Description:
Accurate temperature estimation in lithium-ion batteries is fundamental to ensure safety, performance, and lifetime in electric vehicles and energy storage systems. Traditional methods rely on thermal sensors or physics-based models, which can be costly or computationally demanding. Recent advances in physics-informed machine learning (PIML) offer a promising alternative by embedding the governing electro-thermal equations directly into the learning process. The goal of this thesis is to develop and validate a physics-informed deep learning model capable of estimating the internal temperature distribution of Li-ion cells using limited sensor data. The student will compare purely data-driven and hybrid physics-informed approaches, assessing accuracy, generalization, and computational efficiency.