Sélectionner une page

AI and digital twins in metallurgy

CIFRE contract


Personnes à contacter par le candidat




Multiscale materials modeling, and more precisely simulations at the mesoscopic scale, constitute the most promising numerical framework for the next decades of industrial simulations. of microstructure evolutions.

In this context, the efficient and robust modeling of evolving interfaces like grain boundary networks is an active research topic, and numerous numerical frameworks exist. In the context of hot metal forming, a new promising front-tracking (FT) method was recently developed . This PhD will focus on exploring Machine Learning strategies for different applications to enhance the solutions proposed within IGIMU® for data generation and exploitation.

First, 3D representative polycrystalline microstructure reconstruction from 2D data will be explored by GAN based methods [3]. Secondly, use of supervised DNN and Deep Reinforcement Learning will be explored to build fast surrogates on top of high-fidelity simulation data generated by the new developed front tracking method.

These tools shall enable the automatic causal interpretation of microstructural singularities such as abnormal grain growth. The developments will be validated thanks to pre-existing experimental and numerical data concerning the evolution of grain boundary interfaces during recrystallization and related phenomena for different materials. They will also be integrated in the DIGIMU® software.