Multi-variable study/optimization of a novel geothermal-driven poly-generation system: Application of a soft-computing intelligent procedure and MOGWO

Abstract

Due to the operational conditions of flash-binary geothermal cycles, it is difficult to design a multiple heat recovery technique for its waste heat. To address this deficiency and reuse its waste heat in different stages, this study proposes a novel poly-generation model considering parallel and series waste heat recovery. The prime target is to maintain the temperature and enthalpy level of the waste heat from each stage to another and provide a unified framework, leading to producing the main required products. The model comprises a transcritical carbon dioxide Rankine cycle, a desalination subsystem, a single-effect absorption refrigerator, and a low-temperature electrolyzer. To this end, a multi-aspect feasibility study is conducted from energy, exergy, and economic viewpoints, and a multi-criteria optimization is applied in four different scenarios using an artificial neural network in tandem with a multi-objective grey wolf optimization. The most suitable state of operation of each optimized scenario and the efficient scenario is defined through the TOPSIS approach. It was found that the most influential parameter affecting the performance metric was the geothermal water inlet temperature. Besides, the second scenario, i.e., exergy efficiency/total investment cost rate, resulted in the best state of operation. Finally, the optimum objective functions were gauged at 29.59% and 72.72 $/h, respectively. In this situation, the optimum net output power, cooling load, freshwater production rate, and hydrogen production rate were equal to 1666.0 kW, 1029.0 kW, 146.4 m3/day, and 1967.0 m3/day, respectively. Besides, the sustainability index and levelized cost of products were obtained to be 2.11 and 0.0602 $/kWh.