Search

The work presented in this paper deals with the study of the influence of the operating parameters of welding on observables which are the geometrical characteristics of the melted zone. In this study, the method implemented is of the SMAW (Shielded metal arc welding) manual arc type and is used to assemble the pipes through the filling of a V-shaped chamfer between two tubes 1219.5 mm in diameter, 13 mm thick, and a material of grade X70 steel. In order to increase the productivity and reliability of the chamfer filling operation, we used the experimental designs method, on the one hand to estimate the effects of the operating parameters and their interactions on several functions characterizing the morphology of the weld bead, and on the other hand, to provide a mathematical model that links the welding process operating parameters to the objective functions studied.

Inverse heat transfer problems deal with the estimation of parameters or functions appearing in the mathematical formulation of problems in thermal sciences, by utilizing measurements of dependent variables of the formulation. Inverse problems are extremely useful for the indirect measurement of thermophysical properties, in particular for challenging situations involving high temperatures, where coupled multi-physics phenomena and nonlinearities must be taken into account. In this paper, basic inverse problem concepts are reviewed. Solution techniques within the Bayesian framework of statistics are briefly described and applied to two inverse problems related to the authors� experience on the estimation of thermophysical properties at high temperatures.

The hardness of AA 2024 is mainly dependent of the precipitation state in the material. This one will vary through the process of friction stir welding (FSW) which generates heat and deformations. The most important effect will be the thermal excursion which greatly affects the nature and the distribution of precipitates and so the mechanical properties of the material. Three Myhr & Grong-type submodels have been used in this study in order to simulate the variation of hardness in AA 2024-T3 FSW welds. These models allowed to simulate the hardening by growth of S-precipitates and the softening by coarsening and dissolution of GPB zones / co-clusters or S-precipitates. Finally, the natural ageing was taken into account following the Robson model. The complete model has been calibrated with isothermal data found in the literature and still has to be optimised. Nevertheless, preliminary results show the coherence of the model when performed on isothermal data. The model has been also applied to predict FSW hardness profiles that are compared to those found in the literature.

International audience; This work presents the first step concerning the design of a very high temperature inductive furnace dedicated to the characterization of molten metals in the 1 600 K-2 800 K temperature range. One major constraint is to provide homogeneous temperature within the tested sample and thus avoid thermal gradients as well as magnetic fields. For that, the choice of materials for the furnace design has been made and validated through both a magneto-thermal simulation and some in situ temperature measurements of the sample with both thermocouples until 1 600 K, and with a visible-NIR pyrometer used for the upper temperatures. Experimentally, the pyrometer has been calibrated using the melting temperature measurement of a pure nickel sample placed in the furnace. Finally, the furnace heating capacity limit has been tested and reached during the niobium sample melting. During this test, technical difficulties occurred due to physico-chemical contamination from the crucible.

International audience; The work presented in this paper relates to the optimization of operating parameters of the welding by the experimental design approach. The welding process used is the hybrid laser-MAG welding, which consists in combining a laser beam with an MAG torch, to increase the productivity and reliability of the chamfer filling operation in several passes over the entire height of the chamfer. Each pass, providing 2 mm deposited metal and must provide sufficient lateral penetration of about 0.2 mm. The experimental design method has been used in order to estimate the operating parameters effects and their interactions on the lateral penetration on one hand, and to provide a mathematical model that relates the welding parameters of welding to the objective function lateral penetration on the other hand. Furthermore, in this study, we sought to the identification of the set of optimum parameters sufficient to comply with a constraint on the quality of weld bead. This constraint is to simultaneously obtain a total lateral penetration greater than 0.4 mm and an H/L ratio less than 0.6. In order to obtain this condition, the multi-objective optimization (for both response functions) of a weld bead by the implementation of the plans method using two categories of Experiments Plans, on two levels has been used: the first is a complete experimental design (CED) with 32 tests and the second a fractional experimental design (FED) with 8 tests. A comparative analysis of the implementation of both types of experiments plans identified the advantages and disadvantages for each type of plan.

This study presents the joining of steel (DP600) and PA66GF (Polyamide 6.6 reinforced with glass fibers) with a 6 kW Nd:YAG Laser.The phenomena at the interface between the two materials are fundamental elements of analysis in the behavior of the assembly. To understand the impact of process parameters, an experimental plan with different powers and joining speeds is presented. For a better understanding, a specific thermal measurement with micro-thermocouples (wire diameter: 50 µm) is added in the middle of the thermal affected area. The mechanisms leading to a good adhesion are identified, and will allow us to optimize the joining process and increase the understanding. Shear tensile tests are presented to characterize the strength of the assembly. The presented method is compared with bonded assembly as reference.This study presents the joining of steel (DP600) and PA66GF (Polyamide 6.6 reinforced with glass fibers) with a 6 kW Nd:YAG Laser.The phenomena at the interface between the two materials are fundamental elements of analysis in the behavior of the assembly. To understand the impact of process parameters, an experimental plan with different powers and joining speeds is presented. For a better understanding, a specific thermal measurement with micro-thermocouples (wire diameter: 50 µm) is added in the middle of the thermal affected area. The mechanisms leading to a good adhesion are identified, and will allow us to optimize the joining process and increase the understanding. Shear tensile tests are presented to characterize the strength of the assembly. The presented method is compared with bonded assembly as reference.

