Composite Structures 157 (2016) 207221

Contents lists available at ScienceDirect

Composite Structures
journal homepage: www.elsevier.com/locate/compstruct

Numerical study on the free vibration and thermal buckling behavior

of moderately thick functionally graded structures in thermal
environments
Ramkumar Kandasamy a, Rossana Dimitri b, Francesco Tornabene c,
a
b
c

BLK 301, #09-64 Jurong East Street 32, Singapore 600301, Singapore
Department of Innovation Engineering, University of Salento, Via per Monteroni, 73100 Lecce, Italy
DICAM Department of Civil, Chemical, Environmental and Materials Engineering, University of Bologna, Viale del Risorgimento, 2, 40136 Bologna, Italy

a r t i c l e

i n f o

Article history:
Received 8 August 2016
Accepted 28 August 2016
Available online 31 August 2016
Keywords:
Vibration
Thermal buckling
FGM
FSDT
FEM

a b s t r a c t
This paper is aimed at studying the free vibration and thermal buckling behavior of moderately thick
functionally graded material (FGM) structures including plates, cylindrical panels and shells under thermal environments. A numerical investigation is performed by applying the finite element method (FEM).
A formulation based on the first-order shear deformation theory (FSDT) is proposed for the purpose,
which considers the effects of the transverse shear strain and rotary inertia. A graded concept is
employed to allow the material property to vary gradually inside the elements. The proposed FGM
structures are characterized by two constituents (ceramic and metal) whose material properties are
dependent on the temperature and vary continuously throughout the thickness according to a power
law distribution proportional to the volume fraction of the constituents. Two different sets of power
law distribution are used to describe the volume fraction of the constituents, based on a single, or four
parameters. Based on a parametric analysis, we demonstrate the potentials of the proposed method
through its comparison with results available from the literature and by means of a convergence study.
Several numerical examples are further presented to investigate the effects of material compositions,
geometrical parameters, specified thermal loading and boundary conditions on the free vibration and
thermal buckling behavior of these structures. The effect of initial thermal stresses on the vibration
behavior is also investigated for plate and shell structures.
2016 Elsevier Ltd. All rights reserved.

1. Introduction
Structural elements such as plates, cylindrical panels and shells
are widely used in civil, marine and offshore engineering. In addition to the respect to fatigue requirements, these structures also
need to satisfy the dynamic and stability requirements [1]. The
application of coatings in marine and offshore engineering is
becoming even more popular in order to protect structures against
thermal environments, which can significantly influence the
mechanical properties of the materials and structures. Because of
severe service conditions and greater expectations on the life time,
thermal stability and durability of these structures are the major
concern for the industry [2]. Thus, it is crucial to understand the
vibration and thermal buckling characteristics in designing these
structures [3]. Due to such adverse operating conditions, the marine and offshore structures should have, in good balance, enough
Corresponding author.
E-mail address: francesco.tornabene@unibo.it (F. Tornabene).
http://dx.doi.org/10.1016/j.compstruct.2016.08.037
0263-8223/ 2016 Elsevier Ltd. All rights reserved.

thermo-mechanical strength to accommodate both thermal

induced loadings and general mechanical forces. Materials of these
structures should ideally exhibit low density, high strength, good
thermal and corrosion resistance and high toughness, etc. Hence,
new materials such as functionally graded materials (FGMs) are
being designed to meet such demands. The basic physics behind
the FGM concept is that the functionality of a particular material
system can be tailored by appropriately combining two or more
materials [4]. One example is represented by a ceramicmetal
FGM which can combine the optimum properties of ceramics and
metals, resulting in a desired property gradation in spatial
directions [5]. The ceramic provides the high thermal resistance
due to its low thermal conductivity [6], while the ductile metal
constituent minimizes fracture due to its greater toughness. Thus,
FGM structures are able to maintain the structural integrity while
reducing thermal stresses, residual stresses and stress concentrations. For the thermal buckling analysis of FGM plates, Javaheri
and Eslami [79] investigated the buckling of FGM plates subjected
to uniform in-plane compressive and thermal loads using a

208

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

variational approach based on the classical plate theory. Later on,

the same authors employed some equilibrium and stability
relationships to study the thermal buckling behavior of simply
supported FGM plates using a higher-order shear deformation theory (HSDT) [10]. In order to take into account the complete effects
of higher-order deformations, various higher-order plate theories
have been presented in the literature [1116]. Apart from higherorder theories, some researchers have analyzed thermal buckling
characteristics of FGM structures by different numerical
approaches. For example, the buckling behavior of functionally
graded (FG) plates under uniaxial mechanical and thermal loading
was investigated by Zhao et al. [17] using element-free kernel particle Ritz method. A simple and efficient formulation relied on the
framework of non-uniform rational B-spline (NURBS) based isogeometric analysis (IGA) was proposed by Tran et al. [18] for thermal
buckling of FGM plates. Zhang et al. [19] investigated the mechanical and thermal buckling behaviors of ceramicmetal FG plates by
using a local Kriging meshless method. A finite strip method (FSM)
was applied for analyzing the buckling behavior of rectangular FG
plates under thermal loadings by Ghannadpour et al. [20]. Very
recently, Ovesy et al. [21] analyzed and extended the application
of the finite strip method to the analysis of post-buckling behavior
of FG plates subjected to thermal loadings. The research on thermal
buckling of FG cylindrical shells was introduced by Shahsiah and
Eslami [22,23]. In their analysis the material properties were considered to be independent of the temperature. Shen [24] presented
the thermal postbuckling behavior of FG cylindrical shells made of
materials dependent on the temperature. He adopted a classical
shell theory with a von Karman-Donnell-type of kinematic nonlinearity. Similarly, the thermal buckling behavior of such shells was
studied with different numerical approaches. For example, Pourhamid et al. [25] examined the thermal analysis of the FGM cylinder
by super element methods. Yas and Aragh [26] studied the
three-dimensional steady-state response of a FG fiber reinforced
cylindrical panel. A differential quadrature method was employed
in [26] to solve three-dimensional thermo-elasticity equations.
Very recently, Shen and Xiang [27] investigated the thermal
post-buckling of cylindrical panels made of carbon nanotube
reinforced composite (CNTRC) resting on elastic foundations and
subjected to a uniform temperature rise. In the same work, the
HSDT was adopted with a von Karman-type of kinematic
nonlinearity. For the vibration characteristics of the FGM plates,
Yang and Shen [28] developed a differential quadrature-based
semi-analytical approach capable of incorporating general boundary conditions to examine the vibration and dynamic response of
FGM thin plates. Then, Yang et al. [29] extended their work to solve
the large amplitude vibration behavior of pre-stressed FGM
laminated plates by using Reddy HSDT. Woo et al. [30] provided
an analytical solution for the nonlinear free vibration behavior of
plates made of FGMs. For the vibration characteristics of FGM
shells, Najafizadeh and Isfandzibaei [31] studied the free vibration
of thin FGM cylindrical shells with a ring support, by using the Ritz
method based on the first-order and higher-order shear deformation shell theories. Loy et al. [32] investigated the vibration characteristics of a FG cylindrical shell composed of stainless steel and
nickel using the Love shell theory and RayleighRitz method. Wang
and Sheng [33] analyzed the vibration, buckling and dynamic stability of FGM cylindrical shells embedded in an elastic medium and
subjected to mechanical and thermal loads based on the first-order
shear deformation theory (FSDT). Tornabene et al. [34] studied the
dynamic behavior of FG conical, cylindrical and annular shell structures using the FSDT. Shen [35] carried out an investigation on
vibrational behavior of a FGM cylindrical shell of finite length,
embedded in a large outer elastic medium and in thermal environments based on a HSDT. Furthermore, Tornabene [36] studied the
dynamic behavior of moderately thick FG conical, cylindrical shells

