International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 2, February 2013)

Finite Element Analysis of EBW Welded Joint Using

SYSWELD
PadmaKumari.T1, VenkataSairam.S2
1,2
Department of Mechanical Engineering, JNTUK, Kakinada- 533 003, Andhra Pradesh, India
Abstract— Computational simulations based on finite During welding, localized heating takes place with high
element method are very useful tools in predicting welding thermal gradients as a consequence. As a result of heating,
distortions and residual stresses at the early stage of product the material expands but the expansion is prevented by the
design and welding process development. Therefore process surrounding material and the clamping conditions. This
design changes may be made to compensate for welding
leads to the generation of thermal stresses in the material.
effects before the commencement of the actual fabrication
process. This compensation strategy would save considerable In the area of elevated temperatures, since the yield point is
cost and time over the approach of post welding corrections of lowered, plastic strains are developed in weld region.
distortions and improvement of mechanical properties. Because of high temperatures and subsequent cooling,
However, simulation of welding is such a complex phenomena undesirable residual stresses and deformations are
that it includes mechanical, thermal and metallurgical effects. produced.The residual stresses in the weld region are
The present work involves the finite element analysis of a butt normally tensile in nature and close to the yield stress of
weld plate by using commercial finite element software called the material which leads to brittle fracture, fatigue, and
SYSWELD. Because of the persuasive applications, Inconel stress corrosion cracking.
718 is a considered for investigation. Numerical simulation of
Residual stresses and deformations in welding are
residual stresses and distortions are computed accurately
taking into account the interactions between heat transfer, contradictory in nature. A high degree of clamping
metallurgical transformations and mechanical fields under condition in welding results in small deformations
different clamping conditions and clamp releasing times. (distortions) but high residual stresses where as a low
degree of clamping condition produces low residual
Keywords— FEA, residual stresses, SYSWELD, Heat stresses but large deformations. As a consequence the
source model, EBW Inconel 718 welding reliability of a design is assessed primarily on the
basis of residual stress analysis, whereas feasibility in
I. INTRODUCTION manufacturing is assessed mainly on the basis of
Welding as a fabrication technique finds innumerable deformation analysis. In the present work, the effect of
applications in manufacturing industries having a direct three different clamping conditions on residual stresses and
influence on the life of the components and their thermal distortions is studied.
and mechanical behavior during service. Because of high
temperatures generated during welding and followed by II. LITERATURE SURVEY
cooling of the welded metal, undesirable residual stresses In the literature, finite element analysis (FEA) has been
and deformations are produced in welded components. used widely by many researchers [Bibby et al. (1992),
Therefore, it is of paramount importance to simulate the Goldak et al. (1991), Zhang et al. (2006). Wikander et al.
process of welding to delineate the ensuing residual (1996), Breiguine et al.(1992), Gundersen et al. (1997),
stresses and deformations and predict the behavior of Lindgren L-E et al.(1988)], to perform welding simulations
welded structures. Because of the inherent complexity of and to predict residual stresses in different types of welded
geometry, boundary conditions and nonlinearity of material joints and materials. Prediction is very difficult due to the
properties in welding, not too many studies have been complex variations of temperature, thermal contraction and
conducted in the literature to simulate the welding process. expansion, and variation of material properties with time
However, rapid growth in computer processor technology and space. Furthermore, modeling of the weld process must
and efficient numerical methods have made computational account for the specialized effects of the moving weld arc,
weld mechanics solve an increasing number of problems material deposit, and metallurgical transformations [Goldak
that are of interest to diversified industries. Typical et al.(1985)].
applications include nuclear reactors, automotive,
aerospace structures, defense, micro electronics, etc.

