Improved Resonance Treatment for WIMS-Library-based nTRACER Direct Whole Core Calculation

Abstract

The resonance treatment methods based on the resonance integral data are improved for the applications to the nTRACER direct whole core calculation code in which the WIMS-IAEA library is to be optionally used. The Dancoff factor based equivalence theory is implemented via the background cross sections (XSs) evaluated by the enhanced neutron current method. A new resonance interference treatment method, based on the intermediate resonance (IR) representation of the flux spectrum, is introduced to generate the resonance interference factors (RIFs) on-the-fly. The concept of resonance escape probability is also introduced to incorporate the spectrum reduction effect below each resonance in the low energy range. The temperature feedback effect on resonance self-shielding is incorporated by a linear interpolation on the RIFs. The performance of the WIMS library based nTRACER calculation is examined by comparing the groupwise XSs with McCARD tallied XSs for homogeneous configurations. A maximum reactivity error amounting about 50pcm is observed in the resonance range. For 3x3 mini-core and quarter core problems based on the VERA benchmark specifications, the assembly and pin power distributions are compared with the KENO-VI results. With the in-scattering based transport correction which is newly added to the epithermal range of the primary scattering isotopes, the error in the power distribution is reduced significantly. The performance of the on-the-fly RIF method is examined for homogeneous mixtures with four resonant nuclides. RIFs are compared and verified with reference RIFs generated from the two ultrafine group slowing down calculations for a resonant nuclide: one in mixture and the other in isolated configuration. The RIFs and the XSs show good agreement with the McCARD results. The multiplication factors for the homogeneous configurations show good agreement with the reference. The incorporation of temperature feedback effects shows good agreement. The performance of the method for the heterogeneous configuration is examined for VERA benchmark pin cells. The XSs are compared for VERA pin cell (1A) with McCARD while the reactivity is examined for all other pin cell problems. The results show good agreement with the reference. The method is further examined for heterogeneous burned fuel configuration with only four major resonant nuclides. The method gives large improvements in XSs. The method shows quite good results for the homogeneous as well as heterogeneous problems.