Determining the Uncertainty Associated with Retrospective Air Sampling for Optimization Purposes
D. J. Hadlock
NUREG 1400 contains an acceptable methodology for determining the uncertainty associated with retrospective air sampling. The method is a fairly simple one in which both the systemic and random uncertainties, usually expressed as a percentage error, are propagated using the square root of the sum of the squares. Historically, many people involved in air sampling have focused on the statistical counting error as the deciding factor of overall uncertainty in retrospective air sampling. This paper looks at not only the counting error but also other errors associated with the performance of retrospective air sampling. By placing the various errors in the same units (e.g., percentage error) it is possible to determine the overall uncertainty for a specific air sample. In the case of this paper, the overall uncertainty when analyzing a "typical" air sample for Pu-239 at the 0.10 and 10 derived air concentration (DAC) levels using a gas proportional counter was evaluated. For the examples in this paper it was found that the counting error dominated the overall uncertainty at the 0.10 DAC, whereas at the 10 DAC level the counting error, although significant, did not dominate. Once this analysis is performed, it is possible to optimize the overall uncertainty associated with the air sample based on the knowledge of which errors influence the overall uncertainty the most. This optimization process allows resources to be directed where they can do the most good. In the examples used in this paper, it was determined that for the 0.10 DAC sample, a simple increase in the daily background count time from 10 to 30 min provided a reduction in overall uncertainty of about 15%.
This abstract was presented at the 37th Annual Midyear Meeting, "Air Monitoring and Internal Dosimetry", Workplace Air Monitoring, Part 1 Session, 2/8/2004 - 2/11/2004, held in Augusta, GA.