ARTICLE IN PRESS

International Journal of Mechanical Sciences 52 (2010) 819828

Contents lists available at ScienceDirect

International Journal of Mechanical Sciences

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

Buckling of rectangular plates with various boundary conditions loaded by

non-uniform inplane loads
Sarat Kumar Panda, L.S. Ramachandra 
Department of Civil Engineering, IIT Kharagpur 721302, India

a r t i c l e in fo

abstract

Article history:
Received 20 December 2008
Received in revised form
20 December 2009
Accepted 18 January 2010
Available online 25 January 2010

In the present paper, buckling loads of rectangular composite plates having nine sets of different
boundary conditions and subjected to non-uniform inplane loading are presented considering higher
order shear deformation theory (HSDT). As the applied inplane load is non-uniform, the buckling load is
evaluated in two steps. In the rst step the plane elasticity problem is solved to evaluate the stress
distribution within the prebuckling range. Using the above stress distribution the plate buckling
equations are derived from the principle of minimum total potential energy. Adopting Galerkins
approximation, the governing partial differential equations are converted into a set of homogeneous
linear algebraic equations. The critical buckling load is obtained from the solution of the associated
linear eigenvalue problem. The present buckling loads are compared with the published results
wherever available. The buckling loads obtained from the present method for plate with various
boundary conditions and subjected to non-uniform inplane loading are found to be in excellent
agreement with those obtained from commercial software ANSYS. Buckling mode shapes of plate for
different boundary conditions with non-uniform inplane loadings are also presented.
& 2010 Elsevier Ltd. All rights reserved.

Keywords:
Buckling
Non-uniform inplane loading
Parabolic loading
Ritz method
Galerkin method

1. Introduction
Often, plates are a part of complex structural system and hence
load coming on it may not be always uniform. For example, in the
case of I-beam or wide anged beam subjected to bending
moment at the ends or lateral loads on the ange, the web of the
beam is subjected to non-uniform inplane loads. The load exerted
on the aircraft wings, or on the stiffened plate in the ship
structures or on the slabs of a multi-storey building by the
adjoining structures usually is non-uniform. The type of distribution in an actual structure depends on the relative stiffnesses of
the adjoining elements. Behaviour of structures subjected to nonuniform inplane compressive loading and shear loading is
important in aircraft, civil and ship-building industries. Much
work has been reported in the literature on the buckling of
rectangular plates subjected to uniform inplane loading. However,
very few papers deal with the buckling of plates subjected to nonuniform inplane loads. Buckling of plates subjected to sinusoidal
[1] and parabolic [2] inplane compressive loading was obtained
by earlier researchers based on unrealistic inplane stress
distribution. Wang et al. [3] have adopted Galerkin procedure
with Legendre polynomials as shape function to analyse buckling
of rectangular plates subjected to linearly varying inplane edge
compressive load with two loaded edges simply supported, one

 Corresponding author. Tel.: + 91 3222 283444; fax: + 91 3222 282254.

E-mail address: lsr@civil.iitkgp.ernet.in (L.S. Ramachandra).

0020-7403/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijmecsci.2010.01.009

side free and the other side simply supported, clamped or

rotationally restrained. Biggers and his co-workers have exploited
the stiffness-tailoring concept to improve the buckling load
capacity of plates subjected to, compressive load [4], and shear
load [5]. Whereas, Baranski and Biggers [6] have used the same
concept to study the postbuckling response of damaged composite plates. In a companion paper Xie and Biggers [7] have
extended the stiffness-tailoring concept to improve the compressive buckling loads and ultimate loads of at pates and curved
panels with cutouts. Buckling of moderately thick composite
plates subjected to partial edge compression was studied by
Sundaresan et al. [8] within the framework of nite element
method. Solving the prebuckling equations, authors obtained
stress distributions within the plate and hence evaluated the
geometric stiffness matrix. Bert and Devarakonda [9] studied
buckling analysis of simply supported rectangular Kirchhoff plate
subjected to sinusoidal distribution of inplane loading by superposition method based on more realistic but approximate stress
distribution. In recent years, Kang and Leissa [10,11], Leissa and
Kang [12] presented exact solutions for the Kirchhoff plate having
two opposite edges simply supported subjected to linearly
varying inplane loading. They have considered all other possible
boundary conditions on the unloaded edges. As the loaded edge is
simply supported, authors assumed the transverse displacement
(w) to vary as sin((mpx)/a) (where a is the size of the plate along
x-direction and b along y-direction) and reduced the governing
partial differential equation to an ordinary differential equation in
y with variable coefcients, for which an exact solution was

ARTICLE IN PRESS
820

S. Kumar Panda, L.S. Ramachandra / International Journal of Mechanical Sciences 52 (2010) 819828

obtained in terms of power series (i.e., method of Frobenius).

