CHAPTER2

REVIEW OF LITERATURE

2.1 Introduction

Submerged Arc Welding (SAW) is one of the major metal fabrication

techniques in industry due to its reliability and capability of producing good quality

weld. The ability to join thick plates (as thick as 1.5 inch) in a single pass, with high

metal deposition rate has made this process useful in large structural applications.

Indeed various research works have been explored on various aspects of submerged

arc welding, yet investigations are still being carried on to study the phenomenon that

occurs during the process of submerged arc welding, and many other related matters,

so that the process becomes controllable more precisely, and can be monitored well,

both manually as well as automatically.

In submerged arc welding, various process parameters interact in a complicated

manner, and their interactions influence the bead geometry, bead quality as well as

metallurgical characteristics and mechanical properties of the weldment. Acceptability

of the weldment depends on various quality characteristics that confirm functional

requirements of the welded joint in the intended area of application. In most of the

cases, quality of the weld is left to depend on the past experience and working skill of

operator. But, with the advent of automation, it is now possible to design a machine

capable of selecting optimal process parameters to provide desired yield.

Research in the field of submerged arc welding is not new. The major trends of

research related to the field of submerged arc welding include the following: (a)

Chemical aspect of weldment (b) Electrode melting rate and weld bead (c)
13
Metallurgical and mechanical properties of the weldment (d) Process modelling,

simulation and their application in SAW (e) Reclamation of SAW.

2.2 Chemical Aspect of Weldment

Remarkable attention has been employed to study different aspects of weld

chemistry. It is quite understandable that compositions of base metal, electrode wire,

and flux impose profound effect on mechanical properties of the welded joint, which in

turn depends on microstructure of weld metal and heat affected zone (HAZ). Literature

provides ample scope to review the study in evaluating the effect of compositional

variation of base metal, wire electrode as well as welding flux on different aspects of

submerged arc weldment. Attempts have also been made by the researchers to

understand the nature of chemical reaction and element transfer during submerged arc

welding. Belton et al. (1963) remarked that the study of slag-metal reactions during

SAW of steel was made difficult by the complex temperature cycle which the metal

and slag experienced. They mentioned that vigorous stirring and mixing took place at

some favorable reaction temperature. This reaction temperature was lower than 2000 0C

for slag high in FeO but low in SiO2. The authors reported that higher weld pool

temperature was maintained with the increase in silica content of the slag whereas bead

width to depth of penetration ratio increased significantly with the increase in FeO

content in slag.

The effect of flux composition on weld metal in submerged arc welding of

QT35 alloy steel was studied by Bennet and Stanley (1966). They found that weld

14
metal tensile strength was independent of flux composition and heat input unless

alloying elements had been deliberately added to the flux. Weld metal impact

properties, however, had been affected remarkably by flux composition as well as heat

input. The authors mentioned that SiO2 and MnO in the flux would influence the levels

of Si and Mn in the weld metal, influencing the impact strength of the weldment. They

also discussed about transfer of sulphur and phosphorous between metal and slag.

Ferrera and Olson (1975) found that the basic constituents like CaO, CaCO 3,

CaF2 in flux produced shallow penetration whereas MgO led to deep penetration. The

similar conclusion came from the work of North (1978). Schwemme and Olson (1979)

reported about higher penetration due to increase in SiO 2 content of the flux. It was

reported that sodium and potassium salts, and other elements, which improved arc

stability and reduced cathode spot wandering, generally increased penetration.

Several investigators like Schwemmer and Olson (1979) and Davis and Bailey

(1982) had shown that a flux with higher viscosity would tend to confine the molten

weld pool, thus increasing the heat input for a given area and resulting into deeper

penetration. Chai and Eagar (1981) developed a theory for predicting slag-metal

equilibrium during submerged arc welding with fused neutral fluxes. The proposed

theory was capable of predicting the gain or loss of Mn and Si over a wide range of

flux electrode-base plate compositions. Their study represented a significant

advancement in the ability to predict and control the extent of the reactions related to

Mn and Si transfer in the weld pool. Chai and Eagar (1982) showed that the stability of

metal oxides during submerged arc welding was not directly related to their

15
thermodynamic stability. Some otherwise chemically stable fluxes might decompose

into suboxides in the presence of welding arcs, thereby providing higher levels of O 2 in

weld metal than those oxides, which did not form suboxides. It was shown that CaF2

would reduce the amount of oxygen in the weld metal, but the effect might be due to

dilution of the metal oxide rather than due to a direct chemical reaction. The effect of

CaF2 in reducing the level of weld metal oxidation was dependent upon the stability of

the metal oxide.

Mitra and Eagar (1984) studied the transfer of Cr, Si, Mn, P, S, C, Ni and Mo

between the slag and the weld pool of submerged arc weldment made with calcium

silicate and manganese silicate fluxes. They observed a strong interaction between Cr

and Si transfer but no interaction with Mn. The manganese silicate flux produced

lower residual sulphur whereas the calcium silicate fluxes were more effective for

removal of phosphorous. The effective oxygen reaction temperature was found to lie

between 17000C and 20000C for all elements studied. The study also revealed the

evidence of Cr and Mn loss by metal vaporization. In another report, Mitra et al.

(1984) gave a quantitative theory which was able to predict weld metal composition on

the basis of slag-metal reactions during submerged arc welding of steel.

Bead-on-plate welds were produced using twenty four fused reagent grade

submerged arc welding fluxes, selected from three flux systems, SiO2-MnO-FeO,

SiO2-MnO-CaF2 and SiO2-CaO-CaF2 by Jang and Indacochea (1987). The welds were

processed using AISI 1010 steel coupons, and a Lincoln L-50 (AWS type A5.17)

welding wire with a constant heat input of 3.0 kJ/ mm. The three flux systems were

16
selected because of their different oxygen potentials and their ability to produce welds

with a wide oxygen range (70 to 1400 p.p.m.). Qualitative and quantitative

metallography and chemical analysis were performed on the welds. Inclusion

morphology and volume fraction were observed to be affected by flux composition.

Inclusions of 1m and greater size are associated with grain boundary and blocky

proeutectoid ferrites, while inclusions of 0.6 m and smaller size are linked with the

presence of acicular ferrite.

Polar et al. (1991) evaluated the relative effects of thermo-chemical reactions

on the transfer of elements, particularly Mn, from the flux to the weld metal in

submerged arc welding of steel. Davis and Bailey (1991) described that the complex

change in composition during submerged arc welding of C-Mn steels were related to

flux composition and weld metal inclusions which formed the final reaction products.

They analyzed the chemical factors controlling the transfer of elements into the weld

pool and also observed that high temperature reactions in the arc plasma were followed

by reactions in the slag and the weld pool. Gupta and Arora (1991) used five different

commercial fluxes in their investigation of submerged arc welding of mild steel, and

found that the welding parameters and the flux basicity appreciably affected the depth

of penetration and width of the weld bead. The reinforcement was also influenced by

welding parameters, but not appreciably by flux basicity. With respect to HAZ width,

the authors observed lower values with highly basic fluxes as compared to lower

basicity fluxes, at lower welding current range. The investigators, however, could not

find any definite trend of HAZ width with reference to flux basicity and felt that

17
different widths of HAZ obtained with different fluxes (other welding parameters

being kept constant) might be due to different proportions of heat shared by the molten

flux and the weld metal.

