Numerical study of electrode geometry effects

on resistance spot welding

P. S. Wei*1 and T. H. Wu2
This work numerically investigates the geometrical effects of the electrode containing a coolant hole
on transport variables during resistance spot welding. The model accounts for transient magneto-
fluid mechanics, heat and species transport, bulk resistance in workpiece, and film and constriction
resistances at contact interfaces. The computed results show that electrode cooling due to the
coolant hole influences transport processes during not only freezing and cooling periods but also
heating and melting periods. Electrode cooling can be interpreted by thermal resistance,
depending on the shapes of the electrode and coolant hole within the growing thermal diffusion
layer. Major factors affecting electrode cooling are different in distinct time stages. In most cases,
enhanced electrode cooling due to the coolant hole decreases the electrode temperature and
nugget growth rate. A decrease in the electrode face radius strongly increases nugget growth rate
and workpiece temperature and decreases electrode temperature, whereas its effect on cooling
rate is insignificant. The optimum design and sensitivity analysis of the electrode shapes subject to
required cooling rate, nugget growth rate and welding time are revealed.
Keywords: Resistance spot welding, Electrode geometry, Electrode cooling, Coolant hole

Introduction Glebov23 modelled and estimated the efficiency of a

coolant hole for a water cooling system. Cooling efficiency
Resistance spot welding is a complicated process was found to be determined by the bevel angle of the
involving coupled electrical–thermal–mechanical–metal- coolant hole and the radius ratio between the inner pipe
lurgical phenomena.1–10 The electrodes apply force to and coolant hole. Cooling effect increased with increasing
clamp the workpieces. They then provide electric current radius ratio and decreasing distance between the outlet of
through the workpieces and post-weld cooling of the the inner pipe and the bottom of the coolant hole.
weld nugget. The characteristics of electrodes are closely Latypova et al.24 proposed a new cooling system for the
related to weld quality. In addition to the large internal coolant hole to alleviate shortcomings such as dead zones
stress exerted by workpieces, high temperatures in and incomplete cooling for the flow of water coolant. The
electrodes result in electrode wear, tip contamination cooling system included two oblique sections on the outlet
and mushrooming, leading to a decline in weld quality of the feed pipe and the heat insulating layer on its external
related to insufficient or irregular nugget formation, surface to produce a more efficient cooling effect, and also
expulsion or spattering, etc., and reduced electrode installation provided by a whirler situated inside the inner
life.11–18 Electrode temperature thus can be used to pipe to enhance the flow and heat transfer rates of the
control the quality of a weld nugget.19 coolant water. The proposed cooling system can reduce the
Electrode temperatures have been measured by Kim temperature by 40uC at the edge of the truncated electrode
and Eagar20 using both high speed cinematography and after welding 40 spots. Lai et al.25 proposed a special
the infrared emission monitoring method. They found sphere shaped electrode cap with an inlet pipe and coolant
a relatively even temperature field in the horizontal hole characterised by the diameter and height of the
direction. The observed isothermal lines within the cooling pipe. It was found that the cooling effects can be
truncated region of the electrode seemed to be different enhanced by proper arrangement of structure parameters.
from predictions.1,4,7,21,22 They also found a critical Rao et al.26 numerically simulated electrode cooling using
electrode face thickness, above which heat conduction in a new cone–fin weld cap design. The new cone–fin design
the electrodes controls the maximum electrode surface not only lessened the stagnation flow near the cap but also
temperature, and below which convective heat transfer increased the cooling area, leading to large reductions in
at the water coolant interface is rate determining. temperatures at the underside of the weld cap. A
roughened fin surface was also suggested to provide
1
nucleation sites that can enhance the heat transfer between
Department of Mechanical and Electro-Mechanical Engineering, National
Sun Yat-Sen University, Kaohsiung, 80424 Taiwan, ROC the cap and the cooling water. Based on heat transfer
2
Department of Mechanical Engineering, Yung Ta Institute of Technology theory, cooling through the coolant hole in different
and Commerce, Pingtung, 909 Taiwan, ROC shapes, however, can be evaluated simply by empirical and
*Corresponding author, email pswei@mail.nsysu.edu.tw relevant heat transfer coefficients. Interactions between the

ß 2013 Institute of Materials, Minerals and Mining

Published by Maney on behalf of the Institute
Received 30 March 2013; accepted 30 May 2013
DOI 10.1179/174329313X13759662488396 Science and Technology of Welding and Joining 2013 VOL 18 NO 8 661
 Wei and Wu Numerical study of electrode geometry effects on resistance spot welding