and annular plates based on the FSDT using a four-parameter

power-law distribution. There were also other interesting studies
[3749] on the vibrational behavior of cylindrical shells. A lot of
studies on the thermal buckling and vibrational behavior of FGM
structures are available in the literature. However, most of them
are restricted to: one-parameter power-law distributions
[[7]-[33,35,37,42]-[50]], power-law indexes with one exponential
law exponent [51], one-parameter power-law indexes for 3D
FGM plate analysis [52], or one-parameter with two simple
power-law distributions [34]. Only a few studies have considered
four-parameter power-law functions to feature the constituents
of a volume fraction [36]. Additionally, thermal buckling and vibration studies of plates, cylindrical panels and shells, taking into
account the temperature-dependent material properties with
four-parameter power-law index, are rarely reported in the literature. In this work, we present a numerical method to study the
thermal buckling and vibration behavior of FGM structures including plates, cylindrical panels and shells in thermal environments.
The focus is on the thermal buckling and vibration characteristics
of moderately thick FGM structures made of ceramicmetal.First
we present the finite element formulations based on the FSDT. In
our formulations, we assume that the material properties are
dependent on the temperate, and we adopt three different
power-law distributions to model the ceramic volume fraction
based on four different parameters or a single parameter. Our formulation has been implemented in the ANSYS code, tested and
compared with existing methods on the thermal buckling and on
the vibration characteristics of FGM structures. Thus, we examine
the influence of the initial stresses due to temperature on the
vibration behavior of FGM plates, cylindrical panels and shells for
different volume fractions of the constituents. We also investigate
the effects of the material composition in terms of volume fraction,
geometrical parametric ratios, thermal loading and boundary conditions on the thermal buckling and vibration characteristics of
FGM structures.
2. Finite element formulation
In this section, we describe the finite element formulation proposed for the study of thermal buckling, free vibration and initially
stressed vibration behavior of the plates (Fig. 1a and b), cylindrical
panels and shells (Fig. 2ac). The study considers moderately thick
FGM structures, where the FSDT is used to account for the
transverse shear deformation and rotary inertia. The respective
coordinate system is established for each FGM structure, in which
z represents the thickness direction.
2.1. Moderately thick FGM plate element
The formulation presented here is an extension of a previous
work [42], by introducing the transverse shear deformation effect
for the FGM plate, within a FSDT. Let define the displacement field
as

8 9
>
<u >
=

8
9
>
< zwx >
=
v
v 0 zwy Hqe
>
>
>
: >
:
:
; >
;
; >
w
w0
0
8
9
>
< u0 >
=

where u0 ; v 0 and w0 are the middle surface displacements of the

plate in the x; y and z directions; wx and wy represent the transverse
normal rotations of the reference surface with respect to the y and x
axes, respectively, with

fqe g fu0

v0

w0

wx

wy gT ;

3
1 0 0 z 0
6
7
H 4 0 1 0 0 z 5
0 0 1 0 0

209

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

Fig. 1. (a) Schematic diagram of the FGM plate and (b) nodal displacements and grading scheme.

Fig. 2. (a) Schematic diagram of the FGM cylindrical panel, (b) nodal displacements and grading scheme and (c) schematic diagram of the FGM cylindrical shell.

The associated straindisplacement relationships of the plate

read

8
9
>
< ex >
=

ey

>
:c >
;
xy

8 @u
0
>
>
< @x

9
>
>
=

@v 0
@y

>
>
>
: @u0 @ v 0 >
;
@y

8 0 9
>
< ex >
=

@x

8
9
>
< kx >
=
feg e0y
z ky ;
>
>
>
>
: 0 ;
:
;
kxy
cxy

8 @wx
>
@x
>
< @w

9
>
>
=

@y
>
>
: @wx

>
;
@wy >

@y

cyz
cxz

( @w

@y

@w0
@x

wy
wx

)
;

feg fe0 g zfkg; fcg fc0 g;

A B
B

"

D

fe0 g
fkg

"


fNT g
fM T g

#
;

where fN T g and fMT g are the thermal stress and moment

resultants.
The transverse shear force resultant can be written as

fQ g ASij  fc0 g;

where feg0 is the mid-plane strain vector, fkg is the mid-plane curvature vector, and fc0 g is the transverse shear strain vector. The
membrane stress resultants fNg and the bending stress resultants
fMg for the FGM plate under thermal environments can be related
to fe0 g and fkg through the constitutive relations given by

8
9
>
< Nx >
=
Ny
Aij  fe0 g Bij  fkg  fNT g;
>
>
:
;
Nxy

3
)

c0yz
;
c0xz

fNg
fMg

@x

(
fcg

8
9
>
< Mx >
=
My
Bij  fe0 g Dij  fkg  fM T g;
>
>
:
;
M xy

The constitutive relations based on the Eqs. (6)(9) are

expressed as

3 2
32 0 3 2 T 3
A B 0
fNg
fe g
fN g
6
7 6 B D 0 76
7 6
7


fMg
fkg
4
5 4
5 4 fMT g 5;
54
fQg
fc0 g
0 0 AS 
f0g

10

where the matrices Aij ; Bij  and Dij  (i, j = 1,2,6) refer to the extensional, bending-extensional and bending stiffness coefficients.
These coefficients are defined as

fAij ;

Bij ;

Dij g

l Z
X
k1

hk

hk1

Q  f1;

z;

z2 gdz

11

210

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

Similarly, the matrix ASij  (i, j = 4,5) refers to the transverse

shear stiffness coefficients and is defined as

fASij g S

l Z
X

hk

Q  dz

12

hk1

k1

where Q  represents the generally isotropic lamina properties of

the kth layer, and S denotes the transverse shear correction coefficient assumed as [43]

6

m1 V f 1 m2 V f 2

:

section. The thermal stress resultant fN T g and thermal moment

resultant fM T g are expressed as
l Z
n
o
X
NT ; MT S

hk

hk1

Q fag Tz1; zdz

k

14

where Tz is the temperature rise and fagk is the vector of the coefficients of a linear thermal expansion in the kth layer for the FGM
plate consisting of a certain number of layers. The displacement
components are approximated by the product of the shape function
matrix Si  and the nodal displacement vector as follows

fqei g fu0i ; v 0i ; w0i ; wxi ; wyi gT ;

fqg

4
X
Si fqei g

15

i1

The superscript e of fqei g in Eq. (15), refers to the element where

the variable has to be determined. The procedure here adopted for
computing the structural stiffness matrix, mass matrix, geometric
stiffness matrix and thermal load vector for each element is given
in [42], which is based on a variational energy method, here not
repeated for the sake of brevity.

0
0

0
0

0
0

7
7
7
7
7
7
5

Q 66

Q 44

Q 55

19

For isotropic materials, the reduced stiffnesses Q ij are defined as

Q 11 Q 22

13

In Eq. (13), V f 1 and V f 2 denote the volume fraction and m1 and m2

denote the Poissons ratio of each material in the entire cross-

k1

Q 12
Q 22

Q 11
6
6 Q 21
6
fQg 6
60
6
40

Q 66

Ez; T
;
1  m2 z; T

Ez; T
;
21 mz; T

Q 12

mz; TEz; T
;
1  m2 z; T

20

Q 44 Q 55 Q 66 ;

21

where E is the Youngs modulus and m is the Poissons ratio. The displacement field considered in this study is based on the FSDT and is
represented in the following form

8 9
>
>
<u =

8
9
>
< zwx >
=
v
v 0  zwh ;
>
>
>
: >
:
:
; >
;
; >
w
w0
0
8
9
>
>
< u0 =

22

where u0 ; v 0 and w0 are the displacement components of the midsurface along the axial, circumferential, and radial directions,
respectively, while wx and wh refer to the slope in the x-z and h-z
planes. The kinematic relationships between strains and displacements for a cylindrical panel or shell are

8
9
8
9
@ 2 w0
>
>
@u0
>
>
2
>
>
>
>
@x
ex
>
>
>
>


<  @x
<
=
=
2

@
w
@
v
1
0
0
@
v
eh 1R @h0 w0  z R2 @h2  @h
;
>
> >
>
>
 2
>
:
>
>
: @ v 0 1 @u0  >
>
;
>
cxh ; >
>
>
: 1 2@ w0  @ v 0 ;
R @h
@x
8
>
<

9
>
=

cxz
chz

wx

wh

@w0
@x
@w
1
0
 vR0
R @h

@x@h

23

@x

)
:

