DELAMINATION ANALYSIS OF COMPOSITES USING A

FINITE ELEMENT BASED DISCRETE DAMAGE ZONE

MODEL

Xia Liu, Ravindra Duddu, Haim Waisman

Columbia University
610 Seeley W. Mudd Building, 500 West 120th Street
New York, NY 10027

ABSTRACT
We perform delamination analysis in laminated composites in 2D and 3D using the discrete
damage zone model within the framework of the finite element method. In this approach, spring-
like elements are placed at the laminate interface and damage laws are used to prescribe both
interfacial spring softening and bulk material stiffness degradation to study crack propagation.
The irreversibility of damage naturally accounts for the subsequent reduction of material
stiffness once the material is loaded beyond the elastic limit. The model is implemented in the
commercially available ABAQUS software via the user element subroutine (UEL). Numerical
results for 2D mixed-mode and 3D mode-I delamination are presented. The results for the
benchmark examples show good agreement with those obtained from virtual crack closure
technique (VCCT), which validates the method. This discrete method is particularly suitable
when the material nonlinearities in the continuum surrounding the crack tip are significant.

1. INTRODUCTION
Delamination and debonding are the two major failure modes in laminated composites [1], but,
often damage may accumulate in the weaker matrix leading to matrix cracking. Therefore,
specialized methods are needed for the numerical analysis of fracture and delamination in
composites so as to obtain reasonable results. To this end, we have recently proposed a discrete
damage zone model within the frame of the finite element (FE) method that is further explored
herein. This method falls into the class of the cohesive zone models (CZMs), however, it is
different from the widely popular continuum cohesive zone models (CCZMs), which is
addressed later in this section.

The numerical approaches used to study fracture and delamination in composites may be broadly
classified into: (i) linear elastic fracture mechanics (LEFM) based [2, 3] and (ii) the cohesive
zone modeling based. The LEFM approaches have been widely used to tackle delamination
problems due to its simplicity and effectiveness [4– 6]; however, such an approach requires an
initial crack and a negligible size of nonlinear zone at the crack tip [7]. In order to overcome
these limitations, the cohesive zone modeling approach was proposed initially by Barenblatt and
Dugdale [8, 9]. The concept introduces a bounded stress within the cohesive zone at the crack
tip. This method has been widely used in simulations for composite materials and dynamic crack
growth [10–13]. However, there are some inevitable computational issues associated with the
CCZMs, such as the shape of the softening region of the cohesive laws and the integration
scheme may cause convergence difficulties, oscillation and mesh sensitivity [11–13].
In order to deal with the computational issues associated with the CCZMs, discrete interfacial
spring elements were introduced by Cui and Wisnom [14], instead of continuum cohesive
elements. Later, Xie and Waas [15, 16] developed this idea by proposing the discrete cohesive
zone model (DCZM), which employs bilinear traction-separation laws as constitutive laws in the
discrete spring elements at the finite element nodes of the interface. Their approach is shown to
overcome the computational issues of the continuous CZMs. However, this discrete cohesive
zone model does not consider the material nonlinearities in the continuum finite elements
surrounding the crack tip and so it precludes matrix cracking as a failure mode. Further, this
DCZM has only been tested for bilinear traction separation laws and the corresponding negative
tangent stiffness of the interface element may lead to convergence issues.

The current discrete damage zone model (DDZM) is different from the above mentioned models
in three respects: (i) it employs spring-like elements at FE nodes whose softening behavior is
given by a damage law; (ii) it considers damage of the surrounding continuum so as to include
the effects of matrix cracking on the macroscopic mechanical behavior; and (iii) it uses the non-
negative secant modulus in order to avoid convergence issues. The DDZM defines the force-
separation law of the interfacial spring elements based on continuum damage mechanics and not
particularly from any cohesive zone model. For homogeneous isotropic material, the same
damage law is implemented in both the interfacial spring and bulk continuum elements; whereas,
for a laminated composite, the damage law or parameters may be different in the springs and in
the continuum. This paper is organized as follows: Section 2 describes the DDZM formulation,
the procedure for parameter identification and the numerical implementation. Section 3 presents
the results for two benchmark examples involving 3D mode-I and 2D mixed-mode delamination
and discusses their convergence to the LEFM solution. In addition, we present results of fracture
analysis under a three-point bending test to study fracture in adhesive joints. Section 4 details the
conclusions and the directions of future research.

