated with the float length and tightness of

the weave. According to this model the nesting will increase with an increase in yarn

spacing and with a decrease in float length of the weave. Teresa et al. [30] analysed the

resin impregnated plain weave fabric samples after static and vibration compaction and

studied the degree of nesting along with different tow geometry parameters. They also

observed a decrease in nesting with a decrease of yarn spacing.

In the previous research work, multilayer fabric compression has been studied on the

macro-level using mechanical methods for both dry and wet samples but the meso-

structure of multilayer fabrics has only been investigated for the laminated composite


109

samples. No work has yet been reported on the meso-structure of dry fabrics under

compressive loading. As dry and wet fabrics have different compaction behaviour [31-

33], so the study of dry fabric for the meso-structure is important to predict the

structural changes during compression and for making accurate simulation tools. This

research work concentrates on the analysis of deformation behaviour of dry fabric

samples at macro and meso-scale using mechanical and geometrical analysis of the

fabric samples. The scope of the present work is to study the changes in the tow cross-

section, tow waviness, nesting of the layers, yarn packing fraction and inter-tow voids

in the six layer dry fabric stack using CT under in-situ loading at low pressures. For this

objective a dry preform was compressed between two clear polycarbonate plates at low

pressures and scanned by CT to study the above mentioned parameters.

4.2 Nesting factor and tow geometry parameters

The nesting of layers in a laminate is an important effect which changes the thickness

and the permeability of the laminate resulting a change in fibre volume fraction and

mechanical properties of the final product [29]. The nesting is defined in terms of

nesting factor [34] or nesting coefficient [35].Nesting factor can be calculated by using

equation 4.1.

∑ (4.1)

Where NF is the nesting factor, Ts is the stack thickness and Ti is the individual ply

thickness.

In Figure 4.2a, the thickness of single ply is represented by “H”. In case of two plies of

individual thickness H sit exactly upon top of each other without shifting, the nesting


110

factor will be 1 i.e. no nesting (Figure 4.2b). The thickness of the stack in this case will

be 2H, the sum of the thickness of the individual plies. If there is shifting and nesting

between individual plies the nesting factor will be less than 1 (Figure 4.2c). The stack

thickness will not be equal to the sum of the individual ply thicknesses, but due to layer

shifting and nesting it will be less than the sum of the individual layer thickness by a

factor “N”, where N is the nesting of the layers. The stack thickness will be 2H-N

(Figure 4.2c). Nesting can be calculated by using the stack thickness (Ts) and the

individual layer thickness (Ti) as given in equation 4.2.

∑ (4.2)

Figure 4.2. Layer thickness (a) Single layer, (b) 2 layer without nesting, (c) 2 layer

with shifting and nesting


4.3.1 Material

The material used for the mechanical and microstructure analysis was E-glass plain

woven fabric with a warp density of 4.8/cm and weft density 4.4/cm. The warp and weft

(a)

(b) (c)


111

count was 650 Tex and density of the fibres was 2.60 g/cm3. The numbers of filaments

in the yarn were 1960.


Mechanical testing was done by using Instron testing machine. On the Instron 5569, two

metal platens were used for compaction of the woven fabric sample. The area of the

sample was 15 cm2. Testing of the samples was done by using static test method instead

of dynamic control. In dynamic testing the machine is moved continuously at a constant

speed up to certain pressure level and the pressure thickness curve is recorded. Whereas

in static test method the machine is moved to a defined pressure level and stopped at

that pressure level for certain defined period of time before proceeding to the next

position. As already discussed in previous research that due to relaxation or

rearrangement of the fibres the load decreases with time to maintain a constant

thickness or thickness decreases to maintain constant load due to relaxation or

rearrangement of the fibres [36-38]. In the present work, static test method was adopted

with pressure control to test the fabric sample so that at constant pre-defined pressures,

a constant thickness of dry preform may be achieved. In this method the machine was

moved to the desired pressure level at a constant speed of 1mm/min and then it was

hold for 5 minutes at that pressure level before moving to the next position. This 5

minute hold was considered adequate for this type of loading [30, 39]. Additionally a

compliance curve was also taken before the testing of the fabric samples so that minor

compliances due to the machine may be removed. The pressure thickness results for

single layer and for the stack of six layers were recorded.


112

4.4 Tomography and in-situ compression set up

For the compression of dry preform a compression rig (Figure 4.3) was made to

compress the preform within tomography machine for a meso-structure analysis of the

fabric sample. The rig comprised two polycarbonate compression plates 60x30x12 mm

in length, width and thickness respectively. Two side screws were used to compress the

plates from both sides and two thickness gauges were placed on either side to keep the

thickness uniform on both sides. The edge to edge distance of the two side screws was

40 mm, The sample size of dry fabric compressed between two plates was 40x30 mm.

The bending strength of the plates was estimated by applying beam bending theory to

ensure that there was no significant bending of the plates under load. Once mounted on

the x-ray scanner, the side screws were tightened to compress the fabric preform to the

desired pressure. From the pressure - thickness curve, the thickness value against

desired pressure was taken and the slip gauge of thickness corresponding to that

pressure was put on both sides of the plates.


113

Figure 4.3. Compression Rig (a) Compression rig fixed on the tomography stage,


The process of CT involves collecting a large set of radiographs (projections) of the

sample during a single rotation. Together with a small number of calibration images,

these images are reconstructed into a 3D volume which represents the attenuation

through the sample. A Nikon Custom 320 Bay tomography system was used for this

analysis. The current was then adjusted to 110µA and the voltage was adjusted to 80 kV

and the white grey level was kept at 62500: 3500 projections were taken for each

tomograph.These images were then reconstructed using Metris X-Tek CT Pro and taken

to the Avizo 6.3 software for analysis.



Figure 4.4 shows the average thickness results of the one and six layer fabric with the

pressure up to 600 kPa tested on Instron testing machine with static test method. It is

clearly evident from Figure 4.4 that the thickness of both the single layer and the six


114

layers preform decreased with application of pressures. The average thickness per layer

of the six layers stack is lower than the single layer at the same pressures which is

considered due to nesting of the layers. It can also be seen from the Figure 4.4 that the

thickness of the single layer and six layer stack is different at the start of the loading and

after 5 minutes hold at the same pressure. This is due to relaxation of the fibre as

already discussed. The thickness results changed at the start and became constant with

in 5 minutes of hold.

Figure 4.4. Pressure thickness curves of different layers


Figure 4.5 represents the different tow geometry parameters which were calculated

during the meso-structure analysis using CT. These parameters are also described in the

table 4.1 for both warp and weft yarns separately. The thread spacing P calculated is

also equal to half of the wavelength (λ/2).A sample size of 80 yarns for both warp and

weft yarns at different slices was observed for the tow geometry parameters.


115

Figure 4.5. Yarn geometry parameters

The tomography technique has the advantage that the specimen can be seen throughout

the 3D thickness at any slice so both the warp and the weft yarns were seen at different

slices with in the compressed fabric stack to check the tow parameters and the nesting

of the layers.

Table 4.1. Parameters used to describe the fabric geometry

Parameter Warp Weft

Yarn width a1 a2

Yarn thickness b1 b2

Yarn area A1 A2

Crimp

c1 c2

Crimp amplitude h1 h2

Yarn shift S1 S2

Thread spacing P1 P2

Crimp percentage is calculated by using the equation 4.3.

(4.3)


116

Figure 4.6 represents the 3D stack of dry preform which was reconstructed using Avizo

software.

Figure 4.6. 3D view of the six layer dry preform

Table 4.2 presents the mean values of various tow geometry parameters calculated

during compaction at different pressure levels. For the multilayer preform, the stack

thickness reduction at the initial load of 4 kPa was due to nesting of the layers and as

the pressure was increased from 4 kPa to 45 kPa (Figure 4.7a, b), there was slight

reduction in the thickness of the single layer which was due to crimp amplitude

reduction in the warp yarns, whereas the change in the yarn height and area was

insignificant. The crimp was higher in the warp yarns compared to the weft yarns which

can also be seen in the Figure 4.8. When the pressure level was increased to 45kPa

(Figure 4.7b) there was a decrease in the warp crimp percentage , and the crimp

amplitude whereas there was a slight increase in weft crimp percentage and crimp

amplitude, which is most likely due to the crimp balance or crimp interchange of the

warp and the weft yarns, this phenomenon of crimp interchange during compression

when crimp in one type of yarn is more than the other type of yarn was discussed in

detail by lomov et al. [40]. There was no change in yarn widths of both the warp and

weft tows at this stage and only a slight decrease in the yarn height and yarn area of


117

both warp and weft yarns was seen. At this pressure level nesting was the prominent

reason for the stack thickness reduction. The phenomenon of nesting was varying at

different slices with in the stack resulting a change in the single layer thickness at

different points with in the compressed fabric sample. At the third pressure level of 100

kPa (Figure 4.7c) there was a decrease in yarn area and yarn height and again there was

a reduction in the crimp and crimp amplitude of the warp yarns and an increase in weft

yarns at this pressure level. Also fabric layers were packed further and nested closer to

each other. In all these pressure levels the effect of layer nesting on stack thickness

reduction was dominant and the single layer reduction was mainly due to crimp

amplitude reduction at the 45 and 100 kPa pressure levels. This observation is in

agreement to the experimental results of Saunders et al. [16]and Teresa at al. [30] on

plain woven resin impregnated samples in which they also reported the nesting of the

layers as dominant factor in the initial mode of compression. When the pressure level

was increased to 100 kPa (Figure 4.7c), inter- tow voids were reduced. No significant

change was found in yarn widths upon an increase in pressure even to 100 kPa. From

Table 4.2, if the warp data (area, height) is compared with the weft data, the warp yarn

is likely to deform more as compared to the weft yarn. The reason may be due to the

number of the warp yarns per unit area being greater than the weft yarns so that there

was less nesting in the warp yarns compared to the weft yarns, giving rise to more

deformation of the warp yarns.


