Solid oxide fuel cell simulation model tuning using operating conditions dependent optimization techniques

This paper discusses the mathematical model and simulation of Solid Oxide Fuel Cell (SOFC), where the conventional and reversible SOFC are studied. Their performance is studied at the steady state to get the V-I polarization curves, as well as in case of changing the electric current drawn from the cell to get the change in voltage over time. Unlike most existing studies, the focus is on the time needed by the cell to reach voltage stability after a change, in preparation for studying the cell integration into an electrical network, and all of these studies are carried out when changing the cell operating conditions such as temperatures and the flow rate of reactive gases. Then optimization techniques are used, such as Walrus Optimization Algorithm (WaOA), Secretary Bird Optimization Algorithm (SBOA), Chaos Game Optimization (CGO) and Teaching-Learning Based Optimization (TLBO) to increase model accuracy and make its results closer to the published laboratory results. The results demonstrate that WaOA achieves the lowest modeling error among the investigated optimization techniques and provides the most accurate model calibration. The proposed WaOA-based tuning reduces the modeling error by more than five orders of magnitude compared to curve-fitting approaches reported in the literature, and is further employed to optimize reactant flow rates, resulting in an output power increase of approximately 6% compared to nominal operating conditions.