Material Accountability and Its Verification: A Special Example of Multivariate Statistical Inference

Avenhaus, R. & Nakicenovic, N. ORCID: https://orcid.org/0000-0001-7176-4604 (1975). Material Accountability and Its Verification: A Special Example of Multivariate Statistical Inference. IIASA Research Report. IIASA, Laxenburg, Austria: RR-75-025

Preview

Text RR-75-025.pdf Download (1MB) | Preview

Abstract

The IAEA nuclear material safeguards system consists basically of two different parts. One is the data verification scheme: the operators of nuclear plants report all relevant data on nuclear material processed in the plant to the safeguards authority. These data are then verified by the safeguards authority with independent measurements. The other part is the material accountability scheme: in case there are no significant differences between the operator's and the safeguards authority's data, all of the operator's data are taken for the nuclear material balance establishment.

The purpose of this paper is to evaluate the overall probability of detection of this system in case someone tries to divert material. This evaluation takes into account the different diversion strategies available. It is complicated because the two decision functions on which the evaluation is based -- the difference between operator's and inspector's data and the difference between book and physical inventory -- are stochastically dependent. Exact formulas are derived and applied to a realistic case; it is shown that with a good approximation, one may neglect the correlation and thus, use simplified formulas.

Actions (login required)

View Item