Comparative Calculational Evaluation of Array Criticality Models

Abstract

The NB/sup 2//sub N/, albedo, interaction potential, weighted interaction Oak Ridge Gaseous Diffusion Plant solid-angle, s/v surface density, fraction critical surface density, and equilateral hyperbola methods are compared by attempting to apply them to 14 different example arrays. The example arrays are distinguished by the type of fissile material in the array units and by the number of units in each array. Arrays with units of Pu(95) metal, U(93) metal, damp U(93)O/sub 2/, dry Pu(95)O/sub 2/, U(93)O/sub 2/(NO/sub 3/)/sub 2/ solution, Pu(95)(NO/sub 3/)/sub 4/ solution, or U(5)O/sub 2/F/sub 2/ solution are considered. It is concluded that if the user has access to a Monte Carlo code such as KENO, he is better off using it than any of the other models for criticality evaluations of specific problems. The remainder of the models are most useful in establishing possible designs for a fissile storage or processing facility. The albedo method is found to be the preferred solid-angle technique, while the s/v surface density method and the NB/sup 2//sub N/ method are the most comprehensive and useful of the semiempirical techniques. The s/v method is easier to apply and covers more possible contingencies (e.g., array flooding or persons in arrays), while the more » NB/sup 2//sub N/ method is more difficult to use but has added flexibility, broader applicability, and yields more analytical scaling relations between array parameters. « less