Effect of surface blast treatment on fatigue behaviour of out of plane gusset welded joints

Thesis for the Degree of Doctor of Philosophy Effect of Surface Blast Treatment on Fatigue Behaviour of Out-of-Plane Gusset Welded Joints by Le Van Phuoc Nhan Department of Civil Eng

Thesis for the Degree of Doctor of Philosophy

Effect of Surface Blast Treatment on Fatigue Behaviour of Out-of-Plane

Gusset Welded Joints

by

Le Van Phuoc Nhan

Department of Civil Engineering

The Graduate School Pukyong National University

August 2009

Effect of Surface Blast Treatment on Fatigue Behaviour of Out-of-Plane

Gusset Welded Joints

이음부의 피로거동에 미치는 영향)

Advisor : Prof Dong Uk, Lee

by

Le Van Phuoc Nhan

A thesis submitted in partial fulfillment of the requirements

for the degree of

Doctor of Philosophy

in Department of Civil Engineering

The Graduate School Pukyong National University

August 2009

Effect of Surface Blast Treatment on Fatigue Behaviour of

Out-of-Plane Gusset Welded Joints

TABLE OF CONTENTS

TABLE OF CONTENTS i

LIST OF TABLES vii

LIST OF FIGURES x

LIST OF PHOTOS xvii

SYMBOLS .xxi

DEFINITIONS xxiv

ABSTRACT xxviii

CHAPTER 1: GENERAL 1.1 Introduction 1

1.2 Background 4

1.2.1 Fatigue Failure of Welded Joints 4

1.2.2 Fatigue Life 11

1.2.2.1 Fatigue Crack Initiation Life 13

1.2.2.2 Fatigue Crack Propagation Life 13

1.2.3 Factors Effect on Fatigue Strength of Welded Joints 16

1.2.3.1 Effect of Weld Quality 16

1.2.3.2 Effect of Size 17

1.2.3.3 Effect of Residual Stress 20

1.2.3.4 Effect of Surface Roughness 21

1.2.3.5 Effect of Composition 21

1.2.3.6 Effect of Post Weld Treatment 23

1.3 Objectives 23

1.4 Summary 24

CHAPTER 2: PREVIOUS RESEARCHES 2.1 General Methods for Improving Fatigue Life 25

2.2 Methods of Modification of Weld Toe Geometry 26

2.3 Methods of Modification of Residual Stress Distribution 28

2.4 Methods of Fatigue Strength Improvement Applied in Steel Structures 36

2.4.1 Tungsten Inert Gas (TIG) Dressing Treatment 36

2.4.2 Grinding Treatment 38

2.4.3 Peening Treatment 41

2.4.3.1 Hammer Peening Treatment 41

2.4.3.2 Needle Peening Treatment 42

2.4.3.3 Shot Peening Treatment 43

2.4.4 Ultrasonic Impact Treatment (UIT) 46

2.5 Surface Blast Treatment 48

2.6 Summary 54

CHAPTER 3: FATIGUE TESTS 3.1 Fatigue Test 1 55

3.1.1 Materials 55

3.1.2 Detail of Specimen and Loading History 56

3.1.3 Results of Fatigue Test 58

3.2 Fatigue Test 2 60

3.2.1 Materials 60

3.2.2 Detail of Specimen and Loading History 61

3.2.3 Results of Fatigue Test 62

3.3 Fatigue Test 3 (Bending Test) 65

3.3.1 Materials 66

3.3.2 Detail of Beam 67

3.3.3 Nominal Stress Measurement 69

3.3.4 Hot Spot Stress (Geometry Stress) Measurement 72

3.3.5 Fatigue Test Program 78

3.3.6 Results of Fatigue Tests of As-Welded and Blast-Treated Specimen 80

3.4 Weld Toe Profile 82

3.3.1 Measurement of Radius and Flank Angle of Weld Toe 82

3.3.2 Results of Radius and Flank Angle of Weld Toe 82

3.5 Residual Stress 86

3.5.1 Measurement of Residual Stress 86

3.5.2 Results of Measurement of Residual Stress 86

3.6 Fatigue Crack Initiation and Propagation 88

3.6.1 Fatigue Crack of Out-of-Plane Gusset Welded Joints 88

3.6.2 Fatigue Crack of As-Welded Beam and Blast-Treated Beam (BT1 Beam) 92

3.7 Summary 98

CHAPTER 4: NUMERICAL ANALYSIS

4.1 Stress Concentration Factor 99

4.1.1 Determination of Stress Concentration Factor by Experimental Formula 99

