摘要
针对轴承振动信号具有非线性、非平稳性以及故障特征提取困难鸽问题,提出了一种基于EEMD-SVD与支持向量机的轴承故障诊断方法。首先,利用集成经验模态分解方法将轴承振动信号自适应地分解为多个本征模态函数分量。然后,根据哨度准则选取6个本征模态函数分量,并将其构成的矩阵进行奇异值分解得到特征向量。最后,将特征向量输入支持向量机进行故障诊断。利用凯斯西储大学的轴承数据进行了试验,并与BP神经网络进行了对比,结果验证了本文方法的有效性。
The vibration signal of bearings is nonlinear and non一stationary and it is difficult to extract fault features from the vibration signal of bearings.In order to address the problems aforementioned,a fault diagnosis methodology based on ensemble empirical mode decomposition(EEMD),singularity value decomposition(SVD)and support vector machine(SVM)is proposed.Firstly,with the EEMD method,bearing vibration signals are adaptively decomposed into a number of intrinsic mode functions(IMF).Secondly,six IMF components are selected according to the criterion of kurtosis,and singular value sequences regarded as eigenvectors are obtained with the method of SVD.Lastly,eigenvectors serve as import of SVM so that faults of the bearing are recognized.Experiments are conducted with the bearing datasets of the Case Western Reserve University and comparison is carried out with BP neural network,which fully verifies the effectiveness of the proposed method.
作者
李东
刘广璞
黄晋英
张安安
LI Dong;LIU Guangpu;HUANG Jinying;ZHANG Anan(School of Mechanical Engineering,North University of China,Taiyuan 030051,China)
出处
《机械设计与研究》
CSCD
北大核心
2019年第6期123-127,共5页
Machine Design And Research
关键词
轴承
集成经验模态分解
奇异值分解
支持向量机
故障诊断
bearing
ensemble empirical mode decomposition
singularity value decomposition
support vector machine
fault diagnosis