Professional Documents
Culture Documents
Accepted Manuscript
Accepted Manuscript
PII: S0263-8223(13)00154-2
DOI: http://dx.doi.org/10.1016/j.compstruct.2013.04.002
Reference: COST 5101
Please cite this article as: Thai, C.H., Ferreira, A.J.M., Carrera, E., Nguyen-Xuan, H., Isogeometric analysis of
laminated composite and sandwich plates using a layerwise deformation theory, Composite Structures (2013), doi:
http://dx.doi.org/10.1016/j.compstruct.2013.04.002
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Isogeometric analysis of laminated composite and sandwich plates using a
layerwise deformation theory
Abstract
We present an isogeometric finite element formulation for static, free vibration and buckling analysis
of laminated composite and sandwich plates. The idea behind this work is to associate an isogeometric
analysis (IGA) with a layerwise theory [A.J.M. Ferreira. Analyis of composite plates using a layerwise
deformation theory and multiquadrics discretization. Mech Adv Mater Struct 2005;12(2):99-112]. Isogeo-
metric analysis based on non-uniform rational B-spline(NURBS) basic functions were recently proposed to
preserve exact geometries and to enhance very significantly the accuracy of the traditional finite elements.
B-splines basic function (or NURBS) is used to represent for both geometric and field variable approxima-
tions, which provide a flexible way to make refinement and degree elevation. They enable us to achieve
easily the smoothness with arbitrary continuity order compared with the traditional FEM. The layerwise
theory assumes a first-order shear deformation theory in each layer and the imposition of displacement con-
tinuity at the layers interfaces. This permits to remove shear correction factors and improves the accuracy
of transverse shear stresses. Intensive numerical studies have been conducted to show the highly efficient
performance of the proposed formulation.
Keywords: Isogeometric analysis, laminated composite and sandwich plates, layerwise theory
1. Introduction
Composite and sandwich structures have extensively been used in various engineering disciplines such
as aerospace engineering, automotive engineering, civil engineering, etc. Laminated composite structures
are often made of several orthotropic layers with different materials stacked together to achieve superior
properties such as high stiffness and strength-to-weight ratios, long fatigue life, wear resistance, lightweight,
etc[1, 2]. On the other hand, the sandwich structures are a special case of laminated composite structures
in which the difference of material properties between core and face sheets are very large. To use them
effectiveness, a good understanding of the bending behavior, stresses distribution, dynamic response and
classical laminate plate theory (CLPT) [3] is only suitable for thin plates, which the shear deformation effect
is neglected. The first-order shear deformation theory (FSDT) [4] can be applied for both moderately thick
and thin plates but the accuracy of solutions strongly depends on shear correction factors. To bypass the
limitations of FSDT, the higher-order shear deformation theories (HSDT) [5, 6, 7] have been developed.
These theories disregard shear correction factors and yield highly accurate solutions for displacements and
transverse shear stresses for thick plates. Classically, first-order and higher-order theories use the equivalent
single-layer models (ESL), which consider the same degrees of freedom for all laminate layers. However,
in the practical application of sandwich plates, the difference of strength between core and face sheets
are often large. These theories can face certain drawbacks to predict accurately the bending behavior and
dynamic response. In fact, most of them do not correctly represent the transverse shear stresses and high
frequencies. Alternatively, layerwise theories [8, 9, 10], regarding independent degrees of freedom for
each layer, have been devised. Herein, the generalized layerwise model proposed by Reddy [9] is perhaps
the most popular. As the simplified Reddy’s model, a layerwise displacement model reported in [11] is
adopted in this study. The proposed model assumes a first-order shear deformation theory in each layer
and the imposition of displacement continuity at the layers interfaces. In addition, several other equivalent-
single-layer models for laminated plates have been proposed accounting for zig-zag effects and fulfillment
of interlaminar continuity. Among these the one by Mau [12], Chou and Carleone [13], Di Sciuva [14],
Toledano and Murakami [15] and Ren [16] are herein mentioned. Mixed layer-wise and equivalent-single-
layer theories based on Reissener Mixed Variational Theorem have been discussed by Carrera [17, 18, 19].
A historical review encompassing early and recent developments of advanced theories for laminated beams,
plates and shells was revisited in [20]. Interested readers are addressed to that last paper for a more complete
Nowadays the finite element method has become a very powerful and reliable tool for engineering and
science problems. In the FEM, a discretized geometry obtained through the so-called meshing process is
needed. This process often leads to geometrical errors even using the higher-order FEM. Also, the com-
munication of the geometry model and the mesh generation during an analysis process that aims to provide
the desired accuracy for the solution is required and this constitutes a time-consuming part in the overall
analysis-design process, especially for industrial problems [21]. To overcome this disadvantage, Hughes
2
et al. [21] have recently proposed a new computational method which is coined Isogeometric Analysis
(IGA) to closely link the gap between Computer Aided Design (CAD) and Finite Element Analysis (FEA).
It means that the IGA uses the same basis functions to describe both the geometry of the computational
domain and the approximate solution. Being different from the Lagrange polynomials of the standard FEM,
isogeometric approach utilizes more general basis functions such as B-splines and Non-Uniform Rational
B-splines (NURBS) that are common in CAD geometry. The exact geometry is therefore maintained at
the coarsest level of discretization and re-meshing is performed on this coarsest level without any further
communication with CAD geometry. Furthermore, B-splines (or NURBS) provide a flexible way to perform
refinement (or h-refinement), and degree elevation [22]. They allow us to achieve easily a smoothness of
arbitrary continuity order in comparison with the FEM. Our present work will thus exploit the advantage of
high continuity order of the IGA in the composite sandwich plate analysis based on the HSDT. Isogeometric
analysis has been applied to a wide range of practical mechanics problems such as structural vibrations [23],
linear and non-linear elasticity and plasticity [24], and structural shape optimization [25], Kirchhoff-Love
shell [26, 27, 28], Reissner-Mindlin plate/shell [29, 30], laminated composite plates based on HSDT [31],
rotation-free shells [32] and further developments [33, 34], to mention only a few.
In this paper, a novel numerical approach using a NURBS-based isogeometric approach associated with
the layerwise deformation theory is formulated for static, free vibration and buckling analysis of laminated
composite and sandwich plate structures. A plate model based on the layerwise approach can produce more
accurate solution than the FSDT and HSDT in transverse shear stresses. It is therefore very suitable for
the analysis of laminated composite and sandwich plates. The NURBS basis functions are used in both ge-
ometry representation and analysis. Several numerical examples are illustrated to show the effectiveness of
the present method. It is found that obtained results match well with the exact three-dimensional elasticity,
The paper is arranged as follows: a brief on the B-spline and NURBS surface is described in section
2. Section 3 presents a formulation of a NURBS-based isogeometric analysis for laminated composite and
sandwich plates. Several numerical examples are provided in section 4. Finally we close our paper with
3
2. A brief of NURBS functions and surfaces
Let Ξ = [ξ1 , ξ2 , ..., ξn+p+1 ] be a nondecreasing sequence of parameter values, ξi ≤ ξi+1 , i = 1, ..., n + p.
The ξi are called knots, and Ξ is the set of coordinates in the parametric space. If all knots are equally spaced
the knot vector is called uniform. If the first and the last knots are repeated p + 1 times, the knots vector
is described as open. A B-spline basis function is C∞ continuous inside a knot span and C p−1 continuous
at a single knot. A knot value can appear more than once and is then called a multiple knot. At a knot of
multiplicity k the continuity is C p−k . Given a knot vector, the B-spline basis functions Ni,p (ξ ) of order p =
1 ξi ≤ ξ < ξi+1
Ni,0 (ξ ) = (1)
0
otherwise
The basis functions of order p > 0 is defined by the following recursion formula [35]
ξ − ξi ξi+p+1 − ξ
Ni,p (ξ ) = Ni,p−1 (ξ ) + Ni+1,p−1 (ξ ) with p = (1, 2, 3, ...) (2)
ξi+p − ξi ξi+p+1 − ξi+1
For p = 0 and 1, the basis functions of isogeometric analysis are identical to those of standard piecewise
constant and linear finite elements, respectively. In IGA, basis functions with p ≥ 2 are emphasized [21].
