PHYSICAL INVENTORY SAMPLING

Inventory sampling practices and procedures significantly impact the validity of physical inventories of nuclear materials. While statistically valid sampling techniques can provide a defensible and cost effective means for evaluating the status of an inventory, inappropriate sampling applications are often employed resulting in indefensible inventory statements. Part of the problem is that the assumptions and parameters associated with a sampling methodology are not carefully developed, fully documented, and strictly enforced during the conduct of the inventory. Often lacking in a sampling plan is a detailed description of inventory populations and strata, a clear definition of a defective inventory item, and the action to be taken if a defective item is found in the sample. This paper discusses the objectives of inventory sampling and identifies the necessary elements of an inventory sampling plan. It highlights the benefits of carefully planned and executed sampling techniques and discusses the risks and potential pitfalls associated with traditional applications of statistical sampling to physical inventories. The discussion is presented in operational terms using hypothetical and real-world examples. Common-sense approaches that are statistically valid are presented while emphasizing the need for a sufficient sample size in an environment of shrinking human resources in the Department of Energy (DOE).