You are on page 1of 24

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/303624761

Projectile geometry-Online

Data · May 2016

CITATIONS READS

0 192

2 authors:

Aswani Kumar Bandaru Suhail Ahmad


University of Limerick Indian Institute of Technology Delhi
32 PUBLICATIONS   253 CITATIONS    78 PUBLICATIONS   517 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Ballistic View project

Varicomp View project

All content following this page was uploaded by Aswani Kumar Bandaru on 29 May 2016.

The user has requested enhancement of the downloaded file.


2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

International Journal of Applied Mechanics


Vol. 7, No. 3 (2015) 1550039 (23 pages)
c Imperial College Press
DOI: 10.1142/S1758825115500398

Effect of Projectile Geometry on the Deformation


Behavior of Kevlar Composite Armors
Under Ballistic Impact

Aswani Kumar Bandaru∗ and Suhail Ahmad


Department of Applied Mechanics
Indian Institute of Technology Delhi
New Delhi, Delhi 110016, India
∗aswani006@gmail.com

Received 5 September 2014


Revised 20 February 2015
Accepted 20 February 2015
Published 22 June 2015

A failure model based on the three-dimensional strains in a composite layer with


improved progressive damage modeling has been implemented to predict the deformation
behavior of composite armors subjected to ballistic impact. The present model comprises
mainly of two parts. First, quadratic strain based failure criteria are presented to pre-
dict the initiation of failure modes. Second, the post damage softening behavior and
degradation of the material stiffness is measured by damage evolution law. The model
has been implemented within ABAQUS/Explicit as a user defined subroutine VUMAT.
The validity of the model has been carried out by performing computational analysis
of different composite armor materials such as Kevlar 29 and Kevlar 129 impacting
with cylindrical-hemispherical nosed and 120◦ conical projectiles. It transpires that the
predictions from the present model are in good agreement with the experimental and
numerical observations available in the literature in terms of back face signature (BFS)
for both the targets and projectiles. Further, the model has been implemented to study
the effect of projectile geometry on the velocity time histories of the projectile, residual
velocity and ballistic limit velocity. BFS values showed good agreement for conical pro-
jectile while for hemispherical projectile it is slightly low. The combination of Kevlar 129
armor and hemispherical projectile shows higher ballistic limit compared to that of the
other combinations.

Keywords: Kevlar; ballistic impact; projectile; ABAQUS/Explicit; back face signature;


ballistic limit.

1. Introduction
Composites have fully emerged as important engineering materials for applications
where weight of the components is an important consideration. Structures like body
armor should be light in weight having more energy absorbing capability. Developing
a low cost and light weight body armor with good impact resistance is the primary

∗ Corresponding author.

1550039-1
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

A. K. Bandaru & S. Ahmad

objective of the body or vehicle armors research [Cunniff, 1992]. Laminated com-
posites have greater potential in developing light weight body armors. Important
parameters that influence the ballistic impact response of laminated composites are
projectile geometry, thickness of the target, stacking sequence of the composite, ini-
tial impact velocity. Extensive literature is available on the ballistic impact behavior
of laminated composites comprising of various aspects such as, effect of projectile
types, wide range of velocities [van Hoof, 1999], target geometry, stacking sequence
[Cantwell and Morton, 1989; Cheeseman and Bogetti, 2003; Naik et al., 2005; Zhao
et al., 2007; Bandaru and Ahmad, 2014] and damage induced due to impact loading
[Yen, 2002, 2012; Sevkat et al., 2009]. As the test conditions and the geometry are
different in those studies, results cannot be compared.
Parameters like back face signature (BFS), ballistic limit, residual velocity, etc.
are involved in the characterization of composite body armor under ballistic impact.
Body armor standards require that a projectile should be stopped within the armor
and the penetration depth of the armor should not exceed 44 mm [NIJ standard-
0101.04, 2000]. If the penetration depth exceeds this value wearer can undergo blunt
force trauma [Bazhenov, 1997]. To meet these protection requirements for typical
ballistic threats, approximately 20–50 layers of fabric are required [Young et al.,
2003].
Maximum permanent displacement at the back face of the armor during impact
is described as the back face signature (BFS). BFS is the one of the important perfor-
mance factors for assessing body or vehicle armors [NIJ standard — 0101.04, 2000].
At the early phase of ballistic impact, the laminate material undergoes compres-
sion under the projectile and shear deformation takes place through the thickness.
As the top layers get compressed, cone formation develops on the back face of the
target [Naik and Shrirao, 2004]. It causes more strain in the top layers than that of
the bottom layers. Figure 1 shows the deformation and cone formation on the back
face of the composite armor during the ballistic impact.
Composite armors tend to damage under ballistic impact loading and strength
of the armor degrades without causing any visible damage. Experimental and

Point of impact

Direction of projectile

Secondary yarns

Primary yarns

(a) Plan of the back face (b) Side view

Fig. 1. Deformation of a composite during ballistic impact [Morye, 2000].

