PHD Forestry

2.
07 Damage Modeling in Composite Structures

MV Donadon and SFM de Almeida, Instituto Tecnológico de Aeronáutica-ITA, São José dos Campos, São Paulo, Brazil
2014 Elsevier Ltd. All rights reserved.
2.07.1 Introduction 111

2.07.2 Intralaminar Damage Modeling 113
2.07.2.1 3-D Orthotropic Stress–Strain Relationship 113
2.07.2.2 Effective Stresses 114
2.07.2.3 Longitudinal Failure Modes 115
2.07.2.4 Transverse Failure Modes 116
2.07.2.5 In-Plane Shear Failure 119
2.07.2.6 Smeared Cracking Approach 120
2.07.2.7 Objectivity Algorithm 121
2.07.2.7.1 Crack Modeling in the Continuous Medium 121
2.07.2.7.2 Displacement and Traction Vectors in the Singular Band 122
2.07.2.7.3 Energy Dissipation within the Band 122
2.07.2.7.4 Determination of the Function f and the Characteristic Length l* 124
2.07.3 Interlaminar Damage Modeling 127
2.07.4 Formulation 128
2.07.4.1 Kinematics 128
2.07.4.2 Constitutive Laws 129
2.07.4.2.1 Mode I Delamination 129
2.07.4.2.2 Modes II and III Delamination 130
2.07.4.2.3 Mixed-Mode Delamination 130
2.07.5 Applications 132
2.07.5.1 Mesh Sensitivity Study 132
2.07.5.2 In-Plane Coupon Test Simulations 133
2.07.5.3 Delamination Coupon Test Simulations 135
2.07.5.3.1 Double Cantilever Beam 135
2.07.5.3.2 Edge-Notched Four-Point Bending 136
2.07.5.3.3 Mixed-Mode Bending 137
2.07.5.4 Impact-Induced Damage Prediction 143
2.07.6 Summary and Conclusions 145
Acknowledgments 146
References 146
2.07.1 Introduction
Damage in composite structures is a very complex phenomenon that can occur through a different number of failure mechanisms
such as fiber breakage, fiber buckling, matrix cracking, fiber–matrix debonding, and delamination either combined or individually.
The increasing computational resources have allowed reliable prediction of such phenomena with a certain degree of accuracy by
using the finite element method. However, there is still a lot of work to be done in this field to better understand the physics of
failure in order to improve numerical failure models for composite materials. The damage modeling in composites can be broadly
divided into four different approaches:
l Failure criteria approach
l Fracture mechanics approach
l Plasticity approach
l Damage mechanics approach
a) The failure criteria approach

Failure criteria approaches were initially developed for unidirectional materials and restricted to the static regime. They are
divided into two categories: interactive and noninteractive criteria (1). Noninteractive criteria assume that the failure modes are
decoupled and specific expressions are used to identify each failure mechanism. In the stress-based criteria, for example, each and
every one of the stresses in the principal material coordinates must be less than the respective strengths; otherwise, fracture is said to
have occurred. In a similar fashion, the maximum strain failure criteria state that the failure occurs if one of the strains in the
principal material coordinates exceeds its respective failure strain.
Comprehensive Materials Processing, Volume 2 http://dx.doi.org/10.1016/B978-0-08-096532-1.00210-7 111

On the other hand, interactive criteria assume an interaction between two or more failure mechanisms, and they describe the
failure surface in the stress or strain space. Usually stress or strain polynomial expressions are used to describe the boundaries for
the failure surface or envelope. Any point inside the envelope shows no failure in the material. Several interactive failure criteria can
be found in the open literature, such as the Tsai and Wu (2) criterion, Tsai-Hill criterion, and Hoffman criteria among others (1,3,4).
Nevertheless, the disadvantage in using such polynomial criteria is that they do not say anything about the damage mechanisms
themselves; therefore, modified versions have been used to distinguish between failure modes such as Hashin failure criteria (5) and
Chang–Chang failure criteria (6,7).
The main applications of the failure criteria–based damage models include impact-induced damage prediction (8–16),
prediction of residual strength in notched laminated composites (6,7), and delamination in composite laminates (17,18).
The disadvantage in using stress-based criteria for composite materials is that the scale effects relating to the length of cracks
subject to the same stress field cannot be modeled correctly (11,19). In the failure criteria approach, the position and size of the
cracks are unknown. For these reasons, the fracture mechanics approach may be more attractive.
b) The fracture mechanics approach
Fracture mechanics considers the strain energy at the front of a crack of a known size and compares the energy with critical quantities
such as the critical strain energy release rate. The fracture mechanics approach has been used to predict compression after impact strength
of composite laminates (20–23). In such models, the damaged area is replaced by an equivalent hole, and the inelastic deformation
associated with fiber microbuckling that develops near the hole edge is replaced with an equivalent crack loaded on its faces by a bridging
traction that is linearly reduced with the crack closing displacement. The diameter of the hole is obtained from X-radiographs and/or
ultrasonic C-scan images. The results showed good correlation between analytical and experimental values.
Another potential application of the fracture mechanics approach is its indirect use to predict progressive delamination in
composites (24–27). In such models, stress–displacement constitutive laws describe the interfacial material behavior, and fracture
mechanics concepts are used. The area defined by the constitutive relationship is equal to the fracture energy or energy release rate,
and once the stresses have been reduced to zero, the fracture energy has been consumed and the crack propagates. Linear and
quadratic energy-based interaction criteria may be used to describe the crack propagation in mixed-mode delamination.
Comparisons were made with experimental and closed-form results, and good agreement was obtained.
However, the fracture mechanics approach cannot be easily incorporated into a progressive failure methodology because its
application requires an initial flaw. A possible solution is to adopt a hybrid approach by using a stress- or strain-based criterion for
the failure initiation and a fracture mechanics approach for the failure propagation.
c) The plasticity approach
The plasticity approach is suitable for composites that exhibit ductile behavior such as boron/aluminum, graphite/PEEK, and
other thermoplastic composites. The damage model formulation based on the plasticity approach uses an orthotropic stress–strain
relationship combined with the classical flow theory of plasticity and a failure criterion (28). The material constitutive law is
assumed to be elastic–plastic, and it has two stages. The first stage is the postyield and prefailure where an orthotropic plasticity
model is used to model the nonlinear material behavior. The second stage is the postfailure where brittle or ductile failure modes
start to occur.
d) The damage mechanics approach
The damage mechanics approach has been investigated by many researchers in recent years, and its application to damage
modeling in composites has been shown to be efficient. The method was originally developed by Kachanov (29) and Rabotnov
(30), and it has the potential to predict different composite failure modes, such as matrix cracking, fiber fracture, and delamination.
The approach consists of defining thermodynamically consistent internal damage variables dij that describe the crack concentration
and the decohesive process within a representative volume element (RVE) of the material. The internal damage variables dij usually
vary from zero (for undamaged or intact material) to one (fully damaged material), and they degrade locally the material stiffness
matrix according to the predominant failure modes.
Ladeveze and Dantec (31) proposed an in-plane model based on damage mechanics to predict matrix microcracking and fiber/
matrix debonding in unidirectional composites. Two internal damage variables were used to degrade the ply material properties,
one of them associated with the transverse modulus and another with the in-plane shear modulus. A linear elastic-damage behavior
was assumed for tensile and compressive stresses, and a plasticity model was developed to account for the inelastic strains in shear.
Damage evolution laws associated with each failure mechanism were introduced, which relate the damage variables to strain energy
release rates in the ply. Tension, compression, and cyclic shear tests were performed to determine the constants required in the
damage-development laws. Comparisons with experiments were performed by the authors, and a good correlation between
numerical and experimental results was obtained.
Johnson (32) applied the model suggested by Ladeveze and Dantec for the prediction of the in-plane damage response of fiber-
reinforced composite structures during crash and impact events. The damage model was implemented for shell elements into the
PAM-CRASH explicit finite element code. An experimental program was carried out in order to validate the model.
Williams and Vaziri (33) implemented an in-plane damage model based on continuum damage mechanics (CDM) into
LS-DYNA3D for impact-induced damage simulation in composite laminates. The model was originally developed by Matzenmiller
et al. (34). The model considers three damage parameters: the fiber failure damage parameter, the matrix failure damage parameter,
Damage Modeling in Composite Structures 113
and an extra damage parameter to account for the effect of damage in shear response. Individual stress-based criteria for each failure
mechanism were used to detect damage initiation. The damage growth law adopted by Matzenmiller is a function of the strain, and
it assumes an exponential form. The stress/strain curve predicted by this damage function is a Weibull distribution that can be
derived from statistical analysis of the probability of the failure of a bundle of fibers with initial defects. Impact simulations with
different energy levels were performed, and the performance of the proposed model was checked against the Chang–Chang failure
criteria and experimental results. Some limitations of the model were pointed out by the authors. First, the response is predicted by
a single equation and, as a result the loading and postfailure responses, cannot be separated, thus restricting the versatility of the
model. Also, there is a dependence of the damage growth on the loading rate as well as mesh size.
In his subsequent work, Williams et al. (35) proposed a plane-stress CDM model for composite materials. The model was
implemented for shell elements into LS-DYNA3D explicit finite element code. Also, the model is an extension of their previous work
(36), and it deals with some issues related to the limitation of the model suggested by Matzenmiller et al. (34). The approach adopted
by the authors uses the substructuring concept whereby one integration point is used for each sublaminate and composite laminate
theory is used to obtain its effective material properties. The model assumes a bilinear damage growth law, and the damage process has
two phases, one associated with matrix/delamination damage and another with fiber breakage. The driving force for damage growth is
assumed to be a strain potential function, and the threshold values for each damage phase are experimentally determined.
Pinho et al. (37) proposed a three-dimensional failure model to predict failure in composites. Their model is an extension of the
plane-stress failure model proposed by Davila and Camanho (38), and it accounts for shear nonlinearity effects. The model was
implemented into LS-DYNA3D finite element code, and the authors obtained good correlation between numerical predictions and
experimental results for static standard in-plane tests. However, the authors addressed neither strain rate effects nor strain locali-
zation problems associated with irregular mapped meshes.
The following sections present a detailed description of damage modeling in composite structures by using a hybrid approach
that combines CDM with stress and fracture mechanics–based approaches in a unified way. This approach allows prediction of
failure initiation and damage progression within an energy-based framework. In order to provide a comprehensive understanding
of the damage modeling approach, the chapter has been divided into two parts: (1) intralaminar damage modeling, and (2)
interlaminar damage modeling. The first part of the chapter presents a detailed description of a three-dimensional ply failure model.
The second part of the paper describes a constitutive damage model to predict mixed-mode delamination in composite laminates.
Finally, some application examples of the damage models described in the previous sections are presented; the results are discussed,
and conclusions are drawn.
2.07.2 Intralaminar Damage Modeling
The model for intralaminar failure presented in this section is based on the CDM approach, together with the following
assumptions (39):
l Cracks are assumed to be smeared over a RVE of the material.
l Four internal damage variables, d1 ; dtm ; dcm ; d12 ˛ ½0; 1, were introduced at the lamina level to quantify the crack concentration
within the RVE. d1 and d12 are the internal damage variables associated with fiber and in-plane shear-dominated failure modes,
respectively. dtm and dcm are the internal damage variables associated with interfiber-dominated failure modes due to combined
tension/shear and compression/shear loading, respectively.
l The maximum stress failure criterion is used to detect failure initiation in tension/compression in the fiber direction. The Hashin
failure criterion (5) is used to detect failure initiation in the transverse direction due to combined tension and shear loading.
A stress-based criterion proposed by Puck and Shurmann (40) is used to detect transverse compression failure.
l The damage evolution laws are based on strains with their ultimate strain values defined as a function of the fracture energies per
unit volume of damaged material by using a smeared cracking approach.
2.07.2.1 3-D Orthotropic Stress–Strain Relationship

