Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 173 (2017) 190 – 197

11th International Symposium on Plasticity and Impact Mechanics, Implast 2016

Simulation and Analysis of Ballistic Impact using Continuum

Damage Mechanics (CDM) Model
Manoj Kumara*, Utkarsh Deepb, P M Dixtc
a
Research Scholar, Department of Mechanical Engineering, Indian Institute of Technology Kanpur, India
b
Dual Student, Department of Mechanical Engineering, Indian Institute of Technology Kanpur, India
c
Professor,Department of Mechanical Engineering, Indian Institute of Technology Kanpur, India

Abstract

Ballistic Impact is the study of behaviour of ductile targets subjected to projectile impact. In this work, simulation and
analysis is done by using the Continuum Damage Mechanics (CDM) model on ballistic impact of steel (IS2062: 2006
GR E410W A) target using a rigid projectile. The Continuum Damage Mechanics model is implemented in
ABAQUS/Explicit through a user defined material model subroutine (VUMAT). The developed VUMAT is validated
by comparing the simulated results in tensile test with experimental results. The effect of the temperature and the strain
rate are included in the hardening model as well as in the damage growth model. A critical damage criteria is used to
predict the fracture initiation and spread of fracture is simulated by tracing the path of failed elements.
©2017
© 2016TheTheAuthors.
Authors. Published
Published by Elsevier
by Elsevier Ltd.is an open access article under the CC BY-NC-ND license
Ltd. This
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility ofthe organizing committee of Implast 2016.
Peer-review under responsibility of the organizing committee of Implast 2016
Keywords: Dynamic Fracture; Continuum Damage Mechanics; Ballistic Impact.

1. Introduction

Ballistic impact is one of the special case of impact problems. In such problems, a high force is applied over a short
time period (called shock loading) when two or more bodies collide. Such a force usually has a greater e
effect than a lower force applied over a proportionally longer period. This fact makes the study of ballistic impact an
interesting and also an important subject with respect to the safe design of structures. Ballistic impact refers to the
phenomenon of a high velocity impact of a small mass.

* Manoj Kumar. Tel.: (+91) 512 259-7627; fax: (+91) 512 259-7408.
E-mail address:mkumar@iitk.ac.in

1877-7058 © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of Implast 2016
doi:10.1016/j.proeng.2016.12.057
 Manoj Kumar et al. / Procedia Engineering 173 (2017) 190 – 197 191

