Machine learning-aided optimal energy management of a Solar-to-X energy system based on hydrogen production/storage and CAES

Abstract

The present work introduces a highly integrated smart solar-based system for reliable and sustainable electricity and cooling productions without relying on environmentally and economically unsustainable battery storage. Hydrogen functions as an energy carrier in this system, storing surplus solar energy via the thermochemical vanadium chloride cycle and facilitating stable operation amid intermittent solar energy. The system is also integrated with compressed air energy storage and high- and low-grade thermal energy recovery subsystems to generate power and cooling via absorption cycles. Compressed air energy storage facilitates the storage of surplus solar energy in mechanical form, allowing for power generation during times of diminished solar availability. It recuperates the compression heat via intercoolers and an aftercooler, which is subsequently utilized to power the absorption power cycle and the single-effect absorption chiller, substantially enhancing exergy efficiency and guaranteeing no waste of valuable heat. The proposed smart integration's thermodynamic/economic/environmental indicators are comprehensively assessed to analyze the practicality. Then, optimal energy management/conversion is achieved through machine learning-aided multi-criteria optimization by applying a non-dominated sorting genetic algorithm. The main goal of this work is to optimize the thermo-economic and exergy performance of a novel solar-driven hybrid energy system integrating Compressed air energy storage and a vanadium-chlorine hydrogen production cycle using a machine learning-aided NSGA-II optimization strategy to minimize cost and maximize exergy efficiency. The results show that the system can store more energy from solar fields in the Compressed air energy storage and hydrogen storage systems, making it better able to harvest energy from fields. Heliostat and vanadium chloride cycles have the maximum exergy destruction because of the temperature difference and chemical reaction. Under the most optimal conditions, the system achieves a 65.8 % exergetic round trip efficiency and a unit product cost of $16.3 per GJ.