BUCKLING OF COMPOSITE

LAMINATES UNDER IN-PLANE

BOUNDARY CONDITIONS:
ANALYTICAL AND EXPERIMENTAL
STUDIES
A thesis submitted in fulfilment of the requirements
for the degree of Master of Technology

by

Reshma Prasad

DEPARTMENT OF CIVIL ENGINEERING

INDIAN INSTITUTE OF TECHNOLOGY KANPUR

July 2015
 ii

ABSTRACT

The composite plates have a wide range of application in the field of civil and aerospace
engineering owing to the advantage of being able to resist various kinds of loading despite
being light weight. The composite laminates may be subjected to direct compressive
load or other types of loads may induce compressive stresses on the plate. Also, in many
practical situations, the plates may be restrained along its edges when subjected to in-plane
compressive loads. Hence, thin composite laminates are susceptible to buckling failure.
The present study deals with the analytical and experimental buckling study of glass-epoxy
flat laminates subjected to in-plane uniform compressive loads. Analytical study has been
carried out in ABAQUS 6.11 considering various factors affecting the critical buckling
load such as aspect ratio and fibre orientation. Uniform compressive load is applied to
two edges while the other edges are kept unloaded but restrained. The in-plane stress
distributions and the critical buckling load are estimated. It has been noted that the
edges of the plate are subjected to non-uniform stress distributions and that the in-plane
restraints affect the stress distribution. It is observed that the crossply laminates offer
more resistance to buckling than the laminates with other fibre orientations irrespective
of the aspect ratio. For the same fibre orientation, plates with lower aspect ratio are less
susceptible to buckling. Further, experimental studies have been conducted on square
and rectangular flat composite plates fabricated in-house. An experimental set-up was
designed to simulate the boundary condition as realistic as possible with minimum error.
The plates are simply supported with all the edges restrained in in-plane and subjected to
uni-axial uniform compressive loading along the loaded edges. The experimental study has
been carried out for glass-epoxy flat laminates with three aspect ratios and three different
fibre orientations. The result obtained from the experiment is compared with the output
from ABAQUS. Good agreement of results has been observed for rectangular plates and
square laminates for uni-directional laminates.
 ACKNOWLEDGEMENT

I would like to express my heartfelt thanks to my Thesis supervisors Dr. Samit Ray
Chaudhuri and Dr. K.K. Bajpai for guiding me through the work and for the valuable
advices /instructions given that helped me to gain more insight about the work. I am also
thankful for the constant encouragement and support. I take this opportunity to thank
Kush Synthetics who sponsored the fabric used for the fabrication of laminates. A special
thanks to all the supporting staffs at the Structural Engineering Laboratory for helping me
with the experimental setup, fabrication and testing of composite laminates. I would like
to thank Mr. Atul Nigam and Radheshyam who helped me in designing the experimental
set-up by providing practical suggestions. Special thanks to Sunil Bhaiyyaand Rajkumar
Bhaiyya, staff of structural laboratory who have worked hard to make good laminates and
helped me to complete the work. I take this oppurtunity to also thank Mohit sir and
Abrar Bhaiya who helped with the data recording and testing.I would also like to thank
Vinay sir for helping with the inventory.

My completion of this project could not have been accomplished without the support of
my friends in structures, who helped me with valuable suggestions when I thought there
is no way ahead. I extend my sincere gratitude to my good friends in and out of college
who were always around to lighten my mood, encourage me, to make me feel good and to
make me believe in myself.

Finally I am expressing my deepest gratitude to my caring, loving and supportive family,

especially to my mother who helped to keep my spirit high and always boosted me and
my father for the encouragement given to me though they were miles away.

iii
Contents

Certificate i

Abstract ii

Acknowledgement iii

Contents iv

List of Figures vi

List of Tables viii

Abbreviations ix

1 Introduction and Literature Review 1

1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 Buckling of Thin Plates . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Previous Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Theoretical Expanse . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.2 Experimental Expanse . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Objectives of the study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Organization of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2 Analytical Study- Buckling Analysis 10

2.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Buckling Solution Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 Description of Specimen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4 Modelling of Composite Plates . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.5 In-plane stress distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.6 Effect of Aspect Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.7 Effect of Fibre Orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

iv
Contents v

3 Experimental Study 25
3.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2 Hand Lay-up Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.3 The Fabrication Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.3.1 The Heating Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.4 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.5 Methods of Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.6 Test Set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.6.1 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.6.2 Specimen Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.6.3 Material Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.6.4 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.7 Results and Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.7.1 Unidirectional laminates . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.7.2 Cross-ply Laminates . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.7.3 Angle plies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

4 Summary and Conclusions 49

4.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.3 Scope for Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

A The Experimental Results: Graphs 53

A.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
A.2 Unidirectional Laminates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
A.3 Cross-ply Laminates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
A.4 Angle ply Laminates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

Bibliography 91
List of Figures

1.1 A typical composite plate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1.2 Buckling of a plate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.1 The fibre orientations of the laminate . . . . . . . . . . . . . . . . . . . . . 18

2.2 The in-plane boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3 The variation of normal stresses in X-direction . . . . . . . . . . . . . . . . 19
2.4 The variation of normal stresses in Y-direction . . . . . . . . . . . . . . . . 20
2.5 The variation of shear stresses . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.6 Typical variation of stresses across the laminate thickness . . . . . . . . . . 22
2.7 Effect of aspect ratio on buckling resistance . . . . . . . . . . . . . . . . . . 22
2.8 Various fibre orientations considered in analysis . . . . . . . . . . . . . . . . 23
2.9 Variation of buckling load with fibre orientation . . . . . . . . . . . . . . . . 24

3.1 The bottom and top plates for fabrication . . . . . . . . . . . . . . . . . . . 38

3.2 Temperature control module . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.3 The heating rods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.4 The glass fibre used for fabrication of laminate . . . . . . . . . . . . . . . . 38
3.5 Procedure of fabricating laminates in steps . . . . . . . . . . . . . . . . . . 39
3.6 Placement of heating rods for curing . . . . . . . . . . . . . . . . . . . . . . 39
3.7 The curing of laminate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.8 Initial support setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.9 Boundary condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.10 Schematic diagram of experimental setup . . . . . . . . . . . . . . . . . . . 41
3.11 Details of the experimental setup . . . . . . . . . . . . . . . . . . . . . . . . 42
3.12 The final experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.13 Position of LVDTs along the plate . . . . . . . . . . . . . . . . . . . . . . . 43
3.14 The load deformation curve of mild steel . . . . . . . . . . . . . . . . . . . . 43
3.15 Load Deformation Curve for unidirectional laminate . . . . . . . . . . . . . 44
3.16 The load deformation curve of square laminate . . . . . . . . . . . . . . . . 44
3.17 The load deformation curve for rectangular laminates . . . . . . . . . . . . 45
3.18 The load-deformation curve for orthotropic plates with aspect ratio 0.75 . 45
3.19 The load versus displacement . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.20 The load versus displacement of orthotropic plates with aspect Ratio 1.5 . . 46
3.21 The load versus displacement of angle plies with aspect ratio 0.75 . . . . . . 47
3.22 The load versus displacement of angle plies with aspect ratio 1 . . . . . . . 47

vi
List of Figures vii

3.23 The load versus displacement for rectangular laminates of aspect ratio 1.5 . 48

A.1 Load Deformation Curve for U75S1 . . . . . . . . . . . . . . . . . . . . . . . 57

A.2 The Load vs Displacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
A.3 Load Deformation Curve for U01S1 . . . . . . . . . . . . . . . . . . . . . . . 59
A.4 Load Deformation Curve for U01S2 . . . . . . . . . . . . . . . . . . . . . . . 60
A.5 Load Deformation Curve for U01S3 . . . . . . . . . . . . . . . . . . . . . . . 61
A.6 Displacement Profiles for U01S2 and U01S3 . . . . . . . . . . . . . . . . . . 62
A.7 Load Deformation Curve for U15S1 . . . . . . . . . . . . . . . . . . . . . . . 63
A.8 Load Deformation Curve for U15S2 . . . . . . . . . . . . . . . . . . . . . . . 64
A.9 Load Deformation Curve for U15S3 . . . . . . . . . . . . . . . . . . . . . . . 65
A.10 Displacement Profiles for U15S1, U15S2 and U15S3 . . . . . . . . . . . . . . 66
A.11 Load Deformation Curve for O75S1 . . . . . . . . . . . . . . . . . . . . . . . 67
A.12 Load Deformation Curve for O75S3 . . . . . . . . . . . . . . . . . . . . . . . 68
A.13 Load Deformation Curve for O75S4 . . . . . . . . . . . . . . . . . . . . . . . 69
A.14 Displacement Profiles for O75S3 and O75S4 . . . . . . . . . . . . . . . . . . 70
A.15 Load Deformation Curve for O01S2 . . . . . . . . . . . . . . . . . . . . . . . 71
A.16 Load Deformation Curve for O01S3 . . . . . . . . . . . . . . . . . . . . . . . 72
A.17 Displacement Profiles for O01S2 and O01S3 . . . . . . . . . . . . . . . . . . 73
A.18 Load Deformation Curve for O15S1 . . . . . . . . . . . . . . . . . . . . . . . 74
A.19 Load Deformation Curve for O15S2 . . . . . . . . . . . . . . . . . . . . . . . 77
A.20 Load Deformation Curve for O15S3 . . . . . . . . . . . . . . . . . . . . . . . 78
A.21 Displacement Profiles for O15S1, O15S2 and O15S3 . . . . . . . . . . . . . 79
A.22 Load Deformation Curve for Q75S1 . . . . . . . . . . . . . . . . . . . . . . . 80
A.23 Load Deformation Curve for Q75S3 . . . . . . . . . . . . . . . . . . . . . . . 81
A.24 Displacement Profile for Angle plies with aspect ratio 0.75 . . . . . . . . . . 82
A.25 Load Deformation Curve for Q01S1 . . . . . . . . . . . . . . . . . . . . . . . 83
A.26 Load Deformation Curve for Q01S2 . . . . . . . . . . . . . . . . . . . . . . . 84
A.27 Load Deformation Curve for Q01S3 . . . . . . . . . . . . . . . . . . . . . . . 85
A.28 Displacement Profile for Angle plies with aspect ratio 1 . . . . . . . . . . . 86
A.29 Load Deformation Curve for Q15S1 . . . . . . . . . . . . . . . . . . . . . . . 87
A.30 Load Deformation Curve for Q15S2 . . . . . . . . . . . . . . . . . . . . . . . 88
A.31 Load Deformation Curve for Q15S3 . . . . . . . . . . . . . . . . . . . . . . . 89
A.32 Displacement Profiles for Q15S1, Q15S2 and Q15S3 . . . . . . . . . . . . . 90
List of Tables

2.1 Buckling load (in kN) obtained for various fibre orientations . . . . . . . . . 17

3.1 The details of specimen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

3.2 Buckling load values for unidirectional laminates . . . . . . . . . . . . . . . 36
3.3 Buckling load values for cross-ply laminates . . . . . . . . . . . . . . . . . . 37
3.4 Buckling load values for angle-ply laminates (0/45)s . . . . . . . . . . . . . 37
3.5 Summary of the buckling load . . . . . . . . . . . . . . . . . . . . . . . . . . 37

viii
Abbreviations

CLT Classical Lamination Theory

FEA Finite Element Analysis
FEM Finite Element Method
HSDT Higher Order Shear Deformation Theory
LVDT Linear Variable Differential Transformer

ix
Chapter 1

Introduction and Literature

Review

1.1 Overview

A plate is a three dimensional structure where the thickness is very less compared to in-
plane dimensions. A composite laminate is made up of matrix and reinforcement. Figure
1.1 shows a typical composite laminate. Usually, the glass, carbon or any fabric form the
reinforcement and the binding materials or adhesives such as epoxy resin form the matrix.
The matrix binds the fabric together. A number of laminae when stacked up forms a
composite plate. The fibre in the plate can have various orientations as desired according
to the required strength.

