PEMFC parameter identification using a Reverse-Search Chaos Differential-Evolution Whale Optimization Algorithm

This study proposes a meta-heuristic approach, the Reverse-Search Chaos Differential-Evolution Whale Optimization Algorithm (RSCDWOA), to accurately estimate unknown parameters in semi-empirical models of proton exchange membrane fuel cells (PEMFCs). Accurate modeling and parameter identification are crucial for optimizing PEMFC performance, and the RSCDWOA algorithm is designed to address the challenges posed by the nonlinear, multivariable nature of the PEMFC voltage-current characteristic curve, mitigating the risk of local optima convergence and reducing computational time compared to traditional optimization methods. This study introduces a new algorithm for efficiently identifying unknown PEMFC model parameters, employing a threestage process: Differential Evolution, Exploration 1, and Exploration 2. These stages work synergistically to minimize the discrepancy between measured and estimated data, leading to optimal solutions. The algorithm's performance is validated by extracting seven unknown parameters from twelve commercial PEMFC systems, with comparative analysis against established optimization methods demonstrating its superior convergence speed, stability, and accuracy. Furthermore, simulation results under dynamic operating conditions confirm the algorithm's robustness and high computational efficiency in parameter identification. The running time of sheet 1 for RSCDWOA is the least, which is 0.198984s, and it accounts for approximately 42%, 72%, 86%, 80%, 70%, 60%, 70%, 59%, 67%, and 39% of the runtimes of RIME (4.40233s), GNDO (4.213444s), GBO (8.186774s), GBO (8.186774s), AO (4.937394s), EDO (9.469845s), GMO (5.347436s), PSO (5.469166s), GA (6.174074s), and WOA (9.537138s) respectively. This indicates that RSCDWOA is significantly more efficient in terms of runtime compared to other algorithms in the study.