335
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 2, February 2013)
To simplify the welding simulation, it is computationally
efficient to perform the thermal and mechanical analyses
separately. Physically, it is assumed that changes in the
mechanical state do not cause a change in the thermal state.
That is, a change in stress and strain does not cause a
change in temperature. However, a change in the thermal
state causes a change in the mechanical state. Computation
of the temperature history during welding and subsequent
cooling is completed first, and then this temperature field is
applied to the mechanical model to perform the residual
stress analysis. Some of the works where one-way coupling
approach is used are given in references [Andersson et al.
(1978), Marcal et al. (1974), Karlsson et al. (1986), Bibby Fig. 1. Gaussian conical distribution heat source model [Goldak et al.
et al. (1985), Wikander et al.(1996), Lindgren et al.(1988)]. (1989)]
Oddy et al.(1992) state that the heat generated by the
Fig. 1 shows the 3D-Gaussian conical distribution,
plastic deformation is much less than the heat introduced
where the power density deposited region is maximum at
by the weld arc itself. Therefore, the thermal analysis may
the top surface of work piece, and is minimum at the
be performed separately from the mechanical analysis.
bottom surface. Along the work piece thickness, the
In the present work, thermal analysis is carried out
diameter of the power density distribution region is linearly
followed by mechanical analysis and the electron beam
decreased. However, the heat density at the Z-axis (central
welding (EBW) is considered for investigation as it finds
axis) is kept constant. The power density distribution at any
widespread applications diversified industries. EBW offers
plane perpendicular to the Z-axis may be expressed as [Wu
a significant number of advantages compared to other
et al. (2006)]:
fusion welding techniques which include high penetration,
narrow heat affected zone, flexibility to weld a variety of qv(r,z) = qo exp(-3r2/r2o) (1)
materials, less distortion, less energy per unit length and its
vacuum environment. Where q0 is the maximum power density, r0 is the radius
of the heat source at z and r is the radial coordinate of
III. HEAT SOURCE MODELING interior point. The height of the conical Gaussian heat
source is H = ze -zi, the z-coordinates of the top and bottom
Before simulation, the heat source needs to be fitted for surfaces are ze and zi, respectively, and the radii at the top
the welding process. Generally for arc welding processes, and bottom surfaces of the heat source at z = ze and z = zi
double ellipsoid model Goldak et al. (1984) is chosen in are re and ri, respectively. The parameter r0 is decreased
which there are two ellipsoids. In cases where the fusion linearly from the top to the bottom surfaces of the cone
zone differs from an ellipsoidal shape, other models are region and it can be stated as:
used. For the deep penetration electron beam welds, a
conical distribution of power density which has a Gaussian ro(z) = re – (re-ri) (ze-z)/(ze-zi) (2)
distribution in radial direction and a linear distribution in
A. Welding parameters:
axial direction yields more accurate results. Therefore
Gaussian conical distribution is considered for modeling. In this work, two butt plates are joined by using electron
beam welding and the process is simulated for the
prediction of residual stresses and distortions taking the
following parameters into consideration: Welding speed -
1.2 m /min, Voltage - 130KV, Current- 8mA and Energy
per unit length of weld - 55 kJ/mm. The effect of clamping
and the clamping releasing time on welding angular
distortions and residual stresses is investigated.

336
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 2, February 2013)
B. Material used: In the course of welding process, addition of filler
Because of a considerable number of applications, material is modeled using the element birth-and-death
Inconel 718 is considered for analysis. It is a nickel-based technique. During the thermal analysis, deactivated
super alloy that is well suited for applications requiring elements are reactivated sequentially when they come
high strength in cryogenic temperatures. It also exhibits under the influence of the welding torch. For the
good tensile and impact strength. The filler material is subsequent structural analysis, birth of an element takes
assumed to be the same as the parent material. The place at the solidification temperature. Melting and ambient
properties of this material, viz., thermal, metallurgical, and temperatures are set as reference temperatures (temperature
mechanical properties are directly secured from the at which thermal strain is zero) for thermal expansion
SYSWELD material database [SYSWELD manual, 2010]. coefficients of filler and base metals respectively.
For the transient simulation to complete, first thermal
C. Modelling and Meshing: analysis is carried out followed by mechanical analysis. On
Initially, the geometrical model of the welded assembly the basis of the parameters defined in heat input fitting,
is created as shown in Fig.2. Each plate has the dimensions transient analysis is run for a time period which included
of 200mm length x 100mm width x 2.5mm thickness. The the weld run and sufficient time for the plate to cool down
geometrical model is then meshed with the tetrahedral solid to ambient temperature after welding. The analysis has
element. Totally 17625 three dimensional elements are required 118 time increments and taken approximately 2.5
generated for meshing. In order to apply convection on the h to complete. The whole temperature history is saved for
surfaces of the plate with its surroundings, 6880 two the subsequent mechanical analysis. Then the mechanical
dimensional elements are extracted from the solid mesh. analysis is carried out using an isotropic hardening model
Then, since the weld source is to be modelled with the one and small displacement theory. Annealing is assumed to
dimensional elements, totally 80 one dimensional elements occur at 1400 0C. It is the temperature at which the material
are generated from the two dimensional mesh. During melts and the finite element code, SYSWELD, assumes
meshing, special care is taken such that finer mesh is that the material loses its hardening memory, i.e. the effect
provided at the areas of higher concentration such as weld of prior work hardening is removed by setting the plastic
line and heat effected zone and the coarse mesh is provided strain to zero. Due to large nature of the output files
at the remaining parent material. produced by the mechanical transient analysis, it is not
practicable to record every time step. Therefore, only the
last steps of the welding process and the cooling period are
recorded.

