Int J Mech Mater Des (2012) 8:51–69

DOI 10.1007/s10999-011-9175-1

Static solutions for functionally graded simply

supported plates
Sandeep S. Pendhari • Tarun Kant •
Yogesh M. Desai • C. Venkata Subbaiah

Received: 20 April 2011 / Accepted: 19 October 2011 / Published online: 15 November 2011
 Springer Science+Business Media, B.V. 2011

Abstract In this article mixed semi-analytical and 1 Introduction

analytical solutions are presented for a rectangular plate
made of functionally graded (FG) material. All edges of A functionally graded (FG) material is a micro-
a plate are under simply supported (diaphragm) end engineered composites where the composite and
conditions and general stress boundary conditions can structures are continuously varied in thickness direc-
be applied on both top and bottom surface of a plate tion of a structural component so that an optimum
during solution. A mixed semi-analytical model consists distribution of properties can be obtained depending
in defining a two-point boundary value problem on the functional requirements and are therefore, free
governed by a set of first-order ordinary differential from interface weaknesses typically consists in
equations in the plate thickness direction. Analytical laminated composites and sandwiches. The great
solutions based on shear-normal deformation theories feature of this particular material is elimination or
are also established to show the accuracy, simplicity and reduction of delamination failure associated with the
effectiveness of mixed semi-analytical model. The FG traditional laminated components at their interfaces
material is assumed to be exponential in the thickness where a material property varies suddenly. FG
direction and Poisson’s ratio is assumed to be constant. material offers many advantages over the monolithic
material and traditional laminated composites and
Keywords FG plates  Mixed semi analytical sandwiches, and therefore, extensive use of FG
method  ODE  PDE  BVP material has been seen in the field of aerospace and
nuclear technologies. However, directional composi-
tional variation of the constituents within FG mate-
rials makes analysis and design more challenging
S. S. Pendhari (&)
Structural Engineering Department, Veermata than traditional materials.
Jijabai Technological Institute, H.R. Mahajani Over the years, a number of approaches/models
Marg, Matunga, Mumbai 400019, India have been developed and presented to analysis of
e-mail: [email protected]
structural components with FG material under trans-
T. Kant  Y. M. Desai verse/thermal/electric loads. A comprehensive review
Department of Civil Engineering, Indian Institute on FG materials can be found in Tanigawa (1995).
of Technology Bombay, Mumbai 400076, India Three dimensional (3D) elasticity solutions based on
the solution of partial differential equations (PDEs)
C. V. Subbaiah
Department of Civil Engineering, Indian Institute with appropriate boundary conditions are valuable
of Technology Bombay, Mumbai 400076, India because they represent a more realistic and closer

123
52 S. S. Pendhari et al.

approximation to the actual behaviour of the struc- loading and temperature fields using Von-Karman
tures. Sankar (2001) has presented two dimensional theory. The material properties of FG materials are
(2D) elasticity solution under plane strain conditions assumed to vary according to power-law distribution
for simply supported FG beams subjected to sinusoidal of the volume fraction of the constituents through the
load by assuming exponential variation of Young’s thickness. Further, analytical solution is presented by
modulus through the thickness of beam. With the same Woo et al. (2005) for post-buckling analysis of
assumption, a mixed semi-analytical solution along moderately thick FG plates and shells under edge
with analytical solutions based on shear deformation compression loads and a temperature field. Bodaghi
theories for simply supported FG beams under plane and Saidi (2010) presented analytical approach based
stress condition of elasticity have been presented by on a HOST to determine critical buckling loads of
Pendhari et al. (2010). Sankar and Tzeng (2002) thick FG rectangular plates.
extended the elasticity solutions for a FG beams Praveen and Reddy (1998) developed finite element
subjected to thermal loads. (FE) model for static and dynamic analysis of FG
Kashtalyan (2004) developed exact 3D elasticity ceramic–metal plates with Von-Karman type nonlin-
solutions for simply supported FG plate subjected to earity. Power-law dependence of material properties
transverse loading by assuming exponential variation in the thickness direction of plate is assumed. Reddy
of Young’s modulus through the thickness of plate. By and Chin (1998) derived boundary value problem
assuming power-law variation of the volume fractions (BVP) by using FOST to study the dynamic thermo-
of the constituents through the thickness of simply elastic response of FG cylinders and plates. The
supported FG plate, exact solutions are presented by presented formulation accounting for the coupling
Vel and Batra (2002) for mechanical and thermal with 3D heat conduction equation for a FG plate.
loading. Governing PDE for thermo-mechanical defor- Further, Reddy (2000) extended same formulation for
mation is reduced to a set of coupled ordinary third order shear deformation plate theory. With the
differential equations (ODEs) in the thickness co- help of classical plate theory (CPT), Shen (2002) and
ordinate, which is then solved by using the power series Yang and Shen (2003) studied large deflection and
method. Anderson (2003) solved sandwich composites post buckling response of FG plates with temperature
with FG core components exactly for circular patch dependent material properties. Chakraborty and
loading. Parametric studies on different degrees of core Gopalakrishnan (2003) have developed a FE model
stiffness at the face sheet interface are also presented. based on the FOST. Ma and Wang (2004) presented a
The shear stiffness and shear correction factors relationship between the third order shear deformation
associated with first order shear deformation theories solutions of axisymmetric bending and buckling of FG
(FOSTs) were calculated by Nguyen et al. (2008) for circular plate with isotropic circular plates based on
FG simply supported plates under cylindrical bending. CPT. GhannadPour and Alinia (2006) studied the
Matsunaga (2009) developed higher order shear large deflection behaviour of FG plate with power-law
deformation theories (HOSTs) for displacements and distribution of the volume fraction of constituents by
stresses in FG simply supported plates subjected to using CPT. The fundamental equations for rectangular
thermal and mechanical loads. Khabbaz et al. (2009) FG plate are obtained using Von-Karman theory and
used energy concept along with first and HOSTs to solution is obtained by minimization of the potential
evaluate large deformation and through thickness energy (PE).
stresses of FG plates. Kang and Li (2009) presented In the present article, mixed semi-analytical model
non-linear behaviour of cantilever beam subjected to developed by Kant et al. (2008) has been reformulated
end force by using large and small deformation for 3D stress analysis of simply supported FG plates
theories. Analytical solutions for piezoelectric FG under transverse loads. The model use of the formation
half-spaces under uniform circular surface loading are of two-point BVP governed by coupled first-order
presented by Han et al. (2006). The effect of different ODEs along the thickness coordinate of a plate
exponential factors of the FG materials on the field (Kantrovich and Krylov 1958). In addition to this,
response is demonstrated in detail. analytical solutions based on higher order shear-
Woo and Meguid (2001) developed series solutions normal deformation theory (HOSNT) are also devel-
for large deflections of FG plates under transverse oped and presented.

123
Static solutions for functionally graded simply supported plates 53

8 9 2 38 9
2 Formulations > rx > C11 C12 C13 0 0 0 > ex >
>
> >
> >
> >
>
> >
> 6 0 7 > ey >>
>
>
r y >
> 6
C22 C23 0 0 7>> > >
<r >
> = 6 7> >
> >
A FG plate (Fig. 1), simply supported on all its four z 6 C33 0 0 0 7< ez =
¼6 7
edges is considered. A right-handed orthogonal co- >
> sxy > 6
> C44 0 0 77>> c >
>
> >
>
6
6 7>> xy >>
>
ordinate system (x, y, and z) is chosen such that the >
> s >
> 4 symmetric C55 0 5> >
> c > >
>
> xz > > xz >
: > ; : >
;
plate occupies a domain X in the x–y plane and z-axis syz C66 cyz
is normal to the plane. The top surface of a plate is Or
loaded only with transversely distributed load and it r ¼ Cij e
 
can be expressed as,
X ð3Þ
mpx npy
pðx; yÞ ¼ p0mn sin sin ð1Þ
a b where,
mn

where, m and n = 1,3,5,……. and other surfaces are E0 ekz ð1  t2 Þ

C11 ¼ C22 ¼ C33 ¼
free from any stresses. The 3D equations of equilib- ð1  3t2  2t3 Þ
rium are, E0 ekz ðt þ t2 Þ
C12 ¼ C13 ¼ C23 ¼
orx osxy osxz ð1  3t2  2t3 Þ
þ þ þ Bx ¼ 0 E0 ekz
ox oy oz C44 ¼ C55 ¼ C66 ¼
osyx ory osyz 2ð1 þ tÞ
þ þ þ By ¼ 0 ð2Þ
ox oy oz
and k¼  ln EE0h is the Gradation factor, Eo is the
oszx oszy orz
þ þ þ Bz ¼ 0 Young’s modulus at the bottom of the beam, Eh is
ox oy oz
the Young’s modulus at the top of the beam, t is the
where, Bx, By and Bz are the body forces per unit Poison’s ratio and, general 3D linear strain–displace-
volume in x, y and z directions, respectively. ment relations are,
It is assumed that the Poisson’s ratio is constant ou ou ov
through the thickness and variation of Young’s ex ¼ cxy ¼ þ
ox oy ox
modulus through the plate thickness is given by ov ou ow
EðzÞ ¼ Eo ekz (k = gradation factor). Further, it is ey ¼ cxz ¼ þ ð4Þ
oy oz ox
assumed here that FG material is isotropic at every ow ov ow
point. Therefore, constitutive relations for FG plate ez ¼ cyz ¼ þ
oz oz oy
can be written as,
2.1 Mixed semi analytical model

An attempt is made in this section to extend the semi

y
analytical model developed by Kant et al. (2008) for
z stress analysis of FG plate under transverse loads. The
models is based on the formation of two-point BVP
h governed by a set of first-order ODEs,
P(x,y)
Eh d
yðz) = Aðz)yðz) + pðz) ð5Þ
dz
E0 e λ z in the domain 0\z\h with any half of the primary
b
unknowns prescribed at the top and bottom surface of
a plate. In Eq. 5, yðzÞ is an n-dimensional vector of
E0
fundamental variables whose number (n) equals the
x
a
order of PDE, AðzÞ is a n  n coefficient matrix (which
is a function of material properties in thickness
direction) and pðzÞ is a n-dimensional vector of non-
Fig. 1 3D plate domain subjected to transverse loads homogenous (loading) terms.

