Thin–Walled Structures 154 (2020) 106784

Contents lists available at ScienceDirect

Thin-Walled Structures
journal homepage: http://www.elsevier.com/locate/tws

An investigation of the ballistic performance of independent ceramic target

M.K. Khan a, M.A. Iqbal a, *, V. Bratov b, c, N.F. Morozov b, c, N.K. Gupta d
a
Department of Civil Engineering, Indian Institute of Technology Roorkee, Roorkee, 247667, India
b
Institute for Problems in Mechanical Engineering, Russian Academy of Sciences, Saint Petersburg, Russia
c
Department of Theory of Elasticity, Saint Petersburg State University, Saint Petersburg, Russia
d
Department of Applied Mechanics, Indian Institute of Technology Delhi, New Delhi-110016, India

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

Keywords: The ballistic experiments have been carried out on alumina 99.5% bare ceramic plates for studying the initiation
Ballistic resistance and propagation of brittle fracture and the resistance offered by the target. The ceramic tiles of size, 100 mm �
Bare ceramic target 100 mm, and thickness, 5 mm, were impacted by the ogival nosed hardened steel 4340 projectiles of diameter
Fragmentation
10.9 mm and mass 30 grams at velocities in the range 52–275 m/s. The ceramic fragments were carefully
Alumina 99.5%
Steel 4340
collected to examine the cracking patterns at the front and the back surfaces. The cracks developed in the target
were studied extensively to develop more insight into the fracture mechanism. The energy absorbed by the target
has been studied and correlated with respect to the fracture mechanism of the target. Numerical simulations have
been performed on a commercial finite element code and the experimental findings have been reproduced in
order to further understand the fracture and fragmentation behaviour and its influence thereon the ballistic
characteristics of the target. The Johnson-Holmquist-2 (JH-2) constitutive model has been used for simulating
the material behaviour of ceramic and the Johnson-Cook (JC) elasto-viscoplastic material model for simulating
the behaviour of the steel projectile. The behaviour of ceramic target under oblique impact was explored
numerically. The damage in the projectile was found to be higher in case of oblique impact. Both the experi­
mental and numerical findings have described an increase in the crack density with the increase in the incidence
velocity of the projectile. The average size of the fragments has also been found to be reduced with the increase in
the projectile incidence velocity.

1. Introduction The ductile backing thus serves as the reinforcement imparting ductility
and tensile strength to the protective system [1,2]. The ballistic per­
The primary requirements for an armour material are low density, formance of ceramic has been investigated in the available literature
high strength and high ductility. Ceramics have emerged as one of the considering it as a front layer in a bi-layer armour. The available studies
most suitable armour materials due to their low density, high on bi-layer ceramic could be categorized based on the backing layer
compressive strength and high hardness. A relatively low fracture material as metallic bi-layer armour [3–16] and composite bi-layer ar­
toughness and less tensile strength however ceases the possibility of mour [17–21]. The composite armor is meant to replace the traditional
using the ceramic as an independent armor structure. However, in spite metallic armor as it reduces the weight of the body armor for facilitating
of the associated limitations responsible for a premature fracture, the the movement of the personnel. The light weight armour when used in
ceramic when used as the front layer of a composite armour with a military vehicles reduces the fuel consumption and ensures the safety of
ductile backing is considered to be most effective in defeating the armor the vehicle by reducing the load on the underlying weaker terrain. A
piercing projectile. The function of ceramic layer is to shatter or blunt significant pre-eminence of ceramic based composite armour has been
the projectile and thus restrict the perforation process. Although, the noticed [22] over conventional single layer metallic armor at velocities
ceramic front layer is completely fractured during this process, the higher than 250 m/s. The design of composite armor involves deter­
backing layer helps in keeping the comminuted ceramic in position for mining the optimum thickness ratios of ceramic and backing plate for
further energy dissipation. The remaining kinetic energy of projectile is obtaining the best protection level at minimum weight. The armours
subsequently absorbed in the plastic deformation of the backing plate. made-up of ceramic and aluminium plates of consistent properties but

* Corresponding author.
E-mail addresses: [email protected], [email protected] (M.A. Iqbal).

https://doi.org/10.1016/j.tws.2020.106784
Received 27 January 2020; Received in revised form 1 May 2020; Accepted 12 May 2020
Available online 6 June 2020
0263-8231/© 2020 Elsevier Ltd. All rights reserved.
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Fig. 1. (a) A schematic top view of pneumatic gun setup (b) Target holding frame with (1) Front steel plate (2) Rear steel plate (3) Slot for holding the target (4) Side
thick glass panel (5) Top thick glass panel (6) Thin fabric cover.

2
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Fig. 2. (a) The schematic of the projectile (b) Ceramic target held in steel frame.

