ws-1949 sr 13 00 + 0 00

Perpamw fournals Ltd.

A FIRST-PLY FAILURE ANALYSIS OF

COMPOSITE LAMINATES

J. N. REDDY and A. K. PANDEY

Department of Engineering Science and Mechanics, Virginia Polytechnic Institute and State University.
Blacksburg, VA 24061, U.S.A.

(Received 18 June 1986)

Abstract-A unite-element computational procedure is developed for the first-ply failure analysis of
laminated composite plates. The procedure is based on the first-order shear deformation theory and a
tensor polynomial failure criterion that contains the maximum stress. maximum strain, the Hill. Tsai-U’u
and Hoffman failure criteria as special cases. By specifying the desired criterion, a first-ply failure analysis
of composite laminates subjected to in-plane and/or bending loads can be achieved. A number of problems
are presented to evaluate these failure criteria when applied to laminates subjected to in-plane and or
bending loads.

NOTATION 1. 1NTRODLKl’IOIY

extensional, flexural-extensional and

The use of composite materials in automobile and
Aexural stiffnesses (ij = I. 2,6) space structures in the last two decades has increased
plate planform dimensions along the .I not only in number but also in the breadth of appli-
and _Ydirections, respectively cations. The applications range from medical pros-
layer elastic moduli in directions along thetic devices, sports equipment, applicances, auto-
fibers and normal to them, respectiveiy
components of the strength tensors in motive parts, to high performance military structures
failure criteria (e.g. ground. underwater and space vehicles). The use
layer in-plane and thickness shear moduli of composite materials in structural applications is
total thickness of the laminate dictated by the outstanding strength, stiffness and,
thickness of ith layer
low specific gravity of fibers, low maintenance costs,
shear correction coefficients associated
with the JZ and .r= planes, respectively and the flexibility in tailoring the stiffness and
(i = 4, 5) strength in the preliminary design stage of complex
stress couples and stress resultants, re- vehicle structures. Use of composite materials in
spectively (i = 1.2,6) advanced underwater and space vehicles can result in
shear stress resultants (i = I, 2)
plane-stress reduced stiffness coefficients significant increase in payload, weight reduction,
(i.i = I, 2.6) range and speed, maneuverability, fuel efficiency and
shear stress strengths of a lamina in yz, .YZ safety. Functional requirements and economic con-
and XJ planes, respectively siderations of design require designers to use reliable
shear strain strength of laminae in the y;,
.YZand X_Yplanes, respectively and accurate but economical methods of determining
components of the compliance matrix stresses and identifying failure mechanisms.
displacement components in the X,Y, z The freedom to tailor the stiffness and strength of
directions, respectively filamentary composites in the preliminary design
nodal values of displacements u, c, w
stage of complex vehicle structures demands further
(i=l,X...,n)
position coordinates in Cartesian system attention to details which either have no counterpart
compressive stress stength of tamina in the in metallic structure design, or are considered of
s. y, z directions, respectively secondary importance. For example, changing lami-
tensile stress strength of lamina in the nate thickness by terminating internal plies. which is
s, .v, I directions, respectively
compressive strain strength of lamina in
also known as a ply drop-off, is a common design
the X. _F.2 directions, respectively detail for achieving efficient stiffness distribution in
tensile strain strength of lamina in the the tapered composite panels (e.g. wing skins). Also,
.r, J. I directions, respectively inserting or terminating internal plies at the con-
column vector of generalized nodal dis-
nections between components is a version of the ply
placements
strain components (i = 1.2.. . ,6) drop-off problem. Ply drop-offs initiate interlaminar
orientation of the nrth layer stresses, and filamentary composites are particularly
(??1= f, 2.. , L) vulnerable to even low interlaminar stresses. Detailed
stress components (i = I. 2,. . . ,6) analyses inco~orating interlaminar stress predictions
unite-element interpolation functions
(i = 1.L.... are required in order to assess if such stresses reduce
fi)
rotations of a transverse normal about the strength below design loads.
v and I coordinates, respectively. Establishing the compression strength of

371
372 J. S. &ODY and A. K. P.ASDEY

filamentary composites is important for structural scopic or macroscopic points of view. From a micro-
integrity of lightweight vehicle structures. Com- scopic point of view, failure of the composite material
pression strength can be more limiting than tensile is attributed to the damage on the molecular level. A
strength in governing the design. The ply drop-off damage as defined in [a] is ‘a collection of perma-
mentioned in the previous paragraph, for example. is nent microsctructural changes brought about in the
weaker in compression than tension. .Mechanisms material by a set of irreversible physical (or chemical)
explaining compression strength are not completely processes resulting from application of thermo-
understood. but fiber microbuckling, fiber kinking mechanical loadings’. Analysis of the composite
and matrix splitting are very plausible explanations. materials at a microscopic level is difficult and in-
These micromechanical details are very difficult to tractable.
incorporate on the structural design and analysis The macroscopic behavior of composite materials
level. In spite of the difficulty, a means of incorpor- can be deduced from the microscopic behavior [j]. In
ating micromechanical strength and failure mech- practice, macroscopic behavior is determined from
anisms in the macroscopic analysis is desirable for the load-displacement data of a given test specimen,
rational design. Mechanistic rather than phenom- without the deep understanding of the activities at
enologically based failure criterion would provide the microscopic level.
means for a more rational strength and failure anal- One important area of concentration, besides the
ysis (of course, phenomenological studies are needed failure load, is the mode of failure. Laminated com-
to formulate mechanistic failure criteria). Mech- posite may fail by fiber yielding, matrix yielding, fiber
anistic failure criteria bridge the gap between micro- breakage, delamination of layer or by fracture. The
and macromechanical analyses, and provide the first three failure modes depend on the constituent’s
structural analysis with ability to isolate specific strength properties, whereas delamination is basically
failure modes. due to improper stacking sequence of laminae. Frac-
The present study deals with the first-ply failure of ture is caused by the pre-existing voids and cracks in
composite laminates. The failure criteria used are the constituent material. Macroscopic failure criteria,
macroscopic in nature. The primary objective of the which are discussed here, are based on the tensile,
present study is to incorporate various commonly compressive and shear strengths of the individual
known macroscopic failure criteria into finite-element lamina.
computational procedure and to evaluate various The most popular failure criteria, as discussed by
criteria when applied to composite laminates in Soni [6], are maximum stress criterion, maximum
stretching and bending. The following literature re- strain criterion and quadratic polynomial criteria
view forms a background for the present study. such as the Tsai-Wu, Chamis, Hoffman and Hill
Failure criteria for composite materials are more criteria. The maximum stress criterion and maximum
difficult to postulate because they are more involved strain criterion are very simple and are called inde-
in terms of structural and material complexity in pendent failure mode criteria. Failure surfaces for
comparison to isotropic materials. Physical proper- maximum stress and strain criteria are rectangular in
ties of the constituent phases can cover extreme stress and strain space, respectively [2]. The failure
ranges. Reinforced materials are high modulus elastic surface for the quadratic polynomial criteria is of
and brittle fibers. whereas the binder materials are ellipsoidal shape. Figure I shows the schematic
ductile and viscoelastic. failure surfaces for maximum stress, maximum
There exist a number of failure criteria and a brief strain and quadratic polynomial criteria. Maximum
description of the criteria developed over the last 50 stress and strain criteria have some physical basis
years is given by Sandhu [I]. A survey of failure whereas the polynomial criteria are mathematical in
criteria as applied to composite materials is presented nature.
by Tsai [2] and by Tsai and Hann [3]. Failure of Different quadratic polynomial criteria differ in the
composite materials can be regarded from micro- way polynomial constants are determined. The most

c+ &
Y
Y
CT

xT
xC
X
X,C CT xC xT

-_
CC yC yC

(a) Maximum strain criterion (b) iiaximum stress criterion (c) Quadratic polynorial criterion