The orthotropic material law that relates second Piola-Kirchhoff stress S to the Green-St. Venant strain e in the local material
coordinate system is written as
S ¼ T t CTe [1]
where e ¼ [e11, e22, e33, e12, e23, e31] is the Green-St. Venant strain vector and T is the transformation matrix given by
2 3
k2 l2 0 0 0 2kl
6 7
6 l2 k2 0 0 0 2kl 7
6 7
6 7
6 0 0 1 0 0 0 7
T¼6
6
7
7 [2]
6 0 0 0 k l 0 7
6 7
6 0 7
4 0 0 0 l k 5
kl kl 0 0 0 k2 l2
where k ¼ cos(b) and l ¼ sin(b). b is the in-plane rotation angle around the Z-direction which defines the orientation of the material
axes with respect to the reference coordinate system. The compliance matrix C1 is defined in terms of the material axes as
2 3
1 v21 v31
6 E11 0 0 0 7
6 E22 E33 7
6 7
6 v v32 7
6 12 1 7
6 0 0 0 7
6 E11 E22 E33 7
6 7
6 7
6 v13 v23 1 7
6 0 0 0 7
6 E 7
6 11 E22 E33 7
C 1
¼6
6
7
7 [3]
6 1 7
6 0 0 0 0 0 7
6 G12 7
6 7
6 7
6 7
6 1 7
6 0 0 0 0 0 7
6 G23 7
6 7
6 7
4 1 5
0 0 0 0 0
G31
where the subscripts denote the material axes, that is,
vij ¼ vx0i x0j [4]
Eii ¼ Ex0i [5]
Since C is symmetric
Vij Vji
¼ [6]
Eii Ejj
The Green-St. Venant strain tensor e is given by

1 t
e¼ F FI [7]
2
where F is the deformation gradient and I is the identity matrix. The Cauchy stress tensor sij can be determined in terms of the
second Piola-Kirchhoff stress tensor S as follows:
sij ¼ J1 FSFt [8]
where J is defined as the Jacobian determinant, which is the determinant of the deformation gradient F. The present formulation will
predict realistic material behavior for finite displacements and rotation as long as the strains are small.
2.07.2.2 Effective Stresses

The effective or damaged stresses are defined as stresses transmitted across the damaged part of the cross section in an RVE of the
material. Based on the isotropic damage theory as originally proposed by Kachanov (29), an orthotropic relationship between local
stresses acting on the damaged configuration can be written in terms of the local effective stresses in the undamaged configuration at
ply level as follows:
8 92 38 9
>
> sd1 >> > s1 >
>
> >
>6 ð1 d1 ðε1 ÞÞ 0 0 0 0 0 7>> >
>
>
> >
> > >
>
> >
>6 7>>
>
>
dcm >
> s2 >
>
d >6 7>> s2 > >
>
>
> >
>6 0 1 d t εt 0 0 0 0 7 > >
>
> >
>6
m m 7>>
>
>
>
>
>
> >
>6 7 >
> d c >
>
>
< sd3 > =6 0 0 1 0 0 077 >
< s 3 =
m >
6
6 7 [9]
> >6 7> dm >c
>
> sd12 >
>6 0 0 0 1 dtm εtm ð1 d12 ðg12 ÞÞ 0 07 >
> s12 > >
>
> >6
>6 7 >
> >
>
>
> >
> 7>> >
>
> >6 07 > dcm > >
> sd >
> >6
> 0 0 0 0 1 dtm εtm 7>>s >
> >
> >6 7 > 23 >>
>
> 23 >
> 5>> >
>
> >4
> > c >
> >
>
>
: d ; > 0 0 0 0 0 1 : d ;
s13 s m
13
where d1, dtm , and d12 are internal variables introduced to quantify the damage concentration within the RVE (41,42). A detailed
description of the physical meaning and definition of these variables will be given in the following sections. The stress components
with the superscript dcm are the stresses degraded locally on the action fracture plane. Details about the determination of the local
action plane and the local degradation procedure are given in Section 2.07.2.4.
This orthotropic relationship between degraded and intact stresses ensures that the material stiffness matrix is positively
defined during the degradation process. Moreover, it is physically based. By inspecting carefully eqn [9], one can notice that
transverse compression damage dcm degrades the through-the-thickness stresses s3, s23, and s13. From the physical point of view,
this relationship incorporates the main features observed in the experiments, such as:
l It accounts for fiber damage effects due to tensile and compression loadings by means of the internal variable d1.
l It provides the coupling between shear-induced damage and matrix cracking during the degradation process using the internal
damage variables d12, dtm , and dcm .
l It accounts for damage coupling between the normal stress s2, out-of-plane shear stress s23 (transverse shear cracking, which
leads to delaminations), and in-plane shear stress s12 for matrix cracking predictions.
2.07.2.3 Longitudinal Failure Modes

The failure index to detect fiber failure in tension is given by
s1
F1t ðs1 Þ ¼ 1 [10]
Xt
In order to detect the catastrophic failure in compression related to the total instability of the fibers, the maximum stress criterion
is used to detect damage initiation,
js1 j
F1c ðs1 Þ ¼ 1 [11]
Xc
where Xt and Xc are the longitudinal strengths in tension and compression, respectively. When one of criteria given above is met,
damage commences and grows according to the damage evolution law proposed by Donadon et al. (43),
d1 ðε1 Þ ¼ dt1 ðεþ c t þ c
1 Þ þ d1 ðε1 Þ d1 ðε1 Þd1 ðε1 Þ [12]
where dt1 ðεþ c
1 Þ and d1 ðε1 Þ are the contributions of the irreversible damage due to fiber breakage in tension and fiber kinking in
compression, respectively.
εt1;0 h i
dt1 ðεþ
1Þ¼1 1 þ k21;t ðεþ þ
1 Þ 2k1;t ðε1 Þ 3 [13]
εþ
1
εc1;0 h i
dc1 ðε
1Þ¼1 1 þ k21;c ðε
1 Þ 2k1;c ðε1 Þ 3 [14]
ε1
with
εþ1 ε1;0
t
k1;t ðεþ
1Þ¼ [15]
ε1;f εt1;0
t
ε1 ε1;0
c
k1;c ðε
1Þ¼ [16]
εc1;f εc1;0
where εt1;0 and εc1;0 are the failure strains in tension and compression, respectively. εþ
1 ¼ maxðε1 ðtÞ; ε1;0 Þ and ε1 ¼ maxðjε1 ðtÞj; ε1;0 Þ
t c
are the maximum achieved strains in the strain time history in tension (s1 > 0) and compression (s1 < 0), respectively. ε1;f and εc1;f
t
are the final strains in tension and compression, which are written as a function of the tensile fiber breakage and compression fiber
kinking fracture toughnesses, respectively, as follows,
2Gtfiber
εt1;f ¼ [17]
Xt l
2Gcfiber
εc1;f ¼ [18]
Xc l
where Gtfiber and Gcfiber are the intralaminar fracture toughnesses associated with fiber breakage in tension and compression,
respectively, and l* is the characteristic element related to the size of the process zone. Experimental procedures and data reduction
schemes to characterize the intralaminar fracture toughness for composites can be found in Ref. (44). Figure 1 shows a plot of the
damage evolution law and the softening behavior in terms of normalized stress (sq/Xq) associated with the failure mode p in respect
p p p
to kq;p ðεq Þ ¼ ðεq εq;0 Þ=ðεq;f εq;0 Þ, where p ¼ c for fiber kinking failure and p ¼ t for fiber breakage in tension in the q-direction,
where q ¼ 1 for the fiber direction.
The proposed damage evolution law d1 ðε1 Þ results in a thermodynamically consistent degradation procedure, regularizing
damage and combining stress, damage mechanics, and fracture mechanics–based approaches in a unified way, similarly to the
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
Normalized stress
0.1 Damage evolution law
0
0 0.25 0.5 0.75 1
Kq,p
Figure 1 Damage evolution law and softening behavior.
approach proposed by Miami et al. (45). Moreover, these laws, compared to the widely used bilinear law, ensure smoothness at
damage initiation and fully damaged stress onsets leading to a more stable numerical response.
2.07.2.4 Transverse Failure Modes

The transverse failure modes consist of transverse matrix cracking either in tension or compression. Based on the Worldwide Failure
Exercise experimental results (46), Pinho et al. (37,47) found that the failure envelope defined between the transverse stress s2 and
in-plane shear stress s12 is accurately described by quadratic interaction criteria. Thus, a failure index based on the interactive
quadratic failure criterion given in Refs. (37,47) has been used to predict tensile transverse matrix cracking. For tensile matrix
cracking, a failure index based on an interactive quadratic failure criterion written in terms of tensile and shear stresses is proposed in
the following form,
2 2 2
s2 s23 s12
F2t ðs2 ; s23 ; s12 Þ ¼ þ þ 1 [19]
Yt S23 S12
Once the above criterion is met, the proposed expression for damage growth due to tensile matrix cracking is given by
εtm;0 h i
dtm εtm ¼ 1 t 1 þ k2m;t εtm 2km;t εtm 3 [20]
εm
with
εtm εtm;0
km;t εtm ¼ t [21]
εm;f εtm;0
where εtm is defined as the resultant strain, which is given by

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
εtm ¼ ε22 þ g2s;m [22]
with
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
gs;m ¼ g212 þ g223 [23]
where εtm;0 and stm;0 are the damage onset resultant strain and stress, respectively; that is,
εtm;0 ¼ εtm jF t ¼1 [24]
2
stm;0 ¼ stm jF t ¼1 [25]

2
In order to account for damage irreversibility effects, εtm ¼ maxðεtm ðtÞ; εtm;0 Þ must be used in eqn [20], where εtm is the maximum
achieved resultant strain in the strain time history. The derivation of the resultant failure strain εtm;f associated with tensile/shear
matrix cracking is based on a power law criterion, which accounts for interactions between energies per unit of volume of damaged
material within the RVE subjected to tensile and shear loadings. The power law energy criterion is given in the following form:
!l !l
gmt s
gm
t
þ s ¼1 [26]
gm c
gmc
with l ¼ 1 for UD laminates. The resultant stress in the transverse direction (or matrix direction) due to combined tensile and shear
loadings is given by
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
stm ¼ s22 þ s212 þ s223 ¼ s22 þ stm [27]
The tensile and resultant shear stress components can be written in terms of the resultant stress as follows:
st2 ¼ stm cosðqÞ [28]
st ¼ stm sinðqÞ [29]
where q is the angle defined between the resultant stress and tensile normal stress in the transverse direction; that is,
q ¼ acos½maxð0; st2 Þ=stm . In a similar way, the tensile and resultant shear strain components are given as follows:
εt2 ¼ εtm cosðqÞ [30]
gts;m ¼ εtm sinðqÞ [31]
For the damage evolution law given by eqn [20], the specific fracture energies associated with tensile and shear stresses are,
respectively, given by
f
Zε2 stm;0 εtm;f cos2 ðqÞ
t
gm ¼ st2 dε2 ¼ [32]
2
0
f
Zgs;m stm;0 εtm;f sin2 ðqÞ
s
gm ¼ st dg ¼ [33]
2
0
where the areas under the stress–strain curves defined by the proposed polynomial damage evolution laws are identical to the one
defined by the widely used bilinear softening law given in Refs. (43,37). Substituting eqns [32] and [33] into eqn [26], we obtain the
following expression for the final strain due to the combined tensile and shear stress state:
" !l !l #1
cos2 ðqÞ sin2 ðqÞ
l
2
εm;f ¼ t
t
t
þ s
[34]
sm;0 gm c
gm c
where gm t and g s are the critical specific fracture energies. These energies are related to the intralaminar fracture toughnesses. By
c mc
using a smeared cracking formulation (48) and assuming that for UD laminates the values of intralaminar toughnesses associated
with tensile matrix cracking and shear matrix cracking are comparable with mode I and mode II interlaminar fracture toughnesses,
a relationship between specific critical fracture energies and intralaminar fracture toughnesses can be written as
GIc
t
gm ¼ [35]
c
l
s GIIc
gm ¼ [36]
c
l
where l* is the characteristic length associated with the length of the process zone for each particular failure mode. A detailed
description of the characteristic length calculation will be presented in the following section.
The failure index to detect matrix cracking in compression failure is based on the criterion proposed by Puck and Schurmann
(40,49). Their criterion is based on the Mohr–Coulomb theory, and it enables the prediction of fracture planes for any given stress
state related to inter-fiber-failure modes. This criterion is currently the state of the art to predict the transverse compression response
of composite laminates. The failure index to detect matrix cracking in compression based on the failure criterion proposed by Puck
and Schurmann (40,49) can be written as follows:
!2 2
snt snl
F2c ðsnt ; snl Þ ¼ þ 1 [37]
SA23 þ mnt snn S12 þ mnl snn
Here, the subscripts n, l, and t refer to the normal and tangential directions in respect to the fracture plane direction. S12 is
the in-plane shear strength, and SA23 is the transverse shear strength in the potential fracture plane (action plane), which is
given by (50)

