基于变分模态分解的单通道信号盲源分离方法
|
摘要
针对单通道情况下传统盲源分离方法难以恢复源信号的问题,提出一种基于变分模态分解(VMD)的单通道信号盲源分离方法。首先对单通道信号进行变分模态分解(VMD)获得一系列本征模态函数(IMF)分量,将单通道信号和其IMF分量构成多维信号,然后采用主成分分析法估计源数,依据估计的源信号数目重组多通道观测信号,最后利用改进的变步长等变自适应分离(VSEASI)算法实现信号的盲分离。将所提出方法应用于齿轮和轴承的单通道信号仿真研究,仿真结果表明,该方法能够有效地分离出齿轮和轴承信号,解决了单通道信号盲源分离问题。
Aiming at the difficulty of blind source separation(BSS) method cannot solve the problem of single channel blind source separation, a new algorithm based on variation mode decomposition(VMD) was proposed in this paper. First, the observed signal was decomposed into instrinsic mode function(IMF)components by variation mode decomposition(VMD),then the obstained IMFs and the original observed signal are combined to compose new multidimensional signals. Secondly, the application of principal component analysis could accurately estimate the number of source signals, multiple channel observation signals are reconstructed based on estimated source signals. Finally, the variable step equivariant adaptive separation via independence(VSEASI) algorithm is adopted to solve the BSS problem. This method is applied to the simulation research of bearing and gear in order to correctly separate their source signals. Simulation research indicates that it can well solve the difficult problem of single channel blind source separation.
Aiming at the difficulty of blind source separation(BSS) method cannot solve the problem of single channel blind source separation, a new algorithm based on variation mode decomposition(VMD) was proposed in this paper. First, the observed signal was decomposed into instrinsic mode function(IMF)components by variation mode decomposition(VMD),then the obstained IMFs and the original observed signal are combined to compose new multidimensional signals. Secondly, the application of principal component analysis could accurately estimate the number of source signals, multiple channel observation signals are reconstructed based on estimated source signals. Finally, the variable step equivariant adaptive separation via independence(VSEASI) algorithm is adopted to solve the BSS problem. This method is applied to the simulation research of bearing and gear in order to correctly separate their source signals. Simulation research indicates that it can well solve the difficult problem of single channel blind source separation.
引文
[1] Belaid S, Hattay J, Naanaa W, Aguili T. A new multi-scale framework for convolutive blind source separation. Signal Image and Video Processing, 2016; 10(7):1203-1210.
[2] Domanov l, De Lathauwer L.Generic uniqueness of a structuredmatrix factorization and applications in blind source separation. IEEE Journal of Selected Topics In Signal Processing,2016;10(4):701-711.
[3] Haile M A, Dykas B. Blind source separation for vibration-based diagnostics of rotorcraft bearings. Journal of Vibration & Control, 2016; 22(18):3807-3820
[4] 毋文峰, 陈小虎, 苏勋家. 基于经验模式分解的单通道机械信号盲分离[J]. 机械工程学报, 2011, 47(4):12-16.
[5] 崔玲丽, 吴春光, 邬娜. 基于EMD与ICA的滚动轴承复合故障诊断[J]. 北京工业大学学报, 2014, 40(10):1459-1464.
[6] Flandrin P, Rilling G, Goncalves P. Empirical mode decomposition as a filter bank[J]. IEEE Signal Processing Letters, 2004, 11(2):112-114.
[7] Dragomiretskiy K, Zosso D. Variational Mode Decompositi-on[J]. IEEE Transactions on Signal Processing, 2014,62(3):531-544.
[8] 刘长良, 武英杰, 甄成刚. 基于变分模态分解和模糊C均值聚类的滚动轴承故障诊断[J]. 中国电机工程学报, 2015, 35(13):3358-3365.
[9] 朱永利, 贾亚飞, 王刘旺,等. 基于改进变分模态分解和Hilbert变换的变压器局部放电信号特征提取及分类[J]. 电工技术学报, 2017, 32(9):221-235.
[10] 陈东宁, 张运东, 姚成玉,等.基于变分模态分解和多尺度排列熵的故障诊断[J/OL]. 计算机集成制造系统,http://kns.cnki.net/kcms/detail/11.3619.TP.20170523.1019.002.html.
[11] 李振璧,王康,姜媛媛. 基于串音误差与分离度的变步长EASI盲源分离算法[J]. 计算机工程, 2017, 43(8): 108-113.
[12] 高丽, 张天骐, 侯瑞玲,等. 基于改进EASI算法的DS-CDMA信息码与伪码盲估计[J]. 电讯技术, 2012, 52(8):1275-1281.
[13] 陆建涛, 成玮, 訾艳阳,等. 变步长等变自适应盲源分离算法[J]. 西安交通大学学报, 2015, 49(12):83-89.
[14] 柳晓云, 马增强, 张俊甲,等. 基于变分模态分解与独立分量分析的轴承故障特征提取方法[J]. 济南大学学报(自然科学版), 2017, 31(3):261-267.
