International Journal of Mechanical Sciences 130 (2017) 41–51

Contents lists available at ScienceDirect

International Journal of Mechanical Sciences

journal homepage: www.elsevier.com/locate/ijmecsci

Analytical approach for dynamic instability analysis of functionally graded

skew plate under periodic axial compression
Rajesh Kumar∗, Subhajit Mondal, Shyamal Guchhait, Rimen Jamatia
Department of Civil Engineering, Indian Institute of Technology Kharagpur, Kharagpur 721302, India

a r t i c l e i n f o a b s t r a c t

Keywords: Analytical studies on the dynamic instability analysis of a functionally graded (FG) skew plate subjected to
Dynamic instability uniform and linearly varying in-plane periodic loadings with four different types of boundary conditions are
Rayleigh–Ritz method presented. The total energy functional of the FG skew plate is formulated based on Reddy’s third order shear
BCOPs
deformation theory (TSDT) and this functional is mapped from the physical domain to computational domain
Gram–Schmidt process
using transformation rule. The boundary characteristics orthonormal polynomials (BCOPs) are generated for
TSDT
FG skew plate different boundary conditions using Gram–Schmidt process, which satisfy the essential boundary conditions of
skew plates in the computational domain. The energy functional is converted into a set of ordinary differen-
tial equations (Mathieu–Hill equations) using Rayleigh–Ritz method in conjunction with BCOPs. The solution
of Mathieu–Hill equations describes the dynamic instability behavior of skew plate. The instability regions are
traced using Bolotin method. The effect of skew angles, power-law distributions, span-to-thickness ratios, aspect
ratios, boundary conditions and static load factors on the instability region of FG skew plates are presented. The
result indicates that the width of instability region become narrow with the increase in skew angle. Moreover,
the time history response and corresponding phase plot in the unstable and stable region is studied to identify the
instability behavior such as existence of beats, bounded and unbounded response, and effect of forcing amplitude
and its frequency on the response.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction
Over the decades, functionally graded materials (FGMs) are increas-
ingly used in the defence, aerospace, nuclear reactors and other high-
performance structures [1,2]. The FGM has several advantages over FRP
composite structures based on material characteristics and thus, they
are used more frequently in high-performance structures. The FGM is
made up of ceramic and metal having a smooth variance of material
properties throughout the thickness. The uses of functional graded ma-
terial in aerospace/nuclear industries are well known. Therefore, the
static and dynamic instability of FG skew plates is essential for design
aspects and its performance. Extensive research has been done on the
static [3–6] and dynamic instability [7–11] for the structural element
made of isotropic and composite type of materials. Recently, Jha et al.
[12] provided a critical review on the studies of FG plates.
The FG skew plate starts deflecting outward under in-plane com-
pressive loading because of the extensional-bending coupling like un-
symmetric cross-ply laminated skew plates. Ng et al. [13,14] studied
the dynamic instability behavior of FG cylindrical thin shells under
uniform periodic axial loading. Shen [15] analyzed the bending per-
formance of a simply supported FG rectangular plate under transverse
load and pointed out that the FG plate behaves like un-symmetric lam-
inated plate. Since the neutral axis of the FG plate do not pass through
the middle plane the plate, it may not possess the bifurcation point,
the shift of the neutral axis needs to be considered in the analysis of
the FGM plate [16]. The effect of various factors such as gradient in-
dex, temperature, aerodynamic loading on the dynamic instability of a
FG plate was studied by Lanhe et al. [17] using differential quadrature
method.
The nonlinear free flexural vibration of functionally graded skew and
rectangular plates was performed by Sundararajan et al. [18] under ther-
mal loadings. Ganapathi and Prakash [19] discussed the influence of the
parameters (such as aspect ratio, volume index, skew angle etc.) on the
critical buckling load of simply supported FGM skew plate. Pradyumna
and Bandyopadhyay [20] studied the dynamic instability of FGM shell
under in-plane periodic load and temperature. Torki et al. [21] inves-
tigated the flutter behavior of a FGM cylindrical shell subjected to dis-
tributed axial forces. Asnafi and Abedi [22] studied the dynamic stability
of a FG plates subjected to random lateral loads. Sofiyer and Kuruoglu
[23] studied the instability behavior of FG shell structures and sandwich
cylindrical shell with FG core under uniform periodic axial loads using
the Galerkin’s method.

∗
Corresponding author.
E-mail address: rajeshiitkgp@iitkgp.ac.in (R. Kumar).

http://dx.doi.org/10.1016/j.ijmecsci.2017.05.050
Received 3 January 2017; Received in revised form 4 May 2017; Accepted 20 May 2017
Available online 7 June 2017
0020-7403/© 2017 Elsevier Ltd. All rights reserved.
R. Kumar et al.
Fig. 1. (a) FG skew plate geometry and uniform periodic axial compression. (b) The position of mid plane and neutral plane for FG skew plate. (c) Different types of linearly varying
in-plane edge loadings.
The literature on the dynamic instability of FG skew plate is very lim-
ited and required details exploration of design aspects and better perfor-
mance. In the present study, analytical investigation carried out on the
dynamic stability of FG skew plates under uniform and linearly in-plane
varying loads using Rayleigh–Ritz Method in conjunction with BCOPs.
The total energy functional of the FG skew plate is formulated based
on Reddy’s TSDT and transformed from physical domain to computa-
tional domain using transformation rule. The total energy functional (in
computational domain) is then reduced to ordinary differential equa-
tions (Mathieu–Hill equations) using Rayleigh–Ritz method in conjunc-
tion with BCOPs. The BCOPs are generated by orthogonalization to the
boundary characteristic polynomials (BCPs) via Gram–Schmidt process.
The BCPs are then taken as a product of two-dimensional independent
set of polynomials and the basis functions. The basis functions are devel-
oped by taking the product of boundary equation of skew plate, which
satisfy the essential boundary conditions. The boundaries of instability
regions are plotted using Bolotin method [25]. The effect of different
skew angles, various types of linearly varying loadings, static load fac-
tors, boundary conditions, span-to-thickness ratios and aspect ratio on
the instability regions are studied.
2. Mathematical formulation
The functionally graded (FG) skew plate is made by mixing of ce-
ramic and metal materials in which the bottom surface is metal and the
42
International Journal of Mechanical Sciences 130 (2017) 41–51
top surface is ceramic. The FG skew plate of length “a”, width “b” and
thickness “h” under uniform loading shown in Fig. 1(a).
The material distribution along the thickness direction of FG plate
is not symmetric about the middle plane. Therefore, the neutral plane
may not coincide with the middle plane. Fig. 1b shows the position of the
middle plane and the neutral plane of the skew plate at any cross section.
In the present case the reference plane taken as a neutral plane for the
analysis of functionally graded skew plate using elementary theory to
avoid the coupling between the stretching and bending deformation. In
order to determine the position of the reference plane from the middle
plane of the functionally graded skew plate for which the first moment
of modulus of elasticity with respect to reference plane is equated to
zero [26,27] as follows,
ℎ∕2
𝐸(𝑧)(𝑧 − 𝐶)𝑑𝑧 = 0 (1)
∫−ℎ∕2
the position of reference plane (neutral plane) from middle plane may
be defined as
ℎ∕2
∫−ℎ∕2 𝐸(𝑧)𝑧𝑑𝑧
𝐶= (2)
ℎ∕2
∫−ℎ∕2 𝐸(𝑧)𝑑𝑧
Therefore, distance (z) of the top surface and the bottom surface
from the reference plane are 𝑧𝑡 (= ℎ2 − 𝐶) and 𝑧𝑏 (= − ℎ2 − 𝐶), respec-
tively. The effective material properties (modulus of elasticity and
R. Kumar et al. International Journal of Mechanical Sciences 130 (2017) 41–51

