轨道交通车辆轴承故障诊断系统的研究
|
摘要
随着我国轨道交通运输的快速发展,机车作为承载乘客的直接载体,其安全与否直接与乘客的生命安全息息相关。近年来轨道交通车辆设备的复杂程度不断提高,故障率相应上升,而轴承作为轨道交通车辆最重要的部件之一,如何对其进行高效、快速而准确地诊断是值得研究的一个重要问题。
本文在认真总结现阶段世界轨道交通故障诊断技术的基础上,总结分析了轴承故障发生情况,并找出故障发生的规律和特性。基于上述的需求分析,本文采用了小波、希尔伯特,神经网络等作为故障诊断方法对轨道交通车辆轴承进行了诊断,首先对采集到的信号进行降噪处理,再对降噪后的信号进行小波变换,对小波变换后的信号分别进行了:1)希尔伯特变换;2)提取特征向量,并将其输入到神经网络进行故障诊断。实验结果表明:
(1)采用小波包和Hilbert包络谱分析,能有效地识别滚动轴承的故障特征,说明该故障特征提取方法是行之有效的;
(2)采用的小波包和BP、Elman、RBF神经网络进行的滚动轴承故障诊断,能有效地识别故障特征,此方法同样适用于轴承等旋转机械的故障诊断;
(3)无论从时间、还是输出准确程度上来看,RBF神经网络都明显优于BP神经网络。
最后,对轨道交通车辆轴承故障诊断系统进行了软件实现。首先对各功能模块进行了说明,然后展示了系统的运行界面。应用希尔伯特变换找出了故障频率;结合小波与神经网络诊断出了故障类型,实现了对机车轴承的智能诊断,保证了机车的安全、快速运行,此系统具有较好的实用性。
With the development of rail transport in our country, rail vehicles safety or not is closely related to passengers’safety. In recent years, the complexity of rail vehicles equipment is increasing, and the failure rate is rising. Therefore, how to diagnose the bearing of rail transit vehicles efficiently, rapidly and accurately is an important issue to be resolved.
In this paper, based on the rigorous summarization to resent rail transit fault diagnosis technology in the world, we have conducted and analyzed the fault occurrence of the Bearing, and identified the rules and characteristics of fault knowledge. Based on the above needs analysis and research, wavelet transform, Hilbert transform, Neural Network was used in this paper, Firstly, rolling bearing signal is denoised. Then, three-layer wavelet packet is adopted to decompose the signal and reconstruct energy eigenvector. Last, 1) Hilbert transform, 2) fault samples of wavelet packet energy eigenvectors are used as neural network input parameters to realize intelligent fault diagnosis. Examples with real data demonstrate:
(1) The experimental result proves that the fault characteristic extracted from improved wavelet packet and Hilbert transform is in accord with the one analyzed from theory, and the fault feature extration method is effective.
(2) The practice example shows that the trained BP, Elman, RBF can diagnose this kind of rolling bearing faults, the method has fair prospects of application for the rotary machine fault diagnosis.
(3) The generalization capability of RBF is superior to that of BP. Meanwhile, In the training time, RBF is also superior to that of BP network
The characteristics of the bearing and extraction methods was presented, the advantages and disadvantages of some methods was brief introduced, then the system's overall architecture was proposed, and each part is described in detail.
Finally, this paper has developed the fault diagnosis system of the bearing of rail transit vehicles. First, functional modules are described, and then system operational interface is presented. In this paper, the fault frequency was found by Hilbert transform; The wavelet and neural network was used to diagnose the kind of fault of the bearing, it realizes intelligent fault diagnosis, and then ensures the safety of rail transit vehicles, fast operation, this system has good practicability.This has important practical value.
