GNMSA: An Accurate Parameter Estimation of Semi-empirical Proton Exchange Membrane Fuel Cells Model using A Hybrid Artificial Intelligence–Based Optimization Approach

The PEMFC parameter estimation problem has recently piqued the interest of several researchers due to its significance in reaching better PEMFC modeling. However, existing PEMFC parameter estimation algorithms encounter significant difficulty in solving this problem because of its complex and highly nonlinear nature. Therefore, this paper presents a new hybrid approach, termed GNMSA, for addressing this problem more accurately. This hybrid approach combines the exploration operator of the generalized normal distribution optimization with the exploitation operator of the mantis search optimization (MSA), resulting in this strong variant that has high abilities to alleviate stagnation in local optima and expedite convergence speed. Six PEMFC stacks of varying characteristics are employed in this study to validate the performance of GNMSA at various difficulty levels. Furthermore, it is compared to around thirteen recently published optimization techniques according to several performance metrics to demonstrate its effectiveness and efficiency. According to the experimental findings, GNMSA is the best alternative for predicting unknown PEMFC parameters because it can obtain considerably better and different results on all the studied PEMFC stacks.