different thickness ratios have been seen to offer best performance significantly more damage than that observed in the bare plates. The
against 7.62 mm AP projectiles when the alumina/aluminium thickness numerical simulations performed on LS-DYNA code successfully repro­
ratios were more than or equivalent to unity [3]. The theoretically op­ duced the cracking pattern and the damage evolution in ceramic
timum ratio of 2.5 has been suggested to provide an armour with near explaining the usefulness of the restraining effect. An analytical model
optimum performance. A combined experimental and numerical study has been developed for predicting the behaviour of ceramic tiles without
performed by Lee and Yoo [8] on alumina/aluminium composite target any backing plate at the useful range of incidence velocities, including
against 7.62 mm AP projectiles described no significant difference in the below the Bernoulli limit [2]. The novel algorithm of the proposed
performance of the target for the thickness ratios between 1.5 and 3. model enables the erosion of the projectile tip when it is in contact with
However, the optimum thickness ratio was identified as 2.5. the target at the zero level of penetration. Thus, the capability of the
The studies focussing primarily on the fracture and fragmentation of model to reproduce the impact event at relatively low incidence veloc­
the ceramic are limited in the open literature. The long rod penetration ities is a fundamental step towards an optimal design of the ceramic
mechanism of steel encased ceramic (A12O3, SiC, B4C and TiB2) target at target with minimum thickness of the backing layer.
a range of incidence velocities (0.8–1.4 km/s) described that the The ballistic investigation of an independent ceramic target, there­
crushing of a small volume of ceramic in the vicinity of impact zone and fore, explains the mechanics of the evolution and propagation of damage
the subsequent flow of small fragments (in the lateral direction) which certainly influence its capability to defeat the projectile in a
constituted the major defeating target mechanism [23]. When an ogival multilayer configuration and also improves understanding about the
nosed projectile strikes the hard surface of the ceramic, its nose is computational modelling of its structural behaviour.
blunted and the resultant effective cross-sectional area is increased, A detailed experimental and numerical investigation has been re­
causing the reduction in the compressive stress in the ceramic. The ported in this manuscript on the fragmentation and ballistic evaluation
square nosed striker was seen to be more effective in penetrating the of alumina 99.5% independent ceramic target against 30 grams ogival
ceramic as less energy was dissipated in the deformation of the striker. nosed hardened steel projectile at normal incidence velocities in the
When the hardness of the striker material was more than that of the range, 52–275 m/s. The ceramic tile of planner dimensions, 100 mm �
ceramic, the penetration capability of ogival nosed striker is seen to have 100 mm, and thickness, 5 mm, was impacted by the projectile close to
recovered subsequent to the blunting process [4]. The failure mecha­ and above the ballistic limit velocities. The incidence and the residual
nism of the AD95 ceramic confined with 4340 steel under different projectile velocities were measured through high speed videography.
conditions of confinement (front, side and back) described that the re­ The damage induced in the target was quantified based on the number
straint provided against the ejection of fragments by introducing a front and the size of the cracks. A detailed numerical investigation has been
cover plate could substantially improve the ballistic performance in carried out on ABAQUS/Explicit code to further explore the propagation
comparison to that of the restraint provided against the radial expansion of damage in the ceramic target. Simulations accurately reproduced the
[10]. residual projectile velocities and the magnitude of damage in the target
The studies about the ballistic performance of ceramic based bi-layer as well as the projectile. The numerical investigation was subsequently
armour are indeed important in respect of the functional utility of the carried out to understand the performance of the ceramic against obli­
ceramic, however, in order to understand and improve the contribution que impact.
of ceramic in a bi-layer armour a further insight is required to be
developed about the ballistic properties of the bare ceramic. The 2. Experiments
development and propagation of damage in alumina ceramic with and
without a thin fiber-glass wrapping was studied by Rahbek et al. [24] at The ballistic experiments have been carried out in the Impact Lab­
below perforation velocities against M61 7.62 mm AP projectiles. The oratory at IIT Roorkee for studying the fragmentation behaviour of 5
experimental results demonstrated that the covered plates experienced mm thick alumina 99.5% targets against ogival nosed steel projectiles.

3
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Fig. 3. Penetration process recorded by high speed camera at a velocity of 275 m/s at (a) 0, (b) 400, (c) 650 and (d) 900 μs

from the vessel with the help of a pneumatic actuator. A robust target
holding setup consisting of two thick steel plates {marked 1 and 2 in
Fig. 1(b)} supported on a rigid steel frame was designed. The purpose of
the front steel plate (1) was to arrest the fragments ejected from the front
surface of the target and it was provided just at the muzzle end with a
central hole of diameter, 50 mm. The rear steel plate (2) provided at a
distance of 200 mm from the front plate (1) had a slot for holding the
target {marked 3 in Fig. 1(b)} of span, 100 mm � 100 mm. After
inserting the ceramic tile in the slot, the edges of the plate were tight­
ened for arresting the movement of plate and maintaining fixity at the
supports. The clear span of the tile was 95 mm � 95 mm. The two other
sides and the top of the steel frame were covered with the thick glass
panels. The glass panels providing a transparent shield between the two
steel plates helped in recovering the fragments ejected from the front
surface of the target without obstructing the view for the high speed
camera. The space behind the rear steel plate was also covered by fixing
the side as well as the top glass panels onto the rigid steel frame. The
transparent glass panels facilitated the recovery of the fragments ejected
from the rear surface of the ceramic target without obstructing the view
for the high speed video recording of the projectile perforation {marked
as 4 and 5 in Fig. 1(b)}. The far end of the steel frame was covered with a
very thin fabric to avoid any possibility of missing the ejected fragments.
Two Phantom V411 high-speed cameras were installed to record the
perforation phenomenon and to measure the projectile velocity. One of
the high speed video camera (Camera-1) was placed normal to the line of
sight of the projectile in front of the target holding (rear) steel plate to
record the perforation phenomenon and to measure the incidence and
residual velocities, see Fig. 1 (a). The second camera (Camera-2) was
focused at the back surface of the target to record the damage propa­
gation and the ejection of the fragments from the target. The frame rate
Fig. 4. Meshing of projectile and target. of both the high speed cameras was maintained between 40,000–50,000
per second with a resolution of 500 � 150 for Camera-1 and 256 � 256
The total of nine ballistic experiments were conducted by carefully for the Camera-2.
varying the incidence velocity of the projectile and the response of the The schematic of the projectile is shown in Fig. 2 (a). The diameter of
target corresponding to each incidence velocity was studied. The pro­ the projectile shank was 10.9 mm, total length 52.6 mm, and mass 30 g.
jectiles were fired with the help of a pneumatic gun shown in Fig. 1 (a). The projectile was heated for 2 hours at a constant temperature, 820 C,
�

