针对运载火箭复杂系统的故障检测难以建立准确的数学模型的问题，研究了基于数据驱动的数据挖掘异常检测算法，对多种数据挖掘算法在运载火箭发动机异常检测的应用进行了研究和分析，提出了基于混合概率密度统计的多策略异常检测评价算法。该算法基于非监督学习的算法挖掘火箭发动机不同参数间的正常关联模型，火箭发动机早期的异常数据会引起正常关联模型的破坏，引入混合概率密度统计的多策略异常检测评价机制，可以有效屏蔽参数测量故障对系统故障检测的影响，从而更加准确给出系统异常程度。使用发动机历史试车数据作为样本进行特征模型的训练，使用一元、多元和混合概率密度模型对存在异常的发动机试车数据进行了实时异常检测的实验验证。实验结果表明，相比传统基于阈值和规则的异常检测算法，基于概率密度统计的多策略异常检测算法不仅可给出系统的正常和异常的状态，还可计算各参数和整个系统的异常值，为运载火箭进一步的故障诊断提供更加灵活的参考。
Considering the fact that modeling a complex system such as the launch vehicle is particularly difficult, anomaly detection methods based on data driven are studied. This paper proposes a multi-strategy anomaly detection method based on statistics of probability density distribution. The unsupervised learning method is used to excavate the model for normal correlation between different parameters of the rocket engine. As the early anomaly data of the rocket engine will cause damage to the normal correlation model, the multi strategy anomaly detection and evaluation mechanism is introduced to effectively shield the influence of parameter measurement faults on fault detection of the system, so that system anomaly can be detected more accurately. We use the historical data of engine tests as the training samples to train the characteristic models, and experimental verification of anomaly detection of engine test data in real time is carried out by using the one-element model, the multi-element model and the model based on hybrid probability density distribution. The result shows that compared with the traditional thresh and norm based anomaly detection methods, the multi-strategy anomaly detection method can give not only the normal ad anomaly states of the system, but also the anomaly scores of each parameter and the whole system, thus providing a more flexible reference for further fault diagnosis of launch vehicles.
[1] 李洪. 智慧火箭发展路线思考[J]. 宇航总体技术, 2017, 1(1):1-7. LI H. Thinking on the development route of intelligent rocket[J]. Astronautical Systems Engineering Technology, 2017, 1(1):1-7(in Chinese).
[2] 宋征宇. 运载火箭远程故障诊断技术综述[J]. 宇航学报, 2016, 37(2):135-144. SONG Z Y. A review of remote fault diagnosis technology for launch vehicle[J]. Journal of Astronautics, 2016, 37(2):135-144(in Chinese).
[3] 刘成瑞, 张庆振, 任章. 基于扩展故障树的运载火箭故障诊断专家系统[J]. 宇航学报, 2008, 29(6):1936-1941. LIU C R, ZHANG Q Z, REN Z. Fault diagnosis expert system for launch vehicle based on extended fault tree[J]. Journal of Astronautics, 2008, 29(6):1936-1941(in Chinese).
[4] 张素明, 安雪岩, 颜廷贵, 等. 大型运载火箭的健康管理技术应用分析与探讨[J]. 导弹与航天运载技术, 2013(6):33-38. ZHANG S M, AN X Y, YAN T G, et al. Application and analysis of health management technology for large launch vehicles[J]. Missiles and Space Vehicles, 2013(6):33-38(in Chinese).
[5] 周东华, 刘洋, 何潇. 闭环系统故障诊断技术综述[J]. 自动化学报, 2013, 39(11):1933-1943. ZHOU D H, LIU Y, HE X, A review of fault diagnosis technology in closed loop system[J]. Acta Automatica Sinica, 2013, 39(11):1933-1943(in Chinese).
[6] 李璨, 赵小卓, 耿辉, 等. 一种用于运载火箭测试发射过程的多策略故障诊断系统:中国, CN105486526A[P]. 2016-04-13. LI C, ZHAO X Z, DI H, et al. A multi strategy fault diagnosis system for launch vehicle test and launch process:China, CN105486526A[P]. 2016-04-13(in Chinese).
