摘要
飞行器包括航天飞机、运载火箭、航空固定翼飞行器和低空旋翼直升机等等。飞行器健康管理技术对于监测飞行器关键部件工作状态、评估其状态退化趋势、预测其剩余使用寿命、保证其安全运行具有重要意义,是执行状态基维修策略乃至自治性维修保障的基础和前提。时有发生的航空航天事故也不断地给人们以警醒:开展飞行器健康管理关键技术研究的需求十分迫切。
在飞行器健康管理体系中,监测是一个极其重要的主题,也是诊断、预测和减缓等其它主题的基础。但在飞行器状态监测的实施过程中,常常存在先验知识缺乏、故障样本稀缺和故障模式不完备的问题。针对这些问题,本文以某型液体火箭发动机的涡轮泵为研究对象,开展了面向飞行器健康管理的新异类检测方法研究,主要研究工作包括:
1.深入分析了提高新异类检测器推广能力需要考虑的因素。较为系统地研究了常见新异类检测方法的基本原理,对一些性能优异的新异类检测器进行了数值仿真分析。面向健康管理,总结了新异类检测器的设计原则,提出了三种借助新异类检测开展状态监测的故障检测策略。
研究表明,新异类检测可以依据对已知观测样本的学习,实现对已知和未知异常的识别。新异类检测器的设计应综合考虑稳健性的折衷、推广性能、计算的复杂度和适用于不同情况的故障检测策略等。
2.深入研究了涡轮泵状态监测的单类支持向量机新异类检测方法。
(1)详细介绍了两种具有不同几何解释的单类支持向量机基本原理,结合数值仿真,深入分析了单类支持向量机的关键参数对其性能的影响,并据此给出了单类支持向量机故障检测方法的参数设置原则。
(2)应用单类支持向量机对涡轮泵历史试车数据进行了检测分析,结合单类支持向量机的参数设置原则和对涡轮泵试车数据分析的交叉验证,优化了单类支持向量机的参数,降低了虚警率。研究表明,单类支持向量机的性能与其参数设置密切相关;单类支持向量机的原理性误差和特征量的一致性不足会造成较大的检测虚警,通过增大高斯带宽系数和使用偏置缩放因子可以显著地降低虚警率。
3.提出了一种用于涡轮泵试车数据在线检测的双偏置单类支持向量机自适应在线检测算法。
(1)提出了用于求解单类支持向量机的序贯最小优化算法,降低了求解单类支持向量机的时间和空间复杂度。
(2)在序贯最小优化算法的基础上,提出了一种双偏置单类支持向量机自适应在线检测算法,消除了单类支持向量机的原理性误差和因工况变化等外界随机因素造成的误差,剔除了异类样本对检测模型自适应更新的贡献,避免了检测器随异类样本自适应更新的现象。
(3)利用提出的双偏置单类支持向量机自适应在线检测算法对涡轮泵历史试车数据进行了检测分析,同时也完成了对该方法的验证。
研究表明,序贯最小优化算法可以显著提高单类支持向量机的训练效率,双偏置单类支持向量机自适应在线检测算法能够在无虚警的情况下有效识别涡轮泵的异常状态,包括叶片断裂/脱落和转子碰摩。
4.提出了一种基于增量聚类和单类支持向量机的完备样本区域描述方法,用于新异类检测和涡轮泵状态监测。
(1)提出了一种具有样本压缩功能的增量聚类算法,用于解决完备样本区域描述所涉及的大样本学习问题。
(2)提出了一种集成增量聚类和单类支持向量机的完备样本区域描述方法,建立了涡轮泵时域统计特征的完备样本区域描述模型。
(3)构建了涡轮泵故障检测系统,将提出的完备样本区域描述方法在该系统中进行了集成,使用涡轮泵历史试车数据验证了该方法的有效性。
研究表明,集成增量聚类和单类支持向量机的完备样本区域描述方法能够从大量样本中增量式地提取均匀分布、张满目标区域且大小可控的代表样本集,实现样本集的压缩;能够建立用以描述完备样本分布区域的边界。对涡轮泵试车数据的检测结果表明,该方法能够在无虚警的情况下有效识别涡轮泵的叶片断裂/脱落、转子碰摩、传感器失效等故障。
5.提出了基于并联单类支持向量机和串联单类支持向量机的故障诊断方法,将新异类检测拓展应用到了故障诊断,使用涡轮泵试车数据验证了单类支持向量机故障诊断方法的有效性。
研究表明,与各种基于支持向量机两类分类器的故障诊断方法相比,单类支持向量机故障诊断方法的训练样本重复使用率更小,诊断效率更高,扩展性更好,而且能够识别未知状态和已知状态。仿真数据及涡轮泵试车数据的分析结果表明了该方法的有效性。
Health management (HM) technologies for flight vehicle, such as space shuttle, carrier rocket, fixed-wing aircraft and helicopter, are of significance to enable operating condition monitoring, degradation trend assessment, remaining useful life prediction and safe operation assurance. Thus HM is considered the premise and foundation that supposes condition based maintenance, even autonomic logistics. Occasionally occurred accidents warn us constantly that it is imperative to research on key techniques of HM.
