Bearing Fault Diagnosis Method based on Sensitive Component and MCPG

Abstract

Aiming at the problem that is difficult to accurately identify rolling bearing faults， a fault diagnosis method based on sensitive components and Multi Convolution Pooling Group （MCPG） is proposed. Firstly， the Empirical Mode Decomposition （EMD） is used to decompose the original signal into multiple Intrinsic Mode Function（IMF）， and the discrete Fréchet distance is used as the measurement index， the fault sensitive components are selected as the fault data sources representing different fault types. Then， a MCPG deep neural network architecture is proposed， and sensitive data sources are used to train and test the model to achieve the data-driven bearing fault diagnosis. Through experimental verification， it is proved that the method has good recognition effect on different types of vibration data （different speeds， different damage types， different damage degrees）.

Keywords

• Fault diagnosis
• Rolling bearing
• EMD
• Discrete Fréchet
• Convolutional neural networks