Chawki 01
Chawki 01
Chawki 01
Chawki Tahri, Christophe Bertoni, Eric Feulvarch, Helmut Klocker & Jean-
Michel Bergheau
To cite this article: Chawki Tahri, Christophe Bertoni, Eric Feulvarch, Helmut Klocker & Jean-
Michel Bergheau (2020): Numerical simulation of the resistance braze welded assembly of
a copper Inconel 601 ground electrode and a steel shell, Numerical Heat Transfer, Part A:
Applications, DOI: 10.1080/10407782.2020.1777776
1. Introduction
Lower emissions, lower fuel consumption, and higher ignition voltage, with respect to the thermal
loads: requirements continue to grow for automotive industries. The main function of an ignition
system inside an engine remains the same: ignite the air/fuel mixture at the right time with the
appropriate ignition energy. In gas engines, the combustion of air/fuel mixture is started by spark
plugs which produce, for a 4 cylinder engine with 3000 rpm, more than 6000 sparks per minute
inside the combustion chamber [1]. A spark plug is made mainly by a ground electrode and a
shell. To withstand high temperatures, the ground electrodes are formed by a copper core and a
high temperature corrosion resistant Nickel-based alloy mantel. Inconel 601 was chosen for its
exceptional resistance to oxidation and corrosion at high temperatures. It was proved by Buscail
et al. [2] that the Inconel 601 forms a protective oxide layer, and retains much of its strength at
elevated temperatures. The copper core used in this study is a Cu-OF high purity copper [3].
Both materials are tamped, cold worked, annealed and finally welded to a steel shell (FR10 [4])
using a resistance braze welding technique, which consists of three principle phases called accost-
ing, welding, and hold as shown in Figure 1a. This technique is frequently used in automo-
tive [5–9].
This article deals with the resistance braze welding of a bimetallic ground electrode (GE) made
of copper coated with Inconel 601, and a shell made of steel, the chemical compositions of which
CONTACT Chawki Tahri Chawki.tahri@enise.fr University of Lyon, ENISE, LTDS, UMR 5513 CNRS, 58 Street Jean Parot,
42023 Saint-Etienne, France.
ß 2020 Taylor & Francis Group, LLC
2 C. TAHRI ET AL.
are given in Table 1. The ground electrode (GE) may be welded to the shell using an AC current
with two frequencies of 50 Hz and 5000 Hz, or a DC current. The resistance braze welding pro-
cess in this case involves electromagnetic, thermal and mechanical phenomena and their interac-
tions. In addition, all the physical properties are strongly temperature dependent (Tables 2–4).
Experimental optimization of the process requires prototyping and measurements which are
extremely expensive and time consuming, and finally, very few solutions can be used. For these
reasons, Finite Element simulations can save time and reduce cost-spent on the development of
products or processes. Furthermore, the simulations give more detailed information of the differ-
ent phenomena involved and thus lead to a better understanding of the problem. In this article,
the commercial finite element code SysweldV R was used to simulate the process.
For the numerical simulation of the resistance braze welding using low and high frequencies,
and for complex geometries (presence of spacing between copper and Inconel 601), the literature
is not very abundant. Resistance welding has been addressed largely in the literature to the resist-
ance spot welding. The first numerical model of the resistance welding process was developed by
Greenwood in 1961 [10]. The goal was to study the thermal history and the formation of the
molten zone, and the Heat Affected Zone (HAZ). Years later, Rice and Funk [11] established a
finite difference model to study the thermal history, but also the influence of pulsation on the
formation of the weld nugget. The materials studied were nickel, kovar, and stainless steel. All
physical data were temperature dependent. An empirical relationship between temperature and
sheet/sheet contact resistance had been proposed in this work. Kaiser et al. [12] continued the
work of Rice and Funk [11] to evaluate contact resistance. For this purpose, two sheets were used
(bare sheets of HSLA steel and low carbon steel). The results show that heating is faster in HSLA
steel than in low carbon steel, so pointing the influence of the electrical resistivity and the ther-
mal conductivity of the steels. The first finite element model for the resistance spot welding was
developed by Nied [13]. He used the commercial finite code ANSYS and introduced a quarter
axisymmetric model. The model was able to predict the electrode and the workpiece deformations
NUMERICAL HEAT TRANSFER, PART A: APPLICATIONS 3
as well as the stress distributions along the interface. Six years later, Dickinson et al. [14] used
ANSYSV R to model the resistance spot welding. They found that for dissimilar materials, the nug-
get forms in the low conductivity workpiece rather than that with higher thermal conductivity.
