Content uploaded by Obeid Obeid
Author content
All content in this area was uploaded by Obeid Obeid on Jul 10, 2019
Content may be subject to copyright.
1
A parametric study of thermal and residual
stress fields in lined pipe welding
Obeid Obeid1, a, Giulio Alfano1,b, Hamid Bahai1,c, Hussam Jouhara1,d
1College of Engineering, Design and Physical Sciences, Brunel University, UB8 3PH
Uxbridge, UK
aobeid.obeid@brunel.ac.uk, bgiulio.alfano@brunel.ac.uk, chamid.bahai@brunel.ac.uk,
dhussam.jouhara@brunel.ac.uk
ABSTRACT
Welded lined cylindrical structures such as boilers, pressure vessels and transportation pipes
are widely used in the oil and gas industries because an inexpensive outer layer is protected
from corrosion by a thinner expensive layer, which is made of a corrosion resistant alloy
(CRA). Welding in the lined pipe is of two different types, where the first one, so called weld
overlay (lap-weld), is deployed to seal the liner with the outer pipe whilst the other one,
known as girth welding (butt-weld), is deposited to join two specimens of lined pipe together.
Therefore, the precise prediction of the thermal and residual stress fields due to the
combination of two different types of circumferential welding is a major concern regarding
welded lined pipes to avoid sudden failure during service. Six parametric studies have been
conducted primarily to examine the influence of welding properties (weld overlay and girth
welding materials), geometric parameters (weld overlay and liner) and welding process
parameters (heat input) on the thermal and residual stress fields. All predicted results
obtained from a 3-D FE model based on the ABAQUS code are validated against small-scale
experimental results. Furthermore, in this study, the effect of mesh size has been investigated.
Keywords: Lined pipe; Weld overlay; Girth welding; Thermal history; Residual stress
Nomenclature
Half-length of heat source (mm)
Depth of heat source (mm)
Half-width of heat source (mm)
Total strain
Elastic strain
Plastic strain
Thermal strain
Convective heat transfer coefficient (W/m2K)
Current (Amperes)
NT11
Temperature (°C)
Power density (Wm-3)
2
Total heat input (W)
Radial distance of the heat torch centre from the pipe axis (mm)
S, S22
Hoop residual stress (N)
S, S33
Axial residual stress (N)
Welding time (s)
Current pipe temperature (°C)
Ambien temperature (°C)
Welding speed (mm/s)
Voltage (volts)
Z
Axial direction starting from the WCL (mm)
2-D
Two-Dimensional
3-D
Three-Dimensional
A/D
Analog-to-digital converter
BM
Base material
CRA
Corrosion resistant alloy
C-Mn
Carbon-Manganese
FE
Finite element
FZ
Fusion zone
HAZ
Heat affected zone
TFP
Tight-fit-pipe
TIG
Tungsten Inert Gas
WCL
Weld centre line
WM
Welding material
Angle of moving torch around the pipe (Rad)
Welding efficiency (%)
Stefan-Boltzmann constant
Effective radiation emissivity ((W/m2 K4)
Von Mises stress (Pa)
1. Introduction
Welding, in general, is a reliable process widely used in industry to join two specimens
together with a high strength bond. In particular, oil and gas applications depend significantly
on welding. Although it is a necessary process, the main problems of using lined pipe
welding arise from the high temperatures at which two completely different filler materials
are deposited into two different welding grooves, the weld overlay and girth welding grooves,
which in turn lead to higher residual stresses concentrated in the two fusion zones (FZ) and
heat affected zones (HAZ) [1]. Therefore, predicting the locations and magnitudes of residual
stresses after completing the lined pipe welding operation is important to determine the
reliability and integrity of welded structures. A lot of research work has been conducted to
study the isothermal and residual stress fields induced by only single circumferential welding.
Karlsson and Josefson [2] studied the effect of thermal field, residual stresses and radial
shrinkage in a single-pass butt-welded C-Mn pipe (Carbon-Manganese) using 3-D FE
models. To enhance the accuracy of the numerical solution in the welding process, Teng and
Chang [3] also developed 3-D FE models to study temperature and stress fields in carbon
3
steel welded pipe with respect to the wall thickness variations. Deng and Murakawa [4]
presented 2-D and 3-D FE models to validate the numerical thermal history and residual
stress fields in multi-pass stainless steel pipe with experimental results. In their study, the
results of both 2-D and 3-D models are consistent with the experimental results. Yaghi et al.
[5] produced an axisymmetric thermomechanical FE model to predict the residual stresses in
a circumferentially butt-welded P91 steel pipe. Moreover, the effect of phase transformation
from austenite to martensite is considered in the simulation. Two methods were used to
measure the residual stresses along the outer surface by means of X-ray diffraction and deep-
hole drilling techniques. Dehaghi et al. [6] used an axisymmetric 2-D model to join a nozzle
with a pipe in a power plant reactor due to complexity of welding processes, buttering and
heat treatment,. It is pointed out that the buttering and heat treatment leads to reduce residual
stresses in the nozzle and pipe, respectively.
A few research investigations have studied circumferential welding subjected to parametric
factors. Brickstad and Josefson [7] used axisymmetric FE models to simulate multi-pass
circumferential butt-welding of stainless steel pipe. In particular, the residual axial and hoop
stresses across the wall thickness were discussed according to the variation in weld
parameters, namely pipe size, heat input, weld metal yield stress and inter-pass temperature.
The effect of the yield stress of the welding material on the residual stresses was investigated
by Deng et al. [8]. Beyond the weld metal and its vicinity, significant discrepancies exist
between the numerical and experimental results because of initial residual stresses produced
by pre-heat treatment. Malik et al. [9] discussed the effect of welding speed on residual
stresses. This study proves that a lower welding speed leads to a greater heat input.
Consequently, the residual stresses increase because the FZ and HAZ become wider. The
model developed by Zhao et al. [10] was used to study the effect of heat input and layer
number on the residual stresses in a dissimilar butt-welded pipe where one pipe was made of
austenitic stainless steel (S30432) and the other one was made of martensitic steel (T92).
Their study states that a decrease in heat input would lower the tensile residual stresses in the
S30432 steel more than those in the T92 steel. The possible reason is attributed to the yield
stress which is much less in the S30432 steel than in the T92 steel. A 3-D FE numerical
model was carried out by Velaga and Ravisankar [11] to study the effect of sixteen different
geometrical conditions of heat source on welded austenitic stainless steel pipe. The results
point out that there is a slight effect on the temperature history and weld pool size whilst
there is no considerable influence on residual stress distributions.
4
However, there are no detailed experimental or numerical studies conducted for lined pipe
welding. Furthermore, no study has investigated the influence of different factors on lined
pipe welding. Consequently, Obeid et al. [12] presented a new procedure to simulate a typical
lined pipe process including the weld overlay and girth welding. Furthermore, a sensitivity
analysis to determine the influence of the cooling time between weld overlay and girth
welding and of the welding speed has been conducted thermally and mechanically.
