MODAL PROPERTIES OF SINGLE-EDGE CRACKED

RC COLUMN STRENGTHENED WITH FRP:

ANALYTICAL AND NUMERICAL STUDIES

A Thesis Submitted

in Partial Fulfillment of the Requirements

for the Degree of

Master of Technology

by
VEMURI V R P S KUMAR

DEPARTMENT OF CIVIL ENGINEERING

INDIAN INSTITUTE OF TECHNOLOGY,KANPUR

MAY 2015
 iii

ABSTRACT

The use of fibre reinforced polymer (FRP) materials for retrofitting existing structures is gain-

ing importance. The advantages are high strength to weight ratio, ease of installation and low

maintenance costs. Typically, the structural members are being subjected to dynamic loads such

as moving and impulse loads in the case of bridge piers. Hence knowledge of dynamic properties

of structural elements retrofitted with FRP is useful. In this study, the modal properties of a cracked

reinforced concrete (RC) column strengthened with FRP are studied. At first, continuous cracked

beam theory is used to formulate the free vibration equations of motion for a cracked concrete

column. A solution methodology is proposed to estimate the frequencies and mode shapes using

Rayleigh’s quotient. A numerical study is then performed in ABAQUS/CAE v 6.10 to validate

the results from formulation. The effect of crack parameters on the modal properties is studied.

Further, to study the dependence of mode shapes and frequencies on the damage location, pertur-

bation formulation is used. The expressions for difference in mode shapes and frequencies as a

function of crack depth and crack position are obtained. By analysing these expressions, it has

been observed that the difference in mode shapes and frequencies is sensitive to the curvature of

corresponding mode shapes depending on the location of damage. These findings are illustrated

using the results of the cracked column. The continuous cracked beam theory is then modified to

incorporate the effects of steel reinforcement and FRP confinement. A similar solution method-

ology based on Rayleigh’s quotient is also used here. Further, a finite element analysis is carried

out by modelling a cracked column along with steel and FRP constraints in ABAQUS/CAE v 6.10

to compare the frequencies and mode shapes. This comparison indicates that the results from the

formulation match reasonably well with the findings from ABAQUS/CAE v 6.10. Also, it is ob-

served that the effect of steel bar’s restraints and FRP confinement should be considered for better

estimation of modal properties . In particular, it has been found that the frequencies of the cracked

RC column will increase when these effects are taken into consideration.
Dedicated to Amma, Nanna and Akka
 v

ACKNOWLEDGEMENT

I sincerely thank my thesis supervisors Dr.Samit Ray Chaudhuri and Dr. P. M. Mohite for their

guidance and motivation the whole time, which enabled the successful completion of this work.

I am really thankful to my professors Samit Ray Chaudhuri, Vinay.K.Gupta, Durgesh.C.Rai,

Sudib.K.Mishra, D. Yadav, Sudhir Mishra and S.K.Chakraborty. It is with great sincerity that I am

thanking them for their love and willingness to help and more over to introduce me the vast world

of knowledge.

I am really thankful to my lab mates Kanta di, Sanjukta di, Koushik da, Gourab, Dikshant for

creating such a wonderful environment.

I have been extremely fortunate to have a group of lovely friends without whom my stay in

IIT would be rather dull. The time spent with Prasanna, Valli, Naveen, Rahitya and others was

extremely refreshing.

Finally I would express my gratitude towards my parents, my sister and all my well wishers,

relatives and friends to provide me moral support.

Vemuri V R P S Kumar
 vi

Contents

Abstract iii

Acknowledgement v

Contents v

List of Figures ix

List of Tables xiii

1 Introduction 1

1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.3 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.4 Organization of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 Free Vibration of Cantilever Column with Crack 7

2.1 Brief Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 Theoretical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2.1 Cracked Beam Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2.2 P-delta Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.3 Equation of Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.3.1 Geometry of Cantilever Column . . . . . . . . . . . . . . . . . . . . . . 10

 vii

2.3.2 Displacement and Strain Fields . . . . . . . . . . . . . . . . . . . . . . 10

2.3.3 Energy Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.3.3.1 Kinetic Energy . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.3.3.2 Strain Energy . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.3.3.3 Potential Energy . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.3.4 Hamilton’s Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.3.5 Calculation of Exponential Decay Rate α . . . . . . . . . . . . . . . . . 16

2.4 Solution using Rayleigh’s Quotient . . . . . . . . . . . . . . . . . . . . . . . . . 18

2.4.1 Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2.4.2 Solution Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.4.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.4.3.1 Dependence of Modal Properties on Crack Parameters . . . . . 23

2.5 Finite Element Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2.5.1 Geometry and Material Properties . . . . . . . . . . . . . . . . . . . . . 25

2.5.2 Boundary Conditions and Meshing . . . . . . . . . . . . . . . . . . . . 25

2.5.3 Analysis Step and Mesh Size . . . . . . . . . . . . . . . . . . . . . . . . 26

2.5.4 Damage Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2.6 Comparison of Theoretical and Numerical Modal Properties . . . . . . . . . . . 27

2.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

3 Free Vibration of Cracked Column with FRP and Steel Reinforcement 39

3.1 Brief Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

3.2 Compatibility of Deformations . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

3.3 Strain Displacement Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

3.4 Energy Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.4.1 Kinetic Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.4.2 Strain Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

 viii

3.4.3 Potential Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.5 Solution using Rayleigh’s Quotient . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.5.1 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

3.6 Numerical Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

3.6.1 Geometry and Material Modelling . . . . . . . . . . . . . . . . . . . . . 47

3.6.2 Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

3.6.3 Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

3.6.4 Meshing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

3.7 Comparison of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

3.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

4 Effect of Crack Position and Crack Depth on Modal Properties 67

4.1 Brief Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

4.2 Perturbation Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

4.3 Modal Properties of Cantilever Column . . . . . . . . . . . . . . . . . . . . . . 70

4.4 Numerical Illustration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

4.4.1 Effect of Crack Position on Frequencies . . . . . . . . . . . . . . . . . . 71

4.4.2 Effect of Crack position on Mode shapes . . . . . . . . . . . . . . . . . 72

4.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

5 Summary and Conclusions 81

5.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

5.2 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

5.3 Scope for future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

References 83
 ix

List of Figures

2.1 Distribution of secondary moments due to P-delta effect: (a) Configuration of

cantilever column; (b) Distribution of secondary moments . . . . . . . . . . . . 28

2.2 Geometry of cantilever column with an open-edge crack . . . . . . . . . . . . . 29

2.3 Variation of exponential decay constant with crack depth ratio . . . . . . . . . . 29

2.4 Effect of crack depth ratio on first frequency of cantilever column for different

crack positions of xc /L = 0.1, 0.25 and 0.4 . . . . . . . . . . . . . . . . . . . . 30

2.5 Effect of crack position on fundamental frequency of cantilever column for differ-

ent crack depth ratios of a/d = 0.1, 0.15 , 0.2 and 0.25 . . . . . . . . . . . . . . 30

2.6 Comparison of first mode shapes for uncracked and cracked cantilever column for

crack position xc /L = 0.25: (a) Crack depth ratio a/d = 0.2; (b) Crack depth ratio

a/d = 0.3; (c) Crack depth ratio a/d = 0.4 . . . . . . . . . . . . . . . . . . . . . 31

2.7 Comparison of second mode shapes for uncracked and cracked cantilever column

for crack position xc /L = 0.25: (a) Crack depth ratio a/d = 0.2; (b) Crack depth

ratio a/d = 0.3; (c) Crack depth ratio a/d = 0.4 . . . . . . . . . . . . . . . . . . 32

2.8 Comparison of first mode shapes for uncracked and cracked cantilever column for

crack depth ratio a/d = 0.25: (a) Crack position xc /L = 0.1; (b) Crack position

xc /L = 0.2; (c) Crack position xc /L = 0.3 . . . . . . . . . . . . . . . . . . . . . 33

 x

2.9 Comparison of second mode shapes for uncracked and cracked cantilever column

for crack depth ratio a/d = 0.25: (a) Crack position xc /L = 0.1; (b) Crack position

xc /L = 0.2; (c) Crack position xc /L = 0.3 . . . . . . . . . . . . . . . . . . . . . 34

2.10 ABAQUS model of cracked concrete column . . . . . . . . . . . . . . . . . . . 35

2.13 Convergence of finite element result . . . . . . . . . . . . . . . . . . . . . . . . 35

2.11 A typical C3D10 element in FE analysis . . . . . . . . . . . . . . . . . . . . . . 36

2.12 ABAQUS model of cracked cantilever column . . . . . . . . . . . . . . . . . . . 36

2.14 Comparison of analytical and finite element fundamental mode shapes : (a) Case

I; (b) Case II; (c) Case III . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.1 Variation of fundamental frequency ratio for concrete and RC columns with crack

position ratio for: (a) CDR = 0.1; (b) CDR = 0.2; (c) CDR = 0.3; (d) CDR = 0.4 . 54

3.2 Variation of second frequency ratio for concrete and RC columns with crack posi-

tion ratio for: (a) CDR = 0.1; (b) CDR = 0.2; (c) CDR = 0.3; (d) CDR = 0.4 . . . 55

3.3 Variation of fundamental frequency ratio (ωCRF /ωU ) for RCF1, RCF2, RCF3

with crack position ratio for: (a) CDR = 0.1; (b) CDR = 0.2; (c) CDR = 0.3; (d)

CDR = 0.4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

3.4 Variation of second frequency ratio (ωCRF /ωU ) for RCF1, RCF2, RCF3 with

crack position ratio for: (a) CDR = 0.1; (b) CDR = 0.2; (c) CDR = 0.3; (d) CDR =

0.4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

3.5 Fundamental mode shape for CDR 0.3 and CPR 0.1 for (a) Concrete, RC, RCF1,

RCF2 and RCF3; Difference in mode shapes for (b) Concrete and RC, (c) Concrete

and RCF1, (d) Concrete and RCF2, (e) Concrete and RCF3 . . . . . . . . . . . . 58

3.6 Fundamental mode shape for CDR 0.4 and CPR 0.1 for (a) Concrete, RC, RCF1,

RCF2 and RCF3; Difference in mode shapes for (b) Concrete and RC, (c) Concrete

and RCF1, (d) Concrete and RCF2, (e) Concrete and RCF3 . . . . . . . . . . . . 59
 xi

3.7 Second mode shape for CDR 0.3 and CPR 0.1 for (a) Concrete, RC, RCF1, RCF2

and RCF3; Difference in mode shapes for (b) Concrete and RC, (c) Concrete and

RCF1, (d) Concrete and RCF2, (e) Concrete and RCF3 . . . . . . . . . . . . . . 60

3.8 Second mode shape for CDR 0.4 and CPR 0.1 for (a) Concrete, RC, RCF1, RCF2

and RCF3; Difference in mode shapes for (b) Concrete and RC, (c) Concrete and

RCF1, (d) Concrete and RCF2, (e) Concrete and RCF3 . . . . . . . . . . . . . . 61

3.9 FRP shell model in ABAQUS with partitions . . . . . . . . . . . . . . . . . . . 62

3.10 Assembly of RC Column with FRP . . . . . . . . . . . . . . . . . . . . . . . . . 62

3.11 Details of reinforcement in column . . . . . . . . . . . . . . . . . . . . . . . . . 63

3.12 Meshed view of FRP cover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

3.13 Comparison of analytical and numerical cracked column first mode frequencies

for (a) RC column, (b) RCF1, (c) RCF2, (d) RCF3 . . . . . . . . . . . . . . . . 64

3.14 Comparison of analytical and numerical cracked column second mode frequencies

for (a) RC column, (b) RCF1, (c) RCF2, (d) RCF3 . . . . . . . . . . . . . . . . 65

4.1 Variation of first mode shape difference at crack location with crack position ratio 73

4.2 Variation of second mode shape difference at crack location with crack position ratio 73

4.3 Variation of third mode shape difference at crack location with crack position ratio 73

4.4 First mode shape of cantilever column: (a) mode shape, (b) slope of mode shape,

(c) curvature of mode shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

4.5 Second mode shape of cantilever column: (a) mode shape, (b) slope of mode

shape, (c) curvature of mode shape . . . . . . . . . . . . . . . . . . . . . . . . . 74

