Full Text

INVESTIGATION OF DIFFERENT GEOMETRIC
STRUCTURE PARAMETER FOR HONEYCOMB

TEXTILE COMPOSITES ON THEIR
MECHANICAL PERFORMANCE
A thesis submitted to the University of Manchester for the degree of

Doctor of Philosophy
In the Faculty of Engineering and Physical Sciences
By
Xiaozhou Gong
MAY 2011
1
LIST OF CONTENTS
LIST OF CONTENTS............................................................................................ 2
LIST OF PUBLICATIONS.................................................................................... 8
LIST OF FIGURES................................................................................................ 9
LIST OF TABLES.................................................................................................. 15
DECLARATION..................................................................................................... 17
COPYRIGHT STATEMENT................................................................................ 18
ABSTRACT.............................................................................................................. 19
ACKNOWLEGEMENT.......................................................................................... 20
CHAPTER 1 INTRODUCTION.................................................................... 21
1.1 Description of the problem....................................................... 23

1.2 Research aim and objectives..................................................... 24
1.3 Thesis layout............................................................................... 27
CHAPTER 2 LITERATURE REVIEW........................................................ 28
2.1 Classification of Cellular Solids……………………………... 28

2.1.1 Honeycomb structure…………………………………………... 29
2.1.2 Foam structure…………………………………………………. 31
2.2 Main features of cellular solids………………………………. 31
2.2.1 Low density……………………………………………………. 33
2.2.2 Stiffness and strength of cellular solids………………………... 35
2.2.3 Open porosity structure and its application……………………. 41
2.2.4 Thermal insulation property…………………………………… 44
2.3 Manufacturing of honeycomb structure……………………. 45
2.4 Mechanical performances of cellular solids…………………... 47
2.4.1 Previous studies on cellular solids’ mechanic performances…... 48
2
2.4.2 Dynamic impact with different velocities……………………... 54
2.4.3 Energy absorption of cellular solids…………………………… 56
2.5 Textile Honeycomb Composites………………………………. 58
2.5.1 3D woven fabrics……………………………………………… 60
2.5.2 3D honeycomb fabrics………………………………………… 61
2.5.3 Structure parameters for textile honeycomb composite…………. 64
2.6 Applications of textile honeycomb composite on PPE……...... 66
2.7 Comments……………………………………………………… 68
CHAPTER 3 DESIGN OF 3D HONEYCOMB FABRICS............................ 70
3.1 Design of 3D honeycomb weaves................................................ 70

3.1.1 Representation of woven honeycomb structures......................... 72
3.1.2 Layer connection methods........................................................... 73
3.1.3 Weave creation............................................................................. 74
3.2 Design of 3D honeycomb fabrics.............................................. 77
3.2.1 3D honeycomb fabrics................................................................ 77
3.2.2 Design details for 3D honeycomb fabrics................................... 81
3.2.2.1 Cell opening angle, …………………………………………... 81
3.2.2.2 Different cell size at the same number of layers......................... 82
lb
3.2.2.3 Length ratio of cell walls ( )…………………………………
lf 83
3.2.2.4 Similar sample thickness with different cell size........................ 85

3.3 Manufacturing of 3D honeycomb fabrics................................ 86
3.3.1 Weft density of the 3D honeycomb fabric.................................. 86
3.3.2 Parameter specifications for 3D honeycomb fabric in the
weaving process............................................................................. 87
3.3.3 Honeycomb fabric production…………………………………. 91
CHAPTER 4 CREATION OF HONEYCOMB COMPOSITES AND TEST

SAMPLE PREPARATION……………………………………. 95
3
4.1 Fabric opening and consolidation……………………………. 95
4.1.1 Fabric opening............................................................................. 95
4.1.2 Fabric impregnation..................................................................... 100
4.1.3 Textile honeycomb composite..................................................... 102
4.2 Fabrication of woven honeycomb composite………………… 103
4.3 The sample groups…………………………………………….. 104
4.4 Summaries……………………………………………………..... 108
CHAPTER 5 EXPERIMENTAL DATA ANALYSIS ON TEXTILE

HONEYCOMB COMPOSITES................................................. 109
5.1 Low velocity drop weight impact tests……………………….. 109

5.1.1 Basic principle of low velocity drop weight impact……………. 109
5.1.2 The set-up of the low velocity impact instrument…………….... 111
5.1.3 Test procedure………………………………………………….. 114
5.2 Preparation for test…………………………………………… 115
5.2.1 Specimens of textile honeycomb composites…………………... 115
5.2.2 Impact test setting……………………………………………….. 115
5.3 Impact test results……………………………………………... 116
5.3.1 Data processing…………………………………………………. 116
5.3.1.1 Basics for low-velocity impact test……………………………... 118
5.3.1.2 Force attenuation………………………………………………... 119
5.3.1.3 Acceleration of the impactor……………………………………. 120
5.3.1.4 Characteristics of the transmitted force………………………… 121
5.3.1.5 Energy absorption performance………………………………… 122
5.4 Experiment results…………………………………………….. 126
5.4.1 Various experiment results during impact procedure…………... 126
5.4.2 Experiment results for energy absorption………………………. 127
5.4.3 Experiment results for force attenuation factor (fatt)……………. 128
5.5 Structure and properties of textile honeycomb composites…. 129
4
5.5.1 Structure parameters and performance indices…………………. 129
5.5.2 Grouped sample experimental performance……………………. 129
5.5.2.1 Cell size and its experimental performance (8L3P60, 8L4P60,
8L5P60, 8L6P60).………………………………………………. 129
5.5.2.2 Opening angle and its experimental performance(8L6P30,
8L6P45, 8L6P60, 8L6P75, 8L6P90)…………………………….. 134
5.5.2.3 Length ratio of cell walls and its experiment performance
lb l
(  1 : 8L3P60,8L(4+3)P60, 8L(6+3)P60; b  1 : 8L(3+6)
lf lf
P60,8L(4+6)P60,8L6P60) …………………………….………… 140

5.5.2.4 Honeycomb composites with similar thickness and their
performance (4L6P60, 6L4P60 and 8L3P60) …………………… 146
5.5.3 Discussions on composite density and composite thickness…….. 149
5.5.3.1 Composite volume density………………………………………. 150
5.5.3.2 Composite thickness…………………………………………….. 153
5.6 Conclusions……………………………………………………... 155
CHAPTER 6 EXPERIMENTAL DATA ANALYSIS ON TEXTILE

HONEYCOMB COMPOSITE IMPACTED WITH
LARGER MASS AND LOWER VELOCITY………………. 155
6.1 Low velocity impact test setting by Instron Dynatup Model

156
8200 drop weight impact testing instrument………………….
6.1.1 Assembly of Instron Dynatup Model 8200 drop weight impact

testing instrument………………………………………………... 156
6.1.2 Testing procedure………………………………………………... 159

6.1.3 Classifications of textile honeycomb composites………………..
159
6.1.4 Impact setting for Instron Dynatup Model 8200 system…………
160
6.2 Tested results and discussion………………………………….. 161
6.2.1 Cell size and its experimental performance (8L3P60, 8L5P60, 161
5
8L6P60)…………………………………………………………
6.2.2 Opening angle and its experimental performance (8L6P30,
8L6P45, 8L6P60 and 8L6P75)…………………………………
164
6.2.3 Different length ratio of bonded and free wall and its experiment
performance ( lb  1 : 8L(3+6)P60, 8L(4+6)P60, 8L6P60;

lf
lb
 1 : 8L3P60, 8L(4+3)P60, 8L(6+3)P60)…………………… 166
lf
6.2.4 Comparison of the results between two different loading

170
conditions…………………………………………………………
6.2.4.1 Samples with different cell size (8L3P60, 8L5P60, 8L6P60)…… 170
6.2.4.2 Samples with different opening angle (8L6P30, 8L6P45,
8L6P60, 8L6P75)…………………………………………………
173
6.2.4.3 Samples with different length ratio of free and bonded wall
( lb  1 : 8L(3+6)P60, 8L(4+6)P60, 8L6P60;

lf
lb
 1 : 8L3P60, 8L(4+3)P60, 8L(6+3)P60)……………….......... 176
lf
6.3 Summaries……………………………………………………… 181
CHAPTER 7 FEA ON TEXTILE HONEYCOMB COMPOSITES……….. 182
7.1 FEA Based on 2D Honeycomb Composite Models…………… 183

7.1.1 Creation of 2D models for textile honeycomb composites……… 183
7.1.2 Meshing the geometrical models and the impactor……………… 186
7.1.3 Meshing the impactor……………………………………………. 187
7.1.4 Materials………………………………………………………… 188
7.1.4.1 The tensile test of a single layer composite……………………… 188
7.1.4.2 Material properties………………………………………………. 189
7.1.5 Boundary conditions applied to the honeycomb composite
models…………………………………………………………… 190
6
7.1.6 Impact setting for FEA of 2Dmodels……………………………. 191
7.1.7 Results and discussions of FEA based on 2D models………… 192
7.1.7.1 Introduction of performance indices…………………………… 192
7.1.7.2 Classifications of the FE composite models…………………… 195
7.1.7.3 Simulated results…………………………………………………. 196
7.1.7.4 Deformation area under cylinder impact………………………… 202
7.1.7.5 History of dynamic contact force……………………………… 209
7.1.7.6 Energy absorption performance………………………………… 213
7.1.7.7 Comparison betwen ball and cylinder impact………………… 214
7.1.7.8 Validation of the simulation results with experiment results…… 222
7.2 FEA of 3D Textile Honeycomb Composites………………… 223
7.2.1 Creation of the geometric models……………………………… 223
7.2.2 Boundary conditions…………………………………………… 225
7.2.3 Set-up of 3D FE models………………………………………… 225
7.2.4 3D FE results and discussions…………………………………… 226
7.3 Summaries on FEA…………………………………………… 232
CHAPTE 8 CONCLUSIONS AND FUTURE WORK…………………….. 234
8.1 Conclusions…………………………………………………… 234

8.2 Recommendations for Further Research Work……………… 237
REFERENCE…………………………………………………………... 240
7
LIST OF PUBLICATIONS
(i) Xiaogang Chen, Ying Sun and Xiaozhou Gong. (2008). Design,
Manufacture, and Experimental Analysis of 3D Honeycomb Textile Composite Part I:
Design and Manufacture. Textile Research Journal, vol.78. no.9, pp.771-781
(ii) Xiaogang Chen, Ying Sun and Xiaozhou Gong. (2008). Design, Manufacture,
and Experimental Analysis of 3D Honeycomb Textile Composites, Part II:
Experimental Analysis. Textile Research Journal, vol.78, no.11, pp.1011-102
(iii) Xiaogang Chen, Ada Gong, Ying Sun, Daniel Yu. (2006). 3D Honeycomb
Textile Composites for Impact Protection. In: Kang,T.J. ed. International fibre
conference, vol. A4-1, May-June, Korea
(iv) X.Chen and X.Gong.(2008). Manufacture and Characterization of Exatra-

light 3D Hollow Textile Composite, ECCM13 Conference, Jun, Sweden
(v) Xiaogang Chen, Xiaozhou Gong and Shijun Tang.(2008). Design,

Manufacture, and Analysis of 3D Honeycomb Textile Composites. ECCM13
Conference, Jun, Stockholm，Sweden
8
LIST OF FIGURES
Figure 2-1 Examples of cellular solids (Gibson and Ashby, 1997) 30

Figure 2-2 Schematic illustration of honeycombs structure will different cell
31
shape (Gibson and Ashby, 1997)
Figure 2-3 The range of properties available to the engineer through foaming
32
(Gibson and Ashby, 1997)
Figure 2-4 A chart showing material Young’s modulus and density where each
material class occupies a characteristic field on the chart (Pflug and
34
Vangrimde, 2003)
Figure 2-5 Honeycomb structure with hexagonal cells (Gibson and Ashby,
36
1997)
Figure 2-6 Typical compressive stress-strain curves for cellular solids under in-
plane compression (Gibson and Ashby, 1997) 37
Figure 2-7 Geometric parameters of the honeycomb cell from Gibson and
38
Ashby (1997)
Figure 2-8 A schematic diagram shows the way the stress-strain curve changes
40
with t/l (Gibson and Ashby, 1997)
Figure 2-9 Schematic detailed description of the honeycomb sandwich
42
structure (Abbadi et al., 2009)
Figure 2-10 Different Sandwich core types (Herrmann et al., 2005) 43
Figure 2-11 Examples for sandwich application A380 (Herrmann et al., 2005) 43
Figure 2-12 Expansion manufacturing process (Bitzer, 1997) 46
Figure 2-13 Corrugation manufacturing process (Bitzer, 1997) 46
Figure 2-14 The peak stresses generated in foam of three densities in absorbing
49
the same energy, W (Gibson and Ashby, 1997)
Figure 2-15 Typical time-load pulses from uniaxial crushing tests for Redwood
50
specimens (Reid and Peng, 1997)
Figure 2-16 Crushing of a honeycomb in the X1 direction, where v is the initial
53
crushing velocity (Ruan et al., 2005)
9
Figure 2-17 Sketch of spacer fabric construction 59
Figure 2-18
3D woven composites (a) cylinder and flange; (b) egg crate
structure; (c) turbine rotors; and (d) various complex shapes woven 60
preforms (Mouritz et al., 1999)
Figure 2-19 Woven architectures used in 3D woven composites (Yi and Ding,
61
2004)
Figure 2-20 A schematic diagram of woven fabric with multilayer (Takenata et
63
al., 1991)
Figure 2-21 Cross section view of the honeycomb fabric in 3D form (Yassar,
64
1999)
Figure 2-22 Parameters of single honeycomb cell (Tan and Chen, 2005) 65
Figure 2-23 Schematic diagram of a 6-layer honeycomb structure (Sun, 2005) 66
Figure 3-1 Region division of a honeycomb structure 71
Figure 3-2 Selection of weave 73
Figure 3-3 Honeycomb structure 2L1P 75
Figure 3-4 Honeycomb structure 4L3P 76
Figure 3-5 8L6P with different opening angle 82
Figure 3-6 Different cell size for 8-layer composites 83
Figure 3-7 lb
Honeycomb structure with length ratio of cell walls (  1)
lf 85
Figure 3-8 lb
Honeycomb structures with length ratio of cell walls (  1)
lf 85
Figure 3-9 Structures with same thickness and different cell size 86
Figure 3-10 Weave lifting plan for 8L3P, 8L4P, 8L5P and 8L6P 90
Figure 3-11 The dobby weaving machine 92
Figure 3-12 Card punching 92
Figure 3-13 Photograph of one sample fabric weaved from loom 94
Figure 4-1 Honeycomb fabric opening devices 96
Figure 4-2 Illustration of the thickness (T) of the honeycomb structure 97
Figure 4-3 Illustration of a four-layer honeycomb composite 98
10
Figure 4-4 Photos of textile honeycomb composite with different cell size 103
Figure 4-5 Specimens with different opening angle 105
Figure 4-6 Specimens with different cell sizes 105
Figure 4-7 lb
Specimens with different length ratios (  1)
lf 106
Figure 4-8 lb
Specimens with different length ratio of cell walls (  1)
lf 107
Figure 4-9 Specimens with same thickness 107

Figure 5-1 Schematic diagram for in-plane low velocity impact test 110
Figure 5-2 Dropping hammer system for impact test of specimens 111
Figure 5-3 The impactor and the anvil 112
Figure 5-4 Charge amplifier used in the tests 113
Figure 5-5 Snapshot of the resultant curves for force and acceleration displayed
113
in Nicolet Windows
Figure5-6 Data processing flow chart for experimental data analysis 116
procedures
Figure 5-7 Measured acceleration curves for 8L6P60 121
Figure 5-8 Measured transmitted force curves for 8L6P60 122
Figure 5-9 The response of contact force against displacement 123
Figure 5-10 Trapezoidal method to calculate the energy absorption 124
Figure 5-11 Evaluation curves of velocity, displacement and energy absorption
125
for 8L6P60
Figure 5-12 Comparison of transmitted force-time diagram 130
Figure 5-13 Comparison of value of peak transmitted force diagram 130
Figure 5-14 Comparison of contact force-displacement diagram 131
Figure 5-15 Comparison of energy absorption and structure displacement
131
diagram
Figure 5-17 Comparison of value of peak transmitted force diagram 136
Figure 5-19 Comparison of energy absorption and structure displacement 137
11
diagram
Figure 5-20 Energy dissipation direction diagram 139
Figure 5-23 Comparison of energy absorption diagram 145
Figure 5-26 Influence of volume density on honeycomb composites 151
Figure 5-27 Influence of composite thickness on honeycomb composites 152
Figure 6-1 Instron Dynatup Model 8200 drop weight impact testing machine 158
Figure 6-2 Contact force and energy absorption behaviour of samples with
162
different cell size
Figure 6-3 Contact force and energy absorption behaviour of samples with
165
different opening angle
Figure 6-4 Energy absorption of samples with different length ratio of free and
170
bonded wall
Figure 6-5 Contact force-displacement curves of composite with different cell
171
sizes
Figure 6-6 Energy absorption under different impact situation (samples with
173
different cell size)
Figure 6-7 Contact force-displacement curves for composites with different
175
opening angles
Figure 6-8 Energy absorption under different loading conditions (samples with
175
different opening angle)
Figure 6-9 Contact force-displacement curve and energy absorption diagram
for the sample with different length ratio of bond and free wall
178
( lb  1 )
lf
Figure 6-10 Contact force-displacement curve and energy absorption diagram

180
for the sample with different length ratio of bond and free wall
12
( lb  1 )
lf
Figure 7-1 Meshing of a cell 187

Figure 7-2 Meshed impactors (a) the cylinder and (b) the sphere 188
Figure 7-3 Stress-strain behaviour of cotton/epoxy sheet and steel 190
Figure 7-4 Schematic illustration of boundary conditions for the FE impact
191
model
Figure 7-5 Estimation of deformed cross-section represented by a trapezoidal
193
shaped area (use 8L6P60 as an example)
Figure 7-6 Deformation of models with different cell sizes under impact
204
energy of 6J, 8J and 10J
Figure 7-7 Comparison of deformation area in models with different opening
206
angles
Figure 7-8 Comparison of deformation area ratio of models with different cell
lb 207
wall ratio (  1 ).
lf
Figure 7-9 Comparison of deformation area in models with different cell wall
lb 208
ratio (  1)
lf
Figure 7-10 Dynamic contact force of models under the impact energy of 8J 212
Figure 7-11 Peak contact force from cylinder impact 213
Figure 7-12 Validation of energy absorption between FEA and experiment
214
results
Figure 7-13 Comparison of contact force-time response of 8L3P60 under 8j by
215
cylinder and ball impact
Figure 7-14 Comparison of structure deformation under dynamic impact for
220
model 8L3P at 8J impact (a) by cylinder impact (b) by ball impact
Figure 7-15 Dynamic contact force of models under ball impact at 8J 221
Figure 7-16 Comparison between ball and cylinder energy absorption capability
222
Figure 7-17 Comparison of contact force between experiment and simulated 224
13
(2D) results for 8L3P60
Figure 7-18 Created honeycomb model 225
Figure 7-19 FEA and experiment results from 3D scale 229
Figure 7-20 Relationship of input force and transmitted force (a) 8L3P60 and
(b) 8L4P60 through FE simulation 231
14
LIST OF TABLES
Table 2-1 Density of cellular solids and solid material (Gibson and Ashby, 34
1997)
Table 3-1 List of fabric types with weaving quantity and design angle 79
Table 3-2 Experimental design outline in groups 80
Table 4-1 Calculated sample height and distance between wire and other 99
design parameter
Table 4-2 Cell geometric parameters for testing specimens 102
Table 5-1 Experiment results from impact test 126
Table 5-2 Experiment results for the energy protection 127
Table 5-3 Experiment results for force attenuation 128
Table 5-4 Experiment results of the energy dissipated along vertical and 139
horizontal direction
Table 5-5 Experiment results (length ratio of cell walls) 141
Table 5-6 Experiment results (samples with similar thickness) 146
Table 5-7 Volume density, sample thickness, energy absorption and peak
149
transmitted force of different composites
Table 6-1 Experiment results from impact (samples with different cell size: 161
8L3P60, 8L5P60 and 8L6P60
Table 6-2 Experiment results from impact (samples with opening angle: 164
8L6P30, 8L6P45, 8L6P60 and 8L6P75)
Table 6-3 Experiment results from impact (samples with different length ratio 167
of bonded and free wall)
Table 7-1 Schematic illustrations with structural parameters of 12 geometric 184
models
Table 7-2 Mechanical properties of materials 189
Table 7-3 Impactor mass, impact velocity and impact energy 191
Table 7-4 Details of FE models 192
15
Table 7-5 Effect of cell size on models under cylinder impact 196
Table 7-6 Effect of cell opening angle in the models under cylinder impact 198
lb
Table 7-7 Effect of cell wall ratio (  1 ) on the models under cylinder
lf
200
impact
lb
Table 7-8 Effect of cell wall ratio (  ) on the models under impact 201
lf
Effect of cell size on its maximum displacement and energy

Table 7-9 216
absorption for textile honeycomb composite models under ball
impact
Table 7-10 Dimension of cylinder impactor 225
Table 7-11 Dimension and cell parameter for the models 226
16
DECLARATION
No portion of the work referred to in the thesis has been submitted in

support of an application for another degree or qualification this or
qualification of this or any other university or other institute of learning.
17
COPYRIGHT STATEMENT
1. The author of this thesis (including any appendices and/or schedules to this
thesis) owns certain copyright or related right in it (the „Copyright‟) and s/he has given
The university of Manchester certain rights to use Copyright, including for
administrative purpose.
2. Copies of this thesis, either in full or in extracts and whether in hard or

electronic copy, may be 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.
3. The ownership of certain Copyright patents, designs, trade marks and any and
all 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 Rights 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 Rights and/or Reproducitons.
4. Further information on the conditions under which disclosure, publication and

commercialisation of this thesis, the Copyright and any Intellectual Property and/or
Reproductions described in it may take place is available in the University IP Policy
(see http://www.campus.manchester.ac.uk/medialibrary/policies/intellectual-
property.pdf), 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.
18
ABSTRACT
Textile honeycomb composites, with an array of hexagonal cells in the cross section, is
a type of textile composites having the advantage of being light weight and energy
absorbent over the solid composite materials. The aim of this research is to investigate
the influence of the geometric parameters on textile honeycomb composites on their
mechanical performances under low velocity impact, which can be used to help
designer control over the textile honeycomb composites.
Four groups of textile honeycomb composites, involving 14 varieties, have been

systematically created for the experimental analysis. The geometric parameters of the
honeycomb composites, including the cell opening angle, cell size, cell wall length ratio
and structural parameters such as composite thickness, composite volume density are
studied for their influence on the honeycomb composites under low-velocity impact.
Followed by experimental work, honeycomb composites with 12 varieties are modelled
by finite element method (FEM) to further investigate the honeycomb structure
performance under various loading condition including different impact energy (6J, 8.3J
and 10J) and impactor shape (cylindrical and spherical).
The 3D honeycomb fabrics are successfully manufactured and converted into textile
honeycomb composites. It was found through the experimental and finite element
analysis (FEA) that changes in geometric and structural parameters of the textile
honeycomb composites have noted influences on the energy absorption, force
attenuation and damage process of the structure. The length ratio of cell wall and the
cell opening angle are the most effective parameters for controlling the energy
absorption of the composites and composites with medium cell sizes tend to have more
reliable mechanical performances under various loading conditions. And it is also found
in FEA that cylindrical impacts are more threatening to human beings than the ball
shaped impact. The methodology has been established by using FEM to investigate the
composites more systematically in the current study. This helps to provide a faster and
economic design cycle for the honeycomb composites, which can substantially decease
the time to take products from concept to the production.
19
ACKNOWLEDGMENT
I would like to thank my supervisor, Dr.Chen, Xiaogang, for his guidance, help and
support throughout this research project.
Many thanks also go to Mr. Zadoroshnyj, A. from material centre for his helping in
producing resin/cotton composite and Mr. Steve from mechanical centre for setting
equipment and assisting test the composites patiently.
Financial support from school of science and physics (EPS) and Dr.Chen, Xiaogang’s
research budget is gratefully acknowledged.
My special thanks go to Mr. Robson, M.T. for his encouragement and support initially
from the starting of this degree. I would like to acknowledge the helping of Mr.Yin’s
family for their helping personally during the study and Dr.Wei,H. & Dr.Wang,H.W.
Mr.Lee,P., Dr. Smith,M. and Mrs.Tina for their assistants during later thesis writing
stage. A lot of thanks to my friends Mrs.Chen,L. and Mrs.Shen,J.L. in China for their
endless love and encouragements to help me emotionally and finically in the late stage
of the Phd study. Thanks also go to Mrs.Zhen,J.C. and his parents in China for their
great help during my study.
Thanks also extend to my dear colleges including Dr.Yu, Daniel, Mrs.Sun,Y.,

Mr.Wang,J.F., Dr.Ali, Dr.Chris, Dr.Sun D.M., Mr.Tang, S.J. Dr.Yang,D., Dr.Zhou,F.L.
Dr.Wang,Y., Mrs. Zhao,L.R., Mr.Zhou,Y., Mr. Zhu,F.Y., Mr.Ako,J.A.F., Mr. Bilal etc.
in the Textile and Paper division in the School of Materials for providing a pleasure
working atmosphere.
I would like to express my deepest gratitude to my dearest parents for their endless love,
patience, encouragement, and huge supports during the study.
20
CHAPTER 1
INTRODUCTION
Textile composites are made from two most important parts, the textile perform and
the matrix. Various forms of textiles have been used as performs and reinforcements
for textile composites. The advantages of textile composites have been widely
recognised and utilised for appropriate applications. In general, unidirectional and 2D
woven textiles are the main forms used in creating composite materials although new
composites have been created from using different types of 3D textiles structures for
improved performance. Such 3D composite reinforcements are advanced in that they
possess structural integrity and fibre continuity, and for these reasons they have
attracted much attention in research and in applications. 3D honeycomb structures,
which can be found in nature ranging from the spines of a porcupine to the stem of a
plant of reed, have many features that are important for many of the composite
applications, as have been described by Gibson and Ashby (1997) . Composites made
from this type of 3D reinforcements can be super-light, energy absorbent, voluminous
as well as being strong.
Lightweight materials with comparable properties such as high energy absorbency

compared to traditional materials have always been a favourable choice for many
applications in the aerospace and automotive sectors (Schmueser and Wicliffe, 1987;
Tao et al. 1993). Examples include woven ‘H-joint’ connectors for joining honeycomb
sandwich wing panels on the Beech starship in the aircraft (Wong, 1992) and floor
beams in trains and fast ferries, and so on (Mouritz et al., 1999). Lightweight materials
are of interests for energy absorption and protection against trauma impact, where
capabilities of the materials in impact energy absorption and in impact force
attenuation become important. Theoretical analysis on the 3D honeycomb composites
were carried out (Tan and Chen, 2005; Tan et al. 2007; Yu and Chen, 2006) and the
21
results suggested that the honeycomb composites have advantages over other types of
cellular materials in energy absorption and force attenuation under in-plane direction.
Excellent energy absorption is one of honeycomb composites’ important mechanical

characteristics and investigations have been carried out in various fields of engineering
dealing with different materials. For instance, energy absorbing sub-floor structures for
aircraft have been studied and large composite sub-floor structures applying cellular
structures as box element core have been developed for commuter and transport
aircraft (Herrmann et al., 2005). Energy absorption of paper cellular structures has
been studied by dynamic and static compression tests and the results showed that
increasing the loading speed and number of the layers could increase energy
absorption accordingly (Kobayashi et al. 1998). A number of investigations into
honeycomb structures have been carried out using aluminum alloy materials and for
instance the mechanisms governing in-plane crushing of hexagonal aluminum
honeycombs have been investigated with finite size honeycomb specimens crushed
quasi-statically between parallel rigid surfaces (Papka and Kyriakides, 1998).
Increasing interest in textile composites has been developed on the account of their
attractive properties of light-weight and high energy absorption capability for a variety
of applications (Mouritz et al., 1999). For many composite applications, such as those
in automobile, aerospace and aircraft sectors, reduction in component weight is highly
desirable. A wide range of fabrics available for composite reinforcement in the field of
textile structural composites has been reviewed by Wang and Zhao(2006), Bibo and
Hogg (1996). They summarised the different forms of textiles which were used for
composite reinforcement including different impact conditions and general material
variables such as fibre and resin type, and they pointed out that there was sufficient
information available to indicate that control of fibre organization by the use of textiles
might be an effective method of optimising composite properties for specific end use
properties. Recently, it was reported (Qiu et al., 2001) that 3D cellular matrix
composites were fabricated and their structural and mechanical properties were
investigated and compared to the 3D regular matrix composite. The former composite
has higher specific tensile strength, greater specific tensile modulus, lower specific
22
flexure strength, and higher specific impact energy absorption. Composites reinforced
by a new class of knitted structures have been designed (Cox and Davis, 2001) to
maximize the total energy absorbed during tensile failure. Levels of energy absorption
achieved reach approximately 40MJ/m3 or 25J/g. With optimisation, levels of 120-
200MJ/ m3 or 75-130J/g seem feasible. Textile composite panels reinforced with
integrally woven 3D fabric have been investigated at University of Manchester (Zic et
al., 1990) and their mechanical properties were found comparable to those of softwood,
aluminium alloy, and steel.
1.1 Description of the problem
Numerous investigations have been carried out to investigate the impact energy
absorption of the 3D honeycomb composites, but little has been reported in the
literature in relation to the influence of the structural parameters on impact energy
absorption. However, woven honeycomb composites used for low velocity impact
energy absorption were developed and produced at the University of Manchester over
the recent years (e.g. limb protection intended for the riot police). Previous research
has indicated that the mechanical properties of textile based honeycomb composites
can be engineered and controlled by selecting appropriate structural parameters. In-
depth analysis such as cell height and cell size in the woven honeycomb structures has
been carried out and the results show that the thick panel with 60°expansion/opening
angle leads to optimal performance in energy absorption (Tan and Chen, 2005). Wang
and Zhao (2006), and Bibo and Hogg (1996) have reviewed a wide range of fabrics
available for composite reinforcement in the field of textile structural composites. A
systematic work has been reported to engineer and characterizes 3D honeycomb
composites for impact applications at the University of Manchester (Wu, 2003). Chen
and Wang (2006) worked on the mathematical modeling of integrated cellular woven
preforms and on a CAD tool for designing cellular fabrics with various structural
parameters. Tan and Chen (2005) and Yu and Chen (2004) carried out FE analysis
adopting the quasi-static and dynamic approaches respectively reporting on the
influence of structural parameters on the mechanical behaviour of 3D honeycomb
composites. Besides these, little experimental and numerical studies on impact on
23
textile honeycomb composites were reported, which is an important gap in the study in
order to gain comprehensive understanding on 3D honeycomb textile composites.
There is a lack of systematic investigation to examine the effect of the parameters of
the honeycomb reinforcement and the composite performance in energy absorption
and force attenuation.
1.2 Research Aim and Objectives
Textile technology is capable of creating 3D honeycomb woven structures without the

need for weaving machine modifications. Based on the research that has been carried
out (Tan and Chen, 2005; Chen et al., 2007; Tan et al., 2007), this present research is
set to investigate the geometric parameters of honeycomb textile composites such as
the cell dimension, cell opening angle, ratio of cell walls on the composite
performance through experimental and numerical study. This study will start with the
design and manufacture of 3D honeycomb fabrics as reinforcements and will then
move to characterise honeycomb composites in terms of the behaviours and
performance. Evaluation the honeycomb composites for their energy absorption, force
attenuation under the influence of structural parameters of the composites will be
followed up by theoretical and experimental study of the 3D honeycomb composites
for their mechanical properties.
The aim of this research is to investigate how geometric and structural parameters of
honeycomb composites would affect mechanical performances under low velocity
impact. The outcome of the investigation could be useful to help design protection
products against trauma impact, such as shields for improved protection. This study
will focus on three aspects.
The first aspect of the study is to manufacture the honeycomb composites from the 3D
honeycomb fabrics with appropriate resin and hardener according to their designed
angle and cell size. The objectives for this part of the research are described as follows.
1) to develop a simple yet effective method for making honeycomb textile
24
composite. In the current study, a set of apparatus [see Chapter 4] has been developed
which will allow fabrics to be stretched into 3D and by adjusting the height of the
apparatus, the geometric parameters of the 3D textile honeycomb composite can be
altered;
2) to design and optimize the geometric parameters for the 3D honeycomb

structure and this includes cell height, cell size, cell opening angle and length ratio of
cell walls. Four groups of honeycomb composites, involving 14 varieties, will be
systematically created in order for the future experimental analysis of the honeycomb
composites to be carried out;
3) to establish a procedure for the sake of manufacturing 3D textile honeycomb

composites and this procedure will be able to guide future practical production
processes for similar composites. This procedure is expected to allow easy
reproduction in the future.
The second aspect of this study is to conduct low-velocity impact tests on the 3D
honeycomb textile composite samples in groups according to their cell size, cell
opening angle, cell bonded and free wall ratio and cell volume density, and the
acquired data will be analysed for the investigation of mechanical behaviour and
energy absorption properties of the honeycomb composites, which will lead to the
optimal geometries for honeycomb textile composites. The objectives for this part of
the research are:
1) to set up the testing equipment for the low velocity impact tests. The initial
impact velocity and impact energy will be targeted to divide the experiment into
different groups according to the impact energy level;
2) to conduct the low-velocity impact tests on the 3D honeycomb textile

composites in order to obtain the associated data for analyzing the mechanical
properties and energy absorption behaviour of the composites;
25
3) to analyse the experimental data to investigate how the structural and
geometric parameters of the 3D honeycomb textile composite affects its mechanical
and energy absorption behaviour that impact on the protection capability of the
composites; and
4) to carry out tensile tests on the single piece of shell from the honeycomb
composite to obtain material properties for the future theoretical and numerical study.
The third aspect of the research is to use the finite element method (FEM) to
investigate the impact performance of different geometrically optimised 3D
honeycomb textile composites and to validate the numerical results with practical
experimental results. This part of the work will lead to the establishment of a design
procedure for engineering 3D honeycomb composites. The objectives for the third part
of the research are as follows:
1) the first objective is to establish geometrical models for the 3D honeycomb

textile composites. Because the geometric parameters of single cell honeycomb textile
composite has been established by Tan and Chen (2005), the present study starts by
creating the geometric models based on the previous research findings and will expand
the single cell to multi-cell matrix of the composite. Such geometric models will then
be used for various FE analyses.
2) the second objective is to create models of three sets of honeycomb textile

composite structures with different cell size, opening angles, and cell bonded wall
length to free wall length in order to examine their mechanical and energy absorption
behaviour. The results are expected to provide comprehensive details in mechanical
performance between the modelled honeycomb composite and the physical
honeycomb composite specimen.
3) the third objective is to validate the theoretical result with the experimental result
to seek out the similarity from both models.
26
1.3 Thesis Layout
After this introductory chapter, Chapter 2 will present a review of the literature.
general introduction of honeycomb structure and its applications in the field, review of
mechanical behaviour including the energy absorption performance of the honeycomb
structure under impact, theoretical equations reported by previous researchers on
honeycomb textile composites and the design and manufacture of 3D honeycomb
woven fabric in the weft direction.
Chapter 3 presents works on designing and manufacturing 3D honeycomb fabrics;
Chapter 4 describes the creation of honeycomb composites and test sample preparation;
Chapter 5 reports on the results and analyses of the low velocity impact test under
dropping hammer system;
Chapter 6 reports on the experimental data analysis on textile honeycomb composites

impacted by larger mass and lower velocity impactor with comparing the result with
the results from Chapter 5;
Chapter 7 presents a finite element analysis (FEA) on honeycomb composites;
Chapter 8 ends the thesis with conclusions and future recommendations.
27
CHAPTER 2
LITERATURE REVIEW
Cellular solids such as sandwich panels have been used as advanced materials in
aerospace, automobile and marine industries for decades (Torre and Kenny, 2000; Meo
et al., 2003; Kim and Chung, 2007; Shin et al., 2008) due to their unique combination
of properties derived from their cellular structures. Scientists and engineers have paid
more and more attentions to cellular solids since new techniques for making ceramic
and metallic foams have widened the range of man-made materials and the diversity of
their applications (Gibson and Ashby, 1997). Textile reinforced honeycomb composite
(Sun, 2005, Tan and Chen, 2005) can be regarded as a kind of cellular solid due to its
hollow core structure and as an innovative product, much interests have been drawn on
it to find out its mechanical performance under various loading conditions(Tan and
Chen, 2005; Yu and Chen, 2006; Tan et al., 2007) .
This chapter presents a literature review on cellular solids including textile honeycomb
composites in the following aspects, which are (1) classification, applications,
mechanical and non-mechanical features of cellular solids (2) honeycomb structure
manufacturing techniques (3) the mechanical performances of cellular solids under
various impact conditions (4) the energy absorption analysis of cellular solids (5) the
basic concept of three-dimensional (3D) fabrics and structural parameters for textile
honeycomb composite (6) the application of 3D honeycomb fabrics on personal
protection equipment (PPE).
2.1 Classification of Cellular Solids
Cellular solids which are made from very diverse materials including wood, polymers,
metals, ceramics, glasses and composites and they are used in a broad classes of
application.
28
For example, cellular solids are notably used as core materials for sandwich structures
(Torre and Kenny, 200) and they are widely used as energy absorbers and shock
protectors in packaging industry too (Wang, 2009; Pflug and Veopoest, 1999; Pflug et
al., 2002). Besides those applications, cellular solids are also equipped as thermal
insulation for housing, for refrigeration and for high temperature equipment (Ashby and
Mehl, 1983) as well as floatation and buoyancy-aids in the ship (Gibson and Ashby,
1997)
Although there are a variety of cellular solids exist in nature and man-made, according
to their cell structures, could be classified as honeycomb structure and foam structure.
It is noted that in the current study, if the cellular solids exhibit a honeycomb structure,
it is called ‘honeycomb’ and if it shows up as a foam structure, it will be called ‘foam’.
2.1.1. Honeycomb structure
Gibson and Ashby (1997) defined honeycomb as a two-dimensional array of polygons

that pack to fill a plane like the hexagonal cells of a beehive as shown in Figure 2-1(a).
29
(a) Two-dimensional honeycomb (b) Three-dimensional foam with open cells
(c) Three-dimensional foam with closed cells
Figure 2-1 Examples of cellular solids (Gibson and Ashby, 1997)
In nature such as balsa wood, the honeycomb structure exists from the frequent
deviation from regularity caused by the way in which the individual cells nucleate and
grow, and the rearrangement that take place as they are developing. However, in many
man-made honeycombs, the honeycomb structures come in many different shapes and
sizes such as triangles, squares or hexagons and they are regular in pattern which are
shown in Figure 2-2 (Gibson and Ashby, 1997).
30
Figure 2-2 Schematic illustration of honeycombs structure will different cell shape
2.1.2 Foam structure
True honeycomb structures are relatively rare and the structures used mostly in
sandwich panels are as core material made by hexagonal aluminium (Abbadi et al.,
2009). More commonly, in cellular solids, the cells are polyhedral which are packed in
three-dimensions to fill the space and such cellular solids are called ‘foam’ and Figure
2-1 (b) and (c) gives two photographs of man-made foams. The foam structure can be
divided into open-celled foam structure which contains the cell edge only and closed-
celled foam structure when the faces of the cell is solid and each cell is sealed off from
its neighbors. Incidentally, some foam structure is partly open and partly closed.
2.2 Main Features of Cellular Solids
Both honeycomb and foam exhibit a unique combination of properties which are
derived from their cellular structures. In Figure 2-3, Gibson and Ashby (1997) stated
that comparing to the fully dense solids; cellular solids with hollows inside their
structure provide some outstanding features such as low density, unique stiffness and
strength according to their loading direction, open porosity structure and low thermal
31
conductivity. These four enormous extensions of properties create applications for the
cellular solids which cannot be easily filled by fully dense solids.
With a low density and open pore structure, a cellular solid can be used to design light,
stiff components such as sandwich panels and large portable structures (Torre and
Kenny, 2000). The cellular solids can also exhibit a low stiffness and strength behaviour
depend on their loading direction (Miltz et al., 2003) and this unique feature makes
them ideal for cushioning and energy absorption applications (Shaw and Sata, 1966).
Additionally, with low thermal conductivity, it allows cellular solids to be used as
disposable coffee cups to refrigerated trucks for the modern buildings.
Figure 2-3 The range of properties available to the engineer through foaming (Gibson
and Ashby, 1997)
32
2.2.1 Low density
Gibson and Ashby (1997) states that the relative density of the cellular material could be
1
calculated by
 2 , that is, the density of the cellular material, 1 divided by that of the
solid from which the cell walls are made,  2 . Polymeric foam used for cushioning,
packaging and insulation have relative densities which are usually between 0.05 and 0.2
and cork is about 0.14; most softwoods are between 0.15 and 0.40. Special ultra-low-
density foam can be made with a relative density as low as 0.001.
According to Figure 2-4: the material property charts from Pflug and Vangrimde (2003),
it can be seen that the density of ‘foam’ and ‘honeycomb’ lie near the bottom left of the
chart which is below 0.4kg/dm3 (that is 0.4g/cm3)while metals positioned near the top
right (over 100 times denser than cellular solids), which indicates metals own a high
density and high modulus; fine ceramics such as aluminium or concrete are less dense
than metal but stiffer still. Table 2-1 further exampled the density of alumina ceramic
honeycomb and rigid polyurethane foam, comparing with solid ceramics and solid steel.
The results indicate that the cellular solids can reduce the weight of the material
dramatically.
Takenata et al., (1991) specified that for honeycomb structured woven composite, the
density of the composite material is 0.03 to 0.2g/cm3. If the density is lower than
0.03g/cm3, sufficient high compression strength is difficult to attain. On the other hand,
if the density is higher than 0.2g/cm3, the mechanical performance of the composite
material can be significantly increased but the weight-reducing effect is declined.
33
Figure 2-4 A chart showing material Young’s modulus and density where each material
class occupies a characteristic field on the chart (Pflug and Vangrimde, 2003)
Table 2-1 Density of cellular solids and solid material (Gibson and Ashby, 1997)
Material Density,  (g/cm3)

