Radial basis function networks applied to DNBR calculation in digital core protection systems

Cited 50 time in webofscience Cited 0 time in scopus
  • Hit : 345
  • Download : 0
The nuclear power plant has to be operated with sufficient margin from the specified DNBR limit for assuring its safety. The digital core protection system calculates on-line real-time DNBR by using a complex subchannel analysis program, and triggers a reliable reactor shutdown if the calculated DNBR approaches the specified limit.. However, it takes a relatively long calculation time even for a steady state condition, which may have an adverse effect on the operation flexibility. To overcome the drawback, a new method using a radial basis function network is presented in this paper. Nonparametric training approach is utilized, which shows dramatic reduction of the training time, no tedious heuristic process for optimizing parameters, and no local minima problem during the training. The test results show that the predicted DNBR is within about +/-2% deviation from the target DNBR for the fixed axial flux shape case. For the variable axial flux case including severely skewed shapes that appeared during accidents, the deviation is within about +/-10%. The suggested method could be the alternative that can calculate DNBR very quickly while guaranteeing the plant safety. (C) 2003 Elsevier Science Ltd. All rights reserved..
Publisher
PERGAMON-ELSEVIER SCIENCE LTD
Issue Date
2003-10
Language
English
Article Type
Article
Keywords

NEURAL-NETWORK; PREDICTION

Citation

ANNALS OF NUCLEAR ENERGY, v.30, no.15, pp.1561 - 1572

ISSN
0306-4549
URI
http://hdl.handle.net/10203/85211
Appears in Collection
NE-Journal Papers(저널논문)
Files in This Item
There are no files associated with this item.
This item is cited by other documents in WoS
⊙ Detail Information in WoSⓡ Click to see webofscience_button
⊙ Cited 50 items in WoS Click to see citing articles in records_button

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0