Accepted Manuscript

Isogeometric analysis of laminated composite and sandwich plates using a

layerwise deformation theory

Chien H. Thai, A.J.M. Ferreira, E. Carrera, H. Nguyen-Xuan

PII: S0263-8223(13)00154-2
DOI: http://dx.doi.org/10.1016/j.compstruct.2013.04.002
Reference: COST 5101

To appear in: Composite Structures

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

Chien H. Thaia , A.J.M. Ferreirab , E. Carrerac , H. Nguyen-Xuana,d,∗

a Division of Computational Mechanics, Ton Duc Thang University, Hochiminh City, Vietnam
b Departamento de Engenharia Mecanica, Faculdade de Engenharia, Universidade do Porto, Rua Dr. Roberto Frias, 4200-465
Porto, Portugal
c Department of Aeronautics and Aerospace Engineering, Politecnico di Torino, Corso Duca degli Abruzzi, 24, 10129 Torino, Italy
d Department of Mechanics, University of Science, VNU, 227 Nguyen Van Cu Street, HCMC, Vietnam

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

buckling loads of the plates [1] are necessary.

∗ Corresponding author. Email address: nxhung@hcmus.edu.vn (H. Nguyen-Xuan)

Preprint submitted to Elsevier April 11, 2013

 Several laminate plate theories have been applied for analysis of composite and sandwich plates. The

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

review on relevant topics.

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,

analytical or semi-analytical and other numerical solutions.

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

some concluding remarks.

3
2. A brief of NURBS functions and surfaces

2.1. Knot vectors and basis functions

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 =

0 are defined as follows



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}

and Ξ = {0, 0, 0, 0, 0, 12 , 1, 1, 1, 1, 1}, respectively.

2.2. NURBS surface

The B-spline curve is defined as

n
C (ξ ) = ∑ Ni,p (ξ ) Pi (3)
i=1

where Pi are the control points, n denotes the number of control points and Ni,p (ξ ) is the pth -degree B-spline

basis function defined on the open knot vector.

4
 1 1

0.9 0.9

0.8 1 0.8

0.7 0.8 0.7

0.6 0.6 0.6

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

Figure 1: 1D and 2D quadratic B-spline basis functions.

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0
0 1/2 1

Figure 2: 1D and 2D cubic B-spline basis functions.

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0
0 1/2 1

Figure 3: 1D and 2D quartic B-spline basis functions.

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


tensor-product B-spline surface is defined as

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

traditional notation of a “node” I [30] and rewrite Eq. (4) as follows

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

that NIb (ξ , η) is a B-spline shape function.

Non-uniform rational B-splines (NURBS) are obtained by augmenting every point in the control mesh

PI with the weights wI . The control weight is constructed as follows

n×m
wg (ξ , η) = ∑ NIb (ξ , η) wI (6)
I=1

The NURBS surfaces are then defined by

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

surface with 4 × 4 elements is demonstrated in Fig. 4.

3. An isogeometric formulation for laminated composite and sandwich plates using layerwise theory

3.1. The displacements, strains and stresses in plates

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

model is also easily extendible for a general laminate layer.

The displacement field for the middle layer is given as

Figure 5: 1D representation of the layerwise kinematics.

(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)

w(2) (x, y, z) = w0 (x, y)

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,

for the middle layer (layer 2).

From Fig. 5, the corresponding displacement field for the upper layer (layer 3) is given as

(2) (3) (3)

u(3) (x, y, z) = u0 (x, y) + h22 βx + h23 βx + z(3) βx
(2) (3) (3)
v(3) (x, y, z) = v0 (x, y) + h22 βy + h23 βy + z(3) βy (9)

w(3) (x, y, z) = w0 (x, y)

And the corresponding displacement field for the lower layer (layer 1) is proposed as

(2) (1) (1)

u(1) (x, y, z) = u0 (x, y) − h22 βx − h21 βx + z(1) βx
(2) (1) (1)
v(1) (x, y, z) = v0 (x, y) − h22 βy − h21 βy + z(1) βy (10)

w(1) (x, y, z) = w0 (x, y)

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

The membrane components are given by

   
∂ u0

 ε m(k) 
 

∂x


 xx

 
 
 

