NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No.

119 September 2018

UDC 669 . 14 . 018 . 292 : 621 . 791 . 763 . 1 : 681 . 3

Technical Report

Finite Element Simulation of Resistance Spot Welding

Process for Automotive Steel
Hideki UEDA* Manabu FUKUMOTO
Hiroki FUJIMOTO Kazuo OKAMURA
Shota KIKUCHI Eisuke NAKAYAMA
Tohru OKADA Masanori YASUYAMA

Abstract
This work aims to develop the FEM analysis system for evaluating the overall phenom-
ena during resistance spot welding. The framework of this system is the finite element cal-
culation based on the incrementally coupled electrical- thermal-mechanical simulation pro-
cedure. Using this analysis system, the weldability lobes of mild steel and high strength steel
are predicted, which shows the validity and capability of this model. In addition, this analy-
sis system is applied to three-dimensional model simulation door opening, and influences of
the shunt, member rigidity and plate gap on welding are examined. Moreover, it is applied
to the estimation of the mixing ratio of molten metal, and then the flow curves and fracture
limit in the weld metal are calculated. Predicted fracture modes and joint strengths show
good agreement with the experimental results.

1. Introduction nugget growth under arbitrary sheet joining/welding conditions. 1, 2)

Resistance spot welding (hereinafter referred to as “spot weld- Furthermore, based on the initiatives, we have developed a coated
ing”) is widely used as a joining method in the assembling process steel plate analysis model taking into account the melting and evap-
of automotive bodies and parts. Spot welding has been used for oration of the coating, and also a spattering prediction model inside
more than 100 years. Due to the high productivity and low cost, spot of which molten metal starts to be spattered. This has enabled weld-
welding is still a major joining method in the field of automotive ability lobe prediction for actual steel sheet assemblies including
body manufacturing. At the same time, the need to meet the require- coated steel sheets. 3)
ments regarding a reduction in weight and crash safety of automo- In the past, studies were conducted assuming the most funda-
biles is increasing the production rate of high tensile strength steel mental welding conditions, i.e., cases in which strip specimens (test
plates/sheets used for automobile members. The use of hot-stamped coupons) were single-spot welded in a laboratory (stationary weld-
steel plates/sheets of which tensile strength after quenching reaches ing machine). In addition, analyses using axially symmetric models
1 470 MPa has also expanded. with the upper or lower electrode center axis used as the symmetri-
Against this backdrop, the number of locations where steel cal axis have been prevalent. This means that studies have been lim-
plates are joined into assemblies by spot welding has increased per ited to cases in which the influence of the electrode axis misalign-
vehicle body structure, complicating the determination of appropri- ment or steel sheet deformation on the weldability is considered to
ate welding conditions for all joining locations. Given the situation, be small. However, when welding for assembling an actual automo-
we have developed a spot welding analysis system based on an in- tive body is assumed, various factors may affect welding conditions;
crementally coupled analysis method between the electric, tempera- such factors include current shunting caused by contact with an ad-
ture, and stress fields using the finite element method analysis (here- jacent already-welded point or another member, presence of sheet
inafter referred to as the “FEM analysis”), as part of our continued gaps caused by less accurate component forming, and the rigidity in
initiatives to establish basic technology for accurately forecasting a the vicinity of a welding area determined by the shape of the com-

* Senior Researcher, Materials Reliability Research Lab., Steel Research Laboratories

1-8 Fuso-cho, Amagasaki City, Hyogo Pref. 660-0891

- 103 -
NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No. 119 September 2018

Fig. 1 Schematic illustration of the incrementally coupled electrical-thermal-mechanical analysis procedure

