The main purpose of this thesis is to improve the fundamental knowledge of non-isothermal tempering of martensite phase and its effects on the reduction in hardness (softening) with respect the base metal occurring at the heat affected zone (HAZ) of resistance spot welded dual-phase (DP) steels. This thesis also aims at understanding the influence of HAZ-softening on the joint performance of various DP steel grades. The tempering of martensite occurring at the sub-critical HAZ (SC-HAZ) of resistance spot welded DP600, DP780 and DP980 steels has been systematically evaluated by microhardness testing through Vickers indentation and the degree of tempering has been correlated to the HAZ-softening. From the joint performance analysis of similar and dissimilar steel grade combinations assessed through standardized testing methods, three important issues have been targeted: a) the joint strength (maximum load to failure), b) the location of failure (failure mode), and c) the physical characteristic of the weld that determines certain type of failure (weld nugget size). In addition, a partial tensile test has been conducted in order to evaluate the initiation of failure in dissimilar steel grade combinations. It has been shown that HAZ-softening lowered the weld size at which transition from interfacial to pullout failure mode takes place along with increased load-bearing capacity and higher energy absorption. Thus, it is concluded from mechanical testing that HAZ-softening benefits the lap-shear tensile joint performance of resistance spot welded DP steels by facilitating pullout failures through failure initiation at the SC-HAZ (tempered region). Instrumented nanoindentation testing was employed to further investigate HAZ-softening along the SC-HAZ by evaluating individual phases of ferrite matrix and tempered martensite islands. Although the ferrite matrix presented a slight reduction in hardness at nanoscale, higher reduction in hardness (softening) resulted for tempered martensite; thus confirming that tempered martensite is the major contributor to softening at micro-scale. A comparison between nanohardness and microhardness testing made at different distances from the line of lower critical temperature of transformation (Ac1) allowed revealing the actual extension of the SC-HAZ. In this regard, good correlation was obtained between nanohardness results along the SC-HAZ and the microstructural changes analyzed by electron microscopy (i.e., the tempering of martensite occurring at various distances far from Ac1 was correlated to low temperature tempering of dual phase steels). An in-depth analysis of the tempering of martensite phase at high temperature in DP steel subjected non-isothermal conditions i.e., rapid heating, extremely short time at peak temperature and rapid cooling (resistance spot welding), has been carried out mainly through analytical transmission electron microscopy (TEM). In addition, an isothermal tempering condition (i.e., slow heating and long time at peak temperature) in DP steel has been evaluated for complementing the analysis. Both non-isothermal and isothermal conditions have been correlated to the softening behaviour. TEM analysis of the base metal in the DP steel indicated that the morphology of the martensite phase is dependent on its carbon content, and its tempering characteristics are similar to that of equal carbon containing martensitic steel. The isothermally tempered structure is characterized by coarsening and spheroidization of cementite (θ) and complete recovery of the martensite laths; whereas precipitation of fine quasi-spherical intralath θ-carbides, coarser plate-like interlath θ-carbides, decomposition of retained austenite into elongated θ-carbides, and partial recovery of the lath structure were observed after non-isothermal tempering of DP steel. This difference in tempering behaviour is attributed to synergistic effect of delay in cementite precipitation due to higher heating rate, and insufficient time for diffusion of carbon that delays the third stage of tempering process (cementite coarsening and recrystalization) during non-isothermal. The finer size and the plate-like morphology of the precipitated carbides along with the partial recovery of the lath structure observed after non-isothermal tempering strongly influenced the softening behaviour of DP steel. The chemical analysis of θ-carbides through extraction replicas for three different DP steels revealed that the chemistry of the carbides is inherited from the parent DP steel during non-isothermal tempering at high temperature confirming that non-isothermal tempering DP steel is predominantly controlled by carbon diffusion.

