一种基于压缩感知和小波信息熵的滚动轴承特征提取方法

为降低滚动轴承在线监测和故障诊断过程中振动信号采集、传输、存储和处理负担，基于压缩感知理论和小波包分析技术，提出一种基于压缩感知和小波信息熵的滚动轴承特征提取方法，用于滚动轴承故障诊断。应用部分哈达玛矩阵采集振动信号实现压缩，通过小波包分解提取滚动轴承状态特征，计算其小波信息熵作为故障诊断特征。在标准数据集上进行振动信号特征提取，并采用四种分类方法完成故障诊断实验，结果表明本文特征提取方法能够在较高的数据压缩率条件下，保持较高的故障诊断精度，适用于滚动轴承在线监测和故障诊断。

A Rolling Bearing Feature Extraction Method Based on Compressed Sensing

In order to reduce the burden of vibration signal acquisition, transmission, storage and processing in the process of online monitoring and fault diagnosis of rolling bearings. Based on the theory of compressed sensing and wavelet packet analysis technology, a feature extraction method of rolling bearings based on compressed sensing and wavelet information entropy is proposed for fault diagnosis of rolling bearings. PartHadamard matrix is used to collect vibration signals to achieve compression, and the state features of rolling bearings are extracted through wavelet packet decomposition, and the wavelet information entropy is calculated as the fault diagnosis feature. The vibration signal feature extraction is performed on the standard data set, and four classification methods are used to complete the fault diagnosis experiment. The results show that the feature extraction method in this paper can maintain a high fault diagnosis accuracy under the condition of a high data compression rate. It is suitable for online monitoring and fault diagnosis of rolling bearings.