  
∂ v0
 εyym(k) 
=
 ∂x 
(14)
   
m(k)
   ∂ u0

+ ∂∂vx0

 εxy 
 
 

∂y

The bending components can be expressed as

   
(k)

 ε b(k) 
 
 ∂ βx
∂x


 xx

 
 
 

  (k)

∂ βy
 εyyb(k) 
=
 ∂y 
(15)
(k)
   (k)

b(k)
   ∂ βx ∂ βy

+

 εxy 
 
 

∂y ∂x

The membrane-bending coupling components for layers 1, 2, and 3 are given as

(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

for plane stress by

   (k)  
(k) (k)
σ1 Q Q12 Q16 0 0 ε1
 11

 
 
 


 
    

    
 (k)  
 Q12 Q22 Q26 0
 (k) 

 σ 
 0  
 ε2 

 2

 
     

    
(k) (k)
τ12 =
 Q61 Q62 Q66 0 0   γ12  (19)

 
    
(k) (k) 
    
 0 0 0 Q55 Q54  
   


 τ13 

  γ13 

 
    




 (k)
   (k) 
0 0 0 Q45 Q44
  
τ23 γ23

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

(k) (k) (k) (k)

(k) E1 (k) ν12 E2 (k) E2
Q11 = (k) (k) , Q12 = (k) (k) , Q22 = (k) (k)
1−ν12 ν21 1−ν12 ν21 1−ν12 ν21 (20)
(k) (k) (k) (k) (k) (k)
Q66 = G12 , Q55 = G13 , Q44 = G23

(k) (k) (k) (k) (k)

where E1 , E2 are the Young modulus in the 1 and 2 directions, respectively, and G12 , G23 , G13 are the
(k) (k)
shear modulus in the 1-2, 2-3, 3-1 plane, respectively, and ν12 and ν21 are Poisson’s ratios.

The laminate is usually made of several orthotropic layers. Each layer must be transformed into the laminate

coordinate system (x, y, z). The stress - strain relationship is given as

    
(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).

3.2. Weak form

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

taken into account in the weak form as

! !
Z 3  T Z 3
 T Z
(k) (k) (k) (k)
∑ δ ε p D̄ε p dΩ + ∑δ γ Cγ dΩ = ∇T δ w0 σ̂0 ∇w0 dΩ (25)
k=1 k=1
Ω Ω Ω

 
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

Therefore w10 = w20 = w30 = w, Eq. (26) can be expressed as

 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.

The in-plane strains and shear strains are written as

 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

For the geometrical strains, one writes

n×m
εg = ∑ BgI qI (30)
I=1

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)

For free vibration analysis, one has

K − ω 2 M q = 0,


(32)

And for buckling analysis, we have

(K − λcr Kg ) q = 0 (33)

where K is the global stiffness matrix

  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

and the global geometrical stiffness matrix Kg is

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

points are used. The material parameters are assumed as

• Material I:

E1 = 25E2 , , G12 = G13 = 0.5E2 , G23 = 0.2E2 , ν12 = 0.25, ρ = 1.

• Material II: [37]

Face sheets:

E1 = 25E, E2 = E, G12 = G13 = 0.5E, G23 = 0.2E, ν12 = 0.25.

15
 Core:

E1 = 0.04E, E2 = 0.04E, E3 = 0.5E, G12 = 0.016E, G13 = G23 = 0.06, ν12 = 0.25.

• Material III: [38]

E1 = 40E2 , G12 = G13 = 0.6E2 , G23 = 0.5E2 , ν12 = 0.25, ρ = 1.

• Material IV : [39]

E1 = 2.45E2 , G12 = G13 = 0.48E2 , G23 = 0.2E2 , ν12 = 0.23, ρ = 1.

4.1. Static analysis

4.1.1. Three-layer sandwich square plate subjected to uniform load

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

layer (core), which has the following properties

 
0.999781 0.231192 0 0 0
 
 
 0.231192 0.524886 0 0 0 
 
 
Q̄core = 
 0 0 0.262931 0 0 

 
0 0 0 0.266810 0
 
 
 
0 0 0 0 0.159914

and two outside layers (skins) are calculated as

Q̄skin = RQ̄core

The normalized displacement and stresses of the plate are defined as

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.