1 Sketch of resistance spot welding and coordinate system

electrode containing a coolant hole and workpiece during of the axisymmetric axis and bottom electrode–workpiece
resistance spot welding are also ignored in these studies. surface. Workpieces are binary alloy with an equal
The shape of an electrode plays an important role in the dimensional thickness L~/2 and radius r~b . Welding current
transport processes during resistance spot welding. is the AC imposed on the top of the electrode in a radius of
Electrodes with a small face radius increase nugget r~o . The major assumptions31 are as follows:
growth rate and primarily control the nugget width.27 (i) geometry of the electrode is specified by the
Since concentrated electric current density unavoidably dimensional face radius r~e and truncated length
occurs near the face edge, a lower cone angle gives rise to of the electrode L~1 , whereas that of the coolant
stronger electromagnetic stirring fluid flow in the weld hole is determined by the distance between the
nugget28,29 and melting through the surface.28 Li et al.30 tip of the coolant hole and the truncated
found that low cone angles enhance the cooling effect of location of the electrode L~2 , truncated length
the electrode with increased mass and lower electrode L~3 and total length L~4 for maximum radius r~c of
temperature and thus have a better electrode life. Their the coolant hole7,20,32–35
findings were verified by electrode wear experiments. (ii) the mushy zone occurring for alloy solidification
However, low cone angle electrodes produced smaller is a porous medium in local thermal and phase
nuggets and needed a larger welding current to compen- equilibrium; this type of workpiece has success-
sate for the heat dissipated by the electrodes due to their fully been modelled by mixtures in phases, for
excellent cooling capability. example in different solidification processes of
In this work, the cooling effects of electrodes containing castings and resistance spot welding36–38
coolant holes of different shapes on transport processes (iii) heat generations at interfaces due to electrical
during resistance spot welding are extensively investi- contact resistance can be modelled as local
gated. Although functions of electrode cooling have been volumetric heat sources5,27
explored in the past, a quantitative and systematical study (iv) contact resistance is composed of constriction
accounting for transient magneto fluid mechanics and and film resistances; this type of resistance has
heat transfer interacting between electrodes and work- been successfully used to simulate and confirm
pieces and relevant bulk and contact resistances is still the measured trends of dynamic resistance;38
lacking. Extending previous work,28 this study presents contact resistance affected by contamination or
deep insight into the effects of electrode cooling on deformation of the workpiece resulting from
optimising the electrode shape to control weld quality. electrode load can be included in the modelling
of constriction and film resistances
(v) an unsteady electromagnetic model is used;31 an
System model and analysis unsteady electromagnetic field is relevant for
Resistance spot welding, as illustrated in Fig. 1, is in a cylin- resistance spot welding of ferromagnetic materials
drical coordinate system with the origin at the intersection and superalloys with high magnetic permeabilities

Science and Technology of Welding and Joining 2013 VOL 18 NO 8 662

 Wei and Wu Numerical study of electrode geometry effects on resistance spot welding

