A comparative performance analysis, working fluid selection, and machine learning optimization of ORC systems driven by geothermal energy

Abstract

Geothermal-energy-based desalination plants that utilize reverse osmosis (RO) are evaluated in detail with the aim of converting the thermal energy of geothermal flow into electricity based on three configurations of the organic Rankine cycle (ORC). These systems include basic ORC, parallel ORCs, and series ORCs. The performance of the proposed systems is evaluated by developing thermodynamic and economic models. A parametric study is conducted to assess the impact of main design parameters on the performance. Moreover, a multi-objective optimization procedure is implemented, by coupling the genetic algorithm (GA) with an artificial neural network (ANN), to identify optimal cycle configuration and working fluid(s) under three different heat source temperatures. For a specific geothermal temperature (135 °C), the highest exergy efficiency in the basic scheme is obtained by R1233zd (E), while ammonia yields maximum power generation. For parallel configuration, irrespective of the heat source temperature, the combination of R1233zd (E)-Ethane always leads to the highest exergy efficiency, while comparison results at 135 °C indicate that the combination favored by power generation is R1234ze (Z)-Ethane. For the series configuration, optimization results under the heat source temperature of 135 °C indicate that the ammonia-R1234ze (Z) combination is a suitable candidate under the economic objective, while ammonia-ammonia is recommended as the preferred choice thermodynamically. In addition, comparing the optimization results of the parallel configuration with the basic design reveal considerable improvements of about 150% and 60% in power generation and water production, respectively. The same comparison between series and basic configurations revealed an improvement of about 30% in power generation.