Eagar 021

Ph.D.
Thesis
On
SOME STUDIES ON DESIGN AND DEVELOPMENT OF

AGGLOMERATED FLUXES IN SUBMERGED ARC WELDING
For the Award of the Degree

of
DOCTOR OF PHILOSOPHY
IN
MECHANICAL ENGINEERING
BY
AJAY KUMAR
2K08/PhD/NITK/l 186-M
Under the Supervision of
Dr. Sachin Maheshwari, Professor Dr. Hari Singh, Professor

Dept. of Manufacturing Processes Dept. of Mech. Engg.
& Automation Engg. NIT, Kurukshetra
NSIT, Delhi
Submitted to
DEPARTMENT OF MECHANICAL ENGINEERING
NATIONAL INSTITUTE OF TECHNOLOGY
(Institution of National Importance)
KURUKSHETRA-136119, INDIA
April, 2013
9-
^76 ./ 3
CANDIDATE'S DECLARATION
I, hereby, certify that the work which is being presented in thesis entitled "SOME
STUDIES ON DESIGN AND DEVELOPMENT OF AGGLOMERATED FLUXES
IN SUBMERGED ARC WELDING" in partial fulfilment of the requirements for the
award of the degree of "Doctor of Philosophy" in MECHANICAL ENGINEERING",
submitted to the Department of Mechanical Engineering, National Institute of
Technology, Kurukshetra, is an authentic record of my own work carried out from
May, 2008 to till date under the supervision of Dr. Hari Singh, Professor, Department of
Mechanical Engineering NIT, Kurukshetra and Dr. Sachin Maheshwari, Professor,
Department of Manufacturing Processes and Automation Engineering, NSIT Delhi, India.
The matter embodied in this thesis has not been submitted by me for the award of any
other degree or diploma.
Date: o^-oh '^°^^

£^
(Mr. AjayKumar)
CERTIFICATE
This is to certify that the thesis entitled "SOME STUDIES ON DESIGN AND
DEVELOPMENT OF AGGLOMERATED FLUXES IN SUBMERGED ARC
WELDING" which is being submitted by Mr. Ajay Kumar (Registration number
2K08-NITK-PhD-1186-M) for the award of the degree of Doctor of Philosophy is a
record of bona fide research work carried out by him. Mr. Kumar has worked under our
guidance and supervision and fulfilled the requirements for the submission of this thesis,
which to our knowledge has reached the requisite standard. The results contained herein
have not been submitted, in part or in full, to any other University or Institute for the
award of any degree.
Dr. Sachin Maheshwari, Professor Dr. Hari Singh, Professor

Dept. of Manufacturing Processes Dept. of Mech. Engg.
& Automation Engg. NIT, Kurukshetra
NSIT, Delhi
ACKNOWLEDGEMENTS
The present work which stands in the form a thesis has been one of the greatest endeavors
of my academic pursuit and is a culmination of scholarly and enlightening advices and
blessings of my respected teachers to whom I owe this endeavor of mine.
In the completion of this thesis, 1 received substantial and most willing cooperation from
people. Therefore, I wish to record my deep sense of gratitude to every one of them who
made this research both pleasant and possible.
I express my deep sense of gratitude to Dr. Hari Singh, Professor, Department of
Mechanical Engineering, National Institute of technology, Kurukshetra, first and
foremost, for his kind consent to accept me his disciple and secondly for his invaluable
erudite explanations which refined my work.
My thanks are especially due to my co-supervisor, Dr. Sachin Maheshwari, Professor,
Department of Manufacturing Processes and Automation Engineering, Netaji Subash
Institute of technology, Delhi, who has always been a constant source of encouragement
to me.
I also express my gratitude to Dr. Sudhir Saxena, Professor and Head, Department of
Mechanical Engineering, National Institute of Technology, Kurukshetra for providing me
all possible help. I am also thankful to all the faculty members of the Department for their
help, inspiration and encouragement which went a long way in the successful completion
of my thesis.
I would fail in duties unless I acknowledge my profound sense of gratitude to Dr. O. P.
Agarwal, Managing Director and the entire management of NIET, Greater Noida for their
blessings. 1 am also thankful to all my colleagues, especially Dr K P Singh, Dept. of
Chemistry and Dr Vipin Kumar, Dept. of English, Noida Institute of Engineering &
Technology, Gr. Noida for their support and encouragement.
It goes without saying that I would have not been able to maintain all things properly
without the support and cooperation of the entire staff of Welding Engineering
Laboratory of NSIT, Delhi and of B.Pharma, Chemical Engineering and Mechanical
Engineering of Noida Institute of Engineering & Technology, Gr. Noida.
in
I also express my sincerest thankfulness to Dr. Sudhir Kumar, Professor and Head,
Department of Mechanical Engineering, NIET, Greater Noida for his valuable and
practical guidance at critical moments.
I would also like to thank Mr. Mahipal Singh, Ex. Sr. Lab Technician and Mr. Manoj
Kumar, Asstt. Registrar, NIET, Greater Noida for their valuable support.
My long cherished dream would have remained unaccomplished without the blessings of
my parents who always encouraged me and never allowed my morale to sink. My wife,
son and daughter showered endless love and unconditional support in accomplishing this
research.
Finally, I owe a debt of gratitude to the Almighty!
Date: OV ^^- °^'^^3

(L--
(AJAY KUMAR)
Place: Kurukshetra
IV
ABSTRACT
The present investigation is an attempt to develop a methodology for design,

formulation, development and modeling of agglomerated fluxes during submerged arc
welding. The agglomerated fluxes that are used for welding of low carbon steel plates
have CaO, Si02 and AI2O3 as main constituents and MnO, CaF2, NiO, MgO and Fe-Cr as
minor alloying constituents. These fluxes are used for this study. The amounts of main
constituents are fixed and the ranges of alloying constituents are designed on the basis of
ternary and binary phase diagrams. Thirty two fluxes are formulated for five variables as
five constituents of agglomerated fluxes with the help of response surface methodology
technique. These fluxes are prepared using agglomeration method. The alloying
constituents of fluxes play an important role during welding because the variation in their
composition affects physical properties of fluxes which in turn change the mechanical
properties of weld joints.
The physical characteristics of developed agglomerated fluxes namely bulk density, grain
size,flowabilityand viscosity are measured with the help of different setups. The results
of these physical properties offluxesare correlated with the standard values or data.
The developed agglomerated fluxes are analyzed with the help of X-ray diffraction
(XRD) and differential thermal analysis (DTA). The XRD analysis is helpful to analyze
the different crystalline phases and oxides present in developed agglomerated fluxes. The
DTA is used to analyze endothermic temperatures of these developed fluxes. Thus, fluxes
so developed are characterized using XRD and DTA.
The weld joints are produced on low carbon steel plates by using developed agglomerated
fluxes. The characteristics of these weld joints, like tensile strength, yield strength,
elongation, impact strength, hardness and weld microstructures are determined with the
help of different testing equipment. The mathematical modeling equations of tensile
strength, impact strength, elongation and hardness are developed through the response
surface methodology to correlate the actual experimental values with the predicted values.
The beads on plate on low carbon steel plates are laid to obtain the bead morphology
parameters like weld bead penetration, weld bead width, and weld bead reinforcement.
The mathematical modeling of weld bead penetration, weld bead width, and weld bead
reinforcement is done with the help of response surface methodology technique.
The optimum values of the mechanical properties and bead morphology dependent
parameters are determined with the help of response surface methodology and desirability
functions.
The pads of five high beads on low carbon steel plates are laid to study the element
transfer behavior during welding. Transfer of C, Mn, Si, Cr, Ni, S and P is studied in the
present work.
The results obtained clearly stipulate that it is possible to model the behavior of fluxes
with the help of response surface methodology. The fluxes have been developed and
characterized for welding of low carbon steel plates.
The thesis comprises six chapters. The first chapter introduces submerged arc welding
process and fluxes used during this process. The second chapter contains review of the
work done by previous researchers and points out some of the drawbacks and
discrepancies in their analyses. This chapter also details the drawbacks and discrepancies
identified by researchers under the title "gaps and opportunities". The third chapter
describes about design, formulation and development of fluxes. The fourth chapter
introduces the basic theories of measurement of characteristics of fluxes and describes the
procedure of joining welds by using developed fluxes. It also discusses different weld
joint testing methods and describes the bead geometry and shape relationship by using
different developed agglomerated fluxes. The fifth chapter discusses the results in detail
and finally this report is concluded in the sixth chapter with the recommendation for
future work.
VI
TABLE OF CONTENTS
Description
Candidate's Declaration I
Certificate a
Acknowledgements Hi
Abstract V
Table of Contents vii

List of Figures xiii
List of Tables xxii
Nomenclature xxvi
List of Acronyms xxvii
TER 1 INTRODUCTION 1- 19
1.1 Submerged Arc Welding (SAW) 1
1.1.1 Introduction 1
1.1.2 Principle of operation 2
1.1.3 Advantages 5
1.1.4 Limitations 6
1.1.5 Process applications 6
1.1.5.1 Types of metals 6
1.1.5.2 Joint configurations 6
1.1.5.3 Industrial uses 7
1.1.6 Equipment and supplies 7
1.1.7 Power sources 8
1.1.7.1 Constant voltage DC type 8
1.1.7.2 Constant current DC type 9
1.1.7.3 Alternating current type 10
1.1.8 Process parameters 10
vii
1.1.8.1 Welding current 10
1.1.8.2 Electrode polarity 10
1.1.8.3 Current density 11
1.1.8.4 Welding voltage 11

1.1.8.5 Travel speed 11
1.1.8.6 Electrical stickout 12

1.1.8.7 Flux layer depth 12
1.1.8.8 Type of flux and particle distribution 12
1.2 Submerged Arc Welding Fluxes 12
1.2.1 Classification of submerged arc welding fluxes 14

1.2.1.1 Fused fluxes 14
1.2.1.2 Bonded fluxes 15
1.2.1.3 Agglomerated fluxes 15

1.2.2 Classification relative to basicity index 16
1.2.3 Comparative consumption of fiised and agglomerated fluxes 18
1.2.4 Particle size of submerged arc welding flux 19
CHAPTER 2 REVIEW OF EXISTING WORK 20-36
2.1 Introduction and Previous Work 20
2.2 Present Work 35
2.3 Identified Gaps in the Literature 35
2.4 Impact of Proposed Research in Academia/Industry 36
CHAPTER 3 DESIGN, FORMULATION AND DEVELOPMENT OF FLUXES 37- 46
3.1 Design of Fluxes 37
3.1.1 Selection offluxconstituents 37
3.1.2 Flux composition design 37
3.2 Response Surface Methodology 39
3.2.1 Introduction 39
Vlll
3.2.2 Methodology 40
3.2.3 Central composite second order rotatable design 40
3.3 Development of Fluxes 43
3.3.1 Mixing of flux constituents 43
3.3.2 Preparation and baking of fluxes 44

CHAPTER 4 CHARACTERISTICS OF FLUXES AND WELD JOINTS
(EXPERIMENTAL PROCEDURE) 47-66

4.1 Characteristics ofFluxes and their Measurement 47
4.1.1 Flowability 47
4.1.2 Grain size 48
4.1.3 Bulk density 49

4.1.4 Viscosity 50
4.1.5 X-Ray diffraction (XRD) of developed agglomerated fluxes 52
4.1.6 Differential thermal analysis (DTA) of developed agglomerated

Fluxes 53
4.2 Development of Weld Joints 54

4.2.1 Material and equipment 55
4.2.2 Welding procedure 56
4.3 Characteristics of Weld Joints 58
4.3.1 Tensile strength 58
4.3.2 Impact strength 59
4.3.3 Hardness test 61
4.3.4 Weld joint microstructure 62
4.4 Bead Morphology and Shape Relationship 64
4.5 Element Transfer Study 65
CHAPTERS RESULTS AND DISCUSSION 67-1
5.1 Introduction 67
IX
5.2 Physical Properties of Fluxes 67
5.2.1 Flowability 67
5.2.2 Grain size 69
5.2.3 Bulk density 69
5.2.4 Viscosity 71
5.2.5 X-Ray diffraction of developed agglomerated fluxes 72
5.2.6 Differential thermal analysis ofdeveloped agglomerated fluxes 77
5.3 Characteristics of Weld Joints 78

5.3.1 Modeling of we Id joint characteristics 80
5.3.1.1 Modeling of tensile strength 80
5.3.1.1.1 Effect of flux constituents on tensile strength 82
5.3.1.2 Modeling of percentage elongation 85

5.3.1.2.1 Effect of flux constituents on percentage
elongation 88
5.3.1.3 Modeling of impact strength 90
5.3.1.3.1 Effect of flux constituents on impact strength 94
5.3.1.4 Modeling of Vickers hardness 96

5.3.1.4.1 Effect of flux constituents on hardness 99
5.3.2 Optimization of mechanical properties using desirability function 102
5.3.2.1 Desirability function 102
5.3.2.2 Single response optimization using desirability function 103
5.3.2.2.1 Optimal solution using desirability function 105
5.3.2.2.2 Multiple response optimization using

desirability fiinction 115
5.3.3 Weld joint microstructure 130
5.4 Bead Morphology 130
5.4.1 Modeling of weld bead penetration 132
5.4.1.1 Response surface modeling of penetration 132
5.4.1.2 Effect of flux constituents on penetration 135
5.4.2 Modeling of weld bead reinforcement 137
5.4.2.1 Effect of flux constituents on reinforcement 140
5.4.3 Modeling of weld bead width 142
5.4.3.1 Effect offluxconstituents on bead width 145
5.4.4 Optimization of bead morphology using desirability function 148
5.4.4.1 Single response optimization of bead morphology using

desirability function 148
5.4.4.2 Optimal solutions using desirability function 149
5.4.4.3 Multiple response optimization of bead morphology

using desirability function 158
5.4.4.3.1 Model: Penetration, reinforcement and

bead width 158
5.5 Element Transfer Study 162
5.5.1 Carbon (C) transfer 163
5.5.2 Manganese (Mn) transfer 163
5.5.3 Silicon (Si) transfer 163
5.5.4 Chromium (Cr) transfer 164
5.5.5 Nickel (Ni) transfer 164
5.5.6 Sulphur (S) transfer 164
5.5.7 Phosphorus (P) transfer 165
5.5.8 Effect of element transfer on mechanical properties 165
CHAPTER 6 CONCLUSION AND SCOPE OF FURTHER RESEARCH 167-172
6.1 Conclusion 167
6.1.1 Bulk density, grain size,flowabilityand viscosity 167
6.1.2 XRDandDTA 168
XI
6.1.3 Modeling ofweldjoints and bead morphology 168
6.1.3.1 Tensile strength 168
6.1.3.2 Hardness 169
6.1.3.3 Elongation 169

6.1.3.4 Impact strength 169
6.1.3.5 Optimal parameters (Mechanical Properties) using
single response optimizations 170
6.1.3.6 Optimal parameters (Mechanical Properties) using
multiple response optimizations 170
6.1.4 Weld joint microstructures 171
6.1.5 Bead morphology 171
6.1.5.1 Optimal parameters (Bead Morphology) using single and
multiple response optimizations
6.1.6 Element transfer

6.2 SCOPE OF FURTHER RESEARCH
REFERENCES
LIST OF PUBLICATIONS (JOURNALS)
APPENDIX A Tables of Sieving Results for Agglomerated Fluxes
APPENDIX B Figures of Grain Size Histogram for Agglomerated Fluxes
APPENDIX C Figures of XRD Graph for Agglomerated Fluxes
APPENDIX D Figures of DTA Graph for Agglomerated Fluxes
APPENDIX E Tables of ANOVA of Mechanical Properties and Bead

Morphology for Agglomerated Fluxes
APPENDIX F: Figures of Weld Joint Microstructure for Agglomerated Fluxes
APPENDIX G: Figures of Weld Bead Profile for Agglomerated Fluxes
xu
LIST OF FIGURES
Figure No. Title Page No.
Figure 1.1 Equipment for Submerged Arc Welding 3

Figure 1.2 Typical Weld Pool Dynamics of A Submerged Arc Weld 4
Figure 1.3 Schematic Representation of The Equipment for Arc Welding
With Twin-Wire Electrode 5
Figure 1.4 Effect ff Welding Current on Deposition Rate 8
Figure 1.5 Volt-Ampere Output Relationship for a Constant Voltage Power
Source 9
Figure 1.6 Typical Volt-Ampere Characteristics of A "Drooping" Power Source
With Adjustable Open Circuit Voltage 9
Figure 1.7 Comparative Consumption of Fused And Agglomerated Fluxes 19
Figure 3.1 Phase Diagram of CaO-Si02-Al203 System 38
Figure 3.2 Central Composite Rotatable Designs In 3X-Variables 42
Figure 3.3 Ball Mill Setup 44
Figure 3.4 Mixture taken out from Ball Mill (Pre-fmal stage) 46
Figure 3.5 Prepared Flux (Final stage) 46
Figure 4.1 Setup of Flowability 48
Figure 4.2 Setup for Grain Size Measurement 49
Figure 4.3 Setup for Bulk Density Measurement 50
Figure 4.4 Setup of Viscosity Measurement 51
Figure 4.5 Setup of X-Ray Diffractometer, (Source: IIT, Roorkee, India) 53
Figure 4.6 Figure 4.6: Setup of TG-DTA (Source: IIT, Roorkee, India) 54
Figure 4.7 Welding Specimen 55
Figure 4.8 Submerged Arc Welding Machine 56
Figure 4.9 Correct way of Bead Deposition 57
Figure 4.10 Welded Low Carbon Steel Plates 57
Figure 4.11 Dimensions of Specimen for Tensile Test 58
Figure 4.12 Orientation of Specimen from the Weld Plate 59
Figure 4.13 Setup for Tensile Strength
Figure 4.14 Dimensions of Charpy Impact Test Specimen gQ
xni
Figure 4.15 Prepared Weld Metal Charpy Impact test Specimen 60
Figure 4.16 Impact Testing Machine 61
Figure 4.17 Hardness Testing Machine 62
Figure 4.18 Set up of Polishing Machine 63
Figure 4.19 Set up of Metallurgical Microscope 64
Figure 4.20 Weld Beads on Low Carbon Steel Plate 65
Figure 4.21 Specimen for Element Transfer Study 66
Figure 5.1 Effect of Bulk Density on Flux Consumption 71
Figure 5.2 Viscosity Trend of Flux (AGF1101) 72
Figure 5.3 Normal Probability Plot of The Residuals for Tensile Strength 82
Figure 5.4 Predicted Vs. Actual Value of Tensile Strength 82
Figure 5.5 Effect of CaFi and MnO on Tensile strength 83
Figure 5.6 Effect ofMgO and MnO on Tensile Strength 83
Figure 5.7 Effect of NiO and MnO on Tensile Strength 84
Figure 5.8 Effect of Fe-Cr and MnO on Tensile Strength 84
Figure 5.9 Normal Probability Plot of The Residuals for Percentage
Elongation 86
Figure 5.10 Predicted Vs. Actual Percentage Elongation Value 87
Figure 5.11 Effect of CaF2 and MnO on Percentage Elongation 88
Figure 5.12 Effect of MgO and CaFj on Percentage Elongation 88
Figure 5.13 Effect of NiO and CaFa on Percentage Elongation 89
Figure 5.14 Effect of Fe-Cr and CaF2 on Percentage Elongation 90
Figure 5.15 Normal Probability Plot of the Residuals for Impact Strength 92
Figure 5.16 Predicted vs. Actual Impact Strength 93
Figure 5.17 Effect of CaFa and MnO on Impact Strength 94
Figure 5.18 Effect of CaF2 and MgO on Impact Strength 94
Figure 5.19 Effect ofNiO and CaF2 on Impact Strength 95
Figure 5.20 Effect of Fe-Cr and CaF2 on Impact Strength 96
Figure 5.21 Normal Probability Plot of the Residuals for Vickers Hardness 98
Figure 5.22 PredictedHardness vs. Actual Value of Vickers Hardness 98
Figure 5.23 Effect of NiO and Fe-Cr on Vickers Hardness 99
Figure 5.24 Effect of CaF2 and MgO on Vickers Hardness 100
Figure 5.25 Effect of MnO and NiO on Vickers Hardness 101
Figure 5.26 Effect of MnO and Fe-Cr on Vickers Hardness 101
xiv
Figure 5.27 Ramp Function Graph of Desirability for Tensile Strength 111
Figure 5.28 Bar Graph of Desirability for Tensile Strength 112
Figure 5.29 Ramp Function Graph ofDesirability for Elongation 112
Figure 5.30 Bar Graph of Desirability for Elongation 113
Figure 5.31 Ramp Function Graph ofDesirability for Impact Strength 113
Figure 5.32 BarGraphofDesirability for Impact Strength 114
Figure 5.33 Ramp Function Graph of Desirability for Hardness 114
Figure 5.34 Bar Graph of Desirability for Hardness 115
Figure 5.35 Ramp Function Graph ofDesirability for Tensile Strength and
Elongation 118
Figure 5.36 Bar Graph of Desirability for Tensile Strength and Elongation 119
Figure 5.37 Ramp Function Graph ofDesirability for Tensile and Impact
Strength 122
Figure 5.3 8 Bar Graph of Desirability for Tensile Strength and Impact
Strength 122
Figure 5.39 Ramp Function Graph ofDesirability for Tensile Strength and
Hardness 125
Figure 5.40 Bar Graph of Desirability for Tensile Strength and Hardness 125
Figure 5.41 RampFunctionGraphofDesirability for Tensile strength,
Elongation, Impact Strength and Hardness 128
Figure 5.42 Bar Graph ofDesirability for Tensile strength. Elongation,
Impact Strength and Hardness 129
Figure 5.43 Normal probability of residuals for penetration 134
Figure 5.44 Predicted penetration vs. actual value penetration(mm) 134
Figure 5.45 Effect of CaF2 and MnO on penetration 135
Figure 5.46 Effect of MgO and CaF2 on penetration 136
Figure 5.47 Effect of NiO and CaF2 on penetration 136
Figure 5.48 Effect of Fe-Cr and CaF2 on Penetration 137
Figure 5.49 Normal Probability of Residuals for Reinforcement (mm) 139
Figure 5.50 Predicted Reinforcement Vs. Actual Value Reinforcement (mm) 139
Figure 5.51 Effect of CaF2 and MnO on Reinforcement 140
Figure 5.52 Effect of MgO and CaF2 on Reinforcement 141
Figure 5.53 Effect of NiO and CaF2 on Reinforcement 141
Figure 5.54 Effect of Fe-Cr and CaF2 on Reinforcement 142
XV
Figure 5.55 Normal Probability of Residuals for Width (mm) 144
Figure 5.56 Predicted Width vs. Actual Value Width (mm) 145
Figure 5.57 Effect of CaF2 and MnO on Bead Width 146
Figure 5.58 Effect of MgO and CaFj on Bead Width 146
Figure 5.59 Effect of NiO and CaF2 on Bead Width 147
Figure 5.60 Effect of Fe-Cr and CaF2 on Bead Width 147
Figure 5.61 Ramp Function Graph ofDesirability for Penetration 155
Figure 5.62 Bar Graph of Desirability for Penetration 155
Figure 5.63 Ramp Function Graph of Desirability for Reinforcement 156
Figure 5.64 Bar Graph of Desirability for Reinforcement 156
Figure 5.65 Ramp Function Graph of Desirability for Width 157
Figure 5.66 Bar Graph of Desirability for Width 157
Figure 5.67 Ramp Function Graph of Desirability for Penetration,
Reinforcement and Width 160
Figure 5.68 Bar Graph ofDesirability for Penetration, Reinforcement
and Width 161
Figure B . I Grain size Histogram for AGF1101 192
Figure B .2 Grain size Histogram for AGF 1102 192
Figure B .3 Grain size Histogram for AGFl 103 193
Figure B .9 Grain size Histogram for AGF1109 195
Figure B . 10 Grain size Histogram for AGF 1110 195
XVI
Figure B . 19 Grain size Histogram for AGF1119 198
Figure B.24 Grain size Histogram for AGF 1124 200
Figure B.28 Grain size Histogram for AGFl 128 201
Figure C.l XRD Graph of AGFl 101 Flux 203
Figure C.2 XRD Graph of AGF 1102 Flux 203
Figure C.3 XRD Graph of AGFl 103 Flux 204
Figure CIO XRD Graph of AGFl 110 Flux 207
Figure C. 11 XRD Graph of AGF 1111 Flux 208
Figure C.12 XRDGraphof AGF1112 Flux 208
Figure C.l 8 XRD Graph of AGFl 118 Flux 211
xvii
Figure C.21 XRDGraphof AGFl 121 Flux 213
Figure D.l DTAGraphof AGFl 101 Flux 219
Figure D.2 DTA Graph of AGFl 102 Flux 219
Figure D.3 DTA Graph of AGF1103 Flux 220
Figure D.4 DTA Graph of AGF 1104 Flux 220
Figure D.IO .DTA Graph of AGFl 110 Flux 223
Figure D. 11 DTA Graph of AGFl 111 Flux 224
Figure D.l 5 DTA Graph of AGFl 115 Flux 226
Figure D. 17 DTA Graph of AGF 1117 Flux 227
XVlll
Figure F.l Weld Jo nt Microstructure of AGF 1101 Flux 240
Figure F.2 Weld Jo nt Microstructure of AGFl 102 Flux 240
Figure F.3 Weld Jo nt Microstructure of AGF 1103 Flux 241
Figure F.4 Weld Jo nt Microstructure of AGF 1104 Flux 241
Figure F. 10 Weld Jo nt Microstructure of AGFl 110 Flux 244
Figure F.l2 Weld Jo nt Microstructure of AGFl 112 Flux 245
xix
Figure F.25 Weld Joint Microstructure of AGFl 125 Flux 252
Figure F.33 Microstructure of Base Metal 256
Figure G.l Weld Bead Profile for AGFl 101 257
Figure G.2 Weld Bead Profile for AGF1102 257
Figure G.3 Weld Bead Profile for AGF 1103 258
Figure G.4 Weld Bead Profile for AGFl 104 258
Figure G.IO Weld Bead Profile for AGFl 110 261
Figure G. 11 Weld Bead ProfileYor AGF 1111 262
Figure G. 12 Weld Bead Profile for AGF 1112 262
Figure G. 13 Weld Bead Profile for AGFl 113 263
Figure G.l 8 Weld Bead Profile for AGFl 118 265
XX
Figure G.27 Weld Bead Profile for AGF1127 270
XXI
LIST OF TABLES
Table No. Description Page No.

Table 1.1 Composition Ranges of Various Fluxes 18
Table 3.1 Flux Constituents and their Ranges 39
Table 3.2 Components of Centre Composite Second Order Rotatable
Design 41
Table 3.3 Flux Constituents Used in the Experiment and their Levels 43
Table 3.4 Chemical Composition of Fluxes and their Percentage Weight 45
Table 4.1 Relationship Between Angle of Repose ( ^) and Flow
Properties of Powders 48
Table 4.2 Features of X-RayDiffractometer 52
Table 4.3 Features of TG-DTA (Thermal Analyzer) 53
Table 4.4 Chemical Composition of Welding Electrode and Base Plate 55
Table 4.5 Welding Unit Specifications 56
Table 4.6 Fixed Welding Parameters 57
Table 5.1 Angle of Repose for Developed Agglomerated Fluxes and their
Flow Properties 68
Table 5.2 Values of Bulk Density, Flux Consumption and Penetration 69
Table 5.3 Quantity of Oxides In % Wt. for Developed Agglomerated Fluxes 74
Table 5.4 Crystalline Phases and Crystal Structure in Developed
Agglomerated Fluxes 75
Table 5.5 DTA Analysis for Endothermic Temperatures 78
Table 5.6 Tensile, Impact and Hardness Properties of the Base Metal 78
Table 5.7 Tensile, Impact and Hardness Properties of the Weld
Metal for CaO - Si02 - AI2O3 Based Flux System 79
Table 5.8 Pooled ANOVA of Tensile Strength 81
Table 5.9 POOLED ANOVA of Percentage Elongation 85
Table 5.10 POOLED ANOVA of Impact Strength 91
Table 5.10 POOLED ANOVA of Vickers Hardness 97
Table 5.12 Range of Constituents and Elongation for Desirability 103
Table 5.13 Range of Constituents and Tensile Strength for Desirability 104
Table 5.14 Range of Constituents and Hardness for Desirability 104
xxii
Table 5.15 Range of Constituents and Impact Strength for Desirability 104
Table 5.16 Optimal Solutions for Desirability (Elongation) 105

Table 5.17 Optimal Solutions for Desirability (Tensile Strength) 106
Table 5.18 Optimal Solutions for Desirability (Hardness) 108
Table 5.19 Optimal Solutions for Desirability (Impact Strength) 109
Table 5.20 Optimal Sets of Process Parameters Using RSM and Desirability
Function 111
Table 5.21 Range of Constituents and Responses for Desirability (Tensile
strength and Elongation) 116
Table 5.22 Set of Optimal Solutions for Tensile Strength and Elongation 116
Strength and Impact Strength) 119
Table 5.24 Set of Optimal Solutions for Tensile Strength and Impact
Strength 120
Strength and Hardness) 123
Table 5.26 Set of Optimal Solutions for Tensile Strength and Hardness 123
Strength, Elongation, Impact Strength and Hardness) 126
Table 5.28 Set of Optimal Solutions for Tensile Strength, Elongation,
Impact Strength and hardness) 127
Table 5.29 Point Prediction at Optimal Setting of Responses (Tensile
Strength Elongation, Impact Strength and Hardness) 130
Table 5.30 Bead Profile Parameters of Different Samples 131
Table 5.31 Pooled ANOVA for Penetration 132
Table 5.32 Pooled ANOVA for Weld Bead Reinforcement 138
Table 5.33 Pooled ANOVA for Weld Bead width 143
Table 5.34 Range of Constituents and Penetration for Desirability 148
Table 5.35 Range of Constituents and Reinforcement for Desirability 148
Table 5.36 Range of Constituents and Width for Desirability 149
Table 5.37 Optimal Solutions for Desirability (Penetration) 149
Table 5.38 Optimal Solutions for Desirability (Reinforcement) 151
Table 5.39 Optimal Solutions for Desirability (Width) 152
Table 5.40 Optimal Sets of Process Parameters Using RSM and
XXlll
Desirability Function 154
Table 5.41 Range of Constituents and Responses for Desirability
(Penetration Reinforcement and Bead width) 15 8
Table 5.42 Set of Optimal Solutions for Penetration, Reinforcement and Width 159
Table 5.43 Point Prediction at Optimal Setting of Responses (Penetration,
Reinforcement, and Bead Width) 161
Table 5.44 Elements (wt %) in weld metal 162
Table A. 1 Sieving Results for AGFl 101 Flux 181
Table A.2 Sieving Results for AGFl 102 Flux 181
Table A. 14 Sieving Results for AGFll 14 Flux 185
XXIV
Table A.29 Sieving Results for AGF 1129 Flux 190
Table E. 1 ANOVA of Tensile Strength 236
Table E.2 ANOVA of Percentage Elongation 237
Table E.3 ANOVA of Impact Strength 238
Table E.4 ANOVA of Vickers Hrdness 239
Table E.5 ANOVA of Penetration 240
xxv
NOMENCLATURE
Symbol Unit Description

e Degree Angle of Repose
h mm Height of the Heap of Flux
r mm Radius of the Base of the Heap of Flux
A Wavelength
V Volt Welding Voltage
I A Welding Current
V cm/min Welding Speed
d mm Nozzle to Plate Distance
H kJ/mm Heat Input Per Length
- g/cc Bulk Density
- "C Temperature
- Mpa Tensile Strength
- Joule Impact Strength
xxvi
LIST OF ACRONYMS
Acronyms Explanation
SAW Submerged Arc Welding

DCEP Direct Current Electrode Positive
DCEN Direct Current Electrode Negative
PIAZ Heat Affected Zone
AC Alternating Current
B.I Basicity Index
RSM Response Surface Methodology
XRD X-Ray Diffraction
DTA Differential Thermal Analysis
AGF Agglomerated Flux
ANOVA Analysis of Variance
VHN Vickers Hardness Number
DC Direct Current
AC Alternating Current
[a] Activity/ Concentration of

Substances
xxvn
CHAPTER 1
INTRODUCTION
1.1. Submerged Arc Welding (SAW)

1.1.1 Introduction
Since the inception of submerged arc welding process in early 1930s in USA and USSR
simultaneously, a wide variety of fluxes have been developed and a large number of flux
electrode combinations have been in use for welding of various types of steels. The flux
formulations have, however, been done primarily on the basis of empirical data as well as the
practical experience. This approach has been found to be time consuming and has limited
applicability because it lacks the necessary scientific basis required to formulate the flux.
Engineering requirements demand innovative approaches to the formulation and manufacture
of welding consumables. The use of higher strength steels for applications requiring both
higher strength and toughness requires better control of the weld-metal composition. The
possibility of hydrogen damage increases with the use of these high-strength low-alloy steels
and the need to reduce the moisture absorption of the welding flux is essential. New stainless
steel alloys, such as duplex stainless steel, are requiring new consumables and methods to
predict ferrite content. The advancement of titanium alloys for corrosion resistant structural
applications will need new consumables. Similarly new aluminum alloys, such as the
aluminum-lithium alloys, will advance with acceptable welding consumables and practice.
The use of sensors combined with microprocessor control to increase the productivity of
automatic welding processes and improve the reliability and integrity of the production welds
will require more stringent control on the flux composition. Fluxes and wires created for use
with microprocessor-controlled automatic welding processes will need to be compositionally
consistent and homogeneous.
The welding consumable plays a very complex role in the welding processes. Components of
these consumables must protect the weld deposit, generate acceptable plasma, stabilize the
arc, influence the bead morphology and promote the ability to weld out of position. Fluxes
must melt in a specific temperature range, have a density lower than the weld deposit, refine
the weld metal and be easily removed. Welding consumables affect the weld both physically
and chemically.
Even though the patents and product literature illustrate continual advances in welding
consumable technology, the science of welding consumable behaviour is still evolving. It
would be very cost effective at this time to use a sound scientific approach (observation,
characterization, correlation, modeling, etc.) to establish a fimdamental basis to the
engineering of the manufacture and use of welding consumables.
The properties of welded joints were earlier determined by the study of microstructures of the
weld metal and heat affected zone. However, it was realized later that gas - slag - metal
reactions play a significant role in controlling the properties of the welded joint. It was
fiirther emphasized that flux type and composition play an unquestionably important role in
controlling the gas - slag - metal reactions in submerged arc welding which in turn affect the
weld metal chemistry.
Based on the different types of commercial fluxes being used for welding of different types
of steels for various applications, a number of researchers have reviewed the submerged arc
welding fluxes and classified them into different categories based upon their chemical
composition, method of manufacture, chemical nature and use. However, such classifications
of fluxes do not give any idea regarding formulations of the fluxes and selection of flux wire
combination for any particular application.
1.1.2 Principle of operation
Submerged arc welding (SAW) is an arc welding process in which arc is concealed by a
blanket of granular flux [1-3]. Heat for SAW is generated by the arc between consumable-
wire and the workpiece. The arc is maintained in a cavity of molten flux or slag, which
refines the weld metal and protects it fi"om atmosphere contamination. Alloy ingredients in
the flux may be present to enhance the mechanical properties and crack resistance of the
weld deposit. Figure 1.1 shows setup for automatic SAW. An electrode is continuously fed
into the joint by mechanically powered drive rolls. A layer of granular flux, just deep enough
2
to prevent flash through, is deposited in front of the arc. Electrical current, which produces
the arc, is supplied to the electrode through the contact tube. The current can be either direct
current (dc) with electrode positive (reverse polarity), with electrode negative (straight
polarity), or alternating current (ac). Figure 1.2 shows the melting and solidification sequence
of SAW. After completion of welding and solidification of weld metal, the unused flux may
be screened and reused. The solidified slag may be collected, crushed, resized, and blended
back into new flux. Recrushed slag and blends of recrushed slag with unused or virgin flux
are chemically different from new flux. Blends of recrushed slag may be classified as
welding flux, but cannot be considered the same as the virgin flux [4].
Hspper \ Wire
Comttf
System **^
Eleetede Lead
Torch
mt&Am 6iTm€\0SBm
i:^
gjii|g.|i:-':.::''' g-t..::'::'...:a-.:.:.
Bsm Mital
Figure 1.1: Equipment for Submerged Arc Welding [5]

Hux Hqpper
Gmndar Mux; Blanks

Molten Flux
SdUdtned ujuddl M d r e n w d d Arc With Base metal
Figure 1.2: Typical Weld Pool Dynamics of a Submerged Arc Weld
Submerged arc welding is adaptable to both semiautomatic and fully automatic operation,
although the latter, because of its inherent advantages, is more popular. In semiautomatic
welding, welder controls the travel speed, direction, and place of the weld. A semiautomatic
welding gun is designed to transport the flux and wire to the operators, who weld by
dragging the gim along the weld joint. Semiautomatic electrode diameters are usually less
than 2.4 mm to provide sufficient flexibility and feed ability to the gun assembly. Manually
guiding the gun over the joint requires skills, because the joint is obscured from the view by
the flux layer. In automatic SAW, travel speed and direction are controlled mechanically.
Flux may be automatically deposited in front of the arc, while the unftised flux may be
picked by a vacuum recovery system behind the arc.
To increase deposition rate or welding speed, more than one wire can be fed simultaneously
into the same weld pool. Figure 1.3 shows the twin arc process, in which the two electrodes
are fed into the same weld pool while sharing a common power source and contact tip. In
tandem arc SAW, multiple electrodes are arranged with one in front of other. Each electrode
has an independent power supply and contact tip. The spacing, configuration, and electrical
nature of the electrodes may be arranged to optimize welding speed and bead shape.
mms.
JJ
I
ra
HHIIIIIIIMMP
•••••••JHi
.FEEDPRrVE:
MBCHANISM
CONTACT
TUBB
n POWER
SOURCE
^
WORKPIECE
5
Figure 1.3: Schematic Representation of the Equipment for
Arc Welding With Twin-Wire Electrode [6]
1.1.3 Advantages
(a) The arc is under a blanket of flux, which virtually eliminates arc flash, spatter,
and fume (thus making the process attractive from an environmental standpoint).
(b) The high current density increases penetration and subsequently the need for edge
preparation decreases.
(c) High deposition rates are possible in SAW.
(d) The flux acts as a scavenger and deoxidizer to remove contaminants such as
oxygen, nitrogen, and sulphur from the molten weld pool. This helps to produce
sound welds with excellent mechanical properties.
(e) Low hydrogen weld deposits can also be produced.
(f) The shielding given by the flux is not sensitive to wind as in shielded metal arc
welding and gas metal arc welding.
(g) Minimal welder training is required.
(h) The slag is collected, reground, and sized for mixing back into new flux as
prescribed by manufacturers and standard procedures.
1.1.4 Limitations
(a) The initial cost of wire feeder, power supply controls, and flux-handling
equipment is high.
(b) The weld joint needs to be placed in the flat or horizontal position to keep the flux
positioned in the joint.
(c) The slag must be removed before subsequent passes can be deposited.
(d) Because of high heat input, SAW is commonly used to join steels of more than
6.4 mm thickness.
1.1.5 Process applications

1.1.5.1. Types of metals
The submerged arc welding is commonly used to join plain carbon steels. Alloy steels can be
readily welded with SAW if care is taken to limit the heat input as required to prevent
damage to the heat-affected zone (HAZ). Low-heat-input procedures are available for
welding alloy steels and heat-treated steels to prevent grain coarsening and cracking in the
HAZ. Submerged arc welding can be used to join stainless steels and nonferrous alloys. It is
also commonly used to produce a stainless or nonferrous overlay on top of a base metal.
1.1.5.2. Joint configurations

The most common weld deposits made with SAW are groove, fillet, and plug welds, and
surfacing deposits. For groove welds, the characteristic of deep penetration of SAW plays a
role in joint selection. Plate up to 15.9 mm thick can be completely welded from one side
using a square butt joint with a 0.8 mm root opening. Beveled joints should be used for
multiple-pass welds on plate thickness greater than 15.9 mm. Baking bars should be used to
prevent loss of flux, melt through, and to ensure ftill-penetration weldments on one-sided
joints. Although groove welds are generally made in the flat position, it is possible to develop
special procedures to weld horizontally.
Single-pass fillet welds with throat sizes up to 7.9 mm thick and multiple-pass fillet welds are
usually made in the horizontal position. Single-pass fillet welds over 7.9 mm thick are
usually made in the flat position. Submerged arc plug welds are made in the flat position by
pudding the electrode into the center of the hole until the welds are complete. Surfacing
welds are made with both wire and strip electrode. Surfacing is done to rebuild worn parts
with a wear-resistant material or to overlay a plain steel part with stainless steel or other
alloys [7-9].
The surface to be welded should be kept clean and free from grease, oil, paints, moisture, and
oxides to prevent porosity. All slag from tack welds or previous layers should be removed.
Tack welds should be positioned so that the submerged arc weld completely melts out the
tack. The workpiece should be properly cleaned to minimize distortion and the need for tack
welds.
1.1.5.3. Industrial uses

The SAW is used for joining thick steel sections; it is primarily used for ship-building, pipe
fabrication, pressure vessels, and structural components for bridges and buildings. Other than
joining, the SAW is used to build up parts and overlay with stainless or wear-resistant steel
(for example, rolls for continuous casting steel, pressure vessels, rail car wheels, and
equipment for mining, mineral processing, construction, and agriculture).
1.1.6 Equipment and supplies

