Wind power plant short-term wind power prediction modeling method based on wavelet analysis and multi-model AdaBoost deep network

Abstract

The invention discloses a wind power plant short-term wind power prediction modeling method based on wavelet analysis and a multi-model AdaBoost deep network. On the basis of analyzing the relationship between wind power and meteorological factors, wavelet multi-scale analysis and entropy and nonparametric estimation methods are firstly used for respectively inspecting time-frequency domain feature distribution, uncertainty and randomness of wind power data, and the wavelet multi-scale analysis, entropy and nonparametric estimation methods are used for reasonably dividing subsets so as to ensure that training samples fully excite all modes of a system. Secondly, nonlinear manifold learning is adopted to extract nonlinear features of the wind power data, and dimensionality reduction is achieved so as to reduce calculation complexity; and finally, a short-term wind power combined prediction model is created with high prediction precision, low calculation complexity and strong robustnessin combination with a long-term and short-term memory neural network with an optimized structure. Accurate and reliable wind power prediction can be provided for a wind power plant, and guarantee is provided for coordination control and power grid dispatching of large-scale wind power grid connection.

Description

Technical field [0001] The invention belongs to the technical field of wind power generation, and particularly relates to a short-term wind power prediction modeling method for wind farms based on wavelet analysis and a multi-model AdaBoost deep network. Background technique [0002] In recent years, the development of renewable energy has gradually become the consensus of the international community. Wind energy is a kind of renewable energy generated by the work of air flow. China has abundant wind energy resources. According to statistics from the Meteorological Department, the current wind energy that can be developed and utilized reaches more than 1 billion kilowatts. Accurate and reliable short-term wind power forecasting plays an important role in smart grid dispatch, which can reduce the economic loss caused by grid integration and reduce the risks of grid transmission and integration. Due to the instantaneity, randomness and uncertainty of wind speed distribution, short...

AI Technical Summary

Problems solved by technology

However, the currently widely used feedforward neural network cannot effectively acquire data with a time series structure, which makes it impossible to apply the content related to the previous sequence in practical applications.

Insufficient generalization ability of feedforward neural network in predictive modeling of wind power data

In addition, predictive modeling of feedforward networks as static neural networks cannot accurately describe the dynamic performance of the system

Images