Recent Advances in Modeling and Experiments of Kevlar Ballistic Fibrils, Fibers, Yarns and Flexible Woven Textile Fabrics - A Review

Review
Textile Research Journal

0(00) 1–27
Recent advances in modeling and ! The Author(s) 2016

Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
experiments of Kevlar ballistic fibrils, DOI: 10.1177/0040517516646039
trj.sagepub.com
fibers, yarns and flexible woven textile
fabrics – a review
Subramani Sockalingam1,2, Sanjib C Chowdhury1,

John W Gillespie Jr1,2,3,4 and Michael Keefe1,2
Abstract
Ballistic impact onto flexible woven textile fabrics is a complicated multi-scale problem given the structural hierarchy of
the materials, anisotropic material behavior, projectile geometry–fabric interactions, impact velocity and boundary con-
ditions. Although this subject has been an active area of research for decades, the fundamental mechanisms, such as
material failure, dynamic response and multi-axial loading occurring at the lower length scales during impact, are not well
understood. This paper reviews the recent advances in modeling and experiments of Kevlar ballistic fibrils, fibers, yarns
and flexible woven textile fabrics pertinent to the deformation modes occurring during impact and serves to identify
topics worthy of further investigation that will advance the basic understanding of the phenomena governing transverse
impact. This review also explores aspects such as homogeneous versus heterogeneous behavior of yarns consisting of
individual fibers and the inelastic transverse behavior of the fiber, which is not considered in the previous review papers
on this topic.
Keywords
ballistic fiber, ballistic impact, finite element analysis, wave propagation
The textile fabrics possess a hierarchical multi-scale

High-performance aramid (KevlarÕ , TwaronÕ ) and architecture from fibrils to fibers to yarns (tows) to
polyethylene (PE; SpectraÕ , DyneemaÕ ) fibers are multi-layer fabrics, as shown in Figure 2(a). Kevlar
widely used in ballistic impact applications, such as (p-phenylene terephthalamide, PPTA) is an aromatic
soft body armor1,2 and gas turbine engine containment polyamide (i.e., aramid) type polymeric material.
systems,3 in the form of flexible textile fabrics. The tex- In the aromatic structure of the Kevlar chain, adjacent
tile fabrics are also used as a backing material in an phenylene rings are connected through the amide group
armor system with a ceramic strike face. A generic (see Figure 2), while the phenylene rings stay preferably
fabric armor system subjected to impact and some of
the physically observed mechanisms and deformation
modes in the fabric are shown in Figure 1. Impact velo- 1
Center for Composite Materials, University of Delaware, USA
cities lower than the wave speed of the material are the 2
Department of Mechanical Engineering, University of Delaware, USA
typical range for armor backings. Hence, shock wave 3
Department of Materials Science and Engineering, University of
effects are not considered in this review. The impact Delaware, USA
4
response is often dependent on the projectile geometry, Department of Civil and Environmental Engineering, University of
impact velocity, material behavior and boundary con- Delaware, USA
ditions. Generally speaking, the fabric layers closer to
Corresponding author:
the impact site are subjected to local transverse com- Subramani Sockalingam, Center for Composite Materials, University of
pression and shear and the layers in the rear are loaded Delaware, Newark, DE 19716, USA.
in in-plane tension. Email: [email protected]
Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

2 Textile Research Journal 0(00)
Figure 1. Mechanisms, deformation and failure modes in a generic fabric armor at multiple length scales: (a) multiple fabric layers
impacted (figure courtesy of Chocron et al.4); (b) yarn pull out from fabric edges during ballistic impact testing (figure courtesy of
Nilakantan et al.5); (c) yarn pull out from fabric edges predicted by homogenized yarn model (figure courtesy of Nilakantan et al.6);
(d) slipping through of a sharp nosed projectile; (e) spherical projectile caught by the fabric (figure courtesy of Nilakantan et al.5);
(f) longitudinal and transverse wave propagation in a yarn, transverse compression and shear; (g) fiber–fiber interactions – spreading
and flattening of fibers within a yarn7; (h) fibrillated failure of Kevlar KM2 single fiber.
in trans-stereo-isomeric conformation. The founda- the chains; the only interactions are non-bonded van
tional element of Kevlar fiber is the crystalline lattice. der Waals (vdW) and coulombic. In terms of bonding
Figures 2(b) and (c) shows the orthorhombic PPTA strength, vdW and coulombic interaction are weaker
crystal structure proposed by Northolt and Van than the hydrogen bond and hydrogen bonds are
Aartsen,8 Northolt9 and Tashiro et al.10 This unit cell weaker than the covalent bond that exists between dif-
contains two molecular repeating units per crystal lat- ferent atoms in the chain. Therefore, from a mechanics
tice, one at the center of the cell and four one-fourths at perspective, the crystal is anisotropic and is stronger in
each corner of the cell. In the bc lattice plane, hydrogen the c-axis, weaker in the b-axis and weakest in the
bonds are formed between the neighboring NH and CO a-axis. In the hierarchical scale, the crystals evolve
end groups due to the close proximity of the chains. into macromolecules/fibrils and the fibrils evolve into
In the ab plane, there is no hydrogen bonding between filaments/fibers. Therefore, the fiber properties are

Sockalingam et al. 3
Figure 2. (a) Multi-scale architecture of Kevlar K706 fabric with KM2 fibers. Kevlar crystal lattice8–10: (b) ab plane view; (c) bc plane
view. Atom color: gray – carbon, white – hydrogen, red – oxygen, blue – nitrogen. (Color online only.)
superior in the oriented fiber direction compared to the total of two per unit cell. The precise morphology of
transverse direction. Kevlar fibers have a thermal fibers is a function of many factors, including the struc-
decomposition temperature of 500 C 11 due to the ture prior to deformation and the process by which
high stability of aromatic structures. orientation was achieved. The molecular alignment
PE fibers are made up of extremely long chains of PE developed during orientation strongly influences the
(monomer unit > 250,000 per molecule). The chains properties of the resulting products; higher levels of
become highly oriented along the fiber axis during the orientation magnify the anisotropic response to exter-
deformation process induced by shear, compression or nal forces. Unlike Kevlar, the melting temperature of
tension. At the crystal level, PE exhibits three types of PE is low (145 C).16 Prevorsek et al.16 reported that
unit cell – orthorhombic, monoclinic and hexagonal.12– the projectile fiber frictional interactions result in an
15
The orthorhombic unit cell is by far the most increase in temperature at the projectile/armor interface
common. Each orthorhombic unit cell contains a com- above the melting point of high-performance PE fiber.
plete ethylene unit from one chain segment and part of However, they noted that due to the short time scale
four others from surrounding chain segments, for a during ballistic impact, the temperature rise is limited

Figure 3. Transverse impact.
to a very small region and its effects on the armor per- Theory of transverse impact
formance cannot be detected. Predicting the mechanical
and impact response of these materials and thus design-
Transverse impact: classical Smith theory
ing the protective/containment system is a challenge. The The theory of transverse impact onto a homogenous
impact experiments on yarns and fabrics are not only yarn, shown in Figure 3, was developed by Smith
expensive but also complicated due to the fiber–fiber et al.21,22 The transverse impact initiates a longitudinal
qffiffiffi
and yarn–yarn interactions. Therefore, a combined strain wave that propagates at a speed given by c ¼ E
numerical modeling and experimental approach is typi- and a transverse wave (‘V’ shaped) that propagates
cally adopted to understand the system response. With with a speed an order of magnitude lower than the
advances in experimental techniques and computational longitudinal wave. The angle made by the transverse
power, some of the recent researches have focused on wave (Lagrangian transverse speed U) with the hori-
understanding the fundamental mechanisms during zontal is a constant for the region 0 x Ut. That
impact, particularly at lower length scales. is, the transverse wave is a straight line assuming ten-
The review papers published on this topic include the sion as the only restoring force in the fiber for this
works by Cheeseman and Bogetti,17 Tabiei and analytical solution.
Nilakantan18 and David et al.19 These review papers
introduce the key factors that influence the ballistic per- It is noted that the one-dimensional (1D) analytical
formance of fabrics from a multi-scale perspective such solution for transverse impact onto a fiber or a homo-
that the fabrics comprised of woven yarns can be mod- genous yarn21,23 assumes linear elastic small strains
eled as a homogeneous orthotropic elastic material and with tension in the fiber as the only restoring force
that tensile properties in the yarn direction govern and long fibers (to avoid complications of wave reflec-
dynamic response during impact. There have been lim- tions). The boundary condition is constant velocity (V)
ited reviews on the responsible mechanisms associated impact at a point. Therefore, they do not consider
with different length scales where yarns are considered the effect of projectile geometry, interlaminar shear
as heterogeneous (e.g., KM2 600 denier yarn has 400 (in the plane of fiber direction–impact direction) or
hundred individual fibers), fibers within an impacted the wave propagation through the thickness of the
yarn are loaded non-uniformly, subjected to multi-axial fibers that are loaded in compression directly under
loading and failure is governed by the inelastic trans- the impactor.
verse behavior of the fiber. The goal of this paper is to The strain (implicitly) in the fiber behind the long-
review the recent research in modeling and experiments itudinal wave and the Euler transverse wave speed are
of important mechanisms and deformation modes of given by equations (1) and (2)
these materials (primarily focused on plain woven
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi
aramid fabrics) at different length scales. The theory of
V ¼ c 2" "ð1 þ "Þ "2 ð1Þ
transverse impact is first reviewed. Then the impact
experiments and mechanical characterization experi- pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ments at different length scales relevant to the deforma- ulab ¼ c "ð1 þ "Þ " ð2Þ
tion modes experienced by the material during impact
are discussed. Finally, modeling of ballistic fibrils to where " is the strain in the fiber and V is the impact
multi-layer fabrics is reviewed. The authors recognize velocity. The transverse wave speed increases when
the complexity of penetration mechanics of armor,20 it interacts with a reflected longitudinal wave.22
which will not be considered in detail here. Alternately, the transverse wave speed cs, derived by

Cole et al.24 and Wang25 for the case cs c, is given by observed experimentally. This study clearly shows that
the interactions between the impactor and fibers
c1=3
are important factors governing failure of the fiber as
cs ¼ ðVÞ2=3 ð3Þ
2 well as the experimental set-up required for accurate
characterization.
The ‘V’ angle made by the transverse wave with the It should be noted that classical theory and modified
horizontal direction, as shown in Figure 3, is given by theory do not differentiate between a single fiber and a
equation (4)26 yarn, as they consider only axial tensile loads and the
longitudinal wave speed. As will be discussed later,
1=3
2V the transverse impact of a yarn consisting of many
¼ tan1 ð4Þ interacting fibers is very different. Although the 1D
c
solution provides an understanding of wave propaga-
Classical theory (equation (1)) predicts an impact tion during impact, a more thorough model is needed
velocity of 849 m/s to break the fiber for a failure to understand the fabric system response. Phoenix and
strain of 3.55% for Kevlar KM2 fiber.1 It should be Porwal29,30 developed an analytical solution for a two-
noted that breaking speed is defined as the speed at dimensional (2D) isotropic membrane impacted by a
which the fiber/yarn fails immediately upon contact cylindrical projectile. Their solution provided good
with the projectile. Although experimental breaking agreement with the multi-ply fabric experimental data
speed for single fibers is not reported in the literature of Cunniff31 and closely followed Cunniff’s empirical
to our knowledge, breaking speeds for yarns (KM2, curves. However, a fabric-based membrane is orthotro-
DyneemaÕ SK-65 and PBOÕ ) reported by Walker pic and idealizing fabric architecture as a membrane
and Chocron27 are up to 40% lower than the classical does not allow the lower length scale energy absorbing
breaking speed. mechanisms to be modeled accurately (e.g. yarn com-
pression, yarn pull-out, windowing and friction
between yarns and impactor, etc.).
Transverse impact: modified theory
Walker and Chocron27 extended the classical Smith
Cunniff equation
theory to account for large strains using the first
Piola–Kirchhoff stress measure in the undeformed con- Cunniff31 empirically related the ballistic performance
figuration. For the small strain case their solution is of the armor system (both fabrics and composites of
given by different areal densities) to the fiber mechanical proper-
ties, specifically the product of fiber-specific toughness
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
pffiffiffiffiffiffiffi and acoustic wave speed given by
V ¼ cðE11 Þ3=4 2 E11 ð5Þ

V50 Ad Ap
, ¼0 ð6Þ
where E11 is the axial component of the Green strain. ðU Þ1=3 mp
This extension still assumes axial tensile modulus and
density are the only properties affecting the solution. sffiffiffiffiffi!1=3
The experimental observations of lower breaking 1=3 f "f Ef
ðU Þ ¼ ð7Þ
speed for yarns compared to the classical theoretical 2f f
speed is explained as due to the waves from the edges
of the flat faced projectile that interacts at the center of where V50 is the ballistic limit, Ad is the areal density,
the yarn. Also, homogenized yarn computational Ap is the projectile presented area, mp is the projectile
models showed bounce of the yarn in front of the pro- mass, f is the fiber tensile failure strength, "f is the fiber
jectile4,28 due to elastic collision. Therefore, a relation- tensile failure strain, f is the fiber density and Ef is the
ship is derived between the impact velocity, bounce fiber tensile modulus assuming the fiber behavior is
velocity and strain in the material for a flat faced pro- linear elastic. He plotted V50 against the dimensionless
jectile.27 When the bounce velocity is equal to the parameter that relates the projectile characteristics to
A A
impact velocity, the breaking speed is predicted to be the armor configuration md p p and showed good correla-
11% less than the classical breaking speed. However, tion with equation (7) for different armor systems,
when the bounce velocity is twice the impact velocity except ultra high molecular weight polyethylene
(upper bound for an elastic collision), the breaking (UHMWPE) spectra. It is argued that this deviation
speed is predicted to be lowered up to 40% due to the may be due to the lower melting point of the thermo-
increase in the axial strain. It should be noted that plastic spectra fibers.1 These models also assume that
bouncing of the yarn (in KM2 and DyneemaÕ ) is not transverse compression of the fibers does not contribute