Orientador: Itamar Ferreira / Tese (doutorado) - Universidade Estadual de Campinas. Faculdade de Engenharia Mecanica / Made available in DSpace on 2018-07-26T22:44:31Z (GMT). No. of bitstreams: 1 Damasco_Astrid_D.pdf: 55690110 bytes, checksum: 6c2d5f5f7cbef8c9f338896b128300d0 (MD5) Previous issue date: 1995 / Resumo: A motivação deste trabalho surgiu do fato da indústria automobilística usar em larga escala o processo de soldagem a ponto por resistência elétrica para a produção de juntas de chapa/chapa e de chapa/tubo com diferentes espessuras, revestimentos e composição química, sendo que ainda há uma série de dificuldades práticas associadas a algumas combinações chapa-tubo- revestimentos e, inclusive, com poucos dados disponíveis na bibliografia especializada. Foi feito, neste trabalho, a caracterização e análise do processo de soldagem a ponto por resistência elétrica para juntas de chapa sobre chapa e de chapa sobre tubo constituídas de aço baixo carbono, com diferentes composição química, espessura e revestimento. Para tanto, foram utilizadas chapas de aço ASTM 1005, de 0,95 mm de espessura, zincadas e tubos de aço ASTM 1023, de seção quadrada 30X30 mm2, com espessuras de 1,5 mm, 2,0 mm e 3,0 mm, todos fosfatizados. Com esses materiais foram obtidas juntas de chapa/chapa e chapa/tubo com diferentes combinações de espessuras ( 0,95/1,5 mm; 0,95/2,0 mm e 0,95/3,0 mm) e revestimentos (zinco/fosfato; zinco/zinco e decapados). As juntas de interesse foram submetidas ao processo de soldagem a ponto por resistência, cujos parâmetros de processo foram previamente determinados com base na chapa mais fina (0,95 mm de espessura), de modo a se obter intervalos de corrente de soldagem para cada combinação de espessura e revestimento.Os parâmetros de processo estudados foram corrente de soldagem, tipo de eletrodos e força entre eletrodos. As juntas obtidas foram analisadas quanto aos seus aspectos externos (identação, deformação e expulsão), aspectos internos (macroestrutura, diâmetro de lente, penetração, vazios e trincas), propriedades mecânicas (resistência à tração axial e torque máximo) e modo de fratura macroscópica. Para as juntas de chapa/chapa foram obtidos intervalos de soldagem maiores que 2,0 kA, garantindo ampla faixa de soldabilidade, para as diferentes combinações de espessuras de chapas e dos revestimentos de chapas estudados. Também foi possível obter intervalos de soldagem maiores que 2,0 kA para as juntas de chapa/tubo, com tubos de 2,0 mm e 3,0 mm de espessura nas diferentes combinações de revestimentos estudadas. A soldagem em tubos de 1,5 mm de espessura foi viabilizada através do desenvolvimento de quatro diferentes processos, sendo que todos os quatro resultaram em intervalos de soldagem acima do valor mínimo de 1,0 kA, a partir do qual um processo de soldagem a ponto por resistência elétrica é considerado viável. Esses processos envolvem a utilização de alguns dispositivos desenvolvidos pela autora, e que levaram a valores razoavelmente altos de carga de tração e de torque máximo. Em alguns casos, os pontos de solda obtidos apresentaram vazios no interior da lente de solda ou trincas na superfície da chapa de 0,95 mm. Contudo, todos os pontos obtidos dentro dos intervalos de soldagem apresentaram valores de carga de tração máxima e de torque máximo maiores que o mínimo prescrito na literatura especializada / Abstract: The motivation for this work came from the fact of the automobile industry uses on a large-scale the resistance spot welding of low-carbon steel of sheet/sheet and sheet/tube joints, with different composition, thickness, and coating, and there are a few information on these complex joints in the bibliography. A characterization and analysis ofthe resistance spot welding process for sheet/sheet and sheet/tube joints of low-carbon steel, with different composition, thickness, and coating has been conducted. For that, ASTM 1005 steel zinc coated sheet, 0.95 mm thick, and ASTM 1023 steel square tube, 30x30 mm2, 1.5, 2.0, and 3.0 mm thick, in a phosphorised condition have been utilized, for obtaining sheet/sheet and sheet/tube joints