𝑎
density) according to the rule of mixture can be written as ⎛⎛𝑁𝑥 ⎞ ⎛𝑀𝑥 ⎞ ⎛𝑀𝑥 ⎞⎞ ℎ∕2−𝐶 ⎛𝜎𝑥 ⎞
⎜⎜𝑁 ⎟, ⎜𝑀 ⎟, ⎜𝑀 𝑎 ⎟⎟ = ⎜𝜎 ⎟(1, 𝑧, 𝑓 (𝑧))𝑑𝑧 (11)
𝐸(𝑧) = 𝐸𝑚 𝑉𝑚 (𝑧) + 𝐸𝑐 𝑉𝑐 (𝑧) ⎜⎜ 𝑦 ⎟ ⎜ 𝑦 ⎟ ⎜ 𝑎 ⎟⎟ ∫−ℎ∕2−𝐶 ⎜ 𝑦 ⎟
𝑦
(3) ⎝⎝𝑁𝑥𝑦 ⎠ ⎝𝑀𝑥𝑦 ⎠ ⎝𝑀𝑥𝑦 ⎠⎠ ⎝𝜏𝑥𝑦 ⎠
𝜌(𝑧) = 𝜌𝑚 𝑉𝑚 (𝑧) + 𝜌𝑐 𝑉𝑐 (𝑧)
ℎ∕2−𝐶
where, subscripts m and c represent the metal and ceramic constituents, (𝑄𝑎𝑥 , 𝑄𝑎𝑦 ) = (𝜏𝑥𝑧 , 𝜏𝑦𝑧 )𝑓 ′ (𝑧)𝑑𝑧 (12)
∫−ℎ∕2−𝐶
respectively. Here, the volume fraction of the metal and ceramic phase
is described by the following power-law distribution, where Nx , Ny , Nxy and Mx , My , Mxy are the force and moment resul-
( ) tants; 𝑀𝑥𝑎 , 𝑀𝑦𝑎 , 𝑀𝑥𝑦
𝑎 are the additional moment resultants due to ad-
𝑧+𝐶 1 𝑛
𝑉𝑐 = + , 𝑉𝑚 = 1−𝑉𝑐 (4) ditional changes of curvature 𝜙𝑜1,𝑥 𝜙𝑜2,𝑦 𝜙𝑜2,𝑥 + 𝜙𝑜1,𝑦 ; 𝑄𝑎𝑥 , 𝑄𝑎𝑦 are transverse
ℎ 2
shear force resultants. The force, moment, additional moment and shear
The volume fraction varies through the thickness based on the power
resultants are related to strains by [24]
index n. In the present investigation, the Poisson’s ratio (𝜈) is assumed
to be constant across the thickness. ⎧𝑁 ⎫ ⎡𝐴 𝐴12 𝐴16 ⎤⎧𝜀𝑜𝑥 ⎫ ⎡𝐵11 𝐵12 𝐵16 ⎤⎧−𝑤𝑜,𝑥𝑥 ⎫
The linearly varying in-plane edge loading may be expressed as ⎪ 𝑥 ⎪ ⎢ 11 ⎪ ⎪ ⎪ ⎪
⎨𝑁𝑦 ⎬ = ⎢𝐴12 𝐴22 𝐴26 ⎥⎨𝜀𝑜𝑦 ⎬ + ⎢𝐵12 𝐵22 𝐵26 ⎥⎨−𝑤𝑜,𝑦𝑦 ⎬
𝑁𝑥𝑥 = 𝑁0 (1 − 𝜆( 𝜂𝑏 )). In this expressions four different values of𝜆 yield ⎪𝑁𝑥𝑦 ⎪ ⎣𝐴16 ⎥⎪ 𝑜 ⎪ ⎢ ⎥⎪
⎩ ⎭ 𝐴26 𝐴66 ⎦⎩𝛾𝑥𝑦 ⎭ ⎣𝐵16 𝐵26 𝐵66 ⎦⎩−2𝑤𝑜𝑥𝑦 ⎪
⎭
different in-plane loadings distribution like a uniform (𝜆 = 0), trape-
zoidal (𝜆 = 0.5), triangular (𝜆 = 1), partial tension (𝜆 = 1.5) and in-plane 𝑜
⎡𝐶11 𝐶12 𝐶16 ⎤⎧ 𝜙1,𝑥 ⎫
bending (𝜆 = 2). Different types of linearly varying in-plane edge load- ⎪ ⎪
+ ⎢𝐶12 𝐶22 ⎥
𝐶26 ⎨ 𝜙2,𝑦𝑜
⎬
ings are shown in Fig. 1c. ⎢ ⎥⎪ 𝑜
⎣𝐶16 𝐶26 𝐶66 ⎦⎩𝜙1,𝑦 + 𝜙𝑜2,𝑥 ⎪
⎭
The total energy of FG skew plate is formulated based on the third-
order shear deformation theory [24]. In this theory no shear correction ⎧𝑀 ⎫ ⎡𝐵 𝐵12 𝐵16 ⎤⎧𝜀𝑜𝑥 ⎫ ⎡𝐷11 𝐷12 𝐷16 ⎤⎧−𝑤𝑜,𝑥𝑥 ⎫
⎪ 𝑥 ⎪ ⎢ 11 ⎪ ⎪ ⎪ ⎪
factors is required because of the parabolic distribution of the transverse
⎨𝑀𝑦 ⎬ = ⎢𝐵12 𝐵22 𝐵26 ⎥⎨𝜀𝑜𝑦 ⎬ + ⎢𝐷12 𝐷22 𝐷26 ⎥⎨−𝑤𝑜,𝑦𝑦 ⎬
shear stress and satisfy the zeros stress conditions at the bottom and top ⎪𝑀𝑥𝑦 ⎪ ⎣𝐵16 ⎥⎪ 𝑜 ⎪ ⎢ ⎥⎪
⎩ ⎭ 𝐵26 𝐵66 ⎦⎩𝛾𝑥𝑦 ⎭ ⎣𝐷16 𝐷26 𝐷66 ⎦⎩−2𝑤𝑥𝑦 ⎪
𝑜
⎭
surface of plate. The displacement components u, v and w in the x, y
𝑜
and z directions, respectively, at a distance ‘z’ away from the reference ⎡𝐸11 𝐸12 𝐸16 ⎤⎧ 𝜙1,𝑥 ⎫
⎪ ⎪
plane (i.e. mid plane) can be defined as [24], + ⎢𝐸12 𝐸22 𝐸26 ⎥⎨ 𝜙𝑜2,𝑦 ⎬
⎢ ⎥⎪ 𝑜
𝑢 = 𝑢𝑜 + 𝑧𝜑1 + 𝑧3 (4∕3ℎ2 )[−𝜑1 − 𝑤𝑜,𝑥 ] ⎣𝐸16 𝐸26 𝐸66 ⎦⎩𝜙1,𝑦 + 𝜙2,𝑥 ⎪
𝑜
⎭
𝑣 = 𝑣𝑜 + 𝑧𝜑2 + 𝑧3 (4∕3ℎ2 )[−𝜑2 − 𝑤𝑜,𝑦 ] ⎧𝑀 𝑎 ⎫ ⎡𝐶 𝐶16 ⎤⎧𝜀𝑜𝑥 ⎫ ⎡𝐸11 𝐸16 ⎤⎧−𝑤𝑜,𝑥𝑥 ⎫
(5)
𝐶12 𝐸12
𝑤 = 𝑤𝑜 ⎪ 𝑥𝑎 ⎪ ⎢ 11 ⎪ ⎪ ⎪ ⎪
⎨𝑀𝑦 ⎬ = ⎢𝐶12 𝐶22 𝐶26 ⎥⎨𝜀𝑜𝑦 ⎬ + ⎢𝐸12 𝐸22 𝐸26 ⎥⎨−𝑤𝑜,𝑦𝑦 ⎬
⎪𝑀 𝑎 ⎪ ⎣𝐶16 ⎥⎪ 𝑜 ⎪ ⎢ ⎥⎪
⎩ 𝑥𝑦 ⎭ 𝐶26 𝐶66 ⎦⎩𝛾𝑥𝑦 ⎭ ⎣𝐸16 𝐸26 𝐸66 ⎦⎩−2𝑤𝑥𝑦 ⎪
𝑜
⎭
Here u0 , v0 and w0 are displacement component on a reference plane.
𝜙1 and 𝜙2 are rotations of the cross sections initially perpendicular to 𝑜
⎡𝐹11 𝐹12 𝐹16 ⎤⎧ 𝜙1,𝑥 ⎫
the x and y axes, respectively. Upon the rearrangement of displacement ⎢ ⎥ ⎪ 𝑜 ⎪
+ 𝐹12 𝐹22 𝐹26 ⎨ 𝜙2,𝑦 ⎬
components [28], ⎢ ⎥⎪ 𝑜
⎣𝐹16 𝐹26 𝐹66 ⎦⎩𝜙1,𝑦 + 𝜙𝑜2,𝑥 ⎪
⎭
𝑢 = 𝑢𝑜 − 𝑧𝑤𝑜,𝑥 + 𝑓 (𝑧)𝜙1 𝑜 { 𝑎} [ ]{ 𝑜 }
𝑄𝑦 𝐻44 𝐻45 𝜙2
𝑣 = 𝑣𝑜 − 𝑧𝑤𝑜,𝑦 + 𝑓 (𝑧)𝜙2 𝑜 (6) = (13a–d)
𝑄𝑎𝑥 𝐻45 𝐻55 𝜙𝑜1
𝑤 = 𝑤𝑜

where, 2.2. Strain energy components

𝜙01 = 𝜑1 + 𝑤0,𝑥 ; 𝜙02 = 𝜑2 + 𝑤0,𝑦 Now, the various components of strain energy of FG skew plate are
(7)
𝑓 (𝑧) = 𝑧[1 − (4∕3)(𝑧∕ℎ)2 ] expressed as,
𝑇
The strain–displacement relations at a distance ‘z’ away from the ⎛⎧𝑁 ⎫ ⎧𝜀𝑜 ⎫⎞
𝑎 𝑏
reference plane of a plate can be written as, 1 𝑚 ⎜⎪ 𝑥 ⎪ ⎪ 𝑥𝑜 ⎪⎟
Membrane energy (𝑈 ) =
2 ∫0 ∫0 ⎜⎨𝑁𝑦 ⎬ ⎨𝜀𝑦 ⎬⎟𝑑 𝑥𝑑 𝑦 (14)
⎜⎪𝑁𝑥𝑦 ⎪ ⎪𝛾 𝑜 ⎪⎟
𝜀𝑥 = 𝜀𝑜𝑥 − 𝑧𝑤𝑜,𝑥𝑥 + 𝑓 (𝑧)𝜙1 𝑜,𝑥 ⎝⎩ ⎭ ⎩ 𝑥𝑦 ⎭⎠
𝜀𝑦 = 𝜀𝑜𝑦 − 𝑧𝑤𝑜,𝑦𝑦 + 𝑓 (𝑧)𝜙2 𝑜,𝑦 After substituting the stress resultants (Nx , Ny , Nxy ) expression
𝛾𝑥𝑦 = 𝜀𝑜𝑥𝑦 − 2𝑧𝑤𝑜,𝑥𝑦 + 𝑓 (𝑧)𝜙1 𝑜,𝑦 + 𝑓 (𝑧)𝜙2 𝑜,𝑥 (8) (Eq. (13a)) in Eq. (14) then membrane energy can be expressed as,
𝛾𝑥𝑧 = 𝑢,𝑧 + 𝑤,𝑥 = 𝑓 ′ (𝑧)𝜙1 𝑜 𝑇
⎛ ⎧𝜀 𝑜 ⎫ ⎡𝐴11 𝐴12 𝐴16 ⎤⎧𝜀𝑜𝑥 ⎫
𝛾𝑦𝑧 = 𝑣,𝑧 + 𝑤,𝑦 = 𝑓 ′ (𝑧)𝜙2 𝑜 1
𝑎 𝑏
⎜⎪ 𝑥𝑜 ⎪ ⎪ ⎪
𝑈 𝑚
= ⎜ ⎨𝜀 𝑦 ⎬ ⎢𝐴 𝐴22 𝐴26 ⎥⎨𝜀𝑜𝑦 ⎬
and 𝜀𝑜𝑥 , 𝜀𝑜𝑦 and 𝛾𝑥𝑦
𝑜 are reference surface strains and are defined as, 2 ∫0 ∫0 ⎜ ⎪𝛾 𝑜 ⎪ ⎢ 12 ⎥⎪ 𝑜 ⎪
⎝⎩ 𝑥𝑦 ⎭ ⎣𝐴16 𝐴26 𝐴66 ⎦⎩𝛾𝑥𝑦 ⎭
𝜀𝑜𝑦 = 𝑣𝑜,𝑦 , 𝜀𝑜𝑥 = 𝑢𝑜,𝑥 , 𝛾𝑥𝑦
𝑜
= 𝑢𝑜,𝑦 + 𝑣𝑜,𝑥 (9) 𝑇
⎧𝜀𝑜 ⎫ ⎡𝐵11 𝐵12 𝐵16 ⎤⎧−𝑤𝑜,𝑥𝑥 ⎫
⎪ 𝑥⎪ ⎢𝐵 ⎪ ⎪
+ ⎨𝜀𝑜𝑦 ⎬ 𝐵22 𝐵26 ⎥⎨−𝑤𝑜,𝑦𝑦 ⎬
2.1. Contitutive relation ⎪𝛾 𝑜 ⎪ ⎢ 12 ⎥⎪
⎩ 𝑥𝑦 ⎭ ⎣𝐵16 𝐵26 𝐵66 ⎦⎩−2𝑤𝑥𝑦 ⎪
𝑜
⎭
A plane stress constitutive law is used in the present study. The stress 𝑇
⎧𝜀𝑜 ⎫ 𝑜
⎡𝐶11 𝐶12 𝐶16 ⎤⎧ 𝜙1,𝑥 ⎫⎞
and strain are related as, ⎪ 𝑥⎪ ⎢𝐶 ⎪ ⎪⎟
+ ⎨𝜀𝑜𝑦 ⎬ 𝐶22 𝐶26 ⎥⎨ 𝜙𝑜2,𝑦 ⎬⎟𝑑 𝑥𝑑 𝑦 (15)
⎪𝛾 𝑜 ⎪ ⎢ 12 ⎥⎪ 𝑜
{𝜎} = [𝑄]{𝜀} (10)
⎩ 𝑥𝑦 ⎭ ⎣𝐶16 𝐶26 𝐶66 ⎦⎩𝜙1,𝑦 + 𝜙2,𝑥 ⎪
𝑜
⎭⎠
⎟

where, {𝜎}T = {𝜎 x 𝜎 y 𝜏 xz 𝜏 yz 𝜏 xy }and {𝜀}T = {𝜀x 𝜀y 𝛾 xz 𝛾 yz 𝛾 xy } 𝑇

⎛⎧𝑀 ⎫ ⎧−𝑤𝑜 ⎫⎞
Here, {𝜎}T and {𝜀}T are the cartesian components of stress tensor 1
𝑎
𝑏
𝑏
⎜⎪ 𝑥 ⎪ ⎪ ,𝑥𝑥 ⎪⎟
Bending energy (𝑈 ) = 𝑀𝑦 ⎬ ⎨−𝑤𝑜,𝑦𝑦 ⎬⎟𝑑 𝑥𝑑 𝑦
2 ∫0 ∫0 ⎜⎜⎨
and corresponding strain at a point, respectively. The force and moment (16)
⎪𝑀𝑥𝑦 ⎪ ⎪−𝑤𝑜 ⎪⎟
resultants of the FG skew plate can be defined as ⎝⎩ ⎭ ⎩ ,𝑥𝑦 ⎭⎠

