Fractal analysis of daily solar radiation time series with the aim of typical meteorological day's classification

Awareness of the dominant patterns of radiation fluctuations in any region is important in terms of planning for the use of solar energy potential. In this study solar radiation time series’ signals has been investigated from a new perspective, based on the fractal geometery (self-similar). In this regard the fractal dimension of daily solar radiation time series was calculated with Minkowski-Bouligand method and with clearness index consideration used for classification the days of year in terms of radiation properties. Results showed that in the years of 2014, 2015 and 2016 the highest Probability of occurrence was belong to class 3 (overcast sky) with 55.2%, 52.07% and 44.8% respectively. The Days belonging to class 1 (clear sky) and class II (semi-cloudy sky) are ranked second and third in terms of probability of occurrence in all three statistical years.The results of this new method, that classes the days of the year based on solar radiation behavior, can considered as an essential step in photovoltaic cells installation and performance analysis in one region.