Composite Structures 284 (2022) 115130

Contents lists available at ScienceDirect

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

Application of discrete shear quadrilateral element for static bending, free

vibration and buckling analysis of functionally graded material plate
Imam Jauhari Maknun a, Sundararajan Natarajan b, Irwan Katili a, *
a
Universitas Indonesia, Civil Engineering Department, Depok 16424, Indonesia
b
Department of Mechanical Engineering, Indian Institute of Technology Madras, Chennai 600036, India

A R T I C L E I N F O A B S T R A C T

Keywords: This paper presents a discrete shear quadrilateral (DSQ) element to study the static bending, free vibration, and
Discrete shear quadrilateral buckling analysis of functionally graded material (FGM) plate structures. The effective properties of the FGM
Functionally graded material plate are computed using the rule of mixtures. The plate kinematics is based on Resinner-Mindlin plate theory
Reissner-Mindlin plate
with discrete shear constraints introduced to relate the kinematical and the independent shear strains. The ef­
Static bending
Free vibration
ficacy of the present element is demonstrated with a few examples for Static bending, free vibration, and me­
Mechanical and thermal buckling chanical and thermal buckling. The influence of various parameters such as the plate aspect ratio, material
gradient index, and boundary conditions are systematically studied. From the study, it is obvious that the DSQ
element yields accurate results and converges optimally.

1. Introduction complex shapes. Considerable research has been focussed on employing

the finite element method for analyzing plates with plate kinematics
Functionally graded materials (FGM) are a new class of engineered based on Reissner-Mindlin plate theory [6–7]. Originally, the Reissner-
material that has received considerable attention in the recent past. The Mindlin plate theory was introduced to model thick plates that show
FGM consists of two or more materials, which volume fraction of its considerable shear deformation through the thickness of the plate.
constituent materials gradually varied along a particular direction Theoretically, when the plate thickness approaches zero, the Reissner-
[1–5]. This specific feature of the FGM allows designers and engineers to Mindlin plate theory becomes Kirchhoff-Love plate theory. However,
custom tailor material properties for specific needs and has found pro­ this is not feasible with a numerical approach, and the framework suffers
found applications in a wide range of applications, especially where the from shear locking syndrome when Reissner-Mindlin plate theory is
structure is exposed to combined thermo-mechanical loading. Typical employed to model thin plates [6–7]. This has also attracted researchers
constituents of FGM are ceramic and metallic phases, wherein the to develop techniques to suppress shear locking phenomenon, such as
ceramic phase provides thermal stability and the metallic phase pro­ selective integration [8–9], reduced integration [10–11], assumed nat­
vides structural strength. Further, the gradual variation of material ural strain (ANS) [12–13], MITC element [14–15], to name a few. The
property eliminates the possibility of stress singularities or displacement discrete shear triangular (DST) element [16,17] and discrete shear
discontinuity [5] through the thickness of the plate. This is in contrast to quadrilateral (DSQ) element [18,19] are also proposed using the ANS
laminated composites, which show discontinuity in the material prop­ method. The DST and DSQ elements give overall good behavior for
erty, for example, through the thickness of the plate. analyzing isotropic and composite thin to thick plates, but the transverse
To support the development and the application of FGM for engi­ shear contribution is a bit complicated, and patch tests for very thick
neering structures, a thorough understanding of its response is essential. plates are not fully satisfied. On another front, based on the discrete
With analytical solutions available only for certain classes of problems Kirchhoff-Mindlin (DKM) approach, Katili [20–21] proposed discrete
and boundary conditions, a common practice is to use numerical tech­ Kirchhoff-Mindlin Triangle (DKMT) element and discrete Kirchhoff
niques. Of the available techniques, the finite element method (FEM) is Mindlin quadrilateral (DKMQ) element to analyze thick to thin plate
the most widely adopted due to its strong mathematical analysis and bending problems. The DKMT and DST elements converge to the DKT
robustness in treating different boundary conditions and modeling element [22] in thin plate problem, while the DKMQ and DSQ elements

* Corresponding author at: Universitas Indonesia, Civil Engineering Department, Depok 16424, Indonesia.
E-mail address: [email protected] (I. Katili).

https://doi.org/10.1016/j.compstruct.2021.115130
Received 15 September 2021; Accepted 19 December 2021
Available online 1 January 2022
0263-8223/© 2022 Elsevier Ltd. All rights reserved.
I.J. Maknun et al. Composite Structures 284 (2022) 115130

z phases are related by VC + VM = 1 and VC is expressed as:

( )
Ly z 1 n
Vc = + (1)
h 2

where h is the total thickness of the plate, n is the power-law index

y parameter, and z is the coordinate along the thickness direction that
h/2 ceramic varies from − h/2 ≤ z ≤ + h/2. Fig. 2 shows the volume fraction vari­
Lx ation with different power index n.
- h/2 By using the rule of mixtures, the effective material property through
metal the thickness of the plate can be computed using the following expres­
x sions:
( )
Fig. 1. The geometry of a rectangular FGM plate in the Cartesian coordi­ z 1
E(z) = (EC − EM ) + n + EM
nate system. h 2
( )
z 1 n
ρ(z) = (ρC − ρM ) + + ρM (2)
h 2
1 ( )
0.9 z 1
α(z) = (αC − αM ) + n + αM
h 2
0.8 n=0.1
0.7
Volume Fraction VC

n=0.3 where E represents the effective material property (Young’s Modulus), ρ

0.6 is represents the density and α represents the thermal coefficient
n=0.5
0.5
n=1 expansion. The value of n equal to 0 and ∞ represents a fully ceramic
0.4 and metal plate, respectively. Fig. 2 shows the schematic distribution the
n=2
0.3 volume of the ceramic phase through the thickness of the plate for
n=5
0.2 different plate thickness.
n=15
0.1
0 3. Formulation of DSQ element for FGMS application
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
Non- dimensional thickness (z/h) DSQ element was proposed by Batoz and Lardeur in 1990 [18,19].
The detailed formulation of this element with 3 DOF’s per node has been
Fig. 2. Variation of volume fraction. presented in [37]. To apply the DSQ formulation for FGM, we have to
include the in-plane displacements in the formulation due to continuous
converge to the DKQ element [23]. Using free formulation approach, change in the material through the thickness of the plate, which results
recently, Katili proposed a modified DST and DSQ element called DSTK in membrane-bending coupling. To accomplish this, the formulation of
[24] and DSQK [25]. Katili [20,21,24,25] demonstrated that the DKMT, the DSQ element is augmented with 2 more DOF’s per node, thus making
DKMQ, DSTK, and DSQK elements pass patch test to machine precision the total DOF’s as 5 DOF’s per node. The plate kinematics is based on the
and yield accurate results for thin to moderately thick plates. It has been first-order shear deformation, where the DOF’s are two in-plane dis­
numerically demonstrated that these elements are free from shear placements, two rotations, and one transverse displacement. These will
locking phenomena. This was further extended to plate and shell for be detailed in this section. By employing the modified Hu-Washizu
isotropic and composite structures in [26–37]. principle, we have:
The main objective of this paper is to extend the discrete shear
Π = Πint − Πext
technique to quadrilateral meshes and study the performance in the case ∫
of static bending, free vibration, and buckling (mechanical and thermal) Πext = w fz dA (3)
of functionally graded material plates. Regarding the DSQ element re­ A

sults in [37], it is interesting to evaluate its performance for FGM. The Πint = Πmint + Πbint + Πmb
int + Πsint
proposed element has 5 degrees of freedom, two in-plane displacements, With
one transverse displacement, and two rotations. The paper is organized
as follows: Section 2 presents the theoretical formulation of functionally Membrane energy :
∫
graded material plates. The DSQ element for FGM plates is presented in 1
Πmint = 〈e〉[Hm ]{e}dA
Section 3. The details of free vibration and buckling analysis are detailed 2
in Sections 4 and 5. The numerical results are then shown in Section 6.
A

For comparison, the results from [2,38–42] are employed. Concluding Bending energy :
∫
remarks are proposed in Section 7. Πbint =
1
〈χ〉[Hb ]{χ}dA
2
2. FGM materials
A
(4)
Membrane bending energy
∫ ∫
A functionally graded material plate (Fig. 1) with dimensions (Lx × Πmb
1
〈e〉[Hmb ]{χ}dA +
1
〈χ〉[Hmb ]{e} dA
int =
Ly × h) in the Cartesian coordinate system (x-y-z) is considered for this 2 2
A A
analysis. The homogenization technique considered in this work is based
Shear energy
on power-law distribution. The material on the top surface of the plate is ∫ (〈 〉 { })
assumed to be rich in ceramic phase, while the bottom surface of the 1
Πsint = γ [Hs ] γ dA
plate is metal-rich. The effective properties are computed by using the 2 − −

rule of mixtures. Let Vi (i = C, M) be the volume fraction of the con­

A

stituent materials, where the subscripts C and M refer to ceramic and The membrane strain e and curvature χ are given by:
metallic phase, respectively. The volume fraction of the respective

