基于改进小波算法的轴承信号降噪研究

Denoising of roll bearing signal based on improved wavelet algorithm

针对传统阈值函数降噪在阈值处不连续和存在恒定偏差的问题,采用层层递进的改进方法对小波阈值和小波阈值函数做了去噪研究。小波阈值的改进是基于不同分解尺度小波阈值算法,改进阈值考虑了轴承故障信号在不同分解层上的去噪特性和噪声在各分解层上的分布情况。小波阈值函数改进是基于对数逼近阈值函数算法,改进阈值函数考虑了传统软阈值函数的不连续性和硬阈值函数的偏差性。从传统阈值函数、改进阈值函数和传统阈值、改进阈值选取上看,传统阈值函数结合改进阈值去噪效果最好,其次是传统阈值函数结合改进阈值,最后是传统阈值函数结合传统阈值。在小波去噪算法中,改进小波阈值法可获得较好的去噪效果。

According to the poor de-noising effect of the traditional wavelet method in roll bearing signal,improved wavelet method and improved wavelet threshold function are proposed.The improved threshold takes account of the de-noising characteristics of roll bearing signals at different decomposition levels and each decomposition layer noise distribution.The improved threshold function takes into account of the discontinuity of the traditional soft threshold function and the deviation of the hard threshold function.From the traditional threshold function,improving threshold function,further improving threshold function and traditional threshold and improving threshold selection,the traditional threshold function combined with improved threshold de?noising effect is best,the second is to further improve threshold function and improve threshold,then it is the traditional threshold function combined with improved threshold.Finally,the traditional threshold function is combined with the traditional threshold.Experimental simulation results in the wavelet de?noising improved algorithm(wavelet improved threshold,wavelet improved threshold function),the improved wavelet threshold method can obtain better denoising effect.