基于二维多尺度时频散布熵的滚动轴承故障诊断方法  

A two-dimensional multi-scale time-frequency distribution entropy based rolling bearing fault diagnosis method

作　　者：郑近德[1] 李嘉绮 潘海洋[1] 童靳于[1] 刘庆运[1] ZHENG Jinde;LI Jiaqi;PAN Haiyang;TONG Jinyu;LIU Qingyun(School of Mechanical Engineering,Anhui University of Technology,Ma’anshan 243032,China)

机构地区：[1]安徽工业大学机械工程学院,安徽马鞍山243032

出　　处：《振动与冲击》2023年第8期215-225,共11页Journal of Vibration and Shock

基　　金：国家自然科学基金(51975004);安徽省自然科学基金项目(2008085QE215)。

摘　　要：多尺度散布熵(multi-scale dispersion entropy,MDE 1D)是一种有效衡量一维振动信号复杂性特征的非线性动力学分析法,但其仅能反映振动信号时域中的复杂性特征,无法完整反映振动信号频域的非线性动力学信息。为此,在二维散布熵(two-dimensional dispersion entropy,DE_(2D))的基础上,提出二维时频散布熵(two-dimensional time-frequency dispersion entropy,TFDE_(2D))用于衡量时间序列的时频复杂性特征。同时,为更完整地反映时频分布在不同尺度下的复杂信息,受多尺度粗粒化启发,将传统粗粒化方法拓展到二维多尺度粗粒化,提出了二维多尺度时频散布熵(two-dimensional multi-scale time-frequency dispersion entropy,MTFDE_(2D)),用来量度振动信号时频分布的多尺度复杂性特征。在此基础上,将其应用于滚动轴承故障诊断中的非线性特征提取,提出一种基于MTFDE_(2D)和萤火虫优化支持向量机的滚动轴承智能诊断方法。最后,将所提方法应用于滚动轴承试验数据分析,并与现有方法进行对比。结果表明,所提方法不仅能有效地提取故障特征,实现不同轴承故障类型和故障程度的有效诊断,且诊断效果优于对比法。Multi-scale dispersion entropy(MDE 1D)is an effective nonlinear dynamics analysis method to measure the complexity characteristics of one-dimensional vibration signal,but it can only reflect the complexity characteristicsin the time domain of the vibration signal,and cannot completely reflect the nonlinear dynamics informationin the frequency domain of the vibration signal.To this end,two-dimensional time-frequency dispersion entropy(TFDE_(2D))was proposed based on the two-dimensional dispersion entropy(DE_(2D))to measure the time-frequency complexity characteristics of time series.Meanwhile,the traditional coarse-grained method was extended to two-dimensional multi-scale coarse-grained to reflect the complexity of the time-frequency distribution at different scales more completely,and the two-dimensional multi-scale time-frequency dispersion entropy(MTFDE_(2D))was proposed to measure the multi-scale complexity characteristics of the time-frequency distribution of vibration signal.On this basis,an intelligent diagnosis method for rolling bearings based on MTFDE_(2D) and firefly optimized support vector machines was proposed to extraction of nonlinear features in rolling bearing fault diagnosis.Finally,the proposed method was applied to the analysis of experimental data of rolling bearings and compared with the existing methods.Results show that the proposed method can not only be effective in extracting fault characteristics and realize effective diagnosis of different bearing fault types and degrees,but also has better diagnostic effects than the compared methods.

关 键 词：时频散布熵 多尺度时频散布熵 滚动轴承 萤火虫优化支持向量机 故障诊断 

分 类 号：TH165.3[机械工程—机械制造及自动化]