Study of single side single pass submerged

arc welding using reusable backing strip

N. R. Mandal* and Rajiv Maiti
An attempt has been made to establish a submerged arc welding procedure that will enable the
production of welded butt joints in thin steel plate, having proper fusion and reinforcement
geometry in a single welding run. In this work, the combined effect of the basic welding
parameters, i.e. voltage, current, and speed, along with the effect of thickness, on weld
penetration have been studied. Also an attempt has been made to develop a flux filled reusable
backing strip. In the present investigation submerged arc welding flux in its powdered form has
been used without application of any external pressure and also without any additives.
Experiments have been carried out extensively on 6 mm and 8 mm thick C–Mn steel plate
specimens. This form of backing strip has shown great promise towards achieving single side
single run welding. The welds achieved in this single side single pass welding process are not yet
100% satisfactory, however, the results indicate the feasibility of the process to achieve quality
welds to meet relevant quality standards.
Keywords: Single side welding, Submerged arc welding, Joint geometry, Backing strip, Heat input

Introduction to establish that submerged arc welding can be gainfully

utilised to weld steel plates as thin as 6 and 8 mm. In the
A major portion of welding activity in a shipyard prevailing shipbuilding practice, the belief is that SAW is
comprises the butt welding of large flat plates for not an effective technique for welding such steel plates.
fabrication of flat panels. The conventional method of The ultimate goal of a welding engineer is to develop a
welding such panels consists of several operations such suitable welding procedure for a given welding technique
as welding from the top, turning the panel over, gouging such that full penetration is achieved, along with
of root, followed by final welding. Usually these panels adequate top and bottom reinforcement. In order to
have very large overall dimensions making it difficult for achieve this, all the relevant process variables need to be
such operations. Where turning over is physically not suitably selected. Dryaton,1 Doherty et al.2 and Yang
possible, overhead welding becomes the only alternative. et al.3 studied the influence of process parameters and
This results in an uneven weld bead and a slower rate of the relationship between process variables and bead
production. The uneven bead shape calls for additional geometry in the case of submerged arc welding. Joarder,
grinding operations, resulting in further man hour Saha and Ghose4 studied the effect on microstructure of
requirements. The genesis of this investigation is based the heat affected zone and submerged arc weld metal in
on the requirement to improve productivity by cutting the case of plain carbon steel. Shin et al.5 reported an
down total welding time. Implementation of single side analytical method for predicting through thickness
welding will result in improvement in productivity and distribution of residual stress in a thick plate undergoing
this will be further enhanced if the desired deposition multipass welding. Miyazaki et al.6 reported one side
can be achieved by single run. However, this has to be submerged arc welding using flux copper backing.
achieved while maintaining the desired level of quality. Malin7 discussed the effect of joint geometry and
Since the welding process leads to very complex physical welding variables in root weld formation using modified
and metallurgical changes in the workpiece, it is practically refractory flux one sided two electrode singlepass/
impossible to establish an exact mathematical model of the multipass welding. His work was based on using
same. An attempt therefore has been made in this study to thermosetting backing flux. DuPont et al.8 conducted a
establish a submerged arc welding procedure that will study on the arc and melting efficiency of various
enable the production of welded butt joints having proper welding processes and concluded that with higher arc
fusion and reinforcement geometry in a single welding run efficiency higher travel speeds are possible. Bonifaz9
through experimental investigation. Also the objective was developed a simple 2D finite element model to calculate
not only the transient thermal histories but also the sizes
of fusion and heat affected zones in single-pass arc welds
Department of Ocean Engineering and Naval Architecture, IIT Kharagpur, considering an arc efficiency of 0?95. Arc efficiency of
India 0?95 was reported by Goldak et al.10 for submerged arc
*Corresponding author, email nrm@naval.iitkgp.ernet.in welding. Single pass submerged arc welding with three

