Composites Part B 156 (2019) 72–87

Contents lists available at ScienceDirect

Composites Part B
journal homepage: www.elsevier.com/locate/compositesb

An exact three dimensional solution for bending of thick rectangular FGM T

plate
Zahra Vafakhah, Bahram Navayi Neya∗
Babol Noshirvani University of Technology, Civil Engineering Department, P.O. Box 484, Shariati Ave, Babol, Mazandaran, 47148-71167, Iran

A R T I C LE I N FO A B S T R A C T

Keywords: This study uses the displacement potential function, DPF, to present an exact solution for bending of functionally
Exact solution graded, FGM, thick plates. The plates are simply-supported rectangular, isotropic, and nonhomogeneous, with
Thick rectangular plates exponential variation of Young's modulus along their thickness Using the DPF, the governing equations are
FGM simplified to second and fourth order partial differential equations solved by using the separation of variables
Displacement potential function
method and applying exact boundary conditions. The solution is applicable to any plate with no restriction on its
thickness as well as to all FGM plates. The results are compared with other existing analytical and numerical
results for thin, moderately thick, and thick plate. Comparisons show excellent agreement making the method
applicable and reliable for various kinds of material properties, thickness ratios and loading conditions without
any simplifying assumptions such as for strain or stress distribution in plate thickness. Various inhomogeneity,
thicknesses, and aspect ratios are also investigated with respect to plate responses with results showing that
increase in inhomogeneity, and decrease of aspect ratio and FGM plate thickness result in decrease in the di-
mensionless transverse displacement of the plate. Moreover, the effects of increasing inhomogeneity on trans-
verse displacement are more significant for thicker plates and higher aspect ratios.

1. Introduction considered variations in modulus of elasticity in the forms of power,

Functionally graded materials (FGMs) are a class of advanced
composite materials with nonhomogeneous microstructure and variable
properties in one or more dimension. Classical composite material
structures have discontinuity of properties in layers interface, and as a
result, stress concentration occurs at high temperatures, particularly
where two metals with different coefficients of expansion are used
[1,2], and functionally graded materials have emerged as a solution to
this problem. Preparation of FGMs occurs in such a way that the
properties of the constituents change continuously, so there is no spe-
cific interface between two materials and the characteristics of two or
more materials of the composite remain [3,4]. FGMs are usually made
from ceramics and metals with material properties such as the Young's
modulus vary from a metal-rich surface to a ceramic-rich surface [5],
with the metallic part providing the required fracture toughness and
ceramic part supplying thermal resistance [6].
In recent years, many studies have been conducted on analysis and
modelling of FGMs, eventually leading to impressive developments in
studies on FGM plates. For example, a bending solution for FGM plates
using classical plate theory (CPT) combined with a spline finite-strip
method has been presented by Beena and Parvathy [7]. They
sigmoid, and exponential functions and obtained deflection and stress
results validated by a Fourier series solution [3]. Nguyen et al. [8], used
first-order shear deformation theory (FSDT) and the finite-element
method to analyze FGM plates. In that study the plates were analyzed
under mechanical and thermal loads and the method's accuracy was
corroborated. In 2013, bending and free vibration of thick rectangular
FGM plates using simple first-order shear-deformation theory were in-
vestigated by Thai and Choi [9], resulting in a theory with identical
accuracy to that of FSDT theory but with fewer unknowns. Xiang and
Kang [10] used different higher-order shear-deformation theories
(HSDTs) developed by Levinson [11], Touratier [12], Karama [13],
Aydogdu [14], and Mantari [15] to model FGM plates. They also solved
governing equations using numerical methods and their results, eval-
uated in comparison with other available solutions, demonstrated high
accuracy of their method for thin FGM plates. In 2015, a higher-order
hyperbolic shear deformation theory for analyzing FGM plates was
proposed by Nguyen [16]. Nguyen used Hamilton's principle to derive
equations of motion solved using Navier-type solutions. The results of
his study were in good agreement with different HSDTs with results
close to those of quasi-3D theory in many cases. Quasi-3D theories are
higher-order deformation theories that consider both shear strain and

∗
Corresponding author.
E-mail addresses: Zahra_vafakhah@yahoo.com (Z. Vafakhah), Navayi@nit.ac.ir (B. Navayi Neya).

https://doi.org/10.1016/j.compositesb.2018.08.036
Received 5 July 2018; Accepted 9 August 2018
Available online 16 August 2018
1359-8368/ © 2018 Elsevier Ltd. All rights reserved.
Z. Vafakhah, B. Navayi Neya
thickness stretching effects that also take into account both in-plane and
out-plane displacement functions. Analysis of bending and free vibra-
tion of FGM plates using various higher order shear deformation the-
ories were conducted by Thai and Choi [17]. In their study, HSDTs were
based on cubic, sinusoidal, hyperbolic, and exponential variations of in-
plane displacements through the thickness of the plate. All their results
showed that the theories considered were not only close to one another
but also in good agreement with other available HSDTs, quasi-3D, and
exact 3D solutions. In 2013, Thai and Kim [18] presented a bending
analysis of FGM plates using quasi-3D sinusoidal shear deformation
theory. They also compared their results with FSDT, HSDTs, other
quasi-3D theories with more unknowns, and exact 3D theory. It was
observed that if thickness-stretching effects are considered, HSDTs will
produce good results for thick plates but inaccurate results for very
thick plates. One year later, bending and free vibration of FGM plates
by considering hyperbolic variations of all displacements across the
thickness were studied by Thai et al. [19]. This proposed theory ob-
tained more accurate results than quasi-3D sinusoidal theory that were
in excellent agreement with exact 3D solutions even for very thick
plates. In 2015, a quasi-3D tangential shear deformation theory for
bending analysis of the FGM plates was presented by Mantari and
Soares [20]. In their work, they derived governing equations and
boundary conditions using the principle of virtual work that were
solved via Navier-type solution. That study also concluded that their
theory presented good results with fewer unknowns and was compar-
able to other quasi-3D theories requiring more unknowns.
While various theories with different accuracies have been em-
ployed by researchers to study behavior of functionally-graded mate-
rials, the increased importance of application of these materials makes
development of more accurate theories essential. For two-dimensional
theories in which variations of structural parameters like stress, strain,
and displacement are considered in just two directions, assumptions
such as ignoring the effects of normal or transverse shear strains or
considering special distribution for stresses or displacements can cause
errors in obtained responses. Even using quasi-3D theories, although
effects of both normal and transverse shear strains are considered, the
exact response cannot be obtained because of special distributions of
stresses or displacements. Furthermore, with increase in plate thickness,
since errors become greater and more inevitable, using three-dimen-
sional elasticity equations for analysis of FGM plates is recommended
despite their extremely complexity.
In 2004, Kashtalyan [21] studied a three-dimensional elasticity so-
lution for bending of simply-supported rectangular transversely-loaded
FGM plates. To evaluate the solution, results for FGM plates were
compared with available homogeneous isotropic plate results, showing
that when there are no nonhomogeneous parameters, responses of FGM
plates will converge to responses of homogeneous isotropic plates. In
2007, Zenkour [22] studied exponentially-graded thick rectangular
using trigonometric 2-D and 3-D elasticity solutions and compared
exact 3-D solutions with those from trigonometric 2-D theory and HSDT
theories, demonstrating the accuracy of his method. In 2010, a three-
dimensional static analysis of thick functionally graded plates using a
meshless local Petrov-Galerkin (MLPG) method was published by Va-
ghefi et al. [5], obtaining results in excellent agreement with available
analytical and numerical solutions. In 2014, Zhang et al. [23] presented
a 3-D elasticity solution for static bending of thick FGM plates using a
hybrid semi-analytical approach - the state-space based differential
quadrature method (SSDQM).
Displacement potential functions represent one of the methods for
solution of elasticity problems. This method helps in solving such
problems by uncoupling the system of differential equations or at least
by decreasing the number of governing differential equations [24].
Well-known sets of displacement potential functions include those of
Boussinesq, Galerkin, Love, and Neuber-Papkovich for isotropic media,
those of Lekhnitskii, Lekhnitskii-Hu-Nowacki for transversely isotropic
media, and other well-known displacement functions related and
73
Composites Part B 156 (2019) 72–87
possible to obtain from one another [25]. The majority of these po-
tential functions depend on constitutive relationships and are mainly
applicable only to isotropic media.
In 1940, a series of potential functions for general solution of
transversely isotropic and axial symmetry media for elastostatics pro-
blems were published by Lekhnitskii [26]. Hu (1953) [27] and Nowacki
(1954) [28] extended Lekhnitskii's potential functions to transversely-
isotropic general problems in three-dimensional states. In 2005, Es-
kandari-Ghadi [29] extended Lekhnitskii-Hu-Nowacki methods from
elastostatics problems to elastodynamics. In 2011, Nematzadeh et al.
[30], successfully employed Lekhnitskii-Hu-Nowacki potential func-
tions for bending analysis of transversely isotropic thin, moderately
thick, and thick plates under arbitrary static loading. By modifying the
potential functions for the dynamic case [29], Eskandari-Ghadi and
Amiri-Hezaveh [31] obtained the harmonic response of exponentially-
graded transversely-isotropic half-space media. Moreover, Eskandari-
Gahadi et al., solved elastostatic [32] and elastodynamic [33–37] pro-
blems for both transversely isotropic half-space and full-space by using
DPFs introduced in Ref. [29]. The senior author of this research, Navayi
Neya et al. [38] obtained a three-dimensional solution for simply-sup-
ported thick rectangular plates under a moving load using the potential
functions introduced in Ref. [29]. The Navayi Neya et al. [39] also
presented a free-vibration analysis of transversely isotropic thick rec-
tangular plates using such potential functions [29] and also used the
DPF method on buckling problems of isotropic [40] and transversely
isotropic [41] plates. They obtained solutions for simply-supported
thick rectangular isotropic plates using augmented Love potential
functions, and for the transversely isotropic plates for which the po-
tential function introduced in Ref. [42] was used. While there are many
other studies that have presented exact responses for three-dimensional
analysis of homogeneous and nonhomogeneous half or full-space cases
using potential functions [43–49], there is a very limited number of
studies [22] in the literature regarding exact solutions for solving the
thick FGM plate bending problem.
This study presents an exact solution for bending of simply sup-
ported rectangular transversely loaded FGM plates with arbitrary
thickness. Governing differential equations are obtained using three
dimensional elasticity equations expressed in terms of the DPF in-
troduced in Ref. [31]. These governing equations, solved using the se-
paration-of-variables method, can develop responses that include dis-
placements, stresses, strains, moments, shears, and axial forces obtained
at each point section of plate. A major advantage of this work is that it
determines the exact bending solution of FGM plates using 3D elasticity
equations without simplifying assumptions such as specific strain or
stress distributions in plate thickness or shear correction factors. Since
the obtained results can be used for any arbitrary thickness, they can be
used as a reliable exact solution that can be used to verify results from
other studies that use numerical methods.
2. Theoretical formulation
2.1. Governing equations
A simply-supported rectangular plate with linear elastic and non-
homogeneous isotropic material will be considered. Fig. (1) illustrates
such a plate with dimensions a and b and arbitrary constant thickness h
located in a Cartesian coordinate system (x,y,z) with origin assumed at
the bottom of the surface where z = 0. u, v, w are the displacement
components in the x , y, z directions, respectively.
Based on relationships between stress-strain, strain-displacement
and equilibrium equations [24], 3D elasticity equations in terms of
displacements with absence of body forces can be written as in Eq. (1),
Navier's equations for inhomogeneous media composed of linear elastic
isotropic materials.
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

where
2
□2 = ∇2xy + ( ∂
∂z
+β ) − β2
□2 . □2 = □ 4 (9)
To solve the governing equations, the separation-of-variables
method is used, resulting in potential functions that can be written as
follows:
F = f (x ) g (y ) h (z ) (10)

