Prediction of Monthly Global Solar Radiation Using Evolutionary Neural Network In Tehran city

Predicting monthly Global Solar Radiation (GSR) on a horizontal surface based on meteorological variables is main objective of this study. We apply Monthly mean of maximum air temperature, relative humidity, sunshine hours and wind speed values between 2005 and 2010 for Tehran city in Iran (35_47N, 51_37E). An evolutionary neural network (based on MLP and bee colony) along with different combinations of input variables were applied to consider the effect of each meteorological variable on monthly GSR prediction. The proposed models are especially useful for locations with no available measurement equipment. Statistical parameters such as RMSE, MBE and R2 show acceptable and reasonable accuracy of the developed models.

Keywords:

Global Solar Radiation (GSR) , Prediction , Evolutionary , Artificial Bee colony , Multi-layer perceptron (MLP)