Effective biasing schemes for duct streaming problems.

ABSTRACT

The effective use of biasing for the Monte Carlo solution of a void streaming problem is essential to obtaining a reasonable result in a reasonable amount of time. Most general purpose Monte Carlo shielding codes allow for the user to select the particular biasing techniques best oriented to the particular problem of interest. The biasing strategy for void streaming problems many times differs from that of a deep penetration problem. The key in void streaming is to bias particles into the streaming path, whereas in deep penetration problems the biasing is aimed at forcing particles through the shield. Until recently, the biasing scheme in the SCALE SAS4 shielding module was considered inadequate for void streaming problems due to the assumed one-dimensional nature of the automated bias prescription. A modified approach to the automated biasing in SAS4 has allowed for significant gains to be realised in the use of the code for void streaming problems. This paper applies the modified SAS4 procedures to a spent fuel storage cask model with vent ports. The results of the SAS4 analysis are compared with those of the ADVANTG methodology, which is an accelerated version of MCNP. Various options available for the implementation of the SAS4 methodology are reviewed and recommendations offered.