See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/315382620

Weld Distortion Estimation in Combined Butt and Fillet weld joint using Finite
Element Analysis

Article  in  Manufacturing Technology · June 2016

CITATIONS READS

0 578

2 authors:

Sanket Vaghela Somasundaram Kumanan

Fiat Chrysler Automobiles, Chennai, India National Institute of Technology Tiruchirappalli
4 PUBLICATIONS   10 CITATIONS    246 PUBLICATIONS   1,483 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

My Ph.D work View project

Non traditional manufaturing View project

All content following this page was uploaded by Sanket Vaghela on 23 October 2017.

The user has requested enhancement of the downloaded file.

Indian Journal of Engineering & Materials Sciences
Vol. 24, June 2017, pp. 201-206

Investigations on effect of direction of welding on distortion in combined butt

and filled joint using finite element analysis
S Kumanan* & S J Vaghela
Department of Production Engineering, National Institute of Technology, Tiruchirappalli 620 015, India

Received 5 October 2015; accepted 9 February 2017

Weld distortion is one of the important quality measures. Because of non-linearity of variables and interaction among them,
it is difficult to predict its behavior. In this study, a three dimensional thermo mechanical finite element analysis is carried out
to understand effect of weld direction; and start and end points on angular distortion of combined butt and fillet weld joint
using gas metal arc welding. Seven different weld sequences are considered to study the effects. The geometry is modeled
using visual mesh and simulated using visual weld. Moving heat source is considered as double ellipsoidal volumetric heat
source available in code. Results obtained from numerical analysis depict the effect of different sequences on angular
distortion.

Keywords: Finite element analysis, Angular distortion, Weld sequence, Weld direction, Gas metal arc welding

Welding process is a multi-physic phenomenon and welding procedure, welding sequence, thickness of
several physical cum chemical changes occur during skin plate and stiffener spacing on distortion were
the process. Due to rapid heating and cooling while analyzed by Deng and Murukawa2. Tsai et al.3 detailed
welding, there is metal expansion and contraction in a comprehensive review on weld distortion, where the
vicinity of weld area. The cooling process is a metal parameters were grouped as design related parameters
shrinkage process happen in three stages such as such as plate thickness, stiffness spacing, number of
volume contraction in liquid stage, volume change attachments, mechanical constraints, assembly
during solidification and solid metal contraction1. The sequence; and process related parameters such as heat
expansion of weld metal creates compressive force on input, welding speed and welding sequence. The
material away from weld region and tension during techniques to control distortion were also discussed in
contraction. The volumetric shrinkage in weld area and the review. Murugan and Gunaraj4 analyzed effect of
uneven forces generated on solidification result in inter pass time, number of weld passes and wire feed
distortion. The types of distortion are longitudinal rate on distortion and established that the inter pass
shrinkage, transverse shrinkage, angular distortion, time and wire feed rate were having inverse
buckling and twisting. The longitudinal shrinkage relationship with angular distortion. Number of passes
occurs parallel to weld line, the transverse shrinkage has a direct effect on distortion. Deng et al.5 studied
occurs perpendicular to weld line and non-uniform about the distortion in fillet weld and concluded that
metal contraction through thickness creates angular thinner flanges are more prone to distortion than
distortion. Controlling and predicting weld distortion is thicker ones and also observed that the effect of
still a challenging task in manufacturing industries. striking the arc and point of starting the welding
Minimization of weld induced distortion is a decision influences the distortion. Murakawa et al.6 studied
to trade off among time, cost and other weld quality effect of root gap and tack weld position on distortion
measures. The effect of welding distortion is studied for butt weld and fillet weld. Manurung et al.7
under the following categories design related numerically and experimentally analyzed distortion of
parameters, process related parameters, material combined butt and fillet weld using Sysweld software.
related, pre-welding parameters and post-welding Keivaniet al.8 in their work, considered nine different
parameters as shown in Fig. 1. The effects of heat input, weld sequences to study the distortion and residual
—————— stresses and the effect of reversing weld direction and
*Corresponding author (email: [email protected]) heat flux deposition over time along neutral axis were
202 INDIAN J. ENG. MATER. SCI., JUNE 2017