123
54 S. S. Pendhari et al.

The Eqs. 2–4 have a total of fifteen unknowns functions, the following ODEs are obtained,
u; v; w; ex ; ey ; ez ; cxy ; cxz ; cyz ; rx ; ry ; rz ; sxy ; sxz and syz dumn ðzÞ mp 1
in fifteen equations. After a simple algebraic manip- ¼  wmn ðzÞ þ sxzmn ðzÞ
dz a C55
ulation of the above sets of equations, a set of PDEs dvmn ðzÞ np 1
involving only six primary dependent variables ¼  wmn ðzÞ þ syzmn ðzÞ
dz b C66
u; v; w; sxz ; syz and rz are obtained as follows, dwmn ðzÞ C31 mp C32 np
  ¼ umn ðzÞ þ vmn ðzÞ
ou ow C66 dz C33 a C33 b
¼ þ sxz 1
oz ox C55 C66 þ rzmn ðzÞ
  C33
ov ow C55  
¼ þ syz dsxzmn ðzÞ C13 C31 m2 p2
oz oy C55 C66 ¼ C11  umn ðzÞ
  dz C33 a2
ow rz 1 ou ov  
¼  C31 þ C32 n2 p 2 C13 C32 mnp2
oz C33 C33 ox oy þ C44 2 umn ðzÞ þ C12 þ C44  vmn ðzÞ
  2 b C33 ab
osxz C13 C31 o u C13 mp
¼  C11   rzmn ðzÞ  Bx ðx;y;zÞ
oz C33 ox2 C33 a
   
o2 u C13 C32 o2 v C13 orz dsyzmn ðzÞ C23 C31 mnp2
 C44 2  C12 þ C44    Bx ¼ C21 þ C44  umn ðzÞ
oy C33 oxoy C33 ox dz C33 ab
 
  m2 p2 C23 C32 n2 p2
osyz C23 C31 o2 u þ C44 2 vmn ðzÞ þ C22  vmn ðzÞ
¼  C21 þ C44  a C33 b2
oz C33 oxoy
  C23 np
2
o u C23 C32 o2 v C23 orz  rzmn ðzÞ  By ðx;y;zÞ
 C44 2  C22    By C33 b
ox C33 oy2 C33 oy drzmn ðzÞ mp np
¼ sxzmn ðzÞ þ syzmn ðzÞ  Bz ðx;y;zÞ
orz osxz osyz dz a b
¼   Bz ð6Þ
oz ox oy ð9Þ
The above PDEs defined by Eq. 6 can be further Further, the secondary variables rx ;ry and sxy can be
reduced to a coupled first-order ODEs by using double expressed as a function of the primary set of variables
Fourier trigonometry series expansion for primary as follows,
displacement variables satisfying the simply support   hmpi
C13 C31 X mpx npy
end conditions on all four edges. rx ¼ C11  umn ðzÞ sin sin
C33 a a b
X mpx npy
mn

uðx; y; zÞ ¼ umn ðzÞ cos sin  X hnpi

C13 C32 mpx npy
mn
a b þ C12  vmn ðzÞ sin sin
X C33 b a b
mpx npy mn
vðx; y; zÞ ¼ vmn ðzÞ sin cos ð7Þ
a b C13 X mpx npy
mn þ rzmn ðzÞ sin sin
X mpx npy C33 mn a b
wðx; y; zÞ ¼ wmn ðzÞ sin sin  X
a b C23 C31 hmpi mpx npy
mn
ry ¼ C21  umn ðzÞ sin sin
C33 mn
a a b
and from the basic relations of theory of elasticity, it  X
C23 C32 hnpi mpx npy
can be shown that, þ C22  vmn ðzÞ sin sin
X mpx npy C33 mn
b a b
sxz ðx; y; zÞ ¼ sxzmn ðzÞ cos sin
mn
a b C23 X mpx npy
þ rzmn ðzÞ sin sin
X mpx npy C33 mn a b
syz ðx; y; zÞ ¼ syzmn ðzÞ sin cos ð8Þ hnpi
a b X mpx npy
mn sxy ¼ C44 umn ðzÞ cos cos
X mpx npy b a b
mn
rz ðx; y; zÞ ¼ rzmn ðzÞ sin sin hmpi
mn
a b X mpx npy
þ C44 vmn ðzÞ cos cos
mn
a a b
Substituting Eqs. 7–8 and its derivatives into Eq. 6
and noting orthogonality conditions of trigonometric ð10Þ

123
Static solutions for functionally graded simply supported plates 55

Load term
The Eq. 9, defines the governing two-point BVP in

Include

Include
ODEs through thickness of the laminate in the domain

Delete
Delete
Delete
0 \ z \ h with stress components known at the top
and bottom faces. The basic approach to the numerical
integration of the BVP defined in Eq. 9 and the

p (x, y)
associated boundary conditions when it contains no

Y61
Y62
Y63
Y64
rz
boundary layer effects, is to transform the given BVP
Y51 into a set of IVPs—one non-homogeneous and n/2
Y52
Y53
Y54
homogeneous. The solution of BVP defined by Eq. 9
syz

0
is then obtained by forming a linear combination of
one non-homogeneous and n/2 homogeneous solu-
Y41
Y42
Y43
Y44
sxz

0 tions so as to satisfy the boundary conditions at z = h

(Kant and Ramesh 1981). This gives rise to a system of
Y31
Y32
Y33
Y34
wT

n/2 linear algebraic equations, the solutions of which

w
Final edge; z = h

determines the unknown n/2 components, X1, X2 and

X3 (Table 1) at the starting edge z ¼ 0. Then a final
Y21
Y22
Y23
Y24
vT
v

numerical integration of Eq. 9 produces the desired

results. Numbers of successful and well-tested numer-
Y11
Y12
Y13
Y14
uT

ical algorithms are available in literature for solution

u

of IVPs expressed by ODEs. Displacement and stress

0 (known)

boundary conditions on all four edges of a 3D FG plate

Known

are detailed in Table 2.

rz

0
0
0

2.2 Analytical models based on HOSNT

0 (known)

In order to reduce 3D elasticity problem to a 2D plate

Known

problem, the displacement components in all three

syz

0
0
0

direction of rectangular Cartension co-ordinate system

uðx; y; zÞ; vðx; y; zÞ; and wðx; y; zÞ at any point in the
0 (known)

plate domain are expanded in Taylor’s series in terms

Known

of thickness co-ordinate. The higher order displace-

sxz

ment fields considered here in the formulation are,

0
0
0
0 (assumed)
0 (assumed)
0 (assumed)

uðx; y; zÞ ¼ uo ðx; yÞ þ zhx ðx; yÞ þ z2 uo ðx; yÞ þ z3 hx ðx; yÞ

1 (unity)

vðx; y; zÞ ¼ vo ðx; yÞ þ zhy ðx; yÞ þ z2 vo ðx; yÞ þ z3 hy ðx; yÞ

X3
w

wðx; y; zÞ ¼ wo ðx; yÞ þ zhz ðx; yÞ þ z2 wo ðx; yÞ þ z3 hz ðx; yÞ

Table 1 Transformation of a BVP into IVPs

ð11Þ
0 (assumed)
0 (assumed)

0 (assumed)
1 (unity)

Table 2 Boundary conditions (BCs)

Edge BCs on BCs on
Starting edge; z = 0

X2
v

displacement field stress field

x = 0, a w=0 rx ¼ 0
0 (assumed)

0 (assumed)
0 (assumed)

sxz ¼ 0
1 (unity)

x = a/2 u=0
x = 0, b w=0 ry ¼ 0
X1

syz ¼ 0
u

x = b/2 v=0
z=0 – rz ¼ 0; sxz ¼ 0; syz ¼ 0
Final
Intg.

z=h – rz ¼ p0mn ; sxz ¼ 0; syz ¼ 0

1
2
3
4

123
56 S. S. Pendhari et al.