43
R. Kumar et al. International Journal of Mechanical Sciences 130 (2017) 41–51

After substituting the moment resultants (Mx , My , Mxy ) expression and

(Eq. (13b)) in Eq. (16) then bending energy can be expressed as,
⎡1 − 𝜈 ⎤
𝐸(𝑧) ⎢ 2 ⎥0
𝑇 𝑄= (𝑖, 𝑗 = 4, 5)
⎛⎧−𝑤𝑜 ⎫ ⎢ 1−𝜈 ⎥
(24)
𝑎 𝑏 ⎡𝐵11 𝐵12 𝐵16 ⎤⎧𝜀𝑜𝑥 ⎫ 1 − 𝜈2 ⎢ 0 ⎥
1 ⎜⎪ ,𝑥𝑥 ⎪ ⎢𝐵 ⎪ ⎪
𝑈 = 𝑏 −𝑤𝑜,𝑦𝑦 ⎬ 𝐵22 𝐵26 ⎥⎨𝜀𝑜𝑦 ⎬ ⎣ 2 ⎦
2 ∫0 ∫0 ⎜⎜⎨
⎪ 𝑜 ⎪ ⎢ 12 ⎥⎪ 𝑜 ⎪
⎝⎩−2𝑤𝑥𝑦 ⎭ ⎣𝐵16 𝐵26 𝐵66 ⎦⎩𝛾𝑥𝑦 ⎭
The total strain energy can be written as,
𝑇
⎧− 𝑤 𝑜 ⎫ ⎡𝐷11 𝐷12 𝐷16 ⎤⎧−𝑤𝑜,𝑥𝑥 ⎫
⎪ ,𝑥𝑥 ⎪ ⎪ ⎪ 𝑈 = 𝑈 𝑚 + 𝑈 𝑏 + 𝑈 𝑎𝑏 +𝑈 𝑠 (25)
+ ⎨−𝑤𝑜,𝑦𝑦 ⎬ ⎢𝐷 𝐷22 𝐷26 ⎥⎨−𝑤𝑜,𝑦𝑦 ⎬
⎪−2𝑤𝑜 ⎪ ⎢ 12 ⎥⎪
⎩ 𝑥𝑦 ⎭ ⎣𝐷16 𝐷26 𝐷66 ⎦⎩−2𝑤𝑜𝑥𝑦 ⎪
⎭
The potential energy in term of in-plane stress resultant are expressed
as,
𝑇
⎧− 𝑤 𝑜 ⎫ 𝑜
⎡𝐸11 𝐸12 𝐸16 ⎤⎧ 𝜙1,𝑥 ⎫⎞
⎪ ,𝑥𝑥 ⎪ ⎢𝐸 ⎪ ⎪⎟ 𝑎 𝑏
+ ⎨−𝑤𝑜,𝑦𝑦 ⎬ 𝐸22 𝐸26 ⎥⎨ 𝜙𝑜2,𝑦 ⎬⎟𝑑 𝑥𝑑 𝑦 (17) 𝑉 =−
1
⎪−2𝑤𝑜 ⎪ ⎢ 12 ⎥⎪ 𝑜 2 ∫0 ∫0
⎩ 𝑥𝑦 ⎭ ⎣𝐸16 𝐸26 𝐸66 ⎦⎩𝜙1,𝑦 + 𝜙2,𝑥 ⎪
𝑜
⎭⎠
⎟
𝑇
⎛⎧ 𝜕 𝑢 0 ⎫ ⎧ 𝜕 𝑢 0 ⎫⎞
Additional bending energy due to additional change in curvature ⎜⎪ ⎪ ⎪ ⎪⎟
⎜⎪ 𝜕𝑥0 ⎪ ⎪ 𝜕𝑥0 ⎪⎟
(Uab ) = ⎜⎪ 𝜕 𝑢 ⎪ ⎪ 𝜕 𝑢 ⎪⎟
⎜⎪ 𝜕𝑦 ⎪ ⎡𝑛𝑥𝑥 𝑛𝑥𝑦 0 0 0 0 ⎤⎪ 𝜕𝑦 ⎪⎟
𝑇 ⎜⎪ 0 ⎪ ⎪ ⎪⎟
⎛⎧𝑀𝑥𝑎 ⎫ ⎧ 𝜙𝑜1,𝑥 ⎫⎞ ⎢𝑛𝑥𝑦 𝑛𝑦𝑦 0 ⎥⎪ 𝜕 𝑣0 ⎪⎟
1
𝑎 𝑏
⎜⎪ 𝑎 ⎪ ⎪ ⎪⎟ ⎜⎪ 𝜕 𝑣 ⎪ ⎢
0 0 0
⎥⎪
⎜⎨𝑀𝑦 ⎬ ⎨ 𝜙𝑜2,𝑦 ⎬⎟𝑑 𝑥𝑑 𝑦 (18) ⎜⎪ 𝜕𝑥 ⎪ ⎢0 0 𝑛𝑥𝑥 𝑛𝑥𝑦 0 0 ⎥ 𝜕𝑥 ⎪⎟
2 ∫0 ∫0 ⎜⎪𝑀 𝑎 ⎪ ⎪𝜙𝑜 + 𝜙𝑜 ⎪⎟
× ⎜⎨ 0 ⎬
⎢0 𝑛𝑥𝑦 𝑛𝑦𝑦 0 ⎥⎨ 0 ⎬⎟𝑑 𝑥𝑑 𝑦
⎝⎩ 𝑥𝑦 ⎭ ⎩ 1,𝑦 2,𝑥 ⎭⎠ ⎜⎪ 𝜕 𝑣 ⎪ ⎢0
0 0
⎥⎪ 𝜕 𝑣 ⎪⎟
⎜⎪ 𝜕𝑦 ⎪ ⎢ 0 0 0 𝑛𝑥𝑥 𝑛𝑥𝑦⎥⎪ 𝜕𝑦 ⎪⎟
After substituting the moment resultants (𝑀𝑥𝑎 , 𝑀𝑦𝑎 , 𝑀𝑥𝑦
𝑎 ) expression ⎜⎪ 𝜕 𝑤 0 ⎪ ⎣0 0 0 0 𝑛𝑥𝑦 𝑛𝑦𝑦⎦⎪ 𝜕 𝑤0 ⎪⎟
⎜⎪ ⎪ ⎪ ⎪⎟
(Eq. (13c)) in Eq. (18) then additional bending energy can be expressed ⎜⎪ 𝜕𝑥 ⎪ ⎪ 𝜕𝑥 ⎪⎟
as, ⎜⎪ 𝜕 𝑤 ⎪0 ⎪ 𝜕 𝑤 ⎪⎟
0
⎜⎪ ⎪ ⎪ ⎪⎟
⎝⎩ 𝜕𝑦 ⎭ ⎩ 𝜕𝑦 ⎭⎠
𝑇
⎛⎧ 𝜙𝑜 ⎫ ⎡𝐶11 𝐶12 𝐶16 ⎤⎧𝜀𝑜𝑥 ⎫ (26)
𝑎 𝑏 1,𝑥
1 ⎜⎪ ⎪ ⎢𝐶 ⎪ ⎪
𝑈 𝑎𝑏 𝜙𝑜
𝐶22 𝐶26 ⎥⎨𝜀𝑜𝑦 ⎬
2 ∫0 ∫0 ⎜⎜⎨ ⎬
= 2,𝑦
⎪ 𝑜 𝑜 ⎪ ⎢ 12 ⎥⎪ 𝑜 ⎪ where, nxx and nyy are the in-plane stresses resultant along the x and y
⎝⎩𝜙1,𝑦 + 𝜙2,𝑥 ⎭ ⎣𝐶16 𝐶26 𝐶66 ⎦⎩𝛾𝑥𝑦 ⎭
directions and nxy is the shear stress resultant in x–y plane. The kinetic
𝑇
⎧ 𝜙𝑜 ⎫ ⎡𝐸11 𝐸12 𝐸16 ⎤⎧−𝑤𝑜,𝑥𝑥 ⎫ energy can be expressed as,
1,𝑥
⎪ ⎪ ⎢𝐸 ⎪ ⎪
+ ⎨ 𝜙𝑜2,𝑦 ⎬ 𝐸22 𝐸26 ⎥⎨−𝑤𝑜,𝑦𝑦 ⎬
⎪𝜙 + 𝜙 ⎪ ⎢ 12 ⎥⎪ ⎛⎧ 𝜕 𝑢0 ⎫ ⎧ 𝜕 𝑢0 ⎫⎞
𝑜
⎩ 1,𝑦
𝑜
2,𝑥 ⎭
⎣𝐸16 𝐸26 𝐸66 ⎦⎩−2𝑤𝑥𝑦 ⎪
𝑜
⎭ ⎜⎪ ⎪ ⎪ ⎪⎟
⎜⎪ 𝜕𝑡 ⎪⎡0 𝜌1 0 0 0 ⎤⎪ 𝜕𝑡0 ⎪⎟
𝑇 ⎜⎪ 𝜕 𝑣0 ⎪⎢ ⎪ ⎪⎟
⎧ 𝜙𝑜 ⎫ 𝑜
𝐹16 ⎤⎧ 𝜙1,𝑥 ⎫⎞ ⎥⎪ 𝜕 𝑣
⎡𝐹11 𝐹12 ⎜⎪ ⎪ ⎥⎪ 𝜕𝑡 ⎪ ⎟
1,𝑥
⎪ ⎪ ⎪ ⎪⎟
+ ⎨ 𝜙𝑜2,𝑦 ⎬ ⎢𝐹 𝐹22 ⎥
𝐹26 ⎨ 𝜙𝑜2,𝑦 ⎬⎟𝑑 𝑥𝑑 𝑦 (19) ⎜⎪ 𝜕𝑡 0 ⎪⎢⎢0 0 𝜌1 0 0 ⎥ 𝜕 𝑤0 ⎪⎟
⎢ 12 ⎥⎪ 𝑜 𝑎 𝑏 𝜕 𝑤
⎪𝜙𝑜 + 𝜙𝑜 ⎪ ⎜⎪ ⎪⎢ ⎥⎪ ⎪⎟
⎣𝐹16 𝐹26 𝐹66 ⎦⎩𝜙1,𝑦 + 𝜙2,𝑥 ⎪
𝑜 ⎟ 𝑇𝐸 =
1
⎩ 1,𝑦 2,𝑥 ⎭ ⎭⎠ 2 ∫0 ∫0 ⎜⎜⎨ 𝜕𝑡 ⎬ ⎥⎨ 𝜕𝑡 ⎬⎟𝑑 𝑥𝑑 𝑦 (27)
⎪ 0 ⎪⎢⎢0 0 0 𝜌3 0 ⎥ 𝜕𝜙0 ⎪⎟
⎪
⎜⎪ 𝜕𝜙2 ⎪⎢ ⎥⎪ 2 ⎪⎟
⎜⎪ 𝜕𝑡 ⎪⎢ ⎥⎪ 𝜕𝑡 ⎪⎟
⎜⎪ 0 ⎪⎣0 0 0 0 𝜌3 ⎦⎪ 0 ⎪⎟
Shear energy due to shear deformation (𝑈 𝑠 ) ⎜⎪ 𝜕𝜙1 ⎪ ⎪ 𝜕𝜙1 ⎪⎟
{ }𝑇 { } ⎜⎪ 𝜕𝑡 ⎪ ⎪ 𝜕𝑡 ⎪⎟
𝑎 𝑏 ⎛ 𝑄𝑎 𝜙𝑜1 ⎞ ⎝⎩ ⎭ ⎩ ⎭⎠
1 ⎜ 𝑥 ⎟𝑑 𝑥𝑑 𝑦
= (20)
2 ∫0 ∫0 ⎜ 𝑄𝑦 𝑎
𝜙𝑜2 ⎟ where,
⎝ ⎠
ℎ∕2−𝐶 ℎ∕2−𝐶
After substituting the shear resultants (𝑄𝑎𝑥 , 𝑄𝑎𝑦 ) expression (Eq. (13d))
𝜌1 = 𝜌(𝑧)dz and 𝜌3 = 𝜌(𝑧)𝑧2 dz (28)
in Eq. (20) then shear energy can be expressed as, ∫−ℎ∕2−𝐶 ∫−ℎ∕2−𝐶
(
𝑎 𝑏 { 𝑜 }𝑇 [ ] { 𝑜 })
𝑠 1 𝜙2 𝐻44 𝐻45 𝜙2
𝑈 = 𝑜 𝑑 𝑥𝑑 𝑦 (21)
2 ∫0 ∫0 𝜙1 𝐻45 𝐻55 𝜙𝑜1 2.3. Transformation relation