118

Table 4.2. Tow geometry parameters measured by X-ray tomography

Warp Weft

4 kPa 45 kPa 100 kPa 4 kPa 45 kPa 100 kPa

(STD) (STD) (STD) (STD) (STD) (STD)

a (mm) 2.08 2.07 2.1 1.65 1.65 1.66

0.16 0.2 0.18 0.08 0.08 0.08

b (mm) 0.33 0.32 0.29 0.41 0.4 0.38

0.03 0.03 0.03 0.03 0.03 0.03

A (mm2) 0.52 0.51 0.48 0.49 0.48 0.46

0.05 0.04 0.04 0.03 0.03 0.03

C (%) 3.5 2.8 2.3 1.02 1.03 1.05

1.01 1.12 1.16 0.67 0.61 0.59

h (mm) 0.57 0.54 0.49 0.23 0.22 0.22

0.07 0.06 0.06 0.06 0.05 0.03

P (mm) 2.1 2.1 2.11 2.27 2.26 2.28

0.11 0.16 0.15 0.09 0.12 0.13

S (mm) 0.53 0.71 0.72 1.1 1.15 1.12


119

Figure 4.7. Tomographic sections through the stack of six layers at different

pressures with warp cross-sections on the LHS and weft cross-sections on the RHS

at (a) 4 kPa, (b) 45 kPa, and (c) 100 kPa


120

Figure 4.8.Yarn isolated from the tomograph showing crimp in crossing yarns, (a)

warp yarn, (b) weft yarn

Table 4.3. Nesting measurements for the 6 layer preform

Nesting

Pressure Shift 1st 2nd 3rd 4th 5th 6th Stack

(kPa) (mm) layer layer layer layer layer layer Nesting

4 1.1 0.92 0.91 0.92 0.9 0.88 0.9 0.91

45 1.15 0.77 0.74 0.72 0.75 0.75 0.78 0.75

100 1.12 0.73 0.67 0.68 0.66 0.68 0.75 0.69

Table 4.3 presents the nesting measurements for 6 layers made from tomographic

images with random nesting at different slices. There was only a minor change in the

nesting of the individual layers at a pressure level of 4kPa but at the increased pressure

levels of 45kPa and 100kPa there was more deviation in the nesting results for the 6

individual layers. For the top and bottom layers at these two pressure levels least nesting

was observed, which may be due to them being in contact with the compression plates.

(a)

(b)


121

(a) (b)

Figure 4.9. Tomographic weft cross-sections showing layers with and without shift

at a) 45 kPa and b) 100 kPa

To investigate the effect of layer shifting and nesting on inter-tow voids, slices of dry

preform were taken into consideration where there was no shifting and nesting of the

tows and where some tows were shifted and nested closer to each other within the stack

of the six layers at pressure levels of 45 kPa and 100 kPa. It can be seen that at 45 kPa,

there was no shifting in the three bottom layers where the tows sit exactly upon each

other (Figure 4.9a), and there was shifting in the top three layers. In the bottom layers

bigger flow channels were present. The maximum area of the channels in the bottom

layers with no shifting was 0.43 mm2 while in the top layers where there was shifting,

the maximum area of the channels was 0.25 mm2. At the pressure of 100 kPa (Figure

4.9b), the layer were sitting upon each other with no shifting. The maximum area of the

channels in the top three layers was 0.23 mm2. In the bottom three layers where there

was the highest level of shifting and nesting, the maximum flow channel area was

0.1mm2. It can also be seen from the Figure 4.9 that at the point of perfect nesting, the

adjacent layers are giving almost no gaps for the resin infiltration. The results show that

channel area for resin flow decreases with the nesting of the layers. The higher the

nesting, the smallest the channel size and vice versa. The same behaviour was observed


122

by Schell et al. [41] for the resin impregnated sample. It also shows that when the

pressure increased from 45 kPa to 100 kPa the layers packing became more prominent.

4.5.2.1 Nesting factor analysis

The nesting factors were calculated from the tomographic images and the mechanical

testing separately at each pressure level using equation 4.1.For the tomographic nesting

factors, the single layer thickness was calculated from individual layers from the stack

of six layers and average layer thickness was taken from the average stack layer

thickness. In mechanical testing the nesting factors were calculated from the thickness

results of a single layer and stack of six layers taken from the Instron testing machine.

The nesting factor results calculated from both the tomography and the mechanical

testing were compared. In Figure 4.10, the nesting factors from mechanical testing and

tomographic images for six layer stack are presented. It can be seen that the nesting

factors calculated from the mechanical compressions are higher than the nesting factors

calculated from the tomographic images. The reason for this difference may be because

the single layer thickness for both mechanical and tomographic results differs due to

nesting of the layers within the stack of six layers in tomography. Also the nesting

factors for top and bottom layers are higher which mean lower nesting in both the top

and the bottom layers than the stack of six layers which may be the top and bottom layer

being in contact with the plate.


123

Figure 4.10. Nesting factors from mechanical testing and tomographic analysis

From Figure 4.10, it can also be seen that the decrease in the nesting factors was higher

when pressure was increased from 4 kPa to 45 kPa and there was lesser decrease upon

increase of pressure to 100 kPa for tomographic images. For the mechanical testing

there is no such sharp decrease in nesting factors. Also it was observed that the

deviation at different slices in the nesting factors of six layers was high when seen

throughout the sample in the 3D stack. The tomographic analysis of dry preform gives

an actual picture of the nesting in the preform during compaction.

4.5.2.2 Quantification of inter-tow voids using image analysis

The permeability of the fibre preform is mainly determined by the inter-tow voids and

the intra-tow voids has only a negligible effect on the permeability of the fibre preforms

[42] so the study of the inter-tow voids is significant to make accurate permeability

models. In the previous chapter 3, the inter-tow voids were calculated for single layer


124

preform. Here in this chapter the inter-tow voids or channels at different slices for both

warp and weft yarns and for the 3D structure of six layer preform were analysed and

quantified in Avizo 6.3 software by applying a threshold method as both the yarns and

the voids have different densities so they were easily distinguished by the threshold

method. They were labelled as different materials and in material statistics their

respective volumes were calculated.

Figure 4.11. Flow channels at (a) 4 kPa, (b) 45 kPa, and (c) 100 kPa

Figure 4.11 shows the presence of flow channels (inter-tow voids) in between the tows

at pressure levels of 4 kPa (Figure 4.11a), 45 kPa (Figure 4.11 b) and 100 kPa (Figure

4.11 c). These images were taken from the material segmentation tool in the Avizo 6.3

software for the weft cross-section at different slices. From these images it can be seen

that the flow channels are not connected in the series and there is a gap in these

channels which is due to nesting of the layers that results in breakage in pores

connectivity in these images. When the layers nest perfectly, the pores connectivity

breaks at that point, resulting no passage for resin flow.


125

Figure 4.12. 3D representations of inter-tow voids at (a) 4kPa, (b) 45kPa, and (c)

100kPa

Figure 4.12 represents the 3D view of the inter-tow voids at the three different pressure

levels. It can be seen that flow channels reduced when the pressure was increased from

4 kPa to 45 kPa and 100 kPa. It was noticed that these voids decreased sharply with an

increase in pressure. Almost 60~70 percent decrease in inter-tow voids was observed. It

was also seen that voids were higher in the area between the two tows and lesser in the

centre of the tow intersection when these voids were checked for different slices for

both the warp and the weft yarn cross-sections. Also interestingly the connectivity of

these resin channels was found to break at points of perfect nesting giving no passage to

resin flow. Figure 4.13 shows the inter-tow voids for both warp and the weft yarn cross-

sections at different slices and also the total inter-tow voids present in the 3D stack at

different pressure levels.


126

Figure 4.13. Inter-tow voids at different slices and average inter-tow voids

4.5.2.3 Yarn packing fraction

Packing fraction of the yarn was calculated by using equation 4.4

(4.4)

where ‘Vf’ is the volume of the fibres in the yarns and ‘Vy’ is the total volume of the

yarns in the stack. Here also the yarn volume was calculated by selecting the yarns by

applying the threshold and fibre volume was calculated by the density of the fibre given

in the material section. Table 4.4 shows the yarn packing fraction against the

corresponding pressures. There was only a slight increase in the yarn packing fraction

with increase in pressure from 4-100 kPa, which shows that the composites

manufactured using vacuum infusion will provide low packing fraction materials and


127

the thickness reduction and fibre volume fraction change is mainly due to reduction of

inter tow voids and nesting of the layers.

Table 4.4. Yarn packing fraction

Pressure (kPa) Yarn packing fraction

4 0.49

45 0.51

100 0.55

4.6 Conclusions

In the present work, changes in the tow cross-section, waviness, layer nesting and inter

tow voids has been studied in dry woven preform under in-situ compressive loading

using CT up to pressure range of 100 kPa being representative of VI. It was seen that at

initial pressure level of 4 kPa the thickness reduction of the preform was due to layer

nesting though the nesting was low in magnitude. Higher inter-tow voids were present

in the layers at this pressure level. As the pressure was raised to 45 kPa the layers nested

closer to each other and the thickness reduction was mainly due to this nesting effect.

There was no significant change in the tow cross-section of both the warp and the weft

at this pressure level. The thickness reduction of the individual layers was due to crimp

amplitude reduction which was again very low. At increased pressure level of 100 kPa,

it was observed that the tow cross-section deformed and again the nesting of the layers

was major contributor to stack thickness reduction. Here also the tow waviness was

responsible for single layer thickness reduction. At the increment of pressures from 4

kPa to 45kPa and 100 kPa, it was observed that the inter-tow voids reduced drastically.

The layer nesting was playing an important role on the magnitude of the resin channels.


128

It was noticed that at pressure levels of 45 & 100 kPa and at the point of perfect nesting

there was no space for resin infiltration. However, bigger resin channels were found

when layers were sitting upon each other without nesting. Yarn packing fraction was

also calculated at these pressures and it was seen that there is no significant increase in

the yarn packing fraction for the vacuum infusion region. The nesting factors were also

calculated from average layer and stack thickness using tomographic images and these

nesting factors were compared with nesting factor values from mechanical testing. It

was found that the nesting factors calculated from tomographic analysis are lower than

the nesting factor calculated from the mechanical test results. It is evident from the

tomographic study of dry preform that nesting and inter-tow void reduction are major

sources of stack thickness reduction at low pressures. This work will help the

researchers to understand the compaction process of dry preform at low load regime in a

better way and data of the inter-tow voids will be used to make permeability model in

the future.

References

[1] Summerscales J, Searle TJ. Low-pressure (vacuum infusion) techniques for

moulding large composite structures. Proceedings of the Institution of Mechanical

Engineers, Part L: Journal of Materials Design and Applications. 2005;219(1):45-58.



1998;19(2):198-216.


1946;37(12):T285-92.


129



1992;26(16):2330-47.



