Simulation and sensitivity analysis of controlling

parameters in resistance spot welding

The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.

Citation Kim, Euiwhan, and Thomas W. Eagar. “Simulation and Sensitivity

Analysis of Controlling Parameters in Resistance Spot Welding.”
Metals and Materials International 21, no. 2 (March 2015): 356–364.

As Published http://dx.doi.org/10.1007/s12540-015-4312-2

Publisher The Korean Institute of Metals and Materials

Version Author's final manuscript

Citable link http://hdl.handle.net/1721.1/105532

Terms of Use Creative Commons Attribution-Noncommercial-Share Alike

Detailed Terms http://creativecommons.org/licenses/by-nc-sa/4.0/

Met. Mater. Int., Vol. 21, No. 2 (2015), pp. 356~364
doi: 10.1007/s12540-015-4312-2

Simulation and Sensitivity Analysis of Controlling Parameters in

Resistance Spot Welding
1,* 2
Euiwhan Kim and Thomas W. Eagar

1
Department of Systems Engineering, Ajou University, Suwon 443-749, Korea
2
Professor of Materials Engineering and Engineering Systems, Massachusetts Institute of Technology,
Cambridge, MA02139, U.S.A.
(received date: 25 June 2014 / accepted date: 13 September 2014)

This study was performed to investigate the fundamental parameters controlling the nugget growth. The
parameters were categorized into four groups, i.e. material parameters, electrical parameters, thermal parameters
and geometrical parameters. In order to quantify the sensitivity of nugget growth to changes in these parame-
ters, a numerical model which incorporates the electrical, mechanical and thermal contact was developed.
As a result, a sensitivity index table was constructed and analyzed to ascertain the relative importance of these
characteristic parameters. It was found that the most important factor in determining the variability of nug-
get growth behavior is the ratio of contact radius to electrode radius and the ratio of electrode radius to the square
of specimen thickness. In general for a variation of 10%, the geometrical parameters are most important,
followed by the material parameters. The electrical parameters and the thermal parameters are the least import-
ant. The importance of contact at the faying interface is greater for the contact area than for the contact resis-
tance.
Keywords: metals, welding, interfaces, computer simulation, nugget growth

1. INTRODUCTION numerical analyses were performed, there are few studies to

investigate the controlling parameters of the process overall
To investigate the fundamental phenomena in the resis- [10-16]. In this research, in order to investigate the sensitiv-
tance spot welding process, the process system was qualita- ity of nugget growth to changes in each parameter, a numer-
tively modeled in a separate research program and four ical model which incorporates the electrical, mechanical and
groups of characteristic parameters were derived.[1,2] The thermal contact was developed and the simulation results
four groups are the material parameters, geometrical param- were analyzed quantitatively to develop a fundamental under-
eters, electrical parameters and the thermal parameters. These standing of the controlling parameters in the process.
parameters and the qualitative effects on the lobe width and
the energy input is outlined in Table 1 [1]. Among these, some 2. NUMERICAL MODEL
parameters are controllable while others are inherent to the
system. The difficulties confronted in assessing the weldabil- An axisymmetric two dimensional model was developed
ity of spot welding are mainly due to the combined effects using the ABAQUS code, a finite element computer algorithm.
of these parameters. Furthermore, there is some variability During welding three different physical processes occur
in each parameter from weld to weld. The material proper- simultaneously, i.e. thermal, electrical and mechanical pro-
ties may not always be the same even though the material cesses are present. For accurate calculation of the nugget
classification is the same. The electrode contact area and the growth behavior all three physical processes must be incor-
surface condition of the electrode also changes during the porated simultaneously. The pressure for the mechanical
welding sequence. All these uncertainties affect the weld- contact model was a combination of electrode force and the
ability of a material to a greater or lesser extent, yet in most thermal load. At first the contact area at room temperature
cases, it is very difficult to experimentally quantify the effect was estimated from a mechanical model. Then a calculation
of variations in each parameter [3-9]. Even though various of the temperature field with the thermal load was performed
using the contact area obtained from the mechanical model.
*Corresponding author: keuiwhan@ajou.ac.kr Even though the calculation shows plausible contact behav-
KIM and Springer ior, the true contact in the welding process can deviate from
 Simulation and Sensitivity Analysis of Controlling Parameters in Resistance Spot Welding 357