The pneumatic gun had a pressure vessel of 20-litre capacity connected in a high-temperature furnace and oil quenched subsequently for
to a 5 m long smooth steel barrel of inside diameter, 11.9 mm. The obtaining the hardness, 52 HRC.
projectile was accelerated by instantaneously releasing the air pressure The nose of the projectile was damaged after hitting the target in all

4
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Fig. 5. (a) JH-2 Strength model (b) JH-2 Damage model (c) JH-2 Pressure model [28].

the experiments. Therefore, for each experiment a new projectile was trials of numerical simulations and based on the numerical studies
employed. The projectile was collected after the experiment from a available in the literature [12]. The remaining portion of the target plate
robust steel catcher box filled with the cotton rag. The alumina ceramic was meshed with continuum three dimensional eight node reduced
tile had its dimensions, 100 mm � 100 mm, and the thickness 5 mm, see integration (C3D8R) elements of size 0.8 mm � 0.8 mm � 0.8 mm. This
Fig. 2 (b). The initial velocity of the projectile was varied between 52 configuration of elements resulted in a total number of 4,11,892 ele­
and 275 m/s. The different stages of a typical penetration process ments in the whole target of span 95 mm � 95 mm, and thickness, 5 mm.
recorded by a high speed camera at an incidence velocity of 275 m/s are The linear C3D8R elements of size 0.4 mm � 0.4 mm � 0.4 mm was
shown in Fig. 3. The fragments ejected from the rear surface of the target considered to discretize the body of the projectile with a total number of
had very high density in comparison with the fragments ejected from the elements, 84,440. The finite element model of the target and projectile is
front surface of the target. shown in Fig. 4.
The interaction between the projectile and the central portion of the
3. Numerical simulation target was modeled using surface to surface contact with kinematic
contact algorithm assuming negligible friction between the projectile
A 3D numerical simulation model was developed using Lagrangian and the target. This assumption was considered based on the high
finite elements in ABAQUS/Explicit finite element code. Both the target incidence velocity of the projectile, small target thickness and the brittle
and the projectiles were modeled as the deformable continuum. The material of the target. The target plate was restrained against all degrees
central portion of the target of dimension 47.5 mm � 47.5 mm was of freedom at its periphery.
meshed with continuum three dimensional eight node reduced inte­
gration brick elements (C3D8R) of size 0.4 mm � 0.4 mm � 0.4 mm.
This is an optimum size that was identified after performing a number of

5
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Table 1 ceramic. The polynomial equation of state (EOS) evaluated the current
JH-2 constitutive model parameters for alumina 99.5% [19]. state of pressure as a function of volumetric change, the specific strength
Material parameters Numerical values model provided the equivalent strength for both intact and fractured
3 material, and the damage model was responsible for the transition of
Density (kg/m ) 3700
material from an intact to a fractured state. The graphical representation
EOS Polynomial of the JH-2 model is provided in Fig. 5.
Bulk modulus, K1 (GPa) 130.95
Pressure constant, K2 (GPa) 0
The normalized equivalent stress is defined as;
Pressure constant, K3 (GPa) 0 � �
σ * ¼ σ *i D σ *i σ *f (1)
Strength model JH-2
Shear modulus, G (GPa) 90.16
Hugoniot elastic limit (HEL) (GPa) 19 where σ*i is the normalized intact equivalent stress, and σ *f , the
Intact strength constant, A 0.93
normalized fracture stress, and D is the damage (0 � D � 1.0).
Intact strength exponent, N 0.6
Strain rate constant, C 0 The normalized intact equivalent stress is given in equation (2) and
Fracture strength constant, B 0.31 the normalized fractured equivalent stress in equation (3);
Fracture strength exponent, M 0.6
Normalized maximum fractured strength, σ*fmax 0.2 σ *i ¼ AðP* þ T * ÞN ð1 þ Cln_ε* Þ (2)
Pressure at HEL (GPa) 1.46
σ *f ¼ BðP* ÞM ð1 þ Cln_ε* Þ (3)
Failure model JH-2
Damage constant, D1 0.005
Damage exponent, D2 1 where A, B, C, M and N are material constants and the normalized
Bulking factor, β 1 pressure P* ¼ P=PHEL where P is the actual pressure and PHEL is the
pressure at Hugoniot elastic limit (HEL). The HEL is the net compressive
stress corresponding to the uniaxial strain (shock wave) exceeding the
Table 2 elastic limit of the material. The normalized maximum tensile hydro­
Johnson-Cook model parameters Steel 4340 [12]. static pressure is T * ¼ T=PHEL , where T is the maximum tensile hydro­
Material parameters Numerical values static pressure the material can withstand and the dimensionless strain
Density, ρ0 (kg/m ) 3
7770 rate, ε_ * ¼ ε_ =_ε0 , where ε_ is actual equivalent strain rate and ε_ 0 is refer­
ence strain rate considered as 1 s 1.
EOS Polynomial
Bulk modulus, K1 (GPa) 159 The hydrostatic pressure before and after the damage was expressed
Specific heat, Cr (J/KgK) 477 through equations (5) and (6), respectively;
Strength model JC P ¼ K1 μ þ K2 μ2 þ K3 μ3 (4)
Shear modulus, G (GPa) 77
Static yield strength, A (GPa) 0.95
Strain hardening constant, B (GPa) 0.725 P ¼ K1 μ þ K2 μ2 þ K3 μ3 þ ΔP (5)
Strain hardening exponent, n 0.375
Strain rate constant, C 0.015 where K1 is a bulk modulus and K2, K3 are the pressure constants. The
Thermal softening exponent, m 0.625 volumetric strain μ ¼ ρ=ρ0 1 where ρ is current density and ρ0 is the
Melting temperature, tm (K) 1793
initial density of the material. ΔP is pressure increment due to bulking of
Reference strain rate, ε_ 0 1
the material due to accumulation of damage.
Failure model JC The damage criterion for the fracture model is same as is used in the
Damage constant, D1 0.8 Johnson-Cook [26,27] model;
Damage constant, D2 2.1
Damage constant, D3 0.5 X Δε
Damage constant, D4 0.002 D¼ (6)
εfp
Damage constant, D5 0.61

