A Short-Term Load Forecasting Method Based on Composite Asymmetric Stochastic Fluctuation Model

Abstract

The invention discloses a short-term load forecasting method based on a composite asymmetric random fluctuation model, step 1: load time series F t Decomposition; Step 2: load forecasting method based on asymmetric SV model; Step 3: forecasting and forecasting effect evaluation. A short-term load forecasting method based on a composite asymmetric random fluctuation model provided by the present invention improves the fluctuation equation of the random fluctuation model, and relaxes the ε by adding an asymmetric item or an indicative function t , η t The uncorrelated assumption takes into account the impact of the asymmetric effect of fluctuations from two dimensions in the model, and at the same time considers the smooth transition of the asymmetric transition stage. In this way, the information contained in the second moment of the load time series can be discovered, and the accuracy of load forecasting can be improved.

Description

technical field [0001] The invention relates to a short-term load forecasting method based on a composite asymmetric random fluctuation model, which belongs to the technical field of electric load forecasting. Background technique [0002] At present, short-term load forecasting is of great significance to the economic dispatch operation of power grid. With the advancement of my country's power system construction, a large number of new energy sources on the power supply side are continuously connected, a large number of new equipment on the grid side, especially large power electronic equipment, are continuously put into operation, and user behavior on the demand side presents more complex patterns and characteristics. These factors power load Time series exhibit more complex volatility characteristics. [0003] The Stochastic Volatility (SV) model is an effective load forecasting model, and the classic SV model requires ε t Represents a random perturbation of the measure ...

AI Technical Summary

Problems solved by technology

Although the classical stochastic fluctuation model can describe the volatility of the load time series, it cannot reflect the fluctuation asymmetry of the load time series due to the symmetry of its equation structure

[0004] Some studies have shown that there is an asymmetry in the fluctuation of power load. The positive shock and negative shock of the same unit have different effects on the future power load fluctuation. Information to improve prediction accuracy

Asymmetric SV model by adjusting ε t , η t The correlation matrix of the load time series can describe the volatility of the load time series, but in some cases the asymmetric characteristics of the load time series are complex, and the model needs to be further promoted and improved

Images