Historical model evaluation data requirements

Abstract

Several studies about tank waste contents have been published using historical records of tank transactions and various analytical measurements. While these records offer a wealth of information, the results are questionable until error estimates associated with the results can be established. However, they do provide a direction for investigation. Two principal observations from the studies are: (1) Large quantities of individual waste types from the various separations processes were widely distributed throughout the tank farms, and (2) The compositions of many of these waste types are quite distinct from one another. A key assumption associated with these observations is that the effects of time and location on the tank wastes are either nominal or not discernable. Since each waste type has a distinct composition, it would benefit all programs to better quantify that composition, and establish an uncertainty for each element of that composition. Various process, disposal, or other decisions could then be made based on current information reducing the need for extended sampling and analysis.