f
where Δε is the increment in equivalent plastic strain and εp is the plastic
strain to fracture
Table 3
The plastic strain to fracture of the material is described as;
Comparison of actual and predicted residual projectile velocities.
S. No. Impact velocity (m/s) Residual velocity (m/s) εfp ¼ D1 ðP* þ T * ÞD2 (7)
Experiments Simulations
where D1 and D2 are the damage constants. The damage parameter D is
1 275 239 212
same as is explained in the Johnson Cook model [26,27]. The details of
2 264 215 201
3 249 182 190 the experiments and calculations required for the calibration of the
4 215 153 151 material parameters of the JH-2 model have been discussed in Holm­
5 179 116 126 quist et al. [28]. The parameters of the JH-2 model for alumina 99.5%
6 112 70 45
used in the present study have been directly obtained from Feli and
7 94 58 23
8 62 31 0
Asgari [19], and these have been presented in Table 1.
9 52 26 0 In order to implement the JH-2 constitutive model through VUMAT
program in Abaqus/Explicit code using the built-in user material, an
element erosion strain (FS) is provided to remove the severely distorted
3.1. Johnson and Holmquist-2 model elements from the mesh. It should be noted that the FS is not a material
parameter but a numerical technique to smoothly run the program and
The Johnson-Holmquist-2 (JH-2) model [25] consisting of a poly­ to avoid severe mesh distortions during a simulation. The fine tuning of
nomial equation of state (EOS), a strength model for intact, partially the erosion strain (FS) is a tedious process and is so carried out that the
damaged and completely damaged material, and a damage model was numerical computations of the material response in a simulation are not
employed for simulating the material behaviour of alumina 99.5% affected. When the equivalent plastic strain exceeds the given value (FS),

6
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Fig. 6. The actual projectile perforation and corresponding finite element computations at different stages.

the element is removed from the mesh. been assumed to have increased progressively, and hence, the 1Damage
The 1Damage ¼ 0 is set (or left blank) to specify the JH-2 progressive ¼ 0 has been considered as per the requirement of the JH-2 model. Thus,
damage model and 1Damage ¼ 1 is set to specify the JH-1 instantaneous the damage is assumed to occur progressively and is governed by Eqn.
damage model, see Abaqus Explicit User’s Manual [29]. As in the pre­ (6).
sent study the JH-2 constitutive model has been used, the damage has

7
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Fig. 7. The projectile (a) before experiment (b) after experiment (c) numerical result.

Fig. 8. (a) Schematic of the fracture conoid (b) typical fracture conoid form in the target during computer simulation.

3.2. Johnson and cook model

h � �n i� � _ pl � �h
ε i
σ 0 ¼ A þ B εpl 1 þ Cln 1 bm
T (8)
The Johnson-Cook (JC) constitutive model describes the behaviour ε_ 0
of metals at large strains, high strain rates and high temperatures [26,
27]. The equivalent von-Mises stress of the Johnson-Cook model is where, εpl is equivalent plastic strain. A, B, C, n and m are material
defined as: parameters measured at or below the transition temperature, T0.

8
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Fig. 9. Variation of the angle of indentation due to change in impact velocity.

T
b is the non-dimensional temperature, defined as;
Table 4
The damage quantification of the target with respect to impact velocity. 0 for T < To
S. No. Impact Velocity (m/ No. of radial Size of the Circumferential b ¼ ðT To Þ
T for To � T � Tmelt (9)
s) cracks cone (mm) ðTmelt To Þ
Experiments Numerical 1 for T > Tmelt
Model
The JC fracture model is defined as;
D d D d
X Δεpl
1 52 8 34 12 – – D¼ (10)
2 62 7 31 10 – – εplf
3 94 7 39 12 48 11
4 112 9 49 11 51 12
5 179 8 44 13 42 15
where D is the damage parameter and Δεpl is an increment of the
6 215 7 42 13 50 14 equivalent plastic strain and εpl
f is the strain at failure.
7 249 7 52 11 47 14
8 264 8 48 11 46 13 h � � � _ pl � �
σm � i ε
9 275 9 52 14 59 13 εplf ¼ D1 þ D2 exp D3 1 þ D4 ln b�
½1 þ D5 T (11)
σ ε_ 0

where σm is the mean stress and σ is the equivalent von-Mises stress. D1 –

D5 are the material parameters. The detailed procedure for the cali­
bration of the material parameters for the JC model has been system­
atically discussed in Iqbal et al. [30]. The JC parameters for the Steel
4340 obtained directly from Ref. [12] have been provided in Table 2.

Fig. 10. Failed ceramic tiles after the impact at (a) 52 m/s and (b) 275 m/s.

9
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Fig. 11. The fragments of target at (a) 249 m/s and (b) 112 m/s velocity.

