转子动平衡计算及转子轴心轨迹识别
|
摘要
饱和汽轮机是船用核动力装置的重要组成部分，而转子是其主要工作部件，通过转子轴心轨迹来识别转子工况，对监测转子状态保证其正常运行，减少事故发生率，具有十分重要的意义。
本文在阅读了大量关于转子动平衡和轴心轨迹识别方法资料的基础上，以船用核动力装置中的饱和汽轮机为对象，运用转子动力学理论，对汽轮机转子的振动特性、转子常见故障及其轴心轨迹特征进行了研究，设计出了汽轮机转子动平衡计算程序，从模式识别的角度研究、设计了基于BP网络的转子轴心轨迹识别系统。本系统以计算机为主体，实时采集机组运行中的各种参数，通过计算机进行实时分析与处理，对饱和汽轮机转子的运行状态进行评估，并进行显示、记录、对异常状态进行报警，并对故障状态进行诊断。本系统集数据采集、状态监测、故障诊断于一体，功能完善，操作方便，对指导饱和汽轮机的运行与维修具有重要的实用价值。
为了对转子故障进行分析，并对转子故障特性进行识别，在转子特性实验台架上利用影响系数法动平衡计算程序进行了转子动平衡实验研究，实验结果表明该程序能正确给出平衡重量及其相位，同时也验证了影响系数法的可行性及程序的正确性。此外还模拟了转子不平衡与转子不对中的多组实验，获取了大量的实验数据，利用这些数据对所设计的基于BP神经网络的轴心轨迹识别系统的网络结构进行了训练生成，利用该识别系统对转子的不平衡、不对中故障进行了在线识别，实验结果表明系统得到的轴心轨迹与故障模式完全匹配，从而验证了根据轴心轨迹进行故障分析与故障诊断的可行性与有效性，也验证了所设计的轴心轨迹识别系统运行的正确性。
The saturated steam turbine is an important part of the nuclear power equipment of ships, and the rotor is a major working part of the turbine. It has very important significance that monitoring rotor condition form orbit of shaft centerline, which can assure the turbine running unfailingly and reducing the rate of accident.
A lot of papers about rotor dynamic balance and orbit of shaft centerline recognition has been read by the author and let as the basis. In this paper, the object of study is the turbine of the nuclear power equipment of ships, based on the study of dynamics specific property of the turbine rotor and common fault and its orbit of shaft centerline character, the system of dynamic balance calculation for turbine rotor and orbit of shaft centerline recognition based on BP network has been designed. In this system the computer is the most important part. The computer can control the real-time data acquisition and has the data of all kinds of parameters processed and analyzed. The system can evaluate the running status of the saturated steam turbine. It also can display and save the processed results. The system can give an alarm to the abnormal status. If the saturated steam turbine has faults the system can diagnose them. Moreover, the system has the complete functions and convenient operations. The system will
play an important action in directing the running and maintaining of the saturated steam turbine.
A rotor experimental platform had been constructed. It provided an experimental platform for fault analysis and fault diagnosis. The experiment for the system of dynamic balance calculation which designed with the method of influence coefficients had been carried
out, the experimental results indicated that the program can calculate the balance mass and phase, the results also showed the feasibility of method of influence coefficients and correctness of the program. Besides, a series of experiments of unbalance and non-alignment had been carried out, a great deal of data had been acquired, the pattern recognition system for orbit of shaft centerline based on BP network had been trained using these data, the system can on-line recognize the faults of unbalance and non-alignment. The experimental results indicated that the orbit of shaft centerline that the system acquired were consistent with the fault modes. The results showed the feasibility and validity of carrying out fault analysis and fault diagnosis with the orbit of shaft centerline. They also showed the correctness of the running of the recognition system .
