Composite Structures 275 (2021) 114489

Contents lists available at ScienceDirect

Composite Structures
journal homepage: www.elsevier.com/locate/compstruct

A numerical study on the influence of quasi-isotropic structures

on the ballistic performance of para-aramid woven composites
Hang Zhou a,b, Shengnan Min c, Xiaogang Chen a,⇑
a
Department of Materials, University of Manchester, Manchester M13 9PL, UK
b
College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China
c
School of Materials Design and Engineering, Beijing Institute of Fashion Technology, Beijing 100029, China

A R T I C L E I N F O A B S T R A C T

Keywords: Aimed at improving the ballistic performance of woven fabric reinforced composites, quasi‐isotropic structures
Para‐aramid fibre were developed. This paper reports on the ballistic impact performance and mechanisms of multi‐ply para‐
Quasi‐isotropic composite aramid/epoxy composites with orthotropic and quasi‐isotropic structures based on numerical exploration. A
Ballistic performance validated yarn level model was established using Abaqus® for the investigation of the ballistic performance
FE simulation
in terms of the total energy absorption and the ballistic mechanisms regarding energy absorbed in different
forms, stress dissipation, damage occurrence time and deformations. The results revealed that para‐aramid/
epoxy woven composites reinforced by quasi‐isotropic structures exhibit improved ballistic performance with
new ballistic mechanisms as the reinforcing structures changing from the conventional orthotropic to quasi‐
isotropic. In this research, up to an 8.9% increase in energy absorption is achieved by the quasi‐isotropic com-
posites comparing with the orthotropic composites under the same areal density, caused by enlarged stress dis-
sipation and deformation areas as the yarn directions increase. Limited by the stress propagation in the
thickness direction, there exists a maximum number of yarn directions in forming the most energy absorbent
quasi‐isotropic panel. This research fills the gap between the current study on quasi‐isotropic woven compos-
ites and their applications for ballistic protections.

1. Introduction ing kinetic energy from the projectile. Cunniff defined primary yarns
as those that are in direct contact with the projectile whereas the sec-
Lightweight hard armours made from textile composites have been ondary yarns are those that are not [5]. Transverse deflection on pri-
applied as a superior system for ballistic protection over the conven- mary yarns is caused by the projectile indentation. The
tional metallic armours, benefitting from the high strength, high flex- interlacements in a woven structure propagate the transverse deflec-
ibility and low density. One of the most commonly used high‐ tion onto the secondary yarns [6,7]. Both primary and secondary yarns
performance fibres, para‐aramid fibre, exhibits modulus of around are reported to play important roles in energy absorption [8,9]. How-
100GPa and strength of 3‐5GPa, which is 5–6 times stronger than that ever, due to the magnitude differences between the stress distributed
of steel [1,2] with the density only around 1/5 of steel. Kevlar® in primary and secondary yarns, secondary yarns are often not fully
(DuPont) and Twaron® (Teijin) are two of the major commercial stressed which hinders the fabric to be fully utilized for energy absorp-
para‐aramid fibre products [3]. Woven fabrics and composites made tion. Efforts have been made to involve more yarns in absorbing the
from such high‐performance fibres are widely used for protective kinetic energy of projectile through structural design. Quasi‐isotropic
armours. Vehicles and soldiers are able to achieve better mobility structures for plied fabrics were experimented, in which fabric plies
without jeopardizing the ballistic performance when equipped with were aligned along different directions compared to only two principal
lightweight high‐performance textile armours rather than conven- directions for orthotropic structures. Wang et al. [10] reported 20%
tional metal materials. increased energy absorption for the quasi‐isotropic panel when com-
When a fabric receives a transverse impact, transverse deflections pared to the structure of multi‐ply woven fabric counterpart. In their
incur in yarns and reach their limits at the yarns’ breakage [4], absorb- research, the improved ballistic performance was indexed by the

⇑ Corresponding author.
E-mail address: xiaogang.chen@manchester.ac.uk (X. Chen).

https://doi.org/10.1016/j.compstruct.2021.114489
Received 17 February 2021; Revised 11 June 2021; Accepted 30 July 2021
Available online 3 August 2021
0263-8223/Crown Copyright © 2021 Published by Elsevier Ltd. All rights reserved.
H. Zhou et al. Composite Structures 275 (2021) 114489

