Finite Elements in Analysis and Design 48 (2012) 13461357

Contents lists available at SciVerse ScienceDirect

Finite Elements in Analysis and Design

journal homepage: www.elsevier.com/locate/finel

A higher order nite element including transverse normal strain for linear
elastic composite plates with general lamination congurations
Wu Zhen a,b, S.H. Lo b,n, K.Y. Sze c, Chen Wanji a
a

Key Laboratory of Liaoning Province for Composite Structural Analysis of Aerocraft and Simulation, Shenyang Aerospace University, Shenyang 110136, China
Department of Civil Engineering, University of Hong Kong, Pokfulam Road, Hong Kong, China
c
Department of Mechanical Engineering, University of Hong Kong, Pokfulam Road, Hong Kong, China
b

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 5 September 2010
Received in revised form
18 March 2011
Accepted 7 August 2011
Available online 31 August 2011

This paper describes a higher-order globallocal theory for thermal/mechanical response of moderately
thick laminated composites with general lamination congurations. In-plane displacement elds are
constructed by superimposing the third-order local displacement eld to the global cubic displacement
eld. To eliminate layer-dependent variables, interlaminar shear stress compatibility conditions have
been employed, so that the number of variables involved in the proposed model is independent of the
number of layers of laminates. Imposing shear stress free condition at the top and the bottom surfaces,
derivatives of transverse displacement are eliminated from the displacement eld, so that C0
interpolation functions are only required for the nite element implementation. To assess the proposed
model, the quadratic six-node C0 triangular element is employed for the interpolation of all the
displacement parameters dened at each nodal point on the composite plate. Comparing to various
existing laminated plate models, it is found that simple C0 nite elements with non-zero normal strain
could produce more accurate displacement and stresses for thick multilayer composite plates subjected
to thermal and mechanical loads. Finally, it is remarked that the proposed model is quite robust, such
that the nite element results are not sensitive to the mesh conguration and can rapidly converge to
3-D elasticity solutions using regular or irregular meshes.
& 2011 Elsevier B.V. All rights reserved.

Keywords:
Higher order theory
Finite element
Transverse normal strain
Angle-ply
Laminated composite and sandwich

1. Introduction
Due to low weight, high strength and rigidity, laminated
composite structures are being widely used in many engineering
elds such as aerospace, automotive and submarines. For safe and
reliable designs, it is necessary to well understand the structural
behavior of laminated composite and sandwich plates. Appropriate computational models ought to be developed for accurately
predicting the responses of these laminated structures. Among
the available approaches, the classical laminated plate theory and
the rst order shear deformation theory have been proposed
to predict thermo-mechanical behavior of laminated plates.
The laminated composites generally possess relatively soft transverse shear modulus. However, the classical laminated plate
theory and the rst order shear deformation theory are unable
to adequately model the relatively large transverse shear deformations, so that they often produce unacceptable errors in predicting
displacement and stresses of thick and moderately thick laminated
composite plates. Although the three-dimensional models are

Corresponding author.
E-mail address: [email protected] (S.H. Lo).

0168-874X/$ - see front matter & 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.nel.2011.08.003

accurate enough, the three-dimensional analysis requires huge

computational efforts, as the number of unknowns, in general,
depends on the number of layers of the laminate [1].
By adding high-order terms of transverse coordinate z to the
in-plane and transverse displacements, the global higher order
shear deformation theories [26] have been widely developed.
Compared to the rst-order theory, the global higher-order theory
does not require the shear correction factors and can account for
the warping of the cross-section. In the global higher-order
theory, however, it is assumed that in-plane displacements and
its derivatives are continuous through the thickness of laminates.
Owing to this assumption, the continuity conditions of interlaminar stresses at interfaces will be violated as continuous strains
at the interfaces are multiplied by varying material properties of
different layers. Previous studies [711] indicated that the global
higher-order theories overestimate the natural frequencies and
buckling loads of the soft-core sandwiches with a vast difference
in material properties and thickness between the faces and the
core, as these models violate the continuity conditions of transverse shear stresses at interfaces. To determine the natural
frequencies and the critical loads of the soft-core sandwiches,
the mixed layerwise theories [710] have been developed.
Although in general the performance of the mixed layerwise

W. Zhen et al. / Finite Elements in Analysis and Design 48 (2012) 13461357

theories is promising, they are computational expensive and

become intractable when the number of laminates increases, as
the number of unknowns depends on the number of layers.
The zig-zag theories, which can a priori satisfy the continuity
of transverse shear stresses at interfaces, generally provide a
reasonable compromise between accuracy and efciency. Carrera
[12] introduced the original work and development of the zig-zag
theory in great details. In recent decade, the zig-zag theories have
been developed to study the mechanical, thermal and electrical
behaviors of smart composite plates/shells [1318]. However, due
to enforcement of the transverse shear stress continuity at the
interfaces, derivatives of transverse displacement are involved in
the displacement eld. Hence, C1 continuity of transverse
displacement at element interfaces will be required for the nite
element implementation. Although the continuity conditions of
transverse shear stresses at interfaces can be a priori satised, the
zig-zag theories are unable to produce accurate transverse shear
stresses directly from constitutive equations. For zig-zag theories,
accurate transverse shear stresses could only be obtained by
integrating three-dimensional equilibrium equation. However,
third order derivatives of transverse displacement are involved
if transverse shear stresses are to be computed by integrating the
3-D equilibrium equations. Hence, global equilibrium equations
related to the full domain were proposed [19], which is rather
complicated and expensive, and even might fail under irregular
mesh conguration. An eight-node element based on the zig-zag
theory has been proposed by Icardi [20] in which different
degrees of freedom are dened at the nodes of a nite element.
Nonetheless, the proposed element consists of rst-order and
second-order derivatives of transverse displacement as nodal
variables, which is rather difcult to handle in a practical analysis.
By virtue of the zig-zag theory, Chakrabarti and Sheikh [21]
developed a six-node nonconforming triangular element for the
analysis of sandwich plates, which violates the normal slope
continuity requirement. Recent study [22] indicated that the
six-node nonconforming triangular element [21] was unable to
produce the accurate dynamic response of the sandwich plates, as
it violates the continuity conditions of the normal slope at the
interelement boundary. Pandit et al. [23] proposed an improved
zig-zag theory in which transverse displacement is assumed to be
quadratic over the core thickness and constant over the face
sheets. Nevertheless, the derivatives of transverse displacement
are involved in the displacement elds. To avoid the use of C1
interpolation functions in the nite element implementation,
articial constraints have to be imposed using a penalty approach.
Subsequently, the improved zig-zag theory was extended to study
the static behaviors of sandwich plate with random material
properties [24].
To avoid using the C1 interpolation function in the nite
element formulation, a C0-type globallocal higher-order theory
[25] was proposed for the static analysis of laminated composites.
Compared to the previous globallocal higher-order theories
[26,27], C0 interpolation functions are only required for the nite
element implementation, as derivatives of transverse displacement are eliminated from displacement eld based on known
shear stress conditions. Numerical results indicated that the
C0-type globallocal higher-order theory neglecting transverse
normal strain [25] can be employed to study the deformations
and stresses of laminated composites subjected to mechanical
loads. However, for thick multilayer composite plates under
thermal loads, transverse normal strain plays an important role.
In view of this situation, this paper aims to develop a C0-type
globallocal theory including transverse normal strain for the
analysis of thick multilayer plates under thermal/mechanical
loads. Furthermore, the proposed model is applicable not only
to cross-ply but also to angle-ply laminated composite plates.

