基于MFE与改进层次原型的轴承故障诊断方法

The diagnosis method of bearing fault based on MFE and improved hierarchical prototype

针对滚动轴承故障特征难以提取和故障特征分类困难的问题,提出了一种基于多尺度模糊熵(MFE)与改进层次原型分类器(IHP)的故障诊断方法。首先,利用多尺度模糊熵从滚动轴承不同状态下的振动信号中提取20种故障特征。其次,引入线性判别分析(LDA)对Hierarchical Prototype进行改进,从而提高故障分类精度。最后,结合多尺度模糊熵与改进层次原型分类器对故障特征进行分类。实验证明,提出的MFE与IHP能有效提取滚动轴承的故障特征,并实现高精度分类。相比于其他故障识别分类器,所提方法有更高的识别精度,分类精度达到了99.29%。

Aiming at the problem of feature extraction and accurate classification of fault features,a fault diagnosis method based on multi-scale fuzzy entropy(MFE)and improved hierarchical prototype classifier(IHP)is proposed.Firstly,the MFE extracts fault features with 20 scales from the vibration signals in different states of the bearing.Linear discriminant analysis is introduced to improve hierarchical prototype,which is called improved hierarchical prototype classifier(IHP).Then,the extracted fault features are classified with IHP.Experiments show that the proposed MFE-IHP algorithm can effectively extract the characteristics of the bearing vibration signal and achieve high-precision classification.Compared with other fault recognition classifiers,this method has higher recognition accuracy,and the classification accuracy reaches 99.29%.