Applying the boundary conditions at y= 0 and b yields the
eigenvalue problem for nding the buckling load. Zhong and Gu
[13] studied buckling analysis of ReissnerMindlin plate with
various thicknesses to width ratio and subjected to linearly
varying inplane load. For the case of linearly varying load the
stress eld within the plate coincides with the applied inplane
load distribution. Gurdal et al. [14] worked on ber orientation
variation for at rectangular composite laminates that possess
variable stiffness properties. The variable stiffness concept
provides exibility to the designer for trade-offs between overall
panel stiffness and buckling load. Recently, Wang et al. [15]
obtained the buckling loads of thin rectangular plates under
parabolic edge compression by differential quadrature (DQ)
method. Authors considered nine possible combinations of
boundary condition of the plate in their study. Jana and Bhaskar
[16] have solved the plane elasticity problem exactly by superposition of Airys stress function represented by Fourier series.
They have also obtained the inplane stress distribution by the
extended Kantorovich method based on the principle of minimum
complementary energy. Using these distributions, authors have
obtained buckling loads for simply supported plate by Galerkin
method for various inplane load distributions.
From the above literature survey it is observed that, buckling
loads of layered composite plates subjected to parabolically
distributed inplane loads are not available in the literature. In
this study, buckling loads of isotropic and composite plates
subjected to non-uniform inplane loads are evaluated for nine
different sets of boundary conditions of the plate considering
higher order shear deformation theory proposed by Reddy [17]. In
the rst step, the plane elasticity problem is solved to evaluate the
stress distribution within the prebuckling range by Ritz procedure. Using the above stress distribution and adopting multi-term
Galerkins approximation, the governing partial differential
equations of plate buckling are converted into a set of homogeneous linear algebraic equations. The critical buckling load is
obtained from the solution of associated linear eigenvalue
problem. For the nine cases of boundary conditions, appropriate
beam functions are used as displacement eld approximation in
Galerkins method. When the two loaded edges are simply
supported and applied inplane load is uniform or linearly varying,
the plate buckles with a particular number of half-waves in the
loading direction depending on the length to width ratio of
the plate and in combination of two or more half-waves along the
unloaded edge. Similarly if the applied inplane loading is nonuniform the buckling mode is a combination of two or more halfwaves in both loaded direction as well as the unloaded direction
independent of boundary conditions. The buckling loads obtained
by the present method are compared with those of Leissa and Kang
[12] and Wang et al. [15] wherever possible. The present results
compare well with the literature values. The present results are
also compared with the buckling loads obtained from commercial
nite element software ANSYS and found to compare well.

2. Formulation
Consider a composite rectangular plate having length a and
breadth b and made up of n layers of equal thickness. The coordinate system is such that the middle plane coincides with the
xy plane and the z-axis is perpendicular to the middle plane.
Using Reddys higher order shear deformation theory, the
displacement eld can be written as
u u0 zw0;x f zf1 ;

v v0 zw0;y f zf2 ;

w wo

where f1 Fx wo;x ; f2 Fy wo;y ; f(z)=z[1(4/3)(z/h)2]; h is the

thickness of the plate. Fx is the rotation of normal to midplane
about y-axis and Fy is the rotation of the normal to midplane about
x-axis due to shear deformation alone. f1 and f2 are, respectively,
the total rotation of normals to midplane about y- and x-axis. The
von Karman nonlinear straindisplacement relations at a generic
point z distance away from the midplane can be written as

ex eox zwo;xx f zf1;x

ey eoy zwo;yy f zf2;y
gxy eoxy 2zwo;xy f zf1;y f zf2;x

where

gxz u;z w;x f 0 zf1 ;

gyz v;z w;y f 0 zf2

3
o o
x, y

The superscript o refers to strain in the middle plane. e e and

goxy are the reference surface strains and are dened as

1
1
eoy vo;y wo;y 2 ; goxy uo;y vo;x wo;x wo;y 4
2
2
The stress strain relations for the composite plate in the
material co-ordinate axes are given by

eox uo;x wo;x 2 ;

fsg Q feg
n
o
fsgT sx sy tyz txz txy ;
n
o
fegT ex ey gyz gxz gxy

where [Q]ij is the reduced stiffness matrix in material co-ordinate

system, {s}T cartesian components of stress at any point and {e}T
are the corresponding strains. The governing partial differential
equations of nonlinear buckling of plate are derived from the
principle of minimum total potential energy and is stated as
ZZ
d1 p  fnxx Nx ;x nxy Nxy ;y du nxy Nxy ;x
R

nyy Ny ;y dv Mx;xx 2Mxy;xy My;yy  nxx Nx w;x
 
nxy Nxy w;y ;x  nxy Nxy w;x nyy Ny w;y g;y dw
Px;x Pxy;y Qxa df1 Pxy;x Py;y Qya df2 gdxdy
Z b
Z b




nxx Nx N x dudy
nxy Nxy N xy dvdy

0
0
Z b
Z b
Z b
@@w

Mx
Px @f1 dy
Pxy @f2 dy
dy
@x
0
0
0

Z b
@Mxy
@w
@w
nxx Nx

Qx
dwdy
nxy Nxy
@y
@x
@y
Z 0a
Z a





nyy Ny N y dvdx

nxy Nxy N xy dudx
0
0
Z a
Z a
Z a
@@w

My
Py @f2 dx
Pxy @f1 dx
dx
@y
0
0

Z 0a 
@Mxy
@w
@w

nyy Ny
Qy
dwdx 0
nxy Nxy
@x
@y
@x
0

6
The force and moment resultants are dened as
00
1 0
1 0
11
0
1
Mx
Px
sx
Nx
Z h=2
BB N C B M C B P CC
Bs C
@@ y A; @ y A; @ y AA
@ y A1; z; f zdz
h=2
Nxy
Mxy
Pxy
txy
Qx ; Qy

h=2

txz ; tyz dz

txz ; tyz f 0 zdz

h=2

Qxa ; Qya

h=2
h=2

where f 0 (z)=(d/dz)(f(z)); Nx, Ny, Nxy, and Mx, My, Mxy are,
respectively, the force and moment resultants; Px, Py, Pxy are

