Model predictive control of alkaline water electrolyzers with storage constraints

The flexibility of alkaline water electrolyzers is limited by requirements on the gas purity (hydrogen in oxygen) to avoid the creation of explosive gas mixtures. To increase operational flexibility, recent works have proposed to control the gas purity by adjusting the lye recirculation rate and the system pressure. These approaches assume that the pressure can be freely set in a given range, which may not be a valid assumption. If the produced gas is stored on-site, the pressure in the electrolyzer is coupled to the pressure in the storage tank. Depending on the size of the gas storage and constraints on the compression equipment, it may then not be possible to set the pressure freely. In this work, we present a model predictive control (MPC) implementation for alkaline water electrolyzers with such constraints. We test the system controlled by MPC against historic wind power data with different demand profiles and storage sizes. Our results show that MPC is effective at simultaneously handling gas purity and storage constraints. However, small storage sizes can reduce the range in which the pressure can be used to control the hydrogen-to-oxygen ratio, leading to flexibility bottlenecks that were previously unaccounted for.