Steel and Composite Structures, Vol. 17, No.

4 (2014) 387-403
DOI: http://dx.doi.org/10.12989/scs.2014.17.4.387 387

Delamination growth analysis in composite laminates subjected

to low velocity impact
Masoud Kharazan 1, M.H. Sadr 1a and Morteza Kiani 2,3b
1
Aerospace Engineering Department & Center of Excellence in Computational Aerospace Engineering,
Amirkabir University of Technology, Tehran, Iran
2
Center for Advanced Vehicular Systems, Mississippi State University, Starkville, Mississippi, USA
3
Engineering Technology Associate (ETA), Troy, MI 48083, USA

(Received December 12, 2013, Revised March 02, 2014, Accepted March 14, 2014)

Abstract. This paper presents a high accuracy Finite Element approach for delamination modelling in
laminated composite structures. This approach uses multi-layered shell element and cohesive zone
modelling to handle the mechanical properties and damages characteristics of a laminated composite plate
under low velocity impact. Both intralaminar and interlaminar failure modes, which are usually observed in
laminated composite materials under impact loading, were addressed. The detail of modelling, energy
absorption mechanisms, and comparison of simulation results with experimental test data were discussed in
detail. The presented approach was applied for various models and simulation time was found remarkably
inexpensive. In addition, the results were found to be in good agreement with the corresponding results of
experimental data. Considering simulation time and results accuracy, this approach addresses an efficient
technique for delamination modelling, and it could be followed by other researchers for damage analysis of
laminated composite material structures subjected to dynamic impact loading.

Keywords: composite materials; intralaminar damage; interlaminar damage; delamination; cohesive

zone modelling; impact loading

1. Introduction

A wide range of fibre-reinforced composite materials are used in aerospace and automotive
industries due to relatively low weight in combination with high specific stiffness and strength
compared to metallic alloys. For example, more than 50% of the primary structure in Boeing 787
has been made of composite materials such as carbon fibre or sandwich panels, and the design of
this airplane will be probably followed soon by Airbus A350 (Jackson et al. 2012). However the
damage of composite materials under impacting by foreign objects is one of the great concerns of
their applications since the impact-induced damages can significantly reduce strength and fatigue
life of composite material structures.

Corresponding author, M.Sc., E-mail: [email protected]

a
Professor, Email: [email protected]
b
Ph.D., E-mail: [email protected]

Copyright © 2014 Techno-Press, Ltd.

http://www.techno-press.org/?journal=scs&subpage=8 ISSN: 1229-9367 (Print), 1598-6233 (Online)
388 Masoud Kharazan, M.H. Sadr and Morteza Kiani

