Rapport

Simplified Approaches to
Buckling of Composite Plates

by
Qiao Jie Yang
THESIS
for the degree of
MASTER OF SCIENCE
(Master i Anvendt matematikk og mekanikk)
Faculty of Mathematics and Natural Science

University of Oslo
May 2009
Det matematisk- naturvitenskapelige fakultet
Universitetet i Oslo
Preface
This thesis has been written to fulfill the degree of Master of Science at the
University of Oslo, Department of Mathematics, Mechanics Division.
I would like to thank Professor Brian Hayman at the University of Oslo and Det
Norske Veritas (DNV), who has been my supervisor during this project. His
knowledge and support is highly appreciated. Then I want to thank Dr. Scient.
student Henrik Mathias Eiding at the University of Oslo for his patient assistance
related to ANSYS programming.
Abstract
Composite structures consisting of plates or plate-like elements are used widely in
windturbine blades and in certain types of ships, particularly naval ships. These
structual elements are often subjected to significant forces. Buckling analyses
are often conducted by FE analyses. But sometimes these analyses are really
complex and make heavy demands on both computer resources and the analysts
expertise. There is a need for simplified but reliable analysis methods.
Both CLPT and FSDT have been applied to the estimation of elastic critical
loads for plates. Thus, the method includes out-of-plane shear deformation.
Further, the method is developed for plates subjected to uniaxial compression
load, both simply supported and clamped edges have been studied. The analysis
method will also cover cases with in-plane biaxial compression, in-plane shear
loading and combined loadings. These are confined to plates with simply supported edges. The case of a plate having an initial geometric imperfection will also
be invstigated and it is been tried to establish the onset of first ply failure.
To validate the methods, FE analyses is performed using ANSYS.
The methods based on FSDT give a better estimation than CLPT. It is best
suited for thin and moderately thick plates. Higher order deformation theories
should be considered for really thick plates. In addition, the methods are limited
to linear cases.
Contents
1 Introduction
1.1
Background and Motivation . . . . . . . . . . . . . . . . . . . . .
1.2
Problem Definition . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3
Contents of The Thesis . . . . . . . . . . . . . . . . . . . . . . . .
2 Classical and First-Order Theories of Composite Plates
2.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2
The Classical Laminated Plate Theory, CLPT . . . . . . . . . . .
2.2.1
Kinematics . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.2
The Material Law . . . . . . . . . . . . . . . . . . . . . . .
2.2.3
Resultant Forces and Moments . . . . . . . . . . . . . . .
2.2.4
Equilibrium Equations in Terms of Displacements . . . . .
The First-Order Shear Deformation Theory, FSDT . . . . . . . .
2.3.1
Kinematics . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.2
The Material Law . . . . . . . . . . . . . . . . . . . . . . .
10
2.3.3
Resultant Forces and Moments . . . . . . . . . . . . . . .
10
2.3.4
Equilibrium Equations in Terms of Displacements . . . . .
11
2.3
3 Analysis of Specially Orthotropic Plates Using CLPT

3.1
3.2
Buckling of Simply Supported Plates under Uniaxial Compressive

Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13
Buckling of Clamped Plates under Uniaxial Compressive Load . .
14
4 Analysis of Specially Orthotropic Plates Using FSDT

4.1
13
Buckling of Simply Supported Plates under Uniaxial Compressive

Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
17
4.2
Buckling of Clamped Plates under Uniaxial Compressive Load . .
19
4.3
Buckling of Simply Supported Plates under Biaxial Compressive

Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
23
4.4
Buckling of Simply Supported Plates under In-plane Shear Load .
25
4.5
Buckling of Simply Supported Plates under Combined Loads . . .
30
4.6
Plates with Initial Geometric Imperfection under Uniaxial Compressive Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
33
4.6.1
Relationship Between Displacement and Applied Load . .
33
4.6.2
First Ply Failure . . . . . . . . . . . . . . . . . . . . . . .
35
5 Finite Element Model
39
6 Presentation of Results
43
6.1
Uniaxial Compressive Load, CLPT and FSDT . . . . . . . . . . .
43
6.1.1
Simply Supported Plates . . . . . . . . . . . . . . . . . . .
43
6.1.2
Clamped Plates . . . . . . . . . . . . . . . . . . . . . . . .
52
6.2
Biaxial Compressive Load, FSDT . . . . . . . . . . . . . . . . . .
59
6.3
In-plane Shear Load, FSDT . . . . . . . . . . . . . . . . . . . . .
69
6.4
Combined Load , FSDT . . . . . . . . . . . . . . . . . . . . . . .
75
6.5
Plates with Initial Geometric Imperfection, FSDT . . . . . . . . .
82
6.5.1
Displacement - Applied Load . . . . . . . . . . . . . . . .
82
6.5.2
First Ply Failure - Tsai-Wu . . . . . . . . . . . . . . . . . .
86
7 Conclusion
93
7.1
Conclusion of Results . . . . . . . . . . . . . . . . . . . . . . . . .
93
7.2
Suggestions for Futher Work . . . . . . . . . . . . . . . . . . . . .
95
A Parameter Definitions [9]
97
A.1 CASE A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
97
A.2 CASE B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
98
A.3 CASE C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
99
B Definitions and Equations
100
B.1 Matrixes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

B.2 Governing Equations for Plates . . . . . . . . . . . . . . . . . . . 102
C Program Codes
105
C.1 Matlab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

C.2 ANSYS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
1
1.1
Introduction
Background and Motivation
Figure 1.1: Windturbines and the norwegian naval ship, Fridtjof Nansen.
Composite structures consisting of plates or plate-like elements are used widely
in windturbine blades and in certain types of ships, particularly naval ships.
These structual elements are often subjected to significant forces such as inplane compressive or shear loading. So understanding and proper application of
composite materials have helped to influence the lifetime and stability of these
constructions. Thus in the design context buckling analysis plays a crucial role.
Buckling analysis or parametric studies are often conducted by FE analyses. But
sometimes these FE analyses are quite complex and make heavy demands on both
computer resources and the analysts expertise. There is a need for simplified but
reliable analysis methods that can readily be used for parametric studies.
1
1.2
Problem Definition
The general aim of the thesis will be the investigation of simplified approaches to
the estimation of failure loads for composite plates under in-plane loadings. The
specific objectives will be as follows:
For a selected number of laminate lay-ups and plate aspect ratios, uniaxial critical loads having different thicknesses shall be estimated using the
Matlab routines developed in the pre-project and also, in some cases, the
ANSYS FE software. The analysis includes simply supported and clamped
edges. Both Classical Laminated Plate Theory (CLPT) and First-order
Shear Deformation Theory (FSDT) shall be applied. The results presented in a grathical (preferably non-dimensional) format and in the form of
tables.
To extend the analysis method to cover cases with in-plane biaxial compression and in-plane shear loading. Only simply supported edges using FSDT
shall be applied. A limited parametric study shall be performed for each
of these loading cases and the results presented graphically and in form of
tables.
The method shall be extended to analyse cases with combined compression
and shear loadings. This will be confined to shear combined with uniaxial
compression loading and to the case of simply supported plates. A limited
parametric study shall be performed and the results shall be presented in
one or more interaction diagrams.
To extend the approach to analyse the plates with initial out-of-flatness. For
uniaxial compression loading, the case of a plate having an initial geometric
imperfection in the form of the first buckling mode shall be analysed such
that the deflections and stresses can be estimated for increasing values
of applied loading. This will be confined to small-deflection (linearised)
buckling theory and plates with simply supported edges, but will include
out-of-plane shear deformation (FSDT). Corresponding FE analyses shall
be performed using ANSYS for some selected cases to validate the method.
Using the imperfection analysis, a suitable material failure criterion shall
be applied to find the value of applied load at which material failure first
occurs (first ply failure). A limited parametric study shall be performed
for square, simply supported plates with a selected lay-up type and the
results compared with those from the ongoing studies in the MARSTRUCT
Network of Excellence.
2
1.3
Contents of The Thesis
Section 2 gives a presentation of theories related to buckling analysis of composite plates. Both classical laminated plate theory (CLPT) and first-order shear
deformation laminated plate theory would be briefly reviewed.
Section 3 deals with the analysis of specially orthotropic laminates using the
CLPT. Both simply supported and clamped plates subjected to uniaxial compressive load will be investigated.
Analysis of specially orthotropic laminates based on the FSDT is devoted to section 4. Here analytical solutions are developed for simply supported and clamped
plates with uniaxial compression. The analysis method will also cover the cases
with in-plane biaxial compression, in-plane shear loading and combined loadings
related to simply supported plates. Further, the investigation is concerned about
plates with an initial geometric imperfection.
Section 5 deals with finite element analysis to validate the present method.
Section 6 contains the results from the analysis. The critical buckling loads are
estimated using Matlab. Corresponding FE analyses have been performed using
ANSYS. The results will be presented both graphically and in tables.
Section 7 contains conclusion and suggestions for further work.
Material properties for a selected number of laminate lay-ups are listed in Appendix A, while Appendix B gives a presentation of some useful expressions and
deduction of buckling equations. A part of Matlab, Maple and ANSYS codes are
listed in Appendix C.
Classical and First-Order Theories of Composite Plates
2.1
Introduction
According to Reddy [1], analysis of composite plates is based on the following

approaches:
1. Equivalent single-layer theories (2-D)
(a) Classical laminated plate theory
(b) Shear deformation laminated plate theories
2. Three-dimensional elasticity theory (3-D)
(a) Traditional 3-D elasticity formulations
(b) Layerwise theories
This section gives a presentation, and explains the main differences between classical laminated plate theory (CLPT) and first-order shear deformation laminated
plate theory (FSDT). The other theories have not been investigated further and
are not a part of this project.
The simplest theory is the CLPT and requires that the Kirchhoff hypothesis
holds, which assumes that plane cross sections remain plane and normal to the
middle-plane during deformations. This implies that the transverse shear strains
vanish. The FSDT is a bit more complicated and is build on the Reissner-Mindlin
hypothesis, where plane cross sections remain plane after deformation, but not
necessarily normal to the refrence plane. This results inclusion of out-of-plane
shear deformation.
2.2
The Classical Laminated Plate Theory, CLPT
2.2.1
Kinematics
The in-plane displacements are related to the normal displacements as follows

[2]:
u(x, y, z) = u0 z
w
,
x
v(x, y, z) = v0 z
4
w
,
y
w(x, y, z) = w0 (x, y)
Figure 2.1: Undeformed and deformed geometries of an edge of a plate under the Kirchhoff
assumptions [1].
So u is displacement in x direction, v is displacement in y direction and w is

displacement in z direction, while u0 , v0 and w0 are displacements of the midplane
in x, y and z directions, respectively. Based on the displacement field above, we
can find the strains as follows:

0
x
x
u

0 y
y
=
=
v
xy
y
x
2 w0
u0
x
v0
2 w0
y2
+z
=
u0 y v0
2w
0
+ x
2 xy
y
0
kx
x
0
ky
+z
=
0y
xy
kxy
5
(2.1)
2.2.2
The Material Law
Definition of tensor strains [2]:
x
L
y
T
= [T ]
1
1
2 LT
2 xy
Using this definition, the stress strain relations are given by [2]:
Q11 Q12
0
x
L
L
y
T
0
T
= [T ]1
= [T ]1 Q12 Q22
xy
LT
0
0 2Q66
2 LT
x
x
x
]
y
y
y
= [Q
= [Q]
= [T ]1 [Q ] [T ]
xy
2 xy
2 xy
11 Q
12 Q
16 x
Q
12 Q
22 Q
26
y
= Q
Q16 Q26 Q66

xy
Inserting of equation (2.1) gives:
11 Q
12 Q
16 0x
Q
x
12 Q
22 Q
26
0y
y
= Q
0
16 Q
26 Q66
xy
xy
Q
11 Q
12 Q
16 kx
Q
12 Q
22 Q
26
ky
+ z Q
66 kxy
Q16 Q26 Q
2.2.3
(2.2)
Resultant Forces and Moments
The stresses in a laminate vary layer to layer. Hence it is convenient to deal

with a simpler but equivalent system of forces and moments acting on a laminate
cross section. Resultant force is obtained by integrating the corresponding stress
through the laminate thickness h [2]:
6
n Z hi x
Nx Z h2 x
X
Ny
y
y
=
dz =
dz
(2.3)
h
h
i1
i=1
2
Nxy
xy
xy i
Similarly, the resultant moment is obtained by integration through the thickness
of the corresponding stress times the moment arm with respect to the midplane
[2]:
Z
n
Mx Z h2 x
h
x
i
X
y
y
My
z dz =
z dz
(2.4)
=
hi1
h2
i=1
Mxy
xy
xy
i
Substitution of equation (2.2) in equations (2.3) and (2.4) gives:
11 Q
12 Q
16
n Z hi
Q
Nx
x
X
12 Q
22 Q
26 dz
Q
0y
Ny
=
0
26 Q66
16 Q
i=1 hi1
Nxy
xy
Q
i
11 Q
12 Q
16
n Z hi
Q
kx
X
12 Q
22 Q
26 dz
ky
+
z Q
i1
i=1
Q16 Q26 Q66 i
kxy
u0
A11 A12 A16
x
v0
A12 A22 A26

=
u0 y v0
A16 A26 A66
+ x
y
2 w0
2
B11 B12 B16
x
2w
y2
B12 B22 B26
+
B16 B26 B66 2 2 w0

xy
11 Q
12 Q
16
Z
n
Q
Mx
x
h
i
X
12 Q
22 Q
26 dz
0y
My
=
z Q
0
16 Q
26 Q66
i=1 hi1
Mxy
xy
Q
i
11 Q
12 Q
16
Z
n
Q
k
h
x
i
X
12 Q
22 Q
26 dz
ky
+
z2 Q
16 Q
26 Q66
i=1 hi1
kxy
Q
i
u0
B11 B12 B16
x
v0
=
B12 B22 B26
u0 y v0
B16 B26 B66
+ x
y
2 w0
2
D11 D12 D16
x
2 w0
y2
D12 D22 D26
+
D16 D26 D66 2 2 w0

xy
7
(2.5)
(2.6)
2.2.4
Equilibrium Equations in Terms of Displacements
From Appendix B.2, the equations (B.1), (B.2) and (B.7) are the equilibrium
equations for a laminated thin plate:
Nx Nxy
+
=0
x
y
Ny Nxy
+
=0
y
x
2 Mxy 2 My
2 Mx
+
2
+
+ p = 0
x2
xy
y 2
2
w
where p = p + Nx xw2 + Ny yw2 + 2Nxy xy
tw2 .
The last equation solves a buckling problem. Insert equation (2.6), we obtain:
4w
4w
4w
4w
4w
4D
(2D
+
4D
)
4D
D
16
12
66
26
22
x4
x3 y
x2 y 2
xy 3
y 4
3 u0
3 u0
3 u0
3 u0
+B11 3 + 3B16 2 + (B12 + 2B66 )
+
B
26
x
x y
xy 2
y 3
3 v0
3 v0
3 v0
3 v0
+B16 3 + (B12 + 2B66 ) 2 + 3B26
+
B
22
x
x y
xy 2
y 3
+p = 0
(2.7)
D11
For specially orthotropic laminates, their constitutive equations satisfy the following conditions [2]:
A16 = A26 = 0
Bij = 0
D16 = D26 = 0
Incorporation of conditions above into equation (2.7) simplifies the equilibrium
equation for specially orthotropic laminates as follow:
D11
4w
4w
4w
+
(2D
+
4D
)
+
D
= p
12
66
22
x4
x2 y 2
y 4
8
(2.8)
2.3
The First-Order Shear Deformation Theory, FSDT
2.3.1
Kinematics
Figure 2.2: Undeformed and deformed geometries of an edge of a plate under the assumptions
of the first-order plate theory [3].
The displacement field for the FSDT based on the assumption from chapter 2.1
and the figure (2.1) can be expressed as [3]:
u(x, y, z) = u0 + zx ,
v(x, y, z) = v0 + zy ,
w(x, y, z) = w0 (x, y)
where:
u
v
= x ,
= y
z
z
which indicate that x and y are the rotations of a transverse normal about the
y and x axes, respectively.
It is convenient to split the strain vector into two parts, where b is the bending
part and s is the shear part [4]:

x
b =
y
=

xy
u0
y
u0
x
v0
y
v0
x
+z
x
y
x
x
y
y
y
x
(2.9)
s =
2.3.2
yz
xz
v
z
u
z
+
+
w0
y
w0
x
y +
x +
w0
y
w0
x
(2.10)
The Material Law
Definition of tensor strains [2]:
L
x
T
y
= [T ]
1
2 LT
2 xy
The relations between stresses and strains are from the relation for linearized
elasticity. For FSDT it is convenient to split it into two parts, bending and
shear. Similar to section 2.2.2, by using the tensor strains, the bending part can
be expressed as:
u0
11 Q
12 Q
16
Q
x
x
v0
12 Q
22 Q
26
y
= Q
y
16 Q
26 Q66 u0 + v0
xy
Q
y
x
Q11 Q12 Q16
x
y
+ z Q12 Q22 Q26

y
16 Q
26 Q
66
x + y
Q
y
x
(2.11)
Then the shear part [2]:
yz
xz
Q
= k 44
Q45

Q
= k 44
Q45

45
Q
55
Q

45
Q
55
Q
yz
xz
y +
x +
w0
y
w0
x
(2.12)
where k is shear correction coefficient.

2.3.3
Resultant Forces and Moments
The resultant force and resultant moment are obtained in the same way as the
CLPT:
10
A11
Nx
Ny
=
A12
Nxy
A16
B11
B12
+
B16
A12 A16
A22 A26
A26 A66
B12 B16
B22 B26
B26 B66
B11
Mx
My
= B12
Mxy
B16
D11
+ D12
D16
B12 B16
B22 B26
B26 B66
D12 D16
D22 D26
D26 D66
u0
y
x
y
u0
y
u0
x
v0
y
+
x
x
y
y
u0
x
v0
y
x
y
v0
x
y
x
v0
x
x
x
y
y
y
x
(2.13)
(2.14)
Equations relating the shear-force resultants Rxz and Ryz to the shear strains xz
and yz can be written as [2]:
2.3.4
Ryz
Rxz
n Z
X
44 Q
45
Q
= k
45 Q
55
Q
i=1 hi1

y +
A44 A45
= k
A45 A55
x +
hi
dz
i
w0
y
w0
x
0
yz
0
xz
(2.15)
Equilibrium Equations in Terms of Displacements
To solve a buckling problem, we need equations (B.4), (B.5) and (B.6) from
Appendix B.2:
Mx Mxy
+
Rxz = 0
x
y
My Mxy
+
Ryz = 0
y
x
Rxz Ryz
+
+ p = 0
x
y
11
w
tw2
where p = p + Nx xw2 + Ny yw2 + 2Nxy xy
Constitutive equations for a specially orthotropic plate with the new displacement
field still satisfy the conditions stated in chapter 2.2.4: A16 = A26 = 0, Bij = 0
and D16 = D26 = 0. In addition, A45 = A54 = 0. In view of these conditions,
equilibrium equations above can be written in terms of the displacement field as
follows:

2 y
2 x
w
2 x
=0
+ D66 2 A55 k x +
D11 2 + (D12 + D66 )
x
xy
y
x

2 y
2 x
2 y
w
D22 2 + (D12 + D66 )
=0
+ D66 2 A44 k y +
y
xy
x
y

y 2 w
x 2 w
+
+ A44 k
+
+ p = 0
A55 k
x
x2
y
y 2
(2.16)
(2.17)
(2.18)
Equations (2.16)-(2.18) are three coupled second-order differential equations with

w, x and y as the three unknows.
12
3.1
Analysis of Specially Orthotropic Plates Using

CLPT
Buckling of Simply Supported Plates under Uniaxial
Compressive Load
Figure 3.1: Plate with uniaxial compression load [1].
For the buckling analysis, we assume that the only applied load is the in-plane
force in x direction. All other loads are zero. From equation (2.8) we put p =
2
2
Nx xw2 = N xw2 . Now the equation that solves the buckling problem is given
by:
D11
4w
4w
2w
4w
+
(2D
+
4D
)
+
D
+
N
=0
12
66
22
x4
x2 y 2
y 4
x2
(3.1)
The plate edges are simply supported so that the transverse displacements at the
edges and resultant moments about each edge are zero. These edge conditions
are the boundary conditions, and mathematically expressed as follows [2]:
x=0:
w(0, y) = 0
Mx (0, y) = 0
x=a:
w(a, y) = 0
Mx (a, y) = 0
y=0:
w(x, 0) = 0
My (x, 0) = 0
y=b:
w(x, b) = 0
My (x, b) = 0
13
A Navier solution of equation (3.1) that also satisfies the preceding boundary
conditions is given by [2]:
w(x, y) =
wmn sin
n=1 m=1
mx
a
sin
ny
b
where wmn are the displacement coefficients, m and n are positive integers.
We now assume that
w(x, y) = wmn sin
mx
a
sin
ny
(3.2)
Substituting equation (3.2) in equation (3.1) gives:
N = D11
m 2
a
+ (2D12 + 4D66 )
n 2
b
+ D22
a 2 n 4
m
(3.3)
Thus, for each choice of m and n there corresponds a unique value of N. The
critical buckling load is the smallest of N, which can be obtained by n = 1 and
m varying.
3.2
Buckling of Clamped Plates under Uniaxial Compressive Load
Still, we assume that the only applied load is the in-plane force in x direction. All
other loads are zero. For plates with all edges clamped we have chosen RayleighRitz method to solve the buckling problem. The method is based on the plates
potential energy. We now split the total potential energy in two parts, bending
and external forces [5]:
= Ub + Up
(3.4)
where
Ub
Z Z h
Z
2
1
1
T
dV =
T D dA
Q dz dA =
h
2
2
V
A 2
A
2 2
2 2
Z bZ a
1
w
w
w
=
D11
+
2D
12
2 0 0
x2
x2
y 2
2 2
2 2
w
w
dx dy
+D22
+ 4D66
2
y
xy
1
=
2
14
(3.5)
1
Up =
2
Z bZ
0
w
x
2
(3.6)
dx dy
The boundary conditions associated with the clamped edges are [1]:
x=0:
w(0, y) = 0
x=a:
w(a, y) = 0
y=0:
w(x, 0) = 0
y=b:
w(x, b) = 0
w(0, y)
=0
x
w(a, y)
=0
x
w(x, 0)
=0
y
w(x, b)
=0
y
A solution that satisfies the preceding boundary conditions is given by [6]:

w(x, y) =
wmn sin
n=1 m=1
mx
a
sin
nx
a
sin2
y
b
where wmn are the displacement coefficients, m and n are positive integers.
The equation above with only one term, m and n varying, is usually enough to
solve the buckling problem. So we assume that:
mx
y
nx
w(x, y) = wmn sin
sin
sin2
(3.7)
a
a
b
Substitution of equation (3.7) in equation (3.4) gives:
h
i
1 4 2
3 b
3 a
1
m2
4
w
m
+
D
+
(
D
+
D
)
11 4 a3
22 4 b3
66 ab
mn
2
2 12
2 2 bm2
3 wmn
N,
m=n
32
a
=
1 4 2
a
3 b
4
2
2
4
wmn D11 16 a3 (n +
4
i 6m n2 + 2m ) + D22 b3
b
+( 1 D + D ) n2 +m2 3 wmn
(n2 + m2 )N,
m=
6 n
66
2 12
ab
64
a
Equilibrium requires that = 0, thus
wmn = 0
wmn
15
=0
wmn
(3.8)
1
2
23 b
4
23 a
2 m2
mn D11 4 a3 m + D22 4 b3 + ( 2 D12 + D66 ) ab
3 bm2
N = 0,
16
a
=
wmn
1
3 b 2
2
wmn [D11 16
(n4 + 6m2 n2 + m4 ) + D22 ba3 2
a3
2
2 1
3 b
+( 2 D12 + D66 ) ab (n2 + m2 ) 16
(n2 + m2 )N] = 0,
a
m=n
m 6= n
(3.9)
Solving equation (3.9) for N, we obtain:
N=
2
2
2 1
4 2 D11 m2
+ 4b4Dm222 a + 16
( D + D66 ),
a2
3b2 2 12
2
16 a2
4
2
2
4
2
D11 2 (n +6n m +m )+D22 3 4 +( 12 D12 +D66 ) 16
3
a
b
n2 +m2
m=n
2
(n2 +m2 )
b2
, m 6= n
(3.10)
Thus, combination of m and n that gives the smallest value of N is the critical
buckling load for a clamped plate.
16
4.1
Analysis of Specially Orthotropic Plates Using

FSDT
Buckling of Simply Supported Plates under Uniaxial
Compressive Load
Figure 4.1: Plate with uniaxial compression load [1].

