Modeling and control of hydrogen production systems through water electrolysis and res power

The increased consumption of fossil fuels has led to scientific interest in establishing a hydrogen-based infrastructure. Currently, there are different electrolysis technologies for H₂ production: i) PEM, ii) alkaline (AFC) and iii) solid oxide (SOEC). PEM electrolysis uses a hydrated polymer membrane, with an operating temperature window of 45-85 °C and reaching power capacities of >500kWs. AFC electrolysis uses a caustic solution as an electrolyte with operating temperatures in the range of 65-220 °C and can also reach high power capacities. Finally, SOEC units use a ceramic electrolyte with an operating temperature of 600-1000 °C and power capacities that can reach several MWs (including thermal exploitation). In terms of costs, PEM and AFC are currently the most economical and flexible solutions.

The aim of this study is the dynamic modeling and control of an electrolysis system. The modeling module is based on submodels that use a combination of electrochemical, thermal, and mass transport equations to fully describe the system operation (see Fig.1). The electrical model takes as input the electrolyzer temperature and the applied current for the voltage calculation. In the mass transfer model, the molar flows/concentrations of water, hydrogen and air (oxygen and nitrogen) are based on the provided electric current. Finally, the thermal model simulates the dynamics of heat and temperature variations as it is provided by the V-I electric model and the applied cooling utilities.

Feedback controllers will be designed for the a) regulation of the hydrogen production by manipulating the applied electric current and b) regulation of the temperature by the manipulation of a coolant flowrate. Conventional P/PI/PID controllers, model-based controllers and optimal controllers are evaluated towards enhanced disturbance rejection and set-point tracking. Performance criteria such as ISE, IAE, ITAE and ITSE will be evaluated for the best controller selection, also based on low power consumption and high efficiency.