ponent. Given this, the spot welding simulation technology was ap-
plied to a 3D model that simulated a door opening as a part of an
actual automotive structure to examine the influence of current
shunting, member rigidity, and sheet gaps on the weldability in
comparison with the single-spot welding of strip specimens. 4)
In addition, the application of the spot welding simulation tech-
nology for the weld area strength evaluation was examined. If the
spot welding nugget diameter or weld pitch of an automobile mem-
ber is different from the value as instructed in the drawing, the weld
area may fracture due to collision deformation. For this reason, the
development of fracture prediction technology is required when de-
signing a member. The joint strength and form of a fracture of a spot
welding joint are significant depending on the load mode on the
weld area. Considering this, a fracturing criterion was created for a
spot weld area using the fracturing strain that took into account the
influence of stress triaxiality (hereinafter referred to as the “fractur-
ing limit value”). By applying the fracturing limit value to the FEM
analysis, a fracturing prediction FEM analysis approach capable of
accurately reproducing the strength and fracture form of a spot-
Fig. 2 History of nugget growth (exp. and CAE) and resistance between
welded joint was developed. 5) Structural members of automobiles electrodes (CAE). 590 MPa steel, thickness = 1.2 mm, electrode
use many sheet assemblies with different materials and thicknesses force = 3.92 kN, current = 7.1 kA
(hereinafter referred to as the “dissimilar material/thickness sheet
assembly”). In order to improve the fracturing prediction accuracy phase transformation in each coupled analysis. The main purpose
for dissimilar material/thickness sheet assemblies, an analysis model for using the coupled analysis method that considers the three fields
was built using the deformation resistance and fracturing limit value is to include the ever-changing influence of the contact pressure and
that were calculated based on the mixing ratio in the nugget molten the diameter of the section where the electric currents flow in the es-
area volume (hereinafter referred to as the “molten area mixing ra- timation of the contact resistance value to specify the resistance heat
tio”) calculated in a spot welding simulation. 6) This article describes at the interface. As represented by curve (a) in Fig. 2, the generation
the details of these initiatives. and growth of a nugget could be accurately predicted using this
method. The contact resistance, considered not to remain for long
2. Establishment of the Spot Welding Analysis System once the current has been applied, is likely to be viewed as unim-
2.1 Overview of the system portant; however, in fact, it is very important. As indicated by curve
Figure 1 shows a schematic view of the basic concept of the in- (b) in Fig. 2, an analysis that disregards the interface resistance heat-
crementally coupled electrical-thermal-mechanical analysis method ing, the resistance value does not increase sufficiently to cause am-
used in this study. 1) The basic flow involves alternately repeating a ple heat generation especially at the start of current application, hin-
coupled electrical-thermal analysis and a coupled thermal-mechani- dering precise prediction of nugget generation.
cal analysis at appropriate small intervals (e.g., 1/2 cycle). This ap- The analysis model has an axisymmetric structure with a pair of
proach is characterized by the consideration of the phase change and electrodes and two to four steel sheets. The sheet thickness and gap

- 104 -
 NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No. 119 September 2018

between sheets can be arbitrarily set. Regarding material property point is determined from the temperature distribution obtained as an
data necessary for the analysis, data owned by Nippon Steel & analysis result; the volume of the molten coating metal is calculated
Sumitomo Metal Corporation products was available including that from the film thickness and density of the coating metal, as shown
of mild steel sheets and 1 180 MPa-class high tensile strength steel in Fig. 3 (a). While assuming that the sheet separation area is filled
sheets. The consideration of surface coating is also enabled by the with the molten metal in the calculated volume, the continuity con-
model newly developed through this study. Modelling of the coated dition is applied considering the range in which the molten coating
steel sheet is described in the next section. This analysis system is metal is present as the diameter of the current flowing section, as
composed of many user subroutines and our own peripheral pro- shown in Fig. 3 (b). When the temperature further rises, and in the
grams based on Abaqus, a suite of general-purpose FEM analysis moment that the temperature in the vicinity of the sheet separation
codes. area reaches the boiling point of the coating metal, the molten coat-
2.2 Coated steel sheet analysis model ing metal pushed to the seat separation area is considered to evapo-
2.2.1 Basic concept of the coated steel sheet model rate. This causes the virtually enlarged diameter of the region where
When a coated steel sheet is spot-welded, the nugget formation currents are flowing to be reduced to the actual contact region diam-
is generally slower than that on uncoated steel sheets. The reason for eter, as shown in Fig. 3 (c). Coating metal characteristics necessary
this is that the coating metal on the steel sheet surface melts due to for this model are the melting point, boiling point, and film thick-
the temperature increased by the applied currents, and then is ness (or the coated amount and density).
pushed out of the contact area, increasing the diameter of the section For the interface between an electrode and sheet, although not
where the currents flow. Furthermore, when the steel sheet tempera- shown in the figure, a completely identical model can be applied.
ture increase progresses to reach the boiling point of the coating For the coating layer and the virtual continuity region, no FEM
metal, the pushed molten coating metal evaporates. This is likely to mesh is defined; the entire coated steel sheet model is analytically
be the mechanism for the disappearance of the phenomenon of di- processed.
ameter increase of the current flowing section. Considering the 2.2.2 Validity verification of the model
above, the analysis model of the coated steel sheet was created. In order to verify the validity of this model, test pieces were pre-
Figure 3 shows schematic views that illustrate the basic concept pared in a laboratory by electrogalvanizing one side of a 1.2 mm-
of the model from the aspect of the interface between sheets. First, a thick 590 MPa precipitation hardened steel sheet using pure zinc in
region where the temperature is above the coating metal melting the amount of 90 g/m2. Using the test pieces, the nugget formation
behavior when undergoing spot welding in the following two cases
was examined: One was a case in which a sheet assembly was con-
ducted by welding two test pieces, with the coated surfaces of the
test pieces overlapping each other (hereinafter referred to as “C/C”);
and the other was a case of a sheet assembly with the same two test
piece structure as C/C, but with the coated surfaces of the test pieces
not overlapping (hereinafter referred to as “N/N”). The two assem-
blies were welded at an electrode force of 2.45 kN while applying
electric current at 7.2 kA for 233 ms (including the up-slope control
for 50 ms). The electrodes used were chromium copper electrodes
of the DR type with 6R40 diameter tips.
Figure 4 shows a graph plotting the comparison results between
the nugget growing histories obtained from the experiment and cal-