Since the only applied load is the force in x direction, from equation (2.18),
2
2
p = Nx xw2 = N xw2 . Based on equations (2.16)-(2.18), the equation set that
solves the buckling problem is given by:

2 x
2 y
2 x
w
D11 2 + (D12 + D66 )
=0
+ D66 2 A55 k x +
x
xy
y
x

2 y
2 x
2 y
w
D22 2 + (D12 + D66 )
=0
(4.1)
+ D66 2 A44 k y +
y
xy
x
y

2w
y 2 w
x 2 w
+
A
k
N
+
+
=0
A55 k
44
x
x2
y
y 2
x2
Boundary conditions for this plate are the same as those for CLPT:
x=0:
w(0, y) = 0
Mx (0, y) = 0
x=a:
w(a, y) = 0
Mx (a, y) = 0
y=0:
w(x, 0) = 0
My (x, 0) = 0
y=b:
w(x, b) = 0
My (x, b) = 0
17
The following double Fourier series are assumed to represent w, x and y [2]:
w(x, y) =
X
X
(4.3)
(4.4)
xmn cos
mx
sin
ny
ymn sin
mx
cos
ny
n=1 m=1
y (x, y) =
(4.2)
wmn sin
n=1 m=1
x (x, y) =
sin
ny
mx
n=1 m=1
where wmn , xmn and ymn are the series coefficients, m and n are positive integers.
For simply supported plates, it is enough to consider one term with m and n
varying from each equation. Substitution of equations (4.2)-(4.4) into equation
set (4.1) gives the following matrix equation:
D11 2 D66 2 A55 k D12 D66

A55 k
D12 D66
D22 2 D66 2 A44 k A44 k
2
2
2
A55 k
A44 k
N A55 k A44 k

xmn 0
ymn
0
=
(4.5)

wmn
0
where =
m
a
og =
n
.
b
By defining
C1 = D11 2 D66 2 A55 k
C2 = D12 D66
C3 = A55 k
C4 = D22 2 D66 2 A44 k

C5 = A44 k
simplify the equation (4.5):

C1 C2
C3
xmn 0
C2 C4
C5
ymn
0
=

2
C3 C5 N + C3 + C5
wmn
0
18
(4.6)
We are seeking non-trivial solutions, thus

C1 C2
C3

C2 C4
C5

C3 C5 N2 + C3 + C5
N=

=0

(4.7)
C1 C52 + C3 C22 + C22 C5 + C32 C4 C1C3 C4 C1C4 C5 2C2 C3 C5

2 (C1 C4 C22 )
(4.8)
The critical buckling load occurs at n = 1, while m can vary.
4.2
Buckling of Clamped Plates under Uniaxial Compressive Load
As the CLPT, the Rayleigh-Ritz method has been used to solve the buckling
problem for a clamped plate. It is convenient to split the total potential energy
in three parts, bending, shear and external forces:
= Ub + Us + Up
(4.9)
where
Ub
Us
Z Z h
Z
2
1
1
T
Qb dz dA =
dV =
T D dA
2 A h2 b
2 A
V
2 2
2 2
Z Z
wb
wb
wb
1 b a
+ 2D12
D11
=
2
2
2 0 0
x
x
y 2
2 2
2 2
wb
wb
+D22
dx dy
+ 4D66
2
y
xy
1
=
2
Tb b
(4.10)
Z Z h
Z
2
1
1
T
Qskj s dz dA =
dV =
T Askj s dA
2 A h2 s
2 A s
V
2
2

Z bZ a
1
ws
ws
=
k
+ A55
dx dy
(4.11)
A44
2 0 0
y
x
1
=
2
Ts s
19
1
Up =
2
Z bZ
0
w
x
2
(4.12)
dx dy
The boundary conditions associated with the clamped edges are still:
x=0:
w(0, y) = 0
x=a:
w(a, y) = 0
y=0:
w(x, 0) = 0
y=b:
w(x, b) = 0
w(0, y)
=0
x
w(a, y)
=0
x
w(x, 0)
=0
y
w(x, b)
=0
y
A solution that satisfies the preceding boundary conditions is given by [6]:

w(x, y) = wb + ws
X
y
x
mx
X
sin
sin2
=
wb sin
a
a
b
n=1 m=1
nx
y
+w
s sin
sin
a
b
(4.13)
where wb and ws are the displacement coefficients for bending and shear, m and
n are positive integers.
We now assume that:
w(x, y) = wb sin
mx
a
sin
x
a
sin
y
b
+w
s sin
nx
a
sin
y
b
(4.14)
Equilibrium requires that = 0, thus:
wb +
ws = 0
wb
ws
This implies

w
b
w
s
=0
=0
20
(4.15)
Substitution of equations (4.9) and (4.14) in equation (4.15) gives two solutions.
One for m 6= 1, and another m = 1. For m 6= 1, equation (4.15) gives following
matrix equation:

H1 + NH2
NH3
NH3
H4 + NH5

0
w
b
=
0
ws
(4.16)
where
H1 =

4
4
2
4
2 2
2
4
2 2
2
D
b
(18m
+
3
+
3m
)
+
D
a
b
(8
+
8m
)
+
16D
a
+
D
a
b
(16
+
16m
)
11
12
22
66
32a3 b3
H2 =
2 b (3m6 + 3 + 3n4 3m4 3m2 + 3m2 n4 12m2 n2 6n2 6m4 n2 )

32a (1 2n2 + m4 2m2 n2 2m2 + n4 )
H3 =
8b (mn3 + (1)m+n+1 mn3 )

3a (1 2n2 + m4 2m2 n2 2m2 + n4 )
H4 =
H5 =
k 2
(A44 a2 + A55 n2 b2 )
4ab
2 b (n2 + n6 2m2 n4 + m4 n2 2n4 2m2 n2 )

4a(1 2n2 + m4 2m2 n2 2m2 + n4 )
Non-trivial solutions if the determinant of the matrix expressed in equation (4.16)

is zero. This leads us to a second-order equation:
(H2 H5 H32 )N 2 + (H1 H5 + H2 H4 )N + H1 H4 = 0
The smallest value of N is given by:
N=
(H1 H5 + H2 H4 )
p
(H1 H5 + H2 H4 )2 4(H2 H5 H32 )H1 H4
2(H2 H5 H32 )
Combination of positive integers m and n gives the critical buckling load.

21
(4.17)
For m = 1, equation (4.15) gives following matrix equation:
G1 + NG2
NG3
NG3
G4 + NG5

wb
0
=
w
s
0
(4.18)
where
G1 =

4
3D11 b4 + 2D12 a2 b2 + 3D22 a4 + 4D66 a2 b2
3
3
4a b
G2 =
G3 =
G4 =
2 b (3n4 12n2 )
16a (n4 4n2 )
8b (n3 + (1)n+2 n3 )
3a (n4 4n2 )
k 2
(A44 a2 + A55 n2 b2 )
4ab
G5 =
2 b (n6 4n4 )
4a(n4 4n2 )
Non-trivial sulotion if the determinant of the matrix in equation (4.18) equals

zero. A result of this, is a second-order equation:
(G2 G5 G23 )N 2 + (G1 G5 + G2 G4 )N + G1 G4 = 0

Solving this, we obtain the smallest value of N:
N=
(G1 G5 + G2 G4 )
(G1 G5 + G2 G4 )2 4(G2 G5 G23 )G1 G4

2(G2 G5 G23 )
The critical buckling load depends on the positive integer n.

22
(4.19)
4.3
Buckling of Simply Supported Plates under Biaxial

Compressive Load
Figure 4.2: Plate with biaxial compressive load [1].
For the buckling analysis, we assume that the only applied loads are the in-plane
compression edge forces. We now define [1]:
Nx = N,
2
Ny = N,
2
Ny
Nx
That implies p = Nx xw2 +Ny yw2 = N xw2 N yw2 from equation (2.18). Based
on equations (2.16)-(2.18), the equation set that solves the buckling problem takes
the form:

2 y
2 x
w
2 x
=0
+ D66 2 A55 k x +
D11 2 + (D12 + D66 )
x
xy
y
x

2 y
2 x
2 y
w
D22 2 + (D12 + D66 )
=0
(4.20)
+ D66 2 A44 k y +
y
xy
x
y

y 2 w
2w
2w
x 2 w
+
+
A
k
+
N
=0
A55 k
44
x
x2
y
y 2
x2
y 2
23
Boundary conditions for simply supported plates are still:

x=0:
w(0, y) = 0
Mx (0, y) = 0
x=a:
w(a, y) = 0
Mx (a, y) = 0
y=0:
w(x, 0) = 0
My (x, 0) = 0
y=b:
w(x, b) = 0
My (x, b) = 0
The following double Fourier series are assumed to represent w, x and y :

ny
mx
sin
(4.21)
w(x, y) = wmn sin
a
b
mx
ny
x (x, y) = xmn cos
sin
(4.22)
a
b
mx
ny
y (x, y) = ymn sin
cos
(4.23)
a
b
Substitution of equations (4.21)-(4.23) into equation set (4.20) gives the following
matrix equation:
D11 2 D66 2 A55 k D12 D66

A55 k
D12 D66
D22 2 D66 2 A44 k A44 k
2
2
2
2
A55 k
A44 k
N + N A55 k A44 k

xmn 0
ymn
0
=
(4.24)

wmn
0
where =
m
a
og =
n
.
b
By defining
C1 = D11 2 D66 2 A55 k
C2 = D12 D66
C3 = A55 k
C4 = D22 2 D66 2 A44 k

C5 = A44 k
C1 C2
C3
xmn 0
C2 C4
C5
ymn
0
=

2
2
C3 C5 N( + ) + C3 + C5
wmn
0
24
(4.25)
We are seeking non-trivial solutions, thus

C1 C2
C3

C2 C4
C5

C3 C5 N(2 + 2) + C3 + C5
N=

=0

(4.26)
C1 C52 + C22 C3 + C22 C5 + C32 C4 C1C3 C4 C1C4 C5 2C2 C3 C5

2 (C1 C4 C22 ) + 2(C1 C4 C22 )
(4.27)
Right combination of m and n gives the critical buckling load. Various values of
will also be investigated.
4.4
Buckling of Simply Supported Plates under In-plane

Shear Load
Figure 4.3: Plate with in-plane shear load [1].
In this section we consider buckling of specially orthotropic plates under inplane shear load, Nxy . The problem does not permit the Navier solution, so
as for clamped plates, we use Rayleigh-Ritz method to solve the problem. When
2w
2w
everything else but in-plane shear load is zero, p = 2Nxy xy
= 2Nxy xy
.
25
For simply supported plates subjected to in-plane shear load, the same boundary
conditions are valid with corresponding expressions:
X
ny
mx
X
sin
w(x, y) =
wmn sin
a
b
n=1 m=1
x (x, y) =
y (x, y) =
n=1 m=1
X
xmn cos
mx
ymn sin
mx
n=1 m=1
sin
ny
cos
ny
We now split the total potential energy functional for the Rayleigh-Ritz method
in three parts (bending, shear and external forces):
(4.28)
= Ub + Us + Up
where
Ub
Us
Z Z h
Z
2
1
1
T
dV =
T D dA
b Qb dz dA =
h
2 A 2
2 A
V
2

Z bZ a
1
y
x
x
=
+ 2D12
D11
2 0 0
x
x
y
2
2

x y
y
+
+ 4D66
dx dy
+D22
y
y
x
1
=
2
Tb b
(4.29)
Z Z h
Z
2
1
1
T
s Qskj s dz dA =
dV =
Ts Askj s dA
h
2
2
V
A 2
A
2
2

Z bZ a
1
w
w
=
+ A55 x +
dx dy
(4.30)
k
A44 y +
2 0 0
y
x
1
=
2
Ts s
1
Up =
2
Z bZ
0
xy
2N
w
x

w
y
dx dy
(4.31)
Substituting the Fourier approximations for w, x and y gives:
Ub
X
X
2
(D11 m2 b2 + D66 n2 a2 )x2mn
=
8ab
n=1 m=1
+(2D12 mnab + 2D66 mnab)xmn ymn

2
+(D22 n2 a2 + D66 m2 b2 )ymn
26
(4.32)
X
k
2
A55 a2 b2 x2mn + A44 a2 b2 ymn
+ 2A55 mab2 xmn wmn
=
8ab
n=1 m=1

2
+2A44 na2 bymn wmn + (A44 n2 2 a2 + A55 m2 2 b2 )wmn
(4.33)
Us
Using the following identities on equation (4.31) [7]:

Z
a
0
mx
px
sin
cos
dx =
a
a
0,
0,
ny
qy
sin
cos
dy =
b
b
m
2a
,
m2 p2
2b n
,
n2 q 2
if m p is an even number
if m p is an odd number
if n q is an even number
if n q is an odd number
We arrive at:
xy
Up = 4N
X
X
X
X
n=1 m=1 q=1 p=1
(m2
mnpq
wmn wpq
p2 )(n2 q 2 )
(4.34)
where m p and n q are odd numbers.

Equilibrium requires that = 0, thus:
= 0,
xmn
m = 1, ..., , n = 1, ...,
= 0,
ymn
m = 1, ..., , n = 1, ...,
= 0,
wmn
m = 1, ..., , n = 1, ...,
(4.35)
Inserting of equations (4.32)-(4.34), equation set (4.35) becomes:

1 2 n2 a
1
1 2 m2 b
D11 +
D66 + kabA55 xmn
4 a
4 b
4

1 2
1
+
mnD12 + 2 mnD66 ymn
4
4

1
kmbA55 wmn
+
4
= I1 xmn + I2 ymn + I3 wmn = 0
=
xmn
27
(4.36)

1 2
1 2
mnD12 + mnD66 xmn
4
4

2 2
1 2 m2 b
1
1 n a
D22 +
D66 + kabA44 ymn
+
4 b
4 a
4

1
knaA44 wmn
+
4
= I4 xmn + I5 ymn + I6 wmn = 0
=
ymn

1
1
kmbA55 xmn +
knaA44 ymn
4
4

2 2
1 kn aA44 1 2 km2 bA55
wmn
+
+
4
b
4
a

mnpq
wpq
+
8Nxy 2
(m p2 )(n2 q 2 )
= I7 xmn + I8 ymn + I9 wmn + I10 wpq = 0
=
wmn
(4.37)
(4.38)
m, n, p and q in equations (4.36)-(4.38) are positive integers, and run from 1 to

. I10 is valid for m p, n q odd numbers, otherwise zero.
Equation set (4.35) can be expressed in matrix form:
x11
x12
..
.
x1n
x21
..
.
xm1
..
.
xmn
y11
..
.
ymn
w11
..
.
wmn
= [MXI]
28
x11
x12
..
.
x1n
x21
..
.
xm1
..
.
xmn
y11
..
.
ymn
w11
..
.
wmn
0
0
..
.
0
0
..
.
0
..
.
0
0
..
.
0
0
..
.
0
(4.39)
A simplified version of equation (4.39) takes the form:

matrix1 matrix2 matrix3 x 0

matrix4 matrix5 matrix6
y
0
=

matrix7 matrix8 matrix9
w
0
(4.40)
If we now assume that m = 1, ..., M and n = 1, ..., N, then this [MXI] is a

3MN 3MN matrix. Entries of [MXI] are based on equations (4.36)-(4.38). To
get a idea how [MXI] looks like, it is convenient to split this huge matrix [MXI]
into 9 small matrixes. [matrix1] has row number 1 to MN, column number 1
to MN in [MXI], and is based on I1 in equation (4.36). This results a diagonal
matrix:
[matrix1] =
I1 (m = 1, n = 1)
0
0
0
I1 (1, 2) 0
..
..
..
.
.
0
.
0
0 I1 (m = M, n = N)
[matrix2] has row number 1 to MN, column number MN + 1 to 2MN, and is

based on I2 in equation (4.36). Also a diagonal matrix:
I2 (m = 1, n = 1)
0
0
I2 (1, 2) 0
[matrix2] =
..
..
..
.
.
0
.
0
0 I2 (m = M, n = N)
Diagonal matrix [matrix3] has row number 1 to MN, column number 2MN + 1
to 3MN, and is based on I3 in equation (4.36):
I3 (m = 1, n = 1)
0
0
I3 (1, 2) 0
[matrix3] =
.
..
.
.
.
.
.
.
0
0
0 I3 (m = M, n = N)
[matrix4] has row number MN + 1 to 2MN, column number 1 to MN, and is
based on I4 in equation (4.37).
[matrix5] has row number MN + 1 to 2MN, column number MN + 1 to 2MN,
and is based on I5 in equation (4.37).
[matrix6] has row number MN + 1 to 2MN, column number 2MN + 1 to 3MN,
29
[matrix7] has row number 2MN + 1 to 3MN, column number 1 to MN, and is
based on I7 in equation (4.38).
[matrix8] has row number 2MN + 1 to 3MN, column number MN + 1 to 2MN,
[matrix9] has row number 2MN +1 to 3MN, column number 2MN +1 to 3MN,
and is based on I9 and I10 in equation (4.38).
All the matrixes are diagonal matrixes except [matrix9] because of the term wpq .
I9 denotes the diagonal entries, while I10 denotes others.
To do the the buckling analysis, we need to find the non-trivial solution of equation (4.40). By solving det[MXI] = 0, we will have the critical buckling load
xy )cr . The accuracy of the result depends on the number of xmn , ymn and wmn
(N
terms. Timoshenko [7] suggests that we divide this system into two groups, one
containing constants xmn , ymn and wmn for which m + n are odd numbers and
the other for which m + n are even numbers. For shorter plates (a/b < 2), it is
enough to consider the second group. For longer plates both groups of equations
should be considered.
4.5
Buckling of Simply Supported Plates under Combined

Loads
Figure 4.4: Plate with uniaxial compressive load and in-plane shear load [1].
For the plate subjected to combined loads, we assume that the only applied
loads are uniaxial compression load in x direction and in-plane shear load. Now
30
introducing a new constant, :
Nxy
Nx
Nx = N,
Nxy = N
The boundary conditions and the corresponding double Fourier series for wmn , x
and y are still valid here. Futher, the calculation procedure is the same as the
section 4.4. Again, we use Rayleigh-Ritz method to solve the problem. Potential
energy due to bending and shear are given by equations (4.29) and (4.30), and
the results are presented in equations (4.32) and (4.33). Up is given by:
Up
#
2

w
w
w
dx dy
+ 2Nxy
Nx
x
x
y
0
0
"Z Z
#
2

b
a
w
w
w
1
dx dy
+ 2N
N
=
2 0 0
x
x
y

X
X
1 2 m2 b 2
N
=
wmn
8
a
n=1 m=1
#
X
X
mnpq
4N
w w
2 p2 )(n2 q 2 ) mn pq
(m
q=1 p=1
1
=
2
"Z Z
b
(4.41)
where m p and n q are odd numbers.

