| Abstract: |
| The pseudo-two-dimensional (P2D) model is a mathematical model that describes the electrochemical processes in Li-ion batteries. Thi model is composed of multiple nonlinear partial differential equations and nonlinear relations, such as the Butler-Volmer equation. In this talk, we investegate the application of Physics-informed neural networks(PINNs) for solving P2D model. Due to the aformentioned nonlinearities in the P2D model, the standard approach often lead to inaccurate solutions. To address these issues, we introduce additional strategies: (1) the incorporation of bypassing terms and (2) the implementation of secondary conservation laws, aimed at improving the stability and accuracy of the PINNs. We first show the efficiency and importance of these strategies through simple toy examples. And then, we present simulation results for the P2D model using PINNs enhanced with our proposed strategies. |
|