Journal of Materials Processing Technology 224 (2015) 135142

Contents lists available at ScienceDirect

Journal of Materials Processing Technology

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

Numerical simulation of alloy composition in dissimilar laser welding

M.R. Nekouie Esfahani, J. Coupland, S. Marimuthu
Optical Engineering Group, Wolfson School of Mechanical and Manufacturing, Loughborough University, Loughborough LE11 3TU, Leicestershire, United
Kingdom

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

Article history: A three-dimensional multiphase computational uid dynamic model was developed to investigate the
Received 10 January 2015 meltpool uid dynamics, dilution and alloy composition in laser welding of low carbon steel and stainless
Received in revised form 1 May 2015 steel. Using the developed model, independent predictions on weld properties are made for a range of
Accepted 4 May 2015
laser parameters, and in all cases the results of the numerical model were found to be in close agreement
Available online 12 May 2015
with experimental observations. The investigation revealed that above certain specic point energy the
materials within the melt pool are predominantly homogenous. A minimum meltpool convention is
Keywords:
required in dissimilar laser welding to obtain weld bead properties suitable for industrial applications. The
Alloy concentration
Composition
present model provides a simple yet effective method to predicting the weld bead alloying concentration
Dissimilar and homogeneity encompassing wide range of materials.
Laser welding 2015 Elsevier B.V. All rights reserved.
Modelling
Microstructure

1. Introduction A schematic diagram of a conduction mode laser dissimilar

welding is given in Fig. 1. A continuous laser beam of sufcient
Laser welding has become an important joining process in auto- intensity is incident upon the work piece surface, at a constant
mated manufacturing, and is now extensively used in automotive, velocity (scanning speed). A fraction of the incident energy is
aerospace, energy, electronic and medical industries (Duley, 1999; absorbed by the work piece leading to the formation of a weld
Nekouie Esfahani et al., 2015). The advantages of laser welding pool. As the laser beam passes through the work piece, the melt
include precise energy control, low thermal distortion, narrow pool extends along the scanning direction and solidies soon after
heat affected zones, high welding speed, deep penetration and, the laser beam moves away.
in contrast with electron beam welding, laser welding does not Several numerical and experimental investigations of laser
require a vacuum chamber. Laser welding of dissimilar material dissimilar welding have been reported by various researchers.
is more complicated than welding of similar materials because Rosenthal (1941) rst proposed a mathematical model of a mov-
of the immense difference in elemental composition and thermo- ing heat source under the assumption of quasi-steady state. After
physical properties of metals (Tomashchuk et al., 2010). However, this work various studies have been performed on laser dissimilar
laser welded joints of low carbon and stainless steels are cur- welding simulation to investigate the nature of heat transfer, melt
rently used in power generation industries and more generally pool convection and residual stress distribution. Ranjbarnodeh
are of interest for joining 3D structures, complex assemblies and et al. (2012), developed a three dimensional model using nite ele-
high precision components. Despite the potential of laser dissimilar ment analysis (FEA) to predict the temperature distribution and
welding, uneven alloying concentration in the weld bead can often fusion zone shape, and concluded that in the absence of melt pool
results in reduced weld strength (Sun and Ion, 1995), unaccept- convection, the temperature distribution in weld bead is asym-
able intermetallic phases and crack formation (Anawa and Olabi, metrical with maximum temperature shifted toward low carbon
2008). Consequently, strategies to predict and control the alloy steel. A comprehensive FEA model was developed by Deng et al.
composition and alloy distribution of the weld bead needs to be (2009) to compute residual stresses in a dissimilar metal pipe joint
identied. with considering cladding, buttering, post weld heat treatment and
multi-pass welding using a simplied moving heat source. The
inuences of cladding, buttering and post weld heat treatment on
the nal residual stresses are also determined after welding pro-
Corresponding author. Tel.: +44 0 1509 227647; fax: +44 0 1509 227 615. cess is nished. Youtsos and Katsareas (2005) developed a FEA
E-mail address: S.Marimuthu@lboro.ac.uk (S. Marimuthu). model to predict thermal and residual stress distributions in a

http://dx.doi.org/10.1016/j.jmatprotec.2015.05.005
0924-0136/ 2015 Elsevier B.V. All rights reserved.
136 M.R.N. Esfahani et al. / Journal of Materials Processing Technology 224 (2015) 135142

Table 1
Thermal properties of the substrate.

