Investigation of sub-cell homogenization for PHWR lattice cells using superhomogenization factors

Abstract

To avoid the computational effort associated with full-core neutron transport calculations, full-core neutronics calculations for Pressurized Heavy-Water Reactors (PHWRs) are usually performed in diffusion theory using an approximate core model, whereby only two energy groups are utilized and two-group neutronic properties (i.e. macroscopic cross sections and diffusion coefficients) are homogenized in two dimensions over large sub-domains, each corresponding to a 28.6 cm x 28.6 cm lattice cell. The lattice cell is the elementary geometrical unit describing the rectangular array of fuel channels comprising the PHWR core. The use of lattice-cell homogenization introduces some computational errors. One possible way to reduce such homogenization errors is to sub-divide the lattice cell into sub-cells and perform sub-cell-level homogenization. In this study, the PHWR lattice cell is divided into 3 x 3 sub-cells. Full-cell-averaged, as well as sub-cell-averaged two-group cross-sections, are generated for subsequent use in an equivalent two group two-dimensional diffusion model. Cross sections with Superhomogenization (SPH) [Hebert, 2009] factors are also utilised in an attempt to improve accuracy. The effect of using different homogenization models (full cell, partial cell, partial-cell with SPH-corrected cross sections) is tested on a two-dimensional partial-core model consisting of 3 x 3 lattice cells (bundles). Results from reference transport model with detailed geometry 69-group are compared with cell-homogenized two-group diffusion results obtained using full-cell homogenization and sub-cell homogenization with and without SPH correction factors. The application of sub-cell homogenization, as well as the use of SPH correction factors, is found to have only a minimal effect on computational accuracy.