虚部噪声辅助LCD方法及其在遥测振动信号处理中的应用  

An image noise assisted Local Characteristic Scale Decomposition method and its application in telemetry vibration signal processing

作　　者：刘学[1] LIU Xue(9 4 Units No. 91550 Army, Dalian 116023, China)

机构地区：[1]中国人民解放军91550部队94分队

出　　处：《振动与冲击》2017年第12期1-6,17,共7页Journal of Vibration and Shock

摘　　要：针对局部特征尺度分解方法(Local Characteristic Scale Decomposition,LCD)的模态混叠以及集成局部特征尺度分解法(Ensemble LCD,ELCD)在集成平均时容易引入新的模态混叠、伪分量和运算量大等问题。提出一种基于虚部噪声辅助局部特征尺度分解法(Image Noise Assisted LCD,INALCD);首先以原信号为实部添加虚部白噪声构成复数信号,然后对复数信号在指定方向上进行投影,求取对称投影象限的基函数,通过投影后虚部白噪声均匀化原信号投影的极值点的分布,辅助信号分解过程中极值点的选取,抑制模态混叠,最后将对称投影象限的基函数进行线性组合消除噪声的影响,避免了ELCD因集成平均带来的相关问题。仿真和实测数据实验结果表明,该方法在降低模态混叠的同时,大大减少了计算量,性能优于LCD和ELCD方法。In order to alleviate mode mixing in the Local Characteristic Scale Decomposition method （LCD） , as well as solve the problem of the ensemble average always resulting in new mode mixing, illusive component, and computational cost increasing in the ensemble LCD （ELCD） method, an image noise assisted LCD method was proposed, First, a complex signal was formed by treating the original signal as the real part and adding white noise as the imaginary part. Then the complex signal was projected in the specified direction to strike the base functions of symmetrical projection quadrant. Through the projection of the imaginary part of white noise, the distribution of original signal extreme points were uniformed, the selection of extreme point was assisted in the decomposition. Finally, the quadrant projection was symmetrical linear combination of base functions to eliminate the effects of noise,which could avoid the problems associated by ELCD due to ensemble average. Experimental results show that the method can reduce mode mixing, while greatly reduces the amount of calculation. Its performance is superior to LCD and ELCD.

关 键 词：遥测振动信号 局部特征尺度分解方法 集成平均 模态混叠 投影 虚部噪声 

分 类 号：TN911.7[电子电信—通信与信息系统]