基于广义形态滤波和相关系数的Hilbert-Huang变换方法  

An Improved Hilbert-Huang Transform Based on Generalized Morphological Filter and Correlation Coefficients

作　　者：刘畅[1] 周川[1] 伍星[1] 迟毅林[1] 

机构地区：[1]昆明理工大学机电工程学院,昆明650093

出　　处：《机械科学与技术》2011年第1期71-75,共5页Mechanical Science and Technology for Aerospace Engineering

基　　金：国家自然科学基金项目(50805071);云南省教育厅科学研究基金项目(08J0009)资助

摘　　要：针对Hilbert-Huang变换(HHT)中噪声引起的模态裂解和虚假模态问题,应用广义形态滤波降噪法和相关系数法改进了HHT。该方法首先将广义形态滤波器作为经验模态分解(EMD)的预处理,抑制噪声干扰;然后利用相关系数法去除虚假固有模态函数(IMF);最后对真实的IMF进行Hilbert谱分析,得到改进的HHT。利用仿真信号检验了改进HHT的效果,并将该方法应用于滚动轴承故障诊断中。实验分析结果表明:该方法能有效地进行滚动轴承早期故障诊断。For the problems of mode fission and pseudo modes caused by noises in Hilbert-Huang Transform（HHT）,an improved HHT is proposed by using a de-noising method based on generalized morphological filtering and a correlation coefficients method.Firstly,a generalized morphological filter was used to process the signal before empirical and mode decomposition（EMD）,with restrained the interferences of noises.Then,a correlation coefficients method was used to eliminate the pseudo intrinsic mode function（IMF）.Finally,the really IMFs were analyzed by using Hilbert spectrum,and the improved HHT was obtained.The advantages of this improved HHT were verified by simulation experiment.The improved HHT was applied to bearing fault diagnosis.Experiment result shows that this method is effective to diagnose the early fault of rolling bearings.

关 键 词：HILBERT-HUANG变换 经验模态分解 广义形态滤波器 相关系数 滚动轴承 故障诊断 

分 类 号：TN911[电子电信—通信与信息系统]