Fig. 1, Fig. 2 and Fig. 3 illustrate a set of one-dimensional and two-dimensional quadratic, cubic and quartic
B-spline basis functions for open uniform knot vectors Ξ = {0, 0, 0, 12 , 1, 1, 1}, Ξ = {0, 0, 0, 0, 12 , 1, 1, 1, 1}
where Pi are the control points, n denotes the number of control points and Ni,p (ξ ) is the pth -degree B-spline
4
1 1
0.9 0.9
0.8 1 0.8
0.5 0.5
0.4
0.4 0.4
0.2
0.3 0.3
0
0.2 1
0.2
1
0.1
0.5 0.1
0.5
0
0 1/2 1 0
0 0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 1/2 1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 1/2 1
5
Given two knot vectors Ξ = {ξ1 , ξ2 , ..., ξn+p+1 } and H = η1 , η2 , ..., ηm+q+1 and a control net Pi, j , a
n m
S (ξ , η) = ∑ ∑ Ni,p (ξ ) M j,q (η) Pi, j (4)
i=1 j=1
where Ni,p (ξ ) and M j,q (η) are the B-spline basis functions defined on the knot vectors Ξ and H , respec-
tively.
In a finite element context, we identify the logical coordinates (i, j) of the B-spline surface with the
n×m
S (ξ , η) = ∑ NIb (ξ , η) PI (5)
I
where NIb (ξ , η) = Ni,p (ξ ) M j,q (η) is the shape function associated with a node I. The superscript b indicates
Non-uniform rational B-splines (NURBS) are obtained by augmenting every point in the control mesh
n×m
wg (ξ , η) = ∑ NIb (ξ , η) wI (6)
I=1
n×m
∑ NIb (ξ , η) wI PI n×m
I=1
S (ξ , η) = = ∑ NI (ξ , η)PI (7)
wg (ξ , η) I=1
where NI (ξ , η) = NIb (ξ , η) wI /wg (ξ , η) are NURBS basis functions. An example of a quadratic NURBS
3. An isogeometric formulation for laminated composite and sandwich plates using layerwise theory
In this paper, a layerwise displacement model [11, 36] is adopted to analyze laminated composite and
sandwich plates. The proposed model is assumed to use a first-order shear deformation theory in each
6
Figure 4: NURBS surface and control mesh.
layer and the imposition of displacement continuity at the layers interfaces. For simplification, a three-layer
laminate plate is chosen to construct the present formulation, as shown in Fig. 5. However, the present
(2)
u(2) (x, y, z) = u0 (x, y) + z(2) βx (x, y)
(2)
v(2) (x, y, z) = v0 (x, y) + z(2) βy (x, y) (8)
7
(2) (2)
where u, v and w are the displacement components at any point (x, y, z); u0 , v0 , w0 , βx and βy denote the
displacement components in the x, y, z directions and the rotations in the y − z and x − z planes, respectively,
From Fig. 5, the corresponding displacement field for the upper layer (layer 3) is given as
And the corresponding displacement field for the lower layer (layer 1) is proposed as
where hk and z(k) ∈ [−hk /2, hk /2] are the kth -layer thickness and z coordinates, respectively.
The in-plane and shear components of strain for layer k are given by
∂ u(k)
εxx(k)
∂x
∂ v(k)
ε (k)
yy
∂y
(k) ∂ u(k) (k)
εxy
=
∂y + ∂∂v x
(11)
∂ u(k) (k)
εxz(k) + ∂ ∂wx
∂z
(k) ∂ v(k) (k)
+ ∂ ∂wy
εyz ∂z
T
(k) (k) (k) (k)
The in-plane strain vector εp = εxx εyy γxy can be rewritten as
ε (k)
ε m(k)
ε mb(k)
ε b(k)
xx xx xx xx
εyy(k) = εyym(k) + ε mb(k) + z(k) εyyb(k) (12)
yy
(k)
m(k) mb(k)
b(k)
γxy γxy
γxy
γxy
8
T
and the transverse shear strain vector γ (k) = γ (k) γ (k) has the form:
xz yz
(k)
γ (k)
βx(k) + ∂ w0
xz ∂x
= (k) (13)
(k)
γyz βy(k) + ∂ w0
∂y
∂ u0
ε m(k)
∂x
xx
∂ v0
εyym(k)
=
∂x
(14)
m(k)
∂ u0
+ ∂∂vx0
εxy
∂y
(1) (2)
ε bm(1)
− h21 ∂ β∂ xx − h22 ∂ β∂ xx
xx
(1) (2)
− h21 h2 ∂ βy
∂ βy
εyybm(1)
=
∂ y − 2 ∂ y
(16)
(1) (1) (2) (2)
h
− h22
bm(1)
∂ βx ∂ βy ∂ βx ∂ βy
−
∂y + ∂x +
εxy
1
2 ∂y ∂x
ε bm(2)
0
xx
εyybm(2)
=
0 (17)
bm(2)
εxy
0
(2) (3)
ε bm(3)
h2 ∂ βx
2 ∂x + h23 ∂ β∂ xx
xx
(2) (3)
h2 ∂ βy h3 ∂ βy
εyybm(3)
=
2 ∂ y + 2 ∂ y
(18)
(2) (2) (3) (3)
h2
+ h23
bm(3)
∂ βx ∂ βy ∂ βx ∂ βy
εxy
2 ∂y + ∂x ∂y + ∂x
9
The constitutive equation of an orthotropic layer in the local coordinate system is derived from Hookes law
where subscripts 1 and 2 are the directions of the fiber and in-plane normal to fiber, respectively; subscript
(k)
3 indicates the direction normal to the plate; and the reduced stiffness components, Qi j , are given by
The laminate is usually made of several orthotropic layers. Each layer must be transformed into the laminate
(k) (k) (k) (k) (k)
σxx
Q̄11 0Q̄120 Q̄16
εxx
(k) (k)
Q̄21 Q̄(k) (k) (k)
σyy
22 Q̄26 0 0
εyy
(k)
(k) (k) (k) (k)
τxy =
Q̄61 Q̄62 Q̄66 0 0
γxy (21)
(k) (k) (k) (k)
0 0 0 Q̄55 Q̄54
τxz
γxz
(k)
(k) (k) (k)
0 0 0 Q̄45 Q̄44
τyz γyz
(k)
where Q̄i j is transformed material constant (see [2] for more details).
Similar to the HSDT model, the present approach does not require the use of shear correction factors.
Here, only symmetric laminated composite plates are considered. So u0 , v0 and related stress resultants can
be ignored.
10
The in-plane strain in Eq. (12) can be rewritten as
ε (k)
ε mb(k)
ε b(k)
xx xx xx
εyy(k) = εyymb(k) + z(k) εyyb(k) (22)
(k)
mb(k) b(k)
γxy γxy
γxy
A weak form of the static model for laminated composite and sandwich plates using the layerwise theory
can be formulated as
! !
Z 3 T Z 3 T Z
(k) (k) (k) (k) (k) (k)
∑ p
δ ε D̄ ε p dΩ + ∑ δ γ C γ dΩ = δ w0 p̄dΩ (23)
k=1 k=1
Ω Ω Ω
T
(k)
where εp = ε mb(k) ε b(k) and γ (k) are in-plane strains and transverse shear strains vector, respec-
tively, and
A(k) B(k)
D̄(k) = ;
B(k) D(k)
h(k) /2
(k) (k) (k)
1, z, z2 Q̄kij dz
R
Ai j , Bi j , Di j = (i, j = 1, 2, 6)
−h(k) /2
h(k) /2
(k)
(1)Q̄kij dz
R
Ci j = (i, j = 4, 5)
−h(k) /2
For the free vibration analysis of laminated composite and sandwich plates using the layerwise theory,
a weak form may be derived from the dynamic form of the principle of virtual work
! ! !
Z 3
T Z 3
T Z 3 T
(k) (k) (k) (k) (k)
∑ δ ε p D̄ε p dΩ + ∑δ γ Cγ dΩ = ∑ δ ũ(k) mũ¨ dΩ (24)
k=1 k=1 k=1
Ω Ω Ω
where
(k) (k)
I0 I1
(k) (k) (k)
R h /2
m(k) = I0 , I1 , I2 = −h(k) /2 ρ (k) 1, z, z2 dz
;
(k) (k) (k)
I1 I2
T
in which ũ(k) = (k)
u1
(k)
u2 and
11
− h21 βx1 − h22 βx2 1
β 0
x
u11 = − h21 βy1 − h22 βy2 ; u12 = β 1
y
; 2
u1 = 0
w0
0
w0
2 h2 2 h3 3 3
2 βx + 2 βx
βx
βx
2
u2 = 2 ; u31 = h2 2 h3 3 ; 3
u2 = 3
βy 2 βy + 2 βy βy
0
w0
0
In the case of in-plane buckling analysis and assuming pre-buckling stresses σ̂ 0 , nonlinear strains are
σx0 0
τxy
where ∇T = [∂ /∂ x ∂ /∂ y] and σ̂ 0 = are the gradient operator and in-plane pre-buckling
0
τxy σy0
stresses, respectively.
