Parameter Estimation of PEMFC Model Via Glider Snake Optimization Algorithm

This study proposes a novel approach based on the Glider Snake Optimization (GSO) algorithm for estimating the seven unknown parameters of proton exchange membrane fuel cell models. The proposed GSO method was tested on three different fuel cell models: Nedstack PS6, BCS 500W and Ballard Mark V. Using the experimental voltage-current (V-I) polarization data of the fuel cells, the seven unknown parameters of the semi-empirical model (ξ₁, ξ₂, ξ₃, ξ₄, λ, Rc, β) were estimated by the proposed GSO method. With the help of the obtained parameters, the output voltage of the fuel cell was calculated with high accuracy. By using the proposed GSO method, Sum of Squared Error (SSE) values of 2.0655, 0.0193 and 0.7598 were obtained for the Nedstack PS6, BCS 500W and Ballard Mark V models, respectively. Comparative results with recent literature methods indicate that the proposed method provides performance improvements of 3.88%, 24.02% and 6.51% over the E-ABC, TDE and EMEDEA algorithms for the Nedstack PS6, BCS 500W and Ballard Mark V models, respectively. The main contribution of this study lies in introducing a recent, effective and practical heuristic optimization approach based on the GSO algorithm for PEM fuel cell parameter estimation. The best SSE values obtained for different PEM fuel cell models and the superiority achieved over current optimization methods reported in the literature demonstrate that the proposed method offers a competitive and effective alternative for estimating semi-empirical parameters based on experimental measurements.