In this study, six cases have been investigated by changing different factors affecting the
quality and results of the welding process. Case A is considered the reference case, where the
weld overlay and girth welding have accordingly been modelled with different materials for
their base metals. In case B, the material of girth welding is the same material used in weld
overlay, namely austenitic stainless steel. Case C considers the effect of neglecting the weld
overlay where the two pieces of lined pipe have been joined solely using girth welding. In
this case, the material of girth welding is the same used in case A. However, the weld overlay
is used to seal the liner with the outer pipe which in turn blocks the gap between the liner and
outer pipe at the pipe ends. The heat input plays a key role in the welding deformation and
the residual stresses [13]. Therefore, in case D, the heat input is lowered to 75% of the heat
input in case A for all welds. In a similar way, the heat input in case E is dropped to 50% of
that in case A. The last case is case F where the liner with weld overlay is not considered.
To study the effect of specific parameters, the other parameters are kept constant and equal to
the values of the reference case (case A). Furthermore, the mesh topology for all FE models
remains with the same arrangement as in case A. The numerical thermal fields and residual
stress distributions are compared against the experimental ones using thermocouples and
residual stress gauges in all cases.
2. Manufacturing procedure
In this study, the specimen of welded lined pipe schematically shown in Fig. 1 was
manufactured from two adjacent pipes. The outer pipe is made of low carbon steel equivalent
to E235 AISI 10305-1, known as C-Mn pipe, with an outer diameter of 114.3 mm and a wall
thickness of 6.35 mm. The inner pipe is made of austenitic stainless steel which is rich in Cr
and Ni, known as AISI304 pipe, with an outer diameter of 101.6 mm and a wall thickness of
1.5 mm. The entire length of welded lined pipe, composed of two components, is 400 mm.
The AISI304 pipe was inserted inside the C-Mn pipe using tight fit pipe (TFP) thermal
5
manufacturing process, which is based on heating the outer pipe and cooling down the inner
one [14].
Fig. 1 Schematic semi-sketch of the welded lined pipe, dimensions in mm.
During welding, the heat source was fixed and the two sections of lined pipe were rotated
with a uniform speed for each pass (one-pass weld overlay and two-pass girth welding). The
weld overlay pass took 240 seconds to complete one revolution. Then, 270 seconds were
consumed as inter-pass time between the weld overlay and girth welding to cool down the
lined pipe naturally to the final maximum inter-pass temperature, which was around 100 °C.
The first and second pass of girth welding required 270 seconds each, too. Also, there was a
second inter-pass time between the two girth welding passes, again of 270 seconds. After the
second girth welding pass, the entire lined pipe took 3000 seconds to finally cool down
naturally to ambient temperature. In all passes, welding began at the central angle and
then progressed through the anti-clockwise circumferential direction to complete one rotation
and stop at the same starting point. Tungsten Inert Gas (TIG) welding was used for
all welds where ER308L stainless steel rod was inserted in the weld overlay groove whilst
E70S2 mild steel rod was utilized to deposit the girth welding. Fig. 2 shows the lined pipe
specimens during the welding overlay and girth welding.
6
(a)
(b)
Fig. 2 Recording the temperatures during the (a) weld overlay and (b) girth welding
To record the thermal history, HI-766F K-type thermocouples, made of AISI 316 stainless
steel, were placed at 6 axial locations with 270° central angle. The thermocouples can record
temperatures up to 1100°C. The maximum accuracy of such thermocouple type is ±2.2°C.
Three thermocouples were mounted on the outer surface (C-Mn pipe) and the others on the
inner surface (AISI304) to record the thermal history at those locations during welding and
cooling as shown in Fig. 3. The thermal history results were recorded and stored every 0.001
second by LabVIEW software via a 24-bit A/D interface (NI 9213). The maximum accuracy
of the A/D interface is ±2.25°C. Consequently, the maximum error in the measured thermal
history is the interval ±4.45°C.
To measure the residual stresses after completing welding and cooling down to ambient
temperature, 14 residual stress gauges with three elements, FRS-2, were mounted also on the
outer surface (C-Mn pipe) and the inner surface (AISI304). The tolerance of the gauge factor
of the FRS-2 gauges is ±1% at room temperature. The residual stresses were also recorded
every 0.001 second using LabVIEW software via a 24-bit A/D interface (NI 9235) with
accuracy ±0.4%. Therefore, the error in the measured residual stress can reach a range of
±1.4%. A reference hole with diameter and depth of 2 mm each was drilled vertically through
the pipe thickness using a high speed milling machine as shown in Fig. 4.
7
Fig. 3 Locations of thermocouples with the welding direction for three passes, dimensions in mm.
Fig. 4 Recording residual stress from rosette gauges, FRS-2.
3. Finite element modelling
In this study, all thermal and structural models have been executed using ABAQUS [15]. Due
to the symmetry around the weld centreline, only one component, in particular, one-half of
the welded lined pipe was modelled. It is clear that the mechanical properties in welding
depend on the temperature whereas temperatures are assumed to be independent of
mechanical deformation. Thus, the thermal analysis is simulated first to obtain the thermal
history with respect to time for all nodes of the welded lined pipe. This thermal history is then
accordingly transferred to the mechanical analysis as thermal loads.
8
In the thermal analysis, the element type is selected as a continuum, three-dimensional 20-
node quadratic brick diffusive heat transfer element, named DC3D20 in ABAQUS. At each
of its 20 nodes, there is one degree of freedom, which is the temperature. In the mechanical
analysis, the element type is a continuum fully-integrated three-dimensional 20-noded
element, named C3D20 in ABAQUS. Each of the 20 nodes undergoes three translation
degrees of freedom to keep one element subjected to 60 degrees of freedom totally. Both the
3-D thermal and mechanical models have the same mesh associated with the same numbers
and arrangements for nodes and elements as shown in Fig. 5. In the reference case (case A),
the model consists of 35220 nodes associated with 7380 elements. Due to the high
temperatures and their high gradients in the FZ and HAZ, a finer mesh can be seen in these
regions of the weld overlay for the inner pipe, and of the two-pass girth welding for the outer
pipe. The weld overlay, AISI304 pipe, girth welding and C-Mn pipe are coloured with red,
light blue, yellow and green, respectively, as portrayed in Fig. 5.
Fig. 5 3-D FE model (case A)
The element birth technique is adopted in the FE models to simulate the deposition of the
filler materials in the weld overlay and girth welding grooves while moving the heat source.
In this study, both base and weld metals have the same thermal-mechanical material
properties except the yield stress, where the weld material is supposed to have higher yield
strength in both C-Mn and AISI304 as tabulated in Table 1 and Table 2, respectively.
9
Table 1 Thermo-mechanical properties of C-Mn [2].
Temperature
(°C)
Density
(Kg/m3)
Specific
heat
(J/Kg
K)
Conductivity
(W/m K)
Thermal
expansion
(x10-5K-1)
Yield stress
(MPa)
Young’s
modulus
(GPa)
Poisson’s
ratio
Base
Weld
0
7860
444
50
1.28
349.45
445.42
210
0.26
100
480
48.5
1.28
331.14
441.29
200
0.28
200
503
47.5
1.30
308.00
416.49
200
0.29
300
518
45
1.36
275.00
376.18
200
0.31
400
555
40
1.40
233.00
325.54
170
0.32
600
592
35
1.52
119.00
172.59
56
0.36
800
695
27.5
1.56
60.00
43.41
30
0.41
1000
700
27
1.56
13.00
14.47
10
0.42
1200
700
27.5
1.56
8.00
9.30
10
0.42
1400
700
35
1.56
8.00
9.30
10
0.42
1600
700
122.5
1.56
8.00
9.30
10
0.42
Table 2 Thermo-mechanical properties of AISI304 [4].