ß 2005 Institute of Materials, Minerals and Mining

Published by Maney on behalf of the Institute
Received 15 June 2004; accepted 4 September 2004
DOI 10.1179/174329305X40633 Science and Technology of Welding and Joining 2005 VOL 10 NO 3 319
 Mandal and Rajiv Maiti Single side single pass submerged arc welding using reusable backing strip

electrodes in tandem was studied by Kasuya et al.11 The thus these effects are well known. In this study, it has
effect of different preheat temperatures on cooling time been attempted to study the combined effect of the basic
of t8/5 and t100 was reported. Arc efficiency of 0?8 was welding parameters, i.e. voltage, current, and speed
considered by Jang et al.12 for FCAW process with Ar– along with the effect of thickness, on weld penetration.
CO2 shielding gas. Heat input was taken as (garc E I)/ To capture the combined effect of these parameters, the
(bL), where b is the width of the weld bead and L is the quantity ‘heat input per unit volume’ has been used in
length in the weld direction of the heat input region, i.e. this investigation.
unit length for 2D analysis. The effect of plate thickness For a given steel composition, its microstructure may
in multipass welding has been reported.13 Reducing the be affected if it is subjected to a thermal cycle. In turn,
plate thickness from 50 to 20 mm led to a large increase depending on the microstructure, the hardness also will
in cooling times due to reduction in heat dissipation be affected. One of the parameters that affects the
capacity. The cooling times t8/5 were drastically affected microstructure is cooling rate. Now it is well known that
by changing heat input through variation of the basic thicker the plate, the higher the cooling rate for a given
welding parameters. The influence of plate thickness on heat input.13 Hence, for a given heat input a thicker
the size of weld pool and HAZ have been reported.14 It plate may exhibit higher hardness compared to that of a
has been observed that the size of weld pool and HAZ in thinner plate subjected to same heat input. Thereby
4 mm specimens are smaller to those of a 3 mm plate thickness does play a role on the resulting material
specimen. Also, a faster thermal equilibrium is achieved hardness.
with the 4 mm specimen. The thicker plate suffers To observe the effect on hardness and fusion depth,
greater heat loss by conduction. Gunaraj et al.15 showed due to certain heat input, the effect of thickness should
that penetration reduces as welding voltage increases, be taken into account. Therefore instead of considering
but bead width and dilution increase considerably with
‘heat input per unit length’, ‘heat input per unit volume’
increasing voltage.
becomes more relevant for consideration. To arrive at
Sullivan16 highlighted the effectiveness of single side
this parameter, the quantity ‘heat input per unit length’
welding in making ship production more competitive.
has been further divided by square of plate thick-
In all these previous studies, mostly multielectrode
ness.17,18 The square term is based on the assumption
multipass submerged arc welding of thick plates has
that the heat is distributed over a volume equal to the
been studied. In this work an attempt has been made to
product of weld metal length traversed by the welding
establish a submerged arc welding procedure that will
torch per unit time, weld metal thickness equal to plate
enable production of welded butt joints using thin steel
plates, having proper fusion and reinforcement geome- thickness, and weld metal width perpendicular to the
try in a single welding run. The combined effect of the welding direction equal to that of plate thickness. To
basic welding parameters, i.e. voltage, current, and arrive at this parameter, a better option would be
speed along with the effect of thickness, on weld dividing the quantity ‘heat input per length’ by
penetration have been studied. Also an attempt has measured cross-sectional area of fusion zone.
been made to develop a flux filled reusable backing strip. However, this approach is dependent on the individual
In the present investigation submerged arc welding flux case. For predictive purpose this approach cannot be
in its powdered form has been used without application used because unless welding is done, the area of fusion
of any external pressure and also without any additives. zone is not known. Hence this approach does not lend
The study also aims at investigating the effect of flux on itself to generalization. In view of this the above
the formation of the bottom reinforcement. Experiments assumption has been made.
have been carried out extensively on C–Mn steel plate of The heat input and heat input per unit volume are
thickness 6 mm and 8 mm. This form of backing strip given as
indicated great promise towards achieving single side gV I
single run welding. The welds achieved in this single side q~ J mm{1
ws
single pass welding process are not yet 100% satisfac-
tory, however, the results indicate the feasibility of the gV I
process to achieve quality welds to meet relevant quality heat input per unit volume~ J mm{3
ws|t2
standards.
where q 5 rate of heat input (J mm21), g 5 arc
efficiency, taken as 0?9 for submerged arc welding, V 5
Process variables arc voltage (V), I 5 welding current (A), ws 5 welding
The process variables and their range of variation speed (mm s21), t 5 plate thickness (mm).
considered in this investigation are given in Table 1.
Various combinations of these process variables were
Table 1 Process variables used in single side single
used to carry out welding. To obtain adequate weld pass welding of C–Mn steel samples
penetration with proper top and bottom reinforcement,
an optimum combination of the variables is selected. Welding current (A) 280–410
Extensive experiments were carried out to evaluate the Welding arc voltage (V) 24–32
Welding speed (mm s–1) 5?8–11?3
combined effect of these process variables on these
Electrode diameter (mm) 3?15 (fixed)
aspects of the welded joint. Weld penetration is indeed Length of stickout (mm) 20 (fixed)
affected not only by voltage and current only, it is as Root gap (mm) 0–4
much affected by welding speed, electrode orientation, Root face (mm) 0
electrode polarity, plate thickness, etc. However, varying Bevel angle (degrees) 10–50
one parameter at a time and keeping the rest constant, Electrode angle (degrees) z10 to 210
Electrode polarity positive (fixed)
the effect of individual parameters can be captured, and

