AIAA JOURNAL

Vol. 41, No. 9, September 2003

Functionally Graded Cylindrical Shell Thermal Instability

Based on Improved Donnell Equations

R. Shahsiah¤
Azad University, 14515-775 Tehran, Iran
and
M. R. Eslami†
Amirkabir University of Technology, 15914 Tehran, Iran

The thermal instability of cylindrical shells of functionally graded material is considered. The derivation of
equations is based on Ž rst-order shell theory and the complete Sanders kinematic equations. The resulting equi-
librium and the stability equations contain the rotations in the x and µ directions and the transverse shear force in
Downloaded by PENNSYLVANIA STATE UNIVERSITY on September 30, 2013 | http://arc.aiaa.org | DOI: 10.2514/2.7301

the µ direction, in addition to the conventional Donnell equations. When it is assumed that the material properties
vary linearly through the thickness direction, the system of fundamental partial differential equations in terms of
the displacement components is established. Instability analysis of functionally graded cylindrical shells under two
types of thermal loads with simply supported boundary conditions is carried out. Results are obtained in analytical
form. The results are validated with the known data in the literature.

Nomenclature thermal gradients used in the aerospace industries. FGMs are mi-
Em , Ec = modulus of elasticity of metal and ceramic croscopically inhomogeneous with mechanical properties that vary
fc = volume fraction of ceramic smoothlyand continuouslyfrom one surfaceto the other.1 Typically,
fm = volume fraction of metal these materials are made from a mixture of ceramic and metal.
h = thickness of cylindrical shell Generally, there are two ways to model the material property gra-
ki j = curvature of middle surface dation in solids: 1) Assume a proŽ le for volume fraction and 2) use a
Mi j = moment resultant micromechanicsapproachto study the nonhomogeneousmedia. For
.Mi j /1 = moment resultant of stable state compositionproŽ le modeling,polynomialrepresentationsincluding
Ni j = force resultant quadratic2 and cubic3;4 variations are used. Other representations,
.Ni j /1 = force resultant of stable state such as exponential functions5;6 and piecewise homogeneous layer
R = mean radius of cylindrical shell representations7;8 have also been used. At the microstructural level,
u = axial displacement a FGM is characterized by transition from a dispersive phase to an
u 1 , v1 , w1 = displacement components of stable state alternative structure with a networking structure in between. Zhai
v = circumferential displacement et al.9 and Nan et al.10 directly address the constitutive relations of
w = lateral displacement FGMs. SpeciŽ cally, Nan et al.10 used an analytical approach to de-
®m , ®c = thermal expansion coefŽ cient of metal and ceramic scribe the uncoupledthermomechanicalpropertiesof metal/ceramic
¯x = rotation of middle surface about the x direction FGMs. Pindera and Freed,11 Pindera et al.,12 and Aboudi et al.13 used
¯x1 = rotation of stable state about the x direction the unit-cell approach to analyze FGMs. A recent special issue on
¯µ = rotation of middle surface about the µ direction FGMs 14 focuses primarily on the micromechanics-basedstudies.
¯µ 1 = rotation of stable state about the µ direction Thermal buckling analysis of perfect cylindrical shells of
.°i j /m = shear strain of middle surface isotropic and homogeneousmaterials and cylindrical shells of com-
.°i j /m 1 = shear strain of middle surface of stable state posite materials based on the Donnell and improved Donnell sta-
²i i = normal strain bility equations are studied by Eslami et al.15 and Eslami and
.²i i /m = normal strain of middle surface Javaheri.16 Eslami and Shariyat17 consideredthe  exural theory and,
.²i i /m1 = normal strain of middle surface of stable state with the full Green nonlinear strain–displacement relations instead
º = Poisson’s ratio of the simpliŽ ed Sanders assumption, formulated the dynamic me-
¾ii = normal stress chanical and thermal buckling of imperfect cylindrical shells. The
¿i j = shear stress higher-order shear deformation theory, including the normal stress,
was used, and the mixed formulation was established to simplify
I. Introduction the approach of both kinematic and forced boundary conditions.

R ECENT studies on new performance materials have been
extended to a new material known as functionally graded
materials (FGMs). These are high-performanceheat resistant mate-
rials able to withstand ultrahigh temperatures and extremely large
Received 30 May 2002; revision received 14 April 2003; accepted for
publication 15 April 2003. Copyright °c 2003 by the American Institute of
Aeronautics and Astronautics, Inc. All rights reserved. Copies of this paper
may be made for personal or internal use, on condition that the copier pay
the $10.00 per-copy fee to the Copyright Clearance Center, Inc., 222 Rose-
wood Drive, Danvers, MA 01923; include the code 0001-1452/03 $10.00 in
correspondence with the CCC.
¤
Ph.D. Student, Mechanical Engineering Department, Science and
Research Campus.
† Professor, Mechanical Engineering Department, Distingiushed Solid
Mechanics Center. Associate Fellow AIAA.
The technique was then improved by the same authors to an ex-
act three-dimensional analysis of circular cylindrical shells based
on the equilibrium equation and the full nonlinear Green strain-
displacement relations.18 The Donnell and improved Donnell sta-
bility equations are employed to present a closed-form solution for
the elastoplastic and creep buckling of cylindrical shells under me-
chanical loads at an elevated temperature.19 Eslami and Shahsiah
determined the critical thermal buckling loads for imperfect cylin-
drical shells.20 They used the Donnell and the improved Donnell
stability equations and two models for imperfection, namely, the
Wan–Donnell and Koiter models. Many postbuckling studies based
on the classical shell theory of composite laminated thin cylindrical
shells subjected to mechanical or thermal loading or their combi-
nations are available in the literature, such as those of Birman and
Bert21 and Shen.22¡24 Relatively few studies involve the application

1819
 1820 SHAHSIAH AND ESLAMI

of shear deformation shell theory to the postbuckling analysis in where

the literature, such as those given by Iu and Chia25 Reddy and
Savoia.26 In these studies, the material properties are considered E cm D E c ¡ E m ; ®cm D ®c ¡ ®m (4)
to be independent of temperature. However, studies of temperature Consider a thin cylindrical shell of mean radius R and thickness
and moisture effects on the buckling loads of laminated  at and h with length L. The normal and shear strains at distance z from the
cylindrical panels are limited in number,27¡31 and all of these stud- shell middle surface are54
ies assumed a perfect initial conŽ guration. Palazotto and Tisler,32
Palazotto,33 Horban and Palazotto,34 Siefert and Palazotto,35 ²x D ²x m C zk x ; ²µ D ²µ m C zk µ ; °xµ D °xµ m C zk xµ (5)
Dennis and Palazotto,36;37 Isai et al.,38 Dennis and Palazotto,39 where ²i j is the normal strain, °i j is the shear strain, and ki j is the
Schimmels and Palazotto,40 Palazottoet al.,41 Chien and Palazotto,42 curvature. The subscript m refers to the strain at the middle surface
and Schimmels and Palazotto43 have done extensive theoretical and of the shell. The indices x and µ refer to the axial and circumferential
experimental work on the stability of composite panels. Their work directions,respectively.The nonlinearstrain–displacementrelations
substantially reduced the gap between the theoretical and experi- according to the complete Sanders assumption (see Ref. 54) are
mental works. Shen44 gave a full nonlinear postbucklinganalysis of ¯
composite laminated cylindrical shells subjected to combined load- ²x m D u ;x C 12 w;x2 ; ²µ m D .v;µ C w/=R C .v ¡ w;µ /2 2R 2
ing of axial compression and external pressure under hygrothermal
conditions. ° xµ m D u ;µ =R C v;x C w;x .w;µ ¡ v/=R; k x D ¡w;x x
Buckling analyses of FGM structures are rare in the literature. ¯
kµ D .v ;µ ¡ w;µ µ / R 2 ;
Downloaded by PENNSYLVANIA STATE UNIVERSITY on September 30, 2013 | http://arc.aiaa.org | DOI: 10.2514/2.7301