Fig. 1 — Factors affecting weld distortion

studied. It is observed that the distortion was minimum

when welding was started from middle and carried
towards end. Iranmanesh et al.9 developed an
algorithm to minimize distortion based on thermal
simulation. It was observed that among other factors
the change in weld sequence affects distortion. Selvan
et al.10 explained out of plane distortion mechanism for
circumferential butt joint of boiler header-nipple and
observed that the distortion in circumferential joint is
different than that of the flat geometries. Welding
direction; and start and stop point was found to have
effect on out of plane distortion. Wen et al.11 studied Fig. 2 — Geometry and displacement measurement line A and B in
and analyzed before and after effects of submerged arc the combined butt and fillet joint weld model
welding process using finite element modeling. Several
attempts have been made to analyze distortion in butt also increases. Among the mitigation techniques,
weld and fillet weld. Limited attempts have been made change in weld sequence could be a least expensive
to analyze distortion in combined butt and fillet option. The effect of different weld sequences in the
joint7,12- 14. combined butt and fillet joint to reduce distortion needs
Several techniques such as controlled preheating, attention. In this research, a finite element model is
heat sinking, thermal stretching, mechanical proposed and analyzed to study the effect of welding
straightening, pre-bending, pre-cambering, post weld sequence on angular distortion for combined butt and
heat treatment have been attempted to control weld fillet joint by gas metal arc welding. The focus is on
distortion in welded structures. Most of the mitigation development of an offline procedure to model and
techniques are costly and requires special attention and simulate the welding process and investigate the effect
skills. As the complexity of geometry increases, the of the sequence and direction on angular weld
complexity of controlling and predicting of distortion distortion.
KUMANAN & VAGHELA: WELD DISTORTION IN COMBINED BUTT AND FILLED JOINT 203

Fig. 3 — Investigated weld sequences for distortion analysis on combined butt and fillet weld joint

Proposed Finite Element Analysis Fig. 2, dimension of flange and web are 200 × 300 × 9
Finite element simulation and analysis were carried mm and 75 × 300 × 9 mm, respectively. Length, width
out in a decoupled approach using exclusive welding and thickness were considered to be in Y, X and Z
simulation software SYSWELD. Thermo- directions, respectively. The weld sequences for
metallurgical, mechanical behavior and hardening laws combined butt and fillet joint are shown in Fig. 3. For
are predefined in material library. The selected sequence 1 to sequence 5 the model consists of total
material type is S355J2G315 an equivalent to mild steel fifteen weld passes, three weld passes at butt joint and
in term of carbon content, which contains 0.2% C, three passes at either side of each of the two webs. In
0.55% Si, 1.6% Mn, 0.035% S and 0.035% P. The sequence 6 and sequence 7, there total 30 numbers of
geometry was created using visual mesh. As shown in weld passes: six passes at butt joint and 6 passes at
204 INDIAN J. ENG. MATER. SCI., JUNE 2017

either side of two webs. The base sequence will be as which are lying between two webs, i.e., weld passes 6,
considered by Manurung et al.7 In the second sequence 10, 14 and 7, 11, 15 as shown in Fig. 3(e). In sequence
as shown in Fig. 3(b) last four fillet weld passes were 6 and 7, the weld passes were divided into two sub-
reversed so the temperature distribution may differ passes. Weld was started from mid plate and carried
from the base sequence. For sequence 3, last eight passes out towards the end point. In sequence 1 to sequence 5,
have been reversed as shown in Fig. 3(c). As shown in the heat flux deposition is symmetrical to one axis,
Fig. 3(d) the sequence 4 can be considered as identical whereas in sequence 6 and 7 the heat flux deposition is
to sequence 2 as the number of passes in opposite symmetrical to two axes. The only difference between
direction are same as that of sequence 2. In this sequence 6 and 7 is that the distance between
sequence instead of last four passes, the passes 8, 9, 10 consecutive weld pass in sequence 7 is higher than that
and 11 were reversed. Sequence 5 was taken by in sequence 6. Modelled seven sequences with suitable
reversing the direction of weld passes of sequence 4, material properties and boundary conditions are then
solved invidually. Fine mesh was created near weld
and heat affected zone areas and coarser mesh away
from weld vicinity to reduce computation time and to
maintain accuracy. The mesh pattern is shown in Fig.
4. Same mesh pattern was used for all welding
conditions. The heat source was modeled using Goldak
double ellipsoidal heat source model suggested by
goldak et al.16, whose value can be calculated by Eqs
(1) and (2).

