基于DSP的同步无刷励磁发电机旋转整流器故障诊断系统研究
|
摘要
目前,应用在同步发电机上的无刷励磁技术,取消了原有励磁系统中滑动接触部分(滑环、碳刷),提高了运行的可靠性,因此近几年同步无刷励磁发电机在工业中的应用越来越广泛。
同步无刷励磁发电机主要是依靠交流励磁机和旋转整流器来实现无刷励磁的,由于旋转整流器安装在电机转轴上,随电机转子一起旋转,其核心器件整流二极管会承受很大的离心力,故障率较高。旋转整流器的失效会导致同步电机不能正常运行,严重时还会对交流励磁机造成一定的损失,因此对旋转整流器进行实时监测与故障诊断是非常重要的。
由于无刷励磁系统取消了滑环和碳刷,旋转整流器的电压和电流不可能直接测量,使其监测和保护十分困难,但又可知当旋转整流器发生故障时,交流励磁机电枢绕组产生的磁场就会发生异常,感应出交流电势也会随着不同的运行状态产生不同的谐波分量。本文通过检测线圈法,对感应电势进行谐波分析,采用综合故障诊断法,对同步无刷励磁发电机旋转整流器故障进行在线故障诊断。
谐波分析使用FFT得到励磁绕组中感应电势的各次谐波幅值,而FFT主要存在有频谱混叠、频率分辨能力、栅栏效应及频谱泄漏等问题。本文使用模拟滤波器;选择合适的采样频率和采样点数;采用加窗法和幅值比值校正法对FFT进行减小误差处理,经过处理效果明显提高。
目前已有的故障诊断法主要有阈值法,神经网络法和模糊模式识别法等,本文对比了他们各自的优缺点,提出了综合故障诊断法,即先用阈值法判断故障是否发生,如有故障发生则使用神经网络法与模糊模式识别法相结合的方法来具体判断故障类型,如没有故障发生则继续监测。
本文使用TMS320F2812作为旋转整流器故障诊断系统的中心处理器,分析感应电势中的各次谐波幅值,采用综合故障诊断法进行故障诊断,使用液晶屏将显示信息,并通过触摸屏完成参数设置和系统操作。
Brushless exciter has been used in generator for many years; it has canceled slip rings and carbon brushes, and improved its operational stability. So in recent years, synchronous brushless excitation generator has been increasingly used in the industrial application.
Synchronous brushless excitation generators mainly rely on the exchange exciter and rotating rectifiers to achieve brushless excitation. Because rotating rectifiers work with the motor shaft, and install on the motor rotor rotation. The core of rotating rectifiers is diode device, it tremendous centrifugal force, so its failure rate is high. The failure of rotating rectifier can lead to the synchronous motor can not operate normally. Therefore the real-time monitoring and diagnosis of rotating rectifier's failure is very important.
Because brushless excitation system has canceled slip rings ands carbon brushes, it's voltage and current can not be measured directly, so it's difficult to be monitored and protected. But it is been known that when failure has happen, the exchange of armature winding excitation generated magnetic field will be abnormal, the AC induction potential will be different with the operation of the state which have different harmonics. In this paper, using detection coil method, harmonic analysis, use comprehensive diagnostics, achieve the monitoring and fault diagnosis of the brushless synchronous generator excitation rotating rectifier.
FFT analysis by the use of harmonic excitation winding induction potential of the various harmonic amplitude, and the major existence of FFT spectrum aliasing, frequency resolution capability, and the effect of the fence spectrum leakage problems. In this paper, using analogue filter; choose suitable sampling frequency and sampling points; using window law and adjusting method to reduce the error of FFT, the results have been markedly improved.
At present, the methods of fault diagnosis mainly are threshold method, neural network and fuzzy pattern recognition, etc. The paper compares the advantages and disadvantages of them and proposes an integrated fault diagnosis, so use threshold method determine whether the fault, if has malfunction in the use of neural networks and fuzzy combination of pattern recognition methods to determine the specific types of faults, if has no faults then continue to monitor.
In this paper, use TMS320F2812 as central processor of fault diagnosis system; induction of electric potential in the various harmonic amplitude, integrate fault diagnosis method fault diagnosis, use liquid crystal to display screen information, and complete control system through the touch screen.