From Eq. (8), Eq. (9) and Eq. (10), the total displacement fields can be rewritten as
3
u= ∑ u(k) (26)
k=1
T
u= w βx1 βy1 βx2 βy2 βx3 βy3 (27)
Using NURBS basis functions, approximate variables are the transverse deflection and the rotations at all
12
control points, which can be expressed as
w
NI 0 0 0 0 0 0
wI
1
βx
0 NI 0 0 0 0 0 1
βxI
1 1
β 0 0 NI 0 0 0 0 βyI
y
n×m n×m
u= β 2 = ∑
0 0 0 NI 0 0 0
2
βxI = ∑ NI qI (28)
x
I=1 I=1
2
βy
0 0 0 0 NI 0 0 βyI2
3 3
0 0 0 0 0 NI 0
β
x
βxI
β3
0 0 0 0 0 0 NI
3
βyI
y
T
where n × m is the number basis functions, NI and qI = wI βxI1 βyI1 βxI2 βyI2 βxI3 βyI3 are ratio-
nal basis functions and the degrees of freedom of u associated with a control point I, respectively.
T n×m T n×m
εp γ = ∑ Bmb1
I Bmb2
I Bmb3
I Bb1
I Bb2
I Bb3
I Bs1
I Bs2
I Bs3
I
qI = ∑ BI qI (29)
I=1 I=1
where
0 − h21 NI,x
0 0− h22 NI,x
0 0
Bmb1
I =
0 0 − h1
2 NI,y 0 − h2
2 NI,y 0 0
h1 h1 h2 h2
0 − 2 NI,y − 2 NI,x − 2 NI,y − 2 NI,x 0 0
h2 h3
0 0 0 0 0 0 0 0 0 0 2 NI,x 0 2 NI,x 0
Bmb2
I = mb3
0 0 0 0 0 0 0 ; BI = 0 0 0 0 h2
2 NI,y 0 h3
2 NI,y
h2 h2 h3 h3
0 0 0 0 0 0 0 0 0 0 2 NI,y 2 NI,x 2 NI,y 2 NI,x
0 NI,x 0 0 0 0 0 0 0 0 NI,x 0 0 0
b1
b2
BI = 0 0 NI,y 0 0 0 0 ; BI = 0 0 0 0 NI,y 0 0
0 NI,y NI,x 0 0 0 0 0 0 0 NI,y NI,x 0 0
0 0 0 0 0 NI,x 0
s1 NI,x NI 0 0 0 0 0
Bb3
I = 0 0 0 0 0
0 NI,y ; B =
I
NI,y 0 NI 0 0 0 0
0 0 0 0 0 NI,y NI,x
13
NI,x 0 0 NI 0 0 0 s3 NI,x 0 0 0 0 NI 0
Bs2
I = ; BI =
NI,y 0 0 0 NI 0 0 NI,y 0 0 0 0 0 NI
T
BI = Bmb1
I Bmb2
I Bmb3
I Bb1
I Bb2
I Bb3
I Bs1
I Bs2
I Bs3
I
where
NI,x 0 0 0 0 0 0
BgI =
NI,y 0 0 0 0 0 0
For static analysis, the formulation of laminated composite and sandwich plates based on the layerwise
theory is obtained as
Kq = f, (31)
K − ω 2 M q = 0,
(32)
(K − λcr Kg ) q = 0 (33)
T
3 mb(k) (k) (k)
B Bmb(k)
B A
Z
T
+ Bs(k) C(k) Bs(k) dΩ (34)
∑
k=1 Bb(k) B(k) D(k) Bb(k)
Ω
Z
f= p̄NdΩ (35)
Ω
in which f is the global force vector and M is the global mass matrix
T
3 (k) (k) (k) (k)
N1 I0 I1 N1
Z
M= dΩ (36)
∑
(k)
k=1 N
(k) (k) N(k)
I1 I2
Ω 2 2
where
14
0 − h21 N 0 − h22 N 0 0 0 0 N 0 0 0 0 0
N11 =
0 0 − h21 N 0 − h22 N 0 0
1
; N2 = 0 0 N 0 0 0 0
N 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 N 0 0 0
N21 =
0 0 0 0 0 0 0
2
; N = 0 0 0 0 N 0 0
2
N 0 0 0 0 0 0 0 0 0 0 0 0 0
h2 h3
0 0 0 2N 0 2N 0 0 0 0 0 0 N 0
N31 =
0 0 0 0 h2
2N 0 h3 ; N3 = 0 0 0 0 0 0 N
2N 2
N 0 0 0 0 0 0 0 0 0 0 0 0 0
Z
Kg = (Bg )T τBg dΩ ; τ = hσ̂ 0 (37)
Ω
in which ρ, h, ω and λcr are the mass density, the thickness, the natural frequency and the critical buckling
load, respectively.
4. Numerical results
In this section, several examples of the laminated composite and sandwich plates for static, free vibra-
tion and buckling analysis are studied. Results obtained are compared with other published ones. For all
numerical examples, quadratic, cubic and quartic NURBS elements integrated with (p + 1) × (q + 1) Gauss
• Material I:
Face sheets:
15
Core:
E1 = 0.04E, E2 = 0.04E, E3 = 0.5E, G12 = 0.016E, G13 = G23 = 0.06, ν12 = 0.25.
• Material IV : [39]
We consider a simply supported sandwich square plate proposed by Srinivas [8] subjected to a uniform
transverse load q, as shown in Fig. 6. The length to thickness ratio (a/h = 10) is given. The core thickness
hc to plate thickness ratio h is fixed at 0.8 (hc /h = 0.8). The laminate sandwich plate is made of one inner
Q̄skin = RQ̄core
a a 1 = σ 1 ( a , a , h )/q,
w̄ = 0.999781w( , , 0)/hq, σ̄xx xx
2 2 2 2 2
2 1 a a 2h 3 2 a a 2h
σ̄xx = σxx ( , , )/q, σ̄xx = σxx ( , , )/q,
2 2 5 2 2 5
1 1 a a h 2 1 a a 2h 3 = σ 2 ( a , a , 2h )/q.
σ̄yy = σyy ( , , )/q, σ̄yy = σyy ( , , )/q, σ̄yy yy
2 2 2 2 2 5 2 2 5
16
Figure 6: Geometry of a sandwich plate.
Table 1: Convergence of the normalized displacement and stresses of a three-layer sandwich square plate laminated plate (a/h = 10)
Method Normalized Meshes
9×9 13 × 13 17 × 17
Quadratic w̄ 258.8287 (0.0546%) 258.8324 (0.0531%) 258.8338 (0.0526%)
Cubic 258.8327 (0.0530%) 258.8342 (0.0524%) 258.8346 (0.0523%)
Quartic 258.8330 (0.0529%) 258.8347 (0.0522%) 258.8347 (0.0522%)
Exact [8] - - 258.97
Quadratic 1
σ̄xx 60.1275 (0.3736 %) 60.1907 (0.2689%) 60.2163 (0.2265%)
Cubic 60.0971 (0.4240 %) 60.1803 (0.2861%) 60.2115 (0.2345 %)
Quartic 60.2528 (0.1660 %) 60.2530 (0.1657 %) 60.2535 (0.1649%)
Exact [8] - - 60.353
Quadratic 1
σ̄yy 38.4479 (0.1120%) 38.4703 (0.0538%) 38.4798 ( 0.0291%)
Cubic 38.4055 (0.2221 %) 38.4526 (0.0998 %) 38.4706 ( 0.0530%)
Quartic 38.4940 (0.0078 %) 38.4944 (0.0088 %) 38.4947 ( 0.0096%)
Exact [8] - - 38.491
17
We first investigate the convergence of the normalization deflection and stresses of the present method
at R = 5. The plate is modeled with 9 × 9, 13 × 13 and 17 × 17 elements as shown in Fig. 7. The exact
solution of this problem was given by Srinivas [8]. Table 1 displays the convergence of the normalization
deflection and stresses. The relative error percentages compared with the exact solutions [8] are given in a
parentheses. It be can observed that obtained results agree very well with the exact solution.