The power supplied to the contact tip may be direct current or alternating current. The details
of different polarities are given as under [10].
(a) Direct current electrode positive (DCEP) provides the best control of the bead shape
because the arc is very stable and the molten weld pool is almost constant. The weld
penetration is better.
(b) Direct current electrode negative (DCEN) provides the highest filler metal deposition rate
as shown in Figure 1.4, but penetration is low due to cathode spot climbing up to the
electrode wire.
(c) Alternating current (AC) minimizes arc blow. This assumes increasing importance of
welding current above 900 A. AC is generally used, where multiple wires are used,
because with DC the magnetic fields of the wires can deflect the arcs.
Ik.
m
Q
S
a
Q
400 500 600 7^ 800 §00 1000

Welding current, A
Figure 1.4: Effect of Welding Current on Deposition Rate [10]
1.1.7 Power sources

The power sources should provide the high amperages at 100% duty cycle, required by the
most 'SAW installation. Generally, welding is done in a range from 400 to 1500 A. Current
as high as 4000 A at 55V, or as low as 150A at 18V may be used [5].
1.1.7.1 Constant voltage DC type

The constant-voltage power supply is self-regulating. It is used with a constant-speed wire
feeder. The wire diameter controls current, the electrical stickout, and the wire-feed speed,
while the voltage is controlled by the power supply. Constant-voltage dc power is the best
choice for the high-speed welding of thin steel. The volt-ampere curve is almost flat as
shown in Figure 1.5.
O
o
>
too 150 200 250 300 350

CURRENT. A
Figure 1.5: Volt-Ampere Output Relationship for a Constant Voltage Power Source [5]
1.1.7.2 Constant current DC type

Unlike constant voltage, constant-current power supplies are not self-regulating. Therefore
they must be used with voltage -sensing variable-wire-feed speed controls. A constant-
current wire feeder monitors arc voltage and adjusts the wire-feed speed in response to
changes in the arc voltage. The wire-feed speed control attempts to maintain a constant arc
length, while the power supply controls the arc length. Constant current power sources of the
rectifier or motor-generator type are shown in Figure 1.6, and are widely used for DC
welding. If the current required exceeds the output of a single machine, two or more rectifiers
or generators of the same type can be connected in parallel.
AMPERES
Figure 1.6: Typical Volt-Ampere Characteristics of a "Drooping" Power Source with

Adjustable Open Circuit Voltage [5]
9
1.1.7.3 Alternating current type
The constant-current output of a conventional ac machine varies with time like a sine wave
drooping through zero with each polarity reversal. The voltage associated with the current is
approximately a square wave. Alternating current is generally supplied by heavy duty
welding transformer of 1000, 1500 or 2000 A capacity. For 'SAW with AC power source,
the open-circuit voltage should be minimum of 80 volts to ensure the re-ignition of the arc
during current reversal and high frequency unit is essentially to be used. Generally, AC is not
used in single-wire welding process. It is used with multi-wire welding process in which
trailing wire is generally connected to AC power-source.
1.1.8 Process parameters

Control of the operating variables in submerged arc welding is essential if high production
rates and welds of good quality are to be obtained. These variables, in their approximate
order of importance, are the following [5].
1.1.8.1 Welding current

Welding current is the most important variable because it controls parameters such as
deposition rate, penetration, and dilution. An increase in welding current at a constant voltage
will decrease the flux-to-wire ratio, while a decrease in current will increase the flux-to-wire
ratio. Welds made at excessively low current will tend to have little penetration and higher
width-to-depth ratios. Welds made at an excessively high current will have deep penetration,
high dilution, more shrinkage, and excess buildup. Low current will also produce a less
stable arc than higher currents.
1.1.8.2 Electrode polarity

The direction of current flow will also affect the weld bead profile. The current may be direct
with the electrode positive (reverse polarity), electrode negative (straight polarity), or
alternating. Reverse polarity is commonly used in this process. For a given set of welding
conditions, reverse polarity will produce wider welds with more penetration at a lower
deposition rate than straight polarity. Straight polarity welding will contribute to narrower
beads with less penetration and more buildup. Because straight polarity reduces base plate
10
dilution, it is frequently used in surfacing applications. For the same welding current, the
deposition rate with straight polarity is higher than with reverse polarity. The bead shape,
penetration, and deposition rate for alternating current fall between those straight and reverse
polarity. Alternating current is used when welding current exceeds 900A and on multiple-
wire applications to reduce arc blow and arc interaction.
1.1.8.3 Current density

In SAW, the current density in the electrode also plays a role in bead shape and penetration.
Smaller-diameter electrodes with a high current density will produce narrower beads with
deeper penetration than larger-diameter electrodes. Larger- diameter electrodes are able to
bridge larger root openings. In case where a given current can be achieved with two different
electrode diameters, the smaller electrode will produce the higher deposition rate.
1.1.8.4 Welding voltage

Like current, welding voltage will affect the bead shape and the weld deposit composition.
Increasing the arc voltage at a constant current will increase the flux-to-wire electrode ratio,
while decreasing the voltage will reduce the flux-to-wire electrode ratio. Increasing the arc
voltage will produce a longer arc length and a corresponding wider, flatter bead with less
penetration. Higher voltage will increase flux consumption, which could then change deposit
composition and properties. Slightly increasing the arc voltage will help the weld to bridge
gaps when welding in grooves. Excessively high voltage will produce a hat-shaped concave
weld, which has low resistance to cracking and a tendency to undercut. Lower voltages will
shorten the arc length and increase penetration. Excessively low voltage will produce an
unstable arc and a crowned bead, which has an uneven contour where it meets the plate.
1.1.8.5 Travel speed

Variations in travel speed at a set current and voltage also affect bead shape. As welding
speed is decreased, heat input per length of joint increases, and the penetration and bead
width increase. The penetration will increase until molten metal begins to flow under the arc
and interfere with heat flow at excessively slow speeds. Excessively high travel speeds will
promote a crowned bead as well as the tendency for undercut and porosity.
11
1.1.8.6 Electrical stickout
In SAW, the current flowing in the electrode between the contact tube and the arc (electrode
extension) will cause F R heating, resulting in a voltage drop across the length of electrode.
This resistance heating and subsequent voltage drop can be used to obtain higher deposition
rates. Normal electrode extension for solid SAW wire is approximately 8 to 12 times the
electrode diameter. As this length increases at a constant current, so does the resistance
heating and the melt-off rate. To compensate for the voltage drop and the increase in wire-
feed speed, the voltage must be increased to obtain a properly shaped bead. Extending the
electrode 20 to 40 times the diameter can increase deposition rates by more than 50% [11].
1.1.8.7 Flux layer depth

Flux layer is another variable that will alter the appearance, penetration, and the quality of
submerged arc weld. If the flux layer is too deep, greater-than-normal amount of flux will be
melted, resulting in weld beads that are narrower than normal. Some surface imperfections
may also appear because gases may be trapped by the deep flux layer. If the flux layer is too
shallow, the arc will flash through, and the bead will have a rough appearance or porosity due
to lack of shielding from the atmosphere. The correct depth of flux is just enough to prevent
flash through. This will allow welding gases to escape while providing adequate protection.
1.1.8.8 Type efflux and particle distribution

The type of flux and particle distribution will also affect the mechanical properties of weld
joint.
1.2 Submerged Arc Welding Fluxes

American welding society defines welding flux as a material used to dissolve or facilitate
removal of oxides and other imdesirable substances. Welding slag has no formal definition,
but is generally used to describe the fused residue after welding. Submerged arc welding
fluxes usually consist of combination of manganese oxide and silica or lime and silica with
additions of various other oxides to produce complex oxide fluxes. Lately, fluxes having
titania, magnesia, alumina, and fluorspar as the major constituents have also been produced.
12
The submerged arc welding flux performs several fiinctions which are intimately connected
with the welding process. These include [12]:
(a) Improvement of electrical conductivity.
(b) Formation of slag, which covers the molten metal and provides shielding to it to achieve
slow cooling rates and also the covering for molten metal to avoid atmospheric
contamination.
(c) Additions of different elements can be made.
(d) Metallurgical influences can be made.
(e) Slag cleans the weld-metal.
A flux must have certain characteristics if the molten metal is to be protected from the
atmosphere. The melting temperature of the flux must be less than that of molten metal so
that no gases are trapped between the flux and the metal. The flux must also remain molten
until the metal has solidified so that the complete fluxing action will occur. Consequently, the
upper limit for the mehing point of welding compositions is generally considered to be about
1300°C when composition is selected from a phase diagram.
An ideal flux is ductile at high temperature to prevent oxidation of the weld metal and brittle
at room temperature to facilitate slag removal. The temperature coefficient of expansion of
the flux should be different from that of the steel to aid in slag removal. The friability and
temperature coefficient of expansion of the flux can be controlled by regulating the
composition of flux.
The flux viscosity of a welding flux must be high enough to give it impermeability to
atmospheric gases and to prevent it from running away from the molten metal and flowing in
front of the arc, leading to possible overlapping by the molten metal. The flux must also be
fluid enough in the welding operation to permit rapid solution of nonmetallic such as oxides
and the evolution of gases from the molten metal.
The flux may remove the oxides present in the weld metal in any of the three ways. The
oxides may be: (a) combined with the welding flux to form a more complex oxide with a
13
lower melting range and decreasing viscosity, (b) dissolved in the molten flux, or (c) most
commonly, reduced electrochemically to a metal component of the deposit.
In addition to removing oxides, the flux is expected to aid in removing sulphur and
phosphorus compounds which may be present. The ability of flux to remove sulphur or
phosphorous components is a function of its composition; fluxes with low sulphur or
phosphorus contents have particularly effective cleansing action. The lime-silica fluxes are
more effective than the manganese oxide-silica compositions in removing sulphur and
phosphorus.
In addition to the functions described above, the physical presence of the flux offers some
other advantages. Spatter losses are negligible, since all of the molten metal is contained
beneath the blanket of flux. Operator discomfort is also greatly reduced because there is no
open arc and much of the heat from the process is contained under the flux.
1.2.1 Classification of submerged arc welding fluxes

Submerged arc welding may be produced in one of three ways: fused, bonded and
agglomerated. Jackson [13] gives a detailed description of this and also the list of advantages
and disadvantages of each method of production.
1.2.1.1 Fused fluxes

Fused fluxes are produced by mixing dried powders followed by melting at a temperature of
1500-1700°C. The molten flux is then chilled either by water shooting or by the use of chill
plates. The final materials are crushed and sized to an upper limit of 12 and 200 mesh
respectively.
Advantages
(a) Extremely good chemical homogeneity.
(b) Fines may be removed without changing composition of flux.
(c) Most of the fused fluxes are a non-hygroscopic product that is vmaffected by
moisture, thus simplifying storage problems.
14
(d) The unused portion may be collected and used again without significant change in
particle sizing or flux composition.
(e) Suitable for highest speed of travel.
Disadvantages
The primary disadvantage is the inability to add de-oxidizers and Ferro-alloys because of the
high temperature used during production of flux.
1.2.1.2 Bonded fluxes

Bonded fluxes are produced by dry mixing finely ground (approx. 100 meshes)
powders and bonding them with additions of potassium silicate or sodium silicate solutions.
The bonded mixture is palletized, dried at low temperature, broken up and screened.
Advantages
(a)The low temperature involved in the bonding process permits the extensive use of metallic
de-oxidizers and ferro-alloys.
(b) The lower bulk density permits the use of a thicker layer of the flux in the weld zone.
(c) Generally it has excellent detachability of fused slag.
Disadvantages
(a) In many cases, fines cannot be removed for fear of altering the flux composition.
(b) Bonded flux may have a tendency to absorb moisture.
(c) The molten flux pool may evolve gases.
1.2.1.3 Agglomerated fluxes

(a) In agglomerated fluxes, finely powdered ingredients are mixed and grovmd dry in a
mixer. The mixture is steadily moistened by spraying with a solution of alkaline
silicate and the mixing is continued. The mixer blades are suitably designed to assist
I agglomeration. The silicate solution initially fills the spaces between the pores of the
particles. When subsequently dried, the water evaporates, leaving the binder as
bridges between the particles. Like low hydrogen electrodes, agglomerated fluxes
have to be baked at around 500°C to remove the last traces of water clinging to the
15
silicate binder. After baking, the flux is graded to a specified granule size by sieving
and packed in water proof containers.
(b) The agglomeration technique of manufacturing fluxes has led to the development of
fluxes with low Si02 content and high CaO, MgO and CaF2 contents.
(c) The higher the flux basicity, the lower are the oxygen and sulphur contents of the
weld metal and the higher its notch toughness value.
(d) Agglomerated basic fluxes of good performance characteristics have maximum
basicity between 2.6 and 3.2.
(e) Basicity higher than 3.2 makes the flux unworkable mainly because of very high
solidification range and very low viscosity.
(f) Agglomerated fluxes normally baked at approx 500°C have a tendency to pick up
moisture when stored under humid atmosphere due to hygroscopic properties of CaO
and MgO. Experience has shown that by baking the flux at 800°C while
manufacturing, the flux becomes insensitive to moisture.
(g) Agglomerated fluxes have lower bulk density than fiised fluxes. This has the
advantage that under identical conditions of current and voltage, less flux is melted
and forms a less slag as compared to ftised fluxes.
(h) The use of agglomerated fluxes exhibits the greatest advantage for hard facing
applications.
1.2.2 Classification relative to basicity index

Apart from the manufacturing method, another common method used to describe submerged
arc welding fluxes is the basicity index, BI. The basicity index is the ratio of strongly bound
metallic oxides to weakly bound metallic oxides [14]. The basicity index is defined thus:
r. • V T J .T^TN CaO + MgO + CaF^+Na,0 +K,0 + BaO + Q.5{MnO + FeO) ,_,

Basicity Index (B.I.) = ^^ 1 (1.1)
SiO^+0.5{TiO^+ZrO^+AlÔ^)
The basicity index is an estimate of the oxygen content in the weld metal and is, therefore,
used to predict weld metal properties. Basic fluxes tend to have lower weld metal oxygen
16
content with good weld metal toughness, while acidic fluxes tend to produce higher weld
metal oxygen content and coarser microstructure with a lower resistance to cleavage. Fluxes
may be considered acidic, basic and neutral as per the following concept [14].
(a) A flux is considered acidic if B.l.< 1
(b) Neutral flux if 1< B.l.< 1.5
(c) Basic flux if 1.5< B.I .< 2.5
(d) Highly Basic if B.l.> 2.5
Acidic fluxes are typically preferred for single-pass welding due to their good operating and
weld bead characteristics. In addition, acidic flux usually has more resistance than basic
fluxes to porosity caused by plate contamination by oil, rust, and mill scale.
Basic fluxes tend to have better impact properties than acidic fluxes. This advantage is
particularly evident on large multiple-pass welds. Highly basic fluxes generally produce weld
metal with very good impact properties in large multiple-pass weldments. Basic fluxes tend
to exhibit poorer welding characteristics than acidic fluxes on single-pass welds, their use
should be limited to large multipass weldments where good weld metal notch toughness is
required.
The composition ranges of various flux types along with their chemical character and
basicity index range are given in Table 1.1.
17
Table 1.1 Composition Ranges of Various Fluxes [15]
Type Character Composition Ranges Expressed as Oxides Basicity
Si02 Ti02 Zr02 A1203 Mn Ca Mg CaF2 Index
0 0 0
Mn-silicate Acid 30- 0-10 35- 0- 0-10 0.5-1.0
40 45 10
Ca-silicate Acid 50- - - 03- - 25- 08- 00- O.S-0.8
High silica 60 05 30 12 02
Ca-silicate neutral 30- - - 10- 0-5 25- 10- Os- 0.8-1.2
medium 40 15 30 15 lo
silica
Ca-silicate basic 30- - - OS- - 35- 00- 10- 1.2-1.7
low siUca 40 OS 45 05 15
Aluminate neutral 10- OS- 00- 20- 10- 00- - OS- 0.5-1.0
Ti02 20 IS 10 50 20 05 IS
Aluminate basic AI2O3 + CaO +MgO = 45% min. 1.2-2.0
basic AI2O3 = 20% min.
Fluoride basic 05- - - 10- - 10- 30- 20- 2.5-3.5

Dasic 12 15 15 40 30
1.2.3 Comparative consumption of fused and agglomerated fluxes

As already explained in the advantages of agglomerated flux that the bulk densities of
agglomerated fluxes is lower than the fused fluxes, which affects the melting of
agglomerated fluxes. Comparative consumption of agglomerated and fused flux for different
currents and voltages is shown in Figure 1.7 [15].
18
600
•s
560
5
<
az 520
480
FUSED FLUX / j i
X / j 490/0 j
3 4.40
O
z
o
400
360
A ^H )
/ \ ' /
CL
320 y/'iV>h\ J/
w 280 ^•^2q<jl,,,.»»-"'^GGl,OMER;ATE0 FLUX
z;
o
o 240^3*1
AMPERES 4O0 500 600 700 800
VOLTS 28 30 32 34 36
SPEED
CM/MIN. 45 45 45 45 45
Figure 1.7: Comparative Consumption of Fused and Agglomerated Fluxes [15]
1.2.4 Particle size of submerged arc welding flux

Particle size is an important property of a submerged arc welding flux. During welding, a
finer flux produces more dense and compact burden over the flux cavity, thereby reducing its
volume. Evolution and escape of gases from the weld bead then becomes more difficult, and
this results in a characteristic aerated slag. Finer fluxes are also less tolerant to oil and rust
because of the inability of the resultant gases to escape. An increased particle size in a flux
can give rise to a humped peaky weld bead with increased penetration. Fluxes for currents
higher than 800 A or for the multiwire applications are supplied to a fine grain size than
similar fluxes for lower current levels. Using coarse flux at very high current can result in
considerable arc instability and inferior weld bead surface. At high current, the flux melting
is increased leading to gas evolution while the viscosity of the slag is very low. Under these
conditions, a coarse, low density flux burden would cause violent eruptions in the arc cavity
which sometimes would rise above the flux cover leading to the defects [15].
19
CHAPTER 2
REVIEW OF EXISTING WORK
2.1 Introduction and Previous Work

Many researchers have studied about the design and development of different fluxes,
interaction between slag and metal, properties of fluxes, flux consumption and bead
morphology for submerged arc welding of steel. In this chapter, the previous work of
researchers, their important findings and some gaps are identified for the current research.
Lewis et al. (1961) developed a new flux and filler wire for submerged arc welding of HY-
steel used in submarine hulls. The developed flux raises weld metal impact properties by
reduction of oxygen, and inclusions, and due to changes in weld metal microstructure.
Investigators used reasonable amount of aluminum as a deoxidizer, during the development
of electrode for submerged arc welding, higher amount of aluminum and titanium as a
deoxidizer did not improve the weld-metal impact properties [16].
Robinson (1961) pointed out that the significant factors controlling SAW electrode
melting rate were current, composition used, electrode extension, electrode diameter and
flux [17].
Pokhodnya and Kostenko (1965) studied about fusion of electrode metal and its interaction
with the slag during submerged arc welding. It is revealed in the study that the electrode
metal is transferred to the weld pool in the form of droplets, but if the current is very high,
the metal may be transferred to weld pool without forming droplets. The variations in the
chemical composition of the weld metal with the mode of metal transfer in the weld pool
were also observed [18].
Bennet and Stanley (1966) used CaO-SiOz-CaFa, CaO-SiOa-AlaOs, CaO-SiOa-TiOa and

miscellaneous commercial fluxes for the submerged arc welding of Q.T.35 steel. The results
were analyzed for these commercial fluxes in terms of weld metal compositions which have
been foimd closely linked with the flux composition, but the contribution of heat input was
20
less. The study revealed that the flux composition affects the mechanical properties like
impact strength and tensile strength very well in the presence of alloying elements or their
oxides. The investigators revealed the role of Si and Mn in the weld metal due to the
presence of SiOi and MnO in the flux. The study also revealed that the level of S and P
depends upon the desulphurising and dephosphorising powers of the welding fluxes [19].
Butler and Jackson (1967) worked on 75 synthetic fluxes and weld tested them to ascertain
if any compositions were suitable for a submerged arc welding flux. A good welding
performance was achieved with the composition of 25%) CaO, 30% TiOj, 40% SiO 2 and
5% CaFj when used with a filler metal containing 1% manganese. The investigators
correlated the welding properties of the various flux compositions with the ternary phase
diagram. The various characteristics, maximum speed at which the synthetic flux welded
satisfactorily, appearance and density of synthetic slag, appearance of fused melt removed
from the weld, electrode melting rate, flux consumption, and ratio of flux consumption rate
to electrode melting rate, were subsequently evaluated. The welding performance improved
with the Si02 constituent. The flux performance was compared on the basis of metallic
particle contents in the microscopic examination of the polished sections and the flux. The
fluxes having high metallic particle contents were found to have better welding performance
than fluxes with the same melting temperature, but lower metallic particle contents [20].
Bennet (1970) investigated the moisture content and the rough appearance of weld surface
by using basic fluxes. The investigator also compared the slag detachability and weld bead
appearance of basic fluxes used in the study with the commercial acid fluxes [21].
Colvin (1970) studied the development of basic submerged arc welding fluxes and discussed
contemporary research on the subject. He stated that the use of basic fluxes in preference to
acid fluxes, in particular non-carbonate basic fluxes with a basicity greater than 2.6, can
appreciably extend the scope of the submerged arc welding process by depositing weld
metal of improved metallurgical quality without detracting the physical and economic
characteristics of the process. By employing alloyed wires with a predetermined excess of
deoxidants, the process becomes extremely robust in the application [22].
21
Drayton (1972) studied the influence of current, current type and polarity, voltage and flux
composition as welding parameters on process measures such as metal deposition rate, weld
bead profile and flux consumption using single wire for submerged arc welding. The bead on
plate welds were made by changing one welding parameter at a time independently. The
study revealed that current, current type and polarity abort the performance to a lesser extent,
while flux composition and voltage influences the process performance significantly. In the
second part of the study, statistically designed experiment was used to determine the effects
and interactions of three levels of voltage, current type, nozzle to plate distance, flux
composition and welding speed on metal and slag deposition rates, wire feed speed, heat
input, penetration, bead width, etc. The study revealed the presence of a narrow band of
optimum parameters setting [23].
Ivochkin et al. (1972) used flUer metal in the form of powder and set the submerged arc
welding process variables on the basis of calculation and experimental method. The
investigators used this powder for two sided automatic submerged arc welding for butt joints
in plate of low carbon and low alloy steels up to 50 mm thick without edge beveling. Due to
its economical viability and high productivity this method of submerged arc welding has
been introduced in production [24].
Palm (1972) described that the dissolved oxygen is the deciding factor in determining impact
properties of weld metal, in particular the transition temperature, and it is controlled by the
molten slag and its active components [25].
Potapov and Babin (1974) investigated an improvement in technological properties in

respect of weld bead appearance, reduced gas evolution, improved arc stability and slag
detachment etc. by using fiised slag as a slag forming base of the agglomerated flux. The
researchers also investigated the transfer of alloying elements from a fused agglomerated flux
into the deposited metal without losses from oxidation if the elements in the flux composition
have equal or lower affinity. Additions of alloying elements with a great affinity will lead to
avoid the contamination of the deposited metal with globules of silica glass [26].
22
Sorokin and Sidlin (1974) studied the effect of transfer of alloying elements from the
electrode into the deposited weld metal. The combined actions of alloying elements and of
the oxidizing action of the coating do not affect the degree of adoption of molybdenum and
tungsten by the deposited metal. The electrode with non-oxidizing coatings reduces the % of
carbon when its content in the electrode is less than 0.07%, whereas element is oxidized at
greater carbon contents. Similarly the electrode with non-oxidizing coating reduces the
silicon when its content in the electrode is less than 0.6%, but at greater contents it is
oxidized [27].
Ferrera and Olson (1975) investigated performance of the MnO-SiO 2 -CaO system as a
welding flux. Fluxes, derived from pure components, were tested using the submerged arc
welding process. Viscosity, arc stability, and weld-bead morphology were reported and
compared with existing accepted concepts for welding flux behaviour. The following
conclusions have been drawn [28]:
(a) As basicity increases for both the MnO-SiO 2 and MnO-SiO 2 -CaO flux system, the
viscosity decreases. When some of the MnO is replaced by CaO, the viscosity
increases, but the major component affecting viscosity is SiO 2.
(b) The basicity of the arc increases with increase in MnO and CaO content. The CaO
increases the stability of the arc more than MnO does. Arc stability behaviour is
found to be controlled by the ionization potential of the metal ions from the flux.
(c) Arc penetration is independent of melting temperature and viscosity for the MnO-
Si02 flux system. Arc penetration is dependent on melting temperature and
composition for the MnO-SiO ^ -CaO system.
(d) The width to depth-of-penetration ratio shows a type of dependence similar to that
found for arc penetration.
(e) The weld bead angle shows dependence on the MnO and CaO content of flux. Weld
bead morphology correlates better with capillary considerations than with kinetic
functions such as melting temperature and viscosity.
Wittstock (1976) selected submerged arc fluxes for carbon and alloy steels. In the
preliminary study he had gone with the historical background of the submerged arc welding
23
fluxes and considered recent needs of submerged arc welding industry. The flux behaviour,
flux chemistry and slag/metal interactions were studied by the investigator before selecting
the flux. The investigator elaborated some guidelines for the selection of electrode, flux and
materials used for welding. It was emphasized that AWS standard may be useful for
comparing fluxes outcomes and the literature be consulted to be sure about the compatability
of the flux electrode combination. The researcher also categorized most of the fluxes for
specific performance groups [29].
Eagar (1978) identified the source of weld metal oxygen contamination during submerged
arc welding. The investigator defined the level of oxygen in submerged arc weld metal which
is controlled by Si02 decomposition in acidic fluxes, whereas the oxygen level of basic
fluxes is controlled by the oxygen potential of the slag determined by the FeO content of the
slag. It was also elaborated that flux basicity index is an indirect measure of the oxygen of
the weld metal [30].
Potapov and Kurlanov (1978) carried out a quantitative evaluation of the basicity of
welding fluxes. In this study, the investigators described results of a comparative analysis of
the bsaicity equations. The basicity equations gave a differentiated evaluation of the effect of
various components of the slag and of their ratios on the activity of silica and manganese
oxide included in the flux composition. The investigators proposed a more accurate formula
for the calculation of flux basicity; the formula takes into accoimt the affinity of acid oxides
for basic and vice versa, in the formation of complicated complex compound in the flux. The
study also revealed the evaluation of the chemical activity of the flux components [31].
North et al. (1978) investigated the influence of flux formulation on the oxygen content of
submerged arc weld deposit. They used low oxygen potential constituents e.g. CaF 2, Al 2 O 3
and CaO, for this study and observed their result for fused fluxes and agglomerated fluxes. It
was found that oxygen contents were markedly lower using fiised fluxes as compared to
agglomerated fluxes. Researchers also investigated that the oxygen pick-up during welding
with CaF2 - AI2O3 - CaO fluxes was determined by gas/metal reactions occurring at the
electrode tip region. No correlation existed between the FeO content of slags and the oxygen
24
content of weld deposit. Further the addition of aluminum to a 45% CaF ^, 35%A12 O 3,20%
CaO flux lowered the oxygen contents of weld deposits to below that of the electrode while
the addition of up to 1.2% aluminum to the flux improved weld notch toughness [32].
Charles and Entrekin (1979) studied the influence of flux basicity on weld-metal structure
and examined the effects of two fluxes of different basicities when used with two
experimental wires and a columbium-bearing base plate. The authors showed that the
toughness of submerged arc weld is highly dependent on microstructure. The study
demonstrated that a low-carbon manganese weld wire alloyed with Mo,Ti and B was capable
of producing the weld-metal microstructure in a Cb-bearing steel plate when used with a
fully basic welding flux but similar wire without Mo, Ti and B was not able to produce welds
of satisfactory toughness due to its lack of hardenability [33].
Schwemmer, Olson and Williamson (1979) studied the effect of addition of iron oxide to
manganese silicate base welding flux on weld penetration. Prepared preflised welding fluxes
were used to conduct weld bead tests on 1020 and 4340 steel. The ionic state of iron in the
flux was determined using Mossbauer analysis and correlated with the arc stability, bead
morphology, capillarity, viscosity, and Mossbauer line widths. It was revealed that an
increase in viscosity, arc stability, surface tension between the flux and liquid metal, and a
decrease in surface tension between the flux and the base metal resulted in deeper
penetration. Researchers also obtained an expression relating penetration to arc stability,
viscosity, and interfacial tensions [34].
Koukabi (1979) developed fused fluxes by using CaFa, AI2O3 and CaO and adding some
more elements like zirconium, vanadium and titanium/boron. The effects of these elements
were studied with reference to mechanical properties, cleanliness and microstructure in
submerged arc welds deposits. The zirconium acting as a deoxidant decreased the volume
fraction of inclusions, improved toughness, ductility and hot workability. Vanadium was
used for precipitation strengthening and titanium/boron was used for improving notch
toughness [35].
25
Chai and Eagar (1980) carried out parametric study of two commercial fluxes namely
calcium silicate and manganese silicate during submerged arc welding. The study of the
factors controlling weld chemistry during submerged arc welding has shown that the weld
deposit chemistry is primarily dependent upon flux composition; however, others seemingly
insignificant operating parameters may produce notable changes in weld chemistry. It was
also concluded that the oxygen potential of the flux is not a uniquely definable quantity nor
can the weld metal mechanical properties be determined solely by the base plate, flux and
electrode chemistry [36].
Davis and Bailey (1980) described how submerged-arc flux composition influences element
transfer. The results are compared with the DINSSS? classification as a means of predicting
element transfer. After the study, the following conclusions are revealed [37]:
(a) Silicon transfer to the weld depended on the flux Si content. Network formers in the
flux, such as Zr, Ti, and Al, increased the transfer of Si.
(b) Sulphur was removed from the weld by all the basic fluxes. Other fluxes could
remove sulphur if they contained sufficient Ca and Mg and if their own sulphur
contents were low.
(c) Carbon was removed from welds by all fluxes, the removal being greatest with high
Si contents fluxes and least with the basic fluxes.
(d) The investigation has shown that compositional changes on welding can be
understood in terms of flux chemistry, although it is not possible to put forward a
simple predictive scheme.
(e) Generalized parameters, such as the DIN 8557 classification and basicity index, can
not be used to predict weld compositions accurately, although they are all usefiil as
rough guides.
Silvinski (1981) investigated the effects of the components of the flux systems CaF2-Si02.
Al203-MgO and some oxides on the density, viscosity and surface and inter-phase tensions of
slag during interaction with type EP 690 metal. The data obtained by the interconnection
between the technological and physico-chemical properties were used as an input in the
development of a new flux for welding high-alloy steels. It was also concluded that the low
26
activity of silicon dioxide in fluxes tends to improve the resistance of welds to hot
cracking [38].
Podgaetskii and Galinich (1981) described a review study on the structures of molten
welding slags. The study established the connection between the chemical composition, the
high-temperature physical and chemical properties, and the structure of molten welding slag.
On the basis of chemical composition molten welding slags were divided into three
categories oxide, salt and sah-oxide. The structures of these three molten slags were
evaluated between their melting points and 2000 °C [39].
Davis and Bailey (1982) described a study of ten separate flux properties at ambient and
elevated temperatures and obtained specific correlations between coefficient of expansion
and slag detachability and between melting point and current capacity. They selected flux
composition, bulk density, grain size and flowability for measurement at ambient
temperature and their effects were correlated with the welding behaviours. Satisfactory
experimental methods were developed to measure all these properties except those involving
prolonged heating of molten flux [40].
Chai and Eagar (1982) studied the stability of metal oxides in terms of oxygen level in the
weld metal by producing binary CaF ^ -metal oxide fluxes. The metal oxides used for this
study were SiOj, MnO, MgO, Alj O3, Naj O, Ti02, Kj O and CaO. The study revealed
that the stability of metal oxides during welding does not directly depend on their
thermodynamic stability because some fluxes despite their good thermodynamic stability
decompose into suboxides during welding. Such oxides produce higher level of oxygen in the
weld metal and reduce weld impact strength. The effect of CaFa in reducing the level of
oxygen in weld metal is dependent upon the stability of metal oxide present. Investigators
also revealed that Si02 and MnO are the primary sources of oxygen contamination in the
fluxes of low FeO. It was finally observed that the formation of suboxides and metal vapour
in the plasma in addition to the stability of oxides play important roles in the reactivity of
flux [41]. ^
27
Snyder and Pense (1982) observed the effects of titanium on the mechanical properties and
microstructure of submerged arc weld metal in HSLA Si-Al-killed low sulphur steels
containing varying amounts of titanium. The study revealed that toughness is strongly
dependent on the titanium and manganese levels. It increases with decreasing titanium
content and increasing manganese content. The toughness also improved by using higher
basicity fluxes because of reduced oxygen level in the weld metal. Molybdenum or
molybdenum plus nickel additions improved toughness. The yield strength and tensile
strength also increased with the carbon, manganese, silicon and titanium but manganese and
silicon exhibited weak effects as compared to carbon and titanium [42].
Terashima and Tsuboi (1982) described work to develop submerged arc welding
consumables for steel plate of tensile strength above 785 N/mm^. A highly basic
agglomerated flux combined with wire containing sufficient Si produced weld metal of high
strength and toughness and low diffusible hydrogen content. A highly basic agglomerated
flux containing adequate amounts of carbonate and fluoride could produce low hydrogen
weld metal. With an agglomerated flux, no nitrogen dissolution into the molten weld pool
from the arc atmosphere was observed. Transverse cracks in multipass butt welds stemmed
from hydrogen induced cracking propagating along prior austenite grain boundaries; the
occurrence of cracking was affected by weld metal strength as well as the diffiisible
hydrogen content [43].
Kohno et al. (1982) developed new fluxes which can consistently microalloy weld metal
with Ti and B. The most probable way seemed to be the addition of Ti and B to the weld
metal obtained by reducing their oxides in fluxes during welding. The developed new fluxes
improved weld metal toughness for HSLA steels. The fluxes exhibited good arc stability and
good slag detachability when used during multipass welding for thick plates. The developed
new fluxes were used during submerged arc welding for the applications of nodes cans of
offshore platforms, a sea berth, LPG tanks and ships [44].
Potapov (1983) described the resuUs of calculating the reactions between slag and metal in
automatic submerged arc welding. The equations given for determining the final amounts of
silicon, manganese, sulphur, and phosphorus for this process make it possible, with an
28
accuracy sufficient for practical purposes, to calculate the increase in the contents of the weld
metal, and consequently also to forecast the properties of the weld metal. For calculation
between slag and metal the welding materials and welding conditions must be known [45].
Davis et al. (1983) studied the effects of titanium and boron additions to submerged arc
welding fluxes. The authors added different boron compounds to commercial submerged arc
welding fluxes to ascertain the transfer of boron to weld metal, and their effects on weld
metal chemistry and microstructure. Small amounts of boron influenced microstructure and
ferro-titanium deoxidant additions promoted formation of acicular ferrite [46].
Indacochea and Olson (1983) studied the relationship of weld-metal microstructure and
penetration to weld-metal oxygen content. The authors revealed that the weld-metal oxygen
content influenced the behaviour of the welding process. It was finally concluded that flux
composition can alter the weld-metal composition, affecting weld-bead morphology,
microstructure and mechanical properties [47].
Mitra and Eager (1984) studied the transfer of Cr, Si, Mn, P, S, C, Ni and Mo between the
slag and the weld pool for submerged arc weld for low alloy and stainless steels with calcium
silicate and manganese silicate fluxes. The following conclusions were drawn [48]:
(a) The transfer of chromium is strongly dependent on the type of flux used. Lime
silicate fluxes produced weld metal with much higher chromium content than
manganese silicate fluxes, although both fluxes contained the same amount of
chromium (III) oxide.
(b) The transfer of silicon is greatly influenced by the basicity index of flux used as well
as by the initial chromium content of the electrode. The lower the basicity index of
flux and the greater the initial chromium content of the electrode, the greater will be
the amount of silicon in the weld metal.
(c) The manganese content of the weld metal depends mainly on the amount of
manganese oxide in the flux and the initial manganese content of the electrode. The
amount of the other alloying elements present does not appear to have significant
influence on the transfer of manganese.
29
(d) The lime based fluxes produced weld metal with a lower phosphorous content than
that produced by using the flux containing manganese oxide. This may be due to the
stronger dephosphorizing power of lime.
(e) The carbon content of the weld metal depended strongly on the type of flux used.
Use of the MnO-Crj 03-Si02 flux resulted in a lower content of the carbon in the
weld metal due to the higher oxidizing potential of the flux. The low carbon
austenitic steels picked carbon from the fluxes.
Lau et al. (1985) studied the sources of oxygen and nitrogen contamination during
submerged arc welding of different CaO-Al ^ O 3 based fluxes of varying basicity index. By
analyzing the oxygen and nitrogen contents at the electrode tip, droplet and weld metal
stages, it was demonstrated that the major source of nitrogen contamination is by reaction
with the atmosphere at the electrode tip. However oxygen contamination is associated not
only with reactions with the atmosphere, but also with flux decomposition and flux
entrapment.
There was no significant change in the nitrogen content between the electrode and droplets
stages, but there was marked increase iri oxygen content as droplets traversed the arc column
for all the fluxes investigated. The differences in behaviour between nitrogen and oxygen
pickup were further illustrated by the influence of current at a fixed heat input. An increase in
welding current led to an increase in oxygen content but a small decrease in nitrogen level in
the weld pool. This was attributed to factor such as flux decomposition, which becomes more
prevalent as the arc temperature increases at higher currents [49].
Lau et al. (1986) studied about gas/ metal/ slag reaction in submerged are welding using
CaO-AÔs based fluxes, and reported that changes in oxygen, aluminum and manganese at
the electrode tipped arc columns were dominated by flux decomposition, while at the metal
stage slag- metal reaction occurred. The reactions were most extensive in MnO containing
fluxes, and resulted in a significant loss of oxygen as well as metallic species such as
aluminum by the serration of oxidized products into the slag phase [50].
30
Tandon et al. (1988) studied the effect of flux characteristics on transformed reaction and
HAZ during submerged arc welding. In this study fluxes based on CaO-TiOj system with
minor additions of Al 2 O 3, MgO and CaF ^ have been designed and manufactured by fusion
method. The physical properties like melting point, melting range and volumetric specific
heat were determined. It was further shown that the desired properties of the welded joint
may be obtained by proper and judicious selection of flux for the given energy input [51].
Burck et al. (1990) evaluated the effects of addition of CaF2, CaO and FeO to manganese
silicate flux system on weld metal chemistry. Comparisons were made between AISI 4340
steel and low carbon steel welds to understand the weld metal chemistry [52].
Mitra and Eagar (1991) reported a critical review of thermodynamic theories of slag-metal
reactions. The authors established a quantitative relationship between the compositions of
welding consumables used, the welding parameters employed, and the resulting weld metal
chemistry. The investigators presented the work in three parts. In part I, the previous work
done by researchers is reviewed, and the strength and shortcomings of prevailing theories are
described along with preliminary experiments which prove or reject certain hypotheses [53].
In part II, an entirely new theory was presented to explain the changes in weld chemistry, and
a kinetic model was formulated to predict weld metal composition [54]. In part III, the
authors presented the verification of the theory through several different experiments [55].
Gupta and Arora (1991) studied the weld bead geometry and heat affected zone which are
generally dependent on the welding parameters and may also be influenced by the chemical
composition of flux. The different flux constituents which impart different physical and
chemical properties to the flux may also affect the weld bead geometry and heat affected
zone. The authors alsoi reported that the depth of penetration, weld bead width and weld
reinforcement depend upon welding parameters and flux basicity [56].
Panday et al. (1994) studied effects of SAW parameters and fluxes on element transfer
behaviour and weld metal chemistry while welding mild steel plate. The weld metal
composition showed, in general, gain of silicon and loss of carbon, manganese and sulphur
elements. The results also revealed that welding current and voltage have an appreciable
influence on element transfer, as well as on weld composition [57].
31
Ana et al. (2003) studied chemical and structural characterization of fluxes for submerged
arc welding. Three flux formulations were preferred using mineral oxides for agglomerating
and sintering process. A commercial agglomerated and sintered flux was used for
comparison. The four fluxes were analyzed chemically by atomic absorption and X-ray
diffraction to determinate the quality and type of oxide formed, as well as to assess the
change in oxidation number after the sintering process at 950 DC of the initial compound. A
differential thermal analysis was also carried out from 1000 to 1350 DC in order to
determine the temperature for phase transformations and melting of the different compounds
formed in the sintering process. It was concluded that the most reactive components react
quickly in the weld pool and might be either absorbed in the slag or retained in the weld as
inclusions [58].
Ana et al (2005) studied the effect of flux composition on microstructure and tensile
property of AISI 1025 steel in SAW process. Three flux compositions were used with a low
carbon electrode. A commercial flux composition was used for comparison. The welding
conditions were kept the same. Microstructure and macrostructure of welds were observed
with light and scanning electron microscopes (SEM) [59].
Kanjilal et al. (2006) developed model by rotatable design technique to study the combined
effects of flux and welding parameters on chemical composition and mechanical properties
of submerged are weld metal. For this study, bead-on-plate weld deposits on low carbon steel
plates were made. The investigators had drawn the following conclusions on the basis of
their study [60]:
(a) Among the welding parameters, polarity has a profound influence on weld metal
chemical composition.
(b) Welding speed influences weld metal carbon content through oxidation reaction;
whereas weld metal sulphur and phosphorus content are affected by dilution of weld
deposit.
(c) Welding current influences weld metal manganese content through slag metal
reaction.
32
(d) Transfer of nickel from flux to weld is found to be impeded by oxides formed during
slag- metal reaction.
Kanjilal et al. (2007) worked on prediction of element transfer in submerged are welding by
developing suitable regression models in terms of flux ingredients with the help of statistical
design of mixture experiments. Bead-on-plate weld deposits were made at fixed welding
parameters using SAW fluxes as per extreme vertices algorithm of mixture experiment in a
CaO-MgO-CaFi-AbOs flux system [61].
Ademola David Adeyeye and Festus A. Oyawale (2008) reported that traditional welding
flux development has been by done by labor intensive experiments. The extensive and
expensive trial and error experimentation is needed because it is often difficult to know a
priori how the flux ingredients interact to determine the operational characteristics of the flux
and find performance of the welded structure [62].
Abhay Sharma et al. (2008) developed a mathematical model of flux consumption during
twin-wire submerged arc welding. During welding a certain amount of flux is consumed and
converted into slag and the flux consumption being a function of process parameters directly
influences the productivity of the process [63].
Vinod kumar et al. (2009) worked on development of cost effective agglomerated fluxes
from waste flux dust for submerged arc welding. Approximately 10-15% of the flux used in
SAW gets converted into very fine particles termed as flux dust before and after welding due
to transportation and handling. If welding is performed without removing these very fine
particlesfiômthe flux, the gases generated during welding are not able to escape resulting
into surface pitting and even porosity. On the other hand, if these fine particles are removed
by sieving, the cost of welding will be increased significantly. This flux dust dumped /thrown
will create the pollution. Therefore, attempts have been made by the authors to develop the
acidic and basic agglomerated fluxes by utilizing wasted flux dust. The fluxes prepared from
the waste flux dust were used without any compromise in mechanical properties and quality
of the welding joint [64].
33
Bang et al. (2009) observed the effects of flux composition on the element transfer, tensile
strength, and impact toughness of the weld metal during submerged arc welding. The study
also observed that both carbon and manganese show negative delta (A) quantity in most
combinations. The negative values indicate the transfer of the elements from the weld metal
to the slag during welding [65].
Kumar et al. (2011) observed the effect of flux, welding current, arc voltage, and travel
speed on changes in micro-hardness and microstructure of the heat-affected zone (HAZ)
during submerged arc welding. The process was optimized with the help of Taguchi
technique. The study concluded that the welding current and type of flux were the most
significant factors leading to changes in microhardness and metallurgical properties. Travel
speed and arc voltage were found to be insignificant in relative comparison [66].
Vinod Kumar (2011) developed flux from waste flux dust for submerged arc welding.
Response surface methodology was used for the development of mathematical models to
predict critical dimensions of the weld bead geometry and shape relationship. The model
accuracy was checked by F-test and t-test. Main and interaction effects of the process
variables like voltage, current, welding speed and basicity index on bead geometry and shape
factors were presented in graphical form [67].
Shen, Oguocha, I.N.A., and Yannacopoulos (2012) investigated that a series of

measurement was carried out on specimens of submerged arc welded plates of ASTM A709
Grade 50 steel to determine how variation in heat input achieved using single and double
wires affected bead reinforcement, bead width, penetration depth, contact angel, heat affected
zone (HAZ) size, deposition area, penetration area and total molten area. The authors
concluded that the bead reinforcement, bead width, penetration depth, HAZ size, deposition
area and penetration area increased with increasing heat input, but the bead contact angle
decreased with it. The electrode efficiency are also correlated with heat input [68] .
34
2.2 Present Work
A number of investigators designed and developed different agglomerated fluxes with minor
additions of alloying constituents during submerged arc welding keeping in mind the
objective of welding industries. Such fluxes are not only economically viable but have good
techno-mechanical properties, less reinforcement, more weld bead penetration also at the
weld joint. The present investigation using CaO- Si02 - AI2O3 based flux systems with minor
additions of alloying constituents like MnO, CaF2, NiO, MgO and Fe-Cr has been taken as
these are the most widely used fluxes at the commercial level. The ranges of these alloying
constituents are designed on the basis of binary and ternary phase diagrams for different
oxide and fluoride systems. These fluxes are formulated by using response surface
methodology technique and developed by agglomeration technique. The physical properties
of these developed agglomerated fluxes are studied with the help of different setups. The
effect of these alloying elements on mechanical properties of weld joints, microstructures of
weld joints, bead morphology, and element transfer behaviour are correlated with the help of
response surface modeling and standard values.
2.3 Identified Gaps in the Literature

