Model-based optimization of alkaline electrolysis systems for hydrogen production

ARPAJOU, CANDELARIA; DIEGO G. OLIVA; IGNACIO SCHMIDHALTER; MIGUEL C MUSSATI; AGUIRRE P.; MOROSUK, TATIANA; MUSSATI, SERGIO F.

VARSOVIA

Congreso; 7" International Conference on Contemporary Problems of Thermal Engineering CPOTE 2022, 20-23 September 2022, Poland; 2022

SILESIAN UNIVERSITY OF TECHNOLOGY

Hydrogen plays a crucial role in the sustainable transformation of the energy systems. It is certainly anessential factor for achieving the decarbonization of different sectors such as industry and transport. Waterelectrolysis using electricity generated from renewable energy sources, mainly wind and solar, is amongthe most environmentally friendly hydrogen production processes. Despite being the most maturetechnology at the moment, there is still room for improvements concerning cell materials, components,dimensions, and the process itself. In this paper, the focus is on an alkaline water electrolysis process.Model-based simultaneous optimization of the geometric dimensions and operating conditions such as celltemperature, electrolyte concentration, applied electrolyte pressure, and current density of an alkalinewater electrolyzer is addressed. To this end, a nonlinear mathematical programming (NLP) optimizationmodel, based on first principles, is developed. Gradient-based deterministic optimization 1s performed. Inaddition to the electrochemical reactions, the phenomena taken into account for the material balance arethe mass transfer of the electrolysis products from the solution to the rising bubbles, the gas crossoverthrough the separator (diaphragm), and the reaction rates at the electrodes. The model 1s firstly validatedusing two reference cases reported in the literature. Then, given process data and specifications as well asbounds on variables, the values of operating conditions and geometric dimensions that maximize cellefficiency are simultaneously optimized. In addition, the influence of critical operating variables on theoptimal solution is investigated. Regarding computational aspects, the model is implemented in GeneralAlgebraic Modeling System (GAMS) software and solved using CONOPT solver