摘要
提出了基于SVR谱和Gabor变换的故障诊断方法。首先采用扫描方式检测SVR谱最大峰值，并提取信号主周期分量，然后在Gabor变换域中，保留主周期分量处展开系数，其余置零，再根据处理后的展开系数重构信号以实现降噪，从而可提取故障特征频率。模拟信号验证了该方法的有效性，为机械系统中振动信号的分析与诊断提供了新方法。
A signal processing algorithm was proposed here for vibration signal analysis in fault diagnosis. The technique is based on Singular Value Ratio Spectrum and Gabor transform. The principal periodic component in the vibration signal was reconstructed automatically with sweep frequency method based on finding the peak value of singular value ratio （SVR） spectrum. The Gabor transform is applied to the signal, and then the Gabor expansion coefficients are processed by the principal periodic component. Then the de-noised signal could be obtained by the Gabor inverse transform. The fault frequency was experimentally evaluated by using vibration signals. The experi- mental results showed that this method is proved valid and efficient, and provides a new research direction for the diagnosis.
出处
《四川工程职业技术学院学报》
2013年第4期42-44,共3页
Journal of Sichuan Engineering Technical College
关键词
SVR谱
GABOR变换
故障诊断
singular value ratio spectrum
Gabor transform
fault diagnosis
