Short-term wind power prediction method based on similar days and FRS-SVM

Abstract

The invention discloses a short-term wind power prediction method based on similar days and FRS-SVM, and belongs to the technical field of short-term wind power prediction methods. According to the method, a discrete Frechet distance is adopted as a similarity criterion, an overall similarity formula between a prediction day and a historical day is defined, and similar day samples are matched according to the overall similarity formula. Sample size of similar day samples is reduced, and similarity of modeling data is improved. Then, aiming at the characteristics of mutual correlation, nonlinearity, irrelevance and the like of original features, the complexity of a training model is increased, a fuzzy rough set (FRS) method is introduced to eliminate redundant features, and the input of anSVM model is optimized; the similar day method and the FRS method preprocess the data on the two levels of the sample number and the sample characteristics at the same time, the training time of the SVM model is saved, and the prediction precision is improved. The invention can effectively improve the deficiency of prediction precision at the prediction inflection point, and has a certain practical value.

Description

technical field [0001] The invention relates to a short-term wind power prediction method based on similar days and FRS-SVM (fuzzy rough set-support vector machine), belonging to the technical field of short-term wind power prediction methods. Background technique [0002] Energy is an important material basis for human social activities. With the rapid development of social economy, the demand and use of energy in human society have increased significantly, and the excessive exploitation and utilization of traditional fossil fuels have caused air pollution and ecological environment damage. At the same time, due to the limited nature of traditional energy sources, we are facing an energy consumption crisis. Therefore, the development and utilization of renewable energy has attracted more and more attention worldwide. [0003] Wind energy, as a clean renewable energy, is increasingly being valued by countries all over the world because of its easy development and low cost....

AI Technical Summary

Problems solved by technology

With the rapid development of my country's wind power industry, the collective data of wind farms is large and comes from a wide range of sources. However, due to problems such as wind abandonment and limited power operation, the data quality is uneven and the value density is low.

A large amount of wind farm observation data not only contains redundant information, but also affects the training efficiency of the prediction model

Images