Welding bimetal pipes in duplex stainless steel

Abstract

Butting bimetal pipes in duplex stainless steel are quite recent and present a set of interesting characteristics especially
for oil and gas transportation, namely weight to corrosion resistance ratio. Gas tungsten arc (GTA) welding is used to join
these pipes, but several problems are identified as lack of penetration and cracking resulting not only from the material
itself, but also from the difficulty to weld in orbital position. In the present work, autogenous GTA pipe welding and pulse
Rapid Arc gas metal arc welding of the pipes were studied. Current intensity, welding speed, electrode diameter, shielding
gas and orbital positions were defined as variables. It was shown that under appropriate conditions, it is possible to obtain
sound welds with proper geometry and defect free. The major limitation to penetration is the outwards flow pattern in the
molten pool driven by the Marangoni effect, as a result of low sulphur content. Sulphur is an active surface element which
reverses the surface tension coefficient to a positive effect. For penetrations of about 2mm, a combination of current intensity
of 170A and a welding speed of 200mm/min with an electrode angle of 30° under a shielding gas protection of He–25%Ar was
defined for narrow groove welding of a J-bevelled pipe in flat position. No defects were detected in the inner layer in duplex
stainless steel.

KeywordsBimetal pipes–Duplex stainless steels–Welding–GMAW–GTAW–Weld bead shape

Full text

ORIGINAL ARTICLE
Welding bimetal pipes in duplex stainless steel
A. M. Torbati & R. M. Miranda & L. Quintino &
S. Williams
Received: 15 March 2010 /Accepted: 5 August 2010
#
Springer-Verlag London Limited 2010
Abstract Butting bimetal pipes in duplex stainless steel are
quite recent and present a set of interesting characteristics
especially for oil and gas transportation, namely weight to
corrosion resistance ratio. Gas tungsten arc (GTA) welding
is used to join these pipes, but several problems are
identified as lack of penetration and cracking resulting not
only from the material itself, but also from the difficulty to
weld in orbital position. In the present work, autogenous
GTA pipe welding and pulse Rapid Arc gas metal arc
welding of the pipes were studied. Current intensity,
welding speed, electrode diameter, shielding gas and orbital
positions were defined as variables. It was shown that under
appropriate conditions, it is possible to obtain sound welds
with proper geometry and defect free. The major limitation
to penetration is the outwards flow pattern in the molten
pool driven by the Marangoni effect, as a result of low
sulphur content. Sulphur is an active surface element which
reverses the surface tension coefficient to a positive effect.
For penetrations of about 2 mm, a combination of c urrent
intensity of 170 A and a welding speed of 200 mm/min
with an electrode angle of 30° under a shielding gas
protection of He–25%Ar was defined for narrow groove
welding of a J-bevelled pipe in flat position. No defects
were detected in the inner layer in duplex stainless steel.
Keywords Bimetal pipes
.
Duplex stainless steels
.
Welding
.
GMAW
.
GTAW
.
Weld bead shape
1 Introduction
1.1 Butting bimetal pipes
Butting bimetal pipes, also known as BuBi pipes, are the
subsequent generation of bimetal pipes after clad pipes.
They are comprised of an outer thick layer of high strength
carbon steel and a thin liner layer of corrosion-resistant
alloy such as austenitic stainless steel, duplex or super
duplex stainless steel, or nickel-based alloy which is in
close contact with the outer layer [1]. In the manufacturing
process, this thin layer is expanded to close contact with the
outer layer by being aligned inside the thicker pipe and then
expanded to the outer layer by an internal hydraulic
expansion force [2].
