Enhancement of REBUS-3/DIF3D for whole-core neutronic analysis of prismatic very high temperature reactor (VHTR).

Abstract

Enhancements have been made to the REBUS-3/DIF3D code suite to facilitate its use for the design and analysis of prismatic Very High Temperature Reactors (VHTRs). A new cross section structure, using table-lookup, has been incorporated to account for cross section changes with burnup and fuel and moderator temperatures. For representing these cross section dependencies, three new modules have been developed using FORTRAN 90/95 object-oriented data structures and implemented within the REBUS-3 code system. These modules provide a cross section storage procedure, construct microscopic cross section data for all isotopes, and contain a single block of banded scattering data for efficient data management. Fission products other than I, Xe, Pm, and Sm, can be merged into a single lumped fission product to save storage space, memory, and computing time without sacrificing the REBUS-3 solution accuracy. A simple thermal-hydraulic (thermal-fluid) feedback model has been developed for prismatic VHTR cores and implemented in REBUS-3 for temperature feedback calculations. Axial conduction was neglected in the formulation because of its small magnitude compared to radial (planar) conduction. With the simple model, the average fuel and graphite temperatures are accurately estimated compared to reference STAR-CD results. The feedback module is currently operational for the non-equilibrium fuel cycle analysis option of REBUS-3. Future work should include the extension of this capability to the equilibrium cycle option of the code and additional verification of the feedback module. For the simulation of control rods in VHTR cores, macroscopic cross section deviations (deltas) have been defined to account for the effect of control rod insertion. The REBUS-3 code has been modified to use the appropriately revised cross sections when control rods are inserted in a calculation node. In order to represent asymmetric core blocks (e.g., fuel blocks or reflector blocks containing asymmetric absorber rods), surface-dependent discontinuity factors based on nodal equivalence theory have been introduced into the nodal diffusion theory option of the DIF3D code (DIF3D-nodal) to improve modeling accuracy. Additionally, the discontinuity factors based on the Simplified Equivalence Theory (SET) have been incorporated as an alternative and may be employed for both the DIF3D-nodal and DIF3D-VARIANT (nodal transport) solution options. Two- and three-dimensional core calculations have been performed using the routines developed and modified in this work, along with cross sections generated from single fuel block and one-dimensional or two-dimensional fuel-reflector model. Generally, REBUS-3/DIF3D results for the core multiplication factor and power distribution are found to be in good agreement with reference results (generated with MCNP continuous energy calculations) particularly when discontinuity factors are applied. The DIF3D-VARIANT option was found to provide a more accurate solution in its diffusion approximation than the DIF3D-nodal option. Control rod worths can be estimated with acceptably small errors compared to MCNP results. However, estimation of the core power tilt needs to be improved by introducing the surface-dependent discontinuity factor capability in DIF3D-VARIANT.