Pandey et al. (1994) investigated the influence of submerged arc welding

(SAW) parameters and flux basicity index on the weld chemistry and transfer of

elements manganese, silicon, carbon and sulphur, five fluxes and different values of

the welding parameters being used for study. The welds were produced as a bead on a

mild steel plate. The weld metal composition showed, in general, gain of silicon and

loss of carbon, manganese and sulphur elements. The results showed that welding

current and voltage have an appreciable influence on element transfer, as well as on

weld composition. On the other hand, except for weld-metal silicon, the basicity index

of the fluxes had only a minor influence. Weldments properties such as strength,

toughness and solidification cracking behaviour were affected by chemical

composition. Data reported in this paper can be used to specify welding variables for

desired weld composition.

Ana et al. (2003) conducted a study for chemical and structural characterization

of fluxes used in submerged arc welding process, which enabled one to quantify the

ions that might be present in the plasma arc due to the fluxes. Their analysis was

capable of predicting reactions that occur in the weld pool. Kanjilal et al. (2004)

developed a prediction model for submerged arc weld metal chemical composition in

terms of flux ingredients with the help of statistical experiments for mixture.

Experiments were performed to obtain bead-on plate weld deposits at some specific

18
process parameters (current, voltage, speed, electrode extension) using CaO 2-MgO-

CaF2-Al2O3 flux system. They reported that some of the individual flux ingredients and

their binary mixtures had predominant effects on weld metal oxygen, manganese,

silicon, sulphur, nickel and carbon contents. The predicted results showed a good

agreement with the experimental results.

Farias et al. (2004) investigated to assess the operational behaviour of

ANSI/AWS A5.1-91 E6013 type electrodes when 0, 8, and 16% of quartz (100%

SiO2) was replaced with Wollastonite (calcium Silicate, 50% SiO 2- 50% CaO) in the

coating composition. It was concluded that the replacement of quartz with

Wollastonite in the coating increased slag basicity and decreased all weld metal silicon

and oxygen contents. It seemed possible to increase the weld metal toughness without

losing the excellent operational behaviour of rutile electrodes. Mohan and Pandey

(2005) analyzed the effect of welding parameters on the Mn, Si, C and Cr content of

the weld metal through quadratic response surface modelling. It was observed that the

welding parameters and basicity index affected the Mn, Si and C content of the weld

metal. Cr content of the weld metal was influenced by the basicity index of the flux

used.

2.3 Electrode Melting Rate and Weld Bead

Literature depicts that ample work has been done to study the effect of

parameter variation, base metal and flux composition on electrode melting rate during

SAW welding. Robinson (1961) pointed out that the significant factors controlling

SAW electrode melting rate were current, composition used, electrode extension,

19
electrode diameter and flux. Tusek (1999) developed a mathematical model for

calculation of the melting rate on the basis of the physical principles of the welding arc

and of the wire extension heating due to current conduction and mutual influence of

the welding arcs. Tusek and Suban (2003) investigated some aspects of multiple wire

submerged arc welding (and cladding) with metal powder addition. They observed that

the use of metal powder increased the deposition rate, welding arc efficiency and

reduced the shielding flux consumption. By incorporating the metal powder addition

technique, it was possible to alloy a weld or a cladding with optimal chemical

elements.

Weld pool size and shape, extent of heat affected zone, microstructure, mode of

metal transfer, droplet size, frequency of transfer, arc force, bead geometry and shape

relationships, burn off rate and equilibrium in submerged arc welding process are

important criteria in determining the weld ability of any metal. Useful predictions

about these characteristics can be made if there exists a correlation between the quality

parameters of the weld and welding process parameters controlling the same.

Kaae (1968) observed that heat affected zone width was influenced by heat

input, thickness, initial temperature and thermal conductivity of the base metal.

Renwick and Patchett (1976) studied the characteristics of the weld bead penetration,

melting rate under variable operating current conditions and found that those increased

with the increase in current. They also investigated that increase in welding voltage

produced flatter and wider bead and increased flux consumption. Gupta and Arora

(1993) studied the effect of welding parameters on weld bead geometry as well as

20
HAZ and concluded that welding parameters and flux basicity appreciably influenced

the depth of penetration, weld bead dimension and width of HAZ at the centre line and

at the toe of the weld bead. Malin (2001) conducted experiments employing modified

refractory flux welding and studied the effects of welding variables on formation of

root (backside) welds, including the root bead (deposit inside the groove) and root

reinforcement (deposit outside the groove). The work also revealed that welding

variables produced profound and sometimes conflicting effects on the root welds

shape. For example, increasing the current increased the deposition rate and the depth

of joint penetration; however, root bead shape deteriorated and slag pockets formed,

which might provoke defects in the weld. In another paper (Part 2) Malin (2001)

established relationships between shape of the root weld and variations in joint

geometry. The work revealed the effects of joint geometry in terms of root opening,

included angle, root face and plate misalignment on root welds including the root bead

and root reinforcement. Murray (2002) established analytical relationships between

welding parameters and process variables by regression and dimensional analysis of

the experimental data. Non-dimensional variables were used to correlate experimental

data to obtain accurate analytical relationships between welding parameters, arc

process variables and bead geometry.

2.4 Metallurgical and Mechanical Properties of Weldment

Metallurgical characteristics of the weld metal as well as heat affected zone

(HAZ) are very important because this directly influence the weld mechanical

21
properties and joint performances. It is well known that the micro structure of the weld

metal is different from the microstructure of base metal as well as HAZ. The weld

microstructures, however, are somewhat different with respect to distributions of

pearlite and ferrite, their amounts, grain sizes etc. depending upon the welding

conditions adopted. In a pass of the welding torch, material is rapidly heated to the

maximum temperature and allowed to cool more slowly by conduction of heat into the

bulk of the parent metal. Phase change can occur depending on the temperature

reached. Sufficiently far from the weld pool, the material remains unaffected. The

region next to the fusion zone where microstructural changes have occurred but no

melting of base metal has effected is known as heat affected zone. Such micro-

structural changes may affect the mechanical properties of the weld and need to be

controlled. The weld metal microstructure is controlled mainly by the cooling cycle. At

lower energy input (i.e. with low level of current), the time for solidification is less.

This rapid cooling promotes smaller grains. With higher energy input, the time

required for solidification decreases, and, therefore, cooling rate slows down which

yields coarse grains. Coarse grains in the microstructure generally indicate lower

hardness and low tensile strength.

Mechanical properties are important characteristics of the weldment that must

confirm to the application feasibility as well as functional requirements of the welded

joint. These include hardness, impact strength (toughness), yield strength, ultimate

tensile strength, percentage elongation, resistance to wear and corrosion, etc. These

mechanical properties greatly depend on weld microstructure, which in turn is related

22
to cooling condition, composition of base metal, wire electrode as well as flux.