ARTICLE IN PRESS
S. Kumar Panda, L.S. Ramachandra / International Journal of Mechanical Sciences 52 (2010) 819828

additional moment resultants due to additional changes of

curvature f1,x, f2,y, (f2,x + f1,y) due to shear deformation. Qxa ; Qya
are additional transverse shear force resultants. nxx, nyy, nxy are the
plate internal stress resultants due to applied end non-uniform
inplane loading and can be determined by the Principle of Least
Work applied to membrane problem. Minimizing the total
potential energy and substituting for force and moment resultants
in terms of displacement components, the partial differential
equations governing the postbuckling analysis of cross-ply
composite plate in displacement variables are obtained and are
given in Appendix A (Eqs. (A.1)(A.5)).
2.1. Plate prebuckling analysis
In the present investigation, parabolically and linearly varying
inplane compressive loads are considered. However, for linearly
varying in-plane load, the stress distribution within the plate
coincides with the applied edge loading. In the case of parabolic
non-uniform inplane loading, the stress distribution within the
plate due to applied inplane loading is obtained by solving the
plate membrane problem. The correct stress distribution within
the plate is the one which satises the boundary condition and
minimizes the membrane strain energy of the plate. The
membrane strain energy of a plate of thickness h of composite
plate is given by
8
92
9
31 8
n
>
ZZ > nxx >
= A11 A12 A16
< xx >
=
h <
6
7
nyy 4 A12 A22 A26 5
nyy dxdy
10
V
>
2 >
:n >
; A A A
:n >
;
xy
xy
A
16
26
66
where
nyy

@2 F
;
@x2

nxx

@2 F
;
@y2

nxy 

@2 F
;
@x@y

Aij

h=2

Q ij dz

11

h=2

where Q ij the transformed reduced stiffness and F is the stress

function. The membrane strain energy is minimized in this study
using Ritz method [18]. The boundary conditions of the plate
membrane problem are given here for parabolically varying
uniaxial inplane load (see Fig. 1) as the following.For
y
y
1 y 0; b N xy 0 N y 0
x 0; a N xy 0 N x 4N 0
b
b
12

where

F0 2N 0



y2 y y2
 2
3 b 2b

which gives
N yy

@2 F0
0;
@x2

N xx 

N xy 

@2 F0
0
@x@y

@2 F0
y
y
4N 0
1
b
b
@y2

The remaining functions F1, F2, F3 are chosen such that the
stresses corresponding to them vanish at the boundary.
Substituting Eqs. (11) and (14) into Eq. (10) and carrying out
integration, an expression in second degree in a1, a2 and a3 is
obtained. Then, the strain energy function V is minimized with
respect to the constants a1, a2 and a3. Then the constants a1, a2
and a3 are evaluated from the 3 algebraic equations resulting
from the condition, (qV/qa1)= 0, (qV/qa2) =0, (qV/qa3)=0.
2.2. Plate buckling analyses
The critical buckling load of composite rectangular plate with
various boundary conditions and subjected to parabolically
varying inplane compressive load is obtained using Galerkins
method. In the present investigation following nine sets of
boundary conditions are considered: SSSS, SSCS, SCSS, CSCS, SCSC,
SSCC, CCSC, CCCS and CCCC, where S stands for simply supported
edge and C for clamped edge. The letters indicate the boundary
conditions on the edge of the plate in the anti-clockwise fashion
starting from the left hand corner. In the Galerkins method, the
out-of-plane displacement eld w(x,y) satisfying the boundary
conditions of the plate is expressed as the product of beam
function as [19]
wx; y

1 X
1
X

Xm xYn y

15

m1n1

where Xm(x) and Yn(y) are the eigen functions of the beam having
the same boundary conditions as that of two opposite edges of the
plate. This choice of functions satises all boundary conditions of the
plate exactly. In present case following beam functions are adopted.
(a) Simply support along two opposite edges, at x =0 and x= a

The stress function is assumed in the form of a series:

F F0 a1 F1 a2 F2 a3 F3 a4 F4   

13

where a1,a2,a3, y are constants and are determined such that the
boundary conditions (12) are satised. In the present case, for
parabolic loading, the stress function is assumed as


y2 y y2
F 2N 0
 2 x2 ax2 y2 yb2 a1 a2 x a3 y   
3 b 2b
14

mpx
m 1; 2; 3; . . .
16
a
(b) Clamped support along two opposite edges, at x= 0 and x= a
ss
Xm
x sin

cc
Xm
x cosxm




x 1
sinxm =2
x 1
coshxm



a 2
sinhxm =2
a 2

m 2; 4; 6; . . .

17
where xm are obtained as roots of
tanxm =2 tanhxm =2 0

18

and

y
y
N x = 4 N0 (1 )
b
b

821

cc
Xm
x sinxm




x 1
sinxm =2
x 1



coshxm
a 2
sinhxm =2
a 2

m 3; 5; 7; . . .

where xn are obtained roots of

tanxm =2tanhxm =2 0

b
x
Fig. 1. Geometry and loading of the plate.

19

(c) Clamped support along the edge, x =0 and simply supported

at x =a
cs
Xm
x sinxm




x 1
sinxm =2
x 1
sinhxm



2a 2
sinhxm =2
2a 2

m 2; 3; 4; . . .

20
where xm are obtained as roots of Eq. (19).

ARTICLE IN PRESS
822

S. Kumar Panda, L.S. Ramachandra / International Journal of Mechanical Sciences 52 (2010) 819828

The functions Yn(y) are similarly chosen by the condition at y= 0

and y= b by replacing x by y, a by b and m by n in respective
equations, where m and n are, respectively, the number of nodal
lines along x and y directions. In all cases, only normal inplane
displacements are allowed and inplane tangential displacements
and out of plane displacements are prevented.The boundary
conditions for SSSS plate are
vo wo Px

nxx Nx N x ;
ny Ny N y ;

fo2

uo wo f1 Py My 0

21

at y b

25

Following displacement elds satisfy the above boundary conditions:

at x 0; a
at y 0; b

at y 0

uo wo f1 Py My 0

ny Ny N y ;

u~ o
Mx 0

w0;y uo vo wo f1 f2 0

v~ o

mpx
npy
Umn cos
cos
a
b
m1n1
j
i
X
X

j
i
X
X

cs
Vmn Xm
xXncs y

m1n1

Following displacement elds satisfy the above boundary conditions:

mpx npy
Umn cos
sin
a
b
m1n1

j
i
X
X

u~ o

j
i
X
X

v~ o

Vmn sin

mpx
a

m1n1
j
i
X
X

~o
w

cos

f~ 1
o

npy
b

f~ 2

ss
Wmn Xm
xXnss y

mpx npy
Kmn cos
sin
a
b
m1n1
j
i
X
X

f~ 2

Lmn sin

mpx
a

m1n1

cos

cs
Wmn Xm
xXncs y

m1n1
j
i
X
X

mpx
npy
Kmn cos
cos
a
b
m1n1
j
i
X
X

cs
Lmn Xm
xXncs y

26

The displacement elds for other boundary conditions can be

assumed by suitably combining the displacement functions
described above.