Fig. 12. (a) Variation in kinetic energy (b) Variation in momentum (c) Percentage variation in kinetic energy and (d) Percentage variation in momentum with
respect to impact velocity.

10
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Table 5 mm, see Fig. 7 (b) and (c), respectively. Therefore, the actual and pre­
The residual velocities for normal and oblique impact for varying impact dicted average length of the broken tip was 5.1 and 3.6 mm, respec­
velocity. tively. In a bilayer ceramic-metal (confined) target, the projectile
S. Impact velocity (m/ Residual Velocity (m/s) continued penetrating the conoid located between the projectile and the
No. s)
Experimental (Normal Numerical
backing plate causing the distribution of the momentum on the wider
impact) Simulation area of the backing plate [8] and resulting in a rapid deceleration and
(Obliquity) severe distortion of the projectile.
0� 15� 30� After each experiment, the pieces of the fractured ceramic target
were placed together to measure the dimensions of the fracture conoid.
1 275 239 212 205 183
2 264 215 201 195 178
The fracture conoid represents the magnitude of damage in terms of
3 249 182 190 188 172 failure zone at the front and the rear surfaces. The schematic represen­
4 215 153 151 147 133 tation of the fracture conoid is provided in Fig. 8 (a). The typical nu­
5 179 116 126 120 80 merical reproduction of the fracture conoid in the target corresponding
6 112 70 45 40 0
to 179 m/s impact velocity is shown in Fig. 8 (b). The cone angle (α) of a
7 94 58 23 15 0
conoid {see Fig. 8 (a)} depends upon the elastic properties of the
indenter as well as ceramic material. The cone angle has been noted to
4. Impact and residual projectile velocities be 63� for the quasi static indentation of glass plate against a spherical
steel ball [31]. Florence (1965) also considered a cone angle of 63� for
The impact velocity of the projectile was varied as 52–275 m/s for the analytical modelling of a bi-layer ceramic-metal target against blunt
studying the propagation of damage and the resultant influence on the nosed projectile impact based on the assumption that the diameter of the
ballistic resistance of bare ceramic target. The projectile was initially fracture conoid is equal to the sum of projectile diameter and twice the
fired at a relatively high incident velocity (275 m/s) and in each sub­ ceramic thickness [32,33]. The experiments on the bi-layer ceramic
sequent experiment the velocity was reduced in order to evaluate the target against ogival nosed projectile described that the cone angle is
lowest possible velocity required for perforation of the target. However, much smaller than the assumption (60–65� ) considered in many
the minimum possible velocity that was achieved through the available analytical studies [8]. In another study, the cone angle of the bare
experimental setup for the given mass of the projectile (30 grams) was alumina ceramic target against blunt nosed projectile impact has been
52 m/s. The numerical simulation was performed corresponding to each reported to be 50� [34].
ballistic test for reproducing the residual projectile velocity and the In the present study, the cone angle of the fracture conoid is found to
quantum of damage induced in the target. The actual and computed be dependent upon the incidence velocity of the projectile. Though, a
residual velocities compared in Table 3 described a close correlation definite correlation could not be established, the cone angle is seen to
between the experimental and numerical results. However, at incidence have increased with the increase in the incidence projectile velocity, see
velocities of 62 and 52 m/s, the experiments witnessed complete Fig. 9. With increase in projectile velocity between 52 and 275 m/s, the
perforation while the numerical simulations predicted rebounding of cone angle increased between 65� and 76� . The fracture zones, “d” and
projectile after causing damage at the front surface of the target. “D”, {see Fig. 8 (a)} measured at the front and rear surfaces of the target,
respectively, have been seen to increase with the increase in the pro­
5. Damage initiation and propagation jectile velocity. The measured diameter of the fracture zone at the front
surface (d) varied from 10 to 14 mm and at the rear surface (D) from 31
When a projectile strikes the ceramic target, the compressive stress to 52 mm. The failure mode of the ceramic target reproduced through
waves originated from the point of impact travel in the radial direction the finite element simulations closely represented the front as well as
and across the thickness of the target. As the magnitude of compressive rear surface fracture zones, see Table 4.
wave exceeds the dynamic strength of the ceramic, the damage initiates The number of radial cracks developed in the target have been found
in the form of micro cracks leading to the communition of ceramic in to increase with the increase in the projectile velocity, see Fig. 10. The
front of the projectile head. After reaching the back surface (free circumferential cracks developed in the target have also been found to
boundary) of the target, the compressive wave is reflected back as a be a function of the impact velocity such that at the lowest impact ve­
tensile wave causing the formation of the circumferential tensile cracks locity (52 m/s) the circumferential cracks could not be distinguished and
concentric to the point of impact. With the advancement of the projec­ the radial cracks played a major role in the failure of the target, Fig. 10
tile, the compressive strength is exceeded beneath the projectile head (a). At the highest impact velocity (275 m/s), however, several rings of
pulverizing the ceramic and resulting in the formation of radial cracks in the through thickness circumferential cracks emerged causing the
the outward direction from the point of impact [23]. disintegration of the target into smaller pieces, see Fig. 10 (b). Although,
In the present study, as the target is relatively thin and unconfined, the magnitude and vicinity of damage in the target decreases with the
the communition of ceramic occurred immediately after the develop­ increase in projectile impact velocity (due to localization of damage
ment of contact, see Fig. 6 (a). The projectile tip is also seen to have with increase in the strain rate), however, an opposite trend has been
eroded simultaneously. The stress wave reached the back surface of the noticed in the ceramic targets subjected to ballistic impact not only in
target within 24 μs of the development of contact, causing early for­ the present study but also in the previous studies on bare [24] and
mation of the rear surface radial cracks, see Fig. 6 (a)–(c). With further confined [12] ceramic targets.
movement of the projectile, the circumferential cracks also became The ejected fragments and the broken pieces of ceramic at relatively
visible at the back surface after 32 μs of the contact, see Fig. 6 (d)–(f). higher impact velocities have been found to be finer in comparison to
The intersection of the radial and circumferential cracks led to the for­ those ejected at lower impact velocities. At higher impact velocities the
mation of the fracture conoid {Fig. 6 (g)–(i)} causing the complete fragments and the broken pieces of ceramic were higher in numbers and
failure of the target. The projectile tip is seen to be significantly distorted smaller in size, Fig. 11 (a), as compared to those at lower impact ve­
by the time it reached the back surface of the target, see Fig. 6 (g). The locities, Fig. 9 (b). The increased fineness of the fragments at higher
projectile recovered after each experiment was also found to have a impact velocities was also reported by Savio et al. [35]. The high kinetic
broken tip. Fig. 7 (a) and (b) show the projectile before and after the energy possessed by the projectile in case of higher velocities was sug­
experimentation, respectively. The average length of the recovered gested to be the possible reason of finer size of the debris [35].
projectile was measured to be 47.5 mm while the average residual
projectile length obtained through the numerical simulations was 49

