Enabling thermal-neutral electrolysis for CO2-to-fuel conversions with a hybrid deep learning strategy

High-temperature co-electrolysis of CO2/H2O through the solid oxide electrolysis cells (SOECs) is a promising method to generate renewable fuels and chemical feedstocks. Applying this technology in flexible scenario, especially when combined with variable powers, requires an efficient optimisation strategy ensure its safety cost-effective long-term operation. To purpose, we present hybrid simulation for accurate fast process SOECs. This builds multi-physics models based on experimental data extends database develop deep neural network genetic algorithm. In case study, thermal-neutral condition (TNC) set as target various operating conditions, where SOEC generates no waste heat needs auxiliary heating equipment. Small peak-temperature-gradient (PTG) inside found at TNC, which vital prevent thermal failure For cell 1023 K 1123 inlet gas temperatures, smallest PTGs reach 0.09 0.31 mm−1 1.13 1.19 V, respectively. Finally, 4-D map presented show interactions among applied voltage, required power density, composition, temperature under TNC. The proposed can be flexibly modified different targets applications energy sector.