The following research gaps have been identified in relation to the present work:
a) A few researchers have worked on the modeling of basic and alloying constituents of
agglomerated fluxes to see their effects on techno- mechanical properties and bead
morphology.
b) A few researchers have worked on the physical properties and characteristics of
submerged arc welding agglomerated fluxes and their correlation with the flux
consumption, penetration and arc stability.
c) The research in this area is excessively classified and data is not readily available due
to the commercial constraints for the development of flux in the welding industries.
d) There is hardly any reported work on using five flux constituents as variables while
modeling of agglomerated fluxes during submerged arc welding.
e) The research area demands more explanation because not much work has been
carried out owing to interdisciplinary nature of research.
35
The above listed points led to the development of a new agglomerated flux which may be
used as a commercial flux in future during submerged arc welding process.
2.4 Impact of Proposed Research in Academia/Industry

a) CaO-AÔs -Si02 based flux systems have been designed for study as these are the
most widely used fluxes at the commercial level. The different constituents namely
MnO, CaF2, NiO, MgO and Fe-Cr, are also added into the basic constituents to
improve techno-mechanical and physical properties of the fluxes.
b) The design is based on ternary and binary phase diagrams depending on major and
minor constituents.
c) Thirty two agglomerated fluxes are formulated oti the basis of response surface
methodology design (RSM).
d) Characteristics of developed agglomerated fluxes are helpful to choose good flux at
commercial level.
e) Test plates are welded for developed agglomerated fluxes.
f) In the present work, element transfer behaviour of the fluxes and other related effects,
e.g. bead geometry and shape relationships, tensile strength, impact strength,
percentage elongation and hardness of weld metal etc. are studied.
g) The mechanical properties and bead morphology characteristics have been modeled
with the help of response svirface methodology. The optimum values of these
characteristics are ascertained by using desirability function.
36
CHAPTERS
DESIGN, FORMULATION AND DEVELOPMENT OF FLUXES
3.1 Design of Fluxes

In the present study, CaO-AliOs - Si02 based fluxes with minor additions of MnO, CaFi, and
MgO have been designed on the basis of binary and ternary phase diagrams for different
oxide and fluoride systems [69]. In these fluxes varying amount of high carbon Ferro-chrome
and Nickel oxide were added to study the transfer of chromium and nickel during welding.
After ascertaining the compositions and constituents of fluxes, these were prepared by
agglomeration using inorganic binders. Details of reaching out at the constituents and
compositions of fluxes, and to manufacture these are discussed below.
3.1.1 Selection of flux constituents
Selection of flux constituents is primarily based on the composition of base plate and
electrode employed for welding. CaO-AlaOs - SiOa based flux systems have been selected for
study as these are the most widely used fluxes at the commercial level. Since fluxes
containing Si02 give rise to dissolved oxygen in the weld metal which adversely affects the
toughness and thus the mechanical properties of the welded joint, addition of NiO and FeCr
has been done to study the transfer behaviour of these elements to the weld pool fi-om the
flux containing the above mentioned constituents. CaFi has been added to control the
viscosity and the melting range of the flux. CaFa along with MgO bring down the liquidus
temperature as is evident in the phase diagram
3.1.2 Flux composition design

The primary concern while designing the composition of fluxes was to keep the liquidus
temperature below 1350 °C. This was arrived at by studying the CaO- SiOa, CaO- AI2O3,
CaO-Si02-Al203, CaF2-CaO-Si02, CaO- Al203-CaF2 and CaO-Al203-MgO phase diagrams.
In case of CaO-Si02 binary system, there are two eutectic points, one is at a temperature of
1435 °C for the composition of SiOa - 65 wt% CaO - 35 wt% and another is at 1460 °C for
the composition SiOa 45- wt%, CaO-55 wt%. For CaO-Al203 system, the minimum liquidus
temperature is 1400 DC for a composition of AI2O3 - 53 wt% and CaO - 47 wt%. From this
37
diagram, it is observed that with increase or decrease in AI2O3 proportion from a certain
value, the liquidus temperature is raised. From CaAbOs-CaFi ternary system, the liquidus
temperature of 1300 DC is minimum for a composition CaFa - 63 wt%, CaO - 15 v^% and
5Ca0.3Al203 - 22 wt%. Here we find that for an increase up to a certain extent in CaF2
proportion there is a decrease in the liquidus temperature, when CaO.AbOs proportion is
kept constant. In the ternary phase diagram of CaO.Al2O3.MgO, it is seen that with the
addition of MgO upto 10%, the melting point of oxide system comes down appreciably.
Therefore, it is clear that the CaF2 and MgO when added in small amoimts generally lower
down the liquidus temperature and higher amounts of CaO and MgO raise it. Certain amount
of AI2O3 addition causes lowering of melting point. Since AI2O3 does not show good arc
stability due to higher ionisation potential, some amounts of K2O has been added in the form'
of K2Si03 to improve the arc stability. K2Si03 also acts as binder. The phase diagram for the
system CaO-Si02-Al203 is shown in Figure 3.1. The details of basic constituents, alloying
constituents and their ranges are listed in Table 3.1.
RANKINITE
3CaO-2SO,rf-"
CoQ •AiPi laCoO-TAtjOj CaO-AyDj CtO-?Mfi3 CoO-eAlgOj f'fljpj

\\'E!GHT *
Figure 3.1: Phase Diagram of CaO-SiOi-AliOa System [69]
38
Table 3.1: Flux Constituents and their Ranges
Basic Constituents &

Alloying Constituents CaO SiOj Al,03 MnO CaFj MgO NiO Fe-Cr
Amount (wt %) 20 35 10 5-10 10-20 5-15 0-10 0-10
3.2 Response Surface Methodology

3.2.1 Introduction
A scientific approach to plan the experiments is a necessity for efficient conduct of

experiments. By the statistical design of experiments the process of planning the experiment
is carried out, so that appropriate data will be collected, analyzed by statistical methods
resulting in valid and objective conclusions. When the problem involves data that are subject
to experimental error, statistical methodology is the only objective approach to analysis. Thus
there are two aspects of an experimental problem, the design of the experiments and the
statistical analysis of the data. These two points are closely related since the method of
analysis depends directly on the design of experiments employed. The advantages of design
of experiments are as under:
(a) Number of trials is significantly reduced.
(b) Important decision variables which control and improve the performance of the
product or the process can be identified.
(c) Optimal setting of the parameters can be found out.
(d) Qualitative estimation of parameters can be made.
(e) Experimental error can be determined.
(f) Inference regarding the effect of parameters on the characteristics of the process can
be made.
In the present work, the response surface methodology has been used to plan the experiments
and subsequent analysis of the data collected.
39
3.2.2 Methodology
Response surface methodology, or RSM, is a collection of mathematical and statistical
techniques useful for analyzing problems in which several independent variables influence a
dependent variable or response, and the goal is to optimize this response [70]. In many
experimental conditions, it is possible to represent independent factors in quantitative form as
given in Equation 3.1. Then these factors can be thought of as having a functional
relationship between the response Y and x/, x^,... ,Xk the k quantitative factors
Y= f (X,, X2, X3, X4, Xk) ± e (3.1)
The function/is called response surface or response function. The residual e measures the
experimental error [70]. For a given set of independent variables, a characteristic surface is
responded. When the mathematical form of / is not known, it can be approximated
satisfactorily within the experimental region by a polynomial. Higher the degree of
polynomial better is the correlation but at the same time cost of experimentation becomes
higher.
For the present work, RSM has been applied for developing the mathematical models in the
form of multiple regression equations for the design and development of agglomerated fluxes
during submerged arc welding. In applying the response surface methodology, the dependent
variable is viewed as a surface to which a mathematical model is fitted. For the development
of regression equations related to various quality characteristics of fluxes for submerged arc
welding, the second order response surface has been assumed as:
Y=b„ +Xbi X, +Xb, xf + Jbii ^. ^j ±^ (3-2)

i=I i=l i<j=2
This assumed surface Y contains linear, squared and cross product terms of variables Xj's. In
order to estimate the regression coefficients, a number of experimental design techniques are
available. Box and Hunter [71] have proposed that the scheme based on central composite
rotatable design fits the second order response surfaces quite accurately.
3.2.3 Central composite second order rotatable design

In this design, standard error remains the same at all the points which are equidistant from
the centre of the region. Let the point (0, 0, —, 0) represent the centre of the region in which
40
the relation between Y and x is under investigation. From the results of any experiment, the
standard error, Cr, of Y can be computed at any point on the fitted surface. This standard error
acts as a function of the co-ordinates Xj's of the point. Because of rotatability condition, this
standard error is same at all equidistant points with the distance p from the centre of region
i.e. for all points, which satisfy the following equation:
xf +xl+.... + xl-p^ = constant (3.3)
Central composite rotatable design is subdivided into the following three parts:
• Points related to 2^ design, where k is the number of parameters and 2 is the number
of levels at which the parameters is kept during experimentation.
• Extra points called star points positioned on the co-ordinate axes to form a central
composite design with a star arm of size a.
• A few more points added at the centre to give roughly equal precision for response Y
with a circle of radius one.
The factor a is the radius of the circle or sphere on which the star points lie. With k >5 the
experimental size is reduced by using half replication of 2^ factorial design. With half
replication, a become 2^'^'^'*. Also, no replication is needed to find error mean square, since
this can be found out by replicating the centre points [72&73]. The components of central
composite second order rotatable design for different number of variables are given in Table
3.2. A pictorial representation of different points for the case of 3 variables is shown in
Figure 3.2.
Table 3.2: Components of Central Composite Second Order Rotatable Design [70]
Variables (k) Factorial Star Points Center Total (N) Value of a
Points (2'') (2k) Points (n)
3 8 6 6 20 1.682
4 16 8 7 31 2.000
5 16* 10 6 32 2.000
6 32* 12 9 53 2.378
* Half replication
41
(0,1.682,0)
(-1, -1,1),
(-1.682,0,0) (1.682,0,0)
(0,0, -1.682)
6 Points at center
(-1, -1, -1)
(0, -1.682,0)
Figure 3.2: Central Composite Rotatable Designs in 3X-Variables [70]
In this study the Design- Expert trial version 8.0.6 software was used for central composite
second order rotatable design with five variables as flux constituents for the development of
agglomerated fluxes during submerged arc welding. The flux constituents as variables—
MnO , CaF2, NiO, MgO and Fe-Cr, and their levels are given in Table 3.3.
Table 3.3: Flux Constituents Used in the Experiment and their Levels
Variable -2 -1 0 1 2
MnO(wt%) 5 6.25 7.5 8.75 10
CaF2(wt%) 10 12.5 15 17.5 20
MgO(wt%) 5 7.5 10 12.5 15
42 Contd.
NiO(wt%) 0 2.5 5 7.5 10
Fe-Cr (wt%) 0 2.5 5 7.5 10
3.3 Development of Fluxes

The development of flux was accomplished in the following stages.
3.3.1 Mixing of flux constituents
The constituents of agglomerated fluxes for submerged arc welding were mixed by the Ball
Mill setup as shown in Figure 3.3. A Ball Mill consists of a cylindrical shell slowly turning
about a horizontal axis and filled to about 1/4'*^ of its volume with solid grinding medium (i.e.
metallic balls etc.). When the ball Mill is rotated, the grinding elements (balls) are carried to
the side of the shell nearly to the top, from where they fall on the particles under gravity. In a
Ball Mill, most of the size reduction is done by impact. The energy expended in lifting the
grinding units is utilized in reducing the size of the particles. Ball Mill can accept a feed size
of 12mm or less and deliver a product size in the range of 50)j.m. The speed of Ball Mill
varies from 60 to 70 r.p.m. As the product size becomes fine, the capacity of mill reduces and
the energy requirement increases. For effective operation of the mill, it should be operated at
65 to 80% of its critical speed [74&75].
Cylindrical Shell
Motor
Figure 3.3: Bail Mill Setup
43
3.3.2 Preparation and baking of fluxes
The methods available for producing fluxes are fusing, sintering or agglomeration. Out of
these three methods, the first two were considered to be unsuitable as the temperature
involved would have decomposed most of the carbonates. The agglomeration method was
favoured because it is a room temperature process except for its final baking operation. The
thirty two fluxes were formulated by response surface methodology and produced on a
laboratory scale by agglomeration by taking different % weights for each combination as
given in Table 3.4. A small batch (2 kg) of weighted quantities of powdered chemicals taken
for each combination was thoroughly mixed in Ball Mill. Approximately 5 % solution of
potassium silicate, the binder, was then sprayed on the mixture. As the moisture content of
the mixture increased, it no longer created dust and the mixing speed was stepped up. The
same procedure is continued for a few seconds. After that the mixture is taken out as shown
in Figure 3.4 and rubbed with the hands properly. The powdered mixture soon began to
coalesce into small spheres which increased in size with further moisture addition and mixing
time. When the granular mixture was visually estimated to have reached 1- 2 mm diameter as
shown in Figure 3.5, the mixing process was stopped. The moist flux was removed and
spread on flat sheet for 24 hrs. The dried flux was baked in an electrically heated furnace for
3 hrs. at 800 DC . The dried flux was allowed to cool down to room temperature and stored
in moisture free boxes. The basicities of these different developed agglomerated fluxes are
inclined towards the basic fluxes.
Table 3.4: Chemical Composition of Fluxes and their Percentage Weight

Flux Chemical composition (wt. %)
code
Run CaO Si02 AI2O3 MnO CaF2 MgO NiO Fe-Cr
AGFllOl 24 20 35 10 6.25 12.5 7.5 2.5 7.5

AGF1102 25 20 35 10 8.75 12.5 7.5 2.5 2.5
AGF1103 26 20 35 10 6.25 17.5 7.5 2.5 2.5
AGF1104 5 20 35 10 8.75 17.5 7.5 2.5 7.5
AGF1105 11 20 35 10 6.25 12.5 12.5 2.5 2.5
AGF1106 4 20 35 10 8.75 12.5 12.5 2.5 7.5
AGF1107 16 20 35 10 6.25 17.5 12.5 2.5 7.5
AGF1108 18 20 35 10 8.75 17.5 12.5 2.5 2.5
AGF1109 9 20 35 10 6.25 12.5 7.5 7.5 2.5
44
Contd.
AGFlllO 13 20 35 10 8.75 12.5 7.5 7.5 7.5
AGFllll 23 20 35 10 6.25 17.5 7.5 7.5 7.5
AGF1112 17 20 35 10 8.75 17.5 7.5 7.5 2.5
AGF1113 19 20 35 10 6.25 12.5 12.5 7.5 7.5
AGF1114 27 20 35 10 8.75 12.5 12.5 7.5 2.5
AGF1115 6 20 35 10 6.25 17.5 12.5 7.5 2.5
AGF1116 21 20 35 10 8.75 17.5 12.5 7.5 7.5
AGF1117 31 20 35 10 5 15 10 5 5
AGF1118 29 20 35 10 10 15 10 5 5
AGF1119 28 20 35 10 7.5 10 10 5 5
^
AGF 1120 J 20 35 10 7.5 20 10 5 5
AGF1121 15 20 35 10 7.5 15 5 5 5
AGF 1122 8 20 35 10 7.5 15 15 5 5
AGF 1123 20 20 35 10 7.5 15 10 0 5
AGF 1124 12 20 35 10 7.5 15 10 10 5
AGF 1125 2 20 35 10 7.5 15 10 5 0
AGF 1126 10 20 35 10 7.5 15 10 5 10
AGF1127 32 20 35 10 7.5 15 10 5 5
AGF1128 1 20 35 10 7.5 15 10 5 5
AGF 1129 7 20 35 10 7.5 15 10 5 5
AGF1130 14 20 35 10 7.5 15 10 5 5
AGF1131 22 20 35 10 7.5 15 10 5 5
AGF 1132 30 20 J5 10 7.5 15 10 5 5
'*-''î-%W:i
Figure 3.4: Mixture Taken Out From Ball Mill (Pre Final Stage)
45
Figure 3.5 Prepared Flux (Final Stage)
46
CHAPTER 4
CHARACTERISTICS OF FLUXES AND WELD JOINTS
(EXPERIMENTAL PROCEDURE)
In this chapter the experimental procedure of characteristics effluxes and weld joints are
described. The details of experimental procedures are given as under.
4.1 Characteristics of Fluxes and their Measurements
The aim of this study is to measure different physical properties of submerged arc welding
flux. Flowability, Grain size. Bulk density and viscosity are selected for the measurement.
These properties throw considerable light in revealing the welding behaviour.
4.1.1 Flowability
The flowability of flux is determined by placing known weight in the setup as shown in
Figure 4.1.The flowability of fluxes is generally assessed by determining the angle of repose
of the granular fluxes. The angle of repose is defined as the maximum angle possible
between the surface of a pile of flux and the horizontal plane. It is determined by allowing a
mass of 200 g of granular flux to flow freely through an orifice from a certain height and
from a conical heap on the horizontal surface. As the heap is formed, the particles slip and
roll over each other until the mutual friction between the particles just balances the
gravitational force. The angle which the heap forms with the horizontal surface is the angle
of repose and is determined by Equation 4.1.
tan0 = ^ (4-1)
r
Where, 0 is the angle of repose, h is the height of the heap of flux and r is the radius of the
base of the heap of flux. In this way, the repose angle of each developed agglomerated flux is
determined by taking the average value. The average angle of repose is compared with
standard values as reported in Table 4.1.
47
Funnel
Support
Angle of Repose Granular Flux
Stand
Figure 4.1: Setup of Flowability
Table 4.1: Relationship between Angle of Repose ( 6) and Flow Properties of

Powders [76]
S.No. Angle of repose(6') in degrees Flow Properties

1. D25 Excellent
2. 25-30 Good
3. 30-40 Satisfactory
4. 40-50 Poor
5. D 50 Very poor
4.1.2 Grain size

The grain size is also one of the important properties of the flux. The flux grain size
distribution influences the weld bead shape and surface quality [38]. The grain size of the
developed fluxes is measured with the help of microcontroller based Electromagnetic Sieve
Shaker (EMS-8, ELECTROLAB make) as shown in Figure 4.2. The microcontroller is used
to set the process time from 1 min to 99 min and the amplitude from 0.5 to 2.5 mm. Sieving
is one of the oldest methods of classifying developed flux by particle size distribution. In the
flux industry, sieving is usually the method of choice for the classification of the coarser
grades of single flux. A sieve separates a specific sample material in two fractions — one is
retained by the sieving media, which is the rejected or oversized material and the other,
48
which is passed through the openings. This particle size sieving for sample preparation is not
an exact science for sampling as it is bound to give errors. To eliminate the errors,
electromagnetic sieve shakers have been introduced [77].
Sie\'e Clamping
knob
Sieve
Computer
Receiver
Front Panel
Figure 4.2: Setup for Grain Size Measurement
The values of grain size for different developed fluxes are measured by taking 300 g sample
of flux. The sample was placed in the top sieve and shaken for 4 min. The fraction remaining
in each sieve was weighed and expressed as a percentage. The sieving results and histogram
for different developed fluxes are drawn. This procedure is followed for all thirty two
developed agglomerated fluxes.
4.1.3 Bulli density

Bulk density is considered an important property of flux when estimating flux usage, as the
weight consumed will be proportional to the bulk density when maintaining constant flux
burden. The behaviour of the arc cavern is also likely to depend on bulk density, as the flux
burden counteracts the pressure within the activity [78]. The aim of this study is to measure
bulk density property of developed agglomerated fluxes used during submerged arc welding,
which reveals how flux composition controls welding behaviour. To measure bulk density,
49
50 g sample of flux is poured into a caliberated container and put inside the Bulk Density
Measurement Setup (ELECTROLAB make, Model ETD-10/20/1020) as shown in Figure
4.3. After settling flux inside the caliberated container, the occupied volume of flux is
calculated. The bulk density of the flux is calculated in gm/cc by dividing weight of flux
taken from the volume occupied by the flux in the caliberated container. The same procedure
is repeated for each flux thrice and the average of these values is determined. In this way, the
bulk density of all developed agglomerated fluxes is ascertained.
Cylinder
Cylinder
holder
Figure 4.3: Setup for Bulk Density Measurement
4.1.4 Viscosity
The viscosity of a welding flux is one of the most important properties. The flux must have
fluidity enough to allow the transfer of elements between flux and metal and allow gases to
escape. But it must also be viscous enough to support the weld puddle. Viscosity of the
molten flux plays an important role in the production of an acceptable welded joint. Viscosity
is not only a measure of its resistance to flow but is also directly related to the rate coefficient
of mass and heat transport in the flux affecting the weld metal protection and heat transfer
during welding. A flux must remove elements and gases from the weld metal by absorbing
atoms or molecules at the liquid metal-slag interface. The viscosity, however, must also be
high enough to protect the weld metal from atmospheric gases. If the flux is too viscous, it
50
will not absorb or transport gases at the metal slag interface resulting in surface pocking of
the weld bead. Selection of the proper viscosity also depends on the position of welding. A
more viscous flux is desirable for vertical and overhead welding to assist reinforcing and
protecting the weld pool. The measurement of viscosity of a developed agglomerated flux
was attempted by the change of viscosity with temperature using Brookfield digital
viscometer (model DV-E). The principle of operation of the viscometer is to rotate a spindle
(which is immersed in the flux) through a calibrated spring. The viscous drag of the fluid
against the spindle is measured by the spring deflection. The spring deflection is measured
with a rotary transducer which generates a torque signal. The measurement range of viscosity
(in centipoises) is determined by the rotational speed of the spindle. The size and shape of
spindle are customized as per the requirement [79]. The measurement of viscosity is so
tedious that it is carried out for only one flux by taking 200 g of a developed flux in a
crucible which is inside the furnace. The viscosity with respect to temperature was recorded
by the setup as shown in Figure 4.4. The viscosity is recorded in the viscometer and the
temperature is recorded by using thermocouple of type R (Pt-Pt and 13% Rh). The values
thus obtained are then plotted.
Shipping
Laboratorv Stand
Spindle
Temperature
indicator
Muffle
Furnace
Thermocouple
Figure 4.4: Setup of Viscosity Measurement
51
4.1.5 X-Ray Diffraction (XRD) of developed agglomerated fluxes
X-ray Diffraction (XRD) is a high-tech, non-destructive technique for analyzing a wide range
of materials, including fluids, metals, minerals, polymers, catalysts, plastics,
pharmaceuticals, thin-film coatings, ceramics, solar cells and semiconductors. XRD has
become an indispensable method for materials investigation, characterization and quality
control. The features of X-ray Diffraction are reported in Table 4.2.
The crystalline phases, different types of ions and their distribution in developed
agglomerated fluxes (AGF1101-AGF1132) were carried out by X-ray diffraction (XRD) as
shown in Figure 4.5, with the monochromatic light wavelength 1.540600 A (Cu Ka).
Table 4.2: Features of X-Ray Diffractometer

XRD (X-Ray Diffractometer): Model D8 Advance (Bruker)
S.No. Features
1. Computer controlled diffractometer used for automatic operation
2. High temperature (RT-1600 : C) in-situ XRD facility
J. Angle range: 51-1201
4. Target ( Cu, Mo, Fe)
5. Scintillation counter detector
6 Lynex eye detector controller
7. Accelerating voltage up to 40kV
8. Electron probe current range up to 40 mA
9. Sample spinners
10. Nine samples can be simultaneously placed in sample holder
11. Leptos and EVA software provide a common look and feel for measurement and
analysis software.
52
Figure 4.5: Setup of X-Ray Diffractometer, (Source: IIT, Roorkee, India)
4.1.6 Differential thermal analysis (DTA) of developed agglomerated fluxes

The Thermo Gravimetric/ Differential Thermal Analyzer (TG/DTA) combine the flexibility
feature of DTA with the proven capabilities of the TG measurement technology providing
property information for a variety of samples. Thus simultaneous TG/DTA system can be
used for such applications as oxidation, heat resistance, compositional analysis and
measurement of ash contents in a sample. This system is also used to meet such needs as
reaction rate and accelerated degrading tests. The features of this machine are specified in
Table 4.3.
The differential thermal analysis (DTA) of developed agglomerated fluxes (AGFl 101-
AGFl 132) DTA was carried out in an AI2O3 crucible from 28 to 1400 "C at a heating rate of
10 C/min as shown in Figure 4.6. Alumina was used as a standard reference for differential
thermal analysis.
53
Table 4.3: Features of TG-DTA (Thermal Analyzer)
TG-DTA (Thermal Analyzer): Model SII 6300 EXSTAR
S.No. Features
1. The TG/DTA can operate in either DSC or DTA mode. In DSC mode it displays heat
flow signal.
2. The horizontal differential balance provides high sensitivity and accuracy.
3. Gas control unit automatically controls the environment of furnace between
measurements.
4. Auto sampler with 30 samples positions provides laboratory automation and
increases productivity.
till 'ilfeii
1 UJ.
> I'.fVi
Figure 4.6: Setup of TG-DTA (Source: IIT, Roorkee, India)
4.2 Development of Weld Joints

The objective of developing the weld joints was to measure different characteristics of weld
metal like tensile strength, elongation, reduction in area, impact strength, hardness and weld
microstructures by using thirty two newly developed agglomerated fluxes. The details of
54
material, equipment, parameters used for the development of weld joints and the
characteristics of weld joints are discussed in the following section.
4.2.1 Material and equipment

Experiments were conducted on low carbon steel plates of 300X300X20 mm size (V-groove
at 45 ) [80] as shown in Figure 4.7. The chemical composition of base plate and welding
electrode is given in Table 4.4. A Submerged Arc Welding Machine (Make: Essar, welding
lab of NSIT, Delhi, India) was used to deposit weld at V-groove and it is shown in Figure
4.8. The specifications of the SAW Machine are given in Table 4.5.
Figure 4.7: Welding Specimen [80]
Table 4.4: Chemical Composition of Welding Electrode and Base Plate
Composition of electrode and base plate C Mn Si Ni Cr

(wt %) (wt %) (wt %) (wt %) (wt %)
Electrode (Auto Rod 12.08L 3.15mm) 0.08 0.5 0.05 0.33 0.83
EL-8 (ESAB)
Low carbon steel plate 0.13 0.5 0.2 0.01 0.02
55
Mild steel
Electrode
Flux Hopper Control Panel
Electrode Holder
Granular Flux Low carbon Steel

Plate
Figure 4.8: Submerged Arc Welding Machine
Table 4.5: Welding Unit Specifications
Travel Speed
Voltage (V) Current (A)
CPRA 800 (Make: Essar) (cm/min)
0-100 0-1500 0-250
4.2.2 Welding procedure

In the present study, multipass butt welds as shown in Figure 4.9 were obtained by
employing automatic submerged arc welding process by using thirty two developed
agglomerated fluxes and EL-8 electrode wire of 3.15 mm diameter on low carbon steel plates
with fixed welding parameters. The fixed welding parameters are selected on the basis of
good slag detachability, weld appearance, arc inifiation, and arc stability. Each pass was
preceded by thorough cleaning. In this way, total thirty two plates were welded as shown in
Figure 4.10. The fixed welding parameters are given in Table 4.6.
56
Welding Runs
Figure 4. 9: Correct Way of Bead Deposition
Figure 4.10: Welded Low Carbon Steel Plates
Table 4,6: Fixed Welding Parameters
Parameter Value
Arc voltage 36 V
Welding current 500 A
Speed 28 cm/min.
Nozzle to plate distance 25 mm
Heat input 3.87kJ/mm
Polarity DCEP
57
4.3 Characteristics of Weld Joints
The weld joints obtained by using developed agglomerated fluxes were assessed through
different characteristics. Before preparing the specimens, 20 mm lengths of weld joints was
discarded from both sides — weld start point and weld finish point. Backing plate was
removed with the help of lathe machine. The various characteristics are described in the
following sections.
4.3.1 Tensile strength

Specimens in the form of round bars of 5.06 mm diameter and 17.9 mm gauge length
prepared on lathe machine were used to evaluate the tensile properties. The specimens for
measuring tensile strength, yield strength and elongation were prepared from the welded
joints as per the standard specification shown in Figure 4.11. Tensile strength test was
performed on low carbon steel and weld joint using tensometer as per the specimen
orientation and size as per standard shown in Figure 4.12. These properties of prepared
samples were measured with the help of tensometer (KIPL-PC 2000) having maximum limit
of 18 kN load and a uniform test speed of 0.8 mm/min as shown in Figure 4.13.
)S5D6
179
-M—H H
All Dimensions in mm
Figure 4.11: Dimensions of Specimen for Tensile Test [80]
58
Welding Direction
"L3- 6- Weld Meta
tt Base Metal Ssecimen
Figure 4.12: Orientation of Specimen from the Weld Plate
Computer
Reading Indicator
Specimen Clamping De\ ice
Figure 4.13: Setup for Tensile Strength
4.3.2 Impact strength

Standard charpy-V-notch specimens in the form of square bar of 10 mm side and 55 mm
length were used to evaluate the toughness of weldments. Five test specimens were taken
from each joint with proper spacing. Groove of 45 ^ angle was made accurately in weld
zone. The specimens of charpy test are shown in Figures 4.14 and 4.15.
59
For impact test, all the specimens were cleaned thoroughly and their surfaces polished
using 600 grit emery paper. The specimens were positioned properly on impact testing
machine of maximum impact energy of pendulum (300 Joules) as shown in Figure 4.16,
and load was released suddenly behind the notch and recorded on a scale having a least
count of 0.2 kgm. From five values, the lowest and the highest ones were discarded and
the other three values were averaged for each type of flux.
10mm
SSrnrTi
\* *\
10mm
1* Weld Metal
\
Figure 4.14: Dimensions of Charpy Impact Test Specimen
Figure 4.15: Prepared Weld Metal Charpy Impact Test Specimen
60
Dial
Pendulum Hammer
Specimen Positioning
Figure 4.16: Impact Testing Machine
4.3.3 Hardness test

The Vickers hardness of the welded joints was measured on Vickers Hardness Testing
machine as shown in Figure 4.17, by applying 30 kgf load for 10 seconds on the prepared
specimens and taking average of three readings at different places on the same specimen.
61
Optical Scale
An\il
yv-^3\^
Figure 4.17: Hardness Testing Machine
4.3.4 Weld joint microstructure

The samples for microstructure study arc cut from the weld joints which arc made with the
help of developed agglomerated fluxes. The scratches on samples were removed with the
help of different grades of emery papers. After that, the samples were polished on polishing
paper of grit 400, 600 and 800. The final stage of polishing is done on polishing machine
shown in Figure 4.18 on velvet cloth with a combination of water and alumina powder for
smooth and fine finish.
The samples were etched after polishing to reveal the microstructure clearly; the acids in the
etchant attack the grain boundaries and give a clear image of the size of the grains. The
samples were dipped in the solution of 2 to 5% HNO3 and 95 to 98% CH3OH for etching and
62
then washed in distilled water. The etched samples are washed thoroughly to remove the
carbon deposits and then dried to study the microstructure under the optical microscope.
The dried weld joints samples were analyzed at 400X magnification using Trinocular
Metallurgical Microscope with CCD camera shown in Figure 4.19.
Figure 4.18: Set up of Polishing Machine
63
CCD Camera
Microstructure Microscope
Display Monitor E
Figure 4.19: Set up of Metallurgical Microscope
4.4 Bead Morphology and Shape Relationship

The samples for bead morphology and shape relationship study are cut in a cross-sectional
way at the mid length of the beads shown in Figure 4.20 which were laid on low carbon steel
plate with the help of developed agglomerated fluxes. The same procedure as explained in
Section 4.3.4 was followed to remove the scratches from the samples and do polishing. The
samples were then etched by the same procedure. The etched samples are washed thoroughly
to remove the carbon deposits and dried to study the weld bead profile using the optical
microscope.
64
Weld Beads
Figure 4,20: Weld Beads on Low Carbon Steel Plate
4.5 Element Transfer Study

To measure the actual concentration of various elements in weld metal, the weld bead was
drilled at four different locations 1, 2, 3 and 4 as shown in Figure 4.21 up to a depth of 1mm
so as to minimize the effect of dilution. The drilled chips are mixed thoroughly and used as
samples for chemical analysis. The sample of 1 g is dissolved into 100 ml of cone. HCl. 10 ml
of this solution is diluted to 100 ml by adding double distilled water in it. This solution is then
tested by an Atomic Absorption Spectrophotometer against the standard solutions to calculate
the wt % of C, Si, Mn, P, S, Cr and Ni elements present in the weld metal.
65
Drilled Holes
^ 3iq?vjpu MO-II
Figure 4.21: Specimen for Element Transfer Study
66
CHAPTERS
RESULTS AND DISCUSSION
5.1 Introduction
The chapter on results and discussion describes the physical properties of fluxes - bulk
density, flowability, grain size and viscosity. The X-ray diffraction (XRD) and
Differential thermal analysis (DTA) study are used to identify different crystalline
structures and endothermic peaks at different temperatures in developed agglomerated
fluxes.
The results of different characteristics of weld joint including tensile strength, impact
strength, elongation, hardness, microstructures, bead morphology and element transfer are
also discussed in this chapter.
5.2 Physical Properties of Fluxes

The physical properties of fluxes are the important characteristics to understand how
submerged arc flux composition controls welding behaviour through these properties for the
developed agglomerated fluxes.
5.2.1 Flowability
The flowability of developed agglomerated fluxes is correlated with the help of repose angle.
The repose angle of developed fluxes was calculated by the procedure given in Section
4.1.1.The average value of angle of repose of each developed flux is compared with the
standard value of angle of repose as reported in Table 4.1. As far as the flow properties of the
fluxes are concerned, twenty eight fluxes out of thirty two are good, one excellent and three
satisfactory. The average values of angle of repose along with the flow properties are listed in
Table 5.1
67
Table 5.1: Angle of Repose for Developed Agglomerated Fluxes and their Flow
Properties
Angle of repose (8) in
degrees Flow
S.No. Flux Code Properties
Oi 82 Average 8
1 AGFllOl 29.16 29.16 29.16 Good
2 AGF1102 28.32 29.37 28.85 Good
3 AGF1103 26.77 26.91 26.84 Good
4 AGE 1104 29.04 30.22 29.63 Good
5 AGF1105 28.04 28.07 28.06 Good
6 AGF1106 25.6 25.7 25.6 Good
7 AGF1107 22.6 23 22.6 Excellent
8 AGF1108 27.93 27.25 27.59 Good
9 AGF1109 26.96 26.7 26.83 Good
10 AGFlllO 30 29.7 30 Satisfactory
11 AGFllll 29.74 28.6 29.17 Good
12 AGF1112 29.02 29.59 29.31 Good
13 AGF1113 29.59 30.61 30.16 Satisfactory
14 AGF1114 28.74 29.35 29.05 Good
15 AGF1115 27.7 27.16 27.43 Good
16 AGF1116 28.81 28.14 28.48 Good
17 AGF1117 27.83 28.66 28.25 Good
18 AGF1118 26.3 26.43 26.37 Good
19 AGF1119 26.17 27.17 26.6 Good
20 AGF1120 27.5 30.3 28.9 Good
21 AGF1121 30.3 27.21 28.7 Good
22 AGF1122 30.1 30.8 30.4 Satisfactory
23 AGF1123 27.17 28.3 27.7 Good
24 AGF1124 25.58 27.65 26.62 Good
25 AGF1125 28.9 28.9 28.9 Good
26 AGF1126 26.17 25.75 25.96 Good
27 AGF1127 29.2 29.56 29.3 Good
28 AGF1128 29.1 29.1 29.1 Good
29 AGF1129 29 29 29 Good
30 AGF1130 29.3 29.2 29.2 Good
31 AGF1131 29.1 29.1 29.1 Good
32 AGF1132 29 29 29 Good
68
5.2.2 Grain size
The grain size of all the thirty two developed agglomerated fluxes is determined with the
procedure outlined in Section 4.1.2. The grain size of these fluxes ranges from 0.355 to 1.18
mm in size, which is a very good range for achieving good quality of welded surface. The
sieving results are given in Tables A.l- A.32 (APPENDIX A), and histograms for different
developed fluxes are shown in Figures B.l - B.32 (APPENDIX B).
5.2.3 Bulk density

The bulk density of developed agglomerated fluxes is calculated with the same procedure as
given in Section 4.1.3. The weld bead penetrations and flux consumption are determined with
the procedure discussed in Section 4.4. The values of bulk density and flux consumption are
given in Table 5.2.
Table 5.2: Values of Bulk Density and Flux Consumption
Bulk Flux
Density Consumption
(g/cc) (g)
Flux MnO CaFi MgO NiO Fe-Cr
Code (wt%) (wt%) (wt%) (wt%) (wt%)
AGFUOl 6.25 12.5 7.5 2.5 7.5 0.83 160
AGF1102 8.75 12.5 7.5 2.5 2.5 0.85 170
AGF1103 6.25 17.5 7.5 2.5 2.5 0.81 155
AGF1104 8.75 17.5 7.5 2.5 7.5 0.84 165
AGF1105 6.25 12.5 12.5 2.5 2.5 0.76 110
AGF1106 8.75 12.5 12.5 2.5 7.5 0.76 115
AGF1107 6.25 17.5 12.5 2.5 7.5 0.84 165
AGF1108 8.75 17.5 12.5 2.5 2.5 0.88 195
AGF1109 6.25 12.5 7.5 7.5 2.5 0.86 180
AGFlllO 8.75 12.5 7.5 7.5 7.5 0.78 140
AGFllll 6.25 17.5 7.5 7.5 7.5 0.79 150
AGF1112 8.75 17.5 7.5 7.5 2.5 0.64 50
AGF1113 6.25 12.5 12.5 7.5 7.5 0.77 130
AGF1114 8.75 12.5 12.5 7.5 2.5 0.9 200
AGF1115 6.25 17.5 12.5 7.5 2.5 0.76 115
69 Contd.
AGF1116 8.75 17.5 12.5 7.5 7.5 0.8 150
AGF1117 5 15 10 5 5 0.8 150
AGF1118 10 15 10 5 5 0.82 155
AGF1119 7.5 10 10 5 5 0.8 150
AGF1120 7.5 20 10 5 5 0.81 150
AGF1121 7.5 15 5 5 5 0.84 165
AGF1122 7.5 15 15 5 5 0.87 190
AGF1123 7.5 15 10 0 5 0.8 150
AGF1124 7.5 15 10 10 5 0.77 125
AGF1125 7.5 15 10 5 0 0.76 110
AGF1126 7.5 15 10 5 10 0.76 115
AGF1127 7.5 15 10 5 5 0.7 90
AGF1128 7.5 15 10 5 5 0.74 100
AGF1129 7.5 15 10 5 5 0.74 100
AGFU30 7.5 15 10 5 5 0.73 95
AGF1131 7.5 15 10 5 5 0.75 105
AGF1132 7.5 15 10 5 5 0.76 110
The bulk densities of developed agglomerated fluxes range from 0.64 to 0.9 g/cc. In this
section the bulk densities and weld bead penetration are correlated with the flux
consumption. The flux consumption increases as bulk densities of developed agglomerated
fluxes increase. The correlation of bulk densities with flux consumption is shown in
Figure 5.1.
70
200
180
'GO
£3 160
o • m'
•-•'•
^ 140
3
ao
120
X
100
80
60
40
0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00

