Parameter fitting of PEMFC models based on adaptive differential evolution

• A parameter fitting approach based on ADE is proposed for PEMFC polarization curve models.
• ADE adjusts its operators and control parameters online by two adaptation schemes.
• Results show that the proposed approach achieves highly precise and robust results when using noise-free or noised data.

Optimal modeling of proton exchange membrane fuel cell (PEMFC) is extremely important for the understanding and interpretation of the overall cell behavior. Intelligent optimization technique is an effective strategy to optimally model PEMFCs because the determination of model parameters is a complex optimization problem. In this paper, an adaptive differential evolution algorithm (ADE) is proposed to accurately fit parameters of PEMFC models. The main contribution is the proposal of ADE for the parameter fitting of PEMFC models. To avoid premature convergence and increase search efficiency, ADE uses two adaptation schemes to dynamically adjust its variation operators and control parameters. The effectiveness and robustness of ADE is verified by applying it to solve benchmark functions and to extract model parameters of four PEMFCs using both noise-free or noised data. Experimental results demonstrate that the proposed method produces more accurate and robust models than several other algorithms and the polarization curves obtained by ADE match well with the experimental data in all cases. This work indicates that ADE is an effective alternative to solve the PEMFC optimal modeling problem.