4.1.2 Determination of Stress Concentration Factor by Finite Element Method (ANSYS Software) 101

4.2 General Method Evaluating Fatigue Life of Welded Joints 106

4.3 Effect of Weld Toe Geometry on Fatigue Endurance Limit 108

4.4 Effect of Residual Stress on Fatigue Endurance Limit 113

4.5 Prediction of Fatigue Life Considering Residual Stress 116

4.5.1 Method of Prediction Fatigue Crack Propagation Life 116

4.5.2 Illustrative Example 128

4.5.2.1 Results of Fatigue Life Prediction 130

4.5.2.2 Comparison with Fatigue Test Results 141

4.5.2.3 Effect of Ratio of Crack Aspect on Fatigue Life 142 4.6 Fatigue Crack Growth Rate 143

4.7 Summary and Conclusions 145

CHAPTER 5: APPLICATION OF RESEARCH IN PRACTICE 5.1 Design Curve 147

5.1.1 Design Curve σr – N Relationship Determined by BS

5400 147

5.1.2 Design Curve σr – N Relationship Determined by IIW/IIS152 5.1.3 Design Curve σr – N Relationship Determined by JSSC156 5.1.4 Design Curve σr – N Relationship Determined by AASHTO 158

5.2 Application in Fabrication 160

CHAPTER 6: CONCLUSIONS 162

APPENDICES 165

A.3.5.1 Fracture sections of Out-of-Plane Gusset Welded Joints 165

A.3.5.2 Crack 2 and Crack 3 of AW beam 166

A.4.5.2 S-N curves of AW and BT1 specimens 169

LIST OF REFERENCES 172

LIST OF PAPERS 184

ACKNOWLEDGEMENT 186

LIST OF TABLES

Table 2.1 Evaluation of improvement methods 31

Table 2.2 Classification and applications of steel shot 52

Table 2.3 Classification and applications of steel grit 53

Table 2.4 Blast treatment conditions 54

Table 3.1 Chemical composition and mechanical properties of steel SM490B 55

Table 3.2 Welding conditions 56

Table 3.3 Blast conditions 56

Table 3.4 Fatigue test program of AW and BT1 specimens 58

Table 3.5 Fatigue test results of AW and BT1 specimens 59

Table 3.6 Fabrication process of test specimens 61

Table 3.7 Fatigue test program of AW, BT1, BT2, BT3 and BT4

specimens 62

Table 3.8 Fatigue test results of AW, BT1, BT2, BT3 and BT4 specimens 63

Table 3.9 Chemical composition and mechanical properties of steel SS400 66

Table 3.10 Strain gauge used for balance of specimens and measurement

of nominal stress 68 Table 3.11 Strain gauge used for hot spot stress measurement 68 Table 3.12 Nominal stress at some locations on the beams 71 Table 3.13 Values of hot spot stress of as-welded specimen and blast-

treated specimens……… 77 Table 3.14 Fatigue test program of AW and BT1 beam 78 Table 3.15 Fatigue test results of AW and BT1 beam 81 Table 3.16 Mean values and standard deviation (SD) of measured weld

profile of out-of-plane gusset joints 83 Table 3.17 Mean values of radius and flank angle measured at gussets of

AW and BT1 beams 84 Table 3.18 Fatigue crack propagation of as-welded beam 97

Table 4.1 Stress concentration factor of out of-plane gusset welded

joints (Experimental Formula) 100

Table 4.2 Stress concentration factor of AW and BT1 beam

(Experimental Formula) 101

Table 4.3 Stress concentration factor of as-welded and blast-treated

specimen of gusset welded joints (FEM) 106

Table 4.4 Fatigue life of AW specimens with a/b = 0.3 132

Table 4.5 Fatigue life of AW specimens with a/b = 0.4 133

Table 4.6 Fatigue life of AW specimens with a/b = 0.5 134

Table 4.7 Fatigue life of BT1 specimens with a/b = 0.3 137

Table 4.8 Fatigue life of BT1 specimens with a/b = 0.3 138

Table 4.9 Fatigue life of BT1 specimens with a/b = 0.3 139

Table 5.1 σ r – N relationships and constant amplitude non-propagating stress range value 148

