Dr. R.S.

Parmar
II II
\ ,. \
•. 'j
l"" )
t'-...,
Other Publication from the Same Author WELDING ENGINEERING
WELDING PROCESSES AND TECHNOLOGY
AND TECHNOLOGY
"The aim of this book is to provide comprehensive knowledge
about the vital field of fabrication including welding, b~azing,
soldering, adhesive bonding, surfacing, metal spraying and thermal
cutting. Apart from giving a survey of all the Welding and Allied
Processes it describes in detail the industrially important welding
processes including arc welding, gas welding, resistance welding,
solid-state welding, electron beam welding and laser welding.
The role of automation and robotic welding is also detailed and
guidelines are provided for making selection of a process to weld a
particular material. It also gives information regarding fabrication
under specificconditions like underwater welding, welding in snow, R.S. Parmar
welding under windy conditions and welding in space. Economics of
welding and safety in welding are also dealt at length. B.A., B.Sc. (Mech. EnggJ, M.E. Hans. (Prod. Engg),
Ph.D. (Welding), FIIW, FIE (India), MISME, MISTE
Apart from the normal description of welding processes, solved Professor of Mechanical Engineering,
Indian Institute of Technology,
numerical problems is a unique feature of this book and same is the New Delhi
.case regarding Practical Exercises which can be introduced for the
benefit ofB. Tech. and M. Tech. students. Indian, British and AWS
Standards are also listed for ready reference in actual fabrication.
Addresses of Education and Research Institutes, User
Organil?ations and also Equipment & Consumable Manufacturers
concerned with Welding and Allied Processes are also listed. L~', . ~ .~/ ;. .•..• :li

-
-, .•.J / '.'~ , ." ."' .•.... L'P
,. •....•.. ~, .• ~ S •.

It is expected that the book will be useful for not only the
Engineering students and Research Scholars but also the Practicing
'."i/\ i~'~l~~,~,=
Engineers. .. .1\ <\. r- A~.'" I,
~ .. L- ~-.:;L.-~ ..• ~._. .~':.r

KHANNA PUBLISHERS
2-B, Nath Market, Nai Sarak,
Delhi-l10006. India
Published by :
Romesh Chander Khanna
for Khanna Publishers,
2-B, Nath Market, Nai Sarak,
Delhi-ll0006. (India)

All Riehts Reserved

{This book or part thereof cannot be translated or reproduced in
any form (except for review or criticism) without the written
permission of the Author and the Publishers.]

7>1l~ie4ti~
~N No: 81-7409-028-2 I
t"
ftt'! 7>1l4J' UJi61l ;41H4J'llllltl
First Edition: 7991
Fourth Reprint: 2005 $"" 'ViI4'! 4n~ 7>41f(J/'tllJ' ;41H4nl'!"t
/IO/'" (J401l IHIl 1H4n,! 1H"lHllnts "6 -+t41'l'inllss
in IH,! t-i61l

Price : Rs. 220.00

Computer Typeset at;

MiotDQ .•,,,,/'icl-
Delhi-51

Printed at: Bright Printers, Phari Bhojla, Delhi-ll0006.

 Preface

This book on Weldi~g Engineering and Technology has been writ-

ten in the light of the suggestions received from a number of faculty
members and practicing engineers to act as a complementary volume
to my earlier book entitled Welding Processes and Technology.
Thinking that I may not be tempted to write once again on some
leftout topics this book has become rather a hefty one and it covers
almost all major aspects of welding engineering and technology.
The treatment of residual stresses and distortion, weldability of
different materials, heavy welded fabrications, etc. is rather lengthy.
An average reader may not be interested to go that deep in these
aspects of the subject. However, the intention is to make this book
useful to no,t only the students but also the fabricators and re~
searchers so the need to give exhaustive coverage to these topics.
With the availability of this book at low price, it is expected this
rather neglected subject of study will acquire its due place in the
engineering curricula ..
The last two chapters cover the very vital and modern tOpIcs of
residual life assessment of the welded components and their failure
analysis as also artificial intelligence and expert systems in welding.
These topics are expected to provoke the reader to go deeper into
" these aspects of study because they can prove to be money spinners
as topics of consultancy work.
The Appendices include a few laboratory exercises for the stu-
dents and the same can be introduced at Undergraduate and
Graduate levels of Engineering Studies for the students to attain
higher level of understanding of this vital .field of technology.
Addresses of major Welding Institutes are also included-as an
Appendix for those who are interested to establish worldwide contacts
with the welding fraternity.
I sincerely wish that this book may come up to the expectations
of every reader who happens to! come across it.
Any suggestions to improve this book will be most welcome.
lIT R.B. Parmar
New Delhi
May 16, 1997
 Contents
\,.~,

1. WELDING AND WELDING PROCESSES 1-61

I
1.1. Classification of Welding and Allied Processes 1
1.2. Cast-Weld Processes 3
1.2.1. Thermit Welding 3
1.2.2. Electroslag Welding 4
1.3. Arc and Flame Welding Processes 5
1.3.1. Seam Welding Processes 5
1.3.2. Arc Spot Welding Processes 18
1.4. Resistance Welding Processes 22
1.4.1. Spot Welding Processes 23
1.4.2. Seam Welding Processes 25
1.4.3. Zonal Welding Processes 29
1.5. Solid-State Welding Processes 33
1.5.1. High Heat Input Processes 33
1.5.2. Low Heat Input Processes 36
1.6. Allied Processes 42
1.6.1. Material Joining Processes 42
1.7. Thermal Cutting Processes 48
1.8. Modes of Welding 52
1.8.1. Manual Welding 53
.•.. 54
1.8.2. Semi-Automatic Welding
1.8.3. Automatic Welding 54
1.8.4. Automated Welding (Flexible Welding System) 55
1.9. Positions In Welding 58

2. HEAT FLOW IN WELDING 62-132

2.1. Temperature Distribution in Welding 62
2.1.1. Temperature Distribution in Arc Welding 63
2.2. Efficiency of Heat Sources 79
2.3. Further Modifications of Temperature Distribution Equations 81
2.3.1. Adams Modification 82
2.3.2. Wells Modification 82
2.4. Heat Flow in Fillet Welds 89
2.5. Heat Flow in Circular Welds 91
2.6. Heat Flow in Resistance Welding 96
2.6.1. Heat Flow in Resistance Spot Welding 96
2.6.2. Heat Flow in Upset Butt Welding 99
2.7. Heat Flow in Electroslag Welding 102
2.8. Heat Flow in Underwater Welding 102
2.8.1. Heat Flow in Wet Underwater Welding 103
 (x) (xi)
...'
195
2.9. Metallurgical Effects Of Heat Flow In Welding 115 3.7.7. A9icular Ferrite
197
2.9.1. Experimental Determination of Cooling Rates in Welding 118 3.7.8. Pearlite
3.7.9. Bainite 197
2.10. Time-'l'emperature-Transformation Diagrams (TT1' Curves) 120 201
3.7.10. Martensite and Retained Austenite
2.11. Continuous Cooling Transformation Diagrams (CCT Curves) 122 ,~ 203
3.8. Microstructure of Multi-run Welds
203
3.9. Effect of Alloying Elements on Microstructure
3. BASIC METALLURGY OF FUSION WELDS 133-212 3.10. Other Microstructural Phases
204
3.1. Introduction 133 3.10.1. Delta Ferrite in Stainless Steel Welds 204
3.2. Weld Metal Zone 135 204
3.10.2. Sigma Phase
3.2.1. General Theory of Solidification of Metals and Alloys 207
3.11. Etchants for Study of Microstructures
(Growth of grains in permanent mould castings) 135
3.2.2. Epitaxial Solidification 145
4. WELDING STRESSES AND DISTORTION 213-350
3.3. Fusion Boundary Zone 150
213
3.4. Heat Affected Zone 152 4.1. Residual Stresses
214
3.4.1. Underbead Zone 155 4.2. Causes of the Development of Residual Str~sses
3.4.2. Grain Refined Zone 156 4.2.1. Effects of Weld Thermal Cycle and Shrinkage on
Residual Stresses 216
3.4.3. Partially Transformed Zone (AI -A3 temperature range) 158
4.2.2. Stresses Generated by Phase Transformation 222
3.4.4. Zone of Spherodised Carbides 160
225
3.4.5. Zone of Unchanged Base Metal 160 4.3. Measurement and Calculation of Residual Stresses In Weldments
226
3.5. Properties of HAZ 160 4.3.1. Destructive Measuring Methods
4.3.2. Non-destructive Residual Stress Measurement Methods 243
3.5.1. Toughne~s of the HAZ 162
163 248
3.5.2. Other Properties of HAZ 4.4. Residual Stresses in Specific Materials and Joints
249
3.6. Heat Treatment and Related Processes in Fusion Welding 164 4.4.1. Longitudinal Residual Stresses in Welds
4.4.2. Transverse Residual Stresses in Weldments 253
3.6.1. Annealing 164
4.4.3. Residual Stresses in Circular Butt Welds 257
3.6.2. Subcritical Annealing 166
259
3.6.3. Quench Annealing 166 4.4.4. Residual Stresses in Flame-Cut Material
3.6.4. Spherodization 168 4.5. Methods of Relieving (or Controlling) Welding Residual Stresses 261
170 262
3.6.5. Normalising 4.5.1. Design Considerations
263
3.6.6. Quench Hardening 170 4.5.2. Material Considerations
172 4.5.3. Preheating
265
3.6.7. Tempering
174 268
3.6.8. Secondary Hardness 4.5.4. Welding procedure
270
3.6.9. Austempering 175 4.5.5. Welding Sequence
271
3.6.10. Martempering 175 4.5.6. Postweld Heat Treatment (PWHT)
273
3.6.11. Stress Relieving 177 4.5.7. Stress Relief By Natural Ageing
3.6.12. Solution Treatment 178 273
4.5.8. Peening
274
3.6.13. Precipitation Hardening or Age Hardening 179 4.5.9. Vibratory Stress Relief
3.6.14. Natural Ageing 181 4.6. Distortion in Welds 276
3.6.15. Strain Age EmbrittJement 182 277
4.6.1. Significance of Material Properties
3.6.16. Temper Embrittlement 183 278
4.6.2. Influences of Welding Processes and Procedures
3.6.17. Graphitisation 185 279
4.6.3. Types of Welding Distortion
3..6.18. Carbide Precipitation and Sensitization 186 4.6.4. Bowing or Longitudinal Bending 322
3.7. Microstructural Products in Weldments 188 4.6.5. Rotational Distortion 326
3.7.1. Delta Ferrite 192 328
4.6.6. Buckling and Twisting
3.7.2. Austenite 192 4.6.7. Welding Suitability Index Based on Distortion 330
3.7.3. Grain Boundary Ferrite 193 4.6.8. Measurement of Distortion 331
3.7.4. Ferrite Side Plates 193 4.6.9. Control of Distortion in WeldmlJlnts 332
3.7.5. Lath Ferrite 195 4.6.10. Correction of Distorted Weldments 341
3.7.6. Polygonal Ferrite 195 4.6.11. Case Studies on Weld Distortions 343
 (xii) (xiii)

7.3.3. Visual Examination for Weldability 435

5. PREHEAT AND POSTWELD HEAT TREATMENT 351-399
352 7.3.4. Component Sampling Tests 435
5.1. Aims of Preheating 7.3.5. Actual Welding Tests 436
5.2. Clip Test 354
7.4. Fabrication Weldability Tests 436
5.3. Methods of Preheating 355 7.4.1. Hot Cracking Tests 436
5.3.1. Flame Heating 356 7.4.2. Cold Cracking Tests 448
5.3.2. Electrical Resistance Heating 357 7.4.3. Weldability Tests Related to Specific Types of Cracking 458'
5.3.3. Advantages and Limitations of Different Preheating 7.4.4. Allied Weldability Tests 462
Methods 360
7.5. Service Weldability Tests 464
5.4, Measurement of Preheat Temperatures 361 7.5.1. Tensile Tests 464
5.5, Preheating Temperatures for Different Materials 362 7.5.2. Nick-Break Test 475
5.5.1. Preheating of Carbon Steels 362 7.5.3. Pillow Test for Resistance Seam Welds 476
5.5.2. Preheating of Low Alloy Steels 363 7.5.4. Bend Tests 477
5.5.3. Preheating Requirements of Aluminium and its Alloys 391 7.5.5. Impact Testing 486
5.6. Postweld Heat Treatment (PWHT) 392 7.5.6. Hardness Tests 488
5.6.1. Examples of Postweld Heat Treatments 393 7.5.7. Fracture Thughness Tests 489
5.7. Preheat and PWHT of Some Specific Steels 398 7.5.8. Fatigue Tests 490
7.5.9. Corrosion Tests 493
7.5.10. Creep Rupture Tests 496
6. CRACKS IN WELDS 400-430
6.1. Introduction 400
8. WELDABILITY OF SPECIFIC l\'lATERIALS 500-658
6.2. Classification of Weld Cracks 401
8.1. Introduction 500
6.2.1. Hot Cracks 401
410 8.2. Weld ability of Carbon Steels 500
6.2.2. Cold Cracking
411 8.2.1. Welding of Low Carbon Steels 501
6.3. Nomenciature, Location and Orientations of Weld Cracks
8.2.2. Weldability of Medium Carbon Steels 503
6.3.1. Weld Metal Cracks 411
8.2.3. Weldability of High Carbon Steels 504
6.3.2. Base Metal Cracks 413
416 8.3. Weldability of High Strength Low Alloy (HSLA) Steels 505
6.4. Factors Contributing to Weld Cracking
8.3.1. Welding of Weathering Steels 507
6.5. Specific Weld Cracks 417
6.5.1. Chevron Cracks 417 8.4. Weldability of Low Alloy Steels 507
6.5.2. Lamellar Cracks 418 8.4.1. Quenched and Tempered Steels 508
422 8.4.2. 9% Nickel Steels 512
6.5.3. Reheat Cracking ~
427 8.4.3. Heat Treatable Low Alloy Steels 513
6.5.4. Stress Corrosion Cracking (SCC)
8.4.4. Chromium-Molybdenum Steels 518
6.6. Case Study 428
8.1-.5. Weldability of Pre coated Steels 522
6.6.1. Material and Recommended Procedure 429
429 8.5. Weldability 4)fStainless St.eels 526
6.6.2. Investigational Procedure Employed
6.6.3. Conclusions 430 8.5.1. Austenitic Stainless Steels 527
8.5.2. Fenitic Stainless Steels 530
8.5.3. Martensitic Stainless Steels 530
7. WELDABILITY AND WELDABILITY TESTS 431-499 8.5.4. Weldability of Austenitic Stainless Steels 530
7.1. Introduction 431 8.5.5. Weldability of Ferritic Stainless Steels 552
7.2. Weldability Assessment 432 8.5.6. Weldability of Martensitic Stainless Steels 554
7.2.1. Thlerance for Metallurgical Defects 432 8.5.7. Precipitation-Hardening (PH) Stainless Steels 558
7.2.2. Thlerance for Operational Defects 433 8.6. Weldability of High Alloy Steels 564
7.2.3. Measurement of Weld ability 433 8.6.1. Weld ability of Nickel-Cobalt Alloy Steels 565
7.3. Weldability Tests 433 8.6.2. Weldability of Chromium-Molybdenum-Vanadium Steels 570
7.3.1. Theoretical Tests 433 8.6.3. Weldability of Maraging Steels 572
7.3.2. Simulated Tests 434 8.6.4. Weldability of Austenitic Manganese Steel 576
 (xiv) (xv)

8.7. Weldabilit.y of Cast Irons 580 701-780

10. WELD INSPECTION AND QUALITY CONTROL
8.7.1. Gray Cast. Iron 580 10.1. Introduction 701
8.7.2. White Cast Iron 581
8.7.3. Malleable Cast Iron 10.2. Visual Inspection and Measurement 702
581 702
8.7.4. Ductile Iron 10.2.1. Equipment
381
10.2.2. Visual Inspection 707
8.7.5. Welding Considerations for Cast Irons 582
10.3. Destructive Tests 709
8.8. Welciability of Nickel Alloys 591
10.4. Non-destructive Tests (NDT) 710
8.8.1. Welding of Nickel Alloys 591
10.401. Liquid-Penetrant Testing 710
8.9. Weldability of Copper and its Alloys 602 10.4.2. Magnetic Particle Testing 716
8.9.1. Weldability Factors 602 723
10.4.3. Eddy Current Testing
8.9.2. Welding of Copper and eu-Alloys 604 10.4.4. Magnetographic Test 729
8.9.3. Post.weld Heat Treatment of Copper Weldments, 609
10.4.5. Radiographic Testing 730
8.9.4. Brazing of Copper and its Alioys 611 10.4.6. Ultrasonic Testing 744
8.9.5. Soldering of Copper and its Alloys 612 ~ ..
\~
10.4.7. Acoustic Emission Testing 761
8.9.6. Other Processes 612 10.4.8. Comparison ofNDT Methods 763
8.10. Weldability of Aluminium and its Alloys 613 10.5. Pressure and Leak Testing 767
8.10.1. Alumini umAlloys 613 10.5.1. Kerosene Test 767
8.10.2. Welding of Aluminium and its Alloys 615 10.5.2. Hydrostatic Pressure Testing 767
8.10.3. GTAW of Aluminium and its Alloys 622 10.5.3. Air Pressure or Pneumatic Testing 768
8.10.4. GMAW of Aluminium and its Alloys 626 10.5.4. Vacuum Testing . 769
8.10.5. Problems Encountered in Welding Aluminium 628 10.5.5. Halide Test 770
8.11. Weldability of Magnesium and its Alloys 631 10.5.6. Helium Test 770
8.11.1. Magnesium Alioys 632 10.5.7. Leak Testing By Radioactive Material 772
8.11.2. General Weldability Considerations for Mg-Alloys 633 10.6. Non-Destructive Testing of Adhesive Joints 772
8.11.3. Welding of Magnesium and its Alloys 634 10.6.1. Defect~ in Adhesive Bonded Joints 773
8.11.4. Problems Associated with Welding of Mg-Alloys 642 10.6.2. NDT Methods Employed 775
8.12. Weldability of Titanium Alloys 643 10.7. Concluding ~emarks 779
8.12.1. Characteristics of Titanium 644
8.12.2. Titanium Alloys 645
8.12.3. Welding of Titanium and its Alloys 646 11. REPAIR AND MAINTENANCE WELDING 781-814
8.12.4. Weldability Rating 648 11.1. Selection of Welding Process for Repair Work 782
8.12.5. Welding Processes and Procedures 650 11.2. 'IYpes of Welding Repairs 783
8.12.6. Problems Associated with Welding of Titanium 655 11.2.1. Repair of Castings and Forgings 783
11.2.2. Repair Welding of Rivetted Structures 794
9. WELD DEFECTS 11.2.3. Maintenance of Railway Carriage Wheels 796
659-700 11.2.4. Filling-in Large Holes 801
9.1. Classification of Weld Defects 659 11.2.5. Maintenance Welding of Earth Moving Machinery and
9.2. General Sources of Weld Detects 661 Mining Equipment 801
9..'3. Arc Welding Defects 662 11.2.6. Repair Welding of Large Structures 811
9.3.1. Surface or Visual Defects 662
9.3.2. Subsurface Weld Defects 674
9.3.3. Acceptance Levels of Arc Welding Defects 682
12. WELD JOINTS, WELD SYMBOLS AND
JOINT DESIGN PRINCIPLES 815-888
9.4. Weld Defects in other than Arc Welding Processes 684 12.1. Types of Joints 815
9.4.1. Resistance Welding Defects 684 816
12.1.1. Angular Limits of Different Joints
9.4.2. Defects in Friction Welding 694
12.2. Types of Welds 818
9.4.3. Defects in Welds of Other Welding Processes 698
12.2.1. Fillet Weld 819
12.2.2. Groove WeIr" 819
 (xuii)

(xui) 13.2.6. Torsional Rejlistance 894

820 13.2.7. Radius of,Gyration (r) 895
12.2.3. Back or Backing Weld 820 13.3. Salient Weld Design Features 899
12.2.4. Flange Weld 820 13.3.1. Arc Welded Joints 899
12.2.5. Plug and Slot Welds 820 13.3.2. Stress Calculations for Welds 900
12.2.6. Spot and Projection Welds 821 13.4. Useful Formulas for Design Under Different Types of Loading 903
12.2.7. Seam Weld 821 13.4.1. Tensile Loading 903
12.2.8. Surfacing Weld 821 13.4.2. Compression Loading 904
12.3. Variants of Butt Joints 823 13.4.3. Bending Loading 906
12.4. Variants of T-joint 823 13.4.4. Shear Loading 907
13.4.5. Torsional Loading 908
12.5. Variants of Lap Joint 824 13.4.6. Shock Loading 911
12.6. Variants of Corner Joints 827
13.5. Solved Example 913
12.7. Variants of Edge Joint 827
12.8. Variants of Weld Types 829
12.9. Combined Welds 830 14. WELD DESIGN FOR FATIGUE I.OADING 933-998
12.10. Selection of Weld Type ~4.L Characteristics of Fatigue Failure 933
833
12.11. Weld Joints for Structural Tubular Connections 837 i4.2. Fracture Appearance 934
838 ~4.3. Fatigue Mechanism 935
12.12. Welding Symbols
12.12.1. Basic Welding Symbols 848 14.4. Types of Fatigue Fracture 938
12.12.2. Solved Examples 858 14.4.L High-Strain Low-Cycle Fatigue Life 939
12.13. Brazing Symbols 860 14.5. S-N Curve Or Wohler Line 941
12.14. NDT Symbols 862 14.6. Factors Affecting Fatigue Life 946
12.14.1. No Side Significance 862 14.6.1. Material Properties 946
12.14.2. Radiographic Testing 863 14.6.2. Joint Configuration 948
12.14.3. Combined Symbols 864 14.6.3. Stress Ratio 951
12.14.4. Number of Examinations 864 14.6.4. Welding Procedure 953
12.14.5. Area of Examination 865 14.6.5. Postweld Treatment 953
12.14.6. Acoustic Emission 867 14.6.6. Loading Condition and Sequence 953
12.15. Principles of Weld Joint Design 867 14.6.7. Residual Stresses 954
12.15.1. General Design Principles 868 14.6.8. Weld Reinforcement Geometry 954
12.15.2. ~asidSpecific Weld Joint Design Principles 878 14.6.9. Environmental Effect 955
878 14.6.10. Weld Defects 957
12.16. Evolving Good Weld Designs
12.16.1. Conversion from Castings to Weldments 882 14.7. Crack Propagation Behaviour ofWeldments Under Constant
12.16.2. Redesign to Avoid Mixing of Weld Types 883 Amplitude Loading 96·i
12.16.3. Designing to Avoid Opening of Notches 884 14.8. Methods of Improving Fatigue Life of Welded Structures 964
12.16.4. Redesign the Whole - Not in Parts 885 14.8.1. Improvement in Design and Welding Procedure 964
12.16.5. Redesigning to Eliminate Unsuitable Members 885
9. 14.8.2. Inducing Surface Compressive Stress 970
12.16.6. Proper strengthening of a Poorly Designed Component 886 980
14.8.3. Protection from Atmosphere
12.17. Importance of Proper and Realistic Specifications 14.8.4. Comparison of Improvement Techniques 981
14.9. Case Studies 982
13. WELD DESIGN FOR STATICLOADING 889-932 14.9.1. Failure of a Steel Pipe Element 983
14.9.2. Failure of a Bali Mill 984
13.1. Designing for Strength and Rigidity 889
13.1.1. Design Formulas 890 14.10. Solved Problems 985
13.1.2. Loading 891
13.2. Material or Section Properties 891 15. FRACTURE TOUGHNESS AND WELD DESIGN 999-1087
13.2.1. Modulus of Elasticity 891
15.1. Types Of Fractures 1000
13.2.2. Rigidity or Stiffness 892 15.1.1. Ductile and Brittle Fractures 1000
13.2.3. Modulus of Rigidity 893
13.2.4. Moment of Inertia 893
13.2.5. Polar Moment of Inertia 894
 (xviii) (xix)

1001 16.5.2. Welding of Off-shore Structures 1180

15.2. Factors Affecting Type of Fracture
1002 16.6. Non-Destructive Testing of Heavy Welded Structures 1181
15.3. Plane Stress and Plane Strain Conditions
16.6.1. Methods and Extent of Examination "1181
15.4.. Modes of Crack Extension 1004
15.5.R~presentation of Plane Stress and Plane Strain Conditions 17. EXPERT SYSTEMS IN WELDING 1183-1200
On Mohr's Circle 1005 17.1. Artificial Intelligence 1183
15.6. Residual Strength 1007 17.2. Expert Systems 1184
15.6.1. Effect of Section Thickness on Residual Strength 1008 17.2.1. The Need for Expert Systems 1184
15.7. Materials and Methods of Determining their Fracture Toughness 1009 17.3. Expert System - Structure and Working Methodology, 1186
15.7.1. Griffith-Irwin Criterion 1011 17.3.1. Knowledge Base 1186'
15.7.2. R-Curve Criterion 1015 17.3.2. Inference Engine 1186
15.7.3. Linear Elastic Fracture Mechanics (LEFM) 1021 17.3.3. User Interface 11th·
15.7.4. Crack Opening Displacement (eon) and Crack Tip 17.4. Types of Expert Systems 1187
Opening Displacement (CTOD) Criteria 1056 17.4.1. Synthesis-Type Expert System 1189
15.7.5. J-Integral Criterion 1071 17.4.2. Analysis-Type Expert System 1190
15.8. Fral;ture Toughness Tests with Impact Loading 1075 17.5. Explanation of Weldex 1192
15.8.1. Charpy V-Notch Test 1075 1195
17.6. Welding Information Network (WIN)
15.8.2. Drop Weight - NDT Test 1076 1197
17.6.1. Knowledge Bases for WIN
15.8.3. Dynamic Tear (DT) Test 1078 1197
17.6.2. Expert Systems in WIN
15.9. Crack Arrest 1080 1198
17.7. Future Developments
15.9.1. Principles and Method of Crack Arrest 1080
17.8. Concluding Remark 1198
15.9.2. Practical Applications of Crack Arresters 1084

16. HEAVY WELDED FABRICATIONS 1088-1182 18. RESIDUAL LIFE ASSESSMENT AND FAILURE
ANALYSIS OF WELDED STRUCTURES 1200-1230
16.1. Boilers and Pressure Vessels 1088
18.1. Residual Life Assessment of Welded Structures 1201
16.1.1. Material Selection 1088
18.1.1. Involvement of External Agencies in FFS and RLA 1202
16.1.2. Fabricatipn of Copventional Pressure Vessels 1098 1203
16.1.3. Fabricatibn of Clad Pressure Vessels 1111
18.1.2. Development of Expertise on FFS and RLA
18.1.3. Justification for FFS and RLA Studies 1203
16.1.4. Weldabili'ty Aspects of Pressure Vessel Steels 1111
16.1.5. Preheat and Postweld Heat Treatment of Pressure Vessels1116 18.1.4. FFS and RLA in Presence of Service Induced Defects 1204
16.1.6. Aluminium Pressure Vessels 1125 18.1.5. Inspection Techniques Applied for FFSIRLA Studies ~206
18.1.6. Analysis of Available Data on Plant History ;1206
16.2. Shipbuilding 1130 11207
18.2. Failure and Failure Analysis
16.2.1. Ship Structure 1130. 1208
18.2.1. Types of Failures
16.2.2. Materials for Shi.pbuilding 1141 18.2.2. Causes and Classification of Failures 1209
16.2.3. Welding Processes for Shipbuilding 1147 1212
18.2.3. Procedure for Failure Analysis
16.2.4. Weld Joints' in Sltipbuilding . 1156 18.2.4. Some Case Histories ofWeldment Failures 1216
16.2.5. Welding Sequence 1158
16.2.6. Automatipn in Shipbuilding 1160 Appendices 1231-1254
16.2.7. Uoit anqlBlock Construction of Ships 1161 A. Weld Quality Test 1~31
16.3. Pipe Welding 1166 B. Fracture Toughness Data 1238
16.3.1. Pipe St¢els and Electrodes 1166 'C. Laboratory Exercises 1241
16.3.2. Types qf Pipe Joints 1168 D. Chemical Elements, Their Symbols and Atomic Numbers 1248
16.3.3. Types 6f Welding 1169 E. Greek Alphabets 1250
16.4. Welded Bridges 1175 F. SI Units and Symbols 1251
16.4.1. Materials for Welded Bridges 1175 G. International Welding and Allied Organizations 1253
16.4.2. Welding of Bridges 1177
16.5. Off-Shore Structures 1178
References 1255-1261
16.5.1. Materials for Off-Shore Platforms 1179 Index 1262-1270'
 ONE
Welding and Welding Processes
Welding is a process ofjoining two or more pieces of the same
or dissimilar materials to achieve complete coalescence. This is
the onlymethod of developing monolithic structures and it is often
accomplished by the use of heat and or pressure. Although in its
present form it has been used since about the beginning of 20th
century but it is fast replacing other joining processes like riveting
and bolting. At times it may be used as an alternative to casting.
Presently welding is used extensively for fabrication of vastly
different components including critical structures like boilers and
pressure vessels, ships, off-shore structures, bridges, storage
tanks and spheres, pipelines, railway coaches, anchor chains,
missile and rocket parts, nuclear reactors, fertiliser and chemical
plants, structurals, earth moving equipment, plate and box
girders, automobile bodies, press frames and water turbines.
Welding is also used in heavy plate fabrication industries, pipe
and tube fabrication, jointing drill bits to their shanks, automobile
axles to brake drums, lead wire connections to transistors and
diodes, sealing of containers of explosives like nitroglycerine,
welding of cluster gears, and the like.
1.1. Classification of Welding and Allied Processes
Although almost all materials (metals, plastics, ceramics, and
composites) can be welded but not by the same process. To achieve
this universality a large number of Welding and Allied Processes
have been developed. Most of the industrially important processes
amongst them classified depending upon the nature of heat source
and its movement resulting in spot, seam or zon,alwelds; or on the
extent of heat generation viz., low heat and high heat, are shown
in Fig. 1.1. This rather unusual type of process classification has
been chosen because often these processes will be referred to
accordingly in the remaining text ..
2 Welding Engineering and Technology Welding and Welding Processes 3
~III 0U~'"
",c
Sd -~ -ou .S
UJ.o
c'"
o..u
cS
a$J~
c-
~g
E
~.~
c,,-
sO::J
Q'l'"
utloI ..0...JU
(ge
-" Brief description and important uses of these processes are
:90'" ec
.~ V
"0 ~.~
C::VI
01'" :§
-.;
given in the following sections.
,~~
.•...
v
.•...
0
::l:
1.2. Cast-Weld Processes
::;J'-
va. g
These processes involve large amount of molten metal
:e g' V\ \-

~.~
<-01
2!< ."
a::: E::J
.••.• CIlI V
-
S.
~ 0'1
.fj
~ c Ccn e resulting in properties close to that of castings. For achieving the
",::;J
o~c ,~c 0' "g g' tloI tloI.~ ~ .S
",.~g::g=>-u=
.c;.0
~'-" .~~·2 ~:~ ~~ 6~eo.2~do~ desired joint properties such welds are usually given
VI
'" VI ~~~
u.oa: Ou::l:a normalisation treatment. Two processes in this class are Thermit
VI
VI
.E ?A
Welding and Electroslag Welding.
~
o
~ CTI g' trI

<-
a..
L-
a ~'- '0
,§ .~
d :6g~~-e
.~~ is] c;'~ o
c
1.2.1. Thermit Welding
"0
r. :;:
.-g'Cl.
01 ~
1 '" ~'" <aJ~~~~ .~
•.
~ Thermit is a mixture of aluminium powder and metal oxide
c:: c:: a.
'"
:~~ ~
" E
~ VI
u
o rn which when ignited results in a non-explosive exothermic reaction.
- 0
.Eo. ~ c ~
~'" '7 CII
UJ
'"
'" "'c
0", 0'" UJ
The heat so generated meltfi and reduces the metal oxide to
'" '"
-~
.~ O'l

LO ~~
o..c E'S
~." 8
c~ - •.
53:
E:§
o~
~"i u3:
-0:.0
•• C
.1::0
••• a:l £ metallic form at a high temperature. This molten metal is used
g' ~': for joining metal parts by pouring it between them resulting in
.;:;r ~a.
"0

.~ what may be termed as a cast-weld joint .

Qj lit . ....r.51 trI
;;j
One of the most used thermit mixtures is aluminium powder
~e ~ gg> ~ g",
"0

c::

3~1'" ~ and ferric oxide which on ignition results in the following reaction.
<Il
VI b./)
'" c::
VI
VI
:.a 8Al + 3Fsa04 ~ 9Fe + 4Al20a + 3310 kJ/mol ...(1.1)
'"
v
o<-
~ ~lgf ~

I~J: ! ~J
:x:S ,5
a.. ~r ~.s ~.s ~
•....
The molten metal obtained has high temperature of about
o
'" u; '!5 !1l:2 ~." 2450°C. This metal is poured into the sand mould surrounding the
"0
'" j~ I ]~"",.2;
_~~ C1:!2\ &! a:l d!. ,3:
-.;
.S
...,
c::

parts to be welded, as shown in Fig. 1.2. The mould is broken soon

~ ~ <Il
tl.I
U -
0 '';:
U S
'OJ
after the solidification of metal is complete and the excess metal
'"
c
C
.~,g g' .~ g' -g 0\
'" " UJ
is removed, by chisel and hammer, to give the necessary shape to
01 UJ
c:: t:.~~ a::~ -::6 <Il
the weld.
VI
'i3
- VI
~j1~j
E~I Ei; ~~~~~ ~j U
'" '"
'"
VI 3 VI ....;
oc ...•
VI
VI
'"
'" vc:: v'"0
ov <-
--0 .- . Clr-
Q'l ',.. O'l
l:liJ
<-
a..
01
~o.. 0_ --0 ._'0 ~
01 Vi
'"
~r"'I &rC £j
U C
c:: a:::
§ '"
'"
~ ~ ~ .,,~ Crucible with
'" thermit mixture
~~ ,h '"
u.VI
VI -:9\ GI ~I ~;I C

"0'" c:~ ~Eg'~

c:: V
CO ~~~~~2'§~
Tapping pin
- <-
Eo..
~g'
lD• .-
"0
~
~ "2~ ~
~:s1v
fa.cW:p-:J
-g
~ §
~ -; U III ~
j
~ao:X~.&IW-
0 ~
~
g'
~
5~ J ~~
~'~ ~ ~:§
Slag basin
Riser
'-v-'" 0 •... 0- .Q .!:! _tloI O)(ClI
«~ 0:9 I I I I I fl. I I I I Section to
E"'lu"'E"''''''''''''''
.,;~ 0.. 0.. UJ UJ3: ...• o3:
be welded
Collar Heating gate
Qj ~ . ~
..!. \oJ E:.o -.= pattern
.!'ii,.
VI E
80..
3:;:
"0 VlULJ'"
.•...'!::~ iiil
e '" 1i::!2

Fig. 1.2. A standard set-up for thermit welding.

4 Welding Engineering and Technology I Welding and Welding Processes 5
Thermit welding is extensively used for joining rails at site, Due to high heat input the weld pool in ESW is usually quite
cable conductors, reinforcing bars for R.C.C. structures, and for voluminous resulting in a weld with properties resembling that of
heavy repairs such as those of broken necks of rollers, and ship a casting which makes post weld heat treatment (PWHT) essential
i'.•terns. t.o achieve the desired metallurgical structure to attain the
1.2.2. Electroslag Welding required mechanical strength.
Electrosiag welding (ESW) is a fusion welding process for Typically ESW is extensively used in the construction of
joining thick workpieces in a single run. This is not an arc welding pressure vessels, press frames, water turbines and hea'vy plate
process though most of the set-up is similar to the usual arc fabrication industries.
welding processes and arcing is required to initiate the process 1.3. Arc and Flame Welding Processes
and may also occur subsequently when the process stability is The welding processes included in this class are those which
disturbed. Apart from the conventional ESW process in which the make use of an electric arc or a flame obtained by burning an
usual electrode wire with contact tube is employed there is a oxy-fuel gas mixture. The size of the weld pool evolved depends
popular variant of the process called Consumable Guide ESW upon the energy input per unit time, and the extent of spread of
Process; Fig. 1.3 shows the process diagram for the latter. An arc or flame, however the volume of the molten metal in the weld
essential feature of the ESW process is that the welding is done pool, at any given time, is much smaller than that obtained in
with weld joint in vertical position. electroslag welding. These processes are either used to produce
welds along seams (e.g., SMAW,GMAW,Oxy-fuel gas welding, etc.)
Wire
reel or just at spots (e.g., stud welding, GTAWspot welding, etc.). Brief
descriptions of industrially important processes in this class
Control panel follows.
1.3.1. Seam Welding Processes
Wire feed motor These processes are mainly used for welding workpieces along
Wire feed straight or curved seams of desired lengths and include all well
drive
known arc and flame welding processes.
1.3.1.1. Carbon Arc Welding
In carbon arc welding heat is produced with an arc between a
carbon electrode and the work, and normally no shielding gas is
Molten slag bat h used. The heat from the arc melts the work material and filler
wire, if required. Fig. 1.4 shows the basic circdit for carbon arc
Pow~r source
Molt~n weld welding. Elcctrodc
pool holdcr
Retaining shoe

./ Water in

,~
Carbon clcctrodc DC
~ Powcr
sourcc
Ground cablc

"
Water out T +
Fig. 1.4. Carbon arc weldmg.
Work (ground) lead To avoid €xcessive heating and consequently accelerated
consumption of carbon electrode it is required to use dc (direct
Fig. 1.3. A set-up for consumable-guide electroslag welding. current) power source with electrode negative.
6 Welding Engineering and Technology Welding and Welding Processes 7

The weld pool produced is normally small and therefore in its protective media to ward off the atmospheric gases from reacting
manual mode carbon arc welding process can be used as an all- with the molten metal pool. The electrode wire diameter may
position welding process. range between 2 and 10 mm. Both ac and de power sources are
Carbon Arc welding can be used for welding copper since it used though de with electrode positive (dcep) is the preferred
can be used at high current to develop the high heat usually choice.
required for the purpose. It can as well be used for welding SAWis mainly used in the downhand welding position in both
automatic and semi-automatic modes. The former is. a more
galvanized steel and repairing of steel castings.
popular mode and a set-up for the same is shown in Fig. 1.6.
1.3.1.2. Shielded Metal Arc Welding WirlZ flZlZdmotor Flux h opplZr
This process employs coated or covered electrodes for
producing an arc to act as a heat source; the covering on buming
provides the necessary shield to protect the molten metal from the
ill effects of oxygen and nitrogen from the surrounding
atmosphere. This process is more popularly known as Stick wirlZ
electrode welding or manual metal are welding and is the single
most used weldine- process in the world. Both ac and dc power
sources can be used equally effectively. Fig. 1.5 shows the basic
circuit diagram for the process.
AC or dc powlZr sourc~ EllZctrodlZholdlZr
and controls

Weld mlZtc.1depth 'Arc'

" . 5ubmc:rglZdarc
Fig. 1.6. Essential elements of an a.utomatic submerged arc welding unit.
The weld joint produced by submerged arc welding is of very
Work high quality and consequently this process finds extensive use in
cabllZ
joining thick plates in long, linear seams as are encountered'in
ships, pressure vessels, bridges, structural work, welded pipes,
and nuclear reactors. '
EllZctrodlZ cabllZ
Fig. 1.5. Basic circuit for shielded metal arc welding. 1.3.1.4. Fusarc Welding
The weld pool produced depends upon the size of the covered This process employs flux coated electrode in which the core
wire is helically wrapped with both left and right handed spirals
electrode and the welding current used and may vary from very of wire shown in Fig. 1.7; the coating fIlls the spaces between the
small to fairly large size. Larger sized pools are used only for spirals. Current to the core wire flows through the contact between
downhand welding. This process is an all-position welding process the contact tube and the outer wire spiral which is partly bare.
and is used for all types of jobs. All metals for which covered Welding current from 200 A to 1000 A can be used depending upon
electrodes are available can be welded by this process. Because the electrode diameter, however, the upper limit for the current is
this is a very versatile process so it is still extensively used in the also set by the ability of the outer. spiral to carry it without
fabrication of ships, bridges, pressure vessels, and structurals; Core First auxiliary 5econd auxiliary Flux
however it is used in its manual mode only. wi rlZ wirlZspiral wirlZ spiral covlZring
1.3.1.3. Submerged Arc Welding
Submerged arc welding (SAW) is a process in which
continuous copper coated spooled wire is used in conjunction with
}Dosegranulated flux poured ahead of the arc so as to provide a
(-----·~31 ~.;~'"
Fig. 1.7. Constructional features of a continuous covered electrode.
 8
----------Welding Engineering and Technology ~elding and Welding Processes 9
overheating and collapse. Long current slide is often used to Gas tungsten arc welding (GTAW)or tungsten inert gas (TIG)
overcome this difficulty. The set-up for Fusarc welding resembles welding employs a non-consumable tungsten electrode with an
the set-up used for automatic submerged arc welding excluding' envelope of inert shielding gas (Argon, helium, etc.) to protect both
the flux supply and recovery system. the electrode and the weid pool from the detrimental effects of
Fusarc welding usually employs an additional shield of CO2 surrounding atmospheric gases.
to enhance the protection of weldpool and thereby greatly Both ac and de power sources are used for GTAW.The tungsten
improves the weld quality.
electrode employed varies in diameter from 0.5 to 6.5 mm and. the
Fusarc welding is more tolerant of joint fit-up, surface and current carrying capacity varies accordingly between 5A and
weather conditions then other open arc processes induding 650A. The welding torch used for carrying current higher than
submerged arc welding. It, therefore, finds extensive use in 100A is normally water cooled. The process is used mainly in its
shipbuilding, structural work or any long straight or circum- manual mode. Fig. 1.8 shows a schematic representation of the
ferential seams. It also finds use for double-sided fillet welds and basic elements of an ac GTAWunit.
pressure vessel fabrication requiring sound welds and good
penetration where visibility of the arc is imperative. GTAWis an all-position welding process and gives the highest
1.3.1.5. Gas 'I\mgsten Arc Welding quality welds amongst the commonly employed arc welding
processes and is, therefore, extensively used for welding most of
the industrially useful metals and alloys usually in thin grades.
Aircraft industry, rocket and missile fabricators, chemical and
nuclear plant fabricators are the typical user industries of this
process.
...; l.3.1.6. Plasma Arc Welding
'0
::l
Plasma is a flow of ionised gas that is obtained by passing a
••
u
..
::J
o
o ® ~
:::
gas through a high temperature arc which results in splitting the
gas molecules to atoms and then to ions and electrons.
.. o
III t.:l
'"
••

l o @ •...
o
c:
.51
...
'"
In plasma arc welding the arcis created between a tungsten
electrode and the workpiece, as in gas tungsten arc welding.
However, the plasma arc is constricted by an outer nozzle through
...
c:
<l.l which the shielding gas flows.
~
c.. Power source used for plasma arc welding is invariably of
l!::
constant current dc type with an open circuit voltage of 70 to 80
volts and a duty cycle of 60%.
.!:!
....
'"
S
c ii
<l.l
There are two variants of the plasma arc welding process
called non-transferred' type and transferred type. In the non-
~..
'"
- 0" j.".-f
u.
•• 8
u..

~..
•.. ~.- -<
u..E Co
oci
transferred type the tungsten. electrode is the "cathode and the
.r.
•••• u ....• torch tip the anode. Such a torch is very similar to oxy-acetylene
l:liJ

i:i:
torch as regards its manoeuvrability as workpiece is outside the
.~u electrical circuit. However, such a plasma arc is less intense
compared with the transferred arc wherein the workpiece is the
anode. The manoeuvrability of the transferred arc is, however,
restricted. But such an arc is very intense and therefore the
process results in higher thermal efficiency with consequential
higher deposition rates as compared with GTAW,Fig. 1.9. Fig. 1.10
shows the set-ups for two modes of the plasma welding arc.
10
Welding Engineering and Technology Welding and Welding Processes 11
2.5
Any gas that does not attack the tungsten electrode or the
" copper nozzle tip can be used for plasma welding. However, argon,
'-
2.0 "" and argon-hydrogen mixture are more commonly employed.
-'=
'- .I"
"
.I"
.I"

"
"
.I"
A major disadvantage of plasma arc welding process is the
C'I
~. 1.5 "
.1".1" "

".I" Plasma arc noise due to the operation of the plasma source. Because of this,
/ " ,./ ,," welding hand-held plasma torches are used to a very limited extent. For
"/ ""
QI
.•..
'-Cl most part, remotely operated plasma sources. are employed.
/ /

" .I"

/ " "/
c: 1.0
a Commercially the major users of plasma welding process are
+= ./ " the aeronautical industry and jet engine manufacturers. Typically
/ /" " /
,,
a
VI

" the process is used for making piping and tubing made of stainless
c..
" steels and titanium.
~ 0.5 "
1.3.1.7. Gas Metal Arc Welding
o In gas metal arc welding (GMAW)process a consumable wire,
o 100 200 300 400 500 600
of 0.8 to 2.4 mm diameter and wound in spool form, is fed at a
Welding C lJrrent , Amps _ preset speed through a welding torch wherein it is provided the
electrical connection and the shielding gas. The arc which is struck
Fig. 1.9. by direct contact between the wire electrode and the workpiece, is
maintained at a constant length by the interaction of electrical
parameters. The power source used is invariably of the rectified
dc type. Both, constant voltage and constant current type power
sources are in use.
Depending upon the work material, the shielding gas may be
High - argon, helium, nitrogen, carbon dioxide, hydrogen, and their
frequency mixtures. When inert shielding gas is used the process is more
supply,--"
Powe r f dc pilot arc
+ starter popularly known as MI9 (metal inert gas) welding and when
CO2 is used as the shielding gas it is referred to as CO2 welding
or MAG (metal active gas) welding.
A node
(a) Transferred arc
GMAW is an all-position semi-automatic welding process
though its automatic versions are also available. A set-up for
Cathode gas
se:mi-automatic GMAWprocess is shown in Fig. 1.11.
GMAWis a very versatile process and can be used for welding
Cooling all metals for which compatible filler wires have been developed.
Power water
sUPply However, its typical applications include medium-gauge fabri-
cation such as structurals, earth moving equipment, plate and box
W-electrode
girders, and automobile bodies. This process has great potentials
Plasma for use with robotic welding systems.
- jet
Anode i1iit
,/III
_ -1.3.1.8. Plasma-MIG Welding
""
This process, as the name implies, has been developed by
~Pi~ combining the features of plasma arc welding and MIG welding
(b) Non -transferred arc processes. It has two variants; one with separate non-consumable
Fig. 1.10. Two modes of plasma arc welding.
tungsten electrode and the other uses the torch nozzle as non-
 12
Welding Engineering and Technology Welding and Welding Processes 13
Fill~r wirt~
Tungst~n
~l~ctrod~ Control tub4/:
H. F. Pow~r
sourcli!

-0~Cl
171
:2
."
III
VI u
c•••
"; "::~
...,
..,
UJ
c.Cl
<:l
....,
<ll
...•
s::
OJ
...;
ff!
i:i:
bil
171

~d
c
'-1lI

~'i
'-
c.?
.S
,_\.
::s ~8.~
0 C..!::

<r: 'fi
OJC
<ll
r:r.,

'0 'C
::J
•..
S
~
en

(0 )
r Plasma
SrClding

~N0%2I'

Bright
gas
gas

vapour jet
Work pi~c~

Contact t ub~

[Plasma gas

~
~Ydln9 9a'
Nozzl~

Work pj~c~
(b)
Fig. 1.12. Plasma-MiG welding torches: (a) with separate non-consumable
electrode: (6) with nozzle as non-consumable electrod.e.
600 6
.r:.
.~ Figures indicate ~k!ctrod~55 66
- stickout in mm 65 C7l

C7l ..j:-~ 55
"'"
~
24 ~
~o,' 055 2
consumable electrode. The essential features of torches used for c: ~C>~ c
these two types are shown in Fig. 1.12. -o 200
'j;;
~ "'048
30 •• MIG
.8
12 'iij
Essentially plasma-MIG welding process differs from the oC. 8.
III ~oo
63 3~ 3330 2t 2~~4
•••
Cl
existing GMAWprocess in that the electrode wire is enveloped in Cl 35 30 28
a plasma sheath which controls heat and droplet transfer in such o Welding cl,Jrr~nt,A
a way that much higher speeds and deposition rates are attained 200 300 400 500 600
than possible with GMAWprocess, as is shown in Fig. 1.13. Welding current vs. d~position rat~
. This process can be used both for welding and surfacing. Most Fig. 1.13. Comparison of deposition rates for MIG
of the materials that can be welded by GMAW can as well be and plasma-MIG welding processes.
welded by this process and at much faster rates. EGW is an arc welding process and gives welds with properties
1.3.1.9. Electrogas Welding close to those obtained by submerged arc welding.
The equipment used for Electrogas Welding (EGW) is similar Electrogas welding uses the vertical orientation of the weld
in appearance to the one used for electroslag welding. However, joint and employs copper 3hoes for retaining the molten metal in
shape at the end of the plate width as in electroslag welding.
 15
14 Welding Engineering and Technology Welding and Welding Processes
.
1.3.1.10.Electron Beam Welding
However, the wire used in EGW is of the flux-cored type which
provides minimal covering to the weld pool. Additional protection In electron beam welding (EBW) a beam of electrons is used
is normally provided by the use of CO2 or argon-rich shielding gas. to melt the metal for welding. The electron beam; emitted from a
The rating of the equipment is similar to that of gas metal arc heated filament, is focussed on to the desired spot on the workpiece
welding equipment. The duty cycle of the power source, however, Oil cooled cathode
insulator
needs to be 100% as it is a continuous operation. The essential
features of a set-up for electrogas welding are shown in Fig. 1.14.
Winz
spool

Drive
motor Fud rolls

Shielding gas Power

source

~Isolated valve
actuator

b{GlaSS and metal

vapour shield
piece nctuator
Starling
Weld

L:.Xd Copper shoes

(a) Set-up • vacuum

Chamber
port

Slog wire

o
Water Weld pool Fig. 1.15. Essential features of an electron beam welding unit.
surface with the help of a focussing coil. The workpiece which is
CO2 Weld metol placed in a vacuum chamber can be moved to create the necessary
Water welding speed.
The penetrating power of the electron beam depends upon the
( b) Process details speed of the electrons which is controlled by the magnitude of
Fig. 1.14. Essential features of a set-up and process details tor electrogas welding. accelerating voltage. Depending upon the accelerating voltage the
Electrogas welding process is mainly used forjoining metals with EBW guns are rated as low voltage and high voltage· types with
a thickness or 12 to 75 mm; more on the lower range. Typically EGW the range of voltages between 15-30 KV and 70-150 KV
is used in shipbuilding, and site fabrication of storage tanks.
16 Welding Engineering and Technology
respectively. Fig. 1.16 shows a schematic representation of a triode
type EBW unit.
The EBW welds are very narrow and can be of the full
penetration type with width to penetration ratio of 1 : 20 compared
with 5 : 1 of shielded metal arc welding, and 2 : 1 of gas metal arc
welding. The energy density of electron beam (EB) being nearly
5 x 108 W/mm2 it is, therefore, possible to melt and weld any
known metal. Due to high energy density of the EB the HAZ is
extremely narrow and high welding speedR can be reached.
Electron beam welding is widely used in the electronics,
n uelear, missile, and aircraft industries. Typical applications of
the process include cluster gears, intricate valve arrangements
rnade of corrosion resistant alloys for automobile industry as well
as preRsure capsules, and missile hull frames. A portable EBW
unit has also been developed for inflight repair welding of
satellites.
1.3.1.11. Laser Beam Welding
In laser beam welding a monochromatic (of one wavelength)
coherent light beam is used as a heat source. A coherent light i~
one in which the waves are identical and parallEI w.biehcan travel
a long distance without loss in intensity or deviation. Laser light
can be ea&ilyfocussed without any decrease in intensity to a very
small s;ot giving a very high energy density which may reach 109
'\IV/mm. Thus, a laser beam like an electron beam, can weld any
known material. Due to very high energy density the HAZ is
extremely narrow and high welding speeds can be attained.
There are three basic types oflasers viz., the solid-state lasers,
semi-conductor lasers, and gas lasers. Although at present the
solid-state Nd:YAG (Neodymium doped Yttrium aluminium
garnet) lasers are the most used lasers in industry but their heat
conversion efficiency is very low usually below 1%. Fig. 1.16 shows
the essential features of one such laser unit. The CO2 lasers with
their heat conversion efficiency of 15-25% are now being
increasingly installed.
Laser beam welding is more versatile than EBW in that it can
weld metals in air, in a gas shield, and in vacuum. It can also weld
through transparent materials as laser beam can easily pass
through such media.
A laser must have a power rating of atleast about 2 KW to be
used successfully for welding, however Nd:YAGlasers of 100 '\IVto
1 KvV power are Ilsed for welding in indtlBtry because they can
achieve pulses with peak power of the order of 10 Iffi'. Although
Welding and Welding Processes 17
Losli:r rod
Rdlli:ctor·
E:nt:rgy monitor
Bli:om dump
pointing
PrOCli:SSshultli:r
Pu!s~ control
unit and
poWli:r supply
Fig. 1.16. Schematic representation of an Nd:YAG laser unit.
C,AfterDawes)
CO2 lasers of power rating up to 25 KW have reportedly been
installed but presently such a unit is a rare piece of equipment
and is extremely expensive.
c.ommercially laser welding is finding use in radio engineering
and electronics where fine wires are often to be connected to films
on micro-circuit boards, solid-state circuits, and micromodules. It
is also expected to be used in high quality precision work as in
aerospace industry and high speed mass production applications
as in automobile industry.
1.3.1.12. Oxy-Acetylene Welding
In this process acetylene gas is mixed with oxygen in the gas
welding torch and is then burnt at tlle torch tip to give a flame
with a temperature of about 3300°C which can melt most of the
ferrous and non-ferrous metals in common use. Fig. 1.17 shows a
standard set-up for oxy-a<,:etylenewelding.
Three types of flames are used in oxy-acetylene welding. The
nature of t.he flame depends upon the ratio of the two gases~ The
neutral flame is most often used for the welding of most of the
materials like low carbon steels, cast steel, cast iron, etc. The
oxidising flame has higher proportion of oxygen than acetylene
and is used for welding of Mn-steel, brass,' and bronze whereas
18 Welding Engineering and Technology Welding and Welding Processes
19
Cylinder pressure Outlet pre:ssure gauge: -
'-lit
gauge Cylinder pressure "'a
Outlet pressure gauge
.5C7l

gauge "-
Flash back Valve Valve
arre:stor

Welding
torch
;
'-
o
~
;e
Qi
Oxygen valve
Co
@ ~
c'"v
C7l

Acetylene
._
"Q'-
G a
."
valve -"8
Hose che:ck
3:'" l:
o:s

valve:s
v ~
:;,-~
_"'00
~
..
0~00 ..s
:::0;:1
Q. 00 lit Co
:;'
...•
Ace:tylene Oxyge:n ~
(maroon) (black) '"
....
Fig. 1.17. A standard set-up for Oxy-acetylene welding. Q
dQ
v0
.: a'"III
'E I-<"Q'':'"C0c... '0.••
'-
••
the carburising flame has higher proportion of acetylene in it and •• '" >
0v~~I I
.!It
"Q
••• C7l C f
c
Ui .::
(.')
~
C7l

is used for welding aluminium, nickel, etc. E

."

The heat transfer to the work in this process is very poor S

..,
(about 30%) and may lead to a wide HAZ around the weld. The ~
welding speed is also' accordingly low.. S
Typical applications of oxy-acetylene welding include welding <ll

-5
of root run in pipe and other multi-run welds, light fabrications C/J
cO
like ventilation and air-conditioning ducts, and motor vehicle ...•
~
repairs. A large percent of general repair work is also done by this bll
process. ri:

1.3.2. Arc Spot Welding Processes

'"
The processes in this class are used to join workpieces within lli:i
a narrow zone of desired shape. -0
.!1c
1.3.2.1. GTAW Arc Spot Welding ~19
.5C7l

In this process the equipment used is basically the same as

for conventional GTAW except that the control system includes
timing device and the torch nozzle is modified to develop a spot
weld at the intended place. GTAWarc spot welding may be done
with ac or dcen (direct current with electrode negative). DCEN is
used for all materials except aluminium for which ac with
continuous superimposition of high frequency (HF) current is
employed. The torch nozzle is made of copper or stainless steel
and is often water-cooled as the arc is enclosed completely within
the nozzle. The torch nozzle, usually about 12 mm inside diameter,
is provided with venting ports to affect gas flow and escape.
rate The shielding
of 2.5 gas used is either helium or argon with a flow
to 4.5 lit/min.
To accomplish a spot weld, the arc is initiated by the HF
discharge for which the outline circuitry is shown in Fig. 1.18. The
arc continues for the preset time and the spot weld'is achieved.
 Welding and Welding Processes 21

20 Welding Engineering and Technology nozzle attached to it. A vented metal nozzle of a shape to suit the
application is fitted to GMAW gun and is pressed against the
Normally no filler metal is used but when required it is fed workpiece at the desired spot. The operation is carried out for a
with the help of special wire feeder. Filler wire addition improves period of 1 to 5 seconds and a slug is melted between the parts to
nugget configuration and helps in overcoming crater cracking. be joined, as is shown in Fig. 1.19. Timing is usually controlled
This process is mainly used in its semi-automatic mode but it automatically with the help of a timer.
can be mechanised and even controlled by numerically controlled No joint preparation is required except proper cleaning of the
(NC) system to achieve high rate of production. overlapped areas. Argon and CO2 are the shielding gases
GTAWarc spot welding is widely used in the manufacture of commonly used for GMAWarc spot welding.
automatic parts, precision metal parts and parts for electronic GMAWarc spot welding process can be used most efficiently
components and appliances. It is particularly useful for appli- for downhand welding position. It can be successfully employed
cations where access to a lap joint can be gained only from orie for horizontal position but fails for overhead welding position.
side. This process does not require a hole to be made in either
1.3.2.2. GMAW Arc Spot Welding member, thus it differs from plug welding in that respect. As the
Normal GlVIAWequipment can be used for making spot welds upper member is required to v••be melted•... through0 and through its
0
Q.
uc0 0> ~ 'C
.3 oS
~
:i5
.w•..
l:I
C1l ~
.bj)
-;;;
;a
between the lapped sheets by providing a special 0v
;:, torch with a'60•..~J ~ 0 c u'"u
< s en
~
.:,
OIl
•... •••
••
eC'l~dclQ,:I
:i5
OIl
•..
Q. •••
,...j
'13 ~
>E) •• ::1
go .!io

e
.!!!
0:(E)
~

......
••v b.O
~5
00
c:
:a
0.11I 0;
~
@ 8-
OIl

~
<8 ~
d
•...
.s
Q.,

Z
'"
OIl

Cd
,~
o
c
<:>

.~
•••
~•...
OIl
v
'"
S
1
III 'E
~
'0.
""
•..
o
~
.~
Cd

E
~ ~ OJ

-B
.~ UJ
o:i
•...•

b.O

~ ti:

\/I
•..
._ w-..!
~ "0

~-= e
 23
22 Welding and Welding Processes
Welding Engineering and Technology
1.4.1. Spot Welding Processes
thickness is normally restricted to 3 mm. The thickness of the
second member is not important; In this class of processes the materials are joined at a spot the
size of which depends upon the design specifications and is
GMAW arc spot welding can be used successfully on alumi- controlled by the electrode size and the magnitude of the welding
nium, mild-, low alloy-, and stainless steels. current. Two main processes fall in this class viz., resistance spot
1.3.2.3. Stud Welding welding and projection welding.
This is a process of welding stud (a headless threaded bolt) or 1.4.1.1. Resistance Spot Welding
stud-like pieces (e.g. bolts, screws, rivets, rods, etc.) to flat pieces In resistance spot welding process overlapping sheets are
like plates. welded by the flow of current between two cylindrical electrodes.
The main equipment for stud welding consists of a stud The main equipment for spot welding is the spot welding machine
welding gun, a time control unit, a dc power s(\urce of 300-600 which consists of a step down transformer, a time control unit, and
amperes capacity, studs and ceramic protective caps called a pair of copper alloy electrodes. As voltage plays no direct role in
Rocker arm
ferrules.
For stud welding the stud is held in the welding gun and a Electrode
ferrule is slipped on it. The stud is then made to touch the cleaned
spot where it is to be welded and the switch, in the form of gun
trigger, is pressed and the process is completed in a. couple of - " F~-~_ ,
seconds. Fig. 1.20 shows the basic features of a stud welding unit. Electrode
n "--"",\
I I 'I Air cylinder
Typical applications of stud welding include steel decks of
ships, for attaching brackets, hangers, cover plates, piping, etc. to
,
III I'-- ....J I
II\ ,J

Transformer
metal workpieces. The process also finds wide use in automotive secondary
rail road machinery manufacturing and construction industries. ~
Current control
1.4. Resistance Welding Processes Foot switch
In all resistance welding processes the heat is generated at
the interface of contacting workpieces due to the resistance offered Air in
to the flow of electric current and is expressed by Joule's Law,
12Rt (a) Setup
Squeeze Forge
H;:;:1"\J ...(1.1)
Force
where,
H;:;:
I;:;:
heat generated, calories,
welding current, amperes (rms),
R ;:;:contact resistance, ohms,
t;:;: time for which the current flows, seconds,
Fo". lh \
W.ld
Time
~'

J;:;: the electrical equivalent ofheat~

1"\ thermal efficiency of the process.
;:;:

The welds produced by resistance welding are normally

without the addition of any filler material and are, therefore, cu,,'"'~ I Time
I
sometimes referred to as autogenous welds.
Resistance welding processes can be divided into three cate-
gories viz., spot, seam, and zonal type welds with some of them ( b) Operation (c) Current and force cycles
falling in more than one category. Brief description of industrially Fig. 1.21. Essential features of a rocker arm type spot welder;
important processes, among them, follows. operation details with current and force cycles.
24 Welding Engineering and Technology Welding and Welding Processes 25

resistance welding it is, therefore, kept low between 5 to 25 volts body like automobile chasis, by making raised portions or
but the current is usually heavy (100-50,000 A); however, it flows projections on one of the components. The projections are made
only for a short duration of time (0.06 to 3 seconds). Application by embossing or intersection (e.g. cross of wires, etc.). There are
of pressure to achieve forge between the two sheets is an essential several types of projections viz., round button or dome type, ring
aspect of the process. Fig. 1.21 shows a general purpose spot type, elongated projection, shoulder projection, and radius
welding machine. projection. The projections act to localise the heat of welding
Spot welding is mainly used for lap welding of thin sheets circuit, because when placed together the sheets touch only at the
particularly in the welding of automobile and refrigerator bodies, points of the projections. Projection collapses due to heat and
and high quality work in aircraft engines. pressure and a fused nugget is formed at the interface.
1.4.1.2. Projection Welding Equipment used for projection welding is similar to that used
Projection welding is a process of joining two sheets or a sheet for spot welding except that the rod electrodes are replaced by flat
and a thick componept, or a small component like nut to a big copper platens as shown in Fig. 1.22; which also shows the force
and current cycles for the process.
Air or hydraulic cylinder Typical applications of the process include projection welding
of reinforcing rings around holes in sheet metal tanks, welding of
threaded studs to backing bar or plate and cross-wire welding.
Cross-wire products include such items as refrigerator racks, grills
Ram of all kinds, lamp shade frames, wire baskets, fencing, gratings
r ---, and concrete reinforcing mesh.
7-1- - --l-~
[ ~ J '. 1.4.2. Seam Welding Processes
r-
''1--=1 II II In these processes the weld is established along a seam so as
I,I 1----..., I I• to make a leakproof joint. The seam weld may be produced by
:1"T
, L ----, I
..1
making partially overlapping spot welds. Apart from projection
welding the main processes in this class include resistance seam
welding, electric resistance butt seam welding (ERW), high
10 ~o6 freque.ncy resistance welding (HFRW), and high frequency
Knee
3o~co7
2,~o8 induction welding (HFIW).
support 1.4.2.1. Resistance Seam Welding
In resistance seam welding wheel electrodes are used to
produce spot welds overlapping to the extent of 25 to 50%. Due to
(a) Setup for projection cum (b) operation details
spot welder. shunting of current through the already made weld, the ,current
required is higher than in normal spot welding. Pressure is
applied to fuse the metal properly into a nugget as in sp<,>twelding.
Fig. 1.23 shows the principle of resistance seam welding alongwith

LEw,~
Fo". " .". \
Tim~
& Cu".nt ~ I Tjm~
I ,
force and current cycles employed in the process.
Seam welding is used for producing leak-proof joints in tanks
and boxes generally required for the automobile industry. This
process is, however, restricted to welding thin materials ranging
between 2.5 and 5.0 mm. Also, it is used mainly for welding
(c ) Current and force cycles materials with low hardenability rating, for example, hot-rolled
grades of low alloy steels. This process is commonly used for
making flange welds for use in water-tight tanks.
Fig. 1.22. Essential features of a projection welder;
operation details with current and force cycle.
26 Welding Engineering and Technology
Welding and Welding Processes 27
• •..
'"
E
•..
0>0
,-
c-
11\

:9C
'" Cl
S
Side 15~
Workpiece
force
Welding
transformer
and controls Mains
Tube bl!ing weldl!d supply
switch Fig. 1.24. ERW process of tube manufacturing.

(a) Set-up The maximum rate of production is limited by the welding

current frequency because as the welding speed is increased
Copper alloy individual current half-cycles eventually lead to spot welding
electrode wheel instead of seam welding. To overcome this difficulty current
frequency is usually increased to 350 Hz to achieve welding speed
of up to 36 m/min. The tube produced by this process has a fin of
upset metal along the weld joint both inside and outside which is
usually removed by installing appropriate cutters on the
RotationForce ~. --~~------
~ __
Time
production line. The tube is cut to the desired lengths by
employing a cutter which moves along the tube and is
synchronised to cut the desired length in the available run in a
given cycle.

n
..>'

1.4.2.3. High Frequency Resistance Welding (HFRW

CUIT.n,Ln Time
Process)
Wddl!d seam
( b) ope ration details (c) Current and fo rce cycles

Fig. 1.23. Resistance seam welding.

1.4.2.2. Electric Resistance Butt Seam Welding (ERW
Process)
Large quantities of steel tube and pipe are manufactured by Contact
resistance butt seam welding from strip which is continuously
edge sheared and rolled into tube of desired diameter before
welding. Alternating current of up to 4000 A at about 5 volt is
introduced across the joint by pressure rolls as shown in Fig. 1.24.
For introducing heavy current directly to the moving electrodes a
rotating transformer with slip rings on the primary side is
employed. Unlike the normal resistance seam welding, current Impl!der
and work motion are continuous in this process. Fig. 1.25. HFRW process of tube manufacturing.
28 Welding Engineering and Technology
In this process the tube is formed by rollers in the same way
, --------------------29
as in ERW process but the current in the range of 50u to 5000 A
at a frequency (f) of up to 500 KHz and a voltage of about 100 volts
is introduced through pr0bes made of copper alloys and silver
brazed to heavy water-cooled copper mounts. Contact tip sizes
range between 15 and 650 mm 2 depending upon the amperage to
be carried. Whereas in ERW process the heat is generated mainly
by the interfacial contact resistance in HFRW process it is
produced by the skin effect due to which the current flows in a
shallow depth of the conductor and is proportional to ..yr.
Pressure rollers, to provide the forging pressure are installed a
short distance down the line from the current probes as shown in
Fig. 1.25. Due to the skin effect the current flow path lies along
the strip through the apex. of vee formed by the faying surfaces
menting at an angle of 4° - 7° as they close to form the tube. The
depth of the heated region is generally less than 0.8 mm and thus
affords the optimal condition for weld joint.
HFR\V process is used to produce pipe and tubing of diameters
ranging between 12 and 1270 mm, and with a wall thickness of
0.25 to 25 mm. Any metal can be welded by this process with a
speed range of 5 to 300 m/min depending upon the wall thickness.
This process can also be used to manufacture spiral and finned
tubes and pipes. Various types of serrated or folded fins can also
be welded to tubes.
1.4.2.4. High Frequency Induction Welding (HFIW
Process)
High frequency induction welding of tubes is similar to high
frequency resistance welding except that the heat generated in the
welding
f Point of /' Welded seam
Impeder
Fig. 1.26. HFIW process of tube manufacturing.
Welding and Welding Processes
work material is by the current induced into it. Because there is
no electrical contact with the work this process can be used only
where there is a complete current path or closed loop wholly within
the work. The induced current flows not only through the weld
area but also through other portions! of the work.
Tube edges are brought together in the same manner as in
ERW and HFRW processes. A water-cooled induction coil or
inductor made of copper encircles the tube at the open end of the
vee as shown in Fig. 1.26. High frequency current flown through
the coil induces a circulating current around the outside surface
of the tube and along the edges of the vee, heating them to welding
temperature. Pressure is applied to accomplish the weld as in
HFRW process.
HFIW process is suitable for tubing made of any metal within
a diameter range of 12 to 150 mm with a wall thickness of 0.15 to
10 mm at a welding speed ranging between 5 and 300 m1min.
HFIW process is not limited to tube manufacture but can also
be employed to make circumferential welds for welding cap to
tube. The process can as well be advantageously used for
manufacturing tubing from coated material, small or thin-walled
tubing; and it eliminates surface marking by electrical contacts.
This process is, however, not suitable for welding high conductivity ,
metals or those which form refractory oxides as there is no
effective mechanism for oxide disposal.
1.4.3. Zonal Welding Processes
In these resistance welding processes heat is generated
simultaneously over the entire zone which is requi,red to be
welded. The processes included in this class are Resistance Butt
Welding, Flash Butt Welding, and Percussion Welding.
1.4.3.1. Resistance Butt Welding
In Resistance Butt Welding or Upset Welding the pieces to be
welded are held in clamps supported on two platens, one of which
is fixed and the other moveable; and form part. of the single-turn
secondary loop of a heavy duty transformer, as shown in Fig. 1.27.
The ends to be welded touch each other before the current is
switched on. A heavy current is then pa~sed from one workpiece
to another and the contacting faces are heated up due to the
contact resistance. The two pieces are pressed together firmly
after the desired welding temperature of 870 to 925°C, for steels,
is reached. The pressing action which results in the increase in
lateral dimension of the workpieces is called upsetting. Upsetting
takes pla~e both during and after the current flow. The upsetting
 30
StCitionary
clamp
- Welding Engineering and Technology
Upsdting
pr~ssut~
~Q (a) S~tup
~ = (b) Finished w~ld
, Welding
Fig. 1.27. Basic features of upset welding.
action results in welding of end faces with squeezing of a part of
the softened metal to form·a fin, which can bo removed later, if
required, by machining.
Resistance butt welding is used for end joir,ing of rods, tubes,
bars, and similar other sections for welding a cross-sectional area
of up to 150 mm2• Wire and rod from 1.25 mm to 30 mm diameters
can be upset weided. Typical application of resistance butt welding
is in wire mills for joining wire coils to each other to facilitate
continuous procesGing.
1.4.3.2. Flash Butt Welding
Flash welding is similar to resistance butt welding except that
it is accompanied by arcing and flashing. Flash welding consists
of one fixed and one moveable clamp to hold and clamp the
workpieces firmly as well as to force them together, a heavy duty
sjngle phase transformer with a single turn secondary, alongwith
equipment to control welding current, movement of the clamp,
force, and time. With a voltage of about 10 volts across the clamps,
heavy current flows along the asperities across the contacting
faces of workpieces. As the points of contact are melted and the
metal is squeezed out in a shower of fine molten droplets, the
contact is broken and arcing takes place across the gap. With
further movement of the clamp, the process of melting, flashing
and arcing repeats itself. Due to flashing contaminants from the
contacting faces are removed and the surfaces are heated to a
uniform temperature. Finally the movement of the platen (or
moveable clamp) is rapidly increased and a high force is applied
Welding and Welding Processes
to achieve a weld with the expelled metal forming a rough fin or
flash around the joint. The flash can be removed by subsequent
31
machining. Basic arrangement for flash butt welding process is
shown in Fig. 1.28.
Fix~d clamp Moveable clamp
Work
_Force
- -'
Moveable platen
....•
transformer
and
cont rols
Flexibl~ I~Qd
A.C.
supply
Fig. 1.28. Basic elements of a set-up for flash butt welding.
Flash Butt Welding requires a heavy power supply, for
example, currents in excess of 100,000 A can flow across the
interface with a power input. up to 200 KVA.Transformer used for
flash butt welding are single phase which can, thus, place an
unbalancing load on normal 3-phase supply from the mains. This
necessitates the use of special transformer which can distribute
the load uniformly.
In flash butt welding the pieces to be welded must be held with
enough force to avoid slipping and that requires a clamping force
of up to twice that of the upsetting force. The upset force is around
70 MPa· for mild steel and nearly four times that for high strength
materials.
Flash butt welding is extensively used for welding mild steels,
medium carbon steels, and alloy steels as well as non-ferrous
metals like aluminium alloys, nimonic alloys (80% Ni + 20% er)
and titanium. Dissimilar metals may be flash welded if their
flashing and upsetting characteristics are similar, for example
aluminium can be flash butt welded to copper or nickel alloy to
steel.
Typical uses of flash butt welding include welding of wheel
rims, cylindrical transformer cases, circular flanges, and seals for
power transformer cases. The aircraft industry utilises flash butt
• 'lMPa = 1 N/mm2.

32
Welding Engineering and Technology
welding to manufacture landing gears, control assemblies, and
hollow propeller blades while the petroleum industry uses oil
drilling with fittings attached by flash welding. Other uses of the
process include welding of rails, steel strips, window frames, and
heavy duty chain links e.g., anchor chains for ships. To avoid
shunting of current the ring-type workpieces are made by welding
two halves of each link simultaneously.
1.4.3.3.Percussion Welding
Percussion welding is an arc welding process of joining,
end-to-end, two parts of equal cross-section. The arc is produced
by a short ,.pulse of electrical energy and pressure is applied in a
percussive manner to produce coalescence simultaneously over
the entire abutting surface. In general percussion welding is the
term used in the electronics industry for joining wires, contacts,
loads. and similar items to a flat surface.
There are two variants of tltis process viz., magnetic force
percussion welding and capacitor discharge percussion welding.
Essential steps involved in the process involves, (i) establishing
an arc between the surfaces, to be joiried, with high voltage to
ionize the gas between the parts or with high current to melt and
vaporise a projection on the part, and (ii) move the_part~ together
pe rcussively with an applied force to extinguish the arc' to
accomplish a weld.
Welding heat is gener~ted by a .high current arc between the
two parts to be joined. The extremely short duration arc limits
mel ting to a very thin layer 011 the two surfaces being welded.
Consequently, there is very little upset or flash on the periphery
of the welded joint. Filler metal is not used nor flux or special
atmosphere required.
33
Welding and Welding Processes
properties in electronic industry. Large contact assemblies for
relays and contactors are often produced by percussion welding.
This process is also used to weld electronic components to
terminals as shown in Fig. 1.29.
1.5. Solid-State Welding Processes
In solid-state welding processes the material to bl: welded is
heated to a temperature below or just up to the solidus. The
coalescence between the parts is achieved under pressure and thus
forging or impact action plays an important role in all these
processes.
Solid-state welding processes may be divided into two groupS
viz., high heat input processes and low heat input processes.
1.5.1. High Heat Input Processes
High heat input solid-state welding processes include Forge
Welding, Friction Welding and Diffusion Bonding.
1.5.1.1. Forge Welding
Forge welding or smith welding is the oldest known welding
process and its use has been reported from about 1400 B.C. By
this process the pieces to be welded are heated to above 1000°C
and then placed together and given impact blows by hammering.
In the more recent form of welding of large components the
pressure is applied by rolling, drawing and squeezing to achieve
the forging action. The oxides are excluded by virtue of design of
the workpieces and or by the use of appropriate temperature as
well as fluxes. Fluxes commonly used for forge welding low carbon
steels are sand, fluorspar, and borax. They help in melting the
oxides, if formed.
Proper heating of the workpieces is the major welding variable
that controls the joint quality. Insufficient heating may not affect
a joint while overheating results in a brittle joint of low strength.
Also, the overheated pieces tend to be oxidised which shows itself
by spongy appearanee. The joints most commonly employed are
scarf, cleft and lap types, as shown in Fig. 1.30.
~~. ~
Scarf
Butt
~
Jump
C \ I l ~
Lap Cleft
Fig. 1.30. Types of joints used in forge or smith welding.
 35
34 Welding Engineering and Technology Welding and Welding Processes

Forge welding is now mainly used in under-developed coun-
tries for welding small agriculture implements and chains, etc.
1.5.1.2.Friction Welding
In friction welding one piece is held stationary and the other
is rotated in the chuck of a friction welding machine. As they are
brought to rub against each other under pressure, they get heated
due to friction. When the desired forging temperature is reached
throughout the rubbing cross-seCtion of the workpieces, the
rotation is stopped suddenly and the axial pressure is increased
to cause a forging action aridlienceWeldmg: -------
The machine used for friction welding resembles a lathe, as
shown in Fig. 1.31, but is strudier than that. The essential
i.c. engine valve heads to sterns, automobile rear-axle hub-end to
axle ca-sing.
1.5.1.3. Diffusion Bonding
In diffusion bonding or diffusion welding a weld is achieved
by the application of pressure, of the order of 5 to 75 N/mm2, while
the pieces are held at a high temperature, normally about 70% of
the melting point in degrees absolute i.e. about 1000°C for steel.
The process is based on solid-phase diffusion which, obviously, is
accelerated with rise in temperature.
Depending upon the extent of diffusion required the process
may be completed in 2 to 3 minutes or may take many minutes or
even hours. The quality of the surfaces to be welded plays an
important role; surfaces machined to a standard of 0.4 to 0.2 !-lm
CLA (centreline average) is usually adequate. The surfa.ces must
be degreased before welding by using acetone or petroleum ether
swab.
Diffusion bonding can be achieved by three methods viz.,
(i) gas pressure bonding,
(ii) vacuum fusion bonding,
(iii) eutectic fusion bonding.
In Gas Pressure Bonding, the parts are held together in an
inert atmosphere and heated to a temperature of 800°C by a
system resembling an autoclave". During heating the high
pressure provides uniform pressure over all the surfaces. This

Heat by
radiation
Fig. 1.31. A continuous drive friction welding machine. conduction
induction or "- "' Vacuum
Door Workpi~c~ pump
features of the machine are that it should be able to withstand Th~rmal
high axial pressure of the order of up to 500 N/mm2 and be able
to provide a high spindle speed of up to 12,000 rpm though the Moving plQt~n --, bQrri~ r
usual working range may rarely exceed 5000 rpm.
Friction welding is a high speed process suited to production
welding. However, initial trials are required to standardise the
process parameters for a given job. Friction welding of two pieces
rarely takes more than 100 seconds though it may be just about
20 seconds for small components.
, Pr~ssur~
Friction welding can be employed to weld most of the metals
and their dissimilar combinations such as copper and steel, Fig. 1.32. Essential features of a set-up for vacuum dif~usion bonding.
aluminium and steel, aluminium and titanium, etc. Typical * An autoclave is a vessel in which chemical reactions take place at high
applications of the process include welding of drill bits to shanks, temperature and pressure.
 36
Welding Engineering and Technology
method is used for bonding non-ferrous metals only because it
necessitates high temperatures for steels.
'In Vacuum Fusion Bonding the parts are held in an intimate
contact in a vacuum chamber. The pressure on the parts is applied
by mechanical means or a hydraulic pump, and heating is done in
the same way as in gas pressure welding. Fig. 1.32 shows the
essential features ofa set-up for vacuum fusion bonding. Avacuum
pumping system which can quickly reduce pressure to about 10-3
torr (mm of mercury) needs to be used. High pressure created by
the use of mechanical or hydraulic means makes it possible to
diffusion bond steels by this method. Successful joining of steel
can be achieved at a70
temperature of about 1150°Cunder an applied
pressure of nearly N/mm2•
In Eutectic Fusion Bonding a thin piece of a particular
material is placed between the surfaces to be welded. This results
in the formation of a eutectic compound by diffusion at an elevated
temperature and the piece may completely disappear and form
eutectic alloy at the interface. The material used for being placed
in-between the two parts is usually in a dissimilar metal foil form
with a thickness of 0.005 mm to 0.025 mm.
Diffusion bonding finds use in radio engineering, electronics,
instrument making, missile, aircraft, nuclear, and aerospace
industries, Typical applications of this process include tipping of
heavy cutting tools with carbide tips or hard alloys, joining of
vacuum tube components, fabrication of high temperat.ure heaters
from molybdenum disilicide for resistor furnace that can operate
in an oxidising atmosphere up to 1650°C. In aerospace industry it
is used for fabricating complex shaped components of titanium
from simpletostructural
components resist wear,shapes.
heat orItcorrosion.
is also used for surfacing
1.5.2. Low Heat Input Processes
Low heat input solid-state welding processes include
Ultrasonic Welding, Explosion Welding, Cold Pressure Welding,
and Thermo-Compression Bonding.
1.5.2.1. Ultrasonic Welding
In ultrasonic welding a metallic tip vibrating at ultrasonic
frequency (i.e. the vibrations which produce sound beyond the
range of human hearing) is made to join a thin piece to a thicker
piece supported on an anvil. The frequency used is mainly around
20 KHz though higher frequencies up to 60 KHz have reportedly
whichused.
been Higher
energy the frequency of vibration higher is the rate at
is transmitted.
Welding and Welding Processe..•• 37
Ultrasonic welding equipment consists of two main parts viz.,
a power source, and a transducer. The power source converts the
50 Hz mains supply to a high frequency electric power and that
is converted by the transducer to magnetic flux and then the
kinetic motion which is amplified through a velocity transformer.
The schematic illustration of the set-up is shown in Fig. 1.33.
H.f.
supply Velocity transformer
Load
1.-1
Work
Anvil
Pivot
Fig. 1.33. Schematic illustration of the set-up for ultrasonic welding.
The transducer and the horn-shaped velocity transformer form
a unit called sonotrode. The tip of the velocity transformer which
is used for welding is made of high speed steel (i.e. steel containing
14 to 20% tungsten and 4% chromium) or Nimonic alloy (80/20
NileI') and is shaped to a spherical contour of about 75 mm radius.
These tips are brazed or welded to the horn.
A velocity transformer' is made oflow-loss, high strength metal
like titanium and is shaped to achieve the desired frequency based
on the relationship,
f= A.E ... (1.2)
where, f = vibrationfrequency,
A. = wavelength,
E = modulus of elasticity of horn material.
Since the welding tip has to be an anti-node thus the length
of the horn has to be a multiple of ').j2 and any support must be
at nodal points at A./4. Avibrator, thus, can operate at one definite
frequency only.
The work to be welded is placed under the sonotrode tip in lap
joint formation and is supported on an anvil. Force is applied on
the sonotrode tip with the help of pneumatic, hydraulic or ,a
spring-actuated device. This set-up can be used for spot and seam r
welds. To make an annular or ring type ultrasonic weld, the force
is applied tangentially on a cylindrical tip to give torsional'
vibration to the welding tip.

38
Welding Engineering and Technology
Due to the ultrasonic vibrations the oxide layer over the metal
is broken and a clean metal to metal contact is achieved. The
temperature at the interface rises to between 35 to 50% of the
absolute melting point temperature of the metal and, thus, a
solid-state weld is achieved. The strength of the weld is 65 to 100%
of the base metal strength. The process is fast and seam welding
with a speed of up to 10 m/min. have been reported. Energy
required (E) for ultrasonic welding unit depends upon the
thickness (t) and hardness (h) of the material to be welded-and
may be calculated from the following relationship,
E = Kt3/2 h3/2 .•. ( 1.3)
. With adequate power a spot weld can be made in less than a
second but the maximum thickness of the thinner piece should not
exceed 3 mm.
Ultrasonic welding can be used for welding thin to thick parts
as well as for welding dissimilar metal combinations like alumi-
nium to steel, aluminium to tungsten; aluminium to molybdenum,
nickel to brass, etc. The major users of the process are semi-
conductor, micro-circuit, and electrical contact industries. It is also
used by automotive and aerospace industries.
TYpical applications of the process include fabrication of small
motor armatures, aluminium and gold lead wire connections to
transistors and diodes, helicopter access doors, dissimilar metal
joints in solar collectors. The unique applications of ultrasonic
welding, however, is welding of containers of explosives like
nitroglycerine, pyrotechnic (fire works), and reactive chemicals.
1.5.2.2. Explosion Welding
In explosive or explosion welding process the weld is achieved
by making one part strike against the other at a very high but
subsonic velocity. This is achieved by the use of explosives usually
of the ammonium nitrate base. The process is completed in
micro-seconds.
Flyer
'Jet
Coilision point
Fig. 1.34. Basic features of a set-up for explosion welding.
39
Welding and Welding Processes
The set-up, in principle, used for explosion welding is shown
in Fig. 1.34. It shows the two plates to be welded placed at an
inclination to each other. The included angle varies between 10
and 100• The thicker plate called the target plate is placed on an
anvil and the thinner plate called the flyer plate has a buffer plate
of PVC or rubber, between it and the explosive charge, for
protection against surface damage. The charge is exploded by a
detonator placed at one end of the flyer plate. When the charge
explodes, the flyer plate moves towards the target plate at a
velocity of 150 to 550 m/sec and the pressure produced at the
interface of the impacting plates by such a high velocity is of the
order of 700 to 7,000 N/mm2• Under such a high velocity and
pressure the metal flows ahead of the joining front acting like a
fluid jet resulting in a bond of the interlocking type as shown in
Fig. 1.35. The interlocking is an essential aspect of an explosion
weld and is the cause of its strength. The weld strength equal to
the strength of the weaker of the two components (metals) can be
achieved.
Weld interface
Flyer plate
Target plate
Fig. 1.35. Nature of another type of interlocking bond for explosion welding.
Explosion welding is normally an outdoor activity and needs
a large area to ward off the persons coming close to the explosion
site particularly when an explosive charge of high strength may
have to be exploded.
Explosion welding can be used for joining dissimilar ..metal
combinations like copper and steel, aluminium and mild steel,
aluminium and inconel (76% Ni + 15% Cr + 9% Fe), aluminium
and stainless, etc. Typical applications of explosion welding
include cladding of thick plates by thin sheets, even foils. Tube to
tube-plate joints in boilers and heat exchangers, valve to pipe
joint, as well as blocking of leaking tubes in boilers can be
successfully achieved by this process.
1.5.2.3. Cold Pressure Welding
Cold pressure welding or cold welding is a solid-state welding
process in which a weld joint is produced solely by the application
of pressure at room temperature. No heat is involved in this
process. The main requirement of the process is that atleast one
 40
Welding Engineering
of the components being welded should be of ductile metal without
much tendency for workhardening. Thus metals with FCC (face
centred cubic) lattice structure. are best suited for this purpose.
Aluminium and copper are the major metals joined by this process.
The amount of deformation is a major factor in cold pressure
welding and it depends upon the properties and thickness of the
metal as well as the type of joint and surface preparation.
The parts to be joined by cold welding are cleaned thoroughly
contaminants on wire
by degreasing, brushing, and scraping to remove any
the surfaces.
Cold pressure welding is used to make usually lap and butt
types of joints. In lap welding the sheet thickness may vary
between 0.2 and 15 mm and the joint is affected with the help of
a single or a double die and the joint shape varies accordingly. The
.die radius varies with the workpiece thickness and the rough rule
is r = (1 - 3) t, where t is the sheet thickness.
Butt joints by cold welding are made by clamping the two parts
to be welded in a split die. Before clamping, however, a short
section is usually sheared from the ends of the parts to expose
fresh, clean surfaces with square ends. Butt welds usually have
higher strength. than the parent metal because the joint is
workhardened
and Technology Welding and Welding Processes 41
Commercial applications of cold welding include can joints
(Fig. 1.36) for packaging food, closing of aluminium cable sheaths,
and cases of semi-conductor devices, lap and butt joints of wires
and busbars for electroiysis cells, communication lines and trolley
wires.
1.5.2.4. Thermo-Compression Bonding
It is a pressure welding process which is employed at
temperatures above 200°C. The process deals with mainly small
components in the electrica1 and electronic industries for welding
fine wires of about 0.025 mm diameter to metal films on glass or
ceramic.
Ther,e are many versions of the process, three out of which are
shown in Fig. 1.37 and are referred to as chisel or wedge bond,
ball bon,d, and parallel gap bond. In the chisel or wedge bond a
wire is deformed under pressure and welded to the film with the
help of wedge shaped indentor.
(1 )
Pie reed indenter
Wedge indenter wire

Fusion
. Pressure for cold welding may be applied, to the properly parting bead
aligned components, with hydraulic or mechanical presses, rollers,
or specially designed manual or pneumatically operated tools. The
rate at which the pressure is applied has no effect on weld joint
properties hence welding can be done at high speed.
, 5ubstrate
(a) Wedge bond (b) Bell bond

5pilt electrically
Punch conducting indenter
Lid Film
Die

(c) Resistance or Parallel-gap bond

, Joint detail Fig. 1.37. Different methods of thermo-compression bonding. CAfter Houldcroft).
In the ball bond a wire is heated by a micro-hydrogen flame
Con ---~--- to form a ball at the wire tip as shown in Fig. 1.37(b), which is
subsequently welded to the heated film on substrate by the
pressure exerted through the pierced indentor.
(a) Before welding . ( b) Afte r welding In the parallel gap bond the wire or strip is pressed to the film
with the help of twin electrode made of high resistance material
Fig. 1.36. Details of can joint produced by cold pressure welding.
 42
Welding Engineering and Technology
like tungsten. The flow of current through the wire or strip heats
it up locally thus keeping the heat confined to small zone around
it.
For all these variants of the process local inert atmosphere is
created around the joint being bonded. Ultrasonic variations
replace heating in some of the applications of all these modes of
the process.
Commercial applications of the process include welding of
noble metals, aluminium, and copper to substrates of glass or
ceramics.
1.6. Allied Processes
The processes allied to welding are of three types viz., (i)
joining processes involving no melting of the parent metal with
consequential lower joint strength; (ii) metal depositing processes
which often employ the welding process and or equipment to lay
or spray material on to a substrate for repair to accomplish certain
desired properties in the base material; and (iti) thermal cutting
processes which help in cutting the material to the desired size for
-welding; alternatively the process may be employed for edge
preparation by gouging. Brief descriptions and typical appli-
cations of the industrially important allied processes are given in
the following sections.
1.6.1. Material Joining Processes
The processes allied to welding used for joining of materials
include Soldering, Brazing, and Adhesive Bonding.
1.6.1.1. Soldering
Soldering is a process of joining metal pieces usually in the
form of overlapped joints by making a filler metal flow into the
gap between them by capillary action. The filler used is called a
solder and has a melting point lower than 450°C.
The solder most commonly used is a compound of tin and lead
in the ratio of 40/60, 50/50 or 60/40 having a melting point between
185 and 275°C, depending upon composition.
Soldering is done by thoroughly cleaning the pieces with the
help of wire brush, emery cloth, file or even steel wool. The pieces
are then fitted closely with a gap of about 0.08 mm between the
mating surfaces. A flux is applied. to the surfaces to be joined so
as to avoid the formation of oxide due to subsequent heating as
also to dissolve any flux still present on t.hem. A commonly
employed general purpose flux is zinc chloride while for soldering
electrical connections resin, being non-corrosive, is best suited.
Welding and Welding Processes 43
After the application of flux, the pieces are heated by any of
the available methods viz., oxy-acetylene torch, soldering iron, hot
plate, electrical resistance, induction heating, oven heating, or dip
heating; Fig. 1.38 shows a general purpose elec~rical soldering iron
50ldlZring iron
Fig. 1.38. Use of electrically heated soldering iron for making a fillet weld.
in common use. The solder is then applied to the gap. It melts and
flows into the interface of the mating surfaces by capillary action.
On cooling it solidifies and provides a joint of adequate strength.
If the gap between the surfaces is small as mentioned above, then
the strength of the joint is more than that of the strength of solder.
However, if a thick layer of solder is deposited then the maximum
strength attained by the joint equals that of the solder. On cooling
the joint is cleaned by hot water to avoid corrosive action of the
flux residue.
Commercially, soldering is used extensively for joining thin
sheets of ferrous and non-ferrous metals where the joint is not
stressed in tension. It is also widely used in the electrical and
electronics industries.
Typical uses of soldering include joining of electrical con-
ductors, and plumbing of copper tubing to copper fittings.
1.6.1.2. Brazing
Brazing is a process of joining metals by using a non-ferrous
filler metal having a melting point above 450°C but below the
solidus of the base metal. No melting of the base metal is involved
and the filler spreads by capillary action between the pieces being
joined.
The workpieces to De brazed are usually' prepared for lap or
butt joints. Both square butt and scarfed butt are used. The
cleaning of the pieces is done by mechanical methods like filing,
grinding, etc. or by the use of chemicals like carbon tetrachloride
(CCI4)·
 44 Welding Engineering and Technology .
.Joints to be brazed are made with small clearances of 0.025
mnl to 0.25 mm. Brazing flux is then applied to dissolve solid metal
oxide still present and to prevent further oxidation. Brazing fluxes
usually contain chlorides, fluorides, and borates of alkali metals.
Borax, however, is one of the most popular brazing fluxes.
Workpi¢ces
Preplaced
filler mdal
Coil
Fig. 1.39. Schematic illustration of induction brazing.
Heating of workpieces is achieved by oxy-acetylene flame,
induction heating or furnace heating. Fig. 1.:39 shows the
induction heating system employed for brazing. The brazing filler
material, if not already placed in position over the joint, can be
'applied in the form of a rod or a wire and melted to make it flow
into the joint by capillary action. The fillers most commonly used
are brass (60/40 Cu-Zn) and silver-copper-zinc- cadmium alloy like
35 Ag, 26 Cu, 21 Zn, 18 Cd.
Residual flux left on the brazed .joint can be removed by
washing with hot water followed by air drying.
Commercially brazing is used widely throughout the industry.
The major industries using brazing, however, are electrical,
electronics, and maintenance industries. Typically brazing is used
for joining carbide tips to the steel shanks for cutting tools.
A variant of Brazing is called Braze Welding or Bronze
Welding.
Braze Welding is a process in which the metal pieces are
joined in the same wayas in brazing but the filler material is made
to flow into the joint gap without the use of capillary action. Base
metal is melted, if at all, to a limited extent.
All joints used for oxy-acetylene welding can be braze welded.
Heat is also applied usually with the help of oxy-acetylene torch.
Welding and Welding Processes 45
However, carbon are, gas tungsten are, and plasma arc can be
utilised equally effectively and without the use of flux. The filler
is dipped in the flux and is melted with the help of flame or arc
to make it flow into the joint gap. The force of the flame can be
used to make the molten filler flow into the desired position. The
fluxes used for bronze welding are propriety type and the filler is
often a copper alloy brazing rod of 60/40 copper-zinc composition.
Joints for braze welding are of the square butt type for sheet
thickness up to 2 mm but need single or double vee preparation
above that thickness. However, efforts are made to eliminate sharp
corners in edge preparation to avoid overheating.
Braze welding was initially developed for repair welding of
cracked or broken cast iron parts, but is now-a-days used
conveniently for joining dissimilar metals like copper to steel,
copper to cast iron, nickel and copper alloys to cast iron and steel.
Typical applications of braze welding include rapid joining of
thin gauge mild steel, welding of galvanised steel ducts using
carbon are, thin sheets to thick parts of cast iron and for joining
telescoping pipes.
1.6.1.3. Adhesive Bonding
In adhesive bonding a metal is joined to another metal or a
non-metal by the use of an adhesive which usually consists of
,'!'
synthetic organic polymers of the thermosetting type, for example
epoxy, and phenol formaldehyde ..
The pieces to be joined are thoroughly cleaned by chemical or
mechanical means. Whereas chemical cleaning may involve
degreasing in a vapour bath followed by dipping in suitable acids,
mechanical cleaning may include shot blasting, grinding, filing,
wire brushing or sanding.
Adhesives ::re applied to the cleaned surfaces by brushing,
spraying, roller coating or dipping. The thickness of the layer of
adhesive, applied depends upon the metal being bonded, type of
adhesIve, solvent used and the strength aimed at, e.g., to achieve
an ultimate glue thickness of 0.025 mm to 0.75 mm anywhere from
0.125 mm to 0.375 mm of20% solid wet adhesive must be applied.
Typical joints used for adhesive bonding include lap, inset,
butt-strap, and tee type, as shown in Fig. 1.40.
~~/D477~
Lap Inset Butt strap Tee
Fig. 1.40. Typical joints used for Adhesive bonding.
 46 Welding Engineering and Technology
The adherends (workpieces) after being joined in the desired
joint configuration are placed under a pressure of 10 to 100 N/cm2
and are cured usually at a temperature of about 150°C for abo~t,
SO minutes. Adhesion is generally due to molecular attraction
between the adhesive and the adherend. Fig. 1.41 shows the joint
mechanism of an adhesive bond.
Interfacial bond layer
Weak boundary layer
7,,,,, ... InterfaCIal bond layttr
~~9!"e)'!=,n~/
~ cleaning of the surface before surfacing. Cleaning method
Fig. 1.41. Joint mechanism of an adhesive bond.
Commercial uses of adhesive bonding include a large number
of applications in manufacture of railway cars, microwave
reflectors, refrigerators, storage tanks, etc. But, by far the major
users of this method are the aircraft and automobile industries.
Typical applications of the process include fastening of
stiffners to aircraft skin, attaching of brake bnings to brake shoes,
and joints in the aircraft wing and tail assemblies.
1.6.1.4. Surfacing
Surfacing or overlaying is the process of depositing filler metal
over the surface of a base metal with a view tu achieving desired
properties which include corrosion resistance, wear resistance,
dimension control and metallurgical needs. Usually four variants
of the process are recognised viz. cladding, hardfacing, build-up,
. and buttering, the aims of which are respectively to provide
increased corrosion resistance, increased wear resistance,
dimensional requirement, and to achieve metallurgical com-
patibility.
Plasma
Nontransferred arc
power supply
Fig, 1.42, Schematic illustration of a set-up for Plasma arc surfacing,
Welding and Welding Processes 47
Surfacing can be done by a number of welding processes like
shielded metal arc welding, submerged arc welding, electroslag
welding, plasma welding, explosion welding, and even oxy-
acetylene welding. Fig. 1.42 shows a set-up for plasma arc
surfacing process. Cladding can be done even by mechanical
rolling operation.
Surfacing by welding is carried out by the usual welding
techniques but quite often the beads are laid overlapping to the
extent of SO to 50% to achieve complete union between them.
Shallow penetration with low dilution but adequate joint strength
are the desired aims of the process. This may necessitate proper
employed will depend upon the material and the surface integrity
of the base metal. Grinding, shot blasting, and chemical cleaning
may be employed to achieve the desired quality of the surface. The
thickness of the material laid usually varies between S'and 5 mm.
Commercially, submerged arc and plasma arc are the most
often used for surfacing. The industries using surfacing are many
including pressure vessel industry, railways, automobile industry,
and the earth moving machinery industry. Apart from overlaying
the inside of the newly made pressure vessels and boilers, the
process is used mainly for reclamation of equipment such as coal
and cement crushing equipment, drill rigs, coal cutters, forges and
press components like dies and punches.
Typical applications of the process include surfacing of engine
valve facings and seats of internal combustion engines, building-
up broken or wornout gear and sprocket teeth, repairing of
digesters used in pulp and paper mills, rock crusher cones and
bulldozer tips.
1.6.1.5. Therttlal Spraying
Thermal spraying is the process of depositing a metallic or a
non-metallic material over a base material to protect it from
corrosion, or to reduce abrasion, erosion, cavitation, or wear. It is
also used to restore the defective or worn surfaces to their original
shape and dimensions.
Thermal spraying process has three main variants viz.,
electric arc spraying, flame spraying, and plasma spraying.
Whereas the electric arc spraying uses material in wire form,
plasma arc spraying uses it in powder form while flame spraying
can use material both in wire and powder forms. Fig. 1.43 shows
the schematic of flame spraying set-up using material in wire
form.
 /'/
...~.,'{"~<:..
,,,_-':i" ',\ ••. .'~.,.,-I k'r> ;:.•.•.

48 Welding Engineering and Technology

Welding and Welding Processes \' 1\ X):~ _ ~ ~~'.. 49
, ... 0 J - "
Wire spool Sprayed metal Electron beam and laser beam use t~~lli:~R -<lensity;(:ra~iation
energy to achieve cutting of metals. Oxy-l\felg1fir'flame in
conjunction with oxygen jet is utilised to initiate and sustain an
exothermic oxidising reaction which generates enough heat and
affects parting of metals particularly low carbon ferrous alloys.
Out of these processes oxy-acetylene flame cutting and plasma arc
cutting are the two major thermal cutting processes, while air
carbon arc process is used both for cutting and gouging in the
/ Flame cones
Spray industry.
gun Base metal Oxy-Acetylene Flame Cutting Process is the most used process
Fig. 1.43. Schematic of a set-up for flame spraying,
for economic and high speed cutting of low carbon steels. In this
process a gas cutting torch, shown in Fig. 1.44, having some
The principle of operation in all three methods of spraying is resemblance to gas welding torch is employed. The gas cutting
that the material to be sprayed is melted by electric arc or plasma Lever
are, or gas flame and is atomised with the help of high pressure Cutting
air or inert gas and is projected on to the base material. The
sprayed material sticks to the base material due to its fluid state For'Oz -0:
Preheating
and high impact. Depending upon the temperature and pressure
the bond between the coating and the base material is of
-0:
-CzHz
mechanical nature or a complete coalescence. Mixture of
For achieving a good bond between the coating and the base Oz+CzHz
material it is very important to prepare the workpiece properly.
Depending upon the nature of base material it may be machined,
A A
shot blasted, chemically cleaned, or even bond coated. Overall a 00 Hole for cutting 02
clean but rough surface gives the best result if it is sprayed ) ~" Mixture of oo0
immediately after preparation. Apart from metals, the base Cutting! Oz+CzHz ~Holes for Oz +CzHz Flame
material can be cloth, leather, wood, concrete, or anyother porous oxygen Seetion at A A.
surface.
Fig. 1.44. Constructional features of a flame cutting torch.
Commercially, the process of thermal spraying is used in
machinery repairs and maintenance, and for providing protective torch not only provides a means to obtain an oxy-acetylene flame
coatings. Material deposited by spraying is usually of much less but also has a separate lever controlled paSij,ageto provide a high
thickness than that deposited by surfacing. The bond in spraying pressure pure oxygen gas jet which impinges upon the heated
is also normally of the mechanical nature whereas in surfacing it metal to cause oxidation and generation of heat by the following
reaction .
. is of the coalscence type.
3Fe + 202 ----+ Fea04 + heat (1120 kJ/mol) ...(1.4)
Typical applications of the process include zinc coating on However, the initiation of this reaction is possible only if the
turbine blades, armature shafts and cam shafts. Decorative work
metal to be cut has achieved the ignition temperature of 870°C or
by spraying includes spraying of furniture, toys and sign boards. above for steel. Once the reaction is initiated the flame is required
1.7. Thermal Cutting Processes . only to sustain it so a neutral flame of low energy is used. The
Thermal cutting processes is a family of processes in which metal which is oxidised (Fea04) has a lower melting point than
heat of an electric are, radiation energy, or an exothermic reaction the melting point of steel, thus the cut is achieved faster than that
is utilised to melt or oxidise a metal at an accelerated rate to by melting of steel.
achieve a cut. There are a number of processes which utllise the The oxygen jet also helps in blowing the oxidised metal or slag
heat of the arc to cut metals and they include shielded metal are, out of the cut or kerf.
ai r carbon are, plasma arc, gas tungsten are, and gas metal arc.
so Welding Engineering and Technology Welding and Welding ~rocesses
51

Cammercially, axy-acetylene cutting pracess is extensively
used far cutting mild and law allay steels far straight ar cantaured
cuts as well as far jaint edge preparatian far welding. It also.finds
though a limited use in cutting cast iran and stainless steels e.g.
in faundries to.remave gates and risers, etc, fram the castings.
In Plasma Are Cutting a metal is parted by melting using a
high velocity jet af ionised hat gas. The equipment employed is
similar to. that used far plasma arc welding, hawever, the gas
pressure used is higher than that emplayed for welding.
The plasma arc cutting tarch is af the transferred plasma arc
type with the warkpiece connected to. the anade of the de pawer
source, as shown in Fig. 1.45. The pawer source used is af the
High
frequency
generator
Tungsten
Resistor electrode
Pilot arc
relay
contact
1~ m Constricting
nozzle
Work
Fig. 1.45. Circuit diagram for plasma arc cutting.
draaping valt-ampere characteristic type with an apen circuit
valtage (OCV) range af 120 to. 400 valts. Higher OCVis used far
cutting thicker sections. The autput current range required is
usually 70 to. 1000 amperes.
The gas used far praducing plasma jet depends upon the metal
to. be cut, for example, carban steel can be cut by campressed air
while mast nan-ferrous metals can be cut by using nitragen,
hydragen, argan and their mixtures.
Almast all metals can be cut by plasma arc cutting but it is
particularly suited to. cutting aluminium and stainless steels. It
can also. be used far stack cutting, shape cutting, and plate
bevelling.
Air Carbon Arc Gouging and Cutting is a pracess af remaving
unwanted metal to. praduce graaves in plates and to.bevel edges
in preparatian far welding. In this pracess heat is praduced by an
electric
blaw arc af
it out between
the cuta by
graphite rad a~d
campressed airthe workpiece
which to)nelt and
may alsOpartially
axidise the material and hence help in lawering its melting paint.
The campressed jet usually follows the arc far blawing aut the
malten metal as shown in Fig. 1.46. Apart fram gauging the
process can also.be used for cutting the metals.
Electrode
Holder
o o
stream
Metal oxid e particles
~
Fig. 1.46. Schematic illustration of the process of air carbon arc gouging.
As axidation af metal is nat an essential requirement af the
pracess therefare all metals can be gauged ar cut regardless afhaw
rapidly they axidise. Material can be remaved appraximately five
times faster by arc gauging than by chipping. A 10 mm graave, far
example, can be gauged at a speed af 60 cm/min. Depth af cut can
be cantralled clasely, and welding slag daes nat deflect ar hinder
the cutting actian as it wauld with cutting taals. The cast af
aperating gauging equipment is generally less than far chipping
hammers ar gas-cutting tarches, and the arc-gauging equipment
also.requires less space. An arc-gauged surface is clean and smaath
and can usually be welded withaut further preparatian.
Mast of the standard arc welding pawer saurces, both ae and
dc, with an apen circuit valtage af 60 volts can be used far air
carban arc gauging and cutting. The 150 to.300.mm lang electrades
used vary fram 4 mm to.25 mm in diameter. Bath bare and capper
caated electrades are used, hawever the latter type finds mare
extensive use because they erade far less during aperation than
the bare electrades and because af better graave unifarmity
achieved by them.
The air pressure used is 55 to. 7r0 N/cm2 with an airflow rate
af 85 to. 145 lit/min. Where campressed air lines are not available,
arc gouging tarches suitable for lightwark can be operated on gas
from campressed gas cylinders at pressures as law as 30 N/cm2.
S2 Welding Engineering and Technology
Welding and Welding Processes S3
Arc gouging torches are generally air-cooled but those
intended for applications at higher currents are water-cooled. 7. Shift the welding head to the position wherefrom the next
welding cycle is to be initiated.
The correct use of arc gouging with carbon-based electrodes
usually causes no carbon pick-up and does not affect corrosion 8. Remove the completed work. This operation may be done
resistance of the parent material. A thin hardened zone may before or after repositioning the welding he&d or both operations
appear, by gouging, in some metals but subsequent welding may be carried out simultaneously.
remelts this zone and reduces the hardness. To accomplish the above task anyone of the following four
Heat penetration is shallower with arc gouging than with techniques may be employed.
oxygen cutting, so arc gouging produces less distortion. (i) Manual welding,
Air carbon arc process is widely used for gouging, joint edge (ii) Semi-automatic welding,
preparation and for removing defective weld metal. It is also used (iii) Automatic welding,
for scraping of metal objects-by cutting. (iv) Automated welding.
1.8. Modes of Welding 1.8.1. Manual Welding
All the afore described welding and allied processes are used It implies that all the eight operations of welding sequence are
in fabrication industries in different modes depending upon th~ carried out manually. Note, however, that stage 4 that is 'the
volume of production involved, nature of joint, material and its relative movement between the welding head and the work' may
thickness, and the accessibility of the spot to be welded. include some mechanical assistance such as a welding mani-
In most of the welding methods the following sequence is pulator which moves the workpiece at approximately the right
required to be followed to achieve the end product assuming, speed for welding. One such manipulator called gravity welder is
however, that the preliminary operations like cleaning, edge shown in Fig. 1.47 in which the welder winds up the weight, then
preparation, and the fixing of tab-in and tab-out plates have
already been accomplished.
1. Assemble parts by tack welding or by employing jigs and
fixtures.
2. Present the assembled workpiece to the machine or vice
versa.
3. Initiate welding by striking the arc for fusion welding or by Bearing
bringing electrodes in contact with the work and switching on the
current for resistance welding. Weld
loint
4. Create relative movement between the welding head and
the work to attain the desired welding speed.
5. Control the welding variables like arc voltage, welding
current, wire feed rate, etc. to control the arc length in the case
of arc welding processes and to control the depth of molten metal Hole in base
and slag pool in Electroslag welding.
6. Stop welding by stopping the relative movement between
the welding head and the work. If weld pool crater is to be filled
then the crater filler will become operative before the current is Fig. 1.47. Gravity motor for work manipulation in welding.
switched off automatically.
controls rotational speed of the circular table by holding the edge
and letting it run through his fingers at the desired speed enabling
54 55
Welding Engineering and Technology Welding and Welding Processes

him to produce neater, continuous welds on circular seams in the automatically. Fig. 1.48 shows a block diagram for a typical
downhand welding position. automatic welding system.
Manual welding is most popular with SMAW,GTAW,plasma In an automatic welding mode stages 1, 2, 7 and 8 are carried
arc welding, and oxy-fuelgas welding processes. out by hand or initiated manually. By above logic, gravity welding"
1.8.2. Sem~-Automatic Welding is classified as a portable automatic welding method.
The automatic welding mode is most popular with SAW and
In this mode of operation stage 5 that is 'the control of welding ESW processes. It is also used, to a limited extent, with GTAW,
variables such as wire feed speed in GMAW or the duration of GMAW,FCAW,and plasma arc welding processes.
current in resistance welding with a gun welder is automatic' but
the means of welding are held in hand. Stage 4, that is 'the relative 1.8.4.Automated Welding (Flexible Welding System)
.. "
movement between the welding head and the work' is normally u
~ .I..
00.
'" .
manual but mechanical means like conveyor belt or work oc
0'-
C
:;:
"5
"0 00.
:;:J

C7'

manipulator may be employed. Thus, GMAWprocess can be used ~"5

in conjunction with gravity motor to improve the quality and
productivity in welding . .!! 8
.0
~'"
Various operations in stages 3 and 6 that is 'initiation and .S!
>.
'"
stoppage operation' can be carried out in sequence automatically tlIl
c::
with the help of a single on-off switch. :.a
Q3
The semi-automatic welding mode is most popular with ~
C..l

GMAW and FCAW.Though it is possible to use this technique with '.g

GTAW, SAW,and ESW processeE'but it is rarely done. El
o
.•..
::I
1.8.3. Automatic Welding '"
c::
'"
It is a welding mode in which at least stage 5, that is 'the ...
o
control of welding variables' and stage 4 i.e. 'the relative movement c::

between the welding head and the work' are automatic. Usually a .S!
.•..
'"
single switch working through a sequencing device operates the d I1l

controls for power and consumables like wire and gas. This may e
also bring crater-filling device, if incorporated, into action 0.
~
C..l

~
El

Controller -5

o00000 00
'" 0
0'=
-c
tIJ
a;
...;<

~
o®o 0.0
IV
bD
i:i:
.$1",
IV
"'C
Q.~
Zoo

Shielding
gas supply
(where required)
Customized fixture •• Refer to author's book entitled, 'Welding Processes and Technology' for
Fig. 1.48. Block diagram for an automatic welding system. description of Gravity Welder.
56 57
Welding Engineering and Technology Welding and Welding Processes

In automated welding mode all eight stages from assembly by increased productivity; other advantages of robots are that of
and transfer of the parts to the welding head are performed consistent accuracy, minimum wastage of materials and stabilized
without adjustment of controls by a welding operator. The welding labour charges.

~~1S•
which may be completed in one or more stages, and the final vertical Vozrtical PG
2G 'fixed Iposition
HFlatposition
Work
PlacementIGVertical
H-L04S
PC
PErotated
PF
PAISO SGVertical
6GR U·S.A. I U.K. down down
ejection of the completed parts, 3G are
and carried
horizontal
without manual intervention. An important
allis out mechanically
--~
position
4G
SG Vertical
Pipe 'fixed aspect
lG
Inclined
position
of automated'
6G Overhead
(not PGup }---
}--
Vertical
position
fixed positio'~
rotated)
Restriction ring
)) })
PGlG3G
-0 t
4G PF

I
Pipe inclined 2G(T,K,or 6G
welding is that the operator need 2G not continuously monitor the ~ ~ Y connections)
operation. Compared to automatic ~ welding.""<0'
this tends to increase
productivity, improve quality, and reduce operator fatigue.· ~HOri%O"'" ~
Fig. 1.49 shows a schematic diagram >~ for an automated welding

system employing mini-computer, multi-prograli:lmer, and a seam

[JIJ PF
W"d~

tracing unit. The automated welding mode is populary used with

SAW, GMAW, and FCAW processes. To a limited extent GTAW,
PAW, and ESW processes are also used in automated modes.
Automated welding systems often employ Adaptive Controls
to accomplish two aims viz., seam tracking and quality control.
Apart from the conventional automated systems this mode of
welding is used for remote welding and robotic welding.
1.8.4.1. Remote Welding
Remote welding is similar to automated welding in that the
welding operator is not at the welding location and may be at a
great distance from it. While conventional automated system is
designed normally for making the same identical weld time after.
time, Remote welding usually involves maintenance operations
where each weld may be different from the previous one. Remote
welding is becoming more widely used with the increased
establishment of nuclear power plants. In general, it is performed
where humans cannot be present because of a hostile atmosphere, ,.
such as where high level of radioactivity exists.
One ofthe typical applications of remote welding is the sealing
of radioactive materials into metal containers. Sealing of fuel
elements and target rods in nuclear industry is also accomplished
by remote welding.
1.8.4.2. Robotic Welding
Robotic welding is a more fascinating aspect of automated
welding as articulated robots * can closely emulate the productive
actions of a man in the welding environment, and within limits
provide an acceptable alternative for performing many of the
monotonous and thus fatiguing tasks that are to be encountered
in industry in abundance. In such situations a rohot can be cost
effective solution to many welding tasks. Apart from cost reduction
* Articulated robot is a robot with flexible joints including wrist movement. Fig. 1.50. Standard welding positions for butt welds in plates and pipes.
 58
Welding Engineering and Technology Welding and Welding Processes 59

Although welding robots can be used for any weld joint for
. IPF
aXIs 2F
I4F02FR(rotated)
i 3FlFRPG US/UK down PG
which software programmevertical
to batch production.45
can be produced
I ' 45·
3F
but it is best suited
•weld: Vertical
horizontal
Vertical position ,,- I
-{. Vertical
vertical
up
up
~
Horizontal vertical
Work
[«:»-
PB
PO
Flat150
Placement
Overhead
I - I-
--[9-
L4S/PA
1F position
(Fixed) 4F
( rotated)
position
IF
SF
. 4S2FR down
Vertical

e
Pipe fixed axis horizontal IF ~ Axis of weld 2F
1.9. Positions In Welding Rotated
2F for welding posItion
I--- Axis

£;;'~~d~ 8
Hori~?ntal vertical of weld
4F
Depending upon the position during welding Ln all -.(
~'tl'al~~' butt and fillet
Pipe ~rizontal -[9- ~ ~
~
welds in plates are classified into four basic groups viz., downhand
or flat, horizontal, vertical, and overhead welds. AA;'~
According to internationally accepted norms a downhand or
flat weld is a weld in any direction on a horizontal surface; a
horizontal weld is one running horizontally on a vertical surface,
a vertical weld is that which runs vertically on a vertical surface
and an overhead weld is one that is deposited above the oper!ltor's
head.

Butt and fillet welds in pipes are classified in a slightly

different way depending upon not only the type of weld i.e. butt
or fillet weld but also whether the pipe is held stationary or
rotated. Further classification is also based on the inclination of
the pipe and whether there is any restricting ring on it to interfere
with the welding operation. All these positions of welds as per US
(UK) and ISO (International Standards Organisation) standards
are given in Figs. 1.50 to 1.54. While Fig. 1.50 gives the standard
welding positions for butt welds in plates and pipes, Fig. 1.51 gives
77min

SG

All dimensions in mm

Fig. 1.51. Detailed configurations for standard welding

positions 5G and GGR in pipes.
Fig. 1.52. Standard welding positions for fillet welds in plates and pipes.
60 Welding and Welding Processes 61
Welding Engineering and Technology
Flat detailed configurations for test welds in positions 5G and 6 GR for
PA pipes. The standard welding positions for fillet welds in plates and
Horizontal vertical pipes are illustrated in Fig. 1.52, and Fig. 1.53 shows the basis on
PB which the main welding positions are defined as per ISO. Fig. 1.54
shows the standard welding positions for stud welding alongwith
their limits of operation.
Horizontal
PC Horizontal
PC

Horizontal

PE
Overhead
Fig. 1.53. Basis for nomenclature of welding positions as per ISO.

15
25
45
( a) Stud welding positions
II ...
r

45·-+-- 45·

< m'5 X5' r

')xl
'r

<'-+-45'
, '"'
:·45
45·

( b) Limits of positions for plate or pipe

Fig. 1.54. Standard stud welding positions and their limits of operation.

TWO
Heat Flow in Welding
All fusion welding processes involve heat flow during welding
to accomplish the desired joint. Depending upon the heating and
cooling cycles involved different types of microstructures are
obtained in weld bead and the heat affected zone (HAZ).This leads
to varying mechanical properties of different zones of a weldment,
necessitating PWHT (post weld heat treatment) to obtain uniform
structure and the required service behaviour. Apart from the'
metallurgical effects of heat flow in welding there are other
phenomena involved including distortion, residual stresses,
physical changes, and chemical modifications. Thus, to achieve a
weldment of desired specifications to perform satisfactorily in
service it is essential to know the effects of heat during welding.
This can well be achieved by knowing the temperature distribution
during welding so as to determine the cooling rates in different
directions with respect to the weld axis.
2.1. Temperature Distribution in Welding
Temperature distribution in welding depends upon the nature
of the welding process used, type of the heat source employed,
energy input per unit time, configuration of the joint (linear or
circular), type of joint (butt, fillet, et.c.),physical properties of the
metal being welded, and the nature of the surrounding medium
i.e. ordinary atmospheric conditions or underwater. Although it is
beyond the scope of this book to analyse all these aspects of heat
:flow in detail but brief descriptions of the following cases are
included.
(A) Arc Welding
(i) Linear butt welds,
(ii) Circular butt welds,
(iii) Fillet welds.
Heat Flow in Welding 63
(B) Resistance .Welding
(i) Upset butt welding,
(ii) Spot welding.
(C) Electroslag Welding,
(D) Underwater Welding.
2.1.1. Temperature Distribution in Arc Welding
Nearly 90% of welding in world is carried out by one or the
other arc welding process, therefore it is imperative to discuss the
problem of temperature distribution in arc welding in the
maximum possible detail to arrive at the best possible under-
standing of the problem. Because linear butt welds are perhaps
the most used type of welds in welded fabrication therefore this
type of joint will be detailed the most.
2.1.1.1. Temperature Distribution in Linear Butt Welds
Heat flow in welding is mainly due to the heat input by
wetding source in a limited zone, and its subsequent flow into the
body of the workpiece by conduction. A limited amount of heat loss
is by way of convection and radiation as well but that can be
accounted for by allotting heat transfer efficiency factor at the
accounting of heat input. So, the problem of temperature distri-
bution can be seen as a case of heat flow by conduction when the
heat input is by a moving heat source. This case can be further
simplified by assuming workpiece of large dimensions to approach
the infinity concept i.e. the temperature at the farthest end of the
workpiece in all directions remains unchanged', This leads to a
condition of quasi-stationary state which can be defined as a
condition in which an observer at the arc will see a fixed
temperature field all around the arc at all times. In other words
under a condition of quasi-stationary state of heat flow isotherms
representing, different temperatures remain at a fixed distance
with respect to the heat source. Mathematically stated it means
~; = 0, where T is the temperature at any time and t is the time
unit. Fig. 2.1 represents the condition of quasi-stationary state
with observation points at A and B for w~lding along centrelineof
the plate.
To determine the temperature at any point, in a workpiece,
during welding, the problem can be solved by considering from the
basic Fourier's Equation of heat flow by conduction,
 64
Welding Engineering and Technology Heat Flow in Welding 65
h

-~
•..
00
c:
c:
i:3
:..:::
ell ~
....
~e••...
if;
:.a
.2S
\\bD\ .~ ~
"0 OJ

....•
~ .5
~'"::l
...•
'"
~
0-
'"
....
d•..'".... t;::
c.. ::l

'6Qj
0- '"0'"
...•
bD

I f
.S
i::-
dUx dUll +dx

Fig. 2.2. A case of one dimensional heat flow by conduction.

Mathematically stated,

dt = _ k . A . aT
q = :!:!l ...(2.1)
ax
where, dQ = quantity of heat conducted in time, dt .
Note: (i) Since the heat flow always occurs in the direction of

decreasing temperature, the temperature gradient (~~) will,

If one dimensional form of heat conduction is considered as
shown in Fig. 2.2, Fourier's Law states that the rate of heat
transfer per unit time, q, is a product of,
(a) area, A, normal to heat flow path,
the temperature gradient at the section i.e. the rate of
(b) ~~,
change of temperature with reference to the distance in the
direction of heat flow,
~c) k, thermal conductivity of the material of the body.
therefore, always be -ve, hence the negative sign on the right hand
side of the equation (2.1).
(ii) The following assumptions have been made in the above
equation, /,
(a) the heat flow in y and z directions is zero,
(b) the temperature in any plane perpendicular to the x-a:Xis
is uniform throughout the plane.
The equation (2.1) represents the fundamental heat con-
duction law for uni-directional flow of heat. As normally the heat
will flow in all the three directions in a given body, therefore a
comprehensive equation must deal with 3-dimensional heat flow.
Three Dimensional Heat Flow Equation
Consider an infinitely small solid cubic body as shown in Fig.
2 . .'3.
Let the edges parallel to the three axes be respectively dx, dy and
dz. The volume ofthe cubicelement is, therefore, given by the relation,
 66
Welding Engineering and Technology Heat Flow in Welding 67

d(
e I
1//(, I
f
I dQZ"dz.
I I dall +dll
/day+dydz
ax

ax tkk· . dy
=~
ax (dx . .aT)
aT) . dz dy .dx
ax

dz). dt
. dt
...(2.6)

From (2.6) and (2.2), we get,

...(2.7)
dQx - dQx ~dx = ax k· ax dv· dt
a ( aT)
Similarly, the net heat gained by the cubic element by
conduction in y and z directions may be obtained as follows,
dQy

a aT
Fig. 2.3. An infinitely small solid cubic body with
heat flow ~y conduction in 3-directions.
and dQz - dQz +dz = az k· az dv· dt
dQy - dQy + dy = ~ t( k . dv . dt
The total heat gained (dQ) by the cubic
~~j
element
... (2.9)
...(2.8)
(dx . dy . dz)
dv = dx . dy . dz ... (2.2) is the sum of the heat gained by conduction in x, y, and z direction.
Let dQx represent the total quantity of heat entering the face Thus, from equations (2.7), (2.8) and (2.9), we get,
area dy . dz in time dt as shown in Fig. 2.3. So, on the basis of dQ = (dQx - dQx + dx) + (dQy - dQy + dy) + (dQz - dQz + dz)
equation (2.1), we have,
dQx aT ax +
= dv {aax ( k· aT) a
cry cry +
(k aT) . az
a (k· aTJ~
az ~. dt
dt = - k (dy . dz) ax ... (2.3) ...(2.10)
Now, the heat gained by the cubic element (dx . dy . dz) can
Note: The gradient in equation (2.3) is expressed as a partIal also be expressed in terms of increase in internal energy dE,
derivative of the temperature T as it is a function of x, y, z and t. expressed as,
Now, a corresponding quantity of heat will be leaving the cubic dT
element (dx . dy . dz). Let it be dQx + dx, the value of which may be dE = m . s . t = (p . dv) . cP . -dt . dt ...(2.11)
obtained by Taylor Expansion of dQx + dx' whereby, where, m = mass of the cubic element,
a s = specific heat of the material of the element,
dQx+dx = dQx + ax (dQx) . dx + . ...(2.4) t = change in temperature of the material of the element,
p = density of material of cubic element,
Ignoring the higher terms of Taylor expansion in equation cp = specific heat of material at constant pressure,
(2.4), the net heat gained by the element (dx· dy . dz) due to
conduction in x-direction will, thus, be, Now, dQ=dE
a
Therefore, equating equations (2.10) and (2.11), we get,
dQx - dQx +dx = - ax (dQx) . dx ...(2.5) dv ~- k - +- k - +- k - dt = c pdv - . dt
(L ax
a (aT)ax a (aT)cry
cry az az
a ( aTJ~ p dt
dT
By substituting the value of dQx from (2.3) in the right hand ...(2.12)
side of equation (2.5), we get, Assuming k uniform in all directions of the cubic element,
equation (2.12) can be written as,
dQ - dQ d = - - - k (dy . dz) -ax dx· dt
x x+ x ax
(J { aT} k -+-+-
[fiTa2T
ax2 cry2 (J2T]
dZ2
·dt=c
P
p-·dt
dt
dT
 69
68 . Welding Engineering and Technology Heat Flow in Welding

cPT cPT ,iT c P dT where, v = the welding speed,

t.e., ax2 + dy2 + az2 =
k
T' it ...(2.13) t = time taken by welding,
starting from the origin, O.
Now, to determine the temperature distribution in the plate
Now, -- = ex ... (2.14) with respect to the tip of the electrode, we are required to change
cpP
equation (2.15) from cartesian coordinate system (x, y, z) to a new
where ex is known as the thermal diffusivity of the material of coordinate system (;, y, z).
cubic element and its unit is m2/sec.
Differentiating (2.16), we get,
From equations (2.13) and (2.14), we get, ...(2.17)
~=1
a2T a2T a2T 1 dT ax
ax2 + ay2 + az2 = ex dt ...(2.15) and at
~=-v ...(2.18)
This is known as Fourier's Equation of three dimensional heat
conduction in solids. Because temperature T is a function of distance; and time t,
To apply equation (2.15) to welding, let us consider the it can be expressed as,
...(2.19)
situation in arc welding as expressed in Fig. 2.4. T=f(;, t)
Differentiating T w.r.t. t, we get,
/ n Electrode
dT _ aT . ~ aT. dt
dt - a; dt + at dt
A aT aT (from 2.18)
Q =-(-v)+--·l
x / a;
=-v a;
aT
at
aT
+at
...(2.20)

Direction
x

of welding
-!ii--'
_
Also,
_aT =__.~
aT aT . 1
=.- d'f:. ( ... a; -=ax)
ax a1; dx al;
Fig. 2.4. Arc welding on a large plate. aT aT
Let 0 be the origin (i.e. the starting point for welding) for the
ax = a~ ...(2.21)
cartesian coordinates (x, y, z). Suppose, we are interested in Differentiating (2.21), we get,
finding temperature at any point A (x, y, z). Because of moving
heat source the distance of point Ais changing every moment with
respect to the arc or the tip of the electrode. If the origin of coordi-
~2T
ax2 =_iL
ax a1;
a2T a2T
l
(~r)=l..
a; (aT)~
a1; dx iJ; 2 .1
= a2T

nate system is shifted from 0 to the tip of the electrode (which is .. -2

ax
=~
a1;'
...(2.22)
assumed to lie in the plane of top surface of the plate) then the
temperature at any point A located at a fixed distancE)from the In the new coordinate system of (;,y, z), because there is no
tip of the electrode remains fixed because of the establishment of change in y and z of cartesian coordinate system, therefore their
a quasi-stationary state and can thus be determined mathe- derivatives also remain unchanged in equation (2.15).
2
matically.
Let point A in the new coordinate system with respect to the Substituting the values of ~dT, x and ~-~
t aaT ax from equations
tip of the electrode have the coordinates (1;", y, z) (2.20), (2.21) and (2.22) respectively in equation (2.15), we get,
.. I; = x - vt ... (2.16)
• 1; is pronounced as Xi.
':;-?"
;PT
o/; ~
-I-
ay +-2
-2
;PT az
cPT = ( - v aT
:I'f:.
0.., +aT)
0 -::;-t -
1ex
..• (2.23)
70 Welding Engineering and Technology Heat Flow in Welding 71

For convenience in further derivations,

1 --=e
a2T _ --
a2<j>
A.u~
...(2.31)
let -=21.
a ... (2.24) aZ2 aZ2
a2T a2T a2T aT
From (2.23) and (2.24), we get, Putting the values of --2' --2' --2 and :'1):' from equations
a1; ay az U,:>

cPT cPT a2T aT aT (2.29), (2.30), (2.31) and (2.28) respectively in equation (2.~6), we
'd1;2 + ay2 + az2 = - 2Au a1; + 21. -at ...(2.25) get,
When a quasi-stationary state is established then with respect
to the tip of the electrode, the temperature at any given distance f\" v e 'I' - f\"ve
('l2 2 - ).v~ th 2'l -
""I):
A.u~ ~u,:>
+ e- A.u~
2 + e
a1;
a2<j» (- A.u~ ay 2 + e
a2<j» (- A.u~ az 2
a2<j»
undergoes no change with time i.e., ~; = o.
= - 2 AU {- Aue-).v~ <j>+ e-).v~~}
Therefore, equation (2.25) changes to,
a2T cPT a2T aT or e- A.u~ {a2t+.
~a1;2 + ~
ay2 + ~
az2 + A2v2e- ).v~<j>_ 2Ave- a1;
--+--+-=-2Au - ... (2.26) a2t+. a2t+.} A.u~ ~at+.

a1;z ay2 dzz al;

= 2A2v2e- A.u~ <j>_ 2Ave- A.u~ ~
Equation (2.26) is thus the differential equation of the quasi- a1;
stationary state of welding . or a2<j>+ a2<j>+ a2<j>_ (A2u2) th = 0 ...(2.32)
Equation (2.26) is more easily handled if T is replaced by the a1;2 ay2 az2 'I'
following expression, Equation (2.32) is a more convenient form of heat flow
T = To + e- ).v~ <j>(I;,y, z) ... (2.27) equation for a quasi-stationary state of welding. It can be used for
determining temperature distribution in specific cases for exam-
where, To = initial temperature of the plate, ple, in a semi-infinite plate which is representative of welding a
and (I;, y, z) = a funcj;ionto be.determined.
<1> large thick plate.
Thus, the aim is to replace T by <1> in equation (2.26). Differen- 2.1.1.2. Temperature Distribution in a Semi-Infinite
tiating (2.27) w.r.t . .;, we get, Plate (3- dimensional case)
x

aT __ 'l - >..ve, th - AVe, ~ ... (2.28)

al; - f\"ue 'I' + e al;
Differentiating (2.28), we get,
--
a2T = 'l2 U 2e -
{I., AVe, th - f\"ue
'I' 'l - AVe, --
d<j>- ",ve
'l - >..ve, -a<j>+ e - AVe, --
d2<j>

al;2 al; al; d~2

~.e. -a2T _
- '"
'l2 u 2e - AVe, th - 2'lf\"ue-
'I' AVe, +e
-a<j>_ AVe, -a2<j> ...(2.29)
~ al; ~
Differentiating (2.27) w.r.t. y, we get,
Semi-circular
aT = e-AVe, ~ isotherms
ay ay
and -=e
(PT -
_ AV~ a2<1> .... (2.30)
ay2 ay2
Similarly, by double differentiation of (2.27) w.r.t. z, we get,
Fig. 2.5. Isotherms in semi-infinite plate with a point heat source on the surface.
12 Heat Flow in Welding 73
Welding Engineering and Technology

Considering a case of laying a single weld bead, using a point J

heat source, on the surface of a very large and thick plate (work-
piece), as shown in Fig. 2.5.
Let us assume that the Z-axis is placed in the direction of
thickness of the plate downwards. For determining the
temperature distribution the solution of equation (2.32) must
satisfy the following conditions.
(1) Since welding is done by a point heat source, the heat flux
through the surface of the hemisphere drawn around the source
must tend to the value of the total heat, Qp, delivered to the plate,
~,
Z Y
as the radius of the sphere tends to zero. If R is the radius of the I:Yz.zz
sphere, then the total heat flowing through the hemispherical I
I"
surface of heat source as given by Fourier's Equation will be,
, ---
I
....
aT •...

q = - 21tR . k . aR (2.33) Y

q ~ Qp as R ~ 0 (2.34)
where R = v"f,,2 + y2 + Z2 (2.35)
(2) Heat losses through the surface of the plate (workpiece)
being negligible, there is no heat transmission from the plate to
the surroundings
aT
l.e., az = 0 for z = 0 and R *" 0 ...(2.36) z

(3) The temperature of the plate remains unchanged at a great Fig. 2.6. Nomenclature of cylindrical coordinate system employed.
distance from the heat source Now, because y is a function of and '" Z

.. T = To 'for R ~ 00 ••• (2.37) aT _ aT . az aT. £I~ ...(2.40)

.. ay - al dy + a", dy
Now, condition (2) i.e. equation (2.36) assigns a semi-circular
form to the isothermf: located at sections parallel to plane YZ (Fig. Now, temperature <l>is a function of y
and z while y and z are
2.5), thus they are dependent only on radial distance 'Z' from the functions of and", therefore equation (2.32) can be converted into
Z

heat source. polar coordinates by determining the double differential of <l>w.r.t.

y and z in terms of and "'. Z
Keeping in view the semi-circular form of isotherms, equation
(2.32) can be written more conveniently in cylindrical coordinates, Using equations (2.38), (2.39) and (2.40), we get,
~,z, ",. as shown in Fig. 2.6.
where, Z2 = y2 + z2 •.• (2.38)
a2<l>
-=cos
ay2
2
"'--+---.
a",Z
sin2",
--+------
a2<l> sin2
l2
a<l>

az
'" a2<l>

a",2

A
Z

From Fig. 2.6, we have, _ sin 2~ . + Elin2", .~ ...(2.41)

y = cos '" Z l a",az Z2 a",
and z = sin '" Z
a2<l> . 2 a2<l> cos2 '" a<l> cos2 '" a2<l>
and -- = SIn "'---- + -'-- "- + -"-- --.

h__
az 2 a",2 Z az Z a",2
l.e., '" = tan-1 (;) ...(2.39) + sin 2'1' .. sin 2~ . ~ ...(2.42)
l d'Val Z2 a",
* IV is pronounced as psi.
 14
Welding Engineering and Technology Heat Flow in Welding 75

Adding (2.41) and (2.42), we get, fabrication, etc.) of finite thickness. Thus, equation (2.48) must be
d2¢1 a2¢1 a2¢1 1 a¢l 1 a2¢1 modified to account for the limited range of Z dimension.
dy2 + az2 = az2 + T al + j2 a'lfi ... (2.43) Let, the plate thickness = g
Substituting (2.43) in (2.32), we get, Thus, no heat losses must take place at level z =g, i.e.
~ a2¢1 1._~ 1. ~_ 2_
as 2 + al~n + l 0--'l + l 2 a'lf 2 - (AV) ¢I- 0 ..'<2.44)
. ~=o for z=g ...(2.51)
But as there is no effect of angle 'If on temperature ¢I,
Equation (2.48) can be further generalised by putting it into
~
another form.
" a'lf = 0
Now,
Hence equation (2.44) reduces to,
a2¢1 ~ 1. ~ ,,2 _ .
fa J a (~ ...(2.52)
+ al 2 + l 0--'l - (AL) ¢ - 0
as 2 ...(2.45) _e -_A_uR
R __
- 00
e
- AUZ ~-
/\,vr l) 1
-,..l"",d_l=
vl:.! +

Equation (2.45) is satisfied by the following function, where _r;;~--z-

r = \IS +y ... (2.53)
e-· AuR
and Jo is the so-caned Bessel Function of First kind and zero order.
¢I == C . R-- ...(2.46) Thus, equation (2.48) can be written as,
, r;;v.--i2"'
where, R = \IS· +l ...(2.47)
T - T (\ = ~.
21th e- AV!; f0 00
e Auz-JT+l.
- J 0 (~
/\,ur l). vl2l.dl+ 1 ....(2 54)
Putting c = 2~' we get {Refer to (2.33) and (2.34)}
To get a solution for finite thickness, g, of plate, let us replace
Q -AuR
in equation (2.54),
T - T0 -- ~
21th e .-AU!; !!._-
R ....(2 48)
/ _ra-:-;-
e" ;.uz11--.r::l by cosh [Av (g - z) \ll + 1 1 ...(2.55)
As already stated, for the present case of semi-infinite plate, sinh [Avg vl2 + 1 1
aT V
az = 0, for z = 0 and R '# 0 ...(2.36) I where cosh and sinh mean hyperbolic cosine and sine respectively.
and T = To for R ~ 00 ••• (2.37) ...(2.56)
Then, az = _ ~21tk e ,- AUE, (A v) 2 . fOO0 M (z) . Jo (Avrl) l.dl
aT
Equation (2.48) when converted to cartesian coordinate system
leads to,
where, M (z) = cosh [AV (g - Z) -vF+l] ...(2.57)
sinh [AV . g --.ll:.! + 11
T = Qp e- u (r-x)/2\x ...(2.49) But, for z =g,
21tkr .
where aT
r=17+y2t7 ...(2.50) M (z) = 0, hence -az = 0 ...(2.58)
k
and Ct=-- ...(2.14) and, for z = 0, M (z) = 1
pCp
hence, a~ ...(2.59)
Equations (2.48) and (2.49) are important relationships for dz = _ ~
21tk e- AUE, . (AV)2 foo0 Jo (Avrl) . l . dl
determining temperature distribution in a semi-infinite plate.
2.1.1.3.Temperature Distribution in a Large (infinite) However, 0.v)2 500
0 Jo CA.lJrl) . I . dl = 500 d [J1 (x) . xl =
0 d; 0 ...(2.60)
Plate of Finite Thickness
Therefore, boundary condition of equation (2.51) is satisfied.
Equation (2.48) can help in determining temperature distri-
bution in semi-infinite plate but the normal cases encountered are However, it. remains to be proved that conditions of equations
(2.33) and (2.35), viz.,
those of large plates (in ship-building and pressure vess'el
 Heat Flow in Welding 77
76 _ Welding En~~~~rin~~nd Technology

aT _[;""'L'-'l-'~ • R = -JS2 + y'" + Z2 = -J(-50)'" + (15)2 + (Or

- 2rrRk aR ~ Qp as R::: ,,~- + y + z ~ 0 ...(2 ..33) = 52.2015 mm = 0.0522015 m ...(ii)
and T = To for R -~ 00 •• (2.35) k = 55 kCaVm.hr.oC
To do so let us write,
= 55 x60
4.2x x601000 =.64167 J/ m.sec....
0C Coo.)
£u
cosh. (a - x) + eX [1 + e -2a + e --4a + ...)
Sinh x -_ eO'" . e-2a _ [".-x
l-e- 2a - f. + e- (2a.- Xl]
Qp =" VI where V = 18 volt, 1= 200 A
n=oo and heat transfer efficiency, ", for
:::e-x + L (e-
n=1
(2na-x) + e- (2na +X» ...(2.61) GTAW=0.65
Qp = 0.65 x 18 x 200 = 2340 J/sec (iu)
Substituting (2.61) in (2.54), we get,
v = 125 mm/min = 1 Ml:~ M = 0.002083 m/sec (u)

T-To=21tIc
Q
e-'Avl; -a-+ nn~lL'" -(e-
[e- i.vR
-~+ n
AvRn
R"--
n
e- AVR'n)~
...(2.62)
~_ 2340 [from (iv) and (iii)]
21tk - 2 x 1t x 64.167
where Rn = {(2ng~z)2 + ~+ yr =5.80 ...(vi)
and R' n ::: ..J<2ng + Z)2 +v+7 1 1 60 x 60
...(vii)
A = 2ex = 0.0625 2 x 0.0625 28800
Equation (2.62) satisfies equations (2.33) and (2.35).
2 x 60 X 60
To sum up it can be said that the following two equations can 50
be used for determining temperature distribution during butt - AU/; = 28800 X 0.002083 X 1000 [from (vii) and (u)]
weLding of large (semi-infinite) plates of infinite thickness and
finite thickness respectively . = 2.99952
Q ·-A,vR
.. e- Aul; = (e)2.99952 = 20.076 ...(viii)
T T _..::!:.I!.... -Avl; _e __ ...(2.48) - AuR - 2.88 x 104 x 2.083 x 10-3 x 0.0522015
- 0 - 21th . e - R e e
~ = ----0-.0-5-2-2-0-1-5'----
T - To::: 2~ . e- Avl; ~R- +n'"'
L~R 1
-_:+ ~R~' ...(2.62) _ (e)-3.1315 _ ...(ix)
Q [ _ AvR n n
n = ocr _ AvR -. AuR' )~ - 0.0522015 - 0.836
The efficacies of these equations can be checked up by solving Putting the values of (vi), (viii), and (ix) in (i), we get,
actual problems.
Q -AvR
Problem 2.1. Calculate the temperature of' a point T-~ o
=..::!:.I!....·e-Av1;
21th
.~-
R
(-- 50,. 15, 0) mm with respect to the arc as origin for laying a single
= 3.80 x 20.076 X 0.836 = 97.37°C
weld bead on a wide steel plate by GTAW process using 200 A and
18 V at a welding speed of 125 mm / min. Assume thermal Let To =20"C
conductivity of steel (K) = 55 KCal/m.hr.°c and diffusivity (ex) :::: .. T = 97.37 + 20 = 117.37°C Ans. 117.37°C
0.0625 m2 / hr. Consider the following two cases, (b) For Semi-infinite plate of 5 mm thickness
(i) the plate of an infinite. thickness (semi-infinite plate), Temperature distribution in a Sem.i-infinite plate of finite
(ii) the plate of 5 mm thickness. thickness is given by (2.62), reproduced as follows.
Solution. (a) For Sem.i-infinite plate of infinite thickness
Q -AuR
T-T o :::~e
21tk I; ---+'"
R ki --+---
n=l
Rn R' n ...(i)
T - To :::_:L
21tk
. e- AU" . ~--
R ... (i) [from (2.48») Q _ Al' [e- AuR n = ""(e- AvR. e-J...UR")~
All other terms in (i) are calculated, in part (a) abov~•.';E!xcept
Now, ~=-50mm, y=15mm, z=Omm the last two terms. For determining those, let'
 Heat Flow in Welding 79
78 Welding Engineering and Technology

n =1
e- IwRn e- MR'.
Putting the values of ~ n + R' n for n = 1 to 10, and other
... Rn = ..,j(2ng - Z)2 + 1;~-7 = 1(2 x 1 x 5 - 0)2 + (- 50)2 + (15)2
terms in equation (i), we get,
= -.)100+ 2500 + 225 = -.)2825= 53.15 mm ...(ii)
Similarly, T - T0 -- 21th
p.e R Me,
£..J --+'" --+--
Rn R' n
R'n = ..,j-(2-n-g-+-z-)2-+-1;-2-+-y2 =Rn = 53.15 mm ...(iii) Q [e- MR nn=1
_ = l1e- MR. e- AUR'.)~
-AvR - AuR' = 5.80 x 20.076 [0.836 + (1.695 + 1.3735 + 0.993 + 0.6565
. e
.. R----W-• n
n_e
. n
+ 0.408 + 0.2432 + 0.1406 + 0.0798 + 0.0449 + 0.0246)]
= 5.80 x 20.076 [0.836 + (5.6591)]
AV = 2.8 X 104 x 2.083 X 10-3 = 58.324 ...(iv) = 5.80 x 20.076 x 6.4951 = 756.3°C
'\",v R n = 58.324x Let To = 20°C
100053.15 =~100
1000' = 31 ...(v).
.. T = 756.3 + 20 = 776.3°C Ans. T(_ 50,15,0) = 776.3°C.
e-AvR. e-(3.1) .
2.2. Efficiency of Heat Sources
Rn = 53.15 = 0.847586122 ...(v£)
1000 It is evident from the solution of Problem 2.1 that to solve the
e- AuR
ne
- AIIR'
n
-. AvR
e'l
temperature distribution problems we require to know the
efficiency (ll) of the heat source used for welding ; where II is
.. --R---
n + R'-n = 2 -~ n = 2 x 0.84758f3122 = 1.695 ...(vii) defined as,
Similarly,
= Energy transferred to the workpiece (2 63)
for n =2,
e- AuRn e"': AuR' n
... (viii) II Energy generated by the heat source '" .
~ n + R' n = 1.37325 Thus, if the efficiency (11) of the heat source is known, the
e-----+
MR. e- AvR'n energy (Q) transferred from it to the workpiece, can be deter-
n =3, R
n -R-'-
n = 0.993
... (ix)
mined. In arc, electroslag, and electron beam welding,
Q == llVI ...(2.64)
e- AvRn
__ e- AuR' •
... (x)
n =4,
n + R' n = 0.6565
Roo where V and I are the arc voltage and welding current respectively.
e- AuRn e- AIIR'n In gas tungsten arc welding (GTAW)with dcen (direct current,
n =5, Rn
---+ R' n = 0.408
. ... (xi) electrode negative) the majority of the heat is produced by
electrons bombarding the workpiece (anode) that is as a result of
e-AuR. e-AuR'n the release ofthe work function and the conversion oftheir kinetic
n =6, ... (xii)
~ n + R' n = 0.2432 energy into heat at the workpiece. In GTAWwith a.c., however,
electrons bombard the workpiece only during straight polarity half
n=7,
e-
---+
Rn
AvR n e- AuR' •
R' n - 0.1406
.. .(xiii) cycle i.e. for a period when electrode is negative thus resulting in
significantly lower arc efficiency. Also, the heat loss to the
e- AuRn e- AuR'n surrounding can be rather high; particularly so for long arc
n=8, ... (xiv) lengths. A part of the heat generated is taken away by the cooling
~- n + R' n = 0.0798
water employed to keep the electrode cool.
e- e - AuR' •
--+
AuR
Rn -R-'
n
In consumable electrode welding processes like SMAW,SAW,
('- = 0.0449
n=9, ... (x~)
GMAWand FCAW, using dcep or a.c. the heat going to both the
e- AuRn e- AuR' •
n =10, -y- n + ~- n = 0.0246
.. .(xvi) * Work Function of a material is defined as the energy required, in electron
volts (eV) or joules, to get one electron released from the surface of the material to
the surrounding space.
80 Welding Engineering and Technology
Heat Flow in Welding
electrode and the workpiece finally lands up on the workpiece 81
""'_'1-.. _
through transfer of molten meta1. Thus, the heat transfer
£110

efficiencies of these processes are high. In SAW process, the heat .I.eUle ~.1. Heat Source Efficiencies of Arc, Beam and
transfer 11 is further increased because the arc remains under a Flame Welding Processes
blanket of flux, the heat loss to the surroundings is, thus, (Mter Sindo Kou)
minimised. s.
No. Welding lD...,..,.,.
'V" ••••• ......•
~_
",,,
The efficiency of heat transfer in ESW is lower than that in Efficiency, 11
1. (%)
SAW, mainly owing to heat loss to the water-cooled copper shoes Gas tungsten arc welding (GTAW)
and, to lesser extent, by radiation and convection from the surface (a) deen 50-80
of the molten slag. (b) a.e. 20-50
In EBW (electron beam welding) process the welds are 2. Shielded metal are welding (SMAW) 65-85
produced by the phenomenon of keyholing. These keyholes act like 3. Gas metal arc welding (GMAW) 65-85
hI ack bodies to the heat source and trap most of its energy; leading 4. Submerged arc welding (SAW) 80-99
to very high efficiency of heat transfer in EBW process. 5. Electroslag welding (ESW) 55-82
In laser welding the heat transfer efficiency can be strongly 6. Electron Beam Welding (EBW) 80-95
affected by the wavelength and energy density of the laser beam, 7. Laser Beam Welding (LBW) 0.5-70
the workpiece material and its surface condition, and the joint 8. Oxy-aeetylene gas welding (OAW) 25-80
design. For example, with well poli>Shed AI or Cu, the surface
reflectivity can be around 99%, i.e. the efficiency can be only about Thermal properties required for solving the temperature
distribution welded
commonly probleP-ls are listed in Table 2.2 for some of the
materials.
1% for a 10.6 11m.continuous wave COzlaser. For steels, especially
when coated with thin layers of materials that enhance absorption
of the beam energy (for example, graphite and zinc phosphate), Table 2.2. Thermal Properties of Some"Commonly
quite reasonable efficiencies can be obtained. When keyholes are Welded Materials
established during laser 7.No.welding, the efficiency
8.
5.
3.
6.
2. .3.2 x 10 can
of the process s. Aluminium
Monel 400"
(m2/sec) Inconel600'"
Carbon
9% 1400
1500
1300
1527
1650
steel
Ni-steel
Austenitic 1063
660
Material
Copper
Ti-alloy steel
Thermal
Volume
229
roC)
35.2 2.7
27.0
41.0
24.9
(k)
(J/m.sec.oC)
4.7
3.9
4.5 .
5.3
1.1
(8.5-10)
384.0
35.2
18.3
point
3.0
4.4
4.0
Melting 9.0
4.7
9.6
Thermal
thermal
8.0
X 9.1
conductivity 6
xdiffusivity
X 106
106 10~
x 10-5
10-5
rather be impressive. 1.4. (pcpl
X
capacity (a)
J/m3.oC
In Oxy-acetylence welding the energy transferred is given by,
. 48 kJ T h
Q (watts) = 11 x litre CzHz x \ C2H., x 36008
= 13.311 VC·,'2H2 ... (2.65)
where V C2H2 = volumetric flow rat.e of CzHz, litlhr,
48 kJ = the heat of combustion of CzHz .
Both these values refer to standard state of 1 atm, and 25°C
t.em pera ture.
The heat source efficiency in Oxy-acetylene gas welding varies
over a rather wide range as the efficiency decreases significantly
600 contains 76% Ni, 15.5%Cr, 8% Fe and 0.5% Mn.
with increasing fuel consumption rate Vc2 H·2 , because of incomplete "''''Monel 400 contains 66.5% Ni, 31.5% Cu, 1% Mn and 1%Fe.
combustion. Efficiency of oxy-acetylene gas welding is also found
to depend on the torch nozzle diameter, welding speed, material 2.3. Further
EquationsModifications of Temperature Distribution
thickness, and thermal conductivity of the workpiece. In Table 2.1
are listed the efficiencies of most of arc. beam and flame welding Different researchers have tried to modify Rosenthal's equa-.
processes. tions to determine more accurately the temperature distribution
under different sets of welding conditions. The ones put forward
• 111m (one micron) = 0.001 mm.
by Adams, and Wells have received wide recognition and are
included in the following sub-sections.
82 Welding Engineering and Technology . Heat Flow in Welding 83
2.3.1. Adams Modification
In the earlier treatment of problem based on Rosenthal's 5 (VdJ+ 2
solution the heat source has been assumed to be a point heat and ...(2.67)
source which is obviously not true, particularly for small sized (3-d tcase) -- ri
V . 4a8 [(TT2m) - (TTlm J~
U

workpieces. Thus, recognising the existence of finite sized weld The main difficulty in the use of heat flow equations is the
pools, Adams used the fusion line as the boundary condition and variation of physical constants (like k, p, C, etc.) with temperature.
modified Rosenthal's equations viz., equations (2.48) and (2.62). Energy absorption within the weld pool by latent heat and its
The following equations were derived by Adams for the peak subsequent release at the tail end of the weld pool on solidification
temperature, Tp, at a distance y from the fusion boundary at the is one reason why actual isotherms around a moving weld pool are
workpiece surface. more elongated than indicated by calculations. Wells equations
For 2-dimensional Heat Flow (2.65) to (2.67) give good corr~lation with experimental data for
__ 1_ = 4.13vygpc 1 (263) low carbon steel and can be adjusted to apply more accurately to
T pop
- T Q + Tm - T0 .... metals having high latent heats by using.a specific heat value
For 3-dimensional Heat Flow corrected by the following relationship.
...(2.68)
1 _ 5.44 1tka. ~ 1
Tp - To - Qp' v [ 2 + (2a) 2 ] + T m - To ...(2.64) where L = Latent heat
C(corrected) (1 + C~ m)
= CO~fusion, Callh-oC
where T m = melting point of workpiece material, C = the relevant constant
To = ambient temperature. Problem 2.2. Determine the average cooling rate for cooling
Other symbols have the usual meanings. from 800°C to 500°C along the weld axis of a 1cm thick steel plate
2.3.2. Wells Modification for a bead-an-plate weld made by GTAW process at a welding speed
of 12 cm/min. using welding current of 175 A at 20 volts. Take the
The relationship between the heat flow and weld bead dimen- melting point of steel at 1527°C, k = 41 W/m.oC, a = 9.1 x 10-6
sions is given by the Wells simplification of Rosenthal's Equation m2 / see and efficiency of GTAW process equal to 0.7.
by the following relationship.
Solution.

where, Qp
Qp = 8kTm +[i ~:J
= heat input, cals/sec. cm of work thickness,
... (2.65) Procedure (i) Using equation (2.65), find d,
(ii) Using equation (2.67), find t,
d = bead width,
Tm = melting point of work material, at _- T1-T2
C''')
m aT t
v = welding speed, Using Wells equation (2.65), we get,
d = weld pool width,
a = thermal diffusivity.
Q = 8kTm 5' + 4a ...(2.65)
From this equation it has been derived that time t for the (1 Vd)
material to cool between two temperatures Tl and T2 (T1 > T2) on Now, Q = llVI = 0.7 x 20 x 175 = 2450 watts
the centreline of the weld for the two dimensional case of complete (for 1 cm thickness of plate)
penetration in a single pass and the 3-dimensional case of a bead ., Q(metrethickness) = 2450 x 100 watts ...(i)
on plate are as follows.
Now in (2.65), we need to find ~~
t - - . 5 [VdJ+ - 2 ~ 2 - ~ 2 (2 66) 120 120 1
(2-d case) - d
V 4a4 [(TT2 J ,Tl J ]
(T .... v ='60 mm/sec = 60 x 1000' mlsec = 500 m/sec
 Welding Engineering and Technology Heat Flow in Welding 85
34
1
-·d .. r = "X2 :; y2 :;z~= x ...(ii)
...(ii) From (i) and (ii), we get,
.. vd =
4a. 4 x 500
9.1 X 10-6 = 54.945d T-To=~
Q
...(iii)
Putting values of different parameters in (2.65), we get,
Now, = T1VI= 0.65 x18 x 200 = 2340 watts
Qp
k=35W/m.oC
2450 x 100 = 8 x 41 x 1527
2450 x 100
(i +1~4.945d) . 125 1
v = 125 mm/mm = 1000 x 60 = 480 m/sec.
_l8 x 41 x 1527 -"5) _ (0.489 - 0.2)
d - 1153.85 - 54.945 By differentiating equation (iii) w.r.t. x, we get,
= 0.00526 m = 5.26 mm ...(iv)
(ii) Using equation (2.67), we have, ax t =~.
aTI 21th -1
x2 .
From (iii) and (iv), we get
...(2.67) ...(v)
TTlm]
t = ~v . [5 (~)8 + 2] [TT2m _ aTI
ax t = _ ~21tk
... 4 (T - Qp
To)21t2k2
2 _ 21tk (TQp- To)2

Putting the values of different parameters in equation (2.67), Also - =v ... (vi)
we have, at I T
ax

Now, cooling rate

aTI ax I t . ax
at x = aT at I T = - 21th Qp
(T - T 0)2 . v
l4 x 9.1 x 10-6 J [from (v) and (vi)] ...(vii)
1
t = 0.00526. 8 500 _ 1527]
800 (a) With no preheat
(5 (~5~0x 0.005261 + 21 [1527
500
aT
at ,..x -__ 21tx 35 -.J':L
m.oC x (500(2340)
- 20)2W(oC)2 x _1_
480 ~sec.
= 500 x 0.00526 [5 (0.2:9) + 2] [3.054 - 1.90875) __ .. (530)2 _1_ °
= 2.63 x 0.431 x 1.145 = 1.297 sees. - 2 1t 35 (2340) x 480 C/sec
:. Cooling rate from 800"C to 500°C is, = - 55 °C/sec negative sign indicates that
aT 800 - 500 300 ° ° temperature decreases in the
at = 1.297 = 1.297 = 231.3 C/sec Ans.231.3 C/sec.
(b) With 250°C preheat
direction of heat flow.
Problem 2.3. Determine the cooling rate at 550°C along weld
axis for welding a wide thick .';teelplate using GTAW process at a
welding speed of 12..'5 cm I min with' a welding current of 200 A at at
aTl' 21tx 35 x (5502340·
x -__ - 250)2 x _1_
480 °Clsec __
- 17.62 °C/sec,
18 volts. Compare the re.c;ult.<;
if the plate has a preheat temperature Thus it can be seen that the cooling rate is significantly
of 250°C. Take heat conversilJn efficiency = 0.65, k = 35 WI m"C reduced by the use of preheating. Preheating is particularly
and the room temperature a.'i20"(;. important for single run welds because in multi-run welds the
Solution. For semi-infiniw plat(~s, the temperature distri- interpass temperature is equivalent to the preheat temperature
bution is given by equation (2.49), i.e., in single pass welding.
2.3.2.L Weld Characteristic
T _ T o -- 21tQ"
. k . r e /.I, 7)/'ln ." (')£ With the processes in which heat is used, the pattern o(energy
Along the weld axis, y:: Z :: 0 conversion to heat and its subsequent dissipation after welding is
86 Welding Engineering and Technology
Heat Flow in Welding 87
a major factor influencing the utilization of the process and the frequently expressed in terms of Joules/mm of weld length for a
properties of the joint. The efficiency of energy conversion, as moving heat source and Joules/second for a stationary heat source.
already discussed, varies considerably from one process to
another. However the efficiency with which the input energy is 2.3.2.2. Weld Bead Dimensions
converted is not of prime importance; greater importance is Based on Rosenthal's equation for 3-dimensional heat flow
attached to efficiency of heat transfer, energy level and energy Christensen et al have derived theoretical relationships between
intensity. These factors influence, the weld dimensions and the welding conditions. using
(i) the welding speed, and dimensionless parameters D and n which are related to each other
(ii) the size of the heat affected zone. as shown in Fig. 2.7, and can be expressed by the following
10
The former has economic implications and the latter may
affect the joint properties. 0,
Not all the heat reaching the work is available for melting the -a.
.c
•..
workpiece. Some heat must be utilized for building up the "0

temperature gradient. With high heat input conditions, allowing VI

high welding speeds, heat dissipated in the workpiece is mini- ~ 0.1

c:
mised but can never be reduced below half the total heat available. o
Vi
c:
Inefficient processes or processes used at relatively low speeds E 0.01
result in losses three to four times greater than this. A measure Ci

ofthe efficiency of utilisation of heat is given for the 2-dimensional

heat flow case by Weld Characteristic devised by Wells. For fusion 0.001

welds this is a non-dimensional term

(2.65) to (2.67). Thus,
r~ encountered in equations
0·001 0'01 0·1 1·0
Operating parameter.n
10 100

Fig. 2.7. Relationship between dimensionless operating parameter n

and the dimensionless weld depth D.
Weld characteristic (w.e.) = ~~ ...(2.69) relationships.
where, V = weld speed, mm/sec,
d = melted width, mm, Dimensionless depth, D =P2: ...(2.71)
a. = thermal diffusivity, mm2/sec. and Dimensionless operating parameter,
The corresponding weld characteristic for resistance spot Q .v
welds is, n= 2 p ••. (2.72)
41ta. pc (T m - To)
d2 where p = weld penetration,
W.C. = 4a.t ...(2.70) Tm = melting point of work material.
where, d = spot diameter, mm, Since, by Rosenthal's equation weld bead is symmetrical with
a. = thermal diffusivity, mm2/sec, respect to the x-axis therefore the width of the weld bead, d = 2p
t=time, sec. and thus the cross-sectional area of the weld bead can be
Welds in which there is a high heat input rate and the heat is determined. Also, equations (2.71) and (2.72) can be equally
effectively used to determine HAZ width by substituting T m with.
therefore used efficiently ~~ will exceed 1. A low heat input rate, Ts i.e. the solidus temperature of the work maLerial as is done in
allowing wasteful spread of heat, will be indicated by a weld the following problem.
characteristic of less than 0.1. For the majority of situations the Problem 2.4. Determine the width of HAZ for welding thick
weld characteristic lies between 0.25 and 1. Heat input is plates of 9% Ni steel using GTAW process at a welding speed of
88 Welding Engineering and Technology Heat Flow in Welding 89

125 mm / min using welding current of 200 A at 20 volt. The 1

melting point of steel is 1400°C and solidus temperature 1325°C.
Solution. n =-------.---------- 2800 x 480
41t(1.1 X 10-5)2 x 3.2 X 106 x (1325 - 20)
(a) For Weld Bead. Using equation (2.72), determine the = 2800 . = 0.918
dimensional parameter, n, 480 x 41t x 1.21 X 3.2
10-10 x X 106 x 1305

n= " Q. v ... (2.72) Putting the value of n in Fig. 2.7, we get the value of dimen-
41t a:" pc (T m - To) sionless depth, D = 0.9.
Now, Now, to determine the penetration of HAZ, use can be made
Q = TlVI = 0.7 x 20 x 200 = 2800 watts ...(i) of equation (2.71), i.e.,
(assuming heat transfer efficiency = 0.7) D=~ ...(2.71)
2a
v = 125 mm/min = M1215~M m/sec = 4~0 m/sec ...(ii.) D·2a
or ... (vi)
Putting values of different parameters in equation (2.72), we p= v
get, Putting the values of different parameters in (vi), we get,
1
n= 2800 x 480 HAZ t t·
pene ra lOn,p =
0.9 x 2 (1.1
1
X 10-5)

4 x 1tx (1.1 X 10-5)2 x 3.2 X 106 (1400 - 20) 480

2800 = 480 x 0.9 x 2 (1.1 X 10-5) = 0.009504 m
= 480 x 4 x 4 x 1.21 X 10-10 x 3.2 X 106 x 1380 = 9.504 mm ...(vii)
= 0.868 :. HAZ width = (vii) - (vi)
Putting the values of n = 0.868 in Fig. 2.7, we get dimension- = 9.504 - 8.448 = 1.056 mm
less depth D = 0.8. Ans. HAZ width = 1.056 mm.
Now from equation (2.71), we have, 2.4. Heat Flow in Fillet Welds
~ ...(2.71) A simple approach to analyse heat flow in T-type fillet welds
D = 2a is to assume that the total heat supplied from the arc is distributed
D·2a ;' in the three plates in the ratio of their thicknesses. Temperature
.'. Weld bead penetration, p = --- v ...(iii) distribution in the three plates forming the fillet joint can then be
Putting the values of D, a and v in the above equation (iii), we determined individually with the help of formulas used for
get, determining temperature distribution for laying bead-an-plate on
moderately thick plates i.e. equation (2.62). This approach implies
that if three plates have equal thicknesses, temperatures at points
p = 0.8 x 2 (IiIX 10-5) = 480 x 0.8 x 2 (1.1 X 10-5) equi-distant from the centre of the weld should be same in the
480 three plates. This, however, does not hold good at early stages of
= 0.008448 m = 8.448 mm ...(iv) heat flow from the weld centre though all three heat distributions
(b) ForHAZ approach similarity as the time passes. This leads to a conclusion
Ts = 1325°C that the bead-an-plate analysis can be applied successfully to fillet
welds for determining temperature distribution except (i) at the
Dimensionless operating factor, n = 2 Q .v ... (v) early stages of welding, and (ii) 0 the points close to the arc. The
41t· a pc (Ts - To) deviation is large when the arc is passing just over the point under
Putting the values of different parameters in (v), we get, consideration. However, it decreases and ultimately the
 90 Heat Flow in Welding 91
Welding Engineering and Technology

temperature distributions show no differences as the time passes, where the weld is laid shares the least amount of heat. This
as shown in Fig. 2.8. Such a deviation at earlier stages of welding condition makes fillet welds susceptible to high degree of
distorticill and non-uniform metallurgical changes compared to
/
800
,Welding vel. :: 2.54mm/see
1 see.current:: 223 A butt welded joints.
I
I 5 ee.voltage :2 5 V . , 2.5. Heat Flow in Circular Welds
.u Primary voltage :: 380 V
Power = 3.9 kw
I~ 2
The highest strain in welded structures form after making
I'
- welds that do not finish on the free edges of the workpiece. This
QI
t.-
::J
C 2 , 3 includes a large group of circular welded joints i.e. the welding of
t.-
QI 10mm away from fillet eentre various types of patches, flanges, nipples, connecting pipes and
E
c.. many other cylindrical components. The strains in such welds may
~ be considerably reduced by altering the design of the weld and the
) fabrication technology employed. To do so, it is imperative to
evaluate the temperature field formed in the component during
welding, and to select correctly the optimum parameters that
I I f r
control this field.
15 o -15 -30 - 45 - 60 -75 -100 -105 -120
lime - Considering the temperature dependence of the thermo-
Fig. 2.8. Time-temperature distribution curves for physical properties of the material i.e. K (T) = Ko (1 + t"T)' and the
three plates of a fillet welded joint. heat transfer from the surfaces of the workpiece, the volume non-
equilibrium distribution of heat in workpiece in the welding of
a fillet joint can be accommodated by introducing a factor which circular joints is described in the cylindrical system of coordinates
approaches unity with time. If equation (2.62) is multiplied by this (r, 'l', z) by the following equation,
factor, the temperature deviations at earlier stages can be duly 1
accounted for finding the true temperature distribution in fillet T = ~ {(";1 + 2(9) - 1) ...(2.76)
welds. For this purpose an exponentially varying factor ofthe form where, Ko = the reference coefficient of heat conductivity W/cmoC,
(1 _Ae-Bt) is considered most appropriate where A and B are cons-
r. = temperature coefficient ofheat conductivity lIoC,
tants, the magnitudes of which depend upon the thickness ratios
of 9 = Krikhofs variable, defined by the following equation.
as,three plates, etc. If equation (2.62) is expressed in a simple form
(T - TO)b =Qf (e:" y, z) ... (2.73)
I'
9= z;f}--
ciJn=1
L Bn . An (z).
(where b stands for bead-on-plate welds) then the modified
relationship is obtained by multiplying equation (2.73) by a factor r + rO - 2rrO COS 'l' - ---r---
mentioned above. Thus, for fillet welds the temperature distribu-
ti.on can be expressed as, , ft
o
jl -
2 2
4at [Vw(t-'t)]
0 Iln2at d't ...(2.77)
- --g2 - }J ~
e
(T - To)r= Qf (e:" y, z) (1 _Ae-Bt) (2.74) where, q = specific power of the welding are, callsec,
or (T-To)l=(T--To)b(1-Ae-Bt) (2.75) g = work thickness, em.
where f stands for fillet welds. Zo . Zo
The values of the constants A and B as reported by Gupta and Iln cos Iln - + SI SIn Iln-
Gupta are, B - g g ...(2.78)
n - 2 S2
A = 0.598 and B :; 0.029 Iln 2 + S 12 + Il" ..- 1 • Iln + 2S'1 SIn2 Iln
COSIln SIn
It is further reported by the same authors that in T-type fillet
welds the vertical plate shares the maximum instantaneous heat
while the back portion of the flange, that is the one opposite to * r, is a Greek letter, and is pronounced zeta.
 92
Welding Engineering and Technology Heat Flow in Welding 93

and An (z) = Jln ~ + 81 sin ~tn ~

cos Jln g g ...(2.79) .
where, ro = radius of the circular weld, em,
2
Vw = welding speed, em/see,
ex = thermal diffusivity, cm2/sec, ~
4
I-ln = roots of the characteristic equation (2.80),
(81 + 82) Jln
tanlln= 2 - ...(2.80)
Jln - 8182
Zo = depth of immersion of the arc under the work surface, em,
81---L
ko
alII
...(2.81)
and 8
2 -
--L
a~
ko ...(2.82)
2S 20 1S
r, em
where, aI' respectively the coefficients ofheat transfer from
a2 -
the surface of the work atz = 0 andz =g, W/cm20C Fig. 2.9. Temperature distribution in sections 1-3 and 2-4 when
the arc has travelled through half the circumference for
and L - Coefficient of linear isation of the boundary condition. welding a circular patch without any backing strip.
Considering specific cases of automatic welding of circular
joints with a radius of 15 em in 8 mm thick plates of AlMg6 alloy 2
at a speed of 36 m/hr (1 em/see) under three different heat transfer'
conditions, viz., •
-i(i).
,
-+
' 1
4
(i) welding without any backing plate,
(ii) welding with a backing plate of steel,
(iii) welding with the lower surface of the plate water cooled.
The coefficient of heat transfer at the bottom of the plate for
the above three conditions were taken respectively as
2.092 x 10-3, 20.92 X 10-3 and 209.2 x 10-3 W/cm2_oC.
Welding was done in a single pass with a 2 mm diameter wire
using a welding current of 308 A with dcep polarity at an arc 2 S 20 15 10 S 0 S
voltage of25 V and a wire feed rate of 500 m/hr (8.33 m/min). The r, em
temperature distribution patterns obtain,ed respectively for the Fig. 2.10. Temperature distribution in sections 1-3 and 2-4 when
three cases were reported to be as shown---inFigs. 2.9, 2.10 and
2.11.
the arc has travehed through half the circumference for
welding a circular patch with a backing plate of steel.
Fig. 2.9 shows the temperature field in sections 1-3 and 2-4. Fig. 2.10 shows the temperature di.stribution when AI-alloy
Jt is seen that in the case of heat transfer from both sides of the plate is resting on a steel plate during welding. It is evident that
plate to the ambient air central part of the plate is heated to high the temperatures are all along reduced, as compared with the first
temperatures. In the course of welding, the heat source moves over case, indicating a higher heat sink effect provided by the steel
th e already heated zone resulting in increased dimensions of the backing plate.
weld pool and that displaces the pool towards the centre of the
circle. Fig. 2.11 shows the temperature distributiun when the AIMg6
plate is water-cooled at the bottom. It is evident that in this case
 94 Heat Flow in Welding 95
Welding Engineering and Technology

700 s ~ 0 for r ~ 00 ••• (2.86)

I The initial conditions are:

600 e = 0 at t = 0 ...(2.87)
500
2
.w
01-
<-
400 Consequently, we obtain, (2 2 2 J
=>

C 300 e = 21tkgt 2at0)nY:IBn

q 10 (rr ~ . An (z) e - r 4:;/ _ Il;~t ... (2.88)
<-
01
~ 200
Where 10 is the modified Bessel Function of the zero order.
~ If the circular source acts for a specific period of time to, the
100 solution for the given problem may be written as follows.
o r +ro ~t1a't
o 5 10 1S 20 ',em
( 2 2 2 J
o S 10 1S 20 2S
I.
[;1 \\\~rl'j/;
z,CM 0
Fig. 2.11. Temperature distribution in a plate when the arc has travelled
-2a't n ~= 1Bn . An (z) e -4O:t-7 d't
e = ~21tk g- Itt - to 10 (rro) 't
If the temperature field is observed from a steady circular heat
source (to ~ 00), we obtain,
... (2.89)

through half the circumference for welding a circular patch

when the plate is water-cooled at the bottom.
S = -:£gn~lBn . An (z) {10 (r~n )Ko (!:inJ8 (ro - r)
the source moving along a circle may be replaced by a circular
source applied to the entire circumference of a circle with radius
"0. however this substitution cannot be applied to earlier two ..
+ 10 ( ) Kogrof.l.n rf.l.n
()g-
8 (r - ro)} ... (2.90)
where Ko is the Modified Bessel Function of the zero order, and
,
cases. In practice, the concept of a circular heat source may be
used with success in calculation of temperature fields in certain
8 is the Heaviside Function :
specific cases. These cases include the heating of a sheet surface
with a rapidly moving electron beam, with high frequency 8=1.
induction coils, the resistance welding of nipples, GMAW at speeds x 2' x=o· ,
above 60 m/hr, etc. 0, xx>
{I, < 0;0 ;
Taking the above considerations into account temperature Fig. 2.12 shows the temperature field in the welding of 8 mm
distribution under the effect of a circular source can be de~ived. thick plate with the circular heat source and without any backing
The differential equation of heat conduction for this case has the strip. The calculated results given in. Fig 2.12 are based on
following form : equation~ (2.76) and (2.88). This figure indicates that at high
temperatures the data calculated with and without the consi-
iio + 1. de + cPe = 1. ae _ ----9-_ .
iJr2 r dr dz2 a dt 21tkr deration of the effect of temperature on thermophysical properties
have considerable differences.
g (r - ro) . g (z - zo) g (t) ... (2.83)
For calculations in the above mentioned cases the AlMg6 plates
and the boundary conditions are, were considered with the values of different parameters as given
ae below.
-a-.;-·-·a10=0 atz=O ...(2.84)
as S = 0.00068, a = 0.5, ko (Th. cond.) = 0.28,
:;--+a2S=O
dz atz=g ...(2.85) L = 0.877, g = 0.81 and Zo = o.
96 Welding Engineering and Technology Heat Flow in Welding 97

heat generated (Q), the longer the time of generation, the larger
the fraction that leaks off. Further, it can be shown mathe-
., -<jr matically that the rate at which the heat leaks away from the weld
..
~
ell
...
is a maximum at the beginning of the weld period (because of cold
workpiece) and that the amount of heat loss is proportional to the
square root of the duration of the weld time.
;;:;J
-+-
::l
...
ell
c.. 0.30 100
E 60
75
t-Q/
100 50
o
o 2 18 20 22 21. 26 30~
0,
>-
r,cm •...
c:

-'5
"0
CIJ

Fig. 2.12. Temperature field resulting from the use of a circular heat sou'rce on e: 0.15
:::J
'u
;.-:=
the surface of the plate of AlMga alloy. o
o 4 8 12 16
Duration,cycles of SOHz currEnt
5 ~
0.10
2.6. Heat Flow in Resis~ance Welding 'i5
VI
:::J

In this section two types of resisting welding processes viz., Cl

spot welding and zonetl w~lding are discussed. The discussion will
be confined to resistance 'l'lpot welding and upset butt welding to
represent the two types of welding processes. -'
2.6.1. Heat Flow in Resistance Spot Welding
There are three important variables in resistance spot welding
VlZ. , current, time, and pressure. In any resistance welding process
the heat generated by the passage of current is given by either of
the following two equations.
(i) Q::; L [2 R'tdt
(ii) Q::; ro IVdt
where, Q::; heat generated, watt-see or joules,
[::; welding current, amps.,
R ::;contact resistance bet.ween workpieces being
welded, ohms,
\! ::;voltage across the secondary, volts,
t ::;total time for which the current flows, sec.
Current, voltage, and resistance all vary with time.
Of the total heat generated i.e. Q, only a fraction is used to
make the weld. The balance leaks into the work, and more so to
the water-cooled copper alloy electrodes of high electrical and
thermal conductivities. It is obvious that, for a given quantity of
ex: 0.05
o
o 1 4. 9 16
T,sec
Fig. 2.13. Variation of melting efficiency with time. in resistance spot welding.
If heat transfer efficiency be taken as the ratio of the volume
of steel actually melted by a fixed quantity of heat to the volume
which could be melted if no heat were lost then the weld time
versus efficiency curvefor Q = 300 cals is shown in Fig. 2.13. This
(2.91) curve is plotted for a theoretical spot weld made by generating
heat at a point source between two steel sheets in contact. It can
(2.92) be seen from this figure that the maximum possible efficiency
obtained, when the heat is generated instantaneously, is 60%. For
300 calories generated in 1 see, the efficiency is only 1%. For a
typical spot welding operation, 3000 calories per second for 0.1
see, the efficiency is about 10%. Thus, it is evident that for
resistance spot welding as well as for other similar processes, for
example, seam welding, projection welding, etc., these must be in-
herently short time, high current processes. Most of the features
of these processes, and many of the industrial problems met with
in their applications, are caused by this limitation.
Considering the case of resistance spot welding, let Q calories
of heat be generated instantaneously at a point within an infinite
 98 Welding Engineering and Technology Heat Flow in Welding 99

body. The temperature distribution about the point, as a function It is therefore imperative to use very high welding current in all
of time and distance is given by, these resistance welding processes.
2
r 2.6.2. Heat Flow in Upset Butt Welding
Q -~
T - To = V? e 4at ... (2.93) In general the welding processes require that a certain
8cp (1tat)
temperature be reached and maintained long enough for the weld
This equation satisfies the following differential equation of to be completed. For example, in upset or resistance butt welding
radial heat flow, the aim is for an interface temperature of the order of the solidus
temperature, Ts, of the work material. This temperature when
at
aT = r2 ex a (r aT]
'Or'Or ... (2.94) reached needs to be held only until the oxide layer at the interface
has been dislodged by fragmentation or diffusion, or until
for the boundary condition that all the heat remains within the sufficient lengths of the workpieces have been heated to permit
body. upsetting.
It is of interest to know the peak temperatures at various radii, Let us consider upset butt welding of two round bars of steel,
especially in the radius of the sphere within which each of length l and cross-sectional area A. Assume that the
interface is raised instantaneously to the solidus temperature,
Tmax = 1485°C, T, of course, equals T max when ~; = O.Differenting
Ts, and that the opposite ends of bars are held at room tempera-
with time equation (2.93), we get, ture. To determine the temperature variation, along the length of
the bar, with time let us consider an element of thickness dx of
A
at
(aT]J. = 8cp (1tat)3/2
Q e- ) L x (r24ext -
4at ... (2.95) 3
2t
\
,'"
"','
At a finite time, and radius not equal to zero, the only term I '. ' To

that can vanish in equation (2.95) is the one within the " I '
" " I
parentheses, and thus, ..•. '"

r2 3
T=T when ----=0
max 4at2 2t
r2 Fig. 2.14. Model for heat flow in resistance butt welding.
or T = T max when t = 6ex ... (2.96) the bar centred x cm from the interface as shown in Fig. 2.14.
Substituting these values in equation (2.93), we get, Assume that the interface has been at temperature Ts long enough
so that the element has a temperature above the room tempera-
T max = ~ ...(2.97) ture, To: but has not attained as yet its steady-state temperature.
8cpr ( 7te6·
J- 312
Applying Fourier's equation, the heat flow at section x is given
Thus, for a given amount of heat input, the peak temperature by,
at any point is inversely proportional to the cube of the distance
of this point from the point at which the heat is introduced.
Putting 3T=1500°C, Q=.300 calories, C=0.12 cal/gm.oC,
q = - kA (~= 1 (2.98) ...
• where, e = T - To,
P = 7.8 g/cm , we get, r = 0.25 cm.
k = thermal conductivity of bar material which is assumed
This locates t = 0 point for the curve of Fig 2.13. Weld time to be independent of temperature, T.
versus efficiency of heat utilisation in spot, seam, and projection
welding is given by the inset curve in Fig. 2.13. It is evident that The partial derivate indicates that the heat flow at a definite
the weld time duration should approach as Iowa value as possible. time t is being considered. At face I of the element, heat is flowing
at the rate qj as given by the following equation,

, iIilL
 100 Heat Flow in Welding 101
Welding Engineering and Technolog:r
e == es at x = 0 t ~ 0,
e :;::0, at x == ± l for all values of t,
qI = - kAax
[ae ~ ax
a (ae] '2 't
a~ dX] ••• (2.99) e == 0 at t == 0 for all values of x between ± l except at x == o.
where the second derivative indicates that there may be a change It is also known from steady-steady-state heat conduction
in gradient as we move from plane x to the boundary plane I. theory that when t == 00 the solution of equation (2.103) must reduce
Similarly at face II, we have, to,

a (ae] 2 ...(2.100) e
--1
e - ~ -l
x
qII == - kA ax + -a.; aX' ...(2.105)
[a9 dX] s
If surface losses are neglected, the difference qI - qII == qn gives The most general solution to equation (2.104) for the period
the rate at which the heat is accumulating in the element, representing the interval before any heat has reached the ends of
the bars is a temperature distribution given by,
.. qn == kA ax2 dx
(a~) ...(2.101) . e =
-es 1 -- 1t
2 J(f"0 e ._u2 d u ...(2.106)
Now, if the heat is accumulating in the differential element,
its temperature must be changing. Taking c as the specific heat x
where
of bar material and p its density, and assuming that these also.do u==...Jat

not change with temperature T, thus qn can also be given by the

following relationship, Values of ~ J: e-u2du known as the Gaussian error function
(erf u) can be obtained from standard mathematical tables.
qn == C . P . A dx at
(del ...(2.102) Equation (2.106) fails if it indicates that the temperature at ± l is
increasing. Thus, to test the valadity ofa temperature distribution
1
where (~~ refers to the rate of change of temperatu~e on the
obtained by its use, it is merely necessary to calculate e at x = l at
the desired time. If et is greater than 5 or 10 degrees (above the
plane x. Equating the right hand sides of equations (2.102) and ambient), the distribution will be inaccurate. Fig 2.15 illustrates
(2.101), we get, 2000
Stul IX;:0.09Scm2, sec Theoretical
o:t =0.1 see AClual
cp at )
(ael==k. ax2
a2e 1500 b:t ;: 1 sr.c
c:t = 10sec
ae k a2e a2e d:t = SOsec
or -at == - . '-- == a -dX2 ...(2.103) r.:I;: OD
cp dX2
.~1000
CD
where a = ~,
cp the thermal diffusivity of the bar materiaL
Equation (2.103) is a general equation for any case of
unidirectional heat flow through a body containing no heat
source(s) and no heat sink(s) other than its own heat capacity. -8 -6
\Vhen extended to three dimensional heat flow system it reads as
'O\t=a
o
-2+-2+-2
dX ay ...(2.104) r'ig. 2.15. Temperature distribution during resistance butt welding.
de (d2e a2e dZ
d20)
the application of this equation for l == 10 cm for the afore-
Now, to solve the differential equation (2.103) \ye must find a mentioned boundary conditions. The actual temperature
solution that satisfies the following boundary conditions. distribution generally looks more like the dotted curve of Fig. 2.15.
 102
Welding Engineering and Technology Heat Flow in Welding 103
2.7. Heat Flow in Electroslag Welding
enclosed by a chamber from which water is excluded under
In electroslag welding (ESW) the heat source is large and pressure. The welding so done is very similar to that carried out
moves slowly. This implies that the material ahead of the moving in open air conditions except that the pressure varies with the
heat Source is preheated to a much higher degree than in normal water depth. Temperature distribution in the work therefore
arc welding so that the heat flow patterns associated with ESW remains similar to that encountered in normal atmospheric
are not likely to be well described by considering point heat source. welding.
Also, the important role of the slag in this process with its resistive Wet underwater welding is carried out in water without any
heating effect should also be taken into account particularly with chamber around the spot to be welded. The process basically
regard to its effects on the shape and size of the heat Source. remains same as used in normal open air welding but the change
Because the slag volume is quite large in ESW thus the from air to water environment results in higher heat losses and
process.
assumption of a point heat source is not repre3entative of this the arc is constricted. In wet underwater welding the heat losses
from the surface of the workpiece are so high that the temperature
Attempts to simulate. the heat· flow of ESW have been made at a short distance from the outer periphery of the weld pool
by different researchers. One such model based on symmetrical remains unaltered which results in the establishment of very
parabolic model to represent the moving molten-solid interface steep thermal gradients. Thus, the isotherms are confined to a
yielded fairly satisfactory results ; typical isotherm shapes as very narrow zone which makes experimental measurements of
calculated from such a model are shown in Fig. 2.16. temperature, at a point, quite difficult. The present discussion
190 about heat flow in underwater welding is limited to wet
Vw = 35.3 mm/min
underwater welding only.
penetration = 2.5mm
2.8.1. Heat Flow in Wet Underwater Welding
175L Cavity In wet underwater welding the heat flow takes place inside
the workpiece by conduction while heat flow by convection
accounts for most of the heat dissipation from the surface of the
workpiece. The temperature distribution inside the solid body,
away from the heat source, is very well accounted for by Fourier's
3-dimensional, heat flow equation (2.15) and by differential
equation of the quasi-stationary state of welding, that is, by
equation (2.26) ; however the heat transfer at the surface of the
cu workpiece is by convection. At the two major bounding surfaces,
+= since the heat is transferred through the laminar boundary layer
>~ 130 of fluid only by conduction, thus at these surfaces equation (2.15)
reduces to,
aT
115 :I
dQ == - k . . dA . oZ ...
(2.107)
This transfer of heat from the surface of the workpiece could
o 12.5 25.0 37·5 50.0 also be represented by Newton's law of cooling as,
Radial distance, m m dQ == h . dA . T ...(2.108)
Fig. 2.16. Calculated isotherm shapes for electroslag welding. Equating the right hand sides of equations (2.107) and (2.108),
we get,
2.8. Heat Flow in Underwater Welding aT h
The process of underwater welding is broadly divided into two az == - kT ...(2.109)
types viz., Dry Underwater Welding, and Wet Underwater
Fig 2.17 shows that equation (2.109) is valid only for z;::.0,
Welding. In dry underwater welding the spot to be welded is
whereas at z ==g, expressicns (2.107) and (2.108) lead to,
 104
Welding Engineering and ~echnology
n termed as 'pool boiling' in which the whole volume of the water
r
B
£
z= 0
( upon the temperature difference between the workpiece and that
y
B z= 9
z to the very high temperature of the arc the water immediately
Fig. 2.17. Plate with equidistance grid spacing and the direction of outwardly
drawn normal at one of the grid points.
dT h
-=+-T
dz k
or, in general, the bound'ary conditions at the two major bounding
surfaces could be expressed as,
~~±~T=O
The quenching caused by the surrounding water in wet
underwater welding results as, already discu"ssed, in the setting
up of steep temperature gradients in the body of the workpiece.
Therefore, the temperature of the plate drops to ambient
comparatively at a short distance from the weld pool. Hence, it is
logical to assume that the temperature at the periphery of the
workpiece must be the same as that of the surrounding water.
A solution of temperature distribution problem for wet
underwater welding would be one that satisfies simultaneously
equations (2.26) and (2.111). To achieve this the values of a
(thermal diffusivity), k (thermal conductivity), and h (surface heat
transfer coefficient) must be known. Though it is possible to use
some average values for a and k without any serious effects on the
final results but the change in the value of h with temperature is
so large, over the temperature range encountered in welding, that
to assume any single value for it is out of question. To arrive at
any conclusion about the value(s) ofh to be used a thorough insight
into the effects of different situations developed in wet underwater
welding on this parameter is required and the same is done in the
following sections.
2.8.1.1. Surface Heat Transfer Coefficient
In underwater SMAWthe arc is surrounded by a very active
vapour pocket which dissociates 12 to 16 times per second from
Heat Flow in Welding 105
around the arc. This keeps the water in the vicinity of the arc
x
always in an agitated state. Thus, it is not the case of what is
involved boils and thus stirs without agitation due to any external
source.
The value of the surface heat transfer coefficient, h, depends
of the surrounding water as has been expressed by equation
(2.108) and the temperature of the work (steeD may vary from
around 2500°0 to the ambient water temperature. Moreover due
around the vapour pocket boils. Owing to the agitation of water
and convection currents the boiling water moves up, comes in
contact with the bulk and gets condensed. Thus, the heat transfer
...(2 ..110) around the arc and consequently the weld pool is high and a very
complex phenomenon involving heat transfer by conduction,
convection and radiation. The convective heat transfer is further
complicated because it involves simultaneously the phenomenon
of boiling and condensation. Boiling heat transfer is itself quite
..,(2.111) complex because apart from the phase change it involves a large
number of variables such as the geometry of the work, the
viscosity, density, thermal conductivity, expansion coefficient,
specific heat of the fluid, the surface characteristics, surface
tension, latent heat of evaporation, liquid pressure, etc. Moreover,
in underwater welding it is not a case of pool boiling but instead
!..I"
it is a case of what is known as 'Local Boiling', or 'Surface Boiling',
or more comprehensively termed as 'Surface Boiling of Subcooled
Liquid'. Before discussing 'Local Boiling' a note on the mechanism
of 'Pool Boiling' is imperative as that forms the basis to which all
deviations will be referred to.
2.8.1.2. Pool Boiling
Consider a heated plate submerged in a pool of water at
saturation temperature (Tsrd)' As the temperature of the work
(T",) is raised the value of the heat transfer coefficient, h, goes on
increasing. Fig 2.18 represents the type of heat transfer data
obtained as the work temperature is increased above ambient. In
t.his figure heat flux (} ) and surface heat transfer coefficient, h,
are plotted against the excess temperature /:1T, where
/:1 T = Tw - Tsat·
As the temperature Tw is raised, /:1 T is increased, convection
currents cause the liquid to circulate and the steam is produced
 106 Welding Engineering and Technology Heat Flow in Welding 107

IntErfacE results in decrease in the heat flux. The vapour film is unstable
NuclEatE boiling Film boiling
evaporation in regime 4, as under the action of circulating currents it collapses
3

,,
4
2 5
I but reforms rapidly. In regime 5 the vapour film is stable and the
'--- "/' I I Ic...I
I..c I
~ ...c; -E _'!- --•.....VI
"- heat flow is the lowest. For values of ~ T beyond 550°C (regime 6)
the temperature of the work surface is quite high and heat transfer
"
N00E \
..c
-oJ
cro
""
\ \ \ -
occurs predominantly by radiation, thereby increasing the heat
flux.
2.8.1.3. Local or Surface Boiling
The boiling process in a liquid whose bulk temperature is
below the saturation temperature but whose boundary layer is
sufficiently superheated that vapours form next to the heated
surface is usually called 'Local Boiling'. These vapour bubbles
break off and begin to rise through the cooler liquid and get
condensed to the liquid phase again. Thus boiling at the heated
Log 6T,·( -
surface is combined with convection at a distance from it and
Fig. 2.18. Heat flux and heat transfer coefficient vs. temperature condensation of vapour at the interface between the boiling
difference in boiling. water at atmospheric pressure. boundary layer and mass of cold liquid. The intensity of vapo-
risation on the wall depends on the degree of superheat of the
by evaporation at the liquid surface. This is represented as regime liquid, the process of condensation is determined by difference
1 ann is called 'Interface Evaporation'. Here, only the liquid is in between saturation temperature and the bulk temperature, that
contact with the heated surface and the heat transfer is due only
to free convection. is by the degree of subcooling of the liquid.
With further increase in ~ T the energy level of the liquid Subcooling from the saturation temperature to the bulk
temperature is a reference parameter that distinguishes surface
adjacent to the work surf~ce becomes high at a number offavoured boiling from pool boiling. If superheating determines the intensity
spots where vapour bubbles are formed. They rise above the plate of vaporisation, subcooling determines the size of the region that
8m-face but condense before reaching the liquid surface. This is is affected by the disturbing action of vaporisation. The greater
known as regime 2. As the temperature of the work is raised the subcooling of the liquid, the narrower is the region where
further, up to point A in the figure, the bubbles become more boiling takes place. Also, the bubbles increase in number while
numerous. The liquid is so hot as not to allow any condensation their sizes and average life time decreases with decreasing bulk
of the bubbles which rise to the free liquid surface and help rapid temperature at a given heat flux. As a result of increase in the
evaporation. This is known as regime 3. Both the regimes 2 and bubble population, the agitation ofthe liquid caused by the motion
3 fall in the category of 'Nucleate Boiling'. of the bubbles is more intense in a subcooled liquid than in a pool
Beyond the point A, representing the critical heat flux, the of saturated liquid and thus much higher heat flux is atta.ined
number of bubbles formed is so high that they form patches of before any vapour film is formed.
vapour film which form and break regularly. This constitutes To appreciate the effect of subcooling on critical heat flux
r.egime 4 and is often termed as 'Unstable Film Boiling'. Any reference may be made to Kutateladze's equation given below.
further increase in heat input to the work results in the formation
of continuous vapour film over the whole body of the workpiece.
This is termed as regime 5 and is called 'Stable Film Boiling'. The
regimes 4 and 5 are also often simply termed as 'Film Boiling'.
,>

where, (qcr)e
;:::;:
+ + 0065 (7)" c:]
critical heat flux at eo subcooling,
=
...(2.1121
Here the number of bubbles formed is so large that they almost (qcr)o critical heat flux at 0° subcooling,
=
cover the whole of the work surface and provide insulating effect.
This counteracts the beneficial effects of agitation by bubbles and Y = specific weight of water at bulk temperature and
operating pressure,
108 Welding Engineering and Technology
I Heat Flow in Welding 109
l'
specific weight of saturated steam at the
==

Ps pressure of standard atmosphere,

operating pressure, t
==

c specific heat of water,

==
I
, hp = boiling heat transfer coefficient at pressure, P,
8 = degree of subcooling, °C, hps boiling heat transfer coefficient at atmospheric
==

r latent heat ofvapOl'isation.

== pressure, Ps
Putting the values of different variables for the actual 2.8.1.5. The Effect of Position of Heated Surface
condition of operation, with bulk water temperature of 30°C it was
found that,
(qcr)e ~ 70"C
----.

(qcr)e,~
---.-.-

O"C
==
•
2.565 to aThe
great
heatextent,
transferupon
coefficient,
the conditions
h, and the
in heat
whichflux
the(~~nerated
vapours separate from the heated surface. These conditions are
depend, I
which means that the critical heat flux increased 2.565 times for most favourable in the case of horizontal heat.ed surfaces, the
70°(; of subcooling. heated side of the surface facing upwards. The aforementioned
equations [(2.112) to (2.114)] hold good for such conditions. If the
Apart from subcooling other major factors that affect heat flux, heated side of the work faces downwards, the conditions in which
and consequently the surface heat transfer coefficient are the vapours separate from the surface deteriorate sharply and the
arnbient pressure, position of heated surface and the motion of the peak heat flux diminishes by as much as 40%. This is because the
fluid. The effects of these factors are discussed as follows. motion of the fluid is only in a thin layer undE;rneath the work,
2.8.1.4. The effect of Pressure the rest of the fluid below that layer remains stationary.
The effect of pressure on the heat transfer coefficient in well Fig. 2.19 represents the commonly accepted nature of
developed nucleate boiling is more or less the same for all liquids. convection currents above and below a horizontally placed heated
According to test data reported by Mikheyev the peak heat flux
(qpeak) first sharply increases, reaches a certain maximum ",:,ith
rising pressure, then drops to zero at critical pressure. If a graph I
i
••
. db
IS

for
rawn etweenp-Pdq

(Per
9.P.'!.'!:..'s..~
qpeak.l
peak l' th
cr an .------'-
qpeak.l . en

==
the curve IS
. a t 1.t s maximum
-

3.2 and J}_ = 0.35. If this data is applied to water

Per
225 ata.) it will be seen that the peak occurs at. a pressure
==

equal to 80 atmosphere. Further, it has been recommended, from

~~
(0 )
lv~l\lJl (b)

the charts plotted for pressures ranging from P 0.2 to P 100 == ==

ata., that the coefficient of heat transfer of water in nucleate ~~/////@'//(/////~V

boiling may be calculated from the following relationship, !c l
h 39 (Do 7j2.3,3 kcals/m2-hr-oC
== ...(2.113) l"ig. 2.19, Free flow of fluid near heated horizontal flat plates: (a and b)
The relationship suggested by Jackob and Hawkins for finding heated surface upwards, (c) heated surface downwards.
the effect of pressure on heat transfer coefficient is given by the flat plate.
equation, 2.8.1.6. The Effect of Motion of Fluid
Apart. from the aforementioned factors, heat transfer
hp = hps (p
:p:))0.4 .;.(2.114) coefficient is considerably influenced by the rate of forced
where, P = pressure of water at the point under consideration, circulation of the fluid. If there be no forced circulation of the liquid
the steam bubbles which are generated on the heated surface grow
to a specific size before detachment thus the heat transfer
110 Welding Engineering
, . and Technology
coefficient isgovemed by the intensity ofvaporisat ion. If the liquid
is made to circulate then the steam bubbles are detached before
they attain critical size. As the fluid density is increased further
the effect of intensity of vaporisation is gradually reduced till it
reaches a value that of free convection in a single phase liquid.
Thus it can be said that at low circulation velocity (w) the intensity
of vaporisation (q,) is predominant and at higher circulation
velocity the effect. of w is predominant or in other words the
coefficient of heat transfer can be expressed as,
h = f (w, qu) ... (2.115)
In the case of underwater SMAW, when an arc is struck
between the electrode and the workpiece a bubble or a vapour
pocket is formed with the arc at its centre. Next to the arc would
probably be a mixture of incandescent gases emitted by the' arc
and superheated steam around this an envelope of saturated
steam in contact with ambient water. The vapour bubble
fluctuates, as stated earlier, 12 to 16 times per second releasing
about 200 cm3 of combustion gases and steam per second. This
leads to enormous disturbance ofliquid in a sufficiently big volume
around the arc.
From the factors discussed above and also because of the wide
variations in the results of many researchers in the field of boiling
heat transfer it is rather difficult to arrive at any standard relation
for predicting boiling heat transfer coefficient for the conditions
under consideration. However, from the above considerations the
different values of heat transfer coefficient, for different ranges of
telnperatures, can be based on the careful study and weightage
given to different factors discussed as follows.
The peak value of 'h'· for pool 'boiling given by Mikheyev is
5 x 104 Kcals/m2hrC,C and that by Kutateladze is about 2.55 x 104
Kcals/m2-hr-oC. Kutateladze has also recommended equation
(2.112) for finding the effect of subcooling on the value of h. Using
this equation with 70°C of subcooling (i.e. for a room temperature
1)( 30°C) h increases, as already stated, by a factor of 2.565 thus
giving peak value of h as 6.3 x 104 Kcals/m2-hr-oC. Taking into
consideration these two values and the agitational effect of the
detaching bubbles around the arc it is considered best to use
Mikheyev's expression ~nd the subcooling factor found by equation
(2.112) to get peak value of h of the order of 12.8 x 104 Kcals/m2
-hr-oC.
The different values of'h' for work temperature up to 121.5°C
could be decided on the basis of results given by different
Heat Flow in Welding 111
researchers. However, for work temperature higher than 121.5°C,
h can be taken as equivalent to peak value mentioned above
because it is not a case of pool boiling and there being considerable
disturbance around the arc the application of the formula of
continuous film to the work being welded appears to be
impractical. Moreover, the use of peak value of h for higher
temperatures also takes care of heat transfer by radiation above
550°C.
Thus, from the above mentioned considerations the values of
surface heat transfer coefficient, h, for different ranges of
temperature, encountered in wet underwater welding may be
taken as follows.
(i) Heat transfer coefficient can be taken as directly pro-
portional to the temperature difference between the work and the
bulk of water for and up to TU) = Tsat = 100°C.
(ii) For interface evaporation, i.e., for the work temperature
between 100.1 and 104.7°C,
h = 896 (~ T)1/3 x subcoolingfactor" Kcals/m2-hr-oC
(iii) For nucleate boiling i.e. for work temperature between
104.8 and 121.5°C,
h = 39 (~ T)2.33 x subcooling factor, Kcals/m2-hr-oC
• (iu) For work temperature above 121.5°C,
r
h = 39 (21.5)2.33x subcoolingfactor, Kcals/m2-hr-oC
The values of heat transfer coefficient on the bottom side of
the work may be taken as 60% that of the value of h on the upper
side of the work.
Having developed the model for determining the value of
surface heat transfer coefficient, it is possible to determine the
temperature distribution in wet underwater welding provided it
satisfies certain boundary conditions.
2.8.1.7. Boundary Conditions
It is required to satisfy the condition that the heat entered the
work at a uniform rate through the arc and that the same is also
conducted away at a uniform rate par.t1y into the body of the
workpiece and partly dissipated to the surroundings. The size of
the arc being finite the area over which it supplies the heat is also
finite. That means arc supplies heat into the heat input zone of
the work from where it flows QYconduction into the body of the
* Subcooling factor = 2.565 for undercooling of 70°C i,e. for bulk temperature
of 30°0.
112 Welding Engineering and Technology 113
Heat Flow in Welding

W"ork;the molten weld pool zone can be taken as the heat input
zone. The shape of the weld pool can be determined practically by
studying the crater shapes obtained by sudden interruption of J
welding process.
Since the work and with that the heat input zone is conti- I eo
n uously moving thus heat is transferred to the work not only by I ..=
e .
conduction from the lower and side bounding surfaces of the e
to .S
:g

molten metal zone but also by the movement of the work. Thus,
the heat balance at the source may be expressed by the following
I ~ ~::a
~ I :2 '" c::

eo. 1/ 8
equation, ., •..
...(2.116) ~-
- 0
I ~
o.

ii~
:2
a.
E-
o.-
-a.;
..lid c::

Q=- k Jf ~~ ds + cp fJ TVn ds
s s Ole
co.
o.-e
.-
~'': I
. .
o.
a.
E
~
~s .~ 8.
.s 0
e-g
e .,
where ~~ is the temperature gradient along the outward drawn ~! I
t- •.•
'o~

!)
o .
normal to the surface element ds, and Vn is the component of
welding velocity in that direction. ; / II
"0

.~
~ ~
<J
"0
Q.)

c::
::s
e ., •..
g o ~r.S
In equation (2.116), the first term i.e. - k Sf ~~ ds, accounts :: 0...c:::
"' .•..
0

/'I
s Q.)

.~ ..c
for the heat going out of the heat input zone by conduction and ~
•.. I Q.)
.~ c::

the second term i.e. cp ff TVnds represents the heat carried away
..c::;.::

......
Q.) Q.)

s ,P" 3d
by that portion of the plate which directly passes through the arc
zone. The sum of these two terms is equated to the heat input per
/' .,
•••

8."0
eQi
Q.)
<J

unit time i.e. Q, which is equal to T1VI, where TJ is the percentage

- .. Q.)

Eo<
~

Q.)

'..c::
of heat going into the heat input zone - the total heat generated 0 .•.•
<"!
C'l
being the product of arc voltage (V) and the welding current (1). o!l
("
.-....-0-.-0......--.0....._0-._
~~ __
~ h __ ~~ __
~
The heat taken away by water from t.he electrode and by way
of steam formation, etc. is generally considered to be about 15% I.I'l
that of the total heat input. Thus, for calculating the temperat~re I I I I ~ C>
.•...

histories in wet underwater welding the heat input into the molten 08 ~ ~ ~ ~ ~ ~
~C> C> C> C> C> C> C> C>
,.....
metal zone is taken as 50% of the observed value of the power
input into the arc. The correspondirig value for open air conditions .•....•... _ ).' aJn~oJadwal
are generally taken as 65%.
Thus, the final solution of the heat transfer model must not temperature histories practically obtained for welding underwater
only satisfy the equation representing the heat conduction in the and the normal open air conditions. These graphs represent the
quasi-stationary state, i.e. equation (2.26) but also the boundary quasi-stationary temperature distributions along a line parallel
conditions expressed by equations (2.111) and (2.116). This can be and 6 mm away from the weld centreline; x = 0 being the centre
done by using the numerical tools like finite difference and finite of the arc.
element metbods. The results obtained by solving the pr0blem by Fig. 2.21 shows the quasi-stationary temperature distribution,
finite difference is shown in Fig. 2.20 which also shows the for the bottom side of the plate type work, along the transverse
 114
Welding Engineering and Technology Heat Flow in Welding 115
Sample No.
1100 SE

~ Experimental 1000
->c-X- Calculated

.~
I 7001- "\~0 400
Q.
•• ~
500
E
,
\
t 800
6001- \\ 300
\ .\..J

~ 600
100
2000 .-C
::J

"-
= 4.0 mm QI
c..
= 4·5 mm
I-~ 400

o 80
10 20 40 50 60 70
o
Time(see)-
(0 I

o
'" E 3
.2 E4.5
o .-c: 755
~.~ 8
a:;
U; I
9.5
15
Ci> 16
20 30
transvers~ (V ) di rection ( mm) _ (b I
Fig. 2.21. Quasi-stationary temperature distribution at the bottom Fig. 2.22. (a) Experimental quasi-static temperature distribution at the back
of the plate in underwater welding-along-y-axis. side of the plate at different points along a transverse section,
(b) the corresponding isotherms in °C.
section at 4 mm from the centre of the are, and perpendicular to 2.9. Metallurgical Effects Of Heat Flow In Welding
the weld centreline.
Using equations given in the earlier sections it is possible to
shows experimental quasi-st.ationary t.emperature
Fig. 2.22(a) determine temperature at any given point during quasi- stationary
distribution on the backside of the work (plat.e) along the weld state of welding and from such a data it is possible to draw thermal.
centreline and along dif~erent parallel lines at 3, 4, 5, 6.5, 7.5, 8, histories for any point of interest. If sufficient number of such
9.5 and 14 mm away from the weld centreline. In Fig. 2.22(b) are thermal histories are known for different points along a transverse
shown the isotherms for the temperature distribution of Fig. section with respect to the weld centre line then such thermal
2.22(a). The isotherms plotted are for temperature range of 100 to histories can be utilised to draw isotherms for different temperatures
IOOOoe at an interval of lOOoe.
keeping the weld pool as the innermost isotherm representing the
 116 Heat Flow in Welding 117
Welding Engineering and Technology

solidus temperature of the material being welded as shown for

slow and fast welding in Fig. 2.23. I~ ~mm -~~-t-50mm~

Direction of welding

W~ld pool
isotherm
TCI
TC2
TC3
TC4
--- TC 5
TC6
(a I E ac kside of plate with positions of thermocouples.

'"
::J a.
'-
e'"
>-
~'"'- Tl
T4
T3
Ts
I
T6
T2
Welding dir~ction -
(a I Slow speed weld ing

Welding
isotherm

Direction of weldmg

---+ tI
; !
Time -

(b) Temperature histofle.s

Welding direction -
6
(b I High speed welding
5
4
Fig. 2.23. Isotherms in quasi-stationary state welding:
3
(a) slow welding, (b) fast welding. 2
~.
__t
From these isotherms it is possible to determine the cooling
rate in any desired direction such as A, B, C, D, E, etc. If such a
cooling rate is supenmposed on time-temperature-transformation
curves or on continuous cooling transformation curves of the
material under consideration, then it is possible to predict the K
micro-structure of the heat affected zone along that direction; from (() Some of the isotherms
which it may be possible to determine the mechanical strength of
the weldment. Thus, it is possible to predict the probable service
behaviour of the welded fabrication .. Fig. 2.24. Experimental determination of cooling rates in welding,
(a)Backside of steel plate with positions of thermocouples·marked on it,
Thus, the first step to predict the metallurgical effects of heat (b) Thermal histories for point along transverse section to welding direction,
flow in welding is to determine the cooling rate for a given set of (e) Different isotherms obtained from thermal histories of (b).
 118 Welding Engineering and Technology
welding conditions. The experimental method of doing so is
described in the following section.
2.9.1. Experimental Determination of Cooling Rates in
Welding
Take a 6 mm thick steel plate of sufficient width and length
(say 200 mm x 300 mm) so that quasi-stationary state will be
established after welding has proceeded through a length of 50
mm. Mark it on the bottom side as shown in Fig. 2.24(a). Drill 3-4
mm deep holes with 1 mm diameter drill bit at points 1,2, 3, ....,
6. Imbed the hot junctions of Alumel-chromel thermocouples in
these holes which are filled up with high temperatu~e brazing
material-using oxy-acetylene brazing torch. The other ends of
these thermocouples are connected to temperature recorders for
recording thermal histories of theRe points during welding on the
top side.
When welding is carried out with the required heat input and
at the desired welding speed then the recorded thermal histories
will resemble the recorc.s shown in Fig. 2.24(b). From these
temperature histories isotherms for temperature Tl> T2' T3' ...,
T6, etc. can be drawn as shown in Fig. 2.24(c). The procedure for
doing so is shown for isotherm T6 and the same may be followed
for all other isotherm; employ interpolation, where required. From
these isotherms cooling rates can be determined in any desired
direction like K, L, M, N, etc. For steels, cooling times from 8000C
to 500°C is highly significant since this is the critical temperature
range in which phase transformations and, therefore, the
mi.::rostructure and properties of the heat affected zone (HAZ) are
characterised. Cooling rate is expressed by the parameter of
cooling time for example tS/5 represents the cooling time between
800 and 500°C, tT max /100 stands for the cooling time between the
maximum temperature of the thermal cycle and 100°C. On the
other hand V300 stands for cooling rate at 300°C.
Fig. 2.25 shows the temperature histories for SMAW,SAWand
ESW with cooling times mar-ked for the range of 800 to 500°C. It
is evident that the cooling rate in SMAWis much higher than the
cooling rates of SAW and ESW. For each steel there is a critical
cooling rate which decides the final hardness of weldment
particularly in its heat affected zone.
2.9.1.1. Critical Cooling Rate
It, is the fastest rate at which steel can be cooled without the
appearance of martensite, or stated conversely, it is the slowest
119
Heat Flow in Welding
1400
t 1200
.C1000
o
-30 -20 -10 o 10 20 30 40 SO 60 70 80 90
TimlZ aftlZr max. tlZmpcraturlZ (s)-
Fig. 2.25. Typical shapes of thermal histories of the underbead zone
with different welding processes. (After Hrivnak).
rate at which the steel can be cooled that will still produce
martensite. Also, it can be said that any rate of cooling faster than
critical cooling rate produces a structure containing pearlite.
Steels of,low harden ability have a high critical cooling rate
and vice versa. Increase in carbon content reduces the cooling rate
i.e. raises the harden ability. The majority of alloying elements,
with the exception of cobalt, have a similar effect. The so-called
air-hardening steels have such a low critical cooling rate that even
slow cooling from the austenitic range produces martensite.
Water-hardening steels on the other hand, because of their high
critical cooling rates, must be rapidly cooled to produce desired
hardening.
Apart from the standard critical cooling rate for a steel there
are other critical rates related to the formation of different
transformation products, for example, p, f, and z cooling rates
represent the cooling rates for the formation of pearlite, ferrite,
and bainite respectively.
2.9.1.2. Transformation Products
Three major constituents formed on cooling steel from its
austenitic state are pearlite, bainite and martensite. Both pearlite
and bainite are formed by processes depending on the diffusion
rate of various alloying elements. Thus, both these structures
require a certain period of time for formation, a period which
varies with temperature. In other words the transformation rate
 120 121
!
Welding Engineecing and Technology Heat Flow in Welding

of austenite depends on the temperature of transformation. Both amount of austenite transformed is 10%, and when it is 99% or
processes are preceded by a certain incubation period which must ,
100% complete; curves for 30,50, 70%, etc. transformation are also
elapse before any transformation takes place. The martensitic often included.
transformation, on the other hand, is practically independent of The complete transformation of austenite requires long test
time, occurring instantaneously when a certain temperature is periods within certain temperature ranges and for this reason the
reached.
time axis of TTT diagrams is usually plotted on a logarithmic
Thus, each composition of steel has its own characteristic way scale. A typical simple TTT diagram for 0.35% plain carbon steel
of transforming when it is cooled at a given rate from the is shown in Fig. 2.26 while Fig. 2.27 shows the TTT diagram for
austenitic state, and there are only two ways to summarise the
effects of differing conditions viz., isothermal transformation tests 800
and continuous cooling tests. Curves derived from the former are
called Time-Temperature-Transformation (TTT) diagrams and
from the latter, Continuous-Cooling-Transformation (CCT) Ide
diagrams. Brief description of these two types of diagrams follows.
2.10. Time-Temperature-Transformation Diagrams (TTT
Curves) .u
A TTT diagram shows in graphical form the time required, at "
•..
::>
various temperatures, for steel in the austenitic state to transform '6400
to ferrite, pearlite, bainite, and/or. martensite. Which of these "•..
Cl.
transformation products are actually formed depends on the E
:t.
transformation temperature. Thus, for example, the austenite in
a medium carbon steel breaks down to ferrite and pearlite in the 200
temperature range of 700 to 500De, whereas if the cooling rate is Mf
somewhat higher, pearlite formation can be partly or wholly
suppressed, and bainite will form at a lower temperature. At
extreme cooling rates, even bainite formation can be suppressed, o
and martensite will form at a still lower temperature. 1 10 102 103
TimlZ. sIZes
TTT diagrams are constructed from the data obtained on small
specimens of the I steel under investigation, which are heated to Fig. 2.26. Approximate form of TTT diagram for a 0.35% plain carbon steel.
slightly above Aa temperature', and cooled to the temperature at an eutectoid commercial steel AISI 1080 which contains 0.79%
which transformation rate is to be studied. The specimen can, for carbon and 0.76°10 manganese. Note that while for 0.35% plain
example, be quenched in a lead or salt bath, and since it is very carbon steel the nose of the starting curve touches the Y-axis
small it very quickly reaches the bath temperature. The specimen indicating no possibility of getting 100% martensitic trans-
is held at this temperature for a certain accurately determined formation even if the cooling rate is extremely high; while for the
time and finally water-quenched. The testing procedure is similar eutectoid steel the nose is shifted towards right indicating that
to the heat treatment method of austempering. Subsequent 100% martensitic transformation is a possibility.
microscopic examination shows the percentage of austenite which
has transformed within the test period and the structure of All the alloying elements used in alloy steel prod~ction
transformation product(s) typical of the test temperature. By influence the eutectoiG. condition of steel, each one lowering the
making a sufficient number of these tests TTT diagrams can be eutectoid composition, but the influence varies from element to
constructed. This usually contains curves showing when the element, for example, some like Cr and Mo raise the eutectoid
temperature while others like Mn and Ni lower it. The relative
'" A~ temperature is ~he upper recryst.allisation temperature for low and effects of different amounts of some additions are illustrated in
medium carbon steels. Refer to Fig. 3.1 or 3.27.
 122 Welding Engineering and T~chnology I Heat Flow in Welding 123
~
r

800
I
.Eutecto~~ temper~~ ._
700 Austenite !
!
;
600
...
.
•U,SOO
'-•••
:;,
e 400
•••
CI.
E Ti Mo
t!!! '300L Austenite
o
Ms Fe 5 10 15
Alloying element, ./.
--.,1VJSO----------- - ~-
"-
100
- -M9o- - - - - - - - - - =~~-=-~.-----:.=- \0)
Ti
M.:utensite

OL.. I I I
0·1 1 10 100 103 104 105
Time il'1 seconds .u.
~1000
Fig. 2.27, Complete TTT diagram for eutectoid steel AISI 1080 •..
•••
containing 0.79%C and 0.76% Mn, Q.

Fig. 2.28. Also, these alloying elements may either retard or :!E
'U
accelerate the decomposition of austenite. If it retards the '~
u
decomposition then the critical cooling rate is reduced and it :;
•••

results in shifting the TTT curves to right or a 'bay' on the nose UJ

of the diagram appears as is shown in Fig. 2.29. However, if the

. decomposition rate of austenite is accelerated the TTT diagram
tends to bulge more towards left, indicating that the steel is more
difficult to harden; this effect is rare and is caused only by cobalt
in certain circumstances.
Although TTT diagram is useful for comparing steels, it does
not predict accurately the results of welding conditions or heat-
treatment because they involve continuous cooling. The CCT
di agrams can therefore be used more effectively for these
purposes.
2.11. Continuous Cooling Transformation Diabrrams
(CCT Curves)
A CCT diagram is a record of the transformation behaviour of
tnat steel under continuous cooling conditions which can be
o S 10 IS
Alloying element, '/.
\ b)
Fig, 2.28. The relative effects of alloying elements on the eutectoid relationships
of steel: (a) effect on eutectoid composition,(b) effect on eutectoid temperature.
',' correlated fairly closely with the kind of continuous cooling
occurring in the vicinity of a weld. From such a diagram it is
possible to determine whether or not martensite or brittle
structure is likely to form under given welding conditions. The
farther to the right and lower the curves on the diagram the more
hardenable the steel and more difficult the welding.
 124

800
Welding Engineering
Austrnitr
and Technology
I
f
,
600
,t.J
••
•..
:l
Ci
'-
~400
E
~

200

o I I I I I I I
0·1 1 101 102 103 104 105
Time in src
l<~ig.2.29. Effect of alloying element.s like Ni and Cr in cnanging the shape of
TIT diagram by developing a bay on the nose.
Tests used for constructing CCT diagrams utilise differing
sizes of round bar to derive cooling conditions equivalent to oil
quenching conditions on round bar. From these tests the effects
on structure at. the middle, surface, and half radius respectively,
of the bar are studied and recorded on a graph. Bar diameter is
represented on the x-axis and temperat.ure on the Y-axis as shown
in Fig. 2..30. The general form of ceT diagram is similar to that
of TTT diagram as shown by comparison of two sets of curves in
Fig. 2.30.
The difference between TTT diagrams and CCT diagrams are
<:>
<:>
CIO
o
<:>

..;to
<:>

perhaps most easily understood by comparing these two forms for
'8 steel of eutectoid composition as shown in Fig. 2..31.The cooling
curves, corresponding to different rates of continuous cooling,
superimposed on TTT and CCT diagrams are also shown in Fig.
2.31. In each case the cooling curves start above the eutectoid
temperature and fall in temperature with increasing times.
Considering the curve marked 1,' it crosses the line repre-
senting the beginning of the pearlite transformation, at the point
marked '(1', at the end of approximately 6 second. The significance
of point 'a' is that it represents the time required to nucleate
:l.' <tJnlondwaJ.
pearlite isothermally at 650°C. A specimen cooled at the rate
represented by line 1, however, reached the 650°C isothermal at
the end of 6 second and was at temperatures above 650°C for the
entire 6 second interval. Because t.he time required to start the
pearlite transformation is longer at temperatures above 650°C
than it is at 650°C, the continuously cooled specimen is not ready
to form pearlite at the end of 6 second. In other words, more time
is needed before transformation can begin under continuous
cooling condition compared with isothermal transformation. Since
126 Welding Engineering and Technology Heat Flow in Welding 127

BOO allow any appreciable amount of bainite to form. In the second

Eutectoid te m perot u re
case it is also to be kept in mind that the rate at which bainite
----------
----------
forms rapidly decreases with falling temperatures. It is generally
assumed, as a first approximation, in drawing a CCT diagram for
an eutectoid steel, that the transformation, along a path such as
600
curve 2, stops in the region where the bainite and pearlite
transformations overlap on the isothermal diagram. Thus, the
microstructure corresponding to path 2 should consist of a mixture
of pearlite and martensite, with possibly a small amount of bainite
.w 500 which may be ignored. The martensite forms from the austenite
~
...
::J
which did not transform to pearlite at higher temperatures.
C
... 400 Different transformation products obtained by cooling an
~
a. eutectoid steel under different continuous cooling rates are
E
represented in Fig. 2.32. The curve marked 'full anneal' represents
~ 300 800
Eutectoidtemperat IJre

200

"'190
100

o
500
0·1 10 100 103 10" 105 5xl05
Time in seconds .u.
~
...
Fig. 2.31. Schematic representation of relationships between CCT and ::J 400
TTT curves for an eutectoid steel. (After Reed-Hilll . "2
•••
a.
in continuous cooling an increase in time is associated with a drop E
~
in temperature, the point at which transformation actually starts
is 'b' which lies to the right and below point 'a'. In the same way
it can be shown that the finish ofthe pearlite transformation, point 200
a, is depressed downward and to the right of point c, the point
where the continuous cooling curve (Curve 1) crosses the line
representing the finish of isothermal transformation. 100
Fig. 2.31 also shows that the bainite reaction does not appear
on the continuous cooling diagram because the pearlite reaction o
lines extend over and beyond the bainite transformation lines. 0·' 1 10 100 103 104 105 5xl05
Thus, on slow or moderate rate of cooling (represented by cooling Time in seconds
'rate curve 1), austenite in the specimens is converted completely Fig. 2.32. Transformation products produced by different continuous
to pearlite before the cooling curve reaches the bainite trans- rates in an eutectoid steel. (After Reed-Hill)
formation range. Because the austenite has already been very slow cooling and is obtained by cooling, suitably austenitized
completely transformed to pearlite, thus no bainite can form. specimen,.in a furnace which has its power supply switched off.
Alternatively, as shown by cooling rate curve 2, the specimen is in Under this rate of cooling the specimen is brought to room
the bainite-transformation region for too short a period of time to
 128 Heat Flow in Welding 129
Welding Engineering and Technology r
temperature in about a day. Here the transformation of the The dashed continuous cooling rate curve between the oil-
austenite takes place at a temperature close to the eutectoid quench curve and water-quench curve represents the critical
temperature and thus the final structure is coarse pearlite and cooling rate curve; any rate of cooling faster than this produces a
similar to that predicted for an isothermal transformation. martensitic structure while any slower rate than that (dashed
line) produces a structure containing some pearlite.
The second' curve, marked normalizing represents a heat
treatment in which specimens are cooled at an intermediate rate \Vhile Figs. 2.31 and 2.32 show schematic representation of
by pulling them out of the austenitizing furnace and allowing them CCT curves for an eutectoid steel Fig. 2.33 shows the actual CCT
to cool in air. In this case, cooling is accomplished in a matter of curves of two plain C-Mn steels with 0.19% C (dotted lines) and
minutes and the specimen transforms in the range of tempera- 0.28% C (full lines). This figure (Fig. 2.33) shows the effect of
tures between 550°C and 600°C. The structure obtained under this relatively small increase in carbon content on the position of the
rate of continuous cooling is again pearlite but much finer in martensite and bainite fields.
texture than obtained in the full annealing treatment.
800 800
The continuous cooling curve marked as 'oil quench' represents I\A3= 796
a still faster rate of cooling, such as might be obtained when a 700
red-hot specimen is quenched by immersion in an oil bath. Cooling Ferrite
at this rate produces a microstructure which is a mixture of 600 ~O
pearlite and martensite.
Finally, the continuous cooling rate curve farthest to the left .~ 500
",1~t,8 SOO·U.
••• •••
and marked 'water-quench' represents a rate of cooling so rapid •... ~ I ...
::J

that no pearlite is able to furm and the structure is entirely E 4.001-Ms 4000...
e I
martensitic .
, •••
I a.
•••

,, -.... a. 300
E 300 E
," 900 ~ ~
600 I-
'" ,4 __Bainite '"
I
·C 500V 200 200
700F-Ac,
600"Aq Austenite
Start 100 I
--- FlI:r rite
Pearlite
o
1 2
I •
ICz
4
I
Ie;"
6 610 20 406080100200 4006 81000
I
4000
Finish Time. seconds
Fig. 2.34. CCT diagram for a C-Mn steel with different
continuous cooling rates marked on it.
Fig. 2.34 shows the CCT curves for a C-Mnsteel with different
critical cooling rates like Z, F, P and E marked on it; the corres-
300 ponding CCT curves for AISI 4340 (C = 0.4%,Mn = 0.70%, Sand
P = 0.040% (each), Si = 0.30%, Ni = 1.80%, Cr = 0.80% and Mo =
200 0.25%) is shown in Fig. 2.35. It is evident from Fig. 2.35. that
cooling rate slower than E produces microstructure consisting of
100 only ferrite and pearlite while cooling rates faster than that
represented by E produce a microstructure consisting of marten-
o
1000 500 200 100 50 20 10 5 2
site, ferrite, pearlite and bainite.
Cooling rate at 700'C ,'C per min The detailed CCT diagram for AISI 1018 steel containing.
Fig. 2.33. CCT curves of plain C-Mn steel with 0.19~7c,C (dotted) 0.18% C, 0.20% Si and 0.45% Mn is shown in Fig. 2.36, which also
and 0.28% C (fulliinesl. (After Easterling)
 Welding Engineering and Technology Heat Flow in Welding 131
130
shows the lines for 10%, 50%, and 90% transformation of
'lustenite.

