Loading…
Research on deep feature learning and condition recognition method for bearing vibration
•A PFC-CNN method is proposed for bearing vibration state recognition.•The PFC-CNN can also identify the fault type under different working loads.•T-SNE tool in the manifold learning method was used to visualize the results.•The bearing fault recognition accuracy of PFC-CNN is validated with CWRU da...
Saved in:
Published in: | Applied acoustics 2020-11, Vol.168, p.107435, Article 107435 |
---|---|
Main Authors: | , , , , , |
Format: | Article |
Language: | English |
Subjects: | |
Citations: | Items that this one cites Items that cite this one |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | •A PFC-CNN method is proposed for bearing vibration state recognition.•The PFC-CNN can also identify the fault type under different working loads.•T-SNE tool in the manifold learning method was used to visualize the results.•The bearing fault recognition accuracy of PFC-CNN is validated with CWRU data.
To obtain comprehensive bearing vibration characteristic information and improved the accuracy of vibration state recognition, this paper proposed the pre-fully connected deep CNN (PFC-CNN) method and based on this, established the bearing vibration state recognition model. Firstly, the CNN's front fully connected structure was design and the front fully connected layer was used to reduce the complexity of the signal and extracted the global characteristics of the signal. On this basis, the fusion feature learning of global features and local features (which got by CNN's local perceptual field) was studied. Secondly, the back-propagation training of the whole network was studied, and the discrimination degree and recognition accuracy of the corresponding features of different state categories were finally improved. Results of experimental study showed that compared with other methods, the proposed method in this paper improved the identification accuracy which can reach 96.25%. |
---|---|
ISSN: | 0003-682X 1872-910X |
DOI: | 10.1016/j.apacoust.2020.107435 |