Fig. I. Schematic failure surfaces for maximum strain, maximum stress and quadratic failure criteria.
 A first-ply failure analysis of composite laminates 373

general criterion for composite materials is the tensor where: ai, a>, a3 are the normal stress components; ad,
polynomial criterion proposed by Tsai [7] and Tsai as, a, are the shear stress components; Xr, Yr, Z, are
and Wu [8J: the Iamina normal strengths in the .Y,y, : directions,
respectively; and R, S. Tare the shear strengths in the
F,u, + Fijdia, + F+apjak + . . .z 1.
yr, xr and .xy planes, respectively. When a,. a:, a, are
Here 0; are the stress tensor components of compressive nature, then they should be compared
(frl = eil, a? = oz2, etc.) in the material coordinates with Xc, Y, and Z,, the normal strengths in com-
and F:, F, and &, etc. are the components of the pression in the .r, y and r directions.
strength tensors: all components are referred to the The maximum stress criterion is an independent
material principal axes. mode criterion, For example, if a lamina fails because
One problem in applying the tensor polynomial of the condition a1 > X,, then it can be said that
criterion is the determination of the parameter Flz. mode of failure is yielding of the fiber in the I
The value of F,, is not unique and can vary from a direction. The maximum stress criterion can also be
negative value to a positive value. Narayanaswami expressed in terms of the tensor polynomial criterion
and Adelman [9] showed that F,z can be taken to be as:
zero for certain laminates, without significant
sacrifice in the accuracy. Other researchers have given (0, - X&r, -t X&Q - Yr)(e, + Y&e, - Zr)
the value of F,, in terms of other strength constants. x (az + Zr)(cr, - R)((J~ + R)(a, - S)(ar f S)
It is also shown by Tennyson [IO] that a cubic form
x (a6 - T)(a6 -I- T) = 0. (3)
of tensor polynomial criterion is suited for analyzing
composite laminates under plane stress conditions Comparing eqn (3) with eqn (1) and ignoring the
and for certain lamination schemes. But the effort in higher-order terms, we obtain
determining the sixth-order strength constant F,,,,
F,,? etc. overshadows the gain in accuracy. Hashin
[ 1I] proposed two different failure criteria, one corre-
sponding to the fiber mode failure and the other
corresponding to the matrix mode failure.

2. A REVIEW OF FAILURE I
THEORIES F,=-$ F,,=-* F,=+
S
In the previous section it was stated that a11 the
failure theories are the degenerate cases of the tensor 4 F2
F,, = - --; F,, = _ y; FzF,,= - y. (4)
polynomial failure criterion, 2
The remaining strength constants are zero.
F,a,+Fzo,+F,a,+2F,zcr,a2+2F,,a,a3

+ 2FzJ~2c73+ F,,af + F,,ai + F,,a: 2.2. Maximum strain criterion

+F,a:+F,,a:fF,a~+...21. (1) In the maximum strain criterion, failure is assumed
to occur if one of the following conditions is satisfied:
Terms associated with a,, a5 and a6 which are F4, F,
and F6 are taken to be zero, since shear strengths are
the same for positive and negative shear stress. It is
also assumed here that there is no interaction between
shear stresses and normal stresses, thus F,6, F16 etc.
become zero. Now we look at other failure criteria e,>ZET; %>T& (5)
which were developed independently but can be
deduced from the tensor polynomial criterion. It where: et, .s2,.sj are the normal tensile strains in the
should be noted that these theories are presented in x, y, z directions respectively; al, E$,r+ are the shear
three-dimensional stress space. All of the components strains in the yz, xz and xy planes respectively; X,,
are referred to the material coordinates. Y,, Z,, are the tensiIe strain strengths in the x, y, z
directions and R,, S,, T, are the shear strain strengths
2.1. Maximum stress criterion
in the yz, XI and xy planes, respectively. Expressing
Failure of material is assumed to occur if any one the criterion in the form of the tensor polynomial, we
of the following conditions are satisfied: obtain

a,>Xr; a,>R (et - X,)@l + X&2 - Y&z + Y&Ifs, - Z,)

x (&I+ Z,,) (~4- RJ (~4f RJ (6~- &I

a,>X,; a,>T (2) x 65 + S6)(86- T,) 6%+ r,) = 0. (61

C.A.S. ?I’,-E
374 J. N. REDDY and A. K. PANDEY

Using the compliance matrix and the constitutive

refation, eqn (6) can be expressed in the form of
eon (I):

(7)

In the above expressions: S,,, Sn etc. are the

components of the compliance matrix; Xr, Yr, Zr are
the tensile stress strengths in the X, y, : directions; Xc,
Y,, Z, are the compressive stress strengths in the X,
y, z directions; R, S, Tare the shear stress strengths
in the yz, xz and xy planes; and FF, Ff, F: are the
expressions given for F,, F2, F, in the maximum stress
criterion. All other strength tensor terms are zero.

2.3. Hifl’s criterion

Hill’s criterion can be expressed as

It should be noted that in Hill’s criterion the stress

terms do not appear as linear terms; therefore, the Fi,
F2 and Fs terms are zero. The values of X, Y, Z are
taken as either Xr, Yr, 2, or as XC, Ye, Z,, depending
upon the sign of cr,, Q~and CT,,respectively. Strength
tensors for this criterion are:

F;= 0
 .4 first-ply failure analysis of composite laminates 37.5

2.4. ffoffman’s criterion

Hoffman’s criterion is expressed as
F 13= - U2JXT&ZTZc
1 1 1 1
_- + --- (fl? - Q,Y Fz3 = - 112& Ycz,.& . (13)
2 ( y,yc ZTZC XT& )
All other strength tensor components are zero.

2.6. Chamis’ criterion

1 1 Chamis’ criterion is expressed as

(- &-&)(u, -uz)’
+2 x,x,+
F$,a,r I (14)

or

Ql 02 K 61(JJ K2, U?fll

-K,,-- - yzr1 (15)
XY ” xz

where K,,, K,,, K,, are the strength coefficients which

Hoffman’s criterion is a special case of the tensor vary from material to material [9]. In the tensor
polynomial criterion for the following choice of the polynomial form the strength tensor components are
parameters 4 and FU:
Fi = 0
+‘; F,= +-“; F,= &-$
T xC T yC C

1 1 1
F,, =-. F,, = Fj3 = -
x,x,’ r, 1 1
zTzC
F4=h FS5=-; FM=7
52

F, = A, F,, = $; F,, =&

&z-K,. F,3=--.
43
F21 =
-2, (16)
XY' xz’
1 I 1
F,?=_i _ ---
2 ( xTxC + yTyC zTzC AI1 other strength tensor coefficients are zero.

3. A REVIEW OF THE EQUATIONS

) OF COMPOSITE LAMINATES

1 1 Plate and shell structures made of laminated com-

F,, = - ! --- (11)
-+
2 ( &Zc Y,Y, posite materials are often modeled as an equivalent
single-layer using classical laminate theory, in which
Other strength tensor terms are zero. the thickness stress components are ignored. The
2.5. The Tsai-Wu criterion classical laminate theory is a direct extension of
classical plate theory based on the Kirchhoff hypoth-
The Tsai-Wu criterion is given by esis for homogeneous plates. This theory is adequate
when the thickness to side ratio is small. However,
laminated plates and shells made of advanced
where filamentary composite materials, like graphite-epoxy,
are susceptible to thickness effects because their
effective transverse shear moduli are significantly
smatler than the effective elastic modulus along the
fiber direction. Refined laminate theories are single-
1 1 1 layer theories in which the transverse shear stresses
F,, = 52 x -; F,, = -
x,x,; YT rc zTzC are taken into account. They provide improved
global response estimates for deflections, vibration
frequencies and buckling loads of moderately thick
composites when compared to the classical laminate
376 J. N. REDDY and A. K. PANDEY

theory. Both classical and refined laminate theories,

being single-layer theories, do not account for local
effects such as the fiber misalignment and tet-mi-
nation, edge effects, geometric and material discon-
tinuities.
The theory of laminated anisotropic plates to be
reviewed here is based on an assumed displacement M6..y+ Ml.; - Ql = 0 (19)
field (see [12]). The theory is known to be able to
predict accurately the global behavior. In this theory where N,, Q, and M, are the stress and moment
it is assumed that the stress normal to the mid-plane resultants defined by
of the plate is negligible when compared to the
in-plane stresses, and straight lines normal to the
plate mid-plane before deformation remain straight
but not necessarily normal to the mid-surface after
h’Z
deformation.
The plate under consideration is composed of a
finite number of orthotropic layers of uniform thick-
(Qi, Qz)=
J
-h’?
(u5.cr,)dz.