Bulk Density (g/cc)
Figure 5.1: Effect of Bulk Density on Flux Consumption
5.2.4 Viscosity
The viscosity of submerged arc welding fluxes is an important property, which justifies the
weld bead penetration behaviour of the arc at a particular point. The viscosity of developed
agglomerated flux AGFllOl was calculated with the same procedure as specified in the
Section 4.1.4. The trend of viscosity of this flux with respect to temperature is shown in
Figure 5.2. It is clear from the graph that the viscosity first decreases sharply with the
increase in temperature fi-om 1100 DC to 1200 DC and then decreases at a slow pace as
temperature increases to 1400 DC. Finally, the viscosity becomes stable at about 1450 DC.
71
60 n
50-
CO
O
40-
O
o 30-
20
10
0-
1100 1200 1300 1400 1500

Temp, in degree centigrade
Figure 5.2: Viscosity Trend of Flux (AGFllOl)
5.2.5 X-Ray Diffrcation of developed agglomerated fluxes

A structural and chemical analysis of crystalline phases of developed agglomerated fluxes
was carried out by X-ray diffraction setup shown in Figure 4.5, Section 4.1.5. In the present
study, developed agglomerated fluxes are analyzed by X-ray diffraction (XRD) to know the
different types of oxides formed and change in oxidation number of metallic centers after
sintering process at around 850 °C. This study reveals the quantity of ions present
(percentage) and melting behaviour of developed agglomerated fluxes for submerged arc
welding process. The important chemical reactions take place during heating of ceramic
oxide which include chemical reaction among the different phases [81], phase transformation
of minerals [82], formation of crystalline phases [83], and chemical and structural
characterization of crystalline phases in agglomerated fluxes for submerged arc welding [58].
72
The X-ray diffraction (XRD) identified crystalline phases as well as quantified the different
types of ions and their distribution in developed agglomerated fluxes (AGFl 101-AGFl 132).
The XRD of thesefluxesare shown in Figures C.l - C.32 (APPENDIX C).
In the XRD results of agglomerated fluxes, it has been observed that a number of crystalline
phases are formed due to chemical interaction among ions with potassium silicate binder.
The oxides like MnO, MgO, SiOi, AI2O3, and CaO were found in compound form, indicating
that they do not react with others. The quantity of different oxides reported in Table 5.3 was
based on XRD peaks with high intensity [84]. The names of different crystalline phases
observed in XRD analysis along with their crystal structures are given in Table 5.4. The
development of different types of silicates and oxides are mainly due to mixing of potassium
73
fL
ti S ^J p
1 s -g :S § i i i :i .§ g t/4
i s e i i g i § § AGFllOl
g § s £ s g AGF1102
• i != £ 1 S i AGF1103'
• s § s 1 § 5 s AGF1104
g • • i £ 5 5 AGFH05
1 g us g g g £ AGFH06
s £ g g £ 5 § AGF1107
•
s C3
s i § 5 5 AGF1108
§ • • • g 1 g § AGF1109
^ § 5
.Jib
s s 5 E £ AGFlllO
JO
s 3
•
i
•
S
2
S
3
1
1
AGFim
AGF1112 1
S te
• 5
i
•
g i 5
1 1 i
AGF1H3
AGF1114
1
t
•
s
oo
i
s
•
g
§
5
i §
1 s
5 1 s
AGF1115
AGF1116
AGF1117
1
• ' 1 5 1 i s AGF1118
s S 1 g § AGF1119
r
g •
•
oo
io
i
i
£
g
1 i £
S s i
§ § 5
AGF1120
AGF1121
ACF1122
11
1 • s
oo
1 s g AGF1123
s • •
1 1 S AGF1124
io g S S i i AGF1125
g g § g i § s AGF1126
5 g S i g s s AGF1127
s i 1 5 g AGF1X28
s St • 1 i i s AG F 1 1 2 9
i g g AGF1130
i g i ' AGF1131
• g g s 1 AGF1132
saxnji pajBjainoi§§vpadopAaQ Joj JM»/„ m sapixQJoXipncnf) :£s aiq^x

Table 5.4: Crystalline Phases and Crystal Structure in Developed Agglomerated
Fluxes
Flux Crystal Structure

Crystalline Phase
Code
AGFHOl Hexagonal
CaAl2 SiOg (Dmisteinbergite)
AGF1125
AGFllOl Mg7 Fey Sif^ O22 (OH)2 (Protoanthrophyllite) Orthorhombic
AGFllOl Triclinic
CaAl2SiQ0^^.5(H2O) (Epistilbite)
AGF1104
AGFllOl Mg2Al2SUOô(OH) .HH2O) (Palygorskite) Monoclinic
Ca2Fe^^^(Fe^^^ Al)2 Monoclinic
AGFllOl
(Si04)(Si207)(0,0H)2. H2O (Julgoldite)
AGF1102 CaAl2 Si2 OQ (Svyatoslavite) Monoclinic
AGF1102 K2CaMg2(Al,Si)36072,28H20(Mazzite) Hexagonal
AGF1102 Monoclinic
AGF1103
AGF1104 Mg2(Al,Fe^)3Si3A10io(OH)8(Sudoite)
AGF1127
AGF1129
AGF1102 Monoclinic
AGF1118 MriySis 012(0^)2 (Leucophoenicite)
AGF1127
AGF1102 Ca3Al2(Si04)3-x(OH)4x(x= 0.2 - 1.5) (Hibschite) Isometric
AGF1104 KAl2(Si3Al)Oio(OH,F)2 (Muscovite) Monoclinic
AGF1104 Triclinic
CttzCMn, Fe++)Fe+-*-+Si50i4(OH)
AGF1107
(Manganbanbingtonite)
AGF1108
AGF1103 Orthorhombic
AGFlllO CaioMg4y4/25iii039.4(//20) (Juanite)
AGFllll
AGF1105 Monoclinic
CasSi207(C0^)2 (Tilleyite)
AGF1109
AGF1104 Monoclinic
AGF1108 Mg9(Si04)4(OH,F)2 (Hydroxylclinotumite)
AGF1120
Ca7(Si04)3(OH)2 (Chegemite)
AGF1124
AGF1112 CaAlSiOîOH) (Vuagnatite) Orthorhombic
AGFlllO Triclinic
Cai4 (Mn^),4Si24058(OH)8.2(H20) (Truscottite)
AGF1115
AGFlllO Ca0.3(Fe^^MgFe^^)3Si4Al40io(OH)2.4(H20) Monoclinic
(Ferrosaponite)
AGFlllO Ca^Si20e,C03(0H,F)2 (Fukalite) Orthorhombic
75 Contd.
AGFllll Orthorhombic
Mg2Si206 (Enstatite)
AGF1120
AGF1113 Triclinic
AGF1119 Ca(Fe'^)4 (S04)6 (OH)2.19(H20) (Calciocopaipite)
AGF1124
AGF1105 Monoclinic
AGFllll
AGF1113
AGF1117
AGF1119 M55(Fe+++)5 (5104)2 (0//,F)2 (Chondrodite)
AGF1121
AGF1122
AGF1123
AGF1125
AGF1114 CaAl2(iAl2Si2)0-iôiOH)2 (Margarite) Monoclinic
AGF1114 Cac,H2Si(,0x8- 6H2O (Plombierite) Monoclinic
AGF1115 Ca4(Mn^)3Si802o(OH)6.2(H20)(Orlymanite) Trigonal
AGF1116
(Fe^)2Mg2Si206 (Clinoferrosilite)
AGF1131
AGF1115
AGF1118
(Mn^'')2Si03(OH)2.(H20)(Ncahwaningite)
AGF1120
AGF1121
AGF1131
AGF1117 (Fe'^)2Mg2Al4Si50i8 (Sekaninaite) Orthorhombic
AGF1117 Triclinic
CaAl2Si20Q (Anorthite)
AGF1121
AGF1117 Monoclinic
(Mg, Fe^^Ca) (Mg, Fe^) Si206 (Pigeonite)
AGF1126
AGF1118 Ca2Al2SiOe(OH)2 (Kamaishilite) Tetragonal
AGF1119 K2(Fe"^,Fe"^"^"^)Si802o(OH)4 (Ferroceladonite) Monoclinic
AGF1120 Ca2Al[(0H)6AlSi02-3(0H)4_3]. 2,5H2O (Stratlingite) Trigonal
AGF1122 CaAl2Si208 (Dmsiteinbergite) Hexagonal
AGF1130 CasSieOxs(OH)2- 2{H20) (Riversideite)
AGF1132
AGF1123 K5Ca8(Si60i5)2 (Si207)Si409(OH). 3(H20) (Charoite) Monoclinic
AGF1126 [Cai4(OH)8] (Si802o)(Si802o)[(Fe^)9(OH)i4] (Tungusite) Triclinic
AGF1127 MgC03.SH20 (Lansfordite) Monoclinic
AGF1127 Monoclinic
AGF1128 Ca2[Mg4(Al, Fe+++)]Si7A10220H2 (Magnesiohomblende)
AGF1129
76
5.2.6 Differential thermal analysis of developed agglomerated fluxes
The differential thermal analysis (DTA) of developed agglomeratedfluxeswas carried out by

the setup shown in Figure 4.6, Section 4.1.6. The DTA curves are shown in Figures D.l -
D.32 (APPENDIX D).
The curves of differential thermal analysis (DTA) show that the most of the oxides were
stable oxides upto 1000 °C. The ranges of DTA curves were foimd between! 000 to 1400 °C.
With few exceptions, mostly reactions were endothermic in nature. The endothermic peaks
were observed at different temperatures for different fluxes as given in Table 5.5. The
endothermic peaks were different for different fluxes because of their different chemical
compositions. These endothermic peaks were comparable with the melting reaction of
different crystalline phases for different fluxes. In some DTA curves zig-zag behaviour was
observed due to gassing or blistering of the glass formation [85]. The anions generally
"^ -t ^^ O-l- "5-4- "X-i- -m—. O I
present in the fluxes are 0 " and F ' . The cations from the flux like Ca , Mg , Al ' Fe ,
Fe^"^' Si'*^ jMn^"^, Mn'^^, Mn'*"^, can react with oxygen forming tetrahedral unit Si04''' of the
silicates in the fluxes. Since calcium and magnesium ions have largest negative heat of
formation energy AGf, they thus react quickly with oxygen in the welding arc. There were
changes in oxidation number in some cationic constituents like Fe^^, Fe ^ and Mn "^, Mn "^.
In most of the fluxes calcium was present which increases the stability of electric arc during
welding. The viscosity offluxesincreases by the presence of corundum and Si02 which were
present in most of the fluxes. In most of the fluxes it has been observed the MnO and SiOi
are jointly present, which improve the thermo-mechanical behaviour of weld.
77
Table 5.5: DTA Analysis for Endothermic Temperatures
Flux Peaks of Endothermic Flux Peaks of Endothermic

Code Temperature (*C) Code Temperature (^C)
AGFllOl 1239.89 AGF1117 1235.58
AGF1102 1234.25 AGF1118 1224.33
AGF1103 1233.58 AGF1119 1224.99
AGF1104 1240.77 AGF1120 1338.85
AGF1105 1226.99 AGF1121 1246.58
AGF1106 1334.85 AGF1122 1189.52
AGF1107 1244.58 AGF1123 1181.28
AGF1108 1191.77 AGF1124 1231.71
AGF1109 1245.84 AGF1125 1282.26
AGFlllO 1248.37 AGF1126 1247.63
AGFllll 1334.85 AGF1127 1242.18
AGF1112 1166.88 AGF1128 1242.18
AGF1113 1229.25 AGF1129 1241.18
AGF1114 1236.77 AGF1130 1240.18
AGF1115 1241.88 AGF1131 1236.18
AGFni6 1227.33 AGF1132 1242.18
5.3 Characteristics of Weld Joints

The weld joints of developed agglomerated fluxes are produced by the procedure given in
Section 4.2. The characteristics of these weld joints and their resuhs are described in this
section. The experimental values of tensile strength, yield strength, percentage elongation,
impact strength and hardness of base metal and weld metal are given in Tables 5.6 and 5.7
respectively.
Table 5.6: Tensile, Impact and Hardness Properties of the Base Metal
Tensile Properties Impact Vickers

Strength at Hardness
Tensile Elongation room (VHN)
Strength % temperature
(MPa)
561 25.8 28 140
78
Table 5.7: Tensile, Impact and Hardness Properties of the Weld Metal for
CaO - Si02 - AI2O3 Based Flux System
Flux Tensile Pro jerties Impact Strength at Vickers

Code Tensile Elongation room temperature Hardness
Strength (MPa) (%) (VHN)
AGFllOl 560 28.56 80 168

AGF1102 570 29.07 65 168
AGF1103 525 26.77 70 157
AGE 1104 530 27.03 95 157
AGF1105 520 26.52 85 159
AGF1106 530 27.03 68 160
AGF1107 575 29.32 75 172
AGF1108 580 29.58 80 174
AGF1109 535 27.28 75 160
AGFlllO 570 29.1 70 171
AGFllll 535 27.28 60 160
AGF1112 540 27.54 85 162
AGF1113 544 27.77 65 163
AGF1114 550 28.07 80 165
AGF1115 545 27.79 99 163
AGF1116 570 29.07 98 171
AGF1117 560 28.56 80 168
AGF1118 571 29.12 95 171
AGF1119 540 27.54 65 162
AGF1120 545 27.79 85 163
AGF1121 540 27.54 70 162
AGF1122 545 27.79 90 163
AGF1123 540 27.54 80 162
AGF1124 560 28.56 85 168
AGF1125 540 27.54 80 162
AGF1126 550 28.05 75 165
AGF1127 545 27.79 82 163
79
Contd.
AGF1128 547 27.89 87 164
AGF1129 546 27.84 88 163
AGF1130 542 27.64 90 162
AGF1131 539 27.48 92 161
AGF1132 540 27.54 90 162
5.3.1 Modeling of weld joint characteristics
The modeling plays a crucial role in the welding industries because it predicts the
approximate properties of weld joint before the welding which saves time, lowers cost and
provides comfort to the human being. In this section, modeling of different characteristics
like tensile strength, elongation, impact strength and hardness reported in Table 5.7 are
described.
5.3.1.1 Modeling of tensile strength

The Analysis of variance (ANOVA) without pooling and pooled ANOVA reported in Tables
E.l (APPENDIX E) and 5.8 respectively show that the model is significant. The f-value of
25.24 and the p-value < 0.0001 did, in fact, demonstrate that this regression was statistically
significant at 99% confidence level. Besides, the fit of the model is checked by the
coefficient of determination R^ which is found to be 0.9117 thus pointing out that only
8.83% of the total variation in the response is not explained by the model. The difference of
predicted R^ and adjusted R^ is less than 0.20 and "Adeq Precision" is greater than 4; this
means that regression model provides an excellent explanation of the relationship between
the agglomerated alloying constituents and the response tensile strength. The associated p-
value for the model lower than 0.05 (i.e. a=0.05, or 95% confidence level) indicates that the
model is statistically significant at 95% confidence level. The lack-of-fit term is not
significant which further validates that the model is significant. Factors MnO, MgO, Fe-Cr,
interaction effect of factor MnO with factor Fe-Cr, interaction effect of factor CaF2with
factor MgO, interaction effect of factor NiO with factor Fe-Cr and second order term of
factor MnO , have significant effect on tensile strength.
After eliminating the non-significant terms, the final response equation for tensile strength is
given as:
80
Tensile Strength (MPa) = +936.17750-33.08933* MnO-22.58167 * CaF2-33.61667 * MgO
-1.81667* NiO +18.06500* Fe-Cr - 2.58600* MnO* Fe-Cr
+2.30700 * CaF2* MgO +0.49900* NiO * Fe-Cr +3.34240* Mn02

(5.1)
Table 5.8: Pooled ANOVA of Tensile Strength
Degree of Mean
Source Sum of Squares Freedom Squares f-value Prob>f
Model 6435.58 9 715.06 < 0.0001 significant
A-MnO 635.51 635.51 22.44 < 0.0001
B-CaF2 35.77 35.77 1.26 0.2732
C-MgO 146.52 146.52 5.17 0.0330
D-NiO 69.02 69.02 2.44 0.1328
E-Fe-Cr 203.58 203.58 7.19 0.0136
AE 1044.91 1044.91 36.89 < 0.0001
BC 3326.41 3326.41 117.44 < 0.0001
DE 155.63 155.63 5.49 0.0285
A2 818.24 818.24 28.89 < 0.0001
Residual 623.15 22 28.33
Lack of Fit 568.32 17 33.43 3.05 0.1109 not significant
Pure Error 54.83 5 10.97
Cor Total 7058.73 31
Std. Dev. = 5.32 R-Squared = 0.9117
Mean = 547.83 Adj R-Squared = 0.8756
C.V.% = 0.97 Pred R-Squared =0.7706
PRESS = 1619.62 Adeq Precision= 20.245
Figure 5.3 displays the normal probability plot of the residuals for tensile strength. The
residuals are falling on a straight line, which means that the errors are normally distributed.
Further, each observed value is compared with the predicted value calculated from the model
as shown in Figure 5.4. It can be seen that the regression model is fairly well fitted with the
observed values.
81
Normal Plot of Residuals
-1.00 0.00
Internally Studentized Residuals
Figure 5.3: Normal Probability Plot of the Residuals for Tensile Strength
Predicted vs. Actual

58000 —
STOOD —
a
^ o
S60.C0-~
^ ^ B
5S0.0D —
5«.00 —
E
530.C0 —
520 OO—
• ^
510.00 —
1 1
1 1 1 1
Figure 5.4: Predicted vs. Actual Value of Tensile Strength
5.3.1.1.1 Effect of flux constituents on tensile strength

Effects of flux constituents on tensile strength are shown.in Figures 5.5 - 5.8. The 3-D
graphs depict that the tensile strength increases as MnO, MgO, CaF2 increase but the
increase in value of tensile strength is considerably more with increase in CaFi as compared
to the increase in MnO and MgO. The effects of NiO and Fe-Cr on tensile strength are
considerably less as compared to the constituents - MnO, MgO andCaF2. The tensile
strength in case of submerged arc welding is justified by the oxygen contents mainly
available in the weld metal. An attempt was made to lower the oxygen and inclusion contents
by modifying a standard flux developed by different researchers. The lower oxygen contents
82
are favorable to the formation of microstructure such as acicular ferrite which uhimately
improves tensile strength [60]. CaF2, which lowers the oxygen contents in the weld metal, is
the main cause for the improvement of tensile strength [16]. The weld metal analysis for
oxygen described by Chai and Eager [41] shows that the level of oxygen increases as MnO
content increases as compared to MgO content which lowers the tensile strength. The NiO
and Fe-Cr do not affect the tensile strength considerably.
IBI
a
u
X/1
d
H
B:CaF2
A: MnO
Figure 5.5: Effect of CaFz and MnO on Tensile Strength
C:MgO A: MnO
Figure 5.6: Effect of MgO and MnO on Tensile Strength
83
0^
I
CZ5
a
H
D:NiO A: MnO
Figure 5.7: Effect of NiO and MnO on Tensile Strength
ISA
e
i.
la
H us
A: MnO
E: Fe-Cr
za «z
Figure 5.8: Effect of Fe-Cr and MnO on Tensile Strength
84
5.3.1.2 Modeling of percentage elongation
The Analyses of Variance (ANOVA) without pooling and pooled ANOVA have been
reported in Tables E.2 (APPENDIX E) and 5.9 respectively and the model was found
significant. The f-value of 25.18 in pooled ANOVA and the p-value < 0.0001 demonstrated
that this regression was statistically significant at 99% confidence level. The value of R^ was
found to be 0.9115, and fulfilled the requirement of ANOVA. R^ must lie between zero and
one, the larger values in this range being more desirable [86]. The values of Probability > f
for the model less than 0.05 (i.e. a=0.05) indicate that the model terms are significant. The
lack-of-fit term is not significant in the model. In this study, the factors MnO, MgO, Fe-Cr,
combined effect of constituent MnO with constituent Fe-Cr, combined effect of constituent
CaF2 with constituent MgO, combined effect of constituent NiO with constituent Fe-Cr and
second order term of constituent MnO, have significant effect.
The mathematical model that describes the final response equation for elongation is given in
Equation .5.2.
Percentage elongation (%) = +47.74258-1.69053* MnO-1.15117* CaF2-1.71367* MgO

- 0.092667 * NiO+0.92233* Fe-Cr-0.13200 * MnO * Fe-Cr
+0.11760 * CaF2 * MgO+0.025400 * NiO * Fe-Cr
+0.17075*MnO2
(5.2)
85
Table 5.9: Pooled ANOVA of Percentage Elongation
Degree
Sum of Mean
Source of f-value Probability >f
Squares Squares
Freedom
Model 16.75 9 1.86 25.18 < 0.0001 significant
A-MnO 1.66 1.66 22.52 < 0.0001 significant
B-CaF2 0.093 0.093 1.25 0.2753not significant
C-MgO 0.38 0.38 5.14 0.0335 significant
D-NiO 0.18 0.18 2.39 0.1362 not significant
E-Fe-Cr 0.53 0.53 7.15 0.0139 significant
AE 2.72 2.72 36.84 < 0.0001 significant
BC 8.64 8.64 116.97 < 0.0001 significant
DE 0.4 0.4 5.46 0.0290 significant
A^ 2.14 2.14 28.9 < 0.0001 significant
Residual 1.63 22 0.074
Pure Error 0.14 5 0.028
Cor Total 18.37 31
C.V. % = 0.97 PredR-Squared =0.7701
PRESS = 4.22 Adeq Precision =20.225
Figure 5.9 describes the normal probability plot of the residuals for percentage elongation.
The residuals falling on straight line indicate that the errors are normally distributed. Figure
5.10 shows the values of percentage elongation predicted by the model and actual values
obtained by the experiment. It is observed that the regression model is fairly well fitted with
the observed values.
86
95
>. 90
—
s01 80
•o 70
ot _
Q.
50
^
ra 30
E
o
20
z 10
5
Internally Studentlzed Residuals
Figure 5.9: Normal Probability Plot of The Residuals For Percentage

Elongation
29.00 •
Actual
Figure 5.10: Predicted vs. Actual Percentage Elongation Value
87
5.3.1.2.1 Effect of flux constituents on percentage elongation
Figure 5.11 depicts the response surface for percentage elongation in relation to the flux
constituents CaFz and MnO. In this figure, the percentage elongation marginally increases
with increasing percentage of CaF2 constituents because it lowers the high oxygen in
submerged arc welds. The value of percentage elongation also increases with the increasing
percentage of MnO constituent because it is a transition metal oxide.
a
o
ISJ
a
o
a>
• * *
a
u
w
A: MnO B: CaF2
Figure 5.11: Effect of CaF2 and MnO on Percentage Elongation
The effect of CaF2 and MgO constituents on percentage elongation is shown in Fig 5.12. The
response surface shows that the value of percentage elongation increases with CaF2
constituent; the reason is explained earlier. The value of percentage elongation also increases
with the increasing percentage of MgO, because it is very stable (non-reactive) in the
presence of CaF2.
88
a
o
•MM
-^
a
o
fi
u
V
PH
»A
C:MgO 1.9 i z a B: CaF2
Figure 5.12: Effect of MgO and CaFi on Percentage Elongation

Figure 5.13 shows the response surface of behaviour of CaFa and NiO constituents. The
effect of NiO on percentage elongation is slightly constant because Ni increases the
toughness of the weld metal which imparts slightly constant elongation at the weld joint. The
presence of NiO in combination of CaF2 is really helpful to increase the percentage
elongation.
ISJ
a
o
v
9i
R
• * *
a
u
u-
D:NiO B:CaF2
Figure 5.13: Effect of NiO and CaFz on Percentage Elongation
89
Figure 5.14 shows the response surface behaviour of CaFz and Fe-Cr constituents. The
effect of Fe-Cr on percentage elongation gives positive impact on percentage elongation
because Fe and Cr are good transition metals.
BS
d
a
o
cs
B
u
u
;•
o
E: FeCr 19 1Z9 B: CaF2
Figure 5.14: Effect of Fe-Cr and CaFi on Percentage Elongation
5.3.1.3 Modeling of impact strength
The analysis of variance (ANOVA) without pooling and pooled ANOVA reported in Tables
E.3 (APPENDIX E) and 5.10 respectively show that the model is actually significant. The f-
value of 20 and the p-value of less than 0.0001 in ANOVA demonstrate that this regression is
statistically significant at 99% confidence level. In pooled ANOVA the model fitting was
checked by the coefficient of determination R^ which was found to be 0.9652, thus pointing
out that only 3.48% of the total variations in the response is not explained by the model.
The values of R^ and adjusted R^ are 0.962 and 0.9169 respectively. The associated p-value
in pooled ANOVA model is lower than 0.05 (i.e. a=0.05, or 95%) confidence level) which
indicates that the model is statistically significant. The lack-of-fit term is not significant as
per the model requirement. In pooled ANOVA the factors MnO, CaF2, MgO, Fe-Cr,
interaction effect of factor MnO with factor CaF2, interaction effect of factor MnO with
factor MgO, interaction effect of factor MnO with factor NiO, interaction effect of factor
90
MnO with factor Fe-Cr,, interaction effect of factor CaF2 with factor MgO, interaction effect
of factor CaF2 with factor NiO, interaction effect of factor MgO with factor NiO, interaction
effect of factor MgO with factor Fe-Cr , interaction effect of factor NiO with factor Fe-Cr
and second order term of factor CaF2, second order term of factor MgO, second order term
of factor NiO, second order term of factor Fe-Cr have significant effect.
Table 5.10: Pooled ANOVA of Impact Strength
Sum of Degree of Mean

Source Squares Freedom Squares f-value Prob>f
Model 3254.5 18 180.81 20 < 0.0001 significant
A-MnO 160.17 160.17 17.72 0.001
B-CaFa 541.5 541.5 59.91 < 0.0001
C-MgO 337.5 337.5 37.34 < 0.0001
D-NiO 24 24 2.66 0.1272
E-Fe-Cr 60.17 60.17 6.66 0.0229
AB 361 361 39.94 < 0.0001
AC 49 49 5.42 0.0367
AD 81 81 8.96 0.0104
AE 306.25 306.25 33.88 < 0.0001
BC 72.25 72.25 7.99 0.0143
BD 56.25 56.25 6.22 0.0269
CD 182.25 182.25 20.16 0.0006
CE 144 144 15.93 0.0015
DE 256 256 28.32 0.0001
B^ 332.32 332.32 36.77 < 0.0001
C' 130.78 130.78 14.47 J 0.0022
D' 64.63 64.63 7.15 0.0191
E' 220.01 220.01 24.34 0.0003
Residual 117.5 13 9.04
Lack of 0.7636(not
Fit 56.67 8 7.08 0.58 significant)
Pure
Error 60.83 5 12.17
Cor
Total 3372 31
Std. Dev. =3.01 R-Squared = 0.9652
Contd...
91
Adj R-Squared =
Mean = 80.75 0.9169
Pred R-Squared
C.V. % = 3.73 =0.6952
Adeq Precision =
PRESS = 1027.62 18.094
After eliminating the non-significant terms, the final response equation for impact strength is
given as in Equation 5.3.
Impact strength (Joule) = + 127.91667-25.7333* MnO+1.70000* CaF2+7.03333*MgO

- 9.33333 *NiO+1.2333* F-Cr+1.52000* MnO *CaF2
- 0.56000 * MnO * MgO+0.72000* MnO * NiO
+1.40000*MnO*F-Cr+0.34000*CaF2*MgO
+0.30000*CaF2* NiO+0.54000*MgO*NiO
- 0.48000* MgO * F-Cr-0.64000 * NiO * F-Cr-0.53667 *CaF2^
-0.33667*MgO2-0.23667*Nio2-0.43667*F-Cr2
(5.3)
Figure 5.15 displays the normal probability plot of the residuals for impact strength. The plot
indicates that the residuals are falling on a straight line, which means that the errors are
normally distributed. Figure 5.16 compares the observed values with the predicted values in
the model. It can be seen that the regression model is fairly well fitted with the observed
values.
92
Internally Student'zed Residuals
Figure 5.15: Normal Probability Plot of the Residuals for Impact Strength

110* —
100.00 — •,
MM —
• eriQO
m^
Bono —
J^
70.00 — Q M-'^
60.00 —
-
50.00 —
1 1 1 1 1
Actual
Figure 5.16: Predicted vs. Actual Impact Strength
93
5.3.1.3.1 Effect of flux constituents on impact strength
Figure 5.17 shows the response surface for impact strength in relation to the flux constituents
CaF2 and MnO. As shown in the figure, the impact strength tends to increase considerably
with increase in the percentage of CaF2 constituent. This result is in conformity with that of
Lewis [16]. As per Lewis the impact strength increases with increasing CaF2 content,
because it lowers the high oxygen content and many inclusions present in the submerged arc
welds. The value of impact strength also increases with increasing percentage of MnO
constituent because it is a transition metal oxide.
9
C ©
s
u
u
R
C
S
A: MnO B: CaF2
Figure 5.17: Effect of CaFs and MnO on Impact Strength
The effect of CaF2 and MgO constituents on impact strength is shown in Figure 5.18. The
response surface plot shows that the value of impact strength increases with CaF2 constituent.
The value of impact strength also increases with the increasing percentage of MgO, because
it is very stable (nonreactive) with the combination of CaF2 which reduces the oxide nature.
94
s
o
I
I
S
C:MgO B: CaF2
Figure 5.18: Effect of CaF2 and MgO on Impact Strength
Figure 5.19 shows the effect of NiO and CaF2 constituents on impact strength of weld joint.
The effect of CaF2 on impact strength has an increasing trend while there is very little effect
of NiO on impact strength because of its nonreactive nature with others.
s
o
a
u
-*<
w
a
a
D:NiO B: CaF2
Figure 5.19: Effect of NiO and CaF2 on Impact Strength
95
Figure 5.20 shows the effect of Fe-Cr and CaFi constituents on impact strength. The
effect of Fe - Cr on impact strength is almost constant due to free metallic in nature.
c
1^
c
u
{/:
u
C
17.50
E: FeCr B: CaF2
Figure 5.20: Effect of Fe-Cr and CaF2 on Impact Strength
5.3.1.4 Modeling of Vickers hardness

The Analysis of variance (ANOV A) without pooling and pooled ANOV A reported in Tables
E.4 (APPENDIX E) and 5.11 respectively show that the model is significant. The f-value of
22.08 and the p-value < 0.0001 did in fact demonstrate that this regression was statistically
significant at 99% confidence level. Besides, the fit of the model was checked by the
coefficient of determination R^ which was found to be 0.9132, thus pointing out that only
8.68% of the total variations in the response were not explained by the model. The associated
p-value for the model is lower than 0.05 (i.e. a=0.05, or 95% confidence) which indicates
that the model is statistically significant. The lack-of-fit term is not significant as it is desired.
Further, factors MnO, MgO, Fe-Cr, interaction effect of factor MnO with factor NiO,
interaction effect of factor MnO with factor Fe-Cr, interaction effect of factor CaF2 with
96
factor MgO, interaction effect of factor NiO with factor Fe-Cr and second order term of
factor MnO, have significant effect.
Table 5.11 Pooled ANOVA of Vickers Hardness

Source Squares Freedom Squares f-value Prob >f
A-MnO 42.67 42.67 17.47 0.0004
B-CaF2 0.67 0.67 0.27 0.6069
C-MgO 28.17 28.17 11.53 0.0027
D-NiO 6.00 6.00 2.46 0.1320
E-F-Cr 16.67 16.67 6.82 0.0163
AD 25.00 25.00 10.23 0.0043
AE 72.25 72.25 29.58 < 0.0001
BC 256.00 256.00 104.80 < 0.0001
DE 16.00 16.00 6.55 0.0183
A2 76.00 76.00 31.11 < 0.0001
Residual 51.30 • 21 2.44
Pure Error 5.50 5 1.10
Cor Total 590.72 31
Std.Dev. =1.56 R-Squared = 0.9132
Mean = 164.09 AdjR-Squared = 0.8718
C.V.% = 0.95 Pred R-Squared =0.7794
PRESS = 130.29 Adeq Precision = 17.096
After eliminating the non-significant terms, the final response can be represented by the
equation:
Vickers hardness = + 293.70000 -12.81333* MnO-6.33333 * CaF2-9.16667 *MgO

- 3.60000* NiO+4.6333* Fe-Cr+0.40000* MnO * NiO
- 0.68000* MnO * Fe- Cr+0.64000* CaF2 * MgO+0.16000 * NiO*Fe-Cr
+1.01867*MnO2
(5.4)
Figure 5.21 displays the normal probability plot of the residuals for Vickers hardness. It is
noticed from this plot that the residuals are falling on a straight line, which means that the
97
errors are normally distributed about the straight line. The actual values of Vickers hardness
are compared with the predicted values for the model in Figure 5.22. It can be seen that the
regression model is fairly well fitted with the observed values.
Internally Studentlzed Residuals
Figure 5.21: Normal Probability Plot of the Residuals for

Vickers Hardness
Actual
Figure 5.22: Predicted vs. Actual Value of Vickers Hardness
98
5.3.1.4.1 Effect of flux constituents on hardness
Figure 5.23 depicts the response surface for Vickers hardness with the interaction of flux
constituents NiO and Fe-Cr. In 3-D figure hardness tends to increase with increase in the
percentage of NiO because Ni element provides good hardness in NiO constituent. The
hardness also increases with the increase of Fe-Cr because Fe and Cr are good transition
metals.
E: FeCr
Figure 5.23: Effect of NiO and Fe-Cr on Vickers Hardness
The effect of CaF2 and MgO constituents on Vickers hardness is shown in Figure 5.24.
The figure shows that the value of Vickers hardness increases with CaF2 constituent. As
per Lewis [16], the techno-mechanical property slightly increases with increasing CaF2
content, because it lowers the high oxygen content and many inclusions present in the
submerged arc welds. The value of Vickers hardness also increases with the percentage
of MgO, because it is very stable (non-reactive) with the combination of CaF2 which
reduces the oxide nature.
99
lie
ee
U
fi
u
17.50
A:MgO 7.50 12.50

B: CaF2
Figure 5.24: Effect of CaF2 and MgO on Vickers Hardness
Figure 5.25 shows the effect of MnO and NiO constituents on Vickers hardness. The effect
of NiO on hardness shows an increasing trend because the presence of Ni in NiO increases
the hardness. The value of hardness also increases with the percentage of MnO constituent
because it is a transition metal oxide.
100
a
u
D:NiO A: MnO
Figure 5.25: Effect of MnO and NiO on Vickers Hardness
Figure 5.26 shows the response surface of Vickers hardness with the interaction of MnO and
Fe-Cr. The value of hardness also increases with the percentage of MnO constituent because
it is a transition metal oxide. The hardness also increases with the increase of Fe-Cr because
Fe and Cr are good transition metals.
!Ji la^L
<
V»
162 J
a
-B
h.
CS 160
E: FeCr A: MnO
Figure 5.26: Effect of MnO and Fe-Cr on Vickers Hardness
101
5.3.2 Optimization of mechanical properties using desirability function
5.3.2.1 Desirability function
The desirability function for single and multiple responses has been described by Derringer
and Suich [87]. The method makes use of an objective function, D(X), called the desirability.
The desirable ranges are from zero to one, least to most desirable respectively. The factor
settings with maximum total desirability are considered to be the optimal parameter
conditions.
The simultaneous objective function is a geometric mean of all transformed responses:
D = (dlxd2xd3x dn) 1/n = [ F T di j
Where, n is the number of responses in the measure. If any of the responses or factors falls
outside the desirability range, the overall function becomes zero.
Desirability is an objective function that ranges from zero outside of the limits to one at the
goal. The numerical optimization finds a point that maximizes the desirability function. The
characterstics of a goal may be altered by adjusting the weight or importance. For several
responses and factors, all goals get combined into one desirability function.
Responses must assigned a low and high value to each goal for simultaneous optimization.
The "Goal" field for responses must be one of five choices: "none", "maximum",
"minimum", "target", or "in range". Factors will always be included in the optimization, at
their design range by default, or as a miximum, minimum or target goal.
Desirability function has been used to determine the optimum parameters for developed
agglomerated fluxes used during submerged arc welding for optimization of mechanical
properties of weld joints like tensile strength, elongation, impact strength and hardness. The
bead morophology parameters like bead penetration, bead reinforcement and bead width are
also optimise in the present investigation. Second order central composite experimental
102
design involving five flux constituents as factors (MnO, CaFa, MgO, NiO and Fe-Cr) each at
five levels has been used to find optimum combination of factors. The single response and
multi response optimizations were achieved through desirability function.
5.3.2.2 Single response optimization using desirability function

The individual response characterstics — elongation , tensile strength, hardness and impact
strength — for the optimization are given in Tables 5.12- 5.15. Goals and limits were
established for each response individually in order to accurately determine their impact on
individual desirability. A maximum or minimum level is provided for each response
characterstic which has to be optimized.
Weights are assigned to give added emphasis to upper / lower bounds or to emphasize a
target value. The default value "1" of weight is assigned to a goal to adjust the shape of its
particular desirability fimction. The default value "3" is selected for importance in order to
give equal importance to all goals.
Table 5.12: Range of Constituents and Elongation for Desirability
Constituents Lower Upper Lower Upper

Goal Importance
<& Response Limit Limit Weight Weight
A-MnO is in range 6.25 8.75 1 3
B-CaF2 is in range 12.5 17.5 1 3
C-MgO is in range 7.5 12.5 1 3
D-NiO is in range 2.5 7.5 1 3
E-Fe-Cr is in range 2.5 7.5 1 3
Elongation maximize 26.52 29.58 1 3
103
Table 5.13: Range of Constituents and Tensile Strength for Desirability

& Response Goal Limit Limit Weight Weight Importance
A-MnO is in range 6.25 8.75 3
B-CaF2 is in range 12.5 17.5 3
C-MgO is in range 7.5 12.5 3
D-NiO is in range 2.5 7.5 3
E-Fe-Cr is in range 2.5 7.5 3
Tensile
Strength maximize 520 580 3
Table 5.14: Range of Constituents and Hardness for Desirability
Constituents & Goal Lower Upper Lower Upper

Response Limit Limit Weight Weight Importance
is in
A-MnO range 6.25 8.75 3
is in
B-CaF2 range 12.5 17.5 3
is in
C-MgO range 7.5 12.5 3
is in
D-NiO range 2.5 7.5 3
is in
E-Fe-Cr range 2.5 7.5 3
maximize
Hardness 157 174 3
Table 5.15: Range of Constituents and Impact Strength for Desirability
Constituents & Lower Upper Lower Upper

Response Goal Limit Limit Weight Weight Importance
Impact Strength maximize 60 99 3
104
5.3.2.2.1 Optimal solutions using desirability function
The aim of optimization is to find a good set of conditions that will meet all the goals. It is
not necessary that the value of desirability is always 1.0 as the value is completely dependent
on how closely the lower and upper limits are set relative to the optimum [88]. A set of 47
solutions is derived for the specified design for response characteristics — elongation ,
impact strength, hardness — and 46 solutions are derived for tensile strength using Design
Expert Software. The set of conditions possessing highest desirability value is selected as
optimum condition for the desired response. The optimal conditions for the different
response are given in Table 5.16-5.19.
Table 5.16: Optimal Solutions for Desirability (Elongation)
S. No. MnO CaFa MgO NiO Fe-Cr Elongation Desirability

