ABSTRACT

During the operations of nuclear facilities and during the environmental remediation and decommissioning of nuclear facilities, radioactive waste is generated which must be assayed. To save on labor and transportation costs, this material is commonly placed in large containers of typically several cubic meters in size [e.g. B-25]. The most common choices of assay are to either extract a representative sample of the contents for laboratory assay, or to use in-situ gamma spectroscopy of the total container. Both of these methods have strengths and weaknesses. In Situ methods determine container activity directly and typically quantify gamma emitters which are then correlated to total activity. Gamma measurement accuracy can suffer from an inadequate efficiency calibration, which can be caused by variations in container size, shape, matrix fill height, or matrix density, but primarily by non-homogeneous distribution of the radioactivity. Laboratory assay methods using a small sample extracted from the container are generally quite accurate for the sample analyzed; but the determination of total container activity requires the assumption that the sample analyzed in the laboratory is truly representative of the total container, which is also heavily influenced by the non-homogeneous distribution of the radioactivity. This evaluation attempts to determine the best way to estimate the activity within the container and gives quantitative estimates of measurement uncertainty for various conditions of radioactivity contained within the container and for various in-toto and sampling strategies. A new feature of the ISOCS [In-Situ Object Counting System] software called IUE [ISOCS Uncertainty Estimator] was used. First, the various parameters were examined which cause uncertainty in the in-toto measurement to evaluate those which are the major contributors and to asses the measurement uncertainty for a uniformly distributed sample. Next, a series of levels of non-homogeneous distributions were analyzed with a variety of potential detector placement strategies. These variables included number of detectors, placement of the detectors, and movement of the detectors. The uncertainty due to non-homogeneity is reduced when the detectors are moved away from the container, when more detector positions are used, and when the detectors are scanned. When the contents of the container are not homogeneous, the sampling uncertainty is likely to be larger than the in-toto measurement uncertainty. For those same non-homogeneous sample situations, various sampling strategies are tried, including type of sample extraction method, size of sample extracted, and number of samples extracted. The conclusion is that if the contents of the container are not homogeneous, the uncertainty due to the sampling process is likely to be larger than the in-toto measurement uncertainty.

KEYWORDS: Gamma spectroscopy, in-situ, in-toto, uncertainty, box, container,, sampling