Generalization of knowledge acquired by a reactor core monitoring system based on a neuro-fuzzy algorithm

R. Kozma, S. Sato, M. Sakuma, M. Kitamura, T. Sugiyama

Research output: Contribution to journalArticlepeer-review

7 Citations (Scopus)

Abstract

The conceptual description of an automatic reactor core monitoring system based on the analysis of signals of in-core and ex-core neutron detectors is given. The applied system utilizes massive parallel computing. The state of the reactor is monitored by multiple artificial neural networks (ANNs) and an algorithm based on fuzzy logics is applied to combine information from different ANNs. Important elements of the method are: (1) structural learning algorithm with forgetting and (2) a priori assessment of the accuracy of learning by the neural network. The results are applied to early detection of boiling anomaly in a nuclear reactor.

Original languageEnglish
Pages (from-to)203-214
Number of pages12
JournalProgress in Nuclear Energy
Volume29
Issue number3-4
DOIs
Publication statusPublished - 1995

Keywords

  • Neutron noise
  • anomaly detection
  • artificial neural network
  • coolant boiling
  • fuzzy consensus
  • structural learning

ASJC Scopus subject areas

  • Nuclear Energy and Engineering
  • Safety, Risk, Reliability and Quality
  • Energy Engineering and Power Technology
  • Waste Management and Disposal

Fingerprint Dive into the research topics of 'Generalization of knowledge acquired by a reactor core monitoring system based on a neuro-fuzzy algorithm'. Together they form a unique fingerprint.

Cite this