In this problem, a short cylindrical object with a flat/ogive/hemi-spherical tip (called projectile) impacts against a
thin large plate (called target). The phenomenon has been very well researched both numerically as well as
experimentally. The experimental study on ballistic impact started long time back, mostly conducted on metals or a
layered combination. With the advancement in technology, there is a shift in the material used from metals like steel
and aluminium to composite material like kevlar and the likes of it, owing to its higher strength to weight ratio. With
the advent of computers, software packages like ABAQUS, LS-DYNA have surfaced which are helpful in solving
wide range of engineering problems. With the help of powerful computational tools, we can easily simulate the ballistic
impact, which would have been a lot costly if we were to carry out experiments.
The maximum velocity at which a projectile fails to penetrate the target is commonly referred to as the ballistic
limit of the target and projectile. At low impact velocity (Group [1]), the penetration mechanism is more dominated
by the dynamic deformation mechanism within the projectile and target. However, at higher velocities, hydrodynamic
effect dominates and the penetration is controlled by the density of the projectile and target.
Ballistic impact has been studied extensively both experimentally and numerically. Most of the experimental studies
have been on hard steel or synthetic fiber like Kevlar, and in some case both of them sandwiched together. But due to
complexity and the cost of carrying out the ballistic impact experiments, it is necessary to have a general solution
technique as supplement to high-precision testing. Numerical simulation using packages like ABAQUS, LS-DYNA,
has made it quite simple to obtain a good enough solution.
Numerous attempts have been made to study ballistic impact experimentally. Golsdmith and Finnegan [2] studied
the penetration, perforation and fragmentation of 0.05-0.25 inch thick 2024 aluminium and SAE 1020 and 4130 steel
alloy plates impacted normally by AISI 52-100 hard steel spheres at velocities from 500 to 8800 ft/sec. The drop in
projectile velocity was found to decrease just above the ballistic limit, then steadily rise with increasing the sphere
speed. Gupta and Madhu [3] conducted a series of experiments on 10-25 mm thick mild steel targeted by spinning
hard core projectile fired at a velocity of about 820m/s. The angle of impact is kept increasing until ricochet occurred.
Velocity drop versus the angle of obliquity curves are presented. Dikshit et al [4] conducted experiments on the
penetration of steel plates of varying hardness in the range HV295-HV520 impacted normally by ogive shaped
projectile with velocity in the range of 300-800m/s. The data and observations relating to the plastic zone size and
shape surrounding the penetrating projectile and the extent of adiabatic shear band (ASB) formation and plate spalling
have also been presented.
Gupta and Madhu et al [5] carried out an experimental study on perforation of single and layered plates of mild
steel, RHA steel and aluminium targeted by hard-core projectile fired at 800-880m/s. Determination of plate thickness
at the ballistic limit is carried out. Borvik et al [6] conducted an experimental study on 12mm thick Weldox 460 E
steel plates impacted by blunt-shaped projectiles in the lower ordnance velocity regime. Numerical validation was
carried out using non-linear dynamic explicit finite element code LS-DYNA implementing the viscoplastic behavior
of the steel and the Johnson-Cook fracture model. Projectiles of 3 different shapes (blunt, hemispherical and conical)
have been used by Borvik et al [7] in gas gun experiments, to penetrate 12mm thick Weldox 460E steel plates.
Conclusions about the dependence of the failure mode on the nose profile of projectile is established. Borvik et al [8]
proposed the use of lightweight material like aluminium, as protective structures, by studying the ballistic impact of
AA5083-H116 aluminium plates using conical-nosed hardened steel projectiles. The values of ballistic limit from
experiments were found to be in good agreement with the one's found from analytical perforation model in open
literature.
Iqbal et al [9] used a pneumatic gun to fire ogive-nosed projectiles at velocities above the ballistic limit on 1mm
thick aluminium plates to study the deformation of plates. Numerical prediction using ABAQUS 6.3 code is done and
seems to be in good agreement. Gupta et al [10] did the experimental and numerical investigation of the effect of the
parameters like the shape of projectile (blunt, ogive, hemispherical), velocity of impact and the target thickness on the
deformation behaviour. ABAQUS finite element code was utilized to obtain numerical results which were in good
agreement with the experimental results.

It is evident from the literature survey that a lot of effort has been put to study the ballistic impact. In most of the
numerical simulations, the Johnson-Cook fracture initiation criterion is used. However, it is an empirical criterion.
Now, it is well established that ductile fracture occurs due to micro-void nucleation, then growth and finally
coalescence into micro-cracks. There is a need to simulate the fracture in ballistic impact problems using a fracture
192 Manoj Kumar et al. / Procedia Engineering 173 (2017) 190 – 197

initiation criterion based on the above micro-structural observations. Lemaitre's continuum damage [11] mechanics
model provides such a fracture initiation criterion. In this model, an internal variable called damage is introduced that
represents the void density at a point. The critical value of damage is used as a fracture initiation criterion. The
objective of the present work is to employ the Lemaitre's Continuum Damage Mechanics [11] model to study the
ballistic process. The damage growth law proposed by Kumar and Dixit [12] will be employed. The critical damage
value will be used for the element deletion. Effects of process parameters on the velocity, the stress, strain, strain rate,
damage and temperature fields will be studied. The process parameters considered are the impact velocity, projectile
shape, friction at the interface and the target thickness..

1.1 Continuum Damage Mechanics, Elasto-plastic Constitution Equation and Damage Growth Law

The damage variable (D) at a point, when assumed isotropic, is defined as the area void fraction in a plane passing
through the point. That is
§ 'A ·
D lim ¨ v ¸ (1)
'Ao0 © 'A ¹

