Internal Temperature Distribution Estimation of Lithium-ion Batteries with Physics-informed Neural Network

The large-scale deployment of lithium-ion batteries necessitates careful thermal management to ensure safety, performance, and longevity. Internal thermal gradients can pose safety risks and accelerate battery degradation, making internal temperature monitoring both critical and challenging. This paper presents a physics-informed neural network approach that enables accurate, real-time, and high-resolution estimation of the internal temperature distribution within LIBs by integrating sparse sensor measurements with governing physical laws. The proposed framework leverages the complementary strengths of data-driven learning and physics-based modeling, allowing a lightweight neural network to match or even surpass the performance of the full-order model in internal temperature estimation. Furthermore, the proposed method demonstrates robustness to imperfect physics-based prior knowledge, offering a practical and scalable solution for advanced battery thermal monitoring.