Statistical analysis of self-affinity of energy dispatch in power generation in the Northeast region of Brazil

Abstract

In this article, we study the fractal dynamics in time series of energy generation in the Northeast region of Brazil, considering a database used by the National Electric System Operator referring to semi-hourly information on the generation of various energy sources. Applying the DFA technique between the years 2020 and 2022, it was found that the dispatch of solar energy fluctuated in all years from superdiffusive behavior in the short term to persistent in the long term, whereas wind generation in all years varied from superdiffusive short-term to long-term subdiffusive. In all observation periods, there was a repeatability of the dispatch behavior of wind and solar sources, which can be explained by the fact that these sources always maintain a standard of availability for the energy matrix. When analyzing the behavior of solar generation, a greater predictability of energy availability was found, which can therefore be considered as a more viable source to make up for the lack of hydraulic energy. This analysis is based on historical time series numbers, not considering the costs of each energy source, environmental license situations, among other factors.