同步无刷励磁发电机主要是依靠交流励磁机和旋转整流器来实现无刷励磁的,由于旋转整流器安装在电机转轴上,随电机转子一起旋转,其核心器件整流二极管会承受很大的离心力,故障率较高。旋转整流器的失效会导致同步电机不能正常运行,严重时还会对交流励磁机造成一定的损失,因此对旋转整流器进行实时监测与故障诊断是非常重要的。
由于无刷励磁系统取消了滑环和碳刷,旋转整流器的电压和电流不可能直接测量,使其监测和保护十分困难,但又可知当旋转整流器发生故障时,交流励磁机电枢绕组产生的磁场就会发生异常,感应出交流电势也会随着不同的运行状态产生不同的谐波分量。本文通过检测线圈法,对感应电势进行谐波分析,采用综合故障诊断法,对同步无刷励磁发电机旋转整流器故障进行在线故障诊断。
谐波分析使用FFT得到励磁绕组中感应电势的各次谐波幅值,而FFT主要存在有频谱混叠、频率分辨能力、栅栏效应及频谱泄漏等问题。本文使用模拟滤波器;选择合适的采样频率和采样点数;采用加窗法和幅值比值校正法对FFT进行减小误差处理,经过处理效果明显提高。
目前已有的故障诊断法主要有阈值法,神经网络法和模糊模式识别法等,本文对比了他们各自的优缺点,提出了综合故障诊断法,即先用阈值法判断故障是否发生,如有故障发生则使用神经网络法与模糊模式识别法相结合的方法来具体判断故障类型,如没有故障发生则继续监测。
本文使用TMS320F2812作为旋转整流器故障诊断系统的中心处理器,分析感应电势中的各次谐波幅值,采用综合故障诊断法进行故障诊断,使用液晶屏将显示信息,并通过触摸屏完成参数设置和系统操作。
Brushless exciter has been used in generator for many years; it has canceled slip rings and carbon brushes, and improved its operational stability. So in recent years, synchronous brushless excitation generator has been increasingly used in the industrial application.
Synchronous brushless excitation generators mainly rely on the exchange exciter and rotating rectifiers to achieve brushless excitation. Because rotating rectifiers work with the motor shaft, and install on the motor rotor rotation. The core of rotating rectifiers is diode device, it tremendous centrifugal force, so its failure rate is high. The failure of rotating rectifier can lead to the synchronous motor can not operate normally. Therefore the real-time monitoring and diagnosis of rotating rectifier's failure is very important.
Because brushless excitation system has canceled slip rings ands carbon brushes, it's voltage and current can not be measured directly, so it's difficult to be monitored and protected. But it is been known that when failure has happen, the exchange of armature winding excitation generated magnetic field will be abnormal, the AC induction potential will be different with the operation of the state which have different harmonics. In this paper, using detection coil method, harmonic analysis, use comprehensive diagnostics, achieve the monitoring and fault diagnosis of the brushless synchronous generator excitation rotating rectifier.
FFT analysis by the use of harmonic excitation winding induction potential of the various harmonic amplitude, and the major existence of FFT spectrum aliasing, frequency resolution capability, and the effect of the fence spectrum leakage problems. In this paper, using analogue filter; choose suitable sampling frequency and sampling points; using window law and adjusting method to reduce the error of FFT, the results have been markedly improved.
At present, the methods of fault diagnosis mainly are threshold method, neural network and fuzzy pattern recognition, etc. The paper compares the advantages and disadvantages of them and proposes an integrated fault diagnosis, so use threshold method determine whether the fault, if has malfunction in the use of neural networks and fuzzy combination of pattern recognition methods to determine the specific types of faults, if has no faults then continue to monitor.
In this paper, use TMS320F2812 as central processor of fault diagnosis system; induction of electric potential in the various harmonic amplitude, integrate fault diagnosis method fault diagnosis, use liquid crystal to display screen information, and complete control system through the touch screen.
