Classification method of voltage disturbance signal based on LMD and machine learning classification and system

Abstract

The invention discloses a classification method of voltage disturbance signal based on LMD and machine learning classification and system. The method comprises the following steps of: acquiring the original voltage disturbance signal, and performing LMD decomposition on the original voltage disturbance signal, to acquire an LMD including the original signal frequency and the physical value of theamplitude, and perform a 3-layer decomposition product function component; using the signal energy value constructed by the decomposition product function component as a input of the neural network; finally, closely judging the recognition result ,to realize the recognition and classification of the voltage disturbance signal, after the neural network training is identified. According to the classification method of voltage disturbance signal based on LMD and machine learning classification and system, the identification and classification of voltage disturbance signals can be realized; it issuitable for the processing of nonlinear signals and non-stationary signals, and is less affected by noise, is not susceptible to over-envelope and under-envelope phenomena, and has the advantages ofhigh signal processing efficiency, high recognition accuracy and good recognition work efficiency.

Description

technical field [0001] The invention relates to a voltage disturbance signal detection technology for power system signal processing, in particular to a voltage disturbance signal classification method and system based on LMD and machine learning classification. Background technique [0002] During the operation of the power system, there are a large number of nonlinear and irregular power signals. At the same time, under the influence of various faults, switch closing operations, lightning strikes, etc., a large number of voltage disturbance signals will be generated. The existence of these signals will distort the amplitude and frequency of the power system voltage, affect the normal operation of power equipment, and then affect The safety of the power system. [0003] At present, there are many kinds of methods for voltage disturbance signal processing, but there are certain defects in the processing of nonlinear signals and non-stationary signals. Fourier transform is ...

AI Technical Summary

Problems solved by technology

Fourier transform is a traditional signal extraction and processing method, but it is not suitable for dealing with nonlinear and non-stationary signals, and there are spectrum leakage and fence phenomena; wavelet and wavelet packet transform are not real adaptive transforms, and must be constructed Decompose the signal on the basis of strict standard functions, otherwise the best decomposition effect cannot be achieved; S transform is the combination and extension of the idea of ​​windowed Fourier transform and continuous wavelet transform. The transform window function is a variable that scales with frequency Gaussian function, so it is greatly affected by noise; Hilbert-Huang Transform (HHT) and Empirical Mode Decomposition (EMD) methods are easily affected by over-envelope and under-envelope phenomena. Influence, there is serious endpoint pollution, which makes the detection of unexplainable values

Images