Condition monitoring is an important theme in the architecture of vehicle HM. And it provides foundation for other themes including diagnosis, prognosis and mitigation. One of the challenges in condition monitoring of flight vehicles or their key components is lack of fault samples and prior knowledge about fault modes. Accordingly, this dissertation takes the turbopump of a liquid rocket engine as object and researches on novelty detection methods oriented to vehicle HM. The detailed contents and innovative work can be summarized as follows.
1. Facts associated with the generalization of novelty detection methods are deeply analyzed. Philosophy of popular novelty detection methods is systematically studied and some excellent methods of them are analyzed via numerical simulation. Oriented to HM, design principles are summarized and three strategies for fault detection with novelty detection methods are presented.
The research shows that novelty detection is able to recognize known and unknown abnormities according to known samples. Some issues, such as robustness trade-off, generalization, computational complexity, detection strategies that suit for different cases, should be considered when develop novelty detectors.
2. One-class support vector machine (OCSVM) based novelty detection method is deeply studied for turbopump condition monitoring.
(1) Two kinds of OCSVMs with different geometrical explanations are introduced in detail. The influence of OCSVM’s parameters to its performance is thoroughly analyzed integrating numerical simulation. And principles to set OCSVM’s parameters are given accordingly.
(2) OCSVM based novelty detector is applied to detect historical test data of the turbopump. OCSVM’s parameters are optimized according to their set principles and cross-validation of detection results, and by which false alarms are reduced.
The above research shows that OCSVM’s performances are closely related to its parameter set. The principle error of OCSVM and the bad consistency of feature vectors are main factors that cause false alarms. These false alarms can be reduced evidently by increasing OCSVM’s Gaussian kernel width and by introducing an offset scaling parameter.
3. An online adaptive novelty detection algorithm based on double-offset OCSVM is presented and applied to the detection of turbopump real test data.
(1) A sequential minimal optimization (SMO) algorithm is introduced to solve the quadratic optimization problem in OCSVM, which reduced the computational complexity of OCSVM.
(2) A double-offset OCSVM online detection algorithm based on SMO algorithm is developed. In this algorithm, detection error caused with OCSVM itself and changing environmental conditions is eliminated. Abnormal samples detected are prevented from contributing to the adaptive update of the detection model.
(3) The online detection algorithm presented above is applied to turbopump historical test data detection, which also verifies this algorithm.
The research shows that SMO algorithm can improve the training efficiency of OCSVM evidently. The online detection algorithm based on double-offset OCSVM is able to detect kinds of novel events of the turbopum, involving vane shedding and rub-impact. And there are not any false alarms.
4. A complete sample region description (CSRD) method that integrates OCSVM with incremental clustering is presented for novelty detection and turbopump condition monitoring.
(1) An incremental clustering algorithm that enables sample compression is presented to solve the learning problem of large samples in CSRD.
(2) A CSRD method integrating OCSVM with incremental clustering is presented and a complete description model is established for the time-domain statistical features of the turbopump.
(3) A turbpump fault detection system is constructed and the CSRD method is integrated into the system. Validity of the CSRD method is validated with turbopump historical test data.
The research shows that the CSRD method is able to extract a more representative sample set from large rude samples incrementally. The representative set distributes uniformly and covers the entire target region. And the size of the representative set is under control. The CSRD method is also able to generate a boundary surrounding the target region. Detection results of turbopump historical test data demonstrate that the method can identify different spikes in vibration signals caused by abnormal events such as vane shedding, rub-impact and sensor faults. And there are not any false alarms.
5. Novelty detection is exploited to fault diagnosis. Diagnosis methods based on parallel OCSVMs and series OCSVMs are presented and validated.
The research shows that compared with diagnosis methods based on SVM, OCSVM based methods have better diagnosis efficiency and expansibility. And they are able to recognize known and unknown conditions. Their validity is verified with the classification of simulation data and the classification of turbopump historical test data.
引文
[1]史上重大航天事故一览.世界科技报道,2009年2月25日:http://tech.icxo.com/htmlnews/2009/02/25/1342603_1.htm.
[2]背景资料:人类历史上的航天灾难一览.新浪网:http://news.sina.com.cn/w/2003-08-23/08331599952.shtml.
[3]搜狐新闻专题——法航客机坠毁.搜狐网:http://news.sohu.com/s2009/fahangkejishizong.
[4]胡茑庆,高经纬,胡雷等.飞行器健康管理[C].国防科学技术大学科学技术委员会年会论文集,2009.
[5] Michael Pecht. Prognostics and Health Management of Electronics. John Wiley & Sons. Inc., Hoboken, New Jersey, 2008.
[6] Stephen Johnson. Introduction to System Health Engineering and Management in Aerospace. First Internaltional Forum on Integrated System Health Engineering and Management in Aerospace. Napa, California, USA, November 7-10, 2005.