Honeycomb (alumina ceramic) 1.4
Foam (rigid polyurethane) 0.032
Solid ceramics (silicon carbide, SiC) 3.2
Solid Steels 7.6-8.1
34
Because of the low density of cellular solids, one of the earliest market for them is in
marine buoyancy where the light closed-cell plastic foam are extensively used as
supports for floating structures and as floatation in boats. Because of their closed cells,
they can retain their buoyancy even when extensively damaged and they are unaffected
by extended immersion in water with their resistances to rust or corrode (Gibson and
Ashby, 1997). Another major use of man-made cellular solids is in packaging because
low density means the package is light which can reduce handling and shipping costs.
Currently, the foam being frequently used in packaging is polystyrene, polyurethane and
polyethylene.
2.2.2 Stiffness and strength of cellular solids
It is important to understand the stiffness and strength performances of honeycombs

when they are used in load-bearing structure. Gibson and Ashby (1997) specified that
generally, if a honeycomb is compressed in-plane that is the plane along X1 and X2
direction in Figure 2-5, the cell wall at first bend, giving linear elastic deformation.
Beyond a critical strain, the cells collapse by elastic buckling, plastic yielding, creep or
brittle fracture, depending on the nature of the cell wall material. Cell collapse ends
once the opposing cell walls begin to touch each other and as the cells closed up, the
stiffness of the structure increases rapidly. When the loading is along out-of-plane
direction, which is along X3 direction in Figure 2-5, the stiffness and strength are much
higher because they require extra axial extension or compression of the cell walls.
35
Figure 2-5 Honeycomb structure with hexagonal cells (Gibson and Ashby, 1997)
Figure 2-6 exhibits the three regimes of behaviour of all cellular solids when they are
undergoing in-plane loading and it can been seen that at strains less than about 5%, the
material is linear-elastic and with the increase of the loading, depending on the
properties of cell walls, the cells began to collapse by elastic buckling, plastic yielding
or brittle crushing (Barma et al., 1978; Ashby and Mehl., 1983; Kurauchi et al., 1984;
Maiti et al., 1984). Collapse progresses at a roughly constant load until the opposing
walls in the cells meet and touch, when densification causes the stresses to increase
steeply (Shaw and Sata, 1966; Papka and Kyriakids,1998; Ruan et al, 2002).
Regarding the out-of-plane loading, Gibson and Ashby (1997) also said that when the
honeycombs are compressed in out-of-plane direction, that is X3 direction in Figure 2-5,
its linear-elastic regime is truncated by buckling (elastic for an elastomer, plastic for a
metal or rigid polymer) and final failure is by tearing or crushing respectively.
36
(a) Elastomeric rubber (b) elastic-plastic metal (c) elastic-brittle ceramic
Figure 2-6 Typical compressive stress-strain curves for cellular solids under in-plane
compression (Gibson and Ashby, 1997)
After the strain-stress curve has been generated by the researchers mentioned above,
Ruan et al. (2002) discovered that plateau stress of aluminium foam has a power law
relationship with the foam’s relative density (Equation 2-1). Not only Ruan et al., early
in 1982, Hilyard (1982) already mentioned in his book that the strength of the cellular
solids including compressive and shear, can be described as a function of the density of
formed material, and Vural and Ravichandran (2003) also found out that the stress of
balsa wood increases with its material density. Moreover, other researchers (Kenny,
1996; Baumeister et al., 1997; Banhart and Baumeister, 1998) drew the same
conclusions in their papers.
A =A0  n [2-1]
Where A is the property of foam,  is the density of foam, A0 is a factor which reflects
the properties of the solid cell wall material and n is an exponent.
It seems the material density is one parameter which affects the stiffness and strength of
cellular solids; however, there are other parameters which have the same influences too.
Early in 1978, Barma et al. (1978) mentioned in their paper that they found out that the
modulus of the foam as well as the yield stress is proportional to t/l, where t is the
thickness of foam struts and l is the length of foam struts, and this means the foam
37
material property is related to their cell thickness and cell size. Ashby and Mehl (1983)
further specified that the mechanical properties (elastic, plastic, creep and fracture) of
cellular solids or foam are affected by their cell geometry.
Decades later, Gibson and Ashby (1997) demonstrate in their book that when a
honeycomb is loaded in-plane, the cell walls bend and it deforms in a linear-elastic way
firstly (Abd EI-Sayed et al., 1979; Gibson et al., 1982). The response can be described
by five moduli numerically: two Young’s moduli E1* and E 2* which are modulus of
compression or tension applied in-plane individually; a shear modulus G12* and two
Poisson’s ratios, 12* and  21

*
. Gibson and Ashby (1997) further identified the
relationship between honeycombs’ Young’s modulus under linear-elastic deformation
with their cell geometric parameters such as cell wall thickness, cell wall length and cell
opening angle in Figure 2-7 and Equation 2-2 to 2-6:
Figure 2-7 Geometric parameters of the honeycomb cell from Gibson and Ashby (1997)
38
in Figure 2-7, t is the thickness of the cell wall; l and h are the length of the cell walls.
Therefore, in the following Equation 2-2 to 2-6, t/l is the ratio of cell wall thickness to
its length and h/l is the length ratio of two walls; θ is the opening angle which has been
defined in Gibson and Ashby (1997)’s work to describe the angle between cell walls.
cos
3
E1*  t 
  [2-2]
Es  l  ( h  sin ) sin 2 
l
E 2*  t  (h / l  sin  )
3
  [2-3]
Es  l  cos3 
2 cos2 
12*    [2-4]
 1 (h / l  sin  ) sin
 1 (h / l  sin  ) sin 
 21
*
  [2-5]
2 cos2 
(h / l  sin )
3
G12*  t 
  [2-6]
E s  l  (h / l ) 2 (1  2h / l ) cos
where E1* and E 2* are the Young’s modulus of the honeycomb composite along X1 and
X2 directions respectively and Es is the Young’s modulus of the honeycomb composite
wall material; 12* and  21

*
are the Poisson’s ratios of the material in the X1 and X2
directions; G12* means the shear modulus of the honeycomb composite.
Gibson and Ashby (1997) also summarized the mechanisms for compressive
deformation of honeycomb structure schematically in Figure 2-8. This figure is a
schematic diagram for a honeycomb loaded in compression in the X1-X2 plane, showing
the linear-elastic, collapse and densification regimes and the way the stress-strain curve
changes with t/l. From Figure 2-8, it shows how the stress-strain curves changes with
39
the increasing of the t/l. It seems that the modulus of the structure goes up with an
increasing in the t/l and the cell walls touch sooner which reduce the strain at which
densification begins. This means, assuming that the honeycomb cells are under the same
thickness (t), if t is kept constant but making the cell wall length (l) longer, in other
words, to make the cell size larger, thus t/l would get smaller which leads to lower stress
but more strain.
Figure 2-8 A schematic diagram shows the way the stress-strain curve changes with t/l
Not only the stiffness and strength are influenced by honeycomb geometric
configuration, the energy absorption characteristics in impact crush of cellular solids are
strongly affected by their geometric configuration too. This can be traced back to the
previous researchers (Wierzbicki, 1983; Wu and Wu, 1997; Yamashita and Gotoh, 2005)
who have investigated the energy absorption performance or the compressive strength
in the out-of-plane crush situation for the aluminium honeycomb panel with different
40
cell geometric parameters such as cell wall thickness and cell size. They concluded that
the use of a smaller cell size and core height with a stronger cell material will enhance
energy absorbing capability of the honeycomb structure, at the same time that
honeycomb structure with smaller size yields a higher compressive strength respectively.
From the work mentioned above, it can be seen that the mechanical properties of the
cellular solids such as stiffness, compression/tensile strength, shear strength, lateral
expansion and energy absorption are determined by their material density and cell
geometric parameters such as: cell size; the ratio of cell wall thickness to length (t/l );
and ratio of bonded to free wall length (h/l).
2.2.3 Open pore structure and its application
Many natural structural materials with open pore structure are cellular solids such as
wood and cancellous bone that can support large static and cyclic loads for a long
period of time (DeBonis and Bodig, 1975; Odgaard and Linde, 1991). Even today,
wood is still the world’s most widely used structural material and the understanding of
the way in which wood’s properties depends on the wood density and on the direction
of loading and this can lead to improved design with wood. Interest in the mechanics of
cancellous bone stems from the need to understand bone diseases and attempts to devise
materials to replace damaged bone. Both wood and cancellous bones are good examples
of taking advantage of the open porosity structure of cellular solids to benefit human
beings.
Not only natural cellular solids are used widely, there are more man-made foam and
honeycombs which are used to perform a truly structural function. The most obvious
example is their use in sandwich panels. The innovative design of the de Havilland
Mosquito (a World War II bomber) used sandwich panels make from thin plywood
skins bonded to balsa wood cores (Hoff, 1951) and in later designs the balsa wood was
replaced by cellulose acetate foam. Man-made honeycomb sandwich panels are
increasingly being used to replace traditional materials in highly loaded applications
(Kim and Chung, 2007; Shin et al., 2008). Figure 2-9 illustrated a typical honeycomb
41
sandwich structure which consist of a thick layer (core) intercalated between thin-stiff
layers (skins) (Abbadi et al., 2009).
Figure 2-9 Schematic detailed description of the honeycomb sandwich structure

(Abbadi et al., 2009)
There are a large variety of sandwich panels that are being applied in structural
engineering such as aerospace, transportation, marine and packaging due to their open
porosity structure which separates the two thin layers and allows for an outstanding
weight specific bending stiffness and reduces the weight of the composite dramatically
(Pflug et al., 2002; Pflug and Vangrimde, 2003; Wang, 2009). Depending on the
loading rate, the mechanical behaviour of sandwich structure could be various. In fact,
they can have a ductile behaviour in case of static loading, but may behave in a brittle
manner and fail catastrophically when subjected to impact loads (Gibson and Ashby,
1997). Figure 2-10 listed a series of sandwich panels with different core types and
among them honeycomb cores with hexagonal cell are characterized by a considerable
rigidity in shear, high crushing stress, almost constant crushing force, long stroke, low
weight and relative insensitivity to the overall loss of stability (Wierzbick, 1983).
In the last decades, sandwich panels have increasingly been adopted in numerous
aircraft structures such as control surfaces, fairings or in the cabin interior. One of the
42
examples is its application for large structures at AIRBUS started in 1983 (Herrmann et
al., 2005) when the A310 was the first aircraft in the AIRBUS fleet to be equipped with
a composite honeycomb sandwich rudder. Ever since, the experience with large
composite structures was extended and there is a broad range of composite sandwich
structures’ application in Airbus aircraft such as belly fairings, ling and trailing edge,
engine cowling etc. and Figure 2-11 shows the details of their application in AIRBUS A
380.
Figure 2-10 Different Sandwich core types (Herrmann et al., 2005)
Figure 2-11 Examples for sandwich application A380 (Herrmann et al., 2005)
43
Due to their unique open pore structure, honeycomb sandwich panels have been
excessively used in packaging industry besides furniture and building industry for the
sake of their favourable cushioning properties (Shaw and Sata, 1966; Wang, 2009) and
especially paper honeycomb products are usually used as cushioning material in logistic
processing to withstand vibration and shock by means of absorb the energy so as to
protect products from damage.
Regarding the sandwich panels, there are a lot of researches have been conducted on the
area of validating new calculation methods and tools, better understanding of effects of
defects, improved and more economic Non Destructive Testing (NDT) capabilities,
advanced core materials, novel manufacturing methods and integration of structural and
non-structural functions for the sandwich panesl and their core structure (Kleineber et
al., 2002; Ley et al., 1999; Andersson and Van den).
Although cellular solids are favourable used as core material for sandwich panels, there
are some drawbacks in them and the significant one is that cells may suffer from
accumulating and condensing water which are trapped to increase the weight and
decrease the mechanical properties for the material (Vavilov et al., 2003; Kleineberg et
al., 2002).
2.2.4 Thermal insulation property
Foam is remarkable for its good thermal insulation and there is a considerable literatures
regarding on this subject (Yee and Duardo, 1983; Glicksman et al., 1987; Micco and
Aldao, 2006).
More foam is used for thermal insulation than for any other purposes. The closed-cell
foam has the lowest thermal conductivity of any conventional non-vacuum insulation
and it is used, for example, in frozen food industry to fill the double skins of refrigerated
truck and railway cars. Gibson and Ashby (1997) states that there are several factors
combine to limit heat flow in foams: the low volume fraction of the solid phase; the
small cell size which virtually suppresses convection and reduces radiation through
44
repeat absorption and reflection at the cell walls; and the poor conductivity of the
enclosed gas.
2.3 Manufacturing of honeycomb structure
Gibson and Ashby (1997) summarized that the honeycomb structure can be made in at
least four ways. The most common way is to press sheet materials into a half-hexagonal
profile and glue the corrugated sheets together. More commonly, glue is laid in parallel
strips on flat sheets, and the sheets are stacked so that the glue bonds them together
along the strips. The stack of sheets is pulled apart to give a honeycomb. Paper-resin
honeycombs are normally made like this that the paper is glued and expanded, and then
dipped into the resin to protect and stiffen it. Honeycombs can also be cast into a mould
and increasingly, honeycombs can be made by extrusion; the ceramic honeycombs used
to support exhaust catalyst in automobiles are made in this way.
Besides the methods to manufacture the honeycomb structure stated by Gibson and
Ashby (1997), in Bitzer’s (1997) book, more detailed manufacturing process to produce
honeycomb core structure was described. According to him, there are two basic
techniques used to convert the sheet material into honeycomb: the expansion process
and the corrugation process. Expansion process is a more efficient technique to produce
the majority of the adhesive bonded cores and the whole method is illustrated in Figure
2-12. For metallic cores, a corrosive resistant coating is applied to the foil sheets, and
adhesive lines are printed. The sheets are cut and stacked, and the adhesive is cured
under pressure at elevated temperature. Then the slices are cut into the required
thickness and expanded. When metallic cores are expanded, the sheets yield plastically
at the node-free wall joints and thereby retain their expanded geometric shape.
45
Figure 2-12 Expansion manufacturing process (Bitzer, 1997)
The procedure for non-metallic honeycomb is slightly different. Here the honeycomb
does not retain its shape after expansion and must be held in a rack. The block web
material contains a small amount of resin that is heat-set in an oven. Most paper cores
will retain their expanded shape. Then honeycomb block, sometimes as large as 4ft by
8ft by 3ft thick, is dipped in liquid resin (usually phenolic or polyimide) and oven cured.
The dipping-curing cycle is repeated until the block is at the desired density.
Bitzer (1997) also described the corrugation method which is illustrated in Figure 2-13,
and this method is the original technique used to fabricate honeycomb core. Although it
is labour intensive, this method is still used for making high density metallic and some
non-metallic cores.
Figure 2-13 Corrugation manufacturing process (Bitzer, 1997)
46
In the corrugation process the sheets are first corrugated, then adhesive is applied to the
nodes and sheets are stacked and cured in an oven. Some non-metallic corrugated
blocks must be brought up to final density by resin dipping to achieve the optimum
resin-to-reinforcement ratios.
Takenaka et al. (1991) specified in his patent that in general, conventional honeycomb
cores are obtained by coating an adhesive in stripes spaced equidistantly on a thin sheet
such as paper, an aluminum foil or a film, laminating and bonding such adhesive-coated
thin sheets, and expanding the bonded structure to form honeycomb-like structure
having a multiplicity of cells. Regarding to use the woven fabric as honeycomb core
structure, he said that normally, a plane woven fabric composed of glass fibers or the
like is used as the sheet material for forming a honeycomb core according to the above
mentioned process and a thermosetting resin such as an epoxy resin is impregnated over
the honeycomb core to form the composite material. However, this kind of honeycomb
does not have a sufficient tensile strength, peel strength and shear strength of the
bonded surface and it is easy to delaminate at the bonding point. Therefore, he invented
a type of woven fabric which is having a multi-layer structure and comprises a plurality
of woven fabric layers that are integrated through combined portions. It is formed by
interlacing warps or wefts of one of adjacent woven fabric layers to construct the textile
honeycomb structure. One of the advantages of this type of integrated multi-layer fabric
is that it can sufficiently solve the delamination problem between the layers for the
honeycomb structure.
2.4 Mechanical performances of cellular solids
Cellular solids are widely used in energy absorption applications against various loading
conditions, and it is necessary that the mechanic performances of the cellular solids are
understood. The following sections have a good review on cellular solids’ mechanical
behaviours when they are under various loading conditions such as statistic and
dynamic.
47
2.4.1 Previous studies on cellular solids’ mechanic performances
Pioneering works on the mechanic properties of honeycomb and foam material,

including compression, are those Gent and Thomas (1959), Shaw and Sata (1966) and
Barma et al. (1978), whereas a book edited by Hilyard (1982) contains a series of
articles which summarize the state of the art, at that time, on polymeric foam.
McFarland et al. (1963), who have studied the crushing behaviour of honeycomb
structure, then developed a semi-empirical model to predict the crushing stress of
hexagonal cell structure subjected to axial loading. This model was later improved to
incorporate both bending and extensional deformation of such cellular structure by
Wierzbicki (1983).
Gibson and Ashby are the two most famous researchers in the area of cellular solids.
They and their co-workers have focused their research attentions on cellular solids since
1980s (Gibson and Ashby, 1982; Gibson et al., 1982; Ashby and Mehl, 1983; Gibson et
al., 1989). In 1997, they published their credited book named ‘Cellular solids: structure
& properties’ and this book has a comprehensive coverage of this subject and
considering both man-made and natural cellular solids. They introduced the deformation
mechanism of the cellular solids including honeycomb and foam along in-plane and out-
of-plane direction and an in-depth description of their material mechanics were carried
out on the aspect of linear-elastic deformation, elastic buckling, plastic collapse, brittle
failure, viscoelastic deformation, creep and densification. They also devoted a chapter in
their book to introduce the selection of materials for low speed impact applications. The
area they investigated which are relating to the current study is that they verified that the
stress and stiffness including lateral deformation with shear performance of the
honeycomb cell structure are affected by their geometric properties. They also
mentioned in their book that the energy absorption of the foam is proportional to the
density of the foam and it is shown in Figure 2-14. From the figure, it can be seen that,
assuming the foam is compressed until they accumulate the same compressive strain (ε),
with the increase of the foam density (ρ), the absorbed energy (W) and compressive
stress (σ) increases too.
48
Figure 2-14 The peak stresses generated in foam of three densities in absorbing the
same energy, W (Gibson and Ashby, 1997)
Other researchers who studies the mechanical performance of natural cellular solids
such as balsa wood in longitudinal and/or transverse direction are particularly associated
with names of Knoell (1966), Soden and McLeish (1976), Easterling et al. (1982),
Vural and Ravichandran (2003a; 2003b) and Reid and Peng (1997).
Knoell (1966) investigated the effects of environmental and physical variables

(temperature, moisture content and ambient pressure) on the mechanical response of
balsa wood. Soden and McLeish (1976) carried out an extensive investigation, which
mainly concentrated on the variation of tensile strength with fiber alignment. They also
reported compressive strength data. Easterling et al. (1982) paid particular attention to
the micromechanics of deformation in their experiments, during which they performed
in scanning electron microscopy (SEM) observations and defined the end-cap collapse
of grains as the dominant compressive failure mechanism in longitudinal direction.
Vural and Ravichandran (2003a, 2003b) documented the compressive strength; plateau
stress and densification strain of balsa wood in its entire density range, identified the
variations in failure mechanisms with density and described simple analytical models to
represent the observed experimental strength data. They also applied the loading force
49
dynamically onto the balsa wood and compared the dynamic data with quasi-static
experiment results and concluded that the initial failure stress is very sensitive to the
rate of loading; plateau stress remains unaffected by the strain rate.
Reid and Peng (1997) investigated the dynamic crushing behavior of several wood
species, including Balsa wood, Yellow Pine, Redwood, American Oak and Ekki,
through the impact of specimens along/across grain direction. Their tests covered a wide
range of impact velocities up to approximately 300m/s, they tested the specimens at a
certain density for each species and stress-strain curves were obtained. They found out
that under dynamic loading especially when the loading is along in-plane direction,
there is a significant enhancement of the initial crushing strength of the specimens if the
velocity is increased and the corresponding time-load curve is drawn in Figure 2-15.
That is to say, for the same cellular solids and here is Redwood, if increases their impact
velocity and here from 80m/s to 150m/s, their crushing strength will sharp up
dramatically, which means the in-plane crushing strength of wood is highly sensitive to
the its impact velocity.
Figure 2-15 Typical time-load pulses from uniaxial crushing tests for Redwood
specimens (Reid and Peng, 1997)
50
Besides natural cellular solids, man-made cellular honeycombs and foam were actively
studied by Reid, Stronge, Shim and Wierzbick, etc. (Reid and Bell, 1982; Reid and
Reddy, 1983; Reid et al., 1983; Reid and Bell, 1984; Su et al., 1994; Klintworth and
Stronge, 1988; Stronge and Shim, 1988; Shim et al., 1992; Abramowicz and
Wierzbicki, 1989). In the following section, Papks and Kyriakides (1998), Yamashita
and Gotoh (2005) will be taken as representatives and their work will be described in
details.
Papka and Kyriakides (1998) studied the response of hexagonal aluminium honeycomb
under in-plane loading. They focus their work on the post-yield behaviour of aluminium
honeycomb structure. Experiment and numerical results agreed well with each other in
their work. According to them, under uniaxial, quasi-static experiment compression, the
force-displacement response is initially stiff and elastic but this terminated by certain
load instability. Localized crushing involving narrow zones of cells in initiated and
subsequently crushing spreads through the material while the load remains relatively
constant. When the whole specimen is crushed and the response stiffens again. Both of
their experimental and finite element analysis confirmed the honeycomb’s limit stress is
depended on its relative density.
Yamashita and Gotoh (2005) studied the quasi-static compression response of

aluminium honeycomb in the thickness direction. They investigated the effect of the cell
shape and the foil thickness on the crush behaviour. The numerical investigation
showed that the cyclic buckling mode takes place in every case and that the crush
strength is higher for the smaller cell angle specimens. This information reveals that the
cell angle affects the mechanic performance of honeycomb structure significantly.
Dynamic impact tests were also performed by Wu and Wu (1997) who have used a gas
gun to study the out-of-plane properties of aluminium honeycombs. In their study,
honeycombs with different cell size, material strength and core thickness were
compared when they were under quasi-static and dynamic impact loading conditions.
The cell size of the specimen they have chosen is 3.2mm and 4.7mm, and finally, they
concluded that in order to make the best use of a honeycomb structure as an energy
51
absorber, honeycombs which is small in cell size and core height, made by a high-
strength material is recommended. Similar to his conclusions, recently, Tang et al.
(2008) produced a 3D glass/polyester resin cellular woven composite with vacuum
aided resin transfer moulding (VARTM) technique and the fibre volume fraction in his
composite is 40% which can significantly thicken the cell wall and increases the
strength of the composite material.
Although Wu and Wu (1997) studied the crushing behaviour of aluminium honeycombs,

in their study, after-test specimens were observed while the whole crushing processes of
the honeycombs were not investigated. The development of finite element (FE)
technology has made it possible for the researchers to see clearly the whole crushing
process such as the stress concentration and distribution etc., to analyse and predict the
composite structure behaviour against various loading conditions more accurately.
In 1998, Abrate (1998) summarized recent modelling techniques for localized impact of
sandwich panels with laminated face sheets, including contact laws, composite
sandwich beam and plate theories, and dynamic spring-mass models. Such analytical
solutions are needed for designing sandwich panels against impact damage. They
provide valuable information for locating damage and establishing criteria for
acceptance or repair of structural components. Since then, a lot of other researchers such
as Ruan et al. (2003), Zheng et al. (2005), Yu and Chen (2006) have conducted impact
researches on various cellular composites by means of FE analysis tools.
Ruan et al. (2003) studied the influences of cell wall thickness of honeycombs and the
impact velocity on the mode of localised deformation and its plateau stress by means of
FE analysis tool: ABAQUS (Hibbitt et al., 1998) along in-plane direction. The
hexagonal cell is modelled with an opening angle equals to 60°and the cell wall length
(l) is 4.7mm with cell wall thickness (t) varies from 0.08mm to 0.5mm; the impact
velocity (v) is increased from 3.5m/s to 280m/s. They reported three different ‘patterns’
of deformation appeared during the impact loading: ‘X’, ‘V’ and ‘I’, depending on the
impact velocity of the honeycombs and the schematic illustrations can be seen in Figure
2-16. When the impact velocity is 3.5m/s, the deformation pattern shows up as ‘X’
52
shape; at 14m/s, it is ‘V’ shape and at 70m/s, the shape is ‘I’. This means the
deformation of honeycombs with hexagonal cell is highly sensitivity to its impact
velocity. Furthermore, they also figured out that both deformation patterns and impact
plateau stresses of honeycombs are related to the cell wall thickness and impact velocity
and the power low is shown in Equation 2-7. Supposing the thickness of the wall (t) are
the same, when the impact velocities (v) is sufficiently high, the impact plateau stresses
(σ/σys) show a good correlation to its impact velocities by a square law. Ruan et al.’s
study again verified that impact velocity and cell wall length are two factors which are
worthwhile to be investigated regarding the mechanical performances of honeycombs.
 t   t  
3 2
t 
 0.8   62   41   0.01  10 6 2 [2-7]
 ys  l    l  l  
In Equation 2-7, σ/σys is the ratio of impact plateau stress (σ) to its yield stress (σys); t/l is
the ratio of cell wall thickness (t) to its cell wall length (l) and v is the impact velocity of
the honeycombs.
(a) Original shape (b) v=3.5m/s (c) v=14m/s (d) v=70m/s
Figure 2-16 Crushing of a honeycomb in the X1 direction, where v is the initial crushing
velocity (Ruan et al., 2005)
Zheng et al. (2005) got similar results regarding the deformation patterns by conducting
2D FE analysis comparing to the results from Ruan et al. (2003)’s work. Besides regular
hexagonal cell shaped honeycombs, Zheng et al. (2005) also studied the honeycombs
with irregular cell shape. Finally, they found out that the deformation in an irregular
53
honeycomb is more complicated than that in a regular honeycomb due to its cell
irregularity, nevertheless, the energy absorption of the honeycombs can be improved by
increasing their cell irregularity. Zheng et al. further investigated the inertia effect on
the deformation of the honeycombs and this will be described more in the next Section
2.4.2.
The current study is a continuous progress from part of Yu and Chen’s work (Yu and
Chen, 2006), when they did some numerical analysis by FE analysis tools (Mac Mentat,
2005) on investigating the shape and material of the impactor. In their study, the
material of the modelled textile honeycombs are glass/epoxy sheet and the shape of the
cell is regular hexagonal. There are a range of impact objects of spherical, rectangular
and cylinder which are modelled as impactor with item mass between 0.4kg to 0.6kg to
represent hand thrown missiles and the impact velocity is 5m/s, 10m/s and 15m/s
respectively. The materials for the impact objects are modelled as wood, iron, glass and
concrete to represent a wooden ball, a short metal bar, a thick glass bottle and a half
house brick. Eventually, they concluded that textile reinforced cellular structure can
provide much better protection, such as energy absorption and force attenuation, than
the current used foam-shell structure limb protector for the policeman. Although Yu and
Chen’s work covers a wide range of the impact objects, they didn’t consider the
structure of the honeycombs themself may bring effect on the protection capability too
and this is one of the reasons to initiate the current research work.
2.4.2 Dynamic impact with different velocities
The impact mechanic response to the impactor could be divided into low-velocity and
high-velocity impact and they differ in their nature which is determined by their impact
velocities and the mass of impactor. However, no clearly defined boundary exists
between these two groups. For that purpose the structural response of the target is taken
into account (Richardson and Wisheart, 1996). At low-velocity impact, the composite
structure has sufficient time to response to the dynamic loading, and the contact
duration between the impactor and the target is relatively long. Consequently, more
energy can be absorbed elastically. This is specifically the case for personal protective
54
equipment (PPE) where the mass of the impact threats is between 0.05kg to 1.0kg and
the velocity are frequently no more than 30m/s (Dionne et al., 2003).
On the other hand, high-velocity impact responses can be characterised by the shock-
wave propagation through the material, where the structure does not have time to
respond (Flanagan et al., 1999). A typical example is a ballistic impact on a military
helmet, where the bullet’s mass may be only a few grams and the velocity is around
360m/s or more (Aare and Kleiven, 2007). When cellular solids are impacted under
high-velocity situation, the loading force, at a given instant, caused the concentration of
deformation in one particular area and it changes the cell shape which leading to a local
strain-rate much larger than the nominal strain-rate and this results in a much localised
damage in the composite (Gibson and Ashby, 1997).
Abrate (1998) also specified that at higher impact velocities, a critical condition will be
reached when local contact stress exceeds local strength, leading to the structure failure,
core interface delaminating and core compression strength failure too.
Whatever the impact velocity is low or high, Gibson and Ashby (1997) specified in their
book that micro-inertia which can be determined by the thickness of the cell wall plays
a significant role in controlling distribution of the crushing in lightweight open-cell
foam and honeycombs. This is based on the conclusions from Klinworth and Strong
(1998) that micro-inertia is associated with rotation and lateral motion of cell walls
when they buckle and tend to supress more compliant buckling modes in order to
increase the crushing stress and diffuse the crushing wavefront respectively.
Zheng et al. (2005) further reported that at a low impact velocity, the deformation of the
honeycombs can be regarded as under quasi-static loading and the inertia effect can be
neglected while under high impact velocity, inertia effects dominated the deformation of
the honeycombs which will cause the transverse band inside the honeycombs.
Schubel et al. (2005) studied experimentally the low-velocity impact behaviour of

sandwich panels consisting of woven carbon/epoxy face-sheets and a PVC foam core.
55
The impact velocity is set up in the range of 1.6m/s to 5.0m/s and the impact energy is
around 7.8J to108J. They compared the test results with an equivalent static loading
result and they found out that low velocity impact was generally quasi-static in nature
except for localized damage. This kind of impact velocity level is more similar to the
current study due to that this research work is developed from an existing project (Yu
and Chen, 2006), which involves replacing the shield foam core with textile honeycomb
composites for UK policemen in order to reduce the weight of the shield and improve
the protection. Considering that in most trauma and slash cases, the projectile velocity
can be categorised into low-velocity impact, therefore, velocities under 30m/s will be
defined as impact velocity for the experiment and FE analysis in this study.
2.4.3 Energy absorption of cellular solids
Honeycomb is one of the most commonly used materials as core material for energy
absorption besides metal tube, conical shell, tube array, foams and woods, etc. (Shih and
Jang, 1989; Kim and Jun, 1992; Chun and Lam, 1997;Gibson and Ashby, 1997; Reddy
and Reid, 1980; Reid and Bell, 1982; Gupta et al., 1997; Tang et al., 2008). When
design the structure for the core material for the purpose of cushioning performance, the
major concern is that this kind of material should be capable of accommodating large
permanent deformation without structure failure and it should show reliable and
controlled ‘load-deformation’ behaviour under dynamic loading conditions (Hernalsteen
and Leblois, 1976).
Johnson et al. (1977) again summarised the requirements for the energy absorption of
composites and they concluded that firstly the composites should utilize the plastic
deformation rather than elastic deformation, as their major energy-dissipation
mechanism. They also said that while providing sufficient energy absorption capacity,
the peak force (thus the peak deceleration) must be kept below the threshold that would
cause damage or injury and the structure deformation stoke should be long, stable,
repeatable and reliable under impact condition. Finally, Johnson et al. (1977) pointed
out that the composites should be light themselves, possessing high specific energy
56
capacity (i.e. energy absorption capacity per unit weight) and thus textile honeycomb
composite with volume density less than 0.2g/cm3 should meet the requirements.
Regarding the energy absorption mechanisms, Gibson and Ashby (1997) said natural
cellular solids such as wood, bones, and leaves have cell walls which are themselves
composites. When these materials are deformed, the fibres were pulled out and unravel
in complicated ways which dissipate a great deal of energy. However, for man-made
cellular solids, there is a number of mechanisms are at work in absorbing energy
(Schwaber, 1973) which related to the elastic, plastic or brittle deformation of the cell
walls. The energy could be converted into localized plastic deformation, heat and a
small part could be remained as kinetic energy as a result bouncing occurs (James and
Stephen, 2001). In general, the absorbed energy (W), up to a strain (ε), can be expressed
as follows:

W =   ( )d [2-8]
0
where σ(ε,) is the stress up to a strain of the deformed structure (Gibson and Ashby,
1997).
Although there are a lot of researchers (Kobayashi et al., 1998; Yasui, 2000; Zhao and
Gary, 1998; Wang and Wang, 2007; Wang, 2009) who have studies different types of
honeycomb core as energy absorber made from various materials: polypropylene (PP),
polyester (PET), paper and aluminium, there is few papers have been published on the
area of woven textile honeycomb cores.
One of the pioneers who studied woven textile honeycomb composite is Chen and Tan
(Tan and Chen, 2005; Tan et al., 2007) who did a lot of work on optimising the
geometric parameters for the textile honeycomb composites theoretically. They
compared various essential structural parameters which affect the energy absorption and
deformation behaviour of the composites and they found that cell opening angle, cell
wall length and cell wall thickness significantly affects the energy absorption and
57
deformation behaviour of the composites. The strain energy density concentrations
appear very seriously around the cell corners during quasi-static impact. Further
investigations about the honeycomb composites assembled with face sheet plies were
conducted theoretically and the results show that deformation and distributed strain
energy density of both outer and inner surfaces of the applied structure are significantly
affected by ply assembly, outer ply material, outer ply thickness, and loading area. Wu
(2003) conducted limited mechanical experiments to validate above findings and found
that the change of the structural parameters can result in significant change of
honeycomb composites’ mechanical performancse. Therefore, systematic experimental
and theoretical investigations are required and this initialized this research project.
2.5 Textile Honeycomb Composites
There are different types of textile composites including 3D woven, braided or knitted
textile reinforced composites. These types of textile reinforced composite materials
have drawn a wide research interests due to their capability of efficiently absorbing
kinetic energy to weight ratio. And the cost is low during manufacturing process and
their damage tolerance is excellent too.
3D woven composites were first developed nearly 40 years ago in an attempt to replace
expensive high temperature metal alloys in aircraft brakes by Mullen and Roy (1972). In
their work, a specialised loom was developed to allow the weaving of hollow cylindrical
preforms in which carbon fibres were aligned in the radial, circumferential and axial
directions. The produced composites display some specific strength and stiffness
properties as well as excellent resistance to thermal deterioration. Mouritz et al. (1999)
state in their paper that: one of the advantages of 3D woven composites is that they have
higher delamination resistance, ballistic damage resistance and impact damage tolerance.
The 3D braided composites have the similar performance as 3D woven composites
besides they have a greater crashworthiness property (Mouritz et al., 1999). Generally,
braided composites have a higher levels of conformability, drapability and structural
integrity, which makes it possible to produce composite structures with intricate
geometric to the near-net-shape. (Ko and Hartman, 1986; Crane and Camponeschi, 1986;
58
Whitney et al., 1971; Gong and Sankar, 1991). However, one of the major limitations
for 3D braided composites is that their maximum preform size is determined by the
braiding machine size, and most industrial machines are only able to braid preforms
with a small cross-section (Dexter, 1996)
3D knitted composites can be divided into 3D knitted sandwich composites, 3D warp

knitted non-crimp composites and 3D near-net-shape knitted composites (Mouritz et al.,
1999). The recently innovation of 3D knitted composites is ‘Spacer Fabric’, which is
produced on circular knitting machines (Bartels, 2003) and its structure is by interlacing
the upper and lower layers of the fabric with chain yarns through thickness direction and
the schematic illustration is shown in Figure 2-17. Sun et al. (2010) specified in their
paper that spacer fabric has excellent air permeability under high areal pressure and this
kind of structure has high interlaminar shear strength, and can prevent the slide between
upper and lower layer. However, because all the yarns are in curves and not in straight
lines, the tensile stiffness and strength of the spacer fabrics are relatively lower than
those of woven fabrics in the in-plane direction.
Figure 2-17 Sketch of spacer fabric construction.
59
2.5.1 3D woven fabrics
The current study is based on 3D woven fabrics; therefore, reviews will be focused on
this type of fabric and specifically, the reviews will be drawn on 3D woven fabric by
means of multi-layer techniques with hollow structure in between.
The applications of 3D woven structure as core materials for textile composites have
been constantly growing, because they possess major advantages over conventional
materials. First of all, they are generally light weight and show no heat degradation
while processing. And since their fibres are interlaced in the cross-wise, lengthwise and
thickness direction they can withstand multi-axially stresses, which is obviously one of
the key requirements for composites employed in industry and engineering (Mohamed,
1990; Mouritz et al., 1999). In Figure 2-18, a range of diverse complex 3D woven
composites is displayed, which could be found in industries like aircraft, automobile and
civil infrastructure.
Figure 2-18 3D woven composites (a) cylinder and flange; (b) egg crate structure; (c)
turbine rotors; and (d) various complex shapes woven preforms (Mouritz et al., 1999)
60
Some of 3D woven composites are produced with specially-developed weaving
machinery (Buesqen, 1995; Dickson et al., 2000). However such special machinery is
expensive and can only weave a limited range of 3D composites, so that producers are
more interested in ways to produce 3D composites on conventional looms, which are
more economical and versatile in usage.
Popular 3D hollow woven structures, which can be produced on standard weaving

looms, are pictured in Figure 2-19. These structures are known as multilayer woven
fabrics, as they are composed of several series of warp and weft yarns which form
distinct layers, one above the other (Ko, 1989). Obviously the number of layers and the
way they interlace contributes mainly to the through-thickness strength and the
mechanical properties of such structures can be relatively easy altered by varying the
density and types of weft, warp and binder yarn and use of different weave pattern
(Chen et al., 1999).
Figure 2-19 Woven architectures used in 3D woven composites (Yi and Ding, 2004)
2.5.2 3D honeycomb fabric
A woven fabric made from one set of warp and one set of weft yarns is regarded as 2D,
whereas fabric structures with obvious thickness due to addition of warp and/or weft
yarns are referred to as 3D. 3D fabric structures may be made in the form of solid or
hollow depending on its applications and 3D hollow structure are used in creating
composites that are bulky, lightweight and energy absorbent. 3D hollow fabric is
61
featured by one or more layers of triangular or trapezoidal cross-sectional shapes. The
hollows or the tunnels are formed between the adjacent fabric sections and the structure
is self-opening. The use of CAD/CAM systems for woven structures has made the
weaving process more efficient and more versatile and the current textile technology is
capable of creating 3D honeycomb woven fabric with no or little need for machine
modifications. A conventional loom equipped with a dobby shedding mechanism will
be sufficient for making this kind of fabric.
Computer representation of 3D woven fabric structure has been developed by early

researchers such as Hoskins (1983), Xu (1992) at UMIST and they have modeled and
visualized thread paths of fabrics in 3D. Then other workers developed the solid model
for fabrics and the most notably is Keefe et al. (1992). Chen et al. (1996) worked on the
mathematical representations of weaves for 2D and 3D structures in detail, and Chen
and Potiyaraj (1999a; 1999b) implemented the mathematical models and created a CAD
package that covers 3D solid structure and backed fabric structures, as well as single
layer structure.
Tracing back in 1991, Takenaka et al. (1991) has invented a woven fabric having
multi-layer structure to form hexagonal and tetragonal cell shape to form the composite
material and he has hold a patent on this kind of structure, which is shown in Figure 2-
20. In his invention, the thickness of the multi-layer woven fabric can be increased by
increasing the number of woven fabric layers to be superposed.
62
(a) woven fabric with multilayer structure (hexagonal shape)
(b) woven fabric with multilayer structure (tetragonal shape)
Figure 2-20 A schematic diagram of woven fabric with multilayer (Takenata et

al., 1991)
Since Takenaka (1991) invented the muli-layer fabric which can form the honeycomb
structure, the automatically generated weaves have been used to control the shedding
mechanisms of conventional weaving machines at UMIST, leading to successful
production of 3D hollow structures and in 1999, Yassar (1999) managed to weave 3D
63
multi-layer fabric by using polyester yarns with conventional 2D weaving process and
he concluded that an opening process is needed to provide thickness and to convert the
fabrics from 2D form into 3D as shown in Figure 2-21.
Figure 2-21 Cross section view of the honeycomb fabric in 3D form (Yassar, 1999)
By using more advanced IT technology, a significant progress was made by Chen et al.
(2004) followed by Chen and Wang (2006) who mathematically modeled the 3D hollow
woven structures and established an algorithm to create the weave diagrams and lifting
plans for the design and manufacture woven honeycomb structure and their algorithms
were implemented in a CAD/CAM software package specially designed to weave this
woven honeycomb fabric.
Sun (2005) pursued on weaving 3D honeycomb fabric by using cotton fabric and 10
fabrics with different layers and picks that were manufactured in the University of
Manchester and these fabrics were used in the current research to convert them into
textile honeycomb composites. The design of the honeycomb fabric will be explained
more in details in Chapter2.
2.5.3 Structure parameters for textile honeycomb composite

In this section, reviews will be focused on introducing the geometric structure of the
honeycomb from its single cell to the whole structure to help readers building up the
64
general ideas of all the parameters describing the textile honeycomb composites in the
current research.
 : opening angle
l b : bonded wall length
l f : free wall length
t b : bonded wall thickness
t f : free wall thickness
h : height of cell
Figure 2-22 Parameters of single honeycomb cell (Tan and Chen, 2005)
A honeycomb structure is composed of an array of hexagonal cells and cell is the basic
component of the honeycomb structure. A cell structure is formed by free and bonded
walls where free walls refer to the cell sides that are free from other sides, whereas the
bonded walls are those having to be bonded together for the formation of the cellular
cross-section. Parameters such as the opening angle, wall thickness, and length of each
wall are used to describe a cell. Figure 2-22 illustrates the parameters of a hexagonal
cell (Tan and Chen, 2005).
Generally, opening angle  is between 0 and 90. Given the assumption that the
thickness of the yarn does not change during weaving, with the weaving method used in
this study, the thickness of free wall tf is approximately twice the yarn diameter, and the
thickness of bonded wall tb is approximately three times of the yarn diameter. The
height of the cell is calculated as h = 2 lf sin (Tan and Chen, 2005).
A repeat of the cross-section of the honeycomb structure is made of two columns, one
being a cell longer than the other. The overall vertical dimension can be described by
the number of cells in either longer or the shorter column and the horizontal dimension
65
are expressed by the number of repeats of the cell columns. Figure 2-23 illustrates a 6-
layer honeycomb structure. From the top to the bottom, the layers are numbered in
sequence from layer 1 to layer 6. The first layer produces a flat surface to form the
topside of first column hexagon. Then, it lowers and cross-links with layer 2 to form the
second column hexagon, as indicated in Figure 2-23. Layer 2 and layer 3 are
interconnected to form the topside of the second hexagon in the first column and then
divided into two. Layer 3 is connected with layer 4 and layer 5 is connected with layer 6
for the second column hexagons. The same layer connect rule applies on the third
column hexagons. The number of layers of any honeycomb structure should be an
integer larger than 2 (Sun, 2005).
Figure 2-23 Schematic diagram of a 6-layer honeycomb structure (Sun, 2005)
Sun (2005) also mentioned in the case of a textile honeycomb structure, the free walls
will be made by a single layer fabric, and the bonded walls will contain the amount of
yarns for making two layers of fabrics. In weaving the honeycomb fabric, the structure
will be woven flattened, leaving the open angle a redundant parameter at the fabric
stage. The opening angle will be used during the consolidation stage when the 3D
honeycomb fabric will be open up.
2.6 Applications of textile honeycomb composite on PPE
A police officer in a public order emergency is in a high risk situation. There is a range
of potential injuries from slips, trips and falls to missile attack and assault. Rioters will
66
pick up anything to use as a missile from wood, metal, bricks to petrol bombs. There is
also a risk through direct contact with suspects during foot pursuit or restraint. For this
reason there is a range of PPE an officer will wear in a public order situation. With
shields cover most of other parts of the body, most of times; the focus of PPE design is
on how to avoid the attack to lower limbs which is not easy to be covered by shield and
often prone to attacks. Therefore, a new material must be developed in order to provide
better protection and textile honeycomb composite is ideal for this kind of application
because it has an extremely good energy absorption performance.
Bajaj and Sengupta (1992) state that there are three main requirements as how a
protector is protecting wearers against impact loading. Firstly, it needs to disperse the
impact energy from impact point to a large area by hard shells, such as armour, shield,
etc. Secondly, it should delay the occurrence time of peak transmitted force to the
wearer. By observing transmitted force-time diagram of the material, flatter curves
usually are preferable to avoid high peak transmitted force. Lastly, it needs to absorb the
impact shock by energy absorption material with deformation.
Med-Eng Inc. (2001) describes the injury threshold force of lower limbs which can be
used as reference values for PPE design (Dionne et al., 2003; Med-Eng System Inc,
2001). For shin, the threshold values is between 4.30KN and 8.93KN, and for knee, it is
between 7.56KN and 10KN. Therefore, PPE should be designed to reduce the impact
force significantly to be lower than the higher limit of such threshold values in a safer
range. To choose a light weight new material with high energy absorption capability and
to reduce force attenuation is drawing more and more attention in the scientific area and
fabric composites with specially designed honeycomb structure matches the criteria
listed above and could be a good candidate for such purpose.
67
2.7 Comments
The literature review has shown that the honeycomb structured composites is a type of
textile woven composite providing excellent energy absorption and shock protection
comparing to other conventional materials, and the weight of this type of composite is
extremely light (Wang, 2009; Pflug and Veopoest, 1999; Pflug et al., 2002). Therefore,
policeman who are working in a high risk situation, are considering to replace the
relatively heavy shield with new materials to reduce its weight and at the same time, to
improve the shield’s protection and energy absorption performance. Textile honeycomb
composites are one of the options which can meet the requirements correspondingly
However, the mechanical performances of honeycomb structured composites seem to be

affected significantly by their geometric parameters. In more details, the composites’
volume density, cell size, cell wall thickness, cell wall length ratio and cell opening
angle are the parameters which have been investigated frequently by researchers (Barma
et al., 1978; Ashby and Mehl, 1983; Gibson and Ashby, 1997) to seek out these
geometric parameter’s relationship with the composites’ mechanical performances such
as strength and stiffness, deformation pattern, damage tolerance, fatigue performance,
etc. However, most of the studies are based on metal and paper honeycomb structure
composites, few literatures have been reported on textile based honeycomb composite.
It has to be noted that the currently used expansion or corrugation techniques for
making honeycomb structure core materials are not suitable for making textile based
honeycomb structure (Bitzer, 1997) because it cause delaminating between the adhesive
layers easily, Therefore an alternative way, which can simplify and integrate the
honeycomb structure to solve the delaminating problem, has to be found in order to get
a more reliable structure performance.
The aim of this research is to investigate how geometric and structural parameters of
textile honeycomb composites would affect the mechanical performance and energy
absorption capability under low velocity impact. Although a considerable body of
knowledge has been generated in the past years about textile honeycomb composites
(Tan and Chen, 2006, Tan et al., 2007), more research is required to develop design
68
guidelines for optimizing material performance by manipulating the honeycomb
composite’s architecture.
In order to take advantage of the attractive features offered by textile structural

composites, there is a need for the development of a sound database and design
methodologies which are sensitive to manufacturing technology. An examination of the
literature indicated that only a limited number of systematic experimental studies have
been carried out on 3D fabric reinforced composites.
Nowadays, the usage of finite element analysis (FEA) method on the structure
performance for the composites becomes more and more popular due to that this
method makes it possible for the researchers to see clearly the whole crushing process
even at an instant time. FEA provides not only valuable information for composite
localised damage but also establish criteria for acceptance or repair of structural
components (Abrate, 1998). While design of the structural components tends to be very
complex and time-consuming in practices, the development of efficient analytical pre-
processors by FEA can decrease cost and make FE modelling an economic and easy-to-
use solution.
69
CHAPTER3
DESIGN OF 3D HONEYCOMB FABRICS
In order to get the textile honeycomb composites, honeycomb fabrics need to be

produced firstly. This chapter introduces the procedure of design and manufacture of 3D
honeycomb fabrics as reinforcement material. It must be mentioned that this part of
work was carried out in collaboration with a fellow researcher in the University of
Manchester (Sun, 2005; Chen et al., 2008).
It has to be noted that in this part of the experimental work Sun (2005) has designed the
weave structure for the multi-layer honeycomb fabrics by using CAD software created
by Chen and Wang (2006). She also manufactured the honeycomb fabrics on a dobby
weaving machine. The author’s part of the experimental work is on the determination of
the geometric and structural parameters for the honeycomb fabrics for the experiment
purpose, and the manufacture of honeycomb composites from these fabrics.
An algorithm has been established to create weaves based on the specification of the 3D
honeycomb composite parameters, and this algorithm has been implemented into a
CAD programme, Hollow CAD© made in the University of Manchester, which gives
accurate solutions for making reinforcing fabrics of this type (Chen et al., 2004). The
current weaving technology is capable of creating the 3D honeycomb woven fabrics
without carrying out machine modification and 3D honeycomb woven fabrics with
various structural parameters can be engineered from commercial loom. Nevertheless, a
good understanding of weaving process and woven structure is still required in the 3D
honeycomb fabric design in order to characterise the performance of the textile
honeycomb composites in the current research work.
3.1 Design of 3D Honeycomb Weaves

One repeat of a honeycomb fabric can be divided into four regions, and they are regions
I, II, III and IV, as shown in Figure 3-1. Region I corresponds to the section of the 3D
70
honeycomb structure where the fabric layers all separated from each other; region II is
where the adjacent layers join together at an alternate interval; region III is the same as
region I; and region IV is again the joining section but the joining layers are different
from that in region II.
(a) Honeycomb fabric woven before opening
(b) Honeycomb structure after opening
Figure 3-1 Region division of a honeycomb structure
Because of the nature of weaving, the honeycomb fabric is woven with all cells
flattened as indicated in Figure 3-1(a), and the honeycomb structure is achieved when
the fabric is opened up after weaving and consolidated as shown in Figure 3-1(b).
According to the definition of a cell, region II and IV correspond to the bonded walls
with length lb and region I and III the free walls with length lf. Note that Figure 3-1(b)
71
shows a selected part of the honeycomb structure that would open up to from Figure 3-
1(a).
3.1.1 Representation of woven honeycomb structures
In this study, a woven honeycomb structure in weft direction can be defined by

specifying the structural parameters. The following general coding format is used to
denote a particular honeycomb structure:
xL(y+z)Pθ
where, x is the number of fabric layers used to form the honeycomb structure; y is the
length of the bonded wall measured in the number of picks; z is the length of the free
wall measured in the number of picks; θ is the opening angle of the hexagonal cells
which varies from 0˚ to 90˚ and opening angle is the structural parameter to investigate
and compare in the group; L is used to denote the ‘layer’; P is used to denote the ‘pick’.
There are situations when the lengths of free and bonded walls are the same, i.e. y=z. In
such a case, the coding format can be reduced to:
xLyPθ
In the above, x, y, z are integers and x>2, y>1, z>1. When θ is not shown, the opening
angle assumes a default size of 60˚.
According to the format discussed above, a 4L6P honeycomb structure stands for a
structure comprising four layers of fabric, where the length of both the free and bonded
walls is six picks and the opening angle is 60˚. As another example 8L6P45 denotes a
honeycomb structure made from eight layers of fabric, where the length of the free and
bonded walls is six picks with the opening angle being 45˚. On the hand, 8L(4+3)P
refers to a honeycomb structure made from eight layers of fabric, where the lengths for
the bonded and free walls are four picks and three picks, respectively, with the cell
opening angle being 60˚.
72
3.1.2 Layer connection methods
In the honeycomb fabric, adjacent layers are connected at given intervals in order to
achieve the honeycomb effect. There may be many different ways for the adjacent
layers to be woven together. Firstly, the two layers can be woven together as a single
layer fabric where the warp density will have to double in this section of the fabric.
Secondly, it can be woven as a double layer fabric connected together either by stitching
or by layer interchange. Thirdly, the orthogonal or angle-interlock 3D weave structures
can be used for this section of fabric.
There could be further ways for joining the two layers together. In the current work,
however, a type of orthogonal structure is used for connection two layers together (see
Figure 3-3(b)). In this particular construction, half of the warp ends are used for binding
whilst the other half simply embedded in the middle without interlacing with the weft
yarns.
Figure 3-2 illustrates the weaves used for the individual layers (free walls) and the
weave construction for the joined layers (bonded walls). Figure 3-2(b) shows that two
single fabric layers are combined to form a condensed single layer as an example.
(a) plain weave for single layers (b) layer joining
Figure 3-2 Selection of weaves
73
3.1.3 Weave creation
After the honeycomb structure is specified, it is important that the weave is created
accordingly. A procedure for creating waves for the honeycomb structure has been
created and it has been implemented into software Hollow CAD© (Chen and Wang,
2006). In this section, two examples of 2L1P and 4L3P structure are used to explain the
principle of the weave generation for honeycomb fabrics. In all examples, the plain
weave is used for single layer fabric section because of its advantages such as the
simplicity, structural integrity, and good acceptance by the technical end users.
The 2L1P structure
The simplest honeycomb structure is with 2 layers and 1 pick in walls which has been
named 2L1P. Figure 3-3(a) shows the honeycomb structure when opened, and (b) is the
illustration of the interlacement between warp and weft yarns for this honeycomb
structure. It can be seen that there are altogether 4 warp ends, ends 1 and 2 being
responsible for weaving the top layer and 3 and 4 for the bottom layer. The four regions
(I, II, III and IV) are a complete repeat along the warp direction. Warp end 1 is taken to
explain the weave creation. In region I, warp end 1 goes above the two picks and
therefore in the weave diagram shown in Figure 3-3(c) the first warp end received two
warp-up marks. When this warp end travels into region II, it goes under both picks and
therefore in the weave diagram the first warp end receives two warp-down marks, which
are indicated as blank grids. In region III, the warp ends travels above the 2 picks again,
and correspondingly in the weave diagram there are two warp-up marks. In the final
region, this warp end is underneath the pick for the top layer but above the one for the
bottom layer. Accordingly, in region IV in the weave diagram, there is firstly and a
blank then followed by a mark. In the same way, following the movement of warp ends
2, 3 and 4 will complete the columns 2, 3 and 4 in the weave diagram respectively.
74
(a) 2L1P honeycomb structure
(b) Cross-sectional view of 2L1P
(c) Weave diagram
Figure 3-3 Honeycomb structure 2L1P
75
The 4L6P structure
(a) 4L6P structure
(b) 3D view of the 4L6P structure for all four regions (c) Weave diagram
Figure 3-4 Honeycomb structure 4L3P
76
This structure comprises 4 layers of plain woven fabric involving eight warp ends and
the interlayer connections are illustrated in Figure 3-4(a). The interlacing details for
each region are shown in Figure 3-4(b), where there are 12 picks involved in each
region. In region II, the four layers of fabrics are organised into two bonded walls,
whereas in region IV the bonded wall is created by joining layers 2 and 3 leaving layer
1 and 4 to form the free walls. Using the same principle explained already, the weave
diagram can be created where there are eight warp ends all together 48 picks, where is
shown in Figure 3-4(c).
In this work, it is assumed that the tunnels formed by the cells run in the weft direction,
although they can be arranged to go in warp direction too. The reason to make this
arrangement is that less healds are needed during the weaving process when the tunnels
are run in the weft direction. Take 4L6P as an example, according to its weaving
diagram in Figure 3-4(c), when the tunnels are run in the weft direction from left to
right in Figure 3-4(c), it is clearly that there are only 8 different design patterns in one
weave repeat, therefore, 8 healds will be needed to lift the warp ends to form the fabric.
However, if the tunnels are run in the warp direction from bottom to the top of the
weaving diagram in Figure 3-4(c), there are 14 different design patterns shows up along
warp direction, therefore, 14 healds will be requested to lift the warp ends for the fabric
production. And the same rule applies to other honeycomb fabric with 6 and 8 layers too.
3.2 Design of 3D Honeycomb Fabrics

3.2.1 3D honeycomb fabrics
In this research, plain weave has been chosen for the first attempt as this is one of the
easiest ways to induce multilayer analysis. It also helps to simplify the manufacture
process by using plain weave structure as it reduces the healds significantly. As shown
in Figure 3-2(a), the angular sides of each hexagon to the axis were created by the plain
weave and most of the top and bottom sides of a hexagon were formed by
interconnecting the two layers, shown in Figure 3-2(b), except the topside of the first
layer and the bottom part of the last layer are just plain weave.
77
In order to investigate systematically the 3D honeycomb composites, 10 honeycomb
fabrics are designed and manufactured, which are 4L6P, 6L4P, 8L3P, 8L4P, 8L5P,
8L6P, 8L(4+6)P, 8L(4+3)P, 8L(3+6)P and 8L(6+3)P. This is based on the following
geometric parameters of a honeycomb structure:
1) Cell opening angle, 
2) Different cell size at the same number of layers
lb
3) Length ratio of cell walls,
lf
4) Similar sample thickness with different cell size
According to the above four structural parameters and weaving capability in the
laboratory, the fabric to be woven is categorized into four comparison groups. An
outline of the designed fabric types and desired structural parameters are listed in Table
3-1 and Table 3-2.
Table 3-1 also shows the actual needs of quantities of all the fabric types to weave.
78
Table 3-1 List of fabric types with weaving quantity and designed angle
Index Fabric l θ Sample Quantity

(m) (º)
1 4L6P 2 60 1
2 6L4P 2 60 1
3 8L3P 2 60 1*
4 8L4P 2 60 1
5 8L5P 2 60 1
6 8L6P 2 30 1
2 45 1
2 60 1*
2 75 1
2 90 1
7 8L(4+6)P 2 60 1
8 8L(4+3)P 2 60 1
9 8L(3+6)P 2 60 1
10 8L(6+3)P 2 60 1
Total 10 types 28 --- 14
where, l represents the total length of the fabric been produced in the unit of m; θ
means the opening angle for the fabric in the future composite design; the fabric which
is marked with ‘*’ will be used for comparison in three different group for the future
experiments.
The fabric types which are categorized into 4 groups for comparison are listed in Table
3-2.
79
Table 3-2 Experimental design outline in groups
Group 1. Opening angles, θ(º) (5 samples)

Index Sample θ (º)
1 8L6P 30
2 8L6P 45
3 8L6P 60
4 8L6P 75
5 8L6P 90
Group 2.Different cell size at the same number of layers (4 samples)

1 8L3P 60
2 8L4P 60
3 8L5P 60
4 8L6P 60
lb
Group 3. The ratio of wall length,  1 (3 samples)
lf
Index lb Sample θ (º)
lf
1 1:2 8L(3+6)P 60
2 2:3 8L(4+6)P 60
3 1 8L6P 60
lb
Group4. The ratio of wall length,  1 (3 samples)
lf
1 1 8L3P 60
2 4:3 8L(4+3)P 60
3 2 8L(6+3)P 60
Group 5. Same sample thickness with different cell size (3 samples)

1 4L6P 60
2 6L4P 60
3 8L3P 60
80
3.2.2 Design details for 3D honeycomb fabrics
Opening angle, cell size, length ratio of cell walls have been regarded as the potential
key geometric parameters to influence the mechanical behaviour of textile honeycomb
composite (Tan and Chen, 2005; Tan et al., 2007) and it is of interest the current
research to further investigate them in details.
Before the textile honeycomb composites are made from honeycomb fabric, the design
details for various fabric types will be explained in the following sections.
3.2.2.1 Cell opening angle, 

The opening angle of cell units is an important parameter for 3D honeycomb composite
structure, which can change the structure thickness and material use efficiency as well
as mechanical properties.
To show the comparability of this group, diagram in Figure 3-5 has shown five open
8L6P woven honeycomb structures in weft views with different angles of 30, 45, 60,
75 and 90.
In Figure 3-5, it is worthwhile to note that the thicknesses of the specimens are different
after the change of the cell opening angles if the number of layers is kept constant (it is
8 in this group). Therefore, in this group, results from comparison may also indicate the
effect from different thicknesses of specimens. Additionally, to achieve a comparable
result in this group, the wall lengths for all samples are all the same by fixing the pick
numbers of all walls to 6.
81
Figure 3-5 8L6P with different opening angle
3.2.2.2 Different cell size at the same number of layers
Cell size is another significant parameter for honeycomb structure besides opening
angle which also influences the energy absorption capacity effectively.
In this study, the change of cell size is made by changing the length of the walls, or in
another word, changing the pick numbers of walls in weaving. With the same density of
each layer, the length of wall will be increased or decreased by adding or reducing picks
number. To show the comparability of this group, diagrams of open honeycomb
structures in weft view for fabric types in this group (8L3P, 8L4P, 8L5P, 8L6P) are
shown in Figure 3-6. Except the change of wall lengths, the number of layers is fixed to
the same at 8 and all cells are with opening angle of 60. The bonded wall length is the
same as the free wall length, lb= lf..
Again, in Figure 3-6, significant change of specimen’s thickness may be observed after
the change of the cell size while the layer number is fixed for all types. Same as the first
82
group, the effect from the change of the specimen’s thickness might combine with the
effect of cell size on the mechanical properties of the specimens in the mechanical tests.
Another point to note out is, due to the limited yarn purchased, 8L7P were not woven in
actual study. But it is believed giving up 8L7P should not influence the comparability of
this group and discussions on the final results of the group.
Figure 3-6 different cell size for 8-layer composites
lb
3.2.2.3 Length ratio of cell walls ( )
lf
In practical use, the bonded wall length lb and the free wall length lf do not have to be
lb
the same. Therefore, it is valuable to investigate how such ratio of might change the
lf
energy absorption capability of the honeycomb composites.
lb
Selectively, current study chose 6 sample types with in two subgroups:
lf
lb
The first subgroup (  1 ) includes 3 sample types: 8L(3+6)P, 8L(4+6)P, 8L6P. The
lf
fact that pick numbers of the free wall for these three types are all 6 indicates there is no
change of the free wall length in this subgroup. Therefore there is no change of the
83
structure thickness. However, by changing the pick number of bonded wall (the bonded
lb 1 2 1
wall lengths), the ratios for these three types can be set as , and respectively.
lf 2 3 1
Apparently, the bonded wall length is shorter than the free wall length in this subgroup.
lb
Similarly, the second subgroup (  1 ) consists of 8L3P, 8L(4+3)P, 8L(6+3)P with
lf
1 4 2
ratios of , , respectively. This time, the bonded wall length is longer than the free
1 3 1
wall length. This time, the pick number of free wall is specified at 3, but changing the
pick numbers of bonded walls to 3, 4 and 6.
All fabric types in both subgroups are with 60 opening angle as shown in schematic
diagrams in Figure 3-7 and Figure 3-8 for two subgroups separately.
Different from the previous two comparison groups, since there is no change of free
wall lengths and opening angles, there is no change of specimen’s thickness within each
subgroup. Therefore, by comparing the results of each subgroup, the difference of
lb
energy absorptions will be from the different solely.
lf
84
lb
Figure 3-7 Honeycomb structures with length ratio of cell walls (  1)
lf
lb
Figure 3-8 Honeycomb structures with length ratio of cell walls (  1)
lf
3.2.2.4 Similar sample thickness with different cell size
To avoid the potential effect from the thickness change, this group which includes 4L6P,
6L4P, 8L3P with the same opening angle 60 is also designed with same specimen’s
thickness. This is because 4l4sin60=6l6sin60=8l8sin60 where l4, l6, l8 are with pick
numbers of 6, 4 and 3 respectively. In such circumstances, changing number of layers is
also changing the cell size at the same time. Thus, the comparison results of this group
are potentially the combined effects from number of layers and cell size, but without the
effect from specimen’s thickness. Figure 3-9 presents the above specifications in weft
view of open honeycomb structures in diagrams.
85
Figure 3-9 Structures with same thickness and different cell size
3.3 Manufacturing of 3D honeycomb fabrics
In order to investigate systematically the 3D honeycomb composites, 10 honeycomb

fabrics are designed and will be manufactured, which are 4L6P, 6L4P, 8L3P, 8L4P,
8L5P, 8L6P, 8L(4+6)P, 8L(4+3)P, 8L(3+6)P and 8L(6+3)P.
This part of thesis presents the weft density of the 3D honeycomb fabric and the
detailed parameter specifications for fabrics in real weaving lab. The lifting plans
designed from Hollow CAD© [Sun, 2005] for each fabric type are also shown in this
section.
3.3.1 Weft density of the 3D honeycomb fabric

Before making the fabrics, the weft density of the overall fabric must be worked out
based on the number of picks specified for each of the cell walls in the fabric
specification and the actual length expected. Suppose that z is the number of picks
specified for a wall of single layer fabric session (picks), l is the required length of the
fabric session (cm), di is the weft density for this single layer (picks/cm), then
z
di = [3-1]
l
If the honeycomb fabric is composed of m fabric layers, then the weft density of the
honeycomb fabric d, will have to be set to
86
m
d=  di
i 1
[3-2]
In the case that all layers have the same weft density, the weft density of the honeycomb
fabric becomes
d = mdi [3-3]
The honeycomb fabrics are designed to have four, six, and eight fabric layers. In all
cases, it was decided that each layer of fabric will have a warp and weft density of 20
picks/inch (7.87 picks/cm). The warp and weft density of the overall honeycomb fabric
can be found by multiplying the number of layers to the warp and weft density per layer,
as described in equation [3-3]. Therefore, the warp and weft densities of these three
honeycomb fabrics are 80, 120, and 160 picks/inch (31.5, 47.2 and 63.0 picks/cm). In
all these designs, the tunnels run in the weft direction, and the length of the cell walls
can be found using the equation [3-1]. With the wall length being 3, 4, 5 and 6 picks, the
actual length of the walls will be 3.8mm, 5.1mm, 6.4mm and 7.6mm respectively.
This research does not intend to examine the influence of fiber type and yarn parameters,
including yarn twist and yarn linear density, on composite properties because all fabrics
are made from the same types of yarn (14.8tex/3, 3×40’s cotton yarns, with the ply
twists of 433turns/m) was used for making the 3D honeycomb fabrics.
3.3.2 Parameter specifications for 3D honeycomb fabric in the weaving process
As part of weaving design, the detailed parameter specifications for fabrics and lifting
plan are important for the honeycomb fabric manufacturing as they will be directly used
in the real weaving loom. However, the opening angles are actually formed in resin-
impregnation process after the fabric been produced.
Take one group of samples with different cell size at the same number of layers as an
example (8L3P, 8L4P, 8L5P and 8L6P). To maximum capability of the weaving
87
machine, the density of each layer is fixed as that for the reed, 20 picks/inch. Therefore
the total density for 8 layers will be 160 picks/inch.
Given the desired width of 12 inch (304.8mm) for the fabrics, the total number of warp
ends can be calculated as total density per inch multiply total width, which is 160
picks/inch inch = 1920. This result implies that 1920 warp yarns should be used for
weaving eight layers 12 inch wide fabrics in this study.
For a weaving repeat which consists of four parts discussed in section 3.1.3, the total
pick numbers for a full circle are calculated as pick numbers at each part times four
parts times number of layers, which is 6×4×8 =192. This indicates that 192 weft yarns
will be woven for a full weaving repeat. The total weft yarns can be determined from
the desired length of the fabric, which is listed in Table 3.1.
Given the weaving weft density of 20 picks/inch and the pick numbers of walls are 6 as
an example, the lengths of walls lb and lf can be calculated as 6 ÷ 20/inch = 0.30 inch
(7.62 mm). It is based on the assumption that the yarns would strictly follow the density
specifications in weaving.
Among the four fabric types, the pick numbers in weft direction are different due to the
different pick numbers in weft. For a full weft weaving repeat, the picks are calculated
for four types as follows:
8L3P: 3×4×8=96 picks;
8L4P: 4×4×8=128 picks;
8L5P: 5×4×8=160 picks;
8L6P: 6×4×8=192 picks.
88
Consequently from the difference of the pick numbers in cell walls, the lengths of
bonded walls and free walls are different. Following the same calculation methods in
the first group, these are calculated as: Given the weaving density of 20 picks/inch and
the pick numbers of walls are 6, the lengths of walls lb and lf can be calculated as 6 ÷
20/inch = 0.30 inch (7.62 mm). It is based on the assumption that the yarn diameter
bears no significant change during weaving.
l8L3P = 3 ÷20/inch = 0.15 inch (3.81 mm)
l8L4P = 4 ÷20/inch = 0.20 inch (5.08 mm)
l8L5P = 5 ÷20/inch = 0.25 inch (6.35 mm)
l8L6P= 6 ÷20/inch = 0.30 inch (7.62 mm)
For comparison, all weaving designs from Hollow CAD© (Sun, 2005) for 8L3P, 8L4P,
8L5P, 8L6P are shown as in Figure 3-10. The weaving design of 8L6P is used for fabric
types in the first group as well.
89
8L3P 8L4P 8L5P 8L6P
Figure 3-10 Weave lifting plan for 8L3P, 8L4P, 8L5P and 8L6P
90
3.3.3 Honeycomb fabric production
The fabric production was conducted in the weaving laboratory of the School of
Materials in the University of Manchester.
A dobby loom with maximum 16 heald frames was used for the fabric manufacture. It
can be used for weaving up to 8 layers of fabrics when the plain weave is used for all
layers. Three weaver’s beams can be used for warp supply, and each beam can contain a
maximum of 1000 warp ends. Punched cards are used to control the dobby shedding
mechanism. Figure 3-11 shows the overall look of the weaving machine.
For warp ends and weft picks in fabric in this study, the white 100% cotton 40’s/3
(14.8tex/3, 3 × 40’s cotton yarns) yarns were selected due to its relatively higher
tension tolerance feature comparing to other fabric materials. Using the same materials
for both warp ends and weft picks is also for the convenience consideration in further
results calculations and comparisons.
Considering the weaving capability of the dobby machine in the lab, 20 picks/inch at
each layer is used for the densities of both warp and weft. For multilayer fabrics, the
total density for the fabric can be achieved by multiplying 20 picks/inch and number of
layers. In order to produce comparable results, the fabric density was kept constant for
all fabrics in this study.
91
Figure 3-11 The dobby weaving machine
(a) card punching machine (b) card punching
Figure 3-12 Card punching
92
A card is a pattern chain that contains the lifting information to control the movement of
the dobby shedding mechanism. A pattern card is done by punching a hole in the card at
the position where the unit square of weaving designs in Chapter 3 is in blue. Since
straight drawing-in heald frames are used for threading, each weaving design described
in Chapter 3 was punched on cards using the card punching machine shown in Figure 3-
12. It is worth noting that the left bottom unit in the designs should be used as the
starting point of punching. After punching the whole circle of the weaving design is
finished, a few extra rows should be produced for card circle formation purpose.
Normally the first two rows are re-punched for overlapping.
As there are 10 multi-layer fabrics were designed for this study which would make 14
honeycomb composites after opening, 10 cards were punched according to the weaving
designs (lifting plans). One of the finished punched cards is shown in Figure 3-12.
In the present study, three beams are used for weaving. A warp density 20 picks/inch
has been chosen for each layer, therefore 4 to 8 layer fabrics should be with the warp
density from 80 picks/inch to 160 picks/inch. The width of the woven fabrics is set to be
12 inches (304.8 mm), therefore the total number of warp ends can be counted as 960 to
1920 correspondingly. The desired length of each fabric determines the length of warp
at each beam. Warping yarns on the beams will be different if the number of warp ends
is different. It depends on the number of layers to be woven if the density is the same. In
this study, warping is required at least for three times: one for 4-layer structure, one for
6-layer structure and the third one for 8-layer structure.
Take 8-layer fabric types as an example, there are eight different fabrics to be woven,
and 8L6P may be woven longer as it is required in five different opening angles when it
is opened. Since all the 8-layer fabrics share one warping of 1920 ends, ten meters for
8L6P, two meters for rest seven kinds of 8-layer fabric types and all 8-layer fabric types
in total give 24 metres for the useful part of the fabrics. Considering the extra three
metres in warping, there will be 27 metres warping length of yarns on each beam at
least for weaving all 8-layer fabrics. Figure 3-13 illustrates the cross section of an
sample fabric which has been produced from the loom.
93
For the whole weaving laboratory work, relatively simpler structures were woven, and
4L6P was the first fabric made with 8 heald frames were used, followed by 6L4P after
adding four more heald frames on the weaving machine. For all 8-layer fabrics designed,
16 heald frames were required. All 8-layer structures woven include 8L3P, 8L4P, 8L5P,
8L6P, 8L(4+3)P, 8L(6+3)P, 8L(3+6)P and 8L(4+6)P.
The successful weaving in the lab proved the correctness and effectiveness of the
weaving design from Hollow CAD© (Sun, 2005). This package, may be the first one of
its type, is potentially important and useful in automatic manufacturing in textile
industries.
Figure 3-13 Photograph of one sample fabric weaved from loom
94
CHAPTER 4
CREATION OF HONEYCOMB COMPOSITES AND TEST SAMPLE