800 A3787·C

Al 697·C

.w
.u QI
•...
ttJ' 500
-
:::J
•.. .•...
::I o
Cl
~ 400 ~ 400 I M5
a. I-QI I M
E
~ 300

100
1 10 100 1000
Time I see
Fig. 2.37. CCT curves for T-l Q and T steel. The hatched area represents
the region of optimum cooling rates. (After Sindo Kou).
o, . 10" .
104 105 A3
5 10 100 103 800
~e~
Time. see c..
.•...

Fig. 2.35. CCT diagram for .AlSI 4340 steel showing transformation products
obtained with different continuous cooling rates. (After Reed-Him.
.~
100
~ 400
QI 300
:::J

700200
600
500
Austenite
C S1 Mn P S Cr Mo Ni AI Nb V
0.11 0.20 0.45 0.02 0.02 -
900 START'
10·'.
50"1.

Lower transformation product

---------
·c
Martensite

o
1 2 4 10 40100 200 400 1000 4000
Time,sec
1090 590 2qa 1qa 5P 29 19 ~ ~ \ Fig. 2.38. A weld CCT diagram for a medium strength C-Mn'steel, using an
COOLING RATE AT Baa ·c I ·c PER MIN austenizing temperature of l400°C. (After Easterling).
Fig. 2.36. CCT diagram for AlSI lOl8steel.'
132 Welding Engineering and Technology

Lastly Fig. 2.37 shows CCT curves for T-1 Q&T steel (C :;:
0.15%, Mn :;:0.80%, Si :;:0.25%, Cr :;:0.5%, Ni :;:0.85%, Mo :;:0.5
%, V :;: 0.05%, Cu :;: 0.30%, and B :;: 0.004%) having a tensile
strength of about 900 N/mm2 and yield strength of 700 N/mm2•
Cooling rates marked as p, f, and z represent the critical cooling
rates for the formation of pearlite (P), ferrite (F), and bainite (B)
respectively. The hatched area represents the region of optimum
THREE
cooling rates.
Although a CCT diagram for a given material gives fairly Basic Metallurgy
accurate prediction about its microstructure but the micro-
structure in the HAZ of a weldment is most accurately obtained of Fusion Welds
from Weld CCT diagrams which are based on much higher
austenizing temperature, usually 1350-1400°C, that correspond 3.1. Introduction
to graili growth zone. Such diagrams are plotted using a weld Although a weldment formed by fusion welding results in the
simulator to produce the appropriate thermal cycle. An example
of such a diagram is shown in Fig. 2.38. However, weld CCT formation of monolithic structure but such a joint varies in
diagrams are quite expensive to produce and are, therefore, metallurgical structure from point to point with consequential
available for much less number of steels and other industrially variation in mechanical properties. This is because welding results
important materials. in the development of a temperature gradient which varies from
the highest temperature encountered in the centre of the weld pool
The best guide to the weldability of a particular alloy steel is to the ambient temperature along the transverse direction to the
thus its weld CCT diagram but if a weld CCT diagram is not weld axis. The extent of the zone so affected depends upon mainly
available a CCT diagram will suffice but even if that is not' .u
available a TTT diagram will give a good guide though not so ci:
E
N
accurate as CCT diagrams. In case TTT diagram is used, the ~ ~ ~
welding becomes more difficult with the shifting of curve to right - -":- SoJi.sii!iia_-~~IQ:':. "_:.: :-..: :..:'~~~ :-
, :-_L!~~i~_-_::::: 3:;. I
~.-_=-j:t
and/or lowering of any nose on the diagram (See Fig. 2.30), since
these features indicate that the transformation is likely to' be ••
... 'I· 1400
delayed under continuous cooling conditions, to a lower tempera- ::J ~'I
:: '= s~iid-:liq-uid
grain t~;iti~n-;o~;--. - -
growth zone
ture at which the metal will be relatively brittle. Most severe weld e•• "
:~ - --:~:j:-+~
- - -1T
0.
" 1200
cracking tends to occur when transformation takes place below
300°C.
E
••
•....
"

:',. -__---- -- -- -~ 1 ~
"'" "
~crystallized zone 1000 : ::t:
Cl
No doubt it is not possible to assess the basic weldability for ••
Q.
:'
,.
a given welding situation from either CCT or TTT diagrams, ,
"
____z~!:,! I
without knowledge of the equivalent weld cooling condition, "
"

however it is possible to compare the relative weldability of two "

different alloys for similar welding purposes by comparing the
relative posItion and sizes of their respective curves.
:. ~,
,
,
unaffected base
le~~~~d_material
'
~~"..e_@g _~'cr
800 -,'
--pQ~tTcilly-tr<i'nsfo-r;;';d- 400
,
+FeJC ' 1
~
_

: CD

200 I- :
0,15 1·0
Fe wt .,. C

Fig. 3.1. Different zones of a steel weldment as represented

on an Iron Carbon equilibrium diagram.
 134 Welding Engineering and Technology
the heat input per unit time, the welding velocity and the physical
properties like the melting point, thermal diffusivity, etc. of the
work material.
Basically a weldment can be divided into three distinct zones
uiz., the weld metal (WM) zone forming the weld bead, the heat
affected zone (HAZ), and the unaffected base metal (BM) zone. For
low carbon steels, which is the most welded material in the world,
the three zones of a weldment can be indicated on an iron-carbon
equili- brium diagram as shown in Fig. 3.1.
In between the WM zone and HAZ lies what is known as the
Fusion zone which is the volume of the parent metal actually
melted, as shown in Fig 3.2, to form a part of the weld metal zone
or the weld bead.
Columnar grains growing
from portly m~lt~d
parent grains( shaded)
Boundary of melting
Fig. 3.2. Fusion zone in a weld.
While the weld metal zone constitutes the weld bead and is a
cast structure, the HAZ is, in a way, the heat treated portion of
the weldment, while the unaffected base metal is the original work
material plus a small zone which has been heated to about 650°C
and hence has undergone a slight change in its grain size and thus
mechanical properties. Depending upon the material composition,
the welding speed, and the amount of heat input, different
microstructures may be expected from the different zones of a
weldment formed by fusion welding. Brief description of the effects
of heating and cooling cycles on the microstructures formed in the
weld metal zone and HAZ and their consequential effects on the
mechanical properties and the service behaviour of a welded joint
are given in the following sections.
Basic Metallurgy of Fusion Welds 135
3.2. Weld Metal Zone
The weld metal zone is formed by the solidification of the weld
pool which itself is formed by the melting of a part of the parent
material plus the additional material that is contributed by the
melting of the electrode, if used.
Although a weld pool resembles, to an extent, a metal mould
casting but there exists a great difference between the two cases.
While in a metal mould the molten metal is poured into a cold or
a relatively cold mould cavity and thus there is usually a great
difference between the molten metal temperature and the
temperature of the mould wall, in the case of welding, however,
there is a gradual change in temperature from the centre of the
weld pool to a short distance away from it. Also, for the
solidification of a casting the molten metal must be undercooled
to provid'e enough free energy to cause nucleation of solid grains
which grow into the melt. The growth of individual grains starts
only after a critical radius is achieved which is dependent on the
degree of undercooling. On the other hand, solidification of the
molten metal in the weld pool starts as soon as it reaches the
liquidus temperature for that material composition. It requires no
undercooHng as the partly melted grains provide the nuclei
wherefrom the growth of grains starts into the solidifying weld
pool, such a mode of solidification is referred to as Epitaxial
Solidification. In this section these two modes of soli~ification are
discussed and the possible microstructures obtained under
different sets of thermal gradients are detailed.
3.2.1. General Theory of Solidification of Metals' and
Alloys (Growth of grains in permanent mould castings)
Solidification of molten metal in a mould occurs by the
nucleation of minute grains or crystals which then grow under the
influence of the crystallographic and thermal conditions that
prevail. The size and character of these grains are controlled by
the material composition being cast and the cooling rate. Growth
ceases when all the available molten metal has solidified.
Basically the reason for the solidification of metals and alloys
is that the arrangement of the atoms in a solid crystal is at a lower
free energy· than that of the same atoms in a molten state. Above
the freezing point, however, the liquid state is more stable. At the
freezing point there is no driving force in either direction, in other
words, the change in free energy is zero and there exists an
equilibrium.
* Free energy, F, is a measure of the driving force of a reaction.
136 Welding Engineering and Technology Basic Metallurgy of Fusion Welds 137

Thus, a crucible of molten metal will not begin to freeze at the The energy tending to produce a nucleus of solid is the
freezing point (since there is then no driving force to produce difference in free energy per unit of volume between the liquid and
solidification), but will begin to freeze at some lower temperature. solid phases. This quantity, called the bulk-free energy change,
The farther the metal is cooled below the liquid-solid equilibrium designated as ~ Fy, and is usually expressed in joules/cm3 for a
temperature, the greater is the driving force to solidify. given temperature of supercooling. At all temperatures below the
It has been shown experimentally that crystals cannot form freezing point, ~ F\, is -ve; that is, the solid phase is then more
spontaneously in a molten metal unless the melt is considerably stable, and therefore has a lower free energy than the liquid phase.
undercooled. The degree of undercooling necessary for The bulk-free energy change associated with the formation of a
spontaneous nucleation of crystals in pure metals is
approximately 20% of the equilibrium freezing temperature when spherical particle of radius r is ~ 1tr3~ F y. The energy required to
it is expressed in degrees Kelvin. However, normally foreign bodies create the new surface is a function of the surface tension, cr, in
are present in the melt which greatly facilitate nucleation and joules/cm2. The net free energy change, ~, is then given by the
enable crystals to form when the melt is undercooled only a few relationship,
DC.
Free energy change = Energy required to form new surface
The reason that solidification does not begin immediately + energy required for volume growth
when metal is supercooled* a fraction of a degree below the liquid-
solid equilibrium temperature is that energy is required to i.e. ~=41tr2cr+~1tr3~y ... (3.1)
produce the new surfaces of the crystal being formed, just as it
takes energy to blow and enlarge a soap bubble against the force The surface energy term is positive in contrast to the bulk-free
ot surface tension. energy term, resulting in the curve of ~ F versus particle size
There are two types of nucleation viz., homogeneous and shown in Fig 3.3. Since different powers of r are involved, the
hetrogeneous. Homogeneous Nucleation is defined as the
formation of a new phase without the help of special nucleation
sites. Such a nucleation can occur in a melt of a pure metal.
Heterogeneous Nucleation is the solid phase crystallization on
ell -,-
.c
:z:<ell
c:n
L..
'"
c::
ell
c:oJ
.e-
>-
+
0

foreign nuclei. This is the major form of nucleation.

ell ell

I I 4 3
3.2.1.1. Homogeneous Nucleation 411 r20"'+'311r t.Fv

In homogeneous nucleation small atomic groups having the

lattice structure of the solid material are constantly forming by
chance and redissolving in any liquid metal melt. The possibility
Nuc leus sIze, r
ot these small accidental particles of solid growing to form true
crystals depends upon the energitics of the growth process .
Now, the temperature at which homogeneous nucleation
occurs is always below the equilibrium freezing point because it
is necessary to overcome the surface tension forces which impt;)de
nucleus growth. In other words the nucleation of the supercooled
grain is governed by two factors viz., (i) free energy available for
solidification process, and (ii) the energy required to form a
liquid-solid (new surface) interface; the latter is also sometimes
called surface energy. Fig. 3.3. Nucleus size versus net free energy change.
* The terms undercooling and supercooling are used interchangeably -- both curve will show a maximum at a critical radius r*. Note that once
signify the cooling of a phase to below the equilibrium transformation temperature, this radius is exceeded, further growth results in a decrease of
for example, liquid water at (-l)OC.
138 Basic Metallurgy of Fusion Welds 139
Welding Engineering and Technology

free energy and can proceed spontaneously at the temperature of necessary for it to exceed the critical nucleus size and grow
supercooling. If a particle has a radius smaller than r*, it will spontaneously.
redissolve because this decrease in size reduces the free energy. Once the critical radius is achieved the crystal starts growing
The size of the critical radius may be found by differentiating at an accelerated rate releasing the latent heat of fusion in the
equation (3.1) with respect to r and setting the result equal to zero process; that results in the rise of temperature and reduction in
t.e., number of new nucleating sites. When severe undercooling is
d IA17\
achieved numerous stable nuclei are developed leaving less
~r = 8nrcr + 41tr2 .AF'v = 0 amount of liquid metal for the growth of crystals resulting in fine
grained structure. However, if the undercooling is less severe the
r* =---
2cr
M'v
...
(3.2) nucleating sites are less and the grains grow to bigger sizes and
at accelerated growth rate as is evident from Fig 3.5.
where r* defines the critical nucleus size.
Equilibrium temperature (m.p.)
The significance of this relationship is that for the critical Te
radius to become very small, AF'v must become large in negative Degree ofling
undercoo
sense; in other words severe supercooling is needed for
homogeneous nucleation. The term crdoes not change much with
-- --- U}
temperature, however, AF'v increases with supercooling and
therefore causes a very small nucleus to be stable, as is shown in Growth rate. G
Fig. 3.4. t-

-
Q)
•...
='
Cl
•...
Q)
Q, Nucleation rate, N
e
.•.. ,,
..
\ Resultant for
{!!

r-\ Tl < Te
· 1 \
\
~F 0 ~
rc ,- r Rate parameter; G,N -
1--- Resultant
Fig. 3.5. Effect of degree of undercooling on nucleation and growth rates.
I for T2 < Tl
3.2.1.2. Heterogeneous Nucleation
- j Most actual castings crystallize by heterogeneous nucleation,
Bulk free energy change the basic reason being that if the new phase can find a foreign
particle to grow upon, it can in effect adopt the relatively large
~Jt r36 Fv radius of the particle as its own. This means that only a slight
degree of undercooling is needed in comparison with that needed
for homogeneous nucleation. Quantitatively, the relation depends
Fig. 3.4. Effect of amount of undercooling on the size of critical nucleus (r").
upon the degree to which the new phase wets the foreign particle.
If there is no attraction between the atoms of the foreign particle
Obviously a nucleus formed spontaneously in the melt by and those of the precipitating phase, then nucleation is not helped.
random atomic motion will probably be only of the size of a few The wall of the mould usually provides many heterogenous
clusters of a few atoms; therefore severe supercooling will be
]40 Welding Engineering and Technology 141
Basic Metallurgy of Fusion Welds

nucleation sites. The best nucleus, of course, is a particle of the

Liquid
precipitate itself.
3.2.1.3. Freezing of Alloys
The discussion in the above section has been mainly .'-: T6
~
concerni.ng the solidification of pure metals. However, in industry :::J
+-
pure metals are much less used while alloys find extensive use. o'-
a.
QJ

The solidification of alloys differs in three principal ways from ~

E
that of pure metals viz.,
Co
Pure B
(i) usually, the freezing of alloys occurs over a temperature Pure Acs 0108 -
range while pure metals have definite melting points,
(a) Phase diagram for a solid solution alloy.
(ii) the composition of the solid which separates first is
different from that of the"composition of the liquid,
(iii) there may be more than one solid phase crystallizing
from the liquid at the same time. Soild \
Liquid
Major alloys are of two types viz., (a) solid-solution alloys, (b)
eutectic alloys; however still another type of alloys known as
peritectic alloys are also sometimes encountered in the co
industrially important steels. ~
o
(a) Solid-Solution Alloys
Two metals which are mutually soluble in their liquid state
behave much like water and alcohoL The liquid alloy looks like a
one-substance liquid. Two metals in an alloy appear much the· Weld pool ( b) Change in concetrotion of B with the progress of
same. This resemblance to a single substance may exist in solid periphery ~ SOlidification
Distance from the weld pool periphery-
Temperature at the time of pourmg. t = 0
alloys also. Such an alloy is said to be a Solid-Solution alloy t '"t,
because the two metals are mutually soluble in each other. IV
Copper-nickel and gold-silver alloys present well known examples Liquidus temp.
of solid-solution alloys. Steel within certain range of carbon also .'-'. TE
Nucleation temp.
gives solid-solution alloys, e.g., 0.6% carbon steel is a solid-solution '-
QJ

:::J

alloy of iron and carbon. "§ I, n ,Ill, IV - Possible temperature

Solidification of Solid-Solution Alloys a.
QJ
gradients at the soild-liquid
interface
E
~ Soild Liquid
If metal B is dissolved in liquid (molten) metal A to form a
liquid alloy, the crystallization can be described by means of a
phase diagram, as shown in Fig 3.6 (a). Let us assume that the Distance from the weld pool periphery-
freezing point of the metal A is depressed with the addition of (c) Change in liquidus and nUCleation temperature with
metal B, as shown by the liquidus' line. In addition, the alloy of change in concentration of B (solute)·
composition Co does not freeze at a single temperature, but instead Fig. 3.6. Weld metal solidification in solid solution alloys.
over a temperature range. parent liquid, but is richer in metal A. The metal B-atoms in the
Another important fact is that as the alloy solidifies the crystal structure of metal A are in solid-solution and distributed
composition of the solid that grows is not the same as that of the among the metal A atoms in an atomic dispersion. How these
effects lead to a solidification pattern different from those
* The temperature at which freezing begins is called the liquidus and the
temperature at which freezing is complete is called the solidus. described for the pure metals are discussed as follows.
142 Welding Engineering and Technology

If a homogeneous melt of composition Co is taken, the first

Basic Metallurgy of Fusion Welds

Thermal
-- 143

crystals to precipitate are of composition Cs' If solidification is Gradient

fairly rapid, so that no diffusion occurs, the liquid at the interface
becomes richer in the solute (metal B) than the liquid away from
the interface. The variation of% ofmetalB with distance is shown

*
(a) Planar growth - columnar grains
in Fig. 3.6 (b) for temperatures between liquidus and solidus.
Assuming that the mould is filled rapidly and that there is no
temperature gradient at the start when time, t = 0, as shown in Thermal
Gradient 11
Fig. 3.6 (c).
Consider the liquidus temperature of composition Co as TE•
Assume that a thermal gradient is established and the liquid alloy )
js supercooled somewhat before crystallisation begins at time t1.
Then, consider the situation at time t:,l'(after appreciable solid has
I bl Spiked growth-cellullJl' or honeycomb structure.
separated). If a temperature gradient similar to that of the pure
metal is assumed then the difference between this situation and
that of pure metal is that the 'liquidus temperature of the Thermal
remaining liquid now varies with distance from the solid-liquid Gradient. III
interface, as shown by the dashed lilie. Near the interface, due to
buildup of component B in the liquid, the freezing temperature is
considerably lower than TE• However, away from the interface, the
(cl Dendritic growth -dendritic structure
equilibrium freezing temperature is still TE, if -no diffusion is
assumed. Fig. 3.7. Effect of thermal gradient on weld metal structure. '

It is to be noted that a region of liquid at point x has a higher
freezing point (dashed line) than the liquid at the interface, while
the actual metal temperature at x is also higher than tha.t at the
interface. However, the difference between the liquidus and actual
temperatures at x is greater than the corresponding difference at
the interface. Under such conditions the liquid at x is said to be
constitutionally supercooled. The bulk free energy of the liquid at
x, which is available for crystallization is therefore greater.
The effect of constitutional supercobling upon crystallization
are of three types, depending upon the degree of supercooling.
(i) If there is only minor supercooling, certain preferred
regions of interface will protrude as spikes into the su-
percooled region and, once started, will grow more rapid-
ly than neighbouring regions as shown in Fig. 3.7 (a).
This will happen both because the driving force for freez-
ing is greater in the supercooled region and that the
spikes will reject solute at their sides, thus delaying
freezing of the side regions. The spikes may result in the
formation of a honeycomb structure.
(ii) If supercooling is greater, the spikes tend to form side
arms, producing a dendritic structure as shown in Fig
3.7 (b).
(Ui) Finally in the case of extreme supercooling, the tempera-
ture difference, TUquidus - Tsolidus, which is maximum at
some distance y in Fig 3.6 (c), may become large enough
to lead to independent crystallization. In this way, ran-
domly oriented (equiaxed) grains may be encountered
towards the central part of an alloy casting. An extreme
situation of this type is represented by Fig 3.8.
Th.ermal 6
Gradient IV A 464
••4
46
~6"
(d) Dendritic cum indepedent nucleation-composite structure
containg dendrites and equiaxed grains.
Fig. 3.8. Extreme case of undercooling resulting in a weld metal structure
consisting of dendrites and equiaxed grains in central section.
 144
Welding Engineering and Technology I Basic Metallurgy of Fusion Welds 145

The relative solubility of the solute is given by the factor K,

C
where K == C:· Typical values of Klie between 0.01 and 0.05. When Liquid
(Ll
both K and thermal gradient are low, even the metal at the centre .w
can be constitutionally supercooled. On the other hand, it may be
noted that the steeper the thermal gradient, the greater the
possibility of planer (columnar) growth. This tends to prevent
-c
~ Tperitectic
•....
OJ
dendritic growth and random crystallization ahead of the freezing a.
E
interface and thereby facilitates metal flow for feeding of ~
solidification shrinkage. In other words, the steeper the thermal 1
gradient, as in chilled castings, the simpler the feeding problems.
The correlation between rate of solidification and thermal
gradient can be very effectively utilised to predict the type of Peritectic composition, Cp
microsturcture that will be obtained for a given set of conditions,
as shown in Fig. 3.9 Composition -

Fig. 3.10. Phase diagram for a peritectic reaction.

I Planar
8-phase to y-phase so retained 8-phase may be encountered in the
G- Thermal gradient
R- Rate of growth 'of microstructure. Otherwise also peritectic reaction is very sluggish
lolid crystals compared with precipitation of y-iron from a liquid. The reason IS
.,/"' that a solid phase 8 has to react with liquid to form a new phase
y. The new phase will coat the remaining 8 as shown in Fig 3.11.
l.:)
For the reaction to continue, diffusion through the solid y is
required, which is a relatively slow process.
Dendritic +
independent Liquid
nu c leation

Fig. 3.9. Interaction between Rand G to get ~~~l~-

different types of weld metal structure.

(b) Solidification of Eutectic and Peritectic Alloys Fig. 3.11. Conceptual representation of peritectic reaction.
The type of microstructure obtained for eutectic and peritectic
alloys will be similar to those obtained for solid-solution alloys
3.2.2. Epitaxial Solidification
except that the effect of phenomena encountered in these alloys In the case of fusion welding the weld pool is formed with or
will effect the final outcome. For example, pretectic reaction, without the addition of filler metal but a part of the parent
depicted in Fig. 3.10, shows that a solid phase and a liquid react material is invariably melted to provide the necessary
to give a completely new solid phase as is the case in low carbon metallurgical bond. In doing so the metal grains along the fusion
steels where liquid and delta (8) phase react to give austenite. boundary are partially melted. The remaining partially unmelted
grains provide the necessary nuclei for the growth of crystals to
L + 8 ---:> y ... (3.3)
begin, as shown in Fig. 3.12. Such a mode of growth is referred to
Now, if the temperature is changing fast, as in cooling of welds, as epitaxial solidil'ication and is a characteristic of weld pool
then diffusion of liquid into 8 is not fast enough to convert all solidification. \
146 Welding Engineering and Technology
Basic Metallurgy of Fusion Welds 147

Weld pool
of the molten metal in the weld pool. This comparison, between
Columnar,dendritic
solidification of a casting and the weld metal, on the basis of free
/ grains
energy concept can be represented as shown in Fig. 3.14.

*
r - cri tical radius

Bulk
energyfree +1
change MO
Nuc leus radius
Fig. 3.12. Epitaxial growth of crystals in welg pocl.
, 1
Now, in general the critical radius r * and the corresponding Casting
free energy of nucleation, 6F., on a planar substrate are given by
the following relationships .
• 2YLC' Tm
Fig. 3.14. Comparison of critical radii (r") for casting and welding:
r = Ml m . t1T ... (3.4) epitaxial solidification in welding does not require new nuclei.
41t1ic·T m 3 Although weld pool solidification from the fusion line starts
and D.F* = ') . (2 - 3 cos 8 + cos 8) ...(3.5)
by the phenomenon of epitaxial growth but all the grains do not
3 (Mlm· t1T)
where, YLC = the surface energy of the liquid/crystal interface,
survive as the grain structure is governed by what is known as
competitive growth, explained as follows.
T m = equilibrium melting temperature,
During'solidification grains tend to grow in the direction
Mlm = latent heat of melting of base material, perpendicular to the solid/liquid interface, since that is the
t1T = degree of under cooling, direction of the maximum temperature gradient and thus the
8 = the contact angle as' shown in Fig. 3.13 maximum force for solidification; The grains also have their own
1LC
preferred direction of growth called the easy growth direction.
Liquid Thus, during solidification, grains with their easy growth direction
( L) parallel to the direction of maximum temperature gradient will
grow more easily and pinch off other grains whose easy growth
direction differs significantly from the direction of the maximum
1 LS
thermal gradient, as shown in Fig. 3.12. This is referred to as the
- Partly melted
competitive growth of grains.
grain in HAZ The weld metal zone may not consist only of the columnar
grains growing epitaxially from the fusion line, instead some may
Fig. 3.13. The contact angle between the solidifying metal (crystal) in weld popl grow independently by the usual phenomenon of nucleation and
and the substrate (partly melted grain in HAZ), growth. This is, however, based mainly on heterogenous
The molten metal of the weld pool is in intimate contact with nucleation and growth. Such grains are normally equiaxed fine
the substrate, thus there is complete wetting of the substrate. grains.
Therefore, the contact angle 8 is zero, leading to D.F" = 0, as per It is often intended to achieve fine-grained structure in the
equations (3.4) and (3.5). As a result grain growth initiates from weld bead because such a structure leads to,
the parent metal substrate without the need of any undercooling (i) reduced susceptibility of the weld metal to solidification
cracking during welding,
148 Welding Engineering and Technology
(ii) fine grains help improve the mechanical properties like
ductility and fracture toughness of the weld metal.
Fine grain structure in weld metal can be developed by the
processes of inoculation, arc oscillation, arc pulsation, and weld
pool stirring.
Weld pool stirring is achieved by applymg alternating magnetic
field parallel to the welding electrode. This results in increased
thermal conductivity of the liquid pool leading to increased heat·
transfer in the weld pool with consequential loss in weld pool
temperature. That helps in heterogeneous nuclei to survive and
significant amount of equiaxed grains are thus produced.
3.2.2.1. Effect of Welding Speed on Grain Structure
At low welding speeds, the weld pool tends to be elliptical,
whereas at high welding speeds the weld pool takes an elongated
shape. Since the boundary of the trailing portion of the elliptical
weld pool is curved, the grains are also curved in order to grow
perpendicular to the pool boundary as shown in Fig 8.15 (a). On
the other hand the boundary of the trailing portion of the
elongated weld pool is essentially straight, the grains are also
straight and growing essentially perpendicular to the pool
boundary, as shown in Fig 3.15 (b). Therefore, for weld metals free
from heterogeneous nuclei, it can be expected that at low welding
speeds columnar grains will curve and grow in the direction of
welding, while at high welding speeds they grow straight towards
the weld centreline. Thus, at high welding speeds abrupt changes
:;~~'~'
1I:":v ~p,,)
(0)
Welding direction
Axial
~P:)
grains
~~
(c)
Welding direction
Low speed
High
Fig. 3.15. Effect of welding speed on grain growth and orientation.
Basic Metallurgy of Fusion Welds 149
in growth directions take place which can be .detrimental to the
toughness ofwelds due to the increased risk ofsegregation at the weld
centreline. This restricts the maximum welding speed that may be
used for welding a given material under specificwelding conditions.
Another type of columnar grains called axial grains can also
grow in the fusion zone. Most of these columnar grains initiate in
the original weld bead and continue along the length of the weld,
ulocking the columnar grains growing in from the fusion lines.
Since, like other columnar grains, these axial grains also have to
grow perpendicular to the weld pool boundary, the band of these
axial grains is sufficiently wide with slow welding speed elliptical
weld pool but is rather narrow with high welding speed elongated
weld pool as shown in Figs 3.15 (c) and (d). The axial grain
structures have been observed in welds of aluminium alloys and
austenitic stainless steels.
3.2.2.2. ?roperties of Weld Metals
Properties most required of weld metals are toughness and
strength. In general strength and toughness of the weld metals do
not match those of corresponding base metals. The ratio of yield
strength to ultimate tensile strength is always higher in weld
metal than in base metal. Therefore, if equal strength is required
of the weld metal and the parent metal in a weldment then the
yield stress of weld metal will have to be higher because with an
equal yield stress, strength of weld metal will be lower.
Properties of weld metals are greatly influenced by the type
of microstructure, grain si.ze, as also by precipitation processes,
ageing, etc. Also, as discussed earlier, microstructure can be
affected considerably by welding parameters like welding speed,
heat input, etc. It is imperative to know that an important role is
played not only by the basic type of microstructure but also by its
compounds. The occurrence of ferrite with aligned 'MAC
! b) (martensite, austenite, and/or carbides), martensite and retained
austenite (MA) aggregates is disadvantageous in that it has low
cleavage fracture energy and can serve as a reinitiation centre for
spreading the cleavage fracture. Particles of second phase i.e.
oxide inclusions, carbides or nitrides, playa double role i.e. both
positive and negative.
Their role is favourable in that,
(a) they promote the formation of an optimum microstructure
of acicular ferrite (AF),
(d) (b) they inhibit grain growth,
(c) nitrides and carbonitrides of alloying elements like Ti, Nb,
V, etc. chemically bind Nz thereby making the weld more resistant
speed
to ageing.
150 Welding Engineering and Technology Basic Metallurgy of Fusion Welds 151

Their adverse effects are that they cause precipitation
embrittlement of welds, a decrease in ductility by aiding initiation
of ductile fracture through coalescence of microvoids at the
particle-matrix interface.
Toughness properties of weldments are worse than those of
rolled or forged steels. In addition to strength and toughness,
weldments are also required to have good fatigue properties. In
general fatigue properties of weld metals are inferior to those of
corresponding base metals. Because fatigue properties are affected
mainly by weld contours as well as by morphology and dispersion
patterns of inclusions therefore to obtain good fatigue properties,
weld metals must be sufficiently pure. Also, because of a lower
carbon content in weld metal in comparison with that in the steel
forming the HAZ and the base metal, the creep strength of weld
metal is always lower.
In conclusion it can be said that the weld metal properties are
affected not only by the chemical composition of the filler material
used, ,but also by the chemical composition of the steel used as the
parent material, the welding process employed, and the weld
cooling rates.
3.3. Fusion Boundary Zone
Different terms are used to define the volume of parent metal
actually melted, Fig. 3.16, to form the weld deposit. Tweeddale
In most metals and alloys, the fusion boundary zone is quite
sharp and may be referred to as fusion line. At the fusion line, the
composition changes from that of parent metal to that of, more or
less, uniform weld deposit. For a manual weld with coated
eleCtrodes, the thickness of fusion boundary zone lies typically
between 50 and 100 microns·. In alloys having a long freezing
range, partial melting may take place in that portion of the HAZ
which is immediately adjacent to the weld deposit and some of the
liquid pockets so formed may become physically continuous with
the weld metal. The alloys that behave in this way are usually
difficult to fusion weld and, except for cast iron, are rarely met
with in practice.
One of the two types of grains viz., fine and coarse, may form
in the fusion boundary zone. In steels, this may depend on the
y ~ y + 0 phase transformation on heating. Referring to Fig 3. 17(a)
T T
1 1
Co = 0·15 .co

Weld pool (- !M+C)--

(0) ( b)

Liquid liquid
liquid + 6
6.•1 liquid + 1
Columnar grains
(dendritic or cellular
1
dendritic)
l
Boundary of melting

Fig. 3.16. Fusion boundary zone in an arc weld.

calls it fusion zone while Kenyon refers to it as fusion boundary.
However, over here in this section the zone under discussion is
one between the weld deposit. and the HAZ and is a partially
melted zone. This zone, referred to as fusion boundary zone, for
low carbon steel welds corresponds to liquid plus delta (L + 0)
ferrite zone on the iron-carbon equilibrium diagram of Fig. 3.1.
Fig. 3.17. Schematic iUustration of microstructure of the fusion
boundary zone: (a) with y ~ y + /) transformation,
(b) with y ~ y+ L transformation occurring on heating.
* 1 micron = 0.001 mm.
152 Welding Engineering and Technology
I Basic Metallurgy of Fusion Welds
153

the alloy composition is such that the base metal in the fusion Depending upon the peak temperature reached the HAZ in
boundary zone transforms to <5-ferriteand liquid at the grain steels can be sub-divided into the following zones, starting from
boundaries. Because of the lower solubility of C and Mn in ferrite t.he weld metal side.
(both are austenite formers) these elements tend to segregate to 1. Underbead Zone i.e. that part of HAZ which is heated to
the melted grain boundaries. On transforming back to austenite beyond the critical temperature of grain growth and
during the cooling cycle, the high degree of segregation at the extends up to the fusion boundary zone.
boundaries provides innumerable nucleating sites for o-ferrite 2. Grain Growth Zone, beyond 1150°C to peritectic
thus reduces grain growth leading to fine grains in the fusion temperature,
boundary zone. It also very effectively modifies the composition of 3. Grain Refined Zone, 950 to 1150°C i.e. beyond A3 up to
the austenite. This, together with the effect of a smaller grain size, grain refined temperature range,
affects the austenite transformation products, compared with the
larger austenite grains away from the fusion boundary zone, by
transforming at the higher temperatures. This accounts for the o
N
more equiaxed o-ferrite structure in the fusion boundary zone.
In the second case, which is typical of plain C-Mn steels, the
-::
<i"
finer grain size in the fusion boundary zone is not observed. Fig. ' I I w ,.,'
<= E
u...
QI ,.,
lj<::)
C) + co~ C)
000c>C) 00 a 0 C?w .-.,
li ~ - - -
c> Cot
3.17 (b) shows that there is no y ~ <5+ y transformation in this
_ •••• CIl
•...

i I;! ~i~!~ i ~I
bJ)
-r-~_I~
"'i ~~ ~,~!~I
.-
case, so that. on heating austenite transforms -4only to (y + L) at the
grain boundaries. Since there is limited segregation in this case,
-i
.95 ~
N
•
w..;{Vi'QI
=--0
..c:
...,
"'0"'0
C S
CIl

the composition of the auetenite in the fusion boundary zone is il~

\\ I. 'I ..\ I I" 6 :::c~
LJ"'"l

,00
QJ .- CIl

(/J'-
::l

not expected to change substantially. Also, transformation during II \ I I i <-0

,., N u .,-
:E~
~'5
cooling of the unmelted part of the grain occurs at similar -Ql1i':
QI E III
_0'
., .,
temperatures to those grains not in contact with the liquid phase.
c:
00-0
L.. ._
$ c0
N'+-o
III '-
(/J

c..c
c:
3.4. Heat Affected Zone -0
~e.c.
._ .•.....0III-
QI
.85
.....d
_ (.)
CIl
0
A heat affected zone (HAZ) of a weld is that part of the welded ~2="~
- 0 ~.:::
d
joint which has been heated to a temperature up to the solidus of ._
c: •••.!:! OJ
..sa
•••• CIl

o '- c: ~§
the parent material resulting in varying degree of influence on '-00
l:JO-N
microstructure as a consequence of heating and cooling cycle. ~
:I:~
g)

••••-0 QI
When metals and alloys without polymorphous transformation* ••••• cd
o •...
(e.g., Cu, Ni, AD are welded, the microstructure in the HAZ c •...
(/J .,

remains unaltered though grain growth or recrystallization may

QI
c: OJ
oN QI c:
.-o ...,::l
0
N_ .~
.- e.,
take place, while in the case of metals and alloys with C
c:
2 ~.w
"'00-

polymorphous transformations (e.g., steels), significant QI "'0 .0::l...,

~
+--o3:..n
microstructural changes take place in HAZ, that in turn influence ad::;::
EQlO'\.
Cf:IbJ)
cO .9
the mechanical properties and consequently the service behayiour 0
>-'-CLJ"
.c ~]
of the welded joint. ~a.
:;:::
d..- Q.I '-...-
,,;,8..
OC'-_
~""O

0- :::l l:J ~.,

._ (/J
•...
ci •...
o
E 0.0.0
--::.. C" (.)
QI
* Polymorphous Transformation: Existence of a given solid material (e.g. +-
~
steel) in two or more stable but different crystal structures (FCC, BCC, HCPl under o
QI
different conditions of temperature, pressure, or method of preparation is termed 0-
as polymorphism. The terms polymorphism and allotropy are used interchangeably
by most authors though they are not exactly same. The conversion of one
polymorphic form to another is referred to as polymorphic transformation.
 154
Welding Engineering and Technology Basic Metallurgy of Fusion Welds 155
J
4. Partially Transformed Zone, 750 to 950°C i.e. between welding processes except ESW, while heating rate is
Al and A3 temperatures, much slower and the high temperature retention time is
5. Zone of Spherodized Carbides, 550 to 750°C i.e. below much longer in heat treating of steels. Fig. 3.19 shows
Ai> the thermal cycles for welding and heat treating of two
6. Zone of Unchanged Base Material, up to 5500C. types of steel with three diffe{'ent ratf;ls of cooling.
3. The high heating rate, together with the short high-
All these zones are shown with reference to the relevant temperature retention time, can also result in the forma-
sections of the iron-carbon equilibrium diagram in Fig. 3.18. tion of non-homogeneous austenite during welding. Such
The final microstructure of a section of a HAZ depends upon a non-homogeneous austenite, upon subsequent rapid cool-
several factors including composition, grain size, peak temperature ing can cause the formation of localized high carbon mar-
attained, heating and cooling rates, etc. Although it is often said tensite colonies. Consequently, the microhardness of the
that the HAZ of a weldment is its heat treated part but there is HAZ tends to scatter over a wide range.
considerable difference between welding and heat treating of, say,
steels. The microstructures and characteristics of the different sec-
tions of HAZ for welds in steels are described in this section.
1. Near the fusion boundary of a weld, where difficulties 3.4.1. Underbead Zone
such as grain coarsening and underbead cracking often
arise, the peak temperature can reach up to 1400°C or This zone lies next to the fusion boundary zone and undergoes
even higher. In the heat treating of steels, on the other austenitic transformation (which is usually presented in the form
hand, the maximum temperature involved is usually of CCT diagram) and grain growth in practically every fusion
around 950-1050°C. welding process employed.
2. The heating rate is very high, and the retention time at Now, grain size is of paramount importance in metals, it being
high temperature ie normally very short in most fusion a key factor in determining the strength and toughness of a
Welding cycles Heat treatment cycles
material. It also has enormous importance with respect to
determining susceptibility of an alloy to cold cracking and reheat
TM
cracking.
Tm The fine size of the grains and the fact that it is impossible to
achieve mechanical equilibrium between grain boundary tensions
in a 3-dimensional arrangement of grain means that the structure
.'""_ A3
contains a certain residual stored energy. Thus, grains with more
'"
'- sides begin to swallow grains with less sides.
.-c;;, With present day structural steels the critical temperature of·
'-
'" grain growth is 1050°C while for microalloyed steels it may go up
a.
E to 1350°C. The degree of grain growth depends on the chemical
t-.•.. composition of steel and the specific heat input in welding. A
relatively small grain growth is observed in SMAW, it is more
significant in SAW, and is maximum in the case of ESW. Grain.
t' = 6 growth in the underbead zone presents one of the major
weldability problems, with all modes of welding, for every metal
t"= 180
1,2,3-Different and alloy because coarse 'primary' grain causes,
cooling rates (i) decreased zone plasticity,
(ii) slowing down of austenitic decomposition that favours
i) Heating time to maximum temperature,Tm=t'
ii) Time above A3 = t' ferrite precipitation in the cleavage planes of austenite
with the formation of widmannstattan microstructure;
Fig. 3.19. Comparison of welding and heat-treatment cycles for steels.
156 Welding Engineering and Technology Basic Metallurgy of Fusion 'Velds 157

(iii)increased susceptibility of steel to cold cracking, stress small. Also, the carbides may not be fully dissolved. The y ~ a
relief cracking, and Equation. transformation on cooling, therefore, tends to produce a fine-
(it!) lowering of strength of underbead zone in metals which grained ferrite-pearlite structure depending upon the factors like
do not undergo polymorphous transformation (Al, Cu, heat input, plate thickness, etc. The effect of heating and cooling
Ni, etc.). cycle on grain growth and refinement in this zone is well depicted
Further it is known that interphase precipita.tion occurs in low by Fig. 3.20. The large grain boundary area tends to promote ferrite
hardenability alloys, especially the Nb-bearing micro alloyed nucleation and the austenite that remains at the grain centres is
steels, and the presence of this phase is likely to adversely effect 1200
toughness. 1100
STEEL COMPo

3.4.2. Grain Refined Zone 0.09·'.Nb

1000
The peak temperature in this zone does not exceed 1150°C
therefore in a ~ y transformation during heating austenite does t 900
aoo
not have time to develop properly and thus the grain size remains .u
• 700
I w
'"0
~ &00
o~
03: .~
4a:
0>0
e_ CIl

'"p ~
'".0
d
~
~
S.£:
.£:- • ;:s
UJ
t-
>-~ ~ •..

00000000000000 .0.2 .a CIl

-~
c~ e
•..
~.s E
~
<1l
...,
100
o
:0000000
4
e'" •••
-.;::E_
.;;
d 1 2 4 (,810 2 4 681012 4681032 4 6810 2 4 6
e!"'C - .~ COOLING TIME. See

--
•••0 6D
e-",.~ (a) In reo\ time
'00::: od
.w
IS>
.....
'"
:53;::;

:>."0 •••
_ •••0>
~l
<:l
>.
u
l>Il
- 1100
MARTE"ITE
Ms ...-
1200
~.
--
r.../
I.--I II--+---IBAINITE
f--- II ZW
~~ rf ~
I !.--L..l-.1 ~_
T C'FERRITE
:C15°
~Ill.£:•..c••• .9
"0
-
AUSTENITE AC3
I--
1000
Cooling. 8
o~c '"d , 900

00 000000 C;;
•.•0
E;:lO>
•••0 ::0
.£: •...
- ",.-
~·a
==
E
.
c<!

.S
.•..
•u
800
11; 700

~0
E ~.~~
oe
•... ~ooling
0>a.3111
.~ E c
o~·o·f
'"
CIl
.J::
•.•...
a:
;:: &00
t= 0 ~.~~
lIe 0::-
d
0000000000000000000 "'_B •...
..c:. OlE0
o
tl
Cl:
a: 500
w
~ ~ 400
o 000 >o:.t:: ~
.0.;:: 1: •...
•.••
0 ••.••
-
ril w
o •... 300
000000000 c::

E~153:
~~

bloN .•..•
f::J"''"'::
C'I
co-)
200
~
~ 5U!~DaH c- .-
._ III'
•••• Cl).I'- ~
t>il
100
o
.w
.....
~ ~.s:S~
E;:l e ~ ,",0 o 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32
N
r- 80 ·0 0 ~ 0 COOL ING TIME At ( 800 - 500) •See -
0::Q.
k1 '-'•....00.00::0
E I b) With reference to temperature drop from 800 to 500·(
~
Fig. :-1.21.Typical austenite decomposition diagrams under welding conditions for
steel of othe~ ~omposition given as inset in (a) above.
158 Welding Engineering and Technology I, Basic Metallurgy of Fusion Welds 159

rich in carbon and transforms to pearlite. This zone tends to be

particularly wide in microalloyed steels, because of the effective-
ness of the carbonitrides in preventing grain growth at these
tem pera tures.
In general the nature of microstructure of the grain refined
zone can be determined accurately by the austenite decomposition
diagram under welding conditions. This diagram may differ from
that of continuous cooling curves primarily in that the structural
changes in the zones of weld joints are characterized by the
presence of gradients : a temperature gradient results in the
formation of cont::entrational and deformational gradients. Fig.
3,21 shows a typical austenite decomposition diagram under weld-
ing conditions for structural steel. Such diagrams in conjunction
with cooling rates for tS/5 (i.e. cooling rate for temperature drop
from 800°C to 500°C), enable us to optimize the parameters of
welding.
arc weld Fig. 3.23. Microstructure of ferrit:c-pearlitic base material.
steels with ferritic-pearlitic structure, Fig 3.23, eutectic pearlite
begins to dissolve in the zone heated beyond temperature At·
QI
L-
:J
.•...
C
L-
QI
a.
E
~

Acicular ferrite
I
Martensite End of tl'ansfo rma tion

Log time -
Fig. 3.22. A CCT diagram for a steel welded by submerged arc welding process.
Although the diagrams of austenite decomposition under
welding conditions represent the most accurate production of
microstructure under given welding conditions, CCT diagrams
approximate them closely. Fig. 3.22 shows a CCT diagram for a
steel welded by submerged arc welding process. Fig. 3.24. Microstructure of chunky pearlite in partially transformed zone.

3.4.3. Partially Transformed Zone (At - Aa temperature
range)
In this zone steel is heated to between Al and A3temperatures
i,e. approximately between 750-950°C. In welding low carbon
\
However, as pearlite to austenite transformation takes through
nucleation and growth, it requires certain time to be completed but
at temperature A3 this process comes to an end. If subsequent
cooling rate is higher, there is no time for the reverse process to
160 Welding Engineering and Technology
take place--carbon fails completely to diffuse back into the former
pearlitic grains. The result is the formation of a rosette structure,
called chunky pearlite, as shown in Fig. 3.24. At higher speeds of
cooling, a former pearlite grain can be quenched to martensite. In
extreme cases a DP (dual phase) ferritic-martensitic microstruc-
ture may form. DP steels of high strength find use in structural
fabrications. Controlled micro-structural changes brought about in
this zone of intercritical temperatures can be utilized to improve
plasticity and toughness of thick plates, and for improving the
fabrication of components for use in nuclear energy installations.
Apart from the phases mentioned above, it is possible to obtain
upper bainite, autotempered martensite or high carbon martensite
by controlling the cooling rates.
3.4.4. Zone of Spherodised Carbides
In this zone the peak temperature rises to about 550-750°C,
and in appearance it is an extension of the partially transferred
zone with reduced occurrence of chunky pearlite. This is because
there is relatively little a ~ y transformation during the rapid
heating cycle, so that the most notable change concerns degrada-
tion of the lamellar pearlite to spheroidal particles of Fe3C as dark
phase. The agglomeration of spherodized cementite particles of
grain boundaries and triple junctions indicates high diffusivity of
carbon through the grain boundaries at these (550-750°C) com-
paratively low temperatures.
3.4.5. Zone of Unchanged Base Metal
The peak temperature in the zone of unchanged base metal
rises up to about 550°C and there appears to be no changes in the
morphology of constituents. However, the combined effect of
heating and residual stresses can cause dynamic strain ageing to
occur. This is caused as a result of moving dislocations sweeping
up interstitial impurities such as C and N. On cooling, these C
and N enriched dislocations are strongly locked in position, thus
embrittling the structure. This problem is accentuated by
additional welding runs, as in multi-run welds, or during reheat
annealing.
3.5. Properties of HAZ
The microstructure of the grain growth zone, above all other
zones in HAZ determines the properties of the welded joint. To
predict the properties of this zone it is required to know the
amount and extent of grain growth and the weld thermal cycle.
The extent or width of the grain growth zone is important in
determining the maximum length of easy crack propagation in
Basic Metallurgy of Fusion Welds 161
relation to fracture toughness. The knowledge about thermal cycle
is important
cooling rate. in determining the extent of grain growth and the
A decline in strength of the HAZ is observed rarely when Q
and T, and thermochemically treated processed steels are welded.
A certain softening may be noted in a zone heated to a temperature
range aroundAl> due to excessive tempering of-the microstructure,
and recrystallisation. In such cases, the specific heat input of the
welding process employed has to be limited so as to keep the width
of the tempered zone as narrow as possible. In welding large
thickness steels by ESW, the softened zone may be shifted close
to A3 temperature or slightly above A3 as the structure in this z'one
is recrystallized to almost purely ferritic form.
In HAZ with temperature above A3 and particularly in the
underbead zone either coarsening of austenitic grains takes place
that leads to decline in strength or structural changes take place
that result in oversaturated bainitic-martensitic microstructure
whichmetal.
base leads to marked strengthening of HAZ as compared with
>
%:
•••
•••
QI
c:
~~
o
%:
--
1mm
Fig. 3.25. Hardness survey across different zones of weldments :
Curve (a) pearlite-free steel, (b) low carbon steel.
162 Welding Engineering and Technology
Basic Metallurgy of Fusion Welds
Because the strength properties of martensitic-bainitic { 163
• WELD
structures depend on the fine characteristic of martinsitic or METAL ZONE OF
POSSIBLE
bainitic laths, the transformational strength outweighs any
EMBRITTLEM ENT
softening brought about by grain growth.
Hardness testing is another method of testing the quality of a
weld as it enables mechanical properties to be determined at very I
f
small distances. Typical hardness curve across weld joints is
shown in Fig. 3.25. :a
w
There are steels which are less sensitive to thermal cycles z
:c
C>
during welding, for example the so called pearlite-free steels (C ~ =>

I
<::>

0.09%) hardly react to the effects of the thermal cycle. Fig. 3.25 •...
:c
shows the hardness pattern plotted across the HAZfor such a steel u
•...
in comparison with the conventional low carbon steel. Thus, it can oz MILD STEEL
I
be said that some steels are more, and others less, sensitive to the
effect of thermal cycle in welding. Majority of the alloy steels show
a martensitic or a martensitic-bainitic structure over a wide time -'
interval of cooling from 800°C to 500°C (tS/5)'
Presently, increased attention is being paid to strengthening
~
w
:::i:

-'
Cl
w
u
~
Z .u00
Cl .u.u.u .u
0000
I
.u
o
of the underbead zone as affected by welding parameters. The W
~
o 0"'00 o
lXl N a> •...• \Q~
I I N
maximum underbead zone hardness expected for non-alloyed I
structural steels can be calculated by the following relationship
...
...
.•..•.
.•.
, ..
put forward by eoe.
HV max = 90 + 1050e + 47Si + 75Mn + 30Ni + 31Gr ...(3.6)
where, different elements are in %ages. Similar relationships are
available for other steels, and different plate thicknesses.
Fig. 3.26. Schematic representation of notch toughness of various Zones
Till recently a rule about maximum hardness ofthe underbead of HAZ of mild steel and high strength low alloy steel.
zone adhered to was 350 HV, or 400 HV for low-alloy steels
operating in the area of creep temperatures but never more than Because this decline in toughness value is also related to
125 HVabove the hardness ofthe base material. No such universal the formation of unfavourable microstructure of the zone,
rule is recognised now. The admissible maximum hardness will it is often designated as transformation e.mbrittlement.
depend on chemical composition of steel and the content of diffused Fig. 3.26 shows a schematic outline of toughness of various
hydrogen. zones in HAZ during welding of mild and low alloy steels.
3.5.1. Toughness of the HAZ Similar to the case of hardness, equations have also been put
Generally strength properties of a HAZ present no serious forward for predicting the toughness of HAZ. However, as
problem; this, however, could not be said of toughness properties. toughness values are affected not only by its chemical composition,
This is because of possible degeneration of plasticity in the HAZ microstructure, and grain size but also by several other factors,
due to the following reasons. prediction of toughness of an underbead zone is rather difficult.
t
3.5.2. Other Properties of HAZ
(i) Ageing which may result in decline in toughness in the I
I
zone below At , temperature i Welding can also cause deterioration of corrosion resistance of
(ii) More conspicuous decline in toughness takes place in steels, for example, stress corrosion cracking (See) may OCCurin
underbead zone particularly due to grain coarsening. structural steels in HAZ in the presence of certain sulphur
compoun,ds especially of H2S, chlorides and alkalis, if hardness of
I the under bead zone is unduly high. For example, the maximum
~~"T''''·":!?:;-,~-~~-~~";_;:,t:H~;.IIl"§',~',~_,;;:,,,,.~_~_~ ~ • _

164 Welding Engineering and Technology

Basic Metallurgy of Fusion Welds
165
hardness value permissible for pipeline conveying sour gases is
240-280 HV:Conditions are even stricter for weld joints in off-shore
constructions compellingthe designer to lower the carbon content in
the steels to avoid putting limits on welding conditions.
Overall it can be said that the properties of the weld metal
and HAZ of a wel.dment will depend upon the composition of the
parent material, type of heat treatment it gets due to the thermal
cycle involved and the microstructure produced as a result of that
thermal cycle.
3.6. Heat Treatment and Related Processes in Fusion
Welding
Depending upon the composition of the parent material,
welding process employed and the associated welding conditions
involved various heat treatment and related processes may take
place or may be made to take place for achieving the desired end
product. Some of the well known processes and treatments 0-83 'c
amongst them include the following.
1. Annealing, 0·4 0·6 0.8 1.0 1.2 1.4 1.6
2. Sub-critical Annealing, %·c.-_
3. Spherodization, I a I Temperature rang~ for full and
4. Normalising, isothermal annealing.

,
5. Quench hardening,
900
6. Tempering, \
7. Secondary hardness, 800 \
Furnace cooling I full annealing I
8. Austempering, \
700
9. ·Martempering,
10. Stress relieving, .,": 600 "- --T
Salt bath
- --"'--,
11. Solution treatment, QI \- Free cooling
12. Precipitation hardening or Age hardening, .-o5500 \ I isothermal
\ \ annealing I
13. Strain age embrittlement, ~ 400
a.. \
t-f
14. Temper embrittlement, 300 \
15. "'Graphitisation, \
\
16. Carbide precipitation and sensitization. \
Brief description of these processes follows. 100 ~
3.6.1. Annealing o
Welding may seriously affect the size and the conditions of the lime -
grains of which the material is composed. Depending upon the
welding process used, the grains of the material may grow to large (bl lime -temperature cycles for full annealing and isothermal
size or they may be distorted due to the stresses set up during annealing processes.
welding and subsequent cooling. Such stresses are corrected by Fig. 3.27. Temperature range and heating and cooling cycles-for
annealing and the grains refined, so that the material becomes full and isothermal annealing processes.
softer and more ductile, and free from residual stresses.
I For annealing or full annealing a steel weldment is heated to
30 to 50°C above the upper transformation temperature (As)which
166 Welding Engineering and Technology Basic Metallurgy of Fusion Welds 167

varies with the carbon content of the steel. It is held at that

temperature long enough for the carbon to distribute its~lf evenly
throughout the austenite. For most practical purposes it is held
at the annealing temperature for 2 ~ minute per mm thickness of
material. The steel is then cooled slowly, preferably in a furnace LJ
or buried in hot ashes or lime so as to cool at a rate of 55°C/hr or •• 600
below. Microstructure of steel obtained with carbon content of QI
'- Sub-critical
Iannealing
0.83% or less is normally grains of pearlite and ferrite. A variant .2500
Cl I
of full annealing called isothermal anneal is sometimes employed. ~400 I
3.6.1.1. Isothermal Annealing E I
~300 I
Full annealing takes considerable time and valuable furnace I
capacity. To overcome these drawbacks isothermal annealing may' 200 I
be employed. In isothermal annealing the weldment is heated to 100
I
and h~ld at the full annealing temperature of 30 to 50°C above I 0.83
o
A3 temperature followed by cooling to and holding in a salt bath o 0·2 0·4 0·6 O.e 1.0 1·2 1.4
at 650°C until the steel attains that temperature and transforms . °/oe_
completely to pearlite and then cooled freely in air. Ia) Temperature range for sub-critical
Annealing temperature range for full and isothermal annealing annealing.
alongwith thermal cycles involved are shown in Fig. 3.27.
3.6.2. Subcritical Annealing
In this process steel is heated to about 650°C; this temperature
is high enough to produce both recrystallisation and uniformity of
structure. After the heating it is cooled freely in air. If a steel with
prior ferrite-pearlite structure is given this treatment .the surface '-'
tension effects around the cementite lamellae are sufficient '- 300 •I 600
700
I to .100I 0200
I-
.2
QIQI
~400
E 500
Cl

contract the latter into rounded carbide grade. When applied to

higher carbon steels this process causes the cementite to assume
a spherical shape that makes it easier to machine, and work it
subsequently. A similar structure is obtained when the steel has F r e e cooling
a prior hardened structure e.g. HAZ, but in this case the carbides
are more finely dispersed. Fig. 3.28 shows the heating and cooling
cycle for subcritical annealing on a time-temperature graph. This
process was formerly called process annealing.
3.6.3. Quench Annealing
Room temperature
It is the treatment used for softening austenitic stainless
steels, including Hadfield manganese steel. In these steels, the Time hours
martensite point is below room temperature, but if the steel is
cooled too slowly from high temperatures, carbides may I bl Time-temperature graph for subcritical annealing.
precipitate and the austenite may transform, at least in part to
ferrite. To soften them, it is theref&re necessary to heat to 1000 to
1100°C, to ensure all the carbides are taken into solution, and then
to air-cool or even water-quench. Fig. 3.28. Temperature range and schematic representation of heating and
cooling cycle employed for sub-critical annealing process.
 169
168 Welding Engineering and Technology Basic Metallurgy of Fusion Welds

3.6.4. Spherodization More recently faster method of spherodization has been

adopted by which the steel is first heated to slightly above the
It is a process by which steels with carbon content more than lower critical temperature (AI) and held there for about two hours.
0.40% are made softer. The steel is heated to just below the lower The heat is lowered relatively fast to a temperature of 625 to
7050C, the cooling rate being as rapid as is possible to attain. The

«I
.5Gl:=
Gl_
8'0
0 •.•
en
-a\
0CL
::l
...
LLIII
-
steel is then held at theEGlGl lower spherodizing temperature,
0E
c:a::
•... Gl

'E I

a;
~rtJ
bJl
d
~8
a;
E
o
d
III
..
.2
..
•...
..c:
11
:>-
Gl '0
E
;= .13
<Il

Fig. 3.29. Typical microstructure of spherodised steel (0.40%C). ],

til
transformation temperature Al (723°C) and held at that e
..
<Il

temperature for a number of hours. This results in the cementite ..c:

.•..
collecting into spherical shaped particles leaving ferritec:>
.•... throughout
the remaining structure as shown in Fig 3.29. Cooling is allowed to .w
......
c:>
N
c:>
-0
a-
Ln
c:>
co
•....
c:>
"'"
c:>
"..,
_- J.' aJn~OJadwal
c:>
1~
take place very slowly throughout the upper part ofthe coolingrange. ~ d
..
~ ca

r r---,..,r-
---j 2 hrs ..
<Il

BOO
/Long hours A ~ ::S
.•..
..
ca
.

\..._----~ q Q)
Co
e
Q)
co
o
,, ,,
~4-15 --J '~', Eo<

...•
~ 500 hrs , ,

,, ,
::l -Ow <':>
~
~<II 400 o~;;:::0 t>D

c..
~ 300 ,, 6 r;:

tS!
,, .w 6
......

200
, c:>
"..,
c:>
100 c:> c:::::t0 C) 0<::;)0 c:> C>
c:> 0000<:)0
••••. -.0 Ln ~ ".., •.••••
C>
ClO

o J.' aJn~oJadwal
Time hours - r
Fig. 3.30. Time-temperature diagram for two types of spherodizil1g processes.
f
170 Welding Engineering and Technology II Basic Metallurgy of Fusion Welds
171

625-705°C, for 4 to 15 hours, depending upon the degree of

spherodization and softness required ...After soaking the steel at
625 to 705°C, it is removed from the furnace or furnace cooled, if
softer and stress-free steel is required. Final hardness depends 01
c:
upon the degree of spherodization and the size of the spheroidal I 'c
cementite particles obtained by this process. Fig ;3.30 shows the '" '" I QI
"C
spherodising temperatures and the time-temperature relationship
c: .5:
I ...
c:I
:i: g'"O I .c
in the two types of spherodising processes. ~ '.2 S I , .c
~ u _ I u
3.6.5. Normalising C"!.2
•• <:1'.D
c:
QI

N ormalising is the process of heating steel 30 to 50°C above

••
~~ a
:::J

i§~ I I ...
o
the upper transformation temperature (A3) for the pa-rticular I I I .•.
composition of steel being normalised to modify the grain size, to Gee III I
I
I
'"
..!!!
u .s. •..
make the microstructure more uniform and usually to improve II II >-
'"
u CU'-
_'1:l
u cu
mechanical properties. It is held at that temperature just
sufficient time for steel to transform to austenite; followed by , QI
E ~
C1>

'<;-5
~El

cooling in still air. In transforming to austenite the pearlite grains i= oo El d

u •••
cu
cu
::l
absorb the ferrite, giving rise to many small grains of austenite C1>
c: ..dO'
........
which, on fairly rapid cooling, transform back to fine-grained 'i:jQl
"C

c: "0

d'"
d
CiE c:I cu
•..
ferrite-pearlite structure. Fig. 3.31 shows the normalising range :J:+=
cu~
on iron-carbon equilibrium diagram and the thermal cycle C1>
c: tlll·-
-
d"O

involved in normalising. :;:::

c:I
'"
....d
QI cu~
Normalising is faster than full annealing and is often used in :J: "'0
.a~
the welding industry to refine any coarse grain structure, to oo 00 oo 0 0 0 C> .c ....
"''<t'
",0
reduce stress after welding or to remove any hard zones in the 0...,., 04 0I"'l
o•... a- -c 0.. ~
ElQi
HAZ. Because of fine-grained structure, the normalised steel has .2;l.2;l
good toughness properties. 4 b.O'"
~ .-d tlll
d
Normalising treatment is particularly applicable to low carbon d'-
N cu d
and some low alloy steels. A typical example of this process is ~ QI.
"0 cu
••• '1:l
'" •..
C1>
normalising of mild steel pipes. The pipes are heated to between ~ --I~ c: ::r:~
900-950°C, this temperature being held for a period of two 6, c:I
... ~J::
"" <)
'd
minutes for pipes up to 100 mm inside diameter, and for five C1> C"?cu
.::l0'
.5
minutes with over 100 mm diameter followed by slow cooling in c:
QI
.!::P
roo..

still air. "C

...
c:I
3.6.6. Quench Hardening :J:
N
In quench hardening steel is heated to 30-50°C above the .w 6 c:I

upper transformation temperature and held at that 0_

oo
(..1.3) I"'l

temperature long enough for the internal structure to transform oo o

to austenite. Overheating and prolonged heating should be o
•...
r- - ~ S
avoided in order to prevent grain growth. Steel is then quenched -- J.' aJn~oJadwal
by immersing it in oil, water or some other cooling media.
Successful quenching converts the austenite to martensite while
slower cooling rates may yield pearlitic structure. The properties
produced will depend upon the carbon content of steel, the
temperature to which it is heated, the time for which it is held at
 173
Basic Metallurgy of Fusion Welds
172 Welding Engineering and Technology
1000
that temperature, and the cooling rate. If carbon, in steel, is less
than 0.15%, no effect of quenching may be observed. If carbon
content is more than 0.83%, no attempt is made to convert the free
cementite to austenite because this constituent is already hard
and the high temperature required to convert it into austenite
would produce a coarse structure which will make steel more t 700
brittle after quenching without any appreciable increase in .w 600
hardness and strength. The hardening capacity of steel increases <II

5500
with the increase of carbon up to about 0.60%. Fig. 3.32 shows the
hardening range for steel and the thermal cycle used for quench ~<II 400
c..
hardening. E 300
When water is used as the quenching medium it is held at· ~
25°C or below and is continuously agitated during the quenching 200
operation to achieve more uniform and faster cooling action. A 5% 100
sodium chloride brine solution provides a more satisfactory cooling
medium for carbon steels; it gives faster and more uniform o
o 0·2 Q.4 0·6 0·8
quenching action and is less affected by increase in temperature. 0/0 ( .-
A 3 to 5% sodium hydroxide quenching bath is also recommended
for carbon steels; it provides even faster cooling rates than sodium (a) Tempering range for steels.
chloride bath.
Oil quenching is resorted to for thin sections of carbon steel
and high alloy steels because of less danger of cracking and High
reduced distortion and quenching stresses. Oil cools steel much
more slowly during the last cooling stage. This is desirable as it
results in much less danger of severe internal stresses, warping
and cracking.
Air cooling is employed for some high alloy steels of the air
hardening type.
3.6.7. Tempering
A quench hardened steel with a structure of martensite is very
brittle. To reduce this hardness steel is given a tempering
treatment in which it is reheated to a temperature of 180 to 650°.0,
depending upon the properties required, followed by air cooling.
This improves toughness and ductility at the expense of hardness Low
o 100 200 300 400 500 600 700
and strength as shown in Fig. 3.33. Sometimes tempering to Tempering temp.;C
within a temperature range of 260-310°C may show a loss of
toughness; that is why this range of temperature is referred to as (b) Effect of tempering on mechanical
Brittle Tempering Range. properties of steel
In general longer the time of treatment at a given tempering Fig. 3.33. Tempering range and the effect of tempering
temperature the better the result from it. It is therefore on mechanical properties of steel.
recommended that at least one hour be allowed at any tempering
temperature for satisfactory results. air tempering furnaces wherein the hot air is circulated around
parts to be tempered. Fig. 3.34 shows a combined hardening and
Tempering of steel is usually accomplished in liquid bath such t
tempering thermll,l cycle as a time-temperature diagram.
as oil, salt or lead. Tempering is also successfully accomplished in
I
 174

Quenching cycle
Welding Engineering and Technology
, ----.-----------------------
Basic Metallurgy of Fusion Welds 175

Hold time Tempering cycle

700
I
Hold time
.~ 600
Quenching
- -o
QI
'-
QI

~ 500 ::l
o
...
~400 I QI
E Q.
~ 300 e
~
Ms

Mf

Time. hrs -

Time
Fig. 3.34. Time-temperature diagram representing a combined
thermal cycle. for quenching and tempering.
Fig. 3.35. Schematic diagram of austempering heat treatment cycle superimposed
3.6.8. Secondary Hardness on isothermal transformation diagram for a medium carbon steel.

In alloy steels which contain certain carbide forming elements 3.6.9. Austempering
such as tungsten, molybdenum, vanadium, etc., tempering after If steel from hardening temperature is supercooled quickly to
hardening produces precipitated carbides in such finely dispersed about 290°C, austenite at this temperature transforms to a fine
form that the hardness is considerably greater than after the pearlitic or bainite structure of uniform hardness of about
original hardening. At the same time, the breakdown of the Rockwell, Rc56. It requires the holding of austenite at 290°C for
martensite results in increased toughness. Thus, for example, high
speed steel reaches its maximum hardness only after hardening
and tempering to about 550°C, by what is known as Secondary
I about one hour to complete this change. This method of tempering
without the formation of martensite is called austempering i.e.,
the direct tempering of austenite. This treatment is also
Hardening Effect. Steels of this type are also, in consequence, sometimes referred to as Hot Quenching. Fig. 3.35 shows a
resistant to softening at quite high temperatures and, therefore, thermal cycle for this process.
are suitable for high temperat.ure service applications.
It is accepted that the hardness of R,,56 obtained by austem-
For such steels a double tempering treatment is usually pering is much tougher than the same steel treated to the same
adopted, the steel being air-cooled between the two operations. hardness by the usual method of quench hardening and tempering.
The advantage of double tempering accrues from the fact that, Also, nOll formation of martensite eliminates much of the danger
after hardening, the steels normally contain a certain amount of of cracking, and reduces the amount of distortion or warping caused
retained (i.e. untransformed) austenite, which transforms to by rapid quenching to room temperature required for the formation
martensite, wholly or in part, after the first tempering treatment. of martensite in normal quench hardening process.
The second tempering treatment breaks down this newly formed 3.6.10. Martempering
martensite and brings about additional secondary hardening. Martempering is carried out by cooling the steel from the
Besides increasing the hardness, this practically eliminates the hardening temperature through the pearlite range to a
presence of untempered martensite.
176 Welding Engineering and Technology 177
Basic Metallurgy of Fusion Welds

3.6.11. Stress Relieving

Stress relieving is carried out after welding to remove or
reduce welding stresses. Total elimination of residual stresses
after welding is possible only by annealing. However, that leads
to excessive softening and reduction in strength therefore
stress-relieving is adopted in which the welded (steel) component
QI
is uniformly heated to below the recrystallisation temperature
...
::I
+-
e
c.
QI

E
900
~ Ms

'700 -.L - - - -
--------
Mf LJ
BooL. Al (723'( )
.i 600~
ECl 500
Time
~ 400
Fig. 3.36. Schematic diagram ofmartempering heat tJeatment cycle E
superimposed on an isothermal transformadon ~ Slow heating
diagram for a medium carbon steel.

temperature at or a little above the Ms temperature (205-260°C),

holding at this temperature until the temperature of steel is
uniform throughout and finally air cooling through the martensite
range. Fig. 3.36 shows a thermal cycle for martempering
treatment. The aim of martempering is to avoid the risk of
cracking due to the thermal stresses set up during rapid cooling
from high temperatures. This also makes it possible to cool slowly
through the martensite range, thus minimising the additional
stresses set up as a result of the volume change which
accompanies the transformation of austenite to martensite
thereby constituting a risk of cracking.
It is a condition of successful martempering that no trans-
formation shall occur during the initial cooling. Ordinary carbon
steel, for example, is hardly suitable since it undergoes rapid
transformation at all temperatures between the pearlite and
martensite ranges. High alloy steels on the other hand, can be
held for long periods at intermediate temperatures without the
rlsk of transformation as is evident from the TTT diagram of Fig.
2.29.
Martempering treatment is also sometimes referred to as step
hardening.
Time -
Fig. 3.37. A schematic thermal cycle for stress-relieving
treatment of low carbon steel components.
(A1)-normally to 550-650°C for mild steel and held there for a
period of 2 ~ minute/mm thickness of the component depending
upon the thickest section present. This is followed by slow cooling,
preferably in a furnace, atleast up to a temperature of about
4000C. Fig . .3.37 shows the thermal cycle for stress relieving
treatment.
Examination of stress relieved components shows that
recrystallisation of the ferrite takes place which contributes to the
relaxation of residual stresses of the welded joint.
Sometimes low temperature stress-relieving is carried out at
about 450°C for a longer period of time. At this lower temperature
very little, if any, recrystallisation of the ferrite takes place. If the
welded component is held at 450°C for one hour, and slowly cooled
 178 Welding Engineering and Technology Basic Metallurgy of Fusion Welds 179

r the excess B atoms will be trapped in the solid solution and the
thereafter, about 40% of the stresses will be relieved, that may be
, regarded as adequate for some purposes.
Further, it may be known that any gases trapped in the cooling
weld may be released from the welded component during the
stress relieving treatment.
Stress relieving is frequently adopted, among other things, for
welded pressure vessels. Usually by this treatment residual
stresses are decreased to a tolerable level without any marked
decrease in mechanical strength or change in microstructure.
However, such a treatment rarely makes the structure completely
free from residual stresses.
3.6.12. Solution Treatment
In some alloys soiid solubility is gradually increased by
heating; Fig. 3.38 shows a part of the equilibrium diagram for a
system of this type. The alloy of composition 'X' will, if slowly
alloy will be in a condition of non-equilibrium. At this stage the
excess B atoms will be scattered throughout the structure and the
overall effect will be to produce improvement in corrosion
resistance and increase in tensile strength with good ductility.
Heating and quenching to produce this effect is called s,olution
treatment, and is the first step towards increasing the strength of
the alloy. Cold working may be carried out after solution
treatment, thereby increasing the strength without any effect
upon the corrosion resistance of the solid solution alloy. Such a
treatment is popularly applicable to some aluminium alloys e.g.,
AI-Cu alloys.
3.6.13. Precipitation Hardening or Age Hardening
If a solution treated alloy is reheated but to a lower
temperature than that required for solution treatment, the atoms
of metal B will have enough energy to start their precipitation

--t
from the solid solution so that the equilibrium is' restored.

soak

QI
L- S!!.oU -"0E.~
:::J
.•..
dL-
W-Sho,t
I .
\
QI
a. Solid I .\
E
QI
~ +I Metal B Quencn .
soh~tion f'

cooling Heat t Quench

Heat
Free

_ ~Slow
~~-~:~ ~ ,rh ~
Solution
Room temp. ...'. '
', ~::
~'......
•.

treatment
A °loB -
Alloy x
Equilibrium diagram
Fig. 3.38. Equilibrium diagram for a solid-solution aHoy and the
process details for its solution treatment.
cooled from its molten state, consist of solid solution of metal B in
metal A, in which are embedded particles of metal B. When this
alloy is reheated to a high enough temperature, the excess metal
B will be taken into the solid solution. Now,ifit is quenched there
will be insufficient time for precipitation of metal B to occur, and
~
Solution Precipitation hardening
treatment
Fig. 3.39. Schematic representation of solution treatment
and precipitation hardening processes.
Provided the temperature is not too high they will not leave the
solid solution but form regions within it where the B population
is higher and so locally strengthen the solid solution. The alloy is
then given a long soaking time, followed by free cooling. Such a
heat treatment is known as Precipitation Hardening or Age
Hardening or simply Ageing. Fig 3.39 shows the thermal cycle for
solution treatment cum precipitation hardening.
 Basic Metallurgy of Fusion Welds 181
180 Welding Engineering and Technology
The overall effect of precipitation hardening treatment is to
produce increased strength and hardness at the expense of
ductility resulting in lower impact values as shown in Fig. 3.40.
Consequently an ageing-resistant metal, for example steel, should
have a stable transition, even ifit is subjected to ageing treatment.
QI Hardening of an alloy by ageing is said to be due to strain in
::l
C:> the crystal structure set up by t.he migration of B atoms (solute)
- •...•
Cl
to for,m local clusters. Most heat treatment alloys are heat treated
by solution treatment followed by precipitation treatment and a
Q. few among these are of naturally ageing type.
.s
If precipitation treatment is continued for too long, the local
aggregation of atoms results in the formation of separate particles
with a crystal structure differing from the matrix. The local strain
in the crystals is thereby relieved and the hardness of the alloy is
decreased and it is said to be overaged. For a given alloy, the higher
the precipitation treatment temperature the sooner the optimum
conditions are reached.
Testing temperature -
Fig. 3.40. Effect of ageing treatment on impact values.
Welding of age-hardened aluminium alloys results in a
Columnar softened zone alongside the weld due to the overageing effect as
shown in Fig. 3.41.
Original 3.6.14. Natural Ageing
grain size
8.dispersion In the case of some alloys, precipitation after solution
of precipitated treatment occurs at room temperature, starting very soon after
particles quenching and it takes about 4 days. This process of precipitation
hardening is called natural hardening. Duralumin (AI+4%Cu.)is
Elong atced grain a typical natural ageing alloy.
ifit is cold Amongst steels mild steel is the most susceptible to ageing. If
rolled aluminium nitrogen is present in steel, iron nitride can be precipitated at
Region of aVERAGING
inwhich the coalescence temperatures below AI' Precipitation of iron nitrides (FeI6NZ)at
of partiClesproduces room temperature is known as Steel Ageing. Ageing can take place
lowest hardness
Regionof grain growth in a zone heated to temperatures around 200-300°C if free
t with gran size incnzasing nitrogen is present in steel.
lJI
lJI with increasing temperature
•••
c: New metallurgical procedures have helped in lowering the
'0
... nitrogen content in steel, or binding it to a stable nitride phase
Cl
:z: (e.g. AIN), and consequently present-day steels are generally not
Highest
susceptible to ageing.
unaffected Precipitates arising in the austenitic zone slow down growth
by heat ofy-grains and thus make steel fine-grained. In microalloyed steels
'0 nitrides and carbonitrides of alloying elements (Ti, Nb, V, AI, etc;)
!i
o
'.::I'
Lowest hardness
chemically bind nitrogen, thereby making the weld more resistaIl;t
to ageing.
Fig. 3.41. Effect of overaging on hardness of aluminium alloy weld.
(Aftnr l(pnvnn',
182 Welding Engineering and Techn()Iogy
Precipitation hardening in austenitic stainless steels. occurs
, Basic Metallurgy of Fusion Welds
through formation of an inter metallic phase with AI, Ti and P.
These steels have corrosion resistance comparable to that of other
stainless steels. Precipitation-hardened austenitic stainless steels
become brittle after being exposed for several thousand hours to
temperatures above 290°C. Their usability is therefore limited to
a temperature below about 320°C in long term operations. In
short-term service, however, these steels possess adequate
strength at temperatures up to 500°C and corrosion resistance up
to about 820°C.
Body centred metals are affected similarly by bombardment
with massive fast particles. 'l'he bombaradment produces atomic
displacement in the lattice, causing important changes in physical
properties. One of these changes in physical properties is marked
rise in transition temperature. This effect is termed irridiation
damage or irradiation ageing. An important resemblance to the
normal ageing is the annealing of the damage so that physical
properties revert to their former values.
3.6.15. Strain Age Embrittlement
The normal precipitation hardening or Ageing process may also
take the form of strain ageing in the first stage of which nitrogen
atoms segregate at dislocation lines, thereby slowing down the
dislocation mobility. In the second stage nitrides are precipitated
at dislocation lines, ~d in the third stage precipitation occurs
throughout· the entire volume. This results in reduced charpy
energy and thus lower fracture toughness. That is why thi!'jprocess
is also referred to as strain-age embrittlement.
Strain ageing is a possibility where the metal has been
strained and heated within the temperature range of lOO-300°C.
This phenomenon is associated with a fine precipitation or
clustering located at dislocation within the ferrite grains which
are observable under favourable conditions by transmission
electron microscopy (TEM).
Nitrogen in a carbon or C-Mn steel may be responsible for
strain-age embrittlement. If a nitrogen-bearing steel is subjected
to plastic strain and simultaneously or subsequently heated at a
temperature of about 200°C, notch ductility is reduced. Steel
produced by Bessemer process normally contains high amount of
N2 that is' why many bridges made with such steels, prior to
Second World War, failed by brittle fracture, and these. failures
were attributed to ~~rain-age embrittlement.
The known association of strain-age embrittlement with
nitr()gen suggests that the particles are nitrides or carbonitrides.
183
Their effect is to shift the charpy-v impact transition temperature
upwards by about 50°C.
Also, it has been proved that if a small crack forms adjacent
to the weld boundary and is then strained and reheated by
successive weld passes, the tip of the crack can be sufficiently
embrittled to initiate a brittle fracture. The root passes of a
multi-pass weld are similarly strained and reheated and this may
also lead to embrittIemeilt.
Austenitic Cr-Ni steels may embrittle at elevated temperature
due to a strain ageing mechanism during post weld heat treatment
(PWHT) or in service.
Strain-age embrittlement may be prevented by the addition of
strong nitride forming elements such as Al or Ti but this remedy
is normally not possible for arc welding electrodes because Al and
Ti burn out in the arc. However, in self-shielded welding, large
amounts crfaluminium and magnesium are often added to the flux
core to provide a vapour shield, and the amount of aluminium in
the weld deposit may be sufficient to prevent embrittlement of the
root passes.
3.6.16. Temper Embrittlement
Ferritic alloy steels may be embrittled by the formation of
unfavourable transformation products, through grain boundary
segregation, called temper embrittlement. If an alloy steel is held
for a period of time within the temperature range 375-575°G or,
in the case of more susceptible compositions, is cooled slowlX
through this range, it may lead to unwanted increase in the impact
transition temperature. The susceptibility of a steel to temper
embrittlement is normally measured by the temperature shift of
either the 55 J impact energy or the 50% fracture appearance
transition temperature (FATT).
Plain chromium steels are highly susceptible to temper embrit-
tlement, but the addition of 0.5% Mogreatly reduces this tendency.
Elements that markedly increase the degree of embrittlement are
Sb, Sn, P, and As, while Mn, V, B, and Si increase susceptibility
slightly. Weld metal of the same composition as the alloy plate is
also subject to temper embrittlement. The effect is reversible, and
steel may be de-embrittled by heating at temperatures of GOO°C
and above. It is caused by segregation of tramp elements to the .
prior austenite grain boundaries, and fractures are typically inter-
granular.
Embrittling and alloying elements all act together, for example
Mn and Si increase the effects of P and Sn. For Q & T, 2.25Cr,....l
184 185
Welding Engineering and Technology Basic Metallurgy of Fusion Welds

140 because on cooling from operating temperature they contain

supersaturated hydrogen, which is a further embrittling factor.
3.6.17. Graphitisation
Under non-corrosive conditions C and C-Mn steel welds may
lose strength due to the phenomenon called graphitisation. This
.<l t-
<t
u..
l-0
80
100
60 happens if steel is held at temperatures above 450°C for long
periods of time, resulting in the decomposition of the cementite,
eventually forming nodules of graphite in a low carbon iron
w 0
20 40 matrix. The breakdown of cementite appears to be promoted by
the addition of aluminium to the steel. Carbon and C- ~Mo steels
are both susceptible to this defect; however, in carbon steel the
graphite nodules tend to be scattered and do not form localised
1
areas of weakness. In C- 2'Mo steel, on the other hand, the
graphite forms eyebrows in the HAZ, and there have been isolated
failures in steam lines from this cause. Cr-Mo steels do not suffer
graphitisation since the carbides in it are more stable. For this
reason. Cr-Mo material is often specified for service at elevated
temperature, particularly for piping, where in other respects
C-- ~MO steel would be adequate.
Fig. 3.42. Correlation between the embrittlement susceptibility factor J and the
fracture appearance transition temperature for long-term isothermal heating. Grain boundary
carbide5.
Mo, steel with Arsenic below 0.02% and Sb < 0.004%,' the
susceptibility to embrittlement may be measured QY what is
known as J factor, where J is given by the following equation .
.J = (Mn + 8i) (P + Sn) x 104 ••• (3.7) I
I
"
Fig. 3.42 shows the correlation between the .J factor and the
fracture appearance transition temperature for long term isother-
~nal heating. The J factor can be reduced by lowering the Si-con-
tent and controlling P, As, Sn, and Sb. The Mn level, however,
cannot be reduced without sacrificing tensile strength. \ i
Temper embrittlement is of concern in the operation of heavy- Bulk Cr
level
walled pressure vessels for nuclear power and petro-chemical
plants. In case cracking is found during an inspection of the shell, r- I I
then in order to assess the integrity of the vessel it is .imperative
to know the fracture toughness. From the initial properties aIicl
the service life and using a correlation such as that shown in Fig.
3..42, it may be possible to calculate the fracture risk. Vessels used
for hydrogenation such as hydrocrackers are especially vulnerable
Cr level
around
carbides ~---'¥ Distance -
Y{
Fig. 3.43. Schematic illustration of the effect of
sensitisation on local chromium level.

!
 186
Welding Engineering and Technology
Basic Metallurgy of Fusion Welds 187
3.6.18. Carbide Precipitation and Sensitization
indicating that standard 18% Cr steel should contain no more than
If an unstabilised Cr-Ni austenitic stainless steel is welded or 0.015% C.
brazed it experiences chromium carbide (Cr23C6)precipitation at
the grain boundaries, Fig. 3.43, resulting in loss of corrosion The microstructure of the steel can affect the degree of
resistance, due to depletion of chromium, in the affected zone. This sensitization. For example, the presence of &-ferrite lowers the
sensitization temperature for austenitic stainless steel. While
tendency towards intercrystalline carbide precipitation is also fully austenitic stainless steel containing 0.05% C has a maximum
known as Sensitization and is affected by different factors sensitization temperature of 750°C, it falls to around 500°C for
including material composition and thickness, heating time, pwht steel with 25% delta ferrite. This is because the Cr- content of the
(post weld heat treatment), type of microstructure, and the ferrite is markedly higher than that ofthe austenite in equilibrium
influence of stabilizing elements. The ill effect caused by the with it. Thus the Cr-rich o-ferrite has a higher resistance to grain
carbide precipitation is referred to as weld decay. boundary attack.
900l-carbon content 1%) The microstructure of stainless steel becomes sensitive to in-
0·80 /'0.062 0.058 0·056
t.ercrystalline corrosion in certain media, when the weld or the HAZ
remains longer in the critical temperature range of 425 to 815°C.
'11 Within these limits of temperature chromium carbide, Cr23C6'can
.'-'
I precipitate at the grain boundaries. For 18/8 Cr-Ni steel with a
\ carbon content of about 0.07%, a dwell time of 6 minutes at 700°C
0)-
<-
700 \ is sufficient to induce sensitization (susceptibility to intercrystal-
o=>

<-
0)
line corrosion). Therefore, if such steels of higher thickness are to
~ 600~-
be welded, they must contain lower carbon contents, with longer
~0) sensitization time. Given in Table 3.1 are maximum plate thick-
nesses with permissible carbon contents, for steel of type 18/8
500 Cr-Ni, which are not prone to sensitization by welding.
Table 3.1. Penn issible Carbon Contents for Different
40-0 Thicknesses of 18/8 Cr-Ni Steels
10s 1min 10min 1h 10h 100h 1000h
Time
Permissible Carbon content
___ (%)
Fig. 3.44. Time-temperature relationships (C-curves) for .formation of harmful s. NO'l .. Plate (mm)
Thickness
amounts of Cr-carbides in stainless steel with various carbon contents. 1 ~1 0.18
2 1--3 0.08
The major factor cadsing carbide precipitation is the carbon 3 3--12 0.06

_L
content of steel. Steels containing 0.08% or less carbon may suffer 4 12--19 0.05
carbide precipitation in thick sections, but not when the thickness
5 19--25 0.04
is 3 mm or less. Steels containing 0.03% or less carbon and those "
stabilised by the addition of Ti or Nb are immune to weld decay 6 > 25 0.03
in all thicknesses. The time-temperature-sensitisation (TTS) or , '.'/

C-curves for various grades of Cr-Ni steels are shown in Fig. 3.44.
For effective control of carbide precipitation in fully austenitic
stainless steel with around 10% Ni, to avoid grain boundary
corrosion, the relationship between maximum Cr and C contents
is given by,
Cr ~ 80C + 16.8 ...(3.8)
The problem of sensitization of weld metal or the HAZ may be
overcome by PWHT, by annealing at 1050-1100°C followed by
cooling in water or air. This results in dissolving the chromium
carbides. The so-called stabilization anneal at a temperature of
900 to 950°C does not solve the problem as it leads to shifting the
sensitization zone; even with low carbon content.

I
,.

188 Welding Engineering

Since PWHT of welded joints is inconvenient and often even

and Technology
I Basic Metallurgy of Fusion Welds

and size when the material is in annealed state as shown in Fig

189

impossible, it is better to utilize austenitic stainless steel with an

extra low carbon content to overcome the problem of sensitization. 3.45. Heating and fusion influence the size of these grains. Ferrite
is soft, weak and ductile, and magnetic up to curie point of iron
3.7. Microstructural Products in Weldments i.e., 770°C.
Welding processes and the associated phenomena, some of Cementite (FeaC) : It is a chemical compound of iron and
which have been described in the previous section, lead to the carbon which contains 6.67% carbon and is formed around 1500°C.
development of different microstructural products in the weld It is the hardest constituent (1200 VHN) found in iron-carbon
metal and the HAZ. alloys.
The mechanical performance of the weldment particularly Cementite can dissolve a wide variety of elements such as Mn
with reference to its strength and toughness will depend upon the and Cr, to give Complex Carbides. It loses its magnetism at 210cC.
type of microstructure obtained in weld metal and the HAZ. Cementite is hard and brittle.
Because steels are the most welded material therefore most of Cementite can be present as free cementite or laminated with
the microstructural products discussed in this section relate to ferrite to produce a structure called pearlite. Free cementite can
steels.
be of two types viz., primary cementite and secondary cementite.
The two basic constituents of steel obtained on slow cooling to
Too:m temperature are ferrite and cementite.
Ferrite or a-iron is almost pure iron but Carbon can be
dissolved in it to form a solid solution ferrite which can contain
up to 0.006% C at room temperature and that can increase to
0.05% at 750°C. But ferrite is capable of dissolving a large number
of second elt:.lllents other than carbon to form a solid-solution.
Etched by suitable reagent (e.g. nitaD, ferrite can be seen
under the microscope as polyhedral grains, fairly regular in shape

Fig. 3.46. Ni-bearing cast iron showing primary cementite in

an austenitic matrix with some martensite.

Primary Cementite solidifies directly from the melt, and occurs

in the form of hard regions or needles (white iron) as shown in
Fig. 3.46 while it outlines the grai.n structure in hypereutectoid (>
Fig. 3.45. Microstructure of ferrite. 0.80%C) steels as shown in Fig. 3.47.
190 Welding Engineering and Technology
r
Fig. 3.47. Structure of 1% carbon cementite network
and pearlite stained by nital etching.
Secondary Cementite precipitates in the solid, at low carbon
contents, at the lower critical temperature (Al) and it occurs along
the ferrite grain boundaries. This type of cementite is frequently
encountered in weld metal with very low carbon content.
Pro-eutectoid cementite also precipitates from solid-state
supersaturated austenite, and at etuectoid point (0.80% C, 723°C)
to give a completely pearlitic structure.
Although ferrite and cementite are the two basic constituents
of steel but in weldments, depending on the composition and
cooling rates and the pwht employed (if any), they are found to
exist in different allotropic forms and morphologies. Accordingly
they influence the mechanical behaviour (strength, toughness,
etc.) of the welded joint.
Different microstructural phases encountered in fusion weld-
ing of steel can be grouped into three types viz., the primary
microstruct.ure, the secondary microstructure, and the tertiary
microstructure.
Primary Microstructure is the one obtained by direct
solidification of the molten metal and usually results in a cast
st.ructure with columnar grains. The length axis of these grains is
nearly parallel to the direction ofthe temperature gradient during
solidification. Sometimes a dendritic sub-structure can be observed
, Basic Metallurgy of F'usion Welds 191
inside the columnar grains. In metals without phase transforma-
tion in the solid-state e.g., aluminium, the primary structure is
usually retained on cooling to room temperature, although changes
in sub-structures may occur during cooling. However, conditions
are more complicated in unalloyed and low alloy steels because of
the occurrence of phase transformation in the solid-state.
'!\vo major primary microstructural products encountered in
carbon steel welds are delta (0) -ferrite and austenite or y-iron.
Secondary Microstructure is the one obtained by trans-
formation of solid phase austenite into different other solid phases
like pearlite, bainite, martensite, grain boundary ferrite, ferrite
side plates, lath ferrite, acicular ferrite, polygonal ferrite, and
retained austensite.
Tertiary Microstructure is encountered only in multi-run
welds where, due to the successive runs, the previous runs are
heated to austenite region and thus result in products of different
fineness affecting the mechanical strength and toughness
properties due to the normalising effect of heating by successive
runs.
Apart from these three categories of microstructural products
the PWHT, if employed, will affect the microstructure. For
example, tempering results in the formation of tempered
martensite with better toughness properties than those of
untempered martensite.
In high Cr-Ni stainless steels another important micro-
structural phase encountered is what is called the sigma (0) phase.
To sum up it can be said that some or other phase combi-
nation(s) of the following microstructures may be encountered in
weld metal and HAZ of steel welds.
1. Delt.a ferrite,
2. Austenite,
3. Grain boundary ferrite,
4. Ferrite side plates,
5. Lath ferrite,
6. Polygonal ferrite,
7. Acicular ferrite,
8. Pearlite or Ferrite-Carbide aggregate,
9. Bainite,
10. Martensite,
11. Retained austenite.
Brief description ~f all these microstructural phases follows.

192 Welding Engineering
3.7.1. Delta Ferrite
Delta ferrite nucleates from the melt if steel contains less than
O.53%C ; however, this phase is stable only above l390°C. Thus,
with the lowering of temperature it changes to austenite. Some
~-ferrite may be retained in microstructure because the change
(L + 3 ~ y), if it takes place through the peritectic reaction is very
slow compared with precipitation of "(-iron from a liquid. The
reason is that a solid phase 3 has to react with liquid to form a
new phase y. The new phase will coat the remaining 0 as shown
in Fig. 3.48. For the reaction to continue, diffusion through the
Liquid metal Liquid
Peritectic --- ------
Reaction := '0
\V- ----
--------
-~-.:. ---
:_-=
.:: ~
..:-
Fig, 3.48, Schematic representation of peritectic reaction
for change of o-iron to y-iron.
solid y is required, which is a relatively slow process thus there is
every possibility of some delta ferrite being retained, till room
temperature, in the weld metal zone. 'I'hi::;may not be of much
consequence in plain carbon or low alloy steel welds but is of
considerable significance in stainless steel welds and will
therefore be discussed later in the section alongwith sigma phase.
3.7.2. Austenite
Austenite or gamma (y) iron is stable at high temperatures
between 1390°C and 910°C. It crystallizes in the FCC cell
structure and is thus more closely packed than the ferrite cell of
BeC structure. The cast austenitic structure in the weld metal
zone is usually regarded as the primary microstructure which is
maintained during cooling until the y ~ ex transition temperature
is reached.
Gamma iron dissolves up to 1.7%C by weight giving rise to a
wide range of y-solid solutions known as austenite, Austeni~e is
non-magnetic and relatively weak. In sl{)vvlycooled steels it cannot
exist at temperatures below 695°C and is fully decomposed when
Basic Metallurgy of Fusion Welds 193
and Technology
the temperature of a plain carbon steel falls below that tempera-
ture.
3.7.3. Grain Boundary Ferrite
During cooling of the weldment, on reaching the y ~ ex
transformation stage, one of the secondary transformation
products is grain boundary ferrite, which marks the original
Fig. 3.49. Grain boundary ferrite enclosing the original austenite grains,
austenitic grains. The primary microstructure therefore remains
recongnisable in as deposited weld metal, particularly at moderate
magnification of up to 100X, as shown in Fig. 3.49.
Grain boundary ferrite (GBF) is often referred to as
pro-eutectoid ferrite. However, this nomenclature is misleading
because other constituents (e.g., ferrite side plates) are also
originally pro-eutectoid phases.
GBF is normally considered detrimental to toughness, because
of its coarse grain structure, compared with acicular ferrite. It has
been reported that cracks in ferritic weld metal normally
propagate along GBF, a process which is intensified by the
presence of brittle pearlitic structures along grain boundaries.
GBF is also sometime called blocky fern'te, and ferrite veins.
3.7.4. Ferrite Side Plates
In a cross sectional view ferrite side plates (FSP) appear as
long needles with aspect (length/width) ratio of at least 20 : 1. This
microstructural constituent grows from grain boundary ferrite
into the original austenite grain as packets of parallel plates, as
'=".:,....:~''''o''',·,,=,.·,:O!:;-:~ ...·,. ~ -= .... _..,
194 Welding Engineering and Technology
Fig. 3.50. Grain side plates structure often growing from former austenite
grain boundaries with plates separated b~' M-A-C aggregates.
shown in Fig. 3.50. In addition, isolated packets of ferrite side
plates can sometimes be seen inside the original austenite grains.
Further polishing of the specimen will, however, reveal that these
i.solated packets are attached to grain boundary ferrite. It is also
observed that the ferrite side plates have a real plate-like
structure, rather than needle or lath like, and are separated from
each other by low angle grain boundaries. IIW (International
Institute of Welding), based on the proposal of Abson and Dolby,
has suggested the nomenclature for ferrite side plates as ferrite
with aligned M-A-C (martensite, austenite, and / or carbide) This
new nomenclature indicates more clearly the composition of this
microstructural constituent, but does not say anything about its
morphology. Ferrite side plates are also sometimes referred to as
Widmanstatten Ferrite.
Ferrite side plates, like grain boundary ferrite, are also
detrimental to toughness. This is linked with the possible presence
of precipitated carbides, retained austenite and martensite along
the ferrite plates. Moreover the ferrite plates are separated by low
angle grain boundaries, causing a much larger effective grain size.
Basic Metallurgy of Fusion Welds 195
3.7.5. Lath Ferrite
It is an intragranular (i.e. within grains) austenite
transformation product resembling lower bainite. It is found
amongst acicular ferrite or side plate structures. Unlike ferrite
side plates, lath ferrite is not attached to grain boundary ferrite.
Lath ferri~e is also detrimental to toughness on the similar
considerations as the ferrite side plates.
3.7.6. Polygonal Ferrite
Ploygonal ferrite occurs in the form of coarse ferrite islands
inside the prior austenite grains as shown in Fig. 3.51. Polygonal
Fig. 3.51. Polygonal ferrite (PF) in the form of an
island surrounded by acicular ferrite.
ferrite like grain boundary ferrite is detrimental to toughness
because of its coarse grain size.
3.7.7. Acicular Ferrite
Acicular ferrite is formed in the interior of the original
austenite grains by direct nucleation from the inclusions resulting
in randomly oriented short ferrite needles with a basket weave
feature as shown in Fig. 3.52. This interlocking nature together
with its fine grain size (0.5 to 5 j.lm)with aspect ratio from 3 : 1
to 10 : 1, provides the maximum resistance to crack propagation
by cleavage. Acicular ferrite is also characterised by high angle
boundaries betwlilen the ferrite grains. This furthlilr reduclils the
chapce of cleavage because these boundaries impede .crack
propagation. It is reported that nucleation of various ferrite
morphologies is aided by non-metallic inclusions; in particular
196 Welding ~.!~~ee.r:ing~~d!e~~ology
Fig. 3.52. Microstructure of acicular ferrite.
oxygen-rich inclusions of a certain type and size are associated
with the intragranular formation of acicular ferrite.
It can be said that acicular ferrite is a fine Widmanstiitten
constituent which is nucleated by an optimum intragranular-
dispersion of oxide-sulphide-silicate particles.
Fig. 3.53. Microstructure of lamellar pearlite.
Basic Metallurgy of Fusion Welds 197
For the various advantages of acicular ferrite microstructure
it is aimed to increase its presence in the weld metal. It is found
that composition control of the weld metal is necessary in order
to maximise the volum(l fraction of ackular ferrite, because
excessive alloying elements can cause the formation ofbainit.e and
martensite.
3.7.8. Pearlite
A microstructure formed with alternate lamellar layers of
cementite and ferrite, Fig 3.53, is called pearlite. It is so named
because when etched, and viewed with the naked eye, it has the
appearance of 'mother of pearl'. Pearlite combines the good
properties of ferrite and cementite. Pearlite grain size is defined
by the distance So between the ceillentite lamenae, which varies
with the heat treatment conditions.
For a ~oarse pearlite the interlamellar dist.ance So can reach
500 to 1000 microns; for a normal pearlite it is around 300 J..1m,
whilst when So is less than .200 J..1m,
it is considered fine pearlite.
A fine pearlite may be 2-3 times as hard as a coarse pearlite.
Lamellar pearlite can take on the spheroidal form after
prolonged heating near Al (723°C); it is'then said to have
coalesced; this is accompanied by a drop in hardness.
Pearlite adversely influences the toughness of ferritic weld
metal, particularly when brittle cementite lamellae are lying in
the crack direction. It appears that the ductile-brittle transition
temperature decreases, more or less linearly, with an incr(~asing
amount of pearlite. Reduction of inter-lamellar spacing is
detrimental, whereas decrease in thickness of the cementite
lamellae is beneficial to toughness because thinner carbides are
more likely to deform than to crack upon stress.
3.7.9. Bainite
Depending upon the cooling rate austenite may transforin into
pearlite, bainite or martensite as shown by the S-curves of Fig.
3.54. The interme4iate transformation (Ar"), with fast and very
fast cooling rates, results in the formation of a structure called
Bainite. While fast cooling rate is said to give coarse or upper
bainite; very fast cooling rates result in fine or lower bainite.
The intermedi~t~ transformation mechanism results in the
formation of bainite if the structure is acicular with continuous
cooling rates of 1000-2000c·C per minute; if the structure is
granular then it results in the formation of pro-bainitic ferrite.
 198
Welding Engineering and Techn~logy Basic MetaUurgy of Fusion Welds 199
r
3.7.9.2. Lower Bainite
S Curve Microstructure may be likened to a ferrite containing small
Ac, carbide grains, aligned parallel to each other and at 60° to the
ferrite axis, as shown in Fig. 3.56.
Ar'

-o -
'-
OJ

::J

<-
'c:
OJ

OJ
Pearlite + bainite
OJ
C.
_
VI Aru
E :::I
~OJ 0

Ms

Ar'"
- .

Martensitic reaction

o
Time -

Fig. 3.54. Ranges of temperature occupied by three type

(Ar', Ar", Arm) of transformations.
3.7.9.1. Upper Bainite
Microstructure consists of ferrite containing carbides aligned Fig. 3.56. Microstructure of lower bainite.
along the ferrite axis, as shown in Fig. 3.55.
The acicular structures may be either upper or lower bainite,
and are found mainly in low alloy steels. Metallographically they
I

I
I.<'ig.3.57. Microstructure of upper bainite, laths of bainite with
(a) cementite, and (b) retained austenite.
Fig. 3.55. Microstructure of upper bainite.
"~",." ••••••••••••..•• ,_ •••• , -_",,,.,,-._, ••.
c~,••
,,,~•• ,•••."••,,,,.. ~••.••
_ •••.
200
appear as Widmanstatten structures, with a grain boundary
outline of carbides or stabilized austenite.
Apart from the above mentioned types of bainitic structures,
bainite may be found in some other forms as well. For example,
upper bainite may exist as,
(i) alternating platelets of pure ferrite and austenite;
(ii) alternating platelets of pure ferrite and ferritic areas
containing variable amounts of carbides.
Also, when long and aligned ferrite laths are found, Fig. 3.57,
with an aspect ratio of approximately 10 : 1 to 12 : 1 with carbide
precipitation between the platelets the structure is taken as upper
bainite.
Lower bainite in eutectoid steel is made up of fine carbides
whilst on the other hand in low carbon steels only the outlines of
the ferrite are discernable.
Fig. 3.58. Needle-like microstructure
-_" .~_~
Welding Engineering and Technology I Basic Metallurgy of Fusion Welds 201
r
3.7.10. Martensite and Retained Austenite
If the steel is quenched <Cooledvery rapidly) from the austenite
condition, the excess carbon will be trapped within the lattice to
form the supersaturated solid solution of carbon in a-iron called
'martensite. Martensite is, thus, a super saturated solution of
carbon in ferrite formed by a mechanism involving neither
nucleation nor growth.
Martensite is a fine, needlike structure, Fig. 3.58, which, due
to the enormous supersaturation that causes a distortion of the
cubic lattice, is very strong and hard, but very brittle. The
martensite needles crystallize along the cleavage planes and show
the three directions of the Widmanstatten structure.
Martensite crystallises in the tetragonal system, with a body
centred unit cell very close to a cube.
When austenite decomposes at very low temperature
« 260°C), the y ~ a transformation still takes place, but the
breakdown rate is too high to allow the carbides to precipitate, so
that the carbon remains in solution in a-iron (ferrite), although
the solubility of ferrite for carbon is extremely low.For this reason,
the solution is supersaturated and unstable.
Martensite is the hardest constituent of steel. Part of this
hardness is attributed to the presence of carbon in supersaturated
solid solution; it is however mainly due to the distortion of the
Fig. 3.59. Microstructure showing retained austenite in a matrix
of martensite. of ferrite with aligned martensite.
--;-;;;;;;r;;iiJliiIliiii!i!ilii!i!ll!iiil!llllll"!!i .• !!!.-"".!i!,,!!!,"l!l!
•.. !!!
.• I,~!!!.!!!
•.. i!!_Ii._Ii!!!.
202
tetragonal lattice, which increases with the increase in carbon
content.
Although the phenomenon of change from austenite to
martensite does not involve nucleation and growth but still takes
place over a range of temperature rather than at a fixed
temperature. Thus, some austenite may be retained at room
temperature, as shown in Fig. 3.59. The amount of retained
austenite may vary from 1% to 30%. The retained austenite may
be transformed to martensite by the tempering operation or by a
method of cold treating, i.e. cooling the steel containing retained
austenite to a temperature between 20°C to -80"C. The cold
treating causes the retained austenite to transform to martensite
without any loss of maximum hardness the steel is capable of
attaining. Cold treating of steel is found suitable for low alloy and
special high alloy steels. Such steels have a tendency to ret'ain
more austenite at room temperature than the common carbon
steels, and respond noticeably to cold-treating methods.
Fig. 3.60. Microstructure
High carbon steel forms needle-like martensite while low
carbon martensite is more lath-like, Fig. 3.60. This low carbon lath
martensite is believed to contain very thin regions of retained
austenite between laths, or pocknts of laths, evidently due to
limited carbon enrichment.
Presence of martensite and retained austenite adversely
influence the toughness of ferritic weld metal.
1!!!!!I ••.. I!l ~
Basic Metallurgy of Fusion Welds 203
Welding Engineering and Technology
Formation of different microstructures in submerged arc
welding of steel under continuous cooling conditions are as shown
in Fig. 3.22.
3.8. Microstructure of Multi-run Welds
In multi-run welds the weld metal consists partly of a tertiary
microstructure i.e. normalised microstructure, as shown
Fig. 3.61. Microstructure showing outline of ferrite network with about
15% ferrite in a matrix of austenite.
schematically in Fig. 3.61. This structure normally has higher
toughness than non-transformed weld metal (the microstructure)
which means measured notch toughness will strongly depend on
notch position.
3.9. Effect of Alloying Elements on Microstructure
The chemical composition of weld metal is determined by
composition of consumables and the parent material, and the
degree of dilution". The influence of alloying elements on
of lath martensite.
microstructure appears to originate from a displacement of the
CCT (continuous cooling transformation) diagram to shorter or
longer times.
Alloying elements which cause formation of more acicular
ferrite, less grain boundary ferrite and ferrite side plates are
beneficial to toughness. In unalloyed and low alloyed ferritic weld
* Dilution is defined as the ratio of the area of parent material melted to the
total area forming th..eweld bead, at a given cross-section.
 iry-
;;~iiIi iiIi ••;j".••.:;..••~; •• ••
, ,,"'!lII,:a•••• ~~"' •••_~
•,•~""',,- ••_••--' T••.
••_,- __ ",,""
..
~"""""""".:;=="-,._" ....•• _"."".,.., ••
~ ••••••• __ •••• _"='",'",,,,"",,"',"'.••••. --.-.------

Welding Engineering and Technology

Z04

metal such formation is achieved if the CCT diagram is displaced

towards longer times i.e. towards the right hand side on the X-axis.
It must, however, be kept in mind that too large a displacement
may cause the formation of martensite and retained auste~ite i.e.
the deposit becomes over-alloyed for the prevailing level of dilution
and cooling rate.
3.10. Other Microstructural Phases
Apart from the different microstructures obtained in ferritic
steels, there are some import.ant phases encountered in stainless
steels. Two of these viz., delta (a) ferrite and sigma phase are
detailed.
3.10.1. Delta Ferrite in Stainless Steel Welds
Austenitic stainless steels find extensive use in power
generation and related industries. The welds of these steels
normally have duplex (i.e. ferrite + austenite) structure that
cont.ains varying amounts of C.-ferrite.It is well known that a small
percentage of 8-ferrite is gcner&lly required in the room Fig. 3.62. The Fe-Cr phase equilibrium diagram showing the
temperature microstructure of austenitic stainless steel welds to presence of sigma (0) phase between 650 and 820°C.
minimise fissuring. Proper control of the amount of a-ferrite in causes for embrittlement of high dlromium steels, including Ni-Cr
austenitic stainless steel wl\lds is therefore critical because too steels.
much b-ferrite (say > 10 vol%) tends to reduce their ductility,
toughness and corrosion resistance, and too little o-ferrite 5 «
The formation of a-phase is temperature and time dependent.
vollJ'Ooften results in cracking during weld metal solidification. For example, at 750°C the ()(~ a transformation requires only ten
minutes or so, while at 820"C transformation is completed after
Delta ferrite is the primary solidification product i.e. it forms 10 hours. Thus, during cooling, the transformation rate which is
directly from the molten metal. The o-ferrite at the core of the very slow around 820°C rapidly increases to reach a maximum at
dendrites, which form at the beginning of solidification, is very about 740°C, as shown in Fig. 3.63.
rich in chromium; the Cr-content however goes on decreasing as
the solidification proceeds. Sigma phase is considerably harder than the original ferrite;
ferrite containing 48% Cr has a hardness of RB97, that increases
The o-phase is readily detected by the microscopic
examination of etched specimens. Electrolytic etching in oxalic to Rc67 after transforming to a phase. Also, a-phase has a higher
acid solution 0.0% in water) brings out the ferrite boundaries, as density than the original ferrite.
shown in Fig. 3.61. without colouring the grains. Alkaline etching In Fe-Cr-Ni alloys a-phase can form quite readily from
(ferrocY1}ide+ K...1\1.n04 + NaOH) in hot solution stains the ferrite a-phase, even at low chromium concentrations. In duplex (a + y)
grains, so that it is easy to distinguish bet\veen austenite and steels, it is mainly the ferrite that transforms to a phase,
ferrite. Both the reagents produce a dark colouration on chromium simultaneously enriching the austenite with nickel. The formation
carbide grains. of a-phase from austenite is much slower; Fig. 3.64 shows the
3.10.2. Sigma Phase needle-like sigma phase formed directly from ~ustenite. The
y - a transformation can be explained on the basis that the weld
The presence of sigma «(j) phase is seen on the Fe-Cr metal is compositionally heterogeneous, the Cr-rich areas forming
equilibrium system, between about 650"C and 800°C, as shown in nuclei for a phase formation.
Fig. 3.62. This hard, brittle, intermetallic compound is one of the
 Basic Metallurgy of Fusion Welds 207
Welding Engineering and Technology /
206
1
The transformation of (ex + y) to (a + y) is reversible. Thus, if
1100 the steel is heated from 800 to 1l00°C after the a phase has formed
the latter will rapidly revert to ferrite of the same composition
1000 and, if the time allowed is sufficient, the austenite and ferrite will
homogenize.
900 In stainless steels a-phase embrittlement occurs in the
temperature range of 650 to 800°C and is essentially dependent
on the time at that temperature. The critical temperature i.e. the
.o...J

temperature corresponding to the maximum rate for the ex - y

- QI
•...
~
--, transformation, is around 720°C. The embrittling effect is limited
~
•...
when the ferrite content is small (2-3%) and no ferrite occurs in
QI
a- the austenite grains; however if ferrite content exceeds 12%, the
S toughness rapidly decreases.
I-QI

The surest way of avoiding a-phase formation in Cr~Nistainless

steel welds is to raise the Ni content, since the Cr content cannot be
reduced from consideration ofmaintaining goodcorrosion and scaling
resistance. For example, 25/20 Cr-Ni steel can be replaced by 23/28
o(-(J . Cr-Ni steel which is not susceptible to a-phase formation. It is even
400 I I I possible to replace, with advantage, the standard 18/8 Cr-Ni steel by
0.01 0·1 10 10
a nickel-rich grade with 16% Cr and 13%Ni.
Rate of sIgma formation per hour
3.11. Etchants for Study of Microstructures
Fig. 3.63. Mean transformation rates for (J phase formation: (a) full lines for Toreveal the different phases, and their morphologies, present
annealed ferrite, and (b) dotted lines for work-hardened ferrite. in a given weldment it is required to prepare the sample as per the
standard metallurgical polishing procedure and then etch them
with appropriate etchants. Etching removes the upper flowed layer
caused by polishing and thus reveals the microstructure of the
material for microscopic/macroscopic examination. The etchants
developed for parent materials are equally applicable to welded
joints; some typical ones for Black and White metallography are
listed in Table 3.2. A number of materials, notably stainless steels,
are best etched electrolytically.
'l'able3.2. Etchants for Black and White Metallography of
Welds in Various Metals and Alloys
(After T.G. Gooch)
-------
Material
Carbon and
Etchant
Nital : 1-5 vol% HN03 in
Comment
Good general purpose etchant;
low alloy steels methanol (or ethanol) can increase to 15% HN03 for
macroetching.
Picral : 4g picric acid in 100 Little attack on prior austenite
ml methanol (or ethano\) ------
boundaries.

Fig. 3.64. Needle-like sigma phase formed in 25/20 Cr-Ni steel along the contd ....
crystallographic planes directly from austenite by heating at 400 hours at 825°C.
~--~

208
--------_.~----------------_._-----_.--------------------
Welding Engineering and Technology

_____~I!:!.erial. E_'-!:..hant -.-f0mme!!:_t _

---·----=r-----·------T----
. 1 litre saturated aqueouil f'sa.5[JuNansa'-reveals prior
wetting
picric acidagent
with (s'Jdium
10 mil segregation.
grain boundaries and
dodecyl benzene
sulphonate)
Stainless 20% H2S04 with O.lglll Electrolytic, 1.5-2V Good general
steels NH4CNS purpose etchant.
109 oxalic acid in 100 ml Electrolytic, -6V Can reveal
H20 sensitivity to inter-crystalline

I corrosion.
SOg FeCla and 120 ml HCI I Macroetch.

Aluminium 10ml
in 480HF, ISml HCI, 25ml I 'Keller's reagent'.
ml H20
and its alloys HN03 in 950 ml H20 I Caution: Note - HF wash off skin
immediately.
15g NaOH in lOOml H20 I Macroet.ch.
under hood. Apply by swabbing
Nickel. 20% H2S04 with O..lg/l Asfol'stai::llesssteel.
Ni-Cr-X and NH4CNS
similar alloys
Ni-Cu (and 1 vol HN03• 1 vol acetic Mix acids first. Keep cool-
Cu-Ni) <tlloys acid. 1 vol acetone unstable and gives off N02-
Copper and 50g FeCI:!, 20ml HCl, High Cu alloys (e.g. Cu-Cr) and
alloys 960ml methyl alcohol. Cu-Ni

'110ml
ammonium0.880 persulphate.
NH3. 10mg Use fresh
90ml H20
20g Fe(N03). 2g AI·bronzes
NH4N03• 2ml HN03, 500
ml H20
Titanium and 10ml HN03, 2ml HF, 88 Caution: Note - HF wash off
alloys ml H20 skin immediFltely.
Keller's reagent See Al

Colour etch ants are used where it is intended to distinguish

between phases by means of staining and consequent colour
contrast. Table 3.3 lists the etch ants used for colour metallography
of some of the industrially important materials used in the
fabrication industry. Fig 3.65 shows the black and white, and
colour metallographic representations of the microstructure of the
same sample.

Fig. 3.65. Microstructure of stainless steel based on

(a) Black and White metallography, and (b) Colour metallography.
 j
f

Table 3.3. Etchants for Colour Metallography of Welds in Industrially Important Metals and Alloys =
III
(After Vander Voort) <II

n'
a-;:
Metal Phase Detected Comments III
Etch Composition
e:
Aluminium alloys (a) FeAl3 0.1-10 ml HF
90-100 ml water*
Attacks FeAl3; other constituents outlined. c.,
!ICi
"<l

(b) Al3Mg2 1-2 ml HN03 Colors Al3Mg2 brown (Al-Mg alloys). Immerse for 15 min for o
....,
98-99 ml ethanol I-percent solution. "'l'
c:
(c) CuAl2 20 ml water Mix one part of reagent with four parts ethanol. Immerse <II

o'
20 ml HN03 until surface is colored. CuAl2 blue. ::s

3 g ammonium molybdate ~
III
(d) CuAl2 200 ml water Immerse for 2 min. CuAl2 violet. (Lienard and Pacque). E:
<II

19 ammonium molybdate
6gNH4Cl
Copper (a) 12 100 ml water Colors 12 in aluminium bronze. Immerse 90 to 150 sec.
10 g sodium thiosulfate (Aldridge)
(b) Alpha One part water Swab etch. Colors alpha in alpha-beta brass.
One part H202 (3%)
One part NH40H
(c) Beta 50 ml saturated aq. Colors beta in alpha-beta brass. Immerse about 3 min.
sodium thiosulfate (Klemm's I reagent)
1 g potassium metabisulfite
(d) Beta 10 g cupric ammonium chloride Add NH40H until solution is neutral or slightly alkaline.
100 ml water Colors beta in alpha-beta brass.
Iron and Steel (a) Ferrite 50 ml saturated aq. Colors ferrite (blue and red), martensite (brown), austenite
sodium thiosulfate and carbide (not affected). Immerse 60 to 120 see after light
1 g potassium metabisulfite pre-etch with nital or picra!. (Klemm's I reagent). N
o
\Q
(contd.)
 a,
1e.
I~
Ol
'1
0~cant
::s
IJQ 20
30
240 11000
25 g(1)Phase
2100
8-15 picric
lead
NaOH gml
ml
gsodium
gcitric Detected
water
water
sodium
acid
gacetate
Bpotassium
sodium
Etch
acid unaffected.
surface
thiosulfate
Immerse
1metabisulfite
Pre-etch
sodium
metabisulfite
Darkens
only.
Pre-etch
Composition
min.
thiosulfate isml
in
lightly
Pre-etch
Etch
yellow-orange,
next. Age nitrite
as
only
electrolytically
'lb
30
is
Mix
40
Add
with 100
solution;
colored.
to
Dissolve 120
and
see;
HNOa
picra!.good
sec.
colors
in (fable
of
to toComments
(Beraha.)
colored.
boiling
with
as-quenched
with
lightly
24 good
at
(1)
for
Cementite
age Colors
in
order Colors
same
ferrite.
decrease
Immerse solution
4picra!.
100
ferrite
hr picra!.
with
6atVfor
given.3.3.ml
dc,
heated
hr.
MaC
will
After
in the of
nital. III
Er
fD-
ferrite.
Immerseto
martensite;
Immerse
picra!.
15 min.
unaffected.
20°C in
(Contd.)
20°C,
be
and
manner
Allow g'
to60
Pre-etch 15
solution;
Immerse
Immerse
Colors
dark
0.5-2
70°_75°C,
darker
pH
solution
~~.
::I MsC
to
each with
as
90
to ill
(BerahaJ
min.
in
(BerahaJ
Nin2
brown
than
etch
tosec.,
2.5-3.5.
about or
solution
good
carbides
bottle. for
in
add
picral use
ferrite; for
Filter
(0.08-0.3
ferrite;
15 to
No.
ferrite
dissolve sec. for
0.530
about
solution
..Pre-etch
Immerse
black.
3e. about
ml
~nS
isAdd
before minutes
unaffected.
until
cementite
10020of
20
yellow
Colors sec.
surface
until
ml
white.
0.2
adding
Fe3C of
(2).
to
or
lightly
Ncm2), g =
rater
blue. chloride
onium
mium
molybdate
Oamonium
aP04
(Beraha).
nitrate (, d."~
bifluoride
S-
IJQ
j s:
~
(d) FeJC N

--_._~....,...,._--~-------------
"'!

!
Table 3.3. (Contd.)
Metal
---~-- Phase Detected
(i) Oxygen
-- --Etch.Composition
---
145 ml water
comments------- .. /
16 g CrOa
BOg NaOH Alkaline chromate etch .to detect oxygen enrichment during g
or before hot working. Heat to 118° to 120°C; immerse for 7 ;
Ij) FeaP 100 ml water to 20 min. Enriched areas are white. (Fine.) ~
FeaC o
10 g KaFe(CN)6 Immerse at 20°C; M7C3 is attacked; Mo2C turns brown to •..••
Alloy carbides 109 KOH
Sigma black, Fe3C is unaffected. Bring to boil. Feap darkens in 10 ~
Delta ferrite see; FeaC is lightly colored after 2 min. For stainless steels, ~.

BO°C to boiling for 2 to 60 min.; carbides t'lrn dark; sigma, t-

(k) Sigma
Carbides
100 ml water
blue; delta, yellow to brown; austenite is unaffected.
(Murakamj's reagent).
immerse at 20°C to reveal carbides, sigma after 3 min. At l:ls:OIl

20 g K3Fe(CN)6
Delta ferrite 20 g KOH For stainless steels, use cold to boiling; carbides turn dark;
(l) Sigma
and
delta,White).
yellow, sigma, blue; austenite is unarTected. (LeMay
100 ml water
Carbides
30 g KaFe(CN)6 For stainless steels, Use at 95°C for 15 sec.; sigma turns
Delta ferrite 30g KOH
reddish brown; delta, dark gray; carbide, black; austenite is
(m) Austenite unaffected. (Kegley).
Heat tint in air at 500 to 7000C,
up to 20 min.
For stainless steels; austenite colors before delta ferrite.
(n) Sigma 10 N KOH
Delta ferrite
(56 g KOH in 100 ml water) For stainless steels, use electrolytically at 2.5 V dc, 200e
for 10 see to color sigma orange and delta light blue and
(0) Sigma 100 ml water
Delta ferrite 20 gNaOH
For stainless steels, use electrolytically at 5V dc, 200C tor
20 see to color sigma orange and delta tan; austenite is
~~~.
brown; austenite in unaffected.
(contd.) I
N
•...
•..•
 Welding Engineering and Technology
212 r
..,0 '" i:f
'" '" c5.sa 0 ~.
.. -'
~ 00 NI':
N a ~-;
.s.;3..,
o _
g~~==
1: ....
'" .., ~t>ll ~ ~ oo. '"
"O~
.., o .D ~~.~~
u
.-
u
It:
'"
' II t>ll

.J6 bE9
•...

g15
.•..•
'" I': .•
as '" .t: •...
u.J:: 0 0,
o<0~..9-0
~.CJ_.D~
::g
ti'"
~.~
C'l
;;] 'N
.•..• '"
~•
.•..• u1l
~ .9 "'$
I': u
Q) Q)
'" ~~~~
Qi' .••.•
'l:l ~ =i ~..c::
.s
•...
,-(~cJ
~••.•§'" 0~
t'"':'
~~
..,~ u 'e~ gj $::(. .g,
.•..•". .t: .D .••• b.O
.;:D
::g
..,
'" .qo :g .s~ -:s .D
~ ~ .s o~
-'
00 -' ::5
~ "Q)'.'. ~
~
~ ~~
E§~15C3
~
'~"en
~.~
£ S
d ~
ell
15]
.•....
'" '"

~§
.s
o-'
..,'.5
'"
~
a a_
'"
C':l

....a
..s •...
<U~
Q) o.E~(;l
~ ...-:::::''''"'S
.~ E ~
Co}..$ Q,) ~
FOUR
Euu.J:: a .., ~§ ~.s
.., > ~• ctf
.S 1l Sl1
..,
Welding Stresses
Q)
o ~ ~'.B . ':'" .§
Co) .., 0'" g u.-"l
..,u ~
..,
•... a~ _
~ Po;~:S
Q) CI) ~
Q;l ~ CI) UJ '" '" ••.• Po .• I': o
Q ~Q) I': •..•
gj .s:; g -...:. ,
;,,-~ a
C'l

'"
.•..•
.••.•

'"
a
.S
.9 '"
<a a
•...
oU.J::'"
'"
.~
Q) '"

;a ~ ~. and Distortion
~ ~.~
$~
>
'" .s ~
C':l
~~
g ~~
•••• »
'"
o
w
Po",
:j
a
t>ll
Q)

~<all
Po
~~.;3~~a il
8 ~ .;3

",..,~ •...
CD
0 <.f
<a::::
d '" :;a ~-g
..s '"
C'l
.s.s ~
rn '.-4
c1JQ).£G>
..c Qi>. .S
.!! ~> ~-: N .~
•...
~.s :-:=
.•..•~ '" The inherent local non-uniform heating and cooling cycle
.1': ] <.0 ~ 15
<l)
p.. {i.., ~.. gs ::5 (f)
0')
I': I': .., I': .•..•
ca .., .•..••..•
~~< ::g~ .J:: b..o.CO) ::5
'"
•
<l)
.•..•
..c
",0
'";1 associated with the fusion welding processes like SMAW;GMAW,
•....
§ ~ b a..,
.•..•
.D .: .~ Q) ~
Q,) ~ d23:!%Q.l
~ GTAW,SAW, ESW, etc. results in complex thermal stresses and
~§a\ ~8
~ u '" •....• ~::g ::5
•...
~a ~ 5 (f)OO •.••;r'Q)ol':
b.D e.a t.""'" ~

strains which may finally lead to' the development of residual

,-. d
"ti
oo
o stresses and distortion in welded structures. Whereas residual
..•..
stresses can reduce the service life of a structure or even cause
- I:-
c::

oo
..,j
t:
.s Q)
.~, catastrophic failures, distortion usually results in misalignment
.~ o g. with consequential difficulties in assembly as well as p'oor
~ ~ :su <Il

o
.;;;
o~ o ..... •...
appearance of the finished product or structure. It is, therefore,
-
~
::I: ~
'" '"
~ u •....•~~ .~ '"
~
....•
imperative to get prior knowledge of the possible effects of welding
-'-"C,) 0' 0
•...
..,
,J:J.
0 g~::I:
"'..,
•...
<~ .... I':
$'" ""

•.....
'ca
d parameters on the structure for the proposed welding procedure
~
~
!\
~ 0
~
.::l
'-'
15

u
::I:
.••.

~::I: i3~~i
...s'
~~ ~ ",0 .....••.•.•-:s::a 00'"",,$ .•.•..•
~eg ~£~ .§
a C':l 0 a
l:i<
<Il
•...
<Il
and set-up.
4.1. RESIDUAL STRESSES
0..•.
::I:
ai o ziC'l 8.s~
00
a .
o a0 ~
8~t>llo.s -' 00 C'l ::I:
~
'" Residual stresses are defined as the stresses which remain
-' -' C'l
Z 0'> -' ..-l 10 ""Ctf 1O.....-t .•..•
I': within a structure when all external loads or reactions are
<Il

"'lj
~ '" ~..,
1)
.•..• ~
<Il
removed, hence they must be self-balanced within the structure
:::
"'"d .- CI)
•...
itself.
~ Q)
<~ I':
~
:.8
Q)
I': .•8 '"$'.
~~ :=
o:l
~ :'E 'ca '"
~~~~~ Since residual stresses exist without external forces, the
Q~ ~'"
lj 0~
t>ll ~
;-
-
;-0'"
•...

~] P:::.~ < ::g 0 .s

resultant force and the resultant moment produced by them must,
~
Q., .50
::g

s
::g

~ s :S ] ~ d
•...
Q)

~
therefore, vanish. It is now well accepted that the residual stresses
'"C
commonly arise from permanent changes in the shape of the body.
S It means that if residual stresses exist in a welded component the
'"
.•..•
rn plastic deformation must have occurred which is associated with
•... '" ;.a
.9 >. a temperature cycle involving temperatures up to the melting
ti .9 <Il
rn
<a '"
:>, <a o point of the material. The residual stresses generated during
a .9
~ ::5
<a e:j welding may hamper the functional efficiency of the component,
]u
Q)
'Cij
..,
I': .~
.•..•
o leading to failure of the engineering structures. It may alsoJead
'"
t>ll Z
::g Z ~
214 Welding Engineering
to brittle fracture of the welded structures causing enormous
damage to resources and loss of human life.
Residual stresses are the major constituents of a stress field
around a crack which may lead to cracking. For example tensile
residual stresses reduce fatigue strength and corrosion resistance
while compressive residual stresses diminish the stability limit.
Also, while tensile residual stresses may initiate the failure due
to fracture, the compressive residual stresses near a weld can
reduce the capacity of the structural member in buckling and
collapsing. The residual stresses in some regions may be as high
as the yield strength of the material. All these facts point to the
importance of the knowledge about residual stresses in a structure
fabricated by welding ..
4.2. CAUSES OF THE DEVELOPMENT OF RESIDUAL
STRESSES
In arc welding the work is raised to the melting point within
a very limited area, b1lt the temperature drops sharply as one
moves from the heated spot. This is due to,
(i) the high thermal conductivity of the workpiece,
(ii) the subsequent rapid dissipation of heat, and
(iii) the insignificant volume of the heated metal.
These conditions prevent the uniform expansion of the metal
and may produce serious internal stresses, distortion and
warping.
a 0 2
03
(0)
+ Tension
-. Compression
aL ~~_
o ----
03
(b I
Fig. 4.1. Stresses set up in a steel plate: (a) on heating, and
and Technology Welding Stresses and Distortion 215
To examine the internal stresses set up in the plate during arc
welding, consider a steel plate divided into three zones 0, 2, 3) of
equal width as shown in Fig. 4.1.
As zone 1 is raised in temperature, it will tend to expand. If
it were free to expand longitudinally its length would soon be
a1b1' However, zones 2 and 3 restrain its longitudinal expansion,
because they are not so hot as zone 1. This will result in
compressive stresses (-ve) in zone 1, and tensile stresses (+ve) in
zones 2 and 3. Zone 1 is now free to expand to a2b2 only and it
may be considered compressed by the amount a1a2 + b2b1 or
subjected to plastic flow.
On cooling, zone 1 will tend to contract longitudinally by the
amount aa1 + bbl' But it will be restrained from contracting by
zones 2 and 3. As a result, zone 1 will be under residual tensile
stresses, and zones 2 and 3 under compressive stresses.
When fully cooled, zone 1 will have the length asbs which is
shorter than the original length ab, resulting in a plastic flow or
permanent deformation of metal.
The t~o halves of a butt-welded plate cut along its axis will
warp yas shown in Fig. 4.2 (a). The butt weld restrains warping.
x x
( b I longitudinal stress
pattern along XX
y I c) Transverse stress
(0) Butt welded plates pattern along Y-Y
(d) Welded plates after
(b) on cooling. cutting along weld X - axis
Fig. 4.2. Stresses and strains in butt welded plates.
216 Welding Engineering and Technology Welding Stresses and Distortion 217

The resulting longitudinal and transverse stresses will be

distributed as shown in Fig. 4.2 (b, c, d). Fig. 4.2 (c) indicates that x
transverse tensile stresses will be set up in the middle of the weld,
and the ends of the weld will be under transverse compressive I 0 6T~O St~
stresses. A---- ,. Section A-A c:
o
Apart from physical effects of the weld thermal cycle it also 'Vi
c:
B
results in structural changes in the metal being welded. The ~
structural changes brought about in the metal by the welding heat
are accompanied by volumetric changes and, therefore, give rise
to locked-up stresses. This happens because on reaching a definite
point in heating or cooling, the structure of steel undergoes a
transformation, th.t.o-.
different structures having different c.
E
densities·. For example when low carbon steel is raised to the point uo
at which its ferritic structure transforms into austenitic its volume
is somewhat reduced because austenite has a higher density than
ferrite. When steels high in carbon are cooled rapidly, the •••..
"'lI.V
111\
'l.IJY
_

austenite changes into martensite at 200 to 350°C. The martensite

is of lower density and. has a greater volume than the austenite.
This change in volume brings about internal stresses.
Thus, the three main causes for the development of residual
stresses in welded structures are:
(1) local heating and cooling of metal,
(2) shrinkage on solidification, and
(3) structural changes on solidification.
Detailed description of these effects on the development of
residual stresses is given in the following sections.
4.2.1. Effects of Weld Thermal Cycle and Shrinkage on
Residual Stresses
In most cases, expansion and contraction, caused by weld
thermal cycle, produce no harmful results; however under certain
conditions they may reduce the strength of the weld and, indeed
of the structure as a whole.
Fig .. 4.3 is a schematic representation of changel? in temp-
eratureand longitudinal stress (ax) i.e. the stress in the welding
direction during welding. The weld bead zone MM' is the region
where plastic deformation occurs. Section AA is ahead of the heat
source and is not yet significantly affected by the heat input. The
temperature change due to welding, 6.T, is essentially zero. Along
* Ferrite (a-phase) is a B.e.e. (body centred cubic) structure with 68% packing
efficiency while austenite (y-phase) is a F.e.e. (face centred cubic) struct.ure with
74% packing efficiency. H.e.p. (hexagonal close packed) structures like Zn. and
some stainless steels also have 74% packing efficiency.
3. Section (- (
(0) Weld Residual
stress
0-0
b. Temperature' c. Stress,crx
change
Fig, 4..3. Schematic representation of changes in temperature
and longitudinal stresses during butt welding.
section BB, which crosses the heat source, the temperature
distribution is rather steep. This temperature distribution
becomes less steep at some distance behind the heat source (e.g.
along section CC) and eventually becomes uniform very far away
from the heat source, along section DD.
Section AA is not affected by the heat input, the thermally
induced stress ax is , therefore, zero. Along section BB, aX is close
to zero in the region underneath the heat source, since the weld
pool does not support loads. In the regions somewhat away from
the heat source (in the lateral direction), stresses are compressive
(i.e. ax is -ve) because the expansion of these areas is restrained
by the surrounding metal of lower temperatures. Because of the
low yield strength of the high temperature metal in these areas,
218 Welding Engineering and Technology 219
Welding Stresses and Distortion

(jx reaches the yield strength of the base metal at corresponding

, tensile. Finally, along section DD, high tensile stresses are
temperatures; ax in the areas farther away from the weld (in produced in and around the weld and the balancing compressive
lateral direction), are tensile and' balance with compressive stresses beyond that. Since section DD is well behind the heat
stresses in areas near the weld. source, stress distribution does not change significantly beyond it,
Along section CC the weld metal and the adjacent base metal and this stress distribution is, therefore, the final residual stress
distribution.
have cooled and hence have a tendency to shrink, thus producing
tensile stresses; that is ax is +V8. As the distance from the weld To understand how section DD experiences the weld thermal
increases, ax first changes to compressive and then changes to cycle qualitative relationship between temperature and stress
c.. may be considered. As the temperature increases the initial
E
<II
.•..
expansion of the section is restrained by material farther away
., from the heat source, which generates compressive stresses in the
f •..
.a .,
section, as shown schematically in Fig. 4.4. The corresponding
"'..::
•.• ~ en stress-strain relationship is also shown in the same figure .
8..s~ The elastic portion of the stress-strain curve is non-linear
a~""
~,~ § because of the decrease in Youngs Modulus (E) with rise in
~~~ temperature. At some critical temperature (point 2 in Fig. 4.4) the
~ <.0 Q)
~ c: ~
~ '0
' flow stress of the section is exceeded, and further heating results
._ ~A
en ~~
<1> .::
::s in a decline in stress as the metal become softer. At a temperature
§~ ~ ~
:a >. ¢:: Q)
of about 700°C for a micralloyed steel, the stress at the section
'" u '" ~ (DD) is practically zero (point 3) and considerable plastic strain
.;;: e;j w §
'"
> a ~."" 1-1 ::::= ~
may occur. At the peak temperature reached by the section the net
~ ., 'ca as
~-Sb:E
strain is represented by the vector 1-74; on cooling the reverse
b+ b occurs. As the temperature of the section decreases, resistance by
I 'o::s!ena
d C1.)~ c.'d
adjacent, hotter material initially reverses the plastic strain
,8 ~ a ~
~'~---~ pattern until a temperature is reached at which the flow
<" .b ~E-.s
tI'J .~ 00 resistance of the section increases and that subjects the section to
::s~UJ""t:l
==::1
._ "'d b <l.) <l)
'"0 tensile stresses (point 5). Further decrease in temperature is then
.~.s~
'" '" '"
8 accommodated by the section in the form of elastic tensile stresses
~~~ until it cools down completely (point 6). Depending upon the
-c:r.boo
lZl
u en .,
'" •.•
rigidity of the structure the residual. stress can (and usually does
"'":
. b~
en C
in thick welded joints) equal the yields stress of the metal. Fig .
"<I'"Oli: 4.5 shows the typical distributions of residual stresses in a butt
,~
rx..
~ weld in low carbon steels.
According to Masabuchi and Martin, the distribution of the
-4
longItudinal residual stresses, ax can be approximated by the
c following equation.
'0
•...
.•..
Vl
W
I _ [ l-(b):l 2] e-.~ (f J
ax(y)-am ... (4.1)
where, am is the maximum residual stress, which is usually as
high as the yield stress of the weld metal. The parameter 'b' is
width of the tension zone of ax (Fig 4.5).
Z20 Welding Engineering and Technology 221
Welding Stresses and Distortion

y are added along the weld as the reaction stress. The external
constraint, however, has little effect on the longitudinal stress (crj.
Normally the plates used in most fabrications are relatively

x -- rJy

rJx'
x
thin « 25 mm) thus the residual stresses developed in the
thickness direction (crz) are not signific'ant and are, therefore,
normally ignored. However, when thick plates are welded,crz may
be significant and therefore accounted for to determine the final
magnitude of residual stresses.
Re!)idual stress distribution patterns in a butt welded thick
y (> 25 mm) free flat plate (i.e. without fixed ends) are shown in Fig. 4.6.
I a) Butt weld
-rf .C(

Compression Tension

e1'"X' .0-
b
- 0-

I b) Distribution of CJx along YY

___ 01_...,,- Curve 2

, --
lensiof'! r . ~CurVl~1'"
Reclctio~ ""-t --
x
stress
x I ,-if$< ]
Compression Fig. 4.6. Residual stress distribution in a butt welded thick free plate.

~
Ie) Distribution of cry,)long X X
Fig. 4.5. Typical distribution patterns of longitudinal (0',,) and transverse
residual stresses in a butt welded joint.
The distribution of the transverse residual stress, cry,along the
length of the weld is shown in Fig. 4.5 (c). As shown, tensile
stresses of relatively low magnitude are produced in the middle
part of the weld, and compressive stresses are produced at the
ends of the weld. If the lateral contraction of the joint is restrained
by an external constraint, approximately uniform tensile stresses
4.2.1.1. Reaction Stresses
The residual stresses caused by non-homogeneous thermal
expansion or contraction are called thermal stresses. Residual
stresses without external support forces aTetermed as Constraint
(O'y)
Stresses while those caused by forces developed due to external
restraint are termed as reaction stresses (in some special cases
these may be referred to as mis-alignment stresses). The constraint
stresses are in equilibrium with themselves while the reaction
stresses are in equilibrium with the reaction forces at the
supports. If a component is supported in such a way that it can
deform freely, only constraint stresses occur. However, if the
component supports restrain the deformation, the reaction
stresses are superimposed on the constraint stresses, as shown
schematically in Fig. 4.7.
I
2Z1 Welding Engineering and Technology Welding Stresses and Distortion 223

10

~. 0 -
0Ew
.•...•
c: 0
.•...•
76
Tmax
TmaxTmax
0
...
:~
f5~Clma" a-R (maxl
t.ts/s=35s
( b) Longitudinal weld

300 400 500 600 700 800

Temperature (.( I

Fig. 4.8. Temperature versus dilation curves for a Q&T steel.

<1'Ro I cr ( (max)
(a I Transverse weld I--lfR (maxj= lfy_
Fig. 4.7. Reaction stresses (ORn) caused by restraining of the transverse edges,
and constraint stress (O"e) in:(a) Transverse weld, and (b) Longitudinal weld.
4.2.2. Stresses Generated by Phase Transformation
An important contribution to residual stress concems the
dilation due to phase transformations, e.g., in transformable
steels. The temperature at which the phase transformation
(y ~ a) occurs during cooling of the weld depends on various
factors, such as the grain size, the peak temperature reached, and
the subsequent cooling rate. The magnitude ofthe volume changes
that occur during the transformation can be obtained e.g. from
weld simulation experiments which employ a dilatometer.
Examples of such change~ in Q and T (quenched and tempered)
steel are shown in Fig. 4.8 as a function of three peak
temperatures. It is observed that higher the peak temperature,
the larger appears to be the dilation. This effect arises because
the lower the y -,) a transformation temperature the larger is the
dilation in transforming from the F.e.C. (y) to the B.C.C. (a) lattice
i.e. from 74% packing to 68%pacldr..gefficiency.Larger grain sized
material also tends to decrease the transformation temperature
because of the reduction in grain boundary surface at which the
B.C.C. ferritic phase can nucli~ate.It, thus, follows that the largest
dilations-an.d hence the largest stress changes--are associated
I
with the martensitic transformation (y ~ CJ!) in which the dilation
is 4% corresponding to 1.4% linear strain.
The additional stresses due to phase transformation interact
with the quenching and shrinkage stresses, as shown schema-
tically in Fig. 4.9. Last row of diagrams in this figure show the
longitudinal stress at a section perpendicular to weld centreline,
as a result of,
(a) Combined effect of shrinkage and quenching,
(b) Combined effect of shrinkage, quenching and phase
transformation, and
(e) Cumbined effect of shrinkage and phase transformation.
It is to be noted from all the three. diagrams, Fig 4.9 (a, b, e),
that the maximum tensile residual stresses are not present at the
centre line of the seam but at the toe. These tensile stress maxima
represent the residual stresses owing to shrinkage. A thumb nIle
for determining the nature of stress at a given section is that in
the areas of the component which cool down last, tensile stresses
occur where thermal stress dominates, and compressive stresses
occur where the transformation stress dominates.
The difference between the maxim urn tensile stress and the
stress value at the weld centre line r.epresents either the surface
quenching type of thermal residual stress, phase transformation
type residual stress or both together. It is to be expected that the
transformation stresses diminish as the area being transformed
becomes wider.
Finally, the following is a summary of the chief factors
responsible for the setting up of residual stresses.
 ==~.:=~==-==""'=====--""-----------------------

/~T
~._=::!!~ ..~.,~
••~,,,,~.,,",_1I
'-",··~-··I

l-
00
C;C
'0-+2Z4 >< 0
+-
=-
w
+-
-0
'-
I-
10
j
0~~~
'" +
-;;:;
'"
~e0
1-0
•...
.•..•
o£.t::•
we::
<l.>
-0
Q.. .-
.~Q)
-IY
+ """"0c~b
b£~
0-00:
..3~
•.•.•

<l.> Cl
.•..•
.Cl<l.>d ""
C""
t,;. cd
";"eJ";;Cl
-
I'~] ~bi>~Cl g
<J Cl""
Cl.Cl
<l.>
gj ~:§
b.ll
x .S
~~~
'" ri: ~~:3
...,~
- ~o
.g §.~Engineering
Welding
~~tti
._.- Jo..4
~ <l.> e
and Technology
<l.>~';'::

Welding Stresses and Distortion 225

.~ ~
<l.> •..•
C

~ E- ~t ~
J ><
7. Expansion and contraction (free or constrained com-
~
ponent),
8. Rate of cooling,
9. Stresses already present in the weld metal.
Internal stresses remain locked up in welded joints, because
contraction in them is usually restrained. But these stresses vary
widely in magnitude from negligible to very high. They may
sometimes exceed the strength of the metal and will then give rise
to cracks in the weld or in the parent metal. Locked up stresses
x are harmful, since they may impair the strength ofthe weld metal;
it sometime happens that the internal stresses require only a
slight addition from external loads to exceed the strength of the
structure and thus produce failure.
Residual stresses in weldments can cause serious specific
problems such as hydrogen-induced cracking and stress corrosion
cracking.
The very severe stress (crt) associated with the weld at the edge
of the plate in the transverse direction is known to produce end-
cracking problems in some steels unless preheat is applied.
4.3. MEASUREMENT AND CALCULATION OF RESIDUAL
STRESSES IN WELDMENTS
Residual stress Measurement Methods

~I ----,- ----I
Experimental Methods Analytical Methods

-I

Destructive
-I-------~I
Methods Non -destructive Methods

1.
2.
Qualities of the parent metal and filler rod or electrode,
Shape and size of weld,
~--r-_
Uniaxial Stress ~~'lxial Stresses Triaxial stresses
x- ray

method
method

Neutron diffraction

3. Comparative weights of weld metal and parent metal, Ultrasonic method

4. Type of joint and method used in making weld (e.g. method l... Magnetostriction
tacking, back-step sequence, etc.) ·1StriPResll ience
cutting method
method t Block cutting
Hole drilling method ~ 3-d
Slice cutting
method
cutting method
5. Heat input into the weldment i.e. on flame size in oxy- Blind hole drilling meth Boring out or
acetylene welding; and current, electrode size and weld- turning off method
Cutting out method
ing speed in arc welding,
6. Type of structure and neighbouring joints, Ring groove or trepanning method
Resilience method

techniques
Moi re f ri nge. and/'a ssoc iated
Fig. 4.10. Classification of measurement methods for residual stresses in weldments.
r~,,,,-:,,,,,,~,,,,'''_>'~,- -_'-_-- __ -T''-" __
'_'''''''''''~_'''h_''""_"'*"'_., ._~~=~_m~~n__""~
226
Residual stresses in weldments are measured both by des-
tructive and non-destructive methods. DestruCtive methods are of
two types depending upon whether there is complete or partial
destruction of the component. The latter includes small holes or
ring grooves in the surface of the component which may permit
restricted further use of the part.
Depending upon whether the residual stresses, to be
measured, are uniaxial, biaxial or triaxial, the measuring methods
can be grouped as shown in Fig. 4.10. These residual stresses can
as well be calculated by analytical (or numerical) methods.
4.3.1. Destructive Measuring Methods
In destructive measuring methods the elongation or short-
ening of a small measuring base on the surface of a component is
determined while the_component is subjected to off-loading by
cutting, drilling, etc. What is measured is the relieved strain or
the resilience of the completely or partially cut-out part .of the
component involved. The relieved stresses are then determined by
back calculations m;rlting use of Hook's Law. This possibility of
back calculations exists only for bodies of reI atively simple
geometry (e.g., beam, plate, cylindrical or spherical shells) for the
elastic behaviour of which theoretical solutions are available.
Resistance strain gauges, detachable strain gauges, and photo-
elastic surface layers are mainly used for making such measure-
ments. The measurement and analysis technique is simplified if
the principal stress direction is known, this may be determined,
say, by means of brittle lacquer.
Temperature constancy is absolutely essential for precise
measurements with all residual stress measurement methods. It
is imperative to ensure temperature constancy during cutting,
drilling, or layer removal operations. When measuring the
resilience of the residual body (i.e. the component remaining, say,
after layer removal), the yield limit must not be exceeded if high
accuracy is demanded. This limits the resilience methods to
measurement of·residual stresses markedly below yield limit.
S<,>meof the residual stress measurement methods are
described in detail in the following sections.
4.3.1.1. Measurement of Uniaxial Welding Residual
Stresses
In many cases it is adequate to determine welding residual
stresses caused by mainly the uniaxial effect. In general, this
uniaxial residual stress is distributed non-uniformly over the
cross-section of the component, for example, the longitudinal
_
Welding Engineering and Technology
I Welding Stresses and Distortion
residual stress in I-girders with web-to-flange weld or the
227
transformation stress at flame cut plate edges.
In the Strip Cutting Method the component is cut in direction
'x' of the residual stress to be measured into a fairly large number
of narrow strips, as shown in Fig. 4.11. From the measured
Fig. 4.11. Schematic illustration of strip ~utting method applied to I-girder.
relieved strain (e,) residual stress (ax) can be determined by using
the following simple equation.
ax = - Eex ... (4.2)
where, E is the Young Modulus for the material of the component.
Cutting of strips is done by a saw and occasionally the strips
are left connected to one end of the flange in a comb-like fashion.
The relieved strain is measured by means of detachable strain
gauges or bonded resistance strain gauges. The robust detachable
strain gauges act by means of measuring balls which are inserted
into the surface of the component at the end points of the
measuring base. The measuring base length of 100 to 250 mm is
comparatively large which makes it necessary to assume that the
residual stress within this length is non-varying. ThE:!more
sensitive resistance strain gauges permit a shorter measuring
base length, however the connecting wires attached. to them may
impede the cutting process.
A point to note in this method is that a part of the strain is
relieved by flame cutting the I-girder prior to strip cutting, shorter
the flame-cut section, higher the amount of stresses relieved.
In the Resilience Method bars with varying uniaxial longi-
tudinal residual stresses are subjected to layer removal by
planing, grinding, etching, spark erosion, or transverse saw cuts
while holding the bar in plane clamping condition, as shown in
-li'r,?"-.:O:»!;l>, !!i> ~~=~!11~--.I!!!!!!I!~I_==-==!III!!III!---_·

Welding Stresses and Distortion 229

228 Welding Engineering and Technology

/~/~/~~~igiT
~-~-=~
Machined off Strain gauges
layer
This resilience method has been used with particular success
for determining residual stresses in flame>cut plate edge.
A special variant of the resilience method is the cutting ofring-
shaped slices from thin-walled cylindrical shells shown in Fig.
longitudinal
Machined off Resilienc e (bl Strain gauges to measure strain wele
layer deflection. w relieved by machining

(a I Measurement of curvature
Saw kerfs ~3t3h
by dial gauge

.'~
~-.----- ';' · r-r-;..J:-
'7
(c) Measurement of strain relieved by
saw cuts
Fig. 4.13. Schematic illustration of a longitudinal weldin a thin-walled
Fig. 4.12. Strain measurement by Resilience Method cylindrical shell with the position of rings to be>marked on it.
(a) measurement of curvature by dial gauge, (b) measurement of strain relieved
by machining, and (c) measurement of strain by saw cutting method. 4.13, followedby measuring the difference in outside diameter and
calculating the local circumferential residual stresses on that
Fig. 4.12. The bar is unclamped between the layer removal basis.
operations and the bar curvature which results from the resilience
is determined by means of spring-back deflection, Fig. 4.12 (a). 4.3.1.2.Measurement of Biaxial Welding Residual
Curvature or strain is measured on the side opposite to wherefrom Stresses
the material is removed. The initial longitudinal residual stress Measurement of biaxial welding residual stresses is quite
(O"x) distribution across the bar height direction Z is calculated, on often undertaken though compared to measurement of uniaxial
the basis of beam bending theory, by the following equation. residual stresses it is a slightly more involved process.
In the Block Cutting Method the orthogonal directions x and
O"x = ~2 [h"2 (~~)+
4hW - 2 S:· W (z) dZ ...(4.3) y of the normal stresses O"x and O"y to be measured are given; these
may, but need not be the principal stresses. The stresses (O"x and
where, E = modulus ofelasticity,
O"y) are assumed to be uniform in the thickness direction, this is
l = bar length over which the deflection (W)occurs,
an acceptable assumption for sheets and thin plates. The strain
h = variable height ofthe component following the removal gauges, for measuring the relieved strain, are usually applied to
of the layers,
ho = initial height ofthe bar,
W = resilience deflection.
Thus, the initial longitudinal residual stress at distance 'h'
from the measuring side (bottom in Fig 4.2) can be determined
with the help of equation (4.3) by putting the value of deflection,
W, measured following the removal of the layers. Alternatively,
t3~E3rnQDDO
residual stress, O"x' can be determined from the measurement side
by relating deflection 'W', to the measured relieved strain, E.
The. edge stress, O"x 0 (strain gauge side),'. which is of
-r-ODDD· DDO
~o 0000000000 0

considerable significance in practice, follows from equation (4.3)

by putting h = ho and W = Wo = 0, thus reducing it to,
L o~DODtljDODOO
Fig. 4.14. Schematic illustration of block-cutting method of stress
measurement applied to plates joined by butt weld.
F
O"xo="3 _/2
(ho) (dW\ ~
dhJ, •••... (4.4)
 Welding Stresses and Distortion 231
230 Welding Engineering and Technology

both sides of the plate. Following this, the plate is cut into direction \3, it is necessary to evaluate the relieved radial strain
rectangular blocks of about 30 mm x 30 mm, as shown in Fig. 4.14. Erin at least three measuring bases (strain gauges) in different
The residual stresses crx and cryare calculated from the measured measuring directions. This can be well accomplished by using a
relieved strains Ex andEy by the use of the following equations. three-element strain gauge rosette.
E . In this method, therefore, strain gauges in the form of a three-
crx = - I-v (Ex+ VEy)
--2 ... (4.5) element rosette are placed in the area under consideration. A
E through-thickness hole is drilled in the centre of the straiil gauge
cry== - ----2 rosette. In this way the residual stresses in the area surrounding
I-v (Ey+ vex) ... (4.6)
the drilled hole are relaxed and the relieved radial strains can be
wherE;lE and v are the Young's Modulus and poisson's ratio respec- measured with a suitable strain recording equipment. The radial
tively. strains measured are then related to the relieved principal
stresses by the following relationships.
\ At least three measuring directions, combined in a three-
element strain gauge rosette, are required for determining the
complete plane stress state . . = --=-
crl2 4A 4B -=
El+ E3± ...f2 -J (E, - E.,)
. 2 + (E"- E.,)
2 ...(4.7)
An extension of the block cutting method for thick plates and where, with reference to Fig. 4.16,
varying stress distribution over thickness consists of an additional crl' cr2= principal stresses present in the structure before drilling
cutting step releasing the surface layer of the block with the strain the hole,
gauges attached.
Hole Drilling Method : According to this popular and well
established method residual stress is determined by measuring
the relieved radial strain when drilling a small through-hole into
the welded plate. The measuring bases extend across the drilled
hole or are located outside the hole as shown in Fig. 4.15.
To completely determine an unknown biaxial residual' stress
state i.e. two principa~ stresses .crland crzand the principal stress
I· - - - - - - + +- - - - - --+
I II
II
II
I •••
I
·0 • balls
~MeaSUring-..........
II
I
I
I
I
0..I
-...•
• II
Fig. 4.16. Definition ofsymbols.
~------.:- +-------+ El,Ez,E3= relieved radial strains at point P or points PI' P2 and P3
(0) I b) respectively,
+------+
I +-------+ A and B == Constants the values ofwhich for a finite-sized gauge

I,0
are given by the followingrelationships.
I
I . ,
~Strain
I gauges -----I I
I
A-= _ (1 +2Ev) a
10
I
I
- II
I
I _ I
I b
(4 8.)
....
I
I II I
J II Ii = - 2E ...(4.9)
+------+ +-------+
Ie) I d) where, E and v are Young's modulus of elasticity, and Poisson's
Fig. 4.15. Schematic illustration of Hole-drilling method of stress measurement: ratio respectively, for the work material.
(a, b) with measuring balls, and (c, d) with strain gauges.
''''~''_.'''="==-=='''''''''''''''-'''''''------------------------~"
232 Welding Engineerin~ and Technology Welding Stresses and Distortion 233

a and 0 are dimensionless, material independent coefficients 2R2

given by the following equations. r2==-D-
...(4.16)
2 1 where, GW = strain gauge width.
a == . r2 - r1 ) . 2"
r (81 - 02) ... (4.10) GW, D, R1' R2' 81and 82 are defined in Fig. 4.17. Further, from
Fig. 4.17, we get,
.r2 +
-b _- n_2 (1 - v) r 2 [2(1(1+-vv)
r1) . ..!. ) (81_ 82)+ (sm
. 281_. sm 282) R . I d' strain gauge element length
1= gauge C1rcera IUS - 2
- 2" 2 - 2 ••• (4,11)
D GL ...(4.17)
r 2r1 2r2
1 (sin 281. cos2 81 sin 282 - cos2 82]
="2-Z-
where, from Fig. 4.16, and R2 -R1 = GL ..(4.18)
D GWI2
...(4.19)
r == Do ... (4.12) Also, tan 81= R1
and t 8 _ GWI2 ...(4.20)
an 2--R-2-
By substituting the values of A, B, E1'~ and E3in equation
(4.7) the principal stresses (01 and (2) can be determined; they are
located at an angle 13clockwise from direction (1) and (3), when
13is positive. The angle 13is computed from the following
relationship :
E3- 2E2+ E1
tan 213== ---- ••• (4.21)
£3-£1
Positioning of Strain Gauges in a Rosette and with
82 Respect to Weld Centreline. To measure the relieved strains
lbl Weld

Fig. 4.17. Typical three-element strain gauge rosette for the hole drilling method.
where, from Fig 4.17, we have,
D = diameter of gauge circle, 135,)-
Do = diameter of the drilled hole. 03 90·
The size of the hole to be drilled, based on ASTM Standard, is
given by the following condition.
D ...(4.13)
,.p- Slra;n ga"g ••
2.5 < Do < 3.4
The other parameters of equations (4.10) and (4.11) are
calculated in the following way.
W= 2xGW ..,(4.14)
D
2R1
r --- ...(4.15)
I- D Fig. 4.18. Positions of 5 strain rosettes pasted on butt welded
plates to determine longitudinal stress.
Z34 Welding Engineering and Technology
Welding Stresses and Distortion 235

for determining longitudinal (and transverse) stresses along a <1L = + <12 sin
<11 COS 13 13 ...(4.22)
transverse section to weld centreline the strain gauge rosettes are
pasted on the plate as shown in Fig. 4.18. and the orientation of and <1T = <12 cos 13 - <11 sin 13 ..,(4,23)
three gauges in the rosettes is as shown in the inset in Fig 4.18. If 13 is -ve
To determine longitudinal stress (<1L) or the transverse stress
(<1T) it is essential to know the sign of angle calculated with the
13
From Fig. 4.19 (b)
help of equation (4.21). <1L = <11 cos (-13) - <12 sin (-13)
If is +ve
13
z.e. <1L = <11 cos 13 + <12 sin 13 which is equation (4.22)
From Fig. 4.19 (a)
and <1T = <12 cos (-13) + <11 sin (-13)
135°
or <1T = <12 COS <11 sin
13 - which is equation (4.23)
13

If (1 is + ve so in general, (1£ = <11 COS + <12 sin 13 13

and <1T = <12 COS <11 sin 13 - 13

6L ;:: «, Cos (1 + 0-2 Sin (1 where angle is taken in absolute terms.

13

o-T ;:: Cf2Cos (1 - Cf, Sin (1

The accuracy of the hole-drilling method is impaired if the hole
efT
is drilled close to the plate edges or close to other drilled holes, as
the method is normally calibrated on the basis of infinite plate or
atleast largely extended plate. Plastic deformation due to hole
drilling can also lead to erroneous results. This can happen if the
built-up residual stresses arecIose to the yield limit (above 0.6
Y.S.) of the material or if the drilling is carried out improperly,
a; <1"2 Such an error could be between lOand 16%, depending on the hole
2700
diameter. This error could be a severe limitation in measurement
<1"L
of welding residual stresses, particularly in as-welded fabri-
cations, where stresses in and around the weld are typically of
135° (01
«2
yield stress magnitude. However, if the errors are accounted for
the hole-drilling method can be successfully:p.sed for determining
stresses in welded joints.
If n is -ve
Air abrasion method is sometimes used as a means of
machining the hole without the introduction of additional residual
CfL;:: 0-, Cos 1-(11- 0-2~in 1-(11 o
stresses in the region of the strain gauges. It can be successfully
o-T used even on hardened or high carbon steels for which normally
;: cr, Cos (1 + Cf2Sin (1
drilling may be unsuitable. However, the equipment for air
crT;: cr, Sin 1- (1) + 0'2CoS (-(11 abrasion process is not portable and it can be used only with the
;:: 0'2 Cos (1 - 0', Sin (1 workpiece in the horizontal or downhand position.
Example 4.1. Find the longitudinal and transverse stresses
CfL across a transverse section in fiat butt welded plates of HSLA steel
0-,
270° if the radial strains on one side of the weld centre line as measured
lbl along the transverse section by 3-strain gauge rosettes are as given
in table 4.1.
Fig. 4.19. Orientation of principal stresses when
angle ~ is:(a) positive, (b) negative.
~'~.~- ••• =- ===-==""'''''' ••••••••••••••••
-

236 Welding l:ngineering and Technology

Welding Stresses and Distortion 237
Table 4.1. Measured Strains Across a Transverse Section
of Butt Welded Plates From equation (4.14)
S. Distance of strain gauge rosette from Radial strain, microns W=2xGW 2x3:::;06 ...(iii)
No. weld centreline (mm) 10 D .
£1 £2 £3
From equation (4.17), we have,
D GL 10 3
2-
4.
3.

~:Take gauge
5. 132
76
104
:~
7__ ~JI
--T-rr
29
17
-~
~ =~
3
3
22
15
=:~
R1:::; 2 - 2 :::;
and from equation (4.18), we get,
R2-R1:::;GL
'2 :::;5 - 1.5 :::;3.5 mm ... (iv)

.. R2 = GL + R1 :::; 3 + 3.5 :::;6.5 mm ... (v)

width (GW) :::;3 mm, gauge length (GL) :::;:3 mm,
gauge factor:::;2, gauge centre diameter (D) :::;10 mm, resistance of From equation (4.15), we get,
each gauge element :::;.119-1200.. Sketch the longitudinal stress 2 xR1 2 x 3.5
r1 :::; D :::; 10 :::;0.7 ... (vi)
versus distance along the transverse section. Young's modulus of
elasticity (E) :::;207 x 103 N/mm2 and Poisson's'ratio (v) ::::0.3, for and from equation (4.16),
the plate material .
Solution. _2R2_2x6.513 ...(vii)
D -
r2 - 10 .
I STEP From equation (4.19), we get,
To determine the drill bit diameter (Do) for drilling holes. tan 81:::;GW/2 = 3/2:::;~
Using equation (4.13) , we have R1 3.5 7
D .. 81 :::;23.19° ...(viii) -----

and from equation (4.20), we have,

2.5 < Do < 3.4
Let, Do = 3.17 mm (i.e. the available drill bit size) t 8 _ GW /2 _ 3/2 _ ~
an 2 - R2 - 6.5 - 13
.. 82 = 12.99° ... (ix)
.. g~:::;3~~7 = 3.14 which lies between 2.5 mm and 3.4 mm Subtracting (ix) from (viii), we get,
Now, therefore from equation (4.13), 81 - 82:::;23.19° - 12.99°:::; 10.2°:::;0.178 radians ...
(x)

go ::::3.14
r:::; (i) Now, substituting values of W, r, rI>r2' 81, 82, v and E in
and Do = 3.17 mm (ii) equations (4.10) and (4.11), we get,
Do=3.17mm 2 1
r = 3.14 a ::::W (r2 - r1) . '2
r' (81 - 82)
II STEP 2 1
:::;0.6 (1.3 - 0.7) . (3.14)2 x (0.178)
Once Do is known, holes can be drilled at the centres of the
rosettes pasted on the. plate; and thus radial strains can be :::; 2 x 0.178:::;01 ... (xi)
0.6 x 0.6 (3.14)2 .
measured. To calculate the principal stresses, and their directions
it is required to use equation (4.7). To employ this equation we are
to know constants it and B, and to determine these constants it is -b -_ m2 (1
- + v) . . 1:.. (1+- v)Y2 (81 - 92) + (sm
r2 [2 (1 .. 281 - sm 282)
required to know the following parameters.
W. rl' rz, Rl' R2' 81 and e2
r2 [Sin
_ 1:.. 2e12r~
. cos2 91 _' sin 2822r~ ~
. cos2 82]~
238 Welding Engineering and Technology Welding Stresses and Distortion 239

Putting the values of E1' £Z, E3 from Table 4.1, and those of
b = 0.62 (1.3
(1 + -0.3) 1 [2 (1(1-+ 0.3)
0.7) . (3.14)2 0.3) x (0.178) constants A and B from (xiii) and (xiv) respectively, we get, for
+ (sin 2 (23.19) - sin 2 (12.99) - 1 2x point 1 at 20 mm from the weld centreline,
(3.14)
a = [(-69) + (-72)] x 10-6 + . -.f2 x 10-6 X
1 4 (_ 0.314 X 10-6) 4 (- 0.755 X 10-6)
2x(0.7)2
(sin 2(23.19) 2x(1.3)2
. cos2 ~3.191_ sin 2(12.99) . cos2 (12.99)] "«-69) - (_2))2 + «-2) - (-72))2
= 141 X 10-6 + x 10-6 ."f2 ~ n
or 1 2 [2 ~ .<~.7 x 0.178 + (sin 46.38 - sin 25.98)
. x.1036. (3.14)
b = 026X 4 (0.314) X 10-6 - 4 (0.755) X 10-6 ( )

= 141 X 10-6 ."f2 X 10-6 V

_ (3.14)2 2 x 0.49
1 (Sin 46.38· 2 x .1.69
cos2 23.19 _ sin 25.98 cos2 12.99J~U 4 x 0.314 X 10-6 4 (0.755) X 10-6

= 0.314 - 0.755 x 96.8969 ]

10-6 - [141."f2
= 0.7325[0.1917 + (0.7239 - 0.4381) __(3.14)
1 <>

1 267.54 2
='4 [449.04 - 181.50] = 4 = 66.885 N/mm ...(xv)
(0.7329 0.98
x (0.9192)2 2 x 1.69
(0.4381) (0.9744)2JnU
--=- r:--
a2 = E1 + £3
4A - --= £z) 2 + (£2 - E3) 2
."f2.
4B ",«E] -
--1-i
= 0.7325 [0.1917 + 0.2858 - (3.14) (0.6241- 0.1231)]
From the above calculations, we get,
= 0.7325 [0.1917 + 0.2858 - 0.Q508] 1
or b = 0.3126 ... (xii) a2 = '4 [449.04 + 181.50]
III STEP = 63~.54 = 157.635 N/mm2 ••• (xvi)

Evaluate A and 13 using equations (4.8) and (4.9) respectively,

at = 66.885 N/mm2
A--
- 2E
II
+ v) a (48)
.... V STEP
a2 = 157.635 N/mm2

Now, E = 207 X 103 N/mm2 and v = 0.3, Now, longitudinal stress (ad will act at angle [3,which can ,be
A = _ (1 + 0.3) x 0.1 = _ 1.3 x 0.1 _ found from equation (4.21),
., 2 x 207 X 103 2 x 207 X 103 E3 - 2ez + £1
or A = - 0.314 X 10-6 mm2 IN ...~xiii) .. tan 2[3= (E3 - E1)

and B=--2E
- b
from (4.9) [(-72) - (2 x -2) + (- 69)] x 10-6
- [(- 72 - (- 69))] X 10-6
_- 0.3126 = _ 0.755 X 10-6 mm2/N ...(xiv) = -72 + 4 - 69 = - 137 = 45.67
2 x 207 X 103 -72 + 69 - 3
IV STEP 2[3= tan-1 45.67
2[3= 88.75°
Evaluate principal stresses a1 and a2 by employing equation or 13= 44.37° ...(xvii)
(4.7). VI STEP
.. --=-
4A -=
...J2" - ' :r:--;::-~
a1 = E1 + E3 +
4B (101 + £2) (E2 - £3) Evaluate lorlgitudinal (ad and transverse
equations (4.22) and (4.23) respectively.
(aT) stresses using
~

240 Welding Engineering and Technology Welding Stresses and Distortion 241

From equation (4.22), we get, I

.....
<h = al cos J3 + a2 sin J3 ~
~ 0
\/I
~E
z'Ein...J-45
C>I

~~-~
'Q
.£E\/I

Putting the values of al, a2 and from (xv), (xvi) and (xvii)
J3
-15 -60
'Q
-30
§
-75 15
respectively, we get, .~ 130
I
aL = 66.85 cos 44.37° + 157.635 sin 44.37° <t
= 47.79 + 110.23
= 158.02 N/mm2
Ef) TensillZ
From equation (4.23), we get, strlZS!?
aT = a2 cos J3 - al sin 13,

Putting the values of al, 0'2 and 13, we get,

aT= 157.635 cos 44.37° - 66.85 sin 44.37°
= 112.68 - 46.75
= 65.93 N/mm2 20 40 80 100 120
Distance. mm _
Following the above procedure we can calculate aL and aT for
different points in table 4.1. The final results obtained are shown
in table 4.2. e CompressivlZ
stress

Table 4.2. Calculated Values of ah a2' 13, <h and

O'T for the Strains Lis.ted in Table 4.1. Fig. 4.20. Longitudinal welding stress (aLl versus distance
from weld axis along a transverse cross section.
S. -]0
10 -1 158.02
166.85
·-55.74
22
-7230'1
-2
Drilled
104
£2
£3 and
aT48
65.93
CJ2 -25.28
89transverse
-36.47
-18.61
39.34
1327
19.6910.87
157.64
-25.46
41.43
-58.90
-40.57
19.75
20
76
8.82
distance
-19.75
-42.80
8.97
-69 17
16.91
44.37 -16.81
-8.45
holeMeasured
radial strains
x (N/mm2)
Longitudinal
Principal
aL 10~ stress
Angle stresses (N/mm2)
E (~O)
complete biaxial state of relieved strain when the material is
I 151-34.14 removed in layers. The care required in layer removal in the case
of the plate is even higher than in Case of the beam. Because the
clamping forces increase as the layer removal progresses thus the
risk of yield limit being exceeded also increases. It is, therefore,
recommended to remove layers only up to approximately the
centre plane of the plate.
Moire Fringe Technique : Some attempts have been made
to apply any of the moire fringe on holographic variants to the
precise problem of residual stress measurement. The major
disadvantage of conventional Moire patterns is that the sensitivity
obtainable from the finest practical grid of 40 lines Imm is too low
The graph between longitudinal welding stress (aL) and the to be of real use in the measurement of elastic strains, although
distance from weld centreline along a transverse section is shown it is particularly suited to the monitoring of plastic strains such
in Fig. 4.20. as those which occur during welding.
The Resilience llJethod extended from bar to plate problems is For application of moire fringe technique to measurement of
particularly suitable for biaxial, homogeneous and depth-variable residual stresses the experimental procedure is again destructive,
welding residual stresses. Knowledge of the spring-back defl.ection but as a laboratory tool it has interesting potentialities with its
alone, however, is not sufficient for solving the biaxial problems. ability to monitor a large area at one go.
Strain gauge rosette~ are therefore used, which indicate the
242 Welding Engineering and Technology
4.3.1.3. Measurement of Triaxial Welding Residual
Stresses
Measurement of triaxial residual stresses involves the
fundamental difficulty that the volume of the relevant material to
be stress-relieved is relatively inaccessible in the interior of the
component. Nevertheless, certain methods can be applied here
also if the principal directions of the residual stress state are
known beforehand, e.g. if they coincide with the main directions
of the geometry of the body.
The simplest step for determining the triaxial stresses consists
in additionally measuring the change in thickness when applying
the block cutting method. It is then possible to determine the
average stress perpendicular to the plane of the plate. However,
there is another three-dimensional cutting method in which the
averaging of residual stress value over the total plate thickness is
avoided.
Three Dimensional Cutting Method : In this method
narrow deep blind holes are drilled to the point of the residual
stress to be measured, as shown in Fig 4.21 (a). This is followed
by positioning miniaturized longitudinal strain measuring
yz\ ~~/)f.z fRihrgro~e"'"Y7m~~
;:;ai!' gal:gef~
~~1~
F/~16/ ;o~I:!:/)
-170':: .J13'''-'~
wPkfZ2a
(b I Hole fitted
elements
with measuring
(al Blind holes
Fig. 4.21. Arrangement for measuring triaxial residual stress in t.he interior of
the component using:(a) blind holes, (b) hole fitted with measuring elements;
strain relief is achieved by cutting ring groove by a trepanning drill.
elements in them. A slightly pre-tensioned bolt with measuring
shank and measuring gauge is a suitable measuring element; and
it is placed in the hole as shown in Fig. 4.21 (b). The cylindrical
material containing the drilled holes alongwith the meas\Jring
elements are then machined out, say, by trepanning. The strain
released in the operation is, thus, measured from which stress can
be calculated.
The axial directions of the three drilled holes are made to
coincide the principal directions of the residual stress. If the two
Welding Stresses and Distortion 243
r
holes in the horizontal plane of the plate (along x and y directions)
become too long to be drilled, then these are replaced by the
inclined holes as shown in Fig. 4.21 (a). As the vertical strain Ez
is known, thus from the relieved strains in the inclined holes (i.e.
Eyz and Exz) the horizontal strains Ex and Ey can be determined by
the following relationships.
1
Exz;:::"2 (Ex + Ez) ... (4.24)
1
and Eyz;::: 2" (Ey + Ez) (4.25)
..•
On one hand it is required that the measuring elements should
be as close as possible to each other to determine the triaxial
residual stress at a given point; on the other hand, however, they
should not be too close to each other as otherwise they may cause
mutual strain relieving. The distance between the ends of the
holes should be approximately 10 times the diameter of the hole
in order to limit this interaction effect. Drilling the hole and
tightening the pre-tensioned bolt in it cause measuring errors
which should be minimised by calibration or by comparative
evaluation.
To improve upon the three hole-cutting methods an attempt
has been made to use only a single hole and apply strain gauge
rosettes to the bottom and wall of the hole.
/)
4.3.2. Non-destructive
Methods
Residual Stress Measurement
Amongst the non-destructive methods used for measuring
residual stresses in weldments, X-ray technique is the most
popular at present.
4.3.2.1. X-ray Method
X-rays are employed for measuring stress in the crystalline
materials. It involves X-ray diffraction measurement of change in
interplanar spacings of a stressed metal which can lead to
information on residual stresses. In simple terms, the lattice
spacing of a particular plane (the actual plane is a function of the
radiation wavelength) is proportional to the angle of reflection of
the X-rays, which is determined either by a back reflection camera
or with the help of a diffractometer, as shown in Fig. 4.22. The
angle of reflection is used to determine the change in the
interplanar spacing compared to the stress-free state; that in turn
is used to determine the sum of the principal stresses at the
surface of the welded material by using the Bragg's Law,
 Welding Stresses and Distortion 245

244 Welding Engineering and Technology r 1

Ex = E (ax - v (ay + az)
1
'-
.&
c
::>
~I QI Ey = E [ay - V (az + ax) I ... (4.27)
8 Vi
1
o c: ;;-
Ez =E [az - V (ax + a) -'
-
+=
'-.2
0;-::
where, Ex' Ey and Ez are the strains in the three directions and ax'
~.§<II -0-
o
u...
..s::
"'"'
ay and az the corresponding stresses. E and v are the Young's
e..§
modulus and Poisson's ratio respectively of the material being
oc:<;::
c:
welded.
0.0. 0
From the above expressions (4.27), the sum of the principal
-
c:
:D
<.> Q)
o:l b.O
..l:: o:l
~ 8
-0 .-
stresses in a surface can be easily determined. For example, if a
~:5 >,~ steel specimen is subjected to principal stresses 0'1 and 0'2 in its
VQl o:l •••
.S.c surface plane, the stress 0'3 normal to that surface is zero. The
~~
>, •.•..•
..0":: normal strain E3' perpendicular to the surface is then given by,
"'"' .-
'- c:~ v
.& Q) Q)
8-0 ...(4.28)
E
QI
Q) ~
•.•-0 + 0'2)
£3 = - E (0'1
'-v.e
00'- :l~
E_
-:::~ II "0
QI
'"
00-0
o:l c:
Q) o:l
and represents a contraction if 0'1 and 0'2 are tensile stresses or an
~"C
-'"
'-QI
'"
c:
8 ~
00
00
Q)"'"'
••• Q)
"'"' 00
0
:l
elongation if they are compressive. The strain E3, is generally
measured by determining the change in interplanar spacing d
::>
00 o:l parallel to the surface under consideration i.e. by taking back-
'" <; ..•...
" QI - :l 0 reflection diffraction patterns in the stressed and unstressed con-
!,r/ oc: 0_ -0
.5
"Cii
Q)
ditions. Precision measurements of the patterns yield interplanar
Ci.
v ~-;;:l spacing ds (stressed) and du (unstressed) for use in the following
o'- /1/
'" -
'E
o
c-iO
C'l ~
.q<.g
relationship,
-a. '"
QI-
QI <l:
E c:"O
..9
o QI
'" ~
cD
0"1 +0'2 =----=------
E d E (ds
!:1 - du)
...(4.29)
.--
v '-
E '"
<lI
v d v du

~- Surface stress in any particular direction may also be

nt.. = 2d sin S
where n = the order of reflected beam, (1, 2, '" etc.)
d =- interplanar lattic spacing,
determined from the two sets of counts (or X-ray photographs) of
...(4.26) the stressed material, one with the beam perpendicular to the
surface and another with it inclined at a known angle to the
surface and lying in the vertical plane fixed by the direction of
interest on the surface. If stress, a~ at point 0 in the direction <l>

S = Bragg angle. of the metal plate of Fig. 4.23, is desired to be determined, it can
The val'Ie of interplanar spacing, d, is used in the elastic be obtained from diffractometer counts (or x-ray photographs)
stress-strain relationship to determine the sum of the principal taken along the Z-direction and 'l' direction. Thus, in terms of
!"tresses. interplanar spacings dz and d'l', the stress in the direction can <l>

Elastic Stress-strain Relationship be expressed as,

_ dlJ,f- dz E
For perfectly elastic homogeneous and isotropic materials, the ...(4.30)
strains for the three dimensional stress system are given by the a~- dZ (1 + v) sin2 'l'
following equations.
246 Welding Engineering and Technology
z
y ractometer, it is necessary to have best focussing condition when
Fig. 4.23. Angular relationship for determining stress component in ljl-direction.
Experimental Procedure
The Bragg ar.gle 8 can be determined, as already stated, by
two methods viz.,
(i) Film method, and
(ii) Diffractometer method.
(i) Film Method: An X-ray striking the surface of the
component produces, after reflection, interference which appears
as an interference ring on the film (Fig. 4.22 (b)). The interference
line of the first order, i.e. closest to the centre, satisfies Bragg's
law with n = 1, thus angle 8 depends on the interplanar spacing d
of the atoms in the crystal lattice and on the wavelength A. of the
X-radiation, the practicable limit for which is 0.002 to 2.0 A, where
A = 10-7 mm.
The Bragg angle 8 can be determined with high precision from
the radius r of the interference ring and the spacing af between
the specimen and the film, by the relation,
8 ~ ~ arc tan (- ~f) ...(4.31)
(ii) Diffractometer Method : The general principles for the
diffractometer method are the same as for the film method. The
specimen surface under examination is mounted in the
parafocussing* position whenever possible in order that the
diffraction peaks will be as sharp as possible. Determination of
the sum of principal surface stresses follows the usual back
reflection technique at an angle 28 which is nearest to 180°. For
back reflection sin 8 = 1, therefore the Bragg's equation
(nil. = 2d sin 8) reduces to nil. = 2d. Now, n can be made unity by
setting the position of the counter tube at first order deflection .
.• beyond focussing position .
Welding Stresses and Distortion 247
On this basis proper radiation involves a choice of A. slightly less
than 2d. On these basis cobalt Ka radiation with a mean
wavelength of Ka doublet 1.7902A is used. With cobalt radiation
usually (310) plane is scanned having 29 nearly equal to 162°.
However, if cobalt radiation source is not available the next best
would be to use FeKu radiation and therefore (hkl) plane having
maximum value of 28 = 145.8° is (220) and the same can be
scanned for measurement of residual stresses.
For the determination of directional stresses in a diff-
the radiation strikes the specimen at an inclined angle such as
the usual 45° angle. This can only be obtained by moving the slit
of the X-ray tube, since at this parafocussing condition the radius
of the focussing circle is much shorter compared to the distance
between the specimen to the slit of the X-ray tube. In some units
this adjustment is not possible and the two distances viz.,
specimen to slit of X-ray tube, and specimen to the slit of the
counter tube, are fixed and equal. In such cases the usual back
reflection technique is employed to determine the interplanar
distance between the atoms.
The measuring area covered in X-ray stress measurements is
0.1 to 1 mm2, the measuring depth being approximately 10
microns*.Although the main advantage claimed for X-ray residual
stress measurement is that it is non-destructive but in practice a
piece has to be cut from the welded joint so as to be accommodated
in the work holder of the diffractometer. Even the conventional
unit with camera and X-ray source cannot be made so portable as
to be moved around weld seams. Thusthe non-destructive nature
of the X-ray method should be considered within these limits.
4.3.2.2.Neutron Diffraction Method
This is a more recently developed non-destructive method in
which neutron beam is used. Neutrons are scattered by the atomic
nuclei whereas X-rays are scattered by the electron shells, with
the consequence that neutrons penetrate far deeper than X-rays
e.g., up to 3 cm in steel or 30 cm in aluminium alloys. Thus,
stresses or strains can be measured in the interior of the
components. Typically, a neutron beam from the reactor core
(approximately 50 mm x 50 mm) is transmitted by a mono-
chromator crystal. Slits, in absorbing cadmium masks are placed
between the component and both the incident and the diffracted
beams, as shown in Fig. 4.24, in order to reduce the volume of the
.• 1 micron = 0.001 mm.
 249
248 Welding Engineering and Technology Welding Stresses and Distortion

Soller
4. Greatly differing residual stresses may occur on the
collimator surface and interior parts of a welded component, par-
ticularly in the case of transforming alloyed metals.
5. Considerable differences exist in respect of the residual
transverse stresses between the longitudinal weld between
strips and the circumferentialweldofa cylindricalor spherical
shell,as a result ofthe constrictionofthe circumferentialweld.
6. Particularly unfavourable state of residual stresses
occur in many cases at weld ends.
7. Considerable differences exist in respect of the residual
stresses between the single-pa,ss weld and the multi-
pass weld; because in a multipass weld all runs except
the one laid by the last pass are normalised by the post
weld heat treatment caused by the subsequent passes.
Although the above factors apply to all fusion welds for
residual stress condition in all directions but it is customary to
study and analyse the longitudinal and transverse stresses in butt
welds, and radial and transverse stresses in circular or ring welds.
Detector
4.4.1. Longitudinal Residual Stresses in Welds
Fig. 4.24. Schematic set-up for Neutron diffraction method of residual stress
measurement as applied to electron beam welded pipe.
material scanned for determining the residual stress. Parallel
scattered beams originate from the crystallographic planes which
have the correct orientation for diffraction. By studying both the
peak location and intensity of the diffracted beam, data on stress
or strain and texture can be obtained.
Other non-destructive residual stress measurement methods
like Ultrasonic Method and the Magnetostriction Method are still
-::J ( a 1 Mild steel. TItanium ( b 1 Aluminium
in developmental stage and it will take sometime before these are
available for widespread application.
4.4. RESIDUAL STRESSES IN SPECIFIC' MATERIALS
AND JOINTS
Residual stresses vary considerably depending upon the type
of joint, welding parameters and the material being welded; '-
however the following statements apply irrespective of these ....

factors.
1. Particularly high residual stresses occur in the weld
metal and HAZ of the fusion welds.
(c 1 High alloy ferritic steel weld I d) Austenitic fIller metal
2. The maximum residual stresses generally reach the
Fig. 4.25. Longitudinal stress patterns for butt welds in different materials.
yield limit of the material being welded. (a) Mild steel as well as titanium, (b) Aluminium, (c) High alloy
3. The residual stresses may vary greatly from point to structural steel welded with ferritic filler metal, and (d) High alloy
point in the weld metal and HAZ. structural steel welded with austenitic filler metal.
"r,.",.,..
••·•• ----- ..------ .. - ...

250 Welding Engineering and Technology Welding Stresses and Distortion 251

LOIlf,ritudinalresidual stresses in welded joints are caused as Table 4.3. Maximum Values of Longitudinal
a result of heating and cooling down of the weldment; in certain Residual Stress in Welds Compared with
cases they are superimposed by opposing transformation stresses. Yield Limit for Different Base Metals
Distribution patterns of longitudinal residual stresses developed S. No. Base Metal Max.
jn centrally placed butt weld between two strips are shown in Fig. Yield Limit, Oy
Longitudinal
(N/mm2)
'II
4.25, for different materials. Residual Stress
In mild and low alloy steels, stress distribution occurs with (CJL max ), (N/mm2)
high tensile stresses, reaching up to the yield point of the material 1 Mild and low alloy steels 210-240 210-240
being welded, in the weld metal and lower compression stresses 2 Austenitic steels, non-hardened 280-300 280--350
alongside it. The complete stress distribution pattern takes the
shape of letter W, as shown in Fig. 4.25 (a). Such a distribution 3 Aluminium alloy (AlMg6) 160 80-120
also occurs in the case of titanium, however the maximum residual 4 Titanium alloys 500-700 300-400
stress developed in that case is somewhat below the yield lim~t.
In aluminium alloys also the maximum stress developed is ~160
'"
below the yield limit, but with a slight trough in the stress pattern ~
corresponding to the weld metal zone, as shown in Fig. 4.25 (b).
In high-alloy steels with ferritic weld metal, the stress in the
""
~

'~NE
80

weld centre is displaced into the compression range, taking an M .:: EO

o -
cZ
shape, as shown in Fig. 4.25 (c); this is as a consequence of the 'g Mild steel
austenite-ferrite transformation at low temperature. If, on the ·~-80
c
o
other hand, an austenite electrode is used, the yield limit, cry, of ...J

the weld metal is reached. A compressive stress trough occurs to

the right and left of that in the HAZ,caused by the transformation a=B
of the base metal after getting heated to above the transformation
temperature Al' Farther to the outside, the relatively high yield
limit, cry, of the base metal is reached as a consequence of heating
to below Al temperature and the stress pattern takes a shape of
160x8
letter W ext.ended on both sides, as shown in Fig. 4.25 (d). Even
farther to the outside, an additional reversal into the low tension
range may occur as a consequence of the complex formation of -.0:
1"////~~////'l;J -160 -80 0 80 160 240
III
residual stresses. ~ Longitudinal residua\stress, OJ N/mm2
~80
The areas of particularly high tensile stresses prone to cold
cracking occur in or close to the weld bead in accordance with the ""
'0

'~NEO
stress patterns presented in Fig. 4.25, depending on the base and ~ E
the filler metal combination and process parameters employed. gz
'6
Maximum values of longitudinal residual stresses relative to ~ 80
'"
co
the yield limits for weldments in different materials are as listed ...J

in Table 4.3. 160

Fig. 4.26. Longitudinal residual stresses in welded mild steel

I-section girder with web-to-flange fillet weld.
252 Welding Engineering and Technology
r Welding Stresses and Distortion 253
The reduction in the maximum stress compared to the yield
limit observed in the case of aluminium and titanium alloys is 4.4.2. Transverse Residual Stresses in Weldments
attributed to the Shear Warping of the cross··section as the heat
source passes through it. The warping, in turn, is the result of the Transverse residual stresses arise directly as a result of the
elasticity of the component. The effect is particularly severe if the transverse contraction of the weld as it cools down, and indirectly
welding speed is low. as a result of its longitudinal contraction. In addition trans-
The longitudinal residual stresses drop to zero at the edges of formation stresses may have some influencing effect. When a
the transverse section or at the weld end. As per St. Venant's rapidly deposited butt weld in transversely restrained long plates
Principle based on theory of elasticity, such a stress drop takes cools down, it results in a transverse residual stress pattern of the
place over a distance that corresponds approximately to the width type shown in Fig. 4.28 (a). But if such a weld is deposited in short
of the plate. However, in comparison the reduction in longitudinal plates the residual stress pattern is of the type shown in Fig
residual stresses in plastically deformed zones takes place more 4.28(b).
rapidly.
The longitudinal residual stresses in three elements of welded I 1 I
I-section (mild steel girder) with web-to-flange fillet welds are
I I I
I II I
shown in Fig. 4.26 while those for low alloy structural steel I I I
I-girder are shown in Fig. 4.27. I ~ I I
I ~ I I
I :;; , I
1'0 ~~ I I
longitudinal
residual stress
Structural steel
I~ e ~
I
cr. max =330 N/mm2
t::: ..:-..:..-:. .JI L tt;; J
(a) I b) (c ) (d)
Outside
of flange
+
I
II
Inside I
of flange I
I
(e)

Fig. 4.28. Transverse residual stresses in rapidly deposited butt weld, (a) long
330 plates, and (b) short plates; in slowly deposited butt welds,(c) long plates, and
(d) short plates; and (e) residual stress pattern for stresses in thickness direction.

A weld produced with slowly moving heat source, by contrast,

results in high transverse tension at the weld end thereby possibly
causing incipient cracks. Transverse residual patterns for slow
welds in long and short plates are shown respectively in Fig. 4.28
(c) and Fig 4.28 (d). The transverse residual stress pattern in the
thickness direction (O'z) along the centreline of a weld is shown in
Fig. 4.28 (e).
600
The longitudinal, and transverse residual stresses (3-d) in a
welded joint with V-groovebutt weld are shown in Fig. 4.29.
Fig. 4.27. Longitudinal residual stresses in welded low In plates with restrained edges, the restrained transverse
alloy structural steel I-section girder.
contraction of the weld and surrounding areas determines the
transverse residual stresses.
254 Welding Stresses and Distortion 255
Welding Engineering and Technology

200 I ness
NiCr Mo I;tnl
z Plot!! thick 6 mm
Welding speed = 6Smm! sec Electron beam weld
0 ~ 100
I \
J ~
'"
I '"
I \ I
<U'"

I \ I
•... E
""Vi
"-
E 0
I \ I 0%
I \ I ::>
::! -100
'"
I
(a I Longitudinal residual stress. cJx
--1-
Iy <U
•...

e: -200
•...
>
<U

'"
§ -300
•...
•....

-400
-10 -5 0 5 10 mm
Distance from weld. y
Fig. 4.30. Surface transverse residual stresses in electron beam
butt welded alloyed martensitic hardened steel.
The variety and complexity of possible distribution patterns
Transverse
shown in Fig. 4.28 in respect of longitudinal residual stresses is
230 N/mm2 also true for transverse residual stresses.
residual stress. rJ'y l--j
z The transverse residual stresses in the circumferential weld
of cylindrical or spherical shells differ essentially from those in
I
flat plates as a result of the super imposed bending stresses
I I caused by the constriction of the shell. A characteristic feature is
I I the tendency to transverse tension at the weld root.
I I
I I In multi-pass butt welds in thicker plates the residual stresses
1 I are irregularly distributed over the plate thickness. For example,
I the last passes to be deposited experience high shrinkage
I
L
Tronsver se
,
-----II L--100
I N/mm2
---1- y resistance in the longitudinal and transverse directions (in the
plane of the plate), with the result that not only high longitudinal
residual stresses occur but also high transverse residual stresses
reSidual stress .rfz
are developed though reduced somewhat by preheating as a
Fig. 4.29. Longitudinal and transverse residual stresses in consequence of the passes deposited earlier. The passes deposited
thick plates with V-grt'ove butt weld. first are subjected initially to tension, and then latter to
The conditions are more complicated with metals subjected to compression, superimposed by strong bending effects in the
microstructural transformation, and a reversal of the stress sign transverse direction. Transverse tension 'occurs at the root of the
may occur. For example, Fig. 4.30 shows the transverse residual multi-pass butt weld deposited from one side where angular
stress pattern, based on X-ray stress measurements, in the surface shrinkage is unrestrained, and transverse compression occurs
of an electron beam welded plat.e made of an alloyed, martensitic where angular shrinkage is restrained. The residual stresses in
hardened steel. The stresses may vary greatly within a confined the direction of plate thickness, by contrast, are negligibly small.
area as a consequence of the concentrated high density heat input The triaxial tensile residual stress state is largely avoided by
and high welding speed. multi-pass welding.
 Welding Stresses and Distortion 257
256 Welding Metallurgy and Joint Design

The rises and drops of the transverse residual stress, during 4.4.3. Residual Stresses in Circular Butt Welds
multi-pass welding of test specimens, in rigid restraint hot Circular welds are often encountered in the fabrication of
cracking test are shown in Fig. 4.31. The reaction stress in the various components e.g. welding of pipes, nipples and patches to
plates. This results in the development ofresidual stresses making
it imperative to analyse them before the component is put to
Restraint varied /"\ service.
by clamping
of altering spacing-l'
bars I II-- Theoretical study involving function-analysis approach on the
basis of the equilibrium and yield condition for the instan-
taneously deposited ring weld in mild steel circular plates leads
to the results shown'in Fig. 4.32. A distinction is made between
r-
I -~--~--,----,
I
~
' ~~
\ I
I
I ,\ \ I
I ~\ \ I
I ~ \ r I
I Ro I
: I
I I I I
(a) Set up for Jig cracking test
IL I~ /II/ I IIII I
400 ....1 .J
SHucturol 51<:«1
(0)
t : 20 mm . V 9roov~
crt
I C
300
~~
(c) II
10<%
- r
VI
~
'- •... 200 r--Ri -+Wpl
~E
c:..§
I Ro
n
oz 1.

t;o 100 (b I
'"'
0:: Fig. 4.32. Residual stress profiles transverse to ring weld in circular
plate of mild steel, (a) three zones marked on a plate, (b) radial and
0 _ tangential residual stres& profiles, and (c) three possible variants of
012345678 residual stress profile in circular welds.
Number of weld layers, Ln -
three zones, the inner zone (zone I), the weld zone (zone II), and
(b) Reaction stresses ,in a butt weld the outer zone (zone III). The overall residual stress pat~rns for
Fig. 4.31. Schematic illustration of a set-up for butt welding under varied the radial (crr) and the tangential (crt) residual stresses are shown
restraint and the corresponding reaction stresses for a specific set of welding in Fig. 4.32 (b). Tangential stresses close to the yield limit and
conditions for a multilayer weld joint in rigid restraint hot cracking test, lower radial stresses, increasing towards outside, occur in the weld
weld results from correlating the measured reaction force to the zone; Fig. 4.32 (c) shows t.hree possibilities of stress distribution
weld cross-section which obviously increases as the number of in this zone.
passes d'eposited increases. The reaction stress is highest at the Further, Fig. 4.32 (b) shows that low biaxial tension or
third and fourth pass and exceeds the yield limit of the base metal. compression prevails in the inner zone (zone I), radial tension and
Consequently, enlarged straining lengths are necessary only up to tangential compression (decreasing towards outside) prevail in the
the fourth pass in the component which is simulated in the test. outer zone (zone Ill).
258 Welding Engineering and Technology Welding Stresses and Distortion 259
400
Heat input weld inner zone and wide outer ring, tensile residual stresses occur in
U6KJ/mm
the inner zone as a consequence of high heating up, whereas with
0"8 -tangential stress
Weld
a larger, less stiffinner zone with a narrow outer ring, compressive
O"r- radial- stress residual stresses are produced in the inner zone as a consequence
,
o
@Stortof
Plate thickness
30mm
14mm of the contraction of the ring weld.
Results of experiments on circular 60° single V butt welds
Q.
'2:
made in 14 mm thick mild steel plates of 200 mm outside diameter
=:
~
a
100--80 80-'00 are presented in Fig. 4.33 for 30, 60 and 90 mm diameter welds.
;;; I The V grooves were filled by CO2 welding using three successive
-100 welding runs with room temperature as the interpass temp-
Radial ditlance ,mm
erature. Residual stresses were measured along a diameter
-200 passing through 90° position with respect to the start of the weld
la I Weld circle diameter of 30 mm
400 which represents the point of almost the severest stresses. Both
Heat input radial and tangential stresses were measured after machining off
300 1·36KJ/mm the reinforcement because such. an action is reported to have a
o Weld
negligible effect on the residual stress level but makes it much
~ 200 60mm
easier to determine them.
••
@St~~ldOf
~ 100
Plate thickness 14mm Further, Fig. 4.33 shows that both radial and tangential
~ residual stresses were maximum around the weld zone. The radial
stress reduces to almost zero as it approaches the outer edge of
~
';;' 01 100 20 9 20 the plate but all through it remains tensile. Tangential stress, on
~-100 Radial distance. mm the other hand, reduces steeply as it proceeds towards the outer
edge and changes over to compressive stresses before reaching
- 200 I b I Weld circle diameter of 60 mm
almost zero at the outer edge of the plate. The stress levels, both
weld for radial and tangential stresses, increase with the increase in
400 weld diameter till the yield limit ofthe material is reached, beyond
Heat input
1.36 KJ Imm which no such increase is possible instead plastic deformation
Weld takes place.
t 300
0..•• ~Startol . 90mm 4.4.4. Residual Stresses in Flame-Cut Material
~ 200 Plate thickness 14 mm
•. O"r
The longitudinal residual stresses in the flame-cut edge can
~ 100 be calculated from the elastic resilience of related plate/strip
;;; resulting from the removal of edge layers. Compressive residual
o stresses occur in the flame-cut surface of carbon steels as a result
~ a o 20 40 100
a:
..
';It

distance ,mm
of martensitic hardening, however severe drop in com'pressive
- 100 I stress and in hardness are observed over a distance of appro-
ximately 0.5 mm normal to the edge while further inside tensile
-200'- Ie) Weld "
Circle diameter of 90mm
residual stresses prevail as is shown in Fig. 4.34 for flame-cut
structural steel. The level of the compressive residual stresses in
Fig, 4.33. Variation of radial (O'r) and tangential (aa) residual stresses along a the flame-cut edge increases with the cutting speed and decreases
diameter for 60° single V butt welds of 30.60.90 mm weid diameters. with the heat output of the flame. No hardening occurs in the case
The actual stress profile depends on the stiffness of the inner of wrought iron (a-iron) and thus no compressive residual stresses
and outer zones, on the ring weld diameter, as well as on the are developed in the flame-cut edge of this material; only tensile
welding and material parameters. For example, with a small, stiff residual stresses originating from thermal strains are produced.
~,.~-'

Welding Stresses and Distortion 261

260 Welding Engineering and Technology
o.-
L
.-~I--------l 300
b-
300
~i 200
,r-------.
1 Structural
,
steel I
I ••

l~
:-~-
:
, I .; 100
1- Flame -cut
o I
o I Flame cut edge
I
I g""e
'0 e 0
Z~200
0"
2-Flame-cut
3-Flame-cut
and annealed
and cold
": i
•.. -100
..• stretched ;;c:
..•
QI
L-
I I ] -200
~ 100 L J 'go - 300
-0 3 2 mm 4 6 B
::>
"0 . I I ~250-~ o
Normal distance from edge.y
'in
QI
L-
\ . " "'J_annealing
After .... L_- r
650'C 10·5 ..1_-
h 2 Fig. 4.35. Longitudinal residual stresses at flame-cut edge of mild steel plate.
oc:
'0
0
.... -" "-
,~ reasons, results in a stress reversal into the tensile range at a low
.-
::>

'0-
c:;
o
....J
-100
'!T-
After cold
-200
o 2 mm 4 6
Normal distance from flame-cut
Fig. 4.34. Longitudinal residual stresses in flame-cut edge of
structural steel with different treatments.
Fig. 4.34 also shows the effect of subsequent therm.al or
mechanical treatment given to flame-cut edges of structural steel.
The residual stress profile perpendicular to the flame-cut edge is
shown in Fig 4.35. It changes little as a result of preheating up to
150°C during flame cutting, but hardening is reduced. Stresses
and hardening are reduced as a result of subsequent sub-critical
annealing (650°C/0.5h); as a consequence of cold stretching (5 to
10%) high tensile residual stresses occur in the flame-cut edge
(Fig. 4.34). The stress reduction to be expected after cold
stretching of ):lomogeneousmaterial does not occur here because
the yield st17ElSSis heavily increased in the flame-cut edge as a
result of hardening. The machining of the non-post-treated
flame-cut edge, which is often undertaken in practice for various
machining depth of approximately 0.5 mm.
4.5. METHODS OF RELIEVING (OR CONTROLLING)
WELDING RESIDUAL STRESSES
Residual stresses caused by welding may have a negative or
a positive effect on the strength of the component depending on
their type, sign, direction and distribution. Triaxial tensile
residual stresses in combination with cracklike defects promote
brittle fracture. Uniaxial or biaxial tensile residual stresses
edge,Z diminish corrosion resistance, however compressive residual
stresses improve the fatigue strength. Components with welding
residual stresses may distort during subsequent machining,
storage, and service loading. It is, therefore. necessary to eliminate
or atleast minimise welding residual stresses to take maximum
advantage from the welded fabrications.
Reducing welding residual stresses consists, of the following
possibilities :
(i) decreasing the level of residual stresses, in particular
the maximum tensile stress level,
(ii) decreasing the zones with high residual stresses,
(iii) decreasing the degree of multi-axiality of residual ten-
sile stresses.
To achieve these aims the following methods are employed.
1. Design considerations,
2. Material considerations,
~

262 Welding Engineering and Technology 263

Welding Stresses and Distortion

3. Preheating, twice applied heat cycle may result in the material exceeding
4. Welding procedure, ductility limit locally, for instance, at notches and defects. A
5. Welding sequence, counter measure is to arrange the transverse welds in a staggered
pattern between continuous longitudinal welds, for instance in
6. Postweld heat treatment; plate panels, Fig. 4.36 (a) or in I-section girder joints (staggering
7. Natural ageing, of flange and web welds) as shown in Fig. 4.36 (b). TILeamount of
8. Peening, staggering should be atleast twenty times the plate thickness.
Another method is to cut a 'mouse hole' in the vertical member so
9. Vibratory stress relieving. as to avoid overwelding at the intersection as shown in Fig. 4.36
Brief description of each one of these methods follows. (c). Cutouts of this type, however, cannot generally be reco-
4.5.1. Design Considerations mmended without reservations in respect of fatigue strength.
The principal design measures for limi'ting welding residual A fillet weld in a cruciform, T, corner, and lap joint is superior
stresses are: to butt weld in respect of residual stresses though not so in respect
(i) smallest sized weld of smallest length,
of fatigue strength. The force flow diversion in the case 6f fillet
weld reduces the stiffness of the joint with the result that the '
(ii) fillet weld favourable over butt weld, transverse residual stresses build up to a lesser extent. The fillet
(iii) prefer division of structure into sub-assemblies. weld is also superior to the butt weld in respect of the required
Larger the amount of weld metal deposited, more the shrink- dimensional tolerances of the component to be fabricated. Ahigher
age and higher the residual stresses. Weld accumulations and degree of transverse and angular misalignment is acceptable in
weld crossings are regarded as particularly detrimental. Where fillet welds than in the case of butt welds-the groove of which is
weld seams are accumulated in parallel, welding residual stresses required to be accurately aligned. The drawback on account of
acting in the same direction and plastically deformed areas stiffness of the butt weld can be counted in design term by an
superimpose to some extent. Where weld seams cross, high adequately large straining length i.e. the workpiece length
residual stresses are built up in two directions. In both cases the transverse to the weld.
Finally, the division of large structures (e.g. ships, bridges,
etc.), into sub-assemblies (or sections) pre-fabricated to specified
dimensional accuracy, is a particularly effective measure for
controlling residual stresses in the completed structure as a whole.
The sub-assemblies can be fabricated in the shop with higher
dimensional accuracy and only the final joining may be done at
site. This leads to reduced overall stresses.
4.5.2. Material Considerations
lalPlate panel (c) Plates stiffened with rib ~l The selected material should have good weldability under the
~ ange
prevalent fabrication conditions and the resulting welded joints
Web
should prove to be free of cracks, resistant to fracture and
adequate to perform satisfactorily under the service conditions.
3 ~p welding Flange Welding residual stresses play an important role in crack
formation and growth. Compressive residual stresses impede
lbl I-section girder 1,2,3 -Welding sequence crack formation, whereas tensile residual stresses, particularly
Fig. 4.36. Methods of avoiding weld crossings and adoption of favourable those of biaxial and triaxial nature, promote it.
welding sequence, (a) staggered welds in a plate panel. transverse welds The influence of the material on welding residual stresses can
to be laid first; (b) I-section girder wi~h staggered flange and we j
welds, transverse staggerelhweld to be laid first; (c) plate stiffened by be comparatively assessed essentially based on material charac-
rib, transverse weld laid first and weld crossing avoided by cut out. teristics like melting point (T m)' diffusivity. (a), coefficient of
~"'OP"' __ "' _
Z64 Welding Engineering and Technology
thermal expansion (a.), modulus of elasticity (E), and the yield
limit (cry) of the material being welded.
The welding residual stresses increase with the increase in
melting point of the material. A high melting point results in
higher stresses and the same holds good for specific heat (c) of the
m~terial. Thus, from the melting point of view, aluminium has
much better weldability than steel or titanium.
The residual stresses incr'ease with increase in thermal
diffusivity (a) of a material which characterises the rate of
temperature equalisation with transient heat flow conditions,
because the effect of thermal conductivity (k), specific heat (c) and
density (p) of the material being combined in a (= c~) Small K
with high cp would be particularJy suitable from the point of view
of 'a', that means steel or titanium have better weldability than
aluminium or copper. A high temperature concentration should be
aimed at, provided it does not have detrimental effect on the
microstructural transformation due to welding.
Welding residual stresses increase with the increase in
coefficient of thermal expansion (a.).Thus, from the point of view
of a., titanium is relatively better weldable than aluminium,
Residual stresses in welds also increase with the increase in
modulus of elasticity (E). From this aspect, aluminium will
develop low residual stresses while steel, titanium and copper
would develop higher residual stresses.
Higher the yield limit, cry, higher the possibility of sustaining
residual stresses both in respect of peak values as also in respect
of the average stress level. The deformation energy stored in the
welded structure may, thus, be increased, promoting the risk of
brittle fracture. The same holds good in respect of increasing the
yield limit at elevated temperatures; this increases the risk of
cracking at elevated temperatures, hence needs to be countered
by high temperature, ductility. Cast iron is less suitable for
welding because of the lack of high temperature ductility. ' vary greatly from the average values of welding suitability indices
It is evident from the above discussion of material charac-
teristics that the tendencies revealed are in some cases contra-
dictory. For example, aluminium on the basis of the melting point
(T m) has better suitability for welding while on the basis <!f.a, a.,
and E it is relatively unsuitable. So, by simple discussion of the
above disc~ssed parameters a clear picture may not emerge as to
the knowledge about weldability of a particular material.
However, an overall evaluation can be made by using the Welding
Welding Stresses llnd Distortion 265
Suitability Index (I.'a)' suggested by Radaj, based on welding
residual stresses relative to brittle fracture, and is given by the
following relationship.
r;.. , a* , a.* , E* . KIC
A.a = . m ' a . a. ' E ' A/C (4.32)
T r>"" •••
where, Tm, a.*,a*, E*, K;c refer to the values ofdifferent parameters
connected with a reference material and, T m' a., a, E, KIC are the
corresponding values of the parameters for the material which
needs t.o be assessed for weldability from the point of view of
residual stresses developed.
In this equation the determining factor on the materials side
for the level of residual stresses is introduced with T m' a, a. and E
which must be set in proportion to cry because higher residual
stresses should be sustainable at a higher yield limit. The higher
yield limit can, however, be exploited without brittle fracture if
the fracture toughness KIC (or for less brittle material the critical
crack opening displacement, 0c) rises in the same ratio. Hence
there follows the factor -?
cry
. *
cry/ cry
cr [KICI K;c]
= .A/C
rr- -
KIC
in equation (4.32).
Considering low alloy steel as the reference material, the
welding suitability indices for different materials are given in
table 4.4
It is clear from table 4.4 that aluminium and copper alloys
appear on an average to be less suitable for welding, whereas
titanium alloys come off well. It may, however, be noted that the
possible embrittlement of titanium as a result of gas absorption
is not included in this assessment. High alloy steel is on an
average as suitable for welding as low alloy steel; the possible
embrittlement of steel as a result of hydrogen, ageing or hardening
is not covered, nickel-alloys are on an average considerably less
suitable for welding than steel. The individual alloy may however
based on residual stress considerations (A.a) given in table 4.4.
The welding suitability index can also be used to quantify the
influence of preheating on residual stresses.
4.5.3. Preheating
\Velding residual stresses can be reduced by preheating. That
it is so is clearly revealed by the welding suitability index given
by equation (4.32). Because with preheating the term Tm in
266
I. '"
'"
Welding Engineering and Technology
, Welding Stresses and Distortion 267

equation (4.32) is to be rep~aced by (T m - Tp) thus A.a increases .

It.) 'tli
~-- ~
•....• 0
to
OC! ~
0
~
~
co
~ C'1
00
~ C")
~
Higher the value of Tp and lower theT m' more the value of A.a'
~ «w This also leads to lesser heat input by welding and reduced width
:0 of the fusion zone. Preheating from this point of view is important
S -----
.:; for materials with high thermal diffusivity, such as aluminium
'" r- <Xl CO') 0 0>
and copper alloys for which the preheat temperatures are respec-
l>()i ~ ~
'-s::- «., -' 0C'! ~
00 0~ 0~
"t:l
1"""4

tively ::;;200°C and::;;700°C. Preheating may also be necessary for

~ reducing residual stresses at elevated temperatures during
~
CIJ

",,-
welding of materials with low ductility at high temperature, for
eJ
'" '" '"
'- example, cast iron. This consideration, however, holds good when
:.a CIJ •.. ~ E 0 the whole component/structure is prehe'ated with unrestrained
= >. .E!.:: E 0 0 0 0 0 0 cooling down. Local preheating may involve additional residual
~>.-
0 ~~,
••• ::l ~
0
<Xl
0
00
0
<.0
C'I
C'I
<::>
00
0
00
;<,:< ~o-...,
U /I /I /I /I /I /I stresses. It is recommended to preheat the largest possible areas
=-e= :.::- i.e. wide areas on both sides of the weld atleast 70-80 mm on each
.t:J
--'---- side of the welded joint. Local preheating, therefore, cannot be
.s
••••

::s-;
00_
b«l~
(IS
CIJ
'-..•..•
;.::
E E
"t:l'
~3;
---
""",
Ii;

~
0
0
r-I
0 0
~
0
ot:>

'9I
0
""I
00
0
00
00
t-I
0
:15
0
0
i i
C'I

CO')
0
0
CO')

;::l
considered as beneficial in all situations.
Preheating of steels for welding is effective in two ways viz.,
=~ ::;:: b>-
(i) it reduces the cooling rate in the HAZ which improves
••--
....
-e_ d
the microstructure, and,
-=- .9
.~ ~ E ~'" (ii) it reduces diffusion of hydrogen which prevents cold
~t:-e
~ (IS';O
<;;:;~ 0 0 ot:> 0 0 ot:>
:.a cracking.
= 0
-e c:l.=
(ll!
c:l-g~
t;;lE~
•...• 0
"" C'I
<.0 ••••
••••••••
CO') •••••

""
'-o
(IS S •.. ~ ';j
'" Preheating slows down the cooling rate leading to reduced
hardness and thus more ductile microstructure. It is usually
---
I:#)~ Q,)
~ >.~ .s
.... --'
..... - ~- 'I' '\' 'I' 'I' 'I' 'I' ~
., intended to keep the hardness values at ::;;350 Hv. The preheating
~
••• (IS
.... '-'
~
E
-~ C,)
~, I
0
-'
0 0 0 0 0 temperature for welding steels is usually up to 300°C and it is
=.. •.. -.-4 ~
0- ,<:: ,.,- •••.• <.0 ..,.
..-..4

ot:>
1"""4

00
-'

CO')
«
o
c::

w selected depending on the carbon equivalent, and ratio of weld

••• CIJ
E--.;tj~ cd .-I'
~.g "'c:l •..• I 11'"""I

x
..-:

x
~

x x
....-l

x x -0 thickness to plate thickness for single pass welds. Preheating to

c::
- =
(IS~
<'l
at least 20°C is a common practice to reduce the risk of brittle
.~ ....0 c;€~ ~ ~ :5
'"
Q)
'"
'"
fracture if the working temperature is below 5°C.
CIJ

>.~ ;:: '" ""E

I::'~ I I •..•
I 0
"'::l
~~...5 '" ~
t- ~ t-
ot:>
to "" ot:>
~'" Preheating is not necessary for austenitic steels because
.::: S lC:) ....-l .,......

~::l
~ 0 martensitic hardening does not occur in these steels.
..,;, CIJ ~~ -0
••••
~~~ 0 -r;;
Q) Gas welding, compared to arc or beam welding, in a way
~EG t I 0 '-o
...
preheats the material due to a large spread of the flame. This
:E 0 0 0 0 tt'J
results in reduced residual stresses in such welds.
.,;:t ~ ~
~
'"
~ -< ~~
'"
~Q,~E I lD"" ..,.
II'"""I....-l0 0
<.0 000 00 0
1"""4....-l....-l ..,.
CO') <'l
..0
---- ~ I ..-- Q)
.,c
..c::
Localised preheating of the weld zone can be done with gas
torches, heating mats or induction coils. Asbestos mats are used
t; _ ~ o
-...:0 _ Qj ..£
0~ for reducing the heat losses and as a protection for the welders.
.§ ~ ~.,
'" Q) 2S -

~ tl ~ ~ ~ * The attained preheating temperature is checked by means of

~ ~.£ ,2 e ~ ~ temperature indicating crayons (i.e. thermal chalks) or adhesive
~ ~ ~ ,S ,2 :u -
~ i1~ ~ thermometers. Preheating of the whole component in a furnace is
.3 i :;;:~ §:~
(5 Z expensive and is therefore rarely resorted to.
268
Welding Engineering and Technology
Welding Stresses and Distortion 269
4.5.4. Welding procedure
Residual stress pattern in a slow weld (i.e. a weld made at each end and compression in the middle section. The intensity of
slow welding speed), as already discussed, produces a transverse load depends on the material characteristics, the width of the
load pattern similar to the one shown in Fig. 4.37 with tension at HAZ, and the length of the continuously directional weld tending
to be pr.oportional to the last factor. In a continuously directional
Weld weld the runs or parts of runs (if the weld is laid manually), are
deposited successively in the same direction, each part run picking
up from where the preceding run finished as shown in Fig 4.38(a).
Waiting for the material to cool between part runs does not help
the situation and infact may aggravate it by making the overall
deposition rate 'slower'. Continuously directional welds of this
kind can leave a high level of residual transverse load or may
cause immediate cracking in the tension zones. Therefore it is
necessary to use a deposition technique that will minimise the risk
from this cause. The accepted method is to use the backstep
Tension Tension welding after setting up t.he joint and tacking as necessary.
+
4.5.4.1. Backstep Welding
In backstep welding, first of all short tack welds of length
.-
~
V'l
V'l

0
Weld i 1 = 2.5t are executed, spaced e*~ 25t apart where t is the plate
Vl thickness. Slower the welding speed closer should be the spacing
between tacks to prevent groove edge displacement. The first weld
layer is deposited using backstep method from one tack to another
Ccimpressior. opposite to the overall welding direction, as shown in Fig. 4.38 (b).
It can be seen that each successive part run finishes where the
Fig. 4.37. Transverse residual stress distribution in the central plane of
butt weld made by continuous slow deposition. preceding one started so that the transverse load pattern is broken
Teck ,velds . up into a low-level more uniform type of residual stress pattern
as shown in Fig. 4.37 (c).
--x~-t __ ~x~e~l~~ Finally the remaining weld runs are applied continuously,
----- 1 2
lal
3 4 --- alternating the welding direction. Welding residual stresses
remain limited if the welding operations are performed in
succession without the part cooling down completely ,in-between
the runs. This is of significant benefit compared to the sectionwise
--x--- - -x--- --x- -- - -x- x~e~l!..n!
_4 - 3 •.....
Ibl
2 ---- - -
1
joining of the structure with the unavoidable· complete cooling,
down of the welded sections before the remaining groove openings
are welded up. Backstep welding is a widely used method for
fabricat.ion of large structures such as ships and storage tanks.
4.5.4.2. Skip Welding
: 0
~
\ ~~~~
A A A J Weld 'to.
r; skip welding is another method of reducing residual stresses
in welded joints. In this method a short length of the weld metal
_+1 T C T C T
Ie} C T T,",;"
Compression is deposited in one part of the seam, then the next length is done
Fig. 4.38. Welding sequences, (a) in continuously directionai welding, some distance away (Fig. 4.39) keeping the sections as far away
(b) backstep welding sequence, and (e) break up of transverse from each other as possible, thus localizing the heat. This method
residual stress with baekstep method. is· very successful in arc welding of cast iron.