[6] Karbhari VM, Simacek P. Notes on the modeling of preform compaction: II-Effect

of sizing on bundle level micromechanics. Journal of Reinforced Plastics and

Composites. 1996;15(8):837-61.



Composites. 1996;15(1):86-122.



1999;59(10):1519-26.





2006;66(16):3254-62.





2002;72(12):1073-78.


130





2001;317(1-2):188-96.



Manufacturing. 2003;34(5):431-44.







[18] Mahadik Y, Hallett SR. Finite element modelling of tow geometry in 3D woven

fabrics. Composites Part A: Applied Science and Manufacturing.41(9):1192-200.









131






Physics. 2004;86(2-3):358-69.


low-pressure permeation of fabrics: Part II ”The variable gap model and prediction of








[27] Yurgartis SW, Morey K, Jortner J. Measurement of yarn shape and nesting in

plain-weave composites. Composites Science and Technology. 1993;46(1):39-50.

[28] Chen Z-R, Ye L. A micromechanical compaction model for woven fabric

preforms. Part II: Multilayer. Composites Science and Technology. 2006;66(16):3263-

72.



2003;63(7):993-1007.


132



Manufacturing. 2008;39(3):488-502.



















Manufacturing. 2006;37(6):868-73.


133



reinforcements. Polymer Composites. 1998;19(5):543-57.


reinforcements during composite processing. Polymer Composites. 1991;12(1):13.








Chapter 5 Compaction and nesting in textile preforms influenced by tow architecture

134

Chapter 5

COMPACTION AND NESTING IN TEXTILE

PREFORMS INFLUENCED BY TOW

ARCHITECTURE


Abstract

This chapter reports macro and meso-scale deformations in textile preforms due to

compaction pressures applied during processing techniques such as vacuum infusion

and autoclave curing. 2D fabrics with a variety of interlacement patterns - plain, twill

and sateen- as well as stitched non-crimp (NCF) fabrics have been investigated. The

study demonstrates the influence of interlacement topology on the compaction and

nesting behaviour of individual plies. In-situ compression of single and multi-layer

fabrics in X-ray Computed Tomography shows interesting meso-scale features within

each ply such as tow waviness and inter-tow voids/channels for resin flow. Composite

laminate thickness, fibre volume fractions and the mechanical properties are influenced

by nesting efficiency as well as tow compaction behaviour under process pressure.

Keywords: weave architecture, compaction, nesting, computed tomography (CT), non-

crimp fabrics


135

5.1 Introduction

Low-cost composites manufacturing techniques based on resin infusion of dry textile

preforms are becoming popular in aerospace, automotive and energy sectors. In these

processes, several fabric plies are draped on the tool surface and subjected to transverse

compaction forces either by vacuum (Vacuum Infusion) or external pressure (in match

mould RTM), and at the same time infused with liquid resin. Compaction of the

preform changes the thickness and the resulting change in the fibre volume fraction of

the composite laminate. Preform compaction changes the tow waviness, which in turn

changes the mechanical properties of the composites [1]. Compaction of the preform

also effects its permeability to resin flow [2]. During autoclave curing, significantly

higher pressures (around 7 bar) are applied resulting in thickness reduction and a

corresponding increase in fibre volume fractions. As a result it is important to

characterise the compaction behaviour during the composite manufacturing process [3].

Extensive research work has been reported in the literature on the compaction behaviour

of different fabrics [4-15]. Van Wyk [16] was probably one of the first researchers to

treat the fibres under compression as a system of bending units. A typical pressure-

thickness curve was given by Mastudaira et al. [5]. The curve consists of three parts,

two linear and one exponential part. The effect of the reinforcement structure on the

compaction behaviour has been studied by various researchers [8, 13, 17-19].

Experimental research on the compression of plain weave fabric can be found in

previous literature [4, 8, 11, 17, 20, 21]. The compression response of non-crimp

stitched fabrics has also been reported [22-25]. The compression response of multilayer

preforms differs from that of a single layer due to nesting of the layers. The nesting

behaviour was discussed by Lomov et al. [26]. Resin permeability through fabrics due


136

to the nesting effect has been studied by various researchers [27-29]. It was observed by

Hoes et al. that nesting is a major source of variation in experimentally determined

permeability values [30]. The nesting of layers has been studied in terms of a nesting

factor [31] and a nesting coefficient [32]. The majority of the existing literature is

focused on plain weave, with very little reported on the influence of interlacement

architecture on compaction behaviour. Additionally, tow geometry analysis found in

the literature was conducted primarily on composite laminates not on dry fabrics.

Commercial fabrics have been used in previous compaction studies with no control on

weave parameters. In this study, several weave architectures were produced on a rapier

loom by keeping the remaining parameters such as inter-tow spacing and tow linear

density constant. This procedure enabled the study of the influence of interlacement

geometry on compaction behaviour, while keeping all the other parameters constant.

5.2 Material & Experimental details

5.2.1 Material

In this study, woven fabrics including plain (1/1), twill (3/1) and 5-harness sateen

weaves were produced with 600 tex, E-glass yarn in both warp and weft directions. In

addition, 0о/90

о and +45

о/-45

о non-crimp fabrics were included in the study [33]. The

fabrics were manufactured such that the yarn linear density (Tex) and inter-tow spacing

(ends or picks/cm) were identical for plain, twill and sateen fabrics. In this way the

influence of weave style on the compaction behaviour can be studied. The specifications

of the fabrics, used are listed in tables 5.1 & 5.2. The unit repeats of the plain, twill and

sateen weaves are as shown in Figure 5.1; the images were created using the TexGen


137

software developed by University of Nottingham [34]. The scanned images of woven

fabrics and non-crimp stitched carbon fabrics are presented in Figure 5.2.

Figure 5.1. Woven fabrics (a) Plain, (b) 3/1 twill, (c) 5H sateen

Table 5.1. Woven fabric specifications

Sample

#

Yarn

count

Type Weave

Ends

/cm

Picks/

cm

Areal weight

(g/m2)

WV-1 650 Tex E-glass 1/1 Plain 4.8 4.4 620

WV-2 600 Tex E-glass 1/1 Plain 4.8 4.4 548

WV-3 600 Tex E-glass 3/1 Twill 4.8 4.4 546

WV-4 600 Tex E-glass 5H Sateen 4.8 4.4 545

Table 5.2. Stitched non-crimped fabric (NCF) specifications

Sample

Yarn

count

Plies

orientation

Tows/cm Stitch yarn

Arial

weight

(g/m2)

NCF-1

12k-

carbon

+45о/-45

о 1.5/1.5

55 dtex

(polyester)

246

NCF-2

12k-

carbon

0о/90

о 2.5/2.5

80 dtex

(polyester)

550


138

Figure 5.2. Scanned images of woven fabrics and non-crimp stitched fabrics

5.2.2 Experimental Methodology

5.2.2.1 Mechanical testing

Mechanical testing of single and multilayer fabrics was conducted using an Instron

5569 universal testing machine (Figure 5.3) with a small-capacity load cell (5kN). The

surface area of the top and the bottom plate was 50 cm2. The fabric samples were cut

into 10x10 cm pieces. As the fabric thickness is small in comparison to the machine

stroke, the accuracy with which the compression strain measured becomes important. In

order to minimise errors, machine compliance as a function of the applied load was

measured and accounted in fabric starin calculations. Two different test methods were

used, namely continuous and static testing. In the continuous testing method, the cross-

head was moved continuously at 1mm/min and the resulting pressure thickness curve

was recorded. During the static test method, the machine cross-head was moved at a


139

speed of 1 mm/min to the desired load and then it was stopped for five minutes. Fibres

tend to relax during compression which result in a decrease in fabric thickness with

time, so five minute hold time was given to allow the fibres to relax so that there is no

further relaxation in the fibres with time and consequently, a stable, uniform thickness is

achieved. This five minute period was considered to be adequate for these type of

loadings [20, 21] . These final thickness values at each pressure level were recorded.

These values were also used to caliberate the compression rig for the meso-structure

analysis.

Figure 5.3. Compression set up

5.2.2.2 Computed tomography (CT) analysis

Meso-scale tow geometry has been studied by several researchers using SEM or

Computed Tomography, primarily on laminate samples. Tow geometry in dry fabrics

has not been studied so far as the fabrics are limp and hence difficult to section or scan.

In the present work, a novel in-situ loading rig has been developed (Figure 5.4) in order

to support dry fabric as well as to apply known compressive loads. The rig consists of


140

two clear polycarbonate plates, 60x35x12 mm in length, width and thickness directions

respectively. Two thickness gauges (slip gauges) were placed between the two plates on

either side in order to control the degree of compression. The fabric was placed

between these two plates and compressed to the desired thickness by tightening the side

screws. The gauge thickness was decided according to the pressure level required, based

on the mechanical tests.

Figure 5.4. Compression rig for in-situ loading

The fabric specimens were scanned on the Nikon Metrology 225/320 kV Custom Bay

system at the Henry Moseley X-ray Imaging Facility, University of Manchester.

Scanning was performed with a silver target using a voltage of 85 kV and a current of

115 μA. The number of projections was set to 3142 acquired of 360. The 3D data set

was reconstructed at full resolution with a voxel size of 13.2 μm along the x, y,


141

and z directions. Image analysis was performed using Avizo® Fire version 7.0.1

software.

The dataset were filtered using a non-local means filter in order to remove noise from

the data. The segmentation, i.e. separation of the pore space and fibres, has been

achieved by a global thresholding approach based on seed region growing. The seed

region growing is one of the simplest region based segmentation methods. It performs a

segmentation of an image wich examines the neighbouring pixels of a set of points,

known as seed points, and determines whether the pixels could be classified to the

cluster of seed points or not. The threshold is made by the user and it is usually based on

intensity, grey level, or colour values. The regions are chosen to be as uniform as

possible.

5.3 Analysis of macro-scale deformations

5.3.1 Woven fabrics

Fabric compression tests were performed on single layers, as well as stacks of four and

six layer fabrics. Figure 5.5 shows the pressure thickness curves for single layer plain,

twill and sateen fabrics. It was observed that, in case of single layer, the thickness of the

plain woven fabric was lowest of the three weaves, under a very small load; twill and

sateen weaves have nearly same thickness (Figure 5.5).