Table 5.2 Mean-line σr – N relationships 149

Table 5.3 Probability factor 149

Table 5.4 Probability factor 150

Table 5.5 Constant, constant amplitude fatigue limit and cut-of limit 153

Table 5.6 Basic allowable stress range for joints subjected to normal stress by JSSC 157

Table 5.7 Basic allowable stress range for joints subjected to normal stress by AASHTO 159

LIST OF FIGURES

Fig.1.1 Stress flow lines and stress concentration 6

Fig.1.2 Stress concentration regions for weldment 8

Fig.1.3 Imperfections and cracks in welded joints 9

Fig.1.4 Residual stress for butt-welded plate 10

Fig.1.5 Schematic S-N curve divided into an initiation and propagation component 12

Fig.1.6 Schematic representation of fatigue crack growth in steel 14

Fig.1.7 Schematic showing relation between initiation life and propagation life 15

Fig.1.8 Effect of weld profile on fatigue strength of transverse welds 16 Fig.1.9 Dimensions relevant to size effects in transverse fillet and butt welded joints 17

Fig.1.10 Fatigue test results showing influence of plate thickness and attachment length 18

Fig.1.11 Effect of plate width on stress concentration factor (SCF) at end of longitudinal attachment (NB-SCF does not include the effect of the weld itself) 19 Fig.1.12 Effective stress resulting from superposition of applied and

residual stress 20 Fig.1.13 Effect of steel strength on fatigue strength 22

Fig.1.14 Fatigue test results from fillet weld in various strength of steel

22 Fig.2.1 Classification of methods of modification of weld toe

geometry 27 Fig.2.2 Classification of methods of modification of residual stress

distribution 29 Fig.2.3 Comparison of improvement methods for mild steel

specimens with transverse non-load-carrying fillet welds 30 Fig.2.4 Comparison of improvement methods for mild steel

specimens with longitudinal non-load-carrying fillet welds 30 Fig.2.5 Comparison of improvement methods for medium strength

steel specimens with transverse non-load-carrying fillet welds

35 Fig.2.6 TIG Dressing 37

Fig.2.7 The weld toe burr grinding technique 39

Fig.2.8 The burr grinding technique, showing depth and width of

groove in stressed plate 40

Fig.2.9 Details of burr grinding weld toe geometry 40

Fig.2.10 Hammer peening technique 43

Fig 2.11 Impact of steel shot on surface of specimen 46

Fig 2.12 Surface of specimen after shot peening treatment 46

Fig 2.13 Shot blasting treatment 49

Fig.3.1 Dimensions of fatigue test specimens 57

Fig.3.2 S-N curves of AW and BT1 specimens (fatigue test 1) 60

Fig.3.3 S-N relationship of AW and BT specimens (fatigue test 2) 64

Fig.3.4 S-N curves of AW and BT1 specimens (fatigue test 1 and 2) 64 Fig.3.5 S-N relationship of AW and BT specimen (fatigue test 1 and 2) 65

Fig.3.6 Detail of test specimen (mm) 67

Fig.3.7 Applied load conditions (size unit of mm) 68

Fig.3.8 Distribution of nominal stress at mid-span of BT1 beam 70

Fig.3.9 Hot spot and hot spot stress 73

Fig.3.10 Locations of strain gauges in measuring hot spot stress 74

Fig.3.11 Extrapolation of hot spot stress (geometric tress) 75 Fig.3.12 Locations of strain gauges used for measurement of nominal

stress and hot spot stress 75

Fig.3.13 Hot spot stress measured at point L3 of the beams 77

Fig.3.14 Fatigue test results of AW beam 79

Fig.3.15 Fatigue test results of BT1 beam 80

Fig.3.16 Measurement of flank angle and radius of weld toe 82

Fig.3.17 Radius and flank angle of weld toe of out-of-plane gusset specimens 84

Fig.3.18 Radius and flank angle of weld toe of AW and BT1 beam 85

Fig.3.19 Distribution rate of radius of weld toe of AW and BT1 beam 85 Fig.3.20 Location of strain gauges and cutting direction in residual stress measurement 87

Fig.3.21 Distribution of longitudinal residual stress of out-of-plane sset stress specimens 87

