Year
2012
Abstract
The major use of passive neutron coincidence and multiplicity counting is to determine plutonium mass in measured items. At the same time this technique is also applicable for measurement of uranium mass when its limitations, predefined assumptions and influencing factors are properly taken into account. The point model which is the base for the passive multiplicity analysis assumes that all emitting neutrons are equivalent, i.e. have the same detection efficiency, fission probability, die-away time etc. This in particular implies homogeneity of a measured item. However, there are instances where uranium may be present as a mixture of metal lumps and chemical compounds, with low-atomic number (Z) elements mixed with matrix material. In these cases, such as those involving scrap and waste items, the heterogeneity significantly reduces the accuracy of this technique in estimating uranium mass. Since the penetration of neutrons with a fission energy spectrum is higher versus moderated neutrons and matrix effects, in reverse, are lower, we propose a technique which uses spontaneous fission neutrons from 238U and takes into account multiplication effects and an (a,n)-source term in the measured item. This paper describes a passive neutron multiplicity method, modified for the measurement of heterogeneous uranium scrap and waste items containing a mixture of uranium metal chunks and chemical compounds of uranium with low-Z material. It also presents an analysis and estimated values of a possible systematic bias of the method caused by different factors. Taking into account the low spontaneous fission rate and high Z of U material, the cosmic ray (CR) spallation neutron influence becomes significant when measuring neutron multiplicities. This makes the widely used “cycle rejection algorithm” based on an assumption of normal distribution of respective count rates inaccurate. The paper examines the changes needed in the criterion and algorithm for the individual acquisition cycle rejection in order to calculate more precise multiplicity net counting rates.