The benefits of choosing BuBi pipes over conventional
metallurgically bonded clad pipes include cost efficiency in
the manufacturing process and better properties of the
interface between outer and liner layer s due to the absence
of fusion bonding and thermal effects [3]. Additionally, it
has the advantage of having various options in selecting
combinations for the outer layer and liner layer; hence, it
results in increased cost-effectiveness by choosing more
corrosion-resistant materials with better mechanical proper-
ties less prone to failure [2]. The thick layer of carbon steel
A. M. Torbati
:
S. Williams
Cranfield University,
Cranfield MK43 0AL, UK
R. M. Miranda (*)
UNIDEMI, Departamento de Engenharia Mecânica e Industrial,
Faculdade de Ciências e Tecnologia, FCT,
Universidade Nova de Lisboa,
2829-516 Caparica, Portugal
e-mail: rmiranda@fct.unl.pt
L. Quintino
IST-UTL Instituto Superior Técnico,
Lisbon, Portugal
L. Quintino
IDMEC, Institute of Mechanical Engineering, TULISBON,
Lisbon, Portugal
Int J Adv Manuf Technol
DOI 10.1007/s00170-010-2889-7

provides adequate strength for the pipelines, while the thin
layer, which is normally 1.5- to 3-mm thick, provides
internal corrosion resistance against corrosive media such
as H
2
S and CO
2
. Replacing metallurgically bonded clad
pipes with BuBi pipes has significantly reduced costs by
about 25% to 40% [3]. Moreover, greater cost-effectiveness
of the process is achieved as a result of the semi-automated
manufacturing process of BuBi pipes. Figure 1 demon-
strates the expansion process and Fig. 2 a flow chart of the
whole manufacturing process.
The main application of BuBi pipes is in the oil and gas
industry. Exxon Mobil has reported installing more than
420 km of bimetal pipes as well as BP, Statoil, etc., since
1994 [2, 3 ].
1.2 Welding BuBi pipes
When butt welding, root pass is the most critical, since it
must fully penetrate and prevent contamination of the
corrosion-resistant liner, so moisture and dirt contaminations
must be avoided. First pass into the liner pipe results in a fine
back surface inside the pipe, and then other passes are
conducted to fill the groove. Figure 3 shows the joint
preparation.
Duplex stainless steel (DSS) consists of bcc ferrite and fcc
austenite with almost equal proportions and has the benefits
over austenitic stainless steels of higher strength and better
corrosion resistance, namely to pitting, crevice and chloride
stress corrosion [4–6]. Duplex stainless steels combine the
good resistance of ferritic stainless steels with the good
ductility and toughness of austenitic stainless steels [7].
Fig. 1 Expansion process for BuBi pipe production [2 ]
Fig. 2 BuBi pipe manufacturing process chart [2]
Fig. 3 Preparation for seal welding of BuBi pipe [2]
Int J Adv Manuf Technol

However, service temperature is restricted to around 280°C
as they are prone to 475°C embrittlement. They are also
susceptible to sigma phase formation above 538°C [6–8].
The composition of DSS consists of 20–30 wt.% Cr, 5–
10 wt.% Ni, 0.03–0.04 wt.% C and other elements such as
Mo, N, W and Cu. Each of these elements influences the
materials’ property, and nitrogen is important as an austenite
stabiliser. Sulphur is present in low contents in order to
prevent hot cracking in the fusion zone [8]. During cooling,
DSS solidify with a fully ferritic structure followed by
austenite nucleation, primarily in grain boundaries of ferrite
and then within the grains at lower temperatures. Two zones
undergo different transformations: the high temperature zone
in the vicinity of the fusion zone that becomes fully ferritic
and the low temperature heat-affected zone close to the base
metal that remains unchanged [8]. Since the cooling rate is
high during welding, the austenite transformation tempera-
ture decreases. Weld metal is fully ferritic in autogenous
welding, and austenite transformation begins at temperatures
below ferritic solvus temperature [9, 10]. In non-autogenous
welding processes, the use of filler materials with austenite
forming or stabiliser elements, such as nickel, increases the
austenite nucleation temperature [6, 8]. Also, increasing the
nitrogen content of the shielding gas prevents fully ferritic
structures in the HAZ.