2
I.J. Maknun et al. Composite Structures 284 (2022) 115130

h
+
∫2
Eb = E(z)z2 dz
h (12)
−
2
( )
n2 + n + 2 EM 3
Eb = (EC − EM ) + h
4(n + 1)(n + 2)(n + 3) 12

h
∫+2 ( )
n
Emb = E(z)zdz = (EC − EM ) h2 (13)
2(n + 1)(n + 2)
− h
2

Fig. 3. Kinematic of DSQ element with 5 DOF’s per node. κ and υ are shear correction factor and Poisson ratio, which are
assumed to be constant in this paper.
The displacement u, v, w, and rotation fields βx and βy (Fig. 3) are
interpolated as follows:
∑
4 ∑
4 ∑
4
u= Ni u i ; v = Ni v i ; w = Ni wi
i=1 i=1 i=1
∑4 ∑
8
βx = Ni βx i + Pk Ck Δβsk (14)
i=1 k=5
∑4 ∑8
Fig. 4. Cosine direction on the side i-j. βy = Ni βyi + Pk Sk Δβsk
i=1 k=5
⎧ ⎫
⎨ u,x ⎬
{e} = v,y (5) where Ni are the bilinear shape functions,
⎩ ⎭
u,y + v,x 1 1
N1 = (1 − ξ)(1 − η); N2 = (1 + ξ)(1 − η)
4 4
⎧
⎨ βx,x
⎫
⎬ (15)
1 1
{χ} = βy,y (6) N3 = (1 + ξ)(1 + η); N4 = (1 − ξ)(1 + η)
⎩β + 4 4
x,y βy,x ⎭
and Pk is the incomplete quadratic function as follow.
where u is in-plane displacement in the x-direction, v is in-plane 1( ) 1 ( )
displacement in the y-direction,βx is the rotation in the z–x plane, and P5 = 1 − ξ2 (1 − η); P6 = (1 + ξ) 1 − η2
2 2
βy is the rotation in the z–y plane. The notation u,x denotes the first (16)
1( ) 1 ( )
derivatives of u with respect to x. The transverse shear strain γ , which is P7 = 1 − ξ2 (1 + η); P8 = (1 − ξ) 1 − η2
2 2
−
interpolated independently, will be detailed later. Ck and Sk are the cosines direction defined in Fig. 4.
The constitutive matrices are: Substituting Eq. (14) into Eq. (5), we have the membrane strain
⎡
1 υ 0
⎤ matrix:
[Hm ] = Dm ⎣ υ 1 0 ⎦ ; D m = Em (7) ⎧ ⎫
1 − υ2 ⎨ ex ⎬
0 0 (1 − υ)/2 {e} = ey = [Bm ]{un } (17)
⎩ ⎭
⎡ ⎤ exy
1 υ 0
Eb
[Hb ] = Db ⎣ υ 1 0 ⎦; Db = (8) ⎡ ⎤
1 − υ2
0 0 (1 − υ)/2 ⎢ ⎥
⎢ Ni,x 0 0 0 0 ⎥
⎡ ⎤ ⎢ ⎥
1 υ 0 [Bm ] = ⎢
⎢ ... 0 Ni,y 0 0 0 ...i = 1, 4 ⎥
⎥
Emb ⎢ ⎥
[Hmb ] = Dmb ⎣ υ 1 0 ⎦; Dmb = (9) ⎣ Ni,y Ni,x 0 0 0 ⎦ (18)
2(1 − υ2 )
0 0 (1 − υ)/2
[ ] Ni,x = j11 Ni,ξ + j12 Ni,η ; Ni,y = j21 Ni,ξ + j22 Ni,η
1 0 κEs
[Hs ] = Ds ; Ds = (10) 〈 〉
0 1 2(1 + υ) 〈un 〉 = ⋯ ui vi wi βxi βyi ⋯ i = 1, 2, 3, 4
with
where Ni ,x and Ni ,y denote the first derivative of the shape function with
∫+2
h
( ) respect to x and y, respectively. j11 , j12 , j21 and j22 are the components of
Em = E s = E(z)dz =
EC − EM
+ EM h (11) the inverse Jacobian matrix, where the Jacobian matrix is defined as
h
n+1 follow:
− 2
[ ] [ ]
J J12 x,ξ y,ξ
[J] = 11 = (19)
J21 J22 x,η y,η

with x and y are the coordinate, which is expressed as:

{ } ∑n=4 { }
x x
= Ni (ξ, η) i (20)
y i=1
yi

3
I.J. Maknun et al. Composite Structures 284 (2022) 115130

Upon substituting Eq. (20) into Eq. (19), we have: [ ] [[ ] [ ]]

[ ] Tβ 6×20 = Tβ1 Tβ2
1 x21 (1 − η) + x34 (1 + η) y21 (1 − η) + y34 (1 + η) ⎡ ⎤
[J] = (21) 0 0 0 0 0 0 0 0 0 0
4 x41 (1 − ξ) + x32 (1 + ξ) y41 (1 − ξ) + y32 (1 + ξ) ⎢ ⎥
⎢0 0 0 0 0 0 0 0 0 0 ⎥
⎢ ⎥
Substituting Eq. (14) into Eq. (6), the bending strain matrix is defined ⎢ ⎥
as follow: [ ] 1⎢ ⎢0 0 0 1 0 0 0 0 − 1 0 ⎥⎥
Tβ1 = ⎢ ⎥
⎧ ⎫ 4⎢0 0 0 0 0 0 0 0 0 0 ⎥
⎨ χx ⎬ [ ] ⎢ ⎥
[ ]{ } ⎢ ⎥
{χ} = χy = Bbβ {un } + BbΔβ Δβsn (22) ⎢0 0
⎣ 0 0 0 0 0 0 0 0 ⎥⎦
⎩χ ⎭
xy
0 0 0 0 1 0 0 0 0 − 1 (29)
⎡ ⎤
where, 0 0 0 0 0 0 0 0 0 0
⎡ ⎤ ⎢ ⎥
⎢0 0 0 0 0 0 0 0 0 0 ⎥
0 0 0 Ni,x 0 ⎢ ⎥
⎢ ⎥ ⎢ ⎥
[ ] ⎢ ⎥ [ ] 1⎢ ⎢0 0 0 1 0 0 0 0 − 1 0 ⎥
Bbβ = ⎢
⎢ ... 0 0 0 0 Ni,y ...i = 1, 4 ⎥
⎥ (23) Tβ2 = ⎢
⎥
⎥
⎣ ⎦ 4⎢0 0 0 0 0 0 0 0 0 0 ⎥
0 0 0 Ni,y Ni,x ⎢ ⎥
⎢ ⎥
⎢0 0 0 0 0 0 0 0 0 0 ⎥
⎣ ⎦
⎡ ⎤ 0 0 0 0 1 0 0 0 0 − 1
P k ,x C k
[ ] ⎢ ⎥ ⎡ ⎤
BbΔβ = ⎢
⎣ ... Pk ,y Sk ... k = 5, 8 ⎥ − C5 (1 − η) 0 − C7 (1 + η) 0
⎦ (24) ⎢
Pk ,y Ck + Pk ,x Sk ⎢ 0 − C6 (1 + ξ) 0 − C8 (1 − ξ) ⎥
⎥
[ ] ⎢ C5 ξ − C6 η − C7 ξ C8 η ⎥
Pk ,x = j11 Pk ,ξ + j12 Pk ,η ; Pk ,y = j21 Pk ,ξ + j22 Pk ,η TΔβ = ⎢
⎢ − S5 (1 − η)
⎥
⎥ (30)
⎢ 0 − S7 (1 + η) 0 ⎥
⎣ 0 − S6 (1 + ξ) 0 − S8 (1 − ξ) ⎦
In the DSQ element, the value is independent of shear strains γ is
− S5 ξ − S6 η − S7 ξ S8 η
defined by the equilibrium equations and constitutive equations. As in
As in [37],Δβsn is defined as:
[37], we can define γ in terms of the nodal variables:
− { }
Δβsn = [An ]{un }; [An ] = [AΔ ]− 1 [Auu ] (31)
{ } { }
γx [ ] [ ]{ }
γ = = Bsβ {un } + BsΔβ Δβsn (25) where,
γy
⎡ [ ] ⎤
With ⎡ ⎤
⎢
〈 C5 S5 〉 BsΔβ (0, − 1)
⎥
1 0 0 0 ⎢〈C [ ] ⎥
[ ] [ ][ ][ ] 2⎢ 0 1 0 0⎥ ⎢ 6 S6 〉 BsΔβ ( + 1, 0) ⎥
Bsβ (ξ, η) = H s Tj (ξ, η) Tβ ⎢
[AΔ ] = − ⎣ ⎥+⎢ [ ] ⎥ (32)
[ ] [− ] [ ][ ] (26) 3 0 0 1 0⎦ ⎢⎢ 〈 C7 S7 〉 BsΔβ (0, +1)
⎥
⎥
BsΔβ (ξ, η) = H s Tj (ξ, η) TΔβ (ξ, η) 0 0 0 1 ⎣ [ ] ⎦
−
〈 C8 S8 〉 BsΔβ ( − 1, 0)
With [Hs] is defined in Eq. (10) and
⎡ [ ] ⎤
⎡ ⎤ 〈 C5 S5 〉 Bsβ (0, − 1)
1− υ 1+υ ⎢ ⎥
[ ] D ⎢1 0 0 0 ⎢〈C [ ] ⎥
2 2 ⎥ S6 〉 Bsβ ( + 1, 0)
(27) ⎢ 6 ⎥
b⎢ ⎥
Hs =
− Ds ⎣ 1+υ 1− υ ⎦ [Auu ]4×20 = [Au ] − ⎢
⎢ [ ] ⎥
⎥ (33)
0 0 1 0 ⎢ 〈 C7 S7 〉 Bsβ (0, +1) ⎥
2 2 ⎣ ⎦
[ ]
[[ ] ] 〈 C8 S8 〉 Bsβ ( − 1, 0)
[ ] tj [0]
Tj (ξ, η) = [ ]
[0] tj [Au ] = [ [Au1 ] [Au2 ] ]
⎡ ⎤ ⎡ ⎤
j211 j212 2j11 j12 (28) 2 2
[ ] ⎢ ⎢ 2
⎥
⎥
⎢0
⎢ 0 − C5 S5 0 0 C5 S5 ⎥
⎥
tj = ⎢ j21 j222 2j21 j22 ⎥ ⎢ L5 L5 ⎥
⎣ ⎦ ⎢ ⎥
j11 j21 j12 j22 j11 j22 + j12 j21 ⎢ 2 ⎥
1⎢
⎢0 0 0 0 0 0 0 − C6 S6 ⎥
⎥
[Au1 ] = ⎢ L6 ⎥
2⎢ ⎥
⎢0 0 0 0 0 0 0 0 0 0⎥
⎢ ⎥
⎢ 2 ⎥
⎢ ⎥
⎣0 0 C8 S8 0 0 0 0 0⎦
L8
(34)
⎡ ⎤
⎢0 0 0 0 0 0 0 0 0 0⎥
⎢ ⎥
⎢ 2 ⎥
⎢0 0 C6 S6 0 0 0 0 0⎥
⎢ L6 ⎥
1⎢
⎢
⎥
⎥
[Au2 ] = ⎢ 2 2 ⎥
2⎢0 0 − C7 S7 0 0 C7 S7 ⎥
⎢ L7 L7 ⎥
⎢ ⎥
⎢ ⎥
⎢ 2 ⎥
⎣0 0 0 0 0 0 0 − C8 S8 ⎦
L8