For a comparison, we calculate the normalized displacement and stresses of the sandwich square plate using
17 × 17 B-spline elements. Obtained results from the present method are compared with those of the FEM
based on HSDT reported by Pandya and Kant [7], the multiquadric radial basis function method (MRBF)
based on the finite point formulation and HSDT by Ferreira et al. [40], the multiquadric radial basis function
method (MRBF) relied on the layerwise deformation theory by Ferreira [11], the multiquadric radial basis
function pseudospectral method (MRBF-PS) based on the layerwise deformation theory by Ferreira et al.
[36], the closed form solution (CFS) based on the inverse hyperbolic shear deformation theory (IHSDT) by
Grover et al. [41], the closed form solution based on the trigonometric shear deformation theory (TrSDT)
by Mantari et al.[42] and the exact solution by Srinivas [8]. The results with respect to various values of
R (R = 5, 10, 15) are given in Table 2. It is observed that the obtained results are in very good agreement
with the exact solution for all displacement and stresses. It is worth noting that when R increased, i.e, skin
properties are very different from core skin properties, the results derived from HSDT [7], IHSDT [41] and
TrSDT [42] solution seem to be not matched well with the exact solution. The results of the present method
4.1.2. Four layer [00 /900 /900 /00 ] square laminated plate under sinusoidally distributed load
Let us consider a simply supported square laminated plate subjected to a sinusoidal load q as shown in
Fig. 8. The length to width ratio is a/b=1 and the length to thickness ratios are a/h= 4, 10, 20, 100. The
plate is modeled by 17 × 17 control points per each side. Material I is used in this example. The normalized
h2
ϖ = (100E2 h3 )w( a2 , a2 , 0)/(qa4 ), σ̄xx = σ (a/2, a/2, h/2)
qb2 xx
h2 h2
σ̄yy = σ (a/2, a/2, h/4); τ̄xy
qb2 yy
= σ (0, 0, h/2)
qb2 xy
h
τ̄xz = qb σxy (0, b/2, 0)
18
Table 2: The normalized displacement and stresses of a square sandwich plate under a uniform load
R Method w̄ σ̄x1 σ̄x2 σ̄x3 σ̄y1 σ̄y2 σ̄y3
5 FEM-HSDT [7] 256.13 62.380 46.910 9.382 38.930 30.330 6.065
MRBF-HSDT [40] 257.110 60.366 47.003 9.401 38.456 30.242 6.048
MRBF-LW [11] 257.523 59.968 46.291 9.258 38.321 29.974 5.995
MRBFPS-LW [36] 258.179 60.063 46.393 9.279 38.364 30.029 6.006
CFS-IHSDT [41] 255.644 60.675 47.055 9.411 38.522 30.206 6.041
CFS-TrSDT [42] 256.706 60.525 47.061 9.412 38.452 30.177 6.035
Exact [8] 258.97 60.353 46.623 9.34 38.491 30.097 6.161
Quadratic 258.8338 60.2163 46.4783 9.2957 38.4798 30.0964 6.0193
Cubic 258.8346 60.2115 46.4765 9.2953 38.4706 30.0900 6.0180
Quartic 258.8347 60.2535 46.5101 9.302 38.4947 30.1094 6.0219
19
The results of the present method are compared with those of the several published ones such as the
FEM based on HSDT by Reddy [5], the finite strip method (FSM) based on HSDT by Akhras et al.[43],
the multiquadric radial basis function method (MRBF) based on a finite point formulation and HSDT by
Ferreira et al. [40], the node-based smoothed finite element method (NSFEM) based on HSDT by Thai et
al. [44], the multiquadric radial basis function method (MRBF) based on layerwise deformation theory by
Ferreira [11], the multiquadric radial basis function pseudospectral method (MRBF-PS) based on layerwise
deformation theory by Ferreira et al. [36] and the elasticity solution given in Pagano [45]. The comparison
is provided in Table 3. It is observed that the present formulation produces very accurate solutions for lam-
inated composite plates with all ratios a/h. The normalized displacement and stresses match well with the
exact solution [45]. For very thick a/h = 4 and thick a/h = 10 plates, the normalized displacement of the
present solution is better than that of HSDT [5, 43, 40, 44] solutions, while the normalized stresses of all
methods have same value. For thin plates, the differences between the present solution and others solution
are not significant. Fig. 9 describes the distribution of stresses through the thickness of the plate with a/h =
4.1.3. Sandwich (00 /core/00 ) square plate subjected to sinusoidally distributed load
Consider a sandwich (00 /core/00 ) simply supported square plate subjected to a sinusoidally distributed
load. The thickness of each face sheet is fixed at h/10. The length to thickness ratios are used as a/h =
4, 10, 20, 50, 100. Material II is used in this example. The plate is modeled by 17 × 17 elements. The nor-
20
Table 3: The normalized displacement and stresses of a four-layer [00 /900 /900 /00 ] laminated square plate under a sinusoidally
distributed load
a/h Method w̄ σ̄xx σ̄yy σ̄xz σ̄xy
4 FEM-HSDT[5] 1.8939 0.6806 0.6463 0.2109 0.0450
FSM-HSDT[43] 1.8937 0.6651 0.6322 0.2064 0.0440
MRBF-HSDT[40] 1.8864 0.6659 0.6313 0.1352 0.0433
NSFEM-HSDT[44] 1.9266 0.7076 0.6303 0.2084 0.0475
MRBF-LW[11] 1.9075 0.6432 0.6228 0.2166 0.0441
MRBFPS-LW [36] 1.9091 0.6429 0.6265 0.2173 0.0443
Elasticity [45] 1.954 0.720 0.666 0.270 0.0467
Quadratic 1.9060 0.6410 0.6248 0.2171 0.0443
Cubic 1.9060 0.6413 0.6251 0.2168 0.0443
Quartic 1.9060 0.6419 0.6257 0.2169 0.0443
21
0.5 0.5
a/h=20 a/h=20
0.4 a/h=10 0.4 a/h=10
a/h=4 a/h=4
0.3 0.3
0.2 0.2
Nomalized thickness z/h
0 0
−0.1 −0.1
−0.2 −0.2
−0.3 −0.3
−0.4 −0.4
−0.5 −0.5
−0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8
Stress, σ x (a/2,b/2,h/2) Stress, σ y (a/2,b/2,h/4)
0.5 0.5
a/h=20 a/h=20
0.4 a/h=10 0.4 a/h=10
a/h=4 a/h=4
0.3 0.3
0.2 0.2
Nomalized thickness z/h
0.1 0.1
0 0
−0.1 −0.1
−0.2 −0.2
−0.3 −0.3
−0.4 −0.4
−0.5 −0.5
−0.05 0 0.05 0.2 0.22 0.24 0.26 0.28 0.3 0.32 0.34
Stress, τ xy(0,0,h/2) Stress, τ xz(0,b/2,0)
Figure 9: The distribution of stresses through the thickness of a four-layer [00 /900 /900 /00 ] laminated composite square plate under
a sinusoidally distributed load.
22
a a h a a h
w̄ = 100h3 E2 w( , , )/q0 a4 , σ̄x = h2 σx ( , , )/q0 a2 ,
2 2 2 2 2 2
a a h h
σ̄y = h σy ( , , )/q0 a , σ̄xy = h σxy (0, 0, )/q0 a2 ,
2 2 2
2 2 2 2
b a
σ̄xz = hσxz (0, , 0)/q0 a, σ̄yz = hσyz ( , 0, 0)/q0 a
2 2
The obtained results of normalized transverse displacement and stresses are compared with the exact elas-
ticity solution by Pagano [45], the nine-node C0 quadratic finite element solution (FEMQ9) based on higher
order zigzag plate theory (HOZT) by Pandit et al. [37] and Chalak et al. [46] and the 3D mixed 18-node
finite element solution (MFEM3D) based on layerwise theory by Ramtekkar et al. [47], as given in Table 4.