本文在阅读了大量关于转子动平衡和轴心轨迹识别方法资料的基础上，以船用核动力装置中的饱和汽轮机为对象，运用转子动力学理论，对汽轮机转子的振动特性、转子常见故障及其轴心轨迹特征进行了研究，设计出了汽轮机转子动平衡计算程序，从模式识别的角度研究、设计了基于BP网络的转子轴心轨迹识别系统。本系统以计算机为主体，实时采集机组运行中的各种参数，通过计算机进行实时分析与处理，对饱和汽轮机转子的运行状态进行评估，并进行显示、记录、对异常状态进行报警，并对故障状态进行诊断。本系统集数据采集、状态监测、故障诊断于一体，功能完善，操作方便，对指导饱和汽轮机的运行与维修具有重要的实用价值。
为了对转子故障进行分析，并对转子故障特性进行识别，在转子特性实验台架上利用影响系数法动平衡计算程序进行了转子动平衡实验研究，实验结果表明该程序能正确给出平衡重量及其相位，同时也验证了影响系数法的可行性及程序的正确性。此外还模拟了转子不平衡与转子不对中的多组实验，获取了大量的实验数据，利用这些数据对所设计的基于BP神经网络的轴心轨迹识别系统的网络结构进行了训练生成，利用该识别系统对转子的不平衡、不对中故障进行了在线识别，实验结果表明系统得到的轴心轨迹与故障模式完全匹配，从而验证了根据轴心轨迹进行故障分析与故障诊断的可行性与有效性，也验证了所设计的轴心轨迹识别系统运行的正确性。
The saturated steam turbine is an important part of the nuclear power equipment of ships, and the rotor is a major working part of the turbine. It has very important significance that monitoring rotor condition form orbit of shaft centerline, which can assure the turbine running unfailingly and reducing the rate of accident.
A lot of papers about rotor dynamic balance and orbit of shaft centerline recognition has been read by the author and let as the basis. In this paper, the object of study is the turbine of the nuclear power equipment of ships, based on the study of dynamics specific property of the turbine rotor and common fault and its orbit of shaft centerline character, the system of dynamic balance calculation for turbine rotor and orbit of shaft centerline recognition based on BP network has been designed. In this system the computer is the most important part. The computer can control the real-time data acquisition and has the data of all kinds of parameters processed and analyzed. The system can evaluate the running status of the saturated steam turbine. It also can display and save the processed results. The system can give an alarm to the abnormal status. If the saturated steam turbine has faults the system can diagnose them. Moreover, the system has the complete functions and convenient operations. The system will
play an important action in directing the running and maintaining of the saturated steam turbine.
A rotor experimental platform had been constructed. It provided an experimental platform for fault analysis and fault diagnosis. The experiment for the system of dynamic balance calculation which designed with the method of influence coefficients had been carried
out, the experimental results indicated that the program can calculate the balance mass and phase, the results also showed the feasibility of method of influence coefficients and correctness of the program. Besides, a series of experiments of unbalance and non-alignment had been carried out, a great deal of data had been acquired, the pattern recognition system for orbit of shaft centerline based on BP network had been trained using these data, the system can on-line recognize the faults of unbalance and non-alignment. The experimental results indicated that the orbit of shaft centerline that the system acquired were consistent with the fault modes. The results showed the feasibility and validity of carrying out fault analysis and fault diagnosis with the orbit of shaft centerline. They also showed the correctness of the running of the recognition system .
