基于一维盲源分离的滚动轴承故障诊断
|
摘要
盲源分离较之传统的信号处理方法在处理弱信号问题上更具优势。针对轴承故障诊断中因条件限制仅能进行单通道信号采集的情况,提出了一种基于总体经验模式分解的一维盲源分离算法。算法先通过总体经验模式分解将信号分解为多个本征模态函数,再根据本征模态函数之间的相关系数重组观测矩阵,最后利用近似联合对角化对矩阵进行盲源分离。通过数据仿真将该方法与小波分析和Hilbert-Huang变换作对比,说明该方法更适于处理低信噪比的轴承故障信号。对滚动轴承进行了故障诊断实验,成功找到了表征内圈故障和外圈故障的特征信息。
Blind source separation(BSS) has more advantage in weak signal extraction than that of some traditional signal processing methods.Aiming at the case that it only can obtain single-channel signal in bearing fault diagnosis because of the limited condition,a one-dimension BSS algorithm based on empirical mode decomposition(EEMD) is presented.First the signal is decomposed into a number of EEMD,then an observing matrix is rebuilt based on the correlation coefficient between the IMFs and processed by joint approximate diagonalization of eigenmatrix(JADE).Compared with wavelet analysis and Hilbert-Huang transform by simulation data,it shows that this method is more suitable for processing bearing fault signal with low SNR.And then in the experiment of bearing fault diagnosis,the feature information of inner-ring and outer-ring fault is found.
Blind source separation(BSS) has more advantage in weak signal extraction than that of some traditional signal processing methods.Aiming at the case that it only can obtain single-channel signal in bearing fault diagnosis because of the limited condition,a one-dimension BSS algorithm based on empirical mode decomposition(EEMD) is presented.First the signal is decomposed into a number of EEMD,then an observing matrix is rebuilt based on the correlation coefficient between the IMFs and processed by joint approximate diagonalization of eigenmatrix(JADE).Compared with wavelet analysis and Hilbert-Huang transform by simulation data,it shows that this method is more suitable for processing bearing fault signal with low SNR.And then in the experiment of bearing fault diagnosis,the feature information of inner-ring and outer-ring fault is found.
引文
[1]吴昭同,杨世锡.旋转机械故障特征提取与模式分类新方法[M].北京:科学出版社,2012:1-19.
[2]樊永生.机械设备诊断的现代信号处理方法[M].北京:国防工业出版社,2009:15-17,39-47.FAN Y SH.Modern signal processing methods for diag-nosis of mechanical equipment[M].Beijing:NationalDefence Industry Press,2009:15-17,39-47.
[3]VERMA A K,SREEIITH B.Rolling element bearingfault diagnosis using adaptive Morlet wavelet filter[J].International of COMADEM,2009,12(4):25-32.
[4]赵志宏,杨绍普.基于相对小波能量的滚动轴承轴承故障诊断[J].电子测量与仪器学报,2011,25(1):44-49.ZHAO ZH H,YANG SH P.Fault diagnosis of rollerbearing based on relative wavelet energy[J].Journal ofElectronic Measurement and Instrument,2011,25(1):44-49.
[5]何志勇,朱忠奎,张茂青.基于小波包域噪声能量分布的脉冲噪声消除[J].仪器仪表学报,2011,32(9):2071-2078.HE ZH Y,ZHU ZH K,ZHANG M Q.Impulsive noiseremoval based on noise energy distribution in waveletpacket domain[J].Chinese Journal of Scientific Instru-ment,2011,32(9):2071-2078.
[6]RAI V K,MONHANTY A R.Bearing fault diagnosisusing FFT of intrinsic mode functions in Hilbert-Huangtransform[J].Mechanical Systems and Signal Process-ing,2007,21(6):2607-2615.
[7]史习智.盲信号处理———理论与实践[M].上海:上海交通大学出版社,2008:1-6,56-59.SHI X ZH.Blind signal processing———theory and prac-tice[M].Shanghai:Shanghai Jiaotong UniversityPress,2008:1-6,56-59.
[8]杨福生,洪波.独立分量分析的原理与应用[M].北京:清华大学出版社,2006:7-29,97-99.YANG F SH,HONG B.Principles and applications ofICA[M].Beijing:Tsinghua University Press,2006:7-29,97-99.
[9]WU Z H,HUANG N E.Ensemble empirical mode de-composition:a noise-assisted data analysis method[J].Advances in Adaptive Data Analysis,2009,1(1):1-41.
[10]陈佳.基于粗糙集理论和人工神经网络的滚动轴承故障诊断[D].成都:西南交通大学,2012.CHEN J.Rolling bearing fault diagnosis based on roughset theory and artificial neural network[D].Chengdu:Southwest Jiaotong University,2012.
[11]刘立君,王奇,杨克己,等.基于EMD和频谱校正的故障诊断方法[J].仪器仪表学报,2011,32(6):1278-1282.LIU L J,WANG Q,YANG K J,et al.EMD and spec-trum correction-based fault diagnosis method[J].Chi-nese Journal of Scientific Instrument,2011,32(6):1278-1282.
[12]陈隽,李想.运用总体经验模态分解的疲劳信号降噪方法[J].振动、测试与诊断,2011,31(1):15-19.CHEN J,LI X.Application of ensemble empirical modedecomposition to noise reduction of fatigue signal[J].Journal of Vibration,Measurement and Diagnosis,2011,31(1):15-19.
