Ultra-short-term wind speed prediction method based on spectral clustering and genetic optimization extreme learning machine

Abstract

The invention relates to an ultra-short-term wind speed prediction method based on spectral clustering and genetic optimization extreme learning machine, comprising: S1: preparing data; S2 preprocessing the prepared data; S3: performing wavelet transformation on the preprocessed data; S4 : Normalize the wavelet transformed data; S5: Select the normalized data through correlation analysis to determine the input variables; S6: Reduce the dimensionality of the input variables generated by S5 through principal component analysis ; S7: cluster analysis on the data after dimensionality reduction in S6 by spectral clustering method, and form an extreme learning machine sample space with the normalized data in S4; S8: use extreme learning machine and genetic algorithm on S7 Data hierarchical prediction of the formed extreme learning machine sample space; S9: Add the hierarchical prediction values ​​to obtain the ultra-short-term wind speed prediction value. The invention realizes ultra-short-term and multi-step prediction of wind speed, improves prediction accuracy, greatly reduces calculation amount, and improves prediction efficiency.

Description

technical field [0001] The invention relates to the technical field of wind speed prediction, in particular to an ultra-short-term wind speed prediction method based on spectral clustering and genetic optimization extreme learning machine. Background technique [0002] Wind power is a kind of clean renewable energy, which is relatively simple to develop and utilize, and has been paid more and more attention by countries all over the world. Effective wind speed prediction is the basic link of wind power generation research, and is the necessary prerequisite and guarantee for the establishment and operation of wind power forecasting and forecasting systems for grid-connected wind farms. High-precision ultra-short-term wind speed prediction for wind farms can effectively reduce grid voltage and frequency fluctuations caused by sudden cut-out of wind turbines, thereby reducing large fluctuations in wind power output and ensuring safe and reliable operation of the power system. ...

AI Technical Summary

Problems solved by technology

However, the traditional neural network learning algorithm needs to manually set the network training parameters, and the gradient descent algorithm is often used to adjust the weight parameters. The learning speed is slow, the generalization performance is poor, and it is easy to generate local optimal solutions, which limits its further application.

Images