Low velocity impact of laminated composite is the main factor that limits widespread
application of this branch of materials especially in aerospace and automotive industries. Impact
loads can result different forms of damages in laminated composite structures. These damages are
classically divided in two main categories.
The first category is identified as intralaminar damages which developing inside plies such as
matrix cracking and fibre breakage. The second category is addressed as interlaminar damages
which develop at the interface between two adjacent plies, called delamination. Many researchers
have experimentally studied composite material structures under impact loading (Davies and
Zhang 1995, Abrate 1998, Reid and Zhou 2000, Davies and Zhang 2004) and few analytical
solutions have been proposed based on simplifying the problem (Zheng 2007). These approaches
often result errors due to oversimplification; consequently, these errors limit the application of the
analytical approaches in damage analysis for composite material subjected to low velocity impact.
By improving computational performance in the last decade, various finite element (FE)
approaches have been introduced to model the damages developed by impact loading. Wang et al.
(2013), Aymerich et al. (2008), Finn and Springer (1993), Kim et al. (2013), Zhou et al.(2013) and
Hosseini-Toudeshky et al. (2013) have studied different kinds of composite structures, either
laminated or sandwich panels under various loading. These researchers used simulation techniques
to investigate modelling of delamination by application of shell, solid, or hybrid shell-solid
elements; however, the advantages and disadvantages of each FE modelling approach are still
under discussion in literature (Dogan et al. 2012). From the viewpoint of industrial application to
design composite structures by FE approach, where the structures are mostly thin, implementation
of shell elements is highly suggested over solid or hybrid models. This can be explained by
considering the complexity and number of elements necessary for modelling a thin structure
through solid or shell elements. If solid elements are used to model a thin structure, the complexity
and number of necessary elements for the model extensively increases. Consequently, shell
elements are highly preferred.
Today, the major commercial software that can be used in crash simulations include:
LS-DYNA, ABAQUS Explicit, PAM-CRASH, DYTRAN, and RADIOSS. Many of researchers
such as (Maio et al. 2013, Heimbs et al. 2009, Feraboli et al. 2011) used LS-DYNA to simulate
the impact response of structures made of different composite materials. Maio et al. (2013) deals
with the prediction of the delamination damage induced by low velocity impact in a laminated
composite plate using progressive damage model (MAT162) implemented in LS-DYNA and
achieved acceptable results for shape and orientation of the observed delamination. However lots
of efforts and investigations are still necessary for improvement damage modelling in composite
laminates under impact loading. These investigations help to achieve comprehensive information
for energy absorption mechanisms, failures modes, and residual mechanical strength of composite
material structures that are needed for improvement the performance of composite structures
(Kiani et al. 2013).
This paper presents a significant approach for delamination growth analysis in the laminated
composite structures by application of shell elements and cohesive-zone modelling. The
procedures and advantages of the current approach are discussed in detail and results are compared
to the experiment presented in the literature (Choi and Chang 1992). In addition, the energy
absorption mechanisms for laminated composite materials under low velocity impact are
discussed.
 Delamination growth analysis in composite laminates subjected to low velocity impact 389

Fig. 1 Comparison between (a) large-mass and (b) small-mass impact (Olsson 2003)

2. Low velocity impact phenomenon

Impact on structures is often classified in two main categories as low-velocity and

high-velocity impacts. However, a similar classification is proposed as boundary controlled and
wave controlled impacts which are associated with the response of the structure to large-mass and
small-mass impactors. A mass-ratio criterion governing the boundary controlled and wave
controlled impact responses has been derived in detail by Olsson (2003). Olsson shows that wave
controlled impact occurs when the mass of impactor is less than one quarter of the mass of the
structure, which waves do not interfere with the boundaries. It is clear that when a structure is
impacted, the elastic waves propagate throughout in the structure from the impact point.
Impact response of a structure is related to the impact duration. If the impact duration is in
order of transition time for dilatational waves, response of the structure is dominated by
dilatational waves. In addition, when the impact duration is longer than the time for the waves to
reach the boundaries of the structure, the response can be studied under quasi-static consideration.
On the other hand, impact velocity is the major parameter that alters the impact response for
composite material structures; therefore, most of the impact responses are investigated under low
and high velocity impact categories. However, an intermediate impact velocity can be considered
between low-velocity and high-velocity impact events.
Different damages that propagate in a structure can change the structural impact responses
especially for composite materials. Indeed, the impact response of a laminated composite structure
is a very complicated phenomenon that is dominated by the interaction of two main categories of
the damage. These two categories of damage and their interaction (Fig. 2) are classically explained
as follows:
 The intralaminar failures, such as damages developing in each ply like matrix cracking,
fibre breakage and fibre matrix debonding
 The interlaminar failures, such as damages developing in each ply interface between two
adjacent ply which called delamination.

In low velocity impact phenomenon, matrix cracks are developed by tensile, compressive and
shear stresses, and usually the cracks distribute within the entire damage region. Fibre compressive
and shear failure are observed locally in contact region, while the local fibre tensile failure
typically develops on the opposite surface and in the area including the large matrix cracks.
390 Masoud Kharazan, M.H. Sadr and Morteza Kiani

Fig. 2 Diagram of the principle of interaction between intralaminar and interlaminar damage of
laminated composite (Chang and Chang 1992)

Delamination is particularly serious failure in the laminar structures because it forms at relatively
low contact loads and plays important role in the flexural stiffness and buckling failure. Different
orientations of the plies within a laminate can intensify the delamination for two adjacent plies due
to the mismatch stiffness at their interface.
The delamination areas are influenced directly by the change in the impact energy. A typical
distribution of delamination due to impact is shown by Fig. 3. At the interfaces between plies with
different fibre orientations, the delamination is usually observed as “peanut shaped” with their
major axis oriented in the direction of the fibres at the lower ply (Zheng 2007).

