A bearing fault diagnosis method, device, medium and product

By combining mathematical morphology and convolutional neural networks, the problem of inaccurate mode segmentation in traditional bearing fault diagnosis methods under strong noise environments is solved, realizing intelligent diagnosis and feature extraction of bearing faults, and improving the accuracy and real-time performance of diagnosis.

CN119046808BActive Publication Date: 2026-06-09NORTH CHINA ELECTRIC POWER UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN Β· China
Patent Type
Patents(China)
Current Assignee / Owner
NORTH CHINA ELECTRIC POWER UNIV
Filing Date
2024-08-09
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Traditional bearing fault diagnosis methods rely on manual inspection or simple spectral analysis, which are characterized by strong subjectivity, low accuracy, poor real-time performance, and difficulty in accurately classifying modes and extracting features in high-noise environments.

Method used

The spectrum is divided using empirical wavelet transform based on mathematical morphology, the autocorrelation kurtosis of the modal components is calculated, and a fault diagnosis model is trained using a convolutional neural network to extract bearing fault features.

Benefits of technology

It improves the accuracy and real-time performance of bearing fault diagnosis, realizes intelligent diagnosis and feature extraction of bearing faults, and enhances the accuracy of mode division in high-noise environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119046808B_ABST
    Figure CN119046808B_ABST
Patent Text Reader

Abstract

This application discloses a bearing fault diagnosis method, equipment, medium, and product, relating to the field of fault diagnosis. The method includes: acquiring bearing information data; performing mode decomposition on the information data based on mathematical morphology and using an empirical wavelet transform spectral partitioning method to obtain a set of mode components; calculating the autocorrelation kurtosis of any mode component in the set; determining an autocorrelation kurtosis feature vector based on all autocorrelation kurtosis values; inputting the autocorrelation kurtosis feature vector into a fault diagnosis model and outputting a diagnosis result; the diagnosis result is either normal or a fault exists; the fault diagnosis model is trained based on a convolutional neural network. This application enables feature extraction and intelligent diagnosis of bearing faults.
Need to check novelty before this filing date? Find Prior Art