See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/356462044

Static, free vibration, and buckling analyses of laminated composite plates via
an isogeometric meshfree collocation approach

Article  in  Composite Structures · November 2021

DOI: 10.1016/j.compstruct.2021.115011

CITATIONS READS

5 192

5 authors, including:

Nhon Nguyen-Thanh Kun Zhou

Nanyang Technological University Nanyang Technological University
41 PUBLICATIONS   2,638 CITATIONS    502 PUBLICATIONS   16,567 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Isogeometric analysis View project

Material development and optimisation of nanocomposite for selective laser sintering View project

All content following this page was uploaded by Nhon Nguyen-Thanh on 04 November 2022.

The user has requested enhancement of the downloaded file.

Static, free vibration, and buckling analyses of laminated composite plates
via an adaptive isogeometric meshfree collocation approach
Jiazhao Huang2, Nhon Nguyen-Thanh2, Zheng Fan1, Kun Zhou1, *
1
School of Mechanical and Aerospace Engineering, Nanyang Technological University, 50
Nanyang Avenue, 639798, Singapore
2
School of Electrical and Electronic Engineering, Nanyang Technological University, 50
Nanyang Avenue, 639798, Singapore

Abstract

This paper aims to develop a isogeometric meshfree collocation (IMC) approach for the

static, free vibration, and buckling analyses of laminated composite plates. As a promising

alternative to Galerkin method, collocation method has been introduced into the isogeometric

analysis framework to reduce the computational cost and achieve high convergence rate. An

adaptive refinement strategy based on the gradient of strain energy is further performed to

generate more accurate results with higher computational efficiency. In this way, laminated

composite plates can be modeled by considering only one element throughout its thickness

with homogenized material properties. The robustness and efficacy of the proposed method

has been demonstrated through a series of benchmark problems. It is found that the proposed

approach can yield more accurate results at a similar convergence rate compared to its

isogeometric analysis collocation counterpart. The effects of several parameters such as the

Young’s modulus ratio, thickness-to-span ratio and orthotropy angle on the structural analysis

of composite plates have been subsequently investigated in detail under different boundary

and loading conditions. The simulation results are found to be in good agreement with

reference solutions.

Keywords: Laminated composite plate; Isogeometric analysis; Collocation method; First-

order shear deformation theory; Adaptive mesh refinement strategy

*Corresponding author: Tel.: +65 6790 5499; Fax: +65 6792 4062
[email protected] (K. Zhou)
1
1. Introduction

Composite structures are merged by layers of materials with different physical and

chemical properties to achieve desirable properties such as high fatigue strength, robust

design flexibility, low self-weight, and specific stiffness. Among these structures, laminated

composite plates are employed in a wide range of engineering applications, including civil,

naval, aerospace and electrical. Owning to the increasing applications of composite materials,

a plethora of methods have been developed for structural analyses. The finite element method

(FEM) is one of the most employed numerical techniques for the study of composite plates

[1, 2]. Setoodeh and Karami [3] employed a 3D layer wise FEM to perform a comprehensive

investigation on thick laminated composite plate. Gherlone and Sciuva [4] studied the

thermo-mechanics of composite plates using the FEM sub-laminates approach. However,

conventional FEM still suffers from drawbacks such as C0 inter-element continuity, tedious

remeshing process and severe mesh distortion at large deformations [5].

As potential alternatives, the isogeometric analysis (IGA) and the meshfree methods can

offer higher continuity and better geometry representation than traditional FEM. Shojaee et

al. utilized classical plate theory to study the free vibration and buckling of composite plates

within the IGA framework [6]. Thai et al. [7] integrated layer wise deformation theory with

IGA for a comprehensive structural analysis of laminated plates. Thai et al. [8] further

studied the composite plate using an inverse trigonometric shear deformation theory. Bui et

al. [9] employed an efficient meshfree method for the vibration analysis of laminated

composite plates based on the classic plate theory. Liew et al. [10] explored the application of

meshfree methods for the nonlinear analysis of composite plates using the first-order shear

deformation theory. Liew et al. [11] further extended this approach to the dynamic analysis of

piezo-laminated composite plates.

Isogeometric and meshfree basis functions are coupled using the consistency condition to

2
take advantages of both IGA and meshfree methods [12]. The local refinement strategy for

this approach can be implemented in a direct meshfree manner [13, 14]. This method has

been successfully applied in several engineering domains such as contact mechanics, fracture

mechanics and phase field modeling [15-17]. However, both IGA basis functions and

meshfree shape functions which are computed based on Galerkin method are generally

computational expensive. Generally, a large number of Gauss points are required for the

numerical integration for Galerkin methods, which could be time consuming.

In contrast, the strong form collocation formulation can be alternatively adopted to

reduce the computational cost by reducing the total number of Gauss points [18, 19]. In the

collocation framework, the strong form of the problem is enforced across a set of discrete

collocation points with the same number to that of control points [20, 21]. Although the

computation time per collocation point may be different from that per Gauss point, the

collocation method is still beneficial in terms of a more effective evaluation of the higher-

order shape functions. Ferreira et al. [22] used wavelets collocation method to investigated

the bending and vibration behaviours of composite plates using first order deformation

theory. Pavan and Nanjunda Rao [23] applied isogeometric collocation method on the

bending analysis of isotropic and composite plates. Fantuzzi and Tornabene [24] expanded

the isogeometric collocation method to investigate laminated composite plate in arbitrary

shapes. Recently, an adaptive mesh refinement strategy has been incorporated into the

coupling method to further enhance the computational efficiency [25, 26]. It is because that

such strategy can automatically mark the domain that requires local mesh refinement to

reduce the computational cost by selecting a proper indicator.

The main objective of the present research work is to exploit potential advantages of the

isogeometric meshfree collocation (IMC) approach in the analysis of composite plate

problems. In this way, the computational cost of integration can be reduced through the

3
strong-form formulation. The Reissner–Mindlin plate theory is adopted to govern the

deformation of laminate composite plates. A comprehensive structural analysis is performed

to study the static, free vibration, and bulking problems of both isotropic and composite

plates under various simulation settings, including geometries, boundary conditions and

material properties. The simulation results obtained by the proposed method are compared

with the reference results to demonstrate the robustness and effectiveness of IMMC approach.