Table 1. Cunniff velocity and theoretical breaking speed for common ballistic fibers
Density Modulus Failure Cunniff Theoretical

Fiber (g/cm3) (GPa) strain (%) velocity (m/s) breaking speed (m/s)
Spectra 1000 0.97 120 3.50 801 1223

Dyneema SK76 0.97 116 3.50 892 1202
Kevlar KM2 1.44 82.6 3.55 682 849
Carbon fiber 1.80 227 1.76 593 744
E-glass fiber 2.89 74 4.70 559 690
to energy absorption or degrade the fiber tensile ballistic resistance of body armor. The V50 ballistic
strength and does not account for the statistical limit is one of the widely used measures of characteriz-
nature of fiber failure. Table 1 shows the Cunniff velo- ing the ballistic performance of armor systems and is
city (equation (7)) and theoretical breaking speed defined as the velocity at which the projectile is
(equation (1)) for some common ballistic fibers with expected to perforate the armor 50% of the time. The
properties from Kim et al.1 It is seen that the theoretical V50 ballistic limit is determined through an experimen-
breaking speed is much higher than the Cunniff velo- tal statistical evaluation of the armor system without
city. While Smith’s equation 1 depends on the wave consideration of material response at lower length
speed and failure strain, Cunniff’s equation depends scales. Consequently, due to the complexity of ballistic
on the wave speed and the fiber-specific toughness impact onto flexible materials involving multiple layers
and thus is more realistic. and length scales, experiments conducted at lower
The relationship between the yarn mechanical prop- length scales can isolate the underlying mechanisms
erties to the ballistic performance of a multi-layer fabric and provide a better understanding. These experiments
is complex32 and it is important to consider the com- can then be used to validate the numerical models at
pressive properties, temperature, fiber–fiber friction the same length scales. This section reviews some of the
and layer–layer interactions. Based on V50 (defined in ballistic impact experiments on fabrics and mechanical
the next section) testing, Cunniff’s equation is found to characterization experiments on these materials at
be in agreement for Kevlar fibers with low failure strain different length scales and strain rates.
(3.6– 4.4%), which showed fibrillated failure.32
However, Cunniff’s equation is not in agreement for
experimental poly(ethylene naphthalate) (PEN) fibers
Fabric length scale experiments
with high failure strain (8–18%), which showed melt- The textile fabrics can be woven from yarns in many
ing,32 and UHMWPE composites,33 which showed a different styles. Dong and Sun35 list five different styles
shear strength-dependent ballistic limit. It should be of plain woven Kevlar fabrics using different Kevlar
noted that melting and shear was not a mechanism fibers with their axial modulus. For example, Kevlar
considered by Cunniff. K706 fabric consists of KM2 fibers, 600 denier yarns
While the analytical models provide useful insights and 34 yarns per inch in the warp and weft directions.
and indicate the relative ballistic performance of fibers, The K706 fabric has a thickness of 0.23 mm and a span
they do not account for many other contributing of 0.747 mm. Traditionally, ballistic impact experi-
factors, including compressive properties, interlaminar ments on the fabric systems focused on the relationship
shear, strain rate effects, fiber-specific failure modes, between the impact velocity and residual velocity for a
projectile–fiber interactions, the effect of projectile given projectile and determining the V50 ballistic limit.
geometry, fiber–fiber and yarn–yarn interactions, However, some of the recent work looked at under-
three-dimensional (3D) stress states, multi-axial load- standing the wave propagation, failure mechanisms
ings and so on. To understand these effects, one must related to projectile geometry and progressive fiber fail-
consider modeling and experiments of these deforma- ure, as discussed below.
tion mechanisms at the appropriate length scales. A qualitative description of the fabric impact
mechanics is provided by Cunniff.36 Longitudinal
Experiments on fibrils, fibers, and transverse waves develop and propagate in the
yarns in a fabric that are in direct contact with
yarns and fabrics
the projectile during impact. The longitudinal waves
At the fabric macro length scale, National Institute of in the primary yarns drive the longitudinal waves in
Justice (NIJ) standard 0101.0634 provides minimum the secondary yarns, as shown in Figure 4, thus provid-
performance requirements and test methods for the ing an effective transverse impact on the secondary

Figure 4. Fabric (plain woven) impact mechanics: (a) top view; (b) side view (displacement exaggerated).
yarns. The secondary yarns constrain the motion of the information can thus be used to validate the fabric
primary yarns, resulting in tensile strain gradient as properties of a fabric length scale model.
observed by the significant curvature during transverse The projectile geometry not only affects the ballistic
deflection of a single ply fabric. The longitudinal wave limit, but also plays a major role on the energy absorp-
c
speed in a fabric is approximately cfabric ¼ pyarnffiffi2 .37 A simi- tion behavior and failure mechanisms.40–43 Yarn pull-
lar explanation is provided by Chocron et al.37,38 based out is reported most for hemispherical and least
on their measured strains in the secondary yarns con- for ogival and conical projectiles, as they can slip
sistent with Cunniff’s description. They measured through the fibers due to their sharp nose, as shown
strains in a multi-layer Kevlar fabric during ballistic in Figure 1(d). Fibrillation is a predominant failure
impact by embedding nickel-chromium (NiCr) wires mechanism for sharp nosed projectiles where splitting
in the secondary yarns. NiCr high-resistivity wire has of fiber occurs along and perpendicular to their length.
the property to change its resistance (and voltage) due This fibrillation is attributed to the breaking of second-
to applied strains. The pyramid transverse wave propa- ary hydrogen bonds in aramid fibers.40 Frictional
gation during the impact of a flexible fabric is experi- mechanisms, including fiber breakage at the yarn
mentally observed for both traditional fabrics28 and cross-overs, flattening and rupture of fibers and yarn
unidirectional fibers held together by a polyurethane pull out, are also physically observed.40 Severe flatten-
matrix (DyneemaÕ HB80).39 The transverse wave ing of fibers is observed for a hemispherical and flat
speed, unlike longitudinal wave speed, depends on headed projectile impact.40 Significant transverse per-
both material properties (longitudinal wave speed) manent deformation of the fibers is observed from
and loading conditions (impact velocity) and is con- post failure investigation of impacted fabrics.44,45 In a
firmed by these experiments. The wave propagation multi-layer plain woven fabric system, the first few

layers are subjected to transverse shear and compres- A nonlinear force displacement response with increas-
sion.36 While the fibers in the front layer closer to the ing stiffness is observed for QS transverse compression
impact site are found to fail in the shear failure mode tests on 28 plies of Kevlar KM2 fabric.60 A bilinear
due to the edge of a cylindrical flat ended projectile, load displacement response is observed during a static
fibers in the rear layer display fibrillation due to tensile transverse indentation of Kevlar fabrics with an initi-
failure.46 Projectile obliquity is also found to influence ally compliant response due to the decrimping of yarns
the ballistic limit.47 These experimental observations on and a stiffer response due to the stretching of yarns.56,61
projectile geometry-dependent failure mechanisms are In-plane fabric shear, necessary for the homogenized
critical in setting up lower length scale experiments to membrane fabric model, is typically determined
better understand these interactions. through a picture frame test.62–64
Several researchers report the influence of boundary The fabric length scale impact experiments reviewed
conditions on the impact response of fabric systems36,48,49 in this section indicate complicated failure modes
and also note slippage of fabric from clamped edges.49,50 and mechanisms observed, including transverse shear
In general, slippage leads to a decrease in primary yarn and compression of the initial layers, pyramidal wave
tension, resulting in higher energy absorption than with- propagation, yarn pull out, inter-yarn friction, progres-
out slippage. Parsons et al.50 report the usage of knurled sive fiber failure, slipping through of the sharp nosed
clamps to minimize slipping of the fabric. The nature and projectiles and fibrillation. Other fabric length scale
manufacturing of the textile fabrics and armor system mechanical characterization experiments discussed
allows infinite possibilities of configurations to achieve above also reveal complicated mechanisms occurring
optimum ballistic performance through a materials-by- at lower length scales, such as crimp interchange,
design approach. The fabric weave type,51,52 weave den- strain rate-dependent properties and failure. Lower
sity,53 number of filaments54 and hybrid fabric panels55 length scale modeling and experiments are needed to
are some of the design variables that affect the ballistic better understand these mechanisms. For instance, to
performance. study the effect of different fabric architecture and
To gain more insight into the fabric deformation inter-yarn friction, yarn length scale modeling resolu-
behavior, researchers conducted mechanical character- tion is required. This, in turn, necessitates yarn length
ization experiments, including quasi static (QS) and scale experiments to characterize the yarn behavior
dynamic tension, in-plane shear and transverse com- needed as input to the numerical models.
pression tests. A bilinear load displacement uniaxial
response – a low modulus decrimping region followed
by a high modulus elongation regime – is observed for
Yarn length scale experiments
silica nanoparticle impregnated fabrics.56 The deforma- A yarn consists of a bundle of minimally twisted fibers
tion in a fabric is governed by yarn interactions at the with fiber counts in the range of 100–10,000. The yarns
cross-over points. While the yarns in the direction of differ in denier, number of filaments and tenacity.
tensile loading straighten, orthogonal yarns undergo Denier is yarn weight in grams per 9000 m11 and tena-
additional crimping. This phenomenon is often referred city is force in grams required to break a yarn.19
as crimp interchange. A strain rate-dependent stiffness, For example, a Kevlar KM2 600 denier yarn consists
strength and failure strain is exhibited by Twaron and of 400 individual fibers. The transverse impact experi-
Kevlar fabrics.57,58 While the stiffness and strength ments reported in the literature are at the yarn length
increased with strain rate, failure strain decreased, scale26 and fabric length scale,40 as it is extremely
with brittle failure at high strain rates and a combina- challenging to capture real time information at the
tion of brittle failure and plastic deformation at low fiber (micron) length scale. The dominant role and con-
strain rates. The energy absorption at high strain tribution of principal yarns (yarns directly in contact
rates is significantly lower than at low strain rates. with the projectile) to energy absorption and projectile
This is attributed to the intermolecular slippage and deceleration is reported by several researchers.17,65
plastic flow at low strain rates resulting in more Fibers belonging to the principal yarns directly under-
energy absorption. neath the impact and at yarn cross-overs are subjected
Fibers directly underneath the projectile and at yarn to significant transverse compressive stresses.
cross-over points are expected to have substantial The current state-of-the-art experimental capabilities
transverse stresses.59 The transverse stresses may be in yarn transverse impact testing does not have the
amplified depending on specific boundary conditions, spatial resolution to track individual single fiber
for example multi-layer impact scenarios with surrogate (KM2 fiber is 12.0 mm in diameter) deformations in
backing, which constrain the yarn out-of-plane displa- real time. The high-speed camera imaging allows mea-
cements.60 However, there is limited investigation of suring transverse wave speed and the failure mode. The
transverse properties of ballistic fabrics in the literature. transverse wave propagation speed is measured,

Figure 5. (a) Transverse wave speed (equation (3)) and wave angle (equation (4)) for KM2. (b) KM2 yarn transverse impact (figure
courtesy of Song and Lu66). (c) Fragment simulating projectile impact (figure courtesy of Chocron et al.28). (d) Blunt nose impact
(figure courtesy of Bazhenov et al.67).
assuming the yarn to be homogeneous (that is, indivi- yarn transverse impact experiments. They further
dual fiber–fiber interactions are not considered). noted that ‘yarn behaves simply as a group of indivi-
The transverse wave velocity from Smith theory (equa- dual fibers instead of a solid component bundling all
tion (2)) is shown to match approximately with the fibers’. While the transverse wave in aramid yarns
experimentally measured velocity for different yarns, appears to make a sharp transition at the horizontal
including KM2, Dyneema HB80 and PBO.28 The direction for a sharp edge impact (fragment simulating
experimentally measured Euler transverse wave speeds projectile (FSP)), curvature and bending of the fibers
and transverse wave angles increase nonlinearly with are observed for a blunt nosed projectile impact, as
impact velocity, as shown in Figure 5(a), using a shown in Figures 5(c) and (d).
wave speed of 7639 m/s for Kevlar KM2 yarns.26 The The yarns failed in a transmitted stress wave mode
influence of the presence of 20% matrix in a (after the development of a transverse V wave) for low-
UHMWPE Dyneema HB80 strip is argued to slow impact energies or a shear ‘plug’ mode for high-impact
qffiffiffiffiffiffiffiffiffi
1 dT energies.68 It is also found that aramid fibers failed in
down the wave speed c ¼ M d" by a factor of 0.91, fibrillation and UHMWPE fibers failed in shear with
where M is the mass per unit length, T is the tensile shear bands and melt damage. Aramid yarns showed
force and " is the strain.39 This wave propagation infor- increased specific energy absorption at low-impact ener-
mation can be used to validate the properties of a yarn gies than at high-impact energies due to increased fibril-
length scale model. Most recently, Song and Lu66 noted lation at low-impact energies. However, UHMWPE
that fibers are impacted progressively along the thick- yarns showed an opposite trend with higher energy
ness direction (see Figure 5(b)) based on their KM2 absorption at higher impact energy. In general,