Moreover, welding process parameters also impose direct/indirect influence on weld

mechanical properties and microstructure.

Charles (1979) examined the effects of two fluxes of different basicities when

used with two experimental wires and a columbium bearing base plate. The flux

basicity was seen to have an effect on the transfer of hardenability elements to the weld

deposit and on the resultant microstructure and impact properties. The use of a fully

basic flux with Mo-Ti-B wires aided in the suppression of proeutectoid ferrite in the

weld metal microstructure and thereby improves the notch toughness of the weld.

Quantitative image analysis was used to document the effects of flux basicity on

microstructure and non metallic inclusions.

Fleck et al. (1986) found that filler material and flux composition in SAW

would influence the growth of austenite considerably. Tandon et al. (1988) observed

that heat input and cooling rate played major roles in deciding the extent of HAZ, the

microstructure and the hardness. They pointed out that higher cooling rate led to

smaller HAZ width. Smith et al. (1989) demonstrated that the notch toughness of the

coarse grained HAZ decreased with increase in energy input. They also found that

stress relieving reduced the notch toughness of both the weld metal and HAZ as a

result of embrittlement caused by carbide precipitation. Reddy et al. (1991)

investigated the effect of electrode polarity and welding current on mechanical

properties in multi pass submerged arc welding. They studied on the influence of

welding current with two polarities over mechanical properties of the weldment. They

23
also found that the hardness value increased in coarse microstructure region than the

reheat refined region of the weld deposit. The authors studied submerged arc welds

made by C-Mn filler wire and concluded from observations that (a) at a given welding

parameter the use of electrode negative polarity reduced the extent of reheat

refinement than that occurred while using electrode positive, (b) at any electrode

polarity higher amount of reheat refinement regions in the weld metal could be

obtained with a decreased welding current. (c) in multipass welding, comparatively

higher hardness resulted in the coarse microstructural region than that in reheat refined

region of the weld deposit at any given welding parameter and (d) yield strength and

ultimate tensile strength were higher with DCEN than that observed with DCEP.

Joarder et al. (1991) made an extensive study on the microstructure of weld

metal and heat affected zone of plain carbon steel plate in SAW. They noticed that

depending on the number, size and distribution of inclusions, the weld metal

microstructure varied. Inclusions were found to decrease in number from top to base of

weld metal and because of the prevalence of varying cooling rates in weld metal, a

wide range of microstructures, such as pearlite, grain boundary ferrite with pearlite and

side plate with cementite along the side plate boundaries were observed. They also

noted that a limited number of large inclusions helped in the formation of acicular

ferrite, whereas a large number of smaller inclusions favoured grain boundary ferrite

formation.

Khallaf et al. (1997) described cracking behaviour during the submerged arc

welding of medium carbon steel plates (0.45% C). They discussed the results of tests

24
made to examine the effect of welding variables, plate fabrication conditions (rolling

reduction ratio) and plate thickness on cracking susceptibility. It was found that the

cracking susceptibility increased with an increase in the welding current and decreased

with an increase in the welding speed or the electrode wire feed rate. It also increased

with increase in the plate rolling reduction ratio and with decrease in the plate

thickness. These results have a practical significance for industrial fields, especially

where welded machine spare parts are concerned.

Bhadeshia (1997) gave quantitative relationships between microstructure and

mechanical properties so that the variety of models could be consolidated and used

directly in the design process. The paper provided some information to deal with a few

of the important mechanical properties and described some of the innovative

treatments under development for the production of better welds or for the exploration

of welding technologies in novel applications.

Zhang and Dorn (1999) investigated the mechanical properties of nuclear

pressure vessel steel, A508CL3, and its welded joints by using the microshear test

method. The fracture toughness of A508CL3 steel and its welds were also estimated.

Moreover, a comparison had been carried out between the conventional test,

microshear test and fracture mechanics test. In addition, the possibility of using the

microshear test on the surveillance program of nuclear pressure vessel embrittlement

due to neutron irradiation had also been considered in detail, and the results indicated

that the microshear test can be used successfully to estimate the degradation of

mechanical properties both for A508CL3 steel and its welded joints. It had been found

25
that the lower the microshear toughness, the smaller the Charpy V-notch impact

toughness and fracture toughness, as well as the tearing modulus. Finally, the results

showed that the microshear test method may be developed as a supplemental test

method.

Eroglu et al. (1999) investigated the effects of coarse initial grain size with

varying heat inputs on microstructure and mechanical properties of weld metal and

heat affected zone (HAZ). It was concluded that the coarse initial grain size had a great

influence on the microstructure, hardness and toughness of HAZ of low carbon steel.

The investigators recommended a higher heat input to obtain maximum toughness of

the HAZ in the welding of grain-coarsened low carbon steels, taking into consideration

the plate thickness. Surian and Vedia (1999) made experiments with all weld metal

deposited with various types of electrodes, and reported the role of different alloying

elements such as manganese, carbon and chromium on the tensile properties, hardness

as well as on the microstructure of the submerged arc weldment. They suggested

criteria for selecting the weld metal composition leading to optimal combination of

tensile strength and toughness. It was concluded that an increase in Mn, C or Cr

individually produced an increase of tensile strength and hardness. Ghosh and Ahmed

(1999) built up an analytical model for quantitative analysis of micro structural

constituents of a multi-pass weld. The authors experimented with multi-pass

submerged arc welding of C-Mn steel blocks and developed software to carry out the

complex analysis of microstructure of the weld along with execution of database, for

26
estimation of mechanical properties using their empirical relationships with the matrix

microstructure.

Christopher et al. (2000) made a comparison between fatigue crack initiation

and growth predictions for a submerged arc welded butt joint in ASTM A517 steel.

The fatigue tests of the butt-welded joint involved a variable amplitude spectrum

consisting of three storm sequences per year. Fatigue crack initiation, coalescence and

growth were monitored using a localised potential drop system. Multiple fatigue crack

initiation sites were found along the weld toe of the specimen, and this was followed

by fatigue crack coalescence. Fatigue life, crack initiation and growth predictions were

carried out using local notch strain and fatigue crack propagation approaches

respectively. A comparison between the predicted and experimental fatigue lives

indicated that the notch strain and fatigue crack propagation approaches were

conservative while the strain life predictions were unconservative.

Eroglu and Aksoy (2000) investigated the effects of initial grain size with the

varying heat inputs on the microstructure and toughness of intercritical heat affected

zone (HAZ) of low-carbon steel. In the welding experiments, SAE 1020 steel

specimens in hot-rolled, in grain-refined and in grain-coarsened conditions were

welded by a submerged arc welding machine with the heat inputs of 0.5, 1 and 2 kJ/

mm. Following the welding, microstructure, hardness and toughness of the intercritical

HAZs of the specimens were investigated. The determination of microstructure and

measurement of hardness in the intercritical HAZs were performed on the samples

taken from the welded specimens, while toughness values were obtained using the

27
weld thermal simulation technique. From the results, a relationship was established

between the heat input, initial grain size, microstructure, hardness and toughness of the

intercritical HAZ. From the results of the toughness tests and microstructural

observation, it was seen that the fine initial grain size was effective on the formation of

ductile phases and on the higher toughness, whereas the coarse initial grain was

effective on the formation of brittle phases and on the lower toughness at the same heat

input. As a result, considering the microstructure, hardness and toughness of the

intercritical HAZ, a higher heat input for both the coarse initial grain size and fine

initial grain size gave good results. However, it was also seen that a lower heat input

can be used in the welding of low carbon steel with fine initial grain size with respect

to the toughness of the intercritical HAZ.

