This article was downloaded by: [CERIST]

On: 24 June 2014, At: 02:03

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number:
1072954 Registered office: Mortimer House, 37-41 Mortimer Street,
London W1T 3JH, UK

Numerical Heat Transfer,

Part A: Applications: An
International Journal
of Computation and
Methodology
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/unht20

Numerical Simulation of
Resistance Spot Welding
Process
Jamil A. Khan, Lijun Xu, Yuh-Jin Chao, Kirkland
Broach
Published online: 29 Oct 2010.

To cite this article: Jamil A. Khan, Lijun Xu, Yuh-Jin Chao, Kirkland Broach
(2000) Numerical Simulation of Resistance Spot Welding Process, Numerical
Heat Transfer, Part A: Applications: An International Journal of Computation and
Methodology, 37:5, 425-446, DOI: 10.1080/104077800274145

To link to this article: http://dx.doi.org/10.1080/104077800274145

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all
the information (the “Content”) contained in the publications on our
platform. However, Taylor & Francis, our agents, and our licensors
make no representations or warranties whatsoever as to the accuracy,
completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of
the authors, and are not the views of or endorsed by Taylor & Francis.
The accuracy of the Content should not be relied upon and should be
independently verified with primary sources of information. Taylor and
 Francis shall not be liable for any losses, actions, claims, proceedings,
demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in
relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study
purposes. Any substantial or systematic reproduction, redistribution,
reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access
and use can be found at http://www.tandfonline.com/page/terms-and-
conditions
Downloaded by [CERIST] at 02:03 24 June 2014
 Numerical Heat Transfer, Part A, 37:425 –446, 2000
Copyright Q 2000 Taylor & Francis
1040– 7782 r 00 $12.00 H .00

NUMERICAL SIMULATION OF RESISTANCE

SPOT WELDING PROCESS

Jamil A. Khan, Lijun Xu, Yuh-Jin Chao, and Kirkland Broach

L aboratory for Heat Transfer in Material Processing, Departm ent of
Mechanical Engineering, Univ ersity of South Carolina, Colum bia,
South Carolina 29208, USA

This paper develops a model to predict the nugget development during resistance spot
welding (RSW) of Al-alloys. The model employs a coupled thermal–electrical–mechanical
analysis and accounts for phase change and convective transport in the weld pool. The
contact area and the pressure distribution are determined from a coupled thermal–mechani-
cal model. The model calculates time varying interface pressure. The knowledge of interface
pressure allows for accurate prediction of interfacial heat generation. Temperature-depen-
Downloaded by [CERIST] at 02:03 24 June 2014

dent thermal–electrical –mechanical properties are used. The predicted nugget shape and
size agree well with experimental data. The proposed model can be applied to predict the
effects of the welding parameters and the electrode shapes on the nugget development.

INTRODUCTION
RSW, illustrated in Figure 1, is a joining process. Two metal sheets are
compressed between a pair of water-cooled copper-alloy electrodes with an exter-
nal applied force , and then an electric current is passed through the sheets via the
two electrodes to generate concentrate d heating at the contact surface. The
contact surfaces in the region of current concentration are he ated for a very short
duration by a pulse of low-voltage , high-amperage current. Due to heat generation
at the contact surface and joule he ating, a molten nugget is formed at the
intersection of the two sheets. After the current flow ceases, the electrode force is
maintaine d for a short duration to allow the workpiece to rapidly cool and solidify.
The contact surface area depends on the electrode diameter, applied force ,
temperature , and metal deformation. Thus, a RSW process involves interactions of
thermal, electrical, mechanical, and metallurgical phenomena.
An accurate prediction of the thermal history in the weld metal is a prerequi-
site for a reliable prediction of weld metal microstructure and welding properties.
Since direct and accurate me asurements of temperature profiles in the weld pool
are difficult, it is important to develop reliable thermal models.
A detailed temperature distribution during the RSW process was first deter-
mined numerically by Bentley et al. w1x . Later, Greenwood w2x used a metallo-
graphic technique to determine the temperature distributions at different stages
during the RSW process in mild steel. Rice and Funk w3x reported a simplified

Received 27 May 1999; accepted 9 November 1999.

The authors acknowledge the financial support for this project by NSF r EPSCoR. We are also
grateful to the NASA Langley Rese arch Center for carrying on the RSW.
Address correspondence to Professor Jamil A. Khan, University of South Carolina, Department
of Mechanical Engineering, Columbia, SC 29208 , USA. E-mail: [email protected]

425
 426 J. A. KHAN ET AL.

