You are on page 1of 10

SCIENCE CHINA

Physics, Mechanics & Astronomy


Research Paper

June 2011 Vol.54 No.6: 11681177


doi: 10.1007/s11433-011-4261-9

Structural behaviour of fibre metal laminates subjected


to a low velocity impact
FAN JiYing, GUAN ZhongWei* & CANTWELL W J
University of Liverpool, Department of Engineering, Liverpool L69 3GQ, UK
Received August 27, 2010; accepted December 30, 2010; published online February 15, 2011

Structural impact tests were first presented to cover typical fibre metal laminates (FMLs) subjected a low velocity projectile
impact, which produced the corresponding load-displacement traces and deformation/failure modes for the validation of numerical models. Finite element (FE) models were then developed to simulate the impact behaviour of FMLs tested. The aluminium (alloy grade 2024-0) layer was modelled as an isotropic elasto-plastic material up to the on-set of post failure stage,
followed by shear failure and tensile failure to simulate its failure mechanisms. The glass fibre laminate (woven glass-fibre reinforced composite) layer was modelled as an orthotropic material up to its on-set of damage, followed by damage initiation
and evolution using the Hashin criterion. The damage initiation was controlled by failure tensile and compressive stresses
within the lamina plane which were primarily determined by tests. The damage evolution was controlled by tensile/compressive
fracture energies combined with both fibre and matrix. The FE models developed for the 2/1, 3/2 and 4/3 FMLs plates made
with 4-ply and 8-ply glass fibre laminate cores were validated against the corresponding experimental results. Good correlation
was obtained in terms of load-displacement traces, deformation and failure modes. The validated models were ready to be used
to undertake parametric studies to cover FMLs plates made with various stack sequences and composite cores.
damage, Hashin criterion, finite element, impact, FMLs, progressive failure
PACS: 02.70.Dc, 81.05.Ni, 46.50.1a

Introduction

Fibre metal laminates (FMLs) are multi-layered materials


based on stacked arrangements of aluminium alloy and fibrereinforced composite materials. Currently, FMLs such as
GLARE (glass fibre/aluminium) and CALL (carbon fibre/
aluminium) are attracting a number of aircraft manufacturers. For example, the aramid fibre/epoxy system, ARALL,
is presently being used in the manufacture of the cargo door
of the American C-17 transport aircraft whilst GLARE is
being used in the manufacture of the upper fuselage of the
A380, an aircraft that is capable of carrying up to 700 passengers.
*Corresponding author (email: zguan@liv.ac.uk)

Science China Press and Springer-Verlag Berlin Heidelberg 2011

The origin of fibre metal laminates is traced back to the


development of built-up structures that took place at the
Fokker aircraft in 1945 since the introduction of bonding in
the F-27 aircraft [1]. Later it was discovered that the bonded,
laminated structure provided a significant improvement in
fracture toughness and damage tolerance over monolithic
material of the same thickness as the laminate [2]. In 1978
Schijve and Vogelesang started with flight simulation tests
on carbon fibre reinforced laminates and aramid fibre reinforced laminates. The test results looked promising. It had
become clear that fatigue in the fibre-reinforced metal
laminates was a delicate process. Subsequently, Fokker aircraft collaborated in a research project on the manufacture
of fibre metal laminates with Delft University of Technology, ALCOA (suppliers of aluminium), ENKA (producer of
aramid fibres), 3M (producer of adhesives), AKZO (manuphys.scichina.com

