卫星在轨运行期间，遥测数据是反映卫星健康状态的重要依据，在卫星故障早期检测到遥测数据的潜在异常对卫星的安全维护具有重大意义。工程上采用的阈值法无法有效检测到门限内的故障征兆，而且目前这一领域的理论研究无法有效地挖掘多维遥测序列的潜在相关性。针对这一问题，采用一种融合主成分分析的相关概率模型的检测方法，以某型号卫星实际在轨遥测数据为对象，深入分析故障案例。通过仿真验证该方法能够在故障早期检测出异常，并对实验结果进行对比和分析。而且，这种方法可以快速地帮助运管人员对早期故障做出诊断，以便地面及时处理，避免发生更大的事故。
During the orbital operation of the satellite, the telemetry data is an important basis for reflecting the health status of satellites. The detection of potential anomalies in telemetry data is of great significance for the maintenance of satellites. Threshold method used in engineering can not effectively detect failure symptoms within the threshold, and the current theoretical research in this field can not effectively tap the potential correlation of multidimensional telemetry sequences. Therefore, this paper, taking the actual on-orbit telemetry data of a satellite as the object, adopts a detection method of correlation probability model that incorporates principal component analysis (PCA), and analyzes the failure case deeply. It verifies that this method can detect satellite's early failure and the results are compared and analyzed. This method can also quickly diagnose the failure so that the ground can handle it in time to avoid further accidents.
[1] Yairi T, Kawahara Y, Fujimaki R, et al. Telemetry-mining:a machine learning approach to anomaly detection and fault diagnosis for space systems[C]//2nd IEEE International Conference on Space Mission Challenges for Information Technology(SMC-IT'06), Pasadena, CA, USA:IEEE Press, 2006:466-476.
[2] Hayden S C, Sweet A J, Christa S E, et al. Advanced diagnostic system on earth observing one[C]//Proceedings of AIAA Space 2004 Conferace and Exhibit, 2004:14.
[3] 彭喜元,庞景月,彭宇,等.航天器遥测数据异常检测综述[J].仪器仪表学报, 2016, 37(9):1929-1945.
[4] Chandola V, Banerjee A, Kumar V. Anomaly detection:a survey[J]. ACM Computing Surveys, 2009, 41(3):1-58.
[5] Song X Y, Wu M X, Jermaine C, et al. Conditional anomaly detection[J]. IEEE Transactions on Knowledge and Data Engineering, 2007,19(5):631-645.
[6] 谢锦生, 郭立, 陈运必, 等. 基于时空惊奇计算的视频异常检测方法[J]. 中国科学院研究生院学报, 2013, 30(1):83-89.
[7] Ma J, Perkins S. Time-series novelty detection using one-class support vector machines[C]//Proceedings of the International Joint Conference on Neural Networks, 2003. Portland, CR, USA:IEEE Press, 2003:1741-1745.
[8] Cohen G, Hilario M, Pellegrini C, et al. One-class support vector machines with a conformal kernel:a case study in handling class imbalance[C]//Structural, Syntactic, and Statistical Pattern Recognition. Springer-Verlag Berlin Heidelberg, 2004, 3138:850-858.
[9] 陈斌. 异常检测方法及其关键技术研究[D]. 南京:南京航空航天大学, 2013.
[10] 李楠,张云燕,李言俊. 一种自旋稳定卫星姿态传感器数据异常的诊断方法[J]. 宇航学报, 2011, 32(6):1327-1332.
[11] Spiegelberg J, Rusz J. Can we use PCA to detect small signals in noisy data?[J]. Ultramicroscopy, 2017, 172:40-46.
[12] Wang X, Kruger U, Irwin G W, et al. Nonlinear PCA with the local approach for diesel engine fault detection and diagnosis[J]. IEEE Transactions on Control Systems Technology, 2008, 16(1):122-129.
[13] Iverson D L, Martin R, Schwabacher M, et al. General purpose data-driven monitoring for space operations[J]. Journal of Aerospace Computing, Information, and Communication, 2012, 9(2):26-44.
[14] Iverson D L. Inductive system health monitoring[C]//Proceedings of the International Conference on Artificial Intelligence & Proceedings of the International Conference on Machine Learning:Models, Technologies & Applications, Las Vegas, Nevada, USA, 2004:605-611.
[15] Ding J W, Liu Y B, Zhang L, et al. An anomaly detection approach for multiple monitoring data series based on latent correlation probabilistic model[J]. Applied Intelligence, 2016, 44(2):340-361.
[16] Zhong S S, Luo H, Lin L, et al. An improved correlation-based anomaly detection approach for condition monitoring data of industrial equipment[C]//2016 IEEE International Conference on Prognostics and Health Management (ICPHM). Ottawa, ON, Canada:IEEE Press, 2016:1-5.
[17] Reshef D N, Reshef Y A, Finucane H K, et al. Detecting novel associations in large data sets[J]. Science, 2011, 334(6062):1518-1524.
[18] Pang J Y, Liu D T, Peng Y, et al. Anomaly detection based on uncertainty fusion for univariate monitoring series[J]. Measurement, 2017, 95:280-292.