Fig. 3 Shapes of impact-induced delaminations (Abrate 1998)

 Delamination growth analysis in composite laminates subjected to low velocity impact 391

3. Theory and background

This study focuses on development an approach that efficiently predicts delamination in a

laminated composite panel under low velocity impact with high degree of accuracy and
inexpensive simulation time. Because of small thickness to length ratio of the panel and negligible
stress in the thickness direction, a 2D modelling with shell approach is suggested. As discussed
before, application of shell elements instead of 3D solid elements can significantly reduce the
simulation time and somehow increase the accuracy and performance for the FE approach.
Following sections explain the two important damage classes in composite materials.

3.1 Intralaminar failure modelling

To model laminated composite plate and considering of intralaminar failures, LS-DYNA

explicit code is used in this study because of large different contact algorithms and various
material models implemented in this software. Many literatures reported a good capability of
Enhanced-Composite-Damage (MAT54) material model in LS-DYNA for composite material
analysis (Feraboli et al. 2011, Kiani et al. 2013); therefore, this material model was selected for
the current study. Note should be taken that MAT54 is only valid for shell element modelling
approach and it has been implemented based on the failure criteria developed by Chang and Chang
(1987).
In the elastic region, stress-strain behaviour of the material for fibre (axial, 1 direction), matrix
(transverse, 2 direction) and shear (12 direction) are given by following equations (Wade and
Feraboli 2012)
1
1   1  v12 2  (1)
E1

1
2   2  v21 1  (2)
E2

1
212   12   123 (3)
G12

where σ and ε are stresses and strains for the mentioned direction, τ12 is the shear stress, E1 and
E2 are the young’s modulus in longitudinal and transverse directions, G12 is shear modulus and α in
Eq. (3) is a weighting factor for nonlinear shear stress term.
Beyond the elastic region, damage occurs when one or more criteria defined by the
Chang/Chang formulation are satisfied. In these four criteria, a special criterion for shear stresses
is not defined but the interaction of shear stresses is included in the failure criteria by using
parameter β in the formulation. It is worth noting that the matrix tensile failure plays the most
important role in the low velocity impact phenomenon; therefore, the material model must
consider the failure in the matrix direction. The Chang/Chang material model in LS-DYNA
includes typical damage criteria for unidirectional and fabric composite materials in all material
directions which can be demonstrated as follows (LS-DYNA 2012, Schweizerhof et al. 1998):
 For tensile fibre failure mode where σ1 ≥ 0
392 Masoud Kharazan, M.H. Sadr and Morteza Kiani

2 2
      0 failed
e 2f   1     12   1
 S  (4)
 Xt   c  0 elastic

After failure E1 = E2 = G12 = v12 = v21 = 0

 For compressive fibre failure mode where σ1 < 0
2
    0 failed
ec2   1   1 (5)
 Xc   0 elastic

After failure E1 = v12 = v21 = 0

 For tensile matrix failure mode where σ2 ≥ 0
2 2
       0 failed
em2   2    12   1 (6)
 Yt   Sc   0 elastic

After failure E2 = v21 = G12 = 0

 For Compressive matrix failure mode where σ2 < 0

  2   Yc        2   0 failed
2 2

ed2        2   12 
2 S   2S   1 Y    S   1 0 elastic (7)
 c   c   c   c  

After failure E2 = v21 = v21 = G12 = 0

where Xt is the strength value for fibre tensile, Xc is the strength value for fibre compressive, Yt is
the strength value for matrix tensile, Yc is the strength value for matrix compressive and Sc is the
shear strength value of the unidirectional ply. Whenever one or more criteria defined by Eqs.
(4)-(7) are satisfied in a ply within the element, the elastic properties for that ply are set to zero,
and consequently if all the plies within an element fail, then the element will be considered as the
failed element and removed from the FE simulation.