Nomenclature Stainless steel Low carbon steel

Dynamic viscosity (kg/ms) 0.03 0.00578

Liquid temperature (K) 1770 1723
Solid temperature (K) 1670 1523
Latent heat of fusion (J/kg) 400,000 272,000
Temperature coefcient of 0.00043 0.0005
surface tension (N/m K)

compared with experimental results obtained under similar

parameters.

2. Formulation and grid structure

The CFD analysis was performed to model the heat transfer,

Fig. 1. Schematic diagram of a laser dissimilar welding process. uid ow and material diffusion, using the nite volume based
code, Fluent. The mathematical model used in this work was based
on the Reynolds-averaged NavierStokes (RANS) time-dependent
equations. The governing equations were composed of the conser-
dissimilar joint between A508 and AISI 304L. They used the ele-
vation of mass, conservation of momentum, conservation of energy,
ment birth and death technique to model the addition of ller
transport equation for turbulence and volume fraction equation
metal to the weld pool. Melt pool convention and shape of the
(modied continuity equation) (Fluent, 2009). The volume of uid
fusion zone subjected to Marangoni forces in different directions
(VOF) model was used to account for two different materials (SS
in dissimilar welding of aluminium and steel was investigated
and CS) inside the met pool. The following simplifying assumptions
by (Chung and Wei, 1999; Wei and Chung, 2000). Phanikumar
made:
et al. (2001) investigated the uid dynamics and mixing of molten
metals in laser dissimilar welding of CuNi and reported that,
despite positioning the beam at the weld seam centre, the dis- The welding takes place in conduction mode (no keyhole forma-
similar fusion zone is mostly asymmetrical. Chakraborty (2009) tion) and the free surface of the melt pool changes according to
extended Phanikumars work to develop a three dimensional model the melt pool convection.
to study conduction mode laser dissimilar welding of CuNi apply- Laser gas dynamic parameters such as, shielding gas jet, noz-
ing nite volume based method and later discussed the signicance zle stand-off and nozzle exit diameter are assumed to have
of turbulence. Esfahani et al. (2014) investigated the microstructure insignicant effects on the thermal history and weld bead shape
and service performance of dissimilar joint between low carbon and characteristics.
austenitic stainless steel and showed that the alloying element con- There is no chemical reaction or oxidation in the melt pool.
centration has got a signicant inuence on the microstructure and The variation in mechanical behaviour during the welding pro-
service performance of the weld. Recently Hu et al. (2012) devel- cess has insignicant effect on uid ow dynamics.
oped a model to predict heat and mass transfer in welding stainless
steel and nickel and concluded that the mass transfer is highest dur- Fig. 2 shows the initial mesh employed for the CFD analysis. A
ing the initial stage of weld pool formation and thereafter decreases variable spacing grid system with a ne grid near the heat source
with time. and a course grid away from the heat source has been used. The
In this work, a three-dimensional computational uid dynamic computational domain has a dimension of length 4 mm, width
(CFD) model was developed to study alloying in the fusion zone 8 mm and thickness 1 mm in this way the model consists of 194,300
of a dissimilar laser welding of low carbon steel (CS) and stainless elements. The ambient temperature was set at 300 K. The mate-
steel (SS). Temperature eld, velocity eld and material concentra- rial properties of austenitic stainless steel (Attarha and Sattari-Far,
tion prole were analysed for various parameters. The calculated 2011) and low carbon steel (BritishStandards, 2006) used in the
alloying concentration and dilution in the fusion zone was analysis are presented in Tables 1 and 2.

Fig. 2. Mesh used for the analysis.

 M.R.N. Esfahani et al. / Journal of Materials Processing Technology 224 (2015) 135142 137

Table 2
Temperature dependent thermal properties of the substrate.

