A Comparison of Thermal-Neutron-Activation Measurements and Monte Carlo Calculations in Light-Water-Moderated Uranium Cells

Thermal disadvantage factors and spectral indexes have been measured in a variety of light-water-moderated lattices. One series contained slightly enriched uranium rods in hexagonal geometry and another series used natural-uranium fuel in slab geometry. The detectors used were ¹⁶⁴Dy, ¹⁷⁶Lu, and ²³⁹Pu. Full energy range (0 to 10 MeV) Monte Carlo calculations with explicit cell-geometry representations were performed using the RECAP program. In addition, thermal energy range (0 to 0.625 eV) calculations were obtained with the Monte Carlo program MARC as well as with the integral transport-theory-code THERMOS. The purpose of these investigations was to test the adequacy of the various water scattering kernels—Nelkin, Koppel, and Haywood—for a broad range of thermal-flux characteristics: from a soft moderator spectrum with a steep spatial gradient to a very hard spectrum which was relatively flat as a function of position. The conclusions obtained were as follows. Calculated spectral indexes using the Haywood kernel were 2 to 3% higher than experiment, on the average, in the fuel region of these cells. Use of the Koppel kernel removed most of this disagreement in the case of ¹⁷⁶Lu but the comparison for ²³⁹Pu remained unchanged. On the basis of these results, the thermal-flux spectrum obtained with the Haywood model appears to be slightly too hard. With regard to the disadvantage factors, good agreement was generally obtained between theory and measurement except for the tightest lattices. The calculated disadvantage factors were found to be insensitive to the kernel model selected. The effects from including thermal-scattering-pattern treatments above P₁ as well as a spatially dependent and anisotropic source-to-thermal description were found to be small in these cells, < 2%.