Note that FEMQ9 solution based on HOZT has 11 degrees-of-freedom per node (dof/node) while MFEM3D
solution based on the layerwise theory has 6 dof/node (3 displacement components and 3 transverse stress
components). As seen, the present results match well with the exact elasticity solution [45] and other solu-
tions for all normalized displacement and stresses. It be can seen that the present approach is a very useful
tool to predict the static response of soft core sandwich structure. Fig. 10 exhibits the distribution of stresses
through the thickness of the plate using IGA corresponding to a/h=4, 10, 20. It is clear that the layerwise
Let us consider a three-layer laminate [00 /900 /00 ] rectangular plate with fully clamped boundary condi-
tions, where all layers of the laminated plate are assumed to be of the same thickness and mass density, and
made of the same linearly elastic composite materials. Material III is used. The normalized frequencies are
defined as ϖ = ωb2 /π 2 (ρh/D0 )1/2 with D0 = E2 h3 /12 (1 − ν12 ν21 ). The geometry domain and meshes
of the laminated composite plate are depicted in Fig. 11. The length to width a/b and width to thickness
b/h ratios are a/b = 1, 2 and b/h = 5, 10, 20, 100, respectively. The normalized frequencies for several
meshes are given in Table 5. It can been observed that the differences of normalized frequencies between
mesh 13 × 13 and mesh 17 × 17 are not significant. Hence, for a comparison with other methods, a mesh
of 13 × 13 elements can be employed. The first eight normalized frequencies of three layer [00 /900 /00 ]
clamped plate are indicated in Table 6 and Table 7 with respect to the length to width ratios a/b = 1, 2.
Obtained results are compared with the moving least squares differential quadrature method (DQM) based
on FSDT reported by Liew [38], the refined three-noded triangular finite element method (FEM-T3) based
23
Table 4: The normalized displacement and stresses of a sandwich (00 /core/00 ) simple supported square plate under a sinusoidally
distributed load
a/h Method w̄ σ̄x σ̄y σ̄xz σ̄yz σ̄xy
4 FEMQ9-HOZT [37] 7.6552 1.5218 0.2506 0.2520 0.1156 0.1468
FEMQ9-HOZT [46] 7.5822 1.5306 0.2581 0.2436 0.1147 0.1445
MFEM3D-LW [47] - 1.570 0.260 0.240 0.108 0.145
Elasticity [45] 7.5962 1.556 0.2595 0.239 0.1072 0.1437
Quadratic 7.6412 1.5127 0.2503 0.2389 0.1044 0.1469
Cubic 7.6411 1.5134 0.2504 0.2386 0.1041 0.1468
Quartic 7.6412 1.5149 0.2506 0.2387 0.1041 0.1468
24
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
Nomalized thickness z/h
−0.2 −0.2
−0.3 −0.3
−0.4 −0.4
−0.5 −0.5
−2 −1.5 −1 −0.5 0 0.5 1 1.5 2 −0.4 −0.3 −0.2 −0.1 0 0.1 0.2 0.3
Stress, σ x (a/2,b/2,h/2) Stress, σ y (a/2,b/2,h/2)
0.5 0.5
a/h=20
0.4 a/h=10 0.4
a/h=4
0.3 0.3
0.2 0.2
Nomalized thickness z/h
−0.2 −0.2
−0.3 −0.3
−0.4 −0.4
−0.5 −0.5
−0.2 −0.15 −0.1 −0.05 0 0.05 0.1 0.15 0.18 0.2 0.22 0.24 0.26 0.28 0.3 0.32
Stress, τ xy(0,0,h/2) Stress, τ xz(0,b/2,0)
0.5
0.4
0.3
0.2
Nomalized thickness z/h
0.1
a/h=20
0 a/h=10
a/h=4
−0.1
−0.2
−0.3
−0.4
−0.5
0.02 0.04 0.06 0.08 0.1
Stress, τ yz(a/2,0,0)
Figure 10: The distribution of stresses through the thickness of a sandwich (00 /core/00 ) square plate under a sinusoidally dis-
tributed load.
25
on global-local higher-order theory (GLHOT) proposed by Zhen and Wanji [48], the collocation method
based on the multiquadric radial basis function pseudospectral (MRBF-PS) and FSDT by Ferreira et al.[49]
as well as the collocation method based on wavelet function and FSDT reported by Ferreira et al.[50]. From
Table 6 and Table 7, it can be seen that the present approach produces very accurate results for all test cases.
Obtained results are very close to those of GLHOT in [48]. First six shape modes of the three-layer plate
Figure 11: Geometry and control points of a three-layer [0/90/0] laminated square plate.
Next, let us consider a four layer [00 /900 /900 /00 ] plate with simply supported boundary conditions.
The effects of the length to thickness a/h and elastic modulus ratios E1 /E2 are studied. The first normalized
frequency derived from the present method is listed in Table 8 corresponding to various modulus ratios and
a/h = 5. The obtained results are compared with exact solutions based on HSDT [51, 52], the moving least
squares differential quadrature method (DQM) [38] based on FSDT, the meshfree based on multiquadric
radial basis functions (MRBF) [53] and the meshfree based on wavelets function [50] based on FSDT. The
first normalized frequency of the present approach is slightly higher than those of other methods. The
influence of the length to thickness ratios is also considered, as shown in Table 9. The obtained results are
compared with those of GLHOT model [48], local higher-order theory (LHOT) [54], Matsunaga [55] based
26
27
Figure 12: Modes shape 1-8 of a three-layer [00 /900 /00 ] laminated clamped square plate with a/b= 1 and b/h= 10.
Table 5: The convergence of non-dimensional frequencies parameter ϖ of three layer [00 /900 /00 ] clamped laminated plate (b/h = 5
and a/b=1)
Method Mesh Modes
1 2 3 4 5 6 7 8
Quadratic 9×9 4.5695 6.5375 8.1565 9.4473 9.4525 11.6903 12.2408 12.8550
13 × 13 4.5687 6.5312 8.1545 9.4214 9.4416 11.6643 12.2286 12.7338
17 × 17 4.5686 6.5300 8.1541 9.4161 9.4405 11.6599 12.2264 12.7156
Cubic 9×9 4.5685 6.5295 8.1539 9.4149 9.4401 11.6589 12.2259 12.7178
13 × 13 4.5685 6.5294 8.1539 9.4138 9.4400 11.6579 12.2254 12.7083
17 × 17 4.5685 6.5294 8.1539 9.4137 9.4399 11.6578 12.2254 12.7077
Quartic 9×9 4.5685 6.5294 8.1539 9.4137 9.4400 11.6579 12.2254 12.7084
13 × 13 4.5685 6.5294 8.1539 9.4136 9.4399 11.6578 12.2254 12.7076
17 × 17 4.5685 6.5294 8.1539 9.4136 9.4399 11.6578 12.2254 12.7076
on the glocal-local higher-order theory, Cho et al. [56] based on HSDT and Thai et al. [44] based on HSDT.
As seen, the present results agree very well with other published ones.
shown in Fig. 13a. Material III is used and various fibre orientation angles α = 00 ; 150 ; 300 ; 450 are studied.
The circular plate has a radius to thickness ratio of 5 (R/t = 5). A rational quadratic basis is used to exactly
describe the geometry of the circular plate. Knot vectors Ξ × H of the coarsest mesh with one element
is defined as follows Ξ = {0, 0, 0, 1, 1, 1}; H = {0, 0, 0, 1, 1, 1}. Data of the circular plate are given in
Table 10. The normalized frequency of the circular plate is defined as ϖ = 4ωR2 /h (ρ/E2 )1/2 . A rational
quadratic basis is enough to model exactly the circular geometry. Coarse mesh and control net of the plate
with respect to quadratic, cubic and quartic elements are illustrated in Fig. 14.