[7] Hess A. Joint Strike Fighter– Real Prognostics– Challenges, Issues, and Lessons Learned: Chasing the Big“P”. First Internaltional Forum on Integrated System Health Engineering and Management in Aerospace. Napa, California, USA, November 7-10, 2005.
[8]张宝珍.国外综合诊断、预测与健康管理技术的发展及应用[J].计算机测量与控制,2008,16(5):591-594.
[9] Fiorucci T R, Lakin D R, Reynolds T D. Advanced engine health management Applications of the SSME real-time vibration monitoring system [R]. AIAA-2000-3622, 2000.
[10] Dvidson M, Stephens J. Advanced health management system for the space shuttle main engine [R]. AIAA-2004-3912, 2004.
[11] Powrie H, Novis A. Gas path debris monitoring for F-35 Joint Strike Fighter propulsion system PHM [C]. Aerospace conference, 2006 IEEE.
[12] Hess A, Fila L. The joint strike fighter (JSF) PHM concept: potential impact on aging aircraft problems [C]. Aerospace Conference Proceedings, 2002-06: 3021-3026.
[13] Augustin M J, Bradley S J. Achieving HUMS benefits in the military environment– HUMS developments on the CH-146 griffon fleet [C]. Proceedings of the American Helicopter Society 60th Annual Forum, Baltimore, Maryland, May 2004: 7-10.
[14] Raini Wright. New IMD HUMS technology improves helicopter lifecycle [C]. Pentagon Brief, Septemper, 2004.
[15] Byington C S, et al. Metrics evaluation and tool development for health and usage monitoring system technology [C]. The American Helicopter Society 59th Annual Forum, Phoenix, Arizona, 2003.
[16] Clark G J, Vian J L, West M E, et al. Multi-platform airplane health management [C]. Aerospace conference, 2007 IEEE: 1-13.
[17]国家中长期科学和技术发展规划纲要(2006-2020年)中国政府门户网站:http://www.gov.cn/jrzg/2006-02/09/content_183787.htm.
[18]胡茑庆,胡雷,秦国军.面向机电系统状态监控的新异类检测技术分析[C].第11届全国设备故障诊断学术会议论文集,2008:143-145.
[19] Markou M, Singh S. Novelty detection: a review Part 1: statistical approaches [J]. Signal Processing, 2003, 83(12): 2481-2497.
[20] Sch?lkopf B, Willianmson R, Smola A, et al, Support vector method for novelty detection[C]. Advances in neural information processing systems 12: proceedings of the 1999 conference, MIT Press, 2000: 582-588.
[21]宣蕾.网络安全定量风险评估及预测技术研究[D].长沙:国防科学技术大学,2007.
[22]屈梁生.全息谱原理与全息动平衡技术//雷源忠.机械科学基础研究20年[M].武汉:武汉理工大学出版社,2007.
[23] Barnett V, Lewis T. Outliers in statistical data [M]. Wiley series in probability and mathematical statistics, John Wiley and Sons Limited, 2nd ed.,1978
[24] Moya M, Koch M, and Hostetler L. One-class classifier networks for target recognition applications [C]. In Proceedings world congress on neural networks, 1993: 797-801.
[25] Ritter G, Gallegos M. Outliers in statistical pattern recognition and an application to automatic chromosome classification [J]. Pattern Recognition Letters, 1997, 18: 525-539.
[26] Jagota A. Novelty detection on a very large number of memories stored in a Hopfield-style network [C]. International Joint Conference on Neural Networks IJCNN-91, 1991, 2: 905.
[27]谢光军.液体火箭发动机涡轮泵实时故障检测技术及系统研究[D].长沙:国防科技大学,2006.
[28]谢光军,李俭川,胡茑庆等.涡轮泵实时故障检测的多特征参量自适应阈值综合决策算法[J].中国机械工程,2005,16(13):1184-1187.
[29]谢光军,胡茑庆,胡雷.涡轮泵实时故障检测的改进自适应相关阈值算法[J].推进技术,2006,27(1):5-8.
[30] Clifton D, Bannister P, Tarassenko L. A framework for novelty detection in jet engine vibration data [J]. Key Engineering Materials, 2007, 347: 305-310.
[31]朱恒伟.液体推进剂火箭发动机地面试车故障检测与诊断研究[D].长沙:国防科技大学,1997.
[32] Biggs R. A probabilistic risk assessment for the space shuttle main engine with a turbomachinery vibration monitor cutoff System [R]. AIAA-90-2712, 1990.
[33] Smyth P. Hidden Markov models for fault detection in dynamic systems [J]. Pattern Recognition, 1994, 27(1): 149-164.
[34] Smyth P. Markov monitoring with unknown states [J]. IEEE Journal on Selected Areas in Communications, 1994, 12(9): 1600-1612.
[35] Miao Q, Makis V. Condition monitoring and classification of rotating machinery using wavelets and hidden Markov models [J]. Mechanical System and Signal Processing, 2007, 21: 840-855.
[36] Zhang X, Xu R, Kwan C, et al. An integrated approach to bearing fault diagnostics and prognostics [C]. Proceeding of the 2005 American Contral Conference, 2005, 4: 2750-2755.
