A Review on Data‐Driven‐Based Remaining Useful Life Prediction Methods for Proton Exchange Membrane Fuel Cells

Proton exchange membrane fuel cells (PEMFCs) are considered as one of the leading technologies for future new energy vehicles due to their advantages, including zero emissions and high efficiency. However, the large‐scale commercialization of PEMFCs is still constrained by durability issues. Remaining useful life (RUL) prediction enables the estimation of degradation rates in PEMFCs, thereby facilitating the implementation of life‐extension strategies and proactive maintenance before the end of their service life. In recent years, data‐driven methods have become a mainstream approach for RUL prediction, as they effectively capture complex degradation patterns that are challenging to model using traditional methods. Therefore, conducting a systematic and comprehensive review of data‐driven methods for RUL prediction is both scientifically significant and critically important. This paper introduces commonly used datasets and data preprocessing techniques for data‐driven methods, outlining key concepts and steps involved in the RUL prediction workflow. It then reviews the research progress on data‐driven methods for predicting the RUL of PEMFCs, including traditional neural network‐based, classical machine learning‐based, and deep learning‐based methods. In addition, by comparing various methods from three perspectives: accuracy, computational efficiency, and applicability to real‐world systems, this paper aims to provide valuable guidance for future research on PEMFC durability.