RapidArc™ is a relatively new variant of gas metal
arc welding (GMAW) for low penetration depths, with
good contr o l over t he a r c st abi li ty at low arc volt ag es ,
hence low heat input. Rapid arc is a refined pulse
process that uses short arc length with controlled short
circuit cycles. This resul ts in better bead profile at higher
speeds, eliminating humpy welds characteristic of higher
travel speeds, produces less spatter, reduces the risk of
underc ut s [9] and improves the welding speed up to
2.5 m/min with improved cost savings. The rapid arc
waveform can be divided in several segments: a pulse,
which is a sudden increase in the current that increases
the arc energy and squeezes a molten droplet extending
from the tip of the electrode; a puddle rise, a ramp down
of current that relieves the plasma force depressing the
puddle allowing it to rise up towards the droplet; the arc
collapse, originating the droplet to contact with the weld
puddle; a puddle repulsion immediately following a short
breaking into the arc, a plasma boost pushes the puddle
away ensuring a separation of the wire tip and the puddle,
resulting in a stable cycling rhythm of the process.
2 Experimen tal procedures
2.1 Materials
The material used in this study is an ASTM A240 S32506
duplex stainless steel pipe with 254-mm internal diameter
and 14-mm wall thickness. Tables 1 and 2 present the
compositions of the base material and the wire filler
material applied in GMAW, respectively.
2.2 Welding
2.2.1 Equipment
The sele cted power source for TIG welding was a
MIGATRONIC A/S TIG COMMANDER which is a three
phase machine based on inverter technology. All functions
can be set on a one control knob. For TIG welding it has
functions such as
– Variable slope down
– Fixed gas pre-flow time
– Variable gas post flow time
– Option of two times operation or four times operation
– HF TIG and LIFTIG
– Digital display
– Water cooled tubes
The power source used for MIG welding is a Lincoln
455/STT which has the capability of suppl ying power for
Table 1 Chemical analysis of S32506 base material (weight percent)
C Si MnP S Cr MoNi W
0.015 0.32 0.44 0.025 <0.005 25.0 3.19 6.34 2.00
Table 2 Chemical analysis of 2205 filler material for GMAW (weight
percent)
Type C Si Mn N Cr Ni Mo
2205 0.02 0.5 1.6 0.17 23.0 8.5 3.1
Fig. 4 Experimental setup for GTA welding of the pipe
Int J Adv Manuf Technol

MIG, Pulsed MIG, STT and pulsed RapidArc. It also has a
feeding system to provide welding wire. This power source
is used on the synergic mode which means that the voltage
and current are altered automatically to compromise with
the defined wire feed speed and travel speed and sustain the
stability of the arc.
Data of frequency oscillation was recorded by an
oscilloscope Yokogawa DL 750 Scopcorder. It recorded
the electrical data from the power supply and monitored the
current, voltage and wire feed speed during the welding
trials. With these values, it was possible to calculate the
heat input for the different welding conditions studied.
The manipulator was a rotator that holds the pipe with
three jaws in any angle from 0° to 90° to the horizontal axis
while revolving it around the pipe’s axis of rotation both
clockwise and anti-clockwise. It had 90° freedom of
rotation around the x-axis to hold the pipe in different
spatial positions and inclinations, and on the jaws, it had
360° freedom of rotat ion around the pipe’s central axis of
rotation.
A gas flow mixer PRC-2000/3000 programmable ratio
control system automatically contr olled the flow of mul tiple
gases. It can mix two or three gas bo ttles in order to get a
different proportion of gases for shielding.
After welding, samples were cut and prepared for
metallographic analysis after polishing and etching with
Kalling reagent. Macrostructure, as well as, defects were
observed using an optical microscope NIKON PRYSM
connected to a computer through a KY-F55B JVC video
camera for data acquisition.
2.2.2 Weldi ng procedure
Before welding, the inner surface was brushed to eliminate
dirt and oxide contamination of the weld zone. The root
pass was performed in all positions with a continuously
rotating torch. To simulate this condition, a manipulator
rotating both clockwise and anti-clockwi se was used. The
torch was fixed for any position defined, and the pipe
rotated to simulate the upwards and downwards torch
travel. Figure 4 show s the experimental setup.
