Optimizing PEM Fuel Cell Models with Bio-Inspired Algorithms

Abstract

With hydrogen gaining traction as an energy source, polymer electrolyte membrane fuel cells (PEMFCs) stand out as a leading clean energy solution. Yet, precise modeling is essential for optimizing their design and performance. To overcome this hurdle, researchers developed the enhanced salp swarm algorithm (ESSA). This bio-inspired metaheuristic innovatively tackles the challenge of identifying unknown PEMFC model parameters. ESSA integrates opposition-based learning, a novel "exploration salps" mechanism, genetic algorithm-style crossover and mutation, and a "survival of the fittest" selection process. These enhancements effectively overcome the limitations of the standard salp swarm algorithm, leading to greater population diversity, faster convergence, improved exploration capabilities, and higher accuracy. Extensive experimental validation on commercial 250W and 500W PEMFC stacks confirmed ESSA's superior ability to provide more accurate and stable parameter estimations with quicker convergence compared to numerous state-of-the-art algorithms, representing a substantial leap in PEMFC parameter estimation and optimization.