PREPARATION
As mentioned previously in Chapter 3, the honeycomb fabrics are in flat form when
manufactured and need opening before impregnation. This chapter will introduce the textile
honeycomb composite production process. This includes the use of an opening device which
has been designed at the University of Manchester and resin impregnation of the fabric to form
composites. Four groups of textile honeycomb composites with different geometric parameters
will be produced, and the division is done according to the cell size, cell opening angle, length
ratio of cell walls and samples with similar thickness but different cell size.
4.1 Fabric Opening and Consolidation
4.1.1 Fabric opening
The fabrics were cut into the size of 20cm×20cm before opening. The schematic diagram of
the opening device is shown in Figure 4-1(a).
Two sets of metal wires, illustrated as A and B in Figure 4-1(a), were used to open the woven
honeycomb fabric and they were laid on top of the surface of four metal bars. The fabric was
hold tightly and then a set of stainless steel wires with 3mm diameter, coated with
polytetrafluoroethlene (PTFE,) were inserted into the top and bottom tunnels of the fabric
respectively. The top and bottom wires were pushed apart to the desired vertical distance by
adjusting the screws at both ends of the opening device. The edges of the wires inserted into the
top layer of the honeycomb fabric were placed onto the surface of two up metal bars and the
edges of the wires which have been inserted into the bottom layer of the honeycomb fabric
were allocated onto the surface of another two lower metal bars. The photograph of the opening
device is captured in Figure 4-1(b).
95
A: wires inserted to the top hexagon in the fabric
B: wires inserted to the bottom hexagon in the fabric
dw: distance between the two metal wires
m: adjustable metal screw
(a) Schematic diagram of the fabric opening device
(b) Photograph of the fabric opening device
Figure 4-1 Honeycomb fabric opening devices
96
Although the honeycomb fabrics for composites could be designed with either even or odd
layers, in the current study, the honeycomb fabrics are woven with even layers. Therefore, in
the following equations [4-1] and [4-2], the (x) is taken as an even integer. The distance
between the lower and upper metal bars, dw, of the fabric opening device is adjustable in order
to open the fabric up to the required opening angle. For given lengths of the cell walls, the
height of opened honeycomb structure relates to the cell opening angle. For an xLyPθ sample
where z=y, if the free wall length is lf, and the free and bonded wall thicknesses are tf and tb
respectively, then the opening angle and the height of the opened honeycomb structure T can be
expressed in Figure 4-2 as follows:
Figure 4-2 Illustration of the thickness (T) of the honeycomb structure
x 
T  xl f sin   2t f    1t b [4-1]
2 
where x is the layer of fabrics involved in the honeycomb structure, and it is an even
integer and x >2.
To achieve a given cell opening angle, the distance between the two sets of the metal wires, dw,
should be
x 
d w  xl f sin  2t f    1tb  2d [4-2]
2 
97
where, d is the diameter of the metal wires and x is an even integer with x >2.
In the above equations, tf is estimated to be the sum of the warp and weft yarn diameters, and tb
is equal to 1.5 of tf .
Figure 4-2 illustrates a honeycomb composite that is made from 4 layers of fabrics. The
distance between the top surface of the up steel wire and the bottom surface of the lower steel
wires (indicated as dw in Figure 4-1) is the critical parameter that determines the opening angle.
Since the wire diameter is known and the desired height of the honeycomb structure is also
decided, then this distance can be calculated by subtracting the thicknesses of the top and
bottom layers of the honeycomb structure. The thickness of the top layer (tf) and the bottom
layer (tb) of the honeycomb composite were measured by ruler and the calculated heights of the
various textile honeycomb composites and the distance between the up and lower wires were
calculated according to Equation [4-1] and [4-2] and listed in Table 4-1.
The calculation procedures are outlined as follows, taking 4L6P with opening angle θ=60°as
example (Figure 4-3).
Figure 4-3 Illustration of a four-layer honeycomb composite
The length of the free walls (lf) and the bonded walls (lb) of a cell can be calculated from the
weft density and pick numbers in the walls. For 4L6P, the weft density is 7.87 picks/cm, and
the free and bonded walls both contain 6 picks in them. Therefore, the lengths of the free wall
(lf) and bonded walls (lb) are both 6(picks)÷7.87(picks/cm) =7.62mm. The number of fabric
98
layers (x) of 4L6P is 4. The measured thickness of the free wall (tf) is 0.78mm and that of the
bonded wall (tf) is 1.09mm. Accordingly, the height T of the 4L6P (opening angle θ=60°)
composite, according to Equation [4-1], is
x  4 
T= 4 ×lf ×Sin θ +2×tf +  1  t b = 4 ×7.62×Sin 60°+2×0.78+   1  1.09 = 29.04 mm.
2  2 
The distance between the wires, as a result of subtracting dlayer from T is 22.65 mm.
Based on the designed weft density of each of the fabric layer (7.87picks/cm), all the relevant
structural parameters are summarised in Table 4-1, taking d=3mm (where, d is the diameter of
the metal wires).
Table 4-1 Calculated sample heights and distance between wire and other design parameter
θ lb+2lf tf tb T dwire
Sample
(°) (mm) (mm) (mm) (mm) (mm)
4L6P 60 22.9 0.78 1.09 29.04 23.04
6L4P 60 15.2 0.75 1.12 30.13 24.13
8L3P 60 12.2 0.92 1.18 31.77 25.77
8L4P 60 15.2 0.85 1.04 40.00 34.00
8L5P 60 19.1 0.80 1.02 48.99 42.99
8L6P 30 22.9 0.80 1.10 35.37 29.37
8L6P 45 22.9 0.79 1.10 47.97 41.97
8L6P 60 22.9 0.82 1.11 57.75 51.75
8L6P 75 22.9 0.80 1.10 63.77 57.77
8L6P 90 22.9 0.82 1.10 65.90 59.90
8L(3+6)P 60 19.1 0.85 1.07 57.69 51.69
8L(4+6)P 60 20.3 0.84 1.07 57.67 51.67
8L(4+3)P 60 12.7 0.80 1.05 31.14 25.14
8L(6+3)P 60 15.2 0.79 1.12 31.33 25.33
99
Where in Table 4-1, θ the is the opening angle of the cell; lb and lf are the lengths of the bonded
and free walls; tb and tf are the thickness of the bonded and free walls; T is the specimen height;
and dwire is the wire distance.
4.1.2 Fabric impregnation
The solution for fabric consolidation was made as a mixture of resin and hardener. The criteria
for selecting the resin and hardener are determined by the requirement that the solution should
penetrate and wet all the layers of the fabric in order to form a continuous rigidified composite
structure, also that the composite should cure within a reasonable period of time, for example,
within 24 hours. Hence, the viscosity of the solution is quite important in order for it to adhere
to the specific fabric materials uniformly. It is worth noting that the strength of the resin and
hardener is also a factor in influencing the capability in energy absorption of the textile
honeycomb composite, a performance that is important to seek in analysis (Miravete, 1999; Wu,
2003).
Following the comparison of the three major resin systems by Wu (2003), the selection of the
resin and hardener is as follows:
Resin: LY5152 – Epoxy phenol novolak resin (60-70%)

Botanediodiglycidy (34-42%)
Hardener: HY5052 – 2,2-dimethy 1-4,4 methylenebis (cyclohexylamine) (50-60%)

Isophorone diamine (35-45%)
2,4,6-tris (dimethyl-aminomethyl) phenol (1-5%)
The mixing ratio of the resin and the hardener is LY5152:HY5052=100:38.
There are many methods to impregnate the reinforcing fabrics, such as the vacuum process,
spraying, rolling and brushing (Wu, 2003). Due to the complexity of the reinforcing fabrics in
this work, the brushing method was chosen in the current study to convert the soft fabric into a
honeycomb composite.
100
The detailed procedure is described as follows: the fabric was placed flatly when the
resin/hardener mixed solution was brushed onto both sides of the fabric. After all sections of
the fabric had been wet through, the impregnated fabric was opened using the fabric opening
device described in Figure 4-1. The impregnated fabric was left for 30 minutes in the ambient
atmosphere and after the 30 minutes interval, the impregnated fabric was turned over to try to
achieve a uniform distribution of the resin/hardner solution inside the cells of the composites.
During the curing procedure, the specimen was placed in the fume cupboard for quicker
hardening. It took around 24 hours for the composite to be cured. Finally, reducing agents were
used to release the wires from sticking to the fabric during the curing process. Tapes of
polytetrafluoroethlene (PTFE), which has high resistance to adhesion, were also wrapped onto
the wires before being used to open the fabric to avoid the adhesion of the wires to the fabric in
the consolidation process.
However, it has to be noted that because the resin was hand painted onto the honeycomb fabric,
and there are no facilities to strictly control the resin being evenly distributed inside every cell
of the composites, therefore, it caused the thickness of the cell walls various to each other
among all the honeycomb composites and the detailed thickness of cell free wall (tf) and
bonded wall (tb) are listed in Table 4-1. Another index to show how many percentage of fabric
and resin are contained inside the honeycomb composites is defined as following:
M fabric
R  100 (%) [4-3]
M composite
where the R means the fabric/resin ratio; Mfabric is the weight of the fabric and Mcomposite is the
weight of the honeycomb composite.
The calculated fabric/resin ratios (R) are listed in Table 4-2, and it seems that the specimen of
4L6P60 and 8L3P60 contains more resin than the rest of the samples. This can affect the
performances of the resulting composites, and for example, if the sample coated with a thicker
resin provides a better force protection and it is hard to explain whether this is caused by the
101
composite’s structure optimization or by its heavy coating. The issue should be considered
when conducting the data analysis in the later sections.
Table 4-2. Fabric/resin ratio for the honeycomb composites
Sample Weight of Honeycomb Weight of Honeyocmb Fabric/Resin

Fabric (Mfabric)(g) Composite (Mcomposite)(g) Ratio (R)
4L6P60 3.10 32.46 9.55%
6L4P60 4.31 40.30 10.69%
8L3P60 5.34 40.76 13.10%
8L4P60 5.52 39.52 13.97%
8L5P60 6.26 41.57 15.06%
8L6P30 5.87 61.61 9.53%
8L6P45 6.12 40.76 15.01%
8L6P60 5.87 40.25 14.58%
8L6P75 5.74 45.52 12.61%
8L6P90 6.38 52.07 12.25%
8L(3+6)P60 5.33 37.87 14.07%
8L(4+6)P60 5.59 50.49 11.07%
8L(4+3)P60 5.31 32.84 16.17%
8L(6+3)P60 5.96 32.15 18.54%
4.1.3 Textile honeycomb composite
After the resin/hardener solution had cured thoroughly, the manufactured textile honeycomb
composites were cut into small specimens with the dimensions of 60mm ×120mm for the
future impact testing. For illustration purposes, textile honeycomb composites with different
cell sizes (8L3P, 8L4P, 8L5P) are shown in Figure 4-4.
102
8L3P 8L4P 8L5P
Figure 4-4 Photos of textile honeycomb composite with different cell size
4.2 Fabrication of Woven Honeycomb Composite
Fourteen groups of textile honeycomb composites with different parameters were manufactured
from 10 types of honeycomb fabrics. Each group of composites has nine samples and they were
cut into size of 60mm×120mm.
The cell structure geometric parameters for these testing specimens are listed in Table 4-2.
However, the specimen dimensions are according to the real samples been made and some
modifications were applied which restricted by the engineering equipment provided.
It is noted that the real composite height (T) in Table 4-3 is different from the calculated values
in Table 4-1. In opening the honeycomb structure, two metal wires of diameter of 3mm were
used for each cell. The metal wires used to lift the top and bottom cells of the woven
honeycomb fabric provide a width of 6mm. Therefore, for the top and bottom layers of the
textile honeycomb composite, the bonded wall (lb) is 6mm in length which causes the length of
free walls (lf) to change from the calculated value. Take sample 4L6P for an example. The
length of lb+2lf is 22.86mm according to Table 4-1 and therefore, the length of lf should be
22.86mm÷3=7.62mm. However, in the real case, the lf is (22.86 - 6) ÷ 2=8.43mm. This is one
reason that has caused the inaccuracy in the height (T) of the composites. Another reason
contributing to this problem is that when the metal wires were handled to separate the surfaces
of the woven honeycomb fabric, they may not be exactly centrally allocated in the cell. Thus, a
result of lf could be even different from above calculated 8.43mm. The third reason to cause this
103
problem is that manually to measure the size of the composites by ruler could bring the
inaccuracy in the final data acquired.
Table 4-3. Honeycomb geometric parameters for testing specimens (by real measurement)
θ dwire
Sample T (mm)
(°) (mm)
4L6P60 67 29.2 23.2
6L4P60 65 30.4 24.4
8L3P60 60 31.6 25.6
8L4P60 62 40.4 34.4
8L5P60 60 49.2 43.3
8L6P30 38 35.7 29.7
8L6P45 45 48.3 42.3
8L6P60 64 58.0 52.0
8L6P75 70 64.1 58.1
8L6P90 80 66.2 60.2
8L(3+6)P60 53 57.1 29.7
8L(4+6)P60 61 57.1 29.7
8L(4+3)P60 66 31.6 25.6
8L(6+3)P60 61 31.6 25.6
4.3 The Sample Groups
The above fourteen textile honeycomb composite with different parameters were categorized
into four groups for the future analysis. These groups and their features are described as
following:
104
Group 1: Composites with different opening angles
All the composites in this group were made from 8L6P honeycomb fabric, which are composed
of 8 fabric layers and 6 picks in each cell wall with different cell opening angles of 30°, 45°,
60°, 75° and 90° respectively. Consequently the thickness of the sample increases as the
opening angle gets larger.
8L6P30° 8L6P45° 8L6P60° 8L6P75° 8L6P90°
Figure 4-5 Specimens with different opening angle
Group 2: Composites with different cell sizes
8L3P60 8L4P60 8L5P60 8L6P60
Figure 4-6 Specimens with different cell sizes
All the composites in this group are based on eight-layer structures but the lengths of the
hexagonal cell walls are changing from 3, 4, 5, to 6 picks which results in a change in cell size.
The opening angle for this group of composites is 60°for all the samples. When the length of
105
the wall gets longer, the specimen gets thicker. Figure 4-6 clearly illustrates the change of the
cell size.
Group 3: Composites with different length ratios of cell wall
The third group of composites was made from an eight-layer fabric, but with different length
ratios of free wall to bonded wall. The cell opening angle is 60°for all the samples. In this
lb l
group, the composites are divided into two subgroups with  1 and b  1 respectively.
lf lf
The first subgroup includes three samples: 8L(3+6)P60, 8L(4+6)P60, and 8L6P60. They share
the same free wall length of 6 picks, whereas the bounded wall lengths change from 3 picks to
4 picks to 6 picks. These composites are shown in Figure 4-7. They are made to have the same
thickness since the thickness, which is determined by the free wall length and opening angle, as
they are all made from 8 layers of fabrics.
It needs to mention that due to an inaccurate operation in using the fabric opening device, the
thickness of 8L(3+6)P60 composite is slightly thinner than the other two in the group. This is
evident in Figure 4-6 and may affect the test result.
8L(3+6)P60 8L(4+6)P60 8L6P60
lb
Figure 4-7 Specimens with different length ratios (  1)
lf
106
The second subgroup includes composites 8L3P60, 8L(4+3)P60 and 8L(6+3)P60, where the
free wall length in all is 3 picks and the bonded wall length is 3, 4, and 6 picks respectively.
Figure 4-8 shows the photo of these three composites.
8L3P60 8L(4+3)P60 8L(6+3)P60
lb
Figure 4-8 Specimens with different length ratio of cell walls (  1)
lf
Group 4: Composites with the similar thickness
The last group of composites was made from fabrics with different numbers of layers and
different cell wall lengths, but they were opened to very similar composite thickness, with the
opening angle being 60°in all cases. They include composites 4L6P60, 6L4P60, and 8L3P60,
which are shown in Figure 4-9 with significant cell size variation.
4L6P60 6L4P60 8L3P60
Figure 4-9 Specimens with same thickness
107
4.4 Summaries
In this chapter, a device which has been used to opening the woven honeycomb fabric has been
designed and introduced at the University of Manchester for this research purpose. The resin
was impregnation onto the fabric to form textile honeycomb composites. However, by hand
painting the resin, it is not evenly distributed all over the fabric and this caused variation in the
measured value of composite’s height (Tsample) and wire distance (dwire). Therefore, more
advanced technology such as vaccum-assistant-resin transfer-moulding can be used to
consolidate the textile honeycomb composite in the future.
Fourteen textile honeycomb composites with different geometric parameter are made from ten
types of woven honeycomb fabric and they are divided into four groups with different cell size,
opening angle, length ratio of cell walls and similar thickness but different cell size. These
composites will be used in the next experiment stage to investigate their mechanical
performance against low velocity impact test.
108
CHAPTER 5
EXPERIMENTAL DATA ANALYSIS ON TEXTILE HONEYCOMB

COMPOSITES
As mentioned in the previous chapters, the three-dimensional (3D) textile fabrics were
consolidated into textile honeycomb composites with different geometric parameters
and were designed in four groups, i.e. the cell size group, the opening angle group, the
length ratio group, and the similar thickness group. It is of academic interest to
investigate the influence of these parameters on the impact performances on these
honeycomb composites. This chapter aims to analyse and compare the impact
performances between the groups of the textile honeycomb composites.
This chapter will start by describing the low velocity impact instrument, also known as
the dropping hammer system, and followed by the presentation and analysis of the
experimental results from testing the textile honeycomb composites. A discussion is
also carried out to evaluate the effect of the geometric parameters on the performance of
the composites.
5.1 Low Velocity Drop Weight Impact Tests

5.1.1 Basic principle of low velocity drop weight impact
For low velocity drop weight impact, the assumption is that the friction between the
impactor assembly and the rails it drops along can be neglected and there is no energy
loss while the potential energy is converted to kinetic energy. According to the energy
1
conservation law, the kinetic energy ( mv 2 ) carried by the impactor assembly (known
2
as the external energy) at the start of the impact should be the same as the sum of energy
absorbed by honeycomb composites through deformation (known as internal energy),
energy transmitted through the honeycomb composite, and other forms of energy during
109
the impact process. In an ideal situation where there is no fracture, this can be
mathematically described as follows:
s
E   Fload ds [5-1]
0
where, ∆E is the change in the kinetic energy carried by the impactor assembly, Fload is
the loading force or contact force applied to the composite at a given time during the
impact, and s is the deformation depth. If the composites are fractured, the energy taken
to fracture the composite cell walls must also be counted in this equation.
In the present study, the low velocity impact test was conducted along the in-plane
direction of the cells in the honeycomb composites. As indicated in Figure 5-1, x1 is the
warp direction, x3 the weft direction, and x2 the thickness direction. Although it is
possible to impact the composite along any of the principle axes to evaluate the
mechanical behaviour (Wierzbicki , 1983; Zhang and Ashby, 1992), in the current study,
the impact comes in x2 direction which represents an in-plane impact in relation to the
honeycomb cells for an intended application.
(a) An overall view (b) Loading on a cell
Figure 5-1 Schematic diagram for in-plane low velocity impact test
110
5.1.2 The set-up of the low velocity impact instrument
A photograph of the drop weight impact instrument (the dropping hammer system) is
shown in Figure 5-2 with the major components indicated. The relationship between the
impactor and the anvil is shown in Figure 5-3.
Figure 5-2 Dropping hammer system for impact test of specimens
The major parts of this experimental set-up pictured in Figure 5-2 include:
111
a. Impactor and accelerometer: a hardwood tup (a blunt wood cylinder, 30mm diameter), a
steel holder which holds the hardwood tup and an accelerometer embedded in the
impactor, which links to the amplifier via a wire. The mass of the impactor is 0.55kg.
b. Steel tube: to provide a guidance track for the impactor sliding inside smoothly and
ensure that the impactor strikes in the upright position. .
c. Force transducer: to detect and measure the transmitted force during the experiment. It
is embedded in the anvil under the specimen. The collected signal will be send to the
charge amplifier and then to the data recorder via another data wire.
Figure 5-3 The impactor and the anvil
d. Charge amplifier: two identical charge amplifiers were used to detect and amplify the
experimental signals in volts and within a proper range (Instrument model: KISTLER,
Type 5009). Signals from the charge amplifier were recorded and a relative interval of
time (5μs) was given by the high speed data recorder (Nicolet 500) and a computer with
the appropriate software (Nicolet Window) was set up to record and display the data
(Figure 5-4). Channel 1 was used to acquire data for the transmitted force, and channel
2 was set up to collect the acceleration data.
112
Figure 5-4 Charge amplifier used in the tests
e. High-speed data recorder (Nicolet 500) and computer: used together with software
(Nicolet Windows) for data recording and processing, as well as displaying.
Figure 5-5 Snapshot of the resultant curves for force and acceleration displayed in
Nicolet Windows
113
5.1.3 Test procedure
All kinds of specimens are trimmed into the size of approximate 60mm × 120mm. This
is determined by the dimension of the anvil and the diameter of the impactor, which is
30mm.
The test procedures are summarised as follows:
 Pre-test and make sure that everything is ready: this includes the settings of the charge
amplifiers and Nicolet Windows;
 Place the specimen correctly: the specimen should be kept tight on the anvil to keep the
experimental conditions of each specimen identical and comparable with each other;
 Measure and mark the distance of tub falling height: the height of the tub is determined
according to the desired velocity which is approximate 5.5m/s when the impactor hits
the specimen surface. It is 1.62m from the releasing position to the anvil surface during
the experiment. Therefore, the falling height which represents the distance from
releasing position to specimen surface will be around 1.54m subtracting the height of
the specimen.
 Release the thread: The raising/releasing thread was bundled together with the signal
wire for the acceleration. When releasing the thread, special care needs to be taken for
the wire to avoid potential damage, which transfers the detected signal of acceleration
from the accelerometer to the charge amplifier. The wire travels at the same speed as the
impactor, and its catching with the edge of the steel tube may cause damage to the wire.
 Check and save the recorded data: in two separate channels, the voltage signals from
two charge amplifiers are converted and displayed on the computer screen using Nicolet
Windows. Channel 1 was set for transmitted force and channel 2 was set for impactor
acceleration. The voltage values can be saved to computer separately. The real force and
acceleration magnitudes should then be calculated from these voltage values,
multiplying by the scale factors of the charge amplifiers previously described.
114
5.2 Preparation for Test
5.2.1 Specimens of textile honeycomb composites
The production of textile honeycomb composites and their group division has been
explained in the Chapter 4. Fourteen different textile honeycomb composites with
various geometric parameters have been created in the present study. The composites
are with the following variations:
 Different opening angle (8L6P30, 8L6P45, 8L6P60, 8L6P75, 8L6P90)

 Different cell size (8L3P, 8L4P, 8L5P, 8L6P)
 Different free wall to bonded wall ratio:
lb
 1 : (8L3P, 8L(4+3)P, 8L(6+3)P)
lf
lb
 1 : (8L6P, 8L(4+6)P, 8L(3+6)P)
lf
 Different cell size with same thickness in total (4L6P, 6L4P, 8L3P)
5.2.2 Impact setting for the dropping hammer system (v0 =5.5m/s)
The current research is developed from a previous project (Yu and Chen, 2006) on
textile honeycomb composite materials for the riot police as limb protectors. In this
research, Yu and Chen specified that the projectile impact velocity is 5.5m/s and the
projectile mass is 0.55kg, resulting in the impact energy of 8.3J.
To match their work, in the present investigation, the impactor weighing 0.55kg was
positioned on a rail at a height of 1.54m above the top surface of the specimens. The
impact velocity of the impactor is therefore 5.5m/s, thus the impact energy is 8.3J. At
the same time, the signals were set to be recorded in every 5μs interval.
115
5.3 Impact Test Results
5.3.1 Data processing
The data processing method was introduced in this section, taking 8L6P60 for example.
To help with a clear illustration, a data processing flow chart was drawn and shown in
Figure 5-6.
Raw Data including Transmitted Force – Time Diagram and

Acceleration – Time Diagram:
a
F
t
t
Zero Resetting:
(Peak transmitted force and its arrival time can be retrieved from
Transmitted Force – Time diagram at this stage. The peak transmitted
force attenuation factor can be calculated.)
F a
t
: t
116
Velocity Calculation Based on Acceleration Integration:
a v
t
t
Contact Force equals Acceleration multiplied by Mass:
a Fc
t t
Displacement Calculation Based on Velocity Integration:

v s
t t
Calculation of Energy Absorption in Vertical Deformation from

Integration of Contact Force – Displacement Diagram:
Fc E
s
t
Figure 5-6. Data processing flow chart for experimental data analysis procedures
117
During the impact tests, there are a few experimental factors which influence the test
accuracy such as the irregular shape of honeycomb structure of some specimens, uneven
resin coating on different sides of the specimens, the disturbance during the impactor
releasing and the friction of the tube track to the impactor. Such listed factors may result
in differences of initial impact velocity when the impactor hits the specimen top surface
in magnitude and direction. Therefore, although about 10 specimens were tested for
each type of composite, to improve the data accuracy, only results from three tests with
the most repeatability were selected for further data processing.
In the following detailed data processing procedures, introductions will be put

forwarded on basic principles for low-velocity impact test, force attenuation,
acceleration of the impactor, characteristics of the transmitted force and energy
absorption performance of the textile honeycomb composites.
5.3.1.1 Basics for low-velocity impact test
For impact tests based on the drop-weight principle, the impact energy depends on the
mass of the impactor assembly m and the height of the impactor assembly h over the
specimen. The impact energy K is express as:
K=mgh [5-2]
where, g is the gravitational acceleration of 9.8 m/s2.
The velocity of impact when impactor head first touches the specimen may be
expressed as:
v0 = 2 gh [5-3]
where, h is the height from which the impactor starts dropping onto the top surface of
the specimen, v0 is the initial impact velocity when the impactor first hits the top surface
of the specimen, and g is the gravity acceleration of 9.8 m/s2.
118
It is assumed that there is no energy loss during the fall of the impactor assembly. The
accelerometer detects the changes in acceleration during the impact process, which
reflects the changes in the normal contact force between the impactor and the specimen.
It also describes the movement of the impactor too. Energy absorption by the specimen
can be obtained from the specimen deformation. The ratio of energy absorbed by the
specimen to the impact energy can be used as a measure of the specimen’s capability for
energy absorption.
5.3.1.2 Force attenuation
In a typical impact on a honeycomb composite, the impact energy is absorbed by the

elastic and plastic deformation as well as the collapse of the cells (Gibson and Ashby,
1997). The impact energy that is not absorbed by the honeycomb composite may cause
damage to materials and structures beneath the composite by exerting a transmitted
force downwards.
The transmitted force can be detected by a load cell embedded in the anvil. Compare the
transmitted force with the normal impact force acting on the anvil directly, without the
involvement of the specimen, leads to the definition of impact force attenuation.
The attenuation factor (fatt) is used to demonstrate the force-blocking effectiveness of

the specimen, and it is defined as (Dionne et al., 2003):
 F 
f att  1  trans   100 (%) [5-4]
 F 
where, Ftrans is the transmitted force through the specimen and F is the impact force
acting directly on the anvil. A value of 100% for the attenuation factor corresponds to
no force being transmitted underneath and a value of 0% indicates that all the force has
been transmitted.
119
Therefore according to the maximum transmitted force obtained by the impact testing
with and without specimen, the force attenuation can be calculated by using the
maximum transmitted force as the Ftrans (with sample) and F (without sample). Research
report from Med-Eng System Inc.(2001) shows that fatt value of many commonly used
material for blunt impactor are within 20%-30%, however, the value for all textile based
honeycomb composites are above 90%.
In the current dropping hammer system, the mechanical test without specimen with the
same experimental setup had obtained 17.5KN as F (without sample) and it will be used
to calculate fatt for all textile honeycomb composites.
Take 8L3P60 for example as listed in Table 5-3, fatt is calculated as:
 0.95 
f att  1    100  94.2 (%)
 17.5 
This indicates that during the low velocity impact in the current experiment, about 94.2%
impact force was attenuated by the 8L3P60 while only 5.8% impact force has been
transmitted to the anvil. This effectiveness shows the potential of such textile
honeycomb composite in PPE industries.
5.3.1.3 Acceleration of the impactor
The impactor experiences a deceleration when it strikes the specimen; the deformation
and collapse of the honeycomb cells contribute to the overall deformation of the
honeycomb composites as a whole. The deceleration curves from three impact tests on
the 8L6P60 together with the averaged curve are shown in Figure 5-7.
120
Figure 5-7 Measured acceleration curves for 8L6P60
It can be seen that the deceleration reached its peak about 7ms during the impact and the
impact process completed within about 20ms. The fluctuation in the curves is believed
to have come from the deformation and collapse of the honeycomb cells.
5.3.1.4 Characteristics of the transmitted force
The load cell embedded in the anvil detects and picks up the force signal transmitted
through the specimen. The magnitude of the transmitted force perceived from beneath
the material, in this case the honeycomb composites, is an important indicator for
protective capability of the honeycomb composite. Figure 5-8 takes 8L6P60 as an
example and retrieved the transmitted force curve from three impact tests. The
transmitted force was measured at an interval of 5μs.
121
Figure 5-8 Measured transmitted force curves for 8L6P60
The measured transmitted force curve is important as it shows the peak transmitted
force and the peak transmitted force strike time. To be considered as material in PPE
design, the peak transmitted force should be designed lower than the threshold force of
chin and knees to avoid damage to wearer’s bones. The peak transmitted force arrival
time is another important factor as the later the peak transmitted force strikes, the more
is the reaction time for the wearer to escape from the attack.
5.3.1.5 Energy absorption performance
Velocity v and the displacement y for the impactor can be derived by integrating the
acceleration curve once and twice, respectively, leading to
v  2 gh   adt [5-6]
and,
y  2 gh  T0   ( adt)dt [5-7]
where T0 is the duration of the impact.
122
The impact force, also referred to as the contact force, can be calculated using Newton’s
second law of motion and it is expressed as:
Fcontact  M  a [5-8]
where M is the mass of the impactor (0.55kg in current study) and a is the measured
deceleration.
C ontact Force
0.4
0.3
Contact Force(KN)
0.2
0.1
0
0 5 10 15 20 25 30 35 40
-0.1
Di spl ace me nt(mm)
Figure 5-9 The response of contact force against displacement
Figure 5-9 shows the contact force-displacement curve for a typical sampled composite,
where the area under the curve gives the energy absorption (E):
E   Fcontact dy [5-9]
Energy absorption can be calculated by integrating the closing area of contact force-
displacement curve in Figure 5-9. In details, the trapezoidal method was used for the
123
numerical integration in calculating the energy absorption and it can be schematically
interpreted as Figure 5-10:
Figure 5-10 Trapezoidal method to calculate the energy absorption
By dividing the closed area between contact force - displacement curve and the x-axis
(Figure 5-10) to infinite small trapezoidal, every single interval area (equals to every
interval energy absorption) can be calculated and the sum of all will give the area of the
closed curve (that is also the total energy absorption deformed by the specimen).
1 n
E= S i =  ( Fi  Fi 1 )( yi 1  yi )
2 i 0
[5-10]
where E is the absorbed strain energy, Si is the trapezoidal area, Fi is the contact force
applied on the specimen which can be calculated from Equation 5-10 and yi is the
displacement increment at each time interval of 5μs caused by the impact force. The
absorbed strain energy is calculated as the integration of contact force multiples
displacement increment at each time interval.
124
Adjusting the units appropriately, the temporal evaluation curves for velocity,
displacement and the energy absorption in vertical deformation for each composite type
can be plotted together as that for 8L6P60 in Figure 5-11. As a general feature
description, although the velocity of the impactor is reduced due to the resistance from
the specimen, the impactor keeps going deeper into the specimen and the energy is
therefore absorbed by the specimen deformed vertically.
10
9
Velocity
Velocity (m/s), Displacement
8
7 Displacement
(cm), Energy (J)
6 Energy
5
4
3
2
1
0
-1 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04
Time (s)
Figure 5-11 Evaluation curves of velocity, displacement and energy absorption for
8L6P60
125
5.4 Experiment Results
5.4.1 Various experiment results during impact procedure
According to the travel height of the specimen (h), the initial velocity (vo) equals to
2 gh and the results are shown in the Table 5-1. By comparing the absorbed strain
energy with the kinetic energy, the energy absorption ratio can also be easily calculated.
Table 5-1 Experiment results from impact test
Sample Density Thickness h vo K E E/K

3
(g/cm ) (mm) (mm) (m/s) (J)) (J) (%)
4L6P60 0.081 29.2 1582 5.57 8.53 6.31 73.90
6L4P60 0.118 30.4 1589 5.58 8.56 7.18 83.94
8L3P60 0.155 31.6 1579 5.56 8.50 7.46 87.78
8L4P60 0.099 40.4 1568 5.54 8.44 7.78 92.13
8L5P60 0.101 49.2 1566 5.54 8.44 7.83 92.73
8L6P60 0.072 58.0 1558 5.53 8.41 7.54 89.69
8L6P30 0.115 35.7 1580 5.56 8.50 8.41 98.94
8L6P45 0.083 48.3 1568 5.54 8.44 8.30 98.37
8L6P75 0.085 64.1 1556 5.52 8.38 7.33 87.45
8L6P90 0.073 66.2 1545 5.50 8.32 7.05 84.69
8L(4+3)P60 0.121 31.6 1579 5.56 8.50 7.33 86.26
8L(6+3)P60 0.112 31.6 1580 5.56 8.50 7.28 86.26
8L(3+6)P60 0.089 35.7 1564 5.54 8.44 8.21 97.27
8L(4+6)P60 0.086 35.7 1555 5.52 8.38 7.17 86.05
In this table, E is the strain energy being absorbed because of the structure deformation,
and E/K is the percentage of absorbed energy divided by kinetic energy, h is the travel
height of the impactor and vo is the initial velocity when the impactor hits the sample, K
1
means the kinetic energy of the impactor which can be calculated by mv 02 .
2
126
5.4.2 Experiment results for energy absorption
The energy absorption results compared to the potential gravity energy which can be
calculated by using mgh (where g: gravity acceleration) without specimen are shown in
Table 5-2. From the results, it can be seen that 8L6P30 (E/K1=96.89%), 8L6P45
(E/K1=95.65%), 8L(3+6)P60 (E/K1=94.58%) shows a good performance of energy
absorption.
Table 5-2 Experiment results for the energy absorption
Sample h1 K1 E E/K1
(mm) (J) (J) (%)
4L6P60 1610 8.68 6.31 72.70
6L4P60 1610 8.68 7.18 82.78
8L3P60 1610 8.68 7.46 85.96
8L4P60 1610 8.68 7.78 89.58
8L5P60 1610 8.68 7.83 90.17
8L6P60 1610 8.68 7.54 86.90
8L6P30 1610 8.68 8.41 96.89
8L6P45 1610 8.68 8.30 95.65
8L6P75 1610 8.68 7.33 84.43
8L6P90 1610 8.68 7.05 81.18
8L(4+3)P60 1610 8.68 7.33 84.47
8L(6+3)P60 1610 8.68 7.28 83.88
8L(3+6)P60 1610 8.68 8.21 94.58
8L(4+6)P60 1610 8.68 7.17 82.58
Where in the table, h1 is the travel height of the impactor without sample, K1 is the
kinetic energy when the impactor hit the anvil without sample, E is the absorbed energy
by the structure deformation and E/K1 is the percentage of the absorbed energy divided
by kinetic energy (without sample).
127
5.4.3 Experiment results for force attenuation factor (fatt )
From the force attenuation point of view in Table 5-3, 8L6P60 (fatt=97.9%), 8L6P75
(fatt=98.3%), 8L6P90 (fatt=98.4%), 8L(4+6)P60 (fatt = 97.6%) illustrates a very good
force protection. A possible reason for causing this phenomenon is that the thickness of
above sample being mentioned are relatively higher than the rest of the samples (See
Table 5-1), which decrease the transmitted force attenuate underneath the specimen and
provide it a longer peak arrival time.
Table 5-3 Experiment results for force attenuation
Sample t Ftrans F fatt