Table 11 summarizes the normalized frequencies of circular plate derived from the present approach in
comparison with an alternative alpha finite element method with discrete shear gap technique (Aα-DSG3)
based FSDT by Thai et al. [57], the node-based smoothed finite element method (NS-FEM) with discrete
shear gap technique (NS-DSG3) based on HSDT by Thai et al. [44] and the moving least squares differential
quadrature method (MLSDQ) based on FSDT by [38]. It can be seen from Table 11 that the present method
leads to upper bound solutions compared to other published ones. Fig. 15 plots the first six mode shapes of
28
Table 6: A non-dimensional frequencies parameter ϖ of a [00 /900 /00 ] clamped laminated plate (a/b=1)
b/h Method Modes
1 2 3 4 5 6 7 8
5 Quadratic 4.5687 6.5312 8.1545 9.4214 9.4416 11.6643 12.2286 12.7338
Cubic 4.5685 6.5294 8.1539 9.4138 9.4400 11.6579 12.2254 12.7083
Quartic 4.5685 6.5294 8.1539 9.4136 9.4399 11.6578 12.2254 12.7076
DQM[38] 4.447 6.642 7.700 9.185 9.738 11.399 11.644 12.466
FEM-T3[48] 4.450 6.524 8.178 9.473 9.492 11.769 12.395 12.904
MRBF-PS[49] 4.5141 6.5080 8.0361 9.3468 9.3929 11.5749 12.0611 12.6564
Wavelets[50] 4.4466 6.6422 7.6996 9.1851 9.7393 11.3988 11.6448 12.4655
100 Quadratic 14.5073 17.8251 26.3596 38.0799 39.8541 41.9547 44.8746 54.0661
Cubic 14.4493 17.3999 24.2826 35.3281 37.8714 39.4790 43.5094 50.5589
Quartic 14.4490 17.3973 24.2522 35.1254 37.8692 39.4758 43.4910 49.6245
DQM[38] 14.666 17.614 24.511 35.532 39.157 40.768 44.786 50.297
FEM-T3[48] 14.601 17.812 25.236 37.168 38.528 40.668 45.724 53.271
MRBF-PS[49] 14.4305 17.3776 24.2662 35.5596 37.7629 39.3756 43.4874 51.7685
Wavelets[50] 14.4455 17.5426 25.1868 37.8851 39.5489 39.6519 44.0026 54.1828
29
Table 7: A non-dimensional frequencies parameter ϖ of a [00 /900 /00 ] clamped laminated plate (a/b=2)
b/h Method Modes
1 2 3 4 5 6 7 8
5 Quadratic 2.9697 4.3128 5.5254 6.1323 6.3885 7.7513 8.0973 8.7502
Cubic 2.9693 4.3123 5.5234 6.1302 6.3865 7.7484 8.0897 8.7420
Quartic 2.9693 4.3123 5.5233 6.1302 6.3865 7.7483 8.0895 8.7419
DQM[38] 3.045 4.248 5.792 5.905 6.535 7.688 7.729 9.176
FEM-T3[48] 2.953 4.288 5.595 6.096 6.446 7.796 8.053 9.005
MRBF-PS[49] 2.9679 4.2575 5.5406 6.0225 6.3620 7.6737 7.9414 8.7482
Wavelets[50] 3.0453 4.2483 5.7921 5.9046 6.5354 7.6901 7.7292 9.1778
100 Quadratic 5.1874 10.6614 11.1961 14.8481 20.0264 22.1089 22.8177 24.4300
Cubic 5.0901 10.4442 10.5453 14.2316 19.2994 19.6122 21.8736 22.1045
Quartic 5.0899 10.4430 10.5413 14.2279 19.2889 19.5742 21.8626 22.0706
DQM[38] 5.105 10.527 10.583 14.324 19.567 19.701 22.148 22.237
FEM-T3[48] 5.144 10.407 10.929 14.706 18.954 20.799 22.205 23.703
MRBF-PS[49] 5.0844 10.4349 10.5527 14.2538 19.2727 19.8125 21.9359 22.3671
Wavelets[50] 5.1102 10.5341 10.8046 14.4741 19.7425 20.7458 22.3679 23.1373
30
Table 8: A non-dimensional frequency parameter ϖ = ωa2 /h (ρ/E2 )1/2 of a [00 /900 /900 /00 ] SSSS laminated square plate
(a/h=5)
Method E1 /E2
10 20 30 40
Quadratic 8.3355 9.6143 10.3659 10.8775
Cubic 8.3354 9.6142 10.3658 10.8775
Quartic 8.3354 9.6142 10.3658 10.8775
MRBF-FSDT[53] 8.2526 9.4974 10.2308 10.7329
Wavelets-FSDT[50] 8.2794 9.5375 10.2889 10.8117
DQM-FSDT [38] 8.2924 9.5613 10.320 10.849
NSDSG3-HSDT [44] 8.2171 9.4631 10.1967 10.6983
Exact-HSDT[52, 51] 8.2982 9.5671 10.326 10.854
Table 9: A non-dimensional frequency parameter ϖ = ωa2 /h (ρ/E2 )1/2 of a [00 /900 /900 /00 ] SSSS laminated plate (a/b=1 and
E1 /E2 = 40)
Method a/h
4 5 10 20 25 50 100
Quadratic 9.3864 10.8775 15.2087 17.6934 18.0951 18.6844 18.8514
Cubic 9.3864 10.8774 15.2085 17.6928 18.0942 18.6809 18.8380
Quartic 9.3864 10.8774 15.2085 17.6928 18.0942 18.6809 18.8379
Zhen and Wanji [48] 9.2406 10.7294 15.1658 17.8035 18.2404 18.9022 19.1566
Whu and Chen [54] 9.193 10.682 15.069 17.636 18.055 18.670 18.835
Matsunaga [55] 9.1988 10.6876 15.0721 17.6369 18.0557 18.6702 18.8352
Cho et al. [56] - 10.673 15.066 17.535 18.054 18.670 18.835
Thai et al. [44] 9.2215 10.6982 15.0725 17.6620 18.0875 18.7139 18.8827
Table 10: Control points and weights for a circular plate with a radius R = 0.5
i 1 2 3 4 5 6 7 8 9
√ √ √ √ √ √
xi −√ 2/4 − 2/2 −√2/4 √0 0 √0 √2/4 2/2 √2/4
yi 2/4 √0 − 2/4 √2/2 0 −√ 2/2 2/4 √0 − 2/4
wi 1 2/2 1 2/2 1 2/2 1 2/2 1
31
Figure 13: Geometry and element mesh of a circular plate.
Figure 14: Coarse mesh and control points of a circular plate with various degrees: a) p=2; b) p=3 and c) p=4.
32
Table 11: A non-dimensional frequencies parameter ϖ = ωa2 /h (ρ/E2 )1/2 of a circular 4-layer [θ 0 / − θ 0 / − θ 0 /θ 0 ] clamped
laminated plate
θ0 Method Mode
1 2 3 4 5 6
00 Quadratic 22.9159 30.4104 41.8363 44.6467 52.4910 56.1835
Cubic 22.9131 30.3926 41.7512 44.6381 52.4653 55.8996
Quartic 22.9130 30.3922 41.7489 44.6377 52.4645 55.8883
Aα-DSG3[57] 22.0941 31.3552 43.7003 43.8558 45.1307 53.7838
NS-DSG3[44] 21.8936 30.164 42.23612 42.44091 51.14067 57.1407
MLSDQ[38] 22.211 29.651 41.101 42.635 50.309 54.553
33
Figure 15: Six mode shapes of a four-layer [450 / − 450 / − 450 /450 ] clamped laminated circular plate with R/h= 5.
34
4.2.3. Elliptical plate
Let us consider a three layer [00 /900 /00 ] elliptical plate with a fully clamped boundary using Mate-
rial IV . The elliptical plate radiuses are a=5 and b=2.5 as shown in Fig. 16, respectively. The normalized
frequency is defined by ϖ = ωa2 (ρh/D0 )1/2 with D0 = E2 h3 /12 (1 − ν12 ν21 ). This problem was numer-
ically studied by Chen et al. [39] using the Element Free Galerkin method (EFG) and classical laminated
plate theory (CLPT). The normalized first eight frequencies of the three-ply [00 /900 /00 ] laminated clamped
ellipse plate are presented in Table 12 with various length to thickness ratios. For the thin plate, the present
solution is also compared well with the EFG one. The first eight modes shape of three-layer [00 /900 /00 ]
For a comparison, the buckling load factor is defined as λ̄ = λcr a2 / E2 h3 where a, h, E2 and λcr are
the edge length, the thickness of the plate, the elastic modulus and the critical buckling load, respectively.
A simply supported four-layer cross-ply [00 /900 /900 /00 ] square laminated plate is subjected to uni-
axial compression as shown in Fig. 18a. The plate is modeled with 9 × 9, 13 × 13 and 17 × 17 elements.
