报告详情
Fault diagnosis of mechanical equipment using a gradient information-constrained generative adversarial networks
编号：113
访问权限：仅限参会人
更新：2025-11-10 15:36:58 浏览：34次
张贴报告
报告开始：暂无开始时间（Asia/Shanghai）
报告时间：暂无持续时间
所在会场：[暂无会议] [暂无会议段]
暂无文件
摘要
The performance of mechanical equipment may decrease due to long-term operation under harsh working conditions, so to ensure the normal operation of the equipment, it is necessary to accurately diagnose the health status of key components. To improve the fault diagnosis performance of mechanical equipment in the case of scarce fault data, a gradient information constrained generative adversarial network is proposed to enhance the fault diagnosis capability by data augmentation. Firstly, the encoder is integrated into the discriminator to extract effective features from the original samples, constructing a generative adversarial network with stronger data synthesis capabilities. By generating fault samples that are similar to real samples, the effectiveness of fault diagnosis under sample scarcity can be improved. Secondly, a gradient information constraint mechanism is constructed based on the information bottleneck theory to improve the training stability of the generative adversarial network. By imposing constraints on the mutual information between the input data and the deep features of the discriminator, the feedback gradient of the discriminator to the generator can be effectively adjusted, further promoting the stable training of the network structure. The experimental verification on the bearing dataset shows that the proposed method has excellent ability in small sample diagnostic tasks.
关键词
fault diagnosis, mechanical equipment, generative adversarial networks, small sample, gradient information
报告人
shaowei liu
Overall Technical De Xi'an Modern Control Technology Research Institute稿件作者
全部评论
重要日期
-
会议日期
11月21日
2025
至11月23日
2025
-
10月20日 2025
初稿截稿日期
-
11月23日 2025
注册截止日期
主办单位
IEEE Instrumentation and Measurement Society
South China University of Technology
South China University of Technology
承办单位
South China University of Technology
联系方式
- Icsmd
- ic******@163.com
历届会议
-
2023年11月02日 中国 China
International Conference on Sensing, Measurement & Data Analytics in the era of Artificial Intelligence -
2023年11月02日 中国 Xi'an
2023 International Conference on Sensing, Measurement & Data Analytics in the era of Artificial Intelligence -
2022年11月30日 中国 Harbin
International Conference on Sensing, Measurement & Data Analytics in the era of Artificial Intelligence -
2021年10月21日 中国 Nanjing
2021 International Conference on Sensing, Measurement and Data Analytics in the era of Artificial Intelligence -
2020年10月15日 中国 Xi'an
International Conference on Sensing, Measurement and Data Analytics in the era of Artificial Intelligence
发表评论