1550039-2
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

Effect of Projectile Geometry

numerical simulations of ballistic impact on Kevlar-29 were investigated with


STAGNAG-2920 fragment simulating projectile (FSP) by Silva et al. [2005]. Bal-
listic limit and damage modes were predicted, however, this work did not focus on
the BFS, that occurs during the ballistic impact.
BFS of different Kevlar targets was estimated with two different types of projec-
tile by Gower et al. [2008] through experimental tests and LS-DYNA simulations.
The predicted BFS from LS-DYNA simulations showed error of 33% with experi-
mental results but predictions were within the acceptable limit. The BFS for hemi-
spherical projectile was slightly low. The striking velocity of the projectile in the
experiment was between 130 m/s and 250 m/s. Sevkat et al. [2009] performed exper-
imental studies on S2 glass/epoxy to study the damage patterns occurred due to
ballistic impact. Numerical simulation were carried out in LS-DYNA by developing
a nonlinear orthotropic model based on maximum stress theory to model the failure
in fiber and matrix. The experimental results were compared with the numerical
results and also with Chang–Chang failure theory. Kumar et al. [2010] computation-
ally studied the ballistic impact response of thick Kevlar/epoxy composite targets
when impacted with cylindrical projectile. The effect of mass and the diameter of
the projectile on the ballistic response was studied in terms of ballistic limit and
velocity time histories. Even though parametric study was carried out, the influence
of these parameters on BFS was not studied. A finite element model was developed
by Jeoung et al. [2010] for simulating the high velocity impact response of Kevlar
29/Phenolic plate. Tie break and surface-to-surface eroding contact algorithm was
used for simulating the interaction between the projectile and target. Islam et al.
[2011] applied an erosion approach to avoid the problem of element distortion and
proposed a range of erosion parameters for numerical simulation of high velocity
impact of concrete targets by impacting with steel projectiles. Liu et al. [2012]
performed simulations on steel and aluminum alloy plates of different thicknesses
subjected to ballistic impact with steel projectiles. Residual and ballistic limit veloc-
ities were calculated by adopting modified Johnson-Cook model. Islam et al. [2012]
modified the Holmquist–Johnson–Cook model by simplifying the strain rate and
pressure-volume relationship. The modified model was implemented for concrete
under high velocity impact and residual velocities, penetration depths and failure
patterns of the target plates were validated with experimental findings. Singh and
Singh [2014] discussed the effect of projectile characterization, fiber orientation, on
modeling of the ballistic impact response of composites by LS-DYNA. However, in
their review more discussion on the effect of projectile geometry was not discussed.
Since last decade, modeling of the ballistic impact phenomenon using com-
mercial finite element packages like ABAQUS/Explicit and LS-DYNA has been
increased. Through numerical simulations, armor designers may explore and inves-
tigate the ballistic performance of composite armors in an economic way. Both
ABAQUS/Explicit and LS-DYNA are efficient in modeling the ballistic impact
problem. Several authors [Patel et al., 2014; Heimbsa et al., 2014; Xin and Wen,

1550039-3
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

A. K. Bandaru & S. Ahmad

2015] used ABAQUS/Explicit to model the ballistic impact behavior of composite


materials and succeeded. The work by Fu et al. [2013] reported that the differ-
ence of the results obtained from both the simulation softwares is almost negligi-
ble. Similarly, Bandaru and Ahmad [2014] performed numerical simulations using
ABAQUS/Explicit and good validation was reported with LS-DYNA simulations
and experimental results.
The above studies were concentrated on the penetration of the projectile into tar-
get, effect of the target and projectile parameters, damage due to ballistic impact
and determination of ballistic limit velocity. But studies on the effect of various
parameters on the deformation behavior of composite armors are very few. Investi-
gation of the ballistic impact response of several composite armors impacting with
different types of projectile through experiments is time consuming, requires more
manpower and is not economic. Numerical simulations can lighten the expenses
incurred during the development of light weight composite armors. Through numer-
ical simulations, armor designers may explore and investigate the ballistic perfor-
mance of light weight composite armors with different combinations of projectile
and target.
BFS is one of the important performance factors for the assessment of body
armors. Effect of projectile geometry on the deformation behavior of composite
armors with progressive damage modeling has not been reported in the open lit-
erature. Therefore, in the present study, ballistic impact response of woven Kevlar
composite armor is investigated by including the progressive damage behavior of the
armor. The progressive damage model is based on the three-dimensional strains in
composite layer. The damage model is implemented within ABAQUS/Explicit as a
user defined subroutine with solid elements (C3D8). Quadratic strain failure criteria
are used to predict the various failure modes such as fiber failure, matrix cracking,
fiber shear punch and delamination. Composite armors made of Kevlar 29 (K-29)
and Kevlar 129 (K-129) with polyvinyl butyral are considered with the thickness of
9.5 mm (MIL H 44099A). Two rigid projectiles namely, 7.5 mm diameter hardened
steel 120◦ cylindrical-conical nose (conical) projectile and 9 mm hemispherical nose
projectile (hemispherical) are considered. BFS is measured for different projectiles
at various impact velocities and compared with experimental and numerical results
reported in the literature. After attaining a successful validation with the present
model, it is further extended to obtain the velocity time histories and ballistic limit.
Also, the effect of projectile geometry on the ballistic performance of the composite
armors is discussed.

2. Progressive Damage Model for Composites


In the following section, failure in laminates is predicted by a set of proposed failure
criteria based on the progressive damage model. The reduction of material proper-
ties is calculated based on the property degradation model. Different failure modes
implemented are presented in their mathematical forms. A user defined subroutine

1550039-4
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

Effect of Projectile Geometry

is developed based on the progressive damage model to assess the ballistic impact
response of Kevlar composite armors. Damage in composites, occurs in two phases:

(i) Damage initiation and


(ii) Damage propagation (or evolution).

Damage initiates, when the failure criterion is satisfied for a particular element.
Physically, the damage initiation indicates the beginning of formation of micro
cracks in an element. After damage initiation, the calculation of damage variable
starts. This is the damage evolution phase where micro cracks either expand or
new micro cracks generate. When the damage variable reaches a critical value, the
element is deleted from the simulations. This critical value is considered as 0.99 for
the present study.

2.1. Damage initiation


The damage initiation or failure criteria are similar to those given by Yen [2012].
In this failure model, failure initiation criteria and damage evolution laws are con-
sidered to account for the major fiber failure modes such as tensile, compressive,
punch shear and crush. In addition, two matrices related failure modes like in-plane
shear and through the thickness delamination are also implemented. This failure
model is based on the three-dimensional strains in a composite layer with improved
progressive failure modeling capability. This failure criterion can be used to effec-
tively simulate the fiber failure under high strain-rate and high pressure ballistic
conditions. The failure criteria adopted in the present study is as follows:

fi − ri2 = 1 (1)

where parameters fi are functions based on the quadratic interaction between


strains and ri are damage thresholds which are initially unity. The following are
the different failure modes implemented in the present study.

2.1.1. Fiber failure modes


Three fiber failure modes are considered in the present subroutine: (a) fiber shear
punch, (b) fiber failure in compression and (c) fiber crush. E11 , E22 , E33 are the
elastic constants indicating in-plane fill (11 direction), in-plane warp (22 direction)
and out-of-plane (33 direction) directions respectively, G12 , G23 , G13 are the shear
moduli in 12, 23 and 13 planes respectively, and ε11 , ε22 , ε33 are the strains in 1, 2
and 3 directions respectively.