Kostrivas and Lippold (2000) developed a technique that would allow

evolution of the fusion boundary to be studied under controlled thermal conditions. In

their report, non-dendrite equiaxed grain microstructures were simulated in Li-bearing

Alloy 2195. Wojnowski et al. (2000) worked for metallurgical assessment of the

softened HAZ region during multipass submerged arc welding of Cr-Mo-V steels.

Their study focused on the effect of multiple thermal cycles on the development of the

softened zone. Peng et al. (2001) studied the variations of elements between wire and

weld metal, and discussed the effects of Mo, B and Ti on the microstructure and

mechanical properties. The study concluded that fine and uniformly distributed

inclusion particles in the weld metal could promote the formation of acicular ferrite.

TiN particles could pin the prior austenite grain boundaries of the heat affected zone

28
and inhibit the growth of austenite grains. Boron was found to segregate in the prior

austenite grain boundaries and suppress the growth of pro-eutectoid ferrite.

The relationship between microstructure and toughness of weld deposits was

studied by means of hardness, Charpy-V notch and metallographic tests in specimens

cut transversely to the weld beads (Jorge et al., 2001). Qualitative and quantitative

analysis of microstructural constituents and fine phases were made by light optical and

scanning electron microscopy, respectively. In this investigation, a selective etching

method was employed in order to distinguish the martensiteaustenite (M/A)

microphase constituent from carbides. The results showed that chromium impairs

impact toughness, although it promotes an increase in the percentage of acicular ferrite

(AF). In addition, it was observed that an increase in carbon content promoted a further

decrease in impact toughness due to the higher volume fraction of the M/A constituent.

Wang et al. (2002) investigated the effects of the heterogeneous microstructure

at the base metal, the heat affected zone and weld metal on hydrogen permeation in

thermo-mechanical controlled rolling steel weldments. The base metal with equiaxed

refined ferrite and scattered fine grain pearlite has the highest permeation rate and

effective diffusivity. The HAZ with bainite showed the lowest values of the

permeation rate and effective diffusivity. Weld metal yields a higher permeation rate

coupled with an intermediate diffusivity value. The hydrogen apparent solubility was

low for the base metal, intermediate for the HAZ and high for the weld metal. The

mechanisms of hydrogen diffusion path and hydrogen traps were discussed and

experimentally confirmed using the hydrogen microprint technique. The high

29
diffusivity paths and the hydrogen trapping site were the grain boundary and the

ferrite/carbide interfaces for both the base metal with refined ferrite and the HAZ with

bainitic microstructure. The spaces among the basket weave acicular ferrite, where the

martensite and the retained austenite constituents present, were the main hydrogen

trapping sites for the weld metal.

Heat affected zone affects the mechanical and metallurgical properties of the

weldment. Heat affected zone mainly depends on heat input which is influenced by the

welding parameters viz.: voltage, welding current and welding speed (Wu and Gao,

2002). An increase in voltage and welding wire-feed rate and a decrease in welding

speed increase the heat affected zone (Stuck and Stout, 1972).

Basu and Raman (2002) experimented to obtain bead-in-groove weld under

isoheat input conditions by submerged arc welding using quenched and tempered (Q &

T) HSLA steel, a suitable welding wire and an agglomerated basic flux (Basicity index

3.1). The heat input was adjusted by varying welding current and welding speed with

the machine operated at constant voltage mode. It was noticed that all the parameters

namely, prior austenitic grain size, inclusion characteristics, cooling rate influencing the

volume fraction of intragranularly precipitated acicular ferrite in the weld showed

significantly different dependence at a specific heat input depending upon the welding

current and speed combinations used. They defined a new cooling rate parameter based

on the weld nugget cross-sectional area and fusion zone/heat affected zone interface

length. Utilizing multiple regression analysis, a correlation between acicular ferrite

content and the different welding parameters was established with 90% correlation

30
coefficient. That correlation could be utilized in selecting the trial welding parameters

for similar grades of steel substrates and consumables with the aim to maximize the

acicular ferrite content. In the course of investigation performed by Gulenc and

Kahraman (2003), worn parts were welded using submerged arc welding process with

various wires and fluxes. These welded parts were subjected to pin on disk wear tests

under different loads to examine the changes in the hardness and microstructures. They

concluded that weld hardness and wear resistance were dependent on the chemical

composition of the weld wire and flux. Ana et al. (2005) focused on the effect of flux

composition for the microstructure and tensile properties of submerged welded AISI

1025 steel. In their study, three flux compositions were used with a low-carbon

electrode. The study revealed that the presence of acicular ferrite was detected for welds

of fluxes with the highest content of titanium oxide. The presence of acicular ferrite was

found to influence the yield and ultimate tensile strength of the weldment. The

elongation and percentage of reduction of cross sectional area were affected by the

inclusion volume percentage. Predicted values obtained from computer programme, for

tensile properties and microstructure were compared with that obtained through

experiments.

Kolhe and Datta (2005) conducted a detailed study on the microstructure, phase

analysis, mechanical properties and HAZ width of submerged arc weld in multipass

joint and heat affected zone of 16 mm thick mild steel plate. The bulk hardness, impact

energy and micro hardness of the multipass welded joint were tested by Rockwell

hardness testing machine, Charpy V Notch testing machine and Vickers micro

31
hardness tester. The various sub zones in the microstructure observed in the HAZ were

spheroidized, partially transformed, grain refined and grain coarsened. The variation in

hardness of weld metal, fractured surface and base metal were compared. The authors

made an attempt to establish correlation between hardness and microstructures as

observed in the weld and HAZ.

The aim of the study conducted by Quintana et al. (2003) was to evaluate the

capacity of various agglomerate fluxes to transfer to the welding bath the elements Cr,

Mn and C from alloy loads formed with different ratios of FeCr, FeMn and graphite.

To do this, a vitreous matrix was synthesised and melted in an electric arc furnace,

using a formulation of the minerals, and alloyed loads were formed with different

ratios of the components established using a McLean Anderson experimental design.

Shan et al. (2004) conducted study to investigate microstructure and surface

hardness of the cladding on samples prepared under different welding conditions. The

results showed that the retained austenite in the cladding increases with the increased

welding current and reduced travel speed. The wear behaviour of the clad was studied

using ball-on-disk tribometer. Wear mechanism was analyzed based on the analysis of

the worn surfaces both the clad and ball by scanning electron microscope (SEM)

equipped with an energy dispersive spectrometer. The scanning electron microscopic

examination of the worn surface showed that a layer of oxide film formed on the worn

surface. Oxidation wear mechanism controls the wear process. The spalling of the

oxide was caused by the repeated rubbing fatigue mechanism.