International audience; This work focuses on the integration of an electrical conduction heating of circular blank in a bulge test device. This device will be used to characterize the thermomechanical behaviour of Usibor (R) 1500 under biaxial deformation at very high temperature (to 930 degrees C). First a thermoelectric model using COMSOL Multiphysics (R) was developed to study the heating of a rectangular blank. This model is validated by comparing the calculated temperatures with thermocouples measurements. Secondly electrical field optimization is approached to obtain a fast and uniform heating of a circular blank.

In this article, we report successful preparation of dense [(Na 0.5 K 0.5 ) 1− x Sr x ](Nb 1− x Ti x )O 3 ( x =0.005–0.100) ceramics by ordinary sintering in air. The dependence of phase structure on doping content of SrO and TiO 2 has been determined by the X-ray diffraction technique. It was found that the crystal structure changed from orthorhombic to tetragonal at x ≈0.040. Dielectric study revealed that the dielectric relaxor behavior was induced by doping of SrO and TiO 2 into (Na 0.5 K 0.5 )NbO 3 . The samples in the composition range from x =0.005 to 0.020 exhibited excellent electrical properties, piezoelectric constant of d 33 =90–96 pC / N , electromechanical planar and thickness coupling coefficients of k p =26.6–32.5% and k t =39.8–43.8%. The results show that the [(Na 0.5 K 0.5 ) 1− x Sr x ](Nb 1− x Ti x )O 3 ceramics are one of the promising lead-free materials for electromechanical transducer applications.

This article shows the estimation of a 2D heat transfer coefficient. The study is a continuation of our previous investigations [P. Le Masson, D. Carron, P. Rogeon and J.J. Quemener (1999). Identification du coefficient de transfert lors d’un essai Jominy instrumente. SFT, 99, 57–62; P. Le Masson, T. Loulou, E. Artioukhine, P. Rogeon, D. Carron and J.J. Quemener (2001). A new approach for the estimation of a convection heat transfer coefficient during a “metallurgical Jominy end-quench ” test. 68th Eurotherm Seminar. Poitiers, France.]. A cylindrical steel bar (L = 100 mm and d = 25 mm) is heated up to 900°C and then quenched by a water jet sprayed on its lower end. The problem under analysis is multifaceted. First, during the exchange phase between water and solid surface, phenomena like vaporization, boiling or forced convection occur. Second, the metallurgical phase changes of the considered material have to be taken into account in the examination of the heat treatment problem. The study is conducted ...

The understanding of phenomena responsible for defects in the solidified welded joint is still a major issue. Very few 3D numerical models have been proposed to study heat transfer and fluid flow during laser welding process. Indeed, the modeling of vaporization effects and keyhole instabilities requires particular numerical methods and developments. For that purpose, a model is developed in order to offer an adaptive and predictive tool, using the commercial code Comsol Multiphysics®. The model takes into account the three phases, namely: the vaporized metal in the keyhole with the plume generation, the liquid metal in the melt pool and the solid metal base. The coupled energy and momentum equations are solved in the three phases. The free surface of the keyhole is tracked with the fixed mesh level set method. The phase change during vaporization is treated with an original approach. Results obtained with the complete model are analyzed in a 3D configuration representing a fusion line and compared to experiments. Then, some feasibility tests are carried out in industrial configurations.The understanding of phenomena responsible for defects in the solidified welded joint is still a major issue. Very few 3D numerical models have been proposed to study heat transfer and fluid flow during laser welding process. Indeed, the modeling of vaporization effects and keyhole instabilities requires particular numerical methods and developments. For that purpose, a model is developed in order to offer an adaptive and predictive tool, using the commercial code Comsol Multiphysics®. The model takes into account the three phases, namely: the vaporized metal in the keyhole with the plume generation, the liquid metal in the melt pool and the solid metal base. The coupled energy and momentum equations are solved in the three phases. The free surface of the keyhole is tracked with the fixed mesh level set method. The phase change during vaporization is treated with an original approach. Results obtained with the complete model are analyzed in a 3D configuration representing a fusion line and compared to exp...

