Neural simulation of plastic phenomena using MeshGraphNets
1 : ESI Group-UZ Chair of the National Strategy on Artificial Intelligence. Aragon Institute of Engineering Research (I3A). Universidad de Zaragoza.
2 : ESI Group chair. PIMM Lab. ENSAM Institute of Technology.
ESI Group chair. PIMM Lab. ENSAM Institute of Technology
3 : CNRS@CREATE LTD
* : Corresponding author
This work investigates the use of MeshGraphNets, a deep learning framework, to simulate plastic forming processes. Our database consists of synthetic data coming from high-fidelity finite element simulations. The deep learning model accurately captures the complex behavior of plastic phenomena for non-monothonic loading trajectories.
Examples will be provided that demonstrate the robustness and accuracy of our approach in out-of-distribution cases, thus proving the feasibility of neural simulators for industrial use.
| Subject : | : | oral |
| Topics | : | 5-7 - Data driven approaches of materials structures and processes |
| PDF version | : |  PDF version |
