Forward-weighted CADIS method for variance reduction of Monte Carlo reactor analyses

The FW-CADIS method for optimizing global distributions has been extended to k-eigenvalue calculations for reactor analyses and initially tested on a generic 2-D PWR core model. The method requires two approximate discrete ordinates calculations (one forward and one adjoint) to generate consistent source biasing and weight window parameters for the subsequent MC simulation and does not require any modifications to existing MC codes. Although more testing and thorough analysis of results are still needed, the potential of this method to improve the efficiency and reliability of MC calculations for reactor analyses appears promising. Planned next steps include (1) application to a 3-D core model, (2) implementation to automatically use the converged fission source distribution from the forward Denovo calculation as the starting source for the MC calculation, (3) optimization for directly calculating reaction rates, and (4) more thorough results analysis and testing.