Using an integrated artificial neural networks model for predicting global radiation: The case study of Iran

Abstract

In this article we have used artificial neural networks for predicting solar global radiation by using climatological variables in the locations that no direct measurement equipment is available. Numbers of climatological and meteorological parameters were considered for this purpose and monthly data provided for 6 years (1995-2000) in 6 nominal cities in Iran. Separate model for each city is considered and the quantity of solar global radiation in each city is calculated. The average value of acquired results of these models has shown high accuracy about 94 % and the mean absolute percentage error (MAPE) of the models was 6.7 %. The results of these models have proved the capability of ANN models for predicting solar radiation in the locations that there is no any measurement equipment. The results of this study have shown a better accuracy than other conventional prediction models that have been used up to now.