[13]刘佳,杨士莪,朴胜春.基于EEMD的地声信号单通道盲源分离算法[J].哈尔滨工程大学学报,2011,32(2):194-199.LIU J,YANG SH E,PIAO SH CH.The single channelseismic-acoustic signal blind-source separation methodbased on EEMD[J].Journal of Harbin EngineeringUniversity,2011,32(2):194-199.
[14]王宇,迟毅林,伍星,等.基于盲信号处理的机械噪声监测与故障诊断[J].振动与冲击,2009,28(6):33-39.WANG Y,CHI Y L,WU X,et al.Machine noise mo-nitoring and fault diagnosis based on blind signal pro-cessing[J].Journal of Vibration and Shock,2009,28(6):33-39.
[15]郭科,陈聆,陈辉,等.基于JADE算法的地震资料随机噪声盲分离方法研究[J].地学前缘,2011,18(3):302-309.GUO K,CHEN L,CHEN H,et al.The method ofseismic random noise blind separation based on JADE[J].Earth Science Frontier,2011,18(3):302-309.
[2]樊永生.机械设备诊断的现代信号处理方法[M].北京:国防工业出版社,2009:15-17,39-47.FAN Y SH.Modern signal processing methods for diag-nosis of mechanical equipment[M].Beijing:NationalDefence Industry Press,2009:15-17,39-47.
[3]VERMA A K,SREEIITH B.Rolling element bearingfault diagnosis using adaptive Morlet wavelet filter[J].International of COMADEM,2009,12(4):25-32.
[4]赵志宏,杨绍普.基于相对小波能量的滚动轴承轴承故障诊断[J].电子测量与仪器学报,2011,25(1):44-49.ZHAO ZH H,YANG SH P.Fault diagnosis of rollerbearing based on relative wavelet energy[J].Journal ofElectronic Measurement and Instrument,2011,25(1):44-49.
[5]何志勇,朱忠奎,张茂青.基于小波包域噪声能量分布的脉冲噪声消除[J].仪器仪表学报,2011,32(9):2071-2078.HE ZH Y,ZHU ZH K,ZHANG M Q.Impulsive noiseremoval based on noise energy distribution in waveletpacket domain[J].Chinese Journal of Scientific Instru-ment,2011,32(9):2071-2078.
[6]RAI V K,MONHANTY A R.Bearing fault diagnosisusing FFT of intrinsic mode functions in Hilbert-Huangtransform[J].Mechanical Systems and Signal Process-ing,2007,21(6):2607-2615.
[7]史习智.盲信号处理———理论与实践[M].上海:上海交通大学出版社,2008:1-6,56-59.SHI X ZH.Blind signal processing———theory and prac-tice[M].Shanghai:Shanghai Jiaotong UniversityPress,2008:1-6,56-59.
[8]杨福生,洪波.独立分量分析的原理与应用[M].北京:清华大学出版社,2006:7-29,97-99.YANG F SH,HONG B.Principles and applications ofICA[M].Beijing:Tsinghua University Press,2006:7-29,97-99.
[9]WU Z H,HUANG N E.Ensemble empirical mode de-composition:a noise-assisted data analysis method[J].Advances in Adaptive Data Analysis,2009,1(1):1-41.
[10]陈佳.基于粗糙集理论和人工神经网络的滚动轴承故障诊断[D].成都:西南交通大学,2012.CHEN J.Rolling bearing fault diagnosis based on roughset theory and artificial neural network[D].Chengdu:Southwest Jiaotong University,2012.
[11]刘立君,王奇,杨克己,等.基于EMD和频谱校正的故障诊断方法[J].仪器仪表学报,2011,32(6):1278-1282.LIU L J,WANG Q,YANG K J,et al.EMD and spec-trum correction-based fault diagnosis method[J].Chi-nese Journal of Scientific Instrument,2011,32(6):1278-1282.
[12]陈隽,李想.运用总体经验模态分解的疲劳信号降噪方法[J].振动、测试与诊断,2011,31(1):15-19.CHEN J,LI X.Application of ensemble empirical modedecomposition to noise reduction of fatigue signal[J].Journal of Vibration,Measurement and Diagnosis,2011,31(1):15-19.
[13]刘佳,杨士莪,朴胜春.基于EEMD的地声信号单通道盲源分离算法[J].哈尔滨工程大学学报,2011,32(2):194-199.LIU J,YANG SH E,PIAO SH CH.The single channelseismic-acoustic signal blind-source separation methodbased on EEMD[J].Journal of Harbin EngineeringUniversity,2011,32(2):194-199.
[14]王宇,迟毅林,伍星,等.基于盲信号处理的机械噪声监测与故障诊断[J].振动与冲击,2009,28(6):33-39.WANG Y,CHI Y L,WU X,et al.Machine noise mo-nitoring and fault diagnosis based on blind signal pro-cessing[J].Journal of Vibration and Shock,2009,28(6):33-39.
[15]郭科,陈聆,陈辉,等.基于JADE算法的地震资料随机噪声盲分离方法研究[J].地学前缘,2011,18(3):302-309.GUO K,CHEN L,CHEN H,et al.The method ofseismic random noise blind separation based on JADE[J].Earth Science Frontier,2011,18(3):302-309.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心