32
In the work of Pathak and Datta (2004), a three-dimensional transient finite element

analysis of heat transfer in arc welding had been done to predict the different zones of

microstructures. The problem was solved by taking several important factors into

consideration (such as temperature dependence of material properties, a Gaussian

distribution for the welding arc, enthalpy, etc.). The model was indirectly verified by

conducting some experiments on submerged arc welding. The weld metal zone, coarse

grained zone and fine grained zone were theoretically estimated by using three-

dimensional finite element analysis and compared with the experimentally measured

values. It was found that theoretically estimated zones corresponded well to the

experimentally measured zones. A correlation was also developed to estimate the arc

radius from the weld metal zone.

Jerzy and Pawel (2005) analyzed the influence of heat input in submerged arc

welding (SAW) of duplex steel UNS S31803 on kind and quantity of welded butt

joints defects. Defects were identified by a radiographic method. Radiographs had

been classified on the basis of Polish and European Standards. The mechanical

properties of the joints and value of ferrite share test had been done. The effect of

welding heat input influence on mechanical properties of test joints using heat input

from 2.5 to 4.0 KJ/mm was analyzed. It was shown that submerged arc welding of

duplex steel with the heat input from 2.5 to 4.0 KJ/mm had no negative influence on

mechanical properties of the joints. Experiment showed that welding with heat input

up to 3.0 kJ/mm reduces welding defects of joints, e.g. slags, lack of a joint penetration

for plates thickness of 1023 mm, as well as sticks, cracks, and a thorough decrease of

33
other defects existence. Usage of larger welding heat input provides the best joints

quality but decreases the joints control and repair costs.

Bhole et al. (2006) investigated the effects of additions of nickel (Ni),

molybdenum (Mo), and Ni and Mo together on the impact toughness of API HSLA-70

steel by submerged arc welding and microstructural factors which affect the impact

toughness. Ni additions resulted in low impact toughness and an increased fracture

appearance transition temperature in weld metal. The above influences of Ni should be

attributed to the formation of acicular ferrite (AF) suppressed by increasing the Ni

content. Conversely, the combined presence of Ni and Mo in the weld metal decreased

the volume fractions of grain boundary ferrite (GBF) and promoted formation of high

toughness of AF. The increase of Mo content created an acicular ferrite-predominant

weld metal microstructure with impressively improved toughness. Mo addition of

0.881 wt. % in the weld metal gave the optimal impact toughness at 450C with a

microstructure of 77% AF and 20% granular bainite.

Kanjilal et al. (2006) developed rotatable designs based on statistical

experiments for mixtures to predict the combined effect of flux mixture and welding

parameters on submerged arc weld metal chemical composition and mechanical

properties. Bead-on-plate weld deposits on low carbon steel plates were made at

different flux composition and welding parameter combinations. The results showed

that flux mixture related variables based on individual flux ingredients and welding

parameters have individual as well as interaction effects on responses, viz. weld metal

chemical composition and mechanical properties. In general, two factor interaction

34
effects were higher than the individual effect of mixture related variables. Amongst

welding parameters, polarity was found to be important for all responses under study.

In the study of Abilio et al. (2007), a normalised fine grain carbon low alloy

steel, P355NL1 (EN10028-3), intended for service in welded pressure vessels, where

notch toughness is of high importance, had been investigated. Applications with this

steel usually require the intensive use of welds. The influence of the automated

submerged arc welding in the mechanical performance was investigated. In this paper

the low and high cycle fatigue and crack propagation behaviours were compared

between the base and welded materials. Several series of small and smooth specimens

as well as cracked specimens made of base, welded and heat affected materials,

respectively, were fatigue tested. Strain, stress and energy based relations for fatigue

life assessments, until crack initiation, were evaluated based on experimental results

and compared with the base and welded materials. Finally, the fatigue crack

propagation behaviours were compared with the base, welded and heat affected

materials.

Prasad and Dwivedi (2008) investigated the influence of the submerged arc

welding (SAW) process parameters on the microstructure, hardness, and toughness of

HSLA steel weld joints to analyze the results on the basis of the heat input. The SAW

process was used for the welding of 16 mm thick HSLA steel plates. The weld joints

were prepared using comparatively high heat input (3.0 to 6.3 KJ/mm) by varying

welding current (500700 amperes) and welding speed (200300 mm/min). Results

showed that the increase in heat input coarsens the grain structure both in the weld

35
metal and heat affected zone (HAZ). The hardness had been found to vary from the

weld centre line to base metal and peak hardness was found in the HAZ. The hardness

of the weld metal was largely uniform. The hardness reduced with the increase in

welding current and reduction in welding speed (increasing heat input) while the

toughness showed mixed trend. The increase in welding current from 500 to 600

amperes at a given welding speed (200 mm/min or 300 mm/min) increased toughness

and further increase in welding current up to 700 amperes lowered the toughness.

Scanning electron microscopy of the fractured surfaces of impact test specimen was

carried out to study the fracture modes. Electron probe micro analysis was carried out

to investigate the variation in wt. % of different elements in the weld metal and HAZ.

Babu and Nattarajan (2008) described an experimental work carried out to

evaluate and compare corrosion and its inhibition in SA 516 Gr.70 carbon steel boiler

weldments made by submerged arc welding process with four different heat inputs

exposed to hydrochloric acid medium at 0.1 M , 0.5 M and 1.0 M concentrations. The

parent metal, weld metal and heat affected zone are chosen as regions of exposure for

the study carried out at 1000C. Electrochemical polarization techniques such as Tafel

line extrapolation and linear polarization resistance had been used to measure the

corrosion current. The role of hexamine and mixed inhibitor (thiourea + hexamine in

0.5 M HCl), each at 100 ppm concentration was studied in these experiments.

Microstructural examination, surface characterization using scanning electron

microscopy and X-ray diffraction had also been made on samples in order to highlight

the nature and extent of film formation.

36
Kolhe and Datta (2008) conducted a detailed study on the microstructure, phase

analysis, mechanical properties and HAZ width of submerged arc weld metal multi-

pass joint and heat affected zone of 16 mm thick mild steel plate using trinacular

metallurgical microscopy, with Onix Vision Video Microscope Projection System

including camera and image analysis software. The bulk hardness, impact energy and

microhardness of a multipass welded joint were tested by Rockwell hardness testing

machine, Charpy-V Notch test and Vickers microhardness test. The various sub zones

in the microstructure observed in the HAZ of submerged arc weld were spheroidized,

partially transformed, grain refined and grain coarsened. The variation in hardness of

weld metal, fractured surface and base metal were compared with the microstructure to

get a defect free weld, and it was also correlated with the microstructure of weld metal

and heat affected zone. The main purpose of this work was to investigate and correlate

the relationship between the various parameters, mechanical properties and

microstructure of single V butt joint of mild steel plate, and also to perform the

phase analysis of the multipass welded joint to get defect free welded structures.