NOMENC LATURE
c specific heat e radial direction width of first control
g volume fraction, or acceleration due volume
to gravity n kinematic viscosity
h enthalpy f electric potential
j electric current density r density
ª
J electric current density vector s electrical conductivity
k thermal conductivity j radial transformed coordinate
K perme ability z axial transformed coordinate
K0 perme ability constant
p pressure , stress
p0 uniform pressure load at the top of the Subscripts
electrode
p dimensionless contact pressure , p r p 0 cw cooling water
r radial coordinate e electrode
R contact resistance E grid east point
R0 static contact resistance l liquid
Downloaded by [CERIST] at 02:03 24 June 2014

S source term liq liquidus

t time N grid north point
T temperature P grid point
u axial ve locity components s solid
v radial ve locity components S grid south point
ª
V velocity sol solidus
z axial coordinate w workpiece
a heat transfer coefficient W grid west point
b T thermal expansion coefficient ` ambient

Figure 1. Schematic drawing of RSW.

 SIMULATION OF RESISTANCE SPOT WELDING PROCESS 427

one-dimensional he at transfer model using the finite difference method to predict

the temperature history. Gould w4x me asured the nugget growth by using metallo-
graphic techniques and proposed a finite difference based one-dimensional he at
transfe r model. This model demonstrate d the importance of radial he at transfer,
which inhibits the calculation of nugget expansion. This model did not account for
nonuniform current density distribution. The numerical models reported by Nied
w5x and Tsai w6x had specific limitations of the commercial software package.
Two-dimensional axisymmetric he at transfe r models for analyzing the RSW process
were reported by Cho and Cho w7x , Han et al. w8x , and Wei et al. w9, 10x , where the
finite difference method was used. Huh and Kang w11x developed a three-dimen-
sional thermoelectrical model using a finite element code. A model considering
transport of mass, momentum , energy, species, and magnetic field intensity was
reported by Wei and Ho w12x. These models focused on heat transfer response only.
Gupta and De w13x and Feng et al. w14x recently presented a coupled thermal ] elec-
trical ] mechanical model for the RSW process of steel. However, the thermal
contact resistance between the workpiece and electrode was not accounte d for. In
Downloaded by [CERIST] at 02:03 24 June 2014

neither of the models (Gupta et al. w13 x and Feng et al. w14x ) was the convection in
the mushy and melt zones considered, nor was reported the effect of interface
friction coefficient. Browne et al. w15, 16x proposed a he at conduction model for the
RSW process of aluminum, and neglected the electrical contact resistance at the
workpie ce r electrode (W-E ) interface. The W-E interface contact resistance for
Al-alloys can be of the same order of magnitude as that of the faying surface w17 x .
Khan, Xu, and Chao w18x and Xu and Khan w19x developed the coupled
thermal ] electrical ] mechanical models for the Al-alloy RSW process. These mod-
els included the thermal contact resistance of the W-E interface , the effect of
interface friction coefficient w18x , the contact resistance of W-E interface and the
faying surface , and the contact resistance variation with contact pressure and
temperature. Recently, a transient thermal analysis of the electrodes was reported
by Yeung and Thornton w20x to determine the temperature history of the elec-
trodes. However, these models w18 ] 20x consider only he at conduction and do not
account for convection in the molten pool. A re alistic thermal model for analyzing
and simulating the RSW process necessarily should include a thorough he at
transfe r analysis that includes convection in the mushy and molten zone , electrical
field analysis, thermoelastic ] plastic analysis, actual variation of contact resistance ,
phase change , and temperature-dependent material properties.
In the present study a finite element based commercial code is used for the
temperature-dependent solid mechanics modeling; and a control volume based
finite difference code , developed by the authors, is used for the thermal and
electrical field solution. The finite difference code accounts for the convective
terms in the molten weld pool. The present paper integrates the interdependence
of all the important fe atures during the RSW process. Initially the stress field is
obtained from the thermal ] mechanical analysis for the applied electrode force
using the commercial ABAQUSTM 5.8 code. (Details are presented by Khan et al.
w18x .) The specified temperature field at the beginning of the program is the initial
workpiece temperature. The contact areas in the W-E interface and faying surface
are obtained from this stress field. The contact area determines the interfacial
electrical contact resistance responsible for heat generation at the interface. Next,
 428 J. A. KHAN ET AL.

the temperature field is obtaine d from the fully coupled thermal ] electrical analysis
based on the contact area at the W-E interface and the faying surface. The thermal
history from the thermal ] electrical model is used as the input data for the
thermal ] mechanical model analysis. The contact pressure obtaine d from the
thermal ] mechanical model is passed into the fully coupled thermal ] electrical
model. The contact resistance of both W-E interface and the faying surface are
especially treated with artificial interface elements that are used to simulate the
volumetric he at source because of the contact resistance. At the interface the
material properties for thermal ] electrical analysis are artificially imposed for
physically reasonable simulation. The stress field is updated at every 1 r 8 cycle (1
cycle s 1 r 60th of a second) , and the resulting values are input into the thermal
model, whereas the temperature field is updated from the thermal ] electrical
model to the thermal ] mechanical model of the ABAQUSTM code. The overall
solution provides the stress field, the electric potential field, the current density
distribution, and the transient temperature history. The propose d model was
validated with data obtained from experiments performed at NASA, Langley. The
Downloaded by [CERIST] at 02:03 24 June 2014

experimental results were compared with that of the predicted results.