引文
[1] 黄文虎，夏松波，刘瑞岩等．设备故障诊断原理、技术及应用．科学出版社，1996：1-8页
[2] Moradian AM A, Chow M P, Osborne R L, Jankin M A. Turbin Aid. Turbine Artificied Intelligence, Diagnostic. WHEC, owp 5266
[3] 百木万博等．故障诊断、异常诊断极其对策．振动监测、机械振动讲演文集，机械工业部郑州机械研究所，1984
[4] 张瑞林．机械故障诊断技术发展现状及展望．第二届全国机械设备故障诊断学术会议论文集，1988
[5] G. A. Carpenter and S. Grossberg. Chemical transmitters in selforganizing pattern recoguition architectures. On Neural Network,1990, Vol(2):30-33P
[6] T. Kohonen. Self-organization formation of topologically correct feature maps. Biolog Cybernetics, 1982, Vol(43):59-69P
[7] 蔡元龙．模式识别．西安电子科技大学出版社，1992：1-17页
[8] W. C. McClloch and W. Pitts, A Logical Calculus of the Ideas Immanent in Nervous Activity. Bulletin of Mathematical Bio-Physica. 1943, Vol(5):115-133P
[9] 何述东，黄献青，黄心汉．多层前向神经网络结构的研究进展．控制理论与应用，1998，15(3)：19-23页
[10] 陆颂元．汽轮发电机组振动．中国电力出版社，2000：1-18，105-207页
[11] 钟一鄂，何衍宗，王正，李方泽．转子动力学．清华大学出版社，1987：1-40页
[12] 施维新．汽轮发电机组振动及事故．中国电力出版社，1998：1-325页
[13] 刘泽明，崔亚辉，王惠武．微机化刚性转子动平衡测量装置的研制．机械科学与技术，1996，15(4)：614-616页
[14] 王莹，杜永祚．转子动平衡及振动测试仪的研制．河北机械，1996(2)：39-43页
[15] 邱海，屈梁生．转子动平衡中的相关平衡面问题．化工机械，1999：82-85页
[16] Songyuan Lu. Nonlinear Vibration Characteristics with Large unbalance of Turbo generators Rotor Systems. Vibration Conf. 1995, Vol(37):133-148P
[17] W kellenberger. Shall a Flexible Rotor Be Balance in N or (N+2) Plance. Vibration Conference and International Design Automation conference, 1971, Vol(1):108-121P
[18] M. L. Adams. Large Unblance Vibration Analysis of Steam Turbine Generators. EPR1 Report CS-3716, 1984,319-327P
[19] 戴汝为．人工智能．化学工业出版社，2002：32-47页
[20] 王永骥，徐健．神经元网络控制．机械工业出版社，1998，1-155页
[21] 吴简彤，王建华．神经网络技术及其应用．哈尔滨工程大学出版社，1998：1-69页
[22] 虞和济，陈长征，张省．基于神经网络的智能诊断．冶金工业出版社，2000：104-138页
[23] 韩捷，张瑞林．旋转机械故障机理及诊断技术．机械工业出版社，1997：91-135页
[24] R. A. Collacott. Mechical Condition Monitoring and Fault Diagnosis. Chapman and Hall. london, 1977:11-74P
[25] Hoskins. J. C, Himmelblau. D. M. Artificial Neural Network Modes of knowledge Representation in Chemical Engineering. Computers chem. Eneng, 1998Vol(14):6-14P
[26] 黄秀珠．大型旋转机械振动故障特征分析和诊断．电力部热工研究所，1996：15-60页
[27] 刘雄，赵振毅．转子监测和诊断系统．西安交通大学出版社，1996：31-91页
[28] F. Rosenblantt. Two theorems of statistical reparability in the perceptron, in mechanization of Thought Processes. Proceedings of a Symposium Held at the National Physical laboratory, 1978, Vol(1):421-456P
[29] S. Grossberg. Adaptive Pattern classification and universal recoding. Biology Cybernetics 1976Vol(23):187-202P
[30] F. Rosenblantt. Principles of neurodynamics: Perceptroud and Theory of Brain Mechanisms. Washington, DC: Sparttan books, 1962:1-35P
[31] G. A. Carpenter and S. Grossberg. Chemical transmitters in selforganizing pattern recognition architectures. On Neural Network, 1990, Vol(2):30-33P