As we know, the total potential energy is given by = Ub + Us + Up . Equilibrium
requires that = 0, thus:

1 2 n2 a
1
1 2 m2 b
D11 +
D66 + kabA55 xmn
4 a
4 b
4

1 2
1
+
mnD12 + 2 mnD66 ymn
4
4

1
kmbA55 wmn
+
4
= J1 xmn + J2 ymn + J3 wmn = 0
=
xmn
31
(4.42)

1 2
1 2
mnD12 + mnD66 xmn
4
4

2 2
1 2 m2 b
1
1 n a
D22 +
D66 + kabA44 ymn
+
4 b
4 a
4

1
knaA44 wmn
+
4
= J4 xmn + J5 ymn + J6 wmn = 0
=
ymn

1
1
kmbA55 xmn +
knaA44 ymn
4
4

2 2
1 kn aA44 1 2 km2 bA55 1 2 m2 b
+
N wmn
+
4
b
4
a
4 a

mnpq
wpq
+
8N 2
(m p2 )(n2 q 2 )
= J7 xmn + J8 ymn + J9 wmn + J10 wpq = 0
=
wmn
(4.43)
(4.44)
m, n, p and q in equations (4.42)-(4.44) are positive integers, and run from 1 to

. I10 is valid for m p, n q odd numbers, otherwise zero.
Equations (4.42)-(4.44) can be expressed in matrix form:
x11
x12
..
.
x1n
x21
..
.
xm1
..
.
xmn
y11
..
.
ymn
w11
..
.
wmn
= [MXJ]
32
x11
x12
..
.
x1n
x21
..
.
xm1
..
.
xmn
y11
..
.
ymn
w11
..
.
wmn
0
0
..
.
0
0
..
.
0
..
.
0
0
..
.
0
0
..
.
0
(4.45)
A simplified version of equation (4.45) is given by:

mat1 mat2 mat3 x 0

mat4 mat5 mat6
y
0
=

mat7 mat8 mat9
w
0
(4.46)
Now this [MXJ] is based on equations (4.42)-(4.44). If m = 1, ..., M and n =

1, ..., N, then [MXJ] is a 3MN 3MN matrix. Just like section 4.4, we can now
split [MXJ] in 9 smaller matixes, where [mat1],...,[mat8] are diagonal matrixes
based on expressions J1,...J8 in equations (4.42)-(4.44). J9 denotes the diagonal
entries in [mat9], while J10 denotes others.
To find the the critical buckling load N, we need to solve det[MXJ] = 0. The
accuracy of the result depends on the number of xmn , ymn and wmn terms.
4.6
Plates with Initial Geometric Imperfection under Uniaxial Compressive Load
4.6.1
Relationship Between Displacement and Applied Load
It is normally unrealistic to assume that a plate is perfect. We consider now a

simply supported plate with an initial deformation, wint . Usually such deformation occurs by production faults or welding faults in assembly of the plate. When
the plate subjected to in-plane stresses, it will receive an additional deformation,
w. Thus, the total deformation is wtot = wint + w. In this section we assume that
the only applied load is the in-plane compressive force in x direction. This means
2w
2w
tot
tot
= N x
We are now interested in analysing the response
p = Nx x
2
2 .
under increasing load. In other words, finding the relationship between applied
load and the displacement. Based on equations (2.16)-(2.18) from section 2.3.4,
the equation set that solves the problem is given by:

2 y
2 x
w
2 x
=0
+ D66 2 A55 k x +
D11 2 + (D12 + D66 )
x
xy
y
x

2 y
2 x
2 y
w
D22 2 + (D12 + D66 )
=0
+ D66 2 A44 k y +
y
xy
x
y
A55 k
x 2 w
+
x
x2
+ A44 k
y 2 w
+
y
y 2
33
2 wtot
=0
x2
(4.47)
For simply supported plate, we have the usual boundary conditions with corresponding double Fourier series:
ny
mx
sin
x (x, y) = xmn cos
a
b
mx
ny
y (x, y) = ymn sin
cos
a
b
and
wtot (x, y) = w(x, y) + wint (x, y)
mx
ny
m x
n y
i
i
= wmn sin
sin
+ wi sin
sin
(4.48)
a
b
a
b
where wmn , xmn and ymn are the unknown series coefficients. wi is a given imperfection amplitude at centre.
If we now assume that m = mi and n = ni , substitution of x , y and wtot into

equation set (4.47) gives the following matrix equation:
D11 2 D66 2 A55 k D12 D66

A55 k
D12 D66
D22 2 D66 2 A44 k A44 k
2
2
2
A55 k
A44 k
N A55 k A44 k
where =
m
a
og =
n
.
b
0
xmn
ymn
0
wmn
N2 wi
(4.49)
By defining
C1 = D11 2 D66 2 A55 k
C2 = D12 D66
C3 = A55 k
C4 = D22 2 D66 2 A44 k

C5 = A44 k

C1 C2
C3
0
xmn
C2 C4
C5
ymn
0
=
C3 C5 N2 + C3 + C5
wmn
N2 wi
34
(4.50)
Now solving equation (4.50) for xmn , ymn and wmn :

N2 wi
wmn =
C2 C3
C2 C3
C5
C5 C
C1
1
C2
C4 C2
ymn
C32
C1
(4.51)
+ C3 + C5 +
N2

C5
wmn
=
C2
C4 C21
xmn =
C2 C3
C1
(4.52)
C2 ymn C3 wmn
C1
(4.53)
Relationship between displacement and applied load for a point (x, y) is given by
equation (4.48). Inserting of wmn obtained in equation (4.51), we arrive at:
wtot (x, y) =
N2 wi
C2 C3
C2 C3
C5
C5 C
C1
1
C2
C4 C2
1
+ wi sin
m x
i
sin
C32
C1
+ C3 + C5 + N2
n y
i
sin
mx
a
sin
ny
b
(4.54)
for given N, wtot is the only unknown.
4.6.2
First Ply Failure
For composite materials, strengths in different directions can vary widely. For
example, a unidirectional lamina could withstand a lot more of tension along
the fibres, compared to tension perpendicular to the fibres. So it is interesting
to calculate the allowable strength for composites. The failure criteria discussed
here are limited to first ply failure, which gives a conservative estimate of the
strength of the laminate.
There are a several models which can be used to calculate the Failure Index (FI).
The one we have chosen is a widely used criterion, Tsai-Wu. In its most general
form, it can be written as [5]:
F I = Fij i j + Fi i = 1
where i, j = 1, . . . , 6.
35
(4.55)
For an orthotropic material in the 2-D plane stress state, the Tsai-Wu failure
criterion becoming:
2
F I = F11 12 + F22 22 + F66 12
+ 2F12 1 2 + F1 1 + F2 2 = 1
(4.56)
This has four quadratic strength parameters, analogous to the modulus components, and two linear strength parameters which account for the differences in
the tensile and compressive strengths. Five of the six strength parameters are
obtained from simple mechanical tests, and they are given by:
F11 =
1
,
Xt Xc
F2 =
F22 =
1
,
Yt Yc
F1 =
1
1
Xt Xc
1
1
1
, F66 = 2
Yt Yc
S12
where
Xt = Tension strength along the x-axis
Xc = Compression strength along the x-axis
Yt = Tension strength along the y-axis
Yc = Compression strength along the y-axis
S12 = Shear strength in the xy-plane
The sixth parameter, F12 , represents the interaction of two stress components,
1 and 2 , in a combined strength test. A biaxial test must be conducted to
determine the F12 . This is a much more difficult test to perform experimentally.
In practice, we usually set the normalised interaction term (F12

) to a fixed value,
then F12 can be found by:
F12
=
F12
F11 F22
These are just default values. Often F12

= 0.5 or F12
= 0.
In practice, instead of equation (4.56), there is a another way to define Tsai-Wu

criterion. By introducing a factor R [5], equation (4.56) becoming:

2
R2 F11 12 + F22 22 + F66 12
+ 2F12 1 2 + R (F1 1 + F2 2 ) 1 = 0
R2 1 + R2 1 = 0
36
(4.57)
Solving this, we will have positive R given by:

p
2 + 22 + 41
R=
21
(4.58)
1/R equals 1 gives first ply failure.

Now we are going to analyse at which material failure first occurs for a simply
supported plate with an initial geometric imperfection. Stresses in each ply or
ply i is given by [5]:
0
11 Q
12 Q
16
Q
x
x
0y
y
Q12 Q22 Q26
=
0
16 Q
26 Q66
xy i
xy
Q
i

11 Q
12 Q
16
Q
kx
12 Q
22 Q
26
ky
+ zi Q
(4.59)
16 Q
26 Q
66
Q
k
xy
i
First, we need to find the midplane strains 0 for the entire laminate by solving:
A11 A12 A16 0x

Nx
0
Ny
= A12 A22 A26
(4.60)
0y
Nxy
A16 A26 A66
xy
For a applied load Nx = N, equation (4.60) becoming:
0
1
A11 A12 A16
N
x
0
0
= A12 A22 A26
0y
A16 A26 A66
0
xy
(4.61)
Then from section 2.3.2, using FSDT, we know that the midplane curvature for
entire laminate is given by:
kx
x
y
ky
=
(4.62)
y

x + y
kxy
y
x
where:
x (x, y) = xmn cos

y (x, y) = ymn sin
mx
a
mx
37
sin
cos
ny
b
ny
b
For a given point (x, y) equation (4.62) becoming:

ny
mx
sin
sin
xmn m
kx
a
a
b
ymn n
sin mx
sin ny
ky
=
b
a
b

kxy
xmn n
+ ymn m
cos mx
cos
b
a
a
ny
b
(4.63)
where xmn and ymn are given by equations (4.52) and (4.53) in section 4.6.1.
and coordinate z for each ply. Thereafter calculate lamina
Now determine [Q]
stresses x , y and xy along the x and y axes in each ply by substituting equations
(4.61) and (4.63) into (4.59).
To produce a first ply failure, we need to scale up (or down) the applied load N
until we find the Failure Index F I = 1 in equation (4.56), or find the inverse of
the "strength ratio" 1/R = 1 in equation (4.57).
38
Finite Element Model
For verification of the present methods, a variety of plate dimensions and three
cases with different lay-ups have been considered. For more details, see Appendix
A. Computed results by the present methods have been compared with finite
element (FE) analyses using ANSYS. Case A and B are modelled using SHELL281
elements, while case C SHELL91.
FE analysis is performed in several steps:
Choose the correct element type (SHELL91 or SHELL281) and apply the
lay-ups.
Build up the model with geometry and material properties. Then set the
element size (25 for all cases).
Define the boundary conditions and apply the load.
Static analysis followed by a buckling analysis. List the eigenvalues and the
corresponding buckling modes.
For analyses with imperfection:
Perform the steps above.
Do A non-linear analysis with initial imperfection (= 1. buckling mode). To
show this non-linear behavior, we have to apply a new load which is much
larger than buckling load. It is been chosen twice the critical buckling load
for all cases.
Having the results plotted (load-displacement).
For first ply failure, we have to add the failure criteria before the non-linear
analysis.
Then investigate the plies and try to find the load which gives first ply
failure. "Inverse of Tsai-Wu Strength Ratio Index" has been used.
To do the analyses, it is very important to define the correct boundary conditions
and applied loads for each case. We explain the conditions using figures.
39
1. For simply supported, uniaxial loads, the boundary conditions and the
applied load are defined below.
uz=0

uz=0
ux=0
applied load
y
6
-x
uz=0
connect the nodes on this line

uz=0
uy=0
Figure 5.1: Applied boundary conditions and load for simply supported plate.
2. For clamped edges, uniaxial loads, the boundary conditions and the
applied load are listed below.
uz=0
rotx=roty=rotz=0

uz=0
y
ux=0
rotx=roty=rotz=0 6

-x
applied load
uz=0
rotx=roty=rotz=0

uz=0
uy=0
rotx=roty=rotz=0
Figure 5.2: Applied boundary conditions and load for clamped plate.
3. Now, the boundary conditions and the applied loads for simply supported, biaxial load.
40
applied load, y
uz=0
connect the
nodes on this line

?
applied load, x
uz=0
ux=0
-x
uz=0

uz=0
uy=0
Figure 5.3: Applied boundary conditions and loads for simply supported plate subjected to
biaxial load.
4. For simply supported, in-plane shear.
ux=0
t

?
uz=0

-

?
ux=0
uy=0
these vectors
- are applied load
6
uz=0
t?

uz=0
-x

6

uz=0
Figure 5.4: Applied boundary conditions and load for simply supported plate subjected to
in-plane shear load.
The main idea is to apply each node a point load. The value is given by:
total load
number of element number of nodes per element
41
(Using "nlist" in ANSYS will return a list of nodes.)

5. For simply supported, combined loads (uniaxial and in-plane), the
boundary conditions and the applied loads are defined below.
uz=0

-
uz=0 ?
ux=0
uy=0
these vectors
- are applied load
6
y
6

?

?
t

-x

uz=0
uz=0
6
connect the nodes on line

6

uniaxial
Figure 5.5: Applied boundary conditions and load for simply supported plate subjected to
combined loads.
42
Presentation of Results
6.1
Uniaxial Compressive Load, CLPT and FSDT
This section contains results from both simply supported and clamped edges
using CLPT and FSDT. Corresponding FE analyses have been performed using
ANSYS. Results are presented in the form of tables and graphs. Three cases with
different lay-ups and thickness have been investigated. Their material properties
are listed in Appendix A.
6.1.1
Simply Supported Plates
From section 3.1, using CLPT, the critical buckling load is given by equation (3.3).
From section 4.1, using FSDT, the critical buckling load is given by equation (4.8).
The results are estimated using the routines developed in Matlab.
CASE A - 500500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Case A - Simply supported, 500 500mm

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
175.0 (n=1,m=1)
174.5 (n=1,m=1)
172.2
1372.0 (n=1,m=1) 1356.4 (n=1,m=1)
1323.7
4600.9 (n=1,m=1) 4484.7 (n=1,m=1)
4332.4
10871.0 (n=1,m=1) 10394.0 (n=1,m=1)
9947.4
36578.0 (n=1,m=1) 33159.0 (n=1,m=1)
31236.0
Table 6.1: Case A, simply supported - 500500.
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Failure in percent - Simply supported, A 500 500mm

CLPT - FSDT CLPT - ANSYS
FSDT - ANSYS
0.3%
1.6%
1.3%
1.2%
3.6%
2.5%
2.6%
6.2%
3.5%
4.6%
9.3%
4.5%
10.3%
17.1%
6.2%
Table 6.2: Case A, simply supported - 500500. Failure in percent.
43
The first buckling mode for case A-1 (500 500) is plotted in figure (6.1).
Figure 6.1: Left one shows first buckling mode for case A1, simply supported plate,
500500mm, based on FSDT. Right one is from ANSYS.
Now, to show the results graphically, it is been chosen to introduce two new
parameters, affine plate buckling coefficient, k0 , and generalized rigidity ratio, D
[8]:
k0 =
Ncr b2
2 D11 D22
D12 + 2D66
D =
D11 D22
The modified buckling coefficient is given by:
k0 2D
Futher, equation (6.1) is shown in figure below, plotted against b/t.

44
(6.1)
FSDT
2.6
ANSYS
CLPT
k02D
2.4
2.2
1.8
1.6
0.01
0.02
0.03
0.04
0.05
t/b
0.06
0.07
0.08
0.09
0.1
Figure 6.2: Left one shows simply supported plate, case A, 500500mm. Uniaxial buckling
coefficient (modified) k0 2D vs t/b. Right one is from the article "Generic Buckling Curves
for Specially Orthotropic Rectangular Plates" written by Brunelle and Oyibo [8]. Here modified
buckling coefficient is plotted against plate affine aspect ratio a0 /b0 using CLPT.
According to Brunelle and Oyibo, we should have got a straight horizontal line
with modified buckling coefficient equals to 2 on CLPT. But for this lay-up, we
have got a D11 and D22 which result a plate affine ratio smaller than 1. Maybe
this is because the lay-up contains many nonzero degree plies. For example, for
case A1 500500:
D11 = 1.8820 106 ,
D22 = 0.7557 106
a
(D22 )1/4
a0
=
0.8
b0
(D11 )1/4
b
This gives us a modified buckling coefficient greater than 2. The results from
FSDT and ANSYS are under CLPT, which is logical.
45
CASE A - 2000500

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
165.7 (n=1,m=3)
165.4 (n=1,m=3)
163.9
1284.5 (n=1,m=3) 1274.8 (n=1,m=3)
1253.7
4291.7 (n=1,m=3) 4219.9 (n=1,m=3)
4119.8
10123.0 (n=1,m=3) 9826.7 (n=1,m=3)
9527.2
33998.0 (n=1,m=3) 31850.0 (n=1,m=3)
30494.0
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Failure in percent - Simply supported, A 2000 500mm

FSDT - ANSYS
0.2%
1.1%
0.9%
0.8%
2.5%
1.7%
1.7%
4.2%
2.4%
3.0%
6.3%
3.1%
6.7%
11.5%
4.4%
Table 6.4: Case A, simply supported - 2000500. Failure in percent.

For 2000 500, it is chosen to show the first buckling mode for case A-4.
Figure 6.3: Left one shows first buckling mode for case A4, simply supported plate,
Again, equation (6.1) has been used to show the results graphically.
46
FSDT
2.3
ANSYS
2.2
CLPT
2.1
k02D
2
1.9
1.8
1.7
1.6
1.5
1.4
0.01
0.02
0.03
0.04
0.05
t/b
0.06
0.07
0.08
0.09
0.1
Figure 6.4: Left one shows simply supported plate, case A, 2000500mm. Uniaxial buckling
for Specially Orthotropic Rectangular Plates" written by Brunelle and Oyibo [8], using CLPT.
We see that the results for CLPT 2000500 give a better match. (For case A1,
a0 /b0 3.2. According to the article, k0 2D should be a little higher than 2.)
CASE B - 500500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Case B - Simply supported, 500 500mm

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
179.4 (n=1,m=1)
178.9 (n=1,m=1)
173.8
1484.2 (n=1,m=1) 1467.1 (n=1,m=1)
1412.4
5064.3 (n=1,m=1) 4933.9 (n=1,m=1)
4678.8
12070.0 (n=1,m=1) 11526.0 (n=1,m=1)
10766.0
40955.0 (n=1,m=1) 37007.0 (n=1,m=1)
33688.0
Table 6.5: Case B, simply supported - 500500.
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Failure in percent - Simply supported, B 500 500mm

FSDT - ANSYS
0.3%
3.2%
2.9%
1.2%
5.1%
3.9%
2.6%
8.2%
5.5%
4.7%
12.1%
7.1%
10.7%
21.6%
9.9%
Table 6.6: Case B, simply supported - 500500. Failure in percent.
47
2.4
FSDT
ANSYS
2.2
CLPT
k02D
1.8
1.6
1.4
1.2
0.01
0.02
0.03
0.04
0.05
t/b
0.06
0.07
0.08
0.09
0.1
Figure 6.5: Left one shows simply supported plate, case B, 500500mm. Uniaxial buckling
For case B1, 500500:
D11 = 1.6413 106 ,
D22 = 0.9412 106
a
(D22 )1/4
a0
=
0.87
b0
(D11 )1/4
b
which results k0 2D greater than 2 according to Brunelle and Oyibo. For

CLPT plotted in Matlab, we see that case B1 lies little higher than 2.
CASE B - 2000500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
179.4 (n=1,m=4)
178.9 (n=1,m=4)
173.9
1484.2 (n=1,m=4) 1467.1 (n=1,m=4)
1428.7
5064.3 (n=1,m=4) 4933.9 (n=1,m=4)
4763.2
12070.0 (n=1,m=4) 11526.0 (n=1,m=4)
11018.0
40955.0 (n=1,m=4) 37007 (n=1,m=4)
34736.0
48
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Failure in percent - Simply supported, B 2000 500mm

FSDT - ANSYS
0.3%
3.2%
2.9%
1.2%
3.9%
2.7%
2.6%
6.3%
3.6%
4.7%
9.5%
4.6%
10.7%
17.9%
6.5%
Table 6.8: Case B, simply supported - 2000500. Failure in percent.
Figure 6.6: Left one shows first buckling mode for case B4, simply supported plate,
2.4
FSDT
ANSYS
2.2
CLPT
k02D
1.8
1.6
1.4
1.2
0.01
0.02
0.03
0.04
0.05
t/b
0.06
0.07
0.08
0.09
0.1
Figure 6.7: Left one shows simply supported plate, case B, 2000500mm. Uniaxial buckling
49
For case B1 2000500, a0 /b0 3.5. This value can result a modified buckling
coefficient greater than 2 according to Brunelle and Oyibo.
CASE C - 500500
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
Case C - Simply supported, 500 500mm

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
22787.0 (n=1,m=1) 21638.0 (n=1,m=1)
20634.0
13658.0 (n=1,m=1) 13161.0 (n=1,m=1)
12630.2
4189.6 (n=1,m=1) 4118.8 (n=1,m=1)
4001.6
1829.1 (n=1,m=1) 1811.2 (n=1,m=1)
1771.4
844.0 (n=1,m=1)
839 (n=1,m=1)
824.6
171.8 (n=1,m=1)
171.4 (n=1,m=1)
169.7
Table 6.9: Case C, simply supported - 500500.
Failure in percent - Simply supported, C 500 500mm
FSDT - ANSYS
5.3%
10.4%
4.9%
3.8%
8.1%
4.2%
1.7%
4.7%
2.9%
1.0%
3.3%
2.2%
0.6%
2.4%
1.7%
0.2%
1.2%
1.0%
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
Table 6.10: Case C, simply supported - 500500. Failure in percent.
FSDT
2.15
ANSYS
2.1
CLPT
2.05
k 2D
2
1.95
1.9
1.85
1.8
1.75
1.7
1.65
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
t/b
Figure 6.8: Left one shows simply supported plate, case C, 500500mm. Uniaxial buckling
50
Since a0 /b0 equals to 1 for all thicknesses, we will now have a straight line with
value 2 for CLPT. The result is in accordance with Brunelle and Oyibo.
CASE C - 2000500
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
22787.0 (n=1,m=4) 21638.0 (n=1,m=4)
20974.0
13658.0 (n=1,m=4) 13161.0 (n=1,m=4)
12813.2
4189.6 (n=1,m=4) 4118.8 (n=1,m=4)
4043.2
1829.1 (n=1,m=4) 1811.2 (n=1,m=4)
1785.7
844.0 (n=1,m=4)
839 (n=1,m=4)
829.8
171.8 (n=1,m=4)
171.4 (n=1,m=4)
170.3
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
Failure in percent - Simply supported, C 2000 500mm

