Multi-location model for the estimation of the horizontal daily diffuse fraction of solar radiation in Europe

Hourly and daily solar radiation data are crucial for the design of energy systems based on the solar source. Global irradiance, measured on the horizontal plane, is, generally, available from weather station databases. The direct and diffuse fractions are measured rarely and should be analytically calculated for many geographical locations. Aim of this paper is to present a multi-location model to estimate the expected profiles of the horizontal daily diffuse component of solar radiation. It focuses on the European (EU) geographical area joining data from 44 weather stations located in 11 countries. Data are collected by the World Radiation Data Centre (WRDC) between 2004 and 2007. Different analytic functions, corre- lating the daily diffuse fraction of solar radiation to the clearness index, are calculated and compared to outline the analytic expressions of the best fitting curves. The effect of seasonality on solar irradiance is considered developing summer and winter scenarios together with annual models. Similarities among the trends for the 4 years are, further, discussed. The most adopted statistical indices are used as key per- formance factors. Finally, data from three locations not included in the dataset considered for model development allow to test the proposed approach against an independent dataset. Obtained results show the effectiveness of adopting a multi-location approach to estimate solar radiation components on the horizontal surface instead of developing several single location models. This is due to the increase of the geographical range of applicability without a significant decrease of the accuracy. The percentage gap between single and multi-location model performances is lower than 20% for the most of the statis- tical indices adopted for both the model development and ex-post validation through the three indepen- dent datasets.