increased backface deformation area resulted from the superimposi- gated with details through FE simulation in various aspects including
tion effect of different inter‐ply angles. Min et al. reported that the stress dissipation and material deformation.
quasi‐isotropic UHMWPE (Ultra High Molecular Weight Polyethylene)
multi‐ply fabrics were 15% more energy absorbent than their orthotro- 2. Methodologies
pic counterparts [11]. This improvement was attributed to the
increased stress dissipation area. 2.1. Establishment of quasi-isotropic woven construction
Regarding textile reinforced composites subjected to high‐velocity
impacts, the structures of textile reinforcements have been reported The structural change of textile panels from orthotropic to quasi‐
to play important roles on the ballistic performance [12]. Similar to isotropic results in change in yarn directions. Compared to the yarns
the dry fabrics discussed above, energy absorption mechanisms of oriented in the two principal directions in an orthotropic structure,
the textile reinforced composites were reported to include cone forma- yarns are laid along more directions in a quasi‐isotropic structure with
tion in the composites, yarn fracture and the frictional energy absorp- a uniform angle between the adjacent directions. Due to the fact that
tion [8]. In addition, extra energy absorbing mechanisms include the yarns are orthotropic‐fixed in each individual woven fabric ply, the
matrix cracking and delamination. However, they were reported to orientation of yarns can only be realized by changing the ply directions
absorb energy of small magnitude, e. g. less than 10% of the total in a multi‐ply system. Thus, a quasi‐isotropic woven construction can
energy absorption [6,13]. Quasi‐isotropic composites were also be established through orientating the fabric plies along different
reported to be more energy absorbent in studies mainly focused on directions.
unidirectional composites [14‐16]. Yan and Chen reported improved In a quasi‐isotropic multi‐ply construction, the number of yarn
ballistic performance of quasi‐isotropic woven composites [17]. How- directions representing the degree of quasi‐isotropy is directly related
ever, this research only studied composites with 3 and 4 plies and to the ply number. The simplest quasi‐isotropic woven construction is
detailed mechanisms were not fully reported. A comprehensive a 2‐ply system in which yarns in the two plies are orientated along 4
research in this regard is still necessary. directions which are 45° apart, as shown in Fig. 1(a), with the yarn
To evaluate the ballistic performance, the projectile velocities dur- directions represented by red lines. Through increasing the yarn direc-
ing ballistic tests were measured, providing quantitative results reflect- tions, the whole construction can be made more quasi‐isotropic. In
ing the ballistic performance of different target materials [18]. In order to achieve more yarn directions, a higher ply number is required.
parallel, numerical method has also been widely used to study the bal- For instance, a 4‐ply quasi‐isotropic construction can be established
listic performance of materials, which is able to extract information with yarn orientated along 8 directions with an angle of 22.5° in
that is usually hard to obtain through experiments. Finite‐element between, and this is shown in Fig. 1(b). Obviously, this 4‐ply construc-
(FE) models with scales various from laminar‐level, yarn‐level to tion is more quasi‐isotropic than the 2‐ply construction.
fibre‐scale have been popularly applied in studying the ballistic perfor-
mance of textile composites [19‐25]. Compared with the laminar‐level 2.2. Characteristics of impact response
modelling and fibre‐scale modelling, the yarn‐level modelling is able
to offer adequate level of details regarding the yarn‐yarn interaction The impact responses of orthotropic and quasi‐isotropic dry woven
and yarn‐resin interface but with a moderate computation time. It fabrics were studied [11,18]. Due to the orthotropic characteristics of
was developed by explicitly modelling the yarns and resin in the com- the woven fabrics, the transverse deformation area in aligned ballistic
posite. The yarn‐level modelling has been applied and reported with panels when receiving a ballistic impact would exhibit a pyramid
good agreements with experimental results in many researches [24‐ shape with the diamond vertices in the primary yarn directions as seen
27]. in Fig. 2(a). Through plying fabrics along different directions in a
This research extends the study of developing quasi‐isotropic quasi‐isotropic multi‐ply fabric panel, the primary yarns in upper ply
woven fabric reinforced composites for improved ballistic perfor- press on the secondary yarns in the lower plies. Thus, the central trans-
mance. The aim of this research is to obtain a comprehensive under- verse deformation area would form a cone with roughly a round base,
standing on the influence of quasi‐isotropic structure on ballistic as illustrated in Fig. 2(b), due to the superimposition effect. Increased
performance of the composites and on the energy absorbing mecha- deformation area is produced as the base shape changes from diamond
nisms induced by the structural change. The ballistic performances base to round base, indicating more material involvement for energy
are evaluated in terms of energy absorptions of orthotropic and absorption.
quasi‐isotropic woven composites. Ballistic mechanisms are investi-

Fig. 1. Illustration of yarn directions in quasi-isotropic woven constructions. (a) A 2-ply quasi-isotropic construction; (b) A 4-ply quasi-isotropic construction.

2
H. Zhou et al. Composite Structures 275 (2021) 114489

fabric reinforced composites, this investigation is carried out in terms

of energy absorption, stress dissipation and deformation.

