摘要
针对轴承振动的非平稳性特点和频谱成分的混杂性,提出了基于小波的信号自适应阈值降噪法。自适应阈值降噪法首先对信号进行离散正交小波多层分解,对分解后的各层细节系数中模小于某阈值的系数进行处理,然后将处理完的小波系数再进行反变换,重构出经过降噪后的信号。用仿真信号进行降噪处理,结果表明:通过选择合适的小波基和阈值选择规则,可以实现信号的完美降噪;实测轴承振动信号用小波降噪方法进行预处理,提高了信噪比,进一步作频谱分析得到了故障特征信息,为诊断决策提供了依据。
For the nonstationarity and spectrum chaos of the bearing vibration signal, threshold denoising based on wavelet decomposition was put forward. In this method, signal was decomposed into multi-layer, processing the detail coefficients according as the threshold, then reconstructing to get the denoised signal by the wavelet coefficients. The simulated signal was denoised, the result demonstrated that fine denoising could be carried out through selecting suitable wavelet and threshold ruler. The denoising method was employed to preprocess the real vibration signal of bearing, improving the signal noise rate. Fault characteristic was gained by the following frequency analysis, it approved foundation of diagnosis decision-making.
出处
《噪声与振动控制》
CSCD
北大核心
2007年第5期100-103,106,共5页
Noise and Vibration Control
关键词
振动与波
非平稳性
小波分解
阈值选择
信号降噪
特征提取
vibration and wave
nonstationarity
wavelet decomposition
threshold selection
signal denoising
characteristic extraction