Temperature
(°C)
Density
(kg/m3)
Specific
heat
(J/kg K)
Conductivity
(W/m K)
Thermal
expansion
(x10-5K-1)
Yield stress
(MPa)
Young’s
modulus
(GPa)
Poisson’s
ratio
Base
Weld
0
7900
462
14.6
1.70
265
438.37
198.50
0.294
100
7880
496
15.1
1.74
218
401.96
193
0.295
200
7830
512
16.1
1.80
186
381.5
185
0.301
300
7790
525
17.9
1.86
170
361.25
176
0.310
400
7750
540
18.0
1.91
155
345.94
167
0.318
600
7660
577
20.8
1.96
149
255.71
159
0.326
800
7560
604
23.9
2.02
91
97.41
151
0.333
1200
7370
676
32.2
2.07
25
28.41
60
0.339
1300
7320
692
33.7
2.11
21
16.23
20.00
0.342
1500
7320
700
120
2.16
10
12.17
10
0.388
The latent heat for C-Mn steel is set to be 247 kJ/kg between the solidus temperature of 1440
°C and the liquidus temperature of 1560 °C. For stainless steel (AISI304), the latent heat is
assumed to be 260 kJ/kg between 1340 °C and 1390 °C, solidus and liquidus temperatures
respectively. Consequently, the melting point for C-Mn is 1500 °C while it is 1365 °C for
AISI304. The initial temperature of the lined pipe and the weld bead is set at room
temperature.
3.1. Thermal Analysis
During welding, the heat transfer, a combination of heat loss due to radiation and convection,
occurs upon external surfaces exposed to the environment. Radiation loss is dominating in the
weld zone and its vicinity whereby the temperature magnitudes are near or over the melting
temperature. Convection loss is dominating away from the weld zone. The Stefan-Boltzman
law and Newton’s law are applied to model the radiation and convection heat loss,
respectively. In this work, the thermal boundary conditions are applied on all external
10
surfaces of the lined pipe. The total heat loss, , is a combination of radiation, ,
and convection, , losses given as follows:
(1)
(2)
(3)
(4)
where is the convective heat transfer coefficien, is the current temperature at
the pipe surface, is the ambient temperature, is the effective radiation emissivity,
is the Stefan-Boltzmann constant and is the total combined heat-transfer coefficient.
As the lined pipe is composed of two different materials, each material is characterised by
different coefficients governing heat transfer with the room atmosphere, as shown in Table 3.
Table 3 heat transfer parameters
Parameters
C-Mn
AISI304
(W/m2 K)
8
5.7
0.51
0.75
(W/m2 K4)
5.67×10-8
5.67×10-8
A FILM user subroutine [15] has been coded in FORTRAN to implement in ABAQUS the
above expression for the total heat-transfer coefficient, which is Eq. (4), for liner and outer
pipe accordingly. It is worth noting that ABAQUS allows one single user-subroutine to be
written for both materials by simply specifying which surface each condition applies to.
The power density, , transmitted from the heat source to the lined pipe and weld regions is
modelled by a Gaussian distribution as a function of position and time, , in an ellipsoid
(welding pool) with centre that is taken as [12]:
(5)
where is the energy input rate which is given by the product of the current , voltage
and the weld efficiency , is the radial distance of the heat torch centre from the pipe
axis, is the angle from the start/stop point (where = 0°). Welding parameters, and
are the semi-axes of the ellipsoidal welding pool in directions, , and , respectively.
Equation (5) has been implemented in ABAQUS by coding the DFLUX user-subroutine. The
position of the weld torch is calculated first in DFLUX according to the welding time.
Thereafter, the power density, q, is computed at each integration point.
11
The numerical values for the variables used in the power density distribution in Eq. (5) are
illustrated in Table 4 for each welding material.
Table 4 Heat source and welding parameters.
Parameter
Symbol
Weld overlay
1st pass
girth welding
2nd pass
girth welding
Half-length of arc (mm)
4.9
6.2
6.2
Depth of arc (mm)
1.5
2.62
2.85
Half-width of arc (mm)
4.9
5.57
5.66
Welding current (A)
110
220
234
Voltage (V)
22
22
22
Welding speed (mm/s)
1.3
1.26
1.33
Welding time (s)
240
270
270
Welding efficiency
70%
70%
70%
3.2. Structural analysis
The same FE mesh used in the thermal analysis is employed in the mechanical analysis apart
from the boundary conditions and element type. Herein, the nodal temperature histories read
from the thermal output file are considered thermal loads for each increment in the
mechanical simulation. At each structural step, an automatic time increment is executed and
geometrical nonlinear effects (large deformation) have been incorporated in the FE model.
During the lined pipe welding process, the effects of volumetric change and the change in the
yield stress value (the transformation plasticity) due to the metallurgical phase
transformation, namely the martensitic phase transformation, have been neglected in this
work because the volume dilation [16] and the reduction in the yield stress value [5] due to
the phase transformation is small. Therefore, the increment of the total strain, , has been
broken down into three components as follows [12]:
(6)
where
,
and
are the elastic, plastic and thermal strain increment, respectively.
The elastic behaviour is defined by the isotropic Hook’s law. For the plastic behaviour, the
Von Mises,, yield criterion has been used with an associated flow rule with respect to the
three principal stresses, , and , given as below [10]:
(7)
12
The thermal strain is the result of expansion and contraction of the line pipe materials and is
governed by the temperature-dependant thermal expansion coefficient reported in Tables 1
and 2.
The Bauschinger effect should be taken into account in the structural analysis because all
material nodes are under the influence of multiple thermal loading and unloading. In the
kinematic hardening rule, the Bauschinger effect considers that the size and shape of the yield
surface keep the same with translating in the stress space. Consequently, a linear kinematic
hardening rule has been assumed for both materials C-Mn and AISI304 [12], with the
hardening parameter obtained from the temperature-dependant yield stress reported in Fig. 6,
when the plastic strain of C-Mn [9] and AISI304 [8] is equal to 1%.
Fig. 6 Yield stress of the base material (BM) and welding material (WM) for C-
Mn steel [9] and AISI304 [8] corrosponding to 1% hardening.
The mechanical boundary conditions are applied to restrict the axial movement on the
circumferential symmetry-plane. As the lined pipe is not clamped during welding, lateral and
transversal restrictions are employed at the lined pipe end to prevent rigid body motion
whereas the free expansion and contraction are allowed over the entire lined pipe.
