A New Approach to Estimate Importance for Weight Window in Forward Monte Carlo Calculations

摘要

A new importance estimation method, which is based on the adjoint function definition, has been proposed especially for the weight window (WW) technique of MCNP, which is well known as one of the most powerful variance-reduction techniques in Monte Carlo codes. The method employs the scattering point base importance estimation, unlike the WW generator (WWG) of MCNP for the point detector function. Every scattering point has an adjoint contribution to the detector, with which a space-, energy-, and angle-dependent importance for WW could be estimated. From the numerical test calculations, the basic performance was confirmed to be better than WWG by comparing figure-of-merit values. It would be expected that the performance of WWG would be well improved by using the present method instead of the current MCNP routine of accumulating the detector contribution for the F5 tally. The presently proposed method would be a strong tool to estimate the importance applicable to various variance-reduction techniques in Monte Carlo codes.