Artificial Neural Networks Model for Short Term Forecasting Global Irradiation at Center Station in the Nine Station Photovoltaic

This article will studies of forecasting global irradiation (GI) for the short term at one station photo voltaic system (PV) which the station is located at center point between the eight other photovoltaic stations. Short-term forecasting model which is known as Artificial Neural Networks (ANNs) is divided into two models, the first model without taking into account temperature and the second model based on data from meteorological temperature. In this article proposes forecasting which use a combination of exponential smoothing models as preprocessing of data and ANNs as a technique for predicting GI 60 minutes ahead. The simulation results with ANNs forecasting model will be compared with measured data. The result performance of our scheme show good results and have a satisfactory accuracy which can be obtained RMSE value 6.28% without consider of temperature and 5.18% consider data temperature, respectively.