Fig.3.22 a-N curves of AW and BT1 specimen of fatigue test 1 with with Δσ = 150 MPa 89

Fig.3.23 a-N curves of AW and BT1 specimen of fatigue test 2 with with Δσ = 135 MPa 90

Fig.3.24 a-N curves of AW and BT1 specimen of fatigue test 2 with with Δσ = 150 MPa 91

Fig.3.25 Fatigue crack propagation rate of AW beam 95

Fig.4.1 Parameters used in determination of stress concentration factor 100

Fig.4.2 SOLID92 geometry 102

Fig.4.3 SOLID92 stress output 102

Fig.4.4 Radius of weld toe used to model the for welding 103

Fig.4.5 The plain plate subjected to applied stress and the redistribution of residual stress after crack initiation formed……… 120

Fig.4.6 The distribution of residual stress in out-of-plane gusset welde welded joints 121

Fig.4.7 The residual stress distribution in the depth direction of blast-treated specimen 122

Fig.4.8 Diagram of fatigue propagation life calculation (Method a) 126 Fig.4.9 Diagram of fatigue propagation life calculation (Method b) 127 Fig.4.10 S-N curve of AW specimens with a/b=0.3 135

Fig.4.11 S-N curve of AW specimens with a/b=0.4 135

Fig.4.12 S-N curve of AW specimens with a/b=0.5 136

Fig.4.13 S-N curve of BT1 specimens with a/b=0.3 140

Fig.4.14 S-N curve of BT1 specimens with a/b=0.4 140

Fig.4.15 S-N curve of BT1 specimens with a/b=0.5 141

Fig.4.16 S-N curves obtained from fatigue test and prediction 141

Fig.4.17 S-N curves of AW specimens with ratios of crack aspect of 0.3,

0.4 and 0.5, and initial crack size of 0.03mm and 0.04mm 142 Fig.4.18 S-N curves of BT1 specimens with ratios of crack aspect of

0.3, 0.4 and 0.5, and initial crack size of 0.03mm and 0.04mm 143 Fig.4.19 Fatigue crack propagation rate of out-of-plane gusset welded

joints of AW and BT1 specimens 144 Fig.5.1 Fatigue resistance S-N curves of BS5400 151

Fig.5.2 S-N curves of AW and BT specimens and BS5400 151

Fig.5.3 Fatigue resistance S-N curves for m = 3.0, normal stress (IIW)

155 Fig.5.4 S-N curves of AW and BT specimens and IIW 155

Fig.5.5 Fatigue design S-N curves for joints subjected to normal stress

(JSSC) 156 Fig.5.6 S-N curves of AW and BT specimens and JSSC 157

Fig.5.7 S-N curves of AW and BT specimens and AASHTO 159

Fig.A.1 S-N curves of AW specimens with ratios of crack aspect of 0.3,

0.4 and 0.5, and initial crack size of 0.05mm and 0.06mm 169 Fig.A.2 S-N curves of AW specimens with ratios of crack aspect of 0.3,

0.4 and 0.5, and initial crack size of 0.07mm and 0.08mm 170 Fig.A.3 S-N curves of BT1 specimens with ratios of crack aspect of

0.3, 0.4 and 0.5, and initial crack size of 0.05mm and 0.06mm 170 Fig.A.4 S-N curves of BT1 specimens with ratios of crack aspect of

0.3, 0.4 and 0.5, and initial crack size of 0.07mm and 0.08mm 171

LIST OF PHOTOS

Photo 2.1 TIG dressing equipment and a partially dressed weld 37

Photo 2.2 Burr Grinding equipment 38

Photo 2.3 Disc Grinding 39

Photo 2.4 Hammer Peening equipment 41

Photo 2.5 Needle Peening (equipment and operation) 42

Photo 2.6 Steel shot used for shot peening 44

Photo 2.7 Shot peening equipment 45

Photo 2.8 Shot peening treatment 45

Photo 2.9 Set of the changeable working heads 47

Photo 2.10 Ultrasonic Impact Treatment (UIT) 48

Photo 2.11 Steel grit and shot used in blasting 50

Photo 2.12 Shot blasting machine 50

Photo 2.13 Blast operation 51

Photo 2.14 Blast room 51

Photo 3.1 Fatigue test set up 57

Photo 3.2 As-welded beam test setup 66

Photo 3.3 Set up of measurement of nominal and hot spot stresses 69

Photo 3.4 Detail of strain gauges for measurement of

nominal and hot spot stresses 76

Photo 3.5 Fracture section of AW and BT1 specimen of fatigue test 1 with Δσ = 150 MPa 89

