VERIFICATION OF THE DIFFUSION AND TRANSPORT SOLVERS WITHIN DIF3D FOR 3D HEXAGONAL GEOMETRIES

A G Nelson,M A Smith,F Heidet

doi:10.1051/epjconf/202124710030

Abstract

The DIF3D code (DIFfusion 3D) has been a workhorse of fast reactor analysis work at Argonne National Laboratory for over 40 years. DIF3D was primarily built in the late 1970s as a three-dimensional multigroup diffusion equation solver operating on semi-structured grid geometries. In the mid-1990s, transport capabilities needed for high-leakage reactor configurations were added to DIF3D with the variational anisotropic nodal transport approach. Recent reactor design activities at Argonne are requiring that a thorough verification of the Argonne Reactor Computation (ARC) codes be performed. With DIF3D being central to the entire ARC system, the verification efforts are focused on the 3D Cartesian, 3D triangular, and 3D hexagonal core geometry options of DIF3D. Validation activities, while needed for the ongoing design activities at Argonne, are handled at a project-specific level. This paper summarizes the verification work so far on the forward and adjoint forms of the fixed source, inhomogeneous fixed source, and k-eigenvalue steady state transport and diffusion equations as implemented specifically for 3D triangular and hexagonal geometries in DIF3D. Since analytic solutions of the neutron diffusion and transport equations are either limited in scope or not possible, this verification required multiple tiers of problems unique to each solver and geometry type, each testing features independent and complementary arguments for why this separate testing of functionalities is acceptable. This separate testing was also supplemented with a high-level integral check of each the diffusion and transport capabilities and applicable geometries.

Highlights

The DIF3D code (DIFfusion 3D) was primarily built as a three-dimensional solver of the multigroup diffusion equation for semi-structured grid geometries using a finite difference spatial differencing methodology [1]
Later work, motivated by the research and development to support fast spectrum reactors that increase proliferation tolerance by eliminating blanket assemblies, the DIF3D-VARIANT solver [3] was added to DIF3D in the mid 1990s, and later updated [4], that extends the nodal concept to transport via a variational three-dimensional transport method applying spherical harmonic expansions in angle
The DIF3D-FD diffusion solver can be applied to every geometry type in DIF3D except for hexagonal and hexagonal-Z geometries, where a triangular-Z mesh geometry is instead used to represent hexagonal reactor designs

Summary

Introduction

The DIF3D code (DIFfusion 3D) was primarily built as a three-dimensional solver of the multigroup diffusion equation for semi-structured grid geometries using a finite difference spatial differencing methodology [1]. This capability still exists and is commonly referred to as DIF3D-FD. Later work, motivated by the research and development to support fast spectrum reactors that increase proliferation tolerance by eliminating blanket assemblies, the DIF3D-VARIANT solver [3] was added to DIF3D in the mid 1990s, and later updated [4], that extends the nodal concept to transport via a variational three-dimensional transport method applying spherical harmonic expansions in angle. The DIF3D-Nodal diffusion and DIF3D-VARIANT solvers only support 2D and 3D dimensional Cartesian and hexagonal geometries

Objectives

Results

Conclusion