Temporal global solar radiation forecasting using artificial neural network in Tunisian climate

The temporal prediction of the solar radiation is very important for the operation of any solar energy system technology and completing data set. Based on meteorological parameters, the artificial neural network (ANN) can bring a technical solution for the prediction problems. In this paper, we developed an ANN for the south Tunisian climate to predict global solar radiation. Five years of record from January 2008 to December 2012 and five zones were selected to train and test the neural networks. The root mean square error (rmse), the coefficient of correlation (r) and the mean absolute error (Mae) are used to evaluate our model. The results show that the errors for the temporal forecasting varies between: 0.952 ≤ r ≤ 0.988, 4.9% ≤ rmse ≤ 10.1% and 1847 ≤ mae (j/cm2) ≤ 4657. Keywords— temporal forecasting, solar energy, artificial neural network, ground station meteorological parameter, south Tunisia