Table 1. Effect of parameters on the lobe width and energy input

(after Refs. 1, 2)
Increase in Lobe Width Energy Input
MATERIAL PARAMETER
Thermal conductivity k b + +
Bulk electrical Resistivity  b - -
Heat Capacity CP + +
CP /  b + +
kb/ b + +
GEOMETRIC PARAMETER
2
b/L ? +
D/b - +
ELECTRICAL PARAMETER
Rc / Rb + ?
Rcf / Rce + ?
Fig. 2. Current flow model.
THERMAL PARAMETER
hw / R + +
hc / kb + + For a given contact area the welding current was redistrib-
hc / R + + uted at each time step. Figure 2 shows the concept of this cur-
Subscript c : contact b : electrode radius rent distribution scheme with a representative electrical circuit.
Subscript b : work piece R : total electrical resistance In this way the model can consider differences in the tem-
Subscript w : water cooling L : specimen thickness
Superscript f : faying interface D : contact radius at faying interface perature dependent electrical resistivity of different materials.
Superscript e : electrode h : heat transfer coefficient Even though the electrical resistivity of the copper electrode
is one order of magnitude smaller than that of the steel, it was
included in the analysis as the current density becomes higher
near the contact tip [17]. Convective water cooling is assumed
to have a fixed heat transfer coefficient of 0.02 W/mm2 °C.

Fig. 1. Schematic comparison of the current flowing area and the

mechanical contact area.

the model. For this reason, if the calculated nugget size did
not match the experimentally measured nugget size, the con- Fig. 3. Electrical contact resistivity at electrode interface (after Ref. 17).
tact area at the faying interface was modified until a closer
match was obtained.
As for the thermal model, this model incorporates the
redistribution of current density caused by the uneven tem-
perature field and the size of the mechanical contact. Figure 1
shows a schematic of the current conducting area. Here the
meaning of the contact size is somewhat different from the
mechanical contact size calculated from the contact model.
In this case the contact size is the area where the current can
flow. For example, the current conducting area of a zinc coated
material is different from the numerically calculated area
due to the formation of a molten zinc halo around the weld
nugget. Fig. 4. Electrical contact resistivity at faying interface (after Ref. 17).
358 Euiwhan Kim and Thomas W. Eagar

The electrical contact resistivity and the contact heat transfer radius. The ratio is almost constant at 1.2. Thus the contact
coefficient at the electrode interface were taken from refer- radius at the faying interface is 20% larger than the elec-
ence 18. The electrical contact resistivities are shown in Fig. 3 trode face radius. This will result in 30% lower average cur-
and Fig. 4 [18]. rent density at the faying interface than at the electrode
Temperature dependent properties such as bulk electrical interface. Thus, by the nature of mechanical contact the cur-
resistivity, heat capacity, thermal conductivity, heat of fusion rent density is much higher at the electrode interface. Table 4
and electrical and thermal contact properties were used in this shows the effect of specimen thickness on the contact area.
analysis. The values for these properties were varied accord- As the specimen thickness increases, the ratio of contact radius
ing to both published and measured data. to electrode radius increases. The increment of contact radius
The sheet steels used in these simulations and experiments is proportional to the specimen thickness. Thus, it is seen
are designated as follows. AMBR: bare steel, AM35: electrogal- that the electrode force has little effect on the contact size for a
vanized with 35 g/sq. m of zinc per side, AM68: electrogalva- given material thickness. However, these results are applicable
nized with 68 g/sq. m of zinc per side, AM100: electrogalvanized only to the very early stages of the welding.
with 100 g/sq. m of zinc per side, G40 and G60: hot dip gal- Figures 5 and 6 show cascade plots of the contact pressure
vanized. at the faying interface and at the electrode interface during
welding. As welding progresses, the contact pressure at the
3. RESULTS AND DISCUSSION center increases due to thermal expansion in the electrode
and the work piece. In contrast, the contact pressure at both
3.1. Contact behavior the periphery of the faying interface and the electrode inter-
In general, at the start of welding, the location for the maxi- face decreases resulting in the loss of the mechanical seal.
mum contact pressure at both interfaces and the ratio of Due to the larger displacement in the center, the contact size
maximum pressure to average pressure is listed in Table 2. at the faying interface decreases as shown in Fig. 7. This may
Table 2 also shows the contact size at the faying interface. be related to expulsion of weld metal.
The electrode force did not cause any significant change in As the electrode force increases, the contact size at the
contact size. Thus, it is seen that the electrode force has the faying interface becomes larger as in Fig. 7. The force has a
most effect on the electrical and thermal contact properties strong effect on the contact area at the faying interface par-
rather than on the contact area during the early stages of weld- ticularly in the early stages of welding. A minimum contact
ing. Table 2 lists the ratio of contact radius to the electrode area is observed when the nugget grows to a size comparable

