基于第二代小波变换的振动信号去噪与故障诊断  

Denoising Method and Fault Diagnosis of Vibrating Signal Based on Second Generation Wavelet Transform

作　　者：唐贵基[1] 王誉蓉[1] 郭新富 胡爱军[1] 范德功[1] 

机构地区：[1]华北电力大学机械工程学院,保定071003 [2]中石油宁夏销售分公司,银川750001

出　　处：《汽轮机技术》2006年第4期295-297,共3页Turbine Technology

基　　金：华北电力大学博士学位教师基金项目(项目编号:92104392)。

摘　　要：旋转机械故障振动信号存在不同形式的波形特征,传统小波去噪中,小波分解的结果与所采用的小波基函数有关,选用不适当的小波基函数会冲淡振动信号的局部特征信息,而造成原始信号的部分有用的细节信息丢失。为了克服上述缺陷,提出一种基于第二代小波变换的振动信号预处理方法,即针对分析信号的局部特征,以预测方差最小为目标,对每个样本选择最佳的预测算子,使小波基函数始终能够匹配信号的局部特征。仿真试验表明,该方法克服了传统小波去噪中降噪信号丢失了部分细节信息的缺点,不仅可有效地去除故障诊断振动信号的噪声,而且能够保留信号的局部信息。There is different characteristics of signal in the fault vibrating signal of rotating machinery. In the denoising of traditional wavelet tansform, the result of wavelet decomposing is related with wavelet basis function. And the adopted wavelet which is ill - suilted for original signal can delute the local characteristic of vibrating signal , and lost some useful detail information of original signal. In order to overcome mentioned limitation, a pre - processing method based on second generation wavelet transform for vibrating signal is adopted. According to the local characteristics of signal and the selection criterion of minimizing the squared error ,an optimal predicting operator is selected for a transforming sample so that the used wavelet basis function can always fit the local characteristics of the signal. The simulation showed that the proposed method could overcome the denoising disadvantage of traditional wavelet transform. It not only can filter noise from original signal effectively, but also can hold local characteristics of original signal for fault diagnosis in the denoised signals.

关 键 词：第二代小波变换 预处理 小波基函数 预测算子 

分 类 号：TH17[机械工程—机械制造及自动化]