2. FORMULATION AND IMPLEMENTATION

2.1 Discrete Interface Element

The proposed DDZM is developed from the perspective of damage mechanics. Therefore, the
material failure is interpreted as stiffness degradation due to damage evolution in both interface
and bulk elements. Figure 1 shows a typical mesh for the double cantilever beam (DCB) with
standard quadrilateral finite elements in the continuum and with interface elements at the finite
element nodes along the interface. These interface elements are multi-dimensional spring like
elements resisting displacement along and perpendicular to their axis. In this article, the analysis
needs only two-dimensional springs and the implementation is detailed in Section 2.3.

Now, let us derive the force-separation law for a one-dimensional spring element by considering
the damage model proposed by Mazars [17] for brittle materials,

𝛿 𝐴 [1]
𝐷 = 1 − (1 − 𝐴) −
𝛿 exp (𝐵(𝛿 − 𝛿 ))

where 𝐷 is the damage variable, 𝐴 and 𝐵 are parameters, 𝛿 is the interfacial separation, and 𝛿
is the critical interface separation required to initiate damage. We assume the parameter 𝐴 = 1.0
 Interface

Figure 1. Schematic diagram showing the finite element mesh for the mode-I delamination
analysis of a double cantilever beam. The ‘x’s indicate the integration (Gauss) points and the
dots indicate the finite element nodes and the dashed line indicates the delamination surface.

for simplicity of the expressions without loss of generality. The behavior of this damage model is
shown in Figure 2(a). Using equation 1, the force–separation (𝐹 − 𝛿) relation at a point 𝛿 ∗ ∈
(𝛿 , 𝛿 ), may be written from the 𝐹 − 𝛿 curve’s geometry shown in Figure 2(b) as,

𝐾 𝛿∗ [2]
𝐹 ∗ = 𝐾 ∗ 𝛿 ∗ = (1 − 𝐷)𝐾 𝛿 ∗ =
exp (𝐵(𝛿 ∗ − 𝛿 ))

where 𝐾 is the initial undamaged stiffness of the spring and 𝛿 is the maximum separation
in the spring just before complete failure (rupture). Note that for the current exponential model
𝛿 = ∞. Thus, the general one-dimensional force–separation law in the spring element with
exponential softening is given by,

𝐾 𝛿, if 𝛿 ∈ [0, 𝛿 ]
𝐹= 𝐾 𝛿
, if 𝛿 ∈ ( 𝛿 , ∞] [3]
exp 𝐵(𝛿 − 𝛿 )

(a) Damage vs. separation (b) Force vs. separation

Figure 2. The damage and mechanical behavior of the current discrete interface element.
Thus, the procedure to derive the force-separation law from a given damage model is
straightforward. The behavior of the springs in multi-dimensions can be similarly written for
each directional component assuming there is no interaction between the normal and tangential
components. Then, the interaction under mixed-mode fracture can be incorporated using the
combined fracture energy criterion, which is detailed in the Section 2.4. In this article, we
consider only Mazars damage model, however, the formulation is general and other damage
models may also be employed to derive appropriate force–separation laws.

2.2 Parameter Identification

The parameters that need to be identified include the critical separation for damage
initiation, 𝛿 , the undamaged spring stiffness, 𝐾 , and the damage coefficient, 𝐵, along and
perpendicular to the axis of the spring. Now, let us consider a 2D interface element and
determine the parameters from the three principles of linear elastic fracture mechanics. First, the
spring forces along (normal) and perpendicular (tangential) to its axis are maximum when the
corresponding separations reach the critical values, that is,

𝑑𝐹 [4]
=0
𝑑𝛿

𝑑𝐹 [5]
=0
𝑑𝛿

where the subscripts 𝑛, 𝑡 denote the normal and tangential components, respectively. Second, the
maximum tractions (𝑇 , 𝑇 ) correspond to the respective cohesive strengths (𝜎 ,𝜏 ),

