Computational analysis of the Discrete Ordinates Adjoint Function applicability in VVER Monte Carlo modelling

Abstract

The possible use of the adjoint function as an automated source for weight windows generation in the case of VVER-1000 benchmark transport calculations is presented in this paper. The numerical solution for obtaining the adjoint function is performed using the Discrete Ordinates code TORT. Interface software is used to translate the adjoint function into weight windows’ input parameters. A deep penetration problem has been modelled for the VVER-1000 Mock Up, built at the NRI/Rez’s LR-0 critical assembly for VVER-1000 benchmark studies. The neutron flux is evaluated in three locations placed on the azimuthal symmetry axis, radially from the reactor core to the shielding. The calculations are performed in three different approaches: by MCNP code using the weight windows obtained through the adjoint function, by MCNP code using weight windows created by the author through the standard procedure and by the TORT code. The results are compared in terms of integral fluxes and the validity of the adjoint function in variance reduction for VVER reactor Monte Carlo calculations is justified.