[15] Venkatanarayana K, Manoj V V R, K. Swathi K S. Enhanced Face Recognition based on PCA and SVM[J]. International Journal of Computer Applications, 2015, 117(2):40-42.
[16] 谷千伟,金炜东,余志斌. 基于EEMD和ICA的单通道列车信号盲分离[J]. 计算机应用研究, 2014, 31(5):1551-1553.
[17] 潘城, 田社平, 颜国正. 超定独立分量分析及其在结肠压力信号分析中的应用[J]. 上海交通大学学报, 2010, 44(11):1595-1599.
[18] Ypma A. Learning methods for machine vibration analysis and health monitoring[D]. Delft: Delft University of Technlolgy,2001.
[19] 孟宗, 梁智, 宗振威,等. 基于EMMD的欠定旋转机械振动信号盲源分离[J]. 计量学报, 2013, 34(4):305-310.
[20] 孟宗, 梁智. 基于EMMD和BSS的单通道旋转机械故障诊断方法[J]. 仪器仪表学报, 2013, 34(3):635-642.
[21] 马增强, 李亚超, 刘政,等. 基于变分模态分解和 Teager 能量算子的滚动轴承故障特征提取[J]. 振动与冲击, 2016, 35(13):134-139.
[2] Domanov l, De Lathauwer L.Generic uniqueness of a structuredmatrix factorization and applications in blind source separation. IEEE Journal of Selected Topics In Signal Processing,2016;10(4):701-711.
[3] Haile M A, Dykas B. Blind source separation for vibration-based diagnostics of rotorcraft bearings. Journal of Vibration & Control, 2016; 22(18):3807-3820
[4] 毋文峰, 陈小虎, 苏勋家. 基于经验模式分解的单通道机械信号盲分离[J]. 机械工程学报, 2011, 47(4):12-16.
[5] 崔玲丽, 吴春光, 邬娜. 基于EMD与ICA的滚动轴承复合故障诊断[J]. 北京工业大学学报, 2014, 40(10):1459-1464.
[6] Flandrin P, Rilling G, Goncalves P. Empirical mode decomposition as a filter bank[J]. IEEE Signal Processing Letters, 2004, 11(2):112-114.
[7] Dragomiretskiy K, Zosso D. Variational Mode Decompositi-on[J]. IEEE Transactions on Signal Processing, 2014,62(3):531-544.
[8] 刘长良, 武英杰, 甄成刚. 基于变分模态分解和模糊C均值聚类的滚动轴承故障诊断[J]. 中国电机工程学报, 2015, 35(13):3358-3365.
[9] 朱永利, 贾亚飞, 王刘旺,等. 基于改进变分模态分解和Hilbert变换的变压器局部放电信号特征提取及分类[J]. 电工技术学报, 2017, 32(9):221-235.
[10] 陈东宁, 张运东, 姚成玉,等.基于变分模态分解和多尺度排列熵的故障诊断[J/OL]. 计算机集成制造系统,http://kns.cnki.net/kcms/detail/11.3619.TP.20170523.1019.002.html.
[11] 李振璧,王康,姜媛媛. 基于串音误差与分离度的变步长EASI盲源分离算法[J]. 计算机工程, 2017, 43(8): 108-113.
[12] 高丽, 张天骐, 侯瑞玲,等. 基于改进EASI算法的DS-CDMA信息码与伪码盲估计[J]. 电讯技术, 2012, 52(8):1275-1281.
[13] 陆建涛, 成玮, 訾艳阳,等. 变步长等变自适应盲源分离算法[J]. 西安交通大学学报, 2015, 49(12):83-89.
[14] 柳晓云, 马增强, 张俊甲,等. 基于变分模态分解与独立分量分析的轴承故障特征提取方法[J]. 济南大学学报(自然科学版), 2017, 31(3):261-267.
[15] Venkatanarayana K, Manoj V V R, K. Swathi K S. Enhanced Face Recognition based on PCA and SVM[J]. International Journal of Computer Applications, 2015, 117(2):40-42.
[16] 谷千伟,金炜东,余志斌. 基于EEMD和ICA的单通道列车信号盲分离[J]. 计算机应用研究, 2014, 31(5):1551-1553.
[17] 潘城, 田社平, 颜国正. 超定独立分量分析及其在结肠压力信号分析中的应用[J]. 上海交通大学学报, 2010, 44(11):1595-1599.
[18] Ypma A. Learning methods for machine vibration analysis and health monitoring[D]. Delft: Delft University of Technlolgy,2001.
[19] 孟宗, 梁智, 宗振威,等. 基于EMMD的欠定旋转机械振动信号盲源分离[J]. 计量学报, 2013, 34(4):305-310.
[20] 孟宗, 梁智. 基于EMMD和BSS的单通道旋转机械故障诊断方法[J]. 仪器仪表学报, 2013, 34(3):635-642.
[21] 马增强, 李亚超, 刘政,等. 基于变分模态分解和 Teager 能量算子的滚动轴承故障特征提取[J]. 振动与冲击, 2016, 35(13):134-139.