4.6 Third mode shape of cantilever column: (a) mode shape, (b) slope of mode shape,

(c) curvature of mode shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

4.7 Effect of crack position on first frequency ratio . . . . . . . . . . . . . . . . . . 75

4.8 Effect of crack position on second frequency ratio . . . . . . . . . . . . . . . . . 76

4.9 Effect of crack position on third frequency ratio . . . . . . . . . . . . . . . . . . 76

 xii

4.10 Difference in first mode shapes between cracked and healthy column for crack

position 0.1L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

4.11 Difference in first mode shapes between cracked and healthy column for crack

position 0.3L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

4.12 Difference in first mode shapes between cracked and healthy column for crack

position 0.5L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

4.13 Difference in first mode shapes between cracked and healthy column for crack

position 0.9L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

4.14 Difference in second mode shapes between cracked and healthy column for crack

position 0.1L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

4.15 Difference in second mode shapes between cracked and healthy column for crack

position 0.2L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

4.16 Difference in second mode shapes between cracked and healthy column for crack

position 0.5L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

4.17 Difference in second mode shapes between cracked and healthy column for crack

position 0.8L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

4.18 Difference in third mode shapes between cracked and healthy column for crack

position 0.1L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

4.19 Difference in third mode shapes between cracked and healthy column for crack

position 0.3L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

4.20 Difference in third mode shapes between cracked and healthy column for crack

position 0.5L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

4.21 Difference in third mode shapes between cracked and healthy column for crack

position 0.7L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
 xiii

List of Tables

2.1 Parameters of cantilever column . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.2 Details of mesh sizes used in convergence study . . . . . . . . . . . . . . . . . . 26

2.3 Comparison of fundamental frequencies from analytical and finite element results 27

3.1 Parameters of Cantilever Column . . . . . . . . . . . . . . . . . . . . . . . . . . 50

3.2 Nomenclature used in results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

3.3 Frequency ratio (ωC /ωU ) in terms of crack depth at different crack positions for

concrete column . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

3.4 Frequency ratio (ωCR /ωU R ) in terms of crack depth at different crack positions

for RC column . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

3.5 Frequency ratio (ωCRF /ωU R ) in terms of crack depth at different crack positions

for column RCF1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

3.6 Frequency ratio (ωCRF /ωU R ) in terms of crack depth at different crack positions

for column RCF2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

3.7 Frequency ratio (ωCRF /ωU R ) in terms of crack depth at different crack positions

for column RCF3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

3.8 Average percentage error in first frequency ratio with respect to numerical result

for different crack depth ratios . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

3.9 Average percentage error in second frequency ratio with respect to numerical re-

sult for different crack depth ratios . . . . . . . . . . . . . . . . . . . . . . . . . 53

 xiv

4.1 Mode shapes and curvatures of first three modes of cantilever column . . . . . . 70
Chapter 1

Introduction

1.1 Overview

Large number of buildings, bridges and structural elements are facing the need for restoration and

rehabilitation. The reasons include ageing of structural elements, under estimation of design loads,

construction flaws, demand for increased service loads owing to a change in its use, changing code

requirements and damage to a structure due to several actions such as over loading, earthquake,

wind and blast loading. Various strengthening techniques are available for such purpose. These

are increasing the cross-section by casting additional reinforced concrete (section enlargement

technique), putting additional steel plates to the tension side in case of beams and providing steel

casings in case of columns, increasing the pre-stressing etc. The drawbacks of these techniques in-

clude adding weight to the existing structure, causing disruption to normal activity and limitations

in shape and size of pre-cast steel sections due to handling difficulties. These issues can easily be

handled by use of polymer composites for repair and rehabilitation.

Fibre-reinforced polymers (FRP) are a composite material made of a polymer matrix and fi-

bres. There are different kinds of fibres commonly used in civil engineering application including

glass, carbon and aramid. Applications of FRP can also be found in many areas such as aerospace,

defence, marine, corrosion resistant equipment, and automotive sectors. FRPs are available in sev-
2

eral forms such as bars, cables, two dimensional and three dimensional grids, strips, sheets, plates

and pultruded sections. FRP bars can be used as reinforcement in concrete instead of steel. FRP

grids, strips, sheets and plates are used for external reinforcement and confinement. Pultruded

sections are used as bridge decks. The advantages of using FRP for repair and retrofit purposes

are listed as follows:

• FRP has higher ultimate strength and lower density than steel. Also, these materials are

having high strength to weight ratio. Hence use of FRP does not lead to significant increase

in weight to the existing structure. Also, member dimensions are least affected by use of

FRP.

• These materials are flexible and are available in long lengths. This feature adds advantages

in terms of installation as joints are not required, material can take irregular shapes and

curved boundaries are easily tackled.

• These materials are durable and hence, requirement for maintenance due to environmental

issues is low. Even if materials are damaged, it is relatively easy to repair by application of

one or more additional layers.

• Because of less weight, handling and installation become easier. Drilling is not required in

the structure to support FRP layers. This reduces the risk of damage in the existing structure.

• During installation of FRPs, regular activities can take place with no or minimal disruption.

This is very important in case of bridges with heavy traffic movement.

Inspite of aforementioned advantages, there are some disadvantages which limits the usage of

FRP. These can be listed as follows:

• FRP is a brittle material. It shows linear elastic performance until failure.

• When compared to other traditional civil engineering construction materials, FRP has rela-

tively higher material and fabrication costs.

 3

• FRP materials are combustible and have low fire resistance. Loss of strength begins at a low

temperature in comparison to steel.

• In many countries like India, design codes are not available for use of FRP in structural

applications.

In recent years, FRP composite materials are identified as a good alternate to the conventional

methods for rehabilitating deteriorated structures. Repair of sections such as beams and columns

is very common. In many of these cases, the structural members are being subjected to dynamic

loads, such as in case of bridge piers subjected to moving and impulse loads. The static behaviour

of FRP strengthened members has been extensively studied in the last few decades.

1.2 Literature Review

The focus of this study is placed on dynamic behaviour of cracked beam.

A comprehensive review on the vibration of cracked structures can be found in Dimarogonas

(1996). The review provides the vibration analysis of cracked beams, plates, rotors, turbine blades,

pipes and shells. The analysis methods for cracked beams based on modelling the crack has been

classified as equivalent reduced section models, lumped flexibility models, and continuous cracked

beam models. The equivalent reduced section method takes into account the effect of crack using

a local bending moment and a reduced section. Lumped flexibility models considered the cracked

beam as number of beam segments connected with rotational springs at crack location. Besides

rotational springs, linear springs were also used to model the crack.

The continuous cracked beam model was developed by Christides and Barr (1984) to incor-

porate the parameters of the crack i.e., location and depth of the crack in equation of motion

of cracked beam. It was done for transverse vibration of symmetric double edge cracked Euler

Bernoulli beam using Hu-Washizu-Barr variational principle. The partial differential equations

obtained from this model contains two unknowns, which are to be calculated by correlating ana-
4

lytical results with experimental results. These unknowns are particular for a given geometry and

boundary conditions of the beam. Shen and Pierre (1990) used the Christides and Barr (1984)

cracked beam model to find the natural modes of simply supported beam. They used Galerkin

technique to solve the partial differential equations and compared the results with finite element

solution. It was concluded that the exponential decay constant can be calculated using finite el-

ement results without the use of experimental results. Shen and Pierre (1994) extended the con-

tinuous cracked beam model of Christides and Barr (1984) for symmetrical double edge cracks

to beams with single edge crack. The exponential decay constants for simply supported and can-

tilever beams were calculated from finite element results and compared with experimental results.

Chondros et al. (1998), using Hu-Washizu-Barr variational principle and by eliminating the incon-

sistency between strain field and displacement field, developed continuous cracked beam model

similar to the model of Christides and Barr (1984). Later, they extended their work to find the first

natural frequency of a uniform beam with rectangular cross-section containing a breathing crack.

Carneiro and Inman (2002) have extended the Christides and Barr (1984) approach to Timoshenko

cracked beams. It was concluded that Timosenko beam theory for cracked beams is advantageous

when compared to Euler-Bernoulli beam theory for short beams. Behzad et al. (2005) developed

a new approach for dynamic analysis of cracked beams. They have considered a model similar to

that of Shen and Pierre (1994) and developed a theoretical basis similar to that of Chondros et al.

(1998) to calculate the constants in the partial differential equations of motion. The validity of

the obtained results has been confirmed by comparison with finite element results. Behzad et al.

(2008) has developed a new linear theory for bending stress-strain analysis of a beam with edge

crack. The required constants for this method were obtained from fracture mechanics. The results

were compared with finite element results. Behzad et al. (2008) used the proposed linear theory

to develop equation of motion for cracked beam and solved for frequencies and mode shapes of

a simply supported beam with a crack. The results are compared with those obtained from finite

element method. Heydari et al. (2014) extended the theory proposed by Behzad et al. (2008) for a
 5

cracked Timoshenko beam by considering shear deformations and rotational inertia. They solved

the equation of motion for forced vibration when subjected to a concentrated force.

Tedesco et al. (1999) conducted an FE analysis on dynamic performance of bridge girders

repaired with FRP plates on tension side of bridge girders. They concluded that the increase

in natural frequencies due to strengthening is insignificant. Stallings et al. (2000) have studied

the field dynamic performance of a bridge rehabilitated by external bonding of FRP plates to

the bridge girders. They indicated that the increase in natural frequencies is small. Hamed and

Rabinovitch (2005) developed an analytical model to investigate the dynamic behaviour of RC

beams strengthened with externally bonded composite materials. They developed equations of

motion and used Newmark’s time stepping method to obtain the desired solution. Chen and Liu

(2014) developed the equations of motion for a single edge cracked RC beam strengthened with

FRP. They used the cracked beam model proposed by Shen and Pierre (1994) and concluded that

the effect of steel bar’s restraints and FRP reinforcement on natural frequencies should not be

ignored.

1.3 Objectives

From the aforementioned literature, it may be noted that the studies related to the dynamic be-

haviour of FRP strengthened cracked concrete members are limited. Based on literature review,

the objectives of this work are described as follows:

• To develop free vibration equations of FRP strengthened cracked concrete members based

on dynamic equilibrium and compatibility of deformations.

• To carry out a numerical study in order to validate the analytical formulation.

• To investigate analytically, the dependence of first and higher order frequencies and mode

shapes on the position and depth of crack.

6

1.4 Organization of Thesis

This study is presented in five chapters. The chapters are organized as follows:

Chapter 1: In this chapter, an overview of this study is presented. The literature available in

this field along with the objectives of this study are also discussed in this chapter.

Chapter 2: The second chapter deals with the formulation of free vibration equation of mo-

tion for cracked concrete column. Mathematical solution to find the modal properties has been

presented. Numerical study has been performed to verify the results from the formulation.

Chapter 3: In this chapter, the free vibration equation for cracked reinforced concrete (RC)

column strengthened with FRP is developed. The equation is solved for natural frequencies and

mode shapes. Also numerical study has been performed to validate the analytical results.

Chapter 4: This chapter presents the analytical study of the effect of crack position on the

modal properties by using a perturbation formulation.

Chapter 5: In this chapter, a summary of the work done is presented along with the conclusions

drawn from the study. Scope for future work is also presented in this chapter.
Chapter 2

Free Vibration of Cantilever Column

with Crack

2.1 Brief Overview

In this chapter, the equation of motion for free vibration of a cantilever column is formulated. The

effects of mass at free end, presence of crack at random location and p-delta effects are considered

for this purpose. Rayleigh’s quotient is derived and a solution procedure is presented to find the

frequencies and mode shapes. A finite element simulation of damaged columns is also performed

to verify the analytical solutions.

2.2 Theoretical Background

2.2.1 Cracked Beam Theory

The displacement field for a cracked beam in flexural vibration is defined inBehzad et al. (2008)

as follows:
8



w = w(x, t), 






v = 0, (2.1)






u(x, z, t) = u0 (x, t) − zψ(x, t) + ∆(x, z, t)h(z)


where u, v, w are displacement components along x, y, z axes respectively; u0 (x, t) is longitu-

dional displacement of deviation line along x-axis and ψ(x, t) is the slope of plane sections below

the deviation line; ∆(x, z, t) accounts for the additional displacements of the plane sections above

the deviation line; h(z) is unit step function with a value of zero for z ≤0 and 1 for z > 0. The

function ∆(x, z, t) is assumed in Behzad et al. (2008) as follows:

|x − xc |
−α
∆(x, z, t) = ϕ(z, t)e d sgn(x − xc ) (2.2)

where α is dimensionless exponential decay rate and sgn is sign function. The function ϕ(z, t),

which is found by equating normal stress at crack faces to zero, can be written as

z !
d −β
ϕ(z, t) = u0,x (xc , t) − u0,x (xc , t)e d − zψ,x (xc , t) (2.3)
α

The strain field is extracted from the displacement field as

∂u
x =
∂x
z ! |x − xc |
−β −α
= u0,x − zψ,x − u0,x (xc , t) − u0,x (xc , t)e d − zψ,x (xc , t) h(z)e d (2.4)

The functions u0,x and ψ,x are obtained by equating the average shear strain in each cross section
 9

to zero as given here.

|x − xc | 

−α 
u0,x = zψ,x + k3 zh ψ,x (xc , t)e d 


(2.5)
|x − xc | 
−α


ψ,x = w,xx + k5 w,xx (xc , t)e d 


The equation of motion for free vibration of a cracked beam is as follows:

|x − xc |
 
∂2 −α
EI w,xx + kw,xx (xc , t)e d  + ρAw,xx = 0 (2.6)
∂x2

with

k3 k4 Azz h k4
k = k5 − + (k2 Ac z c z h (1 − k6 ) − Icy ) (2.7)
I I

In Equation 2.7 A is the cross section area of the beam, Ac is the area of cracked face and Ah

is the area of healthy part of cross section. I is the moment of inertia of cross section about

horizontal axis passing through centroid of cross section, Icy is the moment of inertia of crack

face about deviation line, z is the vertical coordinate of the centroid of cross section, z c is the

vertical coordinate of the centroid of cracked part of cross section and z h is the vertical coordinate

of the centroid of healthy part of cross section.