1347

Based on the proposed model, the six-node C0 triangular element

is formulated for the thermal/mechanical analysis of thick multilayer plates with general lamination congurations.

2. Higher-order globallocal theory

In order to model the zig-zag shape distribution of in-plane
displacement through the thickness of laminates, a higher-order
in-plane displacement eld is dened by superimposing thirdorder local displacements to the global cubic displacement elds.
To include the transverse normal strain effect, the transverse
displacement is assumed to be linear across the thickness for
thermo-mechanical problems of multilayered plates. The initial
displacement elds are given by
k
uk x,y,z uG x,y,z u kL x,y,z u^ L x,y,z,
k
vk x,y,z vG x,y,z v kL x,y,z v^ L x,y,z,

wk x,y,z wG x,y,z:

The global displacement components in Eq. (1) are given by

3
X

uG x,y,z u0 x,y

zi ui x,y,

i1
3
X

vG x,y,z v0 x,y

zi vi x,y,

i1

wG x,y,z w0 x,y zw1 x,y:

The local displacement components of the kth ply can be

written as follows:
2

u kL x,y,z zk uk1 x,y zk uk2 x,y,

2

v kL x,y,z zk vk1 x,y zk vk2 x,y,

k
3
u^ L x,y,z zk uk3 x,y,
k
3
v^ L x,y,z zk vk3 x,y,

in which, zk akz bk, ak 2=zk 1 zk , bk zk 1 zk =zk 1 

zk , and x, y and z are the global coordinates of the plate. The
reference plane (z 0) is taken at the mid-surface of the laminate.
The local coordinates for a layer are denoted by x, y and zk, where
zkA[ 1,1]. The relationships between the global coordinates and
the local coordinates are depicted in Fig. 1.
In Eq. (1), there are 6n 10 variables in the displacement elds,
where n denotes the number of layers of the composite plate. In
order to reduce the layer-dependent variables, the interlaminar
continuity compatibility conditions of in-plane displacements
[26] and transverse shear stresses will be employed. The interlaminar continuity conditions of in-plane displacements can be

Fig. 1. Schematic diagram for the laminated plate and coordinates.

1348

W. Zhen et al. / Finite Elements in Analysis and Design 48 (2012) 13461357

at interfaces,

expressed as

k1
txz
zk tkxz zk ,

u kL x,y,zk u k1
x,y,zk
L
k
k1
u^ L x,y,zk u^ L x,y,zk

v kL x,y,zk
k
v^ L x,y,zk

v k1
x,y,zk
L
^v k1
x,y,zk
L

3k 2
u1
F1
6 7
6
6 u2 7
6 G1
6 7
6
6 u3 7
6 H1
6 7
6
6 7 6
6 v1 7
6 L1
6 7
6
6 v2 7
6 M1
4 5
4
v3
N1

where k 2,3,4,. . .,n:

k1
tyz
zk tkyz zk :

Applying the interlaminar continuity condition and free condition at the bottom surface of transverse shear stresses, local
displacement variables for the kth ply can be expressed as

F2

F3

F4

F5

F6

F7

F8

F9

F10

F11

F12

F13

G2

G3

G4

G5

G6

G7

G8

G9

G10

G11

G12

G13

H2

H3

H4

H5

H6

H7

H8

H9

H10

H11

H12

H13

L2

L3

L4

L5

L6

L7

L8

L9

L10

L11

L12

L13

M2

M3

M4

M5

M6

M7

M8

M9

M10

M11

M12

M13

N2

N3

N4

N5

N6

N7

N8

N9

N10

N11

N12

N13

8 19
u >
>
>
>
> 11 >
>
>
>
u2 >
>
>
>
>
>
>
>
>
1
>
>
u
>
3 >
>
>
>
>
>
>
>
>
u
>
>
1
>
>
>
>
3k >
>
>
>
>
>
u
F14
2
>
>
>
>
>
7 >
>
>
u3 >
G14 7 >
>
>
>
>
>
7 >
@w
>
1
=
< @x >
H14 7
7
,
7
1
v
L14 7 >
>
> 1>
>
7 >
>
>
> v1 >
M14 7
>
> 2>
5 >
>
>
> 1>
>
>
N14 >
>
> v3 >
>
>
>
>
>
>
>
v
>
> 1>
>
>
>
>
>
>
>
>
> v2 >
>
>
>
>
>
>
>
v3 >
>
>
>
>
>
>
>
@w
;
: 1>

@y

Using the interlaminar continuity conditions of in-plane displacement, 4(n 1) variables are eliminated from the initial
displacement eld. The interlaminar continuity conditions and
free surface conditions of transverse shear stresses can be used
to further reduce the layer-dependent variables. For angle-ply
composite plate, the transverse shear stresses for the kth ply are
given by

tkxz z Q44k ekxz z Q45k ekyz z,

tkyz z Q45k ekxz z Q55k ekyz z,

where Qijk are the transformed material constants with respect to

the global coordinates for the kth layer and transverse shear
strains are given by

where coefcients Fik , Gki , Hik , Lki , Mik and Nik (i1  14) are given in
Appendix.
Employing the stress conditions of zero transverse shear
stresses at the top surface, @w1 =@x and @w1 =@y can be, respectively, expressed as
@w1
r1 u11 r2 u12 r3 u13 s1 u1 s2 u2 s3 u3 c1 v11 c2 v12 c3 v13
@x
d1 v1 d2 v2 d3 v3 ,
@w1
e1 u11 e2 u12 e3 u13 f1 u1 f2 u2 f3 u3 g1 v11 g2 v12
@y
g3 v13 h1 v1 h2 v2 h3 v3 ,

where the coefcients in Eq. (10) can be found in the Appendix A.