THEOR ETIC AL ANALYSIS

Heat Transfer Analysis

The unsteady he at flow during the RSW process has been modeled as an
axisymmetric problem. The governing equations for transient heat flow for the
workpiece are expressed as follows:
Continuity equation
­ r ª

­ t
q= ? (r V ) s 0 (1 )

Momentum equation

­ r u ª vl ­ p
­ t
q= ? (r Vu s = ? ( r v l= u ) y r
) K
(u y us) y
­ z
q r gw b T
( T y Tsol ) x ( 2 )

­ r v ª vl ­ p
­ t
q= ? (r V v s = ? ( r v l=
) v
)yr
K
( v y vs ) y
­ r
(3 )

According to the Kozeny] Carman equation, the isotropic permeability is defined

as follows:

gl3 q b
Ks K0 (4 )
( 1 y gl ) 2

where K 0 is a constant accounting for the mushy region morphology , which is

taken to be 1.0 = 10 y3 , and b merely is a computational constant introduced to
avoid division by zero, which is selected to be 1.0 = 10 y6 .
 SIMULATION OF RESISTANCE SPOT WELDING PROCESS 429

Energy equation

­ r h ª

­ t
q= ? (r Vh )
s = ? ( G = h ) q = ? w G = ( h s y h ) x y = ? w r ( h 1 y h ) ( V y Vs ) x q S w (5 )

G s
k
cs
Sw s
j w2
s
?
1
cs
j w2
s
s j w2 ?
(
R w ? p ? rw2
e w
q
s
1
w ) (6 )

For the electrode the governing energy equation is states as follows:

­ (r T)
s = ? (G = T ) q S e (7 )
­ t
ke j e2
?
1 j e2
s j e2 ?
(
R e ? p ? re2 1
)
Downloaded by [CERIST] at 02:03 24 June 2014

G s Se s q (8 )
ce s ce s e e s e

Electrical Field Analysis

During the RSW process, the electrical current density distribution is needed
for calculating the he at generation because of the electrical contact resistance and
the Joule he ating. The flow of electrical current is described by conservation of
electrical current:
ª
= ? Js 0 (9 )
ª
By defining an electrical potential, f , is J s y s = f , the equation can be rewritten
as
= ? (s = f ) s 0 (10 )

Mechanical Analysis
The determination of W-E interface contact diameter, faying surface contact
area, and stress distribution are important aspects of the numerical analysis. The
transient strain histories in this coupled thermal ] mechanical problem are solved
simultaneously at every node by utilizing the commercial ABAQUSTM 5.8 code.

Contact R esistance
The contact resistance is the sum of a low resistance metallic contact and a
high resistance film. Savage et al. w21, 22 x experimentally studied static and dynamic
contact resistance of a series of spot welds. Thornton et al. w17, 23 x me asured the
electrical contact resistances of several Al-alloys as a function of the applied
current load with different surface conditions.
 430 J. A. KHAN ET AL.

A relationship for contact resistance R as a function of temperature and con-

tact pressure is proposed by Khan et al. w18x :

I
w 5
í
Rs
max 0 ,
( )
Tsol y T
Tsol y T 0
? ( R 0 y R sol ) q R sol r p n T F Tsol
(11 )

J
w max 0 ,
( )
T liq y T
T liq y Tsol
? ( R sol y R liq ) q R liq r p
5 n
T ) Tsol

where n , R 0 , R sol , R liq are selected to be constants during the simulation of the
RSW process and are taken to be 0.1, 125 m V , 5 m V , and 0 m V , respectively.

Numerical Procedure for the Electrode

Downloaded by [CERIST] at 02:03 24 June 2014

The general governing equation for the electrode is

­ r Q 1 ­ ­ Q ­ ­ Q
­ t
s
r ­ r ( ) ( )
G r
­ r
q
­ z
G
­ z
q S0 (12 )

As mentioned earlier the thermal and electrical fields are numerically solved using
the finite difference method. To accommodate the special geometric shape and the
cone angle of the electrode a coordinate transformatio n is used to obtain a
rectangular mesh in the computational domain. In general, we encounter four
different shapes in the electrode; these shapes and their transformation s are shown
in Figure 2. j and z are the transformed coordinates:

z sz 0 F z F z0
r y a1 ? z y a 3 (13 )
j s 0 F j F r2
a2 z q a4

­ z ­ z ­ j 1 1
s1 s0 s s sM
­ z ­ r ­ r a 2 z q a4 a2 z q a4
­ j y a1 ( r y a1 ? z y a 3 ) ? a 2 (14 )
s y
­ z a2 ? z q a4 ( a2 z q a4 ) 2

s y a1 ? M y a 2 j M s y ( a1 q a 2 j ) M

­ Q ­ Q ­ j ­ Q ­ z ­ Q ­ Q
s q s y ( a1 q a 2 j ) M q (15 )
­ z ­ j ­ z ­ z ­ z ­ j ­ z

­ Q ­ Q ­ j ­ Q ­ z ­ Q
s q sM (16 )
­ r ­ j ­ r ­ z ­ r ­ j
 SIMULATION OF RESISTANCE SPOT WELDING PROCESS 431
Downloaded by [CERIST] at 02:03 24 June 2014

Figure 2. Domain transformation for electrode shape.