Fig. 3 Modeling of coated sheet

(a) Calculation of volume of melted metal, (b) Enlarging current path, Fig. 4 Comparison of nugget growth history between experiments and
and (c) Restoring current path calculations for coated sheets

- 105 -
NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No. 119 September 2018

culation. The sheet assembly of N/N showed growth behavior simi- However, distinguishing the electrode force specified as welding
lar to that of non-coated steel sheets. The sheet assembly of C/C conditions and the force from the molten area dependent on the nug-
showed slower nugget growth. The analysis results are in agreement get size but not affected by axial misalignment is not necessary be-
with the above experimental results. Figure 5 shows an excellently- tween the analysis and actual welding. Accordingly, the following
reproduced difference in the cross-section of each sheet assembly can hold.
obtained in the middle of welding between the measurement result FS = FS* (5)
and calculation result. From the results described above, the validity FN = FN* (6)
of this coated sheet model was confirmed. In conclusion, by applying Eq. (4), Eq. (5), and Eq. (6) to the differ-
2.3 Spatter prediction model ence between Eq. (2) and Eq. (3), the following formula is obtained
This subsection describes the development of a prediction model as a relational expression that should hold in an analysis under the
of spatter (so-called expulsion), which occurs from the interface be- condition of actual spatter occurrence.
tween sheets. First, force balance is considered for a steel sheet that FC = (1 − k) . FS (7)
is being welded as shown in Fig. 6. Assuming that the force from an In order to predict the generation of spatter by determining that
electrode is FE, the force from the molten area (nugget) is FN, and Eq. (7) holds, two conditions need to be satisfied: One is that the
the force from the steel sheet in front of the electrode is FC, the fol- history of FC is obtained as the analysis result; and the other is that
lowing force balance equation in the sheet thickness direction is sat- the value of coefficient k is already known. For the condition regard-
isfied. ing FC, the total of the contact pressure values at the corona bond
FN + FC = FE (1) shown in Fig. 7 in all steps of the coupled analysis should suffice.
Under an ideal condition without axis misalignment, i.e., in an The total can be evaluated using the following formula:
analysis in which the axial symmetrical state constantly holds, FE in rC
FC = 2π ∫ r r . pC ( r ) dr (8)
Eq. (1) is equal to electrode force FS. Given this, the following is re- N

written as Eq. (1). where r is the distance from the nugget center, and rN and rC are the
FN + FC = FS (2) nugget radius and contact radius, respectively. Contact pressure
For actual welding, the balance relationship equal to Eq. (2) holds. pC (r) can be directly obtained as the analysis result. Coefficient k
However, axial misalignment and falling over of an electrode are can be determined by comparing FC of Eq. (8) with the experimental
unavoidable. It is reasonable to consider that force from an electrode result using some welding conditions.
is somewhat smaller than the force applied to the electrode (i.e., Figure 8 shows the determination procedures. First, an experi-
electrode force). Given this, the following is the actual force balance ment involving changing currents for a certain value of electrode
condition, force FS is conducted to examine the time when spatter occurs in
FN* + FC* = k . FS* (3) each current based on the change in resistance during welding. Next,
where k is a coefficient of 1 or less, and the subscript “ * ” is provid- analysis is conducted under the same conditions as those used in the
ed for distinction from the above ideal state in the analysis. Since experiment to obtain a history of FC. From these results, FC value at
the contact force between sheets should be zero when actual spatter the spatter generation time is read to apply to Eq. (7) indicating the
occurs, the following formula can be used for obtaining spatter
causing conditions.
FC* = 0 (4)

Fig. 7 Contact force FC in the simulation model

Fig. 5 Cross sections of weldments after 100 ms

Fig. 8 Evaluation of parameter k in eq. (7) by plotting measured spatter

Fig. 6 Force balance in a welded sheet in idealized conditions time on the history of calculated FC

- 106 -
 NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No. 119 September 2018