experimentally observed breaking velocity of yarns is Table 2. Properties of Kevlar KM288,89 and Spectra 900 single
much smaller than the theoretical breaking velocity.27 fiber.90
This is attributed to the projectile–fiber interactions,
Aramid Polyethylene
multi-axial loading and progressive and statistical Property (Kevlar KM2) (Spectra 900)
nature of fiber failure which, is not accounted for in
the analytical solution, as discussed earlier. Other trans- Density (g/cm3) 1.45 0.97
verse impact experiments on yarns include the work Diameter (mm) 12.0 38.0
by Field and Sun,69 Heru Utomo and Broos70 and Axial tensile modulus E3 (GPa) 84.62 117.21
Bazhenov et al.67 Transverse modulus E1 (GPa) 1.34 -
The QS yarn behavior and strength has been inves- Longitudinal shear 24.40 -
tigated by several researchers;48,71–74 it can feed into modulus G13 (GPa)
homogenized yarn resolution models, discussed later. Poisson’s ratio n31 0.60 -
In general, a linear behavior is displayed by the yarns Poisson’s ratio n12 0.24 -
and a lower strength is observed at higher gage lengths. Axial tensile strength 3.88 2.60
The yarns also exhibit a statistical distribution in
Elongation (%) 4.50 3.60
strength due to inherent strength distribution and mis-
Axial compressive strength 0.68 -
alignment of the fibers leading to progressive filament
failure at different locations of the gage length.72 Longitudinal qffiffiffiffi 7639 10,992
wave speed (m/s) c ¼ E3
Dynamic yarn tensile tests are often conducted using
a split Hopkinson tension bar. Dynamic tensile testing
of yarns indicated weak dependence of strength on
strain rate (800 s–1) for Kevlar and a strong dependence and mechanisms are governed by fiber–fiber inter-
for Twaron and Zylon fibers.75 In another study, an actions, fiber properties and microstructure. To better
increase in modulus, strength and failure strain is understand the fundamental mechanisms during
observed at a strain rate of 480 s–1 compared to QS impact, research in the direction of understanding
for Twaron yarns.73 In addition, for Twaron yarns, fiber–fiber interactions, failure initiation and propaga-
longer fibrils are observed at low strain rates and tion and post-failure investigation is needed.
shorter fibrils at high strain rates, which are explained
due to inter-chain slippage secondary bond failure and
Fiber length scale experiments
primary bond failure, respectively.73 A similar increase
in the tensile properties is observed for Kevlar 49 yarns High-performance polymer fibers lend themselves to
for strain rates in the range of 20 to 100 s–1.74 ballistic applications due to their superior specific
The strength, strain energy and failure mechanism of strength, specific modulus and energy absorbing prop-
poly para-phenylene terephthalamide (PPTA) and erties. The properties of an aramid (Kevlar KM2) and a
UHMWPE fiber bundles are more strongly dependent PE (Spectra 900) single fiber are listed in Table 2.
on the fiber microstructure and morphology than These fibers, in general, exhibit transverse isotropy
the strain rate.76 Fibrillation is the failure mechanism with linear elastic behavior in longitudinal tension
for PPTA fibers at both low and high strain rates, and nonlinear inelastic behavior in transverse (diame-
whereas crazing is the dominant mechanism for tral) compression.87,88
UHMWPE fibers at high strain rates. Some yarns There is an increasing interest to understand the
show strong rate-dependent tensile properties, while strain rate-dependent tensile response of ballistic fibers
others show weak dependence. These results clearly as they are subjected to high strain rates during impact.
show that the properties strongly depend on the micro- QS and dynamic single fiber tensile tests88,91–94 are often
structure and morphology of the fibers that require used to characterize the fibers in axial tension. Most of
further investigation. the recent researches on dynamic fiber tensile testing are
QS5,77–82 and dynamic yarn pull out,83,84 influence of conducted using a miniaturized Kolksy bar.88 The
twist85 and inter yarn friction86 studies are other experi- PPTA fibers, in general, show a strain rate-dependent
ments conducted at the yarn length scale to understand strength behavior. In addition, the fibers exhibit gage
the energy absorption during impact. Most of the yarn length dependent strength, statistical distribution of
transverse impact experiments reported in the literature strengths and diameter dependent strength at high
are focused on determining the transverse wave velocity strain rates.95,96 Kim et al.95 measured an increase of
and the ‘V’ angle from the images obtained. Both the 14% average tensile strength at high strain rate (HSR)
transverse impact and mechanical yarn characterization loading (103 s–1) compared with QS loading for Kevlar
experiments treated yarn as a homogeneous material. 29 fibers. However, Cheng et al.88 showed that the long-
However, the experiments indicate the failure modes itudinal tensile properties of Kevlar KM2 fibers are

Figure 6. Schematic of (a) single fiber transverse compression, (b) torsional pendulum longitudinal shear,88 (c) micro-compres-
sion104 and (d) elastica loop test.89
insensitive to strain rates. Fibrillation is observed as the to the Mullin’s effect in rubber materials, Singletary
main failure mode for both QS and HSR loading. et al.100 attributed the observed force deflection pattern
Figure 1(h) shows the image of a fibrillated Kevlar for staple fibers as cracking of the fiber, followed by
KM2 fiber subjected to QS tension. UHMWPE fibers closing of the cracks under compression that leads to
also display a strain rate-dependent strength behavior.97 the overall inelastic response. They also hypothesized
For instance, Dyneema SK76 fibers exhibited an overall transverse yielding followed by fibrillation for the force
increase in tensile strength with strain rates (1–103 s–1). deflection pattern observed for 1.5 denier Kevlar 29
A transition from viscoelastic to elastic behavior is fibers. These fibers exhibit nonlinear inelastic behavior
exhibited by the fibers from low to high strain rates in transverse compression with a small transverse elas-
with fibrillation at high strain rates. Even with these tic limit.102,103 Wollbrett-Blitz et al.103 determined the
experiments, failure initiation and propagation leading transverse compression elastic limit for Kevlar 29 fibers
to final failure of a single fiber is not well understood. as 0.25 N/mm, below which no permanent deformation
Similar to yarn axial tensile results, these single fiber was observed. However, most of the finite element (FE)
results also show that the properties strongly depend modeling reported in the literature at different length
on the microstructure and morphology of the fibers scales (fiber, yarn and fabric) assumes a transversely
that require further investigation. isotropic linear elastic behavior for the material. The
Researchers recognized the importance of under- role of inelastic transverse compression fiber behavior
standing deformation modes (see Figure 6) other during impact is not well understood.
than axial tension that are occurring during an The general approach adopted to estimate the trans-
impact. Several researchers have investigated the verse modulus is to use the load–displacement measure-
single fiber transverse compression response of Kevlar ments in conjunction with the elastic small strain
fibers, including Kawabata,98 Singletary et al.99,100 and analytical solution based on the Hertzian contact
Cheng et al.87 and limited investigations on UHMWPE models between parallel plates.105,106 An elastic modulus
fibers.101 The experimental results reported by these of 1.34 0.35 GPa and 2.45 0.40 GPa is estimated for
researchers indicate that fibers exhibit nonlinear and KM287 and Kevlar 2999 fibers, respectively. Both geo-
inelastic response under large compressive strains. metric and material nonlinearities are embedded in the
While Cheng et al.87 compared this nonlinear behavior material response, requiring the measurement of contact

width to determine the accurate material behavior. They also reported a significant 50% reduction in
However, the difficulty associated with the measurement tensile strength for imperfectly compressed fibers due
of contact width for smaller fiber diameters is noted by to misaligned grips. The effect of fibrillation caused by
Cheng et al.87 and Kotani et al.107 For larger fiber dia- transverse and axial compression on longitudinal shear
meters, Pinnock et al.108 estimated the transverse elastic is not well understood. A finite longitudinal shear stiff-
moduli and Poisson’s ratio using the analytical solution ness results in a flexural wave during impact and is
for diametral compression109 and contact width.106 The discussed later.
observed force displacement behavior and large residual The effect of twisting of fibers caused by weaving on
strains in transverse compression can be a significant their strength is recently studied by longitudinal testing
energy absorption mechanism in ballistic impact applica- after pre-twisting.113 The study indicates that at low
tions. Although there is some understanding of the QS and intermediate rates, the Kevlar KM2 fibers retained
transverse compression response of fibers, the HSR 95% of its tensile strength for a preloaded shear strain
response and the role of transverse properties during of 0.15. However, for shear strains greater than 0.15 a
impact is not well understood. drastic reduction in the tensile strength of the fibers
The analytical treatment of classical theory, discussed (30% retention in strength at a shear strain of 0.45) is
earlier, assumes axial tension is the only force in the observed due to the surface flaws. A 20% decrease in
fiber. However, experimentally observed curvature in strength for warp fibers at all high strain rates com-
the transverse deflection of a fabric and weaving process pared to unwoven (virgin) fibers and a 35% decrease
may induce axial compression and shear. Single fiber in strength at QS strain rate114 is observed. Hudspeth
axial compressive moduli and strength for different et al.115 also observed a similar reduction in tensile
high-performance fibers, including Kevlar strength of Dyneema SK76 fibers subjected to torsional
KM2 (0.68 GPa), is reported by Leal et al.89 (see strains and developed a biaxial torsion-tension failure
Figure 6(c)). Fibers are found to have different moduli surface. The tensile strength of pre-torsioned fibers is
in compression than in tension (Ec/Et ¼ 0.61 for KM2) . found to be slightly higher at high strain rates than QS
Kevlar KM2 fibers in elastic loop experiments (see loading. These initial experimental studies clearly show
Figure 6(d)) loaded in flexure are damaged in compres- the significant effects of multi-axial loading on failure
sion through the formation of kink-bands that causes and may serve as input to fiber-level resolution models
fibrillation of the fiber microstructure. Micro cracking to accurately predict failure.
and fibrillation due to kink bands in Kevlar KM2 fibers Hudspeth et al.116 examined the failure surfaces of
are also observed, resulting in a 10% loss in tenacity.110 KM2 and DyneemaÕ SK76 single fibers subjected
The aramid Kevlar fibers also have finite longitudinal to different indenters and failure angles (range of
shear moduli, 29 GPa for Kevlar 29111 and 24.4 GPa for 0–50 deg), including a razor blade, 0.30 caliber FSP
Kevlar KM2,88 based on the torsional pendulum experi- and a round indenter. They observed a shear failure
ment112 (see Figure 6(b)). While most of the literature with minimal axial splitting of the fiber for a razor
on the impact of flexible textile composites focuses on blade indenter and a fibrillation dominated failure for
tension (axial loading), the consideration of bending a round indenter for all tested angles. The FSP indenter
and its effects is limited. showed a transition in the fracture surface from fibril-
The multi-axial stress state of the fiber is expected lation failure at low angles to shear failure at high
under ballistic conditions. However there have been angles. Their study also showed approximately con-
limited studies on the multi-axial loading response of stant failure strain for a razor blade and round inden-
fibers in the literature. Cheng et al.87 investigated the ters, whereas there were decreasing failure strains for
response of Kevlar KM2 fibers for bi-axial loading con- the FSP indenter with increasing angles. Hudspeth
ditions by conducting both longitudinal tensile tests et al.117 showed a reduction in the KM2 fiber failure
after transverse compression and transverse compres- strain due to the multi-axial stress state and local stress
sion tests after longitudinal tension. They reported concentration induced by the fiber angle around the
insignificant degradation of tensile strength after trans- FSP indenter.
versely compressing the fibers to a nominal strain Although traditionally fiber failure during impact is
of 0.48, thus allowing the fibers to resist impact thought to be tensile dominated, recent research as
loading. Their results on the latter tests indicated long- reviewed in this section looked at other deformation
itudinal pre-tension increases the transverse stiffness modes, including transverse compression, longitudinal
for nominal strains greater than 0.1. In a later study, shear and multi-axial loading. Characterization of
Lim et al.96 reported a reduction of about 10% fibers at different loading conditions and strain rates
tensile strength (tested at a strain rate of 1500 s–1) for are essential to understand and accurately predict
Kevlar and Kevlar 129 fibers after being transversely fiber failure during impact. However, there is limited
compressed to about 80% of the fiber diameter. research on the strain rate-dependent multi-axial