24

The force and moment resultants can be computed as

2.2. Moderately thick FGM shell element

Consider the functionally graded cylindrical panel and shell

with an outer radius R0 , and inner radius Ri . The structure is
assumed to be moderately thick with thickness h and length L as
shown in Fig. 2. The x-axis is taken along the axial direction, the
circumferential arc length embrasses an angle h, and the z-axis is
directed radially inwards. The analysis considers a moderately
thick FGM cylindrical panel and shell, where the FSDT is used to
account for the transverse shear deformation [43]. In this study,
the thin shell element formulation developed by Stanley et al.
[44], has been extended to moderately thick elements by including
the effect of the transverse shear deformation. For a
moderately thick cylindrical panel or shell, the constitutive
relation is given by

frg fQ gfeg:

16

The stress vector, strain vector, and reduced stiffness matrix are
defined as

frgT frx
fegT fex

rh sxh sxz shz g;

eh cxh cxz chz g;

17
18

where rx and rh are the stresses in the x and h directions,

sxh is the shear stress on the x-h plane, ex and eh are the strains in
the x and h directions, and cxh is the shear strain in the x-h plane.
Similarly, sxz and shz are the transverse shear stresses and the
corresponding strains are cxz and chz . The reduced stiffness matrix
is defined as

fNx

Nh

fM x

Mh

Nxh g

h=2

h=2

Z
Mxh g

frx

h=2
h=2

rh sxh gdz

frx

25

rh sxh gzdz

26

Similarly, the transverse shear force resultant can be expressed

as

Z
fQ xz

Q hz g

h=2

h=2

Sfsxz

shz gdz

27

where S denotes the shear correction factor and is herein assumed

equal to 5/6 [34]. Substituting Eq. (16) into Eqs. (25)(27) and using
Eqs. (22)(24), the stress resultants are expressed in terms of the
mid-plane displacements as

8
9 2
Nx >
A11
>
>
>
>
>
6 A21
>
>
>
>
N
h
6
>
>
>
>
>
> 6
>
60
>
>
Nxh >
>
>
>
>
6
>
<M >
= 6B
x
6 11
6
>
6 B21
M
h >
>
>
>
>
6
>
>
>
>
60
>
>
Mxh >
>
6
>
>
>
>
6
>
>
>
>
40
Q
>
>
xz
>
>
:
;
0
Q hz

A12

B11

B12

A22

B21

B22

A66

B66

B12

D11

D12

B22

D21

D22

B66

D66

A66

A66

38
ex 9
>
>
>
>
>
7>
>
eh >
>
7>
>
>
>
>
7>
>c >
>
7>
>
xh >
>
>
7>
<k >
=
7>
x
7
7
>
7>
> kh >
>
7>
>
>
>
7>
>
>
7>
> kxh >
>
>
7>
>
>
>
5> cxz >
>
>
>
:
;

chz

28
where Aij ; Bij , and Dij (i, j = 1, 2 and 6) are the extensional, coupled
and bending stiffness, respectively. The stressstrain relation for a

211

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

generally isotropic material including the temperature effects based

on FSDT is given by

8
rx
>
>
>
>
>
>
< rh

9
>
>
>
>
>
>
=

Q 11

6
6 Q 21
6
sxh 6
60
>
>
>
>
6
>
>
s
>
>
xz
>
> 40
>
>
:
;
0
shz

Q 12

Q 22

Q 66

Q 44

91
9 8
308
ex >
a x DT >
>
>
>
>
>
>
>
>
>C
>
7B> e >
>
>
> >
0 7B>
a h DT >
>
>
>
h >
<
<
=C
=
7B
C
7
C
B
0 7B cxh  0
C
>
>
>
>
>
>
> >
7B>
C
>
>
>
cxz >
0
0 5@>
>
>
A
>
>
>
>
>
>
>
>
:
;
; :
chz
0
Q 55
0

29
where rx and rh are the normal stresses and sxh is the in-plane
shear stresses. ex and eh are the normal strains and cxh is the inplane shear strain. The coefficient of thermal expansion in the
two principal directions (x; h) is ax and ah . DT is the temperature
variation from a stress-free state. The displacement components
are approximated by the product of the shape function matrix Si 

and the nodal displacement vector: fu0i ; v 0i ; w0i ; wxi ; wyi gT . The
procedure for calculating the structural stiffness matrix and mass
matrix of each element includes the transverse shear deformation
as in [44]. All the details about the finite element formulation can
be found in [44]. By applying the main concepts of [42] for the plate,
we compute the geometric stiffness matrix and thermal load vector
for the cylindrical panels and shells for our study.

Eq. (32). On the other hand, the bottom surface of the structure is
rich of metal, whereas the top surface is rich of ceramic by setting
a = 1 and b = 0 in Eq. (33). One can get various power-law distribution profiles by modifying the parameters a; b; c and p. Additionally, a single-parameter power law distribution [42] is proposed
here to understand the influence of various power law functions
for the given constituents of volume fraction on the thermal buckling and vibration behavior
n

FGM3 : V c 2z h=2h ;

34

where n is the volume fraction index. For n 0, the volume fraction

of ceramic prevails and when the value of n is increased, the content
in metal increases in the FGM. Using the above equations, the effective Youngs modulus E, the mass density q, the coefficient of thermal expansion a and Poissons ratio m of the FGM plate, cylindrical
panel and shell read
FGM1
c p

E Ec  Em 1  a1=2 z=h b1=2 z=h  Em ;

35

q qc  qm 1  a1=2 z=h b1=2 z=hc  qm ;

p

36

a ac  am 1  a1=2 z=h b1=2 z=hc  am ;

37

m mc  mm 1  a1=2 z=h b1=2 z=hc  mm ;

p

38

c p

39

2.3. Analytical model of FGM properties

FGM2

In this analysis, it is assumed that the composition of the FG

plate, cylindrical panel and shell varies along the thickness direction. These structures consist of two constituents: ceramic and
metal. These constituents are graded through the thickness, from
one surface to the other and their compositions are shown in Figs. 1
(b) and 2(b). In this study, two four-parameter and one singleparameter power-law distributions are considered for the volume
fraction of ceramic. Since FGMs are mainly used in high temperature environments, we assume that the constituent materials have
temperature-dependent properties, which may be expressed as
[48]



P P0 P1 T 1 1 P1 T P 2 T 2 P3 T 3 ;

30

where P1 , P0 , P1 , P2 and P3 are constants in the cubic fitting relation

of the materials property. This expression allows the higher-order
effects of the temperature on the material properties to be readily
discernible, because the volume fraction is a spatial function,
whereas the properties of the constituents are functions of the temperature. The combination of these functions yields to the following
effective material properties of FGMs

Peff P c V c Pm V m 1;

31

where Peff is the effective material property of the FGMs, Pc and P m

are the temperature-dependent material properties of the ceramic
and metal, respectively, which are related by: V c + V m = 1. The properties of the FG plate, cylindrical panel and shell are assumed to
vary throughout the thickness, according to the following power
law [36]

E Ec  Em 1  a1=2  z=h b1=2  z=h  Em ;

p

q qc  qm 1  a1=2  z=h b1=2  z=hc  qm ;

40

a ac  am 1  a1=2  z=h b1=2  z=hc  am ;

41

m mc  mm 1  a1=2  z=h b1=2  z=hc  mm ;

42

FGM3
n

E Ec  Em 2z h=2h Em ;

43

q qc  qm 2z h=2hn qm ;

44
p

a ac  am 1  a1=2 z=h b1=2 z=hc  am ;

45

m mc  mm 2z h=2hn mm ;

46

2.4. Temperature across the thickness

A steady-state one-dimensional heat conduction analysis is carried out to evaluate the temperature distribution across the thickness of the FGM plate, cylindrical panel and shell based on some
specified boundary conditions in temperature. This is essential
for the material properties which vary with the temperature. The
convective and radiation heat conduction are neglected. The governing differential equation for the steady-state heat conduction
along the thickness direction is



d
dTz
Kz
0
dz
dz

c p

32

c p

33

where K is the effective thermal conductivity of the material [42]

which varies as
FGM1

FGM1a=b=c=p : V c 1  a1=2 z=h b1=2 z=h  ;

FGM2a=b=c=p : V c 1  a1=2  z=h b1=2  z=h  ;

where p is the power-law index varying as 0 6 p 6 1, and the

parameters a; b and c define the material variation profile through
the thickness. For example, the material distribution in the FGM
plate, cylindrical panel and shell is continuously varied such that
the bottom surface (-h/2) of the structure is rich of ceramic, whereas
the top surface (+h/2) is rich of metal, by setting a 1 and b 0 in