condition for spatter generation. Then, k for electrode force FS is ob-

tained.
Coefficient k by its nature should be affected by a welding ma-
chine difference or experimental variation. Despite that, in the ex-
perimental scope performed by us, the value was around 0.9 for any
electrode force. Therefore, empirically, Eq. (7) for the spatter occur-
rence condition determination is further simplified as the following
formula:
FC = 0.1FS (9)
The validity verification for the spatter occurrence prediction model
introduced as described above is explained in the next section.
2.4 Weld lobe prediction
In order to verify the validity of the spatter occurrence prediction
model obtained as described in the previous section, a weld lobe
was predicted. A weld lobe is a distribution chart of the nugget di-
ameter when the horizonal axis is the current and the vertical axis is
the electrode force or time period during which currents are applied.
The chart indicates the appropriate welding condition range. The
materials used for the study were three types of steels: Mild steel,
590-MPa DP steel, and 980-MPa DP steel. Welding of two 1.2 mm-
thick sheets was tested and analyzed. The conditions were varied
with four levels of electrode force (1.47, 2.45, 3.43, and 4.41 kN)
and with currents changed by 0.5 kA in the range from 3.0 to 11.0
kA. The time period of current application was 233 ms for all elec-
trode force levels. The electrodes were of the DR type, the same as
those described in 2.2.2.
Figure 9 shows charts of the weld lobes experimentally obtained
for the three steel types, each compared with the weld lobe predicted
through the analysis. The predicted weld lobes using solid lines
were represented by regions surrounded by the contour line of the
DN value of which the nugget diameter (DN) is 4 √ t (t: sheet thick-
ness) and the contour line of FC value satisfies Eq. (9). For each
steel type, the prediction weld lobe favorably agrees with the shaded
experimental result. The validity of the spatter occurrence prediction
model in this study was thus confirmed.

3. Study on the Influence of Various Factors on the

Spot Weldability at the Door Opening
3.1 Analysis model and analysis conditions
Assuming spot welding at the door opening, an assembly con-
sisting of three sheets of side panel outer (hereinafter referred to as
“SPO”), reinforcement (hereinafter referred to as “REINF”), and A
pillar inner (hereinafter referred to as “APLR-INN”) is considered
in this chapter. Such a sheet assembly is regarded as one of the most
difficult sections for spot welding in an automotive body due to the
high sheet thickness ratio that is generally 4 or more. Another reason
is that there is concern about sheet gaps caused by less-accurate
press forming of the member or less-accurate assembling with other Fig. 9 Comparison of weldability lobes between measurement and pre-
dicted
components, as well as the concern about current shunting caused
by neighboring existing welding points.
Figure 10 shows the analysis model simulating the door open- front window and one on the flange in the same position as the
ing, while Table 1 shows the sheet assembly specification. In order welding point to be added for the study at intervals of 25 mm, as
to simplify the study, members should be straight. As shown in Fig. shown in Fig. 10 (b). At the existing welding points, mesh vertices
10 (a)(b), an 80-mm range in the longitudinal direction was deter- were shared in the range of a 5-mm in diameter circle to mechani-
mined as the analysis region due to the limited analysis scale. How- cally constrain the three sheets, while allowing currents to flow (cur-
ever, in order to consider the influence of current shunting due to the rent shunting). For the interfaces between the electrode and sheet
rigidity of the member or existing welding points on the weldability, and between sheets, mechanical contact conditions, contact heat,
the cross-section of the member is modeled in accordance with the and electric resistance were defined. All displacements on the front
actual vehicle body. In addition, as existing welding points, three window edge and out-of-plane displacement on the edge in the lon-
5-mm in diameter welding points were provided on the side of the gitudinal direction were constrained as shown in Fig. 10 (b).

- 107 -
NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No. 119 September 2018

Fig. 10 Finite element model of a door opening

Table 1 Sheet combination

SPO Mild steel / 0.8 mm

REINF 590 MPa-class steel / 1.0 mm
APLR-INN 590 MPa-class steel / 1.8 mm

Table 2 Welding conditions

Electrode force 3.5 kN

Weld time 22 cycles (60 Hz)
Weld current 7.0 kA

The study on the influence of sheet gaps is described later. First,

a DR type electrode with a tip diameter of 6 mm and a tip of R40
mm was vertically placed on the flange on the door opening side of Fig. 11 Nugget growth histories
the model without a sheet gap, and welding was analyzed under the
conditions shown in Table 2. The welding center was determined in
a position in the center of the flange in the longitudinal direction, 6
mm from the flange edge. In the following subsections, the influ-
ence of various factors on the weldability at the door opening is ex-
amined, while comparing the obtained analysis results with the
analysis results of test coupons.
3.2 Influence of current shunting caused by existing welding
points
Figure 11 shows a graph indicating nugget growth histories at
the interfaces with sheets of the door opening model, compared with
a history in the case of single-point welding of a 50 mm × 30 mm
test coupon (Fig. 12) under the same conditions. These results show
that the door opening model is slower in nugget formation, and that
the final nugget diameter is considerably small. The main cause for
Fig. 12 Finite element model of a coupon test piece
the results is considered to be the influence of current shunting
caused by existing welding points. Specifically, as is clear from the
current time history at the welding point shown in Fig. 13, only 60 current input to the door opening model. From this result, the occur-
to 80% of the input currents of 7.0 kA flowed at the welding point rence of current shunting can be confirmed at all modeled existing
on the door opening model. The current shunting can be understood welding points. However, for the three points on the flange on the
from the contour chart of current density. front window side, the amount of currents that were shunted was
Figure 14 shows the current density distribution in the 5th cycle very small, with a large part of the entire shunting currents concen-

- 108 -
 NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No. 119 September 2018