­
­ z ( )
G ?
­ Q
­ z
s G M2
­
­ j ( ( a1 q a 2 j ) 2
­ Q
­ j )
­ 2
Q ­ 2
Q
y 2 G ( a1 q a 2 j ) M qG (17 )
­ j ­ z ­ z 2

1 ­
r ­ r ( )G ? r
­ Q
­ r
s
M3
j q ( a1 ? z q a 3 ) M ­ j ( ­
G ?
M
j
q a1 ? z q a 3
­ Q
­ j ) (18 )

­ r Q
­ t
s
G M3
j q ( a1 z q a 3 ) M ­ j
­
( ) M1 j
­ Q
­ j

­
(( ) ) ­ Q
2
a1
qG M2 q a2 j 2
­ j j ­ j
(19 )
­ 2
Q
qG qS
­ z 2

­ 2
Q 1 a1 z q a 3
S s y2 G ( a1 q a 2 j ) M q S0 M1 s q
­ j ­ z M j

The discretization of the transformed equation is obtained by using the control

volume formulation and the central difference approximation. The transforme d
governing equations can be represented in the following general form:
­ r Q
­ t
s A1
1 ­
j ­ j ( )
A2G j
­ Q
­ j
q A3
­
­ j ( A 4G j 2
­ Q
­ j ) ( ) q
­
­ z
A 5G
­ Q
­ z
q S (20 )
 432 J. A. KHAN ET AL.

Where A 1 , A 2 , A 3 , A 4 , A 5 , and S are functions of coordinates j and z . The

source term S represents the terms of electrical contact resistance , joule he ating,
and cross derivatives. The equation can be expressed in the following discrete
form:

aP Q P s aE Q E q aW Q W q aN Q N q aS Q S qb (21 )

where

A1, P ? A2, E ? G E ? j E A 3,P ? A4, E ? G E ? j 2

E G N ? A5, N
aE s q aN s
j P ? (D j ) ? (d j ) (D j ) ? (d j ) (D z ) ? (d z )
A 1 , P ? A 2 ,W ? G W ? j W A 3 , P ? A 4,W ? G W ? j 2
W G S ? A5,S
aW s q aS s (22 )
j P ? (D j ) ? (d j ) (D j ) ? (d j ) (D z ) ? (d z )
Downloaded by [CERIST] at 02:03 24 June 2014

r P r 0
P
a P s a E q aW q a N q a S q b s SP q Q 0
?
D T P
D t

where S P denotes the source term evaluated at the grid point P. Joule he ating
exists throughout the control volume; heat generation because of electrical contact
resistance exists at the W-E interfaces. The cross-derivative term is discretized by
using the central difference approximation.

Boundary Conditions
For the numerical calculations, the electrode is divided into four different
domains; these domains are shown in Figure 3. The boundary conditions for
solving the governing equations for the electrodes and the workpieces are as
follows.

Figure 3. Axisymmetric coordinate system for the RSW

process.
 SIMULATION OF RESISTANCE SPOT WELDING PROCESS 433

Heat transfer boundary conditions

­ T
yke sa `
( T y T` ) (23 )
­ n GH

­ T
yk e sa `
( T y T` ) (24 )
­ r IH

­ T
yke sa (T y Tc w ) (25 )
­ z LK
cw

­ T
yk e sa ( T y Tc w ) (26 )
­ r JK
cw

­ h
yk w sa `
( h y c s T` ) (27 )
­ z GD
Downloaded by [CERIST] at 02:03 24 June 2014

­ T ­ T ­ h ­ h
s s s s0 (28 )
­ z JI ­ r LF ­ r AF ­ r CD

yk e
­ T
­ z FG
sa we ( )hw
cs
y Te (29 )

y
kw ­ h
cs ­ z FG
sa we ( ) hw
cs
y Te (30 )

u F D s u C D s v F D s vC D s 0 (31 )

­ u ­ u ­ v ­ v
s s s s0 (32 )
­ r AF ­ z AC ­ r AF ­ z AC

Electrical fields boundary conditions

­ f ­ f ­ f ­ f ­ f
s s s s s0 (33 )
­ r AF ­ r CD ­ r HI ­ r JK ­ r LF

­ f ­ f ­ f
s s s0 (34 )
­ z BC ­ z GD ­ z LK

­ f
s0 (35 )
­ n GH

­ f ­ f
s s s (36 )
e
­ z GF
w
­ z GF

­ f
s s j0 (37 )
e
­ z JI

f < AB s 0 (38 )
 434 J. A. KHAN ET AL.