Temperature (K) Stainless steel Low carbon steel

3
Specic heat (J/kg K) Density (kg/m ) Conductivity (W/m K) Specic heat (J/kg K) Density (kg/m3 ) Conductivity (W/m K)

273 462 7900 14.6 444 7872 45.9

373 496 7880 15.1 472 7845 44.8
473 512 7830 16.1 503 7816 43.4
573 525 7790 17.9 537 7740 41.4
673 540 7750 18.0 579 7733 38.9
873 577 7660 20.8 692 7669 33.6
1073 604 7560 23.9 837 7578 28.7
1473 676 7370 32.2 860 7440 28.6
1573 692 7320 33.7 863 7380 29.5
1673 696 7320 76.8 863 7324 29.5
1773 700 7320 120 863 7268 29.5

Based on the volume fraction values, appropriate properties are where hc is the heat transfer coefcient, T is the ambient temper-
assigned to each control volume within the domain. The tracking of ature and q(x,y) is the laser heating source given by
the interface between the phases is accomplished by the solution  
of a continuity equation for the volume fraction of the secondary q (x, y) = Px,y 1 rf (8)
phases. For the secondary phase (CS) and primary phase (SS), the
where rf is the reectivity of the material and Px,y is the Gaussian
volume fraction equation has the following form:
heat ux, which is given by
 
1 
Secondary Phase (2 2 ) + (2 2 v) = 0 (1) P 2r
2 t Px,y = exp (9)
rb2  rb2
Primary phase 1 = 1 2 (2)
where P is the total laser power, r is radial position within the beam
where , , t, v are volume fraction, density, time, and uid velocity and rb is the beam diameter.
vector respectively. Subscripts 1 and 2 represent the primary and The uid ow in the weld pool is driven by the combination
secondary phase, respectively. of surface tension, viscous force and buoyancy force (Phanikumar
A single momentum equation is solved throughout the domain, et al., 2001). On the top and bottom surfaces, the shear stress ()
and the resulting velocity eld is shared among the phases such caused by the variation of surface tension due to temperature dif-
that. ference is given by

(v) + (vv) = p + ( ( v)) + g Sw (3) = ST (10)
t T
where p, , and g are static pressure, molecular viscosity, gravi-
where  is surface tension gradient and s T is surface temperature
tational body force. The momentum sink (Sw ) due to the reduced T
gradient. During the computation, the surface tension gradient is
porosity in the mushy zone takes the following form
expressed as a function of the surface temperature. The shear stress
2 given by Eq. (10) is applied to the momentum equations (Eq. (3)).
(1 )
Sw = v Amush (4) A variable time step method using CFL number was utilized near
3 +
the VOF interface to ensure stability and convergence of the com-
where is a small number (0.0001) to avoid division by zero in the putational process. The new time step t was estimated, such that
solid region, a default value of 105 is used as mushy zone constant the CFL equation (Eq. (11)) is satised in all the elements (Ikushima
(Amush ), and the liquid fraction is dened as: and Shibahara, 2014).



1 T > T1
t < 0.5
d
(11)
v
= (T Ts ) / (T1 Ts ) Ts T T1 (5)



0 T < Ts where d is the element size v is the norm of velocity eld in the
element.
with Tl , the temperature of the liquid and Ts the temperature of the The mixing of materials (CS and SS) is primarily due to melt
solid. The value of , ranges between 0 and 1, dening the extent pool convection, which is inuenced by surface tension gradient,
of melting. viscous and buoyancy forces. The weld bead surface topology is
A single energy equation is solved throughout the domain. The predominantly inuenced by the direction of the melt pool move-
energy equation is written in terms of the enthalpy (H). ment (Marimuthu et al., 2013). In the CFD model, the nodes on
the free surfaces (top and bottom of the weld bead) were relo-
(H) cated according to the melt pool velocity and direction (Amara and
+ (vH) = (k T ) (6)
t Fabbro, 2010). Adaptive mesh renement was employed in the CFD
where k is the thermal conductivity and T is the temperature. simulation in order to predict and track the new surface topology at
Heat loss due to convection and radiation is considered over all every time step. A code written as user-dened functions (UDF) in
the surfaces and a Gaussian heat ux (Eq. (9)) was used as the input the C programming language was used in order to apply a heat ux
laser heat source. Heat ux input with heat loss due to convection as a transient boundary condition on the top surface of the weld
and radiation (Mazumder and Steen, 1980) is expressed as (Eq. (9)) and to track the weld bead surface prole (coordinates of
X,Y,Z points of each node in the fusion zone). The UDF was then
k T = q (x, y) hc (T T ) (7) dynamically linked with the Fluent CFD solver.
138 M.R.N. Esfahani et al. / Journal of Materials Processing Technology 224 (2015) 135142

