JAXA Repository / AIREX 未来へ続く、宙(そら)への英知
title機械学習を用いた宇宙機の故障の予兆検知
Other TitleDetection of failure sign of spacecraft using machine learning
DOIinfo:doi/10.20637/JAXA-RR-18-008/0002
Author(jpn)梅津, 里香; 杉江, 卓哉; 長瀬, 雅之; 湖海, 亮; 竹島, 敏明; 海老沢, 研; 満田, 和久; 山本, 幸生
Author(eng)Umezu, Rika; Sugie, Takuya; Nagase, Masayuki; Kokai, Ryo; Takeshima, Toshiaki; Ebisawa, Ken; Mitsuda, Kazuhisa; Yamamoto, Yukio
Author Affiliation(jpn)株式会社セック; 株式会社セック; 株式会社セック; 株式会社セック; 宇宙航空研究開発機構宇宙科学研究所(JAXA)(ISAS); 宇宙航空研究開発機構宇宙科学研究所(JAXA)(ISAS); 宇宙航空研究開発機構宇宙科学研究所(JAXA)(ISAS); 宇宙航空研究開発機構宇宙科学研究所(JAXA)(ISAS)
Author Affiliation(eng)Systems Engineering Consultants Co., LTD.; Systems Engineering Consultants Co., LTD.; Systems Engineering Consultants Co., LTD.; Systems Engineering Consultants Co., LTD.; Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency (JAXA)(ISAS); Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency (JAXA)(ISAS); Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency (JAXA)(ISAS); Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency (JAXA)(ISAS)
Issue Date2019-03-08
Publisher宇宙航空研究開発機構(JAXA)
Japan Aerospace Exploration Agency (JAXA)
Publication title宇宙航空研究開発機構研究開発報告
JAXA Research and Development Report
VolumeJAXA-RR-18-008
Start page11
End page20
Publication date2019-03-08
Languagejpn
eng
Abstract宇宙機は安全性や信頼性が厳しく問われ, 未然に危険を予知し, 事故防止に繋げる運用環境が望まれる. 宇宙機の運用データに機械学習の技術を適用して故障解析を行い, 不具合を未然に検知する技術を蓄積し, 安定した宇宙機運用に資するための研究を行っている. X線天文衛星「すざく」の運用データを用いた電源系機器の故障の予兆検知を試み, 一定条件下での故障の予兆を検出することが可能であることを確認した.
Safety and reliability of the spacecraft are very important. It is desirable to predict possible failures of a spacecraft in advance and to have an operational environment leading to accident prevention. By applying failure analysis utilizing machine learning technology to the operation data of a spacecraft, we accumulate techniques to detect defects beforehand and are doing research to contribute to stable operation of the spacecraft. We tried to detect a sign of failure of the power supply using operational data of the X-ray astronomical satellite 'Suzaku'. As a result, we suggest that it is possible to detect a sign of failure under certain conditions.
Description形態: カラー図版あり
Physical characteristics: Original contains color illustrations
Keywordsmachine learning; failure analysis; spacecraft; Suzaku
Document TypeTechnical Report
JAXA Category研究開発報告
NASA Subject CategorySpacecraft Propulsion and Power
ISSN(online)2433-2216
SHI-NOAA1830035002
Report NoJAXA-RR-18-008
URIhttps://repository.exst.jaxa.jp/dspace/handle/a-is/911107


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.