(ms) (KN) (KN) (%)
4L6P60 6.94 1.27 17.5 92.3
6L4P60 3.05 0.50 17.5 97.0
8L3P60 3.52 0.95 17.5 94.2
8L4P60 3.15 0.60 17.5 96.3
8L5P60 7.18 0.55 17.5 96.7
8L6P30 4.24 0.49 17.5 97.0
8L6P45 7.35 0.42 17.5 97.5
8L6P60 4.91 0.35 17.5 97.9
8L6P75 7.69 0.29 17.5 98.3
8L6P90 6.48 0.27 17.5 98.4
8L(4+3)P60 3.97 0.77 17.5 95.3
8L(6+3)P60 4.65 0.66 17.5 96.0
8L(3+6)P60 4.47 0.40 17.5 97.6
8L(4+6)P60 7.46 0.40 17.5 97.6
Where in this table, Ftrans is the peak transmitted force which has been detected by the
force transducer in the dropping hammer system, t is the peak transmitted force arrival
128
time, F is the peak transmitted force being captured during the impact without specimen,
and fatt is the attenuation factor calculated by Equation 5-4.
5.5 Structure and Properties of Textile Honeycomb Composites

5.5.1 Structure parameters and performance indices
As illustrated in the four comparable groups, the structural parameters investigated

include the cell size at the same number of layers, the opening angle of cell, the ratio of
lb
walls lengths ( ) and the specimen with same thickness but different volume density.
lf
The parameters are used to categorize the composite types into four groups so that each
group has only one changing parameter.
Two most important performance indices were selected to describe the effectiveness of
the textile honeycomb composites in the following data analysis and they are the peak
transmitted force and the energy absorption performance. The peak transmitted force
performance is defined as the maximum value of the transmitted force detected by the
force transducer. The energy absorption performance is defined as the energy absorbed
by honeycomb structure deformation. Due to the existence of the fluctuations in
measured data profiles, the two selected indices are with better reliability and more
direct physical meaning, to describe the effectiveness of textile honeycomb composites.
5.5.2 Grouped sample experimental performance

5.5.2.1 Cell size and its experimental performance (8L3P60, 8L4P60, 8L5P60, 8L6P60)
All honeycomb composites in this group are made from eight layers of fabrics involving
four regular hexagonal cells, where the six walls of each cell have the same length. The
opening angle of the cells in these composites is 60°. With the same weft density for all
reinforcing fabric sections, the cell wall length changes from three, four, five, to six
picks, resulting in cells with increasing sizes. These composites are 8L3P60, 8L4P60,
8L5P60, and 8L6P60 among which 8L3P60 is the thinnest and 8L6P60 is the thickest.
129
The transmitted force performance and the energy absorption performance are
illustrated in Figure 5-12 and Figure 5-14. The peak transmitted force value and the
energy absorption value have been demonstrated in Figure 5-13 and Figure 5-15
respectively.
1
8L3P60
0.9
8L4P60
0.8 8L5P60
0.7 8L6P60
Transmitted Force (kN)
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.005 0.01 0.015 0.02 0.025
Time (s)
Figure 5-12 Comparison of transmitted force – time diagram (samples with different
cell size)
1 0.95
0.9
Peak Transmitted Force (KN)
0.8
0.7
0.6
0.6 0.55
0.5
0.4 0.35
0.3
0.2
0.1
0
8L3P60 8L4P60 8L5P60 8L6P60
Figure 5-13 Comparison of peak transmitted force value (samples with different cell
size)
130
0.9
8L3P60
0.8
8L4P60
0.7 8L5P60
8L6P60
0.6
Contact Force (kN)
0.5
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Displacement (cm)
Figure 5-14 Comparison of contact force – displacement diagram (samples with

different cell size)
Energy Absorption Displacement(%)

10 80
75.86
9 70
Structure Displacement(%)
8
Energy Absorption(J)
60
7 57.50
46.88 57.14 50
6
5 40
4 30
3
7.46 20
2 7.78 7.83 7.54
1 10
0 0
8L3P60 8L4P60 8L5P60 8L6P60
Figure 5-15 Comparison of energy absorption and structure displacement diagram

(samples with different cell size)
131
Regarding the transmitted force-time curve of 8L3P60 and 8L6P60 in Figure 5-12 and
Figure 5-13, it is noticed that 8L3P60 has a higher transmitted force that reaches
0.95KN, however, the whole deformation time is quite short, only about 8ms. It shows a
high peak transmitted force instantly at 3.5ms. In contrast, 8L6P60 has a longer
deformation time (21ms) but the peak transmitted force is quite lower about 0.35KN.
Thus it shows much lower peak transmitted force and cell collapsed with a longer stroke.
Impact force attenuation is dependent on the cell size of honeycomb composites as

demonstrated in Figure 5-12. It is clear that honeycomb composites with larger cells
perform better; the altitude of the peak transmitted force is much reduced and the curve
is smoother. It is also evident that the maximal transmitted force occurs later, in general,
as the cell size becomes larger. This is a favourable property for material intended for
body and limb protection against trauma impact; composites with larger cells allow
more time for the human body to react to impact, hence reducing the risk of more
serious injuries.
The result in contact force-displacement curve in Figure 5-14 illustrated that 8L3P60
experienced less than half collapses of the cell and a higher loading force comparing to
samples with big cell (8L5P and 8L6P). However, with the similar initial impact
velocity, 8L6P60 displayed a lower loading force and stroke at a longer distance
collapse. It can be explained that with the increase of the cell size, the maximum impact
distance is increased too as the thickness of the specimen changed respectively
according to the cell size. Therefore, the larger the single cell, the deeper the specimen
can be impacted with. However, results from Figure 5-14 and Figure 5-15 shows the
energy absorption among 8L3P60(E=7.46J), 8L4P60(E=7.78J),, 8L5P60(E=7.83J),
8L6P60(E=7.54J) are quite similar. Hence, changes of the cell size are not a major
factor that will affect the energy absorption performance of the specimen.
It is seen from the contact force-displacement curves in Figure 5-14 that composites
with smaller cell sizes have high impact modulus and conversely those with larger cell
sizes have low impact modulus thus it leads to harder and softer composite materials
property. This information suggests that for honeycomb composites with the same
132
number of cells in a column, cell sizes can be used as the key parameter for altering the
softness of the composite material.
Therefore, with the consideration of the material property of 8L3P60 and 8L6P60,
8L3P60 has smaller cells that are very rigid on handling whereas 8L6P60 is much softer
on handling but much bulkier in volume. As the energy absorption among 8L3P60,
8L4P60, 8L5P60, 8L6P60 are similar under the impact, the physical performance of
transmitted force and deformation time became the major factor to judge the protection
ability of the specimen. It seems the softer the textile honeycomb sample, the less
transmitted force will be encountered (as 8L6P60 is very soft in handing), and it takes
longer deformation time for the whole process. Respectively, rigid sample (8L3P60)
will have a higher transmitted force and shorter deformation time, but it is less bulky.
Figure 5-15 also shows the altitude of energy absorption of samples with different cell
size and their structure displacement. It seems that the energy absorption is similar for
8L3P60, 8L4P60, 8L5P60 and 8L6P60. However, the structure displacement for
8L6P60 is much deeper than the rest of the samples. This indicates that there is a larger
structure deformation for 8L6P60 while the strain energy absorption has not been
significantly increased. The reason to explain it might be that the thickness of 8L6P60 is
higher than the rest of samples, therefore, it provide more spaces for the impactor to
strike through which leads to a deeper structure displacement vertically.
With reference to the transmitted force and energy absorption, composites 8L3P60 and
8L6P60 absorb similar amount of energy, but their peak transmitted force and striking
time are significantly different. 8L3P60 is associated with the higher transmitted forces
and it has the smaller thickness and is more rigid. By contrast, 8L6P60 is much bulkier
and softer as a material.
Ideally, the honeycomb textile composites with softer handling and less bulk are sought
for improving the protection ability of the impact force. Structure such as 8L3P60 who
holds smaller cell size therefore its bulkiness is ideal in the PPE application. Even better
is that if there is a similar composite whose handling property is less rigidity than that
133
of 8L3P60 which will be more suitable to be used for protection purpose because the
softer the composite material, the more striking time it will take for the structure to
react against outside attack. And this will give more timing for the human being to be
prepared against impact from outside.
5.5.2.2 Opening angle and its experimental performance (8L6P30, 8L6P45, 8L6P60,
8L6P75, 8L6P90)
8L6P30, 8L6P45, 8L6P60, 8L6P75 and 8L6P90 are grouped to investigate the
difference of opening angle which intend to affect the mechanical performance of the
textile honeycomb composites. These composites are made from the same fabric as
reinforcement but have different opening angles, as indicated by the last two digits in
the codes.
Generally, the deformation timing of the specimens with different angles is quite similar
and this can be seen in the transmitted force-time curve in Figure 5-16. However, the
specimens with smaller opening angles (8L6P30, 8L6P45) own a higher transmitted
force at 0.49KN (8L6P30) and 0.42KN (8L6P45) comparing to the rest of the specimens.
It also demonstrates that the trend of transmitted force-time curve for 8L6P60 and
8L6P75 are very similar and the peak transmitted force value is closely matched too.
However, when the opening angle is more than 75º, the peak transmitted force is
decreasing to 0.29KN (8L6P45) which has been shown in Figure 5-17.
Figure 5-16 also show that whilst the peak transmitted force decreases as the composite
opening angle increases, it is clear that the peak forces tend to be smoothed when the
composites get thicker, due to the enlargement of opening angle. Referring to the
maximal transmitted force in Figure 5-17 for the five honeycomb composites in this
group, it demonstrates a clear correlation between increasing opening angle and the
decreasing maximal transmitted force. The thickness of the composites is believed to
have an important role to play in this together with the cell geometry in the cross-
section of the composites.
134
From the data analysis, it seems that if the specimen’s opening angle is less than 60°,
the cell will have a much higher transmitted force leading to a poor protection
performance. However, when the cell opening angle is designed between 60°and 75°,
there is a smaller transmitted force coming through and the height of the composite is
medium heighted. Therefore, if using the honeycomb composites as PPE, cell opening
angle between 60°and 75°will be the first choice when good protection is required.
According to the contact force-displacement curve in Figure 5-18, it is illustrated that

when the impactor reaches the maximum impact distance, 8L6P30 and 8L6P90 has a
reduced collapse distance while 8L6P45, 8L6P60 and 8L6P75 have a longer collapse
distance. However, considering the thickness of specimen 8L6P30 (31.6mm) is less than
half of that of 8L6P90 (66.2mm), the relative collapse distance for 8L6P90 is very small.
This is due to when cell-opening angle reaches 90°, the cell free wall and bonded wall
forms a shape of rectangular, so that the cell will be less flexible. Looking at the contact
force-displacement curve between all the specimen, 8L6P60 and 8L6P75 are much
flatter and lower than rest of the samples in the same group, which means opening angle
between 60°and 75°causes a lower loading force during the impact testing.
Generally, there is no obvious trend can be seen the contact force-time curves to
indicate the opening angle as a major mechanism by which the impact energy was
absorbed. Therefore, other structural parameters such as the thickness of the honeycomb
composites must have played important roles, too.
135
0.6
0.5
0.4
8L6P30
8L6P45
8L6P60
0.3
8L6P75
8L6P90
0.2
0.1
0
0 0.005 0.01 0.015 0.02 0.025 0.03
Time (s)
opening angle)
0.6
0.49
0.5
Peak Transmitted Force(KN)
0.42
0.4
0.35
0.29 0.27
0.3
0.2
0.1
0
8L6P30 8L6P45 8L6P60 8L6P75 8L6P90
Figure 5-17 Comparison of peak transmitted force value diagram (samples with
136
0.45
8L6P30
0.4 8L6P45
8L6P60
0.35 8L6P75
8L6P90
0.3
Contact Force (kN)
0.25
0.2
0.15
0.1
0.05
0
0 1 2 3 4 5 6
Displacement (cm)
Figure 5-18 Comparison of contact force-displacement diagram (samples with different

opening angle)
Energy Absorption Displacement(%)
10 120
9 99.17 100.63 99.66 98.44
100
8
7
80
6
60.91
5 60
4 8.41 8.3 7.54 7.33 7.05
40
3
2
20
1
0 0
8L6P30 8L6P45 8L6P60 8L6P75 8L6P90
Figure 5-19 Comparison of energy absorption and structure displacement diagram

(samples with different opening angle)
137
It is illustrated in Figure 5-19 that the amount of energy absorbed among specimens
with different angles show the trend that with the increase of the cell opening angle, the
energy dissipated inside the honeycomb structure start to reduce accordingly. This is
believed to be because the textile honeycomb composites with smaller opening angles
present lower resistance to the bending of its angled cell walls and therefore create
larger deformation.
In Figure 5-19, the structure displacement ratio for the specimen with the smallest angle
(8L6P30) is the largest, accordingly, the energy absorption for 8L6P30 is the highest too.
Opening angle exceeding 45° can stops the impactor going further deeper with
reasonable a energy absorption behaviour. However, if the opening angle continues
enlarging and exceeding 75°, the striking distance goes shallow, accordingly, the energy
absorption reduces significantly.
As mentioned in Section 5.3.1.5, the energy absorption was obtained by integrating the
area underneath contact force-displacement curve and this can be treated as the strain
energy absorbed along vertical direction. However, the total kinetic energy should all
dissipate during the impact process and therefore the rest of the energy was dissipated in
other forms. Taking the value of energy dissipated in the vertical direction and in other
forms in Table 5-4 and Figure 5-20 as reference, it is believed that 8L6P30 can absorb
the most energy in the vertical direction and only 0.09J energy was dissipated in the
other forms, while 8L6P90 can only absorb 7.05J energy in the direction of vertical and
the rest of them (1.27J) were transmitted in the other forms. In another words, the
format of the energy absorption changes when the cell opening angle changes and it
brings the trend that more energy was dissipated in other forms rather than strain energy
along vertical direction when the cell opening angle is enlarged. However, in term of
using the honeycomb structured composites as PPE, it will be better if more impact
energy is absorbed as strain energy due to the deformation of the composite structure,
138
therefore, if less energy is absorption in other forms, it will be better for the protection
purpose.
Table 5-4 Results of the energy dissipated along vertical and in other forms
Sample K E1 E2
(J) (J) (J)
8L6P30 8.50 8.41 0.09
8L6P45 8.44 8.30 0.14
8L6P60 8.41 7.54 0.87
8L6P75 8.38 7.33 1.05
8L6P90 8.32 7.05 1.27
1 2
where, K is the impactor kinetic energy which is calculated by mv 0 , E1 is the strain
2
energy being absorbed along the vertical direction and E2 is the strain energy being
absorbed in other forms. E2 is calculated using E1 subtract by K.
Energy Dissipation Format
9 0.09 0.14
Other forms of
8 0.87 1.05 1.2 energy
Absorbed Energy (J)
7
6 Vertical
5
4 8.41 8.3
7.54 7.33 7.05
3
2
1
0
8L6P30 8L6P45 8L6P60 8L6P75 8L6P90
Figure 5-20 Energy dissipation direction diagram (samples with different opening angle)
139
lb
5.5.2.3 Length ratio of cell walls and its experiment performance (  1:
lf
lb
8L3P60,8L(4+3)P60, 8L(6+3)P60;  1 : 8L(3+6) P60,8L(4+6)P60, 8L6P60)
lf
Two subgroups of textile honeycomb composites, based on the eight-layer construction,

have been created for the study of influence of cell wall length ratio on their impact
performance. The first subgroup involves 8L3P60, 8L(4+3)P60, and 8L(6+3)P60 where
the free wall length is kept at three picks and the bonded wall length takes the values of
three, four and six picks. In the second subgroup involves 8L(3+6)P60, 8L(4+6)P60,
and 8L6P60, the free wall length is constant at six picks while the length of the bonded
walls changes from three, four, to six picks. The cell opening angle for all the
composites is 60°.
The results from statistic data analysis are summarized in Table 5-5. The change of the
lb l
cell wall length ratio between the case for  1 and b  1 are plotted according to
lf lf
their peak transmitted force in Figure 5-21 and energy absorption performance in Figure
5-22.
140
Table 5-5 Results for samples with different cell wall length ratio
Sample θ lb Ftrans fatt E t
lf
(°) (KN) (%) (J) (ms)
lb
1
lf
8L3P 60 1 0.95 94.20 7.46 3.52

8L(4+3)P 60 3:2 0.77 95.3 7.33 3.97
8L(6+3)P 60 2 0.66 96.0 7.28 4.65
lb
1
lf
8L(3+6)P 60 1:2 0.40 97.6 8.21 4.47

8L(4+6)P 60 2:3 0.40 97.6 7.17 4.46
8L6P 60 1 0.35 97.9 7.54 4.91
where, θ is the opening angle of the composite cell following that lb represent the
lf
length ratio of bonded and free wall. Ftrans is the peak transmitted force accumulated
during the dynamic impact and fatt is the force attenuation factor which has been
specified in Equation 5-4. E is the absorbed strain energy in vertical direction and t is
the striking time when the peak transmitted force arrived.
141
1
8L(3+6)P60
0.9 8L(4+6)P60
8L6P60
0.8 8L3P60
8L(4+3)P60
8L(6+3)P60
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.005 0.01 0.015 0.02 0.025
Time (s)
length ratio of cell walls)
In Figure 5-21, considering the free wall length (lf) of the specimen, when it is
constructed by 6 picks: 8L(3+6)P, 8L(4+6)P and 8L6P, the curve goes flat and long
which means the specimen has encountered a low transmitted force and a long stroke
time, however, the difference of the peak transmitted force among these three samples is
not so significant. This kind of results revealed that the free wall length (lf) has played a
significant role in reducing transmitted force, and increasing the lb can also leads to the
lf
reduction of transmitted force. Previous research work has explained this phenomenon
with the reason that free walls are mainly against the vertical loads on the top of the
structure and it helps dissipating the force across the cell network more efficiently, thus
much more deformation appears than that of the horizontal cell wall (Tan and Chen,
2005).
142
lb
However, in Figure 5-21 it is also show that when  1 , the specimen will transfer a
lf
high transmitted force and a short stroke time with composite cell free wall (lf) formed
by 3 picks of 8L3P, 8L(4+3)P and 8L(6+3)P. It is also demonstrated that more force has
lb
been attenuated with a shorter bonded wall sample when  1 . It is indicated that
lf
under the circumstance of that bonded wall (lb) longer than free wall (lf), changes of the
length of lb does influence the mechanical performance of the textile honeycomb
composite and to sum up that shorter bonded wall increases the transmitted force
correspondingly. The reason to cause the variety of the transmitted force may be due to
the bending behaviour of the bonded wall after being impacted along in-plane direction.
Practically, this kind of phenomena could be used when designing the textile
honeycomb textile composite for the protection purpose that the longer the bonded wall
(or the higher ratio of lb ) can generate lower transmitted force.

lf
Figure 5-22 presents the contact force-displacement curves of the samples under the
impact velocity at about 5.5m/s. Figure 5-23 illustrates the comparison of the energy
lb
absorption value and vertical displacement ratio between subgroup  1 and subgroup
lf
lb
1 .
lf
The results from Figure 5-22 clearly show that the two subgroups perform very
lb
differently. When  1 , the displacement of the impactor goes shallow but the contact
lf
lb
force runs very high. On the contrast that, when  1 , the curves occupies more
lf
displacement but the contact force is much lower. This lead to a conclusion that the
lb l
composite with  1 demonstrates a high impact modulus and composite with b  1
lf lf
143
lb
has a ductile behaviour. With the  1 subgroup, it is clear that 8L3P whose cell wall
lf
ratio is 1:1 shows the highest impact modulus and 8L(6+3)P, with a length ratio of 2:1,
displays the lowest impact modulus in this subgroup.
lb
For the subgroup with  1 , a ductile behaviour is demonstrated for all three samples
lf
involved, with the 8L6P60 being the most ductile. This could lead to a good resistant to
the impact loads as materials with a ductile performance normally exhibits a greater
resistance to impact loads than do brittle material (James and Stephen, 2001).
Figure 5-23 shows, however, that the total amount of impact energy absorbed by the
lb l
composites of  1 is about the same while for the composites with b  1 , there are
lf lf
subtle differences in their energy absorption. There is no significant relationship

between energy absorption and their structure displacement in each subgroup and the
subtle differences in the absorption of impact energy among the composites, especially
lb
in the  1 subgroup, require further investigation.
lf
The above results can be a good instruction for the design of PPE. Instead of changing
the whole cell size, by simply modifying the length ratio of cell bonded and free wall, it
can avoid increasing the protector’s weight significantly while improve the mechanical
performance of the honeycomb composites the same time. But it is believed that there
will be a limitation of increasing this ratio ( lb ), and it means the designer cannot
lf
increase the ratio lb dramatically to achieve higher energy absorption with the same
lf
weight. The current research work can only prove that change of this ratio up to 2:1 is
still effective to improve the energy absorption capability of the textile honeycomb
composite, however, to find out the critical ratio value higher than 2:1 with optimal
energy absorption requires more experimental investigations.
144
0.9
0.8
0.7
8L(3+6)P60
8L(4+6)P60
0.6 8L6P60
8L3P60
Contact Force (kN)
8L(4+3)P60
0.5 8L(6+3)P60
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Displacement (cm)
Figure 5-22 Comparison of contact force-displacement diagram (samples with different

length ratio of cell walls)
Energy Absorption Displacement

10 120
9
97.22 95.83 100
8
7
74.14 80
71.88
6
5 56.25 60
46.88
4
40
3
2 7.46 7.33 7.28 8.21 7.17 7.54

20
1
0 0
8L3P60 8L(4+3)P60 8L(6+3)P60 8L(3+6)P60 8L(4+6)P60 8L6P60
Figure 5-23 Comparison of energy absorption diagram (samples with different length
ratio of cell walls)
145
5.5.2.4 Honeycomb composites with similar thickness and their performance (4L6P60,
6L4P60 and 8L3P60)
This group of 3D textile honeycomb composites includes 4L6P60, 6L4P60 and 8L3P60,
whose thickness is 29.2mm, 30.4mm and 31.6mm respectively, and their thinness is very
similar in practical term. However, the volume density of the honeycomb composites
are various as 4L6P<6L4P<8L3P, which are listed in Table 5-6. In this section, data
analysis will be focus on these samples who are similar in thickness but different in
their volume density. The results from experiments are summarized in Table 5-6.
Table 5-6. Experiment results (samples with similar thickness)
Sample θ (°) Density Ftrans fatt E t

3
(g/cm ) (KN) (%) (J) (ms)
4L6P 60 0.081 1.27 92.3 6.31 6.94
6L4P 60 0.118 0.50 97.0 7.18 3.05
8L3P 60 0.155 0.95 94.2 7.46 3.52
Where in this table, θ is the cell opening angle; Ftrans is the peak transmitted force and t
is the peak transmitted force arrival time, fatt is the attenuation factor and E is the
absorbed energy.
Figure 5-24 displays the transmitted force against the impact time. It is shown that
4L6P60 was crushed by the impact leading to a large transmitted force and the whole
sample was totally crushed during the test visually too. Composite 8L3P60 is associated
to a higher transmitted force than 6L4P60 because the former has a higher impact
modulus, which relates to the volume density of the composites.
The behaviour of 4L6P60 is quite abnormal comparing to the rest of the samples from
their test results. 4L6P60 is very soft in handling and the whole sample is crushed after
the impact, which allows the impactor to go through the sample and touches the anvil
146
underneath during the impact, therefore, the transmitted force is relatively higher at
1.27KN and from Table 5-3 it can be seen that the transmitted force factor (fatt) is only at
92.3% which is the lowest among all the rest of 13 samples. Besides the poor behaviour
of transmitted force, the energy absorption capability of 4L6P60 is also very limited at
only 6.31J in Table 5-6, and it is the lowest among all the tested textile honeycomb
composites. The energy absorption of uncrushed samples of 6L4P60 and 8L3P60 are
almost the same at 7.18J and 7.46J and this is about normal.
From Table 5-7, it can be seen that the volume density of the three composites in this
group is in the order of 4L6P60<6L4P60<8L3P60. In creating engineering material, low
density is one of the features sought for the material. The data analysis here would
suggest that a lightweight honeycomb composite must be combined with strong cell
walls in order for the composite to be of engineering significance. For honeycomb with
particularly low densities such as sample 4L6P60, it is more important for the cell walls
to have high strength.
147
1.6
4L6P60
1.4
6L4P60
1.2 8L3P60
0.8
0.6
0.4
0.2
0
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04
-0.2
Time (s)
Figure 5-24 Comparison of transmitted force – time diagram (samples with similar
thickness)
1.2
4L6P60
1
6L4P60
0.8 8L3P60
Contact Force (kN)
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3
Displacement (cm)
Figure 5-25 Comparison of contact force - displacement diagram (samples with similar
thickness)
148
5.5.3 Discussions on composite volume density and composite thickness
So far, analysis has been carried out for the influence of the structural parameters,
including cell opening angle, cell size, cell wall length ratio, and composite volume
density, on the impact characteristics of the textile honeycomb composites. It is
important to be able to engineer the composites with required properties by
manipulating the structural parameters. However, changes in structural parameters will
have to lead to alteration of the volume density and thickness of the honeycomb
composites. Therefore, it is necessary to see how the volume density and the thickness
of all composites would affect the impact performance, particularly the impact energy
absorption by the honeycomb composites and the transmitted force through the
honeycomb composites. The volume density and the thickness of the samples were
measured and are listed in Table 5-7.
Table 5-7. Volume density, thickness, energy absorption and peak transmitted force of
different textile honeycomb composites
Sample Density Thickness E Ftrans
(g/cm3) (mm) (J) (KN)
4L6P60 0.081 29.2 6.31 1.27
6L4P60 0.118 30.4 7.18 0.50
8L3P60 0.155 31.6 7.46 0.95
8L4P60 0.099 40.4 7.78 0.60
8L5P60 0.101 49.2 7.83 0.55
8L6P60 0.072 58.0 7.54 0.49
8L6P30 0.115 35.7 8.41 0.42
8L6P45 0.083 48.3 8.30 0.35
8L6P75 0.085 64.1 7.33 0.29
8L6P90 0.073 66.2 7.05 0.27
8L(4+3)P60 0.121 31.6 7.33 0.77
8L(6+3)P60 0.112 31.6 7.28 0.66
8L(3+6)P60 0.089 35.7 8.21 0.40
8L(4+6)P60 0.086 35.7 7.17 0.40
Where in the table, E is the absorbed energy and Ftrans means the peak transmitted force.
149
5.5.3.1 Composite volume density
The impact energy absorption for all the composites against the composite volume
density is illustrated in Figure 5-26(a), where the navy mark is for 4L6P60 which was
crushed during the impact test. No clear trend can be found in the relationship between
the composite volume density and their energy absorption. As long as the honeycomb
composites are not completely crushed, they demonstrate similar capability for impact
energy absorption. This suggests that for a given level of impact energy, it is possible to
create low density honeycomb composites for impact energy absorption. Work in this
direction will lead to materials of high energy absorption to density ratio, which is
attractive to many engineering applications. Figure 5-26(b) shows the relationship
between the transmitted force and the composite volume density. Apart from the
composite that was crushed during the test, a trend is clearly shown that for the
honeycomb composites the transmitted force increases as the volume density goes up.
As long as the composites are strong enough not to be crushed, lower density
composites will be more capable to attenuate the impact force. This prompts more work
in engineering design of honeycomb composites which are lightweight and
mechanically protective.
5.5.3.2 Composite thickness
Figure 5-27(a) shows that the thickness of the composites which are not demonstrating
obvious influences on the energy absorption performance. Similar to the case of
composite volume density, it seems that all textile honeycomb composites are capable
of absorbing similar amount of impact energy as long as the composite is not
completely crushed. However, if the energy absorption and transmitted force are
considered separately, it seems that the thickness of composites has an obvious affect on
the transmitted force, for example, the thicker the composite, the smaller the transmitted
force.
It is noticed that in Figure 5-27, the navy square mark represents composite 4L4P60
which was crushed totally in the test.
150
9
8
Energy Absorption(J) 7
6
5
4
3
2
1
0
0.05 0.07 0.09 0.11 0.13 0.15 0.17
Composite Density(g.cm-3)
(a) Energy absorption
1.4
Transmitted Force(KN)
1.2
0.8
0.6
0.4
0.2
0
0.05 0.07 0.09 0.11 0.13 0.15 0.17
Composite Density(g.cm-3)
(b) Transmitted force
Figure 5-26 Influence of volume density on honeycomb composites
151
9
7
0
25 30 35 40 45 50 55 60 65 70
Composite Thickness(mm)
(a) Energy absorption
1.4
1.2
0.8
0.6
0.4
0.2
0
25 30 35 40 45 50 55 60 65 70
Composite Thickness(mm)
(b) Transmitted force
Figure 5-27 Influence of composite thickness on honeycomb composites
152
5.6 Conclusions
In this chapter, experimental study of 14 systematically designed and manufactured 3D

textile honeycomb composites was carried out with an emphasis on their impact
behaviour. The 3D textile honeycomb composites were created from integral 3D textile
reinforcements with the required material continuity in the composites.
For 3D textile honeycomb composites, the cell geometries are critical issues for
achieving higher energy absorption capacity. For the cells designed for the experiment,
cell size, cell opening angle, length ratio of cell walls, and also the composite in the
similar thickness with different volume density were selected as variables to explore the
influence of the cell geometry on resultant impact performances. Among all of cell
geometrical parameters being investigated, length ratio of cell wall or the cell opening
angle of the cell structure is the most effective parameter for controlling the energy
absorption performance of the honeycomb textile composites.
These conclusions provide useful information for engineering textile honeycomb

composites against impact:
1). the opening angle of cells in honeycomb composites plays an important role in
determining their properties. For the same fabric, increasing the opening angle results in
a less energy absorbent and less force attenuating honeycomb composite, and vice versa.
2). an increase in cell size of honeycomb composites makes them more efficient in
impact force attenuation. It helps reducing the peak transmitted force and delays its
arrival time. In addition, reducing the cell size can lead to the least bulky engineering
material for similar impact energy absorption.
3). Manipulation of the length ratio of cell walls leads to the creation of two damping
materials with different impact behaviour in absorbing impact energy. Honeycomb
composites with length ratio of bonded wall to free wall ratio more than or equal to 1
153
lb
(  1 ) is associated with high strength. The inverse design leads to composites that
lf
have low modulus and low strength with the same level of impact. However, the total
energy absorption is not affected by the length ratio of cell walls.
4). for the same volume, low density honeycomb composite leads to low contact force
and low transmitted force. However, composites with too low density can be easily
crushed as in the case of composite 4L6P60. As long as the composite is not destroyed
by crushing, they are able to absorb a similar level of impact energy.
154
CHAPTER 6
EXPERIMENTAL DATA ANALYSIS ON TEXTILE HONEYCOMB

COMPOSITES IMPACTED WITH LARGER MASS AND LOWER
VELOCITY
Chapter 5 aims to establish the understanding of the mechanical performances of the

textile honeycomb composites under low velocity impact (v0=5.5m/s and M=0.55kg). In
this chapter, an experimental investigation is reported on mechanical behaviour of the
textile honeycomb composites under the similar impact energy but with a lower velocity
less than 2.0m/s and a bigger mass at 4.52kg. The impact energy in these cases is quite
similar (8.31J versus 8.5J), which indicates that the impact energy is similar while the
energy construction is different. The tested results will be compared with the results
obtained under dropping hammer system from chapter 5 to see how the construction of
the impact kinetic energy influences the final mechanical performances for the textile
honeycomb composites with different geometric parameters. The majority mechanical
performances discussed in this chapter will be focus on contact force and energy
absorption performances.
It has to be noted that as a reference, in terms of impact mechanism, it is reported by

Prashant and Badri (1993) that a heavier impactor will cause an overall 30-50% more
damages to the item comparing to the light weight impact with same impact energy.
Both Delfosse et al.,(1993) and Ujhashi (1993) explained this phenomenon by the fact
that for heavier mass low velocity impact, there are many small superimposed
oscillations due to the plate vibrating against the impactor during contact, therefore
more damages occurs accordingly.
155
6.1 Low Velocity Impact Test Setting by Instron Dynatup Model 8200 Drop
Weight Impact Testing Instrument
In the present study another commercial low velocity dropping weight impact
instrument has been newly bought by university of Manchester and used to conduct the
low velocity impact test. This machine is called the Instron Dynatup Model 8200 drop
weight impact test instrument and is designed for acquiring the tested data by capturing
the impact velocity and load signals which have been transmitted from the test machine
to the data system for analysis. It is a complete system consisting of a drop weight
impact test machine and a data acquisition system which provides a comprehensive load
and energy record from each test. This machine is going to be used to conduct the
impact test with a large impactor mass (4.52kg) and lower velocity (less than 2.0m/s) to
investigate how the textile honeycomb composites structure response under different
impact situations. The test results will be compared with those from dropping hammer
system which have been described in Section 5.3.
6.1.1 Assembly of Instron Dynatup Model 8200 drop weight impact testing
instrument
The Instron Dynatup Model 8200 drop weight impact test machine was used in this
research work to conduct the impact tests (v0<2.0m/s and M=4.52kg) for the textile
honeycomb composites, therefore it is important to introduce its correct assembly firstly.
The Instron Dynatup Model 8200 drop weight impact test machine is a gravity driven
test instrument that is used to test the impact characteristics of an extensive variety of
materials and components over a wide range of impact velocities. The instrument is
capable of testing at velocities up to 4.4m/s and in this research the impact velocity is
about 2.0m/s. Figure 6-1 shows the assembled instrument.
The basic assembly of the Instron Dynatup Model 8200 drop weight impact testing
machine could be described as following and they are illustrated in Figure 6-1:
156
(a). drop weight and tup (impactor): drop weight and tup provide the mass for the
impact testing. The drop weight is a framework with an empty weight of
approximately 3kg of which weights can then be added up to 13.6kg. In the current
work, the impactor mass is 4.52kg. The tup is a device that measures the force applied
to a specimen by the drop weight assembly. It consists of two parts: the tup itself,
which is a load cell for measuring force and the tup insert, which is the tool steel
component that actually strikes the specimen. It has to be noticed that the diameter for
the tup insert is 20mm which is smaller than the one (30mm) used in the previous
experiment (in Chapter5). This could bring differences to the results as the thinner
impactor intends to produce a more localised deformation into the honeycomb
composite specimens than the thicker one.
(b). velocity flag and detector: a mechanical lever is provided to manually release the
drop weight from a pre-selected drop position. This position is set by moving the
clamp frame up and down and then clamping it to the guide columns using the clamp
knobs provided. The release latch is protected against inadvertent release by a guard
over the latch.
157
Figure 6-1 Instron Dynatup Model 8200 drop weight impact testing machine
(c). drop tower framework: the framework of the test machine consists of two guide
columns and the back weldment. The guide columns are 19mm diameter chrome
plated steel rods and the drop weight assembly rides on the guide columns via holes in
its upper and lower guide blocks. The back weldment is a painted section of C-
channel and it provides rigidity and vertical stability to the drop tower.
(d). anvils: anvils are fixtures that hold the specimens during testing. Many different
styles of anvils are available to accommodate various test specifications and
techniques. The anvils sit on the table and are secured in place using standard bolts.
158
6.1.2 Testing procedure
All the specimens were trimmed into approximate size of 60mm × 120 mm as those
samples which have been tested in the dropping hammer system (see Section 5.2.1). The
testing procedure was divided into a few sections:
1. pre-test preparation: this action consists of adjusting the testing assembly including
the drop height, velocity detector, stop blocks and drop weight mass. Position the
specimen on the specimen support fixture to prepare for the test. The instrument first
needs to be set up to conduct a pre-test in order to measure the initial kinetic energy
of up to 9J.
2. performing a test: conduct the test by using the integrated software and release the
impactor in order to conduct the test. Each test needs to be done 3 times and new
samples were replaced each time.
3. capture the data and analysis: tested data were passed to the computer through A/D
digital convertor for future analysis. The Instron Dynatup Model 8200 drop weight
impact testing machine is the hardware to conduct the testing and to amplify and
capture the dynamic transducer output from a high speed impact event. All the raw
data will be interpreted through the associated software which is called Instron-
Dynatup impulse data acquisition software. After the test has been done, analyse the
data and find out its relationship among different textile honeycomb composites and
their mechanical performances.
6.1.3 Classifications of textile honeycomb composites
Same as the composites which have been tested in the previous experiment (see Section
5.2), the honeycomb textile composites are divided into four groups according to their
cell size, opening angle, length ratio of cell wall and samples with similar thickness.
However, because of the shortage of the fabric, 8L4P60 and 8L6P90 has been
eliminated from the composites which will not affect the study on the samples.
159
Therefore, there are twelve different textile honeycomb composites which have been
created and they have been listed as following:
1. Different opening angle (8L6P30, 8L6P45, 8L6P60, 8L6P75),

