Artificial Neural Network Based Prediction of Daily Global Solar Radiation for optimum sizing of solar photovoltaic system

Abstract:

Modeling of solar energy system requires daily value of global solar radiation (DGSR). Due to expensive measuring instrument DGSR is not available for most of the sites. Therefore, the main objective of present study is to predict DGSR on a horizontal surface, based on meteorological variables, using Artificial Neural Network Fitting Tool. Extraterrestrial radiation, clearness index, ambient temperature, day number, sunshine hours and wind speed data between 2009 to 2012 for Hamirpur city in Himachal Pradesh, India are used to predict DGSR. In order to consider the effect of each meteorological variable on daily GSR prediction, six following combinations of input variables are considered:

1. Extraterrestrial radiation as inputs and daily GSR as output.
2. Clearness index as input and daily GSR as output.
3. Ambient temperature as input and daily GSR as output.
4. Day Number as input and daily GSR as output.
5. Sunshine hours as input and daily GSR as output.
6. Wind speed as input and daily GSR as output.
7. Extraterrestrial radiation, clearness index, ambient temperature, day number, sunshine hours and wind speed as input and daily GSR as output.