Revolutionizing proton exchange membrane fuel cell modeling through hybrid aquila optimizer and arithmetic algorithm optimization

Parameter identification in a Proton Exchange Membrane Fuel Cell (PEMFC) entails the application of optimization algorithms to ascertain the optimal unknown variables essential for crafting an accurate model that predicts fuel-cell performance. These parameters are typically not included in the manufacturer's datasheet and must be identified to ensure precise modeling and forecasting of fuel cell behavior. This paper introduces a recently developed hybrid algorithm (Aquila Optimizer Arithmetic Algorithm Optimization (AOAAO)) that enhances the AO and AAO algorithm's efficiency through a novel mutation strategy, aimed at determining seven unknown parameters of a PEMFC during the optimization process. These parameters function as decision variables, and the objective function aimed for minimization is the sum square error (SSE) between the predicted and actual measured cell voltages. AOAAO demonstrated superior performance across various metrics, achieving an SSE minimum in comparison to other compared algorithm. AOAAO's robustness was validated through extensive testing with six commercially available PEMFCs, including BCS 500 W-PEM, 500 W SR-12PEM, Nedstack PS6 PEM, H-12 PEM, HORIZON 500 W PEM, and a 250 W-stack, across twelve case studies derived from various operational conditions detailed in manufacturers' datasheets. For each datasheet, both Current-Voltage (I/V) and Power-Voltage (P/V) characteristics of the PEMFCs scenarios closely aligned with those observed in experimental data, affirming AOAAO's superior accuracy, robustness, and time efficiency for real-time fuel cell modeling. In terms of computational efficiency, AOAAO runtime is significantly faster than all compared algorithms, demonstrating an efficiency improvement of approximately 98%.