| Autor: |
Wu Zhang, Heping Hou, Zhuofei Xu, Dan Yu, Yafeng Zhang |
| Rok vydání: |
2021 |
| Předmět: |
|
| Zdroj: |
Advances in Graphic Communication, Printing and Packaging Technology and Materials ISBN: 9789811605024 |
| DOI: |
10.1007/978-981-16-0503-1_71 |
| Popis: |
Rolling bearing is a kind of important transmission component and widely used in the printing machines. Since the printing press works in high pressures and a complex chemical environment, there are always various faults in transmission component and it is hard to detected or diagnosis due to their non-linearity and non-stationarity in signals. In order to solve the problems mentioned above, this paper proposes a fault diagnosis method based on time-frequency image and convolutional neural networks. The vibration signal of rolling bearing is converted into a series of images from time-frequency analysis and then used to realize the fault diagnosis in printing machine. The images are analyzed and recognized based on the convolutional neural network (CNN). A basic structure of CNN is established and then parameters are studied and optimized. After a training for CNN net, faults with different patterns and degrees are distinguished successfully in experiment. Experimental results show that this method can identify faults effectively with a high accuracy. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
