基于乘积函数相关熵的滚动轴承故障辨识方法
|
摘要
为了提高多工况下对滚动轴承的故障辨识能力,本文提出以乘积函数相关熵为故障特征的滚动轴承故障辨识方法,并利用最小二乘支持向量机(Least Square Support Vector Machine,LSSVM)实现自动辨识.首先对预处理的轴承振动信号进行局部均值分解,提取乘积函数(Product Function,PF),然后计算PF与原始信号的皮尔逊积矩相关系数熵,进而根据离散变量相关熵的估计模型得到乘积函数相关熵(Product Function Correntropy,PFC).以PFC为故障特征,结合LSSVM实现滚动轴承的故障识别.多组工况下的滚动轴承状态辨识实验证实了PFC比经典故障特征具有更高的故障辨识效率;另外改变工况参数提取轴承振动数据,验证了PFC-LSSVM方法具有更好的鲁棒辨识能力.综上所述,本文验证了LMD-PFC-LSSVM方法的高效性和实用性,为提高复杂工况下在线故障诊断能力提供了可靠的技术支持,具有广阔的应用前景.
To improve the ability of roller bearing fault diagnosis under multi rotating condition,this paper proposes a roller bearing fault identification approach that Product Function Correntropy( PFC) is as fault feature and Least Square Support Vector Machine( LSSVM) is applied to implement fault identification of roller bearings under multi-stationary working conditions.Firstly,roller bearing vibration acceleration signal is extracted from motor bearing test bench and the vibration signal is dealt with LMD to extract Product Functions( PF). Then the mathematical model of correntropy of PF and primary signal modified by Pearson Correlation Coefficient Entropy( PCCE) will be calculated,which is named Product Function Correntropy( PFC). Drawing support from LSSVM the safety and failure types identification is achieved.Through the bearing identification experiments in different operating conditions,it is verified that PFC-SVM generates higher diagnosis accuracy than typical fault features. Meanwhile,it is proved that PFC has better robustness than typical fault features when roller bearing operates under mixed condition. Above all,the higher efficiency and availability of LMD-PFC-SVM is confirmed from the experiment consequence. It can be concluded that LMD-PFC-SVM is a reliable technology for roller bearing fault diagnosis online under complicated operating conditions and possesses the broad application prospect.
To improve the ability of roller bearing fault diagnosis under multi rotating condition,this paper proposes a roller bearing fault identification approach that Product Function Correntropy( PFC) is as fault feature and Least Square Support Vector Machine( LSSVM) is applied to implement fault identification of roller bearings under multi-stationary working conditions.Firstly,roller bearing vibration acceleration signal is extracted from motor bearing test bench and the vibration signal is dealt with LMD to extract Product Functions( PF). Then the mathematical model of correntropy of PF and primary signal modified by Pearson Correlation Coefficient Entropy( PCCE) will be calculated,which is named Product Function Correntropy( PFC). Drawing support from LSSVM the safety and failure types identification is achieved.Through the bearing identification experiments in different operating conditions,it is verified that PFC-SVM generates higher diagnosis accuracy than typical fault features. Meanwhile,it is proved that PFC has better robustness than typical fault features when roller bearing operates under mixed condition. Above all,the higher efficiency and availability of LMD-PFC-SVM is confirmed from the experiment consequence. It can be concluded that LMD-PFC-SVM is a reliable technology for roller bearing fault diagnosis online under complicated operating conditions and possesses the broad application prospect.
引文
[1]Tabrizi A,Garibaldi L,Fasana A.Early damage detection of roller bearings using wavelet packet decomposition,ensemble empirical mode decomposition and support vector machine[J].Meccanica,2015,50(3):865-874
[2]秉林,黄仁.机械故障诊断学[M].第3版.北京:机械工业出版社,2007:l-16Bing Lin,Huang Ren.Introduction to machine fault diagnosis[M].3rd ed.Beijing:China Machine Press,2007:1-16
[3]程军圣,郑近德,杨宇,等.基于部分集成局部特征尺度分解与拉普拉斯分值的滚动轴承故障诊断模型[J].振动工程学报,2014,27(6):942-950Cheng Junsheng,Zheng Jinde,Yang Yu,et al.Fault diagnosis model for rolling bearing based on partly ensemble local characteristic—scale decomposition and Laplacian score[J].Jounral of Vibration Engineering,2014,27(6),942-950
[4]Mark G F,Ivan O.Intrinsic time-scale decomposition:time-frequency-energy analysis and real-time filtering of nonstationary signals[J].Proceedings of the Royal Society,2007,463,321-342
[5]Smith S J.The local mean decomposition and its application to EEG perception data[J].Journal of the Royal Society Interface,2005,2(5):443-454
