Artificial ecosystem optimizer for parameters identification of proton exchange membrane fuel cells model

Abstract Modeling and optimization of the proton exchange membrane fuel cells (PEMFCs) raise a crucial challenge due to their characteristics of multi-variability and nonlinearity natures. To ensure an accurate and reliable model for PEMFCs, best values of their uncertain parameters should be defined carefully. A conventional artificial ecosystem optimizer (AEO) and an improved and developed AEO (called IAEO) are used to realize the later aim. In the proposed IAEO, a dynamic crossover pattern is presented to enable the algorithm to achieve better solution, and also prevent the stuck in local optima. Sum of squared errors (SSE) defines the fitness function subjected to set of practical constraints. The proposed IAEO-based algorithm is analyzed and demonstrated on different typical benchmarking PEMFCs modules widely used in the literature. Comprehensive simulations and performance assessments are carried out on the PEMFCs models to affirm the efficacy and robustness of the proposed IAEO based on methodology while simulating the commercial PEMFC stacks behavior in regards to the experimental data. In this context, best values of SSE resulted by IAEO are 0.0116, 0.3359, and 2.1459, for BCS 500-W, 250 W stack, and NedStack PS6, respectively that are very competitive values among other challenging methodologies. This noticeably indicates that the developed IAEO-based method gives better efficiency with the highest robustness and convergence speed compared with the other methods. At a later stage, dynamic performance of PEMFCs stacks are carried out. It can be established that the reported outcomes affirm the superiority and reliability of the IAEO algorithm over the conventional AEO and the other competitors.