Short-term wind power prediction method based on partial least squares regression

Abstract

The invention provides a short-term wind power prediction method based on partial least squares regression. The method comprises the following steps of: analyzing and extracting influencing features and influencing factors of a wind power prediction from a wind farm to form a historical wind power data vector and obtain a training sample set; performing reducing dimension on the training sample set; extracting component with the obtained training samples, carrying out a correlation analysis on input and output, calculating a regression coefficient with a least squares method to establish a partial least squares regression model; and selecting wind power point values as an input vector of the test sample, selecting a corresponding output vector as a true power output value and taking the input vector into the partial least squares regression model to obtain an ultra-short-term predictive value vector of the wind power. The partial least squares regression method is good in multi-regressive problem analyzing ability, and includes methods such as data regression model establishment, principal component analysis and typical correlation analysis, etc., which not only simplifies the prediction model structure, but also improves the prediction precision and enhances the generalization ability of the prediction method.

Description

technical field [0001] The invention relates to a wind power system, in particular to a short-term wind power prediction method. Background technique [0002] With the rapid development of society and economy, countries around the world have increasingly strong demand for energy, which threatens the depletion of traditional fossil energy. The development of renewable energy is imminent for human beings. At the same time, the extensive use of fossil energy has also brought challenges to the social environment. As an important part of renewable energy, wind energy has become one of the important directions of social development by fully exploiting existing wind energy resources and developing clean energy. However, when developing wind energy, there are also very prominent problems, because wind power itself has characteristics such as randomness, fluctuation and intermittent, and these characteristics will make it very difficult for the power grid to absorb wind energy. At ...

AI Technical Summary

Problems solved by technology

However, in actual operation, if there are too many selected features or influencing factors, the structure of the predicted model may be complicated and the training efficiency will be reduced. The prediction accuracy is very important

Images