142

Figure 5.5. Single layer thickness results against pressure

With an increase in pressure, the thickness of all the fabrics started decreasing and at a

pressure of 20 kPa, the sateen fabric thickness became less than the plain fabric and at a

pressure of 30 kPa, the twill weave thickness becomes less than the plain fabric. Upon

increasing the pressure further, the least reduction was observed for the plain fabric and

the highest thickness reduction was evident for the sateen fabric. The plain fabric may

have the lowest initial thickness due to having the highest number of interlacement

points compared to the twill and sateen fabrics. Interlacements in the plain weave result

in higher inter-tow compaction forces and a corresponding reduction in thickness.. Twill

and sateen fabrics have fewer interlacements per unit area and hence a higher initial

thickness. During compression, plain weave offers the greatest resistance due to the

tightness of the weave giving the least reduction in thickness. Chen et al. [13] attributed

this to the bending stiffness of the interlacing tow segments. Of the two remaining


143

structures, sateen has the least number of interlacements and hence the highest reduction

in thickness.

(a)

(b)


144

(c)

Figure 5.6. Compression behaviour of multi-layer fabrics, (a) plain, (b) twill, (c)

sateen

In the case of multilayer preforms, stacks of two and six layers were tested for plain,

twill and sateen fabrics. The average layer thickness results against different pressures

are presented in Figure 5.6 along with single-layer curve. From Figure 5.6, it can be

seen that there was a reduction in the average layer thickness of all these fabrics with

increase in number of layers, which is considered to be due to nesting of the layers. This

observation of reduction of average layer thickness with increasing number of layers is

in agreement with the literature [7, 11, 17, 26] as already mentioned in chapter 4.

It can be observed from Figure 5.7 that prior to the application of pressure the thickness

of the multi-layer plain weave is the lowest. With increasing pressure, the rate of

reduction in thickness was least for the plain fabric and highest for the sateen fabrics

giving rise to the lowest fibre volume fraction in plain fabrics compared to twill and

sateen fabrics at high pressures. Consequently, the thickness of the twill and sateen


145

weaves is almost 20 percent higher than the plain weave at a pressure level of 5 kPa for

both two and six layers stacks. In the case of the two-layer stack, the thickness of the

sateen and twill fabrics became equal to that of the plain weave at pressures of 25 kPa

and 35 kPa respectively. But in the case of the six layer stacks, considerably higher

pressures of 57 kPa and 85 kPa were required to bring the sateen and twill fabrics

respectively equal to the plain weave fabric thickness.

(a)

(b)

Figure 5.7. Different layers thickness results against pressure, (a) two layers (b) six

layers


146

This observation can be attributed to the tight structure of the plain fabric, which

hinders compression compared to the twill and sateen fabrics. This observation is in

contradiction to the study of Saunders et al. [17], in which they found twill fabric to be

the most resitant to compression compared to plain and sateen fabrics. But in their

study, the yarn densities and the counts were not same for all the three type of weaves

which could result in different compaction behaviour of these fabrics. Overall, from

single layer and multilayer fabric study of plain, twill and sateen structures under

compression, it was seen that the plain fabric is the most compression resistant due to its

tight structure and consequently it results in the lowest fibre volume fraction at a given

pressure even though the pressures are characteristically low (<100 kPa) for vacuum

infusion.

5.3.2 Non-crimp fabrics

In the case of non-crimp stitched carbon fabrics, two configurations: +45о/-45

о and

0о/90

о were tested for compression behaviour. In both cases, the plies were stitched

together by a polyester thread. The pressure thickness results are presented in Figure

5.8.


147

(a)

(b)

Figure 5.8. Non-crimp fabric thickness results as a function of pressure, (a) +45о/-

45о, (b) 0

о/90

о


148

For NCF-1 (+45о/-45

о), it was observed that there was a reduction in the average layer

thickness upon movimg from a single layer to 2 but that the addition of further layers

resulted in only a marginal decrease in thickness. In the case of NCF-2 (0о/90

о), which

is similar to woven broadcloth, the thickness reduction with increasing number of layers

was prominent. For non-crimp stitched fabrics, different compaction behaviours were

observed in literature. Only a small thickness reduction was noticed in stitched non-

crimp fabrics for both 0о/90

о and +45

о/-45

о by Hammani [18] while the thickness

reduction was found to be higher in the 0 о

/-45 о

/90 о

/+45 о

plies compared to +45 о

/-45 о

plies [22]. Saunders et al. [17] observed that non-crimp stitched bonded fabrics are easy

to deform compared to plain and satin fabrics.

5.3.3 Nesting of layers

Layer nesting describes how the “hills” and “valleys” of two adjacent layers embed one

into another reducing the average layer thickness relative to that of an individual layer

[32]. The nesting of the layers is quantified in terms of the nesting coefficient [32] or

nesting factor [31]. The nesting factor can be calculated using the equation 5.1 [31].

∑

⁄ (5.1)

is the nesting factor, tack thickness and individual layer

thickness.

If the layers sit upon each other without nesting, the thickness of the two layers stack

will be equal to sum of the individual layer thickness and the nesting factor will be one.

But in the case where the layers shift and come closer to each other, the nesting factor

will be less than one and the thickness of the stack will be smaller than the sum of the


149

thickness of the individual layers such that a lower nesting factor indicates more

efficient nesting.

In this study, the nesting of the layers was calculated in terms of nesting factor (NF) for

multilayer stacks using test results obtained from the compaction of dry preform on

Instron testing machine. Figure 5.9 represents the results of nesting factors calculated

for woven and non-crimp stitched fabrics.

(a)


150

(b)

Figure 5.9. Nesting factors (a) woven fabrics (b) non-crimp stitched fabrics

In the case of woven fabrics, It was observed that the nesting factors are lower for plain

woven fabrics, reflecting better nesting compared to twill and sateen fabrics. The

highest nesting was seen in the six layer plain woven fabric which again can be

attributed to the presence of the shorter float length as already discussed. The highest

nesting factors were observed for sateen fabric. Higher nesting in plain woven fabrics

can be attributed to a shorter float length which allows better nesting . The reason of

better nesting due to shorter float length can be explained by the presence of more peaks

and valleys in the shorter floatt compared to longer float which allow neighbouring

yarns to locate in these valleys as mentioned by Lomov et al.[26]. On the other hand,

the sateen fabric,which have the largest float length, have minimum nesting. This

observation agrees with the study of Lomov et al. [26], according to which the presence

of longer yarn float results in minimum nesting and vice versa. As already mentioned in

the discussion of results of woven fabrics that least fabric thickness reduction was

witnessed in plain fabrics, in both single and multilayer cases. Whereas, better nesting is


151

observed in plain fabric compared to twill and sateen fabrics during nesting factor

calculations. Ideally, the better nesting in plain multilayer fabrics should result into

reduced stack thickness of plain fabric compared to multilayer twill and sateen fabrics

but opposite behaviour was noticed. The reason of higher stack thickness of plain

woven fabric, even in presence of higher nesting, can be returned to least single layer

thickness deformation due to tight structure of plain fabrics. Consequently, higher stack

thickness even in presence of higher nesting in plain fabrics.

For non-crimp fabrics, the measured nesting factor results (Figure 5.9), present better

nesting in case of 0о/90

о compared to +45

о/-45

о configuration. Again the reason may be

due to the difference in tow spacing. The nesting in 0о/90

о configured fabric is even

better than woven fabrics. This can be attributed to the presence of stitching yarn on the

face and the surface of the non woven stitched fabric,which allowed the yarns of the

other layers to nest better. Also, the shifting pattern of the layers during lay-up, plays an

important role in nesting. In ideal situations, the shifting of the layers should also be

considered using CT images during nesting comparison which can give a better

understanding of the nesting phenomenon.

5.4 Analysis of meso-scale deformations

For meso-deformation analysis, Wv-1 fabric was used. Both single layer and multi-layer

stacks were studied for geometrical changes using CT. The CT images were analysed

for tow waviness and inter-tow voids at different cross sections. Additionally, the void

behaviour was investigated for the 3-D structure of single layer and multilayer stack.

Figure 5.10 shows the 3D reconstruction of the single and multi-layer fabric samples.

The cross-sectional images of single layer and multi-layer stack are presented in Figure

5.11 and 5.14. The obtained images were studied for inter-tow voids using Avizo


152

software and the yarn waviness in the cross-sectional images was calculated using

measure software.

Figure 5.10. 3-D reconstruction of single layer (right) and 10 layers (left)

Figure 5.11. CT images of single layer

5.4.1 Tow waviness

Interlacement of warp and weft yarns imparts a certain amount of waviness to these

yarns in the fabric. This waviness is termed as yarn crimp. Due to this crimp, the actual

length of the yarn in the fabric exceeds the length of the fabric. Yarn crimp is calculated

using the equation 5.2.

(5.2)


153

Where P and L are actual yarn length and crimp yarn length respectively as seen in

Figure 5.12. The tow waviness results are presented in Figure 5.13.

Figure 5.12. Tow showing crimp

Figure 5.13. Yarn crimp percentage (a) warp yarn, (b) weft yarn

During the tow waviness study, it was observed that the crimp percentage is higher for

the warp yarns compared to the weft yarns in both single layer and multilayer cases at a

pressure of 100 kPa (Figure 5.13). Additionally, the crimp percentage of ten layer stack

was higher than that of single layer. This is most likely due to the nesting of the layers

in multilayer stack, as yarns of one layer are embedded in the gaps between the adjacent

yarns of the other layer which results in less crimp reduction in the nested areas.

Whereas, in the case of single layer fabric, the warp and weft yarns are in direct contact


154

to the compression plate, giving rise to a higher crimp reduction on application of

pressure. After increasing pressure from 100 kPa to 300 kPa, there was a reduction in

the crimp percentage of the warp yarns and an increase was observed for the weft yarns

for both single layer and ten layer stack (Figure 5.13). This phenomenon can be

attributed to the balancing of the yarns, in which the crimp percentage decreases in

yarns in one direction and increases in yarns going in the other direction on application

of pressure, if the two are not at same crimp levels. This behaviour was discussed in

detail by Potluri et al. [31]. Also at the pressure level of 300 kPa, the crimp percentage

of multilayer stack was higher than the single layer which again can be due to the

nesting of the layers in multilayer stack as already discussed. Moreover, the results of

the crimp percentage showed higher deviation for the multilayer stack compared to

single layer (Figure 5.13). Again the reason for this is due to the presence of nesting in

some areas and absence in other areas within the multilayer stack of ten layers. This

results in different crimp percentage at nested and non-nested regions, consequently,

giving more scattered results of crimp percentage for the multilayer stack.