χ = f1 (x ) g1 (y ) h1 (z ) (11)
Fig. 1. Simply supported FGM rectangular plate. Substituting Eqs. (10) and (11) into Eqs. (7) and (8) produces the
following terms:
μ∇2 u + ( λ + μ) ∂x
∂
( ∂u
∂x
+
∂v
∂y
+
∂w
∂z )+ ( ∂λ
∂x
∂u
∂x
+
∂v
∂y
+
∂w
∂z )+2 ∂μ ∂u
∂x ∂x ∂ 4f
4
∂ 4g ∂ 4h ∂ 2f ∂ 2g ∂ 2g ∂ 2h ∂ 2f ∂ 2h
gh + f 4 h + fg 4 + 2 2 2 h + 2f 2 2 + 2 2 g 2
∂x ∂y ∂z ∂x ∂y ∂y ∂z ∂x ∂z
+
∂μ
∂y ()+ (
∂u
∂y
+
∂v
∂x
∂μ
∂z
∂u
∂z
+
∂w
∂x )=0 ∂2f ∂h ∂2g ∂h ⎞ ∂3h ∂ 2h
+ 4β ⎛ 2 g +f 2 + 4β ⎛fg 3 ⎞ + 4β 2 ⎛fg 2 ⎞
μ∇ v + ( λ + μ) ( + )+ ( + )+
⎜ ⎟ ⎜ ⎟ ⎜ ⎟

2 ∂ ∂u ∂v ∂w ∂λ ∂u ∂v ∂w ∂μ ∂v
∂y ∂x ∂y
+ ∂z ∂y ∂x ∂y
+ ∂z
2 ∂y ∂y ⎝ ∂x ∂z ∂y ∂z ⎠ ⎝ ∂z ⎠ ⎝ ∂z ⎠
∂ 2f ∂ 2g
α − 1
+ ( + )+ ( )=0 − 4β 2 1 ⎛ gh + f 2 h ⎞ = 0
∂μ ∂u ∂v ∂μ ∂v ∂w
+ ⎜ ⎟
∂x ∂y ∂x ∂z ∂z ∂y α2 ⎝ ∂x 2 ∂y ⎠ (12)
μ∇2 w + ( λ + μ) ∂z
∂
( + + )+ ( +
∂u
∂x
∂v
∂y
∂w
∂z
∂λ
∂z
∂u
∂x
∂v
∂y
+
∂w
∂z )+2∂μ ∂w
∂z ∂z ∂2f1 ∂ 2g ∂ 2h ∂h
g1 h1 + f1 21 h1 + f1 g1 21 + 2βf1 g1 1 = 0
+ ( + )+ ( )=0
∂μ ∂u ∂w ∂μ ∂v ∂w ∂x 2 ∂y ∂z ∂z
+ (13)
∂x ∂z ∂x ∂y ∂z ∂y (1)
Since boundary conditions must be satisfied, among the probable
where λ, μ are Lame's constants and ∇2 is the Laplace operator defined
solutions to the differential Eqs. (12) and (13), only trigonometric
by:
functions for f (x ), g (y ), f1 (x ), and g1 (y ) are acceptable.
∂2 ∂2 ∂2
∇2 = 2
+ 2 + 2 ⎧ f = c1 sin (αm x ) + c2 cos (αm x )
∂x ∂y ∂z (2)
⎪ g = c3sin (αn y ) + c4 cos (αn y )
The displacement components can be expressed in terms of the ⎨ f1 = c5 sin (αm′ x ) + c6 cos (αm′ x )
displacement potential functions F , χ (x , y, z ) presented by Eskandari- ⎪
g = c7 sin (αn′ y ) + c8 cos (αn′ y ) (14)
Ghadi [31] in the form shown in Eq. (3): ⎩ 1

∂2F ∂F ∂χ
where c1 to c8 , αm , αn , αm′ and αn′ are unknown coefficients. Substituting
u = −α1 ∂x ∂z − 2β ∂x − ∂y Eq. (14) in Eqs. (12) and (13), fourth and second ordinary differential
∂2F ∂F ∂χ equations in terms of h and h1 can be obtained as:
v = −α1 ∂y∂z − 2β ∂y + ∂x
1 ∂ 4h 1 ∂3h 1 ∂ 2h 2 1 ∂h
w = α2 ∇2xy F +
∂
∂z (2β + ) F ∂
∂z (3) h ∂z 4
+ 4β
h ∂z 3
2
+ (4β 2 − 2αmn )
h ∂z 2
− 4βαmn
h ∂z
4
+ αmn

2 α1 − 1
in which β is an inhomogeneity parameter showing the rate of change + 4β 2αmn =0
α2 (15)
in material properties along the thickness. α1, α2 are defined as follows:
λ+μ λ + 2μ 1 ∂2h1 1 ∂h1
α1 = , α2 = + 2β ′ 2
= αmn
μ μ (4) h ∂z 2 h ∂z (16)
2
In this study, it is assumed that the material properties along the where αmn ′ 2 are:
and αmn
thickness varies exponentially, so the properties of Lame's constant can 2
αmn = αm 2 + αn 2
be formulated as:
′ 2 = αm′ 2 + αn′ 2
αmn (17)
λ (z ) = λ 0 e 2βz , μ (z ) = μ0 e 2βz , 0 ≤ z ≤ h
νE0 E0
Solving Eqs. (15) and (16) using Laplace transforms yields the fol-
λ0 = , μ0 = lowing relationships.
(1 + ν )(1 − 2ν ) 2(1 + ν ) (5)
where λ 0 , μ0 are Lame's constants, E0 is the Young's modulus at the h = c9 e s1 z + c10 e s2 z + c11 e s3 z + c12 e s4 z
origin surface (z = 0) and ν is Poisson's ratio, assumed constant. β is h1 = c13 e s5 z + c14 e s6 z (18)
therefore defined as:
where S1 to S4 are the roots of the Laplace equations.
1 E
β = − ln ⎛ 0 ⎞ ⎜ ⎟

2h ⎝ Eh ⎠ (6) 2.2. Kinematic and static boundary conditions

where Eh is Young's modulus at z = h . Kinematic and static boundary conditions for simply-supported
Substituting Eq. (3) into Eq. (1) results in Eqs. (7) and (8) produces rectangular plates can be written as:
the governing differential equations, second and forth order, respec-
tively, in terms of potential functions of χ , F : ⎧ x = 0, a → w = 0
⎨ y = 0, b (19)
□2χ = 0 (7) ⎩

⎛□ 4 − 4β 2 α1 − 1 ∇2 ⎞ F = 0
⎜ ⎟
⎧ x = 0, a → Mx = 0
xy y = 0 , b → My = 0
⎝ α2 ⎠ (8) ⎨
⎩ (20)

74
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

which MX and MY are bending moments about y and x axes, respec- Table 1
tively. To apply the boundary conditions, the displacements and Dimensionless deflection W ( a b h
, ,
2 2 2 ) of FGM plate subjected to sinusoidal
stresses in Eqs. (21) and (22) are written in terms of f , g , h, f1 , g1 and h1. distributed load for δ = 10 .
s
η Theory β

⎧u = −α1
⎪
( ) g ( ) − 2β ( ) gh − f ( ) h
∂f
∂x
∂h
∂z
∂f
∂x 1
∂g1
∂y 1 0.05 0.15 0.25 0.35 0.5 0.75
⎪
⎨
v = −α1 f ( ) ( ) − 2βf ( ) h + ( ) g h
∂g
∂y
∂h
∂z
∂g
∂y
∂f1
∂x 1 1 1 Quasi-3D [22]
Quasi-3D [52]
0.2790
0.2790
0.2523
0.2523
0.2280
0.2280
0.2060
0.2060
0.1767
0.1767
0.1366
0.1366
⎪
⎪ w = α2 ⎛
⎩ ⎝
( ) gh + f ( ) h⎞⎠ + 2βfg ( ) + fg ( )
∂2f
∂x 2
∂2g
∂y 2
∂h
∂z
∂2h
∂z 2
(21)
Quasi-3D [18]
HSDT [55]
0.2790
0.2816
0.2523
0.2550
0.2280
0.2309
0.2060
0.2093
0.1767
0.1807
0.1366
0.1417
Present work 0.2800 0.2537 0.2299 0.2085 0.1801 0.1414
2 Quasi-3D [22] 0.7015 0.6344 0.5734 0.5180 0.4444 0.3435
( )( ) ( ) ( )( )
2 2 2
⎧ σx = −μα2 ∂ f2 g ∂h − 2μβα2 ∂ f2 gh − 2μ ∂ f2 ∂g1 h1 Quasi-3D [52] 0.7015 0.6344 0.5734 0.5180 0.4444 0.3435
⎪ ∂x ∂z ∂x ∂x ∂y
Quasi-3D [18] 0.7015 0.6344 0.5734 0.5180 0.4444 0.3435
⎪
⎪
⎪
+ λf ( )( )
∂2g
∂y 2
∂h
∂z
− 2λβf
∂2g
∂y 2 ( )
h + 2λβfg
∂2h
∂z 2
+ λfg 3 ( ) ( )
∂3h
∂z
HSDT [55]
Present work
0.7066
0.7040
0.6397
0.6376
0.5795
0.5778
0.5252
0.5239
0.4536
0.4527
0.3560
0.3556
⎪ 3 Quasi-3D [22] 0.8849 0.8002 0.7233 0.6534 0.5605 0.4333