4. Results and discussion
4.1. Temperature response in case A
Incorporating the heat source movement within the heat transfer analysis during welding is
complicated by mathematical and physical issues, because of the need to consider two
different types of welding (weld overlay and girth welding) associated with two different
parent materials at the same time. It is important to validate the FE model experimentally to
verify the accuracy of the moving heat source and heat transfer equations. The macrograph of
cross section at 270° in case A has been taken by means of a microscope where the FZ and
0
100
200
300
400
500
600
0 200 400 600 800 1000 1200 1400 1600
Yield Stress (MPa)
Temperature (°C)
Yield stress of BM (C-Mn)
Yield stress of WM (C-Mn)
Yield stress of BM (AISI304)
Yield stress of WM (AISI304)
13
HAZ boundaries are clearly distinct as shown in Fig. 7. From this figure, it can also be seen
that the predicted FZ and HAZ isotherms correlate well with the crystallization of the lined
pipe cross section. The minimum temperature at the FZ is 1365 °C and 1500 °C for weld
overlay and girth welding, respectively. The HAZ extends in the FZ vicinity to attain a
minimum temperature of 800 °C.
Fig. 7 The numerical FZ and HAZ isotherms and macroscopic examination at the cross section of 270°
central angle (case A).
Consequently, the macroscopic examination and numerical results of FZ and HAZ isotherms
prove that the accuracy is not only in the thermal equations used in this work but also in the
parameters applied to these equations such as welding pool geometries and heat input values.
For typical welding, temperatures in the fusion zones of weld overlay and two-pass girth
welding should be higher than the melting points, 1500 and 1365 °C for C-Mn and AISI304,
respectively, to make filler materials flow through the grooves. Moreover, all points located
on the same circumferential direction should have an identical temperature history.
Fig. 8 shows the numerically computed temperature distributions at 90°, 180° and 270°
central angle during weld overlay where the girth welding has not been deposited yet in case
A. As anticipated, the maximum temperature is achieved at the welding pool centre of weld
overlay, 1634°C, which is beyond the melting point of AISI304, 1500 °C. From this figure, it
can be seen that the thermal histories of weld overlay pool centres at three circumferential
locations, 90°, 180° and 270, have to a reasonable extent the same shape and magnitudes of
the transient thermal cycle.
14
Fig. 8 The thermal history (°C) of weld overlay centre at 90°, 180° and 270° central angle (case A).
Likewise, the numerically computed temperature history at the second pass of girth welding
also has identical distributions circumferentially around the mid-plane (symmetric line) at
three locations, 90°, 180° and 270° in case A. The three curves reach the same peak
temperature, 2076°C, and it could be seen that the weld overlay and first-pass girth welding
and three quarters of the second-pass girth welding have been laid down in their grooves as
shown in Fig. 9.
Fig. 9 The thermal history (°C) of second pass centre of girth welding at 90°, 180° and 270° central angle
(case A).
4.2. Comparisons of thermal results
To compare the thermal results of the reference case (case A) with other cases, Table. 5
shows the peak temperatures at the six thermocouples during the weld overlay, the first and
second pass of girth welding. The numerical temperature gradients are compared with the
results recorded experimentally.
0
200
400
600
800
1000
1200
1400
1600
1800
0 100 200
Temperature (°C)
Time (s)
Temp.90°
Temp.180°
Temp.270°
0
400
800
1200
1600
2000
2400
1100 1150 1200 1250 1300 1350
Temperature (°C)
Time (s)
Temp.90°
Temp.180°
Temp.270°
15
Table 5 Comparison between numerical and experimental results at six location during welding
Case
Pass
Inner surface °C
Outer surface °C
TC1
TC2
TC3
TC4
TC5
TC6
A
Overlay
Num.
540
350
271
446
342
271
Exp.
525
343
265
432
333
263
1-Girth
Num.
798
565
450
709
554
448
Exp.
775
550
441
695
540
435
2-Girth
Num.
929
703
573
918
690
570
Exp.
913
685
562
910
681
558
B
Overlay
Num.
538
525
348
331
271
260
446
430
341
332
271
264
Exp.
1-Girth
Num.
790
775
564
552
450
443
708
690
554
539
448
441
Exp.
2-Girth
Num.
928
911
704
685
574
561
911
895
692
680
570
561
Exp.
C
Overlay
Num.
-
-
-
-
-
-
-
-
-
-
-
-
Exp.
1-Girth
Num.
689
670
482
471
374
362
614
605
472
460
374
361
Exp.
2-Girth
Num.
894
882
669
650
539
525
878
865
658
645
537
525
Exp.
D
Overlay
Num.
408
401
271
265
212
208
345
337
268
262
212
208
Exp.
1-Girth
Num.
610
600
440
431
352
340
552
540
432
422
353
339
Exp.
2-Girth
Num.
716
701
553
542
453
438
701
688
548
535
453
451
Exp.
E
Overlay
Num.
276
270
191
185
151
145
240
232
188
179
151
142
Exp.
1-Girth
Num.
412
405
306
301
248
245
388
375
302
295
248
235
Exp.
2-Girth
Num.
496
485
387
375
320
315
492
485
386
378
320
315
Exp.
F
Overlay
Num.
-
-
-
-
-
-
-
-
-
-
-
-
Exp.
1-Girth
Num.
730
721
531
520
417
412
610
600
475
461
374
362
Exp.
2-Girth
Num.
1010
992
744
732
599
585
875
861
655
645
535
526
Exp.
Comparing the thermal results of case A against case B, it can be observed that the peak
temperatures for all points are close to each other with differences of less than 10 °C at all
points. Consequently, changing the girth welding material to austenitic stainless steel does
not influence the thermal results during welding. This can be attributed to the thermal
properties, namely the specific heat and conductivity, which are close to each other especially
at high temperatures.
In case C, the differences in temperatures measured by thermocouples drop drastically
compared with case A during the first pass of girth welding because the inter-pass
16
temperature is neglected in case C. In the second pass of girth welding, the variations in
temperatures between two cases are significantly narrower.
Decreasing the heat input from the heat source leads to a strong decrease in temperatures
during all welding passes. In more detail, the maximum temperatures predicted and recorded
by thermocouples in cases D and E are significantly lower than their counterparts in case A.
Furthermore, temperatures in case D are larger than those in case E because case D has 75%
of case A heat input, whilst case E just has 50%.
Removing the liner and weld overlay in case F keeps the whole thickness minimized to that
of the C-Mn pipe, equal to t = 6.35 mm. In this case, it can be seen that the peak temperatures
recorded and predicted at each thermocouple during the first girth welding pass are lower
than their counterparts in case A because the inter-pass temperature is not there anymore.
During the second pass of girth welding, the temperatures in case F are higher than those in
case A on the inner surface where the thickness of pipe is 6.35 mm. On the outer surfaces, the
temperatures are much closer to their counterparts in case C.
It can be observed that there is good agreement between the numerical and experimental
temperatures which are within a maximum variation of less than 6%. Thus, the developed
thermal FE models for all cases can be considered to be validated experimentally. Also, the
disparity between the results of case A and other cases becomes larger as the distance from
the WCL decreases. It is also observed from Table 5 that the thermocouple on either the outer
or inner surface located nearest to the WCL experiences a peak temperature higher than those
located farther from the WCL [17].
4.3. Structural response in case A
In case A, the axial and hoop residual stress distributions at 270° central angle are depicted in
Fig. 10(a) and (b), respectively. The bottom row of elements is the liner, AISI304 pipe, with
the weld overlay, whereas the rest of pipe is the backing steel pipe, namely C-Mn pipe, with
the girth welding.