本文在认真总结现阶段世界轨道交通故障诊断技术的基础上,总结分析了轴承故障发生情况,并找出故障发生的规律和特性。基于上述的需求分析,本文采用了小波、希尔伯特,神经网络等作为故障诊断方法对轨道交通车辆轴承进行了诊断,首先对采集到的信号进行降噪处理,再对降噪后的信号进行小波变换,对小波变换后的信号分别进行了:1)希尔伯特变换;2)提取特征向量,并将其输入到神经网络进行故障诊断。实验结果表明:
(1)采用小波包和Hilbert包络谱分析,能有效地识别滚动轴承的故障特征,说明该故障特征提取方法是行之有效的;
(2)采用的小波包和BP、Elman、RBF神经网络进行的滚动轴承故障诊断,能有效地识别故障特征,此方法同样适用于轴承等旋转机械的故障诊断;
(3)无论从时间、还是输出准确程度上来看,RBF神经网络都明显优于BP神经网络。
最后,对轨道交通车辆轴承故障诊断系统进行了软件实现。首先对各功能模块进行了说明,然后展示了系统的运行界面。应用希尔伯特变换找出了故障频率;结合小波与神经网络诊断出了故障类型,实现了对机车轴承的智能诊断,保证了机车的安全、快速运行,此系统具有较好的实用性。
With the development of rail transport in our country, rail vehicles safety or not is closely related to passengers’safety. In recent years, the complexity of rail vehicles equipment is increasing, and the failure rate is rising. Therefore, how to diagnose the bearing of rail transit vehicles efficiently, rapidly and accurately is an important issue to be resolved.
In this paper, based on the rigorous summarization to resent rail transit fault diagnosis technology in the world, we have conducted and analyzed the fault occurrence of the Bearing, and identified the rules and characteristics of fault knowledge. Based on the above needs analysis and research, wavelet transform, Hilbert transform, Neural Network was used in this paper, Firstly, rolling bearing signal is denoised. Then, three-layer wavelet packet is adopted to decompose the signal and reconstruct energy eigenvector. Last, 1) Hilbert transform, 2) fault samples of wavelet packet energy eigenvectors are used as neural network input parameters to realize intelligent fault diagnosis. Examples with real data demonstrate:
(1) The experimental result proves that the fault characteristic extracted from improved wavelet packet and Hilbert transform is in accord with the one analyzed from theory, and the fault feature extration method is effective.
(2) The practice example shows that the trained BP, Elman, RBF can diagnose this kind of rolling bearing faults, the method has fair prospects of application for the rotary machine fault diagnosis.
(3) The generalization capability of RBF is superior to that of BP. Meanwhile, In the training time, RBF is also superior to that of BP network
The characteristics of the bearing and extraction methods was presented, the advantages and disadvantages of some methods was brief introduced, then the system's overall architecture was proposed, and each part is described in detail.
Finally, this paper has developed the fault diagnosis system of the bearing of rail transit vehicles. First, functional modules are described, and then system operational interface is presented. In this paper, the fault frequency was found by Hilbert transform; The wavelet and neural network was used to diagnose the kind of fault of the bearing, it realizes intelligent fault diagnosis, and then ensures the safety of rail transit vehicles, fast operation, this system has good practicability.This has important practical value.
引文
[1]黄采伦,樊晓平,陈特放,列车故障在线诊断技术及应用[M].北京:国防工业出版社,2006
[2]谭富强,城市轨道动车动力系统状态随车检测平台的研究[D].南京:南京理工大学.2004
[3]钟秉林,黄仁,机械故障诊断学[M].北京:机械工业出版社,2006
[4]梅宏斌,滚动轴承振动监测与诊断[M].北京:机械工业出版社1996
[5]韩捷,张瑞林,旋转机械故障机理及诊断技术[M].北京:机械工业出版社,1997
[6]管辉,基于小波分析的滚动轴承故障诊断方法研究[D].太原:太原科技大学,2008
[7] Vas, P: Parameter Estimation, Condition Monitoring, and Diagnosis of Electrical Machines. Clarendron Press, Oxford,1993
[8] Masoud Haji, Hamid, A, Toliyat, Patern Recognition-A Technique for Induction Machines Rotor Fault Detection Eccentricity and Broken Bar Fault. Conference Record of the 2001 IEEE Industry Applications Conference, 2001, 30(3):1572-1578
[9] Nandi, S, Toliyat, H A, Cond ition Monitoring and Fault Diagnosis of Electrical Machines– A Review. IEEE Industry Applications Conference, 1999,1(1):197-204
[10] Penman, J, Sedding, H.G, Lloyd, B.A, Fink, W.T, Detection and Location of Interturn Short Circuits in the Stator Windings of Operating Motors. IEEE Trans. Energy Conv, 1996,4(9).