3. Results and discussion

The main objective of this research was to develop a CFD model

that is not only capable of predicting the laser dissimilar weld-
ing thermal cycle and uid ow but also to predict the alloying
concentration and weld bead homogeneity in the fusion zone. To
get a better insight into the mechanisms of uid ow and material
mixing in the weld pool and to evaluate the robustness of the CFD
model, three models were developed, with specic point energies
of 50 J, 17 J, and 10 J. In-line with the experimental congurations,
the laser spot diameter of the beam was maintained at 0.5 mm. To
achieve realistic results, the beam was moved one element length
for each time step. The CFD analysis was performed for 100 time
steps of which 80 time steps is for the welding phase (correlate to
the number of elements along the length of the weld) and the nal
Fig. 4. Comparison of temperature inside and outside the melt pool for various
20 time steps corresponds to the cooling phase (chosen considering
specic point energies.
the computational resources). The time steps for the beam energies
of 50 J, 17 J, and 10 J were 10, 3.3 and 1.4 ms, respectively. This is in
accordance to the element length along the laser beam traverse
direction. thermal properties, and in all cases, the maximum temperature
To highlight the non-linear heat transfer phenomena in dis- exceeds the melting temperature of the substrate.
similar laser welding, the results are rst presented in the form The temperature dependence of the thermal conductivity of the
of temperature contours overlaid on the weld bead surface pro- materials signicantly inuences the position of maximum tem-
le. Part of the absorbed energy is used to generate the weld pool perature in melt pool and away from melt pool (heat affected zone
and part of it is conducted into the solid base metal. Heat con- (HAZ)). Fig. 4 shows the maximum temperature in the melt pool
duction is the major mode of heat transfer at the initial stage of and 1 mm outside the melt pool. As seen from Fig. 4, in all the
welding, and in the intermediate stage, uid convection becomes cases, the maximum temperature was noticed within the low car-
dominant, and signicantly inuences the heat transfer in vertical bon steel, however in the region away from the molten pool (1 mm
direction. away from melt pool), a higher temperature is noticed over the
Fig. 3ac shows the temperature contours and the correspond- stainless steel. This is attributed to the temperature dependent
ing weld bead proles for specic point energy of 50 J, 17 J and thermal conductivity of the materials. At high temperature (above
10 J, respectively. The plots are generated after the temperature and 1200 K) low carbon steel has a relatively low thermal conductivity
weld bead prole reach a quasi-steady state. As can be seen from and heat accumulates over the low carbon steel, hence the max-
Fig. 3a, a high specic point energy of 50 J, results in a higher tem- imum temperature occurs within the region of low carbon steel
perature which is attributed to longer laser interaction time. The melt pool. At low temperature the thermal conductivity of stainless
XY views clearly show the asymmetric heat distribution observed steel is less (Table 2) and the temperature distribution is noticed
with the dissimilar laser welding. This is due to the difference in accordingly.

Fig. 3. Comparison of temperature distribution for specic point energy of: (a) 50 J, (b) 17 J, (c) 10 J.
 M.R.N. Esfahani et al. / Journal of Materials Processing Technology 224 (2015) 135142 139

Fig. 5. Comparison of Melting and solidication distribution proles for specic point energy of: (a) 50 J, (b) 17 J, (c) 10 J.