Fig. 15 Finite element model of a coupon test piece with an existing

weld

Fig. 13 Histories of current passing through the welding area

Fig. 14 Distributions of current density at 5 cycles

Fig. 16 Nugget growth histories

trating on the nearest welding points. Therefore, when examining

the influence of current shunting on the weldability, the main focus
should be placed on the nearest welding points.
3.3 Influence of the member’s rigidity
Next, the influence of the rigidity of the member on the weld-
ability is studied. When a test coupon is spot-welded, the steel sheet
sustains out-of-plane deformation due to the electrode force, result-
ing in a gap (sheet separation) between the steel sheets. The size of
a sheet separation influences the contact statuses at the interfaces,
and varies depending on the welding conditions, electrode shapes,
steel type, sheet thickness, etc. In addition, the shape of the material
to be welded, meaning the rigidity of the member, is responsible for
the size of a sheet separation as well. In particular, in a member with
a closed cross-section structure like the door opening, the deforma-
tion during welding may be suppressed, showing weldability differ-
ent from that of test coupons. In this section, the analysis results ob-
Fig. 17 Mean current density
tained from a test coupon model (Fig. 15) with the same sheet as-
sembly and existing welding points as the door opening model are
compared with the analysis results of the door opening model to ex- between the two models. Figure 17 shows the results. As the chart
amine the influence of the rigidity that a member has on the weld- clearly shows, the average current density at the weld area directly
ability. involved in the nugget formation was almost equal to each other.
Figure 16 shows a chart comparing nugget growth histories at The sheet assembly and the shape of the component dealt with in
the SPO-REINF interface between the door opening model and test this study are considered to be the reason for no difference in the
coupon model. Whereas the nugget formation in the door opening nugget formation between the two models.
model is slightly slower, the overall growth of the two nuggets show Furthermore, focusing on the sheet separation, an obvious differ-
only a small difference. For the interface of REINF-APLR-INN as ence is found between the two models. For example, Fig. 18 shows
well, there was very little difference between the two. To determine images of the cross-sections taken in the 5th cycle of current input,
the reason for this, the average current density of the weld area at clearly showing that the sheet separation in the door opening model
the SPO-REINF interface was calculated to examine the difference is smaller. This is very important when dealing with a coated steel

- 109 -
NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No. 119 September 2018

Fig. 18 Deformation and sheet separation with temperature distributions at 5 cycles on the cross section
(a) Parallel to (b) Perpendicular to the longitudinal direction (upper: coupon TP model, lower: door opening model)

Fig. 19 Finite element models of single spot joint with gap

sheet. In other words, coating on the steel sheet surface is partially

melted when currents are being applied; The sheet separation is
filled with the molten coating, leading to an apparent increase in the
diameter of the region where the currents are flowing. As a result, Fig. 20 Nugget growth histories
the current density at the weld area is lowered to render the nugget
growth slower. 3) Since this study only deals with bare steel sheets,
no significant difference has been found in the weldability in the
two analysis models. If the SPO was a coated steel sheet as in the
case of actual vehicles, the difference in sheet separation due to the
difference of member rigidity might have affected the weldability to
no small extent.
3.4 Influence of sheet gaps
When the influence of sheet gaps is evaluated in experiments,
test coupons provided with gaps by inserting spacers between sheets
are used in general. However, sheet gaps in a variety of forms can
be contained in an actual vehicle body, and the evaluation using
such test coupons can cover only a small part of the diverse sheet
gaps.
For example, we assumed a case in which a test coupon model
with supports provided at the interval of 50 mm and gaps of 1 mm
as shown in Fig. 19 (a) and the door opening model with the same
gaps of 1 mm as shown in Fig. 19 (b) were used. The two models
were welded under the conditions shown in Table 2 to compare the
nugget growth of the REINF-APLR-INN interface. The nugget
growth histories of the test coupon model and door opening model
significantly differed from each other as shown in Fig. 20. When
compared with the result obtained using a no-gap test coupon, one Fig. 21 Temperature distributions at 1 cycle
of the two models was faster than it, while the other was slower in
nugget growth; the weldability change tendency thus different from ference in heating as shown in Fig. 21. In addition, if there is a coat-
each other was shown. The reason for this is considered to be the ed steel sheet in the assembly, the influence of the sheet separation
difference in the way the gaps were provided, rendering the contact is present as well as described in 3.3. For this reason, the change in
status at each interface different from each other, resulting in a dif- weldability is considered to be more complicated. As explained