[37] Dong M, He D. A segmental hidden semi-Markov model (HSMM)–based diagnostics and prognostics framework and methodology [J]. Mechanical System and Signal Processing, 2007, 21(5): 2248-2266.
[38] Yeung D Y, Ding Y. Host-based intrusion detection using dynamic and static behavioral models [J]. Pattern Recognition, 2002, 36: 229-243.
[39] Satnam Singh, Haiying Tu, William Donat, et al. Anomaly detection via feature-aided tracking and Hidden Markov Models [C]. IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans, 2009, 39(1): 144-159.
[40] Jiang M F, Tseng S S, Su C M. Two-phase clustering algorithm for outliers detection [J]. Pattern Recognition Letters, 2001, 22: 691-700.
[41] Tax D M J, Duin R P W. Data description in subspaces[C]. International Conference on Pattern recognition, 2000, Barcelona.
[42] Yeung D Y, Chow C. Parzen window network intrusion detectors [C]. Proc. International Conference on Pattern Recognition, 2002.
[43] Tarassenko L, Hayton P, Cerneaz N and Brady M. Novelty detection for the identification of masses in mammograms [C]. Proc. 4th IEE International Conference on Artificial Neural Networks, 1995, 4: 442-447.
[44] Tarassenko L, Nairac A, Townsend N and Cowley P. Novelty detection in jet engines [C]. IEE Colloquium on Condition Monitoring, Imagery, External Structures and Health, 1999: 41-45.
[45] Forrest S, Perelson A S, Allen L, et al. Self-non-self discrimination in a computer [C]. Proc. IEEE Symposium on Research in Security and Privacy, Oakland, CA, May 1994: 202-212.
[46] Dasgupta D, Gonzalez F A. An immunogenetic approach to intrusion detection [R]. Division of Computer Science, University of Memphis, Tech. Report CS-01-001, 2001.
[47] Dasgupta D, Majumdar N S. Anomaly detection in multidimensional data using negative selection algorithm [C]. Proc. IEEE Conference on Evolutionary Computation, Hawaii, May 2002: 1039-1044.
[48] Stibor T, Timmis J, Eckert C. A Comparative Study of Real-Valued Negative Selection to Statistical Anomaly Detection Techniques [C]. ICARIS 2005, 2005: 262-275.
[49] Hang X S, Dai H H. An Extended Negative Selection Algorithm for Anomaly Detection [C]. PAKDD 2004, LNAI 3056, 2004: 245-254.
[50] Gonzalez F A, Dasgupta D. Anomaly detection using real-valued negative selection [J]. Genetic programming and evolvable machine, 2003, 4: 383-403.
[51] González F, Dasgupta D, Kozma R. Combining negative selection and classification techniques for anomaly detection [C]. Congress on Evolutionary Computation, IEEE 2002: 705–710.
[52] González F, Dasgupta D, Nino L F. A randomized real-valued negative selection algorithm [C]. Proceedings of the 2nd International Conference on Artificial Immune Systems (ICARIS). LNCS, Edinburgh, UK, Springer-Verlag, 2003: 261–272.
[53] Ji Z, Dasgupta D. Real-valued negative selection algorithm with variable-sized detectors [C]. Genetic and Evolutionary Computation– GECCO-2004, Part I. Volume 3102 of LNCS.,Seattle, WA, USA, Springer-Verlag, 2004: 287–298.
[54] Ayara, Timmis, de Lemos, et al. Negative selection: how to generate detectors [C]. Proceeding of 1st International Conference on Artificial Immune Systems, University of Kent at Canterbury, UK, 2002.
[55] Kim J, Bentley P J. A model of gene library evolution in the dynamic clonal selection algorithm [C]. Proceeding of 1st International Conference on Artificial Immune Systems, University of Kent at Canterbury, UK, 2002.
[56] De Castro L N, Von Zuben F J. An evolutionary immune network for data clutering [C]. Proceeding of 6th Brazilian Symposium or Neural Networks, Rio de Janeiro, Brazil, 2000, 84-89.
[57] Hunt J E, Cooke D E. Learning using an artificial immune system [J]. Journal of Network and Computer Application, 1996, 19(2): 189-212.
[58] Markou M, Singh S. Novelty detection: a review Part 2: neural network based approaches [J]. Signal Processing. 2003, 83(12): 2499-2521.
[59] Hodge J, Austin J. A survey of outlier detection methodologies [J]. Artificial Intelligence Review, 2004, 22(2): 85-126.
[60] Bishop C M. Novelty detection and Neural Network validation [C]. Proceedings of the IEE Conference on Vision, Image and Signal Processing. 1994: 217-222.
[61] Nairac A, Townsend N, Carr R, King S, et al. A system for the analysis of jet system vibration data [J]. Integrated Computer Aided Engineering, 1999, 6(1): 53-65.
[62] Wong M L D, Jack L B, Nandi A K. Modified self-organising map for automated novelty detection applied to vibration signal monitoring [J]. Mechanical Systems and Signal Processing, 2006, 20: 593-610.
