Excerpt from the Proceedings of the COMSOL Conference 2010 Boston

Failure Stress Analysis of Fiber Reinforced of Composite Laminates

under Uniaxial/Biaxial Loading
Zeaid Hasan*1, Feras Darwish2 and Suhayb Al-Absi3
1
Texas A&M University, College Station, Texas, 2, 3Jordan University of Science and Technology, Irbid,
Jordan
*Corresponding author: zeadnws@hotmail.com

Abstract: The main objective of the present composite material systems are usually subjected
work is to perform stress analysis on composite to complex loading that leads to multi-axialstress
laminates under uniaxial/biaxial loading to serve and strain fields at critical surface locations
as a preliminary data for test verification. A which leads for the need of a reliable design
detailed calculation based on the Classical procedure validated rigorously by multi-axial
Lamination Theory was performed for a laminate experimental data in order to ensure satisfactory
with stacking sequence [90/45/-45/0]s. The performance over the predefined service period
material used in the analysis was carbon epoxy and to avoid the use of high safety factors to
(AS4/3501-6). We started by applying a pure cover the high level of uncertainty. The
uniaxial followed by a biaxial load. The finite development of the field of composite materials
element software COMSOL was used to model in structural applications relies on the ability to
the geometry, material, loading conditions, and model and simulate the behavior of these
to perform the stress analysis to validate the materials successfully. Numerous theories have
output results of the analysis. Experimental been proposed to predict the response of
results available in the literature were also used composite materials, but in order to apply these
to validate the results of the analysis. Computer analysis tools for structural design, experimental
software was also developed that enables the validation under a variety of complex loading
user to study general composite layup including conditions is mandatory. As composite materials
hybrid composites under uniaxial/biaxial loading generate complex biaxial and multi-axial stress
conditions. Moreover, several carpet plots were states [1], even for simple uniaxial loading, the
also conducted to show the interacting behavior current practice of using solely uniaxial analysis
of two independent variables such as Youngs is not suitable and, consequently, analysis closer
modulus, Poissons ratio and Shear modulus. to reality is of paramount importance. The
composite testing community has successfully
Keywords: Composite Materials, Biaxial load. developed the ability to characterize the uniaxial
response of fibre-reinforced composite materials.
1. Introduction Several standardized test methods have been
developed to evaluate the in-plane shear, the
Composite material systems are increasingly axial and transverse tensile, and the axial and
used in almost every industrial branch. Several transverse compressive response [2].
applications could be found in the automotive,
aeronautical, and sports sectors, additionally, 2. Macromechanical Analysis of Lamina
there are numerous and ever increasing industrial
uses, including wind turbines, storage tanks, The main difference between a lamina and a
high-speed and precision machinery, gas turbine laminate, where a lamina is a thin layer of a
engines, and medical diagnostic equipment. composite material that is generally of a
Their outstanding mechanical performance thickness on the order of 0.005 in. (0.125 mm),
added to their lowweight and other unique and and a laminate is constructed by stacking a
tailorable physical properties make composites number of such lamina in the direction of the
the material system of first choice for many lamina thickness. Mechanical structures made of
applications. The properties of a composite these laminates are subjected to various loads,
material depend on the properties of the such as bending and twisting. The design and
constituents, their geometry, and the distribution analysis of such laminated structures demands
of the phase. The constituent parts of the knowledge of the stresses and strains in the
structural components manufactured from these laminate. Also, design tools, such as failure
theories, and stiffness models need the values of five. In most structural applications, composites
these laminate stresses and strains. materials are used in the form of thin laminates
Understanding the mechanical analysis of a loaded in the plane of the laminate. Thus,
lamina precedes understanding that of a composite laminae and laminates can be
laminate. A lamina is unlike an isotropic considered to be under a condition of plane
homogeneous material. For example, if the stress, with all stress components in the outof-
lamina is made of isotropic homogeneous fibers plane direction being zero. An isotropic material
and an isotropic homogeneous matrix, the is characterized by an infinite number of planes
stiffness of the lamina varies from point to point of material symmetry through a point. We
depending on whether the point is in the fiber, summarize the number of independent elastic
the matrix, or the fibermatrix interface. constants for various types of materials as
Accounting for these variations will make any follows
kind of mechanical modeling of the lamina very Anisotropic: 36
complicated. For this reason, the Orthotropic: 9
macromechanical analysis of a lamina is based Transversely isotropic: 5
on average properties and considering the lamina Isotropic: 2
to be homogeneous.
2.2 Stress-Strain Relations for a Thin Lamina
2.1 Stress-Strain Relations (Two-Dimensional)