Introducing Eq. (31) to Eq. (22), we have

4
I.J. Maknun et al. Composite Structures 284 (2022) 115130

[ ] [ ]
{χ} = [Bb ]{un }; [Bb ] = Bbβ + BbΔβ [An ] (35) Introducing (2) and (14) to (47), we have:
∏ 〈 〉 { }
Introducing Eq. (31) to Eq. (25), we have =
1
u̇n [m] u̇n (48)
{ } K
2
[ ] [ ]
γ = [Bs ]{un }; [Bs ] = Bsβ + BsΔβ [An ] (36)
− [m] is defined as follow:

Introducing Eq. (18) into Eq. (4), the membrane energy of DSQ [m] = [mu ] + [mv ] + [mw ]+
[ ] [ ] [ ]
element for an element with the area A is given by: mβ + mβ + muβ + (49)
[ x ] [ y ] [ x ]
1 mvβy + mβx u + mβy v
Πmint = 〈un 〉[km ]{un } (37)
2
The mass matrix of the DSQ element in (49) is obtained using the
where standard 2 × 2 Gauss integration scheme. With:
∫ ∫
[km ] = [Bm ]T [Hm ][Bm ]dA (38) [mu ] = ρm {Nu }〈Nu 〉dA
A
∫A
Upon substituting Eq. (35) into Eq. (4), the bending energy of the
[mv ] = ρm {Nv }〈Nv 〉dA
DSQ element can then be written as:
∫
A
1
Πbint = 〈un 〉[kb ]{un } (39) [mw ] = ρm {Nw }〈Nw 〉dA (50)
2 A∫
[ ] { }〈 〉
where the bending stiffness matrix is: mβx = ρb Nβx Nβx dA
∫ [ ] ∫
A
{ }〈 〉
[kb ] = [Bb ]T [Hb ][Bb ] dA (40) mβy = ρb Nβy Nβy dA
A A

Introducing Eq. (18) and Eq. (35) into Eq. (4), the membrane-
〈Nu 〉 = 〈 ⋯ Ni 0 0 0 0 ⋯ i = 1, 4 〉
bending coupling energy for an element with the area A is given by:
〈Nv 〉 = 〈 ⋯ 0 Ni 0 0 0 ⋯ i = 1, 4 〉
1 ( )
Πmb
int = 〈un 〉 [kmb ] + [kmb ]
T
{un } (41)
2 〈Nw 〉 = 〈 ⋯ 0 0 Ni 0 0 ⋯ i = 1, 4 〉 (51)
〈 〉
leading to the coupling term stiffness Nβx = 〈 ⋯ 0 0 0 Ni 0 ⋯ i = 1, 4 〉
∫ 〈 〉
Nβy = 〈 ⋯ 0 0 0 0 Ni ⋯ i = 1, 4 〉
[kmb ] = [Bm ]T [Hmb ] [Bb ] dA (42)
A

Upon substituting Eq. (36) into Eq. (4), the transverse shear energy of +
h
an element is written as follows: ∫2 (ρ − ρ )
C M
ρm = ρ(z)dz = + ρM h
1 n+1
Πsint = 〈un 〉[ks ]{un } (43) h
2 −
2
where the shear stiffness matrix is:
h
∫ + (52)
[ks ] = [Bs ]T [Hs ][Bs ] dA (44) ∫2
2
A
ρb = ρ(z)z dz
h
The total stiffness of the DSQ element is then: −
2
[k] = [km ] + [kmb ] + [kmb ]T + [kb ] + [ks ] (45) (
n2 + n + 2 ρ
)
ρb = (ρC − ρM ) + M h3
The stiffness matrix of the DSQ element in Eq. (45) is obtained using 4(n + 1)(n + 2)(n + 3) 12
the standard 2 × 2 Gauss integration scheme. ∫
[ ] 〈 〉
muβx = ρmb {Nu } Nβx dA
4. Free vibration analysis A∫
[ ] 〈 〉
Mvβy = ρmb {Nv } Nβy dA
For free vibration problems, we have the eigenvalue problem as:
( ) [ ]
∫A
{ } (53)
[k] − ω2 [m] {un } = {0} (46) mβx u = ρmb Nβx 〈Nu 〉dA
A∫
[ ] { }
where [k] is the total stiffness (45), [m] is the mass matrix of element
Mβy v = ρmb Nβy 〈Nv 〉dA
(49), {un } is the eigenvector, and ω is the natural frequency.
For free vibration analysis, the potential energy is supplemented
A

with the system’s kinetic energy. The kinetic energy for free vibration h
analysis is defined by: +
∫2
∏ 1∫ ( )
ρmb = ρ(z)zdz
= (u + zβx )ρ(z) ü + zβ̈x dV+
K
2 h (54)
V
(47)
−
∫ ( ) ∫ 2
1 ( ) 1 ( )
v + zβy ρ(z) v̈ + zβ̈y dV + wρ(z)ẅdV n
2 2 ρmb = (ρC − ρM ) h2
V V 2(n + 1)(n + 2)

