Flow channelization is a commonly observed phenomenon in fractured subsurface media where the flow of fluids is restricted primarily to highly transmissive fracture networks surrounded by a ...low‐permeability rock matrix. The multiscale structural heterogeneity of these networks results in multiscale flow channelization where preferential flow paths form at length scales ranging from the entire system down to the subfracture size. We present an analysis of how one of the largest scales in fractured media, the network density, influences the degree of flow channeling that occurs using an ensemble of semigeneric three‐dimensional discrete fracture network (DFN) simulations. We construct 10 DFNs, whose fracture lengths follow a power law distribution, at four densities for a total of 40 networks. We characterize their structure in terms of the network topology and geometry. Eulerian and Lagrangian observations of the steady‐state flow fields obtained within the networks are used to quantify the degree of flow channelization at the network scale. We introduce a measure for the importance‐ranking/hierarchy of different flow paths in the network using graph‐based analysis of Lagrangian transport by which the degree of flow channeling between networks is compared. These flow observations are then linked to the structural properties of the networks. In general, network‐scale flow channeling decreases as the network density increases. However, at low densities, there is more uniform flow within the entire connected network than in high‐density networks. We also demonstrate how standard transport observables can be used to infer the degree of flow channelization occurring within a fracture network.
Key Points
Flow and transport through an ensemble of networks with varying densities are simulated
Novel measurements of flow channeling degree are introduced
As the network connectivity increases, the degree of network‐scale flow channeling decreases
•Shale gas has revolutionized US energy including prices, consumption, & emissions.•Key questions remain including environmental concerns & extraction efficiencies.•New discoveries are identified ...through data mining & analysis of 20,000 wells.•Findings include learning-by-doing, refracturing, & long-term tail production.•We hypothesize that manipulating tail production could re-revolutionize shale gas.
Shale gas and hydraulic refracturing has revolutionized the US energy sector in terms of prices, consumption, and CO2 emissions. However, key questions remain including environmental concerns and extraction efficiencies that are leveling off. For the first time, we identify key discoveries, lessons learned, and recommendations from this shale gas revolution through extensive data mining and analysis of 23years of production from 20,000 wells. Discoveries include identification of a learning-by-doing process where disruptive technology innovation led to a doubling in shale gas extraction, how refracturing with emerging technologies can transform existing wells, and how overall shale gas production is actually dominated by long-term tail production rather than the high-profile initial exponentially-declining production in the first 12months. We hypothesize that tail production can be manipulated, through better fracturing techniques and alternative working fluids such as CO2, to increase shale gas recovery and minimize environmental impacts such as through carbon sequestration.
•Hydraulic fracturing has increased shale gas production and lowered energy costs.•Water-based drawbacks: poor production, environmental impacts, water shortages.•Supercritical CO2 could enhance ...production while minimizing environmental concerns.•Through theory, modeling, & experiments, we explore CO2 opportunities & challenges.•CO2 has substantial potential to transform shale gas; further research is needed.
Hydraulic fracturing of shale formations in the United States has led to a domestic energy boom. Currently, water is the only fracturing fluid regularly used in commercial shale oil and gas production. Industry and researchers are interested in non-aqueous working fluids due to their potential to increase production, reduce water requirements, and to minimize environmental impacts. Using a combination of new experimental and modeling data at multiple scales, we analyze the benefits and drawbacks of using CO2 as a working fluid for shale gas production. We theorize and outline potential advantages of CO2 including enhanced fracturing and fracture propagation, reduction of flow-blocking mechanisms, increased desorption of methane adsorbed in organic-rich parts of the shale, and a reduction or elimination of the deep re-injection of flow-back water that has been linked to induced seismicity and other environmental concerns. We also examine likely disadvantages including costs and safety issues associated with handling large volumes of supercritical CO2. The advantages could have a significant impact over time leading to substantially increased gas production. In addition, if CO2 proves to be an effective fracturing fluid, then shale gas formations could become a major utilization option for carbon sequestration.
We study the large-scale dynamics and prediction of hydrodynamic transport in random fracture networks. The flow and transport behaviour is characterized by first passage times and displacement ...statistics, which show heavy tails and anomalous dispersion with a strong dependence on the injection condition. The origin of these behaviours is investigated in terms of Lagrangian velocities sampled equidistantly along particle trajectories, unlike classical sampling strategies at a constant rate. The velocity series are analysed by their copula density, the joint distribution of the velocity unit scores, which reveals a simple, albeit hidden, correlation structure that can be described by a Gaussian copula. Based on this insight, we derive a Langevin equation for the evolution of equidistant particle speeds. In this framework, particle motion is quantified by a stochastic time-domain random walk, the joint density of particle position, and speed satisfies a Klein–Kramers equation. The upscaled theory quantifies particle motion in terms of the characteristic fracture length scale and the distribution of Eulerian flow velocities. That is, it is predictive in the sense that it does not require the a priori knowledge of transport attributes. The upscaled model captures non-Fickian transport features, and their dependence on the injection conditions in terms of the velocity point statistics and average fracture length. It shows that the first passage times and displacement moments are dominated by extremes occurring at the first step. The presented approach integrates the interaction of flow and structure into a predictive model for large-scale transport in random fracture networks.
dfnWorks 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 CO2 sequestration are also included.
