Fusion Engineering and Design 86 (2011) 99–105

Contents lists available at ScienceDirect

Fusion Engineering and Design

journal homepage: www.elsevier.com/locate/fusengdes

A study on port plug distortion caused by narrow gap combined GTAW & SMAW
and Electron Beam Welding
Pankaj Biswas a,∗ , N.R. Mandal a , Parameswaran Vasu b , Shrishail B. Padasalag b
a
Department of Ocean Engineering and Naval Architecture, Indian Institute of Technology, Kharagpur, 721302, W.B., India
b
ITER-India, Institute of Plasma Research, Ahmedabad, India

a r t i c l e i n f o a b s t r a c t

Article history: A study of port plug distortion resulting from narrow gap combined GTAW & SMAW and Electron Beam
Received 8 February 2010 Welding was carried out. Thermomechanical finite element analysis of port plug becomes virtually impos-
Received in revised form 2 August 2010 sible because of the requirement of huge number of nodes and elements. Hence an analysis method
Accepted 17 August 2010
based on the concept of inherent strain was used in this work. The computational time required was
Available online 27 September 2010
about 40–50 min only in a Core (TM) 2 Duo, 2.66 GHz computer with 2 GB RAM, which otherwise was not
possible with other conventional computation techniques. As was expected the overall distortion due to
Key words:
EB welding was found to be less compared to that of narrow gap GTAW & SMAW.
Welding distortion
GTAW & SMAW © 2010 Elsevier B.V. All rights reserved.
Electron Beam Welding
Upper port plug
Finite element analysis
Inherent strain method
Narrow gap welding

1. Introduction the residual stresses were introduced to a three-dimensional struc-

The distortion in a welded structure results from the nonuni-
form expansion and contraction of the weld and surrounding base
material, due to the heating and cooling cycle during welding pro-
cess.
Thermal elastoplastic finite element method can be effectively
used to predict welding residual stress and distortion for small or
medium welded structures [1–5]. However this method is inappli-
cable to simulate the welding distortion for large welded structures
because of the large amount of computational time.
For seeking practical methods to predict welding residual stress
and residual distortion, many numerical simulation procedures
[6–10] were developed in the past decades. Ueda and Yuan [6,7]
studied the characteristic distributions of inherent strains in butt,
T and I joints, and employed them to predict welding residual stress.
Michaleris and DeBiccari [8] developed a two-step numerical
analysis procedure to predict welding induced deformation. In their
method, a two-dimensional thermomechanical welding simulation
was performed to determine the welding residual stresses. Then,
∗ Corresponding author. Tel.: +91 9775550568.
E-mail addresses: panu012@yahoo.co.in (P. Biswas), nrm@naval.iitkgp.ernet.in
(N.R. Mandal), parameswaran.vasu@iter-india.org (P. Vasu),
subhasis.panja@iter-india.org (S.B. Padasalag).
ture for elastic analysis.
Jung and Tsai [9] developed a method named plasticity-based
distortion analysis (PDA) and applied it to investigate the relation-
ship between cumulative plastic strains and angular distortion in
fillet welded T joint. In their research, the relationship between
six cumulative plastic strain components and angular distortion
was examined. They also studied the effect of external restraints
and thermal management technique on the relationship between
cumulative plastic strains and angular distortion using PDA proce-
dure [10].
Experimental studies were carried out to investigate the thermal
character and the effects of arc distance and welding parameters
on the angular distortion in asymmetrical double-sided double arc
welding [11] (ADSAW). An experimental study was carried out to
investigate the characteristics of welding deformation of fillet joints
[12]. Meanwhile, a 3-D thermal elastic plastic finite element model
was developed to simulate welding deformation. The influence of
the flange thickness on welding deformation was investigated. All
these methods lacked into the capability of having large structures.
To ascertain the extent of deformation due to the thermal cycles
caused by welding it calls for solving a complex thermal elastoplas-
tic problem, which is non-linear and involves plastic deformation
of the medium at high temperature varying in both time and space.
Analytical solutions turned out to be inadequate [13]. At the same
time conventional numerical techniques proved to be highly time