For selected variables as current density, welding speed,
shielding gas and electrode angle, a matrix was made using
dedicated software for design of experiments, and welds
were conducted according to this matrix. For welding on a
narrow groove joint from both sides, one MIG root pass
was done from outside with 2205 duplex steel wire filler
followed by a re-melting by TIG weld run, from the inside.
Back plates wer e used to sustain the molten metal when
joined from outside by GMAW. The ends of the pipes were
prepared and bevelled, and butts were tack welded to hold
the joint for welding. Figure 5 shows the joint preparation
design.
A set of preliminary tests was conducted to evaluate the
individual effect of the welding parameters as current,
shielding gas, travel speed and electrode angle on the weld
bead shape and penetration as well as defect s, and these
were included to build a matrix of data. Then, by choosing
the best results, a matrix was built for positional welding.
Having established the optimum combination of parameters,
these were applied on an actual bevelled pipe. Table 3 presents
the welding parameters used to run tests on a pipe with the
grove shown in Fig. 5.
Fig. 5 J-bevel joint preparation
Table 3 Welding parameters used for tests on pipe 1
Trial WFS
(m/min)
TS
(m/min)
CTWD
(mm)
Trim Wave
control
Gas
(l/min)
Back shielding
(l/min)
Real WFS
(m/min)
Current
(A)
Voltage
(V)
Power
(W)
Heat input
(kJ/mm)
1 8 0.5 12 1.5 10 16 30 7.62 166.90 24.87 4,792.20 0.5751
2 7 0.5 12 1.5 10 16 30 6.64 152.48 25.12 4,502.31 0.5403
3 7 0.5 12 1.5 10 16 16 6.68 152.25 24.00 4,263.65 0.5116
4 8 0.5 12 1.5 10 16 16 7.62 166.59 24.88 4,764.99 0.5718
5 9 0.5 12 1.5 10 16 16 8.39 163.67 24.54 4,535.96 0.5443
6 7.2 0.4 12 1.5 10 16 16 8.45 163.62 24.49 4,532.88 0.6799
7 7.5 0.4 12 1.5 10 16 16 6.82 154.30 25.21 4,529.72 0.6795
8 7.2 0.4 12 1.5 10 16 16 6.74 154.30 25.00 4,493.24 0.6740
9 8 0.4 12 1.5 10 16 16 6.69 155.79 25.39 4,587.47 0.6881
10 9 0.5 12 1.5 10 16 16 6.69 154.82 25.42 4,584.88 0.5502
11 9 0.5 12 1.5 10 16 16 7.63 165.13 25.02 4,725.60 0.5671
12 9 0.5 12 1.5 10 16 16 7.71 164.70 24.38 4,614.08 0.5537
Int J Adv Manuf Technol

Welds were also performed on a second pipe varying
GMAW parameters (Table 4). For the root pass with gas
tungsten arc welding (GTAW), process parameters were
kept constant as shown in Table 5. Back shielding with
argon at a flow of 16 l/min was used to prevent oxidation.
3 Results and discussion
The objective of this research was to establish the optimum
combination of parameters to weld a bimetal pipe to obtain
a consistent weld shape during the continuous circumfer-
ential pipe welding. Gas metal arc welding was performed
from the outside. A root pass by GTA was made from the
inside, in some trials, to reinforce the weld, assuring full
penetration with a smooth transition between the weld and
the base mat erial on the inner surface to improve corrosion
and fatigue resistance. Additionally, attention was given to
the interface between the two passes in order to prevent
gaps which had a detrimen tal effect on pipe performance
under operating conditions (Fig. 6).
Cross-sections of the fully penetrated GMA-welded
specimens were observed. In the first pipe, specimens
number 5 and 9 (Fig. 7a, b) showed full penetration with
GMAW. Trial no. 11 (Fig. 7c) that also s howed full
penetration was re-melted by GTA welding to analyse the
effect of re-melting the duplex stainless steel of inner layer.
