An intelligent eddy current signal evaluation system to automate the non-destructive testing of steam generator tubes in nuclear power plant

  • Kang, Soon-Ju (Artificial Intelligence Research Division, Korea Atomic Energy Research Institute, P.O.Box 7, Daeduk-Danji, Daejon, Korea, 305-606) ;
  • Ryu, Chan-Ho (Artificial Intelligence Research Division, Korea Atomic Energy Research Institute, P.O.Box 7, Daeduk-Danji, Daejon, Korea, 305-606) ;
  • Choi, In-Seon (Artificial Intelligence Research Division, Korea Atomic Energy Research Institute, P.O.Box 7, Daeduk-Danji, Daejon, Korea, 305-606) ;
  • Kim, Young-Ill (Artificial Intelligence Research Division, Korea Atomic Energy Research Institute, P.O.Box 7, Daeduk-Danji, Daejon, Korea, 305-606) ;
  • Kim, kill-Yoo (Artificial Intelligence Research Division, Korea Atomic Energy Research Institute, P.O.Box 7, Daeduk-Danji, Daejon, Korea, 305-606) ;
  • Hur, Young-Hwan (Artificial Intelligence Research Division, Korea Atomic Energy Research Institute, P.O.Box 7, Daeduk-Danji, Daejon, Korea, 305-606) ;
  • Choi, Seong-Soo (Dept. of electronic Application, Research Center, Korea Electric Power Corporation) ;
  • Choi, Baeng-Jae (Dept. of electronic Application, Research Center, Korea Electric Power Corporation) ;
  • Woo, Hee-Gon (Dept. of electronic Application, Research Center, Korea Electric Power Corporation)
  • Published : 1992.10.01

Abstract

This paper describes an intelligent system to automatic evaluation of eddy current(EC) signal for Inspection of steam generator(SG) tubes in nuclear power plant. Some features of the intelligent system design in the proposed system are : (1) separation of representation scheme ,or event capturing knowledge in EC signal and for structural inspection knowledge in SG tubes inspection; (2) each representation scheme is implemented in different methods, one is syntactic pattern grammar and the other is rule based production. This intelligent system also includes an data base system and an user interface system to support integration of the hybrid knowledge processing methods. The intelligent system based on the proposed concept is useful in simplifying the knowledge elicitation process of the rule based production system, and in increasing the performance in real time signal inspection application.

Keywords