引文
[1]黄耀群,李兴源.同步电机现代励磁系统及其控制[M].成都:成都科技大学出版社,1993
[2]沈标正.电机故障诊断技术仁[M].北京:机械工业出版社,1996
[3]樊俊,陈忠,涂光瑜.同步发电机半导体励磁原理及应用[M].北京.水利电力出版社.1991年第二版
[4]G.F.H.Allen.Brushless Excitation Systems for Synchronous Machines[J].IEEE Transactions on Power Apparatus and Systems v PAS-91 NO.5 Septemberctober.1972.pp.1848-1854
[5]S.I.Loginov,V.I.Valkov,T.V.Bekhtereva,and E.I.Ryabenko.Investigation of Brushless Thyristor Excitation Systems for Medium- and High-Capacity Turbomotors[J].Soviet Electrical Engineering(English Translation of Elektrotekhnika)vol(49).No.2.1978.pp.49-53
[6]A.C.Williamson and B.J.Chalmers.Novel Form of Synchronous Machine Excitation[J].Soviet Electrical Engineering(English Translation of Elektrotekhnika)vol49 No3 1978 p4-9.Electric Machines and Electranechanics vol.1.No.4 July-September.1977.pp.365-376
[7]李伟清,王绍禹.发电机故障检查分析及预防[M].北京:中国电力出版社,1996.
[8]W.F.Wrkght,R Hawley and J.L.Dinely.Brushless thyristor excitation systems[J].IEEE Trans.Systems.1972.pp.1848-1851
[9]刘念.无刷励磁旋转整流器电流波形分析及其故障的微机识别[J].电力系统自动化,1994,18(4):19-26
[10]S.C.Gupta,M.L.Dewal.A Novel Technique for the Measurement of Average Rotor Temperature of Brushless Synchronous Machines[J].IEEE Trans.On Power Apparatus and Systems.Vol.PAS-98,No.4.July/Aug.1979,pp.1238-1243
[11]J.Sottle,F.C.Trutt and A.W.Leedy.Condition monitoring of brushless three-phase generators with stator winding or rotor circuit deterioration[J].Conference Record IAS Annual Meeting(IEEE Industry Applications Society)vol.3.Septernber 30-October 4.2001.pp.1587-1594
[12]黄念慈,黄山,赵莉华.旋转整流器故障检测的一种方法[J].电工技术学报,11(4)1996:50-53
[13]施惠昌,陈泉林,冯玉田,袁建华,王庭山.无刷励磁发电机转子电流及温度监测装置[J].计算机自动测量与控制.NO.5.2001
[14]陈维忠.无刷同步电机转子电流测量及旋转整流器故障检测装置的研究[J].自动化与仪表,200015(2):12-14
[15]刘念.无刷同步电机旋转整流器监测新方法研究[J],电工技术杂志,NO.8.2000
[16]Li Xingyuan,Microprocessor-based fault monitor for rotating rectifiers of brushless AC exciters using a pattern ecognition approach[J].Conference Record-IEEE Instrumentation and Measurement Technology Conference 1 May 10-12 1994.
[17]门馄,沈善德,朱守真,杨常府.综合测辨法用于无刷励磁系统故障诊断的研究[J].电力系统自动化,2001,25(24):26-29
[18]丁康,谢明,王志杰.离散频谱的幅值、相位和频率校正方法对误差分析[J].动态分析与测试技术,1996,6(2):10-29
[19]黄方能,吴玉燕,FFT谐波检测存在的问题[J],广西电力,2005(4):39-41
[20]郑恩让,杨润贤,高森,关于电力系统FFT谐波检测存在问题的研究[J],2006,34(18):52-57
[21]Tommy W S Chow,Hnng Zhou Tan.HOS—based on parametric and parametric methodologies for machine fault detection[J].IEEE Trans.on Industrial Electronics 2000 47(5):051-059
[22]Kral C,Wieser R S,Pirker F,et at.Sequences of field oriented control for the detection of faulty rotor bars in induction machines—the Vienna Monitoring Method[J].IEEE Trans.On Industrial Eiectronics,2000,47(5):1042-1050
[23]蔡自兴.智能控制(第二版)[M].北京:电子工业出版社,2004
[24]Dillon.T.S,Expert System Applications in Power Systems[M].New Jersey:Prentice Hall.1990
[25]Benbouzid M E H,Nejjari H.A simple fuzzy logic approach for induction motors stator condition monitoring[J].IEEE International Conference on Electric Machines and Drives,Cambridge,17-20June,2001:634-639
