Distilling Complex Model Results into Simple Models for use in Assessing Compliance with Performance Standards for Low Level Waste Disposal Facilities
Assessing the long term performance of waste disposal facility requires numerical simulation of saturated and unsaturated groundwater flow and contaminant transport. Complex numerical models have been developed to try to realistically simulate subsurface flow and transport processes. These models provide important information about system behavior and identify important processes, but may not be practical for demonstrating compliance with performance standards because of excessively long computer simulation times and input requirements. Two approaches to distilling the behavior of a complex model into simpler formulations that are practical for demonstrating compliance with performance objectives are examined in this paper. The first approach uses the information obtained from the complex model to develop a simple model that mimics the complex model behavior for stated performance objectives. The simple model may need to include essential processes that are important to assessing performance, such as time-variable infiltration and waste emplacement rates, subsurface heterogeneity, sorption, decay, and radioactive ingrowth. The approach was applied to a Low-Level Waste disposal site at the Idaho National Laboratory where a complex three dimensional vadose zone model was developed first to understand system behavior and important processes. The complex model was distilled down to a relatively simple one-dimensional vadose zone model and three-dimensional aquifer transport model. Comparisons between the simple model and complex model of vadose zone fluxes and groundwater concentrations showed relatively good agreement between the models for both fission and activation products (129I, 36Cl, 99Tc) and actinides (238U, 239Pu, 237Np). Application of the simple model allowed for Monte Carlo uncertainty analysis and simulations of numerous disposal and release scenarios. The second approach investigated was the response surface model. In the response surface model approach, the temporal response of a complex model to an instantaneously-released unit mass of a conservative tracer at a defined point is calculated and stored (the response function). A separate response function is needed for each source-receptor pair. The convolution of the response function and estimated contaminant flux from the source then provides an estimate of the media concentration. The response function approach has the advantage of incorporating all processes included in the complex model into a single discrete function that requires only one run of the complex model. Limitations to the response surface model include the assumption of uniform retardation in the model domain and the assumption that radioactive progeny travel at the same rate as their parent. In highly heterogeneous environments where uniform retardation cannot be assumed, contaminant-specific response functions can be calculated. The response surface model was applied to a two-dimensional regional aquifer model to simulate multiple plumes of tritium originating from multiple sources across the Idaho National Laboratory.
- Research Organization:
- Idaho National Lab. (INL), Idaho Falls, ID (United States)
- Sponsoring Organization:
- DOE - EM
- DOE Contract Number:
- DE-AC07-99ID-13727
- OSTI ID:
- 911936
- Report Number(s):
- INL/CON-06-11671; TRN: US0800226
- Resource Relation:
- Conference: Waste Management 07,Tucson, Arizona,02/25/2007,03/01/2007
- Country of Publication:
- United States
- Language:
- English
Similar Records
Technical Basis and Implementation of Groundwater Pathway Models Used for INL’s RH LLW Disposal Facility Performance Assessment
INEEL Subregional Conceptual Model Report Volume 2: Summary of Existing Knowledge of Geochemical Influences on the Fate and Transport of Contaminants in the Subsurface at the INEEL