论文部分内容阅读
为了有效去除机械故障信号中含有的噪声并提取故障特征,在小波阈值降噪的基础上,提出了一种新型改进阈值函数的第二代小波降噪方法。该方法是利用第二代小波对信号进行分解,采用新型改进阈值函数对分解信号的小波系数做阈值处理,同时引入基于类可分离性测度的降噪评价准则,实现对复杂振动信号降噪效果的评价。将所提出的方法应用于仿真信号和实测信号的分析与评价,结果表明:该方法融合了第二代小波和改进阈值函数的优点,能更好地消除噪声。
In order to effectively remove the noise contained in the mechanical fault signal and extract the fault features, a new wavelet denoising method with improved thresholding function is proposed based on wavelet threshold denoising. The method uses the second generation wavelet to decompose the signal, and uses the new improved threshold function to thresholding the wavelet coefficient of the decomposed signal. At the same time, the criterion of noise reduction based on class separability measure is introduced to realize the noise reduction effect of complex vibration signal evaluation of. The proposed method is applied to the analysis and evaluation of the simulation signal and the measured signal. The results show that this method combines the advantages of the second generation wavelet and the improved threshold function, and can better eliminate the noise.