(vi) thermal, physical and transport properties, of liquid, s and sc are the dimensionless electrical
other than electrical and mechanical properties, conductivity and contact electrical conductivity, m0 and
are chosen as the mean values within the mrl are the free and relative magnetic permeabilities of
temperature range considered; all properties liquid and I represents the welding current. Dimen-
are allowed to be distinct between phases. sionless parameters in equations (1)–(6) are respectively
defined as
Governing equations and boundary ~
nl ~
g Ko
Pr: ~ , Prm : ~l , Da: 2 ,
conditions al al ~
r o
With the above assumptions, the dimensionless con- ~ ~ ~
gb ( T sol  T e ) ~
r 3o T0
tinuity, momentum, energy, species and magnetic field Gr: T ~ , h0 : ~ ~ ,
intensity equations36,37 respectively become n 2l T sol  T e
Lr ~ ~ ~
z+ : (rVm )~0 (1) bs f am,0 I 2 m0 mrl cl rl ~
gl
Lt Fo: ~ ~ , Lo: ~ ~ 2
, C: ~ , S: ~ 2h
,
bT (Tsol {Te ) r a
l l c s r o f
Lrum Pr r(1{gl )2 Lp ~
nl
z+ : (rVm um )~Pr+ : (r+um ){ 3
um { Sc: ~ a (7)
Lt Da gl Lz Dl
Lo LH where Pr, Prm, Da, Gr and Sc are the Prandtl, magnetic
zPr2 Gr½h0 (T{Tsol )zF0 (fla {fl,Tsol
a
){ mH (2)
p2 r Lz Prandtl, Darcy, Grashof and Schmidt numbers, K0 is the
~
permeability constant, ~ nl , ~
al , Dal and ~
gl are the momentum,
Lrvm Pr r(1{gl )2
z+ : (rVm vm )~Pr+ : (r+vm ){ vm { thermal, solutal and magnetic diffusivities of the liquid,
Lt Da gl3 and bT and bs are the thermal and solutal expansion
Lp Lo mr LrH coefficients. An effective contact electrical conductivity in
{ 2 H (3) equation (4) is38
Lr p r Lr
Ef
Lrhm *: sc ~ (8)
z+ (rVm hm )~C+ : (K+hm ) E  Rc Ac
Lt
where Ef and Ac are respectively an effective thickness of
zC+:½K+(hs {hm ){+ : ½r(hl {hm )Vm  the heat source due to thermal contact resistance and the
P "    2 # cross-section through which electric current flows, and E*
Lo 1 1 1 LrH 2 LH
z 2 z z (4) is related to static resistance. Electrical contact resistance
p sc s r Lr Lz for a given control volume containing ni contact spots is
given by
Lrfma Pr Pr
z+ : (rVm fma )~ + : (rfl +fma )z + : ½rfl +(fla {fma ) R1 snHv
Lt Sc Sc Rci ~ ðnHv Þ1=2 zR2 (9)
ni s ni sf
{+ : ½r(f a {f a )Vm 
l m (5)
where Hv isPhardness and the total number of contact
  spots is n~ ni . In this case, the number of contact spots
Lmr H 1
z+ : (mr HVm )~Prm + : +H i
Lt s at the first control volume is chosen to be n1510. s is the
    dimensionless film thickness, and sf : ~ sf = ~
sliq is the
H L 1 H m HVm
zPrm { 2 z r (6) dimensionless electrical conductivity of the film.
r Lr s sr r Dimensionless parameters governing constriction and
where transport variables Vm, um, vm, hm, fma and H are film resistances in equation (9) are respectively defined as
respectively dimensionless velocity vector, velocity com- !1=2
~ ~
ponents in axial and radial directions, enthalpy, solute 1 pH v0 R2 : ~
roH v0
Rl : ~ , ~ (10)
content and magnetic field intensity. Subscripts m, l and s 2R0 ~sliq W W R0 ~sliq
denote mixture, liquid and solid, whereas superscript a
represents solute. All the lengths in equations (1)–(6) are ~ ~
where W is the electrode force and Hv0 and R0 are the
non-dimensionalised by the outer radius of the electrode hardness and electrical contact resistance at the ambient
~ro , whereas time is non-dimensionalised by the time for temperature. In view of the temperature dependent
thermal diffusion ~r 2o = ~al . Dimensionless mixture enthalpy electrical properties and hardness, the electrical contact
~ ~
hm : hm =hf , solid enthalpy hs ~RT: ~csT0 T=hf , liquid resistance model proposed by equations (9) and (10)
enthalpy hl ~RCTzR(1{C)Te z1, mixture thermal involves transient coupling between thermal–mechanical
conductivity K:gl zgs ks , solid thermal conductivity ks effects. The solute fraction in the liquid in equation (5) is
~ ~ ~ ~a given by
: ks = kl , solute content fma : fma = fm,0 , solute content at the
a
solidus temperature fl,Tsol and magnetic field intensity fma
~ fla ~ (11)
H~Hp ~ ro =I, where hf and ~ cs are the dimensional latent 1zfs (kp {1)
~
heat at the eutectic point and the specific heat of solid, T0 , where fla and fs are respectively the solute mass fraction in
~ ~
Tsol and Te are the dimensional initial, solidus and eutectic the liquid and solid fraction and kp is the equilibrium
~ ~a
temperatures, fma and fm,0 are the solute content and partition coefficient. Determinations of the distinct
initial solute content in mixture, gl is the volume fraction regions of the mushy zone, full liquid and solid have

Science and Technology of Welding and Joining 2013 VOL 18 NO 8 663

 Wei and Wu Numerical study of electrode geometry effects on resistance spot welding

been presented by Wei and Yeh.37 Energy and magnetic  

1 LTE rc LTE LTE rc
field intensity equations in the electrodes are respectively cos tan{1 { M z sin tan{1
" N Lj L3 Lj Lf L3
  2
LTE kE ks RC L2 TE 1 2 L TE
~ z zM ~Biw (T E {1) (18)
Lt rE RE Lf2 N2 Lj2
  #
L2 TE 1 dN 1 LTE The surface of the electrode: 0ƒfƒL1 , j51
z2M z {2M
LjLf jNzrc df N Lj 1 LTE 1{re
{ cos tan{1 z
P "    # N Lj L1
Lo LHE LHE 2 HE 1 LHE 2  
z 2 M z z z LT E LT E 1{re
p rE RE sE Lj Lf jNzrc N Lj M z sin tan{1 ~BioEs (T E {1) (19)
Lj Lf L1
(12)

