Ann-based optimization of parabolic trough type solar collector power plant: Bilecik case study

In this study, the design of the Parabolic Trough Type Solar Collector Thermal Power Plant (PTSCTPP) system has been made by considering the conditions of Bilecik province. In the system, parabolic trough type collectors are integrated with ORC cycle. It is also supported by a thermal energy storage system that is stored for use in situations where sunlight is absent or insufficient. A total of 4500 designs have been realized according to different operating parameters determined for five different refrigerants (R-134a, R141b, R600a, R236ea ve R1234ze), these designs were made first and second law analyzes and optimum system design. In the next part of the study, cost analysis was performed for each design. The useful life of the PTSCTPP system has been determined as 20 years and the economic efficiency of the system has been examined by Net Present Value (NPV) method. Finally, all designed systems are modeled using Artificial Neural Network (ANN). ANN model of PTSCTPP system was prepared by using MATLAB program. 4500 parameters have been designed to estimate the energy efficiency, exergy efficiency, net power output and NBD values of the PTSCTPP system. 3250 of this were seperated for training and the remaining 1250 for testing the network. The network architecture was trained for 1000 iterations. As a result, it was found that the values calculated by ANN model and analytically calculated values deviate to an acceptable extent. The most appropriate network architecture was used in the optimization of the designed system. As a result of optimization using ANN model, NBD value of optimum design was determined as 61.018.548.133.27 TL.