Mechanical boundary conditions. For the solution of the temperature-

dependent stress field and the contact surface area an axisymmetric thermal ]
mechanical model is solved using the commercial finite element based ABAQUSTM
5.8 code. A rigid surface is created at the faying surface ABC as shown in Figure 3.
Contact is assumed between the deformable workpiece surface and the rigid
surface contact elements at the faying surface. The structure is assumed to consist
of axisymmetric solid elements. Contact between the electrode and the workpiece
is assumed to be a contact between two deformable surfaces, and these surfaces
are allowed to undergo small sliding. Because of the axisymmetric solid model,
radial displacement is restricted along the centerline AL. Uniform pressure load is
applied at the top JI of the electrode as shown in Figure 3. Other surfaces are
treated as free surfaces. The plastic deformation of the workpiece and the
electrode is accounted for during simulation of the RSW process.

Numerical Technique
Downloaded by [CERIST] at 02:03 24 June 2014

The schematic diagram of the algorithm used for the numerical simulation is
shown in Figure 4. First, the stress field is obtaine d from the thermal ] mechanical
analysis for the applied electrode force using the finite element ABAQUSTM code.
The contact pressure distribution and the area of the contact region for both the
faying surface and W-E interface are obtaine d from the stress field. Then the
temperature field is obtained from the fully coupled thermal ] electrical analysis

Figure 4. Schematic diagram simulation calculation.

 SIMULATION OF RESISTANCE SPOT WELDING PROCESS 435

using the finite difference code developed by the authors. The finite difference
code uses the faying surface and the W-E interface contact areas and the contact
pressure , obtaine d from the finite element code , to determine the contact electrical
resistance in Eq. (11). The faying surface he at generation is based on the contact
area and the resistance at the faying surface. The update d contact pressure from
the thermal ] mechanical model is input into the thermal ] electrical model every
eighth of a cycle. The thermal history from the thermal ] electrical model is then
used as the input data for thermal ] mechanical model analysis. Some examples of
temperature-dependent properties are presented in Figure 5.
For the convection terms, the power law scheme is employed. For velocity
and pressure correction equations, a SIMPLER algorithm w24x is used, and grid
staggering between nodes for the pressure and velocity is employed for numerical
stability.
For convenience of updating data from the finite difference to finite element
model, and vice versa, identical grids are used for both the models. The simulation
default grid for the finite difference method for domain I is 53 = 20; for the
Downloaded by [CERIST] at 02:03 24 June 2014

domain II is 19 = 4; the domain III is 19 = 40; and the domain IV is 10 = 10. Grid
independence tests for both the finite difference and finite element methods were
performed. Grids were refined until converged results were obtained. For the grid
sizes used, the modeling simulation results were determined to be independent of
grid sizes, as shown in Figure 6.

Figure 5. Variation of ( a ) specific he at and ( b )

thermal conductivity of Al-alloy with temperature.
 436 J. A. KHAN ET AL.
Downloaded by [CERIST] at 02:03 24 June 2014

Figure 6. Grid independence validation. ( a ) Default grid calculated nugget; calculated nugget of grid
independence validation.

The convergence criteria of the numerical solution were checked for the
temperature , electrical potential, and velocity iterations. The relative errors be-
tween the two values were always better than 10 y3 for velocity, 10 y5 for tempera-
ture , and 10 y7 for electrical potential.

Validation of the Model

To validate our numerical calculations, actual welding experiments were
performed at the NASA Langley Research Center. Straight electrodes of the
radius nose made of CMW alloys and the workpiece material was 54540 Al-alloy.
The scheduled welding current and the welding pressure varied over time , which
was used for numerical simulation. The experimental nugget and the numerically
predicted nugget are shown in Figures 7a and 7b. The thickness and the width of
the calculated nugget, respectively, are 5.708 = 10 y4 m and 4.9561 = 10 y3 m. The
thickness and width of the experimental nugget are individually 5.675 = 10 y4 m
and 4.8427 = 10 y3 m. The calculated nugget size and shape based on the proposed
 SIMULATION OF RESISTANCE SPOT WELDING PROCESS 437
Downloaded by [CERIST] at 02:03 24 June 2014

Figure 7. Comparison of ( a) calculated and ( b ) experimental nuggets.

thermal model are within 3% of the experimental values. The thermal ] mechanical
model was developed on the basis of Khan et al. w18x and Xu and Khan w19x. In w18 x
and w19x , the mechanical model was validate d by comparing the predicted interface
pressure with the analytical results obtained by Nied w5x and Tsai et al. w6x . The
validation results are not repeated in this paper.