The state of stress for general anisotropic A thin, unidirectional lamina is assumed to
material can be represented by nine stress be under a state of plane stress and the equations
components, ij (where i, j=1, 2, 3) acting on the that govern the stress and strain are shown in Eq.
sides of an element cube. Similarly, the state of (4) where Qij is the reduced stiffness coefficients.
deformation is represented by nine strain For a unidirectional lamina, these engineering
components, ij. Applying symmetry of the stress elastic constants are E1, E2, 12 and G12.
and strain tensors
ij = ji (1) 1 Q11 Q12 0 1
ij = ji (2)
= Q Q22 0 2 (4)
2 12
Thus the stress-strain relation for anisotropic 12 0 0 Q66 12
body is given as follows for a three-dimensional
body in a 123 orthogonal Cartesian coordinate
3. Macromechanical Analysis of
system
Laminates
= ij Cijkl kl ( where=
i, j 1, 2, 3, , 6 ) (3)
Where the 6 6 [C] matrix is called the stiffness It is apparent that the overall behavior of a
matrix which contains 36 constants. For multidirectional laminate is a function of the
Orthotropic material which have three mutually properties and stacking sequence of the
perpendicular planes of material symmetry. The individual layers. The so-called classical
stress-strain relation have the same form as lamination theory predicts the behavior of the
anisotropic material, however, the number of laminate within the framework of several
independent elastic constants are reduced to nine, assumptions and restrictions. Figure 1 shows two
because the stiffness and the compliance terms cross sections before and after loading, we can
are interrelated. An orthotropic material is called observe the deformation that has occurred after
a transversely isotropic when one of its principal loading. Assume u0, v0, and w0 to be the initial
planes is a plane of isotropy, that is at every displacements in the x, y, and z directions,
point there is a plane on which the mechanical respectively, at the midplane and u, v, and w are
properties are the same in all directions. The the displacements at any point in the x, y, and z
stress-strain relations for transversely isotropic directions, respectively. At any point other than
materials are simplified for a two-three planes of the midplane, the two displacements in the xy
isotropy; thus, the stress-strain relations for a plane will depend on the axial location of the
transversely isotropic material are reduced to
 point and the slope of the laminate midplane stiffness matrices, respectively. The extensional
with the x and y directions. stiffness matrix [A] relates the resultant in-plane
forces to the in-plane strains, and the bending
stiffness matrix [D] relates the resultant bending
moments to the plate curvatures. The coupling
stiffness matrix [B] couples the force and
moment terms to the midplane strains and
midplane curvatures.

4. Failure Theories

When dealing with composite materials,

micromechanical failure theories have been
Figure 1. Relationship between displacements through
proposed by extending and adapting isotropic
the thickness of a plate to midplane displacements and
curvatures failure theories to account for the anisotropy in
stiffness and strength of the composite. Lamina
If the strains are known at any point along the failure theories can be classified in the following
thickness of the laminate, the stressstrain in three groups: Limit or noninteractive theories, in
each lamina can be defined as which specific failure modes are predicted by
comparing individual lamina stress or strains
with corresponding strengths or ultimate strains,
Q11 Q12 Q16
x x
for example Maximum Stress and Maximum
= Q Q 22

