Please wait a minute...
浙江大学学报(工学版)
J4  2011, Vol. 45 Issue (8): 1376-1381    DOI: 10.3785/j.issn.1008-973X.2011.08.008
机械工程     
旋转机械振动信号的固有模式函数降噪方法
熊炘, 杨世锡, 周晓峰
浙江大学 机械工程学系,浙江 杭州 310027
IMF-based denoising method for vibration signal in
rotating machinery
XIONG Xin, YANG Shi-xi, ZHOU Xiao-feng
Department of Mechanical Engineering, Zhejiang University, Hangzhou 310027, China
 全文: PDF 
摘要：

针对旋转机械非平稳振动信号中局部低能量噪声的消除问题,提出一种基于固有模式函数(IMF)的振动信号降噪方法.该方法在信号经验模式分解(EMD)的基础上,通过对一阶IMF进行L次随机排序操作,构造观测信号的L个样本序列.根据白噪声各阶IMF的能量密度,计算L个样本序列各自分解所得IMF的阈值.通过样本幅值与阈值的比较,将IMF中过零点区间内极值小于阈值的所有样本点去除,并利用这些阈值去噪后的IMF重构信号.仿真和实验结果表明,本方法对各阶IMF中局部低能量噪声的消除是有效的,且降噪后信号的时频特征显著.
关键词： 经验模式分解;  固有模式函数;  阈值去噪;  旋转机械    
Abstract:

Cancelling low-energy noises of non-stationary vibration signal in rotating machinery based on empirical mode decomposition(EMD) is problematic. Considering this kind of problem, a new method based on intrinsic mode function(IMF) thresholding was presented for locally excluding low-energy noises. In this method, vibration signal was first decomposed into several IMFs using EMD. Then, L sample sequences were constructed in which the sequences were made up of the same signal portion and L different first-order IMFs. L different first-order IMFs were previously obtained by randomly sorting the locations of first-order IMF samples. In order to locally exclude low-energy noises in IMFs, threshold value of each IMF had to be calculated based on the estimation of power densities in white noise only IMFs. The samples which have smaller amplitude than the threshold value in the zero-crossing intervals were considered to be noises and removed. Collect the retained samples and reconstruct signal using thresholded IMFs. Simulation and experiment show that the proposed method can effectively exclude low-energy noises in each IMF. Besides, time-frequency feature of the denoised signal is obvious.
Key words: empirical mode decomposition    intrinsic mode function    threshold denoising    rotating machinery
出版日期: 2011-08-30
:  TN 911.7  
基金资助:

国家自然科学基金资助项目（50675194）；国家“863”高技术研究发展计划资助项目（2008AA04Z410）；国家科技重大专项资助项目（2009ZX04014-101-01）.
通讯作者: 杨世锡,男,教授,博导.     E-mail: yangsx@zju.edu.cn
作者简介: 熊炘(1983—),男,博士生,从事大型旋转机械状态监测与故障诊断的研究工作.E-mail: xiongxinfrank83@zju.edu.cn
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章　　

引用本文:

熊炘, 杨世锡, 周晓峰. 旋转机械振动信号的固有模式函数降噪方法[J]. J4, 2011, 45(8): 1376-1381.

XIONG Xin, YANG Shi-xi, ZHOU Xiao-feng. IMF-based denoising method for vibration signal in
rotating machinery. J4, 2011, 45(8): 1376-1381.

链接本文:

http://www.zjujournals.com/xueshu/eng/CN/10.3785/j.issn.1008-973X.2011.08.008    或    http://www.zjujournals.com/xueshu/eng/CN/Y2011/V45/I8/1376

［1］ 焦卫东,杨世锡,钱苏翔,等.乘性噪声消除的同态变换盲源分离算法［J］.浙江大学学报：工学版,2006,40(4)： 581-584.

JIAO Weidong, YANG Shixi, QIAN Suxiang, et al. Homomorphic transform based BSS algorithm for multiplicative noise reduction ［J］. Journal of Zhejiang University: Engineering Science, 2006, 40(4): 581-584.

［2］ DIPIETROPAOLO D, MLLER H P, TEDESCHI W, et al. Noise reduction in CHD patients by means of BSS ［J］. International Congress Series, 2007, 1300: 217-220.

［3］ 彭志科,卢文秀,褚福磊.新的基于小波变换的振动信号消噪方法［J］.机械工程学报,2006,42(4)：18-22.

PENG Zhike, LU Wenxiu, CHU Fulei. Novel wavelet based method for vibration signal noise cancellation ［J］. Journal of mechanical engineering, 2006, 42(4): 18-22.

［4］ 庞茂,周晓军,胡宏伟,等.基于解析小波变换的奇异性检测和特征提取［J］.浙江大学学报：工学版,2006,40(11)： 1994-1997.

PANG Mao, ZHOU Xiaojun, HU Hongwei, et al. Singularity detection and feature extraction based on analytic wavelet transform ［J］. Journal of Zhejiang University: Engineering Science, 2006, 40(11): 1994-1997.

［5］ WU Z H, Huang N E. A study of the characteristics of white noise using the empirical mode decomposition method ［C］∥ Proceedings of the Royal Society. London: A, Royal Society Publishing, 2004, 460: 1597-1611.

［6］ 曹冲锋,杨世锡,杨将新.基于白噪声统计特性的振动模式提取方法［J］.机械工程学报,2010,46(3)： 65-70.

CAO Chongfeng, YANG Shixi, YANG Jiangxin. Vibration mode extraction method based on the characteristics of white noise ［J］. Journal of Mechanical Engineering, 2010, 46(3): 65-70.

［7］ KOPSINIS Y, MCLAUGHLIN S. Development of EMDbased denoising methods inspired by wavelet thresholding ［J］. Signal Processing, 2009, 57(4): 1351-1362.

［8］ 曹冲锋.基于EMD的机械振动分析与诊断方法研究［D］.杭州：浙江大学,2009： 70-78.

CAO Chongfeng. Study on mechanical vibration analysis and diagnosis methods based on empirical mode decomposition ［D］. Hangzhou: Zhejiang University, 2009: 70-78.

［9］ HUANG N E, SHEN Z, STEVEN R L, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and nonstationary time series analysis ［C］∥ Proceedings of the Royal Society. London. A, Royal Society Publishing, 1998, 454(1971): 903-995.

［10］ FLANDRIN P, RILLING G, GONALVS P. Empirical mode decomposition as a filter bank ［J］. Signal Processing Letters, IEEE, 2004, 11(2): 112-114.

［11］ 屈梁生.机械故障的全息诊断原理［M］.北京：科学出版社,2007: 176-177.
[1] 程卫东,赵德尊　. 用于滚动轴承转频估计的EMD软阈值降噪算法[J]. 浙江大学学报(工学版), 2016, 50(3): 428-436.
[2] 吴丹, 封洲燕, 王静. 微电极阵列神经元锋电位信号的去噪方法[J]. J4, 2010, 44(1): 104-110.
[3] 姚振杰 黄海 陈祥献. 基于经验模式分解的语音端点识别方法[J]. , 2009, 43(4): 705-709.