Daily global solar radiation estimation in desert areas using daily extreme temperatures and extraterrestrial radiation

Abstract

The estimation of daily solar radiation is needed in many studies related to solar power plant placements. To optimize photovoltaic (PV) systems, their placement must be as efficient as possible in terms of the prevailing meteorological conditions. There are situations where radiation data are not available, as in the case of desert areas, suitable for the operation of PV systems. In this work, daily global solar radiation has been estimated in desert areas using Artificial Neural Networks (ANN), where the inputs used are daily minimum and maximum temperatures and extraterrestrial radiation. The ANN model is validated with data from deserts in Chile, Israel, Saudi Arabia, South Africa and Australia. The results show that the average Relative Root-Mean-Square Deviation (RRMSD) value is 13%, the average Relative Mean Bias Difference (RMBE) value is less than 4% and the average correlation coefficient (r) value is about 0.8.