An adaptive mesh refinement strategy is further developed based on the gradient of strain

energy across the plate. The area associated with the gradient of strain energy larger than a

threshold could be automatically targeted for the local refinement to generate more accurate

results with less computational cost.

The paper has been organized as follows. A comprehensive introduction of isogeometric-

meshfree coupling approach is firstly introduced in Section 2, followed by a description of

the Reissner–Mindlin theory for laminated composite plates. In Section 3, the numerical

model for bending, free vibration and buckling analyses of laminated composite plates are

explained in details, followed by the description of collocation scheme for the equilibrium

equations in Section 4. A series of numerical examples of isotropic and laminated composite

plates are presented in Section 5. Finally, the paper is closed with conclusions and future

works in Section 6.

2. Isogeometric meshfree coupling approach

This section provides an overview of the IGA basis functions and meshfree shape

functions based on moving least squares (MLS) approximation, followed by a description of

isogeometric meshfree coupling approach. Both IGA basis functions and MLS-based

meshfree shape functions are denoted as the basis functions for convenience.

Based on the reproducing conditions for IGA basis functions, the one-dimensional (1D)

consistency conditions for B-spline are given by

4
 ( )
n
p (ξ ) = ∑ N ip ξi[ ] p (ξ )

(1)
i =1

{ ( ]) }
T

( ) ( )
2 p
Reproducing point vector p ξi[] can be obtained by 1, ξi[1] , ξi[2] ,..., ξi[
p
, where the

reproducing points ξi[ ] ( l = 1, 2,..., p ) are defined as

l

[l ] S lp Gii++1p  p!
=ξi l = and C lp , (2)
Cpl
l !( p − l ) !

where Gii++1p = {ξi+1, ξi+1, …, ξi+p}. The operator S lp [G] indicates l elements from G are

selected and multiplied together as one term. All the possible terms are then summed up.

Likewise, consistency condition in two-dimensional (2D) domain is obtained as

( )
n
p ( ξ ) = ∑ N ipq ( ξ ) p ξ[I ] , (3)


i =1

( )
where p ξ[I] is the 2D reproducing point vector and is computed as

( ) 
{
p T ξ[I ] = 1, ξ I[ ] ,η I[ ] , ξ I[
1 1
( ] ) , ξ [ ]η [ ] , (η [ ] ) ,..., (ξ [ ] )
2
2
I
1
I
1
I
2
2
I
p
p
( ] ) }.
,..., η I[
q
q
(4)

In this way, ξ I[l ] is equal to (ξi +1 + ξi + 2 ) / 2 and ξi +1ξi + 2 for linear and quadratic reproducing

conditions, respectively.

The MLS-based meshfree basis function can be given by

−1

=
m 
Φ J ( ξ ) p ( ξ ) ∑ W
T

 I =1
1 
(
  
ξ − ξ[I ] p ξ[I ] p T ξ[I ]  W

1
)( ) ( )
ξ − ξ[I ] p ξ[I ] ,

( )( ) (5)

( )
 ξ − ξ[1] affects the smoothness properties of MLS-based
where the weight function W I

meshfree shape functions. As an example, a particular weight functions for cubic splines can

be defined by

5
 2 2 3
 3 − 4r + 4r r ≤ 0.5

(
 ξ − ξ[1]
W J )  4 − 4r + 4r 2 − 4 0.5 < r ≤ 1
= (6)
3 3r 3

 0 r >1

where r can be expressed by the support radius rs of the node ξI as | ξ−ξI |/ rs.

Finally, the meshfree reproducing representation of NURBS basis functions can be

expressed by meshfree basis functions Ψ and weights w as follows:

Ψ J (ξ ) w J
SJ (ξ ) = n . (7)
∑ Ψ (ξ ) w
k =1
k k

3. Static, free vibration, and buckling analyses of laminated composite plates

The kinematics of laminated composite plate based on the first-order shear deformation

(FSDT) is outlined in this section. The governing equations for bending, free vibration and

buckling analyses of the laminated composite plates are presented in the collocation

formulations.

3.1. Displacement and strains

According to the work presented by Ferreira et al. [22], the deformation of a

symmetrically laminated composite plate can be divided into in-plane and bending

displacements. Therefore, each node within the plate is assigned with one transverse

displacement w and two rotations θx and θy. Consequently, the displacement components ux,

uy and uz at the mid-plane of a symmetrically laminated plate in x, y and z directions can be

given by ux = zθx, uy = zθy, and uz = w, respectively. In this way, the strains associated with

the displacements can be derived as

∂θ x ∂θ y
=ε xx z= ; ε yy z= ; ε zz 0; (8)
∂x ∂y

6
  ∂θ ∂θ  ∂w ∂w
z  x + y  ; γ yz =+ θ y ; γ xz =+ θ x .
γ xy = (9)
 ∂y ∂x  ∂y ∂x

3.2. Equilibrium equations

The equations of equilibrium for the bending of a plate can thus be given by

∂M xx ∂M xy
+ 0;
− Qx = (10)
∂x ∂y

∂M xy ∂M yy
+ 0;
− Qy = (11)
∂x ∂y

∂Qx ∂Qy
+ − f ( x, y ) ,
= (12)
∂x ∂y

where Mxx and Myy are the bending moments, Mxy is the twisting moment, Qx and Qy are the

in-plane forces, and f(x,y) is the transverse load applied on the plate.

The stress resultants stated in Eqs. (10) to (12) can be expressed in terms of

displacements as

 M xx   D11 D12 D16   ∂θ x ∂x 

  D   
= M yy   12 D22 D26   ∂θ y ∂y  (13)
  D66  ∂θ x ∂y + ∂θ y ∂x 

 M xy   D16 D26

and

Qx   A55 A45   ∂w ∂x + θ x 

  = ks   , (14)
Qy   A45 A44  ∂w ∂y + θ y 

where ks is the shear correction factor which is set to 5/6, and Aij and Dij are the shear and

bending stiffness components, respectively [7].

