Computation of Flow and Transport in Fracture Networks on a Continuum Grid

Fractures in geologic media are known to provide preferential pathways for fluid flow and transport of contaminants. Computations of flow and transport in fracture networks are usually performed using either a discrete fracture network (DFN) modeling approach or a continuum based approach. The DFN approach explicitly model individual elements of a network and thus is considered to provide accurate estimates of flow and transport. The limitations of DFN method arise from computational constraints in both processing speed and memory, and difficulties in incorporating fracture-matrix interactions. Continuum based approaches assign equivalent hydraulic parameters, derived from statistical properties of the network, to cells of a continuum mesh. Such methods allow for simulation of processes not presently achievable using DFN methods, but can suffer from lack of accuracy especially in predictions of transport behavior of the network. Applying a continuum based approach becomes more challenging for sparse fracture networks where preserving the connectivity of elements and anisotropic behavior of the network is of paramount importance. We present a set of mapping rules and upscaling techniques to develop an improved fracture continuum methodology. The method requires an anisotropic conductivity field for the fracture continuum grid as opposed to assigning a scalar value for each cell. Comparisons with DFN simulations demonstrate the accuracy of the improved fracture continuum method in simulating both fluid flow and transport characteristics, including earlyand late-time tails, for a wide range of fracture density values and grid cell sizes.