Birman45 studied the buckling problem of functionallygraded com- k x µ D .v;x ¡ 2w;xµ /=2R (6)
posite rectangular plate subjected to the uniaxial compression. The
stabilization of a functionally graded cylindrical shell under axial where u, v, and w are the axial and circumferential displacements
harmonicloadingwas investigatedby Ng et al.46 Javaheriand Eslami and the lateralde ections of the shell, respectively,and the subscript
presented the thermal and mechanical buckling of rectangularFGM comma indicates a partial derivative. Terms v;µ and v;x in the Ž fth
plates based on the Ž rst- and higher-order plate theories.47¡50 The and sixth of Eqs. (6) are eliminated in the classicalderivationsof the
buckling analysis of circular FGM plates is given by NajaŽ zadeh Donnell equilibrium and the stability equations. Hooke’s law for a
and Eslami.51 functionally graded cylindrical shell is deŽ ned as1
In this paper, the thermal instability of functionally graded cylin- E. z/ E. z/®.z /T
drical shell is considered. The complete Sanders nonlinear strain– ¾x D .²x m C zk x C º²µ m C º z kµ / ¡
1 ¡ º2 1¡º
displacement relations are used. The resulting equilibrium and sta-
bility equations contain the x and µ components of the rotation E.z / E.z /®.z/T
and the transverse shear force in the µ direction, in addition to the ¾µ D .²µ m C zkµ C º²xm C º zk x / ¡
1¡º 2 1¡º
conventionalDonnell equations. This approximationprovides more
reliable results for the cylindrical shells of relatively longer lengths. ¿x µ D G. z/.°x µ m C zk x µ / (7)
It is assumed that the cylindrical shell is under uniform temperature
rise and radial temperature difference for thermal loading. Simply where E.z / and ®.z /, are the elastic modulus and thermal expansion
supported boundary conditions are assumed. coefŽ cient.
The force and moment resultants expressed in terms of the stress
II. Derivations components through the thickness, according to Ž rst-order shell
theory, are
Consider a cylindrical shell made of FGM. The shell is assumed
to be graded through the thickness direction. The constituent ma-
Z h=2 Z h=2

terials are assumed to be ceramic and metal. The volume fractions Ni j D ¾i j dz ; Mi j D ¾i j z d z (8)
of ceramic f c and metal f m corresponding to the power law are ¡h=2 ¡h=2
expressed as52 Substituting Eqs. (3) and (7) in Eqs. (8) results in the constitutive
law in terms of the displacement components as
f c D [.2z C h/=2h]k ; fm D 1 ¡ fc (1)
N x D ¯1 ²xm C ¯1 º²µ m C ¯2 k x C ¯2 ºk µ ¡ ¯3 =.1 ¡ º/
where z is the thickness coordinate,¡h=2 · z · h=2, h is the thick-
ness of the shell, and k is the power law index, which takes values N µ D ¯1 ²µ m C ¯1 º²x m C ¯2 k µ C ¯2 ºk x ¡ ¯3 =.1 ¡ º/
greater than or equal to zero.52 In this paper, it is assumed that the
variation of the composition of ceramic and metal is linear, k D 1. N xµ D ¯4 °x µ m C ¯5 k xµ
The justiŽ cation for this assumption is given in Appendices A and
B. The value of k equal to zero represents a fully ceramic shell. The M x D ¯2 ²xm C ¯2 º²µ m C ¯6 k x C ¯6 ºkµ ¡ ¯7 =.1 ¡ º/
mechanical and thermal properties of FGMs are determined from
the volume fraction of the material constituents.We assume that the M µ D ¯2 ²µ m C ¯2 º²xm C ¯6 k µ C ¯6 ºk x ¡ ¯ 7 =.1 ¡ º/
nonhomogeneous material properties such as the modulus of elas-
M x µ D ¯5 °xµ m C ¯8 k xµ (9)
ticity E and the coefŽ cient of thermal expansion ® change in the
thickness direction z based on the Voigt’s rule over the whole range where
of volume fraction (see Ref. 53) as Z Z
h=2 h=2
E. z/ z E. z/
E. z/ D E c f c C E m .1 ¡ f c / ¯1 D dz ; ¯2 D dz
¡h=2 1 ¡ º2 ¡h=2 1 ¡ º2
®. z/ D ®c f c C ®m .1 ¡ f c /; º.z / D º (2) Z h=2 Z h=2
¯3 D T .z /E. z/®. z/ d z; ¯4 D G. z/ dz
where subscriptsm and c refer to the metal and ceramic constituents, ¡h=2 ¡h=2
respectively.When volume fractions are substitutedfrom Eqs. (1) in Z Z
h=2 h=2
Eqs. (2), material propertiesof the FGM shell are determined,which z 2 E.z /
are the same as the equations proposed by Praveen and Reddy52 as ¯5 D z G. z/ d z; ¯6 D dz
¡h=2 ¡h=2 1 ¡ º2
E. z/ D E m C E cm [.2z C h/=2h] Z h=2 Z h=2
¯7 D T . z/E.z /®. z/ z dz; ¯8 D z 2 G. z/ d z (10)
®. z/ D ®m C ®cm [.2 z C h/=2h]; º.z / D º (3) ¡h=2 ¡h=2
 SHAHSIAH AND ESLAMI 1821