I
 271
270 Welding Stresses and Distortion
Welding Engineering and Technology
In making large walls from smaller plates (Fig. 4.41) start
3 5 7 2 8 4 6
with the welds within each tier (2, 3, 1 and 4) and then make the
welds between the various tiers (6 and 5); whila for plate deckings
the correct welding sequence is shown in Fig. 4.42.
! 6
I I I 11
--1--1----..'~I
"
1

I
•

I
__II 1 I

I--l-+-
I I 3
19 14
I
I
I
1
I
I
I
I
1
I
I I
I

General direction of welding

Fig. 4.42. Correct welding sequence for plate decking.
Fig. 4.39. Skip welding; arrows and numbers show 4
direction and order of welding.
4.5.5. Welding Sequence
Correct welding sequence can reduce the development of
residual stresses and avoid fracture, it is therefore precisely
specified in the welding s~quence schedule for high quality welded
structures. The supreme principle is to assure 'unrestrained
~~~Et (a I 3

3 62 9514
137812
11
transverse shrinkage' or large 'straining leng~h' for butt welds. In 10
4 1
the case of plate panels with transverse welds arranged in a
staggered pattern to avoid weld intersections, Fig. 4.36 (a), the
discontinuous transverse welds are welded first, followed by the {b I
continuous longitudinal welds. A similar procedure is adopted for Fig. 4.43. Two suggested sequences fQrwelding H-beams.
cylindrical vessels where the longitudinal welds are executed first
followed by the circumferential welds as shown in Fig. 4.40. In In the fabrication of built-up I or H girders (Fig. 4.43), start
with the butt welds on the web and flanges and finish with the

w)7')
Fig. 4.40. Sequence of welds in the fabrication of cylindrical vessels.
plate assemblies with stiffeners, the butt welds between the plates
and between the stiffeners should be welded first, and only then
the fillet welds between plates and stiffeners (Fig. 4.36(b)).
4
Fig. 4.41. Sequence of welds in the fabrication of walls from separate plates.
fillet welds between the web and flanges.
4.5.6. Postweld Heat Treatment (PWHT)
The best method to relieve stresses is to anneal the weldments.
Unfortunately, this can only be accomplished in furnaces which
normally can hardly accommodate welded componentJstructure of
any consequence. Stress relieving treatment, therefore, is obli-
gatory only for essential articles, such as boiler drums and high
pressure vessels. In some cases this heat treatment may result in
warping.
Stress relief treatment that is often given to welded
components is dependent upon its material composition. Table 4.5
lists the temperature ranges useu for stress relieving postweld
heat treatment of various types of materials.
The energy for PWHT mliY be provided by oil, propane or
natural gas or electricity. The heat may be applied locally, or the
welded components may be completely enclosed in a furnace.
272 Welding Stresses and Distortion 273
Welding Engineering and Technology

Table 4.5. Soaking Temperatures for Postweld Stress 4.5.7. Stress Relief By Natural Ageing
Relieving Heat Treatment for Different Materials A welded component can also be stress relieved by allowing it
S. Material to age outdoors. Here the necessary energy, to realign the crystal
Soaking Temperature
No. (OC)
lattice structure, is supplied by small natural ground movements
1. Carbon steel 595-680
and the normal temperature fluctuations. In this ageing process,
2. Carbon - 1;2% Mo steel 595-720
the stress level is reduced about 10% per month during the
summer and 5% per month during the winter; thus a year or more
3. Ih% Cr - Ih% Mo steel 595-720
is required for 80 to 90% stress relief.
4. 1% Cr - Ih% Mo steel 620-730
4.5.8. Peening
5. 11;4% Cr - 11.!%Mo steel 705-760
6. 2% Cr - Ih% Mo steel 705-760
Peening consists of lightly hammering the weld and/or the
7. 21;4% Cr - 1% Mo steel
surrounding parent metal in order to relieve stresses present and
705-770
to consolidate the structure of the metal. It should preferably be
8. 5% Cr - 11.!%Mo (Type 502) steel 705-770 done with a pneumatic tool of moderate force. As an alternative,
9. 7% Cr - lI.!9'o Mo steel 705-760
10. 9% Cr - 1% Mosteel 705-760 it may be carried out with a peening hammer of about ~ kg weight,
11. 12% Cr (Type 410) steel 760-815 while the weld is still hot or immediately after the weld has cooled
12. 16% Cr (Type 430) steel 760-815 down to room temperature. In fact peening should be done when
13. 11/4% Mn - 11.!%Mo steel 605-680 the weld metal is at a temperature either higher than 450°C or
595-680
lower than 150°C, because in the intervening temperature range
14. Low-alloy Cr-Ni-Mo ste~ls it tends to crack due to reduced ductility.
15. 2 to 5% Ni steels 595-650
16. 9% Ni steels 550-585
A great deal of controversy exists as to whether peening is
advantageous or not. Some engineers advocate it because it
17. Quenched and tempered steels 540-550
reduces the residual stresses, others oppose it because other
Source: AWS Handbook. stresses are set up and the ductility of the weld metal suffers. If
PWHT temperatures, usually in the range of 540 to 815°C done reasonably, however, it undoubtedly is of value in certain
reduce intemal stresses and help to soften hardened areas in the instances. For example, in the arc welding of cast iron the risk of
HAZ. The work is heated to a given temperature held at the
fracture is definitely reduced if the short beads of weld metal are
selected temperature for an appropriate soaking time and then lightly peened immediately after they have been laid. Care must
allowed to cool, both heating and cooling being subjected to a be taken in peening hot metal in that the slag particles are not
controlled temperature gradient such as 100-200°C per hour for driven under the surface. Also, the first and last layers of weld
thickness up to 25 mm and slower rates for thicker plate, should not be peened. Peening the first layer could actu ally pierce
depending upon the code being followed. the weld or displace the member sideways. Peening the last layer
is prohibited largely in the belief that cold working would injure
Localized heat can be applied using flexible insulated pads and the weld metal, and there is no subsequent application of heat to
finger electrical heaters. These are available in a variety of shapes anneal it. The points to be taken care of in peening are :
with the elements insulated with ceramic beads and supplied from (i) work hardening effect should be considered before peen-
the welding power source or an auxiliary transformer at 60-80 ing AISI 300 series steels.
volts. The pads are connected in parallel as required' and are
covered over with insulating material to conserve heat. (ii) AISI 400 series steels have relatively poor notch duc-
tility in the as-welded condition; therefore utmost care
Heat can also be applied locally by gas or oil burners, as for should be exercised if peening is attempted. Peening
example on circumferential and longitudinal weld seams of should be employed on each weld bead or layer (except
vessels. Modern burners operate in all positions and can be the first and last) to be effective. The effectiveness
obtained with magnetic feet for easy attachment to the work.
·274 Welding Engineering and Technology
decreases as the thickness of the bead or layer increases,
e.g., peening becomes of doubtful value for deposits of 6
mm or thicker except where heavy blows are permitted.
Metal adjacent to the welds is usu.ally peened, as well as the
weld itself.
4.5.9. Vibratory Stress Relief
Vibratory stress relieving (VSR) is a resonance based method
of relieving residual stress and stabilising metals. This is a
non-thermal method, therefore there is no distortion or chang'e of
tensile strength. Energy is introduced to the component, to be
stress relieved, by means of low frequency (10 to 220 cycles/
second), high amplitude vibrations and can be successfully applied
to parts weighing from ~ kg to equipment weighing u.p to 150 tons.
For welded fabrications, stress relieving can be performed during
welding which is very helpful in preventing residual stress
build-up that can cause weld cracking or distortion of some
sections.
VSR method is claimed to ensure extreme accuracy and
stability without affecting the rigidity of or material properties of
a component. In general, where components have been excited at
op,e or more resonant frequencies, the results have been stress
reduction of 30 to 80%. Where resonance has not been attaine'd
results have been poor., Strain measurements show that non-
resonant vibration works only when high tensiles stresses
predominate whereas resonant treatment works equally well on
high tensile .or compressive stresses or on both. When resonance
is used, stability more than matches that of thermal stress relief
(TSR). The success ofVSR is attributed to cyclic versions of stress
over-load or beneficial effects of vibration on the distorted crystal
lattice of the welded material.
4.5.9.1. VSR Equipment
The VSR machines drive high performance rotating mass
vibrators. The equipment is portable and can be divided into a.c,
and d.c. motor types. At low frequencies either can be used but
with d.c. types frequenCies are limited to 100 Hz that is why most
of the equipment used is of a.c. type with a frequency range of
10-220 Hz producing a top force of about 17 KN. Drive units weigh
as little as 20 kg and vibrator 24 kg. For quality control purposes,
chart recorders are available which can provide a permanent
record of the treatment.
Welding Stresses and Distortion 275
4.5.9.2. VS~ Procedure
The VSR treatment can be given during welding where it can
lower the residual stresses, reduce distortion, refine weld
material, improve dilution, reduce cracking, and increase
deposition rate. It can as well be applied just prior to finish
machini:p.g, if employed. The component to be stress relieved is
supported on rubber isolators and the vibrator is attached at the
edge of the component. A sensor mounted on the component will
identify the resonant conditions as the frequency range is scanned.
A resonant peak occurs when the induced frequency of the vibrator
coincides with the natural frequency of the structure and this can
be seen, felt and displayed on meters and recorders. The optimum
position for the supports is beneath the node lines i.e. still areas
shown during the scan.
Peaks are approached slowly, with a pause at the foot to allow
any critically high stresses to diminish, prior to treating at the
actual peak for the number of minutes, specified for the component
type, size and material. As many of the natural frequencies as
possible are sought and sometimes the vibrator is repositioned to
alter the resonant mode shape. The greater the frequency range,
the better is the treatment as more loading patterns (modes) are
established and different areas are affected. More loading patterns
means that fewer cycles per mode are needed (often as low as 1000
cycles). The higher the frequency, the more complex is the loading
pattern and the more uniform the treatment as panels and limbs ii
are resonated. Service loading patterns may also be simulated.
I"~
The rubber isolators, placed at or near the mode lines, allow I,.
freedom of vibration which is vital for optimum treatment. Rigid Ii
clamps are used to make the vibrator and component to behave :1
as one unit. I,i
Straightening of the welded component, if required, should be
carried out before giving the VSR treatment, but it should be borne
in mind that the process does not normalise or anneal the III
components. Components are not damaged by resonance because
the small movement at the foot of the resonant peak starts to
reduce and redistribute critically high stresses.
Charts of conformity are produced which simplify control.
These charts are in the form of certificates detailing the treatment
or chart recordings detailing resonant response-and duration of
treatment.
It is, of course, a good welding practice to nor'1alise items such
as thick flame-cut edges which are to be machmec! or to ensure
276
Welding Engineering and Technology
good weldability. Thermal stress relieving (TSR) of a detail and
VSR of the finished component is still easier, quicker and cheaper
than TSR of the whole finished component.
Failure during VSR, if it occurs, is not due to fatigue but
mainly due to pre-existing microstructural damage caused by
critical stressing during the cooling of highly constrained fabri-
cation or due to some other welding defects e.g., poor.penetration,
lack of fusion, slag entrapment, or the pre-existing cracks, etc.
In welding situations where vibration is used during the
welding process, it is most compatible with SMAW,GMAW,and
GTAWprocesses. Other welding processes may present formidable
logistic problems.
4.5.9.3. Uses of VSR
The VSR method can be used on a wide range of ferrous and
non-ferrous metals including carbon steel, stainless steels,
aluminium, cast iron, manganese, stellite, inconel, etc. The items
often stress relieved include mild steel fabricated press bolsters,
stainless steel gear boxes, mining and quarrying equipment,
carpet knife blades up to 5 m length, large bed plates, mild steel
components clad in stainless steel used in the paper and chemical
machinery industries which cannot be thermally stress relieved.
However, because VSR does not bring about any metallurgical
change and therefore cannot be used on pressure vessels, pipe
work or in anyother case where such changes are required. Also,
it has limitations in that there is no simple way to tell whether
stresses have been relieved or not, and for that one has to rely
upon the standard methods e.g., X-ray diffraction, hol~ drilling
with strain gauges·or for ferromagnetic mat~rials the Barkhausen
Noise Analysis method.
The time required for the VSlf treatment is a function of the
weight of the component, for example, a component weighing
10,000 Kg requires about 20 minutes while the one 'weighing 50
tons requires about 30 minutes.
4.6. DISTORTION IN WELDS
Any unwanted physical change or departure from specifi-
cations in a fabricated structure or component, as a consequence
of welding, is called welding distortion.
Such departures from specified measurements brought about
by welding not only complicate the fabrication ofwelded structures
but may often reduce the usefulness of the finished structure for
the purpose for which it was designed. Also, the cost of rectifying
Welding Stresses and Distortion 277
welding deformations by straightening distorted components and
trimming joints may, in some instances, be more than that of the
cost of actual assembly welding. Measures are, therefore, taken to
minimise welding distortion or to rectify it after welding.
Distortion is caused by the non-uniform expansion and cont-
raction of the weld metal and adjacent base metal during the
heating and cooling cycle of the welding process. During such a
cycle, many factors affect shrinkage of the metal and make
accurate prediction of distortion difficult.
In arc welding the plate areas are cold at the start of the
welding operation but as welding proceeds heat is dissipated
through the adjacent areas which tend to expand. Behind the
moving heat source, molten weld metal solidifies and cools, thus
tending to contract. Hence, expansion and contraction occur in the
weldment simultaneously, producing a complex stress system
often leading to distortion which is aggravated further by
subsequent weld runs.
Typical values of contraction for mild steel expressed as a
percentage of the volume at 20°C are:
(i) Extraction of superheat 0.8%/100°C,
(ii) Solidification shrinkage (phase change) 3.0%,
(iii) Solid-stage shrinkage 7.35%.
Thus, the total contraction in the weld metal and the adjoining
HAZ will be in the range of 11-15%.
Thermal expansion and contraction will either result in
distortion or, if sufficient restraint is applied, in the development
of internal stresses. These stresses may result in plastic defor-
mation ifthe yield stress is exceeded, or even in the tearing of the
material. In the former case, int.ernal stresses up to yield stress
will remain and any subsequent thermal or mechanical treatment
may lead to distortion. Thus, both the material properties and the
welding processes and procedures affect the extent of distortion.
4.6.1. Significance of Material Properties
Effectsofsomeofthe important material properties like coefficient
ofthermal expansion (ex), thermal conductivity(K), yield strength (Oy),
and modulus of elasticity (E) are discussed as follows.
4.6.1.1. Coefficient of Thermal Expansion
It is a measure of expansion and contraction. The higher the
coefficient of thermal expansion of a material, the greater the
distortion is likely to be.
278 Welding Engineering and Technology
4.6.1.2. Thermal Conductivity
It indicates how readily the heat will spread through the
material. Low thermal conductivity leads to high thermal
gradients, and high thermal gradients lead to high distortion
because distortion depends upon internal or external restraint.
4.6.1.3. Yield Strength
The higher the yield strength of the parent material the
greater the residual stresses available for causing distortion.
4.6.1.4. Modulus of Elasticity
The modulus of elasticity is a measure of stiffness of a material
so that higher the modulus the more it will resist distortion.
4.6.2. Influences of Welding Processes and Procedures
Other than the physical properties of the material, the factors
determining the deformation of a structure during welding are
partly of a teclmological and partly of a constructional nature. For
a given design, the deformations will be determined by the order
of assembly of the components, the welding current and voltage,
the welding sequence, clamping methods and some others. Yet for
a given assembly and a welding procedure employed for a given
type of structure, small dimensional variations even of less
important components, or small variations of shape of minor
members, the dimensions of which are not calculated, may lead to
important differences in the resulting deformations. Obviously, all
the constructional and technological factors influencing welding
deformations can only be properly taken into account by
calculating the effect of each separately.
Overall the amount of welding distortion will depend upon the
volume of the heated metal (i.e. weld metal and the parent metal)
involved. The distribution of this heated metal (i.e. the bead shape)
is also important in causing the type of distortion that results. If
heat is evenly distributed throughout the plate, uniform shrinkage
will occur which is' less of a problem than the angular distortion
caused by non-uniform heating and melting through the plate
thickness. The volume of the heated metal depends upon the
following factors.
(i) Welding processes,
(ii) Amount of weld metal,
(iii) Welding speed,
(iv) Edge preparation and fit-up, and
(v) Welding procedure.
279
Welding Stresses and Distortion
4.6.2.1. Welding Processes
The effects of welding processes and number of weld runs can
be summed as follows.
(i) A concentrated heat source gives less distortion than a
spread heat source,
(ii) High welding speeds (less spread of heat) gives less
distortion,
(iii) Deep penetration welds give less distortion,
(iv) Single-pass welding gives less distortion than multi-pass
welds and the fewest number weld runs should be used
consistent with attainment of adequate properties.
4.6.2.2. Amount of Weld Metal
The amount of weld metal should be kept to a minimum. Weld
contours should be flat or only slightly convex. Intermittent
welding can reduce the amount of weld metal by 75% for
attachment of stiffeners without loss of strength, although the
fatigue behaviour may also need consideration.
4.6.2.3. Welding Speed
Maximum welding speed should be used to minimise heat
spread and build-up, however the solidification of weld metal in
desired orientation must be kept in mind to avoid centreline defect
or weakness.
4.6.2.4. Edge Preparation and Fit-up
Edge preparation should be such as to require the minimum
amount of weld metal and be as uniform as possible to allow for
consistent shrinkage along the joint. Close fit-up also reduces the
amount of weld metal and hence weld metal shrinkage.
4.6.2.5. Welding Procedure
Mechanised, single pass, high speed welds should be made
whenever possible.
Where manual welds are used back-step or skip welding
technique, as shown in Figs. 4.38 and 4.39 respectively, should be
employed. Welding should always be carried out from a clamped
to a free end.
4.6.3. Types of Welding Distortion
Distortion due to welding can be of the following types.
(i) Longitudinal shrinkage,
(ii) Transverse shrinkage,
(iii) Angular distortion,
%80 Welding Engineering and Technology Welding Stresses and Distortion 281

Longitudinal distortion, Bowing or Bending;

(iv) Thus, it is advantageous to channel distortion towards the
Rotational distortion,
(v) shrinkage mode.
(vi) Buckling and Twisting. 4.6.3.1. Longitudinal Shrinkage
All these types of welding distortion are shown in Fig. 4.44. (A) Butt Welds
In general it is easier to make allowance for shrinkage than The amount of longitudinal shrinkage in a butt weld is much
angular distortion and bowing which are much more troublesome.
less than the transverse shrinkage and is of the order of 1~ooth
f-- I ~ Yj-
ta I
I
1
fie
I I
b-
d-
-
Transverse shrinkage
Angular distortion
Bowing
of the weld length. It is given by the following relationship.
3·I·L
6L = 100,000. t
...(4.33)
: : I e- Rotational distortion where 6L = longitudinal shrinkage, mm
b,b
II IiO fa -- Buckling
Longitudinal shrinkage
I = welding curren t, amps
I L = length of the weld, mm
I
I t = plate thickness, mm
I
II
I
Example 4.2 If 6 mm thick plates are Joined in butt
.-La I configuration by shielded metal arc welding using 200 A current,
J
find the longitudinal shrinkage.
Weld Weld Solution. Using equation (4.33), we get,
3·1 ·L
--l~~ = 100,000 . t
-l-~:r Now,
!J.L
1=200 A,
( i} Butt weld in thick plate (ii I Fillet weld
t = 6 mm
3 x 200 xL L
e
6L = 100,000 x 6 = 1000 Ans. -b.
(B) Fillet Weld 1000
Longitudinal shrinkage in carbon steel fillet welds is primarily
a function of cross-section of the joint involved. Restraint is more
effective for thicker and wider plates. The total cross-section of
the welded plates in the transverse section is therefore called the

Slow weld Fast weld

I iii I Welds in thin plates
&z:a Ap

e"':'Y'
_ Aw

Fig. 4.45. Fillet weld section showing the .resisting

(iv) Butt weld in thin sheets cross-section (Ap) and the weld metal cross-section (Aw)
Fig. 4.44. Types of distortion in welds.
 282 Welding
-- Engineering and Technology .- Welding Stresses and Distortion 283

straining cross-section or resisting cross-section and marked, Ap 4.6.3.2. Transverse Shrinkage

in Fig. 4.45. When the ratio Ap to Aw (i.e., the cross-sectional area (A)Butt Welds
of the weld metal) is less than 20, longitudinal shrinkage, SL is
given by the following relationship, The distribution of the longitudinal residual stress, ax, shown
2f.1Aw III Fig. 4.47 is the cause of "transverse shrinkage. Though
SL=T-mm ...(4.34)
p
-1~
Example 4.3. Find the longitudinal
shown in Fig. 4.46.
shrinkage in a fillet weld
r - - ------IT
-i .1-- 6

l~----
75

,
1_
6C
All dim~n5ions in mm

Fig. 4.46. Fillet weld dimensions. shrinkage shown

L - _ ----.JTT
Fig. 4.47. Transverse shrinkage in a butt weld.
-E5t

in Fig. 4.47 is uniform it is not so in butt welded

long plates. The major factors causirig non-uniform transverse
Solution: Using equation' (4.34), we get longitudinal shrinkage in butt welds are:
. 25Aw
shrInkage, 0L = --. (a) Rotational distortion,
Ap
(b) Constraint.
Now, Aw = 2 x Area of each fillet weld
(a) Rotational Distortion : When welding is conducted
progressively from one end of the joint to the other, the unwelded
=2X(~X8x8J=64mm2 ... (i) portion of the joint, ahead of the arc moves, causing rotational
and Ap = (6 x\00 + 6 x 75) distortion and hence transverse shrinkage.
::::600 + 450:::: 1050 mm2 ..• (ii) (b) Constraint: The amount of transverse shrinkage that
From (i) and (ii), we get, occurs in welds is also affected by the degree of constraint applied
~_ 1050 to the weld joint. The amount of shrinkage decreases as the degree
Aw - 64 of constraint increases. In many joints, the degree of constraint is
::::16.4 < 20 y
Thus, equation (4.34) is valid. Transverse
25 xAw shrinkage
Hence, SL:::: Ap :)~
Detail of
25 x 64
::::-1050
-xjoint
=: 1.52 mm Ans. 1.52 mm.
Fig. 4.48. Transverse shrinkage distribution in a slit weld.
Ii
~
'"
284 285
Welding Engineering and Technology Welding Stresses and Distortion

not uniform along the weld, for example, a slit weld; where the
degree of constraint is maximum at the ends and the shrinkage Transverse shrinkage, at = ~~ ... (4.35)
varies accordingly as shown in Fig. 4.48. Shrinkage is also not
where, !i.w = cross-sectional area of weld, mm2
uniform in back-step welding. t = weld thickness, mm.
For a given weld thickness, transverse shrinkage of a weld
increases directly with cross-sectional area of the weld. Fig. Important: When deep penetration welding process e.g. SAW
4.49(a) shows this relationship for a 12.7 mm thick plate. Though is used, the cross-section of the entire fused part of the joint, and
not only the area of the deposited weld metal, must be used in the
5.00 calculations.
E
E Example 4.4. Determine the transverse shrinkage of the weld
a:;
C1>
3.75 shown in Fig. 4.50.
~Cl
c
.<:
~ 2.50
All dimcnsionr. in mm
QJ
'-
v>

>
QJ

Vl
C
Cl
t-'-
o
65 130 195 260
f 19
Area of w~ld (mm2,
(al
.-L
E
E
QI
C7'
Cl
-""

.<:
c
.&;
Ii>
QI 60· double V
'-
Ii>
Fig. 4.50. Fillet weld dimensions for Example 4.4.
>
QI

Ii>
C Solution. Area of weld, Aw = area a + area b + area c
Cl
t-'- o
6 12 18 24 30 36 = (6 x 19) + (~x 19 x 19 j+ ~ (3 x (6 + 19+ 3))
Plate thickness lmm' . 2
I bl
= 114 + 180.5 + 3' (3 x 28)
Fig. 4.49. Transverse shrinkage in butt welds, (a) effect of area of weld = 114 + 180.5 + 56 = 350.5 mm2
cross-section, and (b) effect of plate thickness. and t = 19 + 3 = 22 mm
the graph provides data even for large included angles but angles Using equation (4.35), we get,
above 60° are seldom used in welding. Transverse shrinkage of 60° Aw
single V and double V joints in several plate thicknesses are shown Transverse shrinkage, at = 'lOt
in Fig. 4.49 (b). Shrinkage values shown in both graphs are for 35Q.5 Ans. 1.6 mm.
welds between comparatively free plates. at = '-- 1.6 mm.
Approximate transverse shrinkage for weld angles and sizes Overall, the effects of various procedures on transverse
other than those shown in Fig. 4.49 can be predicted from the shrinkage of butt welds are summed up as given in Table 4.6.
following relationship,
2.86 Welding Stresses and Distortion 287
Welding Engineering and Technology

Table 4.6. Effects of Various Procedures on Transverse . _ leg of fillet (l)

Shrinkage of Butt Welds Transverse shrmkage, 0t - thO1mIC ess 0fItp a e ()t x 1.5 ... (4.37)
S. Procedure Effects on Transverse Shrinkage
Note: In lap welds 1 is usually equal to t.
No. Transverse shrinkage in welds is particularly important when
~_. the shrinkage of individual welds is cumulative as, for example,
1. I Root opening Shrinkage increases with the increase in root
opening in the beam-to-column connections across the length or width of
2. I Joint design Single V produces more shrinkage than dOuble V a large building. Unless allowances are made for transverse weld
joint shrinkage - usually for spreading the joint open by the amount
3. I Electrode diameter Shrinkage decreases as electrode diameter it will contract after welding-the cumulative shrinkage of several
increases beam-to-column connections could be large enough to noticeably
4. I Degree of constraint Shrinkage decreases as degree of constraint shorten the building's dimensions.
increases
4.6.3.3. Angular Distortion
5. I Peening Shrinkage decreases by peening but the effect is
minor Angular distortion occurs at butt, lap, T, cruciform and corner
joints as a result of single-sided as well as asymmetrical
6. I G~ging an~..!epairs_ Shrinkage increases by these operat_io._n_s _ double-sided welding. The extent of angular shrinkage depends on
(B) Fillet Welds the width and depth of the fusion zone relative to plate thickness,
on the type of joint, weld pass sequence, thermo-mechanical
(i) For T-joints with two continuous fillet welds Referring to material properties and the. characteristic parameters of the
Fig. 4.51(a), we have, welding process e.g. heat input per unit length of weld and
distribution of heat source density.
-jq- The dependence on welding speed v, heat input per unit length
of weld qw, and weld penetration h which results for the single weld
pass in mild or low-alloystructural steel, is shown in Fig. 4.52.
VI
Slru:tural slee I
si ng Ie -layer weld
15 0.03
:.c
c:J
•..
c::..
L ~ ...0.Q2

Fig. 4.51 (a). T-joint. with two fillet welds.

f t
~
g
'C
.&::
VI
1.2

•..
. _1~£!2f fillet weld (ll .9
~
...(4.36)
P 1a t e tl·lICk ness (t)
CJ'O

Transverse ShrInkage, Ot - c:
<t
(ii) For Fillet Weld(s) in a Lap Joint: Referring to Fig. 4.51(b)
showing a lap joint with two fillet welds, we have, 4 8 J/ m m 312 16 20 24

-h
-, I~·
~r ----T
__
1+ ~~npu~.ng~/h~
Fig. 4.52. Angular distortion, in single pass welds, a function of
Fig. 4.51 (b). Lap joint with two fillet welds. heat input per unit length of weld q/U and weld or plate
thickness h for d~fferent welding speeds u.
288 Welding Engineering and Technology Welding Stresses and Distortion 289

(A) Butt Welds. Angular distortion often occurs in a butt weld Weld leg length, l = 8 mm
w hen transverse shrinkage is not uniform in the thickness and plate thickness, t = 6 mm.
direction. It can be reduced by using double V joint preparation
and can be minimized, to almost zero, for a butt joint having the " Putting these values in above equation, we get,
following dimensions (Fig. 4.53), Oa = 0.0076 x 1090 x (8)1.3 = 3.15 mm.
(6)
g = 3 mm Exercise. Find the angular distortion in a double fillet weld
f~ ~@:: 1 ~
t]:: 2 mm between 1250 mm wide and 10 mm thick horizontal plate and a

L Jt,
Fig. 4.53. Double-V edge preparation for a butt weld
10 mm thick vertical member if the weld leg length is 10 mm.
(Ans : 1.895 mm).
Multiple Restrained Fillet Welds. In multiple restrained
with almost zero angular distortion. fillet weld (Fig 4.55) the distortion is given by the following
relationship,

and
((1 + ~ t3 J = 0.6
t
where root face, t3 = 2 mm,
root gap, g = 3 mm.
(B) Fillet Welds. Angular distortion in a fillet weld (Fig. 4.54),
is given by the following relationship,
... (4.38)

~LJLJ (a) Free joint

.fro
L
~" --I::dt t t-r=w ::::I--L
.'-i
~ U lAngular U
-<,~lA.ngUla.r
distortion,lSd I::' dlstortlon,lSo
Fig. 4.54. Angular distortion in double fillet T-joint .
. ' 0.0076 . W· l1.3
...(4.39)
Angular dIstortIOn, 8a = t ..,
where, W = flange width, mm lbl Restrained joint
l ~ weld leg length, mm
t = flange thickness, mm. Fig. 4.55. Angular distortion in:(a) free, and (b) restrained fillet welds.
Example 4.5. Find the angular distortion in a double fillet
weld of a T-joint between a flange 1000 mm wide and a vertical L="4<1>-L-2 (4.40)
member when the thickness of both the members is 6 mm and the 8 1 [x 1]2 <1> •••

weld leg length = 8 mm. where 8 = angular distortion, mm

L = span length, mm
Solution. Using equation (4.39), we have
= angular change, radians
<1>

Angular distortion in double fillet weld T-joint, x = distance from weld to the point where distortion is to be
Oa=---"
0.0076 . W·
t
l1.3 determined, mm.
Example 4.6. In multiple restrained fillet welds (Fig. 4.55(b))
Where, the flange width, W == 1000 mm the span length is 1 m and the angular change is go at a distance
of 400 mm from the span end, find the distortion.
 Welding Engineering and Technology Welding Stresses and Distortion 291
290
120
Solution: Equation (4.40) is,
110
L 4
~=.!<\>_[~_.!J2 L 2 <\> c 100

Now, L = 1000 mm, <\>

1t
= 9° = 9 x 180 = 0.1571 rad.
.2
~'- 90
/
'" '\--~"' --
L 1000 /
//I--o
70
•.. 80 // r-- :/~I '\I'\
'I~r--
1'--
'"
L°s,OW.U
,~ ~ "'-~ '" !'......
.....

// /
.....••.•

x = - - 400 = -- .- 400 = 100 mm / "- r-... 0,",

// ~
..•.•..•
r-......
r--:-.
r---.....
'" 1""-
2 2 .S 70
..no.
V
Putting these values in the above equation, we get ~ 60
QI

L8 ="41 x 0.1571-
. [ 1000
100 - 2"
1J2 x 0.1571
c
go'
.c
SO

'-' 40
= 0.0393 - [0.1- 0.5]2 x 0.1571 •..
= 0.0393 - 0.16 x 0.1571 = 0.0393 - 0.025136 .2 30
:;:)
C1'
= 0.014164
~ 20
8 = 0.01416 x L = 0.01416 x 1000 = 14.164 mm
Ans. 14.164 mm upwards 10
o
The amount of angular change in a restrained structure
<\>, o 2 4 6 8 10 12 14 16 18 20 22 24 26 28
(Fig. 4.55) is smaller than the angular distortion, <\>0'in a free joint. Thickness of plate, mm
The amount also changes when rigidity ofthe bottom plate given
<\>

Fig. 4.56. Effect of plate thickness and the amount of weld metal deposited per
by the relation, unit length on angular distortion in weldments made by free fillet welds.
D = Et3 [12 (1- v2)] ..,(4.41)
with other (different diameter) electrodes may not be much
and the length of the span, L, change; where E is modulus of different from the ones given in Fig 4.56.
elasticity (in Kg/mm2) and v is the poisson's ratio of the material
being welded. The relation between free and restrained angular Using 5 mm diameter electrodes, maximum angular distortion
distortion is given by the following empirical relationship. was reportedly observed for a plate thickness of 9 mm. For plates
thinner than 9 mm angular distortion reduced due to more even
<\>0 ...(4.42)
<\>=.--
2D 1 heating of the plate in thickness direction which reduced the
1 +--.- L c
bending moment. For plates thicker than 9 mm, the amount of
angular distortion reduced because of increased rigidity.
where c is a coefficient determined by welding conditions and plate Problem 4.7. A restrained steel structure of the type shown in
thickness, and is given by, Fig. 4.55 is fabricated by shielded metal arc welding; calculate the
~4 maximum distortion of the bottom plate if the plate thickness is 10
~
c----
- W
...(4.43)
mm, span length 1 m, and the leg length for triangular fillet welds
1+-5 is 8 mm. Also, calculate the distortion at 300 mm from the centre
of the span length. Take density of weld metal = 7.85 g / cm3,
where, t = plate thickness, mm deposition efficiency of electrode = 65%, E = 206 GPa and Poisson's
W = weight of the weld metal deposited per unit weld ratio (v) = 0.8.
length, gm/cm.
Solution.
_ weld cross-section (cm2) x density of weld metal
- Deposition efficiency (a) Total area of double fillet weld,

Free angular distortion (<\>0) for SMAW ul?ing 5 mm dia- Aw ::: No. of welds x area of cross-section of each weld
electrode can be obtained from Fig. 4.56 The results for welding
292 Welding Engineering and Technology Welding Stresses and Distortion 293

." (i) . Maximum distortion 8m will be at mid-span and obtained from

= 2X (~ X 8 X 8 J = 64 mm 2
Amount of weld metal deposited per unit length, equation (4.40) by putting x = ;
Aw x density 1
W = deposition efficiency
Thus, 8m = 4 . <I>
.L
1
= 64 x 7.85 = 7 73 / ... (i~). = 4 x 0.0428 x 1000
0.65 . g cm
" lOglOW = lOglO7.73 = 0.888 ... (iii) = 10.71mm Ans. 8m = 10.71mm
(b) Deflection at a point 300 mm from centre
From Fig.4.56, for log W = 0.888 and plate thickness of 10 mm, From equation (4.40), we have,
angular distortion for free structure,
... (iv)
= 80 X 10-3 radians
<1>0 L 4 L 2 <I>
~=.!.<I>_[~_.!]2
From equations (4.42) and (4.43), we get L 1000
-=1+-·-
<1>0
<I>
2D 1
L c
Now, x = 2 - 300=-2- - 300 = 200 mm

.. L8 = 41 <I>
- [ 1000
200 - "21]2 <I>
-1+-----·-- Et3 (1
\... + WI
5) 1 2
- 12 (1 - y2) L t4 =4" <I>
- (0.2 - 0.5) <I>

or = 1+
<1>0 E . (1+ WJ
5 ...(4.44) = i1 <I>
- (0.3)2 <I>
<I> 6 (1 - y2) L t
=4 <I>
- 0.09 <I>
Now, E = 206 GPa = 206 x 109 Pa = 206 x 103 MPa
3 2 206 X 103 2
= 206 x 10 N/mm = 9.81 kg/mm = 21000 kg/mm2 (i -
= <I> 0.09)= <I> (0.25 - 0.09)
=0.16<1>
Also, y = 0.3, t = 10 mm, L = 1000 mm and W = 7.73 g/cm2 Now <I>= 0.0428 (from vii)
Putting these values in equation (4.44), we get 8
.. L = 0.16 x 0.0428
.0 = 1 + 21000 ,[1 + _6.~8J i.e. 8 = 0.16 x 0.0428 xL = 0.16 x 0.0428 x 1000
<I> 6 (1 - 0.32) 1000 10 = 6.848mm Ans. 8200 = 6.848mm.
21 2.256
= 1 + 5.46 . 10
= 1 + 0.868 4.6.3.4. Calculation of Angular Distortion from Basic
Principles of Temperature Distribution
<1>0
= 1 868 ... (v) The estimation of angular distortion detailed in the previous
or
<I> . section is based on the empirical relationships evolved on the basis
<1>0 ... (vi) of statistical data collected under different fabrication conditions
.. <I>
= 1.868 and is, therefore, expected to hold good as long as the parameters
from (iv) in (vi), we get,
Putting the value of <1>0 and variables involved are not widely different from the conditions
80 x 10-3 . under which the data were collected. However, it is also possible
... (vii)
<I>
= 1.868 = 0.0428 radIans to determine angular distortion based on the temperature distri-
bution and other fundamental considerations. Brief treatment for
 294 Welding Engineering and Technology Welding Stresses and Distortion 295

the determination of angular distortion, in Butt and Fillet welds,

based on fundamental consideration of temperature distribution b ::= 2y :::::
O.!84 . qlin ...(4.48)
T • cpT 0
are given in this section. For an approximate determination of the temperature
Basic Considerations variation over the width of the strip in a given t.ransverse
For determining distortion in welds it is first required to know cross-section (normal to the weld axis), one can start from the
the heat input per unit length i.e. linear energy and from that to assumption that the temperature variation follows a normal
distribution law (Fig. 4..57)and can be expressed as, .

D
determine temperature distribution around the weld. T

v 0.24lV!l
Now, Linear energy, qlin;::: ll.::::: v ...(.4.45) (0 )
where, q ::::: heat input, cal/sec
v ::::: welding speed, cmlsec
V:::: arc voltage, volts
I::::: welding current, amps

11 :::::efficiency ofheat transfer.

Also, from equation (2.48), we have,
-.1!....(x+R)
q 20.
T:::::21tkR .e ... (4.46) y y
where T:::::temperature at the point of interest,
q ::: heat input (effective power of the heat source), cal/sec, I b)
coefficient of thermal conductivity of the material to
k :::::
be welded, call cm-sec-°c, y

R
R ::::: + .y2 + z'2-~cm
welding velocity, cmlsec,
v :::::
Fig. 4.57. Isotherm representing the weld pool, and temperature
a = -~, coefficient of thermal diffusivity, cm2/sec, variation across the width of plate.
cp
c = specific heat, cal/g-OC, 2 2

Ty:::::T·e-ny ...(4.49)
p = density, g/cm3,
2a' where, Ty = required temperature at distance, y, along y-axis
b ::: --~,
cpu coefficient ofloss of temperature, which is at a distance Xm from the arc, along x-axis.
Ct.'::::: the coefficient of heat emission, cal/cm2-sec-oC,
T = maximum temperature along the weld axis at a
8:::::plate thickness, cm, distance Xm from the arc,
+y ,em.
.r'Z'"~
r:::::''1x
n :::::
coefficient characterising the degree of dispersion of
temperature distribution.
Now, temperature at high welding speed, v, for points For a point located along the weld axis in the cross-section at
sufficiently far removed from the heat source is given by the
following relationship, a distance Xm from the arc (r = Xm), the temperature is given by,
v
T_ q ..Jfi
J~' - 2a (x +r) - 2ko ...fiWxm
T= q"la .e ...(4.47)
2ko ..J1tvr
_ q . 1
The width of the zone of heat. penetration, i.e., weld width can - 20 -Y1tkcp -Yv, Xm •.• (4.50)
be expressed as,
t
 296 Welding Engineering and Technology Welding Stresses and Distortion 297

V 2 A value of a = 0.000012 is assumed for linear expansion for

= -2ex (0.5 bT m )
Xm ... (4.51) ordinary low carbon and low alloy steels.
.. Putting (4.51) and (4.48) in (4.50), we get after simplification, For more accurate determinations, an allowance is made for
T = 1.66 T m ••• (4.52) the variation of the coefficient of linear expansion with
To determine n, substitute the value of T from (4.52) in (4.49) temperature such that the relative thermal deformation can be
and let Ty = Tin' so that, expressed by,
Or.t = 11475 X 10-9. T + 5.3 X 10-9 . r
(4.55) ...
Ty = 1.66 T m . e- n 2
2 [bTmJ2 For austenitic steel the coefficient of linear expansion is
assumed to be approximately 1.5 times greater and thus
Now, because Tm = Ty a = 0.000018 - 0.000020.
Tm
2
Angular Distortion in Butt Welds
e- n [b 2 J2
= _1_
1.66
=0
.U
c.
When fabricating structures by welding local deformations in
n2b2
the form of buckling and warping occur as well as the general
Tm deformations.
t.e., -4--=0.5 ; The reasons for this are :
or n 2 = 0.52X 4 ... (4.5 3) (i) the two halves of the plate are rotating relative to one
bTm another, about the point of welding (Fig. 4.58(a)) due to
Substituting (4.52) and (4.53) in (4.48), we get,
-0.5 b Tm
Tb = 1.66 T m . e r b
~ J2
... (4.54) IQI~
Using (4.54), a curve can be plotted representing the -o
; nf
•...
lbl L2:_~[
temperature distribution in the cross-section where the element -:;;
located at a distance bTm from the weld axis has a temperature 'i5
•...
Tnt'
The thermal properties of materiels are very strongly depen-
:;o
0>
c:
dent upon temperature. When determining the temperature 40{

reached during welding some mean value is assumed for the Plate thickness,6 --e-
thermal properties and these are indicated in Table 4.7. Fig. 4.58. Angular distortion, (a) for bead-on-plate, and (b) the
Table 4.7.Thermal Properties of Some Important Materials relationship between plate thickness and angular distortion
for different values of energy input per unit length.
S. Low
Cr-Nicarbon
Chromium
Medium
Aluminium
alloy 0.95
0.90
0.09-0.11
0.079-0.096
0.096
0.04-0.06
carbon
steel
u=-
ksteel
0.085
steel
0.035-0.053
0.070-0.079
0.053
0.06
0.63
1.00
0.08-0.09
Material
k steel cp
(Callcm-OC)
steel 0.95
1.14
0.65
cpRed Copper (Callcm3-oC) non-uniform temperature developed through the thick-
(cm2/sec)
ness of the plate and to plastic deformation during heat-
ing (so called angular deformation), and,
(ii) the loss in the strength of the welded plates as a result
of the longitudinal contracting stresses set up during
welding.
The first cause applies in all cases, affecting both thin and
:1
'I'L
thick plates. The variation of the angular deformation, 13, with the
plate thickness is of the form shown in Fig. 4.58 (b). The angular
deformation is small, both for small plate thicknesses, due to the

lil:i
·.'i.,,~
298 Welding Engineering and Technology Welding Stresses and Distortion 299
1.6
relatively uniform heating, and for large thicknesses-for the same
energy input per unit length - due to the greater rigidity of the m : -.!.i.-,.4
plate. Thus the angular deformations are maximum for medium "" Tb
t:.av
thicknesses. Increasing the energy input per unit length from 6 r :a.l.b
<ilinl to qli1l2 displaces the curve towards the larger thicknesses i.e.
1.0
it reduces the angular deformation for small thicknesses and
increases it for large thicknesses.
The second reason applies mainly to thin plates, less than 8
to 10 mm thick, for which the critical stresses leading to the loss
of strength are below the yield point.
RELATIONSHIP BETWEEN ANGULAR DEFORMA-
TIONS AND TRANSVERSE CONTRACTION, AND THE
SHAPE AND DEPTH OF 'rIlE ZONES OF WELD 0 _
PENETRATION
0.2 0.4 0.6 0.8 1.0 0 0'2 0·4 0·6 0·8 1.0
If ~au is the value of transverse contraction, and the angular (01 p/6- (bl
distortion is angle ~, then, Fig. 4.59. Graphical representation of relation between:(a) plO & m, and (b) pi (,
& r for triangular, parabolic and rectangular shaped weld beads.
~o+~l:i
~au = --2- .b ...(4.56) Transverse contradions increase with the increasing depth of
~ -~3 penetration, and attain a maximum value when the fusion extends
and angle ~ = 0 0 . b ...(4.57) over the whole thickness.
where Ao - contraction at top of the plate, It follows further from equation (4.58) that for a given'
~3 - contraction at buttom of the plate. penetration, i.e. for m = constant, the angular deformation ~ is
b - plate width, greater the greater the ratio between the width of the bead and
0- plate thickness.
the plate thickness (%)
Also, if
The transverse shortening for a given penetration (for
m=-L ...(4.58) r = constant) increases with an increasing bead width, b. Angular
aTE. Deformations for other shapes of weld bead can be determined
o
similarly. Values of m and r for a parabolic bead shape are shown
and r =
L\.v ...(4.59) in Fig 4.59 curve 2, and for a rectangular zone of penetration (i.e.
aTb weld bead) in curve 3.
where 'a' is the coefficient of linear expansion, then the relation The corresponding expressions for m and r for a parabolic
fused zone are :
between m & ~ and r & ~ are as shown in Fig. 4.59, where p is
...(4.60)
the weld penetration. It can be seen from these graphs that the
maximum angular deformations occur for incomplete fusion of the m ~ 2 ~ [3 - (1 - k)' - ~ {3- 2 (1 - k)' + 0.6 (1 - k)'l]

I
plate, i.e. ~. < 1.0. For complete penetration (i.e. = 1.0) cor-
responding to butt welding, the angular deformations are less than
those during deposition of a bead with penetration extending to
r = ~ (1- k) [1- -~(1- k)2]
and for a rectangular fused zone,
...(4.61)

...(4.62)
more than half the plate thickness . m =6~(1-~Jel-k)
r
11
i:l

:, • 300 ne Id'109 E'ngmeermg

m . an d T.I.echno Iogy Welding Stresses and DistorUon 301

k - weld bead shape factor

r=~(l-k) ...(4.63) Factor k for different bead shapes are given in table 4.8.
Values for weld bead shape factor, k are shown in Fig. 4.60 Table 4.8. Values of Shape Factor k for
wherein curve 2 is for a parabolic weld bead shape, and curve 3 Different Bead Shapes

,,
for a rectangular weld bead shape.
Rectangular I Semi-circular I Parabolic I Triangular
1.0I ~
'\. \ --
.\\
~\ \ 1
3 \ \
0.6I ....•...•
\2
k 1 0.78 I 0.667--.1 0.5
0.6
..:.::

0.4 The ratio ~ in terms of the cross-sectional area of the fused

zone (weld bead) is given by,
0.2
E- - ~pen_
o - k .b .0 ....
(466)
o
o 0·2 0·4 0·6 0·8 1.0 From equations (4.64) and (4.66), we get,
~:;:J:._
6
E- = qUn.pen (4 67)
o f·k·b·o ....
Fig. 4.60. Effect of weld penetration to plate thickness ratio (P/O) on the
value of weld bead shape factor (k). Since the coefficient k varies with the shape of the fused zone
APPROXIMATE METHOD OF DETERMINING (i.e. weld bead), to ensure the same penetration for differently
ANGULAR DISTORTION IN BUTT WELDS shaped weld beads it is necessary to deposit the weld bead with
different energies per unit length, qlin.pen'
To determine angular deformation and transverse contrac- The angular and transverse deformations can thus be
tions from Fig. 4.59 it is necessary to know the relation between
the depth of penetration, the plate thickness, and the shape of the expressed by the ratio qlin.pen/f.b.o in accordance with equation
zone of penetration i.e. the weld bead shape. For convenience in (4.67). Using the values of m and r obtained from Fig. 4.59 the
use it is advisable to express the depth of penetration in terms of variation of the angular and transverse deformations with the
the unit length energy, qlinpen for the given welding conditions, as energyllength is expressed in the form of curves reproduced in Fig.
4.61 for differently shaped weld beads.
expressed by the following relationship,
1 It is evident from Fig. 4.61 that angular distortions depend
substantially ml the shape of the weld bead for all values of the
Apen = 7' qlinpen ... (4.64)
energyllength, whereas transverse deformations are independent
where, Apen area of weld bead at a transverse cros8.section
==
of the weld bead shape for ratios,
qlinpcn = the energyllength used to fuse the base metal, and
f= coefficient determined by the welding technique and
qlin.pen
(.b.o ~
'.0 ')
.L1
conditions and is equal to 3900 cals/cm3 the bead shape having a small effect only for smaller values.
In the general case, Apen is given by, If one assumes that the most probable shape of the fused zone
Apen = k . b .P ... (4.65) is parabolic, and the width of the bead, b, in terms of the plate
thickness, 8, is,
where, b, p - width and depth of penetration respectively, ern
b=<;·o ...(4.68)
 302
Welding Engineering and Technology \Velding Stresses and Distortion 303

••• p q /in.pen q lin.pen ...(4.70)

-5 = 2600 --:Tf; 2600 . <; . 82
" ~18
==

6> 0 0 0
-"- - ~N
~
0 ~I:g where it is apparent that m is some function of q lin.penl 02. Knowing
"-
"-

"-
'-- ~ "- ',$ <D <Cl Ei ~

~
8
o the relationship between m and~, if only from Fig. 4.59 (a) curve
M
2, one can establish a relation between the angle 13and qlin.pen1f?
oo "OU-g
00" for different values of I; i.e. for different bead widths, as shown in
<:>
C"'-' .D.o
••••• ~
••• .c Fig. 4.62.
o .-u :;0
•.• c..
.:a 0.026
to
0:;00"
8
_ g~.2
tTl~C:
._
•.• 0 ••
•...
a.a:
u -
:;:;
oc::
ell ....
:=:-0
0.022 Ir~
p~
i
I i I '-
o
c::
ell
:...2.
5
,,1~ q 00 <D -{ N o -NM .S
C •
t
<II ••.•
[!o
6 ci o ci (.::>

~]
IS

o
0.018
c::
'0
.D
Ol
~ 0.014
o ~<-O
l::~
c..'
., "
-- clo
&~ ~I~
•.• 1(, 0.010
.- ~c
"§

t-n
='

..c:
I .~ __ c.."
j.--}::. - - ~ .- 5' o oj

e;~.,
.,

'- M _ i- . - -.. '-- +- 8

M ,...;..0 0.002
••• ~ _ CIO <0
.;.: o
, I~' N.L-- - 0 8 o
-' , /
_ -~
~-
o
0 N
o
"'"

ti:
2500 3000 qlin. pen.
Fig. 4.62. Effect of Qll'lpell/62 on angular distortion for
different valuel! of the ratio blo i.e. 1;.
G2
oo
=: .:;:._ -.;. t-- ~ S It is important to bear in mind that the determination of angle
-_J
i •••..•

"-- - - - - 13must not be based on the tot.al energy per unit length used for
heating the plate, but only the fraction used for fusing the base
. II"-..j...- - .~ o--'
I - - •.. 0 o ":"
"! q co N metal to form a part of the weld bead. This is done by introducing
~~,,0... ...:
d ~ -0$
ci d into equation (4.45) a coeffi.cient 11= 0.3 or for fillet welds by
E multiplying the values obtained from equation (4.75) by 0.43.
LINEAR ENERGY FOR I"lLLET WELD (qlin.r)
When making a single pass fillet weld with a leg length Z, the
then 13in equation (4.58) is given by, weight of the metal which must be melted per cm of weld is given
b by the following relationship,
13 = maT-8 = m;aT ...(4.69) Z2

Wf= P . Ar= P . \If' "2

...(4.71)
where m is determined u whkh can be expressed in
by the ratio ~. where, p = density of weld metal, g/cm3
terms of the energy per unit length. Substituting for (4.68) in Ar = weld crosg.-sectional area, cm2
(4.69) gives, for a parabolic shap(~d fused zone,
 304 Welding Stresses and Distortion 305
Welding Engineering and Technology
" -- r
\jI = coefficient bf fulness of fillet weld, a measure of the dehte. = 6mm
0.3
concavity or convexity of the weld surface, varying
between 0.8 and 1.2. 0.2
0.1
Also, the weight of metal deposited/see in a fillet weld is given
by,
ad' I
-
, o
o 200 400 600 800 1000
G==--g/sec
3600 ...(4.72) ~ tt I
J delee• = Smm
>- 0.3
and hence, weight of metal deposited/cm of weld length is, v
ad' I c'" 0.2
G 3600 :-v g/cm ...(4.73)
v 0.1
- 600
==

'Where, ad ==deposition coefficient, g/Amp-hr .;:0

':;;0 200 400 800 1000 1
I
= welding current, Amps
v ==welding speed, cm/sec. '" " delee. =4mm I
From equations (4.71) and (4.73), we get, ':ii 0.3
I c
ad'
3600v == P . \jI .
Z2
2 ~
-o
~ 0.2
0.1
3600 Z2 o
I==--,p'\jI'-
ad 2 ...(4.74) o o 200 400 600 800 1000
-
'"
:x:

I from 1\
.. Putting this value
'We get,
equation (4.74) in equation (4.45),
0.3
0.2
:::::;-
delee.=3mm

-
qlin == 0.24 ad - . p .
(3600
\jI . TV.
Z2)
11 0·1
o
Now, for fillet weld 11 ==0.7 and p ==7.78 g/cm3 o 200 400 600 800 1000
V 2
Welding current, Amps
qZin(ji.llet) == 2350 \jI . - . Z ...(4.75)
ad
Fig. 4.63. Relationship between welding current (I) and heat
More exact values for 11, when determining the area of transfer efficiency (T\) for different diameter electrodes.
penetration in the base metal can be obtained from Shraerman's
with increasing values of qlin.pen/fj2 while for various values of
graphs given in Fig. 4.63 based on diameter of the electrode '-:Ised.
It follows from Fig. 4.62 that when welding with values c; ==~ it varies according to the dashed lines in Fig. 4.62.
f < 0.6 and ~ < 1.5, the angular deformations extend to some It should be borne in mind that as the penetration approxi-
regions of the diagram which. as a very rough approximation, can mates to ~- ==1, the zone shape will vary from parabolic to
be expressed by a linear relationship of the following form, rectangular (especially for small thickness). For a rectangular

10.5 X 10-6 . (!Ji;::..,::

(30 == ... (4.76) zone ~ is given by,
When the depth of penetration exceeds 0.6 of the plate p qlin.pen qlin.pen
...(4.77)
'& ==3900~T8 ==3900· C;. fjt
thickness (i.e. ~ > 0.6) the angular deformation may decrease (.,' b ==1:,. fj as per equation (4.68»
 Welding Engineering and Technology 307
Welding Stresses and Distortion
Thus, the ratio between the energy per unit length, and the Thus, when depositing a bead on a plate of 20 mm thickness
square of the thickness becomes, the angular deformation increases with increasing length of plate
(and bead) up to 50 em. Increasing the plate length ( and bead
qli~p:z: 3900 1; . ~ ... (4.78) length) beyond this has no further effect.
For 1= 550 Amp, V = 34 volts, v = 1.22 em/see, 8 = 2.0 em and
For the corresponding values of 1; and ~ the angular 11 = 0.7, the length X below which one can assume simultaneous

deformation is determined from equation (4.69). deposition over the-bead length, in accordance with equations
(4.79) and (4.45), will be,

I.
2
If one assumes that when ~ > 0.6 there is a gradual transition
X= 2.89 X 10-6 (q~n) v 2
:from parabolic (i.e. ~ =~ = 0.6) to a rectangular zone of pene-

tration (i.e. ~ = ~ = 1.0) the relation between 130and qZin.penl82 will

=:2.89 X 10-6 (0.24Xl~2~: ~~oox 34 J . 1.22
= 5.78 em.
be represented by the dotted lines in Fig. 4.62. Relationships
,..!?etween qZin.penl82 and 130 for rectangular shaped bead zones, Now, the critical length for maximum angular deformation is
obtained by using equation (4.78) for 1;= 0.6 and 0.8 are shown in approximately 9 times X, i.e.
Fig. 4.62 and are numbered (1) and (2).
The above described relations are valid for the condition of
simultaneous deposition of a bead over the full length of the plate.
I"" ~ 9X = 9 x 2.89 X 10-6 (q~.J .v [From (4.80)]
Such conditions can readily be visualised for relatively small plate
lengths (and bead lengths) not exceeding a value X, where X is the = 26 x 10
-6 ~qZinJ2
v 8 . ...(4.81)
distance between the arc and the start of the weld. and the relationship may be aken as generally valid.
In general, X is given by the relationship, The ratio between maximum angular deformation, 13max,and
130(for simultaneous deposition over the full bead length) must
also remain constant, independent of value 130' On the basis of
Xl= - n'8
(qZinJ2
.v ... (4.79)
. experimental results of Baikova, we have,
where, C = specific heat,
13max= 3130 ...(4.82)
p = density,
8 = plate thickness, and The relationship between angular deformation and the length
is given in Fig. 4.64. The angular deformation equals 130for lengths
v = welding speed.
By putting the values of c and p for steel, X is given by the
c:
-
''0
-<
'3
.2Cc
•0
c:
E
C7>
~ f1l
•..
•..
••

following relationship,
-6 qlin
X = 289 x 10 ( 8
J2 . v ... (4.80)
For plate lengths greater than X some part of the plate will
(lmox

always be in an elastic state during welding thereby increasing

its rigidity. Greater rigidity leads to an increase in angular Icrit. 90 X
deformation 130, analogous to increase in thickness.
Plot~l~n9th.l-
Fig. 4.64. Relationship between plate length and angular distortion ..:
 309
308 Welding Engineering and Technology Welding Stresses and Distortion

smaller than X it can be determined from the following approxi- Contraction of elements parallel to the hypotenuse will be
nlate expression, proportional to their lengths. The shrinkage produces a relative
rotation, WI of the two plates adjacent to the weld, and is given
by the following equation,
13l = 0.25· 130(f + 31 ...(4.83) L1
which assumes that 13l varies linearly'between lcrit and X.
WI = 0.7l = 2aT ...(4.85)
4.6.3.5. Deformation in T-and I-Sections When Flange
and Web are Joined by a Single Fillet Weld It is apparent from equation (4.85) that WI is independent of
When a single fillet weld is deposited the deformation of the the leg length of the weld so that the same twist will be produced
flange will .be made up of two components viz., in the flange for any dimension of fillet weld, if the web is fixed.
(i) a twisting due to metal shrinkage in the fillet weld, and The angle, 81> by which the flange is bent is determined by the
(ii) bending due to depth of penetration of the melted zone. depth of the penetration.
These two components are represented in Fig. 4.65. Assuming For a given leg length, l, the total energy per unit length of a
no bending (i.e. complete absence of penetration), the shrinkage weld is given by equation (4.75). The depth of penetration in the
flange is proportional to the energy per unit length and is given
El, by Cle following relationship,
20fl ...(4.86)
qlin.pen = 20 f! + 0w >< 11m . qUnifillet)

where, ofl = flange thickness,

Ow = web thickness,
11m = 0.43 and is a r.<easureof the fraction of the heat used
in melting the base metal, and
qlin = the total energy/length defined by equation (4.45).
0" 011 20
021 (JI~ 011 I 9, Putting20fl :: Ow i, equation (4.86) reduces to,
...(4.86 (a»
q lin.pen = i . 11m . q tin
(e I (d I (e I
Fig. 4.65. Schematic representation of shrinkage From equations (4.75) and (4.86), we get,
and bending in a single fillet weld in T-joint . ...(4.87)
qlin.pen = 2'"Uf! + Uw
",' 11m' 2350",· -ad l
of weld metal during cooling will produce a twist WI' although the 2J ds), given by
flange remains straight [see Fig. 4.65 (a)]. Assuming only pene- The maximum value 20f!
of angle 130
( (for longv we
tration of the flange plate without any fillet weld, then bending. equation (4.82) is,
by an amount 81, will be produced giving the shape as shown in 13max= 3130
Fig 4.65 (b). and putting the value of 130from equation (4.76) in it, we get,
For single pass fillet welds (most cases) , WI can be determined R = 3 ><10 . 5>< 10-6 qlin.pen
02
from the shrinkage of the hypotenuse of the fillet which, for a leg I-'max
length l, is 1.4 l. The contraction on cooling through TOC wiH be, ...(4.88)
L1 = aT· (1.4l) ...(4.84) = 31.5 ><10-6 °
qlinfen

where, a = coeff.oflinear expansion, and Substituting equation (4.87) in (4.88) and putting 11m'; 0.43,
T = Temp. rise of weld metal above the ambient temperature.
we get,
 310
Welding Engineering and Technology ~Hl
Welding Stresses and Distortion

_ -6 , 28fl V [2
For WI < a), I'll will again be negative, but 1'21will also be
31.5 x 10 x 2350 x \jI x 0.43 x -2~~-
13max - 8 ...(4.89)
ad '-~2
ufl + Uw x -- negative (Fig. 4.65 (d». When wI and 81 are equal the flange
Assuming, for average conditions, assumes a position shown in Fig. 4.65 (e), and therefore,
tan \jI::::: 1 I'll = - wI' 1'21= WI - 81 = 0
28
_ ..'1L--08 Yet in all cases,
28fl + 'Ow - . 81:::::I'll + 1'21 ...(4.95)
V
-=2.5 For approximate determinations of the deformations, one can as-
ad sume aT = 0.012.
Putting these vailles in equation (4.89), we get, Values ofY11'and consequently for 1'21'vary within wide limits
A. -
Pmax -
64 x·1O-3. i~
02
...(4.90) and are determined by the conditions at the start of the welding
and the assembly fitting, especially by the number of clamps used.
The angle through which the flange is bent by the deposition Calculated values of I'll and 1'21must, therefore, be considered
Qf the single fillet weld, is then given by either of the following extremely approximate.
two equations. 4.6.3.6. Deformation of a Flange Plate Joined By Double
Fillet Welds to the Web of aT-Section
81::::: - 31.5 X 10-6. Qlin:e'!:.
o ... (4.91) When a flange plate is welded to the web of a unit of a T- or
2 I-Section with two fillet welds, the conditions for deformation are
different from those when only a single fillet weld is used. The
8 = - 64 X 10-3 J (i ...(4.92) difference arises because when the second fillet is laid the flange
Actually, both a tw:ist and a bend are produced, resulting in a is already rigidly attached to the web by the first weld. Even if
total deformation determined by the ratio 8)/w1• the deformation of the flange plate during the first weld does not
set up stresses in the weld, stresses will be set up in both welds
For WI > 81(Fig. 4.65 (c» the flange plate, originally in position during deposition of the second weld.
1-1 will take up pqsition 2-0-2 determined by the angle WI due Let us suppose that after depositing the first weld, the flange
to shrinkage of the weld metal. Penetration of the flange produces has deformed to an extent defined by the angles I'll and 1'21'as
a bend $1' (supposing provisionally that the weakest point is at 0, shown in Fig. 4.66 (a). When weld 2 is deposited the right half of
and the free half of the flange (that not joined to the web) will be the weld deforms angularly and the flange plate will tin ally occupy
bent through an angle aI' relative to the attached half, assuming position 3-0-4 with both welds extended to the same amount. The
position 2-0-3). Thus, relative to the initial position, the left half second weld would thus modify the angle I'll by 1'12and angle 1'21
_of the flange (containing weld 1) will be bent through an angle by 1'22'
'Yll given by,

I'll = WI ... (4.93)

while the other will be twisted through an angle 1'21'given by,
1'21= WI - 81 ...(4.94)
The angles I'll and 1'21are negative when the corresponding
halves of the flange bend downwards. relative to the initial
position. For WI> 81, therefore, I'll will be negative as given by
equation (4.93), and 1'2)positive, as shown in Fig. 4.65 (c). 10 ,
{bl
Fig. 4.66. Schematic representation of shrinkage and bending to cause angular
deformation in a double fillet weld in aT-joint.
312 Welding Engineering and Technology Welding Stresses and Distortion 313

The total deformation due to the two welds will thus be, o - flange thickness,
For left fillet, 11 (1 + 2) = 111 + 112 (4.96) 0web - web thickness.
For right fillet, 12 (1 + 2) = 121 + 122 (4.97) The sign in equation (4.101) is determined by the sign of the
In 11 and 12 - 1 indicates left-side fillet weld and stresses (i.e. + for tensile and - for compressive).
2 indicates right side fillet weld. The angles of twist of the two halves, of the flange, relative to
while in (1 + 2) - 1 indicates first weld and 2 the second weld. their initial positions are:
For example, 112':"- means deformation in left ~ide due to .weld 11 (1 + 2) = (2aT- \32) B - 2aT ...(4.103)
on right side. and 12 (1 + 2) \32 - (2a.T - \32) (G + B) + 2aT - \31
=-= -
...(4.104)
Thus,
81+2 = 81 + 82
where, B=-- l 2
...(4.105)
0.73'o (30web + 2l) + 4
= (r:11 + 121) + (112 + 122)
= (111 + Y12) + .:Y21+ Y22) For Simultaneous Deposition of Welds
Putting the values of (111 + 112) and (121 + 122) from equations During simultaneous deposition of two fillet welds the
(4.96) and (4.97), we get, deformation of the flange will develop which results in the terms
Thus, 81+2 = 11 (1 + 2) + 12 (1 + 2) ...(4.98)
G and B being replaced by the expressions G' and B' respectively
and these are represented as follows.
The mathematical treatment of the elements of these
equations lead to simple expressions both for the two welds (1 and G'= 1 ...(4.106)
2) laid successively or simultaneously as follows. 1+-- l 2
For Successive Welds 0.7"3 (30web + 2l)
o
The total deformation of the flange plate during the successive
deposition of two fillet welds is given by the following expression, and B'- 2. ...(4.107)
- l2
81+2 = - \31 - \32 - (2aT- \32) G ...(4.99)
0.7 03 (30web + 2l) + 2
where, \31' \32= angular deformation for welds 1 and 2 respectively. when the stress in flange (0') equals the yield point, then,
If the two welds are of the same size and laid by the 8 = ± 2ep
13
same welding parameters then \31 = \32; and
1 where, D is determined from equation (4.102).
In this case, on the basis of symmetry, the angles of rotation
relative t.othe initial position of the two halves of the flange plate
+ 2l) must be given by the following relationship.
G ~ 1 + 0.7 (')~3 (30web
4 ...(4.100) 1
or 81+2=-\31-\32±Es·D ... (4.101) 11 = 12 = -2 8 ...(4.108)

where, Es =; (for yield point) Problem 4.7. Find, (a) the longitudinal contraction, and (b)
l2 the angular deflection in the flanges of the main bridge girder of
and D = 0.7"3
o (30web + 2l) ...(4.102) H-Section, shown in Fig. 4.67, if welding is done by automatic
mode with arc voltage of 35 volts, deposition coefficient. ad = 14
In all these equations, gm /Amp-hr; and the heat transfer coefficient for melting the base
l - leg length of fillet weld, metal, 11m = 0.43.
 Welding Stresses and Distortion 315
314 Welding Engineering and Technology

430 x36

4
ii
CTI

',.;r
K: 12
3S2 x28
_ ..J...

430 x 36
Fig. 4.67. Cross-section of main bridge girder of H-section.
Solution. (a) For Longitudinal Contraction K == .area of single fillet == ~ == 43.0 x 3.6
area of web Aw 38.2 x 2.8
From equation (4.75), we have for a fillet weld,
== 1.447"" 1.45 ... (iv)
q lin(fillet) = 2350 . \If. -~~. /2
Cl.d
...(i) and m
flange ~~ickness == ~
== ==. 3.6 == 0 942
web height Ohw 38.2 .
Now, the coefficient of fulness oftillet weld, \If, for welding in
an automatic mode = 1.0. From Fig. 4.68, for k == 1.45 and m == 0.942, we get,
12.1
Also, V == 35 volts, ad == 14 g/Amp-hr. -- == 11.5
oh3
Putting these values in equation (i), we get, o == web thickness == 2.8 cm
where,
35 /2 I/ and h == web height == 38.2 cm
qlin(Jillet) = 2350· 1 . 14' . ca cm
t" J ==} 1.51; oh3. == 11.5 x 2'~nx (38.2)3 == 149577 cm4 •.• (v)
§l The leg of the fillet weld, / == 12 mm, therefore the
energyllength, from (ii) above, is,
IS ~

-1 qli.n(fillet) 5875 (1.2)2 == 8460 caI/em

5875/2 ... (vi)
Fn=Kh&<f~ == ==

Because of the great rigidity of the unit, the total longitudinal

~
... contraction can be determined assuming that the welding
~ \,()16
....• \.J::. sequence has no effect on the final contraction or angular
deformation. It must, however, be borne in mind that the heat flow
12
zones on the flanges partially overlap as shownin Fig. 4.69, and
this is particularly valid for thick plate sections.
The heat flow zone of the first weld, which is not subsequently
overlapped, can be determined with reference to Fig. 4.69, from
the following equation,
F's\ iFs, - F'l- Fll ...(4.109)
o
0·5 1·0 1·5 2.0 2.5 K C, == y;;-
S, = ------ FSI

0 f we b'- where , C, Non-overlapped h~~t flow

of ~reayffirst
the first weldwe!~
f". \g. 4..68 Re I'atlOns I'up be t ween t'.actor K (.l.e. areaarea Total heat flow area
==
of single fillet)
and lz.;
8h
for different values of m.
316 Welding Engineering and Technology Welding Stresses and Distortion 317
'"
0'
c For T-Section
.S!

-f""':-'
...(4.114)
and r~ = 0.64 xx iFsl)
r~ == iFs2
From equation (4.109),
(0 i iFs - F[- FlI
1; == _I ji';"_
'Sl
when Fs == F's 2 i.e. for 1; = 1, we have,
iFs - 2F[
", 1;= I
Fs I
...(4.115)

The area F[ can then be determined directly from the graph

(bl shown in Fig. 4.70.
F
7'2

1.4

1.2
+x
(c)
1·0
!~r
r
:I
Fig. 4.69. Schematic tepresentati'on of overlapping heat zones for
fillet welds, (n) fillet nomenclature, (b) and (c) overlapping
heat jones for different ratios of heat input.

Supposing that the areas FSI and FS2 represent the areas of
semi-circles in Fig. 4.69 (b). The ratio, 1;, of the area of remaining
part of the zone due to the first weld (shaded area in Fig. 4.69 (b), 0·4
to the total zone area must be determined. It follows from the
nomenclature used in Fig. 4.69 (b) that, 0·2 -
F's I = F s I - F[ - FlI ...(4.110)
From the condition, at the point of intersection, both semi- - ~ -0'8 -0·6 -0·4 -0·2
r
0 02 0, 4 Q·6 0·8 '!:!
circles have a common ordinate y, and thus we find that, r
2 2 2 2
rl - Xl r2 - X2== ...(4.111) Fig. 4.70, Relationship between ~r and ~.
r
Also, X2 d - Xl==
...(4.112)
From equations (4.111) and (4.112), we get, For ...(4.116)
~ == 0.5'l1 :: 0.5 ~r == 0 ..5 . ~web r + -l
_ + rl ,- r2 The value of 1; can vary between 0 and 1. When the two heat
Xl - 2d
...(4.113) flow zones completely overlap, I; == 0; and when do not overlap at
d2 - rl2 + r22 alII; == 1.
and X
2----,----.
d2 2d2 2} When three welds are laid, three cases are possible,
----------------------------
318 Welding Engineering and Technology Welding Stresses and Distortion 319

(i) the region of the third weld does not overlap the heat From equation (4.116), we have,
flow region of the preceding welds (Fig. 4.71 (a)),
~r = 0.5. r+ l
.
= 0.5 2.83+21.2=
'Ow 0.625 ... (xi)
(OI~2 3~ . Frx
From FIg. 4.70, for - = 0.625,2' = 0.20
r r1
I hI!. 3c1 .. Fr = 0.20 x 10.24 = 2.048 cm2 ... (xii)
1 2
.. Using equation (4.115), we get by putting values from (ix),
(xii) and (vii),
iFsl- 16 - 2 (2.048)
2Fr
...(xiii)
(Cl~~ 2~
~ = ~-;- 1 22.23
=. = 0.535
Fig. 4.70. Three sequences of laying welds in fabrication of H-section girder.
Now the total longitudinal contraction, along the axis of e.g.,
(ii) the region of the third weld overlaps that of the second due to the two pairs of fillet welds is given by,
weld (Fig 4.71 (b)), and D..c.g'(2) = '-"e.g.
A (1 + J:)
~ x 2
...(4.118)
(iii) the zone due to the third 'weld overlaps that of the first, where, D..c.g. = the longitudinal contraction along the axis of e.g.
Fig 4.71 (c).
- 3.53 x 10-6 X qlinifillet)
...(4.119)
On the basis of above considerations, the value of Fs I can be - A
determined ftom the following ~quation, Putting the values of qUn and A from (vi) and (Ui) respectively,
we get,
Ji's = 1 Z' 340 ~ ... (4.117)
1

A
---+-+---
J qUn
D..
e.g.
= - 3.53 10-6 X 8460
X
416.6
where Z' = distance between the weld and e.g. of the section (19.1 =-7.1684x10-5 ... (xiu)
cm in the pre!,ent case) . .. tlC'g'(21 = Ac.g. (1 + c,) 2 (from 4.118)
Putting the values of A, J and qlin from (iii), (v) and (vi), we Putting the values of !::I.c.g. and!; from (xiv) and (xiii) respec-
get, tively, we get,
F ------
1 D..c.g.(2) = - 7.1684 X 10-5 (1 + 0.535) x 2

Sl - 1 (19.1)2 340 = - 220 X 10.6 •. ,(xv)

416.6 + '149577 + '8460 The total contraction (M:.) of girder of length 10980 mm is,
1 2
...(vii)
therefore, given by,
.- )24 + 0.0024 + 0.0040 = 22.23 cm D..L = D..C.g'(2) xL ... (4.120)

Now, i= = 2 x 3.6
2'6fl = 0 72 ...(viii)
= - 220 X 10.6 x 10980
+ 'Ow (2 x 3.6) + 2.8
2'6ft . =2.4156mm
From (vii) and (viii), we get, Longitudinal Contraction = 2.4156 mm
.. iFs I = 0.72 x 22.23 = 16.00 cm2 ...(ix) (b) For Angular Deflection
For successive deposition of two fillet welds the deformation
Also, from equation (4.114), we have, is given by equation (4.99) i.e.,
r~ = 0.64 x iFs I 81+2= - 1)1- 1)2- (2aT - 1)2)G ...(4.99)
= 0.64 x 16 = 10.24 cm2 Assuming that both fillet welds are made under the same
r1 = r = 3.2 em ... (x) condition (i.e 1)1 = 1)2 = I) then, if aT = 0.012, we have,
320 Welding Engineering and Technology Welding Stresses and Distortion 321

81+2 = - 213- (2 X 0.012 - 13)G qUn.pen = i . 11m . qlin = 0.72 X 0.43 X 8460
The value of can be determined approximately
13 from equation = 2637.8 cal/cm ... (xx)
(4.90) or even more accurately with the help of equations (4.67)
qlin.pen -_ 2637.8 _ 203 5 ca 11cm 3 ...(xxi)
and (4.80) in conjunction with Fig. 4.62. 8fl
2
(3.6)
<) - •

From equation (4.90), we get,

Now ~ ;: ~ length of fillet weld = -L
flange thickness 8fl

where
13max:::64 X 10-3
k::: web thickness, and
(~J £.e. ...(xxii)
1; :::; ~:~ = 0.333
8 :::flange thickness
From Fig. 4.67, we have, :. From Fig. 4.62, we have for 1;= 0.333 and
k::: 1.2 cm
8= 3.6 cm qlin;:en = 203.5,130 = 0.0023
8fl
Hence, from equation (4.82), we have,
., 13m
ax :::64 X 10-3 (~:~ J :::0.007
To determine 130 from Fig. 4.62, it is necessary to determine
... (xvi) 13max= 3130= 3 X 0.0023 = 0.0069 ... (xxiii)
More accurate values can be obtained if the energy per unit
the value of qlin,penl82. length corresponds to the actual welding parameters rather than
Now, from equation 4.86 (a), we have, to average values based on a given weld leg length.
qlin.pen = i ' 11m . qlin Knowing p we can determine 81+2 by finding G from Fig. 4.72.
As 11m::: 0.43 (Given) G 1.0.a
qlin.pen ::: 0.43 ' i . qlin ... (xvii) 0.9
V 2
(from 4.75) 0.8 Y::l.0 G
Now, q linifillet} ::: \jI . 2350 . --
ad , l 0.7 Y:0·5
Y::2.0l
As \jI ::: 1.0, V = 35, ad == 14 g/A-hr 0.6
35 2 0.5
1.0 x 2350 x 14 ' l
qlmifillct} === 0·4
Now leg length, l ='1.2 cm 0.3
. 35 '. 2 0.2
qlin(fillct) = 2350 x 14 x (1.2) 0.1 Y: 1.0
O'S} B
Y:=2..0
Y
= 8460 cal/cm ...(xviii) o
Now from equation 4.86 (a), we have o 0.5 t 1.0 l.S
'6
. 2,----
8fl
£=-
+ 'Ow 2 8fl
Fig. 4.72. Relationship between i and Band G parameters,

and
Flange thickness, 'Ofl::: 36 mm
Web thickness, 'Ow = 28 mm
For 1;= i = ~:~ = 0.333
, 2 x 36 72 and v - 8web - 2.8 - 2 3
£ = ------- :::---.-----.-- == 072 ... (xix) - 1 - 1.2 - .
(2 x 36) + 28 72 + 28 '
Putting the value of q/in from (xviii) and i from (xix) in (xvii), From Fig. 4.72, for ~ = 0.333 and v = 2.3, we get,
we get, G = 0.06 ...(xxiv)
322 Welding Engineering and Technology Welding Stresses and Distortion 323

Hence, putting the values of 13max from (xxiii) and of G from Amount of distortion depends on the shrinkage moment and
(xxiv) in equation (4.99), we get, the resistance of the member to bending, as indicated by its
81+2 = - 213 - (+ 2aT -13) G
moment of inertia.
= - 2 x 0.0069 - (2 x 0.012 - 0.0069) 0.06 >... A Neutral
= - 0.0138 - 0.001026
= - 0.014826 ""- 0.015 Ql+2 = - 0.015 = 0.86° L--75'- "- /
T</~~ '75 -:7 r----.-,::3
C
\-=L : 635--t
The slope of one-half of the flange i.e. Yl (1 + 2) given by equation Y: I : 5mm
(4.103) is, d :.21.34mm Calculated deflection: 3.03mm
Yl (1 + 2) = - 2aT + (2aT - 132) B ...(4.103) Imin : 171315 mm4 (ends up)
Now for same parameters of welding 132 = 131 = 13 (a ).
.• Yl (1 + 2) = - 2aT + (2aT - 13) B --I 1--6
From Fig. 4.72, for 1; = 0.333 and v = 2.3, we find
B = 0.47 .,.(xxv) 1 : 7 mm
t . Neutral
Hence,

••
Yl (1 + 2) = - 2aT + (2aT - 13) B
= - 0.024 + (0.024 - 0.0069) x 0.47
;:= - 0.015963 ""- 0.016

Yl (1 + 2) = - 0.016 = 0.92° Yl (1+2) = - 0.016 = 92°

~o
-----!--L
r axis
L
I L:915=-\
-::\

Calculated deflection: 2.93 mm

4.6.4. BOWING OR LONGITUDINAL BENDING (ends up)
( b)
When the weld line does not coincide with neutral axis of a
weld structure, as shown in Fig. 4.73, longitudinal shrinkage of
,- 5 x 100 plate

C -:::J

p~~
__
[~ dis tortion
[Longitudinal ( mm) \" d (mm) 91.45
L_ I-beam I. r--
L: 1455 - I --j
I: 5mm
~~~J.~ d: 94.59mm 4 Calculated deflection: 0·745mm
I,~ L(mm)---.--~~ 1 : 33602363mm (ends up)
Fig. 4.73. Longitudinal bending of an I-section girder (c )
with a plate welded to top flange.
the weld metal induces bending moments, and longitudinal I-
I : 1Z.5mm::-ir-i------,
8 mm
distortion of the structl1re results. This is also called Bowing or
Cambering and is given by the following equation.
d : 2S.53mm
1 :3642°15
r c: ~
\--L :915--t
Longitudinal Bending = 0.005 ~~L2 mm
where, A = cross-sectional area of the weld, mm2.
...(4.121) L C~~=:J
37.7
mm w:~~5 L Calculated deflection
( ends down)
= -4·585mm

d = distance from e.G. to outermost fibre, mm 11--'25--ll.smm

L = length of the weld, mm (d)
I = moment of inertia of the section, mm4. Fig. 4.74. Four sections considered for calculation oflongitudinal bending.
324 Welding Engineering and Technology Welding Stresses and Distortion 325

Very little distortion is produced by intermittent welding, due Putting the values of Aw, d, L and I, we get,
probably to the fact that longitudinal residual stress does not 2
Teach a high value in a short intermittent weld. AT:::
iJ..I.Jb
0 • 005 x 49 x 7.34 x (915) = 2.93 mm
513213
Problem 4.8. Find the longitudinal bending in the four cases Ans. !:!.Lb = 2.93 mm
shown in Fig. 4.74.
Solution. (c) I-Section With a Plate Welded to Top Flange
Referring to Fig. 4.74 (c), we have,
(a) Angle section Fillet weld leg length, I = 5 mm
Referi'ng to Fig. 4.74 (a), we have, Taking triangular section for the two welds, we have
leg length of the weld, I ::: 5 mm
Assuming triangular weld, Area of cross-section of 2 welds, Aw::: 2 (~x 5 x 5)= 25 mm2
d=+94.59mm
...(i) I::: 33602363 mm4
Area of weld, Aw::: ~ X 5 x 5::: 12.5 mm2 L::: 1455mm
Distance between the neutral axis of the section A ·d·L 2
and the neutral axis of the weld, d::: + 21.34 mm ...(ii) Longitudinal bending,' M..b ::: 0.005 w -
(+ sign indicates that distance d is above the neutral axis
of the section) . Putting the values of Aw, d, L and I, we get,
Minimum moment of inertia of the
!:!. L b = 0 •005 x 25 x 94.59 x (1455)2
33602363
section, Imm::: 177315 mm4 ...(iii)
Length of the section, L :::635 mm ... (iv) = 0.745 mm Ans.!:!.Lb = 0.745 mm
Using equation (4.121), we get, (d) Double Fillet Weld on Inverted T-Section
A ·d ·L2 Referring to Fig. 4.74 (d) and taking weld cross-section as
Longitudinal bending, M..b = 0.005 ~ triangular, we have,
I = 12.5 mm
Putting the values of different parameters from (i)-(iv), we get,
AT = 0 005 12.5 x 21.34 x (635)2 Area of two welds, Aw = 2 (~x 12.5 x 12.5)= 156.25 mm2
Ul..Jb • x 177315 d::: - 25.53
= 3.03 mm Ans. !:!. Lb ::: 3.03 mm (N.B. : -ve sign for d indicates that the neutral axis of the
(b) Strip Welded to Thick Base welds lies below the neutral axis of the section)
Refering to Fig. 4.74 (b), we have, I = 3642025 mm4
Leg length of fillet weld, I = 7 mm L::: 915 mm
Taking weld section as triangular, A ·d·L2
:. Longitudinal bending, M..b = 0.005 ~--
Area of two fillet w~lds, Aw = 2 (~ x 7 x 7J = 49 mm2 t.e. = 0 005 156.25 x 3642025
M..
b'
(- 25.53) X (915)2
d =+ 7.34mm .
1= 513213 mm4 =- 4.585mm
L=915mm (-ve sign indicates that ends of the horizontal plate have been
A ·d ·L2 bent downwards as shown in Fig. 4.74 (d)
.. Longitudinal bending, !:!. Lb = 0.005 x ~ Ans. !:!.Lb ::: - 4.585 mm
 ~26 Welding Engineering and Technology
4.6.5. Rotational Distortion
In rotational distortion, which is observed ahead of the point
of welding, the sheets being butt welded either come closer to each
uther or the distance between them is widened.
Now, relative to a given thickness of material and a given large
temperature difference between a molten weld pool and the
unheated parent material, there tends to be a limit to the rate at
which the heat can be conducted away from the vicinity of a weld.
Thus, if heat input is greater than the limit the isothermal pattern
will take another form. It is to be expected that the difference in
heat flow will cause differences in the transient distortion
patterns.
These differences can be shown to exist by taking sheet or thin
plate of metal and cutting from it four fairly large equal sized
• rectangles to make two pairs of test plates. Now, if one pair is set
up to give an open square butt preparation and it is welded at a
relatively slow rate it will be founct that the joint edges ahead of
the point of welding tend to close in as welding proceeds, as shown
in Fig. 4.75 (a) and may even cross past each other if the
L,
(a I Slo •••.•••.eld ( b I Fast weld
Fig. 4.75. Types of rotational distortion in butt welds when progressively welding
material at widely different heat inputs.
thicknesses get out of line. This necessitates the use of fairly large
pieces of sheet to prevent their saturation with heat. Now, if the
other two pieces are set up in the same way but are welded rapidly,
say with the help of an automatic arc welding head, it will be found
that the plates tend to open out ahead of the point of welding as
Welding Stresses and Distortion 327
shown in Fig. 4.75 (b). Although it is difficult to define slow and
fast welds because the differences are dependent on so many
variables of materials, heat source, etc., but all manual welds are
termed slow welds while the welds made with automatic welding
units are generally termed fast welds.
As stated above the difference in type of distortion encoun-
tered is due to thermal flow patterns in 'slow' and 'fast' welds. A
rough picture of what happens as a result of the combined effect
of some of the metal being heated into its plastic range and heat
flow outwards into the adjacent material is given in Fig. 4.76. In
EXPANDING
(al Slo •••.•••.eld (bl Fast weld
Fig. 4.76 Expanding and contracting zones in arc butt welding for
(a) slow weld, and (b) fast weld.
a 'slow' weld represented in Fig. 4.76 (a) the expansion and
contraction pattern is such that, locally the expanding zones are
wider at Y than at X and are reacting against'the contracting zone
so that the joint faces draw temporarily towards each other ahead
of the point of welding. The plates are thus welded in this position
so that the weld then tends to prevent their return to their original
alignment. In Fig. 4.76 (b), the corresponding effect of a 'fast' weld
is illustrated. In this case X is greater than Y so the joint edges
curl temporarily away from each other and are welded in this
position so tending to be restrained against returning to their
original alignment. It should be borne in mind that this picture of
heat flow is not a precise one; however it does make it possible to
visualise the actual mode of behaviour.
 329
328 Welding Engineering and Technology Welding Stresses and Distortion

In conclusion it can be said that the rotational distortion is (ii) the amount of deformation in buckling distortion is
affected by both heat input and welding speed; as well as by the
location of tack welds, as and when used. The welding sequence
produces complex effects on the rotational distortion and the
distribution of constraint along the weld.
much greater than that in bending.
""'- -.:.- -- -~--
-........

( a I Buckling distortion
~
••..•.... -
4.6.6. Buckling and Twisting
YVhen thin plates are welded, considerable residual stresses
occur in areas away from the weld and cause buckling. co
~
Buckling type of distortion occurs when the specimen length ""

exceeds the critical length for a given thickness, as shown in Fig. ~<II

4.77. Thus, for example, if two panels each 1m wide and 2 m long '0
'0
<II
•...
-;;l
II'l

T Weld Z
c.l
<II

1
200 400 600
Compression- Length, cm
(b I Eight stable distortions observed
I- L I in one specimen
Fig. 4.78. Buckling distortion in thin welded plate, (a) buckled plate, and
E (b) eight stable distortion shapes observed in one specimen.
E.
•...
...•
While the shrinkage
.c. force of the weld is a
II'l
II'l function of the square of VoJe\cl
c
cu
.x the thickness of the parent
...•
:E.
--- material, the resistance to
d'-' buckling is' a function of
:&:
'C the cube of the thickness.
u Thus, buckling due to
0·5 1.0 1.5 2.0 2.5 3·0 welding of a panel
Breadth of plate (Bl. m increases directly as the (0)
(0)
thickness decreases.
Fig. 4.77. Relationship between breadth of plate and critical thickness for
buckling distortion of a butt weld of different aspect i.e. (wlp) rat.ios. Twisting can also be a
are butt welded (.'. B = 2 m, BIL = I), the plate is going to buckle problem with thin
material because of its low
after welding if the thickness is less than 4.6 mm. torsional resistance.
In studying weld distortion in thin plate structures, it is When a weld is made
important to first determine whether the distortion is produced along the centre of a
by bl1ckling or by bending. Buckling distortion differs from member, Fig. 4.79 (a) the
bend;i'g distortion in that, weld area tends to shrink
(i) there is more than one stable deformed shape, as shown and become shorter. The Ie I (d)
in Fig. 4.78, and . effect is the same as if a Fig. 4.79. Schematic representation
of twisting due to welding.
 330 331
Welding Engineering and Technology Welding Stresses and Distortion

turnbuckle is attached at this position, and tightened as shown in For ~, the determining factor on the material side for the level
Fig. 4.79exist.
cannot (b). Under this centrally located tension, a flat rectangle of welding distortion is dependent on Tm . a . a which is modified
in the opposite direction by the modulus of elasticity (E) of the
To satisfy the conditions of a member that has outer edges material and its yield strength (cry) in respect of components
longer than its centerline, the member must twist, as shown in Fig. subjected to residual stresses. The melting temperatures Tin and
4.79 (c). Applying a counter force to untwist the weldment is of no T':r, are taken as relative to a working temperature of O°C.
use. Because when the applied force is great enough to re-establish
th€ original plane, the material snaps into a twist in the opposite
4.6.8. Measurement of Distortion
direction, as shown in Fig. 4.79 (d). Distortion in the post-weld cooled state is determined by
Twisting can be prevented or minimized by one or more of the applying length and angular measuring techniques common in
following methods. practice without any need for specific adaption. Some of the
methods used are shown in Fig. 4.80.
(i)Minimize shrinkage force by good welding practice _
decrease volume of weld metal, and weld at the highest
compatible speed,
(ii) Keep the length of the welded member as short as prac-
tical,
~"""n1~ : /\ ~.'t---~;
Ciii) Incorporate as much resistance to twisting as feasible.
Since the resistance of a plate to twisting is a function
of the cube of its thickness, doubling the thickness in- (,( ;,~:~"""'~::':t'::"'''::''1
creases its resistance by a factor of eight. Torsional resis- (b)~f_\

tance can also be increased, where design permits, by

using closed box sections or diagonal bracing. (f )

4.6.7. Welding Suitability Index Based on Distortion

~~'-~
Similar to welding suitability index based on residual stress
CA,(J.)' equation (4.32), Radaj has also proposed a welding Ig )
Fig. 4.80. Different methods of measuring weld distortion.
suitability index based on distortion (~). It is given by the
following equation Transverse and longitudinal sprinkage are most easily
T* . a* . a* . E . a determined using a measuring tape. Bending or angular shrinkage

a . E *:. cry ...(4.122)

1..£:;: ,"In
can be measured on a measuring plate by means of tensioned
In • a . string (measuring should be performed by supporting the string
where, Tm = melting point of material being welded, in the horizontal plane because of its sag) or bymeans of a straight
edge set against the component 4.80 (a, b, c). The deflection can
a = diffusivity of the material, also be measured continuously in order to determine the bending
a = coefficient of thermal expansion, or angular distortion profile 4.80 (d, e). A more'elaborate method
E = modulus of elasticity, and of measuring angular distortion is shown in Fig. 4.81.
cry = the yield limit. On vertically extended components such as pillars, supports
and tank walls, inclination and deflection are measured by means
while the values ofT;.., a*, a*, E* and cr' refer to those of reference of strings hanging exactly vertically. The tensioning weight of the
material string is immersed into some liquid, as shown in Fig. 4.80 (f>, in
compared. with which the material under consideration is being order to prevent oscillations. Circumferential measurements on
spherical and cylindrical shells are performed by means of a string
332 Welding Engineering and Technology
./),-
,L-J-.L
~'I;)./1hl-hl
.0'=Sin\hl~h2)
V«rnicr
Surface plate height
gouge
Fig. 4.81. Distortion measurement hy height gauge method.
wrapped around the structure, the string being applied tightly to
all sides by means of tensioning rollers which are twisted against
each other by a constant spring force, Fig. 4.80 (g). The change in
circumferential length from which the deflections can be calcu-
lated follows from the twisting angle of the tensioning rollers
applied to the component with weld and compared with the
component without weld.
• 4.6.9. Control of Distortion in Weldments
Since distortions originating during fabrication of welded
structures often lead to unacceptable deviations from the design
dimensions, these deviations must be minimised.
A basic problem in this connection consists in the fact that the
w·eldi.ngresidual stresses and welding distortion, to a large extent,
behave in a contrary way. A component rigidly fixed when welded
displays high residual stresses. If, by contrast, it is welded without
any restraint, the distortion is large and residual stresses are
relatively smaller. It is not, therefore, readily possible to manu-
facture a component which is at one and the same time low in
residual stresses and in distortion, as is often desired in practice.
In case a weldment is required to have definitely no stresses
and no distortion then annealing is the best method but improper
and. uncont.rolled annealing may also result in warping of welds.
However, apart from annealing there are a number of practical
ways of controlling distortion. Some of these are :
Welding Stresses and Distortion 333
1. Least root gap
2. Proper Tack welding
3. Narrow groove section
4. Direction of welding
5. Least metal deposited
6. Welding sequence
7. Back-step welding
8. Block sequence and Cascade sequence
9. Symmetrically simultaneous welding
10. Counter or opposing set up
11. Distortion control in Thin Plates and Sheets
12. Fixing
Brief description of each one of these follows
4.6.9.1. Least Root Gap
The gap between the plates to be welded should be as small
as possible but sufficient for good penetration. Excessive gaps
should be avoided. The included angle should not exceed 60°. For
heavy sections the double vee preparation should be preferred.
4.6.9.2. Tack Welding
The tacking of aligned parts, before seam welding, has the
purpose of achieving particularly accurate manufacture in terms
of shape and dimensions. The tack welds must be sufficiently long
in order to reliably transmit the shrinkage forces which will
otherwise occur. Atack weld length of two to three times the platp.
thickness is recommended. As generally small volume tack welds
are applied to relatively large volume components, the cooling
rates which occur with tacking are high and, as a consequence,
hardening and cracking are frequently the accompanying
phenomepa. These undesired phenomena are countered by
preheating when tacking. There is also a possibility of the
formation of cracks and other defects due to welding over tack
welds. The counter meam; used in multi-layer welding, such as
preheating, slag removal, and removal of other defects are also
employed in the case of tack welds.
4.6.9.3. Narrow Groove Section
The groove cross-section should be as narrow as practically
feasible to permit the smallest possible heat input and filler
material in conjunction with thE)highest possible heat concen-
tration for the welding method used. As a result, the fusion zone
and the plastically deformed zones remain small, and that reduces
distortion in weldment. The selected groove shape has a
i~
I

334 Welding Stresses and Distortion 335

Welding Engineering and Technology

considerable influence on the groove cross-section. The square In welding 1- or H-beam the first to be welded are the joints
groove permits the smallest possible groove cross-section. The U within each web plate and flange, and then the butt joints between
shape is more favourable than the Vee shape. Symmetrical weld the web plates and flanges of a beam. If an 1- or H-beam has
grooves reduce angular shrinkage though residual stresses are stiffeners they should be welded in place before the webs and
increased. For this reason the butt weld double Vee groove may flanges are assembled, as shown in Fig. 4.83.
be used instead of the butt weld single Vee groove (or Y groove).' 3 62
10 9 87"12
514
13
By the similar reasoning, the double-sided fillet weld may be 4 1
selected instead of the single-sided fillet weld. The reduction in
angular distortion is fully achieved, however, only if the welds are
laid simultaneously on both sides or only if the alternative method Fig. 4.83. Welding sequence for an I or H beam with stiffeners.
of multi-layer welding is employed, if possible. For welding a cylindrical vessel consisting of several courses
4.6.9.4. Direction of Welding or tiers, the longitudinal seams should be welded first, followed
The direction of welding should be away from the point of by the circumferential seams, as shown in Fig. 4.84. If welded in
restraint and towards the point of maximum freedom. reverse order, the longitudinal welds will be restrained.
4.6.9.5. Least Weld Metal Deposited
It should be bome in mind that the amount of contraction
increases with the amount of metal deposited. Therefore, no excess
metal should be deposited. All welds should be made to drawings
without increasing the weld cross-section 'just in case'.
4.6.9.6. Welding Sequence
One of the principal methods of controlling distortion in
welded fabrications is by adopting the correct welding sequence. Fig. 4.84. Welding sequence for a reservoir.
By varying welding sequence one can, within broad limits, control
both the sign and magnitude of the eventual deformations. In welding frames of different lengths and thicknesses from
formed mild steel channels and plates the sequence shown in Fig.
In welding job of a large surface area consisting of several 4.85 (a) should be followed. The least amount of distortion would
plates (stich as vessel bottoms, platforms, decks, etc.), the result if welds 1 and 2 are made simultaneously, followed by welds
transverse seams should be welded first followed by the
longitudinal seams, as shown in Fig. 4.82.
~ y
No·1 : ./
No·3No.4 2(c)
PoorBad
N4o.1(bl No.1
v No.
.\.)3"'~--I2 3
, No·4
."
(a) Good
Mild stee~ I ·7~n.<chOl'nels r

1 No2
-r--tH

Fig. 4.85. Welding sequences for welding plates to box section;

(a) a good sequence, (b) a sequence that may cause Bornebending, and
(e) a totally bad sequence.
Fig. 4.82. Welding sequence for a plate decking.
 r
336 Welding Engineering and Technology Welding Stresses and Distortion 337

3 and 4, also made simultaneously. But since thi!' procedure would Tack weld Sl!:quence
5 6 J 7 8 9
not be practical in most cases, the next best procedure is to make
welds 1, 4, 2, 3 in that order, resulting in some camber. t '25 t I
If two welds are made at the same time in a horizontal position
i.e. weld No.1 and No.4 in Fig 4.85 (b) some bending would result (al Tack weld sequence
because the yield point would probably be exceeded. The
remaining two welds would probably not quite pull the member
straight.
Alternatively, if welds across the longer dimension were
chosen as pairs, Fig. 4.85 (c), bending would be about the Y-Y aXIs.
Such a member would have little structural utility. (c) Cove r pass di rection
In the fabrication of larger structures once the assembly
sequence has been chosen, the order of laying the welds cannot 2 J 4 5 6
alter the magnitude of the deformation as the overall effect of a
number of welds is the same as the sum of the separate effects
due to each weld. As a matter of fact, in some instances the earlier
laying of a particular seam may affect the magnitude of the final (b) Back-step welding sequence.
deformation. These cases, however, are .~are and as a first Fig. 4.86. Distortion control by back-step welding sequence;
i:i
approximation one can assume that the effect of each weld is (a) sequence for laying tack welds, (6) scheme for back-step welding sequence,
independent of other welds unless the welds are placed close to and (e) laying of cosmetic run or cover pass.
each other.
~i
IIIii:! weld at one end of the seam to the opposite end (Fig. 4.86 (b».
III It is, however, not possible to give exact sequences for all Finally, the remaining layers are applied continuously, alternating
I;!
I'I' possible cases of fabrication. In each particular case the welder the direction.
should learn the instructions, for a given weldment including the Also, by their lengths, welds are classed into short (up to 300
sequence of welds. In many cases he may follow a sequence of his mm in length), medium length (300 to 1000 mm long), and long.
own choice provided he knows the way residual stresses work in (over 1000 mm) in length. Short welds are deposited from thEj
his job. beginning to the end in one direction. Medium length welds are
4.6.9.7. Back-Step Welding Sequence deposited in increments (1 through 6) from the centre to the ends
Back-step welding is a measure to counteract the, wedge- of the weld or by using the back-step method as shown in Fig.
shaped opening and closing (i.e. rotational distortion) of the weld 4.87. The length of increment is so selected that each one of them
groove ahead of the weld pool, which may especially occur with
long welds and low welding speeds. If the groove opening is
suppressed by prior tack welding or by rigid lateral restraint, this
may result in a permanent roof-shaped arching of the weld joint.
The back-step sequence that has proved satisfactory in practice is
\-1-1 ~
( b) Centre to ends by backstep
(a) Centre to ends by normal
shown in Fig. 4.86. First of all, short tack welds of length method method
l = 2.5t are executed, spaced S ~ 25t apart and by following the
Fig. 4.87. Welding of medium length welds;
sequence oftacking shown in Fig. 4.86 (0); where t is the thickness (a) by increment from centre to end, and (6) by back-step method.
of plate being welded. The spacing between tack welds should be
further reduced if the welding speed is still slower, so as to reliably may be welded with the whole number of electrodes. For
prevent groove edge displacement. The first weld layer is depositing long welds use is also made of the back-step method
subsequently deposited in back-step sequence starting from a tack which promotes good melting through of the initial sections of
welds and diminishing of the workpiece warpage.
r
!

3~8 Welding Engineering and Technology 339

Welding Stresses and Distortion

Back-step welding sequence reduces transverse and longi-
tudinal shrinkage of the weld joint as a whole and is a widely used
method in the fabrication of large structures such as ships, storage
tanks, etc.
4.6.9.8. Block Sequence and Cascade Sequence
Welding
To deposit long welds of great thickness the block or cascade
sequence is used. In welding by Block Sequence Fig. 4.88, the first
and bottom of large I-section girders to execute the chord joints
symm etrically-simul taneously.
The symmetrically opposite longitudinal welds of slender
girders are welded simultaneously to reduce distortion. The
of longitudinal curvature caused by a longitudinal weld on one side
of the girder is only partially compensated for by a weld seam
which is later on welded on the opposite side (as already discussed
in section 4.6.9.6 on welding sequence). To eliminate angular
distortion in double Vee butt joints, layers arranged symmetrically

~~~;~I
Fig. 4.88. Block sequence of welding thick sections.
to the centre plane of the plate should be welded simultaneously.
4.6.9.10. Counter or Opposing Set-up
The warpage caused by intemal stresses may be effectively
controlled by a counter or opposing effect ordistortion. In aT-joint,
layer of the weld is deposited in the centre over a s~ction 200-300 for example, warpage usually occurs due to the contraction of the
mm long. Next, stepping aside 200-300 mm from its beginning this weld as shown in Fig. 4.90 (a).
length is sealed to the beginning of the first layer, the latter is
then overlapped by another layer and welding is completed at a
distance of 200-300 mm from the end of the first layer. All
subsequent layers are deposited in the same order until one of
them attains the effective throat thickness. Shorter lengths are
~=~----~~ (0)

then deposited on sections which have not yet attained the

effective throat thickness.
In welding by Cascade Sequence a length 200-300 long of the
C:::__
I --
~ •
~
c:=:=J
(b)
first layer is deposited at the planned end of the weld, as shown
in Fig. 4.89. Whereupon welding is performed in much the same
way as in welding by block sequence.
1 2 3 I.... Oi~n ~dlng
I( )
I Jl
\ ~'::::"'~ Fig. 4.89. Cascade sequence of welding.
I Fig. 4.90. Warpage in a T-beam and the suggested
method of eliminating it by counter set-up.

By giving the vertical member a slight reverse bend, or counter

4.6.9.9. Symmetrically-Simultaneous Welding
Welding symmetrically-simultaneous reduces distortion.
Welds arranged symmetrically in the component reduce distortion
only if they are laid for the most part simultaneously. Conse-
quently, the circumferential welds of large vessels or tubes are
executed symmetrically-simultaneously by several welders; the
. first layer is, of course, laid by back-step welding sequence.
Welding of large cylindrical components if not performed
simultaneously, would cause the first layer to tear and result in
unacceptable mis-alignment. Two welders are assigned to the top
camber, prior to welding, Fig. 4.90 (b), the weldment will
straighten itself after welding as shown in Fig. 4.90 (c).
In butt welding two plates with a single Vee edge preparation,
it is a good plan to place a backing strip under the joint so that
an increased opening is formed, as shown in Fig. 4.91 (a). After
t7~,5;;ff/////J;/
(0 )
~~~~
(b)
Fig. 4.91. Method of controlling angular distortion in butt welding of two plates.
340
Welding Engineering and Technology
W eldi~g Stresses and Distortion 341
welding the plates will align themselves as shown in Fig. 4.90 (b).
Some other examples of pre shaping to control weld distortion are 4.6.9.11. Distortion Control in Thin Plates and Sheets
shown in Fig. 4.92.
When light gauges are welded, it is advisable to place copper
~lfr-{-..Gll
•...••- __ -J \ chill blocks under the joints. As copper has a high thermal
conductivity, it will abstract much heat from the weld, thus
\ t bl / reducing the heating and warpage or buckling of the plates.
(opper c lamps ~ ,- Joint to be wel ded
/
lal

C':I:"""j
-.-.-+
-t-.-.--.- .. ['m"U'~l
-t---.-.-
Fig. 4.92. Distortion control by preshaping the components; prewelding shap~s'
shown by solid lines, postweld shapes in broken lines, (a) roof shaping of chord,
(b) plane end sections of cylindrical shell, (c) outward building of spherical shell,
with block flange, (d) outward building of pipe ends for circumferential welds,
and (e) inward drawing of pipe at plane end.
The distortion reducing measures also include wedge shaped
aligning of the gap of the weld grooves, as shown in Fig. 4.93, and
Wedge
Fig. 4.93. Use of wedge to avoid closing-in of groove
edges in welding of cylindrical shell.
control of this gap by progressively withdrawing wedge at the end
of the groove. It ensures that the welding groove is held open. This
procedure can, however, be applied only to those welds which are
not too long but with transversely moving groove edges. It is a
wide spread practice in gas welding of copper and aluminium. In
such cases it replaces back-step welding which is not practical in
welding these metals because of the high heat losses.
(opper water tubes
Fig. 4.94. A water-cooled jig for rapid removal of
heat to control distortion in welding sheet metal.
Also, a water-cooled jig shown in Fig. 4.94 is useful to carry
away heat from the welded components. Copper tubes are brazed
or soldered to copper holding clamps and water is circulated
through the tubes during welding. The restraint ofthe clamps also
helps to minimise distortion.
4.6.9.12. Fixing
Fixing the parts, to be joined by welding, in a frame as rigidly
as possible is an effective counter measure against bending and
angular distortion. In many cases, the load from the weight of the
connected parts is sufficient without any special fixing elements.
However, it is possible only to reduce relatively slightly the
back-spring shrinkage when the component is released. The
arching of thin sheet metal plates at the weld pool is minimised
by fixing measures close to the groove. Fixing is largely an
unusable practice in respect of transverse and longitudinal
shrinkage. The very large forces which occur in these cases rapidly
exceed the frictional locking limits of the elements. The transverse
inward displacement of the groove edges is minimised better by
groove gap inserts rather than by fixing.
4.6.10. CORRECTION OF DISTORTED WELDMENTS
If a weldment warps despite the precautions taken, there are
ways and means of correcting the defect using one of the following
two methods.
1. Mechanical methods,
r
342
Welding Engineering
2. Thermal methods.
4.6.10.1. Mechanical Methods
These methods involve presses, jack screws, straightening
rolls, sledges, and special fixtures to bring back the q.istorted
component to the required shape.
4.6.10.2. Thermal Methods
These involve the local heating of a warped spot. Any source
of heat may be employed, e.g.,
(a) An oxy-acetylene torch,
(b) A carbon arc,
(c) Powerful oil or gas burners, etc.
The technique is based on the fact that metal distorts when
heated non-uniformly and contracts on cooling.
Thermal methods apply to metals which do not change their
properties when heated to the corrective temperature.
Heat should always be applied on the convex side of the
warped work. Repeated application may be required if the plate
is badly distorted, working gradually from the middle outwards.
The heating zone in a single pass may be one or two plate
thicknesses. Fig. 4.95 shows distortion correction on the
mushroomed flanges of H-beams.
b=20-30 i!i? b=30-497
~VI ~ --H-~ 1,2,3 -areas
T deep penetration, automatic submerged arc welding, the centre of
Fig. 4.95. Distortion correction of mushroomed flanges of welded H-beam.
With a web thickness of not more than 16 mm and flange
thickness of not more than 30 mm, as in Fig. 4.95 (a), heat is
applied to a single area 20 to 30 mm wide along the axis of the
beapl, raising the temperature to 700-800°C.
When the web thickness is greater than 16 mm, heat is applied
to two areas each 30 to 40 mm wide, running over the web seams,
as shown in Fig. 4.95 (b). Sometime reheating may be required in
this corrective treatment.
and Technology 343
Welding Stresses and Distortion
4.6.11. CASE STUDIES ON WELD DISTORTIONS
Some of the distortion problems commonly encountered in
welded fabrications are discussed in this section.
4.6.11.1. Welding of aT-Section
A manually welded T-assembly, shown in Fig. 4.96 (a) was
distorted later all)':.after welding even though proper specified size
of fillet weld was used. Analysis showed that the centre of'gravity
of the two welds was well above the neutral axis. By changing to
eGot ~eld'J[:M;:'nt Q'~ ~
Neutral'
axis Na t~ure 0 f
distortion
I a I Manual weld
CGof.welds7roment arm ~
'--' t ,.... -
Neutral
axis
I b)
Fig. 4.96. Distortion control in T-assembly by changin,g from manual to
of heating
automatic submerged arc welding process to make deeper welds.
gravity of the welds was lowered substantially, as shown in Fig.
4.96 (b), thus reducing the shrinkage moment. Depth, or throat,
of the weld is the same in the two cases but the weld metal with
deeper welds in the second case is nearer to the neutral axis. In
addition, the higher speed of the automatic welding also reduces
distortion due to the reduced heat input per unit length.
4.6.11.2. Three-Member Welded Column
The welds in the lift truck column shown in Fig. 4.97 (a) are
balanced and can be made in downhand welding position by
merely turning the assembly once. But it leads to excessive
longitudinal distortion of 10.7 mm in 2540 mm length. Analysis
shows that the distance between the centre of gravity of the welds-
and the neutral axis of the section is 17.3 mm. If this distance
could be reduced by a change in design, less distortion would occur.
One way to bring the welds close to the neutral axis is shown
in Fig. 4.97 (b). With this design the distance is reduced from 17.3
 345
344 Welding Engineering and Technology Welding Stresses and Distortion
All dimensions in mm An alternative better procedure might be to weld the cover
I---
I 230 --1 ~~ 45 plate to the bottom flange plate first. If distortion occurs,
straightening of sub-assembly will not be difficult. Now, if this
Neutral sub-assembly and the top flange are welded to the web it will need
1-
63·S ax,s7
Neut.ral U175
'Ji
L
.axi2. -J~'~
only four fillet welds, the centre of gravity of those four welds is
much closer to the neutral axis of the section and would lie above
1- -- --

T',
~..L CG of welds
(0) Poor
Fig. 4.97. Two designs for the lift truck column: (a) excessive distance
between neutral axis of the section and the e.G. of weld,
and (b) reduced distance to reduce distortion.
mm to 14.1 mm for the 2540 mm length. If even this does not lead
to the desired reduction in distortion the column members could
be pre-bent about 8 mm in the opposite direction so that the
assembly would be very nearly flat after welding.
4.6.11.3. Girder with Welded Cover Plate on Lower
-r it, as shown in Fig. 4.98 (b). The deflection due to welding will be
directly proportional to the amount of weld metal and the distance
CG of-welds 'd' between the e.g. of the welds and the neutral axis of the section,
(b) Good and inversely proportional to the bending resistance i.e. moment
of inertia of the section.
By changing the welding sequence from the first method to
the second reduces the number of welds that influence distortion
from 6 to 4. In other words, the welds then have only ~ i.e. two-
thirds the effect on distortion compared to the first method. The
distance 'd' is reduced by about half. Bending resistance, of course,

!
Flange remains constant. Theoretically, the second sequence would thus
The girder in Fig. 4.98 (a) is made of High Strength steel and
needs more flange area on the bottom because the concrete deck produce only (~x.~ = lone-third the distortion, and since the
e.G. of welds in the sec6nd case would be above the neutral axis,
on top is attached through shear lugs and will carry some of the the distortion would be in the opposite direction to that produced
bending moment. by the first sequence i.e. the ends will bend upwards.
4.6.11.4.Buckling of Welded Hopper Boxes
Neutral -
- CG of 4 welds Fig. 4.99 illustrates two hopper boxes with welded corners.
Neutra:"-----'l:f
-CGof 6 welds axis/""
One is made from 15 mm thick plate and the other from 3 mm
aXIs T thick sheet. The welds at the corner shrink and tend to leave
excess metal under compression in the central portion of the
lalPoor IblGood
Fig. 4.98. Distortion control in welded I-section girder wit.h welded cover plate to
lower flange; (a) poor design, and (b) good design.

The strength and allowable stress for such a steel decreases

wi.th increased thickness. Two 19 mm thick plates are used on the
bottom because they have a higher allowable stress and are
stronger than a single 45 mm thick plate, thus saving 18% of the
steel.
If the girder is welded in the conventional way in the shop it
will need four fillet welds in the bottom portion of the girder and
the two in the top leading to an unbalanced condition. In such a t = 15 mm t= 3 mm
fabrication, the centre of gravity of the six welds is considerably
be10w the neutral axis ofthe girder section, and the welding would Fig. 4.99. Hopper boxes, (a) made of 15 mm thick plate·-no distortion, and
(b) made of 3 mm thick sheet-badly distorted.
cause the ends of the girder to move downwards, producing camber.
 \ 347
346 Welding Engineering and Technology Welding Stresses and Distortion
I

panels. Distortion by buckling can result. The size of the corner ~l with the beam turned upside down. By supporting the beam near
its ends during welding, as shown in Fig. 4.100 (a) the sag
welds in the hopper made from 15 mm thick plate indicates that
the shrinkage forces would be much greater than those in the supplements weld shrinkage and increases the camber.
hopper fabricated from 3 mm thick sheet. However, resistance to In case too much bending results from welding by the above
buckling is also much greater in the thicker material because as procedure, the beam can be supported near the centre during
stated earlier that while the shrinkage force of the weld is a welding, as shown in Fig. 4.100 (b). The effect of gravity will then
function of the square of the thickness of the material, the be subtracted from the bending resulting as a consequence of
resistance to buckling is a function of the cube of the thickness. welding, thus reducing the camber of the finished beam.
With the thicknesses used in this case, the plate material has 4.6.11.6.Welding of Large Plate Panels
Large plate panels are fabricated for use as tank bottoms, ship
J = 125 times more resistance to bending or buckling than
( ~~3 the decks or ship bulkheads. The sheet metal plates including the edge
sheet material. . preparation are cut with a high degree or'dimensional accuracy.
4.6.11.5. Control of Camber in Welded Beams In rectangular bottoms or walls the plates are arranged so that
Weld shrinkage can also be used beneficially in fabricating a the longitudinal welds run continuously parallel to the long edges
long beam or girder to produce the desired camber. The effects of whereas the transverse welds are arranged in a staggered pattern
weld shrinkage and gravity can be combined to control the amount (see Fig. 4.82) in order to avoid the crossing of the welds with the
of camber. The technique involves support of the assembly so that
the two effects work in the desired direction, either together to
I
produce a straight member or one with minimum camber. Thus,
supporting the beam near its ends cause!?it to sag at the centre,
and supporting it at the centre and letting the ends overhang
produces a relatively straight beam.
Considering the case of a beam which is to have a cover plate
welded to its lower flange and it is required to have a certain
amount of camber with ends down. It is estimated that the
shrinkage from welding will not provide enough camber. Since the
cover plate is narrower than the flange, it will have to be welded

r;;~==~.~.J( (a) Maximum camber

e
__ITD_~_
Fig. 4.101. Joint design and welding sequence for plate panel of a tank bottom.
particularly unfavourable superimposition of tensile residual
stresses. The shape and arrangement of plates as well as the
welding sequence for a tank bottom are shown in Fig. 4.101. It
JL shows that the transverse seams are welded first. The outsite
plates are provided with radial weld grooves, which are initially
( b) Minimum camber
half-filled and are not completely filled until the shell-to-bottom
Fig. 4.100. Two methods of controlling camber in welded beams; fillet welds have been executed. They are thicker than the interior
(a) end supports to produce maximum camber, and
(b) centre support to produce minimum-camber.

348 Welding Engineering and Technology
plates in order to reduce warpage and to reinforce the lower edge
of the shell. The backing plates used with butt welds also serve
as rei.nforcement besides providing a backing to the weld pool.
The longitudinal welds of slightly tapered bottoms are
exclusively radial. They tend to form circumferential corrugations
as a consequence of the radial residual stresses. This is countered
with flat backing bars placed underneath. The use of high
performance welding is recommended to keep the transverse and
angular shrinkage to a relatively low level.
[11 fabricating an octagonal panel, with the welding sequence
shown in Fig. 4.102, each radial weld is completed before the next
layer 1: Root pass
Structural steel 37-2
la)~~~rL~~~~
'layer2
(b)~~'W~
layer 3: (over pass
! Welding Stresses and Distortion 349
~ij
! interior welds may produce corrugations on the edge side. It is a
widespread practice to brace thin plates on a slightly curved
surface for welding. Peening after welding is advisable. Also,
welding on a root-forming copper backing plate has proved
satisfactory in practice.
4.6.11.7. Patch Welding
In patch welding, dimensionally precise cutting to shape and
favourable welding sequence play important roles. Patches are
inserted or placed overlapping in order to close cutouts which have
arisen as assembly openings, as a consequence of design
modifications. or by removing defective material. Patches placed
overlapping, which present no problem in respect of distortion and
residual stresses, have the drawback of a low fatigue strength, are
prone to crevice corrosion and interrupt the smooth wall surface.
Patches inserted flush, by contrast, require a well planned welding.
sequence if residual stresses and distortion are to remain under
.control. The basic principle is to subdivide the circumference of
+----------+ ~----------+

I
I
I
:
II 11 12
~---------+ +---------+I
(a)

+---------+
6
5
I
I
I I
I:
:

I
I
I
:

lIb)
-r---------t
12
11
'D'
10
D 65,
9-8 4
2
3
I

I
I
III

I 6 5 4 3 I I 20~ I

: 702'
II
I
'6
:\c)
8
5

+- - - - - - - - 7 8 4 3
2
1

+- - - - -
+---------t
I
I

I
I
-t-
II
I
II
lId)
22

23
9:

- - - - -t
10
II

I
I
I

Fig. 4.102. Distortion control in an octagonal plate panel, (a) layer sequence, : I
(b) joint design, and (c) welding sequence.

joint :is tacked and then welded in back-step sequence. Unres-

trained transverse shrinkage is thus possible. Where warpage
may occur as a result of angular and longitudinal shrinkage, this
can be avoided by peening the weld before the next plate is joined
:
\"1 itC~+~ i
~~2
Fig. 4.103. Welding sequence for inserted patch, (a, bl circular and square
I

by welding. Thin plates (t ::;;3 mm) tend to warp more easily than patches, (c) circular patch with two welders simultaneously acting, (dl circular
thick plates. Edge welds may produce dents in the interior; patch with multilayer weld. and (cl rectangular edge patch.
350 Welding Engineering and Technology
the patches into two halves, which are finish-welded in turn, each
in back-step sequence. This assures relatively unrestrained
transverse shrinkage of the 'first circumferential half without
interfering with the twisting of the patch. The tack points on the
opposite side would indicate the transverse shrinkage by tearing.
Suitabl.e welding sequences for circular and rectangular patches
are pr~sented in Fig. 4.103.
Large patches are best welded simultaneously by two welders
as indicated in Fig. 4.103 (c). Thicker patches are welded in the
first layer with a small back-step, in the subsequent layers with
a longer back-step in the alternating direction, as indicated in Fig.
4.103 (d). For double vee groove welds, one side of the circum-
ferential half is completed first. Rectangular patches are well
rounded at corners to enable continuous welding around. The
slight bulging of the patch, which is occasionally employed, does
not offer the aimed-at gain in compliance (i.e. the capacity to yield
under the applied force). Stretching the weld zone by peening is a
useful means of reducing residual stresses in a ductile material.
Local annealing, by contrast, does not offer any major
improvement.
FIVE
Preheat and Postweld
Heat Treatment
5.0. INTRODUCTION
Successful welding of materials of high thermal conductivity
and/or thick sections in any material or ~he welding of harden able
steels requires a controlled rate of cooling because when heated
to a high temperature during welding and cooled rapidly
thereafter, they harden. Welding under such conditions without
due control of cooling rate may produce embrittlement in HAZ
(heat affected zone) parallel to the weld joint. With proper
preheating the rate of cooling is reduced and consequently the
metal in and around the weld bead does not harden. The preheat
temperature for arc welding is the temperature at which the
workpiece must be maintained and below which it must not fall
until the welding is complete.
If steel is welded without preheat than the total drop in
t~mperature will be from about 1540°0 to about 30°0 (room
temperature) i.e. about 1500°0. In case it is preheated, to say
300°0, the drop will be reduced to about 1200°0. This results in
reduced cooling rates particularly at the important intermediate
temperatures of 800 to 500°0. In multi-run welds the succeeding
bead may be deposited on metal that has been preheated (by the
preceding beads). The more rapidly the beads are deposited on
each other, the higher, obviously, is the preheat or interpass
temperature.
If welding is to be carried out at low ambient temperatures
(say at ~ -40°0) the cooling rates will be unusually high,
particularly in the dangerous ranges below 800°0. Now, it is well
known that rapid cooling in the range of 315 to 200°0 may cause
cracks in welds in sqme high strength steels. W-aldingthese steels
352 Welding Engineering and Technology Preheat and Postweld Heat Treatment 353

at unusually low temperatures is risky but this risk can be avoided Although preheating is quite useful in improving the quality
by preheating the steel to above the dangerous range. of a welded joint but it leads to bigger HAZ as shown in Fig. 5.1
PWHT (post weld heat treatment) is intended primarily as a
stress-relief treatment. For welding most of the high carbon as
0 0eg" l;7l
0IIICa.
•...
Cl/
.:0::
s::.
III
10
10
we11as the high alloy steels post heating is as important as, if not
more, preheating . .A.Jthough.preheating does control the cooling'
rate but there is always a chance for the residual stresses to
develop and approach the darigerous le'velaffecting the service life
of a component. If due attention is not paid to these aspects of
preheating and PWHT, it may lead to a catastrophic failure of a
welded component or structure.
o 10
5.1. AIMS OF PREHEATING DistoncQ: ,mm
Apart from what is stated above, the main aims of preheating 10.) Without preheating
are:
1. to reduce the heat losses from the weld area, that in turn
ol;7l

~ 800
reduces the cooling rate of the weld, III
III
2. to reduce cracking by preventing th3 formation of hard Cl/
C
surfaces (due to the formation of martensite in the case '0
of steels), e
•...

.s::. 400
3. to reduce the expansion and contraction rates thus a.
8c
reducing distortion and residual stresses, :0::

4. to burn grease, oil, and scale from the joint area leading
to faster welding speeds, . DistoncQ: ,mm
5. preheating also keeps the weld beads more fluid with
flatter surfaces thus avoids stress concentration due to
notch effect,
6. preheating also brings some steels above the tempera-
ture where brittle fracture might occur during welding,
7. to help allow sufficient time for hydrogen to diffuse out
of the weld and HAZ. This may also modify the
microstructure of the weld, making it less susceptib~e to
H2-embrittlement.
Both the intensity of internal stresses and the cooling rate of
the weld depend upon the geometrical and technological factors
like welding process, current, joint restraint and thickness of the
component. Consequently it is very difficult, if not impossible, to
define accurately the preheating temperature in advance. Inspite
of these limitations; the usual range of preheating temperatures.
lies between 120 and 300°C. However, if extra-low rates of cooling
are required the preheat temperature may have to be raised up
to as high as 600°C.
o 10
I b I With preheating to 250·C
Fig. 5.1. Effect of welding on weldment hardness and the size of weld bead on
HAZ of 1040 steel, (a) without preheating, and (b) with preheating to 250°C.
as well as results in higher welding costs and considerable
inconvenience to the operator during welding.
Moreover penetration is increased with an increased pickup
of undesirable elements from the parent metal into the fusion
zone. Basic electrodes should therefore be used for welding a steel
requiring preheating as they are relatively insensitive to the
effects of parent material admixture.
Slower cooling rates, as a consequence of preheating, result in
reduced tensile strength and increased ductility. Excessive
preheating can have detrimental effect on the impact toughness
properties of a welded component, as is shown in Fig. 5.2. It is
therefore imperative to know accurately the preheat temperature
required for a given material. This may be done with the help of
Clip Test for alloy steels.
 354 Welding Engineering and Technology Preheat and Postweld Heat Treatment 355

Allow the weld to cool down to room temperature and then

break the clip off by hammering, as shown in Fig. 5.4. If the lug
....•
lJ"
bends and, after a number of blows, fails through the weld [Fig
""
~
'"
'"
QI
C
.&:.
0'\
:::J
~
o
-140 -100 - 60 -20 0 20 40 60 (0) (b) (c)

Impact t~st temperature I"C Fig. 5.4. Clip test, (a) method of testing, (b) satisfactory fracture,
and (c) unsuitable conditions.
Fig. 5.2. Effect of impact test temperature on impact toughness of T-1 steel.
5.4 (b»), the test indicates that no serious underbead cracking will
5.2. CLIP l.'EST be caused if the same procedure is followed for actual fabrication
Usually the preheat temperature required for satisfactory at normal room'temperature. If the lug breaks and pulls out some
welding of a given alloy steel is supplied by its manufacturer. of the parent metal, as shown in Fig 5.4 (c) the test indicates that
However, the chemical composition of the low alloy steel may not this particular steel must be preheated.
always be known or it may be required to establish whether The test is therefore repeated by preheating the base metal
preheating is required or not. Clip test is a useful method to make plate to some temperature Tl' If the steel fails the clip test again;
a rapid check in such a case. This test is, however, not applicable
to thin steels but produces good results on heavy sections, 10 mm this is an indication that the preheat temperature must be raised
or more in thickness. further. Failure of clip test is because of the hardened condition
of the HAZ. It may be necessary to repeat the test several times
The test involves the welding of a clip or a lug (a piece of 10\"1 using a higher preheat temperature, up to a maximum of 300°C.
carbon steel) that is 13 mm in thickness and 50 or 75 mm square· If underbead cracking still occurs, then it is imperative to
to the steel plate which is being checked for determining the reconsider the choice of electrode and the welding procedure being
preheat temperature. A finet weld with convex contour, as shown followed.
in Fig. 5.3, is made with the size and type of electrode, the welding The.,cooling rates accompanying a particular preheat
current, and speed that will be used for the actual welding of the temperature bear a constant ratio to those occurring in a plate
unknown steel component. initially '$t room temperature. If the cooling rate for a plate at
room temperature, of Say 24°C, is assigned a value of 1.00, the
Clip or lug 50or7Smm ratios for the other plate/tilmperatures
,/
would be as follows.
Square Plate Temperature roC) / ~<. Ratio of Cooling Rate
Low carbon o / " 1.10
steel
I 24 1.00
Convex fillet weld
93 .'
0.85
205 .; 0.67
Alloy steel
5.3. ' METHODS OF PREHEATING
under test Several methods are possible for localised preheating ; more
popular amongst them include the following:
1. Flame heating,
Fig. 5.3. Test plate dimensions for clip test to determine
2. IndJJ~tioJl";brresistance heating,
preheat temperat4,re of low alloy steel. 3. Infrared-heating, and
 35. Welding Engineering and Technology Preheat and Postweld Heat Treatment 357

4. Quartz lamp heating. flowrates are unusually high, use large diameter hoses, fittings
Out of these four methods flame heating and resistance and glands with 6.25 mm minimum ID (inside diameter); Fig 5.6
heating are most popular and are described in brief in this section. shows the use of multi-head preheating set-up.
5.3.1. Flame Heating
Flame heating is done with the help of oxy-fuel gas heating Pressure Vessel
torehes; one such torch used for manual mode of preheating is she"
shown in Fig. 5.5 (a). Heating torch tips are considered precision

~
la I Manual Heating Torch

Mixing chamber
~
'-.
@
- 02
- C2H2

Fig. 5.6. Preheating a part of pressure vessel shell by oxy-acetylene burners.

Water-cooled flame heating heads, one such is shown in Fig.
5.7, are normally used with special equipment. These heae~ may
be 300 mm or more wide.
I b IMultiflame Heating Heads
Fig. 5.5. Flame heating torch and torch heads, (a) manual heating torch, and
(b) some typical multiflame heating heads used as torch tips.

tools. Some torches have a range of tip ends to cover various flame
sizes; a few such torch tips called multiflame heating heads are
shown in Fig. 5.5 (b). Smaller tips are used for light heating while
larger tips are employed for heavy heating. Multiflame heating
heads distribute heat efficiently over a large area. They are
usually attached on adaptor or an extension from the mixing
chamber [Fig 5.5 (a)].
Various types of flames (carburising, neutral, and oxidising)
can be obtained by adjusting the oxygen-to-fuel ratio but for
I almost all applications the neutral flame is used.
Amount of heat input per unit time should be controlled by
using larger or smaller tips rather than by increasing or
decreasing the gas flow rate alone. Torches and multi-flame heads
are available to operate with gas flow rate up to 330 HtJmin.Where
Water (aoled Head
Fig. 5.7. Water-cooled heating head.
5.3.2. Electrical Resistance Heating
It is generally recognised that the electrical resistance heating
methods offer greater flexibility in meeting the requirements of
temperature distribution and uniformity over a wide range of
preheating applications. In addition, electrical resistance systems
are portable, and attractive in terms of capital cost. These units
when used in their automatic mode reduce the labour costs
considerably.
 358 Welding Engineering and Technology Preheat and Postweld Heat Treatment 359
5.3.2.1. Power Sources
60 AMP welding 60 AMP welding
Specialized power sources with solid-state switching is now plug & Socket plug & Socket
for bottom half
the norm in resistance heating units. When these power sources of heater
for top half of
heater
are used in conjunction with variable voltage control they prove
to be ideal for unusual shapes and sizes. Certain industries such
as aero-engineering and nuclear steam generation have greatly
benefitted from the development of variable voltage facility with
their considerably reduced operating and maintenance costs. 25mm ceramic
These power sources range between 8 KVAtransformers for pipe fiber
butt welds and 132 KVA heat treatment units for on-site
simultaneous multi preheat operations over a small site area with
both fixed and variable voltage.
, mm stainless
5.3.2.2. Temperature Control steel sheet
Open flame gas techniques are still widely practiced although
it is known that these are the least satisfactory from the
temperature control point of view. For improved control of preheat
temperature set-point temperature controllers are used. There are , mm perforated
different types of th.3se co'ntrollers viz., one called on-off type stainless steel sheet
monitors the temperature from the thermocouple and compares,
it to the set temperauture-turning the heater 'on' if it is below Fig. 5.8. Wraparound resistance heating unit for preheating pipes.
and 'off' if it is above that temperature. Most reliable system,
however, employs solid-state switching with phase control which
supplies only that amount of power which is required to keep it
at a constant temperature. Still another is called Burst Fire
Control which involves switching on the thyristor when the supply ",,'
Welding~ '-/ \
voltage is at zero and switching off after a number of supply cycles.
With proper choice of cycle time this unit can be used for accurate
control of preheat temperature even for thin components like thin
wall boiler tubes. Pressure ves~el
drum
5.3.2.3. Applications
i:1

The resistance heating process is particularly advantageous

I
I!i

l for the welding of pipes, axles and similar other shapes. Fig. 5.8
shows a wraparound resistance heater used for preheating pipes.
ir It consists of a nichrome wire element embedded in 25 mm thick
c8ramicfibre blanket which is housed in a stainless steel shell.
I~
i~
'ii
Preheating for Rotating Weld Seams : Electrical resistance Radiation
II heating elements cannot readily be fitted directly onto the heating
workpiece where fixed head, automatic welding system requires Heating element
component rotation, as for circumferential welds in pressure
vessels. Under these circumstances, it is more convenient to
provide non-flexible heating units at the lower part of the Power supply
(l
circumference away from the component, with transfer primarily Fig. 5.9. Electrical resistance preheating set-up for a rotating shell.

360 Welding Engineering and Technology
by radiation, as shown in Fig. 5.9. The heating unit may be either
electrical resistance panels, manufactured as a strip or a coil
element design, or surface combustion gas radiant panels
operating on LPG (liquified petroleum gas) or natllral gas.
Apart from the aforementioned preheating methods special
preheating electrodes which deposit no weld metal but produce a
pool of molten slag are also available. Heavy plates and similar
other components may, however, require furnace or oven
preheating. With a correctly chosen preheating temperature hard
zones can be avoided, irrespective of the hardenability of the
parent material. Special large sized furnaces may be required for
preheating big-sized components.
5.3.3. Advantages and Limitations of Different
Preheating Methods
The advantages and limitations of different preheating
methods can be summed up as follows.
1. Flame Heating
The advantages are low cost and portability while the
drawbacks are; minimal precision and repeatability, non-uniform.
temperature distribution, arid the need for operator skill.
2. Electric Resistance Heating"
Advantages of this process are:
(i) Continuous and even heat can be maintained
throughout the welding operation and during long
breaks;
(ii) Temperature can be adjusted quickly and accurately;
(iii) Welders can work in relative comfort and they need not
stop to raise the preheat temperature;
(iv) Uneven heat can be obtained easily. That includes the
heat needed for the top and bottom halves of a pipe or
where pipes are attached to heavier sections as in valves.
Limitations of the process include,
(i) elements may bumout during preheating process,
(ii) a resistance element may short itself out to the pipe
producing arc spot which can be the cause for crack
initiation in service ..
3. Induction Heating
With induction heating, high heating rates are possible;
temperatures can be controlled within a narrow range; local hot
spDts can be avoided more easily; and the coils last for a long time.
361
Preheat and Postweld Heat Treatment
Amongst the limitations of induction preheating are;
(i) High initial cost,
(ii) Equipment is bulky and thus not easily portable,
(iii) provides limited means to compensate for non-uniform
wall thickness and geometries as well as for different
energy requirements of the top and bottom of a pipe in
the horizontal position,
(iv) the power has to be tumed off during welding,
(v) extra set up required if extra coils are needed.
4. Gas-Flame Generated Infrared Ray Heating
This process uses economical fuel and suitable control
equipment is also available for it. Also, no insulation is used in
the area to be heated.
The only disadvantage is that separate furnaces have to be
used
5. Quartz-Lamp Heating
The advantages of quartz lamp heating are fast response time,
efficiency, cleanliness, fast cooling down and quick turn around.
The limitation of this process is a high initial equipment cost.
Also, quartz lamps are fragile and sensitive to contamination.
When employed for preheating pipes separate furnaces must
either be fabricated or made available for each pipe diameter.
5.4. Measurement of Preheat Temperatures
Correct temperature is an important factor in preheating
operation. There are different methods of measuring preheat
temperature some of which are as follows.
1. Use of tempilsticks or thermal chalks which are crayons
developed for different temperatures. When applied, the
markings melt and change colour at the designated
ternpera tures.
2. Marking the surface with a carpenters blue chalk. A
mark made with such a chalk turns to a whitish gray
when the temperature reaches approximately 330°C.
3. Rubbing 50-50 tin-lead solder on the surface; this starts
melting at IS3°C.
4. Rubbing a pine stick on the heated surface. The pine
stick chars at 330°C.
5. Use of surface thermometers or thermocouples.
For accurate temperature control the current practice is to use
type K thermocouples the hot junction of which is welded directly
 I
362 Welding Engineering and Technology
,)
to the workpieces with the help of portable capacitance discharge
(percussion welding) unit. However, the use of thermocouples
becomes quite complex especially when unequal thicknesses and
or geometry are involved which necessitates independent control
on each side of a weld. Microprocessor Ct;:mtrolledunits have now
been developed to cater for such specific welds.
These methods of measurement can be used depending upon
the material used and accuracy required as well as the availability
of the mode of measurement. Shop welding can ensure better
standards of measurement rather than field welding.
5.5. MATERIALS
PREHEATING TEMPERATURES FOR DIFFERENT
The degree of preheat required will depend upon the material
and its thickness. Some details about preheat temperatures
required
in for the: commonly used following materials are included
this section
(i) Carbon steels,
Ui) Low alloy steels,
(iii) Stainless steels, and
(iv) Aluminium.
5.5.1. Preheating of Carbon Steels
Low carbon steels (::::;
0.27% C) do not form hard zones except
in thick sections of steel with carbon above 0.20%. The steels in
this groups, including structural plates and sections, boiler plates,
etc. are not usually hardenable enough to require any heating
before welding. Preheating is necessary only when welding
exceptionally heavy sections. Then even it needs a preheating
temperature between 100 and 200°C. However, the preheat
temperature should not be below the critical temperature for
transition from ductile to brittle behaviour; particularly so if
welding in winter conditions. Also, it must be borne in mind that
there is a considerable danger of cracking under the influence of
welding stresses if one uses a material with a notch impact value
below 8-10 J. It is, therefore, recommended to refer to the impact
value curve for the material when it is to be fabricated into a
strl1cture under severe stress conditions. Even a temperature
increase of 100e, say from -20°C to -lOoe can, at times, make a
considerable improvement in tl1e service behaviour of the
component.
Preheating may also be necessary when welding carbon steel
of quite low carbon content, for example, if a very thin plate is
Preheat and Postweld Heat Treatment 363
welded to a considerably thicker plate. To correspond with the
thinner plate it is imperative to use a small diameter electrode,
and if welding is carried out at room temperature; the cooling rate
in the heavier plate may be sufficiently rapid to harden even low
carbon steel. Slight preheating (50-100°C) is enough to avoid this
hardening. However, the correct solution ofthis problem is to taper
the heavier section so that sections of nearly equal thicknesses
can be welded at the joint. This eliminates the risk of hardening.
Preheat temperatures for medium carbon steel with a carbon
content of 0.27-0.45% should be between 95 and 370°C depending
on the carbon content. The greater the carbon content, the higher
the preheating temperature. However, for welding most low
carbon steels (in the above mentioned range) it is more economical
to use the controlled heat input process than to employ preheating.
The principle behind this technique is to induce a large volume of
heat into the baE'<:l
metal by welding with high current at low speed
or by making multipass welds. The high current and slow welding
speed build up considerable heat in the metal. This slows down
the rate of cooling and results in subsequent prevention of hard
zones near the weld area.
Plain high carbon steels (C > 0.45%) are difficult to weld as
they form hard martensite in the HAZ to cause severe brittleness.
The best procedure to avoid excessive hardening is to utilise a
controlled system of preheating with a preheating temperature
between 95 and 205°C for steels with a carbon content of 0.45 to
0.60% and 205 to 370°C when the carbon content is over 0.60%.
In multi-run welding the metal deposited during the first run
heats up the base metal. The heat of the next run tempers the
base metal adjacent to the weld bead. Each successive run then
leaves enough heat so that there is no rapid cooling and therefore
no appreciable hardening.
5.5.2. Preheating of Low Alloy Steels
Total alloying content of low alloy steels usually does not
exceed 5%. There are thousands of varities of low alloy steels and
most of them are hardenable by quenching, i.e. high rate of cooling,
due to the formation of hard martensite. Easier it is to
quench-harden a particular type of steel more prone that steel is
to weld-cracking so greater the care required in welding it. The
situation becomes progressively worse with decrease in relative
size of the weld.
Because the danger of martensite formation increases with
hardenability and with the brittleness of martensite formed, so
the difficulty of welding increases with the carbon content of plain
carbon steels and with the equivalent hardenability of hardenable
 365
364 Welding Engineering and Technology Preheat and Postweld Heat Treatment

Si Mn Cu Ni Cr Mo V .
alloy steels. Since the hardenability of low alloy steels varies with
alloy content, a simple approximate relative measure has been Pcm = C + 30 + 20 + 20 + 60 + 20 + 15 + 10 + 5B ...(5.4)
devised in terms of carbon equivalent (C.E.). A number of and the Susceptibility Index is given by,
relationships have been proposed to determine C.E., some of these ...(5.5)
are as follows. S.l. = 12Pcm + IOglO H
where H is the hydrogen level which is determined from Table 5.2.
...(5.1)
•• -/0 20 + 15 + 10
C E = £d. C (%Mn %Ni %Cr + %Mo + %V) Table 5.2. Susceptibility Index Grouping'
For ordinary C-Mn steels, the equivalent proposed is, < 0.38

C.E. = %C +
HzHi
+ -4 4
%Mn %Si

H3
(H) Level (5.2) ... Hydrogen
F
D
BA
G
E
<
CB
C0.33
0.23
0.28
< 0.18
Pcm

Another typical formula for determining the carbon equivalent

of a steel containing not more than 0.5% C, 1.5% Mn, 3.5% Ni, 1%
Cr, 1% Cu, and 0.5% Mo is,
C •.E = at'l0 C + %Mn
6 + .%Ni~Cr
15 + 5 + %Cl.!.
13 + %Mo
4 ....(5 3)
So it is evident that C.E. is only a rough guide and that it where, HI = 5 ml/l00 gm of weld metal
becomes increasingly unreliable as the alloy content and number Hz = 10 ml/100 gm of weld-metal
of elements increase. Hs = 30 ml/100 gm of weld metal..
As a rule it is unsafe to fusion weld thick sections of plain and Susceptibility Index Groupings are as follows.
carbon steel with carbon equivalent more than 0.25%, witnout A = 3.0 E = 4.6-5.0
preheating. Table 5.1 gives a rough guide to preheat temperatures B = 3.1-3.5 F = 5.1-5.5
for severe welding condition such as a small weld being laid on a C = 3.6-4.0 G = 5.6-7.0
long thick plate. D =4.1-4.5
Table 5.1. Preheat Temperatures for Plain and Low Level HI
Alloy Steels The hydrogen level II I pertains to the cleanest consumables
Preheat Temperature, °C that can be used. These might be low-hydrogen electrodes taken
from hermetically sealed containers dried at 370-430°C for one
So 1.
No1. 0.40
Carbon Equivalent, ~ 50
hour, then used within two hours after they are removed from the
2. 0.45 100 oven or the consumables could be clean solid wires used in gas
3. 0.50 150 metal arc welding.
4. 0.55 200 Level Hz
250 The 'Consumables used for hydrogen level Hz could be low-
~_~ 5. I 0.60
0.65 ..
L' 30_1) _
hydrogen electrodes taken from hermetically sealed containers
and used within 4 hours. Submerged arc welding with dry flux
5.5.2.1. Susceptibility Index Grouping Method would also qualify under level H2•
Another method of determining preheat and interpasss Level Hs
temperatures, based on C.E. formulae, is known as Susceptibility
Index Grouping. The carbon equivalent or composition parameter, The hydrogen level Hs pertains to the use of all other
PUP takes account of the different alloying elements by the consumables which do not qualify under levels HI and Hz·
following relationship.
 367
3ti6
----------------------~
Welding Engineering and Technology 1,
Preheat and Postweld Heat Treatment

After determining the Susceptibility Index Grouping the

I ~ /10~~ ~~0 ~0~ 0~~ ~0~ ~~0 ~0~ ~~<:> ~<:> ~~0 ~~<:> ~<:>
~ ~ ~~. ~ ~
000 ~ ~
preheat and interpass temperatures can be determined for
common fillet and groove welded joints for three levels of restraint
--
....
by referring to Table 5.3.
ai
= Low Restraint : Common fillet and groove welded joints in
... which a reasonable freedom of movemeItt of members exists .
,t:J ~I~ ~ ~
~ ~ ~ ~ ~ ~
~ ~ ~ ~
~ ~ ~ ~
~ ~~ ~ g g~ g~ ~g
~ ~
Medium Restraint: Fillet and groove welded joints in which
~
~--
..:=0
a reduced freedom of movement exists due to members already
being attached to structural work .
•••• 0 --
~ .•..
High Restraint : Welds in which there is almost no freedom
o~ .••
r:I
of movement for members jointed, such as repair welds, especially
-
.c ~
<
(1J
:..
.•..
III
~
~I
~
.~
~
~IO~ g~ ,S
~ ~
~ ~
M 0 M~ ~~ ~~
~ g ~~ ~~ ~~ ~g ~g
~
in thick materials.
Problem 5.1. Determine the preheat and interpass
~~ ~
::l'""' ~ temperature by Susceptibility Index Grouping method for a 38 mm
~-
••
(C
Cl.l

~~
0
III
4.l
~I
~~ ~ .
~
I~v ~ <:> 00
0 m m
~ ~ ~V 00 ~
00<:> ~ ~ ~
<:> <:> ~ ~
<:> ~
<:> ~
<:>
thick steel plate of the following composition .
C = 0.23% Si = 0.30%
Mn = 1.35% Cu = 0.20%
a:;J ~
P = 0.04%
~ 4.l ~
(1J 4.l
(1J ;..
~
~ S = 0.05%
g,~ ~I~I~v <:> <:> <:> <:> ~
~ ~ ~ ~ v
<:> ~
~ ~
~
~
~
~
~
v
~
~
~
~
~
~
~
~ The plates are to be fillet welded by using rutile coated
~ electrodes under medium restraint.
~.s
;..

JS (1J
Solution.
~~ Step 1.
;~
..... ~I~
v
~
v
~
v
0
~
0
~
~
v
~
v
0
~
<:> ~
~ ~
~
v
<:> ~
~ ~
g g
•....•
~
Find the Composition Parameter, Pern, using equation (5.4)
(C .•••
Cl.l ::;
Si Mn Cu Ni Cr Mo V
.s=Q:l Pern = C + 30 + 20 + 20 + 60 + 20 + 15 + 15 + 5B
~~
Cl.l

'"'" = _-. O'23 + 0.30

30 + 1.35·
20 + 0.20
20
~~ ~I~ v ~v ~v ~
0 0 ~ ~ ~ 0 ~ ~ ~ <:> ~ ~
S$
~ v v v ~ ~ v v ~ ~ ~ = 0.23 + 0.01 + 0.07 + 0.01 = 0.32 Pern = 0.32
·a ~ ...;
•..
Step 2.
...• '"
::S,s
;..

...
Q,
•..
For finding Susceptibility Index Grouping, refer to table 5.2.
'"
co5 '"
~
Cll
For rutile electrodes, hydrogen level = Ha
aC ~E <:> ~ ~ ~ ~ <:> ~ ~ ~ >Q <:> ~ ~ ~ 10
1j,§. •.•.• c:!, d, J,
M 1\ •....• c:!,
r-4 d, J, ~ ~ c:!,
,...-l J,
d, M ~ ~ .'. For Hydrogen level of Ha and Pern = 0.32, we get
~ f'""'t 1"""4 ('fj f'""'t
o
.c ~ ....
r: Susceptibility Index Grouping = F
~ .s
f'" Step 8.
1s-~ E For fillet weld with medium restraint with Susceptibility
i ~bll '"
~ Cll

~.~ 13 ::a .fi Index Grouping F, the preheat and interpass temperature, by
~~ ~ <..>

~ " I referring to table 5.3, is 150°C.

* Ans: 150°C
368 Welding Engineering and Technology Preheat and Post weld Heat Treatment 369

To determine the preheat and interpass temperature more In a Jominy Test a specimen 102 mm (4 iJ:lch)in length and
accur-ately for a specific low alloy steel it is best to make use of 25.4 mm (1 inch) in diameter is heated to 1l00°C and
the information provided by the Jominy Hardness Test. water-quenched in a jig by means of a water jet of fixed
5.5.2.2. Jominy Hardenability Test characteristics playing upon its lower end, as shown in Fig. 5.10.
Jominy Hardenability Test or End-Quench Test is very widely Thus, the rate of cooling is extremely rapid at the lower end, and
used, at present, to estimate quite accurately the welding decrease towards the upper end. When the specimen cools down,
conditions required to avoid the HAZ (heat affected zone) hardness measurements are made along the cylindrical surface.
har-dening. The changes in hardness with increasing distance from the
quenched end give a picture of the hardenability of the steel. Thus,
low harden ability steels are relatively soft only a few mm from
Top of the quenched end while the air-hardening steels exhibit high
Fixture hardness along the entire length of the specimen.
Since the test conditions are fixed in all respects each point
along the testpiece corresponds to a known cooling rate. Generally,
however, it is not necessary to express the cooling rate in
degrees/see instead it is sufficient to quote the distance in mm
from the quenched end. For example, cooling rate 20 mm distance
from the quench end will be known as '20 mm Jominy distance'
8 . I or 20 J.
1/2in.
(12.7 mml 1/ .n.
13.2 mml II-- I--11-1/8
I i.,
129 mm I Jominy tests are carried out in a similar way in determining
the weldability of a steel and the test piece is always heated to
1100°C before quenching, irrespective of the normal hardening
dr::::::~~~~, ,
T 1,.50
1 - 1/32 in. (26·2 mm I
temperature of the steel.
Tables have been drawn up showing the cooling rates obtained
as a function of different welding variables such as arc voltage,
welding current, welding speed, electrode diameter, plate
Unimpeded thickness, type ofjoint, and degree of preheat. Thus, it is possible

T
water jet to determine the maximum cooling rate which can be tolerated if
I. in.
[1 C12 mm I '-in.(25.10 mml a given hardness is not to be exceeded in the HAZ. For example,
Round specim en it may be shown that a certain steel cannot be cooled faster than
10 mm jominy rate if the HAZ hardness is to remain within safe
limits. The tables thus indicate the welding conditions for a given
2-1/2 in. type of joint to produce required cooling rate in the HAZ.
(64mm)
Although Jominy test results can at times predict HAZ
hardness very accu!ately but still this test has the following
. -L~ IZI,.-( ±2.8-() limitations. ---
J :I 1/2in
• I 'I I
I I
Orifice at7S°n 5°,
rwa'" 1. The HAZ of a weld is not only cooled rapidly but is also
first heated rapidly and has little or no soaking time,
In mm) I
1
whereas the Jominy testpiece is heated more gently and
From quick-opening
. rn1/Z-in.112.7
'"'-
mm) i·d. soaked before quenching.
2. The maximum HAZ hardness can reach values greatly
Valve f in excess of those found at the quenched end of a Jominy
Fig, 5.10. A schematic representation of the set-up for Jominy Hardenability Test.
37\1 Welding Engineering and Technology Preheat and Postweld Heat Treatment 371

testpiece. This is because the alloying elements cannot 6·4

~.t: H V (approx.)
11I111 ,,·0
v·o
lt~ b HI
~.0
0ji 4
diffuse completely during welding. Consequently the HRC
'-
Jominy test specimen, which corresponds to steel in
able results for fine-grained rather than for coarse
-
which diffusion is practically complete, gives more reli-
~
1/1-

~
70
60
50
1150
765
540
389
grained steels. 287
~ 30
3. The rate at which a Jominy bar is heated is slow com- ::: 20 223
pared with welding conditions, and this leads to com- -E
~ 140
10 187
plete transformation and homogenisation of the -0 0 156
C)
O' 5 1 2 3 4 5 6 7 8 9 10 20 30 40
austenite. In welding, on the other hand, the heating ex:

period is only 3 to 15 seconds ; diffusion is thus incom-
plete and, certain alloy carbides, for example; are not
taken fully into solution.
4. A Jominy test bar cannot be austenised at temperatures
above about 1100°C, whereas in the HAZ the maximum
temperature reached is the melting point of the
material, this is a fact of primary importance as regards
the austenisation process.
5. Since a Jominy test bar is held at austenitinsing
temperature for an appreciable period, the prior struc-
ture of the material has little or no influence on the
result. In welding, on the other hand, the prior structure
is of vital importance for the behaviour of the HAZ.
6. The size of the original jominy test bar makes it un-
suitable for testing material under 1 inch (25.4 mm),"
Jominy distances: sixtee~rhs of an-inch
Fig. 5.11. Jominy hardness curve for a rail steel.
Jominy curve for a rail steel containing 0.69% C, 0.15% Mn,
0.016% P and 0.013% S, is shown in Fig. 5.11. Ii shows that if
maxim urn hardness of Rc 40 is allowed in the HAZthen the cooling
rate should not exceed 8 mm jominy distance otherwise harder
constituents will form with consequential increase in hardness.
By referring to the Preheat Temperature Table for this steel it can
be determined as to what size of electrode must be used with a
particular preheat temperature to 'keep the HAZ hardness to a
maximum of Rc 40.
Jominy Curves and preheat temperatures required for
different types and sizes of coated electrodes for butt and fillet
Jominy distances in mm I approx.1

while most welding applications use material about i HV

1000
0.8 1.6 74.8 8 11 1624 HRC lapprox.1
69.0
(6.3 mm) thick. 900~1 0;;;;;; 64.5
7. The Jominy end-quench test is not very suitable for low 1=0-- 3 _I-- ~ Curve No· 0/0 C
carbon steels or the shallow hardening types of higher- 800 ~It
~.
""" ~ '\ 61·5 1 0·83
carbon steels. 7001--5'-- "'" ""''' 1\ 57.5 2 0·56
In the above stated limitations, the Jominy test has been of 3 0·48
600 E6 ~ •.....
""" "~ f\' 53.0 4 0.43
greatest help in connection with the repair welding of fairly deep 5
hardening high carbon steels. Also, Jominy test has been usefully 500~7 T"':: I\: f\ 47.5 0.37
6 0.30
carried out on weld metals to get an indication of risk of its 400~8~ ~ , ....•.
"'""" ~
...••~~r\ ~ 41.0 7 0.20
hardening on rapid cooling. Thus, certain low-alloy steel deposits 8 0.10
are unable to withstand a rapid quench, such as occurs when 300 .•.••....•.•.....•••.. ~ 31·5
welding very heavy sections. Both weld metal and parent material 200 ""':::", ~13.S
will have limited ductility and preheating may be necessary even 100 ' • 1.0
though the parent material may have good weldability. It is most o
important to bear in mind, when welding high strength low alloy 0·5 1 2 3 4 5 6 78910 15
(HSLA) steels with low alloy steel electrodes that the cooling rate Jominy distances; sixt;enths of an-inch
of the weld may require to be limited by preheating. Fig. 5.12. Jominy curves for plain carbon steels of various carbon contents.
I
J7Z Welding Engineering and Technology Preheat and Postweid Heat Treatment 373

weLdsin plates of 13-50 mm thickness for some typical carbon and Plate Material and Weld Metal Compositions
low alloy steels are given as follows. 0.28
0.027
0.010
0.031
SP
·0.010
0.65
0.019
1.47
1.69
Mn
Si
0.015
1.55 0.15
0.16
C
0.10
Fig. 5.12 shows a number
I Weld
Plate of
Plateforfor
for Jominy
Figs.
Fig. 5.13
Fig. curves,
5.13 and
5.13 5.14 illustrating the
changes brought about in the curve by varying only the carbon
content of steels containing no other alloying elements.
Jominy distances, mm I approx.) Jominy distances, mm (approx.) H V (approx.)
HV
0·8 1.6 3.2 4.8 16 HRC 0 0·8 1.6 3.2 4·8 9·6 16 32 64
32 48 80
--
I I 1 I I I , • I I
~~oo 470 1150

-- , t--
I 70
Weld' 400
•.. ~--~ "-"-
\-
I--.-Plate
.-
""~
'I'-- ..•.....••.

..•
..•
"-
I 425
60
50
765
540
389
...>
•.. 40
30 287
20 223
!--...L ! 1 87
300 305 10
'" o 1:50
'" 0·5 1 2 3 4 5 6 78910 20 30 40
QI
c:
"0
•...
Jominy distances: sixteenths of an-inch -
o
:I: 200 200
Composition of steel
C = 0.10% Cr = 0.59%
Mn = 0.76% Ni = 0.42%
100 100 Si = 0.25% Mo = 0.60%
0·5 2 3 4 5678910 20 30 405060
P = 0.09% Cu = 0.60%
S = 0.02% N = 0.002%
Jominy distances in sixteenthsof an inch
Fig. 5.15. Jominy curve for 510 N/mm2 UTS low alloy ship plate.
Fig. 5.13. Jominy curve for 510 N/mm2l)TS steel plate of normal
Mn content and the corresponding weld metal.

-
C Sl Hn P S Jominy distances,mm (approx.) H V (approl<.)
Plate fpr Fig. 5.13 0·15 0·2' 1·47 0·019 0017
Plate for Fill' 5·14 0·16
Weld for Figa.5.13 1.5.14 0·10
0·2.
0·65
1·69
1·55
0·010
0·015
0·031
0·010
HRCO
70
0·8

I I,
1·6

,I ,
3·2 4.8
I II -
9.6 16 32 64
...., 1150
60 765
Jominy distances, mm lapprox.l_ 50 540
0·8 1.6 3.2 4·8 16 32 48 80 HV 40 389
1..70
30 287
400 _- ...
,, 425 20
10
223
187

,
"-
•.. •.. o
0.5 1 2 3 4 5 67 8 9 10 20 30 40
Jominy distancczs:sr:tczlZnths of an-inch
300 305
Composition of steel
C = 0.27% S = 0.017%
M:n = 0.50% Cr = 1.49%
200 200 Si = 0.34% Mo = 0.28%
P = 0.019%
Fig. 5.16. Jominy curve for Cr-Mo heat-treatable aircraft steel.
100 I I I I I I I I I III I I I I II 100
0,5 1 2 3 456 8 10 20 30405060 Figs. 5.13 to 5.25 give typical Jominy curves for a number of
Jominy distances in sixteenths of an inch_ low alloy steels often used for fabrication by welding. For all these
Fig. 5.14. Jominy curve for 510 N/mm2 UTS steel plate of high-Mn
content and ·the corresponding weld metal.
 375
374 Welding Engineering and Technology Preheat and Postweld Heat Treatment

H \J CQPP

, HR CO
SJ.Cl 50
rox.)

..
Jominy
0.8 1.6
-
distances,mm

------......
3.2 4.8
(approx,)
6.4 16 19·2 HRC
70
60
Jominy

0.8 1.6
distances

3.2 4.8 9·6

,mm
16
(appox.)

......•..
32 64
H V (approx.)

1150
765
540
389 40 ;::;'=.:,::"';:":;;i:::!:' ::..~~--=- __ ';.~...: . :: 40 t-....
389
287
c 30 22'3
~ 20
r ,="""- '-~ .•••.,
~ ..••.•••••.•.•••.
•. ~--. J"'. Cl ISO10
187

...
137 30 ",.~ " ~
. lit
:z:: 0
1 2 345678910 203040150
. JE: "'~ 0·5
M
.•..
C
"a
. !---
r;,..,"::- r -I"- Jominy distances: sixteenths of an- inc h
...•.

~;
:; Z23 20 Composition of Steel
:z:

187 10 M~ = 1.84% 8 = 0.013%

81,",0.21%
C = 0.59% \ P = 0.024%
3
IS IJ 0 Fig. 5.19. Jominy curve for 1.84% Mn heat-treatable steel.
o

Jominy
0·5

distances;
-
sixteenthsof
2 3 4
an - inch
5 6 7 8 9101112 Jominy distances , m m (appox.)
H V (approx. )
HRC 0.8 1.6 3.2 4·8 9.6 16 32 64
Composition of base metal steels 70 1150
(1-6) and weld metal (7) 60 765
I 540
1.10' C
1.15
0.80
0.14
0.39
0.98
0.12
0.15
0.16
0.13
0.57
0.41
0.87
0.86
0.65
0.49
0.42
0.12
0.32
0.36
1.09
0.11
Mn0.40
0.79
0.21
1.06
0.18
0.66
0.24
0.16
1.04
8Mo
Cr
i
1.02 389
~ 40 I-.. 287
'C 30
~ 20 223
:: 15010
:z:: 187
o
0.5 1 2 '3 4 5 67 89 10 20 3040 150
Jominy distances: sixteenths of an- inch
Composition of Steel

M~ = 1.16% 8 = 0.014%
Fig. 5.17. Jominy curves for Cr-Mo creep resisting steels (1-6) and weld metal (7). 81
C '"' 0.19%
= 0.44% \ CrP == 0.55%
0.019%
Jominy distances, mm (approx.) Fig. 5.20. Jominy curve for 0.55% Cr-Mn heat-treatable steel.
H V (app rox.) cases the Jominy test bar was austenitised at 1100°C. This is a
HRC 0.8 1.6 3.2 4.8 9.6 16 32 64
I significant point because the majority of published curves refer to
-
I I • I • I I •• I '"
70 1150
'" 60 765 specimens heated to the hardening temperature which is usually
540 below noo°c. The curves are affected by increased hardening
c: 40 389
~ 30 287 temperature, especially in low alloy steels; more austenite is
:l f
Cl
:x: . 10
20
50 2 23
1 87
retained and because of grain growth the steel becomes
o 1 50 deeper hardening. The hardening temperature of noo°c is used
0·5 1 2 '3 l., 5 61& 9 10 20 30 40 for weldability studies and in most cases tests carried out with
Jominy distances: sixteenths of an -inch other hardening temperatures do not give data which can be used
Composition of Steel for this purpose.
Mn = 1.04% 8 = 0.028% Comparison of Figs 5.13 and 5.14 shows the effect of varying
C == 0.43%
8i 0.38% I P = 0.053% Mn-content on hardness.
Fig. 5.18. Jominy curve for 1.04% Mn heat-treatable steel.
376 Welding Engineering and Technology Preheat and Postweld Heat Treatment 377

Jomlny distances, mm (approx.) Jominy distances, mm (approx.)

H V ( appro x.) H V ( approx.)
1.6 3.2 4·a 9.6 16 32 64
HRC
0.8 1· 6 3.2 4·8 9.6 16 32 64
, 1150 H RC
70
0.8
1150
l8 765
765 540
:::50 to- 389
'" 40 287
~
""""'-L::-::r::J 389
5 40
2&7
~ 30 223
::t
~
•... 20
so
10
223 ~ t 30
:x:: 10 187
o 150
187 o 150
0·5 1 2 3,. 5 6 7B 910 20 30 40 0.5 1 2 3 45678910 203040
Jominy distances!sixteenths of an-inch Jominy distanceslsixteenths of an-inch
Composition of Steel Composition of Steel
Mn = 1.20% S = 0.023% Mn = 0.66% S = 0.008%
C == 0.36%
Si 0.32% I CrP == 1.03%
0.025% SiC == 0.21%
0.36% I Cr 1.17%
P == 0.019%
Fig. 5.21. Jominy curve for 1.03% Cr-Mn heat-treatable steel. Fig. 5.24. Jominy curve for 1.17% Cr heat-treatable steel.
i
J 0 min y d s ta n ces I m m (a p pro X. ) Jo'miny -distances, mm (approx.)
H V (app rox.)
HRC 0·8 1.6 3.2 4·8 9.6 16 32 64
70 1150
60 765 1150
540 765

~I
•• 40 389 540
..;
a
30
20
50
t-I 287
223
389
287
:x:: 10 187 223
o 150 187
0.5 1 2 3 45678910 20 3040 150
0,5 1 2 3 4 5 6 7 8 910 20 30 40
Jominy distances:sixteenths of an -inch
Jominy distances:sixteenths of an - inch
Composition of Steel Composition of Steel
Mn = 0.36% S = 0.015%
Si 0.29% CrP == 0.60% Mn = 0.56% Cr = 2.72%
C = 0.51% I 0.034% C == 0.31%
Si 0.18% I NiS == 0.81%
0.005%
Fig. 5.22. Jominy curve for 0.60% Cr heat-treatable steel. P = 0.039%
Jom in)' distances, mm (app rox.) Fig. 5.25. Jominy curve for Cr-Ni heat-treatable steel.
H V (app rox.) All these Jominy Curves (Figs 5.13 to 5.25) are intended for use
HRC
in conjunction with tables 5.4 and 5.5 and facilitate an accurate
0·8 1.6 3·2 4.8 9·6 16 3~4
70 1150 estimation of the welding conditions which will in each case produce
U~ a suitable cooling rate in the HAZ. The maximum permissible hardness
'"
~
~
",,60 20
SO
40
30
..••.. - 389
2&7
223
is first established and the corresponding Jominy distance is found on
the curve for the parent material. The tables then help in knowing the
r::lI

;r: 10 187 preheat temperature to be used to produce the corresponding cooling

o 150 rate, as a function of different variables, viz., section thickness,
0·5 1 2 3 4 5 6 8910 20 30 40
electrode diameter, type of joint, and type of coating, which in turn
Jominy distances:sixteenths of an-inch determine the arc voltage and welding current. Within the range of
Composition of Steel hardnesses normally permissible in HAZ (250-350 Brinell, 25-37
C = 0.38% S = 0.008%
Mn = 0.66% Cr = 0.95% ," Rockwell C or 250-360 Vickers) the calculated welding conditions will
Si = 0.24% Mo = 0.28% generally be found in conformity with actual results.
P = 0.023%
Fig. 5.23. Jominy curve for Cr-Mo heat-treatable steel.
 Preheat and Postweld Heat Treatment 379
378 Welding Engineering and Technology '--'
For Basic Electrodes Operating {Preheating} Temperature in °C
Table 5.4 Preheat Temperatures for Welding for different Joints and Plate Thicknesses in mm
with Basic Coated Electrodes Jominy Electrode 1st Run in U or V Butts Fillets
Distance Dia.
Note: 1. 0 in the Table indicates no preheating.
2. A blank space or - in the Table indicates that the required mm mm 13 25 38 50 1.'3 25 38 50
preheating temperature is unattainable in practice. 1.6 300 450 550 600 400 600 600 -
2.0 250 400 500 500 350 500 600 600
For Basic Electrodes Operating {Preheating} Temperature in °C 2.4 150. 350 450 450 300 450 500 600
f-' or di erent Joints and Plate Thicknesses in mm
1st Run in U or V Butts Fillets
9 3.15
4.0
0
0
300
200
400
300
400
350
200
150
400
300
450
400
500
450
Jominy
Dista nee. I Electrode
Dia. 5.0 0 100 200 250 0 250 350 400
mm 25 38 50 13 25 38 50 5.8 0 0 150 200 0 200 300 350
1.6 0 o ioo 150 0- 100 150 150 1.6 450 - - - 550 - - -
!!!!I'
2.0
2.4
~00 o
o
o
o
100
o
o
o
100
o
150
100
150
100 2.4 200 450 550 - 400 600 - -
3 3.15 0 o o o o o o 100 11 2.0
3.15 350
100 550
400 -
500 -
550 500
300 --
550 - \ ---
4.0 0 o o o o o o o 4.0 0 300 400 450 200 450 550 -
5.0 0 o o o o o o o 5.0 0 150 300 350 100 350 450 550
5.8 0 o o o o o o o 5.8 0 0 200 300 0 300 400 500
1.6 0 100 150 200 10'0 150 150 200 1.6 - - - - - - - -
2.0 0 100 100 150 o 100 150 200 2.0 - - I _
I
2.4 0 o 100 100 o o 100 150 2.4 400 - - -. - - - -
3.15 0 o o 100 o o 100 100 13 3.15 200 - - - - -- - -
4.0 0 o o o o o o 100 4.0 0 600 -- - 300 - - -
5.0 0 o o o o o o o 5.0 0 250 600 - 150 - - -
5.8 0 o o o o o o o 5.8 0 200 400 600 0 500 - -
1.6 0 150 2'00 250 150 200 250 300 1.6 - - - - .- - - -
2.0 0 150 150 200 100 200 250 250 2.0 - - - - - - - -
~~ 2.4 0 100 150 150 o 150 200 200 2.4 - - - - - - - -
5 3.15 0 o 100 150 o 100 150 200 15 3.15 300 - - - - - - -
4.0 0 o o o 0" o 150 150 4.0 200 - - - 550 - - -

I
5.0 0 o o o o o 100 100 5.0 0 400 - - 200 - - -
5.8 0 o o o o o o 100 5.8 0 250 600 - 100 - - -
1.6 150 250 300 400 200 300 350 350 1.6 -- - - - - - - -
2.0 0 200 250 300 200 300 350 350 2.0 - - - - - - - .-
I 2.4 0 150 200 250 100 250 300 300 2.4 - - - - - - .- -
6 3.15 0 100 200 200 o 200 250 300

I-
4.0 0 o 100 150 o 150 200 250 4.0 250 - - - - - - -
5.0 0 o o 100 o 100 150 200 16 3.15
5.0 500
100 -
600 - - -
300 - - I ---
5.8 0 o o o o o 100 ,150
1.6 200 300 350 400 250 400 450 450
2.0 100 250 300 350 200 350 400 400
2.4 0 200 250 300 150 300 350 400
7 3.15
4.0
0
0
150
100
250
150
250
200 .
100
o
250
200
300
250
350
300
i~ ". 350 = = 150 =

r
5.0
5.8
1.6
0
0

250
o
o
350
100
o
450
150
100
450
o
o
300
100
o
450
200
150
250
200
500 .- 500
I. g5.0
5.8
150
3501
0
-
- I -
-

Footnote
-
-
-
300
-
-
-
- I
-
-- JI

2.0 150 300 400 400 250 400 450 500 For Jominy distances fro;n 22 to 38 mm dia preheating is only practicable in
2.4 100 250 300 350 200 350 400 450 prepared U or V butt joints in 13 mm material using 5 or 5.8 mm electrodes as
I 8 3.15 0 200 300 300 100 300 350 400 indicated in the following table :-
100 250 o 200 300 350

l__J' Jomin 5.8

350
250
5.8
200 31 200
5150 25
5 mm 22 ! 250
550
28
Distance ~ --L-

.~:~I
5.0 0 o 100 Piece,oC
I' 200 150 o 150 250 300
I
5.8
4.0 I 0 o I 0 Electrode
Temperature
! 100
Dia.,
o of mm
Work
100 200 250
(contd.) 50 500
-L...-l
 380 Welding Engineering and Technology Preheat and Postweld Heat Treatment 381

Table 5.5 Preheat

t---
i 7 Temperatures
13
11 2.4
5.0
8.0
2.0
3.15
5.8
6.3 for Welding -' mm
- ----- --- -
400
500
200
03.15
For 0o50
550
300
200
250 400
I350
400 13
550
100
200
600
150
300
100
150 350
38
450
500
250
250
25300 0o0mm
13
00Organic,
4.0
4.0
200
350
150
600
250
100
3.15 400
200
100
200
0400
450
250
150
Electrode
I Dia.
300
550
1st
8.0
5.8
600300Operating
450
Run
300
1.6 Butts and
in U or V(Preheating)
Joints Fillets
Plate I Thicknesses
Temperature I °Cinformrn
in Different
89and 2.0
1.6
1.6
1.6 8.0
8.0
5.8
2.4
5.0
3.15
6.3
1.6
4.0 6.3
5.0
Neutral I
Electrodes
with Organic Neutral AcidDistance
and Acidic
5.8
Coated Electrodes I
I
Jominy I 400
600
100 -
I 400
Note: I
1. 0 in the Table indicates nQ preheating. o 250I I 200
200 1500
250
2. A blank space or .-. in the Table indicates that the required
preheating temperature is unattainable in practice. '50 ~oo
,------·-----------T------------------.--.------- ---------·-------·------------------1
Pm- Organic,
I Ner;tral
I. and Acid I' OperatingJoints
(Preheating) Ternperature
and Plate Thicknessesin in f'or Different
°Cmm. iI

I
l Du;tcw_ce Dla.
JCJmi~~ctr;~::trOde!-l;tRun ,
in--U or V-B_~'t;;-. r--·-------F~let~--1 .

I 2.0 0 0 0 I 0 I 0 0 100 I 150

r~5~o
>

It=~;~---
, -t-i.~--rl---l0-.')--==2;---1-1-:-cfo---=r-i~%J--
I,2-4 3.15 I 0 0 I 0 I 0 zg--/
0 I ::.0 II
-11-13580
0 0 j
I 100

i ;) I 4.0 I 0 0 0 0 I 0 I 0 I 0 0 I
i I 5.0 0 I 0 01' 0 I 0 0 I 0 I 0
iI I 8.0
6.3 I 0 0 0 I 0 , 0 I 0 J 0 0 i
! i 5.8 0 0 0 0 0 0 I 0 Wo

r----t--:I--T
I I 2.4- I gr'go
C I a :gn-
0' :gg
1,)0 !-f -t-:g~r:gg
0 I 0 I 100 -II ~~g
150 1
1 I 3_15 0 0 0 0 0 0' 100 I lOG

!i!
I
I

~----+-H --
!
4- I
.

I
I,2-41
5.0
4.0
5.8

2.0
3.15 .I
- g
0,
0
0'

00
I
0
0

100
00
i
,t- 2g,+,gO
0 ° I
I
0
0

0 I 200
150
100/, 0 I 100
150 00
0
0
0
J
I--iio
0
0
0

I
0
0
0

2g0"
200
100
150
I
I
0
0
0

200
I 0
·jo-t- 2~0 _
,
_ 100~!

250 II
150 III 200
200 150
0

5 !
!
5_0
4_0 I 00 00 00 I}
0 II 00 00 0
100! 100
150 i

I 6.3
5.8 0 'I 0 0 0 I 0 0 Ii 0 0
8.0 0 0 0 + 0 0 0 0 0
-- ----. -''1.'6 IOor- 250 -gOO 366- -200 300 350 350

II :3.15
2-4 I 00 I! 0
150 150
200, I 150 1100
200 0 150
I 200 200!
250 250
300
I 6 I 4_0 i 0 0 I 100 i 100 I 0 100 150 I 200
I I 5.0 I 0 ! 0 ! 0 I' 0 , 0 I 0 100! 150
I
i
L __
I -Li
!
5.8
8.0
6_3 II
0 I
O.J.
0 i
0
0
!
I
!
0
0
0 LI
0 I
0O'j ,I
0
0 I
0
0() i
0 I 100
0 _I 0 JI

(Contd,)

(Contd.)
 383

'----Electrode
382 Preheat and Postweld Heat Treatment
-For --Organic,
Welding Engineering and Technology

I
---------
100
450
2eO
350
250
600
0100
25
38
0
20013
50
100
200
500
500
150
300
400
250
400
300
600
3500
----- -
00400
13 mm
4.0
150
100
150
600 Dia.
Electrode
4.0
250
600
600
0
150 1st Run in U or V Joints
Butts Fillets
For Organic, Operating (Preheating)
and Plate Thicknesses
Temperature in °Cinformm
Different I
41200
300
500
00
0
25
50
13
38
400
600
300
200 - Dia.
mm
50450
13
4.0
300
150
100
0
250
550
200 1st RunOperating Joints Fillets
Butts and
in U or V(Preheating) Plate Thicknesses
Temperature in°Cinfor
mmDifferent
Acid
m Jomin.y 31
38
28
34 6.3
5.8
3.15
5.0
2.4
5.0
8.0
5.8
8.0
1.6
3.15
2.0
Distance
1.6 Neutral andmm
Electrodes Acid Jominy
I 500

Solved Problems
Problem 5.2. Determine the preheat temperature for butt
welding, by shielded metal arc welding (SMAW) process, of 50 mm
thick plain carbon steel having a carbon content of 0.30%. Suggest
the electrode diameter and appropriate current setting for rutile
coated electrodes to be used if the HAZ hardness is nQt to exceed
350 VHN (Vicker's Hardness Nuinber).
(Contd.)
 I
384
Welding Engineering and Technolo~y f Preheat and Postweld Heat Treatment 385

Solution Problem 5.4. Determine the preheat temperature for butt

Step 1. Refer to Fig. 5.12, Jominy Curve 6, !
v

welding, by SMAW process, of 25 mm thick 1 ~Cr - ~Mo aircraft

For VHN = 350, the Jominy distance == 5.5 mm (!
Step 2. Refer to Table 5.5, Jominy distance of6.0 ( ... it is safer
to use preheat value for Jominy distance of 6 mm rather than 5 ] welding current range.
mm)
If 3.15 mm diameter electrode is to be used, the preheat il].
temperature required = 250°C and if 4.0 mm dia. electrodes are !/
" one under-consideration.
used, preheat temp. reqd. ;:: 150°C.
Suggestion: It is suggested to use 5 mm diameter electrodes
with a preheat temperature of 150°C which must be maintained
throughout the welding operation till the job is completed.
Current setting for 5 mm diameter electrode = 160-220A
;::200 A (say)
(Refer to table 7.1, Page 176 of Welding Processes and
Technology by R.S. Parmar) Ans: Preheat temperature = 150°C
Electrode dia. = 5 mm
Welding Current = 200 A
Problem 5.3. Determine the preheat temperature for fillet
welding, by SMAW process, of 38 mm thick Cr-Ni-Mo ship steel of
510 N / mm 2 U.T.S. Also, suggest the electrode diameter and
appropriate current setting for basic coated electrodes to be used,
if the HAZ hardness is not to exceed 300 VHN.
Solution ,'I
'il
Step 1. Refer to Fig. 5.15 and find the Jominy distance for Hv ']
heat treatable steel using 4 mm diameter basic coated electrodes if
the HAZ hardness is not to exceed 280 VHN. Also, suggest the
Solution
Step 1. Refer to Fig. 5.16 as it is for steel which is close to the
Jominydistance for HV280 = 6.5 mm
Step 2. Refer to Table 5.4
For a Jominy distance of 7 mm, for butt welding of 25 mm
thick plates, preheat temperature for welding with 4 mm diameter
electrodes = 100°C
Current range for 4 mm diameter electrodes = 140-200A.
Ans. Preheat temp. = 100°C,
Current range = 140-200A.
Problem 5.5. It is required to fillet weld 50 mm thick
2~Cr - ~Ni heat-treatable steel using basic coated electrodes.
Suggest the preheat temperature and diameter of the electrodes if
the HAZ hardness is not to exceed 400 VHN.
Solution
Step 1. Refer to Fig. 5.25 which gives a Jominy curve for

nearly the same composition of steel as is the one under

= 300.
consideration. For 400 VHN, the J ominy distance = 38 mm
It is equal to = 4.0 mm.
Step 2. Refer to Table 5.4 for .Jominy distance of 4 mm, and Step 2. Refer to Footnote to Table 5.4,
plate thickness of 38 mm for making fillet welds. The preheat For J38, the preheat temperature required = 500°C for welding
temperature of 100°C should be used for electrodes of diameter with an electrode diameter = 5.8 mm
2.4 mm or 3,15 mm.
Ans. Preheat temp. = 500°C
Current setting for an electrode of 3.15 mm diameter Electrode size = 5.8 mm dia
== 90 -140A
5.5.2.3. Preheat and Interpass temperatures for Q & T
= 120A (say) Steels
Ans. Preheat temp. = 100°C The quenched and tempered (Q & T) steels were developed in
Electrode dia. = 3.15 mm the early 1950's and have become extremely popular in
Current setting = 120 A. constructional work because of their high strength-to-weight ratio.
Note : If electrodes of 4 mm or bigger diameter are used, no These steels are water-quenched by special techniques from a
preheat will be required. temperature of 815-870°C and tempered at a temperature of
540-595°C. This produces a microstructure of tempered
=0 -0ships,0>0e?buildings,
Q J,
QJ J,
386 ~
0N
0It:>
...•M
=<3
II
N
00>
.0
'"
.0
<l:>
..0 E
::sS
II
'"~
lQ
'0
c.; Q
=
>
<Xl~
'".It:>
...•
.M
It:>
<C
S
IJ,
lC'l":Q
<l:>
E::
'=? 00 0
00 009s>
I'";;1<:=e~
..•
0::-..g
<l:>
<=>
"'"
.~
~<:'
c'<
10
.~
0<Xl
0Q11,400
~ N
136.0INL
S
II
NL
It:>
It:> US
25
19
32
14,200
51
38
9,450
7,500
::s
C'i
<Xl
::s
...•
---
very
,....;
~ 'T
for
0>
Table
l
value
10,600
in J,
equipment.
Fig. 5.2
0.It:>
0It:>
~'"8,270
<0
0
<C
J,
combination Table
N
Temperature
M II:-
...•
0.0
of
:><
6,700
Table
5,120
Preheat ~
shows
highSome
II
:>< I.0
"l:-
aDaD
I~
"'!
5.7 ""
steel "~
...•
<Xl
<C
It:>
ofSome
c.;
:=
It:>
low-temperature
<Xl
J, ~I'iN
."..•
...•
'0>
Minimum
lQ :-
>
T-l
for
~
S
,....;
C
0>
::x::
the
of
<l:>
:x::
strength,
Welding I0
Maximum
strength
"0
...•
II
at
Recommended
thec.;
"
steels N
Preheating
IM
J,
27,600
0It:>22,000
steels
adverseand
15,750
typical
transformation
of the
5
Quenched areHeat
24
52
93
52
024and
66
10
9310
IIHY-80*
107
conventional
different testing
:- HY-130*,**
Tdesignated
18,500 and
toughness.
for
effect
Because
order
steels ...•
of 0...• N
Inputs·
NL**
T-l*
welding
32,300
extensively
Steels
products
and of
inof Thus,
;;125,600
TArc Welding
preheating
their
700
general
Tempered are
(J/cm)
0Postweld
0~ 36,600
Interpa.ss
strength
whichused
veryT·1Temperatures
aDNL
temperatures.
47,600
39,000
T-l, processes.
Hy-80
listed
on Plate
68,000
and NL
Heat50,000
49,600
these
havefor
Steel
the
and
high
require
Alloy inan
yield
200
lQ
2019-22
*Welding
Thickness
Hy-130;
Treatment
69,000
steels
bridges,
table
Steels
NL
35-38
have
93°e,
(mm)
toughness
5.6.
corrosion
excellent
values
Maximum
strength
25-35
Thickness
38-51
(mm)
16-19
while
storage
65,000
95
and
strength
little
** preheat
Maximum
**
Interpass 22-35
22-25
NL;
•••Can
150 be
13 mm,5.8.
<Engineering
5.7. istanks,
resistance.
indicated.
preheating
good
as good
No
and Technology
penstocks,
135°C;
These
indicated
preheating
Ilimit.higher
25% for> T Q 387
(OC)
35joints
mm,
can begood
welded
Qtemperatures
weldability these
temperature
& forPreheat
&149°C
by
should not or
HY-130 16> mm,
beN/mm2
: more51than
13-16
65°C16-22
65°C; above the
(OC)mm,
~I
aD
'=? ''=?
'=?

earth
'=?

moving machines, and

=?

mobile
'=?
construction combined with good ductility, notch toughness, fatigue
I
388 Welding Engineering and Technology Preheat and Postweld Heat Treatment 389

Tables 5.8 and 5.9 show the maximum heat inputs Although it is impossible to define the exact preheating
recommended for welding steels T-1 and Hy-130, while Table 5.10 temperature with accuracy, it generally lies between 100 and
shows the maximum heat input for different thicknesses of the 300°C, although 13%Cr steels containing 0.10-0.12 % carbon can
different grades of the ASTM A514/A517 steel. However, these often be welded without preheating if the thickness does not
he at input values are for general guidance only and for more exceed 10 mm.
accurate results reference must be made to the supplier's data in In general the preheating temperature required for a stainless
this regard. steel depends upon the type of steel uiz., martensitic, ferritic,
Table 5.9. Maximum Heat Inputs (J/cm) austenitic, and duplex stainless steel.
for Welding BY-I30 Steel
Martensitic Stainless Steels : The cooling rate of
Plate Thickness Shielded·Metal Arc Gas-Tungsten or
(mm) Gas-Metal Arc martensitic stainless steels cannot be reduced in practice to the
10-16 15,800 13,800
level required to prevent hardening in the heat affected zones, and
preheating will only lead to an increase in the volume of material
16-22 17,700 15,800
22-35 17,700 17,700
transformed to martensite. As in all other cases of air-hardening
steels, it is best only to preheat to a fairly low temperature, if at
35-102 19,700 19,700
all. It must be admitted, however, that preheating to about 2000C,
Table 5.10. Suggested Maximum Heat Input (J/cm) Limits for i.e. at or just below the martensite point (M.•), is desirable from
Some Q&T Steels (ASTM A5141A517)
the point of view of preventing cracking particularly in multi-run
S. 32
25
13
19
512,800
95
Plate (mm)
Thickness
6.5
150
No
205 limit
22,300
28,000
14,500
10,400
16,800
19,000
4,950
6,500
7,700
9,500
344,000
5,500
6,000
3,600
7,000
48,000
34,400
26,250
37,600
62,000
4,600 20 Preheat and Interpass Temperature, °C welds. This preheat will prevent the formation of martensite in
each successive HAZ as welding proceeds, instead they remain
austenitic and the welder does not have to face the problem of
embrittlement during welding.
The martensitic 16-20% Cr steels containing less than 0.10%
carbon with the addition ofnickel can very often be welded without
preheating. In the as welded condition, these steels contain a high
proportion of Delta Ferrite, while the low carbon martensite which
5.5.2.4. Preheating Requirements of HSLA Steels they form is relatively ductile at room temperature. Both these
Successful welding of HSLA (high strength low alloy) steels factors significantly reduce the risk of cold cracking.
requires consideration of preheat and control of H2 in the welding Martensitic stainless steels with up to 0.15% C are generally
process. Preheating is generally required for all thicknesses over preheated to 200-300oG. As Illuch as 50% of austenite may then
25 mm and for highly rest~ained joints. Specific preheating be retained in the HAZ and the risk of cracking is greatly reduced.
temperatures depend largely on the grade, thickness and the The maximum HAZ hardness, however, increases rapidly with
welding process used and usually range between 40 to 200°C. increasing carbon content. There is thus considerable risk of
It is generally necessary to use low hydrogen electrodes for arc cracking in the steels with 0.15 to 0.25% Gin which local hardness
welding of HSLA steels to preclude the need for preheating except peaks can occur with values of up to and exceeding 500 BHN (550
VHN).
when welding thick and highly restrained sections.
5.5.2.5. Preheating Requirements of Stainless Steels Ferritic Stainless Steels : The ferritic stainless steels
contain from 16 to 28% Cr and 0.05 to 0.25% carbon. The
Preheating of the joint to be welded in steel is the most mechanical propertie.s of these st~els are unfavourable from the
commonly employed technique for reducing the effects of point of view of welding. Below about 100-200°C they have
martensitic formation. The joint is heated to a temperature which
decidedly low ductility, and preheating to about 200°Cis necessary
depends upon various factors and maintained at this temperature to reduce the risk of cracking under welding stresses.
during the whole welding operation.
390 Welding Engineering and Technology
With reference to weldable heat-resisting steels, there is a type
of CI'-steel used for castings for certain special applications, which
contain up to 35% Cr and 0.7-1.2 carbon. This steel is extremely
difficult to weld, and success is possible only by preheating to red
heat. This is necessary in order to allow the stresses to be
dis sipated during welding and not to reduce the HAZ hardening-
as a matter of fact preheating tends rather to increase the
hardness as a result of certain microstructural changes.
Austenitic Stainless Steels: These steels contain 12 to 25%
Cr and 8 to 25% Ni. The carbon content is low, usually not
exceeding 0.10%. The 18/8 stainless steel containing 18% Cr and
8% Ni is the well known example of austenitic stainless steel.
The austenitic structure is obtained by quench-annealing, i.e.
rapid cobling from 1050-1100°C ; water quenchir..g is preferable.
This structure makes the steel soft, yet strong and non-magnetic.
The mechanical properties are not sensitive to heat treatment
because these steels are not hardenable, nor is there any practical
risk of grain growth or consequent embrittlement. The welding
must, however, be carried out in the cold; since a high cooling rate
is necessary to produce satisfactory mechanical properties and
corrosion-resistance, preheating of any kind is quite unsuitable
and the interpass temperature should not exceed 100°C.
Duplex Steel: The most common duplex steel type contains
26% Cr and 4% Ni as major alloying elements. The structure of
these steels is a mixture offerrite and austenite, and the presence
of the former renders it magnetic. These steels are distinctly
brittle and can give rise to cracking under the welding stresses.
In this respect, however, the duplex steels are not so bad as the
ferritic stainless steels.
For successful arc welding 26/4 CrINi electrodes should be
used and the material must be kept as cool as possible and good
rule is to limit the interpass temperature to 100"C maximum.
5.5.2.6. Modelling Approach to Preheating Problems
With the extensive use of computers the latest trend is to
develop models to predict preheat temperat.ure for a given steel
and its thickness. These models are based on relating the
susceptibility of the weld to H2-cracking with hydrogen in weld,
stress on the weld, and the cooling time. The purpose is to predict
the levels of preheat that are necessary to prevent cracking in
various situations. There have been two approaches viz.,
(i) Hardness Control, and
(ii) Hydrogen Control.
Preheat and Postweld Heat Treatment 391
(i) Hardness Control is a means of controlling the micro-
structure of the weld by using hardness as an indicator. The
harder the microstructure, the more likely it is to crack. By
relating the cooling rate necessary to reduce the hardness to an
acceptable level, it is possible to predict preheat temperature that
will ensure hardness low enough to avoid cracking. This approach
is often used for st.eels having fairly low alloy content and
moderate carbon content. If this type of steel is cooled rapidly it
is likely to form a hard microstructure sensitive to cracking. If it
cools slowly, it will form a soft microstructure which will not be
prone to cracking. So the final hardness of steel can be used
successfully to predict the microstructure and the related preheat
ternperature.
(ii) The second method of preventing hydrogen induced
cracking (HIC) is by determining the critical H2 content to avoid
cracking. This so called critical hydrogen approach is used for the
more modern alloyed steels with low carbon contents. In these
steels, the microstructure does not change appreciably with
change in the cooling rate.
5.5.3. Preheating Requirements of Aluminium and its
Alloys
Although aluminium and its alloys can be welded successfully
without preheating by MIG and TIG welding processes but
because of their high thermal conductivity it is often required to
preheat them as that decreases the effects of expansion ,and
minimises the formation of cracks. Aluminium plate 6 mm or more
in thickness should be preheated to a temperature of 150 to 260°C.
Preheating to these temperatures can usually be done by flame
heating using oxyacetylene preheating torch(es). For large or
complicated parts, however, the preheating is done in a furnace.
It is very important that the preheating temperature does not
exceed 260°C. If the temperature goes above this point, the alloy
may be weakened or parts of aluminium may collapse under their
own weight.
The correct preheating temperature may be determined by the
standard thermocouple, etc. or anyone of the following simpler
methods:
1. If a mark is made on the metal with a carpenter's blue'
chalk, it will turn white.
2. If pine stick is rubbed on the metal, a char mark will be '
left on it.
392 Welding Engineering and Technology
3. If the metal is struck with a hammer, no metallic ring
will be heard ..
5.6. POSTWELD HEAT TREATMENT (PWHT)
Postweld heat treatment is intended primarily as a
stress-relief treatment. For welding some of the higher range
carbon steels, PWHT is as important as preheating. Although
preheating does control the cooling rate but the development of
residual stresses always remains a possibility. Unless these
stresses are removed cracks may develop when the workpiece is
cooled to the room temperature, otherwise the part may be
distorted, especially after a machining operation, if employed.
Apart from stress relieving PWHT can also accomplish
stabilisation as well as reformation of structure of the weldment.
Postweld Stress Relieving: The residual stresses developed
due to welding can be quite severe and may lead to failure when
the component is put into service. The simple solution, if
practicable, is to anneal the whole weldment but this is often not
practical due to size and shape of the fabricated component. Full
annealing may also result in grain growth with consequential
reduction in mechanical strength of the component/structure. To
overcome this difficulty the following alternatives are usually
practised.
1. Sub-critical or low temperature annealing, and
2. Local low-temperature stress-relieving heat treatment
in and around the weld.
However, local heating has to be done very carefully, to avoid
uneven heating as well as overheating.
For local heating PWHT, the heat may be applied by portable
hand torches which though cheap are most unreliable. Other
methods of heating may include automatic traversing torches,
fixed specially located burners, low frequency induction heating,
ahd resistance heating with suitably located coils.
Stabilisation of Structure : The rapid cooling associated
with welding often leads to metastable structure which may
change with time, due to natural ageing, with a possibility of
embrittlement of the structure. Full annealing is a simple method
to homogenise the structures but that may cause grain growth,
coarsening of the metallurgical structure and undesirable loss of
mechanical properties in other parts of the component. The
alternative PWHT method for materials like low carbon (mild
steel), and low alloy steels is to normalise the component to give
a relatively stress-free and comparatively stable structure.
Preheat and Postweld Heat Treatment 393
However, if a proper annealing or normalising operation is not
possible, it may be necessary to use a sub,"critical PWHT to give
as uniform a structure as possible; avoiding at the same time grain
growth or precipitation of a brittle phase.
Reformation of Structure : Postweld reformation of the
structure of materials such as mild steel and low alloy steels is
also often desirable and is relatively easy to affect by normalising.
Other materials also respond well to a full heat treatment. For
example, a weldment of a solution treatable material can be given
a full solution treatment followed by agening, if required. An
undesirable elevated temperature precipitate, such as chromium
carbide in austenitic stainless steel, can sometimes be redissolved
by heating the material into an appropriate temperature range,
higher than that which caused the original trouble, and then
cooling it rapidly.
With certain other materials, a totally different approach may
be needed. For example, to eliminate martensite in the vicinity of
a weld in a high manganese austenitic steel, it is desirable to heat
the weldment to about 800°C, into the fully austenitic condition,
and then quench in water to prevent martensite transformation.
It is evident from the above description that to achieve the
desired quality weld different PWHTs are required depending
upon the material of the structure or component involved. Some
of the typical PWHT processes used extensively in the fabrication
industry include the following.
(i) Annealing,
(ii) Sub-critical annealing,
(iii) Normalising,
(iv) Stress relieving,
(v) Quench annealing,
(vi) Hardening and tempering.
All these processes are described in detail in chapter 3.
5.6.1. Examples of Postweld Heat Treatments
PWHT procedure has to be designed for each specific case
depending upon the material composition, size and shape of the
component and the service conditions in which it is to be used.
However, standard procedures can always be laid out depending
upon material composition, and the maximum wall thickness
involved. The following examples give some details of preheat and
PWHT which may be used in the case of heat treatable alloy steel,
martensitic stainless steel, and austenitic stainless steel
respecti vely.
-------- ••••••.••..•••.. '!!!
•••••....••.•
",-",.",-""..• ~~~~~---~~_._._. - ..-.

394 Preheat and Postweld Heat Treatment 395

Welding Engineering and Technology

5.6.1.1. Preheat and PWHT of Heat Treatable Alloy Steel Fig. 5.26.Using AISI*Steel 4130(C= 0.28-0.33%,:Mn =0.40-0.60%,
P = 0.040%, S = 0.040%, Si :.-:0.20-0,35%, Cr = 0.80-1.10, Mo =
Preheat temperatures suggested for alloy steel of different 0.15-0.25%) as an example, the ideal welding procedure is to use
ranges of carbon equivalent (C.E.) are as given in table 5.11.
Table 5.11. Suggested Preheat Temperatures for Alloy Steels a filler metal of the same composition, preheat to 260°C, weld

r
while maintaining a 260"C interpass temperature using low
___
S. No.
+ Carbon equivalent:
(%)
I I
Suggested Preheat Temperature
. ("_C_~ _
hydrogen electrodes, and give a stress·relief heat-treatment at
about 650°C immediately upon completion of welding. The
], up to 0.45 optional postweld heat treatment can then be given in the following
sequence: austenising at 860-875°C, quench, then give tempering
_2_. :2, j Above
0.45 to 0.60
0.60 -1_ I 95-205
205-370 treatment in the temperature range of 425-595°C.
where the carbon equivalent is based on a typical formula used It may be noted from Fig. 5.26 (a) ~hat the HAZ should be
for determining C.E. of a steel containing not more than 0.5% C, cooled down to a temperature slightly below the martensite finish
},5% Mn, 3.5% Ni, 1% Cr, 1% Cu, and 0.5% Mo, and is as follows. temperature (Mf) before being heEl_tedup for stress-relieving. This
%Mn ---+
C E % = %C +----+ %Ni --+---
%Cr %Cu +---- %Mn is to prevent any untransformed austenite from decomposing into
.. , 6 15 5 13 4 ...(5.6) ferrite and pearlite during stress relieving which may
The above suggested preheat temperatures (Table 5.11) are for subsequently transform to untempered martensite when the
arc welding processes and may be affected by thickness of weldment is further cooled to room temperature. In case
workpiece. For oxy-fuel gas welding processes, however, stress-relief PWHT cannot be carried out immediately after
preheating is usually not required owing to the slow cooling rates completion ofwelding, the temperature of the completed weldment
associated with these processes. can be raised to approximately 400°C, which is in the vicinity of
the bainite knee for 4130 steel and most other heat-treatable alloy
Regarding PWHT of the weldments of heat-treatable alloy steels. By keeping the welded structure at this temperature for
steels they are often heated immediately after welding for about 1 hour or less, the remaining austenite can transform to
stress-relief treatment before cooling to room temperature. During bainite, which is more ductile than martensite. Therefore, there will
the stress-relief heat treatment martensite is tempered and the be no danger of the weldment developing cracks on coolingto room
weldment can, therefore, be cooled to room temperature without temperature. Further heat treatment can be given later to optimise
danger of cracking. This can be followed by PWHT to develop the microstructure and properties of the welded structure.
strength and toughness of the weldment. A sketch of thermal In case the heat-treatable alloy steels cannot be given a PWHT
history during welding and postweld stress relieving is shown in and must be welded in heat-treated condition (e.g. in the quenched
and tempered condition), the softening as well as the hydrogen
~.... Undesired
•.. relieving '0
•..
::>
0. cracking of the HAZ can be a serious problem. To minimise the
~
Welding .•..
Welding A1 E I
(b
Mf softening problem, lower heat input per unit length of weld should
/ ~
Preheating
~
Prehec :3Mf I Time - be employed; also the preheat, interpass, and stress-relief
temperatures should be at least 50°C lower than the tempering
temperature of the base metal before welding. Since PWHT ofthe
weldment is not involved, the composition of the filler metal can
be substantially different from that of the base metal, depending
on the required strength of the weld metal.
5.6.1.2. Postweld Heat Treatment for Martensitic
Stainless Steel
Martensitic stainless steels are known to posses a good
combination of their mechanical properties and chemical
Fi.g. 5.26. Schematic representation of (a) desired. and (b) undesired thermal inertness. This latter property permits them to resist corrosion for
hi:;tIJries for welding and postweld stress-relieving of heat-treatable alloy steel. * AISI-American Iron and Steel Institute.
396 Welding Engineering and Technology Preheat and Postweld Heat Treatment 397

use in normal atmospheric conditions in soft water (hydraulic Welds were made in highly restrained heat-treated plates of
turbines), in contact with food products, and in the chemical 15 mm thickness, using basic electrodes of matching composition
industry, especially the petroleum industry. to the base metal. Under these conditions, the pr~heating
Their favourable mechanical properties at high temperatures temperatures necessary to avoid cracking during a natural cooling
find applications for heat exchangers, steam turbines, valves, etc. of the weld is 100-150°C. However, the assemblies were preheated
Marlensitic steels containing 16-20% Cr and 0.1% C rendered at a higher temperature, in order that the temperature at the end
martensitic by the addition of 2-4% Ni show improved corrosion· of the completion of welding is above 260°C, theMs temperature
resistance particularly in se'a water. of the alloy. The structure on completion of welding is thus mainly
Depending upon the temperature at which the welding composed of unstable austenite.
operation finishes and the precautions taken to avoid cracking,
the weld bead can consist of martensite or a mixture of martensite Mter welding, assembly A was maintained for 1 hour at 300°C,
and untransformed austenite. The amount of austenite present then subjected straight away to tempering at 750°C for 4 hour;
immediately on completion of welding is greater at temperatures assembly B was cooled at 150°C, kept for one hour at this
closer to the Ms temperatures and cooling to room temperature temperature and then subjected to the same tempering t.reatment
before the assembly is put into service causing transformation to as assembly A. The corresponding thermal cycles are shown
a fully martensitic structure. diagramatically in Fig. 5.27.
The weld must, thus, be subjected to a tempering operation for
some hours at a temperature generally between 600 and 850°C to Weldin9 Assembly A f Welding Assembly B
obtain properties of strength and ductility comparable to those of
the base metal. From the point of view of the mechanical
characteristics, tempering gives different results for purely
\
, .
~
~
\
.
-o
QI
L. ~ \

\ ! \~ \ / \
\
martensitic structure, and a structure having a high proportion of ::J
, 750 ( ~ \ 750 (
austenite. With purely martensitic. structure, the martensite L.
QI

decomposition gives rise to a structure composed of ferrite and C.