Fig. 4 Mix mode traction-separation law (LS-DYNA 2012)

 Delamination growth analysis in composite laminates subjected to low velocity impact 393

3.2 Interlaminar failure modelling

Interlaminar failure (delamination) plays a significant role in low velocity impact for composite
plates because of altering the energy absorption mechanism and degrading factor of the plate's
stiffness.
To investigate the delamination propagation at interfaces in laminated composite, an extension
of Dycoss Discrete Crack model is used. This extension is defined based on a bilinear
traction-separation cohesive law with quadratic mixed mode delamination criterion and a damage
formulation. This extension defines a fracture model based on cohesive material model MAT138
(MAT_COHESIVE_MIXED_MODE) which is used with a bilinear traction-separation law as
shown in Fig. 4.
In this cohesive model, the ultimate displacements in both normal and tangential directions are
equal to the displacement at the time when the cohesive contact fails completely. Moreover, a
linear stiffness for loading, which is followed by linear softening during the damage, defines an
accurate but simple relationship between the ultimate displacement and peak tractions as well as
the energy release rate (LS-DYNA 2012)
1
GIc  T   If (8)
2
1
GIIc  S   IIf (9)
2
where GIC and GIIc are the energy release rates for normal and shear interface failure, T and S are
the peak tractions in normal and tangential directions.
In this cohesive zone model, the total mixed-mode relative displacement δm is defined as
 m   I2   II2 , where δI = δ3 is separation in normal direction, and  II  12   22 is separation
in tangential direction. The mixed-mode displacement for damage initiation δ 0 is given by
following equation (LS-DYNA 2012)

1  2
 0   I0 II0 (10)
( II0 ) 2  (  I0 ) 2

T S 
where  I0  and  II0  are damage initiation separations in single modes,   II is
KN KT I
mode mixity which is shown in Fig. 4, and the KN and KT are normal and tangential stiffness of the
cohesive zone. The final mixed-mode separation for the power law (α > 0) and for Benzeggagh-
Kenane law (α < 0) (Benzeggagh and Kenane 1996) are given by
1
 
2(1   )  K N 

2
  KT . 2

F       for   0 (11)
0  GIc   G
  IIc
 
 

  
 
GIc  (GIIc  GIc )  .KT
2
2  for   0
F  
  K   2 .K  
0 1 2  N  (12)

  KN  
KT   T

1  1 
2 2

394 Masoud Kharazan, M.H. Sadr and Morteza Kiani

For the cohesive zone model used in this study, the damage of the interfaces is addressed by
unloading paths from loading and softening paths where the unloading path is pointing to the
origin. Fig. 5 shows a cohesive zone which is modelled between two multilayers shell elements by
using contact definition.

4. FE modelling approach

In this paper, a low velocity impact on a rectangular T300/976 graphite/epoxy plate with
clamped minor edges was studied for two different layup configurations, and the ply orientations
as well as the configurations for each specimen used in the investigation are shown in Table 1. In
addition, the transient impact response of both specimens was investigated under following
assumptions:
 No friction between spherical impactor and laminated plate
 Impactor assume as rigid body
 Neglecting the damping effect in laminated structure
 Skip of considering the gravity force during impact.
Numerical simulation of delamination were carried out by using tie-break contact
(*CONTACT_AUTOMATIC_ONEWAY_SURFACE_TO_SURFACE_TIEBREAK) with contact
OPTION = 11 which is based on bilinear traction-separation law that was mentioned in the last
section. Tiebreak contact allows the transmission of both tensile and compressive forces. The
separation of the slave node from the master is resisted by a contact spring until failure, after
which the tensile coupling is removed. The BREAK part of the contact allows modelling of failure
where the spring decouples tensile forces, allowing independent motion of the slave node under
tension. Post failure in all TIEBREAK contacts allows nodes to interact with segments as in
traditional compression only contacts.
Table 2 shows the necessary material properties for the composite material used in this study
(Choi and Chang 1992, Hosseini-Toudeshky 2006)

