A TECHNIQUE FOR COMBINING NEUTRON AND GAMMA-RAY DATA INTO A SINGLE ASSAY VALUE*

Year
1998
Author(s)
Mark M. Pickrell - Los Alamos National Laboratory
David J. Mercer - Los Alamos National Laboratory
T.J. Sharpe - Los Alamos National Laboratory
Abstract
We have explored the potentials of using both neutron and gamma-ray measurements on a single item and combining these data into a single assay value. The purpose was to improve assay capability for sample matrices that are difficult to measure. We chose an empirical approach because we wanted to address difficult-to-measure items for which the assay problem is complex. We used the tomographic gamma scanner; a passive, high-efficiency neutron counter with add-a-source and multiplicity; and an active neutron, californium shuffler to obtain measurements. Twenty-four 200-L drums were measured with various matrices using all three machines. The matrices were chosen specifically to span the difficult-tomeasure assay problems for some or all of the instruments. For example, we measured a drum filled with concrete and another filled with metal. The data from these measurements were analyzed using the alternating conditional expectation algorithm, which is one of a class of generalized additive models.[1-3] Other data fusion algorithms are also possible and are being explored. The intent was to find ways to combine the data that would reduce the matrix-induced measurement error.