(a) Fiber shear punch


Based on the Hashin’s criteria for a unidirectional layer, the quadratic interaction
between the associated axial and through the thickness shear strains is given by

1550039-5
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

A. K. Bandaru & S. Ahmad

following criterion:
 2  2
E11 ε11 G13 ε13
+ − r12 = 0 (2)
S1T S13
 2  2
E22 ε22 G23 ε23
+ − r22 = 0 (3)
S2T S23

where S1T and S2T are the tensile strengths in fill and warp directions respectively,
S13 and S23 are the shear strengths due to fiber shear failure in fill and warp
directions respectively, and r1 and r2 are the corresponding damage thresholds.

(b) Fiber failure in compression


In-plane compressive damage in fill and warp directions is considered by the max-
imum strain criterion, when ε11 is compressive. Effect of transverse compressive
strains on in-plane compressive damage is considered as follows:
 2
E11 ε11
− r32 = 0 (4a)
S1C
 2
E22 ε22
− r32 = 0 (4b)
S2C
x+|x|
where ε11 = −ε11 − −ε33  E  E33
E11 , ε22 = −ε22 − −ε33  E22 and x =
33
2 , S1C and
S2C are the compressive strengths of fiber in fill and warp directions respectively,
and r3 is the associated damage threshold.

(c) Fiber crush


When composite armor impacts by a projectile, high compressive waves are gener-
ated near impact zone with high shear stresses near the vicinity of the adjoining
area between the projectile and target. As the fiber shear punch failure occurs due
to higher shear stresses as considered in Eqs. (2) and (3), fiber crush damage due
to through the thickness high compressive pressure is modeled using the following
criterion:
 2
E33 ε33
− r42 = 0 (5)
S3C

where, S3C is the fiber compressive strength and r4 is the associated damage
threshold.

2.1.2. Matrix failure modes


The lamina can undergo damage due to in-plane shear stress without the occurrence
of fiber breakage. The matrix damage mechanisms considered are damage of the

1550039-6
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

Effect of Projectile Geometry

plane perpendicular to the layer plane and damage of the plane parallel to the layer
plane (delamination).
The perpendicular matrix damage associated with transverse matrix cracking is
modeled using the following equation:
 2
G12 ε12
− r52 = 0 (6)
S12
where S12 is the strength of the matrix in in-plane shear, r5 is the corresponding
damage threshold.
Parallel matrix damage (delamination) is another failure mode due to quadratic
interaction between the stresses in through the thickness directions and is modeled
using the following equation:
 2  2  2 
E 33 ε 3 G 23 ε 23 G 13 ε 13
S2 + + − r62 = 0 (7)
S3T S230 + SSR S130 + SSR
where S3T is the through thickness tensile strength, S230 and S130 are the shear
strengths for tensile ε3 , SSR is the residual sliding strength [Yen, 2012], and r6 is
the associated damage threshold. The damage surface due to Eq. (7) is parallel to
the composite layering plane.
Under compressive strain in through the thickness, ε3 < 0, the damage surfaces
generated due to matrix cracking are considered to be in full contact and the damage
strengths are assumed to be dependent on ε3 similar to Coulomb–Mohr theory [Yen,
2012], i.e.
SSR = E33 tan ϕ−ε3  (8)
where ϕ is the Coulomb’s friction angle.
A user defined subroutine has been developed for the above explained failure
criteria and is compiled and linked with ABAQUS/Explicit to obtain the needed
information regarding the state of the material and the material transient response
during each time step, for each integration point of each element.

2.2. Damage evolution


The preceding section explains the damage initiation through different failure
modes. In this section, propagation of damage is discussed. Composite material
behavior describing the relation between the stresses (σ) and strains (ε) is expressed
as follows:
[σ] = [C][ε] (9)
In Eq. (9), [C] = [S −1 ], where [C] is known as stiffness matrix, [σ] is stress matrix
and [ε] is strain matrix. The elements of [S] are expressed in terms of Young’s
moduli, shear moduli and Poisson’s ratios of the material.
Once damage initiation criteria are satisfied, further increase in the load causes
degradation of the material through reduction in stiffness coefficients. When damage

1550039-7
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

A. K. Bandaru & S. Ahmad

Damaged area

Undamaged area

Fig. 2. Damaged and undamaged areas of a typical cross-section.

occurs in the target due to ballistic impact, a parameter called “damage variable
(i )” for i = 1, . . . , 6 is introduced which is theoretically the ratio of damaged area
and total area of cross-section of specimen as shown in Fig. 2.
Using this damage variable, the material constitutive relation described in
Eq. (9) can be expressed as equation given below:

[σ̃] = [C̃][ε̃] (10)

where, [C̃] = [S̃ −1 ], and [S̃] is expressed as


 
1 ν21 ν31
− − 0 0 0
(1 − 1 )E11 E22 E33 
 
 
 ν12 1 ν32 
 − − 0 0 0 
 E11 (1 − 2 )E22 E33 
 
 
 ν13 ν23 1 
 − − 0 0 0 
 E11 E22 (1 − 3 )E33 
[S̃] = 



 1 
 0 0 0
(1 − 4 )G12
0 0 
 
 
 1 
 0 0 0 0 0 
 (1 − 5 )G23 
 
 
 0 0 0 0 0
1 
(1 − 6 )G13

(11)

The damage variable i (i = 1, . . . , 6) has been is estimated using the following


equation:
m
i = 1 − e(1−rj )/m
(12)

where m is the damage parameter which decides the shape of the damage evolution
curve. More is the value of m, more abruptly the stress components will go down
(soften) after initiation of stresses. Damage increases with loading and remains
constant during unloading. When the damage variable reaches a critical value, the
element is deleted from further calculations. This critical value is taken as 0.99 for
the present study. Otherwise, if it is allowed to reach unity, the constitutive matrix

1550039-8
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

Effect of Projectile Geometry

will become singular. The deletion of failed elements is made by making status
variable (status Mp) in VUMAT code zero.
Also, for a given jth mode, a set of damage variables is updated. This is decided
by using a coupling matrix q defined as:
 
