Optimization of solid oxide electrolysis cells using concentrated solar-thermal energy storage: A hybrid deep learning approach

The Solid Oxide Electrolysis Cell (SOEC) represents a cutting-edge solution for the conversion of CO2 and H2O into syngas, offering significant economic and environmental benefits. However, the process requires substantial high-temperature heat inputs, traditionally supplied by electricity. This study introduces a novel approach leveraging concentrated solar radiation as a renewable heat source for SOEC, addressing the challenge of its inherent fluctuations through the integration of Thermal Energy Storage (TES) systems. We propose a hybrid model that combines multi-physics simulation with a deep learning algorithm, enabling rapid optimization of the electrolysis process under real-time direct normal irradiance conditions. Our findings demonstrate that the inclusion of TES within the system architecture results in a remarkable 53 % reduction in temperature variation rate at the SOEC inlet, ensuring operational stability and efficiency. Furthermore, by fine-tuning capacity parameters, we have developed a control strategy that harmonizes efficiency with safety performance. The robustness of our system is underscored by its resilience to step changes, achieving a 75 % reduction in temperature fluctuations. This research contributes a pioneering method for the real-time optimization and control of SOEC systems, harnessing the power of TES to drive sustainable energy conversion with enhanced reliability and economic viability, facilitating precise and swift predictive capabilities even under dynamic operating conditions.