2.2.2 P-delta Effect

P-delta effects are additional over-turning moments that are generated as a result of axial load

on the column acting through the lever arm caused by the lateral deflections due to the action of

seismic forces. Figure 2.1 shows the distribution of secondary moments that are developed in the

column due to P-delta effect. The moment at any section is given by (in clock-wise sense)

Mx = P (δL − δx ) (2.8)
10

2.3 Equation of Motion

2.3.1 Geometry of Cantilever Column

A prismatic uniform cantilever RC column with an open edge crack is considered with a mass at

the free end. The following assumptions are made while considering the constitutive relations

• The effect of shear deformation is not considered.

• No slip exists between steel bars and concrete.

• Small strain theory is used.

In Figure 2.2, a and xc denote the crack depth and the position of crack from fixed end, respectively

and M denote the mass at the top of the column.

2.3.2 Displacement and Strain Fields

The displacement field for the flexural vibrations of cracked cantilever column can be considered

as 
w = w(x, z)








v=0 (2.9)


 

d 

u = u0 (x) − zψ(x, t) + ∆(x, z, t)U −a−z 

2

where u, v, w are displacement components along x, y, z axes respectively.ψ(x, t) is the slope of

the plane sections for the healthy part of the column. The valus of ψ(x, t) is equal to w,x (x, t), if

the shear stress effect is ignored. But the effect of shear stress is predominant near the crack tip

and it cannot be neglected. In Equation 2.9, ∆(x, z, t) accounts for the additional displacements

of the plane sections above the deviation line; u0 (x) is the longitudinal displacement along the

x-axis and is equal to zero if x-axis coincides with the level of neutral axis for a cross-section;
 11

 
d
U −a−z is unit step function, which is defined as
2


 d

d
  1
 z≥ −a
U −a−z = 2 (2.10)
2  d
 0 z< −a

2

The function ∆(x, z, t) is assumed in Behzad et al. (2008) as follows:

|x − xc |
−α
∆(x, z, t) = ϕ(z, t)e d sgn(x − xc ) (2.11)

To find out the function ϕ(z, t), the condition that the normal stress on the crack face to be zero is

used. Then the value of function ϕ(z, t) can be obtained as

d
ϕ(z, t) = [u0,x (xc ) − zψ,x (xc , t)] (2.12)
α

Now, the strain field can be extracted using Equations 2.9 and 2.12 as



x = u,x = u0,x (x) − zψ,x (x, t) − [u0,x (xc ) − zψ,x (xc , t)]f (x)U



  (2.13)
1 1 d 
γxz = (w,x + u,z ) = w,x − ψ − ψ,x (xc , t)f (x)sgn(x − xc )U 


2 2 α

where

|x − xc |
−α
f (x) = e d (2.14)

By considering the static equilibrium at each cross-section in x-direction, one can write

Z Z
σx dA = 0 ⇒ x dA = 0 (2.15)
A A
12

where A is the cross-section area of the column. By substituting Equation 2.13 in 2.15 yields

I1 u0,x (x) − I2 ψ,x (x, t) − [I3 u0,x (xc ) − I4 ψ,x (xc , t)]f (x) = 0 (2.16)

R R R R
where I1 = dA; I2 = zdA; I3 = U dA and I4 = zU dA. From Equation 2.15, considera-
A A A A

tion of static equilibrium at x = xc gives the following equation.

I5 u0,x (xc ) − I6 ψ,x (xc , t) = 0 (2.17)

R R
let us consider I5 = dA and I6 = zdA where Ah is the healthy cross-section area of the
Ah Ah

column. Combining Equations 2.16 and 2.17, the following result is obtained.

 
I2 I3 I6 I4
u0,x = ψ,x (xc , t) + − ψ,x (xc , t)f (x) (2.18)
I1 I1 I5 I1

The average shear strain in each cross section is assumed to be zero. Thus,

Z
γdA = 0 (2.19)
A

From Equations 2.13 and 2.19 one can write

I3
ψ,x = w,xx (x, t) + ψ,x (xc , t)f (x) (2.20)
I1

Substitution of x = xc in Equation 2.20 yields

 
I1
ψ,x (xc , t) = w,xx (xc , t) (2.21)
I1 − I3

On simplification (i.e., from Equations 2.13, 2.18, 2.20 and 2.21), the strain field for the cracked
 13

column is obtained as

cxx = g(z)w,xx (x, t) − h(z)w,xx (xc , t)f (x) (2.22)

where   
I2
g(z) = −z



I1 
     (2.23)
I2 I3 I3 I6 I4 I3 I6 I1
− −z − −z U

h(z) = 2 + 

I1 I1 I5 I1 I1 I5 I1 − I3

2.3.3 Energy Relations

2.3.3.1 Kinetic Energy

The kinetic energy associated with the concrete column is given by

Z Z Z
1 1 1
Tc = ρc ẇc2 dV + ρc v̇ 2 dV + ρc u̇2c dV
2 2 c 2
Vc Vc Vc
y 1 
∂w
2 y 1  2 2
∂ w
2
= ρc dxdydz + ρc z dxdydz (2.24)
2 ∂t 2 ∂x∂t
Vc Vc

The kinetic energy contribution of mass at the end of the column is given by

ZL  2
1 ∂w
TM = M δ(x − L)dx (2.25)
2 ∂t
0

where δ(x − L) is the Dirac delta function as given here.



 ∞
 if x = L
δ(x − L) = (2.26)

 0

elsewhere

From Equation 2.24 and neglecting the second term for slender column, the integration over the
14

cross-section yields

ZL  2
1 ∂w
Tc = ρc Ac dx (2.27)
2 ∂t
0

2.3.3.2 Strain Energy

The strain energy of the cracked concrete column is given by

y 1
Uc = σ c c dVc
2 xx xx
Vc
ZL 2 ZL !
∂2w ∂2w ∂2w
  
1
= Ec C1 dx + Ec C2 f (x) dx
2 ∂x2 ∂x2 ∂x2 (x=xc )
0 0
ZL !2
1 ∂2w
+ Ec C3 f (x) dx (2.28)
2 ∂x2 (x=xc )
0

where Z 
C1 = [g(z)]2 dA 






A 


Z 


C2 = g(z)h(z)dA (2.29)

A




Z 

2

C3 = [h(z)] dA 



A

2.3.3.3 Potential Energy

The contribution of energy due to the development of secondary moments in the column can be

considered. From Equation 2.8, the potential energy contribution due to P-delta effects is given by

ZL
1 ∂2w
Vpδ = Mx 2 dx
2 ∂x
0
ZL
1 ∂2w
= M g(w − w|(x=L) ) 2 dx (2.30)
2 ∂x
0
 15

2.3.4 Hamilton’s Equation

The Hamilton’s variational principle is given by

Zt1
(δT − δU − δV )dt = 0 (2.31)
t0

where T, U, V are total kinetic, strain and potential energies respectively, of the system and δ is

the variational operator. The total kinetic, strain and potential energies are given by

T = Tc + TM

U = Uc (2.32)

V = Vpδ

From Equations 2.25, 2.27, 2.28 and 2.30, one can write


ZL  
∂w ∂ ∂w ∂


δw δ(x − L) dx

δT = ρc Ac δw + M 

∂t ∂t ∂t ∂t 


0 



ZL " # ! 
∂2w ∂2 ∂2w ∂2


δU = Ec C1 2 (δw) + Ec C2 f (x) (δw) dx (2.33)
∂x ∂x2 (x=xc ) ∂x2 ∂x2 

0 


ZL 


2 2
  
1 ∂ w ∂ 

δV = M g (δw − δwL ) 2 + (w − wL ) 2 (δw) dx


2 ∂x ∂x



0

Here, w|(x=L) is denoted by wL . Substituting these equations in Hamilton’s variational principle

(Equation 2.31), integrating by parts and then by applying the boundary conditions, i.e,

δw = 0 at x = 0 and x = L

δw = 0 at t = t0 and t = t1
16

one can write

Zt1 ZL " !! #
∂2 ∂2w ∂2w ∂2w ∂2w
Ec C1 2 + Ec C2 f (x) + ρc Ac 2 + M g 2 δw dx dt = 0
∂x2 ∂x ∂x2 (x=xc ) ∂t ∂x
t0 0

(2.34)

By the argument that t0 and t1 are arbitrary and δw is arbitrary, Equation 2.34 reduces to

!!
∂2 ∂2w ∂2w ∂2w ∂2w
Ec C1 2 + Ec C2 f (x) + ρc Ac + M g =0 (2.35)
∂x2 ∂x ∂x2 (x=xc ) ∂t2 ∂x2

Equation 2.35 represents the free vibration equation of cracked concrete column with boundary

conditions given by

! !
x=L
∂2w ∂2w 1 ∂2w ∂
Ec C1 2 + Ec C2 f (x) + M g(w − wL ) (δw) = 0 (2.36a)
∂x ∂x2 (x=xc ) 2 ∂x2 ∂x x=0
! !
x=L
∂ ∂2w ∂2w 1 ∂2w
Ec C1 + Ec C2 f (x) + M g(w − w L ) δw = 0 (2.36b)
∂x ∂x2 ∂x2 (x=xc ) 2 ∂x2 x=0

2.3.5 Calculation of Exponential Decay Rate α

The exponential decay rate, α is calculated by using the theory developed by Behzad et al.(2005).

The additional rotation, θ∗ due to the crack when a moment ‘M ’ is applied at the free end of

cantilever is given by

1 − ν 2 M d2 a
θ∗ = Πf (2.37)
Ec 2I 2 d

where

a x h  a i2
f = a F dAci
d d
Aci
a a  a 2  a 3  a 4
F = 1.12 − 1.4 + 7.33 − 13.1 + 14 (2.38)
d d d d d
 17

and Aci is the area of the cracked part of the column. The internal resisting moment, ‘M ’ at a

section when an external loading is applied to the beam is given by

Z
M= Exx zdydz (2.39)
A

Referring to Equation 2.22, the moment-curvature relationship is obtained as,

d2 w
 
M A2
= 1 + f (x) (2.40)
dx2 Ec A1 A1 − A2

where

Z  
I2
A1 = z− zdA
I1
A
Z    
I3 I6 I4 I2 I3 I6 I4
A2 = − − + − − z U zdA (2.41)
I1 I5 I1 I1 I1 I5 I1
A

By integrating the above moment curvature equation and substituting boundary conditions for

cantilever column, one can write

(x − xc ) xc
  

 M  kd α kd −α

 x+ e d − e d if x ≤ xc
 Ec A1 α α


θ(x) = 
(x − xc ) xc
 (2.42)
 M kd kd kd

 −α −α
d e d +2 
 Ec A1 x − α e − if x > xc

 
α α


where

A2
k= (2.43)
A1 − A2

For uncracked healthy cantilever column, the rotation at a section is given by

ML
θu (x) = (2.44)
Ec I
18

From equations of slopes of cracked and healthy columns, the additional rotation due to crack is

given by

(L − xc ) xc
 
M  kd −α kd −α kd ML
θ∗ = θ(L) − θu (L) = L− e d − e d +2 − (2.45)
Ec A1 α α α Ec I

The value of exponential decay constant α is found out by equating the right hand sides of Equa-

tions 2.37 and 2.45. The value of α is a function of the crack depth, crack position and also the

slenderness ratio of the column. To show the variation of exponential decay constant with crack

depth ratio, the parameters in Table 2.1 are considered. Crack position is fixed at 0.1L and the

value of α is plotted against the crack depth ratio as shown in Figure 2.3.

