一种有效的振动信号分离算法

振动信号包含随机成分与周期成分 ,其中的随机成分占据了很宽的频域 ,其能量极易引起其它设备特别是电子设备的共振 ,从而导致设备失效。为评估随机成分对设备的影响 ,需要将周期成分从振动信号中去除。为此 ,首先使用ARMA建模法对振动信号中周期成分出现的频率位置进行估计 ,然后设计阻带带宽极窄 (小于 0 .5Hz)的V型滤波器来完成相应的滤波。实验表明本文中的方法能有效地去除振动信号中的周期成分 ,而完整地保留随机成分

An effective decompose algorithm for vibration signal

The vibration signal is composed of stochastic component and periodic components.The random component has a wide range in the frequency domain,and its energy always evokes resonance of other apparatus(especially the electronic apparatus)and thus causes the failure of apparatus.To evaluate the influences of the stochastic component on apparatus,it is necessary to remove the periodic components from the vibration signal.In this thesis,firstly,the ARMA model building is employed to estimate the frequencies of the periodic interference,then,the digital notch filters,whose block bands are narrower than 1 Hz,are designed according to the periodic frequencies to filter the periodic components out.Experiments show the method presented is effective to eliminate the periodic interference components in the vibration signal.