The International Atomic Energy Agency (IAEA) employs well-established statistical methods to assess the effectiveness of its inspection plans on a multi-defect stratum by evaluating defect detection probability (DP). The defect detection probability is the chance of identifying at least one defect when a defected stratum is subjected to a specific inspection plan. When subjected to a specific diversion scenario by a proliferator, defective items of various types and numbers are introduced into the original stratum depending on the diversion scenario. The defected stratum, when subjected to a specific IAEA inspection plan, has a probability of detecting at least one defect DP, giving us the effectiveness of the inspection plan. Conventionally, deterministic models using statistical distributions are employed to compute stratum-level DPs. But these models are specific to the diversion scenario or inspection plan applied and lack universality in their application. Conditional tree algorithm can provide a solution; however, this approach is met with high computational inefficiency when applied to strata with multiple defects and large sampling size. A modified conditional tree approach is developed which can compute stratum-level DPs for strata with multiple types of defects and large sampling size independent of the diversion scenario. This paper will discuss the development of conditional tree-based deterministic model and its present limitations. Results for several mock strata will be presented and the results will be validated against stochastic model results.
Year
2022
Abstract