Yc 1 sinðfÞ
SA23 ¼ [38]
2 cosðfÞ
with
f ¼ 2qf 90 [39]
where Yc is the transverse compression strength. The fracture angle can be determined either experimentally or alternatively using
eqn [39], where qf maximizes the failure criterion. Following the Mohr–Coulomb failure theory, the friction coefficients can be
determined as a function of the material friction angle as follows:

mnt ¼ tan f ¼ tan 2qf 90 [40]
In the absence of experimental values, an orthotropic relationship for the friction coefficients can be used (40):
mnt m
¼ nl [41]
SA23 S12
The stress components acting on the potential fracture plane are written in terms of angle qf which defines the orientation of the
fracture plane in respect to through-the-thickness direction (direction-3 in the local material coordinate system),

snn qf ¼ s2 m2 þ s3 1 m2 þ 2s23 mn [42]

snt qf ¼ s2 mn þ s3 mn þ s23 2m2 1 [43]

stt qf ¼ s2 1 m2 þ s3 m2 þ 2mns23 [44]

snl qf ¼ s12 m þ s13 n [45]

slt qf ¼ s12 n þ s13 m [46]
where m ¼ cos(qf) and n ¼ sin(qf). qf is the rotation around the local fiber direction (direction-1 in the local material coordinate
system), as shown in Figure 2. It is clear that in order to apply eqn [37], qf must be known. Many authors have defined qf z53
based on experimental results for standard compression tests in UD laminates (37,47,43). This is true and has also been
confirmed by Puck and Schurmann (49,40). However, qf z53 is only valid for uniaxial compression loading. This implies that
qf changes for different stress states and alternatives are needed in order to handle such a problem. By examining eqn [37], it is
possible to see that the criterion also has the potential of predicting transverse intralaminar shear cracking for values of qf
different from 530. The transverse intralaminar shear cracking is a very important failure mode because it leads to delamination
between adjacent layers. In order to tackle this problem, we have used an iterative procedure to compute the fracture plane
orientation for a given stress state. The procedure consists of incrementally varying qf within the interval [90 qf 90 ] for
a given stress state defined at ply level and checking if the failure index for matrix cracking in compression is reached. Once the
Figure 2 Action plane for transverse compression failure mode.

failure index is reached, the local shear components snl and snt acting on the candidate fracture plane are degraded to zero
according to the following damage evolution law:
εcm;0 h i
dcm εcm ¼ 1 c 1 þ k2m;c εcm 2km;c εcm 3 [47]
εm
with
εcm εcm;0
km;c εcm ¼ c [48]
εm;f εcm;0
where εcm is the resultant shear strain on the action plane, which is defined as
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
εcm ¼ ε2nl þ ε2nt [49]
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
scm ¼ s2nl þ s2nt [50]
with

εnl qf ¼ g12 m þ g13 n [51]

εnt qf ¼ ε2 mn þ ε3 mn þ g23 2m2 1 [52]
εcm;0 and scm;0 are the damage onset resultant strain and stress, respectively; that is,
εcm;0 ¼ εcm jF c ¼1 [53]
2
scm;0 ¼ scm jF c ¼1 [54]

2
The resultant final strain for transverse compression failure is defined in terms of mode II interlaminar fracture toughness as
follows:
2Gcmatrix 2GII
εcm;f ¼ ¼ c c [55]
scm;0 l sm;0 l
In order to account for damage irreversibility effects, εcm ¼ maxðεcm ðtÞ; εcm;0 Þ must be used in eqn [47], where εcm is the
maximum achieved resultant strain on the action plane in the strain time history. After degrading the shear stresses acting on
the potential fracture angle, the stresses are rotated back to the local material coordinate system using the following
transformation:
dc
s2m ¼ m2 snn 2mnsnt 1 dcm εcm þ stt 1 m2 [56]
dc
s3m ¼ snn 1 m2 þ 2mnsnt 1 dcm εcm þ stt m2 [57]
dc
s12m ¼ msnl 1 dcm εcm nslt [58]
dc
s13m ¼ mnsnn þ 2m2 1 snt 1 dcm εcm nmstt [59]
dc
s23m ¼ nsnl 1 dcm εcm þ mslt [60]
2.07.2.5 In-Plane Shear Failure

The observed behavior of glass and carbon fiber laminates generally shows a marked rate dependence on matrix-dominated
shear failure modes. For this reason, a rate-dependent constitutive model has been used to model the in-plane shear
behavior. The constitutive model formulation is based on previous work carried out by Donadon et al. (43,51), and it accounts
for shear nonlinearities, irreversible strains, and damage within the RVE. The stress–strain behavior for in-plane shear failure is
defined as
s12 ¼ aG12 g12 [61]
with

G12 ¼ G012 þ c1 ec2 g12 1 [62]
where G012 is the initial shear modulus and c1 and c2 are material constants obtained from static/quasistatic in-plane shear tests. a is
the strain-rate enhancement given by the following law:
g_
12
c3
a¼1þe [63]
where c3 is another material constant obtained from dynamic in-plane shear tests. By decomposing the total shear–strain into inelastic
gi12 and ge12 elastic components, the inelastic shear–strain can be written in terms of the elastic and total strain components as
s12 ðg12 Þ
gi12 ¼ g12 ge12 ¼ g12 [64]
G012
The failure index for in-plane shear failure is based on the maximum stress criterion, and it is given by
js12 j
F12 ðs12 Þ ¼ 1 [65]
S12
The proposed damage evolution law for in-plane shear failure is given by
g12;0 gi12;0
d12 ðg12 Þ ¼ 1 1 þ k212 ðg12 Þð2k12 ðg12 Þ 3Þ [66]
g12 gi12;0
with
g12 g12;0 2gi12;0
k12 ðg12 Þ ¼ [67]
g12;0 gi12;0 g12;f
where g12,0 and gi12;0 are the total strain and total inelastic strain at failure (s12 ¼ S12), respectively; that is,
g12;0 ¼ g12 jS12 [68]

gi12;0 ¼ gi12 S [69]
12
g12,f is written in terms of the intralaminar toughness in shear,

2Gshear
g12;f ¼ [70]
S12 l
In the absence of experimental results for Gshear, it is reasonable to assume Gshear ¼ GIIc for UD plies, where GIIc is the mode II
interlaminar fracture toughness.
2.07.2.6 Smeared Cracking Approach

The idea behind the smeared cracking formulation is to relate the specific or volumetric energy, which is defined by the area
underneath the stress–strain curve, with the fracture energy of the material (48,52). The method assumes a strain-softening
constitutive law for modeling the gradual stiffness reduction due to the microcracking process within the cohesive or process zone of
the material and translates this damaged width to the finite element domain, avoiding strain localization problems and mesh-
dependent solutions. In order to illustrate the method, let us assume a linear-elastic-damageable material model represented by the
bilinear constitutive law shown in Figure 3(a).
For a single finite element (Figure 3(b)) under a uniaxial deformation process, the specific energy (or energy dissipated per unit
volume) is defined by the area underneath the stress–strain curve, which is given by
Zεf
1 1
gf ¼ s d ε ¼ s0 εf [71]
2 2
0
The fracture energy (or energy dissipated per unit area) within a fully failed element can be written in terms of the specific energy
by multiplying the specific energy by a geometric quantity defined as characteristic length (l*), which has a direction aligned with
the loading direction (for isotropic materials); it is equal to the height of the element in this case (see Figure 3(b)), that is,
1
Gf ¼ gf l ¼ s0 εf l [72]
2
The dimension of the elements which defines the characteristic length must satisfy the following condition in order to ensure
material stability:
Gf
l < [73]
gf
Figure 3 Illustration of the smeared cracking concept: (a) bilinear constitutive law, (b) finite element loaded in tension.
The softening modulus H is defined as a function of the characteristic length and the material fracture energy as follows:
s20 l
H¼ [74]
s0 l ε
0 2Gf
2.07.2.7 Objectivity Algorithm

The smeared formulation described in the previous sections relates the specific energy within an RVE with the fracture energy of
the material for each particular failure mode. Since finite elements are volume based, mesh dependency problems will arise as
a result of the mesh refinement. The correction of the postfailure softening slope according to the finite element size, as
reported by Bazant (48), seems an attractive solution for the problem. However, the approach has some limitations. First, the
crack growth direction must be parallel to one edge of the finite element, which is not the case for multidirectional composite
laminates where layers can have arbitrary directions. Second, it cannot handle nonstructured meshes required in most of the
complex finite element models with geometric discontinuities. In order to overcome such limitations and to ensure the
objectivity of the model for generalized situations, a methodology originally developed by Oliver (53) has been used and
extended to handle composite layers (54). The dependence of the characteristic length on the fracture energy, as well as its
mathematical expression, were derived based in the work proposed by Oliver (53). The method ensures a constant energy
dissipation regardless of mesh refinement, crack growth direction, and element topology so that it is still applicable to
nonstructured meshes.
2.07.2.7.1 Crack Modeling in the Continuous Medium

Imagine a singular line in a two-dimensional domain as a continuous material line, across which displacements are continuous but
displacement gradients are discontinuous. The condition for a point belonging to a singular line with unit normal n at this point is
that the determinant of the acoustic tensor in the n-direction be zero; that is (53),

det ni Cijkl nl ¼ 0 [75]
Nonpositive materials bifurcate, producing singular lines, and the equation above permits their direction to be determined at
each point. In the context of standard finite elements of C0 continuity, a singular line can be modeled only by the sides of the
elements, these being the only points in the mesh where displacement gradient discontinuities can be obtained. However, a crack
produces not only displacement gradient discontinuities but also displacement discontinuities. This latter kind of discontinuity
cannot be modeled by a C0 finite element mesh for finite levels of discretization. However, a displacement discontinuity can be
modeled as the limit of two parallel singular lines G and Gþ which tend to coincide with each other. The band delimited by these
lines is known as a singular band, and h is its width.
By assuming an orthogonal curvilinear coordinate system ðx0 ; y0 Þ in the interior of the band, where y0 coordinate lines are parallel
to the singular lines G and Gþ and x0 are the straight coordinate lines. Let uþ ðy0 Þ and u ðy0 Þ be the displacement vectors on Gþ,
and G the relative displacement vector can be written as
uðy0 Þ ¼ uþ ðy0 Þ u ðy0 Þ [76]
as a vector representing the displacement ‘jump’ between the two singular lines and

dðy0 Þ ¼ lim uðy0 Þ ¼ lim uþ ðy0 Þ u ðy0 Þ [77]
h10 h10
If ds0, the singular band is modeling a discontinuous displacement field as the limit of a continuous one. This allows a crack to
be idealized as a limit (with mesh refinement) of a band of finite elements where, by means of some numerical mechanisms, the
condition ds0 is satisfied.
2.07.2.7.2 Displacement and Traction Vectors in the Singular Band

Consider a singular band in the solid, with a width h according to Figure 4.
Along a coordinate line x0 , the displacement vector u can be expanded from its value in the G line using Taylor’s series as

vu
uðx0 ; y0 Þ ¼ u ðy0 Þ þ Dx0 þ O h2 [78]
vx0
and consequently, for a point in the Gþ line.

vu
uþ ðy0 Þ ¼ u ðy0 Þ þ hðy0 Þ þ O h2 [79]
vx0
From eqns [72], [74], and [75] we can write,

vu
uðx0 ; y0 Þ ¼ u ðy0 Þ þ Dx0 þ O h2 [80]
vx0
Dx0
uðx0 ; y0 Þ ¼ u ðy0 Þ þ uðy0 Þ þ 0 h2 yu ðy0 Þ þ fðx0 ; y0 Þuðy0 Þ [81]
h
where f is a function to be determined, which approximates Dx0 =h when h10. From eqns [76] and [81], it can be seen that
f ¼ 00G [82]
f ¼ 00Gþ [83]
The equilibrium across the singular band will be enforced by assuming the traction vector t acting on the plane defined by the
normal n,
ti ¼ sij nj [84]
which is constant in the x0 direction, that is,

tðx0 ; y0 Þ ¼ t þ ðy0 Þ ¼ t ðy0 Þ [85]
2.07.2.7.3 Energy Dissipation within the Band