Observation of similar asymmetric behaviour was noticed in the The hump noticed for the specic point energy of 17 J (Fig. 5)
fusion zone liquid fraction and is presented in Fig. 5 (for the energies is attributed to the difference in magnitude of the velocities in
of 50 J, 17 J and 10 J). The XY views (Fig. 5) represent the liquid frac- stainless steel and low carbon steel. Although the negative surface
tion contour normal to the weld direction. As can be seen from Fig. 5, tension gradient causes an outward ow, a difference in velocity
at low specic point energy, it is predominately the stainless steel magnitudes results in hump formation at the interface of the mate-
which melts; however, with increased beam energy the melting rials. With high specic point energy (50 J) the width of the weld
rate of low carbon steel tends to be higher than stainless steel. This pool increases, which supresses hump and results in a predomi-
is attributed to the relatively low thermal conductivity of stainless nantly uniform weld bead. However, at low specic point energy
steel at initial melting stage, which results in rapid heating and fast (10 J), the melt pool on the surface ow outwards (due to negative
melting at the start of the melt pool formation (before low carbon surface tension gradient) causing a slight depression in the weld
steel reaches the melting temperature). However, higher thermal pool centre The observed trend in surface topology is consistent
conductivity of the stainless steel at high temperature results in with that reported by Mills et al. (1998) and Li et al. (2011).
a higher melting rate at high specic point energy. The increase in It is clear that the uid velocity can signicantly inuence the
thermal conductivity of stainless steel at high temperatures leads to mixing and the homogeneity of the dissimilar weld bead. The extent
an increase in temperature of low carbon steel (compared to stain- of mixing and concentration of the alloying elements depends on
less steel). It is also noted from Fig. 5, that the fusion zone surface the magnitude of surface tension driven uid ow in molten pool.
prole changes with specic point energy. The surface was almost Fig. 7 illustrates the midpoint cross-section (XY) phase eld of the
at for a high specic point energy of 50 J whereas a hump in the weld fusion zone for the specic point energy of 50 J, 17 J and 10 J,
middle of the fusion zone was noticed for a specic point energy of respectively. The phase eld is a good indication of the mixing of
17 J. This can be explained on basis of the melt pool size and ow weld metals inside the weld zone and can be used to estimate the
velocity distribution. alloying concentration and homogeneity of the weld bead (Amara
Fig. 6 shows the melt pool velocity in the midpoint cross-section and Fabbro, 2010). Increase in specic point energy results in higher
(XY) and top surface (XZ). The velocity inside the melt pool was temperature gradient at the melt pool surface. The high tempera-
d
driven by the surface tension gradient of the material, which again ture gradient over the weld pool results in a negative dT (surface
depends upon the surface temperature gradient and not on the tension coefcient changes over temperature) which causes an
maximum temperature. As the specic point energy increases, the intense outward ow by convection (from the centre to the weld
thermal gradient and consequent surface tension force increases periphery) that transports alloying elements from the parent metal
which result in higher uid velocity in the melt pool. The nega- into the molten pool, and consequently results in more homoge-
tive thermal gradient of the surface tension in the melt pool causes nous weld bead. As can be seen from Fig. 7, the two materials have
an outward ow (Fig. 6), which provides efcient transfer of heat undergone extensive mixing for the specic point energy of 50 J and
from the centre to the weld periphery and from the surface to weld 17 J, whereas there was minimal mixing at the low specic point
root. Relatively low uid ow velocities (0.18 m/s) are found at energy of 10 J. Similar trends are noticed in experimental obser-
low specic energy of 10J and considerably higher magnitudes of vation of micro-hardness analysis along the fusion zone, which is
velocities are noticed for a specic energies of 17 J (0.3 m/s) and 50 J shown in Fig. 8.
(0.41 m/s). Also, the maximum velocity was found in the low car- Laser welding experimentation of 1 mm stainless steel and low
bon steel due to high temperature gradients within the low carbon carbon steel were performed at specic point energy of 50 J, 17 J
steel. This increase in melt pool dynamics predominantly with the and 10 J, using a CO2 laser with a 127 mm focal length, 500 W
low carbon steel is the primary reason for an increase in the melting laser power, 2 mm exit diameter, 5 mm stand-off distance and 1 bar
rate of low carbon steel (as can be seen from Fig. 5). Argon shroud gas (Esfahani et al., 2014). The specic point energy
140 M.R.N. Esfahani et al. / Journal of Materials Processing Technology 224 (2015) 135142

Fig. 6. Comparison of top surface velocity vector for specic point energy of: (a) 50 J, (b) 17 J, (c) 10 J.

