Bayesian optimisation for enhanced energy and exergy efficiency in solar multi-generation plants

This paper examines a novel concentrated solar multi-generation plant producing electricity, heating, cooling, drying, and hot water. The system employs molten salts (60% NaNO3, 40% KNO3) for thermal storage, enabling continuous operation through a solar tower, Rankine cycle, organic Rankine cycle, and air-based cycle. Exergy analysis revealed the solar tower and ORC evaporator contribute 47% of system losses. Using Bayesian multi-objective optimisation with Gaussian process models and expected hypervolume improvement achieved significant enhancements: 10.28% in energy efficiency, 7.50% in exergy efficiency, and 8.85% in net power output.