Abstract: In order to overcome the shortcomings of MSE and extract the nonlinear fault features of rolling bearings more accurately, a new nonlinear dynamic analysis method called RCMDE was introduced to the fault feature extraction of rolling bearings. Based on the feature extraction method via RCMDE and support vector machine (SVM), a new method of rolling bearing fault diagnosis was proposed. Through the analysis of rolling bearing experimental data, the proposed method was compared with the multiscale dispersion entropy and MSE based fault diagnosis methods and the results show that the fault diagnosis method may accurately identify the fault locations and severities of rolling bearings, and has a higher fault recognition rate than that of the two methods mentioned.

Key words: dispersion entropy, multiscale sample entropy(MSE), refined composite multiscale dispersion entropy(RCMDE), rolling bearing, fault diagnosis

摘要： 为克服多尺度样本熵的不足，更精确地提取滚动轴承非线性故障特征，将一种新的非线性动力学分析方法——精细复合多尺度散布熵引入到滚动轴承的故障特征提取。在此基础上，提出了一种基于精细复合多尺度散布熵与支持向量机的滚动轴承故障诊断新方法。通过滚动轴承实验数据分析，将所提方法与基于多尺度样本熵和多尺度散布熵的故障诊断方法进行了对比，结果表明：所提方法不仅能精确地识别滚动轴承故障类型和故障程度，而且故障识别率高于另两种方法。

关键词: 散布熵, 多尺度样本熵, 精细复合多尺度散布熵, 滚动轴承, 故障诊断