Degradation prediction and remaining useful life estimation of PEMFCs: Mechanisms, methods, datasets, and challenges

Proton Exchange Membrane Fuel Cells (PEMFCs) represent a pivotal technology for sustainable energy conversion in automotive, portable, and stationary applications due to their high efficiency, rapid start-up capability, and near-zero emissions. However, widespread commercialization remains severely constrained by uncertainties related to operational durability, cost, and reliability. Consequently, accurate degradation prediction and remaining useful life estimation methods have become critical for facilitating predictive maintenance, which can improve reliability, and reduce lifecycle costs. This review synthesizes recent advances in PEMFCs prognostics, which integrate fundamental degradation mechanisms. Degradation mechanisms are categorized into irreversible and reversible mechanisms. In particular, the review provides protection measures against irreversible and reversible degradation. Subsequently, the review systematically compares various prognostic methods, including model-based model, advanced data-driven model, and hybrid degradation model. Moreover, both publicly available and proprietary PEMFCs durability datasets are systematically collected for the first time. Furthermore, key performance evaluation metrics for fuel cell prognostics models are thoroughly discussed. Finally, significant research challenges and promising future directions are identified, which reveal three key opportunities such as physics-informed artificial intelligence, standardized datasets benchmarking, and real-time onboard health prediction. All in all, this review systematically synthesizes fuel cell degradation mechanisms, prediction methods, aging datasets, and evaluation metrics, which provides a foundational reference to accelerate research in durability enhancement and predictive maintenance for next-generation fuel cell systems.