EEMD结合能量特征和小波降噪的轴承故障诊断  

Fault diagnosis of bearings by using EEMD combined with energy feature and wavelet de-noising

作　　者：边杰[1,2] 王平[1,2] 梅庆[1,2] 

机构地区：[1]中航工业航空动力机械研究所,湖南株洲412002 [2]航空发动机振动技术航空科技重点实验室,湖南株洲412002

出　　处：《广西大学学报（自然科学版）》2014年第6期1206-1211,共6页Journal of Guangxi University（Natural Science Edition）

基　　金：中国航空工业集团公司技术创新基金项目(2012B60804R);中国航空工业集团公司航空科学基金项目(08D08002;2014ZD08007)

摘　　要：为解决在强背景噪声条件下滚动轴承故障诊断问题,开展基于能量特征和小波降噪的总体经验模态分解(EEMD)研究。首先以仿真信号为研究对象,对其进行总体经验模态分解,得到9个固有模态函数(IMF)和1个余项(Res),然后考虑各模态函数的能量特征,将分解后的9个IMF分量与原始信号的能量比作为判断标准,剔除附加5个低频分量,最终得到4个有效的IMF分量和1个余项,与仿真信号相符。在仿真信号分析的基础上,对含噪声信号的滚动轴承故障信号进行故障诊断试验研究,采集信号经小波降噪后,利用总体平均经验模态分解并结合能量特征,得到3个IMF分量和1个余项,然后对3个IMF分量进行包络谱分析,提取故障特征频率157.5 Hz,与滚动轴承故障内圈特征频率157.9 Hz相比,误差为0.25%,说明该方法能很好地提取含有噪声信号的轴承故障信息。该研究为强背景噪声下滚动轴承故障信息的提取提供了一种有效的方法。To solve the problem of fault diagnosis of rolling bearing under heavy background noise, an ensemble empirical mode decomposition （ EEMD ） method combined with energy feature and wavelet de-noising was studied. Taking a simulated signal as the research object, ensemble empiri-cal mode decomposition was made to obtain nine intrinsic mode functions （ IMFs） and a residue. Then the energy ratio of the nine IMF components after decomposition and the original signal was used as the standard of judgment by considering energy features of the IMFs. Finally four effective IMF components and a residue were obtained by eliminating the other five redundant low-frequency components, which is consistent with the simulated signal. According to this analysis strategy of sim-ulation signal, a test of fault diagnosis of rolling bearing was taken. After wavelet de-noising the col-lected signals and utilizing the EEMD method combed with energy feature, three IMF components and a residue were obtained. Then, envelope spectrum analysis of the three IMF components was done to extract the fault feature frequency of 157. 5 Hz. Compared with inner race characteristic fre-quency of 157. 9 Hz, the calculating error is 0. 25%, it indicatesthat bearing fault feature can be well extracted from the collected signal with heavy noise. The study proposes an effective method to ex-tract fault features of rolling bearing under the background of heavy noise.

关 键 词：滚动轴承 总体平均经验模态分解 能量特征 小波降噪 故障诊断 

分 类 号：TH133.3[机械工程—机械制造及自动化] TN911.7[电子电信—通信与信息系统]