The direct metal deposition (DMD) laser technique is a free-form metal deposition process, which allows generating a prototype or small series of near net-shape structures. Despite numerous advantages, one of the most critical issues of the technique is that produced pieces have a deleterious surface finish which requires post machining steps. Following recent investigations where the use of laser pulses instead of a continuous regime was successful to obtain smoother DMD structures, this paper relates investigations on the influence of a pulsed laser regime on the surface finish induced by DMD on a widely used titanium alloy (Ti64). Findings confirm that using high mean powers improves surface finish but also indicate a specific effect of the laser operating mode: using a quasi-continuous pulsed mode instead of fully-cw laser heating is an efficient way for surface finish improvement. For similar average powers, the use of a pulsed mode with large duty cycles is clearly shown to provide smoothening effects. The formation of larger and stable melt pools having less pronounced lateral curvatures, and the reduction of thermal gradients and Marangoni flow in the external side of the fusion zone were assumed to be the main reasons for surface finish improvement. Additional results indicate that combining the benefits from a pulsed regime and a uniform laser irradiation does not provide further reduction of surface roughness.

International audience; A complete modeling of heat and fluid flow applied to laser welding regimes is proposed. This model has been developed using only a graphical user interface of a finite element commercial code and can be easily usable in industrial R&D environments. The model takes into account the three phases of the matter: the vaporized metal, the liquid phase, and the solid base. The liquid/vapor interface is tracked using the Level-Set method. To model the energy deposition, a new approach is proposed which consists of treating laser under its wave form by solving Maxwell's equations. All these physics are coupled and solved simultaneously in Comsol Multyphysics®. The simulations show keyhole oscillations and the formation of porosity. A comparison of melt pool shapes evolution calculated from the simulations and experimental macrographs shows good correlation. Finally, the results of a three-dimensional simulation of a laser welding process are presented. The well-known phenomenon of humping is clearly shown by the model.

International audience; The present work deals with the estimation of the time evolution of the weld fusion boundary. This moving boundary is the result of a spot GTA welding process on a 316L stainless steel disk. The estimation is based on the iterative regularization method. Indeed, the three problems: direct, in variation and adjoint, classically associated with this method, are solved by the finite element method in a twodimensional axisymmetric domain. The originality of this work is to treat an experimental estimation of a front motion using a model with a geometry including only the solid phase. In this model, the evolution of this solid domain during the fusion is set with the ALE moving mesh method (Arbitrary Lagrangian Eulerian). The numerical developments are realized with the commercial code COMSOL MULTIPHYSICS coupled with the software MATLAB . The estimation method has been validated in a previous work using theoretical data ([1]). The experimental data, used here for this identification are, temperatures measured by thermocouples in the solid phase, the temporal evolution of the melt pool boundary observed at the surface by a fast camera and the maximal dimensions of the melted zone measured on macrographs. These experimental data are also compared with numerical results obtained from a heat and fluid flow model taking into account surface tension effects, Lorentz forces and the deformation of the melt pool surface under arc pressure.

A complete modeling of heat and fluid flow applied to laser welding regimes is proposed. This model bas been developed using only a (Graphical User Interface) GUI of a finite element commercial code and can be easily usable in industrial R&D environments. The model takes into account the three phases of the matter: the vaporized metal, the liquid phase and the solid base. The liquid / vapour interface is tracked using the Level-Set method. To model the energy deposition, a new approach is proposed which consists of treating laser under its wave form by solving Maxwell’s equations. All these physics are coupled and solved simultaneously in Comsol Multyphysics®. The calculated velocity and free surface deformation are analyzed. An example of simulation leading to the formation of porosity is presented. Finally melt pool shapes evolution are compared to experimental macrographs.A complete modeling of heat and fluid flow applied to laser welding regimes is proposed. This model bas been developed using only a (Graphical User Interface) GUI of a finite element commercial code and can be easily usable in industrial R&D environments. The model takes into account the three phases of the matter: the vaporized metal, the liquid phase and the solid base. The liquid / vapour interface is tracked using the Level-Set method. To model the energy deposition, a new approach is proposed which consists of treating laser under its wave form by solving Maxwell’s equations. All these physics are coupled and solved simultaneously in Comsol Multyphysics®. The calculated velocity and free surface deformation are analyzed. An example of simulation leading to the formation of porosity is presented. Finally melt pool shapes evolution are compared to experimental macrographs.

