[T3] Modeling and Sintering

[15A-T3-04] A Physics-Informed Neural Network Approach for Spark Plasma Sintering Simulation

D. Sicard¹, F. Naimi¹, *M. Ariane¹ (1.Sintermat SAS, France)

キーワード：Deep learning, Physics-Informed Neural Network, Spark Plasma Sintering

Over the last two decades, Spark Plasma Sintering (SPS) has become a major technique for manufacturing advanced materials. Nevertheless, mastering the SPS process is complex and, for a better understanding, requires computational calculation support. SPS simulations are mostly conducted using Finite Element Method (FEM). However, FEM approaches are well-known for being time and cost-consuming due to high computational complexity. On the other hand, emerging approaches such as Deep Learning (DL) could be considered as a viable alternative to traditional modelling and have proven their effectiveness in many fields. But DL remains data-dependent and does not consider the underlying knowledge of the process such as physical laws. Thereby, the interesting properties of Physics Informed Neural Networks (PINN) lie in the ability of the network to include physics-based knowledge of a system. In this study, we investigated the potential of the emerging mesh-free PINN approach as a faster alternative to SPS FEM models.