E
uniformly distributed carbides, resulting in best compromise to be I \. \ / \
.~\---t---,
QI
I- \300 .(
obtained between strength and ductility. On tempering a structure -- ---Ms =260'(
having a high proportion of austenite, austenite decomposes to a \1 \
ferritic structure in 'which carbidp.s precipitate preferentially at Time -
the grain boundaries; in view of the relatively coarse grain size of
the as-welded structures this leads to a rather brittle structure.
10 I Poor (bl Good

Thus, if the welding operation is terminated at a point close to Fig. 5.27. Thermal cycles used for PWHT after welding
a martensitic steel; (a) Poor, and (b) Good.
the Ms temperature and the assembly is immediately placed in the
tempering furnace, the optimum mechanical properties are not After these treatments, the mechanical characteristics of the
achieved, even though the procedure is completely reliable from the welds were as snown in Table 5.12.
point of view of avoiding cracking. The best properties are obtained Table 5.12. Mechanic~ll Properties achieved
either by subjecting the assembly to a complete quenching and by two diffe.rent PWHTs.
tempering treatment when this is possible, or by carrying out the PWHT Tensile Strength, UF Impact. Value
heat treatment as illustrated by the followingexample. (N/mm2) (J/100 mm~)
Consider a martensitic creep-resisting steel having a composition Weld A (tempering carried out on 650 < 10
of 12% Cr, 0.2% C, 1%Mo, 0.4%W, and 0.3% V.After oil quenching austenite)
from l050°C and tempering at 750°C for 4 hours, this steel had a Weld B (tempering csrried out on 750 55
tensile strength of 780 N/mm2 an": a UFOimpact value of 60 J/mm2. martensite)
• The Mesnager U-notch. test specimen has a variC'.ble notch depth and in The microstructure of the weld metals and HAZsrevealed that
France the test results obtained are increased by a suitable factor so that they are the structure of weld A was ferritic with agglomerated carbides,
applicable to across-sectional area of 100 mm2 at the notch root.
-------~--~~~~~~~~~~"""~,~-=~."
..,_
.._----
Preheat and Postweld Heat Treatment 399
398 Welding Engineering and Technology
<:l '"
.;:so I
-
C,.) U
0 0
C,.)
0 ~ 8 .$.~ _.~ ° Cd~
v.C,.)
whereas weld B had the characteristic structure of tempered
martensite. The slightly lower properties of weld B compared with
~ ..•.•.
..•..
I ~.§2~
.....
'" '"
r...----
;;;~C,.) Cl;S
l!')
t::<-
liQ. '1' '1' '1'
Uj
<- <-
.0 c.o .=
'"0
.
..0 ""
'"
'" l.Q
•
'" <-
'1'
t::
.5l '1'
<-
i' i' .~
6 6
'1'
-lO
..0<- t::<-
.5l
·C,.)I
..• <l?,
::l~;l-~ l.Q l.Q l.Q ~<.) .=:~ 'b.tl';) "'lO 0.10
00> 0> 0> ·S 0 l.t:)
0""l.Q Vo> 00>
parent material were attributed to the presence of delta ferrite .•..••••. '0
~~~Q:::
<:l lO l.Q 10
g'<O
'"00>
l.Q '"OlO lO

phase in the weld metal of weld B. This difference could be •••<-

eliminated if a complete heat treatment were applied, as

9 v 11
quenching from I050°C prior to tempering considerably reduces '" -c
'" -Sl
~ 15 t::
v
the quantity of delta ferrite. <:l

e-
.g
~
~ 0
'" ~
'" ..0
ro C,.)
~ 0
<.)
0
<.)
0 °10
<.) <.) C,)
°
C,.) S <.) u
o
Thus, PWHT followed for weld B was the desired one but it
'"
........... t:: <'1 0')
l.Q ~0 o o ~ 10 a
o :b
0 l.Q

-~~ "1j CI.J~ "" 0> <0 .....

~C,.)
f
1::)

.o
""
:.01
L...

requires the knowledge of Ms temperature of the steel under ell "tj0

~ ~~~.•... '" S ~
C'1

ro o
consideration. However, with massive components, it necessitates .•.. ~
<:l

..•.••:j
~

t::

careful control of the temperature of the assembly, because the' 00

;..
..g ~ ~
.- ~ ~
risk of cracking increases when the temperature to which the weld ~ "tj ~ ~ ~ v v
"3
~ '"
~ E: ~ i5 i5
is allowed to cool decreases: Thus, it is advisable to perform a Q. <:3 E:~ ~~
l>o ::l {j"g ~
o C,.) <.) ~ <.) <.) ~Q o C,) C,)
000 C,) C,.)
~
o
-~ .••"" 00
0 o 0
procedural test to define conditions ensuring reliability from the
i
£ .S E:
~.t:'-' ~t:: •••••••••
"" 0 ~ "" lO 000
cracking point of view. Q
"tj .-~
::l~ .S: I .2l I <0
t:: ••••• ~ I' "<I' ..-I ~ ~ "<I'

.... ~ " '" ero

"S
ro '"
5.6.1.3. PWHT of Austenitic Stainless Steels '"
oS
"tj
~'"
>-.:l

Because of the low thermal conductivity, high coefficient of ~

.•.. ~
expansion, and good elevated temperature mechanical strength of ell ~ ~'"
o
i i ii~o
Cl
austenitic stainless steels, welds in thick material can be subjected
CJ:l
~ "' ....•.
o l/:>
lO
"" co:>
o 0 l.Q 0 l.Q
"ti ~ E: l.Q
I I .....
I
to high internal stresses in addition to the normal external ~ t:~~
.- ~ E: VI l.Q ":l
"" .....
\il 10
""
10
'" M
••••
10 VI
C'I
M
.....
stresses anticipated in service. However, owing to their excellent ~ ~
...

ductility, and their insensitivity to notches and stress 'tS

c::
concentration, these steels have remarkable property of adaption ell
.... en en
to applied stresses, and do not generally require any ell
C d
o
~ ~ .9
~
stress-relieving treatment after welding. .Cl
~
<:l ro
'-' lO ~<.l M
It should also be noted that the high temperature '"
Q:::
t;::
"" .0
""
l.Q "" 10 t;:: M . o
mechanical properties of these steels render the low
~
c.;
§~
""',,--
...
·u
v
C.
en
C'-!

VI
I
•....
.0
o ....• C'l VI
""

.0
VI
""

.0
V!
~ .-
8. I'
c:> t-
""

.0
VI
.... VI '" C'-!
temperature treatments which are applied to ordinary or low (5 t:: .0 0
-
t::
aQ ...
:£ :a
alloy steels ineffective. For stress-relieving to be carried out ~ ci:
~ ~
efficiently in a reasonable time, it must be carried out at a ,.Q

sufficiently high temperature, such as above 800°C, and this is ~ •....

•....

often impractical in practice. ~ o(l

.9 vC.
5.7. PREHEAT AND PWHT OF SOME SPECIFIC STEELS ti'-'
Although it is not possible to provide preheat and PWHT data
S'-' c ~
"" ...• C'I C'I 0 o
'"
o....
•.... •.... 00 00 .......
0> ""

-
C':>
o ~ "'i'
for all steels, not to speak of all engineering materials ; however t} OCI

<0
M
:::E <-
.....
"<I' ~ ~ "<I'
<0
00

minim urn preheat and interpass temperatures as well as the '"

'" ~ <:( ...l
.....
.!.
co:>
<:( < lO
<:( lO
<:( ~
<:( ~
PWHT temperature range for 10 widely used steels are listed in ~ U? ::?l lO C,.) ~ ::E ::E ::E U? U?
U5 E-< •...• E-< E-< u.l
E-< E-< U5
;$:
Table 5.13 for specific carbon and thickness ranges. < ~ ~ "" ~ ~ ~ ~ < <
~~ .... C"i M ~ .,.; <0 ~ 00 0; :=;
•....

SIX
Cracks in Welds
6.1. INTRODUCTION
Of all the things that can go wrong with welds, cracks are the
worst. That is why weldability assessment of a material is usually
based on its cracking tende?cy during and after welding.
Cracks occur because the stress at that point in the weldment
ex.ceeds the ultimate tensile strength or ultimate shear strength
of the base metal or weld metal. Cracks cannot occur where the
metal is under compression because it is being squeezed together
instead of being pulled apart. That is one of the reasons why
peening the weld bead and the HAZ (heat affected zone) of a
weldment reduces surface cracking; as peening the weld puts the
surface of the metal under compression. However, peening is not
always allowed by welding codes.
Tensile Cracks are by far the most common. They can occur
either in the weld metal or the base metal. Shear cracks, when
they occur, often occur at an angle to tensile cracks. They may also.
appear in line running off at an angle from the direction of stress.
After the weld metal has cooled, cracks are most likely to occur if
the weld metal or base metal is hard and brittle. In other words,
it has low ductility. That is why steels with more than 0.3 to 0.35
percent carbon in them are very difficult to weld. These
medium-to-high carbon steels have high strength but low ductility.
Once a crack starts, it will continue right through the weld
and the base metal until the stress that causes the cracking is
relieved. Most often, instead of relieving the stress, the crack
makes the section of the structural smaller so that the stress per
unit area becomes even larger. That is why a crack once started
can continue to extend right through the weldment. Ships have
been known to split in two because of a single crack which looked
insignificant at first sight. A crack in a high pressure pipeline can
Cracks in Welds 401
literally race at the speed of sound for several kilometers from one
pumping station to another. Cracks must therefore be never
ignored, even if they are tiny ones.
In multilayer welds, cracking occurs most frequently in the
first layer of the weld metal in the root zone of the weld joint. That
is why so much attention is paid to making a good, solid first pass
in a large joint. If cracks are found in the weldment by using a
non-destructive testing method, the weld metal will have to be
gouged or ground out and the weld remade. One method of
avoiding or minimising cracking in and around welds is to heat
the parent metal before welding. This reduces the tendency to
form martensite in steels during cooling, and in the case of
multipass welding it provides more extensive allotropic
refinement of beads by the normalising' effect of heat of the
succeeding passes.
6.2. CLASSIFICATION OF WELD CRACKS
Cracks may occur in the weld metal or in the HAZ. Sometimes
they originate in one zone and spread into the other. These cracks
are grouped into two categories, viz.,
(i) Hot Cracks, and
<ii) Cold Cracks.
6.2.1. Hot Cracks
Hot cracks occur mainly in the weld bead but sometimes they
may develop in the HAZ. When located in the weld metal they are
referred to as Solidification Cracks while in HAZ they are called
Liquation Cracks. Solidification cracks occur in the weld metal
when the metal is very hot; just below the solidus temperature of
the metal. Such cracks are often interdendritic following the
random path of grain boundaries along segregated concentrations
of impurities deposited there by the solidifying weld metal. A hot
crack will show temper colours on its inner surfaces. This is due
to the oxide films formed there after the initiation of the hot crack.
Hot cracks may also be intercrystalline, following lattice layers
within the grains.
6.2.1.~..Factors Which May Promote Hot Cracking
Some of the important factors which promote solidi.~ication
cracking in weldments include the following.
(i) Solidification structure,
(ii) Segregation,
(iii) High Stress in Weld Metal,
(iv) Material thickness,
 Cracks in Welds 403
I
!/
402 Welding Engineering and Technology

(v) Joint restraint, T

(vi) Width/Depth ratio,
(vii) Weld bead shape,
(viii) Material Composition; High Carbon and Nickel Con-
tents,
(ix) Crack Sensitivity of the electrode,
(x) High Current,
(xi) Excessive dilution of filler metal,
(xii) Preheating-increases liability to cracking,
(xiii) Weld procedure:
(a) high welding speed,
(b) long arc,
(xiv) Crater Cracking-indicates crack sensitivity,
(xv) Poor fit-up.
Other causes are lack of ferrite in austenitic stainless steels, Xo Xmax XE

and incorrect filler in aluminium alloys. X solute

Solidification Structure : Epitaxial solidification causes t~e Fig. 6.1. Non-equilibrium solidification of an alloy, Xo, in a simple binary system.
coarseness of the weld deposit microstructure to be inherited from
the grain growth zone of the HAZ..In general, the long str.aight iron are given in Table 6.1 which shows that the elements most
sided columnar grain structure obtained in high speed welds, likely to segregate in steel are S, 0, B, P, C, Ti, Nand H in that
tends to be weaker under stress than the more equiaxed, finer order. Of these elements, S is often considered to be the most
grain structure of the slow speed welds. The coarser the cell dangerous because it readily combines with iron and manganese
structure, within the columnar grains, the higher the segregatien to form (MnFe)S, a low melting point compound which easily
tendency. However, high speed welding tends to produce finer cell spreads along grain boundaries.
spacing than slow speed welding. Table 6.1. Values of Partitioning Coefficient, K,
Dendritic growth in weld metals, associated with the final of Elements in a-iron
stages of solidification, is the result of high constitutional super Element Al B C Cr Co Cu H Mn Mo
cooling, relatively slow cooling rates and hence larger segregation. K 0.92 0.05 0.13 0.95 0.90 0.56 0.32 0.84 0.80
Such a tendency is greatest at end-craters where constitutional
Element Ni N 0 P Si S Ti W V
super cooling is at its highest.
K 0.80 0.28 0.02 0.13 0.66 0.02 0.14 0.95 0.90
Segregation : Segregation is unavoidable in the solidification
of alloys. It occurs because of partitioning of elements during the As a general rule, slow solidifying welds with coarse
initial stages of solidification. The amount of segregation depends microstructure, such as obtained by electroslag welding, tend to
upon the partitioning coefficient (K), which is defined as, promote centreline segregation.
K'=Xs ...(6.1) High Residual Stresses : The solidifying metal should not be
XL allowed to become excessively stressed in tension otherwise the
crack will follow up behind the weld because of scissor action of
where, Xs and XL are defined by the phase diagram of Fig. 6.1. the expanding and contracting joint.
The greatest partitioning occurs for the smallest value of K. In the Material Thickness: Thicker the workpiece material faster the
case of steels, any alloying constituent in the weld deposit which
exhibits a wide freezing range of composition with iron is likely to cooling after welding and more the chance of formation of harder
have a low value of K. Approximate values of K for some alloys of constituents thus greater the tendency to cracking.
 Cracks in Welds 405

404 Welding Engineering and Technology
Restraint: If high tensile stresses are induced in the weld due
to restraint from jigging or clamping, such that the hot strength
of the weld metal is exceeded then it may lead to the development
of cracks.
If metal-to-metal contact exists between thick plates prior to
welding, the plates cannot move--the joint is restrained. As the
weld cools and contracts, all shrinkage stresses must be taken up
in the weld, as shown in Fig. 6.2 (a). This restraint may cause the
weld to crack, especially in the first pass on the second side of the
plate.
Soft
WirlZ
(al (bl Ic I
Fig. 6.2. Restrained fillet weld in thick plates, (a) all shrinkage stresses taken up
by weld, (b) placement of soft wires between horizontal and vertical members,
and (c) wires flattened and the stress in the weld is released ..
Joint restraint can be minimised by providing a space of 0.8
susceptible to cracking because of the notch, as shown in Fig 6.3
(a) which is a crack initiator. On high quality work, this notch is
back-chipped, as shown in Fig 6.3 (b)., to,
(i) Remove slag or oxides from the bottom of the groove,
(ii) Remove any small cracks that may have developed in
the root bead, and
(iii) Widen the groove at the bottom so that the first bead of
the second side is large enough to resist the shrinkage
that it must withstand due to the rigidity of the joint.
The weld metal tends to shrink in all directions as it cools,
and the restraint from the heavy plates produces tensile stresses
within the weld. The metal yields plastically while hot to
accommodate the stresses; if the internal stresses exceed the
strength of the weld, it cracks, usually along the centreline.
The problem is greater if the plate material has higher carbon
content than the welding electrode as in that case the weld metal
usually picks up additional carbon through admixture with the
base metal. Under such conditions, the root bead is usually less
ductile than the subsequent beads.
Weld Bead Shape : A concave fillet weld may appear to be
larger than a convex weld, but it may have less penetration into
the welded plates and a smaller throat than the convex bead, as
shown in Fig. 6.4. Thus the convex weld may be stronger of the
two, even though it appears to be smaller.

to 1.6 mm
cooling. between
Such spaceStheor two
gapsmembers
can be to allow movement
incorporated during
by providing
rough flame cut edges on the plate or by placing soft steel wire
spaces between the plates as shown in Fig. 6.2 (b). The wire
flattens out as the weld shrinks, as shown in Fig. 6.2 (c). Copper
wire should, however, be not used because that may contaminate
the weld metal.
Although provision for a space between thick plates to be
welded is particularly important for fillet welds, root pass of a butt 101 Concave weld lbl Convex weld
weld in heavy plates usually requires special welding procedure. Fig. 6.4. Comparison of throat size in concave and convex fillet welds.
For example, the root pass on the first side of a double-V joint is In the past, the concave weld was preferred by designers
because of the smoother stress flow it offers to resist a load on the

Xl i~)(· t
joint. Experience has shown, however, that single pass concave
Ba<k-<h;pped ;;
fillet welds have a greater tendency to crack during cooling than
t Notch
t do convex welds. This disadvantage usually outweighs the effect
of improved stress distribution, especially in steels that require
special welding procedure ...
(a I Notch effect ( b I Back - chipped weld

Fig. 6.3. Double Vee butt joint, (a) notch effect on root run, and (b) back-chipping
.- n •• ~;..l nnt.r.h Affect and crack development.
, 406 Welding Engineering and Technology Cracks in Welds 407

When a concave bead cools and shrinks, the outer surface is
in tensi.on and may crack. A convex bead has considerably reduced
shrinkage stresses in the surface area, and the possibility of
cracking during cooling is slight. For multi-pass fillet welds only
the first pass need be convex.
If the design conditions required concave welds for smooth
flow of stresses in thick plates, the first bead (usually three or four
passes are require'd) should be slightly' convex. The others are then
built-up to the required shape.
Width-to-Depth Ratio: The width to depth ratio of a weld bead
should not be excessive, i.e. WID be not less than 2/3. In case the
depth of the weld bead is much greater than the weld width, the
surface of the weld may freeze before the centre does. When this
happens, the shrinkage forces act on the almost frozen centre (the
strength of which is lower than that of the frozen surface) ~nd can
cause a crack that does not extend to the surface, as shown in Fig.
6.7.
17

1f
Crack ~wldth!zJ. DWldth

,. ,'T ..
;yJ"''' p y

Corract
~
lh
Inco rract ~ I L
Wild depth,Weld width Weld depth/Wild width Corra!=t
Incorrect Correct (01 (b I
(al Too concave (blFlator slightly convex
Fig. 6.5. Root bead in a single Vee butt weld, (a) too concave, ""Width
and (b) flat or sliglttly convex bead runs. '" /~ of
A rc gouge Intlrnnl
'" fn~
A concave root bead in a butt weld, as shown in Fig. 6.5 (a)
has the same tendency toward cracking as it does in a fillet weld.
Increasing the throat dimension of th~ root pass, as shown in Fig.
6.5 (b), helps to prevent cracking. Electrodes and procedures
" '/:< ,'v LA-J
-r-l/r-l-
If = =
,"oocrnCk t" ••••-z.- I
"",.;>

>-~D:r

y
fusion
should be used that produce a convex bead shape. A low hydrogen
pr.ocess usually reduces cracking tendencies; if not, preheating ld I
may be required.
Centreline cracking can also occur in subsequent passes of a
multipass weld if the passes are excessively wide or concave. This
can be corrected by putting down narrower, slightly convex beads,
making the weld two or more beads wide, as shown in Fig. 6.6. + YCorract
lel

Ie)
Incorract
r[dT

~fWl Incorrect
Too wide and concave
Incorrect

Washed up too high Flator slightly convex,

Fig. 6.7. Effects of weld width-to-penetration ratio of weld bead and gouged
grooves on development of internal cracks, (a) partial penetration butt welds
with excessive penetration, (b, e) fillet welds, (d) arc gouged grooves, and
(e) partial penetration butt welds with wide weld beads.

Thus, while cracks caused by too high joint restraint or improper

material composition usually appear at the face of the weld, the
(Also, poor slog removal) and concave less than full Width
(Also,good slag removal)
internal cracks that do not reach the surface are usually caused by
improper joint design (narrow, deep groovebutt or fillet welds) or by
(a I ( bl (e) misuse of a welding process that can achieve deep penetration.
Fig. 6.6. Bead configuration in multi-run butt welds. (a & b) incorrect Internal cracks can also be caused by improper joint design or
concave filler passes, and (e) correct filler runs. edge preparation. Consequences of combining thick plate, a deep-
408 Welding Engineering and Technology
penetrating process, and a 45° included angle are shown in Fig.
6.7 (b). More details of a completed fillet weld made with deep
penetration are shown in Fig. 6.7 (c). A too small bevel, and arc
gouging a groove too narrow for its depth on the second pass of a
aouble- V butt weld, can cause the internal crack, as shown in Fig
6.7, (d).
Internal cracks are serious because they cannot be detected
by visual inspection methods. But they can be eliminated if
preventive measures are taken. Penetration and volume of weld
metal deposited in each pass can be controlled by regulating
welding speed and current, and by using a joint design which
establishes reasonable depth of penetration requirements.
Recommended ratios of width of each individual bead to depth of
penetration are between 1.2 to 1 and 2 to 1.
A different type of intemal crack occurs in submerged arc
welding when the width to depth ratio of a weld bead is too large.
Cracks in these so called hat-shaped welds (Fig. 6.7 (e)) are
especially dangerous because radiographic inspection may not be
able to detect them. Thus, the width-to-depth ratio of any
individual bead should not exceed 2 : 1.
Material Composition: High carbon and nickel contents in
steels may lead to weld cracks due to high harden ability of such
steels.
Welding Procedure : High welding speeds and long arc length
. may result in the development of cracks.
Poor Fit-up or Inadequate Tacks: Reduced throat thickness
alongwith concave welds or lack of penetration, and poor fit-up,
as shown in Fig 6.8, are all prime causes of cracking.
Fig. 6.8. Poor fit-up in a fillet weld.
The mechanism by which solidification cracking occurs is one
in which cracks nucleate at carbide matrix interfaces and spread
along the boundary, under the influence of the tens,ile thermal
stresses. Thus, the role of thermal stresses is very important in
solidification cracking. As a general rule, end cracking is
Cracks in Welds . 409
associated with high transverse stresses , while centreline
cracking behind the weld probably results from high longitudinal
stresses.
6.2.1.2. Liquation Cracking (Burning)
Liquation cracking of HAZ occurs mainly when a process,
having a high heat input such as electroslag or submerged arc
welding, is used to weld steels having a high ratio of sulphur and
phosphorous with a low manganese content.
During the manufacture of hot rolled plate, the sulphur and
phosphorous collect around the grain boundaries as low melting
point eutectics, e.g. sulphide eutectic. The prblonged heat from the
welding causes the eutectic to become molten or semi-molten and,
upon cooling, forms as brittle films around the grains in the HAZ.
The stresses induced by the contracting material result in
micro-cracking at the grain boundaries, which reduces. the
fusion
boundary
zone
------1---
!
Fig. 6.9. Schematic representation of liquation cracking
in the HAZ of a fillet weld.
strength and ductility of this zone. Fig. 6.9 shows schematic
illustration of liquation cracking in the HAZ of a weld.
Liquation cracking is dependent on the amount and type of
impurities in the base metal, the volume percent and density of
inclusions, and the degree of restraint. The latter effect is in turn
dependent on the strength and the thickness of the plates being
welded. The dwell time at high temperatures is another important

410 Welding Engineering and Technology
factor affecting liquation cracking; high energy welding processes
thus increase the susceptibility to this problem.
6.2.2. Cold Cracking
C61d cracking or hydrogen induced cracking (HIC) or delayed
cracking, is perhaps the most serious and least understood of all
weld cracking problems particularly in the welding of high
strength steels.
HIC occurs when the following four factors are present
'Simultaneously.
(i) Hydrogen in the weld metal,
(ii) High stresses,
(iii) Susceptible microstructure, e.g. martensite, and
(iv) Relatively low temperature, between 200 and --100°C.
High stresses can be induced by the heat input and constraints
during welding. Martensite, especially hard and brittle high
carbon martensite, is susceptible to hydrogen cracking. As the
martensite formation temperature, Ms, is relatively low HIe tends
to occur at relatively low temperatures and hence the name cold
cracking.
Prior to crack initiation there is usually a time delay, or
incubation time, and crack propagation occurs in a slow, jerky
mode, with further incubation times between steps in the crack
growth. The crack can be both inter-and trans-crystalline in
character, but mainly follows prior austenite grain boundaries.
The initiation of clod cracking is particularly associated with
notches, such as the toe of the weld, or microstructural
in-homogenities which exhibit sudden changes in hardness, such
as slag inclusions. martensite-ferrite interfaces or even grain
boundaries. In common with most other crack growth phenomena,
cold cracking is accentuated in the presence of high restraint weld
geometries and matrix hardening.
The main source of hydrogen in welding is the water vapours
in the welding arc, but its extent depends upon the type of process
employed. Hydrogen in atomic forp:l can easily penetrate steel
causing cracking, blistering and loss of ductility. The atomic
hydrogen can diffuse through the steel lattice and recombine to
form molecular hydrogen in voids such as laminations or inclusion
sites to form blisters. Worse still, it can cause cracking in areas
having high hardness such as weld metal, heat affected zones, and
localised hard spots. Hardness of steel can lead to cracking; it is
recommended that the hardness should be limited to Rc 22
\
Cracks in Welds 411
(Rockwell C-Scale) to preclude sulphide stress cracking often
encountered in petrochemical plants.
Components failed by cold cracking sometimes show a
phenomenon called fisheyes. These are interpreted as circular
areas of quasi-cleavage fracture centred on a large inclusion (e.g. I
MnS) or clusters of small inclusions e.g. silicates or AlzOa particles.
Apart from water vapours, the hydrogen can come from a I
variety of sources viz., corrosion reactions, chemical processing
reactions, purposeful additions (to MIG shielding gas or plasma
gas, etc.) or from the thermal decompos.ition of SMAW coated
electrodes of the cellulosic type. The best solution is to bakeout I
the hydrogen before welding commences, as well as use preheat
and PWHT to reduce hardness and residual stresses ofweldments.
As a rule it should be recognised that no steel is immune to HIC. i
To sum up it can be said that the factors which may promote
cold cracking in weldments include the following.
(i) Hydrogen in the weld metal and HAZ,
(ii) Joint restraint and high thermal severity,
(iii) Presence of impurities,
(iv) Susceptible microstructure of HAZ (e.g. martensite)
(v) High welding speeds,
(vi) Low current density, and
(vii) Weld of insufficient cross-sectional area.
6.3. NOMENCLATURE, LOCATION AND ORIENTATIONS
OF WELD CRACKS
Hot and cold cracks in the weld metal and HAZ are given
different names depending on their location and the orientation
in which they are placed. Most of these are covered by the
following nomenclature.
(i) Longitudinal Cracks,
(ii) Transverse Cracks,
(iii) Crater Cracks,
(iv) Underbead Cracks,
(v) Toe Cracks.
All these kinds of cracks are discussed briefly for the weld
metal and the HAZ (Base Metal) in this section.
6.3.1. Weld Metal Cracks
There are three types of cracks which may occur in the weld
metal. They are crater cracks, longitudinal cracks, and transverse
cracks.
 '-
'--....
'-"

412 Welding Engineering and Technology 'III

,L
Cracks in Welds 413
6.3.1.1. Crater Cracks /
Crater cracks can occur when weJding is interrupted. These I Longitudinal
or centre line
cracks are usually star shaped and progress only up to the edge crack
of the crater, as shown in Fig. 6.10. Such cracks are also called \
Bascz
mcztal

Joint

Cratczr with s~r-shapczd cracks insidcz it

Fig. 6.10. Star-shaped crater cracks often occur at the
points where the electrode is changed.
Paw Cracks for their resemblance to bird's paw. The crater cracks
can be the starting point for longitudinal cracks, particularly when
they occur in the crater formed at the end of the weld as shown
in Fig. 6.11.
Fig. 6.13. Longitudinal crack in a fillet weld.
crater crack formed at the end of the weld as shown in Fig. 6.13
for a fillet weld. Longitudinal crack at the surface of the weld
metal can also be an extension of a crack that started deep down
in the joint, probably in the root zone. That is why the root pass
must be made very carefully.
6.3.1.3. Transverse Weld Metal Cracks
Transverse weld metal cracks are cracks that run across the
weld perpendicular to weld axis, as shown in Fig. 6.14. This is a

end cracking
Fig. 6.11. Crater crack extending intd weld metal to become a longitudinal crack.
When a crater is formed somewhere else in the weld, for
example at the electrode change point in SMAW, it is usually
welded up when the welding is restarted, but not always.
Sometimes fine, star-shaped cracks are seen at various locations
in a weld. Crater cracks are frequently found in weld metal that
ha;:; a high coefficient of thermal expansion; the prime example is
austenitic stainless steel.
Crater cracks can be prevented by starting each weld properly,
correctly restarting at each electrode change point, and by filling
craters by making them to slightly convex shape before breaking
the arc.
6.3.1.2. Longitudinal Weld Metal Cracks
Longitudinal weld metal crack usually running through the
centre of the weld, as shown in Fig. 6.12, often is an extension of
CCC7:cr::E\
Fig. 6.14. Transverse weld metal cracks running across the weld bead.
common kind of cracks, and in some cases such a crack extends
beyond the weld metal and into the base metal. Transverse cracks
commonly occur in weld joints that are very tightly restricted so
that the strl1ctural section or the weld metal cannqt shift to relieve
the built-up stresses. These are often shrinkage 'cracks produced
when the weld metal is cooling and contracts.
6.3.2. Base Metal Cracks
Cracks that occur in base metal are usually located in the HAZ
and run parallel or perpendicular to the weld bead. They almost
always occur because the HAZ has become hard and brittle
possibly because the required preheat and PWHT were not given.
6.3.2.1. Longitudinal Base Metal Cracks

CCCELCE"OI
Fig. 6.12. Longitudinal weld metal crack running through the weld bead.
Longitudinal base metal cracks run parallel to the weld but
lie in base metal as shown in Fig. 6.15 and may be formed as
extensions of bond-line cracks due to poor weld metal penetration
 I
\

414 Cracks in Welds 415 \

Welding Engineering and Technology \

until the weld metal has cooled to room temperature. NDT (non
BaSil mlltal ----- --..------_ destructive testing) methods may be needed to locate them.
6.3.2.3. Underbead and Toe Cracks

CCCCCCEm"Q'
--
Basil mlltal
--
Fig. 6.15. Longitudinal base metal cracks are located in the HAZ
Surface cracks in the base metal, under or near the weld, are
known as underbead cracks. Similar cracks that appear on the
plate surface adjacent to the weld are called toe cracks. Both these
types of cracks, for a fillet weld, are shown in Fig. 6.17.

and run parallel to the weld bead.

TOil crack
into the sides of the butt joint. For fillet welds, longitudinal base
metal cracks can be divided into two groups viz.,
(i) those that start at the toe of the weld, and
(ii) root cracks which proceed from the root of the fillet weld
into the base metal and sometimes even go out the other Undllrbllad crack
side of the joint. Fig. 6.17. Underbead and toe cracks in a fillet weld.
In the case of butt welds, cracks in base metal are more likely Underbead cracking in the HAZ is caused by,
to show up in the HAZ or at the very edge of the weld in the fusion (i) a relatively high carbon or alloy content steel that is
boundary zone between the weld metal and the HAZ. Cracks in allowed to cool too rapidly from the welding tempera-
the HAZ or the weld mehil are associated with high hardenability ture, and
steels when the filler metal and the b~se metal are of entirely (ii) hydrogen pick-up during welding.
different compositions. The combination of different alloys
sometimes can produce unpredictable results in the weld metal, Underbead or toe cracking seldom occurs with the preferred
even though that may be the only way to weld a high hardenability analysis steel given'in table 6.2.
steel.. One method of overcoming this problem is to make a Table 6.2. Preferred Analysis for Steels to be Arc Welded
transition joint by buttering up the joint face with a very ductile S.No. Element
electrode and then using a different electrode to finish the weld. Composition, %
6.3.2.2. Transverse Base-Metal Cracks
Transverse base-metal cracks develop perpendicular to weld
axis and are placed in the HAZ as shown in Fig. 6.16. They are
usually associated with welding of high hardenability steels e.g.
tool steels and certain alloy steels. These cracks may not be seen
BaSil mlltal
\

CCCCCCEQI ~ ~ ? ) BaSil mlltal

Fig. 6.16. Transverse base metal cracks-most often

occur in lo!.igherstren'gth steels.

416 Welding Engineering and Technology
from there into the extremely hot HAZ. If cooling occurs slowly
the process reverses and the hydrogen has sufficient time to
escape through the weld into the air. But if cooling is rapid, some
hydrogen may be trapped in the HAZ next to the weld metal, as
Most of the hydrogen escapes
through the weld intc the air
Adjacent plate is transformed
to austenite when heated
by welding; hydrogen is
soluble in this region
This region remains as
fe r rite whi ch h as no solubility
Difficult for hydrogen for hydrogen
to diffuse any farther
Fig. 6.18. Schematic representation of movements of
hydrogen during arc welding.
shown schematically in Fig. 6.18. The hydrofen is absorbed and
produces a condition of low ductility known as H2-embrittlement.
Slower cooling, achieved by slow welding speed or by
preheating, allows more of the hydrogen to escape and helps
control the problem. Also, the use of low-hydrogen welding
electrodes and fluxes eliminates the major source of hydrogen and
usually eliminates underbead cracking
Rapid cooling rates occur wp.en the arc strikes on a cold
plate-at the start of the weld with no previous weld bead to
preheat the metal. The highest cooling rates occur on thick plate
and in short tack welds. The effect of weld length on cooling rate
can be judged from the fact that a 225 mm long weld made on a
plate at 20°C has about the same cooling rate as a 75 ,mIl! long
weld on a plate that has been preheated to 150°C..
Welds with large cross-sections require greater heat input
than smaller ones. High welding current and slow welding speeds
reduce the rate of cooling and decrease the likelihood of underbead
and toe cracking.
6.4. FACTORS CONTRIBUTING TO WELD CRACKING
From the aforementioned description of cracks in welds it is
evident that the major contributing factors to cracking include the
following.
1. The contraction forces set up in the weld metal on cool-
ing tend to cause cracking in the weld metal and the
base metal. These forces generally increase with the
Cracks in Welds 417
increase in strength and/or hardenability of the filler
metal and the base metal.
2. The susceptibility to cold cracking is proportional to the
H2 content of the welding atmosphere.
3. Greater crack sensitivity is exhibited by high-alloy base
metal and by heavier plate thicknesses.
4. In general cracking initiates in the HAZ of the base
metal, except when the weld metal is of higher hardness
than that of the HAZ.
5. With an open arc and hot humid weather the arc atmos-
phere will contain more hydrogen as water vapour than
in cool, dry weather. Any tendency to mini mise the im-
portance of preheat or possibly of PWHT in hot summer
months could lead to the d(~velopmentof cracks in heavi- i
ly restrained joints. This will be especially true if either
the weld metal or the base metal is hardenable because
of alloy or carbon content.
6. Low heat input with interruption in the welding cycle
tends to aggravate the cracking problem.
7. The welding position and its influence on bead size, heat
input, number of passes, etc. has a direct influence on
the cracking tendency. For example, three o'clockgroove
butt welds are more sensitive to cracking than the
downhand or flat position groove butt welds.
6.5. SPECIFIC WELD CRACKS
Although all types of weld cracks are covered by Hot and Cold
Cracks but some of these are encountered in specific c;onditions or
environment and are associated with some definite'industries or
service conditions. Some well known among the'se include the
following.
(i) Chevron Cracks,
(ii) Lamellar Cracks,
(iii) Reheat Cracks,
(iv) Stress Corrosion Cracks.
Chevron cracks are weld metal cracks, Lamellar and Reheat
cracks are HAZ cracks while Stress Corrosion Cracks may cover
both the weld metal and HAZ. Special characteristics of all these
types of cracks are discussed in this section.
6.5.1. Chevron Cracks
These are intergranular and transgranular cracks which afe
usually associated with basic SAWfluxes and sometimes also with
418 Welding Engineering and Technology
Cracks in Welds 419
basic coated electrodes. These cracks are also known as Staircase
Cracks or 45° Cracks as they occur at 45° to the weld axis.
Chevron Cracks are hot cracks which are caused when the
weld metal is subjected to Ductility Dip Cracking because steel
may lose ductility between 1200°C and 1000°Cwhich is recovered
at about 800°C. This loss of ductility is due to segregation of
sulphur and phosphorous to grain boundaries; this causes
intergranular cracks. The intergranular surfaces (which are
intergranular relative to prior austenite grains) of these cracks
show thermal facets (as shown by oxide colours), indicating that
they have been exposed to high temperature. These intergranular
cracks are joined by transgranular cracks at low temperature by
hydrogen cracking to produce Chevron cracks as shown in Fig.
6.19.
CHEVRON CRACKS

Fig. 6.20. Lamellar tearing adjacent to a tee-butt weld in a structural steel.

Fig. 6.19. Longitudinal section of a submerged arc weld showing Chevron cracks.
Another theory is that cracking is all due to hydrogen
embrittlement but takes place at two different temperature
ranges.
Chevron cracking is most common in welds in relatively thick
sectioned C-Mn steel plate, welded w:ithhigh heat input processes.
Nevertheless, such cracking is not confined to high heat input
welds as they have also been observed in self-shielded cored wire
welds. Although such cracks have been observed in two-pass
submerged arc welds in 40 mm thick plate but they are more usual
in multi-pass welds.
Chevron cracking in C-Mn steel weld metal can largely be
avoided by using very low hydrogen levels i.e. ::; 5 ml/l00g of
deposited metal.
6.5.2. Lamellar Cracks
Lamellar tearing or a lamellar crack, as shown in Fig. 6.20,
occurs in HAZ due to the combination of localised stress and low
ductility of metal in through thickness direction due to the
presence of 81(mgatedstringers of non metallic inclusions parallel
to its rolling direction. Crack initiation occurs at the flattened MnS
or oxide-silicate stringers by the decohesion of such non-metallic
inclusions near the weld, in or just outside the HAZ.
In lamellar tearing the crack appears to be closely associated
with the edge of the HAZ. The horizontal and vertical cracking of
the base metal is a very typical feat~re of lamellar tearing. The
problem occurs particularly when making tee and corner joints in
thick plates, such that the fusion boundary of the weld runs
parallel to the plate surface.
In general susceptibility to lamellar tearing depends upon:
(i) type of joint,
(ii) inherent restraint,
(iii) sulphur and oxygen contents of weldment in HAZ,
(iv) type and morphology of inclusions in the base metal,
(v) hydrogen content of weld,
(vi) the use of thick plates,
(vii) preheating, and .
(viii) the strength of the material of electrode used.
The effect of H2 appears to be associated with embrittlement
rather than with cold cracking. Preheating reduces the
susceptibility to lamellar tearing, especially when H2 potentials
are present. Electrodes of lower strength levels, within design
requirements, provide the weld metal the ability to accommodate
420 Welding Engineering and Technology
Cracks in Welds
---.-- -4U
..

Use steels t.reated with Rare Earth metals (REM) or

(ii)
for contraction strains and therefore improve the resistance to cerium,
lamellar tearing. (iii) Use materials and procedures to get reduced H2-content
Initiation oflamellar tearing may be triggered by the following in weld metal,
factors. (iv) Improvement in design as suggested for some joints i.n
(i) Separation or void formation at the interface between Fig. 6.22. A practical way of avoiding lamellar tearing
inclusions and metal,
(ii) Shattering of inclusion itself,
(iii) Sub-microscopic inclusions playa part in crack genera- (0)
·tion, ~ ~
(iv) Silicate and sulphide inclusions accentuate the situa-
~
Poor Good Better
tion.
The lamellar cracks observed in thick C-Mn steel butt welded
tD thick sections by SAW process are usually located in central
section of the plate or close to HAZ. (b)
Lamellar tearing may also be due to banding i.e. by
segregation of carbon in bands, as shown in Fig. 6.21, in the
Poor Good Better

")~
~
I Buttering
I -
Poor Good

Id I

Poor Good

Fig. 6.22. Joint design improvements to avoid lamellar tearing;

Fig. 6.21. Microstructure of a steel susceptible to lamellar tearing .
dirp.ction of rolling with cracks appearing in ~igh carbon content
zones.
Remedies The remedies suggested to avoid.or atleast reduce
the occurrences of lamellar tearing include the following.
(i) Use steels with low sulphur content i.e. less than
0.007%.
(~) and (b) replace fillets with solid weld metal or forged sections,
. (c) buttering, (d) modified corner joint.
is to employ joint designs that allow the contraction
stresses to act in the rolling direction, rather than the
transverse (through-thickness) direction of the lamellar
tearing susceptible material. Fig. 6.23 shows the use of
such designs.
422 Welding Engineering and Technology
~'l
Fig. 6.23. Improper and proper designs to avoid lamellar tearing in aT-joint.
Industries seriously affected by lamellar tearing include the
following.
(i) Machine tool industry-tee and corner joints of beds and
frames in heavy plates,
(ii) Off-shore platforms,
(iii) Welded-on attachments to boilers and thick-walled
pressure vessels.
6.5.3. Reheat Cracking
Reheat or PWHT cracking is one of the rarest types of cracking
that occur during welding operation. It appears to be restricted to
steels or weld metals containing at least two of the elements Cr,
Mo, V, and B; and it is mainly found in relatively heavy sections.
Reheat cracking is, however, also possible in coarse grained weld
metal and parent steel.
Reheat cracking is a typical problem in corrosion- and. heat-
resisting low alloy steels, i.e. ferritic steels containing Cr, Mo, and
sometimes Vanadium which are used for corrosion resistance and
strength at elevated temperature service for nuclear and fossil
energy applications. Typical examples of susceptible alloys are
1 1 1 1
.' "2Cr"2 Mo 4 V and "2CrMoV steel. Nakamura and Ito have
proposed the following relationships for the effects of alloying
elements on the Reheat Crack Susceptibility (RCS).
(Nakamura) RCS = %Cr + 3.3 (%Mo)+ 8.1 (%V)- 2 ...(6.1)
and (Ito) RCS = %Cr + %Cu + 2 (%Mo)+ 10 (%V)+ 7 (%Nb)
+ 5 (%Ti)- 2 ...(6.2)
when RCS ~ 0 the steel may be susceptible to reheat cracking.
When postweld stress relieving is recommended, the welded
structure is heated to a high enough temperature, for example,
550-6500C that the elastic stresses are rela.xed quickly. SinceCr, .
Cracks in Welds 423
Mo and V are strong carbide formers, carbide precipitation can
occur in corrosion and heat-resistant low alloy steels during
heating to stress-relief temperature or during stress-relief itself.
In these materials, carbides precipitate within the HAZ grains but .
not at the grain boundaries, thus making the grain interior
relatively stronger than the grain boundary. Consequently, when
the residual stresses relax, deformation and rupture are along the
grain boundaries;· resulting in intergranular cracking. Cracks
occur along the pri~r austenite grain boundaries in the HAz in
transformable steels, and along any of the grain boundaries in
austenitic steels. In some welds, reheat cracks can also appear at
coarse, columnar grain boundaries in the weld deposit.
The initiation and growth of reheat cracks take place in the
following steps.
(i) Nucleation cavities on grain facets and triple junctions.
(ii) Growth ofthese individual cavities to form cracks of one
grain facet in length.
(iii) Interlinking of these single facet cracks to form cracks
several grain diameters long which then rapidly lead to
fracture.
In general the important factors that contribute to reheat
cracking are:
(i) A structure having poor ductility (e.g. upper bainite) is
more susceptible to elevated temperature embrittle-
ment,
(ii) Coarse grain is more susceptible to reheat cracking,
(iii) High heat input processes like SAW and ESW result in
more susceptible weldments,
(iv) Reheat cracking usually affects thick sections (> 50mm)
where residual stress is high as cracks form above 400°C.
(v) High level of residual strain with some degree of tri-
axiality,
(vi) Temperature in the strain relaxation (creep) range,
(vii) Use of preheat can substantially increase grain size and
consequently susceptibility to reheat cracking.
Reheat cracking is not confined to the transformable steels;
this problem also occurs in the Ni-based alloys and austenitic
stainless steels. Among the low alloy steels the most susceptible
are the creep-resistant grades of the type Cr-Mo-V; 2.5 Cr-l Mo,
1
lCr-iMo. As a general rule, however, PWHT of the C-Mn steels

424 Welding Engineering
improves fracture toughn(~ss, although low impact toughness has
, been reported for reheated microalloyed steels, particularly when
welded by a high heat input process.
Amongst the low alloy steels, the 0.7 Ni-O.4 Cr-0.6 Mo steels
used, for example in nuclear pressure vessels, are well known to
be susceptible to reheat cracking. Cracks occur in grain growth
zones either as longitudinal macro-cracks, or more often as
microcrack network.
Design codes typically specify post weld heat treatments of
550-6500C for plate thicknesses above 38 mm and for periods of
an hour per 25 mm of thickness. Although these specifications are
based on isothermal relaxation tests but they closely correspond'
to reheat crack susceptibility temperature range.
Cr-Mo-V steel of the following composition,
0.12 C; 0.62 Mn; 0.45 Cr; 0.54 Mo; 0.06 Nb; 0.25 V with small
traces of P and S,
is used for turbine casings and pipework in high temperature
heavy power equipment because of its good creep resistance. This
is, however, extremely prone to reheat cracking.
Reheat cracking is also amongst the serious problems in
welding heat-treatable Ni-base alloys. These alloys are often given
PWHT to relieve stress and develop the maximum strength. In
order to develop its maximum strength, the weldment is
solutionized and then aged. During solutionization the residual
stresses in the weldment are also relieved. However, it may be
realised that ageing may occur in the weldment while it is being
heated up to the solutionization temperature, simply because the
ageing temperature range is lower than the solutionization
~ _----.--J"
Weldingpostweld heat treating
t
i
\~ -d~.
residual stress +
aging -crack solution
••
co..
e cracking.
~
residual
stress
._-----------'~
Fig. 6,24. Schematic representation of welding and PWHT cycle for a
heat-treatable Nickel-base alloy prone to reheat cracking.
Cracks in Welds 425
and Technology
temperature. Since this ageing action occurs before the residual
stresses are relieved, it is harmful and in fact often results in
cracking during PWHT. Such PWHT cracking is also called
Strain-age Cracking because the cracking occurs in high.ly
restrained weldments as they are heated through the temperature
range in which ageing occurs. Fig. 6.24 shows the welding and
PWHT cycle for a heat-treatable Ni-base alloy and the position on
time-temperature graph where reheat cracking may occur.
Reheat cracking in Ni-base alloys is a result of low ductility
in the HAZ accompanied by high strains in that region. It is
reported that stress relaxation can be smoother and more effective
if the base metal has been overaged before welding. This
advantage, however, does not exist in the HAZ of an overaged base
metal. Thus, it can be concluded that the HAZ of a weldment in.a
heat-treatable Ni -base alloy responds to the PWHT quite
differently than the base metal. The base metal tends to stress
relax during the PWHT without being intercepted by precipita.tion
reactions if it is overaged, rather than solution-annealed, before
welding. The stress relaxation of the weld HAZ during the PWHT.
however, is always interrupted by precipitation reactions of the
PWHT.
Reheat cracking in steels is avoided by,
(i) proper material selection by limiting vanadium in thick
sections to :5 0.1,
(ii) designing to minimise restraint; where restraint is un-
avoidable make PWHT after part welding,
(iii) when using higher preheat temperature employ lower
strength.weld metal.
Although reheat cracking occurs most frequently during
" PWHT, but it may also occur in service. Reheat cracking in service
is almost entirely confined to welds in! Cr ~ Mo ~ V steel and is
i
prevented either by using a different steel, such as 2 Cr 1 Mo,
or by controlling the composition of the alloy to minimise the
content of impurities that contribute to the high temperature
In Ni-base alloys Reheat Cracking is avoided by making use
...
TIme
of crack susceptibility C-curves. A crack susceptibility C-cu:rve is
a curve indicating the on-set of PWHT or strain-age cracking on
time-temperature graph. It is usually obtained by isothermal heat
treating of welded circle patches at different temperatures for
426 Cracks in Welds 427
Welding Engineering and Tecltnology

different periods of time and checking to see if the patches have heating. The compositions of waspaloy, Inconel 718 and Rene 41
cracked after heat treating. Such a curve usually resembles the are given in table 6.3.
shape of letter 'c' and, therefore, is called a crack susceptibility Table 6.3. Chemical Compositions of Different
C-curve. Since the ageing rate at the lower end of the ageing Heat-Treatable Ni-Base Superalloys
temperature range is relatively slow, the occurrence of reheat
cracking approaches asymptotically some lower temperature
No.
0.6
Rene
Ti
1.3
14-
3320
S. Waspaloy
0.9
3.0
1.5
10
Ai
Cr
Mo
Co
Fe 18
19=B
41
Zr==0.01
Others
Inconel718 18,
Alloy 0.1
0.1 = 5
0.05
0.07
Cb C Alloying Elements, %
range and the precipitates formed1.2. 3.
at the lower temperature are
dissolved, the occurrence of strain-age cracking approaches some
higher temperature limit asymptotically as shown for Waspaloy
°c
1050

OLL. 1800
WGspaloy

- ,
, -- .. 950
6.5.4. Stress Corrosion Cracking (SCC)
Although mixing of alloying ingredients added from electrode
Ql

5
.•..•.
o'-
1600
Reduced aging
Improved
rate
weldability
HA Z crack s
\' - .•.•..•.........•
850
coverings and fluxes is normally uniform, occasions are known
where alloying ingredients though dissolve in weld pool but due
~ 1400 No cracks 750
to fast cooling of weld insufficient time is available for complete
E mixing and diffusion. Mter cooling the sites of such particles
~
1200 I 650 become tiny regions more highly alloyed than the bulk of the weld
1 10' 102 103 104 105 106 metal, often with hard martinsitic microstructures -whichcan give
Time I minutes problems in service with stress corrosion cracking. In the case of
Fig. 6.25. Crack susceptibility C-curves for Waspaloy and Inconel 718. non-annealed welds, therefore, the maximum hardness is limited
and Inconel 718 in Fig 6.25. At any temperature between the upper to 240-260Hv. Such hard spots have been encountered in
and lower asymptotic limits, there exists a minimum time prior to submerged arc welds when Mn has been alloyed by the use of
ferro-manganese added to the flux. Presumably some patches of
which cracking is c"ertainto occur. Fig. 6.26 shows that no cracking Fe-Mn from the molten slag entered the weld pooljust as it started
solidifying. Occasionally, similar problems are also encountered
with SMAW electrodes and cored wires. Because the hard spots
result form a weld spot that had solidified too quickly for complete
dissolution and homogenisation of weld metal. They can usua]]y
-o
Ql
'-
:::J be avoided by increasing the weld pool temperature or its
L- solidification time by such measures as increasing the welding
a.
Ql
current, reducing welding speed or reducing the width of electrode
E weaving.
~
When carbon steel welded joints are exposed to severely
0·1 1.0 10·0 100·0 corrosive conditions (for example, warm sea water) there is a
Time I min. tendency for the attack to be concentrated upon the weld metal or
Fig. 6.26. Effect of heating rate on cracking tendency during PWHT of on HAZ, and the life of the weld may be a limiting factor. In the
Rene 41 solution-annealed for welding. duties for which carbon steel is generally employed the
occurs if the weldment is heated up rapidly to avcid intersecting preferential corrosion of weld metal is not a problem. Stress
corrosion cracking of the HAZ of the weld, however, will occur, if
the crack susceptibility C-curve. This approach is feasible when it welded joints that have not been stress relieved are exposed to
is possible to rapidly heat the weldment in a furnace or to have a aqueous solutions of hot concentrated caustic soda and the
structure that will not distort excessively owing to non-uniform weldroent is stressed in tension to a relatively high level. Stress
428 Welding Engineering and Technology Cracks in Welds 429

corrosion cracks in carbon steel are typically intergranular while

they are transgranular in austenitic stainless steel. For carbon
steel the most important stress-cracking agents are caustic soda
(sodium hydroxide) and ammonium nitrate, but coal gas liquors,
cyanides and possibly sulfates can act the same way. The limit of
100
Stress re lief requi red to avoid
caustic cracking of carbon steel
welded joints
'-
to- 4080
60
.::
OJ
E'- 20
Cl.
:::lJ
c;
O
QI
to I {bl

Stre ss re lief not requin!d Hot cracks

o I I I I I Shaft
o 10 20 30 40 50 60 70
Caustic soda concentration (°/0 by weight r
Flange section
Fig. 6.27. Caustic soda concentration vs. temperat.ure relationship to determine
the zone in which stress-relief of welded joints in carbon steel is required .
concentration and temperature of sodium hydroxide above which ./". Fatigue cracks
stress-relief of welded joints is required are indicated in Fig. 6.27.
Stress relief is achieved by heating the weld to a temperature Underbead crack
between 600°C and 650°C, holding for one hour per 25 mm of
thickness, with a minimum period of 30 minutes, and cooling at
a rate of not more than 300°C per hour. (c I ----.;;;.~
- Fractured end
In welding austenitic stainless steel severe residual stresses Fig. 6.28. Flange and shaft assembly, (a) before, and (b) after failure;
can be caused due to low thermal conductivity and high coefficient (c) the detailed view of cracked sections.
of thermal expansion. Upon exposure to chloride containing
corrosive media, sec may occur in the HAZ. Such cracking is 6.6.1. lVIaterial and Recommended Procedure
usually transgranular and branching. When stress-relief Steel ~ontainil1g 0.45% C, 0.17% Si ; and 0.73% Mn.
heat-treatment is carried out, the possibility of chromium carbide Recommended preheat and interpass temperatures 204°0, with
precipitation should be carefully considered. Stress relief can, of PWHT at 593-677°C .. Welding electrodes recommended for use
course, be achieved while solutionizing ASS, to dissolve chromium were of low hydrogen type.
carbide, at a temperature level of 1000°C. New residual stresses 6.6.2. Investigational Procedure Employed
may develop, however, during subsequent quenching. The use of
low carbon or stabilized grades can avoid such a problem by Allrpaint was removed and prepared
the area thoroughly .cleaned. FourA
allowing the material to cool slowly after solutionizing. cross-sections were cut and for micI\6inspection.
magnetic particle test showed a crack in the weld bead which on
6.6. CASE STUDY
inspection was found (,0 be a hot crack that had occurred during
This case study is concerned with different modes of cracking the solidification of the weld bead because the electrode could not
resulting in a catastrophic failure as a result of the failure of the tolerate the sulphur and phosphorous present in the steel: All
shaft and a guy-holding flange joint fillet weld shown in Fig. 6.28. cross-sections, examined under the microscope, showed hot cracks
ill both ~he fillet wel.ds indicating the use of incorrect electrodes.
430 Welding Engineering and Technology

Lack of preheat and rapid cooling were indicated after

hardness tests in the HAZ gave readings in excess of 400 VHN,
and martensite was observed in the microstructure of the shaft
material of both welds.
Underbead crack shown in Fig. 6.28 (c) indicates that this
hardenable steel had not been welded with a low H2 electrode
correctly dried to a low moisture content. The presence of H'2 and
the hard HAZ resulted in the development of underbead crack.
Fatigue cracks could be seen in the transition between weld
and shaft; the crack had spread about three quarters around the
shaft. Corrosion and rust on about two-fifths ofthe fatigue fracture
indicated a period of inactivity had allowed rust to form. The rest
of the fatigue fracture showed a matt or slightly shiny surface with
the characteristic annular ring followed by a coarse-grained
surface due to final fracture in a brittle manner.
6.6.3. Conclusions
From the above mentioned observations it appears that the
fatigue fracture of the shaft, the hot cracks, and underbead cold
cracks were the result of incorrect welding procedure.
1. The material was welded without preheating, which
should have been, approximately 200°C to minimise the
cooling rate and e{)I'l.seql.lently
the forming of hard zones;
shrinkage stresses and distortion. The lower cooling
rates also help hydrogen to diffuse out. Thus, slow cool-
ing rate, after welding, was not achieved.
2. Welding was not carried out with low H2 electrodes
otherwise underbead cracking would have been
prevented.
3. No stress relieving treatment was carried out to
eliminate 'hard spots' and refine the structure. PWHT
is essential for a fabrication such as the one under con-
sideration as it was subjected to flexing of the shaft, in
addition to the torsional stress on the flange.
(B) in the total weld bead <A+B) is termed as the %age by which filler metal has
been diluted,
SEVEN
Weldabilityand
Weldability Tests
7.1. INTRODUCTION
The relative suitability of a material for welding is expressed
by the term weldability. Although it can be said that almost an
metals are weldable by one process or the other but not with the
same ease or to the same level of quality. Thus the real criterion
for deciding the weldability of a material is weld quality and the
ease with which it can be obtained. According to AWS (American
Welding Society), Weldability is defined as 'the capacity of a
material or a combination of materials to be welded under
fabrication conditions into a specific, suitably designed structure,
and to perform satisfactorily in the intended service'. The better
the weldability, the easier it is to meet these requirements.
Weldability is difficult to define in quantitative terms as it
compasses the following aspects.
(i) Metallurgical compatibility of the material(s) being
welded,
(ii) Mech(mi~al soundness, and
(iii) Good serviceability of the welded joint.
(i) Metallurgical Com'patability implies that the base
metal(s) and the weld metal can be combined within the
degree of dilution* encountered in a specific process
without the production of undesirable constituents or
phases.
* Dilution in a weld bead containing both filler metal (A) and the base metal

i.e .. % Dilution = A~ B x 100.


432 Welding Engineering and Technology
(ii) Mechanical Soundness. The welded joint must meet
the soundness requirement.s and normal engineering
standards.
(iii) Serviceability Factor. Ability of the finished weld-
ment tQ meet special requirements, for example, low
temperature impact, high temperature stability, etc.
Weldability is almost always considered in relation only to one
particular process at a time. This is because of the complete'
impossibility of finding one simple index to weldability that can
be applied to all processes in all conditions. Welding response of
one particular material will vary widely in relation both to process
and to type of material. Thus a material that welds readily with
one process, say electric spot or seam welding when the material
is thin in section, may be unweldable by the same process when
the material is thick in section. Also, a material might be readily
weldable, in thin section, say with the help of oxy-acetylene flame,
but almost completely unweldable by means of SMAW i.e. coated
.electrode arc welding.
7.2. WELDABILITY ASSESSMENT
BefcJre weldability can be assessed. in relation to a process
three things must be decided, viz.,
(a) the degree of tolerance that. can be allowed for purely
metallurgical defects,
(6) the degree of tolerance that can be allowed for operation-
al defects, and
(c) the type of test that will be suitable in the circumstan-
ces.
7.2.1. Tolerance for Metallurgical Defects
It is well known that the metallurgical structures within a
welded joint and in its immediate vicinity are likely to differ
appreciably from the structure of the parent metal. For
example, with fusion welds the weld deposit will be a cast
structure, tending to show all the defects likely to occur in that
forin of structure and the HAZ is likely to show enlarged grain
size and other structural differences from the unaffected parent
material. It is also likely that the composition ofthe deposit will
differ from that of the parent material. Thus, it is essential that
a decision should be made as to the amount of such defects and
variation that can be accepted and Yf~\ still accept the \~'eld as
completely satisfactory.
Weldability and Weldability Tests 433
7.2.2. Tolerance for Operational Defects
Having determined the minimum metallurgical standards, it
is still necessary to determine the operational standards. Even a
completely automated process is liable to produce defects such as
misalignment, imperfect bonding, poor welding finish, poor weld
shape, etc., hence, as the skill that is required in a particular
process becomes more individual, the more likely it is that defects
of technique will occur. Therefore, a decision has to be taken
concerning the technical standard that is required in making
particular kinds of welds for particular purposes.
7.2.3. Measurement of Weldability
No single test or combination of tests can accurately predict
the performance of most welded structures because the variables
of restraint, fit:up, surface condition. service stresses, etc. are
generally impossible "'toforecasL~.!!gtherefore to duplicate.
Even with these limitations, weldability testing can provide
useful clues as to the precautions such as appropriate selection.
preheat, energy input, joint design and so forth which may be
required for a reasonable degree of confidence in the resulting
welded joint. Thus weldability tests are employed for getting a
quantitative measurement of weldability for a material or
combination of materials.
7.3. WELDABILI'I'Y TESTS
Numerous weldability tests have been devised, all of which
can be classified into the following five groups.
(i) Theoretical tests,
(ii) Simulated tests,
(iii) Visual examination,
(iv) Component sampling tests,
(v) Actual welding tests.
Salient features of all these tests are described in this section.
7.3.1. Theoretical Tests
Theoretical weldability tests include,
(i) Jominy Hardness-Bend Ductility test,
(ii) Prediction of HAZ Hardness from Carbon Equivalent,
etc.
7.3.1.1. Jominy Hardness-Bend Ductility Test
This test predicts the HAZ cooling rate of a given weld from
experimentally based tables that relate cooling rate to,
(a) heat input,
(b) plate thickness, and
434 Welding Engineering and Technology
(c)preheet.
A series of notched be.nd specimens are austenised at 1150°C
and cooled at the calculated rates to duplicate the expected
hardness of the HAZ. If the specimens can be bent through
10°-20° they are considered sufficiently ductile for most appli-
cations.
7.3.1.2.Prediction of HAZ Hardness from Carbon Equi-
valent (for Steels)
In this case a modified carbon equivalent (C.E.) formula,
C••E _- mc %Mn ...(7.1)
70 + 4 + %Ni20 + %Cr
10 + %Cu
40 _ %Mo
50 - %V 10
is combined with a heat flow equation applicable to the given weld.
Thumb rule is that if the calculated HAZ hardness is below VHN
(Vicker's Hardness Number) 300, hydrogen cracking is very un-
likely, while predicted hardness above VHN 400 indicates high
susceptibility to cracking.
Limitations of Theoretical Tests
Theoretical weldability tests suffer from the following limi-
tations.
(i) The theoretical approaches, obviously, do not take the
effects of weld metal into account and since cellulosic
welding electrodes have a much greater tendency to
cause HAZ cracking than low hydrogen electrodes, the
10° bend criterion may not be applicable in many cases.
(ii) The VHN 300 limit is generally considered conservative.
(iii) The effect of preheat may not be correctly rated by these
methods. It has been shown that preheating to 100-
150°C does little to affect the cooling rate in the trans-
formation range; however the Hz-induced crackipg
tendency decreases greatly.
(iv) Calculated weldability formulae may not take into ac-
count all of the welding variables involved.
7.3.2. Simulated Tests
To simulate the heat effects of welding, and thus create a
synthetic weldability tests, two general types of equipments are
available, viz.,
(i) a unit which heats and cools a metal specimen over a
thermal cycle exactly like that found in a particular
weld, and
Weldability and Weldability Tests 435
(ii) a unit which not only heats and cools the test specimen
over the weld thermal cycle, but which can also apply a
controlled tension load to the specimen at any time
during the cycle.
7.3.2.1. Limitations
While the simulated tests provide very useful information
regarding the mechanical properties of various areas within the
HAZ, during as well as after the weld cycle, they cannot account
for residual and reaction stresses, Hz pick-up, and other conditions
which may be imposed on production welds.
7.3.3. Visual Examination for Weldab~lity
Visual examination for assessing wddability of a material
includes the following observations to be made.
(i) Checking the weld size, shape, appearance, and freedom
from performance defects-to ensure that the technique
is of acceptable standard.
(ii) The weld will be examined for signs of cracking or other
defects such as porosity that can be considered to be due
to deficiency in weldability.
Of course it is difficult to separate these two tests completely.
7.3.3.1. Limitations
Examination of this kind does no more than give a preliminary
guide to weldability, since weldability defects are often difficult to
see, even if artificial aids are used, e.g., crack detecting methods
using magnetic particles or fluorescent ink. Also, microcracks or
even large internal cracks are not likely to be readily detected
from outside the weld. Even the quality of craftsmanship or
technique cannot be firmly assessed from such examination
although it forms an essential preliminary step towards that end.
In other words a visual examination of the welded sample is
essential, but only as a first step towards more precise method of
assessment.
7.3.4. Component Sampling Tests
Welding tests made by component sampling are usually
'suck-it-and-see' tests. The welder makes an actual component
which is first examined visually and then, perhaps physically to
destruction, the actual approach depends to a large extent on the
available facilities. For example, if facilities are available,
macrostructures might be studied and mechanical tests made; but
more likely welds would simply be broken up to see the degree of
penetration, etc. Small organisations, rather than large, are more
436 Welding Engineering and Technology

likely to use this form of weldability tests ; but even some of the Weldability and Weldability Tests
larger organisations have been known to use it. 431
7.3.5. Actual Welding Tests (iv) Ring weldability test
An extensive variety of actual welding tests have been devised (v) Hot ductility test.
to measure the weldability characteristics of steels and other Brief description of all these tests is covered in this section.
materials. In general these tests serve one or both of the following 7.4.1.1. Murex Test
purposes.
(i) They are used to calibrate the weldability of particular In this test a fillet weld is made between two test plates 10-15
grades of individual heats of metals and alloys. For this mm thick and of dimensions approximately 50 mm x 70 mm.
purpose the specimen dimensions and welding condi- These plates are rigidly fixed on two supports one of which can be
tions are prescribed and fixed to make the material rotated about an axis in the root of the test weld, as shown in Fig.
sample the only variable. 7.1. Rotation of the movable clamp starts 5 seconds after welding
(ii) They are used to establish a compatible combination of
base material, filler metal, and welding conditions that
will produce welded joints satisfactory for the applica- .
tion.
In effect these two types of tests fall into two groups viz.,
1. Fabrication Weldability Tests
These tests are used to assess fabrication weldability to
evaluate whether a particular material and process can be used
to produce a joint acceptably free of defects.

\
Rotating
2. Service Weldability Tests- clamp -...... Rotation of rotating
These tests measure whether the properties of the welded joint clamp starts 5 seconds
are satisfactory for the intended service or not. after starting welding

7.4. FABRICATION WELDABILITY TESTS

The tests in this category determine the susceptibility of the
welded joint to cracking and can be grouped under three headings
according to the type of cracking that they produce, viz.,
(i) Hot cracking tests,
(ii) Cold cracking testa,
Fig. 7.1. Basic features of a set-up for Murex Hot Cracking Test.
(iii) Tests related to cracks associated with specific fabrica-
tion conditions or structures. begins. Although rotating speed is of the order of l°/sec. but
various rotational speeds are available.
'1.4.1. Hot Cracking Tests
A number of tests have been devised to check the hot cracking The drive mechanism of the machine should be such that it is
tendency of weldments ; some of the well known amongst them not affected by the deformation resistance of the weld metal.
include the following. Susceptibility to hot cracking is indicated by the extent to which
(i) Murex test the weld metal cracks at various rates of applied strain during
solidification. The strain developed is made proportional to the
(ii) Houldcroft test
speed of rotation and is a function of the V-angle, which is always
(iii) Varestraint test initially 90°, The maximum angle that can be obtained is 1200.
Longitudinal cracks occur during rotation however the length of
a crater crack, if present, is not included in measurement.

Welding Engineering
438
This test has been mainly used for assessing cracking
sensitivity of carbon and low alloy steel weld deposits made with
coated electrodes.
7.4.1.2. Houldcroft Fishbone Test
Houldcroft Fishbone test is often used for evaluating the
solidification cracking susceptibility of sheet. materials. This test
is mainly for application to TIG welding of sheet but may be
applied to other similar processes. The test can be applied to any
thin plate or sheet material on which the process is likely to be
used.
Weld
WeldabiIity and Weldability Tests 4:;9
and Technology
The test makes use of the transverse stress pattern (Fig. 7.2)
built up by a progressive fusion weld, to find the critical heat flow
condition for a particular material in a particular thickness of
sheet.
A test piece shown in Fig. 7.3 is made from one piece of
material and has a series of symmetrically arranged slots of
progressively increasing depth of cut in pairs into the opposite side
of the material. The test piece is laid flat on a sheet of carbon
which rests on a water-cooled copper block and is held in position
by Beryllium CoppercIamping strip as shown in Fig. 7.4.
Be-copper Argonarc torch

clamping strip Q/
J:l__
~~~~b~;~;'O~;~~~d
~
Bench
'I',;'Y

Fig. 7.4. Schematic representation

T os t pi '"

of a set-up for Houldcroft fishbone test.

The size of the test piece for any given material depends partly
'"
'" Weld tl on the thermal conductivity ofthe material. Also a larger test piece
..~o
Vl
is used for thicker material. Sr-~cimenwidth is the most critical
dimension and this should be chosen so that in a strip of material
without slots the crack would follow the entire length of the weld
Compression bead. It is important when designing the test piece to have some
material of known crack sensitivity available so that the right
Fig. 7.2. Transverse stress distribution in the central plane
severity of test may be obtained.
of a progressive fusion butt weld. A bead-on-plate weld, without filler rod, is made under
9 slots. o·a mm wide ,7.5 mm aport controlled conditions along the centreline of the test piece,
beginning from the end with the shortest slots and proceeding
right to the opposite end. As tension builds up at the beginning of

- -- the weld a crack is likely to start and will travel along the weld
until the heat build-up in the relatively narrow plate, as

--\ 1----------
10 \ 85
70mm
mm Fig. 7.5. A typical testpiece showing 50% weldability by
HOllldcroft fishbone test for TIG welding an aluminium alloy.
Fig. 7.3. Testpiece dimensions for Houldcroft weld ability test.

440 Welding Engineering and Technology
represented by the distance between the inner ends of the pairs
of slots, causes the stress intensity to fall below the critical value,
as shown in Fig. 7.5. The slot at which the crack is arrested gives
a measure of weldability in a range of ten conditions (nine sets of
slots) ranging from zero weldability with cracking along the full
length to full weldability with no cracking. The crack length from
the starting edge of the test piece is an index of solidification
cracking sensitivity. The ·shaded area in Fig. 7.3 is used only for
starting purposes and is ignored in estimating the result.
In this test the degree of restraint diminishes from start to
finish of the weld due to the increasing depth c;f saw cut, so that
the length of crack has a relation to the degree of restraint
required for crack propagation.
The test is suitable for investigating the effect of parent metal
and filler metal composition on hot cracking and takes no account
whatsoever of the effect of welding variables on hot cracking.
7.4.1.3. (a) The Varestraint Test
-t
Weldability and Weldability Tests 441
The Varestraint test i.e. the variable restraint test requires a
metal plate 275 mm x 100 mm having a thickness of 6 to 13 mm
which is fixed on one side in the welding jig as shown in Fig. 7.6.
The weld metal is then deposited right across the middle of the
plate or alternatively TIG welding process may be employed
without the use of filler.
This test like Murex test utilises external loading to ilnpose
plastic deformation on the test piece while weld bead is being
deposited. The load F is suddenly applied by actuating the loading
yoke, as the centre of the arc passes the point of tangency between
the curved surface of the die block and the fixed end of the
specimen i.e. point A.
In order that the bend radius is not dependent on the
geometrical factors of the weld, the test piece is placed on a
bending block (B) of radius R. Bending of the test plate, with the
weld bead laid on it, produces distortion of the upper fibres of the
weld bead. The severity of deformation causing cracking can be
determined as,

Welding
I
direction
II ---. I
I \
f
CC ( ( ( ( ( II( ( C
-
f--
(AI ----
@
(xI ( (I (.)
I 200 mm
I
@ .. The applied augmented strain, e = ~
where, t = thickness of the test piece,
... (7.2)

E R = radius of the bend.

E
o
-
o A change in the bend radius will also cause a change in the
L size of the distortion of the upper part of the weld bead. If the
applied strain is sufficient, cracks are formed in the weld,bead,
Hot cracking can occur both inside and outside the weld metal.
Top view
Both, the amount of the applied strain and the crack length (either
the total length of all cracks or the maximum crack length) serve
F as an index of crack sensitivity i.e., the evaluation criterion of the
Specimen -
test may be given by the overall length of the cracks as a function
of maximum deformation, e. The relationship between e and crack
length also allows determination of the threshold value of
6mm
deformation, ecrit.at which cracks are initiated and by virtue of this
criterion comparison of various steels or welds can be made. The
higher the value of£crit, the more resistance, to the cracks, is steel.
(b) The Trans-Varestraint Test

\
It is a modified form of the varestraint test. While in the
.p
varestraint test, the axis of the bend is perpendicular to the
Side view direction of the weld and cracks occur vertically to the weld, j n
the Trans- Varestraint test this axis runs parallel to the weld, thus
keeping cracks inside the weld metal resulting in centreline
Fig. 7.6. A schematic representation of the set-up for varestraint test.
442 Weldability and Weldability Tests 443
Welding Engineering and Technology

cracks. The schematic representation of the set-up for Trans-

varestraint test with a bilateral bend is shown in Fig. 7.7.
TIG torch I 0 o
! without filler!

T1
E
200mm
( E
~
~~ o
C>
N
"
Fast bending ''..:
.••...

Former
o o
Fig. 7.7. A schematic illustration of trans-varestraint
test set-up with bilateral bending of test piece. I
I
The evaluation criterion is based on the unit of crack Patch --.:
susceptibility (UeS), where, t--
I
diameter I
I
DeS = 10 (3 - R) ...(7.3) I I
I I
Higher values of ues indicate low weldability. I I
7.4.1.4. Ring Weldability Test
.r
I~--4T' I -, t---I 3
The circular patch or ring weldability test is used for studying
hot c.racking in the weld metal or partially melted zone, because.
I IT\
I :a: ~====': ~l=:r-r mm

it imposes a relatively high restraint in the weld zone transverse
to the weld. The stresses causin~ cracking are not precisely known·
but a material has to be very weldable to pass the test without
cracking. Ignorance about the values of stress is of not much
consequence because in practical applications the usual cwncern
is not whether one weld metal composition is much worse or only
slightly worse than another, but whether or not it cracks.
The test piece is made up of a square piece of sheet or plate,
with sides about 150-450 mm, out of the centre of which is cut a
disc of diameter about one-third of the side of the square. If the
disc can be cut out with a very narrow cut then it can be used in
the test, but if the gap is too wide to give reasonable representation
of a weld gap then a separate slightly larger disc has to be
prepared. Fig. 7.8 shows the schematic of a set-up for such a test
jig while Fig. 7.9 shows a set-up with zero gap for use with TIG
welding.
To make the test, the disc is lightly tacked into the test plate
at, say, 4 posit.ions at regular interval so that. t.he gap is uniform
~---lS0 or200 mm __ ~ll15mm
Fig. 7.8. Schematic illustration of the set-up for ring weldability test.
and representative of an open'square butt preparation. Ifrequired,
single or double Vee edge preparation may also be used. Welding
is started at one of the tacks and is carried on to complete the
circle.
On welding the patch the radial and circumferential strains
imposed on the solidifying weld pool will increase with the
increase in length of the deposited weld and after some interval
both types of strain will reach a maximum, although not
necessarily simultaneously. Thus, during welding a point is
reached when, because of the weakness ofthe material, the strains
are sufficient to initiate a crack. Further, it is assumed that at any
instant the circumferential strains are larger than the radj al
st.rains. This implies that transverse cracks should appear before
centreline cracks ; assuming that the weaknesses are unifor.m
throughout the weld bead.
 Weld ability and Weldability Tests 445

Uncracked portion = 8r
444 Welding Engineering and Technology
Crocked portion = 1360°- SrI

.,' I 150 mm
"\

o~l
Bolts

Transverse
crocking
I
E Centre line
E
<:>
crocking
.-
LI'

Fig. 7.10. Typical test result for patch test in stainless steel.
susceptibility is the interval, measured in degrees, between
individual cracks. This interval is in some way a function of the
strain necessary to cause cracking; the ~maller the i~terval, the
more crack susceptible is the material.
Restraint can be varied in a given plate thickness by adjusting
the plate size and patch diameter. Also, by using a patch of one
material and a plate of another the cracking tendencies of
dissimilar metal joints can be studied.
r·;· ~::~.~~;~~\ Ring weldability test is used in aluminium alloys and low alloy
steels for establishing correct welding procedures and sequences to
avoid restraint cracking. Fig. 7.11 shows an example of a welding
/::?;:·:n

Te5t specimen Grooved backing ring

Fig. 7.9. Jig for ring weldability test in sheet material using TIG welding process~.
Once a centreline crack has started it should continue until
the welding is terminated. This is because the increased strain
brought about by the notch effect at the root of the crack will help
its continuance. A measure of the crack resistance is the angle
Sr, subtended between the start of the weld and the point where
the crack first began, as shown in Fig. 7.10. Likewise the crack
susceptibility is equal to (360-Sr). Cracks at the beginning should
be ignored except where it is obvious that they form part of the
main weld, since the region at the start of the weld is not
representative of the welding conditions obtained in the main part
of the weld.
Fig. 7.11. Welding sequence to help eliminate distortion
Another criterion of crack sensitivity is the occurrence of and cracking in a circular weld.
transverse cracks; in some cases these are the only types of cracks
observed. To a first approximation, a measure of the crack
 "e~~ r
I

446 Welding Engineering and Technology Weidability and Weidability Tests 441
sequence to help eliminate distortion and cracking in a circular 7.4.1.6. Statistical Evaluation of Hot Cracking Tendency
weld. This may be used for comparison if the patch test shows
cracks. Since the setting up of physical facilities to experimentally
determine the hot cracking tendency of a material often meets
7.4.1.5. Hot Ductility Test with innumerable obstacles, parametric equations are frequently
The thermal cycles experienced by the workpieces during used to evaluate susceptibility to hot cracking. Because these
welding can be duplicated in small specimens conveniently for equations are developed from statistical processing of entire series
mechanical testing, by using weld thermal simulator. of targeted range of parameters, they are therefore of limited
By performing high-speed tensile testing during weld thermal applicability. However, these equations are quite useful in the
simulation, the elevated temperature ductility (and strength) of absence of any other guidelines. Some of these parametric
equations are as follows.
metals can be evaluated. This is called the hot ductility test and
is illustrated in Fig. 7.12. This test can be used to investigate the (i) The susceptibility of structural steel welds to hot crack-
hot cracking tendency of the partially melted HAZ in a weld. It ing, derived from the Houldcroft Fishbone test, desig-
has been used extensively for evaluating the hot cracking nated by the number HCS (hot cracking sensitivity) may
susceptibility of nickel-base alloys. be represented by the following equation.
Zero-Ductility Temperature C [ S + P + 25
Si + 100
Ni] 103
HCS = 3M n+ C M0+ V
r+ ...(7.4)
•••
If HCS < 4, the susceptibility" of the material to hot
•...
::l Heat- Affe c ted- Zone cracking is low. However, for high strength steels to
0+-
'"
•...
2
Thermal Cycte have good weldability HCS < 1.6 - 2.
•••
0- (ii) To evaluate susceptibility to solidification cracking the
E
•••
I- equations recommended by Ostrovskay are based on the
evaluation of carbon equivalent (C.E.)

i
(a) For steels with a carbon content of 0.09 to 0.14 %,
C E = C 2S P (Si - 0.4) (Mn - 0.8)
Time - " + +8+ 10 + 12
Ni Cu (Cr - 0.8)
+ 12 + 15 + 15 ...(7.5)
(b) For steels with a carbon content of 0.14 to 0.25%,
c·••. = C 2S P (Si - 0.4) (Mn + 0.8)
+ +8+ 7 + 8
1 Zero-Ductility
2 " 3
temperature
On -Heating 4 ",,' \ • .L:J.

t
Ni Cu (Cr - 0.8)
<~ +8+10+ 10 ...(7.6)
>-'0
0+-
:.=
:+:
C
0 (c) For steels with a carbon content of 0.25 to 0.35%,
\.I:;:
::l ••.• C E - C 2 r.: S ~ (Si - 0.4) (Mn - 0.8)
Cl ::l 9
On- Cooling • " - + .0 + 2.5 + 5 + 6
~
~
• - Ni
6 + Cu
8 + (Cr -8 0.8) '" (7 .7)
5

Testing Temperature_ The greater the C.E., the higher the susceptibility to hot
cracking.
Fig. 7.12. Illustration of hot ductility test results;
(a) HAZ thermal cycle, and (b) temperature vs. ductility curve.
448 Welding Engineering and Technology
Weldability and Weldability Tests 449
For evaluating cracking susceptibility of weld metals of
(iii) 3 Directions heat flow
austenitic stainless steel, the parameter derived from
tests of hot ductility is,
~H = 700 C + 17Cr - 37 Ni - 117 Nb + 29 Mo + 188 ...(7.8)
Steel is susceptible to hot cracking if M-I < O. On the contrary,
I Shim to simuloie
with Mi> 100 steel possesses outstanding hot ductility and thus
has very good weldability.
la)
T
L Test
View on ·AA
thickness
poor fit-up

7.4.2. Cold Cracking Tests Anchor welds I Fir si 1

Weldability is also assessed by the cold cracking susceptibility 75x75x12
of a weldment. Like for hot cracking, there are a large number of
tests developed to determine the cold cracking tendency; some of
the more popular amongst them include the following. -
. --
..- ..
(i) Controlled thermal severity (CTS) test, ~ -
- -
(ii) Tekken test,
(iii) Lehigh restraint test,
A ~-- 1} S"l;on, fa, '''I
(iv) Longitudinal bead-weld test, Test weld (, )
(I) Implant test.
Salient features of all.these tests are described in this section. Ib)
I Q & b I Set - up
7.4.2.1. CTS Test Poli.sh these faces
The CTS test is based on the principle of the lillet welded.joint
particularly for assessing weldability \n relation to steels welded
by arc welding processes for establishing safe welding proced~res
for low alloy steels. Under appropriate conditions the test can
assess a steel parent m~tal, a weld deposit, or a process in terms
of a critical cooling condition related to the number of units of 6.25

mm weld..
the (~inch ) of thickness of material conducting heat away from
The test pieces of the dimensions shown in Fig. 7.13 are bolted
together. Ie) Sections for mirostructural examination
Welds A and B are anchor welds and the test welds are fillet All dimensions in mm

welds of standard size laid at C and D under controlled conditions.

The test piece is set up in a convenient fixture so that the test
weld can be made in the open flat position i.e., with the V in the
upright position and the line of welding horizontal; the whole
set-up is exposed only to normal still air cooling. Each tllst weld
is begun with the test piece at room temperature. Thus, weld C IS
made under bithermal conditions, with two thicknesses of plate
conducting away the heat, whilst weld D is made under tri-
thermal conditions, with three thicknesses conducting away the
heat (the bottom plate conducts in two directions).
Fig. 7.13. Controlled thermal severity test, (a and b) set-up, and
(c) location of test pieces for microstructure study.
The cooling rate is designated by means of a thermal severi ty
number (T.S.N.). TSNI is the thermal severity corresponding to
heat flow along a single steel plate 6.25 mm thick. TSN2 is
obtained in a butt weld between two 6.25 mm plates whilst in a
6.25 mm T-joint, where there are three heat flow paths, the
thermal severity number is 3. The TSN is also increased in
proportion to the plate thickness, so that CTS test pieces, which
TSN6.have three heat flow paths, in 12.5 mm plate would have
also
 Weldability and Weldability Tests 451
Welding Engineering and Technology
450

The thicknesses t and b, of top and bottom plates respectively, 1- 200 mm

can be changed as required for successive tests, the thermal

severity number being calculated from the units of 6.25 mm of Auxiliary Auxiliary
thickness. weld Tested
weld
weld
Thus, in general,
(i) for bithermal welds, E
1 ...(7.9) oE
IJ"\

(TSN)B = 6.25 (t + b)
and (ii) for tri-thermal welds,
1 ...(7.10)
i.A
(TSN)T = 6.25 (t + 2b)
where t and b are in mm.

M
Susceptibility to cracking grows with the increasing gap 60· 60·
between the plates, therefore a shim is often placed between the
two plates to increase the gap in the fillet weld as shown in Fig.
7.13 (a). Apart from thickness of plate, the severity of the test may Auxiliary
also be varied by the hydrogen level in the test welds and the weld
composition of the weld metal. With this test, cracks occur in the
underbead zone or the weld metal.
The assembly is allowed to stand for a period of 72 hours after
which the welds are sectioned to prepare three test pieces from "1.(2 60·
the transverse sections, for macrostructure study, as shown in Fig. Section at B - B
Section at A-A
7.13 (c). -
Standard known quality plates can be used to assess weld- Fig. 7.14. Test piece and edge preparation details for Tekken test.
ability of particular elel,:trode deposits and standard weld deposits are made on both ends and then the test weld of length about 75
can be used to indicate weld ability of particular sheet or plate
mm is made in the central section. When SAW process is used, the
material. .' auxiliary wel~. on one side is left incomplete so as to leave room
Some authorities do not regard the bithermal and tri-thermal for the test weld proper. It is reported by Japanese research
conditions. as giving any sufficiently significant differences in workers that the intensity of the restraint is not much affected by
information so make the test plates of a size to give two bithermal the dimensions of the test pieces.
welds instead of one of each type.
Tekken test is used in selecting welding parameters for the
For effective location of the critical TSN it is necessary to make root run of butt joints. In this test three types of cracks, viz.
several test welds under different conditions, so a complete a, band c may be noted, as shown schematically in Fig. 7.15.
assessment will take several days to finish. These are,
7.4.2.2. Tekken Test (i) the cracks of the type (a) initiating from the fusion
This is a simple butt welding test and has found wide boundary zone of the bottom part of the root run on the
application in determining cold cracking tendency of welds made double-bevel side of the weld edge. These are typical cold
by arc welding processes including submerged arc welding. The cracks extending to HAZ and then turn back to extend
thickness of test plates of the dim ensions shown in Fig. 7.14 are right into the weld metal.
prepared for butt welding with Y-edge preparation having 2 mm
root gap, and a 60° groove angle. First symmetrical auxiliary welds
 WeldabiIity and Weldability Tests 453
Welding Engineering and Technology
452
determine the percentage of crack incidence in relation to welding
parameters. For example, Fig. 7.16 shows the relationship
between preheating temperature and percentage of cracks. The
test weld is cut for examination atleast 48 hours after welding.
Tekken test is suitable for comparing the cold cracking suscep-
tibilities of parent materials.
Plate
7.4.2.3.Lehigh Restraint Test
The characteristic feature of this test for thick plates is that
the degree of restraint is varied by freeing the edges of the test
pieces with a series of sawcuts extending inwards over a distance
Plate X' from the edge ; the depth of the sawcut determines the
restraint.
300 mm

20·
a , b ,c - cracks

Fig. 7.15. Schematic representation of types of cracks (a, b, c)

observed in a Tekken test specimen. --i
12.5

(ii) Cracks ofthe types (b) and (c) initiate in the weld metal
and may join to become a single crack.
:i;f 90-140mm
,...,
-0.

The test procedure involves metallographiC analysis on 5

sections, two of which are obtained by cutting through the initial
and final weld craters. From these analyses it is possible -to
100
A

90
A longitudinal U-groove is milled in a plate about 300
80 mm x 200 mm ; the shape of the groove is shown in Fig. 7.17 and
~ its length (L) is varied in accordance with the plate thickness for
a plate less than 25 mm thick, L = 90 mm and for a plate more
..• than 25 mm thick, L = 140 mm.
~v
Cl
The test piece thickness varies between 12-50 mm and the
root gap of the U-shaped groove measures about 1.6-2 mm. For
-o
•...
v
plates of thickness up to 25 mm, a single U-edge groove is cut,
~ while for thicker plates double-U groove is used.
o
Weld metal is deposited in the grooved preparation and
explored for cracking by visual examination or by magnetic
particle testing or radiography, etc.
400 500 For steels with 0.30% C, cracks occur practically exclusively
100 ZOO 300 in the weld metal but may be initiated in the root or the upper·
Preheat temp.:C part whereas crack sensitive higher carbon or alloy steels show
cracking in the HAZ extending into the weld metal. This test is
Fig. 7.16. Tekken test data represented to correlate preheat
temperature to percentage of weld cracks.
 Weldability and Weldability Tests 455
Welding Engineering and Technology
454
axis of the weld bead, and the cross-section of one-half the tea t
recommended for the selection of electrodes for use with arc plate is ground [Fig. 7.18 (c)] and tested for underbead cracks by
the magnetic particle or metallographic technique. The weld bead
welding processes. on the second half of the test plate is ground flush with the plate
7.4.2.4. Longitudinal Bead Weld Cracking Test surface [Fig. 7.18 (d)] and thc·mexamined in the same way for toe
This test uses a bead-on-plate weld deposited on a steel test cracks around the edge of the weld bead. Cracking is measured
plate of the size 150 mm x 75 mm x 25 mm, partly immersed in after the specimen is aged for 24 hours at 15°C, and is then
water to 6.25 mm of its top surface, with the length dimension in thermally stress relieved at 595°C for one hour to avoid possible
the direction of rolling. grinding cracks. Results are expressed as total length of crack(s)
A bead 100 mm long is deposited in the central part of the' test as a percentage of the test weld length; underbead and toe cracks
are reported separately. For a reasonable evaluation atleast ten
plate, as shown in Fig. 7.18 (a), with a 3.15 mm diameter electrode tests have to be conducted.
Cut through along this line '-- x
7.4.2.5. Implant Test
While the four cold cracking tests described above are
self-restraint tests, the Implant test is a forced restraint test. In
this test, a cylindrical specimen 6--8 mm diameter, and the other
dimensions as shown in Fig. 7.19 (a), is notched and inserted in a

The implant specimen V'\

V\
6+1
<:>

V'\
I a 1 Bead on plate Underbead cracks 6
y y A
-
(a l <:>
cr:

Sheet
A Ic 1 Section along YY
100
-tE+

r ~;;:;
nu/ }--La "--T;. -';;'~k;' 1 I ! ~ I
( b I Welded plate c·ut along
weld centre line
(d 1 Welded bead ground
flush with top surface of plate

Fig. 7.18. Longitudinal bead weld cold cracking test; (a) test plate dimensions,
(b I
-+u· <:>
<:>
<:>
N
<:>
(b) test plate cut longitudinally along weld axis, (c and d) cut and ground test
pieces for underbead and toe cracks.
at a welding current of 100 Amp and an arc voltage of 24-26 volts
Dimensions in mm
and a travel speed of 25.4 em/min. ; E60l0 type electrode is used
for welding to provide a high potential of hydrogen in the arc I Fig. 7.19. Dimensions of the test components for implant cold cracking test.
atmosphere. The specimen is then cut longitudinally along the
l
 Welding Engineering and Technology Weldability and Weldability Tests 457
456

hole in a 20-30 mm thick plate made of the same or similar
material as the cylindrical component. The detailed dimensions of
notch are also given in Fig. 7.19 (a). A weld run is made over the
specimen, with an electrode of the type that is to be used in actual
fabrication, which is located in such a way that its top becomes a
part of the fusion zone and the notch lies in the HAZ, as shown in
Fig. 7.20.
I
load
Test specimen
Fig. 7.20. The notch in the cylindrical component and its
location in HAZ in the implant test.
Mter welding, when the temperature falls below 100°C, a load
is applied to the cylindrical specimen, and the time to failu.re is
determined. A plot of stress versus time to failure gIves an
assessment of hydrogen cracking susceptibility. Fig. 7.21 shows
down to 125°C. It is evident from the plots that the welds made
with low-hydrogen electrodes (E7018-Basic Coated) are less
susceptible to hydrogen cracking than the welds made with high
hydrogen electrode (E7010-cellulosic coating). The GMA weld
made with Ax+ 2% O2 as the shielding gas was least susceptible
to cracking as the only hydrogen encountered was the residual
hydrogen in the base metal, test pieces, and the welding wire.
Thus, this test also permits the critical stress, Scrit' to be determined
at which no fracture or crack initiation will occur anymore.
Another possible procedure is to fix the applied stress equal
to the yield stress of the test piece material and to alter the
thermal conditions like preheat, heat input, postweld heating. [n
this case it is possible to determine the critical cooling rate, tcrit,
between 800 and 500°C, above which no cracking takes place.
Cracking parameter (Pc) for steel with 0.8 to 2.5 % Mn, suggested
by Tanaka and Kitada, is given by the following equation.
%Si %Mn %Cu %Cr %Mn
Pc=%C+ 20 +~+20+30+~ ...(7.11)
No cracks occurred with values of Pc < 0.25%, nor in samples
with a maximum hardness of the underbead zone up to 270
HV10'
Mutsui and Inagaki deduced the following parametric
equation for calculating the critical stress (Scrit) in the Implant
test,
Scrit = 86.3 - 211 PCM - 28.2 ([H] + 1) + 27.3 tS/5

+ 9.7 x 10-3 tlOO[ 1~0 MPa] ...(7.12)

IAr+2%021 where the crackability parameter PCM' as determined by the fol-
lowing equation,
-
o
Mn Mo Ni Cr
PeM = C + 20 + 15 + 60 + 20
V Cu Si
+ Involvod in AE study + 10 + 20 + 30 + 5B ...(7.13)
_No fracture
ranged between 0.16 and 0.28.
200 _
10 100 1000 10000 100000 Also, the content of diffusable H2' as determined by the
• Time to fGilure ( secs) - glycerine test ranged from 1.2 to 28 mV100g and the cooling time
tS/5 was between 2.0 and 5.0 seconds. Cooling time tlOO between
Fig. 7.21. Implant test results for an HSLA pipeline steel.
maximum temperature and 100°C varied over the range 58.5 to
such plots. for high strength low alloy (HSLA) pipeline steel. In 1409 seconds.
this case loading was applied to the specimen when the weld cooled
458
Welding Engineering and Technology Weldahilit:r and Weldability Tests 459

7.4.3. Weldability Tests Related to Specific Types of there should be no defects such as slag inclusion or lack of fusion.
Cracking
Specimens are then cut out of this assembly and machined to a
In this section tests employed to assess weldability with round tensile test bar of 10 mm gauge diameter so that the original
regard to the following types of cracks are described. plate forms the central part of the gauge length, as shown in F~g,
(i) Lamellar tearing 7.22. If the plate is thick enough the whole test piece may be
(ii) Reheat cracking. ma{.-hinedfrom it.
The ductility in a tensile test made on such a specimen is taken
7.4.3.1. Tests for Lamellar Tearing as a measure of susceptibility; material having a through-
The tests employed to check the susceptibility to lamellar thickness ductility less than, say, 25% is regarded as susceptible
cracking include the following. to lamellar tearing.
(i) Through-thickness Ductility test, Steels may thus be classified accordi:r;1gto the required and
(ii) The Cranfield test, guaranteed values of contraction in the through-thickness
(iii) Lehigh Cantilever Lamellar Tearing test,
direction (eJ. The highest quality steel has ez > 35%. Steel resis-
(iv) The Window test, tance to lamellar tearing is especially given by its count of sulphur
(v) The H-test, and sulphide as counted on the basis of planar ratio in the various
layers of the rolled texture.
(vi) The Brodeau test, and.
(vii) The Notched Tensile Strength test.
(ii) The Cranfield Lamellar Tearing Test
In this test a supporting plate of the suitable dimensions is
For want of space, out of the above listed seven tests, only the placed on the test plate and held in position by auxiliary fillet tack
first two are briefly described here.
weld. The dimensions of the test plate are as shown in Fig. 7.23.
(i) Through-Thickness Ductility Test
The most widely used test for determining susceptibility to
lamellar tearing is the through-thickness ductility test. Plates are
welded at right angles to and on the opposite sides of the plate to
be tested. Alternatively round bars may be friction welded thereto .
However, it should be ensured that the strength of the weld metal
•••••c::,<:::.
f~"<'y
Supporting
sheet

is higher than that of the strength of the test plate material, and

Auxi~iorY fillet welrd (Tack I -,

t---L-l Test plate

I 120 _-1
I - Oimtn$ion$ In "''''

Fig. 7.23. Dimensioned view of Cranfield Lamellar Tearing test componentil,

Fig. 7.22. Steps involved in preparing test pieces for
Through-Thickness Ductility ,test for lamellar tearing.
Various layers ofthe test runs are laid on the portion ofthe set-up,
with a step of7 mm. Usually 7 to 12 weld runs are made depending
upon the number of layers to be used in actual fabrication. This
produces stress due to the weld shrinkage in the
through-thickness direction of the test plate; greater the number
of layers higher the stress produced. The shrinkage of test weld
runs also causes an angular distortion of the test plate. ~lith an
angular distortion above 3°, the size of the lamellar crack, if
460 Welding Engineering and Technology Weldability and Weldability Tests 461

produced, increases abruptly. The occurrence of tearing is checked 0'1 ((X2 - (Xl) T
metallographically. El = E1= E A
_1_1 + 1
...(1.14)
7.4.3.2. Reheat Cracking Tests E2A2
A large number of tests have been evolved to determine the where, El = Overall strain in the test material,
susceptibility of a weldment to Reheat Cracking. Presently the (Xl= Coefficient of thermal expansion of test material,
following three tests are more popularly employed.
El = Young's modulus oftest material,
(i) Vinckier test,
Al = Cross-sectional area of test strip,
(ii) Compact Tension test,
(iii) Hot Tensile test. E2 = Young's modulus of stainless steel bar,
Salient features of these tests are described in this section. A2 = Cross-sectional area of stainless steel bar,
(i) Vinckier Test T = Reheat or PWHT temperature.
Thus, the overall strain under which the reheat cracking will
This test is used to evaluate the reheat cracking susceptibility occur can be evaluated.
in corrosion-and heat-resistant low-alloy steels. The test speci-
mens are made by welding two pieces of 50 mm thick plates (ii) Compact Tension Test
The reheat cracking susceptibility can be evaluated by
together with about ~.of the weld completed. The welded samples measuring the extent of crack growth in a compact tension test
are then cut into strips, and the strips are welded to an austenitic
stainless steel bar as shown in Fig. 7.24. This assembly is then
... .....

{al stainless steel bar

.,.- - T
stainless steel
"'.
test specimen
fillet weld

0':,
I bl
Q
I Cr Ho V

SOmmlWI
Fig. 7.24. Test specimen used for Vinckier Reheat Cracking test after test pieces
are welded to stainless steel bar; (a) front view, and (b) top view.
heated and held for 2 hours at the PWHT temperature. Due to ~
reheating by PWHT, the test strips are subjected to tensile
loading, owing to the higher thermal expansion of the stainless x POTENTIAL DROP PROBES

steel bar. The greatest length of the sample in which no cracks are Q CRACk' LENGT H

observed is the measure of susceptibility to reheat cracking.

Alternatively, the following formula can be used to determine the Fig. 7.25. Compact tension test specimen for
overall strain in the test specimen. testing susceptibility to reheat cracking.
462 Welding Engineering and Technology
Weldability and Weldability Tests 4~3
specimen or simply a charpy V-notch specimen, which is loaded Table 7.1. Solderability Tests
and kept at the reheat temperature of, say, about 600°C. Fig. 7.25 S. No. Test Method
shows a compact tension test specimen made from a CrMoV-type Applications
1. Solder bath immersion vertical Wires, terminat.ions, tags, terminals.
("edge") test circuit boards.
base metal welded with coated 2'~CrMo electrodes. Crack 2. Solder bath rotary dip test Printed circuit boards, pla~ed
extensions are monitured using the d.c. potential drop technique. through-holes.
The two m.s. strips attached to the test piece are for conducting 3. Globule test
d.c. through the specimen. Round wires, plated through-holes.
4. Surface tension balance test Wires, terminations, tags, some
(iii) Hot Tensile Test types of circuit boards.
This test is made after first subjecting the tensile test \Vires, terminations, tags, terminaLs,
circuit boards.
specimen to simulated weld thermal cycle. Subsequently a tensile
test is made at 600°C and both the strength and reduction of area
are measured. A combination of strength below that of the base -----------------
6. test) of spread
5'lMeniscus
Area rise rate
Solder Bath Dip Test.
testtest(Sessile dro~Flat -.-----.--------
cm2). surface of sufficient area (e.g. 5

This test involves immersion of a

metal and reduction of area below 20 % indicates susceptibility to
reheat cracking. The test is probably the most generally used since
it is simple and gives quantitative data.
7.4.4. Allied Weld ability Tests
Apart from the different weldability tests describedllisted in
the previous section, tests may also be carried out to determine
solderability and brazeability.
7.4.4.1. Solderability Tests
Solderability testing of components is usually based on
measurement of the time of contact with molten solder to achieve
good wetting under given test conditions. The factors which affect
solderability are:
(i) nature of the substrate surfaces,
(ii) composition and purity of the solder,
(iii) the activity of the flux used, and
(iv) temperature.
In general the electronics industry requires to determine the
solderability of surfaces so that the solder, flux and heating
conditions are held constant while the surfaces of the substrate
are tested. On the other hand flux activity may be tested if a
constant surface, solder, etc. are used in order to determine the
optimum flux to be employed. There are a number of solderability
tests as listed in Table 7.1.
Two of these test methods which 2.re widely used by the
electronic industry are described briefly in this section.
specimen for a given time into a solder bath followedby assessmen.t
for completeness of wetting by the solder. There are several
variations, the simplest being a vertical dip test; dipping rate js
controlled and may change with the different materials under test.
It is also necessary in this test to assess the percentage area of the
surface which is poorly wetted, such as a surface area with not more
than 5% defective wetting and not located in one location.
Rotary Dip Test is an alternative method designed to
simulate the wave soldering of printed circuit boards. Samples cut
from circuit boards are carried through a circular path so that
their lower face makes a contact with the surface of a solder bath;
a succession of samples may be f,rivendifferent contact times by
varying the speed of rotation. In this way the minimum time for
complete wetting by the solder may be determined, as well as any
tendency for dewetting to occur. This test is widely used by
European industries for printed circuit laminates and plated
through-holes, the latter being tested by measuring the time for
solder to rise through the hole to the upper surface of the board.
Solder Globule Test. This test is for round component
terminations. A wire oriented in a horizontal plane descends into
a molten globule of solder of given size so that the globule js
bisected; the time is measured for the solder to rejoin over the top
of the wire. The method is less easily applied to non-round
component terminations, it does not rea<iilyindicate tendency for
dewetting and it may be difficult to apply it to extremely short
terminations. The globule method can be used for solderability
test on plated through-holes (PTH); the globule contacts the lower
face of the PTH and a contact needle located exactly in the plane
of the upper surface of the board, and on the axis of the hole,
detects when the solder has risen through to the top of the hole.
 Weldability and Weldability Tests 4iiS
464 Welding Engineering and Technology
Low carbon steel
7.4.4.2. Brazeability Tests p.t~Uo.\ ~~re~~!.~.!.r~i~--
....--- .... ..,
Similar to that for soldering so also for brazing the most .........

important aspect ofjoint filling by capillary action is that of wetting

of the base metal by the molten filler metal. However, this seems to '"
be less of a problem during brazing persumably because of the high
thermal energy input which can ensure that surface interactions will -..
'"
'"

VI
occur. Therefore, tests of brazeability are rarely carried out except
in the context of investigational work and research.
7.5. SERVICE WELDABILITY TESTS I Plastic deformation
This class of tests is designed to measure mechanical pro-
perties that are considered vital to the satisfactory performance
of the welded joint in service. These tests include;
(i) Tensile tests Uniform elon~ ~
Local necking l C J IA
(ii) Nick-break test Extension
~permanent set
(iii) Pillow test 0·1 set
(iv) Bend tests I<'ig.7 26. Stress versus strain curves for mild steel and harder steels for which
(v) Impact tests proof stress is used as a measure of yield strength.
(vi) Hardness tests Longitudinal weld
specimen
(vii) Fracture toughness tests
(viii) Fatigue tests
(ix) Corrosion tests Both plate - type
specimens have iden-
(x) Creep tests. tical dimensions
For all the mechanical tests the specimen should be taken so
that the welded joint is in the middle, and the reinforcement is
removed flush with the parent metal. The specimens are then
prepared for various tests.
7.5.1. Tensile Tests t
Tensile tests are carried out to determine the ultimate tensile
Transverse
strength (UTS) and yield point under static loading of base metal, weld specimen
weld metal and the welded joint. Percentage elongation (ductility
of joint), over and undermatching, reduction of area as well as
proof stress· for hard steels and some non-ferrous metals can also
be determined. For determining the tensile strength of a base
metal the usual procedure of finding the tensile strength is used 6·4mm or12.Bmm
* Proof Stress: For the harder steels and non-ferrous metals. unlike mild diam. round specimens
steel, there is no sharply defined yielding of the material, an.d a regular curve is depending on t
obtained between stress versus strain. A stress which corresponds to a definite Fig. 7.27. Typical tensile test specimens for evaluation
amount of permanent extension, called proof stress, is commonly obtained as of strength of welded joints.
illustrated in Fig. 7.26. A line AB is drawn from A, so that OA represents 0.1% or
some other permanent set. AB is drawn parallel to the rectilinear portion of the and to determine the tensile strength of weld metal alone or
curve and cut~ it at point B. The stress corresponding to point B is the 0.1% proof welded joint the samples are prepared which are of the types
stress. The material fulfils the speritication if. after the proof stress is applied for
15 seconds and removed. the specimen haR not permanently extended more than shown in Fig. 7.27.
0.1 % of the gauge length.
466 Welding Engineering and Technology
Different types of tensile test carried out to evaluate
weldment include the following:
(i) All weld metal test,
(ii) Longitudinal butt weld test,
(iii) Transverse butt weld test,
(iv) Transverse butt weld with a notch test,
(v) Tension-shear test.
And then there are tension tests for resistance welds.
7.5.1.1. All Weld Metal Test
In a commonly used weld tensile test, the axis of the test bar
is parallel to the axis of the weld, and the specimen is machined
so that the entire bar consists of deposited weld metal with a
diameter of 6.4 mm or 12.8 mm as shown in Fig. 7.27 (a). In
general, the only complication associated with this method is the
possible dilution of the weld deposit by base metal melting. For
that the dilution may be kept to the minimum. But if the actual
procedures are adopted it will give more representative value of
the actual metal.
7.5.1.2. Longitudinal Butt Weld Test
In the longitudinal weld tension specimen, the direction of
loading is parallel to the weld axis, but the test differs from the
all-weld metal specimen in that the gauge cross-section contains
weld, HAZ, and base metal as shown in Fig. 7.27 (b). All of these
zones must strain equally and simultaneously. Weld metal,
regardless of strength, elongates with the base metal until failure
occurs. Poor weld or HAZ ductility may force fracture initiation to
occur at strength levels considerably below that ofthe surrounding
unwelded base metal. On the other hand, there are instances,
where the use of a lower str~ngth but more ductile (and possibly
less crack sensitive) weld metal is preferred. For example, a girth
weld in pipe is stressed less across the weld than along the weld
(nominally, transverse stress is one-half the longitudinal stress
due to internal pressures). In this case, lower strength weld metal
may be completely adequate.
7.5.1.3. Transverse Butt Weld Test
Interpretation of test results for the welded joint as a whole
is complicated by the different strengths and ductilities generally
found in the various regions of the joint. This lack of uniform
properties is particularly important with the transverse test
specimen shown in Fig. 7.28 in which straining and failure will
tend to occur primarily in the weaker areas of the gauge section.
For example, if the weld metal strength exceeds that of the base
Weldability and Weldability Tests 467
a
T-t
3O±.mf __ ~
:~ ---37,0;'
_ to,hcq
machIne grips
=-=--=--5-0B,~~8~-- i
Fig. 7.28. Reduced tensile test specimen for transverse butt weld test,
metal, with resultant necking and failure outside of the weld area.
In such a case, the test indicates that the weld strength is above
the D.T.S. achieved in the specimen but does not give any
indication of the weld ductility. Qualitative comparison of various
weld metals are not possible when failures· of this type occur.
When the weld strength is considerably lower than that of the
base metal, most of the plastic strain occurs in the weld. The
localisation of strain in this case will result in low values of tensile
elongation, as indicated by measurements over the normal gauge
lengths. It is obvious, then, that transverse weld specimens may
provide a measure of joint efficiency in terms of strength, but do
not provide a good measure of weld ductility. As with the elon-
gation, it is also not possible to obtain a reliable measure of Y.S.
across a welded joint since the conventional use of YS. depends
upon uniform straining within the specified gauge length.
Because of the above cited factors, the primary information
gained from the transverse weld test in the D.T.S., YS., and
elongation requirements are generally not specified. Inspite of
these limitations, the transverse weld test is the standard
accepted test for procedure qualification and is also quite widely
used for the same purpose in various codes, since it does not
indicate whether or not the weld strength equals the base metal
tensile strength or some other specified minimum value.
7.5.1.4. Transverse Butt Weld With a Notch Test
In this test the test piece similar to the transverse butt weld
specimen is employed but with a notch machined in the weld metal
portion as shown in Fig. 7.29. This test is used when it is desired
to gain a fair idea of the strength of the weld material without
risking failure outside the weld. No attempt is made to create
special notch conditions ; the notch serves only to create the
maximum stress into the weld material. Strain measurements
may be made on the first three types of test pieces discussed but
would be practically meaningless on the notched specimen. ] t
should also be noted that strength values obtained with the
notched specimen must be used with caution since the results are
468

-- -
Welding Engineering and Technology
Weldability and Weldability Tests 4/19
"..\::.t; •."
I ~.':;,
....
~~~,
...
I Tension shear tests may be used to evaluate the shear
Necked weld
properties of fillet welds. Such tests are usually intended to
represent completed joints in weldments and are therefore
J prepared using similar procedures. Two types of specimens are
usually employed, uiz., transverse and longitudinal, as shown in
1_· ~_-

Fig. 7.29. Necked tensile test piece for use with wedge grips.
obtained by off-standard methods and are likely to be higher then
normal because of the concentrated strain.
Because of the above stated facts, longitudinal weld tests can
provide more information about the composite joint than the
transverse test, particularly when weld and base metal strengths
differ significantly. Of course, a considerably greater length of
welded joint is required for longitudinal than for transverse
testing.
7.5.1.5. Tension-Shear Test
Fig. 7.30.
Amongst transverse shear specimens, double strap fillet joint
specimens Fig. 7.30 (b) are preferred because they are mm"a
symmetrical and therefore the stress state under load better
approaches pure shear. Single lap specimens are generally not
used for plates over 6 mm thick. The data obtained from
transverse fillet weld tests are the weld shearing strengths,
reported as either load per lineal mm of weld or MPa based on the
weld throat.

Centre lines of
vertical plates
shall be in the
same plane

(a) lap jOint

Transverse
Longitudinal
(b) Double strap butt joint

Fig. 7.30. Teansverse type specimens for tensile shear test;

(a) double-fillet lap joint, and (b) double strap fillet welded
butt joints for transverse and Ibngitudinalloading. Fig. 7.31. Test piece and method of cutting test specimens from cruciform type
fillet welded joint for tension-shear test.
470 Weidability and Weldability Tests 41]
Welding Engineering and Technology
Edges as sheared
Also, for transverse fillet weld test cruciform type of test piece
may be used; it is welded and cut into test specimens as per details
given in Fig. 7.31. Tensile test is carried out by the usual procedure I
, I
though failure will take place along the throat of the weld by the
shear mode. _,_~~" I
+ -1..-'-'--- _- _
Longitudinal fillet weld shear test measures the strength of
. \ I I
" I -' I
the fillet weld when the specimen is loaded parallel to the axis of I
Spot- weld I
as shown

I
Ii·
__

I
I J
Fig. 7.33. Test specimen for tension-shear testing of resistance spot weld.
When the sheet thickness becomes large such as 4.8 mm and
(a 1 After welding I b) After machining greater, the wedge grips of the test machine should be off-set to
Fig. 7.32. Longitudinal tilIet welded test specimen for tension-shear test. reduce the eccentric loading which is accentuated by the thickness
of the specimen. Thus a more precise shear load will be imposed
the weld as shown in Fig. 7.32. The weld shearing strength is on the spot weld thereby minimising the tension or peeling
reported as load per lineal mm of weld for welds which fail under component.
test.
The tension-shear test is commonly used in production
All the fillet weld shear specimens are sensitive to specimen assurance testing because it is an easy and inexpensive test to
preparation procedures, as well as welding ,defects like root perform.
notches, undercut, and bead contour. For transverse specimens, it This test is used mainly to establish ultimate shear strength
is recommended that the specimen edges be machined to eliminate when the specimen is tested in tension. When this test is used in
crater effects and to provide a notch-free edge. combination with the cross-tension test, the cross-tension strength!
7.5.1.6. Tension Tests for Resistance Welds tension-shear strength ratio is referred to as a measure of
The following tests may be carried out to determine the ductility. The te&t specimens used for the cross-tension test are
strength of resistance spot welds: shown in Fig. 7.34 and the jig for carrying out this test is shown
in Fig. 7.35.
(i) Tension-shear test,
(ii) Cross-tension test,
(iii) Slug and shear test, I
(iu) Direct tension test,
(u) Peel test -1~-
Tension-Shear Test
The tension-shear test is the most widely used method for
-~~
,
determining the strength of resistance spot welds. The test I
specimen used is as shown in Fig. 7.33. It is made by overlapping
suitable sized coupons (strips) and making a spot weld in the (a I b I
Thickness up to 4·8 mm ( 0·19 in.l
centre of the overlapped area. A tensile test machine is used to
carry out the test. Fig. 7.34. Test specimens for Cross-tension test.
472 Welding Engineering and Technology Weldability and Weldability Test'! 473

welds. The first spot weld is then drilled out or isolated by making
a saw cut across one thickness of sheet, and the test pieces pulled
in a testing machine. A satisfactory weld must fail by the slug
being pulled out of the sheet. However, as the thickness of the
sheets or strength increases, the weld will fracture by shearing
across the nugget at the interface. The breaking load is quoted in
kgs/spot ..
Direct Tension Test
The direct-tension spot weld test is used to measure the
strength of the welds for loads applied in a direction normal to the
spot weld interface. This test is used to determine the relative
notch sensitivity of spot welds and it can be applied to ferrous and
non-ferrous alloys of all thicknesses.
Two types of test specimens are used for direct tension test.
One is the cross-tension specimen (Fig. 7.34) which can be used
for all alloys and all thicknesses. When the sheet thickness is less
than 1 mm, it is necessary to reinforce the specimen to prevent
excessive bending.
center

Fig. 7.35. Jig for cross-tension test on sheet thickness up to 4.8 mm.
Slug and Shear Test
When the sheet thickness is less than 1 mm slug and shear
test is made. This test is performed by pulling apart in tension

1
the two parts of a spot welded test piece as shown in Fig. 7.36, so

+
+ 1- -f-tl -.L
b Fig. 7.37. U-test specimen for Direct Tension test of a spot weld.
The second one is the U-specimen shown in Fig. 7.37. The
U-channels are spot welded back-to-back to form the test
~zX mm
specimens. The specimens are assembled on filler blocks with pull
taps for applying a tensile load to the spot weld through the pin
r3xa
Ja~~a Ja ja/zL
.---~o;;;----
~
..-- -- _._~ . - -]
connections. The maximum load, which causes the weld to fail
either by pulling a slug or by tensile failure across the weld metal
I '<.., I
Saw cut interface, is measured. The direct tension load is normally less
Fig. 7.36. Test specimen for slug and shear test of spot welds.
than the tension-shear l(jad for the same size weld and alloy.
The ratio direct-tension load/tension-shear load is a measure
that a slug of metal is pulled from one of the sheets. The mean
diameter of the slug, for spot welds, must not be less than the of the notch sensitivity of the spot weld. This ratio should be
square root of the sheet thickness. This test is carried out on a greater than 0.5 for the spot weld to be considered ductile or
strip 25 mm wide consisting of two sheets joined by three spot having low notch-sensitivity. Table 7.2 shows typical ratio ranges
474
- -------_._-------_._---- Welding Engineering and Technology
Weldability and Weldability Tests 475
for several commonly spot welded alloys. The values of 0.30 and
lower are indicative of brittle or notch sensitive welds. values of cause may be assessed by visual examination of the fracture fae es.
0.50 and above are typical of welds in low carbon steel, austenitic However, high strength or thicker specimens may fracture at the
stainless steel, and nickel-base alloys, which are less notch interface without producing a plug or nugget. Also, welds in such
sensitive. specimens are tested by twisting the test pieces past each other
Table 7.2. Ratio of Direct-Tension to Tension-Shear
Test Loads for Different Materials

~s. 2.
1.
f!0'

a.
Low ('..arbonsteel
r-~=~_.~~':~ia.l --~
Medium carbon steel (0.2% C)

High strength loVIT

alloy (HSLA) steels
F~
I
0.60--0.99
0.18--0.21Rang<

0.21--0.28

5. F'erritic stainless steel O.25-0.~i3

4. Austenitic stainless steel I 0..55··--0.82
6.. Aluminium alloys i 0..37--043

_____8.
7.
._._. J Titanium
Nickel .__ ..alloys
alloys
__ . __ .
I
..__ ....__ .__ ._._ ..... .J.__....
I _. O.27....·..·().52
0.71..-0.81 .._

The direct-tension test is, however, not normally used in
production control because it requires elaborate steps in preparing
test specimens.
Peel Test
A variation of the direct-tension test is the peel test which is
commonly used as a production control test. The steps involved in
carrying out this test are as illustrated in Fig. 7.38. In this test if
1 required by various boiler and pressure vessel codes, shows up
Fig. 7.as. Steps involved in Peel test; 0) grip tebt pi<,cein a vice, (2) Bend the
test specimen, and (3) Peel the pieces apart with pincers.
the weld nugget tears out of one or other of the sheet it may be
regarded as sound but if it splits open then it is defective and the
Fig. 7.39. Peel test by torsional shearing of a spot weld
in thicker or stronger material.
by holding one of them in a vice as shown in Fig. 7.3.9. In a
satisfactory weld test, the size of the plug or nugget is measured
or correlated with weld sizes having known strengths that are
produced by satisfactory production weld schedule.
7.5.2. Nick-Break Test
One of the methods of testing a fusion weld is to cut a strip
about 20 mm wide at right angle to the weld axis and make a saw
cut down the centre line of the weld 3 to 6 mm deep. By holding
one-half of the specimen in a viee, Fig. 7.40 (a) and giving the
other half a sharp blow with hammer or by bending in a machine
as shown in Fig. 7.40 (b), the weld is broken. This test which is
any centreline defects, such as lack of fusion, gas pockets, slag
inclusion, and the degree of porosity in the weld bead. The defects
are generally examined by visual examination, and should not
have a length of more than 3 mm individually.
Fillet welds may be similarly tested by notching and bending
as shown in Fig. 7.40 (c). The standard length for a specimen in
the fillet nick-break test is 100 to 150 mm. The test is carried out
by simply applying force in the back of the fillet, crushing the angle
flat. One requirement of the fillet nick-break test is that there
should be no tack welds on the other side of the fillet weld. One
test. tack weld will have enough holding power to invalidate the
small
 Weldability and Weldability Tests 417
Welding Engineering and Technology
476
Seam Metal

Excess weld metal

and penet ration
bead left intact

rI
Seam
oj
Seam
II
weld weld

(0)
(bJ
Ic I
Fig. 7.40. Different forms of Nick-break test: (a) for butt-welded thin component,
(b) for butt-welded thick plates, and (c) for a fillet weld.
7.5.3. Pillow Test for Resistance Seam Welds
The pillow test is the most common test for determining the
strength of seam welds ; the procedure for the same is
schematically illustrated in Fig. 7.41. It involves seam welding
'~'
r
Hydraulic
pressure
Fig. 7.41. Schematic illustration of pillow test for seam welds.
two pieces of metal to enclose a cavity. An appropriate pipe fitting
is either put on with a nipple or welded on to the two pieces that
were seam welded together. Hydraulic fluid or air is pumped
through the fitting, expanding the cavity into a pillow shape. The
pressure at which the pillow bursts is recorded and compared to the
fracture strength of the base metal. Failure should always occur in
the base metal and not in the welded seam. If the weld seam
fractures, then the weld will not support that particular metal.
7.5.4. Bend Tests
The quality of the weld, in terms of ductility of the weld metal
and HAZ as well as tests for opening of defects particularly lack
of side wall fusion (side bend), root fusion, and penetration of
welded joint, are most frequently checked by means of a bend test.
Such tests are sub-divided into three types:
(i) free bend test,
(ii) guided bend test,
(iii) controlled bend test.
Bend specimens may be longitudinal or transverse to the weld
i axis and may be bent in simple three-or four point free bending
as shown in Fig. 7.42. The test may be carried a step further by
\
flattening the arms together in a press to complete a 180Q bend as
! shown in Fig. 7.42 (c, d).
\
478 Welding Engineering and Technology
_--..L.- __ ~.
Forme~
Weldability and Weldability Test~
,. .-------- Test 479
piece -

OJ , Specimen
G'

~ i~//
""- I'

.~}--_ ..~._- - +-
T ~/we\d
~LT__
__ [)_+ __ .....
_ '~3~lp~r t-_
(a) Three - point bend test

Stopper

Max. rod. 10% a Fig. 7.44. Controlled bend test details.

\~ Roller support A = 32mm When t:::; 13 mm

I
I
i
to avoid cracks

or greased
shoulder s
A=SOmm\vhen t> 13mm I

( b I Initial bend for free - bend specimens

( Four .. point bending) Ie) Final bend for free ..
bend specime ns
Fig. 7.42. Free bend test; (a) three-point bending, (b) four point bending, ! Face in
and (e) press bending to achieve final U-shape. tension
II
1'1 a
L- ,1 - --..J but 30 mm min.

'--'-'~l ,_-
Shoulder
! l' ,-
Plunger

Roller (alternate I
I Max.
(a)
rud. 10%a
to avoid tracks
Transverse face
SUpports
bend, normal.

.. I ~i-~\

Die

~ //
\

i J.-X/
'.

'weld ~/T
-t ';

r min.

Fig. 7.43. Guided bend test.

 480 Welding Engineering and Technology
Weldability and Weldability Tests
Du~ to inhomogenity of the joint, there is a tendency for free 481
bend test specimens to take up an irregular shape, so that .the
actual radius at various' points differs from the specified value. Side Bend Test: Very thick welded plates are difficult to bend
This defect is overcome to some extent in the guided bend test, in the normal manner, and under these circumstances it may be
Fig. 7.43, but can be avoided most effectively by the use of the both permissible and desirable to make a side bend test. A slice 3
controlled bend test as shown in Fig. 7.44. to 6 mm thick is cut at right angles to the plate surface and to the
weld axis and is then bent in the usual way as shown in Fig. 7.46.
The top and bottom surfaces of a welded plate are frequently
designated as the face and root surfaces respectively. In face bend Side bend testing shows up lack of side wall fusion very well
test, face side is under tension and for the root bend test, weld root but is somewhat less sensitive to face and root defects than the
is in tension as shown in Fig. 7.45 (a and b). normal type of bend test. The relative locations for cutting face,
root-asand
are side-bend
shown in Fig ..test
7.47.specimens in a thick butt welded plate
To be acceptable, a test piece must be capable of bending
through 1800 over a former of given radius (for example, 1~plate
thickness for carbon steel) with no more than surface fissures. The
bend test requirements for steels of different strengths are as
given in table 7.2.
Table 7.2. Bend Test Requirements for
Three-Point Bending For
face bend test
S. No. 432-5254326.2
525-618
Under 5.2
4.2 tt 4t
3t
2t supports Root bend tes t
rms
ss, tof plate Clear space between Side bend test
3 mm1.along the specimen
Minimum .peoifi,d I Dia~", of f~m" in

Weld joint

No cracks allowed to exceed 1.5 mm across and 3 m g

Former
Fig. 7.47. Relatived location of face, root, and side bend test
Weld normally specimens in a butt welded thick plate.
dressed flu s h
10mm Bend Testing of Pipe Welds: The location of tensile and
bend test specimens for welded pipes of 50 mm, 125 mm, and 200
mm diameters are shown in Fig. 7.48. Machine the minimum
amount needed to obtain plain parallel faces over the 19 mm wide
reduced section for tensile test pieces while the test specimens for
bend tests will be 230 mm long and 38 mm wide for 125 mm and
200mm
50 mmdiameter
diameterpipe.
pipes, and 150 mm long and 25.4 mm wide for

Details of Tongue and Bend pieces for testing circumferentiaJ

Supports welds in pipes, and TEMA Test for bend test on tube-to-tubeplate
Fig. 7.46. Side bend test details. joint are shown respectively in Fig. 7.49 and Fig. 7.50.
 6 G positions W

~
tIl to
('l)
='
:3 0.. ::
~fIl
..•.
I~
4
=
b:l fIl itS· -(Top
'-'
•.oq
~3.•.
-e. ='
('l)
..::s '1
('l)
~ •III
tIl
='
o'tl
'1 0
0~. ~('l)
::l'l
0("to
o0:2";;1
Q.0 '1
'1
~r-t-
III ='
III ('l) ~
;3 I
of pipe for SG and
/ Tensile
('l) ('l) '1tIl 'tl
=' 0'1 •.•.
~
('l) <
0..
f'oI.
...,
='
:::; ('l) tIl. tIl
g0oq
("to
"'1 oq Jl r-t- ("to
-('l)
0..
r-t-
l!J tIl
~
tIl ,-....~.-..
'1~e:P-."5l
~0:31llS.
•.•••.•• ('l)
::!.
tIl
;>::l'l
~c-
tIl
0l!J III
><
('l)
tIl
r-t- ~ .
: --0 1ti
~tIlO~
r-t-
~ 1-" ('l)
E;- /
,
Top of pipe for SG and
6 G positions
/ Tensile '<
Side bend '

Pipe wall over 10 rom

Pipe wall 10 rom or less
Top of pipe for 56 and
66 positions

1,2.3.4 - Bend test pieces

Fig. 7.48. Relative locations of tensile and bend test specimens from welded
pipes of 50 mm, 125 mm and 200 mm diameters.

W" 38 mm max.
Surface etched
over weld area Tongue bent by
All corners of tongue
pressure of blows
dressed to radius 1.6 mm
with file

dressed. to radius
All corners 1·6 mm ~
of tongue I~
vith smooth file ~
.•.
'<
>-i
~
'"
~
. I " I. Reinforceme nt

M"
I':30· .
QPprox. I
'J \
\
I '- _•• I
/
/,
rI removed by
filing or
grinding

,
.
I
, "I '1--- III
II . I
View on A-A
Saw cut or
flame cut ~0 I
Outside dia.
•.
I
Tongue prepared for
bend test on each half
o
lup to Qnd including
113 mm)
Hethod of bending tonqui

Fig. 7.49. Bend test on pipes by tongue bend pi~s and the procedure ofi;@sting
484 Welding Engineering and Technology
No cracking
after bending
Fig. 7.50. TEMA test procedure for testing tube-to-tube plate weld.
Elongation is considered to provide the most reliable and
reproducible data. Elongation occurring in the outer fibres, at the
outer radius, is determined from gauge marks that are scribed,
inked, or shot etched on the specimen prior to bending or
approximated from the following formula.
t
e = (2R + t) x 100 ...(7.15)
where, e = elongation, %
t = initial plate thickness, mm,
R = radius of curvature of the inside surface
of the bend, mm.
For testing relatively thin material « 4.8 mm) bending the
specimen over a succession of decreasing radii dies provides a good
measure of ductility. The radius of the last die used before failure
is recorded as the bend radius. In each case, the specimen is forced
to the bottom of a standard V-blockby progressively smaller radius
dies.
Transverse bend tests are useful in qualifying welders and
welding operators because they quite often reveal the presence of
defects that are not detected in tension tests. However, the
transverse test specimen suffers from the same weakness as the
transverse weld tension test in that non-uniform properties along
the length of the specimen can cause non-uniform bending. For
example, very narrow welds such as those made by electron beam
welding may contribute very little information on weld ductility.
Thus, the transverse bend test is sensitive to the relative strengths
of the weld metal, the HAZ, and the base metal.
Several different types of problems can develop in transverse
bend specimens. An overmatching weld strength may prevent the
weld zone from conforming exactly to the bend die radius and may
Weldability and Weldability Tests
485
force the deformation out into the base metal, causing less than
the desired elongation of the weld. With an undermatching weld
strength, the specimen may tend to kink in the weld, resulting in
a sharper bend radius in this area and thus more severe elongation
in the weld. When transverse free bend tests are used for welded
quenched and tempered steels, excessive stretching of the over-
tempered region of the HAZ may lead to early failure in tliis area
before full extension of the weld metal is achieved.
The above problems of weld mismatch can be avoided by using
longitudinal bend specimens in which the weld runs the full length
of the bend test specimen ; the bend axis is then perpendicular to
the weld axis as shown in Fig. 7.51. In this test, all zones of welded
Max. rod. 100/0 a
Supports
Fig. 7.51. Three-point bend testing of longitudinal weld specimen.
joint (weld bead, HAZ, and base metal) are strained equally and
Simultaneously.
in This test is generally used for evaluations ofjoints
dissimilar metals.
It should be noted, however, that tests aimed at detecting gross
weld flaws that are inherently oriented parallel to the weld axis-
such as incomplete fusion, inadequate joint penetration or
undercut-necessitate transverse loading tests. Side bend test
specimens strain the entire weld cross-section and are thus
.especially useful for exposing defects near mid-thickness that
might not contribute to failure in face- or root bend tests. They
are also used for relatively thick sections (over 19 mm), since in
these tests all of the weld thickness may be included in the test.
Bend testing is effective in disclosing brittleness, cracking or
gross defects e.g., lack of fusion, or large slag inclusions in the
weld, or cracking in the HAZ. It will not normally show up
microcracking.
 Weldability and Weldability Tests 43'
486 Welding Engineering and Technology
V-notch in the test bar while the Charpy test piece is 55 mm long.
7.5.5. Impact Testing Both test specimens, i.e., for Charpy and Izod tests, are 10 mm
Impact testing determines the relative toughness of a square.
material, where toughness is defined as the resistance of a metal These days virtually all impact testing of welds is made usi.ng
to fracture after plastic deformation has begun. In an impact test, Charpy type machine and specimen because Charpy testing has
this plastic deformation is begun and finished by the swing of a the advantage that low temperature tests may be made by cooling
weighted pendulum which strikes the test piece as it swings the specimen alone, where as for Izod testing both specimen and
through its path along an arc. The energy that is required to grips must be cooled.
fracture the test piece is recorded in Joules on the scale. The Different types of impact test specimens used for Charpy
tougher the material, the higher the energy absorbed by the test testing are shown in Fig. 7..'52.In British practice a 2 mm deep
piece. V-notch is cut in the specimen but USA standards, which are
Two major tests for determining impact toughness are the Izod frequently used in refinery and petrochemical construction call for
test and Charpy test. Both of these test methods use th~ same a 5 mm deep U-notch or keyhole notch as show in Fig. 7.52
type of machine and both yield a quantitative value in joules. The (b and c). The notch may be placed with its root in the weld, the
Izod test specimen is. generally 75 mm in length with a small fusion boundary, or the HAZ, but most frequently it is placed in
the weld metal itself.
Testing temperature is important and the notch must be cut
0.25 mm exact to minimise scatter of results. Inspection of fracture is
carried out for flaws and fracture surfaces for type of fracture i. e.•
lal fibrous or crystalline. Test can also indicate temper brittleness.
Impact test results from welds show even greater scatter than
)'00
.::ts·\y I· S5
L_~A mm --111O~~t~:jI8mm
;rm with wrought material due to the inhomogenity of the weld
structure. The grain size of the weld deposit, e.g., may vary from
run to run, or according to the position relative to the weld axis.

X2mm
~
l l6mm ..
~
(aIV-notch
I B~ ~
(b I Acceptance levels are, however, similar to those for the wrought
material.
Comparison Between Charpy and Izod Tests
or less Smm 1. Izod test specimen is loaded as a cantilever whereas in
I b IKeyhole notch
Charpy test it is loaded as a beam supported at two ends.
(c I 2. The notch in Izod test specimen is off the centre to
EJ-L
T' facilitate holding in vice whereas the notch is in the
---In~2mm I
=\ no___ -----I 5mm centre of the specimen for Charpy test. The Izod test
IcIU-notchIA.5.1M) specimen is longer for the same reason.
3. The V-notch in Izod specimen faces the load whereas it
is opposite to loading in Charpy test.
l () \ U_L
T
ld I
The results of impact tests give a better indication of
resistance of steel to failure under triaxial stresses and
3mm
=ll='mm 10 ,U-no"h I D.VM' low temperatures than do static tensile tests.

-0-
-1 \-2 mm
1'-----_('\__._ _I
Ie I U-notch (Mesnager I
~-L
I:=.:=IT
2mm
(e I Welded ships and pressure vessels are subjected .00
triaxial stresses and embrittling service temperatures in
much the same way as notched specimens. This condi-
tion exists partly as a result of the design of thestruc-
ture and partly from defects in improperly made :welds.
Fig. 7.52. Different types of charpy impact test specimen.
488 Welding Engineering and Techno'ogy
7.5.6. Hardness Tests
Hardness is the ability of a metal to resist penetration, to
resist abrasive wear, or to resist the absorption of energy under
impact load; accordingly these can be thought of as penetration
hardness, wear hardness, and rebound hardness.
Hardness testing may be used in weld evaluation either alone
or to complement information gained through the tension or bend
tests. Routine testing methods for the hardness testing of metals
are well established. These include the Brinell, Vickers and Knoop
hardness tests, which use the area of indentation under load as
the measure of hardness, and the Rockwell test, which relates
hardness to depth of indentation under load.
Hardness measurement can provide information about the
metallurgical changes caused by welding. In constructional steels,
e.g. rapid cooling from HAZ temperatures may cause the formation
of martensite of much higher hardness than the base metaf. In
HAZ areas where the maximum temperature is lower, however,
the hardness may be low'er than the base metal due to tempering
effects. Welding of cold worked or age-hardened materials may
result in significantly lower HAZ hardness due to recovery and
recrystallisation or overaging. Hardness values in a welded joint
are usually sensitive to such conditi.ons of welding, as the process
used, heat input, preheat or interpass temperature, electrode
composition, and plate thickness.
Hardness testing of welds is performed on ground, polished or
polished and etched cross-sections of the joint area. Indentations
are made in the specific areas of interest, including the weld centre
I the case of pipelines. High hardness coupled with hydrogen can
• Surface traverse
• Quarter-thickness
traverse
Center traverse'
I failure due to localised plastic flow occurring around small
I inclusions. The brittle fracture, on the other hand, shows cleavage
Weld
center line
traverse
Fig. 7.53. Typical hardness traverses for a double-vee butt welded joint.
Weldability and Weldability Tests 489
r
I line, face or root regions of the deposit, the HAZ, and the base
metal. Frequently traverses covering all of these areas are made,
:with indentations spaced at regular intervals along the line of
traverse as shown in Fig. 7.53. Traverses along the weld centre
line from root to reinforcement or weld surface may be used to
determine multipass effects, such as the presence of alloy build-up
in successive passes of submerged arc welds.
Which hardness test is used depends primarily on the
hardness or strength of the material, the size of the welded joint,
and the type of information desired. The Brinell test produces a
large indentation, typically 2 to 5.6 mm in diameter and is thus
suited only for large welds, as in heavy plates. Rockwell test
produces a much smaller indentation more suited for hardness
traverses, but even this test is macroscopic in scale in the sense
that the indentations may be larger than the precise area of
interest (e.g. the fusion line or the coarse grain region of the HAZ).
The Vickers and Knoop tests make relatively small indentations
and are thus well-suited for hardness measurements of the
various regions of the HAZ for fine scale traverses.
Attempts have been made to correlate hardness with strength
for certain materials, however such correlations should be used
with caution when applied to welded joints due to the hetero-
geneous makeup of the joint.
Hardness has also been related to service properties of
weldments, and in some cases maximum weld or HAZ hardnesses
have been specified. This is particularly important for weldments
exposed to hydrogen bearing environments such as sour-crude in
and has caused cracking. However, care should be taken to
overcome the problem of non-standardized methods of hardness
measurements for weldments, since the value obtained can be
quite dependent on the test methods used, including the size of
indentation as well as the location of the indentations in the joint.
7.5.7. Fracture Toughness Tests
It is usual in engineering practice to refer to fractures as ei ther
ductile or brittle. The ductile failure is characterised by the
formation of dimples on the surface caused by cup and cone tYJJe
type failure which appears comparatively flat and featureless and
exhibits little or no evidence of plasticity. Fracture toughness is a
property of the metal which defines its resistance to brittle fracture.
.•.
490 Welding Engineering and Technology
Materials in general are divided' into three categories viz.,
(i) Brittle,
(ii) Elastic/plastic,
(iii) Ductile.
Fracture toughness testing involves only the brittle and
elastidplastic materials.
Fracture toughness study of brittle materials is now fairly well
establi8hed by what is known as Linear Elastic Fracture
Mechanics (LEFM) and critical crack lengths for given stress
conditions can be accurately predicted to know of the service
behaviour of a component while similar quantitative analysis
procedures for the elastic-plastic region are not yet established.
However, LEFM covers a very limited range of materials while a
very wide range of materials come under the category of elastic-
plastic materials which are widely used in engineering
fabrications. It is, therefore, imperative to determine the fracture
toughness values of materials which lie in the elastic-plastic
region. In the absence of any definite quantitative analysis the
necessi ty to determine these values has resulted in the use of
methods for characteris.ation of elastic-plastic fracture behaviour
which are largely empirical and rely on either experience or
service failure correlations ..
In LEFM the critical value of stress intensity factor (K) is
referred to as the fracture toughness, the corresponding fracture
toughness value for the materials of elastic-plastic region is
referred to as J c.
A number of test methods are used to determine the fracture
toughness value for the elastic-plastic region but the following four
methods are more popularly employed.
(i) Charpy V-notch (Cu) test,
(ii) Drop weight nil-ductility temperature (NDT) test,
(iii) The dynamic tear (DT) test,
(iv) Crack tip opening displacement (CTOD) test.
While the Cu and Drop weight tests are ASTM standard
methods, the DT test is a military standard and the CTOD is a
British standard test. All these tests are described in chapter on
Fracture Toughness.
7.5.8. Fatigue Tests
Components subjected to repeated cyclic stress have a
tendency to break at a stress considerably below their ultimate
tensile strength. This type of failure occurring under many
491
Weldability and Weld ability Test,>
repititio~s of load is called Fatigue failure. The stress under which
such a £:ailure takes place is called fatigue strength. Thus, while
designi:t>.gsuch components U.T.S. of the material must be
replaced by the fatigue strength. Fatigue life, on the other hand,
is the number of cycles after which failure occurs with the given
loading conditions.
Fatigue failure is the most common type of failure in service
and accounts for almost 90% of failures of welded components.
Fillet welded construction is particularly troublesome in this
regard. Examples of fluctuating service load include almost all
major welded fabrications, e.g., ships, bri'dges, pressure vessels,
vehicle chasis, off-shore platforms, cranes and earth moving
equipment, crankshafts, railroad wheels, gears of all kinds, power
generation turbines, and many others such as rotating load-
bearing parts in automobiles, trucks, aircrafts, and power plants.
Fatigue failure is progressive. It originates in highly localised
region of stress concentration, at design or fabrication disconti-
nuities. There are many factors that affect fatigue behaviour of
weldments such as joint configuration, stress ratio, material
properties, welding procedure, PWHT, loading conditions, residual
stresses, weld reinforcement geometry, and· the service
environment.
Out of the above listed factors, two have been identified as
being particularly important and most likely to govern fatigue
strength of welded joints, viz., stress range and type of design
detail. The service environment can have an appreciable effect on
the rate of growth and propagation of surface-located fatigue
cracks; e.g., the presence of sea water can greatly accelerate the
crack growth rate in high strength steels and certain titanium
alloys. Other factors such as type of material, minimum stress
level, rate of load, rest periods, service temperature, and other
miscellaneous factors have much less effect on the fatigue strength
and life of structural materials.
Fatigue tests are designed to measure the effect of cyclic-stress
loading on the welded joint. Specially designed test bars and
sometimes complete structures or machine parts and fully welded
sections are tested. A typical machine for fatigue testing with a
pulsating axial stress is shown in Fig. 7.54.
The number of cycles of loading and unloading (N) necessary
for failure to occur by cracking is recorded along with the stress
level(s) and the nature of the stress i.e. tensile or compressive, etc.
Fig. 7.55 shows the results ofa typical fatigue test after a number
of samples have been tested at different load and stress levels until
492

0110 Lever
Welding Engineering and Technology
i WeldabiIity and Weldability Tests

into different classes, viz., A, B, C, D, E, F, G, and W. These types

493

of structures are discussed in detail in chapter on Fatigue of

weldments.
Upper pull head
7.5.9. Corrosion Tests
Tes t specimen Turnbu'ckle
Variable eccentric
Corrosion is a chemical or electrochemical reaction between a
Motor metal and its environment, usually resulting in deterioration in
Lower pull head the properties and/or appearance of the metal. A welded joint will
normally resist corrosion to the same extent as the metal being n
joined.
In addition to the well known uniform attack, there are a
Wilson Fatigue Testing Machine variety of localised forms of corrosion and the oneS particularly
Fig. 7.54. Schematic of a set-up for fatigue testing under a pulsating axial stress. relevant to welded joints are:
1. Galvanic corrosion,
350 2. Crevice corrosion,
3. Intergranular corrosion, and
4. Stress corrosion.
6:280 Galvanic Corrosion may occur wh~n two different metals in
~ contact are exposed to a conductive solution. A difference in
~i2 45
QI electrical potential exists between different metals resulting in
<- Endurance
Vi 210
o n 0 flow of current that leads to the corrosion of the anodic areas.
I Limit Thus, galvanic corrosion can occur when a weld filler metal is of
l:J
175 significantly different composition than the base metal and
occasionally might occur as a result of differences in behaviour
140
5 large cathodic region
104 10 106 107 108 Small
anodic region
N - Cycles of Stress
Fig, 7.55. Fatigue test results shown as S-N line, i.e.,
/ \
stress versus the number of cycles before failure.
they failed. This typical curve or S-N line shows that the fatigue
strength, i.e. the stress at which the metals fail drops as the (a)
number of stress cycles increases. The greater the number of
cycles, lower the fatigue strength. regions
Small cathodic region
Fig. 7.55 also shows that below a certain stress level the A ~~ <-A
material does not fail by cracking, no matter for how many stress
cycles the component is loaded. The maximum possible stress level
with no failure, no matter how many cycles are repeated, is called
the endurance limit of the material. Thus, a welded detail will
have indefinite service life if it is loaded up to or below its Regions where attack may be serious
endurance limit. (b I
Because the type ofjoint has far reaching effect on the fatigue

l
Fig. 7.56. Galvanic corrosion of a welded joint, when; (a) weld metal is les8
strength of a component therefore all welded joints are grouped noble than the base metal, and (b) weld metal is more noble than the base metal.
494 Welding Engineering and Technology
Weldability and Weldability Tests 495

"
between cast weld metal and the wrought base metal. Fig. 7.56 Weld metal
•
shows galvanic corrosion effect on weldments .
Crevice Corrosion is a form of local attack caused by
concentration cell effects within a crevice. The concentration
difference may be of electrolyte composition inside and outside the
a.
I=-_-_0-
/ -I All parts
Uniformly
corroded

Base metal
crevice. Crevices in welded joints occur in a variety of ways-
surface porosity or cracking, slag or weld scale deposits, b. Base metal
inadequate joint penetration, underc~t, and design defects.
1-----..-0"'- -_-_I corroded

=
Like all forms of localised attack, crevice corrosion does not
occur in all metals or their combinations. Some materials are more
c 1= ==t:L._1 Weld metal
corroded

susceptible to it than others. For example, stainless steels and

aluminium are often susceptible to crevice corrosion since they
d 1
__ D~l Base metal' high
temperature HAZ
corroded
rely for their intrinsic corrosion resistance upon an abundance of Base metal' low
e
oxygen in their environment. temperature HAZ
corroded
Intergranular Corrosion. Grain boundaries are sometimes
preferentially attacked by a corroding media. Corrosion generally Fig. 7.57. Schematic representation of different types
of corrosion in welded joints.
occurs because the corrodent preferentially attacks the grain
boundary phase or a zone adjacent to it that has lost an element the density of surface material,
D =
necessary for adequate corrosion resistance, for example, deple- T=
duration ofthe test.
tion of chromium at the grain boundaries in stainless steel. The use of this formula is ideally suited to the conditions
Austenite stainless steel can exhibit this problem when heated to shown in Fig. 7.57.
about 650°C because chromium carbides precipitate at the grain
boundaries when steel is heated to 'that temperature. Test specimen: The test specimen used will vary with the
size and type of product being welded, and the type and size of
Stress Corrosion is caused by the combined effect of tensile weld. The simplest specimen shape is rectangular as shown, for
stress and a corrodent in some instances. Most alloys are suscep- butt welded plates, in Fig. 7.58.
tible to this problem, however the number of alloy-corrodent
combinations that cause it are relatively few. Stresses that lead y
to such corrosion arise from cold working, welding or PWHT, or
they may be externally applied during service.
7.5.9.1. Corrosion Testing of Welded Joints
A welded joint may corrode 'uniformly over its entire surface
or may be susceptible to varying degrees of preferential attack as
shown in Fig. 7.57. The most common metho~ of evaluati~g.
Corrosion resistance is to measure the weight lost/during exposure
to the corrodent and to convert this to an average corrosion rate
using the following relationship.
KW
R =ADT ...(7.16) L
where, R = the corrosion rate in depth of attack/time,
K:: a constant whose value depends upon the unit used,
W:: weight lost by the specimen during test,
A = total surface area of the test piece,
r
z

Weld metal
Fig. 7.58. Typical butt welded specimen for corrosion test.

I
496 Welding Engineering and Technology
The length of specimen (X) will depend upon the availability
of test material and the size ofthe container holding the corrodent.
The dimension (Y) must be sufficiently long to ensure that
material at point C has corrosion resistance identical to unwelded
base metal ; the larger the Y dimension the greater would be the
galvanic corrosion. The Z dimension or thickness of the test piece
will decide the heat input rate and the size of the HAZ. Larger the
Z dimension, the greater the amount of end grains exposed which
will result in excessive end grain corrosion that may introduce
error into average corrosion rates calculated from weight loss
measurement. This is usually prevented by painting the edges
with corrosion resisting rubber based paints.
Test Procedure. Machine or grind all surfaces flush, leaving
no undercut or surface imperfection and polish to a 120 grit finish.
Avoid overheating.
The test results will be based on :
(i) weight loss measurements,
(ii) macroscopic examination,
(iii) measurement of electrical resistance to detect general
and intergranular corrosion,
(iv) measurement of depth of attack,
(v) ultrasonic examination to measure localised surface dis-
continuities and to evaluate intergranular attack.
An alternative approach to the above testing method is the
use of Electro-Chemical tests to evaluate susceptibility to
corrosion attack. One such well established and accurate method
for determining corrosion resistance of stainless steel is Electro-
chemical Potentiokinetic Reactivation CEPR)test.
7.5.10. Creep Rupture Tests
When a load, less than required to produce immediate
fracture, is applied to a material at an elevated temperature, the
material will experience gradual elongation and will ultimately
fracture. The gradual elongation or plastic deformation at higher
temperature is called Creep, while the eventual fracture is called
rupture. Creep can result in putting a component out of shape
much before rupture takes place. If that component is a gear or a
jet engine compressor vane, it may not take long for the creep to
ruin the entire engine.
Elongation
roughly or creep
three stages, as progresses, at a7.59.
shown in Fig. constant temperature,
On initial in
application
of load, the creep or elongation rate is relatively high but conti-
nually decreasing with time. This is termed as the period of
497
Weldability and Weld ability Tests
Rupture
co
Second
stage
&c I stage
First
.2
.....
~
o
Time
Fig. 7.59. A typical creep-rupture elongation curve.
primary or first stage creep. The second stage of creep is generally
an extended period during which a constant and minimum creep
rate occurs. In the third stage, the creep rate again increases,
generally as a result of the onset of necking, and culminates in
rupture of material. Increasing load or temperature shortens the
time for each stage and reduces the rupture life.
Usually creep rupture is a phenomenon associated with exotic
materials. The industries most often facing this problem include
aircraft industry, parts of heat treating furnaces, welded boiler
tubes for a fossil fuel or nuclear power plant. Long term creep
failure can be due to difference in creep ductility between weld
and base metal. There are creep-rupture tests to determine as to
which metals are more resistant to the creep effect than others.
Test Method
The test specimen used to determine creep-rupture strength
is similar to that used for the already described tension tests, It
may be either all-weld-metal or a specimen taken transverse to
the weld joint axis containing some or all of the structural
components of the joint. The advantages, disadvantages, and
limitations of each orientation are similar to that of the tension
test. However, transverse creep tests are seldom conducted since
the variations in elongation in each welded joint region make the
total elongation measurement meaningless.
The test is conducted in an appropriate furnance and load is
generally applied in the form of constant deadweights. In the case
of a creep test, extensometers of some form are attached to the
specimen so that the elongation can be monitored during the
course of the test. In the rupture test, time to failure at a given
applied load is the principal variable measured, although final
elongation and reduction of area are usually determined.
-------------------------~"" "-"---~-~-

498 Welding Engineering and Technology Weldability and Weldability Tests 499

Test Data Reporting and Extrapolation This like parameter methods all substitute temperature for
Creep and rupture data are most often reported as log-logplots time and permit determination of long time strength from shorter
of stress versus time to attain a specified percentage creep time-high temperature tests. This results in great time and cost
elongation or time to rupture at a constant temperature. Such savings, and these methods have been developed to the point that
plots are. generally straight lines as shown in Fig. 7.60. extrapolation can be very accurately achieved.
Certain metals such as high-alloy nickel, molybdenum, and
iron-chromium alloys are much more resistant to creep rupture at
VI
T, <T2<T) = Temperature high temperatures than are carbon steels. These special super-
~
VI

'- alloys also have much higher tensile and yield strengths than
.•...
VI carbon steels at very high temperatures. Collectively they are
0'\ called high-strength high-temperature alloys. These superalloys
-'o can be welded but may need special procedures and precautions.

Log time to rupture

Fig. 7.60. A typical set of rupture data at three temperatures.
Rupture Strength is the maximum stress level needed to cause
a piece of metal to fracture in a specified time at a certain high
temperature. Rupture strength data are generally needed for
equipment that must operate for long periods of time, such as
refinery or power generation equipment. Typical design lives for
these applications are 100,000 to 300,000 hours (i.e. 11 to 34
years). Few tests are conducted this long. In fact even 10,000 hours
tests require over a year and are conducted only on a limited basis.
Results are usually reported as the stress necessary for rupture
in 10, 100 or 1000 hours of continuous loading. The temperature
is always stated along with the stress level. For example, the
rupture strength of an alloy might be 10500 N/cm2 at 1025°C after
100 hours of exposure.
To determine the strength at the design life, some form of
extraploating will have to be used. Perhaps the most widely used
method of extrapolating creep-rupture is predominantly a
diffusion controlled pnenomenon as the straight line log-log plot
indicates. Thus, it can be expressed as an equation from which
various time-temperature parameters can be developed. One such
equation developed by Larson and Miller is,
P = T (C + log tr) ••. (7.17)
where, P = the parameter against which stress is plotted,
C = constant (generally 20),
T = absolute temperature, degrees Rankine,
tr = time to rupture, hours.
 r Weldability of Specific Materials

and copper and silicon each do not exceed 0.60%. Classification of

the carbon steels is principally based on their carbon contents. The
50]

groups comprise low carbon steels (up to 0.30% Carbon), medium

carbon steels (0.30 to 0.50% Carbon), and high carbon steels (0.50 .
to 1.0% Carbon). The first group is often sub-divided into very low
carbon steels (up to 0.15% C) and the mild steels (0.15 to 0.30%
C). Typical uses of these steels along with hardn~ss and
weldability ratings are listed in Table 8.1.
EIGHT Table 8.1. Classification and Characteristics
of Carbon Steels
Weldability
S.
carbon I steel Common
of
Carbon
Medium0.30
Mild
High
Poor 0.50
Structural
(WiN)
0.15
dies,
(low
shapes,
strip,
Special
Machine
Springs, 256
toto
389
183 0.50
1.0and
sha-
Good
0.30
Fair
hydr
ing
gen
plates
parts to excelle
(preheat-
PWH1
welding 101 rails
Good Hardne.5S
and
sheet,
pes,
Typical tools
plates
welding and normally prehel
process,
.g
4. steels 2. carbon
3. 1. Low I 0.15
(%)
max and
ding PWHT
bars
electrodes
recommended)
required)
hydrogen
process
required; wel-
low
Specific Materials No. Name Content

8.1. INTRODUCTION
Although it has been stated in earlier chapters that all metals
are weldable, but some are far more difficult to weld than others.
-1--~-· 'Jypioal~", ] Wdding m~

In view of this, it is vitally important that the welding process and .It
procedures be considered when determining the weldability of a
particular material.
No doubt there are innumerable metals and their alloys which
are fabricated into desired shaped components and structures but
it is beyond the scope of this text to consider all of them. So, in
this chapter it is intended to discuss the salient features regarding
the weldability of more often used materials in engineering
construction. These include,
1. Carbon steels,
2. High strength low alloy steels, t
3. Low alloy steels,
4. Stainless steels,
5. High alloy steels,
6. Cast irons, 8.2.1. Welding of Low Carbon Steels
'7. Nickel and nickel alloys, Generally steels with carbon contents up to 0.30% are readily
8. Copper and copper alloys, joined by common arc welding processes and accou:q.tsfor the
9. Aluminium and aluminium alloys, greatest tonnage of steels used in welded structures. Typical
10. Magnesium and magnesium alloys, applications include storage tanks, structural assemblies, vessels,
11. Titanium and titanium alloys.
machine bases, earthmoving and agricultural equipment, and
Important aspects of welding these materials are described in general fabricated components.
this chapter ..
Steels with very low carbon content, usually up to 0.13% C,
8.2. WELDABILITY OF CARBON STEELS are good welding steels, but they are not the best for high speed
Carbon steels are alloys of iron and carbon in which carbon production welding. The low carbon content and the low
does not .usually exceed 1%, manganese does not exceed 1.65%, manganese content (up to 0.30%) tend to produce porosity. This-
502 Welding Engineering and Technology Weldability of Specific Materials

, ,,,, ,
S()3

/' ,,"
- / a:ic\/\/\./'\ /'
condition is usually corrected by modifying the weldingc: procedure
'
50
+- .:; += 70
slightly-usually by using a slower welding
c:
. oX u u
20 speed. However,
:;; 30
presence of some internal porosity has no detrimentalOL effect
>
c:
'Vi
C7'I
Qj c:
if theL-
Carbon
"g40
~
Qj
e:
equivalent
SO
60
.D
;j
<t
-c
~
u
~ 80
90
-c
..c:
50
~
c:
Qj
0'40
"E0·50
Qj
0·70
"g350
400
150
0·60
500
.~Qj~250
"'- 100
450
300 sensitivity
/
//
"" ;'
Crack .••.. -_
'"
Q/
Q/
'-
Y', <
in 0·30 C7'I
Q/
service requirements of the assembly, standard high speed Bend angle 40 ~
-0
welding procedures may be used. o
.2
Steels with very low carbon content are more ductile and e
easier to form than higher carbon steels. They are used for 30 ~
applications requiring considerable cold forming, such as
stampings or rolled or formed shapes.
Mild steels with 0.15 to 0.20 % carbon content have excellent
-ce
.~

20 ~C7'I
weldability. They seldom require anything beyond standard c:
o
welding procedures, and these can be welded with all types of mild ~c:
steel electrodes. These steels should be used for maximum
production speed on assemblies or structures that require 10 1::
ClJ
extensive welding. CI
o'-
Mild steels at the upper end of the low carbon range i.e. 0.25 ClJ

to 0.30% carbon steels with manganese 1.40% have very good o

. >
<C

weldability but when one or more of the alloying elements is on 0·80

the high side of permissible limit, the welds are susceptible to
underbead cracking particularly in fillet welds. With slightly
reduced speeds and currents, any of the standard electrodes can
be used for these steels. For section thickness up to 8 mm,
standard welding procedures are applicable.
If some of the elements - particularly C, Si, or S - are on the
higher side of the permissible limits, surface pitting may occur.
This problem can, however, be minimised by reducing welding
current and speed.
Although carbon is the most significant alloying element
affecting weldability, the effects of other elements can be
estimated by converting them to their carbon equivalent (C.E.) by,
say, the following formula,
C ••E = m-/0 C + %Mn+
4 %Si
In general the susceptibility to underbead cracking from
hydrogen increases when the C.E. exceeds 0.40; as shown in Fig.
8.1, which relates C.E. to hardness, underbead cracking
sensitivity or weldability based upon slow bend capability of
notched-weld bead test bars.
Although most welding applications of mild steels require no
'preheating, heavy section (50 mm or more) and certain joint
configuration may 'require preheating. Less preheating is
required, when low-hydrogen processes are used. In general, steels
Fig. 8.1. Relationship between composition and under-bead hardness, crack
sensitivity, and notch-bead bend angle for 25 mm thick C-Mn
steel plate using E-6010 cellulose type covered electrodes.
in the 0.25-0.30% C range should be welded with low-H2
electrodes or with a low-H2 process if the ambient temperature is
below 10°C.
8.2.2. Weldability of Medium Carbon Steels
Because hardenability of steel increases with carbon content,
the medium carbon steels (0.30-0.50% C) serve where hardness,
wear resistance, or higher strength are needed. Important uses
for medium carbon steels include Wear plates, springs, and
components for railroad, agricultural, earth-moving and material
handling equipment. These steels are also used extensively in
....(8 1) machinery and tools. Many of the steels are selected for wear
resistance rather than high strength, and parts are frequently
heat-treated to meet desired properties.
A pronounced change in weldability of carbon steel takes place
when the carbon content is in the 0.30 to 0.50% range. As the
carbon content of steel is increased, the welding procedure should
be such as to avoid the formation of large amounts of hard
martensite in the HAZ.
For most applications, medium carbon steels should be
preheated prior to welding to control the cooling rate in the weld
504 Welding Engineering and Technology
metal and HAZ, and thus the formation of martensite. The
appropriate preheat temperature depends upon the C.E. of the
steel, the joint thickness, and the welding procedure. With a C.E.
in the range of 0.45 to 0.60, a preheat temperature in the range
of 100°-200°C is recommended, depending upon the welding
process and work thickness. The interpass temperature should be
the same as the preheat temperature.
Low-H2 welding procedures are manadatory with these steels.
Selection of filler metal for arc welding becomes more critical as
the carbon content increases. Pick up of carbon from a steel
containing 0.5% carbon by dilution will usually result in high weld
metal hardness, susceptibility to cracking, and tendency for brittle
failure. Low heat input is generally recommended for the first few
layers in a multi-pass weld to limit dilution. Higher heat input
can be used to complete the joint. It is a good practice to deposit
the final weld bead(s) entirely on previously deposited weld metal
without melting any base metal. This practice has the effect of
tempering the heat affected zones of the previously deposited weld
bead(s), especially those in the base metal. But this should not be
considered as a substitute for postweld heat treatment.
A stress-relief PWHT is recommended immediately after
welding, especially with thick sections, high joint restraint, or
service conditions involving impact or dynamic loading. Ifpossible,
the welded joint should be heated to stress-relief temperature
without immediate cooling to ambient temperature. Slow cooling
to room temperature following stress-relief is recommended to
avoid development of thermal stresses.
When immediate stress-relief is not possible, the welded joint
should be maintained at or slightly above the specified preheat
temperature for 5 to 7 minutes per mm of joint thickness. This
procedure promotes the diffusion of H2 from the weld zone and
reduces the possibility of cracking during intermediate handling.
However, it should not be considered a substitut.e for an
appropriate stress-relief.
8.2.3. Weldability of High Carbon Steels
High carbon steels contain 0.50 to 1.0 %carbon and Mn ranges
from 0.30 to 1.0%. These steels are almost always used in
hardened condition because they are commonly used for
applications requiring high hardness (maximum hardness
attainable is Rc65) or absrasion resistance, which is imparted by
heat treatment. Typical applications are for metalworking and
wood working tools, drill bits, dies and knives, and for abrasion
resist<l.nt parts such as plough shears and scraper blades. Some
Weldability of Specific Materials 505
r
farm equipment is made from re-rolled rail stock (0.65% C), which
is welded in the as-rolled condition, using preheating, interpass
I heating, and postweld stress relief.
I
The weldability of high carbon steels is poor because of their
high hardenability and susceptibility to cracking in the weld metal
and HAZ. Preheating and interpass temperatures of 2000-315"'C
are required to retard the formation of brittle, high carbon
martensite in the weld.
Use of low-hydrogen process can minimise the degree of
preheating necessary and for material thickness of 2 mm or less
it can eliminate preheating entirely. As ~ thumb rule, preheating
temperatures used with low-hydrogen electrodes can be 45° to
90°C lower than those needed for electrodes other than low-
hydrogen type.
Selection of an appropriate filler metal depends upon the
carbon content of the steel, the weldment design, and service
requirements. High carbon steel filler metals are not normally
produced however low carbon steel filler metal may be found
suitable for many applications. To reduce the pick up of carbon by
weld metal, welding procedures should be designed to keep the
dilution minimum.
The high carbon steel should be welded in annealed condition
and then heat treated. Annealing is recommended prior to repair
welding of broken parts. Successful welding requires development
and testing of specific welding procedure(s) for each application.
A postweld stress relief is recommended particularly for
welding joints in thick sections. The stress-relieving procedure
described for medium carbon steels may be used.
8.3. WELDABILITY OF HIGH STRENGTH LOW ALLOY
(HSLA) STEELS
A groups of low alloy steels that are designed to provide better
mechanical properties, and sometimes greater resistance to
atmospheric corrosion than conventional carbon steels are known
as HSLA steels. They are not considered to be alloy steels in the
normal sense because they are designed to meet specific
mechanical properties rather than a chemical composition. Carbon
content of HSLA steels rarely exceeds 0.28% and is usually
between 0.15 and 0.22%. Manganese content ranges from 0.85 to
1.60% depending on grade and other alloying additions.
Chromium, Columbium, copper, molybdenum, nickel, nitrogen,
phosphorous, titanium, vanadium, zirconium and silicon are us€d
 506
Welding Engineering and Technology
in various combinations but with total combined amount less than
1%.
Strength of HSLA steels is between those of carbon steels and
the high strength Q and T (quenched and tempered) steels. These
steels are normally marketed in the as-rolled condition to meet
required mechanical properties. They may also be supplied in a
controlled-rolled normalised, precipitation hardened condition to
meet specific notch toughness requirements. The minimum yield
strength of these steels are in the range of 410 to 785 MPa and
tensile strength in the 590 to 880 N/mm2 range. HSLA steels are
mainly ferritic steels.
Typical applications of HSLA steels include supports and
panels for truck bodies, railway cars, mobile homes, and other
transportation equipment. Components for tractors, threshers,
fertiliser spreaders, and other agricultural machinery, material
handling and storage equipment; and buildings, bridge decks, and
similar structures, pipelines and pressure vessels.
The weldability of most HSLA steels is similar to that of mild
steels however they often require preheat and also control of
hydrogen in the welding process. Preheating is beneficial in
controlling the cooling rate of the weld to reduce ~r avoid
martensitic transformation in the HAZ. Preheat and interpass
temperatures depend upon the C.E. of the steel, where C.E. may
be determined with the help of the following relationship.
C.E. = %C + ~~n + %Cr + %~o + %V+ ,%Si+ %~i + %Cu
...
(8.2)
When welding two different HSLA steels together, the higher
of the two preheat temperatures should be used. The base metal
on each side of the joint must be preheated to temperature over a
distance equal to the section thickness or 80.mm; whichever is the
larger ...
HSLA steels can be welded by all of the commonly used arc
welding processes. SMAW,GMAW,FCAW,and SAW are used for
most applications. Low-Hz practices should be employed with all
processes.
HSLA steel can be welded using ESW (electro&lag welding)
and EGW (electrogas welding) processes but there are certain
loading restrictions for bridges and other applications. Preheating
is not required with these two processes because oftheir high heat
input characteristics.
HSLA steels can also be joined by resis~ance spot, seam,
projection, upset, and flash butt welding processes. When spot,
Weldability of Specific Materials 507
seam, or projection welding, these steels can be welded with about
the same current and time settings used for low carbon steels.
However, higher electrode force may be needed because of the
higher strength of these steels.
8.3.1. Welding of Weathering Steels
HSLA steels produced with a minimum of 0.20% Cu usually
have twice the atmospheric corrosion resistance of structural
carbon steels. Steels with 4 to 6 times the atmospheric corrosion
resistance of carbon steels are often obtained with addition of Ni
and or Cr, often with more than 0.10% P. These alloys are usually
used in addition to copper. Such steels are 'called weathering steels
and are used for architectural and other structural purposes
where it is desirable to avoid painting for either aesthetic or
economic reasons. The steel protects itself by means of a dense
oxide coating (patina) which forms naturally on the steel surface
when it is exposed to weather.
The weathering steels can be welded by all the arc welding
processes, by gas welding and resistance welding.
To maintain the weather resistance characteristics of the
weathering steel a special welding procedure should be employed.
Use the AWSE-7018 type of covered electrodes to within one layer
of the top of the joint. The top layer should be made with E-
7018-C1electrodes) since the 2 % Ni in the weld deposit will cause
the weld metal to weather to the same extent as the weathering
steel. The C1 suffix weld deposit should be used for the top layer
of any multi-pass weld.
The same concept can be used for GMAW,FCAW or SAW of
weathering steels ..
8.4. WELDABILITY OF LOW ALLOY STEELS
Usually steels containing total combined alloy content (except
carbon) of up to 5% are known as low alloy steels. However,
chromium-molybdenum (Cr-Mo) steels and Nickel steels
containing up to 10% alloying content are also generally
considered in this category. It is obviously not possible to discuss
the weldability of all the steels included in this group, however,
the more important steels amongst these are included for
discussing the salient features of their weldability. They incLude
the following:
(i) Quenched and tempered steels,
(ii) 9% Nickel steels,
(iii) Heat-treatable low alloy steels,
5081

(iv)
Welding Engineering and Technology

Chromium-molybdenum steels,
I Weldability of Specific Materials
I
'0 00 oo ot-
509

(v) Precoated steels. U') a;

~ o

~J
to- •....•

8.4.1. Quenched and Tempered Steels

CIj
>.; d
""
e!,
"" ""
I oI
~ l() <t:>
""
co
<!;>
co
The quenched and tempered (Q & 1') steels are high yield ~
strength constructional steels having a minimum yield strength ~.
of 560 to 880 N/mm2• These steels usually have low carbon
contents (typically 0.10 to 0.25%) and other alloying elements;
- tIl
Cl)
Cl)
~
1:;
CIJ
CIj
o
<':l
""
I
chief among them being Mn, Si, Cr, Ni, Mo, ~ Ti, B, Cu, and Cb .•..
00
E-i lQ
C'l
may be present in very small amount. In general the steels are ~
to-

alloyed to favour the formation of tempe.red martensite and lower ~

bainite structures. = 0<&J

~
These steels are water quenched by special techniques from a a C'G
~ . ~
e ~ s
~ ~ ~
see e
o~
oci
It') ~~~
. 0 l'
temperature of 815° to 870°C and tempered at a temperature of
540° to 595°C. This produces a microstructure of tempered, low
';t:I o
l:':

~..,
Me ..•• it)
00'-'1
cicici
/I \I /I
~~~
coo
II /I /I
o
/I
.a
U
/I
1'.hC'l
~;;;~
0\10
temperature transformation products which have an excellent .s. >~8 >~8 > II ~ /I
>UQ:l
combination of strength and toughness. The heat treatment ?: o o lQ ow
Cl)
greatly increases the yield and tensile strengths of the steeL so .;;
<&J

o9
oi
<&J <&J

'9
o o""
o9
Unlike conventional alloy steels, these steels do not require .c; <=>

additional heat treatment by the fabricator except, in some cases,

00
.....
o
"'"
o
"'l
o d
""

o~
for a stress relief. tIl It') 0 'Q 0
Cl) <J-
The high yield strength Q & T steels are used extensively in ...•

earthmoving and mobile construction equipment, bridges, storage 1:: ,£ ~

o
d:, .J, J,;
t- c!,

tanks, penstocks, ships, buildings; end side and bottom plates for
Cl)
!:l.
o
..§
o
'(;! <:'i
C"~

---------.---"------------
~ 0[,"- ..q;

ore and waste haulage cars, hopper cars; pressure hulls for J.<
~ ~o ~
0 ~ 0 ::;~
It') <:>
~ 00 00 t- <0
submarines, and components for dust collecting equipment. ~ (3 r...
" " 9 ~ 9
Some high-yield strength grades are also available in
=
~
'""5
.!::/
C"j

8-t g-i 0~ 0
0~
abrasion-resistant modifications known as i\.R steels which have tIl
=
E:
It') .. It:> It')
o
.."
~ ..,
high hardness. Although these steels can have yield strength 1200 ...•
.•..
c5 <':l C-:l
d,
<':l
o s o
'9
N/mm2, hardness (up to 400 BHN) rather than strength is their ...•
tIl ~ LO
.....
c:!,
C'l
00 .h
key characteristic. These steels find applications requiring o o d o
•....•

o o
•....•

!:l.
maximum resistance to abrasion - in chutes, hoppers, and dump- S o o o"" 00
truck beds, etc. Good toughness can be combined with abrasion o "" ~ ~
o..;<
d
os
Q ci
resistance in these steels, for use in buckets, cutter bars, scraper ~ c:!,
•....•
J, c:!, J
o
•....•

blades, and impact plates. c-1

ci o
<&J

o
C"l
<&J

ci
8.4.1.1. HY Steels -
oci
Cl)

,.Q
><i

<:
><i

OJ
It')
00
C'l
o
An important sub-group of Q & l' steels is the HY (high yield ~ o
S
C'l
S C'l
o c:!,

strength) steels. The most common arid most available of these is C"j I ~
C'l
o
•....
o
•.....

o
HY 80 which has a minimum yield strength of 80,000 psi (""560
N/mm2). Higher strength grades are HY100, HY130, HY150, and o o00
•....•
00
<':l
•....•
00
It')
•....•

HY180. These steels are intended for nuclear reactor vessels and ~ >- >- >< •.....

:r.: :r.: :r.: :r.: ~

also for use in other critical applications, for example, pressure
hulls for submarines, and deep submergence research and rescue ~- <:"i "..; ..q; ar.i
510 Welding Engineering and Technology Weldability of Specific Materials 511
r
vessels. The chemical composition of five typical Q & l' steels are The maximum temperature should not exceed the sugges ted
given in Table 8.2. minimum temperature by more than 65°C otherwise the upper
The Q and l' steels can be welded by SMAW,SAW, FCAW, limit for acceptable welding heat input may be exceeded because
GTAW, and GMA\V processes. Weld cooling rates for these the relative effects of welding heat input and preheat temperature
processes are relatively high, and mechanical properties of the are the same. If a preheat of less than 40°C is used, precautions
HAZs approach those of the steel in the quenched condition. are necessary to ensure that the surfaces to be welded are free of
moisture.
Reheat treatment, such as quenching and tempering after
welding, is not recommended. 8.4.1.3. Heat Input and Welding Procedure
8.4.1.2. Preheating Welding-with high heat input processes like ESW, EGW and
Preheat for welding Q and T steels must be used with caution multiple wire SAW can completely change the mechanical and
because it reduces the cooling rate of the HAZ. If the cooling rate metallurgical properties of Q and T steels necessitating reheat
is too slow, the reaustenized zone adjacent to the weld metal can treatment of the welded joint (quenching and tempering) to regain
the mechanical properties to the desired levels.
transform to ferrite with regions of high carbon martensite or to
coarse bainite. Both these microstructures lack high strength and The heat input limitations for welding most Q and l' steels are
good toughness. Next to this zone, a band of previously tempered based on toughness requirements because this property is more
steel may be overtempered with consequential decrease in adversely affected from excessive heat input, than strength. Heat
strength. input limitations are applicable to each weld pass and are not,
considered cumulatively. They are also applicable to single arc'
In some cases, a moderate preheat can provide insurance welding processes.
against cracking from several causes. Preheating is recommended Large weld beads characteristically have poor notch
when the section to be welded is thick and highly restrained or toughness. Thus, good practice for welding Q and l' steels is to
some other condition is present that may cause cracking. deposit many small stringer beads. This technique produces weld.'
Suggested minimum preheat and inte,rpass temperatures for steel metal with good notch toughness as a result of grain refining and
T- 1 and the recommended ranges of these temperatures for Q and tempering action of succeeding passes.
l' steels HY80, HYlOOand HY130 are listed in Table 8.3.
Hydrogen is very detrimental when welding Q and l' steels; a
Table 8.3. Recommended Preheat and Interpass very small quantity may cause H2-induced cracking. Since
Temperatures for Welding Some Q and T Alloy Steels moisture is a source of H2' the moisture content of covered
electrodes must be kept below certain levels depending upon the
----- over
Thickness
22-25
to 51
19-22
(mm)
25-351355-150
35-.18
16-19
13--16
38-51
up 95-150
110-150
70
95
10
10
95-135
55-95 , '
- 25-65
55-150
20-150 steel being welded. It should not be allowed to exceed 0.4 wt.%
8.1
I
f-- HY130 °C
Temperature, when welding Q and rcarbon steels having minimum yield
II

-r;--ilrr80 ;IT;;;i-
strength up to 500 N/mm2, and 0.1 wt.% with Q and l' low alloy
steels having a yield strength of 700 N/mm2 and above, such as
HY130 steel. Electrodes contaminated by contact with water,
paint, oil, grease, or other hydrocarbons should be discarded.
Deposition of stringer beads without appreciable transverse
oscillation of the electrode is preferred for welding Q and l' steels.
A weave bead technique generally requires a slower trlivel speed
along the joint, with a corresponding increase in heat input. This,
in turn, produces welds with lower strength and toughness. For
vertical welding, however, slight weaving of the electrode for no
more than two electrode diameters is usually satisfactory.
Postweld stress relief heat treatment is generally not required
to prevent brittle fracture in wel~ed joints in most of the Q and T

512 Welding Engineering and Technology
steels. The need for postweld stress relief heat treatment,
therefore, should be thorougWy investigated for each application
because many alloy steels are designed for service in as-welded
condition.
Also, when weldment of many high-strength alloy steels are
heated above about 510°C, intergranular cracking may take place
in the coarse-grained region of the HAZ. The intergranular
cracking occurs by stress-rupture, usually in the early stage of the .
heat treatment. Susceptibility to cracking increases with
increasing weld restraint and severity of stress concentration.
If postweld stress relief is required, the temperature must not
exceed that used for tempering the steel. In fact a temperature of
about 30"C lower than the tempering temperature is desirable ·to
avoid lowering the strength of the steel.
8.4.1.4. Thermal Cutting
Q and T steels can be readily cut by oxyfuel gas flame cutting
either manually or mechanical. Machine cutting is preferred to
obtain relatively smooth, uniform cut edges fo.~welding. Generally
the cutting conditions are the same for carbon steels and Q and T
steels. If scale of the steel surface causes err12ticcutting action,
the travel speed should be reduced about 10 to 15%, or a cutting
tip of the next larger size should be used. Stack cutting of thin
plates should be avoided because the high heat input tends to
overheat the plate closest to the cutting tip. Q and T steels can
also be cut by Plasma Arc Cutting.
Q and T steels can generally be cut thermally without
preheating. However, the steel temperature should not be lower
than 10°C during cutting. Sections above 100 mm thick require
preheating to about 150°C.
Air-carbon arc gouging may be used to remove welds, portions
of welds, or base metal. Oxyfuel gas gouging, however, should not
be used on Q and T steels because of the excessive heat input and
the slow cooling rates inherent with the process.
8.4.2. 9% Nickel Steels
The 9% Ni-steels are also Q and T steels but are considered
separately because they are intended for different type of service
than the Q and T steels already discussed in the previous section.
The 9% Ni-steels provide high strength and extreme toughness at
very low operating temperatures such as to build tanks and
vessels for containing liquified natural gas which has a
temperature of -160°C. These steels provide good notch toughness
at temperatures down to -196°C, .because their ductile ..to-brittle
Weldability of Specific Materials 513
transformation temperature is below about -195°C. An alternative
to 9% Ni-steels is a 5% Ni-steel which also contains 0.25% Mo that
provides good service properties down to -170°C. Both these types
of steels are welded in the heat treated condi.tion and do not
require PWHT to obtain weld joint properties similar to those of
the base metal.
The 9% Ni-steel can be flame cut using normal oxyfuel gas
equipment and can be welded by SMAW, SAW and GMAW
processes.
For welding by SMAW,high Ni-Cr-Fe type electrodes such as
AWS ENiCrFe-2 type and ENiCrFe-3 type are used. A preheat or
PWHT is not required on material thickn~ss up to 50 mm, however
the base material should be brought to .normal room temperature
of 20°C. When making V or bevel groove welds the minimum
included angle should be 70°. The high Ni-electrodes operate
differently from mild steel or stainless steel electrodes in that they
have low penetration, and do not flow or wash into the side wall
of the weld joint. The electrode should be pointed to place the
molten metal where it is desired.
When SAWprocess is used, the compatible material electrode
wire with neutral type welding flux are used. For thinner
materials C<50 mm) a room temperature preheat of 20°C is used
while for thicker (> 50 mm) material a preheat and an interpass
temperature of 120° to 150°C is recommended.
For welding with GMAWprocess, the high Ni-Cr electrode,
AWStype ERNiCrFe-6, is used and a shielding gas mixture of90%
He + 10% AI' is recommended. The short circuit mode of metal
transfer is employed for welding. The pulse-GMAWprocess can as
well be used for welding 9% Ni-steel.
8.4.3. Heat Treatable Low Alloy Steels
The hardness of many low alloy steels makes it impossible to
fabricate them by welding in the hardened condition because of
their cold-cracking tendencies; in such cases heat treatable low
alloy CHTLA)steels may be found a suitable substitute. The carbon
content of HTLA steels generally range between 0.25 and 0.50%,
compared to 0.10 to 0.25% for the Q and T steels. Due to higher
carbon content these steels have higher strength but lower
toughness than the Q and T steels. They are normally \yelded in
the annealed or normalised or overtempered condition, and then
the entire weldment is heat treated to the desired strength or
hardness. Some8.4.
listed in Table of the industrially important HTLA steels are
 ••••
•..
III
::I
F
IiIil'lCI
;3
HPI;I~;; --- ---II;
IQ.
H-ll
D-6a
H-13
SAE
300M
Air
4130Steel
Rocoloy
Common ,Co
9-4-45
Mn
ii
(V
1688
0.45-0.60
0.15-0.25
1.25-1.75
S270
iX200
0.90-1.20
0.15-0.30
0.40-0.80
0.60-0.90
0.90-1.30
1.15-1.60
Ni
0.70-0.95
0.65-0.90
0.60-0.80
1.45-1.80
1.65-2.00
0.75-1.00
0.30
0.10
0.80-1.10
0.80-1.20
0.20-0.40
0.25
8.00
0.75-1.10
0.40-0.70
Others
Cr
4.75-5.50
4335V
1825
1.90-2.25
0.40-0.60
0.70-0.90
2067
2135
Mo Y.S.
or
1.40-1.75
0.30-0.45
0.05 -
1720
0.20-0.35
1720
4.00
0.90-1.10
UTS
0.30-0.40
(AISI)
1860
2135
1930
i 2067
1998
V 0.03-0.08
AMS
IIS'
min0.30-0.50
0.80-1.20
Designation
Ii'
V64340.42-0.48
0.41-0.46
5: .'1.65-2.00
4.75-5.50
0.45
0.39-0.45
0.30-0.40
0.80-1.10
1860
1480
0.20-0.35
1825
1790
17200.38-0.43
0.31-0.38
0:28-0.33
0.05-0.10
0.10-1.20
0.17-0.23
0.20-0.30
Dl
ChemicalStrength,
Composition,
MPa wt%
6"
I
Ul !

s'
III
::I

-- --------
...••.. .•..•..•....•..•.•.•.•••......•.•..... ---~-~--------_._--_._---._._~-_.- ..__ ---~-------------;

.~ac-> 'tl.,-g ~~nQ:5;';-..~

- -
Temperature: F
N
a a Dl

C> a a-
aa aa
CD

aa a =
....
II'
•...
t:l
(D
j:2..
QQ
(J)
••••
tf.l
III
l'l
III
Dl
CI
~
~ 11

Max . hardness ,H R C
aa a N
a W
.....,a,f:-

,,
n
...,
ID a-
<:>
::J o ..•.
::I
(70.
~. ano
'tJ ..•. iii
,I 6iii0a\\\ 0\~\\\ \
..•.

~r
ID ::I 0 ...•
~~ ::;: ::I
ID
::;:
II)
., ,f:- 3:
..,
<:>
ID a0~3:0
...,
ID
V1\ \ \ C>

0'
'1
>-
....•
en
....•
0l>-
e.:>

o-
01>-
Q.,
CD
a-
...
->
a-N
...•
<:>
;;;::
CD
.-..J ;:
..0
,f:- ~
n'
'C
"0

m
.... 0'
(b

~
 516 Welding Engineering and Technology
The combined carbon and alloy contents of HTLA steels are
sufficient to promote the formation of martensite from austenite
when cooled rapidly to below the appropriate transformation
temperature. The carbon content is sufficiently high to form hard
and thin brittle martensite as shown in Fig. 8.2. During welding,
a portion of the HAZ will transform to austenite. If the ¥leld metal
and the austenitic HAZ are cooled too fast, they will transform to
martensite or a combination of martensite and bainite, as shown
by CCT curves in Fig. 8.3 for AlSI 4340 steel. Steels in this group
(i.e. HTLA steels) are very sensitive to Hz-induced cracking.
Proper preheating and low-Hz electrodes should be used to avoid
underbead cracking.
The best approach to welding HTLA steels is to preheat the
joint area to 315°C or higher so that the cooling rate of the weld
will be slow enough to form softer bainite in preference to hard
martensite. A convenient way of estimating the amount of preheat
temperature required is to use the carbon equivalent (C.E)
concept. A typical formula for determining the C.E. of a steel
containing not more than 0.5% C, 1.5% Mn, 3.5% Ni, 1% Cr, 1%
Cu, and 0.5% Mo is as follows.
C E _ nfC %Mn %Ni %Cr %Cu 4- %Mo ...(8.3)
.• - 70 + 6 + 15 + 5 + 13 ' 4
The suggested preheat temperatures for several ranges
carbon equivalents are listed in Table 8.5.
Table 8.5. Suggested Preheat Temperatures Based
Carbon Equivalents of Steels
S.No, I up>to0.60
0.45-0.60
Carbon 0.45 205-370
Suggested
equivalent 95-205
optional
(OC) preheat temperature
%)
Weldability of Specific Materials
511
also by electron beam welding. LOw-Hz electrodes are recommen-
ded for use in SMAW, and for GMAW the shielding gas suggested
is mm oror Ar-COzmixture
1.6COz below. especially with electrodes of diameter
Various methods of weld backing are used to shield the back
side of thejoint, support the weld metal, and provide the proper
root contour. Successful methods include. the use of ceramic tape
and copper bars with drilled holes for applying shielding gas to
the underside of the joint. However, Cu is not recommended for
backing when a relatively high preheat temperature is requir'ed
because
weld of its
metal rapidto absorption
fusing it. of heat and the likelihood of the
Deposition of filler metal in the joint is usually done with a
relatively low heat input. High heat in.put tends to produce a wide
HAZ with enlarged grain size, and to increase the likelihood of hot
cracking in the weld metal and HAZ. Automatic welding is
preferred over manual welding for linear or simple circumferential
joints because automatic operation produces more uniform welds
with lesser defects due to fewer stop and start positions.
8.4.3.3. Postweld Heat Treatment
The weldment of HTLA steels is often immediately heated for
of stress-relief treatment before cooling to room temperature. During
the stress-relief treatment, martensite is tempered, and the
weldment can therefore be cooled to room temperature without
on danger of cracking. Mter this, the weld can be given PWH1: to
develop the strength and toughness the steel is capable of
attaining, Sketches of desired and undesired thermal cycles
8.4.
during welding and postweld stress relieving are shown in Fig.

Welding Welding

'-01
::>
Q.I
<-
.•.. Stress relieving.;: relieving
o
'- o
'-
Th~ above suggested preheat temperatures are for arc )Velding 01
Co
Q.I

processes. They may be affected by the thickness ofthe workpiece. E A, ~

~ ~ Al
For oxyfuel gas welding processes, however, preheating is usually
not required owing to the slow cooling rates associated with such
welding processes.
8.4.3.2. Welding of HTLA Steels
~T·· Ime T'Ime
Heat treatable low alloy steels can be welded successfully by
arc welding processes like SMAW, SAW, GMAW, GTAW and FCAW; (b l Undesirable PWH T cycle (a) Desirable PWHT cycle
resistance welding processes like spot welding, flash welding, and Fig. 8.4. Schematic represelltation of: (a) desirable, and (b) undesirable the.rmal
cycles, for welding and postweld stress relieving of HLTA steel.
·;•..
·.cll._II.III.-IIII.III.~::;.;~!!'._.~!'I
!'I'-,,·,., ..I•.
III'!, '!,!!!!!,!!".!!l!,,-~!!!._ .. !!!,-:;:-!!!_ .!!l.!.!. !!
'~!!l",::~~, .~':''-~.:"~!,.,.,',.ll!
.. !"·j!!!!_.·Il'i'l'.=!!II!l,l!!'~~!!li!!=m=!!!!'!!!!!:l!_"""" """ ",,,,__ ~

518 Welding Engineering and Technology

Weldability of Specific Materials
519
In case the stress-relief heat treatment cannot be carried out
0l.Q l.Q
0coJ.,9I1
'"
1 ~I
•...•
c:;;
to
I",j
8
-0 J--$.
...•
immediately aft~r welding, the temperature of the completed ~I
~I
61
00
c...1 c:;;
J.,
"'"
<:>
<:><=> "'"
I l.Q

.~ ·9
~'"
I:l.
l.Qr:-
00
to
"'"
~
<:>
<=> c:;; l.Q
<=>
l.Q
'<I'
co
co:l
<:>
0>"'"
C'l
l.Q
....•
<:>
<:>
.....
c-i
"'"
.c:;;
cO...•
...r
<=>
....•
....•
<:>
.....
c:!,<f00
~~
""!
<=>
<f
<
<f
,0
l.Q
to
<=>
c:!, l.Q
c:!, 0~
....•
l.Q
<=>

weldment can be raised to approximately 400°C, which is in the ....

U '"
as Q,)
Z ,0
u~00
.•..
....=
00
Q 0as ~c3
....
.•..
•..
~
en
.Q
0:
0:'

u~ElISElo. I~I£:
0:::l,.Q
1.:
..cl
El
Q,)
<=> Co)
Co)
=s .e-
••••
~
en
Q,)
El
~
vicinity of the bainite 'knee' for most of the HTLA steels. Byl.Q <=>
Q,)
El cQ

holding at this temperature for about 1 hour or less, the remaining

austenite can transform to bainite, which is more ductile than
martensite. Therefore, when the weldment is cooled to room
temperature, no cracking should be encountered. Further heat
treatment can be carried out later to optimise the microstructure
and properties of the weldment.
If a HTLA steel cannot be given a PWHT and must be welded
in the heat-treated condition, the softening, as well as H2-
cracking, of the HAZ can be a serious problem. To minimise the
softening problem, lower heat input per unit length of the weld
should be employed, and the preheat, interpass, and stress-relief
temperatures should be at least 50°C lower than the tempering
temperature of the base metal before welding. Since PWHT is not
to be given, the composition of the filler metal can be substantially
different from that of the base metal to achieve the desired level
of strength in the weld metal.
8.4.4. Chromium-Molybdenum Steels C'JI
'"
...•
<=>
c:;;
<=> .....
"'" <=>
I 0'"c~lto9.Q:;;l.Q
...•
....•
C>
to
l.Q
co:l
"'"
"'"
<=>
<=>
"'"
< =>
c:;;
<=> <=>

~J C"I
c:;;
~
The chromium-molybdenum (Cr-Mo) steels are also popularly C,)/

<=>

<=>

known as chrome-moly steels. All steels in this family contain ~ ,to

9% Cr and ~ to 1% Mo. The carbon content is normally less than
0.2% for good weldability, but these steels have high harden ability.
The Cr provides improved oxidation and corrosion resistance and
Mo increases strength at elevated temperatures·. They are
normally supplied in the annealed or normalised and tempered
condition.
Apart from strength at elevated temperatures, Cr-Mo steels e
do not creep which means that they do not stretch or deform under ;:l
8
long periods of use at high pressures and temperatures. Also, they .~
do not become brittle after extended periods of high temperature I
8
service. The strength levels of these steels extend from about 600 ...•.!. Co) ~
•... ...•
::g ~ 0 ~
N
~00I
0a;
~~.n
~C'l...•
Co)
0 >I
•....
C'l
l.Q
....•
...•
;:.,
Co)
....•
'";'
cO
ea
Co)
Co) •...
O
>:i
....•
Co)
•.. 0
.!.~
::g0
"'"
Cll
•..
;$ .!..!.
c.:. Cd

N/mm2 to 950 N/mm2• Nominal compositions of popular Cr-Mo ....• en

Cll
;:l
steels are given in Table 8.6. ~
Cr-Mo steels are widely used in the petroleum industry and Cll

b.o
l':l
in the steam power generating or elevated temperature i:i5

applications. Another typical use of the Cr-Mo steels-usually in *

"i:£
 ?-.;;';;'-.

520 Welding Engineering and Technology Weldability of Specific Materials 521

the form of tubing - is in the highly stressed aircraft parts.
Generally, the best elevated temperature strength is obtained with
1 I
24 Cr-lMo steel.
8.4.4.1. Weldability of Cr-Mo Steels
The Cr-Mo steels are harden able steels; therefore it is
necessary to provide a welding procedure that includes preheating
to prevent hardening and cracking. Recommended minimum
preheat and int.erpass temperatures for various thicknesses of
Cr-Mo steels are given in Table 8.7. Lower preheat temperatures
may be used if the welding heat input is relatively high or chance
of H2 pick up very low.
Table 8.7. Recommended Minimum Pl"eheat and Inter- I·
pass Temperatures for Welding Cr-Mo Steels with Low I
J.
!
I ESW, EBW, laser welding, friction welding, resistance welding,
and brazing.
I In all arc welding processes it is necessary to closely match
r the weld metal deposit analysis with the -composition of the base
metal. When several grades of Cr-Mo steels are to be welded on
one job, limiting the number of different filler metals used will
simplify material control. Filler metal of the same or slightly
higher alloy content can be used for welding several Cr-Mo steels.
For example, 1iCr-~Mo filler metal can be used for welding
~Cr~Mo, lC~Mo, and liCr~-Mo steels. Similarly, 2iCrlMo filler
1 1 1 1 1

metal can be used for 14C~Mo, 2C~Mo, 24CrlMo, and 3Cr"2Mo

steels.
H2 Covered Electrodes
Covered electrodes for welding the Cr-Mo steels are low-Hz
S.No. 1.
321.
911.Cr-
Cr
Cr-1
1.
-Mo
<":r-1Mo
Cr-1 1.
Mo Mo
Mo
95°C
150°C
205°C 25 *205°C type. Low-H2 electrodes are difficult to use with open root joints,
260°C
5721 Cr-ZMoOver
150°C
Steel
13-25 38°C 120°C
Thickness, mm
7.
4.
24 22 65°C
therefore GTAWprocess is used for making the root pass.
up to6. 9.
13 11
1 Type 309 or type 310 austenitic stainless steel filler metals are
3.
often employed for minor repair welding of Cr-Mo steels. They are
sometimes preferred for those applications where the weldment
cannot be given a PWHT. However, an austenitic stainless steel
filler metal is not satisfactory if the welded joint will be subjected
to cyclic temperature service, or a service temperature where
either carbon migration or sigma phase formation can take place.
The difference in coefficients of thermal expansion of an austenitic
weld metal and ferritic Cr-Mo steel results in internal stress at
the weld interface during cyclic temperature service.
Much of the welding on Cr-Mo steels is done on pipe. For pipe
welding, GTAW process is often used for making the root pass
while SMAW,GMAW,or FCAW can be used for the remainder of
the weld joint. Argon or Ar+(1-5)% O2 is used for shielding with
GMAW process. Backup rings are not used for welding high
pressure, high temperature pressure pipes.
The Cr-Mo steels can be joined by electroslag welding. The
process is primarily used to weld sections that are 75 mm or more
in thickness. One advantage of this process is that the large
quantity of heat generated during welding preheats the base metal
ahead of the molten weld pool. This large build up of heat also
* Maximum carbon content of 0.15%. For higher carbon content, the preheat leads to stress-relieving effect.
temperature should be increased by 55° to HO°C. Lower preheat temperature may PWHT: Because Cr-Mo steels are hardenable, therefore it is
be used with GTAW. --
necessary to provide a welding procedure that includes not only
rhe Cr-Mo steels can be readily welded by SMAW, GTAW, preheating but also PWHT. Welds in Cr-Mo steels with high
I
I
GMAW,SAW, and FCAW processes. They can also be joined by hardenability should be postweld heat treated regardless of
I '
 Weldability or Specific Materials 523
522 Welding Engineering and Technology

dimensions. Also, welded joints to be used in acid or caustic service (iii) , Zinc-rich primers, and
should always be postweld heat treated .. (iv) Lead-tin alloy (Tern plate).
A PWHT of Cr-Mo weldments is used to reduce welding These precoated steels find extensive use in manufacture of
stresses and also to increase the ductility and toughness of the truck bodies, air-conditioning housings, processing tanks,
weld metal and HAZ. Table 8.8 gives recommended stress-relief electrical towers, shipbuilding, etc. Welding is often employed in
fabricating these products.
temperature ranges for Cr-Mo steels.
Table 8.8. Recommended Stress-Relief Temperature 8.4.5.1. Welding of Aluminised Steel
Ranges for Cr-Mo Steels Aluminised steel is produced by coating steel either with
1 675-760
commercially pure aluminium or AI-8% Si alloy using hot dip
S.No. !Cr-
1 Cr-!Mo
I!
5 Cr-l !Mo
Cr-!MoMo
Steel process; and such steels are widely used in tubing and in
4. 2.
6.
.
2"4Cr-1
4
3
2 Mo
2
Temperature range, °C*
635-705
63e>-730
675-745
automobile industry particularly for exhaust mufflers. Aluminised
steels can be joined by arc welding, resistance welding and
soldering. I

For welding by SMAW process, an electrode with basic type

8. I.C~l Mo ~ _
• Temperature should not exceed the tempering temperature
In case welding is interrupted before the joint is completed,
the weld joint must be allowed to slow cool and then be properly
preheated before welding continues. The induction heating system
and resistance heating are popular for both preheating and PWHT
of Cr-Mo piping.
When different grades of Cr-Mo steels are welded together, the
preheat and PWHT temperatures should be based on higher alloy
material while welding electrode, as already mentioned earlier,
can be based on the lower alloy material.
8.4.5. Weldability of Precoated Steels
Steel sheets and plates as well as other product forms are
pre coated with an oxidation-resistant or corrosion-resistant
material to extend the service life of the product. Coatings
commonly used on steels are :
(i) Aluminium (Aluminised steeD,
(ii) Zinc (Galvanized steeD,
covering, such as AWS E7015 or E7016 with a.c. is used. The
formation of Al203 has adverse effect on weld pool leading to
improper wetting, and shaping of weld bead.
For welding with GTAW, filler metal addition should be used
to control the Al content of the weld metal and provide acceptable
weld ductility. For welding aluminised stElel with GMAW, Argon
with a minimum addition of O2 or CO2 to stabilise the arc is
recommended for shielding the weld pool.
Loss of ductility in aluminised weld metal is not a problem
with GMAW. A deoxidised electrode is not required because the
aluminium from coating will deoxidise the weld metal.
Resistance welding of aluminised steel can be done by spot or
seam welding process. Aluminised steel tubing is produced in tube
of the steel. mills using resistance butt welding process. Both 50 Hz and high
frequency currents are used for such welding.
Aluminised steels are cleaned by chemical or mechanical
method to remove oxide film before soldering. A dip in 5%
trisodium phosphate solution followed by water rinse and drying
prepares the steel for soldering.
After cleaning the Al-coated steel should be rapidly heated to
soldering temperature and that is best done by electric or
ultrasonic soldering irons of required heating capacity. Specially
formulated fluxes and suitable solders are commercially available
for soldering aluminised steels. Soldering should be completed
quickly to avoid the formation of aluminium oxide and undesirable
alloying of aluminium coating with the steel.
8.4.5.2. Welding of Galvanised Steel
Zinc coated steels can be welded successfully provided specific
precautions are taken to compensate for the evaporation of zinc
I
524 Welding Engineering and Technology
from the weld zone. Zinc vaporises during welding because its
boiling point is 871°C while the melting point of steel is 1540°C.
Thus zinc volatilizes and leaves the base metal adjacent to the
weld. The extent.of the area affected is dependent upon the heat
input to the work. That is why zinc depleted zone is larger in
slower welding processes like GTAWand oxy-acetylene welding.
The welding processes used for welding galvanized steel
include SMAW,GMAW,GTAW,FCAW, Carbon arc welding, and
resistance welding.
The covered electrodes used for welding galvanized steel sheet
are the rutile and basic types; however c.ellulosictype electrodes
are used fo:!:"welding thicker sections and pipes. Basic coated
electrodes may also be used for welding heavier thickness.
Forehand welding technique is employed to facilitate the
evaporation of zinc ahead of the arc.
In GMAWof galvanized steel highly deoxidised fine wires with
short-circuiting technique are used with 100% CO2 or argon
+25C02 as shielding gas. The amount of spatter is usually higher
than when welding lillcoated steel. This necessitates frequent
cleaning of gun nozzle. Stainless steel 01· bronze wires may be used
to deposit corr08ion-resistant weld metal. Flux-cored arc welding
using highly oxidized wire may be employed with results similar
to those obtained by GMAW.
GTAWprocess can be used but being a slow process not only
results in large zinc depleted areas around the weld but also leads
to contamination of tungsten electrode. Electrode contamination
can be reduced by higher flow rate of shielding gas but that can
be expensive.
Carbon arc process· employing brass (60% Cu, 40% Zn) filler
wire has been widely used for welding zinc coated steels
particularly in fabrication of air-conditioning ducts. Both single
and twin carbon arc torches can be used equally effectively.
Resistance welding of zinc coated steels results in much less
evaporation of zinc than in arc welding processes. But resistance
welding results in zinc pick up by the welding electrode tip and
lowers the current density in the weld zone necessitating
progressive increase in welding current to make satisfactory
welds.
8.4.5.3. Weld Quality
Welds made in zinc coated steels are prone to porosity and
cracking due to entrapment of zinc vapours in the weld metal;
delayed cracking due to stress corrosion may also occur. Cracking
I
Weldability of Specitic Materials 525
is caused by intergranular penetration of zinc into the weld metal
and is sometimes referred to as 'zinc penetration cracking' and
occurs most often across the throat of a fillet weld particularly
when the coating is present at the root of the weld. Such cracking
tends to be less prevalent with SMAWthan with GMAWon 6 mm
Or thicker plates. Cracking can be controlled by allowing the :tine
vapours to escape rapidly ahead of the weld pool by keeping J arge
root gaps.
TClproduce a corrosion-resistant joint the zinc coating must be
re-applied in the zinc-depleted area. This can be done by using
zinc-base paste sticks on heated base metal. Another method of
applying zinc coating is by flame spraying using a zinc spray fiDer
material. The thickness of re-applied zinc coating should be 2 to
3 times the original coating to ensure proper corrosion protection.
8.4.5.4. Welding of Steels Coated with Prefabrication
Primers
An alternative to galvanized coating on steel is the application
of pre-fabrication primers containing 85% to 97% metallic zinc.
They are commonly referred to as weld-through primers. Steels
painted with prefabrication primers ,can be welded by resistance
welding and arc welding processes.
Resistance spot welding is facilitated if the zincz-richcoating
is on the faying surface (i.e. the surfaces to be joined by fusion)
only and not on the surfaces in contact with the electrodes.
Otherwise paint/primer will contaminate electrodes and
necessitate frequent electrode dressing.
Arcing. between faying surfaces, a popping noise, and the
generation of pungent fumes are characteristics spot welding
through Zn-rich paints/primers. When possible, it is best, to weld
through the paint before it has completely dried so that the
welding pressure may displace the paint from the faying surfaces
and provide a more conducting path.
Arc welding of painted steel without first removing the paint
is generally not recommended because the pigments can
contaminate the weld metal. When Zn-rich paint is used on steels
prior to arc welding, porosity may be a problem because of the gas
forming constituents of the paint. Paint may also cause cracking
in welds in steels that are sensitive to H2-induced cracking. Tests
must be made to determine if there are such problems when
welding a particular steel; two such tests are described in
Appendix A.

526 Welding Engineering and Technology
8.4.5.5. Welding of Tern-Plate
Tern-plate i.e. the steel sheet coated with lead-tin alloy is often
used for making gasoline tanks by welding, for use in automobiles.
The process generally applied for welding tern-plate is resistance
welding. If oxy-acetylene or arc welding process is used than the
coating is destroyed by evaporation and it must be replaced by a
process similar t6 soldering.
Safety Practices: Zinc fumes are hazardous to health. Metal
fume fever, also known as Zinc chills, may follow exposure to zinc
fumes released during welding ofgalvanized or painted steels. The
chills are caused by colloidal zinc oxide penetrating to the lungs.
Inhalation of iron oxide fumes may 'also cause metal fume fever.
Threshold limit values (TLV)of iron and zinc oxides in breathing
air are approximately 5.0 mg/m3. Although Al203 fume is presently
considered non-toxic, yet its inhalation should be minimised by
proper ventilation.
Zinc chill illness, when it happens, begir.s a few hours after
exposure or more frequently during the night, and may cause a
sweet taste in the mouth, drY1~essof the throat coughing fatigue,
yawing, weakness, head and body aches, vomiti-.:l.g, chills or a fever
rarely exceeding 102°F (38.9°C). A second attack seldom occurs
during repeated' exposure unless there has been an interval of
several days between exposures.
Precoated steels must never be welded in confined spaces. If
welding of such steels in confined space is unavoidable, adequate
ventilation must be provided or at least hose masks must be
provided to the concemed personnels.
8.5. WELDABILITY OF STAINLESS STEELS
Stainless steels or corrosion-resistant steels are those alloy
steels that have a Cr content of atleast 11%. These steels do not
rust and strongly resist attack by a great many liquids, gases, and
chemicals. Many of the stainless steels have good low-
temperature toughness and ductility. Most stainless steels exhibit
good strength properties and resistance to scaling at elevated
temperatures. All stainless steels contain iron as the main element
and Cr in amounts ranging from about 11% to 30% and 1 to 2%
manganese. Chromium provides the basic corrosion resistance to
stainless steels. A thin film of Cr-oxide forms on the surface of the
metal when it is exposed to the oxygen of the air. This film acts
as a barrier to further oxidation, rust, and corrosion. Exposure to
Weldability of Specific Materials 527
elevated temperatures increases the thickness of the oxide film
and reduces the lustre of the stainless steels.
Stainless steels are commonly divided into the following four
groups:
(i) Austenitic stainless steels,
(ii) Ferritic stainless steels,
(iii) Martensitic stainless steels,
(iv) Precipitation hardening stainless steels.
These groups are listed and discussed in the same order in
which the well known AISI (The American Iron and Steel
Institute) designation numbers are assigned to different stainless
steels.
The first three groups are characterised by the predominant
microstructural phase present viz., austenite, ferrite, and
martensite, when the stainless steel is at room temperature. The
fourth group contains those stainless steels that can be
strengthened by an ageing heat treatment.
Before discussing the weld ability of these steels it is
imperative to discuss their general characteristics and uses.
8.5.1. Austenitic Stainless Steels
The austenitic stainless steels contain a combined total Cr, Ni
and Mn content of 24% or more with Cr content generally above
16%, and Ni often above 7%. The Cr provides oxidation resistance
and resistance to corrosion while Ni and Mn stabilise the austenite
phase sufficiently to retain most or all of it when the steel is cooled
rapidly to room temperature. Austenitic stainless steels are
usually the most corrosion resistant of all the stainless steels and
thus find wide use in chemical plants.
The microstructure of the austenitic stainless steels is either
all austenite, or ferrite in a matrix of austenite. These steels are
grouped into AISI 200 and AISI 300 series. The basic grade is type
302 which contains 18% Cr and 8% Ni. These steels are not
hardenable by heat treatment.
The austenitic stainless steels generally have low yield
strength and high ultimate tensile strength that is why they are
often very ductile. They have excellent properties at cryogenic
temperatures and have higher strengths at 540°C than the 400
series. Nominal compositions of the austenitic and all other grades
of stainless steels are shown in Table 8.9 while the specific uses
of each individual austenitic stainless steel are given in Fig. 8.5.
, ,."-
528 Welding Engineering and Technology

Table 8.9. Nominal Composition of Different Grades

of Stainless Steels
Chemical Analyses of Stainless Steels, per cent
AISI Carbon Manganese Silicon Chromium Nickel Other
No.
Chromiurn-Nickel-Magnesium-Austenitic-oNen-Hardenable
-----
Elements

201 0.15 max. 5.5n.5 1.0 16.0/18.0 3.5/5.5 N2 0.25 max.

202 0.15 max. 7.5/10 1.0 17.0/19.0 4.0/6.0 N2 0.25 max.
Chremiurn-Nickel-Austenitic-N on-Hardellable
301 0.15 max. 2.0 1.0 16.0/18.0 6.0/8.0
302 0.15 max. 2.0 1.0 17.0/19.0 8.0/10.0
302B 0.15 max. 2.0 2.0/3.0 17.0/19.0 8.0/10.0
303 0.15 max. 2.0 1.0 17.0/19.0 8.0/10.0 S 0.15 min.
303Se 0.15 max. 2.0 1.0 17.0/19.0 8.0/10.0 Se 0.15 min.
304 0.08 max. 2.0 1.0 18.0/20.0 8.0/12.0
304L 0.03 max. 2.0 1.0 18.0/20.0 8.0/12.0
305 0.12 max. 2.0 1.0 17.0/19.0 10.0/13.0
308 0.08 max. 2.0 1.0' 19.0/21.0 10.0/12.0
309 0.20 max. 2.0 1.0 22.0/24.0 12.0/15.0
309S 0.08 max. 2.0 1.0 22.0/24.0 12.0/15.0
310 0.25 max. 2.0 1.50 24.0/26.0 19.0/22.0
310S 0.08 max. 2.0 1.50 24.0/26.0 19.0/22.0
314 0.25 max. 2.0 1.5/3.0 23.0/26.0 19.0/22.0
316 0.08 max. 2.0 1.0 16.0/18.0 10.0/14.0 Me 2.0/3.0
316L 0.03 max. 2.0 1.0 16.0/18.0 10.0/14.0 Me 2.0/3.0
317 0.08 max. 2.0 1.0 18.0/20.0 11.0/15.0 Me 3.0/4.0
321 0.08 max. 2.0 1.0 17.0/19.0 9.0/12.0 Ti 5xC min.
347 0.08 max. 2.0 1.0 17.0/19.0 9.0/13.0 Cb+Ta 10xC
min.
348 0.08 max. 2.0 1.0 17.0/19.0 9.0/13.0 Ta 0.10 max.
Chremium-Martensitic--Hardenable
403 0.15 max 1.0 0.5 11.5/13.0
410 0.15 max 1.0 1.0 11.5/13.5
414 0.15 max 1.0 1.0 11.5/13.5 1.25/2.5
416 0.15 max 1.25 1.0 12.0/14.0 S 0.15 min.
416Se 0.15 max 1.25 1.0 12.0/14.0 Se 0.15 min.
420 Over 0.15 1.0 1.0 12.0/14.0
431 0.20 max. 1.0 1.0 15.0/17.0 1.25/2.5
440A 0.60/0.75 1.0 1.0 16.0/18.0 Mo0.75 max.
440B 0.75/0.95 1.0 1.0 16.0/18.0 Mo0.75 max.
440C 0.95/1.2 .1.0 1.0 16.0/18.0 Mo0.75 max.
Chromium-Ferritic--N on-Herdenable
405 0.08 max. 1.0 1.0 11.5/14.5 Al 1.1/0.3
430 0.12 max. 1.0 1.0 14.0/18.0
430F 0.12 max. 1.25 1.0 14.0/18.0 S 0.15 min.
430FSe 0.12 max. 1.25 1.0 14.0/18.0 Se 0.15 min.
446 0.20 max. 1.50 1.0 23.0/27.0 N2 0.25 max.
Martensitic
501 Over 0.10 1.0 1.0 4.0/6.0 Mo 0.40/0.65
502 0.10 max. 1.0 1.0 4.0/6.0 Me 0040/0.65
ns.
equivalent
yi partially
Mn.
hining
is

3
cation
filler
for ing
cation
for
creep
rate
~~~~.~f~~~l~.
heavy
further
welding; ofthan
cuts; 30.9
used305;
resistance;
exchangers,
tially replacedfurnace
--
r--I ,
t--
I--
I--
restriction
used for
of 302 foraiTcraft Higher
equipment.
-metals.
products,cookware..
load-handling
Ti content
/ prevents I _
Similar
for 304L
321
modification
302
305
'"384
by 3028
More
Extra-low
ses
308
-tance
309
ar
such
Llsed
3095
heat
317
316
310
Same
Q corrosion
group;
304
parts,
tion
of
Low
er
to.
Crl
Mn; carbide
welding
3105
except
(Ni
lent
Ma ar
principally
strength
used
316L
the 301
cation
welded
201
Law-carbon
lawer
content
chemical
cantent;
resistant
used
~for
as
compensate
content
has
increased
cars, er)
content;
af
arsillme
precipitation
(Ni
processing
catian
corrosive
sio."
Law
Higher
than
elements.
cowling,
far
to.
of
r_m-
<) carbon
for
--corrosion.
hardeningrate.because
chemical,
high
trailer railraad
resistance;
cambustian
duringwelding.
and
nickel Ni
filler
allay
301; is
constructian.
heat-treating
erl
and.
carban
316
corrasianand
af
Ni-Cr
has
drawingoperatians.
work-hardeningrate;
used
forming for
and
springs,atchitectural
af is
content:
Ma
because a.ntennas.
316,
work-harden-
to
and
308
304
and
trim.
heat
precipita-
cantent;
equiva-
byincrea-
badies,
equip-
310
nickel
impraves
far
wark-
lower
metals
content
higher;
used
has
Ni
high-
high
forming for
par-
high
far -
as '''-
equip-
ratio
madifi-
lower
food-
nearly
during
spin-
severe
higher;
.~nditions far
madifi-
content used
sca_~n.l?_.
creep.
of SI
-
ofg 202
nto chramium
scaling
has
--- 384,
except ratio.
than
excellent
butaUaycontent
lower
Iin Ieprcarban
Iimprovedweldability.
Law
for
p'u
Higher
furnace
s. h eat
- h and
30~
carra-
allayfar
I i nIstili
madifi-
nickel-
parts.
n9
g.
••.•.
,.,e,df~c:-
ment,
allay
food
rHigher
314
high
can
except
Similar
-
in._'!'lvere
furnace
loss
equipment.
e sitent
s tan
resistance c parts.
welding.
carrosian
phatagraphic. than
temperatures:
increases
higher
e Higher
.302
to silicon
310,
t a alloy (Ni and
and Similar
chambers,to. 309,
welding

- 385

Fig. a.5. Specific applicatians of different grades of austenitic stainless sleels.

(Source: Licolan Handbook)
 l
1
I
i
i

Ii

I:
il
i
!

askets.
hardening
(contains
430F improves
imparts
scaling
forming y ifor
chromium
5e);
character--
nit difficult
I a increases
ch romium
For
acid
for 4305e
weldability.
con
istics
usealloy
decorative
Molybdenum
processing
group
as
handling
modification
442
434
446
436
429
430
involved.
corrosion
gAddition and
often colum-
increased
conlent
nozzles,
bium
forming ; lent
intermittent
resistance
light-cul~
and'
to
I
in
-used
of equipment
type;
Basic nitric
in01
used
temperatures;
automotive
hol-working
Free-machining
furnace
especially
heading
resislance
s t sea
u winter
excellent
automotive
d
compounds.
conditioning
Slightly
and
acontent
chemical
tanks,
trim,
;corrosion
thansulfur-
nitric
addi-
-ofacid.
for such
non-
430
442
service,
for
combustion
improvesused
trim.
or
parts,
may more
atroad-
Iwhere
resistance
trim.
Used
less
for
for
this
nfor
iused'
cold-
for
high
and
beg
in
chambers.
t ion
High
I
chromium High
bearing chromium
atmosphere.

Fig. 8.6. Specific applications of different grades of ferritic stainless steel.

(Source: lincolon Handbook)
 530 Welding Engineering and Technology Weldability of Specific Materials 5:31

8.5.2. Ferritic Stainless Steels sg ~ ~

-e~Q)..!!
CCD'
oC).c>cno~~
-e.~cuQ)occU
Q) en-o'-

These are the non-harden able grades of AISI 400 series that ~:~B ~..,. o.=.ii Q)~=
co 0( Q; :; : c: •.•. -'en ~
contain from about 14 to 27% Cr and no Ni. The chemical o Q)
~ .c:
0-
C'G :;)

.E .~ en
--
CD

:...---f ..•.
~ 3:~Q)ftI«JoQ) •
..•.
..•. ~::;(.)..cN;;~~
compositions of the grades of this are also shown in Table 8.9. :';c~.E~ .?- c: ~ ... ;; ..,. .~ - ~

These steels have such a chromium-to-carbon ratio that does not -§.~
=oeenC'lJ
["i~ ~~ ~:;~; ~~
:.=ooQ)·-.c::00'Cii
t.n(.)_c:.c_u:::l>
~
(I)(,)._~>
allow the formation of austenite so they remain ferritic throughout
their normal operating range. They are magnetic in all conditions. I
I

, The basic alloy of this group is Type AISI 430. C 0 CD·"

o-.c::!!~
•.

"ce.n-CD-
Some of the ferritic stainless steels (for example AISI 405) not ;;~-~~~
u
only have good corrosion resistance but also have good resistance
to high temperature environment. Furnace parts are, therefore,
C>
..•.
..•.
-=-:~~..
<.Joo'" (.)

C CJ'! CD -
(ij
Cll

;;;
:
CI) CI)
Q)Q) - ~
.c: _ (:) C "C ,=, '"
often made from these steel. Other principal applications offerritic .~~ ~'~ ::
."
<lJ

stainless steels are automotive and appliance trim and mufflers, :I:U?-CI)::J"c C
.;a
chemical processing equipment and industrial equipment I ;;;
a
requiring a low-cost material with good corrosion resistance built -0 ••.•.•.
-.:.
with limited welding requirements. Fig. 8.6 shows the applications :c~~~~E
en
~ccnc:"'<U"C
cu ~.5 0 OJ:= c: c
.c;;

o (l)CCDQ) ... g.= c'o c: ~

of each individual grade of this group. eO(l)«s=e
..-cu-::J:::J ... ~u13"O~~..c:
tV
:;;~
8.5.3. Martensitic Stainless Steels ~~ ..•. g ~ .a.~ u :: ~~ ~~~;~:;~ -E~ 0
-eo~:~·~
cu ~'---
..•.
•.• ~cnQ)ULL(Q 0.0
(.) c: •.• «I' OJ cu ~.: CJ) • en CI) "'"
These steels, also of the AlSI 400 series, are the Fe-Cr-C alloys l:::.01U,-"C(.)1I) ..c:OlO"CC:U)OJ
.~ ~ c...: :~ .~ ~ ui ~~ '"c:
capable of being heat-treated, like carbon steels and low alloy .~ ~ -& :; : 'e» ~
:J:(.):C3:;)~~ ::J: e.§ ~ en ._ .0 .E 0 "':::c
-Dc:0
CD -;; -

steels, to a wide range ofhardnesses and strength levels. Hardness ~8

to Rc62 and tensile strength to about 2000 N/mm2 can be CD c:
developed. These steels like ferritic stainless steels are magnetic ii
::::.::;

in all conditions. They are not as corrosion-resistant as the

0>0
C _ 0
_ ..•.
c: _ .-
- ~e
_...,._0
=o.;..:"i8~(5
~-c
cu--ri -

o ~
Cll

austenitic or ferritic types. Base alloy for the group is Type AISI
410. Chemical compositions of different gTades of this group are
en -
-cu-Q)(/)
'"C
.c:,,-<O(I)t:;:
-Q)Q)::Jm
o
•
H~..•.
- 0- <n
.c
'-' c
cu,~
E ~.= :s
en tIi
f--l
••
re
..•.
.c:
uc:(I)"Oo>
cu.~
e «s
CD

(I)
C II) .-
. .~ 'c; '5 ~.~
~
••••

~.~
cn
~-
.2
c en0
CI)

.!::!
::>

:~~ ~ ti.
• (.) «S U
also given in Table 8.9.
Martensitic stainless steels are used to resist abrasion in
-
o
..•.
-0(1)
o
=.~ U)
C:.c: Q;c-
c.Q) •
>- en CD :::.:
cuu..c:
~
.c:ca ~ ~_oC,.)Q)
u.E_.l::.
GJ::::_uoc:
~-g~:c~-a
u..E~g'~~
c.
....
~
CUO-°:Ocurn
Q)::sC.~Eo. "0
steam and gas turbine components, and for such applications as en 0 - «I ~ E C>

good quality cutlery, bearings, expensive hunting knives, and ca-~Q)o::l

COcnCL~::;_Q.. -. C"'''C '"0-
ceo ;:~~2~~1i
~ii
4)(1)'--",-0 CD &Ii
blades that cut bark off trees in a paper pulp mill. Applications of ~
I c~ E~ 'E ~ § .~~.:: ~;o «>
individual grades of this group are given in Fig. 8.7. ~~8~~~
f--I ~ ~ Q..~ ~ ~.; vi :: <::n

8.5.4. Weldability of Austenitic Stainless Steels.

::-~E-g~
'" '" '"
~:! z:-g~-:£
..•.
~'§~EU;-;u"",
_C Q)-..oQ)
ii:
'" ~ '" c cu en
;U;.*£.'~'§;U~ .
••• (I) «lI
cuCU-cnQ) E >-
::J:=CI'Icz <I.) C'U ltl C'»
..J::.-o.s:::.c.u-O(IJ
Austenitic stainless steels are readily joined by arc, resistance, ...
.,.
erE
, • __
O:C«II >-
U)
I- ~ ~§ ~'~!i
i:.= ~ ~ ~e .~ ~ ~ ~ e.=:
:J:ou.s:::.o.cu.r=ELl * '" .:g
friction, electron beam and laser beam welding processes. ..•. Q)CI)-c::Q:C
='~~G):E
However, SMAW, GMAW, GTAW, and FcAw processes are ..c
;~o:e~~
Q) :;:, c:::

------------
commonly used. Plasma arc and SAW are also suitable joining t- ~;;.2 <5 ~
i- C1J.2~0l
processes for austenitic stainless steels. Oxyacetylene welding is, I n-o-;U:
however, not recommended except for emergency repairs. A I : ~~3:~
'" -0
neutral or slightly reducing flame is recommended. A proper flux Li'"
C -0"'--
CU-':::O)
<n

I ..,. Q) « Q).S
selection, careful flame control with minimum puddling and I
I
.s=.

~~GJ~_
. _ > c:
proper heat input is required to make quality welds. I 0 (tJ ••• :;; to
•1_- :z cna..ca5 _ ..I
 ~,:=00
vided
sweld
rates
of
hotter.
used
uffix
~al0_~
••
.c
S00
.....
"1:1
<: ·s
-;
~ •.
••e
~~"1:1
~
=
~~
rn ~"1:1
000
532
00
~CI:lQrn
~
IO•.•..•000
'"
8
~~~~1~~~~~~~~~.~
.•....•~
lQ
~
C'?....
(3
'.gx"';C':i
•.•..•~
0
""d"
~~
l::0
~~~~~~~~~~~~~~~~~~~~
C"I
II•..••
~1§§~~~~~~~~~~~~~~~~
'" r:;":i
0 Q, 0
~ oC"i<':)00000000000000000000
o•.•..•
~~~~~~~~~~~~~~~~~~~~
:~00000
..c
<:'
C'? lQ
••.•r:---
lQf;.
c:'3
J,
~ lQ
C"I
....•
cx:i
0000000000000000
lQ
<III::t4

C!
~•...•••...••~OOOOOOOO~1"""l
~
Ci5
~ CD'"
~co
C"I
9
~.000.000
<':) ~••~.lQ
CDCDCDCDCDCDCDCDCDCDCDCDCDCDCD<0
~~~~~
0') 0
<':)
~.•....•
·OOl.Q~·····
.....;
~ 00
co:)
l
0~ 9t-
000000
o~~~~~~~~~~~~~~~~~~~~
~00000000000000000000 <':)
Q
•.•..•
.6~~~.6"-:~~66~~"-:~666
~ lQ
P"""l
06·dC"I'"
<':)
P"""l(j)
lQ
600
<':)
C"I
v trJ trJ~ trJ trJ trJ trJ trJ trJ trJ trJ trJ trJ trJ trJ trJ trJ trJ trJ trJ 00
C!<':) C!C':i0
'" 6•..••
OOO')O)O')(7)Q')
~ ......•
~~
<':)
~~
'"~6C"I
<':) '1'"<':)~C"I0~<':)c!,lQ
•.•..••.•..••.•..••.•..••.•..••.•..••.•..••.•..••.•..•
0
<':) C"I<':)C"I
~~~<':)<':)C'l
....• <':) t-
<':) C"I
0
<':) C"I
0
<':) 7~<':)C!
C'1
~
0
<':)
"'"-i-ie'\l~"'C;ft..qt~o~~-.::tt&Ql.()~
••••••••••••••••••.•••••

~ 0
C"\I
lQ
~
•••••••••••
"'" ~

<':)
~~~~~~~~~~~~~~~~
co:)
~<':)
6..•..••
C'\t
~~
0 ·~~o
<':)
Welding
Cf1
•.•
~ ~~ 0
•.•..••.•..•
lQ
<':)
~
0000
<':)
to
<':)
'0000
•.•..• 00
C!C':i
C'I C'1
<':)
~
<':)
<':)
l.()
0 lQ
<':)
C!Engineering
<':)
CIO
~
C"I C"I
C"I C"I
lQ
<':)
<':)
C"I
~ <':)~ ;;t
~ ~
0<':)
<':)
•••••.••••••••••.••••••.•••••.•••••••••••••••••

Weldability
C':i ;;t
~
I.•.....••...•••...•••...•••...•••...••
~.C"I C!7
.•..•.••.•..• ~OO
~c!,'I~ ~· 'I ~7
C"I
~......
&()
0~
~ •....••....•
<':)
<':) '7~of
00
C"I
<':)~ ·and
~.s
.S ~S~
0~
.•....•
<':)
<':) ~~
0<':)
~~
~ l::
~ ~Technology
<':)
C"')M
'f"""'l

trJ Specific
co:)
~
•.•..••.•..••.•..•.••••••.•..••.•..•
~8 ~ 0~
•.•..••.•..••.•..••.•..••.•..••.•..•
tv:)C'\I
C! ~
<':)
'-'
<':)
~ 6
••••••

co:)
0~
C'?
C'?5.33
.••••••
'P"""f

~ Materials 1~-~3~
; ~~~~~~~22222~2222222222
~!i
~ •. ~I~
Q-
as =
Q Q.,8l c:)
_000
"1:1 Q;
g;§
o~01000
'" rn
v ~~~~~
0"1:1
"1:1
~&3
~bO
'"'5-
Q,0
•....•
I
Q" o~000
00~0t-
0~00~•...••....•
'" 00 0~ 0~00~00~00~00~000
'" C!;
00 00
C!;
00 00
0~ :J::....:l::S-u....:l:J::....:l
00 ~ 0~00~00~0t-0~0....:l
0~ ~0~ CDCDCD
'" '" t-
0~00 0
C!;
00
@
10 ~ 0 0
 N1.0-2.5
lOx
Table Nickel,
8.10P
0.015
0.02
Cb
Single
0.40
3.0-4.0
3.25-4.00
S 0.75-1.50
Ei" ----
~::I0.75
0.6
9.0-10.5
11.5-13.5
Cr
Ni
J4.0-37.0
16.0-16.75
19.0-21.5
19.0-21.0
15.5-17.0
0.60
0.5015.0-17.0
9.0-11.0
Mo0.75
25.0-27.5
14.5-16.5
11.0-12.5
18.5-20.5
0.03
0.15-0.30
8xCmin
contd.
Titanium Cu
0.75Si0.75
C min
1.0-2.5
Mn
0.75
1.0-2.5
19.0-21.0
0.35-0.65 -2.5
19.0-21.5
12.0-14.0
8.0-9.5
4.0-5.0
1.0-1.4
0.6
hTa
0.18-0.25
0.25-0.75
0.400.75
0.03
32.0-36.0
Q.4-o.7
1.0-2.0
0.15
7.5-9.5
2.0-3.0
1.5-2.0
4.5-5.0
0.30-0.65
0.04 plus0.015
0.020
0.02
values
max 0.12
0.07
-0.10
0.25-0.40
0.01
0.10
0.06
9C --
~0.03
IJQ
x0.5
0.05
0.025
0.07-0.13 C
shown
0.08 1.0-2.5
tomin
0.30 percent;
tomaximum
1.0 max.
are0.03
minus the
-
tungsten
0.75
copper content,- percent.
percentages 1.25 to 1.75
except percent.
Composition,
where percentb
otherwise specified.
::I
Q
!\;
::l"
Co
~t') ::I

I~;3
t!I!j
::!.
::I
IJQ
C5' is:
reported0.20h
to the
IJQ
Ei"
~

5 S~
'tl
='
...•
•..
~
'"
!\;
!\;
Low
310S
310
308L
309
308309
308
347
308
310
316
310
201
Table
309
202
309S
carbon
5/12
309
316L
317
308L 919
2347
5120
1919CbLTa
Extra
2nd
309Cb
309Cb
309Cb
18112Mo
Extra ----
119114Mo
Elc
Choice
1stlow
18112Elc
18188.11:
1818
1919Cb
Difficult low Free
Substitue
carbon
Choice Substitute
carbon
Recommended
Recommended
to weldHigh
in heavy
-
machining
for
for
Filler
Silicon
~301
- welding
302
Low carbon
Filler
Metal
sections
-
not
Metal
FreeAlloys
Popular Name
;- Remarks5;
~go
~
recommended-312
t')
for Welding
machining - welding Various
!:I~ .,t')~ is:
Grades of Stainless Steels
not recommended-312 Q

316
(Tabl/!Steel
contd)to be Welded
536 Welding Engineering and Technology Weldability of Specific Materials 537

8.5.4.2. Welding Procedure

The austenitic stainless steels have about 45% higher thermal
coefficient of expansion, higher electrical resistance, and lower
thermal conductivity than the mild steels. High welding speed is
recommended which will reduce heat input, reduce carbide
a:l a:l "0 precipitation, and minimise distortion. The melting point of
., 1"0 "0 "0 11 ] austenitic stainless steels is slightly lower than mild steel.
"f: ~ ~ ~ ~ s s Because of lower melting point and lower thermal conductivity,
~ 8 ~ ~ ~ § § welding current required is usually lower. The higher thermal
~ ~ § § § ~ ::: expansion necessitates t~e ne~d to take special precautions. with
g ~ ~ ~ c; c; regard to warpage and dIstortIOn. Tack welds should-be tWIce as
~ c; c; c; ~ ~ many as normal. Any of the distortion reducing techniques such
:§ ~ ; ; ~ ~ as back-step welding, skip welding, etc. should be used. However,

~
f
.-
~i
~
~ ~~ tl tl
c: I I I
;§ ~ ~ ~
._.-
0:; 0:;
>, >,
on thin materials it is very difficult to completely avoid buckling
and distortion .
.,.,
~ ~ .§ ~0 ~§ ~§ ~§ .S.S
..c:..c: '0E '0
E 8.5.4.3. Weld Metal
~ 8 8 5 5 5 ~ ~ ~ ~ The microstructures of austenitic stainless steels are quite
-: ~ -fo -fo -fo -fo ~ ~ 0 8 different from those of similar wrought base metals. Alloys those
::3 ~::a ::a::a::a ~ ~ i5 c3 are fully austenitic in the wrought form often contain small pools
'" I of ferrite in weld metal of equivalent composition. With some filler
§ 0 ~ ~ metals, such as Type ER 16-8-2 (see Table 8.10), small amounts
::: I c3 I I I : I I I I I c3 I I I ~ ~ of martensite can form in the weld metal. Carbides might also be
~ ~ (".) ~ "2 "2 present in the weld metal, particularly with Ti and Cb stabilised
§- ~ ;;5 ;;5 filler metals, such as ER321 and ER347. The microstructure of
i:l;. as-deposited weld metal is dependent upon its composition and
11 '" solidification rate. The as-welded microstructure can be predicted
~] 0 c3 0'> 0 from the chemical composition of the deposited weld metal using
~ I ~ I I I I I I I I ~ g I I ~ I I
c..l the well-known empirical constitution diagrams uiz., Schaeffler
~ ~ """ and the Delong diagrams.
~ Schaeffler Diagram: Schaeffler diagram shown in Fig. 8.8
~ .~ relates the nickel and chromium equivalents to lines which show the
o c5 ~~.~.~ S
~ """~. S g
~ I """ """ g
""" I I I S """ I I ~
~ ~
.,., ~
.,., %age of ferrite. This diagram is useful for estimating .•the
~ tl microstructure of the weld deposit and the filler metal composItIon
~ ...• required to produce the prescribed amount of ferrite in the deposit.
">.:l ~ OJ < a:l (".) r.:. cJ5 The diagram shows how the microstructure of the weld deposit is
=ci :g 03 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ g ~ ~ g ~ affected by the alloying elements in the stainless steel, based on those
~o ~ ~ """ """"""""" """~ that act like Ni (austenite stabilisers) and those that act like Cr
~ .15 (fenite stabilisers). The Ni-equivalent group includes Ni and the
::: -2 effects of C and Mn. The Cr-equivalent includes the effectsof Cr, Mo,
cO
~.::/ Ol oqIsua:j.nlW-.lO
•••• oqU.laJ-.lr'o
'" oJ . an d Cb. m
SI, .
LO estImate t h'e mIcrostru ct ure 0f a deposl't th e C r an d
~ C"/J Ni equivalents are calculated using the followingequations.
Cr-equivalent = % Cr + % Mo + 1.5 (% Si) + 0.5(% Cb) (8.4)
Ni-equivalent = % Ni + 30(% C) + 0.5(% Mn) (8,5)
 538 Welding Engineering and Technology Weldability or Specific Materials 539
J!
~UiDJad ' a~!JJa~ 'c Show the variation of weld metal composition by marking a
~ ~ line on the Schaeffler diagram. What is the midway composition
N !I-.<:
.•....••.
..z :;~:C: of weld metal as depicted by the marked line?
'S o e-~~
co"0 Solution.
-.:t .J::j
,....
U ••
o ~ ••...
.g ~ (
..
Q.-.-~ •... ~
.~. Using equations (8.4) and (8.5), determine the Cr and Ni
0: ••...••
::J c: C;
-o*-
t'f'\ X
,.... .
c •... o
.-o: ••.. ..c: ~ equivalents of the base metal and the filler metal as follows.
00'-
.D ._ '- l:l
....•. .•... u:' ~
.......• a-
,g~go (a) For Base Metal (AISI 304L)
""~f:e
c::>
tl
.!~.:
)1('-.1::, ~ .~ . ·s Cr-Eq = %Cr + %Mo+ 1.5%Si+ 0.5%Cb
00:"" V) .; ~ e:.;
L. :!. ::s
ii]
<II .
0.. •• 0
0.."' •... ClO~ c= g.e1§~
QI GI:t: •••.• Putting the values of different elements, we get,
~ •• )l N X CTlO:ZEE
Ln o & ~.a.a ~<II
~ S Cr-Eq = 2.0 + 0 + 1.5 x 1 + 0 = 21.5 ...(i)
..z +
•... ff
/"-°SEE
:t::J: •..
:Z:z~£-6
<d-.::l
Sa:; Similarly, Ni-Eq = %Ni + 30%C + 0.5%Mn
N 0 '.0 ~
::E [l·Gi = 8 + 30 x 0.03 + 0.5 x 2
•.• .8
o~ sa '" = 8 + 0.9 + 1.0 = 9.9 ."Cii)
N
'-
-+
S

Ll
•. gj
B (21.5, 9.9)
J! LJ 0..
~] ... (ii-i)

'-'
....
'c
J!
..
'"
<
-o~
•... ~
-c
OJ
~
..
>C

!:.
~
.;;1 .5
-.::I
.......
<II 0
'"
(b) For Filler Metal (AlSI 347)
Cr-Eq = %Cr + %Mo+ 1.5%Si+ 0.5%Cb
= 17 + 0 + 1.5 x 1.0 + 0.5 x 0.5
Ne:; Ci ~
o or- > <d
..d = 17 + 0 + 1.5 + 0.25 = 18.75 ...(iv)
'5
c:r
Ci. <J
'(/')

ex) Q/
.2]
.....• o <Xl
aq
Ni-Eq = %Ni + 30%C + 0.5%Mn
= 12 + 30 x 0.08 + 0.5 x 2.0
:
C> •..•
E o:.c
••.•
::l 'S.!:
C.
••...bO
.

~ 0'- 'E ....

'" .. '"
~
.c c:
u c •• c e
~'~ ~ :L 10
5~ o '" o~
:j co 00 ~•...
e"'
,,'- ..
LJ~~
~.5
Vl~
)(
U'"l
..., IIT":Jf0f;:-l[:i/'/
( UW%X 5,0 +)%X O(+!N %HuaIDA!oba la~P!N <:>
+ 2()

The values obtained are marked on the coordinates of the ....•

o~
diagram and a point is located. The microstructure of the point is )( 40
the one predicted for a deposit of that composition. It is possible <:>
,...,
to plot the composition of the filler wire and the composition of +

the base metal and connect them with a line and the resulting Z
o~
weld would be along this line. By the use of Schaeffler diagram,
it is possible to select a filler metal that will avoid ferrite or - II
c:
martensite, as desired, in the stainless steel weld deposit. The ..!!!
o
>
100
Schaeffler diagram can also be used to predict the composition of a
::;
the weld deposit when welding dissimilar stainless steels. Q,I

Example 8.1. An AISI stainless steel 304L is to be welded 0;

.¥
using filler alloy AISI347. The compositions of base material and .~
:z
the {iller alloy are as follows:
Material Type %C %Mn %Si %Cr %Ni %Ta %Cb 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 3638 40
AISI304L 0.03 2.0 1.0 20.0 8.0 Chromium Equivalent =0/0 Cr +0/0 Mo + 1.5 x %Si + 0·5 x % Cb
AISI347 0.08 2.0 1.0 17.0 12.0 0.5 0.5
Fig. 8.9. Schaeffier diagram representing the weld metal composition by line SF.

I
 Weldability of Specific Materials 541
540 Welding Engineering ane:!Technology

12 + 2.4 + 1.0 == 15.4

== (v)
measurement systems. Therefore, an arbitrary standard or ferrite
number (FN) has been established to designate the ferrite content.
F (18.75,15.4) (vi) At low levels of ferrite number and percentage of ferrite are
Marking these points (B and F) on the 8chaffler diagram of identical, however, the two values diverge significantly when large
Fig. 8.9, we get the required line BF representing the weld metal amounts of ferrite are present in the stainless steel.
composition. The mid-way point M indicates that it represents a Using equations (8.6) and (8.7) the relevant points can be
weld metal composition with 5% ferrite and 95% austenite. marked on the Delong diagram and the weld metal phase
Ans. Ferrite == 5% constituents can be determined.
Austenite == 95%
Example 8.2. Stainless steel AlSI202 is to be welded using a
Delong Diagram. The Delong constitution diagram shown in filler wire of stainless steel grade AISI 308. The chemical
Fig. 8.10 is based on the fact that it takes into account the effect compositions of the two steels are as follows:
c: Material Type %C %Mn %Si %Cr %Ni %N2
::;:
21 AlSI202 0.15 10 1.0 17.0 4.0 0.25
°~x
u:' 20 AISI 308 0.05 2.0 1.0 21.0 12.0
<::>
+
:z 19 Determine the maximum and minimum amounts of ferrite
~ Austenite phase possible in the weld metal.
"x
<::> 18 Solution
""
+ -
'-' ",17 Because the base metal stainless steel contains nitrogen, so
~ '-
').( .~ 16
C);-:::
we determine the Cr-equivalent and the Ni-equivalent by using
"".0 equations (8.6) and (8.7).
.~
:z .E'" 15 (a) For Base metal (AISI 202)
~ ~14
--c:
-5
- 13
Cr-equivalent == %Cr + %Mo + 1.5 (%8i) + 0.5 (%Cb)
== 17.0 + 0 + 1.5 x 1.0 + 0.5 x 0
> Austenite. ferrite == 17.0 + 1.5 == 18.5 ... (vii)
'5
cr
<U Ni-equivalent % Ni +30(% C) + 30( % N) + 0.5(% Mn)
==

== 4.0 + 30 x 0.15 + 30 x 0.25 + 0.5 x 10

~u
QJ

Z == 4.0 + 4.5 + 7.5 + 5.0 == 21.0 ... (viii)

27
17 18 19 20 21 22 23 24 25 26 .. Base composition can be represented by point B i.e.
Chromium equivalent (% Cr +% Mo + '·5 x %. Si+O·5 xO/o Cb I B <18.5,21.0) ... (ix)
( a - stabilisers ) (b) For Filler wire (AISI 308)
Fig. 8.10. Delong constitution diagram. Cr-equivalent == % Cr + % Mo + 1.5(% 81) + 0.5 (% Cb)
== 21.0 + 0 + 1.5 x 1.0 + 0.5 x 0
of nitrogen on the phase balance of austenitic stainless steel. The
== 21.0 + 1.5 == 22.5 ...(x)
Cr equivalent (8.4) and the Ni-equivalent (8.5) equations are thus
modified and are represented as follows. Ni-equivalent == % Ni + 30(% C) + 30(% N) + 0.5 (% Mn)
== 12.0 + 30 x 0.05 + 30 x 0 + 0.5 x 2.0
Cr-equivalent == %Cr + %Mo+ 1.5 (%81)+ 0.5 (%Cb) ...(8.6) == 12.0 + 1.5 + 0 + 1.0 == 14.5 ... (xi)
Ni-equivalent == %Ni + 30 (%C) + 30 (%N) + 0.5 (%Mn) .. the filler wire composition can be represented by point F
...(8.7)
l.e. F(22.5, 14.5) ... (xii)
Percentage of ferrite in 8chaffler's work was determined using
metallographic measurement methods. Magnetic measuring For determining the composition of the weld deposit the two
devices are more convenient but these are secondary points i.e. B(18.5, 21.0) and F(22.5, 14.5) representing base metal

1
542 Welding Engineering and Technology
I the weld joint; for example hardness increases but ductility, notch
apd filler wire compositions respectively are marked on the Delong
diagram of Fig. 8.10; from which it is evident that in the weld
deposit the maximum and minimum ferrite contents are 10.7%
and 0 % respectively. Ans. 00.70(0, 0.0 %)
8.5.4.4. Problems Associated With Welding of Austenitic
Stainless Steels
Specific problems associated with the welding of austenitic
stainless steels include the following.
(i) Ferrite and Sigma phase formations,
(£i) Carbide precipitation,
(iii) Knife Edge attack, and
(iv) Stress corrosion cracking.
Salient features of these problems and the suggested remedial
measures are described briefly in this section.
(i) Ferrite and Sigma Phase Formations. A characteristic
of austenitic stainless steels is their susceptibility to hot-short
cracking after being welded. This is sometimes called
microfissuring. This problem is corrected by using electrodes that
deposit weld metal containing a small amount of ferrite. Thus
recommended electrodes for many standard austenitic stainless
steels may deposit weld metal that contains 3 to 10% ferrite even
though the same steel base metal contains no ferrite. Since ferrite
is magnetic, it is easily detected in an otherwise non-magnetic
weldment. Thus, the composition of the filler metal should be so
selected as to get a weld deposit containing a small %age of ferrite.
However, the ferrite content should not become too high or else
the weldment will have lower than desired impact strength. For
low temperature service the weld metal should have the ferrite
content in the range of 4 to 10%.
Presence of ferrite can, howevf>.r,lead to the formation of
various non-metallic phases like sigma phase and Chi Fhase.
Higher ferrite content tends to accelerate these transformations
and lower the temperature where transformations will occur.
Sigma phase is a crystallographic constituent that forms
slowly at higher temperatures in r.igh chromium (20% or more)
ferrites. Generally sigma phase transformation occurs in the
non-Molybdenum-bearing stainle:;s steels, and both sigma and
chi-phase transformations occur in Mo-bearing stainless steels.
Nucleation of these phases occurs at the interphase Caustenite-
ferrite) boundaries, and growth is primarily toward the centre of
the ferrite and sometimes in austenite itself. These intermetallic
phases have detrimental effects on the mechanical properties of
I
r Weldability or Specific Materials 543
toughness as well as corrosion resistance of steel are decreased.
Because of its slow development in the range of 480° to 900°C
sigma phase is primarily a service problem where long exposure
at high temperatures are involved. It does not normally develop
during welding or even during PWHT. These phases can be
redissolved by heating to above 900°C for a relatively short time.
Complete conversion of sigma to austenite, however, usually
requires heating to as high a ~emperature as 1230°C.
(ii) Carbide Precipitation: No doubt mechanically satis-
factory welds can be made on the Cr-Ni (austenitic) stainless
steels, heating these materials sometimes promotes intergranular
corrosion. When the austenitic stainless steels are heated in the
range of 425°-845°C, the so called sensitising temperature range,
or cooled slowly through that range, carbon is precipitated from
solid solution (mainly at grain boundaries) and unites with
chromium to form Cr-rich carbides (M;3C6)'where M represents
presence of some small amount of iron alongwith chromium. These
Cr-carbides may eontain as much as 90% Cr, taken from the metal
immediately adjacent to the grain boundary. The chromium
content of the adjacent metal is thus reduced, and corrosion
resistance is seriously impaired. This phenomenon is termed
Ca.rbide precipita.tion and the type of corrosion it promotes is
known as intergranular corrosion. Stainless steels having
chromium contents to about 18% are most susceptible to carbide
precipitation. A corrosion-resistant pipe or tube, or a reactor vessel
in a chemical plant can literally come apart at the welded seams,
or right next to them, when carbide precipitation has occurred.
The formation of sensitization region in an unstabilised austenitic
stainless steel during welding can be explained by the thermal
T
Pre c:ipitation range
of chromium carbide
time
•
a
Fig. 8.11. Schematic representation of thermal cycles and the sensitization
region in the HAZ of an unstahilised austenitic stainless steel. .
544 Welding Engineering and Technology Weldability of Specific Materials 545

cycles at different distances from the fusion boundary and the The thermal conditions that produce carbide precipitation can
sensitization range, as shown in Fig. 8.11. At position 'a' the occur during welding-particularly in multi-pass welding, welding
cooling rate through the precipitation range is too high to allow of heavy beads, or when two beads cross each other. The most
precipitation to occur at position 'b', on the other hand, the serious problem when welding stainless steels is to avoid carbide
retention time of the material in the sensitization temperature precipitation.
range is sufficiently long and precipitation takes place. At position Remedies : There are four ways of avoiding carbide precipi-
'c' the peak temperature is too low to permit any precipitation. The tation in stainless steel welds. These are:
above explanation is confirmed by the experimental results shown (a) Thermal treatment,
in Fig. 8.12..
(b) Use of extra low carbon steels,
Area of weld decay
(c) Use of stabilised stainless steels,
---'=1. a
Jg.
bed
· (d) Solution annealing.
Brief descriptions of these methods follows.
(a) Thermal Treatment : One way to avoid carbide
precipitation is to make the weld and the base metal coolvery fast
through the dangerous temperature zone of 845° to 425°C so that
1000 the Cr does not have enough time to react with carbon to form
.w Cr-Carbide. This can be achieved by water-quenching the
.,;
.- 870 weldment ..
::>
+-
o'- (b) Use of Extra-Low Carbon (ELC) Stainless Steels:
a.
OJ Another remedy to the carbide precipitation problem is to use
E stainless steels and electrodes having such a low carbon content
~ 500
that no carbides can be precipitated. The austenitic !;!tainless
steels retain about 0.02% carbon in solid solution under all
conditions. With carbon content increased to 0.08% the amount of
carbon that can be precipitated increases slowly above that
amount, precipitation can occur rapidly when the material is
exposed to the sensitising temperature (425 to 845°C).
10 20 Extra-low-carbon steel plate and electrodes cost slight more
Time. sec. than the normal grades but the results are very satisfactory
Fig. 8.12. Measured thermal cycles based on practical data for Arc Welding AISI particularly when the weldments are to be used in the as-welded
304 stainless steel and the corresponding area of weld decay. condition. Carbide precipitation decreases significantly in
Weld metal containing a small %age of o-ferrite is not changing from say AISI 304 (0.08%C)to AlSI 304L (0.03%C)grade.
susceptible to such severe sensitization as wrought material of (c) Use of Stabilised Stainless Steels: The third solution
similar composition. In weld metal, carbide precipitation occurs is to use austenitic stainless steel base metal that has special
more uniformly throughout the structure at the interphase or elements, usually columbium (or Niobium), or Titanium, or
substructure boundaries in preference to the grain boundaries. tantalum, in it to keep the Cr from forming Cr-carbides. Such
Therefore the degree of Cr-depletion along the grain boundaries stainless steels are known as stabilized steels. Because Cb or Nb,
is lower than the HAZ. Consequently, weld metal is not generally Ti, and Ta have higher affinity for carbon than Cr, therefore in
considered to be so susceptible to intergranular corrosion as the the stabilised stainless steels Cb or Ti or Ta carbides are formed
base metaL Nevertheless, weld decay or intergranular corrosion rather than Cr-carbide. Thus, Cr remains in solid-solution in the
can occur in weld metal exposed to severely corrosive environ- weld metal and base metal and keeps the stainless steel weldment
ment.
546 Welding Engineering and Technology
Weldability of Specific Materials 547
corrosion-resistant. These stabilised stainless steels (e.g. AISI 347 precipitation of Cr23C6 at grain boundaries. Knife-line attack
and AISI 321) can be used in the sensitization temperature range differs from weld decay in two ways:
(425°-845°C) with no effect on corrosion resistance, and no PWHT (i) Knife-line attack occurs in a narrow band in the area
is required. immediately adjacent to the weld metal;
Welding of stabilised stainless steels should be done with (ii) Knife-line attack occurs in stabilised stainless steels.
stabilised steel electrodes to get the best results. Since Cb The knife-line attack of stabilised austenitic stainless steels,
transfers through an arc much more effectively than Ti orYa, such as AISI 321, by corrosion media can be explained with the
stabilisationof electrodes is achieved normally with Cb additions. help of Fig. 8.13. The area corresponding to thermal cycle 'a' is
(d) Solution Annealing: The fourth method of avoiding the
very close to the fusion boundary and is therefore, subjected to
carbide precipitation is to reheat the entire weldment to about both, a very high peak temperature and a very rapid cooling rate
during welding. Since this peak temperature is above the solvus
l040°C and hold it at that temperature for a length of time that
depends on the thickness of the base metal. This solution-anneal
heat treatment will re-dissolve the Cr-Carbide and put the Cr back
into solution. Then the entire weldment must be cooled quickly
through the dangerous sensitisation temperature range (845°- 1400' " NbC
'" '"
/

425°C) so that the problem will not occur again.

This fourth solution can be quite difficult if the weldment is
-"
very large, for example, a chemical tank which will not fit in 1000
heat-treating furnace. T. ·c
800
8.5.4.5. Knife Line Attack
Although stabilised austenitic stainless steels are not 600
susceptible to weld decay, they can be susceptible to a different
type of intergranular corrosion attack, the so-called knife-line 400
attack. Like weld decay, knifeline attack is also due to the 18- 8 Stainless steel
200
0·05 0·10 0·15 0·20
T (,'Yo
of titanium carbide
Prec ipitation range Fig. 8.14. Solvus temperature for CI'23Caand TiC in AISI 304 steel
of titanium carbide as a function of carbon content. The hatched areas indicate
the precipitation temperature range.
temperature of Ti-carbide, Fig. 8.14, Ti-carbide dissolves in this
area. Because of the rapid cooling rate through its precipitation
temperature range Ti-carbide does not re-precipitate during
cooling, thus leaving enough free carbon atoms in this area. When
la) Time the weld is then heated in Cr-carbide precipitation range (for
stress relief or elevated temperature service), Ti-carbide does not
0/1 .b form appreciably, since the temperature level is not high enough
( b) (Fig. 8.14). Consequently, Cr-carbide precipitates at grain
boundaries, and the area is susceptible to inter-granular corrosion
Fig. 8.13. Schematic illustration of: (a) thermal cycles, and (b) knife-line attack attack. The area corresponding to thermal cycle 'b', however, is
region in the HAZ of a stabilised austenitic stainless steel. not susceptible to such an attack owing to its low peak
548 Welding Engineering and Technology
Weldability of Specific Materials 549
temperature. Because of the very high temperature gradient near
the fusion boundary (this is especially true for austenitic stainless precipitation at grain boundaries. Fig. 8.16, for example, shows
steels owing to low thermal conductivity) and the high Ti-carbide the reduction in the rate of intergranular corrosion in AISI 347
dissolution temperature, the region in which Ti-carbide dissolves stainless steel by the addition of REM. Similar results have been
during welding is very narrow. As a result, subsequent -3
_10
intergranular corrosion attack takes place in a very narrow band 30
immediately adjacent to the hatched area of Fig. 8.13 . .I:
Problem 8.3. Explain why knife-line attack can occur in the "'e 2S
SUS 347· REM o SUS 347.011 REM
two-pass weld shown in Fig. 8. 15(a). Suggest two simple ways of ....
SUS 347 .6 SUS 347
...•..
avoiding the problem. C' 20

'C._ h, c::
~":"
., .. ~ •..
t••••
,
:;'y ';:: .:~;
...~.:v~ . o
'iij
;..
.;:.
~{~~: .. ,,'Ii
...•. .

\ I ~.' ..
Knifoz linoz attack
(a) Corrosive (b) (or rosive
Environment
Environment
Fig. 8.15. Two-pass weld of a sensitised stainless steel: (a) susceptible to
knife-line attack, and (b) immune from knife-line attack ..
Solution: During the deposition of the bottom pass as the first
pass, Ti (or Nb) carbide near its fusion boundary dissolves. As
shown in Fig. 8.15(a) the sensitisation temperature range of the
second pass (i.e. the top pass) overlaps the lower part of the fusion
boundary of the first pass (shown as hatched area in the figure)
and causes Cr-carbid~ to precipitate. Since these areas are exposed
to the corrosive environment, they are susceptible to knife-line'
attack. The problem can be corrected simply by reversing the
welding sequence i.e. by making the second pass from the side of
the work piece that is to be exposed to the corrosive environment.
If the second pass cannot be made from that side, however, the
joint design and the welding parameters can be readjusted so ,that
the sensitisation temperature range of the second pass overlaps
the middle rather than the lower part of the fusion boundary of
the first pass, as shown in Fig. 8.15(b). Since the areas where the
Cr-carbide precipitation are not exposed to the corrosive
environment, knife-line attack is no longer a problem.
The second solution for avoiding the problem is the use of rare
earth metals (REM) combination of Lanthanum (La) + Cerium
(Ce) which is reported to effectively reduce the knife-line attack
ofstabilised stainless steels. High resolution eleCtron micrographs
reveal that carbide precipitation within grains is accelerated in
these materials, thus leaving fewer free carbon atoms for carbide
o
•...
•...
uo
o
As w.lded 1 10 100
Sensitizing time I hI
Fig. 8.16. Effect of addition of rare-earth metals (REM) on the knife-line
attack sensitivity of AISI 347 stainless steel.
observed in AISI 321 stainless steel. The addition of small
amounts of REMs in such stabilised stainless steels has been
observed to have no adverse effects on their mechanical properties.
The knife-line attack of stabilised stainless steels can also be
avoided by dissolving Cr-carbide at 10000-11000 C. Since Ti-car-
. bide and Nb-carbide form in this temperature range (See Fig.
8.14), subsequent rapid quenching (as required in the case of
avoiding weld decay) is not necessary. As mentioned previously,
the addition of REMs can also help avoid knife-line attack in these
materials. Finally, as in the case of avoiding weld decay, low carbon
grades are also helpful.
The susceptibility of austenitic stainless steels to weld decay
or knife-line attack can be evaluated by Huey test in which the
sensitised weldment is exposed to boiling 65% HN03 fo.r five 48
hour periods. A faster method is the Streicher test which consists
of polishing a small specimen of weld me nt, through No. 3/0 emery
paper, etching in 10% Oxalic acid for 1.5 minute under an applied
current density of 1A/cm2, and then examining the surface at 250X
to 500X magnification. The specimen is made the anode, and a
stainless steel beaker is used as the cathode.
8.5.4.6. Stress Corrosion Cracking (SCC)
Stress corrosion cracking of austenitic stainless steels can
occur when the material is subjected simultaneously to a tensile
~ri~f"'-

550 Welding Engineering and Technology Weldability of Specific Materials 55!

stress and specific corrosive media. The cracking can be ~ither Methods of controlling see include:
intergranular or transgranular. lntergranular see can occur even (i) a stress-relieving or annealing heat treatment to reduce
though the material is 'insensitive to intergranular corrosion or stress to a safe level;
weld decay.
(ii) substitution of more resistant high-nickel or straight
The ions of the halogen family (fluorine, chlorine, bromine, er-stainless steels; or
and iodine) are largely responsible for the see of austenitic
(iii) removal of chlorides and O2 from the environment.
stainless ste.els. Of the halides, the chloride ion causes the greatest
number of failures. For example see may occur in the weld HAZ. The common test for checking susceptibility to see is by
Such cracking is usually transgranular and branching, as can be immersing a stainless steel sample in boiling 41% magnesium
seen in Fig. 8.17. chloride solution.
8.5.4.7. Postweld Heat Treatment of Austenitic Stainless
Steel Weldments
Most austenitic stainless steel weldments do not require
PWHT. However a heat treatment is sometimes used to improve
corrosion resistance or to relieve stresses, or both. If maximum
corrosion resistance is required a solution annealing treatment
must be given as described in the previous section ..
The optimal solution-annealing temperature depends on the
type of stainless steel, and the soaking time is determined by the
section thickness. Table 8.12 gives the recommended annealing
temperatures for some of the well known austenitic stainless
steels. A short time at the annealing temperature is preferred to
avoid grain growth. As a thumb rule, soaking time at annealing
temperature should be abcut 1 minute per mm of thickness. The
weldment must be cooled rapidly and uniformly, at least through
the temperature range of 900° to 425°e, to retain carbon in solid
solution. Water quenching or spraying is necessary for thick
sections, but air-cooling is suitable for thin sections.
Table 8.12. Recommended Solution-Annealing Temperatures
for Austenitic Stainless Steels
S. Stainless Steel Type Solution-AnnelJ.~ing
No. (AlSI Grades) Temperature roC}
1. 201,202,301,302,303, 304,304L, 305, 308 1010-1120
2. 309,3098,316 1040-1120
3. 316L,317L 1040-1105
Fig. 8.17. Stress corrosion cracking in AISI 316 stainless steel. 4. 1 317 1065-1120
Austenitic stainless steel is also subjected to see in high
concentration of hydroxyl (OIr) ions or in lower concentration
solutions where service conditions can promote dead"zone for the ~47,5. \321 348 955-1065
980-1065
chemical(s). lntergranular see can occur in oxygenated boiling While PWHT below the annealing temperature can reduce
water environments and in pressurised water reactors in stagnant residual stresses, they can also destroy several essential or
borated water environments. desirable properties of austenitic stainless steels. Even short time
552 Welding Engineering and Technology
exposure to temperatures in the 425(>to 900°C range might
sensitize and reduce corrosion resistance of the steel unless it is
of low carbon or stabilised type. There is some evidence that even
low carbon and stabilised. type stainless steels can suffer carbide
precipitation.
PWHT of austeniti<; stainless steels can be complicated when
the weldment is made partly from carbon or low alloy steel and
partly from austenitic stainless steel; for example, a carbon steel
pressure vessel lined with austenitic stainless steel. It is essential
that for any dissimilar steel fabrication, a heat treating cycle
suitable for both steels must be selected. Main consideration must
be given to the tempering of a low alloy steel while preserving the
corrosion resistance and ductility of stainless steel. Stress-
relieving of dissimilar steel weldments is difficult because
differences in coefficients of thermal expansion can re-establish
high residual stresses during cooling, regardless of cooling rate or
procedure.
8.5.5. Weldability of Ferritic Stainless Steels
Ferritic stainless steels can be welded by SMAW process
employing filler metal of matching composition or austenitic
stainless steel and nickel alloys. Othf!r processes used are GTAW,
GMAW,and SAW.However, spot and seam welding ofthlilse steels
is not recommended when weld ductility is the important criteria.
Flash welding is used to join ferritic stainless steels, provided the
low ductility associated with welds in these steels can be tolerated.
The welding quality of ferritic stainless steels is the poorest
01 the AISI types. The welding heats a zone in the base metal above
a critical temperature (955°C) and causes rapid grain growth of
the ferrite. This coarse grain zone lacks ductility and toughness,
and a small amount of martensite may be present, which leads to
increase in hardness.
Excessive grain growth can be avoided, of course,' by using
lower welding heat inputs. It has also been suggested that nitride
and carbide formers such as B, AI,V,and Zr can be added to ferritic
stainless steels to supprE!ssgrain growth during welding.
Ferritic stainless steels, though considered as less weldable
than austenitic stainless steels, have recently drawn greater
attention owing to their lower costs and better resistance to SCC
than austenitic stainless steels. The most commonly used ferritic
stainless steel is perhaps the AlSI 430 grade, which contains 17%
Cr and 0.05-0.12% carbon.
553
Weldability of Specific Materials
The sensitisation of ferritic stainless steels (such as AlSI 430
and AISI 446 grades) is quite different from that of unstabilised
austenitic stainless steels. In fact, the sensitising range for ferritic
stainless steels lies above 925°C, and immunity to intergranular
corrosion is restored by annealing at 650° to 815°C for about 10,
to 60 minutes. These temperatures are essentially the opposite of
those applying to austenitic stainless ·steels. Because of their high
sensitising temperature range, the weld decay in ferritic stainless
steels occurs at the area immediately adjacent to the weld rather
than some small distance away, as in the case of austenitic
stainless steels. Unlike the case of austenitic stainless steels,
lowering the carbon content is not very effective in preventing the
weld decay in ferritic stainless steels. It is reported that AISI 430
grade with a carbon content of as low as 0.009% C was still found
susceptible to weld decay. However, as in the case of austenitic
stainless steels, the addition of Ti or Nb is found to be helpful in
preventing weld decay.
Since the diffusion rates ofC and Cr are much higher in B.C.C.
(ferrite) thart F.C.C. (austenite) lattice, the rapid cooling from
above 925°C during welding does not really suppress the
precipitation of Cr-carbide at the grain boundaries of ferritic
stainless. For the same reason, lowering the carbon content does
not effectively prevent Cr-carbide from precipitating, unless the
carbon content is extremely low, for example, 0.002% in AISI 446
grade.
Because of the excessive grain coarsening and the formation
of acicular martensite at grain boundaries, the notch toughness of
as-welded AISI 430 stainless steel is rather low. Martensite
tempering by PWHT at 800°C improves the notch toughness
significantly. Suppressing the formation of martensite by adding
0.5% Ti or 1% Nb also helps. This is prob'ably because both Ti and
Nb increase 'the stability of o-ferrite, thus suppressing the
formation of austenite. Furthermore, as mentioned previously,
they tend to form carbides at high temperatures. Therefore, the
carbon content is reduced, and tendency to form martensite is
decreased. It is reported that 0 ~ y transformation and, thus, the
formation of martensite can be suppressed by rapid quenching
during welding.
8.5.5.1. Preheat and PWHT of Ferritic Stainless Steels
A preheat of 150°C or higher can be used for ferritic stainless
steels to minimise residual stresses that contribute to weld
cracking. Thin sections need not be preheated. Thick sections (6
 554
Welding Engineering and Technology
mm or more) are preheated to 95"-205°C depending on the
thickness and the amount of restraint. Interpass temperature
should
perabe limited to the lowest practical level above the preheat
tern ture.
The corrosion resistance of low-Cr, Al or Ti stabilised, ferritic
stainless steels (Types 405 and 409) is generally not affected by
the heat of welding. Consequently, these steels are often used in
the as-welded conditions. However, those steels that are higher in
Cr and C (e.g. Types 430, 434, 442, and 446) tend to form
Cr-carbides at grain boundaries in the HAZ; this makes the weld
area susceptible to intergranular corrosion. These steels generally
require annealing after welding to re-dissolve carbides and restore
corrosion resistance.
PWHT of the ferritic stainless steels is conducted to prevent
further grain coarsening; temperatures in the 705° to 845°C range
are used. Care must be taken during heat treatment to minimise
oxidation, and during cooling through the 540° to 370°C range to
avoid embrittlement and loss of toughness. Rapid cooling through
this range is a metallurgical necessity with due consideration to
distortion and residual stresses.
Post heating reduces the hardness, but the objectionable
properties due to the coarse grain structure still remain.
8.5.6. Weldability of Martensitic Stainless Steels
The martensitic stainless steels can be welded although the
weldability is not as good as that of austenitic stainless steels.
Martensitic stainless steels can be welded by 8MAW,GTAW,
GMAW, 8AW, spot welding, flash welding, friction welding, high
frequency welding, and electron beam welding. These steels can
be welded in the annealed, semi-annealed, hardened, stress-
relieved or tempered conditions. The condition of heat treatment
has minimal effect on the hardenability of the HAZ, and thus on
the weldability.
The commercial 13% Cr steel containing more than 0.08%
carbon is the most widely used martensitic stainless steel. In
general martensitic stainless steels with higher carbon contents
are rather susceptible to underbead cracking, especially when Hz
is present during welding. Owing to the internal stresses induced
by the volume increase associated with the austenite-to-marten-
site transformation, underbead cracking can still occur even if
relatively low restraint is employed during welding. As a result,
martensitic stainless steels with a carbon content. above 0.25% to
0.30% are not normally welded.
Weldability of Specific Materials 555
To avoid underbead cracking in martensitic stainless steels,
both preheating and postweld tempering (between 600° and
800°C) are usually employed. In addition, austenitic stainless steel
filler metals are often used. This is because the weld metal so
produced remains ductile, thus reducing the chance of underbead
cracking. Also, because of the greater solubility of Hz in austenite,
Hz can dissolve in the weld metal, and the possibility of under bead
cracking is thus reduced. As in the case of welding heat-treatable
alloy steels, the weldments of martensitic stainless steels are
usually not allowed to cool directly to room temperature upon
completion of welding in order to avoid underbead cracking.
Temper Embrittlement: A problem associated with welding
of martensitic stainless steels is what is known as temper
embrittIement. When these higher carbon stainless steels are
hardened by heat treatment and rapid quenching, they must be
reheated (tempered) to a lower temperature to get back some
ductility. But tempering them in the range of 400°-565°C is not
recommended because it produces low impact properties and the
steel loses some of the corrosion resistance.
8.5.6.1. Preheating
Application of preheat and good interpass temperature control
are the best means of avoiding cracking in martensitic stainless
steels. To control cracking, particularly in martensitic 8.8. over 3
mm thick and with a carbon content above 0.10%, preheating is
almost imperative. Preheating temperature used is usually in the
range of 205° to 315°C.
The martensitic transformation temperature ranges of these
steels are sufficiently high that preheat at 315°C or below has
minimal effect on the hardness of the HAZ or weld metal.
Preheating at a temperature in or above the martensitic
transformation range is not recommended. Martensitic
transformation during cooling to room temperature or following
on immediate postweld tempering operation could contribute to
weld cracking.
The carbon content of the steel is the most important factor
in determining preheating requirements. Joint thickness, filler
metal composition, welding process used, and degree of restraint
are the other consideration. Suggested preheat temperature,
welding heat input, and postweld requirements based on carbon
content are given in table 8.13. Preheating is generally beneficial
regardless of carbon content.
iF~i~~"

Weldability of Specific Materials 557

",7'=",,,;,,,""''''''-''''O'''''''''''''''''-~_F~
556 Welding Engineering and Technology "~
-------------------------------- I
If the carbon content of the steel is 0.20% or more, the
Table 8.13. Suggested Preheat, Welding Heat Input, and weldment should be given a subcritical annealing heat treatment
Postweld Requirements for Martensitic Stainless Steels immediately upon completion of welding, as indicated in table
S. <Content
0.20
0.10
over0.10
toCool
0.50
35
Carbon0.50
0.20
280-335
220-280
min.
Heat
Normal
Normal
Hilrh
280-335
Preheat
Input
("C)
Heat
Heat treatment
Temp.
slowly; Post
heat
treatment
Welding weld
reauired.
treatment
optional
required
Requirements optional 8.13. It should be held at the suggested temperature for 25
minutes per em of weld thickness, with a minimum of one hour.
The weldment can then be air cooled.
When the filler metal composition closely matches that of the
base metal, including carbon content, the weldment can be
quench-hardened and tempered to produce uniform mechanical
properties throughout the weldment.
Whether interruption of preheat is acceptable also depends
upon carbon content. When the carbon content is more than 0.20%; Case Study. Suggest a PWHT for creep-resistant 12% Cr-steel
it is best to maintain the interpass temperature during welding having a composition of 12% Cr, 0.2% C, 1% Mo, 0.4% W. and 0.3
and to heat treat the weldment before it is allowed to cool be1.ow %v:
95°C. Solution: Upon the completion of welding, the weldment
should be held at 150aC to avoid cracking. After holding it at 150aC
8.5.6.2. Postweld Heat Treatment of MSS Weldments
for about one hour, the weldment should then be heat treated at
The aims of PWHT are : 750aC for 4 hours to temper the martensite which had formed, as
(i) to temper or anneal the weld metal and HAZ to decrease
hardness and improve toughness or strength; and
(ii) to decrease residual stresses associated with welding.
Postweld heat treatments normally used for MSS are
Welding
\ QJ
,
Welding

L.
::>
\
QJ
\ +-
subcritical annealing and full annealing. L.
~o
o
•...
\ 750·C
The necessity for a PWHT depends upon the composition of L. a.
QJ

\ I \
a. \ \ E
the steel, the filler metal, and the service requirements. Full QJ

E \ \ ~
annealing transforms a multi-phase weld zone to a wholly ferritic ~
QJ

\ I \ \ I \
structure. This annealing procedure requires proper control of the
complete thermal cycle. It should not be used unless maximum Ms = 260·C
\
r--+-
1 \-
\\
Ms=260·C
------T
\.1.Qill \

softness is required. Typical postweld annealing temperatures are 'L--I

150·C _
given in table 8.14. Time Time
Table 8.14. Annealing Treatments Suggested I b) Undesiroble PWHT cycle
for MSS Weldments (01 Desiroble PWHT cycle
S. 420
414
431
CA-6NM
403,410,416
440A.
CA-15, 440B,
CA-40 ("C)
440C Not 830-885
845-900
675-760
650-735
620-705
650-760
790-815
595~20
620~50
recommended
("C) Fig. 8.18. Schematic illustration of a PWHT thermal cycle for a
Stainless Steel
(AI~l) Type Full Annealing
Subcritical
temperature annealing
range*
temperature range** creep-resistant 12% Cr martensitic stainless steel.
shown schematically in Fig. 8.18(b). The resultant microstructure,
of the HAZ, will be essentially tempered martensite. From the
view point of completely avoiding the possibility of underbead
cracking, it might seem even better to hold the weldment at a
temperature higher than Ms (say 300aC) so that no martensite
would form after welding as shown schematically in Fig. 8.18(a).
During the subsequent heat-treating at 750aC, however, the
untransformed austenitic phase decomposes, and the micro-
structure of the HAZ becomes ferritic with agglomerated carbides
• Aircool from temperature; lowest hardness is obtained by heating near the
top of the range .
precipitated at the grain boundaries. Unfortunately, such a
•• Furnace cool to 595°C; weldment can then be air cooled.
microstructure is rather brittle and in fact inferior to that of
 558 559
Welding Engineering and Technology Weldability of Specific Materials
~C t- 0 0 IId
001O0
001O0
...• ....,~ I C'IC'I
...0...0
1O1O
00 0 00
° 00
00000
000
~8°
I•-...•
8800
I0000
<.>
0t-
00
000
-4-t0'••
00000
•...• •rn
C'I
I,-00000
OOC'llDlD
t-
MMOMM 00
~~
0000
""'000'"
~lO"",lD""
~ad.oc.CLO
OC'l It-
t-lO""'lOlD.-l
I~....-f
,~O
.~
,t-C'I
0
0
00
M
MOO '2
.••.•••
"'"
00· t-M
••••M
.....
o
...•
••••lDlO
00
o:s .~ 0000
•'...•
'".;;"O00
C'I
1O
"'"
00
0
000
O.M
1
••••••••
lOlDlD
lD
lOOO
C'IjM
•....•
00 C'I
'"
~000
«'::l~t-....•"'o::!:
1O1O
0,C'IC'I •C'I00
,00M
1O t-01D
00
00t-t-oo..,r..,r
'" C'lC'I~~
0 0
... Ic4 ~'i:JI'8
::s~~ t-1O
""'''''''''<.0
gi::
...•...• Mu
.s 1O
.~ ~...•..•..~ 0000 1O01O ~~ex;>01O
...•
rn .~
<.> rnrn
t-'.0
O""";C":)
.....cicr;
"'"I::l ...-4
rn
...;
b.O rn""1
a:i
0 ••.•••••.•••

]
"'"
00 00 •.•t-15 ::So CJ)
MMC'llD
'"
0
••••C'I
00 •..• 1O
1O ' <.> 00
E .~ ~
~~dE=: .s
rn
:gI~ ~.t-:e.s~ IC'I
c:::~
ex;>
~~.8 ~~~ai~ cd::s::5
~....-t~lCl.Q .-4~C\!
1""'"1

0000.s..c: ~r;<Z
~; IS r--: >-
~~~
.~ ~~
eY'JC":)

tempered martensite. Therefore, only the first heat treatment~ >IJ*

~~ is
~~
•..••..•
rnrnrnrn
t-lt)~
o:s
rn
~~.8"i3
Co>
r..)
.s S
~
Orn
tl,
....-t,.....~~
o:s ..c:
c3
c::: Ql .•.••
'I"""""4,.-..f ~.,......
Ql
•.• I::l
"1' '9c4
<.>
"'"
T"""I

rn
rn·t\l
~~IS~
••••••

.~
c3
:;;:
g ~ C'l1D01D
Ql
'1'
C':iet:i?""4
""''''''~
'::l-OO
•••
::s:::s
b.O
<~~ ....•
c-i00
t.!.~~~~ ...0

recommended. .~
's~-
~ c.j 0000
"'" "'" M '" :£
~ I
8.5.7. Precipitation-Hardening (PH) Stainless Steels
The fourth category of stainless steels - the precipitation or
I~ co

age-hardening stainless steels - offers a combination of

properties not attainable in other three types of stainless steels.
For example, there are stronger steels; there are steels that have
greater corrosion resistance; and there are steels that are easier
-{IJ

CIl
Ci)

§e
.s
to fabricate compared with other stainless steels viz., ASS, FSS, 00
.s.
and MSS. Quite a few of these steels are proprietary material. {IJ
{IJ
~
rn
The PH steels are used in applications requiring high strength b.O

(850 to 2100 MPa, depending on heat treatment), good fracture ~

....
.8
...
00 1l
toughness, resistance to corrosion comparable to AlSI 302 and
AlSI 304 stainless steels, and resistance to oxidation at elevated re ~
tern pera tures. Z
ca 1l
(,) I;:l
PH stainless steels are grouped into three types depending 'on '8
the structure and the behaviour of the steel when it is cooled from $ ::>
::s
the appropriate solutioning (austenitizing) temperature. These CIl

eQ E-t

types are: martensitic, semi-austenitic, and austenitic.