Substituting Eqs. (9) and (10) into Eq. (1), the nal displacement
elds can be written as

ekxz z

@w0
@w1
2
z
u1 2zu2 3z2 u3 ak uk1 2ak zk uk2 3ak zk uk3 ,
@x
@x

ekyz z

@w0
@w1
2
z
v1 2zv2 3z2 v3 ak vk1 2ak zk vk2 3ak zk vk3 ,
@y
@y
6

uk u0

In Eq. (6), it can be found that derivatives of transverse

displacements are present in the strain components. The derivatives of transverse displacements will be involved in the in-plane
displacement elds, as the interlaminar continuity conditions
of transverse shear stresses are used. As the derivatives of
transverse displacements are required in the in-plane displacement elds, C1 interpolation functions would be required for the
nite element implementation. To avoid using C1 interpolation
functions, the derivatives of lateral displacement have to be
eliminated by employing the stress free conditions at the boundary surfaces.
Using the conditions of zero transverse shear stresses at the
lower surface, the following equations can be obtained:

vk v0

@w0
@w1
H1 u11 H2 u12 H3 u13 H4 u1 H5 u2 H6 u3 H7
,
@x
@x
@w0
@w1
N1 v11 N2 v12 N3 v13 N4 v1 N5 v2 N6 v3 N7
,
@y
@y

3. Finite element formulation

7

where coefcients Hi and Ni (i1  7) can be found in

Appendix A. Recall the transverse shear stress continuity conditions

10

3
X

Fki zu1i

i1
3
X

3
X

Fkj 3 zuj

j1

Cki zu1i

i1

3
X

Ckj 3 zuj

j1

wk w0 zw1 ,

Fki

3
X

Fkr 6 zv1r

r1
3
X
r1

Ckr 6 zv1r

3
X

Fks 9 zvs ,

s1
3
X

Cks 9 zvs ,

s1

11

Cki

where
and
are functions of material constants and the
thickness of the laminated plate, respectively, which are also
shown in Appendix A. Based on known traction conditions at the
upper surface, the derivatives of transverse displacement @wi =@x
and @wi =@y (i0, 1) are eliminated. This is a major breakthrough
in the nite element formulation of composite plates, in which
simple planar C0 nite elements could be used for the interpolation of displacement parameters.

As the rst derivatives of transverse displacement are excluded

from the in-plane displacement elds, the classical six-node C0
quadratic triangular element can be used in the nite element
analysis. The interpolations for the displacement parameters used

W. Zhen et al. / Finite Elements in Analysis and Design 48 (2012) 13461357

in the generalized displacement eld are given by

6
X

u0

Nr u0r ,

6
X

v0

r1

Nr u11r ,

6
X

Nr u12r ,

v12

6
X

r1

w1

6
X

Nr w1r ,

r1

6
X

uj

r1

Nr v11r ,

Nr w0r ,

r1

6
X

u12

r1

v11

w0

r1

6
X

u11

Nr v0r ,

6
X

tyz z 

Nr ujr ,

j 123,

Nr v12r ,

6
X

vj

Nr vjr ,

j 123,

12

r1

3.1. The strain and the stiffness matrix

By virtue of linear straindisplacement relationships, the
strain for the kth layer can be written as


ek @uk B1 B2 B3 B4 B5 B6 de ,
13
e

where d d1 d2 d3 d4 d5 d6 T and di (i 16) are displacement parameters at the ith node. The sixteen global displacement parameters at each node are given by
h

dei u0i v0i

w0i

2h

@txy @sy

dz,
@x
@y

15

where sx, sy and txy are in-plane stresses of the element.

4. Numerical examples

where Ni (2Li 1)Li, N4 4L1L2, N5 4L2L3, N6 4L3L1; Li are area

coordinates, (i1,2,3).

r1

r1

1349

u11i

w1i

u12i

u13i

u1i

u2i

u3i

v11i

v12i

v13i

In this section, the nite element results of laminated composites based on the proposed model are presented. The performance of nite element is assessed by comparing with the threedimensional elasticity solutions and other published results. The
effects of transverse normal strain and stacking sequence are also
studied. The nite element meshes for the static analysis of the
composite plates are shown in Fig. 2. The material constants used
in the examples are given as follows:
Material (1) for laminated composite plates [28]
E1 172:5 GPa,

E2 E3 6:9GPa,

G23 1:38 GPa,

v12 v13 v23 0:25:

G12 G13 3:45 GPa,

Material (2) for laminated composite plates [29]

C11 1:0025E2 ,
C44 0:5E2 ,
v1i

v2i

v3i

C12 0:25E2 ,

C55 0:2E2 ,
iT

C22 25:0625E2 ,

C33 C11 ,

C66 C44 :

Material (3) for laminated composite plates [29]

2

@
@x

6
0
@6
4
0
2

@
@z

@
@y

@
@z

@
@x

@
@z
@
@y

@Ni
@x

6
6 0
6
6
6 0
6
6
6
6 0
6
6 k @N
i
6F
6 1 @x
Bi 6
6 ^
6
6 k @Ni
6 F6
@x
6
6
6 k @Ni
F
6 7 @x
6
6 ^
6
4

@ 3T
@y
@ 7
@x 7
5 ,

C11 32:0625E2 ,
C44 0:2E2 ,

@Ni
@y

@Ni
@x

@Ni
@x

@Ni
@y

Ni

i
z @N
@x

i
z @N
@y

i
Ck1 @N
@y

@Fk1
@z

Ni

^
@Fk6

i
Ck6 @N
@y

i
Ck7 @N
@y

@Fk12
@z

i
i
Fk12 @N
Ck12 @N
@x
@y

C55 0:8E2 ,

C22 1:00195E2 ,

C33 C22 ,

C66 C55 :

Material (4) for laminated composite plates [29]

@Ni
@y

C12 0:2495E2 ,

@Ck1
@z

Ni

k @Ni
i
Fk1 @N
@y C1 @x

@z

Ni

@Fk7
@z

Ni

@Ck6

@z

Ni

@Ck7
@z

Ni

k @Ni
i
Fk7 @N
@y C7 @x

^
Ni

i
Ck6 @N
@x

i
Fk6 @N
@y

@Ck12
@z

^
Ni

k @Ni
i
Fk12 @N
@y C12 @x

3T
7
7
7
7
7
7
7
7
7
7
7
7
7
7 ,
7
7
7
7
7
7
7
7
7
7
7
5

C11 25:0625E2 ,
C44 0:2E2 ,

C12 0:25E2 ,

C55 0:5E2 ,

C22 1:0025E2 ,

C33 C22 ,

C66 C55 :

Material (5) for sandwich plates [30]

Face sheets (h/5  2):
i 126:

By means of the strain matrix B, the element stiffness matrix

K can be readily evaluated using the following equation:

n Z ZZ
X
Ke
14
BT Q k B dx dy dz,

E1 200 GPa,

E2 E3 8 GPa,

G23 2:2 GPa,v12 v13 0:25,

G12 G13 5 GPa,

v23 0:35,

a1 2  106 =K,

a2 a3 50  106 =K:
Core material (3h/5):
Ec1 Ec2 1 GPa,

Ec3 2 GPa,

Gc12 3:7 GPa,

Gc13 Gc23 0:8 GPa,

k1

hk

where Qk is the transformed material constant matrix of the

kth ply.
3.2. Transverse shear stresses from 3-D equilibrium equation
Apart from using constitutive equations, more accurate transverse shear stresses could be obtained by integrating threedimensional local stress equilibrium equations through the thickness of laminated plates, i.e.
Z z
@sx @txy

dz,
txz z 
@y
2h @x

vc12 0:35,

vc13 vc23 0:25,

ac1 ac2 ac3 30  106 =K:

where 1 and 2 denote the in-plane directions and 3 denotes

transverse direction of laminates.
Example 1. Cylindrical bending of laminated composite plate
strips subjected to a sinusoidal loading q q0 sinpx=a.
The normalized displacements and stresses are given by

sx

u~

sx a=2,b=2,zh2
q0 a2

E2 u0,b=2,z
,
q0 h

t xz

s~ x

txz 0,b=2,zh
q0 a

sx a=2,b=2,z
q0

1350

W. Zhen et al. / Finite Elements in Analysis and Design 48 (2012) 13461357

Fig. 2. Finite element meshes of the entire plate. (a) Regular mesh conguration (Mesh sh 1, m  m), (b) irregular mesh conguration (Mesh 2, 24 elements), (c) irregular
mesh conguration (Mesh 3, 24 elements) and (d) irregular mesh conguration (Mesh 4, 24 elements).

