Diagnosis of proton exchange membrane fuel cell system based on adaptive neural fuzzy inference system and electrochemical impedance spectroscopy

Abstract

A new diagnostic method based on adaptive neural fuzzy inference system (ANFIS) and electrochemical impedance spectroscopy (EIS) is proposed for the proton exchange membrane fuel cell (PEMFC) system. Firstly, a new parameter identification method that combines genetic algorithm (GA) and Levenberg–Marquardt (LM) algorithm is proposed to identify the fractional-order equivalent circuit model (ECM), in which the anode impedance, cathode impedance, and mass transfer are all considered. This new method allows better exploitation of the EIS diagrams, and the internal relationships between the fault conditions and the ECM parameters are thoroughly analyzed according to it. Then, based on these relationships, a new diagnostic algorithm based on k-means clustering and ANFIS is designed to precisely identify several faults that can occur in the PEMFC, such as membrane flooding, drying, and mass transfer fault. Finally, the effectiveness of this method is demonstrated experimentally through the exploitation of EIS data under different faults and operating conditions of the PEMFC.