Fig. 7. Comparison of volume fraction of stainless steel for specic point energy of: (a) 50 J, (b) 17 J, (c) 10 J.

was changed by controlling the scanning speed (Esfahani et al., and should be avoided (Baghjari and Akbari Mousavi, 2013), how-
2014). ever high magnitude of specic point energies can result in higher
As can be seen from Fig. 8, the variation of the hardness within HAZ and residual stress in weld bead. The proposed model can be
the weld bead decreases as the beam energy increases which a useful tool, in estimating the homogeneity of the dissimilar weld
suggests that the alloying element concentration in weld bead is zone, prior to the welding.
inhomogeneous for specic point energies less than 17 J. For spe- Fig. 9 shows the comparison of experimental (left side) and
cic point energies of 50 J and 17 J, the percentage variation of simulated (right side) fusion zone prole of dissimilar joint (cross-
hardness within the fusion zone is close to 5%, whereas for 10 J the section along the centre of the sample) for various specic point
percentage variation is in the range of 1516%. Variation of hard- energies. The contours of the simulated results shows the nor-
ness within the weld zone can undermine the efcacy of the joint malise dilution of the low carbon steel. Dilution is dened as the
 M.R.N. Esfahani et al. / Journal of Materials Processing Technology 224 (2015) 135142 141

Fig. 8. Experimental hardness prole of the weld fusion zone.

Fig. 10. Comparison of percentage dilution of the weld bead.

percentage contribution of low carbon steel in the weld fusion zone

and gives a good indication of the alloying concentration in the weld 4. Conclusions
bead. As observed, the weld beam surface prole varies at different
specic point energies which are of particular interest for service A numerical investigation was performed to predict dilution
performance of the dissimilar joint. The variation of surface topol- and homogeneity of a laser welded joint between low carbon and
ogy with different beam energies is attributed to the convection austenitic stainless steel. The important results obtained in this
uid ow in molten pool which again dependants on magnitude work are summarized below.
of surface tension driven uid ow. In agreement with the experi-
mental bead prole, the CFD model predicts the weld bead dilution The dilution and homogeneity signicantly inuences the weld
(% melting of two materials) and surface condition of various spe- bead properties and the developed model can be used to predict
cic point energies. Fig. 10 shows the comparison of experimental them.
and simulated dilution of low carbon steel. It is calculated from the The melt pool dynamics, surface topology and fusion zone dilu-
ratio of the melt area of low carbon steel to the total weld bead tion (alloy mixture) were signicantly inuenced by the thermal
area from the volume fraction value of each element (Fig. 9). The gradient and surface tension of the weld pool.
marginal difference between experimental and simulation results Increase in laser energy result in increased melt pool convec-
should be attributable to the assumptions of constant reectivity tion within the fusion zone. Unlike similar materials, a minimum
and linear thermal conductivity of stainless steel above melting threshold of melt pool convection is essential to achieve a homo-
temperature in the CFD simulation which is not the case in actual geneous weld bead.
experimentation. Such property details were not available for the A predominantly homogenous microstructure and well mixed
simulation. fusion zone was produced with a specic point energy of greater

Fig. 9. Comparison of experimental (left side) and simulated (right side) weld bead cross section proles for specic point energy of: (a) 50 J, (b) 17 J, (c) 10 J.
142 M.R.N. Esfahani et al. / Journal of Materials Processing Technology 224 (2015) 135142

