Analysis of measurements for a Uranium-free core experiment at the BFS-2 critical assembly

Hunter

JNC TN9400 99-049, 74 Pages, 1999/04

JNC-TN9400-99-049.pdf:2.03MB

This document describes a series of calculations that were carried out to model various measurements from the BFS-58-1-I1 experiment. BFS-58-1-I1 was a mock-up of a uranium-free, Pu burning core at BFS-2, a Russian critical assembly operated by IPPE. The experiment measured values of keff, Na void reactivity worth, material sample reactivity worths and reaction rate ratios. The experiments were modelled using a number of different methods. Basic nuclear data was taken from JENDL-3.2, in either 70 or 18 groups. Cross-section data for the various material regions of the assembly were calculated by either SLAROM or CASUP; the heterogeneous structure of the core regions was modelled in these calculations, with 3 different options considered for representing the (essentially 2d) geometry of the assembly components in a 1D cell model. Whole reactor calculations of flux and keff were done using both a diffusion model (CITATION) and a transport model (TWOTRAN2), both using an RZ geometry. Reactivity worths were calculated both directly from differences in keff values and by using the exact perturbation calculations of PERKY and SN-PERT (for CITATION and TWOTRAN2, respectively). Initial calculations included a number of inaccuracies in the assembly representation, a result of communication difficulties between JNC and IPPE; these errors were removed for the final calculations that are presented. Calculations for the experiments have also been carried out in Russia (IPPE) and France (CEA) as part of an international comparison exercise, some of those results are also presented here. The calculated value of keff was 1.1%k/k higher than the measured value, Na void worth C/E values were 1.06; these results were considered to be reasonable. (Discrepancies in certain Na void values were probably due to experimental causes, though the efect should be investigated in any future experiments.) several sample worth values were small compared with calculational uncertaint