FSDT - ANSYS
5.3%
8.6%
3.2%
3.8%
6.6%
2.7%
1.7%
3.6%
1.9%
1.0%
2.4%
1.4%
0.6%
1.7%
1.1%
0.2%
0.9%
0.6%
Table 6.12: Case C, simply supported - 2000500. Failure in percent.
Figure 6.9: Left one shows first buckling mode for case C2, simply supported plate,
51
2.2
FSDT
2.15
ANSYS
2.1
CLPT
2.05
k 2D
2
1.95
1.9
1.85
1.8
1.75
1.7
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
t/b
Figure 6.10: Left one shows simply supported plate, case C, 2000500mm. Uniaxial buckling
a0 /b0 equals to 4 for all thicknesses, we will now have a straight line with value
2 for CLPT. The result is in accordance with Brunelle and Oyibo.
6.1.2
Clamped Plates
For clamped edges, it is been chosen to show the results in tables and their
buckling modes.
CASE A - 500500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Case A - Clamped edges, 500 500mm

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
511.1 (n=1,m=1)
508.3 (n=1,m=1)
491.1
4088.5 (n=1,m=1)
3999.3 (n=1,m=1)
3817.6
13799.0 (n=1,m=1) 13132.0 (n=1,m=1)
12276.2
32708.0 (n=1,m=1) 29960.0 (n=1,m=1)
27282.0
110390.0 (n=1,m=1) 90885.0 (n=1,m=1)
77936.0
Table 6.13: Case A, clamped edges - 500500.
52
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Failure in percent - Clamped edges, A 500 500mm

CLPT - FSDT CLPT - ANSYS FSDT - ANSYS
0.6%
4.1%
3.5%
2.2%
7.1%
4.8%
5.1%
12.4%
7.0%
9.2%
19.9%
9.8%
21.5%
41.6%
16.6%
Table 6.14: Case A, clamped edges - 500500. Failure in percent.
Buckling mode for case A3 is shown below.
Figure 6.11: Left one shows first buckling mode for case A3, clamped plate, 500500mm,
based on FSDT. Right one is from ANSYS.
CASE A - 2000500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Case A - Clamped edges, 2000 500mm

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
331.1 (n=1,m=5)
329.5 (n=5,m=5)
315.9
2610.2 (n=1,m=5) 2561.2 (n=5,m=5)
2438.8
8769.1 (n=1,m=5) 8406.8 (n=5,m=5)
7945.6
20740.0 (n=1,m=5) 19254.0 (n=5,m=5)
18045.4
69843.0 (n=1,m=5) 59323.0 (n=5,m=5)
54548.0
Table 6.15: Case A, clamped edges - 2000500.
53
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Failure in percent - Clamped edges, A 2000 500mm

0.5%
4.8%
4.3%
1.9%
7.0%
5.0%
4.3%
10.4%
5.8%
7.7%
14.9%
6.7%
17.7%
28.0%
8.8%
Table 6.16: Case A, clamped edges - 2000500. Failure in percent.
Figure 6.12: Left one shows first buckling mode for case A5, clamped plate, 2000500mm,
CASE B - 500500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Case B - Clamped edges, 500 500mm

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
511.1 (n=1,m=1)
508.3 (n=1,m=1)
482.4
4088.5 (n=1,m=1)
3999.3 (n=1,m=1)
3729.6
13799.0 (n=1,m=1) 13132.0 (n=1,m=1)
11935.8
32708.0 (n=1,m=1) 29960.0 (n=1,m=1)
26388.0
110390.0 (n=1,m=1) 89199.0 (n=2,m=2)
74482.0
Table 6.17: Case B, clamped edges - 500500.
54
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Failure in percent - Clamped edges, B 500 500mm

0.6%
5.9%
5.4%
2.2%
9.6%
7.2%
5.1%
15.6%
10.0%
9.2%
24.0%
13.5%
23.8%
48.2%
19.8%
Table 6.18: Case B, clamped edges - 500500. Failure in percent.
Figure 6.13: Left one shows first buckling mode for case B2, clamped plate, 500500mm,
CASE B - 2000500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Case B - Clamped edges, 2000 500mm

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
346.1 (n=1,m=5)
344.3 (n=5,m=5)
324.0
2859.7 (n=1,m=6) 2792.9 (n=6,m=6)
2621.2
9679.8 (n=1,m=6) 9179.8 (n=6,m=6)
8542.2
22979.0 (n=1,m=6) 20927.0 (n=6,m=6)
19155.0
77673.0 (n=1,m=6) 63294.0 (n=6,m=6)
55898.0
Table 6.19: Case B, clamped edges - 2000500.
55
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Failure in percent - Clamped edges, B 2000 500mm

0.5%
6.8%
6.3%
2.4%
9.1%
6.6%
5.4%
13.3%
7.5%
9.8%
20.0%
9.3%
22.7%
39.0%
13.2%
Table 6.20: Case B, clamped edges - 2000500. Failure in percent.
56
CASE C - 500500
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
Case C - Clamped edges, 500 500mm

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
69137.0 (n=1,m=1) 61910.0 (n=1,m=1)
53718.0
41439.0 (n=1,m=1) 38299.0 (n=1,m=1)
34012.0
12711.0 (n=1,m=1) 12262.0 (n=1,m=1)
11272.2
5549.7 (n=1,m=1) 5435.8 (n=1,m=1)
5064.2
2560.6 (n=1,m=1) 2529.1 (n=1,m=1)
2374.2
521.2 (n=1,m=1)
519.0 (n=1,m=1)
490.9
Table 6.21: Case C, clamped edges - 500500.
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
Failure in percent - Clamped edges, C 500 500mm

11.7%
28.7%
15.3%
8.2%
21.8%
12.6%
3.7%
12.8%
8.8%
2.1%
9.6%
7.3%
1.2%
7.9%
6.5%
0.4%
6.2%
5.7%
Table 6.22: Case C, clamped edges - 500500. Failure in percent.
Figure 6.16: Left one shows first buckling mode for case C2, clamped plate, 500500mm,
57
CASE C - 2000500
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
Case C - Clamped edges, 2000 500mm

CLPT [N/mm]
FSDT [N/mm]
ANSYS [N/mm]
46335.0 (n=1,m=6) 41424.0 (n=6,m=6)
38064.0
27772.0 (n=1,m=6) 25630.0 (n=6,m=6)
23836.0
8519.1 (n=1,m=6) 8210.8 (n=6,m=6)
7787.0
3719.4 (n=1,m=6) 3641.0 (n=6,m=6)
3481.0
1716.1 (n=1,m=6) 1694.4 (n=6,m=6)
1627.5
349.3 (n=1,m=6)
347.8 (n=6,m=6)
335.6
Table 6.23: Case C, clamped edges - 2000500.
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
Failure in percent - Clamped edges, C 2000 500mm

11.9%
21.7%
8.8%
8.4%
16.5%
7.5%
3.8%
9.4%
5.4%
2.2%
6.8%
4.6%
1.3%
5.4%
4.1%
0.4%
4.1%
3.6%
Table 6.24: Case C, clamped edges - 2000500. Failure in percent.
Figure 6.17: Left one shows first buckling mode for case C5, clamped plate, 2000500mm,
58
6.2
Biaxial Compressive Load, FSDT
From section 4.3, equation (4.27) gives critical buckling load for biaxial compressive load. The results are estimated using the routines developed in Matlab.
= 0.5, = 1 and = 2 have been investigated.
CASE A - 500500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Case A - Simply supported, 500 500mm, = 1

FSDT [N/mm]
ANSYS [N/mm] Failure in percent
87.3 (n=1,m=1)
86.1
1.4%
678.2 (n=1,m=1)
661.8
2.5%
2242.4 (n=1,m=1)
2166.2
3.5%
5196.9 (n=1,m=1)
4973.4
4.5%
16579.0 (n=1,m=1)
15617.0
6.2%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Case A - Simply supported, 500 500mm, = 0.5

FSDT [N/mm]
116.4 (n=1,m=1)
114.8
1.4%
904.2 (n=1,m=1)
882.4
2.5%
2989.8 (n=1,m=1)
2888.2
3.5%
6929.2 (n=1,m=1)
6631.4
4.5%
22106.0 (n=1,m=1)
20824.0
6.2%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm

FSDT [N/mm]
58.2 (n=1,m=1)
57.4
1.4%
452.1 (n=1,m=1)
441.2
2.5%
1494.9 (n=1,m=1)
1444.1
3.5%
3464.6 (n=1,m=1)
3315.4
4.5%
11053.0 (n=1,m=1)
10410.8
6.2%
To show the results graphically, we have chosen interaction diagram. To do that,

we need critical buckling load for Ny . Performing the same procedure as section
59
4.1, we arrive:
Ny0cr =
C1 C52 + C3 C22 + C22 C5 + C32 C4 C1 C3 C4 C1 C4 C5 2C2 C3 C5

2 (C1 C4 C22 )
(6.2)
where
=
m
n
, =
, C1 = D11 2 D66 2 A55 k , C2 = D12 D66
a
b
C3 = A55 k ,
C4 = D22 2 D66 2 A44 k ,
C5 = A44 k
Now, for interaction diagram, x axis denoted by Nxcr /Nx0cr and y axis denoted
by Nycr /Ny0cr .
1
Thinnest
Thickest
0.9
0.8
0.7
Nycr/Ny0cr
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
Figure 6.18: Interaction diagram for case A, Simply supported plate, 500500mm.
60
CASE A - 2000500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm

FSDT [N/mm]
32.5 (n=1,m=1)
32.4
0.3%
251.4 (n=1,m=1)
250.2
0.5%
836.3 (n=1,m=1)
830.4
0.7%
1959.9 (n=1,m=1)
1941.9
0.9%
6461.0 (n=1,m=1)
6376.6
1.3%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Case A - Simply supported, 2000 500mm, = 0.5

FSDT [N/mm]
61.4 (n=1,m=1)
61.2
0.3%
474.9 (n=1,m=1)
472.5
0.5%
1579.7 (n=1,m=1)
1568.5
0.7%
3702.0 (n=1,m=1)
3668.0
0.9%
12204.0 (n=1,m=1)
12044.4
1.3%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm

FSDT [N/mm]
16.7 (n=1,m=1)
16.7
0%
129.5 (n=1,m=1)
128.9
0.5%
430.8 (n=1,m=1)
427.8
0.7%
1009.6 (n=1,m=1)
1000.4
0.9%
3328.4 (n=1,m=1)
3285.0
1.3%
For case with 2000500, the interaction diagram is not a straight line. We need
to calculate Nxcr and Nycr for more values to show the right curve.
61
1
Thinnest
0.9
Thickest
0.8
0.7
Nycr/Ny0cr
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
CASE B - 500500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Case B - Simply supported, 500 500mm, = 1

FSDT [N/mm]
89.4 (n=1,m=1)
86.9
2.9%
733.6 (n=1,m=1)
706.2
3.9%
2467.0 (n=1,m=1)
2339.4
5.5%
5762.8 (n=1,m=1)
5383.0
7.1%
18503.0 (n=1,m=1)
16845.2
9.8%
62
Case B - Simply supported, 500 500mm, = 0.5

FSDT [N/mm]
119.3 (n=1,m=1)
115.8
3.0%
978.1 (n=1,m=1)
941.6
3.9%
3289.3 (n=1,m=1)
3119.2
5.5%
7683.8 (n=1,m=1)
7177.4
7.1%
24671.0 (n=1,m=1)
22460.0
9.8%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm

FSDT [N/mm]
59.6 (n=1,m=1)
57.9
2.9%
489.0 (n=1,m=1)
470.8
3.9%
1644.6 (n=1,m=1)
1559.6
5.5%
3841.9 (n=1,m=1)
3588.6
7.1%
12336.0 (n=1,m=1)
11230.0
9.8%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Intaraction diagram has only been made for the thinnest and thickest plates.
1
Thinnest
Thickest
0.9
0.8
0.7
0.5
ycr
/N
y0cr
0.6
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
Figure 6.20: Interaction diagram for case B, Simply supported plate, 500500mm.
63
CASE B - 2000500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm

FSDT [N/mm]
39.7 (n=1,m=1)
39.4
0.8%
365.8 (n=1,m=1)
363.4
0.7%
1272.7 (n=1,m=1)
1262.1
0.8%
3037.1 (n=1,m=1)
3004.8
1.1%
10111.0 (n=1,m=1)
9958.4
1.5%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
Case B - Simply supported, 2000 500mm, = 0.5

FSDT [N/mm]
75.0 (n=1,m=1)
74.5
0.7%
690.9 (n=1,m=1)
686.5
0.6%
2403.9 (n=1,m=1)
2383.8
0.8%
5736.7 (n=1,m=1)
5675.2
1.1%
19098 (n=1,m=1)
18804.6
1.6%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm

FSDT [N/mm]
20.5 (n=1,m=1)
20.3
1.0%
188.4 (n=1,m=1)
187.2
0.6%
655.6 (n=1,m=1)
650.2
0.8%
1564.6 (n=1,m=1)
1548.0
1.1%
5208.5 (n=1,m=1)
5130.4
1.5%
For case B, 2000500, more values of have been calculated and three thicknesses
have been plotted.
64
1
Thinnest
0.9
Thickest
0.8
Middle
0.7
Nycr/Ny0cr
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
Due to numerical approaches, the curves differ slightly from each other. But the
difference is negligible.
CASE C - 500500
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
Case C - Simply supported, 500 500mm, = 1

FSDT [N/mm]
10819.0 (n=1,m=1)
10316.8
4.9%
6580.4 (n=1,m=1)
6315.2
4.2%
2059.4 (n=1,m=1)
2000.8
2.9%
905.6 (n=1,m=1)
885.7
2.2%
419.5 (n=1,m=1)
412.3
1.7%
85.7 (n=1,m=1)
84.8
1.1%
65
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
Case C - Simply supported, 500 500mm, = 0.5

FSDT [N/mm]
14425 (n=1,m=1)
13755.8
4.9%
8773.9 (n=1,m=1)
8420.2
4.2%
2745.9 (n=1,m=1)
2667.8
2.9%
1207.5 (n=1,m=1)
1180.9
2.3%
559.3 (n=1,m=1)
549.7
1.7%
114.3 (n=1,m=1)
113.1
1.1%
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm

FSDT [N/mm]
7212.5 (n=1,m=1)
6878.0
4.9%
4386.9 (n=1,m=1)
4210.0
4.2%
1372.9 (n=1,m=1)
1333.9
2.9%
603.7 (n=1,m=1)
590.5
2.2%
279.7 (n=1,m=1)
274.9
1.7%
57.1 (n=1,m=1)
56.6
0.9%
66
1
Thinnest
Thickest
0.9
0.8
0.7
Nycr/Ny0cr
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
Figure 6.22: Interaction diagram for case C, Simply supported plate, 500500mm.
CASE C - 2000500
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm

FSDT [N/mm]
7105.6 (n=1,m=1)
7064.6
0.6%
4300.5 (n=1,m=1)
4279.2
0.5%
1336.7 (n=1,m=1)
1332.1
0.3%
586.3 (n=1,m=1)
584.8
0.3%
271.2 (n=1,m=1)
270.7
0.2%
55.3 (n=1,m=1)
55.3
0%
67
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
Case C - Simply supported, 2000 500mm, = 0.5

FSDT [N/mm]
12580.0 (n=1,m=2)
12348.8
1.9%
7618.6 (n=1,m=2)
7499.2
1.6%
2370.2 (n=1,m=2)
2344.6
1.1%
1040.0 (n=1,m=2)
1031.5
0.8%
481.2 (n=1,m=2)
478.1
0.6%
98.2 (n=1,m=2)
97.8
0.4%
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm

FSDT [N/mm]
3660.5 (n=1,m=1)
3639.4
0.6%
2215.4 (n=1,m=1)
2204.6
0.5%
688.6 (n=1,m=1)
686.3
0.3%
302.0 (n=1,m=1)
301.3
0.2%
139.7 (n=1,m=1)
139.4
0.2%
28.5 (n=1,m=1)
28.5
0%
68
1
Thinnest
0.9
Thickest
0.8
0.7
Nycr/Ny0cr
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
6.3
In-plane Shear Load, FSDT
For simply supported plates subjected to in-plane shear loads, the critical buckling load are found by equation (4.40) from section 4.4. Both groups of equations
(m + n odd and m + n even, suggested by Timoshenko) will be considered, even
for a shorter plate (a/b<2). The results are estimated using Matlab, and listed
below for M = 5 and N = 5, which means the total number of xmn , ymn and
wmn terms are 75. For some cases it is been tested for 147 terms (M = N = 7)
and others 243 terms (M = N = 9), but it makes heavy demands on Matlab. It
could run hours without any results.
69
CASE A - 500500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm

FSDT [N/mm]
404.0 (N=5,M=5)
399.6
1.1%
3081.0 (N=5,M=5)
3019.0
2.1%
9885.0 (N=5,M=5)
9616.6
2.8%
22030.0 (N=5,M=5)
21312.0
3.7%
63800.0 (N=5,M=5)
61330.0
4.0%
To show the results graphically, we plot the buckling coefficient, k0 , against t/b.
We know from section 6.1.1, k0 is given by:
k0 =
Ncr b2
2 D11 D22
9
FSDT
8.5
ANSYS
8
k0
7.5
6.5
5.5
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
t/b
Figure 6.24: Left one shows simply supported plate, case A, 500500mm. Shear buckling
coefficient k0 vs t/b. Right one is from the article "Generic Buckling Curves for Specially Orthotropic Rectangular Plates" written by Brunelle and Oyibo [8]. Here shear buckling coefficient
is plotted against plate affine aspect ratio a0 /b0 using CLPT.
Generalized rigidity ratio, D , for case A1-5 are given by:

A1
D = 0.7531
A2
D = 0.6937
A3
D = 0.6754
A4
D = 0.6666
70
A5
D = 0.6579
According to Brunelle and Oyibo, for those D values listed above, we should
have got k0 around 8.5-9.0 with a0 /b0 around 1 using CLPT. We see that all the
results, both FSDT and ANSYS, are beneath this value. Besides using FSDT
gave us a really good approximation to ANSYS. We have got some useful results.
CASE A - 2000500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm

FSDT [N/mm]
198.9 (N=5,M=5)
197.6
0.7%
1524.0 (N=5,M=5)
1505.2
1.2%
5005.0 (N=5,M=5)
4916.0
1.8%
11520.0 (N=5,M=5)
11264.6
2.3%
36140.0 (N=5,M=5)
35106.0
2.9%
The plate with thickness t = 48.0mm has been tested for M = N = 7. We arrive
at 35970 N/mm, which is 0.47% better than the result with M = N = 5. But
it took twice the running time in Matlab.
4.4
FSDT
4.3
ANSYS
4.2
4.1
k0
4
3.9
3.8
3.7
3.6
3.5
3.4
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
t/b
Figure 6.25: Left one shows simply supported plate, case A, 2000500mm. Shear buckling
coefficient k0 vs t/b. Right one is from the article "Generic Buckling Curves for Specially
Orthotropic Rectangular Plates" written by Brunelle and Oyibo [8], using CLPT.
According to Brunelle and Oyibo, we should have got k0 5 for a0 /b0 3.2
using CLPT. Both FSDT and ANSYS gave us a acceptable result.
71
CASE B - 500500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm

FSDT [N/mm]
416.3 (N=5,M=5)
354.6
17.4%
3347.0 (N=5,M=5)
3013.2
11.1%
10898.0 (N=5,M=5)
9964.8
9.4%
24416.0 (N=5,M=5)
22454.0
8.7%
70700.0 (N=5,M=5)
65436.0
8.0%
Case B1 (t = 8.0mm) with M = N = 7 gives 415.6 N/mm. The result is still

not good enough. We notice that the failure percentages for case B are really
high compared with case A. This could be caused by D16 and D26 , which we
assumed to be zero during the calculation. We can make a test to comfirm the
suspicion. Assume that we have a plate just like case B5, but the plate thickness
and the ply thickness are reduces to 1/6 of the original values (like case B1). In
this way the effects from D16 and D26 are reduced. From ANSYS we obtained
420.2N/mm and from Matlab with M = N = 5, we arrived at 439.1N/mm.
The failure in percent is 4.5%. So the inaccuracy partially caused by neglecting
of D16 and D26 .
9
FSDT
8.5
ANSYS
k0
7.5
6.5
6
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
t/b
Figure 6.26: Left one shows simply supported plate, case B, 500500mm. Shear buckling
Generalized rigidity ratio, D , for case B1-5:
B1
D = 0.7892
B2
D = 0.8831
72
B3
D = 0.9196
B4
D = 0.9388
B5
D = 0.9586
Using CLPT, we should have got shear buckling coefficients around 9.0-9.5 for
a0 /b0 1 according to Brunelle and Oyibo.
CASE B - 2000500
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm

FSDT [N/mm]
226.1 (N=5,M=5)
192.5
17.5%
1971.0 (N=5,M=5)
1796.8
9.7%
6667.0 (N=5,M=5)
6197.2
7.6%
15450.0 (N=5,M=5)
14473.4
6.7%
48030.0 (N=5,M=5)
45144.0
6.4%
Here, the inaccuracy partially caused by neglecting of D16 and D26 .