00
The martensitic types are so named because their compositions '3 ~
rn
are balanced to provide a martensitic structure after cooling from
il
{IJ Ql
=
the solutioning temperature. Additional strength is obtained by Q
.... ...
ageing . ....
.•.. .Ql
{IJ
Q .l:l'"8
Semi-austenitic types remain austenitic when initially cooled :a '::l

from the appropriate annealing temperature. Then, the steel is ~ ~ ...0~

reheated to condition the structure so that it transforms to
martensite on cooling to room temperature. Subsequent ageing
-
oQ

e:l
'"

E ~
'" I::l

produces additional strength. ~1l

.~
Q
.s
.6), 'Ul
61,

Austenitic PH stainless steels remain austenitic after cooling Z ~ ~

from the appropriate solutioning temperature. Strengthening is . to
.... ~ ~
obtained by ageing the austenitic structure. The nominal 00 ",1l
compositions of some typical PH stainless steels are listed in table .! I::l Ql

~ol..l::l~
8.15 while the characteristics and specific applications appear in ,.Q
table 8.16. ~ 61,$
.~ B
'" '"
8.5.7.1. High Temperature Properties of PH Stainless Ql '"
Steels ~ ~
il"i3
~o 'ca
All PH stainless steels tend to become embrittled after Ql
.•.•
'"

exposure for thousands of hours at temperatures above about IS

o 0~

285°C. In general, long time exposure in the 370° to 425°C range rn::s
td *
results in sharply increased tensile and yield strengths and
reduced fracture toughness. These steels are generally limited to
 561
560 Welding Engineering and Technology Weldability of Specific Materials

,,;
"0
d a maximum service temperature of about 315°C for long time
d '" zi
'S, rn dQ) exposure.
c:
~
..0
S For short-time service, many of the martensitic and semi-
2
'<;1

-V
..
lie ....
'" austenitic PH stainless steels have suitable mechanical prQperties
.$ d
c:.>
.9 for service temperature up to 480°C. For the austenitic types, good
eai
:a..
::s
~ corrosion resistance can be maintained up to about 815°C. The
¢::
....
..
.61J
::s choice of a steel for a particular maximum service temperature
'"

1
.•..
00
I'll

~
..c:

..S:l
'"
..
11
~"
QJ
d
'<;1
rn

~-
depends upon the service environment
requirements .
and the stress
llIl .g "& .! -5 8.5.7.2. Welding of PH Stainless Steels
ItJ
GI 1:1 eCo ..
rD
~
.::/
~ The weldable types of PH stainless steels can be joined by any
:5 ];
Cll
",-
QJ .!t d of the standard welding processes used to join the conventional
...•
00
CIS
-<:: .9
-a
"2
Cll
8 austenitic stainless steels. GTAW, GMAW, and PAW are well
'" ..c:

re "0
d
~- suited. SMAWcan be used but high joint strengths might not be
.... ~- '" ....
'"
obtainable with some of the PH steels because matching electrodes
Q ~ '"
Q)
S
~ are not available. Suggested filler metals are listed in table 8.17.
~ fl-
<tI
0'1
gj These steels can also be joined by resistance welding processes.
Q
...•
...•
c:
QJ

~ ..
CIS
.s Q)
Co
Table 8.17. Suggested Filler Metals for
Q d ",-
Jg
...• '61J .
Q)
S Welding PH Stainless Steels
Q. d rn '"
(Source: AWS Handbook, 7th Edition)
~
l" Q) Q)
, "0
.... '" 't: Dissimilar
~ Q) -
<: Designation UNS Covered Bare
Q ...,..0
No. electrodes welding PH stainless
lQ "'"0 steels
wire
.~ .~._ d
0· ~
_
ct).- -
gs
Q)
]..§:~
~ ..
c:.>
~ ..
0 _gj d
o
QJ

00
Q.,
.~.-
~~ ~~ <tI b.O"O ~ ~.-
.;"E ~
c:.> •

Martensitic Types
AMS 5826 E or ER309,
•• Co c: .••• ::s d '" ..0 "0 17-4 PH S 17400 AMS 5827B
.8 8 oS E or ER309Cb
"C

~ ~f
~ 'C:; .5
fIJ
rn

Ql..o"O~
.~(,J
.-
.~",.,
(,J
£t and
S 15500
(17-4 PH> or
E 308
(17-4 PH> or
ER 308
a Co •• 15-5 PH
8. :tl ~
':
rIl
~
....
-"0
::s c: 0
:g Q) 1a'·.;1 ::;
AMS 5805C E or ERNiMo-3,
...•
~"'b.O
dOc:.>
Sd"EQJ
S ~.~ Stainless W S 17600 E 308 or
(A-286) or E or ER 309
(IJ
.... ::S~"'b '" .8 .S ENiMo-3
50< ~~.s 's 9..<? ~ S~~ ERNiMo-3
~ ~~~ o::is § rn
~ .••• QJ..o ~ S a
Cll
., cUba,)
..0' •• ~ co ~ ~ - Q-- cU Semi austenitic Types
50<
.•.• c: Co Co> .5 '=, ~
tl0 AMS 5824A E or ER 310,
.~ 6'o.s:
Q) ••••
17-7 PH S 17700 AMS 5827B
-=
t;.;l
Cll
.,
';::
.9
.s _;; .=:
Cl) .•.• ~ 'S, 8'
~ S'O •..• ":a
o d
Co
•• (17-4 PH) (17-7 PH> ENiCrFe-2, or
~ ••
::s
S~f:= ERNiCr-3
~ <.>
o 0 '" 0- o ::s Co..o :: 8 ~ E308, or E309
.... ~ 'C..J::
QJ
.S al .S E308 or AMS 5812C E orER309,
:: e S 15700
-
Io.l
C:l QJ •••••••• g" ~ - tU PH 15-7 Mo E orER310
!Xi Co°O (,J QJ
"'0

1il ._
8. ~ E309 (PH 15-7 Mo)
GI c:; ::s..c c: d c.. .-=: ::l d..o
c:.>

,Q '"
f'/J'- ~
c: rn
.5S1l]0
.~ .s ~
rn
0'1
d ~
'" Co AM350 S 35000 AMS 5775A AMS 5774B E or ER308,
(AM350) E or ER309
~ '-.~

C1,) ~
"'0

•
rn
.,.
rn I C1,) c.>
.-~ .S g (AM350)
(,J QJ
f:io§al (,J
d .- 0'1 AMS 5780A E or ER308,
·fJ ~ § d..o '" ~c:.> AM355 S 35500 AMS 5781A
~

.~ rns :a
Cl:S .~

.-
o
~ ~.- d
~ s::: c:: .~
....
lS "2 .~
rn c:.> ::s (AM355) (AM355) E orER309

.~ d
QJ
r.....- 0 . •••• - Co) '"
·aill 'r;; tl.8'
.•.. :;
v uOJ'';: ~ ~ ~ :: Austenitic Type
§ • _
1"'"
c,) ~ .~
;::
Q.J .••.••

1il E309 or ERNiCrFe-6 or E or ER309,

rn ... ~ ~~ =' g.~.•. 0r:
._ A-286 K 66286 E or ER310
~ ~ :s:
'111
'w = '" rn·.;I E310 ERNiMo-3
~ o ~ c:.>'" ill .2
~~ 6"0
UJ ce ~
cu
o Q)._
•••"0 "0
1; ::: .9 "E ~ 1il
0.",0., ·S ••• '" 0
~ o~ro~
Urn..o..8,
::
< J5~g~
Cl,)

00 8 ~8

~--------------".--
562 Welding Engineering and Technology
The PH stainless steels do not require preheating for welding
them. The martensitic and semi-austenitic types are not subjected
to cracking. However, the austenitic types suffer from HAZ hot
eracking that makes them very difficult to weld. The PH stainless
steels are less ductile and more' notch sensitive than the
conventional austenitic stainless steels. Therefore, sharp stress
concentrations should be avoided in the weldment design and at
welded joints.
8.5.7.3. Weldability of PH Martensitic Stainless Steels
In these steels, martensitic structure (untempered) forms
upon cooling from solution-treating temperature of 815° to 1035°C,
depending upon composition. This form of PH stainless steel is
suitable for machining and other fabrication operations.
Subsequent ageing between 480° and 620°C increases hardness
and strength. Finish-machining operations are seldom required
after heat treatment.
Weldability of the PH martensitic stainless steels is generally
excellent. They are not crack sensitive (carbon content is low) or
susceptible to hot cracking associated with other Cu-bearing
stainless steels. Hot cracking due to carbon may occur, however,
if the martensitic steels are welded to carbon or low-alloy steels.
In such cases, welding methods that minimise weld metal dilution
should be used.
Preheating is not nej::essary to prevent cracks and excessive
hardness in the martensitic PH stainless steels. Heavy sections
or members of highly restrained j'oints are, however, often heat
treated to a tough overly aged condition before welding.
The superior notch toughness of grades such as 15-5 PH and
PH 13-8Mo' make these steels suitable for restrained joint
weldments in any condition. However, regardless of the toughness
level of the steel, notch effects should be eliminated. As with all
high strength steels, residual stresses and unfused notches should
be avoided.
8.5.7.4.Weldability of PH Semi~austenitic Stainless
Steels
Composition of the semiaustenitic grades are balanced so that
the structure is austenitic in the solution treated condition. This
ductile structure permits forming by the conventional processes
used for the 18-8 stainless steels. However, these steels work
harden rapidly and intermediate annealing may be needed for
complex forming or deep drawing. Following fabrication, the
austenitic structure is transformed to martensite by a thermal
Weldability of Specific Materials 563
treatment. Maximum strength is then developed by an ageing
treatment, resulting in precipitation hardening and tempering of
the martensite.
Weldability of the PH semiaustenitic stainless steels is good;
preheating and PWHT are not required. These steels are readily
joined by most arc and resistance welding methods. However,
GTAWshould be used for fusion welding to overcome hot shortness
in single pass welds and to minimise loss of aluminium in the weld
metal.
The heat of welding austenitises and solution-treats (anneals)
the area adjacent to weld. Thus, regardless of base metal
condition. the structure of both the HAZ and the weld metal is
austenitic in the as-welded condition. For this reason, the
semiaustenitic PH steels can be welded in any condition without
reheating and without requiring control of interpass temperature
or cooling rate. Strength of fusion welds, made with welding rods
of similar compositions as the base metal and using recommended
procedures can be expected to be 90 to 100% of the base metal
strength.
8.5.7.5. Weldability of PH Austenitic Stainless Steels
These grades of PH stainless steels retain their austenitic
structure at room temperature following solution heat treatment
and are hardened by a precipitation treatment.
The austenitic grades are the most difficult of the three types
of PH stainless steels to weld. The most successful welds have
been made by GTAW,keeping heat input low. The metal should
be in the solution-treated condition, and multipass welding with
light stringer beads is recommended. When SMAW is employed,
Ni-base filler metals are normally used for welding these steels,
although conventional austenitic stainless steel filler metals can
also be used.
8.5.7.6. Welding of 4-6% Cr-Steels
Steels containing 4 to 6% Cr are not truly stainless though
these are sometimes classified as such. They do have a resistance
to corrosion of 4 to 10 times and resistance to oxidation from 3to
8 times that of mild steels. They have considerable resistance to
oxidation and sulphur corrosion at temperatures up to 595° to
650°C. They are therefore used extensively for hot oil transfer
lines and other oil refinery equipment operating at elevated
temperatures, furnace tubes, heat exchangers, valves and high
temperature steam lines.
A noticeable characteristic of the 4,--6%Cr steels is their
intense air hardening which is proportional to the Cr and C
564 Welding Engineering and Technology
contents. This affects the weldability as both the weld and the
HAZ are apt to be very hard unless preheating and subsequent
annealing or stress-relieving are used.
The most commonly used additional alloy in these steels is Mo
CAISI 502), usually in an amount approximately 0.5%. This
addition of Mo does not appreciably affect the physical qualities
of these steels at room temperatures, but when such steels are
given different heat treatments then the presence of Mo does
in.crease, to a considerable extent, the strength of these steels at
elevated temperatures and is also reported to increase the
resistance to certain types of corrosion. Mo also increases the creep
strength ..
When comparable properties in weld and the base metal are
needed to meet service requirements, the general rule is to use an
AWS E502-15 or E502-16 electrode that will deposit 4-6% Cr of
the approximate analysis as the base metal and to preheat and
then anneal or stress relieve. The use of AWS E309 or E310
electrode is recommended where preheating or PWHT cannot be
used. In this case the HAZ will be hard and brittle but the joint
will withstand some impact or bending stress as the weld metal
deposited by these electrodes will be relatively soft and ductile.
The hard and brittle weld and fusion zone obtained when using
the 4-6% Cr electrode without preheating or annealing will not
stand welding stresses as well as the combination of the hard
fusion zone but softer and clore ductile weld metal as obtained
with E309 or E310 electrodes can.
When using either the 4-6% Cr or stainless steel electrodes,
direct current with electrode positive (deep) is used. The arc should
be maintained as short as possible while employing standard
stainless steel welding procedure. E502 electrodes contain Mo so
that they can be used on all types of 4-6% Cr-steels.
8.6. WELDABILITY OF HIGH ALLOY STEELS
High alloy steels, considered here, are those that contain more
than 5% total alloy additions excluding, however, the stainless
steels, Cr-Mo steels, 9% Ni-steels and the tool steels. Four general
classes of high alloy steels discussed in this section are:
1. Ni-Co Alloy steels,
2. 5% Cr-Mo-Vsteels,
3. Maraging steels, and
4. Austenitic Mn-steels or Hadfield steels.
The first two classes of these steels are also known as Ultra
High Strength Steels which are designed for structural
Weldability of Specific Materials 565
applications having at least 1250 MPa yield strength and 1400
MPa ultirnate tensile strength, combined with good fracture
toughness. They are generally used for critical applications
requiring high strength-to-weight ratios as well as excellent
reliability and consistent response to heat treatment.
Welded joints in ultra high strength steels generally do not
require heat treatment except for a stress-relief in certain
applications. \Vith specific alloys, the wcldmeht may be quenched
and tempered to obtain desired properties.
8.6.1. Weldability of Nickel~Cobalt Alloy Steels
Many of the ultra-high strength Ni-Co steels contain 8--10%
Ni, 4-·14%Co, and small additions of Cr and Mo, as shown in table
8.18. The Ni provides good toughness, improves strength by
solid-solution, and promotes hardenability. Co is added to
maintain a high austenite-to-martensite transformation
temperature (Ms) and thus limit.the amount of retained austenite
in the steel after quenching. It also contributes to solid solution
strengthening.
Cr and Mo, two carbide forming elements, are added to
promote harden ability and strength, but they do contribute to
embrittlement. Additions of these elements are kept low to
maintain good toughness and produce a fairly uniform response
to tempering without pronounced secondary hardening. Si and
other residual elements are kept low for optimum toughness.
The Ms temperatures of these steels are sufficiently high.to
permit self tempering of the martensite as the steel cools through.
the transformation range to room temperature. With relatively
low carbon content, a strong tough martensite is formed
as-quenched. This behaviour contributes to good weldability. As
the carbon content is increased, weldability and toughness of the
steel decrease.
In general Ni-Co alloy steels are designed for structures where
the metal will be loaded to near its yield strength to meet
requirements of structural efficiency and economy. Such
applications include helicopter components, armour, high
performance ships and shallow draft vessels, rocket and missile
cases, airframe and undercarriage components, aerospace and
hydrospace structures, pressure vessels, and other heavy section
structures.
HP9-4-20 steel has been used for pressure vessels, structural
components in aircraft, rotor shafts for metal forming equipment,
and high strength, shock absorbing automotive parts. HP9-4-30
steel is a forging grade alloy which is used for a variety of aircraft
forgings such as engine mounts and landing gear.
 1111

566 Welding Engineering and Technology Weldability of Specific Materials 567

N
N
E=: ~ .0--
N-_
ci
cci
II::r:
.....•
<
.....•
M
.....•
.10 0r...: 0_.
09,.9,.0-------
~ c.1J, 0;
t-'" 00
::>
.....•
,....;--.-------
f-------.----.---
0,....; .~---- 00
IiIi N --,--_.~
..__ N
.__-.....
<::> -,. __
AF1410 steel combines ultra-high strength with superior
8 -------
.....•

0II
<::>

0~ci 0 i.O
<::>
..r -.;<
.0
010 -.,j'
....•

_--- 0d, - N 10 t-
'"
N . I ..--ct
00
.....•
< <::>
.....•
~
Id,
0...
.....•
00
J,
c..;
.....• ::>

..J,
..;.
i,....;
.O
....•

----
<::>
.....•
N 00
.... i-i5
... ~
C,) 0~
Q.,
u:J
i.O
10
ci
~
::r:
0...
..;.
~
-.;<i.O
..r
i.O.....•
10 I 00.....•
.-.
i.O
i.O 53 ~
-.;<
I
ci
.....•
c3
~ 0 Z 0 ,....;
fracture properties and stress corro~ionresistance. The low carbon
<::>
<::> ~ <::>
<::>

content of this steel enables the material to be fabricated easily.

This steel has been designed to compete with titanium alloys at
lower cost and superior strength-to-weight ratio. AF1410 steel I

may be used for aerospace applications, armour plate, gears, dies,

sZ-0 -u0uo$0. 8c"..gr:'
I
.;;; I

~ m
~
=....
.•..
...•
.•..m ~ fasteners, and bearings.
~
'.cI
=
00
'""bQ
.•..
~00
...•
m
cxibQ
IS
...;
...•
~ =
C) ~
.cIC)
c~-
ca
~S- xi

....
8.6.1.1. Welding and Weldability of Ni-Co Steels
Suggested guidelines for designing welded ultra-high strength
steel structures are as follows.
(i) Avoid fillet welded joints, such as T-joints, by the use of
forging and butt joints.
(ii) Place welded joints in low stress locations with good
accessibility for welding and inspection .
(iii) Weld with automatic or mechanised equipment to en-
sure consistent welding conditions.
(iv) Use weld groove designs that minimise the amount of
filler metal required to fill them .
(v) Use premium quality, clean filler metal.
(vi) Provide good supervision for the welding operation.
For acceptable welds in high-strength alloy steels, the
mechanical properties of the weld metal must be equal to those of
the base metal. The combination of strength and toughness gan
only be achieved with weld metal containing very low impurity
levels. Exceptional cleanliness of the filler metal, the joint area,
and the weld metal itself are essential. An essential monitor of
weld metal cleanness is the total O2 content. For HY-180 steel
weldments, for example, O2 content of less than 50 ppm is
necessary.
Welding heat input must be controlled to a low,consistent level
to take advantage of grain refinement in prior weld beads and to
minimise structural degradation of the HAZ.
Multipass welding of Ni-Co steels presents some unusual
metallurgical features. Notably, the reheating cycle of succeeding
weld passes gives rise to cyclic phase transformations. Originally
formed, self-tempered martensite in reheated zones may revert to
austenite. That is why welded joints filled with small weld beads
-'"
'"
....•
have better mechanical properties than joints made with large
Uj. beads. Low energy welding processes provide interpass grain
'"1j refinement. Thus GTAW with cold wire feed is generally
recommended for welding these steels in preference to GMAWor
u:i;;;;: GTAWwith hot wire feed.
 569
558
------_.,----------------- Welding Engineering and Techno!og)' Weldabilit.y of Specific Materials
o o
<.C oo
<.0
~
Ni-Co high strength alloy steels are often welded in the o ...;

breat-treated condition. Heat treatment after welding is not

:::.
6 .,J. cb

normally required, except for stress relief in special situations

because low heat input welding with associated rapid cooling
results in minimum deterioration of propert.ies in the HAZ.
Successful fabrication of these steels is dependent upon the
use of low-H2 welding techniques and matching filler metals with
-fI)
~
~
low impurity levels. Shielding gases should be essentially free of
-!oO

00 oon on
~ oon
3)
P2. N2, and H2• Dry, pm'e argon, helium or mixtures of both can fI)
Q
r.. '1
on
'1
L.'j
'?
LC~

be used. Filler rod or wire must be absolutely clean; no surface

<V
....• r:- t- t-
r:.1 "" "" ""
...•
debris or lubricant can be permitted. Likewise, all joints must be =
..•..
thoroughly cleaned of all traces of cutting flui.d, lubricants, oils 00
and gases, oxide films, scale, and other contaminants. ~~ I ~
ci
-Q
With premium-quality high-strength alloys, only the GTAW
and PAW processes can provide clean' weld metal with suitable
mechanical properties. GMAW is not recommended because O2
levels of weld metals deposited by this process seldoIr.\fall below
=
r.'\l
$.(
.•..
S
/s'<

~
~
.
r£
--t-~
.~ 1.-.
"'1 0~
.....
o
"'l
....•

ocoI
.~
o.
150 ppm. ~ <..::i

EBW can be used where production conditions permit. A 8

S' --=--1 ~ ~
PWHT may be required with some steels because of the absence
::s. oro oM o~
U
~ (Ij oo o
o o
ci
of tempering with single pass welds. •.•...

Weldability of HP9-4-20 Steel o

fIl
= 0
The weldability of HP9-4-20 alloy steel is considered excellent o
.... '"
<=>

~ ~
<=>

~
provided the welding process and procedures prevent weld metal .•..
...•
••• ~ ~ C<

contamination and the heat input is not excessive. Weld metal

having properties equivalent to heat-treated base metal can be
tI:l
o __-=
'---1 .: <=>

S- ,55 ;jl g
obtained with GTAW process using the conditions listed in table O
8.19. Machine welding with cold wire feed is recommended. U ~ I ~ ~ ~
Preheat or PWHT is not required. Cil
c;l
d d 0
...•

--- Table 8.19. Typical.Conditions for GTAWHP9-4-20

and AF 1410 Alloy Steel Plates
S
~
..c:'1
+C-l~ C>1
~
C'I
~
S_ No. Parameter Q
i. Filler metal
- I
HP9·4-20
HP9-4-2H AF14i-o--
AF1410----
Q
~ ~ ~J ~ ~
2.
8.
Filler wire dia., mm
Filler wire feed, mm/min 500-750
1.6 I 500·-550
1.6 -
cO
(l.)

,.Q
.9
1:\
~
.~
~
....•
...•
15
'0::

.;.
5.
Shielding gas
Welding current, A (dc)
Argon
300-350
75% He+25% Ax
160-200
~
-~
Q
'"
:I:
~
~ :I:
~ I

6. Axc voltage, V 10-12 13 l:I)

'7. Travel speed, mm/min 125-250 100

8.
9.
Heat input, KJ/mm
I Max. interpass. temp., °C
0.8-1.0
95
1.32-1.36
'/0
~
cri
---;;---~
...; O'i C"j I
 570
Welding Engineering and Technology
Weldments may be stress relieved at 540°C without signi-
flcantly changing the mechanical properties of the weld metal.
1 Weldability
Toughness
while and yield
the tensile strength
strength of the weld metal tend to increase
decreases.
Weldability of AF1410 Steel
As with other Ni-Co alloy steels, AF1410 steel is best joined
by GTAW or PAW processes using low heat input and multiple
passes. Matching filler metal is recommended. Typical conditions
for mechanised GTAWare listed in Table 8.19. Ageing the weld
metal at 480°C for 2 hours appears to improve yield strength and
toughness.
8.6.2. Weldability of Chromium-Molybdenum_Vanadium
.Steels
The 5% Cr-Mo-V(5CrMoV) ultra-high strength steels are tool
steels that are suitable for structural applications. They are types
Hll, Hll Modified,
given in table 8.20. and H13 alloy steels; their compositions are
All three steels are secondary hardening and develop optimum
properties when tempered above 510°C. These steels are air
hardening, toand therefore require care during welding and heat
.treatment avoid cracking.
Typical applications of Hll modified steel are aircraft landing
gear and frame components, parts for gas and steam turbines,
high strength fasteners, springs, and hot working dies. HI3 steel
is used for dies, punches, shafts, beams, tension bars, and
ratchets.
Preheat and PWHT.
The 5CrMoV steels being air-hardening they form hard
martensite when quenched from the hardening temperature of
lOOaoC. Martensite will form in the weld metal and HAZ unless
appropriate preheat and PWHT procedures are followed to avoid
it. As with all hardenable alloy steels, a martensitic HAZ is very
susceptible to underbead cracking, particularly when H2 is present
during welding.
To avoid weld cracking, 5CrMoV steel should be preheated for
welding. The ideal preheat tempera~~re is one that is above t~e
Ms t.emperature at which .martensite begins to form on cooling.
This temperature is about 2900G for Hll steel and about 3450C
for HI3 steel. When the weld is held at or above this temperature,
it remains austenitic, but it could transform to bainite after a
period. If the weld is cooled rapidly prior to bainitic trans-
of Specific Materials
formation, it transforms to hard martensite and may possibly
crack.
S71
High preheat and interpass temper~tures can cause problems
during welding from oxidation of the joint faces. The oxide may
lead to occasional porosity or lack of fusion in the weld. Preheating
to a lower temperature will avoid this problem, but some austenite
in the weld will transform to martensite. If the weld is cooled
rapidly to room temperature, the remaining austenite will
transform to martensite. The volumetric change accompanying
this transformation will produce local high stresses, and possibly
result in microcracking in the weld and heat affected zones.
The problem can be avoided by raising the temperature of the
weld from the preheat level to about 55°C above the Ms
temperature and holding it at that level for about one hour. This
procedure will transform the remaining austenite to a reasonably
ductile bainitic structure. Cracking would not be expected after
the weld is cooled to room temperature. Further heat treatment
is required to produce a desirable microstructure in the welded
joint.
An alternative is to cool the weld from the preheat
temperature to about 65°C to transform the retained austenite to
martensite. The weld must then be immediately heated to an
appropriate higher temperature to relieve welding stresses and
temper the martensite. The welded joint should not be stress
relieved without the cooling step. Any weld repairs should be made
prior to final heat treatment of weldment.
A preheat temperature of about 230°C can be used without
excessive oxidation or operator discomfort. When welding thin
sections by GTAWa preheat of 150°C may be suitable.
Welding Conditions
The 5CrMoV steels are normally welded in the annealed
condition. When the weldment is to be used for a high strength,
critical application requiring good notch toughness, the cleanness
and welding requirements described in previous section for Ni-Co
steels must be applied with these steels as well. GTAW is the
recommended process for welding.
When non-consumable weld backing is required to provide
desired root reinforcement, the backing must not quench-harden
the weld metal. Quench hardening may cause cracks in the root
of the weld. Metal backing should have relatively low thermal
conductivity and it should be heated to the specified preheat
temperature before welding commences. Copper backing is not
 -
~-------_
512 _._-_._-_._------_._._--_._----_._----
..
Welding Engineering and Technology Weidabiiity of Specific Materials 573

['-?commended. A refractory backing material may be used, The 18 Ni (350) grade may have different Mo and Ti contents,
provided it is not a source of moisture or other contamination. depending on the supplier. Maraging steel A8'l'M A590 also
The finer metal may have the same composition as the base designated as 12-5-3 maraging steel, is designed for nuclear
metal or a lower alloy content, particularly the carbon level. Low applications, but its strength is limited because of the absence of
aJloy steel filler metal will produce weld metal with better ductility cobalt.
and toughness at some sacrifice in strength. Weld metal. Maraging steels are characterised by high strength and
(:omposition
welding 'will depend on dilution as affected byjoint design and
procedure. excellent toughness. The strength is obtained as a result of age
hardening of low carbon martensite.
The face and root surfa.ces of the welded joint should be 8.6.3.1. Physical Properties
machined flush with the base meta! to remove stress Concen-
trations. It may be necessary to provide additional thickness when The physical properties of maraging steels of concern in
designing the weld joint to allow fOT this machining. welding are compared to those of mild steel and austenitic
stainless steel as shown in table 8.22.
8.6.3. Weldability of M~raging Steels
Table 8.22. Comparative Physical Properties of Maraging
Maraging steels are a family of Fe-Ni alloys that are Steel, Mild Steel, and Austenitic Stainless Steel
strengthened by precipitation of one or more intermetallic
.compounds in a matrix of essentially carbon-free martensite. In S. Property Units Maraging I Mild steel Austenitic
No. steel stainless
addi tion to the Ni, these steels generally contain either Co or Cr, steel
Mo, Ti, and AI. The nominal composition of six eommercial
maraging steels are given in table 8.21. 1. Density
2. Caeft'. of tho expo 1!m1(m.kl 10.1 12.8 17.8
Table 8.21. Nominal Compositions of l24°-284°C) (20°-.;00ac) (OO-315"C)
Commercial Maragirig Steels 3. I Thermal conductivity g/cmj:
W/(m.k) 8.0
19.7 7.85
52 8.0
15

-;.NOr St"l ''''''''''.

1. I M>TM A5;---/-
r~~~
- Nominal Composition, wt. %_
Al 4.
5. Electrical resistivity
Melting(20°C)
point Iln.
°Cm~ 0.36 to 0.71
1430-1450 0.17
1520 I 1400-1450
0.72

ASTM
CastGr18Ni
18 Ni A590
A (200)
(350) 8 4
There is not much difference between the properties of
I 18 0.2 0.1 maraging steels and mild steels. However, the coefficient of
4.
2. 3. Gr B
C (250)
(300) 8 5 0.4 0.1 thermal expansion of-maraging steel is much different from that
I 18 9 5 of austenitic stainless steel and that may cause some problems
0.7 0.1
when welding them together.
I 18 12 4 1.3 0.1 Maraging steels being magnetic, care must be taken to avoid
5 arc blow during welding.

•Carbon = 0.03% max.

----'~
I 17 I -
10 5 3
--- 0.3
0.3~0.1 --
0.4
Maraging steels are used for aircraft and aerospace compo-
nents where both high strength and weldability are important
considerations. They are also used for making precision tools
because of th~ low temperature change during ageing. The tools
For optimum properties, the carbon and impurity elements in can be machined to size while the steel is in annealed condition
maraging steels are purposely kept very low. The maximum and then aged to increase strength and hardness.
anlQunt for each element is specified as follows:
Carbon = 0.03% 8.6.3.2. Welding of Maraging Steels
Mn, 8i = 0.10% Maraging steels are most often welded by GTAW process
P & S 7' 0.010% . because it allows good control of heat input and protection of the
weld from oxidation.
j'
i
514 Welding Engineering and Technology Weldability of Specific Materials 575

GMAW process can be used provided the impurity levels of of the base metal and, in some cases, below that achieved with
shielding gas and contamination of filler wire and base material inert gas shielded welding.
can be kept low. Carbon and 0z levels should be kept low; because 8.6.3.3. Weldability of Arc Welded Maraging Steels
c3rbon pick up could result in embrittlement due to the formation
of Ti-carbides at the grain boundaries. 0z pick up may result in The low-carbon martensite formed,in the HAZ on cooling has
the formation of Ti-oxide and a reduction in fracture toughness. a low susceptibility to Hz-induced cracking. This is partly because
the region close to the weld interface is relatively soft in the
SMAW process for welding maraging steels is not widely used as-welded condition. Maraging steels are less sensitive to
because of lower strength and toughness of weld metal. The Hz-embrittlement than are heat-treated low alloy steels at the
electrode coatings are generally basic lime cryolite or lime titania same strength levels.
with low levels of silica.
Susceptibility to cold cracking is also attributed to the residual
Submerged arc welding can also be employed for welding these stress pattern in the weldment. Because the martensitic
steels. The filler wire used usually contains higher Ti content to transformation in these steels occurs at relatively low temperature
deoxidise the weld metal. Basic fluxes without silica appear to be thus the weld metal is under compressive longitudinal stress
----1-1 ---" --.L
suitable. rather than a tensile stress, as shown in Fig. 8.20. The stress
.•....
2--11-- ll-4
"0 - c
•...
::l
VI
.><:
~
Q.I

'iij .§ 50
Q.I

.9 ~
E 30
VI
70
.S:! 0+-
'iij
~-0
~
280
10
10
~SOc30
0
MPa 210
50
140
70
0420420
490
350
210
140
- 70
350 280
60· 490
I
~~g.~
~ .-1-
~~~50-1'5-11- 1·5~2.75
1·5-3

4.5-8 R 1-5-2.3

2 1 0 1 2
Distance. inches
Fig. 8.20. Residual stress pattern in a maraging steel weld.

Dimensions are in mm
Fig. 8.19. Typical joint designs for arc welding mar aging steels.
The weld joint designs used for arc welding these steels are
shown in Fig. 8.19.
EBW in vacuum is particularly suited for joining maraging
steels because of the clean conditions during welding. In addition
to low heat input, narrow HAZ, and small distortion are
advantageous. In general, the weld metal toughness is below that
changes from tension to compression when the weld metal
expands during the phase transformation. At the same time, the
stress in the HAZ changes from compression to tension.
Hot cracking can be caused, in maraging steel weldments, if
the impurity levels are high. The low Mn in these steels makes
them particularly sensitive to sulphur embrittlement. Hot
cracking can occur with sulphur content as low as 0.005%, if the
joint fit-up is poor. With good fit-up, however, sulphur levels up to
0.010% can be tolerated. f
576 Welding Engineering and Technoi?gy
Weldability of Specific Materials 577
The higher strength maraging steels that contain larger
amounts of '1'i are more susceptible to hot cracking and thus The Mn content gen,erally ranges between 11 and 14% and
require greater care in welding. In view of the low austenite- carbon from 0.7 to 1.4%.Cr, Mo,Ni,V,Cu, Ti, and Bi are sometimes
to-martensite transformation temperature and the sensitivity to added singly or in combinations to provide special properties.
hot cracking, maraging steels should be welded without preheat Mn in steel has a strong affinity for 02, S, and C. Therefore,
and the interpass temperature be restricted to 120°C. it acts as a deoxidiser, reduces hot cracking, and contributes to
Control of heat input during welding is also necessary to avoid hardening.
hot cracking and inferior mechanical properties. The minimum 8.6.4.1. Physical and Mechanical Properties
practical heat input should be used. The physical properties of austenitic Mn-steel that affect
With GMAW,hot cracking in the weld metal of fillet welds is weldability are listed in table 8.23. The thermal and electrical
related to the depth-to-width ratio of the weld bead, as well as the characteristics of this steel are similar to those of other austenitic
alloy composition, the root gap, and the degree of joint restraint. steels; coefficient of thermal expansion is about 1.5 times that of
In general, a high depth-to-width ratio of over 0.6 or a large root mild steels. Carbide precipitation and transformation to pearlite
gap increases the tendency to hot cracking, as does increased joint will influence the physical properties in the range of 370° to 760°C.
restraint, say by increase in section thickness. Thermal conductivity is about 25% that of mild steel at room
Filler metal of similar composition to that of base metal, temperature, and this contributes to heat build up during welding.
produced by using vacuum melting techniquef' to obtain low levels The austenite in austenitic Mn-steel is almost non-magnetic thus
of 02' N2 and H2, be used. O2 and Nz should be below 50 ppm and Hadfield steel can be used where a strong and tough non-magnetic
Hz below 5 ppm. The carbon and sulphur levels of the filler must metal is required.
be kept as low as possible. Table 8.23. Physical Properties of Austenitic Mn-Steel
To achieve good toughness at the high strength levels, S.No. Melting
Coeff.
Thermal
Density
Specific
Electricalheat 0.68
1396
oftemperature
thermal
(room20.7
0.50
Value
13.4 Unit
expansion
conductivity
Property
resistivity
7.92 (RT*)
temp.)
(RT) °C
W/m-K
~m/m-K
~Q.m
KJlkg-K
g/cm3
impurities in maraging steel 1. weld metal must be kept at very low
6.
4.
2.
3.
5.
levels. Prior to welding, joint surfaces should be cleaned with clean
flint-free and suitable solvent. A dark surface discolouration would
indicate prior shielding of the weld zone.
8.6.3.4. Mechanical Properties of Maraging Steel
Weldments
Weld metal strength is generally below that ofthe base metal, * RT - room temperature.
particularly with the higher strength grades or maraging steel as Austenitic Mn-steel possesses outstanding toughness but
a result of small patches of austenite in multipass welds. Joint relatively low yield strength. The mechanical properties of
efficiencies based on yield strength of about 95% clm be achieved austenitic Mn-steel between -45°C and 205°C are excellent for
with appropriate welding procedures. The toughness of GTAW many applications. However, at high temperatures these steels
welds can match that of the base metal, though in general it tends may lack the strength and ductility necessary to withstand high
to be somewhat lower. welding stresses. These are not oxidation resistant, and their
8.6.4. Weldability of Austenitic Manganese Steel creep properties are poor in comparison with Cr-Ni austenitic
Austenitic Mn-steel, which is also known as Hadfield Mn-steel steels.
is extremely tough, non-magnetic alloy in which the hardening Ductility of Hadfield steel tends to increase with temperature
transformation that occurs in low alloy steel is suppressed by the until transformation of austenite begins. A hot short range starts
high Mn content. It is characterised by high strength, good between 815° and 870°C and it may extend to the melting
ductility, and rapid work-hardening characteristics, as well as temperature. This behaviour must be considered when
good wear resistance. establishing welding procedures to avoid hot cracking.
The impact properties of austenitic Mn-steel are high. At all
service temperatures experienced in railway track, mining, and
 578 Welding Engineering and Technology
construction service, cast Mn-steel has outstanding toughness.
This is very useful for service at sub-zero temperatures.
Standard austenitic Mn-steel is considered immune to H2-
induced embrittlement, for example, hydrogen up to a level of 6.3
mlllOO gm can be tolerated without adverse effects.
Mn-steels have a very high coefficient of work hardening,
exceeding that of austenitic stainless steels. The maximum
hardness obtainable is about 550 BHN (i.e. 625 VHN). The work
hardening properties make the Mn-steel particularly suitable for
applications where repeated impact loading occurs.
8.6.4.2. Uses
Austenitic Mn-steels are widely used in crushing, earth
moving, and material handling equipment, railroad track and
special applications where non-magnetic properties are important.
Although austenitic Mn-steel has relatively low yield and tensile
strengths in the as-cast condition, it has good wear resistance
under impact and erosion conditions. Gyratory and cone crusher
parts such as concaves, bowl liners and mantles, jaw crushers, and
hammer mill parts are examples of severe impact and abrasion
service. 'Weights of these parts vary from 50 kg to 15 tons. The
Mn-steels are also used for railway frogs, switches and crossings
where impact is severe. The non-magnetic properties make them
useful for parts for electromagnets, induction fumaces, and other
electrical equipment.
Large sections of work-hardened Mn-steel, ranging in
thickness from 3 mm to 13 mm are widely used as replaceable
wear shoes. They are used for railroads for applications such as
pedestals and journals. The shoes provide for vertical motion of
wheels on locomotives and cars while protected by replaceable Mn-
steel liners. The primary advantage of work-hardened austenitic
Mn-steel in this application is the ability to resist metal-to-metal
wear without lubrication. ' shop operations. Heat input into the base metal is less than with
8.6.4.3. Welding of Austenitic Mn-Steels
Since reheating of austenitic Mn-steel may cause carbide
precipitation and some transformation of austenite which
significantly reduces ductility therefore this steel should be
welded by using a process that results in minimal heat input. Arc
welding with a consumable electrode is generally used for joining
and surfacing applications.
Oxyacetylene and GTAW processes are not recommended
because they result in heat build up in the parts during welding.
Weldability of Specific Materials 579
Austenitic Mn-steel electrodes are available for shielded metal
arc welding of these steels. In addition to austenitic Mn-steel
electrodes, several other electrode compositions are available for
welding austenitic Mn-steels. The corrosion-resistant Mn-Cr and
Cr-Ni-Mn compositions are among the most commonly used
electrodes. They offer significant advantages for certain
applications-namely, higher abrasion resistance, ability to
work-harden rapidly under impact, and high tolerance to dilution
when joining austenitic Mn-steel to carbon or low alloy steel.
However, high Cr deposits cannot be cut by ordinary 02-cutting
methods. Austenitic Cr-Ni corrosion-resistant steel electrodes may
also be used to weld austenitic Mn-steels to carbon steels or as a
buttering layer on the carbon steel.
Welding Procedure. Preheat should not be used for welding
austenitic Mn-steels. The interpass temperature of the metal next
to the weld should not exceed about 315°C after a one minute
period to avoid excessive heating and embrittlement of the HAZ.
It should be possible to touch the base metal at a distance of 150
mm from the weld at all times without burning the skin.
The welding heat input should be the minimum required to
obtain complete fusion and joint penetration. A short arc length
and direct current, electrode positive (dcep) should be used to
deposit essentially stringer beads. Intermittent welding may be
used to limit heat build up. The weld bead should be peened while
very hot to upset it and relieve welding stresses.
Multipass welds are preferred to single pass welds because
succeeding passes will temper any martensite that may form in
the preceding weld beads. However, heat input and interpass
temperature restrictions must be strictly followed.
Semi-automatic or automatic arc welding may be done using
solid or tubular electrodes. These methods are advantageous for
SMAW because smaller electrodes and higher travel speeds can
be used.
When welding austenitic Mn-steel to carbon or low alloy steel,
Mn-steel electrodes may be used with proper procedures. Best
results are obtained when the filler metal contains more than 14%
Mn, less than 0.03% phosphorous, and dilution with the carbon or
low alloy steel is 25% or less. Otherwise cracking may occur.
Carbon or low alloy steel filler metal should not be used.
Austenitic Mn-Cr electrodes are preferred when the weld will
be subject to wear and high yield strength is needed. Type
 Weldability of Specific Materials 581

530 Welding Engineering and Technology

AISI 308, 309, 310 or 312 stainless steel filler metal can be used
w hen the welded joint will be exposed to little or no wear and
moderate stresses. The weld interface may be hard and brittle if
the dilution is excessive because of the high carbon in the
austenitic Mn-steel.
8.7. WELDABILITY OF CAST IRONS
'l.Ypicalautomotive parts made of gray C.!. include brake drums,
clutch plates, and cam shafts. Furnace parts, ingot and glass
moulds, and melting pots that operate at elevated temperatures,
are made of gray C.!., as various types of pipe, valves, flanges, and
fittings for both pressure and nonpressure applications.
Gray C.!. can be welded without undue difficulty if certain
precautions are observed concerning preheat and cooling rates
after welding. 'II'

Cast iron is a family of ferrous alloys that normally contain 8.7.2. White Cast Iron
more than 2% carbon and 1 to 3% silicon, as well as small amounts
of sulphur and phosphorus. In general, cast irons have lower White C.!. contains carbon in combined form as iron carbide. ill!

melting ranges than those of steels, have good fluidity when The material is therefore harder and more brittle than ordinary
molten, and undergo moderate shrinkage during solidification and gray C.I. It has a fine grain structure and a characteristic white
cooling. The toughness and ductility of cast iron are lower than appearance on fracture.
those of steel; these low properties limit the applications of cast White C.I. is seldom used as cast, except for parts such as dies,
iron. rollers, and moulds. It is usually cast as the first step in the
There are four major types of cast iron viz., gray cast iron, making of malleable C.!., which is much more useful material.
white C.!., malleable C.!. and nodular C.!. Most cast iron parts White C.!. is so difficult to weld that it is generally considered to
be unweldable.
are made in gray C.!.-a weldable material. Malleable iron and
nodular C.!. are also weldable, but they require postweld heat 8.7.3. Malleable Cast Iron
treatment to restore their normal strength and ductility. White This material is a white C.!. that has been heat treated by a
C.!. is normally considered unweldable because of the absence of long annealing process called malleableizing (usually lasting 4 to
adequate ductility to accommodate thermal stresses in the base 7 days) that changes the combined carbon (cementite) into iron
metal. Approximate composition of all the four types of cast irons and a uniformly distributed temper carbon i.e. ferrite and free
are given in table 8.24. graphite. This heat treatment increases both strength and
Table 8.24. Compositions of Different Types of Cast Irons ductility. Malleable iron is very much tougher than white or gray
C.!. and can be bent an appreciable amount while white and gray
S. Type cast irons break in a brittle manner with very little bending.
No. C Si S P Fe
Malleable iron is used for flanges, pipe fittings, and valve
1. Gray C.I. 2.0-4.0 1.0 min. 0.2 0.6 Balance parts. A number of automotive parts of malleable iron include
2. White C.I. 2.5-4.0 0.3-0.8 Balance steering components, compressor crank shafts and hubs,
3. Malleable C.I. 2.0-3.0 0.9-1.8 0.2 0.2 10.25-1.251Balance transmission and differential parts, connecting rods, and
max. max. universal joints.
4. Nodular C.I. When malleable iron is welded, the beneficial effects of the
;~;..:.........L-~~
3.2-4.11 04-1.610.15
1.8-2.8 0.4 I 0.80 ~~~lance malleableizing treatment are destroyed in the vicinity of the weld
Brief descriptions of the four types of cast· irons are given in by the sudden cooling from the welding temperature. The fusion
the following sections. zone changes back to chilled C.!. and must again be heat treated
to return t.omalleable condition.
8.7.1. Gray Cast Iron 8.7.4. Ductile Iron
Gray C.!. contains considerable free graphite, which gives Although quite similar to gray C.!. in composition, ductile or
broken surfaces a characteristic gray colour. Combined carbon in nodular iron possesses much of the strength and ductility of steel.
gray C.!. is usually less than 0.8%;the remaining carbon is in the The internal notch effect of flake graphite in gray C.!. is a prime
form of graphite. This material is the most common type of C.I.
and is widely used in the machinery and automotive industries.
582 Welding Engineering and Technology Weld ability of Specific Materials 583

factor in reducing its tensile strength, toughness, and ductility. In
ductile iron, this graphite is converted into spheroids that
m:inimise the notch effect and thereby produce a product
combining the strength and ductility of cast steel with the low
cost, wear resistance, and damping characteristics of cast iron.
An essential difference in composition between gray C.!. and
ductile iron is a small amount of magnesium or cerium contained
in ductile iron. These elements cause graphite to form into
spheroids. Also, in gray iron the silicon content is held within
lower limits, since Si has a coarsening effect on flake graphite.
This effect is not observed with graphite nodules.
Some culvert, sewer, and pressure pipe, as well as fittings,
valves, and pumps are cast of ductile iron. The advantages of
making these products from ductile iron are their relatively good
toughness and weldability when compared to similar components
of gray C.!. Ductile iron piping systems are commonly fabricated
by welding.
If Mg is lost from the melted zone of the base metal during
welding of a ductile iron casting, the shape of the graphite along
the weld interface can change. A welding process No.
and procedure
Welding
that minimises heat input and 2.
1.
4.
5.
3. time in the molten state will
minimise graphite degradation.
8.7.5. Welding Considerations for Cast Irons
Welding or braze welding of C.!. is generally used for:
(i) the repair of foundry defects,
(ii) the fabrication of components, or
(iii) the salvage of parts that break or wear out in service.
All types of cast irons except white cast iron are considered
weldable but to a lesser degree than carbon steels. Good joint
strength can be achieved for cast irons with tensile strength up to
and including about 480 MPa.
8.7.5.1. Welding Processes for Cast Irons
The welding processes employed for welding cast irons include
SMAW, GMAW,FCAW,oxy-fuel gas welding, braze welding, and
thermit welding. Main description here is with respect to SMAW
process.
When the casting is welded, the area adjacent to the weld is
raised above the critical temperature' i.e. to about 790°C. The mass
of the casting around the weld tends to draw the heat from the
weld area rapidly. If this cooling of the heat treated zone and the
weld is more rapid than it was when the casting was originally
made, a highly brittle, crack sensitive zone develops. The best
method to decrease the cooling rate is to preheat the casting to
prevent the cast iron from absorbing heat from the weld area.
8.7.5.2. Preheating
The formation of a hard and brittle HAZ can lead to cracking
during cooling in service. Low heat input with arc welding limits
the width of the HAZ but a band of hard brittle iron can still form
adjacent to the weld metal. HAZ hardness can be limited by
preheating in combination with slow cooling after welding. A low
cooling rate during and after austenitic transformation reduces
the amount of martensite and thus the hardness.
The preheat temperature and the extent of preheating depend
upon the type of C.!. being welded, the mass of the casting, the
welding process employed, and the type of filler metal used.
Recommended preheat and interpass temperature ranges are
given in table 8.25.
Table 8.25. Reconunended Preheat and Interpass
Temperature Ranges for Welding Cast Irons
S. Ductile
Gray
Type of
C.!.
40-335
iron
Malleable
castiron
Foriron
Preheat
40-165
40-335-
/Matrix
Pearlitic
Ferritic
microstructure
lnterpass
445-665
220-665
Oxyacetylene
welding
For Arc Temp. range, °C
All cast irons generally need to be preheated when welding by
oxy-fuel gas welding process to reduce the heat input
requirements.
Cracking from unequal expansion can take place during
preheating of complex castings or when the preheating is confined
to a small area of a large casting. Local preheating should be
gradual. Preheating of either a large section of the casting where
the welding is to be done or the entire casting in a uniform manner
is recommended. In any case, the preheating temperature should
be maintained during welding and welding should be completed
before the casting is cooled slowly to the room temperature.
Some castings, however, cannot be preheated. Welding
techniques and electrode selection must therefore take into
account whether or not the casting can be preheated.
 585
Weldability of Specific Materials
584 Welding Engineering and Technology
0U
>,
.-< .-<0.c.c
s::: o:l
0 .~6D2 ° Q,)
~~l.() .•... 00
l.() b.iJ
8.7.5.3. Filler Metal Selection
o::Eoo.::
'IJ .-< .:: Q
..!:l
c:ll
~.c
0::
v
<= .=
-015-0-0
s::: .S'
.•... ~
. g !-...
~
cd
~~:§~
8 0'"~I .S
.S'b
GC"3
0::0::Q:i
t:'::
oo~o
~J1~~
~<t:~~ ~.~
8 8
-----
-0",-0-0
I Q & ~' I ~ Cast iron can be arc welded with nickel, Ni-alloy, mild steel,
Cl
~ ..!..- ~
C"-l

~
'0)
~
"'-.
00"';1 ZZ"":' ,..,
'-'~,~,
""
<..l
and Cu-alloy covered electrodes. Bare gray C.1.electrodes are used
<;:'9';;; ..-I ifJ "'=It::.:1v
'" moo
..•oJ

;.i1r-r- - - a) <1;p:) I, c~ U)" C(fJ:<~

U
"-,
--;<~~ " Ie.
IN N " ::l v ::1v
_. <e
for special applications. Filler metal selection depends upon the
. ___

1:/)
l.()
. ____ 1

s:::
t::.:1t::.:1
Z"":'Z
o::'t"'=l
vv
""
I"o:l V::>vvo::
, ,
,1j
V 0:: 0; s::: '-I ..... type of C.1. to be welded and the application to which it is to be
:s: Z put. In any case, dilution with the C.r. should be kept to a
-- ~'-' >:"1

= ...•
~~
c.. ~t30-
...,..•
0c<I
•.. '-' S I "'=l
minimum. Covered electrodes for welding of cast irons are listed
U
~~I~
~~~
~~~I~
~
al
"1:1
r-~
"ti
~ e•
=
0
c:;
..s ~
=
f;I..
•..
a2
CIS
<l.l
CIS
<l.l
•....
<=
12 I

~
aC,
IS
III ::Q
:§
"tiI::
<$.
IS
"clC
CIS
<r.l
a•.. I r~
"J I il;1table 8.26.
~ e:.> w
8.7.5.4. Joint Design and Surface Preparation
I t§ Joint designs commonly used for arc welding of carbon steels
bD
are generally suitable for welding cast iron sections together. The
r- r- r-
s:::

.-<
1.C,.....jOO
.'-<
'<!' l.() l.()
.-<.-< .-<
l.() l.()
.-< .-< C'Je"'-IC'lCDt-<.or-
:.a
Q)
root opening should be wide enough to permit good fusion into the
~~~
• l.() • LCLcilci
~~~
1CU;
~~ ~~~~~~~ ~ root faces and into a backing plate, if used.
<..l
•....
'" For butt welds with V edge preparation, a 60" to 80° groove
'2'" angle is suitable. For thick sections, a V-groove with 20°-25°
a groove angle and 5-6.5 mm root radius should be used.
-0
'"
-0
Q)
When repairing cracked castings, a hole about 3.5 mm or
:..a
'"
larger in diameter should be drilled at the ends of each crack to
'IJ 'IJ prevent its further propagation. Then, sufficient C.1. should be
>.
o ~ I removed to eliminate the crack and provide room to properly
s: manipulate a welding electrode or torch during welding.
..!:l

u
.2i::E::E::E
Q,)I~ ~ ~ Q:i::E::E::EcJ~::E
ooU)U)o o t:'::~
.!:llOU)U)~i:I)U)~ Q,
~I~
Z
~~
\.o~~:;:;:<t:~~<:
~ooo::E::E::E::E
U)OU)O
~S:t:':::<-t:'::t:'::t:'::t:'::

Uo
~
~
bD
C
To prepare surface for repair welding, it should be heated
uniformly at about 370°C for 30 minutes, or for a shorter time at
:.a
Q) 5400C, using an oxyfuel gas toreh. An alternative is to heat the
~ weld area by depositing the first weld pass and then removing it
<..l
•....
'" by suitable means because the weld metal is generally very porous.
'2
'"
This welding and removal operation is repeated until the weld
a metal is sound. rrhen, the weld is completed in the normal way.
'" Q)

°
-0 •....
'" Castings that have been impregnated with plastic or glass sealant
'"
-0
° -00
'"
~-g1j-g-g
Q,) Q) s... Q) Q)
°
'"
-0 "t:l
0
<I.' ..0
should not be repair welded in any case because the sealant will
I
•....

U <..l 0
•....
.•..•.•..• -0
'"
Q)1:;b..§bb
~gasCIJ~irl ""O""O""O~-gg-g
~ Q) r... Q)

s: produce gross porosity or may inhibit bonding.

~~~ .J:j ......•
u Q)
-""o--
Q) Q,) Q) OJ 888Qj~~~
~
::E
8.7.5.5. Welding Procedure
bD bD bD -0 -0
~~~
•••••••••• Q)
Q)
Q)~~8~~
""0 ""0 Jo...t "'0 ""0

.5 .9 .S ~
Q)
OJ ~
Q)
Qj
o:l
For welding C.1. the current for a particular electrode size
'"
:> :>
00'"
'" Q)
•...• cdOO_OO
~ ~ ~ ~ g; ~
:g:9:g §: Q) g; ~ bD
s::: should be within the range recommended by the manufacturer but
vvo:l o:lc.)v~vU gg",o@o",
;>;>:;:Uo:lvo:l :.a
Q) as low as possible, consistent with smooth operation, desired bead
~
'" contour, and good fusion. When used in other than downhand and
s:::
Q)
;;,
horizontal position, the welding current should be reduced by
'" '" a a
.•..•
Q)
<..l
about 25% for vertical welding and about 15% for overhead
~~ '" Q) '" '" ~ ~ '"
welding.
'<!''<!' ~ ~
...c ..0
"2
.- "a
.- ~
aiQj'Q) :J aa o The electrode should be manipulated so that the bead width
'" Q) Q)
.•....•........
",U)U)
~ ~ v vi
QJ Q) ~ bDb.OgJ
.S.S ~.5.S 2 ~ I
is no greater than 3 times the nominal diameter of the electrode .
I..i,)U"::l0 0
Z Z ._ ._ ._ '?
._ ES..o1"tZ2 s:
s:::
000
on on .0
s::: s::: .- .- '"'r "'t "t

22Q3Q)QJQ)Q)
o~ If a larger cavity must be filled, the sides should be buttered with
@ @ @ <;\1~ZZZZ
Mli.:lMMLO.l!';)
:;:~-;1§:§:§:§:
o 0 .- 0 0 0 0
mO')lOLt:lLl'J<O
vvv
....:l....:l Z v v v VI *
 ill
586 II,
!
Welding Engineering and Technology Weldability of Specific Materials 587 I'

'I,

weld metal first, and then the cavity gradually filled toward the mm dia.) peening hammers. Buttering and studding are other III
centre of the repaired area. With large castings, a backstep methods of getting sound and strong weldments in cast iron. I'
sequence provides a more even thermal distribution. Buttering. Where large V edges are to be welded, it is qill
The Ni-alloy covered electrodes are specifically designed for sometimes beneficial to line the sides with an ESt electrode, and II"

welding of cast irons. Deposited metal from these electrodes have
a carbon content well above the suitability limit. The excess carbon
is rejected as graphite during solidification of the weld metal. This
causes an increase in volume that tends to minimise the weld
metal shrinkage during solidification. That, in turn, reduces
residual stresses in the weld metal and the C.r. heat affected zone.
In general, these electrodes are a better choice because of the
higher strength and ductility of the weld metal.
Mild steel electrodes, such as E7018 and ESt types are used
primarily to repair small cosmetic casting defects where colour
match is desirable and machining is not of major concern. Dilution
and cooling rates should be kept low to avoid increased hardness
of the weld metal and the HAZ.
then fill in the remaining space with mild steel electrode E7018.
This process is called buttering. Shrinkage is lessened since the
layers of deposited metal on each side of the groove can cool
without restraint, and a smaller volume of weld deposit, subject
to restraint, is required for filling the final groove. Ii
Ii
Studding. Because the chilling action of the C.!. increases the
combined carbon which in turn increases hardness and brittleness
that reduces the strength of the C.l. just outside the line of fusion. i,
I,
Welds in C.l., if of sufficient thickness, may be strengthened Ii
mechanically by a process called studding. Steel studs about 6 to I
li;1
10 mm in diameter should be used. The crack in the casting should Ii
be V-cut, drilled, and tapped so that the studs may be screwed into [I
the V or the area of reinforcement. The studs should project about 'I

Mild steel electrodes can be acceptable for joining ferritic 5 mm to 8 mm above the surface, and should be long enough to
ductile or malleable iron castings to mild steel components. be screwed into the casting to a depth of at least the diameter of II
However, the steel weld metal shrinks more than cast iron and the studs. Fig. 8.22(a) illustrates the practice of studding. i
causes stresses to develop at the weld interface. The stresses can r
be severe enough to cause cracking in the cast iron HAZ.Toreduce I
the strain caused by the deposition of long bead, deposit the weld
metal in short lengths as shown in Fig. 8.21 and allow each bead

(J (0 I By studding
(b I By lock formation

(]
Fig. 8.21. Suggested sequence of laying short head lengths in welding.
to cool down before laying the next. For example, weld for 15 to
30 seconds and then allow the weld to cool for 3 to 5 minutes. By
depositing small welds in various parts of the work one weld is
allowed to cooland contract while the operator is depositing metal
in another location. An altemative is to upset or peen the deposited
weld rnetallightly before it has a chance to cool and contract. This
e:auses the weld metal to stretch. In many cases, it is best to use
a combination of short welds and peening. Use small-ball (13-20
Fig. 8.22. Meth:Jds of reinforcing welds in cast iron: (a) by studding, and (b) by
forming round-bottom grooves or locks into the prepared edges.
Cross-sectional area of the studs should be about 25% to 35%
of the area of weld surface. In such cases the strength of the weld
may safely and conservatively be taken as the strength of the
studs. It is considered a good practice to first weld one or two beads
around each stud, making sure that fusion is obtained both with
the stud and C.!. base metal. Straight lines of weld metal should
be avoided, if possible. Welds should be deposited intermittently,
and each bead should be peened before cooling.
Double V-groove edge preparation is advisable where the
casting is of sufficient thickness. In many cases it may be desirable
to produce complete penetration at the fracture. In some cases, it
 588 Welding Engineering and Technology
is a better practice not to stud but to shape grooves or locks into
the casting, with round-nose tool, as shown in Fig. 8.22(b).
Fusion welding, because of localised nature, produces thermal
stresses in the weld area. Inability of the weld metal to yield and
relieve welding stresses can result in cracking in the adja~ent cast
iron HAZ.Also, the weld metal can become hard with low ductility
and it may be difficult to machine unless a suitable postweld heat
treatment is applied.
8.7.5.6. Postweld Heat Treatment of C.I. Weldments
Thermal stress relieving is desirahle for fully restrained
welds, welds intended for use in severe service conditions, and
welds subjected to close machining tolerances. Normally, stress
relieving is done immediately after welding by increasing the
temperature of the entire casting into the 595° to 620°C range and
the casting is held at that temperature for about 2.5 minutes per
mm of thickness. The cooling rate should not exceed 28°C p,~rhour
until the casting has cooled down to about 370°C. Some reduction
in the hardness can be achieved by this stress relieving treatment
but it will be slight because the carbides are stable at 595°C.
Heat treatment at 900°C followed by slow cooling to 540°C .or
lower results in maximum softening and stress relief. To obtain
optimum ductility, the weldment should be given this heat
treatment immediately after welding. Reduction in strength must
al ways be considered when selecting a stress-relieving
temperature.
For the best results with ductile C.L, the welded casting
should be immediately placed in a hot furnace at 595° to 650°C
and the temperature raised to 9000e and the casting be held at
that temperature for 2 to 4 hours. It is then cooled to 705°C, held
there for f:j hours and then cooled down to 595°C in the furnace.
The casting may then be cooledto room temperature in the furnace
or in still air.
If a casting is welded with a high preheat and interpass
temperature of the order of 650°C, it should be slowly cooled by
covering it with an insulating blanket, vermiculite, or hot sand.
Malleable iron castings may be heat treated after welding.
8,.7.5.7.Other Processes for' Welding Cast Irons
Apart from arc welding processes, other processes used for
welding cast irons are :
(i) Oxy-acetylene welding,
(ii) Braze welding,
(iii) Thermit welding.
i
I',
~-~---
i. __
Weld.ability of Specific Materials 589
Brief descriptions of these processes for C.!. are given in this
section.
(i) Oxy-acetylene Welding: Oxyacetylene welding
procedures require large amounts of heat input during both the
preheating and the welding operations. The extensive heating is
a limiting factor in the application of this process to finished or
semi-finished castings where distortion or dimensional stability
can be a problem. On the other hand, the resulting slow cooling
rate lessens the tendency for brittleness in the HAZ. Only limited
success has been achieved in welding malleable iron by this
process.
The joint design is normally a double-V edge preparation with
a 90° groove angle. If welding can be done only from one side, the
single V groove angle may be increased to 120°. The joint groove
must have sufficient width to permit good fusion of the root faces
and uninterrupted torch manipulation. If a deep hole is being
repaired, the groove should be elongated to avoid torch backfire.
For gray iron castings, gray iron welding rods of AWS type
RCI or RCI-A are used. Type RCI-B rods are designed for welding
ductile cast irons but can also be used with gray cast irons.
A flux is required during oxyacetylene welding to increase the
fluidity of the fusible iron silicate slag that forms in the molten
weld metal. This slag is difficult to remove when a flux is not used.
Prefluxed welding rods are commercially available, but the
common practice is to dip the heated end of a bare welding rod
into the flux and transfer it as required to the weld. Excessive flux
can result in harmful slag entrapment.
Fluxes, for welding gray C.!. contain borates or boric acid, soda
ash, and other compounds such as iron oxide, sodium chloride, and
ammonium sulphate. Most fluxes are of proprietary compositions,
but a typical flux contains equal parts of boric acid and soda ash,
2% ammonium sulphate, and 15% powdered iron.
For oxyacetylene welding of cast irons a preheat temperature
between 595° and 650°C is normally employed to compensate for
heat input of the process. Thin sections can be welded with lower
preheat temper:ature because the heat loss into the base metal will
be less than with thick sections.
Welding is performed with a neutral or slightly reducing flame
using welding tips of medium or high flame velocity. The backhand
welding technique is recommended. The tip of the welding rod
590 Welding Engineering and Technology
8 hould be kept in the molten weld pool to minimise slag inclusions
and porosity in the weld metal.
The weld metal should be deposited in layers less than 3 mm
thick. For good fusion with joint faces, the molten pool must be
kept small and torch must be directed on the base metal. Moving
the welding rod about in the molten weld pool helps to float slag
to the surface ..
For ductile cast irons, melting of the base metal should be
restricted to that required for good fusion. An interpass
temperature of between 595° and 650°C should be maintained. If
the temperature falls below this range, porosity may result from
gas evolved by the molten base metal. Immediately after
completion of welding, the casting should be insulated to provide
slow cooling. Stress relieving or annealing of the casting is
recommended.
(ii) Braze Welding: Braze welding with an oxyacetylene
torch can be used to make field repairs to castings, but it is not
normally used to repair new castings because ofpoor colour match.
Copper alloy filler metals commonly used for braze welding are
listed in table 8.26. Joint strengths equivaler.t to those of fusion
welds are possible with gray C.r. but not with ductile or malleable
irons. With ductile iron, a maximum joint efficiency of 80% may
be expected. Corrosion resistance of the joints is generally poor.
V-groove edge preparation similar to those used for arc or
oxyacetylene welding are suitable for braze welding of cast irons.
The Cu-Zn welding rods are used with neutral or slightly
oxidising flame. To prevent excessive oxidation, the molten weld
pool should be kept covered with a thin oxide film. Preheating with
the welding torch is satisfactory for small castings. Large castings
require preheating to a temperature in the range of 315° to 400°C,
or higher prior to welding. Generally a suitable flux is applied to
the welding rod by warming one end with the flame and then
dipping it into the powdered flux. The flux-coated rod is applied
to the joint during welding.
The fact that braze welding is done at a temperature of several
hundred degrees below the melting point of C.r. is an advantage
with respect to dilution. Welding is best done in the flat position.
After welding the casting is covered with insulating material and
allowed to cool slowly to room temperature.
(iii) Thermit Welding: Iron castings can be repaired by
thermit welding, especially large structures such as machine
bases or frames. The shrinkage of weld metal produced by
alumino-thermic reaction is significantly greater than that of gray
or ductile cast iron. For this reason, difficulties can arise when the
Weldability of Specific Materials 591
process is used for the repair of cracks that are longer than 8 times
the section thickness or that do not extend through the section.
The process is essentially the same as that used for the
welding of steel. Cast iron thermit mixture normally consists of
aluminium powder and iron oxide to which 3% ferrosilicon and
20% mild steel punchings are added. Because of the insulation and
slow cooling provided by the sand mould surrounding the joint,
the weld metal is generally somewhat harder and tougher than
ordinary gray C.r. but is machinable. Stress relieving is usually
not required unless a condition of high restraint is encountered.
8.8. WELDABILITY OF NICKEL ALLOYS
Because of their high strength and good corrosion resistance
at high temperatures, Ni-base alloys have become the most
extensively used high temperature alloys. Typical applications for
commercially pure Ni are food processing equipment, caustic
handling equipment, chemical shipping drums, and electrical and
electronic parts.
Ni-alloys are divided mainly into two groups:
1. Solid-solution alloys,
2. Precipitation-hardening alloys.
Different solid-solution Ni-alloys include the following types,
(i) Ni-Cu alloys,
(ii) Ni-Cr alloys,
(iii) Ni-Fe-Cr alloys,
(iv) Ni-Mo alloys,
(v) Ni-Cr-Mo alloys.
while the Precipitation-Hardening Ni-alloys are of the following
types,
(i) Ni-Cu alloys,
(ii) Ni-Cr alloys,
(iii) Ni-Fe-Cr alloys.
Compared with pure Nickel, Ni-alloys have significantly lower
thermal and electrical conductivities. Some of the typical
solid-solution and precipitation-hardening Ni-alloys and their
compositionsat:e given in table 8.27.
These alloys are representative of the larg~ number of
available Ni-alloys, some of which are referred to as Superalloys.
8.8.1. Welding of Nickel Alloys
The welding of Ni-alloys is similar to the welding of austenitic
stainless steels. Nickel and Ni-alloys can be joined by all of the
592 WeldabiUty of Specific Materials 593
Welding Engineering and Technology
J ° =° °~
<D
11
C'l
11
C'l I
arc welding processes commonly used for steel and other metals.
oII 100II 0'0' However, not all of the processes are applicable to every alloy
~I N'"' N N.
11
..c ..c
E-< E-< because of metallurgical characteristics or non-availability of
suitable welding materials. In general, the precipitation
J °oo °°0000° 0
l:.O If:) (,0 <.c
<::>
hardening alloys ,require closer control of the welding process
variables because of the possibility of ageing and the formation of
refractory oxides during welding.
I
M 8.8.1.1. Surface Preparation
00
>. I ~ I Cleaniness is the single most important requirement for
o
0 ° U) C'l .••.•.• ~ l.f:) C"l

- 6 0 000 0 successful joining of Ni and Ni-alloys.

C'-~
U)
C'.! ~ I"""f

0 0~
~
~ ,'_ C'l Oxides should be thoroughly removed from the surfaces to be
..::d
Q ~~ I 0 C'l M ~ M 0
.....• ~ ~ ~ ~ ~ I ~
:>'000000:>. ~ ~ ~ ~ welded because they can inhibit wetting and fusion of the base
-Z
!-
",'<
o 0 metal with the weld metal. Their presence can also cause
~ ~ N ~ ~ subsurface inclusions and poor bnad contour. They may be
~1{3 :>.
o 6 c.o
:::
I I I I I I ..... "tl~
removed by grinding, abrasive blasting, machining or pickling.
.... :::
Ci
~ ::: ~

o
:~I
o
?
r;:: =
.::
.•..
::s
~~ ~
~
0
~ M M °
~ ~ ~ ~
C'l C'l C'l ~
:::
Wire brushes used for cleaning should be made of austenitic
stainless steel. However, they will not remove tenacious surface
m '0 ~ ~ ~ M u) <D ~ ~ e oxides from welds. Those oxides must be removed by grinding with
= 81~
"-o
1:1 ' , •• , , , ••••
rn o ~ ~ 0 00
.-m C:l
~ ~
I"""f .•••••• I"""f ~

d an aluminium oxide or silicon carbide wheel. Any cleaning tool,

"- '-' ~ 0 including wire brushes, should be clean and free of other metals
oc.. '~I
'" ~::l
~
o 0
rn ~
.....• °
~ M
~ .~ ~
0 ~~ I I I I I 'iil
M
Q
.S< 111
~ .that may be possibly transferred to the base metal.
a c5 ~
~ ~
uo
- I 8.8.1.2. Joint Designs
....

8 L()
c-.i ~ ~ S I ~
............ I C
Suggested designs for butt weld joints in nickel and Ni-base
~
Q alloys are as shown in Fig. 8.23. Molten Ni-base alloy weld metals
"s ~16 ~ ~ ~ ex> ;;g
M
o M •....• ~
~ 3,3 max 3- 6

-u
..=

~
=
..£
~
~
<D
L()
c-.i
~° M -.jo

"s
o ~
... ""= ~
cq ~ ~
<::>

C'-.t
~ en
..................
0') ~
~
(j)
00 ~
.••.•.•l(j
o
C'l
i
I..bar~
Temp~rary
backlOg
0-',5' (J. 1~
e~[jing IL
0.75 -,.~
Z
l L
60:- 60-
~
~
-.c
cx5
~
00 ~ ~ ~ 00 ~ 10 0 00 0 ~ 00 ~

,<:80 Y 'Y'::80· y
~ j' - ~
~16 06 06 ~
~~~~~
M
6 6 666
~~ 00 ~ = ~
6 6
M 00 00
;~r:ro,y
5-'6/ ~ ~"Ikl 6-raa'king
-16\ ~" -, ••. ·3 1
/.-\:J ',5
-;r~---~~~_:_-----~--~
~
°
-~--=--=-~-=---~~~- 0
..0
0
gO
M
....
backing bar

l
s::S
0-) ~ oC'l..-o bO~
E:gIO
(..)~IW
Q ~
0
Q)'-~8~~
ct;-50E--
0..., "'",
8(j)a~
s...C:rJJ~;:::=O
000
§;;~
-
zz
1-
13-
••9 0~
I -l
I
I Z
.-
~ =
ro
z~u
..... ~ (Ij
0~ ~~z~
::: -0~
_~ __
ro ~ ~
~
Cl 32

, '.14
I
<.
1,-:=g-1jl
§'2P , g
~--- ~
-:- §M- :---:-:- a
.
:::: ~ t-:
~ :--~--:-:
~
"""'" ....-l ......-I....-l
.£' ::;: ,~ -~~-----:-
~:::
.-I
~
.
U)
OJ' All
T3
dimensions in mm

Fig. 8.23, Suggested designs for arc welding butt joints in Ni and Ni·alloys.
594 Welding Engineering and Technology
do not flow and wet the base metal so readily as do carbon and
stainless steel weld metals. The groove angle must be large enough
te> permit proper manipulation of the filler metal and deposition
of stringer weld beads.
With GMAWprocess, the high amperage used with a small
diameter electrode produces high level of arc force. Such an arc is
not easily deflected from a straight line. Consequently, the joint
design must permit the arc to be directed at all areas to be fused.
D-groove joints should have a 30° bevel angle. This will permit
proper manipulation of the arc to obtain good fusion with the
groove faces.
Other factors also influence joint design, for example, the·
age-hardenable alloys are sluggish when molten, and full
penetration is difficult to achieve in some joints. It is also difficult
to avoid lack of fusion in deep groove welds. Another important
factor is the lack of ductility in the welded joint at or near the
ageing temperature of the alloy. Fillet, flare, and edge welds
should not be exposed to temperature above 540°C, because of the
built-in notch at the root of the joint. If such welds are
unavoidable, the filler metal must be tolerant of notches at
elevated temperatures to avoid cracking of the weld metal during
heat treatment.
8.8.1.3. Preheat and Interpass Temperatures
Preheat is not required nor recommended for welding Ni- base
alloys. However, the area to be welded should be at about 16°C or
above to avoid moisture condensate that may produce porosity in
the weld metal.
The interpass temperature should be low to help minimise
total heat input. Amaximum temperature of93cC is recommended
for corrosion-resistant alloys. Cooling methods used to reduce
interpass temperature should not introduce contaminants that.
will cause weld discontinuities. Examples are traces of oil froni
compressed air or mineral deposits from a water spray.
8.8.1.4. Heat Input Limitations
High heat input during welding may produce undesirable
changes in Ni and Ni-base alloys. Some degree of annealing and
grain growth can take place in the HAZ. The heat input of the
welding process and thE!interpass temperature will determine the
ext.ent of these changes. High heat input may result in consti-
tutionalliquidation, carbide precipitation, or other metallUl'gical
phenomena which may lead to cracking or loss of corrosion
resistance, or both.
WeldabiIity of Specific Materials 595
Stringer' beads should be used to fill the joint and a convex
bead shape helps in minimising the hot cracking tendencies.
8.8.1.5. Welding Processes for Ni and Ni-Alloys
The arc welding processes most often used for joining Ni and
Ni-base alloys are GTAWand GMAWwhile SMAW and SAW are
used in some limited cases.
GTAW: Gas tungsten arc welding is widely used for joining
Ni-base alloys, especially for thin sections and for applications
where a flux residue would be undesirable. This process is best
suited for welding precipitation-hardening alloys.
Argon, helium, or a mixture of the two is normally used for
shielding. Argon is normally used for manual welding. Hydrogen
in amounts up to 5% may be added to Ar to improve cleanliness
with single pass welds. However, H2 addition may cause porosity
in multipass welds with some alloys.
Helium has some advantages over Ar for machine welding thin
sections without the addition of filler metal. Higher welding
speeds are possible. Porosity-free joints in Ni-Cu alloys are more
readily obtained with He, and porosity can be reduced in welds in
commercially pure nickel. With d.c. electrode negative power,
welding speeds can be increased as much as 40% over those with
Ar shielding and the same welding current. However, arc initiation
is more difficult with helium.
During welding, the root of the joint must be shielded from
the atmosphere to prevent oxidation of the weld and base metals.
. This can be done by purging the area with the same gas as used
for welding. The gas may be introduced through a temporary
backing bar or cup, or contained by internal dams in the case of
tubing or pipe.
Arc stability is best when the tungsten electrode is ground to
a point. Cone angles of 30° to 60° with a small flat apex are
generally used.
Direct current electrode negative (dcen) power is recommen-
ded for both manual and automatic welding. A.C. with super-
imposed high frequency power may be used for automatic welding
if the arc length can be closely controlled.
The welding torch should be positioned with the electrode in
a nearly vertical position. If the electrode is more than 35° from
the vertical, air may be drawn into the shielding gas and cause
porosity in the weld metal with some Ni-alloys. The shortest
possible arc length must be maintained to ensure sound welds.
vVhen welding without filler metal, the arc length shOl..',ldnot
exceed 1.25 mm, and preferably be 0.5 to 0.8 mm. When filler
metal is added, the arc length should be adjusted to permit proper

596 Welding Engineering and Technology
manipulation of the filler rod and molten weld pool. The
composition of the fille~ metal for GTAWshould, in general', be
similar to that of the base metal. Nickel filler metals frequently
contain additions, such as titanium, manganese, and columbium,
to control porosity and hot cracking.
Square groove welds can be made in section up to about 2.5
m.m thick in a single pass. When automatic welding methods are
employed, travel speed should be adjusted to avoid a teardrop-
shaped weld pool because such a weld pool is more prone to
ce~treline cracking during solidification than is an elliptically
shripedone.
All solid-solution Ni-alloys, except high silicon casting alloys,
are readily welded by GTAWprocess. Commercially pure nickel
and Ni-Cu alloys require special care to prevent porosity in welds
m.ade without filler metal. Filler metals normally contain
deoxidisers to counteract porosity. Therefore, filler metal additions
are generally recommended when welding these alloys. Ni-Cr and
Ni-Fe-Cr alloys are less susceptible to weld metal porosity. Filler
nletals for these alloys contain alloying additions to overcome the
cracking tendencies ..
GTAW process is widely used for precipitation-hardening
alloys because it provides excellent protection against oxidation
and loss of the hardening elements. Aspecial precaution that must
be observed, especially for alloys high in Al and Ti, is the removal
of refractory oxides that form on the weld bead surface. If the
oxides are not removed, they can cause lack of fusion with
subsequent weld passes.
GMAW: This process can be used to weld all of the solution-
strengthened Ni-alloys except high silicon casting alloys. Many
precipitation-hardening alloys may also be welded but electrodes
of matching composition are seldom used because they do not have
adequate resistance to cracking, particularly with multipass'
welding.
Direct current electrode positive (dcep) should be used with
an electrode diameter of 0.9 mm, 1.2 mm, or 1.6 mm, Constant
potential d.c. power sources are normally used for GMAW.
Argon or argon-helium mixtures are normally used' for
shielding. Ar is normally recommended with spray or pulsed spray
metal transfer. Additiqn of 15 to 20% He to Ar will increase bead
width and decrease penetration. Ar-He mixtures are recommended
for short-circuiting transfer to take advantage of the higher arc
energy due to helium.
As with GTAW, the root of the joint must be adequately
shielded to prevent oxidation. A backing strip may be used for
~ i
WeldabiUty of Specific Mate~ials 597
welding Ni or solution-strengthened Ni-alloys that are not crack
sensitive.
The welding gun is positioned nearly perpendicular to the axis
of the weld, consistent with good visibility of the arc and good
shielding of the weld pool. An arc length of about 6 mm is
considered suitable with spray mode of metal transfer.
SMAW : This process is used primarily for welding
commercially pure Ni and solution-strengthened Ni-alloys. The
precipitation hardening alloys are seldom welded with this process
because the alloying elements that contribute to precipitation
hardening are difficult to transfer across the welding arc.
Minimum thickness of metal welded by SMAW is generally
about 1.5 mm. Ni-alloy covered electrodes are available for all
position welding of most alloy families. However, protection
afforded by the covering is not so good as that proviqed by the gas
shielded processes.
Covered electrodes for SMAWshould have composition similar
to the base metals for which they are intended. They normally
have additions of deoxidising elements like Ti, Mn, and Cb. These
elements also prevent weld metal cracking. Ni-alloy electrodes are
divided into five classes, namely Ni, Ni-Cu, Ni-Cr-Fe, Ni-Mo, and
Ni-Cr-Mo alloys.
The covered electrodes are generally designed for use with d.c.
electrode positive (dcep) power. Each electrode type and size has
an optimum current range for good arc characteristics. Excessive
welding current can cause a number of problems including an
unstable arc, excessive spatter, electrode overheating, and spalling
of the coating. Sizes of3.15 mm and under are suitable for welding
in all positions. Other sizes are restricted to the flat and horizontal
position of welding.
Welding in the flat position should be done with the backhand
technique using a drag angle of about 20°. Stringer beads are
recommended for the Ni-Mo and Ni-Cr-Mo alloys. If a weave
technique is used, it should not be greater than 1.5 times the
electrode diameter. Welding in PC?sitionsother than downhand is
not recommended for these alloys.
SAW: Fluxes are available for submerged arc welding many
of the Ni-alloys. The resulting slag cover will be self-lifting when
used with proper welding ~onditions.
The electrodes used with these fluxes are the same as those
used for GMAW.The 'flux composition must be suited to both the
filler metal and the base metal being welded. An impropertlux
598 Welding Engineering and Technology
can cause slag adherence, weld cracking, inclusions, poor weld
bead contour, and undesirable changes in weld metal composition.
FLuxes used for carbon and stainless steels are not suitable for
welding Ni and its alloys. When recycled flux is used, it should
always be mixed with atleast 50% virgin flux.
In most cases, the corrosion resistance of submerged arc weld
metal is generally lower than that of weld metal deposited with
GTAW and GMAWprocesses. Submerged arc welding of Ni-alloys
can be done with d.c., electrode, positive or negative. Electrode
positive will produce a flat bead with deep penetration and is
generally used for welding while electrode negative provides a
sLightly higher deposition rate and shallow penetration and is
generally used for surfacing.
Oxy-Acetylene Welding: The oxyacetylene flame produces a
sufficiently high temperature for welding commercially pure
nickel and some solution-strengthened Ni-alloys. However, this
process should be used only in those situations where suitable arc
welding equipment is not available. Welds can be made in all
positions with practice ..
The torch tip should be large enough to provide a low velocity,
soft flame. The welding torch should be adjusted with excess
acetylene to produce a slightly reducing flame but .when
Cr-bearing alloys are welded the flame should not be excessively
reducing because the weld metal might absorb carbon.
Flux is required for welding Ni-Cu, Ni-Cr, and Ni-Fe-Cr alloys,
however commercially pure nickel can be welded without flux. The
following mixture may be used as flux for solution-strengthened
Ni-Cu alloys,
Barium fluoride 60%
Calcium fluoride 16%
Barium chloride 15%
Gum arabic 5%
Sodium fluoride 4%
Precipitation-hardening Ni-Cu alloys can be fluxed with a
water slurry of one part lithium fluoride and 2 parts of the above
men tioned flux composition.
The flux is mixed with water to produce a thin slurry. It should'
be applied to both sides of the joint and to the filler rod, and then
aLlo'Wedto dry befor~ welding is started.
- Borax must not be used as a flux when welding Ni-alloys
because it can form a brittle, low melting eutectic in the weld
Weldability of Specific Materials 599
"
which is undesirable. The flux residue must be removed from the
joint for high temperature service.
During welding the molten weld pool should be kept quiet with
the inner cone of the flame just touching its surface. Puddling of
the molten metal should be avoided because that can l'esult in 10s5
of the deoxidisers or exposure of the molten .weld metal to the
surrounding atmosphere. That may result in porous welds. The
hot end of the filler rod should be kept within the flame envelope
to minimise oxidation.
'fhe oxy-acetylene welding process is not recommended for
joining the low carbon Ni-alloys, Ni-lVIoalloys, and the Ni-Cr-Mo
alloys because the material can readily pick up carbon from the
flame and that may reduce their corrosion resistance and high
temperature properties. However, oxyacetylene welding is the only
recommended joining process for the Ni-l0Si-8Cu-2Fe alloy
(Hastelloy D). The welding procedure is similar to that used for
cast iron. The filler rod should be of the same composition as the
base metal. A V-groove with a ~5°bevel angle should be used for
sections over 13mm thick.
With the exception of the Ni-Cu alloy, the precipitation-
hardening alloys should not normally be welded by the oxy-
acetylene process because the hardening elements are easily
oxidised and fluxed away during welding.
Other Welding Processes: Other welding processes which
can be used successfully for welding Ni and Ni-alloys inClude
plasma arc welding, resistance spot and seam welding, flash
welding, projection welding, electron beam welding, laser welding,
brazing, and soldering.
Nickel and Ni-alloys cannot be cut by the conventional oxy-fuel
gas cutting methods. Plasma arc cutting and air carbon arc
gouging are normally used with Ni-alloys.
8.8.1.6. Postweld Heat Treatment
Postweld heat treatments are usually not needed to restore
the corrosion resistance of Ni, Ni-Cu, Ni-Cr, and Ni-Fe-Cr alloys.
However some alloys may require a stress-relief heat treatment
for specific corrosion resistant applications.
The Ni-Cr and Ni-Fe-Cr alloys, like some austenitic stainless
steels, can exhibit carbide precipitation in the HAZ. Such
sensitization does not result in accelerated attack in most
environments. Some alloys are stabilized by additions of titanium
and columbium.
600 Welding Engineering and Technology Weldability of Specific Materials 601

. Welding of Ni-Mo and Ni-Si alloys can influence the corrosion 7

. resistance of the metal in the HAZs. Generally, the weldments
should be solution-annealed to restore corrosion resistance. Low
carbon and low silicon Ni-alloys do not require a PWHT for
6 o Alloy 713 C "
corrosion resistance. 01N 100

Under certain PWHT combinations of temperature and stress,

the HAZ may crack. This is known as Strain-age Cracking.
8.8.1.7. Strain-Age Cracldng 'cF.
o Ouranickel
301

l\-fost of the precipitation-hardening Ni-alloys are subject to E

-
strain-age cracking. Fig. 8,24 illustrates the welding and PWHT
welding
----------~,
Postweld heat treating
.~
c
c:
::>

<i 8';-...
Monel
K-500
......•...
'-
o Udimet 500

~I ,
residual
aging -
stress
crack
+
solution rwew~
Nimonic 8ciAO ,0·'
~ene 41

o "-
=>
+-
o
•..
a.
Q.O

E o
o
Inconel718 M25Z0 Waspaloy.....
Inconel x-750 ,
~
I-
Q.O

o 6 7

Fig. 8.25. Effect of Al and Ti contents on weld ability of

Time-
Fig. 8.24. Schematic illustration of a welding and PWHT cycle and the location
of PWHT cracking point fpr heat-treatable Ni-base alloy.
cycle ofaheat-treatable Ni-base alloy and the occurrence ofPWHT
cracking. Alloys high in aluminium are the most sensitive ·to this
type of cracking. The problem is much less severe in those alloys
where columbium has been substituted for a significant portion of
the aluminium. Columbium retards the ageing reaction.
Consequently, the HAZ can remain sufficiently ductile and yield
during heat treatment to relieve high welding stresses without
rupture. Strain-age cracking may occur during ageing, annealing,
or stress-relieving unless precautions are taken to prevent high
residual stresses.
With a few exceptions, susceptibility of Ni-alloys to postweld
strain-age cracking is closely related to the combined Al and Ti
content. The higher the Al and Ti content, the greater is the
possibility of cracking. The relative weld ability of several
precipitation-hardening Ni-alloys, as affected by Al and Ti
contents, is shown in Fig. 8.25.
Several techniques may be used to either control the.
magnit,ude of the residual stresses or provide a condition more
some precipitation hardening Ni-alloys.
tolerant of the stresses. Such techniques include welding the parts
in an unrestrained condition, application of a suitable PWHT,
welding the joint with an appropriate sequence, and annealing at
an intermediate stage during welding.
The recommended PWHT for precipitation-hardenable
Ni-alloys is to solution-anneal followed by precipitation-
hardening. Precipitation-hardening an as-welded joint increases
the tendency for strain-age cracking. During hardening an overall
volume contraction takes place. Welding stresses coupled with this
contraction greatly increase the likelihood of strain-age cracking.
Solution annealing after welding willrelieve high welding stresses
and decrease the possibility of strain-age cracking. However, it is
important to heat the weldment through the hardening
temperature range rapidly to avoid precipitation reactions.
One PWHT sometimes used to avoid strain-age cracking is to
overage the components. Such a treatment is designed to
precipitate the age-hardening constituent in a massive form. This
reduces the hardness and yield strength of the alloy as well as the
welding stresses that can be developed. Cracking of the weldment
during solution annealing is then less likely to occur because
 603
002 Welding Engineering and Technology j Weldability of Specific Materials

further ageing cannot take place during heating. Alloys containing I :j'"
relatively large amounts of Al and 1'i, such as Udimet 500, have
been successfully welded using this averaging technique.
-gr5
8::$ ~
~ ~
~ ~
~ ~
~ ~
~ ~
~ ~
~ ~
~ ~ ~
~ -
w ~
~ ~~
_
Welding of precipitation-hardened components should be Cj c..,
.::l <§"

avoided. Welding in this condition will result in resolutioning and .t:;

'"
overageing in the HAZ. A PWHT is then necessary to restore the ~
properties of this zone . rn
to::l

.---------
:>.
8.9. WELDABILITY OF COPPER .t"\.ND ITS ALLOYS o .0
~ :~
u
Copper has a FCC (face centred cubic) crystal structure, as do '" o
most of its alloys; that is why they have good formability and
malleability. The principal alloying elements in Cu-alloys are .AI,
III
Q,
Q,
o
{J::.:
~
o
u __
. EO ~ I ~ ~~~
M
~ ~ ~ ~ ~ 0
~ o ~ M
~ ~ ~ ~ ~ ~
~ ~ ~ ~
~
o 1;::: ~
Ni, Si, Sn, and Zn. MO!'!t commercial copper alloys are sohd .•..
E
&..

solutions and show no allotropic or crystallographic changes on .d

b.o ~
heating and cooling. Copper and its' alloys are classified into eight o
='
major groups as follows.
(i) Coppers, -
~
CIS
-C,,)
I::l 0
'1'
(ii) High copper alloys, o bbbbb
~ ~g ~x ~x ~x Sx ~x ~x ~x
... ~o
...• •....• ,...... •.....••....• ....-4

(iii) Brasses (Cu-Zn), x

(iv) Bronzes (Cu-Sn),
~
III
!§&..°I'1''1''1''?'1''1''1'
1>.£1 I ~ ~ co <0 0) ~ <Xl
o...• gX ~X X
:g
X X
'""! '""!
El
-..;;: tl <;;)
"-I
co
~
co 0') 0') ~ co 0') ~
...• ai N ex) ~ ~
(v) Copper-nickels (Cu-Ni) 10..: ~ 'lC! ~ 'lC! •••.• c:-I ••••
o :!S !"l.~ ~ ~ ~ ~
L,,":
~~
•..••..•~
00 C,,) l;\
(vi) Copper silvers (Cu-Ni-Zn), .....
o
(vii) Leaded coppers, rn
II)
(viii) Special copper alloys . ...•
.•.•
'"
III
Physical properties ofCu-alloys important to welding, brazing, Q, ~ co M<0 ~ ~ 0 0') 0') <Xl

and soldering include melting temperature range, coefficient at.

thermal expansion, and thermal and electrical conductivities.
e
~
~
~~ i
~ ~ ~ ~I ~
0')
I ~ i ...•~ ...•~ i ...•~ i i i
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ =
0~ ~I <0
~ m ~ ,......
These properties for widely used copper alloys are given in table -;(J '''::: <0 <0 ~
~ 00 ~
••••~ ~ 0') ~
m ~ ~ 00 ~ ~ 00 ~ = ~
~
...• ~
8.28. rn
:>.
.d
Many copper alloys are solid solutions, i.e. single phase'alloys, ~ /-.
while other alloys consist of two or more microstructural phases. 00
C'I
Two phase alloys harden rapidly during cold working, but ~hey
llsually have better hot working and welding characteristics then
the solid solutions of the same alloy system. Ductility usually
-
,x)
III
.Q
~
,£1
~
<,;"
0«
~ ~~~o_<
0
~ •...• ~ OkOf""'lai~Qf~~~~
t- .....
decreases and yield strength increases as the proportion of second Q.)~C: NdNoo I
§ e § •..• M
,~
'" ~ ~ 0 Q 00 ~ ~ ~
phase in the alloy increases. Q go §: ~ ~ ~§~ § ..;::p:) ..;: _' _' \C,
u 8
8.9.1. Weldability Factors .8' gj
~ 8 ~
3 00

en
~
"i~P:)P:)
~
~ -
e 8 ~~~1
J...
- ~ ~

$-I :s .- Q.) c: d
~

.-
~ c .3 8
= Q) ~ -0 ] ~ '8 ~ Q ~ r.n
The high electrical and thermal conductivities of copper and
So..
Q) co .- c.. !:l.. .- I.U ~ Q) CJJ Q)

certain high-copper alloys have a marked effect on weld ability. ~~e--ot: ~ ~ §fo§ §:§:~
Copper, AI-bronze, commercial bronze (90%), and red brass (85%) >: ~ 0 ~ '" ..c: ..c: ::;;:.- 0 0 ,- j
have high thermal conductivities. Welding heat is quickly ---:. -.r.~ 0
ICtj~ P:)
c-.i _~ ~.~ __ ~_ Co)
It? ~c.O ~[".; cO ,~~-~~
cri ~::: ~ ~I
~.

conducted int.o the bas\~ metal. This may cause lack of fusion in
 504
Welding Engineering and Technology
weldments. Preheating of these alloys will reduce welding heat
requirements for good fusion.
Many copper alloys are hardened by cold working. The
application of heat in any quantity will soften them. Mter welding,
the HAZ will be softer and weaker than the adjacent base metal.
This zone may also tend to hot crack with some coldworked alloys.
Care must be taken when welding the precipitation-hardening
eu-alloys to avoid oxidation and incomplete fusion. Whenever
possible, the components should be welded in the annealed
treatment. and then the weldment given a precipitation-hardening
condition,
Copper alloys with wide liquidus-to-solidus temperature
ranges, such as Cu-Sn and Cu-Ni, are susceptible to hot cracking
at solidification temperatures. Hot cracking in these alloys can be
minimised by,
(i) reducing restraint during welding,
(ii) preheating to tetard the cooling rate and reduce the
magnitude of the welding stresses, and
(iii) reducing the size of the root opening and increasing the
size of the root pass.
Certain elements such as zinc, cadmium and phosphorous
have low boiling points. Vaporisation of these elements during
welding will result in porosity. When welding copper alloys
containing these elements, porosity can be minimised by welding
at high speeds and by the use of filler metal low in these elements.
Surface oxides can cause a serious problem during welding
and soldering of Cu and its alloys. The oxides on AI-bronze, Be-
copper, Cr-copper, and Si-bronze are difficult to remove. The
surfaces to be joined must be clean, and special fluxing or shielding
methods must be used to prevent the film from reforming during
the joining operation.
8.9.2. Welding of Copper and Cu-Alloys
Copper and most of its alloys can be joined by arc welding,
oxyacetylene welding, flash welding, brazing and soldering.
8.9.2.1. Arc Welding
,j.
The processes that use inert gas shielding are generally
preferred although SMAW works well for many non-critical
applications. The steps involved in general include fixturing,
surface preparation, preheating, filler metal selection and joint
design.
1 Weldability of Specific Materials 60S
Fixturing : The thermal coefficients of expansion of copper
and its alloys are about 1.5 times that of steeL This means that
distortion will be greater with Cu-alloys unless appropriate
measures are taken to control it. Suitable clamping fixtures can
be used to position and restrain the components for welding. With
thick sections, frequent tack welds may be used to align the joint
for welding. The ends of the tack welds should be tapered to
ensure good fusion with the first weld beads.
With multi-run welds, the root bead should be rather large to
avoid .cracking. Fixturing and welding procedures must be
designed to limit restraint with Cu-alloys that are hot short and
likely to crack when highly restrained.
Joint backing strips or rings may be helpful when welding
copper and high-copper alloys. Control of drop-thru can be a
problem with these very fluid weld metals. The backing strip
should be made of the same alloys as the base metal or a very
similar one.
Surface Preparation: To make a sound weld, the joint faces
and adjacent surfaces should be clean. Oil, grease, dirt, paint,
oxides, and mill scale should be removed by degreasing followed
by suitable chemical or mechanical cleaning. Wire brushing is not
a suitable cleaning method for copper alloys that develop a
tenacious surface oxide, such as AI-bronzes. These alloys should
be cleaned by appropriate chemical or abrasive methods of
cleaning.
Preheating: The relatively high thermal conductivities of
copper and most of its alloys result in the rapid conduction of heat
from the weld joint to the surrounding base metal. Heat loss from
the weld zone will be greater with thick sections at low
temperatures. It can be reduced by preheating the base metal.
The optimum preheat temperature depends upon the welding
process, the alloy being welded, and the metal thickness or mass;
however the usual range js 55° to 445°C. Thin sections or high
energy welding processes, such as GMAW and EBW, generally
require less preheat then do thick sections or low energy welding
i
i processes, such as GTAWand oxyacetylene welding, as is shown
i
in Fig. 8.26. Copper alloys that have low thermal conductivities,
such as AI-bronze and Cu-Ni alloys, should not be preheated.
High V'leldingcurrent should not be used to compensate for
lack of preheat. It will overheat the weld metal and cause a rpugh
bead surface, undercut, porosity, and oxide inclusions.
Filler Metals: Covered and bare welding electrodes and bare
rods are available for welding copper and its aHoys to themselves
 (\06 Weldability of Specific Materials 607
Welding Engineering and Technology
J
·c (Approx.l Covered electrodes for shielded metal arc welding of Cu-alloys
1400 are available in standard sizes ranging from 2.5 mm to 5 mm in
diameter. Other sizes are also available in certain electrode
1200 classifications. The SMAWprocess can be used to weld a range of
thicknesses of various copper alloys with the appropriate electrode
•
u..
- 1000 ./ // -1565 size and alloy composition.
QI
•.. Shielding Gases: Argon, helium, or mixtures of the two are
::;)
470
1? 800 used for shielding with GTAW,PAW,andGMAW. In general, Ar is
a.
QI
used when manually welding material less t.han 3.3 mm thick.
E ;370
<u 600 However, He or a He-Ar mixture may be needed for manual
.0- 27S welding thin sections of high conductivity copper to take
~
Cl
400
':ldvantage of the high arc energy.
'-OJ
176 Welding Position: Welding is best done in the downhand
~
200 position with all of the arc welding processes. GTAWor SMAWis
80
preferred for welding in position other than downhand, parti-
o -1 e
cularly in the overhead position. GMAWwith pulsed power and
o 3 6 9 24 small diameter electrodes may be suitable for vertical and
Plate thickness ,mm overhead positions with some copper alloys.
Fig. 8.26. Effects of process, shielding gas and base metal thickness Welding Technique: Welding technique depends upon the
on preheat requirements for welding copper. process employed.
and other metals. AWS classification for such filler wires and For GTAWeither forehand or backhand welding may be used.
electrodes are as listed in table 8.29.
Forehand welding is preferred for all welding positions. It can

i
Table 8.29. Filler Metals for Fusion Welding Copper Alloys provi.de a more uniform, smaller bead than with backhand

I
Covered
-==--A",~S_ Class![i:!:~]-
---.---------------
Bare Wire-l
co-;;:r~~-~~;'(l;;;~TBaSemetal applications
_._._-~
---
...
-----.--r--'.--------------
. Coppers -
welding. However, a larger number of beads may be required to
fill the joint.
The joint should be filled with one or more stringer beads or
I<:Cu ERCu Copper narrow weave beads. Wide oscillation of the arc should be avoided
ECuSi
(!lect-rode ERCuSi-A 1Silicon bronze II Silicon bronzes. brasses because it intermittently exposes each edge of the bead to the
ECUSll--C ERCuSn-A Phosphor bronze Phosphor bronzes, brasses atmosphere and consequent. oxidation. The first bead should
ECusn.-A}
ECuNi ERCuNi
penetrate to the root of th(~joint and be fairly thick to provide time
Copper-niche: Copper nickel alloys
ECuAI-A2
for deoxidation ofthe weld metal and to avoid cracking of the bead.
ERCuAl-A2 Aluminum bronze Alunlinum bronzes, brasses,
silicon bronzes, manganese GTAWis best suited for joining sections up to 3 mm thick. It
bronzes can also be used advantageously for thicker sections when welding
ECuAI-B I ERCuAI-A3 Aluminum bronze Aluminum bronzes in a position other than dO'Nnhand.Typical joint designs for GTAW
of copper and its alloys are as shown in Fig. 8.27.
ECuMnNiAI I ERCuMnNiAI
ECuNiAl ERCuNiAl --- Nickel-aluminum bronzes
Manganese-nickel-alwninum In GMAW the filler metal should be deposited in stringer
beads or narrow weave beads using spray mode of metal transfer.
Wide weaving of the electrode may result in oxidation at the edges

______ J' RCuZn-B

RBCuZn-A__
RCuZn:g,
I
I
JJ~~~~~.i.r:.iL
bronzes
Low filming brass Brasses. manganese bronzes
I Brasses. manganese
NllVSI brass brass 1~rasses, copper bronze~
of the bead. Suggested joint designs for GMAWof copper and its
alloys are shown in Fig. 8.28.
608
Welding Engineering and Technology Weldability of Specific Materials 609

to counter balance the low heat input, particularly with alloys

X 1ST
TO; ~ I~~~~d:"m -1_
_ ~
-r;-S--~, ,J~
T
r-075T
:\.."1 max
<~
having high thermal conductivities or with thick sections.
The oxyacetylene flame offers little protection of the weld zone
( A I Edge -flange ( B I Square-groove from the atmosphere. Except for 02-free copper a welding flux is
80-90· generally required to exclude air from the weld metal at elevated
temperatures.
1- ~70·-80 7 31 -r- '< '7 2.5-31 This process, however, should be used only for small
noncritical applications including repair welding, when arc
-T--->~-----.,~
I e I Single-V- groove
~T --j
J_J(>3-35 _~r€
(D) Double-V-groove
~ , welding equipment is not available.
8.9.2.3. Resistance Welding
Copper and high-copper alloys are very difficult to join by
resistance spot and seam welding because of their high electrical
,L5'~~" ;'1 ~t§ff'15' and thermal conductivities. Very high current densities would be

T;;RA~lHI 'L ~D
I E I Single- U-groove
1,5-'
( F) lJouble- U-groove
15-J
required for spot and seam welding and this would cause electrode
overheating, sticking to the work, and rapid deterioration.
Therefore joining of these materials by spot and seam welding is
Dimensions in mrn not recommended. However they can be joined by flash welding.
Fig. 8.27. Some typical joint designs for GTAW of copper and its alloys.
Flash welding techniques produce very good results in almost
all applications and on all copper alloys. Rapid upsetting at
minimum pressure is necessary as soon as the abutting faces are
L _T
---11-- 3
I

~-~~~3
<aO"-90"> molten because of the relatively low melting temperature and
narrow plastic range of cu-alloys. Low pressure is usually applied
f" ( A I Square-g
I ,"--I ----=>'.
roove
1 ,.s-~
L
( B I Single-V-groove
to the joint before the flashing current is initiated so that platen
motion will begin immediately after flashing starts. Termination
of flashing current is rather critical because premature termi-
nation of current will result in lack of fusion at the weld interface
,<8o'-90~ -L L ,\30·-; while excessive flashing will overheat the metal and result in
improper upsetting.

~C3T
L3 ~~3-3'S ,- ~~ ~~
., C' •
Copper alloys with relatively low electrical and thermal
conductivities may be spot, seam, and flash butt welded .
8.9.3. Postweld Heat Treatment of Copper Weldments
Ie) Dou~le-V-groove I 0 I Double- U-groove
Dimensions in mm
The necessity for heat treating after welding depends upon the
base metal composition and the service conditions of the
Fig. 8.28. Some typical joint designs for GMAW of copper and its alloys. weldment. PWHT is usually required if the base metal can be
8.9.2.2. Oxyacetylene Welding strengthened by a heat treatment or if the service environment
Copper and many of its alloys can be welded by oxyacetylene can cause stress-corrosion cracking. Alloys that respond to
welding. The relatively low heat input of the oxyacetylene flame precipitation hardening include some high coppers, some Cu-AI
makes welding rather slow compared to arc welding. Higher alloys, and Cu-Ni castings containing beryllium. If these alloys are
not heat treated, the hardness in the weld area will vary as a
preheat temperatures or an auxiliary heat source may be required result of .ageing or overageing caused by the welding heat.
 610 Welding Engineering and Technology 611
Weldability of Specific Materials
Alloys that are particularly susceptible to stress-corrosion Table 8.31. Annealing Te.mperature Ranges for Copper
cracking are high zinc brasses, Mn-bronzes, and Ni-silvers. and So.me of its Alloys
Stresses induced during forming or welding of those alloys can
. lead to premature failure in certain
l.
Stress-relieving and annealing are the
corrosive environments.
9. postweld heat treatments
4.
13.
7.
3.
5.
6.
12.
14.
--- Aluminium
Red
Yellow
Naval
Muntz
Silicon
Nickel
brass
AIumini brass
brass
metal
Admiralty
Beryllium
S.No.Annealing
Phosphor
Alloy bronze
Silver
urn bronze
brass
bronze
copper
Copper-nickel,
I deoxidised
10%I copper
temperature
designation
roC)
30% range 595-815
595-760
480-675
775-940
595-900
650-815
425-735
425-595
480-705
425-705
370-650

2.
8.
11.
10.
usually employed for copper weldments.
8.9.3.1. Stress Relieving
Stress relieving is accomplished by heating the weldment to
some temperature well below the recrystallisation temperature of
the base metal for a minimum period of one hour. The temperature
and time should be designed to reduce welding stresses to
relatively low values. Heating time must be adequate for the
entire weldment to reach the intended temperature.
Typical stress relieving temperatures for some copper alloys
are given in table 8.30.
Table 8.30. Typical Stress Relieving Te.mperatures for
Weld.ments of So.me Copper Alloys
I I brass
Red Aluminium
Naval
S. No. Silicon
Ol'assalloys
Admiralty
Cu-Ni bronze
bronze
brass
Alloy 540
345
260
290
Suggested stress relieving
designation temp.
(OC)
1

8.9.4. Brazing of Copper and its Alloys

8.9.3.2. Annealing
Annealing and homogenisation treatments are used only for
weldments of harden able copper alloys to produce a met'al1urgical
structure that will respond to heat treatment satisfactorily.
Annealing is carried out at temperatures considerably higher than
those used for stress relieving as shown in table 8.31.
Stress relaxation proceeds rapidly at the annealing tempera-
ture. Extended annealing times or annealing at the top of the
temperature range can cause undesirable grain growth. Excessive
grai~ growth usually reduces tensile strength and can cause other
u:p.desirable metallurgical effects. - for brazing. However, certain precautions are required with
Brazing is an excellent process for joining copper and its
alloys. Surface oxides are easily fluxed during brazing except for
those on AI-bronzes that contain more than 8% AI. Special
techniques are required to braze those AI-bronzes.
Ail the common brazing processes can be used for brazing
copper except for special cases, such as resistance or induction
brazing of copper and its alloys that have high electrical
conductivities.
Both lap and butt joints may be used for brazements. The joint
clearance must provide for capillary flow of the selected brazing
filler metal throughout the joint at brazing temperature. In the
case of electrical conductors the electrical conductivity of the joint
must be considered since the brazing filler metals have low
conductivity compared to copper.
All of the silver (BAg), copper-phosphorous (BCuP), gold
(BAu), and copper-zinc (RBCuZn) brazing filler metals are suitable
612 Welding Engineering and Technology Weldability of Specific Materials 613

specific base metals to avoid embrittlement, cracking or excessive 8.10. WEI.DABILITY OF ALUMINIUM AND ITS ALLOYS
alloying with the filler metal. Aluminium is known for its light weight (density = 2.7 g/cm:$)
The various BAg filler metal may be used with all copper and good corrosion resistance to air, water, oils, and many
alloys. The BAu filler metals are best suited for electronic chemicals. It has thermal and electrical conductivities approxi-
applications where the vapour pressure of the brazing filler metal mately four times greater than steel. This necessitates higher heat
is important. inputs when fusion welding aluminium. Thick sections may need
The BCuP filler metals can be used with most copper alloys, preheating to reduce the heat input needed for welding.
including some Cu-Ni alloys. However, BCuP filler metals are not This metal is highly reflective to radiant energy, including
recommended for brazing Be-copper because the joints will be visible light. Consequently, unlike steel it does not assume d.:trk
porous and low in strength. red colour at 650°C. The lack of colour change makes it difficult
The RB CuZn filler metals may be used to join the coppers, to judge when the metal is approaching the molten state during
Cu-Ni, Cu-Si, and Cu-Sn alloys but they are not recommended for welding. High reflectivity of aluminium also causes serious
torch brazing the AI-bronzes because ofhigh brazing temperatures problems in laser welding this metal.
required for the purpose. Aluminium is non-magnetic so there is no problem of arc blow
in arc welding with direct current. The linear coefficient ofthermal
AWS fluxes of the types 3A and 3B are suitable for use with expansion of aluminium is about twice while melting point is half
B Ag and B CuP filler metals for brazing all copper alloys except that of steel thus the total expansion during welding aluminium
the Al-bronzes. A more reactive Type 4 flux is needed for these and steel components is about the same.
bronzes. Type 3B or 5 flux is required with RB CuZn filler metals Pure aluminium has F.C.C. structure and is very ductile. It
because of their high brazing temperatures. has very low strength particularly in annealed condition. It is,
8.9.5. Soldering of Copper and its Alloys therefore, much less used in its pure form but its alloys are
Copper and most of its alloys are readily soldered with lead extensively used in industrial and general structures.
solders. Most copper alloys are easily fluxed, except for Be-copper 8.10.1. Aluminium Alloys
and Al and Si bronzes. Special fluxes are required with these alloys Aluminium is alloyed principally with copper, magnesium,
to remove the refractory oxides that forms on their surfaces. manganese, silicon and zinc. Small additions of chromium, iron,
Soldering is primarily used for electrical connections, plum- nickel and titanium are sometimes made to specific alloy systems
bing, automotive radiators, and other room temperature to obtain desired properties and grain refinement.
applications. Joint strengths are much lower than those of brazed Table 8.32. Designations for Wrought Aluminium Groups
or welded joints. -~--r---------------'-------'---------'--"----
S. No. Alloy group Series designation
8.9.6. Other Pro'cesses L Aluminium 99.00% min. purity 1xxx
Copper and many copper 'alloys can also be joined by the 2. Aluminium-copper 2xxx
following processes for special applications. 3. Aluminium-manganese 3xxx
1. High frequency resistance welding, 4. Aluminium-silicon 4xxx
2. Friction welding, 5. Aluminium-magnesium 5xxx
3. Cold welding, 6. Aluminium-magnesium-silicon 6xxx
4. DitTusionbonding, 7. Aluminium-zinc 7xxx
5. Hot pressure welding, 8. Others 8xxx
6. Ultrasonic welding,
7. Electron beam welding, A system of four-digit numerical designations, devised by
8. Laser welding. Aluminium Association of USA, is used to identify the various
614 Welding Engineering and Technology
Weldability of Specific Materials 615
wrought Al-alloys. The first digit indicates the alloys group, as
shown in table 8.32. The second digit indicates a modification of 2xxx Series Al-alloys provide extremely high strength when
the original alloy or the impurity limit in the case of unalloyed properly heat treated. These alloys do not produce good corrosion
aluminium. The third and fourth digits identify the alloy or resistance and are, therefore, often clad with pure aluminium or
indicate the aluminium purity. special-alloy aluminium. These alloys are used in the aircraft
industry ..
A temper designation system is used to indicate the condition
of a product. It is based on the sequence of basic treatments used 3xxx series alloys are non heat-treatable, have modern'
to produce the desired mechanical properties. This designation strength and are easily worked.
follows the alloy designation. The basic temper designations are Most of the 4xxx series alloys are non-heat-treatable ana are
shown in table 8.33. Subsequent divisions of the basic letter used for making brazing alloys and welding electrodes.
tempers are indicated by one or more digits following the letter. The Al-Mg-Mnalloys of 5xxx series are widely used for welded
These digits designate a specific sequence of basic treatments. structures, particularly where corrosion is a problem, as in marine
Table 8.33. Basic Temper Designations for Alwninium Alloys applications. Weld ductility of these alloys is excellent, and they
are well suited for applications subjected to impact and shock
S. Temper Treatment / Condition loading.
No. designation
1. F As fabricated The 6xxx series Al-Si-Mg alloys possess medium strength and
2. o Annealed
good corrosion resistance. Alloys 6061 and 6063 are widely used
for welded structures.
3. HI Strain-hardened only
Zinc is the major alloying element of 7xxx series alloys.
4. H2 Strain-hardened and partially annealed Outstanding in this group is 7075, one of the strongest Al-alloys
5. H3 Strain-hardened and t,hermally stabilised available, used in airframe structure and highly stressed parts.
6. W Solution heat-treated Though not easily welded by GMAW or GTAW processes, it \
7. Tl Cooled from an elevated temperature shaping process responds well to electrical resistance welding.
and naturally aged Alloy 7039 is somewhat less strong than 7075 alloy but it is
8. T2 Cooled from an elevated temperature shaping process, readily welded by any of the fusion welding processes. It is used
cold worked, and naturally aged for welded pressure vessels for use at -130°C to -185°C. Alloy7039
9. T3 Solution heat-treated, cold worked, and naturally aged is also used for armour plate under military specifications. Alloys
10. T4 Solution heat-treated and naturally aged in the 7xxx series are also used for portable military bridges.
11. T5 Cooled from an elevated temperature shaping process 8.10.2. Welding of Aluminium and its Alloys
and then artificially aged Aluminium and its alloys can be joined by most fusion and
12. T6 Solution heat-treated and then artificially aged solid state welding processes as well as by brazing and soldering.
13. T7 Solution heat-treated and stabilized Fusion welding is commonly done by GMAW, GTAW, and
14. T8 Solution heat-treated, cold worked, and then artificially resistance spot and seam welding. Plasma and electron beam
ll:ged welding are used in special applications. SMAWand oxy-fuel gas
15. T9 Solution heat-treated, artificially aged, and then cold
worked .
welding is not
quality are applied
notbe commercially.
essential
used forfor the intended service.
highSubmerged a. c
16. T10 Cooled from an elevated temperature shaping process,
cold worked, and then artificially aged. welding may applications where strength aT'd
Solid-state welding processes suitable for joining most
8.10.1.1. Applications Al-alloys are friction, diffusion, explosion, high frequency and cold
lxxx series alloys are primarily used in the electrical and welding, These proc~sses
other metals, particularlyare also
thosesuitable, for j.oiningAl
combmatIons, that toare
fertailD
ndt
chemical industries. metallurgically compatible. Brazing, soldering, and adhesiv~ '
6L6 617
Welding Engineering and Technology Weldability of Specific Materials

bonding are also used for joining Al components. Adhesive bonding 8.10.2.1. Filler Metals for Aluminium and its Alloys
in combination with resistance spot welding, called Weld Bonding, Aluminium welding rods and bare electrodes are generally
produces joints with better strength properties than those used with GTAW,GMAW;and oxy-fuel gas welding processes. In
produced by either process individually. Table 8.34 gives the special cases, these filler metals may be used with other fusion
practical thickness or cross-sectional area that can be joined by welding processes, such as electron beam or plasma arc welding.
various processes. A limited number of Al covered electrodes are also available for
Table 8.34. Practical Thicknesses or Area Ranges for Various SMAW.
Joining Processes Applicable to Aluminium and its Alloys
Resistance
Ultrasonic
Electron
I Gas
Cold metal
Brazing No
No
No
beam 25
150
0.40
Maximum
welding
seam
arc
tungsten
welding-lap
spot limit
·i.5
limit
4.5
(305)limit
3(5)
welding
welding
welding
arc welding
welding
joint --
Thickness, mm0.15
30.5
0.25
Foil
0.5
1.25
or Area,
0.015 mm2
Selection'of an appropriate fillei' metal for a specific appli-
cation depends upon a number of considerations including the
-'IT--
1
Stud
I\ Cold
welding Process Minimum
"
Flash welding-butt
welding joint following.
(i) Base metal composition,
(ii) Joint design,
(iii) Dilution,
(iv) Cracking tendencies,
(v) Strength and ductility requirements,
(vi) Corrosion in service,
(vii) Appearance.
Table 8.35 gives suggested filler metals to satisfy one or more
specific requirements of welded joints in commonly welded Al-
alloys, one filler metal meets the needs for two or more require-
ments.
Table 8.35. Suggested Filler Metals for Commonly Welded
Aluminium Alloys to Provide Specific Requirements
Appearance is often of considerable importance in selecting a
joining method. GMAW,GTAW,PAW,and EBW can provide the
--\--
~', I Basel --.-.----------.-----------------
__ T .~.::!~I!.~end!:9:.jill~~ fo~._ .. , .__
best as-welded bead surface. Joints with excellent appearance and o. \ meta High
strength Good iI ma.tch
\ ductility Colour
after \! Saltwater:
corrosion \ Least
cracking
little or no finishing requirements can be made by brazing or
adhesive bonding. Resistance spot and seam welding and ultra-
sonic welding may mar sheet surfaces. Cold welding requires large
metal deformations to accomplish metallurgical bonding. The
process is not suitable when a smooth surface is needed. The flash
may be removed from flash or stud welds to improve appearance.
~-t:~-
4.
-:~:n::1
3. \ ":~"gl-"~~:'l!':::-
3003 I
5052 \
4043
5356
I
\
1100
5654
i 1100
53,'56
I

\
1100
5554
i
i
4043
5356
5. 508.'3 5183 \ 5356 5183 5183 i 5356
The two welding processes mostly used for Al are GMAWand
GTAW.Both are recommended for most AI-alloys generally GTAW 6.
7. 5086
5454 5356 5356
5554 5356
5554 5356
5554 \ 5356
for thickness up to 3 mm, and GMAW for heavier materials, 8. 5456 5556 5.'356 5556 5556 \I 5356
although it is not a very hard and fast rule. ,

Most often filler metal is used·in conjunction with GTAWwhile 9. i 0061

10. 6063 \. 5356
5356 \! 5356
5356 5654
5356 4043
4043 Ii 4043
4043
it is an intrinsic part of GMAW process. Thus selection of filler 11.
I
7005 i
i
5556 5356
-
5356 5356 I 5356
metal is a vital step in successful welding of Al with these : I

12. I 7039 I 5550 : 5:356 5.'356 5:356 I 5356

processes. __ --1- __.__ ._.__' -1._
 (lIS
'0-c~0-
Welding Engineering and Technology
I Weldability of Specific Materials '"
'"..,
•..
ol
..c:
<I.>
(j)
'"
<I.>
619

!\
::E
'<3
IE
.., '"
<I.>

(j)"O
"'" c "g~
.~
Apart from selection of filler metal for the welding process .;;; ol ~ .~ "'"
o
(GTAW,
them. GMAW)
They there are other steps also which are common to
include: o bD ol o
';;j
o bD
. ~.9
u"O '"
. "'
'"

"=~rfo
....
<I.> «I \

(i) Cleaning of base metal, .,

""

c ._- cu

Ui) Weld backing, .~ -.c-;~~-;0~..e.~~ .•..

C<IJ e- :E:!3
s:: .S .<::

..,
<I.>

~
'"

g)
c;: ~Q) .5 I '-0
~
~
-
u>
U _ <I.>

.£Q QtIS
tIS ::!
(iii) Preheating, and
f
.•..
:5
E!
cC O
.='•..
E! o)
c... '0 ~ '0;::: ec.. o
'-..c:8
E
~
<
,g
~00~
<IJ
E!
100
E!
<IJ _ C
«I 0
_ ol - 0 <I.>
<;;.
> 'm ol
>.E ._
0
••• bll
~.;:~ > ~
(iv) Tack Welding. ~
c:) o o '"0 o 0
e.E S 8 :;l S .s"O S '.a

These are discussed briefly in the following sections. ~.s <1.>_

O::t:::..c
•••
0::
'" .5 0
'" <'l
(j)

0::
ol
8
8.10.2.2. Cleaning of Base Metal '" 0
<I.> -
'" <Ii 0 c:: •• 0- 0
._
0'<:: "'"
o
"0
0'<::
<I.>
- <I.>

o'-' 0- ~~ <1.>'<::
'"
The surfaces to be joined must be clean to obtain good fusion ii! .S
c 0 ~ ~ ~ •..•... .-o 0~ - o
ii:! .-
0
-
'" <I.>

between the filler metal and the base metal. This means that they ~ .-
S.<::
<I.>
02
<0 ol o~i:-
0::
. .<::
._ c
ci ~
<Ii
(j)
::!
M~O .S 0 "0
must be free of relatively thick oxide, moisture, greases, oils, ""t:l
"" l/'J ••• d, ~
••.• "0 •..•:9 Sl! o Sii:! ';;j
o
paints, or any other substance. Many contaminants break down
r...:
M-
at elevated temperatures and produce H2' which causes porosity
£ .~ ~0
c'O
o u
'w
.... t:l -
<E8.21
~
.... d ~-
I •.•
C'l
-
••• ol
-<I.> 1§>-
10 ol a·
S
::l
a
in fusion welds. Degreasing may be done by wiping, spraying, or
a1:g
S ~82>'1' 8 '"
-:
'-f '" <I.> Q)Q)b
S "'.<::
<E ~ .s ~ ... 8
c.."'" >, .s-:=: ~ ::l
dipping in a suitable, safe solvent or by steam cleaning. ~.S ~ S .9
..-0:: .§ ~.s CleO o 8 ~ <a
bll
Surface oxides on aluminium can be removed by action of the ~ "0 o
welding arc as welding progresses. However, it is best to remove .;: ol
~
""

S
the oxide from the surfaces by appropriate chemical or mechanical
methods. Chemical removal is accomplished with alkaline, acid,
t:l
e
...
'"
-
Ql
<I.>

'"
'"
-
Ql
(j)
'"
'"
~~ ol
-
«1-
'a Ql
Ql
.2l
'"
'"
<I.>
•..
•..
.E
or proprietary ~ '" "]] '" 02 '"
detailed in tablecleaning
8.36. solutions. Typical chemical treatments are '.§
<I.>

"a
..,
'" ..9:l
c S '" ..9:l
o
<I.>

::0
~ :!3 "Zal ol

Mechanical oxide removal is usually satisfactory if performed

8 :5
CI.l :@
'3
CI.l
c 0
<:.:::
:5
CI.l ~
10

properly, but it is not generally as consistent as chemical methods. oS

<:.
::!
'"
"0
o
In the most popular method of wire brushing, stainless steel Ec..J
"<l'

~ ot- "<l'
C'l
I
C'1
10 -E
brushes with 0.25 to 0.40 mm diameter bristles are employed. To '" 0 00
...•
00
•....
ot-
'1' <.J) <I.>

E
~ "0
be effective, the brush must be kept clean and free of contami- e::: q
'"
nants.
the It should surface.
aluminium be used with light pressure to avoid burnishing •.. '"
<I.> <a
~ o
<Ii!
"'"E (j)
~ "0 .., 'C
.9 10 "0- «I ~ ;.=:
10- ~
b.O.-:: ~ b.O.-::
.2l
Cleaning of AI components is best done prior to fit-up and •.•
-'0 .2 c
.2l •...• 0- ol
~ 2«1-
6h
"0 .-
«I
... I ~ '" ~~;;~ S
fixturing. Fitted joints may be very difficult to clean. Often ~ o ...,
c .••. I I ~1O I I »
~ .... -
<I.>

-
cd .~ ~

solvents, or contaminants may be trapped in or near the joint and

""
<J
CI.l"O (j)
<J 0"'; 0M'" :a
'x
\ 0£~~
e..
cause internal discontinuities. c :§
:::

£ u~
<Ii! c:: <I.>

8 o10
lj'<
~ ..0.=
u <I.>
..,
10
» "0
8 o
Co)
M~
<I.>
'C
c..
8.10.2.3. Weld Backing .<::
--- o
•..
--+---.----
<1.>"0
c..
»
-0 '"
When full penetration weld is to be made from one side only, .- ~0 r::

use of some type of weld backing may be imperative to control the

amount of reinforcement and the shape of the root surface. Two :::
-
0_
>0:

...
"O<E
»-
u
'8
<J

'8
o
•..
ol
S
....
(j)
ol"O
.S:
o .<::

types of backing may be used viz., permanent and temporary. :; "'"

.<::

"8
ol
~
!::
'-t
.~
(j).21
•.. '"
'0«I •.. <1.>._

Permanent backings are left in place after welding. These may ~ a '"
u
u
ol o
.<::
.<::-

u :3 '';;:; C,) c.. Eo<

.. 0 ~

be in the form of strips or ~xtruded sections of aluminium. Such ._ tfJ ._ '"

backings may, however, sometimes cause lack of fusion at the root

'5
Z
-0
o
='
«I
::!. ..c
» Z._ .b
:;
..z
.<::
o -
~ ';3
0- ::l
CI.l
I~CI.l Ie.. Z ~
 621
tilt) Welding Engineering and Technology Weldability of Specific Materials

of the weld if the backing material is not properly cleaned or if the Joints that require welding from both sides should be back
fit-up is poor. Permanent backing should not be used in corrosive gouged to sound metal before welding the other side. This lessens
service unless all edges of the backing strip are completely sealed. the possibility of lack of fusion or porosity, or both, at the root of
Otherwise, crevice corrosion may take place at the root of the weld. the weld.
Temporary backing is generally in the form of a bar made of The above described four factors are common to both GTAW
CQPper,carbon-steel, or stainless steel. This bar is normally a part and GMAW but the welding techniques differ. This aspect of
of the fixturing. Care should be t.aken during welding to avoid GTAW and GMAW processes for AI and its alloys is discussed
meJting the backing bar and contaminating the AI weld metal. briefly in the following sections.
A groove may be machined into the backing bar to decrease
the chilling effect at the root of the joint for good penetration. The 1- r-1l- r-1l- 1_
grO(lveshould be designed to permit good fusion at the root of the ~---II Lt
weld and to provide the root reinforcement of desired height and
shupe. Groove dimensions usually range from 4 mm to 13 mm wide
and 0.25 mm to 2.25 mm deep. The actual dimensions will depend
T.t~r-----
backing (A I
-tf;
upon the thickness of the section being welded, the joint geometry,
and the welding process.
Backing bars may be used to remove heat from welds in thin
sections. This may permit welding with sufficient current for a
stable arc. Water-cooled Cu-bars provide the greatest chilling
effect; stainless steel bars, the least. However, excess chilling may
result in lack of fusion at the root of the weld.
8.10.2.4. Preheating
Preheat is not normally necessary when fusion welding alumi-
.< 90' -..,'/ -L
nium because it can increase the width of the HAZ and reduce the
mechanical properties of the welded joints in some alloys.
However, thick sections of AI may need preheating. In such cases
the preheat temperature should not exceed 165°C and the time of
application should be held to a minimum. AI-alloys that contain
DE¥: r -~ ~
lEI
1·5-3 J Temporary
backing
--1 13
IF)
TI
I;
Mg in the range of about 3 to 5.5% should not be preheated above
120CC and the interpass temperature should not exceed 165°C.
Exposure of these alloys to temperatures in the range of
120" to 205°C for relatively short times can decrease their
resistance to stress corrosion cracking.
8.10.2.5. Tack Welding and Gouging
-LL
-1 ' r-
-r=l--lB L
I I
'?hZ E1-L
--1\Ba;:i"gt-1--3.--J--B~9
_t lS V t

tlmax10) strip Ltlmax.101 strip

Tack welds are often used to hold component parts in proper IG I • (H)
relat-ive position or to attach tab-in and tab-out at the ends of the
joints. Such welds should penetrate well and have sufficient length \\60
to provide the required strength.
Tack welds may lead to porosity in welds and incomplete 1 2
max Temporary
backing
fusion, if they are improperly made. It is recommended that tack t
wet-ds be removed as welding progresses. If this is not done, the
ends of each tack weld should be tapered gradually to blend
smoothly with the joint faces.
Hote: DimenSiOns
Qre ,n mm
co I 1 ) I J I I K I

Fig. 8.29.Typical joint design for arc welding aluminium and its alloys.

.!-
622
Welding Engineering and Technology
I Weldability of Specific Materials 623

Some other special joint designs, such as those shown in Fig.

8.10.3. GTAW of Aluminium and its Alloys
The joint designs used for welding aluminium are, in general,
similar to those used for steel. However, smaller root openings and
larger groove angles are normally used because AI is more fluid
and welding gun nozzles are larger. Typical weld joint designs used
for joining AI by arc welding are shown in Fig. 8.29.
V-groove joint designs are adequate for butt joints that are
accessible from both sides. As a rule, a 60° groove angle is the
minim urn practical size for section thicknesses greater than 3 mm.
Thick sections may require even larger groove angles, such as 750
or 90°, depending upon the welding process. With thicker plates,
J-grooves are preferred to V-grooves to minimise the amount of
deposited metal.
Ala I,- r15 R
Process
DImensions, mm
GTAW
~ \t1 tfr-----1· A ~
GMAW
8.31, may be employed when welding horizontally to minimise
porosity caused by entrapment of hydrogen.
8.10.3.1. Welding Procedure
GTAW of Al can be achieved using anyone of the following
power supplies :
(i) Conventional AC (50 Hz),
(ii) DC with electrode negative (dcen),
(ii~) DC with electrode positive (dcep),
(iv) Square wave AC (SWAC).
Surface cleaning of the AI takes place when the electrode is
+ve, but penetration is poor. Conversely, penetration is good with
a negative electrode, but there is no cleaning action. A.C. can
provide good cleaning action and acceptable penetration, parti-
cularly SWAC of variable frequency and pulse width. Direct
A B
current can provide good penetration (DCEN) or good cleaning
(DCEP), but not both conditions simultaneously.
5 '·5-3
Electrode Selection: The choice of tungsten electrode

Fig. 8.30. A special joint design for welding from one

,·5 '·5 depends upon the type of welding current selected for the
application.
Pure tungsten (EWP) and Zirconated tungsten (EWZr)
side only for a full penetration weld. electrodes are best suited for conventional AC wherein they keep
A special weld joint design used for Al is shown in Fig. 8.30. hemispherical shape; thoriated tungsten electrode (EW Th-3) may
It· is recommended for GTAW or GMAW when only one side is also be used for this purpose.
accessible and a smooth root surface is required. It can be used EW Th-l or EW Th-2 thoriated tungsten electrodes are
with section thicknesses over 3 mm and in all welding positions. suitable for DC or SWAC. Both have higher emissivity, better
However, this design has a large groove area that requires a current carrying capacity, and longer life than EWP electrodes.
relatively large amount of filler metal to fill the joint. Distortion Consequently, arc starting is easier and arc is more stable.
may be greater than with conventional joint designs. Its principal Shielding Gas: Ar is the most commonly used shielding gas,
application is for circumferential joints in aluminium pipe. particularly for manual welding. It pIovides better arc starting
characteristics and improved cleaning action with alternating
current than does helium.

r'~
1m. AJ
Helium is used primarily for machine welding with DCEN
3·5 I'
3.5maxstjfJiP 60 •. 1·5max power. It permits welding at higher travel speed or with greater
Backing
max ..L I 60 mln ~ ~
penetration than does argon.

U
He-Ar mixtures are sometimes used to take advantage of the
higher heat inputs with He while maintaining the favourable arc

Welded from one side

W to15 to15-
0-1'5Tf1~ ~--=-r0-"5 characteristics of argon. A mixture of 75He-25Ar will permit
higher travel speeds with AC power. Cleaning action is still
Dimensions in mrn
Welded from one or
both sides Welded from both sides acceptable. A mixture of 90% He-lO% Ar will provide better arc
starting characteristics with DC power than does pure helium.
Fig. 8.31. Some special weld joint designs for arc
welding aluminium in horizontal position.
---------------------------------.----
624 Welding Engineering and Technology Weldability of Specific Materials 625

(l) Conventional AC Power
When AC is used in conjunction with shielding of Ar or an
Ar-He mixture, the surface oxide is removed by the arc action.
However, this cleaning action may not be satisfactory when the'
mixture contains 90% or more helium, and preweld cleaning is
usually necessary. Pure He shielding is seldom used with AC
because the arc characteristics are poor.
Proper gas shielding and arc cleaning actions are indicated by
a bright weld bead with silvery border on each side. An oxidised
weld bead may be a result of an unstable arc, low welding current,
poor gas shielding, or excessive arc length.
(ii) Direct Current, Electrode Negative Power
DCEN power for welding Al has distinct advantages compared
to AC power in that short arc length can be made and with thin
sections DCEN permits much higher travel speed than does AC.
The surface appearance of a weld made with DCEN will be
dull rather than bright. This is due to the formation of thin oxide
fihn because the cleaning action of the rEverse polarity (i.e.
<' Y 60· to
90·
electrode +ve) half cycle of AC is absent. But such an oxide layer
can be easily removed by wire brushing.
Argon shielding may be used with DCEN, but penetration will
be less than that with helium.
A square-groove joint configuration may be satisfactory for
some section thicknesses that would usually require a V-groove
joint with AC power. When a V-groove is essential, the root face
may have a greater width and groove angle may be less than those
needed for an AC arc. Examples ofjoint designs for thick sections
are shown in Fig. 8.32.
For manual welding with dcen power the welding torch should
be moved steadily forward, and the filler rod should be fed evenly
into the leading edge of the molten weld pool.
For machine welding square-groove joints, a very short arc is
used and the tip of the tungsten electrode may be positioned at or
below the base metal surface. The position of the tungsten
electrode tip effects weld bead penetration and width as well as
undercutting and weld metal turbulence. The optimum position of
the electrode tip will depend upon the base metal composition and
~ Y
/ 90· thickness and the welding procedure employed.
(iii) Direct Current, Electrode Positive Power

~>-y~} ~(A) Single-V-groove

6 'E~
~
(B) Double-V-groove
6
Welding with deep provides a good surface cleaning action, due
to the mobile cathode spot, and permits welding of thin AI sections
with sufficient current to maintain a stable arc. However, the weld
bead tends to be wide and penetration is shallow. Its application
is limited to sections of up to about 1.25 mm thickness because of
tungsten electrode heating. Argon shielding is recommended
because helium or Ar-He mixture would contribute to electrode
~70· ----y ,<--70· ---y overheating. Use of a suitable weld backing is also recommended.
(iv) Square Wave Alternating Current (SWAC)Power
{35R~~_~ t
A SWAC power is designed to produce d.c. power with an
6 ~_35R~6 ~ arrangement to rapidly shift the polarity to produce a square a.c.
(0 I Double-U-groove wave form of adjustable frequency. In addition, the relative
percentage of electrode -ve time within one cycle of current may
u- groove • L be adjusted w~thin certain limits.
SWAC power combines the advantages of surface cleaning
associated with conventional a.c. power and deep penetration
--- ~ obtainable with dcen power. However, one is gained at the e}:pense
((}5In9"'- ~ A~
3-~f::: 10 of the other. If longer electrode positive time is needud for
acceptable cleaning, penetration will decrease with a specific
Noto, Dimensions are in mm (E) Modified double-V-groove
welding current and frequency. The square-wave form enhances
Fig. 8.32. Typical joint designs for GTA W of thick arc re-ignition during polarity reversal.
aluminium sec:tions with DCEN power.
626 Welding Engineering and Technology
Argon shielding is preferred, but Ar-He mixtures or helium
alone can provide deeper penetration at some sacrifice in cleaning
action.
8.10.4. GMAW of Aluminium and its Alloys
Constant-voltage power source and constant speed electrode
drive are normally used with small diameter electrodes for
welding Al and its alloys. The electrode feed rate is adjusted, to
obtain the desired welding current for good fusion and
penetration. The arc voltage is adjusted to give spray transfer of
filler metal. However, a constant-current power source and
variable speed electrode drive should be used with large diameter
electrodes.
The voltage setting is critical with respect to good fusion with
the groove faces. If the voltage is too high, lack of fusion may occur.
If it is too low, short-circuiting will take place between the
electrode and the workpiece.
Shielding Gas: Argon is the most commonly used shielding
gas for manual welding with spray type of metal transfer. It
provides excellent arc stability, bead shape, and penetration it can
be used in all welding positions. Helium is suitable for machine
and automatic welding with high currents in the flat position.
He-Ar mixtures are sometimes used instead of helium to take
advantage of the arc stahility provided by argon. Such mixtures
may contain from 25% to 90% helium. Arc voltage, penetration,
and spatter increase with increasing helium content..
Electrode :Feeding: Electrode feeding system may be push,
pull, or push-pull type depending upon the wire material and the
mode of welding employed. Both semi-automatic and automatic
modes are used for welding aluminium.
8.10.4.1. Semi-Automatic Welding
For semi-automatic GMAwelding of aluminium the arc should
be started at a location on the joint that will be melted into the
"Weldmetal. Alternatively, it may be st.arted on run-in t.ab.The arc
should not be started on the base metal outside the weld area
because an arc strike might cause a surface. discontinuity that
"Wouldcause eventual failure of the weldment in service.
A good technique is to start the arc in the joint about 25 mm
ahead of the weld starting point. As soon as the arc is initiated, it
is quickly returned to the starting point. The direction of travel is
reversed, and welding is started.
The end of the gas nozzle should be held approximately 20 mm
above the work. However, this distance should be reduced while
"WeldingAl-Mg alloys to avoid loss of Mg by vaporisation.
Weldability of Specific Malerials 627
A short arc length should be used for small fillet welds.
'Selection of correct arc voltage is imperative for quality welds. As
a general rule, harsh-sounding, low voltage arc will cause
excessive spatter and tend to increase porosity. On the other hand,
a soft high voltage arc may res.ult in both incomplete root
penetration and contamination of the weld metal. A compromise
between a harsh, low-voltage arc and a soft, high voltage arc will
usually produce the best results. In multi-pass welding, arc
voltage for the first pass should be set on the low side of the range
to ensure good root penetration. For subsequent passes, the
voltage can be increased to widen the weld bead.
The forehand welding technique is recommended to take
advantage of the arc cleaning action ahead of the molten weld pool.
The electrode should be tilted approximately 7° to 12° from the
vertical and pointed in the direction of welding.
The welding gun should be moved progressively along the joint
without weaving to produce a stringer bead. With thia type of bead,
as much filler as can be controlled without sagging or undercutting
should be deposited in each pass. If undercutting is encountered
with vertical and horizontal butt joints a slight reduction in
welding current is recommended. For fillet welding, a slight
circular motion will tend to agitate the molten pool. This helps in
minimising porosity.
Heat-treatable AI-alloys should be welded using the stringer
bead technique, and the joint should be cooled to below 65°C
between passea. With this technique, the heat input will be
minimised and the HAZ will usually be narrower than when larger
passes are used. On the other hand, some of the non-heat treatable
alloys, such as the 5000 series, may be welded with larger beads
without adversely affecting the strength of the joint.
The biggest diameter electrode that is compatible with the
section thickness, the joint design, and the welding position should
be used. Larger diameter electrodes have a favourable surface-to-
volume ratio, which minimises porosity. They are also economical.
8.10.4.2. Automatic Welding
Automatic or machine welding can be done at higher travel
speeds than are possible with semi-automatic welding. Longer
joints can be made without interrupting welding; this reduces the
number of weld craters and possibility of crater cracks. Higher
welding currents can be used with the maximum amperage limited
only by arc stability or adequate process control. With high
amperages, welds can be made with little or no joint preparation
628 Welding Engineering and Technology
and fewer weld passes. Square groove butt joints in sections up to
13 mm thick can be welded with one pass. Similar joints in 25 mm
thick sections can be welded in the same manner using high
welding currents and large electrodes. It may be necessary to bevel
the plate edges to control the height of weld reinforcement.
Also, sections as thin as 0.5 mm can be welded in automatic
mode with adequate control of the welding variables and good.
fixturing.
Although GTAW.and GMAW are the two main welding
processes employed for welding AI and its alloys other processes
may be used effectively in certain specific conditions. For example,
cold pressure welding is used in joining aluminium cable
sheathing. For certain applications of complex form 'like
automotive radiators and core type plate heat exchangers, brazing
is used for joining aluminium. AI-Si alloys may be used as the
brazing filler material and in some cases it is convenient to use
sheet aluminium clad with the brazing alloy, the whole assembly
being dipped in a bath of molten flux to make numerous joints
simultaneously. Brazing fluxes are normally corrosive and special
procedures may be necessary for their removal. Potassium
fluoro-aluminate based fluxes are, however, relatively non-
corrosive.
8.10.5. Problems Encountered in Welding Aluminium
Two major problems encountered in welding aluminium are
porosity and cracking. The causes for their occurrences and the
remedial measures are described briefly in the following sections.
8.10.5.1. Porosity
Porosity in aluminium fusion welds is due almost entirely to
rejection of hydrogen on solidification of weld bead as is evident
from Fig. 8.33. Gas reactions other than Hz rejection do not occur
Temperature :F
1220 4532
o
o
C7>
629
Weldability of Specific Materials
in the weld pool because of the high affinity of Al for Oz and Nz;
i.e. AI is self-deoxidising.
There are many potential sources of Hz in aluminium welding.
These include oxide film which is hydrated to some degree and
w.hichinvariably has an absorbed film of moisture on it. Lubricant
residues on cold-drawn wire used for GMAWare a source of Hz·
Covered electrodes contain moisture in the covering. Condensation
in gas leads, torch or welding gun of inert gas processes may cause
severe porosity. Thus, cleanliness of wire and welding equipment
is an essential prerequisite for reducing porosity in AI welds. Also,
covered electrodes should be well dried before use.
Porosity in the form of large discontinuous cavities or long
continuous holes may occur due to the use of excessive currents
in GMAW with projected transfer. This defect, often called
tunnelling is due to excessive current causing turbulence in the
weld pool. It may be overcome by limiting the welding current per
pass, by using a gravitational metal transfer technique or by
improving the inert gas shielding.
Fusion welding processes develop porosity to varying degree;
it is most severe with SMAW and least with GTAW.The joint
design and welding position also affect the degree of porosity. In
general, the more difficult the welding position, the greater the
risk of gas entrapment; overhead or horizontal joints are
particularly subjected to this defect. Porosity. appears most
frequently or severely at the start of a weld run.
The effect of porosity is to reduce the strength of the weld
metal therefore in designing welded AI structures or pressure
vessels a joint efficiency factor corresponding to anticipated
porosity level must be used. In the case of fully radiographed shop
fabricated welded structures it is practicable to set up a

50
Boiling point radiographic standard for porosity corresponding to a joint
'":;:: efficiency of 95%. Thus radiographic Al welds can be made in shop
E
u fabrication provided adequate attention is paid to cleanliness of
>- work and welding materials. Sound positional welds may,
~ however, be made in piping where a backing bar is used.
:.c
~
C;
Vl
0.7 Solubility in liquid 8.10.5.2. Cracking
... at melting point Aluminium weldments may be subjected to solidification or
:r Solubility in solid
0.036 liquation cracking and may suffer embrittlement and cracking at
at melting point
temperatures below the solidus. The higher strength alloys may
660 Te'mperature:C 2500 also fail by stress corrosion cracking.
Fig. 8.33. Effect of temperature on hydrogen solubility in aluminium.
630
Welding Engineering and Technology
The solidification cracking of AI and its alloys is associated in
the main with intentionally added alloying elements unlike steel
wherein these are caused by the presence of low melting point
impurities.
Maximum cracking in AI-alloy weldments reportedly occurs
for longest freezing range. In a.long freezing range alloy which is
cooling from the liquidus, the growing crystals are at first
completely separated by liquid and the alloy has no strength. As
the temperature level falls the volume of solid increases relative
to that of liquid, and at some temperature, known as coherence
temperature (Fig. 8.34) the growing crystals meet and cohere.
However, a limited volume of liquid still remains and persist.s
. I-E~- W
Co
<-
OJ
OJJ
O<
"0 -
0
700
550
'r~
500
4 6 12
Silicon (%)
Fig. 8.34. Equilibrium diagram for AI-Si alloys (solid lines) showing also
coherence and nil-ductility temperatures on cooling.
down to the eutectic temperature, causing the metal to be weak.
At the same time, the solidified metal contracts, and is, therefore
subjected to tensile stress which may be high enough (depending
upon the degree of restraint) to cause failure of the weak matrix.
The risk of cracking is greatest when a critically small volume of
liquid metal is present below the coherence temperature. If the
volume of eutectic present is relatively large, incipient cracks are
healed by liquid that flows in from the weld pool. Such healed
cracks are sometimes seen in a macrostructure of AI welds.
Solidification cracking is most severe with weld metal of the
%MgISi(MgzSi) type of composition and the fusion welding ofthis
631
Weldability of Specific Materials
alloy with matching filler metal is practicable only under very low
restraint. The higher strength heat- treatable alloys are also prone
to solidification cracking with the exception of Al-Zn-Mg types,
which are only slightly susceptible. Crack sensitive alloYEfare
usually welded with a filler metal, such as 5Mg, 5Si or 12Si, that
brings the weld metal composition out of the cracking range. In
doing so allowance must be made for dilution of the weld metal
by parent metal.
Liquation cracking in the HAZ may occur when welding high
strength AI alloys of the Duralumin or AI-Mg-Zntypes. This type
of cracking is caused by the presence of low-melting point
constituents in the structure and is associated with relatively low
heat input rates. It may be overcome by using a low melting-point
filler alloy or by increasing the welding speed. Cracking has also
been observed well below the solidus (at about 200°C) in high
strength AI alloys. This effect is also due to the formation of
intergranular films at or near the solidus. These intergranular
constituents cause embrittlement, which manifests itself as
low-temperature cracking.
8.11. WELDABILITY OF MAGNESIUM AND ITS ALLOYS
Magnesium iF>an extremely light weight metal and its al~
lays have high strength-to-weight ratio. It has a density of 1.7
g/cm3. Pure Mg melts at 649°C; has a thermal conductivjty
of about 154 W/m.K while its alloys have thermal conduc-
tivities between 70 to 110 W/m.K. The average coefficient of
thermal expansion for Mg alloys is about 3 x 10-5/oC for a
temperature range of 20° to 400°C.
Mg has low ductility compared to AI. Mg and its alloys are
notch sensitive, particularly under fatigue loading, because oflow
ductility.
Pure Mg has too Iowa strength for engineering use; however,
its alloys are used in a wide variety of structural applications
where light weight is important. Structural applications include
industrial, material handling, commercial, and. aerospace equip-
ment. In industrial machinery, such as textile and printing
machines, Mg alloys are used for parts that operate at high speeds.
Material-handling equipment includes dockboards, grain shovels,
and gravity conveyors. Commercial applications include luggage
casings and ladders. Good strength and stiffness at room and
elevated temperatures combined with light weight make Mg alloys
useful for some aerospace applications.
ti32 Welding Engineering and Technolo~y Weldability of Specific Materials 633

8.11.1. Magnesium Alloys Table 8.38. Temper Designations for Magnesium Alloys
Most Mg alloys are ternary types. The major alloying ele- S No. I Cod;l Description
ments include AI, Zn, thorium and rare earths*. To a lesser ex- letter
tent, Mn and Zr are used singly to form binary alloys. 1. F' As fabricated
The Mg-Al and Mg-Zn alloy groups are suitable for only room 2. o Annealed, recrystallised (wrought productl:! only)
temperature service. Their tensile and creep properties decrease 3. H Strain-hardened
rapidly when the service temperature is about 150°C. The Mg-Th H1 Strain-hardened only
and Mg-rare earth alloys are designed for elevated temperature H2 Strain-hardened and then partially annealed
service. They have good tensile and creep properties up to 370°C. H3 Strain-hardened and then stabilised.
Mg alloys are designated by a combination of letter-number 4. W Solution heat-treated (unst.able temper)
system composed of four parts. Part 1 indicates the two principal T Thermally treated to produce stable tempers other than F, 0, or H
alloying elements by code letters arranged in order of decreasing T1 Cooled and naturally aged
percentage. The code letters are listed in table 8.37.
T2 Annealed (Cast products only)
Table 8.37. Code for Magnesium Alloy Designation System
T3 Solution heat-treated and then cold worked
II
S. No.1 H
K LZ letter
T
S
Q Code
A
E
M Zinc
Silicon
Tin
Lithium
Rare
Aluminium
Silver
Zirconium
Thoriumearths
Manganese.Alloying element T4 Solution heat-treated

I T5 Cooled and artifically aged

T6 Solution heat-treated and artificially aged
T7 Solution heat-treated and stabilised
T8 Solution heat-treated, cold worked, and artificially aged
T9 Solution heat-treated, artificially aged, and cold worked
TIO Cooled, artificially aged, and cold-worked.

An example is AZ80A-T5. The AZ indicates that Al and Zn are

the two principal alloying elements. The 80 indicates that the alloy
contains nominally 8%Al and up to 0.5%Zn. The following Ashows
that this was the first standardized alloy of this composition. The
fourth part i.e. T5 indicates that the product has been cooled and
artificially aged.
8.11.2. General Weldability Considerations for Mg-Alloys
The processes mainly employed for welding Mg and its alloys
Part 2 indicates the percentages of the two principal alloying are gas shielded arc welding and resistance spot welding.
elements in the same order as the code letters. The percentages
are rounded to the nearest whole number. Part 3 is an assigned The Mg-Al-Zn alloys and alloys that contain rare earths or
letter to distinguish different alloys with the same percentages thorium (Th) as the major alloying element have the best
of the two principal alloying elements. Part 4 indicates the con- weldability. Alloys with Zn as the major alloying element are more
dition or temper of the product. It consists of a letter and number difficult to weld. They generally have a rather wide melting range,
similar to those used for aluminium, as shown in table 8.38. which makes them sensitive to hot cracking. With proper joint
They are separated from Part 3 by a hyphen. design and welding conditions, joint efficiencies will range from
60% to 100%, depending upon the alloy.
* A group of 15 similar metals with atomic numbers 57 to 71.
634

Most wrought alloys can be readily welded by resistance spot

Welding Engineering and Technology
I Weldability of Specific Materials

The joint designs suitable for GTAWand GMAWare shown in

635

welding. Castings are, however, not normally resistance welded. I Fig. 8.35. Because of the high deposition rates of GMAWprocess,
8.11.3.Welding of Magnesium and its Alloys a root opening, a bevelled joint, or both, should be used to provide
space for the deposited metal. Increasing the welding speed to
The arc welding processes most often used for welding Mg and maintain a conventional bead size is not acceptable because
its alloys are GTAWand GMAW.The steps involved in fabrication undercutting, incomplete fusion, or inadequate penetration may
by arc welding include the following : result.
(i) Joint design 8.11.3.2. Surface Cleaning
(ii) Surface cleaning
(iii) Filler metal selection Mg alloys are usually supplied with an oil coating, an acid
(iv) Preheating
pickled surface, or a chromate conversion coating for protection
(v) Welding procedure
during shipping and storage. The surfaces and edges to be joined
must be cleaned just before joining to remove the surface
(vi) Stress relieving
protection as well as any dirt or oxide present.
(vii) PWHT.
Oil, grease, and wax are best removed by either washing
These steps are discussed briefly in the following sections. with organic solvents or vapour degreasing in a chlorinated
8.11.3.1.Joint Design 60· hydrocarbon solvent. After all oil or other organic material has
~(typ~
been removed the parts are cleaned by pickling. A final mechani-

+~-I hlIDSj
60·
cal cleaning is normally used before welding to ensure uniform
surface cleanliness. Stainless steel wool or wire brush is recom-
L~ It_-ot03, mended for this purpose.
T ~to5 I- ~ Oto 3~~·5-3 An.oxide film or smut deposited on the surface of weldments
may be removed by wire brushing or by chemical treatment in
I A) Square-groove (B ISingle-V-groove I CIDouble-V-groove
butt jOint butt jOint butt joint a solution of 13 ml of tetras odium pyrophosphate and 10 ml of
sodium metaborate per litre of water operating at 82°C.
Lt ~5'
I / to 60· ~-L 8.11.3.3.Filler Material
The weldability of most Mg alloys is good when the proper
T' 0 01010',+ filler metal is employed. A filler metal with a lower melting point

( D I Fillet
D( '.5-~-rr
-1\--0 to 3
and a wider freezing range than the base metal will provide good
weldability and minimise weld cracking. The recommended filler
(E I Single-bevel-groove IF) Square-groove
corner Joint corner joint T-joint, single weld metals for various Mg alloys are given in table 8.39.
ER EZ33A (RE-Zn-Zr) filler metal is used to weld wrought

L__
gf
and cast alloys designed for high temperature service, either to
themselves or each other. The welded joints will have good
o t
D .-l
3 ~
45·~tYP~([]J_ --j [-3.5 0 tOL
to 6·5
mechanical properties at elevated temperatures.
1-1 It I 1=++ ER AZ92 A filler metal is recommended for welding the room
I Square -g roove
r
( H I Double - bevel-
r temperature service wrought and cast alloys together or to one
(G ( J) Lap joint of the wrought or cast elevated service alloys. It will minimise
T-joi:1t, double weld groove T-Joint
weld cracking tendencies. ER EZ33A filler metal should not be
Dimensions are in mm unitt
used for welding Al-bearing Mg-alloY!:lbecause of severe weld
Fig. 8.35. Typical arc weld joint designs for Mg-alloys. cracking problems.
630 Weldability of Specific Materials 637
Welding Engineering and Technology

Table 8.39. Recommended Filler Metals for The maximum preheat temperature generally should not
Welding Magnesium Alloys exceed the solution heat-treating temperature for the alloy.
S. Magnesium alloy Recommended filler metal
Otherwise, the mechanical properties of the weldment may be
No. I, EZ33A altered significantly.
x
x 12A --1
ER
AZ61A
x
ER
AZ,x - ER
x
x metal
Ba.se
ER
AZIOIA
x
x
x
Solution heat-treatment, or solution heat-treated and aged
castings can be charged into a furnace operating at the preheat
Wrought Alloys x
temperature without damage. They should remain in the furnace
2. AZ31B x
until heated uniformly throughout. Welding should proceed
1.
3.
AZ10A
AZ61A
I x
x
immediately after the castings are removed from the furnace.
4. AZ80A x
C~stings can normally be cooled in still, ambient air after
welding without danger of cracking. However, intricate castings
5. ZK21A x should be cooled slowly to room temperature to avoid distortion
6. HK31A due to non-uniform cooling.
7. HM2IA
8.11.3.5.Welding Procedure
8.
9. I HM31A
MIA Welding procedures for the popularly used arc welding
processes viz., GTAWand GMAWare described in this section.
Cast Alloys
J. AM100A x
GTAW: Gas tungsten arc welding is used for joining Mg
components and repair of Mg castings. It is well suited for welding
2. AZ63A x thin sections. Control of heat input and the molten weld pool is
3. AZ8IA x better than with GMAW.Argon, helium, and Ar-He mixtures can
4. AZ91C x be used as the shielding gas. Typical conditions for manual GTAW
5. AZ92A x of butt joints in magnesium alloys are given in table 8.40.
6. EK4IA Table 8.40. Typical Conditions for Manual
GTAWof Magnesium Alloys
1.
8.
EZ33A
HK31A 1. 2. s. --1
50 231754.0
11A
3.0A
5.0
2250
2A
125
160
200
1754.0
1.5A
Work A
B2.5
thickness
75 100
2.5A
12.5C
6.0
12.5B
9.5 C 2.5
3.0
5.0dia. No.
2.5
4.0
Welding
Joint 1rodof 1.5 (mm)
W-Electrode
35 dia.
Welding (mm)

r-~
I I
9. HZ32A 6. No. (mm)
5.
4.
3.
9.
8.
7.
10.
11. design * .:asse~ c:rrent** ~~
10. KIA
11. QH21A

l~. I
)~.! ZE41A
ZH62A

~]4. i ZK51A
ZK61A
Casting repairs should be made with a filler metal when good
colour match, minimum galvanic effects, or good response to heat
treatment is required.
8.11.3.4. Preheating
The need to preheat the components prior to welding is largely
determined by the product form, section thickness, and degree of * A - Square groove butt joint, zero root opening.
restraint on the joint. Thick sections may not require preheating B -- Single V-groove butt joint, 1.5 mm root face, zero root opening.
unless the joint restraint is high. Thin sections and highly C - Double V-groove butt joint, 2.5 mm root. face, zero root opening.
restrained joints generally require preheating to avoid weld ** With Ax shielding. Helium shielding will reduce the welding current about
cracking. This is particularly true of alloys high in zinc. 20A to 30A. Thorium-bearing alloys will require about 20% higher current.
 639
638 Welding Engineering and Technology Weldability of Specific Materials
50rn/min
Welding Castings: GTAWis generally recommended for the 800
o 5 10 15 20 25 30
i
35
i
40
I45
welding of Mg-alloy castings. Welding is usually limited to the
repair of defects in clean metal including broken sections, sand or 700
blow-holes, cracks, and cold shuts. Repair welding is not a.
rec(lmmended in areas containing gross porosity or inclusions of 'fa. 600
(lxide or flux. Castings that have been originally impregnated for w
u
pressure tightness or that may contain oil in pores should not be
~ 500 ,-
welded. High strength castings used in aircraft structures are <l:

usually heat treated; such castings must be heat-treated again

after repair welding.
The castings should be preheated if the section to be repaired
is thick. Welding of broken pieces should commence at the centre
of the joint and progress towards the end. Medium-sized weld
beads are preferred. Low welding current may cause cold laps,
oxide contamination, or porous welds. Too high welding current
may cause weld cracking or incipient melting in the HAZ. The
filling of holes is usually the most critical type of repair from the
cracking point of view. The arc should be struck at the bottom of o 1800 2000
200 400 600 800
the hole and welding should progress upward. The arc should not o
E \ectrode
be held too long in one area to avoid the possibility of weld cracking
or incipient melting in the HAZ. Fig. 8.36. Melting rates for bare magnesium alioy
electrodes with Argon shielding.
GMAW : Welding can be done with gas metal arc welding
process at speeds 2 to 3 times faster than those with GTAW arc voltage ranges corresponding to each type of metal transfer
process. Higher welding speeds reduce the heat input which, in are 13 to 16 volts for short-circuiting transfer, 17 to 23 volts for
turn, results in less distortion and some improvement in the yield pulsed spray transfer, and 24 to 28 volts for spray transfer.
strength of the joint. Also, the higher filler metal deposition rates The short-circuit transfer mode is used for thin sections, and
reduce welding time and fabrication costs. the spray transfer mode for thick sections. Pulsed-spray transfer
Argon is used as the shielding gas for GMAW.Occasionally, is recommended for the intermediate thick:::lessesbecause there is
mixtures of Ar and He are used to aid filler metal flow and alter less heat input than with continuous spray transfer.
the arc characteristics for deeper joint penetration. Pure helium Table 8.41. Recommended Electrode Sizes for
is undesirable for shielding because it raises the current required GMAWof Magnesium Alloys
for achieving spray transfer and increases weld spatter.
S.
Typical melting rates for standard sizes of Mg-alloy electrodes No. Electrode
(mm) c;~-~~ble Short-circuit base Pulsed-spray
metal thickness I range,
Spray mm
transfer
using deep power are given in Fig. 8.36. This figure illustrates the transfer transfer
operating ranges for the three types of metal transfer, used for
GMAW of Mg-alloys, viz., short-circuiting, pulsed spray, and spray
transfer modes. Without pulsing, the welding amperages between 2. 1.2 1.00 -- 1.75 1.7;.'5
- 3.00 5.0 - 6.25
the short-circuiting and spray transfer ranges would produce 3. 1.6 1.5 - 2.25 2.0 - 6.25 5.0 -7.5
highly unstable globular transfer, which is not suitable for welding 4.
1. 2.4
JU'O I 2.25
0.75 -- 5.00
1.5 5.0- - 2.25
1.5 8.0 4.0 -.& Over
7.5 6.25
Mg alloys.
Like short-circuiting transfer, spray transfer is only stable Recommended electrode sizes for welding various thicknesses
over a limited welding current range. Excessive welding current of Mg-alloys are given in table 8.41. With both spray and pulsed
causes arc turbulence, which must be avoided. The approximate .spray transfer, the lowest welding cost is achieved with the largest

~
 640 Weldillg Engineering and Technology 641
Weldability of Specific Materials

applicable electrode. With short-circuiting transfer, only one or Stress corrosion cracking (SCC) in welded structures usually
two electrode sizes can l;»eused to produce welds with good fu'sion occurs in the area adjacent to the weld bead. It is almost always
and joint penetration. a transcrystalline type of crack. Cracking may be delayed some-
8.11.3.6. Stress Relieving Treatment what by painting. However, this will not ensure crack-free service
High residual stresses from welding will promote stress for long p'eriods, and should not be substituted for stress-reliev-
corrosion cracking in Mg alloys that contain more than about 1.5% ing of the weldment.
AI. Thermal treatments are used with these alloys to reduce Stress-relieving can be accomplished either in a furnace or
residual stresses to safe levels to avoid this problem. Other Mg with a torch. Furnace stress-relieving is preferred. The time and
alloys do not appear to be sensitive to this type of cracking. temperature necessary to stress-relieve weldments of the various
Table 8.42. Recommended Stress-Relieving Mg-alloys and product forms are shown in table 8.42.
Treatments for Mg-Alloys
When a furnace is used, a fixture should be used to support
AZ63A
AZ81A15 I
AMI00A
AZ31B-F
AZI0A-F
AZ31B-O
ZK60A-F
ZK21A-F
M1A-F
AZ31B-H24
HK31A·H24
MIA-H24
MIA-O
AZ92A
AZ61A-F
AZ80A-T5
HM31A-T5
ZK 60A-T5
HM21A-T81
AZ80A-F 60
15
30 150
260
425
400
205
315 the weldment during heating to prevent distortion and correct
I (A)60
Sheet 260 Stress relieving treatment

0q
Time, minutes
260
1. I AZ91C I any warpage. The temperature of large weldments should be
s~ Base metal alloy
1'empe,::!:ture, monitored with thermocouples to make certain that all sections
reach the proper temperature. In torch stress relieving, a
temperature indicating device should be used to avoid overheat-
ing.
8.U.3.7 ..PWHT of Welded Magnesium Alloy Castings
Welded castings are generally given a postweld heat treat-
ment to obtain the desired properties. The appropriate PWHT
depends upon the temper of the casting before welding and
desired temper after welding. Because of the fine grain size and
extensive dispersion of the precipitates in the weld zone,
aluminium bearing castings in the T4 or T6 condition may be
solution heat treated for relatively short heating times after
welding. In the case of AMI00A, AZ81A,AZ91C and AZ92A alloy
castings, the solution heat treating time must not exceed 30
minutes at the PWHT temperature to avoid excessive grain
growth in the weld zone. A protective atmosphere must be used
when the solution treating temperature is above 400°C to prevent
oxidation and active burning of the weldment.
The postweld heat treatments for the various Mg alloys, as
shown in table 8.43, will produce the best weldment properties
and also stress-relieve castings to prevent cracking. If a postweld
solution or temper heat treatment is not required, Al-bearing
castings should be stress relieved. d
~el
.out
this
~high
rcrafts

,-
642 Welding Engineering and Technology
Weldability of Specific Materials 643
Table 8.43. Recommended Postweld Heat Treatments
for Magnesium Alloys (i) Solidification cracking,
220°C (ii) Stress corrosion cracking .
2.
315°C
HK31AT4
None
AZ63A
EK41AT4
~
~
216hr
AZ81A
AZ92A
ZK61AFor
48
ZK51AF'or
HZ32A
AMlOOA
AZ91C
EZ33A
~ii
52-5 hrs
hrs
For
hrs
hrs
hours
hour
hr
hr
hourat
at
FT4
T5
'T4
T4
T6T6
orat
'1'5
atT6
T5
T5
temper
at
.500°C
205°C
150°C
315°C
Recommended
T6
at
at T6
410°C
at
415°C
410°C
415°C
at
For
Desired
330°C;
415°C
205°C
T5
205°C;
243°C;
177°C;
215°C;
220°C
before
415°C F
390°C
Metal
177°C.
.+2215PWHT*
+welding
ForT5
390°C either
after
+ 5
temper
hr
hrs
hrs
hrs
hrs
at
4 hrs
at
at
345°C
330°C + 16
5 hrs
welding hrs atat215°C
177°C 8.11.4.1. Solidification Cracking
at 215°C or 516hrs
260°C
220°C hrs atat220°C·
170°C
. s.Tllag~'i"m Zinc and calcium additions both increase the susceptibility of
Mg alloys to solidification cracking during welding. Zn is a
constituent of a substantial proportion of the alloys; in amounts
of up to 2% it is not harmful, but alloys containing larger
quantities, particularly those with 4 to 6% Zn, have poor
weldability. AI, Mn, and Zr have little effect on this characteristic,
but thorium and rare earth elements are beneficial and tend to
inhibit solidification cracking. Generally speaking the most crack
sensitive Mg alloys are the higher-strength high alloy type which
suffer from cracking both in the weld and at the weld boundary.
8.11.4.2. Stress Corrosion Cracking (SCC)
Aluminium-bearing Mg alloys are susceptible to see in the
HAZ of the welds, and must be stress-relieved, generally at about
250oe, after welding to prevent this type of attack. The zirconium
and thorium-bearing alloys are not susceptible to stress corrosion
attack and do not require stress relief after welding.
There is a risk of fire if Mg is allowed to accumulate in finely
divided form, and proper attention must be paid to cleanliness in
all operations involving cutting, machining, and grinding. Except
in the joining of foil there is no direct risk of fire due to either
fusion or resistance welding of Mg.
Mg alloys are commonly protected against atmospheric
corrosion by means of a chromate dip. The green chromate layer
must, of course, be removed from the vicinity of the joint before
welding.
8.12. WELDABILITY OF TITANIUM ALLOYS
Aircrafts and missiles flying at supersonic and hypersonic
speeds are subjected to extremely high temperatures. Under such
heavy thermal loading, conventional materials such as aluminium
fail to operate satisfactorily because the operating temperatures
are very near their softening or melting points. While steels can
operate satisfactorily at these high temp~ratures, the higher
* Temperatures shown are maximum allowable; furnace controls should be set specific weight of steel limits the use of steel in aircrafts and
so that the temperature doe" not exceed the maximum limit. missiles. Titanium is one metal which weighs 45% less than steel
o 8.11.4. Problems Associated With Welding of Mg Alloys but can operate satisfactorily at temperatures up to 4200e without
4. Two major problems that may be encountered in welding Mg
any significant reduction in its mechanical properties. Hence this
_.~_---l!S are: light weight coupled with adequate mechanical properties at high
temperatures has led to the extensive use of titanium in aircrafts
and missiles.
644 Welding Engineering and Technology
Due to its high corrosion resistant property, nowadays Ti is
being widely used in chemical industries. It is also used in nuclear
industries and deep sea water exploration structures.
8.12.1. Characteristics of Titanium
Titanium is a silver coloured metal that has a close-packed
hexagonal structure, known as a-phase, up to 885°C. Above this
temperature, the metal goes through an allotropic change tQ the
B.C.C. (body centred cubic) structure which is called (3-phase. The
transformation temperilture, sometimes called beta transus, is
affected by the amount of impurities in Ti or by delibrate alloying
to improve properties.
Due to its strong affinity for oxygen Ti forms a stable tenacious
oxide layer on a clean surface, even at room temperature. This
leads to a high degree of corrosion resistance to salt or oxidising
acid solutions and acceptable resistance to mineral acids.
The affinity of Ti for O2 increases with increase in tempera-
ture. At temperatures above 650°C, its oxidation resistance
decreases rapidly, and the metal must be shielded from air to avoid
c.ontamination and embrittlement by O2 and N2•
Pure Ti is very ductile and has relatively low tensile strength.
Oxygen and nitrogen in solution markedly strengthen titanium.
H2 has a definite embrittling effect.
Titanium is relatively light in weight with a density of 4.43
g/cm3 (Al 2.77 and steel 7.75 glcm3) but some of its allQYsattain
a strength of 925 to 1080 N/mm2• It is available in wrought forms
e.g. , plates, sheets, and rods and can be drawn into wires. Ti-alloys
are extremely difficult to machine having a pronounced tendency
to gall up and chip weldment on the edge of the tool. With
conventional tools machining speeds are extremely low and tool
life short. As the melting point ofTi is 1668°C, casting is also costly
and difficult, therefore, the importance of welding for fabrication
of Ti and its alloys into the desired shaped components and
structures.
Physical properties of Ti can be summed up as follows:
(i) Density = 4.43 g/cm3
(ii) Melting point = 1668°C
(iii) Coefficient of thermal expansion at 20PC = 8.4 m-6/m.oC
(iv) Thermal conductivity at 20°C = 16 W/m.K
(v) Specific heat at 20°C = 522 J/Kg.K
(vi) Modulus of elasticity = 117 GPa.
Weldability of Specific Materials 645
8.12.2. Titanium Alloys
Alloying elements are added to Ti to improve properties. An
alloying element stabilises the phase, a or (3,in which it has the
gTeatest solid solubility. For example, AI additions stabilise the
a-phase and raise the allotropic or phase transformation
temperature. On the other hand Cr, Mo, V and some other
elements stabilise (3-phase and decrease the allotropic
transformation temperature. With large amounts of (3-stabilisers,
the (3-phase is stable down to room temperature and below. In
addition, alloying additions influence the rate of transformation
from one phase to another. The mechanical properties of the alloy
depends partially on the phase ratio in the microstructure.
Titanium and its alloys are generally classified on the basis of
the annealed microstructure into the following groups.
(i) Commercially pure titanium
(ii) Alpha and Near Alpha alloys
(iii) Alpha-Beta alloys
(iv) Beta alloys.
8.12.2.1. Commercially Pure Titanium
Various grades of commercially pure Ti are classified on the
basis of minimum mechanical properties and maximum inter-
statial impurities. As the content of interstatial elements like O2,
N2, C, and Fe increases, strength also increases.
8.12.2.2.Alpha and Near-Alpha Alloys
Alpha alloys are not normally heat-treated to increase
strength. They always contain high %age of a-phase in their
microstructure.
These alloys are commonly used where moderate elevated
temperature and strength or creep resistance are required.
8.12.2.3.Alpha-Beta Alloys
Alpha-beta alloys contain mixtures of a and (3phases in their
microstructures and can be strengthened by solution treating and
ageing heat treatments. They have excellent fracture toughness
when annealed and outstanding strength-to-density ratios in the
heat-treated condition. Ti-6Al-4V alloy is the most widely used
a-(3 Ti-alloy.
8.12.2.4.Beta Alloys
Commercial (3-alloys contain a high percentage of (3-phase
stabilising elements, but are not truly single phase; Their
transformation to a-phase is very sluggish and during normal
processing, the a-microstructure is nearly all (3-phase at room
646 Welding Engineering and Technology
Weldability of Specific Materials 647

temperature. Ageing at low temperature increases its strength little effect on mechanical properties of annealed material.
greatly. ~-al1oys are in general characterised by excellent However, the strength of cold-worked material in the HAZ or in
formability into the single phase condition and are often used for brazements is decreased as a result of heating. Therefore, these
this reason. When heat treated they exhibit high strength-to- alloys are normally welded and brazed in the annealed condition.
weight ratios but have relatively low ductility and fracture
tou.ghness. I)-alloys show exceptional work hardening The Ti-O.2Pd alloy is essentially the same as commercially pure
charact(~ristics and are used in fasteners and springs for this Ti from weldability point of view. Ti-O.2Pd and Ti-O.3Mo-O.8Ni
reason. alloys have good resistance to crevice corrosion.
The stabilising effects of different alloying elements on Precautions against iron contamination during preparation
Ti-aUoys are summed up in table 8.44. and welding are same as for commercially pure titanium.
Table 8.44. Effects of Different Alloyillg Ti-5Al-2.5Sn alloy should be considered where good strength
Eleme~ts on Titanium Alloys and toughnesl? are needed with good weldability. ELI (extra low
interstatial) version provides high toughness down to -250°C and
~~-=.~~~~~ers
.==]~_~_·:J;:£~~=~I==--N:.t~~l·--~=
Substitutional Alloying
is suitable for cryogenic applications. To reta.in this property, it
must be joined using processes and procedures that avoid
contamination by C, N2, or 02' It has higher strength at elevated
I
Alumir~ium Chromiu.m Tin temperatures than commercially pure Ti. Therefore, residual
Copper Zirconium welding stresses can be proportionately higher. Stress-relieving of
h'on weldments is commonly recommended.
Manganese
Nickel 8.12.3.3. Near-Alpha Alloys
Palladium The near-alpha alloys exhibit excellent creep strength at
Silicon
Tantalum elevated temperatures. They are weldable, but residual stresses
Tungsten can be high. Therefore, stress relieving of weldments is always
Vanadium recommended. In some cases, special welding procedures are
Interstatial Alloying needed to prevent weld cracking in heavy sections. As with other
alloys, iron contamination degrades creep strength, and should'be

1
1

Nitrogen
--_. ._-_
Oxygen _-~--
Carbon
__ .. ~_._-_.__ .,_.__ ._-_._-'-----_._.-._-_ _--~ .._--~-
Hydrogen
8.12.~J.Welding of Titanium and its Alloys
Important points connected with welding of Ti and its alloys
are briefly described in this section,
8.12.3:1.:-.Commercially Pure Titanium
Commercially pure Ti has moderate strength but good
ductility, It is used primarily for its good corrosion resistanc.e,
formability, and weldability. Filler metals with low iron contents
should be used and all sources of iron contamination during
preparation and welding should be avoided. Thus, steel wire brush
.should not be used for pre-cleaning.
8.12.3.2. Alpha Alloys
The a-alloys and Near-alpha alloys have good weldability
because oftheir good ductility. Welding or brazing operations have
avoided.
.. 8.12.3.4. Alpha-Beta Alloys
The response of 0.-13alloys to heat treatment depends upon
the type and amount of alloying. Welding or brazing of these alloys
may significantly change their strength, ductility and toughness
characteristics as a result of the thermal cycle to which th(~alloy
is exposed.
The low ductility of most 0.-13alloy welds is caused by phase
transformations in the weld metal or the HAZ, or both. Alpha-beta
alloys can be welded with unalloyed Ti or a-Ti alloy filler metal to
produce a weld metal that is low in 13-phase.This improves weld
ductility. However, this procedure does not overcome the low
ductility of HAZ in alloys that contain large amounts of 13-phase.
stabilisers.
Ti-6Al-4Valloy has the best weldability of 0.-13alloys. Ti~lOV-'2
Fe-3Al has good weldability because it is highly 13-stabilisedand
retains a fine grain size during welding.
648 Welding Engineering and Technology Weldability of Specific Materials 649

Alpha-beta alloys that are highly ~-stabilised have limited stainless steels or nickel alloys. Ti alloys are sometimes rated
weldability. They tend to crack when welded under high restraint according to their ability to produce tough, ductile welds in them.
or when minor defects are present in the weld zone, for example, One weldability rating is shown in table 8.45.
Ti-7Al-4Mo and Ti-6Al-6V-2Sn alloys. The resistance to cracking Table 8.45. ~lative Weldability of Titanium and its Alloys
may be improved by preheating in the range of 150° to 175°C
followed by stress relieving. Welding of Ti-8Mn alloy is not S. Beta
Alphaalloys
Alpha-Beta pure A
alloysClass
Commercially
Near-Alpha alloys
alloys Rating
Grade (i) D
B
A
AC
A
B
CWeldability
Ti-13V-ll
Ti-6Al-4V TiCr-3AI
(i) Ti-5Al-2.5Sn
Ti-8Al-IMt>-V
Composition
TItanium Alloy
recommended. 2. No.
4.
3. (ii)
(iv)
(iii)
(ii) Ti-O.2Pd
Ti-6Al-2Cb-l
Mn
Ti-6Al-4V-ELI
Ti-7Al-4Mo Ta-O .8Mo
Ti-5Al-2.5Sn-ELI*
(v) Ti-8
Ti-6AI-6V-2Sn
Ti-6Al-4Zr-2Mt>-2Sn
5. 1.
8.12.3.5. Beta Alloys
Most ~-alloys are weldable in either the annealed or the
heat-treated condition. These include Ti-13V-llCr-3AI,
Ti-8Mo-8V-3Al-2Fe, and Ti-11.5Mo-6Zr-4.5Sn alloys. Weld joints
have good ductility but relatively low strength in the as-welded
condition. They are used most often in this condition because the
welded joint can respond differently to heat treatment than the
base metal and have low ductility after ageing. Ageing can 'take
place if the welds are exposed to ~levated temperature in service.
To obtain welded joints 'with 100% joint efficiency, the alloy can
be welded in the annealed condition.
8.12.3.6. Dissimilar Alloys
Titanium alloys within the same group can generally be * Extremely low interstatial.
welded together. Commercially pure Ti can be welded to a.-alloys A - Excellent; B - Fair to Good; C - Limited to special applications;
and to some a.-~ alloys. Commercially pure Ti filler metal should D - Welding not recommended.
be used for welding such joints. Sufficient weld reinforcement can Alloys rated A or B are considered usable for most applications
provide good strength properties in the joints. in the as-welded conditions. Welded joints in all the weldable
One pl..lcaution that should be taken when welding with alloys should have joint efficiencies near 100%, if the base metals
unalloyed filler metal is against the possible diffusion of residual are initially in the annealed condition. Weldments of many alloys
H2 from the base metal to the weld metal. Otherwise it will result with limited weldability can be given an annealing heat treatment
in the formation of hydrides in the weld metal which may lead to to improve ductility. '!\vo weldable alloys, viz., Ti-'6Al~V and
delayed brittle fracture. Ti-13V-llCr-3AI are heat-treated before welding to increase
8.12.3.7. Compatibility of Titanium with Other Metals their strength properties. An important criterion of weldability is
Titanium can be successfully fusion welded to the other the capacity to produce welds free from defects that would limit
reactive metals zi.rconium (Zr), Cb (Columbium), Ta (tantalum), the effectiveness ofthe weldments in service. Especially important
and hafnium (Hf). However, uniform melting is difficult when is the capacity to resist weld cracking, which is related to restraint
fusion welding it to all but Zr. Ti forms continuous solid solutions imposed by the fabrication, and the strength and toughness of the
with these metals but the welded joints are expected to have alloy. The high strength Ti-6Al-6V-2Sn and Ti-7Al-4Mo alloys
ductility. On the other hand, Ti forms brittle intermetallic have limited weldability because of a tendency for weld cracking
compounds with more common structural alloys of Fe, Ni, Cu and under conditions of high restraint and the presence of minor
AI. Thus, extremely brittle welds are formed when dissimilar discontinuities in the weld metal or HAZ. The near-Beta alloys,
metals are diffused together. Ti-clad steel is produced by explosion for example, Ti-8Mn, are not recommended for welding because of
welding. weld cracking under conditions of moderate restraint or minor
weld zone discontinuities. The resistance to weld cracking of the
8.12.4. Weldability Rating high strength, low toughness alloys can· be improved by main.
Unalloyed Ti and most Ti-alloys are readily welded using taining a preheat of 150° to 175°C during welding and by stress
. procedures and equipment suitable for welding austenitic relieving immediately after welding.
650 Wel.ding Engineering and Technology Weldability of Specific Materials 651

8.12.5. Welding Processes and Procedures The design of a weldment, the types of joints, and joint
Titanium and its alloys can be welded by anyone of the locations can be limited by shielding requirements during welding
following processes: in chamber, positioning of the weldment for welding each joint
1. Gas tungsten arc welding (GTAW) must be considered during the design phase.
2. Gas metal arc welding (GMAW) 8.12.5.2. Pre-Cleaning
3. Plasma arc welding (PAW) Prior to welding, brazing or heat treating Ti components must
4. Diffusion bonding be cleaned of surface contaminants and dried oil, finger prints,
5. Resistance welding grease, paint and other foreign matters should be removed using
6. Flash welding a suitable solvent. Ordinary tap water should not be used to rinse
7. Friction welding titanium parts. Hydrocarbon residues can result in contamination
and embrittlement of titanium.
8. Welding with electromagnetic agitation
9. Electron beam welding If the parts to be welded or brazed have a light oxide coating
10. Laser welding. in the vicinity of the joints, it can be removed by pickling in an
However, the welding processes most often used are the arc aqueous solution of 2 to 4% Hydrofluoric acid and 30 to 40% nitric
acid followed by appropriate water rinsing and drying.
welding processes like GTAW, G·MAW,and PAW. The steps Hydrogen absorption by Ti alloys is generally not a problem
involved in arc welding Ti and its alloys include the following: at bath temperature up to 60°C. The parts should be handled at
1. ,Joint design selection this point with lint-free glove.':lduring assembly in the welding or
2. Pre-cleaning . brazing fixture.
3. Selection of preheat and interpass temperatures Scales formed at temperatures above 600°C are difficult to
4. Protection during welding remove chemically. Mechanical methods such as vapour blasting
5. Welding process selection and grit blasting are used for scale removal; these are followed by
6. Postweld heat treatment. pickling to eI.!:':lurecomplete removal of scales and any contami-
All these steps are diseussed briefly in this section. nated metal on the surface.
8.12.5.1. Joint Design To control porosity in welding operations. the edges to be
joined are often given special treatments such as draw filing, wire
The weld joint design used for welding titanium are similar to brushing or abrading the joint edges. Sheared joint edges
those used for steels. Actual joint design depends upon several frequently require such special treatments to remove entraped
factors including the welding process, mode of operation (manual dirt and small cracks because these edge discontinuities promote
or mechanised), joint accessibility, and inspection requirements. weld porosity.
However, one of the typical edge preparations for titanium is as For extended storage time before welding it may be necessary
shown in Fig. 8.3.' I
to store the parts in sealed bags containing desiccant or in
controlled humidity storage room. The fixturing itself should be
thoroughly cleaned and degreased prior to loading.
8.12.5.3. Preheat and Interpass Temperatures

\6';--- -l~m.
,__ \:70'/ Preheat and interpass temperatures must be kept low
120°C) for welding in the open, to avoid surface oxidation that
(:5,

-L
0-0.25-11- Los. mm can be dissolved into the molten weld metal. Low preheat is
employed to drive off IrlOisture from the surface prior to welding.
Fig. 8.37. Typical joint edge preparation fOl' welding titanium. Prolongeci exposures above 120°C can cause the formation of oxide
Edge preparation should be done by a machining process that layer on the Ti weld and base metal. The oxide film must be
does not contaminate the titanium or leaves embedded particles removed with a stainless steel wire brush or rotary carbide burrs
on the surface. prior to welding to reduce contamination.
652 Welding Engineering and Technology
8.12.5.4. Protection During Joining
Because of the sensitivity of Ti to embrittlement by O2, N2,
and H2' the entire component or that portion to be heated above
260°C must be protected from atmospheric contamination.
Protection or shielding is commonly provided by a high purity
iner: gas cover in the open or in a chamber, or by a vacuum of 104
torr or lower.
During arc welding, Ti must be protected from atmosphere
until it has cooled to below about 425°C. Adequate protection can
be provided by an auxiliary inert gas shielding device, when
welding in open. For critical applications, welding should be done
in a gas-tight chamber, that is thoroughly purged of air prior to
filling with high purity argon, helium, or mixture of the two.
The purity of the shielding gas influences the mechanical
properties of the welded joint. Both water and air are particularly
detrimental. Hence necessary care should be taken to ensure that
no air or vapour is added in the inert gas. The dew point of the
gas should be measured at the welding location. Maximum
permissible moisture limit at the welding point is a dew point of
20°F (-6.7°C). Welding gases have a dew point of -65°F (-54°C)
or lower. The inert gas at the cylinder or the source should be
sufficiently dry to allow a margin for some moisture pick up in the
delivery system. One sYl?temfor checking gas purity is by welding
and bending a scrap before and after welding the workpiece. The
surface appearance and 'degree of bending are good indications of
the gas purity.
The colour of a weld bead on Ti workpiece indicates the level
of contamination or the shielding gas purity. A light bronze colour
indicates a small amount of surface contamination, while shining
blue colour indicates a greater amount of surface contamination.
Neither level of contamination is desirable but they may be
accepted on a single or final weld pass, provided the surface layer
is removed before placing the weldment in service. A white flaky
layer on the weld bead indicates excessive contamination which is
not an acceptable condition. If a white or grey flaky oxide is
present, the gas shielding system should be inspected and the
cause of contamination corrected. The contaminated weld metal
should be removed because it is likely to be brittle.
8.12.5.5. Welding Processes Selection and Procedure
GTAW,GMAW and PAW bc~ng the often used pro~esses for
titanium; these are described in this section.
* 1 torr = 1 mm ofHg.
Weldability of Specific Materials 653
(i) GTAW: Gas tungsten arc welding is used either in the open
atmosphere conditions or in a chamber known as Glove Box
Welding.
(a) Welding in the Open: This process is commonly used to weld
Ti and its alloys, particularly for sheet thickness up to 3 mm.
Welding in open is best done in flat position to maintain adequate
inert gas shielding.
Direct current with electrode negative (dcen) is normally used
with type EWTh-2 tungsten electrodes. Contamination of weld
with tungsten must be avoided to avoid embrittlement.
More care is required while starting and stopping the weld in
the manual welding. Scratch start is not recommended and should
be avoided. Starting tab should be used or use of high frequency
power should be made.
(b) Glove Box GTAW : Ti alloys react quickly with and are
embrittled by O2, N2, and H2 and therefore an effective inert gas
shielding must be maintained when the metal is being welded,
until its temperature has fallen below 500°C. Sufficient protection
can be provided by using a torch with a trailing shield extension,
•Welding direction
Inert gas
inlet
~
Porous metal
diffusion
Torch screen
nozzle
Fig. 8.38. Trailing gas shield torch for gas tungsten arc welding of titanium.
as shown in Fig. 8.38. For smaller components, however, welding
in a sealed chamber is often a more satisfactory alternative.
The Glove Box comprises a rigid clear plastic dome hinged to
a substantial steel topped table. The table top has a ground finish
to provide a complete seal between it and the dome, which can be
clamped to the table by toggle clamps fitted around its perimeter.
Work pieces can be placed inside the dome either through an entry
chamber, in the case of small items, or by unlocking the dome and
swinging it clear of the table.