For an unequal thickness sheet, the nugget forms in the thicker workpiece. This is due to the dis-
tance traveled by current. Thieblemont et al. [15] introduced the first electrokinetic-thermal
model using SysweldV. In order to take into account the important interactions between electrical
R
and thermal phenomena, the problem being coupled by the Joule effect on the one hand and the
dependence of physical properties with temperature on the other, with a strong coupling. In his
study, the electrical and thermal equations are processed simultaneously [15]. Tsai et al. [16]
found that the electrode displacement and its velocity are important parameters in the process.
Vogler and Sheppard [17] studied the temperature history in resistance spot welding. An elec-
trical contact resistance was modeled. They found that the electrical contact resistance affects the
thermal history and the final nugget size. Due to the lack of information, electrical and thermal
contact resistances are not taken into consideration here. Recent research performed by Xu et al.
4 C. TAHRI ET AL.
[18] reported that the heat transfer coefficient at the interface has a significant effects on the nug-
get formation and thermal distribution in the workpiece.
In the present work, electrical and thermal contact resistances are not taken into account.
Indeed, as it can be seen on Figure 1b, the welding phase is preceded by a preheating which
ensures excellent contact condition between the ground electrode and the shell. Perfect electrical
and thermal contact conditions are then assumed.
The presence of eddy currents, which depends on the type of current AC or DC, or the pres-
ence or not of gaps between copper core and Inconel 601 coating, strongly influence the distribu-
tion of currents during the process. Thereby changing the Joule effect along the contact interface,
thus consequently strongly influence the quality of the welded achieved. The study of the resist-
ance braze welding of a composite solids of different materials, and a shell made of steel, with a
DC current, and an AC current with different frequencies and geometries (with and without
gaps) will constitute the contribution of this article. The main question in this article is: how vary
the distributions of temperature using a DC and an AC currents with two different frequencies
(50 Hz and 5000 Hz), and considering the presence or not of gaps between copper and Inconel 601,
during welding? Different sections will lead to an answer for this question; Section 2 details the indus-
trial configurations. Section 3 presents the models used for the simulation of welding of the ground
electrode made of copper coated with Inconel 601 and the shell made of steel. All the phenomena
and their interactions, which need to be taken into account are detailed. Two different models are
presented. The first one is a 3-D non-linear magnetodynamic-thermal model, specially suited for AC
currents. The magnetodynamic and thermal Finite Element analyses are weakly coupled to calculate
the temperature distributions during the process. The second model dedicated to DC currents, enables
the calculation of the temperature distributions during the process through a fully coupled 3-D elec-
trokinetic-thermal modeling. The temperature distributions are then used to simulate the ground elec-
trode collapsing. Section 4 presents and discuss the computed results.
2. Industrial configurations
2.1. Industrial process
The ground electrode, the subject of this study, is made of copper coated with Inconel 601 which
their chemical compositions are given in Table 1. The copper core used in this study is a Cu-OF
high purity copper characterized by a minimum of 99.95% copper [3]. Copper is inserted using
an operation called tamping proceeded by an operation called cold-forming in the aim to obtain
the desired form of the ground electrode (Figure 2). The ground electrode is then passed by an
annealing process under a temperature of 1040 C, during 20 minutes, in a protected atmosphere
containing hydrogen and nitrogen. At the end, the ground electrode is welded to a shell made of
steel. The welding process is composed mainly by three phases which are accosting, welding, and
hold as shown in Figure 1b. The welding phase itself is composed by three sub-phases (Figure
1b). The heating sub-phase’s aim is to ensure a better contact between different parts. Thus, at
the end of this sub-phase, contact is supposed perfect at the interface between the ground elec-
trode and the shell (No contact resistance). The resistance braze-welding is then modeled starting
from the welding sub-phase as shown in Figure 1b.
to 160 kV) under a magnification of 2000 on 10 samples. Table 5 shows the values of gaps
NUMERICAL HEAT TRANSFER, PART A: APPLICATIONS 5
Figure 2. Typical cross section after tomographic observation of the ground electrode.
determined by tomographic measurements on five samples (S1 to S5), in Z1 (Figure 2), after
annealing and cooling to room temperature for both types of samples (with and without gaps).