3. Finite-element (FE) simulation

3.1. Composite construction design

In this research, quasi‐isotropic woven fabric reinforced composites

Fig. 2. Impact faces and deformation contours of woven fabrics. (a) Cone base
were designed with plies evenly angled, which is the most typical
of an orthotropic panel; (b) Cone base of a quasi-isotropic panel [11]
quasi‐isotropic structure. A 3D coordinate system was used to describe
the composite structure with each ply specified by the angle θ accord-
In addition, Min et al. [11] reported a press effect of primary yarns ing to the x axial, shown in Fig. 5(a) with yarn directions represented
in the upper ply on the secondary yarns on the lower plies in quasi‐ by red lines. Orthotropic composites were produced by aligning all the
isotropic composites, causing more secondary yarns to be strained plies along 0°and defined as [0]n where n represents the ply number.
hence absorbing more energy. The claim of press effect was supported Due to the fact that each ply of the woven fabric contains 2 sets of per-
by the shape of deformation. However, the shape change and energy pendicular yarns, quasi‐isotropic woven composites were produced
absorption of a panel are not exactly of the same mechanism. The with an inter‐ply angle of 90°/n. Thus, a 2‐ply quasi‐isotropic panel
stress dissipation also plays an important role in influencing ballistic were produced and defined as [0/45] with an inter‐ply angle of 45°
performance. and a 3‐ply quasi‐isotropic panel were designed as [0/30/60], shown
In a textile composite, yarn movements are constrained due to the in Fig. 5(b) and (c). In order to comprehensively understand the per-
resin impregnation. Unlike the woven fabrics, an obvious diamond formance of different quasi‐isotropic structures suffered from ballistic
base deformation area was not observed in an orthotropic woven com- impacts, composites with more yarn directions were investigated as
posite with few ply number as illustrated in Fig. 3(a), but the orthotro- well. Thus, 4‐ply and 6‐ply composites which can be designed with a
pic characteristics of a woven structure was still kept. The deformation higher degree of quasi‐isotropy were studied as well. The construction
area was cross‐shape based in the orthotropic carbon fibre/epoxy com- designs of quasi‐isotropic woven composites are listed in Table 1.
posite subjected to a high velocity impact through FE simulation [28]
as shown in Fig. 3(b). Thus, based on the work regarding quasi‐ 3.2. FE model establishment
isotropic woven dry fabrics discussed above, a modified illustration
with different deformation area shapes can be established for woven In this research, numerical models established through assembling
fabric reinforced composites, shown in Fig. 4. For orthotropic compos- projectile and the composite models. Composite models were estab-
ites, the deformation area exhibits a cross‐like shape as shown in Fig. 4 lished with yarn‐level details aiming at investigating different mecha-
(a). For quasi‐isotropic composites with few yarn directions, the defor- nisms induced by the change of yarn directions.
mation area will exhibit a shape of polygon which is formed through The projectile was modelled as a solid cylinder based on the real
the superposition of few cross‐like shapes from orthotropic construc- projectiles used in experiments, shown in Fig. 6(a). The diameter
tions, shown in Fig. 4(b). The quasi‐isotropic composite with massive and height of the projectile are both 5.5 mm. The projectile was mod-
yarn directions will result in a deformation area of round shape which elled as a rigid body which does not deform during impact events.
is similar to the dry fabric system, shown in Fig. 4(c). As the shape of To establish the yarn‐level composite models, fabric reinforcement
deformation area converts from cross to polygon and then round, more and resin were modelled separately and then assembled together to
materials will be involved in energy absorption of quasi‐isotropic com- form the textile composites. Fabrics were modelled on the yarn level,
posites which can be reflected by the deformation area. Besides, a new in order to save the computation resource. Composite yarns in this
stress dissipation is expected to be produced as well due to the press research were aramid yarns consisting of filaments and impregnated
effect which has been mentioned in woven fabrics. Aiming to compre- with resin. Thus, relative movements among filaments were not con-
hensively study the influence of quasi‐isotropic structure on woven sidered and the yarns were modelled as homogeneous material. Both

Fig. 3. Deformation area of a fabric reinforced composite. (a) Projectile impact of an orthotropic Twaron/epoxy composite [12]; (b) Simulation of the deformation
area in an orthotropic carbon fibre/epoxy composite [28]

3
H. Zhou et al. Composite Structures 275 (2021) 114489

Fig. 4. Illustrations of deformation areas of woven fabric reinforced composites. (a) Orthotropic composite; (b) Quasi-isotropic composite with few yarn
directions. (c) Quasi-isotropic composite with massive yarn directions.

Fig. 5. Schematics of quasi-isotropic composites. (a) A schematic of the coordinate system with primary yarns marked; (b) Schematic of 2-ply quasi-isotropic
composite [0/45]; (c) Schematic of 3-ply quasi-isotropic composite [0/30/60]

Table 1
Construction designs of quasi-isotropic composites.

Ply number Construction design

2-ply [0/45]
3-ply [0/30/60]
4-ply [0/45]2 [0/22.5/45/67.5]
6-ply [0/45]3 [0/30/60]2 [0/15/30/45/60/75]

warp and weft yarns are flattened in the real fabrics, and a lenticular
cross‐section yarn model developed by Shanahan and Hearle [30]
[39] was adopted as shown in Fig. 6(b). In their model, A2 and B2 rep-
resent the width and height of the lenticular cross‐section of the weft
yarn respectively, while B1 denotes the height of the warp yarn cross‐
section and the width of which is not explicit. This model has been
widely used in the ballistic simulation of woven fabrics, leading to reli-
able simulated results [10,11,18]. Woven fabric reinforcements in this
research were prepared using Twaron® filament yarn (168tex) with
plain weave, and the thread density was 7 threads/cm in both weft
and warp directions. In the geometric model of the reinforcing fabric,
the yarn spacing was minimized indicating the fibre filament in the
yarns were spread laterally to cover the space between adjacent yarns
Fig. 6. Geometric models in simulation. (a) Geometric model of the projectile;
and the yarn path was created to fit the cross‐section. The yarn’s geo- (b) Lenticular cross-section model; (c) Geometric model of yarn with lenticular
metric model, as shown in Fig. 6(c), was established according to the cross-section.
calculations and measurements in experiments. Based on the thread
density of the plain weave fabric, the width of the yarn cross‐section
and half of the wavelength of the yarn were calculated as 1.43 mm. the shape of fabric to fit the fabrics tightly. The reinforcing fabric part
The height of composite yarn cross‐section was measured as 0.2 mm. was assembled with the resin part to form the composite, and the geo-
The resin part of the composite was modelled as a homogeneous metric assembly is shown in Fig. 7, with the fabric represented by the
and isotropic hollow solid. The inner space was created according to yarns.