In the above relations, the terms u; v and w are the 2.2.1 Stress resultant
displacement at a general point (x, y, z) in the plate
domain in the x, y and z directions, respectively. The It is convenient to integrate the stress distribution
parameters uo ; vo are the inplane displacement and wo through the thickness of the FG plate to replace the
is the transverse displacement at a point (x, y) on usual consideration of stresses by considering stress-
reference plane. The functions hx ; hy are the rotations resultants due to which the variations with respect to
of the normal to the reference plane about y and x axes, ‘z’ direction are completely eliminated. Here, the
respectively. The parameters u0 ; v0 ; wo ; hx ; hy and hz membrane, flexure and shear stress resultants of FG
are the higher order terms in Taylor’s series expansion plate are derived as a function of the reference plane
and they represent higher order transverse cross stretching, curvature and shear rotation strain terms,
sectional deformation modes. The terms u0 and v0 respectively.
Q
contribute to higher order inplane modes of deformation The total PE of the FG plate with volume V,
Q
while the terms hx and hy contribute to higher order reference surface A can be written as: ¼UW
flexural modes of deformation. The terms wo and hz Y 1Z t Z t
defines the non-linear variation of the transverse strain ez . ¼ e r dV  d p dA ð13Þ
2   
By substitution of the displacement relations given by V A
Eq. 11 into the strain–displacement Eq. 4 of the classical
where, U is the strain energy stored in the plate, W
theory of elasticity, the following relations are obtained. represents the work-done by externally applied loads
ex ¼ exo þ zvx þ z2 exo þ z3 vx and p is the vector of surface load intensities
ey ¼ eyo þ zvy þ z2 eyo þ z3 vy corresponding to the generalized displacement vector
d defined at the reference plane and these can be
ez ¼ ezo þ zvz þ z2 ezo
ð12Þ expressed as,
cxy ¼ exyo þ zvxy þ z2 exyo þ z3 vxy
cxz ¼ /x þ zvxz þ z2 /x þ z3 vxz r ¼ðrx ; ry ; rz ; sxy ; sxz ; syz Þt


cyz ¼ /y þ zvyz þ z 2
/y þ z3 vyz e ¼ðex ; ey ; ez ; cxy ; cxz ; cyz Þt


where, d ¼ðu; v; wÞt

  
  ouo ovo ouo ovo
exo ; eyo ; exyo ¼ ; ; þ p ¼ðpx ; py ; pz Þt
ox oy oy ox 
   
   ouo ovo ouo ovo
  
The expressions for the strain components (Eq. 12) are
exo ; eyo ; exyo ¼ ; ; þ
ox oy oy ox substituted in Eq. 13. The following relation results
ðezo Þ ¼ ðhz Þ when an explicit integration is carried out through the
     plate thickness.
ezo ¼ 3hz
  Y Z Z
  ohx ohy ohx ohy 1 t dt p dA
vx ; vy ; vxy ¼ ; ; þ ¼  dA 
e r ð14Þ
ox oy oy ox 2
  oh oh oh oh  A A
y y
vx ; vy ; vxy ¼ x
; ; xþ in which,
ox oy oy ox
 
  ohz ohz
vz ; vxz ; vyz ¼ 2wo ; 2uo þ ; 2vo þ  ¼ ðNx ; Mx ; Nx ; Mx ; Ny ; My ; Ny ; My ; Nz ; Mz ; Nz ;
r
ox oy
 oh oh  ; Qx ; Sx ; Qx ; Sx ; Qy ; Sy ; Qy ; Sy Þt
 
 Nxy ; Mxy ; Nxy ; Mxy
vxz ; vyz ¼ z
; z e ¼ ðexo ; vx ; exo ; vx ; eyo ; vy ; eyo ; vy ; ezo ; vz ; ezo ; exyo ;
ox oy
 
  owo owo vxyo ; exyo ; vxyo ; /x ; vxz ; /x ; vxz ; /y ; vyz ; /y ; vyz Þt
/ x ; / y ¼ hx þ ; hy þ
ox oy d ¼ ðuo ; vo ; wo ; hx ; hy ; hz ; u ; v ; w ; h ; h ; h Þt
   ow 
ow
 o o o x y z

/x ; /y ¼ 3hx þ o ; 3hy þ o where,

ox oy

123
Static solutions for functionally graded simply supported plates 57

2 8 9 3 8 9 8 9
Nx Mx Nx Mx
Zh > r Nxy > exyo >
< x> = >
> > >
> >
6N   7 > >
< Mxy = <v >
> =
4 y My Ny M y 5 ¼ ry 1 z z2 z3 dz xyo
>
: ; > ¼ ½ B 
NzMz Nz 0 0 rz >
 >
> Nxy >
4x4
>
> exyo >
>
2 38 9 :  >
> ; :  >
> ;
Z h C11 C12 C13 > ex > Mxy 4x1 vxyo 4x1
< =
6 7 8 9 8 9
¼ 4 C22 C23 5 ey 1 z z2 z3 dz Qx > /x >
>
: ; > >
> > >
>
> =
< > > >
< >
symmetry C33 ez vxz =
o
Sx
Zh ¼ ½ D  and ð15Þ
  >
> Qx >
>
4x4
>
> /x >>
Nxy Mxy 
Nxy 
Mxy ¼ sxy 1 z z 2 3
z dz >
: ; > >
:  ; >
o
Sx 4x1 vxz 4x1
8 9 8 9
Zh > Q y> > / y>
 > >
> > > >
> >
¼ C44 cxz 1 z z2 z3 dz < Sy = < vyz =
¼ ½ E 
o
>
> Qy >
>
4x4
>
> /y >>
Zh : >
> ; :  >
> ;
 Sy 4x1 vyz 4x1
½ Qx Sx Qx Sx ¼ sxz 1 z z2 z3 dz
o
where [A], [B], [D] and [E] are the matrices of FG plate
Zh stiffness whose elements are defined in Appendix. The

¼ C55 cxz 1 z z2 z3 dz stress resultants with * as a superscript represent
o
higher order quantities because of the higher order
Zh terms in the displacement fields (Eq. 11).
 
Qy Sy Qy Sy ¼ syz 1 z z2 z3 dz
o

Zh
 2.2.2 Equilibrium equations and boundary conditions
¼ C66 cyz 1 z z2 z3 dz
o
The governing equations of equilibrium for the stress
Upon integration, these expressions are rewritten in analysis are obtained using the principle of minimum
the matrix form as given below, potential energy (PMPE), which states that for equi-
librium, the total PE must be stationary. In analytical
8 9 8 9 form it can be written as follows,
> Nx > >
> exo >>  
>
> >
> >
> >
> d U  Ws  Wex  Wey ¼ 0
>
> Mx >
> >
> vx > >
>
> >
> >
> >
>
>
> > >  >
>  > > exo >
>
> Nx >
>
>
>
>
>
>
>
>
The individual terms of the above equation are
>
> >
> >
>  >
>
>  > > vx >> evaluated as follows
>
>
> Mx >
>
>
>
>
>
>
>
>
>
>
> Z Z Z
>
> >
> >
> eyo >> 
>
> Ny >
> >
> >
> dU ¼ rx dex þ ry dey þ rz dez þ sxy dcxy
< = < =
¼ ½ A11x11 vy x y z
>
> My >
> >
> >
>
>
> >
> >
> > þ sxz dcxz þ syz dcyz Þ dx dy dz ð16Þ
>
> Ny>
> >
> eyo >
>
>
>
> >
> >
> >
>
>
> My>
> >
> >
 >
>
> >
> >
> vy >> Substituting the appropriate strain expressions using
>
> >
> >
> >
>
>
> Nz >
> >
> ezo >> Eq. 12 and integrating through the thickness to get the
>
> >
> >
> >
>
>
> > > > stress resultants as defined in Eq. 15 and integrating
>
> Mz >
>
>
>
>
> vz > >
>
>
: >
; >
> >
> the resulting expressions by parts transforms the
Nz 11x1 :  ;
ezo 11x1 Eq. 16 into the following form

123
58 S. S. Pendhari et al.