[63] Marsland S. On-line novelty detection through self-organisation, with application to inspection robotics [D]. University of Manchester, 2001.
[64] Frota R A, Barreto G A, Mota J C M. Anomaly detection in mobile communication networks using the self-organizing Map [J]. Journal of intelligent and fuzzy systems, 2007, 18(5): 493-500.
[65]胡茑庆,邱忠,谢光军,胡雷.涡轮泵故障检测的频段能量比SOM算法.国防科技大学学报[J],2005,27(6):93-96.
[66] Bogacz R, Brown M W, Giraud-Carrier C. High capacity neural networks for familiarity discrimination[C]. Proc. ICANN'99, Edinburgh, 1999: 773-778.
[67] Crook P A, Hayes G.. A robot implementation of a biologically inspired method for novelty detection[C]. Proc. towards Intelligent Mobile Robots Conference, Manchester, 2001.
[68] Crook P A, Marsland S, Hayes G, et al. A tale of two filters - online novelty detection[C]. Proc. 2002 IEEE ICRA Conference, Washington DC, May 2002.
[69] Kojima K, Ito K. Autonomous learning of novel patterns by utilizing chaotic dynamics [C]. IEEE International Conference on Systems, Man, and Cybernetics, IEEE SMC '99, 1999, 1: 284-289.
[70] Borisyuk R, Denham M, Hoppensteadt F, et al. An oscillatory neural network model of sparse distributed memory and novelty detection [J]. BioSystems, 2000: 265-272.
[71] Tax D M J, Duin R P W. Support vector domain description [J]. Pattern Recognition Letters, 1999, 20 (11-13): 1191-1199.
[72] Tax D M J, Duin R P W. Support vector data description [J]. Machine Learning, 2004, 54(1): 45-66.
[73] Worden K, Manson G, Allman D. Experimental validation of a structural health monitoring methodology. Part I. Novelty detection on a laboratory structure [J]. Journal of Sound and Vibration, 2003, 259: 323-343.
[74] Worden K, Manson G . Experimental validation of a structural health monitoring methodology: Part II. Novelty detection on a gnat aircraft [J].Journal of sound and vibration, 2003, 259(2): 345-363.
[75] Worden K, Manson G.The application of machine learning to structural health monitoring. Part III Damage location on an aircraft wing [J]. Journal of sound and vibration, 2003, 259(2): 365-385.
[76] Worden K, Manson G.The application of machine learning to structural health monitoring [J].Philosophical transactions of the royal society, a-mathematical physical and engineering sciences, 2007,365 (1851): 515-537.
[77] Worden K, Manson G, Denoeux T. An evidence-based approach to damage location on an aircraft structure [J]. Mechanical systems and signal processing, 2009, 23: 1792-1804.
[78] Hayton P, Sch?lkopf B, Tarassenko L, et al. Support vector novelty detection applied to jet engine vibration spectra [C]. Advances in Neural Information Processing Systems 13, 2001.
[79] Hayton P, Utete S, King D, et al. Static and dynamic novelty detection methods for jet engine health monitoring [J]. Philosophical transactions of the royal society, a-mathematical physical and engineering sciences, 2007, 365(1851): 493-514.
[80] Clifton D, Bannister P, Tarassenko L. Novelty detection in large-vehicle turbocharger operation [C]. IEA/AIE 2007: 591- 600.
[81] Nairac A, Corbett-Clark T, Ripley R, et al. Choosing an appropriate model for novelty detection [C]. Proc. 5th IEE International Conference on Artificial Neural Networks, Cambridge, 1997: 227 - 232.
[82] Oza N, Tumer K, Tumer I. et al. Classification of aircraft maneuvers for fault detection [J]. Lecture Notes in Computer Science, 2003, 2709: 375-384.
[83] Srivastava A. Discovering system health anomalies using data mining techniques [C]. Proceedings of the Joint Army Navy NASA Air Force Conference on Propulsion, Charleston, SC, June 2005.
[84] Hickinbotham S J, Austin J.Neural networks for novelty detection in airframe strain data [C]. Proc. IEEE IJCNN, 2000.
[85] Mackey R, James M, Park H, et al. BEAM: Technology for antonomous self-analysis [C]. IEEE 2001, 6: 2989-3001.
[86] Park H, Mackey R, James M, et al. Analysis of space shuttle main engine data using beacon-based exception analysis for multi-missions [C]. Proceedings of the IEEE Aerospace Conference, IEEE, New York, 2002.03, 6: 2835-2844.
[87] Iverson D L. Inductive system health monitoring [C]. Proceedings of the 2004 InternationalConference on Artificial Intelligence (IC-AI’04), CSREA Press, Las Vegas, NV, June 2004.
[88] Schwabacher M. Machine learning for rocket propulsion health monitoring [C]. Proceedings of the SAE World Aerospace Congress, Dallas, 2005, 114(1): 1192-1197.
[89]黄敏超,张育林,陈启智.神经网络在液体火箭发动机故障检测中的应用(II)[J].推进技术,1999,20(2):1-4.