Ana et al. (2009) presented a study of the effect of TiO2 additions in fluxes on

the mechanical properties and microstructure of the weld metal formed during

submerged arc welding of ASTM A-36 steel plates. Four fluxes with about 9, 12, 15

and 18% Ti were used with a low-carbon electrode. The welding conditions were kept

constant. The microstructure of the weld metal for each flux consisted mainly of

equiaxed ferrite and acicular ferrite. The increase in the percentage of acicular ferrite

and a decrease in its length were observed with an increase in titanium content. The

37
increase in titanium content in fluxes also improved the toughness and ductility of the

welds.

2.5 Process Modelling, Simulation and their application in SAW

In recent years, there has been a significant growth in the use of automatic and

robotic welding systems. These systems can be used effectively when mathematical

models that correlate welding process parameters to the weld bead geometry and shape

relationships are available. Therefore, it becomes important to represent the welding

process behaviour through mathematical models. Literature depicts that work has been

explored on various aspects of modelling, simulation and process optimization in

submerged arc welding. The various approaches for modeling in welding include (i)

Response Surface Methodology (RSM) and Factorial Design (ii) Artificial Neural

Network (ANN) Modelling (iii) Hybrid Optimization Techniques (Unal. and Dean,

1991; Rowlands et al., 2000; Antony and Antony, 2001; Maghsoodloo et al., 2004).

2.5.1 Response surface methodology (RSM) and factorial design for process

modelling

Response Surface Methodology (RSM) is a combination of mathematical and

statistical techniques, and is useful for modelling and analyzing the problems in which

several independent variables influence a dependent variable or the process response.

This technique provides a quadratic second order model that represents mathematical

relationship among various predictors (factors) and the process response. The RSM

38
offers some advantages while solving optimization problem of a response variable

within experimental domain. However, this depends on proper selection of the Design

of Experiment (DOE) (Myers and Montgomery, 1995). Box-Behnken designs are

response surface designs, specially made to require only 3 levels, coded as -1, 0, and

+1. Box-Behnken designs are available for 3 to 10 factors. They are formed by

combining two-level factorial designs with incomplete block designs. This procedure

creates designs with desirable statistical properties, but most importantly, with only a

fraction of the experiments required for a three-level factorial. Because there are only

three levels, the quadratic model is appropriate.

For full factorial design of experiment, as the number of factors and their levels

increase, the number of experimental runs also increases exponentially. This results an

increase in experimental cost as well as time for conducting experiments. Compared to

full factorial design, Box-Behnken design is more efficient to establish process

behaviour by a mathematical model. Moreover, this design reduces experimental run

number thereby saving time and cost.

Chandel et al. (1994) described a software system for computing the size and

shape of bead-on-plate submerged arc welds. The system is based on algorithms

originated by (Mcglone et al., 1978; Chandel, 1988; and Yang et al., 1993).

Essentially the system consists of a specially designed interface for welding / materials

/ design / fabrication engineers, automated plotting for parametric studies, a simplified

data base for storing / editing / retrieving frequently used welding parameters and

pictorial graphics for displaying weld size and shape. Chandel et al. (1997) presented

39
theoretical predictions of the effect of current, electrode polarity, electrode diameter

and electrode extension on the melting rate, bead height, bead width and weld

penetration, in submerged arc welding.

Gunaraj and Murugan (1999) gave a clear idea to show how the relationship

between the input process parameters and the features of weld bead geometry would be

constructed correctly. In their investigation, mathematical models were developed to

study the effects of process variables and heat input on various geometrical aspects,

like width of HAZ weld interface and grain growth as well as grain refinement regions

of the HAZ. In another publication, Gunaraj and Murugan (1999) highlighted the use

of RSM by designing a four-factor, five level central composite rotatable design matrix

with full replication for planning, conduction, execution and development of

mathematical relationships for modelling of the welding phenomena. The models were

helpful for predicting weld bead quality and could also be applied for process

optimization. Gunaraj and Murugan (2000) developed a model using the five level

factorial technique to relate the important process control variables welding voltage,

wire feed rate, welding speed and nozzle to plate distanceto a few important bead

quality parameterspenetration, reinforcement, bead width, total volume of the weld

bead and dilution. The model thus developed was checked for its adequacy with the F-

test. They highlighted quantitatively as well as graphically the main/direct and

interactive effect of the process control variables on important bead geometry

parameters.

40
As a continuation of the study, Gunaraj and Murugan (2000) proposed a mathematical

model to relate the process parameters with the weld bead quality parameters including

total volume of the weld bead. They used MATLAB software package to optimize

(minimize) the weld bead volume. The mathematical models thus developed were

capable of predicting the weld bead quality parameters and setting the process

parameters at optimal values to achieve preferable weld quality and at the same time

high productivity. Murugan and Gunaraj (2005) proposed mathematical models for

submerged arc welding of pipes using five level factorial technique to predict the

critical dimensions of weld bead geometry and shape relationships: height of

penetration, height of reinforcement, width of weld bead, penetration size factor and

reinforcement form factor. The significance and adequacy of the developed models

were tested by using F-test and t-test respectively. The predictions as given by the

models were represented graphically to show direct as well as interactive effects of

selected process parameters like arc voltage, wire feed rate, welding speed and nozzle

to plate distance on important weld bead dimensions and shape relationships.

Sharma et al. (2008) dealt with a shop floor applicable mathematical model for

deposition rate during twin wire submerged arc welding. The salient features of this

model were (i) instead of melting rate, as modelled during past investigations, it

quantified the deposition rate, which is the actual outcome of the process and always

remains smaller than the melting rate because of evaporation or spatters losses, and (ii)

it estimates electrode extension in order to predict the deposition rate which made

model more practical than the models constituted with the help of experimental

41
measurement of electrode extension. The model was more scientific than the

simplified models where contact tube to workpiece distance has been considered as the

electrode extension. A critical review of the relevant past investigation was given and a

mathematical model was developed for deposition rate during the twin wire welding

with both the polarities, i.e., direct current electrode positive and direct current

electrode negative. The model was calibrated using the results of experimental runs

and it was found to be very accurate with very high coefficient of regression and

admissible standard error. The developed model was further validated with extra

experimental runs. The practicality of the considered approach for prediction of

deposition rate can further be used in future research for other consumable arc welding

processes.

2.5.2 Artificial neural network (ANN) for simulation and prediction

Another approach used for modelling, simulation and prediction in submerged

arc welding is the Artificial Neural Network. ANN creates a mapping between set of

inputs and corresponding responses. ANN is very efficient to model and simulate

process behaviour, while the nature of response variation, with varying inputs, is very

much complicated. Depending on huge data set obtained from experiments

(combination of inputs and outputs), ANN itself establishes a correlation among inputs

and outputs, into its internal architecture, that consists of input, output, hidden layers

and connection between the layers (nodes/neuron). ANN then predicts output for a

given combination of factor settings (inputs). The perfection of network prediction or

42
network performance depends on the data set used to train the network, and selection

of its internal features viz. number of hidden layers, number of nodes in a layer,

leaning rate, training algorithm, performance goal and so on, (Widrow and Lehr, 1990;

Chester, 1990; and Bebis and Georgiopoulus, 1994). Adequate data set with optimal

network architecture can predict results with minimum error. An important step in

building the network is selection of input variables. In the field of welding, some

studies have been made with the neurons of the input layer to receive the input process

parameters, while the neurons of the output layer were used to send out the features of

quality characteristics of the weldment viz. weld bead and HAZ geometry, mechanical

properties of the weldment, metallurgical features of the weld metal as well as HAZ.