Table 2. Effect of electrode force on contact size and pressure

Electrode force Radius for maximum contact pressure (mm) Normalized maximum pressure Contact size at faying
(lbs) Electrode interface Faying interface Electrode interface Faying interface interface, D (mm)
500 2.25 1.95 1.54 0.93 2.96
650 2.25 1.95 1.57 0.93 2.96
800 2.25 1.95 1.61 0.94 2.96
*electrode radius : 2.4 mm

Table 3. Effect of electrode size on the contact size at faying interface

Electrode radius, b (mm) Contact radius at faying interface, D (mm) Ratio of contact radius to electrode radius
2.4 2.96 1.23
2.6 3.16 1.22
2.8 3.36 1.20
*electrode force : 500lbs

Table 4. Effect of specimen thickness on the contact size

Specimen thickness, L Contact radius at faying Ratio of contact radius to Ratio of difference in contact radius and
(mm) interface, D (mm) electrode radius, D/b electrode radius to the specimen thickness
0.6 2.80 1.17 0.67
0.8 2.96 1.23 0.67
1.2 3.20 1.33 0.67
1.8 3.48 1.45 0.67
*electrode radius : 2.4 mm, electrode force : 650 lbs
 Simulation and Sensitivity Analysis of Controlling Parameters in Resistance Spot Welding 359

Fig. 7. Change of contact size at the faying interface during welding.

From this simulation it is seen that the current density at

the faying interface changes significantly during the course
of welding. The initially large contact area decreases as weld
progresses due to thermal expansion and then increases
Fig. 5. Change of contact pressure at the faying interface during welding. again as mechanical collapse begins in the work piece due
(a) after 1 cycle, (b) after 5cycle, and (c) after 9 cycle. to the presence of molten metal. This can be related to the
tailing of the lobe curve in the high current short weld time
region. It was seen in reference 19 that welding with high
current generally shows severe localization in the heat gen-
eration pattern [19]. This localized heating is combined with a
very rapid expansion. In this case, the dissipation of heat from
the localized hot spot is small due to the very short weld time.
The localized thermal expansion may also cause loss of mechan-
ical constraint of the nugget envelope due to the asymmetry
of the process. Thus expulsion can occur before the formation
of a proper nugget resulting in a small or negligible current
range.
The zinc coated material will behave in a different way.
The current conducting area is determined by the behavior
of the zinc at the faying interface. Figure 8 shows the exper-
imentally measured growth of nugget size and zinc halo
during welding. The effect of increased current conducting

Fig. 6. Change of contact pressure at the electrode interface during

welding. (a) after 1 cycle, (b) after 5cycle, and (c) after 9 cycle.

to the electrode (within 0.15 second for this simulation case).

After this time the contact size increases due to large defor-
mations in the work piece. Severe indentation begins at this
stage of welding. Fig. 8. Evolution of halo size and nugget size.
360 Euiwhan Kim and Thomas W. Eagar