RESULTS AND DISC USSION

Table 1 shows the physical properties of the workpiece and the electrodes
used for the current study. The material properties shown in Table 1 are at a
reference state of 208 C.
Figure 8 shows the force applied by the electrodes versus welding time during
the experimental RSW process. The applied force was stepped up from about 5500
N to about 14,500 N and held constant until the molten pool was completely
solidified. The large applied electrode pressures prevent molten metal from spilling
over. Figure 9 shows the welding schedule of the applied current versus welding
 438 J. A. KHAN ET AL.

Table 1. Electrodes and workpieces properties

Electrodes Workpieces

CMW Alloy 99.85Cu-0.15Z r Al-alloy 5454O 2.7Mg y 0.8Mn y 0.12Cr

R.W.M.A. Alloy Number 1.15000 Thickness (m) 0.0018034
Spherical Radius (m) 0.2032 Solidus Temperature (8 C) 602
Major Radius MH (m) 0.0079375 Liquidus Temperature (8 C ) 646
Tip Distance from 0.0192051 Latent Heat (kJ r kg) 397
Water LF (m)
Yield Strength (Mpa) 386 Yield Strength (Mpa) 117
Elastic Modulus (Gpa) 130 Elastic Modulus (Gpa) 69.6
Thermal Conductivity 367 Thermal Conductivity 134
(W r m ? K) (W r m ? K)
Density (kg r m3 ) 8890 Density (kg r m3 ) 2680
Electrical Resistivity 18.6 Electrical Resistivity 51
(nV ? m) (n V ? m)
Specific He at (J r kg ? K) 385 Specific Heat (J r kg ? K) 900
Downloaded by [CERIST] at 02:03 24 June 2014

time. The applied current peaks at about 55 kA at around the time the weld pool is
formed and the applied force is the maximum. The applied current is gradually
tapered off to allow for gradual solidification of the weld nugget.
Figure 10 shows the dimensionless contact pressures at workpiece interface
versus dimensionless radial distance. Since axial symmetry is assumed, the figure
shows one half of the interface. The contact pressures at the W-E interface are
shown in Figure 11. The dimensionless contact pressure is a ratio of the stress to
the yield strength at 208 C. The dimensionless radial axis is the ratio of the radial
axis to the electrode major radius. It can be noted from these figures that for the
same electrode major radius the contact area increases with the increase of the
applied force. It also can be noted that the W-E interface pressure is more uniform

Figure 8. Applied force ve rsus we lding time.

 SIMULATION OF RESISTANCE SPOT WELDING PROCESS 439
Downloaded by [CERIST] at 02:03 24 June 2014

Figure 9. Welding current versus welding time (weld schedule ).

than the faying surface pressure. It is evident from the interface pressure distribu-
tion that the interface contact resistance is nonuniform. The predicted contact
pressure is used to calculate the interface contact resistance and thus the he at
generation at the interface.
The stress distribution as obtained from the thermal ] mechanical modeling is
shown in Figures 12 and 13. As the workpiece is heated the stress distribution

Figure 10. Faying surface contact pressure with various applied

forces.
 440 J. A. KHAN ET AL.
Downloaded by [CERIST] at 02:03 24 June 2014

Figure 11. W-E interface contact pressure with various applied

forces.

Figure 12. Faying surface contact pressure with we lding cycles.

 SIMULATION OF RESISTANCE SPOT WELDING PROCESS 441
Downloaded by [CERIST] at 02:03 24 June 2014

Figure 13. W-E interface contact pressure with we lding cycles.

starts to even out at the faying surface as shown in Figure 12. Also it can be noted
that at the succeeding time steps the pe ak pressure also decreases, although the
applied forces on the electrodes are almost the same during the first four cycles of
the RSW process. The contact pressure distribution at the W-E interface follows a
similar trend shown in Figure 13. It is the ability of the current model to calculate
time-varying interface pressure (both at the faying surface and the W-E interface )
that makes the model unique. The knowledge of the interface pressure allows us to
predict more accurately the interfacial he at generation.
The electrical current density of the faying surface is shown in Figure 14. The
electrical current density ratio is the ratio of the electrical density to applied
electrical current density on the top of the electrodes. As can be seen from the
figure , the electrical current density distributions are not the same during the
welding cycles. It can be noticed that the peak current density moves away from the
center of the electrode at succeeding cycles. This along with interface pressure will
affect the nugget formation with nonuniform he at generation.
The proposed model is applied to study the effect of the shapes of the
electrodes on the nugget development. Specifically, the effect of the truncation
angle and the face diameter of the electrode are studied. Two different electrode
face diameters, FG as shown in Figure 3, were used, which are 6.25 = 10 y3 m
(D y 1 ) and 9.525 = 10 y3 m (D y 2 ) , respectively. The taper angles (/ HGD as
shown in Figure 3 ) of the electrode face are varied from 308 to 40 8 . The default
taper angle of the electrode face is selected to be 308 , and the default-applie d force
on the top electrode is 8016 N. The materials of both the electrodes and the
workpiece are the same as those of the experiment. The simulation results using
the straight electrodes with the truncated nose are shown in Figures 15 ] 19. Figure
 442 J. A. KHAN ET AL.
Downloaded by [CERIST] at 02:03 24 June 2014

Figure 14. Faying electrical current density with we lding cycles.