- 110 -
 NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No. 119 September 2018

above, when the influence of sheet gaps on the weldability is stud- tures sustained by the tensile shear joints were nugget interface frac-
ied, it is necessary to examine the exact situation in the case of an tures, while all fractures that the L-shaped joints underwent were
actual vehicle structure. unplugged fractures on the sheet side.
4.2 Creating the fracture prediction FEM analysis model
4. Application to the Prediction of Fracture at Weld 4.2.1 Overview of the analysis model
Areas for Dissimilar Material/Thickness Sheet This section describes the development of the fracture prediction
Assemblies FEM analysis model capable of dealing with the dissimilar material/
4.1 Tensile tests of spot-welded joints thickness sheet assemblies. Figure 23 shows the analysis mesh of
The spot-welded joint tensile tests described in this chapter used the spot-welded area. The shapes of the nugget and the heat affected
five different sheet assemblies; a 270 MPa cold rolled steel sheet zone (HAZ) were obtained using a regression formula using the
(hereinafter referred to as the “270 steel”) that was 2.0 mm in thick- sheet thickness and nugget diameter as parameters. For the deforma-
ness was used for one side, and for the other side sheets of various tion resistance curve and the fracture limit of the base metal and
material and thickness were used. Table 3 shows the five sheet as- HAZ, data for each area of the same material sheet assembly was
semblies together with the tensile test results. The joint shapes were used. For the nugget made up of two different materials, the data of
a tensile shear joint and an L-shaped joint as shown in Fig. 22. The the same material sheet assembly could not be used as-is. For the
target nugget diameter value at the weld area was 4 √ t (t: Thickness nugget data, ultra-small test piece tensile test results 7) were used.
on the sheet side). The tensile tests were performed under the static However, from the viewpoint of simplification and extensibility, the
condition of 10 mm/min. The maximum load shown in Table 3 is study described in this chapter used a linear mixing approximation
the average value of the results of three tests. Fractures were defined formula and a prediction formula for the calculation.
into two types: One was those mainly constituted by shear at the 4.2.2 Calculation method of the deformation resistance curve and
center of the nugget (hereinafter referred to as the “nugget interface fracture limit value
fracture”); and the other was those involving the nugget unplugging The deformation resistance curve was approximated by
fracture (hereinafter referred to as the “unplugged fracture”) resulted Ludwik’s equation. As shown in Eq. (10), in the dissimilar material/
by crack growth around the nugget outer circumference. All frac- thickness sheet assembly, the flow stress of the nugget was calculat-
ed by multiplying Ludwik’s equation of each material by the mixing
Table 3 Combination of steel sheets for spot welded joints, and corre- ratio of the molten area (R(270): 270 steel [2.0 t], R(other): Mating sheet
sponding experimental joint strength [1.2 t]). The fracture limit value was expressed by the Eq. (11) pow-
er function, and their constants a and b were obtained using Eq. (12)
TS joint LT joint
and Eq. (13) with chemical components as parameters. From a1 to
Combination of Joint Joint a10 and from b1 to b10 of Eq. (12) and Eq. (13) are the constants
No. Fracture Fracture
steel sheets strength strength derived using the least-squares method from the fracture limit value
mode mode
(MPa) (MPa) data on the five steel types obtained using the tensile test technolo-
270-steel 270-steel gy 7) of ultra-small test pieces. Regarding the dissimilar material/
A1 8.77 Shear 2.11 Pull out
(2.0 t) (1.2 t) thickness sheet assembly, the chemical composition of the nugget
270-steel 440-steel was estimated from the mixing ratio of the molten area. Equation
A2 10.47 Shear 2.00 Pull out (14) is a calculation example of the C content. As described above,
(2.0 t) (1.2 t)
270-steel 590-steel for a dissimilar material/thickness sheet assembly, the calculation
A3 9.80 Shear 2.52 Pull out method of the molten mixing ratio is important. In this chapter, the
(2.0 t) (1.2 t)
molten mixing ratio was obtained from the results of the sheet thick-
270-steel 980-steel
A4 11.71 Shear 2.40 Pull out ness ratio or spot welding simulation (nugget formation analysis).
(2.0 t) (1.2 t) Then the deformation resistance curve and the fracture limit value
270-steel Boron-steel calculated using respective methods were applied. Lastly, the pre-
A5 11.69 Shear 2.28 Pull out
(2.0 t) (1.2 t) diction accuracy of the FEM analysis was compared.
σwm = R(270) (σy(270) + F(270) ε n(270))
+ R(other) (σy(other) + F(other) ε n(other)) (10)
σwm: Flow stress of the nugget, σy(270): σy of the 270 steel
σy(other): σy of the mating sheet, ε: Plastic strain
F(270), n(270), F(other), n(other): Constants

Fig. 23 Finite element mesh of spot welded joint (different materials

Fig. 22 Tension test specimens of spot welded joint and thickness)

- 111 -
NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No. 119 September 2018