[11] Yazici, B, Kliman, G. B, An Adaptive Statistical Time-Frequency Method for Detection of Broken Bars and Bearing Faults in Motors Using Stator Current. IEEE Trans. On Industry Appl., 1999,2(35).
[12] Abbaszadeh, K, Milimonfared, J, Haji, M, Toliyat, H A, Broken Bar Detection In Induction Motor via Wavelet Transformation. IECON’01: The 27th Annual Conference of the IEEE Industrial Electronics Society 2001
[13] Willsky, A, A Survey of Design Method for Failure Detection in Dynamic Systems. Automatica, 1976,12(1).
[14] Isermann, R, Fault Diagnosis of Machines via Parameter Estimation and Knowledge Processing– Tutorial Paper. Automatica, 1993,4(29).
[15]曾芸,基于小波分析的滚动轴承故障诊断方法研究[D].南昌:南昌大学,2007
[16] Jia Ruijuan, Xu Chunxia, Mechanical Fault Diagnosis and Signal Feature ExtractionBased on Fuzzy Neural Network[J]. Chinese Control Conference, 2008, 234-237
[17]唐秋杭,嵌入式轴承故障诊断算法的研究[D].浙江:浙江大学,2006
[18]梅宏斌,滚动轴承若干故障的诊断与分析[M].电力机车与城市车辆,2005
[19] Nikolaou N G, Antoniadis I A. Rolling Element Bearing Fault Diagnosis Using Wavelet Packets [J]. NDT & E International, 2002, 35 (3): 197-205.
[20] Guo Qianjin, Yu Hai-bin , Xu Aidong, Modified Moret Wavelet Neural Networks for Fault Detection[J]. International Conference on Control and Automation, 2005, 1209-1214
[21] Faa-Jeng Lin, Hsin-Jang Shieh, Po-Kai Huang, Adaptive Wavelet Neural Network Control With Hysteresis Estimation for Piezo-Positioning Mechanism[J]. IEEE Transactions on Neural Networks, 2006, 17 (2),432-444
[22] Huang N E, Shen S P, Eds. Hilbert-Huang Transform and Its Applications. Singapore: World Scientific, 2005.
[23] Pang Peilin, Ding Guangbin, Wavelet-based Diagnostic Model for Rotating Machinery Subject to Vibration Monitoring. Proceedings of the 27th Chinese Control Conference, 2008:303–306.
[24] Cincotti G, Loi G, Pappalardo D, Frequency decomposition and compounding of ultrasound medical images with wavelet packet. IEEE Trans.On medical image, 2001,20(8):764-771.
[25] Liu Hongmei, Wang Shaoping, Ouyang Pingchao, Fault Diagnosis Based on Improved Elman Neural Network for a Hydraulic Servo System. Northern Jiaotong University, 2009
[26]金向阳,张莉,于广滨,基于改进小波神经网络的航空滚动轴承的故障检测[J].中国控制会议,2007, 525-529
[27] Germán Ramos, JoséJ.López, Basilio Pueo, Cascaded warped-FIR and FIR filter structure for loudspeaker equalization with low computational cost requirements. Digital Signal Processing, 2009, 19:393–409.