loading response of ballistic fibers. One of the reasons is scratch force to normal indentation force) is reported120
that the fiber length scale experiments are challenging to increase up to 300% higher than the Kevlar yarn–
due to the associated micron length scale and short time yarn friction of 0.2–0.3. Through these studies, appar-
scales. Numerical modeling can help not only to under- ent fiber–fiber and yarn–yarn friction were identified as
stand the complicated experimental results, but also the key energy dissipative mechanisms due to increased
guide the specimen design for the development of new apparent friction associated with particle gouging.
experimental methods. Fibrillation is a predominant UHMWPE is made up of extremely long chains of
failure mode observed not only in single fiber testing PE (monomer unit > 250,000 per molecule) with a hier-
but also in dynamic yarn tensile testing and fabrics archy of sizes. Macro-fibrils129 consist of bundles of
subjected to impact. The fiber microstructure needs to micro-fibrils, which in turn are composed of bundles
be studied to better understand this failure mode. of nano-fibrils. Nano-fibrils consist of stacks of crystal-
lites separated by thin non-crystalline plates, portions
of which are spanned by inter-crystalline bridges, giving
Ballistic fibrils
a shish-kebab structure.130 PE ballistic fibers exhibit a
The fibrillar core-shell structure of Kevlar aramid fibers shish-kebab crystalline structure reported through
based on PPTA has been studied by various research- AFM characterization by Strawhecker and Cole.131
ers.118,119 An atomic force microscopy (AFM) phase It was mentioned earlier that post failure investiga-
image of a Kevlar KM2 fiber obtained by McAllister tion of impacted fabrics showed deformation and fail-
et al.120 is reproduced in Figure 2. The fibril diameters ure modes, including permanent transverse compressive
are in the range of 100–200 nm for Kevlar 491,21 deformation, fibrillation and transverse shear (cutting)
fibers and 10–50 nm for Kevlar KM2 fibers.120 failure. Aramid and UHMWPE fibers can sustain a
Nanoindentation techniques are used to characterize large amount of transverse compressive strain, likely
the 3D properties of Kevlar KM2,119,122 considering due to their fibrillar network. Figures 7(a) and (b)
the influence of fibril-level heterogeneity on properties. show scanning electron microscopy (SEM) pictures of
The local compressive elastic moduli are dependent on a DyneemaÕ SK76 single fiber and a DyneemaÕ fiber
the number of fibrils loaded by the nano-indentor. deformed by a razor blade, respectively. Fiber is seen to
Using this technique, an indenter size-independent spread along the length of the blade. When subjected
value of 6.2 1.0 GPa is measured for the local trans- to transverse compression, KM2 and DyneemaÕ
verse modulus. The measured local transverse modulus SK76 fibers spread to a great extent, with the com-
is about four times higher than the transverse modulus pressed width to diameter ratio attaining 3.5 and 4.5,
(1.34 GPa) measured at the fiber length scale (see respectively, at 70% nominal strains, as shown in
Table 2). This is because the local modulus is measured Figures 7(c) and (d). These lower length scale experi-
using the unloading slope of a number of fibrils that ments at the fiber and fibril levels provide insights into
were permanently deformed. It should also be noted, in the deformation and failure modes that occur during
transverse compression at the fiber length scale, the impact and allow for the design of new materials with
fiber displays negligible strain recovery upon unload- improved properties. For example, nanoindentation
ing. Using similar nano-indentation techniques, Cole studies124 indicated the highest levels of fiber–fiber
and Strawhecker123 reported local mechanical proper- friction are achieved with a minimum particle size,
ties for four fibers, namely KM2Plus, Twaron, AuTx which then becomes a material design parameter.
and Dyneema. These experiments also clearly indicate the presence
The nanoindentation and nanoscratch124,125 studies of multi-axial loading and the importance of account-
are also important to understand the energy absorbing ing for them in the modeling failure criterion to accu-
mechanisms in particle impregnated fabrics using shear rately predict failure.
thickening fluid,126 silica nanoparticles56,127 and carbon
nanotubes (CNTs),128 among others, to improve the Modeling of fibrils, fibers, yarns,
impact performance. These studies use probes that
effectively mimic the contact in the particle–fiber and
single layer and multi-layer fabrics
reveal the effects of fundamental particle–fiber interac- Validated numerical models with predictive capability
tions on the energy dissipative mechanisms, such as are valuable engineering tools for the design and devel-
friction and particle gouging. For instance, a number opment of protective systems and to gain insights into
of fibrillar deformation mechanisms in KM2 were iden- the mechanisms governing the impact response.
tified to optimize the specific energy of indentation and Mechanical characterization experiments at different
scratching, such as transverse compression of fibrils, length scales and strain rates are imperative in the
axial fibril tension and inter-fibril shear and friction.124 development of high-fidelity numerical models. FE
The apparent friction (defined as the ratio of lateral modeling using commercially available explicit FE

Figure 7. Scanning electron microscopy pictures: (a) DyneemaÕ SK76 single fiber; (b) DyneemaÕ fiber deformed by a razor blade
(figure courtesy of Marissen132) (c) KM2 fiber compressed at 77% nominal strain; (d) DyneemaÕ SK76 fiber compressed at 71%
nominal strain.
codes is the most common numerical modeling techni- capture mechanisms such as yarn–yarn friction,
que to predict the impact response. unraveling of yarns, slipping through of the projectile
and woven architecture, as shown in Figures 1(b), (c)
and (d).
Fabric length scale FE modeling of ballistic fabrics
Another approach is to use an empirically based
Most of the fabric length scale modeling in the litera- constitutive model (based on fabric-level experiments)
ture looked at comparing the model predictions of pro- and model the fabric as an equivalent continuum using
jectile residual velocity to the experimental data. Unit shell elements, as shown in Figure 8(a).138,139 Yet
cell representative volume element (RVE) techniques another approach61 adopted a classical laminated
based on shell element modeling of fabrics have been plate theory to develop a homogenized continuum con-
adopted by various researchers133–136 (see Figures 8(a) stitutive model for Kevlar fabrics using 2D plane stress
and (b)). The approach here is to model fabric at the elements in ABAQUS. Other homogenized modeling
continuum scale accounting for the yarn interactions approaches include the works by Nadler et al.,140
through contact forces computed from equilibrium at Boljen and Hermaier,141 Grujicic et al.142 and Erol
the cross-overs with zero through thickness stress and et al.143 Here the fabric in-plane behavior is decoupled
strain. One of the challenges for membrane modeling from the thickness direction, assuming zero out-
approaches is to develop constitutive models that accu- of-plane normal stress. Given the significant transverse
rately describe the lower length scale deformation compressive deformation observed in fibers and also
mechanisms. A membrane element model137 with predicted by fiber-level models (discussed later), it is
viscoelastic and strain rate-dependent behavior of the imperative for the homogenized models to carry over
Twaron fabric showed good agreement between that information at higher length scales to accurately
the numerical and experimental residual velocity. capture all the sources of energy absorption.
However, the limitations of this approach, as also Most of the fabric-level modeling does not incorpo-
recognized by Lim et al.,137 include the inability to rate the statistical nature of failure. Nonetheless

Figure 8. Modeling at different length scales: (a) homogenized fabric model;139 (b) multi-layer fabric impact;133 (c) homogenized yarn
length scale;144 (d) combined global/local model;145 (e) digital element method (DEM);146 (f) fiber-level beam element fabric model;147
(g) fiber-level three-dimensional (3D) yarn model;7 (h) comparison of homogenized yarn and fiber-level 3D models.7
researchers, in general, showed good correlations of the tensile properties of yarns that are extracted
of model predictions to experiments (for example, resi- from a shoot pack away from the impact location.149
dual velocity) by introducing calibration parameters Strain wave reflection, noted by Freeston and Claus,150
such as element removal criterion (erosion strain). at the yarn cross-overs can increase the strain level
The major disadvantage of these modeling approaches depending on the reflection coefficient. Although the
is the introduction of parameters that must be fit to length and width of the fabric are larger than the thick-
impact experiments on each fabric type of interest. ness, impact is a ‘local’ event causing significant trans-
A key assumption in homogenized modeling is the pro- verse deformation initially, which is largely neglected
blem of interest must have length scales greater than the by the continuum scale homogenized approaches.
characteristic yarn spacing (for example, in a plain In general, information related to lower length scale
weave Kevlar K706 fabric, the yarn spacing is mechanisms, shown in Figure 1, is lost in homogeniza-
0.75 mm). However, the impact event is more ‘local’ tion. A holistic multi-scale modeling methodology
and the relation to the characteristic projectile geome- based on fundamental fiber properties, taking into
try should be taken into account. For instance, a sharp account all the important energy absorbing mechan-
nosed projectile can push through the fibers within a isms, is needed for a truly predictive model. A yarn
yarn40 and also a projectile may ‘window’ through length scale modeling resolution is needed to accurately
the fabric with no yarn failure.148 The ‘local’ nature predict the yarn pull out and inter-yarn friction
of the impact event is indicated by a small degradation mechanisms.

Yarn length scale FE modeling of ballistic fabrics This approach leads to a significantly reduced compu-
The early yarn length scale model developed by tational expense compared to a full 3D yarn length
Roylance151 and Cunniff et al.152 modeled the warp scale model. Similar yarn length scale and multi-scale
and fill yarns with a pair of nodes connected by mass- modeling approaches include the works by Gogineni
less elements, with the nodes being coupled through a et al.,165 Gu,166 Sun et al.,167 Grujicic et al.,142 Jia
spring. The transverse compression of the yarn is et al.,168 Yen et al.169 and Jiang et al.170 There is limited
accounted for with a nonlinear spring and bending stiff- yarn-level multi-scale modeling that takes into account
ness with a torsional spring. The yarn crimping signifi- the fiber–fiber interactions and fiber properties other
cantly slows down the transverse wave speeds and than axial tension. Zohdi and Powell171 and Zohdi172
yarn–yarn interaction via coupling springs has a developed a multi-scale model where yarns are modeled
major effect on the ballistic response.152 Pin-jointed vis- as a 2D network. The behavior of the yarn is based on a
coelastic bar elements153,154 are used to model the single fiber behavior, taking into account the misalign-
woven aramid fabric and are reported to be able to ment of fibers. The random misalignment of fibers
capture the sliding of yarns and hence wedging of the results in progressive failure of fibers and also causes
projectile through the fabric. The pin-jointed approach difference in the response from yarn to yarn. A similar
indicates ballistic response is affected by the coefficient multi-scale approach adopts a handshaking process
of friction between the yarns.155 Novotny et al.,156 between the fabric, yarn and fiber length scales.173,174
using a similar approach, concluded the smallest Although limited, there is some work on the prob-
inter-layer gap results in the highest rate of initial spe- abilistic modeling approach by mapping the statistical
cific energy absorption. The pin-jointed approach does strength distribution of yarns onto the yarn length scale
not necessarily account for the friction at the yarn FE model to capture the probability of fabric penetra-
cross-overs, accurate transverse compression and tion as a function of the projectile impact velo-
fiber–fiber interactions. city.175,176 For the yarn-level orthotropic continuum
The meso-scale yarn-level 3D orthotropic continuum FE models, all the Poisson’s ratios are assumed to be
model is developed by explicitly modeling all the yarns zero and the transverse and shear modulus are assumed
in a fabric using solid FE elements, as shown to be three orders of magnitude smaller than the long-
Figure 8(c).48,144,157,158 Modeling at the yarn length itudinal modulus (and are also an order of magnitude
scale using solid elements allows capturing the yarn lower than the transverse modulus of an individual
undulations and inter-yarn friction directly, unlike the fiber) to represent the thread-like yarn behavior.177
pin-jointed bar element models. Yarn-level FE models An actual 600 denier Kevlar KM2 yarn consists of
also allow understanding the effect of projectile shape 400 individual 12 mm diameter fibers. These continuum
and friction between the yarns during impact.159,160 approaches do not accurately account for the trans-
The yarn–yarn friction is reported144,161,162 to play a verse compression response of yarns where fiber com-
major role in the energy absorption during impact by pression, progressive fiber failure, fiber–fiber contact
increasing the impact load. However, it is reported by and friction and fiber spreading within a yarn are
many researchers that this level of detailed modeling is important physically observed deformation mechan-
computationally very expensive. isms. Again a multi-scale modeling methodology
Researchers also employed multi-scale global/local taking into account all the important energy absorbing
approaches combining yarn and fabric length scales, mechanisms is necessary for a truly predictive model.
shown in Figure 8(d), leading to computationally effi-
cient models. Here the woven architecture is modeled
around the impact site and fabric length scale (mem-
Fiber length scale FE modeling of ballistic fabrics
brane shell elements) away from the impact location.163 At the fiber length scale, Durville178 developed an
The two length scales are connected by tied interfaces. enriched kinematical 3D beam model to represent
Rao et al.164 used 3D solid elements with yarn length deformation of the fiber cross-section considering con-
scale around the impact site (local) and homogenized tact and friction between the fibers. Wang et al.146
solid elements at the far field (global) region. This developed the digital element method (DEM) to simu-
approach maintained the same areal density for both late the ballistic impact of textile fabrics wherein the
local and global regions and matched impedances hundreds of fibers in a yarn are represented as an
across the interfaces to avoid spurious wave reflections. assembly of digital fibers using a few rod elements con-
Nilakantan et al.145 developed hybrid element analysis nected by frictionless pins, as shown in Figure 8(e). The
(HEA), which used solid elements for the yarn length DEM uses contact elements to model contact between
scale around the impact site (local) transitioning into the fibers, the contact stiffness for fiber–fiber
woven shell elements and finally homogenized mem- compression being calculated using the fiber transverse
brane shell elements for the far field (global) region. modulus. A later study implemented a nonlinear