International audience; The direct metal deposition (DMD) with laser is a free-form metal deposition process for manufacturing dense pieces, which allows generating a prototype or small series of near net-shape structures. One of the most critical issues is that produced pieces have a deleterious surface finish which systematically requires post machining steps. This problem has never been fully addressed before. The present work describes investigations on the DMD process, using an Yb-YAG disk laser, and a widely used titanium alloy (Ti-6Al-4V) to understand the influence of the main process parameters on the surface finish quality. The focus of our work was: (1) to understand the physical mechanisms responsible for deleterious surface finishes, (2) to propose different experimental solutions for improving surface finish. In order to understand the physical mechanisms responsible for deleterious surface finishes, we have carried out: (1) a precise characterization of the laser beam and the powder stream; (2) a large number of multi-layered walls using different process parameters (P(W), V(m/min), Dm (g/min), Gaussian or uniform beam distribution); (3) a real time fast camera analysis of melt pool dynamics and melt-pool - powder stream coupling; (4) a characterization of wall morphologies versus process parameters using 2D and 3D profilometry. The results confirm that surface degradation depends on two distinct aspects: the sticking of nonmelted or partially melted particles on the free surfaces, and the formation of menisci with more or less pronounced curvature radii. Among other aspects, a reduction of layer thickness and an increase of melt-pool volumes to favor re-melting processes are shown to have a beneficial effect on roughness parameters. Last, a simple analytical model was proposed to correlate melt-pool geometries to resulting surface finishes.

While hybrid laser welding and coating processes involve a large number of physical phenomena, it is currently impossible to predict, for a given set of influencing factors, the shape of the molten zone and the history of temperature fields inside the parts. This remains true for complex processes, such as the hybrid laser/MIG welding process, which consists in combining a laser beam with a MIG torch. The gains obtained result essentially from the synergy of the associated processes: the stability of the process, the quality of the seam realized, and the productivity are increased. This article shows how, by means of a reduced number of experiments (8), it is possible to predict the shape of the molten zone and the temperature field inside parts, for a given window of influencing factors. This method consists in combining the method of exploratory experimental designs with a numerical modelling of the thermal phenomena that occurs during the process, by using the ‘heat equivalent source” approach [1] , [2] , [3] , [4] . Two validations of this method have been carried out: the first for a set of parameters inside the experimental design, and the other for a set of parameters that lies outside the experimental design, but inside the domain investigated.

International audience; A three-dimensional heat transfer model was developed to predict the temperature fields, the weld geometry and the shape of the solidified weld reinforcement surface during hybrid laser-MAG arc welding of fillet joints. Melt pool deformation due to arc pressure was calculated by minimizing the total surface energy. A series of hybrid welding experiments was conducted on S355 steel for different welding speeds and wire feeding rates. A high speed video camera was used to measure weld pool depression and surface weld pool geometry. Visualization of the weld pool during welding has also allowed for a better understanding of the interaction between the keyhole and droplets. The various weld bead shapes were explained through these observations. The arc pressure, the surface energy distribution, and arc efficiency were evaluated by comparing experimental data and numerical results for a wide range of welding operating parameters. Good correlation was found between the calculated and experimental weld bead shapes obtained for the hybrid laser-MAG arc welding process as well as for laser or MAG alone.

International audience; Literature on intrapreneurship has primarily focused on an optimistic "out of the box" thinking, with an individualistic, non-programming view. This paper proposes a contingent and historical approach, from the industrial revolution, linking the organizational contexts of innovation and intrapreneurship development. From this perspective, instrapreneurship is not only a capability, bt a process that can reveal the gaps within an organization.

This work consists of estimating simultaneously the thermophysical parameters which characterize the thermal behavior of two solids in combined thermal and electrical contact, i.e., the heat flux released at the interface, the partition coefficient of this heat flux, and the thermal contact conductance between the two solids. The parameter estimation problem is performed with the Levenberg-Marquardt method. A detailed analysis of the linear dependency of the contact parameters needed to design a robust estimation tool is presented. Despite a weak correlation between the partition coefficient and thermal conductance, the results show that the estimation is feasible and the method developed is an effective tool for a complete thermal characterization of the contact interface. The article concludes with an elementary Monte Carlo study to validate the developed method for estimating the confidence intervals.

A numerical method for the two-dimensional estimation of a convection heat transfer coefficient is developped for a rapid metallurgical heat treatment: the “Jominy end-quench” test. The estimation algorithm is based on the iterative regularization method and on the conjugate gradient and adjoint methods. The procedure is verified for two different materials. The first material, the Nickel, has no metallurgic transformation. The second one, the 16MND5 steel, has transformations implying high nonlinearities in the material. The estimation is simulated numerically for three types of heat transfer coefficients and theoritical thermal cycles. One of these coefficients can be brought together with an experimental case. For the noised thermal cycles the solution is stabilized by the iterative regularization.

International audience; It is widely accepted that laser reflections can play a critical role during keyhole laser welding. The energy concentration, the mask effects and the laser polarization can directly affect the molten pool dynamic. In this paper a new approach to compute laser reflections is proposed which consists of treating laser under its wave form by solving Maxwell's equations. The method has the advantage to be easily coupled with heat transfer and fluid flow equations and can be immediately transposable in any 2D, 2D axi or 3D configurations. The reliability and limits of this approach are discussed through different numerical examples. The complete model takes into account the three phases of the matter: the vaporized metal, the liquid phase and the solid base. To predict the evolution of these three phases, coupled equations of energy, continuity, momentum and Maxwell are solved. The liquid/vapour interface is tracked using the level-set method. All these physics are solved simultaneously with the commercial code COMSOL Multiphysics®. The calculated temperatures, velocities and free surface deformation are analysed. Examples of simulations leading to the formation of porosity are also presented. Finally, melt pool shapes evolution are compared to experimental macrographs.