155

Figure 5.14. CT images of ten layer stack

5.4.2 Inter-tow voids

The study of the inter-tow voids or the resin channels in textile preforms is important

for making accurate simulation tools as the resin permeability mainly depends on inter-

tow voids [2]. Here in this work, a study of inter-tow voids in 650 Tex, E-glass dry

plain woven fabric, single layer and multi-layer stacks conducted under in-situ

compaction has been presented. The inter-tow voids were investigated in 3D structure

as well as at different 2D slices with in the preform and the behaviour was compared for


156

single layer and multilayer. A slice is a CT image which corresponds to a certain

thickness of the object being scanned. A slice has a thickness of one voxel and the

length and width of the slice depend on the field of view of the sample. It was observed

that the inter-tow voids were higher in single layer compared to multi-layer stack at

pressure level of 100 kPa. At a pressure of 300 kPa, the inter-tow voids reduced in both

single and multilayer fabrics. The magnitude of voids reduction was higher in

multilayer stack compared to single layer (Figure 5.15). This higher voids reduction in

multilayer stack can be attributed to the nesting of the layers in the multilayer stack.

Also, in the multilayer stack, larger resin channels were present in non-nested regions

compared with the nested regions. The inter-tow voids were also investigated at

different slices in single layer and multilayer fabrics. The results of voids percentage,

slice by slice, are presented in Figure 5.16 and 5.17. It was found that the deviation of

the inter-tow voids from the mean results was higher in single layer compared to

multilayer stack. The reason of this higher deviation of the voids in single layer was the

different behaviour of voids in the centre of the tow and in between two tows. As

depicted in Figure 5.18, the absence/lower amount of fibres in-between the two tows

results in higher voids in this region compared to the centre of the tow. This gives

higher deviation of the void results from the mean value in single layer. Whereas, in

case of multi-layer stack, due to shifting and nesting of the layers these spaces between

the two tows are occupied by the top and bottom layers. This results in less deviation of

void percentage from the mean value, compared to single layer. Also, the voids

percentage along the slices represents a periodic shape which is due to sinusoidal

structure of the plain weave (1/1). The minimum voids are found at the intersecting

point of the tows, compared with between the tows where maximum voids are observed.


157

This behaviour is clearly visible in the periodic behaviour of inter-tow voids through the

samples (Figure 5.16 & 5.17).

Figure 5.15. Inter-tow voids in single layer and multilayer stack

Figure 5.16. Inter-tow voids in single layer along slices (a) 100 kPa, (b) 300 kPa


158

Figure 5.17. Inter-tow voids in multilayer stack along slices (a) 100 kPa, (b) 300 kPa

Figure 5.18. Plain-weave repeat presenting, between two tows and centre of the tow

5.5 Conclusions

In this work, 2D woven fabrics having different interlacement patterns with identical

tows and tow spacing were developed and tested for compaction behaviour.

Additionally, non-crimp fabrics with two different tow orientations were included in the

study. In woven fabrics compression of plain, twill and sateen fabrics and highest

compression was observed in sateen fabrics and least in plain woven fabrics. Opposite

behaviour was present in case of nesting where plain woven fabric gave better nesting


159

results. This showed that fabric compression and nesting are influenced by float length

of weave. In case of non-crimp fabric, highest nesting was present in fabrics placed at

+00/90

0. The tow waviness was higher in multilayer layers stacks compared to single

layer at 100 kPa showing higher tow waviness reduction in single layer in vacuum

infusion region.

Increase in pressure beyond 100 kPa resulted more reduction in tow waviness in

multilayer stack.The inter-tow voids were lower in multilayer stack compared to single

layer at 100 kPa, which was due to nesting of layers which reduced inter-tow voids in

multilayer stack. The reduction in voids was also higher in multilayer stack at

application of 300 kPa pressure, here again the nesting of the layers was the reason of

higher voids reduction.

References



2000;31(8):715-21.





1998;19(2):198-216.






160



1999;59(10):1519-26.





2001;317(1-2):188-96.










Composites. 1996;15(1):86-122.



2006;66(16):3254-62.

[14] Chen Z-R, Ye L. A micromechanical compaction model for woven fabric

preforms. Part II: Multilayer. Composites Science and Technology. 2006;66(16):3263-

72.


161



2006;66(2):297-305.


1946;37(12):T285-92.






[19] Luo Y, Verpoest I. Compressibility and relaxation of a new sandwich textile

preform for liquid composite molding. Polymer Composites. 1999;20(2):179-91.





Manufacturing. 2008;39(3):488-502.

[22] Lomov SV, Belov EB, Bischoff T, Ghosh SB, Truong Chi T, Verpoest I. Carbon

composites based on multiaxial multiply stitched preforms. Part 1. Geometry of the

preform. Composites Part A: Applied Science and Manufacturing. 2002;33(9):1171-83.

[23] Lomov SV, Verpoest I, Barburski M, Laperre J. Carbon composites based on

multiaxial multiply stitched preforms. Part 2. KES-F characterisation of the

deformability of the preforms at low loads. Composites Part A: Applied Science and

Manufacturing. 2003;34(4):359-70.


162

[24] Liu XL, Falzon PJ, Sweeting R, Paton R. Effective compressibility and

permeability of multi-layer non-crimp fiberglass reinforcements. Journal of Reinforced

Plastics and Composites. 2004;23(8):861-79.

[25] Lomov SV, Barburski M, Stoilova T, Verpoest I, Akkerman R, Loendersloot R, et

al. Carbon composites based on multiaxial multiply stitched preforms. Part 3: Biaxial

tension, picture frame and compression tests of the preforms. Composites Part A:




2003;63(7):993-1007.



Physics. 2004;86(2-3):358-69.


low-pressure permeation of fabrics: Part IIâ€”The variable gap model and prediction of











163




[33] http://www.sigmatex.com/Home.

[34] http://texgen.sourceforge.net/index.php/Main_Page.

http://www.sigmatex.com/Home

http://texgen.sourceforge.net/index.php/Main_Page

Chapter 6 Deformation of dry and wet preforms under compaction

164

CHAPTER 6

DEFORMATION OF DRY AND WET PREFORMS

UNDER COMPACTION


Abstract

In the present work, the compaction behaviour of textile preforms has been studied for

dry and wet fabrics. Mechanical test results have been used to study the compaction

response of dry and wet preforms at the macro-level. A power-law relation has been

derived to predict the wet fabric thickness from the dry fabric thickness results. Meso-

scale analysis of single layer dry and wet fabrics has been conducted using computed

tomography (CT) under in-situ loading. Tow waviness and resin channels of dry and

wet fabrics have been studied for single layer preforms using CT images. It has been

observed that tow waviness of the fabrics decreases drastically with wettability at low

loads.

Keywords: dry preform, wet preform, compaction, computed tomography (CT)


165

6.1 Introduction

The use of composite materials is becoming popular rapidly as a key structural element

in civil and the aerospace industries as they are easy to handle and have high strength to

weight ratio. For composite manufacturing a number of different techniques are

employed amongst which Resin Transfer Moulding (RTM) and Vacuum Infusion (VI)

are most common. A detailed review of both of these manufacturing processes can be

found in literature [1, 2]. These manufacturing processes involve the compaction of dry

fabrics, the infusion of the resin and compaction under wet conditions and subsequent

curing of the laminate to obtain the final composite material. Compaction of the fabrics

under dry and wet conditions changes the preform thickness which results in a change

in the fibre volume fraction of the preform. The compaction of the preform also affects

the inter-tow voids. The compression behaviour of different fabrics has been studied by

a number of researchers [3-6]. The compaction behaviour of wet fabrics is different

from dry fabric compaction as observed by various researchers [7-11]. In the previous

research work, the compaction behaviour of dry and wet fabrics has been investigated

using mechanical test results. There is no comparative study available on the structural

analysis of dry and wet fabrics. In the present work, the compaction behaviour of dry

and wet fabrics has been studied at macro-level using mechanical test results. A power-

law relationship has been used to predict wet fabric thickness from dry fabric test results.

The meso-structure of single layer dry and wet fabrics has been studied using CT

images.


166


6.2.1 Material

The specifications of fabrics used for the macro study are presented in Table 6.1.

Table 6.1 Material used for mechanical testing

Structure Yarn type Yarn count GSM Fabric ID

Plain weave (1/1) Glass yarn 600-tex 548 P1

Twill weave (3/1) Glass yarn 600-tex 546 T1

Satin weave (5-harness) Glass yarn 600-tex 545 S1

Non crimp fabric ( +45/-45) Carbon yarn 12-k 246 NCF

The specifications of the material used for meso-structure analysis of dry and wet

fabrics are presented in Table 6.2.

Table 6.2 Material used for meso-structure analysis

Structure Yarn type Yarn count GSM Fabric ID

Plain weave (1/1) Glass yarn 650-tex 620 P2


For the mechanical testing of the fabric samples, an Instron 5569 machine was used.

Samples with an area of 50 cm2 were used and the speed of the machine during testing

was 1 mm/min. The load cell used was of 5kN. For the testing of wet fabrics, the

samples were saturated with water before compression. To study the macro-level


167

deformation of the fabrics, a continuous test method was used. In this method the upper

compression plate was moved continuously at a speed of 1mm/min until the final load

was achieved and the pressure thickness curve was recorded. For meso-level

deformation study of the glass fabrics, a static test method was used. The reason for

using the static test method for this study was that there is thickness reduction in the

preform under loading with time due to relaxation of the fibres. Therefore, 5 minute

hold time was given for the fibres to relax so that an unchanged, stable final thickness

may be achieved. It had been observed in previous studies that after 5 minutes loading

at a constant pressure there is no further thickness reduction with time [12]. These

thickness values were also used for calibration of the compression rig used for the CT

analysis. Prior to the testing of the samples, the compliance curve was recorded. The

compliance values were removed from the thickness readings so that the compliance

effect due to the machine was balanced.



The compaction of single layer fabric was performed in dry and wet conditions. For wet

compaction the fabric, saturated with water, was compacted to measure the pressure-

thickness response. The thickness results for dry and wet fabrics at different pressures

are presented in Figures 6.1 to 6.5.


