• Home
  • Current congress
  • Public Website
  • My papers
  • root
  • browse
  • IAC-14
  • D2
  • P
  • paper
  • anomaly detection configured as a combination of state observer and mahalanobis-taguchi method for fault detection, isolation and reconfiguration

    Paper number

    IAC-14,D2,P,4,x22274

    Author

    Dr. Yusuke Maru, Japan Aerospace Exploration Agency (JAXA), Japan

    Coauthor

    Mr. Hatsuo Mori, IHI Corporation, Japan

    Coauthor

    Dr. Shinsuke Takeuchi, Japan Aerospace Exploration Agency (JAXA), Japan

    Coauthor

    Dr. Takayuki Yamamoto, Japan Aerospace Exploration Agency (JAXA), Japan

    Coauthor

    Dr. Satoshi Nonaka, Japan Aerospace Exploration Agency (JAXA), Japan

    Coauthor

    Mr. Noriyoshi Mizukoshi, IHI Corporation, Japan

    Coauthor

    Mr. Takashi Ogai, IHI Corporation, Japan

    Coauthor

    Mr. Kenichi Kimoto, IHI Corporation, Japan

    Coauthor

    Prof. Yoshifumi Inatani, Japan Aerospace Exploration Agency (JAXA), Japan

    Year

    2014

    Abstract
    Institute of Space and Astronautical Science (ISAS) and IHI corporation have been researching FDIR (Fault Detection, Isolation and Reconfiguration) technologies as a part of an integrated study of Advanced Reusable Rocket Vehicles. Real time anomaly detecting function is recognized as one of the most important elements to realize such a vehicle.  
    Currently, bottom-up type anomaly detecting logic structured based on FMEA is usually used for this purpose, but it requires large amount of time and labor.
    To improve this process, ISAS and IHI designed a cutting-edge method for anomaly detection which is configured as a combination of state observer and MT (Mahalanobis-Taguchi) method.
    This method, which can generate a unit space with a small number of normal operation data, will minimize the development efforts.
    In 2012, we performed an engine hot firing test with simulated anomaly conditions. The obtained data was processed with the new method mentioned above, and the simulated anomaly was detected as expected.
    In 2013 test campaign, we had this system on-board and demonstrate its real time performance.
    This paper describes the research history and outline of the system and method which we are currently pursuing.
    Abstract document

    IAC-14,D2,P,4,x22274.brief.pdf

    Manuscript document

    (absent)