Multivariate regression for prediction of solar irradiance

This paper describes regression models to forecast solar irradiance for a short term (or period). The regression models enable the prediction of solar irradiance in minute values over a period of a few days. A single variate regression model is used and various plots obtained between solar irradiance as dependent variable and air temperature and relative humidity as independent variables have been studied. Optimal range for prediction using regression is decided. To obtain accuracy multivariate regression is carried out It also presents new multifunctional relationship between solar irradiance, air temperature and relative humidity. This multifunctional regression relationship gives more accurate results compared to other methods having single variable. In this regression model solar irradiance follows an increasing trend upto a particular temperature after which it shows decreasing trend and hence it has been modeled with three equations.