1 0 1 0 1 0 0
 
0 1 0 1 1 0 0
 
0 0 0 0 1 0 0
 
[q] =   (13)
1 1 1 1 1 1 0
 
 
0 0 1 1 1 0 1
1 1 0 0 1 0 1

The matrix [q] is written here for all modes of failure for completeness. However,
in the present study, the columns of the [q] matrix used are 1, 3, 4 and 6. For a
given mode say jth mode, the entries in jth column are checked. If ith entry in jth
column is one (qij = 1), then damage variable i is changed or updated. Table 1
shows the mode and damage variable updated/changed.
Damage evolution deals with the calculation of damage variables which are
responsible for degrading the material properties of the material. This requires the
calculation of failure strain at which, complete failure of the element occurs. The
strain is calculated using the following equation:
2G
εf = (14)
Slc
where S is the strength of the material in tension/compression or shear based on
the mode of failure and lc is the characteristic length.
Characteristic length is calculated using the objectivity algorithm given by Don-
adan et al. [2008]. According to this algorithm, quantities (ϕi ) are calculated at
nodes based on the orientation and location of the crack. The value of φ is unity if
the node is ahead of crack and it is zero if the node is behind the crack. At the rest
of the locations, the interpolated values of ϕ are calculated using iso-parametric
shape functions (Ni ). Thus, ϕ can be written as
n

ϕ= ϕi Ni (15)
i=1

Table 1. Mode versus damage variables.

Mode Damage variables updated/changed


1 1 , 4 , 6
3 1 , 4 , 5
4 1 , 2 , 3 , 4 , 5 , 6
6 4

1550039-9
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

A. K. Bandaru & S. Ahmad

Characteristic length is calculated based on the gradient of ϕ in the direction of the


crack using the following equation:
 n   
∂Ni ∂Ni
lc = cos(θ) + sin(θ) ϕi (16)
i=1
∂ξ ∂η
where θ is the orientation of the crack w.r.t x-axis.
The tie break surface-to-surface contact is used to simulate the delamination
between the layers. The tie break contact works the same way as a common contact
in compressive loads, but under tension it allows the separation of the tied surface
under a failure criteria given below:
 2  2
σn σs
+ ≥1 (17)
NFLS SFLS
where σn and σs are the normal and shear stresses acting on the layer interface
respectively, while NFLS and SFLS are the normal and shear strengths of the layer
interface, respectively.

3. Numerical Study
The progressive damage model discussed in the previous section is implemented
into ABAQUS/Explicit as a user defined subroutine. With the damage initiation
and evolution, the instantaneous stiffness matrix is updated at each time step and
the response is obtained.
The composite armors considered in the present study are made of K-29 and
K-129 with polyvinyl butyral-phenolic matrix resin (18% volume fraction) having
19 plies equivalent to MIL-H-44099A standard thickness of 9.5 mm. The geometric
properties of the projectile and material constants used for composite material are
illustrated in Tables 2 and 3, respectively. The ballistic response of the armor plate is
investigated through the direct impacts of two rigid and non-deforming projectiles
(cylindrical 120◦ conical 7.5 mm diameter and 9 mm hemispherical nosed) made
up of heat treated AISI 4340 steel are used to enhance the different failures. The
dimensions of the composite plate are 254 mm × 304.8 mm × 9.5 mm, which are the
typical measurements used for body armor.
The plate geometry is partitioned to include a central target location that allows
a finer mesh to be applied near the impact zone. This finer mesh is transitioned to
a coarse mesh at the outer edges of the plate using the sweep mesh technique.
Figure 3 shows the details of the finite element mesh implemented for the projectile
and target in the present study.

Table 2. Geometrical properties of the projectile.

Projectile Mass (g) Length (m) Nose radius (m)


Hemispherical 20 0.04 0.009
Conical 120◦ 20 0.04 0.0075

1550039-10
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

Effect of Projectile Geometry

Table 3. Kevlar 29 and Kevlar 129 properties.

Properties Woven Kevlar 29 Woven Kevlar 129


[van Hoof, 1999] [Gower et al., 2008]
E1 (GPa) 18.5 22
E2 (GPa) 18.5 22
E3 (GPa) 6.0 9
G12 (GPa) 0.77 0.77
G13 (GPa) 5.43 2.715
G23 (GPa) 5.43 2.715
υ12 0.25 0.25
υ13 0.33 0.33
υ23 0.33 0.33
ρ (g/cm3 ) 1.23 1.23
Strength (MPa)
X=Y 1850 800
Z 1200 1200
S12 77 77
S23 = S13 543 1000
Sn 34.5 34.5
Ss 9 9

(a) (b)
Fig. 3. Finite element model. (a) Projectiles and (b) composite laminate.

Each lamina in the laminated composite armor is modeled as a separate entity


with surface to surface tie break contact. The composite armor is discretized with
C3D8 elements (continuum three-dimensional eight noded solid elements) and the
impactor with R3D4 elements (three-dimensional four noded rigid elements) is used.
Both ends of the composite beam are fixed. Therefore, the boundary conditions
imposed are (u = v = w = 0 and θx = θy = θz = 0). The coefficient of friction
between the impactor and the plate is considered as 0.2.

4. Results and Discussion


The progressive damage model adopted in the present study is validated by compar-
ing the computational results with the experimental and numerical results reported
in [Gower et al., 2008]. Variation of BFS with time and velocity have been presented
are discussed and good agreement is observed. Further, velocity time histories, resid-
ual velocity and ballistic limit for different combinations of the projectile and target
are presented as well.

1550039-11
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

A. K. Bandaru & S. Ahmad

4.1. Back face signature (BFS)–time relationship


Numerical predictions of BFS-time relationships are validated for hemispherical
projectile impacting 19 ply K-29 armor with an impact velocity of 146 m/s and
conical projectile impacting K-129 armor with an impact velocity of 244 m/s.
Figure 4 shows the BFS–time relationship for the hemispherical projectile
impacting K-29 armor at 146 m/s. A good agreement can be observed for the initial
slopes of the numerical and experimental predictions. The maximum BFS obtained
from the present study is within the 4.3% of LS-DYNA predicted value while with
experiment results it is 38.1%. The difference of BFS prediction between the LS-
DYNA simulations and experimental tests reported in literature is 40.72%. Despite,
the numerical predictions show too much deviation from experimental findings, how-
ever, the BFS obtained is within the acceptable limits given in the standards [NIJ
0101.04, 2000].
Figure 5 presents the BFS-time relationships for the conical projectile impacting
K-129 armor at 244 m/s. The variation of BFS with time from the present study
is 7.2% more than that of the experimental predictions. The predicted BFS-time
variation is in good agreement with the literature, although the values are slightly
higher. To assess the accuracy of the numerical simulation the initial slope of the
BFS-time graph can be used. There is a slight increase in the slope of the numer-
ical simulation for conical projectile compared to that of the experimental results
reported in the literature.
A little effect can be predicted on an initial slope by changing the material
properties such as compressive strength the through the thickness, membrane strain,
membrane modulus, or modulus in the fiber direction (E1 ), tensile strength and
shear strength along the thickness [Gower et al., 2008].