It can be seen that maximum axial residual tensile stresses are located at the toes of the girth
welding, weld overlay and HAZ on the inner surface as shown in Fig. 10(a). Furthermore, the
axial tensile stresses on the inner surface are balanced by the axial compressive stresses on
the FZ and HAZ of girth welding on the outer surface [18]. Therefore, axial bending
deformation is produced through the pipe cross section. As a result, the diameter of lined pipe
becomes smaller in the FZ and HAZ regions after cooling down to room temperature because
17
of radial shrinkage [19]. Also, it can be seen that the length needed to reverse the tensile
stresses to compressive on the inner surface is narrower than that to reverse the compressive
stresses to tensile on the outer surface.
Turning to the hoop residual stress distributions shown in Fig. 10(b), the absolute values of
tensile stresses in the FZ and HAZ on the inner surface are significantly larger than those of
the compressive stresses in the girth welding region and its vicinity on the outer surface. The
magnitudes of residual axial stresses have a significant influence on the value of residual
hoop stresses [20]. The ranges of reversal stresses on the inner and outer surface are
somewhat close to each other.
(a)
(b)
Fig. 10 (a) Axial and (b) hoop residual stress distributions of case A at 270° central angle.
It is evident that the area of C-Mn steel at which the weld overlay is fixed with the C-Mn pipe
has relatively higher axial and hoop tensile residual stresses which result in von Mises
stresses larger than the yield stress of the C-Mn base material. This region is affected more
than others by the thermal cycles of weld overlay, first pass and second pass of girth welding.
Consequently, it is more likely that a crack initiates at this area as shown in Fig. 11.
18
Fig. 11 Initiation and growth of crack at the area of C-Mn pipe above weld overlay.
4.4. Comparison of structural results
4.4.1. Effect of welding materials on residual stresses
To avoid cracking and corrosion in the FZ, austenitic stainless steel is a proper filler material
to join two specimens together. Stainless steel is more capable of expanding and contracting
naturally during welding because of its larger coefficient of thermal expansion. Conversely,
carbon steel is a good conductor of heat which in turn cools more rapidly and shrinks faster
as the joint cools [21]. Moreover, stainless steel has a better corrosion resistance than carbon
steel because of its chemical composition. As a result, austenitic stainless steel welding is
preferred in the oil and gas industries.
Fig. 12(a)-(d) shows a comparison between the numerical results for cases A and B at 270°
central angle from the start/stop welding point along the axial direction starting from the
WCL, Z=0. The experimental results are also plotted for both cases accordingly using
residual stress gauges. The numerical axial and hoop residual stress distributions on the inner
surface (liner) for both cases are in a good correlation except at the toes of weld overlay and
girth welding (Z ≤ 3.6 mm). Within this zone, the maximum axial residual stress is 593 MPa
at Z = 0.3 mm in case A whilst the maximum one in case B is 529 MPa located at Z = 0.6
mm as shown in Fig. 12(a). Similarly, in the circumferential direction, the maximum hoop
residual stress is 573 MPa, over the yield stress of AISI304 welding material, at Z = 2.1 mm
in case A whereas the maximum one in case B is 481 MPa on the WCL as indicated in Fig.
12(b). On the outer surface, it can be seen that significant discrepancies exist between the
19
numerical results of case A and B in the FZ and its vicinity, for Z ≤ 45 mm, as shown in Fig.
12(c) and (d). Beyond this zone, the results in both cases are almost identical in the axial and
hoop residual stress distributions.
The experimental results recorded on the inner and outer surface with 270° central angle are
consistent with the numerical results in the FZ and HAZ because the initial residual stresses
produced by the manufacturing process (TFP) are removed by the high temperatures of
welding. Beyond this range, numerical temperature magnitudes are significantly lower. As a
result, the initial stresses still remain in the pipe and the experimental results are somewhat
larger.
(a)
(b)
-100
100
200
300
400
500
600
700
020 40 60 80 100 120 140 160 180 200
Axial stress (MPa)
Axial distance (mm)
Case A-FEA
Case B-FEA
Case A-Exp.
Case B-Exp.
-100
100
200
300
400
500
600
700
020 40 60 80 100 120 140 160 180 200
Hoop stress (MPa)
Axial distance (mm)
Case A-FEA
Case B-FEA
Case A-Exp.
Case B-Exp.
20
(c)
(d)
Fig. 12 Comparison of residual stresses at 270° central angle between case A and case B: (a) axial
stress distributions on the inner surface, (b) hoop stress distributions on the inner surface, (c) axial
stress distributions on the outer surface, and (d) hoop stress distributions on the outer surface.
4.4.2. Effect of welding overlay
Omitting the weld overlay allows dust and grease to go inside the gap between the liner and
backing steel pipe. Consequently, these go inside the girth welding and deteriorate the quality
of girth welding by forming voids and inclusions. Therefore, in case C welding is conducted
without weld overlay to study the influence of this factor on the stress behaviour.
Fig. 13(a)-(d) shows the hoop and axial residual stress distributions at 270° central angle on
the inner and outer surfaces for cases A and C. In this figure, the experimental results are also
plotted along the axial distance. It can be observed that there is a significant discrepancy in
the axial residual stress at the WCL where it is 540 MPa and 252 MPa in case A and C,
-700
-600
-500
-400
-300
-200
-100
100
200
020 40 60 80 100 120 140 160 180 200
Axial stress (MPa)
Axial distance (mm)
Case A-FEA
Case B-FEA
Case A-Exp.
Case B-Exp.
-250
-200
-150
-100
-50
50
100
150
020 40 60 80 100 120 140 160 180 200
Hoop stress (MPa)
Axial distance (mm)
Case A-FEA
Case B-FEA
Case A-Exp.
Case B-Exp.
21
respectively. Beyond these weld zones, the axial residual stress distributions in both cases A
and C are much closer to each other as shown in Fig. 13(a). Similarly, in the hoop direction,
there is a difference in the hoop residual stress at the weld zones. Beyond that, the results are
closer to each other in both cases as depicted in Fig. 13(b).
On the outer surface, there are significant discrepancies between the results of axial and hoop
residual stress in case A with their counterparts in case C at the weld zone of girth welding
and its HAZ as shown in Fig. 13(c)-(d). The experimental results are in good agreement with
the numerical results for both cases at the FZ and HAZ but they are larger beyond that
especially at the inner surface due to the effect of initial residual stresses of pre-heat
treatment (TFP).
(a)
(b)
-100
100
200
300
400
500
600
700
020 40 60 80 100 120 140 160 180 200
Axial stress (MPa)
Axial distance (mm)
Case A-FEA
Case C-FEA
Case A-Exp.
Case C-Exp.
-100
100
200
300
400
500
600
050 100 150 200
Hoop stress (MPa)
Axial distance (mm)
Case A-FEA
Case C-FEA
Case A-Exp.
Case C-Exp.
22
(c)
(d)
Fig. 13 Comparison of residual stresses at 270° central angle between case A and case C: (a) axial
stress distributions on the inner surface, (b) hoop stress distributions on the inner surface, (c) axial
stress distributions on the outer surface, and (d) hoop stress distributions on the outer surface. and (d)
hoop stress distributions on the outer surface.