I
where the bars on the quantities refer to edge values.
dU ¼ ½Nyx duo þ Ny dvo þ Qy dwo þ Myx dhx
On integration through the thickness the variation of
x

þ My dhy þ Sy dhz þ Nyx duo þ Ny dvo þ Qy dwo these expressions take the form
I
þ Myx dhx þ My dhy þ Sy dhz dx
    
I dWex ¼ ðNx duo þ Nxy dvo þ Qx dwo þ M  x dhx
þ ½Nx duo þ Nxy dvo þ Qx dwo þ Mx dhx þ Mxy dhy
y
y
þM xy dhy þ Sx dhz þ Nx duo þ Nxy

dvo
þ Sx dhz þ Nx duo þ Nxy
dvo þ Qx dwo þ Mx dhx
þ Mxy  
dhy þ Sx dhz dy  x dhx þ M
þ Qx dwo þ M 
xy dhy þ Sx dhz Þ dy ð22Þ
Z Z
   
oNx oNyx oNy oNxy
 þ duo þ þ dvo and
ox oy oy ox
x y
   
I
þ
oQx oQy
þ dwo þ
oMx oMyx
þ  Qx dhx dWey ¼ ðNyx duo þ Ny dvo þ Qy dwo þ M
 yx dhx
 ox oy  ox oy
x
oMy oMxy 
þ
oy
þ
ox
 Qy dhy y dhy þ Sy dhz þ Nyx
þM duo þ Ny dvo þ Qy dwo
    
oSx oSy oNx oNyx    dh þ M
þM   dh þ S dh Þ dx ð23Þ
þ þ  Nz dhz þ þ  2Sx duo yx x y y y z
ox oy ox oy
      
oNy oNxy  oQx oQy The variational Eq. 17 takes the following form when
þ þ  2Sy dvo þ þ  2Mz dwo
oy ox ox oy the relevant foregoing expressions are substituted for
       
oMx oMyx oM y oMxy its individual terms.
þ þ  3Qx dhx þ þ  3Qy dhy
ox oy oy ox Z Z "   
    oNx oNyx oNy oNxy
oSx oSy   þ duo þ þ dvo
þ þ  3Nz dhz dxdy ox oy oy ox
ox oy x y
   
ð17Þ oQx oQy oMx oMyx
þ þ þ pþz dw o þ þ  Q x dhx
ox oy ox oy
   
Work done by externally applied load can be calcu- oMy oMxy oSx oSy
þ þ  Qy dhy þ þ  Nz þ hðpþ
z Þ dhz
lated by,  
oy ox

ox
 
oy

Z Z oNx oNyx 
oN y

oNxy
dWs ¼ ðpþ þ þ þ  2Sx duo þ þ  2Sy dvo
z dw Þ dx dy ð18Þ ox oy oy ox
  
x y oQx oQy
þ þ  2Mz þ h2 ðpþ z Þ dwo

ox oy
where, wþ is the transverse displacement at any point      
oMx oMyx oMy oMxy 
þ þ  3Qx dhx þ þ  3Qy dhy
at the top surface of the plate and is given by ox oy oy ox
   #
oSx oSy
wþ ¼ wo þ hþ hz þ hþ2 wo þ hþ3 hz þ þ  3Nz þ h3 ðpþ 
z Þ dhz dx dy
ox oy
I h
Therefore after further simplification for h
þ ðNyx  Nyx Þduo þ ðNy  Ny Þdvo þ ðQy  Qy Þdwo
Z Z
x
dWs ¼ fðpþ þ þ 2 
z Þdwo þ ðpz hÞdhz þ ðpz h Þdwo þ ðM yx  Myx Þdhx þ ðM y  My Þdhy
 
x y þ ðSy  Sy Þdhz þ ðNyx  Nyx Þduo þ ðNy  Ny Þdvo

þðpþ 3
z h Þdhz g dx dy ð19Þ   
þ ðQy  Qy Þdwo þ ðM 
yx  Myx 
Þdhx
I h
i
þðMy  My Þdhy þ ðSy  Sy Þdhz dx þ ðNx  Nx Þduo
The work done by the edge stresses is given by
Z Z y
1   þ ðNxy  Nxy Þdvo þ ðQx  Qx Þdwo þ ðM  x  Mx Þdhx
Wex ¼ rx u þ sxy v þ sxz w dy dz
2 þ ðM xy  Mxy Þdhy þ ðSx  Sx Þdhz þ ðNx  Nx Þduo
y z    x  Mx Þdhx
þ ðNxy  Nxy Þdvo þ ðQx  Qx Þdwo þ ðM
i
on an edge x ¼ constant ð20Þ þ ðM 
 xy 
 Mxy Þdhy þ ðSx  Sx Þdhz dy ¼ 0
Z Z
1  
Wey ¼ sxy u þ ry v þ syz w dx dz ð24Þ
2
x z The above Eq. 24 will be an identity only if each of the
on an edge y ¼ constant ð21Þ coefficients of the arbitrary variation vanishes. The

123
Static solutions for functionally graded simply supported plates 59

vanishing of the surface integral defines twelve and, on the edge y = constant
equilibrium equations, while that of the line integrals
uo ¼ uo Nxy ¼ N xy vo ¼ vo Ny ¼ N y
defines the consistent natural boundary conditions that
are to be used with this theory along the two edges. wo ¼ wo Qy ¼ Qy hx ¼ hx Mxy ¼ M xy
Setting the individual integral terms in Eq. 24 to zero,
hy ¼ hy My ¼ M y hz ¼ hz Sy ¼ Sy
the following equations of equilibrium and the con-
or  and or 
sistent boundary conditions are obtained. uo ¼ uo 
Nxy ¼ N xy vo ¼ vo Ny ¼ N y
oNx oNxy  
duo : þ ¼0 wo ¼ wo 
Qy ¼ Qy hx ¼ hx 
Mxy ¼ M xy
ox oy

oNy oNxy hy ¼ hy My ¼ M y

hz ¼ hz
 
Sy ¼ Sy
dvo : þ ¼0
oy ox
ð27Þ
oQx oQy
dwo : þ þ pþ
z ¼0
ox oy 2.2.3 Closed-form solution
oMx oMxy
dhx : þ  Qx ¼ 0
ox oy Navier’s solution technique using the Fourier series
oMy oMxy is used to obtain closed form solution of the 2D
dhy : þ  Qy ¼ 0 plate problem. All displacements and loads acting
oy ox
oSx oSy on the FG plate are defined in terms of Fourier
dhz : þ  Nz þ hðpþ z Þ ¼0 series. The equilibrium equations are solved for
ox oy
oN  oNxy  ð25Þ displacement amplitudes by substituting stress resul-
duo : x þ  2Sx ¼ 0 tants in terms of displacements expanded in Fourier
ox oy
oN 
oN  series.
y xy
dvo : þ  2Sy ¼ 0 For the simply supported boundary conditions, viz.,
oy ox
 At edges x = 0 and x = a
oQ oQy
dwo : x þ  2Mz þ h2 ðpþz Þ ¼0
ox oy vo ¼ 0 ; wo ¼ 0 ; hy ¼ 0 ; hz ¼ 0 ; Mx ¼ 0 ;

oMx oMxy
dhx : þ  3Qx ¼ 0 vo ¼ 0 ; wo ¼ 0 ; hy ¼ 0 ; hz ¼ 0 ; Mx ¼ 0 ;
ox oy
oMy oMxy 
Nx ¼ 0 ; Nx ¼ 0:
dhy : þ  3Qy ¼ 0
oy ox
ð28Þ
oS oSy
dhz : x þ  3Nz þ h3 ðpþz Þ ¼0 At edges y = 0 and y = b
ox oy
The boundary conditions, on the edge x = constant uo ¼ 0 ; wo ¼ 0 ; hx ¼ 0 ; hz ¼ 0 ; My ¼ 0 ;
uo ¼ uo Nx ¼ N x vo ¼ vo Nxy ¼ N xy uo ¼ 0 ; wo ¼ 0 ; hx ¼ 0 ; hz ¼ 0 ; My ¼ 0 ;
wo ¼ wo Qx ¼ Qx hx ¼ hx Mx ¼ M x
Mxy ¼ M xy Sx ¼ Sx Ny ¼ 0 ; Ny ¼ 0:
hy ¼ hy hz ¼ hz
  or   and   or   ð29Þ
uo ¼ uo Nx ¼ N x vo ¼ vo Nxy ¼ N xy
 
wo ¼ wo Qx ¼ Qx hx ¼ hx Mx ¼ M x

 
To satisfy the above boundary conditions the general-
hy ¼ hy 
Mxy ¼ M xy

hz ¼ hz

Sx ¼ Sx ized displacement field is expanded in terms of double
ð26Þ Fourier series as:

123
60 S. S. Pendhari et al.