ness, with principal axes of elasticity oriented arbi- Here a,(i = I,2 . . 6) denote the stress components in
trarily with respect to the plate axes. The .Y and y the laminate coordinates (a, = (T,, Q: = CJ,~,a4 = ur,,
coordinates of the plate are taken in the mid-plane (I~= u,! and (TV= o,~,.).
(CA)of the plate. The displacement field is assumed to If one plane of elastic symmetry parallel to the
be of the form plane of each layer exists, the constitutive equations
for the plate can be written in the form

&(X, y, 2) = o<x;
yf+z$).(x, I’)

Here ur, u2, uj are the displacement in the x, y, t

directions, respectively, u, U, w are the associated The Ai;, B,, D,/ (i.i = 1,2,6), and Ki,(i, j = 4,~) are
mid-plane displacements, and II/, and tiY are the the in-plane, bending in-plane coupling, bending or
rotations about the y and x axes, respectively. The twisting, and thickness-shear stiffnesses, respectively:
strains in the linear theory of plates can be expressed
in the form
Qu
(m)U,z,z*)dz

G-1
2,=x
J m 2”
KiKjQi;l’dz. (22)

Here t, denotes the distance from the mid-plane to

the lower surface of the mth layer, and Ki are the
shear correction coefficients.
The variational form of the eqns (Ig)-(20) is given
by

Note that the transverse shear strains sJ and E$, are

o=
J
n
f6u.xN, + SU,~
Ns + 6c, N6 + c%J,.~
N2 + SW,Q,

constant through the thickness. If a linear distribu-

tion of the transverse shear strains through the
thickness is desired, one must add higher-order terms
in z to the displacements U, and I(~ and/or u,; with
each additional term, an additional dependent vari-
able is introduced into the problem (see [l3, 141).
+
Jr JI-
(&,N,+6u,N~J& + SwVds

+Jr@$~
fM,W (23
Neglecting the body moments and surface shearing
forces, we write the equations of motion in the
presence of applied transverse forces, q, as

Nt .s + NQ = 0 where N,, Mi and Qj are given in terms of the

 A first-ply failure analysis of composite laminates 377

generalized displacements by eqn (2l), and V, ,V,,and stiffness coefficients, and {F) is the column vector
IV,, and A4, and IV, are the shear force, normal and containing the boundary and transverse force con-
tangential in-plane forces, and normal and twisting tributions. The elements of [K] are given in [ll. 151.
bending moments defined on the boundary r, re-
spectively:
5. PROCEDUREFOR FIRST-PLYFAILURE

The finite-element procedure described in the pte-

ceding section can be used to determine the stresses
(in the laminate coordinates) at any point of the
laminate. In general, the laminate coordinates do not
coincide with the material principal axes of the
individual plies. Since the failure criteria described
earlier require the stresses and strains with respect to
the material coordinates of each lamina, a trans-
formation of the stresses and strains from laminate
M, = n,n,M, + 2n,n, M6 + nSvfirMz
coordinates to the lamina material coordinates is
required. The stresses in material coordinates can be
M n,nJM1 - M,) + fwv - “,n.,) Me (24) determined in one of the two ways: (i) compute
111 =

strains in the laminate coordinates, transform them

where fi = (n,, n,) is the unit vector normal to the
to lamina coordinates and use the lamina constitutive
boundary r. The variational formulation indicates
equations to compute stresses; or (ii) compute strains
that the essential (i.e. geometric) and natural bound-
in the laminate coordinates, compute stresses using
ary conditions of the problem are given by:
the lamina constitutive equations with stiffnesses
referred to the laminate coordinates (Q& and then
ESSENTIAL: specify u,, u,, W,$,, +,
transform laminate stresses to lamina stresses. Both
are mathematicalIy equivalent. The second method is
NATURAL: specify N,, IV,, Y, M,,, M, (25) used in the prsent study. If t?,,,denotes the Iamination
angle of the mth layer, then the stresses o:~’ in its
where u, and u,, for example, denote the normal and materiat coordinates can be obtained from the fami-
tangential components of the in-plane displacement
nate stresses crz,cY,cr,,, etc. using the transformation:
vector, u = (u, r).
c\~) = (i, cosz 8, -i- trysin’ 8, -I- 2a,, sin f3, cos 8,
4. A SHEAR-DEFORMABLE
FINITE ELEMENT
aim) = a, sin’ 0, + au co9 i?, - 2a,V,sin B, c0S B,
Here we review a finite-element model associated
with the equations governing laminated composite a&“)= - uXZsin B, + avZcos 6,
plates. Since the details are well documented in the
textbook by Reddy 1151, we only summarize the (Q) = - a,, cos f&4 bvi sin 8,
results.
Consider a unite-element analog, Qlr, of the mid- a$“) = - d, sin 0, cos 0, + by sin 6, cos t3,
plane of the plate, 52. Over a typical element, f2, of the
mesh &, each generalized displacement W is inter- + q,(cos’ 8, - sin”&). (28)
polated spatially by an expression of the form
The Iamina (i.e. material coordinate) stresses or
strains are then used in a chosen failure criterion to
check if each finite element and each lamina has
failed. If the failure criterion is satisfied in a ply of an
where Vi is the value of U at node i, cpi is the element, then the individual contributions, called
finite-element interpolation function at node i, and r failure indices, of each stress component to the tensor
is the number of nodes in the element. For sake of polynomial are computed. The failure indices can be
simplicity, we use the same interpolation function for used to interpret the mode of failure. For example, if
each of the generalized displacements (u,u,w,b(/,,$J the contributions of the al’“) to the tensor polynomial
Substituting eqn (26) into eqn (22), we obtain the is larger than the contributions of other stress com-
following: ponents, then the failure in that ply is assumed to
occur due to the tension or compression in the I
lK1 iA) = i F3. (27) direction depending on the sign of ~1”). The failure
indices also can be used to reduce the stiffnesses in a
Here {AL\)is the column vector of the nodal values of post-first-ply failure analysis.
the generalized displacements, [Kj is the matrix of the The first-ply failure load calculation requires us to
378 J. N. REDDY and A. K. PANDEY

I For load & find the

1 nodal displacements.
set: PC" = PO and DPl = increment

Displacement for Pan =

no

Fig. 2. A flow chart of the computational procedure used to determine the first-ply-failure load.

use an iterative procedure. The procedure involves 6.1. Comparison of the~nite-eieme~t solution with the
solving the stress probiem for an initial load. If one exact soiut~o~
or more elements/plies fail, load is reduced by a To validate the finite-element computer program,
predetermined percentage (20% in the present study) problems that have anaIytica1 solutions are solved.
of the initial load and the failure analysis is repeated. Analytical solutions for cross-ply laminates and
This iteration is continued until the difference be- antis~met~~ angle-ply laminates are given in [12].
tween any two consecutive failure loads is less than The boundary conditions for simply supported cross-
1%. A flow chart of the procedure described is given ply and antisymmetric angle-ply laminates are given
in Fig. 2. in Fig. 3. Composite materials considered here are
made of T300/5208 graphite/epoxy material. Proper-
ties of this material are given in Table 1. The plate is
6. SAMPLE PROBLEM RESULTS
AND DISCUSSION subjected to uniformly distributed load on the top
surface in the z direction. The finite-element solutions
Numerical examples presented below under four for the displacement of the center point of various
different sub-headings are results of the first-ply cross-ply and angie-ply laminates are compared with
failure analysis. Ail of the computations were made the exact solutions in Table 2. The finite-element
on an IBM 3081 computer system in double pre- solutions are in excellent agreement with the exact
cision. solutions.
 A first-ply failure analysis of composite laminates 379

(a) Cross-ply laminates (b) Anti-symmetric angle-ply

Fig. 3. The simply supported boundary conditions for full-plate and quarter-plate models of cross-ply and
antisymmetric angle-ply laminates.