1 8.75 17.5 12.5 2.5 2.5 29.5274 0.983 Selected
2 8.75 17.5 12.5 2.92 2.5 29.5128 0.978
3 6.25 17.5 12.5 7.5 7.48 29.4594 0.961
4 8.75 17.5 12.5 3.97 2.73 29.454 0.959
5 8.73 17.32 12.5 3.1 2.51 29.4324 0.952
6 8.75 17.5 12.5 7.23 2.5 29.3895 0.938
7 8.75 17.5 12.5 7.5 2.5 29.3812 0.935
8 8.75 17.49 12.5 7.44 2.63 29.3753 0.933
9 6.25 17.26 12.25 7.5 7.5 29.3141 0.913 #
10 8.75 16.81 12.5 2.54 2.51 29.3032 0.91
11 8.75 17.5 12.5 7.43 4.82 29.2816 0.902
12 8.75 17.5 12.5 7.01 4.83 29.2685 0.898
13 8.75 17.5 12.5 7.44 5.14 29.2681 0.898
14 6.46 17.46 12.5 7.5 7.1 29.2427 0.89
15 6.25 17.5 12.37 5.51 7.5 29.228 0.885
16 8.75 12.5 7.5 2.5 2.5 29.1515 0.86
17 8.75 16.62 12.5 7.5 2.56 29.0944 0.841
18 6.25 12.5 7.52 7.5 7.5 29.0862 0.839
19 6.25 12.5 7.5 7.5 7.47 29.0852 0.838
20 6.27 12.5 7.54 7.5 7.5 29.0741 0.835
21 6.25 17.5 12.41 3.59 7.5 29.0556 0.829
22 8.75 12.5 7.54 6.2 2.5 29.0326 0.821
105 Contd.
23 7.85 17.5 12.5 7.5 7.49 29.0316 0.821
24 6.25 12.62 7.5 7.18 7.5 29.0294 0.82
25 8.75 12.5 7.5 5.99 2.78 29.0272 0.819
26 8.75 12.5 7.5 6.26 2.72 29.0253 0.819
27 8.71 12.5 7.5 6.1 2.5 29.0053 0.812
28 8.75 12.5 7.5 7.5 2.5 29.0045 0.812
29 8.75 12.5 8.14 2.5 2.5 28.9947 0.809
30 6.25 12.5 7.5 7 7.32 28.9943 0.809
31 8.74 17.5 12.5 4.84 6.72 28.9933 0.808
32 8.75 12.84 7.5 4.83 2.5 28.9933 0.808
33 8.75 17.5 12.5 2.77 5.91 28.965 0.799
34 6.51 12.5 7.51 7.28 7.5 28.9405 0.791
35 8.75 12.5 7.5 7.2 4.2 28.93 0.788
36 6.38 12.5 7.5 6.15 7.5 28.8904 0.775
37 8.75 12.5 7.5 5.98 4.75 28.8671 0.767
38 6.25 12.52 7.5 4.46 7.5 28.7915 0.742
39 6.25 12.5 7.5 3.7 7.5 28.7209 0.719
40 8.75 14.97 12.5 2.5 2.51 28.7189 0.719
41 7.37 12.5 7.5 7.5 7.47 28.6914 0.71
42 8.75 17.43 11.32 2.5 5.03 28.6773 0.705
43 8.75 12.53 7.5 4.49 6.1 28.6577 0.699
44 7.08 12.5 7.5 5.94 7.5 28.6037 0.681
45 6.25 14.55 11.09 7.5 7.5 28.5314 0.657
46 8.13 12.5 7.5 4.97 5.58 28.4701 0.637
47 6.71 17.5 12.5 2.5 4.37 28.4509 0.631
Table 5.17: Optimal Solutions for Desirability (Tensile Strength)
Tensile
Number MnO CaFz MgO NiO Fe-Cr Strength Desirability
1 8.75 17.5 12.5 2.5 2.5 577.248 0.954 Selected
2 8.75 17.49 12.5 7.5 2.5 575.622 0.927
3 8.7 17.5 12.5 3.58 2.5 575.176 0.92
4 8.75 12.5 7.5 7.5 2.51 574.654 0.911
5 8.75 17.5 12.5 4.15 2.84 574.519 0.909
6 8.75 12.5 7.5 7.31 2.5 574.268 0.904
7 6.25 17.5 12.5 2.5 7.5 574.232 0.904
8 8.7 17.46 12.49 4.64 2.5 574.124 0.902
Contd.
106
9 8.75 17.47 12.5 7.5 3.78 574.021 0.9
10 6.26 17.5 12.5 7.03 7.46 573.946 0.899
11 8.75 17.24 12.5 6.66 2.5 573.888 0.898
12 8.71 12.5 7.5 7.5 3.14 573.376 0.89
13 6.34 17.5 12.5 6.85 7.5 573.145 0.886
14 8.75 12.5 7.5 7.5 5.14 573.144 0.886
15 8.75 17.5 12.5 6.96 4.02 573.081 0.885
16 6.25 17.5 12.5 4.56 7.48 573.016 0.884
17 8.75 17.5 12.5 7.5 5.1 572.66 0.878
18 6.25 17.34 12.24 7.5 7.5 572.297 0.872
19 6.48 17.5 • 12.5 7.24 7.5 572.189 0.87
20 8.75 12.5 7.5 7.4 6.84 571.874 0.865
21 8.75 12.55 7.54 7.5 7.47 571.585 0.86
22 6.25 16.89 12.5 7.5 7.5 571.554 0.859
23 8.75 17.5 12.5 6.86 5.37 571.032 0.851
24 8.75 17.5 12.5 7.35 6.26 ^ 571.019 0.85
25 8.75 17.5 12.5 3.59 4.18 570.72 0.845
26 8.67 12.5 7.5 6.69 3.3 570.476 0.841
27 8.75 . 17.5 12.5 5.44 4.73 570.082 0.835
28 8.75 12.5 7.57 4.93 2.5 569.785 0.83
29 8.75 16.29 12.5 7.5 2.85 569.623 0.827
30 8.75 12.5 8.26 7.5 5.7 569.457 0.824
31 8.75 12.5 7.5 3.84 2.5 569.012 0.817
32 8.75 12.5 7.53 6.8 7.21 568.835 0.814
33 6.25 12.5 7.5 7.5 7.5 568.781 0.813
34 8.75 16.44 11.9 7.5 2.5 568.36 0.806
35 8.32 12.5 7.5 7.5 6.92 568.127 0.802
36 8.75 16.15 12.48 2.5 2.53 568.067 0.801
37 8.75 17.46 12.49 3.71 5.06 567.808 0.797
38 6.25 17.5 11.55 2.9 7.5 567.738 0.796
39 6.25 16.69 11.87 7.5 7.5 567.387 0.79
40 8.75 12.5 9.27 7.5 2.59 566.662 0.778
41 6.25 17.41 12.5 2.5 5.01 566.24 0.771
42 6.25 16.66 12.27 2.5 6.32 563.553 0.726
43 6.25 12.5 7.5 5.29 7.45 563.469 0.724
44 7.94 12.5 7.5 6.38 7.17 561.606 0.693
45 6.25 16.33 12.5 2.5 5.61 560.014 0.667
46 6.25 12.5 7.62 2.61 7.5 559.942 0.666
107
Table 5.18: Optimal Solutions for Desirability (Hardness)
S.No. MnO CaF2 MgO NiO Fe-Cr Hardness Desirability

1 8.75 17.5 12.5 7.5 2.5 173.1 0.947 Selected
2 8.75 17.5 12.5 7.47 2.5 173.098 0.947
3 8.75 17.5 12.5 7.43 2.78 173.057 0.945
4 8.75 17.5 12.5 7.48 3.08 173.04 0.944
5 8.75 17.49 12.5 7.15 2.5 172.989 0.941
6 8.75 17.5 12.5 7.5 3.9 172.947 0.938
7 8.75 17.5 12.42 7.45 2.5 172.935 0.937
8 8.75 17.5 12.5 7.47 4.69 172.841 0.932
9 8.75 17.5 12.42 6.95 2.54 172.779 0.928
10 8.75 17.48 12.5 6.99 3.39 172.75 0.926
11 8.75 17.47 12.49 7.5 5.41 172.716 0.924
12 8.75 17.5 12.48 7.5 6.29 172.626 0.919
13 8.75 17.5 12.5 4.87 2.5 172.32 0.901
14 8.75 17.5 12.31 5.92 2.5 172.242 0.897
15 8.62 17.49 12.5 7.5 7.5 172.146 0.891
16 8.75 16.81 12.5 7.5 2.7 171.941 0.879
17 8.75 17.01 12.5 7.5 6.46 171.844 0.873
18 8.75 17.5 12.49 3.32 2.5 171.834 0.873
19 6.25 17.5 12.5 7.5 7.5 171.616 0.86
20 6.25 17.48 12.47 7.5 7.5 171.511 0.854
21 6.25 17.47 12.5 5.83 7.5 171.403 0.847
22 6.25 17.44 12.45 4.07 7.5 171.066 0.827
23 6.53 17.49 12.5 7.5 7.24 170.71 0.806
24 7.95 17.5 12.48 7.5 7.27 170.704 0.806
25 6.61 17.5 12.49 * 7.5 7.27 170.628 0.802
26 8.75 12.5 7.5 7.49 2.5 170.603 0.8
27 8.75 12.5 7.51 7.5 2.53 170.602 0.8
28 6.25 16.96 12.5 5.15 7.5 170.486 0.793
29 8.75 17.5 12.5 4.46 5.5 170.394 0.788
30 8.75 12.5 7.5 6.47 2.5 170.305 0.783
31 8.7 16.2 12.5 7.5 7.49 170.237 0.779
32 8.69 12.5 7.6 7.5 2.55 170.145 0.773
33 8.75 12.5 8.07 7.5 2.5 169.955 0.762
34 7.77 17.48 12.5 6.17 7.5 169.685 0.746
35 8.75 15.59 12.16 7.5 2.5 169.656 0.744
36 8.74 17.32 12.14 2.5 3.56 169.639 0.743
37 8.62 12.54 7.58 7.5 7.5 169.529 0.737
108 Contd.
38 8.75 15.58 11.94 7.41 2.5 169.425 0.731
39 7.89 17.43 12.5 • 5.79 7.5 169.367 0.727
40 8.75 12.5 7.5 4.07 2.95 169.289 0.723
41 6.25 12.5 7.51 7.5 7.5 169.1 0.712
42 7.99 17.5 12.5 5.1 5.13 168.958 0.703
43 6.25 12.5 7.5 4.36 7.5 168.802 0.694
44 6.39 12.5 7.5 7.5 7.44 168.73 0.69
45 6.25 12.72 7.98 7.49 7.49 168.283 0.664
46 8.75 14.33 9.11 7.5 2.5 167.816 0.636
47 6.25 14.74 10.09 2.5 7.1 165.556 0.503
Table 5.19: Optimal Solutions for Desirability (Impact Strength)
Impact
Number MnO CaF2 MgO NiO Fe-Cr Strength Desirability
1 6.25 15.2 12.46 ^ 7.16 2.51 99.2352 Selected
2 6.25 17.5 12.5 7.5 2.5 99.25
3 7.01 16.88 12.24 7.46 3.14 99.4847
4 8.05 16.74 11.7 7.07 3.34 99.7956
5 8.22 17.47 12.43 6.85 5.43 100.367
6 6.28 15.76 12.46 7.29 2.75 99.1034
7 8.72 17.47 10.77 6 5.85 99.8363
8 8.75 17.5 12.5 7.5 7.5 99.75
9 7.19 17.01 12.31 6.86 2.78 99.4465
10 8.35 17.02 12.43 6.13 4.97 99.3039
11 8.75 17.47 12.11 6.62 7.48 99.0925
12 8.72 17.23 11.17 7.13 4.6 102.1
13 8.66 17.43 12.43 5.88 6.7 99.0254
14 8.2 17.35 11.85 7.48 4.63 101.555
15 8.52 16.34 11.66 7.03 2.85 99.4371
16 8.7 17.18 11.12 6.06 6.37 99.2306
17 8.07 16.94 12.36 7.4 3.04 102.796
18 8.09 17.29 11.68 7.38 4.37 100.609
19 8.48 17.15 11.43 6.33 5.09 99.5163
20 8.67 16.6 11.35 7.1 4.29 100.59
21 8.66 16.91 10.91 6.84 5.45 99.7313
22 8.38 17.05 12.48 6.03 3.9 99.7529
23 7.08 15.9 12.43 7.22 2.65 99.7314
24 7.97 17.4 11.77 6.8 4.51 99.1419
Contd....
109
25 6.27 17.49 12.48 7.45 2.56 99.0047
26 6.69 16.11 12.47 7.32 2.82 99.5861
27 8.31 16.14 12.44 7.21 2.71 100.874
28 8.6 16.84 11.85 6.99 6.37 99.0631
29 6.75 15.58 12.38 7.37 2.57 99.4788
30 6.46 15.68 12.26 7.45 2.54 99.3122
31 6.28 16.26 12.39 7.02 2.52 99.0752
32 8.56 16.99 11.69 7.23 5.13 101.449
33 8.69 17.23 11.13 7.23 5.5 101.498
34 7.9 16.31 12.02 7.1 3.3 99.4277
35 6.44 15.04 12.47 7.46 2.51 99.394
36 8.75 17.5 10.13 5.85 5.32 98.7922 0.995
37 8.11 17.5 10.8 7.5 2.5 98.7372 0.993
38 7.07 17.5 12.5 6.34 2.5 98.4038 0.985
39 8.75 17.5 9.57 5.15 7.5 97.5216 0.962
40 8.75 17.5 10.17 4.25 7.5 97.2756 0.956
41 6.25 14.98 11.87 7.5 2.87 96.4763 0.935
42 8.75 17.43 9.37 3.12 7.47 96.1925 0.928
43 8.75 17.43 8.62 2.81 7.5 95.9469 0.922
44 8.75 17.49 8.66 2.66 7.46 95.8075 0.918
45 8.75 17.5 8.86 7.06 7.5 95.4382 0.909
46 8.75 16.37 7.5 2.58 7.5 94.6667 0.889
47 8.57 17.5 9 7.5 6.72 94.5355 0.886
The ramp function graphs and bar graphs (Figures 5.27-5.34) drawn using Design Expert
solver show the desirability for each factor and each response. The dot on each ramp reflects
the factor setting or response prediction for that response characteristic. The height of the dot
shows how desirable it is. A linear ramp function is created between the low value and the
goal or the high value and the goal as the weight for each parameter was set equal to one. Bar
graphs show the individual/ partial desirability functions (d) of each of the responses (tensile
strength, elongation, impact strength, hardness); di varies from 0 to 1 depending upon the
closeness of the response towards target [89]. The bar graph shows how well each variable
satisfies the criterion: values close to one are considered good. Table 5.20 reports the final set
of optimum levels of various process parameters and the predicted values of various response
characteristics.
110
Table 5.20: Optimal Sets of Process Parameters Using RSM and Desirability Function
Constituents"^ Predicted
Response 1 Desirability optimal
MnO CaF2 MgO NiO Fe-Cr
response
Tensile Strength 0.954 8.75 17.5 12.5 2.5 2.5 577.248
Elongation 0.983 8.75 17.5 12.5 2.5 2.5 29.5274
Impact Strength 1 6.25 15.2 12.46 7.16 2.51 99.2352
Hardness 0.947 8.75 17.5 12.5 7.5 2.5 173.1
T
C:MlO« 12.50
TenaloSl»nslli»577.2«
Desirability = 0.954
Figure 5.27: Ramp Function Graph of Desirability for Tensile Strength
111
Desirability
A:MnO
=,
C:MgO
ile Strength
Combined
/ „_ y ^ J
Figure 5.28: Bar Graph of Desirability for Tensile Strength
6.25 8.75 12.50 17.50
A:MnO = 8.75 B:CaF2 = 17.50
7.50
C:MgO = 12.50 D:NiO = 2.50
2.50 29.58
E:Fe-Cr = 2.50 Elongation = 29.5274
Figure 5.29: Ramp Function Graph of Desirability for Elongation
112
Desirability
Figure 5.30: Bar Graph of Desirability for Elongation
B:CaF2= 15.20
C:MgO = 12.46
Impact Strength = 99.2352
Figure 5.31: Ramp Function Graph of Desirability for Impact Strength
113
Desirability
A:MnO
B:CaF2 1 W
*''\ . w
Impact strength
Combined
\ \ - 4 1
.- — y^ -, ,-„xL . . .. , ,^. ... y
0,250 0.750
Figure 5.32: Bar Graph of Desirability for Impact Strength
i.25 8.75 17.50
A:MnO = 8.75 B:CaF2 = 17.50
7.50 12.50 2.50 7.50
C:MgO = 12.50 D:NiO = 7.50
7.50
E:Fe-Cr = 2.50 Hardness = 173.1
Figure 5.33: Ramp Function Graph of Desirability for Hardness
114
Desirability
EiFfrCr
0.9. 7035
Figure 5.34: Bar Graph of Desirability for Hardness
5.3.2.2.2 Multiple response optimization using desirability function

The problem of conflicting responses of single response optimization is overcome by multi
response optimization using desirability function in conjunction with response surface
methodology. Various multi-characteristic models have been developed. Goals and limits
were established for each response in order to accurately determine their impact on overall
desirability. A maximum or minimum level is provided for all response characteristics which
are to be optimized. Weights are assigned in order to give added emphasis to upper or lower
bounds or to emphasize a target value. The importance is assigned to each response relative
to the other responses. Importance varies from the least important (1), to the most important
(5).
Model 1: Tensile strength and elongation

The ranges and goals of flux constituents used as input parameters, and the response
characteristics viz. tensile strength and elongation are given in Table 5.21. Tensile strength
has been assigned an importance of 5 relative to elongation with an importance of 3.
115
Table 5.21: Range of Constituents and Responses for Desirability (Tensile strength and
Elongation)
Lower Upper Lower Upper

Name Goal limit limit weight weight Importance
is in
A-MnO range 6.25 8.75 3
is in
B-CaF2 range 12.5 17.5 3
is in
C-MgO range 7.5 12.5 3
is in
D-NiO range 2.5 7.5 3
is in
E-Fe-Cr range 2.5 7.5 3
Tensile
Elongation maximize 26.52 29.58 3
A set of 47 optimal solutions is derived for the specified design space constraints (Table
5.22) for tensile strength and elongation using Design expert software. The set of conditions
possessing highest desirability value is selected as optimum condition for the desired
responses. Table 5.22 shows the optimal set of condition with higher desirability function
required for obtaining desired response characteristics under specified constraints.
Table 5.22: Set of Optimal Solutions for Tensile Strength and Elongation
Fe- Tensile
No. MnO CaF2 MgO NiO Elongation Desirability
Cr Strength
1 8.75 17.5 12.5 2.5 2.5 579.021 29.5274 0.983 Selected
2 6.26 17.49 12.5 7.5 7.5 577.618 29.4613 0.961
3 6.25 17.5 12.5 7.24 7.5 577.234 29.4411 0.954
4 8.73 17.32 12.5 3.1 2.51 577.152 29.4324 0.952
5 8.75 17.5 12.5 7.5 2.5 576.138 29.3812 0.935
6 8.71 17.5 12.5 2.5 3.26 575.933 29.3691 0.932
7 8.67 17.5 12.5 5.74 2.51 575.735 29.36 0.929
8 6.25 17.5 12.5 6.22 7.5 575.355 29.343 0.923
9 6.5 17.5 12.5 7.5 7.5 575.303 29.3429 0.922
10 8.75 17.5 12.5 5.94 3.58 575.302 29.3389 0.922
Contd.
116
11 6.25 17.5 12.5 6.02 7.5 574.983 29.3236 0.916
12 6.25 17.26 12.25 7.5 7.5 574.722 29.3141 0.912
13 8.75 17.5 12.5 7.45 4.36 574.549 29.3023 0.909
14 8.75 17.5 12.5 6.77 5.75 572.642 29.2056 0.877
15 6.75 17.49 12.5 7.14 7.5 572.462 29.197 0.875
16 8.75 12.5 7.5 2.5 2.5 571.521 29.1515 0.859
17 8.75 12.5 7.5 3.81 2.5 570.765 29.1132 0.847
18 8.75 12.5 7.5 3.69 2.55 570.691 29.1094 0.845
19 8.75 12.5 7.5 3.84 2.6 570.496 29.0994 0.842
20 8.75 16.62 12.5 7.5 2.56 570.479 29.0944 0.841
21 8.75 12.5 7.58 3.92 2.5 570.301 29.0896 0.839
22 8.75 16.66 12.5 7.5 3.2 570.232 29.0824 0.837
23 6.25 12.5 7.5 7.5 7.47 570.114 29.0852 0.836
24 8.75 12.58 7.5 4.07 2.53 570.129 29.0807 0.836
25 6.27 12.5 7.5 7.5 7.49 570.022 29.0805 0.835
26 6.42 17.5 12.5 3.93 7.5 569.313 29.0296 0.821
27 7.36 17.45 12.5 7.5 7.31 569.222 29.0322 0.821
28 8.75 12.5 7.5 5.99 2.78 569.069 29.0272 0.818
29 6.26 12.5 7.5 7.5 7.29 569.003 29.0281 0.818
30 8.75 12.5 7.5 6.26 2.72 569.031 29.0253 0.818
31 8.75 17.5 12.5 5.92 7.44 569.035 29.0213 0.817
32 8.75 12.5 7.5 7.5 2.5 568.622 29.0045 0.811
33 8.75 12.5 7.5 5.31 3.24 568.481 28.9972 0.809
34 8.75 12.62 7.57 6.34 2.5 568.37 28.9914 0.807
35 6.25 12.63 7.59 6.73 7.5 567.736 28.9622 0.797
36 6.25 17.5 12.5 5.39 6.14 567.613 28.9445 0.793
37 8.75 12.5 7.5 6.17 3.89 567.321 28.9388 0.789
38 6.25 12.5 7.52 5.87 7.5 567.122 28.929 0.786
39 8.75 13.4 7.5 2.66 2.5 566.703 28.905 0.779
40 8.75 12.5 7.5 7.49 4.9 566.594 28.9039 0.777
41 8.75 12.5 7.5 6.02 5.49 564.776 28.8098 0.747
42 8.75 12.68 7.5 2.5 4.45 564.092 28.7708 0.735
43 8.56 12.6 7.5 7.5 7.5 562.864 28.7162 0.716
44 8.75 12.5 7.5 4.48 6.18 561.811 28.6565 0.697
45 7.12 17.5 12.5 4.15 5.23 560.831 28.596 0.68
46 8.75 13.33 7.5 3.91 4.71 560.548 28.5907 0.676
47 6.25 14.41 11.64 7.5 7.5 558.66 28.5005 0.645
117
The ramp function graph and bar graph (Figures 5.35 and 5.36) drawn using Design Expert
solver show the desirability for tensile strength and elongation. Bar graph shows the overall
desirability function of the responses (tensile strength and elongation). Desirability varies
from 0 to 1 depending upon the closeness of the response towards target. The bar graph
shows how well each variable satisfies the criterion, values close to one are considered good.
6.25
T
8.75 12.50 17.50
A:MnO = 8.75 B:CaF2 = 17.50
T2.50 7.50
D:NiO = 2-50
T
Tenale Strengffi = 579.021
Desirability =0.983
Elongation = 29.5274
Figure 5.35: Ramp Function Graph of Desirability for Tensile Strength and Elongation
118
Desirability
WIIWIIIIIIIIllI
0.983681
0.982802
0.983351
-^ ^ z:
0.750
r
Figure 5.36: Bar Graph of Desirability for Tensile Strength and Elongation
Model 2: Tensile strength and impact strength

characteristics viz. tensile strength and impact strength are given in Table 5.23. Tensile
strength has been assigned an importance of 5 relative to impact strength with an importance
of3.
Impact Strength)

Tensile
Impact Strength maximize 60 99 3
119
5.24) for tensile strength and impact strength using Design expert software. The set of
conditions possessing highest desirability value is selected as optimum condition for the
desired responses. Table 5.24 shows the optimal set of condition with higher desirability
function required for obtaining desired response characteristics under specified constraints.
Table 5.24: Set of Optimal Solutions for Tensile Strength and Impact strength
Fe- Tensile Impact

No. MnO CaF2 MgO NiO C r Strength Strength
Desirability
1 8.75 17.5 12.41 5.59 2.5 576.632 98.9407 0.964 Selected

2 8.75 17.5 12.5 7.49 2.5 576.122 107.205 0.959
3 8.75 17.5 12.5 6.95 3.04 575.849 105.431 0.956
4 8.68 17.5 12.5 6.81 2.5 575.141 104.342 0.949
5 8.75 17.5 12.5 7.05 3.77 575.045 105.773 0.948
6 8.75 17.32 12.5 7.5 2.5 574.975 106.738 0.947
7 8.73 17.5 12.5 7.5 3.65 574.878 107.167 0.946
8 8.75 17.5 12.5 5.17 3.88 574.743 98.9999 0.944
9 8.75 17.5 12.4 7.34 3.94 574.225 106.417 0.939
10 8.75 17.5 12.45 7.5 4.48 574.125 106.448 0.938
11 8.65 17.49 12.5 5.83 3.36 573.942 100.942 0.936
12 8.75 17.36 12.5 5.21 4.08 573.463 98.9999 0.93
13 8.75 17.5 11.78 5.85 2.52 572.223 99.011 0.917
14 8.75 17.5 12.5 4.64 4.59 573.076 97.3714 0.912
15 8.75 17.5 11.85 7.5 2.96 571.366 105.758 0.907
16 8.75 17.5 12.5 5.54 5.95 571.02 99.7092 0.904
17 8.53 17.5 12.5 7.5 6.44 571.013 100.586 0.904
18 8.43 17.5 12.5 6.12 4.73 569.623 100.853 0.888
19 8.17 17.5 12.5 7.5 4.44 567.588 103.42 0.865
20 8.75 17.5 12.5 4.15 6.55 567.942 95.346 0.838
21 8.75 15.67 12.5 6.95 2.5 564.921 98.5776 0.831
22 8.61 17.5 12.5 4.74 7.5 565.961 94.2669 0.806
23 8.75 17.5 11.86 3.42 5.94 564.336 93.9924 0.786
24 7.04 17.5 12.5 7.5 6.62 568.272 87.3162 0.764
25 6.47 17.5 12.5 7.44 5.4 564.974 89.4277 0.751
26 6.39 17.5 12.5 7.5 5.73 567.062 87.1212 0.75
27 6.28 17.5 12.5 7.5 5.66 567.319 86.6784 0.748
28 6.25 17.49 12.5 7.47 5.84 568.393 85.4553 0.745
120 Contd,
29 6.31 17.5 12.5 7.16 5.52 566.047 87.6478 0.745
30 6.25 17.5 12.5 7.5 4.96 563.554 90.3052 0.745
31 6.37 17.5 12.5 6.87 5.49 565.318 88.0856 0.742
32 6.35 17.5 12.5 6.22 5.73 566.015 86.3755 0.732
33 6.25 17 12.48 7.5 6.74 570.321 80.1226 0.699
34 8.75 15.17 9.9 4.72 2.5 559.499 84.676 0.649
35 8.75 13.95 8.21 5.61 4.94 556.752 83.9579 0.613
36 6.25 12.5 7.5 2.5 7.5 560.896 79.9167 0.612
37 8.75 13.33 7.5 4.81 5.75 558.769 81.7462 0.611
38 8.75 13.14 7.5 4.4 5.73 559.414 81.1512 0.611
39 8.75 13.33 7.5 5.2 5.34 559.962 80.6541 0.611
40 8.75 13.35 7.5 4.94 5.85 558.597 81.8701 0.611
41 8.75 12.99 7.5 4.46 6.04 559.509 80.983 0.61
42 6.26 12.5 7.51 2.5 7.21 559.831 80.3985 0.607
43 8.75 12.61 7.5 3.09 6.26 559.267 80.7451 0.606
44 6.25 12.5 7.5 3.23 7.03 560.614 78.5461 0.593
45 8.56 12.66 7.5 2.5 5.55 558.7 79.2519 0.583
46 6.25 12.5 7.57 3.68 7.36 562.276 76.4351 0.581
47 6.25 12.5 9.27 4.21 6.55 551.84 78.6186 0.51
solver show the desirability for tensile strength and impact strength. Bar graph shows the
overall desirability function of the responses (tensile strength and impact strength).
Desirability varies from 0 to 1 depending upon the closeness of the response towards target.
The bar graph shows how well each variable satisfies the criterion, values close to one are
considered good.
121
T. 12-50
T
17-50
B:CaF2 = 17,50
7.50
T
12.50
C:MgO = 12-41
2.50 7.50
E:Fo-Cf=2.50 Tonsilo Strength = 576,632
Impact Strength = 98.9407
Figure 5.37: Ramp Function Graph of Desirability for Tensile and Impact Strength
Desirability
A:MnO
B:CaF2
w
C:MgO
1 W
D:NiO ^ ^ ^ ^ ^ H
E:Fe-Cr ^ ^ • ^ ^
3874
4 1 1 1 -• 0.99848
,1 1 1 1 963992
'^ \^ ^
1 -'
1 .^^
1
0-0 0.500 0.750 1.000
00 0.250
Figure 5.38: Bar Graph of Desirability for Tensile and Impact Strength
Model 3: Tensile strength and hardness

characteristics viz. tensile strength and hardness are given in Table 5.25. Tensile strength has
been assigned an importance of 5 relative to hardness with an importance of 3.
122
Hardness)

Name Goal Importance
limit limit weight weight
Tensile
Hardness maximize 157 174 3
5.26) for tensile strength and hardness using Design expert software. The set of conditions
possessing highest desirability value is selected as optimum condition for the desired
responses. Table 5.26 shows the optimal set of condition with higher desirability function
required for obtaining desired response characteristics under specified constraints.
Table 5.26: Set of Optimal Solutions for Tensile Strength and Hardness
Fe- Tensile
No. MnO CaF2 MgO NiO Cr Strength Hardness Desirability
1 8.75 17.5 12.5 7.5 2.5 576.118 173.108 0.94 Selected
2 8.75 17.47 12.5 7.5 2.59 575.873 173.054 0.936
3 8.75 17.5 12.5 3.53 2.5 578.391 171.915 0.936
4 8.75 17.5 12.5 2.81 2.5 578.831 171.705 0.935
5 8.73 17.49 12.5 7.5 2.74 575.567 172.969 0.931
6 8.75 17.44 12.48 7.47 2.65 575.531 172.954 0.93
7 6.25 17.5 12.5 7.5 7.5 577.755 171.613 0.923
8 8.75 17.5 12.5 5.9 3.18 575.909 172.371 0.921
9 8.7 17.44 12.48 2.51 2.5 577.644 171.286 0.914
10 6.25 17.37 12.5 7.49 7.48 576.86 171.377 0.908
11 6.25 17.41 12.39 7.5 7.5 576.456 171.241 0.901
12 6.26 17.5 12.5 6.13 7.5 575.119 171.458 0.892
13 8.75 17.5 12.5 7.45 6.56 572.623 172.596 0.892
123 Contd.
14 8.69 17.5 12.5 7.5 6.23 572.462 172.471 0.888
15 6.25 17.5 12.47 5.63 7.5 574.078 171.375 0.88
16 6.25 17.41 12.13 7.5 7.5 574.719 170.719 0.871
17 8.75 17.23 12.07 6.58 2.5 572.339 171.592 0.867
18 6.25 17.04 12.4 7.5 7.5 574.273 170.67 0.865
19 6.25 17.45 12.5 4.47 7.48 571.712 • 171.213 0.852
20 8.75 17.5 12.5 5.53 6 570.932 171.015 0.84
21 6.74 17.5 12.49 7.49 7.23 572.104 170.44 0.838
22 6.75 17.49 12.5 7.5 7.16 571.793 170.355 0.833
23 6.25 17.5 12.5 3.35 7.16 568.77 170.888 0.814
24 8.75 12.5 7.5 7.49 2.5 568.652 170.612 0.807
25 8.75 12.5 7.5 7.04 2.5 568.862 170.463 0.806
26 8.75 12.5 7.5 5.86 2.5 569.52 170.105 0.804
27 7.86 17.5 12.5 7.15 7.5 568.556 170.487 0.803
28 8.75 12.5 7.5 7.46 3.2 568.061 170.518 0.799
29 8.75 12.5 7.5 6.79 3.07 568.359 170.269 0.796
30 6.25 12.5 7.5 7.5 7.5 570.248 169.111 0.788
31 8.67 12.5 7.5 5.84 2.5 568.024 169.647 0.779
32 6.25 12.5 7.5 6.99 7.5 569.31 169.063 0.778
33 8.67 12.51 7.5 3.13 2.5 569.585 168.927 0.777
34 6.25 17.5 12.33 2.5 7.13 566.05 170.472 0.777
35 8.75 16.86 11.41 7.5 2.5 566.354 170.276 0.776
36 8.75 12.5 7.5 7.5 5.2 566.301 170.283 0.775
37 8.75 12.5 7.5 6.57 4.4 566.698 169.834 0.769
38 6.25 12.55 7.81 7.5 7.5 568.548 168.686 0.761
39 8.75 12.5 7.5 6.02 4.52 566.322 169.462 0.757
40 6.25 12.52 7.5 6.4 7.35 567.338 168.764 0.751
41 8.09 17.5 12.5 5.21 5.29 565.689 169.213 0.745
42 6.48 12.93 7.86 7.5 7.5 564.368 167.691 0.696
43 6.25 16.05 10.41 7.5 7.5 561.534 166.893 0.649
44 6.25 14.33 9.71 7.5 7.5 559.388 166.317 0.614
45 6.25 12.52 7.5 2.5 5.69 555.082 167.166 0.59
solver show the desirability for tensile strength and hardness. Bar graph shows the overall
desirability function of the responses (tensile strength and hardness). Desirability varies from
124
0 to 1 depending upon the closeness of the response towards target. The bar graph shows
how well each variable satisfies the criterion, values close to one are considered good.
T 12.50
T
17.50"
B:CaF2 = 17-50
T 2.50
T
7.50
D:NIO = 7.50
T
Tensile Slmngtfi =576,118
N>rdne8B= 173.10!
Figure 5.39: Ramp Function Graph of Desirability for Tensile Strength and Hardness
Desirability
Figure 5.40: Bar Graph of Desirability for Tensile Strength and Hardness
Model 4: Tensile strength, elongation, impact strength and hardness The ranges and
goals of flux constituents used as input parameters, and the response characteristics viz.
125
tensile strength, elongation, impact strength and hardness are given in Table 5.27. Tensile
strength has been assigned an importance of 5 relative to elongation, impact strength and
hardness with an importance of 3.
Table 5.27: Range of Constituents and Responses for Desirability (Tensile strength,
elongation, impact strength and hardness)

Elongation maximize 26.52 29.58 3
Hardness maximize 157 174 3
Impact
strength maximize 60 99 3
Tensile
strength maximize 520 580 5
5.28) for tensile strength, elongation, impact strength and hardness using Design expert
software. The set of conditions possessing highest desirability value is selected as optimum
condition for the desired responses. Table 5.28 shows the optimal set of condition with
higher desirability function required for obtaining desired response characteristics imder
specified constraints.
126
Table 5.28: Set of Optimal Solutions for Tensile strength, elongation, impact strength
and hardness)
Fe- Impact Tensile

0. MnO CaFz MgO NiO Elongation Hardness Desirability
Cr strength strength
8.75 17.5 12.47 5.58 2.5 29.4282 172.484 99.0012 577.066 0.952 Selec
8.75 17.5 12.5 7.5 2.5 29.3797 173.106 107.232 576.11 0.951
8.75 17.5 12.46 6.35 2.5 29.4008 172.688 102.47 576.527 0.95
8.75 17.5 12.5 6.99 3.08 29.3639 172.848 105.589 575.79 0.945
8.75 17.41 12.5 7.42 2.5 29.3553 172.942 106.686 575.628 0.944
8.71 17.49 12.5 7.5 2.5 29.3399 172.847 107.092 575.332 0.94
8.74 17.5 12.5 7.5 3.56 29.3284 172.936 107.246 575.081 0.939
8.75 17.5 12.5 7.35 3.88 29.3223 172.878 106.668 574.954 0.937
8.75 17.5 12.5 7.22 3.82 29.3246 172.821 106.308 575.002 0.937
8.75 17.5 12.5 4.21 2.51 29.4-759 172.12 92.1337 578.005 0.917
8.75 17.5 12.1 6.26 2.75 29.2614 171.853 101.721 573.787 0.913
8.75 17.5 12.26 7.45 4.21 29.2253 172.398 106.272 573.043 0.912
8.75 17.5 12.5 6.69 5.43 29.2208 172.15 103.248 572.949 0.908
8.71 17.5 12.48 7.45 6.54 29.1843 172.424 101.724 572.189 0.905
8.75 17.5 12.5 7.5 7.44 29.173 172.538 99.0528 571.941 0.904
8.75 17.44 12.49 7.5 6.95 29.1696 172.463 100.632 571.885 0.902
8.75 17.5 12.49 7.4 7.25 29.1688 172.438 99.6946 571.867 0.902
8.75 17.49 12.5 4.08 3.21 29.3861 171.602 93.115 576.249 0.901
8.75 17.38 12.45 7.5 7.37 29.1234 172.26 99.0016 570.965 0.89
8.58 17.5 12.5 5.29 3.98 29.1735 171.034 98.9998 572.078 0.885
8.75 17.5 12.5 5.94 6.48 29.1012 171.189 99.7597 570.612 0.873
8.75 17.5 12.5 2.74 2.98 29.4426 171.271 85.3256 577.366 0.856
7.98 17.5 12.5 7.5 6.04 28.9496 170.312 97.6248 567.63 0.825
6.34 17.49 12.5 7.5 7.02 29.286 170.683 78.6689 574.205 0.769
6.25 17.5 12.5 7.5 7.31 29.4158 171.324 75.7652 576.734 0.769
6.33 17.5 12.5 7.5 6.8 29.233 170.379. 80.051 573.174 0.769
8.65 17.5 12.45 4.51 7.5 28.832 168.982 94.1703 565.387 0.769
6.29 17.5 12.48 7.5 7.05 29.3103 170.774 78.0013 574.679 0.768
6.77 17.5 12.49 7.5 7.03 29.1227 170.153 82.5623 571.007 0.767
6.26 17.5 12.5 7.24 7.45 29.4224 171.483 75.196 576.869 0.766
6.25 17.29 12.5 7.34 6.89 29.2142 170.306 79.0625 572.799 0.757
6.25 17.5 12.49 5.7 6.68 29.0909 170.354 80.5907 570.458 0.75
6.28 17.5 12.5 4.17 7.46 29.1169 171.124 76.2012 571.011 0.725
[ 6.25 17.5 12.5 3.82 6.48 28.9114 170.237 80.2146 567.01 0.716
6.25 17.5 12.5 2.87 5.46 28.6689 169.437 81.9729 562.295 0.677
) 8.75 12.5 7.5 6.28 4.89 28.8663 169.503 74.1728 565.885 0.647
127
Contd.
17 8.75 12.52 7.5 6.25 4.98 28.8536 169.426 74.4669 565.636 0.647
18 8.75 12.5 7.5 6.08 4.81 28.8657 169.394 74.295 565.878 0.647
!9 8.75 12.59 7.5 6.3 4.87 28.8431 169.381 74.674 565.43 0.647
W 6.25 12.5 7.5 2.5 7.5 28.604 168.617 79.9167 560.896 0.641
n 8.75 12.68 7.5 4.33 5.2 28.718 167.702 77.679 563.025 0.632
\2 8.75 13.58 7.8 7.5 4.78 28.584 168.553 79.3674 560.341 0.632
\3, 6.29 12.5 7.5 3.4 7.5 28.6691 168.545 77.097 562.133 0.63
t4 6.26 12.5 7.5 4.01 7.5 28.7469 168.735 74.836 563.631 0.626
\5 6.25 15.66 12.5 5.07 7.5 28.6449 168.307 77.41 561.642 0.626
\6 8.73 12.5 7.69 4.97 6.56 28.5936 167.475 78.4682 560.559 0.614
\1 6.25 13.21 7.91 3.57 7.5 28.4521 167.346 77.6291 557.88 0.583
solver show the desirability for tensile strength, elongation, impact strength and hardness.
Bar graph shows the overall desirability function of the responses (tensile strength,
elongation, impact strength and hardness). Desirability varies from 0 to 1 depending upon the
closeness of the response towards target. The bar graph shows how well each variable
satisfies the criterion, values close to one are considered good.
1 T 1 — T
625 875 12S1 \ 17.50
\Mno=are BC*2= 1750
1
7.5D
T
1250
1
250
\ 1
750
CMgOîafl
&NiO = S5fl
T
250
1
7a
1
2GSZ
— \ ^ *|
2ass
EFM:ri=2S0 Bcnfiaian ^ 2942S2
1
157 174
1
63
1
T
Kar«»s*l72«4
lnpKtEt[erolti=93.(012
1
De3fabil(ty=0.952
I
1
S2D 5S0
Ten* *B\]Ih= 577.068
Figure 5.41: Ramp Function Graph of Desirability for Tensile strength, Elongation,
Impact Strength and Hardness
128
Desirability
=F
Figure 5.42: Bar Graph of Desirability for Tensile strength, Elongation,

Impact Strength and Hardness
Table 5.29 gives point prediction of optimal responses at optimal setting of parameters. The
95% CI (confidence interval) is the range in which one can expect the process average to fall
into 95% of the time. The 95% PI (prediction interval) is the range in which one can expect
any individual value to fall into 95% of the time. The prediction interval will be larger (a
wider spread) than the confidence interval since one can expect more scatter in individual
values than in averages. Confirmation experiments were conducted at optimal levels and the
results were found to lie within 95% confidence interval.
129
Table 5.29: Point Prediction at Optimal Setting of Responses (Tensile Strength,
Elongation, Impact Strength and Hardness)
Actual value
(average of
three
95% CI confirmation
Response Prediction 95% CI low high 95% PI low 95% PI high experiments)
Tensile
Strength
(MPa) 543.4920455 538.9950883 547.9890026 531.3538308 555.6302602 542.33
Elongation
(%) 29.527375 29.15805744 29.89669256 28.85342386 30.20132614 28.57
Impact
Strength
(Joule) 88.25 85.60747917 90.89252083 81.1172913 95.3827087 87.25
Hardness 162.9 162.173197 163.626803 159.56937 166.23063 . 162.3
5.3.3 Weld joints microstructures
The weld joints are visualized by metallurgical microscope at different magnifications to see
the weld joint microstructures. The procedure for the measurement of microstructures is
given in Section 4.3.4. The weld joint microstructures have acicular ferrite (light areas),
cementite (dark areas) and only a few inclusions as shown in Figures F.1-F.32
(APPENDIX F). The microstructure of base metal is shown in Figure F.33 (APPENDIX F).
The presence of acicular ferrite increases the value of tensile and impact strength.
5.4 Bead Morphology