654 Welding Engineering and Technology
When the workpiece has been placed inside the chamber, it is
~ealed and the air is exhausted with the help of a rotary vacuum
pUJnp. The chamber is then purged and filled with argon.
Permanently fitted inside the chamber is the TIG welding torch
which the operator can use through any of the built-in glove ports
which gives the unit its trade name i.e. Glove Box GTAW. The
glove boxes are available in various sizes.
The process of welding Ti in an inert gas chamber, i.e. Glove
Box, ensures high quality welds for acro parts.
(ii) GMAW: Gas metal arc welding is more economical than
GTAW because of higher deposition rate, particularly with thick
sections. Welding conditions should produce smoothly contoured
weld that blends smoothly with the base metal.
With GMAW, the droplets of finer metal being transferred
across the arc are exposed to much higher temperature than the
one fed into GTAW molten weld pool. Hence, the weld metal is
more susceptible to contamination by impurities. Therefore, care
must be taken in the design of shielding Ilevice to avoid the
probability of such contamination.
Welding Consumables: Cleanliness and uniformity of the
wire electrode are extremely important. It must be free of dirt,
grease, oil, drawing compounds, and surface imperfections that
may entrap foreign matter.
Shielding gases should be of welding grade. Ar, He, or mixture
of the two are used as shielding gas. Oz, Nz, Hz or COz should not
be added to the inert ga~ because they embrittle titanium. The
inert gas system must be leak-tight to avoid contamination by air
or moisture ..
Filler Metal Transfer: Titanium filler metal can be trans-
ferred by all three modes viz., short-circuiting, globular, and spray.
Globular "transfer is not recommended because it may cause
8Kcessive spatter and incomplete fusion in the weld. Short-
circuiting transfer can be used for welding thin sections in all
positions and thicker sections in positions other than downward.
However, incomplete fusion may occur in thicker sections due to
10w heat input.
Spray transfer is preferred for welding thick sections in the
flat and horizontal positions. Regardless of the type of transfer a
trailing gas shield and gas backing are required.
(iii) Plasma Arc Welding :PAW is an extension of GTAW in
that the arc plasma is forced through a constricting nozzle. Inert
gas shielding of the weld is provided by a shielding gas nozzle and
WeldabiUty of Specific Materials 655
an auxiliary trailing shield as with GTAW and GMAW. Welding is
normally done with a transferred arc using direct current
electrode negative (dcen) supplied by a constant current power
source.
Welding grade AI' is generally used as the plasma and
shielding gas, but He-Ar mixtures are sometimes used particularly
for shielding. Hydrogen must not be added to the inert gas because
of its embrittling effect on titanium.
PAW can be carried out using two techniques, viz., melt-in and
key-hole. The melt-in technique is similar to GTAW. The key-hole
technique provides deep joint penetration for welding square
groove joints inane or two passes, depending on the metal
thickness. The two techniques can be combined for welding groove
joints in thick sections. "
The advantages of PAW are higher welding speeds and lower
filler metal requirements than with GTAW.
8.12.5.6. Postweld H(~a.t Treatment of Titanium Weld.
ments
Relief of welding stresses is generally recommended for
complex weldments to avoid cracking or other undesirable effects
of high residual stresses. Residual welding stresses can be
sufficiently high to promote str£ss corrosion in sensitive alloys.
Also such stresses can cause weldments to have low endurance
limits in high- and low-cycle fatigue.
When residual welding stresses are expected to be a problem,
the structure should be heated to the proper temperature and held
for sufficient time at that temperature to relieve the stresses. The
time and temperature for annealing and stress-relieving residual
stre~ses for weldments in different '1'1 alloys are detailed in Table
8.46. An annealing heat treatment automatically relieves residual
stresses. The minimum temperature range is 540° to 700°C for
short time operations. Selection of time and temperature depends
upon mechanical properties required, as ageing can take place
during stress-relieving of heat·treatable alloys.
Before the weldment is heated, it should be cleaned to avoid
contamination and stress corrosion cracking. Surface discoloura-
tion can be removed by chemical cleaning after heat treatment.
8.12.6. Pro~lems Associated with Welding of Titanium
Major problem associated with Ti-weldments is porosity. It is
a persistent problem and mostly it is caused by gas bubbles formed
during solidification of weld metal ; welding procedures and
 656 Welding Engineering and Technology
Weldability of Specific Materials 657

"" o 10 techniques can affect this. There are a number of sources of

~
'"'"0 ~ .b
"<1'
I
•....•
"" 'f
""
o
J.,
""
l' ""
~~~"<1' 00 l' ~~~
l' ~ ""
...•
<.0

J,9 l'
""

porosity in welds.
o
•....•

t;
-/ e:::
~I""" 0 000 o o.-; ~
.E Presence of H2 in metal in solution is not a prime cause of
~ porosity. Porosity-free welds can be made in Ti when ~ content
.;:;
is within specification limits. Hydrogen content of Ti is specified
oll) 0 o10 0 O°ll)
at 150 ppm or less, and usually it is less than 100 ppm.
'"L ll) <.0 <.0
•.••.•
10 <.0 <.0
ll) ~~2S52~2 "<1'C':l Cl>
10 t- 10 Filler metal with higher H2 content can cause porosity;' Die
~ '".6 6 6 ~J.,6

; rjj ll) 10 t- <lj' l- '? 666

11'--'~~ ll)
00
~ "<1'
"<1' 00
ll) Cl> 00
10 "<1'
66~0J.,0~
;1.;;1.;"<1'~~~<.O
t"j
00 ""
"<1' t- •....•
10
lubricant worked on to the wire during drawing is a major cause
of porosity and hence it must be removed before use by wiping
=--
with solvent-moistened, lint-free cloth.
Q) tI) ..
=c; :.:( Another source of H2 is absorbed moisture. This can be
~ s c:

...... evaporated by heating the metal to 120°C before welding.

~-
'" '"
........ol
'" '"
c:/)
c:/)"'Cl
~

~
::l
OJ l:l
ol
E c: 0
oS
_ ......
'"
ol
'"
ol
'"
ol Moisture in shielding gas seriously impairs weld properties
iii~ "'I ~ :.:( E:.:(:.:(:.:(:.:«
cd s... ••.• = :.:( r:..
::s :.:( r:.. < 0
.!;l 10 ••
......
~ ~ ~
......
other than porosity.
1~ -18
••
•••• •• ~
000
J...i loot ••••• "<1' :.:(
::s
.SO r:..
o 0 .s Cleanliness of the joint faces and surrounding area is a ma.i<
~~ ~ C,.)
~ ~ ~
'" :.:( :.:( :.:(
factor in producing porosity-free welds. A large number of fore;
'"
t;
c:
.. materials can cause porosity. Some of them are as follows :
~ § r:..
::s

~ -a .E (i) Plasticizers dissolved from rubber gloves by solver

I l T
.SO ~ 10 10
especially alcohol, are cause of porosity as are resk
~~ l' ...• ~ in clothes used for wiping joint area.
~~I~ ......•..•
""
...•
I I ""
.
o l'...•00
00 10 110
"C J, ~"'f"""J,'i'
~ r;:'"0 o
~.s -.:: $.< ~ OJ •••••• "" I""

""""'
ci
10""
o ci ~
C'!
00 C'! 'i'
•..• Ui) Fingerprints, dirty rags, and lint-bearing gloves ,
s=
~ .•..
CIl
U"? .,.... 'f"""4 •••••

0 o placed in contact with joint edge before welding.

(iii) Grinding residue embeded in titanium.
cQ~ .•.•
~ 10 0 0
~g~~g~~ 000 As sheared edges can get contaminated by materials th;
~I~
as
. t-I
"<1'* Cl>0* Cl> <.0 Cl>
~~ ~S5 o og
"?
Cl>t-
"" t- 00
o~
t'j ~ 0t-
ll) Cl> Cl>
t'j t"j ~
~:g:g:g:g~g
"? t'j ~ "?
t- t-
10
o
I ll)I
0
t-
10
t-
I cause porosity, removal of all burrs eliminate or greatly reduc(~f
effect of this factor.
t- t-
~-;
- ~
~ <.0
<.0 <.0 t- t- t- t- <.0 t- t- t- t-

co= Therefore, to reduce the problem of porosity it i3 recomm,

.£ to shear the metal only a few hours before welding.
~ ~
'C o
~ Porosity is also influenced by welding procedur
of •• ::s
~o d•• ~
c:
d J5 '" techniques. As welding variables are inter-related it is diffi
t ::2l~~ III
U)
CZ <~~
ll) r:..
identify them; some of them are:
~ If c:~ J5~ ~
=~ <'?~<
"" ...• ""
0 .:: A ~ :> O~O <.oC?..;. (i) improper tack welding technique,
~ =;.>
III
U)
~ ::2l U)
10
0"" N
"l'
ClI 10
:>::2l::>::2l & 9~~
t ••••

~ (ii) wide root openings in weld joints,

.~ ~ <:;1 ~
ClI C'i "<1'"<1'<.0<.0
I I I I d:>"" IIJ:>U':)
•....• ""'.
0 6
e =«I ~I <
~ ex>
:;i!:;i!<
10 <.0 <.0 <Ql
Cl:l
:;i!
C':l
««::2l0
<"?~~<1'0?~ ~"7'-;t0C(
~C':l,...;::2l ~
~
(iii) use of consumable inserts.
ElCllE=:~ ~ ~ ~ ds ~
.r:l E=: E=: E=: E=: E= E=: =a E=: E=: E= :: Other variables affecting porosity in Ti-welds are:
8 ~ ~ .- :;:.~3 ~ s:::;
.r:l ~ ~
:::-:;-:;::;-::;.5:::-:::-:-:;
- - ~ -~~
El

....•
.•.•••.• - -:;)
:) ••.•• ~
_<P - ....•
••.••••.••••.•• (a) heat input,
~ ~
Q. (b) cooling rate,
M ~
C"Ji:;gI"";C'i o*
::s
(c) welding speed,
(d) arc voltage,
 658 Welding Engineering and Technology

(e) shielding gas flow rate.

Welds characteristically made with low heat input are porous.
These include, small tack welds, fillet welds, and partial
penetration groove welds. These can be corrected by the use of
increased heat input.
At times, temperature of the weld pool and its cooling rate
might also cause porosity. NINE
Often porosity can be reduced by the use of copper chills placed
• adjacent to the weld joint. As chilling effect requires high welding
heat, that produces a more fluid weld pool from which dissolved Weld Defects
gases are evolved at a faster rate.
The performance of welded structures or components in
service depends upon the quality of fabrication which in turn is
based on the presence or absence of defects in weld joints. The
term weld defect applies to any departure in welded structure or
welded joints from specified requirements. It has been suggested
by various authorities that the term defect should be reserved for
faults which are likely to be detrimental to the service life or
functioning of the fabricated components, especially if they are
situated in a highly stressed region. On the other hand, the same
faults could be termed imperfections, or flaws, or discontinuities
if situated in regions of low stress. IIW (International Institute of
Welding) also defines discontinuity as, an interruption of the
typical structure of a weldment such as a lack of homogeneity in
the mechanical or metallurgical or physical characteristics of the
material or weldment.
Defects impair the strength of welded joints and may result
in the failure of a whole structure in service. Their principal causes
are failure to follow the requirements of welding specifications,
the use of sub-standard materials, defective welding equipment,
or poor welding skill. Weld defects could as well be due to the fault
of the designer i.e. design or drawing error, or fault of some
manufacturing or shop function, such as material preparation.
9.1. CLASSIFICATION OF WELD DEFECTS
One of the IIW Documents classifies the weld defects into six
groups as follows :
1. Cracks-includes all types of cracks such as crater
cracks, hot cracks, cold cracks, etc.
2. Cavities-including blowholes, porosity, shrinkage, etc.
3. Solid inclusions-including slag, flux, metal oxides,
tungsten, etc.
 Weld Defects 661
660 Welding Engineering and Technology
(ii) Lack of fusion,
4. Incomplete fusion or penetration-including lack of Wi) Lack of penetration,
fusion, lack of penetration, suckback, etc. (iv) Blowholes/Subsurface porosity,
5. Imperfect shape or Unacceptable contour-includ- (v) Shrinkage cavities,
ing dimensional deviations, undercut, underfill, overlap, (vi) Slag inclusion,
excessive reinforcement, excessive penetrat.ion, bead (vii) Tungsten/copper ( Metallic) inclusions,
shape, etc. (viii) Finger penetration,
6. Miscellaneous Defects-including arc strike, excessive (ix) Coarse microstructure.
spatter, rough surface, uneven ripples, pock marks, etc. The causes and remedies for all these defects are discussed in
In the present text, however, all these defects are discussed this text mainly for the arc welding processes, however a limited
under two categories, viz., discussion for similar defects occurring in fabrications using other
(I) Visual/Surface defects, welding processes is also included.
(II) Hidden/Sub-surface defects. 9.2. GENERAL SOURCES OF WELD DEFECTS
Different defects included in these two categories are as The following are general sources of defects for most of the
follows. conventional arc welding processes.
VisuaVSurface Weld Defects 1. Improper selection of process, for example, using a very
Under this category are included the following defects: deep penetrating heat source on a narrow Vee angle so
(i) Surface cracks, .
causing cracking in the root run due to large depth-to-
width ratio.
(ii) Distortion/Warpage,
2. Applying the welding process incorrectly for the par-
(iii) Incorrect or bad bead profile,
ticular application, i.e., incorrect current setting or gas
(iv) Dimensional deviation/Incorrect weld size,
shield, excess weld metal deposition, etc.
(v) Overlaps or rolls,
3. Defects caused by the interaction of the weld metal with
(vi) Undercuts, prior defects in the base metal, e.g., laminations·, and
(vii) Underfill/Suckback, impurities like phosphorous, sulphur, and silicate, etc.
(viii) Excess penetration/Excess reinforcement, that cause brittle and weak zones resulting in, for ex-
(ix) Start-Stop points, ample, lamellar tearing.'
(x) Interruptions, 4. Undesirable metallurgical structures with respect to
(xi) Unfilled craters, grain size and hardness as well as undesirable in-
(xii) Burn-through, clusions such as tungsten oxide and slags. Hydrogen is
(xiii) Surface porClsity, a most undesirable inclusion as it is often the main cause
(xiv) Surface oxidation, of cold cracking in steels.
(xv) Uneven and Coarse ripples, 5. Undesirable shape and size of weld bead due to overfill
(xvi) Pock marksIFlat spots, and/or poor profiles.
(xvii) Excessive spatter, 6. Incorrect joint edge preparations and poor fit-up leading
(xviii) Arc strike, to inaccessibility and lack of fusion, cracking, etc.
(xix) Bad surface appearance, 7. Poor supervision allowing poor workmanship resulting
(xx) High-Iow/off-set. in stray arcing, tool marks, undercuts, inclusions, poor
Hidden or Subsurface' Weld Defects finish, lack of fusion and penetration, and incorrect weld
• Laminations are formed when gas voids or shrinkage cavities in an ~got
This category includes the following defects : are flattened during hot working operations; they generally run parallel to the
surfaces of rolled products.
(i) liidden or Subsurface cracks,

""'"'--
 "<~-<~-,----------

662 Wet'ding~ngineering and Technology Weld Defects 663

shape causing a reduction in fatigue life and joint

strength.
8. Inferior design resulting in inaccessibility for welding
and inspection, hard spots and stress raisers leading to
cracks, difficult welding positions, and poor weldability
of base material which is not notch ductile. (A material
which continues to deform plastically is said to be notch --.........
Transverse cracks
ductile. If it cracks in a rapid brittle manner it is not a ,/
notch tough or notch ductile material).
Apart from these general causes ofweld defects there may also
be some other reasons for the occurrence of certain specific defects.
The causes and the remedies for each of the defects listed under
the categories of surface or visual defects, and subsurface or
hidden defects are described in the following sections for arc
welding as well as the other welding processes.
9.3. ARC WELDING DEFECTS
Arc welding processes being the most popular processes for
fabrication most of the defects studied and analysed are with
respect to these processes.
9.3.1. Surface or Visual Defects
These defects are discussed herElunder in the same order as
they are listed in classification of defects in section 9.1.
9.3.1.1. Surface Cracks
Crater cracks
Cracks are the most dangerous of all weld defects. If a crack
is present in the weld metal or HAZ a dangerous level of stress is
shown to have existed which has been relieved by cracking. Cracks Fig. 9.1. Schematic representation of types of surface cracks in
butt and Tee joints made by arc welding.
are invariably unacceptable. Some micro cracks, however, remain
undiscovered and may corne to light after failure of weldment due the surface. These cracks are shown schematically in Fig. 9.1 for
to other causes. Cracks are usually caused by : butt and fillet welds.
(i) high contraction stresses (being greater than the hot Crater Cracks are caused whenever welding operation is
strength of the weld or parent material); interrupted, there is a tendency for a crack to form in the crater.
(ii) poor fit-up and incorrect welding procedures; These cracks are usually star-shaped and progress only to the edge
(iii) poor edge quality. of the crater. However, these act as the starting points for
longitudinal weld cracks, particularly when cracks occur in the
Deep penetrating processes such as CO2 welding, and SAW, crater formed at the end of the weld.
combined with narrow weld gaps, also cause hot cracking at the
root. High welding speeds are a prime cause of cracking, parti- Crater cracks are found most frequently in materials with high
cularly in aluminium alloy weldments. coefficient of thermal expansion, such as austenitic stainless steel.
However, the occurrence of such cracks can be minimised or
Surface cracks can be seen on the surface of the weld ·using prevented by filling craters to a slightly convex shape prior to'
the visual testing techniques. There are several types of surface breaking the welding arc.
cracks viz., transverse; longitudinal, and crater cracks. There are
also toe cracks in adjacent parent metal which normally come to Detailed treatment of weld cracks is given chapter 6.
664 Welding Engineering and Technology Weld Defects 665

9.3.1.2. Distortion
Uneven heating of workpieces during welding results in the
development of welding stresses which often lead to distortion or
warpage of the welded structure or component.
Distortion can be corrected by proper jigging of joint prior to
welding' or postweld slow cooling or stress-relieving heat treat~ Note: ( denotes convexity.
m ent, and sometimes by peening the weld metal and HAZ if the
( A 1 Desirable fillet weld profiles ( B 1 Acceptable fillet weld profiles
fabrication specifications allow it. Proper sequencing of welding
procedure may also balance stresses and eliminate the problem.
Reducing welding current and the use of small diameter electrodes I'I I, I, I " '"
are the means of reducing heat input with consequential reduced I' I' I' ,
distortion or even the complete prevention of it. The exact method ~'1 '_ ••• ~~~~~'
1 __ ~:: _ I __' ... : ' ' _~
employed will be dictated by the size and shape of the parts as
well as by the thickness of the metal being welded. If the weldment !--Size~
Insufficient
~~ze~ !--Size~ ~size~ ~size~ ~Size~
Excessive Excessive Overlap Insufficient Incomplete
is grossly out of shape, or out of alignment or dimensionally throat convexity undercut leg fusion
incorrect, it is simply a reject. l C 1 Unacceptable fillet weld protl\es
Detailed description of different methods employed for ~ R

reducing or preventing distortion altogether in weldment is given

in chapter 4.
f 8: J=r
®
9.3.1.3. Incorrect Weld Profile
, , t

rroQSJ
Note: R denotes n;~forcement. R
Bad bead profiles have r-,a considerable
performance of a weld under load. '
convexity
, , effect on the
~ pass of a multipass weld
If one
,
/undercut
. Insuff
r-<,
Overlap
lD
i\cie1nAcceptable
..... t
th Excessive
roat ""

profile butt weld

)-<..'
has a bad profile it can cause Excessive
incomplete fusion or slag inclusions,
CS8
even though the subsequent weld passes will partially remelt the
first pass. A bad weld bead profil~ in one pass can cause
subsequent cover passes to be out of size, out of contour, or create
some other weld defects that will be covered up by follow-up
passes, making the problems impossible to see without special test ( E 1 Unacceptable butt weld profiles
procedures. The problem is most severe when the bad pass is the
Fig. 9.2. Acceptable and defective weld profiles for fillet and butt welds.
nrst, or root pass of the weld. That is why highly skilled welders
are often employed to make the root passes on pipelines and less welds. In automatic welding these defects may be caused by
skilled welders follow them up to complete the remaining weld variation in the supply voltage, wire slippage in feed rolls,
runs to fill up the joint groove. The fillet weld is particularly variations in the welding speed because of backlash and lost
vulenerable to the problem of imperfect weld profile.
- ';,1
motion in traction gear, a wrong work angle, etc. In semi-
Requirements concerning incorrect weld profile in finished automatic welding, dimensional defects are usually traced back to
welds are usually included in the specifications and drawings the poor skill of the operator. Fig. 9.3 (a) shows some examples of
involved. Nonconformity with the requirements constitutes a weld incorrect weld sizes and the same can be found out with the help
defect. Fig. 9.2 illustrates various types of acceptable and defective of templates as shown in Fig 9.3 (b).
weld profiles. Undersized welds can be corrected by adding one or more weld
9.3.1.4. Dimensional Defects passes. Oversized fillet welds are not harmful if they do not
interfere with subsequent assembly. However, they are
These include the non-uniform width or height of a weld, t uneconomical and can cause excessive distortion and residual
irregular crests and valleys, and variations in leg length of fillet stress.

I
 '66
Welding Engineering and Technology Weld Defects 667
"

"-
" overlaps are accompanied by lack of fusion, cracks and other
Insufficient Insufficient defects in the overlapped portion of the weld.
throat leg
An overlap tends to produce mechanical notch, parallel to the
weld axis, where stress will build up and start a crack at the sharp
point where the weld metal and base metal come together at the
overlapped area. The built-in cracks are also places where
Fillet weld Fillet weld corrosion often starts which is not seen until it begins to
.' undermine the base metal.
,,~-- .-
....

,.
,~Insufficien
throat
....
Overlaps are not permitted by codes and therefore need to be
chipped off and the weld be ground to proper shape.
Butt weld
9.3.1.6. Undercuts
la I Defective welds
Undercuts are grooves melted into the parent metal adjacent
to the toe of a weld and left unfilled by weld metal, as shown in

Undercut

~
Fig. 9.5. Undercuts in butt and fillet welds.
Fig. 9.5. These are generally located parallel to the junction of
weld metal and base metal at the toe or root of the weld. They
reduce the cross-section of the work and may lead to cracking due
I b I Templates for measuring weld shape and size to notch effect.
Fig. 9.3. Some examples of: (a) dimensional defects in welds, and Undercuts are caused by excessive welding current and arc
(b) their quick detection by the use of templates. voltage, since in such cases the width of penetration is increased,
9.3.1.5. Overlaps and Rolls and that leads to excessive melting of fusion faces. In the case of
fillet welds, undercuts may be caused by the fact that the electrode
Overlaps or rolls are protrusions of weld metal beyond the is shifted more towards the vertical member. UndercutB are often
bond at the toe, face or root of the weld, as shown in Fig. 9.4. They accompanied by overlaps.
In electroslag welding undercuts are mainly due to the wrong
position of the water-cooled moulding blocks. Undercuts are

""ap Fig. 9.4. Overlaps in butt and fillet welds.

nlay be caused by excessive welding current, the wrong tilt of
removed by welding up.
electrode in making fillet welds in the horizontal position, or by
an excessive work angle in welding uphill. In making circum-
ferential seams, overlaps may result when the electrode is offset
too much or too little from the 12 O'clock position. As a rule,
Sometimes undercuts are caused due to arc blow in which case
the remedial action is the use of a.c.
The term undercut is also sometimes used to describe melting
away of the groove face of a joint at the edge of a layer or bead of
weld metal. It forms a recess in the joint face where the next layer
or bead of weld metal must fuse to the base metal. If the weld
metal does not fuse with the metal at the bottom of such a recess
when the next layer of weld metal is applied, voids may be left in
 c.63 Welding Metallurgy and Joint Design
Weld Defects 669
the fusion zone. These voids would be identified as incomplete
fusion. This t.ype of undercut is a subsurface defect and is usually stress raiser
associated with incorrect manipulation of the welding electrode at this point
while depositing a weld bead or layer next to the joint face.
9.3.1.7. UnderfilJJSuckback LR
This is a flaw in which the weld bead has a concave type T

t
contour or lack of fill on the face of the weld or a suckback on the
root of a groove weld. The standard term in both cases is underfill
which is defined as, 'a depression on the face of a weld as shown
in Fig. 9.6 (a) or root surface extending below the surface of the
I a) Overfill in single - and double-V
butt welds.

V(a)
-{ern (6) (c)

Fig. 9.6. Different types of underfill; (u) unfilled weld metal zone
R=MalCimum allowabl<l
r<linforc<lment
(usually 3mm)

at the top of the weld bead ; (b) II suckback of weld metal at the
root of the weld, and (d a suckback of base metal.
adjacent base metal', as shown in Fig. 9.6 (b). Underfill reduces
the cross-sectional area of the weld below the designed amount
and is therefore a point of weakness and potentially a stress raiser
where failure may initiate. I bl fillet weld
Underfill may arise when the operator uses a backi.ng gas
shield or a bed of flux and an increased root opening. An underfill Fig. 9.7. Butt and fillet welds with excessive reillforcement.
is a cause for the rejection of a weldment if it is to operate under
dynamic loads or will be exposed to subzero temperatures in Excessive reinforcement impairs the behaviour of the weld-
service. ment under dynamic loads and is therefore unacceptable. In
Another form of suckback is a depression on the root ride of a automatic welding it may happen if the equipment has been
weld in the case of a partial penetration weld, as shown in Fig. 9.6 improperly adjusted to maintain the desired welding schedule. In
(c). This is caused by the shrinkage of weld bead which sucks back manual welding it may be caused by low welder skill.
the parent material resulting in depression in the surface. There may also be the problem of excessive reinforcement on
Normally an underfill or a suckback is corrected by adding one the root of the weld primarily with open root butt welds; this is

or more additional layers of weld metal in the joint prior to
subsequent processing ..
9.3.1.8. Excessive Reinforcement/Excessive Penetration
Excessive weld reinforcement, as shown in Fig. 9.7, is also
undesirable. It tends to stiffen the section and makes notches that
create stress concentration. Excessive reinforcement is also an
economic waste because it adds to the cost of wdding without
gaining any worthwhile strength. It can be a stress raiser and is
objectionable from an appearance point of view. It is normally
caused by poor fit-up, improper welding technique and excessive
welding current, low welding speed, and type of electrode, etc,
~~.
V~
usually due to excessive penetration in single-sided butt welds as
shown in Fig. 9.8. Excess penetration is often called icicles. It is
Fig. 9.8. Excessive penetration in root run of a groove butt weld.
usually caused because of too wide a root gap or too small a root
face, or if the welding current is too high or when the gas torch
used is far too big for the section thickness being welded. Fig. 9.9

"'-
 (110 Welding Engineering and Technology Weld Defects 671
Undercut Notch eff ect
should be carefully evaluated with respect to service requirement
of the weldment.
9.3.1.11. Unfilled Craters
Such craters appear after a sudden interruption of the arc.
They reduce the cross-section of the weld and may act as source
Unfilled
Lack of fusion crater
Fig. 9.9. A defective butt weld showing the combined occurrenc~ of excessive
reinforcement, excessive penetration, undercut and lack of fusion.
shows the combined occurrence of excessive reinforcement,
excessive penetration, and undercut. The circles indicate the
points of stress concentration. I a I Unfilled c rater in a butt weld
Excessive penetration is particularly troublesome in pipe longitudinal or centre line
welds, where it may cause turbulence and erosion in addition to
restricting flow and is therefore unacceptable.
Unless welding specifications state otherwise assume maxi-
mum allowable reinforcement height to be 20% of the work
thickness or 3 mm whichever is less.
9.3.1.9. Start/Stop Points
These will show up on an exograph or radiograph as an
approximately circular area extending irregularly into the weld
metal, as shown in Fig. 9.10. It could be a type of crater crack that

-.
r--
- - ~"-/

------= - ---"-
~.
----'-""'----.---
~

Fig. 9.10. Schematic representation

---....--...
-
.:::::;::;

- ----
~
d
of stop/start points as seen on a radiograph.
was not properly filled up when tying weld metal into the walls of
• a thick joint. It is caused when the welder lifts the electrode
abruptly from the workpiece instead of smoothly tapering off the
weld.
9.3.1.10. Interruptions
Inadvertent interruptions during welding, though rare, do
occur leading to deterioration of weld'quality by way of shape and
dimensional changes or more seriously in the development of
porosity, lack of fusion, and even slag inclusions. These. latter
defects are discussed in more detail under the next main section
on hidden or subsurface defects.
If such interruptions result in gross bead irregularities they
may result in abrupt change of weld section. Such abrupt changes
of section are potential causes of high stress concentration, and
Underbead' I
Toe
Transverse
I b I Unfilled crater in a fillet weld
Fig. 9.11. Unfilled craters.
of crack formation, as shown in Fig. 9.11 for butt and fillet welds .
In some cases, craters should first be chipped out and then filled
up. In the fabrication of critical components or structures, craters
should be brought out on to finishing tabs i.e., run-out plates.
If no tabs are used, it is necessary to weld up the crater
thoroughly and interrupt the arc at the welded portion of the seam.
Never terminate the crater at the parent metal, otherwise
undercuts may occur.
9.3.1.12. Burn-through
Burn-throughs are the holes burned through the parent metal
in a single pass weld or the root run in multi-run welds. Th8 causes
for their occurrence may include excessive root gap, an insufficient
root face, or excessive welding current coupled with low welding

672 Welding Engineering and Technology
speed. Most often this defect occurs in light gauge material and
in making the first pass in multi-layer welds. Sometimes a
burn-through may be due to a sagging copper back-up plate or bed
of flux. Burn-through welds are a rare occurrence but when they
occur, they are rectified by chipping back and welding up.
9.3.1.13. Surface Porosity
It takes place when the welder uses undried covered electrodes
or electrodes with their shelf-life long expired. This defect lends
itself to a corrective action. In general, surface porosity i.s.an
indication that subsurface porosity may have been in the weld
before it became noticeable as surface porosity.
9.3.1.14. Surface Oxidation
It may happen when welding readily oxidising metals such as
aluminium, titanium, magnesium, and their alloys. Surface
oxidation occurs due to the welding heat unless the surface of the
work near the weld is shielded against exposure to air.
9.3.1.15. Uneven and Coarse Ripples
Uneven and coarse ripples in themselves may appear to have
only an aesthetic value but may often be indicative of inconsistent
solidification pattern which may itself be a consequential effect of
inconsistency in heat input. Thus, this defect may override very
often the effect of int-arnal flaws.
9.3.1.16. Flat spotsIPock Marks
Flat spots are sometimes observed on the weld bead in the
case of gas shielded FCAW and SAW.These are associated with
excessive gas pressure and high welding current, and high
moisture content in the flux. These flat spots, J.r..nown
as gas dents
or pock marks, are also produced when gas is produced by chemical
reactions between the molten flux and the molten weld metal. The
flat spots or gas spots are shown schematically in Fig. 9.12. This
may not appear to be a serious defect, howeve.r it mars the weld
appearance. In case this happens in multi-layer weld it may lead
Pock marks
Fig. 9.12. Pock marks on a butt weld.
Weld Defects 673 -
to lack of fusion at these flat spots which are often in the form of
slight depressions. Proper control of welding current, gas flow
rate, and moisture content of flux can eliminate this defect.
9.3.1.17. Excessive Spatter
Excessive spatter adjacent to the weld is also often a defect
and is unacceptable. This may be caused by arc blow, wrong
selection of electrode, use of bare electrode or heavy welding
current, or the poor technique of welder. Though spatter in itself
may not necessarily be a defect but it impairs the quality of the
metal and the metal surface will not take paint well. Also, spatter
-is indicative of the likelihood of other associated flaws.
Spatter may be prevented by applying a suitable coating (some
of these are available in aerosol form) to the metal adjacent to the
weld.
9.3.1.18. Arc Strikes
This defect is caused when the welder accidently or inadver-
tently strikes the electrode or the electrode holder against the
work, usually adjacent to the weld, causing an unwanted arc; it
can also happen between a faulty ground clamp and work. Such
spots are referred to as arc strikes which can initiate failure in
bending or cyclic loading. They can create a hard and brittle
condition particularly on materials which have a high harden-
ability such as low alloy steels, and medium and high carbon
steels. These are inadvisable even on mild steel when high static
or normal fatigue stresses may be encountered. The repair of such
a damage may be difficult and costly, involving chipping, and
possibly preheating before rewelding in the case oflow alloy steels.
9.3.1.19. Bad Weld Surface Appearance
A bad weld surface appearance indicates a bad workmanship,
whether or not the surface defects are structurally important.
Obvious variations in width of the weld bead, hu:mps in the
surface, occasional depressions, nonuniform ripples, changes in
the height of reinforcement and protrusions i.e. overfilled craters
are all examples of bad surface appearance. Even if they cause no
serious problem in service, they can cost a welder his job.
9.3.1.20. High Low/Off-set
Off-set in a welded joint occurs most often as a consequence
of poor fit-up and is a problem area in a plate or large diameter
pipes and rolled cylinder joint where the edges simply do not
match up, as shown in Fig. 9.13. The flat as well as th.e curved
parts do not match up. It can also result from trying to butt weld
wavy plate. One solution is to use better or more tack welds.

674 Welding Engineering and Technology
Weld
W'/~~
(0 l Lack of fusion or incomplete fusion, shown in Fig. 9.15, may
Fig. 9.13. Offset weld in (a) plate joint, and (b)
Another solution, if the base metal mismatch is caused by heat
distortion, is to use welding jigs and clamps or fixtures that will
hold the base metal in position until the weld is completed.
9.3.2. Subsurface Weld Defects
Subsurface defects are more dangerous than the surface
defects and also they are more difficult and expensive to rectify.
All the subsurface weld defects listed in section 9.1 are
discussed briefly in this section.
9.3.2.1. Subsurface CrackslInternal Cracks
The subsurface or internal cracks are of many types. Some
may be in the weld, some in the HAZ-sometimes called
underbead cracks, as shown in Fig. 9.14. Some cracks are located
at the interface between weld metal and base metal, and some
completely in the base metal. There can also be microsized cracks
as well as macrosized cracks. Sometimes the smaller cracks are
called fissures, or if the cracks are extremely small they are called
microfissures which require special techniques to find them.
Fig. 9.14. Underbead cracks in a fillet weld.
Weld Defects 675
LI_I ...•
All weld cracks are considered dangerous because they create
a serious reduction in strength. They can propagate and cause
sudden failure. They are most serious when impact loading and
low temperature service conditions are involved.
Cracks in welds must be fully chipped back and welded up.
Cracks must never be welded up without chipping back.
9.3.2.2. Lack of Fusion
( bl occur between the parent metal and the weld metal, and also
a pipe joint.
tw£.@,U"OO~~",i~ .
~i,n ~'f fusion
Fig. 9.15. Lack of fusion in butt, ..fillet and lap welded joints.
between the various layers in multi-pass welding. The main
causes of lack of fusion are insufficient welding current, an off-set
of the electrode from the axis of the weld, too high a welding speed,
and improper wearing procedure so that the edges are not melted
thoroughly. Other causes of poor fusion include improper joint
preparation and fit-up (inadequate bevelling and narrow spacing)
as well as improper edge cleaning from slag (after flame cutting),
dirt, rust, scale, etc.
In automatic submerged arc welding, lack of fusion usually
occurs at the beginning of welds. That is why the use of starting
tabs, i.e. run-in plates, is recommended. Sometimes lac~ioffusion
across the section of welds may be due to inadvertent iI1~errl1ptions
in the progress of welding operation ..
Lack of root fusion on pipes is most detrimental as the cold
laps on multi-run welds. Tensile and bend tests during welder
approval tests are used to check and remedy this defect.
Like undercuts, lack of fusion appreciably reduces the
strength of welds and makes welded structures unreliable. It
should be rectified by chipping back and rewelding.
676 Welding Engineering and Technology Weld Defects
677
9.3.2.3. Lack of Penetration \
Lack of penetration or inadequate penetration, shown in Fig.
9.16, occurs most often in the root of the weld and it also becomes

~<:J
Filletweld
.••- - - -/ Lack of
,-, J
Buttweld
-
F:::=-
... flJsion

Fig. 9.16. Lack of penetration in fillet and butt welds.

a built-in crack ready to start running through the base metal,
.•.

weld metal, or HAZ when the weldment is put under stress.
Lack of root penetration is caused by a decrease in welding
current or by an increase in arc voltage and welding speed. In
addition, arc poorly directed with respect to the groove axis and
improper starting of the welding operation after electrode
changing may result in lack of root penetration.
The open area, caused by lack of penetration is a notch inside
the weld at its highly critical point, where cracks are more likely
to start. The open area inside the weld lowers the actual throat
depth or thickness of the weld metal, reducing its true strength.
And these defects are impossible to detect without special non-
destructive test method.
Lack of penetration is not normally accepted on high pressure
containment vessels but, where slight root concavities exist, these
may be permitted. The detl imental effects of lack of penetration
increase with length and width so reducing the total effective weld
area, but a certain amount may be tolerated on lightly stressed
fabrications which are loaded in static mode.
Lack of penetration is critically dependent on its orientation
with respect to the applied stress, being much more detrimental
when transverse to the direction of stress. For pipes, correct
penetration of butt welds is made easier by using consumable
inserts, and by the general use of temporary backing medium.
T
0!/0
~~~/4
O·Smm Max. on 3 mm sheet
0·4mm Max. on '·5 mm sheet
Right
Ie)
Fig. 9.17. Three possibilities of weld penetration in an edge weld:
(a) lack of penetration, (b) excessive penetration, and (c) optimal penetration.
When certain tolerance limits for lack of penetration are
exceeded, the defective portion of the weld should be chipped out
and rewelded. Fig. 9.17 shows the possibilities ofweld penetration
in an edge joint. Fig. 9.17 (a) shows an incomplete root penetration
and is sure to cause weld failure; Fig. 9.17 (b) shows excessive
penetration which could possibly be due to burn-through, frem too
low a welding speed.,or from using a filler rod that is too large for
the section thickness and the type of joint involved. Fig. 9.17 (c)
shows the,jdeal edge weld with a reinforcement of 25% of section
thickness T, and a small reinforcement in the root area.
9.3.2.4. Blowholes/Subsurface Porosity
Blowholes arid porosity are voids, holes or cavities formed by
gas trapped by the solidified weld metal. The sources of trapped
gas may be : ./ .#

It may be noted that permanent backing rings in pipes have (i) rust, dirt, grease, paint or primer on the edges of the
proved to be some of the most frequent causes of fatigue failure parent metal or on the electrode;
as they are one of the worst stress raisers. (ii) damp SAW flux;
(iii) impurities and moisture in the shielding gas;
 Weld Defects 679
678 Welding Engineering and Technology
.. '
(iv) excessive welding speed because of which the g~s en-
.. '
velope around the weld pool becomes ineffective;
(v) a wrong grade of electrode wire, especially in CO2
...
shielded arc welding;
(a) Uniformly distributed \ porosity
(vi) welding dver tack welds made with lightly. coated
electrodes; .---- .... . ....
.........
......
(vii) high welding currents; because high weldIng currents
will burn up the deoxidising elements in SMAW
electrodes at such a high rate that not enough
....... ...
... .• •.
.• ""
• ..

deoxidisers are left to combine with the weld metal (b) Cluster porosity
before it starts to cool thus causing porosity and
---I
.... 1-
blowholes. Distance equal to 1 electrode
The gas pockets can occur as extremely large holes (blow- ........
'.....
.... --

.
holes), several mm in size, in the weld metal or extremely small -. ....
.. ,,"
..
....
.• " ..
holes (porosity) few hundredths of a mm, scattered throughout the ........ "

.
"'

cross-section of a weld. Some types of blowholes reach the surface ... ,,-
"
.. "
"
" .,,"
and appear as pinholes. Some types of porosity are called
wormholes when they are long and continuous. Others are called Direction of welding -
I Porosity sometimes becomes finer
piping, usually long in length and parallel to the root of the weld. as electrode dries out I
Normally porosity is not as serious a defect as cracks primarily
because porosity cavities usually have rounded ends which .will ( c) Starting porosity
not propagate like cracks ..
Porosity in welds is usually of 4 types, viz.,
(i) Uniformly scattered porosity; ..............
(ii) Cluster porosity; I 0000 ° •••••••• • 0 •• I
(iii) Starting porosity;
(iv) Linear porosity. ( d) Linear porosity
Fig. 9.18. Schematic representation of different types of porosity in weld metal.
Uniformly Scattered Porosity looks like scattered voids or
tiny bubbles in the weld metal as shown in Fig. 9.18 (a). They look associated with excessive moisture in the electrode covering. As
like black spots on an X-ray negative. The individual voids seen the electrode gets heated up, the moisture evaporates in the
on the X-ray film can range in size from almost microscopic to 3 remaining covering and thus the porosity drops offin the direction
mm or more in diameter. of welding. When a new electrode is used, the pattern repeats
itself.
Cluster Porosity shown in Fig. 9 .18(b) is usually worse type
of porosity in which pores frequently occur in clusters separated Linear Porosity as shown in Fig. 9.18 (d) occurs most
by considerable lengths of porosity-free weld metal. Such clusters frequently in the root pass of welds and is often regarded as a
are often associated with changes in welding conditions, for special case of inadequate joint penetration. Any defects that line
example changes in arc settings when the welder starts or stop up in a weld are bad and when they occur in the root of the weld,
welding. they are worse. This is because the lined up pores can get
Starting Porosity occurs when stick electrodes are used. In connected together V6ryquickly by short cracks to become one big
this case one gets cluster porosity spaced at distances just about disastrous failure when the welded component is heavily stressed
during service.
equivalent of the weld metal that one stick electrode would
,deposit; as is shown in Fig. 9.18(c). This type of porosity is usually
 680 Welding Engineering and Technology
Porosity is not permitted in the fabrication of pressure and
vacuum vessels nor for containers intended for storage and
transportation of liquid and gaseous products. However, virtually
all welds in aluminium and its alloys have some form of porosity.
Linear porosity along the fusion face often indicates lack of fusion.
and is not permitted. When the porosity exceeds that acceptable
by the code it must be chipped out and rewelded. Care mtlst also
be taken when depositing mechanised welds on to root runs laid
out by SMAW.
The single best way to avoid porosity is to use perfectly clean,
dry welding equipment and electrodes that have either been just
baked in an electrode baking oven (usually at 150°C for 4 hours)
or by using newer moisture-resistant SMAWstick electrodes with
hydrophobic (water-hating) flux coatings. Also, always purge the
shielding gas lines before welding. Excessive welding current and
too long arc lengths should also be avoided.
9.3.2.5. Shrinkage Cavity
A shrinkage cavity is defined as a cavity normally formed by
shrinkage of weld metal during its solidification. It seldom occurs
and that also only when thick plates are welded by arc processes
and the weld pool contains a large amount of molten metal as in
SAW, ESW, etc. In such a case a large cavity may be formed at the
surface which may extend inside the weld. It is also known as
contraction cavity.
9.3.2.6. Slag Inclusion
Slag or non-metallic inclusions refer to solidified molten flux
or electrode covering comprising oxides, sulphides, phosphorus
compounds, and nitrides entrapped in the weld metal and parent
metal which fail to float out to the surface. These inclusions can
be continuous,intermittent, or very randomly spaced. In general
~ heavy current is used. Tungsten inclusions can also be caused by
( a I SI og . 1uSlon
me . in a butt SloQ
weld
(b) Slog inclusions in weld metal
Fig. 9.19. Slag inclusion: (a) in a butt weld, and (b) cross-section
of a weld metal with slag inclusions.
~
Weld Defects 681
the slag inclusions are rounded and unlike cz:acks"donot possess
sharp corners, and for that reason are not quite as dangerous as
cracks. Fig. 9.19 shows the slag inclusion and its orientation in
the weld metal along a cross section.
The most common cause of slag inclusion is inadequate
cleaning of weld metal between passes; as a result slag gets
permanently entrapped in the weld.
During SMAW,slag can also be formed and forced below the
surface of the molten metal by the stirring action of the arc. Slag
may also flow ahead of the are, causing the metal to be deposited
over the slag. Also, with some types of electrodes the slag in
crevices of the previously deposited weld metal will not remelt and
thus gets entrapped in the weld.
Slag can also be present in the molten weld metal for other
reasons such as high viscosity (stiff) weld metal that is not hot
enough to flow properly, rapid solidification; or too low a preheat
temperature, that prevents the slag before the weld metal
solidifies. Gas dents or pock marks may also collect slag.
Slag inclusions can be prevented by thoroughly cleaning the
weld metal between passes. Also, use when available, electrodes
that produce re-fusible slag. If the slag is very sticky and hence
difficult to remove, grind or back-gouge the surface of each pass
before making the next weld run.
Slag inclusions reduce the mechanical strength of welds;
therefore welds with slag inclusions should be chipped back and
, rewelded.
9.3.2.7. Metallic Inclusions
These are mostly tungsten inclusions from the tungsten
electrode in TIG welding, or copper from the contact tip when MIG
welding. Tungsten inclusion occurs when TIG welding is done with
electrode positive, and with a very heavy current either because
the tungsten electrode is dipped into the molten weld pool or
because of overheating and melting of tungsten electrode when
incorrectly sized or shaped electrode, and oxygen contamination
of the shielding gas. Tungsten inclusions are extremely brittle and
fracture readily under stress and can thus act as failure origins.
Copper inclusion can be caused by inadvertent dipping of GMAW
gun nozzle in the molten weld metal. Sometimes it may aleo be
caused by melting of copper backing strip, when used. Both these
inclusions, in steel welds, are readily identified on radiographs as
light areas against the darker background.
 (i82 Welding Engineering and Technology
Weld Defects 683
Copper inclusions are not normally permitted in the weld
either in Al~alloys or steels. Tungsten is not as detrimental as chart as per ASME Division 1 Pressure vessels code is shown in
copper in aluminium alloys and may be permitted in isolated Fig. 9.21.
pockets but may cause problems in carbon/manganese steels and Table 9.1. Recommended Acceptance Levels
low alloy steels due to its embrittling effect. H.F. (high frequency) for Arc Welding Defects
or similar arc striking and correct current and voltage settings are (BS 4871:Part 1:1974 and BS 5500:1976)
essential to avoid the occurrence of this defect.
As wormholes, isolated.
9.3.2.8. Finger Penetration
u vv f-.Cracks
Lack
Lack Surface
* (or
of
ismm.
Shrinkage
therefore
3ellDefect and
1.5
root
type
notlamellar
cavities
inter-run
and
of3penetration
root grooves
mm.penetration
fusion tl-.!J
tears
infusion
* a50and
The
Permitted
mm.
including
*forlinear --
75 mm.shall blend with
reinforcement
maximum --
"
"t:l
v I •..
•..
G "t:l
«l "'" Q)
«l t110,
6mmw*
50
root
Excess
Lack
localised
Uniformlymm
t/4 pores
pores
and
and
concavity
of side -
including
Occasional
mm max.
pro
porosity
fusion
distributed
ell local
rata
3.0 h*
mm.
slight
2%
root
S
Not
mm light
Slight
Linear
Z:t
As
or
ell Not
by
forexcess
depth
4.5 permitted.
area
concavity
t6.0
up is
intermittent
greater
mmdoes
permitted. (as
porosity
mm seen
toare
fornott over
and
for
p~rosity.
allowable. in 50
exceed ato75
permitted
tshrinkage
undercut
parallel
over radiograph)
1.2
mm provided
theisgrooves
mm.
mm up
permitted
axis for
. toof and
the
§ Reinforcement shape Po ...
.~ individual
Co)
e
'S;
thicknesses. ~
inpermitted. group) weld
penetration
may indicate lack ofshape
fusion or lacknot
of
0:: testing
required techniques.
provided the does
Finger penetration, shown in Fig .. 9.20 is a defect the ~aligned
'" , sense Undercutporosity interfere
parent
provided metal
Isolated with
the and
the specified
dressing
depth does non-destructive
is notexceed
not normally
0.4
Overlap
~ Wormholes, isolated
Wormholes, '" Crater
Linear misalignment
pipes
that the weld penetration profile may not be proper or adequate

~~
Fig. 9.20. Weld bead with finger penetration.
to give the desired weld shape. This defect is usually encountered
when plasma welding is employed using keyhole welding
technique or MIG welding using Ar-shielding and high welding
current.
The defect is usually avoided by using Ar-He gas mixture
instead of pure argon as the shielding gas or alternatively, use the
conventional melt-in technique rather than the keyholing
technique ofwelding; however that means the welding current will
have to be set at lower levels.
9.3.2.9. Coarse Microstructure
Coarse microstructure in welds can impair the strength and
impact properties of a weldment affecting its performance ip.
service. Coarse microstructure is caused by excessive heat input
in welding and is, therefore, mainly associated with high heat
input welding processes like ESW, gas welding and single run
SAW when employed with high current settings. This defect can
be rectified by suitable PWHT (postweld heat treat~Emt). In case
PWHT is not practicable due to the size and geometry of welded
component involved, it is best to use multi-run lower heat input
welding procedure so as to normalise the previously laid weld
runs.
9.3.3. Acceptance Levels of Arc Welding Defects
Recommended acceptance levels for different arc welding
defects as per the British standard BS 4871 : Part 1 : 1974 and
BS .'5500: 1976 are as listed in Table 9.1 while the typical porosity
(contd,)

"'--
 684
----------------------------•
Dimension
-
. .
Welding •.Engineering• and Technology
•. - .. • .,
~ . Assorted
, . •I 51
Weld..• Defects .,
... ..•......•. .-.- •. -
..-
• • ..
Largl!
•• . No. of pores 685
~
Defect type Permitted maximum
Individual and parallel to t* 18 mm t> 18:1- 75 mm
0-055 I
weld axis 1* t/2:1- 6 mm 1:1- t/3

I
~1.5 m~_ w:l- 1.5 mm ~.J~ m
""
Aggregate length shall not exceed 8% of
length of group, which in turn shall not
:§ •51 rn.~ gronpt. _ exceed 12t in length .
1- -~ As isolated pores.
~ I Individual and t'andomly
U') oriented (not parallel to
OJ
weld axis)
1]
OJL Non-linear group
Tungsten inclusions
-----
As uniformly distributed or localised porosity.

isolated
I grouped
L _ Copperinclusi~ As isolated pores.
Notul.1ifonnlydistributed
permit_te_.d_,, ~
or localised_ porosity.
Abbreuiations used:
t = parent thickness. In the case of dissimilar thick- 1 := length of defect
ness t applies to the thinner component. h = height of defect
w := width of defect. 4>= diameter of defect
t Individual inclusions within the group should not exceed the sizes for isolated
pores (or individual pores in a group). A linear group is defined as a number of Medium
inclusions in line and parallel to the weld axis where the spacing between their
adjacent ends does not exceed 6 times the length of the longest inclusion within the
group. \Vith parallel groups, all inclusions count towards the aggregate.
9.4. WELD DEFECTS IN OTHER THAN ARC WELDING
0-039
.. .. 100

PROCESSES
"
.,
Although arc welding processes are the most used processes
for fabrication throughout the world but there are a large number
Fine
of other welding processes which are used in specific indust.ries
and they have their own peculiar defects. Major defects in welds Fig. 9.21. Typical porosity chart for: (a) assorted, (b) large,
made by the following widely used processes are discussed briefly (c) medium and (d) fine porosity.
in this section. (iv) Cracks,
(i) Resistance welding processes,
(v) Surface bur_ling,
(ii) Friction welding,
(vi) Electrode skidding,
9.4.1. Resistance Welding Defects (vii) Copper pick up,
Amongst resistance welding processes, spot welding, flash and (viii) Sheet separation,
high frequency induction welding are very widely used and (ix) Excessive penetration,
therefore the defects discu8sed in this section pertain mainly to (x) Edge bulging,
these three processes. (xi) Mi.splaced and mis-shaped nuggets
9.4.1.1. Spot Welding Defects Brief description of these defects follows.
The commonly encountered defA.ctsin this weldiilg process are: (i) Excessive Electrode Indentation
(i) Excessive electrode indentation,
Ui) Metal expulsion,
Excess electrode indentation appears as a deep depression of
the weld surface, opposite to the weld nugget, as shown in Fig.
(iii) Shrinkage, voids and porosity,

•..
686 Welding Engineering and Technology
K«««(««(~:~ «(<««««1
lod,olo\;oo
I a) Excessive indentation I ,Schematic)
Fig. 9.22. Excessive electrode indentation in resistance spot welds:
(a) schematic representation, and (b) a photograph showing
two spot welds with excessive indentation.
9.22. Some indentation is unavoidable since vertical expansion of
the weld metal is immediately forged out by the pressure of the
electrodes and also shrinkage of the weld metal is unavoidable.
The limits of susceptibility generally range from 10 to 20% of sheet
thickness in which the indentation occurs. Deep indentation is
caused by excessive electrode pressure and high welding current.
It results in a loss of strength due to a reduction of metal thickness
at the periphery of the weld area. This defect is particularly
observed with materials having high ductility.
Hammering of the electrodes with excessive force is sometimes
observed to lead to excessive indentation particularly in more
ductile material like aluminium.
Such welded assemblies where this indentation is deep are
fmished by filling the indentation by m-seal or Al metal set
compound followed by painting to improve the final appearance of
the assembly.
The electrodes should be set as close as possible to reduce
'hammering' of the electrode tips. ThIS also prevents the possibility
of fingers being placed between the electrodes.
Toprevent excessive indentation it may be necessary to correct
one or a combination of the following factors.
Weld Defects 687
I (i) Excessive force,
(ii) Improper forging cycle,
I (iii) Improper electrode contour,
(iv) Weld metal expulsion,
1 (v) Excessive surface heating,
(vi) Excessive welding current.
(ii) Metal Expulsion
Weld metal expulsion or splashing as shown in Fig. 9.23 is a
condition in which the molten metal from the nugget-forming zone
Fig. 9.23. A splashed spot weld showing the expelled weld metal.
is expelled by puncturing the corona bond forming the periphery
of the nugget. The flow of current through the nugget causes
turbulence in the liquid metal and if the pressure inside the
nugget is high and the corona forming it is weak, it will result in
splashing. Metal expulsion, thus, puts a top limit on welding
current and is promoted by low electrode force, bad fit-up or lack
of mechanical support as a result of welding close to edge. That is
why a minimum edge distance of (1-1.5)d is recommended; where
d is the diameter of weld nug'get.
It is also possible for expulsion to occur at the .electrode-
to-work interface, because of too rapid or excessive generation of
heat at the interface, when scale, for example, has created a locally
high resistance. Low resistivity metals are prone to surface
splashing because electrode-to-work resistance is a high
I proportion of the total resistance.
I~ To avoid weld metal expulsion, the electrode force and contour
must be sufficient to provide a pressure seal around the weld
nugget and the faying surfaces must be free from contamination.
I
tS88 Welding Engineering and Technology
Weld expulsion is normally accompanied by deep surface inden-
tations, sheet separation, voids, and loss of strength. To avoid weld
metal expulsion. it may be necessary to correct one or a
combination of the following factors.
(i) Insufficient force,
(ii) Excessive welding current,
(iii) Foreign substances at the faying ,surfaces,
(iv) Erratic contact .resistanc.e,
(v) Electrode misalignment,
(vi) Electrode skidding.
(iii) Shrinkage, Voids and Porosity
Spot welds in metal thicknesses of about 1 mm and more may
have small shrinkage cavities in the centre of the weld nugget as
Fig. 9.24. Porosity in aluminium spot weld.
illustrated in Fig. 9.24. These cavities are les8 pronounced in some
metals than in others due to the difference in forging action of the
electrode on the hot metal. These will have hardly any detrimental
effect on the static or fatigue strength of a weld if they are located
entirely in the central portion of the weld nugget where stresses
are essentially zero. However, the cavities that result from heavy'
expulsion of molten metal may take up a very large part of the
fused area and must be considered a defect. A certain number of
expulsion cavities are to be expected in production welding. This
defect is caused by low electrode force, high welding current or
any other condition:;;that produces excessive weld heat. When this
occurs, the weld nugget will not be properly forged during
solidification and shrinkage cavities occur.
I
Weld Defects 689
(iv) Cracks
Internal and external cracks in spot welds, generally caused
by improper thermal and electrode force conditions, are observed
in the weld structure and surface respectively. Cracks on the
surface of the metal lower the corrosion resistance of weldment.
To eliminate or minimise the occurrence of cracks it may be
necessary to correct one or a combination of the following
conditions.
(i) Excessive penetration,
(ii) Insufficient force,
(iii) Delayed application of forging pressure,
(iv) Improper rate of current rise,
(v) Improper electrode cooling,
(vi) Electrode skidding.
(v) Surface Burning
Fig. 9.25 shows surface fusion in spot welds. Surface fusion or
burning may be described as the melting of the metal surface in
Fig. 9.25. Surface fusion or burning in aluminium spot welds.
contact with the electrodes. Weld quality and appearance are
adversely affected when surface fusion or burning persists.
Surface fusion is usually accompanied by deep electrode
indentation.
Surface fusion results in the production of undersized welds
due to heavy expulsion of molten metal, increased cost ofremoving
burrs from the outer surface of work, short electrode life and loss
of production time from more frequent electrode dressing.
Parts must be welded just immediately after chemical cleaning
instead of delaying in taking up the job. Because of such delay,
690 Welding Engineering and Technology Weld Defects 691

scales and oxides will again form on the metal surface in the case may be caused by one or more of the following factors and
of AI, Mg, and their alloys; that may necessitate cleaning again necessary remedial action must be taken to avoid this occurrence.
leading to enhanced costs. To prevent surface fusion it may be (i) Improper choice of surface preparation,
necessary to correct one or a combination of the following (ii) Too frequent electrode cleaning,
conditions. (iii) Condensation of electrode tip,
(i) Dirty or improperly cleaned electrodes, (iv) Excessive welding current,
(ii) Improper surface conditions, (v) Insufficient welding force.
(iii) Improper electrode cooling, (viii) Sheet Separation
(iv) Improper electrode material,
With the spread of heat in the later stages of the weld cycle
(v) Inadequate welding force,
the electrodes begin to sink into the work surface as a result of
(vi) Excessive wel~ time, plastic deformation the sheets begin to separate at the weld edge,
(vii) Electrode skidding, as shown in Fig. 9.27. These effects viz.; indentation and sheet
(viii) Misalignment 'of work. separation set a top limit on current and time.
(vi) Electrode Skidding Spot
Sometimes the electrodes skid over the work surface causing weld
skid marks and impairing the appearance and quality of spot

~~(-~(~(~~
Fig. 9.27. Schematic illustration of excessive sheet separation in spot welding.
Excessive sheet separation is an undesirable condition
because it increases the bending moment on the joint when
stressed in shear. Generally, greater separation is experienced
with the softer materials. To remedy this condition it may be
necessary to correct one or a combination of the following factors.
(i) Excessive electrode force,
(ii) Improper fit-up of parts,
(iii) Incorrect electrode contour,
(iv) Tip misalignment,
Fig. 9.26. Spot weld appearance with electrode skidding.
(v) Excessive weld time,
welds, as shown in Fig. 9.26. Electrode skidding is found to be (vi) Excessive welding current.
m.ure }Huminellt when double bent electrodes are used or the (ix) Excessive Penetration
throat depth is increased with extended horns or both.
This defect can be avoided by using more rigid horn extensions The extent of the nugget or fusion zone into the thickness of
a joined member is called penetration. The weld penetration
and providing jack or support to the lower horn whenever possible. should not be less than 20% nor more than 80% of the sheet
As far as possible straight electrodes should be preferred to bent thickness. To achieve the desired penetration it may be necessary
electrodes.
to correct one or more of the following factors.
(vii) Copper Pick up (i) Excessive welding current,
Copper pick up from the electrode tip by the weld increases (ii) Excessive weld time,
the corrosion susceptibility of the spot weld. This copper pick up (iii) Improper surface preparation,
 Welding Engineering and Technology Weld Defects 693
692
9.4.1.2. Flash Welding Defects
(iv)Improper electrode cooling,
Insufficient electrode force,
(v) Some of the defects encountered in flash butt welding include
(vi)Insufficient electrode contour. the following.
(i) Cracks
Fig. 9.28 shows the variation in weld penetration as a result (ii) Die burns
of variation in heat input.
(iii) Oxides and other inclusions.
Brief description of these defects follows.
(i) Cracks: Hot and cold cracks are encountered in flash
welding depending on the temperature of formation. Cold
cracking, as indicated in Fig. 9.29 can be caused by a combination
Cold crack

Fig. 9.29. Schematic representation of a cold crack in flash butt weld.

Fig. 9.28. Variation in nugget size of spot welds as
a result of variation in heat input.
(x) Edge Bulging or Distortion
Distortion or edge bulge is a condition that occurs when welds
are placed too near a metal edge. The edge distance must be great
enough to provide adequate mechanical strength to resist defor-
mation and conduct heat to reduce edge cracking.
Also, if the electrodes close with heavy impact, sliding of
electrode may take place resulting in edge bulging and production
of non-circular shaped spot welds with deep indentation.
Edge bulging can be minimised by controlling one or more of
the following factors.
(i) Excessive welding force,
(ii) Insufficient edge distance,
(Ui) Electrode skidding,
(iv) Improper fit-up of parts,
of excessive upset and improper temperature profile. Excessive
cooling rates in hardenable steels can cause cold cracking from
intolerable strains acting upon martensitic structures. The most
common form of hot cracking in upset welds occurs as
microfissures in the heated zones, and is known as break-up.
(ii) Die-Burns: In flash welding a form of intergranular
oxidation known as die-burn can occur at clamp locations. This
defect is caused by localised overheating of the workpiece where
it is held in the clamping die. Proper cleaning of the workpiece
surfaces in the clamping area will usually eliminate this prol:-lem.
Too large an initial die opening can result in excessive metal
temperatures near the faying surfaces during flashing. This may
,
I lead to intergranular oxidation as well as non-uniform upsetting
I,
and joint misalignment. It is, therefore, necessary to use suitable
\
I die opening.
! (Ui) Oxides and Other Inclusions: These are caused by the
t entrapment of oxidised metal in flash welding and can be
I;

I prevented by increasing the flashing time, welding current, and

(v) Weld metal expulsion, the forging pressure. Proper cleaning of the faying surfaces before
(vi) Excessive welding current. welding can eliminate this defect.
(xi) Misplaced and Mis-shaped Nugget I In flash welding craters are also sometimes formed on the
These conditions can be avoided by proper heat balance and faying surfaces by the expulsion of molten m~tal during flashing.
aligning the properly shaped electrodes.
I

\yelding Engineering and Technology
tt94
If the flashing voltage is too high or the platen motion is incorrect,
violent flashing causes deep craters on the faying surfaces. Molten
metal and oxides may be trapped in these deep craters which are
not expelled during upsetting.
9.4.1.3. Defect in High Frequency Induction Welds
Pipes for high pressure service, pipes for well castings, and
structural beams have been welded by the high-frequency
induction welding process. A peculiar defect associated with such
pipes is called Hook Crack.
Hook cracks are difficult to detect and rAquire extremely
careful inspection for assurance that they are not present.
Delamination of the pipe wall just outside of the narrow weld
zone can be the cause of hook cracks. The slight upset of the metal
at the weld acts to turn the ends of these usually very small
subsurface cracks in the direction of the metal flow (Fig. 9.30).
The"hook"has turned outward to follow the direction
of metal flow in the outer portion of the upset weld
lone.
Fig. 9.30. Schematic illustration of a Hook Crack in a pipe caused by
pipe-wall delamination after High Frequency Welding.
Because the metal flow of the upset is toward both surfaces the
"hook" of the hook crack may turn toward either the outside or the
inside surface, depending on which is closer to the delamination
and in which direction the greatest amount of upset. occurs.
Generally, upset is greater toward the outer surface, thus causing
hooking to turn outward. Use of the cleanest steel ayailable and
the lowest possible heat input will greatly reduce the possibility
of hook cracks.
9.4.2. Defects in Friction Welding
Friction welding is a solid-state welding· process which is
employed for large scale production of weld joints between bar-like
components or between bar-to-plate by mechanised mode of
Weld Defects 695
operation. Some of the defects encountered in welds made by
friction welding are :
(i) Lack of bonding,
(ii) Non-metallic inclusions,
(iii) Cracks,
(iv) Intermetallic phase accumulation, and
(v) Unusual weld flash.
Brief description of each of these defects is included in this
section.
9.4.2.1. Lack of Bonding
This is a defect in which there is insufficient bonding between
the faying surfaces as shown in Fig. 9.31. This is usually caused
Fig. 9.31. Lack of bonding in a friction welded joint.
by low rotational speed, insufficient time period for third stage of
welding cycle i.e. platic deformation or even due to improper
preparation of the faying surfaces. It can therefore be prevented
by optimising the welding variables and by improved surface
preparation.
9.4.2.2. Non-metallic Inclusion
This defect is caused by solid inclusion(s) of non-metallic
foreign matter in contact area, as shown in Fig. 9.32, .. The
noo- metallic
inclusions
Fig. 9.32. Non-metallic inclusions entrapped between the
faying surfaces during friction welding.
inclusions are entrapped during the forging stage and may include
scale, rust, cutting solution, drawing grease, etc. on the faying
surfaces. If there is a central bore in the workpiece(s) then it may
be caused by soiled centre bore as well. The remedial action

Welding Engineering and Technology
(b)
(c )
Fig. 9.33. Schematic representation of cracks in friction welds:
(a) internal cracks, (b) crack at the sharp-edged transitional area to weld flash,
and (c) cracks inside weld flash, in axial direction.
(iii) at the sharp edged transitional area to weld flash, as
shown in Fig. 9.33(b);
(iv) inside weld flash, in axial direction,as shown in Fig. 9.33 (c).
The cracks in the HAZ and marginal area are caused by the
formation of coarse carbides in these zones. This can be prevented
by affecting the quality of carbides by preheating the workpieces
to reduce the cooling rates after welding.
Internal cracks are encountered in weldments of dissimilar
component materials and can be eliminated only by controlling
the behaviour of intermetallic compounds at the faying surfaces.
Weld Defects 697
Cracks at the sharp-edged transitional area to weld flash are
caused by the use of too high a forge pressure. Thus, the remedial
action involves reducing the forge pressure and adjusting other
welding variables ..
Cracks inside weld flash, in axial direction are also caused by
too high a forge pressure and generation of inadequate heat during
the process. To prevent such cracks it is required to increase the
rotational speed and control the weld cycle time to generate
sufficient heat before applying the final forge pressure.
9.4.2.4. Intermetallic Phase Accumulation
There is a possibility of inter metallic phase accumulation [Fig.
formation or depletion with the contact area. The main
9.34 (a)],
I a) Intemetallic phases inside contact areas
I
I
rI
I b) Intermetallic phases along contact areas
Fig. 9.34. Accumulation of intermetallic phases in friction welds.
causes for this occurrence are incorrect welding variables and
dissimilar component material. The preventive measures include
the proper selection of component materials and optimisation of
welding variables like rotational speed, cycle time and the forging
pressure.
Sometimes carbides, oxides, and nitrides may also accumulate
in the bonding zone along the contact area as shown in Fig. 9.34
(b). This is caused by improper material preparation. The remedial
actions to prevent such an occurrence include greater homogeneity
of component material being welded; proper selection and
adjustment of welding variables to achieve the desired welded
joint.
(j<)S
Welding Engineering and Technology ----
Weld Defects 699

9.4.2.5. Unusual Weld Flash

Sometimes unusual flash formations are observed in friction
welding, as shown in Fig. 9.35. Fig. 9.35 (a) shows flash material
__ Flash

(0) Off-line weld

Fig. 9.36. Off-line weld made by improperly aligned electron beam.

the proper beam alignment and prior trial rw1.Sbefore commen-
cing the production run.
Another defect peculiar to EBW is Bursts. A burst is a
(b) disk-shaped solidification defect that occurs at the electron beam
weld centreline. It. is usually found near a change in mass, for
example, near corners or near adjacent lugs or stiffeners. Fig. 9.37
shows the shape, location and appearance of a burst.

Ti-6AI-I.V. {. of wel~ rPlate 2Smm

(c)

Fig. 9.3'5. Unusual flash formations in friction welded joints.

extruded from between the two friction welded parts, along the
Runoff
entire circumference of the weld joint. Fig. 9.35(b) shows flash tab
material spirally extruded from between the faying surfaces in.one Fig. 9.37. Shape, location, and appearance of a 'burst'
place; the occurrence is usually irregular. Fig. 9.35(c) shows flash in an Electron Beam Butt Weld.
or spiral extrusion conti.nuous along the entire circumference of Bursts can be prevented by reducing the travel speed as the
the weldment. beam nears the change in mass or by welding on the runoff tabs
No definite reasons are known for the occurrence of these to reduce the change in mass.
defects; most often these defects are observed with copper compo- Although, as mentioned above, it is not possible to discuss all
nents. types of defects occurring in different welding processes; however
9.4.3. Defects in Welds of Other Welding Processes table 9.2 gives the consolidated picture of the major defects
Apart from the defects discussed for welds made by arc encountered in all significant industrially important welding
welding, resistance welding, and friction welding, the welds processes.
produced by other processes do have their peculiar defects but it
is obviously beyond the scope of the text to include all such defects.
However, mention may be made of off-line welds (Fig. 9.36) which
are produced while welding with fine narrow beams in Electron
Beam Welding and Laser welding. This is caused by improper
alignment of the electron beam or laser beam with respect to the
interface between the faying surfaces. Such welds are obviously
not acceptable and are therefore rejects. The remedial action is
 ,_,~ii<'S,-:"'(.~c~"''''j];iiffi;;::"",,,''''''_rl'':,~~' ,,_-~_-~~-~----

Welding Engineering and Technology ·"i

700
'" X X X X X X
..., xxxxxx xxx XX
XXXXX.XXX
~
\ I:..)

xx
xx xxxx
.., :::;.:
.\ iO §-
"">:l

;:;
xxxx
Weld Inspection TEN
\
and \ Quality Control

~ 10.1. INTRODUCTION
~ ~ ~.~ x x
.•.. x x
=
.... ~
IZI
~"" -g ~ e x x x x x x x The quality of welding is imperative t.o the reliability and
I;J IZI '"- ",,'" durability of weldments but is not always that .this quality can be
oIII III Q ~.::e:

<g
~
-~l
"
~..:; §
-g .., safe guarded in welding. This is where weld inspection comes into
the picture, especially when welding is done by machines. Thus,
~~ the principal objective of inspection or quality control is to assure
= ~ the high quality of welded structures through the careful exami-
o =
e .... .::i

-a§ nation of the component/parts at each stage of fabrication.

a~
o ~ E: '(;l
x x x x x x x x xxxxxx xxx xxx xxx xxxxx
Apart from the quality of welding, the quality of a weldment
U~ ~
E:
.:::!
depends on certain other factors ~iz.,
~d
Q ~
•....
(i) whether or not a given design is convenient for welding,
.s~~ '"
.: x x the magnitude of residual stress and strain set up by
-
100

~~ ~.~ x x x x x welding heat; and the layout of welded joints from the
U)- :::;

<.> view point of load distribution;

C'i~ .S
ai.g
~ ....
(ii) the quality of the parent metal, electrodes,welding wire,
- il= x x x x flux, and shielding gases, etc.;
.€ x x x
2 x x x x x x x (iii) the quality of work preparation and fit-up.
~
~ In short, the 'quality of welded joints depends upon the
performance of welding equipment, welding procedure followed,
and the skill of the operator.
The inspection personnel must, therefore, ensure that,
OJ) (i) the parent metal, electrodes, welding wire and flux com-
'"
'"
.., ~ .5 OJ) ply with appropriate specifications;
<.>
b.O:; OJ)b.Oi1 ~.S b.O
50
~~.5~~:;~ ~~u ~OJ)
(ii) the joint is prepared and the work fitted up properly
C. .5
.:
l>O
OJ)_
.5 ~ '-' ~ ~
~ b.O ~ "0 "0 <.> ." .- .-
i1 :;:l
b.O
~:.a.S ~ d'.
.5
"0 '" OJ)
with particular reference to whether,
-0.., S ._ :<J ~ ~ ~ ~
._".i1
~
."' c .-
b.O ~
(a) the components are formed and dimensioned
"0 '" ~ ~ b.O"O ~ - "0 .-

:§ ~ ~ «l u OJ 0; ~ 0 «l 3! b.O~'" ~ ~ :; 3! b.O.S - e ~ ~ ~ .-
b.D ~ «l C '" ; ~ ~ ~ ~ ~ ~ c'" Cl) c Ul b.D~'" C "0 ~ «l'" ~ ." 3! within the tolerance limits in accordance with the
~ c '-' ." ro ~ ••.. -'- ~ C." «l C ~ ~ .-'Ql'" ." 6 :s""
:; •...1l
_""bJ:>b.D~~'"O""
"it, -;;:; "0 6 ~ 8c:C:O<1><1>
8 C 3! 3! C0 ~>'lo-oa.>-C:oQ,)
~ OJ):; C 3! ~.~c:: .at::Q,)mO....,;t
«l ~ C ~ blueprint specifications;
C ." C ~ «l .~ ~ ~.- <G ~.9·- ~ C 0 8.a 0'-'-
t l3 g E
~ «l '" 0 ." "0 '"

.; 6 8 B S 15 .~ ~ .~ .a <G ~ ;; 2~ £ ~·3~ 13
~]:a; t 000...
~ ~~ -0uow~~;::l
~ ~ z.~== ~ ~ ~ $
OJ ~ "0 ~

B ~ -g ~ gj gj 8' ~ ~
cno...cnoo~cnU ~~o...~;::lo... w..Jw.5E-\ "0
10! ____________ W
__el~ing Enginel~rin~nd Technolo~[ Weld Inspection and Quality Conh'ol
703
(b) the parts to be welded are spaced properly at the
joint; Low power magnifying glasses are helpful in detecting minute
(c) the fit-up produces no initial stresses; and flaws. Indicators such as contact pyrometers, and crayons (thermal
(d) the fusion faces are free from dirt, rust, grease, etc. temperatures.
sticks) should be used to verify correct preheat and interpass
(iii) the operator uses the correct welding procedure, includ-
ing welding current, electrode size, welding sequence For quantitative assessment of the groove angle, root gap,
and number of weld runs, arc length, welding speed, misalignment prior to welding and measurement of weld
inter-run deslagging, chipping back when a sealing run dimensions after welding can be achieved by using standard
is applied, etc. measuring tools to make various measurements, such as joint
Once the component or structure is fabricated the welds may geometry and fit-up, weld size and reinforcement, and depth of
be inspected and tested, for defects and flaws, in a variety of ways, undercut.
to assist inSpecial templates
inspection and welding gauges are also available
of weldments.
depending on the nature and type of given welded structure, its
application, and the test equipment available. Brief description of some of the important equipment emp-
In general the inspection and testing of welds may include one loyed for visual inspection follows.
or more of the following procedures depending upon the type of 10.2.1.1. Magnifying Glass
structure and the importance of its service.
1. Visual inspection an.d measurement, In order to examine an object in detail, it is brought as near
2. Destructive tests for mechanical strength and toughness, to the eye as possible with a view to obtaining a large visual angle.
3. Non-destructive tests for hidden or sub-:3urface flaws, Because the eye cannot focus sharply on objects closer than
4. Tightness tests i.e., Pressure and Leak Tests. approximately 250 mm, the maximum visual angle obtainable by
10.2. VISUAL INSPECTION AND MEASURF~MENT the unaided eye is limited by the power of accommodation. \Vhen
a convex lens is placed in front of the eye, the visual angle is
Visual inspection is probably the most widely used of all the increased and the eye looks at an enlarged virtual image, as shown
inspection methods. It is simple, easy to apply, non-destructive in
nature, quickly carried out, and usually low in cost. Just because I in Fig. 10.1. The diameter of the pupil of the eye is 2.5 mm for a
light of 5500A wavelength'·. The minimum angular separation of
it is simple and cheap, it sh.ould not be ignored or bypassed. Even
though a weldment is to be inspected by using other testing
-8
I
Image
methods, it should be given a good visual examination. Apart from
helping in detection of surface defects, it may be of great assis-
tance as a guide in other tests, for example, visual examination of Obj"'~R,t;nal
the weld bead may aid in deciding upon the angle of incidence of
an x-ray beam when examining the weld for cracks not visible at
the surface. Careful visual inspection of welds at different stages
of its fabrication can detect about 80 to 90% of the defects .and
(a)
,~--.
Image formed by an eyo? lens

flaws before they are found by more expensive testing methods. Image
10.2.1.Equipment·
The equipment required for visual inspection is quite simple,
but adequate illumination is absolutely essential. Prior to visual
O~:t~~
' ~EYe

inspection the surface of the weld should be thoroughly cleaned

with a steel wire brush and chisel to remove spatter, scale, and dag. ~~-;~, j
When the area to be inspected is not readily visible, the use ( b) Image formed by a convex lens
n'lay be made of mirrors, bf)I'l~scopes,t1ashlights, bundles of fibre Fig. 10.1. Image formed by: (a) a human eye, and (b) a convex lens.
optics, or other aids which may be used to insped normally
inaccessible places. * The sensTti~ity
a wavelength of 5560A.of the hllll1~~'-ey(;varies greatly for different wavelengths.
Under ordinary conditions the eye is most sensitive to yellow green light. which has

Welding Engineering and Technology
----------
704
two point objects resolvable by the eye is about one minute of an
arc .
The magnification of a single lens is given by, ...(10.1)
M agn111catlOn=
'J:':' 250
.----'
f(mm)
where f is the focal length of the lens in mm and 250 represents
the average minimum distance of an object from the normal un-
aided eye. Using equation (10.1), a lens with a focal length of 125
mm has a magnification of 2 written as 2X and is also called 2
power lens. The focal length of a simple magnifying glass and its
working distance are approximatelY the same.
The diameter of the field of view of a simple magnifying glass
is less than its focal length. Suppose that one is examining the
surface of a large specimen for defects; it would take hours to cover
the entire surface using a 20-power glass whose field of view is
just 12.5 mm. The proper procedure would be first to use a low
power glass, marking the doubtful areas, then to inspect the
questionable areas with a high power glass.
To get the best possible performance from a magnifying glass,
it should always be held as close to the eye as possible. This
permits the greatest number of rays from the specimen to enter
the eye. It also cuts down the reflected light striking the surface
of the lens, thus minimising reflection and glare. It should be kept
in mind that as an area is magnified, detail is lost.
10.2.1.2. Borescope
It is an instrument designed to enable an observer to inspect
the inside of a narrow tube, bore, or chamber. Borescopes are
precision-built optical systems having a complex arrangement of
prisms and le:lses through which light is passed to the observer
with maximum efficiency. The light source located in front or
ahead of the object lens provides illumination for the part being
examined. 1'he brightest images are obtained with borescopes of
large diameter and short length. As the length of the borescope is
increased, the image becomes less brilliant because oflight losses.
In most borescopes the observed visual area is approximatelY 25
mm in diameter at 25 mm'distance from the objective ..
Borescopes can be design'ed and constructed to provide any
desired an~le of vision. Fig. 10.2 shows borescopes with four
different angles of view. The retrospective borescope shown in Fig.
10.2(d) provides the only method of accurately visually inspecting
a bore with an internal shoulder.
Weld Inspection and Quality Control 705
"I
,,,, I
I
I
//
\ \ , .••. I
Right angle Foroblique [.,
:a
...•.•..•
Direct
I
~l~l...=1
Retrospective
_
Fig. 10.2. Borescope with different angles of view : (a) Right angle borescope,
(b) Foroblique borescope, (e) Direct system borescope, and
(d) Retrospective borescope.
10.2.1.3.Mirrors
One of the simplest ways of looking into an inaccessible 01'.80.
forbidden area is by means oflarge mirrors. The principal objection:
to their use is the difficulty of getting large mirrors sufficiently free
from surface irregularities. Mirrors have an advantage in that only
a portion ofthe mirror is used for a given object point since the cone
of rays is limited by the pupil of the eye.
All mirror surfaces must be extremely flat ifreflection is to remain
good. Mirrors must be kept clean and free from dust. If the right
number of mirrors is not used; the image will not be erect or the right
side too. Moreover the reflecting power of a single mirror is rarely
above 75% after a period of use, and the reflecting power of two or
more mirrors is obviously much less. However, the reflecting power of
a mirror can be improved by the use of special coatings.
10.2.1.4. Welding Templates and Gauges
Fit-up and finished welds are checked for correct geometry,
size and shape, reinforcement height, misalignment, and depth of
cut with the help of templates and welding gauges.
Fig. 10.3(a) shows three views of a typical welding template
and its use for checking the groove angle and root gap are
illustrated in Fig. 10.3(b).
The finished weld is inspected by means of welding gauges.
One such gauge consists of a set of metal plates with cuts
corresponding to the cross-section of a given size and type of weld.
Fig. 10.4(a) illustrates its use for measuring the dimensions of a
butt weld.
On each such gauge numerals are punched indicating the
thickness of the work in mm for butt welds, and the leg length in
mm for fillet welds. As it is difficult to keep the size of a weld
within precise limits, each gauge has its cuts for the maximum
and minimum size of a given weld. A set of gauges for i;he most
commonly used weld sizes, say 6 to 14 mm, is clamped by a ring
or a bolt as shown in Fig. 10.4(b).
 ,,:,,;'~;;§:§',+~§'¥!~~-"~'~"'!;!olIt
••,,!""•••• ,••.••••".". ---

Weld Inspection and Quality Control 707

Welding Engineering and Technology
706 1
100

Jrnd
10

~~
",.LL
~ ?=r co
•..•

ili
•...
T
100
I (a)S2 IT ~(b)
.",'
.
,~
(c)
60 Z I a a
co
•... ~Sl "''''~V~
(d) (e)

Fig. 10.5. A typical finish weld gauge: (a) its shape and markings,
and illustration of its use for measurement of dimensions in (b) a fillet
weld in T-joint, (e) a fillet weld in lap joint, Cd) a butt weld, and
(e) root gap in a butt weld.

I a) Different views of a welding Fig. lO.5(a) shows the general view of another type of a weld
gaug'e while Figs. 10.5 (b, C, d, e) show its use in checking fillet
welds (T and lap joints), butt weld and root gap in a butt weld

I b} Method of use of .welding template (A) Combination gage

Fig. 10.3. A typical welding template: (a) its three different views, and ( 8 ) Fillet weld gage.-
(b) its method of use for checking joint fit-up.
Fig. 10.6. Two other designs of welding gauges: (a) a combination welding
gauge, and Cb) a typical weld measurement gauge.
respectively. Two other designs of welding gauges viz., a combi-
nation gauge and a set of fillet weld gauges are shown in Fig.
lO.6(a) and Fig. 10.6(b) respectively.
10.2.2. Visual Inspection
Visual examination and measurements are carried out before
welding, during welding, and after welding.
Before welding, workpieces are examined visually in order to
( b) A set of welding gauges detect any dents, burrs, scales, seams or laps, and also dirt, grease,
(Q) Use of welding gauge to . rust, and mill scale on the fusion faces and edge preparation.
measure butt weld dimensions
During fit-up workpieces are checked for the geometry of the edge
Fig. 10.4. Welding gauge.

Welding Engineering and Technology
708
and joint preparation (the root face, the bevel and included angles,
etc.)
During welding it is to be ensured that the designated welding
process is employed and specified welding consumables confor-
ming to storage specification are used. Clamping devices, fixtures,
and locating pins are to be used where required; correct welding
current and proper polarity are to be checked. Preheating and
interpass temperatures should be employed with interpass
cleaning assured by the use of chipping, grinding and gouging
tools. The welder should be of required qualification level and his
work should be actually observed during welding to ensure that
the welder produces the required quality of work. postweld heat
treatment (PWHT), and repair welding should be as per the
specified codes. Warpage, if it occurs should be corrected using the
specified methods like press working, thermal bending, etc.
After welding but prior to final visual inspection the surface
of the weld should be thoroughly cleaned with a wire brush and
chisel to remove spatter, scale, and slag. Then the finished welds
are inspected visually for external defects which may include one
or more of the following:
(i) Appearance of ripples on the weld,
(ii) Lack of fusion at the root of the butt weld,
(Ui) Overlaps,
(iv) Undercuts,
(v) Unfilled craters,
(vi) External cracks,
(vii) Improper size and shape of beads,
(viii) Misalignment (high-low),
(ix) Surface porosity,
(x) Suckback,
(xi) Excessive reinforcement,
(xii) Arc strikes,
(xiii) Excessive spatter.
The appearance of a weld is considered to be good if ripples
are spaced regularly. Irregular ripples as well as the varying width
and height of the weld are an indication that the arc was broken
many times during the welding process that its wattage varied
within broad limits, which fact may result in lack of fusion and
blowh9les. Irregular welds should be tested for internal defects by
one method or the other ..
Hair cracks may be detected with a magnifying glass.
Weld Inspection and Quality Control 709
Craters on the surface of the weld should be thoroughly melted
and filled. Examination of unfilled craters may give an idea of the
welding procedure used by the operator. A very deep and spongy
crater results from an excessive welding current and overheating.
A shallow crater points to a low current and the likelihood of poor
fusion.
Welds are inspected for size and shape by means of welding
gauges already described in section on equipment. Fillet weld size
and method of checking are shown in Fig. 10.7.
~ ~
'--
'-.....
:.-
~ W/////////////N///,
Fig. 10.7. Methods of checking fillet weld size by the use of welding gauge.
Although visual inspection provides a heap of data, care
should be taken when judging the quality ofa weld from the visible
appearance alone. Acceptable surface appearance does not necess-
arily prove careful workmanship, and is not a reliable indication
of subsurface weld integrity. However, all welds must be subjected
to visual inspection, irrespective of the tests tQ be applied
afterwards. "',
10.3. DESTRUCTIVE TESTS
Destructive tests are mechanical tests. In these tests the
welded component or structure is loaded till destruction and thus
that cannot be used after such a test.
Destructive tests are the usual tests employed for testing
soundness, strength and toughness of materials and may include
tensile tests, bend tests, impact tests, etc. These tests ate-used to
qualify welding procedures and welders. Such tests are expensive.
110 Welding Engineering and Technologr
to carry out since they involve the preparation of material, the
actual making of welds, the cutting and machining of weld test
speCimens, and finally testing them to destruction. Most of the
relevant destructive tests have been discussed under the title,
'Service Weldability Tests', in section 7.5.
10.4. NON-DESTRUCTIVE TESTS (NDT)
Non-destructive testing of weldments is the method of
detecting defects or flaws in welds without impairing the
usefulness of the component or structure under test. Although
visual inspection is usually included in the list of NDT methods
but in this section only those methods are included which employ
some specialised equipment and technique, and are not mere
visual examination. Some of these NDT methods include the
following.
1. Liquid-penetrant testing,
2. Magnetic particle testing,
3. Eddy current testing,
4. Magnetographic testing,
5. Radiographic testing,
6. Ultrasonic testing,
7. Acoustic emission testing.
All these NDT methods are briefly described in this section.
10.4.1. Liquid-Penetrant Testing
Liquid Penetrant Testing is a method to detect the surface
openings like cracks, surface porosity, folds, inclusions in stainless
steel, aluminium, magnesium, brass, and many other nonferrous
metals ..
The liquid-penetrant testing technique is based on the ability
of a penetrating liquid to wet the surface opening of a weld defect
and to be drawn into it.
Once the penetrant enters the cracks or other openings, the
rest of the liquid can be wiped or washed away but the liquid in
the cracks/openings remains there. Upon application of a chalk-
like developer blotting paper action takes place with the developer
drawing the dye from the crack and showing it on the surface of
the developer. Thus, the flaw is detected.
Liquid penetrant testing involves the following basic steps
1. Surface preparation,
2. Pentrant application~
Weld Inspection and Quality Control
711
3. -nw~
..',:;}ltime,
4. Removal of excess penetrant,
5. Application of developer,
6. Surface inspection,
7. Post cleaning.
major groups viz.,
Liquid-penetrant testing methods can be divided into two
(i) Dye-penetrant testing,
Ui) Fluorescent penetrant testing.
The major difference between the two types of tests is that
dye-penetrant test utilises visible penetrant, usually red colour
that produces a contrasting indication against the white back-
ground of the developer. For fluorescent penetrant testing the
medium employed is fluorescent liquid i.e. a liquid that glows
when illuminated by ultraviolet (UV) or black light of a
wavelength of approximately 3650A. Both these methods are
SOurce. effective when applied properly; however fluorescent_
equally
penetrant testing metrod is more expensive and it needs UVlight
Some of the internationally renowned liquid dye-penetrants
are Dy-check,
well Met-L-chek,
known fluorescent and Spotchek while Zyglo-pentrex is a
liquid.
10.4.1.1.Surface Preparation
All parts or areas of parts to be inspected must be free from
contaminants that inhibit flow of penetrant into cracks and other
surface openings. Such inhibiting contaminants include dirt,
grease, scab, acids, chromate, rust, welding flux, spatter, paint,
oily films, etc. When only the weld is to be inspected, the weld
cleaned.
bead and adjacent areas within 25 mm of the weld must also be
Of the many possible cleaning methods, solvent cleaning is
most popular. The solvent used should be volatile otherwise it will
dilute or prevent the liquid penetrant from entering into the
surface openings. Some of the popular solvent for cleaning metallic
percWorethylene.
surfaces are acetone, isopropyl alcohol, methylene chloride, and
Shot-blasting is not recommended as surface cleaning method
as this peens-over the cracks resulting in their closure. It is
essential that the part to be tested is thoroughly dried after
cleaning because any liquid residue will hinder the flow of
 Weld Inspection and Quality Control 713
Welding Engineering and Technology I
712 T~'-'
d&Ule10.1. Typical Penetration Times for
Liquid Dye-Penetrants
penetrant into the cracks, etc. If need be, drying may be
accomplished by warming the parts in drying ovens, with infrared
.
All
Aluminium
Steel welds
Dy-chek 10 welds(i)
10-20
10-20
2
Spotcheck Cracks and7-10
Fatigue
Cracks porestime,
cracks
3-5
Penetration
Type of dffect iI (150-320Cj
minutes
3.
lamps or forced hot air. However,
2. the
No. temperature
i

I
of the part,
(ii) Heat-treatment cracks
s. I Material under test
1-
under inspection, must not exceed 52°C prior to the application of
penetrant.
10.4.1.2. Penetrant Application
After the welded component has been cleaned, dried and
cooled to lower than 52°C, the penetrant is applied, to the surface
to be inspected, by dipping, brushing, flooding, spraying, in the
form of red dye or a fluorescent liquid. Small parts are often placed
in suitable baskets and dipped into a tank of a penetrant. On Table 10.2. Typical Penetration Times for Zyglo-Pentrex
larger parts, and those with complex geometries, penetrant can , I
be applied effectively by brushing and spraying. For general S. Material under test I _r~ u/
Tvno ~. defect
No.
application penetrants available in convenient aerosol spray cans 30
are very popular. Aluminium I (i) Weld cracks and pores
30
After application, excess penetrant is drained off and care is I (ii) Fatigue cracks
taken to prevent pools of penetrant on the surface of the 2. Magnesium (i) Weld cracks and pores 30
(ii) Fatigue cracks 30
3. Stainless steels (i) Weld cracks and pores 60
(ii) Fatigue cracks 30
Penetrant
10.4.1.4. Removal of Excess Penetrant
After the required dwell time, the excess penetrant is removed
by the use of solvents or self emulsifiers. The solvent may be
trichlorethylene, perchlorethylene, acetone, or a volatile petro-
Fig. 10.8. The flow of liquid penetrant into a flaw that has reached the surface.
component under test. Because of the low surface tension,
I leum distillate. The former two are somewhat toxic, the latter two
are volatile. Proper safety precautions should, therefore, be
observed for their use.
penetrant is readily drawn into the small surface openings by
capillary action, as shown in Fig. 10.8. If the surface of the Cleaning solvent may be applied by using wipes of clean
specimen is warm when the penetrant is applied, better results lint-free material, repeating the operation until all traces of excess
will be obtained because the surface openings will be slightly penetrant· are removed. These solvents are also available in
expanded and therefore more easily penetrated. aerosol spray canes. Care should be taken not to use excess solvent
To detect very fine flaws the penetrant may be applied more as that may cause the removal of penetrant from the surface
than once tu the same surface. cracks and other openings. The part is then allowed to dry by
10.4.1.3. Dwell Time normal evaporation before applying the developer.
The exact penetration time can be determined only by 10.4.1.5.Application of Developer!
experimentation on the particular specimen that is to be After the excess penetrant has been Jemoved, the developer
inspected. By varying penetration time, the optimum time for a (which consists of fine chalk-like silicates that soak up the liquid
.particular size and kind of defect can be determined and may often penetrant) is applied in order to make the ci~bks and other surface
range between 15 and 30 minutes. openings readily visible by blotting paper action, as shown in Fig.
10.9. The developer draws the penetrant out on each side of the
Table 10.1 shows typical penetration times for Dy-chek and opening increasing the size of the indication.
Spotcheck penetrants as recommended by their manufacturers
while Table 10.2 lists typical penetration times for Zyglo-pentrex.
 Weld Inspection and Quality Control 715
114 Welding Engineering and Technology

-Oeveloper The length of time before the coated area is visually observed
powder for indications should not be less than 7 minutes, or as
Indication recommended by the manufacturer. Developing time begins
Penetrant immediately after the application of dry powder developer or as
soon as a wet developer coating is dry. If bleed out does not alter
the inspection results, development periods of over 30 minutes are
permitted before inspecting the surface for defects.
10.4.1.6. Surface Inspection
Fig. 10.9. Schematic representation of blotting paper action of the developer. Mter applying the developer--thesurface is inspected for flaws.
The two major types of developers are the wet developer and The inspection can be done with either a normal light for visible
the dry developer. Wet developers can be either solvent base or type dyes, or with a black light (UV light) for fluorescent dye
water base; generally both are applied with a pressurised spray. penetrants.
The dry developer is applied in a powder form only after the Visible penetrant indications can be inspected in either
surface is dry while the wet developer can be applied even when natural or artificial white light. A minimum light intensity at fhe
the surface is still damp to the touch. Fig. 10.10 shows the three inspection site of 350 Im/m2 is recommended. Inspection for
steps of the liquid-penetrant test. fluorescent penetrant indications is done in a dark area. Maximum
ambient light of about 30 Im/m2 is allowed for critical inspection
though higher levels may be used for non-critical inspection. UV
light intensity should be a minimum of 800 I-Lw/cm2on the surface
of the component under inspection ..
With either method two types of indications are revealed: true
Crack indications and false indications. True indications are caused by
..,,.
penetrant bleeding out from actual surface openings, as shown
,.

-
(a) Dye penetrant (P)
,."",.
Tight crack or
- -~~=:.~-.=..=--
Cold -shut crack
partially welded lap
~~~
~~~~S
--.---- ...f.::-;;~.:_o.;;;....,

Large crack Crack or Fatigue crack Porosity

01' opening cold shut or lap or pits

I a) Some typical flaw indications

~ Fig. 10.11. Liquid penetrant test results: (a) schematic representation of some
typical flaws, and (b) actual test results with dye penetrant (See plate 10.11).
Surface schematically in Fig. 10.11(a), while Fig. 10.11(b)shows the actual
defect test result with dye-penetrant.
The standard true flaws that are indicated are: the cracks,
fatigue cracks, pits and porosities. A large crack is represented by
a line of some width; such cracks become 'apparent quickly after
the application of the developer. The fatigue crack that is bleeding
(e) Chalk developer (C)
through the surface, is represented by a line of dots and may take
a few minutes to show itself. Porosity is noticeable quickly in that
Fig. 10.10. Three steps involved in inspection by dye-penetrant testing.
716 Welding Engineering and Technology

an indication of porosity comes to the surface almost immediately,

as do indications of large cracks. Porosity is generally indicated
by dots (Fig. lO.ll(a)) that identify the area of porosity.
There are only a few false or non-relevant indications to watch
out for. Excess penetrant liquid in the depression between weld
beads may look like cracks when they are not. If such a problem
is encountered it is best to lightly grind the weld surface smooth
and then retest it.
Proper interpretation of dye-penetrant test results can only be
learnt by experience.
10.4.1.7. Post Cleaning
Post cleaning is employed where the developer or residual may
interfere with subsequent processing or service requirements. It 1il
~
is particularly important where residual inspection materials
might combine with other materials in service to produce corrosion ~
products. A suitable technique, such as simple water rinsing, ~
machine washing, vapour degreasing, solvent soaking, or fjt:lo
"1:1
ultrasonic cleaning may be employed. Part cleaning should be 'S
carried out immediately after inspection to avoid adhesion of g
•...
developer to the surface of the component. Developer should be o
removed prior to vapour degreasing, if employed, because that can
bake the developer onto the surface of the component.
]..
....

Q)

10.4.2. Magnetic Particle Testing

Magnetic particle testing is a nondestructive inspection J
method used to detect surface or near-surface defects or
discontinuities in magnetic materials. The method is based on the
principle that magnetic lines of force, when present in a
ferromagnetic material, will be distorted by an interruption in
material continuity, such as a discontinuity or a considerable
change in chemical composition. For example, if a magnet is bent
and the two poles are joined so as to form a closed ring, no external
poles exist and hence it will have no attraction for magnetic
material. Thus, as long as the part to be inspected is free of cracks
or other discontinuities, magnetic particles will not be attracted.
When a crack or other discontinuity is present, north and south
magnetic poles are set up at the edges of the discontinuity (or
crack). The magnetic particles will be attracted to the poles which
are the edges of the crack or discontinuity, as shown in Fig. 10.12.
Magnetic particle testing is obviously limited to ferrous
metals, nickel, and cobalt alloys; and that also when these
materials are magnetic. It will not work on aluminium, copper,
zinc, austenitic stainless steels, or austenitic high-manganese
steels because all these materials are non-magnetic.
i

- ------
717

----- --
Weld Inspection and Quality Control
Crack particles

N North -- pole
---
1:=
---
--- South
pole S

'-V
Magnetic lines
of force

Fig. 10.12. Schematic representation of magnetic particles

attracted to the edge of the crack.
10.4.2.1.Method of Operation
To achieve the best results in magnetic particle testing it is
essential to orient the lines of magnetic flux in a direction
perpendicular to the discontinuity or defect. This is usually
obtained by employing two types of magnetisation, viz;
(i) Circular magnetisation, and
(ii) Longitudinal magnetisation.
Circular Magnetisation. When electrical current is passed
through a conductor, it produces a magnetic field with magnetic
flux lines at 90° to the current flow, following the right hand rule
of current flow, as shown in Fig. lO.13(a).

Current
flow ~
~
I a) Magnetic field around a canductor carrying current
Magnetic field

I b 1 Circular
e\ectTic
Fig. 10.13. Circular magnetisation : (a) circular magnetic field around a linear
conductor carrying current, and (b) circular magnetic field within the conductor.
When current is passed through a conductor, a magnetic field
is present on its surface as well as around it. However, when
current is passed through a defect-free ferromagnetic conductor,
 '_':""._'_::,-._,,:'
:..~ __..~.::::~-::.o..","=====-,,::-,~,'~---' ,'n __

,Weld Inspection and Quality Control 719

718 Welding Engineering and Technology

Test
magnetic field perpendicular to the flaw is needed. This is created
fixture by the use of a circular coil, as shown in Fig. 10.15.
Thus, in order to check a test specimen fully, both circular and
longitudinal magnetisation must be achieved.
10.4.2.2. Equipment
The basic equipment for magnetic particle testing of small
sized components is similar to the one shown in Figs. 10.14 and
10.15 but for testing large sized jobs portable equipment of the
Defect
two basic types are employed that provide a magnetic field in a
localised area. These are known as,
Current (i) Prod method, and
(a) Principle of operation (ii) Yoke method.
Prod Method. In this method a localised area can be magne-
Contact plate
tised by passing current through the component by means of
Contact plate hand-held contacts or prods, as shown in Fig. 10.16. Manual
clamps or magnetic leeches may be used instead of prods.

( Lines of magnetic force

t
Iron powder
f

t
Current

( b I Set-up
Electric cur rent
Fig. 10.14. Detection of a flaw with circular magnetisation.
such as carbon steel rod, most of the magnetic flux field is confined (-)
within the conductor itself, as shown in Fig. 10.13(b). When a (+)
longitudinal defect, in-line with current flow as shown in Fig. Crocks at 90' to lines C rocks parallel to lines
lO.14(a), is present in the conductor then such a circular magnetic of force WILL show
of. force will NOT show
field is able to detect it.
Longitudinal Magnetisation. When a defect or discontinuity is Fig. 10.16. Prod method of local magnetisation for magnetic particle inspection.
oriented along the circumference of a welded component it may The passage of current through the workpiece creates local,
not be detected by a circular magnetic field instead a longitudinal circular magnetic fields in the area between the contact points.
This method is used extensively for localised inspection of
weldments where the area of interest is confined to the weld zone.
The prods must be securely held in contact with the part to avoid
arcing at the contact points. For this reason a low open circuit
voltage of 2 to 16 volts is utilised. This method provides only a
Magnetic unidirectional magnetic field therefore it is necessary to reorient
flux

Fig. 10.15. Longitudinal magnetic flux produced with the help of a circular coil.

no Welding Engineering and Technology
the contacts at about 90° and remagnetise for complete inspection
of the area.
Yoke Method. In this method localised magnetic field is
induced with the help of a solenoid having flexible extensions of
the core. When the extensions make contact with the component
under test and the coil is energised, the magnetic field of the
solenoid is concentrated in the component between the contact
points. The equipment for this method can be relatively small and
light weight. Another desirable feature of this technique is that
unlike the prod method electric current is not transferred to the
workpiece thus there is no tendency to arc or burn the component.
10.4.2.3.Type and Amount of Magnetising Current
Alternating or direct current may be used for magnetising the
workpieces.
With a.c. only the surface of the metal is magnetised. Thus,
a.c. is effective for locating discontinuities that extend to the
surface, such as cracks, but deeper discontinuities or incomplete
fusion would not be detected. It may, therefore, be used to inspect
welds where subsurface evaluation is not required.
D.C. produces a magnetic field that penetrates throughout the
workpiece and is, therefore, more sensitive than a.c. for the
detection of subsurface discontinuities. Full-wave, 3-phase recti-
fied current produces results essentially comparable to direct
current obtained from generator or batteries.
Half-wave rectified, single phase current provides maximum
sensitivity. The pulsating field increases particle mobility and.
enables particles to line-up more readily in weak leakage fields.
If the particle build-up with d.c. magnetisation is considerably
greater than that with a.c., it is highly probable that the flaw has
a significant depth, and is not just a surface irregularity.
The applied voltage has no effect on the magnetic fields, and
should be kept low to prevent arcing and overheating. Thus, high
amperage, low voltage power is usually employed.
The magnetising current should be of sufficient strength to
indicate all detectable flaws but not excessive to produce irrele-
vant patterns.
The approximate amperage ranges for the various magne-
tising methods are as follows.
1. Longitudinal magnetisation : 3000 to 10,000 ampere-
turns depending on the ratio of the coil and workpiece
diameters.
2. Overall Circular Magnetisation : 40 to 400 Ncm of
workpiece diameter.
Weld Inspection and Quality Control 721
3. Prod Magnetisation : 36 to 60 A/em of prod spacing,
depending on metal thickness.
4. Yoke Magnetisation : The magnetising current must be
sufficient to lift 20 kg weight with d.c. magnetisation,
and 5 kg weight with a.c. magnetisation.
10.4.2.4.Methods of Operation
Two basic methods are normally employed. One is the dry
method, wher.ein dry magnetic powder is sprinkled on the
workpiece, and the other is the wet method, which uses magnetic
particles suspended in a liquid.
Dry powder is more satisfactory on rough surfaces and
provides for the greatest portability of inspection. This method is
likely to reveal a subsurface flaw better than the wet method.
The magnetic particles in the wet method are much smaller
than in the dry method and the wet precess is far more sensitive
to fine surface defects. These factors make wet magnetic .particle
testing better suited for the discovery of fine surface flaws on
smooth surfaces.
The liquid used for wet method could be water or light
petroleum distillate (e.g. white kerosene oil). While both oil based
and water suspensions provide nearly equal sensitivity, the fire
hazard due to arcing is avoided with water suspension. However,
the presence of water near electrical apparatus creates a shock
hazard that must be guarded against.
The magnetic particles are available in two colours, viz., red
and black. The red particles are better for·dark surfaces.
10.4.2.5. Sequence of Operation
The sequence of operation in magnetic particle testing involves
timing and application of the particles and the magnetising
current. Although sometimes residual magnetisation is also used
but normally continuous magnetisation sequence is employed,
with either wet or dry particles, for most applications. The
sequence of operation differs for wet and dry continuous magne-
tisation techniques.
The wet technique is generally used for those components
processed on a horizontal, wet-type testing unit. In practice, it
involves:
(i) bathing the component with inspection medium to pro-
vide an abundant source of suspended particles on the
surface,and
 Weld Inspection and Quality Control 723
Welding Engineering and Technology
722
__ , .... .:TYl-:.:.~~..:.:.
__ --- ....••.•. 'nT~i_~1 Many steel weldments used in the aircraft industry, such as
landing gears and light structural components, are tested by
magnetic particle inspection after the part has been in use for a
long time. This technique is excellent for detecting the beginning
of the tiny fatigue cracks that always start on the surface.
Magnetic particle inspection is also used to test the edges of
thick steel plates prior to welding to detect cracks, laminations,
non-metallic inclusions, and alloy segregations in steel.
On multipass welds, it is sometimes used to inspect each layer
immediately after it has been deposited.
For obtaining good results in magnetic particle inspection it
is imperative that the surface of the weld and surrounding areas
should be clean, dry; free from oil, water, excessive slag, or other
accumulations that would interfere with magnetic particle
movement. Wire brushing, or sand blasting the surface before
testing helps a lot and should be done if the surface is not very
smooth.
10.4.3. Eddy Current Testing
In eddy current inspection of welds, an alternating current is
passed through a coil placed in the proximity of the weld. The
changing current in the coil creates an alternating magnetic field
in the component material. The varying magnetic field in the weld
creates electrical 'eddy' currents in the material. These eddy
currents, which vary with the magnetic field, create their own
magnetic field which interacts with the initial field. The test coil,
or in some cases a separate pick up coil, is electrically monitored
to detect any changes in this field interaction. Discontinuities in
the weld will alter the magnitude and direction of the eddy
currents and thus be detected through the test signal. The signal
is then displayed on an analog meter, digital meter, cathode ray
tube, x-y plotter, or strip chart recorder.
10.4.3.1. Eddy Currents
When a coil carrying a.c. is brought near a metal specimen,
eddy currents are induced in the metal by electromagnetic
induction. Generally the currents are induced in the component
by making it the core of an a.c. induction coil, as shown in Fig.
10.17. A crack in a welded seam can disrupt the eddy current flow
and the magnetic field produced by that current.
The magnitude of the induced eddy currents depends upon the
magnitude and frequency of the a.c., the electrical conductivity,
magnetic permeability, shape of the specimen, the relative position
of the coil and spE;cimen, and the presence of defects or
 Weld Inspection and Quality Control 725
Welding Engineering and Technology
724
-< inhomogeneities in the specimen. Mechanical and thermal treat-
ments of a specimen will affect its electrical conductivity. In the
case of austenitic stainless steels, mechanical and thermal
treatments will affect the magnetic permeability.
10.4.3.2. Skin Effect
The induced eddy currents are concentrated near the surface
of the specimen, resulting in the so called skin effect. This effect
is shown in Fig. 10.18, in which the term standard depth refers
to that depth where the eddy current density is approximately
37% that of the surface. Depth of current variation varies inversely
with the frequency of the alternating eddy currents. The frequency
of the current used may range from 50Hz to 1 MHz depending
upon the type and thickness of the material being tested. Table
10.3 shows typical standard depths of penetration for several
metals and magnetising current frequencies. Normally, a part
being inspected must have a thickness of atleast 2 to 3 standard
depths before thickness ceases to have an effect on eddy current
response.
Table 10.3.Typical Depths of Eddy Current Penetration
and the Required Magnetising Current Frequencies
for Different Materials
1. S. 1No. --.- 1----
16
Lead40:19
64
0.25
0.40
0.45
1.65
High0.9
0.51alloyMetal
2.8
250
0.24
0.03
0.1
Aluminium-B061
Zirconium
0.90
1.40
0.71
0.8
3.3
0.13
0.46
1.8
1.9
1.7
5.6
6.5
KHz
Aluminium-7075
3.7
0.36
0.07
0.12
0.23
0.20
0.10
1.0
1.6
0.05
MHz
b.02
0.42
0.17
Stainless
Copper
Magnesium steel T6304111.3
steel*type
Standard 7.4
3.4
2.1
3.7
3.2
3.1(37%) depth of penetration, mm
0.5
------
- ,----
KHz
8.
4.
3.
5.
7.
6.
2.

Fig. 10.17. Effect of a crack on induced eddy current flow

1.00 ca~-------------
during eddy current testing.
Standa rd depth of pene1ration
where density of eddy turrent
I
•• 37 % of density at surfate
~ 0.80
~..
~
_C d
..
:-
::J

....c
•..
......•.•::J
::J
* The standard depths are for tests without magnetic saturation. When
>0
.."
•.•
>- saturated, the values for ferromagnetic steels are approximately the same as those
for austenitic stainless steels .
-..
.." .."
.."
••
o -o 0.20 10.4.3.3.Test Coils and Probes
-c -c
>- >-
0;::; 0;;
Three general types of test coils are commonly used: the
•• •• o 6 concentric coil, the point probe, and the inside or bobbin type coil.
C) C)
1 4 5. Fig. 10.19 shows these three types of coils used for testing bar-like
o
Depth below surfate components while Fig. 10.20 shows four types of commonly used
Standard depth of penetration
surface probes.
Fig. 10.18. Variation of density of eddy currents as a function
of depth below the surface of a conductor.
 Weld Inspection and Quality Control 727
Welding Engineering and Technology
726
Absohite
Oifferential The concentric coil completely surrounds the specimen and
explores an annulus of the specimen having a width equal to the
.. effective width of the coil. Such coils are primarily used on welded
s~ c;.
c .-pr-----+\ ~·:·:1-~ pipes to inspect longitudinal butt joints. Since the eddy currents

i Vi -t,-- flow in a circumferential direction, longitudinal defects produce

.,..
the most significant change in eddy current flow. The pipe is
c:
passed through the coil on rollers, which makes the technique
...• suitable for automation. The point probe consists of a small coil
that can be placed near the surface of the specimen. The point
c:'"
probe explores an area essentially equal to the cross-sectional area
:.0:-::
.Q ••
0 •••
ofthe probe. The inside or coil bobbin is made to be moved through
CD the tube or pipe.

=~
The test coils are also sometime classified as absolute and
differential. The term absolute is used here to mean that the
~I1:U=-~~_ ~
~ ~
.S: - .~
measurement is made without a direct reference to or comparison
with a recognised standard. In the differential type of test coils
~ two coils connected in series opposition are used. Such an
"e
.. arrangement can be used in either of the two ways. In one of the

i techniques one of the coils surrounds or rests On a standard

.Q
~
... specimen (or the specimen known to be good or free of harmful
I-Btr--·~ .+ defects); the other coil surrounds or rests on the specimen to be
tested. If the specimen being tested is good, the voltage output
from the two coils is zero. However, if the two specimens differ for
Fig. 10.19. Different types of eddy current coils. some reason, there will be an output voltage.
Pencli Horseshoe
Tangent
In the other technique the two coils are arranged coaxially so
Pancake point or U- shape coil that the specimen to be tested goes through both coils, one sectibn
probe coil
probe of the specimen being compared to an adjacent section of the same
specimen. The differential test coil is not very sensitive to gradual
changes in diameter or structural properties but is very sensitive to
short cracks or seams. If a defect is long or large enough to extend
through both cells (i.e. coils), this technique will not reveal the
presence of the flaw, except when·the flaw enters or leaves the coil
sy~tem.The configuration and frequency of the exciting and pick up
coilsmust be varied to fit the different test conditions and specimens.
Test coilsand probes vary greatly in size, depending upon specific
applications. The size and type of test coils are governed by the
material ofthe specimen and the kind and size ofdefectto be detected.
For a coilofgivendiameter, the sensitivityto a defectdecreasesrapidly
as the defect size becomesless than that of the coildiameter.
10.4.3.4. Testing Techniques
'Weld There are two ways of measuring changes that occur in the
magnitude and distribution of eddy currents. Either the resil'ltive
component or the inductive component of impedance of the
exciting (or secondary) coil can be measured.

Fig. 10.20. Four types of commonly used Eddy Current surface inspection probes.
723 Welding Engineering and Technology
Weld Inspection and Quality Control
729
Spacing between the coil and the test specimen is an important
variable as any variation in the coil to specimen spacing or the netic materials. The penetration in ferromagnetic
shape of the specimen will affect the flow of the eddy current and materials may, however, be significantly increased by
hence the test results. using special techniq'les such as magnetic saturation of
Associated with any' electrom·agnetic test signal are thre8 the area being inspected.
lmportant attributes viz. amplitude, phase, and frequency. A 3. Since many variables can affect an eddy current signal
frequency low enough to permit reasonable penetration of current (e.g. permeability, conductivity, probe position, and weld
should be selected. contour, etc.), care must be taken to suppress or separate
Equipments are available which are based on amplitude variables of no concern from those of interest. In many
discrimination, phase discrimination, harmonic analysis, coil cases, this is not easy to accomplish.
design, choice of test frequency, and magnetic saturation. 10.4.4.Magnetographic Test
Some of the defects that are readily detected by eddy current
testing are seams, laps, cracks, slivers, open welds, missed welds, A magnetographic test is a sophisticated weld inspection
misaligned welds, and pinholes. technique. In this test a magnetic field is induced in the weldment
10.4.3.5.Advantages and Limitation being tested by a strong electromagnet which is in the form of a
Magnetic coil
In general a majority of the defects can be detected by eddy
current inspection at high travel speed and low costs. Apart from
these the eddy current inspection has several other advantages
over other inspection techniques.
1. The equipment used particularly with surface probes is
generally light-weight and portable.
2. Welds with surface conditions, such as excessive rough- Power source
ness and minor undercuts, which may result in non-
relevant or false indications, can usually be inspected by /
eddy curnmt techniques without the need to verify the f
relevance of the indication by further processing e.g.
grinding to remove the surface irregularities and retest- ,
ing. Defect
3. Since intimate contact between the weld metal and
probe is not required, painted or coated welds can be ( a I Set up fo r recording data
inspected. This can result in significant savings in the Tape reels
areas of in-service inspections and periodic preventive
maintenance inspections.
4. In some instances, such as the inspection of welded pipes,
the inspection can be partially or completely automated for
a high speed, relatively inexpensive inspection.
There are three general limitations in using eddy current Magnetic /
inspection on welds. tape

1. The test piece i.e. the weld must be an electrical conduc- Video ./
tor. head

2. The depth of inspection is generally limited to 6 mm for Ib I Setup for playback

non-ferromagnetic materials and 0.25 mm for ferromag-
Fig. 10.21. Magnetographic testing of weldmellts : (a) general set-up, and
(b) playback of magnetic tape to detect defects.
730 Welding Engineering and Technology Weld Inspection and Quality Control 731

reel, as shown in Fig. 10.21(a). A discontinuity or a flaw in the Table 10.4. Electromagnetic Spectrum
weld makes the field to deviate, so that an area of reduced S. Radio
Type waves
Visible
X-rays
Gamma ofrays
Ultraviolet
Infrared
Wavelength,rays
rays
rays
rays * I
A Iwaves 0.1-0.005
4000-7500
5-0.0004
7500-4
10-4,000x 106
106_1013
magnetisation is produced on a magnetic
4. No.
2.5.
3. 6.
tape placed over the 1.
weld. When the tape is 'played back' a spike is produced to indicate
a flaw, on the screen of the cathode ray tube as shown in Fig.
lO.21(b).
10.4.5.Radiographic Testing
Radiography is a method of NDT which employs X-rays or
Gamma rays to penetrate a weldment and detect any defects or
discontinuities by the resulting image on a recording film or a
viewing screen.
* A is a symbol for Angstrom-the unit of wavelength-and is equal to 10-8 em.
When a testpiece or a welded joint is exposed to a penetrating
radiation like X-rays or y-rays, the variations in amount of X-rays and y-rays have the following properties:
radiation transmitted through it depend upon, 1. are invisible electromagnetic radiations,
(i) the relative densities of the metal and any inclusions, and 2. can penetrate matter,
(ii) thru-thickness variations. 3. are differentially absorbed,
Non-metallic inclusions, pores, aligned cracks, and other defects 4. travel in straight lines,
result in less or more radiation reaching the recording or viewing 5. produce photochemical effects in photographic emul-
medium. The variations in transmitted radiation produce optically sions,
contrasting areas on the recording film, as shown in Fig. 10.22. 6. ionise gases through which they pass,
7. are not affected by electric or magnetic fields,
8. travel with a velocity of 3 x 1010 em/see,
9. are capable of liberating photoelectrons,
10. cause some substances to fluoresce.
4t;-,QY sou,ce
The essential requirements of radiographic testing, employing
IJ'/Ii\\\
X-rays or y-rays, include the following.

/'i Ii!I \\r X-,a"

1. A source of penetrating radiation, such as an X-ray
machine or a radioactive isotope,
2. The object to be inspected, such as a weldment,
3. A recording film or a viewing screen,
/1111 II\\\ 4. A qualified radiographer (radiologist) to handle the X-
Ii i111 \\\
ray or y-ray unit,
Porosity 5. A means to process the exposed film or operate viewing
/ .i- I /./\. .' \ \ . \J: \'-lntlusirv
., system,
Specimen Film 6. A person skilled in the interpretation of radiographs.
The essential features of these two NDT methods are dealt
~~~~ with separately in the following sections.
x -ray image
10.4.5.1.X-Ray Radiography
Fig. 10.22. Effect of metal thickness and inclusions. etc. on X-ray absorption to
produce differential effects on X-ray film .. X-rays are produced when high speed electrons, in the form of
X-rays and y-rays comprise part of the electromagnetic a beam called cathode ray, strike a metal target placed in an x-ray
spectrum which can be divided on the basis of wavelengths as tube. The velocity at which the electrons strike the target is
shown in Table 10.4. determined by the tube voltage where the tube voltage is the
',:~-:: --.i.::::"::~:::::'::: - _- _:::. ::..~:::._--.;:-_-::.t. __ ...-".::.:.:. _= __
..::- ~_ .....
,.,,"'''''
.. ._._

732 Welding Engineering and Technology Weld Inspection and Quality Control 733

potential difference between the source of electrons, called the

cathode and the target. Since electrons in motion comprise an
81ectric current, the number of moving electrons determine the
magnitude of the electric current which is referred to as tube
current for an x-ray tube.
The energy of penetrating radiation is often expressed in
terms of electron volts (ev)*, kiloelectron volts (Kev), or million
electron volts (Mev). The energy of the x-rays produced at the (a 1 X - ray tube
target is proportional to the square of the mean velocity of the H -
electrons; and the shorter the wavelength of the x-rays, the greater
(b) Electrical connections for X - ray tube
their penetrating power. The velocity v acquired by an electron
because of a difference of potential V can be calculated by setting Fig. 10.23. Coolidge X-ray tube.
the K.E. equal to the work done in accelerating the electron and , In radiography it is desirable to have a source of x-rays which
is given by the following expression. approaches an ideal point source. This can be achieved by proper
1 2-12 design of the filament and focusing cup, but it does complicate the
:2 Mv = 1.6 x 10 V ...(10.2) problem of heat removal. In the usual case, a compromise has to
where, M = mass of each electron, (9.1 x 10-28 g), be made on the size of the focal spot which can be used safely. The
V = velocity of electron, em/sec, finer the focal spot, the better will be the radiographic image.
V = tube voltage, volts. The power requirements for x-ray machines range from 50 to
The intensity of the x-rays is directly proportional to the tube 24000 KV (Kilovolts). A 24,000 KV X-ray machine is capable of
current and depends on the tube voltage raised to the power photographing approximately fjOOmm thick steel.
greater than 2.5. The efficiency of x-ray production is given by the Test Procedure: The x-ray tube, welded test component and
following expression : the photographic film are set up as shown in Fig. 10.24. The x-rays
E = 1.4 X 10-7 ZV ...(10~3) are allowed to fall upon the test specimen. Some of the x-rays are
where, E = Efficiency, %
Z = atomic number of target material,
V:=: tube voltage, volts.
The above expression (10.3) indicates that the efficiency of x- .U--~~~. '-x - ray-tube
ray production is low at low voltages. At 300 KV,only about 3% of iiI
the energy of the electrons is converted to x-rays. The rest of the
H,
1/ I,ll
energy of the electrons at the anode appears in the form of heat. ,,:~x-rays
The amount of heat generated at the target of a tube is III I I \\
proportional to the product of the tube voltage and tube current. ", I \ \\
Consequently, it is necessary to cool the target. II I I \ \ \
There are two types of x-ray tubes which can be used for
1111',
I
radiography, a gas tube and a Coolidge tube. The latter shown in I I IJ\
IIII , I >.- \

Fig. 10.23 is used almost entirely today. The source of electrons

for coolidge-type X-ray tube is a heated tungsten wire filament.
Most x-ray filaments operate in the range of 6 to 15 volts and use'
a current of 3 to 5 amperes to heat the filament. The electron
current from a heated filament is a function of the temperature of
the filament. x-ray film holder
film
-'---*-Electron voli (ev) 'is the kinetic energy that an electron acquires when it is
accelerated in an electric field produced by a difference of potential of 1 volt ... Fig. 10.24. Schematic of a set-up for X-ray examination of welds.
734 Welding Engineering and Technology Weld Inspection and Quality Control 735

absorbed. The extent of this absorption, as already stated earlier, parameters can lead to a large number of defective pieces before
depends on the presence (or otherwise) of voids, foreign inclusions, the defect is detected leading to increase in the number of overall
or cracks in the weld metal. As. a result, the radiation passing scraped components. This is leading to the introduction of
i
through the weld and falling upon a photographic plate or film Radioscopy for quick and on-line detection of welds.
r
behind it will produce areas differing in optical density, dark spots Radioscopy: Radioscopy or real-time radiography is the X-ray
corresponding to denser areas and light spots to some defects in radiography in which the components are inspected for subsurface
the interior of the weld. The quality of the resulting picture defects and the image is obtained on the screen and or on the
depends on the intensity of the radiation source, the angle of recorder while at the same time it is analysed with the help of
inclination of the x-rays, the type and thickness of metal. computers.
X-raying is especially effective in locating cracks, lack of
fusion, undercutting, slag inclusion, porosity, pinholes, and Set-up and Procedure. In a typical radioscopic system,
blowholes. X-rays are used for inspection ofweldments in all types shown in Fig. 10.26 an x-ray beam is projected from a fixed point,
of materials viz. steel, aluminium, magnesium, etc. Radiography x- ray Image
source
is used extensively in the pipeline industry to ensure proper weld processor Video
display
quality. A wide range of X-ray units have been developed for the
purpose. However, X-ray'inspection is a slow and expensive NDT
method. Video camera
I
--0
Penetrameters are used to determine the sensitivity of the I
II
radiograph by placing it on the test piece as shown in Fig. 10.25. L._-lI---- I Reports
III I
printer

h~ x -ray source _ Systems control II L -1----1 Archival

1 storage
.

~~Ih\ ---- Analysis

L. 1.....
.J__ -'.I network
Factory

Fig. 10.26. Schematic of a typical real time X-ray (radioscopic) system.

/1, 1\ \~. Penet rating
//1/\\\\ X-rays and an image detection plane is positioned at a fixed distance from
the X-ray source. A manipulator system, placed between the two,
//'/\\\\
/1,'/\\\\ I moves the inspection object in the x-ray beam. As the object passes
through the X-ray beam, an X-ray shadow is cast upon the image
detector. The image is then converted to visible light, either by a
bare fluorescent screen or by an image intensifier. This image is
viewed by a suitable-light-Ievel camera system, which transmits
the signal to an image processor where it is digitized for
p.n"<am.t~.l
Sp,,'m.n / I/ I; ~I \ \\ \~\ "-J}Hm quantitative· analysis and display.
Fig. 10.25. Use Ofpenetrameter to check the sensitivity of a radiograph. A computer can automatically execute the entire process of
These are made of the same material that is being inspected. Thus, component manipulation, x-ray parameter control, image collection,
if the penetrameter can be seen clearly on the radiograph, any and data anaJysis. Software can deal with the actual welding
change in thickness of the testpiece (2% or more) will be seen operations so that they do not interfere with the imaging process.
clearly. Automated radioscopy has the capability for high-throughput
Although film radiography is slow and expensive, it is one of component analysis due to the simplicity of image gathering.
the most popular NDT methods for locating subsurface defects. In radioscopic systems, particularly automated systems, it is
However, with the introduction of automated welding system the essential to coordinate parts handling and positioning with image
production rate is increasing very fast thus a wrong set of welding

~
736 Welding Engineering and Technology
acquisition. Handling is done either with dedicated manipulators
or robotic systems.
Radioscopy does all that the conventional film radioscopy does
in detection of subsurface defects but adds a few more advantages
viz.;
1. The ability to move the part while viewing its x-ray
image, so that the difficult to detect flaws can be inves-
tigated quickly and feedback provided faster to the
manufacturing floor.
2. Enlargement or reduction of magnification of the
image's can be achieved, as desired.
3. Image-enhancement techniques are applied during in-
spection to better detect and understand weld defects.
4. It has the ability to automate the inspection sequence
leading to increased inspection throughput.
5. With the complex shapes and numerous angles neces-
sary to see the detail required in many welds film radiog-
raphy might be labour intensive, rraterials expensive,
and time consuming.
6. With radiosc9PY, onGcan actually increase fault detec-
tion coverage.
7. With an automated inspection capability, exact part
coverage and repeatability can be provided.
8. Real-time systems can reduce x-ray film costs.
9. Reduction in overall scrap-because defects are found
earlier in the process-and reduction in the number of
defective parts returned ty customers.
10. Radioscopy not only provides real-time inspection and
process control, but can help to create records in
videotape, floppy disk or optical disk format for long
term storage and use.
Apart from the normal uses to which film radiography is put,
radioscopy can be employed for many weld inspection applications,
such as in large scale aircraft manufacturing, which presents
special challenges because therein it is required to inspect a wide
variety of welding configurations rather than a great: number of
just a few different parts.
Because computer is better able to quickly detect features that
would be very difficult and time consuming (if not impossible) for
a human to detect. Inspection throughput is thus enhanced using
computer-based (digital) systems, and in many cases 100% of the
product can now be inspected on-line. With this capability,
Weld Inspection and Quality Control
737
automated radioscopic systems can be used effectively for process
control as well as quality inspection.
The flexibility of part manipulation, the speed of inspection
and feedback, the heightened sensitivity to weld defects and the
ability to automate all or part of the inspection process make
radioscopy an increasingly active tool for weld inspection.
10.4.5.2.Gamma Radiography
Gamma rays are emitted during the disintegration of radio-
active material and like x-rays are electromagnetic radiation.
However, gamma rays have a wavelength which is intermediate
between the shortest and longest x-ray wavelength as is seen from
the electromagnetic spectrum shown in Fig. 10.27. Gamma rays
Electrical I Infrared I Visible
Cosmic
3 x 10'7 3 X 106
104 It xl03
Wavelength. angs troms
Fig. 10.27. Electromagnetic spectrum.
do notdiscrete
have a continuous
more energies. spectrum as x-rays but consist of one or
Gamma rays are obtained either from natural sources like
Radium 226 (Ra226) and Radon 222 (Rn222) or the man-made
isotopes •. Natural Sources are either costly or have short useful
life so man-made
in industrial y-rayisotopes are almost exclusively used these days
radiography.
There are three major factors to be considered when selecting
radioisotopes:
1. the half-life,
2. the gamma ray energy, and
3. the material to be radiographed.
Half-Life: All radioactive sources are characterised by the
property which is known as the half-life of the radioactive
material; it is the time required for half the material to decay.
Fig. shows how
10.28decreases
60 (C060) thetime.
with intensity of a radioactive source Cobalt
* An isotope is one of two or more forms of an element which differ from each
other in atomic number and nuclear "behaviour but are chemically same.
738 Welding Engineering and Technology
WeJd Inspection and Quality Control
739

Radioactive Decay: The amount of radioactive material

remaining after a period of time t can be calculated if the amount
initially present is known, using the expression :
N = Noel- IJ) = NoeC-0.6931/1')
...(10.4)
where, N = No. of atoms of remaining material,
No = No. of atoms of material initially present,
•...
III
0.693
A. = Decay constant = ----r-
00
a..

o 2 3 •• 5 6 7 8 9 10 t = Elapsed time,
Years T = Half-life ofmaterial.

Fig. 10.28. Decay curve of Cobalt 60 (CoBO).

In equation (10.4), T and t must be in the same unit of time.
Example 10.1. If the half-life of Cobalt 60 (Co60) is 5.3 years,
Gamma Ray Energy. The activity ofy-ray sources is usually find the activity of its 10 curie source at the end of two years.
Solution. Using equation 00.4), we have
given in curies, millicuries, or microcuries. A curie is defined in
terms of a disintegration rate; a disintegration rate of 3.7 x 1010 N=Noe (-0;93t)
per second being 1 curie (c). The disintegration rate is not ...(i)
influenced by pressure, ..temperature, or chemical combination. Now, t = 2 years, and T = 5.3 years
Table 10.5 lists the half-life and energies of gamma radiations for Putting these values in (i) above, we get
some of the industrially important y-ray emitters used for weld
(- 0.693 x 2)
inspection. N=Nel 000 5.3 =Ne<-0.2615)=0.77N
Table 1().5.·Half-Life and Energy of Gamma Thus the activity of the Co60 10 curie source has decreased to
Radiation of Some Isotopes 0.77 of the initial activity or to 7.7 C. Ans. 7.7 Curie
S. Radium
Thulium
Cesium
Americium
Xenon
Cobalt
Iridium
Tantalum
Radon
Cerium
Half
Antimony
Europium
1.7
129
5.3
1620
290
470
74.4
3.85
37
60
120
5.3 0.084
Life
years
years
days
years
days
days
days 100
2.0-1.7
1.17-133
1.7
59
0.081
310-600
0.667
0.7-2.2
Isotope
Average
Symbol -
Kev
Mev
Mev
Mev
Kev
CoBO
Energy
48 Tm170Mev
Kevof "(.
Mev
Xel33
Sb124
Cs137
clg
Am241
Ra226
Ra222
Cel44
Ta182
Ir192
Eu155 Emission
Specific Activity: The term specific activity .is used to
indicate the activity of 1relationShip:
gm of a radioactive source and is
expressed by the following
S 'fi t"t 0.693 M
peCI IC ac IVI y = T ...(10.5)
where, M = Number of atoms per gm. of the material,
= Avogadro's number/Atomic weight of the material,
= 6 x 1023 / Atomic weight
T = HaIf-life of the material.
Thus, the specific activity of a given radioactive Source is the
element. of
number Thedisintegrations
time unit usedper unit time per gm of the radio-
is usually seconds.
life is 5.3 years.10.2.Find the specific activity of Cobalt 60 ifits half-
Example
Solution. Now, M = 6 X 1023
At.wt.
At. wt. of CoGO = 60
.. M=-
6 X 1023
 Weld Inspection and Quality Control 741

Welding Engineering and Technology

740
Set-Up and Procedure
A gamma ray source is placed in a capsule enclosed in a lead
Also, T:::: 5.3 years container and is used for weld inspection as shown schematically
:. putting these values in equation (10.5), we get, in Fig. lO.29(a). The gamma rays are emitted from the radioactive
0.693 §2<6~023 source at all times and in all directions, however the container is
provided with a window in the form of a lid so that the rays can
Sp. activity of Co60 :::: 5'.3 disintegrations/year be directed to the desired area of weldment. Also, lead containers
0.693 1022 d" t t' / d (with gamma ray source) with magnetic clamps, shown in Fig .
::::-5 .3 x 3.14 x 107 ISln egra IOns secon 10.29(b) can be employed for inspecting awakwardly located welds.
= 0.04164 X 1015
::::4.164 X 1013disintegrations/second.
Ans. 4.164 x 1013dis/sec
Lead
Gamma ray
container source

1\\
Gamma says 1/ 1'\
""-./,,\\
I' , \ \ ~source
/ I , \ \ (extendable for
" \ \ panoramic shots I
I , , , \ Lead shield

Control rod

Film holder
Screens x -ray film
I al (jeneral set-up

Capsule
Lead container

t:~ ...
QI
'0
"0
.&:.

-~
Beam
opening

QI
.•..
~ Cl

Capsule
0:

Lead cylinder
~Plat'lmml hald••
I tll (jamma ray source with magnetic clamps Fig. 10.30. Two types of isotope cameras.
Fig. 10.29. Gamma ray weld inspection: (a) General set-up: and (b) two methods
of inspecting horizontal and vertical weldments.
--~·'-·--~7"'--·-'-'·-"-'-"-'-'-~-~-------'-'-__·"_

742 Welding Engineering and Technology

Weld Inspection and Quality Control
743
Special devices in which the y-ray sources are permanently ""'_L.. _
fixed are a practical means of handling radiographic sources either
in plant or field. Two such radiography units are shown in Fig.
7. No.
8.
10.30 and are known as isotope 6.cameras. 3.
2.1.These are so designed
5.4.
(volts)
S. 12.5For
mm,
Iridium,
Cerium,
Cobalt,
Cesium,
Thulium,
Tantalum,
Radium,
Europium,2.5-,-62.5
Halfvalue
CoGOTa
Cs137
Ce144
Ir192
Eu thickness
thickness
Ra226
Tm170
Remarks
Radioisotope
Best 12.5-,-62.5
38-,-62.5
50-300
results
182
155
1for
inspection
1.3681.35
(8-9)
Equivalent
- -- - -
steel
x21X x105
Useful xX-ray
x106
Power
105 105
105
(av)
(max.)
Source
that the gamma radiation may be directed in the form of a beam. Aluminium 3-50 steel
mm (mm)
:5 lead
of
range = 1.25 mm
for
Directional beam devices capable of safety housing multi-curie
sources are available commercially. It is recommended that such
devices be used when the activity of the source is greater than 1
curie of C060 or its equivalent.
The intensity of gamma ray sources is lower compared with
an X-ray tube. Consequently, much longer exposures are required
for y-ray inspection. Although the exposure time can be reduced
by using faster films with lower contrast but because the y-ray
radiographs have low contrast so the use of faster films do not
produce the desired result.
Despite the differences in quality of radiation from y-ray and
x-ray sources, the same types of film and intensifying "Screens are
applicable. Likewise, the processing film follows the same
procedure as that used in x-ray radiography.
'"The majorareadvantages
radiographies summed upand limitation
in Table 10.8. of x-ray and y-ray
ABSORPTION: In their interaction with matter, gamma rays
behave essentially like x-rays. The half-value thicknesses for
several materials and radioisotopes are listed in table 10.6; where Table 10.8. Advantages and Limitations of X-rays and
the half-value thickness is the thickness of absorber required to Gamma Rays for Radiography
decrease the intensity of a beam to one-half the incident intensity.
> X-Ray Units I Radioisotopes
Advantages
Table 10.6. Half-Value Thicknesses of Different
Materials for Some Popular Isotopes 1. Radiation can be shut off. 1. Small and portable.
I Lead
Material
Aluminium
Concrete
Water
68.58I
22.10
53.34
91.44
134.62
Ir192 I 6.35 48.26
4.83
Cs127
1 II
81.28 12.45
55.88 Co6O 2. adjustable.
Penetrating power (KV) is 2. No electric power and water supply
required.
s. r-------.
1. 17.27I
I Steel I 11.18
I 3. Can be used on all metals. 3. No electric shock hazards.
4. Provides radiographs with good 4. Rugged.
contrast and sensitivity. 5. Low initial cost.
40.64 ~
6. High penetrating power.
7. Access into small cavities.
8. Low maintenance costs.
9. High constrast on films possible.
Limitations
1, High initial cost.
1. Radiation cannot be shut off even
during storage.
Gamma Ray Sources: Various gamma ray sources, useful in 2. Requires high voltage pOwer
supply. 2. Radiation hazard, if improperly
radiography, and their specific characteristics are listed in Table handled.
10.7. 3. Requires cooling water supply.
3. Penetrating power cannot be
adjusted.

(contd.)
 •..•.,.,.- ..-".--.-
......... " .....• ..-- ---

.. "-:--::'::__:.=~_.,__ "._._--::
__
:;.:,~.::;,:,"~;;;;;;;:::;';:'::;;;.;::~"-
"~~,,,,,,-~""""=C;:x="""'.- •., '--" ...•..•.•~.".•_--_.--

Weld Inspection and Quality Control 745

744
!i.Ray Units
4. Equipment comparatively fragile.
5. Less portable.
6. Tube head usually large in size.
7. Electrical hazard from high voltage.
8. Radiation hazard during operation.
10.4.6. Ultrasonic Testing
Vibrational waves of high frequency that is frequency above
the hearing range of normal human ear are called ultrasonic
waves. This term generally includes all waves having a frequency
greater than about 20,000 hertz or CPS (cycles per second).
Ultrasonic testing is a NDT method in which ultrasonic waves
are introduced into a test object to detect and locate surface and
internal defects or discontinuiti8s. This is because when a sound
beam is directed into the test object it is reflected at interfaces
and discontinuities or defects. The reflected beam is detected and
analysed to detect the presence and location of defects.
10.4.6.1. Types of Wave Form
There are three basic modes of propagating sound through
metals viz., (i) longitudinal, (ii) transverse, and (iii) surface waves.
In the first two modes, waves are propagated by the displacement
of successive atoms or molecules in the metal.
Longitudinal Waves: Sometimes called compressional or
straight waves, represent the simplest wave mode. This wave form
exists when the motions of the particles are parallel to the
mInt
t a I Longitudinal
Fig. 10.31. Particle motion in : (a) longitudinal ultrasonic waves, and
(b) transverse or shear ultrasonic waves.
Welding Engineering and Technology
Radioisotopes direction of sound beam propagation, as shown in Fig. lO.31(a).
Longitudinal waves have relatively high velocity and short wave
4. Radioisotope decays in strength
with time, requiring recalibration length. As a result energy can be focused into a sharp beam with a
and replacement. minimum divergence. They are readily propagated through water.
5. Radiographic contrast generally In Transverse or Shear Waves the principal particle motion is
lower than with X-rays. perpendicular to the direction of sound beam propagation, as shown
6. Cannot be used on all metals, for in Fig. 10.31(b).The velocity of these waves is approximately half
example difficult to use on that of longitudinal waves. Advantages of this wave form are:
aluminium. (i) lower velocities allow easier electronic timings;
(ii) greater sensitivity to small indications.
However, these waves can be easily distorted, and cannot be
propagated through a liquid medium like water.
Surface Waves: These are also referred to as .Rayleigh Waves
and are propagated along the metal surface and are similar to
waves on the surface of the water. These waves have very little
movement below the surface into the metal and are, therefore, not
used for examination of welded and brazed joints.
Ultrasonic longitudinal waves are generally limited in use to
detecting inclusions and lamellar type defects in base metal while
shear waves are most useful in the detection of weld disconti-
nuities because of their ability to furnish three dimensional
coordinates for flaw location, orientation and characterization.
The sensitivity of shear waves is also about double that of
longitudinal waves for the same frequency and search unit size.
It is advisable to pretest the zones in base metal adjacent to
a weld with longitudinal waves to ensure that the base metal does
not contain weld defects that would interfere with shear wave
evaluation of the weld.
10.4.6.2. Wave Generation
The ultrasonic waves are generated by a device called
Transducer
transducer which contains a crystal that vibrates at a natural
frequency in the required range and produces the ultrasonic
sound. Thus, a transducer can change electric energy to sound
energy and vice-versa.
Crystals that vibrate in sequence with an alternating current are
called piezoelectric material, for example quartz. When subjected to
an a.c., a piezoelectric material, will vibrate. An ultrafast a.c.
produces ultrafast crystal vibrations and, therefore, ultrahigh-
frequency sound. The transmitting and receiving crystals used in
transducers are of barium titinate or lead zirconate titinate.
10.4.6.3. Test Frequency
The sound wave frequencies used for weld inspection are
between 1 and 6 MHz, beyond the audiable range. Most weld
waves I b I Transverse waves
746 Welding Engineering and Technology
Weld Inspection and Quality Control 747
testing is performed at 2.25 MHz. Higher frequencies e.g. 5 MHz
will produce small sharp sound beams useful in locating and (ii) By increasing the wave frequency ..
evaluating defects in thin wall weldments. The first method has its obvious physical limitations so the
The frequencies selected for a given test depend upon the second option is normally opted for. That is why most ultrasonic
application, the lower ones being used for coarse-grained cast testers operate with a fairly small transducer and a high
frequency.
metals and rough surfaces, while higher frequencies are used to
detect fine defects such as voids, cracks, non-metallic inclusions 10.4.6.6.Operating Systems
where access from only one side is possible. The smallest flaw size Usually three
that ultrasonic testing will detect is about one-half the sound testing. These are: different systems are employed for ultrasonic
wavelength. (i) Pulse Echo System.
10.4.6.4.Couplants (ii) Through Transmission System,
Ultra-high frequency sound in the range of 1 to 6 MHz cannot (iU) Resonance System.
be heard by the human ear and sound in this frequency range (i) Pulse Echo System: The pulse echo technique is the most
won't even travel very far through the air. That is why liquids
called couplants are applied to the material surface under test to widely used ultrasonic flaw detection technique. In this inspection
get as much of the sound as possible into the work. Typical method a pulsed ultrasonic beam is sent through the couplant into
couplants are water, grease, glycerine, light oil, cellulose-gum Transmitting and
powder mixed with water, and some proprietary materials. receiving crystal
Tronmitting
crystal
A weldment must be smooth and flat to allow intimate contact
between the transducer and the testpiece. Weld spatter, slag, and Receiving
other surface roughening elements should be removed. Weld crystal
reinforcement mayor may not be removed depending upon the
technique of testing.
A
10.4.6.5.Testing Procedure
In ultrasonic testing, transducer is placed on the surface of
the object under test. A couplant is used to help make good contact
between the transducer and the surface of the material. A couplant
performs two major functions, viz., (0)
I b)
(i) it removes the air from between the transducer and the Material
under test
testpiece, and
(ii) it provides medium for the transfer of the sound vibra-
tions. Echo from flaw

When sound waves vibrate through the medium to be tested, Reflection from
the process of attenuation occurs. Attenuation is the loss in a wave lower surface
form resulting from the density of the material. Another problem
of major importance in ultrasonic testing is the divergenc~ of the
beam, or the sound wave. as it travels through the material. As
the wave travels, it diverges or spreads and becomes wider and
consequently becomes weaker. There are two possible ways to
control beam divergence:
(i) By increasing the size of the transducer;
Ic)
Fig. 10.32. Ultrasonic testing: (a) Pulse echo technique, (b) shear wave
technique, and (c) cathode ray tube display with a linear time base (a : b :: A : 8).
748 Welding Engineering and Technology
the specimen. First the beam is reflected from the top of the test
component and then from the opposite face or the back of the
component. The echoes are picked up by a transducer, as shown
in Fig. 10.32(a). The transmitting transducer often also serves as
the receiving transducer, or a separate transducer may be used.
A discontinuity or a flaw in the specimen will also send back an
echo. The time intervals that elapse between the initial pulse and
the arrival of the echoes are measured with a eRa (cathode ray
oscilloscope). In the echo pattern a flaw can be recognised by the
relative position and amplitude of its echo. Although it is a
common practice to use echo amplitude as a measure of defect
severity but now it is recognised that amplitude is also governed
by other factors, pa.rticu1arly flaw orientation and roughness.
There may also be a number of multiple reflections displayed in
the echo pattern.
The resolution in the pulse echo method depends on the
duration of the ultrasonic pulses. The shorter the duration of the
pulse, the thinner the specimen that can be successfully tested.
The best reflections are obtained at normal incidence to the
reflecting surface. For adequate deteC'tion of planar !laws, the
ultrasonic beam must be generally normal to the plane of the flaw,
although if its surface is rough it may be detected from oblique
incidence by angle pulse echo or shear wave technique shown in
Fig. lO.32(b). The shear wave technique is used primarily for
testing thin components. The wave enters the specimen at an angle
and travels in a zig-zag path between top and bottom of sheet. As
in the case of normal incidence so in the shear wave technique,
echoes are produced by flaws and by the edges of the sheet and are
displayed on the eRG screen as shown in Fig. lO.32(c).
(ii) Through-Transmission System: The through trans- mis-
sion system requires the use of two transducers; one for sending
the sound waves, and the other for receiving them. As in pulse echo
system, short pulses of ultrasonic sound are transmitted into and
through the test component. However, unlike the pulse echo sys-
tem, the echoes returning to the sending transducer are ignored.
The amplitude of the transmitted beam received by the receiving
transducer is measured. Discontinuities or defects in the sample
decrease the amplitude of the transmitted beam because of the
reflection and scattering that take place. The principle of this
technique is shown in Fig. 10.33. This method can only be applied
to items whose sides are relative or parallel to each other.
(iii) Resonance System: The resonance system makes use of
resonance phenomenon to measure thickness of material and also
I Weld Inspection and Quality Control 749
Coaxial
cables
I ( RT
and
power Sending
source transducer
Flaw or
'-
discontinuity
(ouplant
Receiving
transducer
Fig. 10.33. Schematic representation of through-transmission
ultrasonic system for flaw detection.
to detect gross discontinuities. In this system ultrasonic waves
from the transducer are transmitted into the testpiece under
inspection and the reflected beam is re.ceived by the same
transducer as in pulse echo system, however the waves are always
continuous longitudinal waves. Wave frequency is varied until
standing waves are set up in the system, as shown in Fig. 10.34,
causing the it~m to resonate or to vibrate at a greater amplitude.
When resonance occurs, there is increase in the energy drawn by
the transducer and that can be indicated by a suitable meter or
by a display on the eRa screen. Resonance occurs whenever the
thickness of the specimen is equal to an integral number of half
wavelengths of the ultrasonic waves. Fig. lO.34(b) shows some of
the different ways in which specimen can vibrate.
Achange in 'the resonant frequency, which cannot be accounted
for with the change in material thickness, is usually a disconti-
nuity or a flaw in the test material itself,
The resonance technique may be used on pieces having two
opposite sides which are smooth and parallel, such as plates,
sheets) tubes, blocks, or bars with squared ends. It may also be
 I
750 Welding Engineering and Technology ~
(

G
Weld Inspection and Quality Control 751
[

Coble
CRO screen
Power
source

Transducer
Couplant

Test
specimen

Standing
wove

Transducer (a) Resonance ultrasonic system Fig. 10.35. Harmonic patterns in resonance ultrasonic system: (a) for solid and
laminated material. (b) for a solid piece with inclined flaw, and (e) harmonics for
a material with a flaw parallel to transducer contact surface.
Incident wave
frequencies. The resonance peaks for the total thickness of the test
' ..... -"" ...•. piece do not appear on the screen, and the presence of flaw is
.....••... ~ indicated. The depth to the flaw below the transducer contact
I=- L =0·5 A-=-I
Reflected wave
\--L =1.5}..--I surface may be determined from the instrument readings, as
shown in Fig. lO.35(c). When the transducer is moved over the

V>q
\--L --I = A
~
..•...

~L=2)'
" --

~
simulated parallel flaw, the appearance of the harmonic pattern
on the screen changes. Comparing Fig. lO.35(a) and lO.35(c), the
resonance peaks are farther apart is the former than the latter
because of greater thickness that is set into motion in the former.
Fig. 10.36 illustrates a case where the flaw is perpendicular
( b) Standing wave patterns
to the surface, the effect on the thickness indications will be small
as shown in Fig. 1O.36(a). However, if the width of the flaw is
Fig. 10.34. Schematic'illustration of: (u) Resonance Ultrasonic system, and
(b) Some of the standing wave patterns in solid material.

used on cylinders 13 mm-915 mm in diameter. This technique for

flaw detection consists of moving the transducer over the specimen
CRO
and watching for significant changes in the resonance peaks.
screen
In Fig. 10.35(a), the crystal (transducer) is over a sound (i.e.
defect-free) section of a steel block and a large number of harmonic
peaks are visible on the screen. When the transducer is over a Test piecr'
non-parallel saw cut, Fig. lO.35(b), the thickness indications on Saw cuts
the screen disappear. A· second resonance technique for flaw ( flaws)
detection may locate and give information on the depth of the flaw. (a)
It can be applied only in those cases where the flaw is flat and
parallel to the transducer contact surface. In this case the metal
between the flaw and transducer may vibrate at its own resonant Fig. 10.36. Resonance harmonics for flaws perpendicular to the transducer
contact surface: (a) when the flaw is narrow. and (b) when the flaw is wide.
152 Welding Engineering and Technology Weld Inspection and Quality Control
753
~..... ~ . -
larger compared to the transducer, an appreciable decrease in
resonance indications may be obtained as shown in Fig. 10.36(b). ~ll) hIgher frequencies may be used so that smaller defects
Factors which sometimes make it impossible to use the closer to the surface can be detected,
resonance technique are the type of material, excessive thickness Uii) it permits angular placement of the transducer, so that
variation in the material within the area contacted by the the beam can be directed at normal to the areas other.
transducer, and closely adherant scale. Excessive thickness wise inaccessible, for contact scanning,
variations in the material within the area contacted by the (iv) it is easily adaptable for automatic and semi-automatic
scanning.
transducer may be due to severe corrosion which creates deep pits.
Inspite of these limitations the resonance technique is sometimes Some of the limitations of immersed scanning are:
used for flaw detection. Typical flaws detected using this technique (i) the specimen must be immersed in a couplant; large
include laminations, shrinkage cavities, porosity, and cracks. Even parts sometimes cannot be conveniently immersed,
intergranular cracks and some carbides are detectable.
(ii) sound energy is lost at the interfaces, necessitating the
]0.4.6.7. Reflection Technique use of more powerful equipment,
In general it is advisable to use the reflection technique (iii) precise orientation of the transducer is required to direct
wherever it can be applied. However, this technique cannot be the sound beam to the specific area to be inspected,
used for testing thin specimens. This is due to the fact that in 1 (iv) diffraction will occur if the sound enters the testpiece
microsecond an ultrasonic wave will travel approximately 6 mm. surface may
at other than normal incidence and mode con.
With thin material the reflected and initial pulses will not be version occur.
spread out sufficiently to see any reflections from defects in the 10.4.6.9.Types of Scanning
specimens. Even then the reflection or pulse echo technique is the Lamination
most widely used ultrasonic testing technique. Typical flaws
detectable by this technique include cracks, inclusions, lamina-
tions, partial penetration welds, voids, and porosity.
10.4.6.8.Scanning Methods Plate
Scanning is the general terms used for the process of moving
the transducer over or along the test specimen. Scanning can be
divided into two types, viz.,
(i) Contact scanning,
(ii) Immersed scanning. lamination in plate
In contact scanning the transducer directly contacts the
specimen through a thin film ofoil or some other suitable couplant. Fro nt surface
The contact technique is widely used because it is well adapted
for manual scanning, and the portability of the transducer makes
it useful for field and preventive maintenance inspection.
A Scan of lamination A l\l~t: Back surface

In immersed scanning the specimen and the transducer are

both immersed in the couplant but are separated from each other. B Scan of lamination Top of plate
The one major requirement for immersed scanning is that the back -Flaw
surface of the testpiece be supported so that the total contact with ~ Bottom of plate
the container bottom is avoided. ( Scan of lamination
The advantages of immersed scanning are:
(i) it permits inspection of rough surfaces without excessive ~~ I Area
seen from
of flow asJ
above
wear of the transducer,
Fig. 10.37. Ultrasonic test data in three Scanning modes.
 Welding Engineering and Technology Weld "Inspection and Quality Control 755

The B Scan shows the reflection from the top and bottom of
the specimen and from flaws as the transducer moves along a line.
The B scan gives a cross-sectional view of the specimen along a
given line as shown in Fig. lO.38(b). The advantage of this type of
presentation is that both the length of the flaw and its depth below
the surface are revealed. The B Scan is used mostly in the medical
applications of ultrasonic testing and is generally not employed
for NDT of weldments.
The C Scan is analogous to a radiograph. The external details
of the test specimen are projected on to a plane that is the CRO
Square wave screen. In the C scan front and back surface reflections are not
used, instead the reflection from the flaw is the only one that is
--,LJrt LJrt L.Jc::- ftime
orm orasdistance
indicator
projected, as shown in Fig. lO.38(c).
A scan
The only other presentation that may be displayed on the CRO
screen is the square wave markers located below the horizontal
Fron t surface sweep (Fig. lO.38(a» on all the visual display scans. The square
wave markers are used as units of time or as units of distance to
Flaw relate the scan to some increment of time or distance.
11'1 HP i~!J\~:\\\l\~:
if! Flaw
10.4.6.10.Typical Results for Pulse Echo Method
II.? ".I','I"1~ll
\-I.II
Jj: I
\11\\ 11;1 Back sur faCE The reflection from the rear surface of the specimen is usually
called the back reflection. Reflections from defects will be indicated
B scan C scan between the initial pulse and the first back reflection indication.
Fig. 10.38. Schematic representation of ultrasonic data by different scanning Reflections from shoulders, flanges, or fillets may also be indicated
modes: (a) A scan, (b) B scan, and (c) C scan. to the left of the back reflection. Depending upon the types of
ultrasonic waves used there will be different indications.
(if any), their positions, and also gives the estimate of their sizes.
The time versus amplitude, as shown in Fig. 10.38(a) is the most (A) Typical Indications with Longitudinal Waues: Some of the
efficient way of revealing the existence of discontinuities. From the typical indications with longitudinal waves are:
location of the pips, the relative depth and size of a flaw in the 1. A sharp indication, Fig. 10.39, usually signifies a crack or
material can be determined. The interpretation ofthis scan is easy. other defect in a position approximately parallel to the face of the
The first pulse indication or pip appears at the left of the CRO transducer.
screen and the back surface reflection is on the right hand side of
the screen. The discontinuities are shown somewhere between the
initial pulse, the front surface and the back surface reflection pip.
The height of the pip on the screen represents the amplitude of the
wave reflection coming from the material being tested. Comparing
the height of the flaw pip with pips from known defects from a
material can establish the size of the defect. The distance from the Fig. 10.39. Pulse echo indication with longitudinal waves for
a defect parallel to the face of the transducer.
pulse to the flaw pip and the distance from the flaw pip to the back
surface reflection are proportional to the elapsed time it takes the 2. A bulbous indication, Fig. 10.40, usually signifies a defect
pulse to o;ravelthrough the material. However, during the initial whose surface lies at an angle to the face of the transducer. It may
pulse, the reflection of the sound wave may be blocked by a defect be either a crack or a void. A discontinuity at an angle to the face
directly under the transdJcer, which can cause a dead zone at the of the transducer will cause a shifting of the indication as the
front edge of the test specimen. Therefore, the flaw pip should be transducer is moved along the specimen.
determined from the reflected back surface of the specimen.

Welding Engineering and Technology
756
Fig. 10.40. Pulse echo indication with longitudinal waves for a
defect at an angle to the face of the transducer.
3. A scattering of sharp indications, Fig. 10.41, which fade in
and out as the transducer is moved usually signifies flaking.
j --.....Ll , ,,,,~ 1....-4-
Fig. 10.41. Pulse echo indicat.ion of flaking.
\ H.'" b
Fig. 10.42. Pulse echo indication of coarse porosity.
4. A scattering of bulbous indications, Fig. 10.42, which fades
in and out as the transducer is moved, usually signifies coarse
porosity.
5. Numerous small overlaid indications, Fig. 10.43, which fade
in and out, and are sharp, usually signify segregation.
Fig. 10.43. Pulse echo indication of segregation.
6. In Fig. 10.44, the path of the reflection from the flange is
shorter than the path to the back face, and hence will be'indicated
to the left of the back reflection. This signal may be mistaken for
Fig. 10.44. Effect of flange on pulse echo indication.
Weld Inspection and Quality Control 757
a defect indication. Skilled operator is required to analyse such
indications. Multiple reflections may appear on the viewing screen
when testing parts having irregular shapes and should not be
confused with multiple backreflections. Ultrasonic vibrations tend
to spread out, giving reflections from the flange and shoulders.
7. In Fig. 10.45, the path to the fillet is larger than that to the
back face of the test specimen. Hence the indication from the fillet
Fig. 10.45. Reflection from fillet as pulse echo indication.
will fall to the right of the first reflection indications. Successive
back reflection indications are equally spaced.
8. A sudden increase in the number of back reflections, Fig.
10.46, without a corresponding decrease in the 'length of the
specimen is usually indicative of the presence of lamjnations.
111ill1t#
Fig. 10.46. Pulse echo indication of lamination ur lack of bond.
9. Reduction, fading, or 10ss of back reflection as the
transducer is moved along the specimen can be due to crack 1..l.t
right angles to the face ofthe transducer. This condition may cause
complete lack of back reflection, when the transducer is directly
above the crack. This is particularly true of such cracks when they
extend into, or close to, the surface of the specimen. An angular
discontinuity or crack whose upper face is at such an angle as to
reflect the vibrations in a direction away from the transducer
causes loss of the back reflection.
Loss of back reflection may also occur when the ultrasonic
vibrations are travelling completely through the specimen' and
into the material upon which it is resting.
~
Fig. 10.47. Lack of back reflection as pulse echo indication.
:.-~;.,:±.~~:,.:~~~~:~.~.Es.q~~=:::~:::~~;' ~1~ -~--,",.'"

f
I
158 Welding Engineering and Technology f
Weld Inspection and Quality Control 759

If the back reflection fails to appear on the screen, as shown

in Fig. 10.47, regardless of the position of transducer, it may
indicate one of the following conditions.
(i) The vibrations are not penetrating to the far side of the
material because of poor surface condition.
(ii) The vibrations are not penetrating to the far side be-
cause of the use of too high a frequency.
(iii) The opposite side of the specimen is rough or pitted.
(iv) The transducer is not in good contact with the test
specimen, which may be because of poor handling or
improper couplant.
(v) Laminations in the material close to the testing surface
also cause loss of back reflection.
(vi) If the opposite side is angular, it will deflect t.he reflec-
tionsaway from the transducer.
10. Porous areas in the material may cause the back reflection
to fade as the transducer passes directly over the area. The back
reflection will reappear when the transducer is moved away from Fig. 10.49. Schematic representation of path of shear waves in a pipe.
the surface directly above the porosity. Large grain size may affect
the back reflection in a manner similar to porosity. pipe is rotating the flaw reflection is readily perceived on the
11.. A noticeable decrease in the amplitude or number of screen by its movement to the right or left, depending upon the
successive back reflections, Fig. 10.48, may also be indicative of direction of pipe rotation. Any false or spurious indications remain
fixed on the screen. Notches on the inside and outside of a pipe
porosity, large grain size, or segregation. are often used as a calibration standard. Notches 3% of the wall
thickness or 0.1 mm deep, whichever,is larger, are commonly u~ed.
This technique, however, cannot be used on small diarileter tubing
because of the blanking of the receiver at the instant of pulse
transmission.
Usually the shear wave transducer is placed in direct contact
F'ig. 10.48. Pulse echo indication of porosity, large grain size. or segregation. with the specimen, but shear waves are also used in immersion
Care must be taken in attempting to deduce the size of a defect testing.
from the amplitude of the back reflection pip on the CRO screen. Weld Inspection
The orientation as well as the size of the defect influence the size The weld bead is often too rough for proper coupling using
of the echo. The defect appears far less serious in the side test contact scanning; therefore it can be used only after the bead has
than it does from the top, this is because the reflecting surface is been removed by machining. Fig. 10.50 shows two techniques for
greater when the defect is perpendicular to the sound beam. inspecting butt welds in plates. In the first technique, Fig.
(B) Shear Wave Applications: Ultrasonic shear or trans- lO.50(a), defects p.ear the surface will not be resolved. Another
verse waves travel in a zig-zag path, as is shown in Fig. lO.32(b). reason why the first technique is unsatisfactory is that a crack
Shear waves will travel in a pipe.by the path shown in Fig. 10.49, will give no reflection if the ultrasonic beam is parallel to the plane
and have been used successfully for flaw detection in pipt~s.The of the crack. The second technique, Fig. 10.50(b), can sometimes
technique used in testing pipe is to ,hold ,the transducer firmly be used and gives satisfactory results. Another technique fOf
against the pipe as the pipe is being rotated. The transducer is testing butt welds using shear waves, is shown in Fig. 10.51. It is
slowly moved laterally. This assures complete coverage of the pipe necessary to move the transducer back and forth while moving the
and aids in interpreting the pattern on the CRO screen. Since the
 d
Weld Inspection and Quality Control 761

b
760 Welding Engineering and Technology
t, - T .

W~ld

(a I
\
I ~~~/ I
(a) Testing technique
T - Transducer T, ,T2 - Transducers

(b)

Fig. 10.50. Two techniques for ultrasonic testing of butt welds in plates.
0, Oz
( b) Display of result
Flaw indications Fig. 10.53. Twin transducer technique for ultrasonic testing of butt welds:
(a) procedure, and (b) indication on CRO screen.

screen is shown in Fig. 10.53(b). The defect indication, D1, is the

Fig. 10.51. Testing of butt weld using shear waves.
:]:t= ~
~dU'"
Fig. 10.52. Pulse echo techniques for testing butt welds in I, Land T-sections.
transducer parallel to the weld. Any loose scale, rust, or dirt must
be removed for a distance of approximately 150 mm from the weld.
It is often advisable to inspect a weld from both sides. Fig. 10.52
shows three techniques for testing butt welds in I, L, and T joints.
Fig. 10.53(a) shows twin transducer method of testing butt
welds. The two transducers are placed on either side of the weld
and are electrically connected in parallel so that each transducer
transmits as well as receives. The pattern observep on the CRO
reflection fr1m the defect received by transducer T1; D2 is the
reflection from the defect received by transducer, T2• The trans-
mitted energy received by both transducers is shown midway
between the two defect indications. The size of the defect can be
estimated by comparing the ratio of the size of the reflected pulse
to the size of the transmitted pulse, to the ratio of the size of the
defect to the cross-section
\
of the beam.
10.4.7.Acoustic Emission Testing
This NDT process relies upon the theory that as flaws
propagate in a material under stress, a sound"is emitted. This
sound (or metal cry) is analysed by a monitor and interpreted as
to the severity of defect. Material inspection consists of attaching
piezoelectric sensors (transducers) to the workpiece, applying a
specific 'level of stress and determining if any flaw becomes active.
As flaws become active, acoustic energy will be emitted. Static
defects cannot be located by this process. An extra transducer can
be placed appropriately to eliminate any extraneous signal (noise)
from being considered.
Acoustic emission (AE) test can detect very small active flaws.
The sound energy is radiated out in a spherical direction from the
flaw. Multiple sensors can be used to analyse these signals and
provide information on the location. Flaws moving even a very
small dimension (of the order of25 x 10-12 mm) in the proper time
range can be recorded. Clusters of atoms with a size range of
25 x 10-6 mm can be detected as they move or relocate under
stress. As the movement of crystallographic planes becomes more
dynamic, sound energy level warns of impending failure. The
762 Welding Engineering and Technology Weld Inspection and Quality Control 763

applied stress can be of any type that causes movement in the
testpiece. For example, thermal stress that appears as a result of
the heat from welding may cause a flaw to become active. Fig.
10.54 shows the schematic illustration of the set-up for AE test for
The recorded data is plotted in the form of a number of stress
cycles versus number of AE surges as shown in Fig. 10.55. This
1500

Visual Monitor Results of meosu ring acoustic response

f 800 of crack propagation in the ring.
700
< •...

",' 600
III
C7\
'-
Recorder ~ SOD
c:
o
Vi
'"
400
Sound pick up 'e
III 300
.•..
Sensor 1200 -1000 KHz I c
~ 200
E
:;,
Underbead crack propagating as z 100

-
a result of rapid cooling will
emit sound energy o
o
500 1000
Ia I Set- up
Number of loading cycles
D/A Recording
comerter device Fig. 10.55. A typical record of number of loading cycles
versus number of AE surges ..
data pertains to the AE detection of fatigue crack growth in a
Filter welded joint of ring type test specimen under cyclicloading. Strong
AE in the first phases of loading could have been caused by initial
settling of the work elements which may be associated with local
softening and hardening in the material of the ring where a multi
axial stress state is developed.
Oscilloscope
( b 1 Block diagram The increase of AE after approximately 500 cycles of loading
Fig. 10.54. Acoustic Emission (AE) testing: (a) Schematic set-up, and is attributed to the formation of the plastic zone at the notch tip
(b) block diagram of essential elements. while the reduction of AE at the 1000th cycle shows that the
formation of plastic zone no longer takes place. The increase of AE
monitoring underbead crack propagation as a result of rapid after approximately 1200 cycles of loading probably shows the
cooling during and after welding. formation of a fatigue crack which propagates as the cyclicloading
In all AE testing the surface of the material is cleaned to is continued.
remove contaminants and rust prior to positioning the piezo·· 10.4.8. Comparison of NDT Methods
electric sensor (200-1000 KHz) which is usually around 1 cm in
diameter. To improve tfansfer of AE waves into the sensor, it is The equipment needed, the field of application with advan-
necessary to use a coJ.plant (for example, wax) and possibly a tages and Iimit{ltions of all the NDT methods discussed so far in
compressive spring for magnetic clamping. this chapter
of these are summed up in Table 10.9 for relative assessment
methods.
 --l
Table 10.9. Comparison of NDT Methods for Weldments
be further examined visually.
inspection. Magnelic particle inspection' '"
have
rulers, which
i.e.,
Economical, expedient,
optical Expedient
some
Welds
Weld
discontinuities
to
surface,
seams.
Most
requiresweld
Portable,
suitable
Relatively
tion
voids
dye Advantages
near
discontinuities
theApplications
surface
i.e.,
equipment
training
May be and
slightly on
relatively
for used
penetrants,
Limitations
inspection
observerlinspector
applications.
surface
the
-Limited
relatively
cracks. little
Parts
Surface
economical
Must
require
out
after
relatively
on surface.
some open
indications.
cracks,
metal
discontinuities porosity,
mask
to results.
discontinuities.
to
large
Parts
open
ine~pensive
be
from must or
conductive
extemal
toequipment
films
and
inspection.
inspection.
ismagnetic
considered
all
rejectable
subsurface. littleporousthe
be
such
applied
parts Thick
non-porous
Results
easily
dnergy
may
must
hide or
as
expedient.
to to be
equipment.
be surface
visual
cleaned
coatings,
surfaces
portable.
indications.
Mostparticle
are
except
interpreted.
Indications
be
coatings
materials.
canfor many
Some for
preserved
cleaned
rejectable
materials. light
on
conditions
acuity
before
ferromagnetic
Inspec- scale,
demagnelized
can
Unlike
detect Requires
and
also
may source.
before
defects.
Shallow of and Indications
transparent
Bleed
depth
only.
smearedthe
materials.
after
mask
mask
.applicatior.s no tape.
electrical
may ~
nparators, light source. Equipment
applications.
n_I_~:requires
.•..•.
I••..•.v •.•• needs
otiiont
use of Inw "'n~t energy
electrical Autnmation
for most 2. Liquid Penetrant
3. Magnetic Particle

M
pr'
eq
mi
co

FI
de
em
cl
so
us'

~
c::
Pr i:.i"
IJQ
In tI1
pi =
IJQ
M
5'
III
ca III
ar .,
5'
IJQ

l
III

:;3
t)
AI :r
in _
required. Probe .does not have to be in intimate masked by part geometry due to sensitivity

II
within a test piece and sensing lack of fusion) as well as some
subsurface inclusions. Alloy contact with test piece. variations. Reference standard required.
the resulting electrical currents
content, heat treatment
(eddy) so induced with a suitable
variations, wall thickness.
probe or detector. Calibration
standards.

~
III
c::
...
=
'"
Table 10.9. contd. 'C
"" III
t)
not be adjusted. and
related
Gamma
"
Permanent
hidden
~
X-ray sources
with
orincomplete
fusion,
licensing
Radiography
records
positioned
source
interpretive
energy on inside
dimensional
output
for magneticWelds
subsurface
Most
at
of
production
(machines).
requirements
requires and
personnel.
highly
oftapes
review
thickness,
(wavelength)
Same
corrosion
Permanent
Adjustable
are expensive.
skiUed
of
penetration, enables
weld
and
accessible
a oflater
pipes,
Energy
including
etc.,
gamma
operating
defects.
results
applications
and
Applications
Advantages
efficient
for discontinuities
itbe
date.
Inspite
cracks,
levelslater
Limitations
facilities
fit-up
Radiation
asunusual
(gamma)
energy rays.
High as
issource
not
record-enables
slag,
and
objects,
Gamma ator will.
isareas
above.
defects,
much
requires
as i.e.,
well
interpretive
with
generally of
levels. awall
apparentiy
porosity,
andtechnique
requires
decay
initial
gamma
considered
periodically where
used
sources
review
special
as aexcept
by
personnel
dosages
lack
safety
over
generally
cost very
sources.
of their
sources,
gamma
higher radiation
in
monitoring
may
parties
be
required.
radiographs.
no portable.
replaced. good
hazard-requires
to
electrical
personnel.
produces
x·ray
Gamma will
a
East
of
NDT
Offers
half·lives
equipment.
skilled
radiography
qualitysourcesbe used
constantNotenergy
European
exposure
method,
Sources
special
permanent
and must
operational
radiation (above).
radiographs
hazard
have record as with
than gamma countries presumably due to the difficulty of ...
o'
=
correct Inte'P!:eta~ion of results.
III
Equipment needs =
Q.
(0
C
~
..•
'<
('l
o
=
...
.,
~

--l
'"
III
 I
I Weld Inspection and Quality Control 767
766
. ...; co "E
Welding Engineering and Technology
Q)'-:t'O
J:: a 0 ~
I 10.5. PRESSURE AND LFAK TESTING
g>-oO(fJO
=~~e~ I -

g~:2~
If Q) -

Pressure and leak testing or Tightness testing is employed for

§}5 ~ '5 -
\
8 g;; g ~ ~ P .~Z inspecting pressure vessels, piping, hydraulic lines, and open
0'--0:...; g..s U) _
containers. Whereas NDT methods are used to inspect individual
i~B~~ \ 8Z~~
I

I.~
.0 C"- ._
~ e::g
:c~~~vi
i~ Q)

I
en Q;

Q;<ncoc»e
gEE.~
~ 1::~Z
E

1!!
~
welds, tightness tests are meant for completed components. Thus,
before tightness testing, all welds should be completely finished
and finish grinding should be done, if required. Leak testing is
"O~>ca'ii~
'E
in g- ~ '" ~
\
often done just before the workpiece is painted and made ready
If!
:;:j
E8U)~o~
<II
1
oQ)"O.Q~
Ijg~g~§
Q)uQ)~() for shipment.
.~"~ ~ ~ ~ I
I~~OE~
1~~~Q).5 The soundness of welds, especially in tanks, pressure vessels,
:6
c: ~ ~ u; ~ .5
~ c: Q) U) - 'Q)Q)
and pipelines, may be tested by the application of internal
8~:sg~~ ~1ij~~=
'5 c.. U):.= 0 pressure or vacuum. Some of the well established tightness tests
~~=~e~~
~::cG~~.::
= oc55a::!? 0
00_
g-<n2:?~5
cc~Uicoo include the following:
cn<iiCC ~~
1. Kerosene test,
~~ ca.g _",-0
c: '" 2. Hydrostatic preSlSuretesting,
.~.9i'~ ~
c: ,J:J c:r Q)
:::E
cu '"
3. Pneumatic or Gas pressure test,
~~!~
i:: -
=' "0 en
4. Vacuum testing,
'6~oo. en '" "
5. Chemical or penetrant tests.
'" _>. c .C:
0>
tJ) 'r; ~
90 '" 0'-
~(O"O:I:
:::l
g ~ 8:
Q) '"
All these tests are briefly described in this section.
c: :: co
:;::;'-c 10.5.1. Kerosene Test
Ii
:u ~
cEg~
~ ..; \ g co ~.~
'" 0
%.5!\ g
.!:! <.>
"0",0-
>- al Kerosene is usually used for testing open vessels. The seam
w
.~ ~::!:
.~ under test is painted with kerosene from one side. Should the seam
~ gi~w
1!:
:5
.~ ~
1;S.!:2
~
-
~
0.
<II
::l
Q)eO
E.2.i":'
have open pores or gas pockets, it will pass through them to show
o•.. up as dark spots on the other side of the seam. For better
"':1 ai~'~7U?:-
""Q),=>,g:= <t
.=
-
0~ Q) '"
<II a. 1::
observation of the test it is a good plan to paint the opposite side
~~~U~ '"
a: .~
en 0
::!:cu-",Q.
r;."'5"ij~ loG CL
of the seam with white wash.
0> "0
~o~~o c: c: 10.5.2. Hydrostatic Pressure Testing
~ ~ ~.~ ~ \ I '§
"0 c:
C'CS

.-gs~:g~
.. c. 0
'"0
"0 .- Hydrostatic testing is often required for boilers, pressure
E~-~~E w«i
3:"'" vessels, pipelines, pipes or valves and fitting sections that can be
§
.'" 8 en -g >- .5 :~ sealed at their open ends. In fact, this test checks not only seams
l
Q.'I

.~ .8 ~ :;<Ii
.1:1
-o~_og' .~~
-'"
U <II
- en
• for tightness but also vessels or pipelines for strength.
~ b ~ ~:5 e°'s In a hydrostatic test the vessel to be tested is filled completely
;o>~_enE:>. ~ 0; e with water and all air bubbles are allowed to escape to eliminate
.S -: § ~ '15
Ui-g c: ~~.~ ~E~
a 13'~'o >-~ s g e
"::ucn
the air pockets. Mter all outlets have been closed with lids, gaskets
i and bolts, a pump or a hydraulic press builds up pressure inside
::!:.S.2CL.QCU

$ ~ .~-g ~ -6 -g vi .~"~g~~ g the vessel. The test pressure is 1.5-2 times the working pressure
~E~«I~Ecng~
cuu(/)Q»c'IJ-o
~ ~ ~~ ~.g-~ of the vessel. This pressure is held for some time, while the vessel
-.
~0'"' -.. ~
E u 'S ~ al <II 5-"0 c ~ g g cu .~ E ~ -'- is carefully examined for possible leaks along the seams. Any
C
0)'"" ~
Q)
(-. tl
.--0
.£l.
Q)
c:
UJ
c:
Q)
5- c~:=a~~~~ ~ ~ 0.. ~ ,2 ~.~
defective Epots detected are marked with chalk, chipped out after
C'CS

Q)Q3::Ju;~~d3c ui- ~,~ -= -g =' 0

EoC>Qlo"'Ol"J0 o~:a~~·~8~ the test, and rewelded.
~ .~ .:: -; -5 '0 ~ .~
.~ c. m a .c i ~ "~-5 « Q) co ~- ~

oca~·cw.g(l)~~
c: ~ ~ ~ .Stio~~'x In the case of pressure vessels, a hydrostatic pressure test is
c: .~ rJ ~ (j; ~ u;
£.E~·i:o~~Ug ~ 2
--
"§ ~~ g ~ ~ .g
supplemented by a hammer test. For this test, the vessel is
~: ~ o'6;.~ ID ~ ~u allowed to stay under the test pressure for 30 to 60 minutes, after
-s~~~~gg\~.~
Q...oraQ)ra",E(L)~
~t3~0E7;;6~O
UJ~'*~J~~~~
768 Welding Engineering and Technology
which the pressure is brought down to the working one, and blows
are struck on both sides and within 15-20 mm of the weld for the
full length of the seam with a hammer weighing 1-1.5 kg. Leaks
are discovered by the appearance of water through the seam on a
limewash coating applied over the surface. This type of test is
applied to boilers, pipelines, and pressure vessels.
Fluorescent dyes are sometimes added to the water to make
a small leak more easily visible. Ultraviolet light is used to make
the leaking liquid show up better. A black light lamp will make
the fluorescent dye stand out so that the leak is hard to miss.
Exceptions apart, only the larger defects such as centre-line
cracks in welds and pinholes are generally revealed by this test.
Fine cracks may not be indicated.
The following precautions are imperative in conducting a
hydrostatic pressure test.
(i) Clean water should be used for the test and the filling
should be slow and steady,
(ii) Cold water with temperature less than 7°C should not
be used and the freezing up of valves, etc. should be
avoided to eliminate brittle fracture risks.
10.5.3.Air Pressure or Pneumatic Testing
The air pressure test is also known as pneumatic or gas pressure
test and sometimes as Chemical Reaction Test. Mter attaching
pressure gauges, all outlets of the vessel are hermetically sealed, and
compressed gas (air, nitrogen, or an inert gas) is admitted into the
vessel from a cylinder through a pressure regulator or from a gas
line under a pressure 25% higher than the working pressure of the
vessel. Each weld is then painted with soap water (suds~ 'and
carefully inspected for bubbles which would indicate leaks.
Sometimes, the sensitivity of the test is improved by pumping
a mixture of air and 1% ammonia into the vessel. The welds are
covered with paper tapes or dressing bandage soaked with
developers, for example, aqueous solution of mercuric nitrate or
mercury nitride, ora solution of white phenolphthalein in a
mixture of alcohol and water. The ammonia finding its way
through tiny cracks in weldments, if any, reacts with the developer
and changes its colour. After 3-5 minutes dwell time under
pressure, the tape or bandage is removed from the seam and
carefully inspected. Black or violet spots on the tape or bandage
will indicate leaky spots in the weld. A set-up for the ammonia
test is shown in Fig. 10.56.
Pneumatic pressure testing is potentially a very dangerous
method and should be avoided, if possible. In case it has to be used
then the following precautions must be taken:
-
Weld Inspection and Quality Control 769
-
Ammonia
Weld
Air
Paper or bondage
Fig. 10.56. Schematic of the set-up for tightness test with ammonia-air mixture.
1. Brittle materials should not be tested by this method.
2. Gases entering the vessel under test should have a
temperature that would avoid any condensation.
3. Sealing should be adequate and the vessel should be
enclosed by wat.er, retaining walls or a pit, away from
personnel or building to avoid injury in case of blast.
4. The volume of air in the test vessel should be reduced.
if possible, to reduce its potential energy; this can be
done by placing metal or hardwood cores inside the ves-
sel.
5. Shock loading and hammering of test vessel should be
avoided. Also, inspection should be avoided whilst the
maximum pressure is on.
10.5.4.Vacuum Testing
In testing welds for soundness by a vacuum method, a vacuum
chamber, usually made of plexiglass for ease of observation, is set
up on one side of the weld to be tested, as shown in Fig. 10.57.
The weld is painted with a coat of froth forming detector (for
example, soap water), the vacuum chamber is held firmly on the
weld seam and a vacuum is produced. If the weld is unsound i.e.
leaky,astheshown.
vacuum causes air to leak out, forming soap bubble or
froth
The vacuum test is applicable not only to weldments of an open
type but also to those in which welds are accessible from one side
only.
';:~ •• :; •• c.;..';;;1:·;:;w,;- ""~:;;::~kW'''~~;'lil;,;'',~;,·,,.,,~~,
•..• _
770 Welding Engineering and Technology
Three-way cock
_ To vacuum
Au ylic
Soap bubble Vacuum chamber
Fig. 10.57. Set-up for tightness test under vacuum.'
10.5.5.Halide Test
In the Halide Test the vessel is filled with a mixture of air and
a penetrant (for example Freon-12, sulphur hexafluoride, tri-
chloroethylene, carbon tetrachloride, chloroform, iodoform, etc. or
halogen gases i.e. chlorine, fluorine, bromine, and iodine). The
pressure inside the vessel is raised slightly above atmospheric. A
hand prod (probe) is moved along the weld seam. Any penetrant
leaking out is picked up by a flaw detector which then produces
an audio-signal in earphones or causes the pointer of a meter to
deflect. The basic principle used in this detector is the fact that
the emission of positive ions from a heated (SOO°C)platinum
surface is increased in the presence of halogen gases and most of
their compounds.
This instrument is very sensitive and can detect a leak so
small wherein only 0.01 oz (0.2Scc)of Freon leaks out in an year.
This corresponds to a halogen gas concentration of 1 part per
million.
This test is employed in the case of critical st,ructures.
However, the instrument employed does not give quantitative data
and can give only an approximate indication of the size ofthe leak.
Also, this test is very sensitive to background 'gases, for example
cigarette smoke may effect the device and lead to erroneous
results.
10.5.6.Helium Test
The helium test is similar to the Halide Test'but has a number
of different techniques for its use; the probe technique and snifter
technique are more widely used. In both these techniques it is
Weld Inspection and Quality Control 771
f
necessary to have clean test specimens since dirt, moisture, scale,
and oil may easily seal comparatively large leaks.
pump 10.5.6.1.Probe Technique
In the probe technique, set-up for which is shown in Fig. 10.58,
the test vessel is continuously evacuated by the auxiliary vacuum
plastic
Helium probe
Foam rubber
System
under Auxiliary
test pump
Throttle
valve
Leak
detector
Fig. 10.58. Helium leak testing, probe technique.
pump. A continuous sample of evacuated gas, from the vessel
under test, is passed into the leak detector through the throttle
valve. The throttle valve is used to control the amount of gas being
sampled, in order not to exceed the operating pressure of the leak
detector, which is recommended by the instrument manufacturer.
High operating pressures can be tolerated but will result in some
loss of sensitivity. To detect leaks, a fine jet of helium, such as that
obtained from a hypodermic needle·, is passed over the exterior
surface of the vessel. Helium gas will be drawn into any opening
through the walls of the vessel and r~gister on the leak detector
as a visible or audible indication. By using a small stream of
helium, it is possible to locate precisely the position of an opening
and get a quantitative measure of its size by referring to standard
leak calibrated in cc/sec.
The pressure in the test vessel and the probing speed affect
the sensitivity of the test. For example, increased pressure in the
test vessel increases the sensitivity of thfl test.
10.5.6.2.Sniffer Technique
In the sniffer technique the vessel to be tested is filled with
helium or a mixture of helium and air to a pressure slightly higher
than atmospheric. The surface of the test object is then scanned
---*'-Needle used on injection-~Y-;·inge.
'.-'.~'"
_.. . """".0."_=- ~_.::

772 Weld Inspection and Quality Control

Welding Engineering and Technology

System pressurised
with He -Air mixtu re
leak
detector

Fig. 10.59. Helium leak testing by sniffer technique.

with a sniffer connected to the leak detector, as shown in Fig.
10.59. Helium flowing out through any openings will be sucked
into the leak detector system by sniffer and shown by a meter or
through audible indication.
10.5.7. Leak Testing By Radioactive Material
There are several techniques of detecting leaks by the use of
radioactive materials. In one such technique used for testing leak
tightness of nuclear plants the system to be tested is filled with a
radioisotope solution having an initial concentration of about 0.01
microcurie (Ilc) per millilitre (m!) of water. All welds, valve
bonnets, and vent plugs are wrapped with absorbent tapes. The
pressure in the system is raised to 1.5 times the design pressure
and held for 6 to 8 hours. Then the system is cycled several times
before depressurising. The tapes are removed and counts taken
with the help of radiation detector. From the counts taken the
leaks, if any, can be detected and located.
For detectirlg leaks in very long welded mains or oil pipelines,
still another technique is employed. A quantity of radioactive
solution is injected into the pipeline from the input end. After
sufficient time has elapsed for the injected solution to have
travelled 2 to 2.5 km, a battery operated radiation detector and
recorder are put into the lir..e.This unit is carried along by the
flowing liquid. Position along the line where the radioactive
solution has leaked out are located by radiation detector. Small
gamma-ray sources are placed at convenient intervals against the Adhesive
outside of the pipeline as distance markers. The detection unit is
removed at the other end of the line, and from the record the
leakage points can be detected.
10.6.NON-DESTRUCTIVE TESTING OF ADHESIVE
JOINTS
Adhesive bonding is increasingly being used in industry
because it offers considerable advantages over' other methods of
fastening in that dissimilar metals can be joined, stresses get
distributed over large areas, machining for assembly can be
eliminated, and considerable weight reduction is achieved. Fig. 10.60. Some typical defects in an adhesive joint.
 W eld Inspection and Quality Control 775
774 Welding Engineering and Technology

10.6.1.1. Porosity
Porosity is caused by entrained air and volatiles in the
adhesive, an.d is present in most bond-lines to some extent. cell wall
Honeycomb cell wall
10.6.1.2. Cracks
Cracks are caused by incorrect curing(and/or thermal
shrinkage) or by the applied stresses, either monotonic or 'cyclic. (b) Good
ia} Excellent
10.6.1.3. Voids
Skin
Voids in the adhesives are caused by entrapment of air or gas
by the pattern of laying the adhesive, or can be due to insufficient , T ~
adhesive. being applied. Large voids cannot be caused by volatiles,
unless something is very wrong with the adhesive system.
10.6.1.4. Unbonds
Adh";"]
Honeycomb
cell wall
Unbonds are a form of void often caused when adhesive is Ie I Poor
applied to one adherend only and unevenly, or where some degree Fig. 10.61. Adhesive bonds between the honeycomb cell wall ;
of cure has already taken place in the adhesive, and consist of an (a) an excellent joint, (b) a good joint, and (c) a poor joint.
area where the surfaces are close, but not touching. lO.6l(a). However, Fig. lO.61(c) shows a case in which the adhe-
10.6.1.5. Disbonds or Zero-Volume Bonds sive has formed a thin layer between the skin and the core (this
Disbonds can occur during manufacture due to the presence of a is weak and will easily break) but has run down the honeycomb
contaminant, such as grease, on an aclherend. The surfaces of such a cell wal~ and has formed a fillet. Another case in Fig. 10.61(c)
disband are touching, but will not transfer any.significant load from shows tH~t the adhesive layer on the facing skin has not wetted
the adherend to the adhesive. Disbands can also occur as a result of the cell wall and so gives a weak bond. Defects such as shown in
Fig. lO.61(c) may be found by monitoring the whole area. 017her
impact or environmental degradation after manufactm'e. forms of defect in honeycomb sandwich construction are due to
10.6.1.6. Weak Adhesive Layers lack of attachment between the core and the skin. This may be
This is less obviousbut potentially very serious defect giving poor due to several factors, or lack of adhesive, causing skin-core dis-
cohesive properties, and results from either incomplete mixing, bands. In themselves, none of these defects may prove seriously
incorrect formulation, or from insufficient cure of the adhesive.
10.6.1.7. Adhesion }'ailure detrimental
show poor preparation
to short-term
and manufacture
strength of the
andjoi~.
wil\ provide
However,
sitesthey
for
fatigue crack initiation. '
Adhesion failure or a failure of a weak bond, between the 10.6.2. NDTMe,thods Employed,
adhesive and the adherend, often results from poor surface
preparation or the presence of a contaminant on an adherend. Testing is carried out prior to bonding or afterl bonding. Befch-e
Apart from the usual lap joint, the other major form of joint the adhesive is applied to the joint, the adherend surfaces should be
used with structural adhesives is that employed in bonding a prepared by washing, abrasion, chemical etching, etc. Proper
honeycomb core to skins to form a Sandwich Construction. The cleaning ofjoint edges is crucial in mak,inga goodbond. The presence
resulting mesh of fine joints, which has to take both shear and of excessive amount of water vapour, hydrocarbons and other
direct loading (tensile or compressive), holds the structure contaminants will result in poor quality of adhering surfaces.
together. Fig. 10.61 illustrates some of the possible forms of In some high performance applications, a primer is applied to
skinlhoneycomb bonds. Fig. lO.61(a) shows a well-filled joint in the surface immediately after preparation if the part has to be
which there is a generous adhesive fillet, while in Fig. iO.61<b), stored for more than a few hours before the joint is to be made.
the adhesive has been extruded from between the skin and the The NDT methods employed for inspection after bonding
core, but has left a generous fillet which will still carry the loads include the following:
and is probably as strong if not stronger than the joint in Fig.
 Welding Engineering and Technology Weld Inspection and Quality Control 777
776

(i) Ultrasonic testing, Large differences in the acoustic impedance of the materials
cause a large proportion of the energy to be reflected. Since a
(ii) Sonic vibrations and Mechanical Impedance testing,
(iii) Thermography.
defect containing air or anyother low density substance will have /
a very low acoustic impedance relative to the adhesive or
10.6.2.1.Ultrasonic Testing adherend, the ultrasonic pulse will be almost totally reflected.
Voids and disbonds can be readily detected by ultrasonic A display of the magnitude of the reflected echoes is used to
testing and the technique has the potential for locating very small indicate the presence of defects. Such defects are usually made on
defects like porosity. As the ultrasonic pulse wave propagates A-scan. Fig. 10.62 shows an A-scan from a good single-lap adhesive
through the joint, a p&rt of its energy is reflected at each boundary joint and Fig. 10.62(b) shows an A-scan from a similar joint but
or discontinuity, the amount depending on the acoustic impedance with a disbond at the top adhesive/adherend interface. The
of the materials on' either side of the boundary. The acoustic reflections from the lower interface are no longer present in the
impedance, Z, can be 'defined as, defective joint. Also, reflections from the top adhesive/disbond
Z =cp ...(10.6) interface decay more slowly than those from the top adhesive/
adherend interface of the sound joint. since more energy is
where, c = velocity of sound in the material, and reflected at the disband.
p= material density. Because there is a large difference in the acoustic impedance
Reflections from top adherend/adhesive between air and solid materials, care has to be taken to propagate
ultrasonic energy from the transducer to the structure under test .
.-~,/ l "" interface Commonly, the structure and transducer are immersed in a water
bath. The ultrasonic wave then propagates across the water filled
gap (typically 25-100 mm depending on the transducer) into the
-g
testpiece. Alternatively, a water jet transducer can be used in
~ \ \, .. ~\ Reflections from bottom
adhesive/adherend interface
a. which the ultrasonic wave propagates along a moving column of
E
Cl water(jet). Care must be taken to ensure that the couplant (or
Ci
C anyother liquid such as water or fuel) is not allowed to penetrate
0- the defect, as the presence of the liquid reduces the reflection
Vi
coefficient and the defect becomes much more difficult to detect.
Top surface reflection
The liquid may also remain in the joint region or even be absorbed
o 1 2 3 4 5 by the structure if that is made of composites. This moisture can
Time,JJs - then cause environmental degradation, and so is to be avoided
(0) For well bonded joint wherever possible. In some cases, especially when using woven
material and absorbent fibres such as Kelvar, it is best to test dry.
A map of the defects can be produced by scanning the surface
of the structure. The amplitude of a particular echo, such as that
from the bottom adherend-adhesive interface is measured as the
QI
"0 probe traverses the structure. Changes in the echo amplitude
=:l Transducer indicate the presence of defect, and a record of defect location can
:'!::
Q. be obtained by plotting amplitude -against position i.e. in C-scan
3.zmm1-
t
E
Cl I

as shown in Fig. 10.63. The resolution of small defects, such as

Ci
c porosity, is improved by decreasing the distance between the scan
Vi
C7'
lines, but this also increases the jnspection time.
When using through-transmission technique one transducer
o 2 3 4 5 is positioned on each side of the structure and the magnitude of
Time, JJS - the transmitted signal is then used to indicate defects. The
I b) For defective joint

Fig. 10.62. A-scan records for a single lap joint; (a) well bonded. and (b) defective.
 Welding Engineering and Technology
778
~n.nnr
\Ir'.'
VOid
t, " t1\ Plan of
~ U U U U L_. Pgi~ti'~
Adherends
F""" <~''''
-~:rAdh"i"
(ross
Void
section
~
of joint
Fig. 10.63. A typical C-scan of an adhesive joint with a void.
presence of a defect causes the signal at the receiving transducer
either to reduce or to disappear. This method is often used for
inspecting the bond between the top and bottom skins and the core
of a honeycomb structure.
Ultrasonic impedance and resonance tests are also used to
detect small voids and disbands, but they are less sensitive and
will not generally detect such small defects as porosity.
10.6.2.2. Sonic Vibrations and Mechanical Impedance
Testing
Sonic 'Iribration methods, which effectively measure the local
stiffness of the structure can be used for non-destructively testing
adhesive joints. Disbonds reduce the local stiffness o(a structure
as measured perpendicular to the surface. Instruments of this type
typically operate at frequencies between 1 and 30 KI-iz,compared
with 0.1-25 MHz for ultrasonics. Usually, only disbonds or voids
can be found, the minimum detectable size depending on the type
and depth of defect. Although the minimum detectable defect size
is larger than that obtained using t.he ultrasonic techniques, the
tests are often more convenient since they do not require a
couplant between the transducer and the test structure.
Commercially available mechanical impedance measuring
instruments generally take measurements at a single pre-set
frequency, typically between 1 and 10 KHz. The impedance is
highest over good areas, and lowest over disbonded areas. As the
structure itself becomes more flexible, the impedance of a defective
Weld Inspection and Quality Control 779
zone can be higher or lower than that of a good zone, depending
on the frequency. One advantage of this type of test is that instead
of using a couplant, a dry point contact is used between the
transducer and the structure. This contact has a finite stiffness
which must be kept as high as possible, to maximise the sensitivity
of the technique.
10.6.2.3.Thermography
Thermography methods can be used to indicate the presence
of voids as they provide a high resistance to heat flow. Commonly,
the steady state transfer of heat is monitored, as shown in Fig.
10.64, using an infrared ray camera for scanning to provide a
Heat
Adherends
Adhesive -
Reverse
side --7 -- ...",-- ,
Temperature profile on reverse
side
Fig. 10.64. Schematic illustration of the principle of infrared ray
thermography for detecting voids in adhesive bonds.
pictorial impression of the voids through anomalies in the temp-
erature distribution.
The sensitivity of this method can be improved by using a
video recorder to store the transient thermal response. In this way,
defects in conducting materials whose effects on the temperature
distribution is short-lived may be detected.
X-ray radiography is not used for non-destructive testing of
adhesive bonds because disbonds cannot be detected with metal
adherends as the absorption of x-rays by the metal is very much
higher than that by the adhesive and the disbands. Neutron
radiography is more promising as the neutrons are absorbed by
the adhesive rather than the metal, but neutron sources are very
large and expensive, and are impracticable for other than specia-
lised use.
The non-destructive measurement of adhesion strength is
currently not possible.
10.7. CONCLUDING REMARKS
Although quality control in welded fabrications implies the
detection and elimination of weld defects as well as setting
procedures for avoiding their occurrences in future constructions,
780 Welding Engineering and Technology

however, now there is an increased interest in what is termed as

Fitness for Purpose (FFP) validation of welded construction. FFP
is that standard of fabrication which is purposely chosen to be the
right level of material and fabrication quality for each application
having regard to the risks and consequences of failures; it may be
contrasted with the best quality that can be achieved within a
given set of circumstances as that may be needlessly uneconomical
for service conditions which are less demanding. A characteristic
of the FFP approach is that it requires t.obe defined before hand ELEVEN
according to the known facts and by agreement with purchasers.
Thus, FFP relates well to the quality assurance (QA) approach,
since QA aims to be comprehensive, and makes provision for
Repair anc!
updating its own procedures. Maintenance Welding
Repair and maintenance work covers a very wide variety of
welding applications ranging from the routine surfacing of dipper
teeth on a power shovel or the rebuilding of track shoes on a
crawler tractor to weld repairs which are strictly one-of-a-kind and
often of an emergency nature, for example a broken power shovel
boom. Construction equipment like power shovels, particularly the
large ones, create delays in an entire operation, such as a mine,
and cost extremely large amounts of money while they are out of
operation. Similar considerations apply to oil-drilling rigs,
off-shore platforms, cement kilns, steel rolling mills, electric power
generators, and other production equipment.
The volume of work involved, in repair and maintenance
welding, can be gauged from the fact that approximately 18% of
the welders in USA are engaged in this type of work. The repair
welder needs to be a very versatile and adaptable individual to
tackle the many different jobs he may be called to deal with. At
times, the welding repair work is so specialised that the welder
can earn fat sums in comparatively short times. For example, in
oilfields independent workers called rig welders have their own
equipment and often are highly skilled welders who work on a
daily or hourly rate and make more money than most ofthe oilfield
workers they help.
The economics of weld repairing are usually very favourable
and this applies to the smallest or the largest weld repair job.
Some weld repair jobs may take only a few minutes and others
may require weeks for proper preparation and welding. Even so,
the money involved in a repair job is often much less than the cost
of a new part.
782 Welding Engineering and Technology
11.1. SELECTION OF' WELDIl"'~"GPROCESS FOR REPAIR
WORK
Weld repair is commonly used to improve, update and rework
parts so that they equal or exceed the usefulness of the original
part.
For successful repair welding, it is absolutely essential to gain
knowledge about the type, specification, or composition of the
metal that is required to be repair welded, because it has a great
influence on the service life of the repaired part.
In general a part made of any metal that can be welded can
also be repair welded or surfaced. For example, parts made oflow-
carbon and low-alloy steels can be repaired without adversely
affecting the service life of the part. On the other hand, high
carbon steels may be weld repaired but must be properly heat
treated if they are to provide adequate service life. Weld repairs
on cast iron parts subjected to repetitive heating and cooling, such
as machinery brakes, furnace sections, etc., mayor may not
provide adequate service life because they fail originally from this
type of service and thus due to metallurgical changes the weld
may fail again without providing the required service life..Thus,
except for emergency. situations, it is not wise to repair cast iron
parts of this type.
Broken, cracked or defective castings, ferrous and non-ferrous,
form a large proportion of welding repairs, but there are also worn
parts to be rebuilt and broken vehicle frames, leaky tanks or
pipeworkto be repaired. Hardfacing is another important aspect
of maintenance that concerns the welder working in quarries, steel
plants and other places where severe wear often necessitates
considerable repair welding. The unlimited variety of repair work
calls fOr not only a wide knowledge of the weldability of metals
but also of welding processes-often a.pplied under very
unfavourable working conditions.
In repair welding there are usually limitations in the selection
and use of a welding process because of the non-availability of
equipment for a one-time job and the necessity of obtaining
equipment quickly for emergency repair work. This limits the
selection and it is for this reason that the shielded metal arc
welding process, the GMAW process, the GTAW process,
oxyacetylene welding and torch brazing are most commonly used.
However, for many routine and continuous types of repair work
submerged arc welding is widely used for building up the surface
of worn parts. Also, ESW (electroslag welding) process has been
often used to repair and resurface parts for hammer mills, for
construction equipment, and for rebuilding rolls for steel mills.
Repair and Maintenance Welding 783
Thus, there is a difference in the selection of the welding process,
on the one hand, for the routine, continuing type of repair and
surfacing work, and on the other hand for the one-of-a-type or
breakdown emergency repair job.
11.2. TYPES OF WELDING REPAIRS
Since there are innumerable types of components and
structures of varied design made from different materials for a
large number of service conditions, it is not possible to standardise
the procedure for repairs. Thus, the decision about welding repair
may have to be taken under exacting conditions without any
previous knowledge to handle such jobs. Hence it is beyond the
scope of any text book to describe repair procedures for all possible
conditions and situations. However, weld repair and maintenance
work may be divided into the following categories:
1. Repair of castings and forgings,
2. Repair and maintenance of medium sized structures and
components,
3. Maintenance of earthmoving machinery and mining
equipment,
4. Repair of large structures.
11.2.1.Repair of Castings and Forgings
The reclamation of broken, worn and faulty castings and
forgings is one of the most valuable applications of welding. The
savings that can be affected by satisfactory repair are frequently
quite considerable, while the cost of repair may be very little. The
procedure adopted for rep~~irsdepends upon the type of fracture
involved in repairs. Usually three types of fractures are
encountered in such repairs.
Fig. 111. ,\ typical bar-type fracture.
784 Welding Engineering and Technology Repair and Maintenance Welding 785

The simplest is the free 'bar' fracture, shown in Fig.11.1, in

which each part can be separated for the joint to expand and
contract by welding. The weld metal in such a joint is restrained
only if the weld is very long, in which case lengthwise contraction
of the weld metal will be resisted by the joining edges and this
may cause cracking.
The second type of fracture is the same as the first one except
for its location in a wheel rim or spoke, machine frame or other
part where expansion at the fracture might cause sufficient
lengthening of the member to create stress at other points in the Fracture
casting; contraction of the fracture will also be resisted by the
remainder of the casting. For a successful weld repair of this type,
the fracture edges should be prepared and must then be separated
before welding to allow for expansion and subsequent weld
contraction.
If the casting is fairly springy, the separation may be achieved
by packing or jacks at convenient points, as indicated at A in Fig. B

Fig. 11.3. Fracture in a spoke or cast iron wheel.

11.4. A wedge is then driven into the groove so as to make a gap
of 2 to 3 mm. In welding, care must be taken to melt the wedge
--
/
/
,,
..-

(
,
I

Fig. 11.2. A typical frame fracture.

11.2 - avoiding excessive force, which may cause fractures at
other points. Another way is to apply heat to expand other parts
of the casting in order to open the joint, as shown at B. The points
at which heat should be applied should be in line with the fracture
point, and the heat must be maintained during welding.
Mter welding, jacks must be released or heated points allowed
to cool so that the welded portion contracts without restraint.
Even if the crack opened when the fracture occurred, it will Fig. 11.4. Welding-up a crack in a spoke of a pulley.
be desirable to open the crack further to allow for weld contraction. completely. The length of the wedgejutting outside the spoke must
Another example ofthis type of fracture is shown in Fig. 11.3; here be cut away. Mter the weld metal contracts, the spoke which has
the crack may be opened by jacking at A and/or heating at B, been set apart pulls together without leaving any residual
otherwise weld contraction may pull in the rim or, if the rim is stresses. Welding without inserting a wedge may result in a new
stiff enough to resist contraction, the weld may fracture again. crack either in the spoke or in the rim, because the metal of the
Alternatively, the cracked portion is cut away, and the edges are spoke will restrain the contraction of the weld metal. Instead of a
bevelled to an included angle of 60° on both sides, as sliown in Fig. wedge, ends of the cracked spoke can be set apart by heating the
adjacent spokes and a part of the rim as shown by the dotted lines
;:::~,;,:~!~,-'~:~F.'~~~~'::"~;
Welding Engineering and Technology
786
in Fig. 11.4.The usual source ofheat is the oxy-acetylene gas flame
or even charcoal in a box of perforated iron suspended suitably to
heat the desired sections.
For a boss crack, as shown at C in Fig. 11.3, heating at D will
open the fracture; while for a rim fracture, as shown in Fig. 11.5
Fig. 11.5. Rim fracture in a cast iron pulley.
heating at B will open the break; or for a large, light wheel (pulley),
heating on a line C-C across the boss may be sufficient.
The third type of fracture is where it is entirely surrounded
by metal as in cylinder block cracks, holes in crank cases, cracked
pump bodies, and so on. In such cases, localised expansion in t.he
weld zone creates a wedge effect, which may extend the 'crack or
convert it into a complete fracture. Total preheating is desirable
in order to lessen "this risk; although, using a low heat input
technique, it may be possible to make an arc welding repair with
moderate preheating, depending on the rigidity of the casting and
the welder's experience and skill. Similarly, an oxy'··acetylene
bronze weld may be made at black heat, whereas fusion welding
would necessitate heating to red heat.
11.2.L1. Repair of an Engine Crankshaft
Fig. 11.6 illustrates an engine crankshaft with a crack in one
of its crankpins [Fig. 1l.6(a)], and the steps involved in repair
welding the same. The procedure adopted is as.follows.
The cracked crankpin is entirely cut away together with parts
of the webs, and a round bar 100 mm in diameter and 200 mm
long is inserted as shown in Fig. 1l.6(b). The remaining parts of
Repair and Maintenance Welding 787
100
Crock
(a) -3
co
(b I
Welds
t,t~-~I Fig. 11.6. Sequence of repair welding in a cracked crankshaft.
the webs and the ends of the round bar are bevelled to an included
angle of 70° or 80° on two sides, i.e. in a double vee pattern. After
welding, the bar is turned on a lathe to the diameter of the
crankpin.
The grooves are gradually filled with narrow straight beads
using electrodes 4mm in diameter. During welding the shaft
should be checked at short intervals for possible deformation; this
should be done by means of a reference line scribed along the shaft
and two shifting gauges 'a' at the ends of the shaft, with their
points put on the scribed line as shown in Fig. 1l.6(c).Warping, if
any, can be controlled by changing the direction in which beads
are laid or by turning over the shaft and depositing beads from
the other side. After filling the grooves, a reinforcement is also
deposited.
A very-important point in the repair of a crankshaft is to keep
it in proper alignment. Even the slightest misalignment will spoil
the work, because the shaft will wobble. As a precaution against
misalignment, crankshafts are sometimes repaired while carried
by the engine bearings and rotated as the wdding progresses. Any
warping will immediately prevent the shaft from rotating in the
bearings. One method for correcting the warpage. is to deposit
beads on the opposite side.
 Repair and Maintenance Welding 789
Welding Engineering and Technology
788
strip and nut on the outside (for heavy gauges). Alternatively, the
11.2.1.2.Repair Welding of Cracked Plate Work patch may be held in place by tack welds.
Operators often have to weld up cracks in plate work like In some crack welding operations, precautions must be taken
boilers, vessels, etc. Before welding up such a crack, it should be to prevent cracks due to contraction of the deposited weld metal.
machined or gouged to a vee groove, as shown in Fig. 11.7, to These measures mainly aim at providing conditions under which
obtain' full penetration. the contraction is facilitated.
11.2.1.3.Preparation of Iron Castings for Welding
To attempt a repair on a casting without dismantling it from
its associated components is rarely satisfactory because
accessibility for the preparation of the fracture and control of the
welding procedure is most important.
(alOriginal crock (bl Veed -out crock Adequate cleaning of the casting is necessary for the success
of most repairs, especially for non-ferrous arc welding and oxy-
Fig. 11.7. P~eparation of a crack in plate work for repair welding. acetylene braze welding. Grease and oil may be removed with a
The crack should also be drilled at the ends to prevent its degreasing solvent or vapour, or if solvent is not available with
propagation during bevelling and welding. Whenever it is possible,
it is recommended to apply a sealing run on the root side of the
weld. Before doing so, overflows and slag should be removed from
the root of the weld.
When a plate work develops a cluster of cracks, the faulty area 90·
is cut out and a patch is welded in. Fig. 11.8 suggests a technique
Plate ~
(0 I Single vee for joints accessible
from one side only.
3
~~--~~
~

(b I Double vee edge preparation for

parts that can be turned over
and welded from both sides
4t " e. g. lever. pulley. etc.

•••
2
~ .
Fig. 11.8. Patch welding and its sequence for. plate work.
for patching. The corners of the patch are rounded, because (c) Joint groove surfaces studded
stresses are usually concentrated at sharp corners, giving rise to to improve weld metal bond.
cracks. The patch should be slightly dished so that the contraction
during welding will just about take up the dishing and straighten
out the plate, thus relieving all strains. Fig. 11.8 also shows the
sequence (indicated by 1, 2, 3 and 4) for patch welding. The four
clamps show an easy method of holding the patch in place, simply
putting a bolt through from the inside and clamping down with a Fig. 11.9. Different ways of preparing cast iron joint edges for arc welding.

Welding Engineering and Technology
790
petrol or paraffin, but only preheating will remove deeply
penetrated oil. Corroded areas should be chipped out or ground to
clean and sound metal.
Cracks may be traced by swabbing with paraffin or kerosene,
afterwards wiping dry and rubbing chalk on the surface to show
up any oil retained in the crack. Water-jacket cracks may be traced
by line-washing the suspected area and filling the water space
with kerosene mixed with a dye penetrant .Sprayed fluorescent
penetrant dyes are also excellent for crack detection. Cracks
should be 'stopped' by drilling holes about 3mm in diameter
beyond the visible ends of the cracks, and for metal over 3 mm
thick joint edges should be veed to 90° by grinding or chipping, or
after preheating (for oxy-acetylene welding) by scraping out with
a welding rod. Metal over 12mm thick should be double-veed (if
both sides are accessible); welding from both sides minimises heat
input and distortion. Gouging electrodes may be used for
preparing joints, provided that care is taken to avoid overheating.
For oxy-acetylene braze welding and all types of arc welding,
vees and flat surfaces on thick cast iron can be studded to improve
bond between the weld metal and casting, as shown in Fig. n.9(e).
This is often done when rebuilding broken gear teeth.
When a part is broken through a machined surface or gear
teeth, remachining can be minimised by preparing the joint from
the back, leaving a portion of the original fracture unbevelled for
lining up.
The various parts of broken casting should be located and held
in position by suitable clamping plates or strips of metal, by
bolting to a support or by binding with a wire, this must be done
before preheating.
11.2.1.4.Preheating for Repair Welding
Small parts can be preheated with the oxy-acetylene flame;
for larger castings use LPG or gas burners, taking care to heat
slowly and uniformly, enclosing the casting or the portion to be
heated _ in a brick or sheet-iron furnace built to suit the job; one
such furnace is shpwn in Fig. 11.10. The furnace walls should be
made high enough to allow the casting to be covered with sheet
iron or asbestos. The casting should be packed up to allow heat to
spread under it, and it should be adequately supported to prevent
sagging and distortion. If it will be necessary to move the job
during welding, provision should be made for this when the
furnace is built.
Repair and Maintenance Welding 791
C. I.or steel plate
Fire bricks
Fig. 11.10. Typical temporary preheat furnace built of
firebricks or sheet iron to suit the job.
Gas flames should be baffled so that the heat does not strike
only one spot on the casting. Machined surfaces can be protected
from the flames by suitable arrangement of the casting or by
coating the surfaces with grease and graphite. Cylinder bores can
be protected with asbestos discs at the top and bottom of the bore.
Charcoal is one of the best preheating fuels, heating slowly
and uniformly; it should be packed around and under the casting.
Coke or a blacksmith's fire is not suitable; the heat is too local and
the sulphur fumes produced by it are detrimental to the weld.
Charcoal should, however, be used only in a well-ventilated shop
as the fumes are dangerous.
For oxy-acetylene repair welding, preheat to a dark cherry red;
for arc welding or braze welding (if preheat is considered
necessary), preheat to only blaek hot.
11.2.1.5.Cooling After Repairs
After repair welding, both furnace and casting should be
covered and allowed to cool as slowly as possible. Small castings
can be transferred to a box partly filled with lime and covered to
protect from draughts and to ensu.re slow and uniform cooling.
11.2.1.6.Low Heat Input Arc Welding Procedure
Gas welding a preheated casting follows the usual
oxy-acetylene welding procedure, but arc welding a cold C<1sting
requires patience and an appreciation of the possible effects of
applying local heat to the weld point. To weld a vee joint a bead
should be deposited on one side of the bottom of the vee and then'
on the other; joining these two beads before proceeding to complete
~:.~~~ -~ --

792 Welding Engineering and Technology
f'
the joint. Use small electrodes at the lowest satisfactory amperage;
deposit stringer beads not more than 75 mm long, peen each bead,
and allow the weld zone to cool between deposits so that it never
becomes more than 'hand-hot'.
11.2.1.7. Weld Hardness
If a cold casting is arc welded with mild steel electrodes, the
weld is very likely to be hard and un machinable, due to carbon
pick up from the casting and the chilling effect of the cold casting
on the deposit. If nickel alloy electrodes are used, the weld will
only be hard if there has been excessive melting "f the joint edges;
use a technique that ensures the shallowest possible fusion i.e. the
Repair and Maintenance Welding 793
lowest dilution. Even with pure nickel electrodes, there may be
some hardness at the joint edges where the casting metal has been
chilled.
With oxy-acetylene repair welding, hard spots may be caused
by neglecting to remove oxide during or by dipping a cold welding
rod into the pool, or the weld may have been chilled by draughts
during or after welding.
For welding repairs where hardness is not objectionable or
where a large amount of weld metal is required, mild steel

A-B

I rn//~~J
A-B ~.

(- 0

Fig. 11.11. Corner pillar reinforced by gusset plate. Fig. 11.12. Corner pillar, reinforced by a flat.
794 795
______ . \~elding Engin~eri~lgand Technology Repair and Maintenance Welding

electrodes may be used and will be cheaper than a nickel alloy

electrode but the joint surfaces should first be buttered with a
nickel alloy electrode. However, the softer and more ductile weld
metal deposited by nickel alloy eledrode is more easily machinable
and is less prone to cracking.
11.2.2.Repair Welding of Rivetted Structures
Some of the typical joints encountered in rivetted structures
are shown in Figs.n.ll to 11.15. Fig. 11.11 shows the angle iron
joint of which the pillars are made. It usually fractures where o
uppermost of the three rivets weaken it. The fracture results from
the dynamic stresses, imposed on the frame if the pillar forms,
say, the part of railway cars, when the cars are slowed down
violently on the sidings and yards. The welding of a large piece of
o
o
o
o
~/
10

o
o
o
o
T

Fig. 11.14. Door frame pillar, reinforced by plate on one

side only, used for railway car.
1-- o 8-10mm thick plate to form a stay increases the rigidity of the

:t/
joint to a large extent.
The similar repair by the addition of a flat bar joggled at one
end to form a stay, is more effective the larger it is, as shown in
Fig. 11.12. This is a most satisfactory and aesthetic form of repair
which provides a rigid joint in both directions.
Similar repairs to other types of joints forming pillars in
\ I construction work of different types are illustrated in Fig. 11.13
to 11.15. Fig. 11.16 shows corroded upper part of main girder
\- - ."(U~"",. -' i channel section -- a defect that occurs in many of the open railway
cars, of old design, used to transport coal. The combination of
Fig. 11.13. Intermediate lateral pillar reinforced by plates.
196 Welding Engineering and Technology
~
See tion at AB
J
Fig. 11.15. Intermediate pillar of channel section reinforced by an angle iron.
sulphur and moisture, prevalent in the mining districts, forms
acid that attacks mainly the upper part of the main girder whilst
the remaining structure remains in reasonably good condition.
11.2.3.Maintenance of Railway Carriage Wheels
Probably one of the best known conventional weld surfacing
jobs is to build up the worn flanges of railway carriage wheels
before they are returned, so as to retain the tread diameter. In
large shops this is done by the use of automatic welding set-ups
but in small shops they do it manually. For manual surfacing the
'Wheelis placed on rollers, two to each of the two wheels and the
axle is moved manually by the welder, so that he can always work
in downhand welding position.
I Repair and Maintenance Welding 791
20·
~~
Fig. 11.16. Corroded upper part of main girder channel section replaced by a flat.
For automatic resurfacing the wheels are mostly built-up by
the automatic SAW process. The typical equipment for this
purpose incorporates a stationary suspended welding head, with
the work rotated at the speeg of welding. The head may be of the
single-arc or·the multiple-arc type (i.e. two or three arcs), or two
or more heads working simultaneously may be used.
The building-up of worn carriage wheel flanges affects an
appreciable economy of metal and labour. The relevant codes allow
but a very limited wear on wheel flanges, after which a worn tyre
should either be replaced or turned on a lathe to the nOl'malprofile
by removing a lprge amount of stock.
..
II'

11

,!

798 Welding Engineering and Technology J Repair and Maintenance Welding 799

Fig. 1l.17(a) shows the original contour of a wheel flange at finished diameter, then filling in the remaining tread area. For
A, the contour of the same flange after a period of service at B, the edge or the starting bead, the centreline of the welding wire
and the contour of the flange after reconditioning on a lathe at C. should be positioned about 8mm from the edge of the tread. If the
The cross-hatched area gives an idea of how much metal has been tread has been worn excessively, resulting in a deep groove in the
removed in turning. D shows the contour of a flange for the least centre of the tread, this area should next be built-up to the general
thickness of the tyre. level of the rest of the surface. Each successive weld bead should
B
overlap the preceding one by 50%. This procedure, as shown in
Fig. 11.18, is continued over the entire tread until sufficient metal
has been deposited to permit machining to the desired contour.
'"
.•..
C
01'0
"'01
~.c:
Cl. '"
'2 Cii :l
]
QI
"0
-.c:
'-.- <II
~
<ll
:5
(0) ( bl :§o•...•..
QI- ~
'O::):=:
0
I-
'0
•..
~]~
•..
'"
"0
....
<J
.S
oS
QI•..
gbl)
•..
.-<
'"
~<ll
.....;
:5!
<ll
..!!l
.~
";::
d:;
.!:P
<ll
cO

.•...•..
n
Qlo::)
E
E
ex:>
OCQI
Cl8 ••
J
'0
o
(c I C 01
Fig. 11.17. Building-up the worn railway carriage wheel: ~ ~ .:::
(a) different contours in wearing down and building-up, .~'j 5
(6) built.-up section of wheel. and (c) procedure for depositing weld beads .
c '0
Fig. 1l.17(b) shows a built-up flange, the blackened portion ~.[\u..
~w
o ~
being the deposited metal after machining. In building-up a wheel, •••• _
QI
0>
beads are laid all the way around the tyre in short increments
usually 200-300mm long, as shown in Fig. 1l.17(c). The best
~ol
'" Vlo
o
.•...
-£c
results are obtained in an incline.d position, with the beads laid "en
~
upwards. Each time an increment 200 to 300 mm long is deposited,
the wheel is turned in the direction of welding, and the next
increment is laid. Details of depositing individual weld beads in
resurfacing flange and tread of a carriage wheel are detailed in
the following section.
11.2.3.1.Details of Resurfacing Technique for f'lange
and Tread of a Carriage Wheel --"
01
Often wheels are machined many times without rebuilding ccen
and, as a result, are considerably under their original diameter. u:
The usual practice is to rebuild both wheels of a pair to like
diameters rather than to the original size. Separate procedures
are used to rebuild the tread area and the flange. The tread is No welding should be attempted on the work that is chilled.
resurfaced first, building up the edges to slightly over the desired If the wheels are below about 5°C, heat should be applied slowly,
800 Repair and Maintenance Welding 801
Welding Engineering and Technology

and the work warmed to above 16°C. Rapid cooling after welding
should be avoided. The wheels should not be placed where drafts
or low temperatures can cause sudden cooling. If the ambient
temperature is below 16°C, it is advisable to place the wheels in
a bin of vermiculite, mica, small particles of discarded fused
welding composition or other material that will retard cooling.
Welding interruptions should be kept to a minimum to avoid the
possibility of cracking. For example, it is not advisable to build-up
a surface partially, stop for the night, and restart in the morning.
The tread and flanges are usually rough machined, taking a
light cut. However, machining is sometimes un-necessary if the
submerged arc welding is reasonably smooth.
Since a considerable portion of the resurfacing work is done
on cylindrical surfaces, the relationship between diameter of the
work piece and the welding current or speed are important factors
to consider in preparing for the repair of these parts. These values
800 405 85 105 125 145 165
35 ------ Workpiece dio.
welding cur'~
SAW is also used for rebuilding on flat parts such as backup
chucks, unloading chutes, conveyer wear strips, coke pushers; and
mill guides, etc.
11.2.4.Filling-in Large Holes
Filling- in large holes may also be classed as building-up. This
is done with padding welds. Despite its apparent simplicity, this
technique calls for special attention and proper procedure. Each
bead deposited in a hole must be considered as a part of the parent
metal, and the next bead must unite with it thoroughly, as though
it were being laid on the parent metal. If this is not done correctly,
the weld metal will be only a plug which is easy to knock out. For
good fusion, the electrode should be held at an angle to the inside
Electrod~
185 205 'Z25 245 Cm(opprox.)
100
vs.
75

Workpiece dio. !>o

vs.
welding spe~d ( b I Both sides filling
25
Fig. 11.20. Procedure for filling-in a hole: (a) filling from one side,
and (b) filling from both sides.
100
surface of the hole being filled-in, as shown in Fig. 11.20. When
o 0' , I , r I I --1 I , o the diameter 'd' of the hole is smaller than its depth 'h', the hole
2 10 18 26 34 42 50 58 64 1'Z
should be reamed as in Fig. 1l.20(a) or a washer should be inserted
Diomelu of workpiece ( in )
as in Fig. 1l.20(b), and the hole should be filled from both sides.
Fig. 11.19. Relationship between the diameter of the workpiece and the wE:'lding If this precaution is not taken, the electrode, when inclined at the
current and speed for resurfacing work by submerged arc welding process. desired angle, will bear upon the top edge ofthe hole [Fig.ll.20(b)].
may be readily determined by using the graph shown in Fig. 11.19. 11.2.5.Maintenance Welding of Earth Moving Machi-
The dashed curve gives values 'of diameter versus current. The nery and Mining Equipment I
solid-line curve gives values of diameter versus welding speed. The maintenance of earth moving machinery, mining
It is the consensus of those using this resurfacing method for equipment and similar other articles mainly involve repeated
rebuilding cylindrical objects that deep (direct current, electrode resurfacing with abrasion-resistant overlays of those parts which
positive) welding power gives the best results. The magnetic effect are eroded by repeated rubbing action of soil or mining products.
apparently tends to prevent the puddle from running out, Unlike corrosion-resistant coatings, abrasion-resistant overlays
especially when building up the flange of a steel tyre. In SAW, do not always have to cover the entire work area of a part. That
electrode positive polarity, although giving a slightly lower saves expensive filler metals and extra welding labour and time
deposition rate than electrode negative polarity, is preferred to deposit the overlay.
because of its better arc stability characteristic.
 Welding Engineering & Technology Repair and Maintenance Welding 803
802

The weld bead pattern employed for depositing overlay can

have a marked effect on the service life of the component
depending upon the extent of dilution. The effect of different Weld beads
electrode or wire manipulation techniques on dilution of the filler
metal by the base metal is illustrated in Fig. 11.21. It is best to
Weld beads

x
Fig. 11.23. Weld bead pattern recommended for dipper
or dozer teeth used for dirt, clay or sand.

+
Maximum
base-metal
with stringer
x ,
Burn-in
, ,
Weave Least
with some with less base metal
dilution dilution dilution
11.23. The dirt will fill cavities between the weld beads, giving the
base metal an added protection. Another effective pattern is the
waffle or cross-hatch pattern, shown in Fig. 11.24, used for

beads
Fig. 11.21. Some of the typical electrode weaving patterns We~ bead
used for depositing overlays. patterns

hardface shovel teeth and other products subjected to extreme

wear when they are new before putting them into service.
It is recommended that dipper or dozer teeth working
primarily in rock should be hardfaced with beads running the
length of the tooth as shown in Fig. 11.22. This allows the rock to
ride on the hard metal beads. When working with dirt, clay, or
sand run the beads across the flow of material, as shown in Fig.
Weld
beads
Fig. 11.24. Suggested weld bead pattern for dipper or dozer
teeth to be used both for rock and/or dirt.
combination rock or dirt. Apply hardfacing to the top and sides of
the tooth 50 mm up from the cutting point. Extra-hard tungsten-
carbide-bearing weld metal applied as stringer beads on the
cutting edge of the tooth will prolong service life even more.
Shovel teeth can be repointed using austenitic manganese
bearing bars welded to the carhon steel teeth with austenitic Type

Mn-steel
bar
Fig. 11.22. Weld bead pattern recommended for dipper or
dozer teeth used fqr rocks.
Fig. 11.25. Steps involved in repainting a shovel teeth using manganese-steel bar.
 Welding Engineering & Technology Repair and Maintenance Welding 80S
804

308 stainless steel SMAWelectrodes or GMAWwelding wire. The tooth Before use
bar can then be hardfaced with an appropriate wear-resistant
alloy. The three steps involved in the procedure are shown in Fig.
11.25. Shovel~. ~lCCCa:a:lCCCCZ:ZZ~-
The best procedure for surfacing shovel teeth is to make them
Excessive ..•
self-sharpening. Self-sharpening occurs when the unprotected wear After use
bottom wears faster than the top or sides. Mter initial wear on ~~/
the bottom, hard weld metal on the sides, Fig. 11.26, will retard ~.•.
.•.
.•.

Shovel bucket
tooth
Self - sharpening
by use
Fig, 11.26. Schematic representation of self-sharpening
effect on shovel bucket teeth.
continued wear of the tooth and will result in a self-sharpening
bucket tooth. The top of the teeth should always be protected by
using appropriate weld bead pattern depending upon whether it
is to be used for dirt, clay, and sand or hard rocks (as described
earlier).
It is not recommended to hardface both the top and bottom of
the shovel teeth. Mter the hard metal is worn away on the cutting
edge, the tooth will cavitate as shown in Fig. 11.27 and the
hardfacing will chip off, producing a dull tooth.
Shovel
tooth Before
After
Blunt
'dg.~
Fig. 11.27, Effect of hardfacing both the top and bottom
of shovel tooth is a blunt edge.
Hardfacing just the bottom'of a shovel tooth does not produce
optimal life either. The hardfacing overlay will chip off through
Fig. 11.28. The result of hard facing only the bottom of shovel
tooth is excessive wear on the top.
lack of support as the unprotected top surface wears away, Fig.
11.28.
Harringbone, waffle, and stringer bead patterns are used on
shovel and dragline buckets to fit wear conditions to bucket
contours. These are the same patterns for hardfacing dipper teeth.
The waffle pattern works best in wet sand and dirt as it does with
bucket teeth.
A weld deposit of dot pattern works best on hard-to-face base
metals that should not be overheated, such as austenitic high-Mn
abrasion-resistant steels. Spacing of the dots should be varied by
trial and error to determine the best dot pattern.
Hardfaced buckets working in abrasive slag (as in steel plants)
and rock should be protected with beads running parallel to th~
flow of the abrasive material to let it ride on the hard weld-metal
overlay and not on the base metal of the bucket.
In general the maintenance problems involving surfacing and
hardfacing can be divided broadly into 5 classes based on service
conditions, viz.,
Class I severe impact,
Class II
use very severe abrasion,
Class III corrosion, often with abrasion, and frequently
at high temperature,
Class IV - severe abrasion with moderate impact,
Class V - abrasion with moderate to heavy impact.
use Some typical parts commonly hardfaced are listed as per the
above service classes are as follows.
Class I : In this class of severe impact conditions may be
included,
(i) Crusher hammers,
(ii) Crusher jaws,
(iii) Elevator bucket lips,
806 Welding Engineering ~ Technology Repair and Maintenance Welding 807

(iv) Railroad frogs,

-",
:1:- - -;»8~CoF (v) Shovel bucket fronts,
%04
::W<fl
D::>O
Cl:wou
a:J:lICU
:r:v
uJ:-'~
~~g5.;;
U-wj
~
~z~'"
u~wj
5w~'d
~a:
0'"0:
.-1u
~:n~~
:5°w
c> •••
::>0:
a:a:
~V)-'
'i::l'ti~
li!<~i.j
000.:
II:
~
II:
.......
••.•
a:
....
'"
~
'"III
..
lrol
~~/Xl~ ~U1<fl'"
....
•.
0
0:
~
"
~-,w~,
<fl~
;j
8010'"
.. ..., ~S ~ :.a-'
III ~
d•...
-Wa:-<
Z'"
"''''v>
°J:w~
I&;o
:r:
~
.~
11:-'0",]1
....
's.
,5
....
',p
u - •••.
~>u...'" <Il
Q)
::>\1.1
~i3x:u
~<flX:;;:
Va:
/Xl:;::
~ d.•..
;:
'"
::l 8t:'
Q)
en
0Co
e<I
t:'
.~
a:i
Q)~
t:
'"
;::;
CJ'J
.s~ (vi) Crusher rolls,
(vii) Steel mill wobblers,
..•-
... ",
... '" (viii) Shovel bucket teeth,
~o •..<
~<<fl:d (ix) Dragline buckets.
<01.:>",.,
I&;IXO~ ••• Fig. 11.29 shows some of these parts; the hatch marked areas
li!~IX<~
«u...~ ••
XIX ~~
indicate the zones where the overlay weld metal is applied.
%'"
Class II : Typical components which are used with service
condition with very severe abrasion may include,
(i) Auger bits,
(ii) Bulldozer blades,
(iii) Digging arms,
o.. (iv) Dragline teeth,
(u) Drill collars,
11:
<fl<
~IX~;;;
3o...•
", (vi) Feeder blades,
j
~<•...
"'>U1U (vii) Kiln scoop showers,
iwllCi.j (viii) Pug mill auger and knives,
-<-'VU
:r:w::>;;:
/XlCl:
•••
(ix) Shredder knives,
...•
(x) Post hole auger teeth .
Fig. 11.30 shows these components with the hatch marked
sections to indicate overlayed zones.
~ Class III : Examples of components subjected to corrosion,
'"
<flo often with abrasion and frequently at high temperature include,
;3:-
U<." (i) Arbors,
z_",
vcc~
~wu (ii) Roaster rabble blades,
ll!:I: ••.• Wi) Kiln feed screw bearings,
-<<flu
:r:::>;:
a: a: (iv) Blister bar tong bits,
v:l( (v) Conveyer screws,
(vi) Drag chain feeder blades,
(vii) Exhaust valves,
~ (viii) Flapper valves,
o'"- (ix) Refractory dies,

~1X<fl%",
-", (x) Slag ladles.
3wlXl:;< Sketches of these parts are shown in Fig. 11.31.
::i:l:U1g
'"<fl::E•
d,. Class IV : Components subjected to severe abrasion with
i::>::E;j\l.l
%a:<I&;~
moderate impact include,
V:l:_~ \1.1 (i) Arc doors for ore chutes,
11\
(ii) Asphalt mixer paddles,
(iii) Clampshell bucket lips,
 I~

L ~_
~.
-.-. ........ DIGGING
~t'l
KILN CLASSARMS
HAADFACING
SERVICE: HARDFACING
SHREDDER
KNIVES
SHOVELS SCOOP
HARDFACING
BUllDOZER
BLADES
HAROFACING II OR IV
SERVICE
:T I HAROFACING
=~ ::l ~
.0-
c~
~ ,:
0S'
IJ(l
5'
::lt!.j
.....
.. '""""",,
-l.,. SERVICE:
DRAG
COLLARS SERVIC
DRILL
HAROFACING
AUGER
fI:o
SERVICE:
CLASS
E:ClASS
LINE
KNIVES
PUG
TEETH
HAROFACING
SERVICE
SERVICE: MILL
CLASS
TEETH
CLASS CLASS
11
II11
IIIIAND
. IV
HAAOFACING
5'
IJ(l
IJ(l

:; -

~
~
"Q
~
:;.
III
::l
a.
3:
~
5'
;-
::l
~
HARD FACING
::l
t'l
HARDFACING
CONVEYOR ~
ROASTER
HARDFACING RABBLE
HAROFACING
BEARINGS
HARDFACING
BUSTER BAR
SCREWS :;
~
ARBORS BLADES SERVICE :CLASS!II
SERVICE: CLASS II
(KILN FEED SCREW) TONG BITS c:
SERVICE: CLASS III SERVICE CLASS III SERViCE: CLASS III (l.c.FOR SERVICE OVER
TEMPERATURES 425 C)
fNVOlVINCf S'
OR IV IJ(l

HARD FACING
HARDFACING
DRAG CHAIN HARD FACING
HAROFACING

FEEDER BLADES REFRACTORY SLAG

FLAPPER LADLES
SERVICE ClASS III OR IV DIES
VALVES SERVICE: CLASS III
SERVIC E: CLASS III
SERVICE: CLASS III

Fig. lL31. Some typical machinery parts subjected to corrosion, often with abrasion, and frequently 00
~
at high temperatures, i.e., Class III service conditions. \C
 Repair and Maintenance Welding 811
810 Welding Engineering & Technology

> (iv) Coal recovery augers,

(l.
~ V'l",
;;; (v) Drag chain links,
~ ':>::>a::~ (vi) Dredge spud points,
3E(l.Wu
.:>z '" ~w:::w (vii) Gyratory crusher mantles,
u~< ~
X V'l •.•• OC)w!::!
~O(l.>
~as
(viii) Mixer blades,
:u:o:~ xW~15 Q,
~C) ~~
-e«-'> o
a::_", ,5 (ix) Muller tires,
xa:: '"
o '"••• ~•...
(x) Mill hammers,
(xi) Dredge pump impellers.
1
Cl.)