⎪
( )( )
∂2g ∂h
⎪ σy = −μα2 f ∂y 2 ∂z − 2μβα2 f ∂y 2 h − 2μ ∂x
∂2g
( ) ∂f1 ∂g1
∂y
h1( )( ) Quasi-3D [52]
Quasi-3D [18]
0.8849
0.8849
0.8002
0.8002
0.7233
0.7233
0.6534
0.6534
0.5605
0.5605
0.4333
0.4333
⎪
⎪
+λ
∂2f
( )( )
∂x 2
∂h
g ∂z − 2λβ
∂2f
∂x 2 ( )
( ) ( )
gh + 2λβfg
∂2h
∂z 2
+ λfg 3
∂3h
∂z
HSDT [55]
Present work
0.8909
0.8880
0.8066
0.8042
0.7307
0.7288
0.6622
0.6608
0.5720
0.5710
0.4489
0.4485
⎪
( )( ∂2f
)∂h
⎪ σz = (2μα2 + λ ) 2 g ∂z + (2μβα2 + λ ) f
⎪ ∂x ( )( ) ∂2g
∂y 2
∂h
∂z
gh
⎪
( )( ) ( )( )
2 2
( ) ( )
∂ g 2 ∂3h
∂ f
− 2μ 2 ∂y1 h1 + λf
∂ g ∂h
+ 2μβα2 fg
∂ h
+ μα2 fg 3 Table 2
∂y 2 ∂z 2
⎨
⎪
∂x ∂z ∂z Dimensionless deflection W ( a b h
, ,
2 2 2 ) of FGM plate subjected to sinusoidal

⎪
⎪
− 2λβ ( )
∂2f
∂x 2
gh − 2λfg ( ) ∂2g
∂y 2
h distributed load for δ = 4 .

⎪ η Theory β

⎪
∂f
( )( )( )
∂g ∂h
( )( ) ( )
⎪ τxy = (−2μα1) ∂x ∂y ∂z + (−4μβ ) ∂x ∂y h − μf1 ∂y 2 h1
∂f ∂g ∂2g1

0.05 0.15 0.25 0.35 0.5 0.75

⎪
⎪
⎪
+μ ( )
∂2f1
∂x 2 1 3D [22] 0.3490 0.3167 0.2875 0.2608 0.2253 0.1805
Quasi-3D [22] 0.3475 0.3142 0.2839 0.2563 0.2196 0.1692
⎪ τxy = −λ
⎪
∂f
∂x ( )( ) ∂2h
( )( )
∂z
∂g ∂h
g 2 − μf1 ∂y1 ∂z1 + μα2 3 gh + μα2 ∂x
∂3f
∂x ( ) ( )( )
∂f ∂2g
∂y 2
h Quasi-3D [52]
Quasi-3D [18]
0.3475
0.3475
0.3142
0.3142
0.2839
0.2839
0.2563
0.2563
0.2196
0.2196
0.1692
0.1692
⎪ ∂g
( )( ) ( ) ( )
∂2h ∂f1 ∂h1
⎪ τxy = −λf ∂y ∂z2 + μ ∂x g1 ∂z + μα2 ∂x 2 ∂y h + μα2 f ∂y3 h
⎩
∂2f ∂g
( )( ) ( )∂3g
HSDT [55]
Present work
0.3602
0.3490
0.3259
0.3168
0.2949
0.2875
0.2668
0.2608
0.2295
0.2253
0.1785
0.1763
(22) 2 3D [22] 0.8153 0.7395 0.6707 0.6085 0.5257 0.4120
Quasi-3D [22] 0.8120 0.7343 0.6635 0.5992 0.5136 0.3962
Bending moments Mx and My , respectively, are given by the fol- Quasi-3D [52] 0.8120 0.7343 0.6635 0.5992 0.5136 0.3962
lowing terms: Quasi-3D [18] 0.8120 0.7343 0.6635 0.5992 0.5136 0.3962
HSDT [55] 0.8325 0.7534 0.6819 0.6173 0.5319 0.4150
h Present work 0.8153 0.7395 0.6707 0.6085 0.5257 0.4120
⎧ M = σ z. dz
⎪ x ∫ x 3 3D [22] 1.0134 0.9190 0.8335 0.7561 0.6533 0.5121
0 Quasi-3D [22] 1.0094 0.9127 0.8248 0.7449 0.6385 0.4927
⎨ h Quasi-3D [52] 1.0094 0.9127 0.8248 0.7449 0.6385 0.4927
⎪ My = ∫ σy z. dz Quasi-3D [18] 1.0094 0.9127 0.8248 0.7449 0.6385 0.4927
⎩ 0 (23) HSDT [55] 1.0325 0.9345 0.8459 0.7659 0.6601 0.5154
Present work 1.0134 0.9190 0.8335 0.7561 0.6533 0.5121
Satisfying the kinematic boundary conditions, Eq. (19), coefficients 4 3D [22] 1.1012 0.9985 0.9056 0.8215 0.7098 0.5564
c2 and c4 are obtained as zero, and αm and αn are obtained mπ , nπ , Quasi-3D [22] 1.0968 0.9918 0.8963 0.8094 0.6939 0.5355
a b
respectively. Also, satisfying the static boundary conditions (Eq. (20)), Quasi-3D [18] 1.0968 0.9918 0.8962 0.8094 0.6939 0.5355
mπ nπ HSDT [55] 1.1211 1.0147 0.9186 0.8317 0.7169 0.5599
makes both c5 and c7 equal to zero and αm′ and equal to , , re-
a b Present work 1.1012 0.9985 0.9056 0.8214 0.7098 0.5564
spectively, where m and n are positive integer numbers, resulting in 5 3D [22] 1.1459 1.0391 0.9424 0.8548 0.7386 0.5790
F , χ expressed as follows: Quasi-3D [22] 1.1414 1.0321 0.9327 0.8423 0.7221 0.5573
∞ ∞ Quasi-3D [18] 1.1414 1.0321 0.9327 0.8423 0.7221 0.5573
mπ nπ HSDT [55] 1.1663 1.0556 0.9556 0.8653 0.7458 0.5825
F= ∑ ∑ sin ⎛ x ⎞ sin ⎛ y ⎞ [G1 e s1 z + G2 e s2 z + G3 e s3 z + G4 e s4 z]
⎝ a ⎠ ⎝ b ⎠ Present work 1.1459 1.0391 0.9424 0.8548 0.7386 0.5790
m=1 n=1
6 3D [22] 1.1714 1.0622 0.9633 0.8738 0.7550 0.5919
(24) Quasi-3D [22] 1.1668 1.0551 0.9535 0.8611 0.7382 0.5697
∞ ∞ Quasi-3D [18] 1.1668 1.0551 0.9534 0.8611 0.7382 0.5697
mπ nπ HSDT [55] 1.1920 1.0789 0.9767 0.8844 0.7623 0.5955
χ= ∑ ∑ cos ⎛ x ⎞ cos ⎛ y ⎞ [G5 e s5 z + G6 e s6 z]
⎝ a ⎠ ⎝ b ⎠ (25) Present work 1.1714 1.0622 0.9633 0.8738 0.7550 0.5919
m=1 n=1

in which G1 = c1. c3. c9, G2 = c1. c3. c10, G3 = c1.c3. c11, G4 =

c1. c3. c12, G5 = c6. c8. c13 and G6 = c6. c8. c14 . written as:
Traction boundary conditions can be employed to find unknowns G1
to G6 . ⎧ z = 0 → σz = 0
z = h → σz = −p
⎨
2.3. Surface traction boundary conditions ⎩ = 0, h → τxZ = τyZ = 0
Z (26)

To find the six unknown coefficients G1 to G6, the traction Plate loading p can be described in the form of a double Fourier
boundary conditions at the top and bottom surfaces of the plate under series as in Ref. [50]:
distributed loading p (x , y ) can be used. The traction boundary can be

