Blind source separation-based gear box compound fault diagnosis method

Abstract

The invention discloses a blind source separation-based gear box compound fault diagnosis method and belongs to the gear box fault diagnosis method technical field. According to the method, a single acceleration sensor is adopted collect the vibration signals of a gear box. The method includes the following steps that: SVD noise reduction processing is performed on acquired signals; EEMD signal decomposition is performed on the signals which have been subjected to the noise reduction processing, so that the signals can be decomposed into a plurality of paths of IMF components and residual components; signals are reconstructed, source number estimation is performed, the number of source signals is estimated; according to the result of the source number estimation, a plurality of paths of IMF components of which the variance is large are extracted on the basis of variance, and signals are reconstructed; a JAD blind source separation method is adopted to solve the extracted IMF components, and the source signals are estimated; spectrum analysis is performed, so that the characteristic frequency of the signals is extracted, and therefore, the compound fault diagnosis of the gear box is completed. The method of the invention is simple in principle; the algorithm of the method is easy to implement; and the physical meaning of the signals is clear; and the blind source separation-based gear box compound fault diagnosis method is an effective gearbox compound fault diagnosis method.

Description

technical field [0001] The invention relates to the technical field of gearbox fault diagnosis methods, in particular to a gearbox compound fault diagnosis method based on blind source separation. Background technique [0002] In the research of gearbox fault diagnosis, most of them analyze and diagnose the single fault of the gearbox, and the research on the compound fault diagnosis technology of the gearbox is less; therefore, the compound fault diagnosis is still a difficult and hot issue in the fault diagnosis of the gearbox. In recent years, some scholars have studied the use of GMCA method to separate the fault characteristics from the gearbox composite fault vibration signal, and achieved certain results. However, since GMCA is essentially a multi-input method, that is, multiple sensor signals need to be collected at the same time, which will not only increase the cost of engineering application, but also greatly increase the difficulty of engineering application. Th...

AI Technical Summary

Problems solved by technology

[0005] In mechanical fault diagnosis, there are many signal generation sources in the mechanical system components, and the signal transmission path is complicated. Due to the limitation of the mechanical structure, it is difficult for most machines to collect signals from multiple locations through multi-channel sensors, and the conditions of classic blind source separation are difficult to achieve.

Images