Design and multi-objective optimization of a multi-generation system based on PEM electrolyzer, RO unit, absorption cooling system, and ORC utilizing machine learning approaches; a case study of Australia

Abstract

The geothermal systems' capability to produce various products persuades the design of a novel hybrid system, including power, cooling, freshwater, and hydrogen. Australia, with numerous geothermal sources and the tendency of hydrogen utilization as fuel, is considered a case study. The designed system is analyzed from energy, exergy, and economic viewpoints. Also, the artificial neural network is implemented to optimize the system's operation. Hence, the accuracy of four artificial neural networks in optimizing and predicting systems' performance is compared. Furthermore, four double-objective and four triple-objective optimization scenarios are considered to achieve the best optimum state from different viewpoints. The mean absolute Error value of 2.28×10−14 to predict the exergetic efficiency in the testing procedure is the best algorithm. The system provides 1263 kW net power with 39.89% exergy efficiency and 2.13 years of payback period at the base condition. The exergy efficiency-net present value scenario represents the best performance in which the net power, hydrogen, and freshwater production rates reach 3946 kW, 8.73 kg/h, and 152 kg/h, respectively. Subsequently, the exergy efficiency, payback period, and net present value are estimated at 46.27%, 1.84 years, and 19.52 M$.