PCA-based aircraft conduit vibration signal power frequency noise suppression method

Abstract

The invention discloses a PCA (Principal Component Analysis)-based aircraft duct vibration signal power frequency noise suppression method, which comprises the following steps of: performing PCA preprocessing on a vibration signal in a time window to obtain estimated power-frequency noise, and obtaining corresponding power-frequency noise parameters through the estimated power-frequency noise so as to obtain constructed power-frequency noise; and finally, subtracting the construction power frequency noise from the vibration signal to obtain a noise-pressed vibration signal. According to the PCA-based aircraft conduit vibration signal power frequency noise suppression method, suppression of power frequency noise in the conduit vibration signal can be realized. Compared with other power frequency noise suppression methods, the method is good in noise suppression effect, strong in signal protection capability during signal-to-noise frequency band aliasing and capable of effectively improving the signal-to-noise ratio of the vibration signal.

Description

technical field [0001] The invention relates to the technical field of aircraft duct vibration signal noise reduction, in particular to a PCA-based method for suppressing industrial frequency noise of aircraft duct vibration signals. Background technique [0002] The sensors arranged on the aircraft duct can obtain the vibration signal after hammering the duct, but the vibration signal collected by the aircraft duct in the actual environment has power frequency noise, which seriously affects the judgment of the stress state of the aircraft duct. There are few literatures on the suppression method of power frequency noise of aircraft duct vibration signal. It is concluded that the current methods of suppressing power frequency noise mainly include frequency domain method and time domain method. In the doctoral thesis "Research on Electrostatic Monitoring Technology and Life Prediction Method in Rolling Bearing Wear Area", the notch method is used to suppress power frequency n...

AI Technical Summary

Problems solved by technology

In the doctoral thesis "Research on Electrostatic Monitoring Technology and Life Prediction Method in Rolling Bearing Wear Area", the notch method is used to suppress power frequency noise, but this method is easy to damage useful signals when the signal-noise spectrum is aliased; the article "Based on Harmonic Wavelet and SVD Suppression of Power Frequency Interference of Tilting Pad Sliding Bearings" using harmonic wavelet to extract power frequency bands for singular value decomposition to suppress power frequency noise, but this method ignores the instability of power frequency; the article "Based on spectral interpolation and singular value difference spectrum The denoising method of rolling bearing static monitoring signal" uses the spectral interpolation method to suppress power frequency noise, but this method is easy to filter out some useful signals when interpolating in the 48-52Hz interval; the article "Elimination of Power Frequency Interference in Vibration Signals Based on Independent Component Analysis" The independent component analysis method is used to suppress power frequency noise, but the noise suppression effect of this method depends on the number of observed signal sources; the article "Power Frequency Interference Elimination Method Based on Singular Vector Spectrum" uses the spectrum of singular value vectors to suppress power frequency noise, but the method While filtering out noise, some useful signals will also be filtered out; CN106895985A discloses "Fault Signal Noise Reduction and Reconstruction Feature Recognition Method for High-speed Wire Rolling Mill" proposes a method based on Fourier transform search to characterize the singular value of power frequency noise to suppress power frequency noise, but this method is only suitable for strong power frequency noise environment; the article "cosine function adaptive method to identify and eliminate single-frequency interference" uses the function approximation method to eliminate power frequency noise, but this method ignores the instability of power frequency;

[0003] CN104570118A discloses "a method for automatic identification and removal of industrial interference based on two factors", which proposes to use the sin-cosine weighted approximation method to deal with industrial interference, but this method is applicable to the situation of power frequency and stable amplitude; the article "Based on PCA, the most Optimum-order parallel source power frequency noise suppression" uses the relationship between principal components and power frequency numbers to filter out power frequency noise, but this method also has the problem of damaging useful signals when signal-to-noise frequencies alias; it can be seen that none of the above methods can It is fully suitable for the suppression of power-frequency noise of aircraft duct vibration signals, and pointed out that the existing research on PCA suppressing power-frequency noise has the problem of useful signal loss when the signal-to-noise frequency bands are mixed

Images