Online fault detection and isolation of PEMFC based on EIS and data-driven methods: Feasibility study and prospects

Abstract

Electrochemical impedance spectroscopy (EIS) can be useful for the mechanism analysis and diagnosis of proton-exchange membrane fuel cell (PEMFC) performance degradation. This review summarizes the potential of using EIS for real-time fault detection and isolation of the PEMFC by data-driven methods from the following aspects. First, the data-driven diagnosis strategy of PEMFC based on EIS is overviewed; the typical faults and EIS measurement for data collection are briefly introduced. Then, the application of EIS in the online data-driven diagnosis of PEMFC is analyzed and discussed, focusing on feature extraction from EIS, diagnosis models employing various machine learning methods, and the corresponding EIS features for each machine learning method. Finally, the feasibility of using EIS for online data-driven fault diagnosis of PEMFC is briefly summarized, and the research challenges and prospects are proposed. This review aims to provide inspiration and new insights for future research on online PEMFC diagnosis, prognostics, and health management.