  2 The electrode/workpiece interface: z50, L, 0#r#re

LHE L2 HE 1 2 L HE L2 HE
Lt
~Prm gE ½ Lf2
z
N2
zM
Lj2
z2M
LjLf {K
LT
Lz
~ks BiE (T{TE ) (20)
 
1 dN 1 LHE HE
z
jNzrc
{2M
df N Lj
{
(jNzrc )2  (13)
which reflects discontinuity of temperatures across the
interface. Biot numbers are defined as
~ ~ ~ ~ ~ ~ ~ ~ ~ ~
where dimensionless electrode temperature TE :TE =T0 , hw ro hoET ro hoEs ro hE ro
spatial coordinates j;[r2rc(z)]/N, f;z2L, M;2(drc/ Biw : ~ , BioET : ~ , BioEs : ~ , BiE : ~
kE kE kE ks
dfzjdN/df)/N, N;rs(f)2rc(f), dimensionless thermal
~ ~ (21)
and electrical conductivities kE : kE = ks and sE : ~
sE = ~
sl
~ ~ ~ ~ ~
~
and dimensionless enthalpy RE ~ cET0 =hf . Boundary where hw , hoET , hoEs and hE are respectively heat
conditions associated with convection are that no slip transfer coefficients on the surface of the coolant hole,
condition is applied at the electrode/workpiece interface, the top and side boundaries of the electrode and the
whereas shear stress is free at the interface between the electrode face. Conservation equations (1)–(6), (12) and
workpiece and atmosphere. The boundary condition at (13) are each discretised by the control volume and
the top of the upper electrode is include an implicit finite difference scheme with stag-
gered grids. A grid system of 50653 in workpieces and
1 j2 (1{rc )z2rc j 27672 in the electrode were selected. The dimensionless
HE ~{ sin 2pvt (14)
2(1zrc ) j(1{rc )zrc uniform time step was 261025. The convergence
~ 2 =~ tolerances for global energy and velocity components
where dimensionless current frequency v:vr o al . and enthalpy, concentration and magnetic intensity
Equation (14) is obtained by assuming uniform current fields were 1022 and 1023 respectively. Otherwise, the
across the top surface. Magnetic field intensity in the equations for the electrode were solved again, and the
coolant hole is considered to be negligible. Tangential process was repeated. If solutions converged, tempera-
electric and magnetic field intensities across the interface ture distributions were used to calculate resistances.
between the electrode and workpiece are continuous. Computations then proceeded to the next time.
Magnetic field intensities at surfaces between workpieces
and the atmosphere are satisfied by negligible electric
current in normal directions. Magnetic field intensity Results and discussion
vanishes at the axisymmetric axis. Electrode cooling is This study investigates the cooling effects of electrodes
governed by heat transfer coefficients or Biot numbers at containing coolant holes of different shapes on transport
interfaces. That is processes during resistance spot welding. Electrode
The top surface of electrode:f~L1 zL2 zL3 zL4 , materials have often been classified as copper based
0#j#1 alloys and refractory metal compositions with high
LTE thermal and electrical conductivities and low contact
{ ~BioET (TE {1) (15) resistance. They are used to prevent burning of the
Lf
workpiece surface or alloying of the electrode face.
Cooling effects at interfaces are governed by Biot
The electrode–coolant hole surface and the electrode– numbers, where heat transfer coefficients are chosen
surroundings surface:L1 zL2 zL3 ƒfƒL1 zL2 zL3zL4 from 103–105 W m22 K21 for forced convection in the
1 LTE coolant hole and heat transfer across the interface
~Biw (TE {1), j~0 (16) between the electrode and workpiece and the top
N Lj
boundary of the electrode, and from 10–50 W m22 K21
1 LTE for free convection at interfaces between the electrode and
{ ~BioEs (TE {1), j~1 (17) atmosphere.
N Lj
The isothermals in the workpieces and electrode at
dimensionless times of 0?248 and 0?317 are shown on the
Equation (17) is also applicable forL1 ƒfƒL1 z right and left hand sides in Fig. 2a respectively. The
L2 zL3 at j51. times chosen are respectively the instants when the
The surface of the coolant hole: L1 zL2 ƒfƒL1 thicknesses of the liquidus line dliq50?4 L in the melting
zL2 zL3 , j50 period, and 0 when solidification is complete. It can be