5
I.J. Maknun et al. Composite Structures 284 (2022) 115130

[ ]
5. Buckling analysis Ni,x 0 0 0 0
[Gu ] = … … i=1− 4
Ni,y 0 0 0 0
For buckling analysis, we define the membrane strain energy asso­ [ ]
ciated with the buckling deformation as follow: 0 Ni,x 0 0 0
[Gv ] = … … i=1− 4
∏ 1∫ 1
∫
1
∫ 0 Ni,y 0 0 0
= 〈∇u〉[σ0 ]{∇u}hdA+ 〈∇v〉[σ0 ]{∇v}hdA+ 〈∇w〉[σ 0 ]{∇w}hdA [ ]
σ
2 2 2 0 0 Ni,x 0 0
∫
A
∫
A A
[Gw ] = … … i=1− 4 (59)
1 h3 1 〈 〉 { } h3 0 0 Ni,y 0 0
+ 〈∇βx 〉[σ 0 ]{∇βx } dA+ ∇βy [σ0 ] ∇βy dA (55) [ ]
2 12 2 12 [ ] 0 0 0 Ni,x 0
A A
Gβx = … … i=1− 4
⎡ ⎤ 0 0 0 Ni,y 0
σ0x τ0xy [ ]
[ ]
[σ0 ] = ⎣ ⎦ (56) 0 0 0 0 Ni,x
τ 0
σ 0 Gβy = … … i=1− 4
xy y 0 0 0 0 Ni,x
Using Eq. (14) in Eq. (55), we have: In function of normal forces, equation (56) can be expressed:
∏ 1 ⎡ ⎤
= 〈un 〉[kG ]{un } (57) 0 0
1 ⎣ Nx Nxy ⎦
σ
2 [σ0 ] = 0
(60)
h Nxy Ny0
∫
[kG ] = h [Gu ]T [σ0 ][Gu ]dA+ For mechanical load we have:
A
∫ Nx0 = α1 Ncr ; Ny0 = α2 Ncr ; Nxy
0
= α3 Ncr (61)
h [Gv ]T [σ0 ][Gv ]dA+
For uniaxial load α1 = 1, α2 = 0, α3 = 0, (60) become:
A
∫ [ ]
1 1 0
h [Gw ]T [σ0 ][Gw ] dA+ [σ0 ] = Ncr (62)
(58) h 0 0
A
∫ For biaxial load α1 = 1, α2 = 1, α3 = 0, we obtain:
h3 [ ]T [ ]
Gβx [σ0 ] Gβx dA+ [ ]
12 1 1 0
A [σ0 ] = Ncr (63)
∫[ ] h 0 1
h3 [ ]
12
Gβy T [σ0 ] Gβy dA For linear buckling problem, we have the eigenvalue problem as:
A
([k] − Ncr [kG ] ){un } = {0} (64)

where [k] is the total stiffness (45), [kG] is the geometric stiffness (58),
{un } is the eigenvector and Ncr is the buckling load parameter.
While for thermal load:

∫h/2
1
Nx0 = Ny0 = E(z) α(z)ΔT(z)dz
(1 − υ)
Table 1 − h/2
Material properties.
∫h/2 (65)
Aluminum Zirconia Zirconia Aluminum oxide Unit 1
Nx0 = Ny0 = DT Tcr ; DT = E(z) α(z)dz
(ZrO2)-1 (ZrO2)-2 (Al2O3) (1 − υ)
− h/2
E 70 200 151 380 GPa
υ 0.3 0.3 0.3 0.3 0
Nxy =0
ρ 2707 5700 3000 3800 kg/m3
α 23 × 10-6 10 × 10-6 10 × 10-6 7.2 × 10-6 1/C With α(z) is defined in (2) and by using UTR (Uniform Temperature

Y Y
L L

D C D C
L L

Mesh 4 × 4 Mesh 4 × 4

A B A B
X X

Fig. 5. Simply supported and clamped square plate.

6
I.J. Maknun et al. Composite Structures 284 (2022) 115130

Table 2
The convergence of central displacement wc for SSSS Al/ZrO2-1 square plates.
N n

0 0.5 1 2 5 10 ∞

L/h = 5
4 1.720 2.328 2.723 3.119 3.512 3.822 4.914
8 1.717 2.325 2.720 3.116 3.508 3.817 4.907
16 1.717 2.325 2.720 3.115 3.507 3.815 4.905
32 1.717 2.325 2.719 3.115 3.507 3.815 4.905
64 1.716 2.324 2.719 3.115 3.506 3.815 4.904
128 1.717 2.324 2.719 3.115 3.506 3.815 4.904
[38] 1.722 2.403 2.811 3.221 – – –
[2] 1.703 2.232 2.522 2.827 – – –
L/h = 1000
4 1.421 1.947 2.281 2.595 2.861 3.092 4.060
8 1.422 1.948 2.282 2.595 2.863 3.094 4.062
16 1.422 1.948 2.283 2.595 2.863 3.094 4.062 Fig. 6. The convergence of central displacement wC for SSSS case.
32 1.422 1.948 2.283 2.595 2.863 3.094 4.062
64 1.422 1.948 2.283 2.595 2.864 3.094 4.062
128 1.422 1.948 2.283 2.595 2.864 3.094 4.062

Table 3
The convergence of central displacement wc for CCCC Al/ZrO2-1 square plates.
N n

0 0.5 1 2 5 10 ∞

L/h = 5
4 0.779 1.038 1.210 1.399 1.617 1.776 2.227
8 0.765 1.019 1.190 1.377 1.591 1.746 2.186
16 0.761 1.015 1.185 1.371 1.584 1.739 2.176
32 0.761 1.014 1.183 1.370 1.583 1.737 2.173
64 0.760 1.013 1.183 1.369 1.582 1.736 2.172
128 0.760 1.013 1.182 1.369 1.582 1.736 2.172
[38] 0.774 1.034 1.207 1.404 – – –
[2] 0.777 1.012 1.152 1.313
L/h = 1000 Fig. 7. The convergence of central displacement wC for CCCC case.
4 0.462 0.632 0.739 0.839 0.927 1.003 1.320
8 0.448 0.613 0.718 0.816 0.901 0.974 1.279 have been analyzed. Here ceramic–metal (Al/ZrO2-1) plates are
16 0.444 0.608 0.713 0.810 0.894 0.966 1.269
considered, and their properties, including Young’s modulus and Pois­
32 0.443 0.607 0.712 0.809 0.895 0.965 1.266
64 0.443 0.607 0.711 0.809 0.892 0.960 1.266 son’s ratio, are given in Table.1. Note that Poisson’s ratio υ and shear
128 0.442 0.606 0.711 0.808 0.892 0.964 1.265 correction factor (κ = 5/6) are assumed to be constant.
Due to the symmetry, only one quarter of the plate is considered for
the analysis. A structured quadrilateral mesh N × N, with N = 4, 8, 16,
Rise), ΔT(z) = Tcr . For thermal load (60) become: 32, 64, and 128 is considered for this study. For a simply supported
[ ]
DT 1 0 plate, the boundary conditions are w = βx = 0 on side AB and w = βy =
[σ0 ] = Tcr (66) 0 on side AD, while for clamped supported, the boundary conditions are
h 0 1
u = v = w = βx = βy = 0 on side AB and AD. The symmetry conditions
For thermal buckling, we have the eigenvalue problem as:
are u = βx = 0 on side BC and v = βy = 0 on side CD. This example uses
([k] − Tcr [kG ] ){un } = {0} (67) two different length thickness ratios, viz., L/h = 5 for the thick plate and
L/h = 1000 for the thin plate. Tables 2 and 3 present the central
With Tcr is the critical buckling temperature
displacement of the DSQ element for an FGM plate with fully simple
supports (SSSS) and fully clamped supports (CCCC). The influence of
6. Numerical results and discussion power-law index, n is also systematically evaluated for the following n =
0, 0.5, 1, 2, 5, 10, and ∞.
To understand the performance of DSQ element in static bending, wc EM h 3
3
The results are normalized by wc = 12(1− υ2 )fz L4 10 .
free vibration, and buckling analysis of FGM plate structures, we
consider two different plate configurations, viz., rectangular and skew In the case of the thick plate for n = 0, 0.5, 1, 2, the results from Lee
plates. The influence of plate aspect ratio, L/h, and boundary conditions et al. [38] and Nguyen-Xuan et al. [2] are used as a comparison. It is
on the global response are systematically studied. The analytical or inferred that the DSQ element yields comparable results. From Tables 2
numerical solutions from the literature are used as a comparison to and 3, it can be seen that the power-law index has a strong influence on
demonstrate the formulation’s efficacy. Further, two different combi­ the non-dimensional center deflection increases. This can be attributed
nations of FGM are considered in this study, and the corresponding to the fact that increasing the power-law index increases the metallic
material properties are given in Table 1. volume fraction, thus increasing the flexibility.
Figs. 6 and 7 show the results of central displacement as a function of
L/h. From the Figures, it can be seen that the non-dimensional central
6.1. Static analysis deflection of the plate with a higher percentage of ceramic phase is
lower than that of the plate with a higher percentage of metallic phase.
6.1.1. Simply supported (SSSS) and clamped (CCCC) square plate under The influence of the plate thickness is also studied, and it can be seen
uniform loading that the proposed DSQ element does not show any shear locking
Fig. 5 presents simply supported and clamped square plates, which