4.4.3. Effect of heat input on welding residual stress
The heat input plays a key role in affecting the temperature distributions, which in turn leads
to significant changes in residual stresses. In this section, all welding parameters are kept
constant, such as the welding speed and welding pool geometries. The total heat input, , is
identified as (Watts) where is current (Amps), is voltage (Volts) and is the
weld efficiency. In case A, the total heat inputs are 850, 1700 and 1800 Watts for weld
overlay, first-pass of girth welding and second-pass of girth welding, respectively. Reducing
the heat input has some benefits in reducing consumption of the rod in TIG welding provided
the quality of welding is achieved without porosity (bubbles) in the weld because of the lack
-700
-600
-500
-400
-300
-200
-100
100
200
020 40 60 80 100 120 140 160 180 200
Axial stress (MPa)
Axial distance (mm)
Case A-FEA
Case C-FEA
Case A-Exp.
Case C-Exp.
-300
-250
-200
-150
-100
-50
50
100
150
020 40 60 80 100 120 140 160 180 200
Hoop stress (MPa)
Axial distance (mm)
Case A-FEA
Case C-FEA
Case A-Exp.
Case C-Exp.
23
of fusion. In this section, the influence of heat input on residual stresses has been investigated
through cases D and E. The total heat input is lowered to 0.75 and 0.5 of the heat input of
case A for case D and E, respectively. In more detail, the total heat inputs become 638, 1275
and 1350 Watts for weld overlay, first-pass of girth welding and second-pass of girth welding
in case D, respectively. In case E, the portions of heat input which have been provided to the
weld overlay, first-pass of girth welding and second-pass of girth welding are 425, 850 and
900 Watts, respectively.
Fig. 14(a)-(d) compares the axial and hoop residual stresses from the numerical analysis with
experimental data in the longitudinal direction starting from the WCL at 270° from the
start/stop welding location for cases A, D and E. On the inner surface (AISI304 pipe), the
maximum axial residual stresses in the three cases are located at Z = 0.3 mm at the toe of
girth welding with values of 590, 577 and 352 MPa for cases A, D and E, respectively, as
shown in Fig. 14(a). Turning to the hoop direction, it can be seen that the maximum tensile
hoop residual stresses in cases A, D and E take place at the centre of the weld overlay region,
Z = 2.1 mm, with values of 573, 371 and 502 MPa as shown in Fig. 14(b), respectively. From
Fig. 14(a) and (b), it is observed that the length of the zone with tensile residual stress
becomes narrower by reducing the magnitude of the heat input.
Fig. 14(c) and (d) depicts the axial and hoop residual stress distributions on the outer surface
(C-Mn pipe) for cases A, D and E at 270° central angle with respect to the axial distance. The
maximum axial compressive stresses on the outer surface are located at the WCL with -595, -
561 and -508 MPa for the three cases A, D and E, respectively. The lengths of the zones with
compressive residual stress are slightly close to each other where the zone for case E is still
narrower than others as indicated in Fig. 14(c). It is observed that the magnitude of hoop
residual stress on the outer surface is affected by its axial residual stresses. The larger the
compressive axial residual stress is, the larger the compressive hoop residual stress is [22]. In
a similar way, case E has the narrowest compressive range of all the cases, as shown in Fig.
14(d). Experimental results are in good agreement with their counterparts from numerical
analysis in the FZ and HAZ for all cases.
24
(a)
(b)
(c)
-100
100
200
300
400
500
600
700
020 40 60 80 100 120 140 160 180 200
Axial stress (MPa)
Axial distance (mm)
Case A-FEA
Case D-FEA
Case E-FEA
Case A-Exp.
Case D-Exp.
Case E-Exp.
-100
100
200
300
400
500
600
700
020 40 60 80 100 120 140 160 180 200
Hoop stress (MPa)
Axial distance (mm)
Case A-FEA
Case D-FEA
Case E-FEA
Case A-Exp.
Case D-Exp.
Case E-Exp.
-700
-600
-500
-400
-300
-200
-100
100
200
300
020 40 60 80 100 120 140 160 180 200
Axial stress (MPa)
Axial distance (mm)
Case A-FEA
Case D-FEA
Case E-FEA
Case A-Exp.
Case D-Exp.
Case E-EXP
25
(d)
Fig. 14 Comparison of residual stresses at 270° central angle among case A, case D and case E: (a)
axial stress distributions on the inner surface, (b) hoop stress distributions on the inner surface, (c)
axial stress distributions on the outer surface, and (d) hoop stress distributions on the outer surface.
4.4.4. Effect of liner on welding residual stress
The function of the liner is to protect the inner surface of the C-Mn pipe from corrosion. With
this function, it is made of corrosion resistant alloy CRA, austenitic stainless steel.
Consequently, removing the liner will not only lead to corrosion of the pipe in oil and gas
applications but it will also affect the residual stress behaviour especially at welding regions
[23].
Fig. 15(a)-(d) compares the axial and hoop residual stress distributions on the inner and outer
surface numerically and experimentally for case A and case F in which the liner with weld
overlay is removed. On the inner surface, the axial residual stress at the WCL in case F, 333
MPa, is lower than that in case A, 540 MPa, as depicted in Fig. 15(a). In the hoop direction,
the magnitude of hoop residual stress at the WCL in case F, 364 MPa, is larger than its
counterpart in case A, 203 MPa. With increasing distance from the WCL, the hoop residual
stress distribution drops rapidly in case F whereas the distribution in case A goes sharply up
within the weld overlay region as shown in Fig. 15(b). Furthermore, the extent of the axial
and hoop tensile stresses in case F is relatively narrower, Z = 19 mm, than that of case A, Z =
65 mm, on the inner surface as shown in Fig. 15(a) and (b). This can be attributed to the
absence of the liner and of the weld overlay at the inner surface which in turn slows down the
heat transfer of the exposed surface to the environment.
-250
-200
-150
-100
-50
50
100
150
200
020 40 60 80 100 120 140 160 180 200
Hoop stress (MPa)
Axial distance (mm)
Case A-FEA
Case D-FEA
Case E-FEA
Case A-Exp.
Case D-Exp.
Case E-EXP
26
On the outer surface, the maximum compressive axial stress in case F, -562 MPa, is located
within the FZ, at Z = 2.1 mm, whilst the maximum compressive axial stress in case A is
located at the WCL, -595 MPa, as shown in Fig. 15(c). In both cases, the hoop residual stress
distributions have a wave form as shown in Fig. 15(d). As with tensile stresses at the inner
boundary, the compressive range in case F is relatively narrower than that for case A. In
general, the numerical residual stress results agree reasonably well with the experimental
results obtained by using the hole-drilling strain gauge method.
(a)
(b)
-200
-100
100
200
300
400
500
600
700
020 40 60 80 100 120 140 160 180 200
Axial stress (MPa)
Axial distance (mm)
Case A-FEA
Case F-FEA
Case A-Exp.
Case F-Exp.
-300
-200
-100
100
200
300
400
500
600
700
020 40 60 80 100 120 140 160 180 200
Hoop stress (MPa)
Axial distance (mm)
Case A-FEA
Case F-FEA
Case A-Exp.
Case F-Exp.
27
(c)
(d)
Fig. 15 Comparison of residual stresses at 270° central angle between case A and case F: (a) axial
stress distributions on the inner surface, (b) hoop stress distributions on the inner surface, (c) axial
stress distributions on the outer surface, and (d) hoop stress distributions on the outer surface.
