Rolling Bearing Fault Diagnosis Method Based on Ensemble Intrinsic Time Scale Decomposition Algorithm

Abstract

A rolling bearing fault diagnosis method based on an ensemble intrinsic time scale decomposition algorithm, comprising the following steps: using a displacement sensor to collect a vibration signal of a rolling bearing; using an ensemble intrinsic time scale decomposition algorithm to decompose the collected vibration signal to generate several rotation components and Residual signal; select sensitive rotating components that can reflect fault information from all rotating components; perform envelope spectrum analysis on sensitive rotating components; judge fault type by analyzing envelope spectrum amplitude corresponding to fault characteristic frequency. The invention solves the modal mixing problem of the inherent time scale decomposition algorithm, and lays a good foundation for feature extraction. By calculating the kurtosis, the rotation component that is more sensitive to the fault is selected, and finally by analyzing the sensitive component corresponding to the fault characteristic frequency The envelope spectrum amplitude judges the fault type. The invention can accurately identify rolling bearing faults and is suitable for rolling bearing fault diagnosis.

Description

technical field [0001] The invention relates to a rolling bearing fault diagnosis method. In particular, it involves a fault diagnosis method for rolling bearings based on an ensemble intrinsic time scale decomposition algorithm. Background technique [0002] Vibration analysis is the simplest and most direct method for fault diagnosis of rolling bearings. Typical vibration analysis methods include: wavelet transform, Wigner distribution, empirical mode decomposition, etc., but they all have their own shortcomings. Wavelet transform is not adaptive and the choice of basis function depends too much on the user's experience. The Wegener distribution has high time-frequency resolution, but the appearance of cross terms limits its application. Empirical mode decomposition is an adaptive time-frequency decomposition method, which has been widely used in power machinery fault diagnosis, but this method has over-envelope, modal confusion, end-point effects, and problems caused by...

AI Technical Summary

Problems solved by technology

The first problem is that the added white noise cannot be completely offset. Although the final white noise residual can be reduced by selecting a larger number of added noises, the corresponding calculation time will also be greatly increased, which is contrary to the impact of fault diagnosis on the algorithm speed. requirements

Another problem is that the addition of different white noise may cause the number of rotation components generated by each decomposition to be unequal, which brings difficulties to the final aggregated average.

Images