1 X
X 1
mpx npy 1 X
X 1
mpx npy
uo ¼ uomn cos sin hx ¼ hxmn cos sin
m¼1 n¼1
a b m¼1 n¼1
a b
X1 X 1
mpx npy X1 X 1
mpx npy
vo ¼ vomn sin cos ; hy ¼ hymn sin cos
m¼1 n¼1
a b m¼1 n¼1
a b
1 X
X 1
mpx npy X1 X 1
mpx npy
wo ¼ womn sin sin hz ¼ hzmn sin sin
m¼1 n¼1
a b m¼1 n¼1
a b
X1 X 1 X1 X 1 ð30Þ
mpx npy mpx npy
uo ¼ uomn cos sin hx ¼ hxmn cos sin
m¼1 n¼1
a b m¼1 n¼1
a b
1 X
X 1
mpx npy 1 X
X 1
mpx npy
vo ¼ vomn sin cos and hy ¼ hymn sin cos
m¼1 n¼1
a b m¼1 n¼1
a b
1 X
X 1
mpx npy X1 X 1
mpx npy
wo ¼ womn sin sin hz ¼ hzmn sin sin
m¼1 n¼1
a b m¼1 n¼1
a b

8 9
where uon ; hxn ; uon ; . . .. . .. . .: are called as displace- > ouo ovo >
>
> þ >
>
ment Fourier amplitudes. Only odd values of m and >
> oy ox >
>
>
> >
>
n = 1, 3, 5,…. are taken. 8 9 >
> >
> Nxy > >
> ohy ohx >
>
>
> > > >
Further, substitution of Eq. 12 into stress resultant < Mxy >
> = < oy þ ox >
> =
relationship (Eq. 15), relation between the stress ¼ ½B4x4 ;

resultant and the reference plane strain quantities can > Nxy >
> > >
> ouo ovo > >
>
:  ; > >
> þ >
>
be obtained. Mxy 4x1 >
> oy ox > >
>
> >
>
8 9 >
> >>
ouo > >
> oh  oh >
>
>
> > >
: xþ y >
;
> >
>
>
> ox > >
>
oy ox 4x1
>
> >
> 8 9
>
> oh >
> > owo >
> x > >
> h þ >
> ox >
>
> >
> 8 9
> x
>
> ox > >
>
>
>
> >
> > >
>
>  >> > Q > >
> oh z>
> ð31Þ
8 9 > ou > >
>
x > >  >
> Nx > >
>
>
o >
>
> <S > = >
< 2uo þ >
=
>
> M > >
> >
> ox >
> x ox
>
> > >
> >
> ¼ ½ D 4x4 and
>
>
x >
> > oh >
>  > >
> Qx >
> >
>  owo > >
> > > x > > > > >
>
>
>

Nx > >
>
>
>
>
>
>
>
: ;
Sx 4x1 > 3hx þ ox >
>
> >
>
>
> >
> >
> ox >
> >
> >
>
>
>  >> >
> >
> >
> >>
>
> M >
> >
> ov >
> >
: oh z >
;
>
>
x >
> >
> o >
>
>
> >
> >
> oy >
> ox 4x1
< Ny = < = 8 9
¼ ½ A11x11 ; 8 9 >
> hy þ ow o
oy > >
> My > > ohy > > Q y > > >
> > > > >
> > > >
> > > > < Sy >= >
< 2vo þ oh z >=
>
>
>
>
Ny > >
>
>
>
>
>
> oy > >
>
>
oy
>
>  >> >
> >
>  ¼ ½E4x4  ow 
> My > >  > > Q > > o >
>
> >
> >
> ov >
> > y>
>
: > > 3hy þ oy >
> >
>
> >
> >
> o >
>  ; >
> >
>
>
> Nz >
> >
> oy >
> Sy 4x1 : 
ohz ;
>
> >
> >
> >
> oy
> M > > > 4x1
>
:  >
> z >
;
>
>
>  >>
>
>
> oh y >
> where,
Nz 11x1 >
> >
>
>
> oy >
> 
>
> >
> ouo ohx ouo ohx T X 1 X 1
>
> >
> ¼
>
> h z >
>
>
> >
> ox ox ox ox
> > m¼1 n¼1
> 2wo >
>
>
 >
>
n o
>
> >
> auomn  ahxmn  auomn  ahomn sin a x sin by
: ;
3hz 11x1

123
Static solutions for functionally graded simply supported plates 61

  T 1 X
X 1 After following the above two steps with use of
ovo ohy ovo ohy
¼ orthogonality conditions for trigonometric functions
oy oy oy oy m¼1 n¼1 and collecting the displacement coefficients, one
n o obtains:
buomn  bhxmn  buomn  bhomn
sin a x sin b y 8 9 8 9
1 n o >
> uomn >
> > 0 >
   T X 1 X
>
> >
> >
> >
>
hz w o hz ¼ hzmn womn hzmn sin a x sin b y >
>
> vomn >
>
>
>
>
> 0
>
>
>
>
> womn >
> >
> >
>

m¼1 n¼1
>
> >
> >
> p þ >
>
>
> hxmn >
> >
> z >
>
ouo ohx ouo ohx T X 1 X 1
>
> >
> >
> 0 >
>
¼ >
> >
> >
> >
>
oy oy oy oy >
> hymn >
> >
> 0 >
>
m¼1 n¼1 >
> >
> >
> >
>
n o < hzmn = < hðpþz Þ
=
 
buomn bhxmn buomn bhxmn cos a x sin by ½ X 12x12 uomn ¼ 0 ð32Þ
>
> >
> >
> >
>
  T >
> vomn >
> >
> >
>
ovo ohy ovo ohy X1 X 1 >
> >
> >
> 0 >
>
¼ >
> womn >
> >
> >
>
>
> >
> >
> 2 þ
ðp >
>
ox ox ox ox >
> >
> >
>
h z Þ >
>
n
m¼1 n¼1
o >
> hxmn >
> >
> >
>
  >
> >
> >
> 0 >
>
avomn ahymn avomn ahymn cos a x sin by >
>  >
> >
> >
>
> hymn
> >
> >
> 0 >
>
1 n o >
:  >
; : 3 þ ;
   T X 1 X
hzmn h ðpz Þ 12x1
hx uo hx ¼ hxmn uomn hxmn 12x1
m¼1 n¼1
cos a x sin b y for any fixed value of m and n. The matrix [X] is
 T X 1 X
owo ohz owo ohz 1 the coefficient matrix whose elements are listed in
¼ Appendix.
ox ox ox ox m¼1 n¼1
n o The Fourier amplitudes are obtained by solving
 
awomn ahzmn awomn ahzmn cos a x sin b y Eq. 32. The Fourier displacement amplitudes are then
h iT X
1 X 1 n o used to calculate the generalized displacement com-
hy vo hy ¼ hymn vomn hymn sin a x cos b y ponents and their derivatives. The values of general-
m¼1 n¼1 ized displacement components and their derivatives

owo ohz owo ohz T are then substituted in Eq. 15 to obtain the values of
oy oy oy oy stress resultants. The same displacement values are
X1 X1 n o also back substituted into the strain–displacement
¼ bwomn bhzmn bwomn bhzmn sin a x cos b y relations (Eq. 4) to obtain the values of strain. The
m¼1 n¼1
material constitutive relations (Eq. 3) are then used to
compute the inplane and transverse stresses using 3D
in which, a ¼ mp
a and b ¼ b
np equilibrium equations (Eq. 2).
The following steps are taken to obtain the required
system of equilibrium equations (Eq. 25) in terms of
displacements. 2.2.4 Evaluation of transverse stresses

1. Equations 28–31 are substituted in Eq. 25. The transverse stresses (sxz ; syz and rz ) cannot be
2. The twelve equilibrium equations are multiplied accurately estimated by constitutive relations
npy npy npy
with cos mpx mpx mpx
a sin b ; sin a cos b ; sin a sin b ; (Eq. 3). Here, the transverse stresses are obtained by
npy npy npy
cos a sin b ; sin a cos b ; sin a sin b ; cos mpx
mpx mpx mpx
a integrating the 3D equilibrium equations of elasticity
sin npy mpx npy mpx npy mpx npy
b ; sin a cos b ; sin a sin b ; cos a sin b ; (Eq. 2) for each layer over the plate thickness.
npy npy
sin mpx mpx
a cos b and sin a sin b respectively and After performing numerical integration along the
then integrated between the limits 0 \ x \ a and thickness of a FG plate, first-order and second-order
0 \ y \ b. differential equations are obtained.

123
62 S. S. Pendhari et al.

Zz    3 
orx osxy o uo o3 hx o3 u o3 h
sxz ¼  þ dz þ C1 ð33Þ rz ¼ s1 3 þ s2 3 þ s3 3o þ s4 3x
ox oy ox ox ox ox
!
o 3
o vo 3
o hy 3 
o vo o3 hy
þ t s1 2 þ s2 2 þ s3 2 þs4 2
Zz   ox oy ox oy ox oy ox oy
ory osxy !
syz ¼  þ dz þ C2 ð34Þ o2 hz o2 w o2 h
oy ox þ t s1 2 þ 2s2 2o þ 3s3 2z
o ox ox ox
0 1  
Zz Zz   o3 uo o3 hx o3 uo o3 hx
o2 rx o2 ry o2 sxy þ t s1 þ s2 þs3 þ s4
rz ¼ @ þ þ 2 dzAdz oxoy2 oxoy2 oxoy2 oxoy2
ox2 oy2 ox oy 3 
!
o o o3 v o o3 hy o3 v  o hy
þ s1 3 þ s2 3 þ s3 3o þ s4 3
þ zC3 þ C4 ð35Þ oy oy oy oy
!  
After substitution of Eq. 3 into above Eqs. 33–35, the o2 hz o2 wo o2 hz 1t
þ t s1 2 þ2s2 2 þ 3s3 2 þ 2
final expression for the transverse shear and normal oy oy oy 2
stresses are written as, 
o3 uo o3 hx o3 uo o3 hx
s1 þ s 2 þ s3 þ s 4
 2  oxoy2 oxoy2 oxoy2 oxoy2
o uo o2 hx o2 uo o2 hx !#
sxz ¼ t1 2 þ t2 2 þ t3 2 þ t4 2 o3 v o o3 hy o3 vo o3 hy
ox ox ox ox þ s1 2 þ s2 2 þ s3 2 þ s4 2
 2 ! ox oy ox oy ox oy ox oy
o vo o2 hy o2 vo o2 hy
þt t1 þ t2 þ t3 þt4 þ zC3 þ C4 ð38Þ
oxoy oxoy oxoy oxoy
 
ohz ow oh where,
þ t t1 þ 2t2 o þ 3t3 z
ox ox ox Ez  Eo Ez  Eo zEo
 2 t1 ¼ s1 ¼ 
1t o uo 2
o hx o2 u k k2  k