Science and Technology of Welding and Joining 2013 VOL 18 NO 8 664

 Wei and Wu Numerical study of electrode geometry effects on resistance spot welding

a re50?9, rc50?5, L150?1, L250?5, L350?1, L451?4; b re50?75; c L150?5; d L251; e rc50?1; f L350?5; g L452?8
2 Dimensionless isothermals in workpieces and electrodes at times when nugget thickness dliq50?4 L and 0 for different
geometries of electrodes

Science and Technology of Welding and Joining 2013 VOL 18 NO 8 665

 Wei and Wu Numerical study of electrode geometry effects on resistance spot welding

seen that the nugget shape is nearly a rectangle with the Provided that the heat affected region does not extend to
highest temperature at the centre on the faying surface. the tip of the coolant hole in the early stage, electrode
The temperature in the electrode gradually decreases in cooling due to the coolant hole in the last term is negligibly
the direction away from the workpiece. The temperature small. Electrode cooling dissipated to the coolant hole
gradient indicates that heat transfer is primarily in the increases its magnitude and becomes a rate determining
axial direction near the workpiece, whereas radial heat factor in the late stages. Thickness growth of the heat
transfer increases its magnitude as the coolant hole is ~
affected region or the thermal diffusion layer dd measured
approached. The temperature at the electrode face is from the electrode face can be estimated by scaling
rather uniform. The dimensionless time needed to reach equation (22)39
complete solidification after the power is off is 0?069.
This finding can be used to indicate the magnitude of the ~a
electrode cooling effect. Figure 2b shows that the nugget 1 E ~ ~ ~ 1=2
~t * ~ 2 or dd *ð aE t Þ (26)
grows fast as electrode face radius decreases (refer to dd
Fig. 2a). This is attributed to the high heat generation
which shows that the thermal diffusion layer can reach
resulting from the concentrated current density near the
the top of the electrode as time gets longer. The total
electrode edge.28 Electrode temperature decreases with
cooling rate in the late stages then yields40
decreasing face radius, because the time needed to raise
the temperature is short. Since the dimensionless time ~ ~
~q A~ T EW {T 0
required to reach complete solidification after the power is EW EW ~ ~ (27)
P Lj X 1
off is identical to that of the previous case, the effect of ~ ~z ~ ~
electrode face radius on cooling rate is insignificant. The k j Aj hj Aj
reason for this is that concentrated heat generationed at ~ ~
where q~EW , AEW and TEW are respectively heat flux,
the electrode edge immediately vanishes after the power is
area and temperature at the electrode face, the de-
off. The solidus line readily expands to the workpiece
nominator represents thermal resistances, similar to
surface and melts through, leading to collapse or rupture.
electric resistances for electric circuits problems. Index
Small electrode face radius should be avoided.
of summation j denotes different regions of thermal
In this work, a combination of unsteady one-
resistances. Equation (27) indicates that electrode cool-
dimensional energy equations in the solid region of the ~
workpiece and electrode can be used for interpreting the ing decreases with increasing thickness Lj in the direc-
computed results. That is tion of heat transfer, and decreasing thermal conductivity
~ ~
! kj , heat transfer coefficient hj and cross-sectional area
~
~ c~pi T ~ ~
Lr i i L ~ LT i Aj .
~ ~ ki ~ where i~s, E (22)
L t~ Lx Lx The electrode cooling rate depends on which thermal
resistances are involved within the thermal diffusion
which are respectively satisfied by energy balances at the layer in the course of welding. The thermal diffusion
solid–liquid interface and the surface of the coolant hole thickness can be estimated by 4(~ aE t~)1=2 .41 For the
~ ~ electrode made of copper with a thermal diffusivity of
~ ~ LT s ~ dd
q~{ ks ~ zr~s hsl ~ (23) 1024 m2 s21, the thickness of the thermal diffusion layer
Lx dt is less than the distance between the tip of the coolant
 hole and electrode face, L1zL250?6, for a welding time
~ 
~ LT E  ~ ~ ~ of less than one-third cycle. However, thermal diffusion
{ kE ~  ~ heff (Tw {T0 ) (24)
Lx  ~ easily reaches the top of the electrode, as can be seen on
L
the right side of Fig. 2a. At a dimensionless time of
0?248 corresponding to 30 cycles (50?5 s), the thickness
In equations (23) and (24), q~ represents heat flux from for thermal diffusion is ,30 mm, which is about six
the molten nugget to the melting front at location ~ d, h~sl is times the radius of the electrode. The thickness of
~
the latent heat for melting, L is the top location of the thermal diffusion is thus greater than the total length of
electrode and h~eff is an effective heat transfer coefficient, the electrode: L1zL2zL3zL452?1,6.
which accounts for cooling through the coolant hole and Figure 2c and d respectively shows that electrode
electrode to the surroundings. Substituting equations (23) cooling or cooling rate in the workpiece decreases, and
and (24) and considering a perfect contact surface between nugget growth and electrode temperature increase, as
y
the electrode and solid at location of , an integration of thermal resistance increases due to the increased
equation (22) over the solid region and electrode then leads truncated length of the electrode and distance between
to the tip of the coolant hole and truncated location of the
~ electrode (refer to Fig. 2a). Figure 2e and f shows that
~q~( c~ T ~ ~ ~ dd electrode temperature and nugget growth increase and
ps m z hsl ) rs ~ z
dt cooling rate decreases with decreasing the maximum
ð l~ ð~ radius ratio and increasing the truncated length of the
d ~ ~ ~ ~ d L~ ~ ~ ~ ~ ~ ~
r c T d xz r cpE T E d xz heff (T w {T 0 ) coolant hole. They are attributed to increased thermal
d ~t d~ s d ~t l~ E
ps s
resistance resulting from a decrease in cross-sectional
(25)
area and increase in truncated thickness of the coolant
which indicates that heat transfer from the nugget is hole respectively. Since electrode mass increases, the
balanced by electrode cooling via melting of the solid, increased length of the maximum radius of the coolant
increased sensible heat of the workpiece and electrode, and hole decreases electrode temperature and nugget growth
heat transfer to the coolant hole and surroundings. and increases cooling rate slightly, as shown in Fig. 2g.

