Evaluating the use of surface source banking to accelerate Monte Carlo transport simulations of far-field particle fluxes

Abstract

In order to enhance the verifiability and usability of surface source banking as a far-field flux and dose simulation acceleration method for Monte Carlo neutron transport codes, two surface source stationarity criteria were developed and evaluated. Surface sources were considered defined, or an accurate proxy for fission sources in eigenvalue simulations once enough particles have been banked such that these criteria were met. One criterion utilizes multi-dimensional Shannon entropy to indicate the stationarity of the surface source in physical space and energy. The other criterion uses functional expansions to track the stationarity of Legendre coefficients associated with spatially-dependent banked effective neutron reaction rates with different Z materials. The completion of a test case with an OpenMC model of an MK2 TRIGA facility indicated agreement between the two criteria. Effects of oversampling a surface source that met the stationarity criteria, as well as potential limitations of the surface source banking method itself were also examined via the test case.