The functional of total potential energy, including the membrane, Mµ1 D ¯2 ²µ m 1 C ¯2 º²xm 1 C ¯6 kµ1 C ¯6 ºk x 1
bending, and thermal strain energies, is written, and the Euler equa-
tion is applied to the functional of energy to obtain its stationary M x µ1 D ¯5 °xµ m 1 C ¯8 k xµ1 (14)
value. The result corresponds to the improved nonlinear Donnell
equilibrium equations as Substituting Eqs. (13) and (14) in Eqs. (12) gives the stability equa-
tions in terms of the displacement components as
R N x;x C N xµ ;µ D 0 ³ ´
¯4 ¯2 º ¯5
R N xµ ;x C Nµ ;µ C .1=R/Mµ ;µ C M xµ ;x ¡ .N µ ¯ µ C N xµ ¯ x / D 0 R¯1 u 1;x x C u 1;µ µ C v1;µ x ¯1 º C C ¯4 C C ¯1 ºw1;x
R R 2R
R M x ;x x C 2M xµ ;xµ C M µ ;µ µ = R ¡ N µ ³ ´
¯2 º ¯5
¡ R¯2 w1;xx x ¡ w1;µ µ x ¡ D0 (15a)
¡ [R N x ¯x ;x C .¯x ;µ C R¯µ ;x /N xµ C N µ ¯µ ;µ ] D 0 (11) R R
³ ´ ³ ´
Here, ¯x D ¡w;x and ¯µ D .v ¡ w;µ /=R are the rotationsin the x and ¯5 ¯2 º ¯5 ¯8
µ directions,respectively.The Ž rst and third equationsof the system ¯1 º C ¯4 C C u 1;µ x C C C R¯4 v1;x x
2R R 2 4R
of Eqs. (11) are identical with the Donnell equilibrium equations in
terms of force and moment resultants, whereas the second equation ³ ´ ³ ´
¯1 ¯2 ¯6 ¯8
Downloaded by PENNSYLVANIA STATE UNIVERSITY on September 30, 2013 | http://arc.aiaa.org | DOI: 10.2514/2.7301

is different. Terms 1=R Mµ µ , M xµ ;x , and .N µ ¯ µ C N x µ ¯ x / are added C C 2 C 3 v1;µ µ C w1;x xµ ¡¯5 ¡ ¯2 º ¡

to the conventionalDonnell equations.The reasonis that v is ignored R R R 2R
in the expression for ¯ µ in the Donnell equations. The presence of ³ ´ ³ ´
these terms indicates that the rotations along the x and µ directions ¯1 ¯2 Nµ ¯2 ¯6
C C 2C 0 w1;µ ¡ ¡ 3 w1;µ µ µ D 0 (15b)
and the transverse shear force in µ directions are included in the R R R R2 R
equilibrium equations. It is customary to neglect these terms in the
equilibrium equations. The result of ignoring the terms provides ³ ´
¯2 º ¯5
justiŽ ed buckling loads for the short cylindrical shells. For the long R¯2 u 1;x x x C C u 1;x µ µ ¡ ¯1 ºu 1;x
cylindricalshells,the eliminationof these terms results in signiŽ cant R R
error in the buckling load estimation. ³ ´ ³ ´
The stability equations are obtained by consideration of the sec- ¯6 º ¯8 ¯2 ¯6
C ¯2 º C C ¯5 C v1;µ x x C C 3 v1;µ µ µ
ond variation of the total functional of strain energy or the force R 2R R2 R
summation method. The displacement components are related to ³ ´
the terms representingthe stable equilibriumand the terms of neigh- ¯1 ¯2 ¡ ¢
boring state. Accordingly, the forces Ni j and the moments Mi j are ¡ C 2 v1;µ C 2¯2 º C R N x 0 w1;x x C R¯6 w1;x x x x
R R
divided in two terms representing the stable equilibrium and the
neighboringstate. Through the linear strain–displacement relations, ³ ´ ³ ´
the expression for total potential function is obtained. This expres- 2¯6 º ¯8 2¯2 Nµ ¯6
¡ C w1;µ µ x x C C 0 w1;µ µ ¡ w1;µ µ µ µ
sion, via the Taylor expansion, results in the sum of Ž rst and second R R R2 R R3
variations of the total potential energy. Applying the Euler equation
to the second variation of the total potential energy function results ¯1
¡ w1 D 0 (15c)
in the stability equations R

R N x1 ;x C N xµ1 ;µ D 0 Equations (15) are a coupled set of three partial differentialequa-

¡ ¢
R N xµ1 ;x C Nµ1 ;µ C .1=R/M µ1 ;µ C M xµ1 ;x ¡ Nµ0 ¯µ1 C N xµ0 ¯ x 1 D 0
¯ proved Donnell stability equations. These terms are omitted in the
R M x 1 ;x x C 2M xµ1 ;x µ C M µ1 ;µ µ R ¡ N µ1 ¡ R N x0 ¯ x1 ;x
¡ R N xµ0 ¯µ1 ;x ¡ N xµ0 ¯x 1 ;µ ¡ Nµ0 ¯µ1 ;µ D 0
In Eqs. (12), terms with the subscript 0 are related to the state of
equilibrium and terms with the subscript 1 are those characterizing
the state of stability. Note that whereas the equilibrium equations
are nonlinear, the stability equations are linear. The terms with the
subscript 0 are the solution of the equilibriumequation for the given
load. The linearized strains and curvatures in terms of the displace-
ment components are
²xm 1 D u 1;x ; ²µ m 1 D .w1 C v1;µ /= R
° xµ m 1 D v1;x C .1=R/u 1;µ ; k x 1 D ¡w1;x x
¯ w1 D C 1 sin ¸x sin nµ ;
kµ1 D .v1;µ ¡ w1;µ µ / R 2 ; k xµ1 D .v1;x ¡ 2w1;xµ /=2R
The forces and moments associated with the stability state are
N x 1 D ¯1 ²xm 1 C ¯1 º²µ m 1 C ¯2 k x 1 C ¯2 ºk µ1
Nµ1 D ¯1 ²µ m 1 C ¯1 º²x m 1 C ¯ 2 kµ1 C ¯2 ºk x 1
tions for the dependent functions u 1 , v1 , and w1 . They are the im-
proved Donnell stability equations in coupled form. The underlined
terms in these equations represent the extra terms related to the im-
classical Donnell stability equations. Equations (15) are in x, µ , and
z directions, respectively.
(12) Consider a cylindrical shell with simply supported edge condi-
tions. The boundary conditions at x D 0; L are54
w1 D w1;x x D v1 D u 1;x D 0 (16)
The one-termapproximatesolution of the system of Eqs. (15), using
the boundary conditions (16), may be assumed as54
u 1 D A1 cos ¸x sin µ
v1 D B1 sin ¸x cos nµ; 0·x·L
0 · µ · 2¼ (17)
(13)
where A1 , B1 , and C 1 are constantcoefŽ cientsand ¸ D m¼ =L, where
m D 1; 2; 3; : : : ; and n D 1; 2; 3; : : : : System (15) is made orthogo-
nal with respect to the approximate solutions (17) according to the
Galerkin method. The resulting system of equations with constant
coefŽ cients is

N xµ1 D ¯4 °xµ m 1 C ¯5 k x µ1 b11 A 1 C b12 B1 C b13 C 1 D 0; b21 A 1 C b22 B1 C b23 C 1 D 0