Figure 7: Meshes and control net of a square 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

are in very good agreement with the exact solution.

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

displacement and stresses of a square plate are defined as

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

10 FEM-HSDT [7] 152.33 64.650 51.310 5.131 42.830 33.970 3.397

MRBF-HSDT[40] 154.658 65.381 49.973 4.997 43.240 33.637 3.364
MRBF-LW [11] 158.380 64.846 48.443 4.844 43.390 33.306 3.924
MRBFPS-LW [36] 158.912 64.993 48.601 4.860 43.491 33.409 3.341
CFS-IHSDT [41] 154.550 65.741 49.798 4.979 43.400 33.556 3.356
CFS-TrSDT[42] 155.498 65.542 49.708 4.971 43.385 33.591 3.359
Exact [8] 159.38 65.332 48.857 4.903 43.566 33.413 3.500
Quadratic 159.4058 65.1921 48.699 4.8699 43.6322 33.4853 3.3485
Cubic 159.406 65.1854 48.6971 4.8697 43.6218 33.4787 3.3479
Quartic 159.4059 65.2297 48.7334 4.8733 43.6486 33.5008 3.3501

15 FEM-HSDT [7] 110.430 66.620 51.970 3.465 44.920 35.410 2.361

MRBF-HSDT [40] 114.644 66.919 50.323 3.355 45.623 35.167 2.345
MRBF-LW [11] 120.988 66.291 47.899 3.193 46.292 34.890 2.326
MRBFPS-LW [36] 121.347 66.436 48.011 3.201 46.385 34.965 2.331
CFS-IHSDT [41] 115.820 67.272 49.813 3.321 45.967 35.088 2.339
CFS-TrSDT [42] 115.919 67.185 49.769 3.318 45.910 35.081 2.339
Exact [8] 121.720 66.787 48.299 3.238 46.424 34.955 2.494
Quadratic 121.7766 66.6431 48.1323 3.2088 46.5282 35.0624 2.3375
Cubic 121.7765 66.6348 48.1303 3.2087 46.5172 35.0560 2.3371
Quartic 121.7764 66.6792 48.1672 3.2111 46.5455 35.0797 2.3386

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, 10 and 20, respectively.

Figure 8: Geometry of a square laminated plate under sinusoidally distributed load.

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-

malized transverse displacement and stresses are given as follows:

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

10 FEM-HSDT[5] 0.7149 0.5589 0.3974 0.2697 0.0273

FSM-HSDT[43] 0.7147 0.5456 0.3888 0.2640 0.0268
MRBF-HSDT[40] 0.7153 0.5466 0.4383 0.3347 0.0267
NSFEM-HSDT[44] 0.7246 0.5609 0.3909 0.2812 0.0288
MRBF-LW[11] 0.7309 0.5496 0.3956 0.2888 0.0273
MRBFPS-LW [36] 0.7303 0.5487 0.3966 0.2993 0.0273
Elasticity [45] 0.743 0.559 0.403 0.301 0.0276
Quadratic 0.7296 0.5474 0.3957 0.2995 0.0273
Cubic 0.7296 0.5477 0.3959 0.2989 0.0272
Quartic 0.7296 0.5482 0.3963 0.2990 0.0272

20 FEM-HSDT[5] 0.5061 0.5523 0.311 0.2883 0.0233

FSM-HSDT[43] 0.506 0.5393 0.3043 0.2825 0.0228
MRBF-HSDT[40] 0.507 0.5405 0.3648 0.3818 0.0228
NSFEM-HSDT[44] 0.5089 0.5433 0.3050 0.3051 0.0234
MRBF-LW[11] 0.5121 0.5417 0.3056 0.3248 0.0230
MRBFPS-LW [36] 0.5113 0.5407 0.3073 0.3256 0.0230
Elasticity [45] 0.517 0.543 0.309 0.328 0.0230
Quadratic 0.5109 0.5395 0.3066 0.3268 0.023
Cubic 0.5109 0.5397 0.3068 0.3251 0.0229
Quartic 0.5109 0.5403 0.3071 0.3253 0.023

