Day-ahead electricity price probability prediction method based on dynamic network quantile model

Abstract

The invention discloses a day-ahead electricity price probability prediction method based on a dynamic network quantile model. Characteristic points of the method are as follows, the method comprisesthe following steps: designing a comprehensive influence factor according to an association relationship between different influence factors and an electricity price sequence to select an electricityprice similar day, designing a dynamic network quantile electricity price prediction model, and predicting a nuclear density electricity price probability based on the dynamic network quantile electricity price prediction model. The prediction precision is high, and the operation time can be greatly shortened; the model is less affected by price fluctuation related risks, and has strong toleranceto singular values. The method has the advantages of scientificity, reasonability, applicability, good effect and the like.

Description

technical field [0001] The invention relates to the field of smart grid and power data analysis, and is a day-ahead electricity price probability prediction method based on a dynamic network quantile model. Background technique [0002] Since the advent of the competitive power market, electricity price forecasting (EPF) has gradually become an indispensable and important link in the process of making investment decision-making mechanisms for electricity sales companies. Among them, the day-ahead electricity price, as a reference price for energy trading decisions, has attracted much attention because it can directly affect the revenue of electricity sales companies. With the rapid development of smart grid, a large number of distributed renewable energy connected to the grid has become an important part of a strong smart grid. However, the intermittency and uncertainty of renewable energy have a major impact on the price-oriented electricity market, which will greatly redu...

AI Technical Summary

Problems solved by technology

[0003] The existing day-ahead electricity price forecasting method mainly obtains the predicted daily electricity price by inputting historical daily electricity price data into models such as SVM and BP neural network. Forecasting ignores the daily periodicity of electricity prices. If data inconsistent with the predicted daily electricity price characteristics is input into the forecasting model, it will not only increase the amount of calculation, but also lead to a decrease in the accuracy of the day-ahead electricity price prediction; (2) due to the Most of the research is not on probability forecasting, but on point forecasting, which has certain limitations. Probability forecasting can reflect the fluctuation range of electricity price changes in detail, and provide decision makers with richer information, which is highly regarded by academic circles and operators. (3) Among the existing probabilistic forecasting models, the quantile regression neural network probabilistic forecasting model (QFNN) is widely used in load and wind power probabilities because it does not require prior distribution assumptions and can provide stable forecasting information. During prediction, however, FNN needs to preset the number of network nodes, and setting too many or too few network nodes will reduce the prediction accuracy of the model

So far, there are no literature reports and practical applications of the day-ahead electricity price probability prediction method based on the dynamic network quantile model

Images