?:
.d Fig. 11.32 shows the sketches of the parts with hatch-marked
>- >
....
'r; zones to indicate the sections to be surfaced/hardfaced .
a:: -
Cj
.:>w
z>.n'"'" o
'fii
Class V : Components with service conditions involving
uoa::~ as abrasion with moderate to heavy impact include,
-euwu •...
.cas
"'wC)'"
li!a::::>!:! d0
'"' (i) Baffle plates,
~~«;:;; •..•
Cl.)

(ii) Bulldozer trunions,

ou
Cl.) ._

1lI
> ."t::

(iii) Gear teeth,

Cl.)",
'" Cj
.s 3 (iv) Picks,
V'l?: ] ,~
W'"
\:a'a:'"
<.>
.~
~
OJ
(v) Shovel boom heels, ,.
'Z_~J:J11'J (vi) Wheels for mine cars, skips, cranes, etc.,
~;~ ~ t::
> ~w~ J!l '" (vii) Ladle pins,
-':r~
Uo-'--c
"'-';;
-::> ••.• -0
:!..Jcc
;
_
gj
(viii) Clutch lugs,
;r;w-'•..•.U
OX •..• ~'" ~ 'Cl.) '"
, (ix) Ditcher driver sockets,
i:!.nwUi ,5· ..• (x) Shovel rollers.
i~:o:!:! -5
~~g5
UID",
as
8 Fig. 11.33 shows the sketches for these parts; hatch marked
ca
<.>
zones indicate the sections to be surfaced or hardfaced~
." '~ Some of the components listed in above service classes are
~i'!: e common to two or more of them and the same is mentioned in the
V'l>
':»Z'"•.•.'" en0 sketches .
W;;; zo:«-e ,
u~...IV)
~-,O-e ...•
u«O
uO~d
~< w:3"" 11.2.6•.Repair Welding of Large Structures
::x:"\:u ~o:o:!::!
c>L&J>
•..•
. A successful repair of a large structure (i.e. ball mill and the
~(l.a:ki
-eV'la::!::! :: C);;; 15 ,~
tire assembly) depends on how closely the analysis and
x«w~
>< ••.•
::> '" ti:.
~'" 0:
U performance stages of the repair are linked together by the
welding engineering development efforts. Certain rules that can
be applied in most cases for repair welding oflarge steel structures
> include the following: .
'"
wi'!: o'"
.:>V'l •..•.
~
.:> .,,>
1. Repair of a large structure is typically of an urgent and
uOx ...•
3Ea::::> .•
~wz;
UC)-c critical nature since the failure of the structure may
Souu :0°.... have a devastating effect on the industrial or financial
a:CwUi :cO:o "
.• ua::!::!
xa::oi'C
eW(l.U
xo::>~
(l.>
activity, jeopardize human safety, and have serious
«:5:l¢
Ill'" economic impact.
u. :l¢
F'"'~-'~:-'

812 Welding Engineering & Technology Repair and Maintenance Welding 813

2. In most cases, the only alternative to repair is the re-

> > placement of a substantial portion of the structure or
" -,,,,
.n",

!-'W~ ~-,If)~
~ even the entire structure. This is usually associated with
UWWu uwffi~ considerable replacement cost, prolonged schedule dis-
i::>ri.;
oo~u
~>....J'"
"'o-,u ruption and, therefore, considerable sales losses.
cero>
~.no~ i~ocr:~
x 0 '&,1 . ....,:1 3. Typically, repair is performed in the field under un-
co'" '" .....,
'"
favourable conditions in a very compressed time frame

> w > ..c:a

i ~
which demands much subjective on-the-spot human in-
tervention compared to fabrication of a new structure.
." > If)~ ...., 4. Many large structures can only be repaired once without
~ ~ ~a:~:5 .~
'-'OWu facing risk of significant damage to the structure.
u \J')(j •• x: .. :::
••.a:u ••• o .
~Gi.J li!W o~
._ en
11.2.6.1. Scope of Repair
••ro::> ~ •.... §
-
Q _U
~a...;: xua.a:
x ce ••• ~~
!:=V')~ A typical scope of repair may be divided into three basic stages
III o o :::
....,0 vi3, analysis, development, and performance.
"0 '"

~'" .-
~ Analysis stage may include the determination of the possible
~ :B~ cause of failure and the assessment of a stress situation in the
:l en
rn ",-
area to be repaired.
> .n> .;!l>
r-", •••. en
en The development stage may include an evaluation of structural
U I- 1/)",
--
,,<.:)'" «l

0..":::
=::J~
= W ~~ u-' .... >,U design and the suitability of that design for possible welding repair
UWlUU
::x: ;j •...
'"
~ •....~w
o a: 5'= auu
~- >
.-::: "-' and development of repair approach, main requirements and
~ <~> x::Ja:
-,'"
..<::
'"
E
«l
0.. detailed specification for assembling, welding, and inspection.
:X:UJa:~
u'"
«l

<.:)-11I E .-
- >,
> Performance stage indudes all repair activities at a shop
. -
«l
'" «l
'" and/or the field.
~..<::
....,
'" 11.2.6.2. Repair Approach
a
o
en The main objective of welding repair is to extend the service
Cl:V'):; M
M
life of a failed structure by using welding. Actually there are two
"Wz", ,..; ways of meet.ing this objective, vie,
a:::::~d -> .
.....

ifOzi.J
O-'Zv .n"'''' .!:P (i) Research oriented approach,
~-'::J> ~z~~
--0-, ~
X::Jo::ffi ua."u (ii) Pragmatic approach.
co •..• '"
oif UJ
-J
...Jlaj
,",U In the Research Oriented Approach major emphasis is on the
~ 0:5>
:z: ct V1f:j
-'_Ill study of the probable cause(s) of failure turning it into extensive
failure and metallurgical analyses, and the determination of
stresses in failed area can become a complete and extended stress
~>
analysis investigation. Often such a data is rarely utilised in
actual repairs though in many cases the analysis is needed to
~<~
U -J ...•• > provide support for the repair effort. Considering the high cost of
•• ~u
~ww
ce-,u ~ V") !f'(3~ such an analysis the main objective should be to generate
_ -' ;,i •.•••
xu.ce
••u.> sufficient input data to support development of a comprehensive
<~ ~w z'"d
iLw ~~..
W"
co Q
ce r •••.•
~ and realistic repair approach and welding procedure.
~ ~ ?:~~ ::E ••• The Pragmatic Approach in contrast to Research Oriented
'"
Approach, is more typical in fabrication and repair shops. The
814 Welding Metallurgy and Joint Design
main emphasis here is on the performance aspects of repair, while
the analysis stage is neglected.
Unfortunately in both these approaches the development stage
and thus the welding engineering aspect of repair is not usually
given adequate attention. A professional welding engineer is an
important link between a research organisation, which lacks
repair performance expertise, and a fabricator relying only on the
collective experience of the shop personnel. In fact, industrial.
practice shows that the probability of successful repair depends
mainly on how closely the analysis and performance stages of
repair are linked together by the welding engineering develop-
ment efforts.
The approach best suited for the most efficient and successful
welding repair of large structure is thus, a balanced approach in
which
1. Scope and nature of the analysis are determined by the
necessity of welding repair.
2. The professional welding engineer is in the centre of the
developmental effort.
3. Performance of repair is carried out under the welding
engineer's supervision and strictly according to a
detailed welding procedure with little room for im-
proviziltion by shop personnel.
The extent of data collection, and its analysis as well as the
procedure actu~lly adopted in a given instance will obviously vary
from case to case.
TWELVE
Weld Joints, Weld Symbols
and Joint Design Principles
In a welded fabrication welds are used to join different parts.
Such junctions of parts are called weld joints and are defined as
the locations where two or more members are to be jointed. The
loads in a welded structure are transferred from one member to
another through welds placed in such joints. Parts being welded
to produce a joint may be in the form of rolled plate, sh~et, shapes,
pipe, or they may be castings, forgings, or billets . It is the
placement of these memhers that helps in defining the different
types of joints.
12.1. TYPES OF JOINTS
There are five basic types ofjoints, Fig. 12.1, for bringing two
members together for welding. These are:
1. Butt joint, 2. T-joint,
3. Lap joint, 4. Corner joint,
5. Edge joint.
1. Butt Joint: It is a joint between two members lying
approximately in the same plane.
2. T-Joint : It is a joint between two members located
approximately at right angles to each other in the form of aT.
3. Lap Joint: It is a joint between two overlapping members.
4. Corner Joint: It is a joint between two members located
approximately' at right angles to each other in the form of an angle.
5. Edge Joint: It IS a joint between the edges of two or more
parallel or mainly parallel members.
When more than two members are brought'together other
designs may b~ used or there may be a combination of the five
basic weld joints: The most popular such multi-member weld joint
is the cross or cruciform joint, which is defined as a joint between
three members at right angles to each other in the form of a cross.
* A billet is a semitinished steel product of rectangular cross-section varying
between 40 x 40 to 150 x 150 mm2•
816 Welding Metallurgy and Joint Design Weld Joints, Weld Symbols and Joint Design Principles 817 .

Butt Joint
171 1. Butt Joint 135° - 180° ~r--., 13~··r------.J
180~.3/'·· ''-»
'.~

B u tt Joint
--,

Corner Joint

2. T-Joint 5° - 90° .J
T - Joint

Tee joint

~~

Lap joint 3. Lap Joint 0° - 5° Lap Joint

Edge joint
~
Fig. 12.1. Five basic types of weld joints.
4. Corner Joint 30° - 135°

Corner Join t

~~L~
Fig. 12.2. A cruciform or double-Tee weld joint.
It is actually a double-T joint as shown in Fig. 12.2. The joint 5. Edge Joint 0° - 30°
describes the geometry in cross-section of the members to be
welded.
12.1.1.Angular Limits of Different Joints
St.ructural members and components can meet at all possible Edge Joint
angles but there are angular limits to the definition of each joint.
These limits for the five types of weld joints as follows :
 Weld Joints, Weld Symbols and Joint Design Principles 819

Distinction between the different welds is important because

allowable design stresses are often defined on the basis of the weld
type.
12.2.1. Fillet Weld
This is the most commonly used weld. type and is so named
because of its cross-sectional shape. The fillet is regarded as being
on the joint and is defined as a weld of approximately triangular
cross-section joining two surfaces nearly at right angles to each
other. Details of the fillet weld are shown in Fig. 12.4.

(i) Groove weld

Bond line

~
(iii) Fillet weld Fig. 12.4. The fillet weld nomenclature.
Fillet welds are used to join comer, T-and lap joints because i
I
they are economical and require no edge preparation.
~~ H
( vi 1 Surfacing weld
12.2.2. Groove Weld ilil
"'l
I

I v 1 Plug or slot weld

This is the second most popular weld type employed in welded
joints. It is defined as, a weld made in the groove between two
members to be joined. The groove weld is regarded as being in the
\ ,~ ,~

(vi 1 spot or projection weld ( viii 1 Seam weld Groove angle

Fig. 12.3. The eight types of welds.
1. Fillet weld, Bevel
angle
2. Groove weld,
3. Back or backing weld,
4. Flange weld,
5. Plug or slot weld,
6. Spot or projection weld,
7. Seam weld,
8. Surfacing weld. Fig. 12.5. The groove weld nomenclature.
 Welding Engineering & Technology
820
joint. There are seven basic groove weld designs based on the type
of edge preparation, and they can be used as single or double
welds. The details of a groove weld are shown in Fig. 12.5.
12.2.3. Back or Backing Weld
It is a weld made on the back side or root side of a previously
made weld. The root of the original weld is gouged, chipped, or
ground to sound metal before the back or backing weld is made.
This improves the quality of the weld joint by assuring complete
penetration.
12.2.4. Flange Weld
Flange weld for edge or corner is used for joints between light
gauge or sheet metal parts. In both cases, parts must be prepared
for these specific joint details.
12.2.5. Plug and Slot Welds
These welds are used with prepared holes. If the hole is round,
it is considered a plug weld; if it is elongated, it is considered a
slot weld.
Plug weld is a weld made in a circular hole in one member on
to another member underneath it as shown in Fig. 12.3(v). For
structures with plate thickness less than 10 mm the weld may be
puddled in with no specific attention to root fusion in the comers
this happens automatically. For thicker plates slot weld technique
should be used.
Slot weld is a means of providing joint area between two flat
members, for example, a cover plate to a beam flange. It is really
a fillet welded up joint with the slot width (W) to plate thickness
(t) never greater than 3 but the slot should never be less than 25
mm wide. The ends of the slot should be semicircular to allow a
continuous and round weld. The spacing between slots should not
be less than 4W.
Both these welds, plug and slot, are shown by the same
welding symbol.
12.2.6. spot and Projection Welds
Both these types of welds can be made by different welding
processes which change the actual weld. For example, when the
resistance welding process is used, the weld is at the interface of
821
Weld Joints, Weld Symbols and Joint Design Principles
the members being joined. If arc welding, electron beam welding
or laser welding process is used the weld melts through one
member into the second member. Both these welds are shown by
one symbol.
12.2.7. Seam Weld
This weld in cross-section looks similar to a spot weld and its
geometry is also influenced by the welding process, the same way
as in the case of spot and projection welds. There are no prepared
holes in either the spot or seam weld.
12.2.8. Surfacing Weld
This. is a weld that consists of one or more stringer or weave
beads deposited on base metal as an unbroken surface. It is used
to build up surface dimensions, to provide metals of different
properties, or to provide protection of the base metal from hostile
environment. It is not used for joining different parts but is a
bead-on-plate type weld.
To describe or specify a weld joint it is necessary to describe
both the type of joint and the type of weld. Variation in each type
ofjoint is also caused by the variation in edge preparation.
This leads to considerable variation in each type of weld joint.
Some of these variants are described in the following sections.
12.3. VARIANTS OF BUTT JOINTS
The butt joint, as defined earlier, is used to join the ends 0·1'
edges of two plates or surfaces located approximately in the same
plane with each other. Preparation of the edge varies according to
the thickness of the material and the welding process used. Light
gauge sections require only 90° sheared edges with no spacing
between them.
Materials ranging from 9mm to I3mm thick, that can be
welded only from one side, should be reduced either as a single-V
or a single-V joint. However, thicker plates are prepared from both
sides. The V-shaped type ofjoint is more satisfactory and requires
less filler material than the V-type groove. However, it is generally
more expensive to prepare a V-shape rather than the straight
edged V or bevel. .
Some of the commonly employed variants of butt joint, based
on ~dge preparation are shown in Fig. 12.6
 Weld Joints, Weld Symbols and Joint Design Principles 823
Welding Engineering & Technology
822 ,L --J,._ T • 3- 6 mill
12.4. VARIANTS OF T-JOINT
G-3mm
T-joint are used to weld two plates or sections whose surfaces
I IT:) T '" •••
I I~t=--+-,
1-
are at approximately right angles to each other. The various types
of T-joints are shown in Fig. 12.7.
( i) Close Square Butt I ii I Open Square Butt
60·
60· to 90· T .5 -12 mm
G - 2 mm

O
(i
~

ii) Single Vee Butt

8 T_3-6mm

(iv) Single Vee Butt-with gap

l'--~'~{
Plain
...L1-1.5

T
T ~ 10 mm
mm

~
R
~1.5

Single
T =-5-12mm

bevel
mm

~ ') t
Double
H
1.5-2.5mm ~~1.5mm T>12mm

bevel
}

60·

I ~ 'D ~QT>12mm
T G = 3 mm 3mm
lJT>20mm
-L 1·5-3mm l.5-Jmm H
l-b...f3mm
T>40mm

~
60·
~
60·
~ ,--- ~ }' ~.

J-g roove Double J

" ) Double Vee Butt with no gap I vi) Double with gap.
60· Fig. 12.7. Some popular variants ofT-joint edge preparation.

T> 12 mm
The included angle of bevel used for T-joints is approximately
G - 3 mm
T > 15 mm
one-half the included angle used for butt joints. Plates or surfaces
should have good fit-up in order to ensure uniform penetration
so· .
and fusion. Edge preparation of the vertical member is the saine
as that of the butt joint, as mentioned earlier, in regard to

r-;
( viii) Butt Zed
20· thickness and bevel.
I vii) Smm
12.5. VARIANTS OF LAP JOINT
T>40m.
As the name implies, the lap joint is used in joining two over~.
Smm lapping plates so that the edge of each plate is welded t.o the
surface of the other. The overlapping portion is called the lap.
The width of the lap may be 3 to 5 times the thickness, T, of
~ the plates to be welded. Welds are usually run on each side of the
I X) Double - U Butt lap.
20· No edge preparation is required for a lap joint, and the joint
is easy to line up. But a lap joint has the following drawbacks :
T>30mm (i) some metal is wasted on the lap,
(ii) the courses or tiers of cylindrical products to be lap-
welded have to be of different diameters.
Common lap joints as shovv'llin Fig. 12.8 are single lap, double
~~ lap, and off-set or joggled lap joints. The single-welded lap jomt
T>20mm ~ does not develop full strength but it is preferred to a butt joint for .
(xi) Unequal Double -U Butt I XII).. Composite Butt 60· T > 30mm

Fig. 12.6. Some variants of butt joint based on edge preparation.

824 Welding Engineering & Technology Weld Joints, Weld Symbols and"Joint Design Prin. :ples 825

. jI3-SlT-l

)--TLD\ Single-welded lap joint

I 1 ~-\ +
::::::::, tr".
Double-welded lap joint
--

T ~ 10 mm

\ T LJ_\_~ !
( i) Vee - Bevelled ( ii I U - grooved
-
Joggled or Offset lap joint
Fig, 12.10. Edge preparations for closed corner joint in heavy plates.
Fig. 12.8. Common types of welded lap joints.
some applications. A good example of this is in tubular construc-
tion where one tube telescopes into the other and is lap-welded. (A) (orner - to-corner
In essential weldments lap joints must be welded from both
sides. The lap joint should not be used on plate thickness above
10 mm. Also, double-welded lap joints should have a minimum
overlap of about 5 times the base metal thickness to limit joint
rotation under load. (B) Partially lapped

12.6. VARIANTS OF CORNER JOINTS (B)

A corner joint is used to join the edges of two sheets or plates
whose surfaces are at an angle approximately 90° to each other.
It is commonly used in the construction of boxes, tanks, frames, (() Partially lapped with both
machine tool beds, and other similar items. Welding can be done inside and outside fillets
on one or both sides, depending on position and type of corner joint
used. Shown in Fig. 12.9 are the more common corner joints.

I D) Bevel edge preparation

for thick sectiortS

Open-type corner
(Iosed-type corner
Open-type corner welded from both sides IE) J-Edge preparation for
thick sections
Fig. 12.9. Three types of commonly used corner joints.
The closed corner joint is used to join lighter sheets where
strength required is not too high. Little or no filler metal is needed,
and edges are just melted together. If it becomes necessary to I ~I Bevel ed'Je preparation plus
make a closed corner joint with heavier plates the lapped portion H\Slde fillet-for thick sections
of the plates is V-bevelled or V-grooved to permit penetration to
the root of the joint, as shown in Fig. 12.10.
Fig. 12.11. Six popular variants of welded corner joints.
826 Welding Engineering & Technology 827
Weld Joints, Weld Symbols and Joint Design Principles
The open corner joint is the type used for heavier plates. The The inside fillet weld can be added to corner joints of designs
two edges of the plates are melted down, and sufficient filler metal D and E as well. Design the combined weld size with reference to
h'l added to round out the corner. The joint may be welded on the the thickness of the thinner member because the corner will not
outside and then reinforced by adding a bead on the inside. be stronger than that member. Also, that will minimise the weld
Although above mentioned three types of corner joints are metal requirement to achieve lowest cost joint.
popularly employed but in a more detailed evaluation of such a 12.7. VARIANTSOF EDGE JOINT
joint six possible configurations, shown in Fig. 12.11 are The edge joint consists ofjoining two parallel plates by means
considered. The salient features of these configurations are as of a weld. This type of joint, shown in Fig. 12.12, is often used in
follows. 1T
1. Corner-to-corner Joint (A) --L
This design, Fig. 12.11(a),
position and usually requires
is seldom best. It is difficult to
fixturing. The root pass must be
I -':J2 , I
Sheet metal preparation for edge butt joint
made with small electrodes at lov.•..current to avoid unacceptable
melt-through, unless backing can be provided. It uses a large
amount of weld metal but has a finished contour.
2. Partially Lapped Corner Joint (B)
This design, Fig. 12.11(b), is easy to assemble. It uses half the
weld metal of joint(A),and it needs no backing. It has, however,
lower strength because its effective throat is smaller. Its
appearance looks finished, perhaps m01:'eso than joint A.
3. Partially Lapped Corner Joint With Both Inside and Edge joint for
thin sheets
U Edge joint for
heavy plates
Outside Fillet (C) Fig. 12.12. Three common types of edge joints.
This design, Fig. 12.11(c), is as easy to assemble as the joint sheet metal work. The two edges can be easily and quickly melted
B. It has the same effective throat as joint A and so the same down, eliminating the need for any filler metal. Other variants
strength. It requires half the weld metal of joint A i.e. the same are also shown in the figure. In heavy plates, where bevelling the
as required for joint B, but it locates the welds more efficiently. It edges is done to get deeper penetration, some filler metal is
decidely looks an unfinished weld joint. needed.
4. Corner Joint for Thick Sections (D) No special preparation is necessary,for thin sheets except to
This design, often specified, uses a partial joint-penetration clean the edges, bend them up on a press brake, and tack weld
single-V groove weld, Fig. 12.11(d). The joint preparation of both them into position for welding. Heavy plates require bevelled
members is required. edges in order to secure good depth of penetration and fusion in
5. Alternative Corner Joint for Thick Sections(E) the side welds.
This corner joint, Fig. 12.11(e), is prepared with a J-groove or 12.8. VARIANTSOF WELD TYPES
even a V-groove. In the figure J-groove is shown prepared in the
outside plate, because .that will significantly reduce shrinkage
stresses in that plate. If the J-groove is placed in the inside plate
it can result in through-thickness shrinkage which often results
in lamellar tearing.
6. Another Corner Joint for Thick Sections (F)
This design with its added fillet weld on the inside corner,
makes a neat and economical corner joint, Fig. 12.11(f). The bevel
preparation, however, should preferably be made in the outside
Mitre Convex
plate to reduce harmful through thickness shrinkage. Concave

Fig. 12.13. Three variants of fillet welds.

 Weld Joints, Weld Symbols and Joint Design Principles 829
828 Welding Engineering & Technology
(i)convex,
Most of the welds made by arc welding are either groove welds (ii) concave,
or fillet welds. The grooye welds are almost always with convex (iii) mitre or straight line.
contour of reinforcement, however the fillet welds as shown in Fig.
12. .13, may have the following thI;ee variants for their reinforce- Fillet weld size is measured by the length of the legs of the
ment: largest right angled triangle that may be inscribed within the fillet
Actual throat
weld cross-section as shown in Fig. 12.14. The effective throat, a
better indication of weld shear strength, is the shortest distance
between the root of the weld and the diagrammatical weld face.
With a concave weld face, the actual throat may be larger than
the effective throat of the weld, while for the mitre weld face there
is no difference between the two sizes. In any case, the strength
of a fillet weld is based on the effective throat and the length of
the weld i.e. on the effective area of the weld.
Apart from these variations in fillet welds such welds are also
often used in conjunction with groove welds to form combined
welds.
(A) Convex tillet weld
12.9. COMBINED WELDS
Combined partial-joint-penetration groove and fillet welds, as
shown in Fig. 12.15, are useful for many joints. Occasionally, this
Actual throat
and effective
throat

( i) Bevel edge preparation ( it) J - edge preparation

Theoretical throat
( B) Concave fillet weld
Fig. 12.15. Typical combined groove- and fillet-welded joint.
practice is questioned because of differences in recommended
welding electrodes for the two types of welds. The practice,
however, is sound.
Actual The desired effective throat ofcombined groove and fillet welds
throat
can be obtained by adjustment of the groove dimensions and the
fillet weld leg lengths. However, consideration must be given to
the accessibility of the root of the joint for welding, and to stress
concentrations at the toes of the fillet weld. When a partial-joint-
penetration groove weld is reinforced with a fillet weld the
minimum effective throat is used for the design purposes. The
effective throat of the combined welds is not the sum of the
effective throats of each weld. The combination is treated as a
Theoretical throat
single weld when determining the effective throat.
Fig. 12.14. Weld size for convex, concave and mitre fillet welds.
830 Welding Engineering & Technology Weld Joints, Weld Symbols and Joint Design Principles 831

12.10. SELECTION OF WELD TYPE 20

Frequently the designer is faced with the question of whether
to use fillet or groove welds. Often cost is a major consideration
in the final selection.
.•...
Double-fillet welds, Fig. 12,16(a) are easy to apply and require 0u
<II
a:;
OJ
:E0 1 a
OJ

no special edge preparation. These can be made using large

diameter electrodes with high welding currents for high deposition ...•....
',', ....
,\ ,

rates. In comparison the double-bevel groove weld, Fig. 12.16(b),

~.'" ----=:=-
?'l/'
~~~

o oLL 12.5

lA) I-S--j IBI Ie)

Fig. 12.17. Relative costs of full strength groove and fillet welds in plates.
Fig, 12.16. Comparison of fillet welds (A) and groove welds (B & C),
has about one-half the cross-sectional area of the fillet welds. The combined double-bevel-groove and fillet weld joint, shown
in Fig. 12.18, is theoretically a full strength weld. The plate edge
However, it requires edge preparation and the use of small is bevelled to 60° on both sides to a depth of 30% of the thickness
diameter electrodes to make the root pass. of plate. After the groove on the each side is welded, it is reinforced
In Fig. 12.16(c) is shown a single-bevel-groove weld that
requires about the same amount of weld metal as the double-fillet
weld of Fig. 12.16(a). Thus, it has no apparent economic advantl:!.ge
but has some disadvantages, for example, the single-bevel weld
requires edge preparation and low deposition root pass. From a
designer's point of view, however, it does offer direct transfer of
force through the joint, which means that it is probably better than
fillet welds under cyclic loading.
Double fillet welds having a leg size equal to 75% of the plate
thickness would be sufficient for full strength. However, some
codes have lower allowable stress limits than other codes for fillet
welds and may require a leg size equal to the plate thickness. The I -I- .\
cost of a double fillet welded joint may exceed the cost of a Fig. J 2, .W. Combined groove and fillet welds with partial joint
single-bevel-groove weld in thick plates. Also, if the joint can be penetration but capable of full strength.
positioned so that the weld can be made in the flat or downhand with a fillet weld of equal area and shape. The total effective throat
position, a single-bevel-groove weld would be less expensive than of welds is equal to the plate thickness. This partial-joint-
a double fillet weld.
penetration weld has only about 60% of the weld metal of that in
The development of curves based on the best determination of a full-strength double fillet weld. It does require edge preparation,
the actual cost of joint preparation, positioning, and welding such but wide root face permits the use oflarge diameter electrodes and
as sho'Wn in Fig. 12.17, is a technique for determining the plate high welding ·currents. It is recommended for submerged arc
thickness where a double-bevel-groove weld becomes less costly. welding to achieve deep joint penetration.
The intersection of the fillet weld curve with the groove weld
curve is the point of interest. The validity of the information is Full strength welds are not always required in a weldment
dependent on the accuracy of the cost data used in cor..structing and cost can often be reduced by using smaller welds where
the curves. applicable and J}ermissible. With equal effective throats a fillet
832 Welding Engineering & Technology 833
Weld Joints, Weld Symbols and Joint Design Principles
weld shown in Fig. 12.19(a) requires twice the weld metal than Applicable welds
that needed for a 45° partial-joint-penetration, single-bevel-groove Square-groove J~groove
V-groove F~re-V-groove
Bevel-groove Flare-bevel-groove
U-groove Edge-flange
Braze
10mm
(AI Butt joint
Applicable welds
Fillet Flare-bevel-groove
Square-groove Edge-flange
V-groove Corner-flange
Bevel-groove Spot
U-groove Projection
J-groove Seam
F~re -¥-groove Braze
Weld area = 100 mm2 Weld area = 50 mm2
IA) (B) (B) (orner joint
Applicable welds
Fillet
I ]-groove
Flare- bevel-groove
Slof
Plug Spot
Square-groove Projection
Seam
A Bevel-groove
Braze

(() T-joint
Applicable welds
2
Weld area = 50 mm2 Weld area = 57.735 mm Fillet J-groove
(( ) (D) Plug
Slot
F~re- bevel-groove
Spot
Bevel-groove Projection
Seam
Fig. 12.19. Comparison of different welds with equal effective throat. Braze
weld, as shown in Fig. 12.19(b). The latter weld may not be as
(0) Lap joint
economical as a fillet weld, however, because of the cost of edge
Applicable welds
preparation. Also, some welding codes limit the effective throat of
Square-groove Edge-flange
this type of weld to less than the depth of the bevel with certain Bevel-groove (orner-flange
welding processes because of incomplete root penetration. V-groove Seam
U-groove Edge
J-groove
If a single-bevel-groove weld is combined with a 45° fillet weld,
Fig. 12.19(c), the cross-sectional area of the same effective throat
is also about 50% of the area of the fillet weld in Fig. 12.19(a). fEIEdgejoint
Here the bevel depth is smaller than it is with the single-
bevel-groove weld in Fig. 12.19(b).A similar weld with a 60° groove Fig. 12.20. Different types of weld joints dnd the corresponding applicable welds.
angle and an unequal leg fillet, but with the same effective throat, 12.11.WELD JOINTS FOR STRUCTURAL TUBULAR
Fig. 12.19(d), also requires less weld metal than fillet weld alone. CONNECTIONS
This joint allows the use of higher welding currents and larger Apart from welded joints in plates and sections, tubular
diameter electrodes to obtain deep root penetration. members are widely used for structures fabricated by welding. For
Different types of joints and the corresponding applicable example, tubular members are being used in structures such as
welds for each one of them are shown in Fig. 12.20. drill rigs, space frames,trusses, booms, off-shore platforms, and
earth moving and mining equipment. They have the advantage
834 Welding Engineering & Technology 835
Weld Joints, Weld Symbols and Joint Design Principles
Heel
strength and stiffness. However, if a connection is to be made with
a complete-joint-penetration groove weld, the weld usually must
be made from one side only and without backing because the size
or configuration, or both, will prevent access to root side of the
~~x
weld. Special skill is required for making tubular connection using
complete-joint-penetration welds from one side.
Typical details for branch connections for circular structural
hollow sections for butt welds (wall thickness::;; 30 mm) and fillet
corneJ
( Bl Box sections

x
, Toe zone"
Maximum limit of 90· T ..Y

-
: =...1- .•
T.conflgu3rtlon
" 10•
10
Side zone Ii
I ' ,
o __ L:L_~ _
't'I

(Cll-connection (0) Y- connection

,f,
,'1', ",)
\
'"

, I
H min - T

'f
(E) K-connection <D KIT-K)
(F 1K'combination

Fig. 12.21. Welded tubular connections, their components and nomenclature.

(After AWS-Vol. 5, 1984 Ed.)
'@ KlT-Yl
connections
2~~]d-O
~=r
,---.J(I
1-2.5.~_
2.-3mm

d<D
H

of minimising deflection under load because of their greater
rigidity when compared to standard structural shapes. Various
types of welded tubular connections, their component
designations, and nomenclature are shown in Fig. 12.21.
With structural tubing, there is no need for cutting holes at
intersections and, as a result, the connections can have high
Oetail at y
Fig. 12.22. Details for typical branch connections with butt welds
in circular structural hollow sections.
welds are shown respectively in Figs. 12.22 and 12.23 while for
butt welds in rectangular structural hollow sections these are
shown in Fig. 12.24.
836 Weld Joints, Weld Symbols and Joint Design Principles 837
Welding Engineering & Technology

Saddled branch
x
x

D -~ J....y
-d~. o

Laleg length '-2.5mm

-1-_
2-3 mm
Detail at x e a 60·_ 90·

at Z

Crimped or straight H min" T

cut branch
2-3 mm
e < 60·

/
i
.-1------ --i-_.
Fig. 12.23. Details for typical branch connections with fillet welds in circular
structural hollow sections. (After Hicks).
Note: The angle of intersection 9 of the axes of the circular
hollow sections should not be less than 30° unless adequate
efficiency of the joint has been demonstrated.
With relatively small, thin wall tubes, end of the brace tube
may be partially or fully flattened. The end of the flattened section
is trimmed at the appropriate angle to abut against the main
member where it is to be welded. This design should only be used
with relatively low load conditions because the load is concen-
trated on a narrow area of the main tube member. The flattened
section of the brace member must be free from cracks.
Detail at Y Detail at Z
Fig. 12.24. Details for typical branch connections with groove welds in
rectangular structural hollow sections. (After Hicks).
All type of welds and welded joints are represented by symbols
on the drawing sheets depicting the complete details of the
component or structure to be fabricated. The most common types
of these weld symbols are discussed in the following section.
12.12. WELDING SYMBOLS
Standard symbols are used to indicate desired welding and
brazing information on engineering drawings. These are used to
convey the design requirements to the shop in a concise manner.
For example, a symbol can be used to specify the type of weld,
groove design, welding process, face and root contour, sequence of
welding, length of weld, effective throat, and other informations.
However, when all the required information cannot be conveyed
by a symbol alone, supplementary notes or dimensional details, or
both, are provided to the welder. The design engineer must ensure
 838 Welding Engineering & Technology Weld Joints, Weld Symbols and .Joint Design Principles 839

that all the details and specifications are fully presented on the Reference line
drawing with the help of welding symbols.
12.12.1. Basic Welding Symbols
The terms weld symbol and welding symbol have different
meanings. A weld symbol indicates the required type of weld, and
the basic weld symbols are shown in Fig. 12.25. The welding
symbol includes the weld symbol and supplementary information.
Tail~
------
(Other

(Arrow
side')

side) // / Arrow

I-
r- I I I -.-----
vII I
l 0- a<- IIiII II~r-
E>
c:;
goo·
I

CD~~~
c:;

b0Kq8
1I '".x 'ij. Ia --lI I ------
IwI L- I
r--
I~
----.
I--J
:---
1/'-~!J~---1
----1
L.-L
D
IIIIIIu::~o.*,
I ,--I
J

~
r- c:

~
'" '-
II "0en
c:l ..0
c:

~~ ~~.....2
I u..Cl
Vl
\L
.•...
:;JCl.. a IO:°Vi
:::>
III.E
E

ij0
Vl
'-
c:l
'" '-Evc..:x; I.LJCl
'"
"'" I
Fig. 12.26. Essential elements of a basic welding symbol.
..0
E-<'",,;,
.<:: as shown in Fig. 12.26. Note that the reference line is always
ii:
SC'ol
c-i
.",
.....•
"'W

Z'"
.a (II
'"a.
c::l -0
.~
<.l
in'
..>::i
OJ

~'->-
VI
i:
0- '"vVI'"
c::l
::>
't"""
<- ..c
~
c:
0c
+'
0-
~ ~ '"~--c;
a~
a'"'":gu::
a.
'00'0
..0
(II
",..0
c:
:§
~~
>-
..c.•....
.J!! ~(II
o.ci
<Ii
horizontal on all drawings and that the arrow is always inclined
Qj ~
,~

'Vi
E'0
.•.. .2!
..0
to the reference line at one end. The angle of inclination of the
arrow or whether it is pointing upwards or downwards has no
significance. That means the reference line has two sides viz, the
arrow side and the other side.
The arrow points to a spot on the drawing where the welding
is to be done. Many times it is difficult to put the arrow on the'
same side as the actual spot where the weld is to be made, so the
two sides of the reference line are made use of. Any welding I
I
instructions on the arrow side of the reference line are to be made I,
"

on the same side of the part where the arrow points to. Any
welding instructions on the other side of the reference line apply
Ii
to the opposite side of the part. The arrow side is always under
the reference line, no matter which way the arrow points. TEe
other side is always on top of the reference line. Welds on both I',
sides of a joint are shown by placing weld symbols on both sides
of the reference line.
I
I
,- -----<~- 2 B3Y ./ /" -<f:p
i
I

Fig. 12.27. Welding symbols with references in tails. ii'

I

When a specification, process, test, or other reference is !I

!1

needed to clarify a welding symbol, the reference is placed in a

tail on the welding symbol, as shown in Fig. 12.27. The letters
CJP in tail of the arrow are used to indicate that a complete joint
penetration weld is required. The type of weld or joint preparation
All welding symbols have a minimum of three basic parts : may be optional. The tail may be omitted when no specification,
(i) a reference line, process or other reference is required with a welding symbol.
(ii) an arrow, Including the basic three elements discussed above, a welding
(iii) a feathered tail symbol may consist of the following 8 elements :
1. a reference line, 2. an arrow,
840 Welding Engineering & Technology Weld Joints, Weld Symbols and Joint Design Principles 841

3. a tail, 4. basic weld symbol, Fig. 12.28 shows stand~d locations of elements of a vvelding
5. dimensions and other data, symbol. However, all elements mentioned above need not be used
6. supplementary symbols, on a welding symbol unless required for clarity. "
7. finish symbols, Significance of the elements of a welding symbol are briefly
8. specification, process, or other references. described as follows .
All these elements have standard locations with respect to 12.12.1.1. Weld Symbols'
each other. Fig. 12.25 shows the basic weld symbols used above, below, or
_I c
-
on the reference line. Some weld symbols have no arrow or other
:!! ~'o side significance. However, supplementary symbols used in .
0\0'- "0
.0
:§~o E conjunction with these weld symbols may have such significance .
'0
cli:n'"
>-
••. .0 ..••••• III For example, welding symbols for resistance spot and seam
••• E c0.5 •••.Eo welding have rio side significance, as shown in Fig. 12.29 .
-
~
>-
III
u- •••
•••
'f :g
••• ~~E
!
o~ •..c ••• ~
0 / /' .0
III

••••••'0
w ~ oc( :';'iij
"0 ••• ~"O RS~ <€SEW
~..¥
::JC '"0.,co _ -- "ii
Fig. 12.29. No arrow or other side significance.
u'VI w ~ 0\ lL :8
a>. 12.12.1.2. Weld Dimensions
.5
••
:.;
- III -
~ ~.c
u8 l r1J

•••c"o O.-:::a. Dimensions of a weld are shown on the same side of the
°
o.~w ~Q.~y
C
.Ef
"0

reference line as the weld symbol. The size of the weld is shown
cu~
~"OJ ~o.~ 0.. a;
~
'" to the left of the weld symbol, and the length of the weld is placed
••• 0\
C •••
g.!!o.
0
~ I •...

~
o on the right. If a length is not given, the weld symbol applies to
that portion of the joint between abrupt changes in the direction
.!!! .ft
-

'0 a
<I)
of welding or between specified dimension lines. If a weld symbol
30lS 3 (]IS ~ ~ .r. ....•.. w ..£
is shown on both sides of the reference line, dimensions must be
c)~o_ (~3HJ.o)(,"mHlyl
~ j ~
•..
•..
.-.-
oeL.
C C •••
t!~~ ~ •..
~ 13 g
0
° .... Q)

given for each weld even though both welds are identical.
Examples of dimensioning for typical fillet welds are shown in Fig.
c •••C
./L:J«
••• OIl ~

~(S30IS
"O"Og' _
0::
~~
HJ.OS) Z 11
- g "0
ZOo
Eo~'O'"
••••••
::t
~~.cc~_ 0'"
UJOlllO"''E
~
cs

12.30. /
.0 .0 ._ '"
E E ::
>->--
III III _
"0

en
c:
.s
10 ~ 60 5 f7 100
6 150
~
OII::J
••• 0 OIl
'0 cO
C'I
10 60 "'- 6 [:> 150 5 --,~
.- c
.~ 0 ~"ii
a.s .~ ••• ",;
...•
Dimensions in mm
l&-o~_u Fig. 12.30. Weld size and length on welding symbols.
u 0\.•.£'" ::••• ~
t:Ul

C"O UJ :8
E C
••• Ifa weld in a joint is to be intermittent, the length of the weld
~°
.- C >- ••.
and the space between the welds are placed to the right of the
0.0\
o
-Ci.
::J
•••
.-C'Oc~u -"'-
III •••
"Ow...
weld symbol with the latter in the parenthesis as shown in Fig.
8 •.. t-~~"O ~:g 12.31(a). Alternatively the length of the weld and the pitch
a:.2 .~E ~:lI
~
u •••
'iij '0 (centre-to-centre spacing) are placed to the right of the weld
::J C •...
o c
:: •••
C
'0 IDo symbol as shown in Fig. 12.31(b & c).
~ 'i .!!! The location of the symbol for specifying groove-weld root
opening, groove angle, plug or slot weld filling depth, the number
842 Welding Engineering & Technology
Weld Joints, Weld Symbols and Joint Design Principles 843

A welding symbol to represent a full penetr;ition butt weld by

a welding procedure to be agreed is 2.
12.12.1.3.Supplementary Symbols
Fig. 12.32 shows supplementary symbols that are used on a
,,,elding symbol. They complement the basic symbols and provide

r
I
around
or weld 50(100) a~
additional requirements or instructions to the welder.
jEXt\mine
"'-
..--..
'--"
Concave
weld
B'ucking,
Convex
~
~ r Melt-thru
space
~ Field po}' --~ F!ush Contour
orflat

(0,) Lengt.h of gap between intermittent welds.

Fig. 1'~.32. Supplementary welding symbols.
Su: note
(i) Weld-AU-Around Symbol
A weld that extends completely around a joint is indicated by
the weld-all-around symbol. Fig. 12.33 shows two examples of its
use. The weld can be in more than one plane as is shown in Fig.

i
o

H
Symbol V
weld~ Desired ~Ids
Note: Ifrequired by actual length of the joint, the length of the
increment of the welds at the end of the joint should be
'/)
increased to terminate the weld at the end of the joint.
(6) Length and pitch of increments of chain intermittent welding.

Symbol

C~'l
Desired weld
lalf·i-Beam to plate weld
Symbol

It of weld
Desired we,lds
Note: If required by actual length of the joint, the length of the
increment of the welds at the end of the joint should be
increased to terminate the weld at the end of the joint.
(c) Length and pitch of increments of staggered intermittent welding.
Fig. 12.31. Welding symbols for different types of intermittent welds.
Desi red \01 eld
0fj Symbol

of welds required in a joint, and other dimensions are shown in

Fig. 12.28. I b) WeLd in several planes around periphery
Fig. 12.33. Two examples of the use of weld-ail-around symbol.
844 Welding Metallurgy and Joint Desi~n
Weld Joints, Weld Symbols and Joint Design Principles. 845
12.33(b). This symbol, however, should not be used if more than
one type of weld is required to make the joint.
~ <see
(ii) Field Weld Symbol .
/ note
/' 2::]
.. <see
note
Ji'ield welds are made at the erection site, not in the shop. Each R - Backing removed
of-these welds is designated by a field weld symbol (i.e. a flag) Note; Materials and dimensions

I
a ffer welding
of backing as specified
which' is always placed above the reference line at its junction with (aJSacking symbol

""../1/ "V-~note
--1---{ /See ~
~ Double-bevel-groove
Double- V- groove
Fig. 12.34. Use of field weld symbol. Note; Material
and dimensions of spacer as specified
arrow end pointed away from the arrow towards the tail, as shown (b' Spacer symbol
in Fig. 12.34. i·
Fig. 12.36. Examples of the use of: (a) backing, and (b) spacer symbols.
(iii) Melt-Through Symbol
The melt-through symbol is used to show complete joint removed after welding. The backing type, material and dimensions
should be specified in a note.
penetration (CJP) with root reinforcement on the backside of the
A welding symbol for a typical joint with a spacer strip inserted
V. " . ---U----/" in the root of the joint is shown in Fig. 12.36(b). It is a modified
groove weld symbol having a rectangle withIn it. The material and
I
a Placement of melt tt,ru symbol dimensions of the spacer strip should be specified in a note.

1.
f

(v) Contour Symbol

&
·\~~>~t
M' ....,';?i....
1· / u3 l-(l
3/\
~
Ia'
/'
Contour without finishing
-$J
-
f C
Desired weld
(b) Reinforcement wIth melt-thru
Fig. 12.35. Use of melt-through sYxbbol: (a) method of placement,
(b) height oOnelt-through reinforcement.
Symbol

and
/6 G

G- Grinding
M

"&

C
/
J'

R C
I
M - Machining
C- Chipping
weld to be made from one side only. The reinforcement is shown by
( b) Contour with finiShing
placing the melt-through symbol on the side of the reference line
Fig. 12.37. Use of contour symbols for welds.
opposite the weld symbol, as shown in Fig. 12.35(a). The height of
root reinforcement can be specified to the left ofthe symbol, as shown A contour symbol is used on a welding symbol to indicate the
in Fig. 12.35(b), if the amount of reinforcement is critical. shape of the finished weld. Welds that are to be made approxi-
(iv) Backing and Spacer Symbols mately flat, convex or concave withbut subsequent finishing are
A backing symbol is placed above or below the reference line represented by adding the flush, ~onvex, or concave contour
to indicate that a backing ring, strip, or a consumable insert is to ' symbol to the weld symbol as shown in Fig. 12.37(a). Welds that
be used in making the weld. It must be used in combination with are to be finished by mechanised means are depicted by adding
a groove weld symbol to avoid mix up with a plug or slot weld. A

~_=:\
welding symbol for a typical joint with backing is shown in Fig.
12.36(a). It is a combination of groove weld symbol on one side of
the reference line and a backing symbol on the opposite side. An ~ -,1L"
R may be placed within the backing symbol if the backing is to be
~ -~" I
Fig. 12.38. Weld svmbol construction.
,r--..-,..--
 Weld Joints, Weld Symbols and Joint Design Principles 847
846 Welding Engineering & Technology

I'-l01 E: T he: size: of the 1ill~ weld is not staled

• preparotioo of thi: main joint.
fp
L__ ~ ~_
unles~ 11 ddf~rs from
-----~
~~
H~ot dic\oh:d by the i:dg~
both the appropriate contour symbol and the required finish
symbol to the weld symbol, as shown in Fig. 12.37(b).
12.12.1.4.Construction of Symbols
Bevel-, J-, fillet, corner-Hange, and flare-bevel-f:,'l'ooveweld
symbols are constructed with the perpendicular leg always to the
( i) Compound Yield (fillet superimposed on single-bevel butt) left. When only one member ofajoint is to be prepared for welding,
the arrow is pointed with a definite inclination toward the member
unless the preparation is obvious. The arrow need not be inclined
towards any if either member may be prepared. These features
are illustrated in Fig. 12.38.

( ii) Compound
JJ ~
~~~

weld (fillet superimposed on single-bevel butt)

root Side (see note to Fig'( I) above}
with fillet weld on
When combination of welds is to be specified to make a joint,
the weld symbol for each weld is placed on the welding symbol.
Examples of such symbols are shown in Fig. 12.39.
12.12.1.5.Multiple Reference Lines
Two or more reference lines may be used with a single arrow

(iii) Compound 'Held (fillet

\ 1~
superimrcsed en smgl\!-J buttl filet
(see note to Fig.! \) above)
weld on root Side
to indicate a sequence of operations, as shown in Fig. 12.40.

2nd 3rd

Weld symbol
Operation
Operation2nd 1st -~
3 rd -V-·:../
/ /~
h,"'O'O

CP~ (a) I b) (c)

••
Fig. 12.40. Weld symbols with multiple reference lines.
~~ 12.12.1.6.Processes
( IV) Partially-penetrated single-bevel butt with superimposed fillet lsee not Fig.! i II
TO
Letter designations are used in the tail of a welding symbol to
indicate the appropriate welding process. The more frequently
used welding designations are listed in table 12.1.
[---, S. No.
Table 12.1. Welding Process Designations
Welding Process __" . Designation
1. Shielded metal arc welding SMAW
I v) Compound weld (fillet superimposed on partially penetrated single-bevel buttl
2. Submerged arc welding SAW
(a) Welds to join unequal sections. 3. Gas metal arc welding GMAW
4. Flux-cored arc welding FCAW
5. Gas tungsten arc welding GTAW
6. Plasma arc welding PAW
~ 7. OxyfueJ"gas welding OFW
Symbol 8. Electron beam welding EBW
9. Resistance spot welding RSW
Desired weld h. plate weld
(e) st;.,o 10. Resistance seam welding RSEW
Note: The size of the weld is not stated u :llo>\sit differs from that 11. i Laser beam welding I LBW
dictated by the edge preparation of tile main joint. --'-'--'---.---.-- __ • • ..__ " . ••• '. __ .~ •.• .l ..• • •__ .__ •__ •

Fig. 12.39. S,rmbols for some compound welds.

848 Welding Metallurgy and Joint Design Weld Joints, Weld Symbols and Joint Design Principles 849

12.12.2. Solved Examples is 10 mm and with no root opening. Sketch the desired weld and
Mter designing a joint, a welding symbol can generally be used show the welding symbol if the effective throat is to be 13 mm.
to specify the welding process to be used and the outline of the Solution. The desired weld and its symbol are shown in Fig.
procedure to be adopted. This is explained with the help of the 12.43.
following solved examples.
Example 12.1. An H-section placed perpendicular on a square 0- root opening
plate is to be welded all around using a fillet weld with a leg length
of 8mm by CO2 welding. The weld is to be made at site. Show a '\-60·
s"ketch of the weld and its welding symbol.
Solution. Fig. 12.41(a) shows a view of the desired weld and
its welding symbol is shown in Fig. 12.4Hb).
Squore plote
M
timm
!5 ! ~
I
)"mm
SA~31

C02
welding (a) Desired weld (b) Symbol
Fig. 12.43. Desired partial penetration butt weld and its symbol.
H- Section
Example 12.4. A rod is to be welded to a plate by single
flare-bevel groove weld using gas metal arc welding process. Sketch
(0) DeSired weld I b) Symbol
the weld and show the welding symbol if the diameter of the rod
(a) Desired weld (b) Symbol is 18mm and effective weld throat 5 mm.
Fig. 12.41. Desired weld and its symbol.
Example 12.2. Two plates are to be welded with a single- V Solution. The desired weld and its symbol are shown 'in Fig.
12.44.
outt using SMAW process. A sealing run is also to be laid which
is required to be ground flush. Sketch the desired weld and its
symbol.
Solution. The desired weld and its symbol are shown in Fig.
12.42.

-Plate

(a) Desired weld

r==t=.•~ ! I

(b) Symbol
4"AW
Fig. 12.42. Desired weld and its symbol.
Example 12.3. Two plates 30 mm thick are to be Joined by
partial penetration single- V groove weld using submerged arc
welding process. The groove angle is 60°, the depth of preparation
(a) Desired weld (b) Symbol
Fig. 12.44 Desired single flare-bevel groove weld and its symbol.
Example 12.5. Two plates are to be Joined by square butt weld
with a root gap of 2 mm using submerged arc welding process.
Sketch the desired weld and its symbol it a melt-through with a
reinforcement height of 3 mm is to be achieved.
Solution. The desired weld and its symbol are shown in Fig.
12.45.
 '1

Weld Joint. ••, Weld Symbols and Joint Design Principles 851
850 Welding Engineering & Technology
Example 12.7. Two flat plates are to be welded with a single- V
butt weld. The groove angle is 60° and a root gap of 2 mm is used.
A backing strip is also to be used which is,
(a) to be held in position with double fillet welds, and
(b) to be removed after welding.
I (jD I The process to be employed is shielded metal arc welding.
2mm --H- L3mm Sketch the desired welds and their symbols.
Solution. The desired butt welds and their symbols with: (a)
(a) Desired weld (b) Symbol
fixed, and (b) removable backings are shown in Fig. 12.47.
Fig. 12.45. Desired square butt weld with melt-through
of required height of reinforcement.
Example 12.6. A T-joint is to be made between two plates ~600J
using, (a) chain intermittent double fillet welds of Bmm leg length
and length 50 mm with a centre-to-centre spacing of 150 mm, and
(b) staggered intermittent double fillet weld of leg length 8 mm,
Ir--------:.<, V. I
and length of 75 mm with a centre-to centre spacing of 200 mm. Fillet
weld --iIIr- 2 mm
Sketch the desired welds and their symbols, if the welds are to be
made by CO2 welding process at site. (a) Desired weld Note: Material and dimensions of
backing as specified.
Solution. The desired weld and their symbols are shown in
Fig. 12.46. symbol

Removable
backing

(a) Desired weld (b) Symbol (b) Desired weld

Symbol for weld with removable backing.

(01 Welding
Fig. 12.47 Desired single vee butt weld: (a) with permanent backing.
1-75 and (b) with removable backing.
Example 12.8. (a) Give the welding symbol for the edge-flange
weld shown in Fig. 12.48, to be made by gas metal arc welding.
(b) Sketch the desired weld for the welding symbol shown in
Fig. 12.49. Name the weld.
Solution. (a) The symbol for the edge flange weld of Fig. 12.48
is shown in Fig. 12.50.
(c) Desired weld (d) Symbol (b) The sketch of the desired weld represented by symbol of
Fig. 12.46 Desired intermittent welds, (a & b) chain welds Fig. 12.49 is shown in Fig. 12.51. It is a corner-flange weld.
and (c & d) staggered welds.
852 Welding Engineering & Technology Weld Joints, Weld Symbols and Joint Design Principles 853

~'

~ \45--;

Desired weld
~~--It--·,-,::-!---.~~~
Section of desired weld
Radius of flange -1·5 mm Desired weld
Height of flange above point of tangency - 3.5 mm Size - 25 mm Depth of filling - 20 mm
Weld thickness - 2.5 mm
Angle of countersink - 45· Pitch (centre-to-centre spacingj-100mm
Fig. 12.52. Desired plug welds.
/
Fig. 12.48 Desired edge flange weld.
See detail A

Symbol Symbol
Fig. 12.53. Symbol for slot welds.
2S [10 1\00
Symbol Fig. 12.50. Symbol for the desired
Fig. 12.49. Symbol of a weld. edge-flange weld of Fig. 12.48.

l5R

Symbol
Fig. 12.54. Symbol for desired plug weld of Fig. 12.52.
2.5

Desired weld
Radius of flange -1·5 mm
Height of flange above point of tangency - 3.5 mm
Weld thickness - 2.5 mm
Fig. 12.51. Desired weld for the symbol of Fig. 12.49.
Example 12.9. (a) Sketch a welding symbol for the plug welds Desired weld

r
shown in Fig. 12.52.
(b) The welding symbol for two slot welds joining two
overlapping plates is shown in Fig. 12.53; show different views of
the desired weld so that the symbol is fully explained. ~. t2j.'. _ /f~
i'
j 1
Sec.A-A
Solution. (a) Fig. 12.54 shows the welding symbol for the plug A
Dl?Pth of filling -12
welds of Fig. 12.52. -625
Oetail A
Fig. 12.55. Desired slot welds with details for the svmbol ••hnwn in F;a 1? !;'l
 Weld Joints, Weld Symbols and Joint Design Principles 855
854 Welding Metallurgy and Joint Design
the minimum distance of the centre of the weld from the sheet edge
(b) Fig 12.55 shows the details of desired slot welds of Fig. is 13 mm. Sketch the desired weld details and the corresponding
12.53.
welding symbol if the total number of spot welds to be made is 5.
Example 12.10. 1Wo plates are to be lap welded by arc spot Solution. Fig. 12.57 shows the desired weld and the corres-
welding using gas tungsten arc welding process. The diameter of ponding welding symbol.
each spot weld at the faying surface is 6 mm, pitch i.e.
centre-to-centre distance between successive welds is 50 mm and a Example 12.12. Fig. 12.58 shows the details of arc seam
minimum distance from the edge of the plate is 25 mm. Draw a welds, sketch a welding symbol to represent this joint.
welding symbol to represent this joint if a total of 9 spot welds are
to be made.
Solution. Fig. 12.56 shows the desired arc spot welded lap
joint and the corresponding welding symbol.

GTA76

Desired
A
welds
\i'
Section A- A
Desired weld

Size lot faying surface) -13 mm

Length - 50 mm .
Pitch l center - to -center spacing)-75 mm

Size lot faying surface)- 6 mm

Fig. 12.58. Details of desired arc seam weld.
Number of spot welds - 9
Pitch (center - to - center spacing)- 50 mm
Solution. Fig. 12.59 gives the required welding symbol for the
welded joint shown in Fig. 12.58.
Note: Size can be given in pounds or newtons per spot rather than the diameter.

Fig. 12.56. Desired arc spot welds and their symbol.

Example 12.11. 1Wo overlapping sheets are to be joined by
resistance spot welds. The size of the weld nugget is to be 6 mm and

A (5) ~ I
I

A
~'-I-2s-4
-1
Desi red welds
~
I'-13 ,,' L"
Section A- A
-j~ I- / '
Symbol
Pitch (center -to-center
0 _ 2S

spacing)- 25
.5W
Symbol
Note: Size can be given in pounds per linear
in. or newtons per millimeter
Fig. 12.59. Symbol for arc seam weld of Fig. 12.58.
Example 12.13. Sketch the details of the welded joint for
Size (at faying surface)- 6 mm
Distance from center of first spot weld
which Fig. 12.60 is the welding symbol. Name the weld.
Number of spot welds - 5
to edge -13 Solution. Fig. 12.61 gives the details of the desired welded
Note. Size can be given in pounds or newtons per spot rather than the diameter.
joint. It is a resistance seam weld.
Fie:. 12.57. Desired resistance spot welds and their symbol.
856 Welding Engineering & Technology Weld Joints, Weld Symbols and Joint Design Principles 857

See Note
8 V25-5~RSEW
first
weld

Depth of preparationlarrow side)- 20 mm

Depth of preparationlother side)- 13 mm
Second Root opening - '·5 m m
weld Groove angle I arrow side): 60
Symbol Groove angle I other side)= 55
point
Note, If required by actual length of the joint.
the welds at the end of the joint should Fig. 12.62. Desired double-vee full penetration of weld.
be increased to terminate the weld at
the end of the joint.
Fig. 12.60. Symbol for the welded joint.

Note: BG-Back gouge to sound metal

~
Symbol CJP - Complete joint penetration
Section A-A
50 Fig. 12.63. Symbol for the weld of Fig. 12.62.
II of weld
,\60°;
1---"6--,
Desired weld
Size I at faying surface) - 8 mm
Length - 25 mm
Pitch I center- to-center spacing) - 50 m m
Insert
Fig. 12.61. Desired resistance seam weld joint for the symbol of Fig. 12.60.
I a) Part assembly before welding
Example 12.14. Represent the welding symbol, for the double

,---r
vee full penetration. butt joint shown in Fig. 12.62. The first weld
is to be back gouged to sound metal before achieving complete joint
penetration (CJPJ by laying the second weld. / -O,.(SAW
Solution. Fig. 12.63 shows the desired welding symbol for the
weld shown in Fig. 12.62. I
Example 12.15. A single-V butt weld is to be made, using ( b) Welding symbol
fusible insert and a groove angle of 60°, by submerged arc welding
process. Sketch the part assembly before welding and the welding (a) Part assembly before welding (b) welding symbol
Fig. 12.64 Single-V butt weld with fusible insert and its symbol.
symbol for the desired weld if the reinforcement is to be ground
flush.
 "'-'

858
Welding Engineering & Technology Weld Joints, Weld Symbols and Joint Design Principles 859
Solution. Fig. 12.64(a) shows the weld joint edge setting with Locations of elements of brazing symbol for a lap joint with
a fusible insert before welding, and Fig. 12.64(b) shows the fillet weld are shown in Fig. 12.67.
corresponding welding symbol.
Example 12.16. Sketch the desired compound weld joints the Cl- Clearance
welding symbols for which are shown in Fig. 12.65. L - Length of overlap
S - Fillet size

I Q I Symbol
\*-
I b I Symbol
CITctLjt
Desired braze
t.
Symbol
Fig. 12.65. Symbols of compound welds. Fig. 12.67. Locations of different !llements of a brazing symbol.
Solution. Fig. 12.66 shows the desired welds for the two Example 12.17. Draw a lap braze joint with a fillet weld
welding symbols of Fig. 12.65. having a lap length of6.5 mm and a clearance of 0.025-0.075 mm,
if the leg length of the fillet weld is 1.5 mm and the process
employed is torch brazing. Also, show the brazing symbol for the
desired braze joint.

j~.

(a] Desired
r
weld for Fig.12.65Ial
I
Ibl
I-----.I~
Desired weld for Fig.12.65Ibl
I
Solution. The desired braze joint and the corresponding
brazing symbol are shown in Fig. 12.68.

Fig. 12.66. Desired groove cum fillet welds for symbols of Fig. 12.65.
12.13. BRAZING SYMBOLS
Brazing symbols are constructed the same way as the welding
sYlnbols i.e. by using basically a reference line, an inclined arrow
and a tail for the reference line. Joint clearance can be indicated
on the brazing symbol and the process designation, listed in table
12.2, is indicated in the tail.
L _~
1~'SI015
--"
0·025 - 0·075
~_ I
Table 12.2.Latter Designations for Brazing Processes Desired braze
s. -'lbrch
Dip
Furnace
l. I Diffusion Process
Induction
brazing
brazing
Resistance
Infrared brazing
brazing
Designation
brazing
raybrazing
brazing
I FB
IB
DFB
DB
IRB
RB
TB
Symbol

Fig. 12.68. Desired lap joint with fillet braze and its brazing symbol.
Example 12.18. (a) Sketch the brazing symbols for the
following braze joints shown in Fig. 12.69.
(i) scarf joint,
(ii) flare-bevel-groove and fillet joint,
(iil) Double-flare-bevel groove and fillets joints.
(b) Sketch braze joints for the brazing symbols shown in Fig.
12.70. Name the joints.
 860
Welding MetaJlurgy and Joint Design Weld Joints, Weld Symbols and Joint Design Principles 861
See

l, O.025-:~0·
30"n~~Bg·
LO.025- 0.075
Desired braze

( a) Scorf joint
(a) Symbol for braze joint of Fig.12.69/a)
-tt-T
r 0·025 -0.075
, ~ ,: ~::2S~B075
L-L- -f I b) Symbol for braze joint of Fig·12.69Ib)

TIT I-~ . 1-5

S- Rodiu'
of tangency
from p.int
Desired braze
( b) Flar e - bevel - groove and fillet joint

/ c) Symbol for braze joint of Fig.12.69 / c)

Fig. 12.71. Brazing symbols: (a) for scarf joint, (b) flare-bevel-groove
S - Radius of tube and fillet joint, and (c) double flare-bevel-groove and fillet joints for
desired braze joints of Fig. 12.69 (a, b, c) respectively.
Desired braze

Ie) Double flore - bevel -groove and fillet joints .

Fig. 12.69. Desired braze joints (a) scarf~ (b) flare-bevel-groove and fillet joint,
and (c) double flare-bevel-groove and fillet joints. la)
•...

Desired
~
.1°.025

braze for
- 0·075
..
symbol of Fig.12.70 10)

~ ~B
I ~O'02S-0075 ~ n~~B ~O'050-0125
-I 0·025 -0.125

Symbol (0)
SymbOl (b)
Fig. 12.70. Brazing symbols for the desired braze joints. I b) Desired braze for symbol of Fig.12.70 lb)
Fig. 12.72. Desired braze joints for symbols of Fig. 12.70.
Solution. Ca)The brazing symbols respectively for braze joints symbols by using an additional reference line or by specifying the
of Fig. 12.69(a, b, c) are shown in Fig. 12.71Ca, b, c).
test method in the tail of the welding symbol.
Cb) The desired braze joints for the symbols of Fig. 12.70(a, b) NDT symbols are constructed the same way with the help of
are shown respectively in Fig. 12. 72Ca, b). a reference line, an arrow and a tail, as is the case with welding
12.14.NDT SYMBOLS and brazing symbols. Locations of different NDT elements are
Symbols for non-destructive testing of weldments are also shown on a standard NDT symbol as indicated in Fig. 12.73 and
used on the engineering drawing to specify the method of the abbreviations for different NDT methods are listed in table
12.3.
examination. NDT symbols may be combined with welding
 !
862 Welding Engineering and Technology 863
Weld Joints, Weld Symbols and Joint Design Principles

to be examined
Number of examinations ~

Rd.erenu line ---. \( N) in field RT

...
VI
••••••
J::"O ••• Examine all
Tail
o'ij\ "0

';j L L:=EXl1mine
Length of section
/ around symbol
Specification T
~ •.. J::
or other /'
r&!ferenc:e"""""'- "'''0
•...
<VI
- o
lXl

1==YI
Fig. 12.73. Standard locations of NDT symbol elements. Fig. 12.75. NDT symbol to specify the direction of radiographic examination.

Table 12.3. Non-destructive Testing Symbols 12.14.3.Combined Symbols

I
S. No. 1'ype of Examination Abbreviation / Symbol NDT symbols may be combined with welding symbols as
shown in Fig. 12.76(a), or with each other if a part is to be
1. Acoustic emission test AET
examined by two or more methods, as shown in Fig. 12.76(b).
2. Eddy current test ET Where an examination method with no arrow or other side
3. Leak test LT significance and another method that has side significance are to
4. MT be used the NDT symbols may be combined as shown in Fig.
Magnetic particle test
12.76(c). When a specification or other reference is used with an
5. Neutron radiographic test NRT NDT symbol, the reference is placed in the tail as shown in Fig.
6. Penetrant test PT 12.76(d).
7. I Proof test PRT
RT <B
8. I Radiographic test MT
MT ~ ""_ -U1~
-~-UT ,
UT
(a) Combined NO T and welding symbols
10.
9. I Visual test test
Ultrasonic VT
~ RT
12.14.1.No Side Significance - PT+ET ..../ LT '"
When the NDT may be performed from either side or has no RT+PTJ
arrow or other side significance, the basic examination symbol is RT + LT
centred in the reference line as shown in Fig. 12.74.
I bl Combined NOT symbols

-RT~ ~UT+RT"", LT+PRTRT1

Fig. 12.74. NDT symbol for 'No side significance'. ..tr..-UTJ
LT
12.14.2. Radiographic Testing
(e) Combined side and no-side significant symbols
The direction of radiation (X-rays or Gamma rays) may be
shown in conjunction with radiographic test (RT) and neutron
radiographic test (NRT) symbols. The direction of radiation may A 12'-/' RT /' ,,<
/'- --P-T--"~54
be indicated by a special symbol and line located on the drawing
at the desired angle, as shown in Fig. 12.75. (dlNOr symbol with reference
Fip' 12.76. Combined NDT and welding symbols ; and symbols for use of
 Weld Joints, Weld Symbols and Joint Design Principles 865
864 Welding Engineering & Technology
circle at each change of direction. The type of NDT to be used in
12.14.4. Number of Examinations the enclosed area is designated with the appropriate symbol, as
When several examinations are to be made on a joint or part shown in Fig. 12. 78(a).
at random locations, the number of examinations is given in (ii) Area of Revolution: For NDT of areas of revolution, the---·-
parenthesis, as shown in Fig. 12.77.

"'"-" HT
(4) 6
(9)
PT / area is indicated by using the examine-aIl-around symbol and
appropriate dimensions. In Fig. 12.78(b) the upper right symbol
indicates that the bore of the hub is to be examined by magnetic
particle testing for a distance of 75 mm from the flange face. the
lower symbol indicates that an area of revolution is to be examined
Fig. 12.77. Symbols for a number ofND tests to be conducted. radiographically. The length of the area is shown by the dimension
12.14.5. Area of Examination line.
NDT of areas of parts is indicated by one of the following The symbol shown in Fig. 12.78(c) indicates that a pipe or tube
methods: is to be given an internal proof examination and an external eddy
current examination. The entire length is to be examined because
(i) Plane Areas: To indicate on a drawing, a plane area to be no Hmiting dimensions are shown.
examined, the area is enclosed by straight broken lines with a
,....&--4 12.14.6.Acoustic Emission
RT ,.. ...•
"'- I Acoustic emission test is generally applied to all or a large
r---Y
I I <P' qr- --ED portion of a component, such as a pressure vessel or a pipe. The
<!r---0 I
&--0
I
(a) Plane areas
I

<P-~€>
--UT
1F=.
/-----
------~%
-------
- _n.
AET

Fig. 12.79. NDT symbol for acoustic emission testing.

(b) Area of revolution
HT 75 symbol shown in Fig. 12.79 indicates acoustic emission exami-
nation of the component without specific references to locations of
the sensors.
Example 12.19. A single-bevel butt weld with a groove angle
of 45°, root gap of 2 mm and a root face of 3 mm is to be made
between two 25 mm thick plates using submerged arc welding
process. The weld is to begiven a sealing run which is to beground
flush.
- one side
RT--

E~:?;:::::~~-,PRr ~ ET I ge:'
2---i T
r+s
(oj Desired weld (b) Combined welding and NOT
(c) Area of revolution - both sides symbol
Fig. 12.78. NDT symbols for examination of: (a) plane areas, Fig. 12.80 Desired single-bevel butt weld and its welding .cum NDT symbol.
and (b & c) area of revolution.
866 Welding Engineering & Technology
Weld Joints, Weld Symbols and Joint Design Principles 867
The weldment is to be examined by three NDT methods viz,
visual examination followed by penetrant testing and then- the 12.15. PRINCIPLES OF WELD JOINT DESIGN
gamma ray radiography. Show the combined welding and. NDT Although different factors like production operations, manu-
symbols. facturing costs, product performance, appearance, and customer
Solution. The desired weld and its symbol along with NDT acceptance influence the weld joint design but overall it can be
symbols are shown in Fig. 12.80. said
designthatare:three major factors that influence the selection of joint
Example 12.20.Sketch the desired weld for the symbol shown
i.n Fig. 12.81. (i) Load requirements,
3 rd operation (ii) Ease of welding, and
(iii) Cost.

(JP CJP - Complete joint penetration The joint design principles that are important in achieving the
BG - Back gouge to sound metal desired weld with ease at the least cost can be divided into two
BG categories, viz,
(i) General design principles, and
(ii) Basic/ specific design principles.
Fig, 12.81 Combined welding and NDT symbol for the desired weld. 12.15.1. General Design Principles
Solution. Fig. 12.82 gives the desired weld for the symbol of The design principles that influence the general considera-
Fig. 12.81. In this, the operations are: ' tions in selecting a joint design are:
1st operation - Make the first weld beyond the root face and back gouge (BG) l.The design should satisfy strength and stiffness require-
it to sound metal,
ments. Over designing wastes materials a,nd increases production
Hnd operation - Make the second weld to give complete joint penetration and shipping costs.
(CJP),
[lInd operation - Visually examine both sides and then carry out ultrasonic 2. The safety factor used in design should not be unduly high.
examination (UT) followed finally by radiographic testing
(RT). resist3. bending.
Deep symmetrical sections should be used to efficiently

4. Rigidity may be provided with welded stiffness to minimise

the weight of material.
--- / First weld
5. Tubular sections or diagonal bracing should be used for

~I weld
-
.~.o
~
N,
3
resisting torsional loading, Fig. 12.83.

Fig. 12.82. Desired weld for symbol of Fig. 12.81.

Details of Welded Joint
Root gap = 1.5 10m
Root face = 3 10m
Groove angle for first weld = 60°
Groove angle for second weld = 55°
Depth of preparation for first weld (arrow side) = 20 10m
Depth of preparation of other side i.e. second weld = 12 10m
Thickness of plates = 35 10m
Fig. 12.83. Torsional loading is well resisted by : (a) box tubular l'lp.d.inn

Welding Engineering & Technology
868
6. Standard rolled sections, plate, and bar should be used for
economy and availability. This also helps in that the welding ofeach
member can be balanced about the neutral axis.
7. Plan the design to minimise the number of pieces. This will
reduce assembly time and the amount of welding.
8. Weld metal costs can be reduced for thick plates by specifying
J- and U-groove edge preparation.
9. Consider air carbon arc gouging, flame gouging, or chIpping
for back weld preparation.
10. Create a corner by bending or forming rather than by
welding two piece together.
11. Bend flanges on plate rather than welding flanges on it.
12. Use a casting or forging in place of a complex weldment to
simplify design and reduce manufacturing costs.
13. Use a surfacing weld on an inexpensive component to
provide wear resistance or other desired properties instead ofusing
expensive alloy component.
~4. Rollrings from bar stock instead ofcutting from heavy plate,
Fig. 12.84.
00\
((J)
I b) Roll ringslOstead
(a 1 Flame cut rings from fro m plate
thick plate
Fig. 12.84. Roll rings instead of cutting from plate.
12.15.2. Basic/Specific Weld Joint Design Principles
The weld joint design should be selected primarily on the basis
of load requirements and most of the time the structure that is to
be fabricated will dictate the type of joint to be used. However,
variables in design and layout can substantially affect the costs.
Generally the following rules apply.
1. A joint of the type that will require the minimum amount of
weld metal should be chosen. Use the smallest practical root opening
and groove angle. For example, if a 20° groove can be welded
Fig. 12.85. Possibilities of making a butt weld with different groove angles.
,'I
I - - . '- -----"'---~,,:.:;:;:;::;;; -'C'::':<lll ,
"I'·
I "
Weld Joints, Weld Symbols and Joint Design Principles 869
satisfactorily there is no reason for having greater groove angles,
Fig. 12.85. This not only reduces distortion and residual stresses but
also the need and cost ofpostweld stress-relieving and straightening.
2. On thick plates, use double V- or U-groove welds to reduce
the amount of weld metal and to control distortion. In some cases
a double-V weld will require half the weld metal than that required
for a single-V weld; as shown in Fig. 12.86.
Single and
double V
Fig. 12.86. Schematic representation of weld metal required for single and
double Vee groove welded butt joints.
The use of automatic welding, with its deeper penetration, will
reduce the bevelling required on many hand-welded joints.
3. Make the maximum use ofthe weldedjoint. Several parts can be
jqined with one pass weld. For example, use one weld in place oftwo or
three welds to joint three parts at one location,as shown in Fig. 12.87.
~ ~ S ~_ -.--'" ~
~ 11 I 1'1
of cutting (a I Bad design ( b) Good design
Fig. 12.87. Joining three parts by welding:
(al bad weld design, and (b) good weld design.
4. Fillet welds are always designed on the basis of shear stress
on the throat.
5. Flat 45° fillet welds are more economical than concave fillet
welds, as the strength is still dependent on the actual throat
distance, as shown in Fig. 12.88.
~"~~
Fig. 12.88. Configurations of three types of fillet welds.
6. Curved intersections, such as welding two round bars side by
side or the Vbetween two bent sections, Fig. 12.89,should be avoided
if possible, since they are more difficult to make and often take a
considerable amount .:f weld metal.
 •••-,-_ •.~..•,.';"":-'''"'-,=.,..,:"_ •.",,,._•••• ~

m
870 Welding Engineering & Technology Weld Joints, Weld Symbols and Joint Design Principles 871

7. When planer is available, the V-groove is preferred for

thicker plates as it requires less weld metal than conventional
bevel type preparation. This also reduces distortion.
S, The joint can be made better, faster, and more economically
if it is easily accessible as is shown in Fig. 12.90(b). Fig. 12.90(c)
shows that welding of a hollow circular section too close to the
(Q) ( b) (c) sides does not allow proper electrode positioning. This may result
Fig. 12.89. Typical welds which should be avoided, if possible. in acceptable welds for average work but will.be bad for leakproof
III welding.
co ~
t°
'0 ~::: In butt welds, however, accessibility is generally gained by
0 __ " = compromising between groove angle and root opening, Fig. 12.91.
~~>='i'E
'0. ~~ •••'0
C7\0._ '0 c:
.SU _Ill
a 0E c "0

~Jld
cw '"
.2.c~.,.,
0.-0.0'0
Oc: 0
.~ U'J.- U;.; ~ ~
O=4W;:'u > u
i;~cE.!! ..= ;
!:! •.•
--t.l~c
O. U1~~

~~.~ ~~
~C:1Il
~.-
~-g .
0"<::;::
.•.. ~'"
al "l:l ~ Right (4-5') Wrong ( 22 1/2")
O'l
!:::
~ ~ .•..0
'Oo>

~"O~
~" " . ~.~
0l •..•'O
oQl::'"
Fig. 12.91. Effect of groove angle and root opening on welding accessibility.
c:o- t:j' "' •• ~
~bl)
''-.' ~·c ~ ~
-0 .. I:: 0
'" .- .s 9. Joints should be made with the best possible fit-up because
15:=~o joints with gaps larger than necessary are costly to fill.
.2lS~e ~ ~~
:!Z:!Z"O
~ ~ u
•••0. •••0.
'0 •.. '"
>.><
'iij ••• c c 15.s"C. 10. The weld size should be correct. The leg size of a fillet weld
£20-: I.J
~.od
I:: '-'~ I::
fW_- 0
~~~ .•..
._:;:: should be £th of the plate thickness if full strength is required. If
u'" •••
'0
C
~ ~ 8-
-"'8 rigidity is the main requirement, 50% of the full strength of the
0
o\i).o.o
.28">-
fil
4>
•..
~
u u
'" '"
weld is sufficient, Fig. 12.92.
.•...
: 0.2:~ .0;, '" I:: •.•
~o~
gUJbD
"'~I:: Force
~ :g,~
'" '- 4>
25 25
~ 0 ~
'L/'l
~ cO ~.s
o ~, O'l '-
.1::
C'-l
... 0 '-
•..
~.l!!
~," '- 0
.•..
b.D or;;

u ••• rz..c.

!!~ 25 2S
•••
C'"OIl - ...
••••••
.DC
;;;0-
- ~'
.

"
""")~ ,~/" .•.
"'"
, E.><
I
~ E
Co _
c_
c
'5
~
'6 ..c
I a I Fu\l- strength weld desi gn ( b I 50 % strength design
;,c .•. for rigidity
>-
° ....
-c~
""
E:iC
'" c
"tJE
.•.
-0 >
•...
••• Fig. 12.92. Adequate fillet weld size: (a) for full strength,
and (b) for rigidity only.
-~.c
o
'- c:
.•.
~
t/
.•.
~.c
~~
i:ij~
•••
--------~---------------------------------------------------------!!!!!-!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!,,!~!"""!,'!'"~-!!!·-!""!,~!""'~"~~
i

872 Welding Metallurgy and Joint Design

Weld Joints, Weld Symbols and Joint Design Principles 873

11. Where intermittent fillet welds are satisfactory, Fig. 12.93,

they can be used as a means of reducing costs in hand welding (0)
oper ations.
I, ~- ~ ~
Poor Good Better

(b)

Fig. 12.93. Intermittent fillet weld to reduce cost. Poor Good Better
12. For material 6mm or more in thickneas the mInImUm
length of intermittent welds shall be 8 times their normal size but
not less than 50 mm; the maximum length shall be 16 times their
normal size but not more than 150 mm.
(Cl~
The maximum centre-to-centre distance between the adjoining ~
intermittent welds shall be 32 times the thickness of the thinner
plate, but in no case shall the clear spacing between intermittent
fillets be greater than 300 mm. Poor I.....--G-o-od----

13. Large fillets require more than one pass for manual
welding. For example a 9.6 mm fillet instead of a 8 mm fillet
requires at least two passes, Fig. 12.94, with a 44% increase in (d)
,I

II Poor Good
9.6~ 8~ I
9·6 8 Fig. 12.95. Edge preparation to avoid lamellar tearing in a corner joint.
(After Lancaster)
440/0 more weld metal·
only 20 stronger
% ' material to over.
minimi3e lamellar tearing, especially in thickness of
50 mm and
Fig. 12.94. Size effect of a fillet weld on its cost and strength,
cost though only 20% stronger. Distortion, internal stresses, and 16. For T-joints, where fatigue is not a primary concern, use
the need for subsequent stress-relieving can be reduced by keeping double fillet welds to minimise through-thickness shrinkage
the weld material to a minimum. strains and to spread the transfer of applied forces, Fig. 12.96.
14. Where possible, use a square groove joint together with a
welding process capable of deep joint penetration or welding Good Poor
within the groove, for example electroslag, electrogas, or narrow
gap welding.
15. For corner joints in steel, Fig. 12.95, specify a depth of
preparation of the groove to intersect the full thickness of the (a) Double fillet weld (b) Groove weld
Fig. 12.96. Double fillet weld: (a) is better than groove weld (b) for T-joint when
fatigue is not of primary concern.

814 Welding Engineering
17. Avoid joints where it is difficult to obtain fusion at the roo~.
For example, welds joining two surfaces at less than 30° to each
other are often unreliable ..
18. Specify welds of minimum size and length but adequate
for the forces to be transferred. The size of a fillet weld is especially
important because the amount of weld metal required, for fillet
with equal legs, increases as the square of the increase in leg
length.
19. For equivalent strength a continuous fillet weld of a given
size is usually less costly than a larger sized intermittent fillet
weld. Also, there are fewer weld terminations that are potential
sites for initiation of defects.
20. To derive maximum advantage of automatic welding it may
be better to use one continuous weld rather than several short
welds; design for straight or circular welds, Fig. 12.97.
Fig. 12.97. Try to design sections round or straight
so that automatic welding may be used.
21. The size of the weld should always be designed with
reference to the size of the thinner member and it should be placed
in the section of least thickness, Fig. 12.98.
=O=rJ (0)
Bad Good Bad
Groove weld (bl Fillet weld
Fig, 12.98, Size of a weld should be determined with referenc3 to the thinner
member, for example (a) groove weld, and (b) fillet weld.
22. Welding of stiffners or diaphragms should be limited to
that required to carry the load, and should be based on expected
Weld Joints, Weld Symbols and Joint Design Principles 875
& Technology,
out-of-plane distortion of the supported components under service
loads as well as during transportation and handling.
23. A large structure should first be fabricated in two or more
subassemblies because they can be worked on simultaneously thus
reducing overall time of fabrication. This also provides better
access for welding and may also permit automatic welding.
24. Joints in thick sections should be welded under conditions
of least restraint; for example, prior to installation of stiffners.
25. Weld the more flexible sections first to facilitate
straightening that might be required before final assembly.
26. As-welded joints that have uniform appearance are
acceptable for many applications. Therefore, the surface of weld
need not be machined smooth or flush unless that is required for
some other reason. Smoothing a weld is an expensive operation.
27. Fillet welds are generally applicable where stresses are
low, and the required weld size is less than about 16 mm. If the
load would require a fillet weld of 16 mm or more, a groove weld
should be used possibly in combination with a fillet weld to provide
the required effective throat.
28. Fillet welds may be used in skewed T- or corner joints
having an included angle between 60° and 135°. Beyond th(3se
limits, a groove weld should be used.
29. Avoid double-T or cruciform joints whenever possible. Such
joints have maximum locked up stresses.
30. For corner joints when bevel edge preparation is used,
prepare the thinner member whenever possible.
31. Plug or slot welds, or fillet welds in holes or slots, shall
not be used in highly stressed members unless absolutely
necessary. They shall be used where subjected principally to
. shearing stresses or' where needed to prevent buckling of lapped
parts.
32. Joints requiring the highest efficiency must specify full
penetration weld.
33. Whenever possible require only one member of a joint to
have bevel joint preparation.
Good
34. Weldments should be designed so that the position in
which the welds are made shall have the following order of
preference :
(a) Fillet welds: flat, horizontal fillet, horizontal, vertical,
overhead.
(b) Groove welds: flat, vertical, horizontal, overhead.
35. In dynamic loading situations the joint profile is much
more significant and it is important to avoid sudden change in
section as that may act as stress raiser. This is particularly
 ~,_,,_,,'o>"'_bu"""~~••__r~·'",
,__ '_ ••.• ",.
_ .,~_'T

Welding MetalIurgy and Joint Design Weld Joints, Weld Symbols and Joint Design Principles 877
876

38. Welds placed in transverse shear are 30% stronger than

Weld Weld
those placed in longitudinal shear, Fig. 12.102.

I-x-V===I
(a I Poor
I ~
(b I Good
1/ 39. Where the design permits, fillet welds are used in
preference to groove welds for reason of economy.
Longitudinal shear

Fig. 12.99. Poor (a) and good (b) welding designs between thick and thin sections.
x
important in joints subjected to fatigue loading. Fig. 12.99 shows Transv.erse
the poor and good weld designs between thick and thin sections. shear
36. Dissimilar thickness may also be joined using lap and fillet x

configurations but it may be necessary to equalise the cross-

section at the joint to enable satisfactory edge or butt welds to be
performed, as shown in Fig. 12.100.
Fig. 12.102. Welds placed in transverse shear are
30% stronger than these placed in longitudinal shear.
40. For a T-joint, a single-bevel-groove weld would be less
expensive than a double fillet weld.
41. Build up composite sections by welding Fig. 12.103, to
reduce machining and material costs.

•••
I
J

~
__
~1:JmO.,---k
Fig. 12.100. Machined preparation to equalise thickness
in tube-to-heavy plate weld.
:37. In positional applications non-symmetrical preparations
may be necessary to improve joint profile and avoid excess weld
~ - --
metal, as is shown in Fig. 12.101.

D E~~--1~~i:~~
D
Fig. 12.103. Welding machined components to build composite sections.
42. Sometimes a bend can be used in place of a weld, Fig.
12.104, to substantially reduce fabrication costs.
If from amongst the above listed, the relevant principles are
Fig. 12.101. Unequal edge preparation for horizontal followed the designer can arrive at a weld joint which is expected
butt weld in vertical plates. to serve well in the service conditions for which it is designed.
 .~ ~~_.C""'~""""""_"'" =~""_"".'''''~;:::,===",,,=, ::::.,
..~..::::.~
.•.~..~_~
...~..,!:~>,!:,._.~
.. ~..~~~~ .•._~.,~,_~s.'~"_~'
.~'''~.,,~
.••.••,,,. ~

Weld Joints, Weld Symbols and Joint Design Principles 879

878 Welding Engineering and Technology

12.16. EVOLVING GOOD WELD DESIGNS

Although a designer may follow the general basic and even the
speci11cdesign principles but may still
not produce a satisfactory product. This
happens because probably one of the
most serious errors in welding design
practice is the tendency of the designer
to make the new component assume an
appearance strikingly similar to one
whose parts have been riveted, cast or
forged. This tendency may be .attribut.ed
to the inert.ia of human mind. There is
no merit in following this line of action
unless there is some definite advantage
in doing so. The weld design should be
evolved taking into consideration all
the desired aims of the final product.
This may often take considerable time
and a number of trials. Some cases are
dis'cussed in this section which may Fig. 12.104. Consider whether a
prov]'d e gul'd ance In
. evoIvIng
. a good corner should be welded or bent.
welding design.
12.16.1. Conversion from Castings to Weldments
Quite often cast components are converted to welded compo-
nents as that often leads to considerable savings. Three typical
cases are described in this section.
12.16.1.1. Redesign of Cast Iron Motor Base
Fig. 12.105 shows the original cast iron base weighing 310kg
which was required to be redesigned so as to be produced as a
welded component. Three attempts were made to arrive at an
op~imal design.
Fig. 12.106. First redesign of casting shown in Fig. 12.105 ; welded steel base
(128 kg) with three stiffeners on 380 mm centres.
The second attempt at redesigning was based on the moment
of inertia of the cast iron base. Ifmoment ofinertia of the cast iron
base can be found it can be multiplied by the equivalent rigidity
factor (40%). With the redesign based on an equivalent m.o.i., the
base could be made from 8 mm thick steel plate, instead of 10 mm
of the first redesign, bent into the form of a channel as shown in
Fig. 12.107. Bending eliminates preparing the edges for welding,
as well as welding three pieces together. However, it may increase
the manufacturing cost if only a few components are to be made,
but would result in lower cost if several are to be made at the same
time. Note that an additional stiffener is required because of thin-
ner top panel.
~
8mm ---
-
~.8mm --
Fig. 12.107. Second redesign of casting in Fig. 12.105; welded steel·base
(128 kg) and four stiffeners on 305 mm centres.
The ultimate design arrived at was as shown in Fig. 12.108.
Here flanges have .been bent into the plate at the bottom to give
still greater rigidity. This permitted a further reduction in plate
thickness to 6mm but requires still another stiffener. The weight
has finally been reduced to 113 kg, i.e. to 36.5% of the original.

~ .~6.3mm
16mm
Fig. 12.108. Final redesign of casting in Fig. 12.105 ; welded steel base
Fig. 12.105. Original cast iron motor base (310 kg.). (113 kg) and five stiffeners on 255 mm centres.
Fig. 12.106 shows the first redesigned base. Only two 12.16.1.2.Redesigning of a Gear Housing
operations were required for its fabrication, shearing and welding. Fig. 12.109 is the original cast iron gear housing while Fig. 12.110
The weight was reduced to 128kg i.e. to 41.3% ofthe original cast shows its first and second redesigns as a welded component.
base and the cost was about 30 to 35% that of the casting ..
880 Welding Engineering & Technology
Weld Joints, Weld Symbols and Joint Design Principles 881

Casting

Fig. 12.109. Original cast iron gear housing.

The casting illustrated is a bottom section of a gear housing. It
support.s the upper part ot; the housing, holds the oil supply, and (al Cast housing
provides holes for attaching the upper part with 10mm diameter studs,
In the first attempt at redesign as a welded component the Welded steel
bottom panel thickness was reduced to 10 mm but to provide
adequate width for 10mm diameter studs, 25 mm thickness was
used for the side member of the housing. To make the housing oil
tight the side members were continuously fillet welded to the
bottom plate on the outside; they also provided intermittent fillet
welds on the inside. Welds were also required to be used at corners.
The designed component served the function wen but 25 mm thick
i:lide member was obviously a wasteful use of material.
Oil pan

(bI Welded housing

Fig. 12.111. Complete gear housing; (a) original grey iron casting,
and (b) suggested welded replacement.
The welded gear housing of Fig. 12.111 is considerably
different from the original cast gear housing and much cheaper
than that or the alternatives suggested by the first designer. If
this new design can be integrated with the overall machine it will
(al ( bl obviously be more economical than the other alternatives.
Fig. 12.110. Welded redesigns of cast iron gear housing of Fig. 12.109. 12.16.1.3.Redesign of Steel Casting used for
In the second redesign, Fig. 12.110(b) the side member used was Transmitting Heavy Loads
11 mm which was thick enough to provide a bearing surface for the Fig. 12.112 shows the original design of a steel casting with a
top portion ofthe gear housing. Bars were weldedto the side members very large force, F, exerted on it.
to accommodate the hold-downstuds. Although this design served the
purpose well and was economicalcompared with the original casting
but the line of thinking of the designer is just based on the shape and
geometry of the original casting. When another designer was given
the requirements ofa gear housing without showingthe original cast-
ing, he arrived at the design shown in Fig. 12.111 (b) along with the
original complete cast iron gear housing, Fig. 12.111 (a).
-p F-u3 ~
Fig. 12.112. Original steel casting for transmitting large force.
 === ""'"""-,u;,;,""""'-I«oilc~.

882 Welding Engineering & Technology Weld Joints, Weld Symbols and Joint Design Principles 883

-r ~i) <
t
( a) First welded design

(a) Grignal design (b) Poor redesign (() Good redesign

Fig. 12.114. Redesigned weld joint to avoid mixing of weld types and weld joints;

( b) Second
F~)
welded design
(a) initial design, (b) first redesign which caused mix up of weld types,
and (c) final corrected joint design that avoided mix ups.
accommodate another component which fouled with a part of the
lower plate. This led to a suggestion of completely removing a

-t F~
---
I B
section of the lower plate to get a weld joint of the design shown
in Fig. 12.114(6).Then the needed cutout was made on the drawing
board but this created a problem by mixing of weld joints and weld
types. The weld in Fig. 12.114(6) changes abruptly, at the cutout
o
from a fillet to a groove weld. The groove weld would obviously
(c) Final. welded design require cutting a bevel into the vertical plate; for welding that a
Fig. 12.113. Steel casting of Fig. 12.112 redesigned as a welded component; smaller diameter electrode would be required to make the stringer
(a) first weld-based design, (b) second weld-based design, and bead in the groove joint portion. A simple method of joining the
(e) final weld-based design. base to the vertical plate ·had, thus, been destroyed.
The first redesign as a welded component shown in Fig. The joint design was revised to permit continuous fillet
12. 113(a) is just a copy of the original casting. Because of the large welding as shown in Fig. 12.114(C). This led not only to reduced
force transferred from part C to part A thrOl. lgh part B, all these fabrication costs but also gave a much better appearance to the
welds must be full-penetration groove welds. Heavy welds were weld joint.
needed because plates are thick. The procedures for welding were 12.16.3.Designing to Avoid Opening of Notches
important because plates are of Q and T steel. Thus, high strength
weld metal was used. A fillet weld between rim and disc of a gear shown in Fig.
In the second redesign the component was made of plates A. 12.115(a) resulted in high stress concentration at the root of the
which were welded to plate B by fillet welds as shown in Fig.
12. 113(b). Thus, the force could be directly transmitted through
the same plates into the desired opposite direction. However, the
design was further improved by using thinner plates - both A
and B; additional plate C was used to provide bearing surface for
the force when it is applied to plates A. As this design res'ulted in
the welds being not very critical the size was reduced further;
however plate D was added to provide proper stiffness to part C,
as shown in Fig. 12.113(c). Also, high strength weld metal was not
required. Welding costs were lowered and the welds were no more
critical as they have been used to connect secondary portions.
12.16.2. Redesign to Avoid Mixing of Weld Types
An engin~ designed a weld joint shown in Fig. 12.114(a), ( a) Poor ( b) Good
however it was required to be subsequently redesigned to Fig. 12.115. Fillet weld between rim and disc of gear;
(a) poor design. and (b) good design.

884 W.elding Engineering
weld under service conditions. This was because the notch formed
by the root of the weld is transverse to the direction of applied
force. The problem was overcome by increasing the size of the fillet
to get a weld joint shown in Fig. 12.115(b).
The improvement in design was brought about by shifting the
notch at the root of the weld from transverse to parallel direction
with respect to the direction of stress.
1~.16.4.Redesign the Whole - Not in Parts
The welded component shown in Fig. 12.116 is a redesign of a
cast housing for an earth tamper·. Cast designs are usually broken
Fig. 12.116. Weld-based design of a cast housing for earth tamper.
down into units that are bolted together. The original casting had
a pair of brackets, one on each side of the housing, to which
handles were bolted. Fig; 12.1)7 reveals th'at a similar pair of
brackets was welded on in the redesigned component, imitating
unnecessarily a needed provision in the cast design. There is, of
[a I Poor design r b I
Good design
Fig. 12.117. Welded structure; (a) unsatisfactory design for torsionalloads.'-'
and (b) satisfactory design with tube members eliminated.
* Earth tamper is used for rlUIlming down road material.
Weld Joints, Weld Symbols and Joint Design Principles 885
& Technology
course, no point in bolting the handles to the brackets as they
could be welded directly to the housing. It is, thus, imperative to
consider the whole component for redesigning as a welded
fabrication rather than part by part.
12.16.5.Redesigning to Eliminate Unsuitable Members
The welded structure shown in Fig. 12.117(a) appeared
adequate for loads coming upon it in service. The 150 mm diameter
tube in the centre being the backbone of the structure helped to
maintain alignment of parts during fabrication. In service, the
weldments performed satisfactorily until it encountered an
unexpected torsional load which could possibly repeat itself
occasionally. When this happened again the weldments failed at
the fillet welds joining the tubes to the ribs.
One method to overcome the problem was to increase the fillet
weld size until the weldment no longer failed in service. Analysis
of the problem, however, showed that the ribs would give little
resistance to torsional loads. Consequently, almost all the
torsional load was transferred through the fillet welds into the
tube. The tube ,and the welds were thus severely over stressed.
In the redesigned component shown in Fig. 12.117(b), the 150
mm tube and the side plates were replaced with two large
horizontal channels. These and three intersecting channels from
rigid box sections when joined to the flat members provided a
satisfactory welded design construction. Elimination of unsuitable
tube members solved the problem.
12.16.6.Proper strengthening of a Poorly Designed
Component .
Fig. 12.118(a) shows a gooseneck used in an earth moving
trailer. When the original produ~tion unit was field-tested, the
gooseneck failed and was repaired with the patch plates.
Subsequently, the patch plates were added to the goosenecks on
the production line itself and became a part of the accepted design.
Thus, what was originally no more than an expedient method of
correcting an error became a method of design ..
Reinforcing a member by patch-plate method is not necessarily
good, even though it works. A better way of solving the problem
is to analyse the component, discover the cause of weakness, and.
strengthen the beam in more scientific manner, as is shown in Fig.
12.118(b). The analysis of the bend section shows that dq.ring
) loading the neutral axis of the beam shifts inwards in the curved
region greatly increasing the tensile bending stress in the inner .
flange. Therefore, what is needed is a thicker flange in this regioIf:
 886 887
Welding Engineering & Technology Weld Joints, Weld Symbols and Joint Design Principles
Patch plate --j60!"m l-

101 (a) Correct welding symbol for plates at 90·

(a) Original design witt. patchplates welded on it

( b) Wrong welding symbol for plates at 60·

I c) Correct welding symbol for plates at 60·

( b) Improved design
Fig. 12.118. Gooseneck used in earth-moving trailor ; (a) reinforced
with patchplates. and (0) improved design - reinforced with
suitable engineering practice.
The outer flange has a much lower compressive bending stress
and does not need to be increased in thickness.
12.17.IMPORTANCE OF PROPER AND REALISTIC
SPECIFICATIONS
When designs are to be altered, the designer should carefully
check the changes to be made in the welding specifications and
symbols. Otherwise serious errors could occur. For example, for
many years a company had been welding intersecting plates at
90° with a complete-penetration double-bevel groove joint shown
in Fig. 12.119. Full strength was needed which was achieved by
this design. Later when a job in which plates of the same size
intersected at 60° was undertaken the design engineer unthin-
kingly indicated the same welding symbol that resulted in the
Fig. 12.119. Full penetration double-bevel groove joint; (a) plates intersecting
at 90°, (6) weld as a result of symbol of (a) used for plates intersecting at 60°.
and (e) correct welding symbol and weld for plates intersecting at 60°.
production of a weldment shown in Fig. 12.119(b) that involved
the deposition of 12.9 kg/m of costly weld metal.
Fig. 12.119(c) shows the proper way of specifying how plates
intersecting at 60° are to be welded with a full-penetration,
double-bevel groove weld. A considerable saving in weld metal is
affected simply by shifting the central position of the joint so that
rather than bevelling 28mm on each side the bevel is 15 mm on
the inside face of the joint and 45 mm on the outside face. Th:is
reduced the amount of weld metal, excluding the reinforcement,
from 12.9 kg/m to 8.6 kg/m, a net saving of 33.3%.
Realistic Specifications. Fig. 12.120 shows a section of a
welded frame made from 200 mm, 5.2 kg channel sections. There
are some 25 mm thick plates welded to the top of the frame to
serve as pads between the machinery and the frame. The design

! I
 I' 'i

f 'I,

---- ~ 'I
. SSM Weld••• Engl••••••• a.d T"bnology .',
-r
25mm
!
• 'I
.
-L
11 50-100

C 200 x 5.5 #:
Fig. 12.120. Details of a section of a welded frame
as a result of wrong welding symbol.
engineer has indicated a Smm intermittent fillet weld, 50 mm in
length with 100mm centre-to-centre distance between welds. The
instructions are unrealistic, since tl-le welder is required to put
11 mm welds on the toe of a flange whic~'.is only 6mm thick. The
designer obviously committed the easy error of thinking of the
thickness of the 25 mm thick plate.
Such pitfalls should be avoided i.n weld join-..design.
THIRTEEN
Weld Design for Static Loading
In the case of weld design for static loading the ultimate
tensile stresses of the weld metal are close to those found for
corresponding parent metal while conditions are different in the
case of dynamic or alternating stresses.
Investigations have shown that stress concentrations do occur
in welds, for example at the heel and the toe of the fillet welds,
where stresses may be high even when the load is relatively smalL
As the load increases, the yield point will soon be reached at the
point of stress concentration, where slight yielding will occur, thus
transferring the maximum stress to another point of the weld.
This process continues until the stress distribution over the whole
load carrying section is fairly uniform. Such balancing of the stress
distribution takes place under static loads and, for the purpose of
stress calculations, does not affect the Imal stress in the weld to
any considerable extent.
13.1. DESIGNING FOR S'ffiENGTH AND RIGIDITY
A weld design may require strength only or strength plus
rigidity. All designs must have sufficient strengths so that the
members will not fail by breaking or yielding when subjected to
usual operating loads or reasonable overloads. Strength designs
are common in road machinery, farm implements, motor brackets,
and various types of structures. If a weldment design is based on
calculated loading, design formulas for strength are used to
dimension the members.
In certain weldments such as machine tools, rigidity as well
as strength is important, since excessive deflection under load
would ruin the precision of the product. A design based on loading
also requires the use of design formulas for finding the weld
dimensions or the sizes of members.

89& Welding Engineering & Technology
Some parts of a weldment are classed no load parts, meaning
that they serve their design function without being subjected to
loading much greater than their own weight. Typical no load
members are dust shields, safety guards, enclosures for aesthetic
purposes, etc. Only casual attention is required to be paid to
strength and rigidity in the design of such components.
13.1.1. Design Formulas
The design formulas for strength and rigidity always contain
terms describing load, member (size, shape, cr, E, I, Z, etc.) as also
stress and strain. If any two of these terms are known, the third
can be calculated.
A load is a force that stresses a member. The result is a strain
measured as elongation, contraction, deflection, or angular twist.
A useful member is designed to carry a certain type of load within
a certain allowable stress or strain. In designing within the
allowable limits, the designer should select the most efficient
material and the most efficient section, size, and shape. The
properties of the material (e.g. strength, modulus of elasticity,
ete.), the section (area, length, moment of inertia, section modulus,
radius of gyration, etc.) determine the ability of a member to carry
a given load.
The design formulas in use, developed for various conditions
and types of members, are innumerable and obviously cannot be
listed. However, the use of such formulas may be illustrated by
the problem of obtaining adequate stiffness in a cantilever beam.
The problem obviously involves the amount of deflection at the
end of the beam under a concentr':lted load, as shown in Fig. 13.1.
The following deflection formula may be used,
FL3 ...(13.1).
L\ = 3f£!
where, F = load (force) causing deflection, L\,
L = length ofthe cantIlever,
E = modulus of elasticity of cantilever material,
I = moment of inertia of the cantilever.
l·- ··~l
1 L -.--t F
Fig. 13.1. Deflection (6) of a cantilever beam under a
concentrated load F applied at the extreme end.
I
I
~ I
i!
Weld Design for Static Loading 891
Since it.is desirable to have the least amount of deflection, the
equation (13.1) makes it clear"that E and I should be as large as
possible. The commercial engineering material having the highest
\J
modulus of elasticity is steel, with a value of 207 x 109 N/m2
(30 x 106 psi) for E. The choice of material, thus, often falls on some
or the other type of steel. Once the material is selected the only
other factor requiring a decision relative to deflection is I i.e.
moment of inertia of the cantilever which depends upon the shape
of cross-section. Thus, to keep L\ small a section with a large I will
be needed, whatever be the shape of the section. It should,
however; be noted that the designer must select a shape to
accomplish the best design at the lowest cost. The standard rolled
steel sections, thus, often fulfill most of the requirements.
13.1.2.Loading
There are five basic types of loads uiz., tension, compression,
bending, shear, and torsion. Whatever be the type of load, when
it is applied to a member, the member becomes stressed. The
stresses cause strains or movements within the member, the
extents of which are govemed by the modulus of elasticity of the
ma~rial. Since a load always produces stress and strain, some
movement always occurs.
. Failing stress
Allowable or Workmg Stress = co a fietf.
y act or
13.2. MATERIAL OR SECTION PROPERTIES
For solving the design problems it is essential to have
knowledge about the properties of the material and the section
selected for the design. Some of the more commonly employed
properties include the following :
(i) Modulus of elasticity,
(ii) Poisson's ratio,
(iii) Modulus of rigidity,
(iu) Moment of inertia,
(v) Polar moment of inertia,
(ui) Section modulus.
13.2.1.Modulus of Elasticity
In designing steel members and complex structures a material
with high ductility is generally desired because it is capable of
accommodating appreciable strains and redistributing stresses
without premature fracture.
For ductile materials, the selection of allowable strength
values of a material is based on its stress us. strain diagram. One
such diagram for mild steel is shown in Fig. 13.2.
~ I -
892 Welding Metallurgy and Joint Design i

Weld Design for Static Loading 893

U.T.S.
--------------- Poisson's Ratio. For metallic materials Poisson's ratio may vary
between 0.25 and 0.35 while for rubber it approaches 0.45.
,_\ 13.2.3.Modulus of Rigidity
For tangential or shear stress, the constant ratio of stress to
strain is called Bulk Modulus or tQe Modulus of Rigidity and is
. Force
denoted by the symbol G. Both E and G have umts -Area .
t Vl
~ In analysing the shear properties of a given material the
b :;) modulus of rigidity G can be easily determined by the following
'" relationship if the corresponding modulus of elasticity in tension
-
'"
QI
•..
Vl
E and Poisson's ratio, 1-1, are known.
E
G ="2 ( 1 +1-1) ...(13.3)
13.2.4.Moment of Inertia
Strain, £. - The second moment of area of a section is more popularly
Fig. 13.2. Stress-strain diagram for mild steel. though incorrectly called the moment of inertia of the section and
is usually denoted by the symbol I. It expresses the sum of the
In dealing with stress strain curve, a portion of which is linear, products found by multiplying each element of the area by the
the modulus of elasticity (E) is defined by the slope of the linear square of its distance from a given axis. This axis lies in the plane
part. In other words modulus of elasticity of a ductile material is of the cross-section. The moment of inertia has a unit of length to
the ratio of stress to strain and is a constant value within the the fourth power usually in mm-1.
elastic limit which for practical purposes may be regatded as the
same as the yield point. Whereas the moment is the tendency towards rotation about
an axis, the moment of inertia of the cross-section of a structural
The modulus of elasticity represents stiffness of a material in member is a measure of the resistance to rotation offered by the
the elastic range and under such conditions the numerical value geometry and size of the section. Thus, the moment of inertia is a
of E in tension and compression are essentially .the same. useful property in solving design problems where a bending
13.2.2.Rigidity or Stiffness moment is involved.
The amount of stress that a material can withstand is a 'rhe moment of inertia is needed in solving- any rigidity
measure of its rigidity or stiffness. problem in which the member is a beam or a long column. It is a
13.2.2.1.Poisson's Ratio measure of the stiffness of a beam.
Poisson's ratio is defmed as the ratio ofthe strain in a direction In working with the moment of inertia ofa section, the neutral
perpendicular to the load, to the strain in the direction of load and axis of the section must be located. The neutral axis-represents a
is usually represented by the symbol 1-1 (mu). Therefore, plane of zero strain and therefore zero stress.
=_ lateralstr~ (13 2) There are different methods of determining moment of inertia
1-1 longitudinahtrain .... of a section. but the use of simplified formulas for finding the
The -ve sign indicates that the lateral strain has an opposite moment of inertia of standard sections is the most often used in
sign to the longitudinal strain. design problems. In using these formulas, be sure to take the
Thus, a measure of stiffness of the material in the direction moment of inertia about the correct axis. Notice that the moment
normal to the elastic strain in tension or compression is the of inertia for a rectangle about its I)eutral axis, Fig. 13.3 (a), is,
 / -------- 1
I

894 Welding Metallurgy and Joint Design J Weld Design for Static Loading 895
I-b-j r-b---j where, R = torsional resistance,
T W = width of the flat section,
t = thickness ofthe flat section.
"0

(0)
Neutral axis_.1 TO .
(b)
Axis of
reference
Once the torsional resistance of an open section has been
found, the angular twist may be calculated by the formula :
TL
e = Es . R ...(13.7)
Fig. 13.3. Rectangular section with different axis of reference.
where, Es = modulus of elasticity in shear,
bd3
In = 12
...(13.3) T=Torque
= applied force (P) x moment arm offorce (e) (i.e. the
while the moment of inertia for the same rectangle about its base perpendicular distance from the centre of rotation to
line, Fig. 13.3 (b), is, the line of force)
L = length offla.t section,
Ib = bd!..
3 ...(13.4) R = torsional resistance.
This equation is also valid for a round shaft.
Knowledge of moment of inertia is also required for finding Section Modulus(Z)
the value of the polar moment of inertia (J), and the section Section modulus (Z) is defined as the ratio of moment of inertia
modulus (Z) and thus has a role in solving simple strength design (1) with respect to the neutral axis divided by the distance (c) from
and rigidity design problems .. the neutral axis to the outermost fibre of the section, i.e.,
13.2.5. Polar Moment of Inertia z=-Ic ...(13.8)
The polar moment of inertia (J) equals the sum of any two
moments of inertia about axes at right a~gles to each other. The In an unsymmetrical section, while there is only one value of
polar moment of inertia is taken about an axis which is the moment of inertia with respect to a given neutral axis, the
perpendicular to the plane of the other two axes. Thus, section modulus can have two distinct values corresponding to two
different distances of the extreme fibres from the neutral axis. Two
J = Ix + Iy ... (13.5) different values of section modulus are therefore found in all
13.2.6. Torsional Resistance unsymmetrical sections; the outer face having the greater value
Torsional resistance (R) has largely replaced the less accurate of c will have the lower value of section modulus, Z, and thus the
polar moment of inertia in standard design formula for angular greater stress. Since it has the greater stress, this is the value
needed for design purpose and is the one usually available in
I W I tables of rolled sections.
Section modulus is a measure of the strength of the beam in

1 II
Fig. 13.4. Cross-sectional dimensions of a flat section whose
torsional resistance, R, is t.o be determined.
twist of open sections. It should be employed where formulas have
been developed for the ty'pe of section involved. The torsional
resistance of a flat section, as shown in Fig. 13.4 may be
approximated by the formula:
Wt3
R= 3-
bending. Thus, the sectional property is extremely useful in
engineering calculations and may be derived mathematically for
various shapes of cross-section.
Polar moment of inertia is used in determining the polar.
section modulus (J/c) which is a measure of strength under
torsionalloa'ding of round solid bars and closed tubular shafts.
13.2.7.Radius of Gyration (r)
The radius of gyration (r) is the distance from neutral axis of
a section to an imaginary point at which the whole area of the
...(13.6) section could be assumed to be concentrated but still having the
same moment of inertia. This property of a section is used
 tb oottom .
3Y6+I+ I~
I.
rlc;:I,iI,!Ii 2tb2
iI 2d)
"-"
iIY~
IIII,tb3
itd Il' ~
:r
IIQ
::I
Co
~
-c+d)----------
II .,-l
IIQ
Q :i" 6III
6Ei" iiyy -Table
I I13.1.Properties
,
12 (b, :f of! Thin Sections100 I~(contd.)

:0
36 (b
(b(4b d)jIi Y6IiIII'
bottom top top ::I
::I
IIQ td2
tb2
td2
~(b+3d)
td3
right
!d3 tb2
tm.J
(b
(2b
12(b+6d)
td3
6"(b+4d) 1 Yy+ 2d)
Iside
++4d)
+d)
T I•~(2b
12(b+d)
tb3
td2 (b
(4b+ +d1
"3(2b
i (3b I +d) I12(6b+d)
t1fr2
4dj +d)
3 (b + d)
td2I +d) I

'v'
6" (4b+I+d)
~(b+'6d)
~(b 3d) I .t; I(3b +d) I

'ui' 'm'
I t I
Section I

~
~
c;:
Table 13.1.Properties of Thin Sections (contd.)
'='
Where thickness (t) is small, b = mean width, and d = mean depth of section ~
---. -
'"

down from
"3
I
"3(b+d)
top b+2d
tb2d2
t3 (b+d) I I down
"3 120(b+2d)
2t3(b
t3
b+d
2(b+d) 0t3(2b+d)
(b12
12 from
+(b
+d)(b
2d)
2tb2d2
"3(2b+d)
b2
top
++d)
2d) 12 d2
(b +d)from top
down
- - 0.7071 r
ciQ'
::I
0'
(b +d) Ixy "(d3/3) 0b+d (2b +d)
rx 2 (b +d)~d3(4b+d) ~ 2t1fr3
~d2(3b
~b2(b+3d) b2(b+d2+d) .,
~d2 (6b +d)
~ 6 (2:3+d) d2
I ~ 12 (:2+ d) CIl
.•.
Il'
c:
rl

cIl'
l:""

Q.
Ei'
IIQ

oc
\C
...•
1'1""""" ----- r
!

898 Welding Engineering and Technology Weld Design for Static Loading 899

primarily in solving column problems. It is found by taking the 13.3. SALIENT WELD DESIGN FEATURES.
square root of the moment of inertia (1)divided by the area (A) of Design engineers must accept at the outset that with modern
the section and is expressed in mm. welding equipment in conjunction with modern means of
inspection and testing, the trustworthiness of welds is at par with
Thus, r=~ ... (13.9) that of anyother reliable means of load transmission. However,
The radius of gyration is important because it indicates to a unlike other means of jointing like screws, bolts, or rivets its
certain extent the ability of a section to resist buckling. production is part of the manufacturing/fabrication process in the
It is necessary to use the smaller of the two moment of inertia assembly shop, and it has to be designed for each particular job.
about X-X and :?-Y axes to get radius of gyration used in the In order to decide upon size, shape, and type of a welded joint
it is necessary to know:
(i) the ultimate and permissible safe stresses for the
unsupported lengt~(~-I
slenderness ratio jf for
the the design
column andofr columns;
the radiuswEere L is the
of gyration. various types of welds, and
Table 13.1 lists the properties of regular thin sections while (ii) the load to be transmitted by each weld, in order to
table 13.2 lists the properties of typical irregular thin sections. calculate the stresses in the weld and to determine for
Table 13.2. Properties of Typical Irregular Thin Sections
Where thickness (t) is small, b = mean width, and each particular case the required type and size of weld.
d = mean depth of section It should be realized that it is not only the size and shape of
the weld which determine its mechanical properties, but also the
procedure used to produce it. For example, to obtain good welds
conditions of accessibility must be favourable and an endeavour
I _td2 [kb2 + (k + 1)2 bd3 + d2]
3 k;:5!:.
b should be made to enable the welder to work in the down-hand
I %-
b 'I welding position. Points such as these contribute to the obtaining
I C _ d (b +d) of a weld the mechanical properties of which are as close as
possible to those which can theoretically be expected.
T
L ---j--ft--L
Tt--
d - - - x

Cb Sb=
td [ kb2 + (k + 1) 2bd + d2]

b - d (k + 1). + 2d
b+d 3.Ct = 3b tk (kb + d)
To obtain the desired quality welds the size, shape, and
welding procedure for the various welds must be indicated on the
-I ~~ ~~1- 2 St;: td[kb2+<k+ 1) 2bd+d2]
kb +d 3 3
drawing for a welded construction, and no doubts must exist nor
any decisions be left to the judgement of the worker in the shop.
13.3.1.Arc Welded Joints
Groove welds, with the exception of those used for joining thin
1%;: t~; (k3 _ 3k2 + 3k + 1) + tb;2 plates or those produced by means of deep penetration electrodes,
_mm~ k;:5!:.
d
require plate preparation by either bevelling or gouging. This is
j,
1 f
d - ---
TElt-b~~ S% = t~
2

(k3 _ 3k2 + 3k + 1) + tbd

often an additional operation after cutting a plate to the required
shape. It can be carried out by flame cutting, flame gouging, or
machining. Usually no plate-edge preparation is required for the
\4-c---.l a2 + 2cd +d2 execution of fillet welds. It is obvious, therefore, that in general,
Cb ;: 2 (a + b + e + d) joints using fillet welds are produced more economically than
those using butt welds.
In butt welds the danger of stress concentration is less
n-------~----x
I+--b----.l
I~
T I ;: t (a3 + 3ed2 + d3)
% 3
t (a2 + 2ed + d2f..
4 (a + b + e + d)
pronounced, especially if reinforcements are removed; but it has
been found that, in general, butt welds produce greater residual
stresses. The shape of fillet welds, however, favours the creation
of stress concentrations either at the toe or at the root of the weld.
"
,I,

II,
 Weld Design for Static Loading 901
900 Welding Engineering & Technology

This point is of particular importance if the welds are to be

subjected to dynamic loading .
These· general considerations have to be kept in mind for
..=-
-N

guidance only. The final decision on the most suitable type of weld
must be taken individually for each case, and after considering
the general layout and the loads to be transmitted by the weld.
Fig. 13.5. Throat thickness in a butt weld joining two plates
The permissible working stresses in welds depend upon the of unequal thicknesses.
following factors :
1. the loading conditions i.e. static, dynamic, alternating, equal to the thickness t2 of the thinner plate as shown in Fig. 13.5.
pulsating, etc. In such cases the taper of the thicker plate should have a slope of
2. the type of weld i.e. groove weld, fillet weld, etc.
3. the shape of the weld i.e. convex or concave weld.
4. the surface finish of the weld i.e. as welded, ground, or
machined, etc.
1 in at least 5. For stress calculations of 45° fillet welds with the
leg length l, the theoretical value of the throat thickness, or
0.707 l, Fig. 13.6, is used. In actual practice the throat thickness
*
of fillet welds has to be kept within the limits of 0.6l and 0.9l.
5. the method of producing the weld, and the nature of any
treatment applied afterwards, i.e. welding procedure,
post weld heat treatment, peening, etc.
Leg length
Although in the case of static loading the ultimate tensile
stresses of the weld metal are close to those found for the
corresponding parent metal, conditions are different in the case of
dynamic or alternating stresses.
It has been observed by photo-elastic investigations that stress
concentrations do occur in welds, for example,. at the heel and toe
of fillet welds, where stress may be high even when the load is
relatively small. As the load increases, the yield point will soon be
reached at the point of stress concent~ation, where slight yielding
will occur, thus transferring the maximum stress to another part
of the weld. This process continues until the stress distribution
over the whole load-carrying section is fairly uniform. Such
Leg length, I Throat thickness
balancing of the stress distribution takes place under static loads, = 0.707 I
and for the purpose of stress concentrations, does not considerably
Fig. 13.6. Throat of a fillet weld in aT-joint.
affect the final stress in the weld. In the case of alternating
stresses, however, sufficient time is not available for redistribution The reinforcement of a butt/groove weld and the additional
and ultimate balancing of stresses, and thus fractures occur under thickness created by the sealing run, Fig. 13.7, should not be taken
lower loads. into account when calculating the stress-carrying section of the
13.3.2.Stress Calculations for Welds weld, because the reinforcement is zero at the side of the weld,
and the sharp corner at this point might even produce stress
The critical dimension with regard to the strength of a fillet concentrations. In fact, the most perfect groove welded butt joint
or a butt weld is the throat thickness. This is the minimum is the one which is produced by depositing a reinforcement and a
thickness of the weld on a straight line passing through its root. sealing run and afterwards machining these down to the thickness
The throat thickness 'h' of a butt joint (with a groove weld) of the plate. The minimum height of reinforcement in the centre
connecting two plates of unequal thickness t1 and t2 is, therefore, of the weld should be :
902 Welding Engineering & Technology Weld Design for Static Loading 903

In the case of welds generally used in engineering structures,

the lack of uniform stress is usually taken into account by the
values of the recommended permissible stresses as shown in Table
13.4 for fillet welds in steels.
Table 13.4. Allowable Unit Loads on Steel Fillet Welds
1.6
3.2
4.8
8.0
6.4
11.1
12.7
15.9
22.2
25.4
19.020.10
18.25
12.80
14.60
585.25
6.40
21.90
19.15
27.40
9.5 36.50
32.00
32.65
49.35
43.90
76.75
85.25
115.05
138.40
Weld 16.50
0.95
16.45
10.05
30.20
8.20
27.40
Size.50
116.90
9.15
24.65
7.30
32.85
21.90
40.15
45.70
25.55
29.20
41.15
50.25
32.85
62.30
43.80
38.40
54.80
44.80
51.10
80.40
73.10
65.75
58.40
51.10
65.80
76.75
109.70
102.30
127.95
146.15
98.65
120.60
89.50
140.60
131.50
153.45
100.45
63.95
91.45
82.20
109.60
-
21.90
10.95
63.95
70.35
57.55
87.70
175.35
160.75
480
690
620
550
825
760
5~.85
415 Strength level of weld metal, MPa
(mm) S. Allowable Unit loads, MPa
5.
11.
12.
10.
3.
4.
6.
7.
8.
9.
2. 1. No.
Fig. 13.7. A groove welded butt joint with
reinforcement and sealing run deposit.
(i) up to and including 32 mm groove weld size-not less
than 10% of the weld size,
(ii) above 32 mm groove weld size-not less than 3 mm.
The throat section which carries the maximum stress is
determined by the throat thickness and the effective length, L, of
the weld.
Whereas the throat. sectional area of a fillet weld is the product
of its throat thickness and effective length~ the amount of
deposited weld metal grows with the square of the throat thickness
and only with the first power of the length. For equal throat
sectional area, therefore, long weldF'with small throat thickness
require less weld metal than short welds with large throat
thickness. Consequently former are more economical than the
latter. 13.4. USEFUL FORMULAS FOR DESIGN UNDER
A certain minimum size of a fillet weld is, however, required DIFFERENT TYPES OF LOADING
for each material thickness, in order to avoid cracking in the weld Some useful mathematical relationships required in
due to differential contraction between weld and parent metal, and calculation of stress, etc. under different types of loading are
usually lies between 3 and 8 mm for the most often used plate included in this section.
sections as shown in Table 13.3.
13.4.1. Tensile Loading

l
Table 13.3. Recommended Minimum Fillet Weld Sizes for Steel
The simplest type of loading on a member is tension. A tensile
S. Section thickness load applied axially in line with the centre of gravity of the section
No. (mm) Minimum (mm)
fillet weld size will result in tensile stress given by,
1. 3~ 3 p
2. 6-13 5 crt =A ...(13.10)
3. 13-19 6 where, P = tensile force applied to the member,
4. over 19 8 A = area of cross-section at right angles to line of force,
crt = unit tensile stress.
Although the stress distribution in welds is very rarely
uniform and not as simple and accessible to accurate calculation The strain of the member under tension is given by,
as is the case in the parent metal, experiments have shown that, crt
...U3.11)
for the sake of design calculations, uniform stress distribution is E= E
seriously influenced by conditions of de~ignor loading.
 904 Welding Engineering & Technology Weld Design for Static Loading 905
:'>.

where, e = tensile strain,

crt = unit tensile stress
E = modulus of elasticity.
The total elongation or displacement is given by,
x---- - rx =
/ I;
Elongation = e . L ...(13.12) .
where L = length ofthe member.
13.4.2. Compression Loading
y
Very few members fail by crushing, or exceeding their ultimate
compression strength. However, a compression force requires l:ry
designing against buckling. If a compression member such as the I
column in Fig. 13.8(a) is loaded through its e.g.,the resulting stresses
are simple axial stresses. But because of its slenderness (measured
by the ratio of its unsupported length, L, to its least radius of
gyration, r) the column will start to move laterally at a stress lower
III' I ry = A
IflY
than its yield strength, as shown in Fig. 13.8(b). This results in I
buckling of column and the development of bending moment which I
leads to final failure of column as the loading persists.
P P, + P2
I
Y

~ Fig. 13.9. Radius of gyration' about the

f.
X-X and Y-Y axes of a column with I-section.
I'I
II
The design of compression members or columns is by trial and
I:
error and table 13.5 lists the formulas employed for their design.
Table 13.5. Allowable Compressive Stress
I:
(II Range of Llr Average allowable compressive unit stress
I'I values (oj
III
~
OtoCc
P PI + P2 O=[,_(~)
. 2Cc2 ]~
FS
Ie l Vertical (bl Buckled
column column 0= 149 X 106
Cc to 200
Fig. 13.8. Compressive loading: (a) straight column with concentric load,
and (b) with increased load, the column tends to move laterally. (~J
In design of columns, area of cross-section and radius of
gyration are important factors. The former decides the
[I~ compressive load that the column can carry and the latter where, Cc = -V.2rc2E
~
Vy
5 3 (~)
and FS =_+ 8Cc
3 --~
(KL
r
8Cc3
J
'Ii indicates, to a certain extent, the ability of the section to resist
'1'[1 buckling. If a section has two different radii of gyration, as for Note: 1. For very short columns this provides a factor of safety 1.67; for longer
columns this gradually increases (by 15% max,) up to 1.92.
,I,
I-section shown in Fig. 13.9, it is necessary to use the lower value
to .account for the worst condition. 2. FS = Factor of safety; Oy = Yield strength and K = effective length factor.
!ll:
~lili'i
 906 Welding Engineering & Technology Weld Design for Static Loading 907

13.4.3. Bending Loading To satisfy strength and stiffness requirements beams should
In theory of bending of beam, be deep.
M E ax
13.4.4.Shear Loading
-y =R = Y- ... (13.13) Fig. 13.10 illustrates the shear forces in the web of a beam
under load. They are both horizontal and vertical and create
where, M = external bending moment at the section, Mxz' under
consideration, Load

ax = bending stress at a perpendicular distance y from the

neutral surface,
R = radius of curvature of the beam i.e. the radius of
curvature of the neutral plane * of the beam", Vertical
E = Young's modulus ofthe beam material. shear
1
Most often in equation CI3.13) ax is replaced by fb or f to
indicate bending stress. Therefore, this equation becomes,
Reaction
M _t_E ...(13.14) at support
Horizontal
I-y-R shear

Equation (13.14) is the well known Bending or Flexture

~~~
Equation and can be used for finding the bending stresses in the
beam, and the deformation of beam. It is obvious that the bending
1f- --I ~
str€ss, ax, is zero at the neutral axis (an imaginary line passing
through the centroid of the section and parallel to the axis of the
beam) and maximum at a point in the section that has a maximum
L -=-=--:~ ~~ ---
distance from the neutral axis of the surface. Diagonal Diagonal
tension compression
For design of beams, sometimes only Ymax is considered in Fig. 13.10. Sheat forces in the web of a beam under load.
equation 13.14, giving fmax and the formula is written as, diagonal tension and diagonal compression. Tension is not a
M problem, since there is hardly any chance of the tensile stress
z· =fmax ...(13.15) reaching a value high enough to cause failure. If the diagonal
compression reaches a high enough value, however, the web could
where, Z = -L__
is the section modulus. buckle as shown in Fig. 13.11, especially ifit is very deep or very
Ymax thin.
13.4.3.1. Strength and Stiffness of Beams
For the beams,
Strength is proportional to Z
and t'ff: .
eLoad
S I ness z.e. -D-flect'Ion oc: EI Web buckles

'" Neutra.l Plane of a beam is defined as a plane whose length remains

un~hflnged during the deformation.
Fig. 13.11. Mode of initial failure of beam in compression flange.
 908 Welding Engineering & Technology Weld Design for Static Loading 909
a
where, R == torsional resistance,

I
W == width ofthe flat section,
Flange t == thickness of the flat section.
Once the torsional resistance of an open section is found, the
y
angular twist may be calculated by the formula,
, = Vay
In
_--L T·L
e == E s . R ...(13.18)
where, e == total angular twist, radians,
Flange
T == torque, m-kg;
Es == modulus of elasticity in shear, MPa,
Fig. 13.12. Shear force on the fillet welds R == torsional resistance, mm4.
connecting the flanges of a beam to the web. Each part of an open section will twist the same angle as the
The unit shear force on the fillet welds joining flanges of the whole member. The unit angular twist, <1>, is equal to the total
beam, Fig. 13.12, to the web can be calculated by the following angular twist, e, divided by the length, L, of the member.
equation: e
<l> == L ...(13.19)
V·a·'\!
(==_-=--..L
['n Knowing the unit angular twist, it is possible with the
where, V == external shear force on the member at the desired following relationship to find the resulting shear stress on the
cross-section, surface of the part.
T· t
a == area held by the connecting welds,
y == distance between the e.g. of the area held by the welds 't == t . Es = <l>

where t is the thickness of the section.

• R (13.20) ...
and the neutral axis of the whole section,
1== moment of inertia of the whole section about the If torsion is a problem, closed tubular sections .or diagonal
neutral axis, bracings, as shown in Fig. 13.13, should be used. Closed tubular
n == number of welds used to hold this area.
To satisfy the strength and stiffness requirements the shape
of the compression flange of a beam should be such that it will not
buckle easily. So, it should not be too wide and too thin.
13.4.5. Torsional Loading
Torsion creates problems in the design of bases and frames.
The torsional resistance of an open section is very poor. The
torsional resistance of a built-up section is approximately equal to
the sum of the torsional resistances of the individual flat parts
that make up the member. The torsional resistance of a flat
section, as shown in Fig. 13.4, may be approximated by the
following formula:
(a) Closed box section (b) Box section with diagonal

R == 3
Wt3
...
(13.17)
bracing
Fig. 13.13. Methods of resisting torsion: (a) use of closed box section, and
(b) use, of diagonal bracing.
~ r

910 I 911

sections can easily be made from existing open chan.nel or

Welding Engineering & Technology
! Weld Design for Static Loading

into convenient lengths. The ratio of these individual lengths

I-sections by intermittently welding flat plate to the toes of the divided by their corresponding thicknesses is determined and
rolled sections, thus closing them in. This will increase the summed up. Torsional resistance, R, is then obtained by the
torsional resistance several hundred times. An existing frame may following equation,
be stiffened for torsion by welding in cross bracing at 45° to the 4A2
a.xis of the frame.times.
This also increases the torsionaf resistance by R = -d- ...(13.21)
several hundred
I,--!
ts
Fig. 13.15 gives formulas for calculating the torsional
resistance of three commonly used closed tubular sections.
L~ts 13.4.6. Shock Loading
With static or dead load, Wd, the deflection (0) of a beam is
given by,
Fig. 13.14. A general closed tubular section.
0= Wd ·Z3
...(13.22)
The torsional resistance of any closed tubular shape, Fig.. where, Z = length of beam,
13.14, can be determined by drawing a dotted line through the E = modulus of elasticity of beam material,
mid thickness all the w['v around the section. The area enclosed I = moment of inertia of beam section.
by the dotted line, or mean dimension, is A. Divide this section bd3
...(13.23)
J--
I I b ---II l:A
R :: 4 A2 :: 4 (bd)2 :: JL+JL
~+~ 2 b2 d 2
Now, 1= 12
... from (13.22) and (13.23), we get,
-, ts tb td tb td Wd ·Z3
...(13.24)
d Stress at ~ of b, 0= 4E.b.d3
=-L T Since volume V of the beam with depth d = bdZ ...(13.25)
T -- - __T
b - 2Mb - 2 bdtb
... from (13,24) and (13.25), we get,

R ::2JTr3t
0=--
Wd
4EVd2
·Z4
...(13.26)

...(13.27)
Hence, stiffness, W0d = 4E . V (dJ2
Z2

. Wd·Z
Now, bendIng stress, fb=~
...(13.28)
t-- b I
where, section modulus, Z =
bd2
(5 ...(13.29)

'l./././////////~=r 2tb 2d 2 ... from (13.28) and (13.29), we get,

t d =b+d 3 Wd·Z
R ...(13.30)
-L fb = 2bd2
Thus, from (13.30) and (13.25), we get,
Fig. 13.15. Torsional resistance, R, of three popular
closed tubular sections. 3 Wd ·Z2
fb= 2V~ ...(13.31)
 913
Weld Design for Static Loading
912 Welding Engineering & Technology
13.5. SOLVED EXAMPLES
2 d
I.e., ...(13.32) Some typical problems of weld joint design for static loading
Wd = "3-fb' V· l2 conditions are presented as solved examples in this section.
Now, if shock energy U is applied to the beam, causing a Problem 13.1. Determine the weld throat (twY in the case of
deflection 0, the equivalent static load to cause the same deflection fillet and groove welds for the dimensions of the joint given in Fig.
is given by Wd, 13.16. Compare the amount of metal deposited in each case and
the load carrying efficiencies of the two Joints for P = 100 KN and
where, _d_=
W·o U ...(13.33) Q = 70 KN loads considered separately. Allowable tensile and shear
2
2U stresses are 175 N/mm2 and 125 N/mm2 respectively.
.or
Wd=T ...(13.34)
In comparing the shock loading of similar structures of say p
mild steel and cast iron, though the permissible stress for the 50
former is higher than for the latter, its higher modulus of elasticity,
and consequent smaller deformation for a given load, will to some ----
a
extent offset its initial advantage. The permissible shock load U
for the beam can be found from equations (13.32) and (13.34).
From (13.32) and (13.34), we get,
2U 2 d
T ="3fb . V· l2 ... (13.35)
fb d
U=3.0.V'l2 ...(13.36) (0 I Fillet welds
I b) Groove welds

Thus, the permissible shock load is greater, the greater the Fig. 13.16. Two types of welds for aT-joint:
permissible stress and greater the deflection of the beam under (a) fillet welds, and (b) groove weld.
the load.
Solution.
The importance 'of the factor l~, which Koenigsberger
. called (A)For tensile load, P = 100KN
the degree of compactness can be seen from equations (13.32) and (i) Fillet weld Design (Fig. 13.17a)
W A fillet weld is assumed to fail in shear along the weld throat.
(13.36). The stiffness Od depends upon a material factor (modulus Weld throat area, Aw = weld throat X weld length x no. of welds
of elasticity), the volume of material employed, and the square of =tw' L· n
the degree of compactness. The strength i.e. the maximum
permissible load determined by the permissible bending stress, = tw x 50 x 2 = 100 tw ••• (i)
depends upon a material factor (the permissible maximum Max. allowable shear stress, fa = 125 N /mm2
bending stress), the volume of material employed and the degree :. Max. force on the weld throat =Aw . fs
of compactness.
= 100 tw x 125 ...(ii)
It should not be overlooked that stiffness is not always the = maximum tensile load
decisive factor in a structure. In cases where shock loads are likely
to occur, high stiffness against deformation may be dangerous, as 100 x tw x 125 = 100,000
the energy of the shock has then to be absorbed over a short _ 100,000 -8mm
_ ...(iii)
distance, and the stressing of the structure must be or tw-.d
correspondingly high .
 914

... Weld leg length = tw x -f2 :::8 x -f2

Welding, Engineering & Technology
r Weld Design for Static Loading 915

:::11.3 mm::: 12 mm (say) (ii) Groove Weld Design

... (iv)
1 r
Stress in groove weld under shear load, Q
Volume of the weld metal::: 2" x tw x hyp x L x n
! :::~wQ 70,000
= 12 x 50 = 116 .7 N/ mm 2
... ( Xl ')
8 x --I~~ x 50 x 2
::: ~ X
Allowable shear stress::: 125 N/mm2
:::400 x -vT44+ 144 :::400 \lr~r88
3 So the weld is safe.
:::6788 mm ... (v)
(ii) Groove weld Design (Fig. 13.17b) Volume of the weld metal is still equal to 3536 mm3 as in the
first case i.e. for tensile loading.
S ld P 100,000 2
tress on we :::t~:::
w 12 x 50 :::166.7 N/mm ... (vi) Thus, for shear loading tlllet weld is superior and has 10.4%
Allowable tensile stress :::175 N/mm2
Hence the weld is safe. ,---3536
(3536 - 3168 x 100 ) lesser weld metal.
6U 2 Ans : Choose fillet weld for shear loadjng and groove weld for
Volume of weld metal
:::360 x nr x n x L tensile loading under the given loading conditions.
50 22 2 Problem 13.2. Design a double fillet weld for lead P :::8 KN
::: -360 X --7 x (6 + 3) x 2 x 50
applied as shown in Fig. 13.17. Take maximum allowable shear
50 22 stress in weld as 115 N/mm2.
::: -- X -- x 81 x 100
360 7
:::3536 mm3 ... (vii)
Comparing (v) and (vii) it is evident that the fillet weld has
almost 92% 3536 x 100 extra weld metal. The fillet weld
leg length is\.as large as the pJate thickness and is really not ~
(6788 - 3536
good design ill this case. Thus, for the tensile load the groove weld
is a more suitable weld design.
(B) For Shear Load Q = 70KN
(i) Fillet Weld Design
Fig. 13.17. A fillet welded cantilever. /
tw' n· fs == Q
~.e"
Solution. This system of loading will result ill giving a
tw x 50 x 2 x 125 :::'70,000 combination of axial load and shear force.
or 70,000
100 x 125 :::5.6 mm
tw::: ...(viii) Whilst the maximum bending stress occurs at the extreme
fibre where the stress is zero, the maximum shear stress occurs
.. Leg length, lw ::: 5.6 x -{2 ::: 8 mm ...(ix) at the neutral axis and is, according to Timoshenko', 50% greater
than the average shear stress.
Vol. of the weld metal :::~ X tw X (",[z'Z+zT) . L . It

::: ~ X 5.6 (..J8'2+SX) x 50 x 2

.. Bending stress, fb ;l
max ::: w ...(i)
an d s hI"ear stress, Is
3 A'
max::: 2" P
w
..
..·eu)
::: ~ X 5.6x -{128 x 100::: 3168 mm3 ... (x)

• Timoshenko - Strength of Materials. D. Van Nostrand Co., New York.

 917
Weld Design for Static Loading
916 Welding Engineering & Technology
\= 120 "';12+ 42 = 495
. Ixx tw·L3/12 , tw tw
Smce, Zw = L/2 X no. of welds = _.- X 2 495
t ·L3 t ·L2 t.e., tw = 115 = 4.30 mm
=_w--x2x2=_w-- :. leg length = ...f2 . tw = ...f2 x 4.30
12· L 3
and . Aw = tw . L . n = tw . L . 2 = 2 . tw . L = 6.1 mm = 7 mm (say)
Ans : tw =. 4.30
mm
tw X 502 .
lw = 7.00 mm
:. Zw = 3 ...(m) and Aw = 2· tw . 50 = 100 tw (w)
From (i) and (iii), we get, Problem 13.3. Design a double-sided fillet weld for a lever
mounted on shaft with dimensions and loadings as shown in Fig.
I' max _
Ib Z-
- Plw _ 8000 x 100 _ 960
tw' (50) 2 - w -t- ...(v ) 13.18. Take maximum allowable stress as 115 N/mm2•
3
and from (ii) and (iv), we get,
3 P 3 8000 120 .
t 40KN
fs max='2' T='2'
w 2. t w .50 =-t-w ... (Vt)

The maximum shear stress occurs in the centre of the weld

while the maximum bending stress occurs at the ends of the welds.
From (v) and (vi) it can be seen that the bending stress is greater ,
than
fb
the shear stress if 1 > ~. As L is at least 6 leg lengths,
> fs if 1 > 1.5 leg length. So shear stress need not be considered.
j--SOO ---l
Fig. 13.18. A laver mounted on a shaft with double fillet weld.
Thus, fb = Allowable shear stress Solution. Torque on shaft due to a force of 40KN
= F x lever arm = 40 x 500 = 20 KN-m ... (i)
( '.' weld will eventually fail in shear)
i.e. from (v), we get, The welded joint transfers this torque to the shaft.
960 = 115 Let us take weld length (on first approximation)
tw = circumference of shaft
. 960 835 = 7t . d = 7t X 80 mm
I.e., tw = 115 =. mm 80
or leg length = ...f2 . tw = ...f2 . 8.35 = 12 mm Torque arm =r=-=40mm 2
Now 12 mm leg length is too large therefore actually we may Let weld throat = tw
consider combined effect of stress. Allowable shear stress, fs = 115 N/mm2
:. Total allowable torque for double fillet joint
2 fb max = fs x weld length x weld throat x torque arm
t.e., f max = ~ fs max + ( 2 l2 x no. of welds
2 2
= fs . 7td . tw • r . 2
or 80
tw x 2)
+ ( tw960
, 115 = '\} (120J = 115 X (7t x 80) x tw x "2 x 2 N-mm
= 10 "';(12)2+ (48)2 = 2.313 tw KN-m .... (ii)
. tw
1"""'""-------------------------------------.-------
918 Welding Engineering & Technology
Weld Design for Static Loading 919
Equating (i) and (ii), we get,
2.313 tw = 20
1
= -- x 13 x 13 x 1td
2
t.e . 20
= -2.313 = 8.65 mm
tw = 21246mm3 ... (iv)
.'. weld leg length = tw x --!2 = 8.65 -{2 = 12.23 mm Similarly, weld metal in second case
= 13 mm (say) 1
= -- x 10 x 10 x 1t X 80
Now, we can revise the calculations by taking the weld length 2
measured along the centre of the weld throat, as shown in Fig. = 12571 mm3 ... (v)
13.19.
Percentage reduction of weld metal in second case
8·65 mm
= 21246 - 12571 x 100 = 40.83%
&--- ---- ----- -- ---~
I'• 80 8.65{2 mm
+80mm
Fig. 13.19. Revised weld length.
-II
----1
Problem 13.4. Design a fillet weld for a connection between a
hollow rectangular section and a plate as shown in Fig. 13.20. Take
maximum allowable stress as 160 N/mm2•
As tw = 8.65 mm
.. dia. along centreline of weld throat
80 8.65 2
= +~x
= 80 + 12.23 = 92.23 mm
Now, equating applied torque to allowable torque, we get, T} ,,,quo
S KNm
1t . d' . r' . (2tw) . a = 20 x 106
100 x SO
20 X 106 86 49 N/ 2 ...

a =
1t . 92.23·
J
-2- .2 . (8.65) =.
[9223 mm ...(m)
Fig. 13.20. A hollow section fillet welded aU around on·to a plate.
Solution. This problem may be solved by two different
For an allowable shear stress of 115 N/mm2, tw = 8.65 mm methods .
.. for a stress of 86.49 N/mm2, First Method
x 86.49 = 6.51 mm
tw = ~.;: A simple method is to assume that the shear load is taken on
the vertical i.e. web welds and the moment is taken on the
= 7 mm (say)
horizontal or flange welds. Then,
.. Weld leg length = tw x -{2 = 6.51 x -{2 = 9.2 mm
For the Web Welds
= 10 mm (say)
fs x throat area ofwelcl x no. of welds = shear force
We could then call for 10 mm leg length fillet welds which will
b€ a considerable saving in weld metal compared to a leg length l.e., fs x twl X LI x n = 12 KN ... (i)
pf 13 mm. or 160 x twl x 100 x 2 = 12,000
Weld Metal Reduction
.. twl = 160"12,000.
• "" .. n 0372
• mm ." ("u)
Weld in first case = ~ x altitude offilletweld x base of fillet .. twl = 0.37 mm
x weld length The magnitude of twl is too small for practical welding.
 921
92~ Welding Engineering and Technology Weld Design for Static Loading

For Flange Welds . I 416667

SectlOn modulus, -, - = 100 = 8333 tw
Z = Ymax
As these welds are assumed to take the moment, therefore,
Max. allowable stress on effective weld throat area x moment
2
arm = Bending moment So, if the shear stress in the weld is not to exceed 160N/mm2,
i.e., O'b x (tw2' L2) x q= 5 KN-m ...(iii) Z· 0' = 5 KN-m
Now, as the lower flange weld is in compression thus only the 8333 tw x 160 = 5 x 106
upper flange weld is to take up the total moment, 5 X 106 .,..
., tw = nnnn .• ~"'f"\ = 3./5 mm
.. L2=50mm
q = 100 mm(distance between top and bottom flange welds) This is by ignoring the shear load.
O'b = 160 N/mm2 If shear load is taken into account
Putting these values in (iii), we get, . 5 x 106 600 ... (i)
Stress due to bendmg = 8333 x tw - ~
160 x ·50) x 100 = 5 X 103 X 103 N-mm
(tw2
12KN
5 x 103 X 103 Stress due to shear
(200tw + 100tw)
.. tw2 = 160 x 50 x 100 = 6.25 mm
_ 12,000 _ 40 ...(ii)
.'. Flange weld leg length = tw2 . -.f2 - 300tw - tw
= 6.25 -.f2 = 8.34 mm
Total shear stress + (ii)2
= .."f(i)2
In practice we would use 10mm leg length fillet weld all
around the joint.
Second Method ~"(6~OJ
601.33
+(~~j
A more refined calculation takes into account the effective = ...(iii)
bending modulus of the weld. The moment of inertia (1) is tw
calculated using weld throat as the section thickness as shown Now, max. allowable shear-stress
below in tabular form . = 160 N/mm2 ... (iv)

Weld section50
(mm)
Ymax
100
--
.2·(mm)4
tw tw'
Ay2 Weld
200tw
100tw
.] (50)2
(mm)4
-
Weld
1003 configuration .. From (Ui) and (iv), we get,
00 tw (mm)2
= 166667·12tw
ange weld 601.33 = 160 . t = 601.33 376
tw .. w ~ ~~ . mm
:. weld leg length = tw . -.f2
= 3. 76-.f2= 5.32 mm
= 6 mm (say)
As is evident the shear load makes no difference to the weld
size. By using the weld section modulus concept the weld leg
length obtained is 6 mm compared with 10mm by the first method.
Problem 13.5. Calculate th~ safe value of load P, for the
cantilever structure shown in Fig. 13.21. Maximum allowable
stress on the weld is 160 N / mm2• Take L1 = 500 mm,
, Total 1= 2,50,000 tw + 166667 tw = 416667 tw mm4 L2 = 100 mm, l = 100 mm, and t = 20 mm.
 922 Weld Design for Static Loading 923
Welding Engineering & Technology
As the maximum value of f.l occurs at the neutral axis where
fbis zero, the maximum combined bending and shear stress will
usually occur at the extreme fibre where fb and fs2 have their
maximum values. The combined stress at the extreme fibre is
given by,
j:
Imax -2" 1"
= fbmax + 2" f?
bmax + 4· t,2
s2max
... (viii)

where, fbmax = fb1 + fs1

=--+ l'
1'£1 1.5-
[from (v) and (vi)]
Zw Aw
= l' x t/650q, + 1.5J:.-
[2 [ .t
Fig. 13.21. An angular bracket butt-welded to a plate:
(a) by singlt! Vee groove weld, and (6) by double Vee groove weld. = § x 500 x l' + 1 5 ----.E
(100)2 X 20 . 100 x 20
Solution. (a) For Fig. 13.22(a)
t x [3 = (302000
+ 1.5)1'= 31.5 l'
2000 ... (ix)
---- 2
S ec t·IOn modu Ius Z = ---12 = ----
t .[ ... (i) and 1'£2
(from
, l/2'w 6 fs2max = -22"
l t (3l + 1.8t) vii)
Area across the weld throat,
Aw =It ... (ii) = l' x 100 (3 x 100 + 1.8 x 20)
(100)2 x (20)"
(b) For Fig. 13.22(b) _ l' x (300 + 36) 16.81' ... (x)
(t1 + t2) ·l3/12 [2. t ... - 2000 x 20 - 2000
ZW=--U2'-=-6 (m) From (viii), (ix) and (x), we get,
and Aw=l·(t1+t2)=[·t (iv)
Thus, from design point of view there is no difference whether fmax = -2- + 1-Mfj2----4iz
fbmax
+
2" 'I'b-max Is2max
it is a single vee groove weld or a double vee groove weld.
Now, r31.51') -------
1'L1 61'£1
Max. bending stress, fb mox = -Z--
w = -2-
l t
... (v) 160 = \2000J
2 1 -." rr.;;(31.5PJ2---~-
+ 2" I ~OO + x (16.8)21'2
(2000)2
and M~x. stress due to direct shear, - 2000 2000
3 P
fs 1max = 2" . X---
w _ 15.75 P + 0.51' x 46.06
- 2000 2000
31'
_ (15'75j1' + 0.51' "';992.25+ 1128.96
=2Ti ... (vi) =~8.781'
2000
Max. torsional shear stress,
l' = 2000 x 160 = 8252
1'L? • 38.78
fs2max = -2-2
l . t (31' + 1.8t) ... (vii)
= 8.25 KN = 8 KN (say)
---"-Refer to-'Design---CorWelding' by F. Koenigsberger, Longman Green & Co.,
Ans.: 8KN
London, 1948.
 r
I
i
I

924 Welding- Metallurgy and Joint Design

Weld Design for Static Loading 925
Problem 13.6. A bracket supporting an eccentric load of 3 2
100KN is to be fillet welded to the flange of a vertical column, as = (125)
shown in Fig. 13.22; determine the weld leg length if the ultimate 3 _. (625)
500 + ~ 12 x (1000)
tensile strength of the deposited weld metal is 420 N/mm2. = 813802.06 + 5208333.2 = 6022135 mm3 ... (iv)
Now, Horizontal component of twisting,
250 d
mm 100K N T·-
H-~2~ _~,=
fh=--Jw
2
...
(v)

r~ 1 Max.combined
forces where, T=P·L
(.,' L = 250

[j--
.
250;x - CG x = 100,000 x 250 mm)
~y d/2 = 25 x 106 (N-mm) ... (vi)
.,---f
mil ~'''1 : t,"-!"
... From (v), (vi) and (iv), we get,

,
~. '." .
fv fh-_ (25 x_M_~
106).__
(2~0)_ -531.9N/mm ...(vii)

~fVfr~fs T·C
Fig. 13.22. A bracket welded to a column face by a fillet weld. Now, vertical component of twisting, fll = _T
w
yr
Solution. The point of maximum combined unit forces is at
right ends of the top and bottom horizontal welds. t.e., _ (25 X 106) x 93.75 = 389.19 N/mm ... (viii)
fll - 6.0221 X 106
The twisting force caused by the eccentric loading is divided Load
into horizontal (fh) and vertical (f) components. The distance from' Also, vertical shear force fs = w~ldlength
c.g. to the point of combined stress, Cyro for this shape of connection 100,000 = 200 N/mm
is given by the following relationship. t.e., ... (ix)
fs = (125 + 250 + 125)
b~+~ . Now,
Cyr = nl.. , J •••(t)
Resultant force, fr = fJ1 + (fll + fs)2]1/2 ...
(x)
From Fig. 13.22 • b = 125 mm, d = 250 mm Putting the va1ues of fh, fll and fs from (vii), (viii) and (ix)
Putting these values in (i), we get, respectively in (x), we get,
125 (125 + 250) .. fr = [(531.9)2 + (389.19 + 200)2}]112
= (2 x 125) + 250 = 93.75
Cyr ... (u)
= [282922.33 + (589.19)2]112
Now, polar moment of inertia (J w) of weld section = [282922 + 347144.85]1/2
_b3(b+2d) d26b d) (OO.
= (630067.18)112 = 793.n N/mm ...(xi)
- 3 (2b + d) + 12 ( + ... m) The allowable shear stress on the effective area of weld metal
Putting the values of band d in (iii), we get, having D.T.S. of 420 N/mm2 is as follows:
't = 0.3 x 420 =126 N/mm2 ... (xii)
J w = (125)3
3
. .£125+ (2 x 250)] (250)2(6
(2 x 125 + 250) + 12 x
125
+
250)
Now, effective weld throat, tw = fr't ... (xiii)
926 Welding Engineering & Technology Weld Design for Static Loading
927
Putting the values of fr and 't from (xi) and (xii) respectively in 450.000
(xiii), we get, tw = 904 ~ 95 = 5.263 mm
" weld leg length,
t = fr. = 793.77 == 630 mm
w 't 126 . l[ == tw . '-12 = 5.263 x ...J2 == 7.44 mm
Assuming an equal leg length fillet weld all around, the leg =8mm (say)
length (l[) of the weld is,
(b) To check the length of weld (C) within the connection x-x,
l[ =:: tw . ...[2- = 6.3-,f2 == 8.95 mm (Fig. 13.23b)ofrequired
the flanges force. the force of the web (C) out into
to transfer
(C) as shear
=9mm (say)
Ana : Fillet weld leg length == 9 mm A
Problem 13.7. Determine the weld sizes on tf1:e truss Force in web, Fw =F x A; ...(i)
connection shown in Fig. 13.23. Take maximum allowable shear where, Aw = Web cross-section
stress on weld metal as 95 Nlmm2, and At == Total cross-section of I-section

©
H- 1_450 KN

III
A373 steel
E60 welds
Now, F == 450 KN; Aw = 1180 mm2 andAt = 4755 mm2
Putting these values in (i), we get,
-~I~
'II
112x152 Fw = 450,000 x ~;:~ == 11672 N ...(ii)
III Shear force,
ff I;
F Fw/no.ofwelds
~ 2x150:U2X150-t 't'==-== _

®~
j 2x300~
I ~
a
tt
i
;!