M x 1 D ¯ 2 ²xm 1 C ¯2 º²µ m 1 C ¯6 k x1 C ¯6 ºk µ1 b31 A 1 C b32 B1 C b33 C 1 D 0 (18)

 1822 SHAHSIAH AND ESLAMI

³ ´³ ´
where E c C Em hºn¸ h¸n
³ ´ b21 D C
¯4 n 2 ¯2 º ¯5 2 1 ¡ º2 12.1 C º/
b11 D ¸2 R¯1 C ; b12 D n¸ ¯ 1 º C C ¯4 C
R R 2R ³ ´³ ´
Ec ¡ Em h 2 ¸n h 2 º¸n
C C
¯5 ¯2 º 12 4R.1 C º/ R.1 ¡ º 2 /
b13 D n 2 ¸ ¡ ¯ 1 º¸ ¡ R¯2 ¸3 ¡ n 2 ¸
R R ³ ´³ ´
³ ´ E c C Em h R¸2 hn 2 h 3 ¸2
¯5 ¯2 º b22 D C C
b21 D ¸n ¯1 º C C ¯4 C 2 12.1 C º/ R.1 ¡ º 2 / 96R.1 C º/
2R R
³ ´³ ´
Ec ¡ Em n2h2 h 2 ¸2 h3n2
¸ 2 ¯8 2n 2 ¯2 ¯1 n 2 ¯6 n 2 C C C
b22 D ¸2 ¯5 C C ¸2 R¯4 C C C 6 R .1 ¡ º /
2 2 4.1 C º/ 4R .1 ¡ º 2 /
3
4R R2 R R3
³ ´³ ´
2 2 ¯8 ¯1 n¯2 n 3 ¯2
2 n 3 ¯6 E c C Em h 3 ¸2 n hn
b23 D ¡¸ n¯5 ¡ ¸ n¯2 º ¡ ¸ n ¡n ¡ 2 ¡ 2 ¡ 3 b23 D¡ C
2R R R R R 2 48R.1 º/C R.1 ¡ º2 /
¸n 2 ¯2 º ¸n 2 ¯5 ³ ´³
b31 D R¯2 ¸3 C C C ¯1 º¸ Ec ¡ Em h 2 n3 h 2 ¸2 nº h 2 ¸2 n
R R ¡ C C
Downloaded by PENNSYLVANIA STATE UNIVERSITY on September 30, 2013 | http://arc.aiaa.org | DOI: 10.2514/2.7301

12 R 2 .1 ¡º /2 1¡º 2 2.1 C º/
¸2 n¯ 6 º ¸2 n¯8 n 3 ¯2 n 3 ¯6 ´
b32 D ¸2 n¯2 º C C ¸2 n¯5 C C C
R 2R R 2 R3 h2n h3 n3
C 2 C
n¯1 n¯2 R .1 ¡ º /
2 2R .1 ¡ º 2 /
3

C C 2 ³ ´
R R
E c C Em hº¸
b31 D
4 2 2¸2 n 2 ¯6 º
2 ¸2 n 2 ¯8 2 1 ¡ º2
b33 D R¯6 ¸ ¡ 2¸ ¯2 º ¡ ¸ R N x 0 ¡ ¡
R R ³ ´³ ´
2 4
Ec ¡ Em h 2 ¸3 R h 2 n 2 ¸º h 2 ¸n 2
2n ¯2 n ¯6 ¯1 C C C
¡ ¡ 3 ¡ (19) 12 .1 ¡ º /2 R.1 ¡ º / 2R.1 C º/
2
R2 R R
µ ¶
Substituting Eqs. (3) in Eqs. (10) gives E c C Em h 3 ¸2 n hn
b32 D C
2 48R.1 C º/ R.1 ¡ º 2 /
h h2
¯1 D .E c C E m /; ¯2 D .E c ¡ E m / ³ ´µ
2.1 ¡ º 2 / 12.1 ¡ º 2 / Ec ¡ Em h 2 ¸2 nº h 3 ¸2 nº h 2 ¸2 n
£ ¤ C C C
1
¯3 D T E m ®m h C E m ®cm h C E cm ®m h C E cm ®cm h 1 1 12 1¡º 2 2R.1 ¡ º / 2 2.1 C º/
2 2 3
¶
h h2 h2 n3 h3 n3 h2n
¯4 D .E c C E m /; ¯5 D .E c ¡ E m / C 2 C C
24.1 C º/ 24.1 C º/ R .1 ¡ º 2 / 2R 3 .1 ¡ º 2 / R 2 .1 ¡ º 2 /
³ ´µ ¶
h3 E c C Em h h 3 ¸2 n 2
¯6 D .E c ¡ E m / b33 D ¡ C
24.1 ¡ º 2 / 2 R.1 ¡ º / 24R.1 C º/
2
µ ¶
E m ®cm h 2 E cm ®m h 2 E cm ®cm h 2 ³ ´µ
¯7 D T C C Ec ¡ Em h 3 ¸4 R h 2 ¸2 º h 3 ¸2 n 2 º
12 12 12 C ¡ ¡
6 4.1 ¡ º 2 / 1 ¡ º2 2R.1 ¡ º 2 /
h3 ¶
¯8 D .E c C E m / (20) h2 n2 h3 n4
48.1 C º/ ¡ ¡ ¡ ¸2 R N x 0 (21)
R .1 ¡ º /
2 2 4R .1 ¡ º 2 /
3
In Eqs. (20), the coefŽ cients ¯3 and ¯ 7 are obtained for the uniform
temperature rise, where T is a constant value through the thickness. Equations (18) are a system of homogeneous equations that have a
When T is a functionof the radius coordinate,¯3 and ¯ 7 are obtained nontrivial solution only for discrete values of the thermal load 1T .
from Eqs. (10). Equations (20) are substituted in Eqs. (19), and the The determinant of the coefŽ cients A 1 , B1 , and C 1 are set equal to
members of determinant of the matrix of coefŽ cients become zero, as
³ ´³ ´
E c C Em ¸2 Rh hn 2 ­ ­
b11 D C ­
b11 b12 b13 ­
2 1¡º 2 12R.1 C º/ ­ ­
­
b21 b22 b23 ­
D0 (22)
³ ´³ ´ ­ ­
E c C Em hºn¸ hn¸ ­
b31 b32 b33 ­
b12 D C
2 1¡º 2 12.1 C º/
³ ´³ ´ The resulting equation may be solved for the buckling load. The
Ec ¡ E m h 2 ºn¸ h 2 n¸ critical buckling load is the minimum value of the determinant for
C C values of m and n, the longitudinal and circumferential buckling
12 R.1 ¡ º / 4R.1 C º/
2
waves.
³ ´
Ec C Em hº¸
b13 D ¡
2 1 ¡ º2 III. Uniform Temperature Rise
³ ´³ ´ The initial uniform temperature of the shell is assumed to be
Ec ¡ E m h2n2¸ h 2 ¸3 R h 2 n 2 ¸º Ti . Under simply supported boundary conditions, temperature is
C ¡ ¡
12 2R.1 C º/ 1 ¡ º 2 R.1 ¡ º 2 / uniformly raised to a Ž nal value T f such that the shell buckles.
 SHAHSIAH AND ESLAMI 1823

