Water Failure Diagnosis of PEMFC in Actual Operating Conditions With High-Frequency Impedance Features

The water failures decline the performance of proton exchange membrane fuel cell (PEMFC) system remarkably in practical applications. Nevertheless, precise diagnosis of water faults remains challenging due to their intricate nature and the difficulty in distinguishing similar water states. This complexity is further enlarged for diagnosing the high-power stacks. Motivated by this, a water fault diagnostic method is proposed in this article based on the high-frequency electrochemical impedance spectroscopy (EIS) features. This approach employs 6 EIS features within the 157 Hz to 2 kHz frequency band. The sensitivity of each feature to water state is elucidated utilizing the entropy-weight method, thereby enhancing the discrimination capability to similar water states. Furthermore, a novel autonomous fuzzy clustering algorithm is exploited for clustering analysis of the water failures from practical experiments. The proposed method achieves 93.10% fault recognition accuracy during the application on the commercial PEMFC system under a varying-load condition. By narrowing down the frequency range, the proposed method reduces the diagnostic time by 84.86% while improving the accuracy of diagnosis significantly.