Table 1
Convergence rate of in-plane stress sx a=2,b=2,z for [151/  151] plate (Mesh 1,
a/h 4).
z/h

 0.5
 0.2
0.0 
0.0
0.2
1.0

Present

3-D [31]

44

88

12  12

14  14

 1.0292
 0.0371
0.4719
 0.4926
0.0240
1.0812

 0.9992
 0.0356
0.4582
 0.4771
0.0235
1.0481

 0.9925
 0.0353
0.4552
 0.4738
0.0234
1.0411

 0.9910
 0.0353
0.4546
 0.4732
0.0234
1.0396

 0.5
 0.2
0.0 
0.0
0.2
1.0

Present

 0.9960
 0.0373
0.4526
 0.4719
0.0260
1.0446

3-D [31]

96 elements

384 elements

600 elements

 1.005
 0.0367
0.4595
 0.4785
0.0240
1.0532

 0.9975
 0.0356
0.4570
 0.4755
0.0236
1.0456

 0.9966
 0.0355
0.4566
 0.4751
0.0237
1.0448

a/h

sx

a

b h
2 ,2 ,2

txz 0, 2b,0

Table 2
Convergence rate of in-plane stress sx a=2,b=2,z for [151/  151] plate (Mesh 2,
a/h 4).
z/h

Table 3
Comparison of nondimensional stresses at strategic points of laminated composite
plate [01/901/01].

Present (12  12)

Exact [28]
10
Present (12  12)
Exact [28]
20
Present (12  12)
Exact [28]
50
Present (12  12)
Exact [28]

t~ xy

txy a=2,b=2,z
q0

t~ xz

1.1763
1.1678

0.3531
0.3576

0.7343
0.7369

0.4196
0.4238

0.6540
0.6580

0.4375
0.4374

0.6299
0.6348

0.4854
0.4415

txz 0,b=2,z
q0

The following boundary conditions proposed by Bogdanovich

and Yushanov [31] are used.
 0.9960
 0.0373
0.4526
 0.4719
0.0260
1.0446

u0 v0 w0 w1 0, at x 0;

u0 w0 w1 0, at x a:

Material (1) is used in this example for various models

considered, and the entire plate is modeled in the analysis of
the static response of angle-ply laminated plates. In the Figures,
acronyms have been used. GLHTI represents the results obtained

W. Zhen et al. / Finite Elements in Analysis and Design 48 (2012) 13461357

found that the proposed model is applicable not only to thick

plates but also to thin composite plates. In Fig. 3, convergence rate
of in-plane stress by different meshes for three-layer [301/ 301/301]
plate (a/h4) is also presented. It is found that converged results
close to the 3-D elasticity solution [31] could be obtained from
regular 12  12 meshes and irregular meshes of 384 elements, for
which mesh congurations had little inuence on the accuracy of
results. Effect of transverse normal strain on transverse shear
stresses has been shown in Fig. 4. Numerical results show that
GLHTI including transverse normal strain agree well with the 3-D
elasticity solution [31]. However, GLHTN neglecting transverse
normal strain is less accurate compared to 3-D elasticity solution.
For the ve-layer [ 151/  451/01/301/151] plate, the results are,
respectively, presented in Figs. 58. It is noted that results of
GLHTI are in good agreement with the 3-D elasticity solutions
[32]. However, the results (FSDT) obtained from the rst order
theory seem to be less accurate in comparison with the 3-D
elasticity solutions.

from the proposed six-node triangular element based on the

C0-type Global Local Higher-order Theory, and 16 variables are
dened at each node. GLHTN denotes the results obtained from
the six-node triangular element based on the C0-type Global Local
Higher-order Theory neglecting transverse normal strain [25], and
13 variables are dened at each node. FSDT represents the results
obtained from the six-node triangular element based on the First
order Shear Deformation Theory, and there are 5 variables in the
displacement eld of this model. A sufx has been introduced to
the acronyms:  C and E denote, respectively, transverse shear
stresses obtained directly from constitutive equations and by
integrating 3-D equilibrium equation; M1 and M2 denote,
respectively, results obtained from Mesh 1 and Mesh 2. It is
noted that results of GLHTN and FSDT are produced by the
authors based on their own computer codes.
The convergence characteristics of the GLHTI for an angle-ply
moderately thick [151/ 15] plate (a/h 4) are shown in
Tables 1 and 2. It is found that stresses can rapidly converge to
within 1% of the 3-D elasticity solutions [31] by rening regular or
irregular meshes. In Table 3, comparison of nondimensional
stresses at strategic points of laminated composite plate [01/901/01]
with different length-to-thickness ratios has been presented. It is

Example 2. Simply-supported laminated composite plate subjected to a doubly sinusoidal transverse loading q q0 sinpx=a
sinpy=b.

0.5

0.5

3-D [31]
GLHTI (4x4)
GLHTI (8x8)
GLHTI (12x12)

0.4
0.3

0.3
0.2

0.1

0.1

z
h

0
-0.1

0
-0.1

-0.2

-0.2

-0.3

-0.3

-0.4

-0.4

-0.5
-1

-0.5

0.5

-0.5
-1

x (a / 2, b / 2, z)

0.5
3-D [31]
GLHTI (24 elements)
GLHTI (96 elements)
GLHTI (384 elements)
GLHTI (600 elements)

0.4
0.3
0.2

-0.5

0.5

x (a / 2, b / 2, z)

0.5

0.4
0.3
0.2

3-D [31]
GLHTI (24 elements)
GLHTI (96 elements)
GLHTI (384 elements)
GLHTI (600 elements)

0.1

0.1

z
h

3-D [31]
GLHTI (96 elements)
GLHTI (384 elements)
GLHTI (600 elements)

0.4

0.2

z
h

1351

z
h

-0.1

-0.1

-0.2

-0.2

-0.3

-0.3

-0.4

-0.4

-0.5
-1

-0.5

x (a / 2, b / 2, z)

0.5

-0.5
-1

-0.5

0.5

x (a / 2, b / 2, z)

Fig. 3. In-plane stress by different meshes for three-layer [301/  301/301] plate (a/h 4). (a) Mesh 1, (b) mesh 2, (c) mesh 3 and (d) mesh 4.