elastic-plastic model for the contact forces between the model predictions were consistent with an analytical
digital elements and concluded that energy loss due to solution derived based on the theory of flexural waves
fiber transverse plastic deformation significantly affects in thin rods for an infinite Euler–Bernoulli beam sub-
the ballistic resistance of a fabric.179 While this jected to a constant velocity impact at its mid span.
approach may be effective in eliminating interpenetra- They further studied the yarn transverse impact7 at
tions, it does not account for the Poisson’s ratio effect the fiber length scale by explicitly modeling all the
in the plane of the isotropy, evolution of a large contact fibers in a yarn using a 3D FE model, shown in
width observed experimentally96 and contact forces Figures 8(f) and (g). At the short time scale the fibers
between fibers. undergo significant transverse compressive deformation
In an approach similar to the DEM, the 3D beam (80% true compressive strains at 500 m/s impact)
elements shown in Figure 8(f) are employed to model sufficient to induce fibrillation with little spreading
Kevlar KM2 fibers with a user-defined contact algo- resulting in multi-axial stress states, as shown in
rithm to specify the transverse properties.147 This Figure 9. The fiber-level model revealed complicated
approach indicated fiber transverse properties, fiber– dynamic fiber kinematics and mechanisms with a dis-
fiber and fiber–projectile friction play a major role in persive flexural wave and a spreading wave. The flex-
the penetration resistance and hence ballistic perfor- ural wave results in bending of fibers and significantly
mance of the fabric. The most recent studies use inde- higher axial stresses under the impact and at the
pendently measured fiber properties to study these clamped end compared to a homogenized yarn-level
interactions. Nilakantan148 used solid elements to model. The curvatures due to bending in the fiber are
model all the 400 KM2 fibers within a yarn subjected significant enough to cause axial compressive kinking
to impact by a hemi-cylindrical projectile and observed and fibrillation under the impact. The fibers experience
flattening and spreading of fibers. At the impact velo- multi-axial stress states, non-uniform loading and pro-
cities considered (100–120 m/s), the models predicted a gressive failure, depending on their spatial location.
spreading wave that propagated along the length of the The multi-axial loading, flexural wave induced axial
yarn resulting in progressive spreading of the fibers. compression, fiber bounce and progressive loading are
The models also predicted the projectile residual velo- mechanisms that may explain experimentally observed
city being sensitive to the transverse and shear proper- lower breaking speed for yarns than in the classic ana-
ties. However, the FE discretization used for fibers by lytic solution. However, the model does not account for
Nilakantan148 is not adequate to capture significant the misalignment and statistical failure of fibers.
deformations in transverse compression at experimen- There has been limited investigation of fiber length
tal yarn breaking speeds. None of these numerical scale modeling in the literature due to the computa-
approaches have been experimentally validated through tional intensity of the models. However, with advances
direct correlation with transverse compressive loading in computing power more research is being focused on
of a single fiber. the lower length scales to understand the fundamental
Sockalingam et al.180,181 developed fiber length scale deformation and failure mechanisms. Most recently,
FE models to understand the complicated yarn trans- Recchia et al.183 developed a filament model of 600
verse impact experimental results. To gain insight into denier KM2 yarn incorporating different levels of
the fiber-level mechanisms, they systematically studied twists in the fibers using beam elements. Up to a certain
the QS and dynamic compressive response of a single level of twist (3 twists per inch (TPI)), the yarn tensile
fiber and yarn by explicitly modeling all 400 fibers (of strength is found to increase due to outer filaments
diameter 12 mm) in a 600 denier KM2 yarn with the being stressed more than the inner ones, whereas a
mesh discretization determined from the single fiber high level of twist (10 TPI) reduces the ultimate tensile
study. The study revealed a large extent of fiber spread- strength due to high local stress concentrations. Xia
ing due to the fiber–fiber contact interactions. The sig- and Nadler174 noted that misalignment of fibers affects
nificant role of fiber–fiber contact in spreading, the stiffness and failure of yarns. The fiber-level model-
frictional interactions and deformation of individual ing resolution allows studying the effect of diameter
fibers predicted by the model were consistent with and number of fibers, statistical strength distribution,
experimental observations. Comparison to a homoge- degree of twist and so on, for different loading scenar-
nized yarn model indicated the need for a nonlinear ios and hence the impact response.
material model to accurately capture the transverse Fiber-level modeling indicates progressive failure of
compression response. fibers, significant transverse compression and curva-
A study by Sockalingam et al.182 on single fiber ture-induced compressive kinking leading to fibrillation
transverse impact using a 3D FE model predicted a failure. There is also experimental evidence on
dispersive flexural wave mode due to the finite long- fibrillation of fibers subjected to transverse compres-
itudinal shear modulus of the KM2 fiber. The FE sion, compressive kinking and axial tension. To better

Figure 9. Contours of short time scale compressive strains at 500 m/s impact.7
understand fibrillation and fibril failure, a lower length interactions in the system. For computational efficiency
scale modeling approach is necessary. in both the length- and time-scale, united atom and
coarse grain approaches are adopted in the classical
molecular simulations.186 Since experimentation at the
Inelastic material response
fibril length scale is very challenging, molecular simula-
Most of the modeling work discussed in this section so tion is an excellent alternative tool to characterize
far assumes an elastic behavior for the material at dif- the materials and get insight into the behavior of the
ferent length scales (fabric, yarn and fiber). However, materials under impact conditions.
the fiber transverse compression QS experimental Molecular simulations on Kevlar reported in the
results clearly indicate a very small elastic limit literature mainly studied the elastic and strength prop-
(1.25% strain for KM2 fiber) and an inelastic behavior erties under tensile and compressive loads rather than
of the material with negligible strain recovery upon dedicatedly focusing on the multi-axial loading and
unloading.184 The transverse compressive response of ballistic performance. Recently, Grujicic et al.187–189
the fibers at high rates of deformation is not well under- studied the properties of PPTA using classical MD
stood. The role of this inelastic transverse compression simulations. They used the non-reactive force field
behavior and the associated energy absorption during COMPASS (Condensed-phase Optimized Molecular
impact is also not well understood. The inelastic Potentials for Atomistic Simulation Studies)190 to
compressive behavior may result in dissipative effects represent the atomic interactions. In one study,187
and may influence the evolution of axial strains in the they investigated the effects of microstructural and
material and hence its failure. Therefore, to accurately topological defects such as chain ends, inorganic-
predict energy absorption and failure, future work must solvent impurities, chain misalignments and sheet
consider the inelastic material constitutive behavior at stacking fault on the strength, ductility and stiffness
high rates of loading. of PPTA fibril. The presence of these defects decreases
the mechanical properties of PPTA fibrils. However,
Fibril length scale molecular dynamics modeling some specific properties are found to be particularly
sensitive to the presence of certain defects. For exam-
of ballistic fibrils ple, longitudinal tensile properties are found to be the
Recently there has been an increasing interest in char- most sensitive to the presence of chain ends, in-sheet
acterizing the material properties under dynamic load- transverse properties to the presence of chain misalign-
ing using molecular simulations due to the advances in ments and cross-sheet transverse properties to the pre-
computational power. Simulations with the ab initio sence of sheet stacking faults. From the fibril model
principle are more accurate but computationally very properties, they proposed a fiber and yarn-level mate-
expensive, even with state-of-art computational cap- rial model considering fibril-level crystallographic and
ability. Classical molecular dynamics (MD) simulations morphological defects.188
that solve Newton’s equations of motion185 are less In another study,189 they investigated the effects of
expensive. The empirical force fields used in the classi- prior-axial compression and torsion on the longitudinal
cal MD simulations are derived based on experimental tensile behavior of PPTA fibrils and the role of various
and ab initio calculations. The accuracy of the simula- micro-structural/topological defects affecting this beha-
tion depends on the choice of force field for atomic vior. In the presence of defects, prior-axial compression

Figure 10. (a) Scanning electron microscopy picture of kink bands on a single KM2 fiber.192 (b) Kinking predicted by the molecular
dynamics model.192
and torsion degrades the longitudinal tensile strength. Thomas et al.194 and Lomicka et al.195 used MD simu-
However, prior-axial compression and torsion do not lations to predict the properties of Kevlar fibrils. The
degrade the longitudinal tensile strength of a perfect MD models predicted higher elastic modulus in tension
PPTA fibril, since there is no permanent change in than the experimental values, which they attributed to
the PPTA micro-structure after removal of these pre- the lack of defects in the model. MD models not only
loads. Determining strength properties using the non- provide insight into how molecular and fibril-level
reactive force fields is questionable, since these force properties may affect the fiber properties, but also
fields do not capture the bond breaking phenomena. assist in designing materials with improved properties.
Therefore, analyzing the structure and mechanical For instance, Thomas et al.194 used fragments of CNTs
properties of Kevlar with a reactive force field is neces- along with Kevlar fibrils in their MD simulations and
sary to further our understanding on the deformation predicted a 10% higher tensile failure stress compared
and failure mechanisms. to a Kevlar-only system. As discussed earlier, inter-
There is some evidence in the literature indicating fibrillar shear is an important mechanism during
the presence of compressive kink bands in polymer nanoindentation and in energy absorption associated
composites under ballistic impact,191 indicating the with particle-impregnated fabrics. However, modeling
multi-axial loading. The curvature predicted by the efforts to investigate and better understand this
single fiber transverse impact model182 is physically mechanism are scant. Recchia et al.196 developed a
induced by wrapping single KM2 fibers around a nanoscale FE model to account for the fibrillar micro-
Boron fiber with a curvature of 18 mm. A character- structure of Kevlar fibers to predict the axial tensile
istic kink band density of 70 per mm and an approx- response of the fiber. They used a cohesive zone
imate angle of 60o are observed. Figure 10(a) shows a model to mimic the inter-fibrillar interaction forces.
SEM picture of the kink bands.192 The MD model of Based on a parametric study, they showed fibril
Kevlar subjected to axial compression, as shown in response is strongly dependent on the cohesive interac-
Figure 10(b), indicates that kinking initiates through tions. Alternatively, MD simulations may be used to
the lateral deflection of the Kevlar chains in a direction rigorously determine the inter-fibrillar response, which
normal to the plane containing the hydrogen bonds. can then be used in fibril-level FE models.197
While MD models predict kinking at high rates of load- Unlike Kevlar, there are numerous molecular
ing, further experimental studies are needed to under- simulation studies on PE.198–203 Some of them used
stand the kink band formation at high rates. However, united-atom and coarse grain approaches to reduce
the ratio of the characteristic distance between the the computational cost. Most of the simulations
kinks to the model thickness is found to be in good focused on the characterization of microstructures,
agreement with the experiment. elastic and strength properties, thermal properties and
Fibrillation of fibers under impact is a predominant deformation of the bulk amorphous polymer rather
failure mechanism that is not well understood. than focusing on structural ballistic PE fibers, which
Recently, Mayo and Wetzel193 studied the cut resis- have a hierarchical structure consisting of a crystalline
tance and failure of single fibers and reported signifi- and an amorphous phase. There is limited work on
cant deformation, delamination and progressive fibril characterizing the behavior of the PE fibers under
failure of Kevlar fibers, depending on the blade angle. dynamic impact and the associated deformation and

damage mechanisms. Mattsson et al.204 conducted quan- understood. This understanding will provide
tum and classical MD simulations of PE under shock opportunities to design better materials to achieve
compression to study the accuracy of the different optimum impact performance. There are oppor-
empirical force fields. Recently, Chantawansri et al.205 tunities to better understand the HSR transverse
also conducted shock Hugoniot calculation of polymer compression response and the role of transverse
using the quantum and the MD methods. They pro- properties of the fiber during impact.
duced the principle shock Hugoniot curve and compared (iv) Fibrillation is an important failure mode for the
this with the available experimental results. high-performance Kevlar and PE ballistic fibers,
which is poorly understood. Molecular simulation
of fibrils along with nano-scale experimentation
Conclusions may help to provide insights into the failure process.
The deformation and failure mechanisms during the (v) Characterization of fibers subjected to multi-axial
ballistic impact of a flexible woven textile fabric are loading at different strain rates can improve the
complicated, involving multiple length and time scales predictive capability of the models. This necessi-
affected by anisotropic material behavior, projectile tates experiments with multi-axial loading at dif-
geometry–fabric interactions, impact velocity and ferent strain rates to accurately predict fiber
boundary conditions. This paper reviewed the recent failure.
research in modeling and experiments of important (vi) Most of the modeling approaches do not consider
mechanisms and deformation modes of these materials the experimentally observed inelastic transverse
(particularly Kevlar and PE) at different length scales compression behavior of the fiber. Future model-
and strain rates. Although fiber-specific toughness and ing efforts at different length scales must consider
longitudinal wave speed are important properties this to accurately predict the energy absorption
affecting the ballistic performance, this review indicates and failure.
other properties and mechanisms, such as fiber trans- (vii) Most of the modeling approaches do not incor-
verse compression, longitudinal shear, multi-axial load- porate the statistical nature of failure. To improve
ing response, flexural wave, different longitudinal the predictive capability of the models it is impor-
tensile and compressive moduli, fibrillation, statistical tant to incorporate the statistical nature of mate-
fiber failure and fiber–fiber interactions, play a major rial failure.
role in the energy absorption during impact. Important (viii) In general, element removal (erosion strain) is
information governing energy absorption is lost while used in ballistic impact FE models to model mate-
homogenizing to higher length scales. The transverse rial failure. Erosion strain is used to remove the
impact on fabrics requires a multi-scale modeling heavily distorted elements that would otherwise
approach from fibrils to multi-layers. In general, there significantly reduce the computational time step.
is limited research reported in the literature on the However, erosion strain is known to be mesh
experiments and modeling of these materials at lower dependent and can alter the results. To accurately
length scales (fiber-level and lower). Some key points predict failure, mesh objective/regularization
based on the review are as follows. techniques and nonlocal approaches such as peri-
dynamics206,207 may be considered.
(i) Some of the complex mechanisms observed in (ix) Modeling at the lower length scales helps one to
fabric length scale impact experiments are as fol- understand the responsible load transfer and
lows: transverse shear and compression failure of energy absorption mechanisms. Opportunities
the initial layers; pyramidal wave propagation; exist to understand the effect of the different
yarn pull out; inter-yarn friction; progressive fiber longitudinal tensile and compressive
fiber failure; slipping through of the sharp nosed moduli, fiber misalignment, twist, undulations,
projectiles; spreading and flattening of fibers and statistical strength distribution, strain rate-
fibrillation. Lower length scale experiments are dependent multi-axial failure, transverse perma-
necessary to better understand these mechanisms nent deformation and so on. It is also imperative
and the associated energy absorption. to account for these mechanisms in order to accu-
(ii) Fiber–fiber interactions and progressive fiber rately predict material failure.
failure during the transverse impact of a yarn (x) Morphology-dependent properties and failure
are not well understood, as most of the impact during impact are observed. To develop new pro-
experiments reported in the literature focused on tective materials, it would be useful to conduct
transverse wave speed and angle. MD simulations at high strain rates to understand
(iii) The load transfer and energy absorption mechan- the morphology change in the fiber microstructure
isms at lower length scales are not well and associated energy absorbing mechanisms.

