基于变分模态分解近似熵和支持向量机的轴承故障诊断方法

Bearing Fault Diagnosis Method Based on VMD-AE and SVM

针对滚动轴承振动信号的非平稳特征,以及现实中难以获得大量典型故障样本的情况,提出了一种基于变分模态分解的近似熵和支持向量机的故障诊断方法。首先,通过VMD将原始振动信号分解为若干个频率尺度的本征模态分量;然后,计算各个IMF分量的近似熵并组成特征向量;最后,将上述特征向量输入支持向量机进行训练,并判断轴承的工作状态和故障类型。分析结果表明:与EMD及LMD相比,VMD近似熵与支持向量机相结合后,诊断精度得到了较大的提高,更适用于轴承故障的自动化诊断。

According to the non-stationary characteristics of rolling bearing vibration signals and the difficulty to ob-tain a large number of typical fault samples in reality,a fault diagnosis method is proposed,which uses the Variational Mode Decomposition （VMD）approximate entropy and the support vector machine （SVM）.Firstly,the original vibra-tion signals are decomposed into several Intrinsic Mode Functions （IMF）by VMD.Then,the approximate entropies of each IMF are calculated,which are taken as feature vector.Finally,input the feature vector for training support vector machine and determine the bearing status and fault type.The experimental results show that,compared with EMD and LMD,the VMD combined with approximate entropy and support vector machine is more suitable for the automatic diag-nosis of bearing fault and can greater improve diagnostic accuracy.