[32] T. kohonen. Self-organized formation of topologically correct feature maps. Biology Cybernetics, 1982, Vol (43:59-69P
[33] Yange Liu, Wei hiu, Yimo Zhang. Inspection of defects in optical fibers based on back-propagation neural networks. Optic Com
mcations, 2001, Vol(1):369-378P
[34] Eiji Yoshida, kiyoshi Shizuma, Satoru Endo, Takamitsu Oka. Application of neural networks for the analysis of gamma-ray spectra measured with a Ge spectrometer. Neural Instruments and Methods in Physics Research, 2002, Vol(1):557-563P
[35] Luo Dehan, Lan Wenhui, Huang Sheng. Development of an On-line Condition Monitoring System Based on Neural Networks. Journal of UESTof China, 1999, Vol(28):527-532P
[36] S. Khanmohammadi, I. Hassanzadeh, H. R. Zarei Poor. Fault diagnosis competitive neural network with prioritized modification rule of connection weights. Artificial Intelligence in Engneering, 2000, Vol(14):127-132P
[37] 刘占生，张新江，夏松波，怀进杰．轴心轨迹特征的提取方法及其在旋转机械故障诊断中的应用．汽轮机技术，1997(39)：303-305页
[38] 刘占生，张新江，杨建国，夏松波．转子轴心轨迹故障诊断特征识别方法研究．哈尔滨工业大学学报，1998(30)：22-25页
[39] 韩西京，史铁林，李录平，杨叔子．故障诊断中几种征兆自动获取技术研究．华中理工大学学报，1997(25)：53-55页
[40] 汪家铭．用轴心轨迹/位置诊断机器故障．热力发电，1994(5)：63-65页
[41] 赵林度，盛昭瀚，张静．汽轮发电机组轴心轨迹自动识别系统的开发．汽轮机技术，1997(39)：329-332页
[42] 韩西京，史铁林，李录平，杨叔子．旋转机械轴心轨迹的自动识别．振动、测试与诊断，1997(17)：20-25页
[2] Moradian AM A, Chow M P, Osborne R L, Jankin M A. Turbin Aid. Turbine Artificied Intelligence, Diagnostic. WHEC, owp 5266
[3] 百木万博等．故障诊断、异常诊断极其对策．振动监测、机械振动讲演文集，机械工业部郑州机械研究所，1984
[4] 张瑞林．机械故障诊断技术发展现状及展望．第二届全国机械设备故障诊断学术会议论文集，1988
[5] G. A. Carpenter and S. Grossberg. Chemical transmitters in selforganizing pattern recoguition architectures. On Neural Network,1990, Vol(2):30-33P
[6] T. Kohonen. Self-organization formation of topologically correct feature maps. Biolog Cybernetics, 1982, Vol(43):59-69P
[7] 蔡元龙．模式识别．西安电子科技大学出版社，1992：1-17页
[8] W. C. McClloch and W. Pitts, A Logical Calculus of the Ideas Immanent in Nervous Activity. Bulletin of Mathematical Bio-Physica. 1943, Vol(5):115-133P
[9] 何述东，黄献青，黄心汉．多层前向神经网络结构的研究进展．控制理论与应用，1998，15(3)：19-23页
[10] 陆颂元．汽轮发电机组振动．中国电力出版社，2000：1-18，105-207页
[11] 钟一鄂，何衍宗，王正，李方泽．转子动力学．清华大学出版社，1987：1-40页
[12] 施维新．汽轮发电机组振动及事故．中国电力出版社，1998：1-325页
[13] 刘泽明，崔亚辉，王惠武．微机化刚性转子动平衡测量装置的研制．机械科学与技术，1996，15(4)：614-616页
[14] 王莹，杜永祚．转子动平衡及振动测试仪的研制．河北机械，1996(2)：39-43页
[15] 邱海，屈梁生．转子动平衡中的相关平衡面问题．化工机械，1999：82-85页
[16] Songyuan Lu. Nonlinear Vibration Characteristics with Large unbalance of Turbo generators Rotor Systems. Vibration Conf. 1995, Vol(37):133-148P
[17] W kellenberger. Shall a Flexible Rotor Be Balance in N or (N+2) Plance. Vibration Conference and International Design Automation conference, 1971, Vol(1):108-121P
