Power system low-frequency oscillation mode identification method based on noise-like signal VMD decomposition

Abstract

The invention provides a power system low-frequency oscillation mode identification method based on noise-like signal VMD decomposition, and relates to the technical field of power system low-frequency oscillation. The method comprises firstly acquiring noise-like data existing in the operation of a power system, performing fast Fourier transform (FFT) on the noise-like data to obtain a spectrogram of a noise-like signal, and determining the number of intrinsic modes of the VMD decomposition according to the number of peaks in the spectrogram; secondly decomposing the noise-like signal by using a VMD method to obtain the frequency center and the bandwidth of each intrinsic mode to separate the frequencies of the noise-like signal; finally fitting the respective intrinsic modes (IMF) by using a mode identification method to obtain a potential oscillation mode parameter in the noise-like signal and identify the low-frequency oscillation mode of the power system. The method can directly process the noise-like data in the normal operation of the power system, decomposes the potential low-frequency oscillation mode component, reduces the influence of the noise signal on the mode identification, and achieves a purpose of pre-warning the low-frequency oscillation by identifying the decomposed IMF.

Description

technical field [0001] The invention relates to the technical field of low-frequency oscillations in electric power systems, in particular to a mode identification method for low-frequency oscillations in electric power systems based on VMD decomposition of noise-like signals. Background technique [0002] With the continuous expansion of power system scale, the problem of low-frequency oscillation has become increasingly prominent. The emergence of Wide Area Measurement System (WAMS) provides new means for the monitoring, analysis and control of large-scale interconnected power grid systems, and provides a basis for the online identification of low-frequency oscillations. [0003] In the actual power system, the occurrence probability of obvious disturbance is relatively small, and it cannot be manipulated by humans. The amount of data is limited, and the identification results are difficult to reflect the current operating characteristics of the power grid in a timely and ...

AI Technical Summary

Problems solved by technology

[0004] However, traditional identification methods are not suitable for directly analyzing noise-like data

At present, the commonly used low-frequency oscillation noise reduction methods include empirical mode decomposition (EMD), wavelet transform, etc. These methods have slow operation speed and high misjudgment rate, and it is difficult to achieve effective noise reduction filtering effect.

Images