j
i
X
X

j
i
X
X

m1n1

m1n1

f~ 1

~o
w

npy
b

22
3. Numerical results and discussion

The boundary conditions for CCCC plate are

o

at x 0; a

at y 0; b

w0;x uo vo wo f1 f2 0
w0;y uo vo wo f1 f2 0

3.1. Prebuckling analysis

23

Following displacement elds satisfy the above boundary conditions:

N x 4N 0

mpx npy
u~ o
Umn sin
sin
a
b
m1 n1
j
i X
X

j
i X
X

v~ o

Vmn sin

m1 n1

~o
w
o
f~ 1

j
i X
X
m1 n1
j
i X
X

h
y i
Nx N0 1Z
;
b

cc
Wmn Xm
xXncc y

Kmn sin

m1 n1

npy
sin
b

mpx npy
o
f~ 2
Lmn sin
sin
a
b
m1 n1
i
X

j
X

24

The boundary conditions for plate (CSCS) with clamped support at

x= 0, y= 0 and simple support at x= a, y=b are
o

w0;x uo vo wo f1 f2 0
nxx Nx N x ;

v w Px

fo2

y
y
1
b
b

27

Linearly varying loads are dened as

mpx npy
sin
a
b

mpx

In present case, non-uniform inplane loads are assumed to

vary according to parabolic and linearly varying functions.
Parabolically varying inplane load is represented by (see Fig. 1)

at x 0
Mx 0

at x a

y A 0; b

28

By taking various values of Z, we obtain different inplane

load distribution (uniform (Z = 0), trapezoidal (Z =0.5), triangular
(Z =1), partial tension (Z = 1.5) and pure bending (Z =2.0)). Initially
the plate membrane equations are solved to determine the
stress distribution within the plate as described in Section 2.1.
In the present case, the stress function is represented as a
truncated series with four terms (14). After evaluating
the constants ai (i =1,2,3), the stress distribution within the
plate are obtained. The explicit expressions for constants ai for the
case of isotropic plate are given below. Similar expressions
in the case of composite plate are given in Appendix A
(Eqs. (A.7)(A.9))

a1

N 0 33a8 79:44a6 b2 241:9662a4 b4 79:44a2 b6 33b8

14
12
b 5:1844b a2 36:4435b10 a4 29:0127b8 a6 36:4435b6 a8 5:1844b4 a10 b2 a12

29

a2

N 0 114:4a7 197:6a5 b2 27:2a3 b4 3:52ab6

b14 5:1844b12 a2 36:4435b10 a4 29:0127b8 a6 36:4435b6 a8 5:1844b4 a10 b2 a12

30

a3

N 0 3:52a6 27:2a4 b2 197:6a2 b4 114:4b6

b14 5:1844b12 a2 36:4435b10 a4 29:0127b8 a6 36:4435b6 a8 5:1844b4 a10 b2 a12

31

a/b
0.4
m= 1

0.5
m =1

0.6
m= 1

0.7
m =1

0.8
m= 1

0.9
m= 1

1.0
m =2

Leissa and Kang [12]

Energy method [12]
Present methodn
Present method

93.247
93.2
93.305
93.231

75.910
75.9
75.943
75.879

69.632
69.6
69.652
69.588

69.095
69.1
69.108
69.036

72.084
71.9
72.093
72.007

77.545
77.3
77.545
77.443

75.910
75.9
75.943
75.774

a/b

g=1

0.4
m=1

0.5
m =1

0.6
m= 1

0.65
m= 1

0.7
m= 1

0.8
m=1

0.9
m= 1

1.0
m= 2

1.2
m= 2

1.4
m=2

Leissa and Kang [12]

Energy method [12]
Present methodn
Present method

174.4
175.0
174.5
174.3

145.2
145.0
145.3
145.1

134.8
135.0
134.8
134.6

133.7
133.8
133.7
133.6

134.6
134.7
134.6
134.4

141.0
141.0
141.0
140.8

152.0
152.1
152.0
151.8

145.2
145.0
145.3
144.9

134.8
135.0
134.8
134.4

134.6
135.0
134.6
134.1

a/b

g =2

0.3
m= 1

0.4
m=1

0.5
m =1

0.6
m= 1

0.7
m= 2

0.8
m= 2

1.0
m= 2

1.2
m= 3

1.5
m= 3

2.0
m=4

Leissa and Kang [12]

Energy method [12]
Present method*
Present method

464.5
467.0
467.2
466.9

400.4
402.0
401.5
401.3

391.5
392.2
392.1
391.8

411.8
412.2
412.1
411.7

422.5
424.0
424.1
422.4

400.4
402.0
401.5
400.3

391.5
392.0
392.1
391.0

400.4
402.0
401.5
399.6

391.5
392.0
392.1
389.7

391.5
392.0
392.1
387.9

Without Shear Deformation and inplane displacements.

ARTICLE IN PRESS

g =0

S. Kumar Panda, L.S. Ramachandra / International Journal of Mechanical Sciences 52 (2010) 819828

Table 1
Comparision of dimensionless buckling load coefcient ki for SCSC rectangular plate (a/h = 100) subjected to linearly varying inplane load.

823

ARTICLE IN PRESS
824

S. Kumar Panda, L.S. Ramachandra / International Journal of Mechanical Sciences 52 (2010) 819828

For linearly varying inplane load the computation of inplane

stress distribution is unnecessary since the internal stress
distribution coincides with the applied inplane load distribution.