6.2. Laminates subjected to transoerse loading ing. Table 3 contains a list of laminates analyzed here,
A rectangular laminated composite plate is ana- although other laminates have also been used.
lyzed for transverse loading. The plate is made of A quarter of the plate is modeled because of the
T300/5208 graphite/epoxy material, and the dimen- biaxial symmetry, lamination and boundary condi-
sions and boundary conditions are shown in Fig. 4. tions (see Fig. 3). Various finite-element meshes are
Uniformly distributed load is applied on the top used to study the convergence of the solution. Four-
surface of the plate in the negative z direction. A node and nine-node isoparametric rectangular ele-
number of laminates are analyzed under such load- ments are used. Different meshes of the quarter of the
plate are shown in Fig. 5. Tables 4-12 give the
first-ply failure load (in p.s.i.) and failure location for
Table 1. Material properties of various laminates listed in Table 3. Result for various
T300/5208 graphite/epoxy composite
material
meshes are given for the maximum stress, maximum

Properties Values

E, 19.2 x IO6p.s.i. Table 3. List of laminates analyzed for the

E2 1.56 x IO6p.s.i. first ply-failure load
El 1.56 x IO6p.s.i. Lamination
Given
G,z = G,, 0.82 x IO6p.s.i.
name scheme
GZ 0.49 x IO6p.s.i.
012 = 013 0.24 A [45/-45/451,
U1, 0.49 B [45/-4S/90/0/45/90/~5 01,
x, 219.5 x 10’ p.s.i. C [45/-45/O/90/45/0/-45 901,
XC
246.0 x 1O’p.s.i. D [45/0/-45/0/~5/9O/Oi45]s
Y,=Z, 6.35 x IO’p.s.i. E [45/0/-45~0/-45/0/45i0],
Yc=Zc 6.35 x IO3p.s.i. w90/90/01,
R 9.80 x IO’ p.s.i. E [O/901,
S=T 12.6 x 10’ p.s.i. ~0/90/0/901,
Ply thickness h, 0.005 in./ply I [45/-45/45/-453,

Table 2. Comparison of exact and finite-element solution, z-displacement (in.)

for laminated composite plate (9 by 5in.)
Type Uniformly FEM Exact
Lamination of distributed solution solution
scheme laminate load (psi.) (in.) (in.)
[O/901, cross-ply 0.1 1.886 I .884
[O/90/0/901, cross-ply 0.1 0.135 0.134
[0/90/90/01, cross-ply 0.1 0.230 0.229
[45/-45/45/-45], anti-symm. 0.1 0.1087 0.1086
angle-ply
[15/-15/15/-151, anti-symm. 0.1 0.2515 0.2515
angle-ply
[45/451, anti-symm. 0.1 1.6010 1.6006
angle-ply
[l5/-151, anti-symm. 0.1 2.6040 2.6039
angle-ply
380 J. N. &xxv and A. K. PASDEY

Fig. 4. Geometry, computational domain and boundary conditions for composite laminate under
transverse load.

strain, journal, Tsai-Wu and Tsai-Hill criteria. The by element num~r followed by ply number. in the
maximum stress (independent) criterion, in which case of the nine-node element, Gauss points are also
each stress component is compared with correspond- given to denote exactly the point where failure will
ing strength term, is also included. The first-ply occur. For the four-node element there is only one
failure loads are in psi. The failure location is given Gauss point in the center and for the nine-node

2x2 mesh of 1x1 mesh of nine-node

linear elements quadratic elements

1 1
4x4 mesh of 2x2 mesh of nine-node
linear elements quadratic elements
Fig. 5. Various types of linear and nine-node quadratic element meshes.
 A first-ply failure analysis of composite laminates 381

Table 4. First-ply failure load for out of plane loading for [45,‘-45/45], laminate (A)
Four-noded element Nine-noded element
Theory used 2x2 4x4 8x8 1x1 2x2 4x4
Max. stress 0.6376 0.4360 0.2968 0.5048 0.3960 0.2768
(failed location) 1. 3t 3, 3 1.3 l,3(l. I): 1,3(1. I) 1.3(1. I)
Max. strain 0.7200 0.47 I2 0.3184 0.5672 0.4264 0.2960
4,3 16. 3 64 3 I. 3 (1. I) 4.3 (2, 2) 16,3 (2.2)
Hoffman 0.6328 0.4328 0.2944 0.5016 0.3928 0.2744
193 I. 3 1, 3 I, 3(l. I) I.3 (1, I) I, 3 (I. I)
Tsai-Wu 0.6728 0.4656 0.3168 0.5296 0.4216 0.2944
193 I.3 1.3 1,3(1.1) I,3 (1. 1) I.3 (I, 1)
Tsai-Hill 0.2792 0.1752 0.1184 0.2352 0.1624 0.1128
1, 1 1. L 1. 1 I, 1 (I. I) 1, 1(1, I) I. 1(I. I)
Max. stress 0.6696 0.4608 0.3112 0.5264 0.4160 0.2888
(independent) 193 1.3 1.3 l,3(l, 1) l,3(1. I) 1.3(1. I)
t Failure location, given as: element number, ply number.
$ Failure location, given by: element number, ply number (Gauss point).
Note: ply number starts from bottom to top.

Table 5. First-ply failure load for out-of-plane loading for [45/-45/90/0/45/90/-45/0], laminated (B)
Four-noded element Nine-noded element
Theory used 2x2 4x4 8x8 1x1 2x2 4x4
Max. stress 22.858 20.38 1 18.872 20.565 19.900 18.642
(failed location) 1, 16t I, I6 I, 16 I, 16(l. I): 1, 16(1, 0 1, 16(1,1)
Max. strain 26.778 23.778 21.004 24.207 22.524 20.460
I, 16 16, I6 64, I6 1, 16 (I. I) 4, 16 (2.2) 16, 16(2,2)
Hoffman 22.677 20.202 18.704 20.399 19.724 18.476
I, I6 I, I6 I, I6 1. 16(1, 1) 1, l6(l, 1) 1, 16(L I)
Tsai-Wu 24.434 21.984 20.384 20.028 21.479 20.133
I, 16 1, 16 I. 16 1, 16 (1, 1) 1, 16(1, 1) I, 16(1. I)
Tsai-Hill 9.060 7.596 6.986 8.034 7.388 6.903
1, 1 1, 1 1. I 1, I (I, I) 1, 1(I, I) 1. 1(I, I)
Max. stress 24.55 I 22.324 20.628 22.176 21.816 20.367
(independent) I, I6 I, 16 I. I6 1, 16 (1, 1) 1, 16 (I. I) 1, 16(1, I)
t Failure location, given as: element number, ply number.
$ Failure location, given by: element number, ply number (Gauss point).
Note: p!y number starts from bottom to top.

Table 6. First-ply failure load for out-of-plane loading for [45/-45/0/90/45/0/-45/90], laminate (C)
Four-noded element Nine-noded element
Theory used 2x2 4x4 8x8 1x1 2x2 4x4
Max. stress 20.381 18.053 16.695 18.322 17.629 16.485
(failed location) 1, l6t I, I6 1, I6 1, l6(1, 1): 1, 16(1, 1) 1, 16 (131)
Max. strain 23.853 21.572 19.160 21.532 20.477 18.677
I, I6 16, I6 64. I6 1, 16 (I, I) 4, 16 (2.2) 16, I6 (2,2)
Hoffman 20.22 1 17.893 16.546 18.175 17.472 16.338
1, 16 I, 16 1, 16 1, 16 (1, I) I.16 (I, I) I, 16(1, I)
Tsai-Wu 21.781 19.477 18.037 19.616 19.031 17.808
I, I6 I, 16 I, I6 1, 16(1, 1) 1, l6(1, I) 1. 16(1, 1)
Tsai-Hill 8.095 6.725 6.175 7.186 6.541 6.100
I, I 1, 1 1, I 1, I (I, I) 1, 1(I, I) 1, 1(I, 1)
Max. stress 21.853 19.722 18.216 19.712 19.280 17.983
(independent) 1, I6 I, I6 I, I6 1, l6(2,2) 1, 160, I) 1,16(l, 1)
t Failure location, given as: element number, ply number.
$ Failure location, given by: element number, ply number (Gauss point).
Note: ply number starts from bottom fo top.

element there are four Gauss points. Figure 6 shows strain criterion and the Tsai-Hill criterion, predict
the scheme used to number the Gauss points for a almost the same first-ply failure load and location.
nine-node element. The ply numbering scheme is also The maximum strain criterion predicts a different
shown in the same figure. It is observed that with failure location but the same first-ply failure load.
mesh refinement, first-ply failure load converges to a The Tsai-Hill criterion predicts a much lower value
lower value. All failure criteria, except the maximum of first-ply failure load and a different failure lo-
382 J. N. REDDY and A. K. PANDEY