2. Different cell size (8L3P, 8L5P, 8L6P),
3. Different free wall and bonded wall length ratio:
lb
 1 : (8L(3+6)P, 8L(4+6)P, 8L6P),
lf
lb
 1 : (8L(6+3)P, 8L(4+3)P, 8L3P),
lf
4. Different cell size with similar thickness (4L6P60, 6L4P60, 8L3P60).
However, it is found that during the current impact tests, the samples with 2 layers and 3
layers (4L6P and 6L4P) are crushed instantly. The energy absorption and contact force
generated from crushed samples will not be comparable to those samples which haven’t
been crushed; therefore, samples of 4L6P and 6L4P are obsoleted from the late results
discussion in Section 6.2
6.1.4 Impact setting for Instron Dynatup Model 8200 system
The test was completed by using Instron Dynatup Model 8200 drop weight impact
testing machine and all of the specimens were impacted at the initial velocity around
2.0m/s with the impactor tup mass of 4.52kg to ensure the initial impact energy is
around 8.5J. A set of data including time, contact force, velocity, displacement etc.,
were captured after the impact test and finally, every single contact force value were
aligned to form the curve with the interval time as x-axis. Additionally, every single
contact force value was also aligned with the displacement to evaluate the contact force
and energy absorption capacity.
160
6.2 Tested Result and Discussion
6.2.1 Cell size and its experimental performance (8L3P60, 8L5P60, 8L6P60)
Samples with different cell size from small to big (8L3P60, 8L5P60 and 8L6P60) are
constructed to investigate how the mechanical performance achieves under the similar
impact energy around 8.5J with a larger mass (M=4.52kg) and lower velocity
(v0<2.0m/s) impact. Table 6-1 has listed some of the experiment results after the impact.
Table 6-1 Experiment results from impact (samples with different cell size)
Composite θ Fmax E1 Smax Tpeak vo E2 E1

(%)
Type (°) (KN) (J) (mm) (ms) (m/s) (J) E2
8L3P 60 0.8 7.3 14.2 15.3 1.8 7.7 95.0

8L5P 60 0.4 7.8 37.2 8.8 1.9 8.2 95.8
8L6P 60 0.8 5.3 29.7 15.9 1.9 8.2 64.9
In Table 6-1,θ is the opening angle of the cell structure; Fmax means peak contact force
and Tpeak is the arrival time for this contact force; E1 is the absorbed energy and E2 is the
1 2
impact kinetic energy calculated by mv0 (where m=4.52kg); vo is the initial impact
2
velocity; Smax is maximum displacement of the composites; and energy absorption ratio
E1
is calculated by  100 (%).
E2
After the specimens were impacted by an impactor that has a lower velocity and a
bigger mass, the contact force-displacement curves of 8L3P, 8L5P and 8L6P are created
which are demonstrated in Figure 6-2(a). It is noticed that both 8L3P and 8L6P have
encountered high peak contact force. 8L3P is more rigid because the small cell size
makes this composite dense and hence having higher modulus, and this is the reason
that 8L3P is associated to high contact force. In the case of 8L6P, the peak contact force
161
happens as a result of crush of the composite. It is evident from Figure 6-2(a) that 8L6P
is much softer than 8L3P, and this is due to the bigger size of the cells.
Different Cell Size
1
0.9
0.8
0.7
Contact Force(KN)
0.6
8L3P60
0.5
8L5P60
0.4
8L6P60
0.3
0.2
0.1
0
-0.1 0 5 10 15 20 25 30 35 40
Displacement(mm)
(a) Contact force-displacement curve
Energy Absorption(%) Displacement(mm)

120% 37.2 40
35
100%
Max Displacement(mm)
Energy Absorption(%)
29.7 30
80% 95.0% 95.8%
25
60% 20
14.2 64.9%
15
40%
10
20%
5
0% 0
8L3P60 8L5P60 8L6P60
(b) Comparison of energy absorption
Figure 6-2 Contact force and energy absorption behavior of samples with different cell
size
162
When the samples are impacted under the current loading situation, the contact
responses of 8L6P60 are totally different comparing to the results from small mass and
higher velocity impact in Section 5.5.2.1. During previous impact tests under dropping
hammer system in Figure 5-13, 8L6P shows a ductile performance comparing to the rest
samples. This indicates that samples with big cell size are more sensitive to the loading
conditions and tend to be easily damaged if the impactor mass increases.
Figure 6-2(b) shows the energy absorption for samples (8L3P, 8L5P, 8L6P) with
different cell sizes. As 8L6P tends to be damaged during the impact, it is obsoleted from
the current discussions. For samples of 8L3P and 8L5P, which are stronger enough to
resist the incoming force, the energy absorption between them are very similar.
Figure 6-2(b) also shows that the maximum displacement ratio for 8L3P is less than the
rest samples at only 14.2%, in other words, 8L3P has encountered a smaller deformation
than that of 8L5P. This could be due to that the cells for 8L3P are very dense thus it
makes 8L3P hard to be deformed. Therefore, under the same impact, the depth of
deformation for 8L3P is relatively shallow correspondingly.
With reference to the contact force, max displacement ratio and energy absorption; it
seems that although the samples of 8L3P and 8L5P absorb a similar amount of energy,
8L3P absorbs the energy in a way of high contact force and less deformation while
8L5P absorbs the energy in a way of low contact force and more deformation.
Composite materials with a ductile behavior under impact are more attractive for the
cushioning purposes; therefore, it further verified that samples with medium cell size
can perform better under various loading conditions.
163
6.2.2 Opening angle and its experimental performance (8L6P30, 8L6P45, 8L6P60
and 8L6P75)
The design of the composite with different opening angle from 30°to 90°aims to figure
out the mechanical performances of the textile honeycomb composites under a large
mass and lower velocity impact (v0<2.0m/s & M=4.52kg). As discussed in the previous
work in Section 5.5.2.2), the composites with the opening angle between 60°and 75°
own a better protection capability when it is stroke under the velocity of 5.5m/s with the
impact mass of 0.55kg, the present research is targeted to find out is there any similarity
or variety when the composite is impacted under the current condition. Ideally five
samples with opening angle at 30°, 45°, 60°, 75° and 90° should be prepared for the
testing in order to compare them with the previous tests in Chapter 5, however, due to
the fabric shortage, samples with 90° opening angle is obsoleted in the current
experiment. Four samples with opening angle at 30°, 45°, 60°and 75°were impacted
with different velocities around 2.0m/s and the tested results are listed in Table 6-2 for
the data analysis purpose.
Table 6-2 Experiment results from impact (samples with different opening angle)

(%)
Type (KN) (J) (mm) (ms) (m/s) (J) E2
8L6P 30 0.9 7.1 22.5 12.4 2.0 8.6 82.9

8L6P 45 0.9 7.3 21.7 9.5 2.0 8.8 83.2
8L6P 60 0.8 5.3 29.7 15.9 1.9 8.2 64.9
8L6P 75 0.4 3.2 29.1 14.6 1.9 8.5 38.1
164
Different Opening Angle
1.2
Contact Force(KN) 0.8

8L6P30
0.6
8L6P45
8L6P60
0.4
8L6P75
0.2
0
0 5 10 15 20 25 30 35
-0.2
Displacement(mm)
(a) Contact force-displacement curve
Energy Absorption (%) Max Displacement(mm)
90% 83.2% 35
82.9%
80% 29.7 29.1 30
Max Displacement(mm)
70% 64.9%
22.5 25
60% 21.7
50% 20
38.1%
40% 15
30%
10
20%
5
10%
0% 0
8L6P30 8L6P45 8L6P60 8L6P75
(b) Comparison of energy absorption
Figure 6-3 Contact force and energy absorption behavior of samples with different
opening angle
165
Figure 6-3(a) shows the contact force-displacement curves of samples with different
opening angles. It seems composites with 30°, 45° and 60° opening angles were
destroyed during the current impacts because significant sharp increases of the contact
forces are observed in Figure 6-3(a) after the composites are deformed for a period time.
The reason to cause this could be the impactor touches the anvil thus increases the
contact force suddenly.
However, from Figure 6-3(a) and (b), it seems samples with 75°opening angles have
encountered a lower contact force and a smaller absorbed energy. This is against basic
principle due to the fact that if the contact force of the composite is lower which means
the composite is easier to be deformed and this will lead to a larger structure
deformation in the result of absorbing more impact energy unless the impact stops
instantly. However, in Figure 6-3(a), the displacement of the 8L6P75 is long compared
to other composites. It seems that experiment errors might have occurred during the
testing, thus, 8L6P75 will not be included in the result discussions in the current study.
6.2.3 Different length ratio of bonded and free wall and its experiment
performance ( lb  1 : 8L(3+6)P60, 8L(4+6)P60, 8L6P60; lb  1 : 8L3P60,

lf lf
8L(4+3)P60, 8L(6+3)P60)
8L(3+6)P, 8L(4+6)P and 8L6P of which lb  1 as well as 8L3P, 8L(4+3)P, 8L(6+3)P of

lf
which lb  1 are grouped to investigate the effect of length ratio of bonded and free
lf
wall on the composite mechanical performance under current loading condition. The
results from experiments are summarized in Table 6-3 and their contact force-
displacement curve and absorbed energy were illustrated in Figure 6-4.
166
Figure 6-4(a) shows the contact force-displacement curve of the samples with lb  1 . It
lf
is evident from Figure 6-4(a) that samples with longer free walls such as 8L(3+6)P have
shown a relatively ductile performance comparing to the rest samples. As it has been
mentioned in Section 6.2, sample of 8L6P is destroyed during the current impacts;
therefore, it seems to increase the free wall length of the composite cell can help to
reduce the structure failure sufficiently.
Nevertheless, the contact force responses for 8L6P is in a distinctly different way
comparing the results from dropping hammer system in Section 5.5.2.3 where 8L6P
shows the most ductile performance. This means, textile honeycomb composite with an
even length ratio of cell walls (lb:lf=1:1) tend to be more impactor weight sensitive. The
reasons to cause this need further investigation in the future work.
Table 6-3 Experiment results from impact (samples with different length ratio of cell
walls)

(%)
Type (KN) (J) (mm) (ms) (m/s) (J) E2
lb
1
lf
8L(3+6)P 60 0.3 7.9 39.8 22.4 2.0 8.8 90

8L(4+6)P 60 0.7 8.0 19.8 24.1 2.0 9.0 88.7
8L6P 60 0.8 5.3 29.7 15.9 1.9 8.2 64.9
lb
1
lf
8L3P 60 0.8 7.3 14.2 15.3 1.8 7.7 95

8L(4+3)P 60 0.5 7.9 27.2 11.9 1.9 8.1 96.5
8L(6+3)P 60 0.7 7.4 21.0 20.4 1.9 7.9 93.5
167
Comparing to the samples with lb  1 , Figure 6-4(b) indicates that the subgroup of
lf
samples with lb  1 (8L3P60, 8L(4+3)P60 and 8L(6+3)P60) not only resist the contact
lf
force sufficiently without structure failure but also finish the impact with a bouncing
process towards the end of the impact. This can be seen from the returning of the curves
towards the end of the curving in Figure 6-4(b) and this is the result of the impactor
being bounced back as the striking distance has reached a certain depth then reduced
back to a short distance. Normally, contact force-displacement curve like this means the
samples haven’t been striking through during the impact process as the structure is
flexible and strong enough to accumulate the incoming force.
It has to be mentioned that the bouncing process towards the end of the impact will
release a small amount of impact energy. However, compared to the impact energy
which has been absorbed during the whole impact process, the released impact energy
can be neglected (Sun, 2005; Yu and Chen, 2006).
Considering the free wall length of the composites, are all composed by 3 picks which
makes the composites cell very small in sizes. Therefore, 8L3P, 8L(4+3)P and
8L(6+3)P are very dense in their material property which can provide more resistance to
the incoming force accordingly.
168
Different Wall Ratio (lb/lf<=1)
0.8
Contact Force(KN)
8L(3+6)P60
0.6
8L(4+6)P60
0.4
8L6P60
0.2
0
0 5 10 15 20 25 30 35
-0.2
Displacement(mm)
(a) Contact force-displacement curve for subgroup lb  1

lf
Different Wall Ratio (lb/lf>=1)
0.9
0.8
0.7
0.6
Contact Force(KN)
8L3P60
0.5
8L(4+3)P60
0.4
8L(6+3)P60
0.3
0.2
0.1
0
0 5 10 15 20
-0.1
Displacement(mm)
(b) Contact force-displacement curve for subgroup lb  1

lf
Energy Absorption
120%
95% 96.50% 93.50%
100% 90%
88.70%
80%
64.90%
60%
40%
20%
0%
8L(3+6)P60 8L(4+6)P60 8L6P60 8L3P60 8L(4+3)P60 8L(6+3)P60
Composite
(c) Comparison of energy absorption
169
Figure 6-4 Energy absorption of samples with different length ratio of free and bonded
wall
Figure 6-4(c) clearly demonstrated that subgroup with lb  1 has a slightly better
lf
capability for energy absorption comparing to subgroup with lb  1 under current

lf
impact situation The subtle differences in the absorption of impact energy among the
composites, especially in the lb  1 subgroup, require further investigation.

lf
6.2.4 Comparison of the results between two different loading conditions
The discussions in the following sections will be focus on the mechanical performances
of textile honeycomb composites between the current experiments and the test results
from Chapter 5. It will briefly compare the contact force and energy absorption behavior
between these two different loading conditions.
6.2.4.1 Samples with different cell size (8L3P60, 8L5P60, 8L6P60)
Samples with different cell sizes from small to big have been impacted under two
different loading conditions, Figure 6-5 (a);(b);(c) have compared their contact force-
displacement curves in individual cases.
From Figure 6-5(a), it seems there is a significant curve returning towards the end of the
impact when the composites is under bigger mass impact (M=4.52kg) with a lower
impact velocity (v0=1.78m/s), this means, the impactor is bouncing back when the
impact finishes. However, the red curve, which represents the contact responses under
another loading condition doesn’t show up this bounce process at the end of the impact.
170
8L3P60
Impact Velocity=1.78m/s Impact Velocity=5.5m/s
Conatact Force(KN)
0.8
0.6
0.4
0.2
0
-0.2 0 5 10 15 20
Displacem ent(m m )
(a) 8L3P
8L5P60
0.45
0.4
0.35
Contact Force(KN)
0.3
0.25
0.2
0.15
0.1
0.05
0
-0.05 0 5 10 15 20 25 30 35 40
Displacem ent(m m )
(b) 8L5P
8L6P60
1
Contact Force(KN)
0.8
0.6
0.4
0.2
0
-0.2 0 10 20 30 40 50
Displacem ent(m m )
(c) 8L6P
Figure 6-5 Contact force-displacement curves of composite with different cell sizes
171
Figure 6-5(b) shows the contact force responses of composites with medium cell size
(8L5P) and it seems the composites under both loading conditions are capable of resist
the incoming forces. The depth of deformations for the samples under current impacts is
deeper than those under light weight impacts. The can be explained by that the heavier
impactor will strike deeper into the composites than the light weight impactor. From
Figure 6-5(b), it can also be seen that the trend of both contact force curves are similar
under two different loading conditions, this means, 8L5P performances stable under
both impacts. This is good for the purposes of PPE as reliable performances of the
composites are requested in practice.
The sample of 8L6P is destroyed under heavy weight impacts which have been
explained in Section 6.2.1, and from Figure 6-5(c), it clearly shows that 8L6P is
sufficient to resist the incoming loading when the impactor is light weight.
Regarding the energy absorption performances of the composites, the kinetic energy
should be absorbed as much as possible to prevent damage underneath. Generally, the
more impact energy is absorbed by the composite the lower the acceleration and
damage of the protected item. Figure 6-6 illustrates the influences of the composite cell
sizes on their energy absorption performances under two different impacts. Assuming
the composites are strong enough to resist the impacts, 8L3P and 8L5P both absorb
more kinetic energy under heavier weight impacts.
Combined with composites’ contact force and energy absorption performances under
both loading conditions, it seems samples with small and big cell sizes (8L3P and 8L6P)
show different contact force performances obviously. This means, the discrimination of
performances under different impact situations for the composites with small and big
cell sizes are very significant and this is bad in their application. When the designer
172
choose materials for protection purposes, it is vital the materials should have stable
mechanical performances under all kinds of exposed impact situations.
Comparison of energy absorption under different impact

situation
120%
95% 95.80%
Energy Absorption (%)
100%
85.96% 90.17% 86.90%
80%
64.90%
60%
40%
20%
0%
8L3P60 8L5P60 8L6P60
v<2.0m/s v=5.5m/s
Figure 6-6 Energy absorption under different impact situation (samples with different
cell size)
6.2.4.2 Samples with different opening angle (8L6P30, 8L6P45, 8L6P60, 8L6P75)
Figure 6-8(a);(b);(c);(d) show the contact force-displacement curves under two different
loading conditions individually. It is obviously that the contact forces from these two
experiments are very different. Samples constructed by small to medium opening angle
at 30°, 45°and 60°encountered structure failure during the heavier weight impact and
this has been mentioned in Section 6.2.2 already. The red curves from light weight
impact are much ductile and have generated lower contact force for all the composites.
Although composites at 75°opening angle seems to have a closer peak contact forces
under both loading conditions and they are sufficient to resist the incoming forces too,
however, the large opening angle significantly increases the composites’ thickness
which makes the composites very bulky in practice.
173
It seems, whatever the opening angles of the composites are, if the impact weight
increases dramatically, and here from 0.55kg to 4.52kg, the honeycomb structures will
lost their efficiency to resist the incoming forces. In another word, by only adjusting the
opening angle of the cells to strengthen the honeycomb composite materials against the
impact loadings is not sufficient enough to resist the incoming forces, others methods
should be found out to enhance the composite structure performances against impacts
too.
8L6P30
1.5
Contact Force(KN)
0.5
0
0 10 20 30 40
-0.5
Displacem ent(m m )
(a) 8L6P30
8L6P45
1
Contact Force(KN)
0.8
0.6
0.4
0.2
0
-0.2 0 5 10 15 20 25 30 35
Displacement(mm)
(b) 8L6P45
174
8L6P60
0.8
Contact Force(KN)
0.6
0.4
0.2
0
0 10 20 30 40 50
-0.2
Displacem ent(m m )
(c) 8L6P60
8L6P75
0.4
Contact Force(KN)
0.3
0.2
0.1
0
0 10 20 30 40 50
-0.1
Displacem ent(m m )
(d) 8L6P75
Figure 6-7 Contact force-displacement curves for composites with different opening
angles
Comparison of energy absorption under different impact

situation
120%
96.9% 95.7%
100% 86.9% 84.4%
82.9% 83.2%
80%
64.9%
60%
38.1%
40%
20%
0%
8L6P30 8L6P45 8L6P60 8L6P75
v<2.0m/s v=5.5m/s
Figure 6-8 Energy absorption under different loading conditions (samples with
175
The test results under light weight impacts in Section 5.5.2.2 have mentioned that when
the opening angles of the composites increase up to 90°, the free wall and bonded wall
of the cells form a shape of almost rectangular and it reduces the flexibility of the
composite which restricts the deformation of the composites and causes less strain
energy been absorbed. It seems under current heavier weight impacts, the honeycomb
structures don’t deform much either. However, from Figure 6-8, it seems the energy
absorption under heavier weight impacts for the samples (8L6P75) are less than those
under light weight impacts. The reasons to cause it are not clear at the moment and
more tests will be needed to explain it in the future.
6.2.4.3 Samples with different length ratio of free and bonded wall ( lb  1 : 8L(3+6)P60,
lf
8L(4+6)P60, 8L6P60; lb  1 : 8L3P60, 8L(4+3)P60, 8L(6+3)P60)

lf
Figure 6-9(a);(b);(c) illustrates the contact force-displacement curves for the samples
with lb  1 . It seems only the samples with the longest free wall (8L(3+6)P) can resist
lf
the heavy weight impact and haven’t been destroyed. The contact force responses in
Figure 6-9(b) and (c), are both showing that the rest two samples are encountering
structure failure significantly. This indicates that increasing the cell free wall on the
assumption of that the bonded wall is in fixed lengths, can strengthen the honeycomb
structure against the incoming force. The energy absorption performances in Figure 6-
9(d) state that 8L(3+6)P absorbs similar energy under two different loading conditions.
176
8L(3+6)P60
0.5
Contact Force(KN)
0.4
0.3
0.2
0.1
0
-0.1 0 10 20 30 40
Displacem ent(m m )
(a) 8L(3+6)P
8L(4+6)P60
Impact Velocity=2.0m/s Imapct Velocity=5.5m/s
0.8
Contact Force(KN)
0.6
0.4
0.2
0
-5 0 5 10 15 20 25 30 35 40
-0.2
Displacem ent(m m )
(b) 8L(4+6)P
8L6P60
0.8
Contact Force(KN)
0.6
0.4
0.2
0
0 10 20 30 40 50
-0.2
Displacem ent(m m )
(c) 8L6P
177
Comparison of energy absorption under different
impact situation
100% 94.6%

90.0% 88.7% 86.9%
82.6%
80%
64.9%
60%
40%
20%
0%
8L(3+6)P60 8L(4+6)P60 8L6P60
v<2.0m/s v=5.5m/s
(d) Comparison of energy absorption
Figure 6-9 Contact force-displacement curve and energy absorption diagram for the
sample with different length ratio of bond and free wall ( lb  1 )

lf
Figure 6-10(a);(b);(c) list the contact force-displacement curves for the samples with
lb
 1 , which have been impacted under both loading conditions. It seems composites
lf
with a slightly difference in their wall ratio (lb:lf=4:3) have a more similar contact force
responses under various loading conditions, this means, they are more reliable in their
mechanical performances when the impact situations change.
178
8L3P60
Conatact Force(KN)
0.8
0.6
0.4
0.2
0
-0.2 0 5 10 15 20
Displacem ent(m m )
(a) 8L3P60
8L(4+3)P60
0.6
Contact Force(KN)
0.4
0.2
0
0 5 10 15 20
-0.2
Displacem ent(m m )
(b) 8L(4+3)P60
8L(6+3)P60
0.8
0.7
Contact Force(KN)
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.1 0 5 10 15 20 25
Displacem ent(m m )
(c) 8L(6+3)P60
179
Comparison of energy absorption under different
impact situation
120%

100% 95.0% 96.5% 93.5%
86.0% 84.5% 83.9%
80%
60%
40%
20%
0%
8L3P60 8L(4+3)P60 8L(6+3)P60
v<2.0m/s v=5.5m/s
(d) Comparison of energy absorption

(e)
Figure 6-10 Contact force-displacement curve and energy absorption diagram for the
sample with different length ratio of bond and free wall ( lb  1 )

lf
However, from the shape of contact force-displacement curve in Figure 6-11, it seems
that the structure of 8L3P60 and 8L(4+3)P60 is capable enough to bounce back the
impactor when it is under large mass and lower velocity impact (navy cure in Figure 6-
11(a)(b)) and therefore the curves were returning backwards at the end of the impact
Generally, the energy absorption for the composites in the subgroup of lb  1 shows a
lf
higher value under light weight impact in Figure 6-10(d). Further investigations will be
required to seek out the reasons to cause this phenomenon in the future.
180
6.3 Summaries
The present chapter investigates the composites with different geometric parameters
under heavy weight and lower velocity impact in order to compare their mechanical
performances with another impact situation of light weight and higher velocity in
Chapter 5, using various impact indices, including contact force, energy absorption and
maximum structure displacement. The different performance indices are discussed as
follows.
Generally speaking, most of the textile honeycomb composites are impactor weight
sensitive regarding their mechanical performances. Under heavy weight loading
conditions, only those composites with small to medium cell sizes or larger opening
angles are providing sufficient resists to the impact forces.
Composites with medium cell sizes (8L5P) have more stable mechanical performances
under various exposed impact conditions. Although samples with smaller cell sizes are
capable of resist the impact, it encountered a higher loading force under heavy weight
impacts and this will accelerate the composites and cause more damages to the item
underneath. It seems that if slightly increases the bonded wall length (lb), provided that
the composites are strong enough to resist incoming forces, can bring more stable
performances to the composites when they are under various loading conditions.
There are a few composites with free wall length of 6 picks (8L6P, 8L6P30,8L6P45,
8L(4+6)P) are all destroyed during heavy weight impact, this indicates, big cell sized
composites are very easy to be destroyed under heavy weight loading conditions.
Therefore, under more critical loading conditions, it is necessary to find a way to
enhance the big cell sized composites’ wall material or structures to strengthen their
structure performances correspondingly.
181
CHAPTER 7
FEA ON TEXTILE HONEYCOMB COMPOSITES
The discussion on mechanical behaviour based on the results from experiments

described in the previous chapters has revealed complex interactions among the
structural parameters of the honeycomb composites and the impact performances
including the energy absorption, transmitted force, and impact force. In addition to the
experimental analysis, finite element (FE) method is used to model the performance of
the 3D textile honeycomb composites and to study the influence of the geometrical
parameters on the performance of the 3D honeycomb composites. This chapter reports
on the FE simulations with the aim to examine (i) the composite performance under
impact with different energy levels and (ii) the influence of the impactor shape on the
composite performances. Standard tensile test was conducted to obtain the material data
of the cotton/epoxy single layer sheet for the finite element analysis (FEA).
Ideally, textile honeycomb models in 3D should all be created for the FEA. However,
the amount of calculation is usually too big for the computer hardware and software to
handle, therefore, in the current work, 3D models for limited composites are created and
used for FEA in order to save calculation time. As a simplified method, honeycomb
models in 2D are created as it is reported that the 2D models are also be able to be used
to evaluate effectively the performances of the modelled objects, though the accuracy is
lower than in the case of 3D models (Yu and Chen, 2006).
The purpose of the FEA based on the 2D models is to examine the mechanical
performances of the composites under impact with different impactor shapes i.e., the
cylindrical and the spherical and they are two of the most commonly used projectile
shapes (Yu and Chen, 2006). The impact energy (E) is assumed to be 6J, 8.3J and 10J,
182
of which E=8.3J is to simulate the impact energy in the experiments (E=8.3J-8.5J).
Impact energy levels 6J and 10J are designed for FE analysis only.
3D geometrical models are more complicated but closer to reality. Three models are
created for 8L3P, 8L4P and 8L6P honeycomb composites only.
Experimental results are used to validate the simulation results for both models in 2D
and 3D.
7.1 FEA Based on 2D Honeycomb Composite Models

7.1.1 Creation of 2D models for textile honeycomb composites
Strictly speaking, before modelling the textile honeycomb composites, there are a lot of
variants which should be considered for the cotton/epoxy single layer composites in the
micro-structural scale including fabric weave structure, fabric weave density, fibre and
matrix material, yarn placement, yarn size and type etc. However, to model the
honeycomb composites in such a micro-structural detailed is extremely difficult, time
consuming and almost impossible to carry out in practice. Therefore, in the current
study, these cotton/epoxy single layer composites are modelled as homogenous and
isotropic sheet which is the most common method to conduct macro-mechanical
analyses for woven composite materials (Antonio, 2000), without involves the micro-
structure fabric and yarn. This makes the FEA on honeycomb composite models to
become simpler.
Also, it has to be noted that for simplicity, the cotton/epoxy single layer composites was
assumed to be homogeneous and isotropic in the FE model, rather than anisotropic as it
should be. Assumptions like this have been widely used by other researchers in the FEA
for textile woven composites (Xu et al., 1995; Yu and Chen, 2004; Tan and Chen, 2005;
Tan et al., 2007). On the other hand, it is almost impractical to investigate
experimentally the material property of the single cotton/epoxy sheet along thickness
direction due to the shortage of techniques to obtain the strain-stress behaviour for the
single layer sheet. This assumption, however, would lead to useful results for
183
establishing understanding of the impact performances of the honeycomb textile
composites.
The composite model design aims to reflect the fundamental features of the textile
honeycomb composites, such as overall dimension, material of single layer composites
and structure design. As mentioned in the previous chapter (see Section 4.1.1), the
honeycomb composites were made by impregnating the 3D honeycomb fabrics with
resin and there are all together 14 composites with different geometric parameters in the
experimental investigations. The average dimensions of the sample composites are
120mm (L) × 60mm (W) with varying thickness because of the opening angles and the
increased layers of the specimens.
In this part of FEA, twelve 2D models of 4-layer honeycomb composites with different
lb
cell sizes, cell opening angles, and cell wall ratios ( ) are created. The models of
lf
honeycomb composites with 2 –layer and 3-layer (4L6P and 6L4P) are obsolete from
the current FEA because there are errors in running these 2 models and the software
stops working. It could be because the honeycomb structures with fewer layers are too
weak to resist the incoming force, in another words, the force applied on the impactor
model exceeds the tolerance of honeycomb model. The detailed schematic illustrations
with structural parameters are listed in Table 7-1.
Table 7-1 Schematic illustrations with structural parameters of 12 geometric models
Cell Structure θ lb lf h w Models

(°) (mm) (mm) (mm) (mm)
8L3P 60 4.07 4.07 31.60 58.78
8L4P 60 5.08 5.08 40.54 58.98
184
8L5P 60 6.35 6.35 49.23 55.94
8L6P 30 7.62 7.62 35.71 52.09
8L6P 45 7.62 7.62 48.31 54.61
8L6P 60 7.62 7.62 58.02 57.48
8L6P 75 7.62 7.62 64.12 57.84
8L6P 90 7.62 7.62 66.23 56.54
8L(6+3)P 60 7.62 3.81 31.60 55.96
8L(4+3)P 60 5.08 3.81 31.60 56.54
8L(4+6)P 60 5.08 7.62 57.10 52.46
8L(3+6)P 60 3.82 7.64 57.10 52.46
In Table 7-1, θ is the opening angle of the cell; l b and lf are the length of the bonded
wall and the length of the free wall of the cell measured in mm. tb and tf represents the
185
thickness of the bonded and free walls of the cell; h indicates the height of the
honeycomb composite and w is the width of the composite, both measured in mm.
7.1.2 Meshing the geometrical models and the impactor
The creation of FE models, also known as meshing, is an important step in the FEA.
The selection of element type and meshing quality would influence the accuracy of the
FEA results. FEA are known to lose accuracy when the original mesh becomes highly
distorted. A refinement of the mesh which supports the capture of nonlinear material
effects is often required due to the fact that there is always a potentially large relative
deformation in the FE impact simulation. However, the refinement of the FE mesh is
linked to an increase of the computational costs. In order to retain the efficiency of the
calculation in the present work, the mesh size would only be refined in the impact
vicinity of the models, whereas the rest area uses a coarser mesh and this method has
been used frequently in 3D FE model meshing as more complicate shell refinement will
be needed and more meshed element will be involved.
The software (Marc.Mentat, 2005a) is used for the model creation as well as the FEA.
The individual cell walls are created by defining the surfaces first then each surface is
meshed individually using the 4-node axisymmetric elements with full integration. The
element type uses ‘Type 10’, which is regarded as suitable for dynamic contact analysis.
It is a four-node, isoparametric arbitrary quadrilateral element for axisymmetric
applications. The term ‘node’ describes the edge points of an element and the location
of the element in the 2D space. The nodes on the top or bottom surface of the cell wall
usually are used as reference points for the subsequent data collection. The meshing of a
cell is shown in Figure 7-1.
186
Figure 7-1 Meshing of a cell
7.1.3 Meshing the impactor
Two shapes are used to represent the impactor in the 2D FE modelling, which are
cylindrical and spherical. The cylindrical impactor will impact at the centre of the top
surface of the honeycomb composite models. The geometry of the cylinder impactor is
30mm (d) × 99mm (L), the diameter of the cylindrical shaped impator are modelled as
the same size as used in experiment. The material of the impactor is assumed to be steel
instead of wood which has been used in the experiments, thus, the length of the
cylindrical impactor was shortened in order to keep the same impactor mass as that used
in the experiments for0.55kg. The mesh size is refined to match the fineness in the
impact vicinity of the top surface of the honeycomb model. The element type used for
the cylindrical impactor is the 4-node quadrilateral element because this element suits
dynamic contact analysis. Figure 7-2(a) shows the 2D projection of the meshed cylinder
impactor.
A ball impactor is also used for the FE impact analysis. The diameter of the ball
impactor is 32.4mm and the impactor material is assumed to be steel. The mesh size is
187
refined to match the fineness in the impact vicinity of the impact model. The element
type been is defined the same as that in the cylindrical impactor. The meshed spherical
impactor is shown in Figure 7-2(b).
(a) (b)
Figure 7-2 Meshed impactors (a) the cylinder and (b) the sphere
7.1.4 Materials
FEA requires the material properties, including the Young’s Modulus, material density
and stress-strain behaviour, to be specified. As has been mentioned in the previous
chapters, the honeycomb composites samples were made from woven fabrics with
cotton yarns and epoxy resin. Therefore, cotton/epoxy composite properties were used
as the material for cell walls of the honeycomb composites in the FEA and they are
assumed to be homogenous and isotropic. The material properties for the cylinder and
ball shaped impactors were selected to be that of steel (James, 2001).
7.1.4.1 The tensile test of a single layer composite
The mechanical properties of the cotton/epoxy single layer sheet required for the FE
modelling were determined by using the ‘grab-test’ procedure described in standard
ASTM D3039-95 (2004), which requires 6 rectangular specimens of predefined
dimensions. Grab-test is chosen to conduct the current tensile test instead of strip-test
188
because the testing conditions are similar to the load application on the specimen in
practical use (Bassett et al., 1999; Pan et al., 2003). To prepare the test samples, a plain
woven cotton fabric impregnated with epoxy resin (see section 4.1.1) were cut into a
rectangular shape with the size of 25mm×250mm according to the ASTM standard.
INSTRON 4505 tensile tester was used for the tensile testing; the maximum load cell
capacity of the tensile tester is 50KN. For the test, the crosshead was set to move at a
rate of 2mm/min. Each specimen was then clamped with a grip width of 25mm, and the
gauge length between the clamps was set to be 150mm. To ensure a secure grip an
emery cloth was folded around the ends of the specimen.
7.1.4.2 Material properties
The Young’s modulus and stress-strain curve of the cotton/epoxy composite were
obtained from the above mentioned tensile test and the specific density, ρ, of the
cotton/epoxy single layer sheet was measured too. Table 7-2 lists the basic mechanical
properties of the cotton/epoxy composite required for the FEA and the steel impactor
(James, 2001).
Table 7-2. Mechanical properties of materials
Material Properties Cotton/Epoxy Steel (James,2001)

Specific Density, ρ (kg/m3) 2.0×103 7.85×103
Young’s Modulus, E (GPa) 0.23 200
The determined stress-strain curve of the cotton/epoxy material is displayed in Figure 7-

3 together with the curve for steel (James, 2001). In the subsequent FEA, the cell walls
of the honeycomb composites will be defined as the cotton/epoxy composite material
and the impactor will be specified as steel.
189
Stress-Strain behaviour of Cotton/Epoxy composite and Steel
800
700
Tensile Stress [GPa]

600
500
Cotton/Epoxy
400
Steel
300
200
100
0
0 2 4 6 8 10 12 14 16 18 20
Tensile Strain [%]
Figure 7-3 Stress-strain behaviour of cotton/epoxy sheet and steel
7.1.5 Boundary conditions applied to the honeycomb composite models
In order to carry out FEA on the composites, boundary conditions have to be defined to
simulate the clamping and placement of the real test specimen. This will provide certain
degrees of freedom to allow a natural collision response to the impact. In the current
case, the constraints are defined based on the conditions used in the impact test of
textile honeycomb composites, where the specimen is placed and fixed by adhesive
tapes on the anvil. The honeycomb composite model can only deform corresponding to
the impact load and will not rotate. This boundary conditions are added by fixing the
edges of the honeycomb composite models with the constraints having their effects in x
and y directions. An illustration of the FE model and the constraints is shown in Figure
7-4.This constitution provides sufficient local flexibility of the honeycomb models for
the impact response.
190
Figure 7-4 Schematic illustration of boundary conditions for the FE impact model
7.1.6 Impact setting for FEA of 2D models
The analysis is for low-velocity impacts, which is the same loading condition as that in
the experiments. The level of impact energy has been selected to be 6J, 8.3J and 10J.
The impactor shapes are cylindrical and spherical and the weight of the impactor is
0.55kg. It is observed from experiments that 25ms is sufficient enough for the
honeycomb composites to finish an impact process, therefore, the simulation time in the
current FEA is defined as 25ms and the time step is defined as 100 as an interval. The
material for the honeycomb composite is assumed to be isotropic, the Young’s Modulus
is 0.23GPa, and the specific density of the material is 2×103 (kg/m3) as has specified in
Section 7.1.4.2. The levels of impact energy and velocity are displayed in Table 7-3,
and Table 7-4 gives details of the FE models.
Table 7-3 Impactor mass, impact velocity and impact energy
Impactor Mass(kg) Impact Velocity (m/s) Impact Energy (J)

4.67 6
0.55 5.5 8.3
6.03 10
191
Table 7-4 Details of FE models
Impactor Impact
Group Models
Shape Energy (J)
8L3P60, 8L4P60, 8L5P60,
Cell Size
8L6P60
8L6P30, 8L6P45,8L6P60,
Opening Angle
8L6P75, 8L6P90
lb Cylinder 6, 8.3 and
1 8L(3+6)P60,8L(4+6)P60, and Ball 10
Ratio of Bonded lf 8L6P60
and Free Wall lb 8L(4+3)P60, 8L(6+3)P60,
1
lf 8L3P60
7.1.7 Results and discussions of FEA based on 2D models
In this section, 2D FEA results are given and discussed regarding the mechanical
performances for the models including deformation area, the depth of deformation ,
dynamic contact force and the energy absorption of the composites.
FEA results are also compared to the experimental results for validating the models.
7.1.7.1 Introduction of performance indices
Before introducing the FEA results, it is necessary to describe some of the performance
indices that are to be used in this analysis. Such indices include composite deformation
area, energy absorption density and the depth of deformation in the composite models.
Deformation area in the composite models under cylinder impact
In investigations for performances of the honeycomb models under impact, the shape
and dimension of the deformed volume bear important information that indicates the
composite behaviour. In 2D FEA, a trapezoidal shaped area, which represents the cross-
192
section of the deformed volume of the composite, is used as a performance index to
represent approximately the manner of deformation for the honeycomb models.
It is assumed that the ability for energy absorption of a textile honeycomb composite is
related to the deformation area in the 2D simulation and the shape of this area may
indicate how the energy is absorbed during the impact. For example, for the same
energy absorption, it can be shallow and wide or it may be deep and narrow.
Figure 7-5 shows a deformed cross-section of the honeycomb composite due to impact.
It is reasonable to assume a trapezoidal shaped area to represent the deformed area.
Figure 7-5 Estimation of deformed cross-section represented by a trapezoidal shaped

area (use 8L6P60 as an example)
193
The deformed area, S, in Figure 7-5 can be calculated by the following equation:
( Lt  Lb )  h'
S [7-1]
2
where S is the area of the trapezoidal shape, Lt and Lb are the lengths of the top and
bottom edge for the trapezoidal shape, h‟ represents the height of the trapezoid. The
calculated trapezoidal area leads to a numerical expression of the deformation area.
Deformation area ratio (%), denoted by R, can be used to express the percentage of the
deformation area against the whole composite cross-section area of the model, which is
defined as follows.
S
R  100 (%) [7-2]
w h
where w is the original width and h is the original height of the honeycomb structure in
2D with the unit of mm, and their value have been listed in Table 7-1.
It has to be noticed that the deformation area generated above is based on the 2D
models and therefore, the calculated deformation area ratio is only a way to
approximately evaluate the deformation dissipation capability of the model in general
and more accurate deformation area are expected to be generated in 3D models in the
future.
The depth of deformation