14

12
Back Face Signature (mm)

10

4
Present study
2 LS-DYNA [Gower et al., 2008]
Experimental [Gower et al., 2008]
0
0.0000 0.0001 0.0002 0.0003
Time (Sec)

Fig. 4. K-29 impacting with hemispherical projectile at 146 m/s.

1550039-12
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

Effect of Projectile Geometry

25

20
Back Face Signature (mm)

15

10

5 Present study
LS-DYNA [Gower et al., 2008]
Experimental [Gower et al., 2008]
0
0.0000 0.0001 0.0002 0.0003
Time (sec)

Fig. 5. K-129 impacting with conical projectile at 244 m/s.

Kevlar 29-Hemispherical
25 Kevlar 129-Hemispherical
25 146 m/s, 175 m/s, 205 m/s
250 m/s, 300 m/s, 350 m/s 150 m/s, 200 m/s, 244 m/s
400 m/s, 450 m/s, 500 m/s 300 m/s, 350 m/s, 400 m/s
450 m/s, 500 m/s
20 20
Back Face Signature (mm)
Back Face Signature (mm)

15 15

10 10

5 5

0 0
0.0000 0.0001 0.0002 0.0003 0.0000 0.0001 0.0002 0.0003
Time (sec) Time (sec)

(a) (b)
Fig. 6. BFS time histories (a) K-29 and (b) K-129.

Figures 6(a) and (b) show the time histories of BFS and corresponding maximum
value of BFS for hemispherical projectile impacting K-29 and K-129 armors at
different impact velocities. It indicates that the dynamic displacement of the armor
increases with increase in the impact velocity.

1550039-13
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

A. K. Bandaru & S. Ahmad

4.2. BFS–impact velocity relationship


BFS is obtained at various impact velocities for K-29 and K-129 armors. Validation
of BFS obtained from the present model is shown in Figs. 7 and 8 along with the
influence of projectile geometry.
Figure 7 shows the comparison between the numerically predicted BFS from the
present model with the experimental data for hemispherical and conical projectiles
impacting K-29 armor at different velocities. It can be seen from Fig. 7(a) that the
BFS predictions from the present model for K-29 armor are in good agreement with
LS-DYNA results showing an error of 3.26%.
Also in Fig. 7(b), the BFS predictions from the present study for conical pro-
jectile impacting K-29 armor are in good agreement with the experimental results
showing an error of 6–12.2%, while with LS-DYNA predictions showing an error
of 13.5%. LS-DYNA predictions of BFS reported by Gower et al. [2008] showed an
error of 33% with the experimental results while in the present study it is 33.7%.
This indicates that the predictions from the present model are also acceptable.
Figure 8 shows the comparison of numerically predicted BFS from the present
model with the experimental results for hemispherical and conical projectiles
impacting K-129 armor at different velocities. Figure 8(a) shows the BFS predictions
from the present model for K-129 armor when impacted with hemispherical projec-
tile are in good agreement with experimental results showing an error of 14.23%.
It can be seen from Fig. 8(b) that the BFS predictions from the present study for
conical projectile impacting K-129 armor are very close to experimental results with
an error of 7.2–11.4%, while with LS-DYNA predictions it is 11.52%.

25
Kevlar 29-Hemispherical Kevlar 29- Conical
20 Present study Present study
Experimental [Gower et al., 2008] Experimental [Gower et al., 2008]
LS-DYNA [Gower et al., 2008] LS-DYNA [Gower et al., 2008]
20
Back Face Signature (mm)
Back Face Signature (mm)

16

15

12
10

8 5
125 150 175 200 225 125 150 175 200 225 250
Impact Velocity (m/s) Impact Velocity (m/s)

(a) (b)
Fig. 7. Variation of BFS with impact velocity for K-29. (a) Hemispherical projectile and (b)
conical projectile.

1550039-14
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

Effect of Projectile Geometry

25
20 Kevlar 129-Hemispherical Kevlar 129-Conical
Present study
Present study
Experimental [Gower et al., 2008]
Experimental [Gower et al., 2008]
20 LS-DYNA [Gower et al., 2008]

Back Face Signature (mm)


Back Face Signature (mm)

16

15

10
12

5
150 160 170 180 190 200 125 150 175 200 225 250
Impact Velocity (m/s) Velocity (m/s)

(a) (b)
Fig. 8. Variation of BFS with impact velocity for K-129. (a) Hemispherical projectile and (b)
conical projectile.

The predicted BFS values for hemispherical projectile impacting K-29 and K-
129 armors are low due to increased delamination within the plane. The BFS values
predicted in the present study are validated with the experimental and numerical
results (LS-DYNA) reported in the literature and are shown in Tables 4–7. Good
agreement between different results establishes the validity of the VUMAT code
developed.
From Tables 4–7, it is observed that the BFS values observed from the present
model are in good agreement with experimental results and LS-DYNA simulations

Table 4. Kevlar 29 with hemispherical projectile.

Velocity (m/s) Present study Back face signature (mm)


[Gower et al., 2008] Error (%) [Present study]
LS-DYNA Experimental Error (%) LS-DYNA Experimental
146 10.07 9.65 16.28 40.72 4.35 38.1
175 11.57 11.79 17.75 33.57 1.86 34.8
205 13.77 14.28 19.19 25.58 3.57 28.2

Table 5. Kevlar 29 with conical projectile.