www.springerlink.com

Fan J Y, et al. Sci China Phys Mech Astron

factures of fibres and aerospace chemicals) and the Dutch


National Aerospace Laboratory. In 1983, Aramid Aluminium Laminates (ARALL) were first manufactured by
ALCOA. In 1987, patent on the second generation of laminates, based on R and S2 glass fibres was filed by AKZO. A
partnership between AKZO and ALCOA started to operate
in 1991 to produce and commercialise Glare [1].
It is well documented that FMLs combine the durability
of metals with the impressive fatigue and fracture properties
of fibre-reinforced composite materials [35]. However, the
first generation of thermosetting-based fibre-metal laminates suffers a number of key limitations including long
processing cycles, low interlaminar fracture toughness
properties as well as difficulties associated with repair. In an
attempt to overcome many of these problems, researchers at
the University of Liverpool have developed a number of
novel FMLs based on thermoplastic matrices [6,7]. Thermoplastic-based fibre-metal laminates offer a number of
advantages, including very short processing time, ease of
forming, improved chemical resistance, excellent reparability and superior interlaminar fracture toughness properties.
Extensive testing on a glass fibre-reinforced polypropylene
FML has shown that this system offers an excellent resistance to low and high velocity impact loading conditions
[6].
A number of studies were conducted to evaluate the low
and high velocity impact behaviour of two types of fibre
metal laminates (Arall and Glare). Vlot and Fredell [8] undertook a series of low velocity impact tests on Arall and
Glare laminates using a drop-weight impact tester with an
impactor mass of 575 g capable of achieving a maximum
velocity of 10 m s1. Their results showed that Glare offers
a stronger resistance to low velocity impact loading than a
monolithic aluminium alloy and a carbon fibre/thermoplastic
composite. Vlot [9] conducted static indentation and low
and high velocity impact tests on monolithic Al 2024-T3
and 7075-T6, various grades of FMLs and composites, with
a circular clamped test area. It was found that composite
materials were very brittle and required little energy to create the first fibre failure in the material. The energy to create
the first crack for FML with aramid and carbon reinforced
FML (ARALL) was comparable to fibre reinforced composite materials and relatively low compared to Al 2024-T3
and R-glass reinforced FML (Glare). The higher velocity
tended to give a slightly larger dent depth. The FMLs had
approximately the same dent depth after impact as the
monolithic alloy. Vlot et al. [10] also conducted a series of
low velocity impact tests Glare laminates and monolithic
aluminium and found that Glare exhibited a minimum
cracking energy that was fifteen percent greater than that of
monolithic aluminium. The impact damage resistance of
these fibre metal laminates increased with increasing glass/
epoxy content. A number of standard drop weight tests on
GLARE and non-clad 2024-T3 were also undertaken [11],
which indicated that GLARE showed higher resistance to

June (2011) Vol. 54 No. 6

1169

cracking than non-clad 2024-T3 for the complete range of


thicknesses. The better impact performance of GLARE was
attributed to a favourable high strain rate strengthening
phenomenon and relatively high failure strain of glass fibres.
Abdullah and Cantwell [12] investigated impact behaviour of a glass fibre-reinforced polypropylene FML and
found that the FML offered an excellent resistance to both
low and high velocity impact loading conditions. The results showed that the laminates were capable of absorbing
significant energy through extensive plastic deformation in
the aluminium and composite layers and localised micro
cracking in the GFPP, indicating that FMLs could offer excellent resistance to dynamic loading. An early study [3]
also suggested that FMLs based on a polypropylene fibre
reinforced polypropylene matrix could offer an excellent
resistance to localised high velocity impact loading. Caprino
et al. [13] performed low-velocity impact tests on fibre
metal laminates made of 2024-T3 sheets and S2-glass/epoxy
prepreg layers. Various impact velocities, masses, and energies were applied in the tests to investigate the influence of
these parameters on the impact response. Similar tests were
also performed on monolithic 2024-T3 sheets of equivalent
thickness for comparison purposes. Empirical relationships
for calculations of the maximum contact force, energy, and
residual displacement as a function of the maximum displacement were established based on the experimental data.
Modelling of composite damage subjected to impact
within the intermediate strain rate regime may be generally
categorised into four approaches [14], i.e. (1) failure criteria,
(2) fracture mechanics, (3) plasticity or yield surface, and (4)
damage mechanics. The Tsai-Wu [15] failure criterion describes the failure surface in stress or strain space. However,
it is a significant disadvantage to use stress-based failure
criteria to model brittle materials as the scale effect in relation to the crack length in the same stress field cannot be
modelled properly. Fracture mechanics or energy based
approaches may be more effective in modelling matrix
cracks or delaminations [1618], although it is difficult to
incorporate them into to a progressive failure mechanism.
Vaziri and Olson have proposed a combined plasticity and
progressive failure approach to simulate damage of composites [19]. Using gradient plasticity theories, Molinari and
Ravichandran [20] modelled steady plastic shocks generated
by planar impact on metalpolymer laminate composites.
Damage mechanics approaches through various definitions
have been used to determine the virgin material with no
damage to the fully disintegrated material with full damage
[2123].
In general, there are three existing different approaches
for modeling fibre reinforced composite metal laminates
[24]: the micro-level approach, in which an individual fibre
and fibrematrix interfaces are studied, the meso-level
characterization, in which individual plies are modelled and,
finally, the macro-level approach, in which the effect of the
complete homogenised laminate is studied. Lee et al. [25]