The following mechanical properties are assumed, E11 = E22,

G23 =G13 = G12 =E22/2.5, n12 = 0.25 for isotropic plate and E1 = E2 = 25,
G12 =G13 = 0.5E2, G23 = 0.6E2, n12 = 0.25 for composite lamina in the
analysis. Dropping the nonlinear terms in the plate nonlinear
equations, the plate buckling equations are obtained. Using
displacement elds given in Section 2.2 and adopting Galerkins
method, the governing partial differential equations of plate
buckling equations are converted into a set of linear homogeneous algebraic equations. For a nontrivial solution this is posed
as an eigenvalue problem, solving which critical buckling loads
are obtained. To validate the present formulation, the dimensionless buckling load coefcients of a SCSC plate obtained by the
present method (neglecting shear deformation and considering
only w displacement) are compared with that of Leissa and Kang
[12] in Table 1 for Z = 0, 1 and 2 and for various aspect ratios. The
buckling loads obtained by the present method considering
higher order shear deformation and inplane displacements are
also given in the table. It is observed that the present results
without shear deformation compare well with the energy method
values. For uniform (Z = 0) compressive load, the present results
compare well with that of Leissa and Kang. For triangular (Z = 1)
and pure inplane (Z =2) loading, the present results obtained
without shear deformation compare well with that of Leissa and
Kang results for a/b o1.0. The dimensionless buckling load
coefcients for a simply supported isotropic plate (h/a= 0.01)
obtained by the present method are given in Table 2 for uniform
and parabolic load distributions. In case of plates with aspect
ratios a/b= 1 and 3 and subjected to uniformly distributed inplane
loads the converged value of buckling load is obtained by
considering one term in the displacement eld approximations.
For parabolically distributed loads, 6 terms are required to obtain
the converged buckling load in the case of square plate. However
for the plate with aspect ratio a/b= 3, 15 terms are required to
obtain the converged buckling load. The number of terms
required to obtain the converged buckling load also varies
depending upon the boundary conditions. In all further
calculations 36 terms are considered.The variation of
dimensionless buckling load coefcients ki( = Ncrb2/p2D) of a
SCSC isotropic plate (a/h =100) with two opposite simply
supported edges subjected to linearly varying inplane load
against aspect ratio of the plate is shown in Fig. 2, for various

=2

400
Buckling Coefficient

3.2. Buckling analysis

450
m=1

350

m=2

m=3

m=4

m=6

m=5

300
250

= 1.5

200
150

= 1.0

100

= 0.5
=0

50

m=1

m=2

m=3

m=4

0
0.0

0.5

1.0

1.5
a/b

2.0

2.5

3.0

Fig. 2. Variation of buckling coefcients of SCSC plate with the aspect ratio (a/b)
for different inplane load distributions.

Fig. 3. Buckling modes for SCSC rectangular isotropic plate under uniform inplane
load distributions (Z = 0).

Table 2
Dimensionless buckling load of SSSS rectangular isotropic plate with uniform and parabolic in-plane loadings.
a/b= 1

a/b = 3

Uniform

Parabolic

Uniform

Parabolic

Mode
(m  n)/term

Buckling
coefcient (ki)

Modes
(m  n)/terms

Buckling
coefcient (ki)

Mode
(m  n)/term

Buckling
coefcient (ki)

Modes
(m  n)/terms

Buckling
Coefcient (ki)

[1  1]/ (1)

3.997
(1,1)

[1  1] /(1)
[2  2] /(4)
[2  3] /(6)
[3  2] /(6)
[3  3] /(9)
[4  3]/(12)
[3  4]/(12)

5.252
5.251
5.250
5.241
5.241
5.241
5.241

[1  1]/ (1)

3.997
(3,1)

[1  1] /(1)
[3  3] /(9)
[4  3]/(12)
[5  3]/(15)
[6  3]/(18)
[6  4]/(24)
[6  5]/(30)
[6  6]/(36)

5.633
5.632
5.622
5.547
5.547
5.547
5.547
5.547

Ki( =Ncrb2/p2D).

ARTICLE IN PRESS
S. Kumar Panda, L.S. Ramachandra / International Journal of Mechanical Sciences 52 (2010) 819828

825

values of Z. It is observed from the gure that for very long

(a/b43) plates, the buckling loads remains the same. For uniform,
trapezoidal and triangular loading (i.e., Z = 0, 0.5 and 1) the
buckling mode in the loading direction is four half-waves (m) for
the plate aspect ratio a/b= 2.8, where as for partial tension and
pure bending the buckling mode is ve and six half-waves (m)
respectively (see Fig. 2). The three dimensional buckling modes
with contour plots (lines of constant displacements) for all the
above cases is shown in Figs. 36.Nine different plate boundary
conditions are considered with parabolic inplane load
N x 4N 0 y=b1y=b distributions. For all the boundary
conditions, the aspect ratio a/b is varied up to 3. Effect of shear
deformation is shown in Figs. 79 for plate with SSSS, SCSS and
SCSC boundary conditions respectively for length to thickness
ratios a/h= 100, 50, 20 and 10. The buckling loads are calculated
for all these boundary conditions by considering 6 terms along
x-axis and 6 terms along y-axis in the multi-term Galerkin
method from convergence considerations. However, for SSSS plate
less number of terms is required to obtain converged buckling
loads as shown in Table 2. In Figs. 79, m= 1, m =2,y, indicate the
Fig. 6. Buckling modes for SCSC rectangular isotropic plate under pure inplane
bending load distributions (Z = 2.0).