By considering Eq. (10) to (14), the motion for the bending of plate can be further

derived as follows:

7
 ∂ 2θ ∂ 2θ x ∂ 2θ ∂θ 2
∂ θy ∂θ 2 2

D11 2x + 2 D16 + D66 2x + D16 2y + ( D12 + D16 ) + D26 2y −

∂x ∂xy ∂y ∂x ∂xy ∂y
(15)
  ∂w   ∂w 
ks  A55  + θ x  + A45  + θ y   =
0;
  ∂x   ∂y 

∂ 2θ x ∂ 2θ x ∂ 2θ x ∂ θy 2
∂ θy ∂ θy 2 2

D16 2
+ ( 12 66 )
D + D + D26 2
+ D66 2 + 2 D26 + D22 2 −
∂x ∂xy ∂y ∂x ∂xy ∂y
(16)
  ∂w   ∂w 
ks  A45  + θ x  + A44  + θ y  =
0;
  ∂x   ∂y 

∂    ∂w   ∂w    ∂    ∂w   ∂w   
ks  A55  + θ x  + A45  + θ y    + ks  A45  + θ x  + A44  + θ y   +
∂x    ∂x   ∂y    ∂y    ∂x   ∂y    (17)
f ( x, y ) = 0.

For the free vibration of a laminated composite plate, the equilibrium equations are

defined as

∂M xx ∂M xy
+ − I 2 λ02θ x ;
− Qx = (18)
∂x ∂y

∂M xy ∂M yy
+ − I 2 λ02θ y ;
− Qy = (19)
∂x ∂y

∂Qx ∂Qy
+ − I 0 λ02 w ,
= (20)
∂x ∂y

where λ0 is the frequency of natural vibration of the plate. I0 and I2 in Eqs (18) to (20) are the

mass inertias which can be defined in terms of the density ρ as

h /2 h /2
=I0 ∫ ρ dz and I 2
=
− h /2
∫
− h /2
ρ z 2 dz . (21)

By considering Eqs. (13), (14) and (18) to (20), the harmonic solution of this problem can be

given as follows [27]:

8
 ∂ 2θ ∂ 2θ x ∂ 2θ ∂θ 2
∂ θy ∂θ
2 2

D11 2x + 2 D16 + D66 2x + D16 2y + ( D12 + D16 ) + D26 2y −

∂x ∂xy ∂y ∂x ∂xy ∂y
(22)
  ∂w   ∂w 
ks  A55  + θ x  + A45  + θ y   =
− I 2 λ02θ x ;
  ∂x   ∂y 

∂ 2θ x ∂ 2θ x ∂ 2θ x ∂ θy
2
∂ θy 2
∂ θy 2

D16 2
+ ( 12 66 )
D + D + D26 2
+ D66 2 + 2 D26 + D22 2 −
∂x ∂xy ∂y ∂x ∂xy ∂y
(23)
  ∂w   ∂w 
ks  A45  + θ x  + A44  + θ y  =
− I 2 λ02θ y ;
  ∂x   ∂y 

∂    ∂w   ∂w   
ks  A55  + θ x  + A45  + θ y   +
∂x    ∂x   ∂y   
(24)
∂    ∂w   ∂w   
ks  A45  + θ x  + A44  + θ y   = − I 0 λ02 w .
∂y    ∂x   ∂y   

The equilibrium equations for the buckling analysis are similar to those for the free

vibration analysis. Only the equations for shear forces should be modified as

∂Qx ∂Qy ∂2w ∂2w ∂2w

+ − N xx 2 − 2 N xy
= − N yy 2 , (25)
∂x ∂y ∂x ∂x∂y ∂y

where Nxx, Nxy and Nyy are the in-plane forces.

Through the substitution of Eq. (14) into Eq. (25), the transverse equilibrium should be

reformulated by

∂    ∂w   ∂w    ∂    ∂w   ∂w   
ks  A55  + θ x  + A45  + θ y    + ks  A45  + θ x  + A44  + θ y   +
∂x    ∂x   ∂y    ∂y    ∂x   ∂y   
(26)
∂2w ∂2w ∂2w
N xx 2 + 2 N xy + N yy 2 =0.
∂x ∂x∂y ∂y

3.3. Boundary conditions

The equilibrium equations for structural analysis of laminated composite plates presented

in the previous are only valid within the interior domain of a plate. By denoting direction

cosines of the normal at the boundary as φx and φy, the equations for stress resultants

9
specified at Neumann boundary of a plate can be obtained by

M 
 M nn   ϕ x ϕ y2 2ϕ xϕ y   xx 
2

 = 2
M yy  (27)
 M nt   −ϕ xϕ y ϕ xϕ y ϕ x − ϕ y   M 
2

 xy 

and

Qx 
Q n = ϕ x ϕ y    , (28)
Qy 

where M nn and M nt are the bending moments and Q n is the shear force specified at the

boundary. It should be noted that the subscripts n and t represent the normal and tangential

directions, respectively [28].

The three boundary conditions for an arbitrary edge applied in this work are given by

• Simply supported: w = 0, Mn = 0, and φt = 0;

• Clamped: w = 0, φn = 0, and φt = 0;

• Free: w = 0, Mn = 0, and Mnt = 0.

4. Collocation schemes

In IMC approach, the discretized strong form equations are collocated on the physical

images of the Greville abscissae. The unknown variables mentioned in the previous section

can be given by a linear combination of isogeometric meshfree coupling approach basis

function Φ as

n m
ϕ (ξ=
h
,η ) ∑∑ Φϕ (ξ ,η ) ϕ
=i 1 =j 1
ij
ij
, (29)

where φ denotes the unknown field variables wh, φxh or φ yh .

The discrete approximation in the above equation results in a residual R for the

governing equations of equilibrium and Neumann boundary conditions for the symmetrically

laminated plate. By considering the Eqs. (10) to (14) and residuals together, the residuals for

10
the static bending in the interior domain Ω can be given as

=Ω ∂θ 2
∂ θx
h
∂θ 2 h
∂ θy 2 h
∂ θy 2
∂ θy
h 2 h 2 h

R φx D11 2x + 2 D16 + D66 2x + D16 2 + ( D12 + D16 ) + D26 2 −

∂x ∂xy ∂y ∂x ∂xy ∂y
(30)
  ∂wh   ∂wh 
ks  A55  + θ xh  + A45  + θ yh   ;
  ∂x   ∂y 

 φΩ = D ∂ θ x + ( D + D ) ∂ θ x + D ∂ θ x + D ∂ θ y + 2 D ∂ θ y + D ∂ θ y −
2 h 2 h 2 h 2 h 2 h 2 h

R 16 12 66 26 66 26 22
∂x 2 ∂xy ∂y 2 ∂x 2 ∂xy ∂y 2
y

(31)
  ∂wh h  ∂wh 
ks  A45  + θ x  + A44  + θ yh   ;
  ∂x   ∂y 

 hw ∂    ∂wh h  ∂wh   
=R k A
 s  55  + θ x  + A45  + θ yh    +
∂x    ∂x   ∂y   
(32)
∂    ∂wh h  ∂wh   
k A
 s  45  + θ x  + A44  + θ yh    +f ( x, y ) .
∂y    ∂x   ∂y   

Ω
Similarly, residuals R for the free vibration and buckling analyses can be obtained by
Γ
substituting Eq. (25) into Eqs. (26) to (28). Residual R for the boundary Г can be given

through the substitution of Eq. (25) into Eqs. (23) and (24).