The weld bead profiles on carbon steel plates for developed agglomerated fluxes during
submerged arc welding process are measured using the optical microscope by the procedure
given in the Section 4.4. The weld bead profiles are shown in Figures G.1-G.32
(APPENDIX G). The bead morphology parameters width, weld bead penetration, weld
reinforcement are measured for each sample and their values are given in Table 5.30.
130
Table 5.30: Bead Profile Parameters of Different Samples
Width Penetration Reinforcement

Flux code
(mm) (mm) (mm)
AGFllOl 25.29 5.5 5.16

AGF1102 23.26 5 5.58
AGF1103 25.8 5.63 4.36
AGF1104 21.36 5.36 3.61
AGF1105 24.24 7 3.07
AGF1106 23.14 6.5 4.1
AGF1107 23.61 5.49 4.1
AGF1108 19.27 4.25 4.55
AGF1109 23.13 5 4.01
AGFUIO 21.13 6 4.85
AGFllU 19.73 6 5.34
AGF1112 18.01 11 6.58
AGF1113 22.01 6 4.62
AGF1114 21.94 3.12 4.03
AGF1115 21 6.93 4.73
AGF1116 20.4 6 5.02
AGF1117 22.81 5.78 4.43
AGF1118 23.34 5.54 2.73
AGF1119 26.32 5.77 5.08
AGF1120 24.2 5.63 4.13
AGF1121 21.55 5.5 3.99
AGF1122 20.55 4.68 3.87
AGF1123 22.87 5.97 3.88
AGF1124 22.98 6.25 4.73
AGF1125 20.55 7 6
AGF1126 21.06 6.5 4.6
AGF1127 22.86 8 4.27
AGF1128 22.1 7.3 4.5
AGF1129 22.3 7.2 4.2
AGF1130 21.9 7.4 4.4
AGF1131 21.8 7.1 4.3
AGF1132 21.92 7 4.2
131
5.4.1 Modeling of weld bead penetration
The modeling of experimental values of weld bead penetration mentioned in Table 5.30 is
discussed in this section.
5.4.1.1 Response surface modeling of penetration
The analysis of variance (ANOVA) without pooling and pooled ANOVA reported in Tables
E.5 (APPENDIX E) and 5.31 show that the model is actually significant. The f-value of
11.02 and the p-value < 0.0001 did, in fact, demonstrate that this regression was statistically
significant at 99% confidence level. The values of R^ and adjusted R^ are 94.79% and
85.32% respectively. This means that regression model established an excellent relationship
between the agglomerated alloying constituents and the response penetration. The associated
p-value for the model is less than 0.05 (i.e. a=0.05, or 95% confidence), and indicates that
the model is considered to be statistically significant. As desired the lack-of-fit term was not
significant. Further factors such as CaF2, MgO , NiO, interaction effect of factor MnO with
factor CaF2, interaction effect of factor MnO with factor MgO, interaction effect of factor
MnO with factor NiO, interaction effect of factor CaF2 with factor MgO, interaction effect of
factor CaF2 with factor NiO, interaction effect of factor CaF2 with factor Fe-Cr, interaction
effect of factor MgO with factor NiO, interaction effect of factor MgO with factor Fe-Cr,
squared term of factor MnO, squared term of factor CaF2, squared term of factor MgO and
squared term of factor NiO have significant effects.
Table 5.31: Pooled ANOVA for Penetration

Source Squares Freedom Square f-value Prob>f
Model 51.8 17 3.05 11.02 < 0.0001
A-Mno 0.027 0.027 0.096 0.7607
B-CaF2 1.63 1.63 5.91 0.0291
C-MgO 1.42 1.42 5.14 0.0397
D-NiO 1.44 1.44 5.21 0.0386
E-Fe-Cr 0.18 0.18 0.65 0.4329
AB 1.85 1.85 6.69 0.0215
AC 7.26 7.26 26.28 0.0002
132
Contd.
AD 1.38 1.38 4.99 0.0422
BC 2.59 2.59 9.38 0.0084
BD 10.69 10.69 38.68 < 0.0001
BE 4.88 4.88 17.67 0.0009
CD 3.71 3.71 13.41 0.0026
CE 2.61 2.61 9.44 0.0083
A' 3.31 3.31 11.96 0.0038
B' 3.11 3.11 11.26 0.0047
C' 6.72 6.72 24.32 0.0002
D' 1.46 1.46 5.27 0.0377
Residual 3.87 14 0.28
0.1315 not
Lack of Fit 3.24 9 0.36 2.84 Significant
Cor Total 55.67 31
Mean =6.17 AdjR-Squared = 0.8461
C.V.% = 8.52 PredR-Squared = 0.4176
After eliminating the non-significant terms, the final response equation for penetration is
given as:
Penetration (mm) = - 33.27750+3.23933 * MnO+1.27833 * CaF2+4.07433 * MgO

-1.44367 * NiO+0.64533 * Fe-Cr+0.10880 * MnO * CaFi
-0.21560 * MnO * MgO+0.094000 * MnO * NiO
- 0.064400 * CaF2 * MgO+0.13080 * CaFj * NiO
- 0.088400 * CaFa * Fe-Cr-0.077000 * MgO * NiO
+ 0.064600 * MgO * Fe-Cr-0.21413 * MnO2-0.051933 * CaF2^
-0.076333 * Mg02-0.035533 * Nio2
(5.5)
The normal probability plot of residuals for penetration is shown in Figure 5.43. The
residuals fall on a straight line, which means that the errors are normally distributed. The
model for penetration resulted in good agreement between actual and predicted values. The
comparisons between actual penetration values and predicted values are shown in Figure
5.44.
133
95
>. 90
3 80
(0 70
o
60
sS
"TO 30
E 20
o
z 10
5
-3.00 -2.00 0.00 3.00
Figure 5.43: Normal Probability of Residuals for Penetration

12.00 —
/^n
10.00 —
8.00 —
6.00 —
4.00 —
2.00 —
1 1 1 1 1 1
2.00 4.00
Actual
Figure 5.44: Predicted Penetration vs. Actual Value Penetration(mm)
134
5.4.1.2 Effect efflux constituents on penetration
In the present study, the behaviour of developed agglomerated fluxes with respect to the
different constituents was described. To understand the concept of weld bead penetration the
behaviour of viscosity with respect to different constituents should be understood. The
viscosity also affects weld bead morphology [4]. A flux with a high viscosity will tend to
confine the molten weld pool, thus increasing the heat input for a given area and resulting in
deeper penetration. Viscosity of basic fluxes has been reported to decrease with increasing
percentage of CaFi, MgO, MnO and NiO [30, 32, 34, & 15]. Figures 5.45-5.47 illustrate that
the penetration decreases at the higher values of CaF2, MgO, MnO and NiO, but in case of
Fe-Cr, it remains constant as shown in Figure 5.48. Therefore, the study concludes that the
developed agglomerated flux constituents follow the same trends as discussed by the
investigators.
I
I
I
§
04
MM
A: MnO B: CaF2
«E iza
Figure 5.45: Effect of CaF2 and MnO on penetration
135
a
a
1IA
C:MgO B:CaF2
Z9 tza
Figure 5.46: Effect of MgO and CaFi on Penetration
S
S
o
u
a
OH
B: CaF2
D:NiO
Figure 5.47: Effect of NiO and CaF2 on Penetration
136
i
I
Is
B: CaF2
E: FeCr
Figure 5.48: Effect of Fe-Cr and CaFa on Penetration
5.4.2 Modeling of weld bead reinforcement

The analysis of variance (ANOVA) of pooled ANOVA reported in Table 5.32 shows that the
model is actually significant. The f-value of 37.51 and the p-value < 0.0001 did, in fact,
demonstrate that this regression was statistically significant at 99% confidence level. The
values of R^ and adjusted R^ are 95.38 % and 92.83 % respectively. This means that
regression model demonstrates an excellent relationship between the agglomerated alloying
constituents and the response penetration. The associated p-value for the model is less than
0.05 (i.e. a=0.05, or 95% confidence), and this indicates that the model is statistically
significant. As desired the lack-of-fit term was not significant. Further factors such as A-
MnO, C-MgO, D-NiO, AE, interaction effect of factor MnO with factor E-Fe-Cr, interaction
effect of factor B-CaF2 with factor D-NiO, squared term of factor MnO, squared term of
factor B-CaF2, squared term of factor D-NiO and squared term of factor E-Fe-Cr have
significant effects.
137
Table 5.32: Pooled ANOVA for Weld Bead Remforcement
Degree
Sum of of f-Value Prob > f
Source Mean Squares
Squares Freedo
m
Model 13.41 11 1.22 37.51 < 0.0001 (Significant)
A-MnO 0.72 0.72 22.09 0.0001
B-CaF2 0.041 0.041 1.26 0.2755
C-MgO 0.68 0.68 21.03 0.0002
D-NiO 1.25 1.25 38.37 < 0.0001
6.28E-
E-Fe-Cr 2.04E-04 2.04E-04 03 0.9376
AE 1.51 1.51 46.37 < 0.0001
BD 1.08 1.08 33.13 < 0.0001
A^ 3.73 3.73 114.84 < 0.0001
B^ 0.55 0.55 16.98 0.0005
D^ 1.02 1.02 31.4 < 0.0001
E^ 2.92 2.92 89.73 < 0.0001
Residual 0.65 20 0.032
Lack of Fit 0.4 15 0.026 0.52 0.8485 (not Significant)
Pure Error 0.25 5 0.051
Cor Total 14.06 31
Std.Dev.= 0.18 R-Squared = 0.9538
C.V. % = 3.88 Pred R-Squared = 0.8778
PRESS = 1.72 Adeq Precision = 26.876
Final equation in terms of actual factors:

Reinforcement (mm) = +21.27958-2.78267 * MnO-0.88000 * CaF2-0.067500 * MgO
-0.23400 *NiO+0.23267 * Fe-Cr
- 0.098200 * MnO *Fe-Cr+0.041500 * CaF2 * NiO
+ 0.22747 * Mn02 +0.021867 * CaFf -0.029733 * NiO^
+ 0.050267 * Fe-Cr^
(5.6)
The normal probability plot of residuals for reinforcement is shown in Figure 5.49. The
residuals fall on a straight line, which means that the errors are normally distributed. The
model for reinforcement resulted in good agreement between actual and predicted values.
138
The comparisons between actual reinforcement values and predicted values are shown in
Figure 5.50.
Actual
Figure 5.49: Normal Probability of Residuals for Reinforcement (mm)
Figure 5.50: Predicted Reinforcement vs. Actual Value Reinforcement (mm)
139
5.4.2.1 Effect of flux constituents on reinforcement
Figures 5.51-5.53 depict the response surfaces for bead reinforcement with the interaction of
flux constituents CaFi and MnO, CaF2 and MgO and CaFi and NiO. In 3-D figures
reinforcement tends to increase at the higher percentage of CaFa, MnO, MgO and NiO. The
penetration at higher percentage decreases as suggested by different investigators and this
higher percentage favors to reinforcement. The lower percentage of these constituents gives
good penetration and less reinforcement which is desirable in the welding and it also saves
welding cost. The reinforcement in case of Fe-Cr remains constant as shown in Figure 5.54.
c
0)
17.50
16.50
A: MnO B: CaF2
Figure 5.51: Effect of Cap2 and MnO on Reinforcement
140
c
<u
C: Mgo B: CaFi
Figure 5.52: Effect of MgO and CaFi on Reinforcement
E
<u
p
<u
D:NiO B: CaF2
Figure 5.53: Effect of NiO and CaFi on Reinforcement
141
c
0)
E
(U
E: FeCr B: CaF2
Figure 5.54: Effect of Fe-Cr and CaFi on Reinforcement
5.4.3 Modeling of weld bead width

The analysis of variance (ANOVA) of pooled ANOVA reported in Table 5.33 shows that
the model is actually significant. The f-value of 37.51 and the p-value < 0.0001 did, in fact,
demonstrate that this regression was statistically significant at 99% confidence level. The
values of R^ and adjusted R^ are 90.55%) and 84.58 % respectively. This means that
regression model gives an excellent relationship between the agglomerated alloying
constituents and the response penetration. The associated p-value for the model is less than
0.05 (i.e. a=0.05, or 95% confidence), and indicates that the model is considered to be
statistically significant. As desired the lack-of-fit term was not significant. Further factors
such as B-CaF2, D-NiO, E- Fe-Cr, interaction effect of factor A-MnO with factor B-CaF2,
interaction effect of factor A-MnO with factor E- Fe-Cr, interaction effect of factor C-MgO
with factor E- Fe-Cr, squared term of factor B-CaF2, squared term of factor C-MgO, squared
term of factor D-NiO and squared term of factor E-Fe-Cr have significant effects.
142
Table 5.33: Pooled ANOVA for Weld Bead width
Sum of Degree of f-
Source Squares Freedom Mean Squares Value Prob > f
Model 100.11 12 8.34 15.17 < 0.0001 (Significant)
A-MnO 1.56 1.56 2.84 0.1084
B-CaF2 9.28 9.28 16.87 0.0006
C-MgO 0.29 0.29 0.53 0.4762
D-NiO 21.51 21.51 39.12 < 0.0001
E-Fe-Cr 6 6 10.91 0.0037
AB 4.88 4.88 8.88 0.0077
AE 2.67 2.67 4.86 0.04
CE 9.8 9.8 17.82 0.0005
B^ 14.99 14.99 27.27 < 0.0001
C^ 5.34 5.34 9.71 0.0057
D^ 6.08 6.08 11.07 0.0035
E^ 15.8 15.8 28.74 < 0.0001
Residual 10.45 19 0.55
Lack of
Fit 8.44 14 0.6 1.5 0.3449 (not Significant)
Cor Total 110.55 31
Std. Dev. 0.74 R-Squared 0.9055
Mean 22.25 Adj R-Squared 0.8458
C.V. % 3.33 Pred R-Squared 0.7342
PRESS 29.39 Adeq Precision 16.967
Final equation in terms of actual factors:
Width (mm) = + 36.89833+1.79400 * MnO-2.34217 * CaF2+0.69033 * MgO
+ 0.34750 * NiO-1.26283* Fe-Cr-0.17680 * MnO * CaFi
+ 0.13080 * MnO * Fe-Cr+0.12520 * MgO * Fe-Cr+0.11398* CaFj^
- 0.068017 * MgO2-0.072617 * NiO^-O.l 1702* Fe-Cr^

(5.7)
143
The normal probability plot of residuals for width is shown in Figure 5.55. The residuals fall
on a straight line, which means that the errors are normally distributed. The model for width
resulted in good agreement between actual and predicted values. The comparisons between
actual width values and predicted values are shown in Figure 5.56.
Figure 5.55: Normal Probability of Residuals for Width (mm)
144
28.00
Actual
Figure 5.56: Predicted Width vs. Actual Value Width (mm)
5.4.3.1 Effect efflux constituents on bead width

Figures 5.57-5.59 depict the response surfaces for weld bead width with the interaction of
flux constituents CaFa and MnO, CaFi and MgO and CaF2 and NiO. In 3-D figures width
tends to decrease at the higher percentage of CaFa, MnO, MgO and NiO. The lower
penetration and higher reinforcement at higher percentage of these constituents do not affect
the bead width. The bead width upto some extent remains constant in case of Fe-Cr but it is
reported that at higher percentage bead width decreases as shown in Figure 5.60
145
17.50
A: MnO B: CaF2
Figure 5.57: Effect of CaFz and MnO on Bead Width
C:MgO B:CaF2
Figure 5.58: Effect of MgO and CaFj on Bead Width
146
D:NiO B: CaF2
Figure 5.59: Effect of NiO and CaFi on Bead Width
E: FeCr B: CaF2
Figure 5.60: Effect of Fe-Cr and CaFj on Bead Width
147
5.4.4 Optimization of bead morphology using desirability function
In this section the optimization of bead morphology parameters viz penetration,

reinforcement and weld bead width using desirability is discussed.
5.4.4.1 Single response optimization of bead morphology using desirability function

The individual response characterstics — penetration , reinforcement, and bead width — for
the optimization are given in Tables 5.34- 5.36. Goals and limits were established for each
response individually in order to accurately determine their impact on individual desirability.
A maximum or minimum level is provided for each response characterstic which has to be
optimized.
Weights are assigned to give added emphasis to upper / lower bounds or to emphasize a
target value. The default value " 1 " of weight is assigned to a goal to adjust the shape of its
particular desirability function. The default value " 3 " is selected for importance in order to
give equal importance to all goals.
Table 5.34: Range of Constituents and Penetration for Desirability

& Response Goal Limit Limit Weight Weight Importance
B-CaFi is in range 12.5 17.5 3
Penetration
(mm) maximize 3.12 11 1 1 3
Table 5.35: Range of Constituents and Reinforcement for Desirability
Upper Lower Upper Lower

Name Goal Limit Limit Weight Weight Importance
A-MnO is in range 6.25 8.75 1 1 3
B-CaF2 is in range 12.5 17.5 1 1 3
C-MgO is in range 7.5 12.5 1 1 3
Contd.
148
D-NiO is in range 2.5 7.5 1 1 3
E-Fe-Cr is in range 2.5 7.5 1 1 3
Reinforcement
(mm) minimize 3 6 1 1 3
Table 5.36: Range of Constituents and Width for Desirability

Width
(mm) minimize 19 27 3
5.4.4.2 Optimal solutions using desirability function

A set of 47 solutions is derived for the specified design for response characteristics —
penetration , reinforcement, and bead width — using Design Expert Software. The set of
desired response. The optimal conditions for the different responses are given in
Tables 5.37-5.39.
Table 5.37: Optimal Solutions for Desirability (Penetration)
No. MnO CaF2 MgO NiO Fe-Cr Penetration Desirability

1. 8.75 17.5 7.5 7.5 2.5 10.5967 0.949 Selected
2 8.71 17.5 7.58 7.49 2.5 10.5344 0.941
3 8.67 17.5 7.55 7.5 2.5 10.53 0.94
4 8.75 17.5 7.5 7.5 2.67 10.5239 0.94
5 8.75 17.5 7.5 7.5 2.72 10.5029 0.937
6 8.6 17.5 7.59 7.5 2.5 10.4766 0.934 "
7 8.75 17.2 7.5 7.5 2.5 10.3864 0.922
8 8.75 17.26 7.76 7.5 2.51 10.2914 0.91
9 8.75 17.5 7.57 6.97 2.5 10.255 0.905
149
Contd.
10 8.75 17.5 7.5 7.16 2.85 10.2519 0.905
11 8.75 17.5 7.83 7.5 2.97 10.2395 0.903
12 8.55 17.5 7.52 7.06 2.5 10.2392 0.903
13 8.34 17.5 7.51 7.5 3.16 10.0789 0.883
14 8.75 17.5 7.5 6.59 2.5 10.0607 0.881
15 8.75 16.94 7.61 7.13 2.5 9.96046 0.868
16 8.75 17.5 7.5 7.5 4.08 9.93589 0.865
17 8.75 17.5 7.5 6.42 2.56 9.93128 0.864
18 8.74 17.49 7.5 7.5 4.17 9.89039 0.859
19 8.75 17.5 7.5 6.46 2.71 9.88872 0.859
20 8.75 16.99 7.5 7.5 3.76 9.75729 0.842
21 8 17.5 8.86 7.5 2.5 9.5114 0.811
22 8.75 17.5 9.21 7.29 2.5 9.42884 0.801
23 7.27 17.5 7.5 7.5 2.55 9.38621J 0.795
24 8.75 17.5 7.5 7.4 5.47 9.30286 0.785
25 8.75 17.5 7.54 7.5 5.97 9.13747 0.764
26 8.74 17.45 9.71 7.5 2.5 9.09046 0.758
27 8.75 16.53 7.5 7.5 4.95 9.07011 0.755
28 7.27 17.5 8.83 7.31 2.5 8.97705 0.743
29 6.25 12.5 12.4 2.5 7.5 8.95785 0.741
30 6.25 12.5 12.47 2.65 7.5 8.92832 0.737
31 6.25 12.56 12.27 2.5 7.5 8.88605 0.732
32 6.25 12.62 12.48 2.5 7.36 8.88224 0.731
33 6.25 12.5 12.5 2.83 7.5 8.86665 0.729
34 6.25 12.5 12.12 2.5 7.47 8.8557 0.728
35 8.75 16.66 7.5 7.5 5.8 8.8471 0.727
36 6.46 12.5 12.12 2.5 7.5 8.7592 0.716
37 6.46 12.5 12.5 2.93 7.5 8.7158 0.71
38 6.25 12.5 12.5 2.5 6.61 8.6825 0.706
39 6.72 12.5 12.41 2.5 7.5 8.67153 0.705
40 6.38 12.5 11.73 2.5 7.5 8.66259 0.703
41 7.91 17.5 7.5 5.03 2.5 8.64149 0.701
42 6.25 12.5 11.58 2.5 7.39 8.61176 0.697
43 6.25 12.8 12.5 3.22 7.5 8.59164 0.694
44 6.92 17.23 7.5 6.66 2.5 8.53499 0.687
45 6.25 12.6 12.5 2.5 5.91 8.40191 0.67
46 8.35 16.49 7.5 7-.5 6.47 8.36113 0.665
47 8.75 17.5 9.77 4.19 2.5 7.40919 0.544
150
Table 5.38: Optimal Solutions for Desirability (Reinforcement)
S.No. MnO CaF2 MgO NiO Fe-Cr Reinforcement Desirability

1 7.11 17.5 12.5 2.5 4.64 3.4136 0.862 Selected
2 7.13 17.5 12.5 2.5 4.62 3.41363 0.862
3 7.07 17.5 12.5 2.5 4.52 3.41426 0.862
4 7.16 17.5 12.5 2.5 4.72 3.41476 0.862
5 7.23 17.5 12.5 2.5 4.81 3.41615 0.861
6 7.24 17.5 12.5 2.5 4.63 3.41733 0.861
7 6.95 17.5 12.5 2.5 4.46 3.41838 0.861
8 7.21 17.49 12.5 2.51 4.91 3.41885 0.86
9 7.25 17.48 12.5 2.5 4.52 3.42015 0.86
10 6.91 17.5 12.5 2.5 4.41 3.42163 0.859
11 7.11 17.05 12.49 2.5 4.65 3.42351 0.859
12 7.04 17.46 12.5 2.53 4.58 3.42484 0.858
13 7.07 17.5 12.5 2.56 4.46 3.43524 0.855
14 7.02 17.3 12.15 2.5 4.37 3.44337 0.852
"15 6.69 17.49 12.5 2.5 4.19 3.44591 0.851
16 6.76 17.5 12.5 2.5 3.72 3.45319 0.849
17 7.09 17.46 12.5 2.63 4.49 3.45944 0.847
18 7.5 17.5 12.5 2.5 5.64 3.45972 0.847
19 7.45 16.77 12.5 2.5 5.69 3.47928 0.84
20 7.06 17.5 12.5 2.68 4.93 3.48002 0.84
21 7.59 16.95 12.5 2.5 5.83 3.49312 0.836
22 7.15 17.29 12.5 2.79 4.94 3.51482 0.828
23 7.11 17.06 10.92 2.5 4.8 3.5311 0.823
24 7.18 17.5 12.5 2.89 4.8 3.54558 0.818
25 7.22 17.5 10.57 2.5 4.89 3.5467 0.818
26 6.52 17.5 12.5 2.5 2.76 3.5614 0.813 •
27 6.95 17.43 10.13 2.5 3.69 3.61049 0.797

28 7.26 14.82 12.47 2.5 4.21 3.62227 0.793
29 7.29 16.89 10.26 2.5 3.95 3.623 0.792
151 Contd.
30 7.61 17.4 8.91 2.5 5.58 3.71134 0.763
31 7.06 13.6 12.5 2.77 3.7 3.87595 0.708
32 7.11 15.43 12.5 4.24 5.01 3.89543 0.702
33 7.12 17.5 12.5 4.16 4.68 3.90287 0.699
34 7.08 13.13 12.5 7.5 4.59 3.94686 0.684
35 7.01 12.97 12.5 7.5 4.32 3.95063 0.683
36 6.93 12.77 12.5 7.5 4.58 3.9552 0.682
37 7.15 13.53 12.48 7.5 4.89 3.95626 0.681
38 7.28 13.5 12.5 7.5 4.92 3.95714 0.681
39 7.24 12.5 12.5 7.46 4.64 3.96214 0.679
40 7.16 12.51 12.36 7.5 5.02 3.9677 0.677
41 7.11 14.01 12.5 7.5 4.88 3.97166 0.676
42 7.06 13.99 12.5 7.5 4.88 3.97225 0.676
43 6.87 12.5 12.5 7.5 4.85 3.97296 0.676
44 6.88 15.49 12.5 4.6 5.19 3.98049 0.673
45 7.18 13.24 12.5 6.79 4.63 4.02642 0.658
46 7.13 14.41 12.5 6.01 4.69 4.04576 0.651
47 6.95 14.19 12.5 5.97 4.43 4.05199 0.649
Table 5.39: Optimal Solutions for Desirability (Width)
S.No. MnO CaF2 MgO NiO Fe-Cr Width Desirability

1 8.75 16.83 12.5 7.35 2.56 18.9979 1 Selected
2 8.66 16.8 7.53 7.41 7.49 18.9865 1
3 8.74 16.9 12.42 7.5 2.55 18.9369 1
4 8.72 17.15 12.47 7.36 2.52 18.9917 1
5 8.3 16.74 7.51 7.48 7.5 18.9837 1
6 8.67 17.48 7.55 7.47 7.5 18.969 1
7 8.71 16.97 12.35 7.49 2.51 18.9856 1
8 8.09 16.72 7.5 7.5 7.5 18.9996 1
9 8.08 16.36 7.5 7.49 7.5 19.0062 0.999
152 Contd.
10 7.87 16.27 7.5 7.5 7.5 19.0281 0.996
11 6.25 15.09 7.5 7.5 7.5 19.0449 0.994
12 8.75 16.17 7.5 7.5 7.4 19.0456 0.994
13 6.37 15.15 7.5 7.49 7.5 19.0621 0.992
14 8.75 17.44 12.21 7.46 2.5 19.0725 0.991
15 6.76 15.35 7.5 7.5 7.5 19.0788 0.99
16 6.81 15.55 7.5 7.5 7.5 19.0794 0.99
17 7.07 15.73 7.5 7.5 7.5 19.0795 0.99
18 7.26 15.59 7.5 7.5 7.5 19.086 0.989
19 6.26 15.18 7.5 7.5 7.46 19.0914 0.989
20 7.1 15.68 7.5 7.48 7.5 19.0985 0.988
21 7.8 15.53 7.5 7.5 7.5 19.1079 0.987
22 6.31 15.16 7.5 7.37 7.5 19.1435 0.982
23 6.33 14.3 7.5 7.49 7.49 19.1535 0.981
24 6.79 16.6 7.5 7.5 7.5 19.205 0.974
25 6.29 15.25 7.5 7.23 7.5 19.2435 0.97
26 8.75 16.61 7.52 7.5 7.11 19.2445 0.969
27 8.75 15.56 12.5 7.22 2.5 19.2896 0.964
28 8.75 16.8 12.22 7.5 2.85 19.3027 0.962
29 8.75 15.85 7.5 7.1 7.5 19.3155 0.961
30 8.35 16.66 8.15 7.5 7.5 19.3221 0.96
31 8.75 17.5 8.17 7.34 7.5 19.3799 0.953
32 8.43 15.87 12.5 7.5 2.83 19.4689 0.941
33 8.75 16.65 12.38 7.5 3.33 19.5542 0.931
34 7.77 17.28 12.5 7.5 2.5 19.7518 0.906
35 8.75 17.44 7.5 5.86 7.5 19.9023 0.887
36 8.75 15.89 9.63 7.5 7.5 20.0143 0.873
37 8.73 17.08 7.5 7.5 5.91 20.0622 0.867
38 8.75 17.03 12.5 7.5 7.5 20.212 0.849
39 8.75 17.47 12.47 5.03 2.5 20.2615 0.842
40 8.56 16.91 12.38 7.5 7.44 20.2828 0.84
153
Contd.
41 6.3 15.22 12.5 7.5 7.5 20.392 0.826
42 6.69 15.58 12.42 7.5 7.27 20.5556 0.806
43 8.75 17.03 7.5 7.5 4.09 20.6937 0.788
44 6.25 15.59 7.5 4.09 7.5 20.7529 0.781
45 6.25 16.61 12.47 5.44 7.5 21.8641 0.642
46 6.25 15.14 12.5 2.5 7.49 22.2847 0.589
47 8.75 17.11 8.23 2.62 2.5 22.4559 0.568
The ramp function graphs and bar graphs (Figures 5.61-5.66) drawn using Design Expert
solver show the desirability for each factor and each response. The dot on each ramp reflects
the factor setting or response prediction for that response characteristic. The height of the dot
shows how desirable it is. A linear ramp function is created between the low value and the
goal or the high value and the goal as the weight for each parameter was set equal to one. Bar
graphs show the individual/ partial desirability functions (d) of each of the responses
(penetration, reinforcement, and bead width); di varies from 0 to 1 depending upon the
closeness of the response towards target [89]. The bar graph shows how well each variable
satisfies the criterion: values close to one are considered good. Table 5.40 reports the final set
of optimum levels of various process parameters and the predicted values of various response
characteristics.
Table 5.40: Optimal Sets of Process Parameters Using RSM and Desirability Function
Predicted
Constituents - > A B C D E
Desirability optimal
Response | MnO CaF2 MgO NiO Fe-Cr
response
Penetration (mm) 0.949 8.75 17.5 7.5 7.5 2.5 10.5967

Reinforcement
(mm) 0.862 7.11 17.5 12.5 2.5 4.64 3.4136
Width (mm) 1 8.75 16.83 12.5 7.35 2.56 18.9979
154
T
C:MBO»7,S0
T
Penetration = 10.5967
Desirabiriy = 0.949
Figure 5.61: Ramp Function Graph of Desirability for Penetration
Desirability
A;Mno
•HP—
C:MgO
D:NiO
E:Fe^r
sai
0.9 «816
Combined
=F =F
Figure 5.62: Bar Graph of Desirability for Penetration
155
C;Mgo = 12.50
Reinforcement = 3,4137
Desirability =0.862
Figure 5.63: Ramp Function Graph of Desirability for Reinforcement
Desirability
B:CaF2
_ 1
iwmmmm\mwim
E:Fe-cr
inforcement
1 1
Combined
~ ^
1 1
^ > ^ ^ ^
i 1 1 1
Figure 5.64: Bar Graph of Desirability for Reinforcement
156
12.50 17.50
B-.CaF2 = 17.15
2.50 7.50
D:NiO = 7.36
T
W(Jth = 18.9917
Figure 5.65: Ramp Function Graph of Desirability for Width
Desirability
C:MgO
ZiZl —p^
I
0.500
—r
1.000
Figure 5.66: Bar Graph of Desirability for Width
157
5.4.4.3 Multiple response optimization of bead morphology using desirability function
In this section multiple response of the bead morphology parameters viz penetration,
reinforcement and bead v^ddth is discussed.
5.4.4.3.1 Model: Penetration, reinforcement and bead width
The ranges and goals of flux constituents used as input parameters and the response
characteristics viz. penetration, reinforcement and bead width are given in Table 5.41.
Penetration has been assigned an importance of 5 relative to reinforcement and bead width
with an importance of 3 because the groove weld should be filled with the minimum passes.
Table 5.41: Range of Constituents and Responses for Desirability (Penetration,

Reinforcement and Bead width)

is in
A-MnO range 6.25 8.75 1 3
is in
B-CaF2 range 12.5 17.5 1 3
is in
C-MgO range 7.5 12.5 1 3
is in
D-NiO range 2.5 7.5 1 3
is in
E-Fe-Cr range 2.5 7.5 1 3
Penetration maximize 3.12 11 1 5
Reinforcement minimize 3 6 1 3
Width minimize 19 27 1 3
5.42) for penetration, reinforcement and bead width using Design expert software. The set of
desired responses. Table 5.42 shows the optimal set of condition with higher desirability
fimction required for obtaining desired response characteristics under specified constraints.
158
Table 5.42: Set of Optimal Solutions for Penetration, Reinforcement and Width
S.No. MnO CaF2 MgO NiO Fe-Cr Penetration Reinforcement Width Desirability
I 1 8.05 16.70 7.50 7.50 5.25 8.6746 4.74184 20.7242 0.629930651 Selected
1 2 8.03 16.72 7.50 7.50 5.20 8.68446 4.73894 20.7574 0.629922426
3 8.01 16.74 7.50 7.50 5.25 8.66366 4.73428 20.7439 0.629852975
1 4 8.03 16.71 7.53 7.50 5.23 8.66939 4.7364 20.7468 0.629782776
' 5 8.04 16.70 7.50 7.49 5.22 8.6796 4.74173 20.742 0.6297139
6 8.01 16.70 7.50 7.49 5.15 8.67869 4.73172 20.7921 0.629649494
1 7 8.06 16.71 7.50 7.49 5.31 8.66431 4.74726 20.6935 0.629497183
J
7.99 16.67 7.50 7.50 4.90 8.74243 4.73202 20.914 0.629465298
8
1 9 8.04 16.91 7.50 7.49 5.21 8.77541 4.77651 20.7649 0.62916208
1 10 8.02 16.56 7.50 7.50 5.62 8.46675 4.70658 20.5298 0.628999859
11 8.06 16.75 7.54 7.50 4.84 8.84265 4.77275 20.916 0.628859688
' 12 8.00 16.68 7.66 7.50 5.00 8.69806 4.72165 20.9072 0.628788162
1 13 8.09 16.56 7.50 7.48 5.30 8.61521 4.73481 20.6928 0.628667687
1 14 8.06 16.52 7.50 7.50 4.87 8.72782 4.73831 20.8857 0.628662793
^ 15 7.94 16.78 7.50 7.50 4.65 8.84652 4.7494 21.0451 0.628606908
1 16 7.99 16.57 7.57 7.50 5.66 8.43584 4.69619 20.5442 0.628332525
1 ^^ 7.91 16.85 7.50 7.50 4.57 8.88086 4.75432 21.0995 0.628066691
18 8.04 16.61 7.83 7.50 5.24 8.58586 4.70245 20.8323 0.627635147
' 19 7.83 16.77 7.50 7.50 4.56 8.78306 4.71403 21.1417 0.627401546
1 20 7.98 16.71 7.50 7.50 4.42 8.92194 4.77509 21.0971 0.627283314
7.99 16.49 7.50 7.50 5.92 8.31846 4.69495 20.3553 0.627061451
1 ^1
22 7.95 17.10 7.50 7.50 4.22 9.17014 4.85286 21.2132 0.6246.13411
1 23 7.84 16.67 7.81 7.50 3.87 8.92366 4.75947 21.3618 0.621711096
7.83 16.90 8.09 7.50 3.78 9.0028 4.7904 21.428 0.619262031
1 ^"1
25 8.28 16.26 7.79 7.50 6.58 8.17088 4.76921 19.9432 0.619176109
1 26 7.86 16.58 8.65 7.50 6.20 8.05203 4.61959 20.6222 0.614777777
1 27 7.90 16.96 8.52 7.50 3.69 8.9682 4.8157 21.4066 0.614695353
28 7.83 16.58 7.50 7.50 6.77 7.94957 4.75865 19.7411 0.61280797
' 29 7.69 15.62 7.50 7.50 5.63 7.79896 4.50264 20.6852 0.612033943
1 30 8.06 17.50 7.50 7.11 5.35 8.77647 4.86573 21.0527 0.608461609
7.58 16.66 9.81 7.50 3.14 8.25246 4.65145 21.3834 0.600843613
1 ^1
32 7.48 16.34 9.48 7.50 2.50 8.32881 4.7291 21.3505 0.596135073
1 33 8.14 17.02 7.50 6.62 4.45 8.77014 4.84989 21.6075 0.594369361
1 ^^ 8.11 15.06 8.70 7.50 7.50 7.39962 4.65111 19.885 0.590104385
35 8.46 15.21 8.23 7.50 7.49 7.57024 4.78396 19.6444 0.589256273
' 36 6.70 13.82 12.50 2.50 7.50 8.04586 4.3201 22.6186 0.585135662
1 37 6.65 13.80 12.41 2.50 7.50 8.07562 4.35252 22.6203 0.583568271
, 38 6.52 13.68 12.50 2.50 7.03 8.11352 4.28458 22.8695 0.582700531
159 Contd.
39 6.63 13.84 12.50 2.53 6.92 7.93482 4.17658 22.8917 0.5818944
40 6.66 13.86 12.04 2.50 7.50 7.97292 4.36308 22.6134 0.5772766
41 6.76 13.74 12.50 3.14 7.50 7.94198 4.41274 22.6336 0.570066106
42 6.46 13.77 12.50 2.51 5.13 7.64189 3.90244 23.3423 0.56922414 -•
43 6.95 13.75 11.45 2.50 7.50 7.75954 4.29859 22.7319 0.567336971
44 6.91 13.99 12.50 3.47 7.50 7.68892 4.3676 22.5416 0.563734638
45 6.33 14.30 12.50 2.50 3.57 7.18895 3.78888 23.0273 0.562957451
46 6.39 14.32 12.50 2.50 3.25 7.09866 3.78756 22.8978 0.562234288
47 7.14 14.46 11.52 4.42 7.50 7.30474 4.42909 22.2311 0.545961971
solver show the desirability for penetration, reinforcement and width. Bar graph shows the
overall desirability function of the responses (penetration, reinforcement and bead width).
Desirability varies from 0 to 1 depending upon the closeness of the response towards target.
The bar graph shows how well each variable satisfies the criterion, values close to one are
considered good.
EFe-c-aa
VMHl>2D.72U
Dea[ability •'0.630
Figure 5.67: Ramp Function Graph of Desirability for Penetration, Reinforcement and
Width
160
Desirability
Figure 5.68: Bar Graph of Desirability for Penetration, Reinforcement and Width
Table 5.43 gives point prediction of optimal responses at optimal setting of parameters. The
95% CI (confidence interval) is the range in which one can expect the process average to fall
into 95% of the time. The 95% PI (prediction interval) is the range in which one can expect
any individual value to fall into 95% of the time. The prediction interval will be larger
(a wider spread) than the confidence interval since one can expect more scatter in individual
values than in averages. Confirmation experiments were conducted at optimal levels and the
results were found to lie within 95% confidence interval.
Table 5.43: Point Prediction at Optimal Setting of Responses (Penetration,

Reinforcement, and Bead Width)
Actual value
(average of
three
95% CI 95% PI 95% PI confirmation
Response Prediction 95% CI low high low high experiments)
Penetration 6.63
(mm) 7.1875 6.788818668 7.586181332 5.991456 8.383543996
Reinforcement 4.3
(mm) 4.18625 4.053312443 4.319187557 3.78743733 4.585062672
22.8
Width (mm) 22.92625 22.37756706 23.47493294 21.2802012 24.57229883
161
5.5 Element Transfer Study
The procedure of elements transfer study is reported in Section 4.5 and percentage of
different elements along with the values of carbon equivalent in different samples by using
developed agglomerated fluxes during submerged arc welding is given in Table 5.44.
Table 5.44: Elements (wt %) in Weld Metal
S.No. Flux Code C (%) Si Mn(%) P (%) S (%) Cr Ni êquivalent

(%) (%) (%)
1 AGFllOl 0.10 0.33 0.87 0.026 0.00886 0.57 0.36 0.3818
2 AGF1102 0.11 0.33 0.78 0.022 0.01108 0.65 0.54 0.3973
3 AGF1103 0.07 0.33 0.37 0.028 0.01079 0.71 0.28 0.2944
4 AGF1104 0.07 0.33 0.64 0.024 0.02145 0.74 0.60 0.3534
5 AGF1105 0.08 0.23 0.31 0.026 0.01564 0.79 0.42 0.3098
6 AGF1106 0.10 0.23 0.54 0.026 0.01516 0.85 0.39 0.3793
7 AGF1107 0.09 0.23 0.54 0.026 0.01267 0.89 0.41 0.3778
8 AGF1108 0.07 0.23 0.35 0.032 0.00810 0.95 0.36 0.3369
9 AGF1109 0.08 0.15 0.30 0.022 0.01395 0.58 0.84 0.2733
10 AGFlllO 0.11 0.18 0.69 0.024 0.00725 0.63 0.79 0.3783
11 AGFllU 0.10 0.18 0.51 0.024 0.02643 0.71 0.65 0.3508
12 AGF1112 0.11 0.15 0.31 0.026 0.01790 0.74 0.83 0.3367
13 AGF1113 0.11 0.15 0.57 0.027 0.01113 0.79 0.95 0.3930
14 AGF1114 0.11 0.15 0.36 0.026 0.00970 0.83 0.88 0.3643
15 AGF1115 0.11 0.15 0.34 0.026 0.02387 0.89 0.78 0.3704
16 AGF1116 0.10 0.15 0.57 0.030 0.00806 0.95 0.90 0.4138
17 AGF1117 0.11 0.15 0.43 0.024 0.01628 0.65 0.59 0.3327
18 AGF1118 0.11 0.15 0.47 0.027 0.01204 0.7 0.61 0.3498
19 AGF1119 0.12 0.10 0.46 0.022 0.01252 0.74 0.46 0.3603
20 AGF1120 0.10 0.11 0.47 0.022 0.02494 0.79 0.47 0.3527
21 AGF1121 0.08 0.14 0.52 0.023 0.02624 0.83 0.52 0.3515
22 AGF1122 0.07 0.19 0.41 0.022 0.02120 0.84 0.41 0.3245
23 AGF1123 0.10 0.23 0.44 0.023 0.01542 0.95 0.44 0.3839
24 AGF1124 0.07 0.19 0.52 0.032 0.01062 0.71 0.52 0.3196
25 AGF1125 0.07 0.14 0.31 0.029 0.01434 0.74 0.31 0.2833
26 AGF1126 0.10 0.28 0.73 0.024 0.02938 0.79 0.73 0.4096
27 AGF1127 0.08 0.14 0.57 0.023 0.02281 0.83 0.67 0.3636
28 AGF1128 0.07 0.14 0.56 0.020 0.02281 0.82 0.65 0.3494
29 AGF1129 0.08 0.12 0.57 L0.023 0.02270 0.93 0.63 0.3818
30 AGF1130 0.07 0.13 0.55 0.022 0.02280 0.83 0.62 0.3486
31 AGF1131 0.08 0.12 0.54 0.023 0.02275 0.81 0.67 0.3538
32 AGF1132 0.08 0.13 0.53 0.023 0.02387 0.87 0.61 0.3630
162
5.5.1 Carbon (C) transfer
The carbon content of all weld-beads ranges from 0.07 to 0.12%. The carbon pick-up has
occurred from CaCOs, which yields CO2 on dissociation. The CO2 being unstable at high
temperature dissociates into CO and O. Oxygen reacts with the element having high affinity
for it, whereas CO, which was present in the vicinity of the weld bead at relatively high
temperature, dissociates into C and O2. This C is picked up by weld metal because of
diffusion. The variations in mechanical properties of weld metal during submerged arc
welding using developed agglomerated fluxes follow usual trends with the percentage of
carbon pick up.
5.5.2 Manganese (Mn) transfer