IV. RESULTS AND DISCUSSIONS

Three different clamping conditions are analyzed: i) Un-
clamped ii) Clamped but released just after welding and iii)
Clamped but released after cooled down to ambient
Fig 2. Butt joint meshed model
temperature. The output parameters like temperature,
residual stresses, metallurgical transformations, distortions
D. Analysis Procedure: and plastic strains can be obtained from the FE code used
Before simulation, heat input fitting is done and the as a particular point of time. As an example, the
parameters of the heat sources are adjusted in such a way temperature distribution obtained by simulation at the
that the shape obtained is approximately the same as the 98sec and the transverse stress distribution at the end of
shape of the molten zone. A cone-shaped source is fitted welding are shown in Figs. 3 & 4 respectively. Figs. 5 & 6
for the electron beam welding process as discussed in show the variation of transverse and longitudinal residual
section 1.2. To simulate the weld deposit, a simplified stresses along the weld centre line at the top and the bottom
FORTRAN program provided within the SYSWELD code surfaces of the plate for the three clamping conditions
is used to specify the activation and de-activation of the respectively. The analysis of transverse residual stresses at
proper elements. This is accomplished through specifying the top surface show that in the (i) and (ii) cases, the
an offset parameter, in the direction of the weld arc, which distributions of transverse residual stresses are more or less
activates elements as the weld arc approaches. similar, and are tensile along the weld line.

337
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 2, February 2013)
But in the (iii) case, transverse residual stresses are  In the present work, Electron beam welding is
compressive at weld start and stop locations, and they are considered for simulation. The effect of three
tensile at the middle part of the weld line. The results of different clamping conditions on the residual stresses
transverse residual stresses at the bottom surface show that and the distortions are studied. From the simulation
in (i) and (ii) cases, the distribution of transverse residual results, it is concluded that more the clamping the
stresses are similar and they are compressive along the higher the residual stresses and lesser the distortions
weld line. But in the (iii) case, the transverse residual while lesser the clamping lesser the residual stresses
stresses are compressive at the weld start and stop and higher the distortions. .
positions, and they are tensile at the middle part of the weld  Among the three cases considered, the third case
line . The results indicate that in (iii) case, the distributions yields the minimum transverse residual stresses at the
of longitudinal residual stresses at the top and bottom upper surface of the plate while it is reversed with
surfaces are almost same. The above analysis indicates that regard to the bottom surface. Since the plate is
the prevalence of clamping during welding but unclamping unclamped after cooling down to ambient
after cooling down to ambient temperature reduces the temperature, it increases the longitudinal stresses at
distortions significantly as shown in Figs 7 & 8, but it the bottom surface due to the self weight of the plate
increases the transverse residual stresses. as it acts as a fixture to the bottom surface. Hence the
third case is the most preferable condition in regard to
the residual stresses.

Fig. 3 Temperature dist at 98sec during welding

Fig 4. Transverse residual stress distribution

V. CONCLUSION
The following conclusions are drawn from the present
Fig 5 . Transverse stress distribution at the top surface (a) and bottom
work: surface (b)

338
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 2, February 2013)

Fig7.(a). Distortion contours in Y direction Fig7.(b). Distortion contours in Z direction

Fig 8 . Distortion of plate away from weldline at the top surface (a)
and bottom surface (b)

REFERENCES
Fig 6 .Longitudinal stress distribution at the top surface (a) and [1] Goldak J.A., Oddy A., Gu M., Ma W., Mashaie A and Hughes
bottom surface (b) E(1991). Coupling heat transfer, microstructure evolution and
thermal stress analysis in weld mechanics, IUTAM Symposium,
Mechanical Effects of Welding, June 10-14,Lulea Sweden
[2] Rosenthal, D(1946). The theory of moving sources of heat and its
application to metal treatments, Tra ASME, vol. 68, p 849-865

