Glass Property Models and Constraints for Estimating the Glass to be Produced at Hanford by Implementing Current Advanced Glass Formulation Efforts

Abstract

Recent glass formulation and melter testing data have suggested that significant increases in waste loading in high-level waste (HLW) and low-activity waste (LAW) glasses are possible over current system planning estimates. The data (although limited in some cases) were evaluated to determine a set of constraints and models that could be used to estimate the maximum loading of specific waste compositions in glass. It is recommended that these models and constraints be used to estimate the likely HLW and LAW glass volumes that would result if the current glass formulation studies are successfully completed. It is recognized that some of the models are preliminary in nature and will change in the coming years. In addition, the models do not currently address the prediction uncertainties that would be required before they could be used in plant operations. The models and constraints are only meant to give an indication of rough glass volumes and are not intended to be used in plant operation or waste form qualification activities. A current research program is in place to develop the data, models, and uncertainty descriptions for that purpose. A fundamental tenet underlying the research reported in this document is the attempt to be less conservative than previous studies when developing constraints for the estimation of glass to be produced by implementing current advanced glass formulation efforts. The less conservative approach documented herein should allow for the estimate of glass masses that may be realized if the current efforts in advanced glass formulations are completed over the coming years, and are as successful as early indications suggest they may be. Because of this approach, there is an unquantifiable uncertainty in the ultimate glass volume projections due to model prediction uncertainties that must be considered, along with other system uncertainties, such as waste compositions and amounts to be immobilized, split factors between LAW and HLW, etc. v Summary Efforts are being made to increase the loading of Hanford tank wastes in glass while maintaining adequate processability, regulatory compliance, and product quality. These efforts have significantly expanded the composition regions and waste loadings of glasses beyond the point used in current project planning models. The effort documented in this report is aimed at evaluating the current glass formulation, property, and processing data, and to use the data to develop a non-conservative set of constraints and property models that can be used to estimate the amount of glass that would be produced at Hanford if the current advanced waste glass formulation efforts were to be successfully completed according to current plans. An accurate method of estimating glass volume to be produced from Hanford tank waste is important for making informed decisions regarding the appropriate process options to pursue, as well as estimating the likely cost and schedule for tank waste cleanup mission completion. To help gain an accurate estimate of glass volume, gl...

Effects of heating rate, quartz particle size, viscosity, and form of glass additives on high‐level waste melter feed volume expansion

“…The conversion occurs in the cold cap, a reacting mass that floats on the surface of molten glass . Together with the waste loading in the glass, the rate of conversion (the rate of melting) is a major factor in the glass production efficiency . It is influenced by the mineralogical form of the glass additives, the particle size of solids, the loading and type of waste, and the melter feed rheology .…”

Reactions during melting of low-activity waste glasses and their effects on the retention of rhenium as a surrogate for technetium-99

“…Recent small-scale melter tests with seven representative Hanford LAW glass feeds (simulated waste plus additives) for the WTP showed that the fraction of technetium retained in glass (referred to as retention) varied from 18% to 66% depending on the feed composition [12][13][14]. Although the volatilized 99 Tc can be captured in the off-gas treatment system and recycled to the melter to increase the retention, the recycle stream also contains other volatile components such as sulfur (S), chlorine (Cl), and fluorine (F) that decrease the loading of waste in the glass [17] and, thus, increase the glass volume. Maximizing the incorporation of 99 Tc into glass is desirable also to minimize the fraction that needs to be treated as a secondary waste and disposed in non-glassy waste forms [18,19].…”

Nonlinear logistic regression mixture experiment modeling for binary data using dimensionally reduced components

“…Further details about the data sets compiled to form the 747-point database are contained in Table 2.1 of Vienna et al 15 The portion of the database for the example in this article is available as supporting information. This example is used in the rest of the article to illustrate fitting mixture experiment models expressed in terms of DRCs using NLR (ie, NLR-DRC models).…”