Development of a decision-oriented tool for operational management of a PEMWE

The hydrogen production via electrolysis represents a key pathway for decarbonizing shipping and contributing to carbon neutrality goals, as it enables the generation of green hydrogen with potential zero emissions. The electrolysers system management directly impacts on the cost-effectiveness of green hydrogen production and thus enhancing the overall competitiveness of the electrolysis technology. In the context of Hydrogen Valleys development hydrogen production via Proton Exchange Membrane Water Electrolysis (PEMWE) is gaining increasing attention, with approximately 20% more facilities adopting PEM technology compared to alkaline ones, especially in renewable energy sources (RES) powered plants. However, in large-scale industrial applications, for a determined market demand and in the case of coupling PEMWE with RES, optimal management strategies for hydrogen production are required to ensure effective operation under dynamic power supply conditions. From literature, it is highlighted that in order to ensure the economic viability of hydrogen production, a holistic approach that consider all interconnected variables of PEMWE under dynamic operation should be considered. In this context, data-driven approaches, such as Machine Learning and Digital Twins models, will play a key role in managing electrolysis systems under dynamic power conditions, enabling strategic decisions to be made for electrolyser operation. This research is to develop a comprehensive numerical model of a PEMWE offering insights into system management under variable load conditions driven by RES availability, electricity market price, and weekly hydrogen demand, while avoiding operational conditions that could lead to increase the energy intensity of the process. The work it is placed in a wider framework in which the resulting datasets, obtained through the model numerical simulation, are used to train a physics-based Digital Twin.