The presence of such gaps can change the distribution of currents in the workpiece, thereby chang-
ing the Joule effect along the contact interface and consequently strongly influencing the quality of the
welded achieved. Thus, two extreme cases will be studied; with an initial gap, prior to welding, of
40 mm (experimental measurement), and without an initial gap between copper core and Inconel 601.
Where H is the magnetics field, J is the electric current density, E, the electric field and B,
the magnetic induction. Equation (1a) is the Maxwell-Ampere law, Eq. (1b) the Maxwell-Faraday
law, and (1c) the Maxwell-Gauss law for magnetism.
Equations (1) must be completed by behavior law. The Ohm’s law links the electric current
density to the electric field:
J ¼ rE (2)
Where r is the electric conductivity that depends on temperature. The magnetization law
relates the magnetic induction B to the magnetic field H: The Frohlich-Kenelly model is used in
the present work (Chaboudez et al. [21]):
B0
B ¼ l0 þ H (3)
kHk þ Ha
Where l0 is the magnetic permeability of void and B0 and Ha are two temperature dependent
adjustable parameters. lðkHk, T Þ ¼ l0 þ kHkþH
B0
a
is the material magnetic permeability and invert-
ing Eq. (3), one can define the magnetic reluctivity ðkBk, T Þ such that H ¼ ðkBk, T Þ B:
For the magnetization law, a saturation curve is defined at each temperature and the character-
istic value is linearly interpolated between temperatures (Table 6) using the magnetization law
defined previously.
Introducing the magnetic vector potential A from Eq. (1c), it follows:
B ¼ curl A ð4aÞ
div A ¼ 0 ð4bÞ
Where Eq. (4b) is the Coulomb’s gauge introduced for the unicity of A: From Eqs. (1b) and
(4a), it follows:
Combining Eqs. (1a), (1b), (2), and (4), one obtains the (A,V) formulation of the magnetody-
namic problem:
8
>
> @A
>
> r þ gradV þ curlð:curlAÞ ¼ 0 ð5aÞ
>
> @t
<
@A
>
> div r þ gradV ¼0 ð5bÞ
>
> @t
>
>
: div A ¼ 0 ð5cÞ
One can note that the partial derivative Eqs. (5a) and (5c) hold in the whole space while Eq.
(5b) only holds in conductive media. But in order to bound the space domain of study, boundary
conditions can be applied. The boundary conditions associated with Eqs. (5) are:
The Finite Element formulation of Eqs. (5) with associate boundary conditions can be found
in [22] and [23] and will not be recalled here. Each node of the mesh supports 4 degrees of free-
dom (V and the 3 components of A) in conductive media and 3 in the other media (air).
For taking account of the air surrounding the workpiece, several methods can be used. One
method consists in meshing the air around the workpiece. But the air domain is physically
unbounded and the mesh is. It is therefore recommended to take air domain dimensions ten
times larger than the dimensions of the workpiece [20] which strongly increases the number of
nodes of the problem, or to use special finite element formulations based on space transforma-
tions in order to take “infinite-medium” conditions into account [24]. In addition, as far as non-
linear magnetodynamic applications are concerned, the classical finite element method requires to
assemble element vectors and matrices associated with the air at each iteration of a Newton pro-
cess, though the air is a linear medium. And meshing the air can also be a heavy task.
In the present work, in order to avoid meshing the air, we propose to use a method coupling
finite elements in the conductive media and boundary elements in the air. Such an approach has
been proposed by Fetzer et al. [25], Bergheau and Potier [26], Pascal et al. [27], Bergheau et al.
