Faire avancer la sûreté nucléaire

La Recherchev2


Monte Carlo techniques implemented in the MORET 5 code, description and validation



Email :

Mot de passe :

Titre du congrès :13th UK Monte Carlo User Group Meeting (MCNEG 2007) Ville du congrès :Teddington Date du congrès :28/03/2007

Type de document > *Congrès/colloque

Mots clés > criticité, Monte Carlo, MORET, woodcock

Unité de recherche > IRSN/DSU/SEC/LERD

Auteurs > BERNARD Franck, DOUCET Michel, FORESTIER Benoit, JACQUET Olivier, MISS Joachim, VERBOOMEN Bernard

Date de publication > 29/03/2007


  The MORET code is a three dimensional criticality Monte Carlo code which calculates the effective multiplication factor (keff) of any configuration more or less complex as well as reaction rates in the different volumes of the geometry and the leakage out of the system. The multigroup version of MORET is used in the APOLLO2-MORET 4 standard route of CRISTAL, the French criticality package. It is the most commonly used Monte Carlo code for French criticality calculations.
The MORET code includes various simulation techniques (the conventional method, the stratified sampling, the super-history powering, the use of fission matrix and the use of importance function), each corresponding to a different strategy for redistributing neutron sources at the beginning of each cycle. These options have been implemented into the MORET code in order to study the difficulty, usually connected to the choice of the first generation sources in some particular configurations, and known as “problem of loosely coupled systems” [2]. All those options are available in the MORET code but only the conventional method has been widely validated. Full validation and performances studies of the other techniques are performed and should lead to some user recommendations.
Beside this, a recent development consists in the implementation of a new simulation option designed to perform neutron tracking using a method known as the Woodcock tracking method [3]. The advantages of this method include fast simulation in complex geometry and relatively simple handling of complicated geometrical objects. The main drawbacks are the inability to use the track length estimator and the bad performances obtained, in some configurations, when the system under study exhibits large cross section variations. Such method has been successfully implemented in the multigroup version of MORET 5, allowing the user to specify which part of the geometry should be treated using the Woodcock tracking method.
The purpose of this presentation is double:
·        Firstly to describe the various Monte Carlo strategies implemented in the MORET code, and to show the results of the validation concerning 3 of them: the conventional method, the stratified sampling and the super-history powering. This work is based on the extensive validation database (more than 1800 benchmarks) used for MORET in the framework of the CRISTAL project,
·        Secondly to describe the Woodcock tracking method and its implementation in the MORET code. The preliminary results concerning the performances of this method and first results regarding the gain on computation time obtained for various geometries and parameterization of the method will be presented.