5. Mesh convergence analysis
The FE mesh density plays a key role in determining the accuracy of numerical results. To
assess such accuracy, a coarse mesh analysis has been used for both the thermal and the
mechanical analyses for case A. The coarse mesh model consists of 14000 nodes associated
with 2880 elements. The element type is DC3D20 and C3D20 in the thermal and mechanical
analyses in ABAQUS, respectively. Also, the element birth technique is adopted in the FEM
coarse model to simulate depositing the filler materials in the weld overlay and girth welding
while moving the heat source. The coarse mesh size is equal to or larger than 1.5 times the
normal mesh size utilized in this study for case A (see Fig. 5) with the coarse mesh model
being composed of 40 circumferential elements instead of 60 elements, as shown in Fig. 16.
-700
-600
-500
-400
-300
-200
-100
100
200
020 40 60 80 100 120 140 160 180 200
Axial stress (MPa)
Axial distance (mm)
Case A-FEA
Case F-FEA
Case A-Exp.
Case F-Exp.
-300
-250
-200
-150
-100
-50
50
100
150
020 40 60 80 100 120 140 160 180 200
Hoop stress (MPa)
Axial distance (mm)
Case A-FEA
Case F-FEA
Case A-Exp.
Case F-Exp.
28
Fig. 16 Coarse 3-D FE model (case A)
In the thermal analysis, Fig. 17 compares the temperature distributions during weld overlay
for the coarse mesh model, denoted as 1.5h, against the normal mesh model, denoted as 1h, at
90°, 180° and 270° central angle. The maximum temperature is achieved at the welding pool
centre of weld overlay which is 1650°C in the coarse mesh model and 1634°C in the normal
mesh model of case A.
Fig. 17 The thermal history of weld overlay centre at 90°, 180° and 270° central angle for coarse and
normal mesh.
Likewise, the temperature fields of the second pass of girth welding also have very similar
distributions around lines of symmetry (WCL) at three locations, 90°, 180° and 270° central
angle, for the coarse and the normal mesh models. The peak temperature for the coarse mesh
is 2085°C whereas the peak temperature for the normal mesh is 2076°C, as shown in Fig. 18.
0
200
400
600
800
1000
1200
1400
1600
1800
050 100 150 200 250
Temperature (°C)
Time (s)
Temp.90°-1h
Temp.180°-1h
Temp.270°-1h
Temp.90°-1.5h
Temp.180°-1.5h
Temp.270°-1.5h
29
Fig. 18 The thermal history of second pass centre of girth welding at 90°, 180° and 270° central angle
for coarse and normal mesh.
One may note that there is a very good correlation in the thermal fields between the coarse
mesh and the normal mesh models. As a result, the residual stress distributions on the inner
and outer surfaces for the coarse mesh model should also be consistent with the results of the
normal mesh model of case A. Fig. 19(a)-(d) plots the residual stress comparisons between
the coarse mesh model and the normal mesh model at 270° central angle.
0
400
800
1200
1600
2000
2400
1100 1150 1200 1250 1300 1350
Temperature (°C)
Time (s)
Temp.90°-1h
Temp.180°-1h
Temp.270°-1h
Temp.90°-1.5h
Temp.180°-1.5h
Temp.270°-1.5h
-100
100
200
300
400
500
600
700
020 40 60 80 100 120 140 160 180 200
Axial stress (MPa)
Axial distance (mm)
Inner-Axial-1h
Inner-Axial-1.5h
-100
100
200
300
400
500
600
700
020 40 60 80 100 120 140 160 180 200
Hoop stress (MPa)
Axial distance (mm)
Inner-Hoop-1h
Inner-Hoop-1.5h
(a)
(b)
30
Fig. 19 Residual stress distributions for coarse and normal mesh models at 270° central angle: (a)
axial stress distributions on the inner surface, (b) hoop stress distributions on the inner surface, (c)
axial stress distributions on the outer surface, and (d) hoop stress distributions on the outer surface.
Consequently, the normal mesh used in cases A, B, C, D, E and F can be considered
appropriate to obtain accurate numerical results thermally and mechanically.
6. Radial shrinkage
Moving the heat source circumferentially to deposit the filler materials is mainly responsible
for the radial shrinkage during lined pipe welding. In fact, the magnitudes of heat input
provided during three welding passes are quite enough for the filler materials to flow through
welding regions. Thus, a series of radial expansions is produced due to uniform high
temperatures through the pipe thickness. After completing the welding process, subsequent
radial contractions take place during solidification and cooling down to room temperature. As
a result, a local inward deformation in the weld zones results in a simple linear bending in
conjunction with compressive stresses over the outer surface, which are balanced by tensile
stresses on the inner surface. Moreover, the magnitude of radial shrinkage is significantly
affected by the magnitude of axial stresses. Radial deformations on the inner surface of the
lined pipe for six cases at 270° central angle with respect to the longitudinal direction starting
from the WCL are plotted in Fig. 20. It is noticeable that the case with a large axial tensile
-600
-500
-400
-300
-200
-100
100
200
020 40 60 80 100 120 140 160 180 200
Axial stress (MPa)
Axial distance (mm)
Outer-Axial-1h
Outer-Axial-1.5h
-350
-250
-150
-50
50
150
020 40 60 80 100 120 140 160 180 200
Hoop stress (MPa)
Axial distance (mm)
Outer-Hoop-1h
Outer-Hoop-1.5h
(c)
(d)
31
stress at the WCL has a large radial shrinkage. In other words, larger residual axial tensile
stresses on the inner surface lead to larger bending moments at the WCL in conjunction with
radial shrinkage.
Fig. 20 Radial shrinkage for six cases on the inner surface at 270° from the WCL.
7. Conclusions
In this study, 3-D FE models have been developed and experimental tests have been
conducted to study the influence of a number of factors on the thermal and structural
response in lined pipe welding. These factors include welding properties (weld overlay and
girth welding materials), geometric parameters (using weld overlay and liner) and welding
process parameters (heat input). In detail, the thermal history and residual stress distributions
have been studied for particular locations on the inner and outer surfaces in comparison with
their experimental counterparts, measured using thermocouples and residual stress gauges.
Based on the results, the following main conclusions can be drawn:
(1) The numerical thermal results are consistent with the experimental results with a
variation of less than 6%. Furthermore, the discrepancies between the thermal results
of reference case A and other parametric cases decrease by heading far away from the
WCL along the axial direction.
(2) The tensile stresses on the inner surface are balanced by the compressive stresses on
the outer surface at the FZ and HAZ to produce local inward deformation through the
pipe cross section. The area at which the weld overlay is fixed with the C-Mn pipe is
affected by high thermal cycles, which in turn lead to higher hoop and axial tensile
residual stresses and possible cracks forming.
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
020 40 60 80 100 120 140 160 180 200
Displacement (mm)
Axial distance (mm)
Case A
Case B
Case C
Case D
Case E
Case F
32
(3) Changing the material type for girth welding from carbon steel to stainless steel leads
to enhanced corrosion resistance and a reduction in the axial and hoop residual
stresses on the inner and outer surfaces at the FZ.
