Prediction of global solar radiation using nonlinear auto regressive network with exogenous inputs (narx)

Solar energy is regarded as the most vital non conventional energy sources among all fossil fuels. For the purpose of modeling of solar energy conversion device solar radiation data is most essential. Hence, accurate measurement or prediction of solar radiation data is important for a particular location. The proposed system uses a Nonlinear Autoregressive Network with Exogenous Input (NARX) model to predict solar radiation. NARX approach is used to estimate daily global solar radiation by using meteorological parameters such as sunshine duration, temperature, humidity. Solar data in Bhubaneswar, India have been used in the present case. The data for the period of 2002-2005 are used for training the NARX network while the data for the year 2006 is used for testing. The results of NARX network are evaluated on the basis of mean square error and regression coefficient.