[28, 29], and is applied here to the resistance braze welding simulation. The benefits of the
method lie in the fact that the boundary operator is calculated once at the beginning of the calcu-
lation and that only the workpiece needs to be meshed for both magnetodynamic and thermal
analyses. Besides, the air in this method is taken into account as an infinite medium. The main
drawback of the method is that the FEM-BEM coupling matrix is a non-symmetric big matrix
linking all the nodes of the workpiece boundary that may thwart a part of its benefits.
dH J2
q divðkgradT Þ ¼ 0 (6)
dt r
Where q is the volumic mass, H, the enthalpy, k, the heat conductivity and T, the tempera-
2
ture. In Eq. (6), Jr represents the volumic density of power dissipated through Joule effect.
Boundary conditions can be:
T ¼ T p on @XT
kgradT:n ¼ qðT, t Þ on @Xq
But as the welding process involves high temperature, more complex boundary conditions can
also be applied such as for the modeling of radiation exchanges including multi-reflection phe-
nomena between all the parts of the device (workpiece … ) [26].
8 C. TAHRI ET AL.
Mechanical contact is defined between interacting faces using an algorithm which strictly enforces
a non-penetration condition. Perfectly sliding conditions are assumed at the interface. For ther-
mal contact, a geometric search is performed to find the element on the master surface onto
which each element on the slave surface is subjected. Heat transfer across the gap is based on the
gap conductibility, which can be calculated using a variety of models. The simplest model defines
the gap conductibility as the inverse of the gap distance g multiplied by the thermal conductivity
of the material in the gap, which is air in our case, ðkg ):
hgap ¼ kg =g (9)
10 C. TAHRI ET AL.
Figure 5. Considered geometries (a) 3-D geometry, (b, c) overall dimensions, and (d) geometry of the ground electrode with
and without initial gap between the copper and Inconel 601.
The chemical compositions of different materials used in this study are presented in Table 1.
All the mechanical, thermal, and electrical properties, which are temperature dependent, needed
are determined from literature and presented in Tables 2–6.
Figure 7. Loads applied for (a) magnetodynamique and electrokinetic, (b) thermal, and (c) mechanical analysis.
Figure 8. Illustration of the interface L between the ground electrode and the shell, and the study zones S1, S2, and S3.
Figure 9. Temperature distributions ( C) for (a) MDT-U-5000, (b) MDT-I-5000, (c) MDT-U-5000-G, and (d) MDT-I-5000-G.
Figure 10. Comparison of temperature evolutions ( C) between (a) MDT-U-5000 and MDT-I-5000, (b) MDT-U-5000-G and MDT-I-
5000-G, (c) MDT-U-5000 and MDT-U-5000-G, and (d) MDT-I-5000 and MDT-I-5000-G.
14 C. TAHRI ET AL.
Figure 11. Temperature distributions ( C) for (a) ECT-U-50, (b) ECT-U-50-G, (c) ECT-I-50, and (d) ECT-I-50-G.
to S3) were defined at the interface (L) between copper and Inconel 601 (Figure 8). Figure 10a
shows a comparison of temperature evolutions in these areas over time between MDT-U-5000
and MDT-I-5000. In the absence of a gap, heating appears to be more homogeneous in intensity
control. The temperature difference between the copper core and the Inconel 601 coating is less
important. In addition, the time required to reach the melting temperature of copper is almost
the same for both control modes. Figure 10b shows a comparison of temperature evolutions in
the three previously defined time zones between MDT-U-5000-G and MDT-I-5000-G. In the
presence of a gap, a difference in heating was detected between the copper core and the Inconel
601 coating in both cases. This reflects the heterogeneity of heating. Figure 10c shows a compari-
son of temperature evolutions in the three zones as a function of welding time for MDT-U-5000
and MDT-U-5000-G. This shows the influence of the presence of gaps on the heating distribution
in the weld area. Figure 10d shows a comparison of temperature evolutions versus time for
MDT-I-5000 and MDT-I-5000-G. A significant difference in heating is observed between the cop-
per core and the Inconel 601. It can be concluded that the presence of gaps leads to a different
heating distribution.
To summarize, in the absence of gaps, heating appears to be more homogeneous. The most
homogeneous heating, with a frequency of 5000 Hz-AC, is obtained in absence of gaps, and in
intensity adjustment (MDT-I-5000). Note that non-homogeneity of heating can lead to expulsion
of copper.