In this study, Mn content of all weld-beads ranges from 0.30 to 0.87%. The Mn content in the
weld metal is always helpful additive in iron alloys. The presence of Mn works as
desulphurization agent because of its greater affinity with the sulphur than iron. The probable
slag- metal chemical reaction is:
FeS + MnO = MnS + FeO (5.8)
2Mn [Metal] + SiOi (slag) = 2MnO(Slag) + Si [Metal] (5.9)
Where [ ] denotes the concentration in weld metal and () denotes the concentration in slag.
Oxidation of Si02 at slag metal interface takes place according to the following reaction:
Si02(Slag) + 2Fe (Metal) = 2FeO (Slag) + Si (Metal) (5.10)
This reaction shifts towards right at higher temperatures. Electrochemically iron oxide can be
produced at the anodes. Thus it is produced on weld pool with electronegative polarity and
on electrode wire with electropositive polarity. This difference can be attributed to the
combined effect of electrochemical and pyrochemical mechanisms.
5.5.3 Silicon (Si) transfer

In this study. Si content of all weld-beads ranges from 0.10 to 0.33%). This study shows that
the Si content present in weld metal is always lower than that present in original flux. This is
due to the transfer of Si from weld metal into slag. It is evident that Si is more prone to the
163
oxidation losses than Cr and Ni. The Si distribution under basic slag can be described by the
anionic equilibrium as per following reaction:
[Si]+2[0]+2(02-)=(SiO/") (5.11)
and the equilibrium quotient is:
K = [aSiO/"] / [aSi]. [aO]^. [aO^f (5.12)
5.5.4 Chromium (Cr) transfer

In this study, Cr content of all weld-beads ranges from 0.57 to 0.95%). The presence of Cr
always strengthens weld metal and reduces the corrosion activity because it has been
subjected to the oxidation losses, and it is less prone to oxidation than Mn and Si.
5.5.5 Nickel (Ni) transfer

In this study, Ni content of all weld-beads ranges from 0.28 to 1.48%. The presence of Ni in
weld metal shows good hardness because it often improves fracture toughness.
5.5.6 Sulphur (S) transfer

In this study S content of all weld-beads ranges from 0.00725 to 0.02938%. The S transfer to
the weld metal can be explained as per the following reaction:
The sulphur equilibrium between slag and metal can be expressed by ionic equations such as:
[S] + (0'-) = (S'-) + [0] (5.13)
The equilibrium constant for the above equation can be given by:
Ks = [aS^-.] [aO] / [aO^"] [aS] (5.14)

From which the sulphur distribution ratio is simply:
[as2.]/[as]=Ks.[ao2.]/[ao] (5.15)
If Henrian behaviour is assumed:
[Wt. % S^-] / [Wt. % S] = K. [Wt. % O^"] / [Wt. % O] (5.16)
164
Thus, the reaction clarifies that the desulfurization will be promoted when (aO^") is high, i.e.
oxidizing power of flux is high. The content of sulphur in steels is a big source of hot
cracking, and Mn is added to minimize the effect of cracking. Mn has more affinity than S,
so 1 to 2% Mn is added in steel to nullify the effect of Sulphur so the chance of hot cracking
is reduced. The reaction is as follows:
Fe + S + Mn = MnS + Fe (5.17)
5.5.7 Phosphorus (P) transfer

In this study, P content of all weld-beads ranges from 0.02 to 0.032%. The presence of P
gives strength and also improves corrosion strength in weld metal. This behaviour of P
transfer can be explained as per the following equilibrium:
The anionic phosphorous equilibrium is represented by the equation:
2[P] + 5[0] + 3(0^-) = 2[P04^-] (5.18)
for which the equilibrium constant is given as:

Kp = [a P 0 4 ' f / [ap]2.[ ao]'.[ a O ' f (5.19)
From above expression it is clear that the basic slag will promote dephosphorization because
it facilitates higher [a O^"], which promotes the above equation in forward direction.
5.5.8 Effect of element transfer on mechanical properties

Table 5.44 shows the wt% of different elements present in the welded joint by using
developed agglomerated flux (AGF 1101- AGF 1132). The carbon equivalent to every
welded joint is evaluated by the following equation.
Cequivaient =C+Mn/6+Si/24+Ni/40 + Cr/5 (5.20)
The carbon equivalents of weld joints range from 0.273 to 0.413, which can be correlated
with the mechanical properties viz. tensile strength and toughness of welded joints [58].
C, Mn, Si, Ni, and Cr, are the metallic contents, expressed as a percentage. It has been
165
reported [90] that an equivalent carbon higher than 0.45 had a high susceptibility to cold
cracking after welding and more chance to form martensite which is very hard and brittle.
The carbon equivalents to all the welded joints by using developed agglomerated fluxes are
lower than the value of high susceptibility of cold cracking which indicates that these
developed agglomerated fluxes are less prone to cold cracking at the welded joints. The
details of mechanical properties of weld joints using developed agglomerated fluxes and base
metal is given in Tables 5.6 and 5.7 (Section 5.3). The experimental values of the mechanical
properties of the weld metal using developed agglomerated fluxes, which are superior to the
base metal demonstrates that these fluxes may be used commercially in the future.
166
CHAPTER 6
CONCLUSIONS AND SCOPE OF FURTHER RESEARCH
6.1 Conclusions
In the present study agglomerated fluxes for the submerged arc welding are designed,
formulated and developed in the laboratory. The physical characteristics of developed fluxes,
mechanical properties of weld joints, microstructure of weld joints, element transfer analysis
and bead morphology are studied by using these developed agglomerated fluxes during
submerged arc welding. The experimental results are analyzed and correlated with the help of
modeling and standard data. On the basis of results and discussion, the following conclusions
are drawn.
6.1.1 Bulk density, grain Size, flowability and viscosity

The present study concludes the following points:
(a) The Flowability (Ref. Table 5.1) of developed fluxes are good (28 nos.), excellent
(Olno.) and satisfactory (03 nos.). The good results of Flowability are related to the
ease with which the flux can be delivered to the joint preparation ahead of the arc
during welding [40].
(b) The grain size of these developed agglomerated fluxes ranges from 0.355 to 1.18 mm
(Ref. Tables A.l- A.32, APPENDIX A). The range of grain size of these fluxes
provides good weld bead shape and surface quality and can be used with different
range of operating current.
(c) As the bulk density of the fluxes increases, the flux consumption increases
(Ref. Table 5.2). This can be attributed to the fact that increase in bulk density can be
considered as equivalent to increase in specific heat of the developed flux.
(d) Low bulk density fluxes reduce overall welding cost and these fluxes facilitate higher
weld bead penetration.
(e) The newly developed agglomerated fluxes can be used for commercial purposes in
welding industries.
167
(f) The welding behaviour of these fluxes in terms of arc initiation, arc stability and weld
appearance was good (Ref. Figure 4.10).
(g) The viscosity of developed fluxes has a decreasing trend at higher temperatures
(Ref. Figure 5.2).
6.1.2 XRD and DTA

The study identified the quantity of different ions present in the weld pool, different type of
oxides and probable crystalline phases in developed agglomerated fluxes during submerged
arc welding. Thus, the effect of compositions of fluxes on thermo-mechanical behaviour of
weld metal was known and similarly these phases present in developed fluxes also affect the
mechanical properties, bead morphology and element transfer analysis of weld joints.
The curves of differential thermal analysis (DTA) show that the most of the oxides were
stable oxides up to 1000 °C. The ranges of DTA curves were found betweenlOOO to 1400 °C.
6.1.3 Modeling of weld joints and bead morphology

Response surface methodology (RSM) was applied for developing the mathematical models
in the form of multiple regression equations correlating the dependent parameters viz. tensile
strength, elongation, impact strength, hardness, penetration, reinforcement and weld bead
width with the independent parameters (MnO, CaFi, MgO, NiO, Fe-Cr) in development of
agglomerated fluxes for submerged arc welding. Using the model equations, the response
surfaces have been plotted to study the effects of process parameters on the performance
characteristics. From the experimental data of RSM, empirical models were developed and
the confirmation experiments were performed, which were found within 95% confidence
interval. The regression equations also correlated the dependent response with the
independent variables through RSM. The mathematical regression equations of these
dependent parameters with independent parameters are given in Section 5.3 and 5.4. The
conclusion about dependent parameters is summarized as follows:
6.1.3.1 Tensile strength

(a) The tensile strength of weld joints are compared with the base metal and in few cases it is
greater, same and in few cases it is less than the base metal (Ref. Table 5.6 & 5.7).
168
(b) In modeling, the analysis of variance revealed that the factors MnO, CaF2 and MgO have
the significant effect on tensile strength. Factors NiO and Fe-Cr do not affect the tensile
strength significantly.
(c) The tensile strength of the wêld metal can safely be predicted through response surface
methodology as the same has been achieved in the present study.
6.1.3.2 Hardness
(a) The hardness of weld joints is compared with the hardness of base metal and the
hardness of all the weld joints is greater than the base metal (Ref. Table 5.6 & 5.7).
(b) From this study, it is concluded that the flux constituents influence the hardness of
welded joint during submerged arc welding.
(c) The study revealed that the hardness increases with the wt% of constituents, i.e.
MnO, CaF2, MgO, NiO , Fe-Cr . The major change in hardness occurred due to MgO
constituent because of its non reacting nature. The CaF2 constituent has less effect on
hardness as compared to other constituents because it reduces oxygen level due to
interaction with Si02.
(d) The study is thus helpfiil to design the flux for the submerged arc welding
applications to get the appropriate hardness of the welded joint.
6.1.3.3 Elongation
(a) The elongation of weld joints is compared with the elongation of base metal and the
elongation of all the weld joints is more thanthe base metal (Ref. Table 5.6 & 5.7).
(b) The analysis of variance revealed that the factors MnO, MgO, and Fe-Cr have the
positive impact on elongation but CaF2 and NiO show little impact on elongation.
(c)
6.1.3.4 Impact strength
(a) The impact strength of weld joints is compared with the impact strength of base metal
and the impact strength of all the weld joints is focused more than the base metal
(Ref. Table 5.6 & 5.7).
(b) The impact strength increases with the increasing percentage of MnO, MgO and CaF2
constituents.
169
(c) The analysis of variance revealed that the flux constituents during submerged arc
welding affect the impact strength significantly.
(d) The results indicate that the impact strength can be modeled with the help of response
surface methodology approach as predicted values are well fitted with the actual
values.
6.1.3.5 Optimal parameters (Mechanical Properties) using single response

optimizations
Optimal sets of process parameters, predicted optimal response and desirability value for
single response optimization are summarized in the table.
Predicted
Constituents-^-
optimal Desirability
Response \ MnO CaF2 MgO NiO Fe-Cr
response
Tensile Strength
(MPa) 8.75 17.5 12.5 2.5 2.5 577.248 0.954
Elongation (%) 8.75 17.5 12.5 2.5 2.5 29.5274 0.983

Impact Strength
(Joule) 6.25 15.2 12.46 7.16 2.51 99.2352 1
Hardness (No.) 8.75 17.5 12.5 7.5 2.5 173.1 0.947
6.1.3.6 Optimal parameters (Mechanical Properties) using multiple response

optimization
The optimal values of process parameters for multi response optimization using response
surface methodology and desirability fiinction are reported in the table.
Constituents -••
Desirability
Response | MnO CaF2 MgO NiO Fe-Cr
Tensile Strength and

Elongation 8.75 17.5 12.5 2.5 2.5 0.983
Impact strength 8.75 17.5 12.41 5.59 2.5 0.964
170
Contd.
Hardness 8.75 17.5 12.5 7.5 2.5 0.94
Tensile strength,
elongation, impact
strength and hardness 8.75 17.5 12.47 5.58 2.5 0.952
6.1.4 Weld joint microstructures

(a) The microstructures of weld joints are uniform and have light areas of Ferrite and
dark areas of cementite as microstructure of base metal.
(b) The uniform microstructure of weld joints and less inclusion support the superior
mechanical properties of the welded joints.
6.1.5 Bead morphology

(a) The experimental values of weld bead penetrations, reinforcement and bead width are
in close agreement with predicted penetration values of the model for the developed
agglomerated fluxes.
(b) The modeling and analysis may be an effective tool for welding industries to predict
the penetration, reinforcement and width behaviour in weld bead profile with respect
to newly developed agglomeratedfluxesfor submerged arc welding.
(c) The study reveals that the penetration, reinforcement and width of weld bead are not
only controlled by welding parameters but also depend on the flux constituents
involved during the agglomeration of fluxes.
(d) The analysis of variance reveals that the bead penetration and width decrease at the
higher percentage of MnO, CaFj, MgO and NiO constituents.
(e) The analysis of variance reveals that the bead reinforcement increases at the higher
percentage of MnO, CaFa, MgO and NiO constituents.
(f) Fe-Cr constituent is not prominent for penetration.
6.1.5.1 Optimal parameters (Bead Morphology) using single and multiple response
optimization
Optimal sets of process parameters, predicted optimal response and desirability value for
single and multiple response optimization are summarized in the table.
171
stituents^
Fe- Predicted optimal response Desirabili
onse| MnO CaF2 MgO NiO
Cr
8.75 17.5 7.5 7.5 2.5 10.5967 O.S
;ration
7.11 17.5 12.5 2.5 4.64 3.4136 0.8
ibrcement
8.75 16.83 12.5 7.35 2.56 18.9979
h
xation Penetration Reinforcement Width
ibrcement 8.05 16.70 7.50 7.50 5.25 0.6299306
Vidth 8.6746 4.74184 20.7242
6.1.6 Element transfer

a) The carbon equivalents to all weld metals by using developed agglomerated fluxes during
submerged arc welding process are within the range of high cold cracking susceptibility
which prevents the formation of martensite (Ref. Table 5.44).
b) The range percentage of elements in the weld metal by using developed agglomerated
fluxes during submerged arc welding process is under the limit of percentage elements
value prescribed by American Welding Society (AWS).
c) Mechanical properties of most of the weld metal using newly developed agglomerated
fluxes are superior to the base metal (Ref. Table 5.7).
d) The resuks of elements transfer study show that the newly developed fluxes are as good
as the commercial ones.
6.2 SCOPE OF FURTHER RESEARCH

In the present study, newly developed agglomerated fluxes are used during submerged arc
welding process on thefixedwelding parameters with direct current electrode positive. These
newly developed agglomerated fluxes may also be used with variable process parameters as
well as with tv^n wire submerged arc welding process. The present study may be used as a
reference for research to be carried out in future in this area.
172
REFERENCES
1. Brien, O. R. L. Welding Handbook, AWS, 1991, Vol. 11,191-232.

2. Olson, D. L. et al. Submerged Arc Welding, Metal Handbook, American Society for
Metals (6), 1983, 9* ed, 114-152.
3. The Procedure Handbook of Arc Welding, The Lincoln Electric Co., 1973, 6.3-1 to
6.3-24.
4. Specification for Carbon Steel Electrodes and Fluxes for Submerged Arc Welding,
AWS, 1989, A5.17-89,17-18.
5. American Welding Society, Welding Handbook, 1990, Vol.2, Eighth Edition, 1-41.
6. Tusek, Janez; D. Sc. Mathematical Modelling of Melting Rate in Twin-wire Welding,
Journal of Materials Processing Technology, 2000,100, 250-256.
7. Brown, R. S. Arc Welding of Stainless Steels, Metal Handbook, American Society
for Metals, 1983, Vol.6, 9*. Ed., 327-329.
8. Antony, K. C. Hardfacing, Metal Handbook, American Society for Metals Vol.6, 9
Ed., 1983, 784.
9. Thomas, D. R. Weld Overlays, Metal Handbook, American Society for Metals Vol.6,
9* ed., 1983, 804-819.
10. Nadkarani, S.V. Modem Arc Welding Technology, 1982, First Edition, 207-305.
11. How To Make Single Electrode Submerged Arc Welds, Publication S604, 1991. The
Lincon Electric Co., 11.
12. Buttler, C. A.; Jackson, C. A. SAW Characteristics of Cao-Tio2-Sio2 System, Welding
Journal, October 1967, Research Supplement, 448s-456s.
13. Jackson, C. E. Fluxes and Slag in Welding, December 1973, WRC, Bulletin, No. 190.
14. Tuliani, S. S.; Boniszewski; Eaton, N. F. Notch Toughness of Commercial
Submerged Arc weld Metal, 1969, Weld Met. Fab., Vol.8, 327-339.
15. Vishvanath, P. S. Submerged arc welding fluxes. Indian welding Journal, 1982,1-11.
16. Lewis, W. J.; Faulkner, G. E.; Rieppel, P. J. Flux and Filler -Wire Developments for
Submerged -Arc Welding HY-80 Steel, Welding Journal, 1961, 337s-345s.
173
17. Robinson, M. H. Observation on Electrode Melting Rate During Submerged Arc
Welding, Welding Journal, 1961, 40: 503s-515s.
18. Pokhodnya, I. K.; Kostenko, B. A. Fusion of Electrode Metal and its Interaction With
the Slag During Submerged Arc Welding, 1965, Avt. Svarka, No. 10, 16-22.
19. Bennet, A. P.; Stanley, P. J. Fluxes for the Submerged-Arc-Welding of Q.T.35 Steel,
British Welding Journal, 1966, 59-66.
20. Butler, C. A.; Jackson, C. E. Submerged Arc Welding Characteristics of the
CaO-Ti02 -SiOj System, Welding Journal, 1967,448s-455s.
21. Bennet, A. P. Metallurgy Using Basic Fluxes, Metal Construction and British
Welding Journal, 1970, 523-527.
22. Colvin, P. Basic Submerged -Arc Welding Fluxes, Metallurgia, 1970,45-50.
23. Drayton, P. A. An Examination of the Influence of Process Parameters on Submerged
Arc Welding, Welding Research International, 1972, Vol.2,1-19.
24. Ivochkin. et al. Automatic Submerged Arc Welding using Filler Metal in Powder
Form, Svar. Proiz., 1972, Vol. 2,26-30
25. Palm, J. H. How Fluxes Determine the Metallurgical Properties of Submerged Arc
Welds, Welding Journal, 1972, 358s- 360s.
26. Potapov, N. N.; Babin, S. A. Special Features of the Transfer of Alloying Elements
into the Metal in Deposition under a Fused Agglomerated Flux, Svar. Proiz., 1974,
Zentner, 1976, Vol.3,22-25.
27. Sorokin, L. I.; Sidlin, Z. A. Transfer of Elements from the Electrode into the
Deposited Metal, Svar. 1974, Proiz., Vol.5, 5-6.
28. Ferrera, K. P.; Olson, D. L. Performance of the MnO-SiOj-CaO System as a
Welding Flux, Welding Journal, 1975, 21 ls-215s.
29. Wittstock, G. G. Selecting Submerged Arc Fluxes for carbon and Low Alloy Steels,
Welding Journal, 1976, 733-741.
30. Eagar, T. W. Sources of Weld Metal Oxygen Contamination During Submerged Arc
Welding, Welding Journal, 1978.
31. Potapov, N. N.; Kurlanov, S. A. Quantitative Evaluation of the Basicity of Welding
Fluxes, Svar. Proiz., 1978, Vol.9, 4-7.
174
32. North, T. H.; Bell, H.B.; Nowicki, A.; Craig, I. Slag/Metal Interaction, Oxygen and
Toughness in Submerged Arc Welding, Welding Journal, 1978, 63s-75s.
33. Charles, H.; Entrekin. J. R. The influence of Flux Basicity on Weld -Metal
Microstructure, Metallography, Elsevier, 1979,295-312, 295-312.
34. Schwemmer, D. D.; Olson, D. L.; Williamson, D. L. The Relationship of Weld
Penetration to the Welding Flux, Welding Journal, 1979, 153s- 160s.
35. Koukabi, A. H. Properties of Submerged Arc Deposits-effects of Zirconium,
Vanadium and Titanium/Boron, Metal Construction, 1979, 639-642.
36. Chai, C. S.; Eagar, T. W. The Effect of SAW Parameters on Weld Metal Chemistry,
Welding Journal, 1980, 93s- 98s.
37. Davis, M. L. E.; Bailey, N. How Submerged Arc Flux Composition Influences
Element Transfer, Conference on Weld-Pool Chemistry and Metallurgy, The Welding
Institute U. K., 1980, Paper-34, 289-310.
38. Silvinski, A. M.; Kopersak, V. N.; Lolokha, A. K.; Yushchenko, K. A. The Physico-
Chemical and Technological Properties of CaF2- SiOa- AliOs-MgO Flux Systems,
Avt. Sevarka, 1981, No.7, 31-35.
39. Podgaetskii, V. V.; Galinich, V. I. The Structure of Molten Welding Slags, Avt.
Sevarka, 1981, No.7, 36-45.
40. Davis, M. L. E.; Bailey, N. Properties of Submerged Arc Fluxes- a Fundamental
Study, Metal Construction, 1982, 202-209.
41. Chai, C. S.; Eager, T. W. Slag Metal Reactions in Binary CaF ^ -Metal Oxide Welding
Fluxes, Welding Journal, 1982, 230s -232s.
42. Snyder, J. P.; Pense, A.W. The Effects of Titanium on Submerged Arc Weld Metal,
Welding Journal, 1982, 201s-21 Is.
43. Terashima, H.; Tsuboi, J. Submerged Arc Flux for Low Oxygen and Hydrogen Weld
Metal, Metal Construction, 1982, 648-654.
44. Kohno, R.; Takami, T.; Mori, N.; Nagano K. New Fluxes of Improved Weld Metal
Toughness for HSLA Steels, Welding Journal, 1982, 373s-380s.
45. Potapov, N. N. Calculating the Principal Reactions between Slag and Metal in
Submerged Arc Welding, Automatic Welding, Avt. Svarka, 1983, No.2, 29-31.
175
46. Davis M. L. E.; Pargeter, R. J.; Bailey, N. Effects of Titanium and Boron Additions to
Submerged Arc Welding Fluxes, Metal Construction 15,1983, 338-344.
47. Indacochea, J. E.; Olson D. L. Relationship of Weld-Metal Microstructure and
Penetration to Weld-Metal Oxygen Content, Materials for Energy "Systems, 1983,
Vol.5. No.3,139-148.
48. Mitra, U.; Eager, T. W. Slag Metal Reactions During Submerged Arc Welding of
Alloy Steels, Metallurgical Transactions A, 1984, Volume 15A, 217-227.
49. Lau, T.; Weatherly, G. C; Mc Lean, A. The Sources of Oxygen and Nitrogen
Contamination in Submerged Arc Welding Using CaO-Al2 03 Based Fluxes,
Welding Journal, 1985, 343s-347s.

50. Lau, T.; Weatherly, G. C ; Mc Lean, A. Gas/metal/slag Reactions in Submerged Arc
Welding using CaO-AbOs Based Fluxes, Welding Journal, 1986, 31s.
51. Tandon, S.; Kaushal, G. C ; Gupta, S. R. Effect of Flux Charateristics on HAZ During
Submerged Arc Welding, Int. Conf on Welding Technology, University of Roorkee,
India, 1988,11-65-11-71.
52. Burck, P. A.; Indacochea, J. E.; Olson, D. L. Effects of Welding Flux Additions on
4340 Steel Weld Metal Composition, Welding Journal, 1990, 115s-122s.
53. Mitra, U.; Eagar, T. W. Slag-Metal Reactions During Welding: Part I, Metallurgical
Transactions B, 1991, Vol. 22B, 65-71.
56. Gupta S. R.; Arora Navneet. Influence of Flux Basicity on Weld Bead Geometry and
HAZ in Submerged Arc Welding, Indian Welding Journal, 1991,127-133.
57. Pandey, N. D.; Bharti, A.; Gupta, S. R. Effect of Submerged Arc Welding Parameters
and Fluxes on Element Transfer Behaviour and Weld-Metal Chemistry, Journal of
Materials Processing Technology 40,1994, 195-211.
58. Paniagua, Ana. Ma.; Estrada-Diaz, Paulino.; Lopez-Hirata, Victor, M. Chemical and
Structural Characterization of the Crystalline Phases in Agglomerated Fluxes for
176
Submerged-arc Welding, Journal of Materials Processing Technology 141, 2003,
93-100.
59. Paniagua, Ana. Ma.; Victor, M.; Saucedo, Marible, L. Influence of the Chemical
Composition of Flux on the Microstmcture and Tensile Properties of Submerged-arc
Welds, Journal of Materials Processing Technology 169, 2005, 93-100.
60. Kanjilal, P.; Pal, T. K.; Majumdar, S. K. Combined Effect of Flux and Welding
Parameters on Chemical Composition and Mechanical Properties of Submerged-arc
Weld Metal, Journal of Materials Processing Technology 171, 2006,223-231.
61. Kanjilal, P.; Pal, T. K; Majumdar, S. K. Prediction of Element Transfer in
Submerged Arc Welding, Welding Journal, 2007, Vol.86, 135s -146s.
62. Adeyeye, A. D.; Oyawale, Festus, A. Mixture Experiments and Their Applications in
Welding Flux -Design, Journal of the Braz. Soc. of Mechanical Science & Engg.,
2008, No.4, Vol. XXX, 319- 326
63. Sharma, Abhay; Arora, Navneet; Mishra, Bhanu, K. Mathematical Modeling of Flux
Consumption During Twin-wire Welding, Int. J. Adv. Manuf Technology, 2008,
1114-1124.
64. Kumar, Vinod; Mohan, Narendra; Khamba, J. S. Development of Cost Effective
Agglomerated Fluxes from Waste Flux Dust for Submerged Arc Welding,
Proceedings of the congress on Engineering, 2009, Vol 1.
65. Bang, Kook-soo; Chan, Park; Jung, Hong-chul; Lee, Jong-bong. Effects of Flux
Composition on the Element Transfer and Mechanical Properties of Weld Metal in
Submerged Arc Welding, Met. Mater. Int., 2009, Vol.15, No.3, 471-477.
66. Kumar, P.; Satish, A.; Bhattacharya, A.; Duvedi, R. K. Effect of Process Parameters
on Microhardness of Heat Affected Zone in Submerged Arc Welding, Proceedings of
the Institution of Mechanical Engineers, Part B, Journal of Engineering Manufacture,
2011, Vol.225, 711-721.
67. Kumar, Vinod. Modeling of Weld Bead Geometry and Shape Relationship in
Submerged Arc Welding Using Developed Fluxes, Jordan Journal of Mechanical and
Industrial Engineering, 2011, Vol.5, No. 5, 461-470.
177
68. Shen , S.; Oguocha, I. N. A.; Yannacopoulos, S. Effect of H e a t Input on
W e l d Bead Geometry of Submerged Arc W e l d e d ASTM A709 Grade 50
steel joints. Journal of Materials Processing Tech., Vol. 212, 2012,286-295.
69. Amulf, Muan; Osbom, E. F. Phase Equilibria Among Oxides in Steel Making,
Addison-Wesley, USA, 1965.
70. Cochran, G.; Cox, G. M. Experimental Design, Asia Publishing House, New Delhi,
1962.
71. Box, G. E. P.; Hunter, J. S. Multifactor Experimental Design (J. Ann. Math.) 1957,
Statistics, 28.
72. Akhanazarova, S.; Kafarov, V. Experiment Optimization in Chemistry and Chemical
Engineering, Mir Publishers, Moscow, 1982.
73. Douglas, C. M. Design and Analysis of Experiments, (John Wiley, Fifth Ed.), 2007,
427-440.
74. McCabe, Smith. Unit Operations of Chemical Engineering, (McGraw-Hill, 7*
Ediition), NY, 2005, 985-986, 992-993.
75. Coulson; Richardson. Chemical Engineering Vol-2, 4* Edition, Asian Books Pvt.
Ltd, ND, 1991, 83-86.
76. Agarwal, S. P.; Khaima, Rajesh. Physical Pharmacy (2"*^ Edition), CBS Publishers &
Distributors, India. 44-45.
77. Electromagnetic Sieve Shaker, Instruction Manual, Model No: EMS-8, Electrolab,
Navi Mumbai, India.
78. Jackson, C. E. Fluxes and Slags in Welding, Bulletin 190. New York: Welding
Research Council, 1973.
79. Brookfield Digital Viscometer, Model DV-E, Manual No. M/98-350-G0307,
Brookefield Engineering Laboratories, Inc. USA.
80. Sachin Maheshwari. (1997) Element Transfer Behaviour of Fluxes During
Submerged Arc Welding, Ph.D Thesis, Mechanical engineering department, IIT
Delhi (India).
81. Singer, S. S. Industrial Ceramics, Chapman & Hall, London, 1963,236-257.
82. Allen, A. W. Optical Microscopy in Ceramic Engineering, Proceedings of the Third
Berkeley International Materials Con- ference, Berkeley, CA, 1966.
178
83. Redwine, R. H. and Conrad, M.A. Microstructures Developed in Crystal- lized
Glass Ceramics, Proceedings of the Third Berkeley International Materials
Conference, Berkeley, CA, 1961.
84. CuUily, B. D., Elements of X-ray Diffraction, Addison-Wesley, USA, 1978, 407-
417.
85. Gordon, P. K., Microstmcture of Complex Ceramics: Proceedings of the Third
Berkeley International Materials Conference, Berkeley, CA, 1966.
86. Prakash, S.; Palani, kumar, K. Modeling for Prediction of Surface Roughness in
Drilling MDF Panels Using Response Surface Methodology, Journal of composite
Materials, 2011, Vol. 45,16,1639-1646.
87. Deringer, G. and Suich, R. Simultaneous Optimization of Several Response
Variables, Journal of Quality Technology, 1980,12, 214-219.
88. Aggarwal, A.; Singh H.; Kumar, P.; Singh, M. Modeling of Machining Parameters
and Cooling Conditions in Hard Turning of AISI P-20 Tool Steel Using Response
Surface Methodology and Desirability Graphs. International Journal of Machining &
Machinability of Materials, 2008 ,4(1), 95-110.
89. Aggarwal, A.; Singh H;, Kumar, P.; Singh, M. Simultaneous Optimization of
Conflicting Responses for CNC Turned Parts Using Desirability Function.
International Journal of Manufacturing Technology and Management, 2009, 18, 319-
332.
90. Lancaster J F. Metallurgy of welding, Alden Press Ltd., London, 1980, 25-30.
179
LIST OF PUBLICATIONS (JOURNALS)
1. Kumar, A.; Singh, H.; & Maheshwari, S. Effect of bulk density of developed
agglomerated fluxes on flux consumption & penetration during submerged arc
welding. Trends in Mechanical Engineering & Technology , December 2012,
Vol. 2, Issue 3, pp 6-11.
2. Kumar, A.; Singh, H; & Maheshwari, S. Modeling and Analysis by Response
Surface Methodology of Hardness for Submerged Arc Welded Joints using
Developed Agglomerated Fluxes, Indian Journal of Engineering & Materials
Sciences, Vol. 19, December 2012, pp. 379-385. (Impact Factor: 0,223)
3. Kumar, A., Singh, H. & Maheshwari, S. "Modeling of bead Penetration behaviour
for varying constituents of agglomerated fluxes during submerged arc welding",
Journal of Mechanical Engineering (i-Manager Publications). Vol. 2, No.4,
August-October, pp. 25-30.
4. Kumar, A., Singh, H. & Maheshwari, S. "Review study of submerged arc welding
fluxes". Journal of Mechanical Engineering (Article in Press).
5. Kumar, A., Singh, H. & Maheshwari, S. "Modeling and analysis of elongation for
submerged arc welded joints by using developed agglomerated fluxes", Journal of
ENGINEERING MANUFACTURE, Institution of MECHANICAL
ENGINEERS, Part -B (Article in Press, Impact Factor: 0.73).
6. Kumar, A., Singh, H. & Maheshwari, S. "XRD and DTA analysis of developed
agglomerated fluxes for submerged arc welding", Journal of Metallurgy (Hindawi
Publishing Corporation). Vol. 2013, Article ID 827491, January, pp. 1-8.
7. Kumar, A., Singh, H. & Maheshwari, S. "Effects of developed agglomerated flux
composition on element transfer of weld metal during submerged arc welding",
Indian Welding Journal (IIW, under review).
180
APPENDIX A
Table A.l: Sieving Results for AGFllOl Flux

SIEVING R E S U L T S
Retained Range
Sieve sUas Weight [%{ Particle size P3 [%]

[mml
Pan 30.6 <: 0.3SS 30.6
0.3S5 21.44 0.35S - 0.6 21.44
0.6 17.1 0.6 - 0.85 17.1
0.85 2S.9 0.85-1.18 25.S
1.18 4.96 1.18-2 4.96
2 0 >2 0
Table A.2: Sieving Results for AGF1102 Flux

SIEVING RESULTS
Retained Range
Sieve size Weight |%! Particle size P3 [%]

[mm] (mml
Pati 30.76 < 0.355 30.76
0.355 25.22 6.355 - 0.6 25.22
0.6 19.95 0.6 - 0.85 19.95
0.85 20.83 0.85-1.18 20.85
1.18 3.22 1.18-2 3.22
2 0 >2 0

SIEVING RESULTS
Retained Range
S i e v e size W e i g h t r%l ParUcle size P 3 [%|

tmnri] [mmj
Pan 14,49 «: 0.3SS 14.4S
0.35S 25.21 0.3SS - 0 . 6 25,21
0.6 25.21 0.6 - 0.8S 2S.21
0.8S 32.71 0.85 - 1.18 32.71
1.18 2.38 1.18-2 2.38
2 0 > 2 O
181
SIEViNO RESULTS
Retained Range
Sieve size Weight {%J Particie size P3[%]

tmrn] |mm]
Fan 44.3e < 0.355 44.36
0.355 19.35 0.355 - 0.6 19.35
0.6 16.37 0.6 - 0.»5 16.37
0.85 17.2T 0.86-1.18 1T.27
1.18 2.65 1.18-2 2.65
2 O >2 O

SIEVING RESULTS
Retained Range
Sieve size Weight [%] Particle size P 3 [%|

[mm] [mm]
Pan 34.82 •e 0.355 34.82
0.355 5.25 0.355 - 0.6 5.25
0.6 29.18 0.6 - 0.85 29.18
0.85 28.92 0-SS- 1.18 28.92
1.18 1.82 1.18-2 t.S2
2 0 > 2 0
SIEVING RESULTS
Retaineid Range
Sieve size Weight {%] Particfe size P3 l%l
(mmj Imm]
Pan 6.72 < 0.355 6.72
0.35S 29.09 0.355 - 0.6 29.09
0.6 29.59 0.6 - 0.85 29.59
0.85 26.79 0 . 8 S - 1.18 26.79
1.18 7.81 1.18-2 7.81
2 0 >2 0
182
SIEVING RESULTS
Retained Range
Sieve size W e i g h t [%] Parttcle size P 3 I%]
[mm] Irrtml
Pan 4-59 < 0-355 4.59
0.3S5 6.98 0 . 3 5 5 - 0.6 6.98
0.6 10.13 O.© - 0.85 10.13
0.85 19.98 O . 0 5 - 1.18 19.98
1.18 46.26 1.18-2 46.26
2 12.06 » 2 12.06

SIEVING RESUi-TS
Retained Range
Sieve size W e i g h t [%] P a i t i c l e sizre P 3 [%]
[mm] [mmj
Pan 36.09 « 0.355 36.09
0.355 26.27 0.355 - 0.6 26.27
0.6 17.55 0.6 - 0.85 17.55
0.85 16 8 0 . 8 5 - 1.18 16.8
1.18 3.3 1.18 - 2 3.3
2 O > 2 0

SIEVING RESULTS
Retained Range
Sieve size W e i g h t [%] Particle size P 3 [%]
[mm] [mm]
Pan 16.93 < 0.355 16.93
0.355 25.22 0 , 3 5 5 - 0.6 ^C ^ j . .^..j^.
0.6 21.24 0.6 - 0-85 21.24

0.85 26.61 0 . 8 5 - 1.18 26.61
1.18 10 1.18-2 10
2 O > 2 O
183
Table A.10: Sieving Results for AGFllOlO Flux
SIEVINO RESULTS
Retaiined Range
Sieve size W e i g n t [%] Particle size P 3 £%]

tmm) [mmi
Pan 4569 -s 0 . 3 5 S 45.69
0.355 20.67 0.355 - 0.6 20.67
o.e 15.83 0.6 - 0 . 8 5 15.83
0.85 16.62 0 . 8 S - 1.18 16.62
1.18 1.18 1.18-2 1.18
2 O > 2 O
Table A . l l : Sieving Results for A G F l l O l l Flux

iîfciviiMiai t<t=±»ui-i:>
Re-tainedl Ran<sie
Sieve size W e i g h t [%1 Particle size P 3 (%]
[mmj [mmJ
Pan 36.95 < 0.355 36.95

0.355 23.76 0.355 - 0.6 23.76
0.6 19-9 0.6 - 0 . 8 5 19.9
0.85 17.,23 0.85-1.18 17.23
1.18 2.17 1.18-2 2,17
2 O > 2 O

SIEVING RESULTS
N
Retained Range
Sieve size W e i g h t [%] Particle size P 3 1%]
[mm| tmm|
Pan 43.32 < 0.355 43,32
0.355 23.01 0.355 - 0.6 23.01
0.6 15.86 0.6 - 0 . 8 5 15.86
0.85 16.11 0 . 8 5 - 1.18 16..11
1.18 1.7 1.18-2 1.7
? 0 > 9 O
184
SIEVING RESULTS
Retatmecf RanQi6
Sieve size W e i g h t [%] Pa;riac!e size P3 t%]
Emml tmm]
Pan 44.78 -= 0.35S 44.78
0.355 22.4 0.355 - 0.6 22.4
0.6 15.43 0.6 - O.SS 15.43
0.85 16.28 O.S5 - 1.18 16.28
1,18 1.11 1.18 - 2 1.11
2 O s- 2 O
Table A.14 Sieving Results for AGF11014 Flux

SIEVING RESULTS
Retained Range
Sieve size W e i g h t p/o] Particle size P3[%I

[mm] [ m mJ
Pan 47,94 < 0.356 47.94
0.3S5 25.64 0.365 - 0.6 26.64
0.6 17.16 0.6 - 0.8S 17.16
0.85 7.96 0 . 8 5 - 1.1S 7.96

1.18 1.29 1.18-2 1.29
2 0 > 2 0

SIEVINO RESUt_TS
Retained Rartoe
Sieve size W e i g h t [%.] Particste s i z e P 3 [%]
[mm| |mmj
F*stri 17.67 "< 0 . 3 5 S 17.67
0.355 39.24 0.355 - 0.6 39.24
0.6 2S.S6 0 . 6 - O.SS 2S.86
O.SS 1S.8 O.SS-1.18 1S.8
1.18 1.42 1.1S - 2, 1.42
2 O - 2 O
185
SIEVING RESULTS
Retained Range
S i e v e sSze Wfifigît (%1 Particle size P 3 P^J

[mml Inmm]
Pan 29.67 <: 0 . 3 5 5 29.67
0.355 31.88 0.35S - 0.6 31.88
0.6 20.1T 0.6 - 0.85 20.17
O.S5 17.03 0 . 8 5 - 1.18 17.03
1.18 1.25 1.18-2 1.25
2 O > 2 O

SIEVING RESULTS
Retafned Ranse
Sieve size VVeight [%i Rarticte s i z e P 3 [%]
Pan 8.71 < 0.355 8.71

0.3SS 3 5 .S7 0 . 3 5 5 - O.e 35.97
0.6 32,3 0.6 - 0.85 32.3
0.85 21,4^1 0.85- t.18 21.41
1.18 i.ei 1.18 - 2 1:.Q1
2 o > 2 0
SIEVING RESULTS
R e t a i ft e d Rang®
Sieve size Weight (%] Particle size P 3 £%J
Irrtm]
Pan 33.46 -= 0 . 3 5 5 33.46
o^.ass 30,.28 0..3SS - 0 . 6 30.28
0.6 16.21 0.e - 0.85 18.21
0.8S 15.5 0.85 - 1.18 15.5
IJIS 2.56 1-18 -a 2.56
2 O > 2 O
186
SIEVilSiG RESUI-TS
Retained Range
Sieve size W e i g h t [%-] Partfcrfe s i z e P 3 [%|

[mm] Imm]
f*an 16.01 < 0.3S5 le.oi
0.3S5 32.07 0 . 3 S 5 - O.© 3207
o.e 25.58 0.6-0.8S 25.58
0.S5 23.22 0 . 8 5 - 1.18 23.22
1.18 3.12 1.18-2 3.12
2 O 3» 2 O

SIEVINO RE3Ui--r3
Retained 'R.^n^&
Slove size W e i g h t i'V^J, :F>arUiCte s i z e R3 t%]

