Methods for estimation of covariance matrices and covariance components for the Hanford Waste Vitrification Plant Process
The high-level waste (HLW) vitrification plant at the Hanford Site was being designed to transuranic and high-level radioactive waste in borosilicate class. Each batch of plant feed material must meet certain requirements related to plant performance, and the resulting class must meet requirements imposed by the Waste Acceptance Product Specifications. Properties of a process batch and the resultlng glass are largely determined by the composition of the feed material. Empirical models are being developed to estimate some property values from data on feed composition. Methods for checking and documenting compliance with feed and glass requirements must account for various types of uncertainties. This document focuses on the estimation. manipulation, and consequences of composition uncertainty, i.e., the uncertainty inherent in estimates of feed or glass composition. Three components of composition uncertainty will play a role in estimating and checking feed and glass properties: batch-to-batch variability, within-batch uncertainty, and analytical uncertainty. In this document, composition uncertainty and its components are treated in terms of variances and variance components or univariate situations, covariance matrices and covariance components for multivariate situations. The importance of variance and covariance components stems from their crucial role in properly estimating uncertainty In values calculated from a set of observations on a process batch. Two general types of methods for estimating uncertainty are discussed: (1) methods based on data, and (2) methods based on knowledge, assumptions, and opinions about the vitrification process. Data-based methods for estimating variances and covariance matrices are well known. Several types of data-based methods exist for estimation of variance components; those based on the statistical method analysis of variance are discussed, as are the strengths and weaknesses of this approach.
- Research Organization:
- Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- AC06-76RL01830
- OSTI ID:
- 215713
- Report Number(s):
- PNNL-11000; PHTD-C93-05-01P; ON: DE96008239
- Resource Relation:
- Other Information: PBD: Mar 1996
- Country of Publication:
- United States
- Language:
- English
Similar Records
Preliminary Feed Test Algorithm for the Hanford Waste Vitrification Plant product composition control system
Preliminary IHLW Formulation Algorithm Description