首页 | 本学科首页   官方微博 | 高级检索  
     


ARIANEXPERT: A knowledge based system to analyze ARIANE's mission data
Authors:Jean-Marc Brenot   Yann Parrod   Didier Aubin  Christian Parquet
Affiliation:

a MATRA ESPACE, Z.1. du Palays, 31 rue des Cosmonautes, 31077, Toulouse Cedex, France

b ARIANESPACE, Boulevard de l'Europe, B.P. 177, 91006, Evry Cedex, France

Abstract:The ARIANE launcher post mission analysis is done at ARIANESPACE. This activity is called the ‘level 0 post flight analysis’ (PFA) and is carried out after each launch by about 60 engineers who are working together under the leadership of ARIANESPACE.

The PFA is one of the most critical of ARIANE operations, for several reasons:

• - The launch rate (8 a year for ARIANE 4) leaves a very short time to carry out all the verification work. Moreover, the PFA is a mandatory step before authorizing the next launch.
• - The complexity of the ARIANE launcher results in a very high demand on the PFA engineers. Moreover, there are problems of availability of people with relevant expert knowledge (characterized by a substantial staff turn-over during the 10 year life duration of ARIANE 4) which could potentially result in errors or omissions.

It is very important to be able to take into account the experience of the preceding flights and to record the results and the knowledge accumulated for each launch.

• - The quality and the reliability of the PFA mainly depends on the accessibility of data and on the used methodology.

Because the PFA is still largely done manually, and does not benefit from improved methodologies and advanced technologies providing computerized support for data processing and diagnosis, ARIANESPACE has sponsored MATRA ESPACE for the development of a knowledge based system, called ARIANEXPERT, for supporting the PFA activity. This system combines AI techniques and numerical analysis techniques, together with advanced graphical capabilities.

A prototype has been delivered in April 1990 and has been used since 6 months by ARIANESPACE during real PFAs. Several lessons have been drawn from this operational experience and are described in this paper. They concern:

• - The utility and justification of the use of AI techniques mostly coming from the explanation capabilities and the stress put on capturing the expert knowledge.
• - The difficulties associated with the integration of such systems in the exploitation of ARIANE due to the introduction of very new tasks.
• - The user point of view which evolved from reluctant to convinced.
Keywords:Procedural expert system   operator assistant   space operations   diagnosis   data analysis   fault detection
本文献已被 ScienceDirect 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号