𝐹 [6]
𝑇| = lim =𝜎
∆ → ∆𝑎

𝐹 [7]
𝑇| = lim =𝜏
∆ → ∆𝑎

Owing to the finite element discretization, the force in each spring acts on an area of ∆𝑎, which
depends on the characteristic size of the mesh. Third, the areas under the normal and tangential
force–separation curves should correspond to the mode-I and mode-II critical fracture energy
release rates, that is,

𝐹 (𝛿 ) [8]
𝑇 (𝛿 , 0)𝑑𝛿 = 𝑑𝛿 = 𝐺
∆𝑎

𝐹 (𝛿 ) [9]
𝑇 (𝛿 , 0)𝑑𝛿 = 𝑑𝛿 = 𝐺
∆𝑎
Using the above six equations 4–9, we can calibrate the six parameters of the exponential
softening law, 𝐵 , 𝛿 , 𝐾 , 𝐵 , 𝛿 and 𝐾 in 2D. The final expressions for the parameters after
some mathematical simplification are,

𝐺 𝐺 [10]
𝛿 = ; 𝛿 =
𝜎 exp (1) 𝜏 exp (1)

1 1 [11]
𝐵 = ;𝐵 =
𝛿 𝛿

𝐺 𝐺 [12]
𝐾 = 𝐵 ∆𝑎 ; 𝐾 = 𝐵 ∆𝑎
exp (1) exp (1)

2.3 Numerical Implementation

In this section, we shall describe briefly the numerical implementation of the discrete damage
zone model within the framework of the finite element method. Let us consider the behavior
around the crack tip as shown in Figure 3. Under an applied load, as the extensions of interface
springs increase the interface separates and the crack opens. Eventually, under a sustained load
this will lead to the complete failure of the springs. In Figure 3, spring 6 has completely failed
whereas springs 1, 2 have zero extensions and so all these springs have zero force. The springs 3,
4, 5 have non-zero extensions and non-zero forces.

Figure 3. Sketch of the discrete damage zone model (DDZM) in a domain with an initial crack

In a general case scenario, when the point of fracture initiation in a domain is not known, every
node of the finite element mesh is defined as double nodes and discrete spring elements are
placed between these double nodes. For benchmark problems, the potential crack path is already
known and hence there is no need to put springs in the whole domain. The proposed DDZM is
implemented in the commercial finite element analysis software ABAQUS. The discrete spring
elements are defined via the user defined element subroutine UEL. Figure 4 shows the element
definition and node numbering convention. Even for the case of 3D delamination analysis, the
interface forces can be resolved into normal and tangential components and the following
scheme can be implemented using the principles of vector and tensor calculus. For brevity and
simplicity, we describe the implementation for a general 2D spring element that is not aligned
with the global coordinate axes. Note that the initial length of these spring elements is zero since
the springs are attached between double nodes. After deformation, the displacement vector in
global Cartesian coordinates (𝑥 , 𝑥 ) is given by,

𝒖 = (𝑢 , 𝑢 , 𝑢 , 𝑢 )T [13]

In local coordinates (𝑥 , 𝑥 ), the vector is given as,

𝒖 = 𝑹𝑢 = (𝑢 , 𝑢 , 𝑢 , 𝑢 )T [14]

where a “ T ” in the superscript denotes transpose and 𝑹 is the anticlockwise rotation matrix is
given by,

𝑐𝑜𝑠𝜃 𝑠𝑖𝑛𝜃 0 0 [15]

𝑹 = −𝑠𝑖𝑛𝜃 𝑐𝑜𝑠𝜃 0 0
0 0 𝑐𝑜𝑠𝜃 𝑠𝑖𝑛𝜃
0 0 −𝑠𝑖𝑛𝜃 𝑐𝑜𝑠𝜃

𝑢
𝑢′2
𝑢′1 +
𝑢
+ +
+
+ Continuum
𝑢 +
+
𝑢′4 +
𝑢′3 +
𝑢 + +

Global Local +
+
Interface
𝑥 + +
𝑥 + +
𝑥 Continuum
+ +
𝜃 + +
𝑥
+ +
+

Figure 4. Sketch of the discrete spring element placed at the nodes of the continuum elements at
the interface and the definition of degrees of freedom in local and global coordinates. The
inclination of the spring 𝜃 defines the angle of coordinate rotation.