Derived from laser cladding, the direct laser metal deposition (DLMD) process is based upon a laser beam–powder–melt pool interaction and enables the manufacturing of complex 3D shapes much faster than conventional processes. However, the surface finish remains critical, and DLMD parts usually necessitate postmachining steps. Within this context, the focus of our work is to improve the understanding of the phenomena responsible for deleterious surface finish by using numerical simulation. Mass, momentum, and energy conservation equations are solved using comsol multiphysics® in a 2D transient model including filler material with surface tension and thermocapillary effects at the free surface. The dynamic shape of the molten zone is explicitly described by a moving mesh based on an arbitrary Lagrangian–Eulerian method (ALE). This model is used to analyze the influence of the process parameters, such as laser power, scanning speed, and powder feed rate on the melt pool behavior. The simulations of a single layer and multilayer claddings are presented. The numerical results are compared with experimental data, in terms of layer height, melt pool length, and depth of penetration, obtained from high speed camera. The experiments are carried out on a widely used aeronautical alloy (Ti–6Al–4 V) using a Nd:YAG laser. The results show that the dilution ratio increases with increasing the laser power and the scanning velocity or with decreasing the powder feed rate. The final surface finish is then improved.

International audience; Le travail présenté dans cet article concerne l'étude de l'estimation de l'évolution d'un front de fusion dans le cadre de la modélisation de l'apport d'énergie lié à un procédé de soudage TIG (Tungsten Inert Gas) statique sur une durée de 20s (le projet BA3 [1]). L'objectif étant d'estimer l'évolution du front de fusion, une instrumentation optimale dans des conditions particulièrement difficiles doit être développée. Cet article présente notre réflexion quant à la définition de cette instrumentation, la méthode de mise en oeuvre ainsi que les résultats obtenus.

Hybrid laser-MIG arc welding presents very interesting advantages as compared to laser welding or arc welding used separately. It is known that improved productivity results from higher welding speeds, thicker materials, joint fit-up allowance, better stability of molten pool, improvement of joint metallurgical quality. In order to use efficiently these techniques for industrial production, it is necessary to precisely understand the complex physical phenomena that govern this welding process. This understanding is also necessary if one wants elaborate adequate simulations of this process. This paper will discuss preliminary results that concern the analysis of the main physical processes controlling this hybrid welding. We will also discuss the numerical approach we use for modelling this process.Hybrid laser-MIG arc welding presents very interesting advantages as compared to laser welding or arc welding used separately. It is known that improved productivity results from higher welding speeds, thicker materials, joint fit-up allowance, better stability of molten pool, improvement of joint metallurgical quality. In order to use efficiently these techniques for industrial production, it is necessary to precisely understand the complex physical phenomena that govern this welding process. This understanding is also necessary if one wants elaborate adequate simulations of this process. This paper will discuss preliminary results that concern the analysis of the main physical processes controlling this hybrid welding. We will also discuss the numerical approach we use for modelling this process.

Compared to laser or arc welding used separately, hybrid laser-MAG arc welding show decisive advantages when higher welding speeds, thicker materials, joint fit-up allowance, better molten pool stability, increase of metallurgical joint quality are required. The understanding of the main relevant involved physical processes are therefore very necessary if one wants for example elaborate adequate simulations of this process. Also for an efficient use of this process, it is necessary to precisely understand the complex physical phenomena that govern this welding technique. This paper will first discuss different results that concern the analysis of the main physical processes controlling this hybrid welding. Melt pool depression due to arc pressure is an important physical phenomenon that defines the location where the arc deposits it energy onto the melt pool surface. We will discuss this effect from related experiments and show how it can be taken into account by using corresponding 3D numerical simulations describing the thermal and hydrodynamic field of the melt pool by taking into account its 3D free surface.Compared to laser or arc welding used separately, hybrid laser-MAG arc welding show decisive advantages when higher welding speeds, thicker materials, joint fit-up allowance, better molten pool stability, increase of metallurgical joint quality are required. The understanding of the main relevant involved physical processes are therefore very necessary if one wants for example elaborate adequate simulations of this process. Also for an efficient use of this process, it is necessary to precisely understand the complex physical phenomena that govern this welding technique. This paper will first discuss different results that concern the analysis of the main physical processes controlling this hybrid welding. Melt pool depression due to arc pressure is an important physical phenomenon that defines the location where the arc deposits it energy onto the melt pool surface. We will discuss this effect from related experiments and show how it can be taken into account by using corresponding 3D numerical simulatio...