þ t1 2 þ t2 2 þ t3 2o zEz  t1 zEz Ez  Eo s1
2 oy oy oy t2 ¼ s2 ¼ 2  
!# k k k 3 k
o2 h o2 v o o2 hy o2 vo o2 hy
þt4 2x þ t1 þ t2 þ t3 þ t4 þ C1 2
z Ez  2t2 2
z Ez 2zEz
oy oxoy oxoy oxoy oxoy t3 ¼ s3 ¼ 2  3
k k k
ð36Þ  
Ez  Eo s2 z3 Ez  3t3
þ2  2 t4 ¼
k4 k k
  2  3 2
 
z Ez 3z Ez 6zEz  
Ez  Eo s3
o uo o2 hx o2 uo o2 hx s4 ¼ 2  3 þ 4  6 3
syz ¼ t t1 þ t2 þ t3 þ t4 5 k
oxoy oxoy oxoy oxoy k k k k
 2 2 2  2 
o vo o hy ov oh  ¼ Eo 2 ekz
in which, EðzÞ and Eo ¼ 1t
Eo
þ t1 2 þ t2 2 þ t3 2o þ t4 2x 1t 2
oy oy oy oy
 
ohz owo ohx
þ t t1 þ 2t2 þ 3t3
oy oy oy
 2 3 Numerical investigation
1t o uo o2 hx o2 uo
þ t1 þ t2 þ t3
2 oxoy oxoy oxoy
 A computer code is developed in FORTRAN 90 by
o2 hx o 2 vo o2 hy o2 uo o2 hx incorporating the present mixed semi-analytical and
þ t4 þ t1 2 þ t2 2 þ t3 2 þ t4 2 þ C2
oxoy ox ox ox ox HOSNT formulation for the stress analysis of rectan-
ð37Þ gular FG plates under transverse loads. Numerical

123
Static solutions for functionally graded simply supported plates 63

Table 3 Normalized stresses (rx ; ry ; sxy ; sxz ; syz ) and the transverse displacement (w)  of square FG plate under sinusoidal transverse
load
a=    
h Source r
x =ry a2 ; b2 ; h sxy ð0; 0; hÞ sxz =syz ðmaxÞ w a2 ; b2 ; h

2 Semi-analytical 0.3791 0.0833 0.2361 10.8456

HOSNT 0.3652 0.0908 0.2341 11.7217
5 Semi-analytical 0.2702 0.0832 0.2393 4.9165
HOSNT 0.2570 0.0915 0.2389 5.3749
10 Semi-analytical 0.2583 0.0823 0.2404 4.3203
HOSNT 0.2448 0.0905 0.2403 4.7407
50 Semi-analytical 0.2547 0.0820 0.2407 4.1433
HOSNT 0.2412 0.0901 0.2407 4.5528

Fig. 2 Through thickness 1.00 1.00

variation of a inplane a/h=2.0 a/h=2.0
normal stress rx ,
b transverse shear stress
sxz , c inplane shear stress sxy 0.75 0.75
and d transverse
Sem i-analytical
displacement (w)  for simply
supported square FG plate
z z HOSNT
under sinusoidal load 0.50 0.50

0.25 0.25
Sem i-analytical
HOSNT τ xz
σx
0.00 0.00
-0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.0 0.1 0.2
(a) (b)
1.00 1.00
a/h=2.0 a/h=2.0

0.75 0.75

z z
0.50 0.50

0.25 0.25
Sem i-analytical
HOSNT Sem i-analytical
τ xy HOSNT w
0.00 0.00
-0.2 -0.1 0.0 0.1 7 8 9 10 11 12
(c) (d)

123
64 S. S. Pendhari et al.

Table 4 Normalized stresses (rx ; ry ; sxy ; sxz ; syz ) and the transverse displacement (w)
 of rectangular FG plate under sinusoidal
transverse load
a=      
h Source x a2 ; b2 ; h
r y a2 ; b2 ; h
r sxy ð0; 0; hÞ sxz ðmaxÞ syz ðmaxÞ w a2 ; b2 ; h

2 Semi-analytical 0.6545 0.3872 0.1080 0.3791 0.1896 20.0471

HOSNT 0.6536 0.3461 0.1184 0.3767 0.1884 21.7327
5 Semi-analytical 0.5536 0.2916 0.1060 0.3837 0.1919 11.8021
HOSNT 0.5463 0.2562 0.1165 0.3833 0.1917 12.9252
10 Semi-analytical 0.5424 0.2796 0.1053 0.3848 0.1924 10.8808
HOSNT 0.5341 0.2447 0.1157 0.3847 0.1923 11.9460
50 Semi-analytical 0.5391 0.2759 0.1050 0.3851 0.1926 10.5999
HOSNT 0.5304 0.2412 0.1154 0.3851 0.1926 11.6477

Fig. 3 Through thickness 1.00 1.00

variation of a inplane a/h=2.0 a/h=2.0
displacement u, b inplane
normal stress rx ,
c transverse shear stress sxz 0.75 0.75
and d transverse shear stress
syz for simply supported
rectangular FG plate under
z z
sinusoidal load 0.50 0.50

0.25 0.25

Sem i-analytical Sem i-analytical

HOSNT u HOSNT σ x
0.00 0.00
-0.2 -0.1 0.0 0.1 0.2 -0.2 0.0 0.2 0.4 0.6
(a) (b)

1.00 1.00
a/h=2.0 a/h=2.0

0.75 0.75

z Sem i-analytical z Sem i-analytical

HOSNT
0.50 HOSNT 0.50

0.25 0.25

τ xz τ yz
0.00 0.00
0.0 0.2 0.4 0.0 0.1 0.2
(c) (d)

123
Static solutions for functionally graded simply supported plates 65

investigations on various simply supported rectangu- in which a bar over the variable defines its normalized
lar FG plates have been performed to establish the value. The following sets of material properties are used:
accuracy of the simplified model presented in the Material set 1 (Praveen and Reddy 1998; Reddy
preceding sections of the paper. Comparison is 2000)
demonstrated between present semi-analytical and
Eo = 70 GPa (Aluminum)
analytical solutions based on shear-normal deforma-
Eh = 151 GPa (Zirconia)
tion theories (HOSNT).
Following normalizations are used here in all t ¼ 0:3
numerical examples for the uniform comparison of the
Material set 2 (Sankar 2001)
results.
Eo = 1 GPa
a or b rx ; ry ; sxy
s¼ rx ; ry ; sxy ¼ Eh/Eo = 5, 10, 20 and 40
h p0 s 2
100Eh h w 3
sxz ; syz t ¼ 0:3
w¼ 4
sxz ; syz ¼ ð39Þ
p0 a p0 s
Illustrative examples considered in the present
Eh ðu; vÞ rz
u; v ¼ rz ¼ work are discussed next in detail.
p0 hs3 p0

Table 5 Normalized stresses (rx ; ry ; sxy ; sxz ; syz ) and the transverse displacement (w)
 of square FG plate under uniformly distributed
transverse load for different gradation factors
a=  
h Eh= r
x = ry a2 ; b2 ; h sxy ð0; 0; hÞ
E0
Semi-analytical HOSNT Semi-analytical HOSNT

2 5 0.6341 0.5903 0.0993 0.1230

10 0.7931 0.7355 0.0727 0.0899
20 0.9868 0.9119 0.0523 0.0647
40 1.2186 1.1233 0.0369 0.0457
5 5 0.5091 0.4803 0.1102 0.1232
10 0.6365 0.6000 0.0839 0.0938
20 0.7938 0.7476 0.0629 0.0702
40 0.9863 0.9282 0.0463 0.0517
10 5 0.4921 0.4654 0.1114 0.1229
10 0.6150 0.5815 0.0857 0.0945
20 0.7670 0.7250 0.0649 0.0715
40 0.9536 0.9012 0.0482 0.0532
a=  
h Eh= sxz =syz ðmaxÞ  a2 ; b2 ; h
w
E0
Semi-analytical HOSNT Semi-analytical HOSNT

2 5 0.5280 0.6479 23.1297 24.7710

10 0.5778 0.6881 31.9892 33.9580
20 0.6367 0.7541 43.6258 46.1202
40 0.6998 0.8129 58.7232 61.9529
5 5 0.5061 0.5095 11.8598 12.9185
10 0.5366 0.5509 16.9658 18.4424
20 0.5763 0.6026 24.1295 26.1991
40 0.6142 0.6353 34.0076 36.9109
10 5 0.5092 0.5179 10.5360 11.5496
10 0.5331 0.5476 15.1162 16.5607
20 0.5616 0.5739 21.6259 23.6848
40 0.6043 0.6235 30.7213 33.6431

123
66 S. S. Pendhari et al.

