基于变分模态分解谱熵的电机轴承退化状态识别方法

Degradation Status Identification Method of Motor Bearing Based on Variational Modal Decomposition Spectral Entropy

为更好的表征电机轴承的退化状态,对电机轴承退化特征提取方法进行了研究。结合变分模态分解(Variational modal decomposition, VMD)和信息熵理论,提出了基于VMD分解谱熵的退化状态识别方法。对不同损伤程度的轴承振动信号进行VMD分解,分别计算其在不同尺度下的复杂度度量能谱熵、奇异谱熵和边际谱熵,以其作为退化特征向量。通过建立相关向量机退化状态识别模型实现轴承的退化状态识别。仿真信号和轴承实测信号均验证了VMD分解谱熵对轴承退化状态的表征能力。

In order to characterize the degradation state of motor bearing,the degradation feature extraction method are studied. Combining variational modal decomposition (VMD) and information entropy,a degradation state identification method based on VMD decomposition spectral entropy was proposed. The VMD decomposition was performed for bearing vibration signals with different damage levels,and its complexity indicators of power spectral entropy,singular spectral entropy and marginal spectral entropy in different scale were computed. The three indicators were taken as degradation feature vector. Then,the degradation state identification model is established based on relevance vector machine to allowing the degradation state of bearing to be identified. The analysis results of simulation signal and practical bearing vibration signal demonstrate the ability of the proposed method.