1170

Fan J Y, et al. Sci China Phys Mech Astron

investigated the penetration and perforation behaviour of a


6061-T6 aluminium plate and a C12K33 carbon fibre reinforced 6061-T6 aluminium metal-matrix composite plate
subjected to projectile impact using an explicit finite element code, LS-DYNA3D. Perforation of the plate was
found to occur under all of the studied impact conditions.
The deformation behaviour of the plate and projectile as
well as the projectile post-perforation velocity and the deceleration of the projectile were strongly dependent on the
plate properties and impact velocity. Guan et al. [26] investigated the impact response of fibremetal laminates based
on a woven polypropylene (PP) fibre reinforced PP composite and two types of aluminium alloy 2024-O and
2024-T3 at a velocity up to 150 m s1. The PP-based composite was modelled as an isotropic material with a specified tensile cut-off stress to allow for the automatic removal
of failed elements. The failure modes and the maximum
permanent displacement versus normalized impact energy
were both in good agreement with experimental data.
Payeganeh et al. [27] developed explicit FE models to investigate the contact force history, deflection, in-plane
strains and stresses of 2024-O 2/1, 5/4 and 2024-T3 2/1, 5/4
FMLs subjected to low-velocity impact. Failure shear strain
and tension cut-off stresses were specified as failure criteria
for aluminium layers. The failure of fibre laminate was
simulated using TsaiWu failure criterion by specifying
tensile cut-off stress based on the ultimate tensile stress of
the fibre. The interaction between the impactor and the plate
was modelled with the use of a two degrees-of-freedom
system, consisting of springs-masses. Chois linearised
Hertzian contact model [28] was used in the impact analysis
of the hybrid composite plate.
This paper aims to first present the structural impact tests
on typical fibre metal laminates (FMLs) subjected to low
velocity projectile impact, and then the corresponding finite
element (FE) models to simulate the impact behaviour of
FMLs tested. The aluminium (alloy grade 2024-O) layer was
modelled as an isotropic elasto-plastic material up to the
on-set of post failure stage, followed by shear failure and
tensile failure to simulate its failure mechanism. The glass
fibre laminate (woven glass-fibre reinforced composite)
layer was modelled as an orthotropic material up to its
on-set of damage, followed by damage initiation and evolu-

June (2011) Vol. 54 No. 6

tion using the Hashin criterion. The damage initiation was


controlled by failure tensile and compressive stresses within
the lamina plane which were primarily determined by tests.
The damage evolution was controlled by tensile/compressive
fracture energies that combined both fibre and matrix. The
FE models developed for 2/1, 3/2 and 4/3 FMLs plates
made with 4-ply and 8-ply woven glass fibre laminate cores
were validated against the corresponding experimental results.
Good correlation was obtained in terms of load-displacement traces, deformation and failure modes. The validated
models were ready to be used to undertake parametric studies to cover FMLs plates made with various stack sequences
and composite cores.

2 Experimental work
The fibre metal laminates were based on the 2024-O aluminium alloy sheets and the woven glass fibre prepreg
EP-124-44-40. Figure 1 shows the stacking arrangement of
a 2/1 (with 4-ply composite cores) FMLs in the picture
frame mould. One 0.5 mm thick aluminium alloy sheet, four
plies of 0.1 mm thick glass fibre prepreg sheet and another
0.5 mm thick aluminium alloy sheet were placed one by one
in the 200 mm240 mm picture frame mould. The stacking
sequence of the glass fibre-based FMLs investigated is
shown in Table 1.

Figure 1 The stacking arrangement of a 2/1 (4-ply) FML in the mould.

Table 1 Summary of the glass fibre-based FMLs investigated


Laminate

Configuration

Nominal thickness (mm)

2/1

2 aluminium layers + 1 composite layer (4-ply)

1.4

2/1

2 aluminium layers + 1 composite layer (8-ply)

1.8

3/2

3 aluminium layers + 2 composite layers (4-ply)

2.3

3/2

3 aluminium layers + 2 composite layers (8-ply)

3.1

4/3

4 aluminium layers + 3 composite layers (4-ply)

3.2

4/3

4 aluminium layers + 3 composite layers (8-ply)

4.4

Fan J Y, et al. Sci China Phys Mech Astron

The low velocity impact tests on the woven glass fibre


based FML plates were conducted using the drop-weight
test rig as shown in Figure 2. For the 2/1 FMLs (with 4-ply
and 8-ply composite cores), the impactor mass was 2 kg.
For the 3/2 FMLs (with 4-ply and 8-ply composite cores)
and the 4/3 FMLs (with 4-ply and 8-ply composite cores),
the impactor mass was 5.5 kg.
The perforation processes were recorded using a high
speed video camera, as shown in Figure 3, from which the
impact displacement could be calculated using the software
Pro-analyst. The incident energy was varied by changing
the height from which the head was released. The impact
energy was increased systematically until complete perforation occurred. The data processed would be used to validate
computer models developed.