The stiffness of the spring is a function of damage variable 𝐷 in both directions, along (𝑛) and
perpendicular (𝑡) to its axis,
 𝐾 = (1 − 𝐷 )𝐾 [16]

𝐾 = (1 − 𝐷 )𝐾 [17]

Note that the above two equations give the expressions for the non-negative secant moduli,
which avoid the convergence issues associated with the possibly negative tangent moduli. In
order to implement this interface element via ABAQUS user element subroutine UEL, the
element stiffness matrix, 𝑲 (AMATRX), and the residual force vector, 𝒓 (RHS), in global
coordinates need to be defined. The corresponding expressions are given by,

−𝐾 (𝑢 − 𝑢 ) [18]
⎧ ⎫
−𝐾 (𝑢 − 𝑢 )
𝒓′ = 𝑹T
⎨ 𝐾 (𝑢 − 𝑢 ) ⎬
⎩ 𝐾 (𝑢 − 𝑢 ) ⎭

𝐾 0 −𝐾 0 [19]
0 𝐾 0 −𝐾
𝑲 = 𝑹T 𝑹
−𝐾 0 𝐾 0
0 −𝐾 0 𝐾

Note that, in addition to UEL subroutine for the discrete element, we have implemented the user
material subroutine UMAT for incorporating damage in the continuum elements. A scalar
damage variable is used to implement Mazars damage in the continuum. Again, for brevity, the
numerical implementation of the UMAT is omitted in this section since it is quite
straightforward.

2.4 Mixed-Mode Fracture Criterion

In most of the engineering applications, delamination involves both mode-I and mode-II
opening. Therefore, we need to specify the failure based on mixed-mode fracture energy
criterion. Herein, we use a linear criterion for mixed-mode interaction, previously suggested by
Wu and Reuters [19], is given by,

𝐺 𝐺 [20]
+ = 1.0
𝐺 𝐺

where 𝐺 and 𝐺 are evaluated as,

𝐹 (𝛿 ) 𝐹 (𝛿 ) [21]
𝐺 = 𝑑𝛿 ; 𝐺 = 𝑑𝛿
∆𝑎 ∆𝑎

If the criterion in equation 20 is met for an interface element, then we consider that it has failed
and the damage in the spring is set as 𝐷 = 𝐷 ≈ 1.0.
 3. NUMERICAL EXAMPLES AND RESULTS
In this section, we present three examples including one benchmark examples of 2D mixed-mode
fracture known as the single leg bending (SLB) problem [18]. A convergence study is performed
for the SLB problem to show the influence of cohesive strength in comparison to the results
obtained from LEFM. The other two examples include 3D mode-I delamination in a double
cantilever beam (DCB) problem and fracture at adhesive joints in a simply supported beam
(SSB) problem. Table 1 summarizes the material properties used in each example. The SLB
problem specimen consists multiple layers made of two different composite materials (L1 and L2)
and the SSB problem compares a weak and strong interface, as given in Table 1.

Table 1. Summary of material properties used in each example.

Material 𝑬𝟏𝟏 𝑬𝟐𝟐 𝝊𝟏𝟐 𝑮𝑰𝑪 𝝈𝒎𝒂𝒙 𝑮𝑰𝑰𝑪 𝝉𝒎𝒂𝒙

(𝐍/𝐦𝐦𝟐 ) (𝐍/𝐦𝐦𝟐 ) (𝐍/𝐦𝐦𝟏 ) (𝐍/𝐦𝐦𝟐 ) (𝐍/𝐦𝐦𝟏 ) (𝐍/𝐦𝐦𝟐 )

SLB (L1) 9.1e6 1.5e6 0.51 5.6 2.0e3 9.3 5.0e3

SLB (L2) 2.5e6 1.5e6 0.14 5.6 2.0e3 9.3 5.0e3

DCB (3D) 8.00e6 8.00e6 0.00 1.2 4.0 - -

SSB (weak) 2.4e3 2.4e3 0.25 3.0 3.3e1 3.0 3.3e1

SSB (strong) 2.4e3 2.4e3 0.25 3.0e2 3.3e2 3.0e2 3.3e2

3.1 2D Mixed-Mode Delamination: Single Leg Bending

Herein, we validate and illustrate the convergence of the DDZM for mixed-mode delamination.
The set up of the single leg bending (SLB) benchmark problem is shown in Figure 6. The beam
is simply supported at the left and right ends. A crack of length 𝑎 is prescribed as shown so as
to initiate delamination. The dimensions of the beam are as follows: 𝑎 = 0.27 mm, 𝐻 = 0.0884
mm, 𝐻 = 0.2876 mm, 𝐿 = 3.0 mm, 𝐿 = 4.8 mm, 𝐿 = 6.0 mm.