To Ž nd the critical 1T D T f ¡ Ti , the prebuckling forces should be reduces to 1Tcr D 0:632h=R®. This result is obtained and given by
found. Solving the membrane form of the equilibrium equations, Eslami et al.15
using the method developed by Meyers and Hyer,55 gives the pre-
buckling force resultants as V. Nonlinear Temperature Distribution
¡ Through the Thickness
N x 0 D ¡[1T =.1 ¡ º/] E m ®m h C 12 E m ®cm h A cylindrical shell with inner surface temperature Ta and outer
1 1
¢ surface temperature Tb is considered. Temperature distribution
C E cm ®m h C E cm ®cm h
2 3 across the thickness is obtained by solving the differential equa-
tion of thermal conduction as
Nµ 0 D N x µ 0 D 0 (23) µ ¶
@ 2 T .r/ 1 @ K .r / 1 @ T .r /
When coefŽ cients from Eqs. (21) are substituted into Eq. (22) and C C D0 (28)
the result is simpliŽ ed, the Ž nal expression for 1T becomes @r 2 K .r/ @r r @r

.1 ¡ º/.b11 b23 b32 C b13 b22 b31 ¡ b12 b23 b31 ¡ b13 b21 b32 / .1 ¡ º/
1T D ¡ ¢ ¡ ¡ ¢
¸2 R E m ®m h C 12 E m ®cm h C 12 E cm ®m h C 13 E cm ®cm h .b11 b22 ¡ b12 b21 / ¸2 R E m ®m h C 12 E m ®cm h C 12 E cm ®m h C 13 E cm ®cm h
Downloaded by PENNSYLVANIA STATE UNIVERSITY on September 30, 2013 | http://arc.aiaa.org | DOI: 10.2514/2.7301

» ³ ´µ ¶ ³ ´µ ¶¼
Ec C Em h h 3 ¸2 n 2 E c ¡ Em h 3 ¸4 R h 2 ¸2 º h 3 ¸2 n 2 º h2 n2 h3 n4
£ ¡ C C ¡ ¡ ¡ 2 ¡
2 R.1 ¡ º / 24R.1 C º/
2 6 4.1 ¡ º / 1 ¡ º
2 2 2R.1 ¡ º / 2 R .1 ¡ º /
2 4R .1 ¡ º 2 /
3

(24)

The uniform buckling temperature difference is obtained by mini- where K .r / is the coefŽ cient of thermal conductionof the function-
mizing Eq. (24) with respect to m and n, the number of longitudinal ally graded cylinder. We may assumed a linear distributionfor K .r/
and circumferential buckling waves. For pure isotropic metal or in the radial direction, as discussed in the Appendices, by
pure isotropic ceramic in relations (10), ¯2 D 0, ¯5 D 0, and ¯7 D 0
and Eq. (24) reduces to 1Tcr D 0:316h=R®. This result is given by K .r / D K 0r (29)
Eslami et al.15
where K 0 is a constant value. The solution of Eq. (28) is obtained
IV. Linear Temperature Change by assuming a distribution for temperature of the form
Through the Thickness
T .r/ D b1r t1 C b2r t2 C b3 (30)
Becausethe shellthicknessis thin, the temperaturevariationalong
the radial direction through the thickness may be approximated by
a linear function as where b1 , b2 , b3 , t1 , and t2 are constant coefŽ cients. Substituting the
solution given by Eq. (30) in Eq. (28) and solving for T .r / gives
T . z/ D .1T = h/.z C h=2/ C Ta (25)
T .r/ D b4 =r C b3 (31)
where 1T D Tb ¡ Ta , where Ta is temperature of inner surface and
Tb is temperature of the outer surface of functionally graded shell. To obtain the constant coefŽ cients b3 and b4 , the boundary condi-
The membrane solution of the equilibrium equations, using the tions are assumed as
method developed by Meyers and Hyer,55 results in the prebuck-
ling force resultants T D Ta at r D ra
¡
N x 0 D [¡1T =2.1 ¡ º/] E m ®m h C 13 E m ®cm h T D Tb at r D rb (32)
¢ where ra and rb are the inner and outer radii of the cylindrical shell,
C 13 E cm ®m h C 16 E cm ®cm h
respectively. When the boundary conditions are used, the tempera-
Nµ 0 D N x µ 0 D 0 (26) ture distributionacrossthe thicknessof the cylindricalshell becomes

When coefŽ cients from Eqs. (21) are substitutedin Eq. (22) and the .Ta ¡ Tb /ra rb rb
T .r/ D ¡ .Ta ¡ Tb / C Ta (33)
result is simpliŽ ed, the Ž nal expression for 1T becomes hr h

2.1 ¡ º/.b11 b23 b32 C b13 b22 b31 ¡ b12 b23 b31 ¡ b13 b21 b32 / 2.1 ¡ º/
1T D ¡ 1 1 1
¢ ¡ ¡ ¢
¸2 R E m ®m h C E m ®cm h C E cm ®m h C E cm ®cm h .b11 b22 ¡ b12 b21 /
3 3 6
¸2 R E m ®m h C E m ®cm h C 13 E cm ®m h C 16 E cm ®cm h
1
3
» ³ ´µ ¶ ³ ´µ ¶¼
Ec C Em h h 3 ¸2 n 2 Ec ¡ Em h 3 ¸4 R h 2 ¸2 º h 3 ¸2 n 2 º h2 n2 h3 n4
£ ¡ C C ¡ ¡ ¡ ¡
2 R.1 ¡ º 2 / 24R.1 C º/ 6 4.1 ¡ º 2 / 1 ¡ º 2 2R.1 ¡ º 2 / R 2 .1 ¡ º 2 / 4R 3 .1 ¡ º 2 /
(27)

The linear buckling temperature difference is obtained by minimiz- The prebuckling forces are obtained by solving the mem-
ing Eq. (27) with respect to m and n, the number of longitudinaland brane form of the equilibrium equations, using the method devel-
circumferential buckling waves. For pure isotropic metal or pure oped by Meyers and Hyer.55 The prebuckling force resultants are
isotropic ceramic in relations (10), ¯2 D 0, and ¯5 D 0, and Eq. (27) then
 1824 SHAHSIAH AND ESLAMI

¡2 A0 E m ®m rav C h=2 A 0 E m ®cm rav C h=2

Nx0 D ¡
1¡º rav 1¡º rav
µ
A 0 E cm ®m rav C h=2 A 0 E cm ®cm h 2
¡ ¡ 2 ¡ hrav
1¡º rav h .1 ¡ º/ 4
¶
2 rav C h=2 A 0 E cm ®cm r av C h=2 B 0 E m ®m h
C 2r av ¡ ¡
r av 2.1 ¡ º/ rav 1¡º