100 FEM-HSDT[5] 0.4343 0.5507 0.2769 0.2948 0.0217

FSM-HSDT[43] 0.4343 0.5387 0.2708 0.2897 0.0213
MRBF-HSDT[40] 0.4365 0.5413 0.3359 0.4106 0.0215
NSFEM-HSDT[44] 0.4345 0.5384 0.2706 0.3183 0.0211
MRBF-LW[11] 0.4374 0.542 0.2697 0.3232 0.0216
MRBFPS-LW [36] 0.4348 0.5391 0.2711 0.3359 0.0214
Elasticity [45] 0.4347 0.539 0.271 0.339 0.0214
Quadratic 0.4343 0.5377 0.2704 0.3359 0.0214
Cubic 0.4343 0.5382 0.2707 0.3354 0.0213
Quartic 0.4343 0.5387 0.271 0.3357 0.0213

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

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
−0.8 −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

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

model reflects very well the mechanical characterization of sandwich plates.

4.2. Free vibration analysis

4.2.1. Square laminated plates

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

10 FEMQ9-HOZT [37] 2.2002 1.1483 0.1086 0.3158 0.0570 0.0709

FEMQ9-HOZT [46] 2.1775 1.1528 0.1143 0.3058 0.0575 0.0705
MFEM3D-LW [47] - 1.159 0.111 0.303 0.055 0.071
Elasticity [45] 2.2004 1.1530 0.1104 0.300 0.0527 0.0707
Quadratic 2.2011 1.1480 0.1088 0.2998 0.0518 0.0712
Cubic 2.2011 1.1485 0.1089 0.2992 0.0512 0.0712
Quartic 2.2011 1.1497 0.1090 0.2993 0.0513 0.0712

20 FEMQ9-HOZT [37] 1.2254 1.1055 0.0694 0.3342 0.0392 0.0509

FEMQ9-HOZT [46] 1.2121 1.1103 0.0742 0.3272 0.0399 0.0508
MFEM3D-LW [47] - 1.110 0.070 0.317 0.036 0.051
Elasticity [45] 1.2264 1.110 0.0700 0.317 0.0361 0.0511
Quadratic 1.2262 1.1074 0.0696 0.3180 0.0364 0.0512
Cubic 1.2262 1.1079 0.0696 0.3166 0.0352 0.0511
Quartic 1.2262 1.1090 0.0697 0.3168 0.0352 0.0511

50 FEMQ9-HOZT [37] 0.9341 1.0948 0.0566 0.3403 0.0333 0.0445

FEMQ9-HOZT [46] 0.9234 1.0997 0.0611 0.3300 0.0321 0.0443
Elasticity [45] 0.9348 1.099 0.0569 0.323 0.0306 0.0446
Quadratic 0.9347 1.0972 0.0568 0.3293 0.0353 0.0447
Cubic 0.9348 1.0978 0.0568 0.3223 0.0299 0.0446
Quartic 0.9348 1.0989 0.0569 0.3225 0.0299 0.0446

100 FEMQ9-HOZT [37] 0.8917 1.1093 0.0547 0.3412 0.0324 0.0434

FEMQ9-HOZT [46] 0.8814 1.0982 0.0592 0.3426 0.0322 0.0433
Elasticity [45] 0.8923 1.098 0.0550 0.324 0.0297 0.0433
Quadratic 0.8921 1.0955 0.0548 0.3500 0.0465 0.0438
Cubic 0.8924 1.0964 0.0549 0.3231 0.0291 0.0437
Quartic 0.8924 1.0975 0.0549 0.3234 0.0291 0.0437

24
 0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2
Nomalized thickness z/h

Nomalized thickness z/h

0.1 0.1
a/h=20 a/h=20
0 a/h=10 0 a/h=10
a/h=4 a/h=4
−0.1 −0.1

−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

Nomalized thickness z/h

0.1 0.1
a/h=20
0 0 a/h=10
a/h=4
−0.1 −0.1

−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

with b/h = 10 are plotted in Fig. 12.

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.

4.2.2. Circular plate

In this example, a circular four-layer [θ / − θ / − θ /θ ] laminated plate with a clamped boundary, as

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

a circular four-layer [45/ − 45/ − 45/45] clamped laminated plate.

