An intelligent thermodynamic/economic approach based on artificial neural network combined with MOGWO algorithm to study a novel polygeneration scheme using a modified dual-flash geothermal cycle

Abstract

Flash-based geothermal cycles correspond to environmentally friendly and cost-effective processes in a renewable framework and provide an opportunity for combined cycles. However, these cycles are characterized by significant energy losses and their waste stream’s low/medium operational temperature is the principal defect for managing multiple generation arrangements without assisting other energy resources. Hence, the main aim of this study is to propose a novel polygeneration scheme, integrated with a dual-flash geothermal cycle equipped with self-superheaters, able to mitigate the discussed defect. A new coupled series and parallel design of energy recovery is established, allowing to increase the compatibility of combined cycles and enable a larger production. This design encompasses a single-effect refrigeration cycle, a modified transcritical CO2 cycle, a polymer electrolyte membrane electrolyzer, and a thermal desalination cycle. The proposed process is examined from thermodynamic, sustainability, and economic (exergoeconomic and net present value analyses) points of view. Besides, a detailed sensitivity study is conducted by which the trend of performance variables in response to the increasing five main decision parameters is viewed. Afterward, an intelligent approach relying on an artificial neural network is built to learn and validate the behavior of defined objective functions (exergetic efficiency and products’ levelized cost). Moreover, a multi-objective grey wolf optimization (MOGWO) procedure endeavors to optimize the operation of the system. According to the results of this study, flash tank 2′s inlet pressure is the effective parameter, and its mean sensitivity index equals 0.289. Besides, the aforementioned objectives are gauged at 37.45% and 0.0625 $/kWh, respectively.