1352

W. Zhen et al. / Finite Elements in Analysis and Design 48 (2012) 13461357

0.5

0.5
0.4

3-D [31]
2-D [31]
GLHTI-C (12x12)
GLHTN-C (12x12)

0.3
0.2

0.3
0.2
0.1

0.1

z
h

z
h

0
-0.1

-0.1

-0.2

-0.2

-0.3

-0.3

-0.4

-0.4
-0.5
-0.14

3-D [32]
GLHTI-M1 (12x12)
GLHTI-M2 (384 elements)
FSDT-M1 (12x12)

0.4

-0.12

-0.1

-0.08

-0.06

-0.04

-0.02

-0.5
-20

0.02

-15

-10

-5

10

15

20

~x (a / 2, b / 2, z)

yz(0, b / 2, z)

Fig. 6. In-plane stress for ve-layer [  151/  451/01/301/151] plate (a/h 4).

0.5
0.4

3-D [31]
2-D [31]
GLHTI-C (384 elements)
GLHTI-E (384 elements)

0.3
0.2

0.5
0.4
0.3

0.1

z
h

0.1

z
h

-0.1
-0.2

0
-0.1

-0.3

-0.2

-0.4
-0.5
-0.14

3-D [32]
GLHTI-M1 (12x12)
GLHTI-M2 (384 elements)
FSDT-M1 (12x12)

0.2

-0.3
-0.12

-0.1

-0.08

-0.06

-0.04

-0.02

0.02

yz(0, b / 2, z)

-0.4
-0.5
-6

-4

-2

~xy (a / 2, b / 2, z)

Fig. 4. Transverse shear stress for three-layer [301/  301/301] plate (a/h 4). (a)
Mesh 1 and (b) mesh 2.

Fig. 7. In-plane stress for ve-layer [  151/  451/01/301/151] plate (a/h 4).

0.5
0.5

3-D [32]
GLHTI-M1 (12x12)
GLHTI-M2 (384 elements)
FSDT-M1 (12x12)

0.4
0.3

0.4

0.2

0.2

0.1

z
h

0.1

z
h

-0.1

0
-0.1

-0.2

-0.2

-0.3

-0.3

-0.4

-0.4

-0.5
-1.5

3-D [32]
GLHTI-M1 (12x12)
GLHTI-M2 (384 elements)

0.3

-1

-0.5

~(0, b / 2, z)
u

0.5

1.5

Fig. 5. In-plane displacement for ve-layer [ 151/ 451/01/301/151] plate (a/h4).

-0.5

0.5

1.5

~xz (0, b / 2, z)
Fig. 8. Transverse shear ve-layer [  151/  451/01/301/151] plate (a/h 4).

W. Zhen et al. / Finite Elements in Analysis and Design 48 (2012) 13461357

1353

Table 4
Boundary conditions for examples 2 and 3.
Boundary
conditions

x constant

y constant

Simplysupported
boundary
Symmetric
axis

w0 w1 v11 v12 v13 v1 v2 v3 0 w0 w1 u11 u12 u13 u1 u2 u3 0

u0 u11 u12 u13 u1 u2 u3 0

0.5

v0 v11 v12 v13 v1 v2 v3 0

0.5

3-D [29]
GLHTI (12x12)
ZZT [29]

0.4
0.3

0.3
0.2

0.2
0.1

z
h

3-D [29]
GLHTI (12x12)
ZZT [29]

0.4

0.1

z
h

-0.1

-0.1

-0.2

-0.2

-0.3

-0.3

-0.4

-0.4

-0.5
-0.015

-0.01

-0.005

0.005

-0.5
-0.04

0.01

-0.02

Fig. 9. In-plane displacement through thickness of ve-layer plate with different

thickness at each ply (a/h 4).

0.4
0.3

3-D [29]
GLHTI (12x12)
ZZT [29]

0.2
0.1

0.1

z
h

-0.1

-0.1

-0.2

-0.2

-0.3

-0.3

-0.4

-0.4

-0.5
-0.6

0.06

0.5
3-D [29]
GLHTI (12x12)
ZZT [29]

0.2

z
h

0.04

Fig. 11. In-plane shear stresses through thickness of ve-layer plate with different
thickness at each ply (a/h 4).

0.5

0.3

0.02

xy (a / 2, b / 2, z)

u (a / 2,0, z)

0.4

-0.4

-0.2

0.2

0.4

0.6

0.8

y (0,0, z)
Fig. 10. In-plane stresses through thickness of ve-layer plate with different
thickness at each ply (a/h 4).

The normalized displacements and stresses are given by

u E2 h2 u=q0 a3 ,

s y , t xy sy , txy h2 =q0 a2 ,

t xz txz h=q0 a:

Owing to symmetry in geometry, material properties and

loading conditions, one-quarter of the laminated plate is considered in this example. The boundary conditions used are given in
Table 4.
To further assess the performance of the nite element based
on the proposed model, a ve-layer plate with different thickness

-0.5
-0.05

0.05

0.1

0.15

0.2

0.25

0.3

0.35

xz(a / 2,0, z)
Fig. 12. Transverse shear stresses through thickness of ve-layer plate with
different thickness at each ply (a/h 4).

and material properties at each ply is studied. The plies of the

composite plate are of thickness 0.3h/0.2h/0.15h/0.25h/0.1h and
of materials 4/2/4/3/2. Distributions of displacement and stresses
through the thickness are plotted in Figs. 912. Due to the rapid
changes of material properties through the thickness direction,
the postprocessing method from the zig-zag theory proposed by
Cho and Choi [29] did not produce accurate results of in-plane
displacement and transverse shear stresses. However, the present
nite element based on the proposed model could produce more
accurate displacement and stress distributions.

1354

W. Zhen et al. / Finite Elements in Analysis and Design 48 (2012) 13461357

0.5
0.4
0.3

z
h

0.5
3-D [30]
GLHTI (12x12)
GLHTN (12x12)

0.4
0.3

0.2

0.2

0.1

0.1

z
h

-0.1

-0.1

-0.2

-0.2

-0.3

-0.3

-0.4

-0.4

-0.5
-30

-20

-10

10

20

30

3-D [30]
GLHTI (12x12)
GLHTN (12x12)

-0.5
-1.5

-1

-0.5

Fig. 13. Transverse displacement for sandwich plate under thermal loads (a/h 4).

0.4

0.1
0
-0.1
-0.2
-0.3
-0.4
-50

-40

-30

-20

-10

10

20

x (0,0, z)
Fig. 14. In-plane stress for sandwich plate under thermal loads (a/h 4).

