The constant progress of laser technologies has amplified the interest in welding for the assembly of metal parts. Current laser sources allow a wide variety in the choice of wavelengths and energy distributions in order to guarantee quality welds. Nevertheless, the great diversity of parameters complicates the search for optimal operating conditions to avoid the appearance of certain defects. The objective of this thesis is to develop a Multi-Scale numerical approach in order to predict at the mesoscopic scale the formation of defects induced by the instabilities of the liquid pool (porosity) and at the part scale the defects related to stresses and deformations. This numerical approach will be based on the use of Multiphysics models allowing us to understand the operating conditions minimizing the defects and thus will make it possible to reduce the number of experimental tests to define the optimal parameters. A reflection will also be carried out to propose a user-friendly digital tool coupling these multi-scale and multi-physical approaches within the same software for easier industrial use.
The numerical developments will be based on previous work carried out within the IRDL (thermo-hydro-dynamic model of laser welding) and IREPA LASER (thermo-mechanical model in Additive Manufacturing) with COMSOL Multiphysics® software. The first step will be to understand the behavior of the melt pool according to the different operating parameters: energy distribution (Gaussian, top hat, annular, etc.), and beam oscillation (amplitude, frequency). Different welding configurations can be studied (edge to edge, transparency), as well as cases of heterogeneous welding. This study will aim to identify the operating conditions minimizing melt pool instabilities (humping, porosity). The calculation of the temperature field at the melt pool scale (Meso) will serve as input data for a Macroscopic thermomechanical model to predict distortions and stresses without resorting to source calibration steps. At the same time, experiment/test campaigns will be conducted to validate the predictive side of the models. The measurement of material properties will also be considered to obtain reliable models.
Progress of the CIFRE thesis and location
This CIFRE thesis will begin within the IRDL laboratory (Institut de Recherche Dupuy de Lôme) in Lorient attached to the Université Bretagne Sud for the development of the models, and the measurement of material data necessary for the models. The thesis will continue at IREPA LASER near Strasbourg, to conduct experiment/test campaigns and validate the model using instrumented models.
You have a general engineering or university equivalent education with a specialization in mechanics and heat transfer. You have skills in numerical simulation by Finite Elements ideally with COMSOL Multiphysics. Knowledge of welding processes or metal additive manufacturing would be a plus. You are autonomous and curious. You have an aptitude for integrating into a team and a capacity for dialogue. You are fluent in English.
Thesis start date
Muriel CARIN (PR), Mickaël COURTOIS (MCF), Stephen CADIOU (MCF), Vaibhav NAIN (IREPA LASER), Frédérique MACHI (IREPA LASER)
Institut de Recherche Dupuy de Lôme Lorient (50%) and IREPA LASER (50%)