Mechanical fault diagnosis method based on multi-sensor multivariate data fusion

Abstract

The invention discloses a mechanical fault diagnosis method based on multi-sensor multivariate data fusion. The method comprises the following steps: acquiring multivariate vibration signals x1(t), x2(t),..., xn(t) of a bearing acquired by multiple sensors; preprocessing the multivariate signal by using a non-local mean denoising method; adopting noise-assisted adaptive projection eigentransformation multivariate empirical mode decomposition (NAAPITMEMD) method to obtain a plurality of IMF (intrinsic mode function) groups which are the same in length and have the same appearance sequence of same-frequency components of the signals in each group; calculating a fault correlation factor (FCF) by using correlation analysis, and selecting effective IMFs; and finally, extracting a fault characteristic frequency through spectrum analysis to realize mechanical fault diagnosis. The common problems of power imbalance and mode aliasing in multi-element signal processing can be effectively relieved, and the accurate mechanical fault characteristic frequency can be obtained.

Description

technical field [0001] The invention belongs to the field of mechanical fault diagnosis, and in particular relates to a mechanical fault diagnosis method based on multi-sensor multivariate data fusion. Background technique [0002] Due to the complex structure of rotating machinery and equipment, it is prone to failure under harsh working conditions in the industrial field. In the field of mechanical fault diagnosis, the vibration signal collected by a single sensor contains limited information, especially when multiple faults occur in different parts of the mechanical equipment, the vibration signal collected from a single channel often misses some fault characteristic information, resulting in Diagnostic error. Using multiple sensors to collect multiple vibration signals at different locations can extract more accurate mechanical fault characteristic frequencies. Therefore, multivariate data fusion methods for multivariate signals, especially for synchronous decompositio...

AI Technical Summary

Problems solved by technology

However, due to the different positions of the multi-sensors in the rotating machinery, the multiple signals collected by the multi-sensors often have power imbalances between different channels, and the traditional MEMD method cannot effectively alleviate the problem of power imbalances between channels. Unable to achieve the expected signal processing effect

Images