基于信号空间压缩感知算法的机械故障诊断  

Mechanical fault diagnosis based on a signal space compressed sensing algorithm

作　　者：杜佳兵 唐刚[1] 王华庆[1] DU JiaBing TANG Gang WANG HuaQing(College of Mechanical and Electrical Engineering, Beijing University of Chemical Technology, Beijing 100029, China)

机构地区：[1]北京化工大学机电工程学院,北京100029

出　　处：《北京化工大学学报（自然科学版）》2017年第5期85-89,共5页Journal of Beijing University of Chemical Technology(Natural Science Edition)

基　　金：国家自然科学基金(51405012)

摘　　要：为了解决压缩感知(CS)重构算法通过重构稀疏系数求解原始信号的重构精度不高的问题,提出一种基于信号空间的压缩采样匹配追踪算法。首先在冗余字典中求解原始信号的最优表示空间,然后在最优表示空间中利用迭代算法直接求解原始信号,最后以轴承故障振动信号为例进行实验验证。结果证明本文算法提高了信号的重构精度,可以为增强机械振动信号的故障检测能力提供依据。Compressive sensing （CS） theory makes use of sparse characteristics that allow a signal with a low sam- piing rate to be reconstructed. However, because compressive sensing algorithms solve the original signal by recon- structing the sparse coefficients, the precision of the reconstructed signal is difficult to guarantee. In order to solve this problem, this paper presents a new signal space compressive sensing algorithm. The algorithm solves the opti- mal representation space of the primary signal and then reconstructs the original signal directly by means of an itera- tive algorithm. The optimal representation space is composed of vectors which are optimal linear representations of the original signal. It has been verified by experiment that the method improves the precision of the signal and en- hances the fault detection capability.

关 键 词：压缩感知(CS) 稀疏性 信号空间 故障检测 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置]