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

10 Quadratic 7.6184 10.3081 14.5667 14.8933 16.2407 19.5759 20.6089 22.6955

Cubic 7.6174 10.3005 14.5647 14.8624 16.2343 19.5511 20.5153 22.6870
Quartic 7.6174 10.3004 14.5646 14.8618 16.2342 19.5506 20.5127 22.6868
DQM[38] 7.411 10.393 13.913 15.429 15.806 19.572 21.489 21.620
FEM-T3[48] 7.484 10.207 14.340 14.863 16.070 19.508 20.716 22.489
MRBF-PS[49] 7.4727 10.2544 14.2440 14.9363 15.9807 19.4129 20.6868 22.1851
Wavelets[50] 7.4106 10.3944 13.9128 15.4403 15.8061 19.5797 21.4892 21.6855

20 Quadratic 11.1376 14.0913 20.1837 23.8802 25.5749 28.8250 29.5899 36.2476

Cubic 11.1344 14.0677 20.0766 23.8709 25.5541 28.4887 29.5095 35.9751
Quartic 11.1343 14.0671 20.0737 23.8707 25.5535 28.4773 29.5072 35.9657
DQM[38] 10.953 14.028 20.388 23.196 24.978 29.237 29.369 36.266
FEM-T3[48] 11.003 14.064 20.321 23.498 25.350 29.118 29.679 36.624
MRBF-PS[49] 10.9680 13.9636 20.0983 23.3572 25.0859 28.6749 29.1620 35.8138
Wavelets[50] 10.9528 14.0360 20.4533 23.1974 24.9827 29.2795 29.6910 36.5184

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

10 Quadratic 4.1029 6.7268 8.0337 9.7274 10.2712 12.5040 13.3972 14.2019

Cubic 4.1014 6.7241 8.0241 9.7182 10.2633 12.4918 13.3629 14.1780
Quartic 4.1014 6.7241 8.0239 9.7180 10.2632 12.4915 13.3622 14.1774
DQM[38] 2.9680 4.2576 5.5408 6.0225 6.3620 7.6730 7.9411 8.7462
FEM-T3[48] 4.119 6.705 8.240 9.916 10.212 12.671 14.066 14.082
MRBF-PS[49] 4.0924 6.6205 8.0953 9.7047 10.0482 12.3575 13.5224 13.8453
Wavelets[50] 4.1410 6.6167 8.3563 9.8970 9.9677 12.4443 13.6695 14.1332

20 Quadratic 4.7734 8.9301 9.7188 12.4820 14.9758 17.3042 17.4716 19.1488

Cubic 4.7685 8.9199 9.6863 12.4507 14.9438 17.1804 17.4268 19.0326
Quartic 4.7684 8.9196 9.6855 12.4498 14.9427 17.1772 17.4253 19.0293
DQM[38] 4.779 8.840 9.847 12.511 14.703 17.300 17.673 19.429
FEM-T3[48] 4.813 8.954 9.968 12.768 14.960 17.764 18.041 19.993
MRBF-PS[49] 4.7593 8.8318 9.7221 12.4153 14.7156 17.2484 17.3088 19.1064
Wavelets[50] 4.7790 8.8425 9.8600 12.5212 14.7160 17.3161 17.7502 19.4969

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.

(a) (b) (c)

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

150 Quadratic 23.2623 31.6951 43.7880 44.8951 53.9244 58.5518

Cubic 23.2593 31.6780 43.7111 44.8855 53.8963 58.2989
Quartic 23.2592 31.6775 43.7087 44.8851 53.8954 58.2887
Aα-DSG3[57] 22.261 31.792 42.713 44.603 53.059 59.669
NS-DSG3[44] 22.5458 31.98826 43.21025 44.8315 53.60512 60.1677
MLSDQ[38] 22.774 31.455 43.350 43.469 52.872 57.386

300 Quadratic 24.2441 35.4999 44.9133 50.0683 56.3119 66.4637

Cubic 24.2403 35.4822 44.9019 50.0077 56.2737 66.2844
Quartic 24.2399 35.4812 44.9012 50.0052 56.2719 66.2771
Aα-DSG3[57] 23.699 36.454 43.269 51.819 56.851 63.917
NS-DSG3[44] 23.9464 36.5811 43.7539 52.0899 57.1992 64.8231
MLSDQ[38] 24.071 36.153 43.968 51.074 56.315 66.220