1.00 1.00 1.00

a/h=2.0
Semi analytical
Eh/E0=5 z
Eh/E0=10
Eh/E0=20
0.75 Eh/E0=40 0.75 0.75

z z
0.50 0.50 0.50

0.25 0.25 a/h=2.0 0.25 a/h=2.0

Semi analytical Semi analytical
Eh/E0=5 Eh/E0=5
Eh/E0=10 Eh/E0=10
Eh/E0=20 Eh/E0=20
σx Eh/E0=40 τ xz Eh/E0=40
u 0.00 0.00 0.00

-0.2 0.0 0.2 0.4 0.6 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8
(a) (b) (c)
Fig. 4 Through thickness variation of a inplane displacement u, b inplane normal stress rx and c transverse shear stress sxz for simply
supported square FG plate under uniformly distributed load for different gradation factors

3.1 Example 1 of 2 are shown in Fig. 3. The HOSNT solutions are

used for comparison.
A square FG plate with simply supported end condi-
tion on all four edges and subjected to bidirectional 3.3 Example 3
sinusoidal load has been considered to show the
effectiveness of mixed semi-analytical model over A simply supported square FG plate under uniformly
the simplified plate models. Material set 1 is used. distributed load with various gradation factor (k) is
The normalized inplane normal stress (rx , ry ), inplane considered in this example to study the effect
shear stress (sxy ), transverse shear stress (sxz ) and of gradation factor and to show the capability of
transverse displacement (w) for different aspect ratios presented models to handle the different kind of
are presented in Table 3. Moreover, through thickness loading. Material set 2 is used. The normalized inplane
variations of inplane normal stress (rx ), inplane and normal stress (rx ; ry ), inplane and transverse shear
transverse shear stresses (sxy and sxz ) as well as stress (sxy ; sxz ; syz ) and transverse displacement (w)
transverse displacement (w) for an aspect ratio of 2 are for different aspect ratios and for different gradation
depicted in Fig. 2. The analytical solutions (HOSNT) factors (k = 5, 10, 20 and 40) are detailed in Table 5.
are used for comparison. The HOSNT solutions are used for comparison.
Through thickness variations of inplane displacement
3.2 Example 2 (u), inplane normal stress (rx ) and transverse shear
stresses (sxz ) for an aspect ratio of 2 are shown in
A simply supported rectangular FG plate under Fig. 4.
bidirectional sinusoidal load has been considered
here. Material set 1 is used. The normalized
inplane normal stresses (rx , ry ), inplane and trans- 4 Concluding remarks
verse shear stress (sxy , sxz , syz ) and transverse
displacement (w) for different aspect ratios are A simple mixed semi-analytical model developed by
detailed in Table 4. Through thickness variations of Kant et al. (2008) is extended here for 3D stress
inplane displacement (u), inplane normal stresses (rx ), analysis of simply supported FG plate. A two-point
transverse shear stresses (sxz and syz ) for an aspect ratio BVP governed by a set of linear first-order ODEs is

123
Static solutions for functionally graded simply supported plates 67

formed by assuming all primary variables in the form A6;6 ¼ A2;2 A6;7 ¼ A2;3 A6;8 ¼ A2;4 A6;9 ¼ A2;5
of trigonometry functions along the inplane direction. A6;10 ¼ A2;6 A6;11 ¼ A2;7
The solutions ensure the fundamental relationship
between stress, strain and displacement fields within A7;7 ¼ A3;3 A7;8 ¼ A3;4 A7;9 ¼ A3;5
the elastic continuum. No any simplifying assump- A7;10 ¼ A3;6 A7;11 ¼ A3;7
tions are made through the thickness of a FG plate.
A8;8 ¼ A4;4 A8;9 ¼ A4;5 A8;10 ¼ A4;6
Analytical solutions based on shear-normal deforma-
tion theory (HOSNT) are also developed and pre- A8;11 ¼ A4;7
sented for comparison and to show the effectiveness, A9;9 ¼ A1;1 A9;10 ¼ A1;2 A9;11 ¼ A1;3
simplicity and accuracy of a newly developed mixed
semi-analytical model over the other simplified plate A10;10 ¼ A2;2 A10;11 ¼ A2;3
models. The main feature of mixed semi-analytical A11;11 ¼ A3;3
model is that the governing equation system is not  
transformed into an algebraic equation system, thus where, Eh ¼ Eo
1t2 ek and Eo ¼ ð1t
Eo
2 Þ and Ai,j =

the intrinsic behaviour of the physical system is Aj,i, (i, j = 1 to 11)

retained to a greater degree of accuracy. The coefficients of matrix [B] are,
1t 1t
B1;1 ¼ A1;1 B1;2 ¼ A1;2
2 2
Appendix 1t 1t
B1;3 ¼ A1;3 B1;4 ¼ A1;4
2 2
The coefficients of matrix [A] are, 1t 1t 1t
B2;2 ¼ A2;2 B2;3 ¼ A2;3 B2;4 ¼ A2;4
Eh  Eo hEh  A1;1 2 2 2
A1;1 ¼ A1;2 ¼ 1t 1t
k k B3;3 ¼ A3;3 B3;4 ¼ A3;4
h2 Eh  2A1;2 h3 Eh  3A1;3 2 2
A1;3 ¼ A1;4 ¼ 1t
k k B4;4 ¼ A4;4
A1;5 ¼ tA1;1 A1;6 ¼ t A1;2 A1;7 ¼ tA1;3 2
A1;8 ¼ tA1;4 A1;9 ¼ A1;5 and Bi,j = Bj,i (i, j = 1 to 4)
A1;10 ¼ A1;6 A1;11 ¼ A1;7 [D] and [E] matrix are same as [B] matrix.
The coefficients of vector {I} are
h4 Eh  4A2;3
A2;2 ¼ A1;3 A2;3 ¼ A1;4 A2;4 ¼ qh  qo h q  I1 h2 qh  2 I2
k I1 ¼ I2 ¼ h I3 ¼
A2;5 ¼ tA1;2 A2;6 ¼ t A2;2 A2;7 ¼ tA2;3 k1 k1 k1
3 4
A2;8 ¼ tA2;4 A2;9 ¼ A2;5 h qh  3 I3 h qh  4 I4 h5 qh  5 I5
I4 ¼ I5 ¼ I6 ¼
A2;10 ¼ A2;6 A2;11 ¼ A2;7 k1 k1 k1
6
h qh  6 I6
h5 Eh  5A3;3 I7 ¼
A3;3 ¼ A2;4 A3;4 ¼ A3;5 ¼ tA1;3 k1
k
The coefficients of matrix [X] are
A3;6 ¼ tA2;3 A3;7 ¼ t A3;3 A3;8 ¼ tA3;4 A3;9 ¼ A3;5
X1;1 ¼ A1;1 a2 þ B1;1 b2
A3;10 ¼ A3;6 A3;11 ¼ A3;7
X1;2 ¼ A1;5 ab þ B1;1 ab X1;3 ¼ 0
h6 Eh  6A4;3 X1;4 ¼ A1;2 a2 þ B1;2 b2
A4;4 ¼ A4;5 ¼ tA1;4 A4;6 ¼ tA2;4
k X1;5 ¼ A1;6 ab þ B1;2 ab
A4;7 ¼ tA3;4 A4;8 ¼ t A4;4 A4;9 ¼ A4;5
X1;6 ¼ A1;9 a
A4;10 ¼ A4;6 A4;11 ¼ A4;7
X1;7 ¼ A1;3 a2 þ B1;3 b2
A5;5 ¼ A1;1 A5;6 ¼ A1;2 A5;7 ¼ A1;3 A5;8 ¼ A1;4 X1;8 ¼ A1;7 ab þ B1;3 ab
A5;9 ¼ A1;5 A5;10 ¼ A1;6 A5;11 ¼ A1;7 X1;9 ¼ 2A1;10 a

123
68 S. S. Pendhari et al.

