摘要
考虑电力电子装置中噪声因素对故障检测和识别的影响,用多尺度小波变换和小波包变换对信号进行消噪。小波变换是将信号分解成高频和低频成分,低频成分包含信号的主要性能,高频含较多的噪声,将高频平滑后再重建即可消噪。然后利用Matlab对电力电子装置故障暂态信号进行消噪处理,比较各种消噪方法的优缺点,并结合理论分析和模拟结果讨论了阈值的选取方法和原则。仿真结果表明,基于小波变换的消噪方法是一种提取有用信号、展示噪声和突变信号的实用方法。
Considering the noise effect on the fault identification and detection in power electronics, multi-scale wavelet transform and wavelet packet transform are employed to eliminate the noise of signal. The signal was decomposed into high or low frequency elements by wavelet transform. The low frequency elements include the main performances of the signal. But the high frequency elements consist of mere noise, reconstructing the signal can be realized after flatting the high frequency elements. And Matlab is used to de-noise for transient signals of power electronics faults, comparing the effects and advantages of the methods, and discusses the principle and measure of threshold selection. Simulation results show that the methods based on wavelet transform is practicable for getting usefulness signal, setting out noise and break signal.
出处
《变流技术与电力牵引》
2007年第1期48-51,共4页
Converter Technology & Electric Traction
基金
湖南省自然科学基金(06FD007)
