Abstract
This article presents a new Hybrid Grasshopper Optimization Algorithm (HGOA) that can be used to model Proton Exchange Membrane Fuel Cells (PEMFCs) with high accuracy and optimize multi-objectives. PEMFCs play a central role in clean energy systems because of their high efficiency, zero emission, and their use in transportation, portable power, and stationary energy solutions. Nevertheless, they have nonlinear electrochemical dynamics and are sensitive to operating conditions, which are major challenges in parameter identification and performance prediction. Current metaheuristic algorithms are usually plagued by premature convergence, inefficient computation and instability when solving such complex, multi-variable systems. The new HGOA overcomes these drawbacks by combining elite retention, opposition-based learning, feasibility repair, and local search with the standard Grasshopper Optimization Algorithm (GOA). These modifications make convergence faster, more diverse in solutions and more robust in parameter tuning. Validation was performed on seven PEMFC test cases (FC1-FC7) with various operating conditions. The findings indicate that HGOA has better accuracy with the lowest Absolute Error (AE = 0.0026), Relative Error Percentage (RE% = 0.0613%) and almost zero Mean Bias Error (MBE) recorded in all cases. Moreover, HGOA is computationally more efficient, stable and predictively reliable than nine state-of-the-art metaheuristic algorithms, such as GOA, ECO, RIME, EO, and PO.