47

c p

48

c p

49

K K c  K m 1  a1=2 z=h b1=2 z=h  K c ;

FGM2

K K c  K m 1  a1=2  z=h b1=2  z=h  K c ;

212

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

FGM3
n

K K c  K m 2z h=2h K m ;

50

The temperature-based boundary conditions are specified by

enforcing the temperature on the top surface T top , and the temperature on the bottom surface T bottom . The ambient temperature is
normally considered, unless other temperature values are prescribed. The finite element formulation and its complete procedure
are explained in [42] and here not repeated for brevity.
2.5. Modeling of moderately thick FGM plate, cylindrical panel and
shell in ANSYS
Some commercial finite element softwares such as ANSYS,
ABAQUS etc. can be used to model and analyze the FGM structures,
see i.e. [50,51]. However, there are a few challenges which include
the assignment of the homogenized material properties to elements according to the power-law distributions and the modeling
of the actual material behavior of FGMs directly in the software
[25]. Moreover, it is usually complex to compute the dependence
of the properties of FGM structures on the temperature within each
single layer. Here we simulate numerically the FGM structures in
the ANSYS code by implementing a user-subroutine based on the
ANSYS Parametric Design Language (APDL). 8-node 3D elements
(SOLSH190, [53]) are used to analyze the FGM structures as shown
in Fig. 3.
The material properties vary though the thickness of the structure and can be described with the above mentioned fourparameters or single-parameter power law distributions. Similarly,
the temperature distribution across the thickness of the FGM plate,
cylindrical panel and shell based on some fixed temperature
boundary conditions in ANSYS is evaluated as described in Section 2.4. More details can be found in [42].
2.6. Solution procedure
The linear buckling technique allows one to obtain the critical
buckling load and the corresponding deformed shape of the modeled structure. To obtain this result, the condition of neutral equilibrium between external loads and internal reactions is searched
by solving the equation: Kfqg K s  kK g fqg 0; where K
is the total stiffness matrix, K s  is the stiffness matrix calculated
in small-displacement ranges, K g  is the geometric stiffness matrix
corresponding to a reference load, k is the eigenvalue, (i.e., the load
factor multiplier of the reference load which leads to the critical
value), and q is the displacement vector. In order to analyze the linear thermal buckling behavior, the following eigenvalue problem is
considered: jK s  kK g j 0. The buckling eigenvalues and buckling modal shapes are computed using an iteration technique.
The product of k and the initial guess value DT gives the critical
buckling temperature T cr . Similarly, the natural frequencies are
obtained by using an iteration technique for the eigenvalue
problem of the form: jK  x2 Mj 0. By defining the structural
stiffness, we can express the initial stress due to the thermal load,
and the mass matrices [54]. Thus, the governing equations for the
free vibration and stability of the FGM plate, cylindrical panel and
shell in the presence of the initial stress reads:
jK K g   x2 Mjfqg 0, where K is the global stiffness matrix,
K g  is the global geometric stiffness matrix, M is the global mass
matrix, fqg is the vector of nodal degrees of freedom and x is the
frequency. The method of solution can be described as follows.
First, we compute the initial stress for a given loading by using
the nodal load vector. Second, we define the geometric stiffness
matrix by using the initial stresses obtained from the first step.
Third, we obtain the mass matrix in a consistent way. Finally, we

Fig. 3. A typical FGM (a) plate, (b) cylindrical panel and (c) shell.

perform the eigenvalue analysis to determine the frequency of

vibration under the initial stress.
3. Numerical results and discussion
In this section, we perform a convergence study and verify the
accuracy of the method before proceeding with a detailed numerical investigation. Thus, several numerical examples for the thermal buckling, free vibration and thermally stressed vibration
behavior of FGM plate, cylindrical panel and shell are carried out,
accordingly. The computed results are then discussed in detail.
The material properties of each constituent depend on the temperature in the present numerical study, including the Youngs modulus, Poissons ratio, thermal expansion coefficient, and thermal
conductivity, as given in Table 1.
3.1. Convergence and comparison of results
First, a convergence study is performed to determine the finite
element mesh size and a suitable number of homogeneous layers
to represent the FGM plate, cylindrical panel and shell. Two cases
are herein chosen for the convergence study: the thermal buckling
of a FGM plate and the free vibration of a FGM cylindrical shell.
Similarly, we consider two different cases to validate our method
and formulation: the thermal buckling of a FGM plate/cylindrical
shell and the free vibration of a FGM cylindrical panel/shell. Finally,
the numerical results obtained with the present method are compared with those available from the literature.

213

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

Table 1
Temperature-dependent material properties of ceramic and metal.
Material properties (adapted from Ramkumar et al. [42]).
Materials

P 1

P0

P1

P2

P3

Zirconia
Al2 O3
Ti-6Al-4V

0
0
0

244.27e9
349.55e9
122.56e9

1.317e3
3.853e4
4.586e3

1.214e6
4.027e7
0

3.681e10
1.673e10
0

Zirconia
Al2 O3
Ti-6Al-4V

0
0
0

0.2882
0.2600
0.2884

1.133e4
0
1.121e4

0
0
0

0
0
0

qkg=m3

Zirconia
Al2 O3
Ti-6Al-4V

0
0
0

5700.0
3750.0
4429.0

0
0
0

0
0
0

0
0
0

a1=k

Zirconia
Al2 O3
Ti-6Al-4V

0
0
0

12.766e6
6.826e6
7.578e6

1.491e3
1.838e4
6.638e4

1.006e5
0
3.147e6

6.778e11
0
0

KW=mK

Zirconia
Al2 O3
Ti-6Al-4V

0
1123.6
0

1.70
14.087
1.0

1.276e4
6.227e3
1.704e2

6.648e-8
0
0

0
0
0

EN=m2

3.1.1. Convergence study

3.1.1.1. Thermal buckling of FGM plate. A FGM plate made of
Al/Alumina [17] with a1 /b1 = 1 and a1 =h 100 is here chosen
for the convergence study. This plate is clamped at all the edges
(here labeled as CCCC) by constraining all the degrees of freedom
(DOFs), and it is subjected to a uniform temperature rise. The
FGM of the type FGM3 is here selected for the purpose. Five models
with different finite element mesh density (i.e. 10  10; 20 
20; 30  30; 40  40 and 50  50) are considered for the
convergence study, together with five layer-models (i.e.
5; 10; 15; 20 and 25 layers). The converged lowest critical buckling temperatures are listed in Tables 2 and 3 along with the
related element size and layer, respectively. As visible from Tables
2 and 3, the results remain almost unchanged for meshes with a
number of elements above 40  40 elements and a number of
layers greater than 20. This suggests to use a mesh with 40  40
elements and 20 layers across the thickness, for conservative
reasons.
3.1.1.2. Free vibration of FGM cylindrical shell. Now we consider a
FGM cylindrical shell (as shown in Fig. 3c), made of Al/Zirconia
[36], with L 2 m; h 0:1 m and R 1 m. This cylindrical shell
is clamped at one end and maintained free on the other end (here
labeled as CF shell). The FGM3 -type is considered once again for the
convergence analysis, while adopting four different finite element
discretizations (i.e. 20  20, 30  30, 40  40 and 50  50)
together with four layer-models (i.e. 5; 10; 15 and 20 layers).
The first natural frequency is reported in Tables 4 and 5 for a
power-law index n 1. Based on the results, it is reasonable to

Table 2
Convergence study for varying mesh refinements of a FGM Al/Alumina plate.
Power law index

Critical buckling temperature rise (DT cr )

10  10

20  20

30  30

40  40

50  50

n=1

22.74

21.66

21.52

21.43

21.43

Table 3
Convergence study for varying number of layers of a FGM Al/Alumina plate.
Power law index

n=1

Critical buckling temperature rise (DT cr )

5 layers

10 layers

15 layers

20 layers

25 layers

21.56

21.41

21.38

21.37

21.37

Table 4
Convergence study for varying size of elements of a FGM Al/Zirconia cylindrical shell.
Power law index

Lowest natural frequency (Hz)

20

30

40

50

n=1

150.22

150.20

150.19

150.19

Table 5
Convergence study for varying number of layers of a FGM Al/Zirconia cylindrical shell.
Power law index

n=1

Lowest natural frequency (Hz)

