Nouvelles évolutions majeures dans le code Monte Carlo MORET. Développement d'une version ponctuelle
Titre du congrès :ICNC 2007
Ville du congrès :Saint-Petersbourg
Date du congrès :28/05/2007
The version 5.A.1 is the next release of the Monte Carlo code MORET whose multigroup version is included in the French criticality package CRISTAL. It combines all the capabilities already available in the multigroup version (a powerful modular geometry, graphics productivity tools, many tracking algorithms, a wide validation database
) with many new and enhanced features.
Why such new developments?
Most modern criticality safety Monte Carlo codes like MORET have reached a good level of satisfaction concerning the simulation algorithms implemented. Discrepencies on final results are mainly due to the nuclear data used. That is why, knowing and controlling the way of processing the nuclear data libraries used by these codes is nowadays crucial, and important for the strategic objectives of the IRSN, which carries out research, analysis and work within the fields of nuclear safety in France.
So far, the cross-sections used by the MORET code (developed by IRSN) could only be taken from external multigroup libraries resulting, by default, from preliminary calculations performed with the APOLLO2 assembly code (processed from the JEF evaluation). It has been then decided to plan new developments in MORET 5 to control the whole process of criticality calculation, from the reading of nuclear data to the Monte Carlo simulation. The purpose of this paper is to present the two main new features and first results related to this new version:
Criticality calculations will be able now to be performed either with a multigroup or a continuous energy treatment;
The implementation of the Woodcock tracking method should improve computation time for non-homogenized configurations with many small volumes.
Moreover, in criticality safety studies, one often needs to perform a huge number of calculations because many different parameters may vary, with various impacts on the Keff. It is obviously interesting for such studies to use homogenized configurations to reduce the number of volumes, and hence the calculation time of the Monte Carlo simulation. In these cases the multigroup version of the MORET code used in conjunction with a deterministic assembly code is a very efficient approach:
to treat the local heterogeneity with simple 1D or 2D cells and to apply related models for self-shielding and possibly homogenisation ;
to treat the overall heterogeneity and geometrical complexity with 3D Monte Carlo tracking without wasting computation time in treating local details since they have been taken into account in the previous step.
However, this multigroup approach may sometimes give incorrect results, in particular for configurations with very steep neutron flux heterogeneities, and when the self-shielding models inherent to the multigroup treatment are not well adapted. The continuous energy approach is therefore essential.
This paper describes the main topics to ensure proper development of a continuous-energy version of MORET:
control of the basic nuclear data tools to generate continuous energy cross-section libraries for MORET from any evaluation (JEF, ENDF/B
adaptation of the tracking routines in MORET to the new point-wise description (needs in memory, anisotropy and thermal treatment
study of ways to reduce simulation time (including the Woodcock tracking strategy).
Besides, another way to increase the capabilities of MORET was followed by implementing the reading of cross-sections from other multigroup libraries. First results are also presented in this paper.