Performance Evaluation of Developed Empirical Models for Predicting Global Solar Radiation in Western Region of India

For the prediction of solar radiation, various models have been developed in literature with sunshine hours as prime parameter. However, an extensive amount of work is still required correlating the sunshine hours and other parameters. Seven models each for predicting mean global radiation and mean diffuse radiation has been derived in this research work for city Amravati (20.9374 degrees N, 77.7796 degrees E), correlating with sunshine hours, average temperature and relative humidity. Relative humidity and temperature together with sunshine can prove to be a best alternative for predicting solar radiation at places where direct measurement is unavailable. Both linear and non-linear correlations developed are compared with conventional models and measured values available for last five years from the meteorological data. The validation is done using several statistical parameters viz. mean absolute percentage error (MAPE), mean bias error (MBE), root mean square error (RMSE) and R-2. The performance evaluation of these models gives the best suitable model for both global and diffuse solar radiation for the city Amravati, Maharashtra.