The convergence of the normalized critical buckling load of a simply supported four-layer cross-ply square
laminated plate with a/h = 10 and E1 /E2 = 40 is provided in Table 13. The efficiency and accuracy of
the present method for various elastic modulus are considered. When a/h=10 and various elastic modulus
35
Table 12: A non-dimensional frequency parameter ϖ of a [00 /900 /00 ] fully clamped laminated ellipse plate
a/h Method Mode
1 2 3 4 5 6 7 8
5 Quadratic 14.158 19.976 27.143 28.862 34.955 35.162 41.876 43.841
Cubic 14.157 19.969 27.114 28.855 34.943 35.062 41.840 43.567
Quartic 14.157 19.969 27.113 28.855 34.943 35.059 41.839 43.552
100 Quadratic 19.327 31.265 49.765 52.989 64.717 88.135 90.223 99.767
Cubic 18.755 29.332 44.792 46.508 60.792 65.623 79.382 87.538
Quartic 18.752 29.307 44.596 46.485 60.722 64.581 79.011 87.324
Quadratic (CLPT) 19.796 30.201 46.412 54.697 67.452 69.591 86.011 99.172
Cubic (CLPT) 18.793 29.428 44.848 46.642 60.959 64.930 79.784 88.661
Quartic (CLPT) 18.790 29.413 44.763 46.551 60.867 64.757 79.339 87.621
EFG (CLPT) [39] 18.81 29.58 44.99 46.72 61.34 65.14 79.99 87.23
36
Figure 17: Eight mode shapes of a four-layer [00 /900 /00 ] clamped laminated ellipse plate with a/h= 10.
37
Table 13: The convergence of normalized critical buckling load of a [00 /900 /900 /00 ] simply supported cross-ply square plate with
a/h = 10 and E1 /E2 = 40
Method Mesh
9×9 13 × 13 17 × 17
Quadratic 23.66161 23.65992 23.65959
Cubic 23.65943 23.65943 23.65943
Quartic 23.65943 23.65943 23.65943
E1 /E2 ratios are used, the obtained results are compared to the 3D elasticity solution [58], the mesh-free
method based on radial point interpolation method (RPIM) with radial basis function and HSDT [59] and
the FEM solution based on HSDT [60, 51], as shown in Table 13. It is again seen that the present approach
shows a good competitor to other methods. Also, the normalized critical buckling loads become larger when
Next, the effect of the length-to-thickness ratios (a/h) subjected to the uniaxial compression load is also
considered for four layer [00 /900 /900 /00 ] simply supported square plates. Table 15 summarizes the nor-
malized critical buckling load of the present method and other methods such as the FEM based on FSDT
[61], the FEM based on FSDT [62] and HSDT [62]. Good agreement is obtained for the present solution.
Figure 18: Geometry of laminated composite plates under axial and biaxial compression.
Finally, we consider a three-layer symmetric cross-ply [00 /900 /00 ] simply supported plate subjected to
the biaxial buckling load as shown in Fig. 18b. Various length-to-thickness a/h and elastic modulus E1 /E2
ratios are studied in this example. Table 16 and Table 17 show the normalized critical buckling loads with
respect to various modulus and length-to-thickness ratios. The obtained results are compared with those of
38
Table 14: Normalized critical buckling load of a simply supported cross-ply [00 /900 /900 /00 ] square plate with various E1 /E2
ratios
Method E1 /E2
3 10 20 30 40
Quadratic 5.4094 10.0011 15.4442 19.9101 23.6599
Cubic 5.4089 10.0006 15.4437 19.9096 23.6594
Quartic 5.4089 10.0006 15.4437 19.9096 23.6594
Meshfree-HSDT [59] 5.412 10.013 15.309 19.778 23.412
FEM-HSDT [60] 5.114 9.774 15.298 19.957 23.340
FEM-HSDT [51] 5.442 10.026 15.418 19.813 23.489
Elasticity [58] 5.294 9.762 15.019 19.304 22.881
Table 15: Normalized critical buckling load of a simply supported cross-ply [00 /900 /900 /00 ] square plate with various ratios a/h
Method a/h
10 20 50 100
Quadratic 23.6599 31.8288 35.3945 36.0130
Cubic 23.6594 31.8267 35.3813 35.9617
Quartic 23.6594 31.8267 35.3813 35.9616
FEM-FSDT [61] 23.409 31.625 35.254 35.851
FEM-FSDT [62] 23.471 31.707 35.356 35.955
FEM-HSDT [62] 23.349 31.637 35.419 35.971
the finite element formulation based on FSDT [63], the finite element method based on HSDT [51] and the
meshfree method based on both FSDT and HSDT [59]. It can be seen that the present method is a good
competitor to the other methods for various modulus ratios and length to thickness ratios. The normalized
critical bi-axial buckling loads are increased with respect to increasing the modulus ratio E1 /E2 .
5. Conclusion
An isogeometric finite element formulation with a layerwise theory has been proposed for static, free
vibration and buckling analysis of laminated composite and sandwich plates. The present layerwise dis-
Table 16: Biaxial critical buckling load of a simply supported cross-ply [00 /900 /00 ] square plate with various modulus ratios
Method E1 /E2
10 20 30 40
Quadratic 4.9710 7.5699 8.8806 10.0066
Cubic 4.9707 7.5665 8.8772 10.0033
Quartic 4.9707 7.5665 8.8772 10.0033
FEM-FSDT [63] 4.963 7.588 8.575 10.202
FEM-HSDT [51] 4.963 7.516 9.056 10.259
39
Table 17: Biaxial critical buckling load of a simply supported cross-ply [00 /900 /00 ] square plate with various ratios a/h
Method a/h
2 5 10 15 20
Quadratic 1.5015 5.4592 10.0066 12.0984 13.1033
Cubic 1.5015 5.4586 10.0033 12.0907 13.0893
Quartic 1.5015 5.4586 10.0033 12.0907 13.0892
Meshfree-HSDT [59] 1.457 5.519 10.251 12.239 13.164
Meshfree-FSDT [59] 1.419 5.484 10.189 12.213 13.132
FEM-HSDT [51] 1.465 5.526 10.259 12.226 13.185
placement model assumed a first-order shear deformation theory in each layer and the imposition of dis-
placement continuity at the layers interfaces. It does not require the use of shear correction factors as in the
first-order shear deformation theory model. Weak forms of the static, free vibration and buckling models
for laminated composite and sandwich plates using the layerwise theory were derived. Obtained results
showed high accuracy for all test cases from the thin to thick plates. Specially, transverse shear stresses de-
rived from the proposed method are more accurate than those of several first-order and higher-order theories.
In this work, we restrict to the use of the layerwise theory. However, the extension of our formulation
to other theories such as generalized laminated plate theory [9], the unified formulation for the multilayered
plates [17, 18] and 3D elasticity solution [64] are straightforward. The authors believe that the method
presented herein can be very promising to provide an effectively alternative method of traditional finite
elements for analysis of laminated composite and sandwich plate structures in practice.
Acknowledgement
This research is funded by Vietnam National University HoChiMinh City (VNU-HCM) under grant
number B2013-18-04.
References
[1] J.R. Vinson. The Behavior of Sandwich Structures of Isotropic and Composite Materials. USA:
[2] J.N. Reddy. Mechanics of laminated composite plates and shells theory and anlysis, (Second edition).
40
[3] P. Bose and J.N. Reddy. Analysis of composite plates using various plate theories. Part 1. Formulation
[4] J.M. Whitney and N.J. Pagano. Shear deformation in heterogeneous anisotropic plates. Journal of
[5] J.N. Reddy. A simple higher-order theory for laminated composite plates. Journal of Applied Mechan-
[6] T. Kant and B.S. Manjunatha. An unsymmetric FRC laminate C0 finite element model with 12 degrees
[7] B.N. Pandya and T. Kant. Higher-order shear deformable theories for flexure of sandwich plates-finite
[8] S.Srinivas. A refined analysis of composite laminates. Journal of Sound and Vibration, 30:495–507,
1973.
[10] H. Murakami. Laminated composite plate theory with improved in-plane responses. Journal of Applied
[11] A.J.M. Ferreira. Analysis of composite plates using a layerwise theory and multiquadrics discretiza-
[12] S.T. Mau. A refined laminate plate theory. Journal of Applied Mechanics, 40:606–607, 1973.
[13] P.C. Chou and J. Corleone. Transverse shear in laminated plate theories. American Institute of Aero-
[14] M.Di Sciuva. An improved shear-deformation theory for moderately thick multilayered shells and
[15] A. Toledano and H. Murakami. A composite plate theory for arbitrary laminate configuration. Journal
41
[16] J.G. Ren. A new theory of laminated plate. Composites Science and Technology, 26:225–239, 1986.
