Achieving Higher Accuracy in the Gamma-Ray Spectroscopic Assay of Holdup

Good nuclear material accountability practices include unbiased measurements with minimum uncertainties. This paper describes new procedures for use with traditional holdup analysis methods based on gamma-ray spectroscopy. The procedures address two important sources of bias inherent in gamma-ray measurements of holdup. Holdup measurements are performed with collimated, shielded gamma-ray detectors. The measurement distance is chosen to simplify or generalize the deposit geometry to that of a point, line, or area in order to accomplish the measurement and analysis of thousands of deposit locations during relatively short, routine inventory periods. The quantitative holdup result is based on the net count rate of a representative gamma ray. Room background and equipment attenuation are addressed explicitly in the generalized geometry holdup (GGH) approach. Until now, the GGH approach has not addressed either the finite width of the point or line deposit compared to the measurement distance or the effects of gamma-ray self-attenuation. In both cases the effect is a systematic negative bias, explaining the systematically low results of gamma-ray holdup measurements. The new, generalized analysis corrects for bias that arises from both the finite-source and self-attenuation effects, relying on an additional empirical parameter. Used at three Department of Energy facilities and elsewhere, the new analysis operates within the GGH formalism to achieve two important goals: 1. Self-consistent reduction of the error incurred for an individual deposit. 2. Randomization of the propagated errors (to eliminate bias in total holdup). The algorithms are simple, general, and readily automated for use plant-wide. This paper presents the new correction procedures and algorithms and analyzes the sensitivity of the holdup result to the uncertainty in the empirical parameter when one or both corrections are made. It uses specific examples of the magnitudes of finite-source and self-attenuation corrections to actual field measurements and discusses automation of the correction procedures.