11dfnWorks is a parallelized computational suite to generate three-dimensional discrete fracture networks (DFN) and simulate flow and transport. Developed at Los Alamos National Laboratory over the past five years, it has been used to study flow and transport in fractured media at scales ranging from millimeters to kilometers. The networks are created and meshed using dfnGen, which combines fram (the feature rejection algorithm for meshing) methodology to stochastically generate three-dimensional DFNs with the LaGriT meshing toolbox to create a high-quality computational mesh representation. The representation produces a conforming Delaunay triangulation suitable for high performance computing finite volume solvers in an intrinsically parallel fashion. Flow through the network is simulated in dfnFlow, which utilizes the massively parallel subsurface flow and reactive transport finite volume code pflotran. A Lagrangian approach to simulating transport through the DFN is adopted within dfnTrans to determine pathlines and solute transport through the DFN. Example applications of this suite in the areas of nuclear waste repository science, hydraulic fracturing and CO 2 sequestration are also included.

No abstract

We investigate how the choice of injection mode impacts transport properties in kilometer‐scale three‐dimensional discrete fracture networks (DFN). The choice of injection mode, resident and flux‐weighted, is designed to mimic different physical phenomena. It has been hypothesized that solute plumes injected under resident conditions evolve to behave similarly to solutes injected under flux‐weighted conditions. Previously, computational limitations have prohibited the large‐scale simulations required to investigate this hypothesis. We investigate this hypothesis by using a high‐performance DFN suite, dfnWorks, to simulate flow in kilometer‐scale three‐dimensional DFNs based on fractured granite at the Forsmark site in Sweden, and adopt a Lagrangian approach to simulate transport therein. Results show that after traveling through a pre‐equilibrium region, both injection methods exhibit linear scaling of the first moment of travel time and power law scaling of the breakthrough curve with similar exponents, slightly larger than 2. The physical mechanisms behind this evolution appear to be the combination of in‐network channeling of mass into larger fractures, which offer reduced resistance to flow, and in‐fracture channeling, which results from the topology of the DFN.

Although hydraulic fracturing has been used for natural gas production for the past couple of decades, there are significant uncertainties about the underlying mechanisms behind the production curves that are seen in the field. A discrete fracture network‐based reservoir‐scale work flow is used to identify the relative effect of flow of gas in fractures and matrix diffusion on the production curve. With realistic three‐dimensional representations of fracture network geometry and aperture variability, simulated production decline curves qualitatively resemble observed production decline curves. The high initial peak of the production curve is controlled by advective fracture flow of free gas within the network and is sensitive to the fracture aperture variability. Matrix diffusion does not significantly affect the production decline curve in the first few years, but contributes to production after approximately 10 years. These results suggest that the initial flushing of gas‐filled background fractures combined with highly heterogeneous flow paths to the production well are sufficient to explain observed initial production decline. These results also suggest that matrix diffusion may support reduced production over longer time frames.

We characterize how different fracture size‐transmissivity relationships influence flow and transport simulations through sparse three‐dimensional discrete fracture networks. Although it is generally accepted that there is a positive correlation between a fracture's size and its transmissivity/aperture, the functional form of that relationship remains a matter of debate. Relationships that assume perfect correlation, semicorrelation, and noncorrelation between the two have been proposed. To study the impact that adopting one of these relationships has on transport properties, we generate multiple sparse fracture networks composed of circular fractures whose radii follow a truncated power law distribution. The distribution of transmissivities are selected so that the mean transmissivity of the fracture networks are the same and the distributions of aperture and transmissivity in models that include a stochastic term are also the same. We observe that adopting a correlation between a fracture size and its transmissivity leads to earlier breakthrough times and higher effective permeability when compared to networks where no correlation is used. While fracture network geometry plays the principal role in determining where transport occurs within the network, the relationship between size and transmissivity controls the flow speed. These observations indicate DFN modelers should be aware that breakthrough times and effective permeabilities can be strongly influenced by such a relationship in addition to fracture and network statistics.