[18] M. L. Adams. Large Unblance Vibration Analysis of Steam Turbine Generators. EPR1 Report CS-3716, 1984,319-327P
[19] 戴汝为．人工智能．化学工业出版社，2002：32-47页
[20] 王永骥，徐健．神经元网络控制．机械工业出版社，1998，1-155页
[21] 吴简彤，王建华．神经网络技术及其应用．哈尔滨工程大学出版社，1998：1-69页
[22] 虞和济，陈长征，张省．基于神经网络的智能诊断．冶金工业出版社，2000：104-138页
[23] 韩捷，张瑞林．旋转机械故障机理及诊断技术．机械工业出版社，1997：91-135页
[24] R. A. Collacott. Mechical Condition Monitoring and Fault Diagnosis. Chapman and Hall. london, 1977:11-74P
[25] Hoskins. J. C, Himmelblau. D. M. Artificial Neural Network Modes of knowledge Representation in Chemical Engineering. Computers chem. Eneng, 1998Vol(14):6-14P
[26] 黄秀珠．大型旋转机械振动故障特征分析和诊断．电力部热工研究所，1996：15-60页
[27] 刘雄，赵振毅．转子监测和诊断系统．西安交通大学出版社，1996：31-91页
[28] F. Rosenblantt. Two theorems of statistical reparability in the perceptron, in mechanization of Thought Processes. Proceedings of a Symposium Held at the National Physical laboratory, 1978, Vol(1):421-456P
[29] S. Grossberg. Adaptive Pattern classification and universal recoding. Biology Cybernetics 1976Vol(23):187-202P
[30] F. Rosenblantt. Principles of neurodynamics: Perceptroud and Theory of Brain Mechanisms. Washington, DC: Sparttan books, 1962:1-35P
[31] G. A. Carpenter and S. Grossberg. Chemical transmitters in selforganizing pattern recognition architectures. On Neural Network, 1990, Vol(2):30-33P
[32] T. kohonen. Self-organized formation of topologically correct feature maps. Biology Cybernetics, 1982, Vol (43:59-69P
[33] Yange Liu, Wei hiu, Yimo Zhang. Inspection of defects in optical fibers based on back-propagation neural networks. Optic Com
mcations, 2001, Vol(1):369-378P
[34] Eiji Yoshida, kiyoshi Shizuma, Satoru Endo, Takamitsu Oka. Application of neural networks for the analysis of gamma-ray spectra measured with a Ge spectrometer. Neural Instruments and Methods in Physics Research, 2002, Vol(1):557-563P
[35] Luo Dehan, Lan Wenhui, Huang Sheng. Development of an On-line Condition Monitoring System Based on Neural Networks. Journal of UESTof China, 1999, Vol(28):527-532P
[36] S. Khanmohammadi, I. Hassanzadeh, H. R. Zarei Poor. Fault diagnosis competitive neural network with prioritized modification rule of connection weights. Artificial Intelligence in Engneering, 2000, Vol(14):127-132P
[37] 刘占生，张新江，夏松波，怀进杰．轴心轨迹特征的提取方法及其在旋转机械故障诊断中的应用．汽轮机技术，1997(39)：303-305页
[38] 刘占生，张新江，杨建国，夏松波．转子轴心轨迹故障诊断特征识别方法研究．哈尔滨工业大学学报，1998(30)：22-25页
[39] 韩西京，史铁林，李录平，杨叔子．故障诊断中几种征兆自动获取技术研究．华中理工大学学报，1997(25)：53-55页
[40] 汪家铭．用轴心轨迹/位置诊断机器故障．热力发电，1994(5)：63-65页
[41] 赵林度，盛昭瀚，张静．汽轮发电机组轴心轨迹自动识别系统的开发．汽轮机技术，1997(39)：329-332页
[42] 韩西京，史铁林，李录平，杨叔子．旋转机械轴心轨迹的自动识别．振动、测试与诊断，1997(17)：20-25页