5 layers

10 layers

15 layers

20 layers

150.20

150.19

150.18

150.18

use a mesh of 40  40 elements and 15 layers across the thickness

for the analysis.
3.1.2. Validation study
3.1.2.1. Thermal buckling of FGM plate and cylindrical shell. In this
section, we analyse two cases in order to validate the proposed
method. First, we compare the buckling result based on the present
moderately thick FGM plate element with the results found by
Zhao et al. [17]. The FGM plates considered for validation are
characterized by: a1 /b1 = 1, 20x20 elements, mixed material
Al/Alumina of the type FGM3 ; a1 /h = 50 and a1 =h 100, a double
boundary conditions, i.e. simply supported edges (SSSS) or
clamped edges (CCCC). These plates are subjected to a uniform
temperature rise. Table 6 shows, comparatively, the thermal buckling temperatures given by the present method and those found in
[17]. From Table 6, a reasonably good agreement between the
results can be oberved, which validate the potentials of the proposed approach.
Second, to ensuring the accuracy of the proposed method, we
compare our results in terms of thermal buckling of FGM cylindrical shells made of material properties dependent on the temperature, with results from Shariyat et al. [49] and Shen [50]. A FG
cylindrical shell made of Si3N4/SUS304 and subjected to a uniform
temperature rise is herein considered. The shell geometry is characterized by R=h = 400, L2 =Rh 300 and h = 1 mm, whereas the
material properties are assumed to be nonlinear functions of the
temperature, as reported in [49]. The FGM type is taken as FGM3 .
The main results for the thermal buckling of FG shells are given

214

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

Table 6
Comparative critical buckling temperature rise (DT cr ) for an Al/Alumina FGM plate.
n0

n 0:5

n1

n2

n5

BCs

a=h

Present

[17]

Present

[17]

Present

[17]

Present

[17]

Present

[17]

SSSS

50
100

70.1
17.5

67.9
17.3

39.7
9.9

38.2
9.8

32.3
8.0

31.1
7.9

28.3
7.1

29.6
6.9

28.7
7.2

29.6
7.4

CCCC

50
100

186.3
46.8

175.8
44.1

105.7
26.5

99.1
24.8

86.0
21.5

82.3
20.7

75.4
18.9

71.0
18.4

76.4
19.2

74.56
19.1

Table 7
Comparative results in terms of buckling temperature rise (DT cr ) with respect to
power-law index for FGM cylindrical shells.
Power law
index, n

0.0
0.5
1
2
5

Critical buckling temperature rise (DT cr )

TID

TD

Present

[49]

[50]

Present

[49]

[50]

80.2
94.2
99.2
108.2
125.4

36.4
81.6
88.8
117.2
143.3

86.6
106.5
118.4
132.4
150.2

78.9
87.1
92.5
101.1
116.0

32.3
58.6
79.4
97.0
125.6

82.2
99.8
110.2
122.2
137.1

ation of the natural frequencies by Tornabene [36] and those

obtained from the present method. In addition to the proposed
finite element method, the present free vibration results are compared with those ones obtained with the ANSYS code by applying
our user-subroutines. A clamped-free FGM cylindrical shell with
the same geometry parameters is also considered. The main results
obtained with our method in terms of the lowest natural frequency, are compared to those ones given by ANSYS, as plotted
in Fig. 4. From Tables 8, 9 and from Fig. 4, it is worth noticing a very
good agreement between all the results.
3.2. Parametric studies

in Table 7, where a good agreeement is observed between the

results provided by the proposed method and those ones from
[50] for temperature-independent (TID) and temperaturedependent (TD) material properties. Only some discrepancies arise
between our results (based on FSDT) and those ones of [49]. This
could be related to the presence of the shear deformations which
were neglected in [49].
3.1.2.2. Free vibration of FGM cylindrical panel and shell. Two examples are now chosen to verify the proposed method. First, the FGM
cylindrical panel has L 2 m, h = 0.1 m, R = 1 m, h 1200 ; with a
combined Al/Zirconia as in [36]. We verify the free vibration characteristics for the CFFF (one end at the axial direction is clamped
and the other three edges are free) boundary conditions. The
FGM here considered is of the type FGM2a=b=c=p . Table 8 compares
the natural frequencies by Tornabene [36] and those ones computed from the present method.
Second, a clamped-free FGM cylinder is considered with the
same geometry parameters as previously assumed, in order to
verify the natural frequencies. Table 9 shows a comparative evalu-

3.2.1. Thermal buckling of FGM plate

3.2.1.1. Influence of a1 =h ratio and boundary conditions. This section
deals with the effect of the span-to-thickness ratio a1 =h, (where a1
remains constant and h varies) on the critical buckling temperature
rise of a FGM plate clamped (CCCC) or simply supported (SSSS) at
each side. The FGM plate is subjected to a uniform temperature rise
(UTR), and is made of a mixed material aluminum-alumina [17],
i.e. a1 /b1 1, and the FGM is of the type FGM2a=b=c=p
(a 1; b 0:5; c 2; p varying). The variation of the critical
buckling temperature rise with respect to the a1 =h ratio for the
two boundary conditions (CCCC and SSSS) is computed and shown
in Figs. 5 and 6, respectively. It is obvious that as the plate thickness decreases, the critical buckling temperature rise decreases.
Based on a comparison between Figs. 5 and 6, it is also clear that
the buckling temperature rise for a CCCC plate is much higher than
that related to the SSSS plate, as expectable. The buckling temperature also decreases for increasing values of the power-law index
from 0 up to 10. This is due to the fact that for large values of
the power-law index, the material properties of the plate become
quite similar to the material properties of a metal with higher
thermal expansion coefficients in comparison to the ceramic,

Table 8
Comparative results in terms of natural frequency of FG cylindrical panel (CFFF) as a function of the power-law index. a 0:8; b 0:2; c 3.
FGM type

FGM 2

Mode no.

1
2
3
4

p0

p1

p5

p 50

Present

[36]

Present

[36]

Present

[36]

Present

[36]

61.5
95.9
151.2
243.0

61.0
94.8
153.1
241.4

61.0
94.7
151.2
241.7

60.0
93.4
150.4
237.6

61.5
95.9
151.2
243.0

61.0
94.8
153.1
241.4

61.0
94.7
151.2
241.7

60.0
93.4
150.4
237.6

Table 9
Comparative results in terms of natural frequency of FG cylinder (CF) as a function of the power-law index. a 1; b 0:5; c 2.
FGM type

FGM 2

Mode No.

1
2
3
4

p0

p1

p5

p 50

Present

[36]

Present

[36]

Present

[36]

Present

[36]

151.79
151.79
220.34
220.34

152.93
152.93
220.06
220.06

150.46
150.46
220.39
220.39

151.10
151.10
217.83
217.83

147.16
147.16
219.12
219.12

147.58
147.58
215.22
215.22

142.84
142.84
208.79
208.79

144.55
144.55
210.43
210.43

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

Fig. 4. Comparative results in terms of the lowest natural frequency of a CF FG

cylindrical shell as a function of the power-law index, FGM 2a=b=c=p a 1; b 0:5;
c 2.

Fig. 5. Variation of the critical buckling temperature rise with a1 =h ratio for CCCC
boundary conditions.

215

Fig. 6. Variation of the critical buckling temperature rise with a1 =h ratio for SSSS
boundary conditions.

Fig. 7. Variation of the critical buckling temperature rise with respect to power-law
index for 3 different FGM types (a 1; b 0:5; c 2, SSSS boundary conditions).

independently of the boundary conditions. Thus, one could easily

control the critical buckling temperature of the FG plate by tuning
the volume fraction.
3.2.1.2. Influence of a1 =h ratio and boundary conditions. In order to
understand the thermal buckling behavior variation with respect
to the ceramic volume fraction V c and its type, additional studies
have been carried out on the aluminum-alumina [17] FGM plate.
Three different FGM types are chosen for the study, i.e., FGM1 ,
FGM2 and FGM3 . Moreover, a1 /b1 and a1 =h are set to 1 and 50,
respectively. For FGM1 and FGM2 , the following parameters are
selected, i.e. a = 1;b = 0.5;c = 2, with p varying. For the case of
FGM2 , the selected parameters yield to a mixed ceramic/metal
material in the ratio 50:50 at the bottom surface, while the top surface is made of ceramic. The opposite condition is reached by the
FGM1 with the same parameters. We repeat the analysis for a
square plate with all sides clamped (CCCC) or simply-supported
(SSSS), and subjected to a uniform temperature rise. The computed
critical buckling temperature rise for various volume fractions and
boundary conditions (CCCC and SSSS) is shown in Figs. 7 and 8.
Based on the results, it is found that, the critical buckling

Fig. 8. Variation of the critical buckling temperature rise with respect to power-law
index for 3 different FGM types (a 1; b 0:5; c 2, CCCC boundary conditions).

