Artificial neural network based assessment of grid-connected photovoltaic thermal systems in heating dominated regions of Iran

Abstract

In this paper, an artificial neural network (ANN) is developed to assess hybrid photovoltaic thermal (PVT) systems for grid-connected (GC) electricity generation, space heating and domestic hot water providing in heating dominated regions of Iran. To do so, monthly and annual performance of a 5 kWp GCPVT system is simulated for a single-family house. The simulation results show that the GCPVT system is very promising whereas the annual yield factor varies from 1506 kWh/kWp to 1891 kWh/kWp. Also, an appropriate solar fractions for covering hot water are achieved in a range from 74.5% to 49.4%. A multilayered perceptron feed-forward neural network which is trained by Levenberg-Marquardt algorithm is used to predict AC electrical energy and solar thermal output of the GCPVT system. The developed ANN is based on global horizontal irradiance, ambient temperature, ambient relative humidity and wind speed as inputs. The proposed configuration of ANN presents a high accuracy in predicting output energy of the GCPVT system according to minimum mean square error and maximum correlation coefficient. Analysis of variance is performed to determine the significant control parameters influencing the output energy of the GCPVT system. © 2020 Elsevier Ltd