基于排序模式相异性分析的轴承健康监测

发布时间：2018-05-10 10:06

本文选题：排序模式分析 + 信息散度　；参考：《中国机械工程》2017年06期

【摘要】：排序模式分析方法通过相空间重构将一维振动时间序列映射到排序模式概率分布,来揭示序列内部结构的复杂性变化,为微弱信号特征提取提供了一种新视角。将排序模式分析和信息散度相结合,提出一种排序信息散度指标,用于对设备不同运行状态下的振动信号在高维相空间中排序模式概率分布的差异性进行量化分析,并用于轴承内圈不同损伤程度评估和轴承全寿命退化趋势分析。结果表明,与传统的时域统计指标及小波熵、近似熵、排序熵等非线性复杂度指标相比较,所提出的排序信息散度指标具有较好的故障程度量化分析性能,对轴承早期故障退化更加敏感,且稳定性好、计算效率高,利于工程实现。
[Abstract]:The method of sorting pattern analysis maps one-dimensional vibration time series to the probability distribution of sorting mode through phase space reconstruction to reveal the complexity of the internal structure of the sequence and provide a new perspective for feature extraction of weak signals. By combining sorting mode analysis with information divergence, a sort of information divergence index is proposed, which can be used to quantify the difference of the probability distribution of sorting mode in high-dimensional phase space of vibration signals in different operation states of equipment. It is used to evaluate the different damage degree of bearing inner ring and to analyze the whole life degradation trend of bearing. The results show that compared with the traditional time-domain statistical index, wavelet entropy, approximate entropy, sort entropy and other nonlinear complexity indexes, the proposed ranking information divergence index has better performance in quantitative analysis of fault degree. It is more sensitive to the early failure degradation of bearing and has good stability and high calculation efficiency.
【作者单位】：燕山大学电气工程学院;
【基金】：河北省高等学校科学技术研究重点项目(ZD20131080) 国家自然科学基金资助项目(51505415) 河北省自然科学基金资助项目(F2016203421) 中国博士后科学基金资助项目(2015M571279)
【分类号】：TH133.33

【相似文献】