75
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

Table 3 ∂2h mπ 3
Dimensionless deflection W ( a b h
, ,
2 2 2 ) of FGM plate subjected to sinusoidal
⎧ Z = 0, h → τxz = −
⎪
( )λ
mπ
a ∂z 2
+μ ( )
nπ
b
∂h1
∂z
− μα2 ( )h
a
⎪ nπ 2
distributed load for δ = 2 .
⎪ − μα2 a
mπ
( )( )
b
h
η Theory β
⎪ =0
⎨ ∂2h nπ 3
0.05 0.15 0.25 0.35 0.5 0.75 ⎪ Z = 0, h → τyz = − ( )λ
nπ
b ∂z 2
−μ ( )
mπ ∂h1
a ∂z
− μα2 ( )h
b
⎪ mπ 2
1 3D [22] 0.5769 0.5247 0.4766 0.4324 0.3726 0.2890 ⎪ − μα2 b
⎪
nπ
( )( )
a
h
Quasi-3D [22] 0.5730 0.5180 0.4678 0.4221 0.3611 0.2771
⎩ =0 (30)
Quasi-3D [52] 0.5731 0.5181 0.4679 0.4222 0.3612 0.2771
Quasi-3D [18] 0.5731 0.5181 0.4679 0.4222 0.3612 0.2771
According to Eq. (30), function h1 and accordingly potential func-
Quasi-3D [51] 0.5779 0.5224 0.4717 0.4256 0.3648 0.2793
Quasi-3D [53] 0.5777 0.5222 0.4716 0.4255 0.3640 0.2793 tion χ are both zero, so it is unnecessary to find coefficients G5 and G6 .
Quasi-3D [54] 0.5750 0.5198 0.4694 0.4236 0.3624 0.2780 To find coefficients G1 to G4 by satisfying the surface boundary
HSDT [55] 0.6363 0.5752 0.5195 0.4687 0.4018 0.3079 conditions, Eq. (26)1 and (26)2 the following equations are obtained.
HSDT [54] 0.6351 0.5741 0.5185 0.4679 0.4004 0.3075
Present work 0.5769 0.5247 0.4766 0.4324 0.3727 0.2890 2 nπ 2 ∂h
2 3D [22] 1.1944 1.0859 0.9864 0.8952 0.7726 0.6017 ⎪ a ∂z ( )
⎧ z = 0 → σz = −(2μα2 + λ ) mπ fg ∂h − (2μα2 + λ ) f ( )g
b ∂z
Quasi-3D [22] 1.1879 1.0739 0.9700 0.8754 0.7493 0.5757 ⎪ ∂2h
Quasi-3D [52] 1.1880 1.0740 0.9701 0.8755 0.7494 0.5758 ⎪ + 2μβα fg
2 ∂z 2
Quasi-3D [18] 1.1880 1.0740 0.9701 0.8755 0.7494 0.5758 ⎪ ∂3h mπ 2 nπ 2
Quasi-3D [51] 1.1940 1.0794 0.9750 0.8799 0.7537 0.5786 ∂z ( )
⎪+ μα2 fg 3 + 2λβ a fgh + 2λβf b gh = 0 ( )
Quasi-3D [53] 1.1939 1.0793 0.9749 0.8798 0.7530 0.5785 ⎨ 2 ∂h ∂2h
Quasi-3D [54] 1.1907 1.0765 0.9723 0.8775 0.7511 0.5771 ⎪ z = h → σz = fg ⎡−(2μα2 + λ ) αmn ∂z + 2μβα2 ∂z2
HSDT [55] 1.2776 1.1553 1.0441 0.9431 0.8093 0.6238 ⎪ ⎣
HSDT [54] 1.2763 1.1541 1.0431 0.9422 0.8079 0.6234 ⎪ ∂3h 2
⎪ + μα2 3 + 2λβαmn h⎤
Present work 1.1944 1.0859 0.9864 0.8952 0.7726 0.6017 ∂z ⎦
3 3D [22] 1.4429 1.3116 1.1912 1.0811 0.9333 0.7275 ⎪ +p = 0
⎩ (31)
Quasi-3D [22] 1.4354 1.2977 1.1722 1.0579 0.9056 0.6961
Quasi-3D [52] 1.4354 1.2977 1.1722 1.0579 0.9057 0.6962 By solving the above system of equations, Eq. (30)1 or (30)2 and Eq.
Quasi-3D [18] 1.4354 1.2977 1.1722 1.0579 0.9057 0.6961
(31), unknown coefficients G1 through G4 can be found.
Quasi-3D [51] 1.4421 1.3037 1.1776 1.0627 0.9104 0.6992
Quasi-3D [53] 1.4419 1.3035 1.1774 1.0626 0.9096 0.6991
Quasi-3D [54] 1.4386 1.3005 1.1748 1.0602 0.9076 0.6976 3. Numerical results and discussion
HSDT [55] 1.5341 1.3784 1.2540 1.1329 0.9725 0.7506
HSDT [54] 1.5327 1.3861 1.2530 1.1320 0.9712 0.7503
Present work 1.4429 1.3116 1.1913 1.0812 0.9334 0.7275
This section presents various numerical results for simply-supported
4 3D [22] 1.5515 1.4101 1.2807 1.1623 1.0035 0.7824 rectangular FGM plates bending under transverse loads for different
a b
Quasi-3D [22] 1.5435 1.3954 1.2605 1.1376 0.9739 0.7487 thickness ratio (δ = h ), aspect ratio (η = a ), and inhomogeneity para-
Quasi-3D [18] 1.5435 1.3954 1.2605 1.1376 0.9739 0.7487 meter (β ) values. The values selected for β do not necessarily represent
HSDT [55] 1.6458 1.4885 1.3455 1.2157 1.0437 0.8060
Present work 1.5515 1.4101 1.2807 1.1624 1.0035 0.7824
a certain material, but are used to consider the effects of inhomogeneity
5 3D [22] 1.6065 1.4601 1.3261 1.2035 1.0391 0.8102 on plate stresses and displacements. Poisson's ratio is assumed constant
Quasi-3D [22] 1.5982 1.4449 1.3052 1.1780 1.0086 0.7754 at 0.3 and dimensionless displacements and stresses are defined ac-
Quasi-3D [18] 1.5982 1.4449 1.3052 1.1780 1.0086 0.7754 cording to the following relationships:
HSDT [55] 1.7025 1.5397 1.3919 1.2576 1.0798 0.8340
Present work 1.6065 1.4601 1.3261 1.2035 1.0391 0.8102 3
6 3D [22] 1.6377 1.4885 1.3518 1.2269 1.0593 0.8261
⎪ a p0 (
⎧W = 10E40 h w a , b , z
2 2 )
Quasi-3D [22] 1.6294 1.4731 1.3307 1.2010 1.0282 0.7906 ⎪ E0 h3
Quasi-3D [18] 1.6294 1.4731 1.3307 1.2010 1.0282 0.7906 ( b
⎪ wc = 12(1 − ν) a4p0 w x , 2 ,
h
2 )
HSDT [55] 1.7347 1.5688 1.4182 1.2815 1.1003 0.8500 ⎪
Present work 1.6377 1.4885 1.3518 1.2269 1.0593 0.8261
⎪ (
⎪ wc = Gh w ph a , b , z
2 2 )
⎪
∞ ∞
⎪ cu = Gh u
ph (
0,
b h
,
2 4)
mπ nπ
p (x , y ) = ∑ ∑ pmn sin x sin y ⎨ σ = h2 σ ( , , z)
a b
m=1 n=1
a b (27) ⎪ x a2p0 x 2 2
⎪
where pmn is a Fourier coefficient as defined below [50]:
a b
⎪ σx c = σx p 2 2
⎪
( , , h)
⎪ σy = h2 y a b σ ( , , z)
b a ⎪ a2p0 2 2
4
pmn =
ab
∫ ∫ p (x , y) sin mπ
a
nπ
x sin ydxdy
b
⎪ h
⎪ τxz = ap0 xZ τ (0, , z )
b
2
0 0 (28) ⎩ (32)

For some typical loads, pmn can be obtained as follows [50]:

3.1. Verifications
16p0
⎧ (m , n = 1,3, …) uniform distributed load
pmn = π 2mn
To verify and evaluate accuracy of the present study's outcomes, the
⎨ p0 sinusoidal distributed load
⎩ (29) results for FGM plates subjected to sinusoidal load are compared with
the work of others, including: the exact 3D solution of Kashtalyan [21],
If shear stress conditions are satisfied on the top and bottom surfaces Brischetto [56] and Zenkour [22], the quasi-3D solutions of Zenkour
of the plate, Eq. (26)3, the following equations result: [22], Mantari and Guedes Soares [51,52], Thai and Choi [53], Akavci
and Tanrikulu [54], Thai and Kim [18], and Carrera et al. [57], the

76
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

Table 4
Dimensionless normal stress σy ( a b
, ,
2 2 )
h of FGM plate subjected to sinusoidal distributed load for δ = 10 .

η Theory β

0.05 0.15 0.25 0.35 0.5 0.75 1 1.25 1.5

1 Quasi-3D [18] 0.2196 0.2345 0.2503 0.2671 0.2944 0.3640 – – –

Quasi-3D [52] 0.2196 0.2345 0.2503 0.2671 0.2944 0.3640 0.4065 0.4775 0.5603
Quasi-3D [51] 0.2196 0.2345 0.2503 0.2671 0.2944 0.3640 0.4065 0.4775 0.5603
Quasi-3D [54] 0.2142 0.2285 0.2438 0.2601 0.2866 0.3370 0.3964 0.4664 0.5485
HSDT [54] 0.2063 0.2205 0.2356 0.2516 0.2776 0.3266 0.3838 0.4504 0.5281
Present work 0.2072 0.2214 0.2356 0.2525 0.2786 0.3277 0.3851 0.4521 0.5302
2 Quasi-3D [18] 0.4552 0.4867 0.5201 0.5555 0.6126 0.7201 – – –
Quasi-3D [52] 0.4552 0.4867 0.5200 0.5554 0.6126 0.7201 0.8449 0.9898 1.1580
Quasi-3D [51] 0.4552 0.4867 0.5200 0.5554 0.6126 0.7201 0.8449 0.9898 1.1580
Quasi-3D [54] 0.4466 0.4773 0.5098 0.5443 0.6002 0.7058 0.8289 0.9725 1.1397
HSDT [54] 0.4351 0.4650 0.4968 0.5305 0.5852 0.6884 0.8088 0.9493 1.1129
Present work 0.4349 0.4648 0.4965 0.5303 0.5851 0.6884 0.8092 0.9502 1.1143
3 Quasi-3D [18] 0.5514 0.5897 0.6302 0.6733 0.7427 0.8730 – – –
Quasi-3D [52] 0.5514 0.5896 0.6302 0.6733 0.7427 0.8730 1.0240 1.1990 1.4017
Quasi-3D [51] 0.5514 0.5896 0.6302 0.6733 0.7427 0.8730 1.0240 1.1990 1.4017
Quasi-3D [54] 0.5418 0.5791 0.6187 0.6608 0.7289 0.8570 1.0061 1.1797 1.3813
HSDT [54] 0.5290 0.5653 0.6039 0.6449 0.7114 0.8368 0.9832 1.1540 1.3528
Present work 0.5284 0.5647 0.6033 0.6444 0.7109 0.8365 0.9832 1.1545 1.3540

Table 5 Table 6
Dimensionless normal stress σy ( a b
, ,
2 2 )
h of FGM plate subjected to sinusoidal Dimensionless normal stress σy ( a b
, ,
2 2 )
h of FGM plate subjected to sinusoidal
distributed load for δ = 4 . distributed load for δ = 2 .
b/a Theory β η Theory β