7
I.J. Maknun et al. Composite Structures 284 (2022) 115130

SSSS ( n = 1) CCCC ( n = 1)
-1.5 -1.5
-2 -1.5 -1 -0.5 0 -2 -1.5 -1 -0.5 0

-2 -2

Log ( ES )
Log ( ES )

-2.5 -2.5

-3 -3

-3.5 -3.5

-4 L/h=5 -4 L/h=5

L/h=1000 L/h=1000

-4.5 -4.5
Log ( L k ) Log ( L k )

Fig. 8. The convergence DSQ element by using the s-norm test in a square plate.

syndrome.
The s-norm test is used to understand the convergence behavior of
the plate structure. The details of this norm can be found in [37], where
the relative error is computed by:
∏ ∏FE 2
‖ Ref
int − int ‖
Es = ∏ 2
‖ Refint ‖

Fig. 8 presents the convergence of the DSQ element for the FGM plate
Fig. 9. Skew plate for static analysis. with n = 1 by using s-norm test. The reference solutions are taken from
the DSQ element results with N × N = 128 × 128. The figure shows that
Table 4 the DSQ element results well for both thin and thick plate problems.
The convergence of central displacement wc for SSSS Al/ZrO2-1 skew plates (L/h
= 100).
Table 5
N n The convergence of central displacement wc for CCCC Al/ZrO2-1 skew plates (L/
0 0.5 1 2 5 10 ∞ h = 100).

θ ¼ 30 ◦ N n
4 4.245 5.707 6.550 7.352 8.233 9.063 12.129
0 0.5 1 2 5 10 ∞
8 2.823 3.846 4.471 5.057 5.611 6.106 8.079
16 2.471 3.377 3.946 4.477 4.949 5.361 7.059 θ ¼ 30◦
32 2.371 3.245 3.798 4.314 4.765 5.155 6.774 4 2.597 3.447 3.903 4.339 4.909 5.472 7.419
64 2.344 3.209 3.757 4.271 4.717 5.101 6.696 8 1.116 1.501 1.723 1.932 2.162 2.379 3.188
128 2.344 3.208 3.758 4.273 4.721 5.105 6.696 16 0.730 0.993 1.155 1.305 1.447 1.574 2.085
[38] – – – – – – 6.747 32 0.634 0.867 1.014 1.151 1.272 1.377 1.812
θ ¼ 45◦ 64 0.614 0.841 0.985 1.120 1.237 1.337 1.755
4 9.197 12.493 14.506 16.396 18.214 19.847 26.278 128 0.611 0.837 0.980 1.115 1.231 1.331 1.746
8 7.949 10.861 12.686 14.395 15.921 17.252 22.712 θ ¼ 45◦
16 7.611 10.415 12.191 13.849 15.297 16.548 21.745 4 4.273 5.725 6.548 7.331 8.235 9.095 12.209
32 7.509 10.282 12.042 13.687 15.115 16.343 21.455 8 2.673 3.629 4.212 4.760 5.289 5.765 7.636
64 7.498 10.267 12.027 13.675 15.106 16.330 21.423 16 2.257 3.083 3.602 4.088 4.520 4.897 6.447
128 7.529 10.308 12.077 13.735 15.175 16.405 21.511 32 2.149 2.943 3.445 3.915 4.323 4.675 6.141
[38] – – – – – – 21.544 64 2.125 2.910 3.409 3.876 4.279 4.625 6.070
θ ¼ 60◦ 128 2.120 2.903 3.402 3.868 4.270 4.615 6.056
4 15.333 20.933 24.436 27.715 30.669 33.256 43.810 θ ¼ 60◦
8 14.652 20.052 23.469 26.662 29.448 31.856 41.864 4 6.039 8.161 9.425 10.617 11.844 12.969 17.255
16 14.468 19.813 23.206 26.377 29.120 31.480 41.334 8 4.775 6.515 7.599 8.616 9.540 10.352 13.644
32 14.427 19.760 23.150 26.320 29.057 31.407 41.219 16 4.435 6.069 7.103 8.069 8.913 9.643 12.672
64 14.457 19.800 23.198 26.381 29.132 31.488 41.306 32 4.346 5.952 6.972 7.925 8.749 9.458 12.417
128 14.514 19.876 23.287 26.484 29.250 31.616 41.468 64 4.324 5.923 6.940 7.890 8.709 9.413 12.354
[38] – – – – – – 41.467 128 4.319 5.917 6.933 7.882 8.700 9.402 12.340

8
I.J. Maknun et al. Composite Structures 284 (2022) 115130

element. Consider a skew plate made up of Al/ZrO2-1 with various skew

angle θ, as shown in Fig. 9. The skew plate (the length L and thickness h)
is subjected to a uniform transverse load fz, and the thickness of the plate
is assumed to be 100.
Two boundary conditions used are soft simply supported (u = v = w
= 0 on the side ABCD) and fully clamped (u = v = w = βx = βy = 0 on the
side ABCD).
The convergence of the central displacement is normalized by wc =
16wc EC h3
12(1− υ2 )fz L4
103 .
Table 4 and Table 5 present the central displacement of a soft simply
supported and fully clamped skew plate for different power index n and
skew angle (θ = 30, 45, and 60). From the result, we found that the DSQ
element gives good results. Figs. 10 and 11 show the results of the DSQ
Fig. 10. Simply supported (SSSS) skew plate θ = 45. element for the SSSS case with θ = 45◦ and θ = 60◦ . It can be seen that
starting from N = 16 the curves are flat. The DSQ element gives good
convergence results with a small number of elements.
Fig. 12 presents the convergence behavior result using the s–norm
test. Using n = 2 and the DSQ element results with N × N = 128 × 128 as
reference solution, the DSQ element gives a uniform and optimal
convergence for thick and thin plate problems.

6.2. Free vibration analysis

6.2.1. Square plate

For free vibration analysis, consider the simply supported (SSSS) and
clamped (CCCC) Al/Al2O3 square plate shown in Fig. 5. Due to the
symmetry, only one-quarter of the plate is considered for the analysis.
The boundary and symmetry conditions used are similar to the problems
Fig. 11. Clamped (CCCC) skew plate θ = 60. analyzed in Section 6.1.1. The boundary conditions for the hard simply
supported plate are u = v = w = βx = 0 on side AB and u = v = w = βy =
0 on side AD, while for the clamped supported plate, the boundary
6.1.2. Soft simply supported (SSSS) and clamped (CCCC) skew plate under
conditions are u = v = w = βx = βy = 0 on side AB and AD. The sym­
uniform loading
In this test, the influence of the skewness of the plate on the global metry conditions are u = βx = 0 on side BC and v = βy = 0 on side CD
response of FGM plate structure is studied using the proposed DSQ (Fig. 5). The influence of the boundary conditions and the plate thick­
ness on the free vibration behavior if systematically studied

SSSS (n = 1) CCCC (n = 1)
-2 -1.5 -1 -0.5 0 -2 -1.5 -1 -0.5 0
0.5 0.5
Log ( ES )

Log ( ES )

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5

-2 -2
L/h=5
L/h=5
L/h=1000 L/h=1000
-2.5 -2.5
Log ( L k ) Log ( L k )

Fig. 12. The convergence DSQ element by using the s-norm test in a skew plate.

