PhD opportunities

Numerical metallurgy supporting arc welding processes: Virtual materials for mastering defects.

Thesis proposal

Area of expertiseComputational mechanics and Materials
Doctoral SchoolSFA - Sciences Fondamentales et Appliques
SupervisorM. Charles-Andr GANDIN
Co-supervisorM. Michel BELLET
Research unitCentre for material forming
KeywordsSolidification, Grain growth, phase transformation, mechanic, welding NDT technologies, numerical modelling
AbstractThe CEMEF laboratory has developed during the past years efficient numerical methods in level-set approaches to simulate fusion welding processes. We can cite the research activities developed by O. Desmaison (2013) and S. Chen (2014). Numerical models to simulate arc-welding processes have been proposed including the melting of the base and added metals in a single approach. These models give access to the development of the weld bead and demonstrate the effect of each process parameters. In addition O. Desmaison has developed a thermo-mechanical modelling of welding process including the solution of conservation equations giving access to the evolution of temperature, fluid flow and stress fields endured by the material. At a smaller scale, S. Chen has developed a modelling of grain structure evolution formed by epitaxial grain growth (see Fig.) on the pre-existing microstructure. This modelling has been developed in a multi-scale approach named CAFE (Cellular Automaton Finite element) where the thermo-mechanical evolution of the weld bead and the grain growth development are coupled. This approach gives the possibility to investigate the development of each individual grain and to highlight the effect of grain growth competition on the final texture.

This latter approach will be re-used and continued in the framework of this PhD activity. The aim is to extend the present model in order to consider enhanced models of grain growth kinetics based on thermodynamics coupling and fluid flow effect. Simultaneously, this project aims at modelling the development of defects in welding processes such as hot-cracking. In this field, innovative modelling of crystal plasticity should be developed to follow the evolution of grain deformation during solidification and cracks initiation in intergranular domains. At the same microscale, the development of brittle phase should also be investigated. In addition, virtual microstructure obtained with CAFE models should be exploited considering NDT applications.

All the developments will be validated considering experimental observations of academic partners (ICB, LMGC) also in order to enrich the numerical models. In addition final simulations will also be developed by end-user partners (CEA-LIST, EDF R&D, ArcelorMittal) also considering the time required to integrate numerical developments in TRANSWELD software.
At the end, all the numerical developments will be part of the collaborative C++ library CIMLIB for scientific computing developed at CEMEF during the past years. This library dedicated to FE computations of forming processes gathers the numerical activities of PhD student at CEMEF. The PhD student will benefit also from the models and numerical methods developed by other students in this library (remeshing, numerical resolution, parallelized computation ).
ProfileEngineer / Master student in the field of material science, mechanic or applied mathematics.
FundingFinancement par crdits ANR