The material properties of the composite laminates depends on a lot of factors. A general
composite laminate can be considered as an anisotropic material as the fibre can be oriented
randomly or designed to orient in a particular direction depending on the need.

Composite plates have wide range of applications in the field of aerospace engineering (as
components of air planes, space crafts etc.), in the field of civil engineering (as parts of
bridges, buildings, pipelines etc.) and also as components of ships and rail vehicles. The
fact that composite plates are light in weight and can resist large amount of forces makes

1
Chapter 1. Introduction and Literature Review 2

these plates suitable for various applications. The other properties that make the use of
composite plates advantageous in various applications over others are their good damage
tolerance and impact resistance.

1.1.1 Buckling of Thin Plates

A composite plate is said to be thin when the width-to-thickness ratio is large compared
to its thickness (preferably >10). Depending on the purpose or application of composite
plates, various types of loading can be thrust upon it. The loads may cause the plates to
bend or buckle even before the desired purpose is served.

Buckling is a phenomenon that the thin plates are easily susceptible to. In-plane com-
pression loading or shear loading or torsional loads can cause the buckling of thin plates.
Figure 1.2 depicts the buckling of an isotropic plate under compression.

Being a thin structure, it is susceptible to failure by buckling under the action of com-
pressive forces. Various types of loads can exert an in-plane compressive load on the plate
with end restraints. So, it is important to ensure that the buckling resistance of the plate
is more than the load it is resisting. The buckling load of a plate (i.e., the maximum load
that the plate can take before it starts to buckle) depends on various parameters such as
aspect ratio, presence of cut-out, boundary conditions and the type of loading.

The buckling behaviour of laminated composite plate can be evaluated similar to the
isotropic plates. Like the isotropic plates, it also involves solving an eigen value problem
associated with a set of governing homogeneous differential equations and a prescribed set
of boundary conditions. Method of solving a buckling problem of a composite laminate
plate poses extra challenge owing to the complications inherent to laminated materials.

1.2 Previous Studies

Although the application of composites in modern structures is a recent trend in engineer-

ing, it grew abruptly due to its wide range of applications. Numerous studies have been
conducted on the buckling and post-buckling behaviour of composite plates including both
Chapter 1. Introduction and Literature Review 3

analytical and experimental studies. Following subsections present various theoretical and
experimental studies carried on buckling of composite plates till now.

1.2.1 Theoretical Expanse

A lot of theoretical studies involving the buckling of composite plates deal with the primary
issue of reducing the computational effort in calculating the critical buckling load and
ultimate strength carrying capacity of the plate. These analytical studies on buckling of
plates have been summarized in many standard text books.

In most of these studies, Classical Laminate Theory (CLT) and First Order Shear De-
formation Theory (FSDT) have been used to determine the critical buckling load. CLT
is an extension of the Kirchhoff Plate theory, which was originally proposed to model
homogeneous isotropic plates. It is one of the simplest displacement based theories for
the analysis of composite plates. In this theory, it is assumed that the cross sectional
plane normal to the centroidal plane before deformation is perpendicular even after the
deformation. Here, the transverse shear effects are neglected and thus, the plates can
be modeled as 2D structures. Although this helps in reducing the computational effort,
this theory cannot be used for thick laminates. Owing to this drawback of CLT, a lot of
theories have been developed. Better theories considering the effect of transverse shear
were proposed by Reissner and Mindlin. The Mindlin plate theory is the most commonly
used FSDT. In this theory, it is assumed that the displacement perpendicular to the plane
of the plate is constant and straight lines normal to the mid-plane remain straight, but
not necessarily perpendicular to the mid-plane. A shear correction factor is also imposed
to correctly predict the deflection patterns. Most of the investigations mentioned in the
standard text books are based on CLT and FSDT. In all these studies, it is often assumed
that the plate is allow to move freely in lateral directions and thus, the boundary condi-
tions will be expressed only in terms of transverse displacement and its derivatives. This
assumptions implies an uniform in-plane stress distribution and all stresses other than the
applied traction are negligible.

The effect of initial geometric imperfection is not included in most of these studies to
achieve computational efficiency. Yang et al. (2013) carried out a numerical study on
Chapter 1. Introduction and Literature Review 4

simply supported rectangular plates under uni-axial compression using small deflection
theory. They had intended to take into account the failure and degradation of the com-
posites, out-of-plane shear deformation and post buckling deformations along with the
geometric imperfections. Both thick and thin plates of two different lay-up were investi-
gated.

The composite lay-up was considered to be symmetric so as to reduce the computational

effort (as the B matrix: bending-stretching coupling matrix will be zero). It was found
that the ply degradation model, where the progressive damage of plies were considered,
provided more accurate results.

One of the failure modes undesired in composite laminates is the failure due to delamina-
tion or debonding. Various studies have been conducted to investigate the effect of debond-
ing on the load bearing capacity of the laminates. Senthil et al. (2013) used quasi-isotropic
CFRP (carbon fibre reinforced polymer) specimens with clamped and free boundary con-
ditions at the loaded and unloaded edges. Artificial debonds were inserted in the laminate
and the effect of these on the load bearing capacity was studied. From the numerical
studies, it was observed that as the debond size increased, the buckling load decreased.

Husam et al. (2010) developed a semi-analytical model for buckling analysis of simply sup-
ported orthotropic square plates with multi-blade stiffeners subjected to in-plane loading.
The effects of cut-outs were also studied. Various design variables considered in this opti-
mization were the cut-out size, cut-out location, the fibre orientation angles, and number
and location of stiffeners. The uniaxial, biaxial and shear in-plane loading were considered
in this study. It was found out from the study that the presence of stiffeners can increase
the buckling load up to 5 times for uni-axial loading.

Buckling analysis of rectangular laminated composite plates with arbitrary edge supports,
which were subjected to biaxial compression loading, were studied by Qing-Qing et al.
(2005). Higher Order Shear Deformation Theory (HSDT) along with the Rayleigh-Ritz
method with special displacement functions were used to evaluate the buckling behaviour of
the plate. They considered symmetric composite plates with elastic edge supports. It was
found from the study that the buckling load increases with the rigidity of edge supports.
A comparison of the FSDT and HSDT was also carried out and it was found that HSDT
Chapter 1. Introduction and Literature Review 5

provided good results. Also, the results deviated by quite a large amount when the plate
is thick. Roberts et al. (1999) did studies on orthotropic FRP plates subjected to in-plane
and out-of-plane loads with various boundary conditions in order to develop analytical
solutions for unstiffened, hat-stiffened and sandwich plates. Though good agreement was
found in the analytical and experimental results for unstiffened plates, the same could not
be obtained for the sandwich and hat-stiffened laminates.

Closed form solutions were presented by Herencia et al. (2010) for buckling of long anisotropic
plates with flexural/twist anisotropy for various boundary conditions using Rayleigh Ritz
method. The composite plates were assumed to be simply supported (SS) at short edges,
and clamped, SS or elastically restrained along the long edges. In this analytical study,
various fibre orientations were also considered.

A semi-energy based approach was presented by Chai (1991) for laminated composite
plates under uniaxial compression. The plate was assumed to be simply supported and
Von-Karman’s fundamental equation for large deflections of plates and energy in the buck-
led plate system were considered in this approach. The numerical study was carried out
on antisymmetric crossply and angle plies with 45-degree orientation.

1.2.2 Experimental Expanse

Numerous experimental studies have been reported on the buckling of composite plates.
The main challenge is to simulate the boundary conditions to a good degree of accuracy.
Various experimental investigations have been done on curved composite plates with core
and stiffeners. The current literature review is limited to the buckling analysis of a single
composite plate as the thesis deals only with the experiment on individual plates.

Chailleux et al. (1975) performed experimental study of buckling of laminated composite

columns and plates using Southwell Plot method. Aluminium-boron plates with various
orientations were studied with a fibre volume of 0.55 with unloaded edges free or hinged.
The knife edge were used with a rubber tape and thin strips of steel lamina placed in be-
tween the vertical knives. Along the loaded edges, grooves were provided to accommodate
the plate edges.
Chapter 1. Introduction and Literature Review 6

Dietrich et al. (1978) studied the plastic buckling of rectangular plates that are simply
supported on all edges and subjected to uniform compressive forces on two opposite edges.
In this experimental investigation, the side edges were seated in two columns consisting of
steel tubes. Inside the tubes there were 48 disks of 0.25 inch thick containing a groove of
0.1 inch wide and 0.1 inch deep. On the lower and upper edges of the plates, grooves were
cut out in two solid bars.

Another method to ensure continuous supports at the edges were developed in Cambridge
University and called as the finger principle. It provided simple supports at the unloaded
edges and ensured that the axial thrust was not transferred to the longitudinal supports.
Bambach and Rasmussen (2004) tested the rectangular plates supported on three sides
and the remaining edge free with the finger supports. The fingers were placed very close
to each other, and thus, they can be assumed to provide continuous support.

An experimental study on isotropic plates retrained in the in-plane directions were con-
ducted by Prajapat (2011). The set-up used in the study was fabricated in house. Thin
plates of 1.2mm with and without cut-outs were tested using this simple set-up. Grooves
were provided along all the edges to support the plate. The height of the set-up was kept
as constant while the width was changed to test various aspect ratios. This was done by
providing slots at the top and bottom supporting plates to accommodate the change of
location of one vertical supporting plate, while the position of other was kept fixed.