0920-3796/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.fusengdes.2010.08.040
100 P. Biswas et al. / Fusion Engineering and Design 86 (2011) 99–105

consuming and thereby prohibitively expensive in real life situa-

tions.
To minimize computational time and cost, axi-symmetry mod-
elling approach [21] is adopted widely for welding simulation,
recent investigations have advocated in favor of three-dimensional
solid modelling without considering the axi-symmetry for better
prediction of distortions and residual stresses [14].
The increasing power and capability of computer systems mean
that direct, detailed computational simulations are feasible to some
extent. However, simplifications seem to be necessary in practice
[15], as noted by Josefson et al. [16] and Lindgren [17], it is impor-
tant to keep computational times within a reasonable industrial
time frame.
In this work a comparative study of port plug distortion that may
result from narrow gap combined GTAW & SMAW and Electron
Beam Welding was carried out. The FE model based on inherent Fig. 1. Detail dimensions of upper port plug.
strain [18] generated due to welding was developed to predict
the welding distortion of full scale upper port plug. The compu- where ˛ is the thermal expansion coefficient; c is the thermal
tational time required was about 40–50 min only in a Core (TM) 2 capacity;  is the density. Here the coefficients are taken as [6]:
Duo, 2.66 GHz computer with 2 GB RAM, which otherwise would K = 8.6 × 10−7 cm3 /J, ˇ = 2.8 × 10−6 cm3 /J.
have been not possible with other conventional computation tech- Longitudinal inherent strain εL and transverse inherent strain
niques. Though EBW is a low distortion welding process but the εT are defined as follows:
results herein strongly indicate that the proposed welding proce-
Vx Vy
dure using narrow gap GTAW & SMAW is also a good candidate for εL = and εT = (5)
Ai Ai
fabrication of port plugs. Apart from the cost factor of EBW, one
should also consider the possibility of any welding defect coming Ai is the zone area where Vx and Vy are distributed. For predicting
into place while carrying out EBW. In such an event it will be a very welding deformation, using different area in a certain region has
expensive and difficult process to carryout such repairs. Whereas if no effect on deformation analysis if the total volume of inherent
any such thing happens using the combined GTAW & SMAW weld- strains and the location of its center do not change [19,21].
ing procedure, the repair techniques are simple and inexpensive. The rate of heat input q was calculated based on welding current,
welding voltage, welding speed and number of weld passes as given
below,
2. Analysis methodology

n
V I i i
The welding process being a transient phenomenon, thermo- q= (J/m) (6)
(ws)i
mechanical finite element analysis of port plug becomes virtually i=1
impossible because of the requirement of huge number of nodes
where q is the heat input per unit length (J/m), n is the number
and elements. Hence an analysis method based on the experimen-
of weld pass,  is the welding efficiency, V is the welding voltage
tally verified concept of inherent strain [20,24] was used in this
(volt), I is the welding current (amp) and ws is the welding speed
work.
(m/s).
In fusion welding processes a structure undergoes locally a large
thermal cycle. The thermal expansion of the weld metal and nearby
3. Welding procedure and parameters
areas is restricted by the surrounding cold metal. This leads to for-
mation of residual plastic strains in the weld metal and nearby
The GTAW with argon shielding is one of the most suitable meth-
area. These plastic strains are referred to as inherent strains. These
ods for welding of various grades of stainless steels. However one
strains are considered to be responsible for causing welding defor-
major limitation of the GTAW is its low deposition rate. Therefore
mations. Once the relation between the welding heat input and the
GTAW in conjunction with SMAW was considered for fabrication
inherent strain distribution is established, the residual stress and
of the port plug. The dimensions of upper port plug are shown in
the deformation can be calculated by elastic analysis using inherent
Fig. 1. The details of joint configuration and edge preparation are
strain as initial strain.
given in Ref. [24].
These strains can be divided as longitudinal inherent strain and
GTAW with argon shielding and DCEN polarity was taken for
transverse inherent strain. Let the volume subjected to longitudinal
root pass. After root pass a finishing weld pass to be carried out
and transverse inherent strains per unit length be expressed as Vx
above the root pass weld deposit. The welding parameters for root
and Vy respectively. It is assumed that Vx and Vy are related to rate
pass and finishing pass are shown in Table 1.
of heat input q as shown in Eqs. (1) and (2) respectively:
Filler runs to be carried out using SMAW. The 1st two filler
Vx = Kq (1) passes to be carried out using 2.5 mm diameter electrode and the
remaining welding to be done using 4 mm diameter electrode. Total
Vy = ˇq (2) number of 22 filler runs was considered. The welding parameter to
be used in different passes of SMAW are given in Table 2.
The range of coefficients K and ˇ are given in Eqs. (3) and (4) respec-
tively: 4. Co-ordinate system
˛
K = (0.255–0.335) (m3 /J) (3) The co-ordinate system adopted for the analysis of upper port
c
plug is shown in Fig. 2. The fixed reference point was represented as
˛ (0, 0, 0), as shown in Fig. 2. The Z-coordinate was considered along
ˇ = (0.255–1.0) (m3 /J) (4)
c the length of upper port plug.
 P. Biswas et al. / Fusion Engineering and Design 86 (2011) 99–105 101