3 Finite element modelling


ABAQUS/Explicit [29,30] was used to develop numerical
simulations of various FMLs subjected to low velocity impact. Two materials need to be dealt with, i.e. woven glass
fibre laminate and aluminium alloy 2024-O. They behave
quite differently. The former is fibre based with very limited
ductility up to failure. The latter can be specified by a nominal

1171

June (2011) Vol. 54 No. 6

yield strength with an ultimate extension of 20%. Therefore,


different constitutive models are required to simulate their
behaviour.
3.1

Modelling the aluminium

The aluminium was modelled as an elasto-plastic material


with rate-dependent behaviour. The strain rate decomposition is
d d el d pl .

(1)

For a rate dependent material, the relationship follows


the uniaxial flow rate definition as:

pl h q, pl , ,

(2)

where h is a known function, q is the equivalent stress, pl


is the equivalent plastic strain and is the temperature.
For the aluminium alloy 2024-O used in this research
programme, the isotropic hardening data are given in Table
2. The density of the aluminium is taken as = 2690 kg m3.
Plastic strain can be calculated as follows:

pl total

total
, y .
E

(3)

If the isotropic hardening is defined, the yield stress can


be given as a tabular function of plastic strain. The yield
stress at a given state could be simply obtained from this
table of data, and it remains constant until the plastic strain
reached the next value in the table (Table 2).
The rate-dependent hardening curves in terms of the
static relation can be expressed as:
pl , pl y pl R pl ,

(4)

where pl and R are the equivalent plastic strain and a


stress ratio, which are

pl

Figure 2 The drop-weight impact rig and the clamped specimens.

t
0

2
pl : pl dt , ,
3

(5)

Table 2 Isotropic hardening data for the aluminium alloy 2024-0 used in
this programme
Yield stress (MPa)

Figure 3 The data recording system using a high speed video camera.

Plastic strain

76

100

0.0074

113

0.0117

122

0.02

150

0.0363

176

0.07

183

0.1

186

0.2

1172

Fan J Y, et al. Sci China Phys Mech Astron

R y ,

element is removed from the mesh or its stiffness is set to


zero.
The tensile failure criterion assumes that failure occurs
when the pressure stress p becomes more tensile than the
specified hydrostatic cutoff stress cutoff, i.e.

(6)

The maximum stress ratio of 2.5 was worked out based


on the minimum and maximum stresses shown in Table 2,
and the strain rate of 25 s1 was calculated by dividing the
failure strain (0.2) by the damage time period of 0.008 s.
Both shear failure and tensile failure are used to simulate
the failure mechanism of aluminium layer in FMLs. The
former is based on the value of the equivalent plastic strain
at element integration points; and failure is assumed to occur
when the damage parameter sh exceeds 1. sh is defined as

sh


pl

pl
f

1
ii cutoff ,
3

(8)

In this criterion, the hydrostatic pressure stress is used as


a failure measure to model a dynamic spall or a pressure
cutoff. For the aluminium alloy 2024-O used, the hydrostatic cutoff stress cutoff is set to 279 MPa. The element
removal mechanism is similar to that offered by the shear
failure criterion.

(7)

( pl ) f is the strain at failure which must be determined

3.2

experimentally, which is set to 0.2. When the shear failure


criterion is met at an integration point, all the stress components will be set to zero and that material point fails. If all of
the material points at any one section of an element fail, the
11 1 E1

22 12 E1
33 13 E1

0
12

0
13
0
23

June (2011) Vol. 54 No. 6

Modelling the woven glass fibre laminates

Before damage initiation, the woven glass fibre laminate is


modelled as an orthotropic elastic material by specifying the
engineering constants, which is

21 E2

31 E3

1 E2

32 E3

23 E2

1 E3

1 G12

where vij characterizes the transverse strain in the j-direction,


when the material is stressed in the i-direction. In general,
vij and vji are related by vij/Ei=vji/Ej. For the glass fibre laminates used in this research programme, the values of three
modules E1, E2, E3; Poissons ratios v12, v13, v23; and the
shear modules G12, G13, and G23 are shown in Table 3. The
density of the glass fibre is taken as =1800 kg m3. As the
glass fibre laminate layer is produced by placing fibre in a
090 pattern, the material behaviour within the laminate
plane is similar. In addition there is no need to separate the
fibre and resin. Therefore, longitudinal and transverse
strengths in tension and compression are taken as the same
figures respectively, so do shear strengths.
However, damage initiation which refers to the onset of
degradation at a material point is modelled based on
Hashins failure criteria [31,32]. These criteria employ four
damage initiation mechanisms, i.e. fibre tension, fibre compression, matrix tension and matrix compression. Taking
11 , 22 and 12 as longitudinal, transverse and shear effective stress tensor components within the glass fibre lamina