√
… (1)
√
Fig.4 — Meshpattern and clamping condition in proposed Finite
Element Analysis

Table 1 — Angular distortion 

Position of point relative to Y Left side of flange Right side of flange
direction dz (mm) θ (degree) dz (mm) θ (degree)
Front 2.13 0.65 1.41 0.43
Middle 2.82 0.87 5.31 1.64
End 2.13 0.65 1.53 0.47
2.36 0.73 2.75 0.85

Fig. 5 — Displacement measurement to access the angular distortion in combine butt and fillet joint
KUMANAN & VAGHELA: WELD DISTORTION IN COMBINED BUTT AND FILLED JOINT 205

Table 2 — Comparison of angular distortion of the proposed Table 3 — Experimental validation on angular distortion of flange
model with experimental values reported in literature  and web on combined butt and fillet joint using GMAW
Sequence Position Experiment 3 D analysis Average angular distortion (degree)
Position
(degree) (degree) Experiment Simulation
(Manurung et (Proposed FEA Left side 0.79 0.71
Flange
al.7) model) Right side 0.77 0.68
Sequence 01 Left side of 0.88 0.73 Left side -1.17 -1.22
Web
flange Right side -0.53 -0.48
Right side of 0.92 0.85
flange

Fig. 7 — Vertical displacement along line-Bto access theangular

distortion in combine butt and fillet joint
Where, is density is specific heat, k is thermal
Fig. 6 — Vertical displacement along edge line A to access the
conductivity and Q is internal heat source. The
angular distortion in combine butt and fillet joint specimen is clamped at four points in all direction as
shown in Fig. 4. The clamping condition is common in
√
… (2) all welding conditions.
√
Results and Discussion
Where and are power density in front and rear Displacement was measured along Line-A and
quadrant of heat source. a and b are the width and depth Line-B in Z direction, as shown in Fig. 2. Three points
of heat source respectively. Q is net heat input, which at front, middle and at the end are chosen on each line.
is given by product of arc power (V×I) and arc Angular distortion was calculated as explained by
efficiency (ƞ). and are fractional factor for heat Manurung et al.7 using formula θ=sin-1 (dz/L), as
deposition in front and rear quadrant respectively. shown in Table 1, where dz is displacement in Z
Thermal boundary condition; Heat transfer from direction of point and L is the distance as shown in Fig.
specimen to atmosphere was modeled by Eq. (3). 5. These three angular distortion values are averaged on
the left side and right side; and used to validate the
… (3) proposed FEA model as shown in Table 2. The angular
distortion along Line-A and Line-B is given in Fig. 6
and Fig. 7, respectively. As it can be seen from Fig. 2,
Where, is heat convection transfer coefficient,
Line-A is on left side and Line-B is on right side to the
is thermal emmisivity, is Stefan Boltzmann’s
flanges. From Fig. 6 it can be seen that, the maximum
constant, is body temperature and is surrounding
deflection on Line-A was found 2.94 mm for sequence
temperature. Heat transfer in specimen is given by
2. Minimum deflection was found -0.14 mm for
Fourier’s law of heat transfer by Eq (4).
sequence 6. For sequence 1 to 5 there is a common
trend in distortion, at start and end point it is low, and
at the middle point is high. Distortion for sequence 1,
… (4) 2 and 4 was found almost similar, where in sequence 3
206 INDIAN J. ENG. MATER. SCI., JUNE 2017