Other specimens showed incomplete penetration and were
welded from the inner side with GTAW.
Figure 8 shows the interpenetration effect of both passes,
the outside one with MAG and the internal one with TIG.
No defects were observed, and a smooth surface of the
second pass was seen, whic h is beneficial r egard ing
corrosion and fatigue behaviour.
The effect of sever al variables on the bead geometry
ofaweldinTIGweldingofduplexstainlesssteelwas
evaluated and discussed as follows.
3.1 Effect of welding speed and current intensity
Current intensity and travel speed directly affect heat
input of the welding process. As the travel speed
increas es, less he at is tran sfe r red into th e fusion zo ne
per time unit. So the amount of molten metal is low, and
hence less penetration is achieved. The same is observed
for current intensity since higher currents induce more
depression in the molten metal, which causes turbulence
in the weld pool and undercuts. The results for the effect
of the current and travel speed with different shiel ding
gases are depicted in Figs. 9 and 10, and these comply
with results from other researchers [10–12]. However, the
increase of penetration is not linear since for low currents,
Lorenz force is not the dominant molten flow driving
force, but Marangoni effect is, which causes outwards
flow. Since the m aterial has a low sulphur content, reverse
Marangoni effect due to sulphur is absent. In this case,
reverse Marangoni effect is due to the oxygen content in
the weld m et al when carbon dioxide is u se d a s a sh i eld ing
gas. This effect is usually more pronounced at lower
current intensities, while Lorenz force is the dominating
factor at higher current levels.
Adjusting welding speed and current intensity is insuffi-
cient to achieve good bead shape with adequate penetration
and width. The effect of the low level of surface active
elements such as oxygen in gas mixture and sulphur in welded
material significantly influences the bead shape [10–14].
Table 4 Welding parameters used in GMAW on pipe with J groove
Trial WFS
(m/min)
TS
(m/min)
CTWD
(mm)
Trim Wave
control
Gas
(l/min)
Real WFS
(m/min)
Current
(A)
Voltage
(V)
Power
(W)
Heat input
(kJ/mm)
1 10 0.5 11 1.5 10 16 9.58 188.33 24.83 5,209.887 0.6252
2 9.8 0.5 11 1.5 10 16 9.30 185.99 24.86 5,143.675 0.6172
3 9.8 0.8 11 1.5 10 16 9.23 187.26 23.38 4,952.209 0.3714
4 9.8 1 11 1.5 10 16 9.25 188.41 23.05 4,950.702 0.2970
5 9.8 0.95 11 1.5 10 16 9.16 186.10 23.33 4,917.364 0.3106
6 9.8 0.95 11 1 10 16 9.35 181.35 20.71 4,199.567 0.2652
7 9.8 0.95 11 1 10 16 9.28 184.69 20.64 4,289.015 0.2709
8 9 0.9 11 1 10 16 8.65 167.18 20.55 3,778.55 0.2519
9 9 1 11 1 10 16 8.56 165.43 20.68 3,794.611 0.2277
10 9 0.8 11 1 10 16 8.51 167.55 20.28 3,826.418 0.2870
Table 5 Welding parameters for root pass with GTAW
Current
(A)
Travel
speed
(mm/
min)
Position Shielding
gas
Flow
rate
(l/min)
Electrode
angle
Electrode
gap (mm)
170 200 Flat He–25%Ar 14 30º 2
Int J Adv Manuf Technol

3.2 Effect of shielding gas
Shielding gas is an operating parameter with the potential of
affecting weld bead geometry, besides changing the chemical
composition of the fusion zone. When aiming to increase the
heat input, additional heat can be generated by using gases
with higher ionisation potential, such as helium or hydrogen,
or additional elements that produce an exothermic reaction
during dissociation, as CO
2
. This can be used in case of low
current density or high welding speeds to obtain the same
results in terms of bead geometry. Suitable selection of the
shielding gas all ows to protect the fusion zone from the
surrounding atmosphere and to achieve a better arc
characteristic [15, 16].