Another experimental study was conducted by Shrivastava and Singh (1999) on Glass Fi-
bre Reinforced Polymer (GFRP) Composite plates with unidirectional fibre orientation.
The study involved both experimental and analytical investigations. The experimental
study was carried out for combinations of various clamped and free boundary conditions.
Uniform compressive loading was applied along one direction. Various unidirectional lam-
inates with aspect ratio between 0.5 and 4 were tested. It was found from this study that
the effect of boundary condition on the buckling load increases with the increase in aspect
ratio.
 Chapter 1. Introduction and Literature Review 7

1.3 Objectives of the study

Most of the previous studies dealing with the buckling of thin plates were concerned about
the out-of-plane or flexure boundary conditions. Very little emphasis was given to in-
plane boundary conditions whether it was experimental investigation or a theoretical one.
However, from a practical point of view, when a plate is used as a structural member, it
may be subjected to different restraints along the edges. This may include restraints in
the plane of the plate as well. These in-plane restraints influence the movement of the
plate edges in the plane of the plate. As a result, pre-buckling stress patterns may become
non-uniform even when the plate is subjected to uniform loading.

The objectives of the present study are as follows:

1. To develop numerical model and perform the buckling analysis of glass-epoxy com-
posite plates subjected to uni-axial compression with in-plane restraints. The em-
phasis is to understand the effect of ply orientations and effect of aspect ratio on the
buckling load.

2. To develop numerical model and perform the buckling analysis of glass-epoxy com-
posite plates subjected to uni-axial compression with in-plane restraints. The em-
phasis is to understand the effect of ply orientations and effect of aspect ratio on the
buckling load.

3. To perform an experimental study to evaluate the buckling loads and compare with
analysis results.

1.4 Organization of Thesis

This study has been divided into four chapters.The content of each chapter is explained
in brief.

Chapter 1: This chapter gives an overview about the composite laminates and buckling of plates.
A review of literature is also presented on the buckling of composite laminates.
 Chapter 1. Introduction and Literature Review 8

Chapter 2: The analysis of the composite plates with in-plane restraints using a finite element
analysis commercial software (well-known, widely used) called ABAQUS is depicted
in Chapter 2. Laminates with various fibre orientations and aspect ratios are ana-
lyzed.

Chapter 3: In this chapter, an experimental study with three different fibre orientations and
aspect ratios is carried out on glass-epoxy composite laminates. The method of
fabrication of the laminates are explained in detail. The experimental set-up and test
procedure are also discussed in this chapter. Finally, results obtained are compared
with those from the analysis.

Chapter 4: A brief summary of the work along with the conclusions derived from the study are
presented in this chapter. The future scope of the work is also mentioned here.
Chapter 1. Introduction and Literature Review 9

Figure 1.1: A typical composite plate

Figure 1.2: Buckling of a plate

Chapter 2

Analytical Study- Buckling

Analysis

2.1 Overview

In-plane boundary conditions may play an important role in determining the critical buck-
ling load of a composite plate depending upon the loads applied on the laminate. In this
chapter, ABAQUS 6.11 is used to evaluate the buckling load of composite laminates, which
are simply supported and restrained in in-plane directions. The buckling load is obtained
by carrying out eigen value analysis. The buckling modes and the in-plane stress distri-
butions are studied to see how the in-plane restraints affect the buckling load. Also, the
effect of fibre orientations and the change in aspect ratio is investigated.

2.2 Buckling Solution Approach

The governing equation for the buckling problem of composite laminates resembles that
of the isotropic plates. For a symmetrically laminated plate subjected to in-plane loads,

10
Chapter 2. Analytical Study 11

the governing equation is given by,

∂4w ∂4w ∂4w ∂4w ∂4w

D11 + 4D 16 + 2(D 12 + 2D 66 ) + 4D 26 + D 22 =
∂x4 ∂x3 ∂y ∂x2 ∂y 2 ∂x∂y 3 ∂y 4
(2.1)
∂2w ∂2w ∂2w
Nx 2 + 2Nxy + Ny
∂x ∂x∂y ∂y

where the D terms represents the values from the constitutive matrix, w the out of plane
displacement and the N values represent the stresses in the direction noted by the sub-
script. For the unidirectional loading, the expression can be simplified as Ny = 0.

For solving the aforementioned governing equation, Rayleigh Ritz method or any other
methods can be adopted. Here, an appropriate displacement function is assumed accord-
ing to the boundary condition and the equation is solved as an eigen value problem. But
in the classical approach, the stress at any point is considered equal to the applied in-plane
stresses, which may not always be the case. When the plates are restrained at its edges, it
will naturally influence the in-plane stress distribution and the buckling resistance of the
plate. Owing to the complexity of the problem, the laminates were modelled in ABAQUS
6.11 to reduce the calculation effort.

2.3 Description of Specimen

In this study, glass-epoxy composite plates of three different aspect ratios are modelled
and their behaviour under uniform compressive load is noted. The plates are modelled for
sizes of 150 x 200, 200 x 200 and 300 x 200 ie, of aspect ratios 0.75, 1 and 1.5 respectively.
The sizes denote the unsupported length of the specimens. The laminate is assumed to
be made up of 4 layers each layer of 1 mm each approximately. For each of the sizes
depicted here, five varied fibre orientation were chosen. The unidirectional laminate where
all the plies are arranged in the same direction viz., the main fibres are oriented along the
direction of load (0)2s . The crossply symmetric laminate where the fibres are oriented in
both normal and perpendicular to loading direction (0/90)s . These two ply orientations
Chapter 2. Analytical Study 12

are shown in Figure 2.1. The other fibre orientations are obtained by replacing the 90o
plies with other angle plies such as 30o , 45o and 60o .

The boundary condition considered in the study is that laminate is simply supported on
all edges for flexure. To verify the importance of in-plane restraints, the displacements
are restrained in-plane. The normal displacement u along the loaded edges and normal
displacement v along unloaded edges are restrained in this particular case. The schematic
diagram of the boundary condition sought is depicted in Figure 2.2

2.4 Modelling of Composite Plates

There are various methods to model a composite laminate. It can be modelled microscop-
ically, indicating the matrix and the fibre will be modelled separately and their individual
properties are used in the analysis. It can also be modelled by the method of smear mod-
elling where the composite material is modelled as an equivalent homogeneous material
with either stacked or single configuration. The method adopted to model composite lam-
inates in this study is layered modelling. In this method, each layer is a composite (fabric
+matrix together) and it is laid on as plies. This implies that each layer or ply will have
homogeneous property which is a mix of the properties of fibre and matrix.

The material properties of the composite laminates were obtained using the following
equations:

The longitudinal modulus is given by,

E11 = Ef vf + Em vm (2.2)

where E and v represent Young’s modulus and volume fraction, respectively; f and m
represent fibre and matrix, respectively.

Major Poisson’s Ratio:

Chapter 2. Analytical Study 13

ν12 = νf vf + νm vm (2.3)

The transverse modulus:

Ef Em
E22 = (2.4)
Ef v m + Em v f

The shear modulus:

Gf Gm
G12 = G12 = (2.5)
Gf v m + Gm v f

The values obtained from these equations are as follows: E in longitudinal direction =
36.6 × 103 N/mm2 ; E in transverse direction = 6.12 × 103 N/mm2 ; Poisson’s ratio, ν =
0.285 and shear modulus, G = 1.981 × 103 N/mm2 .

ABAQUS 6.11 provides different types of element types for modelling two dimensional
structures. For shell structure whether thick or thin three types of elements can be used:
S8R5, S4 and S4R. The most common type of element used for modelling a composite
laminate is S8R5. It is an 8 noded doubly curved thin shell element. It has 5 degrees
of freedom and reduced integration which helps in preventing the ill-conditioning in the
analysis. This element also increases the computing efficiency without loss of accuracy in
results by using 5 degrees of freedom instead of 6 in the integration.

2.5 In-plane stress distributions

When the plate is simply supported and not restrained for in-plane displacement, there
will be uniform stresses in the plate if,unidirectional uniform load is applied on the plate.
But, when the edges of the laminate are restrained against in-plane displacement, some
resistance is exerted on the plate resulting in non-uniform stress distribution. Thus in-
plane stresses are developed in the laminate unlike the classical boundary condition.
Chapter 2. Analytical Study 14

Figures from 2.3 - 2.5 shows the normal and shear stress distribution for unidirectional
compressive loading of the unidirectional laminate for the bottom layer of the plies. The
variation of the normal stresses in the X direction (σxx ) for aspect ratios 0.75 and 1.5
are depicted in Figure 2.3. It can be seen that the stress distribution in the plate is
not uniform. Even though the load is applied only along one edge of the laminate, it is
observed that the stress is present along the unloading edge also. Figure 2.4 plots the
variation of normal stresses along the Y direction (σyy ) for the aspect ratios 0.75 and 1.5
respectively. This is due to the fact that the unloaded edge is restrained from moving in
the in-plane direction and this restriction is applying in-plane stress along the unloaded
edges also due to Poisson’s effect.

It can also be seen from Figure 2.5 that shear stresses are present in the laminate at all
the four corners. This implies that while two corners are in tension, the other two are in
compression. But this is observed only at the corners. Towards the centre of the laminate,
the effect of shear stress seems to be negligible. This may be due to the fact that the
loading becomes uniform towards the centre of the plate.

A typical plot of the variation of stresses across the thickness is shown in Figure 2.6. It
can be observed that there is abrupt change in stress values at the ply boundaries due to
the change in fibre orientation similar to the behaviour of other composite structures .This
particular plot is of plies with angles at 60 degrees (0/60)s .As the laminate as 4 layers of
1mm each, the can be noted that the stress value changes at 1mm, 2mm and 3mm.

2.6 Effect of Aspect Ratio

The effect of aspect ratio was studied for the various fibre orientations of the lami-
nate.Three aspect ratios of the composite laminates were modelled. The plates with sizes
200 x 150, 200 x 200, 200 x 300 i.e, with aspect ratios 0.75, 1 and 1.5 respectively. Fig-
ure 2.7 depicts the variation of buckling load for various aspect ratios for a particular
fibre orientation. For comparison all the fibre orientations are plotted in the same graph.
The orthotropic laminates shows more resistance to buckling as can be observed from the
plot. The unidirectional and angle ply at 45 degrees laminates have similar resistance to
Chapter 2. Analytical Study 15

buckling and the plates with fibres oriented at 30 degrees has the least buckling resistance.
Angle plies at 60 degrees has higher resistance to buckling when the aspect ratio is 0.75
but reduces as the aspect ratio increases. It is interesting to note that at aspect ratio
of 1.5, almost all the laminates have similar amount of resistance against buckling. The
buckling resistance of orthotropic laminate is about 40% more than that of unidirectional
laminate at aspect ratio of 0.75 but the differnce comes down to 9% at the aspect ratio of
1.5.

