基于EEMD的液压系统振动信号相关主分量分析效果研究  

Study on Analysis Effect of Related Principal Component of Vibration Signal of Hydraulic System Based on EEMD

作　　者：陈灏[1] 张梅军[1] 柴凯[1] 黄杰[1] 

机构地区：[1]解放军理工大学野战工程学院,江苏南京210007

出　　处：《机床与液压》2015年第1期185-188,135,共5页Machine Tool & Hydraulics

基　　金：国家自然科学基金资助项目(51175511)

摘　　要：为了研究液压系统振动信号经EEMD分解前几阶IMF分量能否代表EEMD分解信号的主要成分,提出了EEMD分解的相关主分量分析,研究了EEMD相关主分量分析的效果。运用EEMD分解信号,得到其IMF分量,计算出各分量与原信号的相关系数,从中找出信号的相关主分量,通过对主分量进行Hilbert包络谱分析,并与原信号的Hilbert包络谱比较来验证EEMD相关主分量分析效果。通过对实测信号研究表明,某液压缸连续信号经EEMD分解后,与原信号相关性较大的5阶IMF分量:IMF1、IMF2、IMF5、IMF6、IMF7包含原信号主要成分,能表示该信号EEMD分解的主分量,而前5阶却不能完整表示原信号的主分量;对液压缸冲击信号分析发现,EEMD分解的前两阶IMF分量:IMF1、IMF2与原信号相关性较大,能够准确地表示原信号所包含的频率成分和信息,能表示信号EEMD分解的主分量,代表EEMD分解信号的主要成分。因此,EEMD相关主分量分析能突出EEMD分解的主要频率成分,对液压系统振动信号分析效果良好。In order to study on the vibration signal of hydraulic system through EEMD decomposition of main components of order IMF components can represent the decomposition of EEMD signals, analysis of principal component decomposition of EEMD was pres-ented, and the effect of EEMD related principal component was researched. EEMD was applied to decompose the signal to get the IMF component. The correlation coefficient of each component of the original signal was calculated to find the relevant principal component signal. Through the Hilbert envelope spectrum analysis of principal component and comparing with the Hilbert envelope spectrum of the original signal, the analysis effect of EEMD related principal component was validated. The study of the measured signal shows that af-ter the continuous signals of a hydraulic cylinder by EEMD decomposition, with the original signal correlation between the larger of the five order IMF component：IMF1, IMF2, IMF5, IMF6, IMF7 were contained in the original signal components. They were able to represent the major components of the decomposition of EEMD signal, and the first five orders were not able to fully represent the main component of the original signal. It was found by the hydraulic cylinder impact signal analysis that the former two order IMF： IMF1、IMF2 components of the EEMD decomposition had larger correlation with the original signal. The frequency components and the infor-mation contained in the original signal were able to be represented to indicate the principal component decomposition of EEMD signal and also to represent the main components of EEMD decomposition. Therefore, EEMD related principal component analysis can high-light the main frequency components of EEMD decomposition, which has good analysis effect to the vibration signals of the hydraulic system.

关 键 词：EEMD 相关系数 包络分析 主分量 

分 类 号：TH17[机械工程—机械制造及自动化]