Table 1
Welding parameters for root and finishing pass GTAW.

Welding pass type Current (A) Voltage (V) Welding speed (mm/min) Filler wire diameter (mm)

Root pass 170–180 10.5–11 75 1.6

Finishing pass 160–165 9–10 150 1.6

Table 2
Welding parameters for filler runs using SMAW.

Welding run Electrode diameter (mm) Current (A) Voltage (V) Welding speed (mm/min)

1st and 2nd 2.5 150–170 24–27 160

3rd to 22nd 4 160–190 26–28 150

The distortion analysis of the port plug was carried out for two
cases: (i) without restraint and (ii) with restraint. All the calcu-
lated extent of distortions is essentially deviation from the designed
dimensions.

5. Distortion analysis for combined GTAW & SMAW

The heat input considered in the analysis was based on the weld-
ing parameters given in Tables 1 and 2 respectively. Based on this
heat input, the inherent strains were calculated and imposed on all
the weld seams. Considering these strains as initial strains, elastic
analysis was carried our using ANSYS finite element package. The
analysis was carried out for two cases:

(i) Without the restraining effect of tack welds.

(ii) With restraining effect of tack welds. Fig. 3. Finite element model and discritization of port plug.

5.1. Case I: without the restraining effect of tack welds

For finite element modelling brick elements were taken with
fine meshing in the weld zone and gradual coarsening of meshing
In this case all the four plates were assembled and held in posi-
were implemented away from weld region. In the present scheme
tion by means of tack welds only. However the restraining effect
the total number of elements and total number of nodes were
of tack welds were not considered, i.e. the individual plates were
226,300 and 294,517 respectively. The finite element model and
not restrained in any way from deformation caused due to welding.
discritization views are shown in Fig. 3.
Figs. 5 and 6 represent the distortion patterns in X and Z directions
The zone area Ai (i.e. the summation of Aw , At and As ) where
respectively.
longitudinal inherent strain εL and transverse inherent strain εT are
Fig. 5 shows the distorted shape of the side inclined plates. One
distributed is shown in Fig. 4. These strains are distributed through
can observe inward deformation of the side plates resulting in a
the entire length of the weld line.
concave shape. The deformation is maximum at the front end of
All the calculated extent of distortions is essentially deviation
the port plug as shown in Fig. 5. The individual plates will bend
from the designed dimensions.
inwards maximum by 6.966 and 7.147 mm. Thereby total reduction

Fig. 2. Coordinate system. Fig. 4. Zone area Ai .