(4) Omitting the weld overlay leads to a significant reduction in the axial and hoop
residual stresses at the FZ on the inner and outer surface but the detrimental effect of
leaving a gap between liner and C-Mn pipe should be taken into account.
(5) Reducing heat input produces lower residual stresses at the FZ and its vicinity on the
inner and outer surface.
(6) The extents of tensile and compressive stresses on the inner and outer surfaces
become significantly narrower by removing the liner.
(7) Increasing the residual axial tensile stress leads to an increase in the radial shrinkage
at the WCL.
(8) Increasing the element size to 1.5 times the normal one used in this work does not
result in a significant change in the results for the thermal conditions and residual
stresses.
Acknowledgments
The authors would like to thank all the lab technicians, and in particular Mr Paul Yates, Mr
Ali Ahmadnia, Mr Keith Withers and Mr Guy Fitch, for their help and technical advice
during the experimental work conducted at Brunel University. The first author also wishes to
gratefully acknowledge the financial support provided by Al-Baath University and the British
Council.
References
[1] Obeid, O., Alfano, G. and Bahai, H., 2014. Analysis of the temperature evolution during lined pipe welding.
In Advanced Materials Research (Vol. 1016, pp. 753-757). Trans Tech Publications. doi:
10.4028/www.scientific.net/AMR.1016.753
[2] Karlsson, R.I. and Josefson, B.L., 1990. Three-dimensional finite element analysis of temperatures and
stresses in a single-pass butt-welded pipe. Journal of pressure vessel technology, 112(1), pp.76-84. doi:
10.1115/1.2928591
[3] Teng, T.L. and Chang, P.H., 1998. Three-dimensional thermomechanical analysis of circumferentially
welded thin-walled pipes. International Journal of Pressure Vessels and Piping, 75(3), pp.237-247. doi:
10.1016/S0308-0161(98)00031-3
33
[4] Deng, D. and Murakawa, H., 2006. Numerical simulation of temperature field and residual stress in multi-
pass welds in stainless steel pipe and comparison with experimental measurements. Computational
materials science, 37(3), pp.269-277. doi: 10.1016/j.commatsci.2005.07.007
[5] Yaghi, A.H., Hyde, T.H., Becker, A.A., Sun, W., Hilson, G., Simandjuntak, S., Flewitt, P.E.J., Pavier, M.J.
and Smith, D.J., 2010. A comparison between measured and modeled residual stresses in a
circumferentially butt-welded P91 steel pipe. Journal of Pressure Vessel Technology, 132(1), p.011206.
doi: 10.1115/1.4000347
[6] Dehaghi, E.M., Moshayedi, H., Sattari-Far, I. and Arezoodar, A.F., 2017. Residual stresses due to cladding,
buttering and dissimilar welding of the main feed water nozzle in a power plant reactor. International
Journal of Pressure Vessels and Piping, 152, pp.56-64. doi: 10.1016/j.ijpvp.2017.05.009
[7] Brickstad, B. and Josefson, B.L., 1998. A parametric study of residual stresses in multi-pass butt-welded
stainless steel pipes. International Journal of Pressure Vessels and Piping, 75(1), pp.11-25. doi:
10.1016/S0308-0161(97)00117-8
[8] Deng, D., Murakawa, H. and Liang, W., 2008. Numerical and experimental investigations on welding
residual stress in multi-pass butt-welded austenitic stainless steel pipe. Computational Materials Science,
42(2), pp.234-244. doi: 10.1016/j.commatsci.2007.07.009
[9] Malik, A.M., Qureshi, E.M., Dar, N.U. and Khan, I., 2008. Analysis of circumferentially arc welded thin-
walled cylinders to investigate the residual stress fields. Thin-Walled Structures, 46(12), pp.1391-1401. doi:
10.1016/j.tws.2008.03.011
[10] Lei Zhao, Jun Liang, Qunpeng Zhong, Chao Yang, Biao Sun, Jinfeng Du, 2014. Numerical simulation on
the effect of welding parameters on welding residual stresses in T92/S30432 dissimilar welded pipe,
Advances in Engineering Software, 68, pp.70-79. doi: 10.1016/j.advengsoft.2013.12.004
[11] Velaga, S.K. and Ravisankar, A., 2017. Finite element based parametric study on the characterization of
weld process moving heat source parameters in austenitic stainless steel. International Journal of Pressure
Vessels and Piping. doi: 10.1016/j.ijpvp.2017.09.001
[12] Obeid, O., Alfano, G. and Bahai, H., 2017. Thermo-Mechanical Analysis of a Single-Pass Weld Overlay
and Girth Welding in Lined Pipe. Journal of Materials Engineering and Performance, 26(8), pp.3861-3876.
doi: 10.1007/s11665-017-2821-5
[13] Ma, N., Li, L., Huang, H., Chang, S. and Murakawa, H., 2015. Residual stresses in laser-arc hybrid welded
butt-joint with different energy ratios. Journal of Materials Processing Technology, 220, pp.36-45. doi:
10.1016/j.jmatprotec.2014.09.024
[14] Focke, E.S., 2007. Reeling of tight fit pipe. TU Delft, Delft University of Technology.
[15] ABAQUS Documentation, release 14. Dassault Systèmes, Providence, RI.
[16] Deng, D., 2009. FEM prediction of welding residual stress and distortion in carbon steel considering phase
transformation effects. Materials & Design, 30(2), pp.359-366. doi: 10.1016/j.matdes.2008.04.052
[17] Yaghi, A.H., Tanner, D.W.J., Hyde, T.H., Becker, A.A. and Sun, W., 2011, May. Abaqus Thermal
Analysis of the Fusion Welding of a P92 Steel Pipe. In SIMULIA Customer Conference (pp. 622-638).
[18] Hempel, N., Bunn, J.R., Nitschke-Pagel, T., Payzant, E.A. and Dilger, K., 2017. Study on the residual
stress relaxation in girth-welded steel pipes under bending load using diffraction methods. Materials
Science and Engineering: A, 688, pp.289-300. doi: 10.1016/j.msea.2017.02.005
34
[19] Nasim, K., Arif, A.F.M., Al-Nassar, Y.N. and Anis, M., 2015. Investigation of residual stress development
in spiral welded pipe. Journal of Materials Processing Technology, 215, pp.225-238. doi:
10.1016/j.jmatprotec.2014.08.009
[20] Xu, J., Chen, J., Duan, Y., Yu, C., Chen, J. and Lu, H., 2017. Comparison of residual stress induced by TIG
and LBW in girth weld of AISI 304 stainless steel pipes. Journal of Materials Processing Technology. doi:
10.1016/j.jmatprotec.2017.05.014
[21] Benson, D., 2014. Tips for Successfully Welding Stainless Steel to Carbon Steel. Welding journal, 93(5),
pp.54-56.
[22] Hemmatzadeh, M., Moshayedi, H. and Sattari-Far, I., 2017. Influence of heat input and radius to pipe
thickness ratio on the residual stresses in circumferential arc welded pipes of API X46 steels. International
Journal of Pressure Vessels and Piping, 150, pp.62-71. doi: 10.1016/j.ijpvp.2017.01.001
[23] Hilberink, A. (2011). Mechanical behaviour of lined pipe (Doctoral dissertation, TU Delft, Delft University
of Technology).