Advancing battery prognostics and health management: Challenges and future perspectives

Battery prognostics and health management (PHM) research has advanced along three primary directions: physics-based modeling, data-driven approaches, and hybrid methods. Despite these developments, this field appears to be approaching a bottleneck. Fundamental aspects, such as internal degradation mechanisms, material evolution, multi-physics coupling, and innovative diagnostics, remain underexplored. This has led to homogenized research, where innovation is concentrated on model tuning, rather than deeper understanding of battery systems. This paper provides a systematic review of battery PHM research from mainly the past decade, highlighting major advances, identifying key challenges, and outlining future directions. We focus on three critical questions: how to balance data-driven and physics-based approaches; how to construct cross-scale, multi-physics health models; and how to integrate machine learning with mechanistic insights and experimental validation. Our aim is to offer an interdisciplinary perspective that advances battery PHM research beyond predictive accuracy, toward mechanistic understanding and reliability-oriented design.