In the strain energy expressions, FG plate stiffness’s Aij , Bij , Cij , Dij , The domain of the skew plate in the x–y plane is transformed into
Eij , Fij and Hij are defined as, unit square plate domain in the 𝜉–𝜂 plane (see Fig. 2) by the transfor-
mation rule, which is given in Eq. (29).
ℎ∕2−𝐶
(𝐴𝑖𝑗 , 𝐵𝑖𝑗 , 𝐷𝑖𝑗 ) = ∫−ℎ∕2−𝐶 𝑄𝑖𝑗 (1, 𝑧, 𝑧2 )𝑑𝑧 (𝑖, 𝑗 = 1, 2, 6)
𝑥 = 𝑎𝜉 + 𝑏𝜂 sin 𝜓; 𝑦 = 𝑏𝜂 cos 𝜓 (29)
ℎ∕2−𝐶
(𝐶𝑖𝑗 , 𝐸𝑖𝑗 , 𝐹𝑖𝑗 ) = ∫−ℎ∕2−𝐶 𝑄𝑖𝑗 (1, 𝑧, 𝑓 (𝑧))𝑓 (𝑧)𝑑𝑧 (𝑖, 𝑗 = 1, 2, 6) (22)
The relation between first order derivative in the physical domain
ℎ∕2−𝐶
and first order derivative computational domain are given by
𝐻𝑖𝑗 = ∫−ℎ∕2−𝐶 𝑄𝑖𝑗 𝑓 ′ (𝑧)𝑓 ′ (𝑧)𝑑𝑧 (𝑖, 𝑗 = 4, 5)
⎡ 𝜕 ⎤ ⎡ 1 ⎤⎡ 𝜕 ⎤
⎢ 𝜕𝑥 ⎥ ⎢ 𝑎 0 ⎥⎢ ⎥
where, 𝜕𝜉
⎢ 𝜕 ⎥ = ⎢ tan 𝜓 sec 𝜓
⎥⎢ 𝜕 ⎥ (30)
⎡1 ⎢ ⎥ ⎢− ⎥⎢ ⎥
𝜈 0 ⎤ ⎣ 𝜕𝑦 ⎦ ⎣ 𝑎 𝑏 ⎦⎣ 𝜕𝜂 ⎦
𝐸(𝑧) ⎢𝜈 1 0 ⎥
𝑄= ⎢ ⎥ (𝑖, 𝑗 = 1, 2, 6) (23)
1 − 𝜈2 ⎢ 1−𝜈⎥ The relation between second order derivative in the physical domain
⎣0 0
2 ⎦ and second order derivative computational domain are given by

44
R. Kumar et al. International Journal of Mechanical Sciences 130 (2017) 41–51

Fig. 2. Mapping of the skew plate domain into a unit square plate domain.

⎡ 𝜕 ⎤ ⎡ 1
2 𝜕2 where, N∗ is the total number of terms and Uj , Vj , Wj , Kj and Lj are
0 0 ⎤⎡ 2 ⎤
⎢ 𝜕 𝑥2 ⎥ ⎢ 𝑎2 ⎥ 𝜉 ⎥
⎢ 𝜕 ∧ ∧ ∧ ∧ ∧
⎢ 2 ⎥ ⎢ tan2 𝜓 unknown constants. 𝛼𝑗 (𝜉, 𝜂), 𝛽𝑗 (𝜉, 𝜂), 𝜒𝑗 (𝜉, 𝜂), 𝛾𝑗 (𝜉, 𝜂) 𝑎𝑛𝑑 𝛿𝑗 (𝜉, 𝜂) are
⎢ 𝜕 ⎥=⎢ sec2 𝜓 2 tan 𝜓 sec 𝜓 ⎥⎢⎢ 𝜕 2 ⎥⎥
⎢ 𝜕 𝑦2 ⎥ ⎢ 𝑎2
−
𝑎𝑏 ⎥⎢ 2 ⎥ (31) boundary characteristics orthonormal polynomials (BCOPs), which are
𝑏2 ⎥⎢ 𝜕 𝜂 ⎥
⎢ 2 ⎥ ⎢ tan 𝜓 sec 𝜓 generated by orthogonalization of boundary characteristic polynomials
⎢ 𝜕 ⎥ ⎣− 2 0 ⎥⎢ 𝜕 2 ⎥
⎣ 𝜕 𝑥𝜕 𝑦 ⎦ 𝑎 𝑎𝑏 ⎦⎣ via Gram–Schmidt process [29]. The boundary characteristics polyno-
𝜕 𝜉𝜕 𝜂 ⎦
mial (Fj (𝜉,𝜂)) over the domain 0 ≤ 𝜉 ≤ 1, 0 ≤ 𝜂 ≤ 1 are generated by
The transformation rule and derivative relation in Eqs. (29)–(31) are taking the product of the basis functions and two-dimensional indepen-
used to transform the strain energy in Eqs. (14)–(21), potential energy dent set of polynomials follows as,
in Eq. (26) and kinetic energy in Eq. (27) from x–y plane to 𝜉–𝜂 plane
and given by 𝐹𝑗 (𝜉, 𝜂) = Θ(𝜉, 𝜂, 𝑝, 𝑞, 𝑟, 𝑠) ∗ 𝑓𝑗 (𝜉, 𝜂), 𝑗 = 1, 2, 3..., 𝑁 ∗ (39)
1 1 where,
1
𝑈 ∗𝑚 = ℜ1 (𝑢𝑜,𝜉 , 𝑢𝑜,𝜂 , 𝑣𝑜,𝜉 , 𝑣𝑜,𝜂 )𝑐 𝑜𝑠𝜓𝑑𝜉𝑑𝜂 (32)
2 ∫0 ∫0
Θ(𝜉, 𝜂, 𝑝, 𝑞, 𝑟, 𝑠) = 𝜉 𝑝 (1 − 𝜉)𝑞 𝜂 𝑟 (1 − 𝜂)𝑠 (40)
1 1
1
𝑈 ∗𝑏 = ℜ2 (𝑤𝑜,𝜉𝜉 , 𝑤𝑜,𝜉𝜂 , 𝑤𝑜,𝜂𝜂 , 𝜙𝑜1,𝜉 , 𝜙𝑜1,𝜂 , 𝜙𝑜2,𝜉 , 𝜙𝑜2,𝜂 )𝑐 𝑜𝑠𝜓𝑑𝜉𝑑𝜂 (33) and,
2 ∫0 ∫0 { }
𝑓𝑗 (𝜉, 𝜂) = 1, 𝜉, 𝜂, 𝜉 2 , 𝜉𝜂, 𝜂 2 , 𝜉 3 , 𝜉 2 𝜂, 𝜉𝜂 2 , 𝜂 3 , ... (41)
1 1
∗𝑎𝑏 1
𝑈 = ℜ3 (𝑤𝑜,𝜉𝜉 , 𝑤𝑜,𝜉𝜂 , 𝑤𝑜,𝜂𝜂 , 𝜙𝑜1,𝜉 , 𝜙𝑜1,𝜂 , 𝜙𝑜2,𝜉 , 𝜙𝑜2,𝜂 )𝑐 𝑜𝑠𝜓𝑑𝜉𝑑𝜂 (34) In Eq. (40), the value of p, q, r and s depends on the boundary condi-
2 ∫0 ∫0
tions. In particular, the value of p = 0, 1 and 2 assigned for free, simply
1
1 1 supported and clamped boundary, respectively. The Gram–Schmidt pro-
𝑈 ∗𝑠 = ℜ4 (𝜙𝑜1 , 𝜙𝑜2 )𝑐 𝑜𝑠𝜓𝑑𝜉𝑑𝜂 (35) cess for generating BCOPs are as follows
2 ∫0 ∫0
𝑗−1
∑
1 1
𝑉∗ =−
1
ℜ5 (𝑛𝑥𝑥 , 𝑛𝑥𝑦 , 𝑛𝑦𝑦 , 𝑢𝑜,𝜉 , 𝑢𝑜,𝜂 , 𝑣𝑜,𝜉 , 𝑣𝑜,𝜂 , 𝑤𝑜,𝜉 , 𝑤𝑜,𝜂 )𝑐 𝑜𝑠𝜓𝑑𝜉𝑑𝜂 (36) 𝜒1 (𝜉, 𝜂) = 𝐹1 (𝜉, 𝜂), 𝜒𝑗 (𝜉, 𝜂) = 𝐹𝑗 (𝜉, 𝜂) − 𝛼𝑗𝑖 𝜒𝑖 (𝜉, 𝜂), (42)
2 ∫0 ∫0 𝑖=1