5.2
FSDT
5
ANSYS
4.8
k0
4.6
4.4
4.2
3.8
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
t/b
Figure 6.27: Left one shows simply supported plate, case B, 2000500mm. Shear buckling
According to Brunelle and Oyibo, we should have got k0 around 5.5-6 for a0 /b0
around 3.5-4 using CLPT.
73
CASE C - 500500
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm

FSDT [N/mm]
45810.0 (N=5,M=5)
44190.0
3.7%
28710.0 (N=5,M=5)
27790.0
3.3%
9383.0 (N=5,M=5)
9153.8
2.5%
4194.0 (N=5,M=5)
4111.4
2.0%
1960.0 (N=5,M=5)
1929.2
1.6%
404.2 (N=5,M=5)
400.0
1.1%
7.5
FSDT
ANSYS
k0
6.5
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
t/b
Figure 6.28: Left one shows simply supported plate, case C, 500500mm. Shear buckling
For case C1-6, D = 0.5679 for all thicknesses. Thereby according to figur (6.28)
by Brunelle and Oyibo, k0 8 for a0 /b0 = 1. We see that both curves from
FSDT and ANSYS are beneath this value.
74
CASE C - 2000500
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm

FSDT [N/mm]
31400.0 (N=5,M=5)
30578.0
2.7%
19360.0 (N=5,M=5)
18911.6
2.4%
6181.0 (N=5,M=5)
6066.4
1.9%
2738.0 (N=5,M=5)
2694.0
1.6%
1274.0 (N=5,M=5)
1255.2
1.5%
261.3 (N=5,M=5)
258.1
1.2%
4.8
FSDT
ANSYS
4.7
4.6
k0
4.5
4.4
4.3
4.2
4.1
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
t/b
Figure 6.29: Left one shows simply supported plate, case C, 2000500mm. Shear buckling
Using CLPT, we should arrive at k0 5 for a0 /b0 = 4 according to Brunelle and
Oyibo. Thus, FSDT gives a better approximation to ANSYS than CLPT.
6.4
Combined Load , FSDT
From section 4.5, equation (4.46) gives critical buckling load for uniaxial x direction load combined with in-plane shear. The results are estimated using the
routines developed in Matlab, and computed for M = 5 and N = 5. The tables
below show the results for = 1 only.
75
CASE A - 500500
FSDT [N/mm]
151.0 (N=5,M=5)
148.7
1.5%
1170.0 (N=5,M=5)
1137.7
2.8%
3830.0 (N=5,M=5)
3698.6
3.6%
8780.0 (N=5,M=5)
8416.0
4.3%
27300.0 (N=5,M=5)
25806.0
5.8%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
We show the results graphically in interaction diagram. x axis denotes uniaxial

loads in x direction, while y axis denotes in-plane shear. Since the results will
not give us a straight line, we have to calculate more values of to obtain the
correct curve.
1
Thinnest
0.9
Thickest
0.8
Middle
Nxycr/Nxy0cr
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
76
CASE A - 2000500
FSDT [N/mm]
113.0 (N=5,M=5)
111.7
1.2%
870.0 (N=5,M=5)
852.4
2.1%
2900.0 (N=5,M=5)
2792.4
3.9%
6600.0 (N=5,M=5)
6428.0
2.8%
21100.0 (N=5,M=5)
20300.0
3.9%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
1
Thinnest
0.9
Thickest
0.8
Nxycr/Nxy0cr
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
77
CASE B - 500500
FSDT [N/mm]
154.0 (N=5,M=5)
139.5
10.4%
1260.0 (N=5,M=5)
1165.8
8.1%
4210.0 (N=5,M=5)
3887.0
8.3%
9720.0 (N=5,M=5)
8930.2
8.8%
30300.0 (N=5,M=5)
27518.0
10.1%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
For case B1 and B3, M = N = 3 gives 155.1N/mm and 4227.0N/mm. We see

that reducing terms from 25 to 9 for xmn , ymn and wmn will not effect the results
that much. But this is only the cese of square plates.
1
Thinnest
0.9
Thickest
0.8
Middle
Nxycr/Nxy0cr
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
78
CASE B - 2000500
FSDT [N/mm]
124.7 (N=5,M=5)
109.1
14.3%
1070.0 (N=5,M=5)
984.8
8.7%
3600.0 (N=5,M=5)
3353.4
7.4%
8500.0 (N=5,M=5)
7811.4
8.8%
26711.1 (N=5,M=5)
24548.0
8.8%
Plate thickness
t = 8.0mm
t = 16.0mm
t = 24.0mm
t = 32.0mm
t = 48.0mm
1
Thinnest
0.9
Thickest
0.8
Nxycr/Nxy0cr
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
79
CASE C - 500500
FSDT [N/mm]
18200.0 (N=5,M=5)
17411.4
4.5%
11200.0 (N=5,M=5)
10730.6
4.4%
3530.0 (N=5,M=5)
3432.0
2.9%
1560.0 (N=5,M=5)
1524.5
2.3%
720.0 (N=5,M=5)
711.0
1.3%
148.0 (N=5,M=5)
146.6
1.0%
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
1
Thinnest
0.9
Thickest
0.8
Nxycr/Nxy0cr
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
80
CASE C - 2000500
FSDT [N/mm]
16306.6 (N=5,M=5)
15597.0
4.5%
10000.0 (N=5,M=5)
9581.2
4.4%
3100.0 (N=5,M=5)
3045.6
1.8%
1400.0 (N=5,M=5)
1348.4
3.8%
640.0 (N=5,M=5)
627.4
2.0%
130.0 (N=5,M=5)
128.9
0.9%
Plate thickness
t = 49.98mm
t = 42.14mm
t = 28.42mm
t = 21.56mm
t = 16.66mm
t = 9.80mm
1
Thinnest
0.9
Thickest
0.8
Nxycr/Nxy0cr
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
N
0.5
/N
xcr
0.6
0.7
0.8
0.9
x0cr
81
6.5
Plates with Initial Geometric Imperfection, FSDT
6.5.1
Displacement - Applied Load
Using equation (4.54) in section 4.6.1, we are now able to plot the load-displacement
curve. This equation is valid only when m = mi and n = ni . Now introducing a
magnification factor:
=
1
1 NNcr
where N is applied load and Ncr is the critical buckling load.

If maximum imperfection is wi , then displacement causing by load N is given by:
wtot = wint (x, y) =
wi sin
mi x
a
sin
N
Ncr
ni y
b
(6.3)
Equations (4.54) and (6.3) are plotted below for comparison with ANSYS. Note
that ANSYS includes non-linear effect. Further, we have chosen the imperfection
magnitude at centre (or maximum imperfection) to be 0.1%, 1%, 2% and 3% of b
wi = 0.5mm, wi = 5mm, wi = 10mm and wi = 15mm. Note that in ANSYS,
the applied loads are twice the current critical loads.
CASE A - 500500
For case A, 500500, load-displacement curves are plotted for (x, y) = (250, 250).
180
160
Applied load, N/mm
140
wtot from sec. 4.6.1

wtot with magn. factor
120
100
80
60
40
20
0
10
15
Displacement, mm
20
25
Figure 6.36: Case A1, 500500. Left one shows load-displacement curve based on equation
(4.54) and (6.3) with imperfection 0.5mm. Right one is from ANSYS.
82
180
160
Applied load, N/mm
140

120
100
80
60
40
20
0
100
200
300
400
500
Displacement, mm
600
700
800
(4.54) and (6.3) with imperfection 15mm. Right one is from ANSYS.
CASE A - 2000500
For case A, 2000 500, we now select an arbitrary point. The load-displacement
curves are plotted for (x, y) = (325, 250).
3.5
x 10
3
Applied load, N/mm
2.5
1.5
0.5
50
100
150
200
Displacement, mm
250
300
CASE B - 500500
For case B, 500 500, we plot for (x, y) = (250, 250).
83
5000
4500
4000
Applied load, N/mm
3500
3000
2500
2000
1500
1000
500
0
50
100
150
200
Displacement, mm
250
300
Figure 6.39: Case B3, 500500. Left one shows load-displacement curve based on equation
5000
4500
4000
Applied load, N/mm
3500
3000
2500
2000
1500
1000
500
0
100
200
300
400
Displacement, mm
500
600
CASE B - 2000500
For case B, 2000 500, we have selected point (x, y) = (250, 250).
84
180
160
Applied load, N/mm
140

120
100
80
60
40
20
0
10
15
Displacement, mm
20
25
CASE C - 500500
For case C, 500 500, the load-displacement curves are plotted for (x, y) =
(250, 250).
2.5
x 10
Applied load, N/mm

1.5
0.5
10
15
Displacement, mm
20
25
Figure 6.42: Case C1, 500500. Left one shows load-displacement curve based on equation
CASE C - 2000500
For case C, 2000 500, we have chosen (x, y) = (250, 250).
85
180
160
Applied load, N/mm
140

120
100
80
60
40
20
0
100
200
300
400
500
Displacement, mm
600
700
800
Figure 6.43: Case C6, 2000500. Left one shows load-displacement curve based on equation
We notice that equation (4.54) matches (6.3) perfectly. Since the models are
confined to small-deflection buckling theory, the graphs from Matlab will never
exceed the current critical buckling loads. ANSYS includes nonlinear effects,
thus its load-displacement curves show us postbuckling behavior. The curve
will continue growing although the critical buckling load is reached. We also
see that load-displacement curve from ANSYS becomes a straight line for larger
value of imperfection. For case C1, the load-displacement curve from ANSYS
is "unstable" (see figure (6.42)). There is a bending of the graph. It can be
interpreted as a change of the buckling modes.
6.5.2
First Ply Failure - Tsai-Wu
Now using equations (4.57) and (4.58) to find the inverse of the "strength ratio".
The analysis is confined to square plates. Only case A and B, and their 3-4 outer
layers will be investigated for first ply failure. The results are estimated using
Matlab. Corresponding analysis is performed in ANSYS. The main differences
between these two methods are 3D Tsai-Wu modeling and including of nonlinear effect in ANSYS. It is been suggested S13 = S12 = 65 in appendix A. But
S13 = S12 = 1000 have also been tested, and it seems that it does not give any
noticeable differences. ANSYS has obtained almost identically results on those
two values. Note that the results are calculated for middle of the layers.
86
CASE A
Simply supported, 500 500mm, Using FSDT

t = 16mm t = 24mm t = 32mm t = 48mm
72.3
153
237.5
364.2
24 (-45)
34 (0)
46 (0)
70 (0)
(250,250) (250,250) (250,250) (250,250)
Imp.=0.1%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate
Case A t = 8mm
19.7
12 (-45)
(250,250)
Imp.=0.1%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate
Case A - Simply supported, 500 500mm, ANSYS

t = 8mm t = 16mm t = 24mm t = 32mm t = 48mm
30.8
72.1
147.1
228.1
357.9
12 (-45)
22 (0)
34 (0)
46 (0)
70 (0)
(250,250) (250,250) (250,250) (250,250) (250,250)
Imp.=1%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate
Case A t = 8mm
10.1
12 (-45)
(250,250)
Imp.=1%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate

27.9
47.1
63.1
90.3
125.0
10 (0)
22 (0)
34 (0)
46 (0)
70 (0)
(250,125) (250,250) (250,250) (250,250) (250,250)
Imp.=2%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate
Case A t = 8mm
6.5
12 (-45)
(250,250)
Imp.=2%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate

25
34.4
45.0
56.3
76.5
10 (0)
22 (0)
34 (0)
46 (0)
70 (0)
(250,62.5) (250,125) (250,250) (250,250) (250,250)

t = 16mm t = 24mm t = 32mm t = 48mm
27.2
48.5
75.0
135.4
24 (-45)
36 (-45)
48 (-45)
70 (0)
(250,250) (250,250) (250,250) (250,250)

t = 16mm t = 24mm t = 32mm t = 48mm
15.4
25.4
36.6
62.1
24 (-45)
36 (-45)
48 (-45)
72 (-45)
(250,250) (250,250) (250,250) (250,250)
87

t = 16mm t = 24mm t = 32mm t = 48mm
10.7
17.0
23.8
38.6
24 (-45)
36 (-45)
48 (-45)
72 (-45)
(250,250) (250,250) (250,250) (250,250)
Imp.=3%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate
Case A t = 8mm
4.8
12 (-45)
(250,250)
Imp.=3%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate

22.5
29.4
37.1
43.8
57.5
10 (0)
22 (0)
34 (0)
46 (0)
70 (0)
(250,62.5) (250,125) (250,125) (250,250) (250,250)
We see that the results are not what we have expected. Almost everything from
Matlab differs from ANSYS: Stresses, ply numbers, coordinates. What is most
surprising is that for some cases, the first ply failure stresses from ANSYS are
lower than stresses obtained by Matlab. Since ANSYS includes this non-linear
effect, it should withstand more. The results from the tables are plotted below.
400
400
Imp. 0.1% of width
Imp. 0.1% of width
350
350
Imp. 1% of width
Imp. 1% of width
300
Imp. 2% of width
FPF Stress [N/mm ]
Imp. 3% of width
250
FPF Stress [N/mm ]
300
200
150
150
100
50
50
10
20
30
40
50
60
b/t
Imp. 3% of width
200
100
Imp. 2% of width
250
10
20
30
40
50
60
b/t
Figure 6.44: Case A, 500500. Left one shows first ply stresses, using FSDT, plotted against
b/t. Right one is from ANSYS.
The graphs have the same shape. Both models give almost identically results
for thick plates with small imperfections. Below, it is been included a plot from
ANSYS that show us the Tsai-wu stress distribution.
88
Figure 6.45: Tsai-Wu stress distribution for case A5 with imperfection 3% of the width.
Plotted in ANSYS.
Different stress values and ply numbers from ANSYS and Matlab can be explained
that computation in ANSYS is based on large-deflection theory and 3D TsaiWu modeling, while Matlab is based on small-deflection theory and 2D Tsai-Wu
model. When the plate is almost perfect (large thickness and small imperfection),
this effect decreases. We will then obtain almost similar results from Matlab and
ANSYS. But it is just a hypothesis and need more investigation.
CASE B
Now, we look at case B with the same model.
Imp.=0.1%
Stress [N/mm2 ]
Ply no.
Coordinate
Case B - Simply supported, 500 500mm, Using FSDT

20.1
40.5
44.6
45.9
47.3
6 (90)
14 (90)
22 (90)
30 (90)
46 (90)
(250,250) (250,250) (250,250) (250,250) (250,250)
89
Imp.=0.1%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate
Case B - Simply supported, 500 500mm, FSDT

25.0
70.9
134.2
187.5
191.7
1 (0)
1 (0)
1 (0)
1 (0)
46 (90)
(250,250) (250,250) (250,250) (250,250)
(500,0)
Imp.=1%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate

11.9
18.8
23.3
26.9
31.5
6 (90)
14 (90)
22 (90)
30 (90)
46 (90)
(250,250) (250,250) (250,250) (250,250) (250,250)
Imp.=1%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate

27.5
46.1
75.0
101.9
142.1
8 (0)
16 (0)
24 (0)
32 (0)
48 (0)
(250,375) (250,250) (250,250) (250,250)
(500,0)
Imp.=2%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate

8.6
12.3
15.7
18.5
23.0
6 (90)
14 (90)
22 (90)
30 (90)
46 (90)
(250,250) (250,250) (250,250) (250,250) (250,250)
Imp.=2%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate

25
34.4
49.2
61.3
85.3
8 (0)
16 (0)
24 (0)
32 (0)
48 (0)
(300,437.5) (250,250) (250,250) (250,250)
(500,0)
Imp.=3%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate

6.7
9.2
11.8
14.2
18.2
6 (90)
14 (90)
22 (90)
30 (90)
46 (90)
(250,250) (250,250) (250,250) (250,250) (250,250)
90
Imp.=3%
Stress [N/mm2 ]
Ply no. (degree)
Coordinate

t = 8mm
t = 16mm t = 24mm t = 32mm t = 48mm
23.5
31.3
38.8
45.3
63.0
8 (0)
16 (0)
24 (0)
32 (0)
48 (0)
(300,437.5) (250,312.5) (250,250) (250,250)
(500,0)
90
200
Imp. 0.1% of width
80
Imp. 2% of width
140
FPF Stress [N/mm ]
Imp. 3% of width
Imp. 3% of width
FPF Stress [N/mm ]
Imp. 2% of width
50
40
30
120
100
80
60
20
40
10
0
Imp. 1% of width
160
70
60
Imp. 0.1% of width
180
Imp. 1% of width
20
10
20
30
40
50
60
b/t
10
20
30
40
50
60
b/t
Figure 6.46: Case B, 500500. Left one shows first ply stresses, using FSDT, plotted against
b/t. Right one is from ANSYS.
Still, the results differ from each other. From Matlab, the first ply failure occurs
at 90 degrees ply every time, while ANSYS at ply 0 degree. Now ANSYS have
got stress values much higher than Matlab. Fifure (6.47) shows us Tsai-wu stress
distribution for case B5. Maximum stress occurs at centre.
91
Figure 6.47: Tsai-Wu stress distribution for case B5 with imperfection 2 % of the width.
Plotted in ANSYS.
Maybe it is not surprising that failure first occurs at 90 degrees plies since the
tension is perpendicular to the fibres. This may be the explanation for the low
stress values. But like case A the main reasons why the results did not match
are computation in ANSYS includes nonlinear effects and is based on 3D TsaiWu modeling. This means that 90 degrees plies withstand more and are not
subjected to tension in the same way as 90 degrees plies under 2D Tsai-Wu
model.
92
7
7.1
Conclusion
Conclusion of Results
This thesis results a method of simplified approaches to the estimation of failure

loads for composite plates. The method is based on FSDT, which includes outof-plane shear deformation.
For uniaxial load in x direction, both simply supported and clamped edges have
been considered. In addition, both thin and thick plate theories (CLPT and
FSDT) have been applied. By solving the buckling equation or equation set, we
are now able to estimate the critical buckling load for simply supported plates.
We see that the results based on FSDT are closer to ANSYS results than the results from CLPT. For a relatively thin plate, the CLPT provides a useful result.
But for a thicker plate, including of out-of-plane deformation makes a noticeable
difference. But still for cases with large thicknesses, not even FSDT is a good
enough method. Here we will have benefit of higher order deformation theories.
For clamped edges, Rayleigh-Ritz method has been chosen to determine the critical buckling load. The FSDT gives a better approximation than the CLPT.
But the model gives us a higher failure percentage compared with the cases with
simply supported edges. It has something to do with the effective length, which
is reduced by 1/2 compared to simply supported edges. But this needs more
investigation.
Simply supported plates subjected to biaxial load have been investigated using
FSDT. By solving the buckling equation set, it is now possible to estimate the
critical buckling load using Matlab. Again, the results are acceptable for thin and
medium thick plates, while the discrepancy is too large for the thickest plates. It
is been computed a several values for (relation between Nx and Ny ) to construct
interaction diagrams. For a 500500 plate from any case, the interaction diagram
shows us a straight, sloped line. This indicates that the relationship between
Nxcr /Nx0cr and Nycr /Ny0cr is inversely proportional. For 2000500 plates, it is
more complicated. Here the interaction diagrams are more parable shaped.
For simply supported plates subjected to in-plane shear, the solution is based
on FSDT. Rayleigh-Ritz method has been chosen to estimate the buckling load.
The accuracy of the results depends on the number of xmn , ymn and wmn terms.
We see that for square plates, it is enough to consider 3MN = 3 3 3 = 27 terms,
but preferably M = N = 5, while it is required minimum 3MN = 3 5 5 = 75
terms for long plates. M = N = 7 or even higher number of M and N will
no doubt give us more accurate answer. But calculating the determinant of a
over 100 100 matrix in Matlab takes eternity (the matrixes are programmed
93
in "forloops", which makes heavy demand on Matlab). Results and time taken
into account, M = N = 5 give us a acceptable answer. So for cases A, B and
C, M = 5 and N = 5 have been chosen to be the standard values. We have got
really good approximations compared with ANSYS. But for case B the model
developed in section 4.4 is not good enough, not even including a higher number
of the xmn , ymn and wmn terms. Primarily, this is caused by neglecting of D16
and D26 . A thin plate with many plies has been tested to confirm this suspicion.
For combinated loads, the model has been developed in the same way as for pure
in-plane shear loading. The investigation has been confined to shear combined
with uniaxial compressive loading and to the case of simply supported plates.
Again, Rayleigh-Ritz method has been used to solve the buckling problem. The
accuracy of the results depends on the number of xmn , ymn and wmn terms.
M = N = 5 give us a acceptable answer. So for cases A-C, M = 5 and N = 5
have been chosen to be the standard values. We see that we have got really
good approximations compared with ANSYS except for case B. The inaccuracy
partially caused by neglecting of D16 and D26 .
For plates with an initial geometric imperfection, it has been cinfined to smalldeflection (linearised) buckling theory. The load-displacement curves from Matlab will never exceed the current critical buckling loads. The analysis in ANSYS
includes nonlinear effects. Thus, its load-displacement curves show the postbuckling behavior. The curve will continue growing although the current critical
buckling load is reached. We also notice that load-dispalcement curve becomes
more straight for larger value of imperfection. Sometimes it also shows us the
change of the buckling modes.
Further, the model developed for simply supported plates with an initial geometric imperfection has been applied to establish the onset of first ply failure. Only
case A and B have been investigated. The method is unsuccessful. There are
large discrepancies between ANSYS results and Matlab results. The deviation
can be explained that computation in ANSYS is based on large-deflection theory
and 3D Tsai-Wu modeling, while Matlab is based on small-deflection theory and
2D Tsai-Wu modeling.
Finally, the methods based on FSDT are better than CLPT. It is best suited for
thin and moderately thick plates. Higher order deformation theories should be
considered for really thick plates. For case A and C, the results are good and
acceptable. For case B, the assumption of specially orthotropic laminates will
affect the results in a bad way. The method is also limited to linear cases.
94
7.2
Suggestions for Futher Work
In general, continue the investigation of simplified approaches to the estimation

of failure loads for composite plates under in-plane loading. Extend the analysis
method based on FSDT to cover cases with clamped edges. It is also interesting to
investigate higher order deformation theories. For thick plates, it will cerntainly
give a better estimation of critical buckling load. For plates with an initial geometric imperfection, extend the approach to include large-deflection (nonlinear)
buckling theory. It is also necessary to take a closer check on the first ply failure
part.
95
References
[1] J.N.Reddy
Mechanics of Laminated Composite Plates - Theory and Analysis
CRC Press, USA, 1st Edition, 1997
[2] Bhagwan D. Agarwal, Lawrence J. Broutman and K. Chandrashekhara
Analysis and Performance of Fiber Composites
Wiley, USA, 3rd Edition, 2006
[3] J.N.Reddy
Mechanics of Laminated Composite Plates and Shells
CRC Press, USA, 2nd Edition, 2004
[4] Geir Skeie
Forelesningsnotat i MEK4560 Elementmetoden i Faststoffmekanikk
UiO, Oslo, 2007
[5] Brian Hayman
Forelesningsnotat i MEK4540 Komposittmatarialer og -konstruksjoner
UiO, Oslo, 2008
[6] Dan Zenkert
An Introduction to Sandwich Structures
Stockholm, Student Edition, 2005
[7] Timoshenko and Gere
Theory of Elastic Stability
McGraw-Hill Book Company, USA, 2nd Edition, 1961
[8] E.J.Brunelle and G.A.Oyibo
Generic Buckling Vurves for Specially Orthotropic Rectangular Plates
The Institute of Aeronautics and Astronautics, USA, 1982
[9] B.Hayman, C.Berggreen, C.Lundsgaard-Larsen, A.Delarche,
H.L.Toftegaard, R.S.Dow, J.Downes, K.Misirlis, N.Tsouvalis and C.Douka
Studies of The Buckling of Composite Plates in Compression
MARSTUCT, Norway-Denmark-UK-Greece, 2009
[10] Christian Jensen
Defects in FRP Panels and their Influence on Compressive Strength
Denmark, Master Thesis, 2006
[11] D. Zenkert and M. Battley
Foundations of Fiber Composites
96
Appendix
A
Parameter Definitions
A.1
[9]
CASE A
Triaxial Lay-up: [45/ + 45/0/ + 45/ 45/0]X,S

