基于HSMM的滚动轴承故障预测技术
|
摘要
滚动轴承是旋转机械设备中最常见的部件之一,其运行状态的好坏直接影响到整个设备的工作性能。由于受到载荷、安装、润滑等因素的影响,滚动轴承在工作过程中故障率较高。因此,如何实现对滚动轴承故障的有效预测,对于尽早发现旋转机械设备的故障苗头,减少或杜绝重大事故的发生以及降低维修成本具有重要意义。
本文针对滚动轴承的故障预测问题,在分析其故障机理和演化规律的基础上,开展了基于隐半马尔可夫模型(Hidden Semi-Markov Models,HSMM)的滚动轴承故障预测技术研究。论文主要内容包括:
(1)滚动轴承的故障机理分析与故障演化建模
系统地分析了滚动轴承的故障演化机理。在对滚动轴承正常状态和故障状态的特征频率进行分析的基础上,建立了基于隐半马尔可夫模型的滚动轴承故障演化趋势模型,对滚动轴承全寿命过程中各个退化状态的驻留时间和状态转移概率进行了合理描述。
(2)基于小波能谱熵的滚动轴承故障预测特征提取方法
针对滚动轴承故障预测特征信息提取难的问题,提出了基于小波能谱熵的滚动轴承故障预测特征提取方法。验证结果表明,该方法所提取的小波能谱熵评判指标,可以较好地描述滚动轴承的全寿命过程中的故障演化趋势。
(3)基于隐半马尔可夫模型的滚动轴承故障预测方法
针对隐半马尔可夫模型算法中存在的参数设置不确定性、多样本训练下溢等问题,深入研究了隐半马尔可夫模型的改进算法。在此基础上,提出了以小波能谱熵为预测特征信息的隐半马尔可夫模型故障预测方法。
(4)实验方案设计与实验验证
以滚动轴承实验台为对象,设计了故障预测的实验方案,并通过实验验证了本文所研究的基于隐半马尔可夫模型的滚动轴承故障预测方法,结果表明,该方法可以较准确的实现对滚动轴承的故障预测,具有良好的可行性和有效性。
Rolling element bearings are of paramount importance to almost all rotating machinery. As a consequence of their importance and wide spread use, Rolling element bearing failure is one of the foremost causes of breakdowns in rotating machinery. There are several reasons that a bearing fails, such as improper lubrication and mounting, inverse environment, overload, fatigue, etc. So the fault prognostics of rolling element bearings is important to detect the incipient faults, to optimize maintenance scheduling, to avoid catastrophic failure, to extend machinery life and to reduce costs.
The purpose of the research is aiming to solve the problem of fault prognostics of rolling element bearings. The mechanisms of the faults and the trends of failure process are analyzed systematically. Hence a fault prognostic method of rolling element bearings based on the Hidden Semi-Markov Models(HSMM) is studied in this thesis. The main contents of the research are as follows.
1. The mechanism of the faults of rolling element bearings is analyzed and the model of failure process trend is built.
Firstly, the mechanism of the faults of rolling element bearings is analyzed systematically. Secondly, a Hidden Semi-Markov Model which describles the failure process trend of rolling element bearings is built. Finally, the description of the state duration and the state transition all the life time are obtained.
2. A prognostic feature information extraction scheme based on Wavelet Energy Entropy of rolling element bearings is proposed.
A prognostic feature information extraction scheme is proposed based on Wavelet Energy Entropy in order to solve the problem of the difficulty to extract a prognostic feature information of rolling element bearing. According to the test result, the failure process of rolling element bearings all the life can be describled through the feature extraction method.
3. A fault prognostic method of rolling element bearings based on HSMM is studied.
In order to solve the problem such as the uncetainty of parameter setting, avoiding underflow due to the multi-samples training, a modified algorithm of HSMM is studied in detail. Based on the above researchs, a fault prognostic method of rolling element bearings based on HSMM is studied through taking the Wavelet Energy Entropy of viberation as the prognostic feature information.
4. Experiment and validation.
An experimental plan of rolling element bearings is designed, and the fault prognostic technology of rolling element bearings based on HSMM is validated. Experimental results show that this proposed method is effective and practicable to prognosticate fault and the remaining life of rolling element bearings.
