AI-augmented electrochemical model for lithium-ion battery: Recent advances and perspectives

With the rapid development of electric vehicles and grid-scale renewable integration, the demand for lithium-ion batteries (LIBs) has significantly increased with high expectations on enhanced energy density, cycle stability, and failure resilience. Electrochemical models (EMs), serving as pivotal mechanism-driven analytical frameworks in battery research and applications, demonstrate unprecedented quantitative fidelity in characterizing intricate multi-physics dynamics for the next-generation battery management systems (BMS). The breakthrough innovations in artificial intelligence (AI) driven methods have revolutionized the dynamic modeling of LIBs. However, the deployment of AI-augmented EMs in BMS faces significant identifiability challenges due to strong parameter coupling. In addition, research on model simplification, parameter determination, and dynamic parameter identification remains largely fragmented. There is a lack of a comprehensive review to pave the way for the cross-domain innovations in BMS. To fill this gap, this paper presents a systematic review of the EMs for LIBs and examines the advancements in parameter determination techniques from both experimental measurement and numerical simulation perspectives. Besides, a comprehensive assessment of the progress in parameter identification from the standpoint of dynamic recognition is presented, encompassing both model-based approaches and intelligent methods. Additionally, from the BMS standpoint, the strengths and limitations of existing approaches are evaluated. Finally, a coordinated framework for multi-stage identification needs to be established in the future. The potential of digital twins (DT), deep reinforcement learning (DRL), and large language models (LLMs) in enhancing EMs also warrants further exploration. The purpose of this work is to provide insights and guidance for the future development of EMs in LIB applications.