Ejector validation in proton exchange membrane fuel cells: A comparison of turbulence models in computational fluid dynamics (CFD) with experiment

In order to industrialize proton exchange membrane (PEM) fuel cells and balance of plant components, simulations are utilized to minimize costs during development and maximize performance. A key component in the anode path is the ejector, which is often optimized by 2D CFD to maximize the entrainment ratio. However, most turbulence models for 2D CFD simulations do not accurately predict the ejector's entrainment ratio across the entire operating range. This study involves validating various turbulence models using experimental data from two distinct ejectors. The optimal turbulence model, identified as the Reynolds Stress Model, achieves an average deviation of 6.1% in the entrainment ratio between simulation and experiment for both ejectors. This represents a significant improvement compared to traditional k-epsilon and k-omega turbulence models. The proposed turbulence model minimizes deviations in actual operation, thereby reducing testing and development costs, and contributing to the rapid industrialization of PEM fuel cell technology.