Developing the coyote optimization algorithm for extracting parameters of proton-exchange membrane fuel cell models

Developing a precise semiempirical mathematical model based on multi-nonlinear equations for the proton-exchange membrane fuel cell (PEMFC), which guarantees suitable and accurate simulation of the electrical characteristics of typical PEMFC stacks under various operating scenarios, is the main target of this study. The unknown parameters of the PEMFC model are extracted using a novel efficient optimization technique called coyote optimization algorithm (COA). To validate the effectiveness of the proposed COA-based PEMFC model, two different cases of seven and ten unknown parameters are performed on a commercial PEMFC taken from literature. The sum of squared errors (SSE) between the experimentally measured data and the corresponding computed ones is considered as the objective function. Besides, the effectiveness of the developed algorithm is validated under different operating conditions. Moreover, the results obtained by the application of the proposed COA have been compared with other recent optimization methods reported in the literature, and very competitive results have been provided. Furthermore, parametric and nonparametric statistical analyses are presented to evaluate the accuracy and viability of the developed COA-based PEMFC model.