基于VMD和MP算法的旋转机械故障特征提取  

Rotating machinery fault feature extraction based on VMD and MP algorithm

作　　者：许爱华[1] 闫俊泉 伍旭灿 盖磊[1] 

机构地区：[1]东北石油大学电气信息工程学院,大庆163318

出　　处：《国外电子测量技术》2017年第8期11-17,共7页Foreign Electronic Measurement Technology

摘　　要：旋转机械故障振动信号为不平稳、非线性而且信噪比低的多分量信号,故障特征很难提取。小波分析方法受限于小波基和分解层次的选取,经验模态分解(EMD)等自适应分解方法存在模态混叠和虚假分量的问题,应用在旋转机械故障提取上容易误判。提出了利用匹配追踪算法(MP)改进变分模态分解(VMD)的故障诊断方法。通过构建合适的仿真信号仿真和对实测信号的仿真,验证了该方法的可行性和实用性。Rotating machine fault vibration signal is not stable,non-linear and low signal to noise ratio of multi-component signal,fault characteristics difficult to extract.Wavelet analysis method is limited by the selection of wavelet basis and decomposition level,empirical mode decomposition （EMD） and other adaptive decomposition method exists modal aliasing and false component problem,and is easy to misjudge in the extraction of rotating machinery.In this paper,a fault diagnosis method based on matching tracking algorithm （MP） is proposed to improve variational mode decomposition （VMD）.The feasibility and practicability of the method are verified by constructing the simulation of the simulation signal and simulating the measured signal.

关 键 词：故障诊断 时频分析 VMD MP 

分 类 号：TP14[自动化与计算机技术—控制理论与控制工程]