Science and Technology of Welding and Joining 2005 VOL 10 NO 3 320

 Mandal and Rajiv Maiti Single side single pass submerged arc welding using reusable backing strip

1 Joint geometry with flux filled aluminium backing bar

Experimental procedure
Submerged arc welding was carried out on several test
samples of size 300 mm6240 mm using various combi-
nations of process variables as given in Table 1. A
constant potential 600 A DC power supply was used.
The voltage was measured between the welding torch
2 Single side submerged arc welding with flux filled alu-
and workpiece. The measured voltage thus represents
minium backing trough
the sum of voltage drop across the electrode and arc.
Current was measured by a calibrated shunt in series
with the current carrying cable. The voltage does not steels. The chemical compositions of base metal,
include the loss in the cables. The submerged arc welding electrode wire and flux were obtained using ZAF
machines used in shipyards do not generally have a quantitative method of scanning electron microscopy
mechanism for measuring electrode feed rate and also at having system resolution of 68 eV, 68 eV, and 67 eV
the same time, the feed rate control is directly coupled respectively. The values are given in Table 2. The grain
with the current control. The implication is that size of the flux varied between 0?2 to 1?6 mm.
changing the feed rate automatically changes the More than 70 test samples were welded. In each case
current, or in other words, for a given electrode diameter the welding voltage, welding current, welding speed,
and given current the electrode feed rate is automatically length of stickout, electrode angle and electrode polarity
adjusted. The machine used in the investigation did not were observed. Electrode angle was varied from 10u
have the facility of direct feed rate measurement. forward to 10u backward. Electrode polarity was
Commercially available C–Mn steel of 6 and 8 mm maintained positive and length of stickout was kept
thickness was used to prepare testpieces for welding. 20 mm in all cases. The welding conditions and joint
Run on run off tabs were used to allow for welding to be geometries are detailed in Table 1.
carried out over the entire length of the plate. The root From each welded sample, transverse specimens were
gap between the plates were maintained with the help of cut and the cross-section of each specimen was polished
these tabs at the ends. No tack welding was done in and etched with 2% Nital solution to reveal the weld
between. An aluminium backing bar having deep groove profile. The same samples were used for testing hardness
was used, as shown in Fig. 1. In the present investigation at the three zones, parent metal, HAZ, and deposited
SAW flux in its powdered form was used without metal.
application of any external pressure and also without
any additives. The aluminium metallic trough as shown
in Fig. 1 was filled with flux prior to welding and the Results
trough was held beneath the plates, as shown in Fig. 2. In the following subsections, the effects of various
This flux in the trough supported the molten metal. The process variables on fusion depth and resulting hardness
flux backing being sufficient the backing bars could be of deposited metal and HAZ have been analysed. The
reused without any damage occurring to them during fusion depth is the depth to which parent metal melting
welding. The consumables used in the experiments were has taken place along the plate thickness. The maximum
commercial 3?15 mm diameter mild steel electrode and a possible value of fusion depth is naturally equal to the
semibasic SAW flux suitable for joining such C–Mn thickness of the plate being welded. A nital etched