[6]Principe J.Information theoretic learning:renyi's entropy and Kernel perspectives[M].New York:Springer,2010
[7]成昊,唐斌.基于相关熵的盲源分离算法[J].电子科技大学学报,2013,42(2):214-218Cheng Hao,Tang Bin.Blind source separation Algorithm based on correntropy[J].Journal of University of Electronic Science and Technology of China,2013,42(2):214-218
[8]Bin G F,Gao J J,Li X J,et al.Early fault diagnosis of rotating machinery based on wavelet packets:Empirical mode decomposition feature extraction and neural network[J].Mechanical System and Signal Processing,2012,27:696-711
[9]张淑清,胡永涛,李盼.基于MEMD互近似熵及FCM聚类的轴承故障诊断方法[J].中国机械工程,2015,26(19):2613-2618Zhang Shuqing,Hu Yongtao,Li Pan.Fault diagnosis method for bearings based on MEMD cross approximate entropy and FCM clustering[J].Chinese Journal of Mechanical Engineering,2015,26(19):2613-2618
[10]Rodriquez P H,Alonso J B,Ferrer M A,et al.Review of automatic fault diagnosis systems using audio and vibration signals[J].IEEE Transactions on Systems Man and Cybernetics:Systems,2014,44(05):642-652
[11]Gunduz A,Principe J C.Correntropy as a novel measure for nonlinearity tests[J].Signal Processing,2009:14-23
[12]Santamaria I,Pokharel P P,Principe J C.Generalized correlation function:definition,properties,and application to blind equalization[J].IEEE Transactions on Signal Processing,2006,54(6):2187-2197
[13]Mahmoud H,Jeremy T,Catherine M,et al.Comparison between approximate entropy,correntropy and time reversibility:Application to uterine electromyogram signals[J].Medical Engineering&Physics,2011,33(8):980-986
[14]Zhang Yuan,Qin Yong,Xing Zongyi,et al.Safety region estimation and state identification of roller bearing based on statistical feature extraction[J].Journal of Shock and Vibration,2013,20:833-846
[2]秉林,黄仁.机械故障诊断学[M].第3版.北京:机械工业出版社,2007:l-16Bing Lin,Huang Ren.Introduction to machine fault diagnosis[M].3rd ed.Beijing:China Machine Press,2007:1-16
[3]程军圣,郑近德,杨宇,等.基于部分集成局部特征尺度分解与拉普拉斯分值的滚动轴承故障诊断模型[J].振动工程学报,2014,27(6):942-950Cheng Junsheng,Zheng Jinde,Yang Yu,et al.Fault diagnosis model for rolling bearing based on partly ensemble local characteristic—scale decomposition and Laplacian score[J].Jounral of Vibration Engineering,2014,27(6),942-950
[4]Mark G F,Ivan O.Intrinsic time-scale decomposition:time-frequency-energy analysis and real-time filtering of nonstationary signals[J].Proceedings of the Royal Society,2007,463,321-342
[5]Smith S J.The local mean decomposition and its application to EEG perception data[J].Journal of the Royal Society Interface,2005,2(5):443-454
[6]Principe J.Information theoretic learning:renyi's entropy and Kernel perspectives[M].New York:Springer,2010
[7]成昊,唐斌.基于相关熵的盲源分离算法[J].电子科技大学学报,2013,42(2):214-218Cheng Hao,Tang Bin.Blind source separation Algorithm based on correntropy[J].Journal of University of Electronic Science and Technology of China,2013,42(2):214-218
[8]Bin G F,Gao J J,Li X J,et al.Early fault diagnosis of rotating machinery based on wavelet packets:Empirical mode decomposition feature extraction and neural network[J].Mechanical System and Signal Processing,2012,27:696-711
[9]张淑清,胡永涛,李盼.基于MEMD互近似熵及FCM聚类的轴承故障诊断方法[J].中国机械工程,2015,26(19):2613-2618Zhang Shuqing,Hu Yongtao,Li Pan.Fault diagnosis method for bearings based on MEMD cross approximate entropy and FCM clustering[J].Chinese Journal of Mechanical Engineering,2015,26(19):2613-2618
[10]Rodriquez P H,Alonso J B,Ferrer M A,et al.Review of automatic fault diagnosis systems using audio and vibration signals[J].IEEE Transactions on Systems Man and Cybernetics:Systems,2014,44(05):642-652
[11]Gunduz A,Principe J C.Correntropy as a novel measure for nonlinearity tests[J].Signal Processing,2009:14-23
[12]Santamaria I,Pokharel P P,Principe J C.Generalized correlation function:definition,properties,and application to blind equalization[J].IEEE Transactions on Signal Processing,2006,54(6):2187-2197
[13]Mahmoud H,Jeremy T,Catherine M,et al.Comparison between approximate entropy,correntropy and time reversibility:Application to uterine electromyogram signals[J].Medical Engineering&Physics,2011,33(8):980-986
[14]Zhang Yuan,Qin Yong,Xing Zongyi,et al.Safety region estimation and state identification of roller bearing based on statistical feature extraction[J].Journal of Shock and Vibration,2013,20:833-846