Table 7. First-ply failure load for out-of-plane loading for [.$j O,‘-tj;O;lj.90/0/4j]s
laminate (D)
Four-noded element Sine-noded element

Theory used 2X’ 4x4 8x8 IX1 .2X’ 4x4

Max. stress 17.41’ 15.032 14.151 15.685 11.669 If.015

(failed location) I. 16t 1, 1s I, I5 I. lj(l, I): I. IS(1, I) I. lj(I, I)
I. 16(1, I)
Max. strain 19.168 16.330 lS.392 17.085 15.933 15.247
I, 15 1. 1s I, 1s 1. IS(I, I) I. lS(l,l) I. 15(1, I)
Hoflman 17.274 14.970 14.090 l5.j60 l-1.608 13.956
1, I6 I, 1s 1,15 1. I6 (I, 1) 1, 15 (1, I) 1. i5(1, 1)
Tsai-Wu 18.212 15.544 14.644 16.244 IS.168 f4.504
1, I5 1. 15 1, 15 I. is (1. 1) I. 1s (1, if 1. fS(i, I)
Tsai-Hiff 6.877 5.764 5.389 6.106 5.618 5.336
l. 1 1,I I.1 1, i (1. 1) I. I (I, I) I. I (I, I)
Max. stress 17.805 15.180 14.292 15.874 14.812 14.156
(independent) 1,lS I, 1s I, 15 1, 15 (2,2) I, 15 (I. I) I. 15 (I, 1)
t Failure location, given as: element number. ply number.
$ Failure location, given by: element number, ply number (Gauss point)
Note: ply number starts from bottom to top.

Table 8. First-ply failure load for out-of-plane loading for [.$j ‘0;-45/0/4j 0:4~/0],
laminate (E)
Four-noded element Nine-noded element

Theory used 2x2 4x4 8x8 1x1 2x2 4x4

Max. stress 16.556 14.285 13.365 14.914 13.933 13.226
(failed location) f, 167 1, 1s 1. 15 f. l6(1, it:. 1. lS(1. I) 1. tS(l, I)
Max. strain 18.277 15.535 14.556 16.308 lS.ISf f4.405
1, 15 f,lS 1, I5 f. 15(i, I) 1, iS(l, 1) I, lS(l,l)
Hoffman 16.424 14.224 13.308 I-t.792 13.874 13.168
I, 16 1, 15 I, 15 I. i6(1, 1) I. 15(1, 1) I, 15(i, I)
Tsai-Wu 17.362 14.780 13.840 15.498 14.416 13.696
I, 15 I, I5 I, 15 i.l5(1, I) I, ij(l,l) 1,15(1,1)
Tsai-Hill 6.556 5.466 5.079 5.828 j.234 5.026
1,l 1.1 I,1 1,l (I, 1) 1. I(1, 1) 1, l(1, 1)
Max. stress 16.975 14.429 13.503 lj.146 iJ.072 13.364
(inde~nde~t) I, 1.5 1, 1s 1, IS I. 15(1, 1) 1. 15 (I, 1) I, IS(l, 1)
t Failure location, given as: element number, ply number.
$ Failure location, given by: element number, ply number (Gauss point).
Note: ply number starts from bottom to top.

Table 9. First-ply failure load for out-of-plane loading for [0~90/90~0~laminate (F)
Four-noded element Nine-noded element

Theory used 2x2 4x4 8X8 1x1 2x2 4x4

Max. stress 0.7064 0.6024 0.5816 0.6216 0.5864 0.5768
(failed location) 1,4t 1,4 1.4 1.4(1, I)$ 1,4(1,1) 1,4(1,1)
Max. strain 0.7704 0.6552 0.6312 0.6776 0.6376 0.6280
I,4 I,4 1,4 l,4(1, I) 1,4(1, 1) 1,4(1,.1)
Hoffman 0.7032 0.5992 0.5784 0.6184 0.5832 0.5752
1,4 1.4 I,4 1,4(I, 1) 1.4(1,1) 1,4(l, 1)
Tsai-Wu 0.7320 0.6232 0.6008 0.6440 0.6072 0.5976
1.4 I,4 I,4 J,4(1,1) 1,4(1,1) 1,4(l,l)
Tsai-Hill 0.3800 0.3288 0.3176 0.3352 0.3208 0.3160
1, 1 I.1 i,l 1, IV, 1) 1, i (1, 1) I, f (i,l)
Max. stress 0.7144 0.6088 0.5864 0.6296 0.5928 0.5832
(independent) I,4 I.4 1,4 i,4(1,1) 1,4(l,i) i,4(1,1)
t Failure location, given as: element number, ply number.
$ Failure location, given by: element number, ply number (Gauss point).
Note: ply number starts from bottom to top.
 A first-ply failure analysis of composite laminates 383

Table IO. First-ply failure load for out-of-plane loading for [O/90], laminate (G)
Four-noded element Nine-noded element
Theorv used 2x2 4x4 8x8 1x1 2x2 4X1

Max. stress 0.4568 0.4072 0.3848 0.4056 0.3720 0.3800

(failed location) I,Zt 3,2 5.2 1.2 (I, I): 2,2(l, I) 3,2(1. I)
Max. strain 0.2424 0.2072 0.2080 0.2120 0.2024 0.1992
I,1 I.1 I.1 I. 1(I, 1) I, I (I, I) I, I (I. I)
Hoffman 0.4552 0.4072 0.3832 0.4040 0.3704 0.3784
I,2 3,2 5.2 l,2(l, 1) 2.2(1, I) 3,2(l, I)
Tsai-Wu 0.4696 0.4152 0.3928 0.4184 0.3784 0.3880
I.2 3,2 5,2 I, 2 (I. I) 2.2(1, I) 3,2(1, I)
Tsai-Hill 0.2248 0.2088 0.2056 0.2040 0.2072 0.2040
1, I I.1 I.1 1, 1(I, I) 1, I (I. 1) I. I (I, I)
Max. stress 0.4616 0.4120 0.3864 0.4104 0.3752 0.3816
(independent) I.2 3,2 5,2 l,2 (I, I) 2,2(1, I) 3,2(1, II
t Failure location, given as: element number, ply number,
$ Failure location, given by: element number, ply number (Gauss point).
Note: ply number starts from bottom to top.

Table I I. First-ply failure load for out-of-plane loading for [O/90/O/90j-rlaminate (H)
Four-noded element Nine-noded element
Theory used 2x2 4x4 8x8 1x1 2x2 4x4
Max. stress I .9992 I .8600 I.8184 I .7688 I .8456 I.8104
(failed location) 1,4t 2,4 3,4 l,4(1, I): l,4(1, I) 2,4(1, I)
l,4(2, I)
Max. strain 1.5240 1.3160 I .2760 1.3400 I .2856 I .2696
1, 1 1, 1 1, 1 1, 1(I, 1) 1, 1(I, 1) 1%IU, 1)
Hoffman I .9720 I .8360 1.7928 1.7448 I.8184 I .7848
1.4 294 3,4 1,4(1, I) wr, 1) 2,4(l, 1)
Tsai-Wu 2.2664 2.1000 2.0600 2.0088 2.0840 2.0552
194 294 4,4 1,4(l.l) l,4(2, I) 2,4(l, 1)
Tsai-Hill 0.6744 0.6264 0.6136 0.5976 0.6184 0.6104
1, 1 1, I 2, 1 1, 1(I, 1) 1, 1(I, I) 2, 1(1, 1)
3, I
Max. strew 2.3096 2.1096 2.0504 2.0440 2.0744 2.0420
(independent) 1.4 2,4 494 l,4(l, 1) 2,4(l, I) 2,4(2, I)
t Failure location, given as: element number, ply number.
# Failure location, given by: element number, ply number (Gauss point).
Note: ply number starts from bottom to top.