Additionally, the shielding gas can correct the chemical
composition of the fusion zone; for instance, when welding
austenitic or duplex stainless steels, small additions of
nitrogen can favour the presence of small amounts of high
temperature delta ferrite, increasing the resistance to hot
cracking exhibited by these alloys.
Fig. 6 Schematic representation
of the project objective
Fig. 7 Macrographs of tests 5 (a) and 9 (b) welded by GMA showing full penetration and re-melted from the inner side with GTAW (c)
Int J Adv Manuf Technol

3.2.1 Argon
Figures 9 and 10 show the effect of shielding gas on the
penetration depth. Macrographic observations of welds
produced using pure argon as a shielding gas on this
particular material gives a sound weld free of defects.
However, penetration is unsatisfactory, and the weld beads
are also too wide. Since argon has a low ionisation potential,
it cannot contribute much to the heat input of the welding
process, and this is completely controlled by the current
intensity and the welding speed. Flow pattern is also
outwards which means that increasing the current or
decreasing travel speed will not lead to higher penetration
and extra heat input gives a wider weld bead.
3.2.2 Oxygen
The effect of oxygen as an active surface element was studied
by adding CO
2
to the shielding gas. Three mixtures were
tested: Ar–2%CO
2
,Ar–0.7CO
2
and He–44.5%Ar–1.5%CO
2
.
Based on previous studies [15, 17], the existence of over
100 ppm oxygen in the weld pool reverses the flow pattern.
Oxygen changes the surface tension temperature coefficient
from negative to positive. Thus, increasing the temperature
will increase the surface tension force. This is due to the
formation of an oxide layer on the weld surface which
prevents the weld pool from absorbing more oxygen. In the
argon–helium mixture, 1.5% increased penetration signifi-
cantly up to 4 mm (see Fig. 9). However, CO
2
content
resulted in rapid d egradation of the electrode, w hich
cannot be changed during a continuous orbital welding.
The contamination and deterioration made the electrode
tip blunt led to w eld pool instability, undercuts and
asymmetrical welds.
3.2.3 Hydrogen
Since hydrogen has a high thermal conductivity, the volume
of molten metal increases and a high penetrat ion is
Fig. 8 Macrograph of a specimen of the second pipe showing good
root welds free of geometrical and metallurgical defects
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
Ar
Ar-0.7%CO2
Ar-2%CO2
Ar-20%He-2.5%N
Ar-30%He
He-25%Ar-5%H
He-44.5%Ar-1.5%CO2
Penetration depth (mm)
150 A
200 A
250 A
Fig. 9 Variation of penetration depth with current intensity for
different shielding gases (welding speed=200 mm/min, electrode
angle=60°)
0
0
.
5
1
1
.
5
2
2
.
5
3
3
.
5
Ar
Ar-0.7%CO2
Ar-2%CO2
Ar-20%He-2.5%N
Ar-30%He
He-25%Ar-5%H
He-44.5%Ar-1.5%CO2
Penetration depth (mm)
200 mm/min
400 mm/min
Fig. 10 Variation of penetration depth with shielding gas composition
for different welding speeds (I=150 A, electrode angle=60°)
0
.
1
0
.
12
0
.
14
0
.
16
0
.
18
0
.
2
0
.
22
0
.
24
100 125 150 175 200
Current (A)
D/W ratio
30º
60º
Fig. 11 Variation of D/W ratio with the current intensity and the
electrode angle (welding speed=200 mm/min, shielding gas=argon)
Int J Adv Manuf Technol

achieved [18]. A noticeable effect is observed when
adding hydrogen, that is the penetration of samples
welded with low current was higher than that of samples
welded with high current intensity (Fig. 9). Since
hydrogen increases the heat input, increasing the current
will concentrate the heat in the centre of the weld and the
temperature gradient is steeper. Therefore, the Marangoni
forcewillbemoreeffectiveamplifyingtheoutwardsflow
in the molten pool .