0.05 0.15 0.25 0.35 0.5 0.75 0.05 0.15 0.25 0.35 0.5 0.75

1 3D [22] 0.2247 0.2399 0.2562 0.2736 0.3018 0.3588 1 3D [22] 0.3103 0.3292 0.3495 0.3713 0.4067 0.4741
Quasi-3D [22] 0.2346 0.2510 0.2684 0.2870 0.3171 0.3739 Quasi-3D [22] 0.2955 0.3181 0.3421 0.3675 0.4085 0.4851
Quasi-3D [52] 0.2346 0.2510 0.2684 0.2870 0.3171 0.3739 Quasi-3D [52] 0.2955 0.3181 0.3421 0.3675 0.4085 0.4851
Quasi-3D [18] 0.2346 0.2510 0.2684 0.2870 0.3171 0.3739 Quasi-3D [18] 0.2955 0.3181 0.3421 0.3675 0.4085 0.4851
HSDT [55] 0.2164 0.2316 0.2477 0.2649 0.2927 0.3451 HSDT [55] 0.2522 0.2710 0.2910 0.3123 0.3477 0.4035
Present work 0.2247 0.2399 0.2562 0.2736 0.3018 0.3552 Present work 0.3103 0.3292 0.3495 0.3713 0.4067 0.4740
2 3D [22] 0.2431 0.2591 0.2762 0.2943 0.3238 0.3797 2 3D [22] 0.3200 0.3385 0.3583 0.3796 0.4142 0.4799
Quasi-3D [22] 0.2588 0.2761 0.2946 0.3143 0.3464 0.4079 Quasi-3D [22] 0.3146 0.3376 0.3620 0.3880 0.4300 0.5093
Quasi-3D [52] 0.2588 0.2761 0.2946 0.3143 0.3464 0.4079 Quasi-3D [52] 0.3146 0.3376 0.3620 0.3880 0.4300 0.5093
Quasi-3D [18] 0.2588 0.2761 0.2946 0.3143 0.3465 0.4079 Quasi-3D [18] 0.3146 0.3376 0.3620 0.3880 0.4300 0.5093
HSDT [55] 0.2294 0.2454 0.2624 0.2805 0.3097 0.3647 HSDT [55] 0.2539 0.2723 0.2919 0.3128 0.3469 0.4064
Present work 0.2431 0.2591 0.2762 0.2943 0.3238 0.3797 Present work 0.3200 0.3385 0.3583 0.3796 0.4142 0.4799
3 3D [22] 0.2319 0.2469 0.2629 0.2800 0.3077 0.3602 3 3D [22] 0.3081 0.3252 0.3436 0.3633 0.3953 0.4562
Quasi-3D [22] 0.2493 0.2656 0.2831 0.3017 0.3323 0.3911 Quasi-3D [22] 0.3042 0.3261 0.3493 0.3741 0.4143 0.4904
Quasi-3D [52] 0.2493 0.2656 0.2831 0.3017 0.3323 0.3911 Quasi-3D [52] 0.3042 0.3261 0.3493 0.3741 0.4143 0.4904
Quasi-3D [18] 0.2493 0.2656 0.2831 0.3018 0.3323 0.3911 Quasi-3D [18] 0.3042 0.3261 0.3493 0.3741 0.4143 0.4904
HSDT [55] 0.2162 0.2312 0.2472 0.2642 0.2917 0.3435 HSDT [55] 0.2368 0.2539 0.2721 0.2914 0.3469 0.3788
Present work 0.2319 0.2469 0.2629 0.2800 0.3077 0.3602 Present work 0.3081 0.3252 0.3436 0.3633 0.3953 0.4562
4 3D [22] 0.2247 0.2392 0.2546 0.2710 0.2977 0.3482 4 3D [22] 0.3008 0.3173 0.3349 0.3537 0.3844 0.4426
Quasi-3D [22] 0.2429 0.2586 0.2754 0.2934 0.3230 0.3800 Quasi-3D [22] 0.2974 0.3186 0.3412 0.3653 0.4045 0.4786
Quasi-3D [18] 0.2429 0.2586 0.2754 0.2934 0.3230 0.3801 Quasi-3D [18] 0.2974 0.3186 0.3413 0.3653 0.4045 0.4786
HSDT [55] 0.2082 0.2226 0.2380 0.2544 0.2808 0.3307 HSDT [55] 0.2272 0.2435 0.2610 0.2795 0.3097 0.3634
Present work 0.2247 0.2392 0.2546 0.2710 0.2977 0.3482 Present work 0.3009 0.3173 0.3349 0.3537 0.3843 0.4426
5 3D [22] 0.2206 0.2348 0.2498 0.2659 0.2920 0.3413 5 3D [22] 0.2967 0.3128 0.3299 0.3483 0.3782 0.4350
Quasi-3D [22] 0.2391 0.2545 0.2710 0.2886 0.3176 0.3737 Quasi-3D [22] 0.2935 0.3144 0.3366 0.3603 0.3988 0.4719
Quasi-3D [18] 0.2391 0.2545 0.2710 0.2887 0.3177 0.3737 Quasi-3D [18] 0.2935 0.3144 0.3366 0.3603 0.3989 0.4719
HSDT [55] 0.2037 0.2178 0.2329 0.2488 0.2747 0.3235 HSDT [55] 0.2219 0.2378 0.2548 0.2729 0.3024 0.3549
Present work 0.2206 0.2348 0.2498 0.2659 0.2920 0.3413 Present work 0.2967 0.3128 0.3299 0.3483 0.3782 0.4350
6 3D [22] 0.2181 0.2321 0.2470 0.2628 0.2886 0.3373 6 3D [22] 0.2943 0.3101 0.3270 0.3451 0.3746 0.4305
Quasi-3D [22] 0.2369 0.2520 0.2683 0.2857 0.3144 0.3699 Quasi-3D [22] 0.2912 0.3118 0.3339 0.3573 0.3955 0.4679
Quasi-3D [18] 0.2369 0.2521 0.2683 0.2858 0.3144 0.3699 Quasi-3D [18] 0.2912 0.3118 0.3339 0.3573 0.3955 0.4679
HSDT [55] 0.2010 0.2149 0.2298 0.2455 0.2711 0.3192 HSDT [55] 0.2187 0.2345 0.2512 0.2690 0.2980 0.3498
Present work 0.2181 0.2321 0.2470 0.2628 0.2886 0.3372 Present work 0.2943 0.3101 0.3270 0.3451 0.3746 0.4305

10E0 h3
HSDT solutions of Akavci and Tanrikulu [54], Mantari and Guedes
Soares [55]and Carrera et al. [57], the numerical solutions (boundary
The dimensionless transverse displacements (W = a4p0
w ( a b h
, ,
2 2 2 ))
of the plate center for δ = 10, 4, and 2 are presented in Tables 1–3,
element methods (BEM) and finite-element methods (FEM)) of Vaghefi respectively. The results are in excellent agreement with the exact 3D
et al. [5]. In addition, the results obtained for FGM plates subjected to [22] solution for a moderately thick plate (δ = 10 ) and a thick plate
uniform load are compared with the numerical solutions of Vaghefi (δ = 2, 4 ) with almost these present results virtually identical with the
et al. [5], the quasi-3D solution of Akavci and Tanrikulu [54] and the exact solutions presented in the literature; in some cases there are small
semi-analytical solution of Zhang et al. [23]. and negligible differences.

77
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

Table 7 results, produces relatively different measures for moderately thick

Dimensionless normal stress σx ( a b
, ,
2 2 )
h of FGM plate subjected to sinusoidal plates, with difference values between 5.6 - 10.3 percent that increase
distributed load for δ = 10 . with increase in aspect ratio. The quasi-3D results of σy for square thick
plates with δ = 4 were pretty close to present results, although results
η Theory β
obtained by quasi-3D theory for rectangular plates with the same
0.05 0.15 0.25 0.35 0.5 0.75 thickness are considerably different, with the high difference value
(6.4–9.7 percent) that increases with increase of aspect ratio. Also, the
1 Quasi-3D [22] 0.2196 0.2345 0.2503 0.2671 0.2944 0.3460 σy results of quasi-3D theory for thick plates with δ = 2 is comparably
Quasi-3D [52] 0.2196 0.2345 0.2503 0.2671 0.2944 0.3460 close to the present results except for a rectangular plates with β = 0.75.
Quasi-3D [18] 0.2196 0.2345 0.2503 0.2671 0.2944 0.3460
HSDT [55] 0.2062 0.2204 0.2355 0.2515 0.2774 0.3264
In general, the differences can be attributed to simplifying assumptions
Present work 0.2072 0.2214 0.2365 0.2525 0.2786 0.3277 associated with particular functions in quasi-3D theory for variation of
2 Quasi-3D [22] 0.2441 0.2599 0.2768 0.2949 0.3244 0.3810 shear deformations along the plate's thickness.
Quasi-3D [52] 0.2441 0.2599 0.2768 0.2949 0.3244 0.3810 Comparing HSDT and the present solution shown in Tables 4–6, it
Quasi-3D [18] 0.2441 0.2600 0.2769 0.2949 0.3244 0.3811
can be concluded that HSDT gives close results, with maximum dif-
HSDT [55] 0.2225 0.2378 0.2541 0.2713 0.2993 0.3521
Present work 0.2246 0.2399 0.2561 0.2734 0.3014 0.3543 ference value of 1.1 percent, of σy for moderately thick plates compared
3 Quasi-3D [22] 0.2347 0.2495 0.2654 0.2825 0.3104 0.3645 to the present results. While HSDT theory also gives close results of σy
Quasi-3D [52] 0.2347 0.2495 0.2654 0.2825 0.3104 0.3645 for square thick plate with δ = 4 , for rectangular thick plates with the
Quasi-3D [18] 0.2347 0.2496 0.2654 0.2825 0.3104 0.3645 same thickness and low inhomogeneity, the difference value is not
HSDT [55] 0.2104 0.2248 0.2402 0.2565 0.2829 0.3328
negligible; the maximum difference between HSDT and present σy re-
Present work 0.2128 0.2272 0.2426 0.2589 0.2854 0.3354
sults is 7.8 percent for β = 0.05 and η = 6. Moreover, HSDT compared
to present method, are somewhat different for a thick plate with δ = 2 .
It can also be observed that there is only small differences, about The difference between HSDT and the present results are in a range of
0.3–4.3 percent that can be ignored, between the present results and 14.9–25.7 percent. In fact, HSDT also underestimates dimensionless
quasi-3D [18,22,51–54] theories for a moderately thick plate and thick normal stress compared to the present exact solution for a rectangular
plate. Comparing HSDTs [54,55] with the 3D exact solution shows that thick plate with δ = 4 with low inhomogeneity and δ = 2 . It can be
the results for dimensionless transverse displacement for thickness ra- concluded that the greater differences occur for thicker plates, higher
tios of δ = 4, 10 are very close to one another, with difference ranges aspect ratios and lower inhomogeneity values. In general, σy increases
between 0.1- 0.6 percent for δ = 10 and 0.6–3.2 percent for δ = 4 , al- with an increase in the parameter β and a decrease of the side-to-
though the results for δ = 2 show a comparatively high difference range thickness ratio δ . Also, σy increases as the aspect ratio η decreases (e.g.
η = 6 to η = 2) and reaches to its maximum value at η = 2 ; however, if
of 2.9–10.3 percent. The highest difference was found for the square
η decreases from 2 to 1, σy decreases.
plate with β = 0.05. In fact, for δ = 2 , the dimensionless transverse
h2
displacement is overestimated by HSDT compared with the present The results for dimensionless normal stress, (σx = σ
a2p0 x ( a b
, ,
2 2 )
h )
result; this is due to the thickness stretching effect ignored in HSDT for a moderately thick plate (δ = 10 ) are shown in Table 7. As can be
solution. seen, solutions for both HSDT and quasi-3D are close when compared to
In general, dimensionless transverse displacement decreases when β the present results for dimensionless normal stress, σx , for moderately
or δ increases and increases when η increases. thick plates. From Table 7 it can also be seen that σx increases when
Tables 4–6 present results for dimensionless normal stress, inhomogeneity parameter β or aspect ratio η increases.
h2
(σy = σ
a2p0 y ( a b
, ,
2 2 )
h ), of square and rectangular moderately thick and Results for dimensionless transverse shear stress
thick plates for δ = 10, 4 and 2. These results, when compared with
h
0
( b h
)
τxz = ap τxz 0, 2 , 2 for a moderately thick plate (δ = 10 ) are shown in
exact 3D solution [22], quasi-3D theories [18,22,52] and HSDTs [55], Table 8 where it can be observed that results for the present solution,
exhibit virtually no difference for moderately thick and thick plates, the HSDT, and quasi-3D are close to one another. In other words, HSDT and
conclusion is that there is excellent agreement among solutions. It was quasi-3D solutions give similar results to those from the present solu-
found, however, that quasi-3D, when compared with the present tion for dimensionless transverse shear stress results for moderately