Figure 5. Initial geometry and boundary conditions for single leg bending (SLB) model
To initiate mixed mode delamination, displacement increments are applied at mid-span as shown
in the Figure 5. Note that displacement control is applied in order to capture the softening
behavior of the load–displacement (P–Δ) curve. The discrete spring elements are placed only
along the dashed line, which describes the expected delamination path. Figures 6(a) and 6(b)
show the Mises contour plots on deformed shape at time step “A” when the crack propagates to
the point of load application and at the final time step “B” f the simulation. It can be seen from
Figure 6 that once the lower layer separates it is under zero.

(a) Time step A when the crack propagates to the point of load application

(b) The final time step B of the simulation

Figure 6. Mises contour plot on deformed shape at the application point and the end of the
simulation. The area circled by red dashed line is zoomed in to show the finite element mesh.

Figure 7 shows the P–Δ curves and the behavior at time steps A and B are marked. As the beam
deforms under the applied loading, the extensions (δ) in the interface elements increase and
eventually when a critical displacement is reached, the crack propagates very quickly so as to
cause an abrupt drop in the load bearing capacity of the beam. This sudden drop in load bearing
capacity happens till the crack reaches the application point of load at time step A. From this
point, the crack propagates much slower and the behavior of the beam is close to linear elastic
with a different slope than at the beginning. This behavior can be clearly observed in Figure 7.
Results obtained from the virtual crack closure technique (VCCT) are used as reference LEFM
solutions. The results from the DDZM are in good agreement with those of VCCT, which
validates our implementation; however, the DDZM results depict a slightly less abrupt failure
than the VCCT results owing to bounded stress in the cohesive zone.

Next, in order to demonstrate the convergence of the scheme, we vary the material cohesive
strengths one component at a time. In the results of Figure 7(a) the normal cohesive strength
𝜎 is a constant while in those of Figure 7(b) the tangential cohesive strength 𝜏 is a
constant. The numerical results obtained from higher cohesive strengths converge to the VCCT
results demonstrating the convergence of the DDZM to the LEFM solution.

400 400

300 B 300 B
Load P (N)

Load P (N)
A A
200 200

VCCT solution VCCT solution

100 2 2 100 2
σ = 0.5 KN/mm , τ = 5.0 KN/mm
2
σ = 2.0 KN/mm , τ = 3.0 KN/mm
2 2
2
σ = 2.0 KN/mm , τ = 4.0 KN/mm
2 σ = 1.0 KN/mm , τ = 5.0 KN/mm
2 2
2
σ = 2.0 KN/mm , τ = 5.0 KN/mm
2 σ = 2.0 KN/mm , τ = 5.0 KN/mm
0 0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.00 0.05 0.10 0.15 0.20 0.25 0.30
Displacement Δ (mm) Displacement Δ (mm)

(a) Constant 𝜎 (b) Constant 𝜏

Figure 7. The load vs. displacement curves for the single leg bending problem. The results for higher
cohesive strengths demonstrate convergence to the LEFM solution.

3.2 3D Mode-I Delamination: Double Cantilever Beam

Herein, we demonstrate the applicability of the DDZM for mode-I delamination analysis in 3D.
The initial geometry and boundary conditions are shown in Figure 8. The right end of the beam
is fixed and displacement boundary conditions are applied on both the upper and lower layers at
mid-point on the left end. Each layer is a 2D plate with zero thickness. The dimensions are as
follows: 𝑊 = 1 mm, 𝐿 = 9 mm.

