An enhanced Draco lizard optimizer for accurate parameter extraction of proton exchange membrane fuel cells

Abstract

Abstract Accurate parameter extraction is crucial for the modelling of proton exchange membrane (PEM) fuel cells, which involves complex, non-linear, and multivariate relationships essential for simulation, design, and fault diagnostics. This paper proposes a Modified version of the Draco Lizard Optimizer (MDLO) technique to precisely extract important PEM fuel cell parameters. This hybridization aims to increase optimization efficiency by striking a balance between exploration and exploitation. The efficacy of MDLO is supported by extensive simulations that use three commercially available PEM fuel cell systems to compare its performance to that of the conventional DLO and new metaheuristic optimization approaches, which are Driving Training-Based Optimization (DTBO), Moss Growth Optimization, and Skill Optimization Algorithm (SOA). Best fitness, average fitness, worst fitness, standard deviation, convergence speed, and multiple-comparison test are among the performance indicators that are applied and measured during the course of 55 runs. According to the findings, MDLO provides the best Sum of Squared Errors (SSE) value, greater accuracy, dependability, speed of convergence, and a strong fit for the estimated primary parameters. The runs’ low and consistent SSE values—0.331348 for the 250 W, 1.1698 $$\:\times\:$$ 10 − 2 for the BCS 500 W, and 2.100246 for the NedStack PS6—provide effectiveness and robustness of the MDLO.