Designing a hydrogen gas ejector for 5 kW stationary PEMFC system: CFD- modeling and experimental validation

K. Nikiforow (Corresponding Author), Pauli Koski, Henri Karimäki, Jari Ihonen, V. Alopaeus

    Research output: Contribution to journalArticleScientificpeer-review

    82 Citations (Scopus)


    Ejectors are durable and inexpensive equipment for realizing hydrogen recirculation in proton exchange membrane fuel cell (PEMFC) systems. In the present work, a hydrogen recirculation ejector targeted for high turndown ratio operation in a 5 kWe PEMFC system was designed, manufactured with 3D-printing, and characterized experimentally with both air and humid hydrogen. The ejector was modeled at the experimental conditions with computational fluid dynamics (CFD) assuming 2D axisymmetric flow and with three turbulence models. A systematic comparison of experimental and simulation results was conducted with humid hydrogen at conditions covering the entire operating map up to 6 bar gauge primary pressure. The simulation results deviate on average 60%-70% from the experimental results, the deviation being less pronounced at conditions relevant in PEMFC applications. The SST k-? turbulence model was identified to agree best overall with the experimental data while the RNG and Realizable k-e turbulence models were observed to accurately predict the position of maximum ejector efficiency. Hence, the SST k-? model is more useful for predicting ejector performance while one of the two k-e models should be adopted when optimizing ejector design.
    Original languageEnglish
    Pages (from-to)14952-14970
    JournalInternational Journal of Hydrogen Energy
    Issue number33
    Publication statusPublished - 2016
    MoE publication typeA1 Journal article-refereed


    • CFD
    • ejector
    • hydrogen recirculation
    • PEMFC
    • turbulence models


    Dive into the research topics of 'Designing a hydrogen gas ejector for 5 kW stationary PEMFC system: CFD- modeling and experimental validation'. Together they form a unique fingerprint.

    Cite this