Air supply temperature impact on the PEMFC impedance

摘要

In this work, a neural network is used to optimize the impedance model of a proton exchange membrane fuel cell (PEMFC). In particular, the model is used to diagnose the effect of the air supply temperature on the water management and humidification of the fuel cell, for an interval of air supply temperature of 60–140 °C to predict the hydration state of the fuel cell (flooding and drying). The Multi-Layer-Perceptron through back propagation training algorithms of the neural network model satisfactorily predicts parameters (the mean square error value of ANN is 6.67 e−16). The neural network method is applied to investigate the PEMFC impedance under drying and flooding conditions. The proposed model is validated by the experimental results of the literature at a frequency range of 0.1 Hz–1 kHz, RH = 10%, t = 180 s for drying case and RH = 100%, t = 500 and 1600 s for flooding case. The proposed approach enabled us to identify the five physical parameters of the impedance that exhibited high sensitivity in the PEMFC diagnostic (the membrane resistance, the polarization resistance, the double layer capacitance of the electrode/electrolyte interfaces, the resistance of species diffusion and the time constant of the charges diffusion).