3.2.4 Nitrogen
Nitrogen did not make a significant change in penetration
for high welding speeds (see Figs. 9 and 10) though it
increases penetration at high current intensities and low
travel speed. Its major benefit derives from its abil ity to
stabilise austenite in the weld metal preventing fully ferritic
microstructures.
3.2.5 Helium
Argon has a low ionisation potential in comparison with
some other inert gases as well as helium [17]. So the
addition of such elements to argon can increase the total
amount of heat in the weld pool. Mixtures as 70%Ar–30%
He did not make a significant change in penetration due to
the predominance of Marangoni forces. High heat input
contributes to increase bead width. Increasing the helium
content to obtain a mixture of 75%He–25%Ar introduces
more heat and increase the penetration. However, accord-
ing to [ 17], it does not enhance the D/W ratio considerably
but still sustains a good quality in the weld. This shows
that, although by increasing the heat input higher molten
metal would be achieved, surface tension still plays the
major role because of the low content of sulphur. So, D/W
ratio is limited when higher heat input is introduced into
the weld.
3.3 Electrode angle
Larger electrode angles have been reported to introduce a
more concentrated arc on the weld zone resulting in a higher
penetration in the parent metal [8]. However, it was observed
that by increasing the electrode angle, penetration reduces.
The D/W ratio is also decreased for larger angles (Fig. 11).
For materials with high content of sulphur, flow pattern is
inwards. A more concentrated arc increases the temperature
centre of the weld pool as a result of accelerated emission of
electrons into the weld pool from the arc which generates an
extensive heat. This extensive heat causes a higher temper-
ature gradient from the weld pool centre to the edges which
reinforces the inwards flow pattern and amplifies the
penetration. On the other hand, in low sulphur content, an
outward flow pattern dominates the weld pool circulation.
So, the higher arc concentration is a result of a blunt
electrode tip that increases temperature gradient between
the edges, and the centre enhances outwards flow and
leads to less penetration (see Fig. 12).
As a result, when GTA welding low sulphur content
alloys aiming at higher penetra tions, the use of sharper
Fig. 12 Effect of electrode
angle on bead geometry for
materials with a high sulphur
content and b low sulphur
content
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
12h 3h 6h 9h
Orbital position
Penetration depth (mm)
Fig. 13 Variation of weld penetration with different orbital positions
(welding current=150 A, welding speed=200 mm/min, shielding gas=
He–25%Ar)
Int J Adv Manuf Technol

electrode angles gives more penetration, unless there is an
additional surface active element in the shielding gas, as
oxygen, to inverse the surface tension temperature coefficient
and change the flow pattern.
3.4 Orbital positions
To analyse the effect of welding position on bead geometry,
welds were performed in several orbital positions. High
current intensities or low travel speed p roduce large
quantities of molten metal which do not cause problems
in flat and overhead position. Gravitational force in vertical
positions overcomes the surface tension force and may
create defects in the weld seam. This does not happen with
overhead position since depression force of the arc
neutralises gravitational force.
So it is difficult to have a consistent bead geometry
around the pipe using constant parameters. Instead, it is
more appropriate to use a power source and a rotator which
can modify the current and the welding speed when the
welding position.
Figure 13 depicts the penetration depth for different
orbital positions. The results, however, show that it is
possible to re-melt a penetrated weld using low current
intensity and high travel speeds which gives a small weld
pool and no lack of penetration. This procedure decreases the
productivity but guarantees a good quality inside the pipe.
4 Conclusions
The present study aimed at optimising GTA orbital welding
in butting bimetal pipes in duplex stainless steel that results
in an adequate penetration on the inner surface. The effect
of operating parameters such as current intensity, welding
speed, electrode angle and shielding gas composition in
various orbital positions inside the pipe was investigated.
The following conclusions were drawn:
1. Changing the current and the welding speed does not
increase penetration when the flow pattern is domi-
nantly outwards due to Marangoni effect unless the
current exceeds a value such that the depression force
overcomes the surface tension force. However, this
results in unsatisfactory weld surfaces and undercuts.