Acknowledgments 14. Seto T, Hara T and Tanaka K. Phase transformation and

The views and conclusions contained in this document are deformation processes in oriented polyethylene. Jpn J
those of the authors and should not be interpreted as repre- Appl Phys 1968; 7: 31.
senting the official policies, either expressed or implied, of the 15. Bassett D, Block S and Piermarini G. A high-pressure
Army Research Laboratory or the U.S. Government. The phase of polyethylene and chain-extended growth.
U.S. Government is authorized to reproduce and distribute J Appl Phys 1974; 45: 4146–4150.
reprints for Government purposes notwithstanding any copy- 16. Prevorsek DC, Kwon YD and Chin HB. Analysis of the
right notation herein. temperature rise in the projectile and extended chain
polyethylene fiber composite armor during ballistic
impact and penetration. Polym Eng Sci 1994; 34: 141–52.
Funding 17. Cheeseman BA and Bogetti TA. Ballistic impact into
Research was sponsored by the Army Research Laboratory fabric and compliant composite laminates. Compos
and was accomplished under the Cooperative Agreement Struct 2003; 61: 161–173.
Number W911NF-12-2-0022. 18. Tabiei A and Nilakantan G. Ballistic impact of dry
woven fabric composites: a review. Appl Mech Rev
2008; 61: 010801.
References
19. David N, Gao X and Zheng J. Ballistic resistant body
1. Kim J, McDonough WG, Blair W, et al. The modified- armor: contemporary and prospective materials and
single fiber test: a methodology for monitoring ballistic related protection mechanisms. Appl Mech Rev 2009;
performance. J Appl Polym Sci 2008; 108: 876–886. 62: 050802.
2. Krishnan K, Sockalingam S, Bansal S, et al. Numerical 20. Haque BZG and Gillespie JW. A new penetration equa-
simulation of ceramic composite armor subjected to bal- tion for ballistic limit analysis. J Thermoplast Compos
listic impact. Composites Part B 2010; 41: 583–593. Mater 2013; 28: 0892705713495430.
3. Sharda J, Deenadayalu C, Mobasher B, et al. Modeling 21. Smith JC, McCrackin FL and Schiefer HF. Stress-strain
of multilayer composite fabrics for gas turbine engine relationships in yarns subjected to rapid impact loading:
containment systems. J Aerosp Eng 2006; 19: 38–45. part V: wave propagation in long textile yarns impacted
4. Chocron S, Figueroa E, King N, et al. Modeling and transversely. Text Res J 1958; 28: 288–302.
validation of full fabric targets under ballistic impact. 22. Smith JC, Blandford JM and Schiefer HF. Stress-strain
Composites Sci Technol 2010; 70: 2012–2022. relationships in yarns subjected to rapid impact loading
5. Nilakantan G, Merrill RL, Keefe M, et al. Experimental part VI: velocities of strain waves resulting from impact.
investigation of the role of frictional yarn pull-out and Text Res J 1960; 30: 752–760.
windowing on the probabilistic impact response of 23. Mamivand M and Liaghat GH. A model for ballistic
Kevlar fabrics. Composites Part B 2015; 68: 215–229. impact on multi-layer fabric targets. Int J Impact Eng
6. Nilakantan G, Keefe M, Gillespie JW Jr, et al. An experi- 2010; 37: 806–812.
mental and numerical study of the impact response (V50) 24. Cole J, Dougherty C and Huth J. Constant strain waves
of flexible plain weave fabrics: Accounting for statistical in strings. J Appl Mech 1953; 20: 53–SA-4.
distributions of yarn strength. In: the 1st joint American- 25. Wang L. Foundations of stress waves. Oxford: Elsevier,
Canadian international conference on composites and the 2011.
24th annual ASC technical conference, University of 26. Song B, Park H, Lu W, et al. Transverse impact response
Delaware, Newark, USA, September 2009. of a linear elastic ballistic fiber yarn. J Appl Mech 2011;
7. Sockalingam S, Gillespie JW Jr and Keefe M. Fiber-level 78: 051023-1–051023-9.
tow modeling of Kevlar KM2 subjected to high velocity 27. Walker JD and Chocron S. Why impacted yarns break at
impact. In: Albers E, Beckwith R and Cates SW (eds) lower speed than classical theory predicts. J Appl Mech
SAMPE. USA: Seattle, WA, 2014, pp.2–5. 2011; 78: 051021.
8. Northolt M and Van Aartsen J. On the crystal and mole- 28. Chocron S, Kirchdoerfer T, King N, et al. Modeling of
cular structure of poly-(p-phenylene terephthalamide). fabric impact with high speed imaging and nickel-chro-
J Polym Sci Polym Lett 1973; 11: 333–337. mium wires validation. J Appl Mech 2011; 78: 051007.
9. Northolt M. X-ray diffraction study of poly (p-phenylene 29. Leigh Phoenix S and Porwal PK. A new membrane
terephthalamide) fibres. Eur Polym J 1974; 10: 799–804. model for the ballistic impact response and V50 perfor-
10. Tashiro K, Kobayashi M and Tadokoro H. Elastic moduli mance of multi-ply fibrous systems. Int J Solids Struct
and molecular structures of several crystalline polymers, 2003; 40: 6723–6765.
including aromatic polyamides. Macromolecules 1977; 10: 30. Porwal PK and Phoenix SL. Modeling system effects in
413–420. ballistic impact into multi-layered fibrous materials for
11. Yang HH. Kevlar aramid fiber. New York: Wiley, 1993. soft body armor. Int J Fract 2005; 135: 217–249.
12. Bunn CW. The crystal structure of long-chain normal 31. Cunniff PM. Dimensionless parameters for optimization
paraffin hydrocarbons. The ‘‘shape’’ of the< CH2 of textile-based body armor systems. In: proceedings of
group. Trans Faraday Soc 1939; 35: 482–491. the 18th international symposium on ballistics, San
13. Bank M and Krimm S. Lattice-frequency studies of crys- Antonio, Texas, 1999, pp.1303–1310.
talline and fold structure in polyethylene. J Appl Phys 32. Afshari M, Chen P and Kotek R. Relationship between
1968; 39: 4951–4958. tensile properties and ballistic performance of poly

(ethylene naphthalate) woven and nonwoven fabrics. continuum-level simulations. J Mech Phys Solids 2010;
J Appl Polym Sci 2012; 125: 2271–2280. 58: 1995–2021.
33. Karthikeyan K, Russell B, Fleck N, et al. The effect of 51. Shimek ME and Fahrenthold EP. Effects of weave type
shear strength on the ballistic response of laminated com- on ballistic performance of fabrics. AIAA J 2012; 50:
posite plates. Eur J Mech A Solids 2013; 42: 35–53. 2558–2565.
34. Standard N. I. J. 0101.06:2008. Ballistic resistance of 52. Ke dzierski P, Poplawski A, Gieleta R, et al. Experimental
body armor. and numerical investigation of fabric impact behavior.
35. Dong Z and Sun CT. Testing and modeling of yarn pull- Composites Part B 2014; 69: 452–459.
out in plain woven Kevlar fabrics. Composites Part A 53. Lim JS, Lee BH, Lee CB, et al. Effect of the weaving
2009; 40: 1863–1869. density of aramid fabrics on their resistance to ballistic
36. Cunniff PM. An analysis of the system effects in woven impacts. Engineering 2012; 4: 944.
fabrics under ballistic impact. Text Res J 1992; 62: 54. Lim JS. Ballistic behavior of HeracronÕ -based compo-
495–509. sites: effect of the number multifilaments on high-speed
37. Chocron S, Ranjan Samant K, Nicholls AE, et al. projectiles. Model Numer Simulat Mater Sci 2013; 3: 84.
Measurement of strain in fabrics under ballistic impact 55. Chen X, Zhou Y and Wells G. Numerical and experimen-
using embedded nichrome wires. Part I: technique. Int J tal investigations into ballistic performance of hybrid
Impact Eng 2009; 36: 1296–1302. fabric panels. Composites Part B 2014; 58: 35–42.
38. Chocron S, Anderson CE Jr, Samant KR, et al. 56. Dong Z, Manimala JM and Sun C. Mechanical behavior
Measurement of strain in fabrics under ballistic impact of silica nanoparticle-impregnated Kevlar fabrics. J Mech
using embedded nichrome wires, part II: results and ana- Mater Struct 2010; 5: 529–548.
lysis. Int J Impact Eng 2010; 37: 69–81. 57. Shim VPW, Lim CT and Foo KJ. Dynamic mechanical
39. Chocron S, King N, Bigger R, et al. Impacts and waves in properties of fabric armour. Int J Impact Eng 2001; 25:
DyneemaÕ HB80 strips and laminates. J Appl Mech 2013; 1–15.
80: 031806. 58. Zhu D, Mobasher B and Rajan SD. Dynamic tensile
40. Tan VBC, Lim CT and Cheong CH. Perforation of high- testing of Kevlar 49 fabrics. J Mater Civ Eng 2010; 23:
strength fabric by projectiles of different geometry. Int J 230–239.
Impact Eng 2003; 28: 207–222. 59. Phoenix SL and Skelton J. Transverse compressive
41. Talebi H, Wong SV and Hamouda AMS. Finite element moduli and yield behavior of some orthotropic, high-
evaluation of projectile nose angle effects in ballistic per- modulus filaments. Text Res J 1974; 44: 934–940.
foration of high strength fabric. Compos Struct 2009; 87: 60. Raftenburg MN, Scheidler NJ and Moy P. Transverse
314–320. compression response of a multi-ply Kevlar vest.
42. Montgomery T, Grady P and Tomasino C. The effects of Technical Report ARL-TR-3343. Army Research
projectile geometry on the performance of ballistic fab- Laboratory. Aberdeen Proving Ground, MD, 2004.
rics. Text Res J 1982; 52: 442–450. 61. Manimala JM and Sun C. Investigation of failure in
43. Lim C, Tan V and Cheong C. Perforation of high- Kevlar fabric under transverse indentation using a homo-
strength double-ply fabric system by varying shaped pro- genized continuum constitutive model. Text Res J 2014;
jectiles. Int J Impact Eng 2002; 27: 577–591. 84: 388–398.
44. Prosser RA, Cohen SH and Segars RA. Heat as a factor 62. Lomov SV, Willems A, Verpoest I, et al. Picture frame
in the penetration of cloth ballistic panels by 0.22 caliber test of woven composite reinforcements with a full-field
projectiles. Text Res J 2000; 70: 709–722. strain registration. Text Res J 2006; 76: 243–252.
45. Tan V, Lim C and Cheong C. Perforation of high- 63. Zhu D, Mobasher B, Vaidya A, et al. Mechanical beha-
strength fabric by projectiles of different geometry. Int viors of Kevlar 49 fabric subjected to uniaxial, biaxial
J Impact Eng 2003; 28: 207–222. tension and in-plane large shear deformation.
46. Chen X, Zhu F and Wells G. An analytical model for Composites Sci Technol 2013; 74: 121–130.
ballistic impact on textile based body armour. Composites 64. Badel P, Vidal-Sallé E and Boisse P. Computational
Part B 2013; 45: 1508–1514. determination of in-plane shear mechanical behaviour
47. Shim V, Guo Y and Tan V. Response of woven and of textile composite reinforcements. Comput Mater Sci
laminated high-strength fabric to oblique impact. Int J 2007; 40: 439–448.
Impact Eng 2012; 48: 87–97. 65. Ha-Minh C, Imad A, Kanit T, et al. Numerical analysis
48. Shockey DA, Erlich DC and Simons JW. Lightweight of a ballistic impact on textile fabric. Int J Mech Sci 2013;
fragment barriers for commercial aircraft. In: proceedings 69: 32–39.
of the 18th international symposium on ballistics, San 66. Song B and Lu W. Effect of twist on transverse impact
Antonio, TX, 15–19 November 1999, pp.1192–1199. response of ballistic fiber yarns. Int J Impact Eng 2015;
49. Zeng XS, Shim VPW and Tan VBC. Influence of 85: 1–4.
boundary conditions on the ballistic performance of 67. Bazhenov S, Dukhovskii I, Kovalev P, et al. The fracture
high-strength fabric targets. Int J Impact Eng 2005; 32: of SVM aramide fibers upon a high-velocity transverse
631–642. impact. Polym Sci Ser A Chem Phys 2001; 43: 61–71.
50. Parsons EM, Weerasooriya T, Sarva S, et al. Impact of 68. Carr D. Failure mechanisms of yarns subjected to ballis-
woven fabric: experiments and mesostructure-based tic impact. J Mater Sci Lett 1999; 18: 585–588.

