Power and freshwater are two energy-intensive products, which consume a huge amount of fossil fuels. It is important to supply the aforementioned products using renewable energy sources due to the depletion of fossil fuel resources and environmental issues. This paper investigates the exergy and exergy-economic analysis of water and power production using a small-scale combined cycle encompasses the concentrated photovoltaic thermal (CPVT) solar collectors, a Kalina cycle (KC), and a humidification-dehumidification (HDH) desalination unit. An exergo-economic parametric analysis was first investigated to determine the influence of some pertinent parameters on the exergy efficiency, and specific unit cost of the products. In the second stage, two intelligent forecasting approaches based on the artificial neural network (ANN) and improved particle swarm optimization (PSO) algorithms were utilized for predicting the performance metrics of the studied system. The system was supposed to work at half of the year's hour. The results demonstrated that the shares of CPVT, generator, humidifier, dehumidifier, and condenser in exergy destruction are 84 %, 6 %, 3 %, 2.5 %, and 2 %, respectively. Moreover, the exergy efficiency, and specific unit cost of the products, unit cost of electricity, and unit cost of the fresh water at the design condition were obtained as 23.23 %, 0.0806 $/kWh, 5.44 $/m3, and 31.15 $/GJ, respectively. Besides, the most effective parameter on the exergy efficiency and the specific unit cost of the products was the solar beam radiation, the increment in which from 300 W/m2 to 1000 W/m2 improved the exergy efficiency by 15.21 % and reduced the specific unit cost of the products by 63.16 %. In addition, the increase in the condenser pressure from 15 bar to 22 bar and the generator pinch point temperature difference from 5 °C to 15 °C reduced the exergy efficiency by 8.13 % and 4.03 %, respectively, leading to increasing the specific unit cost of the products by 1.30 % and 4.30 %. The results of modeling showed that hybrid ANN-IPSO models provide the most accurate prediction, highest tendency, and agreement to observation as compared to ANN in terms of (R2| exergy efficiency = 0.9903 and R2| specific unit cost of products = 0.9948) and (RMSE| exergy efficiency = 0.0010 and RMSE| specific unit cost of products = 0.9684).
Read full abstract