Absolute anti-aliasing multi-scale filtering method for rotary mechanical structure complex harmonic dynamic process

Abstract

The invention provides an absolute anti-aliasing multi-scale filtering method for a rotary mechanical structure complex harmonic dynamic process and relates to a mechanical fault diagnosis method. A non-contact type eddy current sensor is installed on a bearing seat of rotary mechanical equipment running at a constant rotating speed for performing de-averaging processing on the collected vibrationdisplacement signals; the frequency, amplitude and phase information of each high-energy harmonic component in the signal are extracted through a redundant Fourier transform feature; the frequency response of a lifting wavelet packet filter with the decomposition depth being J is obtained through a discrete unit pulse response sequence; weighted filtering is carried out on the extracted harmoniccomponents through a wavelet packet frequency spectrum function; the time domain waveform of a wavelet subspace sequence is observed, Hilbert envelope analysis is carried out, and the loss amplitude and the instantaneous frequency are calculated so as to determine the fault feature.

Description

technical field [0001] The present invention relates to a mechanical fault diagnosis method, in particular to an absolute anti-aliasing multi-scale filtering method for a fault feature extraction method based on time-domain and frequency-domain information to perform wavelet packet filtering on signals, and an absolute anti-aliasing multi-scale filtering method for the complex harmonic dynamic process of rotating mechanical structures . Background technique [0002] Vibration displacement signal acquisition is an important data source for mechanical equipment condition monitoring and fault diagnosis, which is realized by eddy current sensors. Since the eddy current sensor is mainly sensitive to low frequency components, the spectrum of the signal is composed of multiple harmonic components with different amplitudes and measurement noise. Since rotating mechanical equipment often operates under complex and severe working conditions, the early failure of mechanical parts is v...

AI Technical Summary

Problems solved by technology

Although the analysis of vibration displacement signals by wavelet transform has achieved fruitful applications, neither convolution in the time domain nor frequency domain can overcome the distortion of the decomposition results caused by the energy leakage of strong harmonic components.

Images