εCR = a σtriaxb (11) Table 4 Comparison of spot weld simulated results

(εCR: Fracture strain, σtriax: Stress triaxiality) Nugget Mixing ratio

Combination of
a = 1 − a1 . C − a2 . Si − a3 . Mn − a4 . P − a5 . S No.
steel sheets
diameter
R(270) R(other)
− a6 . Mo − a7 . Cr − a8 . B − a9 . Ti − a10 . Nb (12) (mm)
. . .
b = − 0.5 − b1 C − b2 Si − b3 Mn − b4 P − b5 S . . 270-steel 270-steel
A1 4.47 0.692 0.308
− b6 . Mo − b7 . Cr − b8 . B − b9 . Ti − b10 . Nb (13) (2.0 t) (1.2 t)
270-steel 440-steel
(a1–a10, b1–b10: Constants) A2 4.50 0.630 0.370
(2.0 t) (1.2 t)
C = R(270) . C(270) + R(other) . C(other) (14) 270-steel 590-steel
(C(270): C content of 270C, C(other): C content of the mating sheet) A3 4.50 0.582 0.418
(2.0 t) (1.2 t)
4.2.3 Calculation using the thickness ratio 270-steel 980-steel
All joints subject to the evaluation were assemblies of the 270 A4 4.28 0.552 0.448
(2.0 t) (1.2 t)
steel (2.0 t) and mating sheets (1.2 t). Therefore, if the sheet thick- 270-steel Boron-steel
ness ratio is used, R(270) = 0.625 and R(other) = 0.375 were determined A5 4.07 0.528 0.472
(2.0 t) (1.2 t)
from Eq. (15) and Eq. (16).
2.0
R(270) = — (15)
2.0 + 1.2
1.2
R(other) = — (16)
2.0 + 1.2
4.2.4 Use of spot welding simulation
With the welding conditions used in this experiment as input pa-
rameters, the molten area mixing ratio was calculated using this
technology. Nugget volumes V(270) and V(other) respectively used the
maximum liquid phase ratio during heating, and the nugget diameter
was determined to be in a range with the maximum liquid phase of
0.8 or more. Molten area mixing ratios R(270) and R(other) were ob-
tained using Eq. (17) and Eq. (18). Table 4 shows the nugget diam-
eter and the molten area mixing ratio obtained from the nugget for- Fig. 24 Tension test specimens of spot welded joint (unit: mm)
mation analysis. Figure 24 shows the cross-section of sheet assem-
bly No. A4 weld joint and the nugget formation analysis result (liq-
uid phase ratio distribution). The nugget diameter obtained as the
analysis result corresponds to the target value (4 √ t = 4.38 mm), and
the molten area mixing ratio is considered to be a reasonable value.
The molten area mixing ratio of sheet assembly Nos. A1 and A2
with the mating sheet (1.2 t) being 270 steel or 440 steel generally
accorded with the case using the sheet thickness ratio.
Contrary to that, sheet assemblies Nos. A3, A4 and A5 showed
different values from the case using the sheet thickness ratio. Since
high tensile strength steel has a higher electricity resistivity at a tem-
perature range between room temperature to 1 000°C than mild
steel, growth of the nugget is faster and the molten area mixing ratio
is higher. 1) The analysis result agreed with the tendency. When using
hot stamp material for the mating sheet, the molten area mixing ra-
tios R(270) and R(other) of sheet assembly No. A5 were almost equiva-
lent. Fig. 25 Flow curves
V(270)
R(270) = — (17) almost no difference between the spot welding simulation result and
V(270) + V(other)
the calculation result using the steel thickness ratio. For this reason,
V(other)
R(other) = — (18) the deformation resistance curve and the fracture limit line almost
V(270) + V(other) accorded with each other as well. Contrary to this, on the results of
4.2.5 Deformation resistance curve and fracture limit value of the Nos. A3, A4 and A5, the difference in the molten area mixing ratio
nugget exercised influence; Figure 25 shows a chart plotting their deforma-
With the sheet thickness ratio and through the spot welding sim- tion resistance curves, while Fig. 26 shows a chart plotting their
ulation (nugget formation analysis), the molten area mixing ratios fracture limit values. The solid lines and broken lines in gray are the
(R(270) and R(other)) were calculated, respectively. Then using Eqs. (10) calculation results using the sheet thickness ratio, and the solid lines
to (14), the deformation resistance curve and the fracture limit value and broken lines in black are the spot welding simulation results.
of the nugget were obtained for each sheet assembly. As shown in The calculation results from the spot welding simulation of Nos. A3,
Table 4, the molten area mixing ratios of Nos. A1 and A2 involved A4 and A5 showed larger deformation resistance and lower ductility

- 112 -
 NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No. 119 September 2018