[mm] [mm]
Rsn 8.81 -= 0 . 3 S 5 S.81
0.355 2S..S4. 0.35.5 - 0 . 6 :2a.'©^
O.e 29.75 0.6 - 0.85 29-7'5

O.S5 26.99 0.85 - 1-18 26.99
1.18 5.91 1.18 - 2 5.91

2 O =• 2 0

SIEVING RESULTS
Retained Range
Steve size Weight [%] Particle size P3 [%]
[mm] [mm]
Pan 13.75 <: 0 . 3 5 5 13.75
0.355 23.03 0.356 - 0.6 23.03
O.e 24.54 0.6 - 0.85 24.54
0.85 31.65 0.85- 1.18 31.55
1.18 7.14 1.18 -2 7-14
2 0 > 2 0
187
SIEVING RESULTS
Retained! Ranse
Siev© size W e i g h t [%I Particle size P 3 [%]

{mm] |rrim]
iPan: 16.64 < 0.355 16.64
0,3SS 38.25 0 355 - 0.6 38.25

0-6 24.76 0.6 - 0.85 24,76
0.85 18.3 0 . 8 5 - 1.18 18.3
1.18 2.05 1.18 - 2 2 . OS
2 O > 2 O

SIEVING RESULTS
Retaineci Rangie

Jmrrii] Imrn]
Pan 7.8 <• 0 . 3 5 5 7.8
0.355 35.72 0.355 - o.e 35.72

0.6 32.4 0 . 6 - O.BS 32.4
O.S5 21.33 0 . 8 5 ~ 1.18 21.33
1.18 2.76 1.18-2 2.76
2 O > 2 O

SIEVING RESULTS
Retained Range

[mm] [mm]
Pan 20.98 < 0.355 20.98
0.355 35.08 0.355 - 0.6 35.08
O.e 25.44 0.6 - 0.85 25.44
0.85 15.63 0 . 8 5 - 1.18 15.63
1.18 2.87 1.18-2 2.87
2 O > 2 O
188
SIEVING RESULTS
Retainecf Range
Steve size W e ight [%] Particle size P 3 I%J

[mm] [mm]
Pan 11.T •= 0 . 3 5 5 11.7
0.355 3 7 . OS 0.355 - 0.6 37.08
0.6 31.19 O.e - 0 . 8 5 31.19
0.85 17.47 0 . 8 5 - 1-18 17.47
1:.18 2.55 1.18- 2 2.55
2 O > 2 O

SIEVING RESULTS
Retained Range
[mm] [mm]
Pan 24.08 < 0-355 24.08
0.355 30.14 0.355 - 0.6 30.14
0.6 23 0.6 - 0.85 23
0.85 19.48 0.85- 1.18 19.48
1.18 3.3 1.18 -2 3.3
2 0 > 2 O

SIEVING RESULTS
Retained Range
S i e v e size W e i g h t [%] P a r t i c l e sJz© P 3 I%]

[mml [mml
Pan 2.. 7 -c 0 . 3 5 5 2.7
0.355 19.44 0.355 - 0.6 19.44

0.6 47.48 0.6 - 0.85 47.48
0.85 22.85 0 . 8 5 - 1.18 22.85
1.18 7.54 1.1S - 2 7.54
2 O > 2 O
189
SIIEVIN6 R E S U t - T S
Retained Range
Sieve size Weight [%] Particle s i z e P 3 I%]

tmm| {mm]
Pan 30.6 < 0.355 • t J T i j f . %jf
0.355 21.44 0-355 - 0.6 21.44

0.6 17.1 0 . 6 - 0.8S 17.1
0.8S 25.9 0 . 8 5 - 1.18 25.9
1.18 4,96 1.18-2 4.96
2 0 > 2 0

SIEVING RESULTS
Retained Range
Sieve size Weight {%] Particle size P3 |%]
[mm] [rranj
Pan 14.49 < 0.355 14.49
0.355 25.21 0.355 - 0.6 25.21
0.6 ^ . i ? - ^ < b 1: 0.6 - 0.85 &'<Z).^ 1
0.8S 32.71 0.85-1.18 32.71

1,18 2.38 1.18-2 2.38
2 0 >2 0

SIEVING RESULTS
Retained Range
s i e v e size Weight r%l Particle size P 3 [%]
Imm] [mm]
Pan 34.82 < 0.355 34.82
0.355 5.25 0.355 - 0.6 S.2S
0.6 29.18 0.6 - 0.85 29.18
0.85 28.92 0 . 8 5 - 1.18 28.92
1.18 1.82 1.18 - 2 1.82
2 0 >2 O
190
SIEVING RESULTS
Retained Range
Sieve size W e i g h t py<,i R a r t i c l e sioze P3 [%]
[mm] [mmJ
Psin 2.7 -= 0 . 3 S S 2.7
0.3SS ->9.4^ 0.35S - 0.6 19.-44

0.6 -47.-48 0.6 - 0.85 4.7.-48
0.8S 22.85 0.85 - 1.18 22.85
1.1S 7.5-* 1.18 - 2 7.54
2 O =- 2 O

SIEVING R E S U L T S
Reta ined Range

Steve size W e i g h t ["y..] Rarticle size P 3 |<S4.1
[mm] tmm}
Pan 4.S9 •< 0 . 3 5 S 4.S9
0.355 6.98 0,3S5 - 0.6 e.98
o.e 10.13 O.e - 0 . 8 S 10.13
0.85 1S.9B 0 . 8 5 - 1.18 19-98
1.18 46.26 1.18 - 2 46.26
2 12.OS » 2 12.06
191
APPENDIX B
WclBIK [%J
so -
30 -
20 -
1 1 : | : : : : : = : : :
1111
•
10 -
I
2 ) 18 085 0.6 0^55 Pan
Sieve (mm)
Figure B.l: Grain size Histogram for AGFllOl
J3
u
tn
r.ie 0.85 0.R 0.:i55 Pan
Sieve (mm)
Figure B.2: Grain size Histogram for AGF1102
192
vv«*»mi«<j
E 50 -
3C5 -
:::,:=:
::|1 1 : 11
It) -
a
. II
i.^& O.S&
1 •
oe
1 1
o.asG Pan
Sieve mm)
50 ~
: : : . : : : .=^^^^^^^^^^^
-DO -
01
•30 ~
la -
^.:--r:i-f^l -i^
•3 ^ \ti oes o.D aass P«II
Sieve (mm)
1
:«
"53
IK> ••
2: i.ie
1]
•ftftS o.s 0 3H5
-•"
Sieve (mm)
Figure B.5: Grain size Histogram for AGFllOS
193
A<3 -
1
• 1-H
a;
«o -
2
. 11 1
1,19 0.155 0,0 0.a55
1
p„
Sieve (mm)
w«^m |>r
«o -
m -
.2-
20 -
• - , 1 ,
1 1 11
fO -
•
? 1.t» a.Ks o.e D,3iSS Pan
Sieve (mm)
Wfltsiht tat
so "
40 -
ax
1 30 -
20 -
10 -
• 1 1 1
2 1.18 0.85 0.6 O-SSS *»•n
Sieve (mm)
194
50 -
5 m -
1 01
30 -
1 1'f- ::
11
JO -
1
10 -
a Lie 0.S5 as 0,355 Part
Sieve (mm)
Figure B,9: Grain size Histogram for AGF1109

W « ^ i*A]
50 •
<Q •
I 30 -
JO -
: . • . 1
10 -
S 1,» O.SS 0.5 0.355 Ptr,
Sieve (mm)
Figure B.IO: Grain size Histogram for AGFlllO

Weight fii\
- - -
i
•• - •
SO ..
4D -
" " • " • " • — —- -
1 3D ~
1
1
20 -
•.. . 1
I
10 -
2 1.18 0.8$
1
o.s 0.35S Can
Sieve (mum)
Figure B.ll: Grain size Histogram for A G F l l l l
195
Wwgh(E%J
i
50 ••
*3 --
• i-H
30 -
20 -
1
ti3 ..
2 118 0B5 ae 0.355 Pan
Sieve (mm)
WrtOhSJ%)
se -
40 -
"53 30 -
20 -
f
1 1J^
1
2 MB OK ti OS-IS P»J.
Sieve (mm)

w*gtitiin
50 -
- —
bl
30 -
10 -
7 lift
Ill
^.*l^ flfi rx.>is P
Sieve (mm)
196
so -
«0 -
20 -
10 -
2 t.18
ll 1
0,85 0,6 0.255 PSfl
Sieve (mm)
50 '
.2
2 1.16
li O.eS 0.6 C.355 P««
Sieve (mm)
50 -
ja£) ~
.2
!53
30 •
.^::::::::= •1:1::::::::::::::::::^:--:
W ••
H 9 • 1 ^B
2 1,1» Q,eS 0-6 0.323 f>an
Sieve (mm)
197
'S
Sieve (mm)
J3
'S
Sieve (mm)
01
• i-H
Sieve (mm)
198
'53
Sieve (mm)
so -
/-—*
^ 40 -
30 ~ - -
20 -
1
10 -
2
. 1T 1
t.lS a.8S 0,6 0. » 5 Pan
Sieve (mm)
]Figure B.22: Grain size Histogram for AGF1122

Wfti8ht[%l
SO -
JO . ,^.: ::::::i| J : :::
I11 l"11l 11 1'

20 -
«0 -
2 %1 S 0.65 0.6 0,S55 Purr
Sieve (mm)

199
I
Sieve (mm)
J3
Sieve (mm)
Sieve (mm)
200
WaiflfrtptJ
KJ -
^—
5OJ
50 "
•
1 1
•50 -
2 M«
1 110.65 0.6
1
0.355 Pon
Sieve (mm)
50 -
40 -
• i-H
1- 1
30 -
11
\ :• :
10 -
2
1 1
!.18 Q.eS OS ossa Pan
Sieve (mjn)
.'-N
SO -
5
bO m
I::1:::^Z::
10 -
1
1 1 1
Ill 1
1
;._._. : : = | : =
2 1,18 ceil C ,0 0.355 Poo
Sieve (mm) Sieve
201
^ SO -
o^
J3
• ^60
H
II1
<D
^ •
20 -
iO -
2
. II 1
1.18 «R5 0.6
•
O.tfS'j P»o
Sieve (mm)
»0 -
/—
40 -
•SO -
20 -
1 1
10 -
1 . i:- :
2 1.1a 0.85 0S 0,555 Pan
Sieve (mm)
wiwaMitti
50 -
30 -
J3
1, 1
20 ~
1 1 1
10 ~
•
2 1. fft 0.85 O.e 0.3S5 Pan
Sieve (mm)
202
APPENDIX C
FLUX-AGFllOl
1.SiO.
2. CaAI^SiÔg
500- 3.Mg,Fe,Si30,,(OH),
4. CaAI,Si,0,^.5H,0
5. MgÂI^Si,0,„(0H).4Hp
6. C a / e
400-
3
300-
200-
100
2 Theta(Degree)
Figure C.l: XRD Graph of AGFllOl Flux
FLUX-AGF1102
1. CaAI^SiÔ.
500- 2. K3CaMgÂl3,Si330,,.28(HÔ)
3. Mg^(AI, F e * n 3 S i 3 AIO,„(OH)3
4. Fe, Ni
400-
6. Ca3AI^(SiO^)3.X(OH)^^(X=0.2-1.5)
§ 300-
200-
100-
I
20 25 30 35 40 45 50 55 70
2Theta(Degree)
Figure C.2 XRD Graph of AGF1102 Flux
203
FLUX-AGF1103
600' LCaSiOg
2.Ca,„Mg,AI,Si,P33.4(Hp)
3. M g 2 ( A I , F e * n 3 S i 3 A I O , , ( O H ) 3
500-
400-
I 300-
200-
100-
—I
20 30 40
50
2Theta(Degree)
Figure C.3: XRD Graph of AGF1103 Flux
FLUX-AGF1104
600-, LCaAI^SigO^g.SCH.O)
2. KAI^ (Si3AI)0,„(0H,F)^
3. C a 2 ( M n , F e " " ) F e " " " S i p , ^ ( O H )
4. Mg,(AI,Fe^^*)3Si3AI^„(OH)3
5. Mg,(SiO,),(OH,F)3
6. Mn^Si3
20 25 30 35 40 45
2Theta(Degree)
204
FLUX-AGF1105
600-,
1.MgÂl2(OH),^Co3.4(Hp)
2.Ca,Si,O,(C03),
500- 3. Mg^SiO^
4. Mgg(SiO^)4(OH,F)2
5. Al,03
^ 400-
•gS 300 -I
200-
100
2Theta(Degree)
FLUX-AGF1106
1. SiO^
2. SiO^
500 - 3. CaCOg
4. (Mg, Fe*")g(siOj2(F. OH)^
5. AIÔ,
^ 400 -
2 3 6. MnO^
I
4
'1 300 -
2 0 0 ^-
^\JKJ J
***w^^^
— I — " ^ — I — • — I — —I
20 25 35 40 45 50 55 60
2Theta(Degree)
205
600 FLUX-AGF1107
1.MgCo3.5(Hp)
500- 2. Al203.(siO^)1.3-2.(H20)2.5-3
3. Ca^CM, Fe**) Fe^^^SiÔ^^COH)
4. N i / e to NigFe
^ 400
(0 300
c
200-
100-
2Theta(Degree)
FLUX-AGFl 108
1.AI,03.(SiO,),3.3.(H,0)
600- 2. Ca2(Mn, F e " ) F e " * S i 5 0 ^ ^ ( 0 H )
3. MgAIÔ^
500-
4. Mg^Sij
5. ca^(siO^)3(OH)^
400-
300-
200-
100-
•tt^f \km WWW^^ww^^
— I — —I
20 30 40 50 60 70
2Theta(Degree)
206
FLUX-AGF1109
1. 03351^0^(003)
600 -
2. OaMg3(Oo3)^
3. Mgg(SiO^)(OH.F)2
500 - 4. Al,03
3 400
300 -
3 4 5
200 -
100 -
y
—I
20 30 40 50 60 70
2Theta(Degree)
FLUX-AGFlllO
600- 1. (Ca. M n n , , S i ^ P 3 3 ( O H ) 3 . ( H p )
2. (Mnn^Si03(0H),.(Hp)
3. Ca^SiO,03(C03)(OH,F)3
4. Ca,„MgÂl3Si^,03,.4(Hp)
5. MnO„
^'|•*S^*^W^^
—1— —T—
40 50 60 70
2Theta(Degree)
Figure C.IO: XRD Graph of AGFlllO Flux
207
FLUX-AGFllll
600-
1. SiO^
2. Mg^SiÔ,
3. CaCO„
4. (Mg,Fen5(SiO,),(F,OH)2
5. Ca,3Mg,AI,Si,,03,.4(HÔ)
2Theta(Degree)
Figure C.ll: XRD Graph of AGFllll Flux
F L U X - A G F l 112
1. KAISi303
600 - 2. caAlsiO^(OH)
3. C a O
4 . AI3O3
5. F e S i
^'^^^mLfft^^
—I—
40 50 60 70
2Theata(Degree)
Figure C.12: XRD Graph of AGFllll Flux
208
FLUX-AGF1113
800
1. SiO^
700
2. CaFe"*(SiOJg(OH)^.19(HÔ)
3. (Mg, Fe%(SiOJ(F,OH)2
600 4. Al,03
500-
400-
2oo-j(||VTif«!frli
200
100-
^*ww^^
—I— —I— —I
20 30 40 50 60 70
2Theta(Degree)
FLUX-AGF1114
700 1. CaSiOj
2. CaAI,(AI^Si,)0^„(OH)
600- 3. (Mnn2Si03(OH)^.(Hp)
4. Al,03
5. ca3H,Si30,3.6HÔ
500-
400-
300-
200
"piffl^^
100 -
—I— —I
20 30 40 50 60
2Theta(Degree)
209
FLUX-AGF1115
1200-, 1. C a ^ M n " * S i 3 0 J O H ) 3 . 2 ( H 3 0 )
1100 2. (Ca. M n " ) , , S i , , 0 , 3 ( O H ) 3 . 2 ( H p )
1000- 4. MgO
900-
800-
700
600
500
400
300
200
100 l^W^^ll^rtWW^^^
0
20 30 40 50 60 70
2Theta(Degree)
FLUX-AGF1116
1. MgSO^.TCHÔ)
2. SiO^
3. (Fe*Mvlg)^Si O
2 2 e
4. Al,03
^*Www^^
-I— —1— —I— —I
20 30 40 50 60
2Theta(Degree)
210
FLUX-AGF1117
800-1
LCaCOg-CHÔ)
700- 2. (Fe**,Mg)ÂI,Sip^3
3. CaAI^SiÔg
600- 4. (Mg. Fe*\ Ca)(Mg, Fe'") S i p
2 6
5. Ca F e " " " 0 ,
2 5
d 500- 6. (Mg, F e n ^ ( S i O , ) , ( F , O H ) ^
S
S 400 -I
300-
200-
2 3 ,4
—T- —I
20 30 40
2Theta(Degree)
FLUX-AGF1118
800 -|
1. CaÂI^SiOgCOH)^
700 - 2. Mn*%Si03(OH)2.(HÔ)
3. Mg^SiO^
600 - 4. AI3O3
5. Mn^(SiO^)3(OH)^
3 500 -
S 400 -
300
200 -
100 -
^HMMtWMUrfititM'WsU i»Jf»w^><»M>iw*<WiM)M'
—I—
20 30 40 so so
2Theta(Degree)
Figure C I S : XRD Graph of AGF1118 Flux
211
FLUX-AGF1119
1. SiO^
600 - 2 . K^Fe^^Fe^^^SigO^^COH)^
3. C a F e " V S O ^ ) g ( O H ) ^ . 1 9 ( H 2 0 )
500 - 4. (Mg. Fe*X{S\0;)^{F,OH)^
5. A l , 0 3
3 400 -
Jo,
300 -
200 -
1 0 0 - 1 I
«Mi#<wM^4yJSi#li'lW''*''^^
—I— —I
20 30 40 50 60
2Theta(Degree)
FLUX-AGF1120
600 1. Ca,AI[(OH)ÂISiO,.3(OH), 3 ] . 2 , 5 ( H , 0 )
2. M g ^ S i p ,
3. M n " % S i 0 3 ( O H ) 2 . ( H p )
500
4. Mgg(SiOJ^(OH,F)^
5. A l , 0 3
5 400-
S 300-
200
100-
4 5 6
-MfMtNwW
20 30 40 50 60
2Theta(Degree)
212
FLUX-AGF1121
600-, 1. CaAI^SiOg
2. Mn*%Si03(OH)2.(H20)
3. ( M g , F e * * U S i O J , ( F , O H L
500-
^ 400-
Si.
300-
.S
200-
100-
—I— —I
20 30 40 50
2Theta(Degree)
600-, FLUX-AGF1122
1. CaAI^SiÔg
2. C a , S i 3 0 , 3 ( O H ) , . 2 ( H p )
500-
3. (Mg. Fe%(SiO^),(F,OH)2
4. Al,03
400- 5. FeS
300-
200-
100- 1
3 4 5
20 30 40 50 60
2Theta(Degree)
213
700-,
FLUX-AGF1123
1. K ^ C a 3 ( S i 3 0 , , ) , ( S p , ) s i , 0 3 ( O H ) . 3 ( H p )
600- 2. CaCOg
3. ( M g , F e ^ \ ( S i O W F . O H ) ,
500-
400-
1 2
•g 300-
200-
100-
2Theta(Degree)
600-, FLUX-AGF1124
1. SiO^
2. CaFe**7SO^),(OH)2.19(HÔ)
500-
3. C a O
4. Al,03
400- 5. C a , ( S i O j 3 ( O H ) 2
300-
200-
1 2
100- I ll
—I— —I
20 30 40 50
2Theta(Degree)
214
600-,
F L U X - A G F l 125
1. CaAI^SiÔg
500- 2. Ca,Si30,3(OH),.2(H,0)
3. (Mg,Fe**)^(SiOJ^(F,OH)2
4. Al,03
^ 400 5. CaMnSiO,
3.
I
•D 300
•s
200-
)iihi/(0«Hîîi^^
20 30 40 50 60
2Theta(Degree)
600-1
FLUX-AGF1126
LSiO^
500-
2. ( M g , Fe"^Ca)(Mg,Fe**)Si203
3.[Ca,,(OH)3](Si30,„)(Si30,„yFe%(OH)^J
400-
300-
200-
100-
—r- —r-
20 30 40 50 60
2Theta(Degree)
215
600-1 FLUX-AGF1127
1. MgC03.5(H20)
500- 2. Ca^[Mg,(AI,Fe***)lSi/.IO^^(OH)^
3. Mg^(AI,Fe^^*3Si3AIO^„(OH)3
4. CaF^
^ 400-1 5. M n ^ ( S i O , ) 3 ( O H ) ,
a 300 -
200-
100-
^''*''«**'^^^^
—r- —I— I
20 30 40 50
2Theta(Degree)
600-,
FLUX-AGF1128
500- 1. Ca,[Mg,(Al,Fe^**)]Si/.IO,,(OH),
2. CaF„
400-
I 300-
200-
100-
^^''*''***'****^^
I I
20 30 40 50
2Theta(Degree)

216
®°°"' FLUX-AGF1129
500-1 1- CaJMg^(AI,Fe*^*)]Si/,io,,(OH)2
2. MgJAI.Fe*^^
'2^ '
Si A l O10^
3 3
^ r O H' L8
3. CaF„
^ 400-
3
(C
^^
'0 300 H
200-j
100-
^^'**^'***^'***^^
' 1 ' 1 « 1 —
20 30 40 50
2Theta(Degree)

600 - ,
FLTJX-AGFl 130
500- 1 . Ca5SigO,3(OH)2.2(H20)
2. Al,03
^ 400-
3
•ffl 300 -
200-
100- 1 2
IMM*I»****«W
• 1 • 1 • 1 ' 1
20 30 40 50 60
2Theta(Degree)
217
FLUX-AGF1131
900-,
1. SiO^
800- 2. (Fe*M\/lg)2Si^03
3. Al,03
700-
^ 600-
% 500-
,2 400-
300
200 ^f^W Jw ^^'^^^^^^

100-1
0 1 1 1 1 1 1 1 1
20 30 40 50 60
2Theta(Degree)
6 0 0 -1
F L U X - A G F l 132
1 Ca^SiÔ^^(OH)^.2(HÔ)
500 -
2. A l , 0 3
3. C a M n S i O •4.
^ 400 -
3
P 300 -I
200 -
i 23
il|^^^W%1lltlyHl^^
1 1 1 , , — • , , 1
20 30 40 so so
2Theta(Degree)
218
APPENDIX D
60-,
FLUX-AGFllOl
P 50
40-J
30-1
20
10
0-
-10-
1239.88562
-20 1 • 1 — I ' 1
1000 1050 1100 1150 1200 1250 1300 1350 1400
Temperature( C)
Figure D.l: DTA Graph of AFGllOl Flux
50 -
F L U X - A G F l 102
40-
30-
20
10 -
-10 -
1000 1050 1100 1150 1200 1250 1300 1350 1400

Temperature(°C)
Figure D.2: DTA Graph of AFG1102 Flux
219
FLUX-AGF1103
50-
40.
I 30-
20
10-
^
0-
-10-
1233.57629
— I • 1 ' 1 ' 1 ' 1 ' 1 ' 1 • 1
1000 1050 1100 1150 1200 1250 1300 1350 1400
Tempearture(°C)
2 0 -,
FLUX-AGF1104
D
I
a
Temperature( C)
220
50-
f FLUX-AGFl 105
p
"^ 40-
^
iD
tP 30-
5
0-
1226.99231
-10- T • y — I • 1
1000 1050 1100 1150 1200 1250 1300 1350 1400
Temperature( C)
F L U X - A G F l 106
1000 1100 1200 1300 1400 1500

Temperature ( C)
221
50-1
40-
FLUX-AGF1107
30-
20-
10-
0-
-10-
-20-
-30-
^0- 1244.57837
-50 1
1000 1200 1400
Temperature ( C)
10-1
FLTJX-AGF1108
0-
-10-
-20-
-30-
-50- 1191.77063
-60 1 —I
1000 1200 1400
Temperature (°C)
222
10-1
FLUX-AGF1109
0-
-10-
-20-
-30
1245.84021
-40-
-50 —I
1000 1200 1400
Temperature (°C)
i o -I
FLUX-AGFlllO
o-
I -10-
5
-20-
-30 -
-40 -
1248.36841
1000 1200 1400
T e m p e r a t u r e (°C)
Figure D.IO: DTA Graph of AFGlllO Flux
223
6 0 -I
FLUX-AGFllll
°Z,
1000 1200 1400

Temperature ( C)
Figure D.ll: DTA Graph of AFGllll Flux
ou —
544.34802 FLUX-AGF1112
40-
D
I 20-
0 -
20-
Vl 166.88
40- 1 > r -
400 600 800 1000
T e m p e r a t u r e ( C)
224
50-
FLUX-AGF1113
40-
8
c 30-
20-.
10-
.<u
0-
-10-
1229.25305
—I
1000 1200 1400
Temperature ( C)
50-,
40- FHJX-AGF1114
30-
20
10-^
0-
-10- 1236.76904
1 — I
1000 1200 1400
Temperature (°C)
Figure D.14: DTA Graph of AFG1114 Flu
225
60-1
FLUX-AGF1115
50-
I
40-
30-
20-
1 10-,
I o-
-10-
1241.88562
-20
—I
1000 1200 1400
Temperature ( C)
10
F L U X - A G F l 116
O-
-10-
-20-
-30-
1353.93823
-40-
1227.32607
-50
1000 1200 1400
226
50- FLUX-AGF1117
I
^ 4o^
30-
20-
10-
0-
-10-
1235.57629
— I
1000 1200 1400
10-1
FLUX-AGF1118
o-
-10-
-20-
-30-
1350.93823
-40-
1224.32507
-50 1
1000 1200 1400
227
I FLUX-AGF1119
-10
1000 1200 1400
Temperature (°C)
60-,
F L U X - A G F l 120
50
40
30-
20-
10-
O-
-10-
1338.8457
-20
1000 1200 1400
Temperature (°C)
228
5-|
O-
F L U X - A G F l 121
-5-
-10-
-15-
-20-
-25-
-30-
-35-
1246.57837
-40 1
1000 1200 1400
Temperature ( C)
10 -,
F L U X - A G F l 122
O-
-10-
-20-
-30 -
-40-
-50- 1189.52063
1000 1200 1400

229
10-1
F L U X - A G F l 123
-10-
-20-
-30-
-40- 1181.28101
: 1 —I
1000 1200 1400
Temperature (°C)
10
F L U X - A G F l 124
o -
-10 -
-20 -
-30 -
-40 -
-50 -
-60
1231.71277
1
1000 1200 1400
Temperature (°C)
230
F L U X - A G F l 125
1000 1200 1400

Temperature (°c)
10-1
o- F L U X - A G F l 126
-10-
-20-
-30-
-40-
-50-
1247.62524
-60 1
1000 1200 1400
Temperature (°C)
231
ioH
FLUX-AGF1127
0
-10-
-20-
-30-
-40-
1312.03577
-50-
1242.18359
-60-
-70 1
1000 1200 1400
Temperature (°C)
10-
f 0-
FLUX-AGF1128
-10-
-20-
-30-
-40-
1312.03577
-50-
1242.18359
-60-
-70 1
1000 1200 1400
Temperature (°C)
Figure D.28: DTA Graph of AFG 1128 Flux
232
10-
F L U X - A G F l 129
0-
-10-
-20-
-30-
-40-
1311.03577
-50-
1241.18359
-60-
-70 1 ' — I —
1000 1200 1400
Temperature (°C)
10-
FLUX-AGF1130
O-
-10-
-20-
-30-
-40-
1310.03577
-50 -
1240.18359
-60-
-70
1000 1200 1400
Temperature (°C)
233
10-
F L U X - A G F l 131
0-
-10-
-20-
-30-
-40-
1306.03577
-50-
1236.18359
-60-
T — I —
1000 1200 1400
Temperature (°C)
10
F L U X - A G F l 132
o
-10
.5
-20
-30
-40 -
1312.03577
-50-
1242.18359
-60-
1 — I
10G0 1200 1400
Temperature (°C)
234
APPENDIX E
Table E.l: ANOVA of Tensile Strength

Sum of Degree of
ource squares freedom Mean square f-value Prob>f
/lodel 6770.10 20 338.50 12.90 < 0.0001 significant
^-MnO 635.51 635.51 24.22 0.0005
J-CaF2 35.77 35.77 1.36 0.2677
:-MgO 146.52 146.52 5.58 0.0376
D-NiO 69.02 69.02 2.63 0.1331
E-F-Cr 203.58 203.58 7.76 0.0177
^B 28.89 28.89 1.10 0.3165
^C 6.13 6.13 0.23 0.6385
AD 107.64 107.64 4.10 0.0678
AE 1044.91 1044.91 39.82 < 0.0001
BC 3326.41 3326.41 126.77 < 0.0001
BD 103.53 103.53 3.95 0.0725
BE 6.13 6.13 0.23 0.6385
CD 5.41 5.41 0.21 0.6587
CE 0.14 0.14 5.359E-003 0.9430
DE 155.63 155.63 5.93 0.0331
A2 810.95 810.95 30.91 0.0002
B2 7.10 7.10 0.27 0.6132
:2 7.10 7.10 0.27 0.6132
D2 56.10 56.10 2.14 0.1717
E2 0.52 0.52 0.020 0.8907
lesidual 288.64 11 26.24
vack of 233.80 6 38.97 3.55 0.0927 not significant
*ure 54.83 5 10.97
vOr Total 7058.73 31
.td. Dev. =5.12 R-Squared = 0.9591
/lean = 547.83 Adj R-Squared = 0.8848
:.V.% = 0.94 Fred R-Squared = 0.1084
'RESS = 6293.35 Adeq Precision =13.889
235
Table E.2: ANOVA of Percentage Elongation
Degree
Sum of of Mean
Source squares freedom square f-value Probability >f
A-MnO 1.66 1.66 24.11 0.0005
B-CaF2 0.093 0.093 1.34 0.2715
C-MgO 0.38 0.38 5.51 0.0387
D-NiO 0.18 0.18 2.56 0.1378
E-Fe-Cr 0.53 0.53 7.65 0.0184
AB 0.073 0.073 1.06 0.3261
AC 0.016 0.016 0.23 0.6435
AD 0.28 0.28 4.07 0.0687
AE 2.72 2.72 39.45 < 0.0001
EC 8.64 8.64 125.24 < 0.0001
BD 0.27 0.27 3.84 0.0758
BE 0.016 0.016 0.23 0.6435
CD 0.014 0.014 0.21 0.6567
CE 4.00E-04 4.00E-04 5.80E-03 0.9407
DE 0.4 0.4 5.84 0.0342
A^ 2.12 2.12 30.7 0.0002
B^ 0.017 0.017 0.25 0.6252
C' 0.018 0.018 0.27 0.6165
D' 0.15 0.15 2.16 0.1699
E' 1.65E-03 1.65E-03 0.024 0.8799
Residual 0.76 11 0.069
Pure Error 0.14 5 0.028
Cor Total 18.37 31
Std. Dev. = 0.26 R-Squared =0.9587
C.V.% = 0.94 Pred R-Squared = 0.0954
PRESS = 16.62 Adeq Precision =
236
Table E.3: ANOVA of Impact Strength

Source f-value Prob>f
squares freedom square
Model 3272.33 20 163.62 18.06 < 0.0001
A-MnO 160.17 160.17 17.68 0.0015
B-CaF2 541.5 541.5 59.76 0.0001
C-MgO 337.5 337.5 37.25 0.0001
D-NiO 24 24 2.65 0.1319
E-F-Cr 60.17 60.17 6.64 0.0257
AB 361 361 39.84 < 0.0001
AC 49 49 5.41 0.0402
AD 81 81 8.94 0.0123
AE 306.25 306.25 33.8 0.0001
EC 72.25 72.25 7.97 0.0166
BD 56.25 56.25 6.21 0.0299
BE 16 16 1.77 0.2108
CD 182.25 182.25 20.11 0.0009
CE 144 144 15.89 0.0021
DE 256 256 28.25 0.0002
A^ 1.83 1.83 0.2 0.6616
B^ 334.13 334.13 36.88 <0.0001
C^ 132.46 132.46 14.62 0.0028
D^ 66 66 7.28 0.0207
E^ 221.83 221.83 24.48 0.0004
Residual 99.67 11 9.06
0.7679(not
Lack of Fit 38.83 6 6.47 0.53 significant)
Pure Error 60.83 5 12.17
Cor Total 3372 31
Std. Dev. =3.01 R-Squared =0.9704
Mean = 80.75 AdjR-Squared = 0.9167
Pred R-Squared =
C.V. % = 3.73 0.6667
PRESS =1124.01 Adeq Precision = 16.575
237
Table E.4: ANOVA of Vickers Hardness
Degree
Sum of of
Source squares freedom Mean square f-value Prob > f 1
Model 558.18 20 27.91 9.44 < 0.0002 significant ,
A-MnO 42.67 42.67 14.42 0.0030
B-CaF2 0.67 0.67 0.23 0.6443 1
C-MgO 28.17 28.17 9.52 0.0104 1
D-NiO 6.00 6.00 2.03 0.1821 ,
E-Fe-Cr 16.67 16.67 5.63 0.0369
AB 0.25 0.25 0.085 0.7767 1
AC 1.137E-013 1.137E-013 3.843E-014 1.0000 ^ 1
AD 25.00 25.00 8.45 0.0143
AE 72.25 72.25 24.43 0.0004
BC 256.00 256.00 86.55 < 0.0001 1
BD 4.00 4.00 1.35 0.2695 1
BE 2.25 2.25 0.76 0.4018
CD 2.25 2.25 0.76 0.4018
CE 1.00 1.00 0.34 0.5727 1
DE 16.00 16.00 5.41 0.0402 1
A^ 76.38 76.38 25.82 0.0004 ,
B^ 0.55 0.55 0.18 0.6759
C' 0.55 0.55 0.18 0.6759 1
D' 7.00 7.00 2.37 0.1521 1
E' 0.38 0.38 0.13 0.7272
Residual 32.54 11 2.96
Lack of Fit 27.04 6 4.51 4.10 0.0716 not significant '
Cor Total 590.72 31
Std. Dev. =1.72 R-Squared = 0.94 49
C.V.% = 1.05 Pred R-Squared = - 0.2244
PRESS = 723.27 Adeq Precision =11.663 1
238
Table E.5: ANOVA of Penetration

Source squares freedom square f-value Prob>f
A-MnO 0.027 0.027 0.1 0.7564
B-CaF2 1.63 1.63 6.19 0.0301
C-MgO 1.42 1.42 5.39 0.0405
D-NiO 1.44 1.44 5.46 0.0393
E-Fe-Cr 0.18 0.18 0.68 0.4259
AB 1.85 1.85 7.02 0.0226
AC 7.26 7.26 27.55 0.0003
AD 1.38 1.38 5.24 0.0429
AE 0.27 0.27 1.01 0.3374
EC 2.59 2.59 9.83 0.0095
BD 10.69 10.69 40.56 < 0.0001
BE 4.88 4.88 18.53 0.0012
CD 3.71 3.71 14.06 0.0032
CE 2.61 2.61 9.89 0.0093
DE 0.57 0.57 2.16 0.1695
A^ 3.4 3.4 12.88 0.0043
B^ 3.2 3.2 12.13 0.0051
C^ 6.84 6.84 25.93 0.0003
D^ 1.52 1.52 5.77 0.0351
E^ 0.13 0.13 0.51 0.4899
Residual 2.9 11 0.26
Lack of 2.27 6 0.38 2.98 0.1254 not significant
Pure 0.63 5 0.13
Cor Total 55.67 31
Std. Dev. =0.51 R-Squared =0.9479
Mean =6.17 Adj R-Squared = 0.8532
C.V. % = 8.32 Pred R-Squared = -0.0508
239
APPENDIX F
Figure F.l: Weld Joint Microstructure of AGFllOl Flux
Figure F.2: Weld Joint Microstructure of AGF1102 Flux
240
^^^
• * •
400X
4r>
# *^
E
241
I . t'S
"'•-^ V -
242
243
* - «
tem ~'^"*4^
Figure F.IO: Weld Joint Microstructure of AGFlllO Flux
244
•"•
<1k.x " i *
400X
Figure F.ll: Weld Joint Microstructure of A G F l l l l Flux
245
246
247
248
'$-
k>'
•*»
^ \ ^
li'lsVv-
mm
I 400X I
Figure F.19: Weld Joint Microstructure of A G F n i 9 Flux
•'le.^ •?»•
*%
4
•^
1\"»
1^ \
400X
249
250
Figure F,23: Weld Joint Microstructure of AGF1123 Flux
251
m
I
* V'fit
$*-
ft..
# 400X
ffli.
'*WKf '^P* • ^MH'
i
r * *•-
• M H ^ P ^ 400X •
1" ', *•«*. mMi^
252
i
lb
.. % *.
., J.
400X
I
253
254
r"
#- ^ \ A ip-
Jll'
%. . c -% iik
* 400X
^s •*^>
255
Figure F.33: Microstructure of Base Metal
256
APPENDIX G
IHIffffinm
cms.
Figure G.l: Weld Bead Profile for AGFllOl
Figure G.2: Weld Bead Profile for AGF1102
257
(iimiiti
cms.
Figure G.3 : Weld Bead Profile for AGF1103
258
a*-*
• ,-.M»|il»ae*-"
259
cms. 1
111 IMMli Ui i ]
cms.
260
IIHIIIIIIt
cms* 1
Figure G.IO: Weld Bead Profile for AGFlllO
261
iin|iit(|iirtf
tcms. It
Figure G.l 1: Weld Bead Profile for AGFl 111
%'
a-^fi,.i"iLg
m<|<tfi|HFiî
Figure G.12: Weld Bead Profile for AGFl 112
262
ini)iiii(inifttHfrmî

263
Figure G.16: Weld Bead Profde for AGF1116

264
265
miM JJJMl
icms. 1

266
Figure G.21: Weld Bead Profile for AGFIHI

267
crns. 1
268
•'m
4'^ .
, -. . ^1,
fc:?-'^;;^'^^^'^-:
269
wmmmm mmuM
nifiniii
•
cms. 21 3
Figure G.27: Weld Bead Profile for A G r i l 2 7
•i^mmfm.
inii(ffnffi
270
271
3l Jg
Figure G.32: Weld Bead Profile for AGFl 132
272

Eagar 021

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Eagar 021

Hochgeladen von

Copyright:

Verfügbare Formate

Ph.D.

SOME STUDIES ON DESIGN AND DEVELOPMENT OF

For the Award of the Degree

Under the Supervision of

Dr. Sachin Maheshwari, Professor Dr. Hari Singh, Professor

Date: o^-oh '^°^^

Dr. Sachin Maheshwari, Professor Dr. Hari Singh, Professor

Date: OV ^^- °^'^^3

The present investigation is an attempt to develop a methodology for design,

Table of Contents vii

1.1.8.3 Current density 11

1.1.8.4 Welding voltage 11

1.1.8.6 Electrical stickout 12

1.2.1 Classification of submerged arc welding fluxes 14

1.2.1.3 Agglomerated fluxes 15

3.3.1 Mixing of flux constituents 43

3.3.2 Preparation and baking of fluxes 44

(EXPERIMENTAL PROCEDURE) 47-66

4.1.3 Bulk density 49

4.1.6 Differential thermal analysis (DTA) of developed agglomerated

4.2 Development of Weld Joints 54

5.2.2 Grain size 69

5.2.3 Bulk density 69

5.3 Characteristics of Weld Joints 78

5.3.1.2 Modeling of percentage elongation 85

5.3.1.4 Modeling of Vickers hardness 96

5.3.2.2.2 Multiple response optimization using

5.4.1.2 Effect of flux constituents on penetration 135

5.4.2 Modeling of weld bead reinforcement 137

5.4.2.1 Effect of flux constituents on reinforcement 140

5.4.3 Modeling of weld bead width 142

5.4.3.1 Effect offluxconstituents on bead width 145

5.4.4 Optimization of bead morphology using desirability function 148

5.4.4.1 Single response optimization of bead morphology using

5.4.4.2 Optimal solutions using desirability function 149

5.4.4.3 Multiple response optimization of bead morphology

5.4.4.3.1 Model: Penetration, reinforcement and

5.5 Element Transfer Study 162

5.5.1 Carbon (C) transfer 163

5.5.2 Manganese (Mn) transfer 163

5.5.3 Silicon (Si) transfer 163

5.5.4 Chromium (Cr) transfer 164

5.5.5 Nickel (Ni) transfer 164

5.5.6 Sulphur (S) transfer 164

5.5.7 Phosphorus (P) transfer 165

5.5.8 Effect of element transfer on mechanical properties 165

CHAPTER 6 CONCLUSION AND SCOPE OF FURTHER RESEARCH 167-172

6.1 Conclusion 167

6.1.1 Bulk density, grain size,flowabilityand viscosity 167

6.1.2 XRDandDTA 168

6.1.3.1 Tensile strength 168

6.1.3.2 Hardness 169

6.1.3.3 Elongation 169

6.1.4 Weld joint microstructures 171

6.1.5 Bead morphology 171

6.1.5.1 Optimal parameters (Bead Morphology) using single and

multiple response optimizations

6.1.6 Element transfer

LIST OF PUBLICATIONS (JOURNALS)

APPENDIX A Tables of Sieving Results for Agglomerated Fluxes

APPENDIX B Figures of Grain Size Histogram for Agglomerated Fluxes

APPENDIX C Figures of XRD Graph for Agglomerated Fluxes

APPENDIX D Figures of DTA Graph for Agglomerated Fluxes

APPENDIX E Tables of ANOVA of Mechanical Properties and Bead