1
1 1 < 𝐹𝑗 (𝜉, 𝜂), 𝜒𝑖 (𝜉, 𝜂) >
𝑇∗ = ℜ6 (𝑢𝑜 , 𝑣𝑜 , 𝑤𝑜 , 𝜙𝑜1 , 𝜙𝑜2 )𝑐 𝑜𝑠𝜓𝑑𝜉𝑑𝜂 (37) 𝛼𝑗𝑖 = , 𝑖 = 1, 2, 3, ..., (𝑗 − 1), 𝑗 = 2, 3, 4,...,𝑁 ∗ (43)
𝐸 2 ∫0 ∫0 < 𝜒𝑖 (𝜉, 𝜂), 𝜒𝑖 (𝜉, 𝜂) >

The details of ℜi (i = 1,2,…6) are given in Appendix A The inner product of the functions 𝜑i (𝜉,𝜂) and 𝜑j (𝜉,𝜂) can be defined
as
2.4. Derivation of orthogonal polynomial 1 1
< 𝜒𝑖 (𝜉, 𝜂), 𝜒𝑗 (𝜉, 𝜂) >= 𝜒𝑖 (𝜉, 𝜂)𝜒𝑗 (𝜉, 𝜂)𝑑 𝜉𝑑 𝜂 (44)
∫0 ∫0
The following displacement fields are considered to satisfying the
essential boundary conditions in computational domain and can be ex- The norm of the function 𝜒 j (𝜉,𝜂) is given by
pressed as,
[ ]1∕2
1 1
𝑁
∑
∗
‖ ‖
∧ ‖𝜒𝑗 (𝜉, 𝜂)‖ =< 𝜒𝑗 (𝜉, 𝜂), 𝜒𝑗 (𝜉, 𝜂)> =
1∕2
𝜒 (𝜉, 𝜂)𝜒𝑗 (𝜉, 𝜂)𝑑 𝜉𝑑 𝜂
𝑢𝑜 (𝜉, 𝜂, 𝑡) = 𝑈𝑗 𝛼𝑗 (𝜉, 𝜂) cos (𝜔𝑡) ‖ ‖ ∫0 ∫0 𝑗
𝑗=1
(45)
𝑁
∑
∗
∧
𝑜
𝑣 (𝜉, 𝜂, 𝑡) = 𝑉𝑗 𝛽𝑗 (𝜉, 𝜂) cos (𝜔𝑡) The normalization has been done by
𝑗=1
∧ 𝜒𝑗 (𝜉, 𝜂)
𝑁
∑
∗
𝜒𝑗 (𝜉, 𝜂) = (46)
𝑤𝑜 (𝜉, 𝜂, 𝑡) =
∧
𝑊𝑗 𝜒𝑗 (𝜉, 𝜂) cos (𝜔𝑡) ‖ ‖
‖𝜒𝑗 (𝜉, 𝜂)‖
𝑗=1
‖ ‖
𝑁 ∗ Substituting Eqs. ((38)–(46)) to Eqs. ((32)–(37)), then total energy
∑ ∧
𝜙1 𝑜 (𝜉, 𝜂, 𝑡) = 𝐾𝑗 𝛾𝑗 (𝜉, 𝜂) cos (𝜔𝑡) functional П (U +V +TE ) are expressed in terms of unknown constants
𝑗=1 (Uj , Vj , Wj , Kj and Lj ). These constants are determined by minimizing
𝑁
∑
∗
∧ the total energy functional using Rayleigh–Ritz method and procedure
𝜙2 𝑜 (𝜉, 𝜂, 𝑡) = 𝐿𝑗 𝛿𝑗 (𝜉, 𝜂) cos (𝜔𝑡) (38) are defined as,
𝑗=1
45
R. Kumar et al. International Journal of Mechanical Sciences 130 (2017) 41–51

⎧{𝑈 } ⎫
⎪{𝑉 } ⎪ |𝐾 ∗ −0.5𝛽𝑁𝑐𝑟 𝐾𝐺 || |0 0 ||
| |
∑ ⎪ ⎪ | | − 𝑝21 | |=0 (55)
[𝐾 − 𝜆𝑀 ]⎨{𝑊 }⎬ = 0 |−𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾∗ | |0 𝑀 ||
(47) | | |
⎪{𝐾 } ⎪
⎪ ⎪ The secondary instability region with second-order approximation
⎩{𝐿} ⎭ is,
|𝐾 ∗ −0.5𝛽𝑁𝑐𝑟 𝐾𝐺 || |𝑀 0||
| |
⎧𝑈1 ⎫ ⎧𝑉 1 ⎫ ⎧𝑊 1 ⎫ ⎧𝐾 1 ⎫ | | − 𝑝22 | |=0 (56)
⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ |−0.5𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾 − 4𝑀𝑝1 ||
∗ 2 |0 0||
𝑈 𝑉 𝑊 | |
⎪ 2 ⎪ ⎪ 2 ⎪ ⎪ 2 ⎪ ⎪𝐾 2 ⎪
⎪. ⎪ ⎪. ⎪ ⎪. ⎪ ⎪. ⎪ |𝐾 ∗ − 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 | |0 0 ||
{𝑈 } = ⎨ ⎬, {𝑉 } = ⎨. ⎬, {𝑊 } = ⎨. ⎬, {𝐾 } = ⎨. ⎬, | 0 | | 0
. | | | |
⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ |−𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾 ∗ − 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 || − 𝑝22 |0 𝑀 0|| = 0 (57)
| |
⎪. ⎪ ⎪. ⎪ ⎪. ⎪ ⎪. ⎪ | | |0
⎪𝑈 ∗ ⎪ ⎪𝑉 ∗ ⎪ ⎪𝑊 ∗ ⎪ ⎪𝐾 ∗ ⎪ | 0 − 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾 − 4𝑀𝑝1 ||
∗ 2
| 0 0 ||
⎩ 𝑁 ⎭ ⎩ 𝑁 ⎭ ⎩ 𝑁 ⎭ ⎩ 𝑁 ⎭ |

⎧𝐿1 ⎫ The third region of instability corresponding to period 2T with first-

⎪ ⎪ order approximation is,
⎪𝐿2 ⎪
⎪. ⎪ |𝐾 ∗ + 0.5𝛽𝑁 𝐾 − 0.25𝑀𝑝2 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 |
{𝐿} = ⎨ | 𝑐𝑟 𝐺 |
⎬ (48) | 3
|=0 (58)
⎪. ⎪ | 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾∗ − 2.25𝑀𝑝24 ||
|
⎪. ⎪
⎪𝐿 ∗ ⎪ |𝐾 ∗ − 0.5𝛽𝑁 𝐾 − 0.25𝑀𝑝2 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 ||
⎩ 𝑁 ⎭ | 𝑐𝑟 𝐺
| 3
∗ − 2.25𝑀𝑝2 || = 0 (59)
| 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾
| 4 |
Eq. (47) constitutes a set of 5 N∗ simultaneous ordinary differential
equations referred as Ritz system. The fourth region instability corresponding to period T, first-order
approximation is,
2.5. Dynamic instability analysis |𝐾 ∗ − 𝑀𝑝2 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 |
| 0 |
| 3 |
|0.5𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾∗ − 4𝑀𝑝21 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 || = 0 (60)
|
The dynamic instability of the FG skew plate under uniform and | |
|0 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾 ∗ − 9𝑀𝑝24 ||
linearly varying in-plane periodic loading 𝑁𝑥𝑥 = 𝑁𝑠 + 𝑁𝑡 cos 𝑝𝑡 is ana- |
lyzed. In the expression of 𝑁𝑥𝑥 , the static and dynamic components can |𝐾 ∗ |
| 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 0 0 |
be defined as, Ns = 𝛼Ncr and Nt = 𝛽Ncr , respectively. Where 𝛼 and 𝛽 are | |
|𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾∗ − 𝑀𝑝23 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 0 |
static and dynamic load factors, respectively, and Ncr is the critical buck- | |=0 (61)
| 0 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾 ∗ − 4𝑀𝑝23 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 ||
ling load. The ordinary differential equation (Mathieu–Hill equation) |
| |
| 0 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾 ∗ − 9𝑀𝑝24 ||
describing the plate dynamic instability can be expressed as, | 0
{ .. } [ ]
[𝑀] 𝛿 + [𝐾] − (𝑁𝑠 + 𝑁𝑡 cos 𝑝𝑡)[𝐾𝐺 ] {𝛿} = {0} (49) where K∗ = [K] − Ns [KG ]. Solving for p4 from (𝐾 ∗ − 2.25𝑀𝑝24 ) and substi-
tuting for p3 of Eqs. (58) and (59) one can calculate third zone of stability
where, [M], [K] and [KG ] and are, respectively the mass, linear and
of period 2T. Similarly, solving for p4 from (𝐾 ∗ − 9𝑀𝑝24 ) and substitut-
geometric stiffness matrices. The critical buckling load is evaluated from
ing for p3 of Eqs. (60) and (61) one can calculate fourth zone of sta-
the solution of eigenvalue problem by omitting the [M] matrix and the
bility of period T. The time history responses and corresponding phase
time dependent term (Nt cos pt) term in Eq. (49). Similarly, the solution
plots are obtained by solving the set of ordinary differential Eq. (49) us-
of the eigenvalue problem omitting the [KG ] matrix gives the natural
ing Newmark’s numerical integration method in conjunction with the
frequencies. On the boundaries of instability regions, the Mathieu–Hill
Newton–Raphson method.
equation (Eq. (49)) has a periodic solution with period 2T or T and
solution can be expressed in the form of Fourier series as,
3. Results and discussions
∑∞ ( )
𝑘𝑝𝑡 𝑘𝑝𝑡
𝛿(𝑡) = 𝑎𝑘 sin + 𝑏𝑘 cos (50)
2 2 Following material properties is considered for the present investi-
𝑘=1,3,5
gations. The Young’s modulus of metal (Em ) and ceramic (Ec ) is 70 GPa
( )
∑
∞
𝑘𝑝𝑡 𝑘𝑝𝑡 and 380 GPa, respectively. The mass density of metal (𝜌m ) and ceramic
𝛿(𝑡) = 𝑏0 + 𝑎𝑘 sin + 𝑏𝑘 cos (51)
2 2 (𝜌c ) is 2702 kg/m3 and 3800 kg/m3 , respectively, and Poisson’s ratio is
𝑘=2,4,6
𝜈 = 0.3. The dynamic instability
√ regions are traced as a dimensionless
where ak and bk are arbitrary constants. Substituting Eqs. (50) and excitation frequencies (𝑝 𝜌𝑐 ∕𝐸𝑐 ℎ2 ) against dynamic load factor (𝛽).
(51) into Eq. (49) and equating the coefficients of identical sin 𝑘𝑝𝑡
2
and
cos 𝑘𝑝𝑡
2
leads to a system of homogeneous algebraic equations in ak and 3.1. Validation of results
bk . For a nontrivial solution, the determinant of the coefficient matrix
of ak and bk must vanish, which leads to eigenvalue problems. Solving For validating the present mathematical formulation, a simply sup-
these eigenvalue problems corresponding period 2T and T gives the prin- ported FG skew plate (a/b = 1, 𝜓 = 0°) under uniform in-plane loading
cipal instability region and secondary instability region, respectively. for different span to thickness ratios (a/h) is considered. The critical
The principal instability region with first-order approximation is, buckling loads obtained from the present mathematical formulation are
| ∗ | compared with the results available in Thai and Choi [30] and reported
|𝐾 ± 0.5𝛽𝑁𝑐𝑟 𝐾𝐺 − 0.25𝑀𝑝21 | = 0 (52)
| | in Table 1. It is observed that results are well matched with the pub-
The principal instability region with second-order approximation is, lished reference value.
Table 2 shows the non-dimensional critical buckling of a simply sup-
|𝐾 ∗ ± 0.5𝛽𝑁 𝐾 −0.5𝛽𝑁𝑐𝑟 𝐾𝐺 || |0.25𝑀 0||
| 𝑐𝑟 𝐺 | ported FG skew plate (a/h = 100, a/b = 1) for different skew angles (𝜓)
| | − 𝑝22 | |=0 (53)
| −0.5𝛽𝑁𝑐𝑟 𝐾𝐺 𝐾∗ − 2.25 𝑀𝑝1 |
2 | | 0 0|| under uniform in-plane loading. The critical buckling load is compared
| |
with the results available in literature [31,32]. It is observed that the
The secondary instability region with first-order approximation is,
critical buckling loads for different skew angles are well matched with
| ∗ |
|𝐾 − 𝑀𝑝21 | = 0 (54) the published reference value.
| |

46
R. Kumar et al. International Journal of Mechanical Sciences 130 (2017) 41–51

Table 1
The critical buckling loads (k1 ) for a simply supported FG square plate.

