摘要
齿轮振动信号往往受到噪声干扰,用传统的小波阈值去噪法去噪时,由于信号中存在不连续点,会在不连续点处引起Pseudo-G ibbs现象。采用平移不变量小波阈值去噪法对信号进行去噪,首先对含噪声信号进行循环平移,然后对信号进行小波阈值去噪,最后对信号进行反平移后平均,得到去噪后的信号。应用这种方法对仿真信号和实验采集的含噪声齿轮故障信号进行去噪,并和小波阈值去噪效果进行对比,结果表明这种去噪方法有效抑制了Pseudo-G ibbs现象,并有更好的去噪效果。
Gear vibration signal is disturbed by noise very often. In the vicinity of discontinuities of signals, traditional wavelet threshold de-noising method may exhibit pseudo-Gibbs phenomenon. In this paper, a translation invariant wavelet threshold de-noising method was presented to eliminate the noises in the signal＇s. Firstly, the signal with noise was translated cyclically. Then the noises in the signal were eliminated with wavelet threshold de-noising. Finally, the signal was translated back and averaged. The simulation signals and gear fault signals collected in the experiment were processed by translation invariant wavelet threshold de-noising. And the denoising effect was contrasted with that of traditional wavelet threshold denoising method. Results show that this method can suppress pseudo-Gibbs phenomenon effectively.
出处
《噪声与振动控制》
CSCD
北大核心
2008年第6期17-19,27,共4页
Noise and Vibration Control
基金
国家自然科学基金资助项目(项目编号:50677017)