4
H. Zhou et al. Composite Structures 275 (2021) 114489

3.3. Material models Table 2

Properties of steel.
For this yarn‐level modelling, properties of projectile, yarn and Modulus of elasticity E (GPa) Poisson’s ratio ν Density ρ (kg/m3)
resin components were defined separately. The projectiles are made
206.8 0.35 7600
from steel in real situation, whose properties are listed in Table 2.
The resin matrix for making the composites in this research was
prepared using epoxy Araldite® LY564 and hardener Aradur®
LY2954. The mechanical properties in Table 3 are collected from pre- Table 3
Properties of resin system.
vious work [32,35]
In real situations, the para‐aramid yarns used in this research are Resin Properties
Twaron® (168tex) from Teijin Ltd. The yarns consist of filaments
Modulus of elasticity E (GPa) 2.6
and were soaked with resin in textile composites. Thus, in this yarn‐ Fracture strain ɛ (%) 4.3
level simulation, mechanical properties of one composite yarn were Yield stress σ (GPa) 0.075
defined as the mixture of properties of dry yarn and resin. The volume Poisson’s ratiov 0.35
Mass density ρ (kg/m3) 1100
of resin component in each ply is 1.671 cm3 while the volume of fabric
Fracture energy (J/m2) 200
is 3.260 cm3 in the model. Thus, the composite yarns ware modelled at Shear modulus G (GPa) 0.975
an 85% fibre volume fraction to satisfy the 56% fibre volume fraction
of whole composite panel measured in experiments. The mechanical
properties of resin and yarns used in this FE model were adopted from
Table 4
the previous reports [31–36], where the material properties were cal- Properties of a composite yarn.
culated according to classic rules of mixtures [40], which are listed in
Table 4 with direction indexes 1, 2 and 3 illustrated in Fig. 6c. Yarn Properties

The failure criteria used in this simulation is ductile damage, which Modulus of elasticity E 1 (GPa) 78
was reported for simulating the high velocity penetration of woven Modulus of elasticity E 2 (GPa) 2.2
fabrics made from high performance polymer fibres [10,11,29]. A yarn Modulus of elasticityE3 (GPa) 2.2
Fracture strain ɛ (%) 3.4
fails when its failure strain is reached, and the damage evaluation is
Yield stress σ (GPa) 1.45
carried out according to the level of fracture energy. In the case of a Poisson’s ratio v12 0.35
para‐aramid fibre, the plastic deformation is not obviously exhibited Poisson’s ratio v13 0.35
within this deformation period [31]. Thus, the onset of the damage ini- Poisson’s ratio v23 0.35
Mass density ρ (kg/m3) 1389
tiation was assumed to be the point of strain where the stress reached
Fracture energy (J/m2) 1000
the yield stress of material. After damage initiation, the damage eval- Shear modulus G12 (GPa) 2.4
uation was specified in terms of fracture energy. Shear modulus G13 (GPa) 2.4
For modelling the fabric‐resin interface, a perfect bonding with no Shear modulus G23 (GPa) 1.1
separation or movements which was reported feasible in previous
work [36], was applied for binding the resin and fabric in this simula-
tion. Inter‐ply delamination was not considered in this simulation due Table 5
to yarn fracture and deformation was reported as the main energy Properties of the inter-ply contact.
absorption mechanisms in a high‐speed impact event, while energy Interface Properties
absorbed by delamination was observed of small magnitude [6]. Thus,
no constrain was set between plies. Young’s modulus (MPa) 2600
Shear modulus (MPa) 45
The inter‐ply contact was modelled with cohesive behaviour, and Maximum nominal normal stress (MPa) 34.5
failure of the contact was modelled according to the quadratic traction Maximum nominal shear stress (MPa) 9
criterion. The contact between plies fails when the maximum stress is Normal Fracture energy (J/m2) 800
reached and the damage evaluation is carried out according to the Shear Fracture energy (J/m2) 200
level of fracture energy. Inter‐ply contact properties were adopted
from previous reports and listed in Table 5 [41‐43].
eight‐node hexahedron element, coded C3D8R. The element size was
3.4. Other model details set to be 0.5 mm for the projectile and yarns and 0.25 mm for the resin
for an accurate simulation on the complex geometries.
The velocity of projectile was set to be 480 m/s in this simulation
following the measured impact velocities in experiments which were 3.5. Model validation
between 450 m/s and 500 m/s. Specimens were clamped around the
edge in real situation, and in this simulation, a fixed edge boundary Validation of FE model was carried out against the experiment
condition was applied according to previous work [10,11]. A frictional results in terms of the morphology of impact face and energy absorp-
coefficient 0.3 was applied for contacts including yarn‐projectile and tion. High velocity impact experiments were carried out using a ballis-
resin‐projectile in simulation [37]. tic penetration test system with infrared detectors to detect the
To achieve an accurate simulation of the transverse isotropic mate- projectile velocities, and the test rig set up is illustrated in Fig. 8(a)
rial, the projectile, yarns and resin in the model were all meshed with and (b). Composite specimens were prepared using Resin Transfer
Molding (RTM) process following the instruction offered by Hunts-
man® which is the epoxy supplier. The area densities and thickness
of specimens are listed below (Table 6).
In the penetration test, specimens were clamped by a clamp tool
illustrated in Fig. 8(c), which is comparable to the fixed boundary con-
dition and model size in simulation. The total energy absorption was
Fig. 7. Geometric assembly of the resin and fabric reinforcement in simula- characterized using the kinetic energy loss of the projectile, calculated
tion with the fabric represented by yarns. according to the impact and residual velocities. Each structure was

5
H. Zhou et al. Composite Structures 275 (2021) 114489

Fig. 8. Illustrations of the experimental setups. (a) Illustration of the penetration test system; (b) Picture of the penetration test system; (c) Illustration of the clamp
tool.

Table 6
Properties of a composite yarn.