11

0
0 22
0
0 33
,
0
0 12
1 G13
0 13

1 G23 23
0

(9)

plane, the initial criteria can be expressed as:


Fibre tension:
2



F 11 12 , 11 0.
XT
SL
t
f

(10)

Fibre compression:
2


F 11 , 11 0.
XC
c
f

(11)

Matrix tension:
2


F 22 12 , 11 0.
YT S L
t
m

(12)

Matrix compression:
2
2
2

22 YC
12
22
F


1
,
YC S L
2 ST 2 ST
c
m

(13)

Table 3 Orthotropic elasticity data of the glass fibre laminates used in this programme
E1 (GPa)

E2 (GPa)

E3 (GPa)

v12

v13

v23

G12 (GPa)

G13 (GPa)

G23 (GPa)

23

23

0.15

0.15

0.15

Fan J Y, et al. Sci China Phys Mech Astron

where XT,XC are the tensile and compressive strengths in the


longitudinal direction, YT,YC are the tensile and compressive
strengths in the transverse direction, SL,ST are the longitudinal and transverse shear strengths, and is a coefficient that
specifies the contribution of the shear stress to the fibre tensile initiation criterion. Here is set to zero, i.e. assuming

1 d f E1
1
Cd
1 d f 1 dm 12 E2
D

June (2011) Vol. 54 No. 6

1173

that there is no shear stress contribution to the initiation of


the fibre tensile failure. In Abaqus, such a criterion can be
used with the element removal to remove failed elements in
relation to fibre and matrix failure in tension and compression.
The damage elastic matrix can be expressed as:

1 d 1 d
f

1 d m E2
0

21

E1

,
0

1 d s GD
0

(14)

which links stress and strain as:

Cd ,

(15)

and where D is the overall damage variable and can be expressed as:
D 1 1 d f

1 d
m

(16)

12 21

In the above equation, df reflects the current state of fibre


damage, dm reflects the current state of matrix damage, and
ds reflects the current state of shear damage, which can be
expressed as:

d tf , if 11 0,
df c
d f , if 11 0,

(17)

d t , if 22 0,
d m mc
d m , if 22 0,

(18)

d s 1 1 d tf

1 d 1 d 1 d .
c
f

t
m

c
m

(19)

The damage parameters for both the fibre and matrix are
calculated based on the critical tensile and compressive
stresses in relation to the corresponding ultimate strengths.
Damage evolution is modelled by the negative slope of
the equivalent stress-displacement relation after damage
initiation is achieved. Fracture energies for fibre tension,
fibre compression, matrix tension and matrix compression
failure modes need to be specified to indicate energies disF
with values
sipated during damage, i.e. G Fft , G Ffc , GmtF and Gmc
set to 40000, 60000, 40000 and 60000 J m2 respectively in
this study. The linear damage evolution is shown in Figure
4. The typical time for the damage process and energy
change is set to 0.008 seconds, which is to ensure the complete perforation takes place. The equivalent stress and
equivalent displacement in Figure 4 are defined based on
the plain stresses and strains on fibre and matrix respectively, in conjunction with the characteristic element length.
The reason for using equivalent displacements rather than
strains is to alleviate the mesh dependency during material
softening. The value of 0eq depends on the elastic stiffness
and strength parameters of the material. The value of feq is

Figure 4 Linear damage evolution.

defined by the strain energy, which is dissipated due to failure and corresponds to the area underneath the curve in
Figure 4.
For the purpose of modelling the interaction between
various parts of pure composites and FMLs subjected to
projectile impact, the general contact interaction was defined between the two neighbouring layers and the surface-to-surface contact interaction was defined between the
projectile surface and the node set of the target centre of
each layer. Contact interaction properties for interactions
between the projectile and the aluminium alloy layer, interactions between the projectile and the glass fibre laminates
layer, interactions between the aluminium alloy layer and
the glass fibre laminated layers, and interactions between
the glass fibre laminated layer were defined respectively and
referred to by the relevant type of interaction.
Typical mesh generations, dimensions, loading and
boundary conditions for typical FMLs (2/1) are shown in
Figure 5. The FML panel is fully fixed at its four sides, i.e.
no displacement and rotation are allowed on its four boundary side faces. Two elements are generated through one
laminate. The typical element dimension ratio in the impact
region is 1.0:1.0:0.3/0.2 (thickness), which was obtained
through mesh sensitivity studies in terms of accuracy of the
modelling output and CPU time consumption. There are
totally 7200 eight-node solid elements and 2391 four-node
rigid surface elements, and the total degrees of freedom of
the model are 201492. An initial velocity is applied to the
projectile to strike the target, with the same velocity and
mass values as those in the experimental work. The size of
an individual time step was controlled automatically based

1174

Fan J Y, et al. Sci China Phys Mech Astron

June (2011) Vol. 54 No. 6

Figure 5 Mesh generations, dimensions (mm), loading and boundary conditions for 2/1 FMLs.

on the numerical convergence. The numerical modelling


results presented in the following part were obtained
through the converged computational jobs.

