A Fourier Decomposition Algorithm Applicable to Feature Extraction of Non-stationary Power Oscillating Signals

Abstract

The present invention provides a Fourier decomposition algorithm suitable for feature extraction of non-stationary power oscillation signals. The Fourier decomposition algorithm and the Hilbert transform method are used to process non-stationary power oscillation signals, which mainly include three parts, which are signal The preprocessing of the Fourier eigenband function, the screening of the Fourier eigenband function, and the extraction of characteristic parameters. The preprocessing of the signal adopts the FFT preprocessing method, and the screening of the Fourier eigenband function is to filter the non-stationary power oscillation signal pretreated by the FFT. grouping to get the Fourier eigenband function, and finally perform the Hilbert transform on the Fourier eigenband function to obtain the oscillation frequency and attenuation factor of the corresponding mode. According to the oscillation frequency and attenuation factor, the damping ratio parameter can be calculated, so that Corresponding suppression measures can be carried out according to the oscillation frequency, attenuation factor and damping ratio parameters, and provide help for online monitoring of low-frequency oscillations in power systems.

Description

technical field [0001] The invention relates to the technical field of power system analysis, in particular to a Fourier decomposition algorithm suitable for feature extraction of non-stationary power oscillation signals. Background technique [0002] With the gradual formation of my country's large regional power grid interconnection, the power grid is getting closer to the operating limit, and the low-frequency oscillation problem will affect the safety and stability of the system. Quickly and accurately identifying the parameters of the dominant oscillation mode in a noisy environment is the key to suppressing the low-frequency oscillation of the system. safe operation is of great importance. [0003] At present, the mode analysis methods based on measured data mainly include: fast Fourier transform (FFT), wavelet analysis, adaptive moving average (autoregressive moving average, ARMA) model method, ESPRIT algorithm, Prony method and empirical model. Empirical Mode Decompo...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the present invention proposes a Fourier decomposition algorithm suitable for feature extraction of non-stationary power oscillation signals, which can overcome the difficulties of traditional Fourier transform in processing non-stationary signals, Prony method is sensitive to noise, and wavelet base in wavelet analysis method It is difficult to select and the modal mixing problem that often occurs in the empirical mode decomposition method is applied to the power system to provide guarantee for the stable operation of the power system

Images