15 shows the distribution of the faying surface pressure (squeeze cycle ) as a

function of the taper angle. The taper angle has almost no effect on the pressure.
Similar behavior was observed also for the electrical current density distribution.
Figure 16 shows the nugget size as a function of the electrode taper angle. From
Figures 15 and 16, it is obvious that the change of the electrode taper angles has

Figure 15. Faying surface contact pressure with

different taper angles.
 SIMULATION OF RESISTANCE SPOT WELDING PROCESS 443

Figure 16. Nugget formation at different electrode face taper

angles.

very little effect on the faying surface contact pressure and the nugget growth.
Figure 17 shows the contact pressures at the faying surface for two different
Downloaded by [CERIST] at 02:03 24 June 2014

electrode face diameters at squeeze cycle and the first welding cycle. For the same
applied force a larger face diameter electrode obviously will produce lower pres-
sure at the interface resulting in lower interface electrical resistance. Interface
current density distribution is similarly affected. As a result it is observed that the
contact area of the faying surface dominates the development and size of the
welding nugget as shown in Figure 18. A larger face radius of the electrode
produces a lower local interface heat generation per unit area, which results in a
smaller thickness of the nugget. This conclusion is more evident from Figure 18.
For the same electrical current (45 kA) the larger face diameter electrode (curve 1 )
produces a much thinner nugget after five cycles than the smaller diameter
electrode has produced after two cycles (curve 3 ). Also, for the same cycle the

Figure 17. Faying surface contact pressure with different

electrode face diameter.
 444 J. A. KHAN ET AL.

Figure 18. Nugget formation because of difference

electrode face diameters.
Downloaded by [CERIST] at 02:03 24 June 2014

smaller diameter electrode with 30 kA current (curve 2 ) produces a much larger

nugget than the larger diameter electrode with 45 kA current. Therefore , we can
conclude that a larger welding electrical current is needed for the bigger contact
area. When the same welding electric current is applied to different face diameters
D y 1 and D y 2, the bigger face diameter will create the smaller nugget as shown
by nugget 1 and nugget 3 in Figure 18.
Figure 19 shows the vectors of velocity in the mushy zone and molten zone.
The maximum liquid velocity is about 2 = 10 y5 m r s. Convection effects related to
interactions between phases in the porous mushy zone and the buoyancy force
from the temperature difference are not significant for the welding nugget forma-
tion during the RSW process.

C ONCLUSIONS

(a) The unsteady axisymmetric numerical thermal model that accounts for
convective he at transfer and phase change is proposed to model the RSW
process. The model employs coupled thermal ] electrical ] mechanical
analysis to predict the nugget development in binary Al-alloys. Combining

Figure 19. Vector of ve locity in the molten zone and mushy zone
(current 45 kA, electrode face diameter D-2).
 SIMULATION OF RESISTANCE SPOT WELDING PROCESS 445

a commercial mechanical code based on a finite element method with a

finite difference based thermal code , we have successfully predicted the
transient temperature history, the extent of the heat-affected zone , inter-
face stress and contact resistance , and nugget diameter and thickness.
The calculated simulation results based on the proposed model were
validated with experimental results.
(b) The convection effect related to interactions between phases in the
porous mushy zone and the buoyanc y force from the temperature differ-
ence is determined to be not significant for the weld nugget formation
during the RSW process. The maximum liquid velocity is about 2 = 10 y5
m r s, which is very small to have a significant convection effect. The
model, including Lorentz force , which may enhance convection, is under
development.
(c) The developed model also is used to determine the effect of the electrode
shape and applied pressure on the nugget growth.
(d) The contact area of the faying surface is a dominant factor in the
Downloaded by [CERIST] at 02:03 24 June 2014

development and size of the welding nugget. The contact area is deter-
mined by the face diameter of the truncate d nose electrodes and by the
applied force of the radius nose electrode. A larger welding electrical
current is needed for a bigger contact area of the faying surface to
produce a comparable nugget.