Table 1 Ply orientations and geometries of the specimens

Specimens Ply orientation Thickness Span length Width
S1 [03 / 903 / 03 / 903 / 03] 2.16 mm 100 mm 76 mm
S2 [454 / ‒458 / 454] 2.30 mm 100 mm 76 mm

Table 2 Mechanical properties for the laminated composite material

E11 (GPa) E22 (GPa) E33 (GPa) G12 (GPa) G23 (GPa) G13(GPa)
156.00 9.09 9.09 6.96 3.24 6.96
ρ (kg/m3) ν12 ν13 ν23 GIc (J/m2) GIIc (J/m2)
1540 0.228 0.228 0.40 157.0 350.0
Xt (MPa) Xc (MPa) Yt (MPa) Yc (MPa) S12 (MPa)
1520.0 1590.0 45.0 252.0 105.0
 Delamination growth analysis in composite laminates subjected to low velocity impact 395

Fig. 5 Cohesive zone modelled between multi-layered shell element (one integration point for
each single ply)

Fig. 6 FE model of the laminated plate and impactor

The impactor was considered as a spherical rigid body made of steel with nominal radius of
6.35 mm and mass of 0.16 kg which is travelled with initial velocity of 3 to 6 m/s for different test
cases. The laminated plate was modelled by using two offset shell elements in thickness and one
cohesive zone between them (lower 0/90 and 45/-45 interfaces). Inside each shell element a
number of sub-layers have been defined in the shell thickness direction which represents the
laminate layup by using one integration point for each single ply. This modelling approach
referred as “layered shell” (Heimbs et al. 2009).
For modelling delamination in a laminated composite structure, one cohesive interface is used.
Based on the different cases investigated in this study, when more than one cohesive interface is
used in the model, bending stiffness of the model is reduced significantly, even no delamination
occurs. On the other hand, bending stiffness for a stacked shell model is reduced by increasing the
number of delamination contact layers.
The laminated plate in this study was modelled by using two offset shell elements through the
thickness and one cohesive zone between them (lower 0/90 and 45/-45 interfaces). This offset
distance is equal to one half of the upper shell thickness plus one half of the lower shell thickness.
To obtain better performance for the shell element application, Belytschko-Tsay element
formulation (ELFORM 2) was used in the FE modelling which can provide a stiffness hourglass
396 Masoud Kharazan, M.H. Sadr and Morteza Kiani

controlling for the elements (LS-DYNA 2012, Belytschko 1984). The mesh size for the model was
considered as 1 mm which is relatively defined based on the thickness of the plate. This
consideration resulted number of 18,950 elements and 19,306 nodes in the FE model. The
spherical impactor was modelled with rigid material model in LS-DYNA (MAT20). Fig. 6 shows
the FE mesh for the panel and impactor that was generated for this study based on the experiment
setup (Choi and Chang 1992). To model appropriate contact algorithm between the impactor and
the laminated composite plate, automatic surface to surface contact was used.

5. Results and discussion

As mentioned in the previous sections, the damage in a laminated panel can be divided into two
main categories which are the main energy absorption mechanism for low velocity impact
phenomenon. These two categories are discussed in this section and other obtained results were
explained and compared to the experimental results.

5.1 Interlaminar failure

The results of the interlaminar failure for the panels discussed in this study are shown by Figs.
7-10. Figs. 7 and 8 show a peanut shape delamination that developed at the lower 0/90 and 45/-45
interfaces which oriented along the lower ply fibre direction of the interfaces. The delamination
prediction was found to be in very good agreement with experiments for size, shape and
orientation (Choi and Chang 1992). Simulation results also show a delamination that oriented itself

(a)

(b)
Fig. 7 Comparison between delamination shape obtained by simulation and experimental results
from (Choi HY and Chang FK 1992) (lower 0/90 interface) for S1 specimen: (a) Velocity
4 m/s; (b) Velocity 6.7 m/s
 Delamination growth analysis in composite laminates subjected to low velocity impact 397

(a)