Finally, the residual equations can be collocated by setting residuals R to zero on the

images of Greville points in physical space as follows:

()
Ω
RwΩ 0,=
= RφΩy 0 on S ξ
RφΩx 0,= ; (33)

()
Γ
=RwΓ 0,= RφΓy 0 on S ξ .
RφΓx 0,= (34)

()
where S ξ denotes the mapping of a Greville point from the parametric space to its

corresponding physical space. It should be noted that Eqs. (33) and (34) are derived for the

description of a free plate with full Neumann boundary conditions. If the edges of plate are

subjected to clamped or simply supported boundary conditions, the corresponding degrees of

freedom should be reinforced by Eq. (34) based on Dirichlet boundary conditions.

11
 5. Numerical results

In this section, the bending analysis of an isotropic plate is presented first to validate the

IMC approach. Consequently, structural analysis problems of laminated composite plates are

investigated. Simulation results obtained by the proposed method is compared with the

reference solutions.

4.1. Bending analysis of isotropic plates

To validate the proposed method, a problem that has an available analytical solution is

considered [29]. In this case, a clamped unitary square plate is under a distributed load with a

function f(x,y) given by

=f ( x, y )
E
12 (1 − υ )
2 {
12 y ( y − 1) ( 5 x 2 − 5 x + 1)  2 y 2 ( y − 1) + x ( x − 1) ( 5 y 2 − 5 y + 1)  +

2

(35)


2
}
12 x ( x − 1) ( 5 y 2 − 5 y + 1)  2 x 2 ( x − 1) + y ( y − 1) ( 5 x 2 − 5 x + 1)  ,


where Young’s modulus E and Poisson's ratio υ are set to 10.92 × 106 and 0.3, respectively.

The analytical solution for the displacement along the z-direction is given as

1 3 2h 2  3
x ( x − 1) y ( y − 1) − y ( y − 1) x ( x − 1) ( 5 x 2 − 5 x + 1) +
3 3 3 3
w( x,=
y)
3 5 (1 − υ ) 
(36)
x ( x − 1) y ( y − 1) ( 5 y − 5 y + 1)  ,
3 3 2


where h is the thickness of the plate.

The problem is solved with equal orders, which means that the same polynomial orders

are applied for displacement and rotations. The validation study is conducted through the

comparison of convergence rates of the proposed method, IGA collocation method and IGA

Galerkin method [30]. The relative error in L2 norm of the displacement is plotted against the

square root of the degrees of freedom. The convergence rate of each approach is

approximated as the slope of each line m . Fig. 1 shows that collocation method generally can

offer a higher convergence rate as compared to Galerkin method. The proposed method is

12
able to provide more accurate simulation results than IGA collocation method. It even can

achieve higher accuracy than IGA Galerkin method when the mesh is sufficiently refined.

This investigation validates the robustness and effectiveness of the IMC approach for the

structural analysis of plates.

Fig. 1. Relative error in L2 norm approximation of displacement w for a clamped square

isotropic plate with a thickness-to-width ratio of 0.1.

To explore the advantage of local refinement flexibility, an adaptive mesh refinement

strategy is implemented to evaluate the bending moments M of the same isotropic plate. The

exact solution for M is further given by [30]

Et 3  3
y ( y − 1) x (1 − x ) ( 5 x 2 − 5 x + 1) +
3
M xx ( x, y ) =−
6 (1 − υ )
2 
(37)
υ x3 ( x − 1) y (1 − y ) ( 5 y 2 − 5 y + 1)  ;
3


13
 Et 3  3
υ y ( y − 1) x (1 − x ) ( 5 x 2 − 5 x + 1) +
3
M yy ( x, y ) =−
6 (1 − υ )
2 
(38)
x ( x − 1) y (1 − y ) ( 5 y − 5 y + 1)  ,
3 3 2


Et 3
y 2 ( y − 1) ( 2 y − 1) x 2 ( x − 1) ( 2 x − 1) .
2 2
M xy ( x, y ) =
− (39)
4 (1 + υ )

An indicator is introduced to automatically identify the area associated with gradient of strain

energy exceeding a threshold. Fig. 2 compares the difference between the final mesh

generated by the adaptively and globally uniform refined mesh strategies, which further

demonstrates that the proposed strategy can save the computational cost by avoiding the fine

mesh across the whole domain. As shown in Fig. 3, the area processing a large gradient of

strain energy is automatically identified by the indicator and refined by the IMC-based

computational algorithm.

(a) (b)

Fig. 2. (a) The globally uniform refined mesh and (b) adaptive refined mesh after 4 steps of
refinement.

14
 Fig. 3. Displacement of a clamped isotropic plate with a thickness-to-width ratio of 0.1.

The comparison between the analytical solution and numerical results obtained by

adaptively and globally uniform refined meshes are illustrated in Fig. 4. The relative error in

E2 norm approximation error of the bending moments M is plotted against the square root of

the degrees of freedom. The convergence rate of each approach is approximated as the slope

of each line m . As expected, the adaptive refinement strategy assists the proposed approach

in achieving more accurate simulation results and higher convergence rate when compared to

the globally uniform mesh.

15
Fig. 4. Relative error in E2 norm approximation of bending moments M for a clamped square
isotropic plate with a thickness-to-width ratio of 0.1.

4.2. Bending analysis of laminated composite plates

After the proposed approach has been validated by the bending analysis of an isotropic

plate, a square laminated plate formed of four plies with ply orientations [0°/90°/90°/0°].

