Correlation of nuclear criticality safety computer codes with plutonium benchmark experiments and derivation of subcritical limits. [MGBS, TGAN, KEFF, HRXN, GLASS, ANISN, SPBL, and KENO

Abstract

A compilation of benchmark critical experiments was made for essentially one-dimensional systems containing plutonium. The systems consist of spheres, series of experiments with cylinders and cuboids that permit extrapolation to infinite cylinders and slabs, and large cylinders for which separability of the neutron flux into a product of spatial components is a good approximation. Data from the experiments were placed in a form readily usable as computer code input. Aqueous solutions of Pu(NO/sub 3/)/sub 4/ are treated as solutions of PuO/sub 2/ in nitric acid. The apparent molal volume of PuO/sub 2/ as a function of plutonium concentration was derived from analyses of solution density data and was incorporated in the Savannah River Laboratory computer codes along with density tables for nitric acid. The biases of three methods of calculation were established by correlation with the benchmark experiments. The oldest method involves two-group diffusion theory and has been used extensively at the Savannah River Laboratory. The other two involve S/sub n/ transport theory with, in one method, Hansen-Roach cross sections and, in the other, cross sections derived from ENDF/B-IV. Subcritical limits were calculated by all three methods. Significant differences were found among the results and between the results and limits currently in the American National Standard for Nuclear Criticality Safety in Operations with Fissionable Materials Outside Reactor (ANSI N16.1), which were calculated by yet another method, despite the normalization of all four methods to the same experimental data. The differences were studied, and a set of subcritical limits was proposed to supplement and in some cases to replace those in the ANSI Standard, which is currently being reviewed.