with different thickness combination (0.95/1.5 mm; 0.95/2.0 mm; 0.95/3.0 mm), and coating (zinc/phosphate; zinc/zinc; and uncoated). The joints were welded by using the welding parameters obtained from the thinest sheet (0.95 mm thick) in order to obtain welding electric current intervals for each thickness and coating combinations. The welding electric current, type of electrode, and welding force have been studied. The joints have been analysed from the following points of view: external aspects (indentation, deformation, and expulsion); inner aspects (macrostructure, nugget, diameter, penetration, shrinkage void, and cracks); mechanical properties (ultimate tensile load and maximum torsional moment), and macroscopic fracture mode. It was observed electric current welding intervals greater than 2.0 kA for all sheet/sheet joints, that means wide weldability for different combinations in sheet thickness and coating. It was possible to observe also current welding intervals greater than 2.0 kA for sheet/tube joints, 2.0 and 3.0 mm thick for all coating conditions. On the other hand, it was possible to obtain good welds in 1.5mm thick tube only by using four different processes, and as a result of these, it was also possible to obtain current welding intervals greater than the minimal value 1.0 kA. Some of the fixtures used in these four processes were developed by the author. In some case, there were shrinkage voids in the nugget or cracks frem 0.95 mm thick sheet surface. Therefore, all nugget obtained in the current welding intervals presented ultimate tensile load and maximum torsional greater than the minimal value recommended in the bibliography / Doutorado / Materiais e Processos / Doutor em Engenharia Mecânica

The automotive industry of today faces ever harder requirements from regulatory bodies to increase the fuel efficiency, reduce the carbon footprint and increase the safety of their vehicles. The problem is being tackled in different ways; one of them being the use of innovative materials to reduce the overall weight while improving the crash safety of the vehicle. One such material is 22MnB5, an ultra-high strength (UHS) boron-alloyed steel, capable of reaching tensile strength of 1900 MPa. The weldability is a vital factor for applying boron steel in an efficient way into a vehicle construction. Resistance spot welding (RSW) is, among the different welding methods, the primary joining methods used within the automotive industry. The main challenges with RSW of UHS boron steel is the narrow welding window and increased risk of expulsion compared to conventional automotive steel. The aim of this thesis was evaluating how the weldability of three-sheet UHS boron steel combinations could be improved by applying different innovative welding methods. The methods investigated where; three-pulsed welding, two-pulsed welding with force profile and using hollow-cone electrodes instead of regular electrodes. The different methods where evaluated with welding experiments and analysis of the nugget diameter, vicker hardness comparison and tensile strength test of welding nugget. The results from this thesis shows that the current window of three-sheet combinations with UHS boron steel can be significantly improved by using hollow-cone electrodes in RSW. The results also showed that the width of the current window varied depending on the depth of the hole in the electrode, a deeper hole improved the current window but also increased the oxide build-up. Applying a force profile with lowered electrode force during the welding sequence provided an improved process window compared to the constant electrode force when welding a three-sheet combination containing AlSi-coated boron steel. A three-pulse welding sequence performed better than the reference two-pulse welding schedule but still not good enough to meet VCC acceptance criteria. / Bilindustrin står idag inför allt hårdare krav från tillsynsmyndigheter förbättra bränsleeffektiviteten, minska koldioxidavtrycket och öka säkerheten på deras fordon. Problemet angrips från ett flertal olika vinklar. varav en ökad användning av innovativa material för att minska den totala vikten samtidigt som fordonets kraschsäkerhet bibehålls eller ens förbättras. Ett sådant material är 22MnB5, ett höghållfast (UHS) borstål, kapabelt att uppnå brottgränser på 1900 MPa.  