168

Figure 6.1. Thickness vs pressure of single layer and two layers of P1 fabric

Figure 6.2. Thickness vs pressure of single layer and two layers of T1 fabric

Figure 6.3. Thickness vs pressure of single layer and two layers of S1 fabric


169

Figure 6.4. Thickness vs pressure of single layer and two layers of NCF

Figure 6.5. Thickness vs pressure of single layer and two layers of plain fabric P2

Two different thickness curves were recorded for each fabric; the thickness curve for

wet fabrics represents the lower thickness values, which shows there was higher

thickness reduction in wet fabrics compared to dry fabrics under the same compression.

The thickness reduction on being wetted occurs on compression of fabric which may be

associated with the crimp amplitude reduction; just lubrication of fibre without

compression does not reduce the fabric thickness. The thickness decrease on wettability

was higher in the low pressure regions compared to higher pressures. The highest fabric


170

thickness reduction in wet fabrics was observed in sateen fabrics and the least thickness

reduction in plain fabrics. The reason for higher thickness reduction in sateen fabrics on

wettability can be returned to the presence of higher float length of the yarns and fewer

intersecting points giving more space for tow spreading resulting in higher thickness

reductions. This phenomenon of higher thickness reduction was observed in all the

compacted fabrics. This study is in agreement with previous research [13, 14].

A power-law relationship was obtained to predict wet preform thickness from dry

thickness results. To achieve this relationship, the negative stretch ratio (in the

compression direction) Td/T0 and Tw/T0 were calculated for both dry and wet fabric,

where T0 is the initial thickness at a low pressure of 3 kPa, which is assumed to be same

for both dry and wet fabrics and Td and Tw are dry and wet fabric thicknesses

respectively at increased pressure levels. The stretch ratio difference of both dry and wet

fabrics (Td-Tw/T0) was plotted against pressure and fitted to the power-law as shown in

equation 6.1. Both experimental and power equation curves were in good agreement

with each other as seen from Figure 6.6

(6.1)

Where X is the pressure, ‘Y’ is the stretch ratio difference (Td-Tw/T0) and ‘a’ & ‘b’ are

constants, equation 6.1 can be modified to derive equation 6.2.


171

Figure 6.6. (Td-Tw)/T0 plotted against pressure for plain (P2) woven glass fabric

(6.2)

From this relation, wet thickness values (Tw) of different fabrics were calculated and

compared with the experimental results. A good agreement was observed between the

experimental and predicted values of these fabrics. The experimental and fitted values

of these fabrics are presented in Figures 6.7 to 6.11.

Figure 6.7. Experimental and predicted wet thickness of plain fabric (P1)


172

Figure 6.8. Experimental and predicted wet thickness of twill fabric (T1)

Figure 6.9. Experimental and predicted wet thickness of sateen fabric (S1)

Figure 6.10. Experimental and predicted wet thickness of NCF


173

Figure 6.11. Experimental and predicted wet thickness of plain fabric (P2)

6.3.2 Tomography and in-situ compression set up

For compression of dry and wet preforms a compression rig (Figure 6.12) was

developed to compress a dry preform so that the fabric under in-situ loading may be

scanned within the tomography machine for meso-structure analysis. The rig comprised

two polycarbonate compression plates, each measuring 60x35x12 mm in length, width

and thickness respectively. Two side screws were used to compress the plates from both

sides and two thickness gauges were placed on either side to keep the thickness uniform.

The edge to edge distance of the two side screws was 35 mm and the sample size of dry

fabric compressed between the two plates was 40x30 mm. For compression at each

pressure level a slip gauge of known thickness was placed between the two plates on

each side. Once on the scanner the side screws were tightened to compress the fabric

preform to the required pressure. From the pressure thickness curve the thickness

against desired pressure was taken and a thickness gauge corresponding to that pressure

was set on both sides of the plates.


174

Figure 6.12. Compression Rig (a) Compression rig fixed on the tomography stage,


The process of CT involves collecting a large set of radiographs (projections) of the

sample during a single rotation. Together with a small number of calibration images,

these images are reconstructed into a 3D volume which represents the attenuation

through the sample. A Nikon Custom 320 Bay tomography system was used for this

analysis. The current was then adjusted to 110µA and the voltage was adjusted to 80 kV

and the white grey level was kept at 62500: 3500 projections were taken for each

tomograph. These images were then reconstructed using Metris X-Tek CT Pro and

taken to the Avizo 7.1 software for analysis. Figure 6.13 represents the sectional views

of the preform, the segmented resin channels and yarns. The 3D reconstruction of the

fabric with resin channels compacted in the plates is shown in Figure 6.14.


175

Figure 6.13. CT images showing slice of (a) warp cross-section, (b) segmented

inter-tow voids and (c) segmented yarns

Figure 6.14. Reconstructed image of single layer fabric compacted in between

polycarbonate plates


The yarn geometry parameters which were calculated during CT analysis of the fabric

sample are shown in Figure 6.15.

Upper and bottom plate

Yarns

Inter-tow voids

a

b

c


176

Figure 6.15. Yarn geometry parameters

The tomography technique has the benefit that the sample can be analysed through the

3D structure at different slices. In the present study, the warp and weft tows were

studied for tow waviness at different slices. The resin channels in single layer dry and

wet fabrics were studied along the slices and in the 3D structure. The results of tow

waviness of dry and wet fabrics were compared at two different pressure levels. Tow

waviness was calculated using equation 6.3.

(6.3)

In equation 6.3, ‘L’ is the crimped length and ‘P’ is the original yarn length without

crimp.

The cross- sectional images of single layer dry and wet preforms are presented in Figure

6.16.


177

Figure 6.16. Cross-sectional images of single layer dry and wet fabrics

6.3.3.1 Tow waviness

The tow waviness results of dry and wet fabrics at 10 and 300 kPa are presented in

Figure 6.17. It can be seen from the results that initially at a pressure level of 10 kPa,

the crimp percentages of the warp yarns in dry fabric were higher than for the wet fabric.

The reason might be that the coefficient of friction decreases in wet fabric on being

wetted due to lubrication of fibres at this lower pressure which results in greater

reduction in the warp yarn crimp percentage of wet fabric compared with dry fabrics.

Also, the crimp percentage yarns in both dry and wet fabrics were higher than the weft

yarns. At a pressure level of 10 kPa, relatively higher crimp percentage was present in


178

the weft yarns of the wet fabric compared to dry fabric. This might be due to the

application of load, the phenomenon of crimp interchange occurs, where crimp in higher

crimped yarn decreases and increases in the less crimped yarn on application of pressure

when both the yarns are oriented in opposite directions. Therefore, at low load there was

higher reduction in the warp yarn crimp of the wet fabric and consequently an increase

in the weft yarn crimp of the same fabric.

(a)

(b)

Figure 6.17. Tow waviness of dry and wet fabrics (a) warp yarns, b) weft yarns


179

At a pressure level of 300 kPa, there was again a decrease in the crimp percentages of

the warp yarns of both dry and wet fabrics. The decrease was higher in the warp yarn

crimp of wet fabric compared to dry fabric, but the difference in crimp percentages of

dry and wet fabrics at this pressure level was low compared to initial pressure level of

10 kPa. At 10 kPa pressure level, the crimp percentage of the warp yarns of wet fabric

was almost 25 % less than the crimp percentage of dry fabric, whereas at 300 kPa, the

difference in crimp percentage for dry and wet warp yarn was 9 %. This shows that at

low pressure, there is more crimp reduction for wet fabrics which contributes towards

the higher fabric thickness reduction in wet fabrics compared to dry fabrics.

6.3.3.2 Inter-tow voids

Dry and wet preforms were also investigated for the resin channels/inter-tow voids and

the results of the inter-tow voids of dry and wet preforms at a pressure level of 10 kPa

and 300 kPa are presented in Figure 6.18. It can be seen from Figure 6.18 that there

were more inter-tow voids in the dry preform compared to the wet preform at a pressure

level of 10 kPa. The greater thickness reduction in the wet preform at this pressure level

also affected the inter-tow voids.


180

Figure 6.18. Inter-tow voids in dry and wet fabric

On application of a higher pressure of 300 kPa, the inter-tow voids reduced in both dry

and wet preforms. Again, at this pressure level, the inter-tow voids were smaller in the

wet preform compared to the dry preform and the reason again is due to the greater

thickness reduction in wet preforms on the application of pressure, which resulted in

more reduction of the inter-tow voids in wet preforms.

(a)


181

(b)

Figure 6.19. Inter-tow voids (a) dry (b) wet fabric at 10 kPa

6.4 Conclusions

The aim of this work was to study the compaction of textile preforms for dry and wet

fabrics. It was observed that there was more deformation in textile preforms upon being

wetted. The fabric compression results at macro-level were fitted with a power-law to

predict wet thickness values from dry fabric results. There was good agreement between

the experimental and fitted result values. At meso-scale, single layer dry and wet fabrics

were studied for tow waviness and inter-tow voids. It was seen that there was a greater

decrease in tow waviness in wet fabrics compared to dry fabrics, which contributed to

the higher thickness reduction in wet fabrics. Also, there was a greater reduction in

inter-tow voids in wet fabric compared to dry fabric.

References

[1] Summerscales J, Searle TJ. Low-pressure (vacuum infusion) techniques for

moulding large composite structures. Proceedings of the Institution of Mechanical

Engineers, Part L: Journal of Materials Design and Applications. 2005;219(1):45-58.


182

[2] Potter KD. The early history of the resin transfer moulding process for aerospace

applications. Composites Part A: Applied Science and Manufacturing. 1999;30(5):619-

21.



1998;19(2):198-216.


1946;37(12):T285-92.



1992;26(16):2330-47.











[10] Debnath S, Madhusoothanan M. Studies on compression properties of polyester

needle-punched nonwoven fabrics under dry and wet conditions. Journal of Industrial

Textiles. 2012;41(4):292-308.


183

[11] Francucci G, Vázquez A, Rodríguez ES. Key differences on the compaction

response of natural and glass fiber preforms in liquid composite molding. Textile

Research Journal. 2012;82(17):1774-85.








Manufacturing. 2006;37(6):868-73.

Chapter 7 Development of an in-situ technique to analyse


184

Chapter 7

DEVELOPMENT OF AN IN-SITU TECHNIQUE TO

ANALYSE THE MESO-STRUCTURE OF DRY

FABRICS UNDER BIAXIAL LOADINGS


Abstract

During the composite manufacturing process textile preforms experience different

forces. Due to the nature of woven fabrics, biaxial forces play an important role on the

tow geometry of the preform. In this research, a biaxial testing machine has been

designed which can measure the biaxial tensile and shear forces applied to the fabric.