11
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Fig. 13. Ricochet of the projectile at 30� angle of incidence at velocities 94 m/s {(a)–(c)} and 112 m/s {(d)–(f)}.

5.1. Energy dissipated by the target Tate [38] proposes that both the kinetic energy and momentum lost
would increase with the increase in impact velocity subsequent to BLV.
The energy dissipation during projectile impact on a target could be In the present study, since the mass of the eroded tip was insignificant
quantified by the loss of kinetic energy during perforation. Three (approximately 0.5 g), therefore, no loss of the projectile mass has been
different hypotheses have been proposed in the literature pertaining to assumed for the calculation of the residual kinetic energy and mo­
the energy dissipation under ballistic impact [36]. The first theory given mentum. The experimental results indicated that the energy absorbed by
by Florence [33] proposes that the energy absorbed by the target at an the target increased with the increase in the impact velocity up to 249
impact velocity higher than that of the ballistic limit velocity (BLV) is m/s. However, at the two highest impact velocities, 264 and 275 m/s,
same as that of the energy absorbed at BLV. The analytical formulation the absorbed energy has decreased, see Fig. 12 (a). The percentage loss
of Recht [37] for predicting penetration in semi-infinite ductile targets of the absorbed energy, on the other hand, was found to decrease
against rigid projectile is based on the hypothesis that the maximum consistently with the increase in the impact velocity, Fig. 12 (b).
momentum transfer occurs at the just stopped condition (BLV) and that
the momentum transfer actually reduces as the impact velocity in­
creases. The projectile mass erosion model for finite thickness plate by

12
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Fig. 14. Simulation results for perforation of projectile at 275 m/s impact velocity (a) Normal Impact (b) Oblique Impact at 15� (c) Oblique Impact at 30� at 5 μs, 40
μs and 75 μs

13
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

Fig. 15. The erosion of the projectile occurred at different oblique angles (a) 0� (b) 15� and (c) 30� corresponding to 275 m/s velocity.