Table 2 Chemical composition of steel plate, filler metal and the flux used in study

Analysis of plate and filler wire Analysis of flux

Element % in plate % in filler wire Compound %

C 0?19 0?18 SiO2 31?23

Si 0?35 0?34 TiO2 1?02
Mn 1?01 0?44 MgO 17?43
Cu 0?04 0?06 CaO 20?65
Al 0?05 0?05 InAs 1?78
P less than 0?027 less than 0?027 Fe2O3 3?26
S less than 0?027 less than 0?027 Al2O3 21?44
MnO 3?18

Science and Technology of Welding and Joining 2005 VOL 10 NO 3 321

 Mandal and Rajiv Maiti Single side single pass submerged arc welding using reusable backing strip

6 Welded sample macrosection of 6 mm thick plate with

coarse mesh flux backing
3 Macrosection of 6 mm welded sample
and forms the slag, which in a molten state provides fluid
support to the molten metal at the root of the joint. For
sufficient slag formation below the root of the weld,
somewhat higher current, than otherwise required,
becomes necessary. Thus it helps with proper root fusion
as well as formation of adequate bottom reinforcement.
Conventional ceramic backing strips are for one time use,
whereas the aluminium backing strip used in this
4 Welded sample macrosection of 6 mm thick plate with
investigation can be used repeatedly, virtually an infinite
fine mesh flux backing
number of times. Because here the molten metal is
actually supported by the flux, the aluminium bar is not
damaged by the heat of the molten metal.
Flux backing has been used by some investigators in
different forms; some have used powdered flux pressed
against the plates with the help of pressure hose, others
have tried flux mixed with phenolic thermosetting
compound, which solidifies under the influence of heat.
5 Welded sample macrosection of 8 mm thick plate with In the present investigation SAW flux in its powdered
fine mesh flux backing form has been used without application of any external
pressure and also without any additives. The aluminium
macrosection of the fusion zone and HAZ of a 6 mm metallic trough shown in Fig. 1 is filled with flux prior to
welded sample is shown in Fig. 3. Some misalignment welding and the trough is held beneath the plates as
was observed in certain test samples and in fact it shows shown in Fig. 2. This flux in the trough supports the
that the flux backing used in the investigation can molten metal. It has been observed that the finer the
accommodate such small misalignments. These levels of mesh size of the flux, the better is the bottom
misalignment do occur in welding ship panels, which are reinforcement, as can be seen in Figs 4 and 5.
often 5 to 10 m in length. Conclusions have been drawn With a coarse mesh flux size, in spite of adequate root
in each subsection considering that other process fusion, the deposited metal at the root takes a concave
variables remain unchanged. shape, as can be seen in Figs 6 and 7. However, this
problem could be overcome by applying a higher
Effect of electrode angle current. Thus this form of flux backing holds great
The electrode angle was varied from 10u backward to promise of achieving single side welding without the
10u forward. However, it has been observed that for requirement of any additional fixture or additives.
thinner plates, i.e. 6 mm and 8 mm, the electrode angle
has no significant effect on the fusion depth and Effect of heat input
deposition pattern of weld metal. The combined effect of welding current, voltage, and
welding speed on fusion depth (weld penetration) for
Effect of root gap and bevel angle two different plate thicknesses is shown in Figs 8 and 9.
The root gap was varied from 0 to 4 mm. For welding The other process variables, as described above, are kept
with a 3?15 mm diameter electrode, it was observed that unaltered to study the effect of heat input for 6 and
a root gap higher than 3 mm resulted in an uneven 8 mm thick plates. Although there is some scatter in the
fusion at the root. A root gap less than 3 mm yielded a data, a definite trend can be observed.
lack of fusion and required deposition at the root. The For 6 mm plate one can observe in Fig. 8, that over a
best result was obtained with a root gap of 3 mm. range of heat input 700 to 1200 J mm21, the increase in
Similarly, with less bevel angle, the top reinforcement fusion depth is only about 0?5 mm. A similar trend is
tended to be excessive with lack of fusion at the root. A observed in the case of 8 mm thick plate, as shown in
higher bevel angle not only resulted in higher angular Fig. 9, where over a range of 1000 to 1700 J mm21, the
distortion but also led to a lack of deposition and a
concave top reinforcement. The optimum bevel angle
was found to be 50u.

