Statistical method application to knowledge base building for reactor accident diagnostic system

Kazuo Yoshida, Masao Yokobayashi, Kiyoshi Matsumoto, Atsuo Kohsaka

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

In the development of a knowledge based expert system, one of key issues is how to build the knowledge base (KB) in an efficient way with keeping the objectivity of KB. In order to solve this issue, an approach has been proposed to build a prototype KB systematically by a statistical method, factor analysis. For the verification of this approach, factor analysis was applied to build a prototype KB for the JAERI expert system DISKET. To this end, alarm and process information was generated by a PWR simulator and the factor analysis was applied to this information to define taxonomy of accident hypotheses and to extract rules for each hypothesis. The prototype KB thus built was tested through inferring against several types of transients including double-failures. In each diagnosis, the transient type was well identified. Furthermore, newly introduced standards for rule extraction showed good effects on the enhancement of the performance of prototype KB.

Original languageEnglish
Pages (from-to)1002-1012
Number of pages11
JournalJournal of Nuclear Science and Technology
Volume26
Issue number11
DOIs
Publication statusPublished - 1989
Externally publishedYes

    Fingerprint

Keywords

  • DISKET
  • Double failures
  • Expert systems
  • Factor analysis
  • Knowledge base
  • PWR type reactors
  • Reactor accident diagnostic systems
  • Reactor simulators
  • Statistical method

ASJC Scopus subject areas

  • Nuclear and High Energy Physics
  • Nuclear Energy and Engineering

Cite this