Results and discussion

The finite element models using the constitutive models and


failure criteria presented in the early part of this paper were
developed to simulate the critical perforation impact tests of
various fibre metal laminates. Figures 6 and 7 show the
simulated and the related experimental load-displacement
traces of 2/1 FMLs plates made with 4-ply and 8-ply composite cores respectively subjected to low velocity impact.
The predicted peak loads for the above FMLs plates were
1788 and 2460 Newtons, respectively, which are 2.4% and
10.0% higher than the experimental results of 1746 and
2236 N respectively. The predicted initial stiffness and the
displacement at the peak load for the targets were in reasonably good agreement with the corresponding experimental results. The predicted perforation energies were
11.69 and 19.41 Joules, respectively. In comparison with
the experimental results of 10.65 and 17.36 J, they were
9.8% and13.2% higher, respectively. However, the simulated

Figure 6 Load-displacement traces for the 2/1 FMLs plate with 4-ply
composite cores subjected to a low velocity (5 m/s) impact.

Figure 7 Load-displacement traces for the 2/1 FMLs plate with 8-ply
composite cores subjected to a low velocity (5.5 m/s) impact.

perforation processes were not as smooth as the experimental ones.


Figures 8 and 9 show the comparison of the simulated
and experimental failure modes of 2/1 FMLs plates made
with 4-ply and 8-ply composite cores subjected to an on-set
perforation impact. The basic features of the experimental
failure modes for all the FMLs plates were well simulated,
in terms of the cross cracks at the rear face and the local
deformation mode at the target centre. Since the only difference between the FMLs plates was the number of composite plies in the core layer, the experimental failure modes

Figure 8 Failure modes of the 2/1 FMLs plates made with 4-ply composite
cores subjected to a low velocity impact. (a) The test; (b) the FE simulation.

Fan J Y, et al. Sci China Phys Mech Astron

Figure 9 Failure modes of the 2/1 FMLs plates made with 8-ply composite
cores subjected to a low velocity impact. (a) The test; (b) the FE simulation.

for these two FMLs plates were quite similar, so did the
simulated ones.
Finite element models of other types of FMLs plates
subjected to a low velocity impact were also developed to
broaden the validation. Figures 10 and 11 show the numerical simulations of the experimental load-displacement traces
for the 3/2 FMLs plates made with 4-ply and 8-ply composite cores respectively subjected to an on-set perforation impact. Very good correlation was obtained between the experimental results and the numerical simulations, in terms
of the overall initial stiffness, the peak load and the perforation process.
The predicted peak loads for these two FMLs plates were
3255 and 4531 N, respectively, which are only 0.6% and
6.0% higher in comparison with the experimental results of
3235 and 4275 N, respectively. The predicted initial stiffness and the predicted displacement at the peak load were
also in reasonably good agreement with the corresponding
experimental results. The predicted perforation energies for
these two plates were 18.88 and 25.38 J, respectively.
Compared to the experimental results of 16.51 and 26.17 J,
they were 14.4% higher and 3.0% lower, respectively.
Figures 12 and 13 display the failure modes obtained
from experimental tests and numerical simulations for the

June (2011) Vol. 54 No. 6

1175

Figure 11 Load-displacement traces for the 3/2 FML plate made with
8-ply composite cores subjected to a low velocity impact from the test and
the FE simulation.

Figure 12 Failure modes of the 3/2 FMLs plates made with 4-ply composite cores subjected to a low velocity impact. (a) The test; (b) the FE
simulation.

Figure 13 Failure modes of the 3/2 FMLs plates made with 8-ply composite cores subjected to a low velocity impact. (a) The test; (b) the FE
simulation.

Figure 10 Load-displacement traces for the 3/2 FML plate made with
4-ply composite cores subjected to a low velocity impact from the test and
the FE simulation.