[26]王洪元,史国栋.人工神经网络技术及其应用[M].北京:中国石化出版社,2002
[27]Juric M.Optimizing genetic algorithm parameters for multiple fault diagnosis applications[J].In:Proc of the Tenth Conference on Artificial Intelligence for Applications,San Antonia,1-4 March,1994:434-440
[28]刘念.旋转整流器故障的神经网络识别研究仁[J].电力系统自动化,I 998＞22(10):31-33
[29]朱新宇.基于故障字典的旋转整流器故障检测方法[J].中国民航学院学报,1999,17(6):16-19
[30]刘念,谢驰.Fuzzy模式识别在旋转整流器故障监测中的应用[J].四川联合大学学报(工程科学版),1999,3(1):47-49
[31]张选利,蔡金锭,刘庆珍.人工智能在电力电子电路故障诊断中的应用[J].福州大学学报.2003.31(3):303-307
[32]刘念,谢驰.无刷励磁同步发电机旋转整流器故障的模糊神经网络诊断[J].继电器,31(8):8-11
[33]薄海涛,白振兴.基于故障树和神经网络的飞机电源系统故障诊断[J].自动化与仪表,2005,4:65-67
[34]程佩清.数字信号处理[M].北京:清华大学出版社,2002:138-150
[35]Sanjit K.Mitra著,孙洪 余翔宇等译,数字信号处理—基于计算机的方法[M].电子工业出版社,2005.1
[36]周浩敏.信号处理技术基础[M].北京:北京航空航天大学出版社,2001:1-5
[37]应怀樵,董书伟,应明,赵增欣.信号处理中“奇妙”的混叠现象[J].现代振动与噪声技术,2002:5-8
[38]应怀樵,沈松,刘进明.频率混叠在时域和频域现象的研究[J].振动、测试与诊断,2006第1期
[39]邱宽民,赵胜凯.DFT与FFT在实际应用时的性能比较[J]_北方交通大学学报,2000,24(5):60-62
[40]丁康,江利旗.离散频谱的能量重心校正法[J].振动工程学报,2001第3期
[41]谢明,丁康.频谱分析的校正方法[J].振动工程学报,1994,7(2):172-178.
[42]刘渝.正弦波频率快速估计方法[J].数据采集与处理,1998第1期
[43]谢明,张晓飞,丁康.频谱分析中用于相位和频率校正的相位差校正法[J].振动工程学报1999第4期
[44]丁康,钟舜聪.通用的离散频谱相位差校正方法fJ].电子学报,2003第1期
[45]Svend Gand,Henrik Herlufsen "windows to FFT Analysis" Bruel&Kjer Book ISSN 007-2621 1987No.3
[46]王涛.基于免疫算法的无刷同步发电机旋转整流器故障诊断技术研究[J],四川大学硕士学位论文
[47]黄耀群,李兴源.同步电机现代励磁系统及其控制[M].成都:成都科技大学出版社,1993
[48]刘念,谢驰,无刷同步发电机旋转整流器故障监测新方法研究[J],电工技术杂志,2000(8):16-18
[49]田景文,高美娟.人工神经网络算法研究及应用[M],北京立功大学出版社,2006
[50]孙增圻等.智能控制理论与技术[M],清华大学出版社,广西科学技术出版社,1997
[51]李人厚.智能控制理论和方法[M],西安电子科技大学出版社,1999
[52]阎平凡,张长水.人工神经网络与模拟进行计算[M],清华大学出版社,2005
[53]Fredric M.Han Ivica Kostanic著,叶世伟,王海娟译,神经计算原理[M],机械工业出版社,2007
[54]虞和济,陈长征.基于神经网络的智能诊断[M],冶金工业出版社,1999
[55]马春阳,李果.基于模糊神经网络的设备故障预测研究[J].噪声与振动控制,2006(6):33-39
[56]梁中华,田茂芹.给予人工神经网络的电力电子电路故障诊断[J],沈阳工业大学学报,2004(26):140-142
[57]马皓,徐德鸿.基于神经网络的频谱分析的电力电子电路故障在线诊断[J],浙江大学学报,1999(33):664-668
[58]许宁,黄之初.神经网络在旋转机械故障诊断中的应用研究[J],矿山机械,2005(8):86-87
[59]黄德双.神经网络模式识别系统理论[M],电子工业出版社,1996
[60]飞思科技产品研发中心.神经网络技术与MATLAB7实现[M],电子工业出版社,2005
[61]孙即祥,现代模式识别[M],国防科技大学出版社,2002
[62]郭桂蓉,模糊模式识别[M],国防科技大学出版社,1993
[2]沈标正.电机故障诊断技术仁[M].北京:机械工业出版社,1996
[3]樊俊,陈忠,涂光瑜.同步发电机半导体励磁原理及应用[M].北京.水利电力出版社.1991年第二版
[4]G.F.H.Allen.Brushless Excitation Systems for Synchronous Machines[J].IEEE Transactions on Power Apparatus and Systems v PAS-91 NO.5 Septemberctober.1972.pp.1848-1854
[5]S.I.Loginov,V.I.Valkov,T.V.Bekhtereva,and E.I.Ryabenko.Investigation of Brushless Thyristor Excitation Systems for Medium- and High-Capacity Turbomotors[J].Soviet Electrical Engineering(English Translation of Elektrotekhnika)vol(49).No.2.1978.pp.49-53
[6]A.C.Williamson and B.J.Chalmers.Novel Form of Synchronous Machine Excitation[J].Soviet Electrical Engineering(English Translation of Elektrotekhnika)vol49 No3 1978 p4-9.Electric Machines and Electranechanics vol.1.No.4 July-September.1977.pp.365-376
[7]李伟清,王绍禹.发电机故障检查分析及预防[M].北京:中国电力出版社,1996.