Svetsbarheten är en vital faktor för att kunna applicera borstål på ett effektivt sätt i en fordonskonstruktion. Inom bilindustrin är motståndspunksvetsning (RSW) den dominanta svetsmetoden. De största utmaningarna med att punktsvetsa höghållfast AlSi-belagt borstål är det har ett generellt smalare svetsfönstren, samt den ökade risken för sprut under svetsprocessen, jämfört med konventionella stål. Målet med denna avhandling var att utvärdera hur svetsbarheten av tre-plåtskombinationer med höghållfast AlSi-belagt borstål kunde förbättras genom att applicera innovativa svetsmetoder. De utvärderade metoderna var; tre-pulsad svetsning, två-pulsad svetsning med applicerad kraftprofil, samt användning av ihåliga elektroder istället för vanliga elektroder. Metoderna utvärderades genom svetsexperiment och analys av svetslobens storlek, vicker hårdhets mätning samt brottgränsmätning av svetsloben. Resultaten från denna avhandling visar att svetsbarheten för tre-plåts kombinationer med UHS borstål kan förbättras avsevärt genom att använda ihåliga elektroder för punktsvetsning. Resultaten pekar också på att förbättringen beroende på hålets djup i elektroden. Ett djupare hål gav större förbättringar men ökade också uppbyggnaden av oxid och restmetall i elektroden.  Genom att applicera en kraftprofil, där elektrodkraften sänktes under svetsprocessen kunde svetsbarheten förbättras för två-puls svetsning, jämfört med att ha konstant elektrodkraft, vid svetsning av en tre-plåtskombination innehållande höghållfast AlSi-belagt borstål. En tre-puls svetssekvens utförde bättre än referenspulssvetsschemat men fortfarande inte tillräckligt bra för att uppfylla VCC-acceptkriterierna.

Resistance spot welding is the main welding method used in the automotive industry to weld thin sheet metal. Today adaptive control systems have been developed for RSW, which means it can adjust the parameters in the weld process automatically during welding. The control systems can register the parameters and properties of the weld in real-time and from that calculate with algorithms how to adjust to give optimal weld conditions. This project is performed at Scania CV AB, Oskarshamn. Conducted in the part of body in white, where an adaptive control system called HCC is used in all weld processes. In this project, HCC was compared to the fixed control system CCR and another adaptive control system named Master mode. First step in the comparison was to create a weld schedule for each control system and test them on two different material combinations. The aim was to quantify gains and benefits that adaptive resistance spot welding systems have on the welding process. Benefits are quantified by examining the parameters and factors such as: weld time, expulsion, robustness, electrode wear and parameters in the control system. The tests were performed by welding as many approved spot welds as possible without tip-dressing the electrode. The experiment followed the requirements from international standards and the Scania standard for resistance spot welding.  The results from the experiment showed that HCC was the most robust process and the spot welds never decreased in size, which CCR and Master mode did. It is possible to weld several different material combinations with HCC, it increases flexibility in production and reduces the time needed to develop new weld schedules. The same schedule can handle many combinations with the same thickness. HCC allows the process to use several pulses and each pulse adds in time. Therefore, the weld schedule should be well developed and optimized to avoid waste in terms of long weld times. The results will give Scania knowledge about the processes and how to further optimize the welding processes in production. The result can also be used as foundation for selection of products or future investments. / Motståndspunktsvetsning