Each pile has the same thickness which is one fourth of the total thickness h. The

orthotropic material properties for the plate are given as E1 = 25E2, G12 = G13 = 0.5E2, G23 =

0.2E2, and υ12 = 0.25, where G is the shear modulus. As shown in Fig. 5, the plate is

πx  π y 
subjected to a sinusoidal load f of the form f ( x, y ) = f 0 sin   sin   , where f0 is the
 a   a 

amplitude of the loading, and a is the length of plate edge. Four thickness-to-width ratios

h/a are investigated in this case which are 0.25, 0.1, 0.05, and 0.01. The displacement and

stresses are normalized in the following expressions as

16
 3 2 2
 a a  E2 h 0 a a h h 0 a a h h
=w w0 = ,  4 ; σ xx σ=
xx  , ,  2 ; σ yy σ yy  , ,  2 ;
 2 2  a f0  2 2 2  a f0  2 2 4  a f0
2
(40)
a a h h  b  h
σ xy =
σ  , , −  2 ; τ xz =
0
xy τ xz0  a, , k =
1,3  ,
 2 2 2  a f0  2  af 0

where k is the order of ply counted from bottom to top, w is the displacement, σ is the normal

stress and τ is the shear stress.

Fig. 5. Schematic diagram of a laminated plate under a sinusoidal loading.

The results achieved by the proposed approach and reference solutions are summarized

in Table 1. The distribution of stresses σxx and σyy vertically through the thickness of the

plate is illustrated in Fig. 6. The effect of higher order polynomials has been shown by

considering three different polynomial orders two, three, and four. It can be found that the

displacement and stresses achieve good values and agree well with the reference solutions

given by Pavan et al. [23] and Thai et al. [7]. The difference observed between the present

results and the values presented by Pagano et al. [31], Ferreira et al. [22] could be attributed

to the difference in the plate theories considered.

17
Table 1. Displacement and stresses of the simply supported composite plate.

h/a Method w σxx σyy σxy 𝜏𝜏xz

Elasticity [31] 1.9540 0.7200 0.6660 0.0467 0.2700
Meshfree collocation [22] 1.9075 0.6432 0.6228 0.0433 0.2166
IGA collocation [23] 1.7315 0.4173 0.5871 0.0218 0.1455
IGA Galerkin [7] 1.7164 0.4055 0.5758 0.0308 0.1399
0.25
IGA-MF coupling Galerkin 1.7140 0.4060 0.5765 0.0309 0.1400
Present method (p=2) 1.7115 0.4063 0.5769 0.0308 0.1398
Present method (p=3) 1.7120 0.4063 0.5769 0.0308 0.1399
Present method (p=4) 1.7141 0.4061 0.5776 0.0307 0.1400

Elasticity [31] 0.7430 0.559 0.4030 0.0276 0.3010

Meshfree collocation [22] 0.6627 0.4989 0.3614 0.0241 0.3181
IGA collocation [23] 0.6716 0.5062 0.3652 0.0204 0.1746
IGA Galerkin [7] 0.6654 0.4983 0.3610 0.0242 0.1669
0.1
IGA-MF coupling Galerkin 0.6644 0.4989 0.3614 0.0238 0.1657
Present method (p=2) 0.6641 0.4997 0.3620 0.0241 0.1665
Present method (p=3) 0.6643 0.4997 0.3620 0.0242 0.1668
Present method (p=4) 0.6700 0.4979 0.3603 0.0280 0.1691

Elasticity [31] 0.5170 0.5430 0.3090 0.0230 0.3280

Meshfree collocation [22] 0.4912 0.5273 0.2956 0.0221 0.3332
IGA collocation [23] 0.4960 0.5321 0.2974 0.0206 0.1828
IGA Galerkin [7] 0.4931 0.5268 0.2953 0.0221 0.1755
0.05
IGA-MF coupling Galerkin 0.4924 0.5274 0.2963 0.0221 0.1744
Present method (p=2) 0.4912 0.5275 0.2958 0.0221 0.1749
Present method (p=3) 0.4927 0.5290 0.2978 0.0219 0.1720
Present method (p=4) 0.4916 0.5278 0.2958 0.0220 0.1741

Elasticity [31] 0.4347 0.5390 0.2710 0.0214 0.3390

Meshfree collocation [22] 0.4335 0.5381 0.2704 0.0213 0.3390
IGA collocation [23] 0.4350 0.5397 0.2710 0.0209 0.1860
IGA Galerkin [7] 0.4354 0.5376 0.2702 0.0213 0.1907
0.01
IGA-MF coupling Galerkin 0.4348 0.5382 0.2705 0.0213 0.1940
Present method (p=2) 0.4116 0.5108 0.2567 0.0202 0.1689
Present method (p=3) 0.4144 0.5144 0.2587 0.0203 0.1736
Present method (p=4) 0.4341 0.5373 0.2703 0.0216 0.1977

18
 (a) (b)

Fig. 6. Through thickness distribution of stresses for the composite plate: (a) σxx and (b) σyy.

The bending analysis of a laminated composite plate with all edges clamped boundary

conditions is also conducted in this section. A uniform load is applied transversely across the

plate. The plate is formed by one orthotropic core sheet and two transversely isotropic face

sheets. The core sheet has a thickness of 0.8h and its material properties are given by E1 =

25E2, G12 = G13 = 0.5E2, G23 = 0.2E2, and υ12 = 0.25. The face sheet has a thickness of 0.1h

and processes material properties as E1 = E2, G12 = G23 = 0.06E2, G12 = 0.016E2, and υ12 = 0.2.

Three thickness-to-width ratios h/a are investigated in this case which are 0.1, 0.05, and 0.01.

It is observed that numerical results obtained by the proposed approach are also in good

agreement with the reference solutions given by the collocation methods. The difference in

the results between FEM and the proposed method could be attributed to the difference

between weak form and strong form formulations.

