Adjoint and forward Monte Carlo coupled weight window generator for variance reduction

Abstract

Among the various variance-reduction methods in Monte Carlo calculations, one of the most widely used techniques is the weight window method. The MCNP code provides a weight window generator (WWG) option. In WWG of MCNP, the importance of a cell is estimated by the virtual sampling method during normal Monte Carlo calculation. But, the performance of WWG tends to deteriorate in deep penetration problems. To enhance the performance of the weight window method, importance estimation by the deterministic adjoint calculation has been proposed. However, this approach is possible only when the related deterministic code and interface program are available. The midway coupling method is a surface tally technique that calculates detector response based on the reciprocity theorem, but it does not provide an importance generator. In this paper, a new weight window generation method, called adjoint and forward Monte Carlo coupled WWG, is proposed to overcome the drawbacks of the current WWG in MCNP and the deterministic adjoint calculation method.