Fig. 26 Spot weld fracture limit curves in FEM analysis model

for the fracture limit. As the strength of the mating sheet (1.2 t) was
higher, such tendency became significant.
4.3 Analysis result
Figure 27 shows bar charts indicating the comparison results
between the experimentally measured joint strength and the joint
strength calculated through the FEM analysis. For the FEM analysis
results, further comparison between the analysis result of the defor-
mation resistance curve and fracture limit value of the nugget using
the sheet thickness ratio (hereinafter referred to as “analysis result
(1)”) and the analysis result of the same (hereinafter referred to as
“analysis result (2)”) using the spot welding simulation was con-
ducted.
While analysis result (1) showed favorable accord with the ex-
perimental results regarding the tensile shear joint, regarding the L-
shaped joint, the difference from the experimental results increased
as the mating sheet (1.2 t) strength became higher; With sheet as- Fig. 27 Comparison of joint strength
semblies Nos. A4 and A5, the difference from the experimental re-
sults reached 10% or more. In analysis result (2), even the L-shaped
joint result showed a less than 10% difference from the experimen-
tal results. Compared with analysis result (1), the prediction accura-
cy improvement effect for sheet assemblies Nos. A4 and A5 was
significant. The fracture forms in both analysis results (1) and (2)
were the same as that in the experimental results. As an example,
Fig. 28 comparatively shows sheet assembly No. A4 in cross-sec-
tion photographs taken during the experiment and in simulation im-
ages of analysis result (2). On the tensile shear joint, a fracture oc-
curred at the nugget interface; on the L-shaped joint, the fracture
that occurred was an unplugged fracture. If the mating sheet (1.2 t)
strength was high, the fracture started inside the nugget. This was
reproduced in the FEM analysis. The results above indicate that the
material characteristic values at the molten area of a dissimilar ma-
terial/thickness sheet assembly that includes a high tensile strength
steel sheet greatly affect the prediction accuracy, making it impor-
tant to choose a suitable estimation method.

5. Conclusion
In this paper, we described the initiatives to develop a highly ac-
curate spot welding FEM analysis system based on an incrementally
coupled analysis method between the electric, temperature, and
stress fields considering the phase change and phase transformation.
The technology achieved high accuracy prediction of the nugget
growth for any sheet assembly under any welding conditions. In ad-
dition, the technology actually considered melting and evaporation Fig. 28 Comparison of fracture mode (No. A4)

- 113 -
NIPPON STEEL & SUMITOMO METAL TECHNICAL REPORT No. 119 September 2018

of the coating, and predicted spatter occurrence involving the mol- steel sheets, sheet assemblies are becoming more complicated, mak-
ten metal being spattered from inside as well. This has allowed the ing it more difficult to determine the spot welding conditions. In ad-
theoretical examination of the weld lobe of an actual sheet assembly dition, since it is becoming increasingly important to study the
including a coated steel sheet. Since the studies were started with welding mechanism under disturbance conditions such as sheet
single-spot welding of test coupons as the evaluation subject, axial gaps, electrode axial misalignment, etc., expectations for the effect
symmetry models were mainly examined. However, due to the need of the use of the analysis technology are increasing.
for considering current shunting, sheet gaps, electrode axial mis-
alignment, etc., the study subject was switched to 3D models. References
As an example, the spot welding simulation technology was ap- 1) Fukumoto, M. et al.: SAE paper. 2003-01-2806, 2003
2) Fukumoto, M. et al.: Preprints of the National Meeting of JWS. No.72,
plied to the member geometry modelling a door opening. An exami-
2003, p.60
nation method of the influence of current shunting, the rigidity of 3) Fukumoto, M. et al.: 2006 JSAE Annual Congress (Spring) Proceedings.
members, and sheet gaps on weldability was proposed; the method No.74-06, 2006, p.9
involved a comparison of the results from single point welding of 4) Fukumoto, M.: Proceedings of National Symposium on Welding Me-
test coupons. Furthermore, as application to the evaluation of weld chanics & Design 2011. 2011, p.267
5) Ueda, H. et al.: Journal of Society of Automotive Engineers of Japan. 44
area strength, the deformation resistance and fracture limit value (2), 727 (2013)
were calculated from the molten area mixing ratio of the nugget ob- 6) Ueda, H. et al.: Journal of Society of Automotive Engineers of Japan. 46
tained through the spot welding simulation. This was able to con- (3), 687 (2015)
tribute to an accuracy improvement of fracture prediction FEM 7) Nakayama, E. et al.: Journal of Society of Automotive Engineers of Ja-
pan. 36 (1), 205 (2005)
analysis modelling of dissimilar material/thickness sheet assemblies.
Along with the progress in strength improvement of automotive

Hideki UEDA Shota KIKUCHI

Senior Researcher Researcher
Materials Reliability Research Lab. Materials Reliability Research Lab.
Steel Research Laboratories Steel Research Laboratories
1-8 Fuso-cho, Amagasaki City, Hyogo Pref. 660-0891

Manabu FUKUMOTO Eisuke NAKAYAMA

Senior Manager Senior Manager, Dr. Eng.
Research Administration Dept. Research Planning Dept.
R & D Planning Div. Steel Research Laboratories

Hiroki FUJIMOTO Tohru OKADA

Senior Manager, Dr. Eng. Senior Manager
Research Planning Dept. Human Resources Div.
Steel Research Laboratories

Kazuo OKAMURA Masanori YASUYAMA

Fellow, Doctor (Energy Science) Chief Researcher
R & D Laboratories Welding & Joining Research Lab.
Steel Research Laboratories

- 114 -