Panel Experiments Simulations

2
Area densities (g/cm ) Thickness (mm) Area densities (g/cm2) Thickness (mm)

2-ply [0]2 0.095 0.83 0.105 0.84

[0/45] 0.092 0.80
3-ply [0]3 0.125 1.23 0.141 1.24
[0/30/60] 0.128 1.20
4-ply [0]4 0.158 1.60 0.177 1.64
[0/45]2 0.162 1.63
6-ply [0]6 0.239 2.33 0.249 2.44
[0/45]3 0.243 2.30
[0/30/60] 0.244 2.30

prepared with three panels and each panel was tested once, then the numbers and more yarn directions are further studied in following sec-
average energy absorption was taken. Fig. 9(a) exhibits the impact face tions aiming to achieve a comprehensive understanding of quasi‐
of one of these panels ([0]3) in experiment and simulation. A good isotropic structures for woven fabric reinforced composites.
agreement on the number of broken yarns was observed between
the simulation and experiment, both exhibiting 4 fractured warp yarns 4.1.1. Energy absorption of 3-ply composites
and 4 fractured weft yarns which are marked with numbers in Fig. 9 The total energy absorption of the composite is characterized using
(a). Fig. 9(b) exhibits the energy absorptions of the woven composite the kinetic energy loss of the projectile in this simulation. It is seen
specimens from the simulations and experiments. Differences range from Fig. 10 that the transition of structure from orthotropic to
from 0.5% which is seen in panel [0/45] to 5.8% in panel [0]4. Over- quasi‐isotropic led to a rapid kinetic energy loss of the projectile,
all, the FE results agree well with the experimental ones, hence validat- and this is shown in Fig. 9 in the time history of the kinetic energy car-
ing the FE model. ried by projectile. By the time when the projectiles fully penetrated the
composites, panel [0/30/60] absorbed 14.66% (16.69 J) of the impact
4. Results and discussions energy. Panel [0]3 absorbed 14.03% (15.97 J) of the impact energy. As
the structure changes from orthotropic to quasi‐isotropic, the total
4.1. 3-ply composites energy absorption increased by 4.5%.
Upon ballistic impacts, energy of the projectile is absorbed by the
In this section, 3‐ply composites are selected to study the influence fabric panels mainly in the form of kinetic energy (KE) and internal
of quasi‐isotropic structure on the ballistic performance. Ballistic per- energy (IE) [6,16]. The rest of energy absorption is frictional dissi-
formance and mechanisms are investigated in terms of energy absorp- pated energy (FD), fracture energy (FE) and energy absorbed by
tion, stress dissipation and deformation. Composites with larger ply delamination. However, simulation results show that delamination

6
H. Zhou et al. Composite Structures 275 (2021) 114489

Fig. 9. Comparison between the experimental and simulated results of the woven composites. (a) Number of fractured yarns, (b) Energy absorption.

Fig. 10. Time history of projectile kinetic energy loss in 3-ply composites.

absorbed about 0.5% of the total energy, which is much smaller than Fig. 11. Energy absorption by different mechanisms in 3-ply woven
other energy absorption mechanisms. Thus, energy absorption mecha- composites.
nisms KE, IE, FD and FE are mainly studied in this research.
Fig. 11 exhibits the energy absorption by different mechanisms in
the 3‐ply composites at the moment when the projectile fully pene- tion areas in the 2nd and 3rd plies. The 2nd ply showed a stress dissi-
trated the composites. The quasi‐isotropic panel [0/30/60] demon- pation area of a polygon shape with outward expanded edges under
strated a 10.6% increased IE then the orthotropic panel [0]3. KE, the superimposition effect of the 1st ply. The 3rd ply displays a round
related to the movement of the panel, was increased by 7.2% over that shape stress dissipation area which is obviously larger than the dia-
of panel [0]3, whilst FD and FE illustrated little difference between the mond shape in panel [0]3 due to the superimposition effects of both
two. The improved total energy absorption was mainly in the forms of the 1st and 2nd plies. According to the measurements, the stress dissi-
KE and IE as the panel construction changes from orthotropic to quasi‐ pation areas increase by 4.1% in ply2 and 8.6% in ply3 as the structure
isotropic. changes from orthotropic to quasi‐isotropic in the 3‐ply woven com-
posites, and the comparisons are illustrated in Fig. 13. As a result of
4.1.2. Stress dissipation and yarn initial damage changing the structure from orthotropic to quasi‐isotropic, more sec-
In a quasi‐isotropic structure, the superimposition of angled plies ondary yarns are strained to absorb the impact energy.
results in the tensioned primary yarns in upper plies due to impact As the primary yarns contact with the projectile directly, the stress
would press on the secondary yarns in lower plies. Aimed at studying dissipations along the primary yarns are focused on for achieving a
this superimposition effect, the in‐plane stress dissipations are investi- detailed insight into the impact response of composites. Aimed at
gated at the moments just before the projectile penetrates the 1st ply of reflecting the situation of stress growth without influenced by the yarn
each panel. Fig. 12 exhibits the stress dissipation contours of 3‐ply damage, stress dissipations are studied at the moment just before any
woven composites at 2.4 µs, the moment just before the projectile pen- initial damage generated on the 1st plies of both panels. In this simu-
etrated the 1st plies of both the orthotropic and quasi‐isotropic panels. lation, 4 warp and 4 weft primary yarns in one composite are con-
Very limit difference is shown between the 1st plies. However, the tacted with the projectile directly as shown in Fig. 14. One of the
quasi‐isotropic composite [0/30/60] exhibited different stress dissipa- warp yarns in the middle of the projectile coverage is taken for the

7
H. Zhou et al. Composite Structures 275 (2021) 114489

Fig. 12. Stress dissipation contours of 3-ply woven composites.