where 'A is an infinitesimal area around the point in some plane and 'Av is the area of voids traces in that plane
contained in 'A .
For a damaged material, the plastic potential is given by (Lemaitre, [9])
F   
F1 V ij , H eqp , D  FD Y , H eqp , D  (2)
where F1 and FD are the plastic potential associated with yielding and hardening and damage respectively. In
equation (2), (-Y) is the dissipative part of the thermodynamic force (-Y) corresponding to the damage rate. It can be
shown that (Lemaitre, 1985) (-Y) is equal to the strain energy release rate (at constant stress) due to damage. Then, (-
Y) becomes the work-conjugate variable corresponding to D that is given by.
V eq 2 ª § Vm · º
2
« 2
Y 1 Q   3 1  2Q  ¨¨ ¸¸ »» (3)
2 E 1  D  « 3 V eq ¹
2
¬ © ¼
Here, E is the Young's modulus, ν is the Poisson's ratio, V m the mean part of the Cauchy stress tensor V ij and V eq is
the equivalent stress related to the deviatoric part V ij ' of V ij . The quantities V m and V eq are given by the relations:
1
Vm V ii (4)
3
3 ' '
V eq V ij V ij (5)
2
 
The ratio V m V eq is called the Triaxiality. Further, in equation (2), H eqp is the equivalent plastic strain (used as the
hardening parameter for the case of strain hardening) that is defined by
2 p p
H eqp ³ dH
p
eq ; d H eqp d H ij d H ij (6)
3
where, d H eqp is called the equivalent plastic strain increment and d H ijp is the plastic part of the incremental linear strain
tensor:

d H ij
1
2
 dui, j  du j ,i  (7)

Here, dui is the incremental displacement vector and the comma denotes the derivative with respect to a component of
the position vector.

The principle of strain equivalent states that the deformation behaviour of a damaged material can be represented by
 Manoj Kumar et al. / Procedia Engineering 173 (2017) 190 – 197 193

the constitutive law of the virgin material (i.e., defects free material) if the Cauchy stress is replaced by the effective
Cauchy stress. The effective Cauchy stress tensor ( V ij* ) is defined as
V ij
V ij* (8)
1  D 
It is assumed that F1 (i.e., the plastic potential associated with yielding and hardening) is given by the von Mises yield
function. Modifying it for a damaged material by using the principle of strain equivalence, using the associated flow
rule for obtaining the incremental plastic stress-strain relation and adding to it the incremental elastic stress-strain
relation, the incremental elasto-plastic stress-strain relationship becomes
dV ij Cijkl
EP
d H klp (9)
where the forth order elasto-plastic constitution tensor is given by
­° Q 9P V ij' V kl' ½°
2P ®G ik G jl  G ijG kl  ¾ 1  D 
EP
Cijkl (10)
°¯ 1  2Q 2  H '  3P  V eq2 °
¿
Here, P is the shear modulus and H ' is the slope of the hardening curve, which is modelled by the Ludwik’s power
law:
Since, Taylor Rod Impact problem is a high strain rate and high temperature process, the effects of
temperature softening and visco-plasticity (i.e. strain rate hardening) on variable yield stress cannot be ignored.
Johnson and Cook [16] extended the Ludwik’s power law and defined variable yield stress σy as:

V y { H H eqp , T , H eqp  V   K H   1  T 1  C ln H 
y o
p n
eq
*m *

V   K H    f (T )  g H 
(11)
p n p
y o eq eq

where V Y 0 is the initial yield stress and K and n are the hardening parameters. The incremental stress in equation (9)
has to be objective. One of the commonly used incremental objective stress tensor is the product of the Jaumann stress
rate tensor and incremental time.
where C and m are material constants and T* is dimensionless temperature defined as
T  TR
T* (12)
TM  TR
Here TM and TR are the melting and room temperatures respectively. The quantity H * is the dimensionless equivalent
plastic strain rate defined by
H eqp
H* (13)
H0
Here H0 is the critical strain rate.

Based on the experiments conducted at IIT Kanpur on IS2062: 2006 GR E410W A steel, the following expression is
proposed for FD (i.e., the potential associated with damage):

exp ^b0  Y `
a0
FD (14)
b0 (1  D)
where a0 and b0 are the material constants that are determined from experiments. As stated earlier, the incremental
damage growth law is obtained as the derivative of FD with respect to ( Y ):
wFD p
dD = (1  D) dεeq (15)
w (-Y )
194 Manoj Kumar et al. / Procedia Engineering 173 (2017) 190 – 197

Then, using the expression (15) for FD , the incremental damage growth becomes
dD a0 exp ^b0  Y ` d H eq
p
(16)
Values of the material constants a0 and b0 , as found from the experiments, are: a0 = 0.0045 and b0 = 2.62MPa-1.
Equation (16) states that the damage growth depends on two continuum parameters: (i) triaxiality through (-Y)
(equation 3) and (ii) equivalent plastic strain.
The damage critical is experimentally found for the steel rod and it is depended on temperature and strain rate as