19
Table 2. Displacement and stresses of the clamped composite plate.

h/a Method w σxx σyy σxy 𝜏𝜏xz

FEM [32] 1.2654 0.5018 0.0550 0.0120 0.2318
IGA collocation [34] 1.2056 0.4302 0.0330 0.0066 0.2472
IGA-MF coupling Galerkin 1.2074 0.4624 0.0505 0.0129 0.2416
0.1
Present method (p=2) 1.2052 0.4622 0.0505 0.0061 0.2395
Present method (p=3) 1.2053 0.4619 0.0505 0.0063 0.2395
Present method (p=4) 1.2050 0.4617 0.0503 0.0059 0.2393

FEM [32] 0.3111 0.5347 0.0094 0.003 0.2400

IGA collocation [34] 0.3070 0.5246 0.0083 0.0017 0.2804
IGA-MF coupling Galerkin 0.3015 0.5195 0.0160 0.0013 0.2648
0.05
Present method (p=2) 0.3071 0.5301 0.0164 0.0014 0.2611
Present method (p=3) 0.3071 0.5298 0.0163 0.0015 0.2610
Present method (p=4) 0.3071 0.5314 0.0161 0.0014 0.2611

FEM [32] 0.2785 0.5347 0.0094 0.003 0.2400

IGA collocation [34] 0.2724 0.5261 0.0072 0.0012 0.2805
IGA-MF coupling Galerkin 0.2703 0.5217 0.0142 0.0029 0.2859
0.01
Present method (p=2) 0.2715 0.5215 0.0143 0.0029 0.2560
Present method (p=3) 0.2725 0.5332 0.0120 0.0028 0.2447
Present method (p=4) 0.2725 0.5332 0.0120 0.0028 0.2447

4.3. Free vibration of laminated composite plates

In this section, the free vibration behavior of a square laminated composite plate is

investigated using IMC approach. All piles of the plate have the identical thickness and

density ρ. Each of these layers is assigned with the dimensionless high-modulus graphite-

epoxy material properties as G12 = G13 = 0.6E2, G23 = 0.5E2, and υ12 = 0.25. Similar to the

first case, the plate consists of 4 plies with the configuration [0°/90°/90°/0°] one. The

Young’s modulus ratio E1/E2 vary from 10 to 40 for the simulation, while the thickness-to-

span ratio h/a for the plate is fixed to 0.2. The frequency can be further normalized as

λn = ( λ0 a 2 / h ) ρ / E2 .

20
Table 3. Normalized natural frequency λn of the simply supported composite plate.

E1/E2
Method
10 20 30 40
Meshfree Galerkin [28] 8.2924 9.5613 10.3200 10.8490
Meshfree collocation [22] 8.2793 9.5375 10.2889 10.8117
Analytical solution [33] 8.2982 9.5671 10.3260 10.8540
IGA-MF coupling Galerkin 8.2983 9.5673 10.3260 10.8541
Present method (p=2) 8.2923 9.5559 10.3152 10.8491
Present method (p=3) 8.2908 9.5539 10.3156 10.8436
Present method (p=4) 8.2999 9.5687 10.3275 10.8546

The simulation results are compared with the analytical solutions by Khdeir and

Librescu [33], meshfree Galerkin solutions by Liew et al. [28], and meshfree collocaiton

results by Ferreira et al. [34]. The simulation results obtained by the proposed approach

match well with both reference and analytical results. The accuracy of the simulation can be

improved by using IMC approximation with high polynomial order. The relative errors

between the analytical and present results are even less than those between the analytical

and IGA collocation solutions. The first four vibration modes of a plate with h/a = 0.1 are

depicted in Fig. 7, which demonstrates that good geometry exactness can be also preserved

by the proposed approach.

Next, the free vibration analysis is performed on a clamped three-layer laminate

[0°/90°/0°] plate with a modulus ratio E1/E2 = 40. The frequency λ0 is further normalized as

Eh3
λn = ( λ0 a 2 / h ) ρ / D0 , where parameter D0 is equal to . Four thickness-to-
12 (1 − υ12υ21 )

width ratios h/a are investigated in this case which are 0.2, 0.1, 0.05, and 0.01. The

numerical results are found to be close to reference solutions. The natural vibration

frequency generally increases as thickness-to-span ratio h/a of the plate decreases.

21
Table 4. Normalized natural frequency λn of the clamped composite plate.

h/a Method Mode I Mode II Mode III Mode IV

Meshfree Galerkin [28] 4.4470 6.6420 7.7000 9.1850
Meshfree collocation [22] 4.4466 6.6419 7.6996 9.1852
Global-local theory [35] 4.4500 6.5240 8.1780 9.4730
0.2 IGA-MF coupling Galerkin 4.6878 6.9441 8.0878 9.6397
Present method (p=2) 4.4706 6.6804 7.7504 9.2405
Present method (p=3) 4.4583 6.6342 7.7172 9.1845
Present method (p=4) 4.4695 6.6738 7.7447 9.2349

Meshfree Galerkin [28] 7.411 10.393 13.913 15.429

Meshfree collocation [22] 7.4108 10.3928 13.9129 15.4292
Global-local theory [35] 7.484 10.207 14.340 14.863
0.1 IGA-MF coupling Galerkin 7.6769 10.7437 14.6335 16.0416
Present method (p=2) 7.4455 10.4422 13.9999 15.5360
Present method (p=3) 7.4430 10.4215 13.9845 15.4800
Present method (p=4) 7.4434 10.4274 13.9874 15.5951

Meshfree method [28] 10.9530 14.0280 20.3880 23.1960

Meshfree collocation [22] 10.9529 14.0279 20.3904 23.1960
Global-local theory [35] 11.0030 14.0640 20.3210 23.4980
0.05 IGA-MF coupling Galerkin 11.2437 14.3953 21.0065 24.2874
Present method (p=2) 10.9873 14.0787 20.5382 23.3236
Present method (p=3) 10.9833 14.0501 20.4039 23.2887
Present method (p=4) 10.9811 14.0881 20.8456 23.0700

Meshfree method [28] 14.6660 17.6140 24.5110 35.5320

Meshfree collocation [22] 14.4342 17.3942 24.3148 35.4087
Global-local theory [35] 14.6010 17.8120 25.2360 37.1680
0.01 IGA-MF coupling Galerkin 14.9576 18.2859 26.0397 35.0712
Present method (p=2) 14.5192 17.4420 24.3716 35.4471
Present method (p=3) 14.4111 17.4226 24.3921 35.2763
Present method (p=4) 14.4672 17.2249 24.4574 36.7353

22
 Mode I Mode II

Mode III Mode IV

Fig. 7. The first four vibration modes of the clamped [0°/90°/90°/0°] laminated composite
plate when E1/E2 and h/a are 40 and 0.2, respectively.

