基于改进多尺度熵与VPMCD的滚动轴承故障诊断

Fault Diagnosis of Rolling Bearings Based on Improved Multiscale Entropy and VPMCD

多尺度熵(Multiscale entropy,MSE)是一种衡量时间序列复杂性的方法,针对其粗粒化过程由时间序列长度变短而导致熵值不精确、波动较大等问题,提出一种改进的多尺度熵(Improved multiscale entropy,IMSE)算法。在此基础上,结合迭代拉普拉斯得分(Iteration Laplacian Score,ILS)特征选择和多变量预测模型(Variable predictive model based class discriminate,VPMCD),提出一种新的滚动轴承智能故障诊断方法。最后,将提出的方法应用于滚动轴承试验数据分析,并与现有方法进行对比。结果表明,提出的方法不仅能够有效地识别滚动状态和故障类型,而且其诊断效果优于现有方法。

Multiscale entropy(MSE)is an algorithm for measuring the complexity of time series.However,in the coarse graining process of MSE,the shortened time series will lead to an imprecise and large fluctuation of the entropy estimation.In this paper,an improved multiscale entropy(IMSE)algorithm is proposed.On this basis,combining with the iterative Laplacian score(ILS)feature selection with variable predictive model based class discrimination(VPMCD),a novel fault diagnosis method of rolling bearings is proposed.Finally,the proposed fault diagnosis method is applied to analyze the experiment data of rolling bearings.Its results are compared with the existing methods.The results indicated that the proposed method can effectively identify the rolling state and fault types of the rolling bearings,and has much better diagnosis effect than the existing methods.