Q 26 y (5)
Strain theories where no interaction among
y 12
different stress components on failure is
xy xy considered. Interactive theories (the Tsai-Hill
Q Q 26 Q 66 and the Tsai-Wu theories) in which all stress
16 components are included in one expression
The reduced transformed stiffness matrix, [Q], (failure criterion). Overall failure is predicted
corresponds to that of the ply located at the point without reference to particular failure modes.
along the thickness of the laminate. For a certain Partially interactive or failure mode based theory
linear strain variation through the thickness, (the Hashin-Rotem) where separate criteria are
which can result from axial and flexural loading, given for fiber and interfiber failures, more
the variation of the modulus Ex from layer to detailed description of these theories can be
layer causes discontinues stress variation. found in [4].
Because of the discontinuous variation of
stresses from layer to layer, it is more convenient 5. Analysis
to deal with the integrated effect of these stresses
on the laminate. Thus, we seek expressions A detailed calculation for a specific case of a
relating forces and moments to laminate composite layup is performed in order to get the
deformation. More detailed discussion could be failure stress by considering two types of failure
found in [3]. criterias which are the First Ply Failure (FPF)
n
criteria in which the laminate is considered failed
=Aij Q ij ( hk hk 1 ) when the first layer (or group of layers) fails, and
k =1 k the iterative Ultimate Laminate Failure (ULF)
criteria where there is no generally accepted
1 n
2 2
=Bij Q ij ( hk hk 1 ) (6) definition of what constitutes such failure but It
2 k =1 k is generally accepted that a laminate is
considered failed when the maximum load level
3 3
( hk hk 1 )
n
1
Q
is reached. The flow diagrams for each failure
=Dij ij
3 k =1 k criteria are shown in Figure 2 and Figure 3
respectively. The calculation was performed
The [A], [B], and [D] matrices defined in Eq. (6) using four different theories that were previously
are called the extensional, coupling, and bending mentioned and the results obtained are compared
with each other. The geometry and loading
condition for the biaxial loading is shown in
Figure 4, the factor n is the ratio of the stress
applied in the transverse direction to that in the
axial direction. A summary of each case is
shown in Eq. (7), as can be observed, when the
value of n = 0, a pure axial load (uniaxial) is
applied which is a special case of the biaxial
loading.
Figure 4. Loading possibilities applied to the
composite laminate

0 Uniaxial _ Load
1 Biaxial _ Load

n= (7)
0 n 1 Biaxial _ Load
n 1 Biaxial _ Load

Table 1: Failure load based on the FPF & ULF

Figure 2. Flow chart for first ply failure criteria criterias for several different failure theories; n = 0,
Tension

Max Max Tsai-Hill Tsai-Wu

Stress Strain
FPF (ksi) 50.5/90 50.5/90 49.9/90 46.6/90
Ply
ULF 115/0 114/0 112/0 108/0
(ksi)/ Ply