Velocity (m/s) Present study Back face signature (mm)


[Gower et al., 2008] Error (%) [Present study]
Numerical Experimental Error (%) LS-DYNA Experimental
155 10.21 8.30 11.64 28.69 23.0 12.2
198 11.12 12.73 13.21 3.63 12.6 15.8
244 24.26 24.11 22.88 5.37 4.9 6.0

1550039-15
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

A. K. Bandaru & S. Ahmad

Table 6. Kevlar 129 with hemispherical projectile.

Back face signature (mm)


Velocity (m/s) Present Experimental [Gower et al., 2008] Error (%)
150 11.35 14.01 18.9
180 15.49 17.81 13.0
198 17.29 19.40 10.8

Table 7. Kevlar 129 with conical projectile.

Velocity (m/s) Present study Back face signature (mm)


[Gower et al., 2008] Error (%) [Present study]
LS-DYNA Experimental Error (%) LS-DYNA Experimental
150 11.69 9.57 12.60 24.04 22.1 7.2
198 13.26 11.98 14.98 20.02 10.6 11.4
244 23.01 22.59 21.34 5.85 1.86 7.8

reported in [Gower et al., 2008]. The difference of BFS for each Kevlar composite
armor when impacted with different projectiles is compared. The error between the
predictions of ABAQUS/Explicit [Present study] and LS-DYNA are in the range of
3–13% showing the predictive capability of the present model.
Figures 9(a) and 9(b) shows the difference between the experimental and numer-
ical simulations for K-29 composite armor when impacted with hemispherical and
conical projectiles. From Table 4 and Fig. 9(a), it can be observed that the BFS
predictions from the present model show an error of 33.7% with experimental results

Kevlar 29 - Hemispherical 37.5 Kevlar 29 - Conical


45
Experimental vs LS-DYNA [Gower et al., 2008] Experimental vs LS-DYNA [Gower et al., 2008]
Present study vs Experimental Present study vs Experimental
Present study vs LS-DYNA Present study vs LS-DYNA
30.0

30
22.5
Error (%)

Error (%)

15.0
15

7.5

0 0.0
150 180 210 160 200 240
Impact Velocity (m/s) Impact Velocity (m/s)

(a) (b)
Fig. 9. Variation of percentage error in BFS predictions for K-29. (a) Hemispherical projectile
and (b) conical projectile.

1550039-16
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

Effect of Projectile Geometry

20 28 Kevlar 129 - Conical


Kevlar 129 - Hemispherical Experimental vs LS-DYNA [Gower et al., 2008]
Present study vs Experimental Present study vs Experimental
Present study vs LS-DYNA
15 21
Error (%)

Error (%)
10 14

5 7

0 0
140 160 180 200 150 200 250
Impact Velocity (m/s) Impact Velocity (m/s)

(a) (b)
Fig. 10. Variation of percentage error in BFS predictions for K-129. (a) Hemispherical projectile
and (b) conical projectile.

and 13.5% with LS-DYNA results. The error between the experimental and numer-
ical results reported by Gower et al. [2008] is 33% and is in close agreement with
the present ABAQUS/Explicit model.
Figure 9(b) and Table 5 illustrates the BFS values obtained from the present
model with an error of 11.33% with experimental predictions and 13.5% with numer-
ical simulations reported in the literature. The error reported in the literature is
12.5%, which shows the close agreement with the present model.
Figures 10(a) and 10(b) shows the difference between the experimental and
numerical predictions for K-129 armor when impacted with hemispherical and con-
ical projectiles. From Table 6 and Fig. 10(a), it can be observed that predictions of
BFS from present numerical results shows an error of 14% with reported experimen-
tal results. As no comparison was reported in the literature between experimental
and LS-DYNA predictions, it is not presented. Figure 10(b) illustrates the differ-
ence for prediction of BFS from the present model with experimental and numerical
predictions. An error of 8.8% has been observed between present and experimental
results and for the present model and LS-DYNA it is 11.52%. The error reported
in the literature is 16.63%.
From all the above discussions it can be concluded that, the BFS values predicted
from the present model are within the acceptable limits. The predicted values are
in good agreement with that of experimental and LS-DYNA results indicating the
accuracy of the present model and the error observed is in acceptable limits.
Figure 11 shows the variation of BFS with impact velocities estimated from
the present material model. From the present study, both the panels showed almost
same BFS values when impacted with conical projectile at low speeds, however, this
changed for higher impact velocities. K-29 armors show reduced BFS compared to
that of K-129 armors when impacted with hemispherical projectile.

1550039-17
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

A. K. Bandaru & S. Ahmad

27
K29-Hemispherical
K129-Hemispherical
24
K29-Conical
K129-Conical
21
Back Face Signature (mm)

18

15

12

6
140 160 180 200 220 240 260
Impact Velocity (m/s)

Fig. 11. Variation of BFS with impact velocity for K-29 and K-129.

4.3. Velocity time histories and residual velocity


In order to investigate the effect projectile geometry on the impact velocity, sim-
ulations are carried out with impact velocities ranging from 300–500 m/s. Impact
velocity as a function of time is presented for the both K-29 and K-129 armors
when they are impacted with both the projectiles. The positive residual velocity
represents the penetration of the projectile through the armor and negative resid-
ual velocity indicates rebounding of the projectile from the armor. Ballistic limit
velocity is also calculated based on the velocity time histories and residual velocity.
Figure 12(a) shows the velocity-time histories for hemispherical projectile
impacting K-29 armor at different impact velocities. From these velocity time

600
160
Kevlar 29-Hemispherical
Kevlar 29-Hemispherical
350, 397, 398
500 399, 450, 500
120
Residual Velocity (m/s)
Impace Velocity (m/s)

400

80
300

200 40

100
0
350 400 450 500
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30
-40
Time (ms) Impact Velocity (m/s)

(a) (b)
Fig. 12. Effect of impact velocity on K-29 impacting with hemispherical projectile. (a) Velocity
time histories and (b) residual velocity.