0.5
0.4
0.3

z
h

-0.1
-0.2
-0.3
-0.4
-16

-14

a0 T0

s x , t xy , t xz ,

sx , txy , txz
,
a0 T0 ET

a0 106 =K:

Taking into account the symmetry in geometry, material

properties and loading conditions, one-quarter of the sandwich
plate is considered in this example. The boundary conditions used
are given in Table 4.
In order to study the effects of thermal stress and displacement of sandwich plates due to transverse normal strain,
three-layer sandwich plate with material (5) is considered. Distributions of transverse displacement and stresses through the
thickness are shown in Figs. 1316. It is found that the proposed
model GLHTI can produce much better lateral displacement
distribution through the thickness. However, lateral displacement
obtained from GLHTN is zero, which is in contradiction with the
exact solution [30]. Due to neglect of transverse normal strain, the
model GLHTN fails to produce accurate in-plane stresses.
The distributions of transverse shear stresses along the thickness
are shown in Fig. 16. Numerical results show that the transverse
shear stresses computed by the proposed model (GLHTI) are in
excellent agreement with the exact solutions. However, transverse shear stress obtained from the model GLHTN is less reliable.

5. Conclusions

3-D [30]
GLHTI (12x12)
GLHTN (12x12)

-0.5
-18

Fig. 16. Transverse shear stress for sandwich plate under thermal loads (a/h 4).

3-D [30]
GLHTI (12x12)
GLHTN (12x12)

0.2

0.1

1.5

The normalized displacements and stresses are given by

0.3

0.2

Example 3. A square sandwich plate (01/core/01) subjected to

thermal load DTx,y,z T0 sinpx=asinpy=b has been analyzed.

0.5

-0.5
-60

0.5

xz (a / 2,0, z)

w (0,0, z)

z
h

-12

-10

-8

xy (a / 2, b / 2, z)
Fig. 15. In-plane stress for sandwich plate under thermal loads (a/h 4).

-6

By introducing transverse normal strain, a C0-type global-local

higher order theory is developed to enhance the analysis of thermal/
mechanical response of thick multilayer plates with general lamination congurations. The proposed model is applicable not only to
cross-ply but also to angle-ply laminated composite plates.
By employing transverse shear free conditions at upper and lower
surfaces and interlaminar continuity conditions of transverse shear
stresses, the layer-dependent displacement variables can all be
eliminated. As a result, the number of variables involved in the
proposed model is independent of the number of layers of laminates.
One major advantage of the proposed model is that C0 interpolation
functions are only required for the nite element implementation, as
the derivatives of transverse displacement have been eliminated from
the general displacement eld. The six-node quadratic C0 triangular
element can be conveniently applied to laminated composite and

W. Zhen et al. / Finite Elements in Analysis and Design 48 (2012) 13461357

sandwich plates of various material characteristics under different

load conditions. It is remarked that the nite element results are not
affected by the mesh conguration, which could rapidly converge to
the 3-D elasticity solution using regular or irregular meshes. Furthermore, effects of transverse normal strain on thermal/mechanical
behaviors of thick multilayer plates have been studied. It is found
that transverse normal strain has a crucial effect on the lateral
displacement and in-plane stresses of thick multilayer composite
and sandwich plates under thermal loads.

S 11 N4 1Zk ,
S 13 N6 3z2k Zk ,

wk
Wk

gk

The work described in this paper was supported by the

University Development Fund (2009) on Computational Science
and Engineering at the University of Hong Kong, the National
Natural Sciences Foundation of China (Nos. 10802052, 11072156),
the Program for Liaoning Excellent Talents in University
(LR201033), and the Program for Science and Technology of
Shenyang (F10-205-1-16).

rk

H1 a1 ,

H2 2a1 ,

H6 3z21 ,

H3 3a1 ,

H4 1,

N4 1,

N5 2z1 ,

For k1, the coefcients are written as

G13

G14

1
F21 F31 F41 . . .F14
0,

H31 1,
L18 1,

Zk

. . .G114

G12 1,

ck

H5 2z1 ,

H7 z1 ,

N1 a1 , N2 2a1 , N3 3a1 ,

N6 3z21 , N7 z1 :

F11 1,

Bk

yk

The coefcients Hi and Ni (i17) are given by

G11

0,

ak1
;
ak
!
Q
Q Q44k1 Q55k 1
1 45k1 45k
;
2
ak
Q44k Q55k Q45k
!
Q55k1 Q45k Q45k1 Q55k ak1
;
2
ak
Q44k Q55k Q45k
!
Q55k1 Q45k Q45k1 Q55k 1
;
2
ak
Q44k Q55k Q45k
!
Q55k1 Q44k Q45k1 Q45k ak1
;
2
ak
Q44k Q55k Q45k
!
Q
Q Q55k1 Q44k 1
1 45k1 45k
;
2
ak
Q44k Q55k Q45k
!
Q44k1 Q45k Q45k1 Q44k ak1
;
2
ak
Q44k Q55k Q45k
!
Q44k1 Q45k Q45k1 Q44k 1
:
2
ak
Q44k Q55k Q45k

Employing the stress free conditions of transverse shear

stresses at the upper surface, @w1 =@x and @w1 =@y can be respectively expressed as
@w1
r1 u11 r2 u12 r3 u13 s1 u1 s2 u2 s3 u3 c1 v11
@x
c2 v12 c3 v13 d1 v1 d2 v2 d3 v3 ,
@w1
e1 u11 e2 u12 e3 u13 f1 u1 f2 u2 f3 u3 g1 v11
@y

where

1
M 19 1,M11 M21 M31 . . .M14
0

r1

N110 1,

r2

1
N11 N21 N31 . . .N14
0:

The coefcients for k41 can be determined from the following

recursive equations:

r3

Fik 2 wk Fik1 21 wk Gk1

31 wk Hik1
i

s1

gk Lk1
2Mik1 3Nik1 Si ;
i

s2

Lki 2 Bk Lk1
21 Bk Mik1 31 Bk Nik1
i
yk Fik1 2Gk1
3Hik1 S i ;
i

Mik1 ,
Mik Lki Lk1
i

Nik Nik1

S 3 H3 W k ,

S4 H4 1Wk ,

S6 H6 3z2k Wk , S7 H7 zk Wk , S8 N1 rk ,
S9 N2 rk , S10 N3 rk ,
S11 N4 1rk , S12 N5 2zk rk ,
S13 N6 3z2k rk ,
S 1 H1 ck ,

S 2 H2 ck ,

S 4 H4 1ck ,
S 6 H6 3z2k ck ,
S 8 N1 Zk ,

S14 N7 zk rk ;
S 3 H3 ck ,

S 5 H5 2zk ck
S 7 H7 zk ck ,

S 9 N2 Zk ,

S 10 N3 Zk ,

s3
c1

where
S1 H1 Wk , S2 H2 Wk ,
S5 H5 2zk Wk ,

g2 v12 g3 v13 h1 v1 h2 v2 h3 v3 ,

L11 L12 L13 . . .L114 0,

Hik Hik1 ;

S 14 N7 zk Zk :