2.7 Effect of Fibre Orientation

The composite laminates are analysed for understanding the effect of fibre orientations
in the in-plane restrained boundary conditions. Five fibre orientations were tested: uni-
directional (0)2s ,orthotropic (0/90)s and angle plies (0/30)s ,(0/45)s ,(0/60)s .

Figure 2.9 shows the effect of fibre orientations on the buckling load in graphical form. The
Y axis shows the buckling load while the X axis depict the angle of the fibre considered.
For the plates with aspect ratio 0.75, the orthotropic plates showed more resistance to
buckling under in-plane compressive loading. The unidirectional and The buckling load
was 38.5kN . The buckling resistance of the uni-directional plate, angle plies at 300 and
450 are comparable in the range of 25-28 kN. . It is interesting to note that the angle plies
at 30 has less resistance compared to the other two laminates. It can be seen that for
all the orientations, the aspect ratio of 1.5 has the least resistance to buckling. Another
observation from the plot is that for the square laminates, although the buckling resistance
values are less for unidirectional laminates, the buckling load increases steadily as the angle
of the orientation increases.

The critical buckling load values obtained from the analysis is tabulated in a tabular form
in Table 2.1. The critical load is expressed in terms of kN.The values obtained for the
fibre orientations unidirectional, cross-ply and angle plies at 45 degrees were then compared
with the results obtained from the experimental study. From the table, it can be observed
that all the cross-ply laminate buckling loads are above the mean critical buckling load
value for the same aspect ratio. For the laminate with aspect ratio 0.75, the angle plies
Chapter 2. Analytical Study 16

at 60o also have buckling load grater than the mean value. Similarly, the unidirectional
laminate of aspect ratio 1.5 has more buckling resistance than the mean value. The mean
buckling load increases by 41.30% when the aspect ratio changes from 1.5 to 1 and it
increases to 66.12% when the ratio changes to 0.75 from 1.5. But, the deviation in mean
buckling load between the laminates of aspect ratios 0.75 and 1 is 41.34%. This implies
that the buckling load increases by a percentage of approximately 41.3% when the aspect
ratio changes from the 1.5 to 1 and then to 0.75.

2.8 Summary

The general trend of reduction in buckling load with increase in the aspect ratio is observed
to follow here. Also, it was noted that when a laminate is subjected to uniaxial compression
and in-plane restraints are provided, it simulates the effect of laminate under bi-axial
compression. In that case, the force applied along the unloaded edges will be less though
not negligible. As a result, the buckling load gets affected.
Chapter 2. Analytical Study 17

Table 2.1: Buckling load (in kN) obtained for various fibre orientations

Aspect Uni- 300 450 600 Cross- Mean Deviation

Ratio directional plies plies plies ply Load Deviation
0.75 27.49 25.82 27.87 33.50 38.58 30.05 +66.12%
1 18.99 18.15 18.82 20.86 29.52 21.26 +41.30%
1.5 18.15 17.32 17.12 17.45 19.81 18.09 -
Chapter 2. Analytical Study 18

Figure 2.1: The fibre orientations of the laminate

Figure 2.2: The in-plane boundary conditions

Chapter 2. Analytical Study 19

(a) Aspect ratio 0.75

(b) Aspect ratio 1.5

Figure 2.3: The variation of normal stresses in X-direction

Chapter 2. Analytical Study 20

(a) Aspect ratio 0.75

(b) Aspect ratio 1.5

Figure 2.4: The variation of normal stresses in Y-direction

Chapter 2. Analytical Study 21

(a) Aspect ratio 0.75

(b) Aspect ratio 1.5

Figure 2.5: The variation of shear stresses

Chapter 2. Analytical Study 22

Figure 2.6: Typical variation of stresses across the laminate thickness

Figure 2.7: Effect of aspect ratio on buckling resistance

Chapter 2. Analytical Study 23

(a) Unidirectional laminate (b) Angle plies at 300

(c) Angle plies at 450 (d) Angle plies at 600

(e) Orthotropic laminate

Figure 2.8: Various fibre orientations considered in analysis

Chapter 2. Analytical Study 24

Figure 2.9: Variation of buckling load with fibre orientation

Chapter 3

Experimental Study

3.1 Overview

In this chapter, the material properties of the constituent materials used for fabrication
(E-glass and epoxy), the specifications of the composite plates fabricated and the process
of fabrication is explained in detail. The composite plates are fabricated by hand lay-up
method which is one of the oldest and simplest methods to cast the plates manually. The
test set-up and the experimental investigation for in-plane restrained boundary condition
are explained in detail in this chapter . The composite plates made up of glass epoxy are
tested for buckling properties for various aspect ratios and fibre orientations. The buckling
load obtained from the ABAQUS is then compared with the experimental results.

3.2 Hand Lay-up Method

There are various methods available to fabricate laminates. Some of the most popular and
commonly practised are hand lay-up method, spray-up method, bag moulding process
etc. The hand lay-up method is also known as contact lay-up method. It is an open
mould process as the layers are placed one above the other and this is best suitable for
thermosetting resins. A typical lay-up process involves three steps: mould preparation,
gel coating and hand lay-up.

25
Chapter 3. Experimental Study 26

The preparation of mould is an important part of the process. The mould should be
smooth. This can be ensured by machine smoothing or polishing the surface. To facilitate
the easy release of laminate from the mould, releasing agent is used. Wax, Polyvinyl
Alcohol (PVA), flurocarbons etc, can be used as releasing agents. When a smooth surface
is preferred, a special resin layer which is a polyester, mineral-filled, pigmented and non-
reinforced layer is applied as the first step. This becomes the outer surface of the laminate.
This step is called gel coating.

After the mould preparation and gel coating is done, the laminate is prepared. The resin
and hardener is mixed in a special ratio as per the requirement and it is spread over glass
fabric which is laid over the gel coat. It is very important to make sure that air is not
entrapped. This is ensured by compacting the layer against the mould. This process is
repeated till all the layers are laid out as per the design.

3.3 The Fabrication Setup

The method adopted to fabricate the laminate plate is the hand lay-up method. In this
process, the laminae are laid one above the other in the desired orientation, coated with
epoxy. Later, the layers are cured by heating and plate is formed. The instrumentation
involves three main components, i.e., heating component, temperature regulator and plate
fabrication set-up.

Figure 3.1 shows the set-up fabricated to cast the composite plates in the laboratory. The
mould plates are made up of 1 inch thick mild steel. On the sides of both the plates,
there is provision provided to insert the heating rods at equal spacing. A slot for inserting
thermocouple is also provided at the middle of the plate to control the temperature. The
inner surface of both the plates are machined smooth. On the other surface, asbestos sheet
is attached with the help of epoxy resin to retain the heat.
Chapter 3. Experimental Study 27

3.3.1 The Heating Mechanism

The composite laminate, sandwiched between the top and bottom base plate, are cured
by heating it under controlled temperature. The base plates are heated with the help of
four heating rods of 10 mm diameter each made up of nickel. These can be connected to
the temperature control module. The module helps to set a particular temperature for
curing and temperature of the plate can be checked with the help of the thermocouple.
For the fabrication purpose, the temperature was set to 600 C (as prescribed by epoxy
manufacturer). Figures 3.2 and 3.3 shows the temperature control module and heating
rods respectively.

Commonly used thermocouple for recording the temperature is Chromal/Alumel thermo-

couple. It can measure the temperature to the upper range varying between 200o C to
1260o C. Chromel and Alumel are alloys. Chromel is made up of 90% Nickel and 10%
Chromium while Alumel consists of 95% Nickel, 2% Manganese, 2% Aluminium and 1%
Silicon.

3.4 Materials

The composite laminate used in this study is made up of E-glass and epoxy. Among the
various glass fibres used for the manufacturing, E-glass is the most common one. The
thermosetting resin used is epoxy(LY556) and hardener, Huntsman Aradur HY951 1N.

A variety of fibres in various forms are available in the commercial market. A particular
fabric is selected based on four parameters: weight, strand direction, weave and fibre
composition. Weight is the amount of fabric present in grams per cubic centimetre. This
parameter is directly proportional to the strength. More the weight of the fabric, more
the strength. Based on the strand direction, the fabric can be generally divided into
three: uni-directional, bi-directional and mat. The unidirectional fabrics have it’s 80-90%
of fabric aligned in one direction and the rest perpendicular to that so that, the fibres
can be hold in place. The bi-directional fabric will have fibres equally distributed in both
directions. It will have same strength in both directions. For mat fabric, the fibres are
Chapter 3. Experimental Study 28

oriented randomly. It will have fabric in every direction and hence it will have strength in
all the directions.

The various fibres commonly used in the market are E-glass, S-glass, carbon fibres and
Aramid. E-glass is the most commonly and economically used glass fabric. S-glass is
chemically modified version of S-glass which provides more strength. Carbon fibre fabric
is characterised by black colour and Aramid fibres (also known as Kevlar) has high tensile
strength, but is poor in compression.

The weave of the fabric is also an important property while selecting the fabric as it is
related to the flexibility of the fabric.(For example, twill weaved fabric is less flexible than
the modified twill. The flexibility allows the fabric to conform to compound curves and is
thus modified twill is suitable for curves.)

For the study, E-glass fibres were chosen. A unidirectional glass fabric of 1250 kg/m3
was sponsored by Kush Synthetics, a company based in Surat, for the research purposes.
Figure 3.4 shows the material. The epoxy required for the fabrication of laminate is made
by mixing the epoxy resin LY556 and hardener HY951 1N in the ratio of 10:1. LY556
resin is a bi-functional epoxy resin, Di-Glycidyl Ether of Bisphenol-A (DGEBA). HY951
is an aliphatic primary amine called Tri-Ethylene Tetra Amine (TETA).

3.5 Methods of Fabrication

There are various methods to fabricate flat plate laminates of thermosetting resin matrix.
They can be broadly classified into fabrication using prepregs and by wet processing. In
wet processing, the laminate is made when the resin is in liquid state and later it is cured
with the help of heating. Hand layup method is one of the wetprocessing methods and as
stated earlier, it is one of the oldest and simplest manual method.

The procedure adopted to fabricate the laminate is based on thickness control. The spacer
bars were used to maintain the required gap between the two base plates. The intention
was to maintain the thickness of the plate around 4 mm. First of all, the base plates and
the spacers, made up of aluminium, are coated with three to four coats of the releasing
Chapter 3. Experimental Study 29

agent. These help in removing the composite plates later. After the coat is dried, the first
layer of epoxy is applied. The epoxy resin is made by mixing the epoxy and hardener in
the ratio of 10:1. The unidirectional fabrics are cut into sizes a little bigger than required
so as to account for the imperfections in cutting the fabric, the workmanship etc.

