DlSCLAlM ERThis report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor Battelle Memorial Institute, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or.usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise doss not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof, or Battelle Memorial Institute.
SUMMARYThe high-level waste (HLW) vitrification plant at the Hanford Site was being designed .to immobilize transuranic and high-level radioactive waste in borosilicate glass. Each batch of plant feed material must meet certain requirements related to plant performance, and the resu!ting glass must meet requirements imposed by the Waste Acceptance Product Specifications (WAF' S; DOE, 1993). Properties of a process batch and the resulting glass are largely determined by the composition of the feed material. Empirical models are being developed to estimate some properly values from data on feed composition. ...
111Estimating uncertainty from process knowledge may be necessary when data are scarce or lacking entirely. A Monte Carlo procedure for this type of uncertainty estimation is illustrated for a hypothetical analytical process; a hypothetical analytical covariance matrix is developed for use in later studies. Measures of strength of belief in simulated uncertainty estimates are required to update or combine these estimates with incoming .information; an approach to assigning such measures is developed.Proper estimation of the uncertainties required for testing processlproduct specifications requires the combination of various components of uncertainty. Satterthwaite's method for combining components of uncertainty and for ass~ssing the precision of the overall uncertainty estimate is recommended.Several batch and glass properties wiU be estimated from empirical models based on feed composition. Overall uncertainty in estimated property values derives both from uncertainty in estimated feed composition and from uncertainty in model coefficients. *A general method for collapsing multivariate composition and model uncertainties into univariate uncertainties for property values is described. Applying this method with the hypothetical analytical covariance matrix indicates that analytical uncertainty can be expected to produce a relative standard deviation between 3.5% and 8% in estimated property values.An dtemative method for assessing the contribution of composition uncertainty to overall uncertainty is discussed.-A method for calculating the sample size required for estimation of univariate uncertainty (with specified prec...