4.4. Buckling of laminated composite plates

Finally, the effect of Young’s modulus ratio E1/E2 on the critical buckling loads of the

composite plates is investigated in this section. The plate has a thickness-to-span ratio h/a of

0.1 and is subjected to simply-supported boundary condition. A pair of uniaxial load Nxx is

applied on the plate along the x axis as shown in Fig 8. Based on this setting, the normalized

critical buckling load factor λc is defined as λc = Nca2/(E2h3), where Nc is the eigenvalue or the

critical load to be evaluated.

23
 Fig. 8. Schematic of the square laminated composite plate subjected to uniaxial loading Nxx.

The dimensionless high-modulus graphite-epoxy material properties are also adopted in

this section for the analysis. The plate is assumed to have three layers of plies with the fiber

configuration of [0°/90°/0°]. Table 5 lists the critical buckling loads for the modulus ratio

E1/E2 varying from 3 to 40. The numerical results match well the reference solutions. The

first four buckling modes of the plate with modulus E1/E2 = 10 are shown in Fig. 9.

Table 5. Normalized critical buckling load factor λc of the simply supported composite plate.

E1/E2
Method
3 10 20 30 40
Elasticity [36] 5.3044 9.7621 15.0191 19.3040 23.8807
FEM [37] 5.3933 9.9406 15.2980 19.6740 22.3400
Meshfree method [38] 5.3869 9.8601 14.9746 19.0175 22.3070
Present method (p=2) 5.4764 9.9518 15.0646 19.1051 22.3920
Present method (p=3) 5.3108 9.7715 14.9000 18.9341 22.2245
Present method (p=4) 5.3958 9.8707 14.9842 19.0261 22.3145

24
 Mode I Mode II

Mode III Mode IV

Fig. 9. The first four buckling modes of the simply-supported [0°/90°/0°] laminated
composite plate when E1/E2 and h/a are 10 and 0.1, respectively.

6. Conclusions

In this study, a isogeometric meshfree coupling approach have been incorporated with

the collocation method for the static, free vibration, and buckling analyses of symmetrically

laminated composite plates. The equilibrium equations for the plate is derived from a strong

form to save computational cost by avoiding the numerical integration. The robustness and

effectiveness of the proposed method have been validated by several benchmark problems.

Although the classic Galerkin method can offer more accurate results than the collocation

approach, collocation approach generally achieves higher convergence rate. Such

convergence rate can be further improved through the adaptive refinement strategy. The

25
highlights of this paper are summarized as follows:

• Isogeometric meshfree coupling approach is integrated with the collocation approach to

reduce the computational cost by reducing the points for integration required by the

classic Galerkin methods;

• The developed IMC approach is able to generate more accurate numerical result with a

higher convergence rate that can be further improved by an adaptive mesh refinement

strategy based on the gradient rate of strain energy across the plate;

• Effects of various material properties and geometric parameters on the mechanical

performance of laminated composite plates, including their static, free vibration, and

buckling behaviors, is thoroughly investigated by the proposed approach.

It can be concluded that the proposed approach can generate more accurate simulation

results with a similar convergence rate when compared to its IGA collocation counterpart.

IMC approach can even achieve high accuracy close to the one provided by Galerkin method

when the mesh is sufficiently refined. The application of the proposed approach on the

structural analysis of composite plate with defects such as cracks and inclusions will be

explored in future work.

Acknowledgements

This research work was conducted in the SMRT-NTU Smart Urban Rail Corporate

Laboratory with funding support from the National Research Foundation (NRF), Singapore,

SMRT, Singapore and Nanyang Technological University, Singapore.

26
References

[1] C. W. Bert, M. Malik. Differential quadrature method in computational mechanics: A

review, Applied Mechanics Reviews, 1996. 49(1): pp. 1–28.

[2] E. Ramm. Finite element procedures in engineering analysis, k-j. Bathe, prentice-hall,
1982. No. Of pages: 735, International Journal for Numerical and Analytical
Methods in Geomechanics, 1983. 7(4): pp. 500–500.

[3] A. R. Setoodeh, G. Karami. Static, free vibration and buckling analysis of anisotropic
thick laminated composite plates on distributed and point elastic supports using a 3-d
layer-wise fem, Engineering Structures, 2004. 26(2): pp. 211–220.

[4] M. Gherlone, M. Di Sciuva. Thermo-mechanics of undamaged and damaged

multilayered composite plates: Assessment of the fem sub-laminates approach,
Composite Structures, 2007. 81(1): pp. 137–155.

[5] K. M. Liew, X. Zhao, A. J. M. Ferreira. A review of meshless methods for laminated

and functionally graded plates and shells, Composite Structures, 2011. 93(8): pp.
2031–2041.

[6] S. Shojaee, N. Valizadeh, E. Izadpanah, T. Bui, T.-V. Vu. Free vibration and buckling
analysis of laminated composite plates using the nurbs-based isogeometric finite
element method, Composite Structures, 2012. 94(5): pp. 1677–1693.

[7] C. H. Thai, H. Nguyen-Xuan, N. Nguyen-Thanh, T. H. Le, T. Nguyen-Thoi, 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, 2012. 91(6): pp. 571–603.

[8] C. H. Thai, A. J. M. Ferreira, S. P. A. Bordas, T. Rabczuk, H. Nguyen-Xuan.

Isogeometric analysis of laminated composite and sandwich plates using a new
inverse trigonometric shear deformation theory, European Journal of Mechanics-
A/Solids, 2014. 43: pp. 89–108.

[9] T. Q. Bui, M. N. Nguyen, C. Zhang. An efficient meshfree method for vibration

analysis of laminated composite plates, Computational Mechanics, 2011. 48(2): pp.
175–193.

[10] K. M. Liew, J. Wang, M. J. Tan, S. Rajendran. Nonlinear analysis of laminated

composite plates using the mesh-free kp-ritz method based on fsdt, Computer
Methods in Applied Mechanics and Engineering, 2004. 193(45): pp. 4763–4779.

[11] K. M. Liew, X. Q. He, M. J. Tan, H. K. Lim. Dynamic analysis of laminated

composite plates with piezoelectric sensor/actuator patches using the fsdt mesh-free
method, International Journal of Mechanical Sciences, 2004. 46(3): pp. 411–431.

