PNNL

Abstract

In the next five years, critical decisions on the future disposition of wastes on the Hanford Site will be made: * barriers to control recharge at the ground surface,* procedures for retrieval and stabilization of tank waste, and* remediation of contaminated sediments.These decisions will be based, in part, on model predictions of contaminant transport in the vadose zone. Our investigation focuses on high-level radioactive waste tanks in the SX Tank Farm in the 200 West Area of the Hanford Site. The historical SX tank wastes were the hottest, highest pH, highest ionic strength, highest aluminum wastes in Hanford single-shell tanks (SST); 10 of these tanks are confirmed or suspected leakers. Over the last two years, an integrated program of scientific and engineering study has been directed at the SX tank farm, including (1) laboratory experiments on waste-sediment interactions, (2) field experiments on the migration of dense, hypersaline solutions, (3) estimates of historical tank leak source-terms, (4) characterization of hydrostratigraphic units, and (5) physical and chemical analyses of soil samples from the SX tank farm. In this presentation, we describe how these disparate data sets have been used to identify detailed process models and parameterizations that are incorporated into simulators of nonisothermal multiphase fluid flow and multicomponent reactive transport. This modeling framework provides a testbed to systematically assess the appropriateness of the identified process representations in the context of site-specific, field-scale properties, and more importantly, observed historical and contemporary behaviors (e.g., hydrology, chemistry). The ultimate goal is to provide a technically defensible basis for the prediction of long-term contaminant behavior. An important technological issue in the comprehensively detailed modeling approach is addressing the computationally intensive calculations that are required.