Table 12. First-ply failure load for out-of-plane loading for [45/45/45/-45],
laminate (I)
Four-noded element Nine-noded element
Theory used 2x2 4x4 8x8 IX1 2x2 4x4
Max. stress 1.7313 1.4484 1.3902 1.5201 1.4126 1.3825
(failed location) 1,4t I,4 I,4 1.4(1,1): l,4(1, I) 1,4(l,l)
Max. strain 2.3015 1.6622 1.5899 1.7710 1.6187 1.5803
I,4 I,4 I,4 1,4(1,1) 1,4(1,1) 1,4(l, I)
Hoffman 1.7179 1.4382 1.3812 1.5086 I .4030 1.3736
I,4 1.4 1.4 l,4(l, 1) 1,4(1; 1) 1,4(1. I)
Tsai-Wu 1.8510 1.5323 1.4683 1.6200 1.4932 1.4593
I,4 1.4 I,4 l,4(1,1) 1,4(1, I) 1,4(1.1)
Tsai-Hill 0.7009 0.6235 0.6056 0.6273 0.6107 0.6024
l,l I,1 1,l I, 1(1, 1) 1, I(1.1) 1, I(19 1)
Max. stress 1.8664 1.5425 1.4772 1.6308 1.5028 1.4683
(independent) 1.4 1,4 I,4 1,4(1, I) 1,4(1,1) 1,4(l, I)
t Failure location, given as: element number, ply number.
$ Failure location, given by: element number, ply number (Gauss point).
Note: ply number starts from bottom to top.
383 J. N. REDDY and A. K. PANDEY

cation. The reason behind predicting a lower value of

first-ply failure load by the Tsai-Hill criterion under
transverse loading can be attributed to the fact that
the second-order strength tensor terms F, are zero in
this case.
It is also observed that for all laminates considered
here failure occurs in the top iayer. This is due to the
fact that the top or bottom layers have the maximum
stress, and in these laminates the maximum stress
occurs in the top layer. Failure occurs in the element
which is adjacent to the center of the plate.
First-ply failure load variation for an anti-
symmetric angle-ply laminate [O/-O/ . . . ] is plotted in
Fig. 7. Lamination angle 6 varies from 0’ to 90..
Three different curves correspond to the total number
of lamina in the laminate (i.e. two, four and six). The
ayer n first-ply failure load plotted here correspond to
Tsai-Wu failure criterion.
Lamina orientation in a laminate effects the
first-ply failure load for transverse loading. Figure 8
Layer 1 shows the variation of the first-ply failure load ob-
tained by failure criteria for [45/O/90],, [45/90/O], and
Fig. 6. Scheme used to number the Gauss points and layers [90/45/O], laminates under uniformly distributed
in a nine-node element. loading. Lamina orientations are changed and corre-

‘-\, n=6

Angle of lamina orientation, 0

Fig. 7. First-ply failure load vs the lamination angle for simply supported antisymmetric angle-ply
laminates under uniformly distributed transverse load (using the Tsai-Wu criterion).
 A first-ply failure analysis of composite laminates 385

l-Maximum stress
2-Maximum strain
3-Hoffman's
4-Tsai-Wu
5-Tsai-Hill
6-Max. stress (ind.)

.23r 16

Fig. 8. Variation of the first-ply failure load with the lamination changes in symmetric laminates subjected
to uniform transverse load.

sponding values of first-ply failure load are plotted for symmetric and cross-ply laminates. The whole
for different failure criteria. plate is analyzed in the case of antisymmetric angle
ply laminates. Dimensions and boundary conditions
6.3. Laminates subjected to in-plane loading of the plate are the same as in Figs 3 and 4, except
The laminated composite plates discussed in the that loading is on the edge perpendicular to the x
previous section are analyzed here for uniformly axis. Various finite-element meshes (see Fig. 5) are
distributed tensile loading on the two edges parallel used to study the convergence of failure loads. All the
to the _r direction. A quarter of the plate is analyzed laminates listed in Table 3 are analyzed. Tables 13-21

Table 13. Firstr-ply failure load for in-plane loading for [45/-45/45], laminate (A)
Four-noded element Nine-noded element
Theory used 2x2 4x4 8x8 Ix1 2x2 4x4
Max. stress 2854.40 2164.32 1908.16 2262.24 1906.08 1701.44
(failed location) 2, It 4, 1 48, I 1,1(2, 11: 4, f(2, 1) l6,3(2,1)
2.3 493 48,3 l,3(2, 1) 4,3(2, 1) 16,IQ, 1)
Max. strain 2947.68 2268.64 1940.48 2441.92 1985.44 1736.64
4,3 16, I 6473 1, f(2, 1) 4, l(2, I) l6,3(2,1)
4, 1 16,3 6491 L3(2, 1) 4,3(2, 1) 16,IV, 1)
Hoffman 2850.24 2156.80 1905.76 2259.68 1906.24 1703.20
2, 1 4,1 48, I 1, ](2, 1) 4, l(2, 1) 16.3(2,1)
2.3 4,3 48,3 I,3 (2, I) 4.3 (2, 1) 16,1(2,1)
Tsai-Wu 2886.72 2218.88 1917.76 2282.24 1905.12 1687.84
2, 1 12, I 56,3 1, IQ, 1) 4, J (2, 1) 16,3(2,1)
273 12,3 56, 1 1,3(2, 1) 4,3C 1) 16,1(2, 1)
Tsai-Hill 2788.80 1803.84 1530.40 2308.32 1710.72 1526.88
1, 1 1.2 1,2 1, I (1, 1) I, 2(1, 1) 1,2(1,1)
193 t, 3 (1.1) 1,2(2,I)
Max. stress 2886.72 2218.88 1917.76 2282.24 1905.12 1687.84
(independent) 2, I 12, I 56,3 1, 1(2, 1) 4, IQ, 1) 16,3(2,1)
2.3 12.3 56. I 1.3(2. 1) 4.312. II 16.l(2.1)

t Failure location, given as: element number, ply number.

$ Failure location, given by: element number, ply number (Gauss point).
Note: ply number starts from bottom to top.
386 J. N. REDDYand A. K. PASDEY

00
X
m
 A first-ply failure analysis of composite laminates

d
X
*o

N
X
N

30
X
m

0
X
d

N
X
r-4
388 3. N. REDDY and A. h. PANDEY

Table 18. Firs:-ply failure load for in-plane loading for [0 90:‘90,‘0],laminate (F)
Four-noded element Nine-noded element
Theory used 2x2 4x4 8x8 1x1 2x2 4x4
Max. stress 4349.12 3599.36 2760.96 3869.44 3335.36 2824.00
(failed location) 2,2t 16,2 64,2 1.2(2,1): 1.2(2,1) 16,2(2,1)
273 16,3 64.3 1,3(2, I) 4,3(2, 1) 16,3(2, 1)
Max. strain 4397.12 3784.00 2881.60 3940.80 34S7.60 2968.00
2.2 16.2 64.2 1.2(2, 1) J.2(2, I) 16,2(2, I)
293 16,3 643 1.3(2, I) 4,3(2, I) 16,3(2, 1)
Hoffman 4352.00 3599.68 2760.97 3572.00 3336.64 2824.64
2,2 16.2 64.2 1,2(2.1) 4.2(2,1) l&2(2,1)
2,3 16.2 64,3 I. 3 (2, 1) ‘4.3 (2, 1) 16.3 (2, 1)
Tsai-Wu 4326.40 3597.44 2760.00 3849.92 3325.12 2818.88
2,2 16.2 64,2 1,2(2.1) 4.2(2.1) 16 3(2 1)
2,3 16.3 64,3 1.3(2, 1) 4,3(2, I) 16:&2: 1)
Tsai-Hill 4352.64 3968.32 3008.00 3934.08 3503.68 3063.04
2.2 16.2 64,2 i.?(l. I) 4,2(2. 1) 16,2(2, 1)
2,3 16, 3 64.3 1.3(1. 1) 4.3(2, 1) 16,3(2, 1)
Max. stress 4363.52 4353.60 3200.96 4064.00 3933.12 3440.64
(independent) 2,2 4.2 64,l I. 2 (2. I) 4.2 (2, 1) 16. 1 (2.2)
2,3 4,3 64,2 1.3 (2. 1) 4, 3 (2, I) 16,2 (2,2)
64.3 16~3(2,2)
64.4 16.4 (2,2)
t Failure location, given as: element number. ply number.
$ Failure location, given by: element number. ply number (Gauss point).
Note: ply number starts from bottom to top.