69. Field JE and Sun Q. High-speed photographic study of 85. Rao Y and Farris RJ. A modeling and experimental
impact on fibers and woven fabrics. In: 19th interna- study of the influence of twist on the mechanical proper-
tional congress on high-speed photography and photonics, ties of high-performance fiber yarns. J Appl Polym Sci
International Society for Optics and Photonics, 1991, 2000; 77: 1938–1949.
pp.703–712. 86. Chu Y, Chen X, Wang Q, et al. An investigation on
70. Heru Utomo B and Broos J. Dynamic material behavior sol-gel treatment to aramid yarn to increase inter-yarn
determination using single fiber impact. In: 25th confer- friction. Appl Surf Sci 2014; 320: 710–717.
ence and exposition on structural dynamics (IMAC- 87. Cheng M, Chen W and Weerasooriya T. Experimental
XXV), Orlando, Florida, USA, 19–22 February 2007. investigation of the transverse mechanical properties of
71. Mulkern TJ and Raftenberg MN. Kevlar KM2 yarn and a single KevlarÕ KM2 fiber. Int J Solids Struct 2004; 41:
fabric strength under quasi-static tension Technical 6215–6232.
Report ARL-TR-2865. Army Research Laboratory. 88. Cheng M, Chen W and Weerasooriya T. Mechanical
Aberdeen Proving Ground, MD, 2002. properties of KevlarÕ KM2 single fiber. J Eng Mater
72. Nilakantan G, Obaid AA, Keefe M, et al. Experimental Technol 2005; 127: 197–203.
evaluation and statistical characterization of the 89. Leal AA, Deitzel JM and Gillespie JW. Compressive
strength and strain energy density distribution of strength analysis for high performance fibers with
Kevlar KM2 yarns: exploring length-scale and weaving different modulus in tension and compression.
effects. J Compos Mater 2011; 45: 1749–1769. J Compos Mater 2009; 43: 661–674.
73. Tan V, Zeng X and Shim V. Characterization and con- 90. Beckwith SW. Composites reinforcement fibers: II-the
stitutive modeling of aramid fibers at high strain rates. aramid and polyethylene families. SAMPE J 2009; 45:
Int J Impact Eng 2008; 35: 1303–1313. 42–43.
74. Zhu D, Mobasher B, Erni J, et al. Strain rate and gage 91. Kim J, Heckert N, McDonough W, et al. Statistical char-
length effects on tensile behavior of Kevlar 49 single acterization of single PPTA fiber tensile properties from
yarn. Composites Part A 2012; 43: 2021–2029. high strain rate tests. In: Song B, Casem D and Kimberley
75. Dooraki BF, Nemes J and Bolduc M. Study of para- J (eds) Dynamic behavior of materials, volume 1. 2015,
meters affecting the strength of yarns. J Phys IV 2006; pp.5–8. Bethel, CT: Springer International Publishing.
134: 1183–1188. 92. Lim J, Chen WW and Zheng JQ. Dynamic small strain
76. Languerand D, Zhang H, Murthy N, et al. Inelastic measurements of KevlarÕ 129 single fibers with a minia-
behavior and fracture of high modulus polymeric fiber turized tension Kolsky bar. Polym Test 2010; 29: 701–705.
bundles at high strain-rates. Mater Sci Eng A 2009; 500: 93. Lim J, Zheng J, Masters K, et al. Mechanical behavior of
216–224. A265 single fibers. J Mater Sci 2010; 45: 652–661.
77. Kirkwood KM, Kirkwood JE, Lee YS, et al. Yarn pull- 94. Hudspeth M, Claus B, Parab N, et al. In situ visual obser-
out as a mechanism for dissipating ballistic impact vation of fracture processes in several high-performance
energy in KevlarÕ KM-2 fabric part I: quasi-static char- fibers. J Dyn Behav Mater 2015; 1: 55–64.
acterization of yarn pull-out. Text Res J 2004; 74: 95. Kim JH, Alan Heckert N, Leigh SD, et al. Statistical ana-
920–928. lysis of PPTA fiber strengths measured under high strain
78. Kirkwood JE, Kirkwood KM, Lee YS, et al. Yarn pull- rate condition. Composites Sci Technol 2014; 98: 93–99.
out as a mechanism for dissipating ballistic impact 96. Lim J, Zheng JQ, Masters K, et al. Effects of gage
energy in KevlarÕ KM-2 fabric part II: predicting bal- length, loading rates, and damage on the strength of
listic performance. Text Res J 2004; 74: 939–948. PPTA fibers. Int J Impact Eng 2011; 38: 219–227.
79. Zhu D, Soranakom C, Mobasher B, et al. Experimental 97. Schwartz P, Netravali A and Sembach S. Effects of strain
study and modeling of single yarn pull-out behavior of rate and gauge length on the failure of ultra-high strength
KevlarÕ 49 fabric. Composites Part A 2011; 42: 868–879. polyethylene fibers. Text Res J 1986; 56: 502–508.
80. Bilisik K. Properties of yarn pull-out in para-aramid 98. Kawabata S. Measurement of the transverse mechanical
fabric structure and analysis by statistical model. properties of high-performance fibres. J Text Inst 1990;
Composites Part A 2011; 42: 1930–1942. 81: 432–447.
81. Gawandi A, Thostenson ET and Gilllespie JW Jr. Tow 99. Singletary J, Davis H, Ramasubramanian MK, et al.
pullout behavior of polymer-coated Kevlar fabric. The transverse compression of PPTA fibers part I
J Mater Sci 2011; 46: 77–89. single fiber transverse compression testing. J Mater Sci
82. Nilakantan G and Gillespie JW. Yarn pull-out behavior 2000; 35: 573–581.
of plain woven Kevlar fabrics: effect of yarn sizing, pull- 100. Singletary J, Davis H, Song Y, et al. The transverse
out rate, and fabric pre-tension. Compos Struct 2013; compression of PPTA fibers Part II Fiber transverse
101: 215–224. structure. J Mater Sci 2000; 35: 583–592.
83. Guo Z, Hong J, Zheng J, et al. Loading rate effects on 101. Ruan F and Bao L. Mechanical enhancement of
dynamic out-of-plane yarn pull-out. Text Res J 2014; UHMWPE fibers by coating with carbon nanoparticles.
84: 1708–1719. Fibers Polym 2014; 15: 723–728.
84. Guo Z, Hong J, Zheng J, et al. Out-of-plane effects on 102. Guo Z, Casem D, Hudspeth M, et al. Transverse com-
dynamic pull-out of p-phenylene terephthalamide yarns. pression of two high-performance ballistic fibers. Text
Text Res J 2014; 85: 140–149. Res J 2015; 1–10.

103. Wollbrett-Blitz J, Joannès S, Bruant R, et al. Multiaxial scratch characteristics of Kevlar fibers. J Mater Sci
mechanical behavior of aramid fibers and identification 2013; 48: 292–1302.
of skin/core structure from single fiber transverse com- 121. Li S, McGhie A and Tang S. Comparative study of the
pression testing. J Polym Sci Part B Polym Phys 2015; internal structures of Kevlar and spider silk by atomic
54: 374–384. force microscopy. J Vacuum Sci Technol A 1994; 12:
104. Leal AA, Deitzel JM and Gillespie JW. Assessment of 1891–1894.
compressive properties of high performance organic 122. McAllister QP. The energy dissipative mechanisms of the
fibers. Composites Sci Technol 2007; 67: 2786–2794. particle-fiber interface in a textile composite. PhD
105. M’Ewen E. XLI. Stresses in elastic cylinders in contact Dissertation, University of Delaware, 2013.
along a generatrix (including the effect of tangential fric- 123. Cole DP and Strawhecker KE. An improved instrumen-
tion). Philos Mag Ser 7 1949; 40: 454–459. ted indentation technique for single microfibers. J Mater
106. Hadley DW, Ward IM and Ward J. The transverse com- Res 2014; 29: 1104–1112.
pression of anisotropic fibre monofilaments. Proc R Soc 124. McAllister QP, Gillespie JW and VanLandingham MR.
London Ser A 1965; 285: 275–286. The energy dissipative mechanisms of particle-fiber
107. Kotani T, Sweeney J and Ward IM. The measurement interactions in a textile composite. J Compos Mater
of transverse mechanical properties of polymer fibres. 2013; 48: 3553–3567.
J Mater Sci 1994; 29: 5551–5558. 125. McAllister QP, Gillespie JW Jr and VanLandingham
108. Pinnock PR, Ward IM and Wolfe JM. The compression MR. The sub-micron scale energy dissipative deforma-
of anisotropic fibre monofilaments. II. Proc R Soc tion mechanisms of Kevlar fibrils. J Mater Sci 2013; 48:
London Ser A 1966; 291: 267–278. 6245–6261.
109. Timoshenko S and Goodier JN. Theory of elasticity. 126. Decker M, Halbach C, Nam C, et al. Stab resistance of
New York: McGraw-Hill, 1951. shear thickening fluid (STF)-treated fabrics. Composites
110. Obaid AA, Yarlagadda S and Gillespie JW Jr. Sci Technol 2007; 67: 565–578.
Combined effects of Kink bands and hygrothermal con- 127. Tan V, Tay T and Teo W. Strengthening fabric armour
ditioning on tensile strength of polyarylate LCP and with silica colloidal suspensions. Int J Solids Struct 2005;
aramid fibers. J Compos Mater 2015; 50: 339–350. 42: 1561–1576.
111. Kollár LP and Springer GS. Mechanics of composite 128. LaBarre E, Calderon-Colon X, Morris M, et al. Effect
of a carbon nanotube coating on friction and impact
structures. Cambridge: Cambridge University Press,
performance of Kevlar. J Mater Sci 2015; 50:
2003.
5431–5442.
112. Tsai C and Daniel I. Determination of shear modulus of
129. McDaniel PB, Deitzel JM and Gillespie JW. Structural
single fibers. Exp Mech 1999; 39: 284–286.
hierarchy and surface morphology of highly drawn ultra
113. Sanborn BD and Weerasooriya T. Effect of strain rates
high molecular weight polyethylene fibers studied by
and pre-twist on tensile strength of Kevlar KM2 single
atomic force microscopy and wide angle X-ray diffrac-
fiber. Technical Report ARL-TR-6403. Army Research
tion. Polymer 2015; 69: 148–158.
Laboratory. Aberdeen Proving Ground, MD, 2013.
130. Pennings A. Bundle-like nucleation and longitudinal
114. Sanborn B and Weerasooriya T. Quantifying damage at
growth of fibrillar polymer crystals from flowing solu-
multiple loading rates to Kevlar KM2 fibers due to
tions. J Polym Sci Polym Symposia 1977; 59: 55–86.
weaving, finishing, and pre-twist. Int J Impact Eng 131. Strawhecker KE and Cole DP. Morphological and local
2014; 71: 50–59. mechanical surface characterization of ballistic fibers via
115. Hudspeth M, Nie X and Chen W. Dynamic failure of AFM. J Appl Polym Sci 2014; 131: 40880 (1–12).
Dyneema SK76 single fibers under biaxial shear/tension. 132. Marissen R. Design with ultra strong polyethylene
Polymer 2012; 53: 5568–5574. fibers. Mater Sci Appl 2011; 2: 319.
116. Hudspeth M, Li D, Spatola J, et al. The effects of off- 133. Ivanov I and Tabiei A. Loosely woven fabric model
axis transverse deflection loading on the failure strain of with viscoelastic crimped fibres for ballistic impact simu-
various high-performance fibers. Text Res J. Epub lations. Int J Numer Methods Eng 2004; 61: 1565–1583.
ahead of print 2015. DOI: 10.1177/0040517515588262. 134. King MJ, Jearanaisilawong P and Socrate S. A
117. Hudspeth M, Chen W and Zheng J. Why the Smith continuum constitutive model for the mechanical beha-
theory over-predicts instant rupture velocities during vior of woven fabrics. Int J Solids Struct 2005; 42:
fiber transverse impact. Text Res J. Epub ahead of 3867–3896.
print 2015. DOI: 10.1177/0040517515586158. 135. Parsons EM, King MJ and Socrate S. Modeling yarn
118. Rao Y, Waddon AJ and Farris RJ. The evolution of slip in woven fabric at the continuum level: simulations
structure and properties in poly(p-phenylene terephtha- of ballistic impact. J Mech Phys Solids 2013; 61:
lamide) fibers. Polymer 2001; 42: 5925–5935. 265–292.
119. McAllister QP, Gillespie JW and VanLandingham MR. 136. Shahkarami A and Vaziri R. A continuum shell finite
Evaluation of the three-dimensional properties of element model for impact simulation of woven fabrics.
Kevlar across length scales. J Mater Res 2012; 27: Int J Impact Eng 2007; 34: 104–119.
1824–1837. 137. Lim CT, Shim VPW and Ng YH. Finite-element mod-
120. McAllister QP, Gillespie JW Jr and VanLandingham eling of the ballistic impact of fabric armor. Int J Impact
MR. The influence of surface microstructure on the Eng 2003; 28: 13–31.