(b)
Fig. 8 Comparison between delamination shape obtained by simulation and experimental results
from (Choi HY and Chang FK 1992) (lower -45/45 interface) for S2 specimen: (a) initial
velocity: 5.89 m/s; (b) initial velocity: 9.02 m/s

Fig. 9 Comparison between delamination size obtained by simulation, experimental data and
results from the literature (Choi and Chang 1992) for S1 specimen

along the fibre direction of the bottom ply of the delaminated interface. Also comparison between
delamination size obtained by simulation, experimental data and results from the literature (Choi
and Chang 1992) can be seen in Figs. 9 and 10.
The growth of delamination areas versus time for both S1 and S2 specimens at the impact
velocity of 6.7 m/s are shown by Fig. 11.
398 Masoud Kharazan, M.H. Sadr and Morteza Kiani

Fig. 10 Comparison between delamination size obtained by simulation, experimental data and
results from the literature (Choi and Chang 1992) for S2 specimen

Fig. 11 Delamination areas versus time (ms) for impact velocity of 6.7 m/s found forS1 and S2
Specimens (lower inter-face)

5.2 Intralaminar failure (fibre/matrix failure)

Matrix cracks are the main energy absorbing mechanism in low velocity impact phenomenon
for laminated composite structures. The intralaminar fibre and matrix damage could not be
evaluated precisely by experiment (Choi and Chang 1992); however, it could be only studied by
post processing the simulation results and plotting the history variables (HV) for the material
model (MAT54). These variables are set between 0 and 1 based on failure modes in the lamina:
fibre tensile mode (HV#1), fibre compressive mode (HV#2), matrix tensile mode (HV#3), and
matrix compressive mode (HV#4).
The simulation results of the reference model show the intralaminar failure mode in matrix
tensile failure (Figs. 12 and 13). The two-colored fringe plot of this failure variable for the
 Delamination growth analysis in composite laminates subjected to low velocity impact 399

Fig. 12 Contour of tensile matrix mode failure (em) at different integration point in lower and upper
Shell model at time (1.7 ms) for the initial impact velocity of 6.7 m/s for the S1 specimen

Fig. 13 Contour of tensile matrix mode failure (em) at different integration point in lower and
upper Shell model at time (1.7 ms) for the initial impact velocity of 5.89 m/s for the S2
specimen

individual integration points across the thickness of the 9th bottom plies for S1 specimen and 12th
bottom plies for S2 specimen are shown by these two pictures. Damage starts from the bottom
plies on the opposite side of the impact surfaces. This region is in tension because of plate bending
and the low tensile strength in the matrix direction (Yt = 45 MPa); consequently the matrix failure
in tensile mode is resulted. This phenomenon with higher damage in the bottom plies can be
observed for 9th and 12th bottom plies as well as for each single shell element (3 and 4 integration
points (IP) in lower shell element and 6 or 8 IP in upper shell element corresponding to each ply
sequence).
400 Masoud Kharazan, M.H. Sadr and Morteza Kiani

Fig. 14 Energy plot for the impact velocity of 6.7 m/s for S1 and S2 specimens

(a) (b)
Fig. 15 Contact force plotfor the initial impact velocity of 6.7 m/s for: (a) S1 specimen; and (b) S2 specimen

Fig. 14 shows the variation in the internal energy of the plate for S1 and S2 specimens, and
shows the amount of the absorbed energy for these two specimens mainly due to intralaminar and
interlaminar damages of the laminated plate.

5.3 Contact force of the impact

The numerical contact force can be plotted by recalculating the impactor’s rigid body
acceleration result versus impact time, and it is shown by Fig. 15(a) for the initial impact velocity
of 6.7 m/s associated with [03 / 903 / 03 / 903 / 03] ply sequence (Specimen S1). Also, Fig. 15(b)
shows contact force plot for S2 specimen. Sudden drop in the force-time curve was expected due
 Delamination growth analysis in composite laminates subjected to low velocity impact 401