Table 2: Failure load based on the FPF & ULF

criterias for several different failure theories; n = 0,
Compression

Max Max Tsai-Hill Tsai-Wu

Stress Strain
Figure 3. Flow chart for ultimate laminate failure FPF (ksi) 71.7/45 71.7/45 68/45 85/45
criteria Ply
ULF 180/90 180/90 167/90 153/90
(ksi)/Ply
6. Results
Table 3: Failure load based on the FPF & ULF
Based on the Classical lamination plate theory,
criterias for several different failure theories; n = 1,
several loading cases were considered with Tension/ Tension
different values of n applied to a symmetric
laminate having the following stacking sequence Max Max Tsai-Hill Tsai-Wu
[90/45/-45/0]s; the ply thickness is 0.005 inch. Stress Strain
The material used is carbon/epoxy (AS4/3501- FPF (ksi) 51/90 51/90 51/90 51/90
6). The minimum load based on the FPF and Ply
ULF criterias are summarized in Tables. (1-6). ULF 51/0 51/0 51/0 51/0
(ksi)/Ply
Table 4: Failure load based on the FPF & ULF the minimum stress occurs in the 45/-45 ply for
criterias for several different failure theories; n = 1, the tension/completion case since these are the
Compression/ Compression least resistive plies regarding the given loading
condition while the compression/tension case
Max Max Tsai-Hill Tsai-Wu gives a reverse behavior keeping the 45/-45
Stress Strain
FPF (ksi) 136/0 136/0 136/0 136/0
plies the least resistive hence the first to fail.
Ply
ULF 136/90 136/90 136/90 136/90 7. Finite Element Verification
(ksi)/Ply
In order to verify the results obtained from the
Table 5: Failure load based on the FPF & ULF previous analysis the is used to implement a
criterias for several different failure theories; n = -1, certain case study. A composite laminate made
Tension/Compression
of Carbon/Epoxy (AS4/3501-6); the model
consists of 8 plies with fiber orientation [90/45/-
Max Max Tsai-Hill Tsai-Wu
Stress Strain 45/0]s. The model implemented was used to
FPF (ksi) 35.8/45 35.8/45 35.8/45 35.9/45 examine the stress of each ply resulting from a
Ply tension stress followed by a compression stress
ULF 85/0 85/0 74/0 67/0 in order to verify the results that were obtained
(ksi)/Ply from the previous analysis. Eight rectangular
plies were used in the analysis. The element used
Table 6: Failure load based on the FPF & ULF failure are free tetrahedral with a total of 12000
criterias for several different failure theories; n= -1, elements in the model. We will apply the failure
Compression/Tension
load that we have obtained from our previous
analysis to one end of the laminate, while the
Max Max Tsai-Hill Tsai-Wu
Stress Strain other end constrained. It was observed that the
FPF (ksi) 35.8/45 35.8/45 35.8/45 35.9/45 principle stress for the uniaxial case 2 = 30 ksi
Ply in the 90 ply exceeded the transverse tensile
ULF 85/90 85/90 74/90 67/90 strength of the material which means that it
(ksi)/Ply failed as shown in Table 7 while all the principle
stresses in the rest of the plies have not exceed
Table 1 and Table 2 show the results of applying the corresponding strength. Table 8 shows the
a pure uniaxial load in tension and compression result of applying a compression stress; it can be
on the proposed composite laminate respectively. observed that the shear stress in the +-45 plies
As shown in the tables, the minimum stress exceeded the corresponding shear strength
occurred in the 90 ply for the tension case as leading them to fail under the given load while
expected since it has the least resistance against the other plies not; these results aggre with the
the load unlike the compression case where the results previously obtained, hence, according to
minimum load occurred in the 45/-45 plies. the result of the FE model, we found that the
Table 3 and Table 4 show the results of applying results that we obtained from the analysis are
a biaxial load in tension/tension case and reliable and correct. In addition, Table 9 shows a
compression /compression on the proposed comparison of the ULF values obtained from
composite laminate respectively. As shown in experimental results found in [5] with those
the tables, the stress is almost equally distributed obtained from our previous analysis and both
throughout all the plies of the laminate since the show good agreement.
laminate is symmetric and balanced while the
composite laminate will fail in the 90/0 plies at
the same time due to the effect of the biaxial
load. Table 5 and Table 6 show the results of
applying a pure biaxial load but this time the
load fraction n has a negative value giving the
possibility of applying tension/compression or
compression /tension to the proposed composite
laminate respectively. As shown in the tables,
 engineering constants of a symmetric laminate
depend only on the proportion of the various
plies in the entire laminate and not on the exact
stacking sequence. Thus, in-plane engineering
constants are a function of the fractional values
, , , where
=2m/N =2n/N =4p/N
N: total number of plies
As we mentioned a carpet plot is a parametric
family of curves with one of the fractions , ,
as a variable and the other two as parameters,
keeping in mind that + + = 1. Such plots for
Youngs modulus, shear modulus, and Poissons
Figure 5: Finite element mesh ration are shown in Fig. 8, 9, 10 respectively for
Carbon/Epoxy material (AS4/3501-6).
Table 7: Failure stress summery of the FEA model for
n=0, tension

Ply 1 2 6 F1t F2t F6

(ksi) (ksi) (ksi) (ksi) (ksi) (ksi)
90 8.6 30.17 0.3 330 8.3 11
45 47.8 4.5 -7.9 330 8.3 11
-45 47.5 4.4 7.9 330 8.3 11
0 120 3.5 0.27 330 8.3 11