X1;10 ¼ A1;4 a2 þ B1;4 b2 X5;8 ¼ A6;7 b2 þ B2;3 a2 þ 2E1;2

X1;11 ¼ A1;8 ab þ B1;4 ab X5;9 ¼ 2A6;10 b þ E1;3 b
X1;12 ¼ 3A1;11 a X5;10 ¼ A6;4 ab þ B2;4 ab

X2;2 ¼ A1;5 b2 þ B1;1 a2 X5;11 ¼ A6;8 b2 þ B2;4 a2 þ 3E1;3

X2;3 ¼ 0 X5;12 ¼ 3A6;11 b þ E1;4 b
X2;4 ¼ A5;2 ab þ B1;2 ab X6;6 ¼ D2;2 a2 þ E2;2 b2 þ A9;9
X2;5 ¼ A5;6 b2 þ B1;2 a2 X6;7 ¼ A9;3 a þ 2D2;2 b
X2;6 ¼ A5;9 b X6;8 ¼ A9;7 b þ 2E2;2 b
X2;7 ¼ A5;3 ab þ B1;3 ab X6;9 ¼ D2;3 a2 þ E2;3 b2 þ 2A9;10
X2;8 ¼ A5;7 b2 þ B1;3 a2 X6;10 ¼ A9;4 a þ 3D2;3 a
X2;9 ¼ 2A5;10 b X6;11 ¼ A9;8 b þ 3E2;3 b
X2;10 ¼ A5;4 ab þ B1;4 ab X6;12 ¼ D2;4 a2 þ E2;4 b2 þ 3A9;11
X2;11 ¼ A5;8 b2 þ B1;4 a2 X7;7 ¼ A3;3 a2 þ B3;3 b2 þ 4D2;2
X2;12 ¼ 3A5;11 b X7;8 ¼ A3;7 ab þ B3;3 ab
X3;3 ¼ D1;1 a2 þ E1;1 b2 X7;9 ¼ 2A3;10 a þ 2D2;3 a
X3;4 ¼ D1;1 a X7;10 ¼ A3;4 a2 þ B3;4 b2 þ 6D2;4
X3;5 ¼ E1;1 b X7;11 ¼ A3;8 ab þ B3;4 ab
2 2
X3;6 ¼ D1;2 a þ E1;2 b X8;8 ¼ A7;7 b2 þ B3;3 a2 þ 4E2;2
X3;7 ¼ 2D1;2 a X8;9 ¼ 2A7;10 b þ 2E2;3 b
X3;8 ¼ 2E1;2 b X8;10 ¼ A7;4 ab þ B3;4 ab
X3;9 ¼ D1;3 a2 þ E1;3 b2 X8;11 ¼ A7;8 b2 þ B3;4 a2 þ 6E2;3
X3;10 ¼ 3D1;3 a X8;12 ¼ 3A7;11 b þ 2E2;4 b
X3;11 ¼ 3E1;3 b
X9;9 ¼ D3;3 a2 þ E3;3 b2 þ 4A10;10
X3;12 ¼ D1;4 a2 þ E1;4 b2
X 9;10 ¼ 2A10;4 a þ 3D3;3 a
X4;4 ¼ A2;2 a2 þ B2;2 b2 þ D1;1 X9;11 ¼ 2A10;8 b þ 3E3;3 b
X4;5 ¼ A2;6 ab þ B2;2 ab X9;12 ¼ D3;4 a2 þ E3;4 b2 þ 6A10;11
X4;6 ¼ A2;9 a þ D1;2 a
X10;10 ¼ A4;4 a2 þ B4;4 b2 þ 9D3;3
X4;7 ¼ A2;3 a2 þ B2;3 b2 þ 2D1;2
X10;11 ¼ A4;8 ab þ B4;4 ab
X4;8 ¼ A2;7 ab þ B2;3 ab
A10;12 ¼ 3A4;11 a þ 3D3;4 a
X4;9 ¼ 2A2;10 a þ D1;3 a
X4;10 ¼ A2;4 a2 þ B2;4 b2 þ 3D1;3 X11;11 ¼ A8;8 b2 þ B4;4 a2 þ 9E3;3
X4;11 ¼ A2;8 ab þ B2;4 ab X11;12 ¼ 3A8;11 b þ 3E3;4 b
X4;12 ¼ 3A2;11 a þ D1;4 a X12;12 ¼ D4;4 a2 þ E4;4 b2 þ 9A11;11
X5;5 ¼ A6;6 b2 þ B2;2 a2 þ E1;1 in which, a ¼ mp np
a and b ¼ b and Xi,j = Xj,i, (i, j = 1 to
X5;6 ¼ A6;9 b þ E1;2 b 12)
X5;7 ¼ A6;3 ab þ B2;3 ab

123
Static solutions for functionally graded simply supported plates 69

References Matsunaga, H.: Stress analysis of functionally graded plates

subjected to thermal and mechanical loading. Compos.
Anderson, T.A.: A 3D elasticity solution for a sandwich com- Struct. 87, 344–357 (2009)
posite with functionally graded core subjected to transverse Nguyen, T.K., Sab, K., Bonnet, G.: First-order shear deforma-
loading by a rigid sphere. Compos. Struct. 60, 265–274 tion plate models for functionally graded materials. Com-
(2003) pos. Mater. 83, 25–36 (2008)
Bodaghi, M., Saidi, A.R.: Levy-type solution for bending Pendhari, S.S., Kant, T., Desai, Y.M., Subbaiah, C.V.: On
analysis of thick functionally graded rectangular plates deformation of functionally graded narrow beams under
based on higher-order shear deformation plate theory. transverse loads. Int. J. Mech. Mater. Des. 6, 269–282
Appl. Math. Model. 34, 3659–3670 (2010) (2010)
Chakraborty, A., Gopalakrishnan, S.: A spectrally formulated Praveen, G.N., Reddy, J.N.: Nonlinear transient thermoelastic
finite element for wave propagation analysis in function- analysis of functionally graded ceramic-metal plates. Int.
ally graded beams. Int. J. Solids Struct. 40, 2421–2448 J. Solids Struct. 35(33), 4457–4476 (1998)
(2003) Reddy, J.N.: Analysis of functionally graded plates. Int. J. Num.
GhannadPour, S.A.M., Alinia, M.M.: Large deflection behav- Meth. Eng. 47, 663–684 (2000)
iour of functionally graded plates under pressure loads. Reddy, J.N., Chin, C.D.: Thermomechanical analysis of func-
Compos. Struct. 75, 67–71 (2006) tionally graded cylinders and plates. J. Therm. Stress. 21,
Han, F., Pan, E., Roy, A.K., Yue, Z.Q.: Response of piezo- 593–626 (1998)
electric, transversely isotropic, functionally graded and Sankar, B.V.: An elasticity solution for functionally graded
multilayered half spaces to uniform circular surface load- beam. Compos. Sci. Tech. 61, 689–696 (2001)
ings. Comput. Model. Eng. Sci. 14(1), 15–30 (2006) Sankar, B.V., Tzeng, J.T.: Thermal stresses in functionally
Kang, Y.A., Li, X.F.: Bending of functionally graded cantilever graded beams. AIAA J. 410(6), 1228–1232 (2002)
beam with power-law non-linearity subjected to an end Shen, H.S.: Nonlinear bending response of functionally graded
force. Int. J. Non Linear Mech. 44, 696–703 (2009) plates subjected to transverse loads and in thermal envi-
Kant, T., Ramesh, C.K.: Numerical integration of linear ronments. Int. J. Mech. Sci. 44(3), 561–584 (2002)
boundary value problems in solid mechanics by segmen- Tanigawa, Y.: Some basic thermoelastic problems for non-
tation method. Int. J. Num. Meth. Eng. 17, 1233–1256 homogeneous structural materials. Appl. Mech. Rev.
(1981) 48(6), 287–300 (1995)
Kant, T., Gupta, A.V., Pendhari, S.S., Desai, Y.M.: Elasticity Vel, S.S., Batra, R.C.: Exact solution for thermoelastic defor-
solution for cross-ply composite and sandwich laminates. mations of functionally graded thick rectangular plates.
Compos. Struct. 83, 13–24 (2008) AIAA J. 40(7), 1421–1433 (2002)
Kantrovich, L.V., Krylov, V.I.: Approximate Methods of Higher Woo, J., Meguid, S.A.: Nonlinear analysis of functionally gra-
Analysis, 3rd edn. Noordhoff, Groningen (1958) ded plates and shallow shells. Int. J. Solids Struct.
Kashtalyan, M.: Three-dimensional elasticity solution for 38(42–43), 7409–7421 (2001)
bending of functionally graded rectangular plates. Eur. Woo, J., Meguid, S.A., Stranart, J.C., Liew, K.M.: Thermo-
J. Mech. 23, 853–864 (2004) mechanical postbuckling analysis of moderately thick
Khabbaz, R.S., Manshadi, B.D., Abedian, A.: Nonlinear anal- functionally graded plates and shallow shells. Int. J. Mech.
ysis of FGM plates under pressure loads using the higher- Sci. 47, 1147–1171 (2005)
order shear deformation theories. Compos. Struct. 89, Yang, J., Shen, H.S.: Nonlinear bending analysis of shear
333–344 (2009) deformable functionally graded plates subjected to thermo-
Ma, L.S., Wang, T.J.: Relationship between axisymmetric mechanical loads under various boundary conditions.
bending and buckling solution of FGM circular plates Compos. B 34(2), 103–115 (2003)
based on third-order plate theory and classical plate theory.
Int. J. Solid Struct. 41, 85–101 (2004)

123