中文摘要：

针对传统阈值处理函数在小波降噪中的不足,在硬阈值、软阈值处理函数的基础上,构造了一种改进的阈值处理函数,并介绍了该函数的降噪原理。基于MATLAB平台,用改进阈值处理函数分别对三种不同类型的仿真信号进行了降噪处理,通过信噪比、均方误差2个衡量指标,研究不同动态参数对三种类型信号降噪效果的影响;并根据实测转子系统振动信号的特点,选择适当的动态参数,将该函数应用到振动信号的降噪中。结果表明：改进阈值函数能够根据信号的不同类型,调节动态参数,优化阈值的处理,能很好地实现转子系统振动信号的降噪。

英文摘要：

Aiming at the deficiency of the traditional threshold processing function in wavelet denoising,a new improved function was proposed based on soft threshold and hard threshold function and the principle of the function was introduced.Firstly,three types of simulated signals were denoised by the improved threshold function in MATLAB,and then,the effect of denoising was compared by the ratio of signal to noise and mean square error when different dynamic parameters were selected. Finally,the improved function with appropriate dynamic parameters was used to denoise based on the characteristic of rotor vibration signal.The results show that based on the difference of signals,the improved threshold function can optimize the processing of threshold through adjusting the dynamic parameters and realize the denoising of the vibration signal of rotor system efficiently.