area in a zinc coated steel is very important on the nugget

growth mechanism. As the current flow area increases at the
faying interface, the heat generation rate decreases in a qua-
dratic manner compared to the heat generation rate at the
electrode interface. This implies that the temperature rise at
the electrode interface is much more rapid than at the faying
interface for coated steel. If the effect of current oscillation
in AC welding is also considered, the amplitude of tempera-
ture fluctuation at the electrode interface can be even greater.
This phenomenon will hinder the gradual formation of a nug-
get and will also make surface expulsion from the electrode
interface easier.
Fig. 10. Effect of changes in the contact diameter at the faying interface.
3.2. Effect of changes in basic variables on nugget growth
The effect of variability in each parameter derived from Table 5. Effect of changes in basic variables on nugget growth time
Table 1 was numerically simulated to see the effect on the Increase in nugget
nugget growth behavior and thus to be used in determining Increase in (by 10%)
growth time (cycle)
the most important parameters in controlling nugget growth. Thermal conductivity, kb 0.50
The simulated material was an electrogalvanized low car- Specific heat, CP 0.75
bon steel AM68. Electrical resistivity, b -1.0
For a given variability in each parameter, the differences Specimen thickness, L -0.25
in weld time required for a nominal nugget size of 2.0 mm Electrode radius, b 4.5
radius was estimated. The total range of variations in each Contact radius, D 2.5
parameter is double the variation of 5% in one direction. Contact resistance at faying interface, Rc
f
-0.001
The difference between the upper and lower times to nugget Contact resistance, Rc -0.126
formation was taken as a measure of change in nugget growth Contact resistance at electrode interface, Rc
e
-0.125
behavior. An increase in nugget development time results in Contact heat transfer coefficient, hc 0.125
a larger energy input requirement. This is equivalent to the
requirement of higher weld current level. This difference is
also representative of the lobe width. A large difference in weld ner the effect of changes in each basic variables on the weld
time is equivalent to the larger lobe width on the current axis. time were simulated. The results are summarized in Table 5.
Figure 9, as an example, illustrates the simulation results The discussion of the effect of changes in basic variables
of nugget growth behavior for the case of variations in elec- are discussed below.
trode diameter. In this Fig. 9, 10% variation in electrode diame-
ter resulted in 4.5 cycles of weld time change. As another 3.2.1. Effect of Material Related Variables
example, Fig. 10 illustrate the effect of changes in the contact The thermal conductivity, kb, electrical resistivity, b, and
diameter at the faying interface. In this case 10% variation heat capacity, CP, were varied in the model. The nugget forma-
resulted in 2.5 cycle changes in weld time. In the same man- tion starts later as thermal conductivity increases. The reason
can be ascribed to a lower heat loss to the electrode and sur-
rounding material. If heat conduction to the electrode is
dominated by interface control, one cannot expect a strong
effect of the thermal conductivity on nugget growth, partic-
ularly when the specimen is thin.
The materials with high electrical resistivity experience a
faster heat build up in the nugget and complete nugget for-
mation in less time. The reason of this increased sensitivity
comes from the cumulative effect of electrical resistivity. As
the electrical resistivity increases with temperature, for a mate-
rial with higher initial electrical resistivity, the faster initial rise
of temperature raises the resistance more rapidly and gener-
ates more heat. Thus the nugget growth curve will become
steeper and produce shorter nugget development time. This
Fig. 9. Effect of changes in the electrode diameter. phenomenon may not be beneficial in terms of stable nugget
 Simulation and Sensitivity Analysis of Controlling Parameters in Resistance Spot Welding 361