102 P. Biswas et al. / Fusion Engineering and Design 86 (2011) 99–105
Fig. 5. Distortion along X-direction without restraint.
in the breadth at the maximum distorted area at the front end along
X-direction will be around 14.133 mm only.
The port plug deformation along its length, i.e. along Z-axis is
shown in Fig. 6. Here it is observed that the top plate will shrink
maximum towards the end of the port plug to a maximum of
7.778 mm. The shrinkage of the top plate at the front end will be
about 4.699 mm. So, the overall shrinkage of the top plate of the
port plug along the length will be about 12.477 mm. The shrinkage
at the bottom plate towards the front end will be around 6.077 mm.
5.2. Case II: with restraining effect of tack welds
In this case 50 mm long root tacks at a spacing of 500 mm, hold-
ing the plates in position, were considered to provide for restraint
against deformation.
The calculated distortion shape in X and Z directions are shown
in Figs. 7 and 8.
One can observe a substantial decrease in distortion compared
to the unrestrained case. However in this case the distortion pattern
of the side inclined plates changed to bulging outwards, i.e. giving
a convex shape as can be seen in Fig. 7. The maximum distortion
occurred at the front end of the port plug. The extent of maximum
deformation was 3.812 and 3.594 mm on the left and right side
plates respectively. Therefore the total reduction in the breadth
towards the front end of the port plug along X-direction was about
7.406 mm.
The overall shrinkage of the port plug along its length (Z-
direction) is shown in Fig. 8. As expected the overall longitudinal
shrinkage reduced substantially both at the top and bottom plates.
The maximum shrinkage of the port plug was about 1.511 mm both
Fig. 6. Distortion along Z-direction without restraint.
Fig. 7. Distortion along X-direction with restraint.
at the front and at the end of the port plug, Therefore the overall
shrinkage of the port plug along its length would be 3.022 mm only.
The analysis considered two extreme cases of restraints, in one
case the plates were considered absolutely free to deform and in
the other case they were restrained from distortion at the tack weld
locations along the length of the port plug. Since in the actual fab-
rication process, tack welding will be done, therefore it is expected
that the extent of deformations as mentioned above in both the
directions will be within the limits of these two cases.
6. Distortion analysis of port plug using EB welding
The analysis was carried out considering Electron Beam Welding
of the port plug with input power of 14.5 kW with a correspond-
ing welding speed of 2 mm/s producing a weld width of 7.5 mm
[22,23]. The material considered was 60 mm thick SS319LN plates
with cooling channels in all the four plates as can be seen in the
figures below. Because of these channels the inherent rigidity of
the plates were somewhat reduced, without which the resulting
distortions would have been even lesser. The zone area Ai in case
of EB welding is shown in Fig. 9.
6.1. Case I: distortion without restraint
In this case it was assumed that all the four plates were assem-
bled and held in position by means of tack welds only. However the
effect of tack welds were not considered, i.e. the individual plates
were not restrained in any way from deformation caused due to
welding.
Fig. 8. Distortion along Z-direction with restraint.
 P. Biswas et al. / Fusion Engineering and Design 86 (2011) 99–105 103

Fig. 12. Distortion along Z-direction without restraint.

Fig. 9. Zone area Ai for EB welding.

ing in a concave shape. However the deformation will be maximum

at the front end of the port plug as shown in Fig. 10. The individual
plates will bend inwards maximum by 4.44 and 4.33 mm. Thereby
total reduction in the breadth at the maximum distorted area along
X-direction will be around 8.77 mm only.
The distortion pattern along the vertical direction i.e. along Y-
axis is shown in Fig. 11. In this one can observe that towards the
end of the port plug, the top plate will tend to lift up to a maximum
of 4.713 mm and the bottom plate at the front end will tend to go
down by a maximum of 3.752 mm.
The deformation of the port plug along its length, i.e. along Z-
axis is shown in Fig. 12. Here it is observed that the top plate will
shrink maximum towards the end of the port plug to a maximum
of 4.93 mm. The shrinkage towards the front end will be about
3.13 mm. So, the overall shrinkage of the port plug along the length
will be about 8.06 mm.

