Physics-informed neural network surrogate modeling of single particle model for lithium-ion batteries

TL;DR

This paper benchmarks different physics-informed neural network architectures, including DeepONet, for fast and accurate surrogate modeling of lithium-ion battery single particle models, highlighting their potential and limitations.

Contribution

It introduces a comparison of three PINN architectures for battery modeling, demonstrating the superior performance of Fourier-enhanced DeepONet in accuracy and speed.

Findings

Fourier-enhanced DeepONet outperforms other PINNs in generalization.

PINNs can significantly accelerate battery model simulations.

Limitations exist for conventional PINNs under dynamic current conditions.

Abstract

Physics-based models play a key role in battery management, yet face challenges in real-time applications due to the high computational cost of solving coupled algebraic-partial differential equations. To accelerate model simulation, this study benchmarks three physics-informed neural network (PINN) architectures for modeling the battery single particle model, including two conventional PINN architectures and a DeepONet-based architecture. Both the accuracy and the generalization of these PINNs have been evaluated and compared under various current conditions. Our results highlight the potential of PINNs in modeling battery physics but also reveal limitations of conventional PINN architectures under highly dynamic current conditions. Among them, the Fourier-enhanced DeepONet achieves superior generalization performance and offers nearly a 10 times speedup compared with numerical solvers. This work provides an example of integrating physics-based models