9
I.J. Maknun et al. Composite Structures 284 (2022) 115130

Table 6 Table 8
√̅̅̅̅̅̅̅̅̅̅̅̅ √̅̅̅̅̅̅̅̅̅̅̅̅
Convergence of natural frequency parameter ω = 10ωh ρc /Ec for SSSS Al/ The natural frequency ω = ωL2 /h ρc /Ec for SSSS Al/Al2O3 square plates.
Al2O3 square plates.
n
N n
Mode 0 0.5 1 5 10
0 0.5 1 2 4 10 ∞
1 DSQ 5.772 4.902 4.420 3.787 3.657
L/h = 5 [40] 5.676 4.821 4.347 3.722 3.592
4 2.139 1.828 1.652 1.498 1.415 1.340 1.088 2 DSQ 13.790 11.750 10.608 9.023 8.680
8 2.118 1.810 1.636 1.482 1.400 1.327 1.077 [40] 13.537 11.539 10.416 8.845 8.504
16 2.114 1.806 1.632 1.478 1.397 1.324 1.075 3 DSQ 13.790 11.750 10.608 9.023 8.680
32 2.113 1.804 1.631 1.478 1.396 1.323 1.074 [40] 13.537 11.539 10.416 8.845 8.504
64 2.112 1.804 1.630 1.477 1.396 1.323 1.074 4 DSQ 19.491 17.488 16.202 12.636 11.521
128 2.112 1.804 1.630 1.477 1.396 1.323 1.074 [40] 20.633 17.639 15.936 13.445 12.887
[2] 2.055 1.757 1.587 – 1.356 1.284 –
[39] 2.112 1.805 1.631 – 1.397 1.324 1.076
L/h = 10
4 0.585 0.497 0.448 0.407 0.388 0.371 0.297 Table 9
√̅̅̅̅̅̅̅̅̅̅̅̅
8 0.579 0.492 0.443 0.403 0.383 0.367 0.294 The natural frequency ω = ωL2 /h ρc /Ec for CCCC Al/Al2O3 square plates.
16 0.577 0.490 0.442 0.402 0.382 0.366 0.294
32 0.577 0.490 0.442 0.401 0.382 0.366 0.293 n
64 0.577 0.490 0.442 0.401 0.382 0.366 0.293 Mode 0 0.5 1 5 10
128 0.577 0.490 0.442 0.401 0.382 0.366 0.293
[2] 0.567 0.482 0.435 – 0.376 0.359 – 1 DSQ 9.838 8.407 7.604 6.448 6.180
[39] 0.577 0.490 0.442 – 0.382 0.366 0.293 [40] 9.633 8.239 7.453 6.306 6.038
L/h = 20 2 DSQ 18.792 16.134 14.620 12.282 11.713
4 0.150 0.127 0.115 0.104 0.100 0.095 0.076 [40] 18.313 15.742 14.265 11.946 11.377
8 0.149 0.126 0.113 0.103 0.098 0.094 0.076 3 DSQ 18.792 16.134 14.620 12.282 11.713
16 0.148 0.126 0.113 0.103 0.098 0.094 0.075 [40] 18.313 15.742 14.265 11.946 11.377
32 0.148 0.125 0.113 0.103 0.098 0.094 0.075 4 DSQ 26.336 22.674 20.565 17.179 16.338
64 0.148 0.125 0.113 0.103 0.098 0.094 0.075 [40] 25.499 21.979 19.935 16.597 15.763
128 0.148 1.2.541 0.113 0.103 0.098 0.094 0.075
[2] 0.146 0.124 0.112 – 0.097 0.0931 –
[39] 0.148 0.125 0.113 – 0.098 0.094 –

Table 7
√̅̅̅̅̅̅̅̅̅̅̅̅
The convergence of natural frequency parameter ω = 10ωh ρc /Ec for CCCC Al/
Al2O3 square plates.
N n

0 0.5 1 2 4 10 ∞

L/h = 5
4 3.245 2.814 2.564 2.326 2.164 2.002 1.650
8 3.219 2.789 2.538 2.299 2.140 1.984 1.637
16 3.213 2.782 2.531 2.292 2.133 1.979 1.634
32 3.211 2.781 2.529 2.290 2.132 1.978 1.633
64 3.211 2.780 2.528 2.289 2.132 1.977 1.633
128 3.211 2.780 2.528 2.289 2.131 1.977 1.633 Fig. 13. Full model of SSSS and CCCC square plates.
L/h = 10
4 0.994 0.851 0.771 0.702 0.664 0.627 0.506
8 0.987 0.843 0.763 0.693 0.656 0.620 0.502
16 0.985 0.841 0.761 0.691 0.653 0.619 0.501
32 0.984 0.841 0.761 0.690 0.653 0.618 0.501
64 0.984 0.841 0.760 0.690 0.653 0.618 0.501
128 0.984 0.841 0.760 0.690 0.653 0.618 0.501
L/h = 20
4 0.267 0.227 0.206 0.187 0.178 0.170 0.136
8 0.265 0.225 0.203 0.185 0.176 0.168 0.135
16 0.265 0.225 0.203 0.184 0.176 0.168 0.135
32 0.265 0.225 0.203 0.184 0.176 0.168 0.135
64 0.265 0.225 0.203 0.184 0.175 0.168 0.135 Fig. 14. Skew plate for free vibration analysis.
128 0.265 0.225 0.203 0.184 0.175 0.168 0.135
w = βx = 0 on side AB and CD u = v = w = βy = 0 on side AD and BC,
√̅̅̅̅̅̅̅̅̅̅̅̅ while for clamped supported, u = v = w = βx = βy = 0 on side ABCD.
The natural frequency is normalized by ω = 10ωh ρc /Ec . Table 6
presents the results for the SSSS case for different L/h ratios and power For both plate types, the DSQ element yields comparable results.
index n. The results from [2] and Hossemi et al. [39] are used as a
comparison. From Table 6, we can see that starting from 16 × 16 mesh, 6.2.2. Skew plate
the results of the DSQ element are close to those of Hossemi et al. [39]. Next, consider a fully clamped (CCCC) skew isotropic plate with
The CCCC case gives similar results (Table 7), where the results are different skew angles, as shown in Fig. 14. The boundary conditions are
stable starting from 16 × 16 mesh. u = v = w = βx = βy = 0 on the side of ABCD. Three different skew an­
Tables 8 and 9 presents a comparison of the full model (Fig. 13) of the gles are investigated. Table 10 compares the non-dimensional frequency
first four modes of DSQ element with the solution by kp-Ritz method computed from the present method using 32 × 32 mesh with that of Woo
[40] for L/h = 10. N × N = 32 × 32 (full model) as reference. For a hard et al. [41]. It is concluded that the DSQ element yields comparable re­
simply supported plate (Fig. 13), the boundary conditions are u = v = sults for L/h = 5 and 1000 investigated in this paper.

10
I.J. Maknun et al. Composite Structures 284 (2022) 115130

Table 10 30

Non - dimensional Frequency ω

( )√̅̅̅̅̅̅̅̅̅̅̅̅̅
The convergence of 5 modes of natural frequency ω = ωL2 /π2 ρh/Db for
25
CCCC isotropic skew plates. Mode 1
20
Mode 30o 45o 60o
15 n=0
DSQ [41] DSQ [41] DSQ [41]
n=0.2
L/h = 5 10 n=0.5
1 6.278 6.238 4.164 4.162 3.235 3.222 n=1
2 8.197 8.071 5.925 5.904 4.992 4.962
5
n=3
3 9.288 9.644 7.278 7.473 6.028 6.014 0
4 10.109 11.376 7.599 7.801 6.454 6.579 30 35 40 45 50 55 60
5 11.693 11.543 7.809 9.224 6.656 8.270
L/h = 1000 Angle skew (α)
1 12.007 12.534 6.611 6.658 4.667 4.671
2 17.467 18.344 10.729 10.826 8.279 8.269
40

Non - dimensional Frequency ω

3 22.732 25.02 14.950 15.230 10.645 10.690 35
4 28.569 32.022 15.736 16.128 12.113 12.119 Mode 2
30
5 29.283 32.931 19.865 20.572 16.792 16.839
25
n=0
20 n=0.2
15 n=0.5
Table 11
( )√̅̅̅̅̅̅̅̅̅̅̅̅ 10 n=1
The convergence of 4 modes of natural frequency ω = ωL2 /h ρc /Ec for Al/
5 n=3
ZrO2-2 CCCC skew plates (L/h = 10).
0
Mode n
30 40 50 60
0 0.2 0.5 1 3 Angle skew (α)
θ ¼ 30o
1 DSQ 27.771 26.485 25.205 24.019 22.540
50

Non - dimensional Frequency ω

[40] 27.502 26.246 24.973 23.789 22.315 45 Mode 3
2 DSQ 37.692 35.975 34.254 32.336 30.553 40
[40] 37.352 35.656 33.944 32.325 30.238 35
3 DSQ 47.237 45.102 42.953 40.919 38.267 30 n=0
[40] 46.061 43.991 41.890 39.887 37.258 25
n=0.2
4 DSQ 55.425 53.316 51.027 48.544 44.436 20
[40] 54.961 52.506 50.002 47.587 44.042 15 n=0.5
θ ¼ 45o 10 n=1
1 DSQ 16.718 15.914 15.125 14.415 13.596 5 n=3
[40] 16.635 15.835 15.047 14.338 13.523 0
2 DSQ 25.350 24.158 22.976 21.893 20.583 30 35 40 45 50 55 60
[40] 25.275 24.086 22.902 21.816 20.509
Angle skew (α)
3 DSQ 33.618 32.058 30.502 29.060 27.270
[40] 33.377 31.826 30.274 28.833 27.054
60
4 DSQ 35.161 33.537 31.914 30.403 28.513
Non - dimensional Frequency ω

