Numerical optimization of ejector for enhanced hydrogen recirculation in proton exchange membrane fuel cells

In proton exchange membrane fuel cell (PEMFC) systems, ejectors enable hydrogen recirculation without parasitic power consumption. However, their performance is highly sensitive to design parameters and operating conditions, often leading to inefficiencies under off-design conditions. This study develops a comprehensive numerical optimization framework integrating computational fluid dynamics (CFD), design of experiments (DoE), regression modeling, and multi-objective optimization to enhance ejector performance. A Box-Behnken design explores five key geometrical parameters, while a quadratic regression model establishes correlations between design variables and performance. Two optimization techniques, Non-dominated Sorting Genetic Algorithm (NSGA-II) and Desirability Function (DF), are applied to maximize the entrainment ratio while maintaining choked flow conditions with Mach number specifically considered at the nozzle throat. Results identify nozzle throat diameter (NTD) and nozzle exit position (NXP) as most critical parameters governing ejector performance. The optimized ejector achieves a 20 % entrainment ratio improvement and enhanced performance across design and off-design conditions. Additionally, optimization suppresses shockwave formation, improving flow stability and recirculation efficiency. This study introduces a novel simulation-based optimization approach for PEMFC ejectors, providing a systematic methodology to improve efficiency and adaptability. The findings advance hydrogen fuel cell technology by improving fuel utilization and operational flexibility, enhancing ejectors viability for real-world applications.