[90] Surace C, Worden K, Tomlinson G.A novelty detection approach to diagnose damage in a cracked beam [C]. Proc. of SPIE, 1997, 3089: 947-953.
[91] Guttormsson S E, Marks R J, El-Sharkawi M A.Elliptical novelty grouping for online short-turn detection of excited running rotors [C]. IEEE Transactions on Energy Conversion, March 1999: 16-22.
[92] Streifel R J, Maks R J, El-Sharkawi M A, et al. Detection of shorted turns in the field of turbine-generator rotors using novelty detectors- development and field tests [C]. IEEE Transactions on Energy Conversation, 1996, 11(2): 312-317.
[93] Petsche T, Marcantonio A, Darken C, et al. A Neural Network Autoassociator for Induction Motor Failure Prediction. 1996. http://citeseer.ist.psu.edu /petsche96neural.html.
[94] Li Y H, Pont M J, Jones N B. Improving the performance of radial basis function classifiers in condition monitoring and fault diagnosis applications where‘unknown’faults may occur [J]. Pattern recognition letters, 2002, 23: 569-577.
[95] Omenzetter P, Brownjohn J M W, Moyo P. Identification of unusual events in multi-channel bridge monitoring data [J]. Mechanical Systems and Signal Processing, 2004, 18: 409-430.
[96] Yan A M, Kerschen G, Boe P, et al. Structural damage diagnosis under varying environmental conditions - Part I: A linear analysis [J]. Mechanical Systems and Signal Processing 2005, 19: 847-864.
[97] Worden K, Sohn H, Farrar C R. Novelty detection in a chaning environment: regrassion and interpolation approaches [J]. Journal of sound and vibration, 2002, 258(4): 741-761.
[98] Timusk M, Lipsett M, Chris K. Mechefske. Fault detection using transient machine signals [J]. Mechanical Systems and Signal Processing 22 (2008) 1724-1749.
[99]廖少英.液体火箭推进增压输送系统[M].北京:国防工业出版社,2007,4.
[100]总装备部情报研究所.1990-2002年各国运载火箭动力系统故障情况[R].情2003002,北京:中国国防科技信息中心,2003.
[101]张振鹏.液体发动机故障检测与诊断中的基础研究问题[J].推进技术,2002,23(5):353-359.
[102] Jue F, Kuck F. Space shuttle main engine (SSME) options for the future shuttle [R]. AIAA-2002-3758, 2002.
[103]边肇祺,张学工.模式识别-2版[M].北京:清华大学出版社,1999.
[104]温熙森.模式识别与状态监控[M].北京:科学出版社,2007.
[105] Vapnik V N.统计学习理论的本质[M].张学工译.北京:清华大学出版社,2000.
[106] Vapnik V N. Statistical learning theory [M]. Wiley, Newyork, 1998.
[107] Davy M, Desobry F, Gretton A, et al. An online support vector machine for abnormal eventsdetection [J]. Signal Processing, 2006, 86(8):2009-2025.
[108] Rabiner L R, Juang B H. An introduction to Hidden Markov Models [C]. IEEE ASSP Magazine, January, 1986: 4-16.
[109] Rabiner L R. A tutorial on hidden markov models and selected applications in speech recognition [C]. Proceedings of the IEEE, 1989, 77(2): 257-285.
[110] Theodoridis S, Koutroumbs K. Pattern recognition [M]. 3rd edition. Beijing: China Machine Press, 2006.
[111] Kohonen T. The self-organizing map [C]. Proceedings of the IEEE 78, 1990, (9): 1464-1480.
[112] Willsky A S. A survey of design methods for failure detection in dynamic systems. Automatica 1976, 12: 601-611.
[113] Cristianini N, Shawe-Taylor J.支持向量机导论[M].李国政,王猛,曾华军译.北京:电子工业出版社,2004.
[114] Burges C J C. A Tutorial on support vector machines for pattern recognition [J]. Data Mining and Knowledge Discovery, 1998 (2):121-167.
[115] Chu Shun-kwong. Scaling up support vector data description by using Core-sets [D]. Department of compute science, Hong Kong university of science and technology, 2004.
[116]柳新民,邱静,胡茑庆,刘冠军.一种改进的无监督学习SVM及其在故障识别中的应用[J].机械工程学报,2006,42(4):107-111.
[117]王慧.基于振动参量的涡轮泵健康状况判别方法的研究[D].北京:中国航天科技集团公司,2003.
[118]陈启智.液体火箭推进系统健康监控技术的演变[J].推进技术,1997,18(1):1-7.
[119]刘刚,曲梁生.应用Bootstrap方法构造机械故障特征库[J].振动工程学报,2002,15(1):106-110.
[120]邱忠.涡轮泵试车数据分析及健康状态判别方法研究[D].长沙:国防科学技术大学,2005.
[121]胡雷,胡茑庆,秦国军,邱忠.单类支持向量机在涡轮泵试车数据分析中的应用[J].振动、测试与诊断,2007,27(2):95-97.