Power law a/h = 100 a/h = 10

Present Thai and Choi [30] Present Thai and Choi [30]

n=0 19.61 19.61 18.52 18.58

n=1 9.77 9.78 9.32 9.34
n=2 7.63 7.63 7.24 7.26

k1 = Nxcr /Em h3 .

Table 2
The critical buckling loads (k2 ) for simply supported FG skew plates (a/h =
100, a/b = 1) for different skew angles (𝜓) under uniform in-plane loading.

Skew angle n=0 n=1

Present Wang [31] Present Ganapati et al. [32]

0° 4.00 4.00 2.20 2.20

15° 4.39 4.39 2.43 2.42
30° 5.90 5.89 3.32 3.29
45° 10.23 10.10 5.92 5.77

k2 = Nxcr b2 /𝜋 2 Dc . Fig. 4. The effect of higher order of approximation on the different instability zones of
FG skew plate.

Fig. 3. First and second order approximation to the principal instability zones of FG
square plate.
Fig. 5. Effect of skew angle on the principal instability zones of FG plates.

In Fig. 3, the principal instability zones of a simply supported (SSSS)

FG plate (a/b = 1, a/h = 100, 𝜈= 0.25, n = 0) obtained from the
present method are compared with finite element method (FEM) results
reported by Hutt and Salam [33]. The principal instability √ zones are
presented as a dimensionless excitation frequency (Ω = 𝑝𝑏2 𝜌ℎ∕𝐷) for
plate against dynamic load factor (𝛽). It is observed from the figure that
the present first order approximation to the principal instability regions
compares very well with the reference instability zones.
3.2. Effect of higher order approximation
First four zones of instability for a simply supported FG skew plate
(a/b = 1, a/h = 100, 𝜓 = 30°, n = 1, 𝛼 = 0) under uniform in-plane
periodic loading is presented in Fig. 4. The odd number (e.g. 1st and
3rd) of instability zone are corresponds to 2T period and the even num-
ber (i.e. 2nd and 4th) instability zone corresponding to T period. It can
readily be observed from the figure that as the dynamic load factor in-
creases the width of all four instability zone increases. The first zone
of instability has the higher width compared to other three instabil-
ity zone for a particular dynamic load factor. Thus, first zone (prin-
cipal zone) has the greater importance from the practical application
point of view. Moreover, the second order approximation for the prin-
cipal and the secondary instability region has less width as compared
to the first order approximation. The width of instability is defined by
ΔΩ = Ωu − Ωl for the given value of dynamic load factor. Where, Ωl and
Ωu are the lower and upper critical value of dimensionless excitation fre-
quencies. The lower and upper stability region is defined by Ω < Ωl and
Ω > Ωu respectively, and unstable region is defined by Ωl ≤ Ω ≤ Ωu (see
Fig. 4). At dynamic load factor (𝛽 = 0.8),
√ the width of the principal
instability zone (1st zone) is Δ𝑝 = 4√.44 𝐸𝑐 ℎ2 ∕𝜌𝑐 and secondary insta-
bility zone (2nd zone) is Δ𝑝 = 0.93 𝐸𝑐 ℎ2 ∕𝜌𝑐 for the second order ap-
proximation. Since the principal instability zone has greater practical
importance the parametric studies are presented only for the principal
instability zone with second order approximation.
3.3. Influence of skew angle
The effect of the skew angle on the principal instability zone of a
simply supported FG plate (a/b = 1, a/h = 100, n = 1, 𝛼 = 0) subjected
to uniform in-plane periodic loading is presented in Fig. 5. It is to be
observed that the width of the principal zone of instability decreases
as the skew angle increases. For a particular dynamic load factor
√ (𝛽 =
0.8) the widths of principal instability zones (Δp) are 3.64 𝐸𝑐 ℎ2 ∕𝜌𝑐 ,

47
R. Kumar et al.
Fig. 6. Effect of shear deformation on the principal instability zones of FG skew plates.
Fig. 7. Effect of linearly varying in-plane loading on the principal instability zones of FG
skew plates.
√ √ √ a particular dynamic load√
3.46 𝐸𝑐 ℎ2 ∕𝜌𝑐 , 2.98 𝐸𝑐 ℎ2 ∕𝜌𝑐 and 2.23 𝐸𝑐 ℎ2 ∕𝜌𝑐 for the skew angle (𝜓)
0°, 15°, 30° and 45°, respectively.
3.4. Influence of shear deformation
The effect of the shear deformation on the principal instability zone
for a simply supported FG skew plate (a/b = 1, 𝜓 = 30°, n = 1, 𝛼 = 0)
under uniform in-plane periodic loading is presented in Fig. 6. It is ob-
served that the influence of shear deformation on the non-dimensional
frequency becomes pronounced when the ratio of span to thickness (a/h)
decrease. Also, the width of the principal instability zone decreases
with the decease of span to thickness (a/h). For a particular dynamic
load√factor (𝛽 = 0.9),
√ the widths of√principal instability
√ zones (Δp) are
4.98 𝐸𝑐 ℎ2 ∕𝜌𝑐 , 4.96 𝐸𝑐 ℎ2 ∕𝜌𝑐 , 4.93 𝐸𝑐 ℎ2 ∕𝜌𝑐 and 4.47 𝐸𝑐 ℎ2 ∕𝜌𝑐 for side
to thickness ratios (a/h) 100, 50, 20 and 10, respectively.
3.5. Effect of different type of linearly varying loadings
Fig. 7 shows the principal instability zone of a simply supported
FG skew plate (a/b = 1, a/h = 100, 𝜓 = 30°, n = 1, 𝛼 = 0.25) under
different types of linearly varying in-plane periodic loadings. Differ-
ent types of linearly varying loads are obtained by considering differ-
ent values of 𝜆 in the loading expressions 𝑁𝑥𝑥 = 𝑁0 (1 − 𝜆( 𝜂𝑏 )). Here,
48
International Journal of Mechanical Sciences 130 (2017) 41–51
Fig. 8. Effect of aspect ratio on the principal instability zones of FG skew plates.
𝜆 = 0.0, 0.5, 1.0, 1.5, 2.0 indicates uniform, trapezoidal, triangular,
partial tension and pure bending type of in-plane load, respectively. It
can be readily observed that the principal instability zone is maximum
for uniformly distributed load (𝜆 = 0) and minimum for pure bending
load (𝜆 = 2). The phenomena can be attributed as the lower buckling
load of FG skew plate for uniformly distributed in-plane edge loading.
For a particular dynamic load√ factor (𝛽 = 0.5), √ the widths of√principal
instability
√ zones ( Δp ) are
√ 3 . 15 𝐸 𝑐 ℎ 2 ∕𝜌 , 2.33 𝐸 ℎ2 ∕𝜌 ,1.56 𝐸 ℎ2 ∕𝜌 ,
𝑐 𝑐 𝑐 𝑐 𝑐
0.64 𝐸𝑐 ℎ2 ∕𝜌𝑐 and 0.04 𝐸𝑐 ℎ2 ∕𝜌𝑐 for different linearly varying load (𝜆
= 0.0, 0.5, 1.0, 1.5, 2.0).
3.6. Effect of aspect ratio
The effect of aspect ratio on the principal instability zone of a simply
supported FG skew plate (a/h = 100, 𝜓 = 30°, n = 1, 𝛼 = 0) under uniform
in-plane periodic loading is presented in Fig 8. Here, length ‘a’ is kept
constant and width ‘b’ is varied to obtain a different aspect ratio of the
skew plate. The Figure shows the principal instability zone for different
aspect ratio (i.e. 0.5, 1.0, 1.5, 2.0) of the plate. It is observed that the in-
stability zones increases as the slenderness of the plate increases and the
forcing frequency increases as the a/b ratio of skew plate increases. For
factor (𝛽 = 0.9),√the widths of principal insta-
√
√zones (Δp) are 3.10 𝐸𝑐 ℎ ∕𝜌𝑐 , 4.98 𝐸𝑐 ℎ ∕𝜌𝑐 , 6.88 𝐸𝑐 ℎ ∕𝜌𝑐 and
bility 2 2 2
7.01 𝐸𝑐 ℎ2 ∕𝜌𝑐 for aspect ratio (a/b) 0.5, 1.0, 1.5 and 2.0, respectively.
3.7. Effect of power law
The effect of power law on the principal instability zone a simply
supported FG skew plates (a/b = 1, a/h = 100, 𝜓 = 30°, 𝛼 = 0) under
uniform in-plane periodic loading is shown in Fig 9. The Figure shows
the principal instability zone for the different power of ‘n’ (i.e. 0.5, 1.0,
2.0 and 4.0) for the skew plate. It is observed from the figure that width
of the principal zone increases with the decrease of the power law (n).
For a particular dynamic load√ factor (𝛽 = 0.9),
√ the widths of√principal
√ zones (Δp) are 4.22 𝐸𝑐 ℎ ∕𝜌𝑐 , 4.52 𝐸𝑐 ℎ ∕𝜌𝑐 , 4.98 𝐸𝑐 ℎ ∕𝜌𝑐
instability 2 2 2
and 5.50 𝐸𝑐 ℎ2 ∕𝜌𝑐 for power law (n) 4, 2, 1 and 0.5 respectively.
3.8. Effect of boundary conditions
Fig. 10 shows that the effect of edge restraint on the principal insta-
bility zone of a FG skew plate (a/b = 1, a/h = 100, 𝜓 = 30°, n = 1, 𝛼
= 0) under uniform in-plane periodic loading. Four different boundary
conditions, e.g. all edges simply supported (SSSS), all edges clamped
(CCCC), simply supported on the loaded edge and the other two edges
R. Kumar et al.
Fig. 9. Effect of power law on the principal instability zones of FG skew plates.
Fig. 10. Effect of edge restraint on the principal instability zones of FG skew plates.
clamped (SCSC), clamped on the loaded edges and the other two edges
are simply supported (CSCS) has been considered in this present study.
It is observed that the edge restraint has a significant effect on the prin-
cipal instability zone of FG skew plate. The frequency of FG skew plate
increase in the following order SSSS<CSCS<SCSC<CCCC at the same
dynamic load factor (𝛽 = 𝑁𝑡 ∕𝑁𝑐𝑟 of SSSS FG skew plate). The width of
instability zone of CCCC skew plate is the lowest among all bound-
ary conditions due to its higher stiffness. For a particular dynamic load
√ (𝛽 = 0.9), the
factor √widths of the√principal instability√ zones (Δp) are
3.28 𝐸𝑐 ℎ2 ∕𝜌𝑐 , 3.42 𝐸𝑐 ℎ2 ∕𝜌𝑐 , 4.89 𝐸𝑐 ℎ2 ∕𝜌𝑐 and 4.98 𝐸𝑐 ℎ2 ∕𝜌𝑐 for dif-
ferent boundary conditions CCCC, SCSC, CSCS and SSSS respectively.
3.9. Effect of static load factor
Influence of static load factor on the principal instability zone of a
simply supported FG skew plate (a/b = 1, a/h = 100, 𝜓 = 30°, n = 1)
under uniform in-plane periodic loading is shown in Fig. 11. It is clear
from the figure the width of the primary instability zone is increasing as
the static load factor (𝛼) increases due to the decrease in the stiffness of
the plate. For a particular dynamic load √ factor (𝛽 = 0.4),√the widths of
principal
√ instability zones (Δp) are 2.65 𝐸𝑐 ℎ2 ∕𝜌𝑐 , 2.90 𝐸𝑐 ℎ2 ∕𝜌𝑐 and
3.57 𝐸𝑐 ℎ2 ∕𝜌𝑐 for different static load factor (𝛼) 0.0, 0.25 and 0.5,
respectively.
49
International Journal of Mechanical Sciences 130 (2017) 41–51
Fig. 11. Effect of static load factor on the principal instability zones of FG skew plates.
Fig. 12. The principal instability zone with constant and varying Poisson’s ratio across
the thickness of FG skew plate.
3.10. Effect of Poisson’s ratio
In the following section, the effect of the Poisson’s ratio on the dy-
namic instability zones of FG plate is studied. The Poisson’s ratios for
metal 𝜈 m = 0.33 and for ceramic 𝜈 c = 0.26 are considered. The Poisson’s
ratio (𝜈) varies across the thickness according to the rule of mixture are
given by
𝜈(𝑧) = 𝜈𝑚 𝑉𝑚 (𝑧) + 𝜈𝑐 𝑉𝑐 (𝑧) (54)
where, subscripts m and c represents the metal and ceramic constituents,
respectively. Here, the volume fraction of the metal and ceramic phase
is described by the power-law distribution (see, Eq. (4)). The volume
fraction varies through the thickness based on the power index n. The
principal instability zone of a simply supported FG skew plate (a/b = 1,
a/h = 100, 𝜓 = 30°, n = 1, 𝛼 = 0) with constant and varying Poisson’s
ratio across the thickness under uniform in-plane periodic loading are
shown in Fig. 12. It is observed from the figure that the width of the
principal instability zone with a varying Poisson’s ratio across the thick-
ness is almost the same as the width of the principal instability zone
with constant Poisson’s ratio across the thickness.
R. Kumar et al. International Journal of Mechanical Sciences 130 (2017) 41–51