•We provide an overview of the HPC DFN flow and transport suite dfnWorks.•Dfnworks combine FRAM, LaGriT, PFLOTRAN, and Lagrangian transport simulations.•Three applications are presented to demonstrate the utility of dfnWorks.
We study first passage behaviors in the flow through three-dimensional random fracture networks. Network and flow heterogeneity lead to the emergence of heavy-tailed first passage time distributions ...that evolve with increasing distance between the start and target planes, and transition toward stable laws. Analysis of the spatial memory of the first passage process shows that particle motion can be quantified stochastically by a time domain random walk conditioned on the initial velocity data. This approach identifies advective tortuosity, the velocity point distribution and the average fracture link length as key quantities for the prediction of first passage times. Using this approach, we develop a theory for the evolution of first passage times with increasing distance between the start and target planes and the convergence towards stable laws.
We study how the interplay between fracture aperture heterogeneity and tracer injection mode controls fluid flow and tracer transport in three‐dimensional (3D) discrete fracture networks (DFNs). The ...direct 3‐D DFN simulations show that tracer injection mode has substantial effects on tracer spreading across all levels of aperture heterogeneity. The key controlling factor for effective transport is the initial Lagrangian velocity distribution, which is determined by the interplay between injection mode and aperture heterogeneity. The fundamental difference between initial Lagrangian velocity distribution and domain‐scale Eulerian velocity distribution plays a vital role in determining anomalous transport. We effectively capture the observed anomalous transport using an upscaled transport model that incorporates initial velocity distribution, stationary velocity distribution, velocity correlation length, and average advective tortuosity. With the upscaled transport model, we accurately capture the evolution of Lagrangian velocity distribution and predict longitudinal spreading in 3‐D DFN.
Key Points
Tracer injection mode has strong impact on transport in 3‐D DFNs across all levels of aperture heterogeneity
Initial Lagrangian velocity distribution determines the late‐time power‐law scaling of a breakthrough curve
A Bernoulli CTRW model effectively captures anomalous transport in 3‐D DFNs
•Novel fabrication methods are combined to make a core-scale micromodel for EOR.•Long micromodels capture core-scale physics such as the formation of oil banks.•Increasing viscosity of displacing ...phase in long micromodels increases the amount of oil recovered at breakthrough.
Fluid injection experiments in rocks, commonly referred to as corefloods, are widely used to study and understand fluid flow in the subsurface. However, visual inspection of flow in cores requires computed tomography machines which may not be widely accessible. We introduce a novel micromodel that is as long as a typical core (40 cm), has adjustable pore structure, and includes 2.5D pore throats that can be used to conduct fluid displacements analogous to those in cores. Flow can be visualized inexpensively in the micromodel with an optical microscope. We performed standard coreflood tests in our micromodel including a tracer test and a steady state permeability test. We also conducted multiphase displacements by injecting aqueous solutions at varying glycerol concentrations to displace oil from the micromodel and observed the effect of the viscosity ratio on macro-scale recovery efficiency. When the injected aqueous solution was less viscous than the resident oil, it fingered through the oil. Fingering was not observed in the cases where the injected glycerol solution was more viscous than the oil. Moreover, as the viscosity of the injected glycerol solution increased, oil was recovered more rapidly. Additionally, we performed surfactant and glycerol floods in short (2.4 cm) and long (40 cm) micromodels that show long chips capture scale dependent physics, such as oil banking, that small chips do not capture. The novel micromodel shows promise as a screening tool for chemical EOR because it captures phase banks that are desirable in corefloods.
We develop a theoretical model for power law tailing behavior of transport in fractured rock based on the relative dominance of the decay rate of the advective travel time distribution, modeled using ...a Pareto distribution (with tail decaying as ∼ time−(1+α)), versus matrix diffusion, modeled using a Lévy distribution. The theory predicts that when the advective travel time distribution decays sufficiently slowly (α<1), the late‐time decay rate of the breakthrough curve is −(1+α/2) rather than the classical −3/2. However, if α>1, the −3/2 decay rate is recovered. For weak matrix diffusion or short advective first breakthrough times, we identify an early‐time regime where the breakthrough curve follows the Pareto distribution, before transitioning to the late‐time decay rate. The theoretical predictions are validated against particle tracking simulations in the three‐dimensional discrete fracture network simulator dfnWorks, where matrix diffusion is incorporated using a time domain random walk.
Key Points
A theoretical model is developed for power law tailing of breakthrough curves influenced by matrix diffusion and heterogeneous advection
A threshold decay rate for the advective travel time distribution is identified,below which matrix diffusion produces tail decay rates >−3/2
Matrix diffusion is implemented in high‐fidelity three‐dimensioal discrete fracture network simulations to confirm theoretical predictions
Fractured reservoirs are complex and multi‐scale systems composed of matrix and fractures. Accordingly, modeling flow in such geological media has been a great challenge. In this study, we ...investigated the effect of scale as well as matrix and fracture network characteristics on the effective permeability (keff) in matrix‐fracture systems under fully saturated conditions. We generated fracture networks, embedded within a matrix of permeability of 10−18 m2, with fracture lengths followed a truncated power‐law distribution with exponent α = 1.5, 2.0, and 2.5. We set fracture permeability equal to 10−16, 10−14, and 10−12 m2 and numerically simulated fluid flow to determine the keff at six fracture densities for 36