growth. The slope of dsb/dt has another very important effect rial at the electrode interface and will make welding difficult.
in its contribution to the redistribution of the current. Due to This is particularly true for welding of thin material. As the
the geometry of the welding system, the center part of the nug- displacement induced by thermal expansion is cumulative
get is usually the highest temperature region. Thus the center in nature, the total thermal displacement of thin material at
part will have the highest electrical resistivity. The higher resis- the center line is smaller than that of a thick material by roughly
tance at the center part will push the current to the periphery the ratio of specimen thickness. The smaller thermal displace-
of the nugget, increasing the temperature in this peripheral ment at the center of contact will make contact at the periphery
region. Thus it is not clear whether the greater slope in tem- easier. This results in a larger D/b value for thin materials. A
perature dependence of electrical resistivity is beneficial or not. larger value of D/b means that a greater percentage of the
The effect of heat capacity is reversed, compared to the heat is generated at the electrode interface as opposed to the
effect of resistivity. For an increase in specific heat, the time faying interface.
required to raise the work piece temperature to the melting
temperature is longer. Since the temperature rise time is longer, 3.2.3. Effect of interface related variables
the temperature field in the work piece has more chance to even In Table 1, two parameter groups are listed as the electrical
out the temperature profile in the radial direction. Thus, once parameter and the thermal parameter. These basic parameters
the nugget starts to form, the nugget can grow faster. in each group are interface related parameters, i.e. electrical
contact resistance at the faying interface, Rcf, and at the elec-
e
3.2.2. Effect of geometrically related variables trode interface, Rc , contact heat transfer coefficient, hc, at
The nugget growth time was estimated for various size of the electrode interface and the convective heat transfer coef-
electrode face radius, b, contact radius at the faying interface, ficient at the coolant interface, hw.
D, and specimen thickness, L. The effect of electrode size is It was seen in reference 19 that the effect of coolant flow
very strong, producing changes in nugget growth time of rate is small compared to the effect of electrode face thick-
4.5 cycles. This change may be great enough to make nugget ness. The effect of flow rate becomes more important as the
formation impossible. However, the effect of variation in the electrode thickness decreases. However, as the heat transfer
work piece thickness is very small. As the thickness increases, at the coolant interface is another complicated boundary heat
the nugget forms earlier. Since the thicker material loses less transfer problem, it was not included in this variability anal-
heat to the electrode, more heat is available for nugget for- ysis [20].
mation. This can be explained by the higher electrode tempera- One important aspect to be noted here is that the contact
ture as measured in reference 17. heat transfer coefficient at the electrode interface is coupled
The contact size, D, includes the total area of the current to the electrical contact resistance. This means that one can-
path. It was seen in Figs. 7 and 8 that electrode force and not change the electrical contact resistance without affecting
zinc halo formation are the primary sources of changes in the thermal contact heat transfer coefficient. In this simula-
contact area at the faying interface. One interesting observa- tion, when the contact resistance at the electrode interface was
tion in these nugget growth curves is the varying effect of increased, the contact heat transfer coefficient was decreased
contact size. Decreasing contact size shows more significant by the same percentage used for the change of resistance.
changes in nugget development time than increasing contact Thus the effect of changes in contact resistance and contact
size. Two reasons can be postulated. The first is the stronger heat transfer coefficient were considered to be the same but
effect of current redistribution when welding a material with in opposite directions.
small contact area. This was related to the temperature depen- The simulation results showed that the change in the con-
dence of the electrical resistivity of the work piece. The other tact resistivity at the faying interface does not produce any
is the quadratic effect of the current density on the heat gen- significant change in nugget growth time. An effect can
eration rate at the faying interface. This phenomenon makes only be seen in the early stages of nugget growth. This can
the effect of contact area very significant in the nugget develop- be ascribed to the rapidly decaying characteristics of contact
ment mechanism. resistance at the faying interface. This needs to be distinguished
If the contact size or the electrode size increases, the total from the ease of welding of materials with high interface contact
resistance in the system will decrease due to increases in the resistance at the faying interface. The effect of contact resis-
current conducting area. This will reduce the welding current tance changes at the electrode interface was more pronounced
and the power absorbed by the work piece. Thus the increasing than the effect of changes at the faying interface, even though
electrode size or the contact size will shift the nugget forma- it is very small.
tion time a greater extent at longer weld times. The effect of simultaneous changes both at the faying
For a given material, the value for D/b is determined by interface and at the electrode interface was also simulated to
the loading condition and the presence of a coating. As a general see the effect of changes in contact resistance, Rc, as a sum
rule, a larger value of D/b will localize the heating of mate- of contact resistances at the faying interface and at the elec-
362 Euiwhan Kim and Thomas W. Eagar