[28]冯象初,甘小冰,宋国乡,数值泛函与小波理论[M].西安电子科技大学出版社,2003
[29]王本永,郭仁宁,李文生,基于小波包分析和BP网络识别的齿轮故障诊断[J].辽宁工程技术大学学报,2002,21(5):659-660
[30] Grossman A.Wavelet transform and edge detection.Stochastic Processes inPhysics andEngineering,Dodrecht Redial,1986
[31]崔宝珍,潘宏侠,小波分析在滚动轴承故障诊断中的应用[J].科技情报开发与经济,2005,15(2):176-178
[32]汪庆年,黄建红,武和雷,小波变换及其在滚动轴承故障诊断中的应用[J].南昌大学学报,2005,27(1):77-80
[33]刘乐平,林凤涛,基于小波包特征向量与神经网络的滚动轴承故障诊断[J].轴承,2008, (4):659-660
[34]郭计云,基于小波分析的滚动轴承故障诊断方法研究[D].太原:中北大学,2007
[35]唐贵基,范德功,胡爱军等,基于小波包能量特征向量神经网络的旋转机械故障诊断[J].汽轮机技术,2006,48(3):215-217
[36]马拉特,信号处理的小波导引[M].北京机械工业出版社,2002
[37]张德丰,MATLAB神经网络应用设计[M].机械工业出版社,2009
[38]骆江锋,龙江启,范进桢,小波包和BP神经网络在齿轮故障诊断中的应用[J].机械传动,2007,31(3):84-86
[39]李运红,张勇涛,裴未迟,基于小波包-Elman神经网络的电机轴承故障诊断[J].河北理工大学学报,2008,30(4):81-85
[40]朱树先,张仁杰,BP和RBF神经网络在人脸识别中的比较[J].仪器仪表学报, 2007,28(2):375-379
[2]谭富强,城市轨道动车动力系统状态随车检测平台的研究[D].南京:南京理工大学.2004
[3]钟秉林,黄仁,机械故障诊断学[M].北京:机械工业出版社,2006
[4]梅宏斌,滚动轴承振动监测与诊断[M].北京:机械工业出版社1996
[5]韩捷,张瑞林,旋转机械故障机理及诊断技术[M].北京:机械工业出版社,1997
[6]管辉,基于小波分析的滚动轴承故障诊断方法研究[D].太原:太原科技大学,2008
[7] Vas, P: Parameter Estimation, Condition Monitoring, and Diagnosis of Electrical Machines. Clarendron Press, Oxford,1993
[8] Masoud Haji, Hamid, A, Toliyat, Patern Recognition-A Technique for Induction Machines Rotor Fault Detection Eccentricity and Broken Bar Fault. Conference Record of the 2001 IEEE Industry Applications Conference, 2001, 30(3):1572-1578
[9] Nandi, S, Toliyat, H A, Cond ition Monitoring and Fault Diagnosis of Electrical Machines– A Review. IEEE Industry Applications Conference, 1999,1(1):197-204
[10] Penman, J, Sedding, H.G, Lloyd, B.A, Fink, W.T, Detection and Location of Interturn Short Circuits in the Stator Windings of Operating Motors. IEEE Trans. Energy Conv, 1996,4(9).
[11] Yazici, B, Kliman, G. B, An Adaptive Statistical Time-Frequency Method for Detection of Broken Bars and Bearing Faults in Motors Using Stator Current. IEEE Trans. On Industry Appl., 1999,2(35).
[12] Abbaszadeh, K, Milimonfared, J, Haji, M, Toliyat, H A, Broken Bar Detection In Induction Motor via Wavelet Transformation. IECON’01: The 27th Annual Conference of the IEEE Industrial Electronics Society 2001
[13] Willsky, A, A Survey of Design Method for Failure Detection in Dynamic Systems. Automatica, 1976,12(1).
[14] Isermann, R, Fault Diagnosis of Machines via Parameter Estimation and Knowledge Processing– Tutorial Paper. Automatica, 1993,4(29).