Photo 3.6 Fracture section of AW and BT1 specimen of fatigue test 2 with Δσ = 135 MPa 90

Photo 3.7 Fracture section of AW and BT1 specimen of fatigue test 2 with Δσ = 135 MPa 91

Photo 3.8 Crack propagation rate of crack 1 at

N = 1,451,370 cycles, Σb = 7 mm 92

Photo 3.9 Crack propagation rate of crack 1 at N = 1,731,620 cycles Σb = 11 mm 93

Photo 3.10 Crack propagation rate of crack 1 at N = 2,173,500 cycles Σb = 26.5 mm 93

Photo 3.11 Stop testing fatigue and drill the holes at the ends of crack 1 94 Photo 3.12 Crack after drilling holes 94

Photo 3.13 Fracture surface from as-welded beam 95

Photo 4.1 Overall view of meshing out-of-plane gusset specimen 103

Photo 4.2 Close-up of meshing out-of-plane gusset specimen 104 Photo 4.3 Applied loading on specimen 104 Photo 4.4 Stress distribution in as-welded specimen 105 Photo 4.5 Stress distribution in blast-treated specimen 105 Photo A.1 Fracture section of AW specimen of fatigue 1 with Δσ = 237.5

MPa 164 Photo A.2 Fracture section of AW specimen of fatigue 1 with Δσ = 100

MPa 164 Photo A.3 Fracture section of AW specimen of fatigue 1 with Δσ = 195

MPa 164 Photo A.4 Fracture section of BT1 specimen of fatigue 1 with Δσ =

237.5 MPa 165 Photo A.5 Crack propagation rate of crack 2 at N = 1,731,610 cycles, Σb

= 2 mm 165 Photo A.6 Crack propagation rate of crack 2 at N = 2,134,470 cycles, Σb

= 5.5 mm 167 Photo A.7 Crack propagation rate of crack 2 at N = 2,364,880 cycles, Σb

= 8 mm 167 Photo A.8 Crack propagation rate of crack 2 at N = 2,055,890 cycles, Σb

= 5.8 mm 168

Photo A.9 Crack propagation rate of crack 2 at N = 2,134,470 cycles, Σb

= 12.7 mm 168 Photo A.10 Crack propagation rate of crack 2 at N = 2,173,500 cycles, Σb

= 14.7 mm 169

SYMBOLS

a depth of a surface crack or semi length of a through crack

a 0 initial depth of a surface crack

a f crack size at failure

b half of length of crack

C constant

F e correction factor for shape of crack

F g correction factor for stress gradient

F s correction factor for surface crack

F t correction factor for finite thickness and width of plate

F general correction factor

K stress intensity factor

K max stress intensity factor caused by σ max

K min stress intensity factor caused by σ min

∆K threshold stress intensity factor range

K t stress concentration factor

m exponent of S-N curve or Paris power law

n number of stress range cycles per truck passage

t thickness of plate

R stress ratio

σ max maximum stress

σ min minimum stress

σ r residual stress

(ADTT) S single-lane ADTT

(ΔF) TH constant amplitude fatigue threshold

∆σ nominal stress range

∆σ eff nominal effective stress range

AW notation for as-welded specimen

BT1 notation for one-time blast-treated specimen

BT2 notation for two-time blast-treated specimen

BT3 notation for three-time blast-treated specimen

BT4 notation for four-time blast-treatment

DEFINITIONS

Allowable stress range The allowable stress range which is

corrected to the basic allowable stress range considering the effects of mean stress and plate thickness

Basic allowable stress range The allowable stress range for a given

stress cycles, which is determined according to fatigue design curve

Basic allowable stress range The basic allowable stress range for

2 million stress cycles at 2x106 cycles cycles

Constant amplitude loading A type of loading causing a regular

stress fluctuation with constant magnitudes of stress maxima and minima

Crack propagation rate Amount of crack tip propagation during

one stress cycle

Crack propagation threshold Fatigue strength under variable

amplitude loading, below which the stress cycles are considered to be non-damaging

magnitude required to cause fatigue

failure of the component

amplitude loading corresponding to infinite fatigue life or a number of cycles large enough to be considered infinite by a design code

