A rotary machine fault feature extraction method based on time-frequency spectrum correlation analysis

Abstract

The rotary machine fault feature extraction method based on time-frequency spectrum correlation analysis comprises the following steps: firstly, collecting a group of vibration signals on rotary machine equipment, and carrying out time-frequency transformation on the vibration signals to obtain a time-frequency spectrum; secondly, selecting an impact feature in the time-frequency spectrum, enabling the impact feature to perform point-by-point translation along a time axis from the initial time of the time-frequency spectrum, continuously masking a time-frequency block in the time-frequency spectrum, calculating a correlation coefficient between the impact feature and the masked time-frequency block, and obtaining a correlation coefficient sequence after translation is finished; finally, performing Fourier transform on the correlation coefficient sequence, and extracting the fault characteristic frequency from the frequency spectrum. The obtained correlation coefficient sequence curvecan visually display the change rule of the impact characteristic in the vibration signal of the rotary mechanical equipment, and the influence of noise and other interference signals is greatly removed; and the frequency spectrum of the correlation coefficient sequence can effectively and conveniently extract the fault characteristic frequency for fault diagnosis of the rotary mechanical equipment.

Description

technical field [0001] The invention relates to a method for extracting fault features of rotating machinery based on time-frequency spectrum correlation analysis. Background technique [0002] Regular shock vibration is a typical feature of fault damage to internal components of rotating machinery, and the frequency of this shock feature is directly related to the specific fault state of rotating machinery. However, due to the fact that there are usually many components in the equipment, there is a certain attenuation in the process of transmitting the fault impact characteristics from the vibration source of the fault damage to the equipment surface along a complex path. In addition, other moving elements in the equipment will also excite corresponding vibrations. Coupled with the interference of working environment noise, the fault impact characteristics that vibration sensors can measure on the equipment are usually relatively weak, especially in the early stage of equip...

AI Technical Summary

Problems solved by technology

However, the time-frequency analysis method of the vibration signal itself cannot reduce the noise. For the fault vibration signal of the rotating mechanical equipment with a low signal-to-noise ratio, only a few shock features with high energy are often highlighted in the time-frequency spectrum, and most shocks with small energy The characteristics are submerged by noise, so the regularity of the impact characteristics cannot be displayed intuitively, and the fault characteristic frequency is also difficult to effectively extract

Images