The fabric is laid on the epoxy resin spread on the plate and another layer of epoxy resin is
spread on it. Special care was taken to dampen the fabric with resin. It is very important
that the fabric is wet throughout and uniformly. With the help of the brush, pressure
is applied on the fabric so that, resin fills the space between the fabric. The process of
fabrication is depicted in various steps in figure 3.5.

The process is repeated until all the four layers are laid properly. The top steel plate is
placed above and is tightened by 12 bolts spaced equally along the edges. The opposite and
diagonal bolts were tightened simultaneously for uniformity. After the bolts are tightened,
the set-up was left for 5 hours so that the epoxy can undergo chemical reaction and cure
on its own, as it is a thermosetting resin. Then the heating rods and thermocouple are
inserted into the holes provided and the temperature is set to 600 C in the temperature
control module for post-curing as per the recommended curing cycle. Figure 3.6 shows
the set-up for heating the specimen. It is then heated for a minimum of 6 hours after
which it is allowed to cure on its own. The laminates are cured for a total of 20 hours
approximately.

The set-up for the curing of the laminates is shown in Figure 3.7. After the plates are
cooled, the set-up is opened and the laminates are taken out and the releasing agent,
which is now a thin film, is removed from the plate. The plate is washed and is cut into
the required size with the help of the diamond cutter and the edges are smoothed using
sandpaper.

3.6 Test Set-up

The main aim during the design of test set-up was to develop a simple set-up which will
represent the required boundary conditions adequately with minimum margin of error. As
the objective of the study was to understand the behaviour of laminates under in-plane
Chapter 3. Experimental Study 30

restraints, the set-up fabricated by Prajapat (2011) was referred. The unloading edges of
the plates were supported with grooved supporting plates. The plate will be arrested in
the grooves and will not be able to move in-plane. But the mechanism couldn’t simulate
the rotation of the unsupported edges.

An improvisation of the set-up is sought, through which, the rotation can also be simulated.
Initially, the idea was to provide the support on unloaded edges with the help of ball
bearings. The idea was inspired from the principle of finger supports. It was sought to
develop a supporting system with ball bearing attached at fixed intervals with grooves.
The plates can be inserted in to the grooves so that the plate is supported at edges.
Figure 3.8 shows the supporting system. But later it was concluded that, between the
two supports, there would be a chance of local buckling to take place and that bearings,
if provided, at such distance apart will not serve the purpose of finger supports. Also, the
required boundary conditions may not be achieved with much precision. Thus, another
idea was sought to provide continuous support along the edges.

3.6.1 Design

The boundary condition indicates simply support on all edges in flexure. But the in-
plane displacements are restrained. The tangential displacement v along the loaded edges
and normal displacement v along unloaded edges are restrained. Figure 3.9 shows the
schematic boundary condition. A setup was designed taking into account, the in-plane
edge restraints required for the study. The schematic diagram of the final setup is shown
in figure 3.10. The cylindrical rod of 70 mm diameter serves as the supporting sides on the
unloaded edges. They are provided with a groove along the length and are installed with
ball bearings on both ends. These cylindrical rods are attached to the bottom and top
supporting plates of 35 mm thick with the help of four circular rods of 20 mm diameter.
The circular rods are connected to the plate with the help of bolts. The bottom plate also
have groove at the centre along the length to accommodate the edge of the plate. The
schematic diagram and cross section of the base plate is as shown in figure 3.11. Thus, the
composite plate is restrained along at the edges but is able to rotate about the support.
Chapter 3. Experimental Study 31

The composite plate, when placed in between the grooves of the cylindrical rod, is able to
rotate about its axis. Hence, the plate is able to rotate when load is applied from top.

The top plate consists of two parts. One part connects the cylindrical rod to the plate
and the other, applies load to the plate. It is designed in such a manner that, the load
is applied on the plate alone and is not transferred to the side supporting rods. The load
bearing part of the plate has grooves on the bottom, so that the composite plate is seated
and the out of plane movement of the edges are restrained along the edges. The final
experimental setup used for the study is shown in figure 3.12.

3.6.2 Specimen Details

The composite plates that are used for the experimental investigation are made of glass-
epoxy. In other words, the unidirectional glass fabric forms the fibre and epoxy resin is the
matrix used to bind the various layers of fabric. The effects of various fibre orientations
and aspect ratio on the buckling load of the composite plates are investigated through
the experiment. For that, three aspect ratios and three fibre orientations were considered.
Table 3.1 shows the details of the specimens tested. For checking the repeatability or
reproducibility of the results, three samples were tested for each specimen.

3.6.3 Material Properties

The fabric used for the experimental investigation is sponsored by Kush Synthetics. The
net area density of the unidirectional fabric was 1250 kg/m3. A volume fraction of 0.5
was decided upon as the resin would not be able to wet the fabric properly, if the volume
fraction of the fabric is more than 0.6. Then, there could be chances of delamination.
Other values are considered to be the same as in the previous chapter.
Chapter 3. Experimental Study 32

3.6.4 Instrumentation

The displacements of the specimen at three different locations were measured with help of
LVDTs. Figure 3.13 depicts the arrangement of LVDTs. Load being applied on the plate
was measured with the help of donut load cell of 500 kN capacity.

The hydraulic pump of ENERPAC with a maximum pressure of 10,000 psi was used to
apply load to the specimen. The load was applied gradually to the load cell which in turn
transferred it to the top loading plate and then to the specimen

3.7 Results and Discussions

The experimental set-up should satisfactorily simulate the required boundary conditions.
In order to check this, mild steel plates were tested and the results were compared with
the analytical result. This exercise also helped in understanding the repeatability of the
test method. For an isotropic plate of size 150 mm x 200 mm x 3.2 mm made up of mild
steel (with mechanical properties E = 2.1X105 N/mm2 and ν = 0.3), the critical buckling
load was found to be 106 kN from the analysis. Comparable results were obtained from
the experiment. Figure 3.14 shows the load deformation curve. The test was done for
three similar specimens and the results obtained were in the range of 10% error. The
plates were tested to obtain the buckling load by applying uni-axial loading from the top
using a hydraulic jack. As the jack was operated manually, there were instances of small
discrepancy in the load application. It can be concluded from this small experiment that
the set-up simulates the boundary condition in a satisfactory manner and the laminates
can be tested in the set-up in a similar manner.

The experimental study was carried out on glass-epoxy laminates of three different aspect
ratios with three fibre orientations and three specimens each. The experimental results
obtained are discussed in the following section. It may be noted that due to the hydraulic
loading mechanism, there were undulations in the load deformation curves. Effort has been
given to remove these undulations by drawing smooth curves. Results are summarized
on the basis of the fibre orientation considered and the output from the experiment is
Chapter 3. Experimental Study 33

compared with the analytical results from ABAQUS. It should be noted that the analysis
was carried out for laminate with thickness 4 mm, but in the experiment, the thickness
of the laminate varies between 3.35 mm - 3.60 mm. Another factor, that needs to be
considered is the modulus of elasticity in longitudinal and transverse direction. While
calculating the moduli, it was presumed that the volume fraction of the laminate will be
0.5 and all the fibres in the lamina will be oriented in the same direction. But, this is not
possible practically as, in unidirectional fabric, 10-15 % of the fibre will be arranged in
perpendicular direction so as to support the fibres in the main direction.

The general observations made from the experiment are:

• From the visual inspection, it was found out that the plates seemed to regain its
actual shape after the loading was removed. Further investigation is required to
ensure whether it has undergone any permanent damage or the plate is still in the
elastic range.

• At the edges, the delamination has started to occur.

• In some cases, the laminates had come out of grooves after buckling.

3.7.1 Unidirectional laminates

In general, the experimental data and analytical result are comparable for 2 out of 3
specimens tested. Load-deformation curve for only a single specimen is depicted in this
Chapter. The rest of the figures are plotted in Appendix A. Figure 3.15 shows the load
deformation curve for the unidirectional laminates with aspect ratio 0.75 tested experimen-
tally. The plate buckled at a load of 27.5 kN. It is depicted by the point where the slope
of the graph changes. From the analytical model from ABAQUS, the buckling load was
found to be 27.49kN. The value from the experimental result is very close to the analytical
with a marginal deviation of 1.8%. Other specimens also showed similar behaviour.

The square laminate of unsupported length of 200 mm is tested with fibres oriented in
a direction along the loading. The critical buckling load for this laminate was found to
be less than that of the laminate with aspect ratio 0.75 as expected. Figure 3.16 shows
Chapter 3. Experimental Study 34

the load deformation curve obtained. The buckling load was found at the point where the
slope of the load deformation curve changes (16.18 kN). Unidirectional square laminates
showed a deviation of 19.8% from the analytical value. The buckling load estimated for
rectangular laminates with aspect ratio 1.5 was almost equal to the square laminates. The
value obtained from figure 3.17 is 18 kN. The estimated buckling load from the analytical
model is 18.15 kN. In general, it can be said that for uni-directional laminates, the results
from experiment matches with the ABAQUS output quite well. Table 3.2 summarises the
findings for uni-directional laminates.

3.7.2 Cross-ply Laminates

A typical load deformation curve obtained from the testing of crossply plate of aspect
ratio 0.75 is given in figure 3.18. As explained earlier, the change in the slope of the load
deformation curve marks the buckling load for a particular laminate. From the analytical
model, the critical buckling load was found to be 38.57kN which shows a deviation of about
26% from the experimental results. The square laminates with crossply fibre orientation
had critical buckling load lower than the estimated value. The analysis predicted a value of
30kN whereas obtained value is 47.50% lesser. Figure 3.19 shows the behaviour of crossply
laminate with aspect ratio 1.

Figure 3.20 shows the load deformation cuve for crossply laminates with aspect ratio of 1.5.
In this case also, two out of three specimens provided comparable results in experiment.
The average deviation of the crossply laminates from the analytical study was found to be
around 16%. From table 3.3 it can be seen clearly that as the aspect ratio increases the
critical buckling load also decreases.

In general, it can be said that the crossply laminates showed greater amount of variation
from the analytical results than compared to unidirectional laminates. Square laminates
deviated more than the rectangular laminates.
Chapter 3. Experimental Study 35

3.7.3 Angle plies

The experiment was conducted for the symmetric angle plies oriented at 45 degrees. From
the analytical model for plates with aspect ratio 0.75, the buckling load was found to
be 27.87 kN. From figure 3.21, the buckling load was obtained to be 26 kN. The angle
ply laminates gave comparable results for the specimens. The average deviation from the
analytical study was less than 10%.

