William L. Kuhn

Development of a leach model for a commercial nuclear waste glass

A leach model is presented for a commonly studied commercial nuclear waste glass, PNL 76-68. Boron release is taken to be a monitor of the reaction rate of the glass, while the actual releases of many other glass constituents into solution during static tests are evidently controlled by solubilities. The reaction rate determined in this way passes from linear to parabolic kinetics over the duration of the experiments analyzed, and boron concentrations in solution are found to be a function of the product of time and surface areato-solution volume ratio. This behavior is found to be explained well by assuming the reaction is impeded by resorption of reaction products onto the reacting surface. Two model parameters are found as functions of temperature by fitting the model to published data. It is concluded that the accumulation of silica near the glass surface in a waste package in a repository could limit the rate of reaction of the glass, but not that the reaction would cease as silica reaches its solubility limit in solution.

Dynamics of Crust Dissolution and Gas Release in Tank 241-SY-101

Due primarily to an increase in floating crust thickness, the waste level in Tank 241-SY-101 has grown appreciably and the flammable gas volume stored in the crust has become a potential hazard. To remediate gas retention in the crust and the potential for buoyant displacement gas releases from the nonconvective layer at the bottom of the tank, SY-101 will be diluted to dissolve a large fraction of the solids that allow the waste to retain gas. The plan is to transfer some waste out and back-dilute with water in several steps. In this work, mechanisms and rates of waste solids dissolution and gas releases are evaluated theoretically and experimentally. Particular emphasis is given to crust dissolution processes and associated gas releases, although dissolution and gas release from the mixed-slurry and nonconvective layers are also considered. The release of hydrogen gas to the tank domespace is modeled for a number of scenarios. Under the tank conditions expected at the time of back-dilution, no plausible continuous or sudden gas release scenarios resulting in flammable hydrogen concentrations were identified.

Leach Models for a Commercial Nuclear Waste Glass

A review of the leaching behavior of 76-68 glass shows that it cannot be explained in terms of diffusion in the glass, which has been the basis for several leach models. Instead, we present two models based on a dissolution rate impeded by surface processes: the accumulation of a protective layer of insoluble reaction products, and adsorption of reaction products on the surface. The resulting predicted time dependences are identical and predict a change from linear to parabolic rate laws for soluble species, which is found to agree with the data over a range of temperatures. Incongruent release is attributed primarily to solubility effects. The relative merits of the models are discussed on the basis of the effect of surface area-to-volume ratio in static leach tests. Their relevance to modeling repository behaivor is discussed.

Technical Basis of Scaling Relationships for the Pretreatment Engineering Platform

Pacific Northwest National Laboratory has been tasked by Bechtel National Inc. (BNI) on the River Protection Project-Waste Treatment Plant (RPP-WTP) project to perform research and development activities. The Pretreatment Engineering Platform (PEP) is being designed and constructed as part of a plan to respond to an issue raised by the WTP External Flowsheet Review Team (EFRT) entitled “Undemonstrated Leaching Processes” and numbered M12. The PEP replicates the WTP leaching process using prototypic equipment and control strategies. The approach for scaling PEP performance data to predict WTP performance is critical to the successful resolution of the EFRT issue. This report describes the recommended PEP scaling approach, PEP data interpretation and provides recommendations on test conduct and data requirements.

Predicting amounts of radiolytically produced species in brine solutions

The corrosion of iron overpack materials for nuclear waste forms in salt is highly dependent upon the species produced by the radiolysis of brines which may contact the metal. A computer code has been developed to predict the concentrations of the radiolytically produced species as a function of time using published values for radiation yields of species, chemical reaction rate constants, and parameters derived from experimental evidence.

A Statistical Intelligence (STI) Approach to Discovering Spurious Correlation in a Physical Model and Resolving the Problem With an Example of Designing a Pulse Jet Mixing System at Hanford

Pulse jet mixing tests were conducted to support the design of mixing systems for the Hanford Waste Treatment and Immobilization Plant. A physical approach (based on hydro-dynamic behavior) and two semi-empirical (SE) approaches were applied to the data to develop models for predicting two response variables (critical-suspension velocity and cloud height). Tests were conducted at three geometric scales using multiple noncohesive simulants and levels of possibly influential factors. The physical modeling approach based on hydrodynamic behavior was first attempted, but this approach can yield models with spurious correlation. To overcome this dilemma, two semi-empirical (SE) models were developed by generalizing the form of the physical model using dimensional and/or nondimensional (ND) variables. The results of applying statistical intelligence (STI) tools to resolve the spurious correlation problem via fitting the physical and SE models are presented and compared. Considering goodness-of-fit, prediction performance, spurious correlation, and the need to extrapolate, the SE models based on ND variables are recommended.Copyright

Development of a source term for radionuclide release from spent fuel in a basalt repository

Abstract An analysis of groundwater transport of radionuclides from a nuclear waste repository requires a “source term,” i.e., a prediction of the rate of release of radionuclides from the repository. A relatively simple source term has been constructed to support a demonstration of analytical methodology being developed for use in repository licensing. For this study the postulated case of release from spent fuel in a basalt repository sheared by a fault was considered. The source term model illustrates the use of constraints in describing release behavior, without referring to a specific detailed waste package design. The technical approach used to construct the source term is discussed to illustrate how mass balance and equilibrium constraints can be used in support of repository licensing activities. Assumptions made because of incomplete understanding of leaching behavior are discussed.

Methods for Quantifying the Uncertainties of LSIT Test Parameters, Test Results, and Full-Scale Mixing Performance Using Models Developed from Scaled Test Data

This report discusses the statistical methods for quantifying uncertainties in 1) test responses and other parameters in the Large Scale Integrated Testing (LSIT), and 2) estimates of coefficients and predictions of mixing performance from models that relate test responses to test parameters. Testing at a larger scale has been committed to by Bechtel National, Inc. and the U.S. Department of Energy (DOE) to “address uncertainties and increase confidence in the projected, full-scale mixing performance and operations” in the Waste Treatment and Immobilization Plant (WTP).

Physical Attributes of Pulse Jet Mixer Operation

Vessels mixed using pulse jet mixers that produce a periodic, rather than steady, flow present challenges with respect to modeling slurry mixing. A PJM is a cylindrical tank within the mixed tank that has a conical bottom with an orifice through which process fluid cyclically enters and is expelled forcefully by pressurizing the air space above the liquid in the PJM. Between pulses, some of the solids settle from the slurry, which nominally is a failure in mixing, but during the pulses (if operated to attain bottom clearing conditions), all of the solids are resuspended and made available for processing or transfer. Overall, mixing is successful if the solids are processed and removed from the vessel as needed when averaged over repeated PJM cycles. This paper describes the physics of pulse jet mixing process based on physical observation during experiments and analysis of experimental concentration profile data obtained during the mixing cycle.Copyright

Waste Tank Size Determination for the Hanford River Protection Project Cold Test, Training, and Mockup Facility

The objective of the study was to determine the minimum tank size for the Cold Test Facility process testing of Hanford tank waste. This facility would support retrieval of waste in 75-ft-diameter DSTs with mixer pumps and SSTs with fluidic mixers. The cold test model will use full-scale mixer pumps, transfer pumps, and equipment with simulated waste. The study evaluated the acceptability of data for a range of tank diameters and depths and included identifying how the test data would be extrapolated to predict results for a full-size tank.