Year
2008
Abstract
Current methods for closing material balances in many manufacturing areas where uranium metal turnings, scrap, chips and small metal pieces are produced can be time consuming and incurs high labor costs. A timely closing of the material balance and a substantial cost saving may be achieved by installing a nondestructive assay system that is capable of rapidly quantifying the amount of uranium contained in the material balance unit. A Monte Carlo evaluation was performed to determine the measurement performance of several nondestructive assay (NDA) techniques that can be applied to this specific situation. The results of the Monte Carlo evaluation can be used to decide whether to implement NDA technology and if so, which NDA system to implement. An additional practical application of a measurement technology in manufacturing operations is to supplement the MC&A methodology of Process Monitoring. Several different NDA techniques were modeled with the Monte Carlo N-Particle eXtended (MCNPX) code. The techniques modeled include delayed neutron examination, active coincidence counting, and active multiplicity. All three of these techniques use an interrogation source to induce fissions in the sample. Measurements of the delayed neutrons or induced fission neutrons are used to quantify the uranium mass in an unknown sample. Past measurement experience has shown that these measurement techniques work very well for the assay of typical uranium items. Unfortunately, the measurement of uranium metal turnings is more complicated than the typical case because the metallic constituents are intermixed with lubricant. The lubricant moderates neutrons which causes a change in the neutron signature from the material. The Monte Carlo evaluation consisted of modeling many different possible uranium masses with varying amounts and distributions of lubricant. Techniques for measuring the amount of neutron moderation were built into the Monte Carlo models to correct for changes in the neutron signature from moderation by the lubricant. The Monte Carlo modeling indicated that by combining the neutron signature and information about the amount of moderation the best assay results were obtained. This paper will describe the MCNPX modeling that was done and discuss the results.