Final Report on Utilization of TRU TRISO Fuel as Applied to HTR Systems Part II: Prismatic Reactor Cross Section Generation

Abstract

In Very High Temperature Reactors (VHTRs), the long mean-free-path and large migration area of neutrons leads to spectral influences between fuel and reflector zones over long distances. This presents significant challenges to the validity of the classic two-step approach of cross section preparation wherein infinite lattice transport calculations are performed on relatively small physical domains (e.g. single assembly) in order to compute homogenized few-group cross sections for whole core analysis.Effects of the inner and outer reflectors challenge the classical two-step approach, while burnable poison locations affect neighboring assemblies as well. Use of transuranics-only (TRU) Deep Burn fuel in a prismatic VHTR (DB-VHTR) presents the additional challenge of producing vastly different neutron spectra between fresh and burned fuel.Modeling the VHTR also requires the ability to treat in the code the hexagonal geometry, the presence of TRISO particles, whose distribution in the fuel compacts is stochastic, and large lattice domains. Currently, DRAGON can treat these features: the hexagonal geometry is in place, the large domains may be treated by the method of characteristics (MOC) or collision probability (CP), and the doubleheterogeneity treatment can be used for TRISO particles.Therefore, in order to complement the analysis and design activities for Deep Burn in prismatic High Temperature Reactors (HTRs), this study was performed in order to provide some initial guidance on the necessary procedures to model the DB-VHTR with sufficient accuracy. The purpose of this study is to provide evaluation of the tradeoffs in accuracy between domain size in the lattice calculation and number of energy groups used in the whole core calculation. The lattice calculations were performed using the DRAGON-4 code and the whole core calculations using the Idaho National Laboratory code INSTANT, both of which are described in the following sections. The accuracy of the whole-core calculations was judged by comparison of Eigenvalues and ¿ssion rate shapes to results from MCNP calculations. The number of groups used in the lattice calculations was fixed at 295 and two different domain sizes were used, a single block and a supercell. Primarily of interest in this study were the effects of the reflector on the spectrum and, in turn, the adequacy of the cross section sets.Also included in this report are results from a preliminary study aimed at determining the effect of the presence of reflector on the burnup characteristics in a peripheral block. Depletion was performed in a supercell in order to observe the differences across a block of interest.The results showed that using the Method of Characteristics (MOC) in DRAGON, both a single block and a supercell leads to good predictions of the Eigenvalue and of the relative fission rates of each fuel block in the core if enough groups are kept in INSTANT (e.g., 26 groups). The supercell path produced the best results compared to the equivalent calculations using the single-block path....