Key points are not available for this paper at this time.
Battery Management Systems (BMS) are crucial for the safe and efficient operation of lithium-ion batteries in applications ranging from electric vehicles to grid storage. While Artificial Intelligence (AI) and Machine Learning (ML) have significantly advanced BMS capabilities, particularly in state estimation and fault diagnosis, the inherent 'black-box' nature of many complex models raises concerns about reliability, trustworthiness, and safety. Explainable Artificial Intelligence (XAI) offers methods to render these AI/ML models transparent and interpretable. This paper provides a comprehensive review of the application of XAI techniques within various BMS tasks. We survey the literature on XAI applied to state-of-charge (SOC), state-of-health (SOH), and remaining useful life (RUL) estimation, as well as fault detection and diagnosis, and charging management. Key XAI methodologies employed in BMS research, such as SHAP, LIME, attention mechanisms, and inherently interpretable models, are discussed. We analyze current trends, identify significant challenges including real-time implementation, evaluation of explanations, and data limitations, and suggest promising future research directions. This review aims to serve as a valuable resource for researchers and practitioners seeking to develop more transparent, reliable, and trustworthy intelligent BMS solutions.
Jong Myoung Kim (Tue,) studied this question.