本文针对滚动轴承的故障预测问题,在分析其故障机理和演化规律的基础上,开展了基于隐半马尔可夫模型(Hidden Semi-Markov Models,HSMM)的滚动轴承故障预测技术研究。论文主要内容包括:
(1)滚动轴承的故障机理分析与故障演化建模
系统地分析了滚动轴承的故障演化机理。在对滚动轴承正常状态和故障状态的特征频率进行分析的基础上,建立了基于隐半马尔可夫模型的滚动轴承故障演化趋势模型,对滚动轴承全寿命过程中各个退化状态的驻留时间和状态转移概率进行了合理描述。
(2)基于小波能谱熵的滚动轴承故障预测特征提取方法
针对滚动轴承故障预测特征信息提取难的问题,提出了基于小波能谱熵的滚动轴承故障预测特征提取方法。验证结果表明,该方法所提取的小波能谱熵评判指标,可以较好地描述滚动轴承的全寿命过程中的故障演化趋势。
(3)基于隐半马尔可夫模型的滚动轴承故障预测方法
针对隐半马尔可夫模型算法中存在的参数设置不确定性、多样本训练下溢等问题,深入研究了隐半马尔可夫模型的改进算法。在此基础上,提出了以小波能谱熵为预测特征信息的隐半马尔可夫模型故障预测方法。
(4)实验方案设计与实验验证
以滚动轴承实验台为对象,设计了故障预测的实验方案,并通过实验验证了本文所研究的基于隐半马尔可夫模型的滚动轴承故障预测方法,结果表明,该方法可以较准确的实现对滚动轴承的故障预测,具有良好的可行性和有效性。
Rolling element bearings are of paramount importance to almost all rotating machinery. As a consequence of their importance and wide spread use, Rolling element bearing failure is one of the foremost causes of breakdowns in rotating machinery. There are several reasons that a bearing fails, such as improper lubrication and mounting, inverse environment, overload, fatigue, etc. So the fault prognostics of rolling element bearings is important to detect the incipient faults, to optimize maintenance scheduling, to avoid catastrophic failure, to extend machinery life and to reduce costs.
The purpose of the research is aiming to solve the problem of fault prognostics of rolling element bearings. The mechanisms of the faults and the trends of failure process are analyzed systematically. Hence a fault prognostic method of rolling element bearings based on the Hidden Semi-Markov Models(HSMM) is studied in this thesis. The main contents of the research are as follows.
1. The mechanism of the faults of rolling element bearings is analyzed and the model of failure process trend is built.
Firstly, the mechanism of the faults of rolling element bearings is analyzed systematically. Secondly, a Hidden Semi-Markov Model which describles the failure process trend of rolling element bearings is built. Finally, the description of the state duration and the state transition all the life time are obtained.
2. A prognostic feature information extraction scheme based on Wavelet Energy Entropy of rolling element bearings is proposed.
A prognostic feature information extraction scheme is proposed based on Wavelet Energy Entropy in order to solve the problem of the difficulty to extract a prognostic feature information of rolling element bearing. According to the test result, the failure process of rolling element bearings all the life can be describled through the feature extraction method.
3. A fault prognostic method of rolling element bearings based on HSMM is studied.
In order to solve the problem such as the uncetainty of parameter setting, avoiding underflow due to the multi-samples training, a modified algorithm of HSMM is studied in detail. Based on the above researchs, a fault prognostic method of rolling element bearings based on HSMM is studied through taking the Wavelet Energy Entropy of viberation as the prognostic feature information.
4. Experiment and validation.
An experimental plan of rolling element bearings is designed, and the fault prognostic technology of rolling element bearings based on HSMM is validated. Experimental results show that this proposed method is effective and practicable to prognosticate fault and the remaining life of rolling element bearings.
引文
[1] HessA, Fila L. Prognostics from the need to reality from the fleet users and PHM system designer/developers perspectives[A]. Proceedings of IEEE Aerospace Conference[C], 2002:2791~2797.
[2] Hardman W, Hess A, Blunt D. A USN development strategy and demonstration results for propulsion and mechanical systems diagnostics. prognostics and health management[A]. Proceedings of the 2001 IEEE Aerospace Conference[C], 2001: 3059~3068.