Table 19. First-ply failure load for in-plane loading for [O/SO], laminate (G)
Four-noded element Nine-noded element
Theory used 2x2 4x4 8x5 Ixf 2X2 4x4
Max. stress 893.78 718.68 649.50 822.41 694.83 634.64
(failed location) 2,2t 4.2 62 1.2(2, I): 2,2(2,1) 4.2 (2, 1)
Max. strain 913.78 734.46 664.19 842.11 711.19 651.00
2,2 4,2 892 1.2(2, I) 2,2(2, I) 4.2 (2, 1)
Hoffman 893.72 718.62 649.38 822.41 694.77 634.58
2-2 4,2 8,2 1.2(2,1) 2,2(2,1) 492 (2, I)
Tsai-Wu 894.4 1 719.91 650.25 822.82 695.29 635.27
2,2 4.2 872 1,2(2, I) 2,2(2,1) 492 (2, 1)
Tsai-Hill 894.76 718.91 649.61 826.79 697.36 637.46
2,2 4.2 &2 l,2(2, 1) 2,2(2,1) 4,2 (2, 1)
Max. stress 896.32 719.20 649.61 834.40 701.63 640.17
(independent) 232 4,2 8,2 l,2(2,1) 2,2(2,1) 4, l(2, 1)
t Failure location, given as: element number, piy number.
$ Failure location, given by: element number, ply number (Gauss point),
Note: ply number starts from bottom to top.

Table 20. First-ply failure load for in-plane loading for [0~90~0/90]~laminate (H)
Four-noded element Nine-noded eiement
Theory used 2x2 4x4 8x8 IX1 2x2 4x4
Max. stress 3029.44 2741A4 2249.60 2744.32 2360.00 2036.48
(failed location) 2,4t 4.4 64,4 1,4(2, 1): 4,4(2, 1) 16,4(2, 1)
Max. strain 3113.28 2809.28 2370.21 2832.96 2464.96 2151.68
2,2 4.2 64,4 1,4(2, 1) 4,4(2, 1) 16,4(2, 1)
Hoffman 3028.16 2740.80 2249.60 2743.04 2359.04 2035.84
2,4 4,4 64,4 1.4(2, 1) 4,4(2, 1) 16,4(2, 1)
Tsai-Wu 3039.68 2747.20 2250.24 2752.64 2365.44 2039.68
2.4 4,4 64.4 194 (2. 1) 4.4 (2, 1) 16.4 (2, 1)
Tsai-Hill 3034.88 2743.36 2471.36 2775.36 2452.16 2182.40
2.4 4.4 64.4 1,4f2, 1) 4,4(2, 1) l6,4(2, 1)
Max. stress 3041.92 2744.64 2600.32 2830.72 2659.84 2578.56
~ind~ndent) 2,4 4.4 8,4 l,4(2, 1) 4,4(2, 1) 4,4(2, 1)
t Failure location, given as: element number, ply number.
$ Failure location, given by: element number, ply number (Gauss point).
Note: ply number starts from bottom to top.
 A first-ply failure analysis of composite laminates 389

Table 21. First-ply failure load for in-plane loading for [45:45 45’451-453, laminate
(I)
Four-noded element Nine-noded element
Theory used 2x2 4x4 8x8 1x1 2x2 4x4
Max. stress 2736.64 2087.20 1593.28 3530.72 ‘113.12 1724.32
(failed location) I, 2: t.1 1,I I.‘(l, 1% 1. 1(I. 1) 1, I (I. I)
1,4 4,4 8.4 1.1(1. 1) 2.4(2, 1) 4,4(2, I)

::: 16, I
13,4 64,l
57,4 1.3(1
I. l(i.2)
>-‘) 4.I(”
3.-t(i,Z)
-,- 7) 16. I(7-.._
13.4(1.‘) ‘)
sym pt. sym pt.
Max. strain 3134.08 2334.88 1730.08 ‘813.20 2344.96 1848.16
I,2 1, I 171 1.2(1, I) I. l(1.1) 1, [(I. 1)
I,4 434 894 1.1(1. I) 2.1(2, 1) 4,4(2, I)
2, 1 13,4 57,4 1. l(1.2) 3.-1(1,2) 13,4(1,2)
2.3 16, I 64,l 1.3(1,2) 4. l(2.2) 16, l(2.2)
sym pt. sym pt.
Hoffman 272944 2078.56 1588.96 2526.40 2105.92 1721.44
1,2 I.1 I.1 I.2(1, 1) I, [(I, I) I. I (I, 1)
4,4 8.4 1.-1(1. 1) 2.4(2, I) 4,4(2. I)
:?Y
2:3 16,l
13,4 64,l
57,4 1.3
l.l(l.2)
(1. 2) 4. I (2.2)
3.4(1,2) 16. I (2.2)
13.4(1,2)
sym pt. s)m pt.
Tsai-Wu 2195.68 2152.00 1629.28 2566.72 2172.16 1753.12
1.2 1.1 I,1 1.2 (I, I) 1. I(1, I) I, l(1, I)
1.4 4,4 8,4 I.l(l. 1) 2.4(2,1) 4,4(2, I)
21 13,4 57.4 1. I (1,2) 3.1(1,2) 13,4(1,2)
2,3 16,1 64,l 1.3fI.2) 4. I (2,2) 16. 1(2,2)
sym pt. sym pt.
Tsai-Hill 3033.28 2195.20 1649.44 2S11.52 2’18.24 1780.48
1,2 1,l 1, 1 I.l(l,l) l*l(l,l) I,i(l,l)
:*‘: 4,4
13,4 8,4
57,4 I. 1(1,2)
I.J(l, 1) 2.412,
3.4(1,2)1) 4,4(2,
13.4(1,2)f)

2:3 16, I 64.1 I.j(l.2) 4.1(2,2) i&1(2,2)

sym pt. s>m pt.
Max. stress 4083.04 2438.56 1756.00 3530.08 X50.08 1887.04
(independent) 1.2 I,1 1, 1 l.‘(l, 1) 1. I(1, 1) I, l(1, 1)
4,4 8,4 1.4(1, 1) 2.4(2, I) 4,4(2, I)
:,‘:
213 16, I
13,4 64, I
57,4 1(1,2)
1.3(1,2)
I. 4.1(2,2)
3.4(1,2) 13,4(1,2)
16,1(2.2)
sym pt. sym pt.
t A whole (and not a quarter) plate is analyzed, so meshes correspond to full plate.
: Failure location, given as: element number, ply number.
5 Failure location, given by: element number, ply number (Gauss point).
Note: ply number starts from bottom to top.

contain the first-ply failure loads and failure (in lb) angle-ply laminates [@/-0/. . . ] are plotted in Fig. 9
locations corresponding to various finite-element for the present loading case. Here the whole plate is
meshes and failure criteria. analyzed because one cannot assume the symmetric
Mesh refinement results in convergence of the displacement field for such laminates. Lamination
first-ply failure load. For the same mesh, the nine- angle 6’ varies from 0” to 90”. A change of lamina
node element gives better convergence of the first-ply stacking sequence does not cause a change in first-ply
failure load. Failure location is indicated by element failure load for in-plane loading. So a laminate which
number followed by ply number. For the nine-node has two lamina each of 0”, 4.5” and 90”, and is
element, Gauss points are also indicated to show symmetric, will have the same first-ply failure load
exact location of failure in an element. It is observed irrespective of where these lamina are stacked in the
that the first-ply failure may occur at more than one laminate.
location in the laminate for in-plane loading. This
observation is different from the transverse loading 6.4. Laminates with hole under in-plane loading
where failure occurs on the top or the bottom of the A rectangular composite plate with hole in the
plate. The weakest plies fail first. All failure criteria, center is analyzed. Dimensions, boundary conditions
including the Tsai-Hill criterion, predict the same and the finite-element mesh are shown in Fig. IO.
first ply failure load. Recall that the Tsai-Hill crite- Since all laminates analyzed in this case were sym-
rion predicted very low values of the first-ply failure metric, only two degrees of freedom (U and v) are
load in the case of transverse loading. taken at each node. Only a quarter of the plate is
First-ply failure load variations for antisymmetric analyzed.

C.&S.
ZS,?--F
 n=6
12,000
\ 5 in.

la,000

-gin--d

(a) Geometry and loading

Elemen
No. 11

2,000

0
0 15 30 45 60 75 90 (b) Computational donlain and boundary
conditions
Lamination angle, (in deg.)