Science and Technology of Welding and Joining 2013 VOL 18 NO 8 666

 Wei and Wu Numerical study of electrode geometry effects on resistance spot welding

3 Dimensionless temperature history at different locations in workpieces for different a L4, b re, c L1, d L2, e rc and f L3

The effects of the dimensionless length of the The periodic line at the bottom denotes AC. It can be
maximum radius of the coolant hole on dimensionless seen that an increase in the length of the maximum
temperature as a function of dimensionless time at radius of the coolant hole slightly reduces the tempera-
different locations in the top workpiece are shown in ture in the workpiece. Slight deviation, however, can be
Fig. 3a. The left and right ordinates respectively re- seen at the centre of the electrode face after dimension-
present dimensionless temperature and welding current. less times .0?12 within the heating period. The

Science and Technology of Welding and Joining 2013 VOL 18 NO 8 667

 Wei and Wu Numerical study of electrode geometry effects on resistance spot welding

4 Dimensionless temperature history at different locations in workpieces for different a BiE, b and c Biw, d BiOET and e
BiOES

temperature increases rapidly to a local maximum of variation of temperature with time is in agreement with
,3?5 near the electrode edge at a dimensionless time of predictions.34 The local peak of temperature disappears
,0?02 (about two to three cycles) in an early stage. This away from the electrode edge. Temperature in regions
is attributed to the high electric current density near the near the axisymmetric axis and faying surface increases
electrode edge.28 The temperature then decreases and slightly in an early stage because of time delay of electric
increases as time further increases. The predicted current density.42 Thereafter, temperature increases

Science and Technology of Welding and Joining 2013 VOL 18 NO 8 668

 Wei and Wu Numerical study of electrode geometry effects on resistance spot welding

welding current provided by Gould43 in welding

AISI 1008, as shown in Fig. 5. In this case, the solute
concentration f^m,0
a
50?32 wt-% of Mn in a Fe–Mn alloy.
This work shows better agreement with the experimental
result than the one-dimensional prediction. The former
result neglected heat transfer in the radial direction. Thus,
the thicknesses of the nugget are overestimated.

