点击次数：

论文类型：期刊论文

发表时间：2013-11-01

发表刊物：Journal of Information and Computational Science

收录刊物：EI、Scopus

卷号：10

期号：16

页面范围：5189-5198

ISSN号：15487741

摘要：In this paper, a new fault feature extraction method based on Intrinsic Mode Function (IMF) envelope sample entropy (SampEn) is proposed for rolling bearings fault diagnosis. First, the Empirical Mode Decomposition (EMD) method is utilized to decompose the vibration signals self-adaptively into a number of IMFs which represent different frequency bands from high to low. Second, the IMF envelope signals are used to highlight the fault-induced information in a structurally simpler and physically more meaningful way than the original signals. Thus, the shortcoming of SampEn assigning high values to uncorrelated random signals can be overcome. Finally, the IMF envelope SampEn serve as a fault feature vector to be input into multi-class classifier of Support Vector Machine (SVM) for identification of different bearing conditions. The experimental results indicate that the proposed approach based on IMF envelope SampEn can identify different fault types as well as levels of severity effectively and is superior to that based on IMF SampEn. ? 2013 Binary Information Press.