Plate aspect ratio: a b
1. AR1: 500 500mm
2. AR2: 2000 500mm
(Sides b are the loaded edges)
Plate thicknesses:
1.
2.
3.
4.
5.
X
X
X
X
X
=1
=2
=3
=4
=6
t = 8.0mm b/t = 62.5

t = 16.0mm b/t = 31.25
t = 24.0mm b/t = 20.83
t = 32.0mm b/t = 15.63
t = 48.0mm b/t = 10.42
Ply thicknesses:
1. t45 = t45 = 0.143mm
2. t0 = 1.714mm
Imperfection magnitude at center: (half sin-wave shaped)
1.
2.
3.
4.
0.1% of b = 0.5mm
1% of b = 5mm
2% of b = 10mm
3% of b = 15mm
Imperfection amplitude equal to full plate width and length.

Material definitions:[MP a]
E1 = 49627
12 = 0.272
G12 = 4800
Xt = 968 ,
, E2 = 15430
, 13 = 0.272
, G13 = 4800
Xc = 915 , Yt
, E3 = 15430
, 23 = 0.607
, G23 = 4800
= 24 , Yc = 118 , S12 = 65 , S23 = 65
97
A.2
CASE B
Quasi-isotropic: [0/ + 45/90/ 45]X,S

1. AR1: 500 500mm
2. AR2: 2000 500mm
Plate thicknesses:
1. X = 1 t = 8.0mm b/t = 62.5
2. X = 2 t = 16.0mm b/t = 31.25
3. X = 3 t = 24.0mm b/t = 20.83
4. X = 4 t = 32.0mm b/t = 15.63
5. X = 6 t = 48.0mm b/t = 10.42
Ply thicknesses: (t is the same for all plies)
1. t = 1mm
1. 0.1% of b = 0.5mm
2. 1% of b = 5mm
3. 2% of b = 10mm
4. 3% of b = 15mm
Material definitions:[MP a] (same as in case A)
E1 = 49627
12 = 0.272
G12 = 4800
Xt = 968 ,
, E2 = 15430
, 13 = 0.272
, G13 = 4800
Xc = 915 , Yt
, E3 = 15430
, 23 = 0.607
, G23 = 4800
= 24 , Yc = 118 , S12 = 65 , S23 = 65
98
A.3
CASE C
Woven Roving: [0]X

1. AR1: 500 500mm
2. AR2: 2000 500mm

Plate thicknesses:
1. X = 51 t = 49.98mm b/t = 10.00
2. X = 43 t = 42.14mm b/t = 11.87
3. X = 29 t = 28.42mm b/t = 17.59
4. X = 22 t = 21.56mm b/t = 23.19
5. X = 17 t = 16.66mm b/t = 30.01
6. X = 10 t = 9.80mm b/t = 51.02

Ply thicknesses: (t is the same for all plies)
1. t = 0.98mm
(tip: Ply thickness can be equal to plate thickness using only one ply
and increasing the number of integration points)
1. 0.1% of b = 0.5mm
2. 1% of b = 5mm
3. 2% of b = 10mm
4. 3% of b = 15mm
Material definitions:[MP a]
E1 = 17180
12 = 0.170
G12 = 3520
Xt = 238.6
S12 = 80.9
,
,
,
,
,
E2 = 17180
13 = 0.270
G13 = 5150
Xc = 324.5
S23 = 60.7
,
,
,
,
E3 = 10800
23 = 0.270
G23 = 5150
Yt = 238.6 , Yc = 324.5
99
B
B.1
Definitions and Equations

Matrixes
Compliance matrix [2]:
[S] =
1
E1
E12L
E212
1
E2
0
0 ,
1
G12
Transformation matrix [2]:
cos2
sin2
2 sin cos
sin2
cos2
2 sin cos ,
[T ] =
sin cos sin cos cos2 sin2
where is the orientation angle.

Stiffness matrix [2]:
[Q] = [S]1
= [T ]1 [Q][T ]
[Q]
Extensional stiffness matrix, A [2]:
n
X
ij )k (hk hk1 )
(Q
Aij =
k=1
Coupling stiffness matrix, B [2]:

n
1X
(Qij )k (h2k h2k1)
Bij =
2 k=1
Bending stiffness matrix, D [2]:
n
1X
(Qij )k (h3k h3k1 ),
Dij =
3
k=1
For FSDT we have to include out of plane shear. In composite laminated plates
the transverse shear stress varies almost quadratically through the thickness.
100
Figure B.1: Example on shear stress variations for different materials [10].
Materialtype
Tykkplate
Homogen
5/6
Kompositt Sandwich
5/6
1
Table B.1: Typical shear correction coefficient [10].

Stiffness matrix for shear [3]:

[Q]shr =
G23 0
0 G13
where Q44 = G23 , Q55 = G13 and Q45 = Q54 = 0.
Transformation matrix for shear [3]:
[T ]shr =
cos sin
sin cos
Extensional stiffness matrix for shear [3]:
Aij = k
n
X
k=1
ij )k (hk hk1 ),
(Q
shr = [T ]1 [Q]shr [T ]shr .

for i = 4, 5 and j = 4, 5. [Q]
shr
101
B.2
Governing Equations for Plates
Figure B.2: Above - A differential element with in-plane froce resultants. Under
- A differential element with moment resultants, shear force resultants and applied
transverse forces [2].
Equilibrium of forces in x dirextion (Figure B.2):

Nx dy + (Nx +
Nxy
Nx
dx)dy Nxy dx + (Nxy +
dy)dx = 0
x
y
Nx Nxy
+
=0
x
y
Equilibrium of forces in y direction (Figure B.2):

102
(B.1)
Ny dx + (Ny +
Ny
Nxy
dx)dx Nxy dy + (Nxy +
dx)dy = 0
y
x
Ny Nxy
+
=0
y
x
(B.2)
Equilibrium of forces in z direction:

1) Without force projections (Figure B.2):
Rxz dy + (Rxz +
Ryz
Rxz
dx)dy Ryz dx + (Ryz +
dy)dx + pdxdy = 0
x
y
Rxz Ryz
+
+p=0
x
y
(B.3)
Figure B.3: Force projections [11].

2) Including force projections (Figure B.3):
Nx
Nyx
w
Nx
w 2 w
2w
Nx w
dy + (Nx +
dx)dy(
+
dx)
=
N
dxdy
+
dxdy
x
x
x
x
x2
x2
x x
w
Nyx
w 2 w
2w
Nyx w
dx+(Nyx +
dy)dx(
+
dy) = Nyx
dxdy+
dxdy
x
y
x xy
xy
y x
Ny
w
Ny
w 2 w
2w
Ny w
dx + (Ny +
dy)dx(
+
dy)
=
N
dxdy +
dxdy
y
2
2
y
y
y
y
y
y y
103
w
Nxy
w 2 w
2w
Nxy w
Nxy
dy+(Nxy +
dx)dy(
+
dx) = Nxy
dxdy+
dxdy
y
x
y xy
xy
x y
Summing all contributions for the in plane forces and we arrive at:
Nx
2w
2w
2w
2w
dxdy
+
N
dxdy
+
N
dxdy
+
N
dxdy
yx
y
xy
x2
xy
y 2
xy
w
Nx Nyx
w
Ny Nxy
+ dxdy(
+
)+
dxdy(
+
)
x
x
y
y
y
x
2w
2w
2w
2w
+ Nxy
= Nx 2 + Ny 2 + Nxy
x
y
xy
xy
We have used equations (B.1) and (B.2) to get rid of parentheses.

Now with equation (B.3), the final expression for the force equilibrium in z direction is then:
2w
2w
2w
2w
2w
Rxz Ryz
+
+ p + Nx 2 + Ny 2 + Nxy
+ Nyx
2 = 0 (B.4)
x
y
x
y
xy
xy
t
where is the surface weight or mass of the plate.
By summing moments about x axis (Figure B.2):
My dx (My +
Mxy
Ryz
My
dy)dx + Mxy dy (Mxy +
dx)dy + (Ryz +
dy)dxdy
y
x
y
+(Rxz +
Rxz
dy
dy
dy
dx)dy Rxz dy + pdxdy
=0
x
2
2
2
When higher-order terms are neglected, this equation simplifies to:

My Mxy
+
Ryz = 0
y
x
(B.5)
By summing moments about y axis (Figure B.2):
Mx dy (Mx +
Mx
Mxy
Rxz
dx)dy + Mxy dx (Mxy +
dy)dx + (Rxz +
dx)dxdy
x
y
x
104
+(Ryz +
Ryz
dx
dx
dx
dy)dx Ryz dx + pdxdy
=0
y
2
2
2
Similarly, we neglect the higher-order terms:

Mx Mxy
+
Rxz = 0
x
y
(B.6)
From equations (B.5) and (B.6), we get the expressions for shear:
Rxz =
Mx Mxy
+
x
y
Ryz =
My Mxy
+
y
x
Now, substitution of equations above in equation (B.4) gives:
2 Mxy 2 My
2 Mx
+
2
+
x2
xy
y 2
2w
2w
2w
2w
+p + Nx 2 + Ny 2 + 2Nxy
2 = 0
x
y
xy
t
C
C.1
Program Codes
Matlab
Simply supported CLPT (uniaxial), Case A

% Frittopplagt kompositt tynnplate Masteroppgaven
% Case A
clear all;
%Materialdata
EL=49627;
ET=15430;
vLT=0.272;
GLT=4800;
a=500;
b=500;
n=1;
m=1;
% Orientering av lag i radianer og tykkelse av lagene
%R=[-45 45 0 45 -45 0 0 -45 45 0 45 -45]*pi/180;
105
(B.7)
%h=[-4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4];
R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
h=[-8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857
2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45
0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-12 -11.857 -11.714 -10 -9.857 -9.714 -8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857
-3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8
9.714 9.857 10 11.714 11.857 12];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45
0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-16 -15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8
-7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2
3.714 3.857 4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714
13.857 14 15.714 15.857 16];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0
45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45
0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-24 -23.857 -23.714 -22 -21.857 -21.714 -20 -19.857 -19.714 -18 -17.857 -17.714
-16 -15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8
-7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2
3.714 3.857 4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714
13.857 14 15.714 15.857 16 17.714 17.857 18 19.714 19.857 20 21.714 21.857 22
23.714 23.857 24];
%Initielverdi
A(1:3,1:3)=0;
B(1:3,1:3)=0;
D(1:3,1:3)=0;
Q(1:3,1:3)=0;
S(1:3,1:3)=0;
%Compliance (S-matirx)
S(1,1)=1/EL;
S(2,2)=1/ET;
S(1,2)=-vLT/EL;
S(3,3)=1/GLT;
S(2,1)=S(1,2);
%Stiffness (Q-matrix)
Q=inv(S);
Q(3,3)=2*Q(3,3);
%for tensorielle tynigner ved rotasjon
for i=1:length(R)
%Beregner T og Q for hvert lag k. Summerer A, B og D
%Transformation matrix (T-matrix)

T(1,1)=cos(R(i))^2;
T(2,2)=cos(R(i))^2;
T(1,2)=sin(R(i))^2;
T(2,1)=sin(R(i))^2;
T(3,1)=-sin(R(i))*cos(R(i));
T(3,2)=sin(R(i))*cos(R(i));
T(1,3)=2*sin(R(i))*cos(R(i));
T(2,3)=-2*sin(R(i))*cos(R(i));
T(3,3)=cos(R(i))^2-sin(R(i))^2;
Qk(:,:,i)=inv(T)*Q*T;
Qk(:,3,i)=Qk(:,3,i)/2;
A=A+Qk(:,:,i)*(h(i+1)-h(i));
B=B+.5*Qk(:,:,i)*(h(i+1)^2-h(i)^2);
%Stivhetsmatrise for tensorielle tyninger

%Gr tilbake til ingenirtyninger etter rotasjon
%Beregner A-matrisen
%Beregner B-matrisen
106
D=D+1/3*Qk(:,:,i)*(h(i+1)^3-h(i)^3);
%Beregner D-matrisen
end
A
B
D
P=pi^2*(D(1,1)*(m/a)^2+(2*D(1,2)+4*D(3,3))*(n/b)^2+D(2,2)*(a/m)^2*(n/b)^4)%Kritisk verdi for P
% Beregning av knekningskoeffisient
knekkoeff=P*b^2/(pi^2*(D(1,1)*D(2,2))^(1/2))
% Knekningskoeffisient
DD=(D(1,2)+2*D(3,3))/((D(1,1)*D(2,2))^(1/2))
data=knekkoeff-2*DD
Clamped CLPT (uniaxial), Case A

% Fastinnspent kompositt tynnplate Masteroppgaven
% Case A
clear all;
%Materialdata
EL=49627;
ET=15430;
vLT=0.272;
GLT=4800;
a=2000;
b=500;
n=3;
m=5;
%R=[-45 45 0 45 -45 0 0 -45 45 0 45 -45]*pi/180;
%h=[-4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857
2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45
0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-12 -11.857 -11.714 -10 -9.857 -9.714 -8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857
-3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8
9.714 9.857 10 11.714 11.857 12];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0
45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-16 -15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8
-7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2
3.714 3.857 4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714
13.857 14 15.714 15.857 16];
R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45
-45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45
45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
h=[-24 -23.857 -23.714 -22 -21.857 -21.714 -20 -19.857 -19.714 -18 -17.857 -17.714 -16
-15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8 -7.857
-7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857
4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714 13.857 14 15.714
15.857 16 17.714 17.857 18 19.714 19.857 20 21.714 21.857 22 23.714 23.857 24];
107
%Initielverdi
A(1:3,1:3)=0;
B(1:3,1:3)=0;
D(1:3,1:3)=0;
Q(1:3,1:3)=0;
S(1:3,1:3)=0;
S(1,1)=1/EL;
S(2,2)=1/ET;
S(1,2)=-vLT/EL;
S(3,3)=1/GLT;
S(2,1)=S(1,2);
Q=inv(S);
Q(3,3)=2*Q(3,3);
for i=1:length(R)
Beregner T og Q for hvert lag k. Summerer A, B og D

T(1,1)=cos(R(i))^2;
T(2,2)=cos(R(i))^2;
T(1,2)=sin(R(i))^2;
T(2,1)=sin(R(i))^2;
T(3,1)=-sin(R(i))*cos(R(i));
T(1,3)=2*sin(R(i))*cos(R(i));
T(2,3)=-2*sin(R(i))*cos(R(i));
T(3,3)=cos(R(i))^2-sin(R(i))^2;
Qk(:,3,i)=Qk(:,3,i)/2;
A=A+Qk(:,:,i)*(h(i+1)-h(i));
B=B+.5*Qk(:,:,i)*(h(i+1)^2-h(i)^2);
D=D+1/3*Qk(:,:,i)*(h(i+1)^3-h(i)^3);

%Beregner A-matrisen
%Beregner B-matrisen
%Beregner D-matrisen
end
A
B
D
if m==n
P=(4*pi^2*(3*D(1,1)*m^4*b^4+3*D(2,2)*a^4+4*m^2*D(3,3)*a^2*b^2+2*m^2*D(1,2)*a^2*b^2))/(3*b^4*m^2*a^2)
else
P=(pi^2/a^2*(n^4+6*m^2*n^2+m^4)*D(1,1)+16/3*a^2/b^4*pi^2*D(2,2)+16/3*pi^2/b^2*(n^2+m^2)*(1/2*D(1,2)
+D(3,3)))/(n^2+m^2)
end
L1=pi^4/(4*a^3*b^3)*(4*D(3,3)*a^2*b^2+2*D(1,2)*a^2*b^2+3*D(2,2)*a^4+3*D(1,1)*b^4);
L2=3*pi^2*b/(16*a);
L3=pi^4*a^4*D(2,2)/(2*a^3*b^3);
L4=pi^4/(4*a^3*b^3)*(8*D(1,2)*a^2*b^2+3*a^4*D(2,2)+16*D(3,3)*a^2*b^2+48*D(1,1)*b^4);
L5=3*pi^2*b/(4*a);
P2=((L1*L5+L2*L4)-sqrt((L1*L5+L2*L4)^2-4*L2*L5*(L1*L4-L3*L3)))/(2*L2*L5)
Simply supported FSDT (uniaxial+biaxial+imperfection), Case A

% Frittopplagt kompositt tykkplate Masteroppgaven
% Case A
108
clear all;
%Materialdata
EL=49627;
ET=15430;
vLT=0.272;
GLT=4800;
G23=4800;
G13=4800;
a=2000;
b=500;
k=5/6;
n=1;
m=3;
ni=1;
mi=3;
alpha=m*pi/a;
beta=n*pi/b;
alphai=mi*pi/a;
gamma=0.9;
x=325;
y=b/2;
Wi=5;
N1=0:1:31300;
%R=[-45 45 0 45 -45 0 0 -45 45 0 45 -45]*pi/180;
%h=[-4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4];
%t=8;
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857
2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8];
%t=16;
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45
0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-12 -11.857 -11.714 -10 -9.857 -9.714 -8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857
-3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8
9.714 9.857 10 11.714 11.857 12];
%t=24;
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45
0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-16 -15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8
-7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2
3.714 3.857 4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714
13.857 14 15.714 15.857 16];
%t=32;
R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45
-45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45
45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
h=[-24 -23.857 -23.714 -22 -21.857 -21.714 -20 -19.857 -19.714 -18 -17.857 -17.714 -16
-15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8 -7.857
-7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857
4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714 13.857 14 15.714
15.857 16 17.714 17.857 18 19.714 19.857 20 21.714 21.857 22 23.714 23.857 24];
t=48;
%Initielverdi
A(1:3,1:3)=0;
B(1:3,1:3)=0;
D(1:3,1:3)=0;
Q(1:3,1:3)=0;
109
S(1:3,1:3)=0;
A1(1:2,1:2)=0;
Q1(1:2,1:2)=0;
%A-matrisen med skjr

%Q-matrisen med skjr
S(1,1)=1/EL;
S(2,2)=1/ET;
S(1,2)=-vLT/EL;
S(3,3)=1/GLT;
S(2,1)=S(1,2);
Q=inv(S);
Q1(1,1)=G23;
Q1(2,2)=G13;
Q1(1,2)=0;
Q1(2,1)=Q1(1,2);
Q(3,3)=2*Q(3,3);
for i=1:length(R)

T(1,1)=cos(R(i))^2;
T(2,2)=cos(R(i))^2;
T(1,2)=sin(R(i))^2;
T(2,1)=sin(R(i))^2;
T(3,1)=-sin(R(i))*cos(R(i));
T(1,3)=2*sin(R(i))*cos(R(i));
T(2,3)=-2*sin(R(i))*cos(R(i));
T(3,3)=cos(R(i))^2-sin(R(i))^2;
%Transformasjon matrise for skjr
T1(1,1)=cos(R(i));
T1(1,2)=-sin(R(i));
T1(2,1)=sin(R(i));
T1(2,2)=cos(R(i));
Qk(:,3,i)=Qk(:,3,i)/2;
Qk1(:,:,i)=inv(T1)*Q1*T1;

