Intelligent method of plant dynamics behavior estimation by effectively applying similar cases

Akio Gofuku, Hidekazu Yoshikawa, Atsushi Numoto

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)


In order to accomplish efficient execution of a dynamic simulation of engineering systems, it is important to construct suitable mathematical models for the simulation. In the construction of the mathematical models, it is necessary to estimate the system's behavior to suppose the phenomena which are needed to be modeled. The case-based reasoning is considered to be a powerful tool to estimate the outline of system's behavior because we often estimate it from the similar cases which are stored as our experience or in literature. In this study, the technique based on similar cases is investigated to estimate the outline of time-responses of several important variables of pressurized water reactor (PWR) plants at a small break loss of coolant accident (SBLOCA). The registered cases in case base are gathered from various reports and the authors' numerical simulations related with SBLOCA of PWR plants. The functions to use in the case retrieval are formed from the characteristic features of SBLOCA of PWR plants. On the other hand, the rules to use in the case refinement are obtained from the qualitative and quantitative consideration of plants' behaviors of the cases in the case base. The applicability of the technique is discussed by two simple estimation trials of plant behavior.

Original languageEnglish
Pages (from-to)548-562
Number of pages15
JournalNippon Genshiryoku Gakkaishi/Journal of the Atomic Energy Society of Japan
Issue number6
Publication statusPublished - 1994
Externally publishedYes


  • PWR type reactors
  • case-based reasoning
  • loss of coolant
  • mathematical models
  • nuclear power plants
  • numerical simulation model
  • ptant behavior

ASJC Scopus subject areas

  • Nuclear Energy and Engineering


Dive into the research topics of 'Intelligent method of plant dynamics behavior estimation by effectively applying similar cases'. Together they form a unique fingerprint.

Cite this