[122]粟塔山.最优化计算原理与算法程序设计[M].长沙:国防科大出版社,2001.
[123] Platt J. Fast training of support vector machines using sequential minimal optimization [M]// Sch?lkopf B, Burges C, SmolaA, et al. Advances in Kernel Methods-Support Vector Learning. Cambridge, MA: MIT Press, 1999: 185-208.
[124]邓乃扬,田英杰.擞据挖掘中的新方法——支持向量机[M].北京:科学出版社,2004.
[125]钟秉林,黄仁.机械故障诊断学-3版[M].北京:机械工业出版社,2006.
[126] Boser B, Guyon I, Vapnik V. A training algorithm for optimal margin classifiers[C]. Proceedings of the 5th Annual ACM Workshop on Computational Learning Theory, ACM Press, 1992, 144-l52.
[127] Osuna E, Freund R, Girosi F. Support Vector Machines: Training and Application[R]. Cambridge, MA: Massachusetts Institute of Technology, AILab, 1997.
[128] Osuna E, Freund R, Girosi F. An Improved Training Algorithm for Support VectorMachines[C]. Proceedings of the 1997 IEEE Workshop on Neural Networks for Signal Processing, 1997: 276-285.
[129] Platt J. Sequential minimal Optimization: A Fast algorithm for training support vector machines [R]. Technical report MSR-TR-98-14, Microsoft Re-search, 1998.
[130]陈栋梁.支持向量机训练算法研究[D].合肥:合肥工业大学,2007.
[131]曾志强.支持向量机的训练与简化算法研究[D].杭州:浙江大学,2007.
[132] Joachims T. Make large-scale SVM learning practical [M]. Advances in Kernel Methods: Suppon Vector Learning, Cambridge, MA: MIT Press, 1999.
[133]胡雷,胡茑庆,秦国军.涡轮泵试车数据单类支持向量机检测算法[J].推进技术,2008,29(2):244-248.
[134]胡雷,胡茑庆,秦国军.双阈值单类支持向量机在线故障检测算法及应用[J].机械工程学报,2009,45(3):169-173.
[135] Coppock H W, Freund J E. All-or-none versus incremental learning of errorless shock escapes by the rat [J]. Science, 1962, 135(3500): 318-319.
[136] Polikar R, Byorick J, Krause S,et al. Learn++:a classifier independent incremental learning algorithmfor supervised neural networks [C]. Proceedings of the International Joint Conference on Neural Networks, 2002, 2: 1742-1747.
[137] Schlimmer J C, Granger R H. Incremental learning from noisy data [J]. Machine Learning, 1986, 1: 317-354.
[138] Syed N A, Liu H, Sung K K. Incremental learning with support vector machines [C]. Proceeding of International Joint Conference on Artificial Intelligence (IJCAI-99), 1999.
[139]廖东平.支持向量机方法及其在机载毫米波雷达目标识别中的应用研究[D].长沙:国防科学技术大学,2006.
[140]孔锐,张冰.一种快速支持向量机增量学习算法[J].控制与决策,2005,20(10):1129～1132,1136.
[141]史朝辉,王晓丹,杨建勋.一种SVM增量训练淘汰算法[J].计算机工程与应用,2005(23):187-189,212.
[142] Bottou L, Vapnik V. Local learning algorithms [J]. Neural Computation, 1992, 4(6): 888-900.
[143] Ralaivola L, Florence d’Alché-Buc. Incremental support vector machine learning: a local approach [J]. Lecture Notes in Computer Science, 2001, 2130: 322-329.
[144] Florence d’Alchd-Buc, Ralaivola L. Incremental learning algorithms for classification and regression: local stratctics [C]. AIP Conf. Proc, September 2, 2002, 627: 320-329.
[145] Cauwenberghs G, Poggio T. Incremental and decremental support vector machine learning [C]. Adv. Neural Information Processing, vol. (13). MIT Press, 2001.
[146] Martin M. Online support vector machines for function approximation [R]. (Tech. ReP. LSI-02-11-R). Catalunya, Spain: Software Department, Universitat Politecnica de Catalunya. 2002.
[147] MA Junshui, Theiler J, Perkin S. Accurate online support vector regression [J]. Neural Computation, 2003, 15(11): 2683-2703.
[148] Diehl C P, Cauwenberghs G. SVM incrcmetal learning, adaptation and optimization [C]. Proceedings of IJCNN-03, 2003: 2685-2690.
[149]李敬.增量学习及其在图像识别中的应用[D].上海:上海交通大学,2008.
[150]段华.支持向量机的增量学习算法研究[D].上海:上海交通大学,2008.
[151]奉国和.支持向量机改进算法及其应用研究[D].广州:中山大学,2005.
[152]萧嵘,王继成,孙正兴,张福炎.一种SVM增量学习算法[J].南京大学学报(自然科学版),2002,38(2):152-157
[153] Li Zhongwei, Zhang Jianpei, Yang Jing. A heuristic algorithm to incremental support vector machine leaming [C]. Proceedings of the 3th International Conference on Machine Learning and Cybernetics, Shanghai, 2004(3): 1764-1767
[154]安金龙,王正欧.一种适合于增量学习的支持向量机的快速循环算法[J].计算机应用,2003,23(10):12-14.