Table 8: Failure stress summery of the FEA model for

n=0, compression

1 2 6 F1c F2c F6
Ply
(ksi) (ksi) (ksi) (ksi) (ksi) (ksi) Figure 6. Carpet Plot for Youngs modulus of
90 -12.4 28.4 -0.3 250 33 11
[90/45/-45/0]s carbon epoxy laminates (AS4/3501-6)
45 -67 -6.8 12.4 250 33 11
-45 -67 -6.8 -12.4 250 33 11
0 -189 -0.04 -0.3 250 33 11

Table 9: Comparison between experimental &

analytical results

Laminate ULF - ULF -

Experimental Previous
(ksi) Analysis (ksi)
[60/-60/0]s 106 115
n= 0

8. Carpet Plots
Figure 7. Carpet Plot for Poissons ratio of
[90/45/-45/0]s carbon epoxy laminates (AS4/3501-6)
Carpet plot is one that illustrates the interacting
behavior of two independent variables, which
among other things facilitates interpolation in
both variables at once, mainly used as a design
tool. A designation of a certain layup is
[0m/90n/+-45p]s, where m, n, p denote the number
of 0,90,+-45 plies, respectively. The in-plane
 Figure 9. Carpet Plot for Shear modulus of
[90/45/-45/0]s carbon epoxy laminates (AS4/3501-6)
Figure 8. Main Interface of the program
9. Computer Software
addition to experimental verification. Computer
The Composite Analyzer is an engineering software was also conducted that enables the
program that analyzes laminated composite user to analyze a broad variety of composite
plates under uniaxial and biaxial loading layups. Also, carpet plots for youngs modulus,
according to classical laminated plate theory. poissons ratio and shear modulus were
Familiarity with such analysis is assumed. Input introduced.
consists of ply material properties, material
strengths, ply fiber orientation and stacking 11. References
sequence, ply thickness, type of composite
(Hybrid: Composite Laminates containing plies [1] Xu, J., Askari, A., Weckner, O., Razi, H.,
of two or more different types of material) and Silling, S., Damage and Failure Analysis of
unit system(English/SI). Output consists of Composite Laminates under Biaxial Loads 48th
apparent laminate material properties, ply AIAA/ASME/ASCE/AHS/ASC Structures,
stiffness and compliance matrices, laminate Structural Dynamics, and Materials Conference
"ABD" matrices, lamina failure load and mode (2007).
based on Maximum Stress, Maximum Strain, [2] Hemelrijck, D., Ramault, A., Lamkanfi, E.,
Tsai-Hill, and Tsai-Wu failure theories, failure Paepegem, W., Lecompte, D., Biaxial testing of
load and mode for the hole laminate based on fiber-reinforced composite laminates IMechE
first ply failure (FPF) also based on ultimate Vol. 222 (2008).
laminate failure (ULF) and Carpet plots for [3] Reddy, J.N., Mechanics of laminated
orthotropic laminates (An orthotropic laminate composite plates and shells: theory and
requires that the number of + angle plies equals analysis CRC Press (2004)
the number of angle plies [ 90 /+-/0]). The [4] Sandhu, R.S., A survey of failure theories of
main interface page of the program is shown in isotropic anisotropic materials Technical
Fig. 11. Report, AFFDL-TR-72-7 I (1972).
[5] Darwish,F., Shivakumar, K.N., Hamoush, S.,
10. Conclusion PERFORMANCE OF PATCH REPAIRED
COMPOSITE PANELS STATIC AND
This paper presented a general approach in FATIGUE AIAA/ASME/ASCE/AHS/ASC
analyzing composite laminates under uniaxial Structures, Structural Dynamics, and Materials
and biaxial loading using two different failure Conference (2006).
criterias. It was observed that the failure [6] Daniel, I.M., Ishai, O., Engineering
occurred in the 90 ply when applying a uniaxial Mechanics of Composite Materials Oxford
load while the failure occurred in the 45 / -45 University Press (2006).
plies in the biaxial case which obtains the least
resistance in that case. Verification using the
FEA software COMSOL was performed in