[3] Engel S J, Gilmartin B J, Bongort K, Hess A. Prognostics the real issues involved with predicting life remaining [A]. Proceedings of the 2000 IEEE Aerospace Conference[C], 2000: 457~469.
[4]梁旭,李行善,张磊,于劲松.支持视情维修的故障预测技术研究[J].测控技术,2007,26(6):5~8.
[5]杨江天,岳维亮.灰色模型在机械故降预测中的应用[J].机械强度,2001,23(3):277~279.
[6] Gill J . Lessons learned from rotary and fixed wing HUMS applications [A].Proceedings of IEEE Aerospace Conference [C], 2000: 423~431.
[7] Greitzer F L, Pawlowski R A . Embedded prognostics health monitoring [A].Proceedings of the International Instrumentation Symposium[C], 2002: 301~310.
[8] Qiu J, Zhang C, Seth B B, Liang S Y.Damage mechanics approach for bearing lifetime prognostics [J].Mechanical Systems and Signal Processing, 2002,16(5): 817~829.
[9] Ray A, Tangirala S.Stochastic modeling of fatigue crack dynamics for online failure prognostics[J] . IEEE Transactions on Control Systems Technology, 1996,4(4):443~451.
[10] Lembessis E,Antonopoulos G, King R E,Halatsis C, Torres J.An online expert system for fault prognosis, computer integrated manufacturing[A].Proceedings of the 5th CIM Europe Conference[C], 1989:371~377.
[11] Biagetti T, Sciubba E. Automatic diagnostics and prognostics of energy conversion processes via knowledge based systems[J].Energy, 2004,29(15):2553~2572.
[12] Garga A K, McClintic K T, Campbell R L, Yang C C, Lebold M S, Hay T A, Byington C S, Hybrid reasoning for prognostic learning in CBM systems[A].Proceedings of IEEE Aerospace Conference [C], 2001:2957~2969.
[13]秦俊奇,曹立军,等.基于动态模糊综合评判的故障预测方法[J].计算机工程, 2005,31(12):172~174.
[14]薛子云,杨江天,朱衡君.机械故障预测模型综述[J].机械强度,2006,28(5):60~65.
[15] Zhang S, Ganesan R. Multivariable trend analysis using neural networks for intelligent diagnostics of rotating machinery [J].Transactions of the ASME,Journal of Engineering for Gas Turbines and Power, 1997,119(2):378~384.
[16] Yam R CM, Tse PW, LiL, Tu P. Intelligent predictive decision support system for condition based maintenance[J].International Journal of Advanced Manufacturing Technology, 2001,17(5):383~391.
[17] Wang P, Vachtsevanos G. Fault prognostics using dynamic wavelet neural networks[A].IEEE Proceedings of Systems Readiness Technology Conference[C], 2001:857~870.
[18] Wang W Q, GolnaraghiM F, Ismail F. Prognosis of machine health condition using neuro fuzzy systems[J] . Mechanical Systems and Signal Processing, 2004,18(4): 813~831.
[19]李富才,何正嘉,陈进.小波域相关滤波法及其早期故障预示应用[J].振动工程学报,2005,18(2)145～148.
[20]邓士娟.基于数据挖掘技术的轴承寿命预测的研究[硕士学位论文].大连:大连海事大学,2006.3.
[21]薛子云.灰色系统和时间序列分析在机车轴承故障预测中的应用研究[硕士学位论文].北京:北京交通大学,2006.3.
[22]刘源泂.低速重载滚动轴承的状态监测与寿命预测方法研究[硕士学位论文].武汉:武汉科技大学,2006.5.
[23]谢锦辉.隐Markov模型(HMM)及其在语言处理中的应用[D].华中理工大学:机械与能源工程学院,1995:39~41.
[24] Qiu H, Liao H T, Lee J. Degradation assessment for machinery prognostics using hidden Markov models[A]. Proceedings of the ASME International Design Engineering Technical Conferences and Computer sand Information in Engineering Conference(DETC2005)[C], 2005:531~537.
[25] Baruah P, Chinnam R B. HMMs for diagnostics and prognostics in machining processes[J]. International Journal of Production Research, 2005,43(6):1275~1293.