[155]胡雷,秦国军,胡茑庆.涡轮泵振动故障实时检测技术与系统.2007年全国振动工程及应用学术会议论文集,杭州,2007,10:17-19.
[156] Niaoqing Hu, Guojun Qin, Lei Hu, et.al. Real-time Fault Detection System for Turbopump of Liquid Rocket Engine Based on Vibration Measurement Signals. Proc. 6th Int. Workshop on Structural Health Monitoring, Stanford, September 11-13, 2007: 590-597.
[157]胡海峰.液体火箭发动机涡轮泵健康监控系统及数据管理的研究[D].长沙:国防科技大学,2004.
[158] Srivastava A N, Mah R, Meyer Claudia. Integrated Vehicle Health Management, Automated detection, diagnosis, prognosis to enable mitigation of adverse events during flight [Technical Plan, Version 2.01]. NASA, Aeronautics Research Mission Directorate, Aviation Safety Program, August 14, 2008.
[159] Weston J, Watkins C. Multi-class support vector machines[C]. Proceedings of ESANN99, Brussels D. Facto Press, 1999.
[160] Widodo A, Yang B S. Review: Support vector machine in machine condition monitoring and fault diagnosis [J]. Mechanical Systems and Signal Processing, 2007, 21: 2560-2574.
[161] Krebel U. Pairwise classification and support vector machines[C]. Advances in Kernel Methods-Support Vector Learning, MA, 1999: 255-268.
[162] Hsu C W, Lin C J. A comparison of methods for multi-classification support vector machines [C]. IEEE Transactions on nearal networks, 2002, 13(2): 415-425.
[163] Bottou L, Cortes C, et al. Comparison of classifier methods: a case study in handwriting digit recognition [C]. In Int. Conf. on Pattern Recongnition, IEEE Computer Society Press,1994: 77-87.
[164]马笑潇.智能故障诊断中的机器学习新理论及其应用研究[D].重庆:重庆大学,2002.
[165]张周锁,李凌均,何正嘉.基于支持向量机的多故障分类器及应用[J].机械科学与技术,2004,23(5):536-538.
[166] Platt J C, Cristianini N, Shawe-Taylor J. Large margin DAGs for multiclass classification [J]. Advances in Neural Information Processing Systems, 2000, 12(3): 547-553.
[167]李昆仑,黄厚宽,田盛丰.一种基于有向无环图的多类SVM分类器[J].模式识别与人工智能,2003,16(2):164-168.
[168]唐发明,王仲东,陈绵云.一种新的二叉树多类支持向量机算法[J].计算机工程与应用,2005(7):24-26.
[169]侯惠芳,刘素华.基于决策二叉树的多类支持向量机在大气质量评定中的应用[J].计算机应用与软件,2005,22(5):136-138.
[170]袁胜发,褚福磊.SVM多类分类算法及其在故障诊断中的应用[C].振动工程学报,2004,17(S):419-421.
[171] Takahashi F, Abe S. Decision-tree-based multiclass support vector machines [C]. Proceedings of the Annual Conference of the Institute of Systems, Control and Information Engineers, 2002, 46: 483-488.
[172] Schwenker F. Hierarchical support vector machines for multi-class pattern recognition [C]. Brigton The 4th Int Conf on Knowledge-based Intelligent Engineering System & Allied Technologies,2000: 561-565.
[173]刘江.基于SVM的多类分类算法研究及其在网维中心故障定位中的应用[D].广州:中山大学,2007.
[174]张晓平.基于支持向量机的多类分类算法研究及在滚动轴承故障识别中的应用[D].太原:太原理工大学,2007.
[175]赵春晖,陈万海,郭春燕.多类支持向量机方法的研究现状与分析[J].智能系统学报,2007,2(2):11-19.
[176]夏鲁瑞,胡茑庆,秦国军.转速波动状态下涡轮泵典型故障诊断方法[J].推进技术,2009,30(3):342-346.
[177] Tolani D, Yasar M, Chin Shin, et al. Anomaly detection for health management of aircraft gas turbine engines [C]. 2005 American Control Conference, Portland, OR, USA, June 8-10, 2005: 459-464.
[178] Brotherton T.; Johnson T. Anomaly detection for advanced military aircraft using neural networks [C]. Aerospace Conference, 2001, IEEE Proceedings, March 10-17, 2001, 6: 3113-3123.
[179] Brotherton T.; Mackey R. Anomaly detector fusion processing for advanced military aircraft [C]. Aerospace Conference, 2001, IEEE Proceedings, March 10-17, 2001, 6: 3125-3137.
[180] Zhang B, Sconyers C.; Byington C, et al. Anomaly detection: A robust approach to detection of unanticipated faults [C]. International Conference on Prognostics and Health Management, 2008, Oct. 6-9, 2008: 1-8.