är den huvudsakliga svetsmetoden som används inom fordonsindustrin för att svetsa tunn plåt. Idag har adaptiva styrsystem utvecklats för RSW vilket innebär att de automatiskt kan justera parametrarna i svetsprocessen under svetsning. Styrsystemen kan registrera parametrarna och egenskaperna hos svetsen i realtid och därmed beräkna med algoritmer hur de bör justeras för att ge optimala svetsförhållanden. Detta projekt är resultatet av ett examensarbete på Scania CV AB, Oskarshamn. Det utfördes i den nya karossfabriken, där ett adaptivt styrsystem som heter HCC används i alla svetsprocesser. I projektet jämfördes HCC med ett konstantströms styrsystem CCR samt ett annat adaptivt styrsystem kallat Master mode. Den primära metoden var att skapa ett svetsschema för varje styrsystem och testa dem på två olika materialkombinationer. Syftet var att kvantifiera vinster och fördelar som adaptiva punktsvetssystem har på svetsprocessen. Testerna utfördes genom att svetsa så många godkända punkter som möjligt utan att formera elektroden. Fördelarna kvantifieras genom att man undersökte parametrarna och faktorerna svetstid, sprut, robusthet, elektrodslitage och parametrar i styrsystemen. Experimentet följde kraven i enighet med internationella standarder och Scania-standarden för punktsvetsning. Resultaten från experimentet visade att HCC var den mest robusta processen och punkterna minskade aldrig i storlek, vilket CCR och Master mode gjorde. Det är möjligt att svetsa flera olika materialkombinationer med HCC, det ökar flexibiliteten i produktionen och minskar den tid som krävs för att utveckla nya svetsscheman eftersom samma schema kan hantera många kombinationer med samma tjocklek. HCC tillåter processen att använda flera pulser, och varje puls adderar tid och svetsschemat bör därför vara välutvecklat och optimerat för att undvika slöseri med avseende på långa svetstider. Resultaten kommer att ge Scania mer kunskap om processerna och hur man kan optimera processerna ytterligare i produktionen. Resultatet kan också användas som grund för val av produkter eller framtida investeringar.

No / Successful joining of dissimilar metals such as Al and Mg can provide significant advantages to the automotive industry in the fabrication of vehicle bodies and other important components. This study explores dissimilar joining of Al–Mg using a resistance spot welding process to produce microstructurally sound lap joints and evaluates the impact of interlayer composition on microstructural evolution and the formation of intermetallic compounds within the weld nugget. The results indicated that mechanically sound joints can be produced, with fine equiaxed and columnar dendrites within the weld nugget. The presence of intermetallic compounds was also confirmed by the variation in the microhardness values recorded across the weld zone.

No description available.

The union of a set of materials by way of Resistance Spot Welding is designed so that once fused together, a substantial amount of intentional, external force must be applied to separate the contents. Therefore, Resistance Spot Welding is often the preferred fusion method in high-volume manufacturing processes. The result of Resistance Spot Welding however is the formation of a weld nugget which is not visible to the naked eye. Destructive and/or ultrasonic methods applied off-line must be used to determine the quality of each weld; both inefficient and expensive processes. The following research analyzes the data fed back during resistance spot weld sequences in-line and establishes a correlation between emitted characteristics and the final quality of a spot weld. The two characteristics researched to segregate weld quality are: the electrical sin wave signature and the acoustic sin wave signature produced during the welding sequence. Both features were discovered to have a direct correlation to the final quality of a weld once cured. By measuring and comparing these characteristics at the source, an opportunity is presented to decrease time and potential defects by confirming the quality of each weld in-process and at the source.

