Year
1998
Abstract
Destructive assay (DA) and Non destructive assay (NDA) are jointly applied to verify declared values of nuclear materials by the Operator. The number of measurements for both methods are calculated mainly based on a zero acceptance number plan to detect possible diversions of 1SQ with specified false alarm probability (y. and non-detection probability fl . In case of DA, however, its another important role is to verify the Operator’s measurement system. The paper describes the quality control technique based on CUSUM statistic to estimate the number of DA samples for the verification of Operator’s measurement system which had so far no prominent statistical background. When this way of calculation (QC based sample size) is introduced together with conventional zero acceptance number plan calculation method, the number by QC based sample size will exceeds that of conventional method for the facility handling small amount of nuclear materials, and the reverse trend for a large facility. Further CUSUM can give prompt warning to a small potential bias which cannot be easily detected by a 3 o rejection rule. Some numerical examples are presented as well as the idea.