5.2. Numerical study of oblique impact with the increase in angle of obliquity. The residual projectile velocities
also decreased with the increase in the angle of obliquity. The effect of
The simulations have also been conducted by changing the angle of angle of obliquity became more prominent at relatively lower velocities
incidence to 15� and 30� obliquities. The incidence velocities at both the such that the projectile at 30� obliquity experienced ricochet when
oblique angles were considered identical to what have been obtained impacted at 112 and 94 m/s.
during the experiments performed at normal incidence. The residual
projectile velocities obtained at both the oblique angles have been Authors statement
presented in Table 5, and these have been compared with the residual
velocities obtained at the normal incidence. M.K. Khan: Formal analysis, Methodology, Software. M.A. Iqbal:
The residual velocities seen to have decreased with the increase in Conceptualization, Writing - original draft, Visualization, Project
the angle of obliquity. This is due to the fact that the path length of the administration, Funding acquisition. V. Bratov: Conceptualization,
projectile perforation (effective target thickness) increased with the in­ Project administration, Supervision. N.F. Morozov: Funding acquisi­
crease in the angle of obliquity resulting in the dissipation of more ki­ tion, Project administration, Supervision. N.K. Gupta, Data curation,
netic energy. The longer duration of interaction between the projectile Visualization, Writing - review & editing
and ceramic target has also led to a higher damage of the projectile. The
simulations were also performed at 45� obliquity, however, the projec­ Acknowledgements
tile did not experience perforation in the considered velocity regime. A
similar increase in the ballistic resistance of bi-layer ceramic-metal [6] Authors gratefully acknowledge the financial support provided by
and ceramic-composite target [1] at oblique impact has been reported the Department of Science and Technology (DST), India and Russian
earlier by the investigators. Foundation for Basic Research (RFBR), Russia through research grant
The reduction in residual velocity was 3% and 14% for 15� and 30� nos. INT/RUS/RFBR/P-232, INT/RUS/RFBR/P361, 16-51-45063, 19-
obliquity, respectively, at 275 m/s incidence velocity. At relatively 51-45016 for successfully carrying out this work.
lower incidence velocity, 94 m/s, the percentage reduction in the re­
sidual velocity was 35% at 15� and 100% at 30� obliquity. At 112 and References
94 m/s velocities, the projectile failed to perforate at 30� obliquity and
toppled down after hitting the target, see Fig. 13. The projectile expe­ [1] Z. Fawaz, W. Zheng, K. Behdinan, Numerical simulation of normal and oblique
ballistic impact on ceramic composite armours, Compos. Struct. 63 (2004)
rienced ricochet at both of these velocities. 387–395, https://doi.org/10.1016/S0263-8223(03)00187-9.
The perfroation process at different angles of obliquity is shown in [2] L.M. Bresciani, A. Manes, M. Giglio, An analytical model for ballistic impacts
Fig. 14 at 275 m/s velocity. The perforation path of the projectile ob­ against ceramic tiles, Ceram. Int. 44 (2018) 21249–21261, https://doi.org/
10.1016/j.ceramint.2018.08.172.
tained from the simulations was 5, 5.2 and 5.8 mm corresponding to 0� , [3] J.G. Hetherington, The optimization of two component composite armours, Int. J.
15� and 30� obliquity, respectively. The damage induced in the target as Impact Eng. (1992), https://doi.org/10.1016/0734-743X(92)90145-J.
well as the projectile was higher at oblique impact and its magnitude [4] B. Wang, G. Lu, M.K. Lim, Experimental and numerical analysis of the response of
aluminium oxide tiles to impact loading, J. Mater. Process. Technol. 51 (1995)
increased with the increase in the angle of obliquity. The nose of the 321–345, https://doi.org/10.1016/0924-0136(94)01604-Y.
projectile experienced significant distortion at oblique impact, see [5] V. Madhu, K. Ramanjaneyulu, T. Balakrishna Bhat, N.K. Gupta, An experimental
Fig. 15. At 30� obliquity, the entire ogival nose of the projectile was study of penetration resistance of ceramic armour subjected to projectile impact,
Int. J. Impact Eng. 32 (2006) 337–350, https://doi.org/10.1016/j.
eroded.
ijimpeng.2005.03.004.
[6] S. Sadanandan, J.G. Hetherington, Characterisation of ceramic/steel and ceramic/
6. Conclusions aluminium armours subjected to oblique impact, Int. J. Impact Eng. 19 (1997)
811–819, https://doi.org/10.1016/S0734-743X(97)00019-5.
[7] R. Zaera, V. S�
anchez-G� alvez, Analytical modelling of normal and oblique ballistic
An experimental and numerical investigation was carried out to impact on ceramic/metal lightweight armours, Int. J. Impact Eng. 21 (1998)
study the damage propagation and the ballistic resistance of alumina 133–148, https://doi.org/10.1016/S0734-743X(97)00035-3.
99.5% bare ceramic target against 10.9 mm diameter ogival nosed 4340 [8] M. Lee, Y.H. Yoo, Analysis of ceramic/metal armour systems, Int. J. Impact Eng. 25
(2001) 819–829, https://doi.org/10.1016/S0734-743X(01)00025-2.
steel projectile. [9] J. L�
opez-Puente, A. Arias, R. Zaera, C. Navarro, The effect of the thickness of the
The cone angle of the fracture conoid was found to be dependent adhesive layer on the ballistic limit of ceramic/metal armours. An experimental
upon the incidence velocity of the projectile. Though, a definite corre­ and numerical study, Int. J. Impact Eng. 32 (2006) 321–336, https://doi.org/
10.1016/j.ijimpeng.2005.07.014.
lation could not be established, an increase in projectile velocity from 52 [10] Z.H. Tan, X. Han, W. Zhang, S.H. Luo, International Journal of Impact Engineering
to 275 m/s has described a resultant increase in the cone angle from 65� an investigation on failure mechanisms of ceramic/metal armour subjected to the
to 76� . impact of tungsten projectile, Int. J. Impact Eng. 37 (2010) 1162–1169, https://
doi.org/10.1016/j.ijimpeng.2010.07.004.
The number of radial cracks as well as the circumferential cracks [11] R. Chi, A. Serjouei, I. Sridhar, G.E.B. Tan, Ballistic impact on bi-layer alumina/
developed in the target increased with the increase in the incidence aluminium armor: a semi-analytical approach, Int. J. Impact Eng. 52 (2013) 37–46,
velocity. The fragments were found to be finer at higher incidence ve­ https://doi.org/10.1016/j.ijimpeng.2012.10.001.
[12] A. Serjouei, R. Chi, Z. Zhang, I. Sridhar, Experimental validation of BLV model on
locities and thicker at lower incidence velocities. The numerical simu­
bi-layer ceramic-metal armor, Int. J. Impact Eng. 77 (2015) 30–41, https://doi.
lations correctly reproduced the magnitude and the pattern of damage in org/10.1016/j.ijimpeng.2014.11.001.
the target and the projectile. [13] M.J. Pawar, A. Patnaik, S.K. Biswas, U. Pandel, I.K. Bhat, S. Chatterjee, A.
The percentage loss of the absorbed energy has been found to K. Mukhopadhyay, R. Banerjee, B.P. Babu, Comparison of ballistic performances of
Al2O3 and AlN ceramics, Int. J. Impact Eng. 98 (2016) 42–51, https://doi.org/
decrease consistently with the increase in the impact velocity. 10.1016/j.ijimpeng.2016.08.002.
The damage induced in the target as well as the projectile increased

