Effect of Internal Aperture Variability on Tracer Transport in Large Discrete Fracture Networks (DFN)

Abstract:

Aperture variability within individual fractures is usually neglected in modeling flow and transport through fractured media. Typically, individual fractures are assumed to be homogeneous. However, in reality, individual fractures are heterogeneous, which may affect flow and transport in fractured media. The relative importance of including in-fracture variability in flow and transport modeling has been under debate for a long time. Previous studies have shown flow channeling on an individual fracture with internal variability, where the fracture is considered isolated from the rest of the fracture network. Although these studies yield some clear insights into the process, the boundary conditions are impractical for field-scale networks, where the realistic boundary conditions are determined by fracture connections in the network. Therefore, flow in a single fracture is controlled not only by in-fracture variability but also by boundary conditions.

In order to address the question of the importance of in-fracture variability, the internal heterogeneity of every individual fracture is incorporated into a three-dimensional fracture network, represented by a composition of intersecting fractures. The new DFN simulation capability, dfnWorks, is used to generate a kilometer scale DFNs similar to the Forsmark, Sweden site. In our DFN model, the in-fracture aperture variability is scattered over each cell of the computational mesh along the fracture, representing by a stationary Gaussian random field with various correlation lengths. The Lagrangian particle tracking is conducted in multiple DFN realizations and the flow-dependent Lagrangian parameters, non-reacting travel time, τ, and cumulative reactivity parameter, β, are obtained along particles streamlines. It is shown that early particle travel times are more sensitive to in-fracture aperture variability than tails of travel time distributions, where no significant effect of the aperture variations and spatial correlation length is observed.