216

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

Fig. 9. Variation of the lowest natural frequencies with a1 =h ratio for CCCC
boundary conditions a 1; b 0:5; c 2.

Fig. 11. Variation of the lowest natural frequencies with the critical buckling
temperature rise for CCCC boundary conditions a 1; b 0:5; c 2.

temperature rise falls down as the power-law index increases,

until reaching its minimum value for all the considered FGM types.
However, the FGM3 type exhibits a lower critical buckling temperature rise than the other cases, in the range of a power-law index
within 0.2 and 5. FGM1 and FGM2 almost maintain the same trend
with a negligible variation for the critical buckling temperature
with respect to the power-law index. Both of them reach a higher
critical buckling temperature than that provided by the FGM3 . All
the plots end up at the same value for all the FGM types when
the power-law index is equal to 0 and 10 (homogeneous isotropic
material). This is expected because the parameters a; b; c and p in
the FGM influences the volume fraction of the constituents. This
increases much more the ceramic content in the range of the
power-law index between 0.2 and 5.
3.2.2. Free vibration study of FGM plate
3.2.2.1. Influence of a1 =h ratio and boundary conditions. The effect of
the ratio of a1 =h (where a1 is maintained unaltered and h varies) on
the lowest natural frequency is studied in this section for CCCC and
SSSS square FGM plates made of aluminum-alumina [17]. The FGM
type adopted is FGM2 . The natural frequency of the FGM plate for
various a1 =h ratios is shown in Figs. 9 and 10. Based on the results,
it can be noticed that the lowest natural frequency decreases up to

Fig. 10. Variation of the lowest natural frequencies with a1 =h ratio for SSSS
boundary conditions a 1; b 0:5; c 2.

Fig. 12. Variation of the lowest natural frequencies with the R=h ratio for CFCF
boundary conditions a 0:8; b 0:2; c 3.

a constant value for high values of a1 =h. In addition, the value of

the lowest natural frequency decreases when the BCs changes from
CCCC to SSSS, as expectable.
3.2.2.2. Effect of temperature on frequency behavior of FGM plate. In
order to give a clear scenario about the influence of the initial
thermal stress on the vibration, additional studies have been performed on the aluminum-alumina square FGM plate. The material
properties are dependent on the temperature, as listed in Table 1.
The ratio a1 =h is maintained equal to 50. The evaluation of the natural frequencies is performed by applying a thermal pre-stress
based on the assumption that the inner and outer surfaces are at
uniform temperature as already explained in Section 2.6. The natural frequency computed for the CCCC FGM plate with a thermal
pre-stress is shown in Fig. 11 for various volume fractions. Based
on these plots, it is clearly shown that the natural frequencies
decrease with an increase in temperature and they become lower
for higher values of the power law index. Furthermore, it is clear
that, the lower the power law index, the higher the buckling temperature. Obviously, this depends on the constituent materials
characterizing the FGM. Further, the characteristic variation of
the natural frequency with respect to the temperature depends

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

217

on the modal numbers. For example, as previously noticed in Section 3.2.1, the critical buckling temperature rise (DT cr ) is equal to
199 for a FGM plate with a1 =h 50, an aluminum-alumina mixed
material, and p 0. For the material, BCs, geometry and power law
index, the lowest natural frequency is calculated as 308.4 Hz (as
given in Section 3.2.1.1). From Fig. 11, it is clearly seen that the natural frequency of the FGM plate drops to its minimum value when
the critical buckling temperature rise reaches the value 199.
3.2.3. Free vibration of FGM cylindrical panel
3.2.3.1. Influence of R/h ratio and boundary conditions. The FGM
cylindrical panel here analyzed has a length of 2 m, radius of
1 m, h 1200 , and is made of Al/Zirconia [36]. The influence of
the R=h ratio and boundary conditions is evaluated on the vibration
characteristics. The FGM type is assumed as FGM2a=b=c=p . Figs. 12
and 13 plot the variations of the lowest frequencies of the FGM
cylindrical panel with respect to the R=h for a CFCF boundary condition (two ends at the axial directions are clamped and the other
two ones are free) and a CCCC boundary condition, respectively.
Between the two boundary conditions here considered, the CCCC
cylindrical panels seem to have an higher fundamental frequency
than the CFCF one.
Furthermore, it can be seen from Figs. 12 and 13 that the lowest
natural frequency of the FGM cylindrical panel decreases as the
ratio R=h increases for both the boundary conditions. The reason
of this phenomenon is due to the reduced bending stiffness of
the panel. It is also shown that the natural boundary conditions,
such as the stress resultants, affect the frequencies of the FGM
cylindrical panel than geometrical (essential) boundary conditions
especially for a CFCF boundary condition. As also expected, Figs. 14
and 15 show that the frequencies increase with respect to its
modal numbers, independently of the boundary conditions and
the volume fraction of the FGM constituents. However, it can be
seen that the boundary conditions and power-law index have some
effects on the modal behavior. For example, in Fig. 15, it is worth
noticing a decay in the frequency of the CCCC panel for mode 3
and increasing values of p. This is less evident in Fig. 14 for the case
of a CFCF panel.
3.2.3.2. Influence of the ceramic volume fraction V c and its type on the
thermally pre-stressed vibration behavior. A FGM cylinder with the
same geometry parameters as in the previous section, is now considered to study the influence of the ceramic volume fraction V c
and its type (case-1:FGM2 and case-2: FGM3 ) on the thermally

Fig. 13. Variation of the lowest natural frequencies with the R=h ratio for CCCC
boundary conditions a 0:8; b 0:2; c 3.

Fig. 14. Variation of the natural frequencies R=h 30 with the mode number for
CFCF boundary conditions a 0:8; b 0:2; c 3.

Fig. 15. Variation of the natural frequencies R=h 30 with the mode number for
CCCC boundary conditions a 0:8; b 0:2; c 3.

pre-stressed vibration behavior of cylindrical panels. The results

for a CCCC boundary condition under a non-linear thermal loading
(NLTD) are given in Fig. 16. The material properties are dependent
on the temperature, as shown in Table 1. The additional parameters are considered, i.e. a 0:8; b 0:2; c 3; and a varying p.
The temperature at top and bottom sides is maintained equal to
1000 K and to the ambient temperature, respectively, for both
cases. In Fig. 16, the natural frequencies are depicted with respect
to different values of temperature rise for varying power law
indexes.
It is clearly visible from Fig. 16, that the natural frequencies
decrease with an increase in temperature for both cases (cases-1
and 2). However, there is not a regular variation in frequencies
with respect to the temperature when we vary the thermal load
in terms of temperature, power law index and FGM composition
in a combined sense. In other words, a similar trend is observed
for case-1 and 2 when p 0 or p 1000, since it behaves as an isotropic material. For cases-1 and 2 with p 0:5 or p 1, there is a
different decrease of the lowest frequency value for increasing
temperatures. This is due to the fact that for case-1 wih p 0:5
or p 1, the cylindrical panel has an higher coefficient of thermal
expansion compared to the case-2. Though we have studied the

218

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

Fig. 16. Variation of the lowest natural frequencies R=h 30 with the critical
buckling temperature rise for CCCC boundary conditions. a 0:8; b 0:2;
c 3; p varying).

