Early weak fault diagnosis method for rolling bearings

Abstract

The invention aims at providing an early-stage weak fault diagnosis method of a rolling bearing. The early-stage weak fault diagnosis method of the rolling bearing includes the following steps that acceleration sensors are installed on a base, the driven end of a shell and the output end of the shell respectively, and vibration acceleration signals of all the acceleration sensors are collected to obtain a vibration acceleration signal z matrix; the vibration acceleration signals are pre-processed with isolated component analysis to achieve separation of the vibration acceleration signals; separation signals comprising fault feature information are selected; fault signal features of the selected separation signals are accurately extracted by the adoption of self-adaption lifting form wavelet transform; the Pwelch method is used for drawing a power spectrum chart, whether the fault feature frequency exists on the power spectrum or obvious peak values exist on frequency doubling places of the power spectrum or not is observed and therefore, whether the rolling bearing suffers early-stage weak-fault or not is judged. The early-stage weak fault diagnosis method of the rolling bearing has the good performances of reserving details and resisting noise, noise is hindered and the impact feature of a fault signal is fully emphasized, and the early-stage weak fault diagnosis method of the rolling bearing has the better early-stage micro-fault feature extraction effect and higher calculation efficiency.

Description

technical field [0001] The invention relates to a rolling bearing fault diagnosis method, in particular to an early weak fault diagnosis method. Background technique [0002] Rolling bearings are key components of rotating machinery, and their health affects the working status of the entire mechanical system. Failure to deal with accidental failures in time will lead to unimaginable consequences. Therefore, the monitoring and diagnosis of bearings is of great significance. Large-scale power equipment, such as aero-engines, generator sets, etc., in harsh environments such as high speed, heavy load and strong impact, the bearings of its core components are easily damaged, and the fault signal is very weak at the initial stage, and is caused by vibration and vibration caused by other moving parts. Overwhelmed by a lot of random noise. Therefore, how to extract transient mutation features from weak signals or signals that have been submerged by noise is the key to accurate iden...

AI Technical Summary

Problems solved by technology

However, there are still some limitations when using these demodulation methods for analysis. The traditional time-domain analysis method and frequency-domain analysis method analyze early weak faults, but the fault characteristic frequency may be overwhelmed by the frequency structure information of other signal components and background noise. , thus affecting the diagnosis results; when the wavelet transform filter decomposes the signal, the length of the profile signal is reduced by half every time it is decomposed. With the increase of the decomposition scale, the profile signal contains less and less information, and the time resolution decreases; Although wavelet is a nonlinear wavelet transform method based on mathematical morphology, it not only retains the nonlinear analysis characteristics of morphology, but also has the multi-resolution characteristics of wavelet decomposition. The operator cannot be changed for the overall signal, but there will always be a certain mutation in the signal, which requires different filtering characteristics at the mutation point of the signal

Images