Conclusions
The conclusions drawn are as follows.
1. Electrode cooling affects temperatures in the
electrode and workpiece and nugget growth during not
only freezing and cooling periods but also heating and
melting periods. Major factors affecting electrode cool-
ing are different in different stages, characterised by the
growth of the thermal diffusion layer.
2. Electrode cooling can be interpreted by thermal
resistances within the growing thermal diffusion layer in
the electrode. Thermal resistances in the course of welding
5 Comparisons of nugget thickness versus welding cur-
should sequentially include contact resistance at the
rent in welding AISI 1008 steel between one-dimen-
electrode face, bulk resistances resulting from the truncated
sional predications and measurements from Gould43
region of the electrode, thickness between the coolant hole
and this study
and truncated location of the electrode, shapes of the
coolant hole and electrode and heat transfer through the
rapidly. The solidus temperature ,5?7 occurs at a coolant hole and electrode to the surroundings. Electrode
dimensionless time of 0?17. The liquidus temperature cooling due to the coolant hole is absent in the early stage.
appears after a dimensionless time of around t50?23. 3. Cooling rate in the late stage, especially after the
After the electric current is off, the molten nugget power is off, is enhanced, as thermal resistance
becomes completely solidified within several cycles. decreases. This is attributed to increases in the maximum
The variations of dimensionless temperature with time radius ratio and Biot number of the coolant hole, and
at different locations are similar for different electrode decreases in the truncated lengths of the electrode and
face radii, as shown in Fig. 3b. An electrode face with a coolant hole and the distance between the tip of the
small radius results in a high temperature and rapid coolant hole and the truncated location of the electrode.
growth of temperatures in all locations during the entire The Biot number and geometries of the coolant hole
range of welding. Figure 3c and d respectively shows dominate electrode cooling in the late stage.
that increases in truncated length of the electrode and 4. An increase in the contact Biot number at the electrode
distance between the tip of the coolant hole and face also enhances cooling, indicating that heat transfer
truncated location of the electrode increase temperature from the workpiece to the electrode is still important.
in the workpiece. The difference in temperatures is 5. An increase in electrode face radius increases the
found near the electrode face at the axisymmetric axis in electrode temperature and decreases nugget growth rate
the early stage. The effects of dimensionless maximum and workpiece temperature. The cooling rate is, how-
radius and truncated length of the coolant hole on ever, changed slightly. Temperatures at the electrode
temperature in the workpiece, however, are insignificant face are rather uniform.
in the early stage, as shown in Fig. 3e and f. 6. In most cases, a decrease in electrode cooling due to
The effects of contact Biot number at the electrode the coolant hole results in an early onset, fast growth of
face on the temperature histories at the same locations in the weld nugget and an increase in electrode temperature.
the top workpiece are shown in Fig. 4a. A low contact 7. An increase in the length of the maximum radius of
Biot number or high thermal contact resistance increases the coolant hole increases the cooling rate slightly, even
the nugget growth rate and decreases the increase in though thermal resistance increases. This can be
attributed to an increase in electrode mass.
temperature in the early stage. Temperature then
8. Designing the shapes of the electrode and coolant
increases rapidly. Even though power is off, a low
hole to meet required cooling rate or nugget growth rate
contact Biot number significantly reduces the cooling
with a specified welding time is challenging. This study
rate. This implies that heat transfer from the workpiece
gives systematical insight into optimum design and
to the electrode is still important. A coolant hole with a
sensitivity analysis of the shape of electrodes containing
high Biot number significantly reduces the temperature
a coolant hole.
in workpiece and nugget growth in the late stage (see
equation (27)), as shown in Fig. 4b and c. Cooling rate
greatly increases, after the power is off, indicating that References
the Biot number of the coolant hole is a rate determining 1. H. A. Nied: Weld. J., 1984, 63, 123s–132s.
factor in the late stage. Figure 4d and e, however, shows 2. C. L. Tsai, O. A. Jammal, J. C. Papritan and D. W. Dickinson:
that the cooling rates through the top of the electrode Weld. J., 1992, 71, 47s–54s.
3. R. W. Messler Jr. and M. Jou: Sci. Technol. Weld. Join., 1996, 1, 1–9.
and atmosphere are respectively small and insignificant. 4. Z. Feng, S. S. Babu, M. L. Santella, B. W. Riemer and J. E. Gould:
This model is confirmed by good agreement with the Proc. 5th Int. Conf. on ‘Trends in welding research’, Pine
measured relationship between nugget thickness with Mountain, GA, USA, June 1998, ASM International, 1–6.

Science and Technology of Welding and Joining 2013 VOL 18 NO 8 669

 Wei and Wu Numerical study of electrode geometry effects on resistance spot welding