An in-situ technique has also been developed to study the meso-structure of the fabrics

under biaxial loading. With the aid of the designed rig, the dry fabric can be scanned

using computed tomography (CT) under in-situ biaxial loadings. CT images of biaxially

loaded glass fabric were captured using this technique. The quality of the CT images

shows the validity of the developed technique for meso-structure analysis.

Keywords: woven preforms, biaxial rig, computed tomography, tow waviness.



185

7.1 Introduction

Textile composites are becoming popular in aerospace, automotive and civil industries

due to their light weight, high strength to weight ratio and ease of handling during

composite manufacturing. In comparison to unidirectional composites, woven fabrics

have advantages due to their higher dimensional stability, better impact resistance and

damage tolerance [1], better drapeability and suitability to manufacture doubly curved

surfaces [2]. Several composite manufacturing techniques such as RTM and

thermoforming involve complex deformations of textile preforms including biaxial

tensile, in-plane shear, transverse compaction and out of plane bending deformations [3].

Various researchers have studied the shear and tensile behaviour of such high

performance fabrics [4-18].

Picture frame [8, 9, 19, 20], Bias extension [3, 16, 18, 21, 22] and Biaxial tests [3, 9]

have been used for shear testing of the fabrics. The deformation behaviour of woven

fabrics is generally of biaxial nature due to the presence of two mutually perpendicular

yarns known as the warp and the weft in the orthogonal fabric. Several researchers have

studied the deformation of textiles under biaxial shear and tensile loadings [9, 23-34].

The study of the biaxial nature of the interactions between yarns in two dimensional

woven fabrics is considered important as it includes the process of crimp interchange

between the yarns, which can affect many structural applications [32]. The study of the

meso-scale geometry during shear and tensile loadings is important during structural

analysis and for making simulation tools.

Hofstee at al. [35] investigated the influence of thermo-forming on the geometry of a

plain woven carbon fabric laminate. They studied the effects of shear deformation on



186

yarn crimp and cross sectional geometry. The fabric samples were stretched and sheared

in a heat chamber above the melting temperature of the resin and the geometry of the

laminate was studied along the warp and weft cross sections using optical microscopy.

Tow geometry under deformation was also studied by Chang et al. [36]. In their study,

they pasted the resin on the surface of the deformed fabric samples under bias extension,

picture frame and biaxial testing. Fabric samples under loading were cured and with the

help of scanning electron microscopy (SEM) images of the samples, different tow

geometry parameters were investigated.

Potluri et al. [3] developed an optical technique for in-plane tow deformation in which

the textile preform was subjected to in-plane stresses ( tensile and shear). The deformed

images were scanned with a flatbed scanner for each deformed state. After application

of a specified biased load, a metallic clamp was placed around the pure shear region

before removing the applied load. Subsequently the fabric samples were impregnated

with epoxy resin and the load was removed after curing of the resin. The laminated

samples were then scanned by SEM and different tow geometry parameters were

calculated.

Analysis of tow deformation in textile preforms subjected to forming forces was

performed by Potluri et al. [28] by using images of laminates obtained by a video

microscope.

All previous research related with the geometry of the fabrics mentioned previously

deals with the analysis of cured fabric geometry after resin infiltration. The laminated

samples may not accurately represent the geometry of the dry fabric as the introduction



187

of the resin into the fabric samples can change the tow geometry of the dry samples [3].

In the present research work, the focus was to establish a technique which may enable

the meso-scale geometry of the dry fabric to be measured under in-situ axial loadings.

To achieve this objective a biaxial testing machine was designed to load the fabric

samples in biaxial directions; this machine is shown in Figure 7.1. In addition an in-situ

rig shown in Figure 7.3 was devloped to hold the deformed geometry between two

frames so that the deformed geometry may be scanned using computed tomography

(CT).

7.2 Development of biaxial testing machine

For testing of the fabric samples under biaxial loadings for both shear and tensile tests, a

biaxial machine was designed by modification to the existing set up of the shear

machine designed by Ciurezu [37]. The machine consists of four linear actuators, which

can move in orthogonal directions and two small load cells of capacity 2.5 kN, which

are mounted on the two perpendicular jaws as shown in Figure 7.1. In the previous set

up of the machine, there was a limitation where it could only control one actuator at a

time and it was not possible to measure the fabric under biaxial loading by moving two

or more actuators simultaneously. In the modified set up of the machine, the

programming has been done using LabVIEW software from National Instruments [38]

to move the actuators with the following combinations:

individual actuators;

two actuators simultaneously;

four actuators simultaneously in the crossing direction.



188

The software, FCT, developed by Festo can also be used to control the linear actuator

motions but it has a limitation in that it can only move one actuator at a time. The

actuators, speed can be varied from 0.1-9.1 mm/min.

Data acquisition from load cells is performed by using the data acquisition facility

(cDAQ 9172) of LabVIEW. Data from the load cells is received in mV/V, which is

converted to Newton force.

Figure 7.1. Rig for biaxial shear and tensile testing



189

Figure 7.2 Biaxial rig with cruciform specimen

7.2.1 Load cell calibration

The loads cells used for acquiring the load values during testing of the samples for shear

and tensile testing need to be calibrated correctly prior to testing in order to achieve

accurate measurements. Here the load cells were calibrated by compressing a 10 kN

load cell in between the two actuators. The LabVIEW software was used for data

acquisition of the output voltage from the load cells .By using the output voltage of the

load cells and the value of the load obtained by the compressed load cell, the calibration

process was completed.

Different readings up to the load value of 2500 N were taken and by applying a trend

line to the curve of the values of the output voltage and load cell readings, the calibrated

values were calculated.



190

7.3 Tomography and in-situ loading rig

Figure 7.3 Rig to grip the biaxial loaded fabric

A rig was designed as shown in Figure 7.3 to clamp the fabric sample under biaxial

loading (Figure 7.2) in such a way that it can be scanned in the computed tomography

machine under in-situ loading without distortion of the biaxial loaded geometry of the

fabric. The rig consists of two square frames of polycarbonate material with dimensions

75 mmx75mm inside the edges. The reason for selecting polycarbonate material was its

low density (1.22 g/cm3) compared to glass yarn (2.60 g/cm

3). The advantage of the low

density of polycarbonate is that fewer x-rays will be absorbed by the frame allowing

more x-rays to pass through the glass fabric giving better quality images. Four holes

were made in the four corners of each frame so that the two frames on the opposite sides

may be tightened with the loaded fabric inside. After the fabric was loaded under a



191

certain load, glue was pasted on the edges of the frame in order to avoid any slippage of

the deformed fabric samples and the screws were tightened on all four sides of the

frames. This clamped fabric was left for 8 hours to allow the glue to dry. During pasting

of the glue on the frame edges, care was taken to avoid any contact of the glue with the

fabric inside the hollow area. After the glue was dried, the fabric outside the clamps was

trimmed and the clamped fabric was taken to the CT for analysis.

The CT analysis was performed on the Nikon Metris 225/320 kV Custom Bay system in

the Materials Science Centre in The University of Manchester. The current and voltage

were adjusted to 100 µA and 60 kV, respectively. For each scan, a total of 3142

projections were taken. These images were then reconstructed using Metris X-Tek CT

Pro software, and the data processed using VSG Avizo 7.1 software.



192

Figure 7.4. CT images of the biaxially loaded dry fabric: a) side view, b) top view

and crossing yarns in two slices

7.4 Conclusion

An in-situ measurement technique has been designed and CT images have been

captured by using this technique which shows the validity of the methodology.

References

[1] Tabiei A, Yi W. Comparative study of predictive methods for woven fabric

composite elastic properties. Composite Structures. 2002;58(1):149-64.

[2] Page J, Wang J. Prediction of shear force using 3D non-linear FEM analyses for a

plain weave carbon fabric in a bias extension state. Finite Elements in Analysis and

Design. 2002;38(8):755-64.

[3] Potluri P, Perez Ciurezu DA, Ramgulam RB. Measurement of meso-scale shear

deformations for modelling textile composites. Composites Part A: Applied Science and

Manufacturing. 2006;37(2):303-14.

[4] Potluri P, Thammandra VS. Influence of uniaxial and biaxial tension on meso-scale

geometry and strain fields in a woven composite. Composite Structures.

2007;77(3):405-18.



193

[5] Boubaker BB, Haussy B, Ganghoffer J-F. Consideration of the yarn–yarn

interactions in meso/macro discrete model of fabric: Part II: Woven fabric under

uniaxial and biaxial extension. Mechanics Research Communications. 2007;34(4):371-8.

[6] Quaglini V, Corazza C, Poggi C. Experimental characterization of orthotropic

technical textiles under uniaxial and biaxial loading. Composites Part A: Applied


[7] T.V. Sagar PP, Hearle JWS. Energy approach to predict uniaxial/biaxial load-

deformation of woven preforms. ECCM 10. 2002.

[8] Mohammed U, Lekakou C, Dong L, Bader MG. Shear deformation and

micromechanics of woven fabrics. Composites Part A: Applied Science and

Manufacturing. 2000;31(4):299-308.

[9] Sharma SB, Sutcliffe MPF, Chang SH. Characterisation of material properties for

draping of dry woven composite material. Composites Part A: Applied Science and

Manufacturing. 2003;34(12):1167-75.

[10] Harrison P, Potluri P, Bandara K, Long AC. A normalisation procedure for biaxial

bias extension tests. International Journal of Material Forming. 2008;1(1):863-6.

[11] Sharma SB, Sutcliffe MPF. A simplified finite element model for draping of woven

material. Composites Part A: Applied Science and Manufacturing. 2004;35(6):637-43.

[12] Hivet G, Vidal-Salle E, Boisse P. Analysis of the stress components in a textile

composite reinforcement. Journal of Composite Materials. 2012;47(3):269-85.

[13] Badel P, Vidal-Sallé E, Boisse P. Computational determination of in-plane shear

mechanical behaviour of textile composite reinforcements. Computational Materials

Science. 2007;40(4):439-48.



194

[14] Launay J, Hivet G, Duong AV, Boisse P. Experimental analysis of the influence of

tensions on in plane shear behaviour of woven composite reinforcements. Composites


[15] Dong L, Lekakou C, Bader MG. Processing of Composites: Simulations of the

Draping of Fabrics with Updated Material Behaviour Law. Journal of Composite

Materials. 2001;35(2):138-63.

[16] Harrison P, Clifford MJ, Long AC. Shear characterisation of viscous woven textile

composites: a comparison between picture frame and bias extension experiments.

