Parameter identification of PEMFC steady-state model based on p-dimensional extremum seeking via simplex tuning optimization method

For proton exchange membrane fuel cells (PEMFCs), accurate physical modeling and parameter identification play significant roles in simulation and control. To accurately and quickly identify unknown parameters in the model, a p-dimensional extremum seeking via simplex tuning optimization method is proposed in this paper. Different from the meta-heuristic algorithm, this method is a traditional algorithm, which can effectively estimate the unknown parameters of PEMFC model by constructing multidimensional space to find the extremum of the objective function and searching for the optimal value. The effectiveness of the proposed algorithm is verified by the estimation experiments of 7 unknown parameters in two groups of cases. The simulation and experimental data of the fuel cell model agree well in all cases. Also, through the study of four commercial stacks and the comparison of other excellent meta-heuristic algorithms, the potential of the proposed method in various operation scenarios is demonstrated, and the proposed algorithm has good performance in precision, convergence speed, and stability. Moreover, the accuracy of the model, the validity, and the robustness of parameter identification are further verified by new fitting and verification experiments. The proposed algorithm is a promising way to the most optimization problems in the engineering field.