Novel control approach for integrating water electrolyzers to renewable energy sources

Green hydrogen can be produced by integrating water electrolysers to renewable energy (RE) sources. This integration encounters the issue of power volatility of renewables that requires advanced control strategies for electrolyser operation. Three main control approaches can be found in the literature based on the control concepts of (i) battery hysteresis cycle, (ii) off-line ahead scheduling, and (iii) frequency response. These approaches do not solve fully the problem of electrolyser operation under power fluctuation conditions. This study introduces a novel integration and control approach for such systems. It is based on model predictive control algorithm to control electrolyser load so that steers the system into energy balance in the real-time. Simplified prediction models are used for both generation and demand sides with a novel compensator for model uncertainty. The approach is tested on a 5 kW microgrid that includes RE emulators and a PEM electrolyser. The results show that fully automated real-time energy balance is achievable for grid connected and off-grid systems. In this approach, the electrolyser will operate at partial load capacity with a higher efficiency and therefore higher hydrogen yield. The energy balance is subject to power conditioning losses and capacity constraints and it is applicable to any mix of fluctuated RE sources.