Q44k1 Q55k Q45k1 Q45k

2
Q44k Q55k Q45k

1
H11 H21 H41 . . .H14
0;

Gki Fik Fik1 Gk1

,
i

S 12 N5 2zk Zk ,

k 2,3,:::n:

Acknowledgement

Appendix

1355

b1 a14 a1 H1 b14 N7 zn 1

,
D
b2 a14 a2 H2 b14 N7 zn 1
,
D
b3 a14 a3 H3 b14 N7 zn 1
,
D
b4 a14 a4 H4 1b14 N7 zn 1
,
D
b5 a14 a5 H5 2zn 1 b14 N7 zn 1
,
D
2
b6 a14 a6 H6 3zn 1 b14 N7 zn 1
,
D
b8 N1 a14 a8 b14 N7 zn 1

D
b N2 a14 a9 b14 N7 zn 1
c2 9
,
D
b N3 a14 a10 b14 N7 zn 1
,
c3 10
D
b N4 1a14 a11 b14 N7 zn 1
,
d1 11
D
b12 N5 2zn 1 a14 a12 b14 N7 zn 1
d2
,
D
2
b N6 3zn 1 a14 a13 b14 N7 zn 1
d3 13
D
a1 H1 b7 b1 a7 H7 zn 1
e1
,
D
a2 H2 b7 b2 a7 H7 zn 1
,
e2
D

1356

e3
f1

W. Zhen et al. / Finite Elements in Analysis and Design 48 (2012) 13461357

a3 H3 b7 b3 a7 H7 zn 1

Ok1 Lk1 Lk7 r1 Lk14 e1 ,

a4 H4 1b7 b4 a7 H7 zn 1

D
a5 H5 2zn 1 b7 b5 a7 H7 zn 1
,
f2
D
2
a6 H6 3zn 1 b7 b6 a7 H7 zn 1
,
f3
D
a8 b7 b8 N1 a7 H7 zn 1
,
g1
D
a9 b7 b9 N2 a7 H7 zn 1
,
g2
D
a10 b7 b10 N3 a7 H7 zn 1
g3
,
D
a11 b7 b11 N4 1a7 H7 zn 1
h1
,
D
a12 b7 b12 N5 2zn 1 a7 H7 zn 1
,
h2
D
2
a13 b7 b13 N6 3zn 1 a7 H7 zn 1
h3
:
D
n
n
n
bi an Lni 2an Min 3an Nin
i an Fi 2an Gi 3an Hi ;
n
n
n
n
D an F7 2an G7 3an H7 H7 zn 1 an Ln14 2an M14
n
n
n
n
3an N14 N7 zn 1 an F14 2an G14 3an H14
an Ln7 2an M7n 3an N7n

Finally, coefcients Fki and Cki are given by

Fki Rki zk Ski z2k Tik z3k Zi ,

Cki Oki zk Pik z2k Qik z3k Z i ,

k
P1k M1k M7k r1 M14
e1 ,
k
k
k
k
Q1 N1 N7 r1 N14 e1 ,
k
Ok2 Lk2 Lk7 r2 Lk14 e2 , P2k M2k M7k r2 M14
e2 ,
k
k
k
k
Q2 N2 N7 r2 N14 e2 ,
k
Ok3 Lk3 Lk7 r3 Lk14 e3 , P3k M3k M7k r3 M14
e3 ,
k
k
k
k
Q3 N3 N7 r3 N14 e3 ,
k
Ok4 Lk4 Lk7 s1 Lk14 f1 , P4k M4k M7k s1 M14
f1 ,
k
Q4k N4k N7k s1 N14
f 1,

Ok5 Lk5 Lk7 s2 Lk14 f2 ,

k
P5k M5k M7k s2 M14
f2 ,

k
Q5k N5k N7k s2 N14
f2 ,

Ok6 Lk6 Lk7 s3 Lk14 f3 ,

k
P6k M6k M7k s3 M14
f3 ,

k
Q6k N6k N7k s3 N14
f3 ,
k
k
k
k
O7 L8 L7 c1 L14 g1 ,
k
Q7k N8k N7k c1 N14
g1

k
P7k M8k M7k c1 M14
g1 ,

Ok8 Lk9 Lk7 c2 Lk14 g2 ,

k
P8k M9k M7k c2 M14
g2 ,

k
Q8k N9k N7k c2 N14
g2 ,
k
k
k
k
O9 L10 L7 c3 L14 g3 ,
k
k
Q9k N10
N7k c3 N14
g3 ,
k
k
k
k
O10 L11 L7 d1 L14 h1 ,
k
Q10
Ok11
k
Q11
k
O12
k
Q12

k
k
P9k M10
M7k c3 M14
g3 ,

k
k
k
P10
M11
M7k d1 M14
h1 ,

k
k
N11
N7k d1 N14
h1 ,
Lk12 Lk7 d2 Lk14 h2 ,

k
k
k
P11
M12
M7k d2 M14
h2 ,

k
k
N12
N7k d2 N14
h2 ,
k
k
k
L13 L7 d3 L14 h3 ,

k
k
k
P12
M13
M7k d3 M14
h3 ,

k
k
N13
N7k d3 N14
h3 :

where
Z4 z,
Z 10 z,

Z5 z2 ,

Z6 z3 ,
2

Z 11 z ,

Zi 0

Z 12 z3 ,

k
e1 ,
Rk1 F1k F7k r1 F14

i a4,5,6,

Zi 0

i a10,11,12,

Sk1 Gk1 Gk7 r1 Gk14 e1 ,

k
e1 ,
T1k H1k H7k r1 H14
k
e2 ,
Rk2 F2k F7k r2 F14

Sk2 Gk2 Gk7 r2 Gk14 e2 ,

k
k
k
k
T2 H2 H7 r2 H14 e2 ,
k
Rk3 F3k F7k r3 F14
e3 , Sk3 Gk3 Gk7 r3 Gk14 e3 ,
k
k
k
k
T3 H3 H7 r3 H14 e3 ,
k
f1 , Sk4 Gk4 Gk7 s1 Gk14 f1 ,
Rk4 F4k F7k s1 F14
k
f 1,
T4k H4k H7k s1 H14
k
k
k
k
R5 F5 F7 s2 F14 f2 , Sk5 Gk5 Gk7 s2 Gk14 f2 ,
k
f2 ,
T5k H5k H7k s2 H14
k
Rk6 F6k F7k s3 F14
f3 ,

Sk6 Gk6 Gk7 s3 Gk14 f3 ,

k
k
k
k
T6 H6 H7 s3 H14 f3 ,
k
Rk7 F8k F7k c1 F14
g1 , Sk7 Gk8 Gk7 c1 Gk14 g1 ,
k
k
k
k
T7 H8 H7 c1 H14 g1 ,
k
Rk8 F9k F7k c2 F14
g2 , Sk8 Gk9 Gk7 c2 Gk14 g2 ,
k
g2 ,
T8k H9k H7k c2 H14
k
k
k
k
R9 F10 F7 c3 F14 g3 , Sk9 Gk10 Gk7 c3 Gk14 g3 ,
k
k
H7k c3 H14
g3 ,
T9k H10
k
k
Rk10 F11
F7k d1 F14
h1 , Sk10 Gk11 Gk7 d1 Gk14 h1 ,
k
k
k
T10
H11
H7k d1 H14
h1 ,
k
k
Rk11 F12
F7k d2 F14
h2 ,