6.2. Case II: distortion with restraint

Fig. 10. Distortion along X-direction without restraint.
In this case it was assumed that all the four plates were assem-
bled and held in position by rigid clamps at three locations along
Figs. 10–12 represent the distortion patterns in X, Y and Z direc-
the length of the port plug. Thereby the nodes at these locations
tions.
were restrained from any deformation. In reality the clamps will
The distortion pattern along the X-direction is shown in Fig. 10.
not allow outward distortion of the plates, however they will still
It depicts the distorted shape of the side inclined plates. One can
remain free to distort inwards.
observe from Fig. 10 that the side plates would bend inwards result-
The calculated distortion shape in X, Y and Z directions are shown
in Figs. 13–15.

Fig. 11. Distortion along Y-direction without restraint. Fig. 13. Distortion along X-direction with restraint.
104 P. Biswas et al. / Fusion Engineering and Design 86 (2011) 99–105
Fig. 14. Distortion along Y-direction with restraint.
Fig. 13 represents the distortion pattern of the side inclined
plates along X-direction. From Fig. 13 one can observe that the
side inclined plates towards the end of the port plug around the
mid-depth will deform inwards resulting in a concave shape. The
maximum distortion would be 1.62 and 1.54 mm. Therefore the
total reduction in the breadth towards the end of the port plug
along X-direction will be about 3.16 mm.
Fig. 14 represents the distortion shape in Y-direction. Here the
extent of distortion of the top and bottom plates of the port plug,
because of the rigid clamps, will be very small, of the order of
0.735 mm only.
The overall shrinkage of the port plug along its length (Z-
direction) is shown in Fig. 15. Here one can observe that maximum
shrinkage of the port plug is 1.96 mm towards the end of the port
plug, whereas the same is 0.635 mm only at the front end of the
port plug. Therefore the overall shrinkage of the port plug along its
length would be 2.595 mm only.
The analysis was carried out considering two extreme clamping
conditions, in one case the plates were considered absolutely free
to deform and in the other case they were assumed to be rigidly
restrained from distortion at three locations along the length of
the port plug. However in actual practice it may not be feasible
to achieve the condition of fully rigid restraints. Hence it can be
concluded that the extent of deformations as mentioned above
in all the three directions will be within the limits of these two
cases.
Fig. 15. Distortion along Z-direction with restraint.
7. Conclusions
The following could be concluded from the above study:
(i) In fabrication of port plug using narrow gap GTAW & SMAW,
the restraining effect of the tack welds led to more than 50%
reduction in the distortion level. Tack welding being part of the
fabrication scheme, the maximum reduction in the breadth at
the front end of the port plug would be between 7.406 and
14.133 mm.
(ii) Similarly the overall shrinkage of the top plate of the port plug
due to narrow gap GTAW & SMAW along the length will be
between 3.022 and 12.477 mm.
(iii) The overall shrinkage of the top plate of the port plug along its
length due to EB welding would be somewhere between 2.595
and 8.06 mm. The maximum deflection of the top plate along
Y-direction will be between 0.735 mm and 4.713 mm. The total
reduction in the breadth of the port plug along X-direction will
be between 3.16 and 8.77 mm.
(iv) Since in actual fabrication with EB welding as well as in case
of narrow gap GTAW & SMAW, the plates will be held together
by means of clamps or adequate tack welds, hence the result-
ing deformations will be rather closer to the lower values as
calculated above.
(v) As was expected the overall distortion due to EB welding was
found to be less compared to that of narrow gap GTAW &
SMAW.
Acknowledgements
The results reported in this paper are from a consultation project
carried out for ITER-India.The authors gratefully acknowledge their
financial and relevant information support to successfully carry out