For a generic deformation process that takes place over a time sð0 s NÞ, the specific energy dissipation (energy per unit volume)
within a closed domain U (see Figure 4) is given by
ZN ZN
0 0 0 0
gf ¼ sij ðx ; y ; sÞdεij ðx ; y ; sÞ ¼ sij ε_ ij ds [86]
0 0
For the uniaxial deformation process, gf would be, for a given point, the area under the stress–strain curve at that point. By taking
the linearized geometric equations, gf can be expressed as
ZN ZN ZN
1 vu_ i vu_ j vu_ i 1 vu_ i vu_ j
gf ¼ sij þ ds ¼ sij ds sij þ ds [87]
2 vxj vxi vxj 2 vxj vxi
0 0 0
(a) (b)
B
h
Figure 4 Analysis within the singular band: (a) crack band; (b) infinitesimal width of the band.
The last integrand in eqn [87] is zero, being the product of a symmetric and antisymmetric tensor, so that
ZN ZN
vu_ i v
gf ¼ sij ds ¼ sij u_ i ds [88]
vxj vxj
0 0
where the Cauchy’s equations for quasistatic processes and negligible body forces ðvsij =vxi ¼ 0Þ have been considered. The total
dissipated energy in the domain U is
Z Z ZN
v
W ¼ gf dU ¼ sij u_ i ds dU [89]
vxj

U U 0
By applying Gauss’s theorem to eqn [89] and using eqn [84], we obtain
Z ZN Z ZN
_
W ¼ sij nj ui ds dG ¼ ti u_ i ds dG [90]
G 0 G 0
Owing to the infinitesimal width of the band, the curvilinear integral in eqn [90] can be evaluated only on the lines Gþ and G
(see Figure 4):
Z ZN
W ¼ ti u_ i ds dy0 [91]
Gþ WG 0
and taking into account eqns [81] and [85] we obtain,

Z ZN Z ZN
W ¼ ti ðy0 ; sÞu_
i ðy 0
; sÞds dy0 þ ti ðy0 ; sÞfðx0 ; y0 Þuðy
_ 0 ; sÞds dy0 [92]
þ þ
G WG 0 G WG 0
The first integral vanishes because the contributions on Gþ and G cancel each other out. Thus, the dissipated energy on U is
Z ZN
W ¼ fðx0 ; y0 Þ ti ðy0 ; sÞuðy
_ 0 ; sÞds dy0 [93]
þ
G WG 0
Now, if the case where U ¼ U is considered; that is, the whole band between points A and B in Figure 4, the total energy
dissipated within the band between points A and B is
0
Z ZyB ZN
W¼ gf dU ¼ ti ðy0 ; sÞuðy
_ 0 ; sÞds dy0 [94]
U yA0 0
Equation [94] establishes that the energy dissipated within the idealized band can be written as a curvilinear integral along its
length. The integrand of eqn [94] represents the energy dissipated per unit of area, which in terms of fracture mechanics is the
fracture energy Gf :
ZN
Gf ðy0 Þ ¼ ti ðy0 ; sÞuðy
_ 0 ; sÞds [95]
0
If Gf is assumed to be a material property independent of the spatial position of the point from eqns [93] and [95], we
obtain
Z Z
W ¼ Gf fðx0 ; y0 Þdy0 ¼ Gf fðx0 ; y0 Þdy0 [96]
Gþ WG G
By applying Green’s theorem and taking into account eqn [86], we can write the energy dissipated within the band as
Z Z
vf
W ¼ Gf dU ¼ gf dU [97]
vx0

U U
The local form of eqn [97] is

vf Gf
gf ¼ Gf ¼ [98]
vx0 l
where
1
vf
l ðx0 ; y0 Þ ¼ [99]
vx0
The parameter l* plays the role of relating the specific energy (per unit of volume) and the fracture energy (per unit of area). l* is
also identified as the characteristic length or crack band width used in existing cracking models (48). For the unidimensional case
and using the proposed Hermitian stress–strain softening law, the specific energy can be written as
s0 εf
gf ¼ [100]
2
where s0 is the material strength and the degraded stress is given by
sd ¼ sð1 dðεÞÞ [101]

where the damage evolution law dðεÞ is given in terms of the strain as follows:
ε0
dðεÞ ¼ 1 1 þ k2 ðεÞð2kðεÞ 3Þ [102]
ε
with
ε ε0
kðεÞ ¼ [103]
εf ε0
Using eqn [100] the failure strain εf can be written in terms of the fracture energy and the characteristic length as
2Gf
εf ¼ [104]
s0 l
2.07.2.7.4 Determination of the Function f and the Characteristic Length l*

In order to apply the theory presented in the previous section to the discretized medium, consider a mesh of C0 continuous
hexahedron solid finite elements (see Figure 5).
A set of cracked elements is determined by using failure criteria for detecting the crack initiation; the crack orientation depends on the
fiber direction. The cracked plane is defined here as a normal vector that is parallel to the fibers for fiber failure and normal to the fiber
direction for matrix failure. The algorithm described in this section for determining the characteristic length was proposed by Oliver
(53), and it has the advantages of calculating the characteristic length for arbitrary crack directions and any finite element geometry.
From eqn [97] the function f has to be continuous and derivable, satisfying eqns [82] and [83]. A simple function defined in the
isoparametric coordinates x and h that fulfills these requirements is
X
nc
fðx; hÞ ¼ Ni ðx; hÞfi [105]
i¼1
Node 8
Node 5 Node 7
Node 6
Node 3
Node 1
Node 2
Figure 5 Determination of the characteristic length for hexahedron elements.

where nc is the number of corner nodes of a virtual plane located at the midplane of the element (nc ¼ 4, for our case), Ni are the
standard C0 shape functions of an element of nc virtual nodes in its midplane, and fi is the value of the f at corner i. If the crack
location inside the element is known, fi takes the value þ1 if the corner node i is ahead the crack, and 0 otherwise. The function
defined by eqn [105] fulfills the required condition of continuity within elements and takes the values þ1 for the nodes on the
boundaries ahead the crack and 0 for the nodes on the boundaries behind the crack (see Figure 6).
In general, however, the exact crack location is not known, and usually only some indication of the onset of cracking and the
crack directions is obtained at the integration points. The following algorithm proposed by Oliver (53) has been used to determine
the characteristic length at each integration point j, as shown in Figure 7.
1. A set of local Cartesian axes x0 and y0 is defined at the center of the element, this being identified by the values of the isoparametric
coordinates (x ¼ 0, h ¼ 0, and z ¼ 0). The direction of the local axis x0 is defined by the normal to the fracture plane, which is the
fiber angle for fiber failure (qj ¼ qf) and (qj ¼ qf þ 90 ) for matrix failure.

x0i cos qj sin qj xi
¼ [106]
yi0 sin qj cos qj yi
2. Values of f at each corner node are established according to their position with respect to the local axis x0 ,y0
ðfi ¼ 1 if x0i > 0; otherwise fi ¼ 0Þ.
3. The characteristic length, at the present integration point j with isoparametric coordinates xj and hj and cracking angle qj is
obtained as
!1 nc
!1
vf xj ; hj X vNi xj ; hj vNi xj ; hj
l xj ; hj ¼ ¼ cos qj þ sin qj fi [107]
vx0 i¼1
vx vy
Figure 6 Finite element band modeling.
Figure 7 Computation of f values at the virtual midplane of the element.

where
8 9 8 9
> vN > >
> vN >
>
>
< vx >
= < vx >
> =
1
¼ Jxy [108]
>
> >
vN > >
> vN >
>
: ; >
: >
;
vy vh
and Jxy is defined as the Jacobian matrix given by
2 3
vx vy
6 vx vx 7
6 7
Jxy ¼ 6 7 [109]
4 vx vy 5
vh vh
where the partial derivatives with respect to the isoparametric coordinates are written as
vx 1 1 1 1
¼ ð1 þ hÞx1 ð1 þ hÞx2 ð1 hÞx3 þ ð1 hÞx4 [110]
vx 4 4 4 4
vx 1 1 1 1
¼ ð1 þ xÞx1 þ ð1 xÞx2 ð1 xÞx3 ð1 þ xÞx4 [111]
vh 4 4 4 4
vy 1 1 1 1
¼ ð1 þ hÞy1 ð1 þ hÞy2 ð1 hÞy3 þ ð1 hÞy4 [112]
vx 4 4 4 4
vy 1 1 1 1
¼ ð1 þ xÞy1 þ ð1 xÞy2 ð1 xÞy3 ð1 þ xÞy4 [113]
vh 4 4 4 4
where the pairs (xi,yi) refer to the global coordinates of the virtual nodes defining the midplane of the element.
For transverse compression failure, a set of cracked elements is determined by using the stress-based criterion defined by eqn [37]
and the cracked plane is defined by the fracture angle qf. The function q is given by
X
nc
fðx; zÞ ¼ Ni ðx; zÞfi [114]
i¼1
where nc is the number of corner nodes of a virtual plane located at the midplane of the element according to Figure 8, Ni are the
linear shape functions defined previously, nc are virtual nodes defining the virtual cracking midplane, and fi is the value of the f at
corner i.
x0 ,z0 is an auxiliary coordinate system defined at the center of the element, this being identified by the values of the isoparametric
coordinates (x ¼ 0, n ¼ 0, and z ¼ 0) with the direction of the local axis x0 is defined by the normal to the fracture plane,
0

xi cos qj sin qj xi
0 ¼ [115]
zi sin qj cos qj zi
Node 8
Node 5 Node 7
Node 6
Node 1 Node 3
Node 2
Figure 8 Computation of the characteristic length for transverse compression.

The values of f at each corner node are established according to their position with respect to the local axes x0 and z0
ðfi ¼ 1 if x0i 0; otherwise fi ¼ 0Þ in a similar way as the one described previously.
The characteristic length associated with transverse compression at the present integration point j with isoparametric coordinates
and xj and zj fracture angle qf is given by
nc
!1