[40] 35.182 33.554 31.921 30.399 28.505 50 Mode 4

θ ¼ 60o
1 DSQ 12.287 11.686 11.100 10.579 10.003 40
[40] 12.212 11.614 11.029 10.511 9.939 n=0
2 DSQ 20.440 19.463 18.500 17.628 16.612 30
[40] 20.349 19.375 18.413 17.542 16.532
n=0.2
20 n=0.5
3 DSQ 25.498 24.290 23.095 22.005 20.707
[40] 25.452 24.245 23.047 21.954 20.659 10 n=1
4 DSQ 28.404 27.065 25.737 24.521 23.058 n=3
[40] 28.226 26.893 25.568 24.354 22.902 0
30 40 50 60
Angle skew (α)
Table 11 and Fig. 15 show the influence of the angle and the power-
law index on the first four fundamental frequencies of Al/ZrO2-2 skew ( )√̅̅̅̅̅̅̅̅̅̅̅̅
Fig. 15. The convergence of 4 modes of natural frequency ω = ωL2 /h ρc /Ec
plates (L/h = 10). The results from Zhao et al. [40] are used as a com­ for the CCCC case.
parison. From the table, it can be concluded that the DSQ element with
32 × 32 mesh gives solutions close to the results obtained by the kp-Ritz
method. simply supported and clamped cases, are considered in this test. The
influence of the plate thickness is also considered. The critical buckling
Ncr L2
6.3. Buckling analysis load is normalized by Ncr = π2 Db . Table 12 shows the results of the
critical buckling load for SSSS and CCCC cases for L/h = 10 and L/h =
6.3.1. Square plate with uniaxial compression 100, which indicate that the DSQ element delivers accurate results.
For uniaxial buckling analysis, consider a square plate Al/ZrO2-1 as Moreover, it is found that the critical buckling load decreases with
shown in Fig. 16. As before, two different boundary conditions, viz.,

11
I.J. Maknun et al. Composite Structures 284 (2022) 115130

uniform temperature. Table 13 presents the critical buckling tempera­

ture by using mesh 32 × 32. As before, the power-law index significantly
influences the critical buckling temperature. The plate with a higher
metallic volume fraction has a lower critical buckling temperature when
compared to that of the plate with a higher ceramic volume fraction. It is
also found that the critical temperatures for a fully clamped plate are
greater than that of a simply supported plate. This is independent of the
power-law index and is attributed to the extra stiffness that the clamped
boundary offers when compared to the simply supported boundary
condition.

7. Concluding remarks
Fig. 16. Uniaxial Bucking analysis.
In this paper, the Discrete Shear Quadrilateral element formulation
has been extended to study the static bending, free vibration, and
Table 12 buckling analysis of the FGM plate. Starting from the reformulation of
The convergence of critical buckling load Ncr = Ncr L2 /π2 Db the DSQ element with 5 DOFs and the numerical results compared to the
N n
available references solution.
The influence of various parameters such as the boundary condi­
0 0.5 1 2 4 10 ∞
tions, plate thickness, and plate skews on the global response is sys­
SSSS L/h ¼ 10 tematically studied. From the numerical study, it can be inferred that the
4 3.788 3.657 3.508 3.405 3.454 3.622 3.788 proposed DSQ element suppresses shear locking and yields accurate
8 3.748 3.617 3.469 3.367 3.416 3.584 3.748
16 3.739 3.608 3.459 3.358 3.407 3.574 3.739
results for all the examples considered in this paper.
32 3.736 3.605 3.457 3.356 3.404 3.572 3.736 Further, it is noted that the formulation is easy to implement in any
64 3.735 3.605 3.457 3.355 3.404 3.571 3.735 existing finite element code and can be extended to polygonal elements.
128 3.735 3.604 3.456 3.355 3.404 3.571 3.735
[42] 3.741 – – – – – –
SSSS L/h ¼ 100 Declaration of Competing Interest
4 4.062 3.909 3.750 3.651 3.720 3.911 4.062
8 4.016 3.864 3.707 3.608 3.677 3.867 4.016
16 4.005 3.853 3.696 3.597 3.666 3.856 4.005
The authors declare that they have no known competing financial
32 4.002 3.851 3.693 3.595 3.663 3.853 4.002 interests or personal relationships that could have appeared to influence
64 4.001 3.850 3.692 3.594 3.662 3.852 4.001 the work reported in this paper.
128 4.001 3.850 3.692 3.594 3.662 3.852 4.001
[42] 4.000 – – – – – –
CCCC L/h ¼ 10
Acknowledgment
4 8.343 8.127 7.813 7.550 7.573 7.871 8.343
8 8.109 7.887 7.569 7.307 7.336 7.640 8.109 The financial support from the Indonesian Ministry of Research and
1 8.051 7.827 7.508 7.246 7.277 7.582 8.051 Technology (RISTEK-BRIN) through the PDUPT program (NKB-200/
32 8.036 7.812 7.493 7.231 7.262 7.568 8.036
UN2.RST/HKP.05.00/2021) is gratefully acknowledged.
64 8.033 7.808 7.490 7.228 7.258 7.564 8.033
128 8.032 7.807 7.489 7.227 7.257 7.564 8.032
CCCC L/h ¼ 100 References
4 10.439 10.068 9.679 9.435 9.603 10.070 10.439
8 10.152 9.774 9.379 9.132 9.302 9.776 10.152 [1] Nguyen-Xuan H, Tran LV, Thai CH, Nguyen-Thoi T. Analysis of functionally graded
16 10.081 9.702 9.306 9.058 9.229 9.704 10.081 plates by an efficient finite element method with node-based strain smoothing.
32 10.064 9.64 9.288 9.040 9.210 9.686 10.064 Thin-Walled Struct 2012;54:1–18. https://doi.org/10.1016/j.tws.2012.01.013.
64 10.059 9.679 9.283 9.035 9.206 9.682 10.059 [2] Nguyen-Xuan H, Tran LV, Nguyen-Thoi T, Vu-Do HC. Analysis of functionally
128 10.058 9.678 9.282 9.034 9.205 9.681 10.058 graded plates using an edge-based smoothed finite element method. Compos Struct
2011;93:3019–39. https://doi.org/10.1016/j.compstruct.2011.04.028.
[3] Chi SH, Chung YL. Mechanical behavior of functionally graded material plates
under transverse load-Part II: Numerical results. Int J Solids Struct 2006;43(13):
3675–91. https://doi.org/10.1016/j.ijsolstr.2005.04.011.
Table 13
[4] Chi SH, Chung YL. Mechanical behavior of functionally graded material plates
Critical buckling temperature for Al/Al2O3 FGM square plate under transverse load-Part I: Analysis. Int J Solids Struct 2006;43(13):3657–74.
BC L/h Tcr https://doi.org/10.1016/j.ijsolstr.2005.04.011.
[5] Talha M, Singh BN. Static response and free vibration analysis of FGM plates using
n=0 n = 0.5 n=1 n=2 n=5 higher order shear deformation theory. Appl Math Model 2010;34(12):3991–4011.
https://doi.org/10.1016/j.apm.2010.03.034.
SSSS 50 DSQ 70.108 39.147 31.992 28.282 29.086 [6] Reissner E. The effect of transverse shear deformation on the bending of elastic
[2] 70.700 39.486 32.272 28.529 29.328 plates. J Appl Mech Eng ASME 1945;12:A69–77. https://doi.org/10.1115/
100 DSQ 17.565 9.807 8.014 7.085 7.289 1.4009435.
[2] 17.719 9.895 8.087 7.149 7.352 [7] Mindlin RD. Influence of rotator inertia and shear on flexural motion of isotropic
CCCC 50 DSQ 184.921 103.304 84.439 74.638 76.695 elastic plates. J Appl Mech 1951;18:31–8. https://doi.org/10.1007/978-1-4613-
[2] 188.283 105.270 86.074 76.072 78.060 8865-4_29.
100 DSQ 46.530 25.981 21.234 18.773 19.307 [8] Hughes TJR, Cohen M, Haroun M. Reduced and selective integration techniques in
[2] 47.497 26.541 21.698 19.180 19.702 the finite element analysis of plates. Nucl Eng Des 1978;46(1):203–22. https://doi.
org/10.1016/0029-5493(78)90184-X.
[9] Malkus DS, Hughes TJR. Mixed finite element methods—reduced and selective
integration techniques: a unification of concepts. Comput Meth Appl Mech Eng
increasing power-law index, which can be attributed to a reduction in 1978;15(1):63–81. https://doi.org/10.1016/0045-7825(78)90005-1.
the ceramic constituent phase. [10] Zienkiewicz OC, Taylor RL, Too JM. Reduced integration technique in general
analysis of plates and shells. Int J NumericalMeth Eng 1971;3(2):275–90. https://
doi.org/10.1002/nme.1620030211.
6.3.2. Square plate with thermal buckling [11] Pugh EDL, Hinton E, Zienkiewicz OC. A study of quadrilateral plate bending
For the last example, consider Al/Al2O3 square plate subjected to a elements with ‘reduced’ integration. Int J Numer Meth Eng 1978;12(7):1059–79.
https://doi.org/10.1002/NME.1620120702.