There is a demand from industry and research laboratories for the thermophysical properties of metals in the liquid state (1700 K - 3000 K). This thesis, therefore, set the goal of developing an experimental device for measuring the density and surface tension of such materials and particularly steels. The two measurement methods are based on the principle of the drop in aerodynamic levitation, heated by a laser and positioned in a vacuum chamber filled with a neutral gas. These criteria are set because of an important requirement about the non-pollution of the drop by foreign bodies (mostly crucible and oxygen). The density is measured by observing with a high speed camera the volume variation of the drop during the cooling and by weighting the solidified drop. This first method gave satisfying results on pure substances in comparison with the literature and was applied to an industrial steel. The surface tension is measured by observing the resonance frequency of the drop oscillating thanks to an acoustic excitation. A particular attention is given to the aerodynamic levitation of drop and developments have been undertaken to cancel the hazardous motions of the sample thus facilitating the measurement. This method gives satisfactory results but requires additional developments to avoid pollution impacting the results obtained., Il existe, venant de l’industrie et des laboratoires de recherches, un besoin de connaissance des propriétés thermophysiques des métaux à l’état liquide (1 700 K – 3 000 K). Cette thèse s’est donc posée comme objectif de développer un dispositif expérimental de mesure de la masse volumique et de la tension superficielle de ces matériaux, et plus particulièrement les aciers. Les deux méthodes de mesure sont basées sur le principe de la goutte en lévitation aérodynamique, chauffée par laser et placée dans une enceinte étanche rempli avec un gaz neutre. Ces critères répondent à une importante exigence en matière de non-pollution de la goutte par des corps étrangers (principalement creuset et oxygène). La masse volumique est mesurée en observant avec une caméra haute vitesse la variation de volume de la goutte lors de son refroidissement et en pesant la goutte solidifiée. Cette première méthode a donné des résultats satisfaisant sur des corps purs en comparaison de la littérature et a été appliquée à un acier industriel. La tension superficielle est mesurée en observant la fréquence de résonnance de la goutte mise en oscillation par une excitation acoustique. Une attention particulière a été portée sur la stabilité aérodynamique de la goutte et des développements ont été entrepris pour annuler les rotations hasardeuses de l’échantillon facilitant ainsi la mesure. Cette méthode donne des premiers résultats encourageants mais requiert des développements supplémentaires pour éviter une pollution impactant les résultats obtenus.

Liquation cracking may occur in the heat-affected zone during welding. Two factors influence this phenomenon: the tensile stresses generated during welding and the potential loss of ductility due to the presence of a liquid film at grain boundaries depending on their chemical composition. Gleeble hot ductility tests have been used to study the combined effect of boron content and holding time on ductility drop in the liquation temperature range of a 316L type austenitic stainless steel. It is shown that high boron contents and short holding times promote the loss of ductility in this temperature range. Secondary ion mass spectrometry has been used inattempt to correlate mechanical results to boron distribution either at grain boundaries or in the bulk. Other hot cracking tests (Varestraint, PVR) have been performed to confirm the influence of boron content on hot cracking sensitivity of AISI 316L stainless steels during welding. Results indicate that cracks appear on all specimens and that the minimum external mechanical loading for liquation cracking decreases with boron content. The higher the boron content is, the more the specimen exhibits tendency to hot cracking. A liquation hot cracking criterion has been determined, based on the results of the hot ductility tests and the simulation of welding tests., La fissuration à chaud par liquation peut se produire dans la zone affectée thermiquement lors du soudage. Deux facteurs influent ce phénomène : les contraintes thermiques dues au gradient de température et la perte potentielle de ductilité due à la présence d'un film liquide aux joints de grains en fonction de leur composition chimique. Des essais de ductilité à chaud (Gleeble) ont été utilisés pour étudier l'effet combiné de la teneur en bore et du temps de maintien sur la chute de la ductilité dans la plage de température de liquation pour l’acier inoxydable austénitique de type 316L. Il est démontré que les teneurs en bore élevées et les temps de maintien courts favorisent la perte de ductilité dans cette plage de température. En complément la spectrométrie de masse à ionisation secondaire a été utilisée pour tenter de corréler les variations de ductilité à la distribution du bore aux joints de grains. D'autres essais de fissuration à chaud en soudage (Varestraint, PVR) ont été effectués pour confirmer l'influence de la teneur en bore sur la sensibilité à la fissuration de l’acier 316L. Les résultats indiquent que des fissures apparaissent sur toutes les éprouvettes et que le chargement mécanique externe minimal pour créer les fissures de liquation diminue avec la teneur en bore. Plus la teneur en bore est élevée, plus le matériau est donc sensible à la fissuration à chaud par liquation. Un critère de fissuration à chaud par liquation a été déterminé en se basant sur les résultats des essais de ductilité à chaud et la simulation des essais de soudage.