Figure 8. Initial geometry and boundary conditions for double cantilever bending (DCB) in 3D
The delamination occurs due to the debonding of the upper and lower layers. The discrete spring
elements are placed between these two layers by connecting double nodes belonging to each
layer. Again, displacement control is used so as to capture the global softening behavior. Figure
9(a) and 9(b) show the Mises contour plots on deformed shape at the end of the first and last time
steps of the simulation, respectively. Upon close observation, one can see the variation of the
stress in the y-direction close to the left edge. This is the effect of a local stress concentration as
the displacement increments are applied at the mid-point of the edge. Far away from the point of
application, the stress is uniform across the width of the specimen.

(a) Delamination at the end of the first time step

(b) Delamination at the end of the final time step

Figure 9. Mises contour plot on deformed shape of double cantilever bending in 3D

 20

16

Load P (N)
12

4
VCCT solution
DDZM solution
0
0.00 0.04 0.08 0.12 0.16
Displacement Δ (mm)

Figure 10. Load vs. displacement curves for double cantilever beam (DCB) debonding in 3D

The load-displacement (P– Δ) curve obtained from the DDZM is plotted against the result from
VCCT, shown in Figure 10. The VCCT result consists of a linear regime followed by a failure
regime exhibiting small fluctuations in load bearing capacity, whereas, the DDZM result depicts
a smoother global failure curve. This is expected since the DDZM assumes a cohesive zone with
damageable interface elements wherein the individual elements gradually failure while
simultaneously redistributing the forces to the surrounding active interface elements.

3.3 Mode-I fracture: Simply Supported Beam Bending

Herein, we consider mode-I fracture of an adhesive joint in a beam element using the DDZM.
The initial geometry and boundary conditions for this simply supported beam are shown in
Figure 11. Both the initial crack and the load are prescribed at mid-span. Displacement boundary
condition is applied in order to deform the beam. The dimensions are as follows: 𝑎 = 10 mm, 𝐻
= 20 mm, 𝐿 = 45 mm, 𝐿 = 22.5 mm, 𝐿 = 90 mm.

adhesive joint interface

Figure 11. Initial geometry and boundary conditions for simply supported beam (SSB). The
adhesive joint interface is indicated by the dashed line.
The discrete spring elements are placed along the adhesive joint interface described by the
dashed line in Figure 11. Mazars damage law [17] governs the behavior of both the interface and
continuum elements. We consider two hypothetical cases, one with weak interface and the other
with strong interface. The materials properties assumed are given in Table 1. Damage contour
plots for these two cases are shown in Figure 12. For a weak interface the failure occurs
primarily in the interface elements, whereas, for a strong interface the damage occurs in the
continuum elements surrounding the crack tip.

Case I: weak interface (hypothetical)

Case II: strong interface (hypothetical)

Figure 12. Damage contour plots for both weak interface and strong interface. The red areas
indicate completely damaged elements.

The DDZM is able to capture both interface failure in the adhesive joint as well as the failure of
surrounding matrix due to microcracking. Thus, this example illustrates the generality of the
current approach.

4. CONCLUSIONS
4.1 Summary

The discrete damage zone model (DDZM) [20] has been implemented to study delamination in
laminated composites. The model has been developed from the perspective of damage mechanics
and so it naturally accounts for the irreversibility of stiffness degradation once the material is
loaded beyond the elastic limit. The model considers both damage of the continuum and the
interface and so it is able to capture multiple failure modes such as debonding and matrix
cracking. For a homogeneous isotropic material the same damage law governs both the interface
element and continuum element stiffness degradation, thus, leads to a unified formulation.
Further, since damage evolution is allowed in both interfacial and bulk elements, the model is
applicable to problems where the point of fracture initiation and the path of fracture propagation
are not know a priori.

This generalized formulation is simple and can be readily incorporated into existing finite
element software such as ABAQUS. The numerical results, of 2D mixed-mode single leg beam
bending and 3D double cantilever beam mode-I delamination, from the DDZM show good
agreement with the reference VCCT results. Moreover, the numerical investigation of the mixed
mode single leg bending test with various cohesive strengths demonstrates the convergence to
linear elastic fracture mechanics. The results of mode-I fracture of adhesive joint interface in
beam elements demonstrate the versatility of the method for problems involving multiple modes
of failure. Thus, the study conducted in the article illustrates the viability of the method for study
failure in laminated composites which is important to predict their durability and performance.

