Reactivity Calculations for the RMC LEU Fuelled SLOWPOKE-2 Reactor

Abstract

Shortly after the commissioning of the low enrichment uranium (LEU) fuelled SLOWPOKE-2 research reactor at the Royal Military College, the trend of the excess reactivity versos uniform core temperature was determined experimentally at a power of 10 W. Although both experiment and previously calculated results displayed similar trends, in that the reactivity variation with temperature exhibited a maximum, there were significant differences between the two sets of rests. The maximum of the calculated reactivity core occurred at a uniform reactor temperature of about 12 C compared to the experimental value of 3 3 %. The absolute values of the reactivities differed by a much greater margin: the system excess reactivity was experimentally determined to be 3.2 mk, while calculations produced reactivity values of about 80 mk out of WIMS-CRNL reducing to about 45 mk from CITATION. Switching from the previously used Winfrith 69 group library to the ENDDF/B-V 89 group library accounted for 71% of the initial reactivity trend discrepancy alone. * Various more detailed core models were simulated in order to resolve the remaining trend differences. Inclusion of the reactor material expansion, beryllium reflector impurities, thermal column and control rod effects all independently increased the remaining discrepancy. The transverse buckling representation in WIMS-CRNL was among the factors suspected for the remaining reactivity trend discrepancy and the reactivity differences between WIMS-CRNL, CITATION and experiment. In previous calculations, a unique transverse buckling value was wed for all neutron energy groups and core temperatures. This transverse buckling term was then combined with the macroscopic cross sections in the main transport calculations in WIMS-CRNL to simulate axial neutron leakage. The computer code DELBUCK was written and developed at RMC to calculate the group and temperature dependent transverse buckling values from the CITATION-calculated fluxes. These buckling values were then included in the successive WIMS-CRNL/CITATION simulations. Although no significant effect was obtained on the reactivity trend with temperature, the difference between WIMS-CRNL and CITATION was reduced to 1 mk and the difference between CITATION and experiment to 16 to 20 mk. This latter reactivity discrepancy has been accounted for by simplifications in the core model representation and the choice of processes in WIMS-CRNL.