Year
1986
Abstract
This paper describes a method for deriving sampling plans for nuclear material inventory verification. The method presented is different from the classical approach which envisages two levels of measurement precision corresponding to NDA and DA. In the classical approach the precisions of the two measurement methods are taken as fixed parameters. The new approach is based on multiple levels of measurement precision. The design of the sampling plan consists of choosing the number of measurement levels, the measurement precision to be used at each level and the sample size to be used at each level. This formulation of the problem is particularly suited to modern NDA instruments. For some material types, safeguards inspectors now have available a range of NDA instruments which can give different accuracies and precisions. In the case of some instruments different precisions can be achieved by varying the count time of the measurement. In these kinds of situations the inspector is presented with a dilemma as to which instrument and which precisions should be used. In making such choices the inspector wishes to achieve a desired assurance with a minimum expenditure of time. This paper presents a formulation of the verification problem which lends itself to these kinds of decisions.