than 17 J for a 1 mm thick dissimilar joint. Irrespective of the laser Hu, Y., He, X., Yu, G., Ge, Z., Zheng, C., Ning, W., 2012. Heat and mass transfer in laser
energy, a signicantly low surface tension coefcient can under- dissimilar welding of stainless steel and nickel. Appl. Surf. Sci. 258, 59145922.
Ikushima, K., Shibahara, M., 2014. Prediction of residual stresses in multi-pass
mine the weld bead homogeneity in dissimilar laser welding. welded joint using Idealized Explicit FEM accelerated by a GPU. Comput. Mater.
The developed model is also applicable for other fusion welding Sci. 93, 6267.
processes, including electron beam welding or arc welding. Li, L., Eghlio, R., Marimuthu, S., 2011. Laser net shape welding. CIRP Ann.Manuf.
Technol. 60, 223226.
Marimuthu, S., Eghlio, R.M., Pinkerton, A.J., Li, L., 2013. Coupled computational uid
References dynamic and nite element multiphase modeling of laser eeld bead geometry
formation and joint strengths. J. Manuf. Sci. Eng. 135, 011004-011004.
Amara, E.H., Fabbro, R., 2010. Modeling of humps formation during deep-penetration Mazumder, J., Steen, W., 1980. Heat transfer model for CW laser material processing.
laser welding. Appl. Phys. A 101, 111116. J. Appl. Phys. 51, 941947.
Anawa, E.M., Olabi, A.G., 2008. Control of welding residual stress for dissimilar laser Mills, K., Keene, B., Brooks, R., Shirali, A., 1998. Marangoni effects in welding. Philos.
welded materials. J. Mater. Process. Technol. 204, 2233. Trans.R. Soc. Lond. Ser. A Math. Phys. Eng. Sci. 356, 911926.
Attarha, M., Sattari-Far, I., 2011. Study on welding temperature distribution in thin Nekouie Esfahani, M.R., Coupland, J., Marimuthu, S., 2015. Microstructural and
welded plates through experimental measurements and nite element simula- mechanical characterisation of laser-welded high-carbon and stainless steel. Int.
tion. J. Mater. Process. Technol. 211, 688694. J. Adv. Manuf. Technol., 18.
Baghjari, S.H., Akbari Mousavi, S.A.A., 2013. Effects of pulsed Nd:YAG laser welding Phanikumar, G., Chattopadhyay, K., Dutta, P., 2001. Modelling of transport
parameters and subsequent post-weld heat treatment on microstructure and phenomena in laser welding of dissimilar metals. Int. J. Numer. Methods Heat
hardness of AISI 420 stainless steel. Mater. Des. 43, 19. Fluid Flow 11, 156174.
British Standards, 2006. Cold rolled low carbon steel at products for cold forming. Ranjbarnodeh, E., Serajzadeh, S., Kokabi, A., Fischer, A., 2012. Prediction of tempera-
Technical delivery conditions. In: BS EN 10130:2006. British Standards. ture distribution in dissimilar arc welding of stainless steel to carbon steel. Proc.
Chakraborty, N., 2009. The effects of turbulence on molten pool transport dur- Inst. Mech. Eng., Part B: J. Eng. Manuf. 226, 117125.
ing melting and solidication processes in continuous conduction mode laser Rosenthal, D., 1941. Mathematical theory of heat distribution during welding and
welding of coppernickel dissimilar couple. Appl. Therm. Eng. 29, 36183631. cutting. Weld. J. 20, 220234.
Chung, F., Wei, P., 1999. Mass, momentum, and energy transport in a molten pool Sun, Z., Ion, J., 1995. Laser welding of dissimilar metal combinations. J. Mater. Sci.
when welding dissimilar metals. J. Heat Transfer 121, 451461. 30, 42054214.
Deng, D., Ogawa, K., Kiyoshima, S., Yanagida, N., Saito, K., 2009. Prediction of residual Tomashchuk, I., Sallamand, P., Jouvard, J.M., Grevey, D., 2010. The simulation of mor-
stresses in a dissimilar metal welded pipe with considering cladding, buttering phology of dissimilar coppersteel electron beam welds using level set method.
and post weld heat treatment. Comput. Mater. Sci. 47, 398408. Comput. Mater. Sci. 48, 827836.
Duley, W.W., 1999. Laser Welding. Wiley, New York. Wei, P., Chung, F., 2000. Unsteady Marangoni ow in a molten pool when welding
Esfahani, R., Coupland, J., Marimuthu, S., 2014. Microstructure and mechanical dissimilar metals. Metall. Mater. Trans. B 31, 13871403.
properties of a laser welded low carbon-stainless steel joint. J. Mater. Process. Youtsos, A., Katsareas, D.E., 2005. Residual stress prediction in dissimilar metal weld
Technol. 214 (12), 29412948. pipe joints using the nite element method. Mater. Sci. For., Trans. Tech. Publ.
Fluent, 2009. 13.0 Users Guide. ANSYS Inc. 490/491, 5361.