3/2 FMLs plates made with 4-ply and 8-ply composite cores
respectively subjected to an on-set perforation impact.
Again, essential features of the experimental failure mode,
such as the cross-crack pattern and localised deformations,
were well simulated by the corresponding finite element
models.
The finite element models were also verified by test results of the FMLs plates made with the higher stacking

1176

Fan J Y, et al. Sci China Phys Mech Astron

number of layers. Figures 14 and 15 show the comparison


between the experimental and the numerical load-displacement traces for the 4/3 FMLs plates made with 4-ply and
8-ply composite cores respectively subjected to a low velocity impact.
The peak loads from the numerical predictions and the
experimental tests for these two FMLs plates were 5034
7443, 4883 and 6817 N, respectively. The former are only
3.1% and 9.2% higher than the latter, respectively. Also the
predicted initial stiffness and the displacement at the peak
load for the two FMLs plates were in reasonably good
agreement with the corresponding experimental results. The
predicted perforation energies were 33.48 and 51.11 J respectively, which are only 5.8% lower and 6.4% higher than
the corresponding experimental results of 35.56 and 48.03 J.
Figures 16 and 17 show the central crosssections from
the failed targets and the numerical failure modes of 4/3
FMLs plates made with 4-ply and 8-ply composite cores
respectively subjected to a low velocity impact. Reasonably
good correlation was obtained between the central
cross-sections obtained from the experimental and the numerical failure modes, in terms of the central crack and the
deformation mode.

June (2011) Vol. 54 No. 6

Figure 16 The central cross-sections of the failed 4/3 FML plate made
with 4-ply composite cores subjected to a low velocity impact. (a) The test;
(b) the FE simulation.

Figure 17 The central cross-sections of the failed 4/3 FML plate made
with 8-ply composite cores subjected to a low velocity impact. (a) The test;
(b) the FE simulation,

Figure 14 Load-displacement traces for the 4/3 FML plate made with
4-ply composite cores subjected to a low velocity impact.

Figure 15 Load-displacement traces for the 4/3 FML plate made with
8-ply composite cores subjected to a low velocity impact.

Figures 18 and 19 show the comparison between the predicted peak load as well as the perforation energy and the
corresponding test results in a chart form. Clearly, very
good correlation was obtained, including the general trend.
In order to draw out the reliable relationship, more points
need to be produced by using validated computer models. In
fact, the finite element models developed are well validated
based on the reasonably good prediction of the failure
modes of the FMLs plates with the increasingly high stacking number, together with good simulations of the load-

Figure 18 Comparison of the predicted peak load and the related test
result for all FML plates studied.

Fan J Y, et al. Sci China Phys Mech Astron

4
5
6
7
Figure 19 Comparison of the predicted perforation energy and the related test results for all FML plates studied.

displacement traces, the initial stiffness, the peak load and


the perforation energy.

9
10

Conclusions

Finite element models have been developed to simulate the


structural behaviour of fibre metal laminates with various
stacking sequences and layer geometries subjected to different low velocity impacts. The corresponding experimental tests have also been carried out to generate data for validating finite element models developed, which cover the 2/1,
3/2 and 4/3 FML plates made with both 4-ply and 8-ply
composite cores.
The aluminium alloy 2024-O and the woven glass fibre
laminate have been simulated with different constitutive
models and failure criteria. The former is modelled as an
isotropic elasto-plastic material with strain hardening. Both
shear failure and tensile failure with critical shear strain and
tension cut-off stress are employed. The latter is modelled
as an orthotropic material with critical tensile, compressive
and shear strengths corresponding to the on-set of the brittle
failure. The subsequent damage evolution is controlled by
fracture energies of the composite. Hashins failure criterion
is used here which is capable of simulating brittle failure
behaviour. The above approaches have produced promising
simulations of fibre metal laminate plates subjected to low
velocity impact. Very good correlation has been obtained
between the numerical simulations and the experimental
results, in terms of load-displacement traces, peak load,
perforation energy, deformation mode and failure mode.
The numerical models are ready to be used for further parametric studies, which cover FMLs targets of other configurations subjected to different loading conditions such as
varying projectile size and striking location.
This work was supported by a PhD studentship of the University of Liverpool and partially supported by the Engineering and Physical Sciences
Research Council (EPSRC).