Fatigue resistance Structural detail’s resistance against

fatigue actions in terms of S-N curve or crack propagation properties

Fatigue strength Magnitude of stress range leading to a

particular fatigue life

FAT Each fatigue strength curve is identified

by the characteristic fatigue strength of the detail at 2 million cycles This value

is the fatigue class (FAT)

crack may initiate due to the combined effect of structural stress fluctuation and the weld geometry or a similar notch

surface at a hot spot (also known as geometric stress)

macro-geometric effects, concentrated load effects and misalignments, disregarding the stress raising effects of the welded joints itself It is also referred to as modified nominal stress

Nominal stress A stress in a component, resolved using

general theories, e.g beam theory

Nonlinear stress peak The tress component of a notch stress

which exceeds the linearly distributed

structural stress at a local notch

dependence of fatigue life N on fatigue strength S

Stress intensity factor Main parameter in fracture mechanics,

the combined effect of stress and crack size at the tip region

maximum and stress minimum in a stress cycle, the most important parameter governing fatigue life

algebraic value of the stress in a particular stress cycle

Structural stress concentration The ratio of structural (hot spot) stress

Factor to modified (local) and nominal stress

Effect of Surface Blast Treatment on Fatigue Behaviour of

Out-of-Plane Gusset Welded Joints

Le Van Phuoc Nhan

Department of Civil Engineering The Graduate School Pukyong National University

Abstract

Surface blasting prior to coating has been widely applied in built steel

structures for cleaning forged surface and increasing adhesive property of

applied coating components The primary purpose of surface blasting

application is to ensure a strong mechanical bond between the substrate and

the coating by the enhanced roughness of the substrate material It has been

found that surface blast treatment significantly effects on fatigue behavior of

welded joints However, this effect is not considered in Fatigue Design Codes

In this thesis, fatigue tests were carried out on five types of

out-of-plane gusset fillet welded joints, including as-welded specimen, one-, two-,

three-, and four-time blast-treated specimens (BT1, BT2, BT3, BT4), and

then the effect of the surface blast treatment on the fatigue behaviour of the

welded joints was studied The radius of weld toe of BT1, BT2, BT3 and

BT4 specimens increased 15%, 21%, 30% and 39%, in comparison with that

of as-welded specimens, whereas the flank angle changed little The tensile

residual stress was induced on the surface of as-welded specimen while the

compressive residual stresses were induced on the surface of blast-treated

specimens The fatigue test results showed that the fatigue life of the

blast-treated specimens is longer than that of as-welded specimens in low stress ranges, even though there is no significant difference in fatigue life between the two types in high stress ranges Over 167% increase in fatigue limit of BT1 specimen could be realized by using surface blast treatment The fatigue test data of BT2 and BT3 specimens are distributed within the confidence lines (mean ± 2s lines) of BT1 specimens This indicates that the fatigue lives of blast-

treated specimens with different blasting times are almost same Compressive residual stress induced on outer layer of specimen is supposed to play important role in increasing fatigue endurance limit of welded joints, while weld toe geometry improvement has minor effect in this process

Prediction of fatigue crack propagation life was performed considering effect of compressive residual stress and redistribution of residual stress due

to crack formation Compressive residual stress will be included if the depth

of crack is still smaller than the depth of compressive residual stress distribution Once crack forms compressive residual stress will relieve and effect of compressive residual stress on fatigue crack propagation life gradually reduce The beneficial effect of compressive residual stress disappears when crack depth exceeds the depth of compressive residual stress distribution The fatigue crack propagation life results obtained with ratio of crack aspect of 0.3 and initial crack size of 0.03mm is nearly approximate with the fatigue test results in high stress range However, there is difference between the fatigue limit obtained by fatigue tests and prediction It is realized that the influence of crack aspect ratio on fatigue crack propagation life of welded joints is not significant

Fatigue crack growth rate was calculated for as-welded specimen and blast-treated specimen (BT1) subjected to stress range of 150 MPa based on formula of Paris Law with modify to account for the threshold The result shows that fatigue crack growth rate of blast-treated specimen is retarded in comparison with that of as-welded specimen This is caused by compressive residual stress induced in blast-treated specimen Fatigue crack growth was also investigated by size measurement of dye marking and beach marks formed by given number of cycles The results indicated that fatigue crack growth of blast-treated specimens were slower than that of as-welded specimen

