Research on bearing fault diagnosis method based on ensemble dropout

Fangcheng Cao

doi:10.1117/12.3007621

25 October 2023 Research on bearing fault diagnosis method based on ensemble dropout

Fangcheng Cao

Proceedings Volume 12801, Ninth International Conference on Mechanical Engineering, Materials, and Automation Technology (MMEAT 2023); 128015X (2023) https://doi.org/10.1117/12.3007621
Event: Ninth International Conference on Mechanical Engineering, Materials, and Automation Technology (MMEAT 2023), 2023, Dalian, China

Abstract

One-dimensional convolutional neural network with dropout step is a common method for bearing fault diagnosis, but the generalization ability and fitting ability of this network are contradictory with parameter variation, resulting in insufficient fault diagnosis accuracy. In order to improve the fault diagnosis accuracy, a scheme based on soft voting method and ensemble dropout is proposed to fuse the advantages of the ability to fit the low dropout probability scheme and generalize the high dropout probability scheme to realize a new convolutional neural network with both of them. Experiments show that the accuracy of the method is significantly improved, the sensitivity of the parameters of this method is high, and there is a specific law of how the parameters affect the effect.

(2023) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Fangcheng Cao "Research on bearing fault diagnosis method based on ensemble dropout", Proc. SPIE 12801, Ninth International Conference on Mechanical Engineering, Materials, and Automation Technology (MMEAT 2023), 128015X (25 October 2023); https://doi.org/10.1117/12.3007621

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available