Table 8
b
Dimensionless shear stress τxz 0, 2 , ( h
2 ) of FGM plate subjected to sinusoidal distributed load for δ = 10.
η Theory β

0.05 0.15 0.25 0.35 0.5 0.75 1 1.25 1.5

1 Quasi-3D [18] 0.2454 0.2450 0.2442 0.2430 0.2405 0.2344 – – –

Quasi-3D [52] 0.2454 0.2450 0.2442 0.2430 0.2405 0.2344 0.2263 0.2162 0.2045
Quasi-3D [51] 0.2454 0.2450 0.2442 0.2430 0.2405 0.2344 0.2263 0.2162 0.2045
Quasi-3D [54] 0.2367 0.2364 0.2359 0.2353 0.2338 0.2300 0.2249 0.2182 0.2102
HSDT [54] 0.2434 0.2430 0.2422 0.2410 0.2385 0.2324 0.2242 0.2140 0.2023
Present work 0.2383 0.2381 0.2376 0.2368 0.2353 0.2315 0.2262 0.2194 0.2113
2 Quasi-3D [18] 0.3927 0.3290 0.3908 0.3889 0.3849 0.3752 – – –
Quasi-3D [52] 0.3927 0.3921 0.3908 0.3889 0.3849 0.3752 0.3621 0.3460 0.3273
Quasi-3D [51] 0.3927 0.3921 0.3908 0.3889 0.3849 0.3752 0.3621 0.3460 0.3273
Quasi-3D [54] 0.3790 0.3787 0.3779 0.3768 0.3744 0.3684 0.3602 0.3496 0.3368
HSDT [54] 0.3896 0.3889 0.3877 0.3857 0.3817 0.3719 0.3588 0.3425 0.3237
Present work 0.3815 0.3811 0.3804 0.3792 0.3767 0.3707 0.3623 0.3515 0.3385
3 Quasi-3D [18] 0.4418 0.4411 0.4396 0.4375 0.4330 0.4222 – – –
Quasi-3D [52] 0.4418 0.4411 0.4396 0.4375 0.4330 0.4221 0.4074 0.3893 0.3683
Quasi-3D [51] 0.4418 0.4411 0.4396 0.4375 0.4330 0.4221 0.4074 0.3893 0.3683
Quasi-3D [54] 0.4265 0.4261 0.4252 0.4239 0.4212 0.4146 0.4053 0.3934 0.3789
HSDT [54] 0.4383 0.4376 0.4361 0.4340 0.4294 0.4185 0.4036 0.3854 0.3642
Present work 0.4293 0.4288 0.4280 0.4267 0.4239 0.4171 0.4076 0.3955 0.3809

78
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

Table 9
Dimensionless transverse displacement wc (h) and dimensionless normal stress σx c of FGM square plate subjected to uniform load.
h Theory E0/ Eh
a

0.1 0.5 1 2 10

0.2 wc BEM [5] 4.0916 8.9751 12.5997 17.6640 39.0605

BEM [5] 4.1596 8.9357 12.6375 17.8397 39.0385
FEM [5] 4.1215 9.0047 12.6134 17.1718 39.1558
quasi-3D [54] 3.8333 8.8724 12.5970 17.7440 38.3330
Present work 3.8608 8.6203 12.1620 17.1860 38.5090
σxc BEM [5] −15.3564 −9.2902 −7.4462 −5.9410 −3.4665
BEM [5] −15.4390 −9.3279 −7.4199 −5.9711 −3.4944
FEM [5] −15.4032 −9.2995 −7.4588 −5.9591 −3.4805
quasi-3D [54] −16.3220 −9.6545 −7.6944 −6.1109 −3.4530
Present work −15.651 −9.2393 −7.3419 −5.8147 −3.3118
0.3 wc BEM [5] 0.9707 2.1378 2.9853 4.1208 8.7134
BEM [5] 0.9688 2.1498 2.9603 4.1098 8.9623
FEM [5] 0.9732 2.1407 2.9792 4.1333 8.7293
quasi-3D [54] 0.8923 2.0834 2.9602 4.1669 8.9229
3D [23] 0.9735 2.1405 2.9795 4.1332 8.7343
Present work 0.8455 1.9410 2.7468 3.8698 8.4327
σxc BEM [5] −7.2230 −4.3084 −3.4496 −2.7499 −1.6449
BEM [5] −7.2034 −4.2943 −3.4959 −2.7556 −1.6499
FEM [5] −7.2639 −4.3378 −3.4681 −2.7673 −1.6499
quasi-3D [54] −7.6576 −4.5062 −3.5748 −2.8235 −1.5731
3D [23] −7.1493 −4.3227 −3.4710 −2.7853 −1.6759
Present work −7.2414 −4.2323 −3.3502 −2.6490 −1.5339

thick plates. It can also be observed that τxz decreases with an increase
in parameter β and a decrease of η.
The dimensionless in-plane displacement ( uc = Gh u
ph
b
2 (0,
,
h
4 )) and
Table 9 presents the dimensionless transverse displacements
the dimensionless transverse displacement (wc = Gh w a b
(,
ph 2 2 ,
h
2 )) for
δ = 3,20,100 are shown in Table 10. HSDT gives a larger error than 5
(
a b
) a b
( )
(wc = Gh w ph 2 , 2 , h ) and stresses (σx c = σx p 2 , 2 , h ) of the FGM plate
percent for thick plate (δ = 3) for both in-plane and transverse dis-
for different E0 ratios. Table 9 shows that, while the results are close to placement; while, the error is negligible for thin plate (δ = 20,100 ). In
Eh
those found in the literature, in some cases, such as the dimensionless fact, HSDT gives more accurate results for thin plates; however, quasi-
h E
transverse displacement results for a = 0.3 and E0 = 0.1 for numerical 3D brings closer results to the present study, for thick plate. This is due
h
solution [5], and the semi-analytical solution [23], the difference value to the neglect of the thickness stretching effect in HSDT. According to
is nearly 15.1 percent and not negligible. Overall, this difference can be Table 10, increase in thickness ratio (δ ) causes displacements increase
attributed to the simplifying assumptions in quasi-3D theory [54] and in FGM plate. Displacement significantly increases for thin plate with
computational error in the numerical solution and the semi-analytical δ = 100 .
solution. From these results, it can be concluded that when E0 ratios
Eh Table 10 Presents dimensionless transverse displacement wc 2 for
h
()
increase, the dimensionless transverse displacements increase while the thick plate with δ = 3. As can be seen in Table 11, The difference be-
dimensionless stresses decrease. tween results of present study and the literature is less than 1 percent.
But for the plate with β = 5∗10−6 , the difference is about 5 percent. In
Table 10 overall, the results are in excellent agreement with the literature re-
Dimensionless transverse displacement wc ( ) and dimensionless in-plane dis-
h
2
sults.
Comparing present results with other methods confirms that the
placement uc ( ) of FGM square plate subjected to sinusoidal load. (β = 0.05).
h
4
accuracy of the presented method is very high. Some small differences
Method δ=3 δ = 20 δ = 100 that can be observed between the present results and other methods are
due to simplifying assumptions such as: considering particular functions
uc wc uc ∗10−2 wc ∗10−3 uc ∗10−4 wc ∗10−6
for variations of shear deformations across the thickness of the FGM
HSDT [57] 0.2273 1.4889 0.7343 1.8388 0.9197 1.1339 plate in HSDT and quasi-3D analytical theories, ignoring plate thickness
Quasi-3D [57] 0.2453 1.4145 0.7356 1.8363 0.9198 1.1340 stretching effects of in HSDTs, and computational error in numerical
FSDT [57] 0.2483 1.4929 0.7358 1.8388 0.9197 1.1339 and semi-analytical solutions.
Present work 0.2456 1.4146 0.7356 1.8361 0.9197 1.1339

3.2. Result and discussion

Table 11 To investigate effects of aspect and thickness ratios and material

Dimensionless transverse displacement wc ( a b h
, ,
2 2 2 ) of FGM square plate sub- properties on the plate response, the results of solution of the governing
jected to sinusoidal load for δ = 3. equations for the bending problem of simply-supported FGM plates
subjected to uniform load are presented in the following sections.
β 3D [21] 3D [56] BEM [5] BEM [5] Present work

5*10−2 1.4146 1.4146 1.4096 1.4229 1.4146 3.2.1. Effect of aspect and thickness ratio
5*10−3 1.3496 1.3496 1.3448 1.3573 1.3496
5*10−4 1.3433 1.3433 1.3384 1.3509 1.3433
In this section, the effects of aspect ratio (η ) and thickness ratio (δ)
5*10−5 1.3426 1.3426 1.3378 1.3503 1.3402 on the response of the FGM plates are investigated. Fig. 2 is a plot of the
5*10−6 1.3426 1.3426 1.3377 1.3502 1.2825 E0 h3
dimensionless transverse displacement (wc = w) versus in-
12(1 − ν ) a4p0
homogeneity parameter (β ) for different thickness ratio (δ ) and aspect

79
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

Fig. 2. Dimensionless central transverse displacement versus parameter β for FGM plate subjected to uniform load.

