Deep learning optimization and techno-environmental analysis of a solar-driven multigeneration system for producing sustainable hydrogen and electricity: A case study of San Francisco

In general, solar-driven systems have substantial heat loss; so that brought designers to consider organic Rankine cycle for heat recovery. However, it still wastes amount of energy in the organic Rankine cycle's condensers. In this context, the present study proposes a novel multigeneration system with the aim of hydrogen and cooling load production, as well as power generation using heat recovery units by considering environmental impacts. The present solar-driven system comprises two power cycles, thermoelectric generator, hydrogen production unit, and absorption chiller subsystem. Performance assessment of the system in terms of energy, exergy, economic, and environmental was executed using a developed programming code, based on a mathematical model. The model's functions were parametrically investigated by considering the most effective parameters. The time-consuming optimization process caused a machine learning optimization procedure to be employed. After optimization, the exergy efficiency, total cost rate, and hydrogen production rate of 13.05%, 101.4 $/h, and 5.15 kg/h were obtained, respectively. For optimized case, the environmental analysis showed that the amount of CO₂ emission reduction rate was resulted in 203.2 kg/h, which is 70 kg/h more than the result of base case. San Francisco, as a case study, was investigated to be considered for the potential of implementing the proposed system. As a result, the hydrogen production rate of 5.94 kg/h and net power generation of 1001 kW were acquired at peak.