时域压缩特征提取及压缩感知在设备状态评估中的应用研究

Time-domain Compression Feature Extraction and Application Study of Compressed Sensing in Equipment Status Assessment

压缩感知是一种新的信号采集与处理框架,其框架中压缩采样过程能够直接获取"压缩"的采样数据。本文中研究了如何利用这些压缩数据提取特征并用于设备的状态评估。首先在压缩感知框架下研究压缩采样数据的特点,研究压缩数据的压缩性与信号的稀疏性的对应关系;接着提出一种时域压缩特征计算方法,用于提取压缩数据的特征信息;最后以滚动轴承为对象,使用时域压缩特征对滚动轴承的运行状态进行评估。使用滚动轴承全寿命周期数据进行实验分析,实验结果表明,时域压缩特征能够准确的判断轴承的运行状态。

Compressed sensing is a new theory of signal acquisition and processing. Base on this theory,compressed data is acquired in the process of compression sampling. In this paper,it is studied that which features is extracted from the compressed data and how these features used for equipment state estimation. Firstly,compressed data is analyzed and the study tried to find the corresponding relationship between compression data and signal sparsity. A time-domain compression characteristics is proposed,and used to extracting characteristic information hidden in the compressed data. Rolling bearing is chosen and time-domain characteristic is used to estimate the operational conditions of the rolling bearing. The method is used to analyze the whole life data of rolling bearing. The results show that the time domain compression feature can accurately estimate the running state of rolling bearings,demonstrate the effectiveness of the proposed method.