trode interface. The results matched well with the combination effect of other parameters may become important.
of simulation results for faying interface change and those
for electrode interface change. Thus far it is seen that a vari- 3.4. Application example of sensitivity index
ation in the electrical contact resistances either at the faying Welding of very thin zinc coated steel is known to be very
interface or at the electrode interface has a small effect on difficult. This is believed to be due to the large b/L2 and D/b
the nugget growth curve. value. The lower value of Rc/Rb is another less important reason.
As the thickness becomes less, more heat flows to the electrode.
3.3. Sensitivity of nugget growth curve to characteristic For ideal contact conditions this may help produce a sound
parameters nugget. In an ideal case, greater heat transfer to the electrodes
The data listed in Table 5 contain information about the will lower the work piece temperature at the electrode inter-
sensitivity of nugget growth to variations in basic parameters. face, making the temperature gradient in the axial direction
From this information Table 6 was constructed to see the effect steeper. This will help increase the temperature difference
of characteristic parameters on nugget growth time. Table 6 between the faying interface and the electrode interface. How-
can be used as an index to ascertain the relative importance ever as b/L2 becomes large, the temperature difference between
of these characteristic parameters and the effect of varia- the faying interface and the electrode interface becomes very
tions on weld time. In constructing this table it was assumed small due to a smaller aspect ratio of the thickness to the
that the variation in the component is less than 10%, which contact radius. In addition, mismatched or tilted electrodes
was the range used in the simulation. Thus the results are only will localize the welding current distribution, resulting in a
applicable to the particular case considered in this work. severely localized temperature. In the welding of thick material,
For example, the effect of kb/b and Cp/b were calculated this localization is believed to be dissipated very quickly due
by considering the maximum possible range. If a 10% to the larger heat conduction path. For thin material, the
increase in kb/sb results from an increase in kb, the increase in conduction path is limited to two dimensions and rapid heat
weld time will be 0.5. However if the 10% increase results conduction from the hot spot is not possible. One good way
from a 10% decrease of b, the increase in nugget growth to help reduce this problem is to increase the heat transfer
time becomes 1.0. Combining the effects of each variable in coefficient at the electrode interface while increasing the
this manner, one can estimate a parameter index. This index electrical contact resistance at the faying interface.
can be used as a measure of the sensitivity of nugget growth In this respect, the ratio of contact resistance at the faying
to changes in the characteristic parameters. interface to the contact resistance at the electrode interface
Among all these characteristic parameters, the geometri- is a very important parameter in the welding of thin materials.
cal parameters show the strongest effects. Table 6 shows that The desirable heat distribution pattern is the one which has
the most important parameter is the ratio of contact size to the highest temperature at the faying interface, particularly
the electrode size. The next is the ratio of electrode radius to at the nugget center. It was also shown in the previous sec-
the square of work piece thickness. The material parameters tion that the effect of contact size at the faying interface is
have intermediate importance. The electrical parameters and the major factor in determining the nugget growth charac-
the thermal parameters show the least importance in this teristics. The larger contact area will reduce the current den-
classification. This is true for the case of low carbon steel. How- sity at the faying interface and will prevent the nugget from
ever, if a larger variability of the electrical parameters and growing in a gradual way. More heat will be generated at
the thermal parameters is considered, these parameters can the electrode interface. Slower heating with a larger faying
be more important than the other parameters. As an example, interfacial area decreases the temperature gradient in the radial
even though the effect of the geometrical parameter is the direction and will make the nugget growth very abrupt.
strongest, if the variability of this parameter is very small, the To verify this concept an experimental lobe for 0.6 mm
thick G40 material was produced. Figure 11 shows the lobe
Table 6. Sensitivity index for the characteristic parameters curve of the original material containing zinc on both surfaces.
Increase in Possible range of change in nugget This lobe curve has only a 0.5 kA lobe width when welded
(by 10%) growth time (cycle) with 8 cycles. Figure 12 shows the lobe of the modified mate-
kb / b 0.5 ~ 1.0 rial. The zinc coating at the faying interface was etched away
CP / b 0.75 ~ 1.0 leaving a bare steel surface, but the zinc at the electrode interface
2
b/L 0.125 ~ 4.0 remained. The purpose was primarily to decrease the con-
D/b -4.5 ~ 2.5 tact size at the faying interface rather than to increase the
Rc / Rb -0.126 ~ 1.0 contact resistance. One has seen that the effect of the contact
f e
Rc / Rc -0.001 ~ 0.125 resistance is much less than that of the contact size. The result
hc / R -0.125 ~ 1.126 shows that the lobe width increases by more than 250%.
hc / kb -0.5 ~ 0.125 For better weldability of a thin material, it is necessary to
 Simulation and Sensitivity Analysis of Controlling Parameters in Resistance Spot Welding 363

type of zinc coating has a smaller electrical contact resistivity.

A reduction of contact size can be made as was done by using
a bare steel surface at the faying interface. However this is
not practical in terms of corrosion protection. Another possibil-
ity of decreasing the contact size is to use a galvannealed surface
at the faying interface. The contact heat transfer coefficient
of galvannealed steel has similar thermal contact character-
istics as bare steel. For this reason the welding of galvannealed
steel is generally reported to have the welding characteris-
tics of bare steel. Thus a material with an electrogalvanized
surface at the electrode interface and a galvannealed surface
at the faying interface would be beneficial in improving the
welding current range. In practice, this is not a very practi-
cal solution to improved weldability.
Changes in electrode size were the most influential among
all the basic variables. An increase of 10% in the electrode
Fig. 11. Lobe curve of 0.6 mm thick G40 hot dip galvanized steel.
radius, which is generally smaller than that commonly observed
in actual welding, delayed the nugget formation by 4 cycles.
This is large enough to make the nugget formation impossi-
ble under practical conditions.