REFERENCES
1. K. P. Bentley, J. A. Greenwood, P. McK. Knowlson, and R. G. Baker, Temperature
Distribution in Spot Welding, British Welding Journal, vol. 12, pp. 613 ] 619, 1961.
2. J. A. Greenwood, Temperature in Spot Welding, British Welding Journal, vol. 6, pp.
316 ] 322, 1963.
3. W. Rice and E. S. Funk, An Analytical Investigation of Temperature Distributions
during Resistance Welding, Welding Journal, vol. 46, no. 4, pp. 175s ] 186s, 1967.
4. J. E. Gould, An Examination of Nugget Development during Spot Welding, Using Both
Experimental and Analytical Techniques, Welding Journal, vol. 66, no. 1, pp. 1s ] 10s ,
1987.
5. H. A. Nied, The Finite Element Modeling of the Resistance Spot Welding Process,
Welding Journal, vol. 63, no. 4, pp. 123s ] 132s , 1984.
6. C. L. Tsai, O. A. Jammal, J. C. Papritan, and D. W. Dickinson, Modeling of Resistance
Spot Weld Nugget Growth, Welding Journal, vol. 71, no. 2, pp. 47s ] 54s, 1992.
7. H. S. Cho and Y. J. Cho, A Study of the Thermal Behavior in Resistance Spot Welding,
Welding Journal, vol. 68, no. 6, pp. 236s ] 244s , 1989.
8. Z. Han, J. Orozco, J. E. Indacoche a, and C. H. Chen, Resistance Spot Welding: A Heat
Transfer Study, Welding Journal, vol. 73, no. 9, pp. 362s ] 371s, 1989.
9. P. S. Wei, S. C. Wang, and M. S. Lin, Transport Phenomena during Resistance Spot
Welding, J. Heat Transfer, vol. 118 , pp. 762 ] 773, 1996.
10. P. S. Wei and F. B. Yeh, Factors Affecting Nugget Growth with Mushy-Zone Phase
Change during Resistance Spot Welding, J. Heat Transfer, vol. 113, pp. 643 ] 649, 1991.
11. H. Huh and W. J. Kang, Electrothermal Analysis of Electric Resistance Spot Welding
Process by a 3-D Finite Element Method, J. Materials Processing Technology, vol. 63, pp.
672 ] 677, 1997.
 446 J. A. KHAN ET AL.

12. P. S. Wei and C. Y. Ho, Axisymmetric Nugget Growth during Resistance Spot Welding,
J. Heat Transfer, vol. 112 , pp. 309 ] 316, 1990.
13. O. P. Gupta and A. De , An Improved Numerical Modeling for Resistance Spot Welding
Process and Its Experimental Verification, J. Manufacturing Science and Engineering,
Transaction of the ASME , vol. 120 , pp. 246 ] 251, 1998.
14. Z. Feng, J. E. Gould, S. S. Babu, M. L. Santella, and B. W. Riemer, An Incrementally
Coupled Electrical-Thermal-Mechanical Model for Resistance Spot Welding, in Trends
in Welding Research: Proc. 5th International Conference on Trends in Welding Research,
pp. 599 ] 644, ASM International, 1998.
15. D. J. Browne , H. W. Chandler, J. T. Evans, and J. Wen, Computer Simulation of
Resistance Spot Welding in Aluminum: Part I, Welding Journal, vol. 74, no. 10, pp.
339s ] 344s , 1995.
16. D. J. Browne , H. W. Chandler, J. T. Evans, P. S. James, J. Wen, and C. J. Newton,
Computer Simulation of Resistance Spot Welding in Aluminum: Part II, Welding
Journal, vol. 74, no. 12, pp. 417s ] 422s , 1995.
17. P. H. Thornton, A. R. Krause , and R. G. Davies, Contact Resistance in Spot Welding,
Welding Journal, vol. 75, no. 12, pp. 402s ] 412s, 1996.
Downloaded by [CERIST] at 02:03 24 June 2014

18. J. A. Khan, L. Xu, and Y. Chao, Prediction of Nugget Development during Resistance
Spot Welding Using a Coupled Thermal-Electrical-Mechanical Model, Science and
Technology of Welding and Joining, vol. 4, no. 4, pp. 201 ] 207, 1999.
19. L. Xu and J. A. Khan, Nugget Growth Model for Aluminum Alloys during Resistance
Spot Welding, Welding Journal, vol. 78, no. 11, pp. 367s ] 372s, 1999.
20. K. S. Yeung and P. H. Thornton, Transient Thermal Analysis of Spot Welding Elec-
trodes, Welding Journal, vol. 78, no. 1, pp. 1s ] 6s, 1999.
21. W. F. Savage , E. F. Nippes, and F. A. Wassel, Dynamic Contact Resistance of Series
Spot Welds, Welding Journal, vol. 57, no. 2, pp. 43s ] 50s , 1978.
22. W. F. Savage , E. F. Nippes, and F. A. Wassel, Static Contact Resistance of Series Spot
Welds, Welding Journal, vol. 56, no. 11, pp. 365s ] 370s, 1977.
23. P. H. Thornton, A. R. Krause , and R. G. Davies, Contact Resistance of Aluminum,
Welding Journal, vol. 76, no. 8, pp. 331s ] 341s , 1997.
24. S. V. Patankar, Num erical Heat Transfer and Fluid Flow, chap. 6, Hemisphere, Washing-
ton, DC, 1980.