•
~~~~~~~~~=:~~=:=~=:=~---
____
I,
"":1:
III.

:.=-r
r
.
~ \
~
.
~ 2x90

_
Aw

we b tho
Aw

Fw/no. of welds
,IC k ness x Required
. overlapping length of
.

-1-'"1- - - - - - - - - - - - -- - •..., - - 111672/2 I-sectIOn by gusset plates

I : : I - 8xL
: :- 300 -~ :
: :.e.. .a: : 111672
L = 8 x 2 x 95 = 73.47 mm
:::75 mm (say)
< 150 mm actually provided., So Safe.

of (A) To find
(c) and the plates(B).
gusset required size of the fillet welds between flanges
(0) Front view (b) Side view
Fig. 13.23. Details of fillet wl~lded truss connections. Total length of connecting weld is, L2,
Solution.
= 4 x 90 + 2 x 300 = 360 + 600 = 960 mm
To find the required size of fillet welds between member
(a) Force on weld = 450,000 ~
(C) and connecting plates (B) == tw2 X L2 X is I

Total length of weld, L :::4 x 150 + 2 x 152 tw2 x 960 x 95 :::450,000/

== 600 + 304 == 904 mm or 450,000
Now, force on weld = L x tw x fs = 450,000 N tw2 == 960 x 95 == 4.93 mm
928 Welding Engineering & Technology 929
Weld Design for Static Loading

.'. weld leg length := tw2 x ...J2 = 4.93...J2 =

6.97 mm P·b
Now, ...(ii)
7mm (say)
= PA=-a-+-b
Ans : Though welds .oftwo different leg lengths i.e. 8 mm and P·a
and PB=-a-+--b
...(iii)
7 mm may be used but for the sake of uniformity all welds may
be made with leg lengths of 8 mm. Taking moments about e.g. we get,
Problem 13.8. Prove that in a welded attachment of PA·a=PB·b ... (iv)
unsymmetrical section the lengths of fillet welds vary indirectly to
their distances from the centroid (e.g.) of the section, Fig. 13.24. :. Shear stress, 'tA on weld A
_ PA
- shear area
p. b 1
Fillet weld £.e., 'tA = (a + b) x 0;707· hA .lA
A
Fillet weld! where, hA = leg length of weld A
B
1A = weld length of weld A .

T •• + b) x 0.707; hA ·lA
'tA = (a
P·b
... (v)

and similarly shear stress on weld B is,

P·a 1 ... (vi)
= -(a-+-b-). -0-.7-0-7-x-h-B
'tB X lB
where, hB = leg length of weld B
lB = weld length of weld B
Thus, the shear displacements BA and BB in welds A and B
respectively are,
'tA
Fig. 13.24. An unsymmetrical angle section fillet welded to a plate. ... (vii)
BA = G . tA
Solution. If a rolled angle iron is to be welded to a plate the where, G = modulus of rigidity of angle iron section,
attachment is to be so arranged that the joint does not rotate, nor E
does the attachment bend when it is loaded axially, thus
conforming to statis1;ically determinate analysis. - 2 (1 +v)
v = Poisson's ratio,
Thus,\ the conditions to be fulfilled are: " E = Young's modulus of weld metal,
(i) the displacement (not the strain) of each weld attach- tA = throat of weld A = 0.707 hA
ment point should be the same,
(ii) the load P is applied through the e.g. of the angle iron From (v) and (vii), we get,
section. . p. b 1
For equilibrium, BA = (a + b) x 0.707hA . lA x G x 0.707hA
p = PA + PR .•• (i) i.e., B _ p. b ...(viii)
:A - «(
where, PA = load acting on weld A
'tB
PB = load acting on weld B Similarly, ...(ix)
BB = G . tB
 930
Welding Engineering & Technology
From (vi) and (ix), we get, Weld Design for Static Loading 931

P·a .. P =; tw(ll + l2) x fs

X
(~+ b) . G· lB
8B =; ... (x)·
p 250 X 103 .•
Now, from condition (i), we have, £.e., II + l2 =; tw . fa=; 95 x 15 x 0.707 ( . tw =; h x 0.707)
8A =; 8~
or II + l2 =; 248.15 mm ... (i)
.. Equating (viii) and (x), we get,
P·b P·a Now, for no twisting, sum of the moments of the loads in the
welds, taken about the e.g. of the section, must be zero, hence,
(a + b) . G· lA (a + b) . G ·lB
II (15 x 0.707 x 95) 50 =; l2 (15 x 0.707 x 95) 125
a ~ .
~.e., b =; lA ... (x~)
thus, II =; 15205 l2 =; 2.5 l2 (ii)
i.e.lengths of fillet welds vary indirectly with their distan- ~.e., II + l2 =; 2.5l2 + l2 =; 3.5l2 (iii)
ces from the c.g. of the attached section.
From (i) and (iii), we get,
Problem 13.9. A force of 250 KN is to be applied to an angle
section welded to a plate using 15 mm leg length fillet welds as 3.5l2 =; 248.15
shown in Fig. 13.25. The centroid of the angle section is 50 mm 248.15
from the upright flange. Find the lengths of two welds II and l2 for ~.e., I2 =; 3.5 =; 70.9 mm
no twisting if the allowable stress in the weld material is not to and II =; 2.5 x 70.9 =; 177.25 mm
exceed 95 N/mm2•
-I /-15 . say II = 180 mm and l2 = 75 mm
Ans : II = 180 mm
l2 = 75 mm
Problem 13.IO •.A 150 mm wide, 12 mm thick steel plate is
welded to another plate with two equal 12 mm fillet welds and
subjected to a pull of 300 KN, as shown in Fig. 13.26. Calculate
the required weld lengths:

125

-,- f 12
I,
-L
L Thick
plate
Fig. 13.25. An angle attachment fillet welded to a plate.
Solution. Total required length of welds = II + l2
Now, Total force (P) on welds = weld throat x weld length
x allowable stress \...--
__~ ==+==
i
ifr 12mm

Fig. 13.26. Fillet welded plates.

932 Welding Engineering & Tecl1nology

FOURTEEN
Weld Design for
Fatigue Loading
A component or structure which survives a single application
of a static load may fracture if the same load fluctuates. Such a
failure is referred to as fatigue failure and it is the most common
type of failure of welded structures involving critical components
like bridges, ships, pressure vessels, press frames, automotive and
aero components. Overall, 50 to 90% of weld joint failures are due
to fatigue loading.
Fatigue failure occurs as a brittle fracture produced by cyclic
stresses which are usually below the elastic limit of the material
of the component. Since there is no appreciable deformation before
final fracture occurs there is no obvious warning that failure is
impending. This trecherous nature offatigue failure is emphasized
by the low stresses involved, stresses which are much lower than
static stresses necessary to produce failure. Fatigue stresses may
be axial, bending, torsional or a combination of these.
14.1. CHARACTERISTICS OF FATIGUE FAILURE
A fatigue failure goes through the following three stages :
(i) the initiation of the crack,
(ii) the slow growth of the crack,
(iii) fracture.
Fatigue cracks usually initiate as microcracks at the surfac\.
where stresses are highest and where corrosive environment and
change in geometry exist. These cracks are self-propagating.
The initiation of the crack is attributed to the stress cycling
which results in the alternating displacement of metal along slip
planes. Thus, the initial cracks form and propagate along the slip
planes at about 45° to the surface. If the load on the member
remains constant the stress increases as the crack becomes large
because of the reduced cross-section of the metal sustaining the
934 Welding Engineering
load. The direction of propagation which is along the slip planes,
for a few mm, at the beginning becomes macroscopically normal
to the maximum tensile stress by turning at right angles to the
column longitudinal axis. A fatigue crack is transgranular;
propagating within grains rather than along grain boundaries.
As the crack progresses the stress on the residual cross-
section increases so that there is a corresponding increase in the
rate of crack propagation. Ultimately, a stage is reached when the
remaining area is unable to support the applied load and the final
rupture occurs.
14.2. FRACTURE APPEARANCE
While the crack is growing, the adjoining faces rub agains:
each other, producing smooth polished surfaces. A brittle cleavage-
type fracture of granular appearance finally results because of the
restraining action of the crack. Thus, fatigue failures usually can
be recognised from the two distinct regions of fracture present,
VlZ.,
the smooth polished area of the original crack,
(i)
a granular area or a region of fibrous appearance of the
(ii)
final fracture.
The degree of polish depends upon the number of cycles of
loading after the initiation of crack and before complete failure
occurs. That is why there is often a tendency for this smooth region
to grow slightly, but progressively, rougher in texture as the
distance from the origin of the crack increases. Careful
examination of this smooth part reveals the existence of concentric
rings called beach, clam shell or conchoidal marks which represent
delays in the fatigue loading cycle. Apart from the beach marks
around the fracture nucleus there are radial lines essentially
perpendicular to the perimeter of the elliptical fatigue crack.
These radial or river patterns are often seen on the final fracture
surfaces and point to the source of the initial fracture location.
Fig. 14.1 shows a-SCberoaticrepresentation of a fatigue crack.
Fig. 14.). Diagrammatic representation of a typical fatigue fracture surface.
& Technology
Weld Design for Fatigue Loading 935
14.3. FATIGUE MECHANISM
Metals are crystalline in nature, which means atoms are
arranged in an ordered manner. Most structural metals are
polycrystalline and thus consist of a large number of individual
ordered crystals or grains. Each grain has its own particular
mechanical properties, ordering direction, and directional
properties. Some grains are oriented such that planes of easy slip
or glide are in the direction of the maximum applied shear stress.
Slip occurs in ductile metals within individual grains by
dislocations moving along crystallographic planes. This creates an
appearance of one or more planes within a grain sliding relative
to each other. Slip occurs under both monotonic and cyclicloading.
Fig. 14.2(a) shows an edge view of coarse slip normally associated
77
','eu,'oo,
(a) (oarse slip
'~;=
Extrusions
( ~=-r-O'l~
( b) Fine slip
Fig. 14.2. Schematic representation of slip in ductile metals
due to external loads ; (a) coarse slip due to static (steady) stress,
and (b) fine slip due to cyclic stress.
with monotonic, static, or high stress loading. Under lower stress
amplitude cyclic loading, fine slip occurs, as shown in Fig. 14.2(b).
Coarse slip can be considered an avalanch of fine movements. Slip
lines shown in these figures, appear as parallel lines or bands
within a grain when viewed perpendicular to the free surface, as
shown in Fig. 14.3.
Both fine and coarse slip are studied by examining a
prepolished and etched specimen by optical and electronic
microscopy technique. Electron microscopy shows that both slip
r
I
i

916
Welding Engineering & Technology Weld Design for Fatigue Loading 937

Ele ctrolytically polished surface
Fig. 14.3. Schematic representation of slip band extrusions.
band intrusions and extrusions occur on the surface of metals when
they are subjected to cyclic loading. Slip band intrusions form
excellent stress concentration zones which can be the locations for
cracks to develop. This slip is primarily controlled by shear stresses
rather than normal stresses. The higher the shear stress amplitude
or larger the number of repetitions, the greater the slip.
Fatigue cracks initiate in local slip bands and initially tend to
grow in a plane of maximum shear range. This growth is quite
small, usually of the order of several grains covering a maximum
distance of a few mm. As cycling of load continues the fatigue
cracks tend to join and grow along planes of maximum tensile
stress range. The two stages of growth of a fatigue crack are called
stage I and stage II and are shown schematically in Fig. 14. 4 by .
Loa d·109 'd' Irectlon
. therefore variation in oxidation of the fracture surface.
Crack
.,
Stage II
Free surface
Fig. 14.4. Schematic of stages I and II trans crystalline
microscopic fatigue crack growth.
a microscopic view of a cracked edge of a compopnent. This figure
shows that a fatigue crack initiates at the surface and grows across
several grains controlled primarily by shear stresses and then
grows in a zig-zag manner essentially perpendicular to, and
controlled primarily by the maximum tensile stress range. Most
fatigue cracks grow across grain boundaries (transcrystalline) as
shown; however, they may sometimes grow along grain boundaries
(intercrystalline).
The fatigue mechanism in high strength or brittle metals may
not contain slip band formation. Microcracks are often formed
directly at discontinuities, such as inclusions or voids, and then
grow along planes of maximum tensile stresses.
An electronic microscopic analysis of fracture surfaces can
reveal a wide rang£: of fatigue crack growth mechanisms two of
which viz, striation formation and microvoid coalescence are
particularly common. Very ductile materials like aluminium and
austenitic stainless steel often display appreciable striations and
and microvoid coalescence. Fatigue striations are a ripple
formation where each ripple represents the crack growth in one
cycle. The converse statement that every cycle produces one
striation is not necessarily valid, since the crack tip stress
resulting from some cycles may be too small to produce an
increment of crack length, especially in a random loading
situation. This fracture phenomenon provides the basis for
attempt at quantifying crack propagation by striation counting
giving a direct measure of crack propagation rate. These striations
are, however, not to be confused with beach marks described
earlier and which are often visible with the naked eye on fatigue
fracture. This latter effect is due to pauses during propagation and
Actually one beach mark can contain thousands of striations.
They often disappear due to the rubbing of surfaces during crack
growth and are also difficult to find in high strength materials.
The fatigue mechanism described above can be summarised
as shown schematically in Fig. 14.5. It shows that slip occurs first
followed by the development of fine cracks which can be seen only
at high magnification. 'l'hese cracks continue to grow under cyclic
loading and eventually become visible to the unaided eye. The
cracks tend to combine so that just a few major cracks grow. These
cracks (or a crack) reach a critical size and sudden fracture occurs.
The higher the stress magnitude, the sooner all events occur.
Cracks may also stop without further growth as a result of
compressive self-stress (residual stress) fields. Thus, fatigue
\consists of crack initiation, propagation, and final fracture. To
r
I
I
938 Weld Design for Fatigue Loading 939
Welding Metallurgy and Joint Design
In low cycle fatigue fracture failure after repeated loading of
less than 105 cycle is the case. There are many structural
A Fracture components in which low-cycle fatigue is usually the problem e.g.
(unstable crack) cracks frequently found in the ship hull are caused by low cycle
Crocks visible
fatigue. This is the case of high-strain fatigue crack initiation.
Visible growth and
with dye penetrants connection of cracks 14.4.1. High-Strain Low-Cycle Fatigue Life
Welds are often found at or near geometrical stress
concentrations in the structure and may in themselves cause
B Slip stress concentration. It is therefore likely that they may

,
VI
VI
QI
Fine cracks visible
at high magnification
----- Fine crocks visible
with naked eye
experience strains beyond yield even when the bulk of the
component is stressed to a low proportion of the yield strength of
the material. High strain fatigue cracks may initiate very quickly
at these points. This has led to the development of design
.:: c Slip Nonpropagating
cracks
philosophy based on strain versus fatigue life of welded
Vl
components and strain-fatigue life curves plotted on log-log scales
are shown schematically in Fig. 14.6 where N or 2N is the number
of cycles or reversals to failureI respectively.
~t
o Slip
.........
... ...

Very fine microscopic

that mayor craCkS)
may not propagate •••

N-number of stress cycles - QI

Fig. 14.5. Schematic representation of the fatigue process. -0

u
VI
en
overcome failure by fatigue, it is imperative to prevellt the crack .3
propagation. I

-
QI
-0 Of/E
14.4. TYPES OF FATIGUE FRACTURE ~
i5..
When a material is subjected to repeated loading, fracture E
o
takes place after a certain number of cycles. The lower the applied c
stress, the larger the number of cycles before fracture takes place.
In the case of fatigue fracture of engineering structures the
-
'0
<-
Vl
Plastic
--- strain
following two types of fracture are important.
(i) High cycle, low stress fatigue,
(ii) Low cycle, high stress fatigue. 2N
In high cycle fatigue, the life of a component is usually several Reversals to failure (log scale I
milli.on cycles. High cycle is a problem in those portions of a Fig. 14.6. Strain-fatigue life curves, for low cycle fatigue showing
structure subjected to fast, repeated loads, such af: points close to total elastic and plastic strain components.
propellers, rotating machinery, and areas under constant Failure criterion for strain-fatigue life curves is based on one
vlbrations. In such cases several million stress cycles can be of the following three factors :
achieved in a relatively short period of time.
~ 1
94() Welding Engineering & Technology ,
I Weld Design for Fatigue Loading 941
a small detectable crack,
(i)
(ii) life to a certain percentage decrease in load amplitude, where, c = - 0.5 to - 0.7 usually - 0.6
(iii) life to fracture. and b = - 0.06 to - 0.14 usually - 0.1
The total strain amplitude in Fig. 14.6 has been resolved into e'r = (0.35 - 1.0) £r
elastic and plastic strain components from the steady state Sometimes e'r and cr'f are taken as £r and af respectively as a
hysteresis loops. As a given life N, the total strain is the sum of rough first approximation.
elastic and plastic strains. Both the elastic and plastic curves can Equation'1-14.1) has been further simplified by Mason and is
be approximated as straight lines. At large strains or short lives, represented as,
the plastic strain component is predominant, while at small S
strains or longer lives the elastic strain component is ~ e = 3.5 E (N)-O.12 + e~·6 (N)-G.6 ... (14.4)
predominant. This is indicated by the straight lines and the where, Su = ultimate tensile stress for the material under
hysteresis loop sizes in Fig. 14.6. The intercepts of the two straight consideration.
lines at 2N= 1 are e'rfor the plastic component and a'r/E for the
elastic component and the corresponding slopes are c and b 14.5. S-N CURVE OR WOHLER LINE
respectively. This provides the following equation for strain- In order to determine the fatigue strength of a particular
fatigue life data of small smooth axial specimens. material under a given loading condition, it is necessary to test
several similar specimens. Each of these specimens is subjected to a
~e=~e+~ep given cyclic stress and the number of loading cycles required to
222 produce failure in each specimen is recorded. The relationship
= i. (2N)b + e'r(2N)c ... (14.1)
between the applied stress i.e. the stress range (Sm= - Smin) and the
number of cyclesto failure, N, is thus obtained. If this data is plotted
in a graphical form it is called an S-N curve or a Wohler line; one
where, ~ = total strain'amplitud'e, such line is shown in Fig. 14.7(a). When such a data is plotted on
log-log scales then the S-N curve for most of the engineering
2 = eIast'ICst'raIn amp l'tI ude = 2E = Iff
~ee ~a aa materials is a straight line as shown in Fig. 14.7(b); most fatigue
data are presented on the basis of log-logrelationships.
T plastic strain amplitude 2
~e
=
~e ~ee
= 2"" -
For plain ferrous metal specimens the S-N curve is almost
parallel to the N-axis after about 2 to 5 million cyclesindicating that
e'r = fatigue ductility coefficient, if the stress were slightly less, the specimen would have an infinite
c = fatigue ductility exponent, life. This limiting stress is called the ~ndurancelimit of the material.
cr'r = fatigue strength coefficient,
Most other materials do not exhibit an endurance limit, although the
S-N curve becomes nearly horizontal as N increases.
b = fatigue strength exponent, The sloping line of Fig. 14.7(b) can be expressed by the
E = modulus of elasticity, relations hip,
k
/),2a stress amplitude.
= aa =
The straight line elastic behaviour can be transformed to, F2:; FI (:: J ...(14.5)
~a b
where, Fl = the fatigue strength computed for failure at nl cycles,
'2: = aa = cr'r (2N) ...(14.2) F2 = the stress (SmaJe - Smin) which produced failures in n2
cycles,
The relation between plastic strain and fatigue life is, K = the slope of the best fit straight line representing the
data.
~p = e'r{2N)c ... (14.3) Constant K varies slightly with the specimen; however a value
of 0.13 has been widely used for butt welds and 0.18 for plates in
axial loading (tensiop and/or compression).
-

Weld Design for I<"atigue Loading 943

942 Welding Engineering & Technology e:

N VI e:
E
E
\

~b
...••
Z-
t:I
i---
..........
.•....
......
r--
.

~f oT~IT------- ~ --Sma,
"TIme
"- OJ
•....
.•....

E
VI \
"-
e: (a)
o Pulsating tension
:::>
E
cE
><

~~ - T\- -f\- - - - -- - - -- -s max

-0
e: ~L ... Time

.~~ lJ __V__L_ - - - - -- --.

E
o .§ Sm;n'· Sma,
:z VI
'-
QI
a.
o 500 1000 1500 2000 2500
Number of cycles to failure in thousands .§ ( b) Alternating

(a) S-N Curve on normal scales - - - Smax

.-
E N -
'-
0EuOJ S
Vl
VI
Zero-to-tension cycle - - - - - - Smin = h Smax
-0:z
....J
OJ
en
...••
VI o~_
F1..E0 ..••
Time
I
I Half tensile
I
----T----
I I Fig, 14.8. Typical stress cycles used in fatigue tests.

F2 000000
t- ~ _ •.•••
I
I
. 4I -
. I
-..- -
~
en
th

I I ,e:
I I ~~
~'"
N .••••. QJ
VI_
c0
Cycles to failure, N (log scale I :>._
QI ~ 'Vi
c:
FlOO,OOO= Fatigue strength for 100,000 cycles .- 40- ~
t:I-S 1-'
F2.ooo,ooo= Fatigue strength for 2.000,000 cycles Wu
'-,_
a.-
Eel
(b) S - N Curve on log -log scale
8~
Fig. 14.7. Wohler line or S-N curve;
(a) on normal scales, and (b) on log-log scales.
-
5b
we:
~e
e:
o
'Vi
VI
W
Only one type of stress cycle is represented on each S-N curve. e: •.... •....
OJW
•••a. a.
E
Therefore, in order to understand the general fatigue behaviour .5 uo
of a material or joint, it is necessary to have one S-N curve for
each type of stress cycle. Some of the typical stress cycles used in Fig. 14.9. Goodman diagram - a composite representation of the effects of
fatigue tests are shown in Fig. 14.8. various types of stress cycles on fatigue life.
I:
The data for various S-N curves can be summarised in a zero-to-tension, to complete reversal. Unless a finite life is stated
Goodman Diagram shown in Fig. 14.9. It provides a composite as the basis for the Goodman diagram, it usually pertains to the
f
repr€sentation of the effects of various types of stress cycles, fatigue limit.
'I
Ii,
I
including those ranging from static tension, through
 944 Welding Engineering & Technology Weld Design for Fatigue Loading 945

Ferrous metals and alloys exhibit a fatigue limit while non- c

V1 Sm = compression
ferrous metals such as aluminium alloys do not, as is evident from
Fig. 14.10. Because of this the strength of metals is sometimes ::l
QI

called the fatigue strength, which is the maximum fluctuating .-

...
VI
C7'
c:
Non -ferrous metals :g
c:
...
.-
QI

«
Steel
I Fatigue I I I I I
\
Log N
'.
~igUe
\ \il
--- Sf

limit
lifE)

'\ 103 10' 105

N
106 107

Fig. 14.11. Effect of mean stress and its nature on fatigue life.
10&

In general it is found that mean stress, S m' is detrimental

when it is tensile and beneficial when it is compressive, as is
evident from Fig. 14.11. The following equations represent the
effects of tensile mean stresses for different materials and
situations.
(i) Modified Goodman equation,
Fig. 14.10. Nature of S-N curve for steels and non·ferrous metals . -;-;-
+-=
Sa Sm
Su
1 ...(14.8)
stress that can be applied over a specified nuxrber of cycles; while .';If

fatigue limit is the maximum fluctuating stress that can be applied (ii) Soderberg <--Iuation,
on infinite number of times to the metal without producing
fracture.
-+--=
Sa Sm
Sf S:y
1 ...(14.9)

A common procedure to partially compare materials is to plot (im Gerber equation,

the un-notched fully reversed fatigue limit, Sf' obtained under ...(14.10)
laboratory conditions, vc::rsusthe. ultimate tensile strength, SUo Sf
Sa Su 2 = 1
+ (Sm)
S where, S:y = tensile YIeldstrength of the material,
The ratio of Sfu is called the Fatigue Ratio. Most of the practical
S __S_r__S_m_=_-_S_m_in_
. S a-2- 2
data indicates that Sfu varies between 0.2 and 0.65. There is a Smax +Smin
tendency to generalise that Sf increases linearly with Su, which Sm= 2
is often not true. It can be said that for steels the following Gerber's parabolic relation~hip between stress range and
generalization holds good. average stress forms the basis for one of the modern design
formulae on fatigue. If Sa is plotted on the Y-axis and the average
0.5 Su for Su ~ 1400 MPa
Sf= (14.6) stress, Sm' on the X-axis a parabolic relationship represented by
and 700 MPa for Su ~ 1400 MPa
Sf= (14.7) (14.10) is assumed. It can be rewritten in the following form.
By suitable surface treatment, however, fatigue limits of the
high strength steels can also .be increased to agree more with ...(14.11)
equation (14.6). s.~s+-[~:n
946 Welding Engineering & Technology Weld Design for Fatigue Loading 947

When completely reversed bending occurs Sm = 0 and the shows that the fatigue strength of materials increases .as the
strength of the material itself increases at a ratio of about 50% of
variable stress, Sa = Sf i.e. the fatigue endurance limit. D.T.S. This relationship is true only when the specimen is polished
14.6.. FAC1.'ORS AFFECTING FATIGUE LIFE and the surface is very smooth. If the material is severely notched,
The fatigue life of a material or welded joint is affected by the endurance limit is reduced drastically. Effects of alloy and
many factors, some of which are : material properties (for some aluminum alloys) on fatigue life of
(i) Material properties, transverse butt joints are shown in Fig. 14.13.
(ii) Joint configuration,
(iii) Stress ratio,
(iv) Welding procedure,
(v) Postweld treatment,
(vi) Loading conditions and sequence,
~
c 5
As wddect square butt joint
Transverse

.
100 a.QIc.-%:

;
VI
VI
1-
Cl
-f ,f!
----
------ , 0
10
...J
500

(vii) Residual stresses,

(viii) Weld reinforcement geometry,
(ix) Environmental effects,
(x) Weld defects.
14.6.1. Material Properties
Fig. 14.12 shows the relationship between ultimate tensile
strength and the endurance limits of various steels. This figure 1
103 104 105 106 107
280 560 840 1120 1400 MPa (approx. ) Number of cycles
140 980 Fig. 14.13. Effects of alloy and material properties on fatigue
strength of transverse butt joints.
.: 120 840 In general the relationship between strain, caused by the
cr
VI applied stress, and the endurance limit of different types of
materials, for low cycle fatigue, is well depicted by Fig. 14.14.
700

oo
o.-
-
.!:
~
560

QI
'-'
420 ~IN
C\
c: o
'-
Cl
:::l
...J

'1::1
c: 280
....

20 140
10 70
40 80 120 160 200 240 Log 2 N
Ultimate tensile strength 1000 lb per sq. in
Fig. 14.12. Relationship between ultimate tensile strength Fig. 14.14. Schematic representation of strain-fatigue
and endurance limits for various steels. life curves for different materials.
948 Welding Engineering and Technology
--------------------------.--.----
Weld Design for Fatigue Loading 949

a:~
14.6.2. Joint Configuration
--I
----- /'/' ./ -,,//' -~-
-~-j- -~~-
Clo~s o~.•In
In developing any fatigue behaviour criterion for'15weldroents,
rollto T - JOints
manuol
hutl.
and
welds
w~ld5
fiI'2t
weldswe
longitudinal
load-carrying
welds
with
manual
LOf"i.gitudinol
the severity of joint geometry is probably the most critical factor;
shown
madebutt
onnoorv~r~
butt
Ids
backing
welds
ors1rip.
outorncfl<::
mode
under(u.l
fillet
fillet
longitudinol

Plain plate
process.
andlood-carrying
if"l T-buH
Wold
fillet
welds
w"lrls,
transverse
wCZ'ld
U..
adjasccnt
positions)
edges
-+--c2J-
Tl'onsvC!rSf:
~lds
or
T ,-:bun
welds
typO
weld!.,
with
mode
welds
ran5,vtrs,€"
mod'l
B5
sc.lf-stu.:ss
we:lcJs with
153
tportial
p~n~trt]tion
WIth
Cruc'lform
Other
Longitudinal =~ ,----.---,-r
non-
butt
bun
on
nds.Trans-
(Nostop/slor1
load-carrying
fillet welds
flat
Cruciform

lor.gituoinol
10 of
positiof"l
butt and
shown
plote
cor
tronsvers£
type
butt

fill",
F
0
C
E
G
A
B
F2
o
Fotig~ 51 rer'ql 1'"'bC5~\..-: ...,n Smo •.
2x·.O'cyr,lt::;, R;: 0 ,MPo
50 lOe 150 200
'

the more severe the geometry, the lower the fatigue strength as ~~~ ~
Transverse: or ~",'\S<~ ~ 19J
shown in Fig. 14.15. The severity level of the longitudinal butt
Longitudinal butt weld _ {\ \_
170
Transverse butt weld -- ( , ;---

Transverse attachment

-~-
--~--
-
Doubletransverse strap

Cruciform ---
100

---
..-
~...
cX
."
E
C>." 5
Vl
QI
lloc: 77
C7' ."
III
III
QI

ti
•...

100
)(
50 ~
66
C7\
o
..j
As -w~ld~d 5063-0
S.R.: 0 .
10
1
10 103 105
Number of cycles
Fig. 14.15. Effect of joint configuration on fatigue strength of AI-alloy 5083. W~ld metal in load

weld is lowest, since the weld is parallel to the direction of loading. CO,frying !Oint5 made,
With parhal penetration
LOAD CARRYING FILLET WELDS
butt or fi lIet welds with W
The severity level ofthe cruciformjoint, on the other hand, is highest
since the welds are not only perpendicular to the loading direction thestressIOg.
\oor
of weld eitherTo\allood
\ran~Ye:r~toporaUel
direction _1; ~~
,,_'- II ~
,,_ J-
44

but also under the most severe load carrying condition. Some carried by weld metal
ocross throat area and
typical welded joints are shown in Fig. 14.16 with accompanying none by bearings between
the plotes.
design fatigue strengths for structural steels at 2 x 106 cycles with 25 50 75 100
PI rClnt of base mltal
R = 0, where R is defined as the ratio between the minimum stress
Fig. 14.16. Effect ~f weld type on fatigue Btrength for structural steel.
 950 Weld Design for Fatigue Loading 951
Welding Engineering & Technology
+
(Smin) and the maximum stress (Smax) and is depicted in Fig.
14.17(a) for constant amplitude cyclic loading and in Fig. 14.17(b)
for different amplitudes of tensile (+ve)and compressive (-ve)
'"
V> loading cycles.
~
OJ

Vl Lap welds have fatigue strengths conside;rablylower tluln that

o of butt welds. This is because lap welds disturb the flow of stresses
Time - more than do butt welds, as is shown in Fig. 14.18.
Stress ratio: (R
A ="Smin
So ISm
ISmox

(a) Nomenclature for cyclic loading

p -- ~-- -- =-- --~--------~
----- -::. -- =-= ,....
-=-_ ::,~;: - - =-
- -
·.t{~l:·, .;:, =., - -..;..-_
-=.. -_- - ' ,~, p
R arithmetical minImum stress R _ algebraic minimum stress
arithmetical maximum stress or - algebraic maximum stress
e::=-=-=:'-~,-------'"'~
p....-- t::.-- - - - -- -----~
St r ess ratio
~- ~- -__- = =---- = -=. -:. ..• -=. -=. =' =- '=t -...,....,
--::.-=.----==-=1
-::. p
Arithmetlrrll Algebrnic Fig. 14.18. Lines of stress flow in butt and lap welds.
r: . .':.flILiTI ~tT('SS minImum s.tress

!.l ': +0·5 !.! = +0.5 The weldments shown in Fig. 14.16 are rather simplified
+2 +2 joints, but they simulate many real parts and structures. In
general, the most used welds are butt, fillet and spot welds with
many different shapes and configurations. :J3othtransverse butt
and longitudinal butt welds are quite common. Fillet welds,
however, are perhaps the most common and may be load carrying
~
+2
= 0 ~ =0
+2
as in class G or non-load carrying as in class F, Fig. 14.16. The
design fatigue strengths given in Fig: 14.16 range from 25 to SO
percent of the un-n()tched base metal strength and provide a
reasonable guidance for actual design usage.
14.6.3. Stress Ratio
"-"
"',...
~ '-/ '"
-',... +1
-2
+2 :l
!.!
~ =~-05
- 05 ::..l =-0.5 It is well known that stress raisers tend to reduce fatigue life
+1
+2
+1 -2 =- 1 =-1
-=l ': ::.l
due to the notch effect. These stress raisers can be mechanical
'"
r-.. such as fillets, undercuts, lack of penetration or fusion. They can
also be metallurgical like microfissures or cracks, porosity,
inclusions, and so on. The shape and size of these weld defects
determine how much stress will be concentrated at a point. Cracks
concentrate the highest amount of stress, and generous fillets with
a smooth, polished surface concentrate the least.
In general, the presence of weld defects, especially those on or
near the surface, reduces fatigue strength. For example, a surface
notch affects the endurance limit of a material much more than a
notch in the interior of the same material.
In most materials, the notch sensitivity tends to increase' with
-2 0 increasing strength. When sharp notches are present, therefore,
0='" _"2=0
it is not unusual to find little or no fatigue strength advantage in
using high strength materials.
lb I Some typical loading cycles
The effects of notches are evaluated by comparing notched
Fig. 14.17. Alternating loading cycles; (a) Nomenclature for constant amplitude versus un·notched S-N data. To determine an S-N curve for a
cyclic loading, awl'{b) some typical loading cycles with different stress ratios, R.
952 Welding Engineering & Te;chnology Weld Design for Fatigue Loading 953

notched specimen, the net stress is plotted. By comparing data, it them. Stress concentrations which occur at the toe of butt and
is then possible to calculate a fatigue notch factor (Kf) where, fillet welds, are common locations for fatigue cracks to initiate.
The toe is at the surface where bending stresses are the highest.
K == fat~e limit, unn~tch_ed Kt at the toe depends upon the geometry of the weld. It is usually
( fatigue limit, notched
larger in fillet welds than in butt welds. Values of Kt have ranged
From the K(, the notch sensitivity factor, q, can be calculated. from essentially 1.0 for butt welds with the reinforcement removed
KI'-l..... up to values of 5 to 7 for sharp discontinuities in fillet welds.
q :::--'--- 19)
Kt -1 '"(14 .•.. 14.6.4.Welding Procedure
where. Kt == stress concentration factor, Welding procedure adopted may lead to stress concentration
S cr or development of defects which will consequently affect the
::: ..:::JL as long as - :::constant == E initiation and development of cracks. For example, if preheating
Snet to
is employed it will help in reducing hydrogen embrittlement and
where, Sp peak stress,
==
hardness of the heat affected zone. This can help in improving the
average stress on the net section.
S net::::: fatigue life of a welded joint.
For fillet welds, Similarly, the use of multi-run welds can lead to improved
Kt == 2.2 to 4.7 at toe weld metal structure due to normalising effects to the successive
== 2.0 to 6.9 at root. runs but on the other hand it can also lead to defects like slag
entrapment and increased distortion. Metallurgical structure will
Fig. 14.19 illustrates the full and partial penetration butt and also depend upon the welding procedure adopted hence may affect
fillet welds with common locations for crack to initiate marked on the fatigue life of a weldment.
Weld metal
14.6.5.Postweld Treatment
Heat affected zone
Postweld heat treatment is usually given to stress relieve the
component or structure and normally involves heating it to a
-- -- suitable temperature, holding it at that temperature for a
predetermined time, followed by uniform cooling. Such a
treatment results in changes in grain structures and dimensional
Weld bead changes which can be injurious to a part as it affects the
Butt reinforcement F itlet microstructure and consequently tensile and impact strengths.
This may affect the fatigue strength of the weldment.
(a J Full penetration welds.
14.6.6.Loading Condition and Sequence
In laboratory, fatigue testing machines usually apply a simple
stress cycle of constant amplitude, but under actual service
conditions the stresses are rarely cyclic and are of complex nature.

-- -- For example, vehicles on road, sea and air are loaded in a random
fashion by their loading and motion. Towers and tall buildings
suffer wind loading of varying intensity. Cranes lifting a range of
loads from various positions will have different fluctuations in
~'-.r loading. Wave loadings on piers, jetties and off-shore platforms
Butt Fillet cause a complicated stress history.
To caiculate the fatigue life under such complicated loading it
I b) Partial penetration welds.
is imperative to calculate or measure the number of cycles at
Fig. 14.19. Weld nomenclature and fatigue crack initiation sites for; various stress ranges. To achieve this use is made of the Linea,.
(a) full penetration welds,. and (b) partial penetration welds.
 954 Welding Engineering & Technology
Weld Design for Fatigue Loading 955

Cumulative Damage Rule which is explained by saying that if for

a stress range 81 the constant amplitude fatigue life is N 1 cycles, cE
:L:.
VI
200
~u
then nl cycles will use a fraction ;; of the fatigue life. In a stress >-
u
history with stress ranges 81, 82, 83"" etc. occurring nv n2, n3 ..., •••

~)( 150
etc. times respectively, failure will occur when, N
.•....

N n22 + N
nl1 + N n33 N= ~ Nn = 1 at failure
nnn "" . ...(14.13) C
c 10o.
d.
.s=
I.
<II
It has been reported that fatigue strength of smooth specimens •...
"ti
I {t is reduced more than indicated by the Linear Cumulative Damage ~ SO
Rule if a few cycles of high stress are applied before testing with .~
.•....
lower stresses. This effect is of particular importance in the case Cl
u..
of notched parts. For example, the life of a specimen with a hole o
100 120 140 160
was 460,000 cycles oflow load after 9.5 cycles of high load but only
63,000 cycles after 10 cycles of high load. Reinforcement angle. 8, degree
14.6.7.Residual Stresses Fig. 14.20. Influence of reinforcement shape on
fatigue strength of transverse butt welds.
When compressive residual stresses exist on or near the
surface of the workpiece, the fatigue s~rength may be increased.
For example, surface hardened stael having a hard skin in 600
compression has improved fatigue life. The compressive stress can
reach the yield strength of the hardened skin which very
effectively prevents the growth of cracks and thus helps in a.c
achieving gain in fatigue strength.
%:
----"-
The welding process can produce both tensile and compressive -
~
<II
•...
VI
400

h =1·5 mm
residual stresses on the surface which may reach the yield E
::;,
strength of the material. As prior tensile stress will reduce the E
fatigue life so it is imperative to eliminate or atleast reduce such '§ 200 ~ -1. h :: 2.3 mm
a stress to avoid failure of a welded joint under much less load
then intended to take.
14.6.8.Weld Reinforcement Geometry
%:

o
T®-, > < -
t-19.1mm
h :: 3.8mm

Resistance to fatigue can be improved by reducing the stress 2 x104 105 106 101
concentration caused by the reinforcement. The influence of Fatigue life, N • cycles
reinforcement angle, e for transverse butt welds is shown in Fig. Fig. 14.21. S-N curves for transverse butt welds in Q and T carbon
14.20; while Fig. 14.21 shows the effect of decreasing steel with UTS of 785 MPa, and R = o.
reinforcement height for double vee but joint in a steel having an 14.6.9. Environmental Effect
ultimate tensile strength of 785 MPa. It shows that the unnotched
fully reversed fatigue limit, 8f' for the transverse butt welded joint There are many environments that affect fatigue behaviour,
with reinforcement removed is very similar to that for the base however, generally engineering structures are subjected to
environments containing air, water, and salt water. Corrosion
material. Therefore, when a joint with a high fatigue strength is fatigue refers to the joint effect of corrosive environment and
required it is imperative to remove the butt weld reinforcement. repeated stressing. The combination of both acting together i.s
 Weld Design for Fatigue Loading 957
956 Welding Engineering & Technology
alloys. Presoaking the fatigue specimens in a corrosive liquid
more detrimental than either acting separately. Corrosive' followed by testing in air often causes a detrimental effect.
environment can be detrimental even under static loads However, the combined effect of environment and repeated
particularly so for higher strength alloys but failure under such stressing causes the most drastic decrease in long-life fatigue
conditions is referred to as stress corrosion cracking. strength.
Laboratory data obtained for four conditions of testing with Many structures in the sea fail because the sea water
constant amplitude loading is shown in Fig. 14.22 in the form of accelerates the fatigue process; Fig. 14.23 shows the combined
effect of weld reinforcement and salt water on fatigue behaviour
of aluminium alloy 5456-H117. A common solution is to design
with a low strength alloy at a low working stress, say below 35
o N/mm2• But that makes the marine structures big and bulky.
VI
Attempts have been made to use higher strength materials for
-
VI
II>
'"
"- marine structures but the effect of marine environment more than
VI cancels out the useful increase in strength.
-oc
C7l
.S: 14.6.10. Weld Defects

-«"-
'"
Regardless of the type of material under consideration. weld
defects may be classified in six main groups :
(i) Imperfect shape,
Ui) Incomplete penetration,
(iii) Cracks,
L- I I (iv) Lack of fusion,
103 105 106 107 108
(v) Voids and porosity,
Applied cycles to failure, N - (vi) Inclusions.
Fig. 14.22. Relative fatigue behaviour under various environmental conditions. Although they may be adjacent to, or emerge at a surface the
SoNcurves. It is seen that while vacuum can have small beneficial last five of these are essentially internal defects. Imperfect shape,
effect at long fatigue life. High humidity air can be quite on the other hand, is an external defect which includes not only
detrimental to fatigue behaviour particularly so with aluminium poor shape i.e. excessive reinforcement and misalignment but also
undercuts and overlaps. The effect of misalignment, whether axial

-
VI
~
VI
VI
QI
I
in
500

100 ~
%:
t
cE'

VI

"-
r L (01
I
E---=?-
.':: 10
VI
50 ~
ox
E
)(
o
E +----
01 01 ( bl
o Transverse butt weld o

}-rr-~-f-
....J ....J
S.R, ;-1.0
1
104 105 10 I•....
~_ r--- -
-i.(: -:---- e-
Number of cycles - Ie)
Fig. 14.24. Butt welds between plates of unequal thicknesses.
Fig. 14.23. Combined effect of weld reinforcement and salt water on fatigue
behaviour of aluminium alloy 5456-H1l7.
r m •••• _

958 Welding Engineering and Technology 959

Weld Design for Fatigue Loading
or angular, is also to reduce the fatigue strength of a welded joint. t = the plate thickness of the thinner plate ifthe plates are
Effect of excessive reinforcement has already been discussed. of different thicknesses.
14.6.10.1. Misalignment For a misaligned joint subjected to a stress, cr, the life should
Considering axial misalignment, it can be caused deliberately, be approximately the same as for an aligned joint subjected to a
as shown in Fig. 14.24(c), or accidentally in which case it is more' 1 stressKt . cr.
often of the form shown in Fig. 14.25 (a). It is usually caused by In the case of angular distortion it has been shown that fatigue
strength decreases with increasing distortion, as might be expected.
It has been reported that for a given level of applied stress there is an
t-------~&-------t (0)
approximately linear relation between log N and the out of
straightness, h [Fig. 14.25(6)]. One of the most common places in
which angular misalignment tends to occur is in the longitudinal seam
8 welds of pipes where it is often referred to as Peaking or Roof-topping
and is usually caused because of improper curvature induced. in the
~-
t-.------- I, ~
lh
----- •. 1-
plate during rolling. The stress concentration factor for a cylindrical
vessel with roof-topping under internal pressure is given by,
(b) K am =1+--
D. t
I-- 1tD ... ( 14.15)
Fig. 14.25. Butt welded joints with; (a) axial, and (b) angular misalignment. 6L2 ( 2L)
incorrect or inadequate jigging before welding, although avoiding
misalignment in thin-walled cylindrical structures with large
diameter to thickness (D/t) ratios can be difficult. Under axial
loading it is clear that the joint will try to straighten, as shown in
Tension Compression

p-l ~
Compression Tension
\-p t

Fig. 14.26. Idealised picture of deformation of misaligned

specimen under axial load.
idealised form in Fig. 14.26, and as a result secondary bending
stresses will be induced. Fig. 14.27. 'Roof-topping' or angular distortion at the
longitudinal seam weld in a pipe or cylinder.
For tests with misalignment up to 16% of the plate thickness
and with axial loading, fatigue life was severely reduced on account where, Kam = stress concentration for angular misalignment,
of the secondary bending stresses. In order to take these stresses 2L = total width of angular misalignment (Fig. 14.27)
into account the following equation has been developed to calculate D = cylinder diameter, t = cylinder wall thickness.
the stress concentration factor, Kt, for a misaligned joint subjected It is also reported that undercuts and overlaps can reduce
to axial load, considerably the fatigue life of welded joints.
Kt = 1+ t
3e

where, e = the distance between the centre of thickness of the two

...
(14.14)
14.6.10.2. Incomplete Penetration
Incomplete penetration and partial penetration welds are
prone to early fatigue failure as the unwelded portion acts as a
abutting plates [Fig. 14.24 (c) and 14.25 (a)] precrack.Ifthe incomplete penetration is buri~d inside the weld,
 960 Welding Engineering and Technology Weld Design for Fatigue Loading 961

as for double-sided butt or fillet welds, the decrease in fatigue to the direction of the applied stress and showed a reduction in
strength as a function of percentage reduction in cross-sectional area fatigue strength of 55% to 65% of 2 x 106 cycles. In general it is
can be expressed as is'shown in Fig. 14.28 for transverse butt welds assumed that cracks produce a considerable reduction in fatigue
in mild steel. These results are based on experimental data obtained strength and thus due importance be given to the control of this
by creating the incomplete penetration of desired length and height. defect in welds.
180
14.6.10.4. Lack of Fusion
•
- ".-~.-~•..•ee\ •-• I,. ,,~
• With weld reinforcement
This defect consists of boundaries of unfused metal, usually
!
...••...
Ol
Cl N<-
::l
V<:>
Ol
u..
><
c::
:+::
Q/ l 40
80 ~-
60
160
100 I With reinforcement machined
~ 140
Q,o
between the parent metal and weld metal or between adjacent layers
~ 120L off
ofmulti-run welds. The main cause ofthe lack offusion is the presence
of foreign matter, e.g. slag or scale on the surface to be welded and
that prevents the metal from reaching fusion temperature. Use of
lower current than required may also lead to this defect.
Depending upon the orientation of this defect to the direction
of the applied stress it can lead to fatigue failure similar in nature
to that obtained in butt welds containing lack of penetration.
14.6.10.5. Porosity
This defect refers to the presence of cavities in the weld metal
caused by gas entrapped during solidification. The cavities may
be widely scattered or may exist in localized clusters. The cavities
~ which are usually spherical in shape may vary from tiny pinholes
(non-detectable by radiography) to large holes of more than 5 mm
20 in diameter.
200
o --L. I -1. I I I
o 10 20 30 40 50 60 .•.~
,,,
Defect severity( reduction of area

Fig. 14.28. Effect of lack of penetration on the fatigue

%)
•..•
ee
"z ,
strength of transverse butt welds in mild steel.

When lack ofpenetration occurs as a surface defect it is expected

140
"- .... - --
....

to produce greater loss ofstrength, for a corresponding severity, than ....

the 'buried' defect because of eccentricity effects. When subjected to -""-

purely longitudinal loading (i.e. parallel to the defect length) this type
of defect, at least when it is continuous, has been shown to have little ClI
:;;J

or no effect of fatigue strength. Under shear loading, however, it has .S!"

c;
been the cause of failures in service. Thus, it is essential that every L.I-

effort should be made to avoid lack of penetration whether it appears

in buried form or as a surface defect.
14.6.10.3. Cracks
Cracks associated with welded joints occur in weld metal or in o
the heat affected zone and may be either parallel or transverse to o 2 4 6 8 10 12 14
the line of the weld. Internal cracks produced in transverse butt Defect severity (reduction of area % )
welded mild steel were parallel to the weld direction and transverse
Fig. 14.29. Effect of porosity on fatigue strength of butt welcis in mild steel.
962 Welding Engineering & Technology Weld Design for Fatigue Loading 963

Porosity in steel may be caused by high sulphur content, welds mainly due to improper cleaning of weld runs. The
excessive moisture in electrode coatings or more often by the relationship between Average Length of slag inclusions and
sucking in of atmospheric nitrogen due to the disturbance of gas Fatigue strength of transverse butt welds in 12.7 mm thick steel
shield around the arc due to excessive arc length in SMAWor large plates are shown in Fig. 14.30. However, the butt welds with slag
nozzle to plate distance in GMAW. inclusion parallel to the applied stress have been found to be
If the defect severity is expressed as the percentage reduction relatively harmless. Also, tungsten inclusions have been found to
in area due to the porosity then the relationship between defect
,
3
severity and pulsating tension fatigue strength is of the shape
shown in Fig. 14.29. It shows that comparatively small amounts •
of porosity produced drastic reduction in fatigue strength, but 2 •.•
•
there was very little effect of further increase in porosity.
••
.!o •
0
+
0
,"
'
As porosity leads to reduced weld area therefore the presence
of reinforcement can help in negating its effect to some extent.,
o ••
•• +0 I,
However, with reinforcement removed the pores form the major .o:o •.,@
",.00 y
stress concentration spots and therefore become the initiating * •. 0' @ X

points for fatigue cracking. The fatigue strength is also much .*~o",,·xo
dependent on pore position than pore size; for example the defects U
QI fC!.cr
. *:•.~o,,@ 0
~*•.•0 x".o
>.
that extend to the surface or which are just beneath it are u
"- 5 + "I). tr;x.·x"X@lb
particularly critical. E
14.6.10.6.Inclusions E 4 -0- .+.J! •
+*~"o'~@
2Q +A .+ .t·x~
Any non-metallic matter included in a weld of steel is normally 3 ••.:l·x." .,.
referred to as slag inclusion. In SMAWwelds, it is usually caused
~I:z:
~-c - •.•• 0 .,,..,,y
,"x.xc •• "
+
c:
by the entrapment of slag from the electrode coatings in multi-run o
+= 2 •... ", ..
c0\ !60"" ~ .x o~"
..':t+v""
ca. AOo@JI."
o 1.5 "+0."
./
•..•
E " Low hydrogen L- *.1.
a.
~
:z: ............•• -
..•• - -

-- _,--
C!.
oX
:5 10
-4
..
••.•.•.

"' ..'
.••0
* ""@+
+·r°x.. Y A
'-'

."
III
~u
>. •
.0

-o
...
u •• 0.
,,"* .;:. x"
+ , o Weld metal A
• Weld metal a

=
..,u
~X
-c
'"

ex: ..
* * "••+0;
* ,:·x+'"
"....
+ Weld metal C
•. Weld metal 0
x Weld melal E
N 5
•• ~>ry • Weld melal F
'0 4 , S',mulated HAZ in mild steel
.c.
.•.. Rutile welds with
@,.
0\
C
• +". y •. Simulated HAZ In as 968(950·C
heat treatment l
'" continuous slag - 3 ••. @•• o
.-L-III line
-!I",.
••.=+ +0 •• S"mulatcd HAZ in as 968 (1100·C
'" Rutile Low hydrogen heat treatment l
,
.. ,..+~v
::) welds welds
0\ 2 @ Tempered HAZ in as 968 (1100:C
~ o 6 Single inclusions *,,@c hf:gt treatmen!, tempered at 650 Cl
C

,' ,
u..
• A Multiple inclusions , Stul to as 968
1.5 ",,0
...
-5 "'0
'1-0
" 0
o
300 400 500 1000 2000 3000
o 2 4 6 8 10
Average defect ,length ( mm) Range of stress intensity factor ,6K MPa {m
Fig, 14.30. Effect of average defect length on fatigue Fig. 14.31. Fatigue crack growth data for structural C-Mn
strength of transverse butt welds in steel plates. steel weld metals, HAZs and base metals.
964 Welding Metallurgy and Joint Design Weld Design for Fatigue Loading 965

have no effect on fatigue strength. This implies that tungsten is
cap~ble of transmitting stress and does.not form a severe stress
concentration spot.
14.7. CRACK PROPAGATION BEHAVIOUR OF
WELDMENTS UNDER CONSTANT AMPLITUDE
LOADING
Fatigue cracks can initiate or grow in the base metal, or HAZ;
however only limited cyclicstress-strain data and low cycle fatigue
data exist for the three weldment regions. The general trend found
for e-N curves for the three weldment regions is that the softer
materials have better fatigue resistance at the low lives while the
harder materials are better at the longer lives as depicted earlier
in Fig. 14.14.
Fatigue crack growth data for weld metals (with yield strength
ranging from 386 MPa to 636 MPa), simulated HAZ and C-Mn
base metals are presented in Fig . .14.31(See page 963). As most
steels used in welded structures have yield strength below 70.0
(iii) Use ofTIG dressing to improve weld profile and remove
entrapped inclusions at the weld toe.
14.8.1.1.Avoiding Severe Stress Concentration
To reduce the stress at the critical section without either
reducing the load or producing a new critical section elsewhere is
often impossible. This can be very well illustrated by two useless
but the so called strengthening techniques. The first is to add local
cover plates to a member over a region which is known to be weak.
This does reduce the stress to a satisfactory level at that section
but introduces new and severe stress concentrations at the ends
of the cover plates. This leads to the use of infinite number of cover
plates, one after the other, without any useful result. The other
stress reducing modification that has often been used with
= == == ==-..,r-----=--=--::.==
•• A

MPa therefore this figure gives representative data for most such .,
"
steels. Eleven different conditions are superimposed in this figure. '.
••
The scatter band for all stress intensity factor ranges varied II
between 2: 1 and 3: 1 with crack growth rate between 10-8 and
(a) Badly designed square frame corner with
3 x 10-5 m/cycle. This small scatter band implies that stage II severe stress concentrallon at the cornerA.
fatigue (see Fig. 14.4) crack growth behaviour is similar in sound
weldments. Similar behaviour is reported to occur in welded steel == =-..:.---.::=;:;=-= = = :.-:...---=
pressure vesseis. I
14.8. METHODS OF IMPROVING FATIGUE LIFE OF "I' 1 /B
WELDED STRUCTURES ,I r B
The fatigue strength of weldments can be improved by II
••
different methods which can be divided into three groups. II
1. Improvement in design and welding procedure, (b) The addition of triangular gussets produces
2. Inducing surface compressive stresses, and little improvement. ThEre are still severe strESS
concentrations at the cornu B.
3. Protection from atmosphere.
Brief description of these methods of improving fatigue life of
welded components and structures are given in the following
sections.
14.8.1.Improvement in Design and Welding Procedure
Improve the actual welding design, procedure and postweld
treatment to get better stress distribution weld shape and/or
microstructure. This may include,
(c) Improved design with reduced stress
(i) Avoiding severe stress concentration, large changes in concentration with WElds separated
from it.
stiffness, intersecting welds and weld ripples,
(ii) Modifying weld notch, if present, by local machining, Fig. 14.32. Different designs for the comer of a welded frame;
(a. b) poor designs, and (c) a good design.
 Welding Engineering & Technology· Weld Design for Fatigue Loading 967
966

disastrous results is shown in Fig. 14.32(a). In order to strengthen
a badly designed square corner in, for example, a chassis frame,
in which failures inevitably occur at the point A a triangular
gusset is welded into the corner as shown in Fig. 14.32tb). This
results in no improvement as it transfers the point of failure to
the weld end at B. The proper remedy is to replace square
connection between the flanges by a shaped plate attached to
'l1embers by butt welds.
Reinforcement
In general the improvement in butt and fillet welded T-joints
can be achieved by forming correct contours as shown in Fig.
14.33.
Intersecting welds can result in serious stress concentration
so should be avoided; alternatively use rat holes, Fig. 14.34, to·
overcome such a situation.
14.8.1.2. Modification of Weld Notches
Weld shape can be modified by local machining to reduce or
eliminate the point of stress concentration. For example, by

~0 i
concentration machining off the reinforcement the strength of the transverse
butt weld can be increased to that of the parent material provided
there are no other significant weld defects.
With fillet welds it is impossible to restore fatigue strength to
that of the parent material by local machining, since there will
still be a change of shape. However, considerable improvement in
fatigue strength can be achieved by reducing the acuity of the
notch created at the toes of the fillet welds and by removing the
slag inclusions which often exist at the weld toe.
. Local machining of the toes of fillet welds is normally carried
out by disc grinder or rotary burr grinder. In order to obtain a
lack of penetration significant improvement in strength by toe grinding it is
stress - concentration lo •••. stress -concentration absolutely essential for the grinding to penetrate into the plate
High surface, as shown in Fig. 14.35, so as to remove the critical toe
•••.
elds •••.
elds
Fig. 14.33. Stress raisers in welded butt and inverse
T-joints and their connections.

- -
I
®/~_: :_~~

Fig. 14.35. Effect of grinding toe of a fillet weld;

(A) Poor shape, and (B) Good shape.
defects. A slight decrease in plate thickness and the corresponding
increase in nominal stress by such machining is hardly of much
consequence.
In the case of transverse non-loading fillet welds the greatest
improvement in strength can be achieved by burr grinding of the
weld to a concave profile-often referred to as full burr grinding.
(a) Intersecting welds I b) Non - intersecting welds
The relevant results are shown in Fig. 14.36. It is evident from
Fig. 14.34. Crnciform joint with: (a) intersecting welds, this figure that the effect of local grinding is to rotate the S~N
and (b) non-intersecting welds.
 968 Welding Engineering & Technology Weld Design for Fatigue Loading 969

• - disc
-0,..\
~\
ground a \
~ "280
320
\ "
300
~\\
\\t-~\
I ,
"A''" ,. . \
~
/J. •"
/J.
""-
"""IA .......
'-- I,
14.8.1.3. TIG Dressing
TIG welding appears to cause no intrusions at the weld toe so
leads to a good fatigue strength of a welded joint made by this
,.. b - .\~~\ ....
Fully burr ground
\..
260 process. However, TIG welding being a slow and costly process
•..•
-E 240 cannot be used to improve the fatigue strength ofjoints welded by
E other process. This is achieved by remelting the already laid weld
z
...•...220 metal to a shallow depth along the weld toe, thereby removing the
intrusions and generally improving the toe shape.
::;
VI
200
Q,I
.: 180
To achieve optimal results by TIG dressing it is best to clean
VI
the weld toe by wire brushing and light grinding before making
-- the TIG runs by placing the TIG torch 0.5-0.15 mm away from
:§ 160
the weld toe. If TIG dressing results in excessive hardening of
'- some steel than that can be substantially reduced by making a
~ 140
§' second TIG run on the basic weld, and 2-3 mm from the first TIG
run.
120
A particular problem is that of stopping and restarting after
accidental stops during TIG dressing, because of the difficulty of
100 avoiding an irregular TIG bead profile. The simplest method is to
re-initiate the arc about 6 mm behind the existing crater, as shown
in Fig. 14.38(a), but the success of this technique depends heavily
80
10S 106 107
Endu ranee (cycles)
Fig. 14.36. Effect of grinding weld toes to concave profile
by rotary burr grinding. and disc grinding.
curves so as to give a relatively large increase in fatigue strength
(a) Ie)
at low stresses but less at high stresses. IbJ

In the case of discontinuous longitudinal butt-welded

attachments on the edge of a stressed plate, two types of local Fig. 14.38. Examples of weld stop and restart techniques
machining can be done; one is machining the corner to a suitable for use in TIG dressing of welds.
radius and the other involves drilling a hole at the ends of the
attachment as shown in Fig. 14.37. It is reported that the optimal on operator skill. Alternatively, the arc can be reinitiated on the
results are obtained for a radius of about 10 mm and a hole surface of the weld being treated and withdrawn into the existing
diameter of about 20 mm. crater, as shown in Fig. 14.38(b). Similarly when deliberately
stopping the TIG weld, the arc can be led from the toe on to the
weld surface, as shown in Fig. l4.38(c).
11111I.111111111111111111111111111111

IIIIIIHn"' •••••••••••••• 111I11

By its nature, TIG dressing is most suited to the treatment of
1 Radius r welds which are transverse to the direction of stress, but it can
also be used to treat the ends of longitudinal welds. As far as butt
welds are concerned the reported increase in fatigue strength at
(a) End of weld radiused. ( b) Drilled hole at weld end 2 x 106 cycles ranges between 18% and 17% but majority of the
results show an increase of 25% to 75%.
Fig. 14.37. Two methods of reducing the stress concentration
at the ends of discontinuous longitudinal butt welds.
 971
970 \"elding Engineering & Technology Weld Design for Fatigue Loading

14.8.2. Inducing Surface Compressive Stress V.S. Y.S.

This can be done by the following methods :

(i) Prior overloading,
(ii) Peening,
(iii) Local compression, and
(iv) Spot heating.
14.8.2.1. Prior Overloading
(0) Weld residual stresses. (b) Residual stresses+applicd
A possible method of improving fatigue strength is to create Ie nsile load.
residual compressive stresses at and around the weld where
fatigue cracks are likely to initiate. In this way even a wholly
tensile applied stress cycles can be made at least partly
compressive provided it is not so high as to relieve the residual V.S. \'.5.
stresses by yielding.
When a butt welded specimen containing stress concentration
points is loaded to the extent that yielding occurs at and around Q)+G)-(D = ®
those points and the load is then removed, residual stresses of
different distribution pattern will be left in the specimen. The way
in which they are formed is illustrated qualitatively in Fig. 14.39
wherein curve (1) represents the initial residual stress pattern for
,--,'
the butt welded plates. If a constant tensile load is applied as prior' (c) Theoretical ® and practical (d) Resultant residual slresse s
otter removal of applied tensile
G) e1Ieclsof applied tensile load.
overload and if yielding is ignored the elastic stress distribution load.
can be represented by curve (2). It will be seen that over the width .'
shown in Fig. 14.39(b) the yield stress is in fact exceeded, which V.S. Y.S.
Service
is impossible, so that the plastic flow occurs in and around the
welded zone. Load is therefore shed from that region to the part
.~
of the specimen which is still elastic, so that the actual stress
distribution due to the combined effect of initial residual stresses .•..
Z.J
.stress~~
._ . ~.--.;I {,'-"
--
and the applied external load is that represented by curve (3).
Now, when the load is removed, the whole specimen will unload
elastically to leave behind final residual stress distribution
represented by curve (4), this being the result of subtracting curve
(2) from curve (3) and adding to curve (1). The magnitude of the
final residual stress is, therefore, obviously dependent on the size
(e) Stress variation with mono-
tonic tensile applied stress.
.__ '-' Ji~ '.
' -'
( 1) Stress variations With fluctuating
of the overload and the stress-relieving treatment which may be service stresses.
given to modify the curve (1).
It is important to keep in mind that the sign of the residual Fig. 14.39. The reformation of residual stresses by prior overloading and the
stress adjacent to the stress concentration point (e.g. a notch) will effect of subsequent loading during service.
always be opposite to that of the overload. Thus, in order to create The behaviour of such a specimen during subsequent fatigue
compressive residual stress adjacent to the stress concentration testing under a lower tensile load is shown in Fig. 14.39(e). If the
points in weld bead and the HAZ it is necessary to apply a tensile elastic distribution (theoretical) due to the load plus the initial
overload. residual stress distribution can be expressed by curve 2, while the
 972
Welding Engineering & Technology .. Weld Design for Fatigue Loading 973
actual distribution is given by curve 3, the final distribution is and for the pre-loaded S-log N curve,
given by curve 4 which is equivalent to curve 3 - curve 2 + curve S +K210g N=A2 ••• (14.17)
1. During a tensile fatigue test the stress will pulsate between
curve 4 and curve 6 which is the stress pattern caused by a ' •. where Kl' K2 and Al are constants for a given type of specimen,
maximum tensile load represented by curve 5. and A2 depends on the preload stress, S l'
If the specimen is subjected to an alternating load of the same Referring to Fig. 14.40 and putting the values of S and N, we
range as shown in Fig. 14.39(1), the stress fluctuations will be get,
between curves 6 and 7. The stress distribution caused by prior (A) For as-welded component
overloading should be such as to cause compressive loading at the
stress concentration points during test loading. From (14.16), we get,
The effect of prior overloading can also be demonstrated in SI +K1log N1 =Al ... (14.18)
terms of S-N curves. For a series of specimens all overloaded at and So + K110g N2 =Al ... (14.19)
the same stress the slope of the S-N curve is less steep than that From (14.18) and (14.19), we get,
for corresponding untreated specimens and diverges from it at the
overload stress, as shown in Fig. 14.40 . So SI
logNI-logN2=--- ...04.20)
K1 K1
(B) For Pre-loaded component
From (14.17), we get,
I
I SI + K210g N1 =A2 ... (14.21)
I and S2 + K210g N2 =A2 ... (14.22)
'"
a.
]'-
'- .•...
'"
QI'"
.•...
c..
:::> $2
So
------L------
! .---~~
~-)jgtS_ 100.
From (14.21) and (14.22), we get,
S2 SI
II I
I
o/~(, ~
00. logNI-logN2 =---
K2 K2
...04.23)
: As welded : ~
I I Now, from (14.20) and (14.23), we get,
------1-II I
I So SI 82 SI
I
I
II
I l' K1 - K1 = K2 - K2
... (14.24)
I I
II I:
I•
£.e.,
. SI ( ](2
So 1 - K1
1) = So
S2 ( K2
1) - Kl1 .. ,(14.25)
II where So and S2 are respectively the fatigue strengths of as-welded
iN, ,N2
2x104 and pre-loaded components; while SI is the preload stress and
Log endurance (cycles)
N1 and N2 the fatigue life at SI and So or S2 respectively. This was
Fig. 14.40. Idealised behaviour of specimens subjected to prior overloading. approximately confirmed by Harrison's results which are shown
Increase in fatigue strength has also been reported for fillet graphically in Fig. 14.41.
welds wherein the S-log N curves were found to be approximately 14.8.2.2. Peening
parallel to each other, as in Fig. 14.40.
Peening is a cold working process in which the surface of the
Assuming them to be straight lines, the equation of the weldment is battered either with a high velocity stream of metal
as-welded S-log N curve could be written as, particles or with a tool operated by pneumatic hammer so as to
S + K1log N =Al ...(14.16) induce compressive residual stresses in the surface layer.
--
---------------------------------
.:
3.4
WeldingEng;n••ring & Tedono'og,
.I ----------------------------------
'0'
We'd ne,' •• Fa"g.e Loading
320
975

300
3.2

3.0
I
I
I 260
280
\
; 240 ",.."~- I'-I
:"'-
-<A 'peened
If
2.8
'Specimens with
transverse gussets
'"As welded -
JL
r\
- 220
r..
200
\\
-\,;;-.,.
02·4 \ 1'\
</1
I
I 180
:.
</1
2.2 ,
..

2.0 :'~ 160

1.8 ~ 140
E
1.6
,= 120
?{.
<.1.
.::>
100
L __ I I I I
1.0 1·2 '.4 1·6 '·8 2.0
52/ So
80
105 106 107
Fig. 14.41. Effect of prior overloading on the fatigu e strength of steel
specimens with non-load carrying fillet welds under puls.'lting tensile loading. EnduranCE: ( cycles)
Fig. 14.42. Effect pf hammer peelling on the fatigue strength of mild steel with
Although peening is perhaps the most wdl-known method non-load carrying transverse fillet welds.
available for improving fatigue strength and hols been used quite
extensively for components like leaf springs, axles and Peening effect induced by the use of explosives on or above the
crankshafts, its application to welded components has been rather surface of the weldment has also shown beneficial results hut the
limited. However, experimental work has indicated that fatigue conditions have to be just right for the method to be used
strength of welded components can improve from 15 to 100% for successfully.
a fatigue life of 2 x 106 cycles. 14.8.2.3. Local Compression
In general the results show that peening increases the fatigue It is a method of inducing compressive residual stresses by the
strength of butt and fillet welds both of transverse and use of dies. When load is applied to the plate through dies, the
longitudinal types. Peening with a pneumatic hammer with solid material is squeezed between them and thus a radial compressive
tool, the method that produces the most severe deformation, gives stress is set up between the material being compressed and that
a substantially greater improvement than either shot peening or surrounding it, as is show-nin Fig. 14.43. If the load applied is
peening with multi-wire hammer.
Experimental work on transverse fillet welds has shown that·
for a fatigue life of 2 x 106 cycles there was an increase in fatigue
strength of about 20% and 85% respectively for peened
components with and without subsequent stress relieving.
The S-N curves, shown in Fig. 14.42, for as-welded and peened
components tend to coincide near the yield stress of the material
and can be interpreted as an evidence that the increase in fatigue
strength is largely the result of beneficial stresses induced by ....
peening. Fig. 14.43. Principle of local compression treatment.
976 Welding Engineering & Technology
Weld Design for Fatigue Loading 97'7

sufficient a state of general yield can be induced; surface is thus service stresses, as shown in Fig. 14.45(a). In this way the notch
indented and the material is squeezed out radially and the is subjected to the maximum possible radial residual compressive
compressed material is left in a state of residual compressive stress.
stress. An approximate theoretical analysis has shown that. the
residual stress distributton is of the form shown in Fig. 14.44. This process is not suited to notches that are not localised,
such as transverse fillet welds. It is because only a part of the
0.4 length of the weld could be subjected to the residual compressive
stress while other parts would lie in the balancing residual stress
field, as shown in Fig. 14.45(b).

-- \ ."
320 \steelT.S.fYSa276HPuI
~A-high m..-IIlS. \I -44with- I
r
6.
\~ \ ~\ ~
'" "\ 300
260 \ "-
-4.~"".
..•••
~, ~ •die- 19
pressed
.•.•...•.. "" ......•...

\~r...iJ.mm lir--..
die. dies \
44mm ..••..
,..
• with
- pressed
..•..•.
s. pressed
mm die. with
dies

~I 0 As welded ..•••.••
"
......•...

VI
•.•• <11 4 R 260
.•....•...
III
<III'"•••.••
'" "'-240
:;,
-0
-
0/-0-0.2
<11
.••... -- --
E
~ 220
z200
'gj >=
0::
-0.4 " t' ... ::::180
<II
•...
-:;; 160
.•....
- 0.6 I
~ 140
<-
<11
- 0.6 0.120
§-
Fig. 14.44. TheoreticaJ residual stress distribution due to
local compression treatment. 100
For successful use of this technique it is essential for the notch
to be localised. Such notches at spot welds or the ends of 80
longitudinal fillet welds are ideally suitable for treatment by this 105 106 107
method. In the case of a spot weld the compressed area can Endurance I cycles)
encompass the whole notch while in the case of longitudinal fillet

I
Fig. 14.46. Effect of local compression treatment of mild and high tensile steel
weld it can be so positioned that the line joining the notch to the specimens with non-load carrying longitudinal fillet welds.
centre of the pressed spot is parallel to the direction of applied

Notch to
\' \,
,+ ')
I
I rI +
\

,,'
,,
I '
Fig. 14.46 shows a comparison of the fatigue strength obtained
by treating similar mild steel and high tensile steel specimens by
local compression method. It is apparent from the figure that local
compression treatment results in a larger strength increase with
Pressed high tensile steel than with mild steeL At 2 x 106 cycles the corres-
,,--\..I spot f" spot
ponding improvements are 109% and 74% respectively. This is
. be treQtl!~r)..._r Pressed
, ...:. '1-_ , Residual
"', ... ,, because it is possible to induce higher residual stresses in the high
+ I
I• ) stress due / tensile steels as the residual stresses are directly proportional to
, ..:-- to pressing,
applle
UirecdO
ti nstresses.
of -J (0 )
II ( 1
(b I
I the yield stress of the material. Like in peening the S-N curves
for as-welded and the ones treated by local compression tend to
coincide near the yield stresses of the relevant materials and
Fig. 14.45. Local compression treatment methods for; diverge from each other at lower applied stresses. In general this
(a) longitudinal, and (b) transverse fillet welds. behaviour is common at all the techniques adopted for
978 Welding Engineering & Technology 979
Weld Design for Fatigue Loading
improvement of fatigue life of welded components by inducing For the success of spot heating treatment it is imperative to
compressive residual stresses. select the right spot for heating. Fig. 14.48 shows the wrong and
This method of improving fatigue life of a welded component
320
is not popular because it requires a large load to be applied to the 300
dies; as to obtain a general yielding the applied load must be at 280
least equal to a compressive stress on the die area of twice the 260
yield stress of the material being treated. Also, it is necessary for 240 \.
both sides of the plate tobe accessible for applying the load and ~220
that may not be always possible.
§ 200"" '"(bl
In view of the high. load requirement local compression is a
method which is more suited for use with aluminium alloys than
with steels.
(a)Wrong
"-

=
Ie) \ ~r\~ I "- ",;;-
\1"\
,A..,
180 ~ 44 mm~ I-
\ e \.=== • \
(b)Corre:cI44mm
VI
:c
L-
160
. e
,"
e- -1
~~::::~::::f~~m
14.8.2.4.Spot Heating "t; 140
.•...

Spot heating is a method of improving fatigue strength by :~

producing desirable residual stress distribution by local heating -: 120
usually with a gas torch. It is similar to local compression in that a.
Q/
a.
the induced residual stress distribution is self-balancing over the ;:) 100
width of the cross-section rather than through the thickness,
however the region actually treated becomes an area of residual
80
tensile, instead of compressive stress. Hence, the useful
compressive stresses are generated farther away from the centre
of the heated spot and their distribution is of the form shown in
Fig. 14.47. 60
+oy
105 10' 107
'Endurane (cycles)
Fig. 14.48. SoN curves for (a) wrong, and (b) correct spot
co selection for spot heating treatment.

c
.;:;;
correct spot selection for spot heating treatment. By using the
~ correct spot, Fig. 14.48(b), an increase of fatigue strength of 53%
~ 1 Radius was reportedly obtained. The heat is applied from both sides with
c...- oxy-acetylene torch fitted with high output nozzle to heat a spot
~ . of 50 to 60 mm diameter for about 10 seconds which usually
§
'gj
Q/
corresponds to a temperature rise of about 260°C in 12.7 mm thick
L-
a. steel plate.
Centre of heated spot
~ Spot heating is also found to slow down the propagation of
fatigue cracks that already exist. The selection of spot is exactly
Fig. 14.47. Qualitative residual stress distribution due to spot heating. similar to that done for increasing the fatigue strength.
The effect of spot heating on S-N curves is similar to other Two main difficulties in the use of this method of spot heating
treatment methods involving the use of residual compressive are to decide precisely where the heat should be applied and what
stresses i.e. it decreases the slope of the SoNcurve; the intersection temperature should be attained to achieve the maximum gain in
of the S-N curves for treated and as-welded specimens occurs fatigue life or fatigue strength. These difficulties are overcome by
approximately at the yield stress of the material. a spot heating method known as Gunnert Method.
 980 Welding Engineering & Technology Weld Design for Fatigue Loading
______________ ._. ...
_0 91U
_

Gunnert Method appears that no definite methodology is yet available for gell(~ral
treatment of components.
In Gunnert method of spot heating, the heat is applied actually 14.8.4. Comparison of Improvement Techniques
to the notch point till it causes plastic deformation but it should
not f~xceedthe material transformation temperature, for example Figs. 14.50 and 14.51 compare how various improvemellt
the suggested maximum temperature for steel is 550°C. After techniques affect the fatigue strength of transverse and
heating- the surface rather slowly to achieve a temperature rise
" 320 , ,{- 0,
')",'.c
f-
r "--1· Hig" - .-ground_
lensile steel .._
",, 0"- I \-LU-- t-.., I "'- .-t-..
IIdresstd -1/ .--
:/ "-T1G
Overloaded2
over a significant region of the underlying layers, it is quenched
by spraying a jet of water at it. In this way the surface is cooled I
'. --
\ ""\

...
\- \ ,'<.... 'ti:d
\ -300
r,IJ~~~I\~d
~
!
at232~/m'"
280 " vl;S "'l!lded
-I IfullV
.•• To.
\111
disc
burr
(~i-=-=
instantaneously but the underlying layers cool more slowly so 260
that, as they shrink, compressive stresses are developed in the 240
previously cooled surface layers. The increase in strength at a -( _ 220
fatigue life of 2 x 106 cycles was reported to be 37%.
In another technique of relieving stresses by the application
of heat, the heat is applied in .two bands to either side of the
N

,
z
~ 200
180

longitudinal weld. The heating cycles produce residual st.resses in ~ 160

Q.I

exactly the same manner as during actual welding. The positions -t; 140
of heat bands are chosen such that the new residual stresses ..•.•.

'E
counteract. and cancel the original ones due to welding as shown :=; 120
•...
li.
Weld Il
Heat
QI
Q.
g 100
I
",,}.
'.C..
-- • "".""""""'."",,,
---Heat Weld ®,
~.~_
._.~
"-"
- '-"::'--.
~_
- ..
. 80

Fig. 14.49. Application of heat bands to reduce welding residual stresses.

in Fig. 14.49. Such a process may not remove the residual stresses
completely but if applied sensibly can reduce them to considerably
low level.
14.8.3. Protection from Atmosphere
Improvement in fatigue life of welded components can also be
achieved by the use of protective coating to avoid corrosion of
weldments, one such method is the use of Plastic Coatings.
14.8.3.1. Plastic Coatings
The latest technique suggested for the improvement of fatigue
strength of welded components is the use of plastic coating
cont.aining some filler. The improvement in fatigue life by applying
plastic coatings is suggested to be primarily due to interaction
between the carbon chains in the plastic and the surface of the
metal i.e. by preventing corrosion. Though it is reported that there
is considerable improvement in the fatigue strength of steel and
aluminium alloy components by the use of plastic coatings but it
60
105 106 107
Endurance ( cycles)
Fig. 14.50. Comparison of improvement methods for mild steel specimens with
transverse non-load-carrying fillet welds.
longitudinal fillet welds respectively. After being subjected to prior
loading the mild steel specimens (except with TIG dressing) were
reportedly stressed at 232 N/mm2 before fatigue testing. The m.ain
conclusions drawn are as follows :
1. It is impossible to produce very large improvements in
the fatigue strength of low strength joints, particularly
in structures which are designed to withstand a large
number of load cycles (~ 2 x 106).
2. The most satisfactory technique to use depends upon the
stress level to which the structure is to be subjected. If
the stress is low and the number of cycles is large, one
260 E
-
a.
982
<-
'E
:<-E
<II
•..
"t:i
:.=
:3"
<II
300
_ ~.•...
240
220
2801
200•\
:l 140
160
120
100
180
Weld Design for Fatigue Loading 9830

320 14.9.1.Failul"e of a Steel Pipe Element

Fig. 14.52 shows a sketch of a pipe element consisting of a
bend in a tubing of 10 mm wall ~hickness,with a flanged connector

,
Overhanging
valve

Welds b

80
-~==1-
6 0 l.....u.1L I I II
105 106 107
Endurance (cyc~s)
Fig. 14.51. Comparison of improvement methods for mild steel specimens with
longitudinal non-load-carrying fillet welds.
of the residual stress techniques is likely to be the most
beneficial; but if the stress is high and the endurance is
relatively short, local machining is likely to prove more
satisfactory than any of the residual stress methods.
3. Wh.en a joint is subjected to a loading spectrum contain-
ing both high and low stresses toe grinding and TIG
dressing may be relied upon; however if the spectrum
does not contain any high stresses, the residual stress
techniques can be very useful.
4. The difference between local compression and spot heat-
ing is very small and peening appears to be better than
either.
14.9. CASE STUDIES
Fatigue failure of welded industrIal installations is not
uncommon. Two such failures are analysecl in this section.
Fig. 14.52. A steel pipe system that experienced fatigue failures at joints c and d.
piece welded to each end of the bent tube. The part was attached,
by two flanged connecting pieces, to a bottle-shaped container and
a heavy overhanging valve weighing 95 kg; The working pressure
was 42 N/mm2, and the operating temperature was 105°C. The
part was subjected to vibrations due to compressors installed at
intervals along the pipe system. The part failed after a service
time of 2500 hours.
Examination of the failed part showed two fractures; fracture
A at position c and fracture B at position d. Both fractUl"es
exhibited radial lines and striations. Analyse the possible causes
of failure ..
Failure Analysis
Examination of the failed part revealed that fracture A
originated in the small craters accidentally formed during arc
welding, at the plane of the junction between the cylindrical and
conical portions of the connector attached to the container.
Fracture B started at the junction of the reinforcement ofthe weld
attaching the connecting piece. The radial lines and striations on
the fracture surface were evident of fatigue failure.
No chemical analysis of the steel composition was available
but microexamination showed that the steel had ferrite-pearlite
type structure and the grains were fine and equiaxed. Many cracks
were observed in the fused metal in the craters and in the
martensitic structure of the heat-affected zone. It appears that
these cracks acted as the starting points for fatigue under the
984 Welding Engineering & Technology Weld Design for Fatigue Loading 985

vibrations caused by the compressors. As for the fracture B, the It was further suggested that a new stress relief method be
notch effect due to reinforcement of the weld with the connector employed to avoid the use of similar tack welds.
determined the position of the crack i.e. at the edge of the weld. 14.10. SOLVED PROBLEMS
14.9.2. Failure of a Ball Mill A few solved examples for determining fatigue life of welded
Fig. 14.53 shows part of a ball mill used for are crushing. The components under different service conditions are included in this
cylindrical shell of the ball mill was 10,300 mm in length, 4400 section.
Old NOw
Problem 14.1. Test data for a butt welded component indiccr..tes
~_._- 10300 - -1 a fatigue life of 1.55 million cycles when the member is stressed to
210 N/mrn2• What would be the fatigue strength at a life ·of
2 x IDe cycles?
Solution. Refering to = (14.5), we have11.

.F 1 =F..,•. (~-JK
11.1

Now,/<\ = 210 N/mm2 1.55 X 106 cycles

11.1 ==

~oo-j
F2 ==? 11.2 = 2 X 106 cycles
:,~%
Putting these values in the above equation, we get,
~~/
,,,., ~ _l
•....

~27.50;?'~~..-;~- .- -r
", (dl Detailed view of joint (bl
( b) (c I 210=F'J(" 1.55 x 106--r
2x106
Fig. 14.53. Failure of a ball mill; (a) overall shell design, (6) fatigue cracks, Now for butt welds, K =: 0.13,
(c) suggested improved joint design, and (d) detailed view of joint (b).

mm in diameter, and 50 mm in thickness. An inner flange with .. F2 = ~1..2.-.

__ :: 210 (1.55)°,13
100 mm thickness and 200 mm width was welded to each end of
--~~-}. 0 -6' I .
the cylindrical shell, as shown in the left half of the figure 14.53(a). (.1.55 x 10 )
6 )0.]3 2
The drum had been protected on the inside by bolted impact plates = 203.1 N/nnn2
made of wear-resistant castingE.\with intermediate ;rubber layers. Ans : Fatigue strength:::: 203 N/mm2
Both the cylindrical shell and flages were made from a killed steel Problem 14.2. An overhead travelling crane has a transverse
with the composition of 0.19% C, 0.25% Si, 0.65% Mn, 0.025% P, load carrying fillet weld (class F) and is being used at a reinforced
and 0.028% S. The entire weight of the mill charge was concrete plant to do the following three basic lifts.
approximately 320 tons, and the drum was operated at a rotational
speed of 14 r.p.m. After about 3000 hours of operation, long cracks - stress at Constant
appeared on the outside surface of the drum. Analyse the causes - TI----------·-·---·-T----~--
1\7~ welded amplitude rife
of failure and suggest remedial measures.
S. Movement I Load I Lifts /
----
detail
88 N/mm2
Failure .Analysis 1.4 X 106 cycles
of concrete to casting bay
The drum was emptied, its interior inspected, and .cracks
ranging from 100 mm to 1000 mm long were observed. It was 2. I Lift and transport empty
I skip back to batching plant
J------r--
1 5000
tOll 8 N/mm2 I > 108 cycLf!s
found that these cracks had originated from the nearby tack welds,
shown in Fig. 14.53(d), which had been made for holding
___ Lift rompl.ted -R C.C. -b•• mJ' 30
3. 1.~~~~rans~~!~.!:... 11 ~ton
Ion I 5000
2000 .--1. > 75,000
j 240 NImm'1 . qyo!••_
insulation during stress relieving of the drum in the course of
manufacture. Apparently, the high notch effect had greatly Determine the life of welded joint.
reduced the fatigue life of the drum. As a remedy it was suggested
that the joint design be modified to the one shown in Fig.14.53(c).
 .",-.-"-""~,,~,,,"~-~,.,,."_---------------------------------------------------------------------------------------
986 Welding Engineering & Technology Weld Design for Fatigue Loading 987

--
Solution. To calculate the fatigue life under varying load u; ~
]
cIi

cycles, use is made of the linear cumulative damage rule vide

~ c:.l
>. <:::>
<:::>~
II <:l
.;;; It:>

t::
N
.....
e'"
....<
i.O
.,..;
.....
§.....
~
co
§ ...•
which if for a stress range of 81 the constant amplitude fatigue c:.l ~ C'l
It:>
<C
~ "<l'
'"
"<I' C'l -.;0

ll::~ co
. n o...•
<C
~~
life is N1 cycles, then nl cycles will use a fraction of z../1 of the X ~~lr,) ~.,
fatigue life. In a stress history with stress ranges of 81,82,83",
CN
..•..
as
"'"
":>
II
.....
I~
II <:l
u;
t::
~
.....

<C
10
e~
.,..; e~
t::"
§
.....
~
co
§ ....
C'l
'" '"
..... "<I' C'l -.;0
etc., occurring nl> n2' n3 ... etc. times respectively failure will occur ~
CIl C/J'"
ll::~ It:> .....

when, Ql
a ~.,
't:l .-:;-
<:::>~ § ;:::0
C § 10
e ;:::0
..... § ~ §
n L N'= 1
n11 + Nnz2 + Nn33 ... N:: ~ ~
N nn '" n ...(14.26)
This is kl1.0Wnas Miner's Rule or Linear Cumulative Damage
-Ql
~

Ql(Q~
(Q
~
II <:l
ll::~
.....
"<I' ~
<C
.....
C'l It:>
<C ~
<C
.....
"<l'
co
C'l
....
~

Rule. "t;; ~~
~., M
Using eq.(l4.26) for the travelling crane, we get,
$.0
.s
QC)
~
II
.....
I~
II <:l
§ M
.....

o'"
10
8 c
,....;
M
C
o'"
§ ~
co
~
....
~ .....
"<l'
co
.....
"<l' C'l ~
L !!..N :: 1.45000 ~
"<I'
cr.t' C'l
\I;l ll::~
+ 5000.+ .75,000
x 106108 .1>.0Q9_ rii
CIl

552 ~
..•..
CIl

:: 1400 + 105 + 75 :: 0.0303 llIl

cx:::I
.8
c2::J
This is the part of life consumed in one year. ..•..
as
eQl
.. Life of welded joint :: r\ r\~M = 33 years
Ans: 33 years.
Problem 14.3. A transv'erse butt welded structural steel plate
~
=
Q
't:l
e e
Ql
with width w = 50 mm and thickness t :: 8 mm is subjected to an CIl
as
axial load Pmin = 2 KN. What value of Pmax do you recommend so ~
that the welded detail has a minimum life of 106 cycles for: CIl

(a) full penetration as-welded;

(b) partial penetration as-welded;
..cl
biJ

~
= ~ B
s::!
<:l
I~
'"

!
..•..
(c) full penetration with reinforcement ground flush. 00
~ '"
Use may be made of the Design Fatigue Data given for steel A ~ ~ II)
...•..
.•
36 in table 14.1
~
as
~
",'
'" ~::.s
so::..c:

'a '" 0
Q)

Solution. ",' ",' ~ '" w ~~

. ..c:
"cf ~~
2l 1ll
[Q (iJ..-a •
~ w ~ - - Cl <.> ..c:a~a>
Plate width :: 50 mm ..• ~
-

~ ~..c:
'"
<1)

§
~ .-
...- r
.... ~ <;:l ~ 0Ql ..c: .';::
~ g ~
U Q) bill
Ii!
Plate thickness = 8 mm
CIl
~ :t:l] •..•~ •.•1;'~ui
Q
Ql
~ g
..5:.0
~§r~ !J~
~"tr)
..Q ..s::s
o ~
Q) °0004

i::'
rn;:r '"'d -=-
a~wC,)
en ~~
a::::(tj-'"0 0I::
e::::

.. Cross-sectional area of plate:: 50 x 8:: 400 mm2

..0 "J:.1 '''''Cj C\1 - '" <;:: ••• tiS ::.- CiI
... (i) ,..;
~.B'"O
<ll Ol Q <ll I:: .3 ~~
-
or;;
...;. ••• <ll <ll ;§1l ..-"'0
P min :: 2 KN (Given)
- ::l", --a~at]
""Cjc;S ~~ :: ••••
...• <ll I:: '" ....
Ql
~ <1) a> r;;~6h r;;C\1~ .- <ll "_
b.D"1:t~"
en 0 """Q..
;<..s::~':Q
••••• <:l

§.::~ §~~ §i: ~ b.O~ ~,g~ ~

S . ::.P min = 2 x 1000 = 5 N/
mw Area 400 . mm
2 ...(ii) .:J
~
?:;§~ ~
I::
o ,
<Jl
Ol j ~ ?:
d'Q)
8 r;:'~c.S
cdS:: <D~cd •.
-t.S~~
rz.. r.h ••.0

....<
~~g~~~ll
---~~ -~ ~ -~ ~ ""
- - -~-~
Bll
lei to
.9 ~~ ~
~
g
~0l1~0l;,a;;
t<l <1>-] ~
00
 988 Welding Engineering and Technology ,'I Weld Design for Fatigue Loading 989

Now, the loading pattern for all the three cases can be depicted 51.4 N/mm2
==
. by Fig. 14.17. Thus, referring to that figure we have,
P max = F2 x area of cross-section = 51.4 x 400
. 8mell -- Smin Smax + 8min
Alternatmg stress 'aS ==----.--.----- 8 ==---------2 ::::20560 ::::20.5 KN ..,(vi)
2 'Tn
P max:::: 20.5 KN
or 8max == Sm + Sa (c) For Full-Penetration Joint with Reinforcement Ground Flush
and 8min = Sm'· Sa
By referring to table 14.1, So:::: 76 N/mm2
As Smin is close to zero therefore data for R =" 0, in table 14.1,
can be taken as valid. .. 8m == 8min + Sa = 5 + 76 = 81 N/mm2
(a) For Full Penetration as-welded Joint and 8max = 8m + Sa == 81 + 76 == 157 N/mm2
Sa:= 41 N/mm:!' (from table 14.1) To find the value of Sm-ax for 106 cycles, we have,
Now, 8m =8min +Sa
.. Sm=5+41=46N/mm2
and 8max:= 8m + Sa:::: 46 + 41 = 87 N/mm2 ... (iii)
F2==F1 N~
(N r
= 171.8 N/mm2
(
=157 2~0~0 6)0'13

This value of Srnax is for fatigue lifE: N) of 2 x 106 cycles. To find P mQX == F x area of cross-section

the value of Smax for fatigue life N2 of 10" cycles, we can use == 171.8 x 400 = 68720:= 68.7 KN ...(uii)
= n (14.5) i.e., Pmax:= 68.7KN
Ans: P mae" := 38 K, 20.5 K, 68.7 KN
F,~F·(~ru(iV) Problem 14.4. A wide flat steel (UT8 = 400 MN / m2) tie-bar
Now, K == 0.13 for butt welds, and F1 = Smax· with a cross-section shown in Fig. 14.54 is loaded 120 times per
day from zero to 180 N/mm2• After 8 years of'service it is found
Putting these values in (iv), we get, that a fatigue crack has developed from the sharp machined corner
/
. ' 0,13
./
F = 8'7 _x__ = 95.2 N Imm2
2 .l(2 106106)
.. P m= == 1"2x area of cross-section
== 95.2 x 400::::38000 N = 38 KN ...(v) (ol
Pmax==38 KN
Weld
(b) For Partial-Penetration as welded Joint
By referring to table 14.1, vve get,
Sa;= 21 N/mm2
.. Sm:=Smin+Sa:.-::5+ 21=26 N/mm2
-~-
((I
Detail under consideration
and Smax == 8m + Sa == 26 + 21:= 47 N/mm2

For finding the value of Smax for 106 cycles, we have,

-~====~====3-
Ti e bar
F 2 == F\ (N
IV~JK =, 47l~~o1f-)
( 61°.13 Fig. 14.54. Cross-sections of flat. tie··bar ; fa) with a fatigue crack at corner,
welded, and (d) on overall view of a tie bar.
(6) fillet welded, (c) groOVE!
 990 Welding Metallurgy and Joint Design Weld Design for Fatigue Loading 991

as shown in Fig. 14.54 (a). Surface fillet welds, Fig. 14.54 (b), are Strain range versus fatigue life (E-N) curve is shown in Fig.
applied to affect a temporary repair. How long might they be 14.55.
expected to last? Would full penetration welds, Fig. 14.54(c) be Using the e-N graph of Fig. 14.56 the life for component with
better? Take E = 200 x lei N I mm2, and stress concentration factor strain range of 0.207% will be, as marked on the figure, equal to
= 2.3. Assume 300 working days in a year. 2.8 x 105 cycles which corresponds to a life of,
Solution. L'fi . years = --
I e In Life in cycles
S No. ofloadings/day x No. of working days per year
(a) Strain range, % = stress concentration factor x Ii x 100 2.8 x 105
= 120 x 300 = 7.78 years
(180 - 0\ x 100 = 0.207%
= 2.3 x ~--~ Thus, appearance of cracks after 8 years of service is as per
200 x 103 expectations.
(b) When the cracked tie-bar is fillet welded as per details
0·6 given in Fig. 14.54(b), then,
[Log i -~Of A]
R of A - Reduction of area
Leg length of each weld =
20 - 10
2 = 5 mm
{H Throat length of each weld = 5 x 0.707
Strain Total throat length for two welds = 2 x 5 x 0.707 = 7.07 mm
range . S max X min cross-aection
{log I Max. stress In fillet weld = - ot a I we ld thr oa t area
____ summation r:t1th. curve is = 180 x 10 x t = 255 N/mm2
'~FatigUe
,_ st 7.07 x t
Slope = 0.12 straight I'Ines
ndlvldual

Slope~
-- where, t is the thickness of tie-bar.
If this weld detail is treated as class F (see Fig.14.16) weld
then its life as per S-N curves of Fig. 14.57 will be 6 x 104 cycles
1 Cycle 10 - Life. cycles ( Log) and the allowable design life will be 40% of it i.e.,
Fig. 14.55. Strain range versus fatigue life curve . .. _ Life in cycles
~
-' LIfe In days - N 0.0 fl oa d'Ings per
. d ay x 0.40
o~ -~
b=>

o 6 x 104
'"
C'>
o
o - 120 x 0.4 = 200 days
oc:L. .."
.•..
=>
o (c) If full penetration welds i.e. groove welded butt joint is
c: ~C

S
a
•...
"-
VI

0> 1.0

c; 0.5
---1-
Cf
"-
E
made, as shown in Fig. 14.54(c), the failure will take place at the
weld toe, as shown. In case it is taken as class E weld detail (see
Fig. 14.16) then its fatigue Bfe can be expected to be based on a
weld throat of 10 mm.
c:
L. 04 _____ . I un. 600 M••lm2
2! Fatigue strength lor Max. stress on weld throat
« 0.2
0.310.207 ASMElIl:::-- I o:~~o-u~. 760M••lm2
_ Smax x area of cross-section
<;ur . gn ~ of A 54·
014 011oyV~ s
forluls
corbon ---=-
ond YLT _~A64·I.
400"1 •• 1 m2 -- -- I. - weld throat area
0.1 2
10 103 104 2.8 Ala' 106 = 180 x 10 x t = 180 NI 2
Life(cycles)
10xt .mm
where t is tie-bar
",
thickness.
I<'ig.14.56. E-N graph for high-strain low-fatigue life components.
 •
Weld Design for Fatigue Loading 993
992 Welding Engineering & Technology
Problem 14.5. A stepped carbon steel shaft (see inset in Fig.
14.58)- has D = 55 mm. d = 45 mm, p = 5 mm and is loaded in

-r----a
BOO
- . --- -·-i----- 3.0

----t--
-----·-----t---·------
-!---rr--1- t--- 2.8
.---.-.---.----.~i ..----.- -
--r--,
.
--t-- r-
--~--------··i----t---l---t . p-~-p
c...
.e
u 2.2

".255~~~i-fifl:-
300 .E
co 2.0
:;::
~ 1.8
'-.. 200
z 160
~ 1.68
~ 1.6
ou
+
-"=
~ 15
c;
'-
C1i ~
Q/
1.4
c...

I---h-,
'" 1 00 Vi 1.2
QJ

~
.2> 80
'"
u_ 60 ,., 0·16 0·20 0·24 0·28
40
Fig. 14.58. Stress concentration factors for stepped shafts in tension.
tension 100 times per day from zero to 175 N/mm2• After 4 years
of service it' is found that a fatigue crack has developed at the step.
7 8
Surface fillet weld was applied all around to affect a temporary
10) 10 10 repair. How long the shaft is expected to last ? Take E =
210 X 103 Nlmm2 and assume 260 working days in a year. Use may
be made of Fig. 14.58 to determine the stress concentration factor.
Fig. 14.57. SoN curves for different classes of welds in BS 5400 steel. Kt. Take safety factor as 2.5 for determining the final expected life.
Referring to Fig. 14.57 for a fatigue strength of 180 N/mm2
and assuming the joint to be an E type weld detail, a life of Fatigue life of carbon steel components may be taken as ~rd the
3.6 x 105 cycles may be expected. The corresponding designed life life of steei BS5400 for use of SoN diagram of Fig. 14.57.
at 40% wiIl be, Solution.
L'~ f 1d d '1 Life in No. ofcyc1es d . ~ t Now, D ::;55 mm, d ::;45 mm, p = 5 mm
he 0 we etal = 10a d'mgsper d ay x eSlgn lac or
5 D _ 55 _ 1 22 a.•..d £. - ..Q.. - 0 11
= 3.6 x 10 x 0.4 = 1200 days .. d - 45 -. H d - 45 - .
Referring to Fig. 14.58, Kt = 1.68
. Sr
Life in years, for 300 working days/year ::; 13200~
= 4 years
It is assumed that in repairing the component. the welds are
., Stram range, % = Kt x Ii
x 100
175
properly finished and ground to ensure tbat -cracks do not initiate = 1.68 x 210 x 103 x 100 = 0.14%
from these points.
994 Welding Metallurgy and Joint Design Weld Design for Fatigue Loading 995

From Fig.14.56, N ow, safety factor = 2.5

for %i1e = 0.14, N = 1 X 105 cycles
. Designed life 40
.'. Life of stepped shaft .'. Expected hfe = S a fie t y fiac t or = 2 .5 = 16 days
_ Life in cycles
- No. ofloadings per day x No. of working days/year Ans.16 days
1 x 105 Problem 14.6.A steel link in a mechanism undergoes reversed
= 100 x 260 =.3.85 years bending stress between 250 N/mm2 and 140 N/mm2. Determine
Thus appearance of a fatigue crack after 4 years of service: is the minimum tensile strength of the steel in order that failure will
as per expectations. not occur.Assume the endurance limit for the steel is 200 N / mm2•
(b) When the shaft is fillet welded aU around, Fig. 14.59, we Also, determine the tensile strength of the steel if the bending is of

~rl
have fluctuating type rather than reversed type. Use Gerber's equation.
Weld
Comment on the nature of results obtained if Modified Goodman's
equation is used to determine tensile strength in both the cases.
Solution.
(a) For Reversed Bending Case
Referring to Fig. 14.60(a), we have,
4~ ~

Fig. 14.59. Cracked shaft repaired by fillet weld all around. +

i'
r
Sr
D-d 55 -45
Weld leg length = -2- = 2 = 5 mm '"
VI
QI
L1
Weld throat length = 5 x 0.707 Z
Vl
o Time!

Total weld throat area = 1t X d x throat length

::::1t X 45 x 5 x 0.707 mm2 (0) Reverser1
I b) Fluctuating
8maxx"4d1t 2
Fig. 14.60. Reversed and fluctuating bending cycles .
t
.. Stress on weld throat = To a I we Id th roa area --t-- 8max = 250 N/mm2
1t 2 8mtn ::::- 140 N /mm2
175X"4d 175x45
= 5xO.707xnd= 5xO.707x4 8 __8,. __8_m_ox_-_8._m_tn
__ 2_5_0_-~(
-_1_4_0~)
=557N/mm2
a-2- 2 - 2
::::195 N/mm2
If it is treated as class F2 weld detail then the life as per S-N
+ 8min = 250 + 2
8 m = 8max 2 (- 140) _ 55 N/mm2
curves of Fig. 14.57 will be 6 x 103 cycles. Taking ~rd of which
Using Gerber's equation, 04.11), we have,
gives a fatigue life of ~ (6 x 103) ::::4 X 103 cycles
.. : Life in cycles
... LIfe of welded detaIl ::::L oa d'mgs per d ay days
4 x 103 s.=s+-r~j] ...( i)

"
:::: 100 ::::40 days where, 8f= endurance limit:::: 200 N/mm2
and 811 = U.T.S. of steel ::::','
996 Welding Metallurgy and Joint Design Weld Design for Fatigue Loading 997

Putting the values in (i), we get, 19Q.+ 55 = 1

200 Su
55 195 5 1
t.e.,
195_ 20+ _(~~)'] S u = 1 - 200 = 200 = 40
.. Su = 55 x 40 = 2200 N/mm2
or Su = 1- 200
(55)2 195 = 200
5 = 40
1 (ii) For Fluctuating Bending Case
.. Su = 55·..J40 = 347.85 = 348 N/mm2 (say) As for case (b) above,
(b) For Fluctuating Bending Stress Sa=55N/mm2
Referrilig to Figl 14.60(b) Sm = 195 N/mm2
Smax:;::250 N/mm2 Putting these values in (ii), we get,
2'
Smin = 140 N /mm 55 195_1
200 + Su -
S e: Smax - Smin = 250 - 140 = 55 N/mm2
a 2 2 195 = 1 _ 55 = 145
Su 200 200
S = Smax+ Smin ::::250 + 140 = 195 N/mm2
m 2 2
Su = 20~ ~_195= 268 N/mm2
Using Gerber's equation, (14.11),
Ans. It is evident from the above results that Modified
Goodman's Equation holds good only in fluctuating bending case
s.~s+-(~:n 2
but not in reversed. bending case.
Problem 14.7. The fatigue life of a welded joint on a shop
crane is 20,000 cycles at ± 420 N / mm2 and 1x 106 cycles at ± 280
N / mm2, the S-log N relation being a straight line between these
,.e., 55~20+ - (~:) ] points. Using Linear Cumulative Damage Rule, calculate the
expected life of the joint at ± 315 N / mm2 if it has already received
20,000 cycles at 365 N/mm2 and another 30,000 cycles at ± 335
Su
(195)2 = 1 _ 200
55.:::: 200
145 N/mm2.

.. Su = 195 x -f200 = 229.02 N/mm

-\1145 2 Solution. Mark the two points corresponding to 20,000 cycles
at ± 420 N/mm2 and 1 x 106 cycles at ± 280 N/mm2 on the semi-log
= 229 N/mm2 (say) graph paper and join them by a straight line as shown in Fig.
(c) Using Modified Goodman's Equation, (14.8) i.e., 14.61.
Sa Sm ..
8+S-=1
f u ...(u) Mark on the figure, the points corresponding to-alternatin~
fatigue strengths of± 365 N/mm2, ± 335 N/mm2 and ± 315 N/mm
(i) For Reversed Bending Case and locate the corresponding lives; which are 8.3 x 104 cycles,
We have, from case (a) above, 2.2 x 105 cycles, and 5 x 105 cycles respectively.
Sa = 195 ~/mm2 Now applying the Linear Cumulative Damage Rule, we have.
Sm = 55 N(mm2 !!2 n2 n3
... (i)
N1 + N2 + N3 = 1 at failure
Putting these values in (ii) above, we get,
 l
998 Welding Engineering & Technology

l\ \ \
I- -
I-
r--
1-
I 425 F
I.
410
f--

0;"""" I..
~E 395
E

i
;Z 380

'''' 365
1\ I
~ 350
~
~ 320
335 FIFTEEN
9' 315
:§
go 290
305
Fracture Toughness and
.+::
g 27:
~
280

260
<i 245
230
215
200
Fatigue life. cycles
Fig. 14.61- Fatigue life versus Alternating Fatigue Strength
for weld details on crane.
where, nl = 20,000 cycles, Nl = 8.3 X 104 cycles,
n2 = 30,000 cycles, N2 = 2.2 X 105 cycles,
na =? Na = 5 x 105 cycles.
Putting these values ill = n (i) above, we get,
20,000 + 30,000 + na = 1
8.3 X 104 2.2 X 105 5 X 105
or _1_ + _1_ + _n~a,---- = 1
4.15 7.33 5 x 105
1
na = 5 x 105( 1 - 4.15 - 7.33
= 5 x 105 (0.6226)
= 3.113 x 105 cycles
= 3 x 105 cycles (say)
Weld Design
15.0. INTRODUCTION
A material when it breaks into two or more parts in the
presence of stress is said to have fractured. Fractures that occur
-- 2.i.105 under elastic loading (i.e. below the yield point) and are
accompanied by very little plastic deformation are termed as
Brittle Fractures. Fracture toughness is a property of the metal
that defines its resistance to brittle fracture.
Designing to prevent brittle fracture involves proper selection
of material and allowable stress levels. The allowable stress is
usually a certain percentage of the yield strength for tension
members and a certain percentage of the buckling stress for
compression members. In general it is recognised that welded
joints almost always contain some discontinuities or cracks. Under
conditions of high mechanical restraint a discontinuity can reduce
ductility of the material to such an extent that it can cause brittle
fracture. Also, cracks may initiate under cyclicservice loads or due
to stress corrosion cracking. Welds are, therefore, required to have
some level of notch toughness to perform satisfactorily in such
service conditions. This aspect of weld design has become more
1) quantitative by the development of lracture mechanics which
basically deals with characterising the interaction between
material stress levels and tolerable crack size, to produce a
fracture-resistant design of any large complex structure, for
example, bridges, ships, pressure vessels, aircrafts, earthmoving
Ans.3 X 105 cycles equipment, etc.
Thus, the main aim of the study of fracture mechanics as a
design criterion is to determine the life of a component once a
crack has initiated and to determine possible means and methods
to slow down the propagation Ofthe initiated or existing crack(s).
 1000 Welding Engineering and Technology
Many designers consider the life of a component (in number
of cycles) to crack initiation as its total life, and any subsequent
life spent in crack propagation as only a bonus. Such an approach,
however, is usually unacceptable on economic considerations. This
chapter is concerned with assessing those situations where cracks
have initiated (usually due to fatigue) in components, and it is
necessary to determine the significance of such cracks with regard
to performance, both in terms of safety and reliability with
satisfying functional requirements and to achieve for it as long a
life as possible.
15.1. TYPES OF FRACTURES
The study of fracture toughness requires the knowledge of the
mechanism(s) by which a fracture occurs. Depending upon the
nature of loading and the type of failure the fractures are broadly
classified into two main categories viz., Ductile and Brittle fractures.
15.1.1. Ductile and Brittle Fractures
The terms ductile and brittle are used to distinguish failures
of materials characterised by high and low toughness respectively.
The most familiar type of ductile fracture is by overload by tension
which produces the well known cup and cone fracture. This
happens gradually and is accompanied by considerable plastic
deformation. The mechanism of ductile fracture under tensile load
. normally involves the initiation, growth and coalescence of
micro-voids to form cracks in a transverse direction at an angle of
45° to the tensile load. When observed under high magnification
the fractured surface shows itself as small dimples which
represent the coalesced voids.
Brittle fracture of metals occurs 'by direct separation along
crystallographic planes due to a simple breaking of atomic bonds
accompanied by very little plastic deformation, and it is usually
associated with a particular crystallographic plane. Purely brittle
fracture is also referred to as cleavage fracture. In other words
cleavage fracture is the most brittle form of fracture that can occur
in crystalline materials and it occurs by the breaking of atomic
bonds. Such a fracture across a crystallographic plane results in
grain cleavage facets of high reflectivity giving the cleavage
fracture a bright shining crystalline appearance.
lntergranular or intercrystalline fracture involves pulling
apart of the individual crystals from their neighbours. Such
fractures are caused by the presence of impurities at the grain
boundaries. Intergranular fracture usually represents a welding
defect like hot crack while the transgranular fracture represents
the overloading of sound metal.
Fracture Toughness and Weld Design lOO}
Other common terms associated with the features of fracture
are shear and fibrous. A shear fracture involves failure along the
slip planes where the crystals become distorted and drawn out to
give a fibrous looking fracture surface.
In actual failures it is very rare that truly brittle or truly
ductile fractures will occur because ductile fracture will have some
traces of brittle fracture and vice-versa. However, the predominant
type is considered the main cause of failure, often mixed mode of
fracture are observed indicating the heterogenity of structure an.d
changing conditions of stresses during fracture.
15.2. FACTORS AFFECTING TYPE OF FRACTURE
The type of fracture (ductile or brittle) produced depends upon
many factors such as material composition, microstructure, grain
size, strain rate, extent of strain hardening, material geometry;
size, shape, and type of notch; residual stresses, welding process
employed, shielding· gas composition, and the weld joint edge
preparation. The salient features of each one of these factors are
discussed briefly in this section.
The change in material composition r'esults in the development
of different microstructural phases with varying mechanical
properties resulting in change in fracture mechanism. For
example, the amount of carbon and other alloying elements
influence the type and extent of different microstructural phases
such as pearlite, bainite, and martensite with the consequential
effect on the fracture tendencies.
For metals in general, the microstructure grain size is a strong
determinant of notch toughness. The grain size has a marked
influence on the yield stress and strain hardening rate and it is,
therefore, logical to expect grain size to influence fracture
toughness. Microstructures having finer grain size are known to
possess high mechanical strength with consequential effect on
fracture toughness. The nature of grain boundaries may also affect
the extent of grain hardening of a material which can influence
the mode of fracture.
Susceptibility to brittle fracture could be induced by triaxial
stress state like imposition of high strain rates. A high speed of
testing generally induces high strain rates, hence the risk to
brittle fracture is enhanced. Rate of strain harden~ng induces
multi-axial stress condition within the material, thuf3 increases
the possibility of cleavage brittle fracture.
Both geometry of cross-section and its size can influence the
nominal strength of a material. Indeed most of the materials show
marked size effect, for example, large sections are more prone to
 1002
Welding Engineering and Technology
brittle fracture. The deterioration in strength of thick sections has
been attributed to increased amount of impurities, flaws, and
discontinuities present in thick section. Profile shape also'
influences the fracture toughness ofa welded joint by the presence
of notches and areas of stress concentration.
The heating and cooling cycles of welding result in the
development of residual stresses. The stress level in a weldment
may be large enough to reach the yield point of the material and
may lead to brittle fracture in the presence of cracks. The
combination of welding residual stresses and material strain
ageing is considered to be a known cause of cleavage fracture. The
residual stresses developed due to welding being a combination of
tensile and compressive stresses; the tensile component of these
locked in stresses act critically at the strained roots of the notches
andbrittle
areas fracture.
of metallurgical imperfections enhancing the tendency
to
Different welding processes result in different rates of cooling
with consequential effect on microstructure, grain size. and
residual stresses. As discussed above, all these factors can
individually and collectively affect the fracture toughness of a
welded joint. A high cooling rate process like MIG welding results
in fine grained weld metal of high hardness and strength with
consequential lowering of impact strength i.e. with low toughness.
On the other hand a low cooling rate process like submerged arc
welding results in comparatively coarser grained weld metal of
medium hardness and strength with consequently higher fracture
toughness.
The shielding gas composition affects the shape and extent of
weld penetration. As the weld metal and HAZ have different
hardnesses thus their tendencies to brittle fracture can be
considerably influenced by the shielding composition.
Type of edge preparation controls the shape of the weld bead
and the number of runs to be deposited thus influences the
microstructure and mechanical properties of the weld metal. The
weld joints between unprepared plates are reported to possess
higher impact values than those having V-edge preparation.
The cumulative effect of all these factors is that a welded
section attains either a plane stress or a plane strain condition
under the given service stresses.
15.3. PLANE STRESS AND PLANE STRAIN CONDITIONS
A loaded component with a pre-existing crack can fracture
under two distinct conditions viz .. plane stress condition and plane
Fracture Toughness and Weld Design 1003
strain condition corresponding respectively to high fracture
toughness and low fracture toughness values.
Plane stress condition corresponds to bulk yielding and is thus
achieved under the following conditions:
(i) ductile material,
(ii) thin section,
(iii) high temperature,
(iv) stresses perpendicular to the outer surface are non-ex-
istent (crz = 0), i.e. the stresses acting on the body are in
one plane.
Under these conditions a large plastic zone is formed at the
tip of the crack requiring expenditure of a large amount of energy
for its propagation.
A crack propagating under plane stress conditions results in
an oblique shear fracture as shown in Fig. 15.1 (a).
o
DD
I aJ Plane stress {b J Plane strain (c I Mixed mode
Fig. 15.1. Fractured body under: (a) plane stress conditions,
(b) plane strain conditions, and (c) mixed mode conditions.
Plane Strain condition corresponds to restricted yielding along
one axis and is achieved under the following conditions:
(i) Brittle/hard material,
(ii) thick section,
(iii) low temperature,
(iv) strains perpendicular to the outer surface are non-exis-
tent, i.e. ez = 0 or in other words the strains produced
under triaxial stresses are along a plane.
Under these conditions the plastic zones formed at the tip of
the extending crack are small requiring lesser energy for crack
propagation than that required in the case of plane stress
1004 Fracture Toughness and Weld Design 1005
Welding Engineering and Technology

conditions. A crack propagating under plane strain conditions
results in a square fracture as shown in Fig. 15.1 (bJ.
A fracture occurring under intermediate conditions is referred
to as mixed mode fracture with plane strain conditions prevailing
inside the section while on and close to the surface it is plane stress
condition, as shown schematically in Fig. 15.1 (c).
Further, plane stress and plane strain conditions can also be
represented by Mohr's Circle however to discuss that it is useful
to first know about different modes of crack extension.
15.4. MODES OF CRACK EXTENSION
A crack in a solid can extend in three different modes as shown
in Fig. 15.2. Mode I is the opening mode, which is the most
common, particularly in fatigue loading. In this mode the
displacement of crack surfaces are perpendicular to the plane of
the crack.
Mode II is the shearing or sliding mode in which the
displacement of the crack surfaces is in plane of the crack and
perpendicular to the leading edge of the crack ..
Mode III or the tearing mode is caused by out-of-plane shear
in which the crack surface displacements are in the plane of the
crack and parallel to the leading edge of the crack.
Combinations ofthese crack extension modes can also occur, Fig.
15.3 shows two such combinations of Mode I and Mode II with single
and double shear. However, because mode I is technically the most
important so the discussion is usually limited to this mode.
15.5. REPRESENTATION OF PLANE STRESS AND
PLANE STRAIN CONDITIONS ON MOHR'S CIRCLE

x x

(0)
"(
Mode III
F mode tearing mode o-mc"

Fig. 15.2. The three modes of cracking.

B
a 0-
Txy

crmin
o-av

-}-c "ic- o-yl

( b)
Fig. 15.4 (a) Coordinate system and stress components ahead of a crack, and
• (b) their representation on Mohr's circle.
Fig. 15.3. Combinations of mode I and mode II of crack extension.
1006 Welding Engineering and Technology Fracture Toughness and Weld Design 1007

Plane stress and plane strain conditions can also be
represented by Mohr's circle as shown in Fig. 15.4.
In order to examine the different deformation behaviours in
plane stress and plane stain conditions, let us consider the Mohr's
circles for mode I crack extension as shown in Fig. 15.5. First note
that for 9 = 0 the stresses O'y and O'x are the principal stresses 0'1
and 0'2' The transverse stress az is always the principal stress
0'3' In the case of plane stress conditions, the maximum shear
stress 't'max' is at plane rotated over angles of 45° from th~·
directions of 0'1 and 0'3 as represented by circle of diameter 0'30'1'
If 0'1 ::::O'y and 0'3 = O'z = 0 (plane stress, 9 = 0) these are planes
through the x-axis subtending an angle of 45° with the X-Z plane,
as indicated in Fig. 15.5 (a). Similarly, circles of diameters 0'3a2
and 0'20'1 may be used to analyse the deformation behaviour as the
element is rotated about the Y-axis and Z-axis respectively.
Y £y
that 't'max is not only much lower than in plane stress, the
maximum shear stress is also on different planes, rotated over 45°
from the direction 0'2 and 0'1 as represented by circle of diameter
0'20'1' If 0'1 = O'y (9 = 0) these are planes through the Z-axis at 45°
from the X-Z plane as shown in F'ig. 15.5 (b). Similarly, circles of
diameters 020'3 and 0'30'1 may be used to analyse the deformation
behaviour as the element is rotated about the Y-axis and X-axis
respectively.
15.6. RESIDUAL STRENGTH
Consider a structure in which a crack develops. Due to the
application of repeated loads or due to a combination of loads and
environmental attack this crack will grow with time. The longer
the crack, smaller the sound metal cross-section to resist the load,
the higher the stress concentration induced by it. This implies that
the rate of crack propagation as a function of time can be
represented by a rising curve as shown in Fig. 15.6 (a). Due to the
presence of the crack the strength of the structure is decreased. it
is lower than the original strength it was designed, and is known
as residual strength. The residual strength of the structure
decreases progressively with increasing crack size, as is shown
diagrammatically in Fig. 15.6 (b). After a time residual strength

x x

:\ilze strength
1 c~ack f te:\iidual
T T - - design strength

Tmax.
- - --service
expected load
highut
I normal
--~--- - ervice load
I I
6) <51 - <1' '<1', - <1' 'failure'
:moy
occur I: failure
-cycles - crack:\iize
(a) -time (b) -time
(a) Plane stress (b) Plane strain
Fig. 15.5. Planes of maximum shear stress for {}close to zero: Fig. 15.6 Relationship between: (a) time and crack size, and
(a) Plane stress conditions, and (b) Plane strain conditions. (b) crack size and residual strength.
In the case of plane strain conditions, al and 0'2 have the same becomes so low that structure cannot withstand accidental high
magnitude as in plane stress conditions. However, the third loads that may occur in service. So, this moment onwards the
principal stress equals v (CJt + 0'2)' For plastic deformation the structure is liable to fail anytime. If such accidental high loads do
requirement of constant volume sets v = 1/2 and thus the stress not occur, the crack will continue to grow until the residual
0'1+0'2. 0'1-0'2 .• strength has become so low that fracture occurs under normal
CJ3 = 2 t.e. 0'2 + 2 as shown m Flg. 15.5 (b). It turns out service loading.
 1008 Fracture Toughness and Weld Design 1009
Welding Engineering and Technology

15.6.1.Effect of Section Thickness on Residual Strength
Keeping in mind that the stress at the crack tip is larger in
plane stress than in plane strain and that the fracture requires a
combination of high stress and high strain; let us consider four
panels of thicknesses Bv B2, Ba, and B4' All panels have the same
length of crack and all panels are loaded to the same stress 01 ;
thus the stress intensities of all panels are equal. Consequently,
the plastic zones in panels are of equal size. This is depicted in
the lower line of Fig. 15.7 which shows the through thickness
sections of-the four specimens. The plastic zones at the stress 01
are indicated by the sectioned or dashed areas.
propagation will occur in the interior of panel Ba. Yet this panel
does not fail due to the effect of the regions of plane stress (lower
crack tip stress) which exist near the testpiece surface and which
are relatively influential in this thinner panel. Panel B2 is also in
the same situation, but this panel has a thickness just equal to
the size of the presently existing plastic zone, which implies that
plane stress has now developed in panel B2•
Further increase of the stress causes failure in panel Ba at a
stress 0a. At 04 the strains at the crack tip in panel B 1 will be so
large that failure occurs. B2, however, does not fail. This is because
the strains in panel B 1 have been larger than those in panel B2
since the stress 01 was reached. Therefore, the strains in panel

-
0'"5
B2 are still insufficient for fracture even though the stresses are
almost the same as in B1• Failure of panel B2 requires a further
residual sfrength of
cracked panels as a increase of stress to 05'
function of thickness Panels of a thickness larger than B4 behave similarly to B4,
'-
'"
Q/
Vi CJ[..

~. and fail at the stress 01, This is the domain of plane strain
f '" fracture, for which valid plane strain fracture toughness can be
"2

0;
J ~ I· KIt dOl1ain
square
obtained. The maximum residual strength is reached by the panel
that develops full plane stress just at the stress at which plane
strain panel fails. Thinner panels have higher strains and fail at

~ ~IB'_
fracture
M1B'~
t--I I----! ~
I' --1
BI B2 B3
Thickness, B
Fig. 15.7. Residual strength as a function of section thickness.
In panels B2' Ba and B4 the height of the plastic zone i:;>still
smaller than the thickness. This implies that yielding in the
thickness direction cannot take place freely, but is restrained by
the surrounding elastic material. Thus, the strain in the thickness
direction is kept at zero i.e. there exists a state of plane strain. In
panel B l' the plastic zone is equal to the thickness and yielding
in the thickness direction is unconstrained. This means that a
state of plane stress can fully develop in panel Bv and the plastic
. zone in B 1 will from now on be larger than in the other panels.
Increase of the stress to 02 will cause failure in panel B4'
because strains and stresses are large enough. Panel Ba is in the
same situation as panel B4, and it is likely that some crack
lower external stress.
A serious objection against the above way of reasoning is that
failure in thin panels is always preceded by slow stable crack
growth and therefore that phenomenon should also have been
considered. Overall picture obtained of the relationship between
section thickness and residual life is, however, quite
representative.
15.7. MATERIALS AND METHODS OF DETERMINING
THEm FRACTURE TOUGHNESS
All engineering materials can be divided into four types from
the point of view of fracture toughness. These are:
1. Linear elastic materials of low yield strength i.e. brittle
materials, for example, glass.
2. Linear elastic materials of high yield strength e.g. HSLA
steels, maraging steels and high strength aluminium
and titanium alloys.
3. Elastic-plastic materials like mild steel
4. Highly plastic materials like lead.
The stress-stram relationships for these four groups ofinaterials
can be represented on relative basis, as shown in Fig. 15.8.
 resista:tce (R-Curve) provides the corresponding energy based
criterion for plane stress conditions.
In the case of elastic materials of high strength a small plastic
x-indicates fracture
I Y.S.~- zone is formed at each end of the crack resulting in resistance to
crack propagation thus resulting in higher toughness of materiaL
VI
VI
GI Fracture toughness in such materials, for plane strain conditions,
.-
VI
•...
( Hi ghly plastic) is determined by Linear Elastic Fracture Mechanics (LEFM)
which is now a very well developed and well established procedure.
The plastic zones created at the tips of the propagating cracks
in elastic-plastic materials (and elastic materials of high yield
strength in plane stress conditions) are much larger than can be
accounted for by LEFM treatment. Fracture toughness for such
Strain -
materials is determined, though with lower level of confidence
Fig. 15.8. Stress-strain relationship for than LEFM, by techniques such as COD (crack opening
different groups of engineering materials. displacement)/CTOD (crack tip opening displacement), J-integl·al
The fracture toughness of ferritic steels exhibits significant and R-Curve.
variations with changes in yield strength level and temperature. Generally materials of very high toughness do not have high
The interaction of toughness with strength level (y.s:) and strength so are much less used in conditions warranting brittle
temperature is illustrated by a 3-dimensional plot of these fracture. No specific method exists for determining fracture
variable in Fig. 15.9. toughness of such high toughness materials. However, such
materials usually obey maximum strain criterion which can be
r Fracture/
energy / Fracture toughness expressed as critical CTOD. Alternative R-Curve method may also
be used in such cases of plane stress conditions.
15.7.1. Griffith-Irwin Criterion
-60'F Fracture mechanics has its origin in the work of Griffith who
(-50'C)
f proposed the principle of energy balance between the strain ene rgy
lost in propagating a crack and the surface energy of the newly
L..
"
'&
•
created fracture surfaces.
€
~ -10~ F
(-75 Cl
Griffith proposed that the brittle fracture of glass occurred by
,!!! the propagation of small surface flaws, and for this to occur the
total energy of the system must decrease in the process, i.e. the
increase in surface energy of the new fracture surfaces must be
less than the decrease in stored elastic strain energy caused by
the growth of the crack.
"80 100 Griffith considered the growth of an elliptical through-
~~/(
,\'l-
550) (6 90) thickness crack in a tensile loaded plate of an elastic material
Note: Figures in parenthesis are approximate. (glass) of unit thickness and infinite width. The plate is stressed
Fig. 15.9. Three dimensional plot showing interrelationship of fracture energy,
to a stress 0' and fixed at its ends as shown in Fig. 15.10 (a). The
temperature, and strength level on fracture behaviour of ferritic steels. load displacement diagram for which is shown in Fig. 15.10 (b).
The elastic energy contained in the plate is represented by the
The crack propagation in truly brittle material like glass area OAB. If the crack extends over a length da the stiffness of
results in cleavage fracture with very little plastic zone, if any. the plate will drop (line OC), which means that some loa!i will be
Fracture toughness of brittle materials is determined by the use relaxed since the ends of the plate are fixed. Consequently, the
of Griffith-Irwin Criterion for plain strain conditions while crack elastic energy content will drop to a magnitude represented by
 ,n

1012 Welding Engineering and Technology Fracture Toughness and Weld Design 1013

t Load 0
E Stored
P6
(+ve)

f1 0
Crack length a
Crack length a+da
Released I Elastic strain
(-ve) I-ase
B re~ nacr2 2

Elongation U = --E-
Fig. 15.10. The Griffith criterion for tlxed grips:
(a) cracked plate with fIXedends, and . ener9~Y
(b) elastic energy release in crack propagation.
area OCE. Crack propagation from a to a + da will result in an Fig. 15.11. Energy changes accompanying the growth of a Griffith crack.
elastic energy release equal in magnitude to area OAC. It can be For plane strain
shown that this release of strain energy (U) is given by, 1rQ2er
'ltera2 R = S - U = 4ays - da - (1- v2) ••• 05.5)
For plane stress: U=~ ... (15.1) dR
. 'ltera2 2 Fracture occurs when da = 0
and, for plane stram U=~ (I-v) ...(15.2) Thus, for plane stress,
where, cr = nominal applied stress;
a = half crack length ; ...(15.6)
E = Your..g'smodulus;
v = Poisson's ratio. fracture stress, err k':'
= (~ Y• J1/2
...(15.7)
Now, the surface energy (8) associated with a central through- or acr = 'Iter
thickness crack of length 2a and unit width (w = 1) is, and similarly, for plane strain,
8= 4ay• .,(15.3)
where, y. = surface energy/area. -:....( 15.8)
The variation with increasing crack length of these energies, fracture stress, crr=. [ 'It (12E
- Y·2
v ) a ]1/2
together with the resultant energy (R) of the system are given in ...(15.9)
Fig. 15.11. This latter curve shows that work must be done on the
plate in order to increase the crack size up to a critical value = ky•
(~:. J/2
(acr)' after which the crack will propagate spontaneously since its or aCr=-2- ...(15. I!)
cr
growth gives an even greater reduction in the total energy of the
system. i.e. cr . crf
a1/2 = (E-vI•)1/2 ...(15.11)
For plane stress This latter quantity i.e. (Ey.)1/2 is a material constant.
1rQ2cr2
R = 8 - U = 4ay. - ~ ...(15.4) In practice engineering materials, even many nominally
brittle ones, deform plastically to some extent and the Grift'ith.
1014 Welding Engineering and Technology Fracture Toughness and Weld Design 1015

Criterion must be modified to allow for local plastic deformation Whilst 1; is fairly readily found experimentally for small
at the crack tip, since this plastic deformation imparts .some specimens, it is not directly applicable to large components and
toughness because of the work carried out. Thus, it has been structures. A more popular parameter which has been developed
proposed that the following modified equation may be used for is the stress intensity factor (K) discussed later under Linear
plane stress conditions. Elastic Fracture Mechanics.
15.7.2.R-Curve Criterion
_ 2E (Ys + Yp) ...(15.12)
crf- [ Tta ]1/2
An R-Curve is a plot of crack growth resistance in a material
as a function of actual or effective crack extension.
where, Yp = a plastic work factor [shown experimentally to be
The Griffith energy criterion for fracture states that the crack
(104-106) Ys]. growth can occur if the energy required to form an additional crack
The plastic work factor Yp is not a materia.l constant, but varies of size da can just be delivered by the system. For a plate of l.mit
with crack length, strain rate and specimen geometry. thickness this condition for crack growth becomes,
d
The essence of the Griffith Theory, i.e. that crack propagation ...(15.19)
da (U-F+W)=O
is an energy conversion process has been extended by Irwin, in an
attempt to find a reliable design criterion for predicting the stress or ~(F_U)=dW
da da
...(l5.19a'
at which a crack will propagate to fracture. This approach involves
a crack extension force, or strain energy release rate, S, which is where, U = the elastic energy contained in the plate,
the stored elastic energy released as a result of crack advancing F = the work performed by the external force,
by a unit area. W = the energy for crack formation.
dU
For plane stress ...05.13) Now, this :a (F - U) = G is the energy release rate or the crack
S= d(2at)
where, t = plate thickness extension force and R = ~: = the crack resistance force. The slow
Now, U = na2a2t ...(15.14) and stable crack growth continues as long as G = Rand
E fracture occurs when G exceeds R.
...(15.15) In a truly brittle material like glass the energy for crack
.. d (Tta2cr2t)
d(2at)
S= E. growth is the surface energy to form the new free surfaces, i.e.
For unit thickness of plate i.e. t = 1 W=2ya
dW .
S= Tta~
E
...(15.16) and R=-=2yda ... C-15.20}

Similarly, for plane strain Equation (15.20) can be evaluated by using 05.16); and 05.20)
2 then yields the following relationship.
Ttacr ...(15.17)
S= E (1 _ y2) ...(15.21)
At a critical value of S, known as fracture toughness (Sc), the w hi ch· IS th e orlgIna
.. crlcG=
rl(2fTta'fiYEtliJC/2't
1 .
n enon.
crack will propagate rapidly. For an elastic crack in an assumed 15.7.2.1.R-Curve for Plane Strain Conditions
infinitely wide plate of an ideal brittle material it can be shown
that If the plastic energy R is the same for every crack increment,
R = dW / da is still a constant. This is proved to be approximately
Sc = 2ys ... (15.18)
the case for plane strain cracks i.e. the specimens containing
indicating that the Griffith and Irwin approach lead essentially to cracks of various sizes appear to fail at the same value of G. This
the same result. critical G value for mode I loading is denoted as,
1016 Welding Engineering and Technology Fracture Toughness and Weld Design 1017
f
• 2 Kic
GIC = (1- V ) IF ...(15.22) r

G ,R f C1j > O""z

Hence, in the case of plane strain,
dW K
R = cIa = GIC •• (15.23)

Thus, from (15.20), (15.21) and (15.23), we get,

H
O'e = ~
(EGrc)1/2
... (15.24)
02
- ..••. R = G1c
__ -- ..... c:r,
.•••. I
The fracture criterion can be depicted graphically as shown in .... "., ....

--
.",

Fig. 15.12. The crack growth resistance R is independent of crack M - L:~-::::'--crz

size and therefore it is represented by a straight horizontal line Oz 01 o -- 6a
R = GIC. The energy release rate is represented by, I OJ

<$1> cf'2 Fig. 15.13. Universal representation of energy criterion.

G = ( l-Y% ) <$l TTa extension f::.a and to the left the initial crack size al' The Glines
E
G.R are again represented by straight lines. At a stress 0'2' for example,
the G line is given by LF. Only point F is a realistic point at LF,
f since the size of the crack is alreadyal' By loading the crack from
zero to 0'2' its G value increases from 0 to F. Further increase of
c stress to 0'1 raises G to point H. Then fracture occurs. Crack
R = G1c extension at the stress 0'1 causes G to follow the line HK. Hence,
/ I /'/'---:;>I G remains larger than R.
/' "....
If a crack of size a2 is loaded from zero to 0'2its G value increases
/' /' /lB from 0 to H (note that LF and MH are parallel lines). At H, crack
~/ I extension occurs: when the stress remains 0'2the energy release rate
o °1 0% -0 follows the line HN, and G remains larger than R.
15.7.2.2.R-Curve for Plane Stress Conditions
Fig. 15.12. Graphical representation of energy criterion for G and R.
So far, R was considered independent of crack length. This is
2 0'27ta approximately true for cracks under plane strain considerations.
GI = (1 - V ) ~ ... (15.;25) However, in the case of plane stress the crack resistance varies
For a particular stress 0'1 the energy release rate is propor- with the amount of crack growth.
tional to the crack size a. Then G is represented by the straight line Consider a crack in a sheet thin enough for plane stress to occur.
OA. If the crack size is ai' the energy release rate at a lower stress When the specimen is loaded to a stress 0'1 the crack starts
0'2is represented by point B. Increase of stress from 0'2to 0'1raises propagating. However, crack growth is stable and fracture does not
G from point B to A. Thus, crack of size al can still extend slowly yet occur. If the stress is kept constant at 0'2the crack extends only
when stress is raised from 0'2to 0'1since at that point A the condition over a small distance and stops. A further increase of the stress is
required to maintain crack growth; although the crack is longer it
G = R is satisfied. A longer crack of size a2 had already reached this can withstand a higher stress. The stress can be increased further
stage at the lower stress 0'2(point C). with simultaneous crack growth at a stress O'e a critical crack size
A more universal representation of the fracture criterion is ac is reached where fracture instability occurs as is illustrated in
given in Fig. 15.13. To the right of the Y-axis is plotted the crack Fig. 15.14.
1018 Welding Engineering and Technology Fracture Toughness and Weld Design 1019
j-'

can be further increased to ai' where the available energy release

rate is given by B. Suppose this value is sufficient for crack growth.
-'-
~

Vl
f
If the crack were to propagate under constant stress, G would
increase according to B-H. This line is lower than the R curve and
K therefore crack growth under constant stress cannot occur.
oc (}= {ifQ Further increase of stress to a2 brings about a crack extenslon
ila2' Both G and R follow the R curve from B to C. Finally, at crc
01
the crack length has become ac and both G and R are at point D.
Crack growth at constant stress ac gives an increase of G according
to the line DF. This line is above the R curve. Since G remains
larger than R, final fracture occurs at point D where,
aG aR
-=_.
aa aaI'
G=R ...(15.26)
OJ ac - crack size
Fig. 15.14. Stable crack growth under plane stress conditions. Apparently, equation (15.26) is the energy criterion for
fracture under plane stress conditions.
At the onset of crack propagation the energy criterion must be Equation (15.26) is a useful fracture criterion if an analytical
fulfilled. During stable crack growth the energy release rate is just
relation for R is available, apart from the relation G---·
7t(J2a
equal to the crack resistance (ifit is lower the crack stops growing, - E
if it is larger fracture instability occurs). The energy release rate
2 . Otherwise the equation cannot be evaluated.
is G = 1t~ a and both a and a increase during crack growth. This Kraft et al have proposed that R is a function of ila only,
means that G increases more than proportional to a. Since independent of ao. Then the R curve is invariant and is the same
G = R it must be concluded that R increases, which is shown for any initial crack length. In many tests the critical crack length
diagramaticalJy in Fig. 15.15. is approximately proportional to the initial crack length i.e.
ac = <:Xao •.. (15.27)
This implies that a certain relation exists between KM and
G,R t LM in Fig. 15.16 and is the same for all tangents. Equation (15.26)
i.e. G = R allows a derivatiqn of a function for R.

G,R 1

-a OJ
I.•
a( a tia2
"I
-tia
R=

Fig. 15.15. The R-curve for plane stress conditions.

SUPPQsea crack of size ai is loaded to stress al. If the crack
were to extend, the available energy released is given by A.
Glc
( plane strain)

-n, I
K
_~
E

G-TT(}("I

J'ti5'-
2

cr:
/ /
./""-

acl
t/
c.•...•.. /
// .••.
6'1
....
A
S'
(
I
II
I
III
:
I
'M
-tia

However, this value is too low for crack growth to occur. The stress Fig. 15.16. The energy concept in plane stress.
 Fracture Toughness and Weld Design 1021
1020 Welding Engineering and Technology

This derivation leads to presumably small in comparison with the energy contained in the
plastic zone. Therefore the R curve must start from zero as
R_ ...(15.28) indicated in Fig. 15.17 ; at zero stress the size of the plastic zone
- [3 (a - ao) (0.-1)
0.
is zero as it requires no energy to form a plastic zone of zero size.
wh~re [3 is constant. 'rhis also means that at any stress different from zero the ellergy
Further, it is known that criterion is fulfilled (apart from the work for void formation).
However, crack growth does not occur, because the stresses and
n~a . plastic strains at the crack tip are still insufficient. The energy
G~ -y- ...(15.29) criterion is a necessary criterion for crack growth, but not a
Equations (15.28) and (15.29) allow an evaluation of the sufficient criterion.
fracture criterion of equation (15.26), The result is, The material at the crack tip is not ready to separate until the
ac . a~/20.= constant (15.30) stresses and strains are large enough to make void initiation and
coalescence possible. When this is the case a fairly large plastic
and ac . aa~/2<x = constant (15.31) zone has already formed. Crack growth can then take place only
Although equation (15.31) follows directly from equation if sufficient energy becomes available from crack growth to provide
(15.30) through the relation ac = <Xao' both equations are obtained the work to form the plastic zone at the neVI crack tip. Crack
by substitution of (15.28) and (15.29) into the two equations of growth cannot occur if this energy condition is not satisfied; it need
(15.26) not occur if it is satisfied.
For a. = 1 the relations reduce to ac . al/2 :=; C. Also, for a. = 1, Apparently, all conditions for crack growth are met at pointB.
the crack resistance R = [3 (= constant. GlC) and there is no slow This is the same point for cracks of sizes a1 and a2' Therefore,
crack growth [equation (15.27)], i.e. a. = 1 for a brittle material. It 2
nail a1
2
nai2 a2
is possible to generalise equations (15.30) and (15.31) for finite Gi1 = E = Gi2 = E ...(15.32)
panel size.
The R curve represents the energy required for crack growth, The onset of stable crack propagation is at a certain value of
In a ductile material this is the work for formation of a new plastic G, namely Gi given by equation (15.32). At Gi crack growth occurs
zone at the tip of the advancing crack, plus the work required for from B to C1 or C2 where G = R. This first discrete crack extension
the initiation, growth and coalescence of microvoids. The latter is is called pop-in, because it is a sudden metastable crack extension,
which is often associated with an audible click. After pop-in
G,R t
gradual slow stable crack growth takes place until fracture follows
at D1 or D2•
15.7.3.Linear Elastic Fracture Mechanics (LEFM)
Linear elastic fracture mechanics is particularly useful in case
of materials which do not exhibit cleavage fracture, but which still I;;

~i
/
, . behave in a brittle manner from an engineering point of view
because only a small sized plastic zone is formed at the tip of the II'
"

0"(2 " 0"(1 / propagating crack. Examples are high strength low alloy (I-ISLA)
/' //6 steels, maraging steels and high strength aluminium and titanium
".; __r/ alloys.
,,/
,__ -
---- 0i2 //
//"
0"i1 LEFM has been used heavily ll1. the aerospace, nuclear, and ship
-6a industries with only a recent extension to the ground vehicleindustry.
LEFM technology is based on an analytical procedure that
relates the stress-field magnitude and distribution in the vicinity
a,1_ _ a, a,
~a" aco0 ----1I _I of a crack tip to the nominal stress applied to the structure, to the
Fig. 15.17. Consequence of invariant R curve.
1022 Welding Engineering and Technology Fracture Toughness and Weld Design 11)23

size, shape, and orientation of the crack or crack-like discontinuity, Z-direction (perpendicular to the X-Y plane) are zero for plane
and to material properties. It refers only to fracture behaviour stress conditions while the normal and shear strains (and shear
under conditions of plane strain. As stated above, this analysis can stress) involving the Z-direction are zero for plane strain
only be applied if the plastic zone at the crack tip is small conditions.
compared with the dimensions of the crack to ensure an overall The equations (15.33) to (15.37) show that the elastic normal
linear behaviour.
and elastic shear stresses in the vicinity of the crack tip are
Using the mathematical theory of linear elasticity and the dependent on r, 8, and K only. The magnitudes of these stresses
Westergaard stress function in complex form, the stresses at any at a given point are thus dependent entirely on K. For this reas on
point near the crack tip can be derived. This result&. in the K is called stress intensity factor. K is not to be confused with the
following equations for the elastic stress components near th.e tip elastic stress concentration factor Kt, which is the ratio of the
of a crack for mode I opening according to the notations shown in maximum stress at a notch to the nominal stress at the notch. The
Fig. 15.18. value of the stress intensity factor, K, depends on the loading, body
configuration, crack shape, and mode of crack displacement. K
used without a mode subscript I, II or III normally refers to mode
cr _ K
I. Units of K are MPa -!ITl i.e. MN m-3/2 and KSi -{[il where
o-y - -rr;;; (For e ::0)
1 MPa -!ITl = 0.91 KSi vrn.
The elastic stress distribution in the Y-direction for e = 0 is
y
shown in Fig. 15.18. As r approaches zero, the stress at the crack
z~x
t "-y tip approaches infinity and thus a stress singularity exists at
~ r = O. Since infinite stresses cannot exist in a physical body, the
~
-JOt-.,-x
-"'xy
elastic solution must be modified to account for crack tip plasticity.
If, however, the plastic zone size ry at the crack tip is small relative
r\oe ~ r r
tip
Crock
x to local geometry (for example, _2:'.t and J::;
a 0.1, where t is plate
0-
thickness), little or no modification to stress intensity factor, K, LS
Fig. 15.18. Elastic stress distribution in the vicinity of needed. Thus an important restriction to the use of LEFM is that
a crack propagating in mode I. the plastic zone size at the crack tip must be small relative to the
...(15.33) geometrical dimensions of the specimen or part. A definite lim] ting
cry K cos 2"l
= -v21tr 8 r1 + sm
. 2"
8 . sm
. 381J
2 condition for LEFM is that nominal stresses in the crack plane
must be less than the yield strength of the work material. In actual
ax = -v21tr
K
K
cos 2"8 [1 - sm
6. 8
. 2"8 . sm. 38]
38 .
2 ...(15.34) usage the nominal stress in the crack plane should be less than
0.8 times the yield strength.
'txy = -v21tr cos 2" . sm 2" . cos 2 ...(15.35)
..,(15.36)
Values of K for various loading and configurations can be
calculated using the theory of elasticity involving both analytical
O'z = 'txz = 'tyZ = 0 for plane stress and numerical calculations along with the experimental methods.
The most common reference value of K is for a 2-dimensional
and o'z = J.L (O'x+ oO'y)l
'txz = 'tyZ = ) I for plane strain
...(15.37)
centre crack oflength 2a in an infinite sheet subjected to a uniform
tensile str~ss 0'. For the infinite sheet,
For validity of the above equations p ::;r::; a must apply (where
p = radius' of curvature of crack tip and a = length of edge crack K = 0' ~ ... 05.38)
or half length of centre crack. Higher order terms exist, but these When K[ reaches the value at which the crack starts to run it
are negligible in the vicinity of the crack tip. It should be noted IS designated as KJC or the critical stress intensity factor and LS
that by definition the normal and shear stresses involving t.he
1024 Welding Engineering and Technology Fracture Toughness and Weld Design 1025

called the fracture toughness of the material. Fracture toughness (i) Crack Tip Plasticity
is the measure of resistance to fast crack propagation when the At the crack tip where r = 0, cry has an infinite value if equation
plate is thick enough, or the yield strength of the material high
(15.33) is rigorously obeyed. In reality such high stress
enough, for plane strain conditions ~o prevail. Hahn has proposed
three categories of failure: magnitudes do not occur because the material in this region
undergoes plastic deformation, thus creating a plastic zone that
(i) linear elastic, surrounds the crack tip. Fig. 15.19 is a schematic representation
(ii) non-linear elastic, and
cry
(iii) plastic instability Infinity
and gives the dividing Jine between categories (i) and (ii) as, \
KJC 1 \
crYB 2 -a
\
(-) = 1.2 ..(15.39) \
and between categories (ii) and (iii) as, \
5 p (stress
Yield) Elastic solution
1\,II
I
transposed to effective
crack tip
(Krc)21.
<JYB a=7 ...(15:40)
I "I'
In practical terms, ductile engineerin~ steels have fractur~
toughness values in the region of 100 N/mm 2, while material with I I" ••.••. Elastic solution
a fracture toughness of about 10 N/mm3/2 would be regarded as II .••...•...
/ --
being in a notch brittle condition.
Krc is a material property which can be found experimentally x
by testing specimens.
Since for the Griffith crack model i.e. elliptical crack in an Fig. 15.19. Plastic zone at the crack tip; plane stress conditions.
infinite plate of unit thickness,
stress intensity factor, K = cr...fiUi and of the change in the distribution of cry caused by localised plastic
cr21ta deformation in the vicinity of the crack tip.
crack extension force, G = ~ The size of the plastic zone, rp' can be estimated from the
then, J(2=~1ta stress field equation (15.33) to (15.37) by treating the problem as
£.e. K2=EG (for plane stress) (15.41) one of plane stress and taking e = 0 and setting cry equal to the
yield strength, cryB, which results in,
and J(2 = EG (for plane strain) (15.42)
(1- v2) r - 1 -Kr
15.7.3.1.Practical Considerations in the Use of Stress For plane stress ...(15.43)
p - 27t ( crYB )2
Intensity Factor Irwin suggested that the plasm zone size under plane strain
The above concepts of stress. intensity factor were developed conditions can be obtained by considering the increase in tensile
for a linear elastic solid in the form of an infinitely wide plate with stress for plastic }ielding caused by plane strain elastic constraint.
simple through-thickness crack configuration. In practice, when Under these conditions, the yield stren.gth is estimated to increase
applying these concepts to metallic materials it is required to allow by a factor of.../3. Consequently the plane strain plastic zone siz€
for, becomes,
(i) ..crack tip plasticity, _/
(ii) finite size of specimen, and
...(15.44)
(iii) variation in crack geometries. 1
rp = 61t (Kr)2
crYB
1026 Welding Engineering and Technology Fracture Toughness and Weld Design 1027

Outside the plastic zone, i.e. where r > 2rp, the stress Now let us assume that a through-thickness crack exists in a
distribution is approximately the same as the elastic stress thick plate. The free surfaces of the plate have zero normal and
distribution around an elastic crack of half length a', where, shear stresses; and therefore the free surface must be in plane
a' = a + rp ...(15.45)
stress conditions. However, the interior region of the plate near
the crack tip is closer to plane strain conditions as a result of
elastic constraint away from the crack. Thus the plastic zone along
the crack tip varies similar to that shown schematically in Fig.
15.20.
For the valid use of the stress intensity factor, based on
LEFM considerations, the relative size of the plastic zone
compared to the crack length and specimen dimensions must be
small. Thus, rp becomes too large when the section stress exceeds
about 0.8 of the yield stress. If the plastic zone size is comparable
to the component dimensions the stress intensity factor is no
longer a valid parameter.
(a) Three dimensional plastic zone
For simple cases transition from plane strain to plane stress
conditions occurs when the plastic zone diameter is equal to the
thickness i.e.,
Plane stress
mode I -2 1t -O"ys =T ...(15.46)
1 (KI)2
(ii) Size of Specimen
The general relationship between fracture toughness, Kc, and
thickness is shown in Fig. 15.21. The nature of fracture and
Crack appearance accompanying the different thicknesses is also shown
=::m schematically for single edge notch specimens. The beach
Tip
markings at the crack tip represent fatigue precracking at low
cyclic stress intensity factor range to assure a sharp crack tip. The
fracture toughness yalues would be higher for dull, or notch type,
crack fronts. It is seen that thin parts have a high value of K.:
accompanied by appreciable shear lips or slant fracture. As the
( b) Overall view thickness is increased, the percentage of shear lips decreases, as
rip (plane strain! c) Edge view
Machine note h
does Kc. This type of fracture appearance is called mixed mode.
For thick parts, essentially the entire fracture surface is flat and
Fatigue crack Kc approaches an asymptotic minimum value. Further increase in
Plastic zone ==rf.,z (plane stress)
Plane thickness does not decrease the fracture toughness, nor does it
strair alter the fracture appearance. The minimum value of fracture
reg~ toughness is called plane strain fracture toughness, KJC. The
subscript I refers to the fact that these fractures occur almost
(d) specimen cross - section entirely by mode I crack opening. The term plane strain is
Fig. 15.20. Schematic representation of plastic zone ahead incorporated here since flat fractures best approach a true plane
of crack tip and its details. strain constraint throughout most of the crack tip region:For thin
1028 Welding Engineering and Technology Fracture Toughness and Weld Design 1029

Low strength, ductile materials are subject to plane strain

fracture at room temperature only if they are very thick.
Therefore, most KIC data have been obtained for the medium and
higher strength materials or for the lower strength materials at
low temperatures.
100 In general a wide range of KIC can be obtained for a given base
alloy. However, a higher yield or ultimate tensile strength
L; generally produces a decrease in KIC for all materials, and thus a
~ 80 greater susceptibility for catastrophic fracture. This is an
:2: important conclusion which too many engineers overlook. The
u
:..:: materials with low fracture toughness are considered to be those
-.::;
<II
60 with KIC less than 60 MN m-S/2•
Q/
C
.&::
0>
(iii) Variation in Crack Geometry
::;)
2 40 Expressions giving the value of K[ for a limited number of
K1c ---
Q/
L.
::;)
crack shapes and positions have been developed and found to have
.•..
u the same basic form, namely,
l:l
L.
u.. cr&"
K[ = ~ ... (15.47)
o where the factor Q accounts for the effect of the crack shape and
o 10 20 30
position neglecting the free surface effect discussed earlier. For a
Thickness IBtmm through-thickness crack in an infinite plate (Griffith moden,
Fig. 15.21. Effect of specimen thickness on fracture toughness.
Q=1.
For the general case, then, of a flaw in a body under stress
sections where appreciable shear lips occur, the crack tip region inducing mode I opening the stress intensity factor can be
most closely experiences a plane stress situation. Thus plastic expressed as,
zone sizes at fractures are much larger in thin parts as compared
to thick parts. Plane strain fracture toughness KJC is considered __ u y
a true material property because it is independent of thickness. K[ = cr -V1t (a + rp) . Q ...(15.48)
Approximate thickness required for steels and aluminium to 1(;
obtain valid KIC values are given in table 15.1. where, rp = plastic zone radius = ~
21tcrys
Table 15.1. Approximate Thickness Required for Valid Krc Tests Q= flaw shape parameter,
S. 2070
1720
690 Yield
1380
1030 345
45
620
19
676
>275
550
448
76 Thickness
- Y= compliance function.
Aluminium
Strength, a"fPa)
cry. (mm)
Three basic types of defects are usually encountered, namely;
(i) flaws which are completely submerged in the section;
(ii) flaws which grow from a surface; and
(iii) through-thickness flaws.
Surface and embedded flaws will generally have high degree
of restraint at the crack leading edge and so plane straiI)
conditions apply. For a submerged crack in a uniform stress field
the equation (15.48) for stress intensity factor reduces to,
 r
1030 Fracture Toughness and Weld Design 1031
Welding Engineering and Technology
I
K[ = 0.6360" (rta)1/2 . Y ...(15.49) i
A surface crack will have a higher stress intensity factor than A
a corresponding buried crack, due to the fact that the restraint c
which controls separation of the two surfaces is reduced and is w
expressed by the equation,
K[ = 0.700" (1ta)1/2 . Y ...U5.50)
Although crack in plates of finite size are of great practical -==-
interest, but for these cases no closed form solutions are available.
The problems are difficult because of the boundary conditions. An ~
approximate solution can be obtained for a strip of finite width
loaded in tension and containing a central crack or an edge crack.
First consider an infinite sheet with a row of evenly spaced B o
infinite collinear cracks as shown in Fig. 15.22. Solution for this
results in,
Fig. 15.23. Stresses on the edges of strip cut from
infinite plate with infinite collinear cracks.
1ta tan 1ta W ...U5.51)
K[ = a -Vrta (w
.-- )1/2 the finite strip behaves as an infinite plate if the cracks are small,
However, a convenient formula for the stress intensity factor for
strip in tension is,
1/2

K[ = a (rta . see ~) ... (15.52)

CT
Fig. 15.22. IInfinite plate with infinite collinear cracks.
If the plate is c)..1talong the lines AB and CD one obtains a
strip of finite widthW, containing a central crack of length 2a. It
is likely that the sQlution of equation (15.51) is approximately
valid for the strip. Ip the case of the collinear cracks a strip of
width W bears stresses along its edges AB and CD as shown in
Fig. 15.23 (Note that shear stresses are zero because of symmetry)
whereas, the edges of a plate of finite size are stress-free.
Supposedly, the stresses parallel to the crack do not contribute
much to K and consequently equation (15.51) can be used as an
approximate solution for the strip of finite size. It appears that
equation (15.51) reduces to K[ = a...fita if ~ -t O. This means that
Cutting the plate with collinear cracks (Fig. 15.22) alongEF
and CD, one similarly arrives at a strip with an edge crack.
Analogous to the central crack problem the solution of equation
(15.51) can be used as an approximation for the edge crack. Again
K reduces to K = a -fiUi for small a/W. However, the stresses acting
on the edge EF t.end to slightly close the crack. Absence of these
stresses in the strip of finite size results in a somewhat larger
displacement of the crack edges. Consequently K is somewhat
higher due to these free edges. The correction factor is in the order
of 12%. Thus, for a small edge crack K is given by,
K[ = 1.12a...fita ...(15.53)
Stress intensity factor and the finite size polynomials for a
number of practical configuration are listed in table 15.2.
Table 15.2. K for Some Practical Geometries
A
a
Ku='~[~aJl~l
K,=a~["';t
II - I'
~J
1032 Welding Engineering and Technology Fracture Toughness and Weld Design' 1033
(J"
where, Q = flaw shape parameter, and
MK = magnification facto:- for deep flaws.

KJ = 1.12 0'...fitQ,- ( small ~ ) 15.7.3.2.Effect of Temperature and Strain Rate on

or KJ = Y 0' Wi Fracture Toughness
2 3 4

(5 A small drop in temperature, or increase in loading rate, can

with y= 1.99- 0.41 ~ + 18.7(~ ) - 38.48 (w J + 53.85(~, J have a drastic effect on fracture toughness. Even when the loading
(1.99 = 1.12 -fit) rate is slow, the strain rate at the tip of a spreading crack will he
much greater than when the crack is non-propagating. Hence
fracture toughness drops as the crack propagation rate accelerates
Kf = 1.12 ~_...fitQ,-( small .; ) and a catastrophic running fracture may develop.
or Kf = yO' .,fa .
w
t Fracture mode
•....
co
(5 with Y = 1.99 + O~~!3~ - 8.48 (~
(1.99 = 1.12 -J1t)
J + 27.36 (~ J OJ
N
transition
'Vi
OJ
c:
Thickness B
oN ~
3 5 v
:0::
2" 2" VI
1:1
C.
Kf = BW2
PS3 [2.9 (E-)t
w - 4.6 (~) + 21.8 (~ ) 2-
~f.....•
;~ ---
7 .JI::
'-'
1:1

U•....

Thickness B
- 37.6lW
(,,)2 +387 (;;J ]
-85 -50 1((-15) 15 50
~ Temperature :C
2
Fig. 15.24. Effect of temperature and loading rate
K _ P a
f - BW1l2 [29.6~ (w)~ -185.61 (~) 9
on crack tip plastic zone size, rp.

2 2 Fig. 15.24 illustrates the sharp change in the crack tip plastic
zone (toughness) with temperature. The fracture toughness
-1017(~) (~7r] variation in temperature transition region is complex since it is

[!]
+655.7 (~) + 63.9
influenced not only by temperature but also by loading rate or
p per unit thickness more specifically, strain rate at the crack tip. In practice the strain
. Kf =p...ria
'0' rate d'ependence offracture toughness can be quite significant. For
Kf mo>: = 1.12 (j) ..[ita
example, at a given temperature the large plastic zone formed
under quasi-static (slow) loading conditions may reduce the
K[ min = 1.12 ~ ...J1ta"'/c mechanical constraint, and therefore improve toughness. On the
other hand dynamic loading may result in a sufficient reduchon
~ <1>= 0
f"/2[ 1-~sin2<p
c2_a2 d<p in ductility (plastic zone) to cause brittle fracture. This
phenomenon is believed to result primarily from the strain rate
J'"

<1>=-+-- 1t a2
31t
9 8 c2 dependence for yield strength of the concerned material Fig. 15.25
,.Gen relation for surface flaw: shows typical temperature results for a low alloy steel.
As the temperature decreases KJC usually decreases. while the
KlC = 1.12 0' (~ )1/2 . MK
yield strength increases. Thus, even though un-notched or
 r
1034 1035
Welding Engineering and Technology I Fracture Toughness and Weld Design

200
t

\\
•• ~ I\.Increasing rno
w I 'f-......• " Ktt
~ ~ Kit-b
II>
Klc cr 0'
:£
II>
150 ...•
>- "Higher .n I \
a..
II>

~ I
250 cf 1\ t t
::lE.;
x-
II>

c: 100
.J:::
0-
::>

.2

~ 50
v
~
...
o I I I I I 10
-150 -100 -50 0 50
Temperatura:':C Lower Kit
Fig. 15.25. Variation of KlC and yield strength with temperature
for low alloy nuclear pressure vessel steel. Crack length,alml-
uncracked tensile strength increases with decreasing Fig. 15.26. Influence of fracture toughness on allowable stress or crack size.
temperature, the flaw or crack resistance can be drastically and strain rate effects are much less marked in face centred cubic
reduced. Increased strain rat.e tend;" to cause changes in KIC metals; for example aluminium, copper and austenitic steels. For
similar to that of decreasing the temperature. That is, higher this reason these metals are often specified for low temperature
strain rates often produce lower fracture toughness, and hence use. Large toughness variations can still be found in such
greater crack sensitivity. KId is the symbol used for critical stress materials, on account of triaxial transitions. Many other
intensity factor for dynamic (impact) loading and plane strain •I specialised environments can alter toughness; for example,
irradiation in nuclear containment vessels, chemical attack, or
f

conditions of maximum constraint. Like Kic it is also expressed

as,
KId = Ccr (G.'
where, C a constant, function of specimen and crack geometries,
=-=
(J nominal stress, MN/mz,
=-=
a = flaw size, mnl.
The allowable stress in the presence of a given crack size is
directly proportional to the fracture toughness, while the
allowable crack size for a given stress is proportional to the square
of the fracture toughness as is .shown in Fig. 15.26. Thus,
increasing KIG has a much"Iarger influence on allowable crack size
than on allowable stress. For monotonic loading on components
containing cracks, a higher fracture touglmess results in larger
allowable crack sizes or larger allowable stress at fracture.
In some metals, notably those showing a body centred cubic
crystal structure such as carbon steel, toughness is highly
sensitive to temperature and strain rate. However, temperature
absorption of gases.
... 05.54) 15.7.3.3.Fracture Toughness Testing
Fracture toughness testing is mainly limited to K1C
determination for the following reasons:
(i) KIC is the most conservative value and is independent of
specimen size;
(ii) crack instability in plane stress is difficult to detect; and.
(iii) the large plastic deformation involved in the fracture of
this section invalidates the analysis used for calculating
Kc·
The practical determination of KIG values involves the
determination of the critical stress intensity factor from a
knowledge of the specimen geometry, nominal stress and crack
length at which fast fracture commences. In practice the onset of
the latter condition may have to be approximated, i.e. the
occurrence of meta-instability in the load/crack extension curv€.
1036 Welding Engineering and Technology Fracture Toughness and Weld Design 1037
r"'1 E
<6 E
The specimen design and testing procedure for valid plane II Vi
strain fracture toughness tests are as per the recommendations of .;'13;
ASTM E399 Test Procedure.
In brief, fracture toughness testing involves subjecting a
1-- ~ ----1 11
S
~
specially designed test piece, in which a crack has been developed r ..c:
'"
l.J

from a machined notch by fatigue, to either bending or tensile 3 '::.J!i

o 'c;j
forces. The applied force is autographically plotted against the o Q .•.•
~13 QI

change in opening mode displacement measured across opposite ,-

~~
"'0

faces of the notch during a rising force test. Crack growth is E:: ._

represented by an incremental increase of displacement without e:E-o~

a.<Ol
an increase in the applied force. The objective is to determine the e ~
force at which a given amount of crack extension takes place. This ... 'a
~2 ...,
os ."
is established from the record of force-displacement in terms of a e:
'"
.c
u
\
on' ~11
given deviation from linearity. The plane strain fracture toughness c5..
.•...
'0
c:
.!!l
o .
.c
..c:
0~
<.>

';;3
can then be calculated using the stress analysis relationship on
'"
t- c:
o'-
end
.s :.0
~
rJ}
E::
-S2S
p > -g3 b5-
.0 QI
.c -'",
Ol.fl ~ a.
g
p dl~
w ",6 '.;:1

." e
c: 11 g

:ll I}i
~o
>- 3-0 ~~
a. co~ .g"t:l
~
,-' a.
on ::l ~
coo
" QI g~
_ "0 Cl)~
>- •• ..Q
::l~

131
:><::

~1l
~ ,
t:.
'"

x- y Plotter -:.13
i:I::i '"

Fracture Toughness
specimen 1010 1TIT
-L3~
~
t-
c-I
u-i
....
::l

t>O

T
H , ,JJ
t£:

appropriate to the specific type of test piece. A diagramatic

representation of a system used for fracture toughness testing is
shown in Fig. 15.27 (A) while 'the details of the standard compact
l~w tension test specimen are shown in Fig. 15.26 (B).
The ASTM test for plane strain fracture toughness testing of
metallic materials suggests that, for standard compact tension
test, the stress intensity is calculated from,
p
...(15.55)
K[ = :B-(W) 112 Y1
p
Universal carrier amplifier
For ~ == 0.600 [See Fig. 15.27 (A)]
(A) Set-up for fracture toughness testing.
1038 Welding Engineering and Technology Fracture Toughness and Weld Design 1039

1/2 3/2 5/2 (i) For L = 4W

Y1 == 29.6 ({t;) - 185.5 (.~ J + 655.7 (W J
_ 1.99-~ (1- ~ )[2.15-3.93U~- .-)+2.7(;~j]
- 1017l(,7/2 + 638.9l
( )9/2~J (15.56)
Also, for a 3-point bend specimen, shown in Fig. 15.28, the
~ ... Y2 -

[1 + 2 (~ )] [1 - ~ ]
...(15.60)

ASTM standard E399 suggests the stress intensity factor may be ForL = 8W
calculated from the equation,
Load
Y2 = 1.107 - 2.12 (~)+ 7.71 (~ J - 13.55 (~ J + 14.25 (~ J
B = 10 or 13 mm ...(15.61)
Now ASTM E399 standard limits state limitations on the
W = 28

L=4W
Machined
notch D range of (~ )with
and (15.~), values
thetoappropriate
be used. In compliance
fact equations 05.57),
(15.55),are
functions, only
~B~
orL=8W recommended to be usedtheover a range offkctions
(~ ) valuesrepresented
between 0.45
and 0.55. In practice, compliance by
equations (15.56), (15.58), (15.60) and (15.61) are applied
Fig. 15.28. 'l'hree point bend test specimen for fracture toughness testing. frequently outside the range for which they are intended. This
6Mal/2. Y2 undoubtedly leads to some error.
...(15.57)
KI = ----:BW"2- No standard test method for dynamic KId testing similar to
where, M == bending moment, and that of ASTM E399 for static testing has yet been established.
Yz = compliance function. Therefore, investigators of the KId toughness testing have
L generally applied the concepts and restrictions of standard E399.
For --=8
W 15.7.3.4.Critical Flaw Size in Fatigue Loading
Knowledge of the plane strain fracture toughness (KIC) of a
Y2 = 1.96 - 2.75 (~ )+ 13.66 (WJ - 23.98 (~ J + 25.22 (;, J material can be used to determine within reasonable accuracy.
either the allowable stress intensity in the presence of a given
...(15.58)
defect or the allowable defect size in the presence of a given stress.
Another equation for determining stress intensity factor for a Thus the critical crack length which when achieved during fatigue
single edge notched 3-point bend (TPB) standard specimen for loading gives final fracture is predictable from the equation,
mode I opening is,
3LP '_'''_H . Yz
KI=---'\1ta ...(15.59) ac = n1(KIC'
cr. YQJ2
.... (15.62)
2BW2
where, L = loading span length, with the symbols having the usual meanings.
P = applied load, Considering the case of an embedded elliptical flaw in the wall
B = thickness ofspecimen, of a pressure vessel under cyclicloading. Fig. 15.29 shows a graph
W == width of specimen, of stress (normal to the flaw versus flaw size factored by Q. the
a = crack length, flaw shape factor) for a given KJC value - the fracture toughness
Y2 = compliance function.
of the material. It is evident that the pressure vessel will fail :in
1040 Welding Engineering and Technology Fracture Toughness and Weld Design 1041

to different repeated stress levels with 0"1> 0"2> 0"3' All specimens
contained the same initial small crack length, and in each test the
b minimum stress was zero. It is seen from the figure that with
...
... higher stresses the crack propagation rates are higher and the
-..
••
...
fatigue life is shorter. The crack lengths at fracture were shorter
at the higher stress levels .
'a
..
Ci. Therefore, the total life to fracture depended on the initial
«
Cl.
crack length, the stress magnitude, and the final fracture
resistance of the material.
A fatigue crack, once formed, moves forward in a controlled
Flaw size. a manner and the rate of crack growth depends primarily upon the
Fig. 15,29. Relationship between failure stress and flaw size.
amplitude of the cyclic stress or strain; the material characteris-
tics; the geometrical configuration; and the environmental
the presence of a crack of size ac• If the vessel is proof pressure conditions. In order to guarantee that cracks which develop and
tested to induce the stress level indicated in Fig. 15.29 then there grow during service do not become catastrophic, it is necessary to
can be no flaw present larger than a1, and the minimum life of the determine the rate of crack growth with as much reliability as
pressure vessel is therefore determined by the crack propagation possible. With this aim in view, various laws have been proposed
life from size a1 to ac' for the rate of fatigue crack growth with stress field, crack size
and material properties and the one most popular among them is
The problem then arises of establishing the rate of crack known as Paris Law and is expressed by the following equation :
propagation for this growth to critical size. If this is possible then
it will enable an estimate to be made of the number of cycles at a
given stress amplitude which will carry the crack from a1 to ac' ~~ = c (tiK[t ...(15.63)
Further more, it should be possible to estimate the crack length where da represents the increment of crack length for an incre·
at any given stage of fatigue loading and thereby gain an ment in number of fatigue cycles, dN; tiKI is the stress intensity
appreciation of the remaining life. This is a significant advance range i.e.
on simple S-N life testing, which gives only total failure life for a tiKI = K[ max - K[ min ... (15.64)'
given stress amplitude ..
15.7.3.5.Fatigue Crack Propagation C and n are experimentally determined constants which are de-
pendent upon material, mean load, and environmental conditions.
Fig. 15.30 shows schematically three crack length versus The average value of n is reported to be 4 with an overall range
applied cycles curves for three identical test specimens subjected of 2 to 10.
03
Clearly, if the Paris Law is applicable then this greatly
facilities the prediction of growth rates in a component or

-o
~
Ol
c:
OJ>OZ>03
structure, because relatively simple laboratory tests can be used
to obtain data which may be directly applied via LEFM to defects
in service, provided that it is possible to analyse service stresses
~ and stress intensity factors. However, the Paris equation does not
u
.><::

describe fully the relationship between da/dN and tiK, being valid
o
u
•...
only for what is known as the intermediate range of growth rates,
(jWv0M Time
eo
being typically 10- 3 to 10- 111m/cycle. In fact the variation of .~~
with tiK tends to be sigmoidal in form, as shown in Fig. 15,31,
Applied cycles, N -
Fig. 15.30. Fatigue crack length versus applied cyc.les. Fracture is indicated by X.
1042 Welding Engineering and Technology
Fracture Toughness and Weld Design 1043

Regime Regime
with a limiting upper bound of Kc (characterising the material
Regime
A B K (
fracture toughness) and a possible limiting lower bound KTH
I c
. /' (characterising a threshold below which crack propagation will not
:z
~
,, ./ ./ proceed).
"- ,/ Average slope = 4 Amongst the numerous models put forward for determining
~o fatigue crack propagation rates not even one of them incorpor~tes
01
I the condition for non-propagating crack and all lack generality. In
-'o __
n =2
I
to 3
any case, the constants and exponents need to be obtained for
-7" 1 In particular materials by actual crack propagation testing.
.
da/dN=((AK) Despite the criticisms associated with the Paris Law, equation
(15.63) is the best accepted relationship at the present time,
because it is the most easily manipulated and that a great deal of
AKTH Log AK experimental data generated fits its general form over a
considerable range of practical importance. Fig. 15.32 shows
Fig. 15.31. Schematic da/dN versus AK curve for fatigue crack propagation typical data for an aluminium alloy from which it may be observed
showing the general sigmoidal form.
that the Paris growth law refers to only a limited section of the
curve, although this section may be extensive.
Microstructural and other features affect the crack growth
rate, but generally the influence is small compared with their
effect on KIC'
As the temperature is increased the rate at which a crack
propagates is also increased as the slope n is invariably reduced
and the constant C increased with increasing temperature.
15.7.3.6. Solved Examples
~u 500
Probleln 15.1. A material has a K[C value, at the service
>.
u temperature of 55 MPa -..rm and a yield strength of 700 MPa. (a)
"- Find the design stress as percentage of yield strength for a
E
c: through-thickness crack shown in Fig. 15.33, if it becomes critical
:z
"0 ~~ ~
"-
o
"0 Through thickness crack
100

50
-LJ-20 K I ::l (j .fTTci

Fig. 15.33. KI value for through-thickness crack.

,,I at a length of 10 mm. Take a safety factor of 2. (b) Find the critical
10 crack length if the same design stress is used but without any safet'}'
4 10 50 . factor. Comment on the two results.
AK (MNm-3/2)
Solution. For a through-thickness crack, the stress intep.sity
Fig. 15.32. Crack propagation data for an aluminium alloy <BS ; RR 58) factor K[ is given by,
at room temperature and with R = 0.05.
1044 Welding Engineering and Technology Fracture Toughness and Weld Design 1045

KJC :::::(Jc -Jrrac ... (i)

t.e. (Jc :::::KJC/-Jrrac ... (ii)
1. Grade 250 1680 87.5
Now, 2ac::::: 10 mm and KJC::::: 55 MPa...fiil
2.Steel
Grade 200
Grade 1540
Yield Strength, O'ys (MPa)t-%K/C (MPa
165 -Jniu)
I' ,

Putting these values in equation (ii), we get, Grade 250 steel was selected for its higher strength, however the.
Krc 55 casing made of it using automatic submerged arc welding failed -
(J :::::--:::::
during hydrostatic test at a pressure of 3.8 MPa at a membrane
-fit (10/2)1/2
1000 stress of 700 MPa. The failure analysis revealed that the fracture
C -fMc initiated from an embedded elliptical crack of 5 mm depth and 35
55
:::::1.77 x 2.236:::::440 MPa mm
60MPalengt~and
'lim. the average fracture toughness of the weldment waR
with a safety factory of 2, design stress, (Jd is, Could the use of Grade 200 maraging steel made the structure
(Jc 440 withstand the desired service stresses? Give detailed mathelnatical
(Jd::::: -
2 :::::-- 2 :::::220 MPa analysis to justify your answer.
220 220 Solution. For the high strength material employed, a wall
thickness of 19 mm put it in plane strain regime.
::::: (Jys X 100:::::700 x 100:::::31.4% of (Jys Now, to calculate the critical crack size using LEFM principle,
(b) Now design stress, (Jd::::: 220 MPa a solution for an internal elliptical crack gives,
Krc :::::55 MPa ..JIll
:. from Krc::::: (Jd -..rita, we get, Kr::::: (J. a1/2 . f(%) ...(i)
where, a :::::
the half crack depth,
K7c (55)2 c :::::
the half crack length
ac:::::-2-::::: 2 m as shown in Fig. 15.34.
(Jd . 1t (220)· 1t

:::::220 . n x 1000 mm:::::19.89 mm

2ac::::: (55
.. Critical crack length
x 1 39.8 mm
19.89:::::
J2 12a~:'~
2ac::::: 39.8 mm
Ans (a) 31.4% of yield stress
(b) 39.8 mm.

Comments : The results show that for design stress of 220

MPawith no safety factory the critical crack length is nearly 4
~2C~
times if the design stress is 440 with a safety factor of 2. So the
two situations are not to be treated as one and the same.
Problem 15.2. An aerospace motor casing of 6600 mm
diameter and 20 m length was designed with a design stress of
.l120 MPa to be fabricated using maraging steel of 19 mm thickness IT
and to withstand proof pressures of 7 MPa. Fig. 15.34. Elliptical crack dimem;ions.

TIvo grades of maraging steel with the following properties

were considered: f( %) is related to complete elliptical integral of the second
kind. However, as
 ___ ~ __ H !
I
1046 Welding Engineering and Technology i
Fracture Toughness and Weld Design 1047

60

and thus equation (i)

% ~
becomes
0, f(~)~(It)1I2 It x 1000
2a = 50 mm, O"c =( 50/2)1/2 - 214.02 MPa
K[ = 0" (Ita)1I2 ... (ii) Similarly, for Grade 200 maraging steel, taking KIC lower by
For the shape of the crack under consideration, the difference 30 MPa ...rm: (to accowit for presence of weld cracks in weldment),
between these two equations is only 4% and so far the sake of as in case of Grade 250 steel, we have
simplicity, equation (ii) is utilised. Based on K[C = 60 MPa ;frll, for K
different crack sizes the membrance stress is, O"c = [~/i give us the following data:
0"
- C
=--~ KIC
(Ita )1/2
... (iii)
For 2a = 0 mm, O"c
(Ita)
= O"ys = 1540 MPa
135
:. for 2a = 0 mm, O"c = O"ys = 1680 MPa 2a=5mm,
60
2a=5mm, 676.8 MPa It x 1000
O"c =( 5/2 )112 = 1523 MPa
O"c _ '---5/2
= -\J It x iooo
135
60
2a = 10 mm, O"c = • m = 478.57 MPa It x 1000
2a = 10mm, O"c =( 10/2)112 = 1077 MPa
135
It x 1000 2a = 15 1um, O"c = --- 1/2 = 880 MPa
( 10/2)'
60
It x 1000
( 15/2)
135
2a ~ 15 mm, 0< ~ ~xJ~-~-~)'" ~ 390.75 MPa
60
2a ::;20 mm, O"c = ( It x 20/2
1000 )1/2
- = 760 _
MPa
135
2a = 25 mm, O"c = • '" = 682 MPa
2a ~ 20 mm, 0<~ (n x ~g~~)"2
60
~ 338.4 MPa
It x 1000
( 25/2)'
135
2a = 25 mm, 0< = (n x ~~~~)'" = 302.67 MPa 2a = 30 mm, O"c = 1/2 = 621 MPa
60
2a = 30 mm, O"c = 1,n = 276.30 MPa It x 1000
( 30/2)
135
(It 1000 )
x 30/2l _
60
2a = 35 mm, O"c = . ," 255.81 MPa It x 1000
2a = 35 mm, O"c =( 35/2)112 = 576 MPa
It X 1000 2a = 40 mm, O"c = 135. '" = 538 MPa
( 35/2)
60
2a = 40 mm, O"c = . m 239.28 MPa It x 1000
( 40/2)
135
It x 1000 2a = 45 mm, = ~'" = 509 MPa
( 40/2) O"c
60
2a = 45 mm, = • m = 225.60 MPa
(It x 45/2)
1000
O"c

It x 1000
( 45/2)
1048
----------------------:::.----:=-----:::..- :::::..
Welding Eng'n ••• lng and Toohnology rI Fradun Toughne ••
---------------------------------- and Weld D~'gn .049
1700
135 C;rod~ 250 ,Y.S. = 1680 IoIPo

C;r.od~ 200.V.S. = 1540 IoIPo

1t x -1000
2a = 50 mm, ac = ( 50/2)112 = 482 MPa
The membrane stress for different values of crack length
(2a),for two grades of maraging steels taking ays = 1680 MPa for \
Grade 250 and 1540 MPa for Grade 200 steel with corresponding
values of KJC = 60 MPa..Jiil and 135 MPa "jffi respectively, are
summed up as shown in Table 15.3. Cl
= 1120 101 Po
a..
Table 15.3.Membrane Stresses for ~
Different Values of Elliptical Crack Lengths
."
Membrane stress. MPa VI
VI
QI
L. slress for 9rod~ 250
.•...
,
.•.•.......•.
(mm) . {rr MPa.
ys = TRIW
, •.... (cry .• = 1540 MPa, VI
30
05 1680
40
20
50
25
35
15
10
45 276
479
214
226
677
391
256
303
239
338
K1CS.=60MI
No. Crack length (2a) ~ Grade 250
'a -.J;n-) I K1C =Grade 200
135 MPa -.Jm) QI
c:
1540 Cl
L.
.l:l
1523 e
1077 L QI

880
760 C;rod~ 250

682 I I with Klc lok~n os 60 MPa{ml

Obsuv~d foilur~
621 crock siz~
I
576 I 24mm
538 I L./ I I I I I
S 10 15 20 25 30 35 40 4S 50
509
Crack size, 20 , mm
482
Fig. 15.35. Design curves for critical defects in 6600 mm diameter motor casing.
(Problem No. 15.2).
This data of Table 15.3 is shown in plotted form in Fig. 15.35.
An intercept of the 700 MPa failure stress for the casing predicts depth to burst a Grade 200 steel casing. This would not have been
a critical defect size, 2a, of about 5 mm which is very close to the possible since tlie plate thickness was 19 mm. Even at the design
observed value also indicated in Fig. 15.35. Besides the Grade 250 stress of 1120 MPa, it would take a flaw of about 9.5 mm in depth
data, the curve for Grade 200 steel is also shown. Although yield to cause plane strain fracture, and that is a fairly large flaw. There
strength of Grade 200 is lower by about 10%that of yield strength is every possibility that a flaw this large in this type of material
of Grade 250 steel, but it has fracture toughness value 165 MPa would be arrested under plane-stress conditions as soon as it grew
--1m for section thickness of 19 mm. The curve for Grade 200 is through the thickness.
drawn by taking KJC = 135 MPa ..fih i.e. lower by about 30 MPa Ans. The Grade 200 maraging steel would have been a more
...fin as in the case of Grade 250 steel. reliable material for this motor casing than the Grade 25D
Significantly, the defect that failed the Grade 250 steel casing maraging steel.
would not have failed a Grade 200 steel casing, and in fact yield Problem 15.3. A pressure vessel of 750 mm diameter is to be
stresses could have been reached without failure. At the failure built to withstand an internal pressure of 35 MPa using high strength
stress of 700 MPa, it would have taken a flaw of about 24 mm in steel with yield strength of 1260 MPa and fracture toughness 242
 1050 Welding Engineering and Technology Fracture Toughness and Weld Design 1051

MPa ...fin. It is intended to prevent fractures caused by any surface where,

{law of maximum depth 13 mm and a/2c ra'tio of 0.25 where a is l!.J2
the crack depth and 2c crack length as shown in Fig. 15.36. Find cr = applied hoop stress, MPa, equal to 2t for vessel sec-
the wall thickness, t, that must be equal to or more than 13 mm. tion shown in Fig. 15.36
t Q = flaw shape parameter as shown in Fig. 15.37
MK= magnification factor for deep flaws, assumed in this ex-
(j = K Ie ..ra:-/,., M X .fTTa
ample to vary linearly between."l.O and 1.6 as ~ (i.e.
L-L-f
A

-j2c f-
Flaw A
CT= Applied
equal to eo
2t
stress.
crack depth/vessel wall thickness) varies from 0.5 to 1.0;

Q = Flaw shope
t < 0.5,
for £. MK"" 1.0, as shown in Fig. 15.38.
parometer , .6

p al 4'~
~ 2c
Section A-A
I
+I
0: pressure vessel with ,'urface flaw: for Problem 15.3.
Fig. 15.36. Section
Mx= Magnification
fac tor for
deep flaws
..¥:
2: ,.5
2v
L..

-o
c:
o
~u 1.3
Solution. The general relation among K1e, cr,and a for surface ~
'c:
flaw as listed in table 15.2 is, go, .2
E
1/2
~E ,.,
KIC=1.12cr(~) ·MK ••. (i)
~
III
T I I < '.0 '.0
O"GIcTys:Negligible 0·5 0·6 0.7 0.8 0.9
a
cTG lerys: 0·60
<TG I cTyS: 0·80
T
<rr;,/r:rys: !·o Fig. 15.38. Assumed magnification factor, MK. for Problem No. 15.3.
Rearranging equation (i), we get,
K1e-{Q ..
u cr = _' .... (u)
'" 1.12 "I/7t . '\fa . MK
"-
o 0.30 Because Q and MK are functions of design stress, an iterative
procedure must be used. Now, KlC = 242 MPa .Jm and a = 13 mm
0.20 Q : [g2_ (0212)( OG/a-ysI2]
Q • Crack-shape paramehr and let us assume ..Q.. = 0.55 and thus from Fig. 15.37, Q = 1.4
g. Complele elliptical
of Ihe second kind
integral crys
0.'0 erG." Gross slress And assuming that MK= 1.0 for the first trial, thus
crys" 0·2 % offset hnsile yield strenglh r
o f
242 x...[fA
0·5 '·0 15 2.0 2.5 I
a. -
1.12 (7t)1/2 1000 . 1.0
Fig. 15.37. Crack shape paramet.er, Q, for surface flaws. cr = ( 13 )1/2
t
1052 Fracture Toughness and Weld Design 1053
Welding Engineering and Technology
242...{fA 242..JIA x 8.77 KJC' -.fQ .
=
1 1.12 x 1.77 cr = 1.12 ~ . ...fa. MK gIves
1.12 X 1.77 X 8.77 242 x --11.38
cr - --------
= 1266.7 MPa. - 1.12 x 1.77 x --113/1000 x 1.48
= 242 x ff:38
Since this stress is greater than yield stress (1260 MPa), use
cr = crys = 1260 MPa and solve for the wall thickness, t, required to \
1
contain the design pressure of 35 MPa. 1.12 x 1.77 x 8.77 x 1.48
= 242 x ff:38 x 8.77 = 850 MPa
t = .E!2...
35 X 750 = lOA mm
== 1.12 x 1.77 x 1.48
2cryS 2 x 1260
For a maximum design stress of 850 MPa, the required wall
Because of the requirement tha,t the wall thickness must be thickness, t, is given by,
at least 13 mm, another iteration must be made.
Second Iteration _ pD _ 35 x 750 _
t - 2cr - 2 x 850 - 15.44 mm
Since t ""a, MK= 1.5 (Refer Fig. 15.36) and assuming Because the calculated thickness is essentially equal to
cr
-=0.8 assumed thickness, a final iteration should show that,
cry.• tassumed -- tcalculated
from Fig. 15.37, Q = 1.33 cr 850
KJC' -.fQ .
Now,
Consequently, cry;= 1260 = 0.675 :. Q = 1.36
cr = 1.12 ~ . ...fa. MK gIves, (Refer to Fig. 15.38)
242...JL33
cr = 1.12 x 1.77 x--l13/1000 x 1.5
242...JL33
and
ta = 15.44
13
= 0.84
= MK= 1.50 (Refer to Fig. 15.37)
1 242 x --11.36
1.12 x 1.77 x 8.77 x 1.5 Thus, cr = ------------1-
= 242 ..fITf x 8.77 = 823 MPa 1.12 x 1.77 x 8.77 x 1.50
1.12 x 1.77 x 1.5 =832 MPa
For a design stress of 823 MPa, the required wall thickness, For a design stress of 832 MPa, the required wall thickness,
t, is given by, t, is,
t = E!2.. = 35 x 750 = 16
2cr 2 x 823 mm t = pD = 35 x 750 = 15 78
2cr 2 x 832 . mm
Based on thickness of 16 mm, iterate a third time to see if the The two values are very close, so we may take t = 16 mm.-
assumed value of MK and Q are equal to the actual values, Ans:r=16 mm
For
T = ~~ = 0.81, MK= 1.48 (Refer Fig. 15.38)
Problem 15.4. An aluminium alloy storage tank 40 mrn wall
thickness is built by GMAW to store LP gas to low temperature.
cr 823 Due to malfunctioning of welding equipment a typical lack of
and for
crys= 1260 = 0.65, penetration defect of 5 mm depth and 10 mm length as shown in
Q = 1.38 (Refer to Fig. 15.37) Fig. 15.39 is caused. Estimate stress iTitensity factor K1 for the
defect if the yield strength of the material is 250 MPa and the
Using MK = 1.48 andQ = 1.38, recalculate the design stress, membrane stress developed is 150 MPa.
t
1054

=W"~~=
150MN/m2
n
Welding Engineering and Technology

+~
Fracture Toughness and Weld Design

FlO(~)=(1-0.619~)

= f - 0.619 .255)= 0.691

mS5

.. ,(ui)
Fig. 15.39. Schematic representation of weld defect in storage tank wall.
(Problem 15.4), Putting these values of Fl (~ ) and F10 (~) in equation (ii), we
get,
Solution. Because the shorter axis of the planar defect is the
more significant dimension so using the infinite plate solution to
obtain a first estimate of K] we get, K] = a\j_ I 1t x 1000'
2.5 Fl' FIO

K] = a...fiUi ... (i)

Now, a= 150 MPa and 2a = 5 mm 150 -fit . ~ l~'gO.. "'1.026 ' "'0.691
=
= 1l.18MPa..[ffi ... (uii)
Putting these values in (i), we get,
Applying an approximate correction for crack tip plasticity, i.e.
K] = 150
replace a by a· in equation (ii),
(1t x 1000
5/2 )1/2 = 13.30 MPa
Applying corrections for the elliptical shape and finite where, a• =a+--2
Kf
... (viii)
geometry i.e. to account for the nearness of the crack tips to the
plate surface (Refer to case 1 and case 10 of Fig. B-1 in
Appendix-B).

K] = a [ 7ta Fl (-W)- FlO (~JJ

~1/2 ...(ii)
and

Now,
21t ays

KI="[ =". F, (~) FlO(~

=--+---~
a • 2.5 (11.18}2
1000 21t. (250}2
r ... (ix)

[from (vii)]

Now, l
F 1 (E-)_
W -

[1-(~) ]
1 lor E-
11.8--- 1.08 va l'dJ: W::;· 091 ... ("')
LU

and
=- 0.00250 + 0.00032

= 0.00282 m = 2.82mm ... (x)

fa·) 1
F10 (~)=FA (max) = [1 - 0.619 (~]
and ...(iu) F1lw
. ~[1-(~r1 roa

r
where, a = half crack depth, from (x)
W = halfplate thickness,
b = half crack length.
Now, 2a = 5 mm, 2W = 40 mm, 2b = 10 mm
~[1- (22~2r
Putting these values in equations (iii) and (iv), we get, = 1
(1 - 0.02942) 08 = 0 961827= 1.033
1.. ... (xi)

(a) 1 1 Putting values of Fl (~) and

we get,
FIO (~ ) from (xi) and (vi) in (ix),
.. F, IV ~[1-(;.gr
=
1
0.9744 = 1.026
r~(1-002;l7)108
...(u}
r
KI~" [=". F, (~) FlO(~ r
1056 "'-. Welding Engineering and Technology Fractus'e Toughness and Weld Design 10S7
1/2
. case of a small crack <Jnet"" <J. Hence, it may be expected that Krc
150 [ 1t X
= ;O~~
X 1.033 X 0.691] can be used until <Jc approaches the yield stress. In practice it
turns out that <Jc should be lower than 66% of yield stress for
11.92 MPa {ill
= Ans. : K[ 11.92 MPa ..Jiil
=::.

LEFM to apply. Also, LEFM does not apply to materials of very

Note: As rp (i.e. Kf/21t<J~3) is much smaller than a in this case, high toughness.
it validates the plasticity correction procedure. It is however
unnecessary to correct the ratio (a/b)_
15.7.4. Crack Opening Displacement (COD) and Crack cr
Tip Opening Displacement. (CTOD) Criteria
Almost all low to medium strength structural steels that are
useq. in the section sizes of interest for large complex structures
such as bridges, ships, pressure vessels, etc. are of insufficient
thickness to maintain plane strain conditions under slow loading Crack
conditions at normal service temperatures. Thus for many
structural applications, the linear elastic analysis used to
calculate fracture toughness (KIC) values is invalidated by the ~YX.f:i;';7.:Gr~~;~it~~:S'JJ2,!;)
formation of large plastic zones and elastic-r lastic behaviour.
LEFM can be usefully applied as long as the plastic zone is
small compared to the crack size. This is usually the case in
materials where fracture occurs at stresses aPi)loeciablybelow the
yield stress .and under conditions of plane strain. When plane
stress coaditions prevail the crack tip plast.ic zone is larger than
in the case of plane strain. If fracture still takes place at stresses cr
which are low in comparison with the yield stress there are ways Fig. 15.40. Schematic represenation of conditions of general yield.
to arrive at a satisfactory solution. If, however, the plastic zone is
larger compared to the crack size, characterised by high fracture Under conditions of general yield shown in Fig. 15.40 plastic
stress and or high crack resistance, LEFM does not apply any flow is no longer contained, but the plastic zone spreads through
longer. the entire cracked section i.e. plastic deformation at the crack tip
One approach to the fracture analysis of these materials is to can occur freely. The crack must be expected to start propagation
use empirical correlations to approximate KiC or KId values. if the plastic strain at the crack tip exceeds a critical value. A
Efforts are also being made to develop elastic-plastic fracture measure for the plastic strain at the crack tip is the crack opening
mechanics concept as an extension of the linear-elastic analysis. displacement (COD). Hence, it is conceivable that fracture takes
Three such concepts are : place when critical value .. of COD is exceeded. This criterion was
(i) COD and CTOD,
first proposed by Wells .
(ii) J-integral, and 15.7.4.1. Crack Opening Displacement
. Wi) R-curve analysis. In 1961, wells proposed that the fracture behaviour in the
Generally speaking, the plastic zone will become large and vicinity of a sharp crack could be characterised by the opening of
spread through the whole cracked section if the net section st.ress the notch faces-namely the crack opening displacement (COD) or
is equal to yield stress i.e., crack tip opening disphicement (CTOD) as shown in Fig. 15.41.
To start with, it should be checked whether Wells' criterion is
<Jnet W~ a ;:::<Jy.;
= <J ... ( 15.65) • in conformity with LEFM or not. In the case of LEFM the elastic
where Gllet in the cracked section is obtained by divid \lg the load
i. •. Wells, A.A., 'Unstable Crack Propagation in Meta' .• - cleavage and fast
by thE:' sectional area of the "'maining uncracked material. In the fracture', Proceedings of Crack Propagation Symposium. Cranfield (U.K.), 1961, pp.
210-230.
------~.-_="c,,=.""'

10S8 Welding Engineering and Technology Fracture Toughness and Weld Design 1059

y Equation (15.70) holds good in the area of LEFM, fracture occurs

I Y' if K] = KIC which according to equation (15.70) is at a constant
I
I value
I LEFM.of CTOD, thus it appears that Wells' Criterion applies in
I
I Use of Wells' Criterion in LEFM would require measurement
II ofCTOD. A direct measurement ofCTOD, however, is difficult and
I virtually impossible in a routine test. It can be obtained indirectly
Crack eroo
II by measuring K] and using equation (15.70). That would imply
--
Ie. - - "t,-- ,
v
I x ac~eptance of the factor 4/rt introduced by the plastic zone
correction. But the magnitude of the plastic zone correction is
I I I subject to doubt. This problem can however be circumvented by
I I 20 I
I -I I substituting equation (15.68) in equation (15.67), leading to
1
2aeff=2a+2rp* -II (neglecting the term in r;2) the following equation,
Fig. 15.41'. Schematic represenation of COD and CTOD.
solution for the crack opening displacement (COD) can still be COD = E
40' ...y
a2 -
E2
x2 +. 16~ (CTOD)2
...<15.71)

used. The displacement of the crack surfaces with reference to l·'ig. According to equation (15.71) the CTOD can be determined
15.41 is given by the equation. indirectly from a measurement of COD (for example at x = 0, the
centre of the crack) without any assumption about the size of the
COD = 2v,= 4a"./22 E
a - x ...<15.66) plastic zone correction. The COD can readily be measured by
means of a clip gauge.
By applying a plastic zone correction r p. it follows that,
Alternativeiy, use is often made of the equations for crack tip
CO 4a "./ •2 T opening that follow from what is known as Dugdale approach or
D= E = (a + rp) -x ...(15.67)
strip yield model analysis. The basic relationship developed by
Dugdale'" is,
where (a + r;) is the effective crack size and where the origin of
the coordinate system is at the cent.re of the crack. The crack tip
opening displacement (CTOD) at the tip of the physical crack is - a- 8
CTOD __ rtE a lognsec (~2 'O'YB
ays' ~) ..• (15.72)
(
found for x = a. Since r; $; a equation (15.67) leads to where, O'YB = yield strength of the material, MPa
1
CTOD '= ~ "./2ar; ... <15.68) a = 2' real crack length, mm
Shifting the origin of the coordinate system to the crack tip a = nominal stress, MPa
yields the general expression for COD, as E = modulus of elasticity of the material, MPa
...(15.69)
COD = ~ "./2aeff· r
Using a series expansion for lognsec (~ . a:s) this expression
becomes,
CTOD then follows from r = r; and aefr= a, leading to equation
<15.68).
2
a = 8ayS • a
[
.! ~.~
2
+ -1-. ~. ~
,4
+ ~ ~. ~
6
Substitution of r; = a ~ yields, n E 2 (2 "yo) 12 (2 "'0) 45 (2 "'0) + ...]
2ays ...(15.73)
4 Kf
CTOD = -rt . -Eays ...(15.70) ** Dugdale, D.S., 'Yielding of steel sheets containing slits', Journal of the
Mechanics and Physics of Solids, 8, 1960, p. 100.
 Fracture Toughness and Weld Design 1&61
1060 Welding Engineering and Technology
particular structure. The advantage of'COD/CTOD approach is
For 0 less thliUl~ 0ys, a reasonable approximation for 0, using that COD/CTOD values can be measured throughout the entire
only the first term of this series is, plain strain, elastic-plastic, and fully plastic behaviour regions,
whereas Krc values can be measured only in the plain strain region
o = _1t_cr-_.a_ ...(15.74) or approximated in the early portions of the elastic-plastic region.
E· Oys As with the Kr analysis, the application of the COD approach
1tcr-a = 0 . E . (T ys,
...(15.75)
or to engineering structures requires the measurement of a fracture
From equation (15.38), for a through-thickness crack oflength toughness parameter, 0c, which is a material property that is a,
2a, function of temperature, loading rate, specimen thickness, and
possibly specimen geometry, i.e. notch acuity, crack length, and
Kr = 0 ...J1ta overall specimen size.
...(15.76)
or 1ta·o2=K; 15.7.4.2.Experimental Determination of CTOD/COD
From equations (15.75) and (15.76), we get Usually, CTOD measurements are made on three point bend
o·E·o =~r
...(15.77) specimens of the type shown in Fig. 15.42. Such specimens are
ys
...(15.78) p
Now
,
E=-
Oys
Eys ~
Putting the value of E from equation (15.78) in equation
(15.77), we get, ~
o· -o;s ..~'
=1.\.1
Eys

...(15.79) B = Thickness of platE matErial

bdng considErEd for structural
i.e., ~Eys = (KrJ2
Oys
application
Also, the strain energy release rate, G is given by, GEnErall y : W = 2B
a = B
G _ 1tcr-a [from (15.29)]
- E Fig. 15.42. COD test specimen.

Thus, from equations (15.29) and (15.75), we get, specially used for tough materials where fracture occurs after
G·E=o·E·o
general yield. The critical displacement (.1c) is determined from
ys
...(15.80) the load versus clip gauge displacement, .1, curves two of which
i.e., G=o,Oys are shown in Fig. 15.43.
At the onset of crack instability i.e. fast fracture, that is when For smooth continuous records in which the applied load rises
Kr reaches Krc, CTOD reaches a critical value, oe· Thus' equation
with increasing displacement to the onset of unstable fracture
(15.79) becomes, (Type a curve, Fig. '15.43) The critical displacement (.1c)should
be taken as the total value corresponding to maximum applied
Oc Krc load including both elastic and plastic components. If onset of
-J
-Eys = ( Oys 2 ...
(15.81)
crack growth cannot be established, the COD for crack initiation
cannot be measured. Formaterial comparison purposes, a CTOD
(om) calculated from clip gauge displacement (am) at the first
Oys
Because (KrcJ2 can be related to the critical crack size in a attainment of a maximum load can be used. However, the results .
particular structure, it is reasonable to assume that the parameter may depend on specimen geometry also.
oc/Eys can likewise be related to the critical crack size in a
1062 Welding Engineering and Technology
Fracture Toughness and Weld Design 1063

0.4
•v v
o

t
~o L 0.3
o
...• oL
.•.... Different symbols for CI
...•
CI vClriety of mCllcriols (Fe.AI.Ti)
.•...

"0
co
;z::
Clip goygcdisplocemcnt. A .E 0.1
o
0::
Fig. 15.43. Force-displacement records for calculation of COD.
Having obtained the critical value of COD it is necessary to o
convert it into CTOD (oc). This is usually done by assuming that o 0·050 0,100 0·150 0·200 0·250 0·300
the deformation to occur by a hinge mechanism, Fig. 15.44 (a), CTOOlmml-
about a centre of rotation at a depth of r (W - a) below the crack
tip as shown in Fig. 15.44 (b) assuming that the whole ligament Fig. 15.45. Relatioship between CTOD and rotational factor. r.
is above yield and thus it may be considered a plastic hinge. The
6clMeasured with clip gauge)

r-Li Knife edge

to support
clip gauge

a
coo~
1_ a ,I rlW-a I

,'
I'
la) I b)
'flr, W-a)

Fig. '15.44. (a) Plastic hinge, and (b) relationship between COD and CTOD ..
W-a
,
Apparent centre
rotational factor r has to be determined experimentally. Thus from of rotation
Fig. 15.44.(b), we get,
CTOD(o ) = COD r (W - a) ...(15.82)
C a + r (W - a)
or COD = a +1 ...(15.83)
CTOD r(W-a)
Fig. 15.46. Representation of the notch profile during bending.
7~-4;>~~~~,*- ".t... _

1064 Welding Engineering and Technology Fracture Toughness and Weld Design 1065

It has been proved experimentally that the rotational factor,

r, increases during loading from almost zero (for elastic material 6
with only limited plastic flow at the crack tip) to a more or less
constant value of 1/3 (for fully plastic ligament), which is often
used in calculations. Relationship between rand CTOD for
000 .~
c:
'Vi Ic:
<IIc: "
0'0u:z: 0" "&
VI
•••
G
•.... ,
.E,..1=0"00 -I.t'I
VIOJ
t5Ol.
oj
•....
1
I
600"",..,
" I.t'I
23" 0·5
6'.0
... .0
"'"
C>
VI • 00 E
••• o,&j OJ

different materials based on experimental data is shown in .Fig.

15.45.
Experimental calibration using test pieces up to 50 mm thick
have shown that, for CTOD range of approximately 0.05-0.5 mm, 0=2.75
the practical range of interest for specimens 50 mm thick, good
amox= 0.066c/£:yc
approximation to the CTOD can be obtained using the equation,
with reference to Fig. 15.46, as
(W -a) /),c L z = 1·75
,I
°c = W + 2a + 3z ...(15.84) amox =0·090cl £ys

This relationship is derived by assuming deformation to occur

by a hinge mechanism about a centre of rotation at a depth of
(W- a) below the crack tip (Fig. 15.46). Test results indicate this I
approximation to the centre of rotation is reasonable, although r 1.0 1.5 2.0 2.5 3.0 3.5 4.0
can be smaller for very small values of /),c.
15.7.4.3. Use of CTOD for Determining Critical Crack I Strain, £: Ity!>
Lengths Fig. 15.47. Design curve for COD, strain, and crack size relationship.
I 1. Determine ~, where e is the actual strain at the location
K1C UsinLthe relationship for aand
through-thickness
for the design stresscrack, t.
= 0" -vrra as
per equation (15.38), equal Eys
to some percentage of the yield strength, then critical length acr being analysed, and Eys' is the yield strain. Thus, for general design
is given by, with stress concentration equal to 1.0,
O"des
acr = constant x -- ...(15.85)
O"ys
(K1C)2 e=E ...(15.87)

Similarly, because -IQ is proportional to -.!2(15.79), For the. condition at a stress concentration factor, multiply the
(KO"ys )2 Eys
0
nominal design strain (Edes = O"~es) by the stress concentration
.. acr = constant x (~) ...(15.86) factor to find the maximum strain, E.
where the values of these constants depend on the appropriate
design stress level and stre~s concentration factors. 2. Knowing ~, determine the non-dimensional COD, <1>, from
Eys
The particular constant is determined from the CTOD design Fig. 15.47.
curve, developed by Burdekin and Dawes·, shown in Fig. 15.47,
as follows. 3. Knowing, <1>, establish the relationship between acr and oc.
as
* Burdekin, F.M. and Dawes, M.G., 'Practical use of Linear Elastic and Yielding
Fracture Mechanics with Particular Reference to Pressure Vessels', Practical Oc
'Application of Fracture Mechanics to Pressure Vessel Technology, Institute of ...(15.88)
Mechanical Engineers, London, May, 1971. aC/' = 21t eys . <I>
1066 Welding Engineering and Technology Fracture Toughness and Weld Design 1067

Now, = ~E. = 210280 ...(viii)

where, acr = ~. through-thicknesscrack length, 2a - hence the E
ys x 103 = 0.00133
factor 2 in denominator of equation (15.88) From (vii) and (viii), we get,
Oc = Critical CTOn for the actual material being used
Oc
at the location being evaluated, measured at the acr = 0.09 x n nn"too = 67.5 Oc
service temperature and loading rate (note, how-
ever, that most Oc values are measured under con- But, Oc = 0.50 (Given)
ditions of slow or static loading) .. acr = 67.5 x 0.5 = 33.75 mm
Eys = yield strain of actual material being used. (b) NowKt=6
Problem 15.5. A welded pressure vessel made of steel having E 1
a yield stress of 280 MPa and modulus of elasticity (E) of 210
. a
-=-
Eys
cr

crys
·Kt=-x
2
6=3 ...(ix)

x 1(j1 MPa. It was designed for nominal design stress of cr = ::.:. From Fig. 15.47, for ~ = 3, <t> = 2.75 ... (x)
Find the critical length of crack in service, if, Eys
(a) the stress concentration factor, Kt, is 4 and CTOD, 0c, at where,
o Oc
<t> = ---- or <t> = ----
the service temperature is 0.50 mm. 21t. Eys . a 21t. Eys . acr
(b) the stress concentration factor be 6 and CTOD, Oc , at the
service temperature 1.25 mm. i.e., a
cr
=---°c
21t . Eys
Solution <t> •

= O_c = 0.05785 . _Oc ...(xi)

Now, cr =--=
crvs
2
(Given) 1tx2.75·Eys Eys
1 I
cr
... (i) Al cryS 280 0 ...(xii)
.. crys = 2"
!. so, . =
EyS ~E == 210 x 103 = 0.0 133
Also, stress concentration factor, Kt = 4 (Given) ... (ii)
E cr
Now, -. =Kt'- ...(iii)
Eys crys

Putting the values of ~ and Kt from (i) and (ii) respectively

crys
in (iii), we get,
_E_=4x.!=2 ... (iv)
Eys 2
From Fig. 15.47, for ..£ = 2, <t> = 1.75 ... (v)
Eys

Now, dimensionless CTOn, <t> = (") 0 ... (vi)

. 1t . Eys . a 0.2

t.e., a
cr
=-----
°c
21t· <t> • Eys o
Oc Oc 0·1 '·0 10·0 100·0 soo.o
= ----- = 0.09 - ...(vii) 20/T
21tx 1.75 . Eys Eys

I Fig. 15.48. Assessment of triaxial constraint in an infinite plate.

 1068 Welding Engineering and Technology Fracture Toughness and Weld Design ]069

From (xi) and (xii), we get,

C1ys = 450 MPa
Oc
acr= 0.05785 x 0.00133 = 43.5 0c ...(xiii) Putting these values in (i), we get,
Now, OC= 1.25mm (Given) 55 1 ' .,
acr = 43.5 x 1.25 = 54.4 mm Cl = (450 'J2 . 0:05 = 0.3 ... (~l)
Ans.: (a) 33 mm and (b) 54 mm. For plane stress conditions, from equation (15.90), we have,
Alternatively, critical crack length for a given structure can be
determined, using Fig~ 15.48 if the values of K1e, C1ys, 0c, Eys' C1des Eys . T = .2 .... ~~l
and the plate thickness are known. (OcJc =1 0.3Cmm, Eys = 0.2% = 0.002 and T =50 mm(0")
Now, Given
The procedure involved is based on the fact that,
Putting these values in (iii), we get,
...(15.8~)
(KICJ2 1 C2 = (0~O~2). 510= 3 ... (iv)
and - . - = constant· ...(15.90) Now, in Fig. 15.48 mark points A and B corresponding to
Eys
C1(y~c)' T
: =.constant
C1
constants C1 and C2 for -- = 0.4.
C1ys
I' Mter determining these constants the same are located on the
two zones of .Fig. 15.48 and the validity of each is checked. For point A
2a = 1.05
Obviously, only one of the two will bevalid and that the ~ value T
is determined. Knowing the section thickness T, acr (i.e. 2a in Fig. 2a = 1.05 x T = 1.05 x 50 = 52.5 mm = acr
15.48) can be determined. acr = 52.5 mm ,
Problem 15.6. The weld metal in a given material has a yield
strength of 450 MPa measured at 0.2%permanent offset. Fracture
toughness tests are carried out in which the material survives to a
K1 level of 55 MPa...Jm: in plane strain and on further increase of
loading a transition occurs. Eventually the crack starts to
propagate at a crack tip opening displacement, 00 of 0.3 mm. 0.7
Estimate critical crack size for infinite plate structures
incorporating such welds. where, b
(i) the structure is 50 mm thick and the applied stress is·O.4of ..•..

the yield strength. b

(ii) the thickness is 10 mm and load is increased to 0.7 of the
yield strength.
Solution
For case (i)
For plane strain condition, from equation (15.89), we have,

... (i)
C1ys T
Now, ~KIC12
IC"; . 1.
55 = C1
MPawn-, T = 50 mm = 0.05 m Fig. 15.49. Figure 15.48 with points A and B as well
as points C and D marked on it.
 Fracture Toughness and Weld Design 1071
1070 Welding Engineering and Technology
2a == 1.7
For point B T
2a .. 2a = 1.7 X T = 1.7 x 10 = 17 mm = acr
-=10
T It is expected that the structure should reach point D with
2a = 10 . T = 10 x 50 = 500 mm = acr acr = 150 mm as a transition will occur from plane strain to plane
acr=500mm
From Fig. 15.49 it is evident that the plane strain fracture O'ys = 0.25 with (KlCJ2
O'ys
0' stress conditions at ...!L T1 < 1.49 (see Fig. 15.49)
could occur at - = 0.4 for a crack length of 52.5 mm. For greater Ans. : (a) Plane strain failure at acr =52.5 mm
O'ys
(b) Plane stress failure at acr = 150 mm
crack lengths, plane strain fracture could intervene before the
situation is relieved by triaxial transition to plane stress 15.7.5. J-Integral Criterion
conditions. A method of obtaining an equivalent to the stress intensity
'(ii) For plane l:ltrain condition, from equation (15.89), we have, factor (K) for elastic-plastic conditions is to calculate the stress
and displacement for the configuration in question, and then
IC 1 integrate strain energy around a closed path that includes the
(KO'ys J2 . T = C~ ... (v) crack tip. The result which is designated by J is a generalised
Now, KIC = 55 MPa..Jm, T= 10 mm:::: 0.01 m, O'ys = 450 MPa relation for the energy release due to crack propagation which may
also be valid if there is appreciable crack tip plasticity. Since J is
Putting these values in (v), we get, not path dependent, it can be determined in the easiest way by
selecting a path along which the integration can be carried out
... (vi) conveniently, for example the edge of the specimen. In this way
C3 = 450 . 0.01 = 1.49
(55 )2 1 the J-integral provides a relatively simple means to determine the
Ca = 1.49
energy release rate for a case where a large plastic zone exists at
For plane stress conditions, from equation (15.90), we have, the crack tip. Thus, this technique can be used to estimate the
fracture characteristics of materials exhibiting elastic-plastic
...(vii) behaviour and is a means of extending LEFM concepts from linear
roc)
C4 =ley.1 1
T elastic (KlC) behaviour to elastic-plastic behaviour.
Now, Oc = 0.3 mm, eys = ~o~
= 0.002, T = 10 rom For linear-elastic behaviour, the J-integral is identical to G,
the energy release rate per unit crack extension. Therefore a J
Putting these values in (vii), we get, failure criterion for the linear-elastic case is identical to the ~c
...(viii) failure criterion. Thus, for linear-elastic plane strain conditions,
C4 = (0~O~2)' 110= 15 1
c4= 15 2 K2
...(15.91)
JIC = OlC = ( - v)
E .1£
Marking points at ...!L = 0.7 corresponding
O'ys
to C3 and C4 values and J is related to CTOD, by the relation,
we get points C and D on Fig. 15.49. The corresponding critical J=m· Oys' 0 ...(15.92)
crack length (acr) in plane stress is, . where m usually lies between 1 and 3.
2a From a more physical view point Jmay be interpreted as the
T--15 potential energy difference between two identically loaded bodies
.. 2a = 15 x T = 15 x 10 = 150 mm = acr having crack size with slightly different lengths,
The corresponding critical crack length 1
in pane straIn. r or J = - -1 . --au ...(15.9.3, ••

--------------
condition, if it could occur, is

-
.•...
j B ~
1072 Welding Engineering and Technology Fracture Toughness and Weld Design 1073

p where, A = area under the load-displacement curve up to the

L various displacements ofinterest, N-m

B = specimen thickness, m
b = uncracked ligament (for a bend bar, W - a), m
Thus, J has a unit of N/m.
Critical values of J are labelled JIC and are taken when crack
extension is first encountered.
p 15.7.5.1. Experimental Test Procedure for Finding J
a Based on their experience with JIC testing Landes and Begley·
o+da recommended the following steps for finding J/C.
1. Take 4 to 6 bend bar specimens deeply notched with
a
W = 0.45 - 0.55.
2. All specimens should be precracked in fatigue with max-
imum loads less than half of the expected specimen limit

6 load and (KE= )< 0.002 in1/2 (0.01 mm1/2).

Fig. 15.50. Interpretation of J-integral.

This definition is shown schematically in Fig. 15.50. Note that "0
o Precrack
6 (deflection) is measured in the load (P) line. -'o end fie
For practical determination of J, several bend bar specimens
shown in Fig. 15.51 (or compact tension specimens) with a deep
Displacement, 6 -

crack J 0.45 to 0.55)asarea loaded

and (~is 2:determined functionto the displacement
of displacement of interest
using the (a) ( b)

following relationship,
2A
J= B .b ...(15.94) o -.
p -. o
o
J1c

4W o

60 - flo -
(c) (d)
Fig. 15.52. Procedure for J measurement: (a) load specimens to various
displacements, (b) measure crack extension, (c) calculate J for each specimen
and plot it versus M, and (d) construct two curves for JIC measurement.

P/2
--_·tP12 • Landes, J.D. and Begley, J.A, 'Test Results from .J-Integral Studies: An
Attempt to Establish a JIC Testing Procedure', A.."TMSTP 560, American Society
Fig. 15.51. Bend bar specimen for J-integral test. for Testing and Materials, Philadelphia, 1974, pp. 170-186.
1074 Welding Engineering and Technology
3. Load each specimen to different displacement values
using displacement control, if possible, Fig. 15.52 (a). It
is desirable to load one specimen to a point where no
actual crack extension has occurred and others to dif-
ferent amounts of crack extension. It may be difficult to
determine this prior to testing; however, if specimens
are loaded individually and each fracture surface is ex-
amined before the next specimen is loaded, this can
easily be done .
4. Unload each specimen and mark the crack. (Heat tinting
is an easy way to mark the crack for steel. For other
materials a dye penetrant or a fatigue mark could be
used).
5. Pull the specimen apart and measure crack extension.
The crack extension should be measured at its maximum
point and taken to include all crack extension from the
fatigue precra.ck to the end of the mark, as shown in Fig.
15.52 (b).
15.7.5.2.Data Analysis
The data obtained from the experimentation may be analysed
in the following steps.
1. Calculate J values from the load (P) versus displacement
(&1) record using the relation (15.94) i.e. J = ~ w~ere
A is the area under the curve up to the point of unloading
[Fig. 1~.52 (a)] with right hand side of the area measured
up to vertical iine at the unloading displacement; b is
the uncracked ligament measured from the end of the
fatigue crack, and B is specimen thickness, as shown in
Fig. 15.51.
2. Plot J versus crack extension as shown in Fig. 15.52
(c).
3. Construct the line J = 2aflow . ~a, as shown in Fig. 15.52
(d), where aflow can be taken halfway between yield and
ultimate stresses.
4. Draw the best fit line to the J versus crack extension
points [Fig. 15.52 (d)). Include only the points. Where
actual crack extension appears only ·as a stretch zone
the point should fall along J = aflow . ~a.
5. Mark JIC at the intersection of the lines constructed in
steps 3 ,and 4.and marked in Fig. 15.52 (d) .
Fracture Toughness and Weld Design uns
15.7.5.3. Specimen Size Analysis
Proper selection of the specimen size for tests !tray be
confirmed as follows :
1. Calculate J/aflow'
2. Compare specimen dimensions a, B, b with J/cr/low as
follows:
0:=-------
(a, B, or b)
...<15.95)
J/aflow
3. 0: should be greater than 25 for a valid specimen size.
15.8. FRACTURE TOUGHNESS TESTS WITH IMPACT
LOADING
A number of fracture toughness tests are used throughout the
world wherein the test specimens are impact loaded. In the
interest of brevity only three such well known tests viz,
(i) Charpy V-notch test,
(ii) Drop weight NDT (nil-ductility transition) test,
(iii) Dynamic tear (DT) test,
are discussed briefly here.
Both the Charpy Cv notch and Drop weight tests are ASTM
standard methods while the DT test is US Military standard.
15.8.1. Charpy V-Notch Test
This test is used to determine the energy absorption in fracturing
a specimen and consists of impact loading a three-point bend bar
which contains a relatively blunt notch. The energy absorbed is
considered to be a measure of fracture toughness. The effect of test
temperature and loading rate on Charpy impact energy are as shown
in Fig. 15.53 (a) while Fig. 15.53 (b) shows possible .orientations of
charpy v-notch test specimens in butt welded plates,
Some investigators and designers do not believe that the Cv test
is a reliable measure of fracture behaviour in real fracture. They
point out that the dull and shallow notch, the small specimen size.
and the short unbroken ligament under the notch do not adequately
represent the sharp, highly constrained defects or discontinuities in
a large structure. However, inspite of these limitations this test
procedure has gained worldWide acceptance and today it is used
routinely for steel specification and quality assurance.
For steels having yield strength greater than 700 MPa charpy
v-notch energy CVN is related to plane strain stress intensity
factor (KIC) by the following relationship:
ays
(KICJ2 = JL
crys [CVN ~J
- 20
...(15.96)
1076 Welding Engineering and Technology
Fracture Toughness and Weld Design 1077

Ductile r-A
>.
a>
•...
--'- r Crack starter
(weld)
ell
c:
-ell

LA
...,
Cl
Q.
e
Loading
rate
mmLI (a) ,..
Test specimen
19 or 25 I BBmm
>.
Q.
~mm
•...
Cl (b) Cross-section at A-A
.&:.
U

Temperature -

(a) Temperature vs impact energy

Fig. 15.54. Cross-section of Drop weight NDT test specimen showing initial crack.
During the test the weld is fractured and the plate is subjected to
a strain of several percent if the base metal does not fracture.
The fracture illustrated in Fig. 15.55 shows a change from
break to no-break performance that evolves sharply within a 6° to

( b) Charpy test pieces

Fig. 15.53. (a) Effect of temperature and loading rate on Charpy

impact energy, and (b) possible orientations of charpy
V-notch test specimens in butt welded plates.
15.8.2. Drop Weight NDT Test
._/ The drop weight test design was developed based on service
failures resulting from a brittle fracture initiation at a small flaw
located in a region of high stress. To stimulate this behaviour, a
specimen consisting of a small plate section containing a brittle
weld region was devised. The specimen, shown in Fig. 15.54, is
dynamically loaded in bending by a falling weight which causes NDT-Nil- ductility
temperature
the specimen to deflect a fixed distance, restricted by "stops".
Fig. 15.55. Fracture appearance of Drop weight NDT specimen.
1078 Welding Engineering and Technology
Fracture Toughness and Weld Design 1079
11°C ,temperature increment. The highest temperature at which
the specimen breaks is termed as the nil-ductility transition (NDT)
temperature. The sharp change between break and no-break
r 178mm 1
performance is an indication of the brittle-ductile transition
phenomenon exhibited by the mild steels. It should be apparent,
however, that an NDT temperature can be obtained only for steels
that do exhibit a sharp temperature transition behaviour, thereby
excluding the high strength steels. Also excluded from the test T I~.
13o!ml nllm
165 ~-1
method are steels that develop a tough HAZ (heat affected zone)
resulting from the brittle weld deposit leading to erroneously low
value for NDT. (a I A standard
_J

test specimen
J
t
16mm

Assuming that at the NDT temperature the plate surface

reached the dynamic yield stress, ayd. corresponding to the testing QJJ
. a
25m~/ 455
te~perature and that the crack geometry was of an 2C ratio of 1
to;4 (.where a is crack length (edge) and Cis half the crack width).
Then for a part-through-thicknesscrack,
Tr
120mm
~

KId = 0.6 ("'Jiii-) a.'ld --L

or KId = 21 ('.1m";,) CJyd ... 05.97)
where, Kld = dynamic critical plane strain stress intensity factor 40Smm
at NDT temperature, MPa"'{inm,
I bl An alternative specimen design
ayd = dynamic yield strength at NDT temperature, MPa
Fig. 15.56. Dynamic tear (DT) test specimen configurations; (a) a standard
Range of the constant in equation (15.97) is 0.5-·0.78 for inch specimen, and (b) an alternative specimen design.
units and 17.5 to 27.5 for the rnm units. This equation appears to
25mmDT
give a realistic approximation to Kld for low strength structural 25mm OT CVN 16mmOTIJ1
steels at their NDT temperature. 6775
* NOT - Nil-ductility J --1350
The NDT temperature has provided a means for correlating temperature
ship fractures and other service failures. f 5420 -·1080
15.8.3~Dynamic Tear (DT) Test

/---
>-
Dynamic tear fracture toughness test was developed to
/
/l
~ 4065 /_ C1.N 16 mm OT
'" 610
characterise the properties ofthe ultra-high strength steels as'well .;
'"
as aluminium and tItanium alloys that do not exhibit a
temperature transition behaviour and the steels of intermediate .c.c
'-I
0Va1 2710
-" 540
<z: 1355
strength which feature a low value of upper shelf energy.
To fulfill the above needs the standard DT specimen shown in
Fig. 15.56 was designed to simulate the critical structqre features
0

=--
/ * f 270

that affect fracture toughness. The specimen thickness may be ___ ~DT ~
adjusted to match that of the structure and thus achieve the - 40 -18 4 26 48 70 92
o
proper mechanical constraint. However, the standard specimen Tempera ~ure:C -
shown in Fig. 15.56 (a) is of 16 mm thickness. The DT specimen
features a deep machined notch with knife-sharpened tip so as to Fig. 15.57. Relationship between CVN, NDT, and DT test results
for a popular ship steel (ABS-C).
1080 Welding Engineering and Technology ), Fracture Toughness and Weld Design 1031

achieve maximum constraint for a given thickness. (Alternate
specimen designs employ a brittle weld region in place of the
machined notch. For steels with a hardness above HRC36,a knife
blade will no longer sharpen the machined notch tip and a fatigue
crack is often used). The unbroken ligament is of sufficient length
to simulate a fully developed fracture. The specimen is tested by
impact loading to account for the strain rate sensitivity of some
metals, and the absorbed energy is recorded. The features combine
to produce a test of high severity.
growing crack the plastic zone increases considerably. It is
conceivable that it becomes of a size in the order of plate thickness,
such that plane stress. conditions develop. This causes a steep rise
in R and thus R may become larger than G, particularly if G is a
decreasing function due to fixed grips. Crack arrest according to
t
G,R G, R t R

Relations between CVN, NDT, and DT test results for one of

the popular ship steels are shown in Fig. 15.57.
15.9. CRACKARREST
Fracture instability occurs, when upon crack extension, the ... ...
,.,
" ,/
elastic energy release, G, remains larger than the crack resistance, ....
....
""
R. The surplus of the released energy,.(G-R), can be converted into
kinetic energy (K.E.). This K.E. is associated with the rapid
--a -6a __ 0 --t!.a
movement of the material at each side of the crack path during
the propagation of a high velocity crack. The difference between Fig. 15.59. Crack arrest due to rising R or decreasing G.
G and R determines how much energy can become available as this principle is shown diagrammatically in Fig. 15.59. Both types
K.E., and consequently it governs the speed at which the crack of arrests i.e. decreasing G or increasing R, can be obtained in real
will propagate through the material. Both G and R represent the structures.
energies associated with a crack extension, da. Hence the total Arrest by increasing R may be promoted by the insertion of
amount of energy that can have been converted into K.E. after a strips of high toughness material; Fig. 15.60 shows the arrest
crack growth da follows from an integration of (G-R) over 6a. This
integral is represented by the shaded area in Fig. 15.58. :t..H , /
" ....
~
I ./

G,R 1
I I
Ekin G

I I

'"
'"
'"
'"
'" '"
'"
-
•••. I R=G,c
- - {. - - - - -R for rate
I sensitive material
I
I
II G,R t Wmin
I I
I I II
r
I II
, II
I

G, R 1

I
I I I-RstriP
I
I
I
\:-.~ ~Y~Q :1 -~Q I
I .."..""
.•.
R
n ."...~
Fig. 15.58. Representation of kinetic energy causing crack extension . .•.'" /
15.9.1. Principles and Method of Crack Arrest
I
I
_ac -6a -0 -6a
I (a) ( b)
Crack arrest may occur when G becomes smaller than R.
Alternatively, it can take place when R rises above G. Due to the Fig. 15.60. Welded crack arrest insert: (a) arrest achieved. and
(b) arrest not achieved.
I
1082 Welding Engineering and Technology Fracture Toughness and Weld Design 1083
.t,
r
behaviour in welded insert. When the crack penetrates the strip
the sudden rise of the crack resistance gives a situation where p strip
G < R. If the KE. is not used for crack extension a sudden arrest I
takes place. Under the circu.mstances where the K.E. can be
I
16!!. »
(0)
consumed as crack driving energy, the crack will grow into the P I rBOlted
strip u.ntil it has reached zero velocity (equality of shaded areas
in Fig. 15.60 (a»). For achieving this the width of the strip should
be in the order of (2rp + B). However, from Fig. 15.60 such a thumb
rule seems hardly applicable since it is likely that the arrest
capacity of the strip depends largely on the KE. and thus on crack
size. The arrest strip in Fig. 15.60 (b) has the same dimensions as
the one in Fig. 15.60 (a), but it is farther away from the crack tip
at the moment of instability. Since area A is smaller than area B G,R t
the crack runs through the insert and the arrest is not achieved
(line n). In the case of small crack (line m) arrest also does not (b) ,./
""

occur, due to steep G line. Though the situation is slightly different

""

""
•.•.·""R=G!c
in the two cases, but the principle remains the same. ".
".
""

01rest
".
It is unlikely that there are many structures where i.t is '"
possible to,make inserts of a high toughness material, particularly
because their spacing has to be small [Fig. 15.60 (b)]. The insert
_0 --60
Fig. 15.61. Crack arrest by bolted strip.
material usually has other un wanted properties such as lower
yield stress than the base material. If the base material was resulting G line and the crack arrest due to decreasing G curve is
selected for its static strength properties, a large number of inserts depicted schematically in Fig. 15.61 (b).
with much lower strength usually cannot be tolerated. It is also
not very attractive to insert rings of a low strength material in a 6
pipe-line. The wall thickness of the insert would have to be larger
because of the lower static strength. Apart from increased
resistance to the flow of the medium conducted by the pipeline,
the larger thickness would not be very efficient in view of the
ME

'
...
E
0\
.x
5

dependence of toughness on thickness. This problem can be solved

by laminating the inserts. Laminates of low thickness can be of
6 .•....
3 I
the same material as the main structure. The laminates will I
~ 2
enforce a plane stress condition at the crack tip and for this reason
I
1I R=EGc __
will have much higher crack resistance.
I
Alternative method of crack arrest by reducing G, under
conditions of constant load, can be achieved by the use of bolted 100 200 300 400 500
crack arrest strip as shown in Fig. 15.61. In the absence of the o(mm) -
strip, points A and B can move apart freely if the crack approaches
the line AB. The presence of the strip contains this displacement. Fig. 15.62. Result from crack arrest test with welded strip.
This means that the strip transmits the force P on the plate Welded strips have also been used in crack arrests and one of
through the bolts. These 'forces reduce the stress intensity factor the results achieved is shown in Fig. 15.62. Arrest took place at
and thus reduce G. The Gloser the' crack tip approaches AB, the G = Ge,so it seems there was no contribution of the K.E. to crack
larger the forces P, making the reduction of G more effective. The
lar.gest reduction occurs if the crack tip is slightly beyond AB. The propagation. It should be noted, however, that Ge was an
I
1084

estimated value. In the application of welded strip instead of Typical test of pressure hull
bolted one, care should be taken that the welds are only at the

Weld

Crock Crock

Os",p strip
Weld

(a) Short strip (b) long strip

Fig. 15.63. Welded strip as crack arrester: (a) welds at ends of strip, and
(b) short intermittent welds along length of strip.
Fig. 15.64. Conventional aircraft 'crack-stopper' construction showing tear
ends of the strip as shown in Fig. 15.63 (a). In the case of long stopper straps in the form of internal bonded-skin doublers.
strip, short welds should be made at short intervals. When a grow beyond the first double plate it intersects and is thus
continuous bead is laid length-wise along the strip, the crack may arrested, as shown in Fig. 15.64. Loss of pressure may occur in
grow into the weld metal and through.it into the strip as shown the aircraft, but the structure is still safe;
in Fig. 15.63 (b). The strip will then lose its effectiveness.
A properly designed crack arrester
following four basic requirements :
system must satisfy the

1. Be fabricated from structural materials with a high level

of notch toughness.
2. Have an effective local geometry.
3. Be located properly within the overall geometry of the
structure, and
Propagating
-
brittle fracture

4. Be able to act as an energy-absorbing system or defor-

mation-restricting system in cases such as gas-pres-
surised line pipes where the high rates of loading
continually deform the structure.
15.9.2. Practical Applications of Crack Arresters
By placing crack arrester at various locations within a Fig. 15.65. Typical design of in-plane crack arrester.
structure, excessive crack growth is arrested before the structure
The crack arresters are also used in ship hull design where
fails. This concept is widely used in the aircraft industry where
welded in-plane crack arresters are used,· as shown in Fig. 15.65.
bonded double plates or crack arresters are attached to the sIqn
Welded in-plane crack arresters are used in welded steel ship hulls
ofthe aircraft. Thus, if a crack occurs in the aircraft skin, it cannot
as integra1..1oad carrying components in conjunction with the
I~

1086 Welding Engineering and Technology Fracture Toughness and Wield Design 1087

IFW:S 36m the large movements of the ends of a ship and the large stored
W
4' or Less
Crock
arresters
elastic energy that continues to drive the crack through the crack
arrester as discussed in the previous section.
Crock Out-of-plane crack arresters may also be very practical because
arresters
01 each their configuration resembles the girders and stiffening members
corner(as commonly used in struCtures. An example of this type of geometry
current
prac1iu) is presented in Fig. 15.67 showing a wide-flange beam welded as
an integral part of the primary structure. Out-of-plane arresters
Typical cross-section of such as this would appear to have several advantages compared
ship hull struc1ures with in-plane crack arresters.

-~
Fig. 15.66. General location of crack arresters in a ship hull.
primary hu~l structure as shown in Fig. 15.66. However, the
arresters are usually made of materials with a much higher level
of notch toughness than the basic material in the hull structure.
For very large cracks, this type of crack arrester may not be very
effective unless the arrester plates are extremely wide, because of

Fig. 15.67. Schematic representation of out-of-plane crack arrester.

, '---
SIXTEEN
Heavy Welded Fabrications
Most of the heavy structures which used to be built earlier by
employing riveted joints are now most often welded. This is
because now it is possible to make high quality welded joints and
the consequential greater confidence in the performance of such
structures in service. Although there are many structures which
can be included in the category of heavy welded fabrications but
in this chapter the discussion is limited to the following :
1. Boilers and Pressure Vessels
2. Ships 3. Pipelines
4. Bridges 5. Offshore structures.
16.1. BOILERS AND PRESSURE VESSELS
Boilers and unfired pressure vessels are used in a large
number in different industries including chemicals and fertilizers,
power plants, cryogenic application, gas storage and the like. The
number involved is so large that Pressure Vessel Technology is
considered an independent field and its importance is recognised
even by International Institute of Welding (nW) by having a
separate Commission for it. Although books have been written to
cover all aspects of this field but in this chapter it is intended to
highlight briefly the following aspects :
1. Material selection,
2. Welding Processes,
3. Fabrication,
4. Preheat and postweld heat treatment.
16.1.1. Material Selection
All pressure vessels can be considered to consist of mainly the
shell, the nozzles, and the tubes. There are four main factors which
determine the selection of materials for pressure vessel shells :
(i) fitness for the application,
(ii) requirements of code,
(iii) cost, and
(iv) availability.
r 8my"".Id.d FabriQ';'M \089
16.1.1.1. Fitness for Service
Three main factors are normally involved in determining the
suitability of a material in service. These are :
(i) Corrosion resistance,
(ii) Resistance to the effect of temperature, and
(iii) Fabricability.
Corrosion Resistance : Acceptable corrosion rates vary
greatly according to the nature of the industrial application and
the severity ofthe environment. In the petroleum and petrochemi-
cal industries it is normal to expect an average life ofabout 10years
for pressure vessels. A typical value for standard corrosion al-
lowance is 3 mm, with a maximum permissible corrosion rate of 0.5
mm per year. Where corrosion rates are extrem~ly high as in chemi-
cal manufacturing industry, no fixed rules about vessel life or ac-
ceptable corrosion allowance can be made. In such cases it may
become economic to accept a short life of even 1 to 2 years for the
vessel.
In some of the pharmaceutical, food and fine chemical industry
the most essential requirement is to avoid the contamination of the
product. In such cases material selection is governed by the rate of
metal pick up by the process fluid and not by the considera.tion of
vessel life. Equipment handling unde -aerated demineralised
boiler feed water must also be selected to avoid contamination.
,'. Stress corrosion cracking is particularly hazardous form of cor-
rosion because it can result in the catastrophic failure of a vessel.
The most common agents promoting such cracking are chlorides
and caustic alkalis in the case of austenitic Cr-Ni steel, and caustic
alkalis, nitrates and H2S in the case offerritic steels. For austenitic
Cr-Ni steel equipment subjected to chloride contamination it is bet-
ter to use a stainless clad steel rather than solid stainless steel. For
caustic alkali stress corrosion cracking of carbon steel, the effective
remedy is to use suitable stress relieving treatment.
Both metallic and non-metallic materials are used for corrosion
resistant pressure vessels, the latter generally as a lining of the
metallic shell. For low temperature applications up to 200°C,
linings of baked or air dried resin or rubber may be used while for
high temperature applicatipns acid resisting bricks or tiles made
of plastic ceramic or carbon may serve the purpose best.
If steel is selected for use then its properties can be modified
during manufacture by the addition of suitable deoxidiser like Si,
Al, Mn, etc. In addition to its use purely as a deoxidiser-, Al may
be added to steel in small quantities in order to refine the grains.
Grain refinement increases the yield strength of the steel and, at
the same time, improves the notch ductility. The creep rupture

1090 Welding Engineering & Technology
strength ofsuch aluminium-treated steel is however inferior to that
of silicon killed or semi-killed steel in the temperature range of
350-450°C. Thus, the optimum selection of carbon steel for low
temperature use will generally be an Al-treated C-Mn steel, while
for elevated temperature service a Si-killed steel, with Al addition
limited to 225 g/ton is to be preferred. However, it must be kept in
mind that steel deoxidised with Al or Mn or Si is prone to lamellar
tearing in case the deoxidation products like manganese silicate
gets unfavourably placed during rolling of steel. For highly critical
components like nuclear pressure vessels vacuum poured steel
should be preferred.
Effect of Temperature on Mechanical Properties :
Material selected for use in pressure vessel fabrication must
be adequate for the service temperature it is intended to be used.
Temperature effects may be grouped in three ranges, viz.,
(i) Sub-zero temperatures, below O°C
(ii) Atmospheric temperatures, 0-20°C
(iii) Elevated temperatures, above 20°C.
(i) Materials for Sub-zero Temperatures
Material for use at sub-zero temperatures is required to possess
good notch ductility. Metals with face-centred cubic lattice structure
i.e. all commonlyused metals except ferritic steels are not much af-
fected by reduced temperature so far as notch ductility is concerned.
However, the main problem of material selection for sub-zero
temperature is to find out as to how far it is possible to use ferritic
steels, generally carbon and C-Mnsteels. Brittle fracture tendency of
such steels increases with increase of thickness. In the refrigeration
industry such materials are commonly specified for temperature as
low as -120°C because ofreduced stress encountered at reduced pres-
sures involved in such circuits.
Grey C.!. may also be used at sub-zero temperatures since its
brittleness is not much affected by reduced temperature; while the
design stresses are correspondingly low. However, cast iron is not
recommended for pressure containers handling inflammable or
toxic fluids.
For higher design stresses, cast steels are used down to
_ 60°C. However, these steels generally contain low carbon with
increased MulC ratio and aluminium addition for grain refinement.
Quenched and tempered C-Mn steels al'e available for use up to
-60°C but low alloy ferritic steels are often used for the tempera-
ture range of -50 to -100°C. 1% Cr 0.25% Mo steel and Ni-steel
containing 1%and 5%Ni have been specified in this range, but the
most widely used alloy is 3.5%Ni-steel which is welded with match-
ing alloy electrodes or less frequently, with austenitic Cr-Ni or Ni-
base electrodes. With ferritic electrodes a careful welding
•
Heavy Welded Fabrications 1091
procedurp. (in particular the avoidance of high heat input rates) is
necessary to get acceptable impact properties in the weld at -100°C.
The non-availability of small fittings in 3.5% Ni-steel can be over-
come by using austenitic Cr-Ni steel fittings which can be welded
to 3.5% Ni steel shell or pipes provided properly qualified welding
procedure is used.
Although 5% Ni-steel can be used down to -120°C, but in
general below -100°C austenitic Cr-Ni steel, 9% Ni-steel, Al and
Cu are used for pressure vessels. All these materials are notch
tough down to the lowest temperature used for industrial plants.
Austenitic stainless steels are readily available and easily fabri-
cated but are more costly than 9% Ni-steel or aluminium. 9% Ni-
steel has a relatively high strength and is welded with Ni base
filler materials. Aluminium - usually the Al-1.25% Mn or one of
the Al-Mn alloys has good availability and is economically competi-
tive; however Al-Mn-Si alloys are avoided in low temperature ap-
plications due to their low ductility. Welding aluminium to pressure
vessel standard is not easy and is thus best done by specialist firms.
(ii) Metals for use at Atmospheric Temperatures
Pressure vessels codes do not, in general, have mandatory
requirements for notch ductility for operating at temperatures
above O°C.Nevertheless special precautions against brittle failure
during fabrication and hydrostatic testing need to be considered
for vessels made from thick plate. The best material from this view
point is one which is least subject to cracking during welding
(particularly hard zone cracking) and cracking during the thermal
stress relief treatment i.e. reheat cracking. If it is possible to use
a C-Mn steel, the fabrication risks will be considerably reduced
but for high pressure vessels weight limitations imposed by
handling and transportation often make it necessary to accept
alloy sections, while Ni additions improve the inherent notch
ductility of the material. Except for very thick sections, however,
vanadium treated steels are widely used which when properly
fabricated are perfectly satisfactory.
Pressure vessels for nuclear reactors raise the additional prob-
lem of CK>ping with an increase in the brittle/ductile transition
temperature during service, primarily as a result of neutron ir-
radiation. Results of specific neutron doses of 4.0 x 1018 and
1.0 x 1019 n/cm2 (fission) have shown increase in slow bend transi-
tion temperature of various steels and weld metals in them. With
one or two exceptions, maximum embrittlement is produced in the
steels and weld metals by irradiation at 130° to 150°C. While the
Si-killed 0.24% carbon steel shows no effect ofirradiation tempera-
ture between 50° and 130°C, the magnitude of the embrittlement
in aluminium grain size controlled and low alloy steels is markedlv
 Heavy Welded Fabrications 1093
1092 Welding Engineering & Technology
Table 16.3.Characteristics of BS 1501 Steels
dependent on the irradiation temperature in the same temperature
range. At irradiation temperatures above 150°C annealing of the Grade (]}fin. UTS, Practice
damage occurs during irradiation, resulting in smaller changes in
transition temperature. However, the temperature ranges in which
141
,
I
N/mm2)
355
-----
Rimming
----------------
Up to 19 mm for hard glass ena-
recovery of most of the damage occurs during irradiation are I

markedly dependent on the type and composition of steels. -'RS 1501'rTen,~ile Grafes r;eoXidatioI melled pressure vessels and tanlts
Function
Whereas some ofthe steels and weld metals show no embrittlement 151 355 i Semi-killed with Up to 100 mm. General purpose
400 Si; Al addition carbon steel for pressure vessels,
after irradiation at 325°C, ot.hers exhibit significant increase In 432
transition temperature after irradiation at this temperature. not more than permitted by BS1500 up to 400°C,
Table 16.1. Selection of Material for Non-Corrosive Service 100 gm/tonne BS 1515 up to 500°C
154 355 Semi-killed with Up to 10 mm for use at ambient
at Sub-Zero and Atmospheric Temperature: Low Values of
----------------
.------
Pressure x- Diameter (other than Refrigeration Equipment)
boiler
3.5%
quality,
grain Thin ofplate
Impact
steel Nitested
Aluminium, forsteel
9%
(say
the 5%
Ni
Ibelow
More
carbonNi
Steel,
Carbon
plate. 13 -
issteel
usable down
-120°C,
conservatively,
mm):
steel
thicker may
--. to Alternatives
2.5%
use Ni
finedown to -60°C
be acceptable
400
432
aluminium temperature

or better.
te: killed steel, B81501-161
-29°C
withouttoimpact
ASME orI test
better
code.
e.g.Material
down to Selection 161 355 Silicon-killed; Al Up to 150 mm. Generat purpose
400 addition not Si-killed steel for ambient and
432 more than 100 elevated temperature. Permitted
gm/tonne up to 510°C by BS1500 and 500°C
by BS 1515
211 400 Semi-killed Up to 100 mm. Higher tensile C-
432 Mn grade for ambient and
463 elevated temperature -15°C to
494 500°C.
213 432 Semi-killed; Up to 100°C. Higher tensile C-Mn
463 niobium treated grade with niobium addition to
494 promote fine grain and hence
improved yield and notch-impact
The pronounced variability of the effects of irradiation at properties. Niobium is not a
deoxidant and so can be added to
temperatures in the range of 250°C to 350°C is of considerable semi-killed steel. Rupture
practical importance since this is just the temperature range in strength inferior to grades 151,
which many pressure vessels, such as those for water-cooled 161, etc. Use -30°C up to about
reactors, have to operate. 350°C.
A typi~al material selection chart for sub-zero and atmospheric 221 400 Silicon killed Up to 150 mm. Higher tensile C-
design temperatures is given in tables 16.1 and 16.2. Steels 432 Mn Si-killed grade. Lower yield
referred to in these tables are further detailed in table 16.3. 463 and impact properties than 213
494
Table 16.2.Selection of Material for Non-Corrosive but creep properties equal to
Grade 161. Usable -15°C to 500°C.
Service at Sub-zero and Atmospheric Temperature:
r--- HigherGrade----
Values -----
of Pressure
design x Diameter 223 432 Silicon killed; AI Up to 150 mm. Otherwise
3g.O°C
or to
224.
BS1501-224
20°C for steel
Ni some may
---
9%
LT.50for BS1515
3.5%
Impact
itemsbeI necessary
Nickel
Material
nickel
tested, steel
Substit~tion
Selection.
steel
A516
grain
Impact
Impact
9% standard
steel.
Ni-steel
fine tested
grain
Fine grain killed steel such
ofsteel
Alterna.tives
tested
may
C- austenitic
semi-killed Cr-
fine
be considered
to ASTM 463
494
addition
more than 100
not properties as Grade 213.

gm/tonne
224 400 Silicon killed; Up to 150 mm. Higher tensile C-
432 AI-treated to Mn fine grain steel with adequate
463 give grain size 5 impact properties down to -50°C.
494 or finer Primarily for low temperature,
elevated temperature properties
inferior to Grade 161. Use -50°C
to about 150°C.
-;"';~""';-;"",-"""-"'"--------'----------------

Heavy Welded Fabrications 1095

Welding Engineering & Technology
The strength of a ferritic steel is dependent mainly on three
factors: the nature of the transformation products i.e. the
structure, the degree of solution hardening, and precipitation
effects. Transformation products comprise ferrite, pearlite,
bainite, martensite, etc. Carbon, C-Mn, and C-Mo steels with
the low carbon contents used in pressure vessel steels are
normally pearlitic, i.e. the structure comprises ferrite and
pearlite.
In selecting materials for use at temperatures in the creep
range, resistance to overaging is particularly important. The
presence of excess carbon above that which is combined with the
alloying elements may result in the precipitation of cementite,
Fe3C, which can be damaging to the creep resistance and carbon
contents in the region of 0.15% appear to be optimum for creep
resistance in low alloy steel. Creep properties are further
improved if the composition is such that appropriate precipitation
occurs during the creep process. As already mentioned
uncombined nitrogen can generate such precipitation in carbon
steel. Also, in CoMosteels carbides are precipitated during creep
thus enhancing the solution hardening effect of Mo. This effect
becomes noticeable with quite small additions of Mo, so that
1.0 t- Ultimate strength residual Mo contents in carbon steel can significantly affect the
30f- rupture strength. For austenitic Cr-Ni steels used in the creep
range similar considerations apply. Optimum creep properties are
20 obtained with 25% Cr-20% Ni steel at higher carbon contents than
Lower yield stress in low alloy steels in the range of 0.40%.
.S To further augument the properties of steel it is necessary to
quench and temper it. Quenching and tempering, apart from the
structural changes induced, generates a fine carbide dispersion
Stress tor combined with a very fine ferrite grain size. Good tensile
Ii 1% creep in properties are thus combined with a low impact transition
-ell
...
'"
100 000
hours temperature .
The other major requirement for satisfactory use at elevated
temperature is that the metal should not deteriorate
30 significantly in service. Metallurgical changes, such as temper
embrittlement, shperoidis,ation, graphatisation, carbide
precipitation, or sigma phase formation may be damaging and
if so must be avoided while corrosion rates must be kept within
reasonable limits.
1 1.00 500
o 100 200 300
Temperature:C

Fig. 16.1. Typical properties of semi-killed carbon C-steel at elevated temperatures.

-;"..,_ ~.'"":·""~~~~"'~,",T,~.~'~· _._~--~-_._~~-~~~"~~~~~~~~
-,,_._. __ ..

1096 Welding Engineering & Technology Heavy Welded Fabrications 1097

Table 16.4. Selection of Steel for Resistance to form iron + graphite to give a structure of low tensile strength.
Scaling in Air or Steam Graphatisation reaction is a long term phenomenon but can be
Alternatives
rapid in HAZ of fusion welds. It is also more rapid with AI-treated
Temperature range Material selection than with Si-killed or semi-killed steels. For the petroleum
(OC)
industry it is recommended that carbon and C-Mo steels
up to 500 Carbon steel equipment exposed to temperatures of 450°C and over should be
Higher Cr-steels, including 18/8 inspected for graphatisation after 35000 hours in service. Cr and

i
500- 550 1 Cr"2 Mo types
550 -600
2 Cr1 1 Mol
600 - 800 18/8 Austenitic Cr-Ni 9CrlMo and 13Cr steels
steel usable up to 700°C
800 - 1050 25/12 Austenitic Cr- 25/20 Austenitic Cr-Ni steel
Ni steel
1050 - 1100 25/20 Austenitic Cr- 20Cr80Ni alloy
Ni steel
Note. In the petroleum industry the upper temperature limits for excessive scaling
of carbon and low alloy steel may be somewhat higher. particularly for low
stressed parts.
Above 600°C, carbon steel is oxidised at an unacceptably high
rate both in air and steam. Chromium in an amount of 3% and
above significantly reduces corrosion rates and virtually all heat-
resisting steels contain chromium for this reason. Table 16.4 shows
a possible scheme of selection of unstressed parts where resistance
to oxidation is the only consideration. Because of its easier
availability and fabric ability, it is often better to specify austenitic
Cr-Ni steel for unstressed parts like internals and liners operating
in the temperature range of 500-600°C. Unstabilised AlSI type 304
steel may be used up to 800°C but if the steel is to be stressed one
of the stabilised grades may be selected because of their higher
tensile strength at elevated temperatures. Welded sections thicker
than about 20 mm in austenitic Cr-Ni steel other than the Mo bear-
ing grade AlSI type 316 or 32%Ni-22%Cr (Incoloy 800) may suffer
stress relaxation cracking during PWHT or in service if there is an
imposed stress. For thick sections type 316 steel would be an op-
timum choice but for the temperature over 650°C it may be sub-
jected to accelerated oxidation rates. Such high rate oxidation is
often associated with stagnant atmospheric conditions, and contact
with insulating materials containing sodium silicate or similar low
melting substances. Where such unfavourable conditions exist it is
imperative to specify the 32%Ni-22%Cr alloy, which is not subject
to either of the above mentioned defects.
Graphatisation is a potential hazard with carbon and C-Mo
steels in which the carbide is cementite, Fe3C, At service
temperatures of 450°C and above cementite may decompose to
Cr-Mo steels containing more than 0.7% Cr are resistant to
graphatisation.
are Mo-free 5% Cr steels suffer from temper brittlement when
held in the range of 325°-625°C, as a result brittle failures have
occurred when these steels were used for heater tubes used in
refineries for processing corrosive oils. This defect can be
eliminated by the addition of Mo;that is why 5% Cr 0.5% Mo steels
are now used for heater tube::>for above mentioned application.
Mo is added to other Cr-steels like 1%CrO.5%Mo and
2.25%Cr1%Mo for controlling temper brittleness as well as for
improving tensile and creep properties. Also, Mo increases
resistance of steels to hydrogen attack.
Hardening due to carbide precipitation at elevated
temperature i.e. secondary hardening has been reported in a few
cases in service but is avoidable by proper PWHT. Cast 25%Cr
20%Ni steel with 0.4% carbon loses room temperature ductility
due to Cr-Carbide precipitation when this alloy is used for
reformer furnace tubes operating at 850°-950°C. Its ductility at
the operating temperature is however adequate. The same alloy
with lower carbon content (0.05-0.25%C) in the wrought form is
only slightly affected by carbide precipitation but in the
temperature range of 600°-850°C it may be embrittled by
precipitation of sigma phase. The Mo-bearing austenitic Cr-Ni
steels may also be subject to sigma phase precipitation but the
effect is less severe and there are grades with higher nickel
content that do not form sigma phase. Carbide precipitation occurs
in unstabilised austenitic Cr-Ni steels if the operating
temperature is within 450°-750°C. Use of a Nb or Ti stabilised
steel minimises this risk.
The metallurgical stabilit.y of austenitic Cr-Ni steels, in
general, improves with increasing Ni content and the 32% Ni 22%
Cr alloy, containing small additions of AI and Ti (Incoloy 800)
represents an optimum combination of stability with good
strength at elevated temperature.
Table 16.5 and 16.6 provide the guidelines for select.ing
materials for elevated temperature applications.
1098 Welding Engineering & Technology· Heavy Welded Fabrications 1099

Table 16.5.Selection of Material for Non-Corrosive Service at 3. Shielded metal arc welding,
Elevated Temperature: Low Values of Pressure x 'Di~meter 4. Gas tungsten arc welding or TIG welding
Temperature Material selection Alternatives 5. Flash butt welding,
20°-450°C Semi-killed carbon steel e.g. Si-killed steel BS 1501-161 I' 6. Oxy-acetylene welding,
BS 1501-151 7. Electric resistance welding (ERW Process),
450°-500°C Si-killed steel BS 1501-161 Low alloy steel 1CrO.5Mo 8. Gas metal arc welding or MIG welding,
500°-600°C 1CrO.5Mo 2.25Cr1Mo 9. Flux cored arc welding (FCAW),
Above 600°C Refractory-lined carbon or Austenitic Cr-Ni steels 10. Special welding processes.
low alloy steel Brief description of some of these processes with particular
Table 16.6. Selection of Material for Non-Corrosive Service at reference to their use in Boiler and Pressure Vessel fabrication
Elevated Temperature: Higher-Values of Pressure x Diameter follows.
(BS 1515Design) 1. Electroslag Welding: For the pressure vessels involving
Temperature Material selection
Silicon killed C-Mn steel BS
-----
Alternatives
Proprietary high tensile alloy
plate thicknesses exceeding 50 mm the most economical welding
process is electroslag welding (ESW). This process has the
20°-450°C
1501-221. For thick plate steel such as Ducol W.30 advantage of welding any desired thickness in single pass using
(say over 75 mm) specify one or more electrodes. The process is fast and comparatively
impact testing simple though it does involve certain unique features like vertical
450°-600°C 2.25Cr1Mo steel 1CrO.5Mo Proprietary steels placement of shell and accordingly the corresponding movement
Above 600°C Refractury-lined carbon or Austenitic Cr-Ni steels of the equipment.
low alloy steel Apart from the factors of speed and simplicity of welding
Note. The economic temperature b:-eakpoints will vary with code, local conditions. resulting in reduction in cost, one of the attractive features of the
steel prices, etc. electroslag procedure is the freedom from defects when operations
16.1.2. Fabrication of Conventional Pressure Vessels are normal.
Having selected the material for the pressure vessel the next ESW has, however, two distinct disadvantages viz,
step is to fabricate it into the desired shape. That involves (i) its application to circumferential seams is rather cum-
mechanical operations (like rolling, press working,etc), assembly, bersome, and
welding, postweld heat treatment (PWHT) and inspection. Typical (ii) that the coarse grain structure produced by this process
CtJst breakdown for these factors may be as follows : requires to be refined by normalising procedure. Apart
(i) Material, 37% from its mechanical drawbacks the coarse grained struc-
(ii) Preparation, mechanical operation and assembly, 47% ture renders ultrasonic inspection more difficult to in-
(iii) Welding and PWHT, 14% terpret satisfactorily.
(iv) Non-destructive testing 2% One procedure for controlling the as-welded grain size depends
Important aspects of welding involving fabrication of main on selecting the welding parameters so that the weld metal fuses
shell, welding of end plates, nozzles and tube-to-tube plate welding, the side walls of the joint preparation without excessive
as well as PWHT are described briefly in the following sections. penetration and with a travel speed so that the weld metal
16.1.2.1. Main Seam Welding solidifies rapidly. The fact that these conditions are attained at
speeds two to four times faster than in the normal procedure is
The processes mainly used for welding different joints in an added advantage as it leads to reduced welding costs.
hoilers and pressure vessels include the following:
Typical mechanical properties obtained by such a refinement
1. Electroslag welding, in a weld of2.25%Cr 1%Mosteel after a stress relief at 650°C were
2. Submerged arc welding, as follows:

I
1100 Welding Engineering & Technology
Yield point 490 N/mm2
Tensile strength 716 N/mm'?
Elongation 24%
Reduction of area 63%
2. Submerged Arc Welding: Due to the technica.l difficulties
and the cost involved in welding circumferential joints by
electroslag welding most of the heavy pressure vessels have their
circumferential seams welded by submerged arc welding:
Submerged arc welding (SAW) is also a more economical
process than electroslag welding for wall thickness of up to about
45 mm. SAWis now a well established process and the variations
in procedure are largely determined by consideration of cost and
ease of application. Reference, however, can be marIe to the proce-
dure whereby all the welding is undertaken from one side. This
usually involves a preparation of the type shown in FiR. 16.2 in··
Pressure vessel
wall
Backing strip
Fig. 16.2. Weld edge prepamtinl1 for multi··layer submerged arc welding.
volving a backing strip which has to be removed, followed by the
necessity afnon-destructively testing the base of the weld preferab-
ly by the magnetic method, to ensure the detection of small surface
cracks arising from the corners where backing strip meets the plate
material.
For achieving higher deposition rates SAW with 2 to 3
electrodes, each having its independent power source, is employed.
This results in faster welding speeds but may lead to two dIstinct
disadvantages:
(i) the deposited metal may not attain full normalisation of
the multi-layer method of single electrode SAW,
(ii) it limits the minimum diameter of the shell that can be
successfully handled be~ause if the diameter is t.oosmall
control of big sized molten metal weld pool may be
beyond the capabilities of the operator and the equip-
ment.
Heavy Welded Fabrications 1101
Though no fixed limits to the minimum diameters are laid, but
an approximate guide is that for 2 electrodes the minimum
diameter is about 1800 mm and for 3 electrodes it is about 2500
mnl.
The recent trend is the use of narrow gap welding wherein the
walls are kept almost parallel or with a slight included angle. This
leads to reduced preparation costs and the weld metal deposit is
considerably reduced leading to faster welding speeds at reduced
cost. This also results in reduced distortion and or residual
stresses.
3. GMAWIFCAW Process: Gas metal arc welding and
flux-cored arc welding are not employed in the fabrication of thick
walled pressure vessels. However, because in the welding of
circumferential welds of small diameter vessels internal access
can be a major problem therefore it is considered perfectly feasible
to deposit the first run or runs by CO2 welding to form a base of
sufficient thickness to permit subsequent submerged arc welding.
Of greater importance, however, is the use of MIG welding with
Ar-shielding for welding of austenitic stainless steels. Owing to
the relatively high cost of this material it is normally used only
for thin shells and for such applications MIG welding with
Ar-shielding is well suited.
4. Shielded Metal Arc Welding (SMAW): In the modern
pressure vessels' fabrication, it is not common to use SMAW for
welding the main seams. Largely this is the result of cost
considerations and also in some instances because of the
non-availability of a sufficient number of qualified operators.
There is also an underlying belief that automatic methods produce
more reliable welds. It is of course, quite erroneous to believe that
SMAWof main seams cannot be made just as satisfactorily as by
an automatic process. This point is well proved when a very large
sized pressure vessel (45m long, 1675 mm internal diameter made
from 135 mm thick plate, weighing about 300 tons) was fabricated
in three pieces and site-welded, in U.K., using SMAWprocess to
give a very satisfactory service. Thus, for many of the large
pressure vessels required in the modern power and petrochemical
industr~es this procedure of fabrication can be very effectively
utilised though there may be certain disadvantages associated
with time and space requirements.
16.1.2.2. Nozzle Welding
The problems involved in the nozzle attachments are some of
the most important in pressure vessel technology particularly as
regards fatigue str~ngth; and this is particularly so for the heavy
.".~=., _"'~._,~~.c.""_,_._,_, __
...~"'__.. -_·_·
----------....,..-----------------------

1102 Welding Engineering & Technology Heavy Welded Fabrications 1103

pressure vessels now in demand by the power and chemical
industries.
There are now two main designs of nozzle attachment viz.,
(i) A nozzle penetrating the full plate thickness i.e. a set
through design;
(ii) A nozzle mere or less set on the plate surface.
In both the cases the design should avoid the so called compen-
sating plates i.e. separate plate material added to the shell to com"
pensate for the loss of strength due to hole made in the plate to
attach the nozzle. Particularly for thick pressure vessels the
provision of compensating plates involves a definite hazard by
facilitating weld metal cracks at the interface between the compen-
sating plate and the shell plate caused by the almost impossible
necessity of avoiding gap there at the assembly stage.
For the set through nozzle a full penetration weld is decidely
better. This arises from the difficulty of making partial
penetration welds of the form shown in Fig. 16.3 free from cracks
Nozzle
at the roots of the welds since such welds have almost of necessity
to be made by SMAW. Also, the stress concentrations produced at
the roots of the weld are a source of weakness to resist the thermal
stresses that inevitably arise when such vessels are used at other
than ambient temperatures.
F611 penetration welds of the design shown in Fig. 16.4 are a
better proposition as in many instances automatic machines could
be advantageously used. For thick materials, however, restraint
placed upon the weld metal can be very high therefore requires a
very careful consideration as regarding material and procedure
selection.
Pressure
vessel wall

Weld Fig. 16.5. Set-on nozzle weld.

Thus, where possible a set-on design such as shown in Fig. 16.5
Pressure can be recommended. If manual welding is involVed the welder has
vessel wall almost the easiest access possible and he has not to work inside a
vessel subjected to a high preheat temperature. Also, fully
automatic machines can be employed to accomplish this work
satisfactorily.
Although the above described nozzle attachments are the
Fig. 16.3. Partial penetration nozzle welds. standard ones but some of the nozzle design variants, including
pipe branches are shown in Fig. 16.6. In this figure only one out of
several possibilities is illustrated in each case as regards weld posi-
tion, type of weld and groove shape alongwith only one specific com-
bination of dimensions. The nozzles (a, b) which butt against the
vessel or are partially inserted, offer little reinforcement acting on
the wrong side of the wall and high notch stresses arise in the acute
angled transition areas. Full penetration of the vessel wall by the
Pressure nozzle pipe (c) improves the reinforcing effect strongly and on the
vessel wall right side of the wall; this method of nozzle attachment, however,
is unfavourable as regards fluid flow and corrosion factors. Necking
out of the vessels wall or of the nozzle pipe (d) does not offer much
reinforcement but reduces the notch effect strongly and moves the
weld outside the zone of high notch stresses. An annular plate
Fig. 16.4. Full penetration nozzle weld.
-~~~~-- Weld;n. En.lneerin. & Te<bnnlogy rHea.,. Welded Fabri<aUons 110S
16.1.2.3.Tube to Tube-Plate Welds
In the petrochemical industries, heat exchangers play a
prominent role, and their successful operation depends, to a large
~~
(0) ( b) ~ IJ extent, on the quality of tube to tube-plate welds so their
production has received considerable attention.
1 Depending upon the service requirements, it is possible that
i the simplest and easiest type ofweld, shown in Fig. 16.8, may serve
i I --
i I -li . I

(e)
~-r
if)
fj'~-t
--+ ~~]~-!
L_.-1l_-1
(g)
:
( h)
II !

Fig. 16.6. Nozzle design variants (including pipe branches) : (a, b) set-on or
partially set-through nozzle pipe, (e) fully set-through nozzle pipe, (d) necked Tube plate
out vessel wall, (e) annular plate reinforcement, (f) section shaped nozzle ring,
(g) unwelded reinforcing ring, and (h) collar reinforcement.
placed around the nozzle pipe on the vessel wall (e) is effective as
regards reinforcement, but is combined with relatively high notch
stresses on the face and nozzle pipe sides. Adjusting the annular
plate to shape is also cumbersome at times. A section shaped ring
(f) fitted using butt welds, is a particularly advantageous solution
which is, however, expensive. A design with a collar welded on (h)
is effective as regards reinforcement but is also notch sensitive;
moreover, for sufficiently good support, the collar needs to go Fig. 16.8. Simplest form of tube-to-tube plate weld.
around and main body of sphere, pipe or tank as well as the nozzle
pipe. An alternative design in the case of bifurcated pipe branches the purpose well. The weld is relatively easy to make either by
is a sickle shaped internal web plate. A reinforcing ring (g) applied manual metal arc welding or by using an automatic welding head,
without welding is suitable only for nozzles on spherical vessels which is normally used on the tube-plate in a horizontal position,
because it is only here that the joint is suitable for the necessarily and requires very little attention. It should, however, be kept in
planar reinforcing ring. It represents an example worth noting for mind that for all tube to tub-plate welds it is not at all wise to
the avoidance of the high notch stresses of a transverse fillet weld expand the tube tightly into the tube-plate; because if it is done
by dropping the weld completely. A good attachment design for some air will be entrapped and that almost definitely leads to weld
welding an eccentric nozzle pipe for draining the vessel is shown in porosity. For the same reason the materials being welded should
Fig. 16.7. be clean particularly requiring freedom from cutting oils, etc.
For making a better weld than the single fillet weld described'
above, it is advantageous to use at least two runs. This is not
necessarily to strengthen it but to overcome the possibility of lack
of fusion at the commencement of welding. However, for critical
+-t---
-t--
I
':
I
----,
t-'-~,'0NOZZle
-10=-- Pressure
:- Weld
vessel
applications such as are encountered in nuclear power plants, this
much precaution to avoid lack of fusion may not be adequate.
Automatic TIG welding machines are now available for use
with the tube-plate in a vertical position. This eliminates the I').eed
for change of position of the vessel to accomplish the weld. A
suitable design for use with such a machine is shown in Fig. 16.9.
Fig.. 16.7. Welded ecc.entric nozzle pipe.
~. ,~,.,"I"c.:.•. '!!
...
-.,!. ,!I--"l!_!11- •. _---,~
.. ~!!
111111111I11I11I
!!!-.!!!- •.

1106
.- ••••••• ---------------

Welding Engineering

Tube plate
& Technology r---
Heavy Welded Fabrications

Typical completed ••••

--=====-
eld configuration
1107

Fig. 16.9. Design for tube to tube plate weld with plate held in vertical position.
This TIG welding machine can be employed for welding most of
the steels (e.g. carbon, low alloy and austenitic' steels) as well as
the non-ferrous metals. It is, however, restricted to welding tubes
_ with a maximum wall thickness of about 3 mm. A CO2 welding
machine can be used for welding tube to tube-plate with a joint
preparation given in Fig. 16.1.0. Where a much stronger connection

/
I a) Back face inset bore ••••
eld ( b) Groove ••••
eld
Recess

Tube plate

•••.
eld

Fig. 16.10. Joint edge preparation for welding with C02 welding machine.
is required, the recess can be increased in depth to about 10 mm
involving three runs of weld metal; for a tube 115 mm o.d. the
arcing time for the complete weld is approximately 5 minute-so
For avoiding crevice corrosion the problem is solved by bore (c) Bock face stub bore ••••
eld
welding a joint of the types shown in Fig. 16.11. Fig. 16.11. Joint preparations for bore welding of
The selection of joint dependS' upon two factors uiz., the cost cervice-free tube-to-tube plate joint.
of machining the tube-plate and the ease with which a weld found Explosive welding is perhaps the latest method being
faulty during fabrication can be removed and- rewelded to the successfully used for tube to tube-plate welding. This process is,
required standard. For heat exchangers necessitating this type of however, more often used for production welding of such joints.
tube to tube-plate weld the accuracy in machining and assembly
are highly important and it is particularly so for the very high Some other tube to tube-plate joints are shown in Fig. 16.12.
Shape (b) is suitable for sealing but larger axial forces can be
quality reql1irements.et:the nuclear power industry.
1108 Welding Engineering & Technology
Heavy Welded Fabrications 1109

B
Ring groove Tub e plate Tube plate

I~~~] R (a)
relief
Stress
groove

---+
( bl
I
I
I ,
/ ----+
I
---1I
c)
I
.
----+
!d I
I
I

~~d~--J 1__
la) ( bl (c)

Fig. 16."12.Some other joint preparations for tube-to-tube plate welds.

carried with fatigue resistance by shapes (a) and (c). The ring
groove in (b), which takes up more space, is intended for relieving
the stresses in the weld.
16.1.2.4. Block Flanges
1__
-1 t--J
lel (f! (gl
Fig. 16.14. Some typical designs of pressu.re vessel and joints.
(h)
-1
Block flanges are compact ring bodies weld jointed into or on 16.1.2.6. Welded-on Bolts
to shell plates, as shown in Fig. 16.13. These flanges offer a Awelded-on bolt, either placed on or penetrating a surface and
jointed by a fillet weld, as shown in Fig. 16.15, is a relatively high

wJIJ~~ I a) ( bl Ic )
~ (d)
~
Ia)
Fig. 16.13. Butted (a, c) and inserted (b, d) types of block flanges. t t)
connection facility via a bolted joint including the machined Fig. 16.15. Different types of welded-on bolts; (a) butted with
jointing plane and usually enclose a circular hole. Design (d) with fillet weld, (b) penetrating with double fillet weld, and (c) penetrating
with double bevel butt joint with fillet welds.
a large rounding radius in the transition area and welds outside
the notch area have the highest level of fatigue strength. Versions quality design. The penetrating type (b) is able to withstand higher
(a, b, c) with fillet welds and with or without a centering shoulder loads than the version which butts against the surface (a). This is
have a lower strength. A further risk of such a design is the risk because of doubling the weld in (b) and thus removing the risk of
of bulging at the unfused slit face at the weld root during lamellar tearing, A double bevel butt joint with double fillet weld
annealing or drilling. The solution of this problem lies in avoiding . (c) is able to withstand the highest loading. A stud-welded bolt;
the unfused slit face or drill ventilation hole as in (a). The unfused when used, has the least load bearing capacity.
slit face should always be sealed to avoid crevice corrosion. Non-critical
Non-critical
component -
16.1.2.5. Vessel Ends component --
2 tor 50mm
whichever is
Some of the typical pressure vessel end joints are shown in greater
Intermediate
Fig. 16.14. Types (a), (e) and (f> are unsuitable in the presence 'of
fatigue loading; types (b), (c) and (g) are suitable subject to certain
Lt1
IntermedlQt~
part t -oo/.-'~-
2 t, or SOmm

L
whichever IS t1
part
-L t,

limitations; types (d) and (h) are well suited to the purpose. In
type (b) the groove is supposed to relieve the notch stress at the
weld root but that is not really the case. Type (g) can be f
I greater

! ~
\
~f
Pressure shell Pressure shell
considerably more loaded than type (c) because of more favourable
force flow and the slit face outside the main flow. Fig. 16.16. Details for attaching non-critical components
by welding to pressure vessel shell.
'''j'''' '''",' ",_'iliiC""·ri"',,"c'·"'''"''''~'''''_ ..' ( --------------------------
i
i
I

Heavy Welded Fabrications 1111

I
1110 Welding Engineering & Technology
U U U I
§.§ u U
U
o oo o
'"
I~
o
U
o o I Examples of welded details for attaching non-critical
o
I~ components to pressure vessel shell are shown in Fig. 16.16.
<Xl •.... •....
,E ~, u., a :r, 16 '" <J:J '" '" '"
>< '" £ £ £ £ £ £
"'~
<l>

C') UIU
<D
'"
'"
C')
<J:J
'" Temporary attachments welded to a pressure vessel be kept to a
:;;~ a. a. a. a. a. c-
:::J :::J

0
:::J :::J :::J

g
:::J
practical minimum.
0 0 .6 '" ;J;
There may be many more welded joints in a pressure vessel
I~ ~I
~~~l
I~ ~::1
c: ~ ~
~ q;
I II but it is beyond the scope of this text to describe all of them.
However, an idea can be formed about weld categories, operating
- III
.,
!:
.,
E
0;
~ .,
'"0E
0toQ;
.0 '"
u 3=E0;
conditions and typical steels used in fabrication of one type of
pressure vessels uiz., the boilers used in power plants; as given in
.!~illu;.0Ue~0
'":::J ., c: ;:;!.
to
.0
0;
>-
.2
00
Q;
:::J 0;
~ c:
~
::E
:::J
0;
E
00 c:
:::J
;:;:.
:::J

di
to
...J
0;
3= .0
VI
::E
0;
(;;
VI
:::J SU
l:?
l:?
Q;
l:?
g'" table 16.7.
.c:C)
'"
00 V>
E to"",
-
.g
- ~ V>
Q)
0;
;:f"' '"
16.1.3.Fabrication of Clad Pressure Vessels
Q
••• ~ Q.
e
C!$
~
., :::J
c:
., '" -
~ .~
'0
'"
'"
l:?
to
~
., To reduce corrosive attack there is a consistent demand for
g '" (ij
E:E
*~ <>
ec-
0;
a; Q; vessels with an internal surface of a corrosion-resistant material
-., .,0
CD
e-
E=i' "6
,e> E :::~ E '" E~ E
"'0 ~ 0; .5~
J:.O
o; -
to
3=
CD 0; which is fr~quently one of the austenitic stainless steels. This
:::J '(3 :::J:=- :::J

§ c:
.,
::E
'"
.!2e
u..,
..
~
UJ
c:=
~.B
:::J ..
'"
a.
00
~ ~
::E.,
c:
.,
::E
protective layer is usually formed by anyone of the three methods
rA
= uiz., roll ~ladding, explosive bonding, and surfacing either by
Q ~ submerged-arc or by a gas shielded welding process.
:= Qi 'c (/)
S::E::E :ll::E ::E
$ Q) ::E

::a E U);§::''C c;;~ u

"C
Such vessels are fabricated only when the wall thickness

"C
= 'O ••..u"C ••.. U ~ ~
:!2 ~caen$.
Q ~ ~u;:;!.
Qi 0 01":=U0
"" ~ '" ::E ~ "'~
o,go I '"0 exceeds 50 mm; and are mainly employed in Nuclear Energy Field.
t;.)
CI
o o .2 0' Another variant of pressure vessel fabrication is the
.9 .!Q c:
~~ Multilayer Vessels. In one of the Jabrication techniques a thin,
eoto ~ 0
c:
~ -~
::E

U e ~ '2
::E
o o 3=~
o ~ o
.. ~ u '"
0 U; -u al-e: ::Eu;
CD

plate of about 6 mm thickness is wrapped around a cylindrical,

~~s
::E .!!! ::E .!!! ~ .!!!

u ~
'" ~
"CQ)C~

~~* ~"OQ)
-g=.. U
~ :::J

o ;:;!.~~ ., to
core and butt welded; the tension applied to tha thin plate is
8~ "'u .. U (/)
..
I» ~
>g '" U.. c:to
~ 1£0; u; ,2. to
'" ,- E C') ..

c: C')
~ '0 ~ --~~ .g -~~g ~~ ~ Q503 --~~g'~.g sufficient to make intimate contact with the inner cylinder. The
~ CD ...- 'c
.. CD '"

a- ~ -8-'"::!Q)"CJQ _u_ ~(j)~VS1:'~ ~;s.!!l ~

~~I~
'(;)C;;

~ ~ u U '" ~ CD
'0>- process can then be repeated until the required thickness of the
...•
••
CD

$ ~ UNi6
Ou; ~ 0 (/) ;S!U"'(/) ...,,,,.,
:::J '" u~~; =.2
]!
U)~'tr..~cn
,,_ ;S
...,Nca §! NCtli u~~u
0 ~ . ~NU;;S! vessel is achieved .
o
CI
I:::J S ,en.. ::Eo; ::> _ :::J

Cl:l

0'0
'" 2., .c:
E '"en ctiCi> '" '0
.,'"E .,
:::J
'0
'":x:
VI .0
.c:
6E
u;
.,..
:::J
3=
'0
c: The advantages of multilayer pressure vessels include the'
~a:.c:
'" ~ 0 '"
c:
'"N
'0 '"CD(/)
'0'" EE a::x:-0 :;; (ij
CD

U(;jCD N.c:~ -0
~l "'- minimisation of risk of brittle failure and the ease of stocking and
CD '" '"
Q, :::J
E .9
00
CD
:;; co to- ••••
~ ~ ~ .~ ~
-to
CD

.~~ .~l
~ 8.
2
'0 VI ~~ ..... ~ ~ .~ tV ~ ., '"
"'.c: c. ca en C. handling the thin plates; however it also involves some difficulties
~ '" '" E ~~~ Jl~ 0
'is E E
., to-E(;jo
c••.•.CDCo..!:
CD C ~~~ E.l!! E
like the problem of making circumferential welds and nozzle
~ <3 '" 0 to c: CD ~ '" a. '0e: .0 m ~.~ m
~
en =' =:l •..
LLw Q) <.J
Q) &al~«.8
Eoo.c:;= 00.,_ (j)£~U5
:::J :::J
attachments. Moreover, the maximum service temperature
t.C recommended for use with such pressure vessels is about 450°C.
-
....• '"
c:'0-~
'0
4i CD _0==
a2 ~0tou
.0
.c:
E
'0;
'"
to
(ij
2c:'is'"~c:0-Q)'" c: 2to
lD 3= Thus, the number of multilayer pressure vessels in use is much
IE:§.
"'a: ==
""'"
,t;J
~ lD'" 2
.0
'0
:::J
Qi
3= .!!!'O '6>
E
.!!!'O
'" ~~
IS~ '" less than the conventional type.
== '0
~
o '0.0 :::J
Q;
~ c: _ 3=
16.1.4.Weldability Aspects of Pressure Vessel Steels
g> ., to .c:
-a;E u
c:
~
li., Weldability is a determining factor in the selection of pressure
c: '" '"
:0<0 '0
:::J-
"" E
Qi
vessel steels. In this context, in general weldability means the
O:::J
3=
(;
c: U ==
o
...Ju
.= lD
:::J ar
::E
ability to fusion weld the material with the minimum' of special
'" procedures, together with minimum susceptibility to weld failures
'O:E in service.
Q)
SJ81!oq U! SPI8M UO!108S U!4.l 'Z
~ ~.~.
$3 ii:'(;)CJ)
_ '" c: Sensitivity of steels to hot and cold cracking is the most
-5 "en
I SJ81!oq U! SPI8M UO!108S lP!4.l -~
(¥')
~ important consideration in relation to welding.
1112 Welding Engineering & Technology Heavy Welded Fabrications 1113
1'1

In plain carbon steels, sulphur is the element which most com-
monly promotes hot cracking and thus it is imperative to maintain
a high Mn/S ratio to avoid this defect. Basic coated electrodes being
more resistant to sulphur cracking are therefore selected to avoid
hot cracking in such steels. Hot cracking has also been observed in
C-Mo and Cr-Mo weld metal and is associated with too low Mn/C
ratio. Also, it has been suggested that the ratio of Mn/Si should be
greater than. 2 for optimum resistance to hot cracking.
Hot cracking affects the filler metal of fusion welds more
frequently than the parent metal for example low carbon fully
austenitic Cr-Ni steel is intrinsically subject to hot cracking. To
reduce the risk of hot cracking, fully austenitic stainless steel of
the 18/12 (CrIND type .is usually welded with a filler whose
composition is balanced to give between 3 t.o 10% ferrite in the
deposited metal. 25 Cr 20 Ni weld deposits with carbon about 0.2%
often contain micro-fissures due to hot cracking but may be used
successfully in applications not subject to high stress, high
temperature or thermal shock. For the more severe duties 25/20
(CrINi) welds which are often used for welding dissimilar metals,
are best replaced by Ni-base fillers. Medium Carbon (0.4% C) 25/20
(CrINi) filler is not, however, subject to hot cracking and is used
successfully for welding centrifugally cast steel of the same
composition for high temperature furnace tubes.
16.1.4.1. Cold Cracking
Cold cracking associated with the presence of martensitic
structure, hydrogen and residual and/or applied stresses has been
responsible for a number of catastrophic failures of pressure
vessels during hydrostatic testing. The cracks form in a hardened
Although presence of hydrogen is usually associated with cold
cracking but low hydrogen in a ferritic steel deposit does not
guarantee freedom from cold cracking though it does minimise
such a risk. Austenitic steel welds in ferritic plate may contain
substantial amounts of hydrogen without any significant risk of
cracking. The diffusivity of hydrogen in austenitic steel is very
low, so that the gas remains trapped in the weld deposit. Also,
austenitic welds have a low elastic limit and tend to yield
preferentially to the ferritic base metal, whereas with ferritic weld
deposits the yield point is invariably higher than that of wrought
plate of the same composition. Austenitic filler is, therefore,
frequently used for welding of 5% Ni-steel, 9% Ni- steel and the
13% or 17% Cr-steel, as well as for welds between ferritic and
austenitic steels.
While selecting steels for a pressure vessel the first rule is to
avoid carbon and C-Mosteels with carbon content over 0.2%. With
ferritic alloy steels the risk of cold cracking increases in general,
with alloy content and with tensile strength. The susceptibility of
a steel to cold cracking depends upon whether or not it forms
martensite; that is why modern high tensile steels are formulated
to avoid the formation of martensite in HAZ and generate a more
favourable structure like lower bainite. Fig. 16.17 illustrates the
0.6
...0
/ /
Poor toug hness

Twinned
morlens'le
L+c
...•..

embrittled region in the weld or HAZ and are, therefore, a

potential source for initiation of brittle fracture. w
II
Because cold cracking requires the presence of martensite in !
.•..
c:
the mf~tallurgical structure it does not usually affect low carbon I,I 2o
steels of thin sections. Thick sections of carbon steel (say 25 mm I >
for 0.2-0.3% C) and higher carbon contents (0.3% and over) may, '3
cr
0.4
however, be subjected to this defect. For any given heat input rate, <lJ

the risk ofmartensitic formation increases with thickness because C

! o
-c
!
it leads to higher cooling rates. For any given thickness, the cooling
rate increases as the size of weld pool decreases thus other things
o'-
w
Hardness 350HV

being equal small weld runs are more likely to promote cold crack- I 0.3
Pearlite
ing than large runs. For ferrtic alloy steels the same rules apply
except that the limiting carbon content is lower. For alloys higher 1 10 100
than 0.5%Cr 0.5%Mo, the steel is sensitive at normally specified I Cooling rate at 300·C, deg. C/sec
carbon content of about 0.15%.
r
F'ig. 16.17. Schematic representation of the effect of welding conditions

I on HAZ cracking and HAZ toughness in C-Mn steels.

~:::G,~~~,~~[l,",,,,"~-"'-:;;",,,;j;i~1lI.ili-",,,,"-",~'';!,i1ll·~l1i''ll1l''~.~Ii-;:l.,=lil~=.'.-·Ill"lIiIiliillim''·~lIl''"'Il·~lO:::. :lllI """ ••• __ ~------------------------,---------------------------------------------------

1114 Welding Engineering & Technology Heavy Welded Fabrications 1115

.'
relationship between microstructure, notch toughness in HAZ and or more, and in the austenitic steel pipework where the wall
crack susceptibility for C-Mn steels, relating these characteristics thickness is 20 mm or greater. A common initiation point is the
also to cooling rate and carbon equivalent (C.E.) of steel where toe of the weld, externally or internally, and the crack may occur
C.E. for C-Mn steels is given by, either in the nozzle or the plate.
eE
....(16 1)• OJ.
=
6 + Cr + Mo
C + Mn 5 + V + Ni 15
+ Cu Steels susceptible to stress relaxation cracking include ferritic
steels containing boron or vanadium, and austenitic Cr- Ni steels
Control of C.E. is desirable for site fabrication of large containing niobium or titanium. However, vanadium bearing
pressure vessels because of the proper control of preheat ferritic steels with nominally 0.25 V have been shown to be
temperature in such conditions. particularly sensitive because during stress relaxation an
intragranular vanadium carbide precipitation forms which
Hydrogen cracks are usually quite short but may even then hardens the grains and concentrates strain in the grain boundary
be hazardous if the metal around the tip of the crack is embrittled. regions. A similar effect occurs with Nb-stabilized austenitic steel
With such an unfavourable combination it is possible for short AISI Type 347, whereas with the AISI 316 Type precipitation is
cracks to initiate a brittle failure in plate which would normally at the grain boundaries, where it is relatively harmless. A 316
be resistant to fracture initiation. In carbon and C-Mn steels steel containing niobium behaves like type 347 and where
embrittlement can occur due to strain ageing, the strain due to resistance to stress relaxation cracking is required such a steel
weld contraction being intensified at the tip of the crack, should not contain any significant amount of Nb or Ti.
particularly in multirun welds where later passes can strain Compositions of Types 347 and 316 steels are as given in Table
pre-existing cracks in the earlier passes. Since the degree of 16.8.
embrittlement from this cause increases with the free nitrogen
content of steel, it is best to select aluminum treated C-Mn steel Table 16.8. Compositions of AlSI 316 and 347 Steels
where optimum resistance to brittle fracture is the main AISI C Mn Si Cr Ni Other elements
consideration. On the other hand, strain age embrittlement can type % % % % %
be removed or significantly reduced by PWHT, thereby reducing 347 0.08 max 2.0 1.0 17-19 9-13 Cb + Ta = (10 x C Min.)
the risk of brittle failure in Si-killed and semi-killed carbon and 316 0.08 max 2.0 1.0 16-18 10-14 Mo = 2-3%
C-Mn steels. That is why standards on pressure vessels usually
permit lower operating temperature with a stress relieved When fillet welding an attachment to the surface of a plate,
pressure vessel than with the corresponding vessel used in the the contraction due to welding sets up strains which have a
as-welded condition. component at right angles to the surface. If these strains are
Ferritic alloy steels may be subjected to embrittlement at the sufficiently high and through-thickness ductility low (due to
tip of the welding cracks if they receive a PWHT at too Iowa non-metallic inclusions), then the plate may fail by lamellar
temperature. Thus steels having higher tempering temperatures tearing. Cracks start at the toe of a fillet and progress in step-like
are clearly preferable because they permit greater tolerance in fashion underneath the weld.
PWHT temperatures. In pressure vessel technology, the hazard of lamellar tearing
Another form of cracking is that due to stress relaxation arises in welding attachments and set-on nozzles to thick-walled
shells. Since the use of set-through nozzles in heavy-wall vessels
promoted by : involves a risk of cracking during PWHT, set-on nozzles are
(i) unfavourable joint design, generally preferred for this class of vessels. At the same time, to
(ii) stress concentration, and reduce the risk of lamellar tearing, a lower limit may be specified
(iii) sensitivity to strain age embrittlement at elevated for the through thickness ductility and the steel may be required
tem pera tures. to be vacuum degassed. Such a vacuum treated steel is much
cleaner than the normal commercial product, and its resistance to
As regarding design, the problem has mainly been lamellar tearing and other deficiencies associated with
encountered with set-through nozzles in high tensile ferritic steel non-metallic inclusions is much better.
where the plate and nozzle thicknesses are in the range of 60 mm
I
 1117
Heavy Welded Fabrications
1116 Welding Engineering & Technology
'tj ~~
S .:- .•..
::!
:::!.- ) o o o.,., 00 00
The corrosion resistance of pressure vessel steels may be
~
~~ •....•
.§ ~ ~ I ~
.S"5 ~
.,.,0 ..... -
.,.,
- "'0
.... ""
.,.,0
•.•.• <:'I ::!

adversely affected by welding, The most important problems of C <:l

'-' •.. ::::s l5. '"

this type are weld decay ofunstabilised austenitic Cr-Ni steels and ~ S
'"
~"_.:+--" ._~
.••tress-corrosion cracking of austenitic steels in general. Weld decay ~~ ., ;;; ;;;
is due to intergranular carbide precipitation in the HAZ of fusion 'tj :3
~
~
•....•
.g
~~~
g)
E::
00
<:'I <:'I
~
00
§
00
0-
<:'I <:'I
•.•.• VI
<:'1<:'1
•.•.• VI
0-
::!
welds of 18/8 (CrINDstainless steel. These sensitised regions may S VI 1\ CO') <':l

be attacked in acidic environments. Weld decay can be avoided by ~ ~]~ .•..

VI VI
VI <:'I
....•
1\
VI c:'I
....•
1\
limiting the carbon content to a maximum of 0.03% or by adding -...
small amounts of titanium or niobium to the steel. The former is ~
~
known as extra low carbon steel and the latter a stabilised steel. s;;p~
Both these types are more costly than straight 18/8 (Cr-Ni) types,
and a compromise which provides adequate resistance to many .1.
-.51
~
Q)
'tj
.::t
"Q='
.b(::!
:::I '-t:l ::!

.§ ~E
E: ..-::'"
C,)
o IQ 81 ~~
00
.•....•
00
<:'10....•
00
o at)
....•.....
00
o at)
....• ....
00
at) 0
•.•.• <:'I

~
environments is to specify AISI Type .:W4 (C = 0.08% max., Mn = !
s..

cS.cl
Q)
E: _.
c
<>
.E
'"

S
;,;;:
<; ~ Q.,
Q., E::
~ ..--+- --
2%, Si = 1%,Cr = 18-20%,Ni = 8-12%).National standards of some .•.. E:

~.s..... 'tj~~~ .,
-
•....•
European countries however, insist on the use of a stabilised or .9 s...!
••. :!'
·s ~ 00
extra low grade for pressure vessels. Also, Type 304 is not suitable •.. E: ss <:'I <:'I
...•.....
<:'I c-.l
....•....•
<:'I <:'I
....•
<:'I <:'I
....• ....•
c<I <:'I
..........
1U,s..c ::2 ~~@ roCO')I VI II VI 1\ VI 1\ VI 1\ VI 1\
for stainless steel clad vessels that must be given PWHT, since ~f~
1:1 rIJ
~~ ..• VI II I
t.he heat treatment will caUS6the precipitation of carbides. Q) rIJ· ••
I

The only means of avoiding stress corrosion cracking at welds Q~

rIJ

Q)l:lIl
~Q).cl
.•••

=
1
in austenitic stainless steel is either to avoid contact with aqueous ~ =.~ I
I
phases containing chlorides or sodium hydroxide, or to stress ~fQ
CljcG::l
relieve at 900o-1000°C. Stress relieving is practicable for piping S Co,s
but is difficult for vessels. Therefore avoidance of contact is the :....S rIJ

Q)

jf~
~ rIJ

best precaution. Contamination by salt water during

transportation and storage must be avoided and fresh water must 6'c =
~ ..•
Q)
II
be used for hydrostatic testing. Critical vessels are hydrotested
with steam condensate or de-ionised water of controlled chloride
=e I
<6 ]'" !
content. Where there is a positive risk of hydroxide or chloride
ions being present at the surface particularly at elevated
-.-4
Q)

..c tl ~ I
<l.l
.0
<l.l
IS
='
"s'1l
]e ]~
c::

temperature, some less susceptible material like monel (66% Ni + ~

'£::

~ ~
33% Cu + 1% Mn) though costly must be used. ~ 1il '"(.)e '" ::l
E CIl ._
0.>"0
::l
16.1.5.Preheat and postweld Heat Treatment of o
c::
.0 c:: 1il Cd
•.. Ql bD§
Pressure Vessels Cd
(.) :g c:: >
:>
e e::l
>.

l
~ "0
Cd
o ,-..
,..,.
The preheat temperature depends upon the composition and o
su
c::
S o ::s
1;>~
s
Cd

="0
thickness of metal being welded as well as upon the welding
I
oc::
..0
l::l
o
..0
Cd..o
s
::s •..
:::s

j~~8 u
•.. •..
process and arc energy being used. As a general rule no welding ~ :'J.
Cd
C,) U ....• ....•
should be carried out when the temperature of the parent metal,
within 150 mm of the joint, is less than 5°C. Austenitic steels, "'<::l
<::l

t-
-g:g
how-ever, do not require preheat for welding. General recommen-
dations regarding preheat for some of the popularly used steels
are as given in Table 16.9.
c3~
,:='..-l I :a
::E~ ""
I~
i
::s

:1
~ I
1118 Welding Engineering & Technology Heavy Welded Fabrications 1119
"0 '"
•..
Q)
.<;: ~
"0
o0.
OJ
Q) 16.1.5.1.Preheating of Tube-Plates

~i
Q) g...
"0
•.
·5 5
<:u
~-<:i5..'
"'-'0 0C<I.----~-
~ "l:t ~ ~ I '<to "0
'- 0. The preheating of tube-plate