%Skjr
A=A+Qk(:,:,i)*(h(i+1)-h(i));
B=B+.5*Qk(:,:,i)*(h(i+1)^2-h(i)^2);
D=D+1/3*Qk(:,:,i)*(h(i+1)^3-h(i)^3);
A1=A1+Qk1(:,:,i)*(h(i+1)-h(i));
%Beregner
%Beregner
%Beregner
%Beregner
A-matrisen
B-matrisen
D-matrisen
A1-matrisen
end
A
B
D
A1
C1=-D(1,1)*alpha^2-D(3,3)*beta^2-k*A1(2,2);
C2=-D(1,2)*alpha*beta-D(3,3)*alpha*beta;
C3=-k*A1(2,2)*alpha;
C4=-D(2,2)*beta^2-D(3,3)*alpha^2-k*A1(1,1);
C5=-k*A1(1,1)*beta;
P=(C1*(C5)^2+alpha*C3*(C2)^2+beta*(C2)^2*C5+(C3)^2*C4-alpha*C1*C3*C4
-beta*C1*C4*C5-2*C2*C3*C5)/(alpha^2*C1*C4-alpha^2*(C2)^2) %Knekningslast i x-retning
P2=(C1*(C5)^2+alpha*C3*(C2)^2+beta*(C2)^2*C5+(C3)^2*C4-alpha*C1*C3*C4
-beta*C1*C4*C5-2*C2*C3*C5)/(beta^2*C1*C4-beta^2*(C2)^2) %Knekningslast i y-retning
110
N=(C1*C5*C5+alpha*C2*C2*C3+beta*C2*C2*C5+C3*C3*C4-alpha*C1*C3*C4-beta*C1*C4*C5-2*C2*C3*C5)
/(alpha*alpha*C1*C4+gamma*beta*beta*C1*C4-alpha*alpha*C2*C2-gamma*beta*beta*C2*C2) %Biaksial
% Imperfeksjon
W=-N1.*alphai.^2.*Wi./((C5-C2.*C3./C1).*(C2.*C3./C1-C5)./(C4-C2.*C2./C1)-C3.*C3./C1+alpha.*C3+beta.*C5+N1.*alpha.^2);
def=W.*sin(m.*pi.*x./a).*sin(n.*pi.*y./b)+Wi.*sin(mi.*pi.*x./a).*sin(ni.*pi.*y./b);
afaktor=(Wi.*sin(mi.*pi.*x./a).*sin(ni.*pi.*y./b))./(1-(N1./P));
plot(def,N1,r,afaktor,N1,--)
xlabel(Displacement, mm)
ylabel(Applied load, N/mm)
legend(w_{tot} from sec. 4.6.1,w_{tot} with magn. factor, 0)
% Beregning av knekningskoeffisient
Pansys=30494;
% Knekningsverdi fra ANSYS (input)
knekkoeff2=Pansys*b^2/(pi^2*(D(1,1)*D(2,2))^(1/2))
% Knekningskoeffisient for ANSYS
DD=(D(1,2)+2*D(3,3))/((D(1,1)*D(2,2))^(1/2));
data=knekkoeff-2*DD
data2=knekkoeff2-2*DD
% den modifiserte knek.koeff

% den modifiserte knek.koeff for ANSYS
Clamped edges FSDT (uniaxial), Case A

% Fastinnspent kompositt tykkplate Masteroppgaven
% Case A
clear all;
%Materialdata
EL=49627;
ET=15430;
vLT=0.272;
GLT=4800;
G23=4800;
G13=4800;
a=2000;
b=500;
k=5/6;
n=5;
m=5;
alpha=m*pi/a;
beta=n*pi/b;
%R=[-45 45 0 45 -45 0 0 -45 45 0 45 -45]*pi/180;
%h=[-4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857
2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45
0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-12 -11.857 -11.714 -10 -9.857 -9.714 -8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857
-3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8
9.714 9.857 10 11.714 11.857 12];
111
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45

0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-16 -15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8
-7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2
3.714 3.857 4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714
13.857 14 15.714 15.857 16];
R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45
-45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45
45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
h=[-24 -23.857 -23.714 -22 -21.857 -21.714 -20 -19.857 -19.714 -18 -17.857 -17.714 -16
-15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8 -7.857
-7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857
4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714 13.857 14 15.714
15.857 16 17.714 17.857 18 19.714 19.857 20 21.714 21.857 22 23.714 23.857 24];
%Initielverdi
A(1:3,1:3)=0;
B(1:3,1:3)=0;
D(1:3,1:3)=0;
Q(1:3,1:3)=0;
S(1:3,1:3)=0;
A1(1:2,1:2)=0;
Q1(1:2,1:2)=0;

S(1,1)=1/EL;
S(2,2)=1/ET;
S(1,2)=-vLT/EL;
S(3,3)=1/GLT;
S(2,1)=S(1,2);
Q=inv(S);
Q1(1,1)=G23;
Q1(2,2)=G13;
Q1(1,2)=0;
Q1(2,1)=Q1(1,2);
Q(3,3)=2*Q(3,3);
for i=1:length(R)

T(1,1)=cos(R(i))^2;
T(2,2)=cos(R(i))^2;
T(1,2)=sin(R(i))^2;
T(2,1)=sin(R(i))^2;
T(3,1)=-sin(R(i))*cos(R(i));
T(1,3)=2*sin(R(i))*cos(R(i));
T(2,3)=-2*sin(R(i))*cos(R(i));
T(3,3)=cos(R(i))^2-sin(R(i))^2;
T1(1,1)=cos(R(i));
T1(1,2)=-sin(R(i));
T1(2,1)=sin(R(i));
T1(2,2)=cos(R(i));
Qk(:,3,i)=Qk(:,3,i)/2;
Qk1(:,:,i)=inv(T1)*Q1*T1;

%Skjr
112
A=A+Qk(:,:,i)*(h(i+1)-h(i));
B=B+.5*Qk(:,:,i)*(h(i+1)^2-h(i)^2);
D=D+1/3*Qk(:,:,i)*(h(i+1)^3-h(i)^3);
A1=A1+Qk1(:,:,i)*(h(i+1)-h(i));
%Beregner
%Beregner
%Beregner
%Beregner
A-matrisen
B-matrisen
D-matrisen
A1-matrisen
end
A
B
D
A1
%L1=pi^4/(4*a^3*b^3)*(3*D(1,1)*b^4+4*D(3,3)*a^2*b^2+2*D(1,2)*a^2*b^2+3*D(2,2)*a^4);
%L2=k*pi^2/(4*a*b)*(A1(1,1)*a^2+A1(2,2)*b^2);
%P2=(18*(36*L1*pi^2+27*pi^2*L2-sqrt(1296*L1*L1*pi^4-1944*L1*pi^4*L2+729*pi^4*L2*L2+262144*L1*L2))*a)
/(b*(243*pi^4-16384))
C1=pi^4/(8*a^3*b^3)*(6*a^4*D(2,2)+4*D(1,2)*a^2*b^2+8*D(3,3)*a^2*b^2+6*D(1,1)*b^4);
C2=270*pi^2*b/(1440*a);
C3=pi^4/(8*a^3*b^3)*4*a^4*D(2,2);
C4=-2560*b/(1440*a);
C5=pi^4/(8*a^3*b^3)*(6*a^4*D(2,2)+32*D(3,3)*a^2*b^2+96*D(1,1)*b^4+16*D(1,2)*a^2*b^2);
C6=1080*pi^2*b/(1440*a);
C7=-2048*b/(1440*a);
C8=k*pi^2/(4*a*b)*(A1(1,1)*a^2+A1(2,2)*b^2);
C9=360*pi^2*b/(1440*a);
U1=C4^2*C6+C7^2*C2-C2*C6*C9
U2=C2*C6*C8+C1*C6*C9+C2*C5*C9+2*C3*C4*C7-C1*C7^2-C4^2*C5
U3=C3^2*C9-C1*C6*C8-C2*C5*C8-C1*C5*C9
U4=C1*C5*C8-C3*C3*C8
F1=pi^4/(32*a^3*b^3)*(18*m^2*D(1,1)*b^4+8*D(1,2)*a^2*b^2+16*D(3,3)*b^2*a^2+16*D(3,3)*b^2*m^2*a^2
+8*m^2*D(1,2)*a^2*b^2+3*D(1,1)*b^4+3*m^4*D(1,1)*b^4+16*a^4*D(2,2));
F2=(-b/192*(18*pi^2*m^6+18*pi^2+18*pi^2*n^4-18*pi^2*m^4-18*pi^2*m^2+18*pi^2*m^2*n^4-72*pi^2*m^2*n^2
-36*pi^2*n^2-36*pi^2*m^4*n^2))/(a*(1-2*n^2+m^4-2*m^2*n^2-2*m^2+n^4));
F3=(-b/192*(-512*n^3*m+512*(-1)^(m+n+1)*m*n^3))/(a*(1-2*n^2+m^4-2*m^2*n^2-2*m^2+n^4));
F4=k*pi^2/(4*a*b)*(A1(1,1)*a^2+A1(2,2)*n^2*b^2);
F5=(-b/192*(48*pi^2*n^2+48*pi^2*n^6-96*pi^2*m^2*n^4+48*pi^2*m^4*n^2-96*pi^2*n^4-96*pi^2*m^2*n^2))
/(a*(1-2*n^2+m^4-2*m^2*n^2-2*m^2+n^4));
G1=pi^4/(4*a^3*b^3)*(3*D(1,1)*b^4+2*D(1,2)*a^2*b^2+3*D(2,2)*a^4+4*D(3,3)*a^2*b^2);
G2=-(pi^2*b*(3*n^4-12*n^2))/(16*a*(n^4-4*n^2));
G3=-(8*b*(-n^3+(-1)^(n+2)*n^3))/(3*a*(n^4-4*n^2));
G4=k*pi^2/(4*a*b)*(A1(1,1)*a^2+A1(2,2)*n^2*b^2);
G5=-(pi^2*b*(n^6-4*n^4))/(4*a*(n^4-4*n^2));
if m==1
P=(-(G1*G5+G2*G4)-sqrt((G1*G5+G2*G4)^2-4*(G2*G5-G3*G3)*G1*G4))/(2*(G2*G5-G3*G3))
else
P=(-(F1*F5+F2*F4)-sqrt((F1*F5+F2*F4)^2-4*(F2*F5-F3*F3)*F1*F4))/(2*(F2*F5-F3*F3))
end
In-plane shear FSDT, Case A

% Case A
clear all;
%Materialdata
EL=49627;
ET=15430;
vLT=0.272;
113
GLT=4800;
G23=4800;
G13=4800;
a=2000;
b=500;
k=5/6;
M=5;
N=5;
%R=[-45 45 0 45 -45 0 0 -45 45 0 45 -45]*pi/180;
%h=[-4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857
2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8];
R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45
0 45 -45 0 -45 45 0 45 -45]*pi/180;
h=[-12 -11.857 -11.714 -10 -9.857 -9.714 -8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857
-3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8
9.714 9.857 10 11.714 11.857 12];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45
0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-16 -15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8
-7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2
3.714 3.857 4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714
13.857 14 15.714 15.857 16];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0
45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45
0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-24 -23.857 -23.714 -22 -21.857 -21.714 -20 -19.857 -19.714 -18 -17.857 -17.714
-16 -15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8
-7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2
3.714 3.857 4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714
13.857 14 15.714 15.857 16 17.714 17.857 18 19.714 19.857 20 21.714 21.857 22
23.714 23.857 24];
%Initielverdi
A(1:3,1:3)=0;
B(1:3,1:3)=0;
D(1:3,1:3)=0;
Q(1:3,1:3)=0;
S(1:3,1:3)=0;
A1(1:2,1:2)=0;
Q1(1:2,1:2)=0;

%Compliance (S-matirx)1/4*k*A1(1,1)*pi^2*n^2*a/b+1/4*k*A1(2,2)*m^2*pi^2*b/a
S(1,1)=1/EL;
S(2,2)=1/ET;
S(1,2)=-vLT/EL;
S(3,3)=1/GLT;
S(2,1)=S(1,2);
Q=inv(S);
Q1(1,1)=G23;
Q1(2,2)=G13;
Q1(1,2)=0;
Q1(2,1)=Q1(1,2);
Q(3,3)=2*Q(3,3);
114
for i=1:length(R)

T(1,1)=cos(R(i))^2;
T(2,2)=cos(R(i))^2;
T(1,2)=sin(R(i))^2;
T(2,1)=sin(R(i))^2;
T(3,1)=-sin(R(i))*cos(R(i));
T(1,3)=2*sin(R(i))*cos(R(i));
T(2,3)=-2*sin(R(i))*cos(R(i));
T(3,3)=cos(R(i))^2-sin(R(i))^2;
T1(1,1)=cos(R(i));
T1(1,2)=-sin(R(i));
T1(2,1)=sin(R(i));
T1(2,2)=cos(R(i));
Qk(:,3,i)=Qk(:,3,i)/2;
Qk1(:,:,i)=inv(T1)*Q1*T1;

%Skjr
A=A+Qk(:,:,i)*(h(i+1)-h(i));
B=B+.5*Qk(:,:,i)*(h(i+1)^2-h(i)^2);
D=D+1/3*Qk(:,:,i)*(h(i+1)^3-h(i)^3);
A1=A1+Qk1(:,:,i)*(h(i+1)-h(i));
%Beregner
%Beregner
%Beregner
%Beregner
A-matrisen
B-matrisen
D-matrisen
A1-matrisen
end
A;
B;
D;
A1;
matrise1=zeros(M*N,M*N);
matrise10=sym(zeros(M*N,M*N));
Nxy=sym(Nxy);
% dpi/dx med x-ledd
for m=1:M
for n=1:N
matrise1(N*(m-1)+n,N*(m-1)+n)=1/4*pi^2*m^2*b/a*D(1,1)
+1/4*pi^2*n^2*a/b*D(3,3)+1/4*k*a*b*A1(2,2);
end
end
matrise1;
% dpi/dx med y-ledd
for m=1:M
for n=1:N
matrise2(N*(m-1)+n,N*(m-1)+n)=1/4*pi^2*m*n*D(1,2)
+1/4*pi^2*m*n*D(3,3);
end
end
matrise2;
115
% dpi/dx med w-ledd

for m=1:M
for n=1:N
matrise3(N*(m-1)+n,N*(m-1)+n)=1/4*k*A1(2,2)*b*m*pi;
end
end
matrise3;
% dpi/dy med x-ledd
for m=1:M
for n=1:N
matrise4(N*(m-1)+n,N*(m-1)+n)=1/4*pi^2*m*n*D(1,2)+1/4*pi^2*m*n*D(3,3);
end
end
matrise4;
% dpi/dy med y-ledd
for m=1:M
for n=1:N
matrise5(N*(m-1)+n,N*(m-1)+n)=1/4*pi^2*n^2*a/b*D(2,2)
+1/4*pi^2*m^2*b/a*D(3,3)+1/4*k*A1(1,1)*a*b;
end
end
matrise5;
% dpi/dy med w-ledd
for m=1:M
for n=1:N
matrise6(N*(m-1)+n,N*(m-1)+n)=1/4*k*A1(1,1)*n*pi*a;
end
end
matrise6;
% dpi/dw med x-ledd
for m=1:M
for n=1:N
matrise7(N*(m-1)+n,N*(m-1)+n)=1/4*k*A1(2,2)*m*pi*b;
end
end
matrise7;
% dpi/dw med y-ledd
for m=1:M
for n=1:N
end
end
matrise8;
% dpi/dw med w-ledd
for m=1:M
for n=1:N
matrise9(N*(m-1)+n,N*(m-1)+n)=1/4*k*A1(1,1)*pi^2*n^2*a/b
+1/4*k*A1(2,2)*m^2*pi^2*b/a;
end
end
matrise9;
116
% dpi/dw med w-leddpq

for m=1:M
for n=1:N
for p=1:M
for q=1:N
if m==p
elseif n==q
elseif rem(m-p,2)==0
elseif rem(n-q,2)==0
elseif rem(m+p,2)==0
elseif rem(n+q,2)==0
else
i=(m-1)*N+n;
j=(p-1)*N+q;
matrise10(i,j)=-8*Nxy*m*n*p*q/((m^2-p^2)*(n^2-q^2));
end
end
end
end
end
matrise10;
matrisetot=[matrise1 matrise2 matrise3; matrise4 matrise5
matrise6; matrise7 matrise8 matrise9+matrise10];
Nxy=solve(det(matrisetot));
double(Nxy)
% velger ut den minste verdien, fortegn spiller ingen rolle
%Beregning av knekningskoeff.
Pansys=35106;
P=36140;
% Knekningsverdi fra ANSYS (input)

% Knekningsverdi fra Matlab (input)
knekkoeff2=Pansys*b^2/(pi^2*(D(1,1)*D(2,2))^(1/2))
% Knekningskoeffisient for ANSYS
DD=(D(1,2)+2*D(3,3))/((D(1,1)*D(2,2))^(1/2))
data=knekkoeff-2*DD
data2=knekkoeff2-2*DD
% den modifiserte knek.koeff

% den modifiserte knek.koeff for ANSYS
Combined load FSDT, Case A

% Case A
clear all;
%Materialdata
EL=49627;
ET=15430;
vLT=0.272;
GLT=4800;
G23=4800;
G13=4800;
a=500;
b=500;
k=5/6;
mu=1;
117
M=3;
N=3;
R=[-45 45 0 45 -45 0 0 -45 45 0 45 -45]*pi/180;
h=[-4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857
2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45
0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-12 -11.857 -11.714 -10 -9.857 -9.714 -8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857
-3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8
9.714 9.857 10 11.714 11.857 12];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45
0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-16 -15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8
-7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2
3.714 3.857 4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714
13.857 14 15.714 15.857 16];
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0
45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45
0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-24 -23.857 -23.714 -22 -21.857 -21.714 -20 -19.857 -19.714 -18 -17.857 -17.714
-16 -15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8
-7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2
3.714 3.857 4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714
13.857 14 15.714 15.857 16 17.714 17.857 18 19.714 19.857 20 21.714 21.857 22
23.714 23.857 24];
%Initielverdi
A(1:3,1:3)=0;
B(1:3,1:3)=0;
D(1:3,1:3)=0;
Q(1:3,1:3)=0;
S(1:3,1:3)=0;
A1(1:2,1:2)=0;
Q1(1:2,1:2)=0;

%Compliance (S-matirx)1/4*k*A1(1,1)*pi^2*n^2*a/b+1/4*k*A1(2,2)*m^2*pi^2*b/a
S(1,1)=1/EL;
S(2,2)=1/ET;
S(1,2)=-vLT/EL;
S(3,3)=1/GLT;
S(2,1)=S(1,2);
Q=inv(S);
Q1(1,1)=G23;
Q1(2,2)=G13;
Q1(1,2)=0;
Q1(2,1)=Q1(1,2);
Q(3,3)=2*Q(3,3);
for i=1:length(R)

T(1,1)=cos(R(i))^2;
T(2,2)=cos(R(i))^2;
118
T(1,2)=sin(R(i))^2;
T(2,1)=sin(R(i))^2;
T(3,1)=-sin(R(i))*cos(R(i));
T(1,3)=2*sin(R(i))*cos(R(i));
T(2,3)=-2*sin(R(i))*cos(R(i));
T(3,3)=cos(R(i))^2-sin(R(i))^2;
T1(1,1)=cos(R(i));
T1(1,2)=-sin(R(i));
T1(2,1)=sin(R(i));
T1(2,2)=cos(R(i));
Qk(:,3,i)=Qk(:,3,i)/2;
Qk1(:,:,i)=inv(T1)*Q1*T1;

%Skjr
A=A+Qk(:,:,i)*(h(i+1)-h(i));
B=B+.5*Qk(:,:,i)*(h(i+1)^2-h(i)^2);
D=D+1/3*Qk(:,:,i)*(h(i+1)^3-h(i)^3);
A1=A1+Qk1(:,:,i)*(h(i+1)-h(i));
%Beregner
%Beregner
%Beregner
%Beregner
A-matrisen
B-matrisen
D-matrisen
A1-matrisen
end
A;
B;
D;
A1;
Nk=sym(Nk);
% dpi/dx med x-ledd
for m=1:M
for n=1:N
matrise1(N*(m-1)+n,N*(m-1)+n)=1/4*pi^2*m^2*b/a*D(1,1)
+1/4*pi^2*n^2*a/b*D(3,3)+1/4*k*a*b*A1(2,2);
end
end
matrise1;
% dpi/dx med y-ledd
for m=1:M
for n=1:N
matrise2(N*(m-1)+n,N*(m-1)+n)=1/4*pi^2*m*n*D(1,2)
+1/4*pi^2*m*n*D(3,3);
end
end
matrise2;
% dpi/dx med w-ledd
for m=1:M
for n=1:N
matrise3(N*(m-1)+n,N*(m-1)+n)=1/4*k*A1(2,2)*b*m*pi;
end
119
end
matrise3;
% dpi/dy med x-ledd
for m=1:M
for n=1:N
matrise4(N*(m-1)+n,N*(m-1)+n)=1/4*pi^2*m*n*D(1,2)+1/4*pi^2*m*n*D(3,3);
end
end
matrise4;
% dpi/dy med y-ledd
for m=1:M
for n=1:N
matrise5(N*(m-1)+n,N*(m-1)+n)=1/4*pi^2*n^2*a/b*D(2,2)
+1/4*pi^2*m^2*b/a*D(3,3)+1/4*k*A1(1,1)*a*b;
end
end
matrise5;
% dpi/dy med w-ledd
for m=1:M
for n=1:N
end
end
matrise6;
% dpi/dw med x-ledd
for m=1:M
for n=1:N
matrise7(N*(m-1)+n,N*(m-1)+n)=1/4*k*A1(2,2)*m*pi*b;
end
end
matrise7;
% dpi/dw med y-ledd
for m=1:M
for n=1:N
end
end
matrise8;
% dpi/dw med w-ledd
for m=1:M
for n=1:N
matrise9(N*(m-1)+n,N*(m-1)+n)=1/4*k*A1(1,1)*pi^2*n^2*a/b
+1/4*k*A1(2,2)*m^2*pi^2*b/a-1/4*pi^2*m^2*Nk*b/a;
end
end
matrise9;
% dpi/dw med w-leddpq
for m=1:M
for n=1:N
for p=1:M
for q=1:N
120
if m==p
elseif n==q
elseif rem(m-p,2)==0
elseif rem(n-q,2)==0
elseif rem(m+p,2)==0
elseif rem(n+q,2)==0
else
i=(m-1)*N+n;
j=(p-1)*N+q;
matrise10(i,j)=-8*mu*Nk*m*n*p*q/((m^2-p^2)*(n^2-q^2));
end
end
end
end
end
matrise10;
matrisetot=[matrise1 matrise2 matrise3; matrise4 matrise5
matrise6; matrise7 matrise8 matrise9+matrise10];
Nk=solve(det(matrisetot));
double(Nk)
First ply failure FSDT, Case A

