One month-ahead forecasting of mean daily global solar radiation using time series models

Abstract

Accurate forecasting’s of solar radiation are crucial to quantifying future solar power production or integration. The integration of solar energy into the electricity grid has become predominant among researchers and scientists due to the current demand for energy as well as the depletion of fossil fuel reserves and environmental impacts. This paper aims to introduce a method for multi month-ahead forecasting (one, two and three month ahead) of monthly mean daily global solar radiation time series and forecasting of a large scale solar radiation data driven. ARMA and ARIMA models are used to predict the coming value of the global solar radiation time series. Both models are applied to stationary and non-stationary time series of the solar irradiation data. Taken into account the advantage of ARMA and ARIMA models, the two models are compared in terms of goodness-of-fit value determining by the AIC and BIC criterion. As a result, the ARMA (2, 1) and ARIMA (0, 2, 1) are selected as optimal models due to the minimize values of AIC and BIC criterion. Based on the statistical error value, ARIMA (0, 2, 1) models give good accuracy in three period ahead forecasting consecutive in comparison with ARMA (2, 1). The lower value of the optimum model indicates that ARIMA (0, 2,1) is more suitable to forecast monthly mean daily global solar radiation for Tetouan city and may be for other locations of similar weather conditions.