1550039-18
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

Effect of Projectile Geometry

500 200
Kevalr129-Hemiospherical
Kevlar129-Hemispherical
310, 323, 325 327,
400 350, 400, 450,
150
Impact Velocity (m/s)

Residual Velocity (m/s)


300
100

200
50

100
0
300 325 350 375 400 425 450 475
0
0.0 0.1 0.2 0.3 0.4 0.5
-50
Time (ms) Impact Velocity (m/s)

(a) (b)
Fig. 13. Effect of impact velocity on K-129 impacting with hemispherical projectile. (a) Velocity
time histories and (b) residual velocity.

histories residual velocities for corresponding impact velocities are measured. From
Fig. 12(a), it can be seen that, at the velocities of 397 m/s and 398 m/s partial pene-
tration is observed and projectile rebounds back. For an impact velocity of 399 m/s,
the residual velocity is very low, but negative, where the projectile is rejected by the
plate. Figure 12(b) represents the residual velocity as a function of impact velocity.
It shows that in the present model, the residual velocity almost becomes zero for an
impact velocity of 399 m/s. At an impact velocity of 400 m/s the residual velocity
is higher than that of 399 m/s. Therefore, the ballistic limit for the combination of
hemispherical projectile and K-29 armor is 399 m/s. The residual velocities for the
initial impact velocities are presented in Table 8.
Velocity time histories for hemispherical projectile impacting K-129 armor at
various impact velocities are shown in Fig. 13(a). At the impact velocities of 323 m/s
and 325 m/s partial penetration is observed and the projectile rebounded back. For
an impact velocity of 327 m/s, the residual velocity is close to zero, indicating that
the maximum kinetic energy of the projectile is absorbed by the target.

Table 8. Impact velocities and corresponding


residual velocities.
Kevlar 29 with hemispherical projectile
Impact velocity (m/s) Residual velocity (m/s)
350 −25.49
397 −2.37761
398 −1.5407
399 −0.37906
400 11.27
450 82.8776
500 146.659

1550039-19
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

A. K. Bandaru & S. Ahmad

Figure 13(b) shows the residual velocity as a function of impact velocity. It


shows that residual velocity approaches zero for an impact velocity of 327 m/s. For
an impact velocity of 330 m/s the residual velocity is much higher than that of
327 m/s. K-129 armor when impacted with hemispherical projectile, the armor can
resist an impact velocity of 327 m/s, beyond which full perforation of the target
takes place, hence, the ballistic limit for the combination of hemispherical projectile
and K-129 armor is 327 m/s. Table 9 shows the details of impact velocities and their
corresponding residual velocities.

Table 9. Impact velocities and corresponding


residual velocities.
Kevlar 129 with hemispherical projectile
Impact velocity (m/s) Residual velocity (m/s)
310 −40.5326
323 −35.1638
325 −6.13758
327 −0.057
330 10.53786
350 50.7455
400 110.516
450 187.445

Figure 14(a) shows the velocity-time histories for conical projectile impact-
ing K-29 armor at different velocities. From these velocity time histories resid-
ual velocities for corresponding impact velocities are measured. At the velocities of
350 m/s and 375 m/s partial penetrations are observed and projectile rebounds back.
For an impact velocity of 377 m/s, the residual velocity is very low, but negative,
where the projectile is rejected by plate.

250
Kevlar 29- Conical
500 Kevlar 29 - Conical
350, 375, 377
378, 379, 400
200
450, 500
400
Residual Velocity (m/s)
Impact Velocity (m/s)

150
300

100
200

100 50

0 0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 325 350 375 400 425 450 475 500 525
Time (ms) Impact Velocity (m/s)

(a) (b)
Fig. 14. Effect of impact velocity on K-29 impacting with conical projectile. (a) Velocity time
histories and (b) residual velocity.

1550039-20
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

Effect of Projectile Geometry

Figure 14(b) shows the residual velocity as a function of impact velocity. It


shows that in the present model the residual velocity has almost become zero for an
impact velocity of 377 m/s. For an impact velocity of 379 m/s the residual velocity is
higher than that of 377 m/s. Thus, for this combination of K-29 armor and conical
projectile the ballistic limit is 377 m/s. Table 10 gives the details of the impact
velocity and their corresponding residual velocities.

Table 10. Impact velocities and corresponding


residual velocities.
Kevlar 29 with conical projectile
Impact velocity (m/s) Residual velocity (m/s)
350 −7.8922
375 −1.02949
377 −0.28783
378 0.367621
379 4.55193
400 6.90582
450 127.588
500 203.276

Table 11 summarizes the ballistic limit values for hemispherical projectile


impacting K-29 and K-129 armors and conical projectile impacting K-29 armor.
The ballistic limit values are calculated from velocity time history curve and from
residual velocity plots. The highest ballistic limit has been observed for the combi-
nation of K-29 armor and hemispherical projectile.

Table 11. Predicted values of ballistic limit.


Target/Projectile Ballistic limit (m/s)
K-29/Hemispherical 399
K-129/Hemispherical 327
K-29/Conical 377

5. Conclusion
In the present study, deformation behavior of Kevlar 29 (K-29) and Kevlar 129 (K-
129) composite armors subjected to ballistic impact is investigated by implementing
a progressive damage model based on the three dimensional strains in a composite
layer. The thickness of the armors considered according to MIL-H-44099A stan-
dard. Various failure modes such as fiber shear punch, fiber failure in compression,
fiber crush, matrix cracking and delamination are considered in this model. Post
damage softening behavior and degradation of the material stiffness is measured
by damage evolution law. The present model can give an insight into the damage
growth and progression that occurs during the ballistic impact of composite armors.

1550039-21
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

A. K. Bandaru & S. Ahmad

The present model predictions are in good agreement with the experimental results
in terms of back face signature (BFS). Finally, by comparing the simulations with
experimental data, it is concluded that the present model is adequate in modeling
the deformation behavior of Kevlar composite armors having standard thickness.
Based on the present work following conclusions are drawn:
• The BFS predictions for conical projectile impacting K-29 armor with the exper-
imental results are in good agreement with maximum error of 12.2% while for
hemispherical projectile the predicted BFS is lower. This endorsed to improved
delamination occurring in the panel.
• The numerical results for conical projectile impacting K-129 armor, predicted
BFS very close to experimental results with maximum error of 11.4% while for
hemispherical projectile BFS values are slightly lower.
• The differences in the BFS attributes the effect of projectile geometry on the
ballistic performance of composite armors.
• Ballistic limit velocity of K-29 with hemispherical projectile is found to be higher
compared to that of other combinations studied.
• The variation of BFS from the present model shows close agreement for conical
projectile impacting both the composite plates, while for hemispherical projectile
it is slightly low. Even though, BFS values are deviated slightly from experimental
results, but they are in the acceptable limits defined in the standards.