Composites Science and Technology. 2004;64(10–11):1453-65.

[17] Boisse P, Zouari B, Gasser A. A mesoscopic approach for the simulation of woven

fibre composite forming. Composites Science and Technology. 2005;65(3–4):429-36.

[18] Cao J, Akkerman R, Boisse P, Chen J, Cheng HS, de Graaf EF, et al.

Characterization of mechanical behavior of woven fabrics: Experimental methods and

benchmark results. Composites Part A: Applied Science and Manufacturing.

2008;39(6):1037-53.

[19] Peng XQ, Cao J. A continuum mechanics-based non-orthogonal constitutive model

for woven composite fabrics. Composites Part A: Applied Science and Manufacturing.

2005;36(6):859-74.

[20] Nguyen M, Herszberg I, Paton R. The shear properties of woven carbon fabric.

Composite Structures. 1999;47(1–4):767-79.

[21] Lebrun G, Bureau MN, Denault J. 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):341-52.



195

[22] Wang J, Page JR, Paton R. Experimental investigation of the draping properties of

reinforcement fabrics. Composites Science and Technology. 1998;58(2):229-37.

[23] Potluri P DA, Young RJ. Biaxial shear testing of textile preforms for formability

analysis. ICCM 16. 2007.

[24] Boisse P, Gasser A, Hivet G. Analyses of fabric tensile behaviour: determination of

the biaxial tension–strain surfaces and their use in forming simulations. Composites Part

A: Applied Science and Manufacturing. 2001;32(10):1395-414.

[25] Pan N. Analysis of woven fabric strengths: Prediction of fabric strength under

uniaxial and biaxial extensions. Composites Science and Technology. 1996;56(3):311-

27.

[26] Buet-Gautier K, Boisse P. Experimental analysis and modeling of biaxial

mechanical behavior of woven composite reinforcements. Experimental Mechanics.

2001;41(3):260-9.

[27] Gasser A, Boisse P, Hanklar S. Mechanical behaviour of dry fabric reinforcements.

3D simulations versus biaxial tests. Computational Materials Science. 2000;17(1):7-20.



2006;66(2):297-305.

[29] Boisse P, Borr M, Buet K, Cherouat A. Finite element simulations of textile

composite forming including the biaxial fabric behaviour. Composites Part B:

Engineering. 1997;28(4):453-64.

[30] Willems A, Lomov SV, Verpoest I, Vandepitte D. Drape-ability characterization of

textile composite reinforcements using digital image correlation. Optics and Lasers in

Engineering. 2009;47(3–4):343-51.



196

[31] Boisse P, Buet K, Gasser A, Launay J. Meso/macro-mechanical behaviour of

textile reinforcements for thin composites. Composites Science and Technology.

2001;61(3):395-401.

[32] Zhu D, Mobasher B, Vaidya A, Rajan SD. Mechanical behaviors of Kevlar 49

fabric subjected to uniaxial, biaxial tension and in-plane large shear deformation.


[33] Xue P, Peng X, Cao J. A non-orthogonal constitutive model for characterizing

woven composites. Composites Part A: Applied Science and Manufacturing.

2003;34(2):183-93.

[34] Hivet G, Boisse P. Consistent mesoscopic mechanical behaviour model for woven

composite reinforcements in biaxial tension. Composites Part B: Engineering.

2008;39(2):345-61.

[35] Hofstee J, de Boer H, van Keulen F. Elastic stiffness analysis of a thermo-formed

plain-weave fabric composite: Part I: geometry. Composites Science and Technology.

2000;60(7):1041-53.

[36] Chang SH, Sharma SB, Sutcliffe MPF. Microscopic investigation of tow geometry

of a dry satin weave fabric during deformation. Composites Science and Technology.

2003;63(1):99-111.

[37] Ciurezu DAP. Characterization of large shear deformation in textile preforms. PhD

thesis. 2007.

[38] http://www.ni.com/labview/.

http://www.ni.com/labview/

Chapter 8 Conclusions and future work

197

Chapter 8

Conclusions and Future work

8.1 Conclusions

In the present research work, the deformation behaviour of textile preforms was

studied under transverse compaction and an in-situ technique was introduced to

measure the meso-structure of dry textile preforms during biaxial tensile and shear

loading.

The deformation under compaction was studied at macro and meso-levels for single

layer and multilayer stacks of different architectures. Glass fabrics and non-crimp

fabrics (NCF) were used in this study. To investigate the meso-structure of glass

fabrics, an in-situ rig was designed which can be used to study the meso-structure of

textile fabrics under compaction loadings using computed tomography (CT).

Additionally, comparisons of dry and wet fabrics were made to investigate the

compaction behaviour of the fabrics after resin infiltration.

The study of single layer preforms was conducted for plain woven glass fabric

(chapter 2). Macro-scale deformation was studied by using mechanical testing of

single layer fabrics on an Instron testing machine using static loadings. The fabric

was tested in the dry and wet states. Higher deformation was observed in wet fabric


198

compared to dry fabric. Also it was observed that fabric thickness dropped to

maintain a constant pressure which was considered to be due to rearrangement of the

fibres. A power-law relationship was derived to predict the final fabric thickness

values for every load and good agreement was observed between the predicted and

the experimental results.

The meso-structure of single layer fabric was studied by compressing the fabric

sample in the designed rig and scanning it by computed tomography (CT) at

increased pressure levels up to 600 kPa. It was observed that fabric thickness

reduction at low loads up to 100 kPa was mainly due to reduction in tow waviness.

The phenomenon of crimp interchange was also observed at low loads. Changes in

the other parameters like tow area, tow thickness and tow width were not prominent

during these loadings. The reduction of inter-tow voids with application of pressure

was greater for pressures up to 100 kPa. On increment of pressure to 300 and 600

kPa, there was noticeable deformation in the tow area and tow thickness which

resulted in higher fibre volume fractions at these loads. CT was seen as an accurate

tool to calculate fibre volume fraction (FVF) as FVF calculated by mechanical test

results and CT were in good agreement.

The deformation behaviour of multilayer stacks was studied for 6 layer plain woven

glass fabric (chapter 4). Tow geometry and inter-tow voids were analysed for a

pressure range up to 100 kPa. It was seen that nesting of the layers was a major

source of stack thickness reduction in this low load regime. The main reason

controlling single layer thickness in this pressure range was reduction in tow

waviness. As for single layer preforms, crimp interchange was noticed in the

multilayer preform. Also it was seen that the resin channels in multilayer preforms

were highly influenced by the presence of nesting in layers. Maximum resin channels


199

were seen in slices where fabric layers were perfectly sitted upon each other and vice

versa. Due to the small decrease in tow area and thickness at these low pressure

values, the increase in yarn packing fractions was not prominent in multilayer stacks.

The deformation behaviour of textile fabrics with different architectures was also

studied at macro-level. Woven fabrics developed with plain, twill and sateen weaves

with the same tow count and tow spacing and non-crimp fabrics (NCF) were

included in this study (chapter 5). It was observed that the float length of the weave

plays an important role in the compression of woven fabrics. Increase in nesting was

observed with shorter float length and with increasing numbers of layers for woven

fabrics. In the case of NCF, it was seen that the fabrics placed at 00/90

0 orientation

exhibited higher nesting. However, no significant increase was observed when the

number of layers was increased to more than two. NCF placed at 00/90

0 showed even

better nesting than plain woven fabrics.

In addition to dry fabrics, the compaction behaviour of wet fabrics was also studied

at macro and meso-scales (chapter 6). Plain, twill and sateen fabrics of glass yarn and

non-crimp fabric of carbon yarn were included in the deformation study at macro-

level. It was observed that there was greater thickness reduction in wet fabrics

compared to dry fabrics. The power-law relation derived to predict wet fabric

thickness from dry thickness results showed good agreement between experimental

and predicted thickness results. The deformation behaviour of single layer wet

fabrics at meso-level was investigated using CT and compared with dry fabric meso-

structures for tow waviness and resin channels. The results showed that there was

higher reduction in tow waviness and inter-tow voids in wet fabrics.


200

In addition to study of the deformation behaviour of textile preforms under

transverse compression, an in-situ measurement technique was designed to study the

deformation behaviour of textile preforms during biaxial tensile and shear loadings

(chapter 7). CT images of plain woven glass fabric captured under in-situ biaxial

loading showed that this technique can be used to accurately measure the meso-

structure of dry textile preforms under biaxial tensile and shear loadings.

8.2 Future work

Research work on the deformation behaviour of dry textiles under forming forces can

be further carried out to analyse the changes at macro and meso scales. Following are

the main aspects of the future research work.

8.2.1 Simulation of the resin permeability during composite manufacturing

Experimental data have been achieved for the voids in single and multilayer fabrics

that now can be used to develop resin simulation tools during the composite

manufacturing process.

8.2.2 Meso-study of NCF and different woven architectures

A macro level deformation study of NCF and different woven fabrics has been

conducted and compared for thickness and nesting of layers. Further studies at meso

level can be performed to analyse the structural changes during compaction and to

study the effect of tow interlacement on the nesting of layers and resin channels.

8.2.3 Deformations of textile fabrics under biaxial shear and tensile loadings

Macro-structure analysis of different fabrics can be performed by using the

developed biaxial rig under biaxial shear and tensile loadings. For biaxial shear


201

loadings, existing high capacity load cells can be replaced by small capacity load

cells so that deformations at low loads may be recorded accurately. With the aid of

in-situ loading device meso structure analysis of fabrics for tow geometry changes

under biaxial shear and tensile loadings can also be performed using computed

tomography (CT).

Publications

202

Publications

1. Conference paper presented in Deformation and Fracture of Composites (DFC-

12) & Structural Integrity and Multi-scale Modelling (SI-6) at Queens’ College

Cambridge 8-11 April 2013University of Cambridge with title “Meso-scale

geometry and nesting of 2D woven fabrics during compaction”.

2. Conference paper presented in ICCM 19, July 2013, Montreal, Quebec, Canada

with title “Meso-scale analysis of 2D glass woven preform under compaction”.

3. “Deformation of single layer dry preforms under compaction” ready to submit in

Composites Part-A.

4. “Deformation of multilayer dry preforms under compaction” ready to submit in

Composites Part-A

5. “Compaction and nesting in textile preforms influenced by tow architecture”

ready to submit in Applied Composite Materials.

investigation of deformation behaviour of dry textiles

Documents