Key words: steel structures, out-of-plane gusset welded joint, fatigue, fatigue

life, surface blasting

CHAPTER 1

GENERAL 1.1 INTRODUCTION

Fatigue strength of weld joints affects durability of steel structures The fatigue strength of welded joints is very low compared with that of base metals Welded joints are vulnerable to fatigue damage when subjected to repetitive load Fatigue failure may occur even under most in service stress [1] The considerable difference between the unwelded components and as-welded components is the fatigue crack initiation dominated the fatigue life

of unwelded components, while the fatigue crack propagation life occupies most of the fatigue life of as-welded components The low fatigue strength of welded joints is caused by stress concentration [2,3,4] and tensile residual stress [3,4,5,6] The stress concentration and residual stress are unavoidable

in welded joints and they can not be eliminated completely

Weld joints are also considered the weakest points of welded structures subjected to cyclic loading because of the presence of global and local stress concentration, and tensile residual stress Fatigue strength improvement is one of the first considerations in design, built and maintenance of steel structures Many methods have been found out and applied in welded structures in order to increase their fatigue strength The fatigue strength improvement methods of welded joints mainly focus on extending the fatigue crack initiation life Post-weld treatment methods have been investigated to

improve fatigue strength of welded joints These methods can be divided into two main groups: weld geometry modification methods and residual stress modification methods [7] Modification methods of weld geometry remove weld toe defect or reduce the stress concentration by achievement of smooth transition between the weld metal and parent material [7,8] while modification methods of residual stress induce beneficial compressive residual stress in the area cracks are likely to initiate [7,9,10,11,12,13,14,15,16] Many researches reveal that both methods increase fatigue crack initiation life therefore fatigue life and fatigue endurance limit are increased

Surface blasting prior to coating has been applied in newly built steel structures for cleaning forged surface and increasing adhesive property of applied coating systems Blasting consists of propelling small steel balls or grits against the outer surface of the object to be treated [17] Normally, this

is done to remove something on the surface such as scale, but it is also done sometimes to impact a particular surface to the object being shot blasted The shot can be sand, small steel balls of various diameters, granules of silicon carbide, etc The surface of specimen will be left the roughness and in the impact area, the compressive stress will be developed This thesis mentions surface blasting treatment by steel shot and grit However, the effect of surface preparation by surface blasting treatment on fatigue behavior of welded joints is not considered in Fatigue Design Codes, such as IIW/IIS, JSSC, BS 5400 or AASHTO The weld toe geometry is improved by surface blasting treatment however the improvement of weld toe geometry has a little beneficial effect on fatigue behaviour of welded joins Compressive residual stress is induced by surface blast treatment and is assumed to be the major

factor effect on fatigue strength of out-of-plane gusset welded joints [17,18,19]

The results of longitudinal residual stress measured show that tensile residual stresses were induced at the outer layer of as-welded specimens (AW) while compressive residual stresses were induced at the outer layer of blast-treated specimens (BT) It is known that residual stresses are detrimental if tensile, and beneficial if compressive [14,15,16,20,11,22], and regions of compressive residual stresses retard the rate of fatigue crack growth while tensile residual stresses produce the opposite effect [21,23,24]

A compressive residual stress induced by post weld treatment is beneficial by eliminating the tensile residual stresses and generating compressive residual stresses, which improves fatigue strength of welded structures [25] However, the initial residual stresses are relaxed and redistributed with fatigue crack growth [20,23,26,27,28,29]

Calculating fatigue life of welded joints with take account of the residual stress is a complex task because stress range and stress intensity factor changes correspond with crack growth This calculation is made by the

principle of superposition based on the effective stress σ eff = σ max(min) + σ r

[24] or effective stress intensity K eff = K σ + K r which included residual stress

distribution caused by crack form and growth [20,23,26,27] (where σ max , σ min

and σ r are maximum applied stress, minimum applied stress and residual

stress, respectively and K σ and K r are stress intensity factor of applied stress and residual stress, respectively)

Compressive residual stress was induced at outer layer of blast-treated specimen and supposed distributing in a shallow depth at surface of specimen

