|
|
|
| Blind Source Separation of Underdetermined Signals Based on Extremum Field Mean Mode Decomposition |
| MENG Zong1,2, LIANG Zhi1,2, ZONG Zhen-wei3, HUI Shao-nan3 |
1. Institute of Electrical Engineering, Yanshan University, Qinhuangdao, Heibei 066004, China;
2. Key Laboratory of Measurement Technology and Instrumentation of Hebei Province, Yanshan University, Qinhuangdao,Hebei 066004, China;
3. Qinhuangdao Institute of Measurement and Testing, Qinhuangdao, Hebei 066004, China |
|
|
|
|
Abstract The ordinary blind source separation (BSS) methods is based on the assumption, which the number of observation signals is no less than that of the source signals. The result of BSS will be relatively poor when the number of observation signals less than source signals. However, the problem of underdetermined BSS, even single observation channel BSS, is common in vibration signals of rotating machinery.To solve the single observation channel problem, a new BSS method based on extremum field mean mode decomposition (EMMD) is proposed.Firstly, by EMMD, the underdetermined observation signal is decomposed to a series of intrinsic mode function (IMF), then the underdetermined observation signal and IMFs compose multi-dimensional signal, to increase the dimensions of observation signals.Secondly, the number of source signals is estimated with singular value decomposition and Bayesian information criterion.〖JP2〗Finally, the characteristic matrix joint diagonalization method based on fourth-order cumulant is used to achieve BSS.The simulation study on rotating machinery fault signal indicates that it can well solve the problem of BSS with underdetermined observation signal.
|
|
|
|
|
|
［1］ Huang N E, Shen Z, Long S R, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and nonstationary time series analysis［J］. Proc R Soc Lond A, 1998, 454（1971）: 903-995.
［2］ 谢平, 王欢, 杜义浩. 基于EMD和Wigner-Ville分布的机械故障诊断方法研究［J］. 计量学报, 2010, 31(5):390-394.
［3］ 孟宗, 顾海燕, 刘利晖, 等. 基于EMD与AR谱的轧机主传动系统故障诊断研究［J］.计量学报, 2011, 32(4):338-342.
［4］ 盖强. 局域波时频分析方法的理论研究与应用［D］.大连：大连理工大学, 2001.
［5］ 焦卫东, 杨世锡, 吴昭同.基于盲源分离的旋转机械干扰消除技术研究［J］.仪器仪表学报, 2004, 25(3):368-371.
［6］ 叶红仙, 杨世锡, 杨将新.基于EMD-SVD-BIC的机械振动源数估计方法［J］.振动、测试与诊断, 2010, 30(3):330-334.
［7］ 申永军, 杨绍普, 孔德顺.一种基于奇异值分解的欠定盲信号分离方法［J］.振动与冲击, 2008, 27(S):157-159.
［8］ 李志农, 刘卫兵, 易小兵.基于局域均值分解的机械故障欠定盲源分离方法研究［J］.机械工程学报, 2011, 47(7):97-102.
［9］ Ypma A. Learning methods for machine vibration analysis and health monitoring［D］. Delft:Delft University of Technology, 2001. |
|
|
|
2013, Vol. 34 