Fig. 13. Time history response and corresponding phase plot of a simply supported FG skew plate (a/b = 1, a/h = 100, 𝜓 = 30°, n = 1, 𝛼 = 0, 𝛽 = 0.3) under uniform periodic axial
compression with excitation frequency parameter (Ωa = 11.42 i.e. Ωl < Ωa < Ωu ) from unstable regions.

Fig. 14. Time history response and corresponding phase plot of a simply supported four layered FG skew plate (a/b = 1, a/h = 100, 𝜓 = 30°, n = 1, 𝛼 = 0, 𝛽 = 0.3) under uniform periodic
axial compression with dimensionless excitation frequency parameter (Ωb = 10.27 i.e. Ωb < Ωl ) from lower stability region.

3.11. Time history response and corresponding phase plot ing has been studied. The width of the principal instability zone is more
than other different instability zones. Hence, the studies of the principal
The characteristic features of the instability regions are studied in instability zone have greater practical importance. As the skew angle
terms of time history response plots. Dynamic instability regions of the increases the width of the instability region becomes smaller. Accord-
FG skew plate are shown in Fig. 4. Fig. 13 shows the time history re- ing to the power law distribution, the width of instability zone increases
sponse and corresponding phase plot of a simply supported FG skew with the decrease of the power law. The width of instability region is the
plate (a/b = 1, a/h = 100, 𝜓 = 30°, n = 1, 𝛼 = 0, 𝛽 = 0.3) under uni- lowest when all edges are clamped compared to all other different com-
form periodic axial compression with the dimensionless excitation fre- binations of boundary conditions. Keeping the same intensity of loading,
quencies from unstable regions (Ωa = 11.42). The initial condition are the width of instability zone is more for uniform loading compared to
w(0) = 0.1 h and 𝑤̇ (0) = 0. It is observed from the figures that the all other linearly varying loading. In the unstable region, the time his-
time history response grows exponentially (i.e. unbounded) and FG skew tory response grows exponentially in a very short duration and FG skew
plate becomes dynamically unstable. The same behavior is observed in plate become dynamically unstable. But in the stable region, the time
phase plots (dw/dt verses w/h) which are spirals with exponentially in- history response becomes bounded and also beats is observed.
creasing amplitude indicating instability.
Fig. 14 shows the time history response and corresponding phase Appendix A
plot of a simply supported FG skew plate (a/b = 1, a/h = 100, 𝜓 = 30°,
n = 1, 𝛼 = 0, 𝛽 = 0.3) under uniform periodic axial compression with 1 2
the dimensionless excitation frequencies from lower stability region ℜ1 (𝑢𝑜,𝜉 , 𝑢𝑜,𝜂 , 𝑣𝑜,𝜉 , 𝑣𝑜,𝜂 ) =
(𝐴11 + 𝐴66 tan2 𝜓 )(𝑢𝑜,𝜉 )2 − 𝐴66 tan 𝜓 sec 𝜓 𝑢𝑜,𝜉 𝑢𝑜,𝜂
𝑎2 𝑎𝑏
(Ωb = 10.27). The initial condition are w(0) = 0.1 h and 𝑤̇ (0) = 0. It is ob- 1 2 2
+ 𝐴66 sec2 𝜓 (𝑢𝑜,𝜂 )2 − (𝐴11 + 𝐴66 ) tan 𝜓 𝑢𝑜,𝜉 𝑣𝑜,𝜉 + 𝐴12 sec 𝜓 𝑢𝑜,𝜉 𝑣𝑜,𝜂
served from the figure that the time history response becomes bounded 𝑏2 𝑎2 𝑎𝑏
and also beats is observed. The maximum displacement of FG skew plate 2 1 2
+ 𝐴66 sec 𝜓 𝑢𝑜,𝜂 𝑣𝑜,𝜉 + (𝐴22 tan2 𝜓 + 𝐴66 )(𝑣𝑜,𝜉 )2 − 𝐴22 tan 𝜓 sec 𝜓 𝑣𝑜,𝜉 𝑣𝑜,𝜂
obtained is same as the initial displacement amplitude. From the phase 𝑎𝑏 𝑎2 𝑎𝑏
plot, it is observed that the skew plate center starts from initial condi- 1
+ 𝐴22 sec2 𝜓 (𝑣𝑜,𝜂 )2 (A.1)
tions (w(0)/h= 0.1, 𝑤̇ (0) = 0) and spirals in till the amplitude reaches 𝑏2
the minimum value. Again, it moves towards the initial condition as the
1
beats are observed in the response. ℜ2 (𝑤𝑜,𝜉𝜉 , 𝑤𝑜,𝜉𝜂 , 𝑤𝑜,𝜂𝜂 , 𝜙𝑜1,𝜉 , 𝜙𝑜1,𝜂 , 𝜙𝑜2,𝜉 , 𝜙𝑜2,𝜂 )= (𝐷11 + 2𝐷12 tan2 𝜓 + 4𝐷66 tan2 𝜓
𝑎4
4. Conclusion 4
+𝐷22 tan4 𝜓)(𝑤𝑜,𝜉𝜉 )2 − (𝐷12 + 2𝐷66 + 𝐷22 tan2 𝜓 ) tan 𝜓 sec 𝜓 𝑤𝑜,𝜉𝜉 𝑤𝑜,𝜉𝜂
𝑎3 𝑏
In this paper, the dynamic instability of functionally graded skew 2 4
+ (𝐷12 +𝐷22 tan2 𝜓 )sec2 𝜓 𝑤𝑜,𝜉𝜉 𝑤𝑜,𝜂𝜂 + (𝐷66 +𝐷22 tan2 𝜓 )sec2 𝜓 (𝑤𝑜,𝜉𝜂 )2
plate subjected to uniform and linearly varying in-plane periodic load- 𝑎2 𝑏2 𝑎2 𝑏2

