Hybrid Modeling Approach for Transport in Fractured Rock, 20-9252

Printer Friendly Version

Principal Investigators
Scott L. Painter
Roland Benke

Inclusion Dates: 04/01-01 - 09/31/02

Background - Many applications in subsurface hydrology involve the transport of solutes through fractured but otherwise low permeability rock. This transport mechanism is important in remediation of fractured formations and in understanding the safety of proposed underground repositories for radioactive or chemical waste. The conventional modeling approach is to use the classical advection dispersion equation to calculate the time and space evolution of contaminant concentrations. However, field investigations and theoretical studies have demonstrated that solutes may be transported in a nondiffusive manner over considerable distances. This observation has important implications for predictive calculations and there is a strong need for alternative modeling approaches.

Approach - This project developed new approaches for extrapolating the results of discrete fracture network (DFN) simulations to the spatial scales of field applications. DFN simulation is an alternative to continuum models that involves stochastically generating networks of discrete fractures and then solving for flow and transport within the synthetic network. The DFN approach does not suffer the same limitations as the advection-dispersion equation, but it is too demanding of computational resources to be of direct practical value in many applications. The approach of this project was to extract statistics on particle velocities from DFN simulations of modest size. These statistics capture relevant information about the transport properties of the network and are then used to construct new and more computationally efficient models for contaminant transport.

Accomplishments - Two- and three-dimensional DFN simulators were developed, tested, and used to generate velocity statistics for a range of fracture network parameters. Results of these simulations were described in a journal article that is accepted for publication in Geophysical Research Letters. Two specific approaches were developed to upscale these results to field-scale applications: an approach based on continuous time random walks, and a continuum model based on the linear Boltzmann transport equation from the kinetic theory of gases. In both approaches, velocity statistics from the DFN simulations are used to calculate model parameters, and the simplified model is then used to extrapolate in a computationally efficient way to model domains of different sizes. Both approaches were successfully compared to global breakthrough curves calculated directly from DFN simulations. The new models were also used to simulate contaminant movement at the spatial scale of field applications, thereby demonstrating the new capabilities to potential clients. These results were summarized in two additional manuscripts for technical journals. These new modeling approaches were successfully used in project work for a European high-level nuclear waste repository program.

2002 Program SwRI Home