B 0 E m ®cm h B 0 E cm ®cm h B 0 E cm ®m h B 0 E m ®cm h

¡ ¡ ¡ ¡
2.1 ¡ º/ 12.1 ¡ º/ 2.1 ¡ º/ 4.1 ¡ º/
(34)
where
A0 D .Ta ¡ Tb /.rb ra = h/; B 0 D Ta ¡ .Ta ¡ Tb /.rb = h/ (35)
Fig. 2 Variation of the critical uniform temperature rise vs h/R.
The coefŽ cients from Eqs. (21) are substitutedin Eq. (22), and upon
Downloaded by PENNSYLVANIA STATE UNIVERSITY on September 30, 2013 | http://arc.aiaa.org | DOI: 10.2514/2.7301

simpliŽ cation, the Ž nal expressions for N x 0 becomes

b12 b23 b31 C b13 b21 b32 ¡ b13 b22 b31 ¡ b11 b23 b32
Nx0 D
¸2 R.b12 b21 ¡ b11 b22 /
» ³ ´µ ¶
1 E c C Em h h 3 ¸2 n 2
C 2 ¡ C
¸ R 2 R.1 ¡ º 2 / 24R.1 C º/
³ ´µ
Ec ¡ E m h 3 ¸4 R h 2 ¸2 º h 3 ¸2 n 2 º
C ¡ ¡
6 4.1 ¡ º /2 1¡º 2 2R.1 ¡ º 2 /
¶¼
h2 n2 h3 n4
¡ 2 ¡ (36)
R .1 ¡ º 2 / 4R 3 .1 ¡ º 2 /
The buckling force is obtained by minimizing Eq. (36) with respect
to m and n, the number of longitudinaland circumferentialbuckling
waves. Then, this expression for N x0 is substituted in Eq. (34) to
provide the buckling temperature difference 1Tcr D Tb ¡ Ta . Fig. 3 Variation of the critical linear radial temperature difference
vs h/R.
VI. Results and Discussion
Consider a ceramic–metal functionally graded cylindrical shell
(Fig. 1) The combination of materials consist of steel and alumina.
Young’s modulus and the coefŽ cient of thermal expansion are for
steel, E m D 200 GPa and ®m D 11:7 £ 10¡6 1=± C, and for alumina,
E c D 380 GPa and ®c D 7:4£10¡6 1=± C, respectively.Poisson’s ratio
is assumed to be 0:3 for steel and alumina. Simply supported bound-
ary conditions are assumed. The variation of the buckling tempera-
ture difference 1Tcr vs the variation of the dimensionless geomet-
rical parameter h=R is plotted for three loading cases in Figs. 2–4.
In Figs. 2–4, three arbitrary values of L=R are considered, and the
mean radius is assumed R D 0:5 m. As can be seen, 1Tcr is increased
as L=R and h=R are increased. The buckling temperatures related
to the uniform temperature rise and linear temperature through the
thicknessfor cylindricalshells made of isotropic material are higher
than the FGM cylindricalshell (Figs. 2 and 3). Comparisonof Figs. 3
and 4 reveals that the buckling temperature associated with the non-
linear temperature distribution across the thickness is greater than Fig. 4 Variation of the critical nonlinearradial temperature difference
the linear temperature distribution. vs h/R.

a) b)
Fig. 1 Distribution of metal and ceramic in the FGM cylinder.
 SHAHSIAH AND ESLAMI 1825

The buckling modes m and n related to the critical uniform tem- Substituting relations (A1) and (B2) in relation (B4) gives
perature rise are checked. Note that as h=R is increased, assuming
constant L=R, the number of bucklingwaves m and n are decreased. Pt D Pm [.h ¡ 2z/=2h] C Pc [.h C 2z/=2h] (B5)
Also, as L=R is increased, with constant h=R, m is increased and n
is decreased. The limiting value for m is 1 and for n is 2. Therefore,
Pt D Pm C Pcm [.h C 2 z/=2h]; Pcm D Pc ¡ Pm (B6)
VII. Conclusions
Equilibrium and stability equations for simply supported func- Here, P may be the modulus of elasticity, thermal expansion co-
tionally gradedcylindricalshells are obtained.Derivationsare based efŽ cient, Poisson’s ratio, or any other material properties being a
on Ž rst-order shell theory the Sanders kinematic relations and the function of graded materials.
improvedDonnellstabilityequations.The bucklinganalysisof func-
tionally graded cylindrical shells under three differentloading cases References
are investigated. The following conclusions are reached: 1 Suresh, S., and Mortensen, A., Fundamentals of Functionally Graded
1) The equilibrium and stability equations are identical with the Materials, Cambridge Univ. Press, London, 1998.
corresponding equations for the pure isotropic cylindrical shell in 2 Fuchiyama, T., Noda, N., Tsuji, T., and Obata, Y., “Analysis of Ther-
form of forces and moments per unit length. mal Stress and Stress Intensity Factor of Functionally Gradient Materi-
2) The critical temperature difference 1Tcr for the functionally als,” Ceramic Transactions, Functionally Gradient Materials, Vol. 34, 1993,
graded cylindrical shell is generally lower than the corresponding pp. 425–432.
Downloaded by PENNSYLVANIA STATE UNIVERSITY on September 30, 2013 | http://arc.aiaa.org | DOI: 10.2514/2.7301

value for the pure isotropic cylindrical shell.
3) The critical temperature difference 1Tcr for the functionally
graded cylindrical shell is increase by increasing the dimensionless
ratios h=R and L=R.
4) The critical temperature difference 1Tcr for a FGM cylindrical
shell under a linear temperature difference through the thickness
is twice of the FGM cylindrical shell under uniform temperature
difference.
5) The critical temperature difference 1Tcr for FGM cylindrical
shell under nonlinear temperature difference through the thickness
is greater than the FGM cylindrical shell under linear temperature
difference through the thickness.
3 Williamson, R. L., Rabin, B. H., and Drake, J. T., “Finite Element Analy-
sis of Thermal Residual Stresses at Graded Ceramic–Metal Interfaces, Part I.
Model Description and Geometric Effects,” Journal of Applied Physics,
Vol. 74, 1993, pp. 1310–1320.
4 Fukui, Y., Yamanaka, Y., and Wakashima, K., “The Stress and Strains
in a Thick-Walled Tube of Functionally Graded Materials Under Uniform
Thermal Loading,”InternationalJournal of Japan Society of MechanicalEn-
gineers, Ser. A, Vol. 36, 1993, pp. 156–162.
5 Delale, F., and Erdogan, F., “The Crack Problem for a Nonhomogeneous
Plane,” Journal of Applied Mechanics, Vol. 50, 1983, pp. 609–614.
6 Noda, N., and Jin, Z. H., “Thermal Stress Intensity Factors for a Crack in
a Strip of a Functionally Gradient Material,” International Journal of Solids
and Structures, Vol. 30, 1993, pp. 1039–1056.
7 Araki, N., Makino, A., Ishiguro, T., and Mihara, J., “An Analyti-

cal Solution of Temperature Response in Multilayered Material for Tran-

Appendix A: Calculation of Volume
Fraction of Metal in Cylinder
Figure 1 shows the distribution of ceramic and metal in the cross
section of cylindricalshell. To calculatethe volume fraction of metal
(Fig. 1a), the volume of metal 8m and the total volume 8m C c must
be used as
sient Method,” International Journal of Thermophysics, Vol. 13, 1992,
pp. 515–538.
8 Tanigawa, Y., “Theoretical Approach of Optimum Design for a Plate of
Functionally Gradient Materials Under Thermal Loading,” Thermal Shock
and Thermal Fatigue Behavior of Advanced Ceramics, NATO ASI Ser. E,
Vol. 241, 1992, pp. 171–180.
9 Zhai, P. C., Jiang, C. R., and Zhang, Q. J., “Application of Three-Phase