Fig. 9. Variation of the &iilure toad with the lnmination ungk? for antisymmetric Fig. 10. The geometry, computational domain zrnd boundary conditions used in the
angie-ply laminates with uniform in-plane edge load (the Tsai-Wu criterion was used). plane laminate problem.
 A first-ply failure analysis of composite laminates 391

Table 22. First-ply failure results of laminates subjected to in-plane loads

First-ply
failure Percentage Ply Element Failure
Theory load errort number: number@ index
[45/-45/453,
(I) Max. stress 912.0 -00.86 193 II I .00670
(2) Max. strain 1012.0 10.00 1, 3 II 1.00470
(3) Hoffman 908.0 - 1.30 1.3 II I.0021 I
(4) Tsai-Wu 920.0 0.00 1. 3 II 1.00182
(5) Tsai-Hill 1008.0 ~45l-45,90,0,45,9~~~5,0]
9.56 II I.00199

(I) Max. stress 6738.4 1.73 3,6, II, I4 11 1.00002

(2) Max. strain 6652.0 0.43 3,6, II, I4 II 1.00118
(3) Hoffman 6760.0 2.06 3,6, II, 14 II 1.001IO
(4) Tsai-Wu 6623.2 0.00 3,6, II, I4 Ii 1.00057
(5) Tsai-Hill 6738.4 1.73 3.6, II. 14 11 1.00140
[45/-45/0/90/45/0/-45/90),
(I) Max. stress 6738.4 1.73 4,8,9, I3 II l.OOOo2
(2) Max. strain 6652.0 0.43 4,8,9, 13 II I.001 18
(3) Hoffman 6760.0 2.06 4,8,9, 13 II I.001 IO
(4) Tsai-Wu 6623.2 0.00 4,8,9, 13 II 1.00057
(5) Tsai-Hill 6738.4 1.73 4,8,9, I3 II 1.00140
[45/0/-45/0/-45/90/0/45],
(I) Max. stress 8294.4 4.80 6, II II I BOO53
(2) Max. srain 8185.6 -0.17 6, II II 1.00080
(3) Hoffman 8423.2 2.72 6, II II 1.00085
(4) Tsai-Wu 8200.0 0.00 6, II II l.O@l71
(5) Tsai-Hill 8387.2 2.28 6, II II 1.00013
[45/0/~5/0/-45/0/45/0],
(I) Max. stress 14550.4 -3.13 1,7, 10, I6 II 1.OoOO5
(2) Max. strain 16849.6 12.17 1,7, 10, 16 II 1.00031
(3) Hoffman 14492.8 -3.51 1,7, IO, I6 II 1.00001
(4) Tsai-Wu 15020.8 0.00 1.7, 10, 16 II 1.00031
(5) Tsai-Hill 14876.8 -0.95 1.7, IO, 16 II 1.00026
t Relative to the load obtained by the Tsai-Wu criterion.
1 Plies are counted starting from bottom to the top of laminate.
§See Fig. 9 for the location of the elements.

Table 22 contains the first-ply failure loads (in lb), 7. SUM,MARY AND CONCLUSIONS
failure locations and corresponding failure indices for
six different laminates. Result for maximum stress A number of existing failure criteria, including the
criterion, maximum strain criterion, Hoffman’s crite- maximum stress, the maximum strain, Hill’s, the
rion, the Tsai-Wu criterion and the Tsai-Hill crite- Tsai-Wu and Hoffman’s, are reviewed and are in-
rion are included. All the failure criteria give essen- cluded in a first-ply failure, finite-element analysis
tially the same result. procedure for the static analysis of laminated com-
Critical elements and critical plies for a posite plates subjected to in-plane and/or transverse
[90/45/-45/O], laminate subjected to uniform edge loads. All criteria are cast in a general tensor poly-
load of 5000 lb are presented in Fig. 11. Failure indices nomial form. The finite-element program allows the
for different elements and plies denote the intensity of user to specify the desired failure criterion and ele-
the stress field. Figure I2 is a visual depiction of the ment type, and yields the first-ply failure results (i.e.
change in first-ply failure load for different lamina failure load, ply and element in which failure occurs).
orientation in a laminate. It is seen that laminates From the present evaluation of various failure
[O/45], and [O/90]s are very strong in axial loading, theories, it can be concluded that all failure criteria
while laminate [90/45], is weak. The first-ply failure are equivalent in predicting the failure when lami-
occurs in element number 11 for all the laminates. nates are subjected to in-plane loads. For laminates
Table 23 compares the in-plane first-ply failure load subjected to transverse load, the maximum strain and
for different laminated plates with and without a Tsai-Hill criteria predict different failure location
hole. The first-ply failure load for composite plates and failure load to the other criteria. A number of
with a hole are much lower than for plates without laminates are studied for the first-ply failure. For
a hole. This is due to the effect of stress concentration antisymmetric angle-ply laminates, which have
at the minimum section of the laminate. bending-stretching coupling, the whole plate is ana-
392 J. N. REDDYand A. K. PANDEY
 A first-ply failure analysis of composite laminates 393

Table 23. Comparison of first-ply failure load for composite 3. S. W. Tsai and H. T. Hahn, Failure analysis of com-
plate with hole and without hole under in-plane loading posite materials. In fneiartic Behavior *j Composite
Material. ASME. New York 119751.
With hole Without hole 4
R. Talreja, A continuum mec~anics’characteri~ation of
Laminate (lb) (lb) ’ damage in composite materials. .A general report.
[45/-45/45h 912.00 1687.84 5. D. C. Drucker. Yielding flow and failure. In /nelasric
[-l5~-45~90~0~45/90/-4510], 6623.20 11475.20 Behaviour of Composite Marerials. Vol. 13. pp. l-12.
[45!-45~0/90:45/0/-l5/90], 6623.20 11475.20 ASME, New York (1975).
[45/0/-l5/0/-45/90/0/45], 8200.0 14086.40 6. S. R. Soni, A new look at commonly used failure
[45/0/-l5;0/~5/0/45/01, 15020.80 17531.20 theories in composite laminates 24th AIAA/ASME/-
ASCEIAHS Sfructures. Strucrural Dynamics and Mare-
rials Conference, Proc., p. I71 (1983).
7. S. W. Tsai, Strength characteristics of composite
lyzed. For such a laminate, failure loads and failure materials. NACA CR-224 (1965).
locations are tabulated. 8. S. W. Tsai and E. M. Wu, A general theory of strength
for anisotropic materials. J. Composite Mafer. 5, 58-80
A natural extension of the present study would be (1971).
to include the geometric non-linearity and a post- 9. R. Narayanaswarmi and H. M. Adelman, Evaluation of
first-ply failure analysis with stiffness reduction. the tensor polynomial and Hoffman strength theories
These extensions will constitute the objective of sub- for composite materials. J. Composire Mater. 11, 366
(1977).
sequent studies. 10. R. C. Tennyson, D. MacDonald and A. P. Nanyaro,
Evaluation of the tensor poiynomjal failure criterion for
Acknoluledgemenrs-The results included in this report were composite materials. J. Composire Mater. 12.63 (1978).
obtained during a study supported by the U.S. Army 11. Z. Hashin, Failure criteria for unidirectional fiber corn-
(IMICOM), Structures Directorate, Redstone Arsenal, Ala- posites. J. Appl. Mech. 47, 329-334 (1980).
bama. The encouragement and helpful discussions with 12. J. N. Reddy, Energy and Yariarional Methods in Applied
Donald Sandidge are gratefully acknowledged. Mechanics (with an introduction IO rhe Finite Elemen!
Merhod). John Wiley, New York (1984).
13. J. N. Reddy, A refined nonlinear theory of plates with
REFERENCES transverse shear deformation. Inr. J. Soli& Sfrucf. 23,
319-330 (1984).
I. R. S. Sandhu, A survey of failure theories of isotropic 14. J. N. Reddy, A simple higher-order theory for lami-
and anisotropic materials, Technical Report, AFFDL- nated composite plates. J. Appl. Mech. 51, 745-752
TR-72-7 I, (1984).
2. S. W. Tsai, A survey of macroscopic failure criteria for 15. J. N. Reddy, A penalty plate bending element for the
composite materials, Technica/ Reporr. AFWAL-TR- analysis of laminated anisotropic composite plates. Int.
84-4025. J. Numer. Meth. Engng 15, Il87-1206 (1980).