450 Quadratic 24.8004 37.7486 44.6872 54.1066 56.9615 67.7756

Cubic 24.7957 37.7292 44.6750 54.0674 56.8998 67.7415
Quartic 24.7952 37.7279 44.6742 54.0659 56.8963 67.7403
Aα-DSG3[57] 24.436 39.175 43.061 57.229 57.369 63.425
NS-DSG3[44] 24.6385 39.3732 43.4140 57.5771 57.7129 64.1297
MLSDQ[38] 24.752 39.181 43.607 56.759 56.967 65.571

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 ]

fully clamped laminated ellipse plate are displayed in Fig. 17.

Figure 16: Geometry and element mesh of a clamped ellipse plate.

4.3. Buckling analysis of composite plates

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.

Material III is used in the examples.

4.3.1. Square plate under uniaxial compression

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

10 Quadratic 17.189 25.736 37.084 39.228 49.194 50.599 61.041 65.915

Cubic 17.184 25.714 36.982 39.196 49.148 50.259 60.917 65.006
Quartic 17.184 25.713 36.979 39.196 49.147 50.245 60.912 64.951

20 Quadratic 18.353 28.363 42.630 44.456 57.268 61.059 73.437 81.281

Cubic 18.329 28.280 42.255 44.321 57.090 59.827 72.973 80.465
Quartic 18.329 28.278 42.246 44.320 57.087 59.779 72.956 80.307

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

the E1 /E2 modulus ratios increase as given in Table 14.

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.

4.3.2. Square plate under 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:

Technomic Publishing, 1999.

[2] J.N. Reddy. Mechanics of laminated composite plates and shells theory and anlysis, (Second edition).

New York: CRC Press, 2004.

40
 [3] P. Bose and J.N. Reddy. Analysis of composite plates using various plate theories. Part 1. Formulation

and analytical solutions. Structural Engineering and Mechanics, 6:583–612, 1998.

[4] J.M. Whitney and N.J. Pagano. Shear deformation in heterogeneous anisotropic plates. Journal of

Applied Mechanics, 37(4):1031–1036, 1970.

[5] J.N. Reddy. A simple higher-order theory for laminated composite plates. Journal of Applied Mechan-

ics, 51:745–752, 1984.

[6] T. Kant and B.S. Manjunatha. An unsymmetric FRC laminate C0 finite element model with 12 degrees

of freedom per node. Engineering Computations, 5:300–308, 1988.

[7] B.N. Pandya and T. Kant. Higher-order shear deformable theories for flexure of sandwich plates-finite

element evaluations. International Journal of Solids and Structures, 24:419–451, 1988.

[8] S.Srinivas. A refined analysis of composite laminates. Journal of Sound and Vibration, 30:495–507,

1973.

[9] J. N. Reddy. A generalization of two-dimensional theories of laminated composite plates. Communi-

cations in Applied Numerical Methods, 3:173–180, 1987.

[10] H. Murakami. Laminated composite plate theory with improved in-plane responses. Journal of Applied

Mechanics, 53:661–666, 1986.

[11] A.J.M. Ferreira. Analysis of composite plates using a layerwise theory and multiquadrics discretiza-

tion. Mechanics of Advanced Materials and Structures, 12:99–112, 2005.

[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-

nautics and Astronautics, 11:1333–1336, 1973.

[14] M.Di Sciuva. An improved shear-deformation theory for moderately thick multilayered shells and

plates. Journal of Applied Mechanics, 54:589–597, 1987.

[15] A. Toledano and H. Murakami. A composite plate theory for arbitrary laminate configuration. Journal

of Applied Mechanics, 54:181–189, 1987.

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

continuity. International Journal for Numerical Methods in Engineering, 39:1797–1820, 1996.

[18] E. Carrera. Evaluation of layer-wise mixed theories for laminated plate analysis. American Institute of

Aeronautics and Astronautics, 36:830–839, 1998.

[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

Reviews, 56:287–308, 2003.

[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.

Computer Methods in Applied Mechanics and Engineering, 196:4160–4183, 2007.