To reduce greenhouse gas emissions, car makers are looking for innovative lightening solutions. Thanks to its very high mechanical properties, the quenchable boron steel Usibor®1500 allows to reduce thicknesses of structural parts. The use of this steel requires a specific forming process named hot stamping. In this process, the blank is heated to induce austenitization microstructure, followed by a forming and quenching step to obtain a completely martensitic structure which confers high mechanical properties on the part. The ANR PRICECAT project consists in developing a process which includes successive forming steps under the same press. In this context, the aim of this thesis is to increase the heating rate using electrical conduction. This technology is suitable for rectangular blank heating, but with a shape blank, the section variations lead to a heterogeneous heating incompatible with a completely martensitic structure. However, in automobile industry, parts have various shapes. So, it was necessary to determine new methods of heating shape blanks by electrical conduction. After a review of the electrical heating methods, a thermoelectric model was developed under COMSOL Multiphysics®. It was validated experimentally on a rectangular blank heating case. This model was used to analyze the influence of different heating parameters and to design two heating devices. Thus, the electrical conduction heating was first integrated to a biaxial expansion device. The tests showed that homogeneous heating of a circular blank is possible during heating and forming steps. The numerical model also led to another heating solution where the shape blank is converted into rectangular strips with minimization of the added surface. This method was used to build a heating demonstrator for a windscreen upright.; Pour réduire les émissions de gaz à effet de serre, les constructeurs automobiles cherchent à alléger lesvéhicules. L’Usibor®1500 permet de réduire les épaisseurs des pièces de structure grâce à ses propriétés mécaniques élevées obtenues suite au procédé d’emboutissage à chaud. Ce procédé comporte une chauffe où le flan (la tôle) est austénisé, puis mis en forme et trempé simultanément pour obtenir une structure entièrement martensitique qui confère des propriétés mécaniques élevées. Le projet ANR PRICECAT consiste à développer un procédé dans lequel la mise en forme se fait par multi- passes dans une presse unique. Dans ce cadre, l’objectif de la thèse est d’augmenter la vitesse de montée en température en utilisant la technologie de chauffe par conduction électrique. Cette technologie est adaptée pour la chauffe de flan rectangulaire, mais lorsque le flan est de forme quelconque, les variations de section mènent à une chauffe hétérogène incompatible avec l’obtention d’une structure entièrement martensitique. Or, les flans rencontrés dans l’industrie automobile ont des formes diverses. Il a donc été nécessaire de déterminer de nouvelles méthodes de chauffe par conduction électrique de flan de forme. Après une revue sur les différentes méthodes de chauffe par conduction électrique rencontrées dans la littérature, un modèle thermoélectrique a été développé sous COMSOL Multiphysics® et validé expérimentalement sur un cas de chauffage de flan rectangulaire. Ce modèle a permis d’analyser l’influence de divers paramètres de chauffe et a servi de base pour le développement de deux dispositifs de chauffe. La technologie de chauffe a ainsi été intégrée à un dispositif d’expansion biaxiale. Les essais ont montré qu’une chauffe homogène d’un flan circulaire était possible même lors de l’étape de mise en forme. Le modèle numérique a aussi conduit à une solution de chauffe où l’on discrétise le flan de forme en bande rectangulaire tout en minimisant la surface ajoutée. Cette méthode s’est concrétisée par la réalisation d’un démonstrateur à l’échelle 1 de chauffe de montant de baie.