[26] Ming Dong, David He, Prashant Banerjee, Jonathan Keller. Equipment health diagnosis and prognosis using hidden semi-Markov Models[J]. Springer-Verlag London Limited(2005), 2005:1~8
[27] Ming Dong, Dvid He. Hidden semi-Markov model-based methodology for Multi-senor equipment health diagnosis and prognosis[J]. European Journal Operational Research, 2006:858~878.
[28] M.Dong, D.He. A segmental hidden semi-Markov models (HSMM)–based diagnostics and prognostics framework and methodology[J]. Mechanical Systems and Signal Processing(2006), 2006:1~19.
[29]盛兆顺,尹琦岭.设备状态监测与故障诊断技术及应用[M].化学工业出版社,2003:68~70.
[30]董建宁,申永军,杨绍普.轴承故障诊断中的信号处理技术研究与展望[J].轴承,2005,35(12):95~l00.
[31]程军圣.基于小波分析的滚动轴承故障诊断研究[硕士学位论文].长沙:湖南大学,2000.
[32]梅宏斌.滚动轴承振动监测与诊断理论、方法、系统[M].机械工业出版社,1995:102~105.
[33]印欣运,何永勇,彭志科,褚福磊.小波熵及其在状态趋势分析中的应用[J].振动工程学报,2004,17(2):165~169.
[34]徐金梧,徐科.小波变换在滚动轴承故障诊断中的应用[J].机械工程学报,1997,33(4):50~55.
[35]郭晶,陆舒. MATLAB6.5辅助小波分析与应用[M].北京:电子工业出版社,2003:106~120.
[36]何正友,蔡玉梅,钱清泉.小波熵理论及其在电力系统故障检测中的应用研究[J].中国电机工程学报,2005,25(5):38~43.
[37]崔景泰.小波分析导论.西安:西安交通大学出版社[M],1997:70~78.
[38]飞思科技产品研发中心.小波分析理论与实现[M],北京:电子工业出版社,2005:29~60.
[39]杨福生.小波变换的工程分析与应用[M].北京:科学出版社,2000, 11.
[40]胡昌华,张军波,夏军等.基于MATLAB的系统分析与设计—小波分析[M].西安:西安电子科技大学出版社,2000:88~96.
[41]冉启文,谭立英.小波分析与分数傅立叶变换及应用[M],北京:国防工业出版社,2002,49~58.
[42]刘希强,周惠兰,郑治真等.基于小波包变换的滤波方法[J].西北地震学报,1999,21(3):248~253.
[43]张佩瑶,马孝江,王吉军等.小波包信号提取算法及其在故障诊断中的应用[J].大连理工大学学报,1997,37(1):67~72.
[44]杨国安,钟秉林,黄仁,贾民平,许云飞.机械故障信号小波包分解的时域特征提取方法研究振动与冲击[J],2001,20(2):25~32.
[45]傅祖芸.信息论-基础理论与应用[M].北京:电子工业出版社,2001:24~25.
[46]张贤达.现代信号处理方法[M].北京:清华大学出版社,2002:102~103.
[47] Quiroga R Q, Rosso O A, Basar E et al.Wavelet entropy in event-ralated potential:a new method shows ordering of EEG oscillations[J].Biological Cybernetics,2001,84(4):291~299.
[48]封洲燕.应用小波熵分析大鼠脑电信号的动态特性[J].生物物理学报,2002,18(3):325~330.
[49]张文炬,苏清祖.车辆变速器故障诊断的Shannon熵研究[J].农业机械学报,2002,33(1):80~83.
[50] Bahl L R, Jelinek F. Decoding for Channels with Insertions, Deletions, and Substitution Applications to Speech Recognition[J]. IEEE Trans.IT, 1975,2l(2): 401~411.
[51] Baker J K. The DRAGON System-An Overview[J]. IEEE Trans. ASSP 1975, 23(1):24~29.
[52] L.R. Rabiner. A tutorial on Hidden Markov Models and selected application in speech recognition[J], Proc. IEEE, 1989,77(2):257~286.
[53] Nadas A,et al. On a Model-Robust Training Method for Speech Recognition[J]. IEEE Trains. ASSP98, 1988,36(9):1432~1436.
