A general-purpose code for correlated sampling using batch statistics with MCNP6 for fixed-source problems

Abstract

Background Correlated sampling can be applied with batch statistics using MCNP6’s tally fluctuation chart (TFC) to reduce the uncertainty of a difference of tallies between a base case and a perturbed case without modifying the source code. The cases are correlated if they are run with the same random number sequence. Equations implemented Equations used to apply correlated sampling to the difference of two tallies are summarized from previous work. Equations used to apply correlated sampling to a ratio of two tallies, to the difference of two tallies divided by a third tally, and to other combinations of ratios are presented. New computer code A general-purpose, command-line-driven computer program, COSUBS (Correlated Sampling Using Batch Statistics), is presented that reads MCNP6 TFCs from output files and applies these equations. The user specifies the input relative perturbation, if desired, and any normalization tally, if desired. Example problems show how to run COSUBS and interpret the output.