4. CONCLUSION

Understanding the fundamental phenomena of the resis-

tance spot welding process is crucial to improve the process
application. As a first step a lumped parametric model was
developed to derive the characteristic controlling parameters.
In this research, in order to quantify the sensitivity of nugget
growth to changes in these parameters, a numerical model
which incorporates the electrical, mechanical and thermal
contact was developed. A sensitivity index table was con-
structed and analyzed to ascertain the relative importance of
these characteristic parameters.
Fig. 12. Lobe curve of modified 0.6 mm thick G40 hot dip galvanized It was found that the most important factor in determining
steel (coating only on the electrode side). the variability of nugget growth behavior is the ratio of contact
radius to electrode radius and the ratio of electrode radius to
modify the axial temperature profile to produce a high tem- the square of specimen thickness. In general for a variation
perature gradient. Two different methods can be contemplated of 10%, the geometrical parameters are most important, fol-
in this respect: the first is modification of the faying inter- lowed by the material parameters. The electrical parameters
face. A reduction in contact area or an increase in the ratio and the thermal parameters are the least important.
of electrical contact resistance at the faying interface to the It was also found that the ratio of contact radius at the faying
electrical contact resistance at the electrode interface is ben- interface to the electrode radius is about 1.2 at the start of
eficial. The second is tailoring of the axial temperature pro- welding and there is a pressure concentration at the periphery
file by modifying the current wave form. In this case, heat is of the contact at the faying interface and at the edge of the
generated for a certain time and then the current is halted or electrode. However, the pressure concentration decreases as
reduced. The electrode interface cools down more quickly welding progresses, the contact size at the faying interface
due to the heat sink of the electrodes and the faying inter- decreases resulting in very easy expulsion. The importance
face maintains its heat and hence a higher electrical resis- of contact at the faying interface is greater for the contact
tance. Then reheating follows with successive cooling. This area than for the contact resistance. The contact area at the
cycle is repeated till formation of a nugget starts. This scheme faying interface determines the current level and the ease of
should be coupled with a method which enhances the heat welding. The ease of spot welding of bare steel is due to the
flow at the electrode interface. An electrogalvanized surface small contact size, not to the high contact resistance. A 0.6 mm
with a thick coating is a good candidate for this purpose. This thick galvanized sheet steel can be made more weldable through
364 Euiwhan Kim and Thomas W. Eagar

making faying interface modification to have smaller con- 9. Md. A. M. Hossain, Md. T. Hasan, S.-T. Hong, M. Miles, H.-
tact size induced by a harder surface. H. Cho, and H. N. Han, Met. Mater. Int. 19, 1243 (2013).
10. S. J. Na and S. W. Park, Welding Journal 75, 233s (1996).
REFERENCES 11. X. Sun, Welding Journal 79, 244s (2000).
12. X. Sun and M. A. Khaleel, Welding Journal 83, 197s (2004).
1. E. Kim, Sc.D., Analysis of Resistance Spot Welding Lobe 13. D. J. Browne, H. W. Chandler, J. T. Evans, P. S. James, J.
Curve, pp.209-287, Massachusetts Institute of Technology, Wen, and J. Newton, Welding Journal 74, 417s (1995).
Cambridge, U.S.A. (1989). 14. H. A. Nied, Welding Journal 63, 123s (1984).
2. E. W. Kim and T. W. Eagar, SAE Transactions, No. 880278 15. W. Zhang, Welding in the World 50, 29 (2006).
(1988). 16. Y. S. Zhang, J. Xu. M. Lai, and G. L. Chen, Science and Tech-
3. T. Horita, M. Oka, T. Kanamura, K. Yamazaki, and T. Fuji- nology of Welding and Joining 13, 192 (2008).
wara, Welding International 10, 937 (1996). 17. K. S. Yeung and P. H. Thornton, Welding Journal 78, 1s
4. A. Aravinthan and C. Nachimani, Welding Journal 90, 143s (1999).
(2011). 18. E. Kim, International Journal of Korean Welding Society
5. W. L. Chuko and J. E. Gould, Welding Journal 81, 1s (2002). 2, 1 (2002).
6. X. Sun, E. V. Stephens, M. A. Khaleel, H. Shao, and M. 19. E. W. Kim and T. W. Eagar, Welding Journal 68, 303s (1989).
Kimchi, Welding Journal 83, 188s (2004). 20. Z. H. Rao, S. M. Liao. H. L. Tsai, P. C. Wang, and R. Ste-
7. S. Boron, Welding International 12, 932 (1998). venson, Welding Journal 88, 111s (2009).
8. C.-W. Ji, I. Choi, Y. D. Kim, and Y.-D. Park, Korean J. Met.
Mater. 52, 931 (2014).