Fig. 16 Displacement versus time for S1 and S2 specimens under impact velocity of 6.7 m/s

to intralaminar and interlaminar failures of the structure which is represented correctly by the
simulation. As Fig. 15(a) shows, four points on the force-time curve can be identified. Each point
represents failures in the plate:
 Point 1 shows the intralaminar failure mode in matrix tensile failure of the 3th bottom plies
(opposite side of the impact surface).
 Point 2 shows the intralaminar failure mode in matrix tensile failure of the 6th bottom plies.
 Point 3 shows the growth of intralaminar failure with higher rate.
 Point 4 shows the growth of delamination with high rate (as can be seen from Fig. 11) and
matrix compressive failure of the 3th upper plies (impact surface).
 The further drops in the contact force plot are due to plate oscillation after impacting.

The displacement of the plate versus time is shown by Fig. 16. As both Figs. 15 and 16
demonstrate, the time duration for displacements and contact forces versus time are close to each
other and address classification of this impact problem under low velocity impact (refer to Fig.
1(a)). Modelling parameters such as element formulation, number of cohesive zones and mesh size
were explored for results accuracy and simulation time. Considering Belytschko-Tsay element
formulation along with 1mm element size provided the best results which match the experimental
observations. In addition, it was observed that increasing number of cohesive zones results
unrealistic stiffness for the plate.
All the simulations were carried out by a PC with one 2.5 GHz Intel core i5 processor and 4 GB
of RAM. Use of two shell elements with integration points for each single ply significantly
decreases solution time to approximately 20 minutes. The results of the simulation demonstrates
that shell elements with cohesive zone modelling have the potential to be adopted as an effective
simulation tool to predict a complex impact-induced damage patterns in realistic laminated
structures not only for research studies but also industrial applications.
402 Masoud Kharazan, M.H. Sadr and Morteza Kiani

6. Conclusions

In this paper, modelling of interalaminar and interlaminar failure modes in laminated composite
were investigated by using a multi-layered shell element approach. The results showed a high level
of accuracy, which was confirmed with the test results, in terms of size, shape and orientation of
delaminated areas. According to the corresponding test report, the peanut shaped damage which
represents delamination in each layer was correctly simulated for the size and orientation.
Moreover, the major modelling parameters such as material model, number of cohesive zones,
element formulation and element size have been found which result accuracy and stability for the
simulation. According to our investigation, increasing the number of cohesive zones results in
reduction in the stiffness of the laminated plate, and deviates the simulation results from reality. In
addition, using element size equal to the specimen’s thickness while considering Belytschko-Tsay
as the element formulation leads to a better prediction of the delamination, with reduction in
simulation time. This paper demonstrates an efficient approach to studying delamination in
composite material by application of multi-layered shell element and cohesive zones. The
approach uses a low complexity modelling procedure with high accuracy of results to study
delamination in composite materials under low velocity impact.

References

Abrate, S. (1998), Impact on Composite Structures, University Press, Cambridge, UK.