Figure 12. Temperature evolutions for (a) ECT-U-50, (b) ECT-I-50, (c) ECT-U-50-G, and (d) ECT-I-50-G.
coupled model taking into account the temporal evolution of the current. Figures 11a, b show the
temperature distributions with ECT-U-50 and ECT-I-50 respectively. In the absence of gaps, the
heating is located in the copper over a wider area in current control than in voltage control. A
difference in heating is observed between the copper core and the Inconel 601 coating. Figures
11c, d show the temperature distributions with ECT-U-50-G and ECT-I-50-G respectively. High
copper temperatures are not seen in comparison to those in the Inconel 601 coating. The gap plays
the role of an insulator, and therefore the total current flows only in Inconel 601, with, consequently,
a more important heating in the latter. This is observed for both voltage and current control modes.
A larger melted area is observed in intensity control in both cases; with and without gap.
Figures 12a–12d represent, respectively, the evolution of temperature in three different areas
(Figure 8), as a function of time and in four different cases; 1 – ECT-U-50, 2 – ECT-I-50, 3 –
ECT-U-50-G, and 4 – ECT-I-50-G. The results show, first of all, the difference in heating
between copper and Inconel 601 when gaps are present. For ECT-U-50 (Figure 12a), higher heat-
ing is observed in copper. The time required to reach the melting temperature of the copper is of
the order of 20 ms, which corresponds to a period of current. For an ECT-I-50 (Figure 12b), the
heating is initially homogeneous. From 20 ms a difference of heating is observed. This shows the
significant energy dissipated after a period of current (20 ms). The copper melting temperature is
reached after 2.5 period (45 ms). For ECT-U-50-G and ECT-I-50-G (Figures 12c, d), the heating
is located in the Inconel 601 coating and the welding time is the same in both cases (40 ms). The
temperature is higher in the Inconel 601 coating than in copper.
To summarize, in the absence of a gap, heating seems to be more homogeneous. The most
homogeneous heating, with a frequency of 50 Hz-AC, is obtained in absence of gaps, and in
intensity adjustment (ECT-I-50).
16 C. TAHRI ET AL.
Figure 13. Temperature distributions ( C) for (a) ECT-U-DC, (b) ECT-I-DC, (c) ECT-U-DC-G, and (d) ECT-I-DC-G.
Figure 14. Comparison of temperature evolutions between (a) ECT-U-DC and ECT-I-DC and (b) ECT-U-DC-G and ECT-I-DC-G.
A difference in welding time is observed, as well as the temperature reached in copper and in
Inconel 601. This reflects the heterogeneity of heating in case of gaps.
Despite the various hypotheses considered here, such as the perfect contact between the copper
core and the Inconel 601, the Inconel 601 and the steel, and the copper and the steel, the results
obtained with an ECT-I-DC (production configuration) show a good agreement with the actual
welding process. The welding time is about 30 ms, which corresponds to the actual time recorded
on the welding machine.
4.3. Summary
This part presents a summary of the results from the different numerical models of welding.
Table 9 presents the criterion for defining the homogeneity level according to the temperature
difference between the copper core and the Inconel 601 coating. Table 10 presents the different
results of the different numerical models. The results show that the most homogeneous heating is
achieved with a MDT-U-5000 and an ECT-U-DC.
1. Magnetodynamic-thermal, with current AC-5000 Hz, in current control, and with and with-
out gaps (MDT-I-5000 and MDT-I-5000-G),
2. Electrokinetic-thermal, with current AC-50Hz, in current control, and without gaps (ECT-I-50),
18 C. TAHRI ET AL.
Figure 15. Deformed shape after cooling obtained with (a) MDT-I-5000, (b) ECT-I-50, and (c) MDT-I-5000-G.
Figure 16. (a) The typical geometry used. (b) The evolution of displacement along the interface after cooling for different cases.
Figures 15a–15c show, respectively, the deformed shapes after welding obtained with the following
models; 1 – MDT-I-5000, 2 – ECT-I-50, and 3 – MDT-I-5000-G. This proves that, despite the different
temperature distributions, the shapes obtained in the case without gaps between copper and Inconel
601 are quite similar. The use of an AC-5000 Hz current with the presence of a gap between copper
and Inconel 601 shows no significant change in the copper core (Figure 15c). Various plots were made
to understand the displacements of the interface as well as the external shapes of the welding area.
