Short-term prediction of power fluctuations in photovoltaic systems using chaos theory

This paper proposes a novel short-term prediction method for power fluctuations in a photovoltaic (PV) system based on chaos theory. A data vector composed of n data points is embedded in an n-dimensional vector space by Takens' embedding theorem. The time delay is determined at 300 s by autocorrelation analysis, and the embedding dimension is determined to be 5 by the Grassberger-Procaccia algorithm. Next, the PV power in the near feature is predicted by using the local fuzzy reconstruction method. A simulation of the smoothing control of power in a PV system with an energy storage device consisting of electric double layer capacitors (EDLCs) is carried out. The largest reduction in the EDLCs' capacity is 3.81 kJ, which can be achieved on a clear day.