16

Buckling Coefficient

14

12
10

8
m=1

a/h = 100
a/h = 50
a/h = 25
a/h = 10
m=3

m=2

4
Fig. 4. Buckling modes for SCSC rectangular isotropic plate under triangular
inplane load distributions (Z = 1.0).

2
0.0

0.5

1.0

1.5
a/b

2.0

2.5

3.0

Fig. 7. Variation of buckling coefcients of SSSS plate with parabolic inplane

loading for different aspect ratios (a/b) and length to thickness (a/h) ratio.

Fig. 5. Buckling modes for SCSC rectangular isotropic plate under partial inplane
tension load distributions (Z = 1.5).

dominant half-waves along x-direction. With the decrease of

length to thickness ratio, the non-dimensional buckling load
decreases and the curve bends away as shown in the gure. From
Figs. 79, it is clear that the effect of shear deformation increases
with the increase in aspect ratio and end restraint. Thus the effect
of shear deformation is the maximum for SCSC boundary
condition as observed from Fig. 9. It is observed that the plate
buckles into more number of half-waves for the same aspect ratio
as the edge restraint increases from SSSS to SCSC.
The plates with SSSS, SCSS and SCSC boundary conditions, the
plates buckle into two half-waves for a/b ratio beyond 1.325,
1.075 and 0.9, respectively. Buckling into four half-waves is
possible only for SCSS and SCSC plates within a/b= 3. The plate
buckles into three half-waves at a/b ratio 2.475 for SSSS boundary
condition, 1.950 for SCSS and 1.625 for SCSC boundary condition
(see Figs. 79). The buckling modes for SSSS, SCSS and SCSC plate
for parabolic inplane loading are shown in Figs. 1012 for aspect

ARTICLE IN PRESS
826

S. Kumar Panda, L.S. Ramachandra / International Journal of Mechanical Sciences 52 (2010) 819828

16
y

Buckling Coefficient

14
12

a/h = 100
a/h = 50
a/h = 20
a/h = 10

b
x

10
m=1

m=3

m=2

6
4
0.0

0.5

1.0

1.5
a/b

2.0

2.5

3.0

Fig. 8. Variation of buckling coefcients of SCSS plate with parabolic inplane

loading for different aspect ratios (a/b) and length to thickness (a/h) ratio.

16

y
a/h = 100
a/h = 50
a/h = 20
a/h = 10

14
Buckling Coefficient

Fig. 10. Buckling modes for SSSS rectangular isotropic plate under parabolic
inplane load distributions.

12

m=2

m=1

10

m=3

m=4

8
6
4
0.0

0.5

1.0

1.5
a/b

2.0

2.5

3.0

Fig. 11. Buckling modes for SCSS rectangular isotropic plate under parabolic
inplane load distributions.

Fig. 9. Variation of buckling coefcients of SCSC plate with parabolic inplane

loading for different aspect ratios (a/b) and length to thickness (a/h) ratio.

ratio of 2.8. It is observed from Fig. 10 that plate with SSSS

boundary condition buckles in three half-waves for aspect ratio
a/b= 2.8. Figs. 11 and 12 show the four half-waves buckling for
SCSS and SCSC plates. It is observed that for the same aspect ratio,
the SCSS and SCSC boundary condition plate buckles into more
number of half-waves due to the increase in boundary restraint.
Recently, Wang et al. [15] obtained the accurate non-dimensional buckling load coefcients of thin rectangular isotropic
plates under parabolic edge compression by differential quadrature (DQ) method for above nine boundary conditions. However, they have not considered shear deformation in their
analysis. The buckling loads of plates with all the above boundary
conditions are also obtained by commercially available nite
element software ANSYS. In ANSYS, 8 noded SHELL93 element has
been used to discretize the plate geometry. The element has
6 degrees of freedom at each node; translations along x, y and
z directions and rotations about the nodal x, y and z axes.
The present results for isotropic plate with nine boundary

Fig. 12. Buckling modes with for SCSC rectangular isotropic plate under parabolic
inplane load distributions.

ARTICLE IN PRESS
S. Kumar Panda, L.S. Ramachandra / International Journal of Mechanical Sciences 52 (2010) 819828

827

Table 3
Comparison of dimensionless buckling load coefcient ki of isotropic rectangular plate (a/h = 100) with different boundary conditions subjected to parabolic in-plane
loading.
Support

SSSS

SSCS

SCSS

CSCS

SCSC

SSCC

CCSC

CCCS

CCCC

Source

Present
Wang et
ANSYS
Present
Wang et
ANSYS
Present
Wang et
ANSYS
Present
Wang et
ANSYS
Present
Wang et
ANSYS
Present
Wang et
ANSYS
Present
Wang et
ANSYS
Present
Wang et
ANSYS
Present
Wang et
ANSYS

a/b

al.[15]

al.[15]

al.[15]

al.[15]

al.[15]

al.[15]

al.[15]

al.[15]

al.[15]

0.4

0.5

0.6

0.8

1.0

1.5

2.0

3.0

9.654
9.663
9.661
17.02
17.02
17.02
10.06
10.06
10.06
30.65
30.69
30.70
10.52
10.54
10.53
17.28
17.29
17.29
17.58
17.59
17.58
30.54
30.58
30.54
31.01
31.03
31.01

7.271
7.274
7.273
12.03
12.03
12.03
7.877
7.888
7.880
21.01
21.02
21.01
8.652
8.663
8.661
12.49
12.50
12.49
13.04
13.06
13.05
21.35
21.38
21.36
21.74
21.80
21.76

6.078
6.080
6.079
9.401
9.399
9.399
6.938
6.940
6.939
15.82
15.83
15.82
8.087
8.092
8.088
10.07
10.09
10.08
10.94
10.95
10.94
16.42
16.44
16.43
17.15
17.17
17.16

5.211
5.211
5.212
7.035
7.045
7.041
6.692
6.698
6.699
10.95
10.96
10.95
8.877
8.887
8.882
8.231
8.233
8.232
9.814
9.821
9.817
12.12
12.12
12.12
13.66
13.71
13.68

