Envelope demodulation frequency band determination method based on harmonic-to-noise ratio

Abstract

Provided is an envelope demodulation frequency band determination method based on harmonic-to-noise ratio. The method comprises: adhering a vibration acceleration sensor on the bearing pedestal of a tested antifriction bearing or installing a sound level meter on the position close to the upper part of the bearing pedestal, so as to acquire signals; establishing a tree filter bank to perform band-pass filtering on original signals; through Hilbert transformation, obtaining envelope of signals, using logarithm spectrum estimation to estimate envelope signal harmonic-to-noise ratio, and using the harmonic-to-noise ratio value as an index of whether frequency band fault features are significant; and finally, using the frequency band whose harmonic-to-noise ratio is the maximum as an optimal envelope demodulation frequency band, and analyzing the optimal envelope demodulation frequency band to determine occurrence and the type of the fault. The method makes full use of periodicity of impact caused by the fault, and uses the envelope harmonic-to-noise ratio as the index to measure the fault, thereby overcoming defects of a conventional kurtosis index. The method has robustness, and can adaptively extract the optimal filtering frequency band, and is beneficial for realizing antifriction bearing fault feature adaptive extraction and automation of diagnosis and monitoring.

Description

technical field [0001] The invention relates to the technical field of rolling bearing fault diagnosis, in particular to a method for determining an envelope demodulation frequency band based on a harmonic-to-noise ratio. Background technique [0002] Rolling bearing is one of the most widely used general-purpose mechanical components. Its health directly affects the operation quality of the whole machine. It is of great significance to monitor the health and diagnose faults of rolling bearings. Envelope demodulation analysis is an effective method to extract the periodic impact caused by rolling bearings. Through bandpass filtering and Hilbert envelope demodulation of bearing vibration or radiation noise, according to the fault characteristic frequency of the bearing in the envelope spectrum You can judge whether the bearing is faulty and the type of fault. [0003] The frequency band selection of the band-pass filter is the key to the whole envelope demodulation method. ...

AI Technical Summary

Problems solved by technology

[0004] The patent application number is 201410140890.6, in order to solve the problem of inaccurate selection of the envelope demodulation frequency band caused by high-peak pulse interference, it proposes to optimize the selection of the filtering frequency band by sub-band spectral kurtosis averaging, but the method has not overcome The disadvantage that the kurtosis value will decrease as the impact intensity increases

[0005] Patent application number 201410577135.4 sets the periodic threshold of the signal envelope after bandpass filtering on the basis of spectral kurtosis, and sets the kurtosis value of the frequency band below the threshold to zero to suppress the interference of aperiodic impact, but the At the end of the method, the kurtosis value is still used as a diagnostic index, and it does not overcome the disadvantage that the kurtosis value will decrease with the increase of the impact intensity.

Images