Parameter extraction of polymer electrolyte membrane fuel cell based on quasi-dynamic model and periphery signals

Abstract

It is important to extract parameters of a polymer electrolyte membrane fuel cell (PEMFC) using periphery signals. The main contribution of this work is to introduce a simple yet effective method for parameter-extraction basing on a quasi-dynamic model for a single PEMFC and periphery signals. The model includes filling-and-emptying sub-models, which set up relations between periphery signals and internal states, and a static water transferring sub-model for the membrane. The parameter-extraction method with 5 steps for 9 key parameters is proposed, drawing on experiments and algorithms of nonlinear least square (NLS) and neural networks (NN). Comparison of the identified parameters to data in literature shows that, the results in our study are reasonable.A dynamic experiment is carried out to verify the model. Relative errors within [-5, 5]% between simulating and experimental results are observed, showing the effectiveness of the results. Properties of internal states with respect to time and frequency are simulated. A net water transport coefficient β∈[0.13, 0.21] is predicted. The normalized transfer functions of small disturbance signals from the cell current to internal states are low-frequency-pass functions. A cutoff frequency (0.0003–0.37 Hz) and a resonating frequency (3.55 Hz), which retain under different operation conditions, is found.