vf xj ; zj 1 X vNi xj ; zj vNi xj ; zj
l xj; zj ¼ ¼ cos qf þ sin qf fi [116]
vx0 i¼1
vx vz
where
8 9 8 9
> vN > > vN >
>
< >
= >
< vx >
=
vx 1
¼ Jxz [117]
> > > >
: vN >
> ; : vN >
> ;
vz vz
and Jxz is defined as the Jacobian matrix given by
2 3
vx vz
6 vx vx 7
6 7
Jxz ¼ 6 7 [118]
4 vx vz 5
vz vz
where the partial derivatives with respect to the isoparametric coordinates are written as
vx 1 1 1 1
¼ ð1 þ zÞx1 ð1 þ zÞx2 ð1 zÞx3 þ ð1 zÞx4 [119]
vx 4 4 4 4
vx 1 1 1 1
¼ ð1 þ xÞx1 þ ð1 xÞx2 ð1 xÞx3 ð1 þ xÞx4 [120]
vz 4 4 4 4
vz 1 1 1 1
¼ ð1 þ zÞz1 ð1 þ zÞz2 ð1 zÞz3 þ ð1 zÞz4 [121]
vx 4 4 4 4
vz 1 1 1 1
¼ ð1 þ xÞz1 þ ð1 xÞz2 ð1 xÞz3 ð1 þ xÞz4 [122]
vz 4 4 4 4
where the pairs (xi,zi) refer to the global coordinates of the virtual nodes defining the midplane of the element. In-plane shear
cracking is strongly dependent on the fiber orientation within the element. Therefore, the characteristic length associated with in-
plane shear failure has been assumed to be the same as the one defined for fiber failure or failure in the warp direction. For out-of-
plane shear failure modes, cracks are assumed to be smeared over the thickness of the element with a crack band defined between
upper and lower faces of the element, which is equivalent to assume qf ¼ 90 in eqn [116]:
nc
!1
vf xj ; zj 1 X vNi xj ; zj
l xj ; zj ¼ ¼ fi [123]
vx0 i¼1
vz
2.07.3 Interlaminar Damage Modeling
Delamination is an interlaminar failure mode (quite often a precursor to ultimate failure) that takes place between two layers of
dissimilar orientation within the laminate due to high interlaminar stresses acting on the interface between upper and lower layers
defining the interface. The delamination failure modes are usually classified into mode I, mode II, mode III, and mixed-mode
delamination modes, according to the predominant stresses acting on the interface. For instance, for mode I, also defined as
opening mode, the delamination is exclusively due to the through-thickness tensile normal stress, which leads to layer debonding in
the direction normal to the interface. Modes II and III are related to the out-of-plane shear stresses that result in relative sliding
between upper and lower layers. Mixed-mode delamination is a combination of modes I, II, and III.
Different techniques for delamination modeling have been proposed by many researchers in recent years. Approaches based on
stress criteria like those proposed by Lee (3), Kim and Soni (55), Brewer and Lagace (17), Liu and coworkers (18), and Jen et al. (56)
are suitable to model the initiation of delamination. However, they do not predict delamination growth realistically. Moreover, they
require a precise calculation of stresses, and usually the stresses are singular at the crack tip or free edge. Therefore, the determination
of stresses using finite element models becomes mesh dependent. Also, as discussed before, they do not give any information about
the delamination mode involved in the failure process.
Numerical approaches based on fracture mechanics require an initial flaw, and they are used in conjunction with techniques
such as the Virtual Crack Closure (VCC) method for determining the strain energy release rate. The VCC method is based on Irwin’s
assumption that when a crack extends by a small amount, the energy released in the process is equal to work required to close the
crack to its original length. The energy release rates can then be computed from the nodal forces and displacements obtained from
the solution of the finite element model, and crack propagation is simulated by advancing the crack front when the local energy
release rate rises to a critical value (57). The method predicts well the delamination growth. However, as aforementioned, the
structure must be precracked, and different meshes are required for each delamination front as the crack advances.
An alternative and efficient way for delamination modeling that has been widely reported in the literature is to use interface
elements. Interface elements offer the possibility of coupling stress–based criteria and fracture mechanics–based criteria in a unified
way. Therefore, they enable the model to predict both initiation and growth of delamination. For bidimensional problems, interface
elements can be defined as a one-dimensional entity inserted between two adjacent layers. In a similar way, they can be extended to
three-dimensional problems in which the one-dimensional element is replaced by a two-dimensional element connecting adjacent
layers. In elastic cases, the interface elements are very stiff in order to ensure the transference of displacement and traction between
the adjacent layers. To model delamination growth, an interfacial material behavior is assumed to control the relative displacements
and traction between layers; as soon as certain failure criteria are fulfilled, the delamination is allowed to initiate and propagate.
Mi, Crisfield, and Davies (24) proposed a continuous interface element for delamination modeling in fiber composites. The
interface element was embedded between two eight-noded isoparametric plane strain elements. A bilinear softening stress–relative
displacement relationship was assumed for the interface material model, and linear and quadratic interaction criteria were used for
mixed-mode prediction. For unloading conditions, a simple elastic-damage model was adopted in which the material is assumed to
unload directly toward the origin. Excellent agreement was obtained between simulations, experimental, and closed-form solutions
for mode I, mode II, and mixed-mode delamination.
Wisnom and Chang (26) modeled splitting and delamination in notched cross-ply laminates by using nonlinear springs. An
elasto-perfectly plastic material model was assumed for the interfacial springs. According to Wisnom and Chang’s approach, the
springs are defined with a high initial stiffness in the elastic regime; if a certain value of force is reached, further displacements are
allowed to take place at a constant force until the spring breaks. The area under the force–displacement curve of the spring element
divided by the corresponding element area is considered as a fracture energy. A good agreement between the predictions and
experimental results was obtained.
Daudeville and Ladeveze (58) proposed a delamination model based on a damage mechanics approach. In their model,
connecting layers were used to represent the resin-rich interface between two adjacent layers. Three internal damage variables were
used in order to describe delamination associated with modes I, II, and III. The authors studied the delamination in the vicinity of
a straight edge of a specimen under static tension or compression. Good correlation between numerical simulations and experi-
mental results was obtained for the prediction of damaged areas and onset strains.
Based on the works proposed by Daudeville and Ladeveze (58) and Crisfield and Davies (24), Camanho and coworkers (27)
proposed mixed-mode decohesion interface elements to model delaminations in composite laminates. The authors obtained
a good correlation between numerical predictions and experimental results for double cantilever beam (DCB), edge-notched four
point bending (4ENF), and mixed-mode bending (MMB) specimens. Their interface element is currently available in ABAQUS FE
code, and it was later implemented into LS-DYNA3D explicit finite element code via user-defined material models within brick
elements by Pinho et al. (59). An alternative version of their model for dynamic delamination modeling in composites was also
proposed by Iannucci (60).
An alternative constitutive model for prediction of mixed-mode delamination growth in composites is presented in this chapter.
The material model formulation is defined in terms of a stress–relative displacement softening law. Based on fracture mechanics
concepts, the area under the curve defined by the constitutive law is equal to the fracture energy or energy per unit of area, and once
this energy is consumed the crack propagates. In order to simulate the mixed-mode delamination, a stress-based criterion is used for
the failure initiation, and interactive mixed-mode criteria are used to predict damage propagation. The numerical predictions
obtained using the proposed model were validated against experimental results for DCB, 4ENF, and MMB specimens.
2.07.4 Formulation
2.07.4.1 Kinematics
Interfacial material behavior is defined in terms of tractions and relative displacements between the upper and lower surfaces
defining the interface. The relative displacement vector is composed of the resultant normal and sliding components defined by the
relative movement between upper and lower surfaces (see Figure 9). For a single integration point hexahedron solid element, the
relative displacement vector can be written in terms of through-thickness normal strain and out-of-plane shear strains, as follows:
T
dT ¼ f u v w gT ¼ h gxz h gyz h εzz [124]
where u ¼ uT ub, v ¼ vT vb, and w ¼ wT wb. h* is the element thickness of the updated geometry. Following the standard
interface element formulation, the uncoupled linear-elastic constitutive law without membrane effects for the contact element can
be written as
8 9 2 38 9
< sI = Kww 0 0 <w=
s ¼4 0 Kuu 0 5 u [125]
: II ; : ;
sIII 0 0 Kvv v
Deformed configuration
wT
I uT
vT vb
z
II ub
III Undefomed configuration
wb
y x
Figure 9 Three-dimensional contact element. Int. J. Impact Eng. 2012, 43, 63–77.
where sI, sII, and sIII are the interfacial stresses between upper and lower surfaces associated with modes I, II, and III delamination,
respectively.
2.07.4.2 Constitutive Laws

Two different constitutive laws have been used to define interfacial material behavior: bilinear softening law and linear-
polynominal softening law. The advantage of the linear-polynominal over the bilinear is that it is numerically more stable due
to its smoothness on both damage initiation and fully failed displacement onsets.
2.07.4.2.1 Mode I Delamination

The interfacial behavior for mode I opening (sI > 0) is given by
sI ¼ Kww ð1 dI ðwÞÞw [126]

where the damage evolution law dI(w) is defined in terms of normal relative displacements. For a bilinear constitutive law, the
expression for dI(w) is given by
wf w0
dI ðwÞ ¼ 1 [127]
wf w0 w
where w0 ¼ s0I =Kww and s0I is the through-thickness interfacial strength in mode I. Kww is the through-thickness interfacial stiffness
given in terms of the adhesive through-thickness Young’s modulus, that is, Kww ¼ Ezz/h0, where h0 is the initial thickness of the
element associated with the undeformed configuration.
The strain energy release rate associated with mode I delamination is defined by the area underneath the stress–relative
displacement, which in turn is defined by the bilinear constitutive law. That is,
Zwf
s0I
GIc sI dw ¼ [128]
2wf
0
where wf is the relative displacement in which the interfacial stress in mode I is equal to zero (complete decohesion). From eqn
[128], wf can be written in terms of the strain energy release rate as follows:
2GIc
wf ¼ [129]
s0I
For a linear-polynomial constitutive law, the expression for damage evolution law dI(w) is written as
w0
dI ðwÞ ¼ 1 1 þ k2I ðwÞð2kI ðwÞ 3Þ [130]
w
with
w w0
kI w ¼ [131]
wf w0
The derivation of the damage law for a linear-polynomial constitutive law enforces the areas under the stress–relative
displacement defined for both bilinear and linear polynomials to be the same.
The material behavior in compression (sI 0) is assumed to be linear elastic in order to avoid element interpenetration:
sI ¼ Kww w [132]
2.07.4.2.2 Modes II and III Delamination

Similarly to mode I, the interfacial behavior is defined in terms of resultant shear stress–resultant sliding displacement for both
mode II and mode III delamination; that is,
sII ¼ Kuu ð1 dII ðuÞÞu [133]
sIII ¼ Kvv ð1 dIII ðvÞÞv [134]

with the damage evolutions for the bilinear constitutive law given by
uf u0
dII ðuÞ ¼ 1 [135]
uf u0 u
vf v0
dIII ðvÞ ¼ 1 [136]
vf v0 v
where u0 ¼ s0II =Kuu and s0II is the interfacial transverse shear strength in the X–Z plane. Kuu is the interfacial shear stiffness given in
terms of the adhesive shear modulus; that is, Kuu ¼ Gxz/h0 where h0 is the initial thickness of the element associated with the
undeformed configuration. Similarly, v0 ¼ s0III =Kvv and s0III is the transverse shear strength in the Y–Z plane. Kvv is the interfacial
shear stiffness given in terms of shear modulus, that is, Kvv ¼ Gyz/h0.
The linear-polynomial constitutive law results in the following damage evolution laws, for modes II and III, respectively,
u0
dII ðuÞ ¼ 1 1 þ k2II ðuÞð2kII ðuÞ 3Þ [137]
u
where
u u0
kII ðuÞ ¼ [138]
uf u0
Zuf
s0II
GIIc ¼ sII du ¼ [139]
2uf
0
2GIIc
uf ¼ [140]
s0II
and
v0
dIII ðvÞ ¼ 1 1 þ k2III ðvÞð2kIII ðvÞ 3Þ [141]
v
with
v v0
kIII ðvÞ ¼ [142]
vf v0
Zvf
s0III
GIIIc ¼ sIII du ¼ [143]
2vf
0
2GIIIc
vf ¼ [144]
s0III
2.07.4.2.3 Mixed-Mode Delamination

A quadratic stress–based criterion (61) given in the following form was used to detect damage initiation for mixed-mode
delamination:
2 2
maxð0; sI Þ 2 sII sIII
þ þ ¼1 [145]
s0I s0II s0III
The formulation enables the prediction of damage propagation within an energy-based framework. For this purpose, two
distinct energy-based failure criteria available in the open literature have been incorporated into the formulation. The first criterion
is an extension of the power law criterion proposed by Wu and coauthors (62), for mixed-mode I/II. The criterion is written in terms
of interactions between the strain energy release rates and interlaminar fracture toughnesses. It also takes into account the
contribution of mode III in the mixed-mode delamination process:
l l
GI GII GIII l
þ þ ¼1 [146]
GIc GIIc GIIIc
The power law criterion obtained from eqn [23] with l ¼ 1 was found to be suited to predict failure of thermoplastic PEEK matrix
composites, while for fabric-based laminates embedded into epoxy resin system l ¼ 2 is recommended (63). The second criterion
incorporated into the formulation is the B–K criterion proposed by Benzeggagh and Kenane (64), which is given by
h
Gs
GIc þ ðGIIc GIc Þ ¼ GI þ Gs [147]
GI þ Gs
where GIc and GIIc are the mode I and mode II interlaminar fracture toughnesses, respectively. GI and Gs are the strain energy release
rates associated with mode I and resultant mode II/III shear delamination, respectively. It is worth mentioning that the B–K criterion
assumes that GIIc ¼ GIIIc , since there are no test standards currently available in the literature to characterize GIIIc . Under mixed-
mode loading, the resultant displacement vector can be written as
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
d ¼ u2 þ v2 þ w2 [148]
where the components of relative displacements vector are illustrated in Figure 10 and then written as

u ¼ d sin b cos a [149]

v ¼ d sin b sin a [150]

w ¼ d cos b [151]
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
where b ¼ acosðmaxð0; wÞ=dÞ and a ¼ acosðuÞ=ds with ds ¼ u2 þ v2 :
Combining eqns [125,148–151] leads to the following expression for the mixed-mode delamination damage onset displace-
ment vector:

1
Kww cosðbÞ 2 Kuu sinðbÞcosðaÞ 2 Kvv sinðbÞsinðaÞ 2 2
d0 ¼ þ þ [152]
s0I s0II s0III
Now writing eqns [128], [139], and [143] in terms of the relative displacement components and substituting them into the
power law criterion (eqn [146]), we obtain, for any mode ratio, the following expression for final resultant displacement associated
with fully debonded interfacial behavior:

l 1
2 Kww cos2 b Kuu sin2 b cos2 a l Kvv sin2 b sin2 a l l
df ¼ þ þ [153]
d0 GIc GIIc GIIIc
The final resultant displacement associated with the fully debonded interfacial behavior based on the B–K criterion is given by
8 9
> " #h >
>
< 2 2 >
=
1
2
ðKuu cosðaÞÞ þ ðKvv sinðaÞÞ
2 GIc þ sin2 ðbÞðGIIc GIc Þ
>
> Kww cos2 ðbÞ þ sin2 ðbÞ ðKuu cosðaÞÞ2 þ ðKvv sinðaÞÞ2 2 >
1
>
: ;
df ¼ h 1 i [154]
d0 Kww cos2 ðbÞ þ sin2 ðbÞ ðKuu cosðaÞÞ2 þ ðKvv sinðaÞÞ2 2
The resultant damage evolution for the bilinear- and linear-polynomial constitutive laws is given, respectively, by the following
expressions,

df d0
dm d ¼ 1 [155]
df d0 d
Figure 10 Resultant relative displacement vector. Int. J. Impact Eng. 2012, 43, 63–77.
Figure 11 Constitutive model behavior for delamination prediction. Int. J. Impact Eng. 2012, 43, 63–77.
d0
dm d ¼ 1 1 þ k2m d 2km d 3 [156]
d
where
d d0
km d ¼ [157]
df d0
and the mixed-mode stress–relative displacement relationships are given by

8 9 2 38 9
< sI = Kww 1 dm d 0 0 <w=

s ¼4 0 Kuu 1 dm d 0 5 u [158]
: II ; : ;
sIII 0 0 Kvv 1 dm d v

fsg ¼ K dm d fdg [159]
The proposed formulation incorporates a consistent single-damage variable dm (d) for all delamination modes. This enables the
prediction of variable mixed-mode delamination growth without knowing a priori the mixity ratio between different delamination
modes. The constitutive model behavior is illustrated in Figure 11.
2.07.5 Applications
2.07.5.1 Mesh Sensitivity Study
The objectivity algorithm, combined with the intralaminar failure model described in Section 2.07.2, was implemented into
ABAQUS/Explicit within single-integration solid elements as the User-defined Material Model, using the VUMAT Fortran
Subroutine. In order to evaluate the model’s mesh sensitivity, a simple coupon test simulation has been carried out. The dimensions
of the composite virtual coupon were 20 10 2 mm2, which represent a small volume of the material under uniaxial stress. The
virtual coupon was discretized using six different mesh densities, three of them being nonstructured meshes shown in Figure 12.
The composite specimen was continuously loaded in the fiber direction under displacement control to mimic a pseudostatic
loading on the coupon, with mechanical properties given in Tables 1–3, assuming a fiber fracture toughness value ¼ 50 kJ m2. For
comparison purposes, the load–displacement responses for all meshes are represented on a single graph where the dissipated energy
is defined by the area underneath the force–displacement curves. Figure 13 compares the structural response obtained using the
different mesh types.
According to Figure 13, the energy dissipated in the formation of crack is clearly mesh insensitive. The structural responses are
almost identical, ensuring control of the energy dissipation regardless of mesh refinement and element topology. Minor differences
may be attributed to rounding errors within the FE code. For explicit dynamic FE codes, the failure location is defined by both
rounding errors in the uniform stress field associated with the viscosity terms used within the FE code to smear the wave over a series
of elements, and the corresponding wave reflections within the FE mesh. These two effects act simultaneously define a band of failed
elements, which in practice intends to mimic the fracture in the real structural component. Figure 14 depicts the failure locations for
the meshes studied in this section.
Figure 12 Meshes used in the mesh sensitivity study: (a) Mesh-1; (b) Mesh-2; (c) Mesh-3; (d) Mesh-4; (e) Mesh-5; and (f) Mesh-6.
Table 1 Mechanical properties of each ply
E11 (GPa) E22 ¼ E33 (GPa) G12 (GPa) G13 (GPa) G23 (GPa) v12 ¼ v13 v23
100 8.11 4.65 4.65 5.0 0.3 0.4
Table 2 Ply strengths
Xt (MPa) Xc (MPa) Yt (MPa) Yc (MPa) S12 (MPa)
2000 1000 100 160 140
Table 3 Intralaminar fracture toughnesses

2
Gtf ðKJ m Þ Gcf ðKJ m 2 Þ Gtm ðKJ m 2 Þ Gcm ðKJ m 2 Þ Gs ðKJ m 2 Þ
100.0 25.0 2.0 2.0 2.0
2.07.5.2 In-Plane Coupon Test Simulations

In this section, the predictions obtained using the proposed failure model are compared against experimental results at the coupon
level in tension compression and shear. The experimental results were taken from Ref. (54) in which the first author tested
composite laminates manufactured using the Resin Infusion under Flexible Tooling process. The mechanical properties of the
material are listed in Tables 1–3.
The dimensions of the coupons used by the author for the tensile tests in the fiber and matrix directions were
200 20 1.35 mm3 and 200 20 2.25 mm3, respectively, with a gauge length of 100 mm for both cases. For the in-plane shear
tests, the specimen had dimensions of 200 20 2.7 mm3, with a gauge length of 100 mm and lay-up of [þ45 /45 ]3. The
material constants c1 and c2 required by the model to predict in-plane shear behavior were obtained by finding the best-fit between
the model and the experimental shear stress and shear strain curves. A comparison between the experimental and numerical shear
stress–shear strain curves, together with the material model parameters c1 and c2, are presented in Figure 15. As the experiments
were carried out quasistatically, a high value in the order of 109 was assigned to c3 in order to neglect the strain rate effects in the
numerical response in shear.
The lay-ups used for the specimens loaded in tension in both fiber and matrix directions were [0 ]3 and [0 ]5, respectively. The
mechanical properties of each layer are listed in Tables 1–3. Each ply with a nominal thickness of 0.45 mm was individually modeled
using a single solid hexahedral linear element through the thickness direction. The tabs behavior was assumed to be orthotropic
linear elastic. The virtual specimens were assumed to be fixed in one end and loaded quasistatically under displacement control using
the dynamic relaxation method; the meshes used in the tensile and compression test simulations are shown in Figures 16 and 17.
45
40
35 Mesh-1
Mesh-2
Mesh-3
30 Mesh-4
Load (kN)
Mesh-5
25 Mesh-6
20
15
10
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Displacement (mm)
Figure 13 Effect of mesh refinement on the structural response.
Figure 14 Failure location.
Figures 18–22 show the correlation between numerical predictions and experimental results for the tension and compression in
both fiber and matrix directions and in-plane shear tests, respectively.
In general, the numerical predictions agree very well with the experimental results for all simulations carried out at coupon level.
Clearly visible in Figure 18 is the stiffening effect in tension in the longitudinal direction. This effect has not been taken into account
in the model development, and it is due to the fabric architecture, which leads to the development of high interlaminar shear
stresses between fibers located in the fill and warp directions during the loading process. This results in a reduction of the initial
crimp angle of the fabric, therefore increasing the laminate stiffness along the loading history.
Comparisons between the experimental and predicted failure locations are shown in Figures 23–26. The numerical predictions
indicate that the use of sharp corners end tabs leads to failure near the tab due to the high-stress concentrations in these regions. This
effect was also observed in the experiments. The prediction of the failure location for transverse compression, Figure 25, correlates
very well with the experiments, indicating that the failure in transverse direction is a shear-dominated failure mode with a fracture
Figure 15 Comparison between numerical and experimental curves for the in-plane shear behavior.
Figure 16 Finite element mesh used in the tensile test simulation.
plane orientated about 50 in respect to the through-the-thickness direction. In order to avoid element distortion problems, the
failed elements in transverse compression have been deleted from the mesh using an erosion algorithm available in the ABAQUS
Finite Element Code.
2.07.5.3 Delamination Coupon Test Simulations

This section presents the results obtained using the delamination constitutive model described previously. The model was
implemented into ABAQUS/Explicit within single-integration solid elements as User-defined Material Model, using the VUMAT
Fortran Subroutine.
2.07.5.3.1 Double Cantilever Beam

The dimensions of the virtual coupon for the DCB consisted of 170 20 3.4 mm, with a lay-up of (0 )8. A typical DCB specimen
is shown in Figure 27. The arms were modeled using four-nodes shell elements available in ABAQUS. A linear orthotropic elastic
material model was used to model the behavior of the arms, and a layer of interface elements with a finite thickness of 0.021 mm
was placed at the midplane of the virtual coupon to represent a resin-rich area between two adjacent layers. The tie constrains option
Figure 17 Finite element mesh used in the compression test simulation.
45000
40000 Numerical
Experimental
35000
30000
Load (N)
25000
20000
15000
10000
5000
0
0 0.0005 0.001 0.0015 0.002 0.0025
Displacement (m)
Figure 18 Comparison between numerical predictions and experimental results for the fiber tensile tests.
available in ABAQUS was used to connect the interface layer and the arms. The initial crack length for the DCB was assumed to be
60 mm. The mechanical properties for the composite arms are given in Table 7, and the mechanical properties for the resin-rich
interface are presented in Tables 4–6. The simulations were carried out quasistatically using the ABAQUS/Explicit FE code according
to the load configuration depicted in Figure 28. A comparison between experimental results and numerical predictions is depicted
in Figure 29. Figure 30 shows the deformed FE model configuration, where the red color regions correspond to the fully
delaminated area with dm ¼ 1.0. The numerical predictions in terms of load–displacement response and damage extent obtained
using the delamination constitutive model agree very well with experimental results.
2.07.5.3.2 Edge-Notched Four-Point Bending

The schematic test fixture and dimensions used for the 4ENF test simulation are shown in Figure 31. The virtual specimens had
dimensions of 170 20 3.4 mm, with a delamination starter length of 40 mm measured from the edge of the specimen. The
dimensions adopted for SL and SR were 20 and 100 mm, respectively, with the total span length (2L) equal to 120 mm. In order to
obtain a better prediction of the out-of-plane shear stress distribution, the composite specimen was modeled using eight-node
2500
Numerical
Experimental
2000
1500
Load (N)
1000
500
0 0.0002 0.0004 0.0006 0.0008

Displacement (m)
Figure 19 Comparison between numerical predictions and experimental results for the matrix tensile tests.
10000
9000
8000
7000
6000
Load (N)
5000
4000
3000
2000 Numerical
Experimental
1000
0
3E-05 6E-05 9E-05 0.00012 0.00015
Displacement (m)
Figure 20 Comparison between numerical predictions and experimental results for the fiber compression tests.
hexahedric solid elements available in ABAQUS. A linear orthotropic elastic material model was used to model the behavior of the
upper and lower parts of the laminate, and a layer of interface elements with a finite thickness of 0.021 mm was placed at the
midplane of the virtual coupon to represent a resin-rich area between two adjacent layers. The tie constrains option available in
ABAQUS was used to connect the interface layer and the arms. The upper and lower parts of the loading frame were assumed to be
rigid, and a surface-to-surface contact logic available in ABAQUS was used to model the contact between the loading frame and the
specimen. The FE model for the 4ENF test simulation is shown in Figure 32. The mechanical properties for the composite arms are
given in Table 7, and the mechanical properties for the resin-rich interface are presented in Tables 4–6. A comparison between
experimental and numerical results is shown in Figure 33.
2.07.5.3.3 Mixed-Mode Bending

The dimensions for the virtual specimens used in the MMB test simulation was 170 20 3.4 mm, with a lay-up of (0 )8 and
a starter crack length of 37 mm. The specimen was loaded according to the schematic figure shown in Figure 34, with a span length
2L of 102 mm.
The adopted relative position between the lever loading roller and the loading saddle/yolk was varied from 35.90 to 96.8 mm in
order to achieve mode ratios (GII/G) of 75, 50 and 25%. Again, the composite specimen was assumed to be linear orthotropic elastic
3500
Numerical
3000 Experimental
2500
Load (N)
2000
1500
1000
500
0
0 5E-05 0.0001 0.00015 0.0002 0.00025 0.0003
Displacement (m)
Figure 21 Comparison between numerical predictions and experimental results for the matrix compression tests.
5000
4000
3000
Load (N)
2000
1000 Numerical
Experimental
0
0 0.0005 0.001 0.0015 0.002 0.0025
Displacement (m)
Figure 22 Comparison between numerical predictions and experimental results for the in-plane shear tests.
Figure 23 Comparison between (a) predicted and (b) experimental failure locations for the tensile tests.
Figure 24 Comparison between (a) predicted and (b) experimental failure locations for the compression test in the fiber direction.
Figure 25 Comparison between (a) predicted and (b) experimental failure locations for the compression test in the transverse direction.
Figure 26 Comparison between (a) predicted and (b) experimental failure locations for the in-plane shear test.
a0 = 60 mm
170 mm
3 mm
20 mm
Figure 27 Dimensions for the DCB specimen.
Table 4 Mechanical properties of the resin-rich interface
Ezz (GPa) Gxz ¼ Gyz (GPa) h0 (m)
2.97 1.08 2.1e-4