Dcr (H * , T ) ^
0.42 1  0.76T *  6.25(T * )2 1  0.021ln(H * )  ` (17)

1.2 Cut-Off Value of Triaxiality

It is clear from the damage growth law (equation 16) that the damage evolution is greatly dependent on triaxiality.
It has been observed during various experiments that fracture does not occur under high compressive loads. Bridgman
[10] conducted numerous tensile tests on round specimens under different hydrostatic pressures and observed that
fracture does not occur at considerably high value of pressures. This clearly indicates that there must be a cut-off value
on negative triaxiality below which fracture does not occur. Based on the observations of Bridgman[10], Bao and
Weirzbicki[6] conducted various experiments to determine this cut-off value of triaxiality by doing upsetting and
tensile tests at various hydrostatic pressures, and reported that fracture did not initiate for the triaxiality value less than
(-1/3). Therefore, in the present work, damage increment is considered zero for triaxiality values less than (-1/3). Thus,
Vm ½
dD 0, if d (1/ 3) °
V eq
°
s ¾ (18)
­° V eq 2 § V · ½° Vm °
dD ® f ¨ m ¸ ¾ d H eqp , if ! (1/ 3) °
¨V ¸ V eq
¯° 2 ES © eq ¹ ¿° ¿

1.3 Criteria for Element Deletion

A critical damage criteria is used as the criteria for element deletion. Whenever the damage in an element reaches
a critical value (Dc), the element is deleted. The criteria is implemented over the whole domain of the workpiece. The
value of critical damage chosen here for the simulation of Taylor Rod impact is slightly higher than that of equation
(17), so as to incorporate a small delay in the element deletion after the damage reaches its actual critical value.

2. Ballistic Impact Simulation

The simulation of ballistic impact is carried out using the commercial software package ABAQUS/Explicit. The
problem is modelled as a 3D problem. The initial geometry of the plate (target) of thickness 12mm is taken. The overall
dimension of the plate is 500 mm x 500 mm. To reduce the computational time, only half the plate is considered i.e.,
500 mm x 250 mm (Fig. 1(a)). The meshing is done keeping in mind the fact that large dynamic deformation takes
place near the contact region. To implement this, ne meshing is done near the impact zone and coarse in the rest of the
domain. Fine region is meshed using a total of 9,100 C3D8RT (An 8-node thermally coupled brick, trilinear
displacement and temperature, reduced integration, hourglass control) element and having a total of 10,290 nodes. The
rest of the domain is meshed using a total of 62,480 C3D8RT elements and having a total 68,334 nodes. In total, there
are 71,580 C3D8RT elements and 78,036 nodes. As for the flat projectile, it is modeled as a discretely rigid body with
a reference point. In this study, Fig. 1(b) shows the meshing for half the projectile (Rigid) with flat nose.
 Manoj Kumar et al. / Procedia Engineering 173 (2017) 190 – 197 195

Fig. 1(a) : The finite element mesh of the domain Fig 1(b): Meshing of half projectile with flat head

The high velocity impact process is usually assumed to be adiabatic, because heat generated by plastic work in the
localized region does not have sufficient time to be conducted to the surrounding material. The increase in local
temperature T owing to plastic work is given by
E
³ V c V ij dHij
p
T (19)

here, E is the fraction of plastic work converted to heat. It is generally accepted that (Bai and Dodd [13]). The
material behaviour in the elastic region is linear and in the plastic region, it is governed by the associated flow rule
based on the Mises Plastic Potential. The hardening is assumed to be governed by the Johnson-Cook (J-C) hardening
model given by equation (11). The damage growth is governed by the damage growth law of equation (16). The
coefficients of J-C model and damage model for steel, used in simulation are given in tables (1) and (2) respectively.