Effect of flux filled backing strip

Aluminium backing strip filled with flux to provide
support to the molten metal proved to be very effective
compared to the conventional ceramic backing strip. 7 Welded sample macrosection of 8 mm thick plate with
Under the action of the arc, the flux at the root burns coarse mesh flux backing

Science and Technology of Welding and Joining 2005 VOL 10 NO 3 322

 Mandal and Rajiv Maiti Single side single pass submerged arc welding using reusable backing strip

8 Variation of fusion depth with heat input for 6 mm

plate
11 Variation of base metal hardness

speed, the depth of fusion showed a decreasing trend.

From the variation it can be concluded that, to achieve
full penetration in single run submerged arc welding
using a 3?15 mm electrode using 6 mm thick plate, the
welding speed should not exceed about 9?5 mm s21. For
8 mm plate, welding speed should be around 5 mm s–1
or less over the respective heat input range.

Effect on hardness
The samples used for metallographic study were further
used for hardness testing using standard Rockwell B scale
RB with a steel ball indenter with standard load of 100 kg.
The variations of hardness at different zones with heat
9 Variation of fusion depth with heat input for 8 mm input per unit volume are shown in Figs 11, 12, and 13.
plate Ideally the base metal hardness should not change with
change in heat input, as the base metal is not affected by
heat. However the small variation observed in Fig. 11,
can be attributed to experimental scatter.
Similarly, in Fig. 12, as was expected, one can observe
a very small variation of hardness of the fusion zone.
This is because, irrespective of the heat input, the fusion
zone undergoes near identical thermal cycles. In the case
of the heat affected zone (HAZ), one can observe a
wider variation of hardness with heat input, as shown in
Fig. 13. With increase in heat input, the hardness varied
from about 70 to 88.
From the above figures, one can observe that the base
metal hardness was less than 70, whereas in the fusion
zone, with increasing heat input the hardness increased

10 Variation of fusion depth with welding speed

increase in fusion depth is almost negligible. Therefore,

with the above referred to process variables remaining
unchanged, to weld 6 mm thick mild steel plates the rate
of heat input can be kept within 700 to 800 J mm21,
whereas in the case of 8 mm plate the heat input could
be 1100 to 1200 J mm21 to achieve full penetration in
single run submerged arc welding.

Effect of welding speed

For different combinations of welding voltage and
current, the effect of increasing welding speed on the
fusion depth of both 6 mm and 8 mm mild steel plate 12 Variation of fusion zone hardness with heat input per
can be seen in Fig. 10. As was expected, with increasing unit volume

Science and Technology of Welding and Joining 2005 VOL 10 NO 3 323

 Mandal and Rajiv Maiti Single side single pass submerged arc welding using reusable backing strip

