®®®® SIIA Público

Título del libro: Joint International Topical Meeting On Mathematics And Computations And Supercomputing In Nuclear Applications, M And C + Sna 2007
Título del capítulo: LPPF prediction in a BWR fuel lattice using artificial neural networks

Autores UNAM:
JUAN LUIS FRANCOIS LACOUTURE; CECILIA MARTIN DEL CAMPO MARQUEZ;
Autores externos:

Idioma:

Año de publicación:
2007
Palabras clave:

Artificial neural networks; BWR; Fuel lattice design


Resumen:

The amount of CPU time required for a typical fuel lattice calculation is considerable; even more if this type of calculations is attempted for carrying out an exhaustive exploration of a search space in a combinatorial optimization problem. This situation has motivated to look at the possibility of replacing the best estimate HELIOS code by an Artificial Neural Network (ANN), in order to predict the major Boiling Water Reactor (BWR) fuel lattice parameters. In this paper the development of an Artificial Neural Network and some results obtained during the training process and application of the ANN are presented. The obtained ANN is used to predict the maximum local power peaking factor (LPPF) in a typical 10×10 BWR fuel lattice. The LPPF is predicted at beginning of life (0.0 MWD/T), at 40% voids moderator density, fuel temperature of 793°K, and moderator temperature of 560°K. The axial location of the fuel lattice is just above the natural uranium zone, called the power lattice, at the bottom, of the reactor core. Finally the trained Artificial Neural Network is applied in order to predict the LPPF of some BWR fuel lattices, which have been used in the real operation of the Laguna Verde Nuclear Power Plant (LVNNP) in Mexico.


Entidades citadas de la UNAM: