STATE OF THE ART MONTE CARLO MODELING OF ACTIVE COLLAR MEASURMENTS AND COMPARISON WITH EXPERIMENT

摘要

Safeguarding the commercial nuclear fuel cycle is key to complying with international non-proliferation treaties. For fuel assemblies containing low enriched uranium using active neutron coincidence counting at the fabrication facility provides baseline data. Understanding the instrument response across the entire operational range of fuel configurations is a pre-requisite to obtaining accurate quantitative results on fissile mass. Experimental assessment using mock assemblies is limited because the inventory of fuel, including substitute and poison pins, is limited which together with long measurement times means the number of variations that can be studied is rather small. Also mock assemblies may not accurately reflect the geometry and materials of actual fuel because they are laboratory items designed for ease of use and because assembly designs evolve. It is therefore necessary to be able to simulate the performance faithfully so that cases for which a direct experimental calibration is not available can still be assayed with confidence using calibrations extended by calculation. In this work we describe the state of the art capabilities for calculating correlated neutron counting rates available within the Monte Carlo transport code MCNPX. A detailed model of the BWR UNCL-II active collar at Los Alamos National Laboratory has been created and is currently under review. Responses have been calculated for a 6×6 array of BWR pins enriched to 2.34wt% which can also be realized experimentally. Variations in which some pins are replaced by stainless steel dummies and/or by pins containing gadolinium have also been calculated and measured. In comparing the results we discuss the current limits on the absolute accuracy of present day collar simulations for fresh fuel.