Microstructure refinement was observed in the fusion zone of AZ31 magnesium (Mg) alloy resistance spot welds when an inoculant was added, either Ti, Al8Mn5, or Mn. The dependence of inoculant potency on the lattice disregistry between inoculants and matrix, and on the liquid cooling rate was studied. Microstructural characterization was performed via optical microscopy, scanning electron microscopy (SEM), and transmission electron microscopy (TEM). Thin foils containing the interface of the inoculant particles and Mg matrix were prepared using a focused ion beam (FIB) technique. Columnar dendritic structures in the vicinity of the fusion boundary and equiaxed dendritic structures in the central area were observed in the fusion zone of welds in the SA and SB AZ31Mg alloys from different suppliers. However, the columnar dendritic zone (CDZ) was well restricted, and the width of the CDZ and the diameters of equiaxed dendrites were much smaller in the SA alloy than those in the SB alloy due to the earlier columnar-equiaxed-transition (CET) in the SA alloy. The refined microstructure in the fusion zone of the SA alloy welds is attributed to the pre-existence of the larger Al8Mn5 particles of 4-10 microns in length in the SA alloy which act as an inoculant for alpha-Mg heterogeneous nucleation. Fatigue life and dislocation substructure were compared between the SA and SB welds. The SA welds with the refined microstructure displayed an enhanced fatigue resistance compared to the SB welds, when the interfacial failure took place across the fusion zone. The increased number and dispersion of slip systems in the fine-grained SA welds contributed to the improvement of fatigue life. The well-developed columnar dendritic grains were successfully restricted and the coarse equiaxed dendritic grains were efficiently refined by intentionally adding Ti or Mn inoculant particles into the as-received SB alloy welds. The Ti and Mn particles of about 8μm diameter were observed to promote the nucleation of alpha-Mg grains during welding. TEM examinations showed the existence of local orientation relationships between the respective inoculants Ti, Mn, and Al8Mn5 with the Mg matrix. The further lattice matching was observed between the Al8Mn5 particles and Mg. The diameter of the added inoculant should be larger than 1.8 microns to make it a potent inoculant based on the thermodynamic calculation. Microstructural examinations of samples with different inoculant additions and under different cooling rates showed that the inoculant potency was high for the Ti inoculant, medium for the Al8Mn5, but low for the Mn, when the cooling rate was low. This order in the decrease of grain refinement efficiency is inversely proportional with the order of crystallographic lattice disregistry between inoculants and matrix, which is calculated based on a crystallographic matching model. This implies that the lattice disregistry determines the potency of inoculants at the low cooling rates. In comparison, the lattice disregistry did not influence the heterogeneous nucleation, when the cooling rate was high. It could be inferred that an extremely high cooling rate produces a large supercooling, and provides a sufficient driving force for heterogeneous nucleation.

Microstructure refinement was observed in the fusion zone of AZ31 magnesium (Mg) alloy resistance spot welds when an inoculant was added, either Ti, Al8Mn5, or Mn. The dependence of inoculant potency on the lattice disregistry between inoculants and matrix, and on the liquid cooling rate was studied. Microstructural characterization was performed via optical microscopy, scanning electron microscopy (SEM), and transmission electron microscopy (TEM). Thin foils containing the interface of the inoculant particles and Mg matrix were prepared using a focused ion beam (FIB) technique. Columnar dendritic structures in the vicinity of the fusion boundary and equiaxed dendritic structures in the central area were observed in the fusion zone of welds in the SA and SB AZ31Mg alloys from different suppliers. However, the columnar dendritic zone (CDZ) was well restricted, and the width of the CDZ and the diameters of equiaxed dendrites were much smaller in the SA alloy than those in the SB alloy due to the earlier columnar-equiaxed-transition (CET) in the SA alloy. The refined microstructure in the fusion zone of the SA alloy welds is attributed to the pre-existence of the larger Al8Mn5 particles of 4-10 microns in length in the SA alloy which act as an inoculant for alpha-Mg heterogeneous nucleation. Fatigue life and dislocation substructure were compared between the SA and SB welds. The SA welds with the refined microstructure displayed an enhanced fatigue resistance compared to the SB welds, when the interfacial failure took place across the fusion zone. The increased number and dispersion of slip systems in the fine-grained SA welds contributed to the improvement of fatigue life. The well-developed columnar dendritic grains were successfully restricted and the coarse equiaxed dendritic grains were efficiently refined by intentionally adding Ti or Mn inoculant particles into the as-received SB alloy welds. The Ti and Mn particles of about 8μm diameter were observed to promote the nucleation of alpha-Mg grains during welding. TEM examinations showed the existence of local orientation relationships between the respective inoculants Ti, Mn, and Al8Mn5 with the Mg matrix. The further lattice matching was observed between the Al8Mn5 particles and Mg. The diameter of the added inoculant should be larger than 1.8 microns to make it a potent inoculant based on the thermodynamic calculation. Microstructural examinations of samples with different inoculant additions and under different cooling rates showed that the inoculant potency was high for the Ti inoculant, medium for the Al8Mn5, but low for the Mn, when the cooling rate was low. This order in the decrease of grain refinement efficiency is inversely proportional with the order of crystallographic lattice disregistry between inoculants and matrix, which is calculated based on a crystallographic matching model. This implies that the lattice disregistry determines the potency of inoculants at the low cooling rates. In comparison, the lattice disregistry did not influence the heterogeneous nucleation, when the cooling rate was high. It could be inferred that an extremely high cooling rate produces a large supercooling, and provides a sufficient driving force for heterogeneous nucleation.