Followed by the above description of deformation area (S‟), the depth of deformation
for the models can be generated as h‟ from the Figure 7-5. It can also be used as an
performance index to evaluate the impact performance of the honeycomb composites.
In the current study, the depth of deformation ratio (%), denoted by D, is expressed as a
percentage and the equation is shown below:
h'
D  100 (%) [7-3]
h
194
where h is the height of the composite.
Comparing to the deformation area mentioned above, which is a numerical value to

represent the deformation of the model in two-dimension, the depth of deformation of
the model is also a numerical value to express the deformation of the model, however, it
is in one-dimension.
Strain energy density
Strain energy density, defined as the strain energy absorbed per unit volume of the
textile honeycomb composite, is used to describe the energy absorption performance.
The strain energy density is determined by dividing the total strain energy E by the
volume V of the model. Thus the strain energy density, denoted by the symbol e, can be
expressed in the following form:
E
e= [7-4]
V
or,
E= e  V [7-5]
where E is the total strain energy absorbed by the model with the unit of J, and V the
total original volume of the model with the unit of mm3 and V=w  h  t, in which w, h
and t are the original width (mm), original height (mm) and original thickness (mm) of
the honeycomb structure, respectively.
In the current FEA, the software (Mac.Mentat, 2005a) can generate the total strain
energy of the whole model as an output and this enables the calculation of strain energy
density e (J/mm3) easily. The original width (w), original height (h) has been listed in
Table 7-1, and the original thickness (t) of the model is assumed to be 120mm for all the
models, which is the same as for the real composites.
195
7.1.7.2 Classifications of the FE composite models
The textile honeycomb composite models, with 12 varieties, are classified into 4 groups
in the current works as follows:
1). different cell size (8L3P, 8L4P, 8L5P, 8L6P)
2). different opening angle (8L6P30, 8L6P45, 8L6P60, 8L6P75, 8L6P90)
lb
3). different cell wall length ratio of  1 : (8L(6+3)P, 8L(4+3)P, 8L3P)
lf
lb
4). different cell wall length ratio of  1 : (8L6P, 8L(4+6)P, 8L(3+6)P)
lf
7.1.7.3 Simulated results
The simulated results from 2D FEA are listed and categorised into 4 groups according
to their geometric parameters.
Group I. Models with different cell size
In this group, textile honeycomb composites with different cell size (8L3P60, 8L4P60,
8L5P60 and 8L6P60) were modelled in 2D. The schematic of their deformation pattern
under impact energy of 6J, 8.3J and 10J are listed in Table 7-5. Their depth of
deformation ratio (D), deformation area ratio (R), and strain energy density (e) were
recorded in Table 7-5 too.
Table 7-5 Effect of cell size on models under cylinder impact
Impactor Type: Cylinder
Model Original Shape 6J 8.3J 10J
8L3P60
196
Deformed
Pattern
D (%) 28.7 37.4 52.8

R (%) 12.3 14.9 18.4
e (×10-6), J/mm3 21.3 28.6 31.5
8L4P60
Deformed
Pattern
D (%) 39.7 56.0 63.1

R (%) 13.5 18.4 22.4
e (×10-6), J/mm3 16.7 23.8 28.3
8L5P60
Deformed
Pattern
D (%) 42.1 54.4 69.4

R (%) 8.6 15 18.2
-6 3
e (×10 ), J/mm 17.2 23.2 29.8
8L6P60
Deformed
Pattern
197
D (%) 43.4 60.0 82.8
R (%) 12.5 16.8 24.1
-6 3
e (×10 ), J/mm 14.8 19.8 24.8
Group II. Models with different opening angle
Models with same cell wall length but different opening angles from 30° to 90° are
categorised into this group. They are 8L6P30, 8L6P45, 8L6P60, 8L6P75 and 8L6P90
respectively. Table 7-6 illustrates their deformation details after impact.
Table 7-6 Effect of cell opening angle in the models under cylinder impact

Model Original Shape 6J 8J 10J
8L6P30
Deformed
Pattern
D (%) 54.5 88.2 96.9

R (%) 22.7 28.8 43.6
-6 3
e (×10 ), J/mm 25.1 34.6 42
8L6P45
Deformed
Pattern
198
D (%) 43.1 52.7 59.7
R (%) 14.8 18.7 27.1
-6 3
e (×10 ), J/mm 16.7 23.5 27.8
8L6P60
Deformed
Pattern
D (%) 43.4 60.0 82.8

R (%) 12.5 16.8 24.1
e (×10-6), J/mm3 14.8 19.8 24.8
8L6P75
Deformed
Pattern
D (%) 20.3 24.9 25.9

R (%) 4.4 9.1 5.5
e (×10-6), J/mm3 12.1 14.6 15.4
8L6P90
Deformed
Pattern
D (%) 0 2.4 5.2

R (%) 0 2.2 1.5
e (×10-6), J/mm3 0 2.3 3.8
lb
Group III. Models with 1
lf
199
lb
In this group, textile honeycomb composites with  1 (8L3P60, 8L(4+3)P60,
lf
8L(6+3)P60 were modelled in 2D and their FEA results are shown in Table 7-7.
lb
Table 7-7 Effect of cell wall ratio (  1 ) on the models under cylinder impact
lf

Model Original 6J 8J 10J
8L3P60 Shape
Deformed
Pattern
D (%) 28.7 37.4 52.8

R (%) 14.3 18.6 21.5
e (×10-6), J/mm3 21.3 28.6 31.5
8L(4+3)P60
Deformed
Pattern
D (%) 30.7 42.4 57.6

R (%) 6.0 12.9 19.2
e (×10-6), J/mm3 24.5 34.8 40.6
8L(6+3)P60
Deformed
Pattern
200
D (%) 45.1 67.3 76.8
R (%) 11.6 22.5 26.2
-6 3
e (×10 ), J/mm 26.9 31.6 35.2
lb
Group IV. Models with 1
lf
lb
In this group, textile honeycomb composites with  1 (8L6P60, 8L(4+6)P60,
lf
8L(3+6)P60 were modelled in 2D and their FEA results are shown in Table 7-8.
lb
Table 7-8 Effect of cell wall ratio (  ) on the models under impact
lf

8L6P60
Deformed
Pattern
D (%) 43.4 60.0 82.8

R (%) 12.5 16.8 24.1
-6 3
e (×10 ), J/mm 14.8 19.8 24.8
8L(4+6)P60
Deformed
Pattern
201
D (%) 30.9 50 58.5
R (%) 9.4 14.7 19.0
-6 3
e (×10 ), J/mm 14.7 22 26.9
8L(3+6)P60
Deformed
Pattern
D (%) 45.4 63.5 71.3

R (%) 14.3 20.5 23.9
e (×10-6), J/mm3 15.7 22.2 27.3
7.1.7.4 Deformation area under cylinder impact
The deformation areas of the models with different geometric parameters are listed
schematically in Table 7-5 to 7-8. Their deformation area ratio (%) are calculated by
using trapezoidal area with their original area from Equation 7-2. Obviously, different
amount of impact energy leads to different deformation area and generally, models
impacted under higher impact energy (10J) created more deformation area than that
under lower impact energy (6J). It is also noted from the figures in Table 7-5 that when
the models were impacted with the same amount of energy, the deformation area could
vary a lot if the geometric parameters of the model change, and the following section
will discuss this more in detail.
Deformation area of models with different cell sizes
202
30
6J
24.1 8J
25 22.4
Deformation Area Ratio (%)

10J
20 18.4 18.2
16.8
14.9 18.4
15
15
13.5 12.5
10 12.3
8.6
5
0
8L3P60 8L4P60 8L5P60 8L6P60
(a) Deformation area ratio (R)
6J
100
8J
90
Depth of deformation ratio, D (%)
10J
80
82.8
69.4
70
63.1
60
60
50 56 54.4
52.8
40 43.4
37.4 42.1
39.7
30
20 28.7
8L3P60 8L4P60 8L5P60 8L6P60
(b) The depth of deformation ratio (D)
203
100
FE Result(E=8J)
90
Depth of deformation Ratio, D (%)

Experiment Result(E=8J)
80
70
60
50
40
30
8L3P60 8L4P60 8L5P60 8L6P60
(c) Comparison betwwen experimental and 2D FE results
Figure 7-6 Deformation of models with different cell sizes under impact energy of 6J,
8.3J and 10J
The comparison of deformation area ratio (R) and the depth of deformation ratio (D) for
the models from small to big cell size are shown in Figure 7-6(a) and (b). The impact
energy are various at 6J, 8.3J and 10J. At first instance, it can be seen that the
deformation area ratio increases when the impact energy is getting higher, and so does
the depth of deformation ratio are the same. This is understandable as more impact
energy causes more structure deformation accordingly.
Secondly, it is noticed from Figure 7-7(a) that the deformation area ratio for samples
with big cell size (8L6P) are generally larger comparing to the rest samples under
various impact energies and this indicates that big cell sized honeycomb composite
modelsare more easy to be damaged during the impact.
The FE results for the depth of deformation ratio have been compared with experiment
results in Figure 7-7(c) too, and generally, for the sample of 8L4P60 and 8L5P60, the
FE results are only slightly lower than experiment results. While for 8L3P60 and
8L6P60, the simulated the depth of deformation ratio is smaller than the experiment
204
results. However, for both results from FE and experiment, it states that 8L6P60
encounters the deepest vertical displacement and this caused large deformation area
which has been found in Figure 7-6(a)
Combining with deformation area ratio and the depth of deformation ratio for 8L4P60
and 8L5P60 in Figure 7-7 (a) and (b), it seems that models with medium cell size
(8L5P60) has had a relatively smaller deformation area ratio while their depth of
deformation ratio is similar. In another words, it means that if there is an object which
impacts the model and strikes up to a similar distance, it will cause less damage if the
cell size of the model is medium sized.
The results from previous experiments in Chapter 5 (Section 5.5.2.1) already states that
samples with a medium cell size like 8L5P60 is recommended to be used in the
application as it shows up a considerable force attenuation and energy absorption
capability with a reasonable material handling property. Here, the FE results further
noticed that 8L5P60 encounters less damage under impact situation.
Deformation area of models with different opening angle
The photographs of deformation area for the models with different opening angle are
listed in Table 7-6. When the cell opening angle is less than 60°, the deformation area
forms an obvious trapezoidal shape. However, it is clearly shown that when the cell
opening angle exceeds 75° the deformation area is more or less close to a rectangular
shape. This indicates that when the opening angle is large enough, not much of the
impact energy is absorbed by the buckling rigidity of the cell walls. Instead, much of the
impact energy is passed through the cell walls to do work on the other side of the
models. For the sample of 8L6P90, the deformation area is tiny and this means when the
cell opening angle enlarges up to 90°, the model is very stiff and hard to deform.
Similar findings have been mentioned in the experiment results’ discussion too in
Section 5.5.2.2 too.
205
100 43.6
10J
90 8J
80 6J
70
27.1
60 24.1
28.8
50
40 18.7
30 22.7 16.8
20 9.1
5.5
10 14.8 1.5
12.5 2.2
0 4.4 0
8L6P30 8L6P45 8L6P60 8L6P75 8L6P90
Figure 7-7 Comparison of deformation area in models with different opening angles
From Figure 7-7, it is noted that 8L6P75 and 8L6P90 provide a smaller deformation
area ratio than that in the rest of models whatever the impact energy is 6J, 8.3J or 10J.
This is due to the fact that the shape of the cell in models of 8L6P75 and 8L6P90 is
close to rectangular and buckling becomes the main form of deformation. Obviously,
the buckling rigidity of the cell walls is much larger than the bending rigidity. Under all
impact levels, the performance of 8L6P60, 8L6P75 and 8L6P90 is relatively similar
whereas for 8L6P30 and 8L6P45, the performance is more different, with larger impact
energy causing more deformation. It indicates that for larger opening angles the impact
energy level would not cause too much difference in the structure damage.
It must be noted that the deformation area is only the cross-section of the concaved
deformation. A more accurate conclusion should be drawn from the 3D models. In
contrast, if the opening angle is small, the deformation area is more sensitive to the
impact energy level. This phenomenon is worth of further exploration.
206
lb
Deformation area of models with different wall ratio ( )
lf
lb
(i)  1 : 8L3P60, 8L(4+3)P60, 8L(6+3)P60
lf
30
6J
26.2
8J
25
21.5 10J
22.5
19.2
20 18.6
14.3
15 12.9
10 11.6
5 6
0
8L3P 8L(4+3)P 8L(6+3)P
Figure 7-8 Comparison of deformation area ratio of models with different cell wall ratio
lb
(  1 ).
lf
Figure 7-8 shows a direct comparison of cylinder impact at three different impact
energy levels and the models are designed with the bonded and free wall length ratio
lb
more than one (  1 ). It seems the models with the length ratio of 4:3 has encountered
lf
the least damage whatever the impact energy various.
207
Referring to the experiment results in Section 5.5.2.3, it has stated that with the increase
lb
of , the impact modulus of the samples reduces which leads to a better force
lf
attenuation performance. However, the energy absorption between these three samples
is similar. Here, from the FE results, it further reveals that the models with a
lb
considerable which is more than one and less than two will show a better damage
lf
tolerance than the rest models. This information again indicates that by modifying the
bonded and free wall length, it can optimize the mechanical performances of the textile
honeycomb composites sufficiently.
lb
(ii)  1 : 8L6P60, 8L(4+6)P60, 8L(3+6)P60
lf
30
6J
24.1 23.9 8J
25
10J
20 19 20.5
16.8
14.7
15
14.3
12.5
10
9.4
5
0
8L6P 8L(4+6)P 8L(3+6)P
Figure 7-9 Comparison of deformation area in models with different cell wall ratio
lb
(  1)
lf
208
lb
The numerical comparison of the deformation area ratio among the models with 1
lf
is illustrated in Figure 7-9. It can be seen that the model with a medium bonded and free
wall length ratio of 4:6 demonstrates a least deformation area ratio under three different
impact levels. The reasons for this performance are not clear at the moment and further
investigations are needed in the future research work. And according to the experiment
results in Section 5.5.2.3, there is not a significant relationship between the structure
deformation and their transmitted force or energy absorption performance for the
lb
subgroup with  1 either.
lf
7.1.7.5 History of dynamic contact force
The present section investigates the contact force response of honeycomb models under
cylinder impact in order to obtain information how the different geometric design of the
model resists to the impact loads.
Figure 7-10(a) to (d) show the history of dynamic contact force against time for the
models with different cell size, opening angle and cell wall length ratio subject to the
impact energy at 8J. The reference points are taken from the impact centre at the top
surface of the model and the average value has been calculated.
It can be seen that in general, the contact response of the cylinder impact demonstrates
an initial peak contact force and the curve fluctuate towards the end of the impact. This
is because the rigidity of the material provides resistances to the impactor and there are
different yield stresses which are required to deform the structure and value of the stress
are different depends on the structure response of the model. In 2D planner FEA, the
cell walls are buckled and plastic-elastic deformations are occurred during the impact,
which caused the cell walls touched and clued to each other (Tan and Chen, 2005; Yu
and Chen, 2006).
209
Focusing on the cylinder impact model in Figure 7-10(b), it can be seen that all the
curves follow a similar trend at early stages of impact and the trends become more
different towards the end of the impact whilst 8L6P75 and 8L6P90 are exceptions and
their peak contact force appears much later than the rest models. The opening angle of
75° and 90° provides a much stronger structure resistance to the impact due to the
rectangular shaped cells are mainly deformed by buckling mechanism and
correspondingly, the model must response in a different way to the impact load
comparing to others. And at the beginning of the impact, the model responses to the
impact while the structure may not have experienced a large amount of deformation,
which delays the appearance of peak contact force accordingly.
FE Contact Force (Different Cell Size/8J)
0.8
Contact Force(KN)
0.7
0.6
8L3P60
0.5
8L4P60
0.4
8L5P60
0.3
8L6P60
0.2
0.1
0
0 5 10 15 20 25
Time(ms)
(a) Models with different cell size
210
FE Contact Force(Different Opening Angle/8J)
0.4
0.35
Contact Force(KN)
0.3 8L6P30
0.25 8L6P45
0.2 8L6P60
0.15 8L6P75
0.1 8L6P90
0.05
0
-0.05 0 5 10 15 20 25
TimeI(ms)
(b) Models with different opening angle
FE Contact Force(Wall Ratio>=1)
0.8
0.7
Contact Force(KN)
0.6
8L(4+3)P60
0.5
8L(6+3)P60
0.4
8L3P60
0.3
0.2
0.1
0
0 5 10 15 20
Time(ms)
lb
(c) Models with 1
lf
211
FE Contact Force(Wall Ratio<=1)
0.35
Contact Force(KN)
0.3
0.25
8L(3+6)P60
0.2
8L(4+6)P60
0.15 8L6P60
0.1
0.05
0
0 5 10 15 20 25
Time(ms)
lb
(d) Models with 1
lf
Figure 7-10 Dynamic contact force of models under the impact energy of 8J
Figure 7-11 rearranges the models according to their peak contact force under the
impact energy of 8J. It can be seen that generally, models with smaller cell size such as
8L3P and 8L4P provide higher peak contact force than the models with larger cell size
and this indicates that specimen with smaller cell size are more difficult to be deformed.
The experiment results in Section 5.5.2.1 also revealed the similar material properties as
above because it is found that samples with smaller cell size such as 8L3P and 8L4P
had a higher impact modulus and their material handling property is very rigid, which
cause the samples more difficult to be deformed.
212
Peak Contact Force (Cylinder Impact/8J)
0.8
0.7
Peak Contact Force(KN)
0.6
0.5
0.4
0.3
0.2
0.1
Figure 7-11 Peak contact force from cylinder impact
lb
Figure 7-11 also reveals that generally, models with  1 requires a higher force to be
lf
lb
deformed than the models with  1 and this indicates that longer bonded wall could
lf
generate higher resistance to the impact.
7.1.7.6 Energy absorption performance
213
120
100
Energy Absorption Ratio (%)
80
60
40
FE Results
20 Experiment Results
Figure 7-12 Validation of energy absorption between FEA and experiment results
According to the strain energy density (e) from Table 7-5, the absorption energy (E)
from FEA are calculated according to Equation 7-5. The energy absorption ratio (%)
from FEA is worked out by diving the absorpted energy (E) with the initical kinetic
energy (K), which is 8.31J in the current FEA work. The energy absorption ration (%)
from experiment is listed in Table 5-1. These two energy absorption ratio from FEA and
experimentsl are compared with each other in Figure 7-12.
Generally, compared with experiment results from impact tests in Figure 7-12, the
energy absorption value from FEA has a good agreement on tendency expect for
8L6P90. In FEA, the model is idealized and the opening angle of the model is exactly
90°, in fact, in the real experiment, the opening angle of the sample is approximately
around 90° which is shown in Figure 4-4. This factor will lead to the difference of
energy absorption obtained from experimental and FEA.
7.1.7.7 Comparison between ball and cylinder shaped impact
214
The following section will investigate the difference between ball impact and cylinder
impact. The impact energy for ball and cylinder impact is both set as 8.3J with initial
impact velocity at 5.5m/s and the impactor mass is 0.55kg. Discussions will be focus on
it mechanical performance including contact force response and energy absorption
performance.
Schematic illustration of ball impact deformation
Take group of samples with different cell size (8L3P, 8L4P, 8L5P, 8L6P) as an example,
generally, the deformation of ball impact is shallow than the cylinder impact, which is
shown in Table 7-5 and Table 7-9, whilst the contact area of ball impact is larger than
the cylinder impact and this is mainly because the curvature of the ball edge can easily
touches the cell wall during the impact and there are mainly line to line deformation
occurs. In the cylinder impact, the two corners on the bottom of the impactor touches
the cell wall firstly which force the cell wall bending until the whole bottom line of the
impactor touches the original or bended/buckled cell to continue the deformation. The
schematic illustration of the ball and cylinder impact deformation process is listed in
Figure 7-13.
215
Table 7-9 Effect of cell size on its maximum displacement and energy absorption for
textile honeycomb composite models under ball impact
Impactor Type: Ball
8L3P60
Deformed
Pattern
Deformation depth ratio (%) 18.7 30.2 40.6

e (×10-6), J/mm3 10.1 22.4 26.2
8L4P60
Deformed
Pattern

e (×10-6), J/mm3 18.7 25.8 23.5
8L5P60
Deformed
Pattern

e (×10-6), J/mm3 17.3 24.2 30.3
8L6P60
Deformed
Pattern

e (×10-6), J/mm3 16.2 19.7 28.4
216
Deformation process for ball and cylinder impact
It is vital to understand the deformation process when the textile honeycomb composites
are impacted by different shaped objects. The contact force response for cylinder and
ball impact is shown in Figure 7-13 and they are taken at 8.3J impact for the model of
8L3P60.
In order to investigate the impact load transfer and the deformation of cylinder and ball
impact, Figure 7-14 specially give a virtual views of the cell deformation process. The
deformation views are taken at 8J impact. The different time steps are selected
concerning characteristics deformation steps which corresponding to the points on the
contact response diagram with [▪], are shown in Figure 7-13.
From Figure 7-13, it is seen that the individual contact curves are similar in their trends
at the beginning of the impact up to 4ms, but follows different rates. The peak contact
force is higher for cylinder impact than in the case of ball impact. In other words, a
cylinder shaped impactor penetrated more loading force at the beginning of the impact
and this could cause a faster rate of deformation, which may lead to a faster energy
input to the model at the initial stage of the impact and a different effect on the
protection level.
By putting cylinder and ball impact contact response together in Figure 7-13, it is
clearly that both of them can reach their peak contact force at around 2ms and
subsequently the contact force of ball impact starts to decrease then suddenly it
increases from between 4ms and 5ms. Referring to the configuration in Figure 7-14(b)
at 6ms, the ball impactor came in touch with the bonded wall of the cell in the centre of
the layer2 and this could increase the contact force accordingly. The contact force curve
went smoothly down until it meets the configuration [5] and [6], it slightly went up and
down again and virtually this could be the crushed cell wall touched the corner of the
cell on the next layer.
217
Figure 7-13 Comparison of contact force-time response of 8L3P60 under 8j by cylinder
and ball impact
[0] 0ms
[1] 2ms
218
[2] 4ms
[3] 6ms
[4] 8ms
[5] 10ms
219
[6] 12ms
[7] 13ms
(a) By cylinder impact (b) By ball impact
Figure 7-14 Comparison of structure deformation under dynamic impact for model
8L3P at 8J impact (a) by cylinder impact (b) by ball impact
Contact response of ball impact
Regarding the contact response of ball impact in Figure 7-15, it is noticed that the
response of honeycomb structure to the ball impact doesn’t perform identical as that of
cylinder impact, which is shown in Figure 7-13. The difference between this two types
of impact is that the change of loading force is more frequently during the ball impact
and this causes more fluctuation in the contact response in Figure 7-15.
The curvature of the ball is the reason to cause this as it touches the cell wall differently
comparing to cylinder impactor. The curvature leads to the fact that more bending
deformation occurs and it provides more opportunity for the deformed cell wall to
contact each other which various the loading force step by step.
220
The contact force are characterised by a higher fluctuation in the ball impact, which
indicates that more cell material is involved for stopping the impactor and this could
lead to a wider energy dissipation and reduced the level of contact forces and it finally
will decrease the forces transmitted underneath and lower acceleration too. This
information from FE results is useful because it can help the researchers to predict the
performance of textile honeycomb composite under ball impact in their future work.
FE Contact Force(Different Cell Size/8J Ball) FE Contact Force(Opening Angle/8J Ball)
0.7
0.45
0.6 0.4
Contact Force(KN)
0.35
8L6P30
0.5
Contact Force
0.3 8L6P45
0.4 8L3P60
0.25 8L6P60
8L4P60
0.3 0.2
8L5P60 8L6P75
0.2
0.15
8L6P60 8L6P90
0.1
0.1
0.05
0 0
0 5 10 15 20 -0.05 0
-0.1 5 10 15 20 25
Time(ms) Time(ms)
FE Contact Force(Wall Ratio>=1/8J Ball) FE Contact Force(Wall Ratio<=1/8J Ball)
0.7 0.35
Contact Force(KN)
0.6 0.3
Contact Force(KN)
0.5 0.25
8L(4+3)P60 8L(3+6)P60
0.4 0.2
8L(6+3)P60 8L(4+6)P60
0.3 0.15 8L6P60
8L3P60
0.2 0.1
0.1 0.05
0 0
0 5 10 15 20 0 5 10 15 20 25 30
-0.1
Time(ms) Time(ms)
Figure 7-15 Dynamic contact force of models under ball impact at 8J
221
Energy absorption performance of ball impact
35
30
Strain Energy Density (J/mm3)
25
20
Ball
15
Cylinder
10
0
8L3P 8L4P 8L5P 8L6P
Figure 7-16 Comparison between ball and cylinder energy absorption capability
As analysed in the above section, the impactor shape influences the deformation pattern
of the models, therefore, their energy absorption performance is supposed to be
different correspondingly. Figure 7-16 compared the strain energy density (e) between
the ball and cylinder impact among the models: 8L3P, 8L4P, 8L5P and 8L6P. Their
impact energy is around 8J. However, from the figure, it is hard to find a trend to tell
whether the model under ball impact will absorb more or less energy than cylinder
impact and this need further investigation in the future.
Nevertheless, from Figure 7-13 and 7-14, it can be seen that the cylinder impactor
provides the faster strain energy input than the ball impactor and it is likely that the rate
of strain energy induction is related to the contact area of the impactor, of which the
cylinder impactor touches the surface by point to line touch while the ball impact is line
to line touch. And also the larger the contact area the wider distributed is the cell
deformation, which leads to more plastic strains generated within a short period of time
when comparing with an impactor with small contact area. The deformation analysis
also showed that the cylinder impactor deforms the structure faster than the ball impact
222
(Figure 7-13) and this proves that the larger is the contact area, the faster is the strain
energy induction. Although it is difficult to make any reliable conclusions at the current
stage as it seems that an increase in contact area normally leads to an increase in impact
energy loss caused by friction, thermal and other forms of energy conversion, generally,
a faster strain energy induction should lead to higher acceleration of the item underneath,
because more impact energy is exposed underneath within a shorter period of time.
From this point of view, a cylinder impactor would be more threatening to the human
being due to higher acceleration caused.
7.1.7.8 Validation of the simulation results with experiment results
In the current section, the simulated results are put together with correspondent
experiment results to seek out the similarity of them. Figure 7-17 demonstrates the
contact force response of 8L3P60 from 2D FEA and compares it with the results from
experiment, the impact energy are both around 8J with impact velocity at 5.5m/s. It
seems the contact force-time curve shares similarities and their peak contact forces both
climes up around 0.65KN. The striking time for red curve (simulated result) is longer
than that of blue curve (experiment result). And the simulated contact force has
generated a second lower peak contact force while there is only one significant peak
contact force in the experiment curve. The reason to explain this kind of difference
could be that FEA in 2D is a much simple way to evaluate the mechanical performance
of the structure deformation comparing to the real situation, therefore, a perfect match
between experiment and simulation results will not be justifiable due to high complexity
in the experiment testing.
223
Contact Force (8L3P60)
0.8
0.7 Experiment
Contact Force(KN)
0.6 Simulated
0.5
0.4
0.3
0.2
0.1
0.0
0 5 10 15
Time(ms)
Figure 7-17 Comparison of contact force between experiment and simulated (2D)
results for 8L3P60
7.2 FEA of 3D Textile Honeycomb Composites

7.2.1 Creation of the geometric models
In this section, three samples with different cell sizes are modelled in 3D to further
validate their mechanical performance with experiment results.
The actual samples of textile honeycomb composites used for experimental analysis
have been described in Section 4.1.1. The geometric models for FEA are created based
on the fact that the impact is loaded in the centre of the top surface of the honeycomb
model. Figure 7-18 shows a quarter of the created model.
224
Figure 7-18 Created honeycomb model
Due to the tremendous demand on computing resources and timing consuming in

running the models, i.e. one 3D honeycomb model can take up to 3 days to finish one
calculation, it was decided that FEA will only be conducted for model 8L3P60, 8L4P60
and 8L6P60. The dimensions of the models and their cell parameters, and the dimension
of the cylinder impactor are listed in Table 7-10 and Table 7-11.
Table 7-10 Dimension of cylinder impactor
Impact Type: Cylinder

Dimension (mm)
Radium 15
Height 99
225
Table 7-11 Dimension and cell parameter for the models
Cell Configuration Cell θ lb lf tb tf h(single)

Structure (°) (mm) (mm) (mm) (mm) (mm)
8L3P 60 4.07 4.07 0.65 0.97 7.05
8L4P 60 5.08 5.08 0.65 0.97 8.80
8L6P 30 7.62 7.62 0.65 0.97 13.2
where in Table 7-11, θ is the opening angle of the cell; lb is the length of the cell bonded
wall and lf is the length of the cell free wall with the unit of mm; tb and tf means the
thickness of the cell bonded and free wall; h(single) is the height of the single cell with the
unit of mm.
7.2.2 Boundary conditions
The FEA was conducted by the FEA tool (Marc. Mentat, 2005), and it is under the
impact loading along in-plane direction.
Boundary conditions were put along x and y-directions both for honeycomb composite
models and cylinder shaped impactor to restrict the movement of them during impact.
An extra boundary condition was put on the honeycomb composite models at the
bottom surfaces towards z-direction to stop the models falling down during the impact
procedure. The applied boundary conditions are shown in Figure 7-18.
7.2.3 Set-up of 3D FE models
226
A transient dynamic FEA was set up to simulate a transverse impact on the models. The
transient and dynamic nature of the simulation indicates the model behaviour where
time effects play a significant role. The analysis can be adopted to calculate deformation,
contact force and the transient strain and stress distribution within the plate
(Marc.Mentat, 2005a).
The software „MSC.Marc Mentat 2005‟ offers two major approaches to calculate
dynamic performance, namely ‘implicit’ and ‘explicit’. The ‘implicit’ method is used
for relatively large time-steps (>2ms), as it would be the case for a low-velocity impact,
whereas the ‘explicit’ method is characterised by a large number of relatively small time
steps, such as the case for explosive loading. The analysis of the present work was
limited to low-velocity impacts. The levels of the impact energy were selected based on
the capacity of the testing instrument and from the literature (Yu and Chen, 2006). In
order to cover a wide range of impact energies, the FE impact analysis was carried out
at 6J, 8.3J and 10J with a cylinder-shaped impactor at a mass of 0.55kg. From the
experiment results described in Chapter 5, it has shown that an analysis time around
25ms is sufficient to capture the majority important results. Therefore the implicit
‘Single Step Houbolt Operator’ was adopted for the time integration in the present work
and this algorithm is recommended for implicit dynamic contact analysis (Mac Mentat,
2005b).
7.2.4 3D FE Results and Discussions
Transmitted force has been discussed in the previous experiment in Chapter 5 because it
is important in evaluating the protection capability of composites for PPE. In the FEA
part of this thesis, it will be helpful if transmitted force could be also generated and
compared with experiment result to validate and investigate the mechanical
performance of the composite model.
The transmitted force of the simulated models (8L3P60, 8L4P60 and 8L6P60) are
collected and compared with experiment results to evaluate whether the theoretical
results matches the practical results in 3D scale. The data are plotted using the time as
227
the x-axis and they are shown in Figure 7-19(a) and (b). Figure 7-19(c) and (d) compare
the peak transmitted force and arrival time between simulated and experimental results.
Transmitted Force
8l4p 8l6p 8l3p
0.8
Transmitted Force (KN)
0.6
0.4
0.2
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
-0.2
Time(s)
(a) Transmitted force from FEA
Transmitted Force
8L6P60 8L4P60 8L3P60
1.0
0.8
0.6
0.4
0.2
0.0
0.00 0.01 0.02 0.03 0.04
-0.2
Time(S)
(b) Transmitted force from experiment
228
Peak Transmitted Force
By FEA By Experiment
1
Peak Transmitted 0.8
Force(KN)
0.6
0.4
0.2
0
8L3P 8L4P 8L6P
Composites
(c) Peak transmitted force comparison
Peak Arrival Time