References
ABAQUS User’s Manual, Version 6.10.1.
Bandaru, A. K. and Ahmad, S. [2014] “Numerical simulation of progressive dam-
age of laminated composites under ballistic impact,” Safety, Reliability, Risk and
Life-Cycle Performance of Structures and Infrastructures (CRC Press, London),
pp. 4375–4382.
Bazhenov, S. [1997] “Dissipation of energy by bullet proof aramid fabric,” Journal of
Materials Science 32, 4167–4173.
Cantwell, W. J. and Morton, J. [1989] “Comparison of the low and high velocity impact
response of CFRP,” Composites 20(6), 545–551.
Cheeseman, B. A. and Bogetti, T. A. [2003] “Ballistic impact into fabric and compliant
composite laminates,” Composite Structures 61(1–2), 161–173.
Cunniff, P. M. [1992] “An analysis of the system effects in woven fabrics under ballistic
impact,” Textile Research Journal 62(9), 495–509.
Donadan, M. V., Iannuci, L., Falzon, B. G., Hodgkinson, J. M. and de Almeida, S. F. M.
[2008] “A progressive failure model for composite laminates subjected to low velocity
impacts,” Computers and Structures 86, 1232–1252.
Gillespie, Jr., Xiao, J. L. and Gama, B. A. [2007] “Progressive damage and delamination
in plain weave S-2 Glass/SC-15 composites under quasi static punch shear loading,”
Composite Structures 78, 182–196.
Gower, H. L., Cronin, D. S. and Plumtree, A. [2008] “Ballistic impact response of laminated
composite panels,” International Journal of Impact Engineering 35(9), 1000–1008.
Heimbsa, S., Bergmanna, T., Schuelerb, D. and Pentecôteb, N. T. [2014] “High velocity
impact on preloaded composite plates,” Composite Structures 111, 158–168.

1550039-22
2nd Reading
June 19, 2015 8:28 WSPC-255-IJAM S1758-8251 1550039

Effect of Projectile Geometry

Islam, M. J., Liu, Z. and Swaddiwudhipong, S. [2011] “Numerical study on concrete pen-
etration/perforation under high velocity impact by ogive-nose steel projectile,” Com-
puters and Concrete 8(1), 111–123.
Islam, M. J., Swaddiwudhipong, S. and Liu, Z. S. [2012] “Penetration of concrete targets
using a modified Holmquist–Johnson–Cook material model,” International Journal of
Computational Methods 9, 1250056, doi:10.1142/S0219876212500569.
Jeoung, H. A., Khanh, H. N., Yong, B. P., Jin, H. K. and Jin, H. C. [2010] “A numerical
study of the high-velocity impact response of a composite laminate using LS-DYNA,”
International Journal of Aeronautical and Space Science 11(3), 221–226.
Kedar, S. P., Lokesh, D., Jayaram, R. P., Ravi Kumar, G. and N. K. Naik [2012] “Stress
wave attenuation in composites during ballistic impact,” Polymer Testing 31(2), 261–
266.
Lee Young, S., Wetzel, E. D. and Wagner, N. J. [2003] “The ballistic impact characteristics
of Kevlar woven fabrics impregnated with a colloidal shear thickening fluid,” Journal
of Materials Science 38, 2825–2833.
Liu, Z. S., Swaddiwudhipong, S. and Islam, M. J. [2012] “Perforation of steel and aluminum
targets using a modified Johnson–Cook material model,” Nuclear Engineering and
Design 250, 108–115.
MIL-H-44099A:1986. Military Specification, Helmet, Ground troops and parachutists.
Morye, S. S., Hine, P. J., Duckett, R. A., Carr, D. J. and Ward, I. M. [2000] “Modelling
of the energy absorption by polymer composites upon ballistic impact,” Composites
Science and Technology 60, 2631–2642.
Naik, N. K. and Shrirao, P. [2004] “Composite structures under ballistic impact,” Com-
posite Structures 66(1–4), 579–590.
NIJ standard-0101.04 [2000] “Ballistic resistance of personal body armor,” National Insti-
tute of Justice and the National Institute of Standards and Technology.
Patel, S. D., Ahmad, S. and Mahajan, P. [2014] “Reliability analysis of composite plate
under low velocity impact using Gaussian response surface method,” International
Journal for Computational Methods in Engineering Science and Mechanics 15(3),
218–226.
Sevkat, E., Liaw, B., Delale, F. and Raju, B. B. [2009] “A combined experimental and
numerical approach to study ballistic impact response of S2-glass fiber/toughened
epoxy composite beams,” Composites Science and Technology 69, 965–982.
Silva, M. A. G., Cismasiu, C. and Chiorean, C. G. [2005] “Numerical simulation of ballis-
tic impact on composite laminates,” International Journal of Impact Engineering 31,
289–306.
Singh, N. K. and Singh, K. K. [2014] “Review son impact analysis of FRP composites
validated by LS-DYNA,” Polymer Composites, 1-13, 10.1002/pc.23064.
van Hoof, J. [1999] “Modelling impact induced delamination in composite materials,”
Ph.D. thesis, Mechanical and Aerospace Dept., Carleton University, Ottawa, Ontario.
Xin, S. H. and Wen, H. M. [2015] “A progressive damage model for fiber reinforced plastic
composites subjected to impact loading,” International Journal of Impact Engineering
75, 40–52.
Yen, C. F. [2002] “Ballistic impact modeling of composite structures,” Proceedings of the
7th LS-DYNA Users Conference, pp. 6-15–6-26.
Yen, C. F. [2012] “A ballistic material model for continuous-fiber reinforced composites,”
International Journal of Impact Engineering 46, 11–22.
Zhao, G., Wang, Z., Zhang, J. and Cho, C. [2007] “An experimental investigation of pen-
etration failure modes in composite laminates,” Proceedings of the 16th International
Conference on Composite Materials, Kyoto, Japan, July 8–13.

1550039-23

View publication stats

You might also like