Fig. 13. Stress dissipation areas of the 3-ply composites.

study of stress dissipation, and due to the symmetrical characteristic of

the structure, half‐length of this yarn was considered.
Fig. 15 depicts the stress dissipations along the primary yarns in the
two 3‐ply woven composites at 0.6 µs, a moment at which the primary Fig. 14. Yarn path of the selected primary yarn.
yarns of 1st plies in these two panels starts to show damage. Compared
with quasi‐isotropic panel, the stress in the primary yarn of the ortho-
panel [0/30/60] than that of [0]3, and this is displayed in Fig. 15
tropic panel [0]3 built up more rapidly than in its quasi‐isotropic coun-
(b), It is highly possible that the decreased stress magnitudes in pri-
terpart. Just before damage occurs to the primary yarns of 1st plies in
mary yarns are induced by the increased stress dissipation areas in
these two panels, the quasi‐isotropic panel [0/30/60] exhibits a 5.9%
which more secondary yarns are strained. Another possible reason
lower peak stress than the orthotropic panel [0]3, as shown in Fig. 15
can be the more rapid reduction of the projectile kinetic energy in
(a). The stress peak drops by 36% on the primary of the 2nd ply in

8
H. Zhou et al. Composite Structures 275 (2021) 114489

Fig. 17. Calculated backface deformation area of the 3-ply composites at

1.1 µs.

Fig. 18. Transverse deformation of 3-ply composites at 1.1 µs.

Fig. 15. Stress dissipations along the primary yarns in 3-ply woven compos-
ites at 0.6 µs. (a) ply1; (b) ply2; and (c) ply3.

Fig. 19. Energy absorptions of orthotropic and quasi-isotropic composites.

Table 7
Occurrence time of initial damage in the 3-ply woven composites.

[0]3 [0/30/60] the 3rd plies had not been fully generated and dissipated at that early
stage.
Ply1 0.6 µs 0.6 µs
Ply2 0.8 µs 1.0 µs The lower peak stress in primary yarns in the quasi‐isotropic com-
Ply3 1.2 µs 1.4 µs posite [0/30/60] delays the initial damage in comparison to the ortho-
tropic composite [0]3, and the times at which initial damage appeared
in different plies of the two composites panels are shown in Table 7.
The 2nd and 3rd plies demonstrate 0.2 µs delayed initial damage
respectively while no obvious difference was observed in the 1st plies.
the quasi‐isotropic construction. Fig. 15(c) indicates no significant
The later the damage occurs, the further the longitudinal stress wave
peak stress difference in 3rd plies between the two composites, and
propagated, and hence more material were strained and involved for
it is believed that the main reason for this is because the stress in
energy absorption in the ballistic event.

Fig. 16. Backface deformation contours of 3-ply composites at 1.1 µs. (a) [0]3; (b) [0/30/60]

9
H. Zhou et al. Composite Structures 275 (2021) 114489

transverse deformation represents small impact trauma that is pre-

ferred for protection. In this research which allows the panels to be
penetrated, aiming at investigating the creation of deformations with-
out influenced by the yarn damage, both backface and transverse
deformations are checked at the moment just before the initial damage
started to damage in the bottom plies of both composites.
Fig. 16 shows the backface deformations at 1.1 µs, the moment just
before the initial damage generated on the bottom plies of both panels.
Due to the pressing effects between plies of different directions, differ-
ent backface deformation contours are produced for orthotropic and
Fig. 20. Energy absorption of 4-ply woven composites. quasi‐isotropic composites through FE modelling. Orthotropic para‐
aramid/epoxy woven composite [0]3 exhibits a deformation area of
cross‐like shaped in the centre and diamond shape around in this sim-
ulation, shown in Fig. 16(a). For the quasi‐isotropic composite, a
deformation area is produced of polygon shape in the centre with
round shape around, shown in Fig. 16(b). Fig. 17 shows that the defor-
mation area increases by 4.0% in the case of the quasi‐isotropic com-
posites when compared to its orthotropic counterpart under this
specific contour scale.
At the same moment, the quasi‐isotropic panel [0/30/60] produced
2.5% smaller transverse deformation depth than [0]3 due to more
materials are involved in energy absorption (Fig. 18). This suggests
that in a non‐penetration impact event, the quasi‐isotropic composite
panel would produce a shallower trauma which is favoured for the bal-
listic protection.
Fig. 21. Energy absorption of 6-ply woven composites.

4.2. Composites with increased ply numbers

4.1.3. Deformation
Deformation, an important mechanism for absorbing energy in As discussed above, the quasi‐isotropic woven composite is more
impact events, are investigated in terms of backface deformation and energy absorbent than the orthotropic composite, with an improve-
transverse deformation. A larger backface deformation indicates more ment of 4.5% for the 3‐ply composite samples. Aimed at comprehen-
materials are involved in the energy absorption. Moreover, a small sively investigate the improvements of quasi‐isotropic structures on

Fig. 22. Stress dissipation contours of 6-ply woven composites.