[12] H. Zhang, D. Wang. Reproducing kernel formulation of b-spline and nurbs basis
functions: A meshfree local refinement strategy for isogeometric analysis, Computer
Methods in Applied Mechanics and Engineering, 2017. 320: pp. 474–508.

27
[13] D. Wang, H. Zhang. A consistently coupled isogeometric–meshfree method,
Computer Methods in Applied Mechanics and Engineering, 2014. 268(Complete): pp.
843–870.

[14] H. Zhang, D. Wang, W. Liu. Isogeometric meshfree coupled analysis of kirchhoff

plates, Advances in Structural Engineering, 2014. 17(8): pp. 1159–1176.

[15] R. Kiran, N. Nguyen-Thanh, J. Huang, K. Zhou. Buckling analysis of cracked

orthotropic 3d plates and shells via an isogeometric-reproducing kernel particle
method, Theoretical and Applied Fracture Mechanics, 2021. 114: pp. 102993-1–20.

[16] W. Li, N. Nguyen-Thanh, K. Zhou. Phase-field modeling of brittle fracture in a 3d

polycrystalline material via an adaptive isogeometric meshfree approach,
International Journal for Numerical Methods in Engineering, 2020. 121(22): pp.
5042–5065.

[17] N. Nguyen-Thanh, W. Li, J. Huang, K. Zhou. Adaptive higher-order phase-field

modeling of anisotropic brittle fracture in 3D polycrystalline materials, Computer
Methods in Applied Mechanics and Engineering, 2020. 372: pp. 113434-1–31.

[18] D. Schillinger, J. A. Evans, A. Reali, M. A. Scott, T. J. R. Hughes. Isogeometric

collocation: Cost comparison with galerkin methods and extension to adaptive
hierarchical nurbs discretizations, Computer Methods in Applied Mechanics and
Engineering, 2013. 267: pp. 170–232.

[19] D. Schillinger, S. J. Hossain, T. J. R. Hughes. Reduced bézier element quadrature

rules for quadratic and cubic splines in isogeometric analysis, Computer Methods in
Applied Mechanics and Engineering, 2014. 277: pp. 1–45.

[20] C. M. C. Roque, P. Martins. Differential evolution optimization for the analysis of

composite plates with radial basis collocation meshless method, Composite
Structures, 2015. 124: pp. 317–326.

[21] C.-P. Wu, K.-H. Chiu, Y.-M. Wang. Rmvt-based meshless collocation and element-
free galerkin methods for the quasi-3d analysis of multilayered composite and fgm
plates, Composite Structures, 2011. 93(2): pp. 923–943.

[22] A. J. M. Ferreira, L. M. S. Castro, S. Bertoluzza. A high order collocation method for

the static and vibration analysis of composite plates using a first-order theory,
Composite Structures, 2009. 89(3): pp. 424–432.

[23] G. S. Pavan, K. S. Nanjunda Rao. Bending analysis of laminated composite plates

using isogeometric collocation method, Composite Structures, 2017. 176: pp. 715–
728.

[24] N. Fantuzzi, F. Tornabene. Strong formulation isogeometric analysis (SFIGA) for

laminated composite arbitrarily shaped plates, Composites Part B: Engineering, 2016.
96: pp. 173–203.

28
[25] W. Li, N. Nguyen-Thanh, K. Zhou. An isogeometric meshfree collocation approach
for two-dimensional elastic fracture problems with contact loading, Engineering
Fracture Mechanics, 2020. 223: pp. 106779-1–19.

[26] N. Nguyen-Thanh, W. Li, J. Huang, N. Srikanth, K. Zhou. An adaptive isogeometric

analysis meshfree collocation method for elasticity and frictional contact problems,
International Journal for Numerical Methods in Engineering, 2019. 120(2): pp. 209–
230.

[27] J. N. Reddy. Theory and analysis of laminated composite plates. In: C. A. M. Soares,
C. M. M. Soares, M. J. M. Freitas, editors. Mechanics of composite materials and
structures. Dordrecht: Springer Netherlands; 1999. pp. 1–79.

[28] K. M. Liew, Y. Q. Huang, 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, 2003. 192(19): pp. 2203–
2222.

[29] C. Chinosi, C. Lovadina. Numerical analysis of some mixed finite element methods
for reissner-mindlin plates, Computational Mechanics, 1995. 16(1): pp. 36–44.

[30] J. Kiendl, F. Auricchio, L. Beirão da Veiga, C. Lovadina, A. Reali. Isogeometric

collocation methods for the reissner–mindlin plate problem, Computer Methods in
Applied Mechanics and Engineering, 2015. 284: pp. 489–507.

[31] N. J. Pagano. Exact solutions for rectangular bidirectional composites and sandwich
plates, Journal of Composite Materials, 1970. 4(1): pp. 20–34.

[32] J. N. Reddy. Mechanics of laminated composite plates and shells: Theory and
analysis, second edition: CRC Press; 2003.

[33] A. A. Khdeir, L. Librescu. Analysis of symmetric cross-ply laminated elastic plates

using a higher-order theory. II - buckling and free vibration, Composite Structures,
1988. 9: pp. 259–277.

[34] G. S. Pavan, N. K. S. Rao. Analysis of laminated composite plates using isogeometric

collocation method. VII European Congress on Computational Methods in Applied
Sciences and Engineering, Crete Island, Greece. 2016. pp. 1–17.

[35] Z. Wu, W. Chen. Free vibration of laminated composite and sandwich plates using
global-local higher-order theory, Journal of Sound and Vibration, 2006. 298: pp. 333–
349.

[36] A. K. Noor. Free vibrations of multilayered composite plates, AIAA Journal, 1973.
11(7): pp. 1038–1039.

[37] N. S. Putcha, J. N. Reddy. Stability and natural vibration analysis of laminated plates
by using a mixed element based on a refined plate theory, Journal of Sound and
Vibration, 1986. 104(2): pp. 285–300.

29
 [38] A. Ferreira, L. Castro, C. Roque, J. N. Reddy, S. Bertoluzza. Buckling analysis of
laminated plates by wavelets, Computers & Structures, 2011. 89: pp. 626–630.

30

View publication stats