...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Mechanical systems and signal processing >An extension of the infograms to novel Bayesian inference for bearing fault feature identification
【24h】

An extension of the infograms to novel Bayesian inference for bearing fault feature identification

机译：将信息报扩展到用于轴承故障特征识别的新颖贝叶斯推理

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

Recently, based on negentropy of squared envelope (SE) and of squared envelope spectrum (SES), extensions of spectral kurtosis, the infograms including the SE infogram and the SES infogram, were proposed to detect impulsive and cyclostationary transients. Moreover, they have abilities to detect transients in the cases, where impulsive noises exist or relaxation times of repetitive transients are lower than their repetition rate. Nevertheless, the infograms are fast filtering algorithms and cannot achieve optimal Filtering for bearing fault feature identification. This paper aims to extend the infograms to novel Bayesian inference based optimal wavelet filtering for bearing fault feature identification. The innovations of this paper are summarized as follows. Firstly, a state space model of wavelet parameters is presented. Here, wavelet parameters are the states of the state space model. Monotonically increasing guess negentropy measurements are constructed. Secondly, either the SE infogram or the SES infogram is employed to initialize the state space model. Then, considering Gaussian disturbance on wavelet parameters, wavelet parameters are assumed to follow a joint Gaussian distribution. Thirdly, spherical cubature integration based Bayesian inference is introduced to iteratively establish posterior wavelet parameters distributions. At last optimal wavelet parameters are determined from the posterior wavelet parameters distributions so as to conduct optimal wavelet filtering. Two instance studies including simulated and experimental bearing fault data were investigated to illustrate how the proposed Bayesian inference method works. The results show that the proposed Bayesian inference method is convergent and provides more fault signatures than the infogram.
机译：最近，基于平方包络（SE）和平方包络谱（SES）的负熵，提出了光谱峰度扩展，包括SE信息图和SES信息图的信息图，以检测脉冲和循环平稳瞬变。此外，在存在脉冲噪声或重复瞬态的弛豫时间低于其重复率的情况下，它们具有检测瞬态的能力。但是，信息报是快速过滤算法，不能实现用于轴承故障特征识别的最佳过滤。本文旨在将信息报扩展到基于贝叶斯推理的新颖小波滤波，以进行轴承故障特征识别。本文的创新之处总结如下。首先，提出了一个小波参数状态空间模型。在此，小波参数是状态空间模型的状态。构造单调递增的猜测负熵测量。其次，采用SE信息图或SES信息图来初始化状态空间模型。然后，考虑到小波参数的高斯扰动，假设小波参数服从联合高斯分布。第三，引入基于贝叶斯推断的球面培养集成，迭代地建立后验小波参数分布。最后，从后验小波参数分布确定最优小波参数，以进行最优小波滤波。研究了两个实例研究，包括模拟和实验轴承故障数据，以说明所提出的贝叶斯推理方法如何工作。结果表明，所提出的贝叶斯推理方法具有收敛性，并且比信息报提供了更多的故障特征。

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