[54] Rabiner L R,Juang B H. An Introduction to Hidden Markov Models[J]. IEEE ASSP Magazine. 1986,3(1):4~6.
[55] Rabiner L R, et al. On the Use of Hidden Markov Models for Speaker Independent Recognition of Isolated Word from a Medium-Size Vocabulary[J]. AT&T Bell Tab. Tech.J, 1984,63(4):627~642.
[56] Huang X D. Phoneme Classification Using Semicontinuous Hidden Markov Models[J]. IEEE Trans. SP, 1992,40(5):1062~1067.
[57] Huang X D, Ariki Y, Jack M A. Hiddened Markov Models for Speech Recognition[M]. Edinburgh University Press, 1990:456~467.
[58] Huang X D, Jack M A. Unified Modeling of Vector Quantization and Hidden markov Model Using Semi-Continuous Hideden Morkov models[J]. Proc. ICASSP89, 1989:639~642.
[59]姚天任.数字信号处理.华中理工大学出版社[M],1991:316~346.
[60] S. E. Levinson. Continuously variable duration hidden Markov models for automatic speech recognition[J]. Computer Speech and Language, 1986(1):29~45.
[61] M. Ostendorf, V. Digalakis, and O. Kimball . From HMMs to segment models[J].IEEE Transactions on Speech & Audio Processing, 1996,4(5):360~378.
[62] Keiichiro Oura, Heiga Zen , Yoshihiko Nankaku, Akinobu Lee, Keiichi Tokuda.Hidden Semi-Markov Models based speech recognition system using weighted finite-state transducer[J].Proc. ICASSP, 2006(1):33~36,2006.
[63] Kevin P.Murphy.Hidden semi-Markov models(HSMM).www.ai.mit.edu/～murphyk. 2002.9.
[64]中国机械工程学会,范厚军.紧固件手册.南昌:江西科学技术出版社,2004.1.
[65]中国机械工程学会,中国机械工程大典编委会.中国机械设计大典:第二卷,机械设计基础,南昌:江西科学技术出版社,2002.1.
[2] Hardman W, Hess A, Blunt D. A USN development strategy and demonstration results for propulsion and mechanical systems diagnostics. prognostics and health management[A]. Proceedings of the 2001 IEEE Aerospace Conference[C], 2001: 3059~3068.
[3] Engel S J, Gilmartin B J, Bongort K, Hess A. Prognostics the real issues involved with predicting life remaining [A]. Proceedings of the 2000 IEEE Aerospace Conference[C], 2000: 457~469.
[4]梁旭,李行善,张磊,于劲松.支持视情维修的故障预测技术研究[J].测控技术,2007,26(6):5~8.
[5]杨江天,岳维亮.灰色模型在机械故降预测中的应用[J].机械强度,2001,23(3):277~279.
[6] Gill J . Lessons learned from rotary and fixed wing HUMS applications [A].Proceedings of IEEE Aerospace Conference [C], 2000: 423~431.
[7] Greitzer F L, Pawlowski R A . Embedded prognostics health monitoring [A].Proceedings of the International Instrumentation Symposium[C], 2002: 301~310.
[8] Qiu J, Zhang C, Seth B B, Liang S Y.Damage mechanics approach for bearing lifetime prognostics [J].Mechanical Systems and Signal Processing, 2002,16(5): 817~829.
[9] Ray A, Tangirala S.Stochastic modeling of fatigue crack dynamics for online failure prognostics[J] . IEEE Transactions on Control Systems Technology, 1996,4(4):443~451.
[10] Lembessis E,Antonopoulos G, King R E,Halatsis C, Torres J.An online expert system for fault prognosis, computer integrated manufacturing[A].Proceedings of the 5th CIM Europe Conference[C], 1989:371~377.
[11] Biagetti T, Sciubba E. Automatic diagnostics and prognostics of energy conversion processes via knowledge based systems[J].Energy, 2004,29(15):2553~2572.
[12] Garga A K, McClintic K T, Campbell R L, Yang C C, Lebold M S, Hay T A, Byington C S, Hybrid reasoning for prognostic learning in CBM systems[A].Proceedings of IEEE Aerospace Conference [C], 2001:2957~2969.
[13]秦俊奇,曹立军,等.基于动态模糊综合评判的故障预测方法[J].计算机工程, 2005,31(12):172~174.