A feedforward neural network is used to model submerged are welding

processes in hardfacing (Tsai, 1996). The relationships between process parameters

(arc current, arc voltage, welding speed, electrode protrusion, and preheat temperature)

and welding performance (deposition rate, hardness, and dilution) were established,

based on the neural network. A simulated annealing optimization algorithm with a

performance index was then applied to the neural network for searching the optimal

process parameters. Experimental results had shown that welding performance can be

enhanced by using this new approach.

The non-linear relationship between the five geometric descriptors (height,

width, penetration, fused and deposited areas) of a bead and the welding parameters of

SAW had been modelled using neutral networks (Ping et al., 1997). A comparative

study between multi-output networks and single-output networks, each modelling one

43
geometric descriptor, had shown the advantages of single-output networks. The

structure of a conventional feed-forward multi-layer perceptron network with a single

output is modified to accommodate an offset layer which offsets the inputs. This

network, known as the self-adaptive offset network, has definite advantages over

conventional multi-layer perceptron networks. Altogether, 21 single-output neutral

networks had been trained for the four types of SAW welds investigated. These

networks have achieved good agreement with the training data and have satisfactory

generalisation.

Vitek et al. (2000) developed an artificial neural network model to predict

ferrite number in arc welds as a function of composition. The paper included valuable

information for the development of a neural network model, named FNN -1999, which

was capable of identifying the optimum network architecture and network parameters.

Results of the model were presented in Part 2 of the paper. It was shown that the

accuracy of the FNN-1999 model in predicting Ferrite Number was superior to the

accuracy of other available models at that time. The work done by Ridings et al. (2002)

described the application of neural network technique to predict the weld bead shape

for three wire, single pass per side, submerged arc pipeline seam welds. The effect of a

particular welding process parameter (input relevance) to the variation in the final weld

bead shape was also considered.

Kim et al. (2003) applied an intelligent system for the determination of welding

parameters for each pass and welding position, for pipeline welding, based on database

and finite element method, and on two back-propagation neural network models and a

44
corrective neural network model. Experiments using the predicted welding parameters

from the developed system proved the feasibility of interface standards and intelligent

control technology to increase productivity, improve quality and reduce the cost of

system integration. Lightfoot et al. (2005) developed a model of ANN and established

that the carbon content of the steel plate played a key role in the amount of distortion

produced by the welding process. The mechanism of the effect of carbon appeared to

be linked to its effect on grain size, transformation temperature, mechanical properties

and pearlite content, at least. It was established that an increase in carbon content was

beneficial in reducing thin plate distortion caused by welding.

2.5.3 Hybrid optimization techniques of SAW

The common approaches to tackle optimization problem in welding include

multiple regression analysis, response surface methodology, artificial neural network

modeling and Taguchi method. In most of the cases, the optimization has been

performed using single objective function. For a multi response process, while

applying the optimal setting of control factors, it can be observed that an

increase/improvement of one response may cause change in another response, beyond

the acceptable limit. Thus for solving multi criteria optimization problem, it is

convenient to convert all the objectives into an equivalent single objective function.

This equivalent objective function, which is the representative of all the quality

characteristics of the product, is to be optimized.

45
Tarng et al. (2000) used fuzzy logic in the Taguchi method to optimize the submerged

arc welding process with multiple performance characteristics. An orthogonal array,

the signal to noise ratio, multi response performance index, and analysis of variance

were employed to study the performance characteristics in the submerged arc welding

process. The process parameters, namely arc current, arc voltage, welding speed,

electrode protrusion, and preheat temperature were optimized with considerations of

the performance characteristics, including deposition rate and dilution. Experimental

results were provided to confirm the effectiveness of this approach.

Tarng et al. (2002) applied grey based Taguchi methods for optimization of

submerged arc welding process parameters in hardfacing. They considered multiple

weld qualities and determined optimal process parameters based on grey relational

grade from grey relational analysis proposed by Taguchi method. Apart from

desirability function and grey-based Taguchi approach, Genetic Algorithm and Fuzzy

Logic are also found to be useful techniques to solve optimization problem in the field

of welding. Genetic Algorithm was developed in 1980s to emulate the Survival of the

fittest principle introduced by Charles Darwin in his theory of evolution. From this

perspective and since optimization is analogous to fitness or the ability to survive real

world conditions, it makes good sense to apply Genetic Algorithm approach for system

improvement and process/product optimization, as mentioned by Al-Aomar (2002).

Apart from Genetic Algorithm, fuzzy logic also comes into the scenario of solving

optimization problems in material processing technology. Fuzzy logic allows degrees

of truthfulness that measures to what extent a given object is included in a fuzzy set.

46
Fuzzy sets correspond to linguistic variables used in a human language (Wang and

Jean 2006). Xue et al. (2005) reported the possibilities of the fuzzy regression method

in modelling the bead width in the robotic arc welding process. In their paper, they

developed a model for proper prediction of the process variables for obtaining the

optimal bead width.

Desirability function approach coupled with Taguchi method has been used by

some researchers to investigate conditions leading to process optimization. In this

context, application of other hybrid techniques deserves mention. These techniques

are: - (i) Taguchi method coupled with fuzzy logic, (ii) Genetic Algorithm and fuzzy

logic, (iii) Desirability Function Approach coupled with fuzzy logic, (iv) Genetic

Algorithm in combination with response surface methodology, and (v) Taguchi-

Genetic Algorithm (Tsai, 2004). Another approach for optimization is the Controlled

Random Search Algorithm developed by Price (1977). In this algorithm, the new trial

point in search space (parameter) is generated on the basis of a randomly chosen subset

of previously generated points. At each iteration, a simplex is formed from a sample

and a new trial point is generated as a reflection of one point in the centroid of the

other points in this simplex. If the worst point in the initially generated set is worse

than the new one, it is replaced by the later. After some repetition of this technique the

solution clusters around the global minima. The advantage of this algorithm is that, the

objective function needs not be differentiable. It only counts the objective function

value obtained through experiments; mathematical modelling of the process is not

required. Moreover, optimization can be achieved within a limited number of

47
experimental runs. This technique can also be applied for optimization in submerged

arc welding (Data et al., 2007). Besseris (2008) proposed a simple methodology in

solving multi response optimization problems by employing Taguchi methods and a

non-parametric statistical technique.

The multi-component mixture regression model developed by Sui et al. (2006)

was used to investigate physical properties of submerged arc welding flux. The effect

of complex interaction of seven components in agglomerated flux on softening

temperature was analyzed. The results indicated that the interaction of MgO-TiO2-

CaCO3-Al2O3 increased the softening temperature of flux, but the additions of CaF2

and ZrO2 can decrease the softening temperature.

