Hydrogen produced by PEM water electrolysis emerges saturated with water vapor and must be dried before storage or use. The efficiency of the regeneration step strongly governs both cycle duration and thermal input. Induction heating can accelerate this step by generating heat directly within the metallic vessel wall, but the resulting axial temperature field is non-uniform and affects the two adsorbent layers differently. Prior electromagnetic–thermal and packed-bed simulations identified the upper zeolite 13X layer as the main regeneration bottleneck: the lower activated-alumina layer achieved over 99% water removal within 2800–3600 s, whereas the upper zeolite 13X layer reached only 26.2–35.0% removal at 5000 s, with a mean solid temperature of 122–139 °C.
This study extends that work by examining whether purge direction and purge flow rate can alleviate this limitation, adding experimental validation. The dryer is a SUS304 vessel with a dual-layer bed, consisting of activated alumina in the lower 40% and zeolite 13X in the upper 60%. Electromagnetic losses in the vessel wall (from ANSYS Maxwell 3D) are passed to a transient thermal model, and the resulting wall-temperature history drives a 3D packed-bed model in ANSYS Fluent, with water transport described by a linear driving force formulation using layer-specific equilibrium relations.
Co-current and counter-current purge configurations are compared across a range of flow rates. In counter-current flow, purge gas reaches the thermally disadvantaged zeolite layer first; in co-current flow, it is preheated by passing through the alumina layer first but then carries alumina's desorbed moisture through the zeolite, representing a trade-off between thermal management and moisture transport. Comparison metrics include axial bed temperature, layer-wise water removal, outlet moisture history, regeneration time, and purge demand.
Experiments use the same induction-heated dryer, with wall/bed thermocouples and outlet dew-point plus measured purge flow used to estimate cumulative water removal; air serves as the experimental purge gas for safety. The model is first validated against air-purge experiments, then applied with hydrogen properties to predict hydrogen-purge performance. The combined results aim to identify a purge direction and flow-rate range that reduces the zeolite regeneration bottleneck while limiting purge consumption, providing an operating basis for induction-heated hydrogen dryer regeneration.

Hydrogen drying,Temperature swing adsorption,Induction heating,Purge flow direction,Purge flow rate