11
12
13
14
15
16

17
18

19
20
21
22
23
24
25

26
27

28
29
30
31

1
2

Vlot A. GlareHistory of the Development of a New Aircraft Material. Dordrecht: Kluwer Academic Publishers, 2001
Krishnakumar S. Fiber metal laminatesThe synthesis of metals and

32

June (2011) Vol. 54 No. 6

1177

composites. Mater manuf processes, 1994, 9: 295877


Reyes G, Cantwell W J. The mechanical properties of fibre-metal
laminates based on glass fibre reinforced polypropylene. Compos Sci
Technol, 2000, 60: 10851094
Vlot A, Gunnink J W. Fibre Metal Laminates: An Introduction.
Dordrecht: Kluwer Academic Publishers, 2001
Vogelesang L B, Vlot A. Development of fibre metal laminates for advanced aerospace structures. J Mater Process Technol, 2000, 103: 15
Vlot A, Kroon E, Rocca G. Impact response of fiber metal laminates.
Key Eng Mater, 1998, 141-143: 235276
Cantwell W J, Wade G, Guillen J F, et al. The impact response of
lightweight fiber-metal laminates. In: Proceedings of ASME Conference, paper AM16A. New York: ASME, 2001
Vlot A, Fredell R S. Impact Damage Resistance and Damage Tolerance of Fibre Metal Laminates. In: Proceedings of the 9th International Conference on Composite Materials. Madrid: Woodhead Publishing Limited, 1993. 5158
Vlot A. Impact loading on fibre metal laminates. Int J Impact Eng,
1996, 18: 291307
Vlot A, Krull M. Impact damage resistance of various fibre metal
laminates. J Phys, 1997, IV 07 C3: 10451050
Vlot A, Vogelesang L B, Vries T. Towards application of fibre metal laminates in large aircraft. Aircraft Eng Aerosp Technol, 1999, 71: 558570
Abdullah M R, Cantwell W J. The impact resistance of polypropylenebased fiber metal laminates. Compos Sci Technol, 2006, 66: 16821693
Caprino G, Spataro G, Del L S. Low-velocity impact behavior of fiber glass-aluminium laminates. Compos-Part A, 2004, 35: 605616
Iannucci L. Progressive failure modelling of woven carbon composite
under impact. Int J Impact Eng, 2006, 32: 10131043
Tsai S W, Wu E. A general theory of strength for anisotropic materials. J Compos Mater, 1971, 5: 5880
Choi H Y, Wu H Y, Chang F K. A new approach towards understanding
damage mechanism and mechanics of laminated composites due to low
velocity impact: Part IIanalysis. J Compos Mater, 1991, 25: 10121038
Kim R Y, Soni S R. Experimental and analytical studies on the onset of
delamination in laminated composites. J Compos Mater, 1984, 18: 7080
Corigliano A, Mariani S, Pandolfi A. Numerical analysis of ratedependent dynamic composite delamination. Compos Sci Technol,
2006, 66: 766775
Vaziri R, Olson M D. Finite element analysis of fibrous composite
structures: A plasticity approach. Compos Struct, 1992, 44: 103116
Molinari A, Ravichandran G. Modeling plastic shocks in periodic laminates
with gradient plasticity theories. J Mech Phys Solids, 2006, 54: 24952526
Lee U. A theory of continuum damage mechanics for anisotropic
solids. J Appl Mech ASME Trans, 1999, 66: 264268
Laws N, Dvorak G J, Hejazi M. Stiffness changes in unidirectional
composites caused by crack systems. Mech Mater, 1983, 2: 123137
Nguyen B N. Damage modeling of laminated composites by the use of
multilayer volume elements. Compos Sci Technol, 1998, 58: 891905
Linde P, Boer H. Modelling of inter-rivet buckling of hybrid composites. Compos Struct, 2006, 73: 221228
Lee W S, Lai C H, Chiou S T. Numerical study on perforation behavior of 6061-T6 aluminium matrix composite. J Mater Process
Technol, 2001, 117: 125131
Guan Z W, Cantwell W J, Abdullah R. Numerical modeling of the impact
response of fiber-metal laminates. Polym Compos, 2009, 30: 603611
Payeganeh G H, Ghasemi F A, Malekzadeh K. Dynamic response of
fiber-metal laminates (FMLs) subjected to low-velocity impact. ThinWalled Struct, 2010, 48: 6270
Choi I H, Lim C H. Low-velocity impact analysis of composite laminates using linearized contact law. Compos Struct, 2004, 66: 12532
ABAQUS, Theory Manual. Version 6.7. Pawtucket: Hibbitt, Karlsson & Sorensen, Inc., 2008
ABAQUS/Explicit, Users Manual. Version 6.7. Pawtucket: Hibbitt,
Karlsson & Sorensen, Inc., 2008
Hashin Z, Rotem A. A fatigue criterion for fiber-reinforced materials.
J Compos Mater, 1973, 7: 448464
Hashin Z. Failure criteria for unidirectional fiber composites. J Appl
Mech, 1980, 47: 329334