339
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 2, February 2013)
[3] Okerblom, N.O(1958). The calculations of deformations of welded [16] Wu CS, Wang HG, Zhang YM(2006). A new heat source model for
metal structures, London, Her Majesty's Stationary Office, keyhole plasma arc welding in FEM analysis of the temperature
Vinokurov, V. A.(1977) Welding stresses and distortion, The British profile. Weld J 2006;85(12): 284s–91s.
Library Board. [17] Oddy AS, Goldak JA and McDill JMJ(1992). Transformation
[4] Hibbitt H.D., Marcal P.V(1973). A numerical thermo-mechanical plasticity and residual stresses in single-pass repair welds, ASME J
model for the welding and subsequent loading of a fabricated Pressure Vessel Technology, Vol. 114, pp 33-38
structure, Comp. & Struct., Vol. 3, pp 1 145-1 174. [18] Wikander L (1996). Efficient thermo-mechanical modeling of
[5] Friedman, E(1978). Analysis of weld puddle distortion, Welding J welding. Doctoral thesis, Lulea.
Research Suppl. pp 161s- 66s [19] Goldak J., Breiguine V., Dai N., Zhou J(1996). Thermal stress
[6] Andersson, BAB(2001). Thermal stresses in a submerged-arc analysis in welds for hot cracking, AMSE, Pressure Vessels and
welded joint considering phase transformations, J Eng. & Tech. Piping Division PVP, Proceeding of the 1996 ASME PVP Conf.,
Trans. ASME, Vol. 100, 1978, pp 356-362 July 2 1-26, Montreal
[7] Pilipenko A Marcal P(1974). Weld problems, structural mechanics [20] Gundersen (1997). Mathematical modeling of welding. A state of the
programs, Charlottesville, University Press, pp 191-206 art review, SINTEF Material Technology, Trondheim 1997
[8] Karlsson L(1986). Thermal stresses in welding, in R.B. Hetnarski , [21] Lindgren L-E (1992). Deformation and stresses in butt-welding of
Thermal Stresses, Vol. I, Elsevier Science Publishers plates. Doctoral Thesis, Lulea, Sweden.
[9] Karlsson L(1993). Thermo mechanical finite element models for [22] Welding simulation user guide, Sysweld manual (2010), ESI Group.
calculation of residual stresses due to welding, in Hauk et al (eds), [23] Andersson B. and Karlsson L(1981). Thermal stresses in large butt-
Residual Stresses, DGM Informationsgesellschafi Verlag, p 33. welded plates, J. of Thermal Stresses, vol. 4, pp 491-500, 198
[10] Goldak J., Pate1 B., Bibby M(1985). and Moore J Computational [24] Jonsson M., Karlsson L. and Lindgren L.E(1985), Deformations and
weld mechanics, -Structures Materials 61St Panel meeting. stresses in buttwelding of large plates, in R.W.Lewis (ed.),
[11] Goldak J.(1989) Modeling thermal stresses and distortions in welds, Numerical Methods in Heat Transfer, Vol. 111, p 35, Wiley 1985
Proc. of the 2ndInt. Conf. on Welding Research, p 71. [25] Lindgren L.E. and Karlsson L(1988). Deformation and stresses in
[12] Goldak J., Bibby M., Downey D. and Gu M(1990) Heat and fluid welding of shell structures, Int. J. for Numerical Methods in
flow in welds. Advanced Joining Technologies, Proc. of the Int. Engineering, Vol. 25, pp 635-655.
Congress on Joining Research, p 69 [26] Karlsson C.T(1989). Finite element analysis of temperatures and
[13] Smith S.D., A review of weld modeling for the prediction of residual stresses in a singlepass butt-welded pipe influence of mesh density
stresses and distortions due to fusion welding, and material modeling, Eng. Comput., Vol. 6, pp 133-142.
[14] Bibby.(1992) Heat effects of welding: temperature field,heat source [27] Oddy AS, Goldak JA and McDill JMJ(1992). Transformation
modelling residual stress, distortion.Springer. plasticity and residual stresses in single-pass repair welds, ASME J
[15] Goldak, J., Chakravarti, A., and Bibby, M., (1984), "A New Finite Pressure Vessel Technology, Vol. 114, pp 33-38
Element Model for Welding Heat Sources," Metallurgical
Transactions B, 15B, pp. 299-305.20. Ueda Y. and Murakawa H.
Applications of computer and numerical analysis techniques in
welding research, JWRI, Vol. 13, No. 2, pp 165-174

340