Application of rapid miner in ANN based prediction of solar radiation for assessment of solar energy resource potential of 76 sites in Northwestern India

In this study new technique Rapid Miner is used for relevant input variable selection for prediction of solar radiation using different ANN techniques. The prediction accuracy of ANN models developed with artificial neural network fitting tool (nftool), Radial Basis Function Neural Network (RBFNN) and Generalized Regression Neural Network (GRNN) are compared. The Rapid Miner shows that clearness index, extraterrestrial radiation, latitude and longitude are least relevant input parameters and maximum temperature, minimum temperature, altitude, sunshine hour are found to be the most relevant input parameters for solar radiation prediction. The ANN models developed with artificial neural network fitting tool (nftool) give better results than RBFNN and GRNN for solar radiation prediction. The mean absolute percentage error (MAPE) for ANN-1, ANN-2, ANN-3, ANN-4 and ANN-5 are found to be 16.91, 16.89, 16.38, 6.89 and 9.04. The ANN-5 model utilized most accessible input variables so it can be used to predict solar radiation for 41 locations of Gujarat and 35 locations of Rajasthan in Northwestern India. The yearly average solar radiation varies from 4.92 to 5.62 kWh/m2/day for Gujarat and it is varies from 4.66 to 5.54 kWh/m2/day for Rajasthan.