基于改进符号序列熵的滚动轴承退化特征提取  

Degradation Feature Extraction of Rolling Bearings Based on Improved Symbolic Sequence Entropy

作　　者：陈臻禕 胡雄[1] 王冰 CHEN Zhenyi;HU Xiong;WANG Bing(Logistics Engineering College,Shanghai Maritime University,Shanghai 201306,China)

机构地区：[1]上海海事大学物流工程学院,上海201306

出　　处：《轴承》2020年第3期51-55,共5页Bearing

基　　金：国家“八六三”计划(2013A20411606);国家自然科学基金项目(31300783);中国博士后科学基金项目(2014M561458)。

摘　　要：研究并提出基于改进符号序列熵的滚动轴承退化特征提取方法,通过引入阈值因子改进原算法中的符号化方式,降低原算法对于幅值变化的敏感度,提高对噪声和冲击的滤除能力,更有利于挖掘信号中的符号序列变化模式。以IMS轴承全寿命数据集为基础,对改进后的符号序列熵的性能和影响因素进行分析。最后,以工业现场中监测得到的齿轮箱轴承全寿命数据为例验证该方法的实际应用效果。实例分析结果表明,符号序列熵以符号动力学为基础,能够从内部揭示表征序列的复杂度大小,区分不同退化状态下的复杂度。A degradation feature extraction method for rolling bearings based on improved symbol sequence entropy is studied and proposed.The symbolic mode in original algorithm is improved by introducing threshold factor to reduce sensitivity of original algorithm to amplitude variation and improve filtering ability for noise and impact,which is more beneficial to mining symbolic sequence change pattern in signal.Based on IMS bearing full life dataset,the performance and influencing factors of improved symbolic sequence entropy are analyzed.Finally,taking the full life data of gearbox bearings obtained by monitoring from industrial site as an example,the practical application effect of the method is verified.The actual example analysis results show that the symbolic sequence entropy is based on symbolic dynamics,which reveals complexity of characterization sequence internally and distinguishes complexity under different degradation states.

关 键 词：滚动轴承 特征提取 符号动力学 退化 阈值 寿命 在线测量 

分 类 号：TH133.33[机械工程—机械制造及自动化] O313.7[理学—一般力学与力学基础]