Particle Method for Regional-Scale Modeling of Transport in Fractured Rock, 20-R9566Printer Friendly Version
Inclusive Dates: 08/03/05 02/01/07
Background - Low-permeability geological formations are generally considered suitable environments for high-level nuclear waste repositories. Within low-permeability formations, interconnected networks of fractures provide discrete pathways for water movement and associated slow release of radionuclides to the accessible environment. Conventional methods for simulating subsurface transport of dissolved contaminants rely on an effective medium assumption and are not well suited for a scenario with multiple discrete pathways. This project developed an efficient and robust alternative simulation approach that makes direct use of deterministic or stochastic pathway information extracted from discrete fracture network (DFN) simulations to simulate at the repository geosphere scale (approximately 1 to 10 km).
Approach - A multiscale algorithm that uses particles to represent packets of radionuclide mass was developed. The particles are moved along transport pathways according to rules that closely mimic the underlying physical processes of advection, dispersion, and retention processes such as matrix diffusion combined with sorption. Decay and in-growth of decay-product radionuclides are modeled as random events. An adaptive kernel method is used to reconstruct mass discharge from the particle histories. The transport pathways may be extracted directly from DFN simulations or based on regional-scale flow models that use upscaled effective flow properties. When using upscaled flow models, stochastic simulation based on the results of mesoscale (approximately 100 m) DFN simulation may be used to generate subgrid velocity variations along the pathways.
Accomplishments - Tests of the algorithm demonstrated that it is accurate, robust, and computationally efficient. Analyses illuminating the results of the project led directly to a new externally funded project. A new computer code (TDRW) implementing the particle algorithm was developed, verified, submitted for copyright registration, and licensed to two foreign repository programs and one federal regulatory agency. Results of the project were presented at three international conferences. Two journal articles were submitted and accepted for publication.