Utilization of Unsupervised Anomalies Detector as a Tool for Managing the TDRS Constellation at GSFC

Ma, Kenneth Y ; Montoro, Manuel ; Shaw, Harry ; Mihir, Patel ; Woods, Lawrence ; Williams, Thomas ; Steele, Jonathan ; Cunnif, David ; Bonacci, Carissa Brealey ; Miller, Ronald(NASA Goddard Space Flight Center, Greenbelt, MD, United States)

关键词: ANOMALIES;    COMMUNICATION NETWORKS;    DATA MINING;    DATA PROCESSING;    FAILURE ANALYSIS;    GROUND STATIONS;    REAL TIME OPERATION;    RELIABILITY ANALYSIS;    REMOTE SENSING;    SAFETY FACTORS;    TDR SATELLITES;    TELEMETRY;   

RP-ID  :  GSFC-E-DAA-TN73802-2

学科分类：航空航天科学

美国|英语

来源: NASA Technical Reports Server

【 摘 要 】

NASA’s Goddard Space Flight Center (GSFC) operates a constellation of ten geosynchronous Tracking and Data Relay Satellites (TDRS). The mission of the TDRS constellation is to provide relay communications from low-earth orbiting spacecraft to the primary ground station at the White Sands Complex in Las Cruces, New Mexico. Major customers include the International Space Station and Hubble Space Telescope. The NASA Space Network project office at GSFC manages the constellation of spacecraft. The constellation is over 30 years old, and a wide range of technologies and manufacturing techniques are represented on-orbit. Since 1983, the TDRS constellation has recorded thousands of gigabytes of telemetry data. Spacecraft telemetry data has changed throughout the three generations of TDRS spacecraft, however each spacecraft has the same basic functions with some generational enhancements. The constellation includes several spacecraft that have significantly outlived the manufacturer's projected lifetime. This has provided NASA with a significant benefit in terms of return on investment, however it places a burden on efficient management of the assets for maximum life without permitting a TDRS spacecraft to become stranded in its geosynchronous orbital slot. Consequently, the highest level of attention is paid to systems whose failure could strand a TDRS spacecraft in orbit. In this paper, we proposed two stages of analyzing spacecraft anomalies using data mining (DM) to enhance on-going predictions of spacecraft life, subsystem performance, and analysis of subsystem anomalies. The first stage conducts the unsupervised anomaly detector to detect potential anomalies in real-time telemetry data. The second stage introduced telemetry weight (TW) to each telemetry parameter to determine which parameter caused the strongest anomaly. We will present case studies of some of these analyses and how the data can impact decisions on the management of the constellation.