Figure 16 shows the contact interface profile (L) at the end of welding (after cooling) in the
three cases mentioned above (Figure 8). It shows that the lower the frequency used, the higher
the displacement observed at the interface (370 mm with AC-50 Hz, compared to 280 mm with
AC-5000 Hz at the center of the welding zone S2). The presence of gaps does not show any
remarkable displacements compared to previous cases, at the interface (L) (Uy 120 mm).
NUMERICAL HEAT TRANSFER, PART A: APPLICATIONS 19
Figure 17. (a) Typical geometry used for longitudinal displacements and displacement evolutions after cooling for (b) MDT-I-
5000, (c) ECT-AC-50-SG, (d) MDT-I-5000-G, and (e) comparison of longitudinal displacement between MDT-I-5000-G and MDT-
I-5000.
For a better understanding of the longitudinal shape, vertical cross sections were defined (Zi).
Z1, Z2, Z3 and Z4 correspond respectively to the external surface of the welding zone, to the
external interface and to the internal interface between the copper and the Inconel 601 and to the
internal surface of the welding zone (Figure 17a).
Figures 17b–17d show the displacements of nodes in different sections (Zi) respectively for:
MDT-I-5000, ECT-I-50, and MDT-I-5000-G. Without gaps, the copper and the Inconel 601
deform in the welding zone. In addition, the graphs show that external displacements are slightly
greater than internal displacements. This is due to the non-symmetry of heating between the
external and internal sides of the ground electrode. When a gap exists (Figure 17d), only the
Inconel 601 is deformed. This is due to the stiffness of copper because of the lack of heating.
Figure 17e shows a comparison between MDT-I-5000 and MDT-I-5000-G. This shows that the
displacements are higher in the case without gaps than with gaps. This difference is due to the
influence of the rigidity of the copper during accosting.
20 C. TAHRI ET AL.
5. Conclusions
The resistance welding process using two types of currents (alternating and direct), two modes of
current regulation (voltage and current), with two different frequencies, and whether or not there
are gaps between the copper core and the Inconel 601 coating, was analyzed. First the gap
between the copper core and the Inconel 601 coating after annealing was determined experimen-
tally. Two extreme cases are studied; with an initial gap (before welding) of 40 mm between cop-
per and Inconel 601, and without an initial gap. Two kind of models was used. The first is a
magnetodynamic-thermal non-linear 3-D model, specially suited for alternating currents. Finite
element magnetodynamic and thermal simulations are weakly coupled to calculate temperature
distributions during the process. A frequency of 5000 Hz, with/without gaps between the copper
core and the Inconel 601 coating was considered. The second model dedicated to direct or alter-
nating low frequency currents (50 Hz), allows the calculation of temperature distributions during
the process through fully coupled electrokinetic-thermal 3-D modeling. The temperature distribu-
tions are then used to simulate the collapse of the ground electrode during welding.
The results show that the type of current (direct or alternating), the mode of current regula-
tion (voltage or current), the frequency and the gaps strongly influence the distribution of heat in
the welded area. Without a gap, the use of a direct current obtained by correcting an alternating
current of 5000 Hz, in voltage control (ECT-U-DC) generates the most homogeneous heating at
the interface between the ground electrode and the shell. The presence of gaps, limits the current
in the copper and therefore generates a localized heating in the Inconel 601 coating. Therefore,
the results in terms of welding time are in good agreement with reality. The welding time
obtained, with a configuration similar to that used in production (ECT-I-DC-G), is of the order
of 30 ms, which corresponds to the time recorded experimentally on the welding machine. Taking
into account the presence of gaps in the ground electrode before welding, the most homogeneous
heating is obtained with a MDT-U-5000-G.
The mechanical calculations prove that the displacement in the case without gaps is almost
similar for both frequencies (AC-50 and AC-5000 Hz). In the presence of gaps, only the Inconel
601 deforms. This is due to the greater rigidity of the copper which, in the presence of gaps, sup-
ports much lower temperatures than in the case without gap.