4.2 Afterwords

The irreversibility of damage makes this numerical scheme particularly suitable for studying
fatigue behavior of composites. By decomposing the total damage into static and cyclic
components and by defining the corresponding damage laws we can investigate failure under
fatigue loading. Currently, we are working on extending the discrete damage zone model.

5. REFERENCES
1. Robinson, P., Besant, T. & Hitchings, D., “Delamination growth prediction using a finite
element approach.” 2nd ESIS TC4 Conference on Polymers and Composites, Les Diablerets,
Switzerland, 1999
2. Griffith, A.A., “The phenomena of rupture and flow in solids.” Philosophical Transactions of
the Royal Society, London, Series A, 221(1921): 163-198
3. Irwin, G.R., “Analysis of stresses and strains near the end of a crack traversing a plate.”
Journal of Applied Mechanics, 24(1957): 361-364
4. Blackman, B., Dear, J.P., Kinloch, A.J. & Osiyemi, S., “The calculation of adhesive fracture
energies from double-cantilever beam test specimens.” Journal of Materials Science Letters,
10(1991): 253-256
5. Blackman, B.R.K., Kinloch, A.J., Paraschi, M. & Teo, W.S., “Measuring the mode I
adhesive fracture energy, 𝐺 , of structural adhesive joints: the results of an international
round-robin.” International Journal of Adhesion & Adhesives, 23(2003): 293-305
6. Blackman, B.R.K., Hadavinia, H., Kinloch, A.J., Paraschi, M. & Williams, J.G., “The
calculation of adhesive fracture energies in mode I: Revisiting the tapered double cantilever
beam (TDCB) test.” Engineering Fracture Mechanics, 70(2003): 233-248
7. Bazant, Z.P., “Size effect on structural strength: A review.” Archives of Applied Mechanics,
69(9-10)(1999): 703-725
8. Dugdale, D.S., “Yielding of steel sheets containing slits.” Journal of the Mechanics and
Physics of Solids, 8(2)(1960): 100-104
9. Barenblatt, G.I., “The mathematical theory of equilibrium cracks in brittle fracture.”
Advances in Applied Mechanics, 7(1962): 55-129
10. Fish, J., Yu, Q. & Shek, K.L., “Computational damage mechanics for composite materials
based on mathematical homogenisation.” International Journal for Numerical Methods in
Engineering, 45(1999): 1657-1679
11. de Borst, R., “Some recent issues in computational failure mechanics.” International Journal
for Numerical Methods in Engineering, 52(2001): 63-95
12. de Borst., R., “Numerical aspects of cohesive-zone models.” Engineering Fracture
Mechanics, 70(2003): 1743-1757
13. Mi, Y., Cisfield, M.A., Davies, G.A.O. & Hellweg, H.B., “Progressive delamination using
interface elements.” Journal of Composite Materials, 32(1998): 1246-1272
14. Cui, W.C. & Wisnom, M.R., “A combined stress-based and fracture mechanics-based model
for predicting delamination in composites.” Composites, 24(6)(1993): 467-474
15. Xie, D., Salvi, A.G., Sun, C. & Waas, A.M., “Discrete cohesive zone model to simulate static
fracture in 2D triaxially braided carbon fiber composites.” Journal of Composite Materials,
40(22)(2006): 2025-2046
16. Xie, D. & Waas, A.M., “Discrete cohesive zone model for mixed-mode fracture using finite
element analysis.” Engineering Fracture Mechanics, 73(2006): 1783-1796
17. Mazars, J., “A description of micro and macro scale damage of concrete structures.”
Engineering Fracture Mechanics, 25(1986):729-737
18. Davidson, B.D., Kruger, R. & Konig, M., “Three-dimensional analysis of center-delaminated
unidirectional and multidirectional single-leg bending specimens.” Composites Science and
Technology, 54(1995): 385-394
19. Wu, E. M. and Reuter, R. C., “Crack extension in fiberglass reinforced plastics.” University
of Illinois TAM Report No. 275, 1965
20. Xia, L., Duddu, R. and Waisman, H., “A discrete damage zone model for mixed-mode
fracture initiation and propagation.” Engineering Fracture Mechanics, in review