“Safeguards”—Étude d'un modèle stochastique gaussien pour la détection d'un détournement de matière fissile

Abstract

Stochastic model is described which enables us to assess quantitatively the efficiency of safeguards operations in reprocessing or fabrication plants. The a priori assumption, or ‘zero-hypothesis’, is that there has been no diversion of fissile materials, the inspector's task being to invalidate it. This zero-centered model seems more adequate than a deviation centered model as suggested in [1]. The latter is shown to lead to unacceptable statements, whereas the former makes improper charges of diversion impossible. • To detect diversion adequately the inspector will resort to three criteria. • The first one sets an upper bound M for the total mass uncertainty. When the latter reaches M the inspector will take a complete inventory. • The second criterion enables him to decide whether or not an estimated mass balance is compatible with the agreed model. • The third criterion connects the mass uncertainty to the time it lasts; moreover, it settles the number of strategic points within the plant. As an application of the mathematical model developed, systematic cheating strategies are also studied. • First strategy: diversions only. If the thief intends to divert in m operations the maximum possible quantity of fissile material without violating the safeguards criteria, he should remove m - k times the maximum allowable amount y ̂ max and k times some smaller quantity y ̂ 0 which depends upon m. It is shown that the total amount of diverted material cannot exceed 6·31 y ̂ max . • Second strategy: diversions and restitutions. By removing m - k times y ̂ max and by restituting k times some quantity y ̂ 0 depending again on m, an arbitrarily great mass could be diverted, at least theoretically. This renders it necessary for the inspector to assess an upper limit to the positive mass balance y ̂ 0 . Moreover, if the thief feels it necessary to limit his global risk of being caught to less than say ten per cent, the amount he can actually divert will be drastically reduced.