Table 1: Material Properties of IS2062: 2006 GR E410W A steel

Young’s modulus, E (GPA) 

200 1  0.2927  T *  0.3796  T *
2

Poisson’s ratio, v 0.3
Thermal conductivity, k (W/m oC) 48.3  0.023T
Specific heat, c (J/kgoC) 420  0.504T
Thermal expansion coefficient, α ( oC-1) 1.1 x 10-5
Density, ρ (kg/m3) 7862
2
Convective heat transfer coefficient, h (W/m K) 30
Faction of plastic work converted to heat (β) 0.90
Coefficient of friction 0.05

Table 2 :Coefficients of Johnson - Cook hardening model

V 
y o (MPa) K (MPa) n m TM (oC) TR (oC) C
x
H 0 (sec-1)
220 579 0.431 1 1300 25 0.014 1
196 Manoj Kumar et al. / Procedia Engineering 173 (2017) 190 – 197

2.1 Typical Ballistic Impact Simulation Results

The ballistic impact simulation is done to study the distributions of damage at a Projectile Velocity 350 m/s.

Fig. 2: Distribution of damage at three different instances

Fig. 3: Projectile velocity as a function of depth of penetration

 Manoj Kumar et al. / Procedia Engineering 173 (2017) 190 – 197 197

3. Conclusions

The analysis of damage in the Taylor Rod impact test is carried out for IS2062: 2006 GR E410W A steel by
using the continuum damage mechanics model and experimentally obtained damage growth law. The analysis is
done using the ABAQUS/Explicit FE Package and the damage growth law is implemented through a user defined
material model subroutine (VUMAT). The objective of the present work is to study the ballistic impact process
using the Lemaitre's Continuum Damage Mechanics (CDM) model. Based on the results of this study, the
following conclusions can be drawn: Equivalent stress, equivalent plastic strain, damage, triaxiality, temperature
and strain rate are highest in the region of contact between the projectile and the target. Strain rates are of the
order of 105/sec in a few select pockets of region near the contact. Projectile velocity drops rapidly during the first
few millimetres of the penetration of the projectile.

References

[1] Group, W. R. Ballistic impact mechanisms of materials. http://www.virginia.edu/ms/research/wadley/ballistic-impact.html, 2014.

[2] Golsdmith, W., and Finnegan, S. Penetration and perforation processes in metal targets at and above ballistic velocities. International Journal
of Mechanical Sciences 13 (1971), 843-866.
[3] Gupta, N., and Madhu, V. Normal and oblique impact of a kinetic energy projectile on mild steel plates. International Journal of Impact
Engineering 12, 3 (1992),333-343.
[4] Dikshit, S. N., Kutumbarao, V., and Sundararajan, G. The influence of plate hardness on the ballistic penetration of the thick steel plates.
International Journal of Impact Engineering 16, 2 (1995), 293-320.
[5] Gupta, N., and Madhu, V. An experimental study of normal and oblique impact of hard-core projectiles on single and layered plates.
International Journal of Impact Engineering 19, 5-6 (1997), 395-414.
[6] Borvik, T., Langseth, M., Hopperstad, O., and Malo, K. Ballistic penetration of steel plates. International Journal of Impact Engineering 22
(1999), 855-886.
[7] Borvik, T., Langseth, M., Hopperstad, O., and Malo, K. Perforation of 12mm thick steel plates by 20mm diameter projectiles with flat,
hemispherical and conical noses. part i : Experimental study. International Journal of Impact Engineering 27 (2002), 19-35.
[8] Borvik, T., Clausen, A. H., Hopperstad, O. S., and Langseth, M. Perforation of aa5083-h116 aluminium plates with conical-nosed projectile-
experimental study. International Journal of Impact Engineering 30 (2004), 367-384.
[9] Iqbal, M., Gupta, N., and Sekhon, G. Behavior of thin aluminium plates subjected to impact by ogive-nosed projectiles. Defence Science
Journal 56, 5 (2006), 841-852.
[10] Gupta, N., Iqbal, M., and Sekhon, G. Effect of projectile nose shape, impact velocity and target thickness on deformation behavior of
aluminium plates. International Journal of Solids and Structures 44 (2007), 3411-3439.
[11] Lemaitre, J. How to use damage mechanics. Nuclear Engineering and Design, 80(1984):233–245.
[12] Kumar M. and Dixit, P M, A Non-linear Ductile Damage Growth Law. The International Journal of Damage Mechanics 24 (2015) 1070-
1085.
[13] Bai, Y. and Dodd, B. Adiabatic Shear Localization: Occurrence, Theories and Applications. Pergamon Press, New York, 1992..
[14] ABAQUS, 2010. ABAQUS Version 6.10 Documentation, ABAQUS, Inc. and Dassault Systèmes.