[23] J.A. Cottrell, A. Reali, Y. Bazilevs, and T.J.R. Hughes. Isogeometric analysis of structural vibrations.

Computer Methods in Applied Mechanics and Engineering, 195(41-43):5257–5296, 2006.

[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

in Applied Mechanics and Engineering, 197:2732–2762, 2008.

[25] W.A. Wall, M.A. Frenzel, and C. Cyron. Isogeometric structural shape optimization. Computer Meth-

ods in Applied Mechanics and Engineering, 197(33-40):2976–2988, 2008.

[26] J. Kiendl, K.U. Bletzinger, J. Linhard, and R. Wuchner. Isogeometric shell analysis with Kirchhoff-

Love elements. Computer Methods in Applied Mechanics and Engineering, 198(49-52):3902–3914,

2009.

42
[27] J. Kiendl, Y. Bazilevs, M.C. Hsu, R. Wuchner, and K.U. Bletzinger. The bending strip method for

isogeometric analysis of Kirchhoff-Love shell structures comprised of multiple patches. Computer

Methods in Applied Mechanics and Engineering, 199(37-40):2403–2416, 2010.

[28] N. Nguyen-Thanh, J. Kiendl, H. Nguyen-Xuan, R. Wuchner, K.U. Bletzinger, Y. Bazilevs, and

T. Rabczuk. Rotation free isogeometric thin shell analysis using PHT-splines. Computer Methods

in Applied Mechanics and Engineering, 200(47-48):3410–3424, 2011.

[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

Advanced Materials and Structures, accepted, 2012.

[32] D.J. Benson, Y. Bazilevs, M.C. Hsu, and T.J.R. Hughes. A large deformation, rotation-free, isoge-

ometric shell. Computer Methods in Applied Mechanics and Engineering, 200(13-16):1367–1378,

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

Applied Mechanics and Engineering, 200:1892–1908, 2011.

[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-

ing, 209-212:87–100, 2012.

[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

optimal shape parameter. Composite Structures, 86:328–343, 2008.

[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

Mechanics and Engineering, 192:2203–22, 2003.

[39] X.L. Chen, G.R. Liu, and S.P. Lim. An element free galerkin method for the free vibration analysis of

composite laminates of complicated shape. Composite Structures, 59:279–289, 2003.

[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

method. Composites: Part B, 34:627–636, 2003.

[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

using higher-order deformation theory. Computers and Structures, 52(3):471–7, 1994.

[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

discrete shear gap method. Applied Mathematical Modelling, 36:5657–5677, 2012.

[45] N.J. Pagano. Exact solutions for rectangular bidirectional composites and sandwich plates. Journal of

Composite Materials, 4:20–34, 1970.

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

model to the flexure of sandwich plate. Composite Structures, 81:2383–2398, 2003.

[48] W. Zhen and C. Wanji. Free vibration of laminated composite and sandwich plates using global-local

higher-order theory. Journal of Sound and Vibration, 298:333–49, 2006.

[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-

neering, 196:134–146, 2006.

[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

laminated composite plates. Composite Structures, 59(3):385–92, 2003.

[54] C.P. Wu and W.Y. Chen. Vibration and stability of laminated plates based on a local higher-order plate

theory. Journal of Sound and Vibration, 177:503–520, 1994.

[55] H. Matsunaga. Vibration and stability of cross-ply laminated composite plates according to a global

higher-order plate theory. Composite Structures, 48:231–244, 2000.

[56] K.N. Cho, C.W. Bert, and A.G. Striz. Free vibration of laminated rectangular plates analyzed by

higher-order individual-layer theory. Journal of Sound and Vibration, 145:429–442, 1991.

[57] C. Thai-Hoang, N. Nguyen-Thanh, H. Nguyen-Xuan, and T. Rabczuk. An alternative alpha finite

element method with discrete shear gap technique for analysis of laminated composite plates. Applied

Mathematics and Computation, 217(17):7324–7348, 2011.

45
[58] A.K. Noor and Mathers. Shear-flexible finite element method of laminated composite plate. Technical

report, NASA, 1975.

[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

sandwiches. Journal of Engineering Mechanics (ASCE), 120:284–303, 1994.

46