% Case A
clear all;
%Materialdata
EL=49627;
ET=15430;
vLT=0.272;
GLT=4800;
G23=4800;
G13=4800;
a=500;
b=500;
k=5/6;
n=1;
m=1;
ni=1;
mi=1;
alpha=m*pi/a;
beta=n*pi/b;
alphai=mi*pi/a;
x=a/2;
y=b/2;
Wi=0.5;
% Feilkriterier
Xt=968;
Xc=915;
Yt=24;
Yc=118;
S12=65;
% initiell n
% initiell m
% initiell
% x-koordinat
% y-koordinat
% initielle geometrisk feil

%R=[-45 45 0 45 -45 0 0 -45 45 0 45 -45]*pi/180;
%h=[-4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4];
%t=8;
121
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;

%h=[-8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857
2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8];
%t=16;
R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45
0 45 -45 0 -45 45 0 45 -45]*pi/180;
h=[-12 -11.857 -11.714 -10 -9.857 -9.714 -8 -7.857 -7.714 -6 -5.857 -5.714 -4 -3.857
-3.714 -2 -1.857 -1.714 0 1.714 1.857 2 3.714 3.857 4 5.714 5.857 6 7.714 7.857 8
9.714 9.857 10 11.714 11.857 12];
t=24;
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 0 -45 45
0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-16 -15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8
-7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2
3.714 3.857 4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714
13.857 14 15.714 15.857 16];
%t=32;
%R=[-45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0
45 -45 0 -45 45 0 45 -45 0 0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45
0 -45 45 0 45 -45 0 -45 45 0 45 -45 0 -45 45 0 45 -45]*pi/180;
%h=[-24 -23.857 -23.714 -22 -21.857 -21.714 -20 -19.857 -19.714 -18 -17.857 -17.714
-16 -15.857 -15.714 -14 -13.857 -13.714 -12 -11.857 -11.714 -10 -9.857 -9.714 -8
-7.857 -7.714 -6 -5.857 -5.714 -4 -3.857 -3.714 -2 -1.857 -1.714 0 1.714 1.857 2
3.714 3.857 4 5.714 5.857 6 7.714 7.857 8 9.714 9.857 10 11.714 11.857 12 13.714
13.857 14 15.714 15.857 16 17.714 17.857 18 19.714 19.857 20 21.714 21.857 22
23.714 23.857 24];
%t=48;
%Initielverdi
A(1:3,1:3)=0;
B(1:3,1:3)=0;
D(1:3,1:3)=0;
Q(1:3,1:3)=0;
S(1:3,1:3)=0;
A1(1:2,1:2)=0;
Q1(1:2,1:2)=0;

%for First Ply Failure

sigma(1:3,1)=0;
% Spenning
Nk(1:3,1)=0;
% Pfrte laster
toyning(1:3,1)=0; % Tyning
krumning(1:3,1)=0; % Krumning
S(1,1)=1/EL;
S(2,2)=1/ET;
S(1,2)=-vLT/EL;
S(3,3)=1/GLT;
S(2,1)=S(1,2);
Q=inv(S);
Q1(1,1)=G23;
Q1(2,2)=G13;
Q1(1,2)=0;
Q1(2,1)=Q1(1,2);
Q(3,3)=2*Q(3,3);
for i=1:length(R)

122

T(1,1)=cos(R(i))^2;
T(2,2)=cos(R(i))^2;
T(1,2)=sin(R(i))^2;
T(2,1)=sin(R(i))^2;
T(3,1)=-sin(R(i))*cos(R(i));
T(1,3)=2*sin(R(i))*cos(R(i));
T(2,3)=-2*sin(R(i))*cos(R(i));
T(3,3)=cos(R(i))^2-sin(R(i))^2;
T1(1,1)=cos(R(i));
T1(1,2)=-sin(R(i));
T1(2,1)=sin(R(i));
T1(2,2)=cos(R(i));
Qk(:,3,i)=Qk(:,3,i)/2;
Qk1(:,:,i)=inv(T1)*Q1*T1;

%Skjr
A=A+Qk(:,:,i)*(h(i+1)-h(i));
B=B+.5*Qk(:,:,i)*(h(i+1)^2-h(i)^2);
D=D+1/3*Qk(:,:,i)*(h(i+1)^3-h(i)^3);
A1=A1+Qk1(:,:,i)*(h(i+1)-h(i));
%Beregner
%Beregner
%Beregner
%Beregner
A-matrisen
B-matrisen
D-matrisen
A1-matrisen
end
A;
B;
D;
A1;
C1=-D(1,1)*alpha^2-D(3,3)*beta^2-k*A1(2,2);
C2=-D(1,2)*alpha*beta-D(3,3)*alpha*beta;
C3=-k*A1(2,2)*alpha;
C4=-D(2,2)*beta^2-D(3,3)*alpha^2-k*A1(1,1);
C5=-k*A1(1,1)*beta;
Nk(1)=3672;
% Variablene i Matrisen
% Pfrte lasten i x-retning
W=-Nk(1)*alphai^2*Wi/((C5-C2*C3/C1)*(C2*C3/C1-C5)
/(C4-C2*C2/C1)-C3*C3/C1+alpha*C3+beta*C5+Nk(1)*alpha^2);
% Beregne konstanten W i w(x,y). har allerede tatt hensyn til minus-tegnet i Nk(1).
Y=(C2*C3/C1-C5)/(C4-C2*C2/C1)*W;
% Beregne konstanten i phi_y(x,y)
X=(-C2*Y-C3*W)/C1;
% Beregne konstanten i phi_x(x,y)
krumning(1)=-X*alpha*sin(m*pi*x/a)*sin(n*pi*y/b);
krumning(2)=-Y*beta*sin(m*pi*x/a)*sin(n*pi*y/b);
krumning(3)=(X*beta+Y*alpha)*cos(m*pi*x/a)*cos(n*pi*y/b);
toyning=inv(A)*(-1*Nk)
% tyning, ikke tatt hensyn til minus-tegnet

% i Nk(1), dermed m multiplisere med -1
% initielle verdi for de nye S- og Q-matrisen, ikke platen, men LAG nr i

QQ(1:3,1:3)=0;
SS(1:3,1:3)=0;
SS(1,1)=1/EL;
SS(2,2)=1/ET;
SS(1,2)=-vLT/EL;
SS(3,3)=1/GLT;
SS(2,1)=SS(1,2);
123
QQ=inv(S);
QQ(3,3)=2*QQ(3,3);
for i=37

TT(1,1)=cos(R(i))^2;
TT(2,2)=cos(R(i))^2;
TT(1,2)=sin(R(i))^2;
TT(2,1)=sin(R(i))^2;
TT(3,1)=-sin(R(i))*cos(R(i));
TT(3,2)=sin(R(i))*cos(R(i));
TT(1,3)=2*sin(R(i))*cos(R(i));
TT(2,3)=-2*sin(R(i))*cos(R(i));
TT(3,3)=cos(R(i))^2-sin(R(i))^2;
QQk=inv(TT)*QQ*TT;
QQk(:,3)=QQk(:,3)/2;

z=(h(i+1)-h(i))/2+h(i);
%hyden fra nullpunkt til midten av laget
end
QQk;
z;
sigma=QQk*toyning+z*QQk*krumning
% Feilkriterier
F11=1/(Xt*Xc);
F1=1/Xt-1/Xc;
F22=1/(Yt*Yc);
F2=1/Yt-1/Yc;
F66=1/(S12*S12);
F12star=-0.5;
F12=F12star*sqrt(F11*F22);
% Beregner Tsai-wu
faktor1=F11*(sigma(1))^2+F22*(sigma(2))^2+F66*(sigma(3))^2+2*F12*sigma(1)*sigma(2);
faktor2=F1*sigma(1)+F2*sigma(2);
tsaiwu=1/((-faktor2+sqrt((faktor2)^2+4*faktor1))/(2*faktor1))
C.2
ANSYS
In ANSYS, it is been used script and ANSYS main menu.

Simply supported with uniaxial
/prep7
!Geometri
k,1,0,0
k,2,500,0
k,3,500,500
!koordinat til knutepunkt1
124
k,4,0,500
a,1,2,3,4
!areal laget av knutepunktene 1-2-3-4
!mesh
type,1
real,1
aesize,all,25
amesh,all
!shell91 eller shell181

!referer til tykkelsen
!deklarerer elementstrrelse
!mesh areal
!boundary conditions
nsel,s,loc,y,0
d,all,uz,0
d,all,uy,0
nsel,s,loc,x,500
d,all,uz,0
nsel,s,loc,y,500
d,all,uz,0
nsel,s,loc,x,0
d,all,uz,0
d,all,ux,0
!load
nsel,s,loc,x,500
cp,1,ux,all
nsel,r,loc,y,500
f,all,fx,-1
finish
/solu
antype,static
allsel
pstres,on
solve
finish
/solu
antype,buckle
bucopt,subsp,4
solve
finish
/post1
set,list
finish
Clamped with uniaxial

/prep7
!Geometri
k,1,0,0
k,2,500,0
k,3,500,500
k,4,0,500
a,1,2,3,4
!mesh
125
type,1
real,1
aesize,all,25
amesh,all
!shell91 eller ahell181

!mesh areal
nsel,s,loc,y,0
d,all,uz,0
d,all,uy,0
d,all,rotx,0
d,all,roty,0
d,all,rotz,0
nsel,s,loc,x,500
d,all,uz,0
d,all,rotx,0
d,all,roty,0
d,all,rotz,0
nsel,s,loc,y,500
d,all,uz,0
d,all,rotx,0
d,all,roty,0
d,all,rotz,0
cp,2,uy,all
nsel,s,loc,x,0
d,all,ux,0
d,all,uz,0
d,all,rotx,0
d,all,roty,0
d,all,rotz,0
!load
nsel,s,loc,x,500
cp,1,ux,all
nsel,r,loc,y,500
f,all,fx,-1
finish
/solu
antype,static
allsel
pstres,on
solve
finish
/solu
antype,buckle
bucopt,subsp,4
solve
finish
/post1
set,list
finish
Simply supported with biaxial

/prep7
126
!Geometri
k,1,0,0
k,2,500,0
k,3,500,500
k,4,0,500
a,1,2,3,4
!mesh
type,1
real,1
aesize,all,25
amesh,all
!shell91
!mesh areal
nsel,s,loc,y,0
d,all,uz,0
d,all,uy,0
nsel,s,loc,x,500
d,all,uz,0
nsel,s,loc,y,500
d,all,uz,0
cp,2,uy,all
nsel,s,loc,x,0
d,all,uz,0
d,all,ux,0
!load
nsel,s,loc,x,500
cp,1,ux,all
nsel,r,loc,y,500
f,all,fx,-1
f,all,fy,-2
!gamma=1 => -1, gamma=0.5 => -0.5, gamma=2 => -2
finish
/solu
antype,static
allsel
pstres,on
solve
finish
/solu
antype,buckle
bucopt,subsp,4
solve
finish
/post1
set,list
finish
Simply supported with in-plane shear

/prep7
127
!Geometri
k,1,0,0
k,2,500,0
k,3,500,500
k,4,0,500
a,1,2,3,4
!mesh
type,1
real,1
aesize,all,25
amesh,all

!mesh areal
nsel,s,loc,x,0
d,all,uz,0
nsel,s,loc,y,0
d,all,uz,0
nsel,s,loc,x,500
d,all,uz,0
nsel,s,loc,y,500
d,all,uz,0
nsel,s,loc,x,0
nsel,r,loc,y,0
d,all,ux,0
d,all,uy,0
nsel,s,loc,x,0
nsel,r,loc,y,500
d,all,ux,0
!load
nsel,s,loc,x,0
nsel,r,loc,y,0
f,all,fy,-1/60
f,all,fx,-1/60
nsel,s,loc,x,0
nsel,r,loc,y,500
f,all,fx,1/60
f,all,fy,-1/60
nsel,s,loc,x,500
nsel,r,loc,y,500
f,all,fy,1/60
f,all,fx,1/60
nsel,s,loc,x,500
nsel,r,loc,y,0
f,all,fx,-1/60
f,all,fy,1/60
nsel,s,node,,3,41,2
f,all,fx,-1/60
!x=12.5 til x=487.5 , y=0
nsel,s,node,,4,40,2
f,all,fx,-2/60
nsel,s,node,,43,81,2
f,all,fy,1/60
!x=500 , y=12.5 til y=487.5
128
f,all,fy,2/60
f,all,fx,1/60
!x=487.5 til x=12.5 , y=500
f,all,fx,2/60
f,all,fy,-1/60
!x=0 , y=487.5 til y=12.5
f,all,fy,-2/60
finish
/solu
antype,static
allsel
pstres,on
solve
finish
/solu
antype,buckle
bucopt,subsp,4
solve
finish
/post1
set,list
finish
Simply supported with combined

/prep7
!Geometri
k,1,0,0
k,2,500,0
k,3,500,500
k,4,0,500
a,1,2,3,4
!mesh
type,1
real,1
aesize,all,25
amesh,all

!mesh areal
nsel,s,loc,x,0
d,all,uz,0
d,all,ux,0
nsel,s,loc,y,0
d,all,uz,0
129
nsel,s,loc,x,500
d,all,uz,0
nsel,s,loc,y,500
d,all,uz,0
nsel,s,loc,x,0
nsel,r,loc,y,0
d,all,uy,0
!shear load
nsel,s,loc,x,0
nsel,r,loc,y,0
f,all,fy,-1/60
f,all,fx,-1/60
!punktlast i de fire hjrnene
nsel,s,loc,x,0
nsel,r,loc,y,500
f,all,fx,1/60
f,all,fy,-1/60
nsel,s,loc,x,500
nsel,r,loc,y,500
f,all,fy,1/60
f,all,fx,1/60
nsel,s,loc,x,500
nsel,r,loc,y,0
f,all,fx,-1/60
f,all,fy,1/60
nsel,s,node,,3,41,2
f,all,fx,-1/60
!x=12.5 til x=487.5 , y=0
nsel,s,node,,4,40,2
f,all,fx,-2/60
f,all,fy,1/60
!x=500 , y=12.5 til y=487.5
f,all,fy,2/60
f,all,fx,1/60
!x=487.5 til x=12.5 , y=500
f,all,fx,2/60
f,all,fy,-1/60
!x=0 , y=487.5 til y=12.5
f,all,fy,-2/60
!uniax load
nsel,s,loc,x,500
cp,1,ux,all
nsel,r,loc,y,250
f,all,fx,-1
finish
130
/solu
antype,static
allsel
pstres,on
solve
finish
/solu
antype,buckle
bucopt,subsp,4
solve
finish
/post1
set,list
finish
Simply supported with imperfection

/prep7
!Geometri
k,1,0,0
k,2,500,0
k,3,500,500
k,4,0,500
a,1,2,3,4
!mesh
type,1
real,1
aesize,all,25
amesh,all

!mesh areal
!nummrg,all
nsel,s,loc,y,0
d,all,uz,0
d,all,uy,0
nsel,s,loc,x,500
d,all,uz,0
nsel,s,loc,y,500
d,all,uz,0
nsel,s,loc,x,0
d,all,uz,0
d,all,ux,0
!load
nsel,s,loc,x,500
cp,1,ux,all
nsel,r,loc,y,500
f,all,fx,-1
finish
/solu
131
antype,static
allsel
pstres,on
solve
finish
/solu
antype,buckle
bucopt,subsp,4
mxpand,4
solve
finish
/post1
set,1,1
pldisp,1
upcoord,5,on
finish
/solu
fdele,all,all
nsel,s,loc,x,500
!cp,1,ux,all
nsel,r,loc,y,500
f,all,fx,-825000
allsel
antype,static
nlgeom,on
nsubst,100
neqit,10
arclen,on
outres,all,all
allsel
solve
!slette den gamle lasten
!setter p en ny last
!statisk analyse
!ikke-liner analyse, inkluderer store deformasjonseffekter
!lastinkrement(antall steg)
!begrensning p antall iterasjon
!buelengde-alternativ, setter p last gradvis
!resultater for alle lastinkrementer
finish
/post26
nsol,2,721,u,z,uzmaks
xvar,2
!bytter om x- og y-aksen
/grid,1
/axlabel,x,Displacement
/axlabel,y,Applied load
/title,nodenr.721(x=250,y=250)
/plopts,info,off
plvar,1,
!viser last-forskyvningskurve
finish
First ply failure

/prep7
!Geometri
k,1,0,0
k,2,500,0
k,3,500,500
k,4,0,500
132
a,1,2,3,4
!mesh
type,1
real,1
aesize,all,25
amesh,all

!mesh areal
!nummrg,all
nsel,s,loc,y,0
d,all,uz,0
d,all,uy,0
nsel,s,loc,x,500
d,all,uz,0
nsel,s,loc,y,500
d,all,uz,0
nsel,s,loc,x,0
d,all,uz,0
d,all,ux,0
!load
nsel,s,loc,x,500
cp,1,ux,all
nsel,r,loc,y,500
f,all,fx,-1
finish
/solu
antype,static
allsel
pstres,on
solve
finish
/solu
antype,buckle
bucopt,subsp,4
mxpand,4
solve
finish
/post1
set,1,1
pldisp,1
upcoord,5,on
finish
/solu
fdele,all,all
nsel,s,loc,x,500
!cp,1,ux,all
nsel,r,loc,y,500
f,all,fx,-757000
allsel
antype,static
!slette den gamle lasten
!setter p en ny last
!statisk analyse
133
nlgeom,on
nsubst,100
neqit,25
arclen,on
outres,all,all
allsel
solve
finish
134

Rapport

Uploaded by

Copyright:

Available Formats

Rapport

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rapport

Uploaded by

Copyright:

Available Formats

Simplified Approaches to

Buckling of Composite Plates

Faculty of Mathematics and Natural Science

Background and Motivation . . . . . . . . . . . . . . . . . . . . .

Contents of The Thesis . . . . . . . . . . . . . . . . . . . . . . . .

2 Classical and First-Order Theories of Composite Plates

The Classical Laminated Plate Theory, CLPT . . . . . . . . . . .

The Material Law . . . . . . . . . . . . . . . . . . . . . . .

Resultant Forces and Moments . . . . . . . . . . . . . . .

Equilibrium Equations in Terms of Displacements . . . . .

The First-Order Shear Deformation Theory, FSDT . . . . . . . .

The Material Law . . . . . . . . . . . . . . . . . . . . . . .

Resultant Forces and Moments . . . . . . . . . . . . . . .

Equilibrium Equations in Terms of Displacements . . . . .

3 Analysis of Specially Orthotropic Plates Using CLPT

Buckling of Simply Supported Plates under Uniaxial Compressive

Buckling of Clamped Plates under Uniaxial Compressive Load . .

4 Analysis of Specially Orthotropic Plates Using FSDT

Buckling of Simply Supported Plates under Uniaxial Compressive

Buckling of Clamped Plates under Uniaxial Compressive Load . .

Buckling of Simply Supported Plates under Biaxial Compressive

Buckling of Simply Supported Plates under In-plane Shear Load .

Buckling of Simply Supported Plates under Combined Loads . . .

Plates with Initial Geometric Imperfection under Uniaxial Compressive Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Relationship Between Displacement and Applied Load . .

First Ply Failure . . . . . . . . . . . . . . . . . . . . . . .

5 Finite Element Model

Uniaxial Compressive Load, CLPT and FSDT . . . . . . . . . . .

Simply Supported Plates . . . . . . . . . . . . . . . . . . .

Biaxial Compressive Load, FSDT . . . . . . . . . . . . . . . . . .

In-plane Shear Load, FSDT . . . . . . . . . . . . . . . . . . . . .

Combined Load , FSDT . . . . . . . . . . . . . . . . . . . . . . .

Plates with Initial Geometric Imperfection, FSDT . . . . . . . . .

Displacement - Applied Load . . . . . . . . . . . . . . . .

First Ply Failure - Tsai-Wu . . . . . . . . . . . . . . . . . .

Suggestions for Futher Work . . . . . . . . . . . . . . . . . . . . .

A Parameter Definitions [9]

B Definitions and Equations

B.1 Matrixes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

C.1 Matlab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

Contents of The Thesis

Classical and First-Order Theories of Composite Plates

According to Reddy [1], analysis of composite plates is based on the following

The Classical Laminated Plate Theory, CLPT

The in-plane displacements are related to the normal displacements as follows

So u is displacement in x direction, v is displacement in y direction and w is

The Material Law

Definition of tensor strains [2]:

Q16 Q26 Q66

Resultant Forces and Moments

The stresses in a laminate vary layer to layer. Hence it is convenient to deal

Substitution of equation (2.2) in equations (2.3) and (2.4) gives:

A12 A22 A26

B16 B26 B66 2 2 w0

D16 D26 D66 2 2 w0

Equilibrium Equations in Terms of Displacements

The First-Order Shear Deformation Theory, FSDT

The Material Law