and sequence 5 it’s almost half of sequence 1 (2 or 4). simulation model agrees well with experimental
The trend in sequence 6 and sequence 7 is different observations for both flange and web side distortions.
from other sequences. In middle the distortion value is
less compared to start and end point. In overall, Conclusions
distortion for sequence-6 is least on left side. For Line- A finite element method is proposed and used for the
B, from Fig. 7, maximum deflection was found 5.34 analysis of weld distortion. A combined butt and fillet
mm for sequence 2, and minimum was 0.13 mm for weld joint is fabricated using gas metal arc welding to
sequence 6. The trend of deflection was same as that of understand the intricacies of the process and the data
left side for sequence 1 to sequence 5. Distortion of were forwarded for further processing in finite element
sequence 1, 2 and 4 is similar, distortion of sequences modelling. In combined butt and fillet joint for a given
3, 5 and 7 are also similar but less than that of sequence clamping conditions the welding direction and heat
1, 2 and 4. Distortion for sequence 6 is found to be least flux deposition along neutral axis affects the angular
on the right side. On both sides sequence 1, sequence 2 weld distortion of flange and web. In combined butt
and sequence 4 have higher distortion than sequence 3 and fillet joint the distance between consecutive weld
and sequence 5. This can be concluded as the heat flux passes was found to have an effect on weld distortion.
in opposite direction for sequence 1, sequence 2 and The distortion was found less when the distance
sequence 4 is same. For sequence 3 and sequence 5, between consecutive weld pass is more. The proposed
heat flux deposited in reverse direction is more than method could be extended to complex structures
that of sequence 1(sequence 2 or sequence 4). From fabricated with different welding processes with
Figs 6 and 7, it can be inferred that values on right side multiple joints.
are more than that of left side. The joint is
asymmetrical. The difference of maximum values on References
right side and left side was found to be maximum as 1 Gourd L M, Principles of Welding Technology (Edward
2.48 mm and as minimum 1.63 mm. Hence for least Arnold, London), 1995.
2 Deng D & Murakawa H, Comput Mater Sci, 43 (2008) 591-607.
asymmetry, sequence-6 was found best among selected
3 Tsai C L, Park S C & Cheng W T, Weld J, (1999) 156s-165s.
sequences. Different distortion patterns resulted from 4 Murugan VV & Gunaraj V, Suppl Weld J, (2005) 165s - 171s.
different sequences suggest that, there is influence of 5 Deng D, Liang W & Murakawa H, J Mater Process Technol,
welding direction on distortion. For the given set of 183 (2007) 219-225.
sequences, from sequences 1 to sequence 5, the heat flux 6 Murakawa H, Sano M & Wang J, Trans JWRI, 41 (2012) 65-
application was symmetrical about Y axes, while for 70.
sequence 6 and sequence 7 it is symmetrical along X 7 Manurung Y H P, Lidam R M, Rahim M R, Zakaria M Y,
Redza M R, Sulaiman M S, Tham G & Abas S K, Int J Press
axes also. Distortion trend for sequence 6 and sequence Vessel Pipe, 111-112 (2013) 89-98.
7 is similar but the magnitude of sequence 7 is higher 8 Keivani R, Jahazi M, Pham T, Khodabandeh A R & Afshar M
than sequence 6. The large distance between start of R, Int J Adv Manuf Technol, 73 (2014) 409-419.
consecutive weld reduces distortion. For both sides left 9 Iranmanesh M, Azad N & Zabihpoor M, Ships and Offshore
and right, sequence 6 has resulted less distortion. Hence Struct, 9 (2014) 489-497.
it can also be said, symmetrical heat flux deposition 10 Selvan R V, Sathiya P & Ravichandran G, J Manuf Process,
19 (2015) 67-72.
about neutral axes over time results in reduction in
11 Wen S W, Hilton P & Farrugia D C J, J Mater Process Technol,
distortion. 119 (2001) 203-209.
12 Lidam R N, Manurung Y H P, Haruman E, Redza M R, Rahim
Validation of proposed model M R, Sulaiman M S, Zakaria M Y, Tham G, Abas S K, Chau
Two test samples were fabricated as combined butt C Y, Int J Adv Manuf Technol, 69 (2013) 2373-2386.
and fillet joint using gas metal arc welding with 13 Manurung Y H P, Sulaiman M S, Abas S K, Tham G &
established parameters. The base sequence and the best Haruman E, Int J Adv Manuf Technol, 77 (2015) 775-782.
14 Gannon L, Liu Y, Pegg N & Smith M, Mar Struct, 23 (2010)
sequence reported by the proposed FEA model were 385-404.
adopted. The distortion was measured in the flange and 15 SYSWELD Reference Manual, 2012.
on the web using coordinate measuring machine and 16 Goldak J, Chakravarti A & Bibby M, Metall Mater Trans B,
reported in Table 3. It is inferred from the table that the 15 (1984) 299-305.

View publication stats