2. Sharper electrode angles resulted in higher penetration
which is also as a result of the Marangoni effect.
3. Adding CO
2
to argon increased penetration but caused
electrode deterioration and increased the possibility of
weld pool oxidation.
4. Additions of hydrogen significantly increased penetration.
However, the use of hydrogen has an associated safety
and porosity risk.
5. Helium introduces more heat to the fusion zone and
slightly increases penetration.
6. A combination of I =170 A, WS=200 mm/min,
electrode angle=30° and He–25%Ar as the shielding gas
were defined for narrow groove welding of a J-bevelled
pipe in flat position.
7. No welding defects were detected in the inner layer in
duplex stainless steel.
References
1. (1998) Bi-metal line pipe provides improved corrosion protection. Oil
Gas J 58(11). http://www.offshore-mag.com/index/article-display/
24751/articles/offshore/volume-58/issue-11/news/general-interest/
bi-metal-line-pipe-provides-improvedcorrosion-protection.html
2. BUTTING Co. (2008) Bi-metal pipe (BuBi®pipe). http://www.
butting.de/bubi-pipe.html
3. Bingen M, Martensen E (2004) BuBi-pipe [butting-bimetal-pipe]
—the intelligent cost-saving solution. Stainl Steel World 16:14–17
4. Banse J (1998) New material alternative: stainless and nickel
bonded pipes. Stainl Steel World 10(3):48–49
5. DNV Pipeline Committee meeting (2008) Reel installation of clad
steel pipe. www.edvermeulen.com/statoilclad180205.pdf. Accessed
08/01
6. American Welding Society (2004) Welding handbook, vol 2. Welding
Processes, Miami
7. Gunn RN (1997) Duplex stainless steels: microstructure, properties
and applications. Abington Publishing, Letchworth
8. Lippold JC, Kotecki DJ (2005) Welding metallurgy and weldability
of stainless steels. Wiley Interscience, Hoboken, NJ
9. Lincoln Electric Company (2005) RapidArc™—high speed
GMAW welding, NX-2.80. http://content.lincolnelectric.com/
pdfs/products/literature/nx280.pdf
10. Mills KC, Keene BJ, Brooks RF (1998) Marangoni effects in
welding. Phil Trans Math Phys Eng Sci 356(1739):911–925
11. Lu SP, Fujii H, Nogi K (2004) Marangoni convection and weld
shape variations in Ar–O
2
and Ar–CO
2
shielded GTA [TIG]
welding. Mater Sci Eng, A 380(1–2):290–297
12. Lu SP, Fujii H, Sugiyama H (2003) Marangoni convection
and welding pen etrati on in ATIG welding. Trans JWRI 32
(1):79–82
13. Heiple CR, Roper JR (1982) Mechanism for minor element effect
on GTA fusion zone geometry. Weld J 61(4):97
14. Leinonen JI, Ka rjalainen LP (1990) Unexpected weld pool
profiles in GTA [TIG] welding with oxidising shielding gas.
In: David SA, Vitek JM (ed) Recent trends in welding science
and technology proceedings, 2nd International Conference,
Gatlinburg, TN, 14–18 May 1989, ASM International, Materials
Park, OH, pp. 387–390
15. Lu SP, Fujii H, Nogi K (2004) Weld shape comparison with iron
oxide flux and Ar–O
2
shielding gas in gas tungsten arc welding.
Sci Technol Weld Joining 9(3):272–276
16. Shanping L, Hidetoshi F, Kiyoshi N (2005) Effects of CO
2
shielding gas additions and welding speed on GTA weld shape. J
Mater Sci 40(9):2481–2485
17. Lu SP, Fujii H, Nogi K (2005) Influence of welding parameters
and shielding gas composition on GTA [TIG] weld shape. ISIJ Int
45(1):66–70
18. Tušek J, Suban M (2000) Experimental research of the effect of
hydrogen in argon as a shielding gas in arc welding of high-alloy
stainless steel. Int J Hydrogen Energy 25:369–376
Int J Adv Manuf Technol