50
R. Kumar et al. International Journal of Mechanical Sciences 130 (2017) 41–51

4 1 1
− 𝐷22 tan 𝜓 sec3 𝜓 𝑤𝑜,𝜉𝜂 𝑤𝑜,𝜂𝜂 + 𝐷22 sec4 𝜓 (𝑤𝑜,𝜂𝜂 )2 − (𝐸11 + 𝐸12 tan2 𝜓 [3] Birman V. Buckling of functionally graded hybrid composite plates. In: Proceedings
𝑎𝑏3 𝑏4 𝑎3 of 10th conference on engineering mechanics, 2; 1995. p. 1199–202.
[4] Feldman E, Aboudi J. Buckling analysis of functionally graded plates subjected to
2 1
+2𝐸66 tan2 𝜓 )𝑤𝑜,𝜉𝜉 𝜙𝑜1,𝜉 + 𝐸66 tan 𝜓 sec 𝜓 𝑤𝑜,𝜉𝜉 𝜙𝑜1,𝜂 + (𝐸12 + 2𝐸66 uniaxial loading. Compos Struct 1997;38(1):29–36.
𝑎2 𝑏 𝑎3 [5] Alipour M, Shariyat M. Semianalytical solution for buckling analysis of variable
1 thickness two-directional functionally graded circular plates with nonuniform elastic
+𝐸22 tan2 𝜓 ) tan 𝜓 𝑤𝑜,𝜉𝜉 𝜙𝑜2,𝜉 − (𝐸12 + 𝐸22 tan2 𝜓 ) sec 𝜓 𝑤𝑜,𝜉𝜉 𝜙𝑜2,𝜂
𝑎2 𝑏 foundations. J Eng Mech 2013;139(5):664–76.
[6] Jabbari M, Mojahedin A, Khorshidvand A, Eslami M. Buckling analysis of a func-
2 2
+ (𝐸12 + 𝐸66 ) tan 𝜓 sec 𝜓 𝑤𝑜,𝜉𝜂 𝜙𝑜1,𝜉 − 𝐸66 sec2 𝜓 𝑤𝑜,𝜉𝜂 𝜙𝑜1,𝜂 tionally graded thin circular plate made of saturated porous materials. J Eng Mech
𝑎2 𝑏 𝑎𝑏2 2014;140(2):287–95.
2 2 [7] Miranda E, Akkar SD. Dynamic instability of simple structural systems. J Struct Eng
− (𝐸66 + 𝐸22 tan2 𝜓 ) sec 𝜓 𝑤𝑜,𝜉𝜂 𝜙𝑜2,𝜉 + 𝐸22 tan 𝜓 sec2 𝜓 𝑤𝑜,𝜉𝜂 𝜙𝑜2,𝜂 2003;129(12):1722–6.
𝑎2 𝑏 𝑎𝑏2
[8] Patel SN, Datta PK, Sheikh AH. Parametric study on the dynamic instability behavior
1 1 1
− 𝐸12 sec2 𝜓 𝑤𝑜,𝜂𝜂 𝜙𝑜1,𝜉 + 𝐸22 tan 𝜓 sec2 𝜓 𝑤𝑜,𝜂𝜂 𝜙𝑜2,𝜉 − 𝐸22 sec3 𝜓 𝑤𝑜,𝜂𝜂 𝜙𝑜2,𝜂 of laminated composite stiffened plate. J Aerosp Eng 2009;135(11):1331–41.
𝑎𝑏2 𝑎𝑏2 𝑏3 [9] Kumar A, Panda SK, Kumar R. Dynamic instability of composite skew
(A.2) plates using boundary characteristic orthogonal polynomials. Int J Appl Mech
2014;6(06):1450078.
[10] Sahoo R, Singh BN. Dynamic instability of laminated composite and sandwich plates
using a new inverse hyperbolic zigzag theory. J Aerosp Eng 2015;28(4):04014109.
1
ℜ3 (𝑤𝑜,𝜉𝜉 , 𝑤𝑜,𝜉𝜂 , 𝑤𝑜,𝜂𝜂 , 𝜙𝑜1,𝜉 , 𝜙𝑜1,𝜂 , 𝜙𝑜2,𝜉 , 𝜙𝑜2,𝜂 ) = −
(𝐸11 + 𝐸12 tan2 𝜓 [11] Kumar R, Kumar A, Panda SK. Parametric resonance of composite skew plate under
𝑎3 non-uniform in-plane loading. Struct Eng Mech 2015;55(2):435–59.
2 1 [12] Jha DK, Kant T, Singh RK. A critical review of recent research on functionally graded
+2𝐸66 tan2 𝜓 )𝑤𝑜,𝜉𝜉 𝜙𝑜1,𝜉 + 𝐸66 tan 𝜓 sec 𝜓 𝑤𝑜,𝜉𝜉 𝜙𝑜1,𝜂 + (𝐸12 plates. Compos Struct 2013;96:833–49.
𝑎2 𝑏 𝑎3
[13] Ng TY, Lam KY, Liew KM. Effects of FGM materials on the parametric resonance of
1
+2𝐸66 + 𝐸22 tan2 𝜓 ) tan 𝜓 𝑤𝑜,𝜉𝜉 𝜙𝑜2,𝜉 − (𝐸12 + 𝐸22 tan2 𝜓 ) sec 𝜓 𝑤𝑜,𝜉𝜉 𝜙𝑜2,𝜂 plate structures. Comput Methods Appl Mech Eng 2000;190(8):953–62.
𝑎2 𝑏 [14] Ng TY, Lam KY, Liew KM, Reddy JN. Dynamic stability analysis of function-
2 2 ally graded cylindrical shells under periodic axial loading. Int J Solids Struct
+ (𝐸12 + 𝐸66 ) tan 𝜓 sec 𝜓 𝑤𝑜,𝜉𝜂 𝜙𝑜1,𝜉 − 𝐸66 sec2 𝜓 𝑤𝑜,𝜉𝜂 𝜙𝑜1,𝜂 2001;38(3):1295–309.
𝑎2 𝑏 𝑎𝑏2 [15] Shen HS. Nonlinear bending response of functionally graded plates subjected to
2 2
− (𝐸66 + 𝐸22 tan2 𝜓 ) sec 𝜓 𝑤𝑜,𝜉𝜂 𝜙𝑜2,𝜉 + 𝐸22 tan 𝜓 sec2 𝜓 𝑤𝑜,𝜉𝜂 𝜙𝑜2,𝜂 transverse loads and in thermal environments. Int J Mech Sci 2002;44:561–84.
𝑎2 𝑏 𝑎𝑏2 [16] Yang J, Shen HS. Free vibration and parametric resonance of shear deformable func-
1 1 1 tionally graded cylindrical panels. J Sound Vib 2003;261(5):871–93.
− 𝐸12 sec2 𝜓 𝑤𝑜,𝜂𝜂 𝜙𝑜1,𝜉 + 𝐸22 tan 𝜓 sec2 𝜓 𝑤𝑜,𝜂𝜂 𝜙𝑜2,𝜉 − 𝐸22 sec3 𝜓 𝑤𝑜,𝜂𝜂 𝜙𝑜2,𝜂 [17] Lanhe W, Hongjun W, Daobin W. Dynamic stability analysis of FGM plates
𝑎𝑏2 𝑎𝑏2 𝑏3 by the moving least squares differential quadrature method. Compos Struct
1 2 1
+ (𝐹11 +𝐹66 tan2 𝜓 )(𝜙𝑜1,𝜉 )2 − 𝐹66 tan 𝜓 sec 𝜓 𝜙𝑜1,𝜉 𝜙𝑜1,𝜂 + 𝐹66 sec2 𝜓 (𝜙𝑜1,𝜂 )2 2007;77:383–94.
𝑎2 𝑎𝑏 𝑏2 [18] Sundararajan1 N, Ganapathi M. Dynamic thermal buckling of functionally graded
spherical caps. J Eng Mech 2008;134(2):206–9.
2 2 2
− (𝐹11 + 𝐹66 ) tan 𝜓 𝜙𝑜1,𝜉 𝜙𝑜2,𝜉 + 𝐹12 sec 𝜓 𝜙𝑜1,𝜉 𝜙𝑜2,𝜂 + 𝐹66 sec 𝜓 𝜙𝑜1,𝜂 𝜙𝑜2,𝜉 [19] Ganapathi M, Prakash T. Thermal buckling of simply supported functionally graded
𝑎2 𝑎𝑏 𝑎𝑏 skew plates. Compos Struct 2006;74(2):247–50.
1 2 1 [20] Pradyumna S, Bandyopadhyay JN. Dynamic instability of functionally graded shells
+ (𝐹22 tan2 𝜓 +𝐴66 )(𝜙𝑜2,𝜉 )2 − 𝐹22 tan 𝜓 sec 𝜓 𝜙𝑜2,𝜉 𝜙𝑜2,𝜂 + 𝐹22 sec2 𝜓 (𝜙𝑜2,𝜂 )2
𝑎2 𝑎𝑏 𝑏2 using higher-order theory. J Eng Mech 2009;136(5):551–61.
[21] Torki ME, Kazemi MT, Reddy JN, Haddadpoud H, Mahmoudkhani S. Dynamic sta-
(A.3) bility of functionally graded cantilever cylindrical shells under distributed axial fol-
lower forces. J Sound Vib 2014;333(3):801–17.
ℜ4 (𝜙𝑜1 , 𝜙𝑜2 ) = 𝐻44 (𝜙𝑜2 )2 + 𝐻55 (𝜙𝑜1 )2 (A.4) [22] Asnafi A, Abedi MA. Complete analogical study on the dynamic stability analysis of
isotropic functionally graded plates subjected to lateral stochastic loads. Acta Mech
2015;226(7):2347–63.
[23] Sofiyev AH, Kuruoglu N. Domains of dynamic instability of FGM conical shells under
ℜ5 (𝑛𝑥𝑥 , 𝑛𝑥𝑦 , 𝑛𝑦𝑦 , 𝑤𝑜,𝜉 , 𝑤𝑜,𝜂 ) time dependent periodic loads. Compos Struct 2016;136:139–48.
( ) [24] Reddy JN. A simple higher-order theory for laminated composite plates. J Appl Mech
𝑛𝑥𝑥 2𝑛𝑥𝑦 𝑛𝑦𝑦 1984;51:745–52.
= − tan 𝜓 + tan2 𝜓 (𝑢𝑜,𝜉2 + 𝑣𝑜,𝜉2 + 𝑤𝑜,𝜉2 )
𝑎2 𝑎𝑏 𝑏2 [25] Bolotin VV. The dynamic stability of elastic systems. SanFrancisco: Holden-Day Pub-
( ) lication; 1964.
2𝑛𝑥𝑦 2𝑛𝑦𝑦 [26] Morimoto T, Tanigawa Y, Kawamura R. Thermal buckling of functionally graded
+ sec 𝜓 − tan 𝜓 sec 𝜓 (𝑢𝑜,𝜉 𝑢𝑜,𝜂 + 𝑣𝑜,𝜉 𝑣𝑜,𝜂 + 𝑤𝑜,𝜉 𝑤𝑜,𝜂 )
𝑎𝑏 𝑏2 rectangular plates subjected to partial heating. Int J Mech Sci 2006;48(9):926–37.
[27] Zhang DG, Zhou YH. A theoretical analysis of FGM thin plates based on physical
𝑛𝑦𝑦 2
+ sec 𝜓(𝑢𝑜,𝜂2 + 𝑣𝑜,𝜂2 + 𝑤𝑜,𝜂2 ) (A.5) neutral surface. Comput Mater Sci 2008;44(2):716–20.
𝑏2 [28] Soldatos KP. A refined laminated plate and shell theory with applications. J Sound
Vib 1991;144(1):109–29.
[29] Singh B, Chakraverty S. Flexural vibration of skew plates using boundary character-
ℜ6 (𝑢𝑜,𝑡 , 𝑣𝑜,𝑡 , 𝑤𝑜,𝑡 , 𝜙𝑜1,𝑡 , 𝜙𝑜2,𝑡 ) = 𝜌1 [(𝑢𝑜,𝑡 )2 + (𝑣𝑜,𝑡 )2 + (𝑤𝑜,𝑡 )2 ] + 𝜌3 [(𝜙𝑜1,𝑡 )2 + (𝜙𝑜2,𝑡 )2 ] istic orthogonal polynomials in two variables. J Sound Vib 1994;173(2):157–78.
[30] Thai HT, Choi DH. An efficient and simple refined theory for buckling analysis of
(A.6) functionally graded plates. Appl Math Model 2012;36(3):1008–22.
[31] Wang S. Buckling analysis of skew fibre-reinforced composite laminates based on
References first-order shear deformation theory. Compos Struct 1997;37:5–19.
[32] Ganapathi M, Prakash T, Sundararajan N. Influence of functionally graded material
on buckling of skew plates under mechanical loads. J Eng Mech 2006;132:902–5.
[1] Chi SH, Chung YL. Mechanical behavior of functionally graded material plates under [33] Hutt JM, Salam AE. Dynamic instability of plates by finite element. J Eng Mech
transverse load—part I: analysis. Int J Solids Struct 2006;43(13):3657–74. 1971;3:879–99.
[2] 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.

51