138. Bansal S, Mobasher B, Rajan S, et al. Development of small projectile impact. Int J Impact Eng 1995; 16:
fabric constitutive behavior for use in modeling engine 585–605.
fan blade-out events. J Aerospace Eng 2009; 22: 249–259. 155. Zeng X, Tan V and Shim V. Modelling inter-yarn fric-
139. Stahlecker Z, Mobasher B, Rajan SD, et al. Development tion in woven fabric armour. Int J Numer Methods Eng
of reliable modeling methodologies for engine fan blade 2006; 66: 1309–1330.
out containment analysis. Part II: finite element analysis. 156. Novotny WR, Cepuš E, Shahkarami A, et al. Numerical
Int J Impact Eng 2009; 36: 447–459. investigation of the ballistic efficiency of multi-ply fabric
140. Nadler B, Papadopoulos P and Steigmann DJ. armours during the early stages of impact. Int J Impact
Multiscale constitutive modeling and numerical simula- Eng 2007; 34: 71–88.
tion of fabric material. Int J Solids Struct 2006; 43: 157. Blankenhorn G, Schweizerhof K and Finckh H.
206–221. Improved numerical investigations of a projectile
141. Boljen M and Hiermaier S. Continuum constitutive impact on a textile structure. In: fourth European
modeling of woven fabrics. Eur Phys J Spec Top 2012; LS-DYNA users conference, 2003.
206: 149–161. 158. Rao MP, Duan Y, Keefe M, et al. Modeling the effects
142. Grujicic M, Bell WC, Arakere G, et al. Development of of yarn material properties and friction on the ballistic
a meso-scale material model for ballistic fabric and its impact of a plain-weave fabric. Compos Struct 2009; 89:
use in flexible-armor protection systems. J Mater Eng 556–566.
Perform 2010; 19: 22–39. 159. Nilakantan G, Wetzel ED, Bogetti TA, et al. A determi-
143. Erol O, Wetzel ED and Keefe M. Simulation of a textile nistic finite element analysis of the effects of projectile
sleeve on a manikin arm undergoing elbow flexion: characteristics on the impact response of fully clamped
effect of arm-sleeve friction. J Text Inst 2014; 106: flexible woven fabrics. Compos Struct 2013; 95: 191–201.
1135–1146. 160. Nilakantan G and Gillespie JW Jr. Ballistic impact
144. Duan Y, Keefe M, Bogetti TA, et al. Modeling friction modeling of woven fabrics considering yarn strength,
effects on the ballistic impact behavior of a single-ply friction, projectile impact location, and fabric boundary
high-strength fabric. Int J Impact Eng 2005; 31: condition effects. Compos Struct 2012; 94: 3624–3634.
161. Duan Y, Keefe M, Bogetti T, et al. A numerical inves-
996–1012.
tigation of the influence of friction on energy absorption
145. Nilakantan G, Keefe M, Bogetti TA, et al. On the finite
by a high-strength fabric subjected to ballistic impact.
element analysis of woven fabric impact using multiscale
Int J Impact Eng 2006; 32: 1299–1312.
modeling techniques. Int J Solids Struct 2010; 47:
162. Duan Y, Keefe M, Bogetti TA, et al. Modeling the role
2300–2315.
of friction during ballistic impact of a high-strength
146. Wang Y, Miao Y, Swenson D, et al. Digital element
plain-weave fabric. Compos Struct 2005; 68: 331–337.
approach for simulating impact and penetration of tex-
163. Barauskas R and Abraitiene_ A. Computational analysis
tiles. Int J Impact Eng 2010; 37: 552–560.
of impact of a bullet against the multilayer fabrics in
147. Grujicic M, Hariharan A, Pandurangan B, et al. Fiber-
LS-DYNA. Int J Impact Eng 2007; 34: 1286–1305.
level modeling of dynamic strength of Kevlar KM2 bal-
164. Rao MP, Nilakantan G, Keefe M, et al. Global/local
listic fabric. J Mater Eng Perform 2012; 21: 1107–1119. modeling of ballistic impact onto woven fabrics.
148. Nilakantan G. Filament-level modeling of Kevlar KM2 J Compos Mater 2009; 43: 445–467.
yarns for ballistic impact studies. Compos Struct 2013; 165. Gogineni S, Gao X, David N, et al. Ballistic impact of
104: 1–13. Twaron CT709Õ plain weave fabrics. Mech Adv Mater
149. Hudspeth M, Agarwal A, Andrews B, et al. Degradation Struct 2012; 19: 441–452.
of yarns recovered from soft-armor targets subjected to 166. Gu B. Ballistic penetration of conically cylindrical steel
multiple ballistic impacts. Composites Part A 2014; 58: projectile into plain-woven fabric target–a finite element
98–106. simulation. J Compos Mater 2004; 38: 2049–2074.
150. Freeston WD and Claus WD. Strain-wave reflections 167. Sun D, Chen X, Lewis E, et al. Finite element simulation
during ballistic impact of fabric panels. Text Res J of projectile perforation through a ballistic fabric. Text
1973; 43: 348–351. Res J 2012; 83: 1489–1499.
151. Roylance D. Wave propagation in a viscoelastic fiber 168. Jia X, Huang Z, Zu X, et al. Effect of mesoscale and
subjected to transverse impact. J Appl Mech 1973; 40: multiscale modeling on the performance of Kevlar
143–148. woven fabric subjected to ballistic impact: a numerical
152. Cunniff P, Ting CTJ and Roylance D. Numerical char- study. Appl Compos Mater 2013; 20: 1195–1214.
acterization of the effects of transverse yarn interaction 169. Yen C, Scott B, Dehmer P, et al. A comparison between
on textile ballistic response, http://web.mit.edu/roy experiment and numerical simulation of fabric ballistic
lance/www/Ting_98.pdf (1998, accessed 3 November impact. In: proceedings of the 23rd international ballistic
2015). symposium, Tarragona, Spain, 2007.
153. Tan V and Ching T. Computational simulation of fabric 170. Jiang Y, Tabiei A and Simitses GJ. A novel microme-
armour subjected to ballistic impacts. Int J Impact Eng chanics-based approach to the derivation of constitutive
2006; 32: 1737–1751. equations for local/global analysis of a plain-weave
154. Shim VPW, Tan VBC and Tay TE. Modelling deforma- fabric composite. Composites Sci Technol 2000; 60:
tion and damage characteristics of woven fabric under 1825–1833.

171. Zohdi T and Powell D. Multiscale construction and 187. Grujicic M, Bell W, Glomski P, et al. Filament-level
large-scale simulation of structural fabric undergoing modeling of aramid-based high-performance structural
ballistic impact. Comput Methods Appl Mech Eng materials. J Mater Eng Perform 2011; 20: 1401–1413.
2006; 195: 94–109. 188. Grujicic M, Glomski P, Pandurangan B, et al.
172. Zohdi T. Microfibril-based estimates of the ballistic Multi-length scale computational derivation of
limit of multilayered fabric shielding. Int J Fract 2009; KevlarÕ yarn-level material model. J Mater Sci 2011;
158: 81–88. 46: 4787–4802.
173. Xia W, Adeeb SM and Nadler B. Ballistic performance 189. Grujicic M, Yavari R, Ramaswami S, et al. Molecular-
study of fabric armor based on numerical simulations level study of the effect of prior axial compression/torsion
with multiscale material model. Int J Numer Methods on the axial-tensile strength of PPTA fibers. J Mater Eng
Eng 2011; 87: 1007–1024. Perform 2013; 22: 3269–3287.
174. Xia W and Nadler B. Three-scale modeling and numer- 190. Sun H. COMPASS: an ab initio force-field optimized
ical simulations of fabric materials. Int J Eng Sci 2011; for condensed-phase applications overview with details
49: 229–239. on alkane and benzene compounds. J Phys Chem B
175. Nilakantan G, Keefe M, Wetzel ED, et al. 1998; 102: 7338–7364.
Computational modeling of the probabilistic impact 191. Greenhalgh E, Bloodworth V, Iannucci L, et al.
response of flexible fabrics. Compos Struct 2011; 93: Fractographic observations on DyneemaÕ composites
3163–3174. under ballistic impact. Composites Part A 2013; 44:
176. Nilakantan G, Keefe M, Wetzel ED, et al. Effect of 51–62.
statistical yarn tensile strength on the probabilistic 192. Chowdhury SC, Sockalingam S and Gillespie JW Jr.
impact response of woven fabrics. Composites Sci Molecular dynamics modeling of compression kinking
Technol 2012; 72: 320–329. in Kevlar. In: SAMPE conference proceedings,
177. Gasser A, Boisse P and Hanklar S. Mechanical beha- Baltimore, MD, USA, 18–21 May 2015.
viour of dry fabric reinforcements. 3D simulations 193. Mayo J and Wetzel E. Cut resistance and failure of high-
versus biaxial tests. Comput Mater Sci 2000; 17: 7–20. performance single fibers. Text Res J 2014; 84:
178. Durville D. Finite element simulation of the mechanical 1233–1246.
194. Thomas JA, Shanaman MT, Lomicka CL, et al.
behaviour of textile composites at the mesoscopic scale
Multiscale modeling of high-strength fibers and fabrics.
of individual fibers. In: textile composites and inflatable
In: SPIE defense, security, and sensing, International
structures II, Springer Netherlands, 2008.
Society for Optics and Photonics, 2012.
179. Yuyang M, Ao J, An Yu Y, et al. Energy loss due to
195. Lomicka C, Thomas J, LaBarre E, et al. Improving
transverse plastic deformation of fibers in textile fabric
ballistic fiber strength: insights from experiment and
impact processes. In: Liu D (ed.) Dynamic effects in com-
simulation. In: Song B, Casem D and Kimberley J
posites. 2012, p.103. Lancaster: DEStech Publications,
(eds) Dynamic behavior of materials, volume 1. 2014,
Inc.
p.187–193. Bethel: CT Springer International
180. Sockalingam S, Gillespie JW Jr and Keefe M. Modeling
Publishing.
the transverse compression response of Kevlar KM2. In: 196. Recchia SS, Pelegri A, Clawson JK, et al. A hierarchical
proceedings of the American Society for Composites model for Kevlar fiber failure. In: ASME 2013 interna-
(ASC) 28th technical conference, College Park, PA, tional mechanical engineering congress and exposition,
USA, 2013. American Society of Mechanical Engineers, 2013.
181. Sockalingam S, Gillespie JW Jr and Keefe M. On the 197. Chowdhury SC, Staniszewski J, Martz EM, et al.
transverse compression response of Kevlar KM2 using A computational approach for linking molecular
fiber-level finite element model. Int J Solids Struct 2014; dynamics to finite element simulation of polyethylene
51: 2504–2517. fibers. In: American Society of Composites - 30th technical
182. Sockalingam S, Gillespie JW Jr and Keefe M. Dynamic conference, East Lansing, MI, USA, 28-30 September
modeling of Kevlar KM2 single fiber subjected to trans- 2015.
verse impact. Int J Solids Struct 2015; 67–68: 297–310. 198. Boyd RH, Gee RH, Han J, et al. Conformational
183. Recchia S, Zheng JQ, Horner S, et al. Multiscale mod- dynamics in bulk polyethylene: a molecular dynamics
eling of randomly interwoven fibers for prediction of simulation study. J Chem Phys 1994; 101: 788–797.
KM2 Kevlar yarn strength and damage. Acta Mech 199. Brown D and Clarke JH. Molecular dynamics simula-
2015; 226: 1–10. tion of an amorphous polymer under tension. 1.
184. Sockalingam S, Bremble R, Gillespie JW Jr, et al. Phenomenology. Macromolecules 1991; 24: 2075–2082.
Transverse compression behavior of Kevlar KM2 200. Rottler J and Robbins MO. Growth, microstructure,
single fiber. Composites Part A 2016; 81: 271–281. and failure of crazes in glassy polymers. Phys Rev E
185. Allen MP and Tildesley DJ. Computer simulation of 2003; 68: 011801.
liquids. Oxford: Oxford University Press, 1987. 201. Yamamoto T. Molecular dynamics in fiber formation of
186. Chen C, Depa P, Maranas JK, et al. Comparison of polyethylene and large deformation of the fiber.
explicit atom, united atom, and coarse-grained simula- Polymer 2013; 54: 3086–3097.
tions of poly (methyl methacrylate). J Chem Phys 2008; 202. Curgul S, Van Vliet KJ and Rutledge GC. Molecular
128: 124906–1-12. dynamics simulation of size-dependent structural and

thermal properties of polymer nanofibers. Macromolecules mechanics and molecular dynamics. J Chem Phys
2007; 40: 8483–8489. 2012; 137: 204901.
203. Hossain D, Tschopp M, Ward D, et al. Molecular 206. Silling SA. Reformulation of elasticity theory for dis-
dynamics simulations of deformation mechanisms of continuities and long-range forces. J Mech Phys Solids
amorphous polyethylene. Polymer 2010; 51: 6071–6083. 2000; 48: 175–209.
204. Mattsson TR, Lane JMD, Cochrane KR, et al. First- 207. Xu J, Askari A, Weckner O, et al. Peridynamic analysis
principles and classical molecular dynamics simulation of impact damage in composite laminates. J Aerosp Eng
of shocked polymers. Phys Rev B 2010; 81: 054103. 2008; 21: 187–194.
205. Chantawansri TL, Sirk TW, Byrd EF, et al. Shock
Hugoniot calculations of polymers using quantum

Recent Advances in Modeling and Experiments of Kevlar Ballistic Fibrils, Fibers, Yarns and Flexible Woven Textile Fabrics - A Review

Uploaded by

Copyright:

Available Formats

Recent Advances in Modeling and Experiments of Kevlar Ballistic Fibrils, Fibers, Yarns and Flexible Woven Textile Fabrics - A Review

Uploaded by

Document Information

Original Description:

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

Recent Advances in Modeling and Experiments of Kevlar Ballistic Fibrils, Fibers, Yarns and Flexible Woven Textile Fabrics - A Review

Uploaded by

Copyright:

Available Formats

Review

Textile Research Journal

Recent advances in modeling and ! The Author(s) 2016

Subramani Sockalingam1,2, Sanjib C Chowdhury1,

The textile fabrics possess a hierarchical multi-scale

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Figure 3. Transverse impact.

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Density Modulus Failure Cunniff Theoretical

Spectra 1000 0.97 120 3.50 801 1223

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Acknowledgments 14. Seto T, Hara T and Tanaka K. Phase transformation and

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

Downloaded from trj.sagepub.com at TEXAS SOUTHERN UNIVERSITY on May 7, 2016

You might also like