Dans l’industrie automobile, les exigences en matière d’émissions polluantes conduisent à alléger les véhicules, notamment en réduisant l’épaisseur des tôles. Ce travail en partenariat avec ArcelorMittal porte sur le soudage par résistance par point de tôles fines d’acier. L’objectif est d’identifier les phénomènes qui induisent les difficultés de soudabilité opératoire rencontrées avec une combinaison dissymétrique de trois tôles revêtues, incluant une tôle très mince galvanisée de moins de 0,6 mm, une tôle de DP600 et une tôle en Usibor® emboutie à chaud. Des observations par caméra infrarouge montrent que les échauffements initiaux se produisent principalement au niveau des interfaces avec la tôle d’Usibor®1500, et que le noyau se forme du côté de cette tôle, loin de la tôle mince. Les valeurs très élevées des résistances de contact électrique et thermique, mesurées aux interfaces avec la tôle d’Usibor®, sont imputables au revêtement Alusi® et sont à l’origine des forts échauffements initiaux observés à ces interfaces. Un modèle numérique, limité aux aspects électrothermiques et développé sur COMSOL Multiphysics®, a permis de montrer que la zone fondue s’initie très rapidement dans la tôle d’Usibor® 1500 sous l’effet des fortes résistances de contact adjacentes, et que son développement en épaisseur et diamètre est piloté par les évolutions des rayons de contact électrode-tôle. Les résistances de contact entre électrode et tôle mince, le profil du courant de soudage et les rayons de courbure des faces actives des électrodes sont les paramètres prépondérants à optimiser pour améliorer la pénétration du noyau dans la tôle mince. / In the automotive industry, the requirements for polluting emissions lead to light vehicles, especially in decreasing the steel sheet thickness. This work in partnership with ArcelorMittal focuses on resistance spot welding of steel sheets. The aim is to identify the phenomena that induce operating weldability difficulties encountered with an asymmetrical stack of three coated steel sheets, including a very thin galvanized sheet of less than 0.6 mm, a sheet of DP600, and a hot stamped Usibor® sheet. Infrared camera observations show that the initial heating takes placeat the interfaces with the Usibor®1500 sheet, and that the nugget appears inside this sheet, away from the thin sheet. The very high values of the electrical and thermal contact resistances, measured at the interfaces with the Usibor®1500 sheet, are due to the Alusi® coating and are at the origin of the strong initial heating at these interfaces. A numerical model, limited to the electro- thermal aspects and developed with COMSOL Multiphysics®, shows that the nugget is initiated very quickly in Usibor®1500 sheet under the effect of adjacent contact resistances, and that its development is driven by the evolutions of the electrode-sheet contact areas. The contact resistances between the electrode and the thin sheet, the welding current evolution, and the curvature radius of electrode tips are the most efficient parameters to be optimized to improve the penetration of the nugget in the thin sheet.