基于精细复合多尺度熵特征向量相关系数在滚动轴承故障诊断中应用

Rolling Bearing Fault Diagnosis Method based on Refined Composite Multi-scale Entropy Eigenvector Correlation Coefficients

复合多尺度熵(CMSE)是在多尺度熵(MSE)基础上提出来的,它改善了MSE存在的熵值不精确、波动较大等,但不能解决样本时间序列太短引起未定义熵问题。精细复合多尺度熵(Refined Composite Multi-scale Entropy,RCMSE)通过改进算法使熵估计的准确性得到提高,并能降低诱导未定义熵的概率。以此为基础,提出基于RCMSE特征向量关系数的轴承故障识别分类方法。该方法首先利用RCMSE对数据样本生成多尺度熵,计算测试样本与已知故障状态的训练样本的RCMSE相关系数,从而判断测试样本的状态类型。对轴承信号数据进行试验表明,该方法能100%准确的对轴承正常,内圈,外圈和滚动体故障信号识别分类。因此,该方法是一种有效的识别故障特征,可为实际轴承故障诊断提供参考。

The composite multi-scale entropy(CMSE)was proposed based on multi-scale entropy(MSE).Although the CMSE overcomes the disadvantages of the inaccuracy and large volatility of MSE method,it cannot solve the problem of undefined entropy caused by too short sample time series.Refined Composite Multi-scale Entropy(RCMSE)can improve the accuracy of entropy estimation by improving the algorithm,and can reduce the probability of inducing undefined entropy.On this basis,this paper presents a method for bearing fault diagnosis based on RCMSE eigenvector correlation.First of all,the RCMSE is used to generate multi-scale entropy for the data,and the RCMSE correlation coefficient between the testing sample and the training sample is calculated to judge the state type of the sample.Experiments on the bearing signal data show that this method can identify and classify the normal operation signals of bearings and fault signals of the inner ring,outer ring and rolling elements with a 100%accuracy.Therefore,the method can effectively identify the fault characteristics of bearings.This study provides a reference for the actual bearing fault diagnoses.