[17] E. Carrera. C0 Reissner-Mindlin multilayered plate elements including Zig-Zag and interlaminar stress
[18] E. Carrera. Evaluation of layer-wise mixed theories for laminated plate analysis. American Institute of
[19] E. Carrera. Developments, ideas and evaluations based upon the Reissner’s mixed vari- ational theorem
in the modeling of multilayered plates and shells. Applied Mechanics Reviews, 54:301–329, 2001.
[20] E. Carrera. Historical review of zig-zag theories for multilayered plates and shells. Applied Mechanics
[21] T.J.R. Hughes, J.A. Cottrell, and Y. Bazilevs. Isogeometric analysis: CAD, finite elements, NURBS,
exact geometry and mesh refinement. Computer Methods in Applied Mechanics and Engineering,
194(39-41):4135–4195, 2005.
[22] J. Cottrell, T.J.R. Hughes, and A. Reali. Studies of refinement and continuity in isogemetric analysis.
[23] J.A. Cottrell, A. Reali, Y. Bazilevs, and T.J.R. Hughes. Isogeometric analysis of structural vibrations.
[24] T. Elguedj, Y. Bazilevs, V. Calo, and T. Hughes. B and F projection methods for nearly incompressible
linear and non-linear elasticity and plasticity using higher-order nurbs elements. Computer Methods
[25] W.A. Wall, M.A. Frenzel, and C. Cyron. Isogeometric structural shape optimization. Computer Meth-
[26] J. Kiendl, K.U. Bletzinger, J. Linhard, and R. Wuchner. Isogeometric shell analysis with Kirchhoff-
2009.
42
[27] J. Kiendl, Y. Bazilevs, M.C. Hsu, R. Wuchner, and K.U. Bletzinger. The bending strip method for
T. Rabczuk. Rotation free isogeometric thin shell analysis using PHT-splines. Computer Methods
[29] Chien H. Thai, H. Nguyen-Xuan, N. Nguyen-Thanh, T-H. Le, T. Nguyen-Thoi, and T. Rabczuk. Static,
free vibration, and buckling analysis of laminated composite Reissner-Mindlin plates using NURBS-
based isogeometric approach. International Journal for Numerical Methods in Engineering, 91:571–
603, 2012.
[30] D.J. Benson, Y. Bazilevs, M.C. Hsu, and T.J.R. Hughes. Isogeometric shell analysis: The Reissner–
Mindlin shell. Computer Methods in Applied Mechanics and Engineering, 199(5-8):276–289, 2010.
[31] Chien H. Thai, H. Nguyen-Xuan, S.P.A. Bordas, N. Nguyen-Thanh, and T. Rabczuk. Isogeometric
analysis of laminated composite plates using the higher-order shear deformation theory. Mechanics of
[32] D.J. Benson, Y. Bazilevs, M.C. Hsu, and T.J.R. Hughes. A large deformation, rotation-free, isoge-
2011.
[33] N. Nguyen-Thanh, H. Nguyen-Xuan, S. Bordas, and T. Rabczuk. Isogeometric analysis using poly-
nomial splines over hierarchical T-meshes for two-dimensional elastic solids. Computer Methods in
[34] R.N. Simpson, S.P.A. Bordas, J. Trevelyan, and T. Rabczuk. A two-dimensional isogeometric bound-
ary element method for elastostatic analysis. Computer Methods in Applied Mechanics and Engineer-
[35] L.A. Piegl, , and W. Tiller. The NURBS book. Springer Verlag, 1997.
43
[36] A.J.M. Ferreira, G.E. Fasshauer, R.C. Batra, and J.D. Rodrigues. Static deformations and vibration
analysis of composite and sandwich plates using a layerwise theory and RBF-PS discretizations with
[37] M. K. Pandit, A. H. Sheikh, and B.N. Singh. An improved higher order zigzag theory for the static
analysis of laminated sandwich plate with soft core. Finite Elements in Analysis and Design, 44:602–
610, 2008.
[38] K.M. Liew, Y.Q. Huang, and J.N. Reddy. Vibration analysis of symmetrically laminated plates based
on FSDT using the moving least squares differential quadrature method. Computer Methods in Applied
[39] X.L. Chen, G.R. Liu, and S.P. Lim. An element free galerkin method for the free vibration analysis of
[40] A.J.M. Ferreira, C.M.C. Roque, and P.A.L.S. Martins. Analysis of composite plates using higher-order
shear deformation theory and a finite point formulation based on the multiquadric radial basis function
[41] N. Grover, D.K. Maiti, and B.N. Singh. A new inverse hyperbolic shear deformation theory for static
and buckling analysis of laminated composite and sandwich plates. Composite Structures, 95, 2013.
[42] J.L. Mantari, A.S. Oktem, and C.G. Soares. A new trigonometric shear deformation theory for
isotropic, laminated composite and sandwich plates. International Journal of Solids and Structures,
49, 2012.
[43] G. Akhras, M.S. Cheung, and W. Li. Finite strip analysis for anisotropic laminated composite plates
[44] Chien H. Thai, Loc V. Tran, Dung T. Tran, T. Nguyen-Thoi, and H. Nguyen-Xuan. Analysis of lam-
inated composite plates using higher-order shear deformation plate theory and node-based smoothed
[45] N.J. Pagano. Exact solutions for rectangular bidirectional composites and sandwich plates. Journal of
44
[46] H.D. Chalak, A. Chakrabarti, M.A. Iqbal, and A. H. Sheikh. An improved C0 FE model for the analysis
of laminated sandwich plate with soft core. Finite Elements in Analysis and Design, 56:20–31, 2012.
[47] G.S. Ramtekkar, Y.M. Desai, and A.H. Shah. Application of a three dimensional mixed finite element
[48] W. Zhen and C. Wanji. Free vibration of laminated composite and sandwich plates using global-local
[49] A.J.M. Ferreira and G.E. Fasshauer. Computation of natural frequencies of shear deformable beams
and plates by an RBF-pseudospectral method. Computer Methods in Applied Mechanics and Engi-
[50] A.J.M. Ferreira, L.M.S. Castro, and S. Bertoluzza. A high order collocation method for the static and
vibration analysis of composite plates using a first-order theory. Composite Structures, 89(3):424–432,
2009.
[51] A.A. Khdeir and L. Librescu. Analysis of symmetric cross-ply elastic plates using a higher-order
theory: Part II: buckling and free vibration. Composite Structures, 9:259–277, 1988.
[52] J.N. Reddy. Mechanics of laminated composite plates. New York: CRC Press, 1997.
[53] A.J.M. Ferreira. A formulation of the multiquadric radial basis function method for the analysis of
[54] C.P. Wu and W.Y. Chen. Vibration and stability of laminated plates based on a local higher-order plate
[55] H. Matsunaga. Vibration and stability of cross-ply laminated composite plates according to a global
[56] K.N. Cho, C.W. Bert, and A.G. Striz. Free vibration of laminated rectangular plates analyzed by
element method with discrete shear gap technique for analysis of laminated composite plates. Applied
45
[58] A.K. Noor and Mathers. Shear-flexible finite element method of laminated composite plate. Technical
[59] L. Liu, L.P. Chua, and D.N. Ghista. Mesh-free radial basis function method for static, free vibration
and buckling analysis of shear deformable composite laminates. Composite Structures, 78:58–69,
2007.
[60] N.D. Phan and J.N. Reddy. Analysis of laminated composite plates using a higher-order shear defor-
mation theory. International Journal for Numerical Methods in Engineering, 21:2201–2219, 1985.
[61] A. Chakrabarti and A.H. Sheikh. Buckling of laminated composite plates by a new element based on
higher order shear deformation theory. Mechanics of Composite Materials and Structures, 10(4):303–
317, 2003.
[62] J.N. Reddy and N.D. Phan. Stability and vibration of isotropic, orthotropic and laminated plates
according to a higher order shear deformation theory. Journal of Sound and Vibration, 89:157–170,
1985.
[63] M.E. Fares and A.M. Zenkour. Buckling and free vibration of non-homogeneous composite cross-ply
laminated plates with various plate theories. Composite Structures, 44:279–287, 1999.
[64] A.K. Noor, J.M. Peters, and W.S. Burton. Three-dimensional solutions for initially stressed structural
46