10
H. Zhou et al. Composite Structures 275 (2021) 114489

the ballistic performance, woven composites with different ply num- However, decreased sensitivity was observed as the yarn direction
bers are studied. In this investigation, 2‐ply, 4‐ply and 6‐ply quasi‐ increased that made a composite panel with evenly distributed yarns
isotropic composites, denoted by [0/45], [0/45]2 and [0/45]3, are not show quasi‐isotropic features. It indicates that the composite with
selected as they involve the same yarn directions of the [0/45]x the most yarn directions may not always absorb the most energy.
structure.
Fig. 19 exhibits the energy absorption of the 2‐ply, 4‐ply and 6‐ply Declaration of Competing Interest
quasi‐isotropic composites, whose areal densities are the same as their
orthotropic counterparts. The results show that 2‐ply quasi‐isotropic The authors declare that they have no known competing financial
composite [0/45] absorbs 3.6% more energy than the orthotropic interests or personal relationships that could have appeared to influ-
panel [0]2. With the increase of ply number, [0/45]3 absorbs 7.7% ence the work reported in this paper.
more energy than the 6‐ply orthotropic composite [0]6. It indicates
that changing the structure from orthotropic to quasi‐isotropic is more Acknowledgement
effective for composites with increased ply number.
The authors wish to thank Teijin® for providing the para‐aramid
4.3. Composites with increased yarn directions fibres for this research. This research was partially funded by National
Natural Science Foundation of China (Grant No. 11902008).
For quasi‐isotropic woven composites, increased yarn directions
indicate a higher degree of quasi‐isotropy. In the case of 4‐ply compos- References
ites, the higher degree of quasi‐isotropy resulted in the increase of
energy absorption. Panel [0/22.5/45/67.5] with the most yarn direc- [1] Kaw AK. Mechanics of composite materials. 2nd ed. Boca Raton, FL: Taylor &
Francis; 2006.
tions absorbed 5.7% increased energy than the orthotropic panel [0]4. [2] Starley D. Determining the technological origins of iron and steel. J Archaeol Sci
This magnitude exceeds the 3.7% improvement for quasi‐isotropic 1999;26(8):1127–33.
panel [0/45]2 with less yarn directions, and is illustrated in Fig. 20. [3] Scott RA, editor. Textiles for protection. Cambridge: Woodhead; 2005.
[4] Roylance D. Stress wave propagation in fibres: Effect of crossovers. Fibre Sci
A similar behaviour is identified in 6‐ply composites. The increase Technol 1980;13(5):385–95.
of yarn directions caused the increase of energy absorption up to an [5] Cunniff PM. An analysis of the system effects in woven fabrics under ballistic
8.9% improvement in panel [0/30/60]2 than the orthotropic panel impact. Text Res J 1992;62(9):495–509.
[6] Naik NK, Shrirao P. Composite structures under ballistic impact. Compos Struct
[0]6, as shown in Fig. 21. However, a limitation in energy absorption 2004;66(1-4):579–90.
is shown when further increased the yarn directions. Panel [7] Duan Y, Keefe M, Bogetti TA, Cheeseman BA, Powers B. A numerical investigation
[0/15/30/45/60/75] exhibited a decreased energy absorption than of the influence of friction on energy absorption by a high-strength fabric subjected
to ballistic impact. Int J Impact Eng 2006;32(8):1299–312.
panel [0/30/60]2 and [0/45]3 although it contains the most yarn
[8] Lee BL, Walsh TF, Won ST, Patts HM, Song JW, Mayer AH. Penetration failure
directions. mechanisms of armor-grade fiber composites under impact. J Compos Mater
Mechanisms behind the decreased energy absorption in panel 2001;35(18):1605–33.
[9] Roylance D, Wilde A, Tocci G. Ballistic impact of textile structures. Text Res J
[0/15/30/45/60/75] can be reflected by the stress dissipation con-
1973;43(1):34–41.
tours when the projectiles have just penetrated the 1st plies as shown [10] Wang Y, Chen X, Young R, Kinloch I, Wells G. A numerical study of ply orientation
in Fig. 22. In panels [0/45]3 and [0/30/60]2, obvious polygon or on ballistic impact resistance of multi-ply fabric panels. Compos B Eng
round‐like shape stress dissipation areas are exhibited due to the 2015;68:259–65.
[11] Min S, Chu Y, Chen X. Numerical study on mechanisms of angle-plied panels for
superposition of plies along different directions. However, it is not ballistic protection. Mater Des 2016;90:896–905.
the same in panel [0/15/30/45/60/75]. Very significant difference [12] Zhang D, Sun Y, Chen Li, Zhang S, Pan N. Influence of fabric structure and
is exhibited especially in 3rd plies and 4th plies in which panel thickness on the ballistic impact behavior of Ultrahigh molecular weight
polyethylene composite laminate. Mater Des 2014;54:315–22.
[0/15/30/45/60/75] produced square shape stress dissipation areas [13] Shaktivesh, Nair NS, Sesha Kumar CV, Naik NK. Ballistic impact performance of
while others are round. When looked at the stress dissipation contours composite targets. Mater Des 2013;51:833–46.
of all the plies in panel [0/15/30/45/60/75], this panel with angled [14] David-West OS, Alexander NV, Nash DH, Banks WM. Energy absorption and
bending stiffness in CFRP laminates: The effect of 45° plies. Thin-Walled Struct
plies do not produce an evenly distributed stress as shown in other 2008;46(7-9):860–9.
two panels. It indicates that the top ply did not have a press effect [15] Stevanović MM, Marković DS, Pešikan-Sekulić DR. Effects of stacking sequence on
on the bottom ply which can be concluded as a decreased sensitivity. the strength of cross-ply and quasi-isotropic carbon/epoxy laminates. MSF
2004;453-454:465–72.
In this case, although the yarns are evenly distributed along different
[16] Kasano H. Impact perforation of orthotropic and quasi-isotropic CFRP laminates by
directions, the whole panel did not exhibit a quasi‐isotropic feature. a steel ball projectile. Adv Compos Mater 2001;10(4):309–18.
Thus, decreased energy absorption is produced in panel [17] Yan R, Chen X. Aramid/epoxy composites with angle-laid reinforcement
constructions for ballistic protection. J Ind Text 2016;45(5):765–79.
[0/15/30/45/60/75] but not for other 6‐ply quasi‐isotropic compos-
[18] Tabiei A, Nilakantan G. Ballistic impact of dry woven fabric composites: a review.
ites. Detailed mechanisms causing the decreased sensitivity worth fur- Appl Mech Rev 2008;61:010801.
ther investigation. [19] Ivanov I, Tabiei A. Loosely woven fabric model with viscoelastic crimped fibres for
ballistic impact simulations. Int J Numer Meth Engng 2004;61(10):1565–83.
[20] Parsons EM, King MJ, Socrate S. Modeling yarn slip in woven fabric at the
5. Conclusions continuum level: Simulations of ballistic impact. J Mech Phys Solids 2013;61
(1):265–92.
[21] Wang Y, Miao Y, Swenson D, Cheeseman BA, Yen C-F, LaMattina B. Digital
The influence of quasi‐isotropic structures on the ballistic perfor- element approach for simulating impact and penetration of textiles. Int J Impact
mance and mechanisms of para‐aramid/epoxy composites have been Eng 2010;37(5):552–60.
studied through a validated FE model. It reveals that the quasi‐ [22] Sockalingam S, Gillespie JW, Keefe M. On the transverse compression response of
Kevlar KM2 using fiber-level finite element model. Int J Solids Struct 2014;51
isotropic structures are more energy absorbent than the orthotropic
(13):2504–17.
structure. This improvement effect is more dramatic for composites [23] Sockalingam S, Gillespie JW, Keefe M. Dynamic modeling of Kevlar KM2 single
with increased ply numbers. In the case of para‐aramid/epoxy compos- fiber subjected to transverse impact. Int J Solids Struct 2015;67-68:297–310.
ites, at least 3.6% increased energy absorption was produced as the [24] Jia X, Huang Z, Zu X, Gu X, Xiao Q. Effect of mesoscale and multiscale modeling on
the performance of kevlar woven fabric subjected to ballistic impact: a numerical
structure converts from orthotropic to quasi‐isotropic. This magnitude study. Appl Compos Mater 2013;20(6):1195–214.
could be more significant and reaches 8.9% as the yarn directions [25] Jiang W-G, Hallett SR, Wisnom MR. Development of Domain Superposition
increased. This improved effect is resulted from the contact feature Technique for the Modelling of Woven Fabric Composites. Mechanical Response of
Composites, vol. 10, Dordrecht: Springer Netherlands; 2008, p. 281–91.
of angled plies which involves more materials in energy absorption.