Fig. 3. Dimensionless maximum compressive stress, σx c , versus parameter β for FGM plate subjected to uniform load.

ratio (η ). It can be observed from Fig. 2 that, with an increase of
thickness and aspect ratio the dimensionless transverse displacement
increases. Furthermore, if β increases the dimensionless transverse
displacement decreases, and the reduction rate represented by the
graph's slope increases with decreasing δ and increasing η. These var-
iations represent notable effects of higher thickness and aspect ratio on
inhomogeneity parameter changes. Fig. 3 shows the dimensionless
maximum compressive stress (σx p ) function of the inhomogeneity
parameter for η = 1, 2 and different thickness ratios (δ ). It can be ob-
served that dimensionless maximum compressive stress increases when
δ , η, and β increase. Also, from the graph's slope it can be concluded
that greater effects of inhomogeneity parameter variations have a sig-
nificant effect on the responses for the FGM plate with lower thickness
and higher aspect ratio.
Fig. 4 shows the distribution of the dimensionless transverse shear
stress (τxz p ) through the thickness of square FGM plate for variable

Fig. 4. Distribution of the dimensionless transverse shear stress, τxz

p ( , ), through the thickness of the square FGM plate subjected to uniform load (β = 0.5).
a
4
b
4

80
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

Fig. 5. Dimensionless central bending moment versus parameter β for different thickness ratio for square FGM plate subjected to uniform load.

Fig. 6. Dimensionless corner twisting moment versus parameter β for different thickness ratio for square FGM plate subjected to uniform load.

thickness ratio (δ ). As can be seen, the amount of τxz p in top and bottom
surface is zero. Also this figure illustrates, with an increase of δ , the
amount of dimensionless transverse shear stress increases, and it can
also be observed that at higher thickness ratio, variations of the τxz p
through the thickness is more significant.
Figs. 5 and 6 show the dimensionless central bending moment
(Mx /(pa2 )) and dimensionless corner twisting moment (Mxy /(pa2 )), re-
spectively, versus the inhomogeneity parameter for different thickness
ratios in a square FGM plate.
As can be seen in the figures, an increase of β has just a small effect
on variations in dimensionless bending moment and dimensionless
twisting moment, so that, for constant thickness, the results are in close
agreement with one another for different inhomogeneity parameters. At
high thickness (δ = 1) by increase of β both the dimensionless bending
moment and the twisting moment increase, while at low thickness
(δ = 100 ) an increase in β at higher inhomogeneity parameter values
leads to a decrease in the dimensionless bending moment and an in-
crease in the dimensionless twisting moment. Fig. 7 also shows the
distribution of the dimensionless twisting moment along the edge of
y = 0 for a square FGM plate. From this figure it can be observed that
for δ = 1 the maximum values of the dimensionless twisting moment
occurs at the edges, while for the FGM plate with δ = 2 the maximum

Fig. 7. Distribution of the corner twisting moment for square FGM plate subjected to uniform load for different thickness ratio (β = 0.5).

81
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

Fig. 8. Distribution of dimensionless transverse displacement, wc ( , ), through the length of the thick square FGM plate subjected to uniform load (δ = 2).
a
2
h
2

Fig. 9. Distribution of dimensionless normal stress, σx

p ( , ), through the thickness of the thick square FGM plate subjected to uniform load (δ = 2).
a
2
b
2

value moves away from the edge and gradually decreases to zero.
However, by increasing the thickness ratio the maximum dimensionless
twisting moment occurs at the edge. Fig. 7 also clearly shows that
highest dimensionless twisting moment is obtained for the FGM plate
with δ = 1, and it can also be seen that for a FGM plate with δ = 1, the
x 3.2.1.1. Effect of inhomogeneity parameter. Figs. 8–12 display variations
dimensionless twisting moment is negative in the range 0 < a < 0.5 and
x
positive in the range 0.5 < a < 1 , while with an increase of thickness
ratio the sign is reversed so that the dimensionless twisting moment is
x x
positive in the range 0 < a < 0.5 and negative in the range 0.5 < a < 1.
of transverse displacement along the length, and variations through

Fig. 10. Distribution of dimensionless vertical normal stress, σz

p ( , ), through the thickness of the thick square FGM plate subjected to uniform load (δ = 2).
a
2
b
2

82
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

Fig. 11. Distribution of dimensionless shear stress, τxy

p ( , ), through the thickness of the thick square FGM plate subjected to uniform load (δ = 2).
a
4
b
4

Fig. 12. Distribution of dimensionless transverse shear stress, τxz

p ( , ), through the thickness of the thick square FGM plate subjected to uniform load (δ = 2).
a
4
b
4

thickness, of normal and shear stresses for the thick FGM square plate
for different inhomogeneity parameters. Fig. 8 demonstrates that the
dimensionless transverse displacement decreases with an increase in
the inhomogeneity parameter, and that there is maximum
dimensionless transverse displacement at the center of the plate.
Fig. 9 shows that an increasing inhomogeneity parameter results in
compressive and tension maximum dimensionless normal stress
increasing and decreasing, respectively, and zero normal stress moves
to the top surface. Fig. 10 shows that dimensionless vertical stress (σz p )
increases with an increase of the inhomogeneity parameter. It should be
noted that maximum vertical stress has been calculated with 2% error.
Fig. 11 shows that positively or negatively increasing in-
homogeneity parameter causes dimensionless shear stress (τxy p ) to in-
crease or decrease, respectively, and zero dimensionless shear stress
also moves to the top surface. Fig. 12 shows that the distribution of
dimensionless transverse shear stress along the thickness for different

Fig. 13. Dimensionless bending moment (Mx /(pa2 )) on the line y = b/2 for square FGM plate subjected to uniform load (δ = 1).

83
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

Fig. 14. Dimensionless bending moment (Mx /(pa2 )) on the line y = b/2 for square FGM plate subjected to uniform load (δ = 2 ).

Fig. 15. Dimensionless bending moment (Mx /(pa2 )) on the line y = b/2 for square FGM plate subjected to uniform load (δ = 5).

amounts of inhomogeneity parameter is polynomial. Also can be seen
that approximately from bottom (z t = 0 ) to mid surface (z t = 0.5) in-
crease of the inhomogeneity parameter leads to decrease of τxz p and
from mid to top surface (z t = 1) results in increase of τxz p . Furthermore
Fig. 12 shows that the maximum τxz p which occurs at mid surface for
β = 0 , with increase of β takes place at upper surfaces (z t > 0.5). For
instance, the maximum τxz p for β = 1 occurs at z t = 0.66 .
As shown in Figs. 8–12, displacement and stress variations are not
linearly related to the thickness coordinate z because the plate is non-
homogeneous with material properties that are functions of z.
Figs. 13–16 show the distribution of the dimensionless bending
b
moment along y= 2 for different thickness ratios in square FGM plates.
The effect of an increase of inhomogeneity parameter can be seen only
for thick plates (δ = 1,2 ) where for an increase of β the dimensionless
bending moment decreases. It is also clear that the maximum di-
a
mensionless bending moment occurs at x = 2 and that with increase of
thickness ratio the amount of dimensionless bending moment de-
creases.
Figs. 17–20 shows the distribution of the dimensionless twisting
moment along the edge y= 0 for both moderately thick and thick square

Fig. 16. Dimensionless bending moment (Mx /(pa2 )) on the line y = b/2 for square FGM plate subjected to uniform load (δ = 10 ).

84
Z. Vafakhah, B. Navayi Neya Composites Part B 156 (2019) 72–87

Fig. 17. Dimensionless twisting moment (Mxy /(pa2 )) on the line y = 0 for square FGM plate subjected to uniform load (δ = 1).

Fig. 18. Dimensionless twisting moment (Mxy /(pa2 )) on the line y = 0 for square FGM plate subjected to uniform load (δ = 2 ).

FGM plates. Similarly to Figs. 13–16, these figures show that an in-
crease of β affects only thick plates (δ = 1,2 ). For a thick plate with
δ = 1, with increase of β the dimensionless twisting moment decreases.
For a plate with δ = 2 the maximum values of the dimensionless
twisting moment are obtained at x/a = 0.2 and x/a = 0.8. Further-
more, from these figures it can be observed that the zero dimensionless
a
twisting moment occurs at x = 2 except for a plate with δ = 2 and
β = 0, 0.05, 0.25, 0.5 and that there are two other surfaces with zero
value near the edges.
4. Conclusion and recommendation
This paper successfully investigated using the displacement poten-
tial function method to determine the exact bending solution for a thick
rectangular simply-supported nonhomogeneous isotropic FGM plate.
The solution was obtained by solving 2 s and fourth order partial dif-
ferential equations. The governing equations were solved using a se-
paration-of-variable method and applying boundary conditions exactly.
The advantage of the present work in comparison with other studies
described in the literature is that the present exact solution can be

Fig. 19. Dimensionless twisting moment (Mxy /(pa2 )) on the line y = 0 for square FGM plate subjected to uniform load (δ = 5).

