基于VMD与Hilbert谱的旋转机械碰摩故障诊断方法  

Rub-impact Fault Diagnosis of Rotating Machinery Based on VMD and Hilbert Spectrum

作　　者：赵岩 朱均超[1] 张宝峰[1] 邵磊[1] 李季[1] 代煜[2] 

机构地区：[1]天津理工大学天津市复杂系统控制理论及应用重点实验室,天津300384 [2]南开大学机器人与信息自动化研究所,天津300071

出　　处：《振动．测试与诊断》2018年第2期381-386,共6页Journal of Vibration,Measurement & Diagnosis

基　　金：天津市教委科研计划资助项目(2017KJ253)

摘　　要：为了有效地诊断旋转机械中的碰摩故障,提出了基于变分模态分解(variational mode decomposition,简称VMD)与Hilbert谱分析的故障诊断方法。首先,利用VMD将碰摩故障信号自适应地分解为若干个不同频率段的本征模态函数(intrinsic mode function,简称IMF),并与集合经验模态分解(ensemble empirical mode dcomposition,简称EEMD)的处理结果进行对比分析;然后,在VMD分解的基础上,利用Hilbert谱对故障信号的时频特性进行分析。实验结果表明:与传统的频谱分析相比,该方法不但可以准确反映故障信号的频率成分,而且可以反映频率随时间的变化情况;与EEMD相比,该方法可以有效抑制模态混叠,更加准确地反映故障信息,从而验证了该方法的可行性与有效性。According to the rub-impact fault of rotating machinery,the diagnosis method based on variational mode decomposition(VMD)and Hilbert spectrum is studied.The fault signal is decomposed into a series of intrinsic mode function(IMFs)by VMD method.The result of VMD is compared with that of EEMD.The time-frequency characteristic of the fault signal is analyzed by the Hilbert spectrum.The result shows that the proposed diagnosis method can accurately show the frequency components of the fault signal.Compared with the traditional method based on frequency spectrum analysis,the proposed method can show the relationship between frequency and time.Compared with the method based on EEMD,this proposed method can effectively inhibit mode mixing and accurately show the fault information.The experimental result shows that the presented diagnosis method is feasible and valid.

关 键 词：故障诊断 旋转机械 Hilbert谱 变分模态分解 

分 类 号：TH86[机械工程—仪器科学与技术] TP23[机械工程—精密仪器及机械]