Int. J. Impact Eng. 2012, 43, 63–77.
Table 5 Interlaminar strengths
s0I ðMPaÞ s0II ðMPaÞ s0III ðMPaÞ
50.0 100.0 100.0

Int. J. Impact Eng. 2012, 43, 63–77.
Table 6 Interlaminar fracture toughnesses and model parameters
GI c (kJ m2) GII c (kJ m2) GIII c (kJ m2) l h
0.585 3.5 3.5 1.0 1.0

Int. J. Impact Eng. 2012, 43, 63–77.
Table 7 Mechanical properties of the laminate
E11 (GPa) E22 (GPa) G12 (GPa) v12
60.8 58.3 4.55 0.07

Int. J. Impact Eng. 2012, 43, 63–77.
Figure 28 DCB-loading configuration. Int. J. Impact Eng. 2012, 43, 63–77.
50
45
Experimental
40
Numerical
35
30
Load (N)
25
20
15
10
0
0 10 20 30 40 50 60 70
Displacement (mm)
Figure 29 Comparison between numerical and experimental responses for DCB test. Int. J. Impact Eng. 2012, 43, 63–77.
Figure 30 FE model for DCB test. Int. J. Impact Eng. 2012, 43, 63–77.
SL
a0
L L
SR
Figure 31 4ENF-loading configuration. Int. J. Impact Eng. 2012, 43, 63–77.
Figure 32 FE model for 4ENF test.
1500
1400
1300 Experimental
1200 Numerical
1100
1000
Load (N)
900
800
700
600
500
400
300
200
100
0
0 0.001 0.002 0.003 0.004 0.005
Displacement (m)
Figure 33 Comparison between numerical and experimental responses for 4ENF test.
ca
Saddle and yoke

Fulcrum
2h
a
L L
Figure 34 Mixed-mode-loading configuration.
350
Experimental (MMB 25%)
Numerical (MMB 25%)
300 Experimental (MMB 50%)
Numerical (MMB 50%)
Experimental (MMB 75%)
250 Numerical (MMB 75%)
Load (N)
200
150
100
50
0
0 5 10 15
Displacement (mm)
Figure 35 Comparison between numerical and experimental results for MMB tests.
and modeled using four-nodes shell elements available in ABAQUS. A layer of interface elements with a finite thickness of
0.021 mm was placed at the midplane of the virtual coupon to represent a resin-rich area between two adjacent layers. The tie
constrains option available in ABAQUS was used to connect the interface layer and the arms. The MMB upper loading device was
modeled using rigid shell elements. The contact between the upper loading device and MMB specimen was modeled using
constraint equations available in ABAQUS. Comparisons between experimental results and numerical predictions for mode rations
of 75, 50, and 25% are presented in Figure 35. Figure 36 shows the deformed FE model configuration, where the red color regions
correspond to the fully delaminated area with dm ¼ 1.0. A very good agreement for both initiation and damage propagation values
was obtained using the delamination constitutive model.
2.07.5.4 Impact-Induced Damage Prediction

