Advanced ejector design for hydrogen fuel cells using computational simulations and genetic algorithms

Hydrogen recirculation is a key component in Proton Exchange Membrane Fuel Cell (PEMFC) systems, essential for minimizing fuel loss, enhancing efficiency, and supporting water management. This study investigates a passive recirculation strategy using an ejector, an energy-efficient device with no moving parts, making it suitable for applications where simplicity, reliability, and weight are critical, such as in automotive and aerospace sectors. A hybrid computational framework is developed by integrating high-fidelity Computational Fluid Dynamics (CFD) simulations with a Genetic Algorithm (GA) to optimize key ejector parameters, including nozzle radius, exit position, mixing chamber length, and diffuser radius. Optimization is performed directly on full CFD simulations using a high-performance computing cluster, avoiding surrogate models and ensuring accuracy. The model is validated against experimental and theoretical data from the literature. Optimization aims to maximize suction capacity based on a commercial 110-cell PEMFC stack rated at 20.4 kW, operating between 0 and 180 A. The resulting design achieves stable performance from 48 An upwards, significantly broadening the operational range compared to prior ejector configurations. In addition to providing a technical solution for fuel cell gas management, this study establishes a transferable optimization methodology that can support ejector integration in other clean energy systems.