[7] SCHWABACHER M, WATERMAN R. Pre-launch diagnostics for launch vehicles[C]//IEEE Aerospace Conference Proceedings. Piscataway, NJ:IEEE Press, 2008:1-8.
[8] MARTIN R A. Aerospace technologies advancements:Evaluation of anomaly detection capability for ground-based pre-launch shuttle operations. aerospace technologies advancements[M]. Croatia:InTech, 2010:141-164.
[9] FERRELL B, BROWN B, HALL D, et al. Usage of fault detection isolation & recovery (FDIR) in constellation (CxP) launch operations[C]//SpaceOps 2010 Conference. Reston, VA:AIAA, 2010:1-9.
[10] SCHWABACHER M, MARTIN R, WATERMAN R, et al. Ares I-X ground diagnostic prototype[C]//AIAA Infotech@Aerospace Conference 2010. Reston, VA:AIAA, 2010:1-11.
[11] 陈斌, 陈松灿, 潘志松, 等. 异常检测综述[J]. 山东大学学报(工学版), 2009(6):13-23. CHEN B, CHEN S C, PAN Z S, et al. A review of anomaly detection[J]. Journal of Shandong University (Engineering Edition), 2009(6):13-23(in Chinese).
[12] SAIN S R, GRAY H L, WOODWRD W A, et al. Outlier detection from a mixture distribution when training data are unlabeled[J]. Bulletin of the Seismological Society of America, 1999, 89(1):294-304.
[13] LEE H, CHO S. Application of LVQ to novelty detection using outlier training data[J]. Pattem Recognition Letters, 2006, 27(13):1572-1579.
[14] SCHOLKOPF B, PLATT J, SHAWE-TAYLOR J, et al. Estimating the support of high-dimensional distribution[J]. Neural Computation, 2001, 13(7):1443-1471.
[15] DEVENDRA T, MURAT Y, ASOK R, et al. Anomaly detection in aircraft gas turbine engines[J]. Journal of Aerospace Computing Information & Communication, 2015, 3(2):44-51.
[16] SPIROS P, MICHAIL V. Computing correlation anomaly scores using stochastic nearest neighbors[C]//Seventh IEEE International Conference on Data Mining. Piscataway, NJ:IEEE Computer Society, 2007:523-528.
[17] TSUYOSHI I, ANKUSH K, JAYANT K. Sparse gaussian markov random field mixtures for anomaly detection[C]//2016 IEEE 16th International Conference on Data Mining. Piscataway, NJ:IEEE Press, 2017:955-960.
[18] 陈雪巍, 汪轶俊, 何岗, 等. 基于CFD的某运载火箭发动机燃料泵故障诊断分析[J]. 上海航天, 2016, 33(2):68-72. CHEN X W, WANG Y J, HE G, et al. Fault diagnosis analysis of a launch vehicle engine fuel pump based on CFD[J]. Aerospace Shanghai, 2016, 33(2):68-72(in Chinese).
[19] 黄强. 高压补燃液氧煤油发动机故障检测与诊断技术研究[D]. 长沙:国防科技大学, 2012:70-84. HUANG Q. Research on fault detection and diagnosis technology of high pressure afterburning liquid oxygen kerosene engine[D]. Changsha:National Defense University of Science and Technology, 2012:70-84(in Chinese).
[20] 李京浩. 基于数据挖掘技术的液体火箭发动机故障检测和诊断研究[D]. 长沙:国防科技大学, 2007:11-34. LI J H. Research on fault detection and diagnosis of liquid rocket engine based on data mining technology[D]. Changsha:National Defense University of science and technology, 2007:11-34(in Chinese).
[21] 李艳军. 新一代大推力液体火箭发动机故障检测与诊断关键技术研究[D]. 长沙:国防科技大学, 2014:40-63. LI Y J. Key technologies for fault detection and diagnosis of new generation large thrust liquid rocket engine[D]. Changsha:National Defense University of Science and Technology, 2014:40-63(in Chinese).
CJA