Asiabanpour et al. (2004) and Ful-Chiang (2005) used optimization using

desirability function approach in their research work. This approach converts each of

the responses into their individual desirability value. Corresponding to each objective,

these individual desirability values are then accumulated to compute the

overall/composite desirability function, which is to be optimized next (Datta et al.,

2006).

Dr. Genichi Taguchi, a Japanese quality management consultant, developed a

method based on Orthogonal Array experiments, which provide much reduced

variance for the experiment with optimum setting of process control parameters. Thus,

the integration of design of experiments with parametric optimization of process is

achieved in the Taguchi method. Taguchi method uses a statistical measure of

performance called signal-to-noise ratio (S/N). The S/N ratio developed by Dr.

48
Taguchi is a performance measure to select control levels in order to set minimum

noise. S/N ratio takes both the mean and the variability into account. It is the ratio of

the mean (Signal) to the standard deviation (Noise). The ratio depends on the quality

characteristics of the product/ process to be optimized. However, in general, Taguchi

method cannot solve multi-objective optimization problem (Jeyapaul et al., 2005).

Therefore, Taguchi method coupled with grey relational analysis is the appropriate

option. Grey relational analysis is more or less similar to the desirability function

approach. Previous researchers, in various fields of production engineering, have

adopted Taguchi method coupled with grey relational analysis. In this method, a

multiple response process optimization problem is converted to a single response

optimization problem where overall grey relational grade serves as the single objective

function or response function to be optimized.

Application of the Taguchi method in combination with grey relational analysis

has been applied for solving multiple criteria (objective) optimization problem in

submerged arc welding (Datta et al., 2007). A grey relational grade evaluated with

grey relational analysis had been adopted to reveal an optimal parameter combination

in order to obtain acceptable features of weld quality characteristics in submerged arc

bead-on-plate welding. The idea of slag utilization, in subsequent runs, after mixing it

with fresh unmelted flux, had been introduced. The percentage of slag in the mixture

of fresh flux and fused flux (slag) had been denoted as slag-mix%. Apart from two

conventional process parameters-welding current and flux basicity index, the study

aimed at using varying percentages of slag-mix, treated as another process variable, to

49
show the extent of acceptability of using slag mix in conventional SAW processes,

without sacrificing any characteristic features of weld bead geometry and HAZ, within

the experimental domain. The quality characteristics associated with bead geometry

and HAZ were bead width, reinforcement, depth of penetration and HAZ width. Using

grey relational grade as performance index, parametric optimization was performed

yielding the desired features of bead geometry and HAZ. Predicted results had been

verified with confirmatory experiments, showing good agreement. This proves the

utility of the proposed method for quality improvement in SAW process and provides

the maximum (optimum) amount of slag-mix that can be consumed in the SAW

process without any negative effect on characteristic features of the quality of the

weldment in terms of bead geometry.

The work of Kumanan et al. (2007) detailed the application of Taguchi

technique and regression analysis to determine the optimal process parameters for

submerged arc welding. The planned experiments were conducted in the

semiautomatic submerged arc welding machine and the signals to noise ratios are

computed to determine the optimum parameters. The percentage contribution of each

factor is validated by analysis of variance technique. Multiple regression analysis was

conducted using statistical package for social science software and the mathematical

model was built to predict the bead geometry for any given welding conditions.

Karaoglu and Secgin (2008) focused on the sensitivity analysis of parameters and fine

tuning requirements of the parameters for optimum weld bead geometry. Experimental

part of this study was based on three level factorial designs of three process

50
parameters. In order to investigate the effects of process parameters on output

responses, which determine the weld bead geometry, a mathematical model was

constructed by using multiple curvilinear regression analysis. After carrying out a

sensitivity analysis using developed empirical equations, relative effects of input

parameters on output responses were obtained. Effects of all three design parameters

on the bead width and bead height showed that even small changes in these parameters

play an important role in the quality of welding operation. The results also revealed

that the penetration is almost non sensitive to the variations in voltage and speed.

2.6 Reclamation of SAW Slag

Slag produced due to fusion of the portion of the granular flux used during

submerged arc welding, is not reused and is discarded. The possibility of reuse of slag

and the results thereof is a new field of research. Rare or only very few investigations

have been carried out in this encouraging field of submerged arc welding (Datta et al.,

2008). Eagar (1980) investigated on reprocessing and reuse of the fused slag as means

of reducing costs of submerged arc welding of titanium. In general, it was found that

the cost of titanium submerged arc welding did not compare favourably with the costs

of alternate process such as gas metal arc welding of titanium. The author also outlined

the advantages and disadvantages of flux shielded welding of titanium. Beck and

Jackson (1996) concluded that according to code requirement, the properly processed

slag could be reliable and could be used as an alternative for new flux. They further

claimed a saving of 50% of the procured flux by recycled flux. Research related to slag

51
reconsumption in conventional SAW process has been carried out by Moi et al. (2001)

and Pal et al. (2001). Their study introduced the concept of using slag-mix% as a

process variable. The percentage of slag, in the mixture of fused flux and fresh flux,

has been denoted as slag-mix%. The main effects of using slag-mix and interactive

effects of process parameters (including slag-mix%) on features of bead geometry and

HAZ, in terms of bead height, depth of penetration, bead width and HAZ width have

been evaluated through analysis of variance method. But their work did not provide the

optimal factor combination to yield acceptable weldment and the maximum slag-mix%

that can be used during SAW process without adversely affecting bead geometry as

well as HAZ dimensions. They, however, carried out microstructure studies for

comparing the results of conventional SAW and SAW with a mixture of fresh flux and

fused slag.

Motivated by their concept, Sing et al. (2005) investigated to see the effect of

recycled slag on bead geometry in submerged arc welding. The slag was processed by

replenishing with suitable alloying elements/deoxidizers and then converted into new

flux called as recycled flux. Recycled flux was used to study the effect of welding

parameters on bead geometry and shape relationships. Mathematical models were

developed using a two level half factorial technique to predict weld bead geometry and

shape relationships. They concluded that SAW slag could be recycled and acceptable

bead geometry could be achieved with processed slag. As a continuation of the study,

Sing et al. (2006) recycled fused slag by replenishing it with suitable alloying elements

and deoxidizers and by agglomeration. They carried out experiments with this

52
modified recycled flux. The performance of the weldment was checked by chemical

analysis, radiography, mechanical and metallurgical tests. They observed favourable

results. Research related to slag recycling and subsequently reuse in submerged arc

welding is not very rich. Very few researchers have so far been attracted in this field,

but the scope of work is vast and bright. It is significantly important as well. If it can

be experimentally proved that use of slag (mixed with fresh flux) can be consumed

instead of fresh flux only, without producing any harmful adverse effects on various

quality aspects of the weldment, then this technique can be adopted on commercial

basis in practical field in the fitted applications. Additionally, it is always important

and useful to reduce waste and to move towards Waste to Wealth, or finally

approaching towards Zero Waste Concept.

No work so far has been done to utilize the waste flux dust to develop as

agglomerated flux. The present work attempts to develop agglomerated flux from the

flux dust without affecting the quality the weldment.

53