Aymerich, F., Dore, F. and Priolo, P. (2008), “Prediction of impact-induced delamination in cross-ply
composite lminates using cohesive interface elements”, Compos. Sci. Technol., 68(12), 2383-2390.
Belytschko, T., Lin, J. and Tsay, C.S. (1984), “Explicit algorithms for nonlinear dynamics of shells”, Comp.
Meth. Appl. Mech. Eng., 42(2), 225-251.
Benzeggagh, M.L. and Kenane, M. (1996), “Measurement of mixed-mode delamination fracture toughness
of unidirectional glass/epoxy composites with mixed-mode bending apparatus”, Compos. Sci. Technol.,
56(4), 439-449.
Chang, F.K. and Chang, K. (1987), “A progressive damage model for laminate composites containing stress
concentrations”, J. Compos. Mater., 21(9), 834-855.
Choi, H.Y. and Chang, F.K. (1992), “A model for predicting damage on graphite/epoxy laminated
composites resulting from low-velocity point impact”, J. Compos. Mater., 26(14), 2l34-2169.
Davies, G.A.O. and Olsson, R. (2004), “Impact on composite structures”, Aeronaut. J., 108(1089), 541-563.
Davies, G.A.O. and Zhang, X. (1995), “Impact damage prediction in carbon composite structures”, Int. J.
Impact Eng., 16(1), 149-170.
Dogan, F., Hadavinia, H., Donchev, T. and Bhonge, P.S. (2012), “Delamination of impacted composite
structures by cohesive zone interface elements and tiebreak contact”, Cent. Eur. J. Eng., 2(4), 612-626
Feraboli, P., Wade, B., Deleo, F., Rassaian, M., Higgins, M. and Byar, A. (2011), “LS-DYNA MAT54
modelling of the axial crushing of a composite tape sinusoidal specimen”, Composites: Part A, 42(11),
1809-1825.
Finn, S.R. and Springer, G.S. (1993), “Delamination in composite plates under transverse static or impact
loads –a model”, Compos Struct., 23(3), 177-190.
Heimbs, S., Heller, S., Middendorf, P., Hahnel, F. and Weiße, J. (2009), “Low velocity impact on CFRP
plates with compressive preload: Test and modelling”, Int. J. Impact Eng., 36(10), 1182-1193.
Hosseini-Toudeshky, H., Hamidi, B., Mohammadi, B. and Oveysi, H.R. (2006), “Finite element based
prediction of failure in laminated composite plates”, Fract. Nano Eng. Mater. Struct., Proceedings of the
16th European Conference of Fracture, Netherlands, July, pp. 311-312
Hosseini-Toudeshky, H., Saeed Goodarzi, M. and Mohammadi, B. (2013), “Prediction of through the width
 Delamination growth analysis in composite laminates subjected to low velocity impact 403

Delamination growth in post- buckled Laminates under fatigue loading using de-cohesive law”, Struct.
Eng. Mech., Int. J., 48(1), 41-56.
Jackson, P.A., Hunter, J. and Daly, M. (2012), “Jane’s All the World’s Aircraft 2012/2013”, Jane’s Infor-
mation Group.
Kiani, M., Shiozaki, H. and Motoyama, K. (2013), Composite Materials and Joining Technologies for
Composites (Volume 7), “Using experimental data to improve crash modeling for composite materials”,
Proceedings of the 2012 Annual Conference on Experimental and Applied Mechanics, pp. 215-226.
Kim, E.H., Rim, M.S., Lee, I. and Hwang, T.K. (2013), “Composite damage model based on continuum
damage mechanics and low velocity impact analysis of composite plates”, Compos. Struct., 95, 123-134.
LS-DYNA keyword user manual (2012), Version 971 R6.0.0, Vol. 1, February.
Maio, L., Monaco, E., Ricci, F. and Lecce, L. (2013), “Simulation of low velocity impact on composite
laminates with progressive failure analysis”, Compos. Struct., 103, 75-85.
Olsson, R. (2003), “Closed form prediction of peak load and delamination onset under small mass impact”,
Compos. Struc., 59(3), 341-349.
Reid, S.R. and Zhou, G. (2000), “Impact Behaviour of Fibre reinforced Composite Materials and Structures”,
CRC Press.
Schweizerhof, K., Weimar, K. and Rottner, T. (1998), “Crashworthiness analysis with enhanced composite
material models in LSDYNA: Merits and limits”, Proceeding of Fifth LSDYNA International User
Conference, Livermore Software Technology Corp., Livermore, CA, USA, September.
Wade, B. and Feraboli, P. (2012), “Simulating laminated composites using LS-DYNA material model
MAT54 part I: [0] and [90] ply single-element investigation”, University of Washington.
Wang, J., Waas, A.M. and Wang, H. (2013), “Experimental and numerical study on the low-velocity impact
behaviour of foam core sandwich panels”, Compos. Struct., 96, 298-311.
Zheng, D. (2007), “Low velocity impact analysis of composite laminated plates”, Ph.D. Dissertation, The
Graduate Faculty of The University of Akron, OH, USA.
Zhou, J., Guan, Z.W. and Cantwell, W.J. (2013), “The impact response of graded foam sandwich structures”,
Compos. Struct., 97, 370-377.

CC