Fig. 18. Variation of the lowest natural frequencies with the R=h ratio for the CC
boundary conditions.

thermal buckling of the cylindrical panel with different R=h ratios,

we comment only about the results of the critical buckling temperature rise obtained for R=h 30, for the sake of brevity.
3.2.4. Thermal buckling of FGM cylindrical shell
3.2.4.1. Influence of R=h ratio. We now compute the thermal
buckling temperature for FGM cylindrical shells clamped at the
bases, for varying R=h ratios where R remains constant and h is
varied. The FGM cylindrical shell has a length of
L 2 m; h 0:1 m; R 1 m and is made of Al/Zirconia [36]. The
FGM type here considered is FGM2 , where a; b and c are chosen
as a 1; b 0:5; c 2; p varying. The variation of the critical
buckling temperature with R=h is shown in Fig. 17, where it is
visible that a FGM cylindrical shell with a large thickness ratio
undergoes a large critical buckling temperature rise. Fig. 17 also
demonstrates that the buckling temperature rise decreases as in
the FGM case, when the power law index, p changes from 0 to 100.
3.2.5. Free vibration of FGM cylindrical shell
3.2.5.1. Influence of R=h ratio and mode characteristics on the
frequency behavior. The FGM cylindrical shell considered here, is
characterized by L 2 m; h 0:1 m; R 1 m, a mixed Al/Zirconia

Fig. 17. Variation of the critical buckling temperature rise with the R=h ratio for CC
boundary conditions.

Fig. 19. Frequencies associated with the first 8 modes for the CC boundary
conditions, R=h 10.

[36]; whereas a CC boundary condition is applied to the bases of

the shell. The FGM type herein considered is the FGM2 . Fig. 18
shows the lowest natural frequencies as obtained for different
material distributions and varying R=h ratios. From Fig. 18, it can
be observed that the lowest natural frequency decreases as the
R=h ratio increases for each different power-law index. Fig. 19
demonstrates that the frequencies increase for increasing mode
numbers, with a slight decreasing trend as the power-law index
increases.
Based on the studies previously performed, we can obtain the
modal shapes for each type of FGM structures. For brevity, we represent here only the modal shapes of the FGM cylindrical shell with
a CC boundary condition, R=h 10, p = 10 (see Fig. 20).
3.2.5.2. Effect of temperature on frequency behavior of FGM cylindrical
shell. In order to investigate the behavior of the natural frequency
variation with respect to the temperature, additional studies have
been carried out on a Al/Zirconia FGM cylindrical shell (as listed in
Table 1). The geometrical parameters are the same as in the
previous section and the FGM type here chosen is FGM 2
(a 1; b 0:5; c 2; p varying), whereas R=h 10. The same
process already described in Section 3.2.2.2 is now repeated for
an inner and outer surfaces maintained at a uniform temperature,

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

219

Fig. 22. Comparison (UTR and NLTD) of the lowest natural frequencies with respect
to the critical buckling temperature rise of FGM cylinder under CC boundary
conditions.

Fig. 20. Selected modal shapes of the FGM cylindrical shell with R=h 10 for CC
boundary conditions.

of the thermal boundary conditions. Fig. 22 shows a comparison of

the lowest natural frequency with respect to the critical buckling
temperature rise across the thickness for a UTR distribution and
nonlinear temperature distribution (NLTD) pre-stressed FGM cylinder. From Fig. 22 it can be seen that the frequency decreases with an
increasiung temperature, whereas the frequency decreases when
the temperature applied is changed from UTR to NLTD.
4. Summary and concluding remarks

Fig. 21. Variation of the lowest natural frequencies with the critical buckling
temperature rise for CC boundary conditions.

for the evaluation of the natural frequencies of the thermally

stressed FGM cylindrical shell. Fig. 21 shows the natural frequency
of the FGM cylindrical shell constrained at both the ends, and subjected to a thermal pre-stress for various volume fractions. It can
be seen from Fig. 21 that the lowest frequency decreases with
increasing values of the initial rise in temperature. It is evident that
the bending strain energy dominates the meridional stress
resultants due to the temperature rise. However, for higher temperatures close to the buckling temperature, the magnitude of
the meridional stress resultants become higher than the bending
strain energy and causes the frequencies to fall drastically, similarly to the FGM plate case.
In addition, the study based on a uniform temperature (UTR) distribution is extended to induce a free vibration with a fixed
temperature on the top surface equal to 1000 K and the ambient
temperature at the bottom surface, in order to understand the effect

In this paper, a method based on the FSDT has been proposed

for analyzing the thermal buckling and vibration characteristics
of plates, cylindrical panels and shells made of FGMs with two
power law distributions based on four parameters and one power
law distribution based on a single parameter. We also present a
numerical framework which can consider the pre-stress due to
the temperature environment by solving a steady state heat conduction equation. The material properties are assumed to be
dependent on the temperature, and vary continuously through
the thickness according to a power law distribution in terms of
the volume fraction of the constituents. Several numerical examples are provided to verify the accuracy of the present method with
the solutions available from the literature, with a good agreement
between them. We also simulate the FGM cylindrical shell in the
ANSYS code by implementing a user-subroutine based on ANSYS
Parametric Design Language (APDL). The good matching between
our results and those ones from ANSYS, confirms the potentials
of our method. Thus, we examine the effect of the material composition, geometrical parametric ratio, temperature field, power-law
index and temperature-dependence of the material property, on
the vibration characteristics. Based on this parametric study, it is
found that these parameters have a significant effect on the vibration behavior and thermal buckling characteristics of these FGM
structures. The effects can be summarized as follows.
For the case of FGM plate we have the following considerations.
 The FGM plates with a larger thickness ratio undergo a higher
critical buckling temperature rise and the buckling temperature
rise of a clamped plate is much greater than that of a simply
supported plate.
 The buckling critical temperature reaches the maximum value
in fully ceramic plates and decreases with an increasing power
index p because of the material degradation due to the enrichment of the metal constituent.

220

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

 The FGM type in terms of graded index has a significant effect

on the critical buckling temperature.
 The critical thermal gradient T cr of the FGM plates under a nonlinear temperature distribution is greater than that under a uniform one.
 When the FGM plate reaches the critical buckling temperature
rise, the natural frequency decreases up to a minimum value.
For the case of FGM cylinder we can conclude as follows.
 The FGM cylindrical panel/ shell with a larger thickness ratio
undergoes a higher critical buckling temperature rise.
 The frequency decreases with increasing temperatures, whereas
the frequency decreases when the applied temperature changes
from UTR to NLTD.
 The temperature effect is more significant on the pre-stressed
vibration behavior of a cylindrical panel when the material
composition is swapped from a prevailing ceramic to a prevailing metal and vice versa, if compared to the pure free vibration.
 The frequency increases with respect to the modal numbers.
However, as the power-law index increases, the frequencies
decreases. The lowest natural frequency decreases as R/h ratio
increases for different power-law indexes.

References
[1] Tornabene F, Viola E. 2-D solution for free vibrations of parabolic shells using
generalized differential quadrature method. Eur J Mech A/Solids
2008;27:100125.
[2] Le Gac PY, Choqueuse D, Paris M, Recher G, Zimmer C, Melot D. Durability of
polydicyclopentadiene under high temperature, high pressure and seawater
(offshore oil production conditions). Polym Degrad Stab 2013;98:80917.
[3] Almeida CA, Romero JC, Menezes IFM, Paulino GH. On the modeling of
structural dynamics of risers composed of functionally graded materials. In:
Proceedings of the XIV international symposium on dynamic problems of
mechanics (DINAME 2011). 13th 18th March, Brazil; 2011.
[4] Giunta G, Belouettar S. Hierarchical theories for a linearized stability analysis
of FGM beams. In: 52nd AIAA/ASME/ASCE/AHS/ASC Structures, Structural
Dynamics and Materials Conference. 4-7 April. Denver, Colorado; 2011.
[5] Lee WY, Stinton DP, Berndt CC, Erdogan F, Lee Y, Mutasim Z. Concept of
functionally graded materials for advanced thermal barrier coating
applications. J Am Ceram Soc 1996;79(12):300312.
[6] Hasselman DPH, Youngblood GE. Enhanced thermal stress resistance of
structural ceramics with thermal conductivity gradient. J Am Ceram Soc
1978;61(12):4952.
[7] Javaheri R, Eslami MR. Buckling of functionally graded plates subjected to
uniform temperature rise. In: Proceeding of the 4th International Congress on
Thermal Stresses. June 811. Osaka. Japan; 2001.
[8] Javaheri R, Eslami MR. Thermal buckling of functionally graded plates. AIAA J
2002;40(1):1629.
[9] Javaheri R, Eslami MR. Buckling of functionally graded plates under in-plane
compressive loading. ZAMM-J Appl Math Mech 2002;82(4):27783.
[10] Javaheri R, Eslami MR. Thermal buckling of functionally graded plates based on
higher order theory. J Therm Stresses 2002;25:60325.
[11] Croce LD, Venini P. Finite elements for functionally graded ReissnerMindlin
plates. Comput Methods Appl Mech Eng 2004;193:70525.
[12] Matsunaga H. Thermal buckling of functionally graded plates according to a
2D higher-order deformation theory. Compos Struct 2009;90:7686.
[13] Kant T, Babu CS. Thermal buckling analysis of skew fiber-reinforced composite
and sandwich plates using shear deformable finite element models. Compos
Struct 2000;49:7785.
[14] Babu CS, Kant T. Refined higher order finite element models for thermal
buckling of laminated composite and sandwich plates. J Therm Stresses
2000;23:11130.
[15] Zenkour AM. Generalized shear deformation theory for bending analysis of
functionally graded plates. Appl Math Model 2006;30:6784.
[16] Lee YY, Zhao X, Reddy JN. Postbuckling analysis of functionally graded plates
subject to compressive and thermal loads. Comput Methods Appl Mech Eng
2010;199:164553.
[17] Zhao X, Lee YY, Liew KM. Mechanical and thermal buckling analysis of
functionally graded plates. Compos Struct 2009;90:16171.
[18] Tran LV, Thai CH, Nguyen-Xuan H. An isogeometric finite element formulation
for thermal buckling analysis of functionally graded plates. Finite Elem Anal
Des 2013;73:6576.
[19] Zhang LW, Zhu P, Liew KM. Thermal buckling of functionally graded plates
using a local Kriging meshless method. Compos Struct 2014;108:47292.

