Use of Local Variation Method for Nuclear Design Calculations

Abstract

The paper presents a local variation method for solving steady-state equations that describe neutron diffusion. The method is based on a variation principle for steady-state diffusion equations and a direct search for the minimum of a corresponding functional and finite element problem definition. The method helps avoid large matrices typical of the conventional finite element method. Benchmark problem calculations of fuel assembly power show a root-mean-square accuracy of ~2%.