5. J. A. Khan, L. Xu and Y. J. Chao: Sci. Technol. Weld. Join., 1999, 25. X. M. Lai, A. H. Luo, Y. S. Zhang and G. L. Chen: Int. J. Adv.
4, 201–207. Manuf. Technol., 2009, 41, 226–233.
6. B. H. Chang, M. V. Li and Y. Zhou: Sci. Technol. Weld. Join., 26. Z. H. Rao, S. M. Liao, H. L. Tsai, P. C. Wang and R. Stevenson:
2001, 6, 273–280. Weld. J., 2009, 88, 111s–119s.
7. K. Matsuyama: J. Jpn Weld. Soc., 2001, 70, 596–609. 27. P. S. Wei and C. Y. Ho: J. Heat Transfer, 1990, 112, 309–316.
8. N. T. Williams and J. D. Parker: Int. Mater. Rev., 2004, 49, 45–75. 28. P. S. Wei, S. C. Wang and M. S. Lin: J. Heat Transfer (Trans.
9. J. C. Tan, S. A. Westgate and T. W. Clyne: Sci. Technol. Weld. ASME), 1996, 118, 762–773.
Join., 2007, 12, 490–504. 29. Y. B. Li, Z. Q. Lin, S. J. Hu and G. L. Chen: J. Appl. Phys., 2007,
10. S. K. Khanna and X. Long: Sci. Technol. Weld. Join., 2008, 13, 101, 053506.
278–288. 30. Y. B. Li, Y. T. Li, Q. Shen and Z. Q. Lin: Int. J. Heat Mass
11. C. Ma, S. D. Bhole, D. L. Chen, A. Lee, E. Biro and G. Boudreau: Transfer, 2013, to be published.
Sci. Technol. Weld. Join., 2006, 11, 480–487. 31. P. S. Wei and T. H. Wu: Int. J. Heat Mass Transfer, 2012, 55,
12. L. Liu, S. Q. Zhou, Y. H. Tian, J. C. Feng, J. P. Jung and Y. N. 3316–3324.
Zhou: Sci. Technol. Weld. Join., 2009, 14, 356–361. 32. A. De, M. P. Thaddeus and L. Dorn: ISIJ Int., 2003, 43, 238–
244.
13. E. Tolf, J. Hedegard and A. Melander: Sci. Technol. Weld. Join.,
33. U. Dilthey, H.-C. Bohlmann, V. Sudnik, V. Erofeev and R.
2013, 18, 25–31.
Kudinov, Weld. Cutt., 2000, 52, E12–E16.
14. R. Raoelison, A. Fuentes, Ph. Rogeon, P. Carre, T. Loulou, D.
34. I. Ranjbar Nodeh, S. Serajzadeh and A. H. Kokabi: J. Mater.
Carron and F. Dechalotte: J. Mater. Process. Technol., 2012, 212,
Process. Technol., 2008, 205, 60–69.
1663–1669.
35. Y. B. Li, Q. Shen, Z. Q. Lin and S. J. Hu: Sci. Technol. Weld. Join.,
15. A. De, L. Dorn and O. P. Gupta: Sci. Technol. Weld. Join., 2000, 5,
2011, 16, 465–469.
49–57.
36. W. D. Bennon and F. P. Incropera, Int. J. Heat Mass Transfer,
16. S. Fukumoto, I. Lum, E. Biro, D. R. Boomer and Y. Zhou: Weld. 1987, 30, 2161–2170.
J., 2003, 82, 307s–312s. 37. P. S. Wei and F. B. Yeh: J. Heat Transfer (Trans. ASME), 1991,
17. P. Dong, M. V. Li and M. Kimchi: Sci. Technol. Weld. Join., 1998, 113, 643–649.
3, 59–64. 38. S. C. Wang and P. S. Wei: J. Heat Transfer (Trans. ASME), 2001,
18. F. Lu and P. Dong: Sci. Technol. Weld. Join., 1999, 4, 285–289. 123, 576–585.
19. Y. Wu, H. Serizawa and H. Murakawa: Trans. Jpn Weld. Res. 39. A. Bejan: ‘Convection heat transfer’, 2nd edn; 1995, New York,
Inst., 2003, 32, 277–282. Wiley.
20. E. W. Kim and T. W. Eagar: Weld. J., 1989, 68, 303s–312s. 40. F. P. Incropera and D. P. DeWitt: ‘Introduction to heat transfer’,
21. A. De: Sci. Technol. Weld. Join., 2002, 7, 119–124. 2nd edn; 1990, New York, Wiley.
22. K. S. Yeung and P. H. Thornton: Weld. J., 1999, 78, 1s–6s. 41. H. Schlichting: ‘Boundary-layer theory’, 7th edn; 1979, New York,
23. L. V. Glebov: Svar. Proiz, 1982, 1, 33–35. McGraw-Hill .
24. E. Yu. Latypova, S. M. Furmanov, Yu. A. Tsumarev and S. N. 42. P. S. Wei and T. H. Wu: J. Phys. D, 2011, 44D, 325501
Emel’yanov: Weld. Int., 2008, 22, 472–474. 43. J. E. Gould: Weld. J., 1987, 66, 1s–10s.

Science and Technology of Welding and Joining 2013 VOL 18 NO 8 670