[8]W.F.Wrkght,R Hawley and J.L.Dinely.Brushless thyristor excitation systems[J].IEEE Trans.Systems.1972.pp.1848-1851
[9]刘念.无刷励磁旋转整流器电流波形分析及其故障的微机识别[J].电力系统自动化,1994,18(4):19-26
[10]S.C.Gupta,M.L.Dewal.A Novel Technique for the Measurement of Average Rotor Temperature of Brushless Synchronous Machines[J].IEEE Trans.On Power Apparatus and Systems.Vol.PAS-98,No.4.July/Aug.1979,pp.1238-1243
[11]J.Sottle,F.C.Trutt and A.W.Leedy.Condition monitoring of brushless three-phase generators with stator winding or rotor circuit deterioration[J].Conference Record IAS Annual Meeting(IEEE Industry Applications Society)vol.3.Septernber 30-October 4.2001.pp.1587-1594
[12]黄念慈,黄山,赵莉华.旋转整流器故障检测的一种方法[J].电工技术学报,11(4)1996:50-53
[13]施惠昌,陈泉林,冯玉田,袁建华,王庭山.无刷励磁发电机转子电流及温度监测装置[J].计算机自动测量与控制.NO.5.2001
[14]陈维忠.无刷同步电机转子电流测量及旋转整流器故障检测装置的研究[J].自动化与仪表,200015(2):12-14
[15]刘念.无刷同步电机旋转整流器监测新方法研究[J],电工技术杂志,NO.8.2000
[16]Li Xingyuan,Microprocessor-based fault monitor for rotating rectifiers of brushless AC exciters using a pattern ecognition approach[J].Conference Record-IEEE Instrumentation and Measurement Technology Conference 1 May 10-12 1994.
[17]门馄,沈善德,朱守真,杨常府.综合测辨法用于无刷励磁系统故障诊断的研究[J].电力系统自动化,2001,25(24):26-29
[18]丁康,谢明,王志杰.离散频谱的幅值、相位和频率校正方法对误差分析[J].动态分析与测试技术,1996,6(2):10-29
[19]黄方能,吴玉燕,FFT谐波检测存在的问题[J],广西电力,2005(4):39-41
[20]郑恩让,杨润贤,高森,关于电力系统FFT谐波检测存在问题的研究[J],2006,34(18):52-57
[21]Tommy W S Chow,Hnng Zhou Tan.HOS—based on parametric and parametric methodologies for machine fault detection[J].IEEE Trans.on Industrial Electronics 2000 47(5):051-059
[22]Kral C,Wieser R S,Pirker F,et at.Sequences of field oriented control for the detection of faulty rotor bars in induction machines—the Vienna Monitoring Method[J].IEEE Trans.On Industrial Eiectronics,2000,47(5):1042-1050
[23]蔡自兴.智能控制(第二版)[M].北京:电子工业出版社,2004
[24]Dillon.T.S,Expert System Applications in Power Systems[M].New Jersey:Prentice Hall.1990
[25]Benbouzid M E H,Nejjari H.A simple fuzzy logic approach for induction motors stator condition monitoring[J].IEEE International Conference on Electric Machines and Drives,Cambridge,17-20June,2001:634-639
