Mechanical device fault diagnosis method based on saliency test

Abstract

The invention discloses a mechanical device fault diagnosis method based on saliency test, and relates to the technical field of signal processing. The method comprises the steps that the sample characteristic of the reference power spectrum of a mechanical device working in a normal state is acquired, and the sample characteristic of a power spectrum to be tested is acquired when the device worksin a fault state; the test hypothesis that the characteristic data of the power spectrum to be tested are equal to the characteristic data of the reference power spectrum is proposed; saliency test is carried out according to the sample characteristic of the power spectrum to be tested and the sample characteristic of the reference power spectrum; and according to a saliency test result, a signalto be tested is filtered and a fault signal in the signal to be tested is extracted to carry out fault diagnosis. Filtering based on saliency test can effectively filter out a noise signal in a strong noise background. The effective characteristic signal of the fault signal is extracted, so that fault diagnosis more accurate.

Description

technical field [0001] The invention relates to the technical field of signal processing, in particular to a mechanical equipment fault diagnosis method based on a significance test. Background technique [0002] During the working process of the mechanical equipment, if it fails, some parameters of the equipment will change. By analyzing the changes of these parameters, the cause and location of the equipment failure can be determined, and the online monitoring and fault diagnosis of the mechanical equipment can be completed. [0003] However, when the mechanical equipment has just failed, the fault signal is usually very weak. At the same time, due to the complex working environment of the mechanical equipment and various other factors, it is impossible to extract any information from such a strong noise background using analysis algorithms such as Fourier and wavelet transform. Fault features, which will cause the fault signal features extracted early are not obvious, and...

AI Technical Summary

Problems solved by technology

[0003] However, when the mechanical equipment has just failed, the fault signal is usually very weak. At the same time, due to the complex working environment of the mechanical equipment and various other factors, it is impossible to extract any information from such a strong noise background using analysis algorithms such as Fourier and wavelet transform. Fault features, which will cause the early extracted fault signal features to be inconspicuous and often overwhelmed by noise

Therefore, it is necessary to filter the noise signal, analyze it after extracting the effective signal, and effectively extract the characteristic frequency of the signal, so as to effectively diagnose the equipment status. The traditional filtering methods currently used have their own shortcomings. When the value filtering algorithm is used, the calculation time is long and it is powerless for signals with too complicated time complexity, which leads to poor online monitoring and fault diagnosis of mechanical equipment.

Images