For square laminates, it was found that the buckling load obtained from the experiment
is higher than result from ABAQUS. These may be due to the fact that the fibres in
the direction perpendicular to the main fibre direction was not included in the analysis.
Another fact can be the angle of the fibre, which may not be exactly 45o as assumed in
the model. Figure 3.22 shows the load deformation curve for the square laminates. The
average deviation was found to be around 8%.

The longer rectangular laminate with aspect ratio 1.5 was found to buckle at a lower
load than estimated from the analysis. Out of the three specimens, two of them buckled
at a lower load while the other exceeded the estimated critical load from the analysis.
Figure 3.23 shows the load deformation curve of the latter one. Table 3.4 tabulates the
experimental results.

Table 3.5 summarizes the average buckling load obtained from all the specimens, the
estimated buckling load from the analysis, and the deviation. It can be noted that the
unidirectional laminates, (0)2s , gave more or less same value in both the studies. But,
the cross-ply laminates shows a large amount of variation. It can be observed that the
experimental value is less than the analytical value for the cross-ply laminates. This may
be because of the reason that, in analytical study, it was assumed that the fibres are strong
enough to carry the load being applied on them and the laminate fails as a whole (ie, there
will not be any local failure in the laminate). From the experimental results, it can be
said that the laminate may have failed due to some other reason and then it buckled. The
angled plied laminates also gave comparable results for both analytical and experimental
studies.
Chapter 3. Experimental Study 36

3.8 Summary

In this chapter, method fabrication of composite plates, the test setup, instrumentation
and test procedure are described. Further, test results are presented and the buckling
loads are evaluated for various ply orientations and aspect ratios. Further, these results
are compared with the numerically obtained values. It is found that the experimentally
obtained results may deviate significantly from the numerical values based on fiber orien-
tation and aspect ratios. Perhaps the reason for such deviation is different mode of local
failure, which needs further investigations.

Table 3.1: The details of specimen

Sl.No. Breadth Height Average Orientation ID

(mm) (mm) Thickness (mm)
1 150 200 3.58 (0)2s U75S
2 150 200 3.54 (0/90)s O75S
3 150 200 3.36 (0/45)s Q75S
4 200 200 3.43 (0)2s U01S
5 200 200 3.47 (0/90)s O01S
6 200 200 3.49 (0/45)s Q01S
7 300 200 3.41 (0)2s U15S
8 300 200 3.51 (0/90)s O15S
9 300 200 3.59 (0/45)s Q15S

Table 3.2: Buckling load values for unidirectional laminates

Sl.No Specimen ID Aspect Ratio Buckling Load Deviation Remarks

Experimental Analytical
1 U75S1 0.75 27.5 27.47 0.00
2 U75S2 0.75 28 27.47 1.80 %
3 U75S3 0.75 - 27.47 - No data
4 U01S1 1.00 37.5 18.99 +97.47%
5 U01S2 1.00 14.5 18.99 23.64%
6 U01S3 1.00 16.25 18.99 14.7%
7 U15S1 1.50 18 18.15 0.83%
8 U15S2 1.50 10.5 18.15 42.15%
9 U15S3 1.50 19.5 18.15 7.43%
Chapter 3. Experimental Study 37

Table 3.3: Buckling load values for cross-ply laminates

Sl.No Specimen ID Aspect Ratio Buckling Load Deviation Remarks

Experimental Analytical
1 O75S1 0.75 27 35.58 24.11%
2 O75S2 0.75 - 35.58 - No data
3 O75S3 0.75 30 35.58 15.68%
4 O75S4 0.75 22 35.58 38.16%
5 O01S1 1.00 - 29.52 - No data
6 O01S2 1.00 16 29.52 45.80%
7 O01S3 1.00 15 29.52 48.96%
8 O15S1 1.50 18.5 19.81 6.61%
9 O15S2 1.50 9.5 19.81 52.04%
10 O15S3 1.50 14.75 19.81 25.63%

Table 3.4: Buckling load values for angle-ply laminates (0/45)s

Sl.No Specimen ID Aspect Ratio Buckling Load Deviation Remarks

Experimental Analytical
1 Q75S1 0.75 26 27.87 6.70%
2 Q75S2 0.75 - 27.87 - No data
3 Q75S3 0.75 31 27.87 +11.23%
4 Q01S1 1.00 23 18.82 +22.2%
5 Q01S2 1.00 22 18.82 +16.89%
6 Q01S3 1.00 16 18.82 14.98%
7 Q15S1 1.50 28 17.12 +63.55%
8 Q15S2 1.50 13 17.12 24.06%
9 Q15S3 1.50 16.52 17.12 3.5%

Table 3.5: Summary of the buckling load

Sl.No Orientation Aspect Ratio Buckling Load Deviation

Experimental Analytical
1 (0)2s 0.75 27.75 27.47 10.02%
2 (90)s 0.75 26.33 35.58 25.98%
3 (45)s 0.75 28.50 27.87 2.20%
4 (0)2s 1.00 15.38 18.99 19.01%
5 (90)s 1.00 15.50 29.52 47.50%
6 (45)s 1.00 20.33 18.82 8.04%
7 (0)2s 1.50 18.75 18.15 3.30%
8 (90)s 1.50 16.63 19.81 16.07%
9 (45)s 1.50 19.16 17.12 11.95%
Chapter 3. Experimental Study 38

Figure 3.1: The bottom and top plates for fabrication

Figure 3.2: Temperature control

Figure 3.3: The heating rods
module and thermocouple

(a) The top surface of E-glass Fibre (b) The bottom surface of E-glass Fibre

Figure 3.4: The glass fibre used for fabrication of laminate

Chapter 3. Experimental Study 39

(a) Applying first coat of (b) Placing the fabric

epoxy

(c) Compacting (d) Tightening the bolt

Figure 3.5: Procedure of fabricating laminates in steps

Figure 3.6: Placement of heating rods for curing

Chapter 3. Experimental Study 40

Figure 3.7: The curing of laminate

Figure 3.8: The initial support system using ball bearing

Chapter 3. Experimental Study 41

Figure 3.9: Schematic boundary condition

Figure 3.10: Schematic diagram of final experimental setup

Chapter 3. Experimental Study 42

(a) The details of stepped cut

(b) The base plate

Figure 3.11: Details of the experimental setup

Figure 3.12: The final experimental setup

Chapter 3. Experimental Study 43

Figure 3.13: Position of LVDTs along the plate

Figure 3.14: The load deformation curve of mild steel

Chapter 3. Experimental Study 44

Figure 3.15: The Load Deformation Curve of Unidirectional laminate with aspect ratio
0.75 ( at Bottom of U75S1)

Figure 3.16: The load deformation curve for U01S3 of square laminate (Aspect Ratio=1)
Chapter 3. Experimental Study 45

Figure 3.17: The load deformation curve rectangular laminate with Aspect Ratio=1.5
(at Centre, U15S1)

Figure 3.18: The load-deformation curve for orthotropic plates with aspect ratio 0.75
(at top of O75S3)
Chapter 3. Experimental Study 46

Figure 3.19: The load versus displacement at the center of O01S2

Figure 3.20: The load vs displacement for orthotropic plates with aspect ratio 1.5 (at
centre of O15S1)
Chapter 3. Experimental Study 47

Figure 3.21: The load deformation curve of angle plies with aspect ratio 0.75 (at centre
of Q75S1)

Figure 3.22: The load deformation curve of angle plies with aspect ratio 1 (at centre of
Q01S2)
Chapter 3. Experimental Study 48

Figure 3.23: The load deformation curve for angle plies of aspect ratio 1.5 (at centre of
Q15S1)
Chapter 4

Summary and Conclusions

4.1 Summary

Not many experimental studies on buckling of composite plates with in-plane boundary
conditions are available in literature. There are only a few studies, where the displacement
boundary conditions have been taken into account. In the present study, the effect of in-
plane restraints on the buckling of glass epoxy composite plates is studied analytically
and experimentally. Simply supported laminates of square and rectangular cross sections
were subjected to uniform in-plane loading. The effect of various fibre orientations and
aspect ratios of the plates on the critical buckling load has been studied. In the analytical
study, the glass epoxy plates have been modeled in ABAQUS 6.11 and various cases of
in-plane restraints were applied at the loading and unloading edges. Then the buckling
load has been evaluated from the eigenvalues analysis. In the experimental study, simply
supported plates of aspect ratio 0.75, 1 and 1.5 are tested. The test set-up is simple
and could apply in-plane restraint to the plate in a satisfactory manner. The in-plane
displacement is restrained along the unloaded edges. The effect of fibre orientation has
also been studied. This is achieved by fabricating specimens with three different fiber
orientations: unidirectional plate , crossply plate and (0/45)s . The results obtained are
compared with those from the analytical study.

49
Chapter 4. Summary and Conclusions 50

4.2 Conclusions

The present study demonstrates the importance of considering the in-plane boundary
condition along with the out-of-plane boundary conditions. The following observations
are made from the study:

1. The crossply plates show more compressive load carrying capacity compared to the
angle plied and unidirectional laminates. To evaluate the load carrying capacity of
laminates in analytical studies, modes of failure other than buckling were not taken
in to account.

2. The laminates with low aspect ratio are found to resist more compressive load. As
the aspect ratio increases, the unsupported length of the laminate that resists the
load increases and thus, it is more susceptible to buckling.

3. The load deformation curves are found to be varying for different specimens of same
aspect ratio and fiber orientation. This experimental variability may be due to

• fabrication errors: These include not properly wetting the fabric, variation in
curing time, lack of precise control in fibre orientation in manual method. Also,
there might be some geometric imperfections present in the laminate specimens.
In addition, the laminates were fabricated using a setup, which was kept at an
open space. The temperature was tested only at the midpoint of the specimen.
Although the heating rods were equally distributed along the length, this cannot
ensure uniform heating and thus curing of the laminate.

• experimental setup: The setup has its own limitations. Also, the method of
loading the laminate was not perfect. A uniform rate of loading could not be
ensured and thus, the load deformation curve obtained was not smooth. The
experimental study was conducted with limited resource and it was not possible
to modify the setup within the stipulated time.

4. Unidirectional laminates gave comparable results for both analytical and experimen-
tal studies. This implies that the experimental setup could simulate the required
boundary conditions to a good extent. But the same pattern could not be repeated
Chapter 4. Summary and Conclusions 51

for other laminates satisfactorily, particularly for the crossply laminates. This may
be due to the presence of other modes of early failure in experimental investigation,
which were not considered in the analytical study. For example, the fibres in the
laminate might have been damaged even before reaching the estimated buckling load,
whereas in the analytical model, it was assumed that the composite fails in buckling.