[26]王洪元,史国栋.人工神经网络技术及其应用[M].北京:中国石化出版社,2002
[27]Juric M.Optimizing genetic algorithm parameters for multiple fault diagnosis applications[J].In:Proc of the Tenth Conference on Artificial Intelligence for Applications,San Antonia,1-4 March,1994:434-440
[28]刘念.旋转整流器故障的神经网络识别研究仁[J].电力系统自动化,I 998＞22(10):31-33
[29]朱新宇.基于故障字典的旋转整流器故障检测方法[J].中国民航学院学报,1999,17(6):16-19
[30]刘念,谢驰.Fuzzy模式识别在旋转整流器故障监测中的应用[J].四川联合大学学报(工程科学版),1999,3(1):47-49
[31]张选利,蔡金锭,刘庆珍.人工智能在电力电子电路故障诊断中的应用[J].福州大学学报.2003.31(3):303-307
[32]刘念,谢驰.无刷励磁同步发电机旋转整流器故障的模糊神经网络诊断[J].继电器,31(8):8-11
[33]薄海涛,白振兴.基于故障树和神经网络的飞机电源系统故障诊断[J].自动化与仪表,2005,4:65-67
[34]程佩清.数字信号处理[M].北京:清华大学出版社,2002:138-150
[35]Sanjit K.Mitra著,孙洪 余翔宇等译,数字信号处理—基于计算机的方法[M].电子工业出版社,2005.1
[36]周浩敏.信号处理技术基础[M].北京:北京航空航天大学出版社,2001:1-5
[37]应怀樵,董书伟,应明,赵增欣.信号处理中“奇妙”的混叠现象[J].现代振动与噪声技术,2002:5-8
[38]应怀樵,沈松,刘进明.频率混叠在时域和频域现象的研究[J].振动、测试与诊断,2006第1期
[39]邱宽民,赵胜凯.DFT与FFT在实际应用时的性能比较[J]_北方交通大学学报,2000,24(5):60-62
[40]丁康,江利旗.离散频谱的能量重心校正法[J].振动工程学报,2001第3期
[41]谢明,丁康.频谱分析的校正方法[J].振动工程学报,1994,7(2):172-178.
[42]刘渝.正弦波频率快速估计方法[J].数据采集与处理,1998第1期
[43]谢明,张晓飞,丁康.频谱分析中用于相位和频率校正的相位差校正法[J].振动工程学报1999第4期
[44]丁康,钟舜聪.通用的离散频谱相位差校正方法fJ].电子学报,2003第1期
[45]Svend Gand,Henrik Herlufsen "windows to FFT Analysis" Bruel&Kjer Book ISSN 007-2621 1987No.3
[46]王涛.基于免疫算法的无刷同步发电机旋转整流器故障诊断技术研究[J],四川大学硕士学位论文
[47]黄耀群,李兴源.同步电机现代励磁系统及其控制[M].成都:成都科技大学出版社,1993
[48]刘念,谢驰,无刷同步发电机旋转整流器故障监测新方法研究[J],电工技术杂志,2000(8):16-18
[49]田景文,高美娟.人工神经网络算法研究及应用[M],北京立功大学出版社,2006
[50]孙增圻等.智能控制理论与技术[M],清华大学出版社,广西科学技术出版社,1997
[51]李人厚.智能控制理论和方法[M],西安电子科技大学出版社,1999
[52]阎平凡,张长水.人工神经网络与模拟进行计算[M],清华大学出版社,2005
[53]Fredric M.Han Ivica Kostanic著,叶世伟,王海娟译,神经计算原理[M],机械工业出版社,2007
[54]虞和济,陈长征.基于神经网络的智能诊断[M],冶金工业出版社,1999
[55]马春阳,李果.基于模糊神经网络的设备故障预测研究[J].噪声与振动控制,2006(6):33-39
[56]梁中华,田茂芹.给予人工神经网络的电力电子电路故障诊断[J],沈阳工业大学学报,2004(26):140-142
[57]马皓,徐德鸿.基于神经网络的频谱分析的电力电子电路故障在线诊断[J],浙江大学学报,1999(33):664-668
[58]许宁,黄之初.神经网络在旋转机械故障诊断中的应用研究[J],矿山机械,2005(8):86-87
[59]黄德双.神经网络模式识别系统理论[M],电子工业出版社,1996
[60]飞思科技产品研发中心.神经网络技术与MATLAB7实现[M],电子工业出版社,2005
[61]孙即祥,现代模式识别[M],国防科技大学出版社,2002
[62]郭桂蓉,模糊模式识别[M],国防科技大学出版社,1993