A numerical study of the gas-liquid, two-phase flow maldistribution in the anode of a high pressure PEM water electrolysis cell

In this work, the use of a circular-planar, interdigitated flow field for the anode of a high pressure (HP) proton exchange membrane (PEM) water electrolysis cell is investigated in a numerical study. While PEM fuel cells have separated flow fields for reactant transport and coolant, it is possible to operate a PEM electrolysis cell with the anode flow field serving as both. This allows for a simpler system and a thinner design, however sets new and more strict requirements for the flow field to distribute uniformly. For the conducted numerical study, two computational fluid dynamics models are developed; a single-phase flow model for establishing the effect of geometry and a two-phase flow model for studying the effect of dispersed gas bubbles. Both models account for turbulence and heat transport. In the literature, previous studies on square-planar, interdigitated flow fields for PEM fuel cells have demonstrated particular advantages in regard to increasing the limiting current density and achieving a uniform distribution of properties. In the present study, it is however found that the circular-planar shape of the interdigitated flow field causes maldistribution. Moreover, under nominal conditions and below a water stoichiometry of 350, flow and temperature maldistribution is adversely affected by the presence of the gas phase; particularly gas hold-up near outlet channels can cause excessive formation of hotspots. As the water stoichiometry increases above a water stoichiometry of 350, the flow maldistribution increases. Nonetheless, its impact on the temperature distribution is counterbalanced by the overall increase in heat capacity of the flow. Hence, a relative uniform temperature distribution is achieved at and above nominal flow conditions. Finally, in a parametric investigation it is underlined that the predicted formation of hotspots is sensitive to the employed particle diameter of the two-phase channel flow model. The larger the particle size, the more severe the maldistribution of temperature becomes. It is therefore concluded that further experimental validation and research into particle size modeling is necessary.