References
[1] S. Javant, S. V. Hosseini, and S. S. Alaviyoun, “An experimental investigation of spark plug temperature in
bi-fuel engine and its effect on electrode erosion,” Int. J. Automotive Eng., vol. 2, no. 1, pp. 21–29, 2012.
[2] H. Buscail, S. Perrier, and C. Josse, “Oxidation mechanism of the Inconel 601 alloy at high temperatures,”
Mater. Corros., vol. 62, no. 5, pp. 416–422, 2011. DOI: 10.1002/maco.200905519.
[3] Lamineries Matthey, A unit of NOTZ METAL SA, Montagu 38, Case postale CH-2520, LA NEUVEVILLE.
Available: https://www.matthey.ch/en/alliages/cuivre/.
[4] Normalisation Française, AFNOR NF, NF A 35-053, 1984, Siege social 11, rue Francis de Pressense 93571 La
Plaine Saint-Denis Cedex FRANCE. Available: https://www.boutique.afnor.org/norme/nf-a35-053/fil-machine-
en-acier-non-allie-pour-fabrications-realisees-par-formage-a-chaud-ou-a-froid-qualites/article/678153/fa031365
[5] Feulvarch, P. Rogeon, P. Carre, G. Sibilia, and J.-M. Bergheau, “Modelisation du soudage par point:
E.
definition des conditions interfaciales et validation experimentale,” Mecanique Ind., vol. 7, no. 3, pp.
251–263, 2006. DOI: 10.1051/meca:2006039.
[6] E. Feulvarch and J. M. Bergheau, “Resistance spot welding process: Experimental and numerical modeling
of the weld growth mechanisms with consideration of contact conditions,” Numer. Heat Transfer Part A:
Appl., vol. 49, no. 4, pp. 345–367, 2006. DOI: 10.1080/10407780500359760.
[7] P. Rogeon, R. Raoelison, P. Carre, and F. Dechalotte, “A microscopic approach to determine electrothermal
contact conditions during resistance spot welding process,” J. Heat Transfer, vol. 131, no. 2, pp. 022101,
2009. DOI: 10.1115/1.3000596.
[8]
C. Srikunwong, Modelisation du procede de soudage par points. Mecanique [physics.medph]. Français: Ecole
Nationale Superieure des Mines de Paris, 2005.
NUMERICAL HEAT TRANSFER, PART A: APPLICATIONS 21
[9] K. T. Chaouch and P. Rogeon, “Development of weld nugget in dissymmetric assemblies,” presented at the
21st French Congress of Mechanics, Bordeaux Aug. 26–30, 2013.
[10] J. A. Greenwood, “Temperatures in Resistance Spot Welding,” Br. Welding J., vol. 8, no. 6, pp. 316–322,
1967.
[11] W. Rice and E. J. Funk, “An analytic investigation of the temperature distributions during resistance
welding,” Welding J., vol. 57, no. 7, pp. 211s–216s, 1967.
[12] J. G. Kaiser, G. J. Dunn, and T. W. Eagar, “The effect of electrical resistance on nugget formation during
spot welding,” Welding J., vol. 61, no. 6, pp. 167s–174s, 1982.
[13] H. A. Nied, “The finite element modelling of the resistance spot welding process,” Welding J., vol. 63, no.
4, pp. 123s–1132s, 1984.
[14] D. W. Dickinson, C. L. Tsai, and O. Jammal, “Modelling of RSW nugget growth application for the auto-
motive industry,” International Congress and Exposition, Detroit, MI, 1990.
[15] E. Thieblemont, P. Conraux, J.-M. Bergheau, P. Gobez, and J. C. Chevrier, “Couplage electro-thermique -
Application au soudage par resistance par point,” Actes Strucome Paris, pp. 143–156, 1991.
[16] C. L. Tsai, W. L. Dai, D. W. Dickinson, and J. C. Papritan, “Analysis and development of a Real time con-
trol methodology in resistance spot welding,” Welding J., pp. 339s–357s, 1991.
[17] M. M. Vogler and S. D. Sheppard, “A study of temperature histories in resistance spot welding, inter-
national trend in welding science and technology,” Proceedings of the 3rd International conference on
Trends in Welding Research, Gatlinburg, TN: ASM, June 1–5, 1992, pp. 57–61.