[20] Ghannadpour SAM, Ovesy HR, Nassirnia M. Buckling analysis of functionally

graded plates under thermal loadings using the finite strip method. Comput
Struct 2012;108109:939.
[21] Ovesy HR, Ghannadpour SAM, Nassirnia M. Post-buckling analysis of
rectangular plates comprising functionally graded strips in thermal
environments. Comput Struct 2015;147:20915.
[22] Shahsiah R, Eslami MR. Thermal buckling of functionally graded cylindrical
shell. J Therm Stresses 2003;26(3):27794.
[23] Shahsiah
R, Eslami MR.
Functionally
graded
cylindrical
shell
thermal instability based on improved Donnel equations. AIAA J 2003;41
(9):181924.
[24] Shen H. Thermal postbuckling behavior of functionally graded cylindrical
shells with temperature dependent properties. Int J Solids Struct
2004;41:196174.
[25] Pourhamid R, Moghaddam HM, Mohammadzadeh A. Thermal analysis of
functionally graded materials in cylinders and pistons based on super element
method. J Engine Res 2013;30:2536.
[26] Yas MH, Aragh BS. Three-dimensional analysis for thermoelastic response of
functionally graded fiber reinforced cylindrical panel. Compos Struct
2010;92:23919.
[27] Shen H, Xiang Y. Thermal postbuckling of nanotube-reinforced composite
cylindrical panels resting on elastic foundations. Compos Struct
2015;123:38392.
[28] Yang J, Shen HS. Dynamic response of initially stressed functionally graded
rectangular plates resting on elastic foundations. Compos Struct 2001;54
(4):497508.
[29] Yang J, Kitipornchai S, Liew KM. Large amplitude vibration of thermo-electromechanically stressed FGM laminated plates. Comput Methods Appl Mech Eng
2003;192:386185.
[30] Woo J, Meguid SA, Ong LS. Nonlinear free vibration behavior of functionally
graded plates. J Sound Vibr 2006;289:595611.
[31] Najafizadeh MM, Isfandzibaei MR. Vibration of functionally graded cylindrical
shells based on different shear deformation shell theories with ring support
under various boundary conditions. J Mech Sci Technol 2009;23:207284.
[32] Loy CT, Lam KY, Reddy JN. Vibration of functionally graded cylindrical shells.
Int J Mech Sci 1999;41:30924.
[33] Wang X, Sheng GG. Thermal vibration, buckling and dynamic stability of
functionally graded cylindrical shells embedded in an elastic medium. J Reinf
Plast Compos 2007;27:11734.
[34] Tornabene F, Viola E, Inman DJ. 2-D differential quadrature solution for
vibration analysis of functionally graded conical, cylindrical and annular shell
structures. J Sound Vibr 2009;328:25990.
[35] Shen H. Nonlinear vibration of shear deformable FGM cylindrical shells
surrounded by an elastic medium. Compos Struct 2012;94:114454.
[36] Tornabene F. Free vibration analysis of functionally graded conical, cylindrical
shell and annular plate structures with a four-parameter power-law
distribution. Comput Methods Appl Mech Eng 2009;198:291135.
[37] Strozzi M, Pellicano F. Nonlinear vibrations of functionally graded cylindrical
shells. Thin-walled Struct 2013;67:6377.
[38] Ebrahimi MJ, Najafizadeh MM. Free vibration analysis of two-dimensional
functionally graded cylindrical shells. Appl Math Model 2014;38:30824.
[39] Bahadori R, Najafizadeh MM. Free vibration analysis of two-dimensional
functionally graded axisymmetric cylindrical shell on Winkler Pasternak
elastic foundation by first-order shear deformation theory and using Navierdifferential quadrature solution methods. Appl Math Model 2015;39
(16):487794.
[40] Liew KM, Yang J, Wu YF. Nonlinear vibration of a coating-FGM-substrate
cylindrical panel subjected to a temperature gradient. Comput Methods Appl
Mech Eng 2006;195:100726.
[41] Malekzadeh P, Bahranifard F, Ziaee S. Three-dimensional free vibration
analysis of functionally graded cylindrical panels with cut-out using
ChebyshevRitz method. Compos Struct 2013;105:113.
[42] Ramkumar K, Ganesan N. Finite-element buckling and vibration analysis of
functionally graded box columns in thermal environments. J Mater: Des App
2008;222(1):5364.
[43] Zhang LW, Lei ZX, Liew KM. Free vibration analysis of functionally graded
carbon nanotube-reinforced composite triangular plates using the FSDT and
element-free IMLS-Ritz method. Compos Struct 2015;120:18999.
[44] Stanley AJ, Ganesan N. Free vibration characteristics of stiffened cylindrical
shells. Comput Struct 1997;65(1):3345.
[45] Morimoto T, Tanigawa Y, Kawamura R. Thermal buckling of functionally
graded rectangular plates subjected to partial heating. Int J Mech Sci
2006;48:92637.
[46] Wali M, Hajlaoui A, Dammak F. Discrete double directors shell element for the
functionally graded material shell structures analysis. Comput Methods Appl
Mech Eng 2014;278:388403.
[47] Pradhana SC, Loy CT, Lam KY, Reddy JN. Vibration characteristics of
functionally graded cylindrical shells under various boundary conditions.
Appl Acoust 2000;61:11129.
[48] Shariyat M, Asgari D. Nonlinear thermal buckling and postbuckling analyses of
imperfect
variable
thickness
temperature-dependent
bidirectional
functionally graded cylindrical shells. Int J Press Vessels Pip 2013;111
112:31020.
[49] Sharma M, Bhandari M, Purohit K. Deflection of functionally gradient material
plate under mechanical, thermal and thermomechanical loading. Global J Res
Eng Mech Mech 2013;13(7):17.

R. Kandasamy et al. / Composite Structures 157 (2016) 207221

[50] Larson, Palazotto. Property estimation in FGM plates subject to low-velocity
impact loading. J Mech Mater Struct 2009;4(78):142951.
[51] Shen H. Thermal buckling and postbuckling behavior of functionally graded
carbon nanotube-reinforced composite cylindrical shells. Compos: Part B
2012;43:10308.

221

[52] Na K, Kim J. Three dimensional thermal buckling analysis of functionally

graded materials. Compos Part B: Eng 2004;35(5):42937.
[53] ANSYS, Release 15.0 UP20131014, Copyright 2013 SAS IP Inc.
[54] Dhanaraj R, Palaninathan. Free vibration of initially stressed composite
laminates. J Sound Vibr 1990;142(3):36578.