[14]薛子云,杨江天,朱衡君.机械故障预测模型综述[J].机械强度,2006,28(5):60~65.
[15] Zhang S, Ganesan R. Multivariable trend analysis using neural networks for intelligent diagnostics of rotating machinery [J].Transactions of the ASME,Journal of Engineering for Gas Turbines and Power, 1997,119(2):378~384.
[16] Yam R CM, Tse PW, LiL, Tu P. Intelligent predictive decision support system for condition based maintenance[J].International Journal of Advanced Manufacturing Technology, 2001,17(5):383~391.
[17] Wang P, Vachtsevanos G. Fault prognostics using dynamic wavelet neural networks[A].IEEE Proceedings of Systems Readiness Technology Conference[C], 2001:857~870.
[18] Wang W Q, GolnaraghiM F, Ismail F. Prognosis of machine health condition using neuro fuzzy systems[J] . Mechanical Systems and Signal Processing, 2004,18(4): 813~831.
[19]李富才,何正嘉,陈进.小波域相关滤波法及其早期故障预示应用[J].振动工程学报,2005,18(2)145～148.
[20]邓士娟.基于数据挖掘技术的轴承寿命预测的研究[硕士学位论文].大连:大连海事大学,2006.3.
[21]薛子云.灰色系统和时间序列分析在机车轴承故障预测中的应用研究[硕士学位论文].北京:北京交通大学,2006.3.
[22]刘源泂.低速重载滚动轴承的状态监测与寿命预测方法研究[硕士学位论文].武汉:武汉科技大学,2006.5.
[23]谢锦辉.隐Markov模型(HMM)及其在语言处理中的应用[D].华中理工大学:机械与能源工程学院,1995:39~41.
[24] Qiu H, Liao H T, Lee J. Degradation assessment for machinery prognostics using hidden Markov models[A]. Proceedings of the ASME International Design Engineering Technical Conferences and Computer sand Information in Engineering Conference(DETC2005)[C], 2005:531~537.
[25] Baruah P, Chinnam R B. HMMs for diagnostics and prognostics in machining processes[J]. International Journal of Production Research, 2005,43(6):1275~1293.
[26] Ming Dong, David He, Prashant Banerjee, Jonathan Keller. Equipment health diagnosis and prognosis using hidden semi-Markov Models[J]. Springer-Verlag London Limited(2005), 2005:1~8
[27] Ming Dong, Dvid He. Hidden semi-Markov model-based methodology for Multi-senor equipment health diagnosis and prognosis[J]. European Journal Operational Research, 2006:858~878.
[28] M.Dong, D.He. A segmental hidden semi-Markov models (HSMM)–based diagnostics and prognostics framework and methodology[J]. Mechanical Systems and Signal Processing(2006), 2006:1~19.
[29]盛兆顺,尹琦岭.设备状态监测与故障诊断技术及应用[M].化学工业出版社,2003:68~70.
[30]董建宁,申永军,杨绍普.轴承故障诊断中的信号处理技术研究与展望[J].轴承,2005,35(12):95~l00.
[31]程军圣.基于小波分析的滚动轴承故障诊断研究[硕士学位论文].长沙:湖南大学,2000.
[32]梅宏斌.滚动轴承振动监测与诊断理论、方法、系统[M].机械工业出版社,1995:102~105.
[33]印欣运,何永勇,彭志科,褚福磊.小波熵及其在状态趋势分析中的应用[J].振动工程学报,2004,17(2):165~169.
[34]徐金梧,徐科.小波变换在滚动轴承故障诊断中的应用[J].机械工程学报,1997,33(4):50~55.
[35]郭晶,陆舒. MATLAB6.5辅助小波分析与应用[M].北京:电子工业出版社,2003:106~120.
[36]何正友,蔡玉梅,钱清泉.小波熵理论及其在电力系统故障检测中的应用研究[J].中国电机工程学报,2005,25(5):38~43.
[37]崔景泰.小波分析导论.西安:西安交通大学出版社[M],1997:70~78.
[38]飞思科技产品研发中心.小波分析理论与实现[M],北京:电子工业出版社,2005:29~60.
