Uncertainty in rate loss corrections based on counting a periodic pulser

Abstract

Tomographic gamma scanning of waste produces three-dimensional transmission and emission images. These are used to derive item-specific attenuation correction factors that improve the accuracy of non-destructive waste assay. For each vertical layer, data grabs of short duration are acquired as the waste item is rotated and translated. The image reconstruction demands accurate rate loss corrections to minimize assay bias. For this application a pulser was used to perform the necessary rate loss corrections. In this work, we summarize the benefits of the pulser approach and review the basic principles on which the method is based. We extend the treatment to include a derivation of the expression for the uncertainty in the net pulser peak area in the presence of an underlying continuum. We report experimental results, taken using a Canberra model WM2900 Tomographic Gamma Scanner, over a broad range of count-rates and peak-to-continuum ratios. Repeat counts under controlled conditions allowed the correction factor and its variance to be determined and compared against expectations. These results confirm the validity of the correction factor formula and the corresponding expression for its uncertainty. The rate loss analysis has been built into a Monte Carlo Replicate engine to allow the uncertainty to be propagated into the total measurement uncertainty of the final assay.