5.242
5.262
5.250
6.254
6.277
6.285
7.551
7.573
7.582
9.032
9.054
9.062
9.172
9.194
9.194
7.957
7.971
7.973
9.847
9.868
9.867
10.88
10.91
10.91
13.55
13.58
13.57

5.704
5.734
5.739
6.023
6.058
6.065
7.104
7.135
7.135
8.127
8.153
8.163
9.114
9.141
9.130
7.645
7.679
7.678
9.340
9.393
9.378
9.735
9.767
9.761
11.58
11.63
11.61

5.478
5.628
5.531
5.768
5.825
5.825
7.412
7.482
7.465
7.063
7.123
7.121
9.053
9.120
9.080
7.390
7.598
7.420
8.309
9.367
9.323
9.185
9.246
9.230
11.17
11.26
11.21

5.547
5.630
5.621
5.671
5.756
5.744
7.373
7.456
7.403
6.495
6.571
6.557
9.117
9.345
9.226
7.451
7.544
7.492
9.205
9.352
9.235
8.571
8.693
8.634
10.77
10.92
10.79

Table 4
Dimensionless buckling coefcients kc of rectangular cross-ply laminated (0/90/0)
plate (a/h = 100) with different boundary conditions subjected to parabolic inplane loading.
Support

SSSS
SSCS
SCSS
CSCS
SCSC
SSCC
CCSC
CCCS
CCCC

Source

Present
ANSYS
Present
ANSYS
Present
ANSYS
Present
ANSYS
Present
ANSYS
Present
ANSYS
Present
ANSYS
Present
ANSYS
Present
ANSYS

a/b
0.5

1.0

1.5

2.0

2.5

3.0

35.69
35.71
70.38
70.58
35.73
35.83
133.31
133.51
35.77
35.79
70.31
70.45
70.38
70.42
133.21
133.33
133.31
133.71

41.50
41.55
77.54
78.35
44.88
44.92
145.92
146.28
49.90
49.90
80.79
80.85
83.93
84.01
148.60
148.75
150.69
150.83

54.42
55.61
90.73
91.80
73.53
73.75
162.35
162.63
103.66
103.73
105.81
105.97
127.51
127.63
174.49
175.38
193.26
194.21

87.38
88.87
119.68
120.27
146.77
147.53
192.89
192.68
199.42
201.91
162.17
163.37
218.13
219.35
233.96
234.51
296.00
298.12

154.26
156.81
171.89
173.19
224.23
227.41
247.03
248.00
281.31
284.20
249.86
252.18
315.11
318.14
342.35
344.11
406.95
410.17

227.32
229.55
246.06
248.09
301.95
304.66
333.43
337.21
424.48
428.98
336.10
340.27
437.39
443.24
436.43
438.44
525.41
532.82

conditions are compared with the ANSYS results and that of Wang
et al. [15] in Table 3. The small difference of values between the
three results is due to the reason that in the present method shear
deformation and inplane displacement have been considered,
where as Wang et al. [15] have not considered the same. Critical
buckling coefcients kc(Ncra2/E22p2h3) of 3-layered cross-ply

laminated square plate subjected to non-uniform inplane

parabolic compression is given in Table 4 along with the ANSYS
results. Material properties used in the analysis are E1 = E2 =25,
G12 = G13 = 0.5E2, G23 =0.6E2, n12 =0.25. In ANSYS, linear SHELL99
shell element has been used to discretize the plate geometry. The
element has 6 degrees of freedom at each node; translations along
x, y and z directions and rotations about the nodal x, y and z axes.
The present results compare well with that of Wang et al. [15] and
ANSYS results.

3.3. Conclusions
Leissa and Kang used power series (i.e., method of Frobenius)
method to obtain the buckling load of the plate with linearly
varying inplane loads. Authors mentioned in their conclusion that
whenever the inplane edge loading is more general than linearly
varying ((i.e., Ny = f(y)) the method is not fruitful and suggested
rst to solve the plane elasticity problem to determine Nx, Ny and
Nxy. In present study, the buckling load of a composite plate
subjected to parabolically distributed compressive inplane loads
are reported for the rst time. For this case, rst plane elasticity
problem is solved to determine the stress distributions within the
plate. Using the above stress distribution and adopting Galerkins
approximation the critical buckling loads are evaluated. Beam
functions are used as shape functions in the Galerkin technique. It
is observed that, whenever the plate restrained condition
increases, the number of terms required is more to get the
converged buckling load. When the two loaded edges are simply
supported and applied inplane load is uniform or linearly varying,
the plate buckles with a particular number of half-waves in the
loading direction depending on the length to width ratio of the

ARTICLE IN PRESS
828

S. Kumar Panda, L.S. Ramachandra / International Journal of Mechanical Sciences 52 (2010) 819828

plate and in combination of two or more half-waves along

the unloaded edge. Similarly if the applied inplane loading is
non-uniform the buckling mode is combination of two or more
half-waves in both loaded direction as well as the unloaded
direction independent of boundary conditions. As the applied load
is non-uniform we need to take six terms in x-direction and six
terms in y-direction to get the converged buckling load up to
aspect ratio of a/b= 3 for maximum restrained plate i.e., CCCC

where
p1 uo;x 0:5wo;x 2 ; q1 vo;x 0:5wo;y 2 ;
p2 wo;xx ; q2 wo;yy ; r2 2wo;xy ;
o
o
o
o
p3 f1;x ; q3 f2;y ; r3 f1;y f2;x

r1 uo;x vo;x wo;x wo;y ;

A:6

Explicit expression for a1, a2 and a3 for the case of cross-ply

composite plate are

a1

N 0 a8 65:0455a6 b2 346:8788a4 b4 125:4424a2 b6 3:7194b8

0:0303a12 b2 4:2449a10 b4 70:3424a8 b6 222:9a6 b8 135:6604a4 b10 15:7885a2 b12 0:2174b14