[18] L. Xu and J. A. Khan, “Finite element modelling of axisymmetric nugget development during resistance
spot welding,” Trends Welding Research: Proceedings of the 5th International conference, Pine Mountain
Georgia, USA, June 1-5 1998, pp. 616–621, ASM international.
[19] Special Metals, 3200 Riverside Drive Huntington, WV 25705-1771. Available: www.specialmetals.com/tech-
center/alloys.
[20] SYSWELDTM, User’s Manual, ESI Group, France, 2019.
[21] C. Chaboudez, S. Clain, R. Glardon, D. Mari, J. Rappaz, and M. Swierkosz, “Numerical modelling in induc-
tion heating for axisymmetric geometries,” IEEE Trans. Magn., vol. 33, no. 1, pp. 739–745, 1997. DOI: 10.
1109/20.560107.
[22] G. Meunier, ed. The Finite Element Method for Electromagnetic Modeling, London: Hoboken, NJ: ISTE;
Wiley, 2008.
[23] J. M. Bergheau and R. Fortunier, Finite Element Simulation of Heat Transfer, London: Hoboken, NJ: ISTE-
Wiley, 2008, pp. 279, ISBN 978-1-84821-053-0.
[24] J. F. Imhoff, “Modelisation magnetique et mecanique des machines electriques par la methode des elements
finis,” These INPG Grenoble, France, 1989.
[25] J. Fetzer, S. Kurz, and G. Lehner, “The coupling of boundary elements and finite elements for non destruc-
tive testing applications,” IEEE Trans. Magn., vol. 33, no. N 1, pp. 677–681, 1997. DOI: 10.1109/20.560096.
[26] J.-M. Bergheau and F. Potier, “Finite element modeling of coupled radiative and diffusive heat transfers in
non participating media including symmetry and periodicity conditions,” Numer. Heat Transfer, Part B:
Fundam., vol. 40, no. 3, pp. 229–247, 2001. DOI: 10.1080/104077901752379639.
[27] R. Pascal, P. Conraux, and J. M. Bergheau, “Coupling between finite elements and boundary elements for
the numerical simulation of induction heating processes using an harmonic balance method,” IEEE Trans.
Magn., vol. 39, no. 3, pp. 1535–1538, 2003. DOI: 10.1109/TMAG.2003.810430.
[28] J.-M. Bergheau and P. Conraux, “FEM-BEM coupling for the modelling of induction heating processes
including moving parts,” J. Shanghai Jiaotong Univers., vol. E-5, no. N 1, pp. 91–99, 2000.
[29] J. M. Bergheau, “Couplages multiphysiques: applications aux procedes de soudage et de chauffage par
induction”, Modelisation numerique 2 – defis et perspectives, Traite MIM sous la direction de P. Breitkopf et
C. Knopf-Lenoir, Hermes-Lavoisier, UTC, Compiegne, France, 2007, pp. 17–48, ISBN 978-2-7462-1616-7.
[30] S. Kurz, J. Fetzer, and G. Lehner, “An improved algorithm for the BEM-FEM-coupling method using
domain decomposition,” IEEE Trans. Magn., vol. 31, no. N 3, pp. 1737–1740, 1995. DOI: 10.1109/20.
376371.
[31] A. Miton, “Magneto-thermal coupling – SIL coupling procedure – Application example,” SYSTUS Int.
Internal Rep. LDZW/98, vol. 255, 1998.
[32] E. Feulvarch and J. M. Bergheau, “Modelling and numerical simulation of resistance spot welding process,”
In Encyclopedia of Thermal Stresses, R. B. Hetnarski, Ed. Springer, New York, 2014, pp. 3112–3123. ISBN
978-94-007-2738-0.
[33] D. G. Fink and H. W. Beatty, Eds., Standard Handbook for Electrical Engineers, 11th ed., McGraw Hill, pp.
18–21, 1978.
[34] T. W. Eagar, “Resistance spot welding: A fast, inexpensive and deceptively simple process,” Proceedings of
the 3rd International conference onTrends in Welding Research: Gatlinburg, Tennesy, USA, Jun. 1–5,
1992, pp 347–351.