• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International Conference on Mechanical Engineering and Mechanics vol.1; 20051026-28; Nanjing(CN) >Fault Diagnosis Based on Fusion of Multi-channel Observations by Sensors with Independent Component Analysis and Mutual Information
【24h】

Fault Diagnosis Based on Fusion of Multi-channel Observations by Sensors with Independent Component Analysis and Mutual Information

机译：基于独立分量分析和互信息的传感器多通道观测融合的故障诊断

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Independent Component Analysis (ICA), as expanding of Principal Component Analysis (PCA) to higher statistic than second order, has such features as signal processing blindly and redundancy reduction, is fit for de-noising and feature extraction in fault diagnosis. In this paper, a new method for fault diagnosis based on fusion of multi-channel observations by sensors with independent component analysis and mutual information (MI) was proposed, by which redundancy among observations by sensors (vibrations or acoustical signals) can be reduced, high dimensional fault features can be compressed, and, feature extraction can be made simpler and faster. By means of such characteristics as redundancy reduction of ICA and higher statistic than second order of MI, two-step fusion and compression of multi-channel observations by sensors were implemented. Firstly, multi-channel observations were inputted into an ICA processing unit, by which additive noise and interference were removed, and information embedded into observations was re-adjusted. Then, a procedure for feature extraction named MI based Wavelet was used for capturing higher statistical features than second order. Finally, by means of clustering analysis of the fused feature vectors, capacity of the new method in classification for different fault patterns was tested. Results of experiment showed that multi-channel observations by sensors were fused sufficiently, dimension of observations was reduced remarkably, and, at the same time, good performance in classifying different fault patterns was obtained. Thus, it is possible to construct an on-line and real-time fault classifier by the application of this new method to fault diagnosis of a complex machine.
机译：独立分量分析（ICA）是将主分量分析（PCA）扩展到比二阶统计更高的统计量，具有盲目处理信号和减少冗余的功能，适合在故障诊断中进行降噪和特征提取。本文提出了一种新的故障诊断方法，该方法基于具有独立分量分析和互信息（MI）的传感器对多通道观测值的融合，可以减少传感器观测值（振动或声音信号）之间的冗余，高维断层特征可以被压缩，特征提取可以变得更简单，更快。借助于ICA的冗余减少和比MI的二阶更高的统计量等特性，实现了两步融合和传感器对多通道观测值的压缩。首先，将多通道观测值输入到ICA处理单元中，从而消除附加噪声和干扰，并重新调整嵌入观测值中的信息。然后，使用基于MI的小波的特征提取过程来捕获比二阶更高的统计特征。最后，通过聚类特征向量的聚类分析，测试了该新方法在分类不同故障模式方面的能力。实验结果表明，传感器的多通道观测值融合良好，观测值的维数显着减小，同时，在分类不同的故障模式方面取得了良好的性能。因此，通过将该新方法应用于复杂机器的故障诊断中，可以构造一个在线实时故障分类器。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号