A:7

a2

N 0 a7 46:6708a5 b2 146:0881a3 b4 1:6034ab6

0:0087a12 b2 1:2245a10 b4 20:2911a8 b6 64:3181a6 b8 39:1328a4 b10 4:5544a2 b12 0:0627b14

A:8

a3

N 0 a6 175:7171a4 b2 108:2630a2 b4 4:4737b6

12
10
0:0105a b 1:4728a b3 24:4064a8 b5 77:3628a6 b7 47:0696a4 b9 5:4781a2 b11 0:0754b13

A:9

plate. For other boundary conditions less number of terms is

required to get the converged solution. From Figs. 79, it is clear
that the effect of shear deformation increases with the increase in
aspect ratio and end restraint. It is observed that for the same
aspect ratio, the SCSS and SCSC boundary condition plate buckles
into more number of half-waves due to the increase in boundary
restraint (vide Figs. 11 and 12). The buckling loads of isotropic and
composite plates presented for different boundary conditions and
loading conditions in this paper can be used to benchmark future
studies.

Appendix A
Nonlinear governing partial differential equations of cross-ply
composite plate in displacement variables are,
A11 uo;xx A66 uo;yy A12 A66 vo;xy A11 wo;xx A66 wo;yy wo;x A12 A66 wo;y wo;xy 0

A:1
A12 A66 uo;xy A66 vo;xx A22 vo;yy A66 wo;xx A22 wo;yy wo;y A12 A66 wo;x wo;xy 0

A:2
o

D11 wo;xxxx 2D12 wo;xxyy D22 wo;yyyy 4D66 wo;xxyy E11 f1;xxx
o

E12 f1;xyy f2;xxy E22 f2;yyy 2E66 f1;xyy f2;xxy

wo;xx A11 p1 A12 q1 wo;x A11 p1;x A12 q1;x nxx wo;xx nxx;x wo;x
2wo;xy A66 r1 wo;x A66 r1;y wo;y A66 r1;x 2nxy wo;xy nxy;y wo;x
nxy;x wo;y wo;yy A12 p1 A22 q1 wo;y A12 p1;y A22 q1;y
nyy wo;yy nyy;y wo;y 0
A:3

E22 wo;yyy E12 2E66 wo;xxy F66 f2;xx F22 f2;yy F12 F66 f1;xy H44 f2 0

A:4
o

E11 wo;xxx E12 2E66 wo;xyy F11 f1;xx F66 f1;yy F12 F66 f2;xy H55 f1 0

A:5

References
[1] van der Neut A. Buckling caused by thermal stresses. In: High temperature
effects in aircraft structures. AGARDograph 1958;28:21547.
[2] Benoy MB. An energy solution for buckling of rectangular plate under nonuniform inplane loading. Aeronautical Journal 1969;73:9747.
[3] Wang JTS, Biggers SB, Dickson JN. Buckling of composite plates with a free
edge in edgewise bending and compression. AIAA Journal 1984;22(3):3948.
[4] Biggers SB, Srinivasan S. Postbuckling response of piece-wise uniform tailored
composite plates in compression. Journal of Reinforced Plastics and
Composites 1994;13(9):80321.
[5] Biggers SB, Pageau SS. Shear buckling response of tailored composite plates.
AIAA Journal 1994;32(5):11003.
[6] Baranski AT, Biggers Jr. SB. Postbuckling analysis of tailored composite plates
with progressive damage. Composite and Structures 1999;46:24555.
[7] Xie D, Biggers Jr. SB. Postbuckling analysis with progressive damage modeling
in tailored composite plates and shells with a cutout. Composites and
Structures 2003;59(2):199216.
[8] Sundaresan P, Singh G, Rao VG. Buckling of moderately thick rectangular
composite plate subjected to partial edge compression. International Journal
of Mechanical Science 1998;40(11):110517.
[9] Bert CW, Devarakonda KK. Buckling of rectangular plates subjected to
nonlinearly distributed inplane loading. International Journal of Solids and
Structures 2003;40:4097106.
[10] Kang JH, Leissa AW. Vibration and buckling of SSFSSF rectangular plates
loaded by in-plane moments. International Journal of Structural Stability and
Dynamics 2001;1(4):52743.
[11] Kang JH, Leissa AW. Exact solution for the buckling of rectangular plates
having linearly varying in-plane loading on two opposite simply supported
edges. International Journal of Mechanical Science 2005;42:422038.
[12] Leissa AW, Kang JH. Exact solution for vibration and buckling of an SSC
SSC rectangular plate loaded by linearly varying inplane stresses. International Journal of Mechanical Science 2002;44:19259455.
[13] Zhong H, Gu C. Buckling of simply supported rectangular ReissnerMindlin
plates subjected to linearly varying inplane loading. Journal of Engineering
Mechanics 2006;132(5):57881.
[14] Gurdal Z, Tatting BF, Wu CK. Variable stiffness composite panels: effects of
stiffness variation on the in-plane and buckling response. Composite: Part A,
Applied Science and Manufacturing 2008;39:91122.
[15] Wang X, Wang Xinfeng, Shi X. Accurate buckling loads of thin rectangular
plates under parabolic edge compressions by the differential quadrature
method. International Journal of Mechanical Science 2007;49:44753.
[16] Jana P, Bhaskar K. Stability analysis of simply-supported rectangular plates
under non-uniform uniaxial compression using rigorous and approximate
plane stress solutions. Thin-Walled Structure 2006;44:50716.
[17] Reddy JN. A simple higher-order theory for laminated composite plates.
Journal of Applied Mechanics 1984;51(4):74552.
[18] Timoshenko SP, Gere JM. Theory of elastic stability, 2e. New York: McGrawHill; 1961.
[19] Warburton GB. The vibration of rectangular plates. Proceedings of the
Institution of Mechanical Engineering, Series A 1954;168(12):37184.