14
M.K. Khan et al. Thin-Walled Structures 154 (2020) 106784

[14] M.K. Khan, M.A. Iqbal, V. Bratov, N.K. Gupta, N.F. Morozov, Numerical study of [25] G.R. Johnson, T.J. Holmquist, An improved computational constitutive model for
fracture mechanism in ceramic armor under impact load, in: Proc. XLVI Int. brittle materials, in: AIP Conf. Proc., 1994, https://doi.org/10.1063/1.46199.
Summer Sch. Conf. APM 2018, 2018, pp. 134–144. [26] G.R. Johnson, W.H. Cook, Fracture characteristics of three metals subjected to
[15] M. Lee, E.Y. Kim, Y.H. Yoo, Simulation of high speed impact into ceramic various strains, strain rates, temperatures and pressures, Eng. Fract. Mech. 21
composite systems using cohesive-law fracture model, Int. J. Impact Eng. 35 (2008) (1985) 31–48, https://doi.org/10.1016/0013-7944(85)90052-9.
1636–1641, https://doi.org/10.1016/j.ijimpeng.2008.07.031. [27] G.R. Johnson, W.H. Cook, A constitutive model and data for metals subjected to
[16] M.K. Khan, M.A. Iqbal, V. Bratov, N.K. Gupta, N.F. Morozov, A numerical study of large strains, high strain rates and high temperatures, in: 7th Int. Symp. Ballist.,
ballistic behaviour of ceramic metallic bi-layer armor under impact load, Mater. 1983, https://doi.org/10.1038/nrm3209.
Phys. Mech. 42 (2019) 699–710, https://doi.org/10.18720/MPM.4262019_2. [28] T.J. Holmquist, D.W. Templeton, K.D. Bishnoi, Constitutive modeling of aluminum
[17] M.M. Shokrieh, G.H. Javadpour, Penetration analysis of a projectile in ceramic nitride for large strain, high-strain rate, and high-pressure applications, Int. J.
composite armor, Compos. Struct. 82 (2008) 269–276, https://doi.org/10.1016/j. Impact Eng. (2001), https://doi.org/10.1016/S0734-743X(00)00046-4.
compstruct.2007.01.023. [29] ABAQUS/explicit User’s Manual, 2010, version 6.10.
[18] K. Krishnan, S. Sockalingam, S. Bansal, S.D. Rajan, Numerical simulation of [30] M.A. Iqbal, K. Senthil, P. Bhargava, N.K. Gupta, The characterization and ballistic
ceramic composite armor subjected to ballistic impact, Compos. B Eng. 41 (2010) evaluation of mild steel, Int. J. Impact Eng. 78 (2015) 98–113, https://doi.org/
583–593, https://doi.org/10.1016/j.compositesb.2010.10.001. 10.1016/j.ijimpeng.2014.12.006.
[19] S. Feli, M.R. Asgari, Finite element simulation of ceramic/composite armor under [31] C.G. Knight, M.V. Swain, M.M. Chaudhri, Impact of small steel spheres on glass
ballistic impact, Compos. B Eng. 42 (2011) 771–780, https://doi.org/10.1016/j. surfaces, J. Mater. Sci. 12 (1977) 1573–1586.
compositesb.2011.01.024. [32] J.G. Hetherington, B.P. Rajagopalan, An investigation into the energy absorbed
[20] D. Bürger, A. Rocha De Faria, S.F.M. De Almeida, F.C.L. De Melo, M.V. Donadon, during ballistic perforation of composite armours, Int. J. Impact Eng. 11 (1991)
Ballistic impact simulation of an armour-piercing projectile on hybrid ceramic/ 33–40, https://doi.org/10.1016/0734-743X(91)90029-F.
fiber reinforced composite armours, Int. J. Impact Eng. 43 (2012) 63–77, https:// [33] A.L. Florence, Interaction of Projectiles and Compoiste Armor Part II, 1969,
doi.org/10.1016/j.ijimpeng.2011.12.001. pp. 1–69.
[21] N.K. Naik, S. Kumar, D. Ratnaveer, M. Joshi, K. Akella, An energy-based model for [34] S. Sarva, S. Nemat-Nasser, J. McGee, J. Isaacs, The effect of thin membrane
ballistic impact analysis of ceramic-composite armors, Int. J. Damage Mech. 22 restraint on the ballistic performance of armor grade ceramic tiles, Int. J. Impact
(2013) 145–187, https://doi.org/10.1177/1056789511435346. Eng. 34 (2007) 277–302, https://doi.org/10.1016/j.ijimpeng.2005.07.006.
[22] M. Mayseless, W. Goldsmith, S.P. Virostek, S.A. Finnegan, Impact on ceramic [35] S.G. Savio, K. Ramanjaneyulu, V. Madhu, T.B. Bhat, An experimental study on
targets, J. Appl. Mech. 54 (1987), https://doi.org/10.1115/1.3173022 ballistic performance of boron carbide tiles, Int. J. Impact Eng. 38 (2011) 535–541,
introduction pg 1. https://doi.org/10.1016/j.ijimpeng.2011.01.006.
[23] D.A. Shockey, A.H. Marchand, S.R. Skaggs, G.E. Cort, M.W. Burkett, R. Parker, [36] J.G. Hetherington, Energy and momentum changes during ballistic perforation, Int.
Failure phenomenology of confined ceramic targets and impacting rods, Int. J. J. Impact Eng. 18 (1996) 319–337, https://doi.org/10.1016/0734-743X(96)
Impact Eng. (1990), https://doi.org/10.1016/0734-743X(90)90002-D. 89052-X.
[24] D.B. Rahbek, J.W. Simons, B.B. Johnsen, T. Kobayashi, D.A. Shockey, Effect of [37] R.F. Recht, Ballistic Perforation Dynamics of Armor-Piercing Projectiles, 1967.
composite covering on ballistic fracture damage development in ceramic plates, China Lake, California.
Int. J. Impact Eng. 99 (2017) 58–68, https://doi.org/10.1016/j. [38] A. Tate, Extensions to the Hydrodynamic Theory of Penetration, 1985. Kent.
ijimpeng.2016.09.010.

15