ANFIS Modelling and Prediction of Horizontal Global Solar Radiation

Modeling and prediction of energy sources is one of the most important methods to exploit different renewable energy sources particularly solar energy. Owing to dearth of accurate and reliable measured solar energy data especially horizontal global solar radiation, make this issue more important. In recent years, different soft computing approaches were employed to predict solar radiation. In this study, the ANFIS1 model was applied to predict the solar radiation for city of Mashhad. The main idea of this research is to catch the most efficient combination of input parameters to have the most accurate prediction. For this purpose long-time data for city of Mashhad was applied to predict solar radiation. Five combinations of Tmax, Tmin, Tavg and nday including (1) Tavg; (2) nday; (3) Tmin and Tmax; (4) Tmax and nday; (5) Tmax, Tave and nday are tried as inputs to develop ANFIS model. From the attained result, combination of Tmax, Tave and nday are the best ANFIS model for this region. For the testing data, the statistical indicators of RMSE, and R for the designed model are 2.310 (MJ/m2), 0193913 respectively. The results show the capability of the adequate precision for prediction of solar radiation.

کلیدواژه ها:

solar radiation prediction ، ANFIS model ، air temperatures ، artificial neural network