基于瞬时能量熵和SVM的滚动轴承故障诊断  

Fault diagnosis approach for roller bearing based on EMD momentary energy entropy and SVM

作　　者：姚亚夫[1] 张星[1] 

机构地区：[1]中南大学机电工程学院,长沙410083

出　　处：《电子测量与仪器学报》2013年第10期957-962,共6页Journal of Electronic Measurement and Instrumentation

摘　　要：针对滚动轴承故障振动信号的非平稳特性和难以获得大量实际故障样本的情况,提出了一种基于经验模式(EMD)分解的新型故障特征撮方法,并与支持向量机(SVM)相结合实现滚动轴承的故障诊断。该方法首先将振动信号进行小波包降噪,再对去噪信号进行EMD分解,求解分解后各单元的瞬时能量变化,取瞬时能量变化的熵值组成特征向量,最后将其作为支持向量机的输入实现滚动轴承故障分类。经过实验验证,该方法能够有效的识别轴承正常状态、内圈故障、外圈故障以及滚珠故障。According to the non-stationary characteristic of the vibration signals from rolling bearing and the situa- tion is hard to obtain enough the fault samples , a new feature extraction method based on empirical mode decompo- sition（EMD） is proposed , then combine with support vector machine （SVM） to achieve the fault diagnosis. First , the vibration signal will be de-nosing by the Wavelet packet , decompose the de-nosing signal used EMD method , calculate the IMF＇ s momentary energy and the energy entropy , then combine every energy entropy to get the feature vector , finally keep the vector as the input of SVM , the classification result will be given by the SVM. The experi- mental verification show that the method can identify the four kinds of failures （norm, inner, outer, ball） effectively.

关 键 词：轴承 经验模式分解 瞬时能量熵 支持向量机 故障诊断 

分 类 号：TH133[机械工程—机械制造及自动化]