[15]曾芸,基于小波分析的滚动轴承故障诊断方法研究[D].南昌:南昌大学,2007
[16] Jia Ruijuan, Xu Chunxia, Mechanical Fault Diagnosis and Signal Feature ExtractionBased on Fuzzy Neural Network[J]. Chinese Control Conference, 2008, 234-237
[17]唐秋杭,嵌入式轴承故障诊断算法的研究[D].浙江:浙江大学,2006
[18]梅宏斌,滚动轴承若干故障的诊断与分析[M].电力机车与城市车辆,2005
[19] Nikolaou N G, Antoniadis I A. Rolling Element Bearing Fault Diagnosis Using Wavelet Packets [J]. NDT & E International, 2002, 35 (3): 197-205.
[20] Guo Qianjin, Yu Hai-bin , Xu Aidong, Modified Moret Wavelet Neural Networks for Fault Detection[J]. International Conference on Control and Automation, 2005, 1209-1214
[21] Faa-Jeng Lin, Hsin-Jang Shieh, Po-Kai Huang, Adaptive Wavelet Neural Network Control With Hysteresis Estimation for Piezo-Positioning Mechanism[J]. IEEE Transactions on Neural Networks, 2006, 17 (2),432-444
[22] Huang N E, Shen S P, Eds. Hilbert-Huang Transform and Its Applications. Singapore: World Scientific, 2005.
[23] Pang Peilin, Ding Guangbin, Wavelet-based Diagnostic Model for Rotating Machinery Subject to Vibration Monitoring. Proceedings of the 27th Chinese Control Conference, 2008:303–306.
[24] Cincotti G, Loi G, Pappalardo D, Frequency decomposition and compounding of ultrasound medical images with wavelet packet. IEEE Trans.On medical image, 2001,20(8):764-771.
[25] Liu Hongmei, Wang Shaoping, Ouyang Pingchao, Fault Diagnosis Based on Improved Elman Neural Network for a Hydraulic Servo System. Northern Jiaotong University, 2009
[26]金向阳,张莉,于广滨,基于改进小波神经网络的航空滚动轴承的故障检测[J].中国控制会议,2007, 525-529
[27] Germán Ramos, JoséJ.López, Basilio Pueo, Cascaded warped-FIR and FIR filter structure for loudspeaker equalization with low computational cost requirements. Digital Signal Processing, 2009, 19:393–409.
[28]冯象初,甘小冰,宋国乡,数值泛函与小波理论[M].西安电子科技大学出版社,2003
[29]王本永,郭仁宁,李文生,基于小波包分析和BP网络识别的齿轮故障诊断[J].辽宁工程技术大学学报,2002,21(5):659-660
[30] Grossman A.Wavelet transform and edge detection.Stochastic Processes inPhysics andEngineering,Dodrecht Redial,1986
[31]崔宝珍,潘宏侠,小波分析在滚动轴承故障诊断中的应用[J].科技情报开发与经济,2005,15(2):176-178
[32]汪庆年,黄建红,武和雷,小波变换及其在滚动轴承故障诊断中的应用[J].南昌大学学报,2005,27(1):77-80
[33]刘乐平,林凤涛,基于小波包特征向量与神经网络的滚动轴承故障诊断[J].轴承,2008, (4):659-660
[34]郭计云,基于小波分析的滚动轴承故障诊断方法研究[D].太原:中北大学,2007
[35]唐贵基,范德功,胡爱军等,基于小波包能量特征向量神经网络的旋转机械故障诊断[J].汽轮机技术,2006,48(3):215-217
[36]马拉特,信号处理的小波导引[M].北京机械工业出版社,2002
[37]张德丰,MATLAB神经网络应用设计[M].机械工业出版社,2009
[38]骆江锋,龙江启,范进桢,小波包和BP神经网络在齿轮故障诊断中的应用[J].机械传动,2007,31(3):84-86
[39]李运红,张勇涛,裴未迟,基于小波包-Elman神经网络的电机轴承故障诊断[J].河北理工大学学报,2008,30(4):81-85
[40]朱树先,张仁杰,BP和RBF神经网络在人脸识别中的比较[J].仪器仪表学报, 2007,28(2):375-379