Micromechanical Theories to Ceramic/Metal Functionally Gradient Materi-

als,” Ceramic Transactions, Functionally Gradient Materials, Vol. 34, 1993,
8m D 2¼ ri .r ¡ ri /L ; 8m C c D 2¼ri .r0 ¡ ri /L (A1) p. 449.
10 Nan, C. W., Yuan, R. Z., and Zhang, L. M., “The Physics of

Therefore, Metal/Ceramic Functionally Gradient Materials,” Ceramic Transactions,

Functionally Gradient Materials, Vol. 34, 1993, pp. 75–82.
8m 2¼ri .r ¡ ri /L r ¡ ri h=2 ¡ z h ¡ 2z 11
Pindera, M. J., and Freed, A. D., “The Effect of Matrix Microstructure
fm D D D D D
8m C c 2¼ri .r0 ¡ ri /L r0 ¡ ri h 2h on the Evolution of Residual Stresses in Titanium–Aluminide Composites,”
AMD Vol. 40, American Society of Mechanical Engineers, New York, 1992,
(A2) pp. 37–52.
12 Pindera, M. J., Salzar, R. S., and Williams, T. O., “An Evaluation of a

Appendix B: Calculation of Volume New Approach for the Thermoplastic Response of Metal–Matrix Compos-
Fraction of Ceramic in Cylinder ites,” Composites Engineering, Vol. 3, 1993, pp. 1185–1201.
13 Aboudi, J., Pindera, M. J., and Arnold, S. M., “Thermoelastic Response
To calculate the volume fraction of ceramic (Fig. 1b), the volume of Metal Matrix Composites with Large Diameter Fibers Subjected to Ther-
of ceramic 8c and the total volume 8m C c must be used as mal Gradients,” NASA TM 106344, April 1993.
14 Pindera, M. J., Arnold, S. M., Aboudi, J., and Hui, D., “Use of Compos-
8c D 2¼ r0 .r0 ¡ r /L; 8mCc D 2¼r0 .r0 ¡ ri /L (B1) ites in Functionally Gradient Materials,” Composites Engineering, Vol. 4,
1994, pp. 154–180.
15 Eslami, M. R., Ziaii, A. R., and Ghorbanpour,A., “Thermoelastic Buck-
Thus,
ling of Thin Cylindrical Shells Based on Improved Donnell Equations,”
8c 2¼r0 .r0 ¡ r/L r0 ¡ r h=2 C z h C 2z Journal of Thermal Stresses, Vol. 19, 1996, pp. 299–316.
fc D D D D D 16 Eslami, M. R., and Javaheri, R., “Thermal and Mechanical Buckling of
8m C c 2¼ r0 .r0 ¡ ri /L r0 ¡ ri h 2h
Composite Cylindrical Shells,” Journal of Thermal Stresses, Vol. 22, No. 6,
(B2) 1999, pp. 527–545.
17 Eslami, M. R., and Shariyat, M., “A High-Order Theory for Dynamic

With consideration of the Voigt model, the overall mechanical Buckling and Postbuckling Analysis of Laminated Cylindrical Shell,” Jour-
properties are nal of Pressure Vessel Technology, Vol. 121, Feb. 1999, pp. 94–102.
18 Eslami, M. R., and Shariyat, M., “Dynamic Buckling and Postbuckling

Pt 8t D Pm 8m C Pc 8c of Imperfect Orthotropic Cylindrical Shells Under Mechanical and Thermal