Sk11 Gk12 Gk7 d2 Gk14 h2 ,

k
k
k
H12
H7k d2 H14
h2 ,
T11
k
k
Rk12 F13
F7k d3 F14
h3 ,
k
T12

k
k
H13
H7k d3 H14
h3 ;

Sk12 Gk13 Gk7 d3 Gk14 h3 ,

References
[1] J.E.S. Garcao, C.M. Mota Soares, C.A. Mota Soares, J.N. Reddy, Analysis of
laminated adaptive structures using layerwise nite element models, Comput. Struct. 82 (2004) 19391959.
[2] J.N. Reddy, A simple higher-order theory for laminated composite plates,
J. Appl. Mech. 51 (1984) 745752.
[3] A. Tessler, E. Saether, A computationally viable higher-order theory for
laminated composite plates, Int. J. Numer. Mech. Eng. 31 (1991) 10691086.
[4] K. Rohwer, Application of higher order theories to the bending analysis of
layered composite plates, Int. J. Solids Struct. 29 (1992) 105119.
[5] T. Kant, S.R. Marur, G.S. Rao, Analytical solution to the dynamic analysis of
laminated beams using higher order rened theory, Compos. Struct. 40
(1998) 19.
[6] T. Kant, K. Swaminathan, Analytical solutions for the static analysis of
laminated composite and sandwich plates based on a higher order rened
theory, Compos. Struct. 56 (2002) 329344.
[7] M.K. Rao, K. Scherbatiuk, Y.M. Desai, A.H. Shah, Natural vibrations of
laminated and sandwich plates, J. Eng. Mech. 130 (2004) 12681278.
[8] M.K. Rao, Y.M. Desai, Analytical solutions for vibrations of laminated and
sandwich plates using mixed theory, Compos. Struct. 63 (2004) 361373.
[9] J.B. Dafedar, Y.M. Desai, A.A. Mufti, Stability of sandwich plates by mixed,
higher-order analytical formulation, Int. J. Solids Struct. 40 (2003)
45014517.
[10] J.B. Dafedar, Y.M. Desai, Stability of composite and sandwich struts by mixed
formulation, J. Eng. Mech. 130 (2004) 762770.
[11] W.J. Chen, Z. Wu, A selective review on recent development of displacementbased laminated plate theories, Recent Patents Mechan. Eng. 1 (2008) 2944.
[12] E. Carrera, Historical review of zig-zag theories for multilayered plates and
shells, Appl. Mech. Rev. 56 (2003) 287308.
[13] M. Cho, J. Oh, Higher order zig-zag plate theory under thermo-electricmechanical loads combined, Composites: Part B 34 (2003) 6782.
[14] S. Kapuria, G.G.S. Achary, An efcient higher order zigzag theory for
laminated plates subjected to thermal loading, Int. J. Solids Struct. 41
(2004) 46614684.
[15] M. Cho, J. Oh, Higher order zig-zag theory for fully coupled thermo-electricmechanical smart composite plates, Int. J. Solids Struct. 41 (2004)
13311356.
[16] S. Kapuria, A. Ahmed, P.C. Dumir, Static and dynamic thermo-electromechanical analysis of angle-ply hybrid piezoelectric beams using an
efcient coupled zigzag theory, Compos. Sci. Tech. 64 (2004) 24632475.
[17] S. Kapuria, G.G.S. Achary, A coupled zigzag theory for the dynamic of
piezoelectric hybrid cross-ply plates, Arch. Appl. Mech. 75 (2005) 4257.

W. Zhen et al. / Finite Elements in Analysis and Design 48 (2012) 13461357

[18] J. Oh, M. Cho, Higher order zig-zag theory for smart composite shells under
mechanical-thermo-electric loading, Int. J. Solids Struct. 44 (2007) 100127.
[19] J. Oh, M. Cho, A nite element based on cubic zig-zag plate theory for the
prediction of thermo-electric-mechanical behaviors, Int. J. Solids Struct. 41
(2004) 13571375.
[20] U. Icardi, Eight-noded zig-zag element for deection and stress analysis of
plates with general lay-up, Compos. B: Eng. 29 (1998) 425441.
[21] A. Chakrabarti, A.H. Sheikh, Vibration of laminate-faced sandwich plate by a
new rened element, J. Aerospace Eng. 17 (2004) 123134.
[22] S.D. Kulkarni, S. Kapuria, Free vibration analysis of composite and sandwich
plates using an improved discrete Kirchhoff quadrilateral element based on
third-order zigzag theory, Comput. Mech. 42 (2008) 803824.
[23] M.K. Pandit, A.H. Sheikh, B.N. Singh, An improved higher order zigzag theory
for the static analysis of laminated sandwich plate with soft core, Finite
Elements Anal. Des. 44 (2008) 602610.
[24] M.K. Pandit, B.N. Singh, A.H. Sheikh, Stochastic perturbation-based nite
element for deection statics of soft core sandwich plate with random
material properties, Int. J. Mech. Sci. 51 (2009) 363371.
[25] Z. Wu, W.J. Chen, A global-local higher order theory including interlaminar stress
continuity and C0 plate bending element, Comput. Mech. 45 (2010) 387400.

1357

[26] X.Y. Li, D. Liu., Generalized laminate theories based on double superposition
hypothesis, Int. J, Numer. Methods Eng. 40 (1997) 11971212.
[27] K.Y. Sze, R.G. Chen, Y.K. Cheung, Finite element model with continuous
transverse shear stress for composite laminates in cylindrical bending, Finite
Element Anal. Des. 31 (1998) 153164.
[28] N.J. Pagano, Exact solutions for composite laminates in cylindrical bending,
J. Compos. Mater. 3 (1969) 398411.
[29] M. Cho, Y.J. Choi, A new postprocessing method for laminated composites of
general lamination congurations, Compos. Struct. 54 (2001) 397406.
[30] H. Matsunaga, A comparison between 2-D single-layer and 3-D layerwise
theories for computing interlaminar stresses of laminated composite and
sandwich plates subjected to thermal loadings, Compos. Struct. 64 (2004)
161177.
[31] A.E. Bogdanovich, S.P. Yushanov, Three-dimensional variational analysis of
Paganos problems for laminated composite plates, Compos. Sci. Tech. 60
(2000) 24072425.
[32] W.Q. Chen, K.Y. Lee, Three-dimensional exact analysis of angle-ply laminates
in cylindrical bending with interfacial damage via state-space method,
Compos. Struct. 64 (2004) 275283.