By FEA By Experiment
25
Peak Transmitted
Force Arrival
20
Time(ms)
15
10
5
0
8L3P 8L4P 8L6P
Composites
(d) Peak arrival time comparison
Figure 7-19 FEA and experiment results from 3D scale
Transmitted forces from the bottom nodes of the honeycomb composites were collected
to evaluate its force attenuation. Figure 7-19(a) illustrated the theoretical transmitted
229
force results and 7-19(b) shows the comparison experiment results for those honeycomb
composite with different cell size (8L3P of 4.07mm, 8L4P of 5.08mm and 8L6P of
7.62mm).
Figure 7-19(a) depicts that among the modelled samples, 8L3P encounters a much
higher peak force than the rest models. Figure 7-19(b) is the transmitted force collected
from the real experiment which has been described previously in Section 5.5.2.1. The
more detailed comparison of peak transmitted force and peak force arrival time has been
shown in Figure 7-19(c) and Figure 7-19(d). It can be seen that the peak transmitted
force is closely matched for 8L3P and 8L6P and for 8L4P there is a 30% difference
between FE and experiment results. And the simulated peak arrival time is longer than
experiment result and it is shown is Figure 7-19(d).
It seems that the results from FEA and practical work are not exactly identical; therefore,
improvements for FEA are still needed to get more accurate results. The reasons to
cause this could be that there are too many assumptions when setting up the FEA model,
for example: the material of the composite single sheet is assumed to be isotropic
instead of anisotropic as it should be due to the difficulty to get physical properties of
the sheet in the thickness direction; the weave structure of the composite single sheet
are neglected and the unit cell of the composite single sheet is assumed to be beam
structure; the change of the yarn cross section during the interlacing and curvature of
the yarn are also neglected, etc. All these assumptions simplified the FEA model which
decreases the accuracy of the FE results correspondingly. Additionally, the nodes to be
taken to evaluate the transmitted force in the simulation could be different from the
sensor detected from the bottom of the tested sample during the experiment. In view of
this, further investigations are needed in order to improve the simulation accuracy.
230
(a) 8L3P
(b) 8L4P
Figure 7-20 Relationship of input force and transmitted force (a) 8L3P60 and (b)
8L4P60 through FE simulation
231
In order to further investigate how many and how long the input force will be
transmitted through the models, e.g. 8L3P60 and 8L4P60, loading force on the top
surface of the model and transmitted force on the bottom of the sample are captured at
the same time during the dynamic impact. Figure 7-20(a) and (b) illustrates the input
loading force and transmitted force against time as x- axis for 8L3P60 and 8L4P60.
Four nodes on the top centre of the model’s surface are picked to be evaluated as the
input loading force and another four corresponding nodes along the vertical direction on
the bottom of the model are chosen to be collected as the transmitted force. In Figure 7-
20(a) and (b), the green line shows how the loading force reacted on the sample from
the beginning to the end of the impact and the blue line indicates how the transmitted
force attenuated on the bottom of the sample towards the end of the impact. Take
8L3P60 as an example, when the impactor stoke on the surface of the honeycomb
composite, there is a loading force occurring and it climbs up to its peak point at 6.5KN
at 2.6ms. This peak loading force has travelled through the model and showing up as
peak transmitted force in the blue line in Figure 7-20(a) and the travelling time is 4ms.
The difference of the peak force in the green and blue line could be regarded as how
much force has been dissipated inside the model and for 8L3P, there are 5.65KN force
has been avoided to be transmitted underneath. In Figure 7-20(b) for the model 8L4P,
the peak force of the green line is 6.89KN and the arrival time is 3.8ms and the peak
force in the blue line is 0.35KN at 7.2ms. The shifting of this peak force in the green
line to the blue line means it takes 3.4ms for the 8L4P to prevent 6.54KN force to go
underneath.
The shape of green line (loading force) and blue line (transmitted force) in Figure 7-
20(a) and (b) also indicates that during the impact, after the loading force goes up to its
peak point, it starts to decrease and this amount of force has transmitted to the bottom of
the sample vertically, therefore, the blue line (transmitted force) began to increase.
When the blue line (transmitted force) reaches its maximum value, it started to decrease
while the green line (loading force) shows the trend to increase again, which means less
force could dissipate through the structure vertically.
232
Therefore, from comparison of the contact force on the top surface and transmitted
force on the bottom surface of the composite between 8L3P and 8L4P, it can again
determine that big cell size honeycomb structure composite can delay the transmitted
time and less force will be transferred through the impacted body.
However, the generated contact force from FEA is too high comparing to the real
experiment results from Chapter 5 and the reasons to cause this could be the 4 nodes
that are picked to collect the contact force is centrally underneath the impactor which
encounters the highest contact force or it could be the setting of the material properties
for the composite single sheet not very accurate which can’t exactly reflect the real
sample’s material property. Further measures are expected to be taken to improve the
accuracy of FEA in the future work.
7.3 Summaries on FEA
This part of the thesis presents a research on the analysis of textile honeycomb structure
by using FE tool. Simulations have been carried out to validate the existing experiment
results. Mac Mentat (2005a) has shown to be an efficient tool in evaluate the structure
deformation and other mechanical performance such as contact force of this kind of
structure in terms of various geometric parameters of the cell.
Result from the FEA in 2D and 3D firstly revealed that the simulated results agrees
well with the experiment results although there are some inaccuracy occurs due to the
simplification of the simulated model compare to the real textile honeycomb composites.
The first findings by conducting FEA is that samples with medium cell size such as
8L5P owns a better damage tolerance than the rest samples besides its favourable
energy absorption and force attenuation performances concluded from experiments. It
also seems when the cell size is getting bigger such as models of 8L6P, under whatever
allowable impact energy, the damage of the model is similar. Regarding the opening
angle of the honeycomb models, it seems that when the opening angle exceeds 75º, the
impact energy would not cause too much difference in the structure damage either.
233
It also found out that if the bonded wall of the honeycomb model is increasing a little bit,
i.e. 8L(4+3)P60, it helps reducing the structure deformation while doesn’t affect its
energy absorption capability. Therefore, from FEA results, it reveals that models with a
lb
considerable show a better damage tolerance than the rest models.
lf
Comparing the honeycomb models being impacted by cylinder and ball shaped objects,
the FEA results reveals that under cylinder impact, the model penetrates more loading
force at the beginning of the impact and this could cause a faster rate of deformation at
the intitial stage of the impact and it results a faster strain energy induction which
accelerates the item underneath more. The contact force are characterised by a higher
fluctuation in the ball impact and this indicates there are more cell materials are
involved for stopping the impactor and this could lead to a wider energy dissipation and
reduced transmitted force and lower acceleration underneath. Therefore, the conclusion
is that a cylinder impactor would be more threatening to the human being than that of
the ball impactor.
234
CHAPTER 8
CONCLUSIONS AND FUTURE WORK
8.1 Conclusions
The main aim of this research is to investigate how geometric and structural parameters
of textile honeycomb composites would affect its mechanical performance and energy
absorption capability under low velocity impact. The objectives set out for this PhD
research include (1) to develop an effective method to produce 3D honeycomb fabrics
and to establish a procedure to convert the fabrics into textile honeycomb composites,
which will be able to guide future practical production processes for similar composites,
(2) to design and optimize the geometric parameters for the 3D honeycomb structure
which including its cell height, cell size, cell opening angle and length ratio of cell
walls, (3) to conduct the low velocity impact tests which includes setting up the testing
equipment in order to obtain the associated data for analyzing the mechanical properties
and energy behaviour of the composites followed by analysing the experimental data to
seek out the relationship between textile honeycomb composite and its mechanical
performances (4) to create geometrical models for the textile honeycomb composites to
examine their mechanical and energy absorption behaviour through FEA tool. The main
achievements from the research are concluded as follows.
a. Produce 3D honeycomb fabrisc and development of a procedure to manufacture the

textile honeycomb composites
In the current study, ten 3D honeycomb fabrics were successfully produced and a set of
apparatus has been developed to allow the fabrics to be opened into three dimensions
and the geometric parameters for the textile honeycomb composites can be created by
adjusting the height of the apparatus. A detailed procedure has therefore been made
available for fabrics and apparatus handling in order to get desired cell structure. The
235
new established route for making textile honeycomb composite through the research
will be able to guide future practical production processes for similar composites. A
total number of 14 textile honeycomb composites have been successfully produced
throughout the study and this built up the foundations for the experimental
investigations of the textile honeycomb composites
b. Low velocity impact test for the textile honeycomb composites
Two types of impact equipment were used for the impact test on textile honeycomb
composites, i.e. the dropping hammer system and the Instron Dynatup 8200, both are
able to run low velocity impact test. Impact tests on the textile honeycomb composites
were conducted in pre-defined groups including different composite cell size, cell
opening angle, length ratio of cell walls and composites with similar thickness. Data
from the experiments were studied, and analysed results reveals, giving the same
impact energy, the geometric parameters affects the composites’ mechanical
performances significantly. The detailed findings will be specified as following:
1). the opening angle is a key factor affecting the performance of the honeycomb
composites. When the composites are strong enough to resist the incoming force,
lower opening angle resulted in relatively higher energy absorption in vertical
deformations but also higher peak transmitted forces. Composites with a small to
medium opening angle are very sensitive to the heavy weight impacts and they tend
to be easily destroyed comparing to those composites with large opening angles.
2). generally speaking, the bigger the cell size is, the more force will be attenuated
on the composites and the higher the energy will be absorbed. The effect was
proved to be quite significant during the impact. However, composites with big cell
sizes are more easily to be destroyed under heavy weight impacts and the thickness
of composites is also significantly increased.
3). by adjusting the length of bonded and free walls, the mechanical performances
of the composites will be affected. Results indicated that better mechanical
236
performances of the honeycomb structure can be achieved by increasing the bonded
wall length but keeping the same free wall length. This method can be used to
improve the mechanical performances of the composite material without changing
the actual weight and volume of the composites. Also, composites will longer
bonded wall lengths are harder to be deformed under heavier impact loadings.
4). given a desired thickness of the composites to design, effects from changing of
the number of layers are strongly compensated by the effects from cell size change.
Composites with too few number of layers with larger cell size results in weak
resistance to impact and will be easily crushed.
c. FEA on honeycomb textile composites
In the current research, a methodology has been established by using finite element
method (FEM) to investigate the textile honeycomb composites more systematically.
The loading conditions for the honeycomb composites have been assumed to be at
different impact energy levels (6J, 8.3J and 10J) and under different impactor shapes
(cylindrical and spherical) in the FEA. The findings through the investigation are as
follows:
1). the results from FEA are in good agreement with experimental results. This
provides a quick way to access the performances of the textile honeycomb
composites without actually producing the composites and carrying out the physical
tests in practice.
2). it is also found that under whatever impact energy levels (6J, 8.3J and 10J), the
cell size of the composites sufficiently affects the mechanical performance of the
honeycomb structure. The analysis shows that with similar energy absorption and
force attenuation performance, composites with medium cell size owns a better
damage tolerance. Another finding of FEA is that if the bonded wall of the
237
honeycomb composites are increased a little bit, it can help to reduce the structure
deformation while doesn’t affect its energy absorption capability the same time.
3). more important to discover from FEA is that when the honeycomb structure is
under cylindrical impacts, the structure destroys more comparing to that it is under
ball shaped impact. This will accelerates the protected item underneath more and
cause more damages. Therefore, the conclusion is that a cylinder impactor would be
more threatening to the human being than that of the ball impactor. By knowing
these information before hand, FEM allows the entire composite design to be
constructed, refined, and optimized before the design is manufactured, which can
accelerate the testing and development of the textile honeycomb composites.
To summarize, the FEM allows detailed visualization of the textile honeycomb

composites as where the structures are deformed and indicates the distribution of
stresses and displacements of the impacted models. It provides a faster and economic
design cycle, which substantially decreases the time to take products from concept to
the production lines thus to increase the production and increases revenue too.
8.2 Recommendations for Further Research Work
A number of different future works are possible to be investigated to extend the current
findings to a higher level of discovery.
In fabric consolidation, an improved apparatus is required in future work. Due to the

limitations of lab facilities, it is hard for resin impregnating to achieve perfect
specimens with even resin distribution throughout the composite. The designed
thickness, length and angle of cell structure were hard to be achieved exactly. Re-resin
is surely an important factor taking effect on the following mechanical tests. Irregular
cell structure and uneven resins on sides of the specimen might cause fluctuations of the
experimental results. Further investigations to avoid such negative effects should be
explored.
238
In addition, filament yarns that are formed by E-glass fibre or carbon fibre etc. are
usually used in the composite material instead of cotton staple fibre yarns in the type of
applications investigated in this thesis. For filament yarns, their Young’s modulus (E) is
much higher than cotton staple fibre yarns (Yu and Chen, 2006; Tan and Chen, 2005),
therefore, the limitations to use cotton yarns is that it can provide a softer cell wall
material even it is impregnated with same epoxy resin and this will affect the
mechanical performances of PPE compared to the composites made by filament
yarns/epoxy resin. In the future work, filament yarns should be chosen to weave the
honeycomb fabric in the application of PPE as they are more commonly used in fibre
reinforced composite.
For mechanical tests, this study focused on the in-plan loading tests using dropping
hammer system and Instron Dynatup 8200 system. But for real violent crowed
management situations, the impact can take place at any speed and in any directions.
The mechanisms how PPE can resist impact by various kind of foreign objectives are
still of scientific interests and demand lots of experimental investigations or model
simulations in future. In-plane impact experiments in X1 direction or out-plane impact
experiments in X3 direction (see Figure 2-5) can be expected in future too.
During mechanical tests, strong fluctuations of datasets during mechanical tests were
observed. To make results more comparable, same dropping distance to reach the same
initial impact velocity is essential, given an ideal friction negligible tube track. Since
the initial impact velocity depends on the height from free dropping position to
specimen top surface and the anvil is not changeable, it will be more flexible to design a
system with adjustable dropping position to ensure the comparability of initial impact
velocity of different composite types with different thickness. The sensitivity settings of
electronic instruments, such as data recorder and charge amplifiers are essential to
record reasonable good datasets. Give the experience learned in this study, it might be
easier for future work to be carried out with more attention.
Additionally, among ten specimens which have been tested for each type of composite,
only results from three tests with the most repeatability were chosen for the data
239
analysis. This is not an ideal solution as the composites are made with hand brushing
which leads to a large variation in their mechanical performances. In the future, in order
to improve the confidence in the data, a standard data processing procedure dealing
with the collection, analysis, interpretation and presentation of numerical data should be
applied to the experiments.
Although some parametric study was successfully conducted in this research, however,
there are more parameters such as length ratio of bonded and free wall higher than 2:1,
change of different yarns with different density to make fabric with different wall
thickness, etc., can be investigated experimentally and theoretically in the future.
FE analysis has been conducted mainly in 2D, and more 3D models can be analysed to
validate the experiment results or for more deep investigation. The material properties
of the model is set to be isotropic rather than anisotropic as it should be and more
measurement could be done to get the material property of the cotton/epoxy single layer
sheet when more advanced testing equipment is invented. The weave structure is plain
in the current works, however, other weave structure could be modelled in the future to
investigate how the structure of the fabric influences the honeycomb composites
performance too.
The studies proved the feasibility of all engineering procedures and achieve good
experience in practice. It will be very helpful to carry out the above recommended
further investigations in the future.
240
REFERENCE
Aare,M. and Kleiven,S. (2007). Evaluation of head response to ballistic helmet impacts
using finite element method. International Journal of Impact Engineering, vol.34, no.3,
March, pp. 596-608
Abbadi,Ahmed., Kourtsawa,Y., Carmasol,A., Belouettar,S. and Azari,Z. (2009).

Experimental and numerical characterization of honeycomb sandwich composite panels.
Simulation Modelling Practice and Theory, vol. 17, no.10, November, pp.1533-1547
Abd El-Sayed,F.K., Jones,R. and Burgess,I.W. (1979). A theoretical approach to the

deformation of honeycomb based composite materials. Composites, vol.10, no.4, October,
pp.209-214
Abramowicz,W. and Jones,N. (1984a). Dynamic axial crushing of square tubes,

International Journal of Impact Engineering, vol.2, no.2, pp.179-208
Abramowicz,W. and Jones,N. (1984b). Dynamic axial crushing of circular tubes,

International Journal of Impact Engineering, vol.2, no.3, pp.263-281
Abramowicz,W. and Wierzbicki,T. (1989). Axial crushing of multicomer sheet metal

columns, ASME Journal of Applied Mechanics, vol.56, pp.113-120
Abrate,S. (1998). Impact on composite structures. Cambridge University Press,

Cambridge, UK
Anderson,T. and Madenci,E. (2000). Experimental investigation of low-velocity impact

characteristics of sandwich composites. Composite Structures, vol.50, pp.239-247
Andersson,J. and Van den,B.B. NDI on Bonded Sandwich Structures with Foam Cores
and Stiff Skins – Shearography the Answer?, Application Report No. 02.01,
http://www.ndt.net/article/wcndt00/papers/idn389/idn389.htm
Antonio,M. (2000). 3-D textile reinforcements in composite materials. Woodhead

Publishing Ltd, UK. pp.145
Ashby,M.F. and Mehl Medalist,R.F. (1983). The mechanical properties of cellular solids.
Metallurgical and Materials Transactions A, vol.14, no.9, September, pp.1755-1769
ASTM D3039-95 – Standard. (2004). Standard Terminology for Composite Materials.
241
Bajaj,P. and Sengupta,A.K. (1992). Protective Clothing: a critical appreciation of recent
developments. Textile Progress, Journal of the Textile Institute, vol.22, no2, pp.105-132
Banhart,J. and Baumeister,J. (1998). Deformation characteristics of metal foams,

Journal of Material Science, vol.33, pp.1431-1440
Barma,P., Rhodes,M.B. and Salovey,R. (1978). Mechanical properties of particulate-

filled polyurethane foams. Journal of Applied Physics, vol.49, no.10, pp.4985-4991
Bartels,V.T. (2003). Thermal comfort of aeroplane seats: Influence of different seat

materials and the use of laboratory test methods, Applied Ergonomics, vol. 34, pp.393-
399
Bassett,R.J., Postle,R., and Pan,N., (1999). Experimental methods for measuring fabric
mechanical properties: A review and analysis. Textile Research Journal, vol.69, pp.866
Baumeister,J., Banhart,J. and Weber,M. (1997). Aluminium foams for transport industry.
Materials & Design, vol.18, no.4/5, pp.217-220
Bibo,G.A. and Hogg,P.J. (1996). Review the role of reinforcement architecture on impact
damage mechanisms and post-impact compression behaviour. Journal of Material
Science, vol.31, pp.1115-1137
Bitzer,T.N. (1997). Honeycomb Technology: Materials, Design, Manufacturing,

Applications and Testing. Chapman &Hall, London, UK
Buesqen,A. (1995). Shape Weaving New Invention for Producing 3D Shaped Woven
Shells. Technical Textile, vol.38, September, pp.E28-E29
Chen,X., Gisbert,P.J. and Mollst,S.P. (1999). Experimental studies on the structure and
mechanical properties of multi-layer and angle interlock woven structure, J. Text Inst.,
vol.90, no.1, pp. 91-99
Chen,X., Knox,R.T., McKenna,D.F. and Mather,R.R. (1996). Automatic generation of

weaves for the CAM of 2D and 3D woven textile structures. Journal of Textile Institute,
vol.87, pp.356-370
Chen,X., Ma,Y.L. and Zhang,H. (2004). CAD/CAM for cellular woven structure.
Journal of the Textile Institute, vol.95, no.1-6, pp.229-241
Chen,X. and Potiyaraj, P. (1999a). CAD/CAM for complex woven fabrics part II: multi-
layer fabrics. Journal of Textile Institute, vol.90, pp.73-90
Chen,X. and Potiyaraj, P. (1999b). CAD/CAM of orthogonal and angle-interlock woven

structures for industrial applications. Textile Research Journal, vol.69, pp.648-655
242
Chen,X., Sun,Y. and Gong,X. (2008). Design, manufacture and experimental analysis of
3D honeycomb textile composite, Part I. Textile Research Journal, vol.78, no.9, pp.771-
781
Chen,X. and Wang,H. (2006). Modelling and computer-aided design of 3D hollow
woven reinforcement for composites. Journal of the Textile Institute, vol.97, no.1,
January, pp.79-87
Crane,R.M. and Camponeschi,J.E.T. (1986). Experimental and analytical

characterization of multidimensionally braided graphite/epoxy composites. Experimental
Mechanics, September, pp.259-266
Davies,G.A.O., Hitchings,D., Besant,T., Clarke,A. and Morgan,C. (2004). Compression

after impact strength of composite sandwich panels. Composite Structure, vol.63, no.3,
pp.1-9
DeBonis,A.L. and Bodig,J. (1975). Nailed Wood Joints under Combined Loading. Wood
Science and Technology, vol.9, pp.129-144
Delfosse.D., Pageau.G., Bennett.R., and Poursartip.A. (1993). Instrumented Impact

Testing at High Velocities, Journal of Composite Technology & Research, vol.15, no.1,
pp.1-8
Dexter,H.B. (1996). Innovative textile reinforced composite materials for aircraft

structures, Proceedings of the 28th International SAMPE Technical Conference, 4-7 Nov,
pp.404-416
Dickinson,L., Mohamed,M. and Bogdanovich,A. (1999). 3D Weaving: What, How, and

Where. In: 44th International SAMPE Symposium, pp.303-312, May, CA
Dionne,J.P., EI Maach,I., Shalabi,A. and Madris,A. (2003). A method for assessing the
overall impact performance of riot helmets. Journal of applied biomechanics, vol. 19, pp.
246-254
Easterling,K.E., Harrysson,R., Gibson,L.J. and Ashby,M.F.(1982). On the mechanics of

balsa and other woods. Proceedings of the Royal Society of London, no.A383,pp.31-41
Flanagan,M.P., Zikry,M.A., Wall,J.W. and EL-Shiekh,A. (1999). An Experimental

Investigation of High Velocity Impact and Penetration Failure Modes in Textile
Composite. Journal of Composite Materials, vol.33, pp.1080
Gent,A.N. and Thomas,A.G. (1959). The Deformation of Foamed Elastic Materials.

Journal of Applied Polymer Science, vol.I, no.1, pp.107-113
Glicksman,L., Schuetz,M. and Sinofsky,M. (1987). Radiation heat transfer in foam

insulation. Int.J.Heat Mass Transfer, vol.30, no.1, pp.187-197
243
Gibson,L.J. and Ashby M.F.,(1997) Honeycomb solids: structure and properties – 2nd
edition, Cambridge University Press, Cambridge, UK
Gibson,L.J. and Ashby,M.F. (1982). The mechanics of three-dimensional cellular

materials. Proceedings of the royal society, vol.A382, pp.43-59
Gibson,L.J., Ashby,M.F., Zhang,J., and Triantafillou,T.C., (1989). Failure surfaces for

cellular materials under multiaxial loads-I.Modelling. International Journal of
Mechanical Sciences, vol.32, no.9, pp.635-663
Gong,J.C. and Sankar,B.V. (1991). Impact properties of three-dimensional braided

graphite/epoxy composites. Journal of Composite Materials, vol.25, pp.715-731
Gupta,N.K., Prasad,G.L. and Gupta,S.K. (1997). Plastic collapse of metallic conical

frusta of large semi-apical angle. International Journal of Crashworthiness, vol.2, no.4,
pp.349-366
Hernalsteen,P. and Leblois, L.C. (1976). The use of energy absorbers to protect
structures against impact loading. Nuclear Engineering and Design, vol.37, no.3, June,
pp.373-406
Herrmann,A.S., Zahlen,P. and Zuardy,I. (2005). Sandwich structures technology in

commercial aviation. In: Thomas,O.T., Bozhevolnaya,E. and Lyckegaard,A. eds. 7th
International Conference on Sandwich Structures(ICSS-7), August, Aalborg, Denmark,
pp.13-26
Hibbitt, Karlsson and Sorensen. (1998). Inc. ABAQUS/Explicit User‟s Manual, vol.5.8
Hilyard,N.C. (1982). Mechanics of Cellular Plastics. Ed., Macmillan, New York, U.S.A
Hoff Michel. (1951). http://plants,jstor.org/person/bm000024979
Hoskins,J.A. (1983). Congressus Numeratium, 40, pp.63
Ishikawa,T., Matsushima,M. and Hayashi,Y. (1987). Geometrical and material non-

linear properties of two-dimensional fabric composites. AIAA Journal. Vol.25, pp. 107–
113
James,M.G. and Stephen,P.T. (2001). Mechanics of Materials, Nelson Thornes, USA,

pp.894
Johnson,W., Reid,S.R. and Reddy,T.Y. (1977). The compression of crossed layers of

thin tubes. International Journal of Mechanical Sciences, vol.19, no.7, pp.423-428
244
Jones,N. (1989). Structural Impact, Cambridge University Press, Cambridge, UK
Keef,M., Edward,D.C. and Yang,J. (1992). Solid Modeling of Yarn and Fiber
Assemblies. Journal of the Textile Institute, vol.83, no.2, pp.185-196
Kenny,L.D.(1996). Mechanical properties of particle stabilized aluminium foam,

Material Science Forum, vol.217-222, pp.1883-1890
Kim,C.G. and Jun,E.J. (1992). Impact Resistance of Composite Laminated Sandwich

Plates. Journal of Composite Materials, vol.26, pp.2247
Kim,J.S. and Chung,S.K. (2007). A study on the low-velocity impact response of

laminates for composite railway bodyshells. Composite Structures, vol.77, pp.484-492
Kleineberg,M., Wenner,U. and Hanke,M. (2002). Cost effective CFRP-Fuselage

Manufacturing with Liquid Resin Infusion (LRI) – Technologies. Workshop at German
Aerospace Centre (DLR) on Final Project of Black Fuselage, Braunschweig, Germany
Knoell,A.C. (1966). Environmental and physical effects on the response of balsa wood as
an energy dissipater. JPL Technical Report, no.32-944, California Institute of
Technology, Pasadena, CA
Ko,F.K. (1989). Three-dimensional fabrics for composites, Chapter 3 in Textile

Structural Composite, edited by Chou, T.W. and Ko., F.K., Netherlands, pp.129-171
Kobayashi,H., Daimaruya,M. and Kobayashi,T. (1998). Dynamic and static compression

tests for paper honeycomb cores and absorbed energy. JSME international journal. Series
A, Solid mechanics and material engineering, vol.41, no.3, pp.338-344
Kurauchi,T., Sato,N., Kamigaito,O., and Komatsu,N. (1984). Mechanism of high energy

absorption by formed materials-foamed rigid polyurethane and foamed glass. Journal of
Materials Science, vol.19, pp.871-880
Ley,R., Lin,W.C. and Uy,M. (1999). Facesheet Wrinkling in Sandwich Structures.

NASA/CR-1999-208994, Virginia, USA
Maiti,S.K., Gibson,L.J. and Ashby,M.F. (1984). Deformation and energy absorption

diagrams for cellular solids. Acta metal, vol.32, no.11, pp.1963-1975
Marc Mentat. (2005a). Volume B: Element Library, MSC.Marc, Version 2005, April,
USA
Marc Mentat. (2005b). Volume A: Theory and User Information, MSC Marc, Version
2005, April, USA
245
McFarland,R.K. (1963). Hexagonal cell structures under post-buckling axial load, A1AA
Journal, vol.1, no.6, pp.1380-1385
Med-Eng System Inc, (2001). The blunt impact performance of V-Top equipment, Med-
Eng System Inc
Meo,M., Morris,A.J., Vignjevic,R. and Marengo,G. (2003). Numerical simulations of

low-velocity impact on an aircraft sandwich panel. Composite Structues, vol.62, pp.353-
360
Micco,C.D. and Aldao,C.M. (2006). On the Prediction of the Radiation Term in the
Thermal Conductivity of Plastic Forms. Latin American Applied Research, vol.36,
pp.193-197
Miltz,J., Ramon,O. and Mizrahi,S. (2003). Mechanical behaviour of closed cell plastic
foams used as cushioning materials. Journal of Applied Polymer Science, vol.38, no.2,
pp.281-290
Mohamed,M.H. (1990). Three-dimensional Textiles, American Scientist, vol.78,

Nov/Dec, pp. 530-541
Mouritz,A.P., Bannister,M.K., Falzon,P.J. and Leong,K.H. (1999). Review of

applications for advanced three-dimensional fibre textile composites. Composites Part A:
Applied Science and Manufacturing, vol.30, no.12, December, pp.1445-1461
Odgaard,A. and Linde,F. (1991). The Underestimation of Young’s Modulus in

Compressive Testing of Cancellous Bone Specimens. J. Biomechanics, vol.24, no.8,
pp.691-698
Pan,N. (2003). Relationship between grab and strip tensile strengths for fabrics with
roughly linear mechanical behaviour. Textile Research Journal, vol.73, no.2, pp165-171
Papka,S.D. and Kyriakides,S. (1998). Experiments and full-scale numerical simulations

of in-plane crushing of a honeycomb. Acta Materialia, vol.46, no.8, May, pp.2765-2776
Pflug,J., Verpoest,I. and Vandepitte,D. (1999). Folded Honeycomb Cardboard and Core
Material for Structural Applications. Sandwich Construction 5, pp.1-12
Pflug,J., Fan,X.Y., Vangrimde,B. and Verpoest,I. (2002). Development of a sandwich

material with polypropylene/natural fibre skins and paper honeycomb core. Proceeding of
10th European Conference on Composite Materials (ECCM-10), pp.331
Pflug,J. and Vangrimde,B. (2003). New Sandwich Material Concepts – Continuously

Produced Honeycomb Cores, COMPOSIT Workshop on New Material Concepts for
Composite Transportation Structures, Leuven, Belgium
246
Prashant Kumar and Badri Rai. (1993). Delaminations of barely visible impact damage
in CFRP laminates. Composite Structures, vol.23, pp.313-318
Qiu,Y.P., Xu,W., Wang,Y.J., Mohammed,A. and Zikry,M.H.M. (2001). Fabrication and

characterization of three-dimensional cellular-matrix composites reinforced with woven
carbon fabric. Composites Science and Technology, vol.61, pp.2425-2435
Reddy,T.Y. and Reid,S.R. (1980). Phenomena associated with the crushing of metal
tubes between rigid plates. International Journal of Solids and Structures, vol.16, no.6,
pp.545-562
Reid,S.R. and Peng,C. (1997). Dynamic uniaxial crushing of wood. International

Journal of Impact Engineering, vol.19, no.5-6, May-July, pp.531-570
Reid,S.R. and Bell,W.W. (1982). Influence of strain hardening on the deformation of thin
rings subjected to opposed concentrated loads. International Journal of Solids and
Structure, vol.18, no.8, pp.643-658
Richardson,M.O.W. and Wisheart,M.J. (1996). Review of low-velocity impact

properties of composite materials. Composites Part A: Applied Science and
Manufacturing, vol.27, no.12, pp.1123-1131
Ruan,D., Lu,G., Chen,F.L. and Siores,E. (2002). Compressive behaviour of aluminium

foams at low and medium strain rates. Composite Structures, vol.57, no.1-4,July,pp.331-
336
Ruan,D., Lu,G., Wang,B. and Yu,T.X. (2003). In-plane dynamic crushing of

honeycombs-a finite element study. International Journal of Impact Engineering, vol.28,
no.2, February, pp.161-182
Schmueser,D.W. and Wickliffe,L.E. (1987). Impact energy absorption of continuous

fiber composite tubes. Journal of Engineering Materials and Technology, vol.109, no.1,
January, pp.72-78
Schubel,P.M., Luo,J.J. and Daniel,I.M. (2005). Low velocity impact behaviour of

composite sandwich panels. Composite: Part A, vol.36, pp.1389-1396
Schwabr,D.M. (1973). Impact Behaviour of Polymeric Foams: A Review. Polymer-

Plastics Technology and Engineering, vol.2, no.2, pp.231-249
Shaw,M.C. and Sata,T. (1966). The plastic behaviour of cellular materials. International
Journal of Mechanical Sciences, vol.8, no.7, July, pp.469-472
Shih,W.K. and Jang,B.Z. (1989). Instrumented Impact Testing of Composite Sandwich

Panels. Journal of Reinforced Plastics and Composites, vol.8, pp.270
247
Shin,K.B., Lee,J.Y. and Cho,S.H. (2008). An experimental study of low-velocity impact
responses of sandwich panels for Korean low floor bus. Composite Structure, vol.84,
pp.228-240
Soden,P.D. and Mcleish,R.D. (1976). Variables affecting the strength of balsa wood.
Journal of Strain Analysis, vol.114, pp.225-234
Sun,B.Z., Hu,H. and Gu,B.H. (2010). Response of 3D biaxial spacer weft-knitted

composite circular plate under impact loading. Part I: uni-cell and elasto-plastic
constitutive model. The Journal of The Textile Institute, vol. 101, no.1, January, pp.28-34
Sun,Y. (2005). Engineering of 3D textile honeycomb composites for energy absorption.

MPhil thesis, University of Manchester, UK
Tan,X. and Chen,X. (2005). Parameters affecting energy absorption and deformation in
textile composite cellular structures. Material&Design, vol.126, no.5, pp.424-438
Tan,X., Chen,X.,Conway,P.P. and Yan,X. (2007). Effects of plies assembling on textile

composite cellular structures. Material&Design, vol.28, pp.857-870
Tang,Y., Sun,B., Ding,X. and Gu,B. (2008) Mechanical properties of 3-D glass/polyester
resin cellular woven composite under impact loading. Pigment & Resin Technology,
vol.37, no.6, pp.410-415
Torre,L. and Kenny,J.M. (2000). Impact testing and simulation of composite sandwich
structures for civil transportation. Composite Structure, vol.50, pp.257-267
Takenaka et al. (1991). Woven fabric having multilayer structure and composite
structure comprising the woven fabric, US Patent No.5021283
Ujihashi.S. (1993). An intelligent method to determine the mechanical properties of

composites under impact loading. Composite Structures, vol.23, no.2, pp.149-163
Vavilov,V.P., Klimov,A.G., Nesteruk,D. and Shiryaev,V.V. (2003). Detecting water in

aviation honeycomb structures by using transient infrared thermographic NDT. In:
Cramer,K.E. and Maldague,X.P. eds. Proceedings of the international society for optical
engineering (SPIE), vol.5073, April, pp.245-255
Vural,M. and Ravichandran,G. (2003a). Dynamic response and energy dissipation

characteristics of balsa wood: experiment and analysis. International Journal of Solids
and Structures, vol.40, no.9, May, pp.2147-2170
Vural,M. and Ravichandran,G. (2003b). Microstructural aspects and modelling of failure

in naturally occurring porous composite. Mechanics of Materials, vol.35, pp.523-536
248
Wadley,H.N.G. (2002). Cellular Metals Manufacturing. Advanced Enginering Materials,
vol.4, no.10, pp.726-733
Wadley,H.N.G., Fleck,N.A. and Evans,A.G. (2003). Fabrication and structural

performance of periodic cellular metal sandwich structures. Composite Science and
Technology, vol.63, pp.2331-2343
Wang,D.M. (2009). Impact behaviour and energy absorption of paper honeycomb

sandwich panels. International Journal of Impact Engineering, vol.36, pp.110-114
Wang,Y.J. and Zhao,D.M. (2006) Effect of fabric structures on the mechanical

properties of 3-D textile composites. Journal of Industrial Textile, vol.35, no.3, January,
pp239
Whitney,J.M., Stansbarger,D.L. and Howell,H.B. (1971). Analysis of the rail shear test-
applications and limitations. Journal of Composite Materials, vol.5, pp.24-34
Wierzbicki,T. (1983). Crushing analysis of metal honeycombs. International Journal

Impact Engineering, vol.1, no.2, pp.157-174
Wierzbicki,T. and Abramowicz,W.(1983). On the crushing mechanics of thin-walled

structures, ASME Journal of Applied Mechanics, no.50,pp.727-734
Wong,Richard. (1992). Sandwich construction in the starship. Proceedings of the 37th

International SAMPE Symposium, vol.9-12, March, pp.186-197
Wu,E. and Wu,S.J. (1997). Axial crush of metallic honeycombs. International Journal of
Impact Engineering, vol.19, no.5-6, May-July, pp.439-45
Wu,H. (2003). Experimental investigation into woven cellular composite for impact
energy absorption. MSc dissertation, UMIST, Manchester, UK
Xu,H.Y. (1992). Computer representation of fabric structure. MSc thesis, UMIST,

Manchester, UK
Yamashita,M. and Gotoh,M. (2005). Impact behavior of honeycomb structures with

various cell specifications-numerical simulation and experiment. International Journal of
Impact Engineering, vol.32, no.1-4, December, pp.618-630
Yassar,Mehmet. (1999) Chapter 3: consolidation of fabric. Final year project report, the
university of Manchester. Manchester, UK
Yasui,Yoshiaki. (2000). Dynamic axial crushing of multi-layer honeycomb panels and

impact tensile behaviour of the component members. International Journal of Impact
Engineering, vol.24, no.6-7, July, pp.659-671
249
Yee,C.C. and Eduardo,D.G. (1983). Simultaneous Conduction and Radiation in Porous
and Composite Materials: Effective Thermal Conductivity. Ind.Eng.Chem.Fundam,
vol.22, pp.276-282
Yi,H.L. and Ding,X. (2004), Conventional approach on manufacturing 3D woven

performs used for composites, Journal of industrial textiles, vol.34, no.1, pp. 39-50
Yu,Daniel.K.C. and Chen,X. (2006). Simulation of trauma impact on textile reinforced

cellular composites for personal protection. First year report, University of Manchester,
UK
Zhang,J. and Ashby,M.F. (1992). The out-of-plane properties of honeycombs,

International Journal of Mechanic Science, vol.34, pp.491-509
Zhao,H. and Gary,G. (1998). Crushing behaviour of aluminium honeycombs under

impact loading. International Journal of Impact Engineering, vol.21, no.10, November,
pp.827-836
Zheng,Z., Yu,J. and Li,J. (2005). Dynamic crushing of 2D cellular structures: A finite
element study. International Journal of Impact Engineering, vol.32, no.1-4, December,
pp.650-664
Zic,I., Ansell,M.P., Newton,A. and Price,R.W.(1990). Mechanical properties of

composite panels reinforced with integrally woven 3-D fabrics. Journal of the Textile
Institute, vol.4, pp.461-479
Zok,F.W., Rathbun,H.J., Wei,Z. and Evans,A.G. (2003). Design of metallic textile core
sandwich panels. International Journal of Solids and Structures, vol.40, no.21, pp.5707-
5722
Zupan,M., Deshpande,V.S. and Fleck,N.A. (2004). The out-of-plane compressive

behaviour of woven-core sandwich plates. European Journal of Mechnics A/Solids,
vol.23, no.3, pp.411-421
250

Full Text

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Full Text

Uploaded by

Copyright:

Available Formats

INVESTIGATION OF DIFFERENT GEOMETRIC

STRUCTURE PARAMETER FOR HONEYCOMB

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

1.1 Description of the problem....................................................... 23

CHAPTER 2 LITERATURE REVIEW........................................................ 28

2.1 Classification of Cellular Solids……………………………... 28

CHAPTER 3 DESIGN OF 3D HONEYCOMB FABRICS............................ 70

3.1 Design of 3D honeycomb weaves................................................ 70

3.2.2.4 Similar sample thickness with different cell size........................ 85

CHAPTER 4 CREATION OF HONEYCOMB COMPOSITES AND TEST

CHAPTER 5 EXPERIMENTAL DATA ANALYSIS ON TEXTILE

5.1 Low velocity drop weight impact tests……………………….. 109

P60,8L(4+6)P60,8L6P60) …………………………….………… 140

CHAPTER 6 EXPERIMENTAL DATA ANALYSIS ON TEXTILE

6.1 Low velocity impact test setting by Instron Dynatup Model

6.1.1 Assembly of Instron Dynatup Model 8200 drop weight impact

6.1.2 Testing procedure………………………………………………... 159

performance ( lb  1 : 8L(3+6)P60, 8L(4+6)P60, 8L6P60;

6.2.4 Comparison of the results between two different loading

( lb  1 : 8L(3+6)P60, 8L(4+6)P60, 8L6P60;

6.3 Summaries……………………………………………………… 181

CHAPTER 7 FEA ON TEXTILE HONEYCOMB COMPOSITES……….. 182

7.1 FEA Based on 2D Honeycomb Composite Models…………… 183

CHAPTE 8 CONCLUSIONS AND FUTURE WORK…………………….. 234

8.1 Conclusions…………………………………………………… 234

(iv) X.Chen and X.Gong.(2008). Manufacture and Characterization of Exatra-

(v) Xiaogang Chen, Xiaozhou Gong and Shijun Tang.(2008). Design,

Figure 2-1 Examples of cellular solids (Gibson and Ashby, 1997) 30

Figure 4-9 Specimens with same thickness 107

Figure 6-10 Contact force-displacement curve and energy absorption diagram

Figure 7-1 Meshing of a cell 187

Effect of cell size on its maximum displacement and energy

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

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

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

Four groups of textile honeycomb composites, involving 14 varieties, have been

Thanks also extend to my dear colleges including Dr.Yu, Daniel, Mrs.Sun,Y.,

Lightweight materials with comparable properties such as high energy absorbency

Excellent energy absorption is one of honeycomb composites’ important mechanical

1.1 Description of the problem

1.2 Research Aim and Objectives

Textile technology is capable of creating 3D honeycomb woven structures without the

1) to develop a simple yet effective method for making honeycomb textile

2) to design and optimize the geometric parameters for the 3D honeycomb

3) to establish a procedure for the sake of manufacturing 3D textile honeycomb

2) to conduct the low-velocity impact tests on the 3D honeycomb textile

1) the first objective is to establish geometrical models for the 3D honeycomb

2) the second objective is to create models of three sets of honeycomb textile

Chapter 3 presents works on designing and manufacturing 3D honeycomb fabrics;

Chapter 6 reports on the experimental data analysis on textile honeycomb composites

Chapter 7 presents a finite element analysis (FEA) on honeycomb composites;

Chapter 8 ends the thesis with conclusions and future recommendations.

2.1 Classification of Cellular Solids

2.1.1. Honeycomb structure

Gibson and Ashby (1997) defined honeycomb as a two-dimensional array of polygons

(c) Three-dimensional foam with closed cells

Figure 2-1 Examples of cellular solids (Gibson and Ashby, 1997)

2.1.2 Foam structure

2.2 Main Features of Cellular Solids

Material Density,  (g/cm3)

2.2.2 Stiffness and strength of cellular solids

It is important to understand the stiffness and strength performances of honeycombs

Poisson’s ratios, 12* and  21