Model parameters estimation of PEM fuel-cell systems using Genetic Algorithms

This paper presents a novel Fuel Cell (FC) model parameters estimation method based on a Genetic Algorithms (GAs) heuristic optimization approach, a theoretical model and experimental measurements. The experimental behaviour of the FC is reproduced by an hardware emulator controlled to generate the FC I-V characteristic by means of a Ballard Mark V model. The close match between experimental measured results and those achieved by the same model with optimized unknown parameters shows that the use of the proposed strategy is an effective and consistent method for parameters identification and for correctly modelling the behaviour of the fuel cell system. It also proves the reliability of the designed hardware emulator. Significant improvements in terms of identification accuracy and reasonable optimization time are also achieved using the Search Space Reduction Method (SSRM).