Optimizing energy efficiency and emission reduction: Leveraging the power of machine learning in an integrated compressed air energy storage-solid oxide fuel cell system

This research introduces a cutting-edge energy system that combines a solid oxide fuel cell (SOFC) with compressed air energy storage (CAES) to generate compressed air, electrical power, and heat. The system's performance was assessed and enhanced using regression-based machine learning models, concentrating on three main process variables: temperature, current density, and utilization factor. The machine learning models achieved impressive accuracy, with R-squared values greater than 98 %, demonstrating their effectiveness in predicting system performance. The results from multi-objective optimization indicated that the ideal conditions for maximizing energy storage, efficiency, and minimizing emissions include a temperature of 973 K, a current density of 6000 A/m2, and a utilization factor of 0.74. At these optimal parameters, the system reached an energy storage capacity of 28.12 cm3, an efficiency of 64.19 %, and emissions of 274.04 kg/MWh. These results underscore the potential of the integrated SOFC-CAES system to tackle significant energy and environmental issues by enhancing energy efficiency, lowering emissions, and offering a sustainable approach to power generation. The findings from this study contribute to the development of hybrid energy systems and facilitate the transition to more sustainable and resilient energy frameworks.