Improved Mobile-Vit model and its application in rolling bearing fault diagnosis

Huina Mu, Xiaoyun Zeng, Wei Liu*, Jialin Sun, Yeshu Zhang

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In order to improve the efficiency and accuracy of rolling bearing fault diagnosis, FDMT fault diagnosis model was proposed. Firstly, based on the Mobile-Vit model, a FDMT model for fault diagnosis with one-dimensional rolling bearing vibration signal as direct input was designed, and the structure of the FDMT model was given. Secondly, the FDMT model was trained and tested using the CWRU data set to verify the effectiveness of the model. Then, the influence of hyperparameter adjustment on the performance of FDMT model was discussed. Finally, the FDMT model was compared with other deep learning models. The results show that the improved Mobile-Vit model has higher fault diagnosis accuracy.

Original languageEnglish
Title of host publication2024 6th International Conference on System Reliability and Safety Engineering, SRSE 2024
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages308-315
Number of pages8
ISBN (Electronic)9798350356083
DOIs
Publication statusPublished - 2024
Event6th International Conference on System Reliability and Safety Engineering, SRSE 2024 - Hangzhou, China
Duration: 11 Oct 202414 Oct 2024

Publication series

Name2024 6th International Conference on System Reliability and Safety Engineering, SRSE 2024

Conference

Conference6th International Conference on System Reliability and Safety Engineering, SRSE 2024
Country/TerritoryChina
CityHangzhou
Period11/10/2414/10/24

Keywords

  • fault diagnosis
  • FDMT
  • Mobile-ViT
  • rolling bearings

Cite this