5. Aluminium backing strip with deep semicircular

groove filled with the same flux as used for SAW was
found to be very effective without application of any
external pressure, for providing support to the molten
metal and to form the required bottom reinforcement.
6. The dimensions of the aluminium backing bars are
so chosen so that they are never damaged by the arc
action and thereby can be virtually infinitely reused.
7. The finer the particle size of the flux used in the
backing strip, the better is the formation of bottom
reinforcement.
8. The effect of heat input on hardness of the fusion
zone is negligible. In the heat affected zone (HAZ), the
hardness varied from about 70 to 88 with increasing heat
input.
9. The hardness of the HAZ remained almost the
13 Variation of HAZ hardness with heat input per unit
same as that of the fusion zone for heat input rates less
volume
than 22 J mm–3. Hence the possibility of HAZ cracking
is less for heat input rates less than 22 J mm–3.
to about 75, and 85 in the HAZ. However the hardness
of the HAZ remained almost the same as that of the
fusion zone for heat input rates less than 22 J mm–3. References
Hence the possibility of HAZ cracking will be less for a 1. A. Drayton: ‘An examination of the influence of process
heat input rates below 22 J mm–3. parameters on submerged arc welding’, The Welding Institute
Report No. PE/4/72, 1972.
Conclusions 2. J. Doherty and J. C. McGlone: ‘Relationship between process
variables and bead geometry’, The Welding Institute Report No.
The following conclusions can be drawn from the 52/1977/PE, 1977.
extensive experimental study conducted on single side 3. L. J. Yang, R. S. Chandel and M. J. Bibby: Weld. J., 1993, 72, (1),
single pass submerged arc welding of C-Mn steel plates 11s–18s.
4. A. Joarder, S. C. Saha and A. K. Ghose: Weld. J., 1991, 70, (6),
of different thickness. 41s–146s.
1. The joint geometry plays an important role in the 5. Y. Shim, Z. Feng, S. Lee, D. Kim, J. Jaeger, J. C. Papritan and
fusion pattern. It has been found that 50u bevel angle, a C. L. Tsai: Weld. J., 1992, 71, (9), 305s-312s.
root opening of 3 mm, and zero root face provide the 6. T. Miyazaki et al : Htachi Zosen Tech. Rev., 1993, 54, (1), 71s–85s.
best through thickness fusion in the case of single side 7. V. Malin: Weld. J., 2001, 80, (9), 217s to 237s.
8. J. N. DuPont and A. R. Marder: Weld. J., 1995, 74, (12),
welding of 6 mm and 8 mm plates.
406s–416s.
2. For full penetration single run submerged arc 9. E. A. Bonifaz: Weld. J., 2000, 79, (5), 121s–125s.
welding of 6 mm mild steel plate having 50u bevel angle, 10. J. Goldak, A. Chakravarty and M. Bilby: Metall. Trans. B, 1984,
root opening of 3 mm, and zero root face, the rate of 15, (6), 299s–305s.
heat input can be kept within 700 to 800 J mm21, 11. T. Kasuya and N. Yurioka: Weld. J., 1993, 72, (3), 107s–115s.
whereas in the case of 8 mm plate the requirement could 12. G. B. Jang, H. K. Kim and S. S. Kang: Weld. J., 2001, 80, (3),
80s–89s.
be as high as 1100 to 1200 J mm21. 13. N. U. Ahmed and B. L. Jarvis: Weld. J., 1996, 75, (1), 15s–24s.
3. To achieve full penetration, for 6 mm mild steel 14. K. S. Boo and H. S. Cho: J. Eng. Manuf., 1990, 204, (5), 175s–183s.
plate having 50u bevel angle, root opening of 3 mm, and 15. V. Gunaraj and N. Murugan: Weld. J., 2000, 79, (10), 286s–293s.
zero root face, the welding speed should not exceed 16. M. J. Sullivan: Proc. ‘How to competitively weld the 21st century
about 9?5 mm s21, whereas for 8 mm plate, the speed ship’, Norfolk, Virginia, 2000, 31s–51s.
17. Y. Luo, M. Ishiyama and H. Murakawa: Trans. Join. Weld. Res.
should be around 5 mm s21 or less over the respective
Inst. Osaka Univ., 1999, 28, (2), 57s–65s.
heat input range. 18. N. Toshiharu and A. Kazuhiro: ’Development of simulator for
4. For 6 mm and 8 mm plate, the electrode angle has plate bending by line heating’, University of Tokyo Publication,
no significant effect on the fusion depth. Paper No.12, 143–148.

Science and Technology of Welding and Joining 2005 VOL 10 NO 3 324