To reduce greenhouse gas emissions, car makers are looking for innovative lightening solutions. Thanks to its very high mechanical properties, the quenchable boron steel Usibor®1500 allows to reduce thicknesses of structural parts. The use of this steel requires a specific forming process named hot stamping. In this process, the blank is heated to induce austenitization microstructure, followed by a forming and quenching step to obtain a completely martensitic structure which confers high mechanical properties on the part. The ANR PRICECAT project consists in developing a process which includes successive forming steps under the same press. In this context, the aim of this thesis is to increase the heating rate using electrical conduction. This technology is suitable for rectangular blank heating, but with a shape blank, the section variations lead to a heterogeneous heating incompatible with a completely martensitic structure. However, in automobile industry, parts have various shapes. So, it was necessary to determine new methods of heating shape blanks by electrical conduction. After a review of the electrical heating methods, a thermoelectric model was developed under COMSOL Multiphysics®. It was validated experimentally on a rectangular blank heating case. This model was used to analyze the influence of different heating parameters and to design two heating devices. Thus, the electrical conduction heating was first integrated to a biaxial expansion device. The tests showed that homogeneous heating of a circular blank is possible during heating and forming steps. The numerical model also led to another heating solution where the shape blank is converted into rectangular strips with minimization of the added surface. This method was used to build a heating demonstrator for a windscreen upright., Pour réduire les émissions de gaz à effet de serre, les constructeurs automobiles cherchent à alléger lesvéhicules. L’Usibor®1500 permet de réduire les épaisseurs des pièces de structure grâce à ses propriétés mécaniques élevées obtenues suite au procédé d’emboutissage à chaud. Ce procédé comporte une chauffe où le flan (la tôle) est austénisé, puis mis en forme et trempé simultanément pour obtenir une structure entièrement martensitique qui confère des propriétés mécaniques élevées. Le projet ANR PRICECAT consiste à développer un procédé dans lequel la mise en forme se fait par multi- passes dans une presse unique. Dans ce cadre, l’objectif de la thèse est d’augmenter la vitesse de montée en température en utilisant la technologie de chauffe par conduction électrique. Cette technologie est adaptée pour la chauffe de flan rectangulaire, mais lorsque le flan est de forme quelconque, les variations de section mènent à une chauffe hétérogène incompatible avec l’obtention d’une structure entièrement martensitique. Or, les flans rencontrés dans l’industrie automobile ont des formes diverses. Il a donc été nécessaire de déterminer de nouvelles méthodes de chauffe par conduction électrique de flan de forme. Après une revue sur les différentes méthodes de chauffe par conduction électrique rencontrées dans la littérature, un modèle thermoélectrique a été développé sous COMSOL Multiphysics® et validé expérimentalement sur un cas de chauffage de flan rectangulaire. Ce modèle a permis d’analyser l’influence de divers paramètres de chauffe et a servi de base pour le développement de deux dispositifs de chauffe. La technologie de chauffe a ainsi été intégrée à un dispositif d’expansion biaxiale. Les essais ont montré qu’une chauffe homogène d’un flan circulaire était possible même lors de l’étape de mise en forme. Le modèle numérique a aussi conduit à une solution de chauffe où l’on discrétise le flan de forme en bande rectangulaire tout en minimisant la surface ajoutée. Cette méthode s’est concrétisée par la réalisation d’un démonstrateur à l’échelle 1 de chauffe de montant de baie.

Welding simulation can be improved by a better knowledge of molten metals thermophysical properties. This improvement requires characterization using a very high temperature apparatus. After portraying a state of the art on characterization technics adapted to this temperature range (chapter 1), this manuscript details works, principally experimental, dedicated to the developments of a device able to heat metals up to 2 500 °C (chapters 3 and 4) and of a five wavelengths pyrometer in order to measure this temperature (chapters 2, 4 and 5). The device heats by induction a graphite tower, the crucible being placed at its center, and has been dimensioned by a magneto-thermal 2D axisymmetric simulation using the Comsol Multiphysics® software. In parallel, the pyrometer has been developed in order to measure both temperature and emissivity. In the absence of a high temperature blackbody, several experiments were carried out for the calibration of it; they were based on the use of the luminance of pure metals at their melting point. In order to cover the largest temperature range possible, chosen metals were iron, chromium, and even niobium. These experiments showed the efficiency of the pyrometer and the maximal ability of the high temperature device during the niobium melting around 2 500 °C. Finally, the pyrometer has been used on an arc welding process, in which temperature has been evaluated over a 1 000 °C – 2 500 °C temperature range., La simulation numérique du soudage peut être améliorée par une meilleure connaissance des propriétés thermophysiques des métaux à l’état liquide. Un approfondissement de cette connaissance requiert un dispositif capable de conditionner ces métaux à de très hautes températures en vue de les caractériser. Après un état de l’art sur les techniques de caractérisation adaptées à ces échelles de températures (chapitre 1), ce manuscrit détaille les travaux, principalement expérimentaux, dédiés aux développements d’un dispositif capable d’élever des métaux jusqu’à 2 500 °C (chapitres 3 et 4) et d’un pyromètre à cinq longueurs d’onde pour mesurer cette température (chapitres 2, 4 et 5). Le dispositif chauffe par induction une tour en graphite, le creuset étant placé en son centre, et a été dimensionné via une simulation magnéto-thermique 2D axisymétrique à l’aide du logiciel Comsol Multiphysics®. Parallèlement, le pyromètre a été développé dans le but de mesurer simultanément la température et l’émissivité. En l’absence de corps noir à hautes températures, plusieurs expériences ont été menées pour étalonner ce dernier ; elles ont été basées sur la mesure de la température de fusion/solidification de corps purs. En fonction de leur tenue et pour couvrir la plus large gamme de température possible, les métaux choisis sont le fer, le chrome, ou encore le niobium. Les expériences ont montré l’efficacité du pyromètre et la capacité maximale du dispositif hautes températures lors de la fusion du niobium à environ 2 500 °C. Enfin, le pyromètre a été utilisé lors d’une opération de soudage à l’arc, au cours de laquelle la température a été estimée sur une gamme de 1 000°C – 2 500°C.