5. Other sources of variation in the experimentally and analytically obtained buckling

load for different specimens can be because of the following assumptions in analytical
modeling:

• The variation of thicknesses in the laminate: In the analytical study,

it was assumed that each lamina of the composite is of 1mm thick. But the
laminate fabricated in-house had thickness in the range of 3.35 mm to 3.60 mm.
This difference in the assumed and actual value could have some effect on the
buckling resistance.

• Fibre orientation of the fabric: In the analytical study, it was assumed that
all the fibres are aligned in the main fibre direction. But, this was not achieved
in the practical conditions as some nominal amount of fibre (about 5-10%) were
provided in the direction perpendicular to the main fibre so that it could be
held in place.

• Bond between the fibre and resin: In the analytical study, the bond be-
tween the fibre and resin was assumed to be perfect. But this may not be the
case in the actual scenario.

• Material properties: The analytical study was carried out assuming the
manufacturer supplied values of elastic moduli of E-glass fabric and epoxy resin.
No coupon test was conducted to determine the actual material properties of
the laminates. So this might have an influence on the buckling resistance of the
laminate.

4.3 Scope for Future Work

Some of the suggestions for future research are as follows:

Chapter 4. Summary and Conclusions 52

1. The fabrication procedure of laminates is needed to be improved using oven or other

modern devices.

2. The experimental set-up needs to be modified for the repeatability of the results.
The method of loading needs to be perfected.

3. The effect of the size of the specimens with same aspect ratio on the buckling load
can be investigated.

4. The experimental set-up may be modified for other in-plane boundary conditions
and its effect can be studied.

5. The experimental study is limited to the symmetric composite laminates. Asymmet-

ric laminates are needed to be investigated as well.

6. The effect of cut-out on the buckling load can be studied. The cut-outs can be
centered or eccentric. The effect of variation of sizes and shapes of cut-outs can also
be studied.

7. The effect of non-uniform loading on the buckling of plate are also required to be
studied through analytical modeling as well as experiments.
Appendix A

The Experimental Results:

Graphs

A.1 General

In this appendix, all the results obtained from the experiment is depicted in the form of
graphs. The intention behind this section is mention how the buckling load was selected
from the graph and it was not feasible to include all the graphs in the main chapter
(Chapter 3 ).

Glass-epoxy laminates of three different fibre orientations are cast and tested with three
samples each for the same specimen. The experiment is conducted and data is collected
from the DAQ system. The data obtained is processed and graphs are plotted. From these
load deformation curves, it is sometimes difficult to determine the buckling load directly,
as there are bumps in the curve. The deformation profile of the specimen in plotted from
the data from LVDTs at a particular load and from this the zone of buckling can be
determined approximately.

53
Appendix A. The Experimental Results 54

A.2 Unidirectional Laminates

Three specimens U75S1, U75S2 and U75S3 of thickness 3.75mm, 3.36 mm and 3.62 mm
were tested for buckling load. But, due to mistake in recording the data, the U75S3 gave
unusable results. The other two specimens buckled at almost same load which is very
much close to the estimated buckling load (27.49 kN). While the first specimen buckled at
a load of 27.5 KN approximately, the other buckled at a higher value of 28 kN with 1.8%
deviation. The graphs for the laminates are plotted here (Figures A.1 and A.2 ).
The square laminates with fibres arranged in unidirection had thickness’s of 3.37 mm,
3.36 mm and 3.56 mm respectively with specimen IDs U01S1, U01S2 and U01S3. From
the figures of U01S1, it can be seen that there is no deformation of the laminate till the
load reaches 15 kN and afterwards the displacement increases gradually and fails at 37.5
kn while the estimated value is only 19 kN. This might have occured because the load was
applied on the laminate at a greater rate at the starting and thus a sudden load of 15 kN
was imposed on the laminate and later it acted as seating load when the rate of the load
application was made gradual. The other two specimens gave comparable results with a
deviation of a maximum of 4.49 kN. Figures A.3 - A.6 gives the various plots for the
specimens.
Rectangular laminates with aspect ratio 1.5 also when conducted experiment showed
value close to the estimated value of 18.15 kN except for one laminate. The laminates
U15S1, U15S2 and U15S3 had thickness’s 3.50 mm, 3.31 mm and 3.42 mm respectively.
The laminate with ID U15S2 had a buckling load of 10.5kN. The various load deformation
curves are given in Figures A.7 - A.9. The displacement profile of the laminate is given
in Figure A.10.

A.3 Cross-ply Laminates

Similar to the case of unidirectional laminates, three specimens of three different aspect
ratios are tested for cross-ply laminates with fibre orientation ((0/90)s ). The results ob-
tained are explained in detail. Unlike the unidirectional laminates where, the main fibre
direction was same in all laminae, in cross-ply laminates has main fibre in orthogonal
Appendix A. The Experimental Results 55

directions. The experimental result obtained varies from the analytical value by a per-
centage of 6-55 in general. A reason for this could be that when the modulus of elasticity
is calculated, it was assumed that the fibres are oriented only in the main direction. But,
practically this is not possible as some amount of fibre should be there in the other direc-
tion so as to hold the main fibre in place. Usually 10% of fibres will be arranged in the
perpendicular direction while the 90% will be in the main fibre direction.

For the laminates with aspect ratio 0.75, the estimated buckling load was 35.58 kN. Four
specimens, O75S1, O75S2, O75S3 and O75S4 of thickness’s 3.48 mm, 3.49 mm, 3.65 mm
and 3.48 mm respectively was tested for critical buckling load. No proper data was ob-
tained from O75S2 due to error in recording the data. The other three laminates showed
a deviation of upto 38% from the estimated value. Figures A.11 - A.13 shows the load
deformation curves for the three specimens. The displacement profile of the laminate is
given in Figure A.14. The displacement profile of O75S1 is not plotted because the graphs
obtained for load deformation curve is messy and profile may not simulate the actual be-
haviour of the laminate. An interesting observation from the displacement profile of the
O75S4 is that the deformation of laminate is near to 1 mm even when the load is at
20kN.But after that, the value increases to around 6 mm for a load of 26 kN.

Three specimens O01S1, O01S2 and O01S3 of 3.36 mm, 3.46 mm and 3.58 mm thickness
respectively is tested experimentally for aspect ratio 1. The data obtained for O01S1 was
so absurd and cannot be used for the study purposes. The load deformation curves for
the rest of the specimens in given in Figures A.15 and A.16. The estimated value from
the analytical model for square laminates is 29.52 kN and surprisingly, the value obtained
from the experiment varies from this data by about 49%. The displacement profile of the
laminate is given in Figure A.17.
Rectangular laminates with aspect ratio 1.5 are tested experimentally and critical
buckling load was obtained from the load deformation curve. Three laminates of thickness’s
3.46 mm, 3.54 mm and 3.53 mm with specimen IDs O15S1, O15S2 and O15S3 are tested.
Being longer, as expected, the buckling load was found to be lower than all the other
laminates of same orientation. The estimated buckling load is 19.81 kN. O15S1 buckled at
load of 18.5 kN while the other two laminates deviates from the expected value. O15S2 has
Appendix A. The Experimental Results 56

a buckling load of 14.73 kN and O15S2 has a load of 9.5 kN only. All the load deformation
curves are depicted from figures A.18 - A.20. The displacement profiles of all the three
specimens are plotted in figure A.21.
Appendix A. The Experimental Results 57

(a) Load Deformation Curve for U75S1 (with all the (b) Load Deformation Curve at Centre of U75S1
LVDTs)

(c) Load Deformation Curve at Bottom of U75S1 (d) Displacement Profile of U75S1 for Various Loads

Figure A.1: Load Deformation Curve for U75S1

Appendix A. The Experimental Results 58

Figure A.2: The Load vs Displacement at Bottom of U75S2

Appendix A. The Experimental Results 59

(a) Load Deformation Curve for U01S1 (with all the (b) Load Deformation Curve at Centre of U01S1
LVDTs)

(c) Load Deformation Curve at Bottom of U01S1 (d) Load Deformation Curve at Top of U01S1

Figure A.3: Load Deformation Curve for U01S1

Appendix A. The Experimental Results 60

(a) Load Deformation Curve for U01S2 (with all the (b) Load Deformation Curve at Centre of U01S2
LVDTs)

(c) Load Deformation Curve at Bottom of U01S2 (d) Load Deformation Curve at Top of U01S2

Figure A.4: Load Deformation Curve for U01S2

Appendix A. The Experimental Results 61

(a) Load Deformation Curve for U01S3 (with all the (b) Load Deformation Curve at Centre of U01S3
LVDTs)

(c) Load Deformation Curve at Bottom of U01S3 (d) Load Deformation Curve at Top of U01S3

Figure A.5: Load Deformation Curve for U01S3

Appendix A. The Experimental Results 62

(a) Displacement Profiles for U01S2 (b) Displacement Profiles for U01S3

Figure A.6: Displacement Profiles for U01S2 and U01S3

Appendix A. The Experimental Results 63

(a) Load Deformation Curve for U15S1 (with all the (b) Load Deformation Curve at Centre of U15S1
LVDTs)

(c) Load Deformation Curve at Bottom of U15S1 (d) Load Deformation Curve at Top of U15S1

Figure A.7: Load Deformation Curve for U15S1

Appendix A. The Experimental Results 64

(a) Load Deformation Curve for U15S2 (with all the (b) Load Deformation Curve at Centre of U15S2
LVDTs)

(c) Load Deformation Curve at Bottom of U15S2 (d) Load Deformation Curve at Top of U15S2

Figure A.8: Load Deformation Curve for U15S2

Appendix A. The Experimental Results 65

(a) Load Deformation Curve for U15S3 (with all the (b) Load Deformation Curve at Centre of U15S3
LVDTs)

(c) Load Deformation Curve at Bottom of U15S3 (d) Load Deformation Curve at Top of U15S3

Figure A.9: Load Deformation Curve for U15S3

Appendix A. The Experimental Results 66

(a) Displacement Profiles for U15S1 (b) Displacement Profiles for U15S2

(c) Displacement Profiles for U15S3

Figure A.10: Displacement Profiles for U15S1, U15S2 and U15S3

Appendix A. The Experimental Results 67

(a) Load Deformation Curve for O75S1 (with all the (b) Load Deformation Curve at Centre of O75S1
LVDTs)

(c) Load Deformation Curve at Bottom of O75S1 (d) Load Deformation Curve at Top of O75S1

Figure A.11: Load Deformation Curve for O75S1

Appendix A. The Experimental Results 68

(a) Load Deformation Curve for O75S3 (with all the (b) Load Deformation Curve at Centre of O75S3
LVDTs)

(c) Load Deformation Curve at Bottom of O75S3 (d) Load Deformation Curve at Top of O75S3

Figure A.12: Load Deformation Curve for O75S3

Appendix A. The Experimental Results 69

(a) Load Deformation Curve for O75S4 (with all the (b) Load Deformation Curve at Centre of O75S4
LVDTs)

(c) Load Deformation Curve at Bottom of O75S4 (d) Load Deformation Curve at Top of O75S4

Figure A.13: Load Deformation Curve for O75S4

Appendix A. The Experimental Results 70

(a) Displacement Profiles for O75S3 (b) Displacement Profiles for O75S4

Figure A.14: Displacement Profiles for O75S3 and O75S4

Appendix A. The Experimental Results 71

(a) Load Deformation Curve for O01S2 (with all the (b) Load Deformation Curve at Centre of O01S2
LVDTs)

(c) Load Deformation Curve at Bottom of O01S2 (d) Load Deformation Curve at Top of O01S2

Figure A.15: Load Deformation Curve for O01S2

Appendix A. The Experimental Results 72

(a) Load Deformation Curve for O01S3 (with all the (b) Load Deformation Curve at Centre of O01S3
LVDTs)

(c) Load Deformation Curve at Bottom of O01S3 (d) Load Deformation Curve at Top of O01S3

Figure A.16: Load Deformation Curve for O01S3

Appendix A. The Experimental Results 73

(a) Displacement Profiles for O01S2 (b) Displacement Profiles for O01S3

Figure A.17: Displacement Profiles for O01S2 and O01S3

Appendix A. The Experimental Results 74

(a) Load Deformation Curve for O15S1 (with all the (b) Load Deformation Curve at Centre of O15S1
LVDTs)

(c) Load Deformation Curve at Bottom of O15S1 (d) Load Deformation Curve at Top of O15S1

Figure A.18: Load Deformation Curve for O15S1

Appendix A. The Experimental Results 75

A.4 Angle ply Laminates

Even though three different angle plies were modelled in analytical study, only angle plies
at 45o was fabricated to test the results experimentally ( (0/45)s ). Three specimens of
each aspect ratio were fabricated in the laboratory. It was interesting to note that unlike
the other laminates, critical buckling load obtained from the experiment for some speci-
mens exceed the estimated value. This may be due to the fact that the fibres in the other
direction may be helping in increasing the resistance of the laminate against compressive
load. The orientation of the plate may also be not accurately 45o . This may also cause
this result.
Laminates of 0.75 aspect ratio with specimen IDs Q75S1, Q75S2 and Q75S3 (with thick-
ness’s 3.27 mm, 3.28 mm and 3.54 mm respectively) were fabricated and tested for com-
pressive load. Out of which data from Q75S2 has to be rejected due to the error in
recording the data. The rest of the results obtained are closer to the estimated value
(27.87 kN), but one is higher and other is lower than the expected value. Figures A.22
and A.23 gives the load deformation curve for rectangular laminates with low aspect ratio.
The displacement profile of Q75S3 is plotted in figure A.24.

Square laminates with thickness’s 3.61 mm, 3.42 mm and 3.45 mm were cast with IDs
Q01S1, Q01S2 and Q01S3 respectively. The estimate value of buckling load from the ana-
lytical model was found to be 18.82 kN. The values obtained from the experiment deviated
by a maximum of 22.2%. Two of the laminates buckled at a higher value than expected.
This may be because the fibres in the perpendicular direction increasing the strength of
the laminate or as stated earlier the fibres may not be aligned properly while fabrication.
It should also be noted that the plate under study is square laminate. From the above
sections also, it can be seen that square laminates behave a little differently compared to
the rectangular laminates ie, the deviation in buckling load for square laminates are more.
Figures A.25 - A.27 gives the load deformation curve for the square laminates. The
displacement profile of Q01S3 is plotted in figure A.28. Three rectangular laminate
specimens Q15S1, Q15S2 and Q15S3 of thickness’s 3.57 mm, 3.68 mm and 3.52 mm re-
spectively with aspect ratio 1.5 are tested experimentally to determine the critical buckling
load. Q15S1 behaved differently from the other two specimens as can be seen from the
Appendix A. The Experimental Results 76

figures A.29 - A.31 . The estimated load was 17.12 kN and Q15S3 gave a value of 16.52
kN with a deviation of 3.5%. The displacement profile of the specimens are given in figure
A.32.
Appendix A. The Experimental Results 77

(a) Load Deformation Curve for O15S2 (with all the (b) Load Deformation Curve at Centre of O15S2
LVDTs)

(c) Load Deformation Curve at Bottom of O15S2 (d) Load Deformation Curve at Top of O15S2

Figure A.19: Load Deformation Curve for O15S2

Appendix A. The Experimental Results 78

(a) Load Deformation Curve for O15S3 (with all the (b) Load Deformation Curve at Centre of O15S3
LVDTs)

(c) Load Deformation Curve at Bottom of O15S3 (d) Load Deformation Curve at Top of O15S3

Figure A.20: Load Deformation Curve for O15S3

Appendix A. The Experimental Results 79

(a) Displacement Profiles for O15S1 (b) Displacement Profiles for O15S2

(c) Displacement Profiles for O15S3

Figure A.21: Displacement Profiles for O15S1, O15S2 and O15S3

Appendix A. The Experimental Results 80

(a) Load Deformation Curve for Q75S1 (with all the (b) Load Deformation Curve at Centre of Q75S1
LVDTs)

(c) Load Deformation Curve at Bottom of Q75S1 (d) Load Deformation Curve at Top of Q75S1

Figure A.22: Load Deformation Curve for Q75S1

Appendix A. The Experimental Results 81

(a) Load Deformation Curve for Q75S3 (with all the (b) Load Deformation Curve at Centre of Q75S3
LVDTs)

(c) Load Deformation Curve at Bottom of Q75S3 (d) Load Deformation Curve at Top of Q75S3

Figure A.23: Load Deformation Curve for Q75S3

Appendix A. The Experimental Results 82

Figure A.24: Displacement Profile for Q75S3

Appendix A. The Experimental Results 83

(a) Load Deformation Curve for Q01S1 (with all the (b) Load Deformation Curve at Centre of Q01S1
LVDTs)

(c) Load Deformation Curve at Bottom of Q01S1 (d) Load Deformation Curve at Top of Q01S1

Figure A.25: Load Deformation Curve for Q01S1

Appendix A. The Experimental Results 84

(a) Load Deformation Curve for Q01S2 (with all the (b) Load Deformation Curve at Centre of Q01S2
LVDTs)

(c) Load Deformation Curve at Bottom of Q01S2 (d) Load Deformation Curve at Top of Q01S2

Figure A.26: Load Deformation Curve for Q01S2

Appendix A. The Experimental Results 85

(a) Load Deformation Curve for Q01S3 (with all the (b) Load Deformation Curve at Centre of Q01S3
LVDTs)

(c) Load Deformation Curve at Bottom of Q01S3 (d) Load Deformation Curve at Top of Q01S3

Figure A.27: Load Deformation Curve for Q01S3

Appendix A. The Experimental Results 86

Figure A.28: Displacement Profile for Q01S3

Appendix A. The Experimental Results 87

(a) Load Deformation Curve for Q15S1 (with all the (b) Load Deformation Curve at Centre of Q15S1
LVDTs)

(c) Load Deformation Curve at Bottom of Q15S1 (d) Load Deformation Curve at Top of Q15S1

Figure A.29: Load Deformation Curve for Q15S1

Appendix A. The Experimental Results 88

(a) Load Deformation Curve for Q15S2 (with all the (b) Load Deformation Curve at Centre of Q15S2
LVDTs)

(c) Load Deformation Curve at Bottom of Q15S2 (d) Load Deformation Curve at Top of Q15S2

Figure A.30: Load Deformation Curve for Q15S2

Appendix A. The Experimental Results 89

(a) Load Deformation Curve for Q15S3 (with all the (b) Load Deformation Curve at Centre of Q15S3
LVDTs)

(c) Load Deformation Curve at Bottom of Q15S3 (d) Load Deformation Curve at Top of Q15S3

Figure A.31: Load Deformation Curve for Q15S3

Appendix A. The Experimental Results 90

(a) Displacement Profiles for Q15S1 (b) Displacement Profiles for Q15S2

(c) Displacement Profiles for Q15S3

Figure A.32: Displacement Profiles for Q15S1, Q15S2 and Q15S3

Bibliography

ABAQUS (2011). Abaqus 6.11 Online Documentation. Dassault Systemes.

Bambach, M. R. and Rasmussen, K. J. R. (2004). Experimental techniques for testing un-

stiffened plates in compression and bending. Technical report, Society for Experimental
Mechanics.

Chai, G. B. (1991). Large deflections of laminated composite plates. Composites Science

and Technology.

Chailleux, A., Hang, Y., and Verchery, G. (1975). Experimental study of the buckling of
laminated composite columns & plates. International Journal of mech. Sci.

Dietrich, L., Kawahara, W., and Phillips, A. (1978). An experimental study of plastic
buckling of a simply supported plate under edge thrusts. Acta Mechanica.

Herencia, J., Weaver, P. M., and Friswell, M. I. (2010). Closed form solutions for buckling
of long anisotropic plates with various boundary conditions under axial compression.
Journal of Engineering Mechanics.

Husam, A.-Q., Dwairi, H., Shatarat, N., Rosan, T., and Al-Qablan, T. (2010). Stability
analysis of composite panels with stiffeners and circular cutouts. Jordan Journal of Civil
Engineering.

Prajapat, K. (2011). Analytical and experimental study of buckling of plates considering

in-plane boundary conditions. Master’s thesis, Indian Institute of Technology, Kanpur.

Qing-Qing, N., Xie, J., and Iwamoto, M. (2005). Buckling analysis of laminated composite
plates with arbitrary edge supports. Compo.

91
Bibliography 92

Roberts, J. C., G, B., and G.J., W. (1999). Experimental,numerical and analytical results
for bending and buckling of rectangular orthotropic plates. Composite Structures.

Senthil, K., Arokiarajan, A., Palaninathan, R., Santhosh, B., and Usha, K. M. (2013).
Experimental and numerical studies on adhesively bonded cfrp laminates with close
debonds. Composite Structures.

Shrivastava, A. K. and Singh, R. K. (1999). Effect of aspect ratio on buckling of composite

plates. Composites Science and Technology.

Yang, Q. J., Hayman, B., and Osnes, H. (2013). Simplified buckling and ultimate strength
analysis of composite plates in compression. Composites.