A Weighted Deep Learning Model with Sampling Strategy for Unbalanced Bearing Condition Assessment
ID:93 View Protection:ATTENDEE Updated Time:2022-05-16 09:39:21 Hits:399 Poster Presentation

Start Time:Pending(Asia/Shanghai)

Duration:Pending

Session:No Session »

Video No Permission Presentation File

Tips: Only the registered participant can access the file. Please sign in first.

Abstract
滚动轴承在旋转机械中起着极其重要的作用。轴承的独特位置决定了一旦发生故障,不仅会造成巨大的经济损失,还会危及人身安全。因此,轴承失效状态的评估一直是一个热门的研究领域。本文解决了传统轴承状态评估方法存在的两个挑战:一是需要手动从轴承数据中提取特征,过度依赖专业知识;第二个是故障样本和正常样本之间的普遍不平衡,这会影响模型的性能。论文提出了一种带有采样策略的两层 CNN-LSTM 加权深度学习模型(wTLCL-O)来应对这些挑战,它包含两层 CNN,具有过采样策略和加权成本敏感损失函数的两层 LSTM。文章在法国PRONOSTIA轴承失效数据集上进行实验,结果表明wTLCL-O在鲁棒性和稳定性方面优于其他传统算法,提高了非平衡条件下轴承状态评估的预测精度。
Keywords
Speaker
ChenYilin
大连交通大学

Submit Comment
Verify Code Change Another
All Comments
Important Date
  • Conference Date

    May 27

    2022

    to

    May 29

    2022

  • Feb 28 2022

    Draft paper submission deadline

  • May 29 2022

    Registration deadline

  • Jun 22 2022

    Contribution Submission Deadline

Sponsored By
IEEE Beijing Section
China Electrotechnical Society
Southeast University
Supported By
IEEE Industry Applications Society
IEEE Nanjing Section
Contact Information