[39]杨福生.小波变换的工程分析与应用[M].北京:科学出版社,2000, 11.
[40]胡昌华,张军波,夏军等.基于MATLAB的系统分析与设计—小波分析[M].西安:西安电子科技大学出版社,2000:88~96.
[41]冉启文,谭立英.小波分析与分数傅立叶变换及应用[M],北京:国防工业出版社,2002,49~58.
[42]刘希强,周惠兰,郑治真等.基于小波包变换的滤波方法[J].西北地震学报,1999,21(3):248~253.
[43]张佩瑶,马孝江,王吉军等.小波包信号提取算法及其在故障诊断中的应用[J].大连理工大学学报,1997,37(1):67~72.
[44]杨国安,钟秉林,黄仁,贾民平,许云飞.机械故障信号小波包分解的时域特征提取方法研究振动与冲击[J],2001,20(2):25~32.
[45]傅祖芸.信息论-基础理论与应用[M].北京:电子工业出版社,2001:24~25.
[46]张贤达.现代信号处理方法[M].北京:清华大学出版社,2002:102~103.
[47] Quiroga R Q, Rosso O A, Basar E et al.Wavelet entropy in event-ralated potential:a new method shows ordering of EEG oscillations[J].Biological Cybernetics,2001,84(4):291~299.
[48]封洲燕.应用小波熵分析大鼠脑电信号的动态特性[J].生物物理学报,2002,18(3):325~330.
[49]张文炬,苏清祖.车辆变速器故障诊断的Shannon熵研究[J].农业机械学报,2002,33(1):80~83.
[50] Bahl L R, Jelinek F. Decoding for Channels with Insertions, Deletions, and Substitution Applications to Speech Recognition[J]. IEEE Trans.IT, 1975,2l(2): 401~411.
[51] Baker J K. The DRAGON System-An Overview[J]. IEEE Trans. ASSP 1975, 23(1):24~29.
[52] L.R. Rabiner. A tutorial on Hidden Markov Models and selected application in speech recognition[J], Proc. IEEE, 1989,77(2):257~286.
[53] Nadas A,et al. On a Model-Robust Training Method for Speech Recognition[J]. IEEE Trains. ASSP98, 1988,36(9):1432~1436.
[54] Rabiner L R,Juang B H. An Introduction to Hidden Markov Models[J]. IEEE ASSP Magazine. 1986,3(1):4~6.
[55] Rabiner L R, et al. On the Use of Hidden Markov Models for Speaker Independent Recognition of Isolated Word from a Medium-Size Vocabulary[J]. AT&T Bell Tab. Tech.J, 1984,63(4):627~642.
[56] Huang X D. Phoneme Classification Using Semicontinuous Hidden Markov Models[J]. IEEE Trans. SP, 1992,40(5):1062~1067.
[57] Huang X D, Ariki Y, Jack M A. Hiddened Markov Models for Speech Recognition[M]. Edinburgh University Press, 1990:456~467.
[58] Huang X D, Jack M A. Unified Modeling of Vector Quantization and Hidden markov Model Using Semi-Continuous Hideden Morkov models[J]. Proc. ICASSP89, 1989:639~642.
[59]姚天任.数字信号处理.华中理工大学出版社[M],1991:316~346.
[60] S. E. Levinson. Continuously variable duration hidden Markov models for automatic speech recognition[J]. Computer Speech and Language, 1986(1):29~45.
[61] M. Ostendorf, V. Digalakis, and O. Kimball . From HMMs to segment models[J].IEEE Transactions on Speech & Audio Processing, 1996,4(5):360~378.
[62] Keiichiro Oura, Heiga Zen , Yoshihiko Nankaku, Akinobu Lee, Keiichi Tokuda.Hidden Semi-Markov Models based speech recognition system using weighted finite-state transducer[J].Proc. ICASSP, 2006(1):33~36,2006.
[63] Kevin P.Murphy.Hidden semi-Markov models(HSMM).www.ai.mit.edu/～murphyk. 2002.9.
[64]中国机械工程学会,范厚军.紧固件手册.南昌:江西科学技术出版社,2004.1.
[65]中国机械工程学会,中国机械工程大典编委会.中国机械设计大典:第二卷,机械设计基础,南昌:江西科学技术出版社,2002.1.