The free surface is often a preferred site for initiation of a fatigue crack [25] and the crack can not propagate as long as stress under zero and only propagates in region of tensile residual stress [25,30] Therefore compressive residual stress at surface of specimen will eliminate a part of tensile applied stress and extend fatigue crack initiation life, consequently result in improvement of fatigue life of welded joints To evaluate the effect of surface blasting treatment on fatigue behavior of welded joints, fatigue tests were carried out on five types of out-of-plane gusset fillet welded joints, including as-welded gusset specimens and four types of blast-treated gusset specimens, and then the effect of the surface blasting treatment on the fatigue behavior of the welded joints was studied The fatigue test results showed that fatigue life and fatigue limit of welded joints can be improved by the surface blasting treatment However, in high stress range, the effect of blasting on fatigue behaviour is not significant The stress concentration factor was determined

by experimental formula and finite element analysis, and residual stress was determined by cutting method Numerical analysis has been used to evaluate effect of weld toe geometry improvement and compressive residual stress on fatigue limit Fracture analysis provides a better prediction of fatigue crack propagation life considering effect of compressive residual stress In addition, fatigue crack growth rate of blast-treated specimen is retarded in comparison with that of as-welded specimen due to presence of compressive residual stress

1.2.1 FATIGUE FAILURE OF WELDED JOINTS

Welded joints has had a significantly impact on steel structures Welding

technology is a complex fabrication because welding includes many characteristics and nature which effects on behavior of steel structural behavior, such as residual stresses, imperfection, and stress concentration Failures in engineering structures occur predominately at component connections, even in those structures which have been designed, fabricated, and inspected according to Codes [21,31] Fatigue is the process by which a crack can form and then grow under repeated or fluctuating loading The magnitude of loading required to produce fatigue cracking in a component may be much less than that needed to break the component in single application of load [28]

The fatigue failures in welded structures often occur at welded joints, such as cover plate fillet weld terminations, stiffeners, backing bars, and seams and girth weld toe [1,8,32,33,34,35] Previous observations showed that the fatigue failures of welded structures often happened at welded joints and rarely happened at base metal The reason can be explained by the existence of defects of welding (undercut, notch, misalignment, slag inclusion, porosity, crack-like), stress concentration (weld geometry change), residual stresses

In welded structures subjected to cyclic loading, fatigue design is limited

by poor strength detail The poor fatigue strength of welded joints can be explained in terms of three factors following [8]:

The first factor is geometry stress concentration The severity of which

depends on the type of weld and its orientation with respect to the loading direction [8] The stress concentration is the reason that fatigue crack initiates

at all when the applied stress is less than the ultimate strength of the component Any discontinuity in stressed member introduces a stress concentration That is a region where, locally, the level of stress raised above the average [28] Changes in geometry of component that interest the stress field flow lines cause an increase in stress at locations where the flow lines are disturbed most

Peak stress at root of notch

Peak stress at edge of hole

Nominal stress

Fig.1.1 Stress flow lines and stress concentration

The magnitude of the stress concentration, K t, depends on the geometry of the

discontinuity and is defined as the ratio of the maximum local stress, σ max,

and the nominal stress, σ n, remote from the influence of the discontinuity [21]:

n t

K

σ

σmax

= In welded structures, stress concentration occurs at the toes

of the weld, that is the junction between the base and weld metal components The smooth transition between base metal and weld metal there

is, the stress concentration will result in But if there is abrupt change of

section, the stress concentration will be high [28] Fig.1.1 demonstrates the effect of geometry discontinuity on stress concentration

The second factor is crack-like discontinuity Stress concentration can be

introduced by weld discontinuities The existence of crack-like intrusion, undercutting are main factors influence fatigue behaviour of welded joints

Fig.1.2 shows the stress concentration regions caused by geometric

discontinuities in fabricated joints Fig.1.3 presents various types of weld

discontinuities and cracks that result in stress concentration and may cause the initiation and propagation fatigue cracks According to John M Barsom and Stanley T Rolfe, weld discontinuities may be divided into three categories, these three categories are crack-like discontinuities (cracks, lack

of fusion, lack of penetration, overlap), volumetric discontinuities (porosity, slag inclusions) and geometric discontinuities (undercut, incorrect profile, misalignment) [8]