11
H. Zhou et al. Composite Structures 275 (2021) 114489

[26] Duan Y, Keefe M, Bogetti TA, Cheeseman BA. Modeling the role of friction during [34] Park Y, Kim Y, Baluch AH, Kim C-G. Numerical simulation and empirical
ballistic impact of a high-strength plain-weave fabric. Compos Struct 2005;68 comparison of the high velocity impact of STF impregnated Kevlar fabric using
(3):331–7. friction effects. Compos Struct 2015;125:520–9.
[27] Barauskas R, Abraitienė A. Computational analysis of impact of a bullet against the [35] Mse.mtu.edu. Mechanical Properties Data. 2019. http://www.mse.mtu.edu/
multilayer fabrics in LS-DYNA. Int J Impact Eng 2007;34(7):1286–305. ~drjohn/my4150/props.html. (accessed February 9, 2019).
[28] Coles LA, Roy A, Silberschmidt VV. Ice vs. steel: Ballistic impact of woven carbon/ [36] Gerlach R, Pabst A, Petrinic N, Hornig A, Wiegand J, Siviour CR, et al. The
epoxy composites. Part II – Numerical modelling. Eng Fract Mech interface between matrix pockets and fibre bundles under impact loading. Compos
2020;225:106297. https://doi.org/10.1016/j.engfracmech.2018.12.030. Sci Technol 2009;69(11-12):2024–6.
[29] Wang Y, Chen X, Young R, Kinloch I. A numerical and experimental analysis of the [37] Gogineni S, Gao X-L, David NV, Zheng JQ. Ballistic impact of twaron CT709®
influence of crimp on ballistic impact response of woven fabrics. Compos Struct plain weave fabrics. Mech Adv Mater Struct 2012;19(6):441–52.
2016;140:44–52. [39] Shanahan WJ, Hearle JWS. 12—An energy method for calculations in fabric
[30] Hearle JWS, Shanahan WJ. 11—An energy method for calculations in fabric mechanics Part II: Examples of application of the method to woven fabrics. J
mechanics Part I: Principles of the method. J Textile Inst 1978;69(4):81–91. Textile Inst 1978;69(4):92–100.
[31] Teijin. Twaron – a versatile high-performance fibre. 2012. http://www. [40] Young RJ, Lovell PA. Introduction to polymers. 3rd ed. Boca Raton: CRC Press;
teijinaramid.com/wp-content/uploads/2012/02/1090308_Twaron- 2011.
productbrochurefinal_051.pdf. (accessed August 1, 2017). [41] Moure MM, Rubio I, Aranda-Ruiz J, Loya JA, Rodríguez-Millán M. Analysis of
[32] Mouldlife.net 2019. https://www.mouldlife.net/ekmps/shops/mouldlife/ impact energy absorption by lightweight aramid structures. Compos Struct
resources/Other/araldite-ly564-aradur-2954-eur-e-1-.pdf. (accessed February 6, 2018;203:917–26.
2019). [42] Kim J-K, Sham M-L. Impact and delamination failure of woven-fabric composites.
[33] Rao MP, Duan Y, Keefe M, Powers BM, Bogetti TA. Modeling the effects of yarn Compos Sci Technol 2000;60(5):745–61.
material properties and friction on the ballistic impact of a plain-weave fabric. [43] Rubio I, Rodríguez-Millán M, Marco M, Olmedo A, Loya JA. Ballistic performance
Compos Struct 2009;89(4):556–66. of aramid composite combat helmet for protection against small projectiles.
Compos Struct 2019;226:111153.

12