Prediction of Global Solar Radiation in India Using Artificial Neural Network

Abstract

Increasing global warming and decreasing fossil fuel reserves have necessitated the use of renewable energy resources like solar energy in India. To maximize returns on a solar farm, it has to be set up at a place with high solar radiation. The solar radiation values are available only for a small number of places and must be interpolated for the rest. This paper utilizes Artificial Neural Network (ANN) in interpolation, by obtaining a function with input as combinations of 7 geographical and meteorological parameters affecting radiation, and output as Global Solar Radiation (GSR). Data considered was of past 9 years for 13 Indian cities. Low values of error and high values of coefficient of determination thus obtained, verified that the results were accurate in terms of the original solar radiation data known. Thus, ANN can be used to interpolate the solar radiation for the places of interest depending on the availability of the data.