Here, Pt , Pm , and Pc are the mechanical properties of the overall
FGM, metal, and ceramic, respectively. The weighted average of
the mechanical properties is
Pt D Pm f m C Pc f c
(B3)
Loads, Based on the Three-Dimensional Theory of Elasticity,” Journal of
Applied Mechanics, Vol. 66, June 1999, pp. 476–484.
19 Eslami, M. R., and Shariyat, M., “Elastic, Plastic, and Creep Buckling
of Imperfect Cylinders Under Mechanical and Thermal Loading,” Journal
of Pressure Vessel Technology, Vol. 119, No. 1, 1997, pp. 27–36.
20 Eslami, M. R., and Shahsiah, R., “Thermal Buckling of Imperfect Cylin-
(B4) drical Shells,” Journal of Thermal Stresses, Vol. 24, No. 1, 2000, pp. 71–90.
 1826 SHAHSIAH AND ESLAMI
21 Birman, V., and Bert, C. W., “Buckling and Post-Bucklingof Composite
Plates and Shells Subjected to Elevated Temperature,” Journal of Applied
Mechanics, Vol. 60, 1993, pp. 514–519.
22 Shen, H. S., “Post Buckling Analysis of Imperfect Stiffened Laminated
Cylindrical Shells Under Combined External Pressure and Axial Compres-
sion,” Computers and Structures, Vol. 63, 1997, pp. 335–348.
23
Shen, H. S., “Postbuckling Analysis of Imperfect Stiffened Lami-
nated Cylindrical Shells under Combined External Pressure and Thermal
Loading,” International Journal of Mechanical Sciences, Vol. 40, 1998,
pp. 339–355.
24 Shen, H. S., “Thermomechanical Postbucklingof Composite Laminated
Cylindrical Shells with Local Geometric Imperfections,” InternationalJour-
nal of Solids and Structures, Vol. 36, 1999, pp. 597–617.
25 Iu, V. P., and Chia, C. Y., “Effect of Transverse Shear on Nonlinear
Vibration and Postbuckling of Anti-Symmetric Cross-Ply Imperfect Cylin-
drical Shells,” International Journal of Mechanical Sciences, Vol. 30, 1988,
pp. 705–718.
26 Reddy, J. N., and Savoia, M., “Layerwise Shell Theory for Postbuckling
of Laminated Circular Cylindrical Shells,” AIAA Journal, Vol. 30, 1992,
pp. 2148–2154.
Downloaded by PENNSYLVANIA STATE UNIVERSITY on September 30, 2013 | http://arc.aiaa.org | DOI: 10.2514/2.7301
27 Whitney, J. M., and Ashton, J. E., “Effect of Environment on the Elas-
tic Response of Layered Composite Plates,” AIAA Journal, Vol. 9, 1971,
pp. 1708–1713.
28 Snead, J. M., and Palazotto, A. N., “Moisture and Temperature Effects
on the Instability of Cylindrical Composite Panels,” Journal of Aircraft,
Vol. 20, 1983, pp. 777–783.
29 Lee, S. Y., and Yen, W. J., “Hygrothermal Effects on the Stability of
a Cylindrical Composite Shell Panel,” Computers and Structures, Vol. 33,
1989, pp. 551–559.
30 Ram, K. S. S., and Sinha, P. K., “Hygrothermal Effects on the Buck-
ling of Laminated Composite Plates,” Composite Structures, Vol. 21, 1992,
pp. 233–247.
31 Chao, L. P., and Shyu, S. L., “Nonlinear Buckling of Fiber Reinforced
Composite Plates Under Hygrothermal Effects,” Journal of Chinese Institute
of Engineers, Vol. 19, 1996, pp. 657–667.
32 Palazotto, A. N., and Tisler, T. W., “Experimental Collapse Determi-
nation of Cylindrical Composite Panels with Large Cutouts Under Axial
Load,” Composite Structures, Vol. 12, 1989, pp. 61–78.
33 Palazotto, A. N., An Experimental Study of a Curved Composite Panel
with a Cutout, Vol. 972, American Society for Testing and Materials,
Philadelphia, 1988, pp. 191–202.
34 Horban, B. A., and Palazotto, A. N., “Experimental Buckling of Cylin-
drical Composite Panels with Eccentrically Located Circular Delamina-
tions,” Journal of Spacecraft and Rockets, Vol. 24, 1987, pp. 349–352.
35 Siefert, G. R., and Palazotto, A. N., “The Effect of a Centrally Located
Midplane Delamination on the Stability of Composite Panels,” Experimental
Mechanics, Vol. 26, 1986, pp. 330–336.
36
Dennis, S. T., and Palazotto, A. N., “Large Displacement and Rota-
tional Formulation for Laminated Shells Including Parabolic Transverse
Shear,” International Journal of Nonlinear Mechanics, Vol. 25, 1990,
pp. 67–85.
37 Dennis, S. T., and Palazotto, A. N., “Transverse Shear Deformation in
Orthotropic Cylindrical Pressure Vessels Using a High-Order Shear Theory,”
AIAA Journal, Vol. 27, 1989, pp. 1441–1447.
38
Tsai, C. T., Palazotto, A. N., and Dennis, S. T., “Large Rotation Snap
Through Buckling in Laminated Cylindrical Panels,” Finite Elements in
Analysis and Design, Vol. 9, 1991, pp. 65–75.
39 Dennis, S. T., and Palazotto, A. N., “Effect of Nonlinear Curva-
ture Strains on the Buckling of Laminated Plates and Shells,” Interna-
tional Journal for Numerical Methods in Engineering, Vol. 36, 1993,
pp. 595–610.
40
Schimmels, S. A., and Palazotto, A. N., “Nonlinear Geometric and
Material Behavior of Composite Shells with Large Strains,” Journal of En-
gineering Mechanics, Vol. 120, 1994, pp. 320–345.
41 Palazotto, A. N., Chien, L. S., and Taylor, W. W., “Stability Charac-
teristics of Laminated Cylindrical Panels Under Transverse Loading,” AIAA
Journal, Vol. 30, 1992, pp. 1649–1653.
42 Chien, L. S., and Palazotto, A. N., “Dynamic Buckling of Composite
Cylindrical Panels with Higher-Order Transverse Shear Subjected to a Trans-
verse Concentrated Load,” International Journal of Nonlinear Mechanics,
Vol. 27, 1992, pp. 719–734.
43 Schimmels, S. A., and Palazotto, A. N., “Collapse Characteristics
of Cylindrical Panels Under Axial Loads,” AIAA Journal, Vol. 30, 1992,
pp. 1447–1466.
44 Shen, H. S., “Hygrothermal Effects on the Postbuckling of Composite
Laminated Cylindrical Shells,” Composite Science and Technology, Vol. 60,
2000, pp. 1227–1240.
45 Birman, V., “Buckling of Functionally Graded Hybrid Composite
Plates,” Proceeding of the l0th Conference on Engineering Mechanics, Vol. 2,
Elsevier Science, 1995, pp. 1199–1202.
46 Ng, T. Y., Lam, Y. K., Liew, K. M., and Reddy, J. N., “Dynamic Stability
Analysis of Functionally Graded Cylindrical Shell Under Periodic Axial
Loading,” International Journal of Solids and Structures, Vol. 38, 2001,
pp. 1295–1300.
47 Javaheri, R., and Eslami, M. R., “Thermoelastic Buckling of Rectangu-
lar Plates Made of Functionally Graded Materials,” AIAA Journal, Vol. 40,
No. 1, 2002, pp. 162–169.
48 Javaheri, R., and Eslami, M. R., “Buckling of Functionally Graded
Plates under In-plane Compressive Loading,” ZAMM, Vol. 82, No. 4, 2002,
pp. 277–283.
49 Javaheri, R., and Eslami, M. R., “Buckling of Functionally Graded
Plates under In-plane Compressive Loading Based on Various Plate The-
ories,” Journal of Iranian Society of Mechanical Engineers (submitted for
publication).
50 Javaheri, R., and Eslami, M. R., “Thermal Buckling of Functionally
Graded Plates Based on Higher Order Theory,” Journal of Thermal Stresses,
Vol. 25, No. 7, 2002, pp. 603–625.
51 NajaŽ zadeh, M. M., and Eslami, M. R., “First-Order-Theory-Based
Thermoelastic Stability of Functionally Graded Material Circular Plates,”
AIAA Journal, Vol. 40, No. 7, 2002, pp. 444–450.
52 Praveen, G. N., and Reddy, J. N., “Nonlinear Transient Thermoelas-
tic Analysis of Functionally Graded Ceramic Metal Plates,” International
Journal of Solids and Structures, Vol. 35, No. 33, 1998, pp. 4457–4476.
53
Sugano, Y., “An Analytical Solution for a Plane Thermal Stress Prob-
lem in Nonhomogeneous Multiply Connected Regions,” Japan Society
of Mechanical Engineers International Journal, Ser. A, Vol. 33, 1990,
pp. 136–144.
54
Brush, D. O., and Almroth, B. O., Buckling of Bars, Plates and Shells,
McGraw–Hill, New York, 1975.
55
Meyers, C. A., and Hyer, M. W., “Thermal Bucklingand Postbucklingof
Symmetrically Laminated Composite Plates,” Journal of Thermal Stresses,
Vol. 14, 1991, pp. 519–540.
A. N. Palazotto
Associate Editor