85
Z. Vafakhah, B. Navayi Neya
Fig. 20. Dimensionless twisting moment (Mxy /(pa2 )) on the line y = 0 for square FGM plate subjected to uniform load (δ = 10 ).
applied for any arbitrary thickness, including thin, moderately thick,
and thick rectangular plates and for different loading conditions (e.g.
uniform and sinusoidal loading). The solution of the problem was
conducted without simplifying assumptions either in obtaining the
governing equations or in applying boundary conditions, so the results
can be used to provide a benchmark for other studies.
While the numerical results obtained from this study are in very
good agreement with results in the literature obtained by other analy-
tical, semi-analytical, and numerical methods, a few differences were
observed due to simplifying assumptions considered in the other
methods. In general, for thick FGM plate, the method described here
predicts displacements and stresses more accurately than other
methods, and the results represent the effects of the inhomogeneity
parameter and thickness and aspect ratio on bending response, dis-
placements, stresses, and moments of the FGM plate.
The most important findings and recommendation of this study are:
- The increase of inhomogeneity and decrease of aspect ratio and also
thickness of the FGM plate result in decreasing the dimensionless
transverse displacement of the plate.
- The effects of increased inhomogeneity on transverse displacement
are more significant for thicker plates and higher aspect ratios.
- Maximum dimensionless compressive stress increases with in-
creasing inhomogeneity parameter and thickness, and the increased
inhomogeneity zero dimensionless normal stress would be closer to
the top surface.
- Dimensionless vertical stress increases with an increase of the in-
homogeneity parameter.
- Increasing the inhomogeneity parameter causes positive and nega-
tive dimensionless shear stress (τxy p ) to increase and decrease, re-
spectively, and increasing inhomogeneity zero dimensionless causes
shear stress to move to the top surface.
- Dimensionless transverse shear stress through of the thickness of the
FGM plate has a polynomial distribution for thin and moderately
thick plates (δ = 5,10,100 ), but with an increase in plate thickness of
the plate the graph does not show a polynomial distribution.
- In general, displacement and stresses variations are not linearly
related to the thickness coordinate. Because the plate is non-
homogeneous and material properties are functions of z.
- Increase of inhomogeneity has only a small effect on variations of
dimensionless bending moment and dimensionless twisting moment
and only affects thick plates.
References
[1] Mahamood RM, Akinlabi ET, Shukla M, Pityana S. Functionally graded material: an
overview. 2012.
[2] Thai H-T, Kim S-E. A review of theories for the modeling and analysis of
86
Composites Part B 156 (2019) 72–87
functionally graded plates and shells. Compos Struct 2015;128:70–86.
[3] Chi S-H, Chung Y-L. Mechanical behavior of functionally graded material plates
under transverse load—Part I: Analysis. Int J Solid Struct 2006;43:3657–74.
[4] Chi S-H, Chung Y-L. Mechanical behavior of functionally graded material plates
under transverse load—Part II: numerical results. Int J Solid Struct
2006;43:3675–91.
[5] Vaghefi R, Baradaran G, Koohkan H. Three-dimensional static analysis of thick
functionally graded plates by using meshless local Petrov–Galerkin (MLPG) method.
Eng Anal Bound Elem 2010;34:564–73.
[6] Jha D, Kant T, Singh R. A critical review of recent research on functionally graded
plates. Compos Struct 2013;96:833–49.
[7] Beena K, Parvathy U. Linear static analysis of functionally graded plate using spline
finite strip method. Compos Struct 2014;117:309–15.
[8] Nguyen-Xuan H, Tran LV, Thai CH, Nguyen-Thoi T. Analysis of functionally graded
plates by an efficient finite element method with node-based strain smoothing.
Thin-Walled Struct 2012;54:1–18.
[9] Thai H-T, Choi D-H. A simple first-order shear deformation theory for the bending
and free vibration analysis of functionally graded plates. Compos Struct
2013;101:332–40.
[10] Xiang S, Kang G-w. Static analysis of functionally graded plates by the various shear
deformation theory. Compos Struct 2013;99:224–30.
[11] Levinson M. An accurate, simple theory of the statics and dynamics of elastic plates.
Mech Res Commun 1980;7:343–50.
[12] Touratier M. An efficient standard plate theory. Int J Eng Sci 1991;29:901–16.
[13] Karama M, Afaq K, Mistou S. Mechanical behaviour of laminated composite beam
by the new multi-layered laminated composite structures model with transverse
shear stress continuity. Int J Solid Struct 2003;40:1525–46.
[14] Aydogdu M. A new shear deformation theory for laminated composite plates.
Compos Struct 2009;89:94–101.
[15] Mantari J, Oktem A, Soares CG. A new higher order shear deformation theory for
sandwich and composite laminated plates. Compos B Eng 2012;43:1489–99.
[16] Nguyen T-K. A higher-order hyperbolic shear deformation plate model for analysis
of functionally graded materials. Int J Mech Mater Des 2015;11:203–19.
[17] Thai H-T, Choi D-H. Efficient higher-order shear deformation theories for bending
and free vibration analyses of functionally graded plates. Arch Appl Mech
2013;83:1755–71.
[18] Thai H-T, Kim S-E. A simple quasi-3D sinusoidal shear deformation theory for
functionally graded plates. Compos Struct 2013;99:172–80.
[19] Thai H-T, Vo TP, Bui TQ, Nguyen T-K. A quasi-3D hyperbolic shear deformation
theory for functionally graded plates. Acta Mech 2014;225:951–64.
[20] Mantari J, Soares CG. A quasi-3D tangential shear deformation theory with four
unknowns for functionally graded plates. Acta Mech 2015;226:625–42.
[21] Kashtalyan M. Three-dimensional elasticity solution for bending of functionally
graded rectangular plates. Eur J Mech Solid 2004;23:853–64.
[22] Zenkour A. Benchmark trigonometric and 3-D elasticity solutions for an ex-
ponentially graded thick rectangular plate. Arch Appl Mech 2007;77:197–214.
[23] Zhang H, Jiang J-q, Zhang Z-c. Three-dimensional elasticity solutions for bending of
generally supported thick functionally graded plates. Appl Math Mech
2014;35:1467–78.
[24] Sadd MH. Elasticity: theory, applications, and numerics. Academic Press; 2009.
[25] Tran-Cong T. On the completeness and uniqueness of the Papkovich-Neuber and the
non-axisymmetric Boussinesq, Love, and Burgatti solutions in general cylindrical
coordinates. J Elasticity 1994;36:227–55.
[26] Lekhnitskii S, editor. Theory of anisotropic elastic bodies. Holden-Day Publishing
Mir Publication; 1981.
[27] Hu H-C. On the three-dimensional problems of the theory of elasticity of a trans-
versely isotropic body. 1953.
[28] Nowacki W. The stress function in three-dimeusionai problems concerning an
elastic body characterized by transverse isotropy. Relation 1954;1(3).
[29] Eskandari-Ghadi M. A complete solution of the wave equations for transversely
isotropic media. J Elasticity 2005;81:1–19.
[30] Nematzadeh M, Eskandari-Ghadi M, Navayi Neya B. An analytical solution for
Z. Vafakhah, B. Navayi Neya
transversely isotropic simply supported thick rectangular plates using displacement
potential functions. J Strain Anal Eng Des 2011;46:121–42.
[31] Eskandari-Ghadi M, Amiri-Hezaveh A. Wave propagations in exponentially graded
transversely isotropic half-space with potential function method. Mech Mater
2014;68:275–92.
[32] Eskandari-Ghadi M, Pak RY, Ardeshir-Behrestaghi A. Elastostatic Green’s functions
for an arbitrary internal load in a transversely isotropic bi-material full-space. Int J
Eng Sci 2009;47:631–41.
[33] Eskandari-Ghadi M, Pak RY, Ardeshir-Behrestaghi A. Transversely isotropic elas-
todynamic solution of a finite layer on an infinite subgrade under surface loads. Soil
Dynam Earthq Eng 2008;28:986–1003.
[34] Eskandari-Ghadi M, Fallahi M, Ardeshir-Behrestaghi A. Forced vertical vibration of
rigid circular disc on a transversely isotropic half-space. J Eng Mech
2009;136:913–22.
[35] Eskandari-Ghadi M, Sattar S. Axisymmetric transient waves in transversely isotropic
half-space. Soil Dynam Earthq Eng 2009;29:347–55.
[36] Eskandari-Ghadi M, Sture S, Pak R, Ardeshir-Behrestaghi A. A tri-material elasto-
dynamic solution for a transversely isotropic full-space. Int J Solid Struct
2009;46:1121–33.
[37] Eskandari‐Ghadi M, Mirzapour A, Ardeshir‐Behrestaghi A. Rocking vibration of a
rigid circular disc in a transversely isotropic full‐space. Int J Numer Anal Meth
GeoMech 2011;35:1587–603.
[38] P. N. Babagi, B. Navayi Neya, and M. Dehestani, "Three dimensional solution of
thick rectangular simply supported plates under a moving load," Meccanica, pp.
1-18.
[39] Bakhshandeh A, Navayi Neya B, Babagi PN. Benchmark solution for free vibration
analysis of transversely isotropic thick rectangular plates. Acta Mech 2017:1–19.
[40] Moslemi A, Navayi Neya B, Vaseghi Amiri J. 3-D elasticity buckling solution for
simply supported thick rectangular plates using displacement potential functions.
Appl Math Model 2016;40:5717–30.
[41] Moslemi A, Navayi Neya B, Vaseghi Amiri J. Benchmark solution for buckling of
thick rectangular transversely isotropic plates under biaxial load. Int J Mech Sci
2017;131:356–67.
[42] Ding H, Chen W, Zhang L. Elasticity of transversely isotropic materials 126.
Springer Science & Business Media; 2006.
87
Composites Part B 156 (2019) 72–87
[43] Rahimian M, Ghorbani‐Tanha AK, Eskandari‐Ghadi M. The Reissner–Sagoci pro-
blem for a transversely isotropic half‐space. Int J Numer Anal Meth GeoMech
2006;30:1063–74.
[44] Rahimian M, Eskandari-Ghadi M, Pak RY, Khojasteh A. Elastodynamic potential
method for transversely isotropic solid. J Eng Mech 2007;133:1134–45.
[45] Khojasteh A, Rahimian M, Eskandari M, Pak R. Asymmetric wave propagation in a
transversely isotropic half-space in displacement potentials. Int J Eng Sci
2008;46:690–710.
[46] Eskandari-Ghadi M, Sture S, Rahimian M, Forati M. Thermoelastodynamics with
scalar potential functions. J Eng Mech 2013;140:74–81.
[47] Eskandari-Ghadi M, Rahimian M, Sture S, Forati M. Thermoelastodynamics in
transversely isotropic media with scalar potential functions. J Appl Mech
2014;81:021013.
[48] Dehestani M, Vafai A, Mofid M, Ahmadi G. On the dynamic response of a half-space
subjected to a moving mass. Math Mech Solid 2012;17:393–412.
[49] Dehestani M, Vafai A, Mofid M, Szidarovszky F. Computation of the stresses in a
moving reference system in a half-space due to a traversing time-varying con-
centrated load. Comput Math Appl 2013;65:1849–62.
[50] Ugural AC. Stresses in plates and shells. WCB McGraw Hill; 1999.
[51] Mantari J, Soares CG. A novel higher-order shear deformation theory with
stretching effect for functionally graded plates. Compos B Eng 2013;45:268–81.
[52] Mantari J, Soares CG. Generalized hybrid quasi-3D shear deformation theory for the
static analysis of advanced composite plates. Compos Struct 2012;94:2561–75.
[53] Thai H-T, Choi D-H. Improved refined plate theory accounting for effect of thickness
stretching in functionally graded plates. Compos B Eng 2014;56:705–16.
[54] Akavci S, Tanrikulu A. Static and free vibration analysis of functionally graded
plates based on a new quasi-3D and 2D shear deformation theories. Compos B Eng
2015;83:203–15.
[55] Mantari J, Soares CG. Bending analysis of thick exponentially graded plates using a
new trigonometric higher order shear deformation theory. Compos Struct
2012;94:1991–2000.
[56] Brischetto S. A general exact elastic shell solution for bending analysis of func-
tionally graded structures. Compos Struct 2017;175:70–85.
[57] Carrera E, Brischetto S, Cinefra M, Soave M. Effects of thickness stretching in
functionally graded plates and shells. Compos B Eng 2011;42:123–33.