We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
基于注意力机制的滚动轴承故障诊断方法.
- Authors
李秋婷; 王秀青; 解飞; 杨云鹏; 杜文霞
- Abstract
On the basis of deep learning and attention mechanism, a rolling bearing fault diagnosis CAR model based on convolutional block attention module (CBAM) and residual network (ResNet) and a rolling bearing fault diagnosis model based on Transformer (multiheaded self attention) are proposed. CAR model introduces the attention mechanism to improve the ability of feature extraction and residual network to prevent the gradient disappearance or explosion effectively; Transformer model uses multihead self attention to extract the features of fault signals effectively. CAR and Transformer models are applied to fault diagnosis for rolling bearings under various sampling frequencies and rotational speeds, the highest diagnostic accuracy rates of two models are 99. 91% and 99. 85% respectively for bearing datasets from CWRU, and 98. 71% and 99. 85% respectively for bearing datasets from Jiangnan University, both of them being superior to existing research results. The two proposed models are simple in structure and easy to implement, meeting the requirements of high fault diagnosis accuracy in practical applications.
- Subjects
FAULT diagnosis; ROLLER bearings; FEATURE extraction; DEEP learning; SELF; EXPLOSIONS
- Publication
Bearing, 2023, Issue 10, p84
- ISSN
1000-3762
- Publication type
Article
- DOI
10.19533/j.issn1000-3762.2023.10.013