Robust parameter estimation of proton exchange membrane fuel cell using Huber loss statistical function

A key area of research in a promising renewable energy technology called Proton Exchange Membrane Fuel Cells (PEMFCs) focuses on identifying parameters not provided by manufacturers’ datasheets and developing highly accurate models for PEMFC voltage-current characteristics. In this regard, a precise model is crucial for designing effective PEMFC systems. This study aims to discover the seven unknown parameters of the steady-state model for PEMFCs using a recent optimization algorithm called Educational Competition Optimizer (ECO). The ECO is used to reduce the effects of local optimal stagnation and early convergence, commonly observed in literature approaches. The goal is to improve model parameter correctness by reducing errors between experimental and predicted polarization curves. A robust regression fitness function known as Huber loss is used to decrease inaccuracies between experimentally measured voltages and their corresponding calculated values. The present research examines three test cases of well-known commercial PEMFC units as benchmarks under various steady-state operation situations. The simulation results show that the suggested model is significantly more accurate than the best alternative technique and achieves high closeness to the experimental records. The article compares the ECO against current optimizers in the literature to assess its feasibility. Based on the findings of this study, the prospective Huber loss function increases the optimizer’s resilience and robustness compared to steady-state error.