the project.
References
[1] D. Deng, Y. Luo, H. Serizawa, M. Shibahara, H. Murakawa, Numerical simulation
of residual stress and deformation considering phase transformation effect,
Trans. JWRI 32 (2) (2003) 325–333.
[2] D. Deng, H. Murakawa, Numerical simulation of temperature field and resid-
ual stress in multi-pass welds in stainless steel pipe and comparison with
experimental measurements, Comput. Mater. Sci. 37 (3) (2006) 269–277.
[3] T.L. Teng, et al., Analysis of residual stresses and distortions in T-joint fillet
welds, Press. Vess. Piping 78 (2001) 523–538.
[4] D. Deng, W. Liang, H. Murakawa, Determination of welding deformation in
fillet-welded joint by means of numerical simulation and comparison with
experimental measurements, J. Mater. Process. Technol. 183 (2007) 219–
225.
[5] H. Nishikawa, H. Serizawa, H. Murakawa, Actual application of large-scaled FEM
for analysis of mechanical problems in welding, Quart. J. Jpn. Weld. Soc. 24 (2)
(2006) 168–173.
[6] Y. Ueda, M.G. Yuan, Prediction of residual stresses in butt welded plates using
inherent strains, J. Eng. Mater. Technol. 115 (10) (1993) 417–423.
[7] M.G. Yuan, Y. Ueda, Prediction of residual stresses in welded T and I-Joints using
inherent strains, J. Eng. Mater. Technol. 118 (1996) 229–234.
[8] P. Michaleris, A. Debiccari, Prediction of welding distortion, Weld. J. 76 (6)
(1997) 172–181.
[9] H. Jung, C.L. Tsai, Plasticity-based distortion analysis for fillet welded thin-plate
T-joints, Weld. J. 83 (6) (2004) 177–187.
[10] H. Jung, C.L. Tsai, Fundamental studies on the effect of distortion control planes
on angular distortion in fillet welded T-joints, Weld. J. 83 (7) (2004) 213–
223.
[11] Z. Huajun, et al., Fundamental studies on in-process controlling angular dis-
tortion in asymmetrical double-sided double arc welding, J. Mater. Process.
Technol. 205 (2008) 214–223.
[12] D. Dean, et al., Determination of welding deformation in fillet-welded joint by
means of numerical simulation and comparison with experimental measure-
ments, J. Mater. Process. Technol. 183 (2007) 219–225.
[13] M. Adak, N.R. Mandal, Pseudolinear equivalent constant rigidity concept
for analyzing welding residual deformation, Appl. Math. Model. 33 (2009)
2096–2108.
[14] H. Alberg, Berglund, Daniel, Comparison of an axisymmetric and a three-
dimensional model for welding and stress relief heat treatment, in: Proc. of
 P. Biswas et al. / Fusion Engineering and Design 86 (2011) 99–105
the 8th International Conference on Numerical Methods in Industrial Forming
Processes – NUMIFORM 2004, Columbus, OH, USA, June, 2004.
[15] P. Mollicone, et al., Simple thermo-elastic–plastic models for welding distortion
simulation, J. Mater. Process. Technol. 176 (2006) 77–86.
[16] B.L. Josefson, et al., Welding residual distortions in ring-stiffened pipes, J. Off-
shore Mech. Arctic Eng. (1996) 118.
[17] L.E. Lindgren, Finite element modelling and simulation of welding. Part 1.
Increased complexity, J. Thermal Stress 24 (2001) 141–192.
[18] W. Jian-Hua, L.U. Hao, Prediction of welding deformations by FEM based on
inherent strains, J. Shanghai Jiaotong University 12 (2000) 83s–87s.
[19] E. Friedman, Thermo mechanical analysis of the welding process using the finite
element method, Trans. ASME (1975) 206–213.
105
[20] P. Biswas, N.R. Mandal, A comparative study of three different approaches of
FE analysis for prediction of welding distortion of orthogonally stiffened plate
panels, J. Ship Prod. 25 (4) (2009) 1–7.
[21] P. Biswas, N.R. Mandal, Welding distortion simulation of large stiffened plate
panels, J. Ship Prod. 24 (1) (2008) 50–56.
[22] N. Rykalin, A. Uglov, I. Zuev, A. Kokora, Laser and Electron Beam Material Pro-
cessing, Mir Publishers, Moscow, 1988.
[23] E. Koleva, et al., Calculation of weld parameters and thermal efficiency in elec-
tron beam welding, Vacuum 53 (1999) 67–70.
[24] Pankaj Biswas, N.R. Mandal, Parameswaran Vasu, Shrishail B. Padasalag, Analy-
sis of welding distortion due to narrow gap welding of upper port plug, J. Fusion
Eng. Des. 85 (2010) 780–788.