12
I.J. Maknun et al. Composite Structures 284 (2022) 115130

[12] Hughes TJR, Tezduyar TE. Finite elements based upon Mindlin plate theory with Eng Sci Mech 2016;17(6):391–400. https://doi.org/10.1080/
particular reference to the four-node bilinear isoparametric element. ASME J Appl 15502287.2016.1231240.
Mech 1981;48(3):587–96. https://doi.org/10.1115/1.3157679. [29] Irpanni H, Katili I, Maknun IJ. Development DKMQ shell element with five degrees
[13] Macneal RH. Derivation of element stiffness matrices by assumed strain of freedom per nodal. Int J Mech Eng Robotics Res 2017;6:248–52. https://doi.
distributions. Nucl Eng Des 1982;70(1):3–12. https://doi.org/10.1016/0029-5493 org/10.18178/ijmerr.6.3.248-252.
(82)90262-X. [30] Katili I, Maknun IJ, Batoz J-L, Ibrahimbegovic A. Shear deformable shell element
[14] Bathe KJ, Dvorkin EN. A four-node plate bending element based on Mindlin- DKMQ24 for composite structures. Compos Struct 2018;202:182–200. https://doi.
Reissner plate theory and a mixed interpolation. Int J Numer Meth Eng 1985;21: org/10.1016/j.compstruct.2018.01.043.
367–83. https://doi.org/10.1002/nme.1620210213. [31] Katili I, Maknun IJ, Tjahjono E, Alisjahbana I. Error estimation for the DKMQ24
[15] Katili AM, Maknun IJ, Katili I. Theoretical equivalence and numerical performance shell element using various recovery methods. 2017;6:1060–1069. Int J Technol
of T3γs and MITC3 plate finite elements. Struct Eng Mech 2019;69(5):527–36. 2017;6:1060–9. https://doi.org/10.14716/ijtech.v8i6.699.
https://doi.org/10.12989/sem.2019.69.5.527. [32] Katili I, Maknun IJ, Katili AM, Bordas SPA, Natarajan S. A unified polygonal
[16] Batoz J-L, Lardeur P. A discrete shear triangular nine dof element for the analysis of locking-free thin/thick smoothed plate element. Compos Struct 2019;219:147–57.
thick to very thin plates. Int J Numer Meth Eng 1989;28:533–60. https://doi.org/ https://doi.org/10.1016/j.compstruct.2019.03.020.
10.1002/nme.1620280305. [33] Katili I, Maknun IJ, Batoz J-L, Katili AM. A comparative formulation of T3γs, DST,
[17] Lardeur P, Batoz J-L. Composite plate analysis using a new discrete shear triangular DKMT and MITC3+ triangular plate elements with new numerical results based on
plate bending element. Int J Numer Meth Eng 1989;27:343–60. https://doi.org/ s-norm tests. Eur J Mech A/Solids 2019;78:103826. https://doi.org/10.1016/j.
10.1002/nme.1620270209. euromechsol.2019.103826.
[18] Lardeur P. Développement et évaluation de deux nouveaux éléments finis de [34] Maknun IJ, Katili I, Ibrahimbegovic A, Katili AM. A new triangular shell element
plaques et coques composites avec influence du cisaillement transversal. Thèse for composites accounting for shear deformation. Compos Struct 2020;243:
UTC; 1990. 112214. https://doi.org/10.1016/j.compstruct.2020.112214.
[19] Batoz J-L, Dhatt G. Modélisation des structures par élément finis. Volume 2: [35] Maknun IJ, Katili I, Purnomo H. Development of DKMT element for error
Poutres et plaques, Hermes, Paris 1990. estimation in composite plate structures. Int J Technol 2015;6(5):780–9. https://
[20] Katili I. A new discrete Kirchhoff-Mindlin element based on Mindlin-Reissner plate doi.org/10.14716/ijtech.v6i5.1050.
theory and assumed shear strain fields- part I: An extended DKT element for thick- [36] Katili I, Maknun IJ, Batoz J-L, Katili AM. Asymptotic equivalence of DKMT and
plate bending analysis. Int J Numer Meth Eng 1993;36(11):1859–83. https://doi. MITC3 elements for thick composite plate. Compos Struct 2018;206:363–79.
org/10.1002/nme.1620361106. https://doi.org/10.1016/j.compstruct.2018.08.017.
[21] Katili I. A new discrete Kirchhoff-Mindlin element based on Mindlin-Reissner plate [37] Katili I, Batoz J-L, Maknun IJ, Lardeur PA. A comparative formulation of DKMQ,
theory and assumed shear strain fields- part II: An extended DKQ element for thick- DSQ and MITC4 quadrilateral plate elements with new numerical results based on
plate bending analysis. Int J Numer Meth Eng 1993;36(11):1885–908. https://doi. s-norm tests. Comput Struct 2018;204:48–64. https://doi.org/10.1016/j.
org/10.1002/nme.1620361107. compstruc.2018.04.001.
[22] Batoz -L J, Bathe K.-J., Ho L.-W.. A study of three-node triangular plate bending [38] Lee YY, Zhao X, Liew KM. Thermoelastic analysis of functionally graded plates
elements. Int J Numer Meth Eng 1980;15(12):1771–812. https://doi.org/10.1002/ using the element-free kp-Ritz method. Smart Mater Struct 2009;18(3):
nme.1620151205. 035007–789. https://doi.org/10.1088/0964-1726/18/3/035007.
[23] Batoz J-L, Ben Tahar M. Evaluation of a new thin plate quadrilateral element. Int J [39] Hosseini-Hashemi S, Taher HRD, Akhavan H, Omidi M. Free vibration of
Numer Meth Eng 1982;18:1665–78. https://doi.org/10.1002/nme.1620181106. functionally graded rectangular plates using first-order shear deformation plate
[24] Katili I. An improved incompatible DST element using free formulation approach. theory. Appl Math Model 2010;34(5):1276–91. https://doi.org/10.1016/j.
Struct Eng Mech 2021;79:67–81. https://doi.org/10.12989/sem.2021.79.1.067. apm.2009.08.008.
[25] Katili I. An improved incompatible DSQ element using free formulation approach. [40] Zhao X, Lee YY, Liew KM. Free vibration analysis of functionally graded plates
Struct Eng Mech 2021;78:665–79. https://doi.org/10.12989/sem.2021.78.6.665. using the element-free kp-Ritz method. J Sound Vib 2009;319(3–5):918–39.
[26] Katili I, Maknun IJ, Hamdouni A, Millet O. Application of DKMQ element for https://doi.org/10.1016/j.jsv.2008.06.025.
composite plate bending structures. Compos Struct 2015;132:166–74. https://doi. [41] Woo KS, Hong CH, Basu PK, Seo CG. Free vibration of skew Mindlin plates by p-
org/10.1016/j.compstruct.2015.04.051. version of FEM. J. Sound Vib. 2003;268(4):637–56. https://doi.org/10.1016/
[27] Katili I, Batoz JL, Maknun IJ, Hamdouni A, Millet O. The development of DKMQ S0022-460X(02)01536-5.
plate bending element for thick to thin shell analysis based on Naghdi/Reissner/ [42] Wong FT, Erwin, Richard A, Katili I. Development of the DKMQ element for
Mindlin Shell Theory. Finite Elem Anal Des 2014;100:12–27. https://doi.org/ buckling analysis of shear-deformable plate bending. Proc. Eng. 2017;171:805–12.
10.1016/j.finel.2015.02.005. https://doi.org/10.1016/j.proeng.2017.01.368.
[28] Maknun IJ, Katili I, Millet O, Hamdouni A. Application of DKMQ24 shell element
for twist of thin-walled beams: comparison with Vlasov theory. Int J Comput Meth

13