This section presents an application example using the damage models described in the previous sections. The application example
consists of applying the models to predict impact-induced damage in composite laminates. The finite element model consists of
a 102 152 3.6 mm3 rectangular plate, with clamped edges impacted by a steel hemispherical impactor under an impact energy of
18.47 J. The mass and diameter of the steel hemispherical projectile are 1.556 kg and 12.70 mm, respectively. The composite plate has
a cross-ply (0 /90 )s lay-up made of the unidirectional carbon UD tape. Both impactor and plate were modeled using single-
Figure 36 FE model for MMB test.
integration point solid elements available in ABAQUS, and each orthotropic layer with thickness equal to 0.45 mm was individually
modeled using the proposed damage model. The mechanical properties for each composite layer are given in Tables 1–3. A 0.01-mm-
thick interface layer containing the interlaminar constitutive damage model previously described was placed at the plate midplane to
predict delaminated area, as shown in Figure 37. The parameters used for the interlaminar damage model were
Ezz ¼ Gxz ¼ Gyz ¼ 1.0 GPa, s0I ¼ 50 MPa, s0II ¼ s0III ¼ 100 MPa, GIc ¼ 0.6 kJ m2, GIIc ¼ 2.0 kJ m2, GIIIc ¼ 2.0 kJ m2, l ¼ 1.0, and h ¼ 1.0.
A surface-to-surface slide-line contact logic based on the penalty method formulation was defined between the impactor and
composite plate. A viscous hourglass control algorithm was used to avoid spurious hourglass modes arising from the use of under-
integrated elements. Figure 38 shows a comparison between predicted and experimental force–time histories. As seen in the figure,
the numerical prediction using the intra- and interlaminar damage models correlates remarkably well with experiments. The
numerical simulation indicated that compression failure due to high compressive stresses at the contact region, followed by
delamination and failure due to transverse tension at the bottom face, were the predominant failure modes taking place in the
specimens. It can be seen from Figure 39 that the model fully captures the back-face tearing due to tensile bending stress in
the transverse direction. Numerical predictions also indicated that through-thickness failure localization initially occurred near the
midplane of the laminate where the transverse shear stresses were higher, propagating between layers of dissimilar orientation with
a cracking extension direction defined by the fiber orientation in the lower layer at each interface as a result of the combination of
Figure 37 FE mesh used for the impact simulations (a) complete FE mesh, (b) mesh detail in the impact region.
5000
4500 Experimental
Numerical
4000
3500
3000
Force (N)
2500
2000
1500
1000
500
0
0 1 2 3 4 5
Time (ms)
Figure 38 Comparison between numerical and experimental contact force–time histories.
Figure 39 Predominant failure mode for cross-ply laminates: transverse tensile failure at the bottom face (left: predicted; right: experimental).
tensile matrix cracking and transverse shear cracking. It can be seen that the predicted delamination area compares very well with
those obtained from C-Scan images, as shown in Figure 40. Figure 40(a) depicts the layer of interface elements placed at the plate
midplane, where the fully damaged elements with dm ¼ 1.0 were removed from the FE mesh.
2.07.6 Summary and Conclusions
This chapter presented a damage modeling methodology for composite structures. The formulations enable the prediction of failure
modes in composites within an energy-based framework, avoiding both pathological strain localization and mesh dependence
problems that arise from the use of orthotropic strain-softening constitutive laws. The accuracy of the models was checked by
comparing numerical predictions with experimental results in different loading ranges varying from static to dynamic impact
loading. The damage models also predicted most of the features experimentally observed, including fracture plane direction and
damage extent with good accuracy. The inclusion of shear nonlinearities and irreversible strains, coupled with damage, led to
realistic predictions of the structural behavior of composites under shear loading.
The interlaminar constitutive damage model formulation incorporates both bilinear- and linear-polynomial softening laws,
where the main advantage of the linear polynomial over the bilinear is the numerical stability at both damage initiation and fully
damaged stress onsets. Prediction of the onset of delamination is based on an interactive stress-based criterion. Damage progression
is also predicted by means of a single damage variable within an energy-based context by using two distinct interactive energy
Figure 40 Comparison between (a) predicted and (b) experimental delamination area in red color.
criteria. The interlaminar damage model formulation also includes interactions between modes I, II, and III without knowing
a priori the mode mixity ratio. Very good correlation between experimental and numerical predictions in terms of load–
displacement responses and damage extent for DCB, 4ENF, and MMB specimens was obtained using the interlaminar constitu-
tive model.
Finally, the accuracy of the models was checked by simulating impact-induced damage in composite laminates. The predictions
in terms of peak load, stiffness degradation, absorbed energy, and damage extent obtained using the inter- and intralaminar
constitutive models correlates remarkably well with experimental results.
In summary, the combination of intra- and interlaminar damage models previously described allows a realistic prediction of the
experimentally observed failure modes in composite laminates.
Acknowledgments
The authors acknowledge the financial support received for this work from Fundação de Amparo Pesquisa do Estado de São Paulo
(Fapesp), contract numbers 2006/06808-6, 2007/02710-4 and CNPq Grant 305601/2007-5.
References
1. Jones, Robert M. Mechanics of Composite Materials, 2nd ed.; Taylor and Francis, Inc., 1999.
2. Tsai, S. W.; Wu, E. A General Theory of Strength for Anisotropic Materials. J. Compos. Mater. 1971, 5, 58–72.
3. Lee, J. D. Three-Dimensional Finite Element Analysis of Layered Fiber-Reinforced Composite Materials. Comput. Struct. 1980, 12, 319–333.
4. Engblom, J. J.; Havelka, J. J. Transient Response Predictions for Transversely Loaded Laminated Composite Plates. AIAA Paper 89-1302-CP, 1989.
5. Hashin, Z. Failure Criteria for Uni-Directional Fibre Composites. J. Appl. Mech. 1980, 47 (1), 329–334.
6. Chang, F.-K.; Chang, K.-Y. Post-Failure Analysis of Bolted Composite Joints in Tension or Shear-Out Mode Failure Mode. J. Compos. Mater. 1987, 21, 809–833.
7. Chang, F.-K.; Chang, K.-Y. A Progressive Damage Model for Laminated Composites Containing Stress Concentrations. J. Compos. Mater. 1987, 21, 834–855.
8. Shivakumar, K. N.; Elber, W.; Illg, W. Prediction of Low-Velocity Impact Damage in Thin Circular Laminates. AIAA J. 1985, 23 (3), 442–449.
9. Choi, H. Y.; Wu, H. Y. T.; Chang, F. K. A New Approach toward Understanding Damage Mechanics and Mechanics of Laminated Composites due to Low-Velocity Impact:
Part II – Analysis. J. Compos. Mater. 1991, 25, 1012–1038.
10. Davies, G. A. O.; Zhang, X. Impact Damage Prediction in Carbon Composite Structures. Int. J. Impact Eng. 1995, 16 (1), 149–170.
11. Watson, S. A. The Modelling of Impact Damage in Kevlar-Reinforced Epoxy Composite Structures. Ph.D. Thesis, Department of Aeronautics, Imperial College of Science
Technology and Medicine, 1993.
12. Davies, G. A. O.; Hitchings, D.; Zhou, G. Impact Damage and Residual Strengths of Woven Fabric Glass/Polyester Laminates. Compos. Part A 1996, 27 (A), 1147–1156.
13. Wiggenraad, J. F. M.; Zhang, X.; Davies, G. A. O. Impact Damage Prediction and Failure Analysis of Heavily Loaded, Blade-Stiffened Composite Wing Panels. Compos. Struct.
1999, 45, 81–103.
14. Zhang, X.; Davies, G. A. O.; Hitchings, D. Impact Damage with Compressive Preload and Post-Impact Compression of Carbon Composite Plates. Int. J. Impact Eng. 1999, 22, 485–509.
15. Hou, J. P.; Petrinic, N.; Ruiz, C.; Hallett, S. R. Prediction of Impact Damage in Composite Plates. Compos. Sci. Technol. 2000, 60, 273–281.
16. Hou, J. P.; Petrinic, N.; Ruiz, C. A Delamination Criterion for Laminated Composites under Low-Velocity Impact. Compos. Sci. Technol. 2001, 61, 2069–2074.
17. Brewer, J. C.; Lagace, P. A. Quadratic Stress Criterion for Initiation of Delamination. J. Compos. Mater. 1988, 22 (12), 1141–1155.
18. Liu, S.; Kutlu, Z.; Chang, F. K. Matrix Cracking-Induced Delamination Propagation in Graphite/Epoxy Laminated Composites due to a Transverse Concentrated Load. ASTM
STP 1156. In Comp Materials: Fatigue and Fracture; Stinchcomb, W. W., Ashbaugh, N. E., Eds.; American Society for Testing and Materials: Philadelphia, 1993; Vol. 4, pp
86–101.
19. Iannucci, L. Progressive Failure Modelling of Woven Carbon Composite under Impact. Int. J. Impact Eng. 2006, 32 (6), 1013–1043.
20. Soutis, C.; Curtis, P. T. A Method for Predicting the Fracture Toughness of CFRP Laminates Failing by Fibre Microbuckling. Compos. Part A 2000, 31, 733–740.
21. Soutis, C.; Smith, F. C.; Matthews, F. L. Predicting the Compressive Engineering Performance of Carbon Fibre-Reinforced Plastics. Compos. Part A 2000, 31, 531–536.
22. Zhuk, Y.; Guz, I.; Soutis, C. Compressive Behaviour of Thin-Skin Stiffened Composite Panels with a Stress Raiser. Compos. Part B 2001, 32, 697–709.
23. Hawyes, V. J.; Curtis, P. T.; Soutis, C. Effect of Impact Damage on the Compressive Response of Composite Laminates. Compos. Part B 2001, 32, 1263–1270.
24. Mi, Y.; Crisfield, M. A.; Davies, G. A. O. Progressive Delamination using Interface Elements. J. Compos. Mater. 1998, 32 (14), 1247–1271.
25. Chen, J.; Crisfield, M. A.; Kinloch, A. J.; Busso, E. P.; Matthews, F. L.; Qiu, Y. Predicting Progressive Delamination of Composite Material Specimens via Interface Elements.
Mech. Compos. Mater. Struct. 1999, 6, 301–317.
26. Wisnom, Michael R.; Chang, Fu-Kuo. Modelling of Splitting and Delamination in Notched Cross-Ply Laminates. Compos. Sci. Technol. 2000, 60, 2849–2856.
27. Camanho, P. P.; Davila, C. G. Mixed-Mode Decohesion Finite Elements for the Simulation of Delamination in Composite Materials. NASA/TM-2002–211737, 2002.
28. Vaziri, R.; Olson, M. D.; Anderson, D. L. Damage in Composites: A Plasticity Approach. Comput. Struct. 1992, 44, 103–116.
29. Kachanov, L. M. Time of Rupture Process under Creep Conditions. Izy Akad Nank S.S.R. Otd Tech Nauk 1958, 8, 26–31.
30. Rabotnov, Y. N. Creep Rupture. In Proc. XII Int. Cong. Appl. Mech., Standford-Springer, 1968.
31. Ladeveze, P.; Le Dantec, E. Damage Modelling of the Elementary Ply for Laminated Composites. Compos. Sci. Technol. 1992, 43, 257–267.
32. Johnson, A. F. Modelling Fabric Reinforced Composites under Impact Loads. Composites 2001, 32, 1197–1206.
33. Williams, Kevin V.; Vaziri, Reza. Application of a Damage Mechanics Model for Predicting the Impact Response of Composite Materials. Comput. Struct. 2001, 79, 997–1011.
34. Matzenmiller, A.; Lubliner, J.; Taylor, R. L. A Constitutive Model for Anisotropic Damage in Fiber-Composites. Mech. Mater. 1995, 20 (2), 125–152.
35. Williams, Kevin V.; Vaziri, Reza; Poursartip, Anoush. A Physically Based Continuum Damage Mechanics Model for Thin Laminated Composite Structures. Int. J. Solids Struct.
2003, 40, 2267–2300.
36. Williams, K. V. Simulation of Damage Progression in Laminated Composite Plates Using LS-DYNA. In 5th. Int. LS-DYNA Conf., Southfield, Michigan, September 21–22, 1998.
37. Pinho, S. T.; Robinson, P.; Iannucci, L. Physically-Based Failure Models and Criteria for Laminated Fibre-Reinforced Composites with Emphasis on Fibre Kinking: Part I:
Development. Compos. Part A 2006, 37 (1), 63–73.
38. Davila, C. G.; Camanho, P. P. Physically Based Failure Criteria for FRP Laminates in Plane Stress; NASA-TM, 2003.
39. Donadon, M. V.; de Almeida, Sergio F. M.; Arbelo, Mariano A.; de Faria, Afredo R. A Three-Dimensional Ply Failure Model for Composite Structures. Int. J. Aerospace Eng.
2009, 1, 1–22.
40. Puck, A.; Shurmann, H. Failure Analysis of FRP Laminates by Means of Physically Based Phenomenological Models. Compos. Sci. Technol. 2002, 62, 1633–1662.
41. Voyiadjis, G. Z.; Kattan, P. I. Advances in Damage Mechanics: Metals and Metals Matrix Composites; Elsevier: Amsterdam, The Netherlands; New York, 1999.
42. Kattan, P. I.; Voyiadjis, G. Z. Damage Mechanics with Finite Elements: Practical Applications with Composite Tools; Springer: Berlin, Heidelberg, NY, 2001.
43. Donadon, M. V.; Iannucci, L.; Falzon, Brian G.; Hodgkinson, J. M.; de Almeida, Sergio F. M. A Progressive Failure Model for Composite Laminates Subjected to Low Velocity
Impact Damage. Comput. Struct. 2008, 86, 1232–1252.
44. Donadon, M. V.; Falzon, B. G.; Iannucci, L.; Hodgkinson, John M. Intralaminar Toughness Characterisation of Unbalanced Hybrid Plain Weave Laminates. Compos. Part A 2007,
38, 1597–1611.
45. Maimi, P.; Camanho, P. P.; Mayugo, J. A.; Davila, C. G. A Continuum Damage Model for Composite Laminates: Part II – Computational Implementation and Validation. Mech.
Mater. 2007, 39, 909–919.
46. Soden, P.; Kaddour, A. S. Biaxial Test Results for Strength and Deformation of a Range of e-Glass and Carbon Fibre Reinforced Composite Laminates: Failure Exercise
Benchmark Data. Compos. Sci. Technol. 2002, 62, 1489–1514.
47. Pinho, S. T.; Robinson, P.; Iannucci, L. Physically-Based Failure Models and Criteria for Laminated Fibre-Reinforced Composites with Emphasis on Fibre Kinking: Part II: FE
Implementation. Compos. Part A 2006, 37 (5), 766–777.
48. Bazant, Z. P. Crack Band Theory for Fracture of Concrete. Materiaux et Constructions 1983, 16 (93).
49. Puck, A.; Schurmann, H. Failure Analysis of FRP Laminates by Means of Physically Based Phenomenological Models. Compos. Sci. Technol. 1998, 58, 1045–1067.
50. Vural, M. Transverse Failure in Thick S2-Glass/Epoxy Fiber-Reinforced Composites. J. Compos. Mater. 2004, 38 (7), 609–623.
51. Donadon, M. V.; Iannucci, L. Bird Strike Modelling Using a New Woven Glass Failure Model. In LS-DYNA Int. Conf., Michigan, Dearborn, USA, 2006.
52. Pijaudier-Cabot, G. Comparison of Various Models for Strain Softening. Eng. Computat. 1988, 5, 141–150.
53. Oliver, J. A Consistent Characteristic Length for Smeared Cracking Model. Int. J. Numer. Meth. Eng. 1989, 28, 461–474.
54. Donadon, M. V. The Structural Behaviour of Composite Structures Manufactured Using Resin Infusion under Flexible Tooling. Ph.D. Thesis, Deptartment of Aeronautics, Imperial
College London, United Kingdom, 2005.
55. Kim, R. Y.; Soni, S. R. Experimental and Analytical Studies on the Onset of Delamination in Laminated Composites. J. Compos. Mater. 1984, 18 (4), 70–84.
56. Jen, H. H. R.; Kau, Y. S.; Hsu, J. M. Initiation and Propagation of Delamination in Centrally Notched Composite Laminate. J. Compos. Mater. 1993, 27 (3), 272–285.
57. Davidson, B. D.; Schapery, R. A. A Technique for Predicting Mode I Energy Release Rates Using a First-Order Shear Deformable Plate Theory. Eng. Fract. Mech. 1990, 36, 157–165.
58. Daudeville, L.; Ladeveze, P. A Damage Mechanics Tool for Laminate Delamination. Compos. Struct. 1993, 25, 547–555.
59. Pinho, S. T; Robinson, P.; Iannucci, L. Formulation and Implementation of Decohesion Elements in an Explicit Finite Element Code. Compos. Part A 2006, 37, 778–789.
60. Iannucci, L. Dynamic Delamination Modelling Using Interface Elements. Comput. Struct. 2006, 84, 1029–1048.
61. Wisnom, M. R.; Cui, W.; Jones, M. Criteria to Predict Delamination of Unidirectional Glass/Epoxy Specimens Waisted through the Thickness. Composites 1992, 23 (3), 158–166.
62. Wu, E. M.; Reuter, R. C., Jr. Crack Extension in Fiberglass Reinforced Plastics. T. and AM Report No. 275, University of Illinois, 1965.
63. Donadon, M. V. The Structural Behaviour of Composite Laminates Manufactured Using Resin Infusion under Flexible Tooling Process. Ph.D. Thesis, Deptartment of Aeronautics,
Imperial College, London, 2005.
64. Benzeggagh, M. L.; Kenane, M. Measurement of Mixed-Mode Delamination Fracture Toughness of Unidirectional Glass/Epoxy Composites with Mixed-Mode Bending Apparatus.
Compos. Sci. Technol. 1996, 56, 439–449.

PHD Forestry

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PHD Forestry

Uploaded by

Copyright:

Available Formats

2.

07 Damage Modeling in Composite Structures

2.07.1 Introduction 111

a) The failure criteria approach

Comprehensive Materials Processing, Volume 2 http://dx.doi.org/10.1016/B978-0-08-096532-1.00210-7 111

2.07.2 Intralaminar Damage Modeling

2.07.2.1 3-D Orthotropic Stress–Strain Relationship

Eii ¼ Ex0i [5]

The Green-St. Venant strain tensor e is given by

2.07.2.2 Effective Stresses

2.07.2.3 Longitudinal Failure Modes

Figure 1 Damage evolution law and softening behavior.

2.07.2.4 Transverse Failure Modes

where εtm is deﬁned as the resultant strain, which is given by

stm;0 ¼ stm jF t ¼1 [25]

st ¼ stm sinðqÞ [29]

gts;m ¼ εtm sinðqÞ [31]

Figure 2 Action plane for transverse compression failure mode.

scm;0 ¼ scm jF c ¼1 [54]

2.07.2.5 In-Plane Shear Failure

g12,f is written in terms of the intralaminar toughness in shear,

2.07.2.6 Smeared Cracking Approach

2.07.2.7 Objectivity Algorithm

2.07.2.7.1 Crack Modeling in the Continuous Medium

2.07.2.7.2 Displacement and Traction Vectors in the Singular Band

which is constant in the x0 direction, that is,

2.07.2.7.3 Energy Dissipation within the Band

and taking into account eqns [81] and [85] we obtain,

The local form of eqn [97] is

sd ¼ sð1 dðεÞÞ [101]

2.07.2.7.4 Determination of the Function f and the Characteristic Length l*

Figure 5 Determination of the characteristic length for hexahedron elements.

Figure 6 Finite element band modeling.

Figure 7 Computation of f values at the virtual midplane of the element.

Figure 8 Computation of the characteristic length for transverse compression.

2.07.3 Interlaminar Damage Modeling

2.07.4.2 Constitutive Laws

2.07.4.2.1 Mode I Delamination

sI ¼ Kww ð1 dI ðwÞÞw [126]

2.07.4.2.2 Modes II and III Delamination

sIII ¼ Kvv ð1 dIII ðvÞÞv [134]

2.07.4.2.3 Mixed-Mode Delamination

and the mixed-mode stress–relative displacement relationships are given by

Table 1 Mechanical properties of each ply

100 8.11 4.65 4.65 5.0 0.3 0.4

Table 2 Ply strengths

Xt (MPa) Xc (MPa) Yt (MPa) Yc (MPa) S12 (MPa)

2000 1000 100 160 140

Table 3 Intralaminar fracture toughnesses

100.0 25.0 2.0 2.0 2.0

2.07.5.2 In-Plane Coupon Test Simulations

Figure 13 Effect of mesh reﬁnement on the structural response.

Figure 14 Failure location.

Figure 16 Finite element mesh used in the tensile test simulation.

2.07.5.3 Delamination Coupon Test Simulations

2.07.5.3.1 Double Cantilever Beam

Figure 17 Finite element mesh used in the compression test simulation.

2.07.5.3.2 Edge-Notched Four-Point Bending

0 0.0002 0.0004 0.0006 0.0008

2.07.5.3.3 Mixed-Mode Bending

Figure 27 Dimensions for the DCB specimen.

Table 4 Mechanical properties of the resin-rich interface