2.4 Solution using Rayleigh’s Quotient

2.4.1 Formulation

Equation 2.35 is hard to solve by following conventional mathematical methods. Hence, Rayleigh-

Ritz method is used. From Equations 2.25 and 2.27, the total kinetic energy of system is given

by

ZL  2  2
1 ∂w 1 ∂w
T = ρc Ac dx + M (2.46)
2 ∂t 2 ∂t x=L
0

From Equations 2.28 and 2.30, the total potential energy of system is given by

ZL 2 ZL !
∂2w ∂2w ∂2w
  
1
U= Ec C1 dx + Ec C2 f (x)dx
2 ∂x2 ∂x2 ∂x2 (x=xc )
0 0
ZL !2 ZL
1 ∂2w 2 1 ∂2w
+ Ec C3 [f (x)] dx + M g(w − w|(x=L) ) 2 dx (2.47)
2 ∂x2 (x=xc ) 2 ∂x
0 0
 19

By using the method of separation of variables, the solution can be expressed as:

w = W (x)η(t) (2.48)

For free vibration case, the time function can be expressed as

η(t) = eiωt (2.49)

Substituting Equations 2.48 and 2.49 in Equations 2.46 and 2.47, and by equating the magnitudes

of kinetic and potential energies, the Rayleighs quotient R(W ) is obtained as

" 2
d2 W
 2  2
RL RL
 
d W d W (xc )
Ec C1 dx + 2 Ec C2 f (x)dx
0 dx2 0 dx2 dx2
 2 2 #
RL d W (xc ) 2 RL d2 W
+ Ec C3 [f (x)] dx + M g(W − W (L)) 2 dx
0 dx2 0 dx
R(W ) = ω 2 = !
RL
ρc Ac (W )2 dx + M (W (L))2
0
(2.50)

The solution to free vibration problem can be found by assuming trial functions and by substi-

tuting in Equation 2.50. The trial functions assumed must satisfy the geometric boundary condi-

tions. Those are called admissible functions and the set of admissible functions is called generating

set. The trial solution will be of the form

∞
X
W (x) = an φn (x) (2.51)
i=1

where φn (x) are admissible functions and an are unknown coefficients. Some examples of gener-

ating sets for cantilever column are as follows:

20


∞
an xn

 P




 n=2

 P∞ n  nπx o
a 1 − cos


 n



 n=1 L
 ∞ ∞ n  mπx o
W (x) = anm xn 1 − cos (2.52)
P P


 n=0 m=1 L

 ∞ ∞  mπx 
anm xn sin

 P P




 n=1 m=1 L
 ∞ ∞
  nπx  n  mπx o
 P P

 anm sin 1 − cos
n=1 m=1 L L

The convergence and accuracy of solution depends on the generating set used in the Rayleigh’s

quotient (Equation 2.50). The Rayleigh’s quotient contains the terms W,xx (x) and W,xx (xc ) and

for larger crack depths, the solution has large derivatives near the crack tip. So, extracting the val-

ues of W,xx (x) and W,xx (xc ) by direct double derivation of W (x) will lead to error in the results.

In order to reduce the error, the function W (x) is assumed piece-wise by dividing the domain into

small pieces and obtaining the solution by converting the problem in to eigen value problem using

Rayleigh’s quotient and the assumed trial solution. The solution method is explained and derived

in the subsequent section.

2.4.2 Solution Procedure

To find the solution, the concepts of finite element method and Rayleigh-Ritz solution method are

used in combination. The domain is divided in to a number of elements and for each element, the

displacement is assumed as follows:

4
X
W = Ni δi = [N ][δ] (2.53)
i=1
 21

where [δ] represents a column vector comprising of displacements of corresponding element and

Ni (i = 1 to 4) represents Hermitian polynomial shape functions that are given by


x2 x3 
N1 = 1 − 3 2 + 2 3

L L 



2 3

x x 

N2 =x−2 + 2 


L L (2.54)
x2 x3 
=3 2 −2 3

N3 

L L 



x 2 x3 


N4 =− + 2 

L L

The Rayleigh’s quotient (Equation 2.50) can also be represented as follows:

N1 + N2 + N3 + N4
R(W ) = ω 2 = (2.55)
D1 + D2

where
ZL

2
d2 W
 

N 1 = E c C1 2
dx 

dx




0 


ZL  2  2  

d W d W (xc )



N 2 = 2 Ec C2 2 2
f (x)dx 

dx dx 


0




L 2


2
Z  
d W (xc )

2

N 3 = Ec C3 [f (x)] dx


2

dx 


0
(2.56)
ZL
d2 W



N4 = M g(W − W (L)) dx


dx2




0 


ZL




ρc Ac (W )2 dx

D1 = 




0




ZL




M (W )2 δ(x − L)dx

D2 = 



0

Let us assume that [∆] represents the global displacement field, [δ] represents elemental displace-

ment field and [δxc ] represents the displacement field of element consisting the crack. Minimising
22

the Rayleigh’s quotient (Equation 2.55) with respect to [∆], one can write

 
∂R 1 ∂N 1 ∂N 2 ∂N 2 ∂N 3 ∂N 4
= + + + +
∂[∆] (D1 + D2) ∂[δ] ∂[δ] ∂[δxc ] ∂[δxc ] ∂[δ]
 
∂D1 ∂D2
−ω 2 + =0 (2.57)
∂[δ] ∂[δ]

Now from Equations 2.54 and 2.56, the following expressions are obtained.


ZLi
d2 [N ]T d2 [N ]
  
∂N 1 

= 2Ec C1 2 2
[δ]dx = [k 1 ][δ] = [K 1 ][∆] 

∂[δ] dx dx



0




L
Zi  2 

T 2

∂N 2 d [N ] d [Nxc ]



= 2Ec C2 f (x) [δ xc ]dx = [k 2 ][δ xc ] = [K 2 ][∆] 

∂[δ] dx2 dx2 


0




ZLi  2 T 2
 

∂N 2 d [Nxc ] d [N ]


T T
= 2Ec C2 f (x) [δ]dx = [k ] [δ] = [K ] [∆]


2 2 2 2 
∂[δxc ] dx dx 


0 



ZLi  2 T 2
 

∂N 3 d [N ] d [N ]

2 xc xc
= 2Ec C3 [f (x)] [δxc ]dx = [k ][δ
3 xc ] = [K 3 ][∆] (2.58)
∂[δxc ] dx2 dx2 

0 


L

i 
2
Z   
∂N 4 T d [N ]


= 2M g [N ] [δ]dx = [k ][δ] = [K ][∆]

4 4

∂[δ] dx2




0 


L

Zi 

∂D1 T


= 2ρc Ac [N ] [N ][δ]dx = [m1 ][δ] = [M1 ][∆] 

∂[δ]




0 


ZLi



∂D2 T


= 2M δ(x − L)[N ] [N ][δ]dx = [m2 ][δ] = [M2 ][∆] 

∂[δ]



0

where Li is the length of individual element. From Equations 2.57 and 2.58, the eigen problem

can be obtained as

 
[K] − ω 2 [M ] [∆] = 0 (2.59)
 23

where

[K] = [K1] + [K2] + [K2]T + [K3] + [K4] (2.60)

[M ] = [M 1] + [M 2] (2.61)

2.4.3 Results and Discussion

To find the mode shapes and frequencies of cracked cantilever column, Rayleigh Ritz equation

(Equation 2.50) is solved using MATLAB. The geometry of cantilever column with open edge

crack is considered as shown in Figure 2.2. In the numerical examples, geometric data of column

and material properties as defined in Table 2.1 are used. The position and depth of the crack are

taken as parameters to find their effect on natural frequencies and mode shapes of the column.

Table 2.1: Parameters of cantilever column

Length of column, L 20m

d 3.9m
b 4.05m
r 4.5m
Density of Concrete 2450 Kg/m3
Mass at the tip of column 1000 Kg
Modulus of elasticity of Concrete 2.8×1010 N/m2

2.4.3.1 Dependence of Modal Properties on Crack Parameters

The mode shapes and natural frequencies of the cracked column depends on the crack location and

severity of damage (crack depth). To understand this dependence, the first natural frequency ratio

(ω/ω0 ) is studied by varying crack position and crack depth (where ω is the first natural frequency

of cracked column and ω0 is the first natural frequency of uncracked column). Also, the mode

shapes for different damage cases are studied by superimposing them with the mode shapes of
24

uncracked column.

Effect of Crack Depth on First Natural Frequency

In Figure 2.4, the variation of first frequency ratio with respect to the crack depth ratio for

crack position ratios of 0.1, 0.25 and 0.4 is shown. It can be observed from this figure that the

fundamental frequency decreases with crack depth ratio. The rate of reduction is less for cracks

which are away from the fixed end.

Effect of Crack Position on First Natural Frequency

The variation of first frequency ratio with respect to the position of crack for different crack

depths of 0.1, 0.15, 0.2 and 0.25 is shown in Figure 2.5. From this figure, it can be concluded that

the reduction in frequency is more for those cases where the crack is closer to fixed end. As crack

moves away from fixed end, frequency ratio approaches unity. The reduction in frequency is more

for higher crack depths.

Effect of Crack Depth on Mode Shapes

Figures 2.6 and 2.7 show the normalized plot of first two mode shapes of cracked column for

three different crack depths (a/d = 0.2, 0.3 and 0.4) for crack position ratio of 0.25. It can be seen

from these figures that the difference between the cracked and healthy mode shapes increases as

crack depth increases.

Effect of Crack Position on Mode Shapes

The first and second mode shapes of column are plotted for different crack positions of 0.1,

0.2 and 0.3 for crack depth ratio of 0.25 along with respective mode shapes of uncracked column

(Figures 2.8 - 2.9). From these figures, it can be concluded that the variation of mode shapes

of cracked and uncracked column depends upon the position of crack and the variation is not

monotonic in relation to the crack position from the fixed end.

 25

2.5 Finite Element Modelling

Finite element modelling of the cracked column was done to verify the theoretical findings using

ABAQUS/CAE (version 6.10), a general purpose finite element software. Three dimensional de-

formable solid elements were used to model the column. In this study, three different columns

with different crack positions and crack depths were modelled for finite element analysis. The de-

tails of geometric and material properties, and various analysis steps are described in the following

subsections.

2.5.1 Geometry and Material Properties

The geometric and material properties of the finite element models are considered as listed in

Table 2.1. To model the column, a three dimensional deformable solid extrusion model is used.

The cross section of the column is drawn as shown in Figure 2.2 and is extruded to the whole

length of the column. The damage is modelled as a transverse surface crack by a V-shaped groove

through out the whole breadth of the column with depth equal to crack depth and by positioning

the vertex of v-shaped groove at the crack position for a particular damage scenario.

2.5.2 Boundary Conditions and Meshing

The boundary conditions were applied to simulate the fixed end condition of cantilever column

using ‘encastrate’ condition. The encastrate condition is applied to the left surface as shown in

Figure 2.10. In finite element analysis, the accuracy and convergence capability of results depend

on element type, shape and number of elements. For a three-dimensional model with geometric

discontinuities, a tetrahedral element is preferred to a hexahedral element. Also greater the number

of elements in the mesh, the more accurate are the results. As the mesh density increases, the

results converge to a unique solution and computational time reuired for analysis also increases.

The mesh density around the crack also influences the result. In this connection, a compatible and

converging ten-noded quadratic tetrahedron element (C3D10) with six degrees of freedom (three
26

displacement and three rotational) at each node was chosen for current analysis as shown in Figure

2.11. Figure 2.12(a) shows the Abaqus model of the cracked cantilever column and Figure 2.12(b)

shows the details of finer mesh adopted near the cracked region.

2.5.3 Analysis Step and Mesh Size

In the ‘step’ module, an analysis step is added for calculating modal properties. The procedure

adopted is linear perturbation frequency analysis with eigensolver ‘Lanczos’. A convergence study

is done varying the mesh density. Table 2.2 shows the mesh size details through out the column and

also around the cracked region. Also specified is the total number of elements in each mesh size.

Figure 2.13 shows the convergence of fundamental frequency values with the number of elements.

It has been found that the frequency is converging for Mesh 5. The percentage difference in the

values of frequencies between Mesh 5 and Mesh 6 was found to be 0.13%. So, Mesh 5 is adopted

for the numerical study.

Table 2.2: Details of mesh sizes used in convergence study

Overall Mesh Mesh Size

No. of Elements
Size (in m) near Crack (in m)
Mesh 1 5 0.5 193
Mesh 2 2 0.2 1888
Mesh 3 1 0.1 9106
Mesh 4 0.5 0.05 51807
Mesh 5 0.4 0.04 93353
Mesh 6 0.3 0.03 210298

To validate the results of modal properties obtained from the theoretical formulation, columns

with different damage scenarios were modelled in ABAQUS and those results were compared with

theoretical modal properties.

2.5.4 Damage Scenarios

Three damage scenarios were modelled in ABAQUS, which are described as follows:

• Case I: In this damage case, the crack is considered to be present at crack position ratio
 27

(CPR) xc /L = 0.1 from the fixed end of column and crack depth ratio (CDR) a/d = 0.3.

• Case II: This damage case is simulated by considering CPR of 0.2 and CDR of 0.3.

• Case III: In this damage case, a CPR of 0.3 and CDR of 0.4 are considered.

The values of CDR in the damage scenarios were limited to a value of 0.4 as the theoretical

formulation is limited to a CDR of 0.6. Also, high CDR values are practically irrelevant.

2.6 Comparison of Theoretical and Numerical Modal Properties

The modal properties chosen for comparison were the fundamental frequency and mode shape.

Table 2.3 compares the analytical and finite element results of first frequency ratios for all the three

damage cases. Also the percentage difference with respect to ABAQUS result is also presented.

It has been observed that the results are in good agreement. Figure 2.14 shows the comparison

between numerical and analytical normalized fundamental mode shapes for all three damage cases.

A good agreement between those two can be observed.

Table 2.3: Comparison of fundamental frequencies from analytical and finite element results

First Frequency Ratio Percentage Difference

CPR CDR
Theoretical ABAQUS (w.r.t. ABAQUS)
Case I 0.1 0.3 0.8410 0.8251 1.927
Case II 0.2 0.3 0.8597 0.8630 0.382
Case III 0.3 0.4 0.8341 0.8291 0.603

2.7 Summary

In this chapter, the equation of motion for free vibration of a cantilever column is formulated by

considering the effects of mass at free end, presence of crack at random location and considering

P-delta effects. Rayleigh’s quotient is derived for the same and a solution procedure is presented

to find the frequencies and mode shapes. A finite element simulation of damaged columns is also
28

performed to verify the analytical solution. The crack position and depth are taken as parame-

ters and their effect on frequencies and mode shapes is found. The study shows that as the depth

of crack increases, the variation between the frequencies of cracked and uncracked columns in-

creases. Also, as the crack position moves away from the fixed end, the variation decreases. It

can be concluded that the changes in the fundamental frequency is not sensitive to small crack

depths and for crack positions away from the fixed end. From mode shapes, it is observed that

as the crack depth ratio increases, the variation between the cracked and uncracked mode shapes

increases. Further, the variation of mode shapes of cracked and uncracked column depends upon

the position of crack and the variation is not monotonic in relation to crack position from fixed

end. By comparing analytical with finite element results, a good agreement between the two has

been observed.

(a) (b)

Figure 2.1: Distribution of secondary moments due to P-delta effect: (a) Configuration of can-
tilever column; (b) Distribution of secondary moments
 29

(a) Elevation (b) Section 1-1

Figure 2.2: Geometry of cantilever column with an open-edge crack

Figure 2.3: Variation of exponential decay constant with crack depth ratio
30

Figure 2.4: Effect of crack depth ratio on first frequency of cantilever column for different crack
positions of xc /L = 0.1, 0.25 and 0.4

Figure 2.5: Effect of crack position on fundamental frequency of cantilever column for different
crack depth ratios of a/d = 0.1, 0.15 , 0.2 and 0.25
 31

(a) Crack depth ratio a/d = 0.2 (b) Crack depth ratio a/d = 0.3

(c) Crack depth ratio a/d = 0.4

Figure 2.6: Comparison of first mode shapes for uncracked and cracked cantilever column for
crack position xc /L = 0.25: (a) Crack depth ratio a/d = 0.2; (b) Crack depth ratio a/d = 0.3; (c)
Crack depth ratio a/d = 0.4
32

(a) Crack depth ratio a/d = 0.2 (b) Crack depth ratio a/d = 0.3

(c) Crack depth ratio a/d = 0.4

Figure 2.7: Comparison of second mode shapes for uncracked and cracked cantilever column for
crack position xc /L = 0.25: (a) Crack depth ratio a/d = 0.2; (b) Crack depth ratio a/d = 0.3; (c)
Crack depth ratio a/d = 0.4
 33

(a) Crack position xc /L = 0.1 (b) Crack position xc /L = 0.2

(c) Crack position xc /L = 0.3

Figure 2.8: Comparison of first mode shapes for uncracked and cracked cantilever column for
crack depth ratio a/d = 0.25: (a) Crack position xc /L = 0.1; (b) Crack position xc /L = 0.2; (c)
Crack position xc /L = 0.3
34

(a) Crack position xc /L = 0.1 (b) Crack position xc /L = 0.2

(c) Crack position xc /L = 0.3

Figure 2.9: Comparison of second mode shapes for uncracked and cracked cantilever column for
crack depth ratio a/d = 0.25: (a) Crack position xc /L = 0.1; (b) Crack position xc /L = 0.2; (c)
Crack position xc /L = 0.3
 35

Figure 2.10: ABAQUS model of cracked concrete column

Figure 2.13: Convergence of finite element result

36

Figure 2.11: A typical C3D10 element in FE analysis

(a) Full view of the model (b) Close up of cracked region

Figure 2.12: ABAQUS model of cracked cantilever column

 37

(a) Case I (b) Case II

(c) Case III

Figure 2.14: Comparison of analytical and finite element fundamental mode shapes : (a) Case I;
(b) Case II; (c) Case III
38
Chapter 3

Free Vibration of Cracked Column with

FRP and Steel Reinforcement

3.1 Brief Overview

In this chapter, equation of motion for free vibration of cracked column considering the restraints

of steel reinforcement and FRP confinement is formulated. The equation of motion as formu-

lated in Chapter 2 is modified to include the compatibility of deformations (i) between the steel

reinforcement and concrete at the level of steel-concrete interface, and (ii) between the FRP and

concrete at the level of FRP-concrete interface are used. The effect of steel reinforcement and

FRP confinement on the frequencies and mode shapes of cantilever column is also studied. A

finite element simulation of cracked RC column with FRP confinement is also performed to verify

the analytical results.

3.2 Compatibility of Deformations

The compatibility of deformations is introduced by equating the lateral and longitudinal deforma-

tions (i) between steel reinforcement and concrete at the level of steel-concrete interface, and (ii)

between the FRP layer and concrete at the level of FRP-concrete interface. These compatibilities
40

can be expressed as

wc (x) = wf (x) = ws (x) (3.1)


uc (x, zs ) = us (x, zs ) 


(3.2)

uc (x, −zs ) = us (x, −zs )



uc (x, d/2) = uf (x, d/2) 


(3.3)

uc (x, −d/2) = uf (x, −d/2)


Equation 3.1 gives the compatibility of lateral deformations between the concrete, steel and FRP

while Equation 3.2 gives the compatibility of longitudinal deformations between concrete and

steel. Equation 3.3 gives the compatibility of longitudinal deformations between the concrete and

FRP. In Equations 3.1-3.3, wc , ws , wf represent lateral deformations of concrete, steel reinforce-

ment and FRP wrap, respectively; uc , us , uf represent longitudinal deformations of concrete, steel

reinforcement and FRP wrap, respectively; and zs represent the z coordinate of steel reinforcement

layer.

3.3 Strain Displacement Relations

From Equations 3.1 - 3.3, the compatibility of strains at different levels in the cross-section can be

represented as, 
cxx (x, zs ) = sxx (x, zs ) 









c s
xx (x, −zs ) = xx (x, −zs ) 


(3.4)

cxx (x, d/2) fxx (x, d/2)

= 







c f 
xx (x, −d/2) = xx (x, −d/2)


In Equation 3.4, the superscripts ‘c’, ‘s’ and ‘f ’ are used to represent concrete, steel and FRP,

respectively and xx represent strain in x-direction. From Equation 2.22, the strain displacement
 41

relations for concrete, steel and FRP at different levels in cross-section are given by


sxx (x, zs ) = g(zs )w,xx (x, t) + h(zs )f (x)w,xx (xc , t)










sxx (x, −zs ) = g(−zs )w,xx (x, t) + h(−zs )f (x)w,xx (xc , t)



(3.5)

fxx (x, d/2)

= g(d/2)w,xx (x, t) + h(d/2)f (x)w,xx (xc , t) 







f 
xx (x, −d/2) = g(−d/2)w,xx (x, t) + h(−d/2)f (x)w,xx (xc , t)


3.4 Energy Relations

The strain energies associated with the steel reinforcement and FRP confinement are obtained by

assuming that the variation of strain along the thickness is negligible.

3.4.1 Kinetic Energy

The kinetic energy associated with the concrete column is given by Equation 2.24. Similarly, the

kinetic energy relations for steel and FRP are obtained by replacing the subscript in the terms of

Equation 2.24 to ‘s’ and ‘f ’, respectively. The kinetic energy contribution of mass at the end of

the column is given by Equation 2.25. By considering the integration over the cross-section, the

kinetic energies of concrete, steel and FRP are obtained as follows:


ZL  2
1 ∂w



Tc = ρc Ac dx 

2 ∂t 


0 



ZL  2  
1 ∂w

Ts = ρs As dx (3.6)
2 ∂t 

0 


L

Z  2  
1 ∂w 

Tf = ρf Af dx

2 ∂t



0

In Equation 3.6, ‘ρ’ represent the mass density and ‘A’ represent the area of cross section.
42

3.4.2 Strain Energy

From Equation 2.28, the strain energy contribution of concrete column is given by,

y 1
Uc = σ c c dVc
2 xx xx
Vc
ZL 2 ZL !
∂2w ∂2w ∂2w
  
1
= Ec C1 dx + Ec C2 f (x) dx
2 ∂x2 ∂x2 ∂x2 (x=xc )
0 0
ZL !2
1 ∂2w
+ Ec C3 f (x) dx (3.7)
2 ∂x2 (x=xc )
0

In Equation 3.7, the values of C1 , C2 and C3 are given by Equation 2.29. The strain energy

associated with steel reinforcement is given by (referring to strain-displacement relations from

Equation 3.5)

y 1
Us = σ s s dVs
2 xx xx
Vs
ZL 2 ZL !
∂2w ∂2w ∂2w
  
1
= Es S1 dx + Es S2 f (x) dx
2 ∂x2 ∂x2 ∂x2 (x=xc )
0 0
ZL !2
1 ∂2w
+ Es S3 f (x) dx (3.8)
2 ∂x2 (x=xc )
0

where the values of S1 , S2 and S3 are given by


2 2 
S1 = As1 [g(zs )] + As2 [g(−zs )] 






S2 = As1 [g(zs )h(zs )] + As2 [g(−zs )h(−zs )] (3.9)




2 2

S3 = As1 [h(zs )] + As2 [h(−zs )]



The strain energy associated with FRP confinement is given by (referring to strain-displacement
 43

relations from equation 3.5)

y 1
Uf = σ f f dVf
2 xx xx
Vsf

ZL 2 ZL !
∂2w ∂2w ∂2w
  
1
= Ef F1 dx + Ef F2 f (x) dx
2 ∂x2 ∂x2 ∂x2 (x=xc )
0 0
ZL !2
1 ∂2w
+ Ef F3 f (x) dx (3.10)
2 ∂x2 (x=xc )
0

where the values of F1 , F2 and F3 are given by


F1 = Af 1 [g(d/2)]2 + Af 2 [g(−d/2)]2








F2 = Af 1 [g(d/2)h(d/2)] + Af 2 [g(−d/2)h(−d/2)] (3.11)




2 2

F3 = Af 1 [h(d/2)] + Af 2 [h(−d/2)]



3.4.3 Potential Energy

The contribution of energy due to the development of secondary moments in the column can be

considered. From Equation 2.8, the potential energy contribution due to P-delta effects is given by

ZL
1 ∂2w
Vpδ = Mx 2 dx
2 ∂x
0
ZL
1 ∂2w
= M g(w − w|(x=L) ) 2 dx (3.12)
2 ∂x
0

3.5 Solution using Rayleigh’s Quotient

As it is evident from the solution procedure of cracked column without FRP and steel constraints,

a direct solution of equation of motion is not possible. So, a similar solution procedure using

Rayleigh’s quotient is adopted in this case also. The total kinetic energy of the system from
44

Equations 2.25 and 3.6 is given by

ZL  2 ZL  2
1 ∂w 1 ∂w
T = ρA dx + M δ(x − L)dx (3.13)
2 ∂t 2 ∂t
0 0

where,

ρA = ρc Ac + ρs As + ρf Af (3.14)

the total potential energy of the system from Equations 3.7 , 3.8 ,3.10 and 3.12 is given by

ZL 2 ZL !
∂2w ∂2w ∂2w
  
1
U= EB1 dx + EB2 f (x)dx
2 ∂x2 ∂x2 ∂x2 (x=xc )
0 0
ZL !2 ZL
1 ∂2w 2 1 ∂2w
+ EB3 [f (x)] dx + M g(w − w|(x=L) ) 2 dx (3.15)
2 ∂x2 (x=xc ) 2 ∂x
0 0

By defining, 

EB1 =Ec C1 + Es S1 + Es S1 






EB2 =Ec C2 + Es S2 + Es S2  (3.16)





EB3 =Ec C3 + Es S3 + Es S3 


Rayleigh’s quotient, R(W ) for the cracked column is obtained as

" 2
d2 W
 2  2
RL RL
 
d W d W (xc )
EB1 dx + 2 EB2 f (x)dx
0 dx2 0 dx2 dx2
 2 2 #
RL d W (xc ) 2 RL d2 W
+ EB3 [f (x)] dx + M g(W − W (L)) 2 dx
0 dx2 0 dx
R(W ) = ω 2 = ! (3.17)
RL 2 2
ρA (W ) dx + M (W (L))
0

It may be noted that the Rayleigh’s quotients for both the cases while considering and ignoring

the effects of steel and FRP confinements (Equations 2.50 and 3.17) are similar. The difference

is observed in the values of coefficients. The same procedure as it is explained in Section 2.4.2 is
 45

used to find the mode shapes and frequencies of the column.

3.5.1 Results and Discussion

To study the restraint effects of steel reinforcement and FRP confinements on mode shapes and

frequencies of cracked column, a study has been performed by considering the geometrical and

material properties of column as listed in Table 3.1. The Rayleigh-Ritz equation is solved using

MATLAB. The results are presented here. The following nomenclature as represented in Table

3.2 is used to while presenting the results.

Effect of Steel Bar’s Restraints on Natural Frequency of Cracked RC Column :

Figures 3.1 and 3.2 show the variation of fundamental frequency and second frequency ratios, re-

spectively for both concrete (ωC /ωU ) and RC columns (ωCR /ωU R ) as a function of crack position

ratio (CPR) for different crack depth ratios (CDR). It has been observed from the graphs that the

frequency ratio for RC column is on the higher side when compared with that of concrete column.

Also, Table 3.3 presents the values of frequency ratios (ωC /ωU ) for the first three frequencies in

terms of crack depth at different regions for concrete column and Table 3.4 presents frequency

ratios (ωCR /ωU R ) for the first three frequencies in terms of crack depth at different regions for RC

column.

Effect of FRP on Natural Frequency of Cracked RC Column Strengthened with FRP :

In Figure 3.3, the fundamental frequency ratios (ωCRF /ωC ) and in Figure 3.4, the second fre-

quency ratios (ωCRF /ωC ) for columns RCF1, RCF2 and RCF3 are shown as a function of crack

position for different crack depths. It can be observed from these figures that the effect of FRP

strengthening on the natural frequencies is significant. The increase in frequencies is more for

more layers of FRP. The effect of FRP strengthening on frequencies is more in the case of larger

crack depths. For fundamental frequency ratios (Figure 3.1), this effect is more for cracks closer

to the fixed end. Tables 3.5 - 3.7 presents the first three frequency ratios (ωCRF /ωU R ) in terms of

crack depth at different crack positions for RC columns strengthened with FRP.
46

Effect of Steel Bar’s Restraints on Mode Shapes of Cracked RC Column Strengthened with

FRP:

The normalised plots of damaged mode shapes for concrete, RC and RC column strengthened with

FRP are shown in Figure 3.5(a) (first mode shape for CDR 0.3 and CPR 0.1), Figure 3.7(a) (second

mode shape for CDR 0.3 and CPR 0.1), Figure 3.6(a) (first mode shape for CDR 0.4 and CPR 0.1)

and Figure 3.8(a) (second mode shape for CDR 0.4 and CPR 0.1). The respective difference in

mode shapes are shown in Figures 3.5(b) - 3.8(b). It can be observed from these figures that the

difference between the concrete and RC column mode shapes is negligible.

Effect of Steel Bar’s Restraints on Mode Shapes of Cracked RC Column :

Figures 3.5(a) - 3.8(a) shows the normalised plots of damaged mode shapes for concrete, RC

and RC column strengthened with FRP. The difference in mode shapes between RCF1, concrete

column are shown in Figures 3.5(c) - 3.8(c) while Figures 3.5(d) - 3.8(d) and Figures 3.5(e) -

3.8(e) shows the difference in mode shapes between RCF3, concrete column and RCF2, concrete

column respectively. It can be observed from the plots that the neither the inclusion of FRP nor

increasing the number of FRP layers has a significant effect in the mode shapes.

3.6 Numerical Study

Numerical studies were conducted on cracked RC columns strengthened with FRP using finite

element explicit analysis using a commercial package, ABAQUS/CAE (version 6.10). Finite ele-

ment models of concrete, steel and FRP were developed to study the modal properties of cracked

column. The interactions between the steel reinforcement and concrete, and between the FRP

and concrete were defined. The results from numerical study were compared with those from the

analytical study.
 47

3.6.1 Geometry and Material Modelling

The concrete column, steel reinforcement and FRP cover were modelled as three different parts

and then assembled by providing constraints between them. The concrete column was modelled

as three dimensional solid extrusion model and the crack is modelled as a V-notch as described

in Chapter 2. The steel reinforcement is modelled as three dimensional deformable wire element

(truss element). The reinforcement bar is partitioned into three parts depending on the location of

the crack. These partitions were made in order to facilitate the generation of different meshes for

different locations. The FRP cover is modelled as three dimensional extrusion shell element. The

geometry of FRP cover is drawn and extruded to the whole length of column. The face of the FRP

cover that is in contact with the cracked surface of concrete column is partitioned into three parts

matching the crack position. These partitions were again made to facilitate mesh generation and

also to ease the process of defining interaction properties. The FRP shell model with partitioned

face is shown in Figure 3.9. The geometric and material properties of the finite element models

are considered as listed in Table 3.1. The boundary conditions as described in chapter 2 were

considered here also.

3.6.2 Assembly

The concrete, FRP and steel parts are imported from ‘part’ module. ‘Linear pattern’ tool is used

to create the required alignment of reinforcement. Then, datum points are created on the face

of the concrete column corresponding to the location of reinforcement bars. The constraint of

‘coincident point’ is used to assemble concrete and steel bars. The same constraint ‘coincident

point’ is used on suitable points to assemble concrete part and FRP cover. The assembled RC

column with FRP is shown in Figure 3.10. The details of reinforcement inside the concrete part

are shown in Figure 3.11.

48

3.6.3 Interaction

To define the interaction between concrete and steel bars, the constraint ‘embed’ is used. The

steel bars are selected as embedded region and the concrete part is selected as host region. Tie

constraint is used to define the interaction between FRP cover and concrete column. The FRP part

is selected as slave region and concrete part is selected as master region.

3.6.4 Meshing

A ten-noded quadratic tetrahedron element (C3D10) with six degrees of freedom (three displace-

ment and three rotational) at each node was chosen for meshing of concrete column. The mesh

density around the crack was increased for the solution to converge faster. For meshing steel bars,

a two-noded three dimensional truss element(T3D2) was used. STRI65 which is a six-noded tri-

angular thin shell element with five degrees of freedom per node was used to mesh the FRP cover.

Convergence study is done to select appropriate mesh sizes for different parts. A mesh size of

0.4m for concrete column, 0.04m for crack edges, 0.2m for steel bars and 0.4m for FRP cover are

found to be giving converged results.

3.7 Comparison of Results

To compare and validate the analytical solution, 80 column models with different crack positions,

crack ratios and thickness of FRP layers were modelled in ABAQUS 6.10. The first two natural

frequency ratios from numerical and analytical results are compared. Figure 3.13(a)-(d) show the

comparison of first natural frequency ratio between analytical and numerical results for RC, RCF1,

RCF2 and RCF3 columns respectively. It can be seen fro these plots that there is a good correlation

among the results. Figure 3.14(a)-(d) shows similar comparison but for second frequency ratio. It

is also evident from the plots that the results from analytical formulation matches with those that

of the numerical models. Also presented in the Tables 3.8-3.9 are the values of average percentage
 49

errors in the frequency ratios calculated with respect to numerical result. It can be seen from these

tables that these errors are within 2% of numerical results with the exception for the cases of first

frequency ratio of columns RCF2 and RCF3 when a/d is 0.1. It can also be observed that the error

increases with the crack depth for all the cases with the exception when a/d is 0.1.

3.8 Summary

In this Chapter, free vibration equation of motion for cracked column considering the restraints

of steel reinforcement and FRP confinement is formulated by using the continuous cracked beam

theory as described in Chapter 2. This is done by using the compatibility of deformations (i)

between steel reinforcement and concrete at the level of steel-concrete interface and (ii) between

FRP and concrete at the level of FRP-concrete interface. The effect of steel reinforcement and

FRP confinement on the frequencies and mode shapes of cantilever column is also studied. A finite

element simulation of cracked column with steel and FRP is also performed to verify the analytical

results. It has been found from the study that the inclusion of FRP and steel reinforcement will

alter the modal properties of the column.

50

Table 3.1: Parameters of Cantilever Column

Length of column , L 20m

d 3.9m
b 4.05m
r 4.5m
zs1 , zs2 1.875m
As1 , As2 50 no’s of 28mm diameter steel bars
Density of Concrete 2450 Kg/m3
Density of Steel 7850 Kg/m3
Density of FRP 1550 Kg/m3
Mass at the tip of column 1000 Kg
Modulus of elasticity of Concrete 2.8×1010 N/m2
Modulus of elasticity of Steel 2.1×1011 N/m2
Modulus of elasticity of FRP 2.4×1011 N/m2
Thickness of single FRP layer 0.165 mm

Table 3.2: Nomenclature used in results

Nomenclature Details
ωU Frequency of healthy concrete column
ωC Frequency of cracked concrete column
ωU R Frequency of healthy RC column
ωCR Frequency of cracked RC column
ωCRF Frequency of cracked RC column confined with FRP
RCF1 RC column confined with five FRP layers
RCF2 RC column confined with ten FRP layers
RCF3 RC column confined with fifteen FRP layers
 51

Table 3.3: Frequency ratio (ωC /ωU ) in terms of crack depth at different crack positions for con-
crete column

CPR
Modes 0.1 0.2 0.3 0.4 0.5
CDR = 0.1
First Mode 0.969183 0.976669 0.984228 0.990266 0.994657
Second Mode 0.989908 0.999741 0.995991 0.985735 0.978919
Third Mode 0.998305 0.994767 0.984831 0.99206 0.999979
CDR = 0.2
First Mode 0.909016 0.923788 0.945397 0.965014 0.980205
Second Mode 0.978949 0.999329 0.988367 0.957653 0.936514
Third Mode 0.998094 0.987636 0.964543 0.981514 0.9999
CDR = 0.3
First Mode 0.840891 0.859597 0.893414 0.928329 0.957736
Second Mode 0.97461 0.999295 0.979959 0.927548 0.890079
Third Mode 0.998894 0.982068 0.950495 0.974644 0.99974
CDR = 0.4
First Mode 0.770413 0.791384 0.834049 0.882952 0.927777
Second Mode 0.973601 0.999488 0.971472 0.898598 0.844669
Third Mode 0.999454 0.977908 0.940923 0.970337 0.999511

Table 3.4: Frequency ratio (ωCR /ωU R ) in terms of crack depth at different crack positions for RC
column

CPR
Modes 0.1 0.2 0.3 0.4 0.5
CDR = 0.1
First Mode 0.971907 0.978749 0.985649 0.991151 0.995146
Second Mode 0.990775 0.999764 0.996349 0.987004 0.980789
Third Mode Frequency Ratio 0.99845 0.995232 0.986157 0.992755 0.999981
CDR = 0.2
First Mode 0.91887 0.932201 0.951612 0.96911 0.982574
Second Mode 0.981074 0.9994 0.98965 0.962274 0.943409
Third Mode 0.99829 0.988959 0.968245 0.983439 0.99991
CDR = 0.3
First Mode 0.8625 0.879164 0.90902 0.939366 0.964518
Second Mode 0.97771 0.999387 0.982708 0.937376 0.90495
Third Mode 0.999043 0.984408 0.956782 0.977843 0.999773
CDR = 0.4
First Mode 0.811328 0.829706 0.866521 0.907516 0.943883
Second Mode 0.977699 0.999574 0.976517 0.916354 0.871934
Third Mode 0.999566 0.981578 0.950467 0.975081 0.99959
52

Table 3.5: Frequency ratio (ωCRF /ωU R ) in terms of crack depth at different crack positions for
column RCF1

CPR
Modes 0.1 0.2 0.3 0.4 0.5
CDR = 0.1
First Mode 0.975596 0.982674 0.989814 0.99551 0.99965
Second Mode 0.995138 1.004436 1.000899 0.991223 0.984789
Third Mode 1.003077 0.999743 0.990353 0.997181 1.004661
CDR = 0.2
First Mode 0.922219 0.935754 0.95547 0.973254 0.986948
Second Mode 0.985458 1.004071 0.994158 0.966332 0.947159
Third Mode 1.002943 0.99346 0.972416 0.987854 1.004589
CDR = 0.3
First Mode 0.865824 0.882637 0.912765 0.943406 0.968815
Second Mode 0.98218 1.004062 0.987217 0.941438 0.908694
Third Mode 1.003714 0.988937 0.961049 0.982312 1.004451
CDR = 0.4
First Mode 0.814765 0.833256 0.870296 0.911557 0.948166
Second Mode 0.982236 1.004252 0.981043 0.920486 0.875774
Third Mode 1.004243 0.986138 0.954827 0.979601 1.004268

Table 3.6: Frequency ratio (ωCRF /ωU R ) in terms of crack depth at different crack positions for
column RCF2

CPR
Modes 0.1 0.2 0.3 0.4 0.5
CDR = 0.1
First Mode 0.97926 0.986574 0.993955 0.999847 1.004129
Second Mode 0.999477 1.009084 1.005425 0.995418 0.988765
Third Mode 1.007681 1.00423 0.994526 1.001585 1.009317
CDR = 0.2
First Mode 0.925547 0.939286 0.959306 0.977377 0.991296
Second Mode 0.98982 1.008718 0.998643 0.970368 0.950888
Third Mode 1.007572 0.997937 0.976566 0.992247 1.009245
CDR = 0.3
First Mode 0.869127 0.88609 0.916493 0.947426 0.947426
Second Mode 0.986628 1.008712 0.991702 0.945479 0.945479
Third Mode 1.008361 0.993443 0.965293 0.986758 0.986758
CDR = 0.4
First Mode 0.818184 0.836786 0.874055 0.915574 0.952432
Second Mode 0.98675 1.008906 0.985545 0.924596 0.879594
Third Mode 1.008896 0.990675 0.959165 0.984098 1.008922
 53

Table 3.7: Frequency ratio (ωCRF /ωU R ) in terms of crack depth at different crack positions for
column RCF3

CPR
Modes 0.1 0.2 0.3 0.4 0.5
CDR = 0.1
First Mode 0.982902 0.99045 0.998071 1.004158 1.008584
Second Mode 1.003794 1.013709 1.009928 0.999589 0.992717
Third Mode 1.012261 1.008694 0.998675 1.005964 1.013949
CDR = 0.2
First Mode 0.928857 0.942799 0.96312 0.981475 0.981475
Second Mode 0.994159 1.013341 1.003105 0.974382 0.974382
Third Mode 1.012178 1.002391 0.980694 0.996617 0.996617
CDR = 0.3
First Mode 0.872415 0.889524 0.920198 0.951422 0.977341
Second Mode 0.991054 1.013339 0.996164 0.949499 0.916122
Third Mode 1.012984 0.997925 0.969516 0.991181 1.013736
CDR = 0.4
First Mode 0.821583 0.840297 0.877791 0.919572 0.95667
Second Mode 0.991241 1.013536 0.990024 0.928683 0.883392
Third Mode 1.013526 0.995189 0.96348 0.988571 1.013552

Table 3.8: Average percentage error in first frequency ratio with respect to numerical result for
different crack depth ratios

Average percentage error in first frequency ratio

with respect to numerical result
RC column RCF1 RCF2 RCF3
a/d = 0.1 0.3674 1.7333 3.257 4.6
a/d = 0.2 0.354 0.185 0.172 0.677
a/d = 0.3 0.595 0.338 0.692 1.109
a/d = 0.4 0.883 0.4124 1.088 1.23

Table 3.9: Average percentage error in second frequency ratio with respect to numerical result for
different crack depth ratios

Average percentage error in second frequency ratio

with respect to numerical result
RC column RCF1 RCF2 RCF3
a/d = 0.1 0.285 0.788 1.164 1.652
a/d = 0.2 0.3459 0.4123 0.6021 0.9654
a/d = 0.3 0.6649 0.5412 0.8184 1.1855
a/d = 0.4 0.9864 0.7464 0.976 1.385
54

(a) CDR = 0.1 (b) CDR = 0.2

(c) CDR = 0.3 (d) CDR = 0.4

Figure 3.1: Variation of fundamental frequency ratio for concrete and RC columns with crack
position ratio for: (a) CDR = 0.1; (b) CDR = 0.2; (c) CDR = 0.3; (d) CDR = 0.4
 55

(a) CDR = 0.1 (b) CDR = 0.2

(c) CDR = 0.3 (d) CDR = 0.4

Figure 3.2: Variation of second frequency ratio for concrete and RC columns with crack position
ratio for: (a) CDR = 0.1; (b) CDR = 0.2; (c) CDR = 0.3; (d) CDR = 0.4
56

(a) CDR = 0.1 (b) CDR = 0.2

(c) CDR = 0.3 (d) CDR = 0.4

Figure 3.3: Variation of fundamental frequency ratio (ωCRF /ωU ) for RCF1, RCF2, RCF3 with
crack position ratio for: (a) CDR = 0.1; (b) CDR = 0.2; (c) CDR = 0.3; (d) CDR = 0.4
 57

(a) CDR = 0.1 (b) CDR = 0.2

(c) CDR = 0.3 (d) CDR = 0.4

Figure 3.4: Variation of second frequency ratio (ωCRF /ωU ) for RCF1, RCF2, RCF3 with crack
position ratio for: (a) CDR = 0.1; (b) CDR = 0.2; (c) CDR = 0.3; (d) CDR = 0.4
58

Figure 3.5: Fundamental mode shape for CDR 0.3 and CPR 0.1 for (a) Concrete, RC, RCF1,
RCF2 and RCF3; Difference in mode shapes for (b) Concrete and RC, (c) Concrete and RCF1, (d)
Concrete and RCF2, (e) Concrete and RCF3
 59

Figure 3.6: Fundamental mode shape for CDR 0.4 and CPR 0.1 for (a) Concrete, RC, RCF1,
RCF2 and RCF3; Difference in mode shapes for (b) Concrete and RC, (c) Concrete and RCF1, (d)
Concrete and RCF2, (e) Concrete and RCF3
60

Figure 3.7: Second mode shape for CDR 0.3 and CPR 0.1 for (a) Concrete, RC, RCF1, RCF2 and
RCF3; Difference in mode shapes for (b) Concrete and RC, (c) Concrete and RCF1, (d) Concrete
and RCF2, (e) Concrete and RCF3
 61

Figure 3.8: Second mode shape for CDR 0.4 and CPR 0.1 for (a) Concrete, RC, RCF1, RCF2 and
RCF3; Difference in mode shapes for (b) Concrete and RC, (c) Concrete and RCF1, (d) Concrete
and RCF2, (e) Concrete and RCF3
62

Figure 3.9: FRP shell model in ABAQUS with partitions

Figure 3.10: Assembly of RC Column with FRP

 63

Figure 3.11: Details of reinforcement in column

Figure 3.12: Meshed view of FRP cover

64

(a) RC column (b) RCF1

(c) RCF2 (d) RCF3

Figure 3.13: Comparison of analytical and numerical cracked column first mode frequencies for
(a) RC column, (b) RCF1, (c) RCF2, (d) RCF3
 65

(a) RC column (b) RCF1

(c) RCF2 (d) RCF3

Figure 3.14: Comparison of analytical and numerical cracked column second mode frequencies
for (a) RC column, (b) RCF1, (c) RCF2, (d) RCF3
66
Chapter 4

Effect of Crack Position and Crack

Depth on Modal Properties

4.1 Brief Overview

This chapter presents an analytical study on the effect of crack depth and crack position on the

modal parameters of the cracked cantilever column. The predictions are then illustrated with the

help of analytical results obtained in Chapter 2.

4.2 Perturbation Formulation

The equation of motion for free vibration of cracked cantilever column is given by Equation 2.35.

Let us consider the following coordinate transformation:

w(x) = φ̂(x)q(t), (4.1)

Based on the coordinate transformation, Equation 2.35 can be expressed in the form of following

differential equations:
68

" #
d4 φ̂i (x) d2 d2 φ̂i (xp )
+  2 f (x) − λ̂i φ̂i (x) = 0 (4.2)
dx4 dx dx2

d2 q(t)
 
Ec C1
+ λ̂i q(t) = 0 (4.3)
dx2 ρc Ac

where  = C2 /C1 ; φ̂i (x) and λ̂i represent the ith eigenfunction and eigenvalue of the cracked col-

umn, respectively. By considering perturbation expansion, the cracked column eigen parameters

can be expressed as follows:

φ̂i (x) = φ0i (x) + φ1i (x) + o(2 ) (4.4)

λ̂i = λ0i + λ1i + o(2 ) (4.5)

On substituting the above expressions of λ̂i and φ̂i (x) in Equation 4.2 and equating the coefficients

of equal powers of  upto o(), the following equations are obtained:

0th order:
d4 φ0i (x)
− λ0i φ0i (x) = 0 (4.6)
dx4

1st order:
d4 φ1i (x) d2 d2 φ0i (xp )
 
0 1 1 0
− λi φi (x) = λi φi (x) − 2 f (x) (4.7)
dx4 dx dx2

From Ciambella and Vestroni (2015), the following solution is adopted:

(1)
X
φi (x) = φ0i (x) + βik φ0k (x) (4.8)
k6=i
 69

X
φ̂i (x) = (1 − )φ0i (x) −  βik φ0k (x) + o(2 ) (4.9)
k6=i

where
RL d2 φ0i (xp ) d2 φ0k (x)
f (x) dx
0 dx2 dx2
βik = (4.10)
RL d2 φ0k (x) 2
 
0 0
(λi − λk ) dx
0 dx2

and
RL d2 φ0i (xp ) d2 φ0i (x)
f (x) dx
0 dx2 dx2
λ1i = (4.11)
RL 0
2
φi (x) dx
0

From these equations, the difference between the damaged and undamaged eigenvalues (∆λi ) and

normalized eigenvectors (∆φi (x)) can be expressed as,

RL d2 φ0i (xp ) d2 φ0i (x)

f (x) dx
dx2 dx2
∆λi = λ1i =  0 (4.12)
RL 0
2
φi (x) dx
0

d2 φ0i (xp ) d2 φ0k (x)

RL
f (x) dx
X 0 dx2 dx2
∆φi (x) =  φ0 (x) (4.13)
L(λ0i − λ0k )kφ0k (x)k2 kφ0i (x)k i
k6=i

From Equation 4.12, it can be observed that for a particular mode, the change in eigenvalue (∆λi )
d2 φ0i (xp )
with varying crack positions is dependent on the term . By defining this term as the
dx2
curvature of mode shape, it can be concluded that the difference in eigenvalues is maximum if

curvature of the mode shape is having a maximum value at crack location. Also, the eigenvalue of

the cracked system is not affected when crack is present at the location of node of that curvature

mode shape. To find the dependence of eigenvectors on the crack location, the value of ∆φi (x) is

found out in MATLAB by using the first ten mode shapes (k 6= i and k = 1 to 10). The following

mode shape is used for calculations.

 
cosh(βi L) + cos(βi L)
φ0i (x) = (cosh(βi x) − cos(βi x)) + (sin(βi x) − sinh(βi x)) (4.14)
sinh(βi L) + sin(βi L)
70

and βi is solution of the equation

cos(βi L) cosh(βi L) = 1; (4.15)

Figures 4.1-4.3 show the absolute values of difference for first three mode shapes at crack location

as a function of crack position ratio. It can be seen from Figure 4.1 that for the first mode, the

maximum difference occurs when crack is close to fixed end of the column. Figure 4.2 shows that

for second mode, maximum difference in mode shapes occur between crack positions of 0.4L and

0.5L while the minimum difference in mode shape is observed when crack is at 0.2L. From Figure

4.3, it can be observed that for the third mode, the maximum difference occurs at 0.3L and 0.7L

while a minimum is observed at 0.5L. From these figures and from Table 4.1, it can be concluded

that the presence of damage at the node of curvature of the mode will not have significant effect

on that mode. On the other hand, presence of crack at the location of maximum mode shape

curvature, results in maximum effect on that mode.

4.3 Modal Properties of Cantilever Column

In this section, the mode shapes, slope and curvature of first three mode shapes of cantilever

column are presented (Figures 4.4-4.6). Also, maximum mode displacement and node positions

for both mode shape and curvature of mode shape are presented in Table 4.1.

Table 4.1: Mode shapes and curvatures of first three modes of cantilever column

Mode shape Curvature of mode shape

Maximum Node Maximum Node
First mode L 0 0 L
Second mode 0.471L 0.783L 0.529L 0.216L
0.291L 0.503L 0.308L 0.132L
Third mode
0.692L 0.867L 0.709L 0.496L
 71

4.4 Numerical Illustration

To illustrate numerically, the dependence of crack position on the modal properties, the frequency

ratios (ωC /ωU , where ωC is the frequency of the cracked column and ωU is the frequency of the

uncracked column) and mode shape differences are studied for the column geometry as shown in

Figure 2.2 and with geometrical and material properties as listed in Table 2.1.

4.4.1 Effect of Crack Position on Frequencies

Fundamental frequency: Figure 4.7 shows the variation of first frequency ratio with the position of

crack. It can be seen that the maximum variation in frequency is observed when the crack position

is close to the fixed end. As the crack position shifts away from fixed end, the frequency ratio

approaches unity. Also, from Figure 4.4 and Table 4.1, the curvature of mode shape is maximum

near the fixed end and decreases monotonically to zero towards the free end. So, it can be observed

that the dependency of frequency ratio is similar to that predicted from Equation 4.12.

Second frequency: In Figure 4.8, the variation of second frequency ratio with crack position is

presented. It can be observed that the frequency ratio is close to unity at crack positions of 0.2L

and 1L. Also, from Table 4.1, it can be noted that these are the positions of zero curvature. The

frequency variation is maximum at crack position of about 0.5L, which is a position of maximum

curvature and can be confirmed from Figure 4.5.

Third frequency: Figure 4.9 shows the variation of third frequency ratio with the position of crack.

From this figure, it can be seen that the frequency ratio is unity at crack positions of 0.5L, in

between 0.1L and 0.2L, and towards the free end (1L). Local minima in frequency ratio values

are observed at crack positions of 0.3L and 0.7L. These values are in accordance with zero and

maximum curvature positions from Table 4.1.

72

4.4.2 Effect of Crack position on Mode shapes

First Mode Shape: Figures 4.10-4.13 show the difference in damaged and undamaged first modes

for crack positions of 0.1L, 0.3L, 0.5L and 0.9L. It can be seen from these figures that the effect

of crack on mode shapes is less when crack position moves away from the fixed end. This is in

accordance with the behaviour that can be understood from Figure 4.1.

Second Mode Shape: In Figures 4.14–4.17, the difference in damaged and undamaged first modes

for crack positions of 0.1L, 0.2L, 0.5L and 0.8L are resented. The results indicate that the differ-

ence in mode shapes at crack position when the crack is at 0.2L and it is maximum at 0.5L. These

are the positions of zero and maximum curvatures respectively from Table 4.1.

Third Mode Shape: Figures 4.18-4.21 show the difference in damaged and undamaged first modes

for crack positions of 0.1L, 0.3L, 0.5L and 0.7L. It can be observed from these figures that the

mode shape difference is minimum for crack positions of 0.1L and 0.5L, while the difference is

more for crack positions of 0.3L and 0.7L.

4.5 Summary

In this chapter, the effect of crack depth and crack position on the modal parameters of the cracked

cantilever column is studied by using perturbation approach. It was observed that the variation

in frequencies and mode shapes because of the presence of damage is related to the curvature of

the mode shape. The predictions are then illustrated with the help of analytical results obtained in

Chapter 2.
 73

Figure 4.1: Variation of first mode shape difference at crack location with crack position ratio

Figure 4.2: Variation of second mode shape difference at crack location with crack position ratio

Figure 4.3: Variation of third mode shape difference at crack location with crack position ratio
74

Figure 4.4: First mode shape of cantilever column: (a) mode shape, (b) slope of mode shape, (c)
curvature of mode shape

Figure 4.5: Second mode shape of cantilever column: (a) mode shape, (b) slope of mode shape,
(c) curvature of mode shape
 75

Figure 4.6: Third mode shape of cantilever column: (a) mode shape, (b) slope of mode shape, (c)
curvature of mode shape

Figure 4.7: Effect of crack position on first frequency ratio

76

Figure 4.8: Effect of crack position on second frequency ratio

Figure 4.9: Effect of crack position on third frequency ratio

Figure 4.10: Difference in first mode shapes between cracked and healthy column for crack posi-
tion 0.1L
 77

Figure 4.11: Difference in first mode shapes between cracked and healthy column for crack posi-
tion 0.3L

Figure 4.12: Difference in first mode shapes between cracked and healthy column for crack posi-
tion 0.5L

Figure 4.13: Difference in first mode shapes between cracked and healthy column for crack posi-
tion 0.9L
78

Figure 4.14: Difference in second mode shapes between cracked and healthy column for crack
position 0.1L

Figure 4.15: Difference in second mode shapes between cracked and healthy column for crack
position 0.2L

Figure 4.16: Difference in second mode shapes between cracked and healthy column for crack
position 0.5L
 79

Figure 4.17: Difference in second mode shapes between cracked and healthy column for crack
position 0.8L

Figure 4.18: Difference in third mode shapes between cracked and healthy column for crack
position 0.1L

Figure 4.19: Difference in third mode shapes between cracked and healthy column for crack
position 0.3L
80

Figure 4.20: Difference in third mode shapes between cracked and healthy column for crack
position 0.5L

Figure 4.21: Difference in third mode shapes between cracked and healthy column for crack
position 0.7L
Chapter 5

Summary and Conclusions

5.1 Summary

In recent years, FRP composite strengthening is identified as a good alternate to the conventional

methods for rehabilitating deteriorated structures. Repair of sections such as beams and columns

is very common. Typically, the structural members are being subjected to dynamic loads such as

moving and impulse loads in the case of bridge piers. Hence knowledge of dynamic properties

of structural elements retrofitted with FRP is useful for proper design of rehabilitation scheme.

This work is directed towards studying the modal properties of a cracked reinforced concrete (RC)

column strengthened with FRP. At first, continuous cracked beam theory is used to formulate the

free vibration equations of motion for a cracked concrete column. Rayleigh’s quotient is used to

estimate the frequencies and mode shapes. A numerical study is then performed in ABAQUS/CAE

v 6.10 to validate the results from formulation. The variations of frequencies and mode shapes

with depth and location of crack are studied. Further, to study the dependence of mode shapes

and frequencies on the damage location, perturbation formulation is used. The expressions for

difference in mode shapes and frequencies as a function of crack depth and crack position are

obtained. The results from this formulation are compared with the results of cracked column. The

continuous cracked beam theory is then modified to incorporate the effects of steel reinforcement
82

and FRP confinement. A similar solution methodology based on Rayleigh’s quotient is also used

here. Further, a finite element analysis is carried out by modelling a cracked column along with

steel and FRP constraints in ABAQUS/CAE v 6.10 to compare the frequencies and mode shapes.

5.2 Conclusions

Based on this study, some important conclusions that can be drawn are listed as follows:

• The reduction in natural frequencies is found to vary with the position and depth of a crack.

For the column considered herein, it has been observed that for a crack located at 0.1L, the

reduction in fundamental frequency is about 4% and 50% for with crack depths of 0.1d and

0.8d, respectively.

• The frequencies of cracked RC column were found to increase when the restraints of steel

and FRP confinement are taken into consideration. The increase in fundamental frequency

is found to be about 7 - 9% for crack with depth 0.4d located at 0.1L.

• Depending on the crack position and depth, the ability of FRP confinement to alter the

natural frequencies varies. It has been observed that the application of FRP leads to an

increase in fundamental frequency by about 8% in comparison with the frequency of a

concrete column for a crack of depth 0.4d located at 0.1L.

• For the chosen cases as considered here, it has been found that for better estimation of mode

shapes, steel bar restraints are to be considered. However, the effect of FRP on mode shapes

is negligible.

• From the perturbation formulation, it has been observed that the difference in mode shapes

and frequencies is sensitive to the curvature of mode shape depending on the location of

damage. For cantilever members, the first mode is effective in localising the damage if the

crack is closer to the fixed end. Further, for crack at the mid span, the second mode is
 83

effective in localising the damage, while the third mode fails to effectively localise.

The findings of this study are limited to the parameter space considered herein, i.e., the geometry

of column, type of crack, amount of FRP confinement etc.

5.3 Scope for future work

The following areas can be studied in future:

• In this work, the crack is modelled as a single edge crack. Breathing cracks are needed to

be considered as well.

• To represent a realistic scenario, multiple cracks are to be considered.

• Analytical models can be derived by considering large deformation theory and also by in-

cluding the effects of shear deformation and rotational inertia.

• Studies can be performed by considering the effects of adhesive layer and angle of ply for

FRP wrap.

• Also, a study is required to be conducted to investigate the effect of FRP strengthening on

frequencies for various crack locations.

84
References

ABAQUS CAE (2010). “Abaqus/cae user’s manual, version 6.10.” Abaqus Inc., USA.

Behzad, M., Ebrahimi, A., and Meghdari, A. (2008). “A new continuous model for flexural vibra-

tion analysis of a cracked beam.” Polish Maritime Research, 15(2), 32–39.

Behzad, M., Meghdari, A., and Ebrahimi, A. (2005). “A new approach for vibration analysis of

a cracked beam.” International Journal of Engineering-Materials and Energy Research Centre,

18(4), 319.

Behzad, M., Meghdari, A., and Ebrahimi, A. (2008). “A linear theory for bending stress–strain

analysis of a beam with an edge crack.” Engineering Fracture Mechanics, 75(16), 4695–4705.

Carneiro, S. H. and Inman, D. J. (2002). “Continuous model for the transverse vibration of cracked

timoshenko beams.” Journal of Vibration and Acoustics, 124(2), 310–320.

Chaudhuri, S. R. (2008). “Change in instantaneous eigenproperties due to yielding of a structure.”

Journal of Sound and Vibration, 312(4), 754–768.

Chen, D. and Liu, F. (2014). “Free vibration of a single-edge cracked rc beam strengthened with

frp.” Journal of Engineering Mechanics, 140(6).

Chondros, T., Dimarogonas, A., and Yao, J. (1998). “A continuous cracked beam vibration theory.”

Journal of Sound and Vibration, 215(1), 17–34.

86

Christides, S. and Barr, A. (1984). “One-dimensional theory of cracked bernoulli-euler beams.”

International Journal of Mechanical Sciences, 26(11), 639–648.

Ciambella, J. and Vestroni, F. (2015). “The use of modal curvatures for damage localization in

beam-type structures.” Journal of Sound and Vibration, 340, 126–137.

Ciambella, J., Vestroni, F., and Vidoli, S. (2011). “Damage observability, localization and assess-

ment based on eigenfrequencies and eigenvectors curvatures.” Smart Structures and Systems,

8(2), 191–204.

Dimarogonas, A. D. (1996). “Vibration of cracked structures: a state of the art review.” Engineer-

ing Fracture Mechanics, 55(5), 831–857.

Hamed, E. and Rabinovitch, O. (2005). “Dynamic behavior of reinforced concrete beams strength-

ened with composite materials.” Journal of Composites for Construction, 9(5), 429–440.

Heydari, M., Ebrahimi, A., and Behzad, M. (2014). “Forced vibration analysis of a timoshenko

cracked beam using a continuous model for the crack.” Engineering Science and Technology,

an International Journal, 17(4), 194–204.

Roy, K. and Ray-Chaudhuri, S. (2013). “Fundamental mode shape and its derivatives in structural

damage localization.” Journal of Sound and Vibration, 332(21), 5584–5593.

Shen, M.-H. and Pierre, C. (1990). “Natural modes of bernoulli-euler beams with symmetric

cracks.” Journal of Sound and Vibration, 138(1), 115–134.

Shen, M.-H. and Pierre, C. (1994). “Free vibrations of beams with a single-edge crack.” Journal

of Sound and Vibration, 170(2), 237–259.

Stallings, J., Tedesco, J., El-Mihilmy, M., and McCauley, M. (2000). “Field performance of frp

bridge repairs.” Journal of Bridge Engineering, 5(2), 107–113.

 87

Tedesco, J. W., Stallings, J. M., and El-Mihilmy, M. (1999). “Finite element method analysis

of a concrete bridge repaired with fiber reinforced plastic laminates.” Computers & Structures,

72(1), 379–407.