Design and optimisation of alkaline water electrolysers depend on modelling equations. One of these equations relate to the gas hold-up, a key factor influencing performance of electrolysers. In electrolysis setups, a maximum gas hold-up is observed which affects the resulting gas hold-up. Gas hold-up prediction models adhere to this maximum gas hold-up, yet its exact nature remains open to discussion. Additionally, relations of slip velocities within the electrochemical multi-phase flows in electrolysis setups are not well understood yet. Several studies propose empirical models, but the mapping of gas hold-up and slip in terms of operating conditions is limited.
Accordingly, this study involves experimental measurements of gas hold-up across a broad range of operating conditions in terms of current densities and superficial liquid velocities. Gas hold-up prediction models are tested on their validity across these operating conditions. Both an explicit as well as an implicit 1D model are addressed and compared to the existing Coalescence Barrier Model. Validity ranges for specific slip and gas hold-up relations have been identified.
The explicit 1D model shows good agreement with flows that are dominated by liquid flow inertia and have relatively small slip. On another note, it was found that slip is pronounced intensely at higher current densities, which led to the proposal of a slip boosting factor. As its exact origin is unclear, it raises more questions on the validity of existing slip functions at higher current densities. A transition towards slug flow might explain this, but should be studied further.
Further results conclude a consistent and physically sensible maximum gas hold-up of ⟨ε⟩_max = 0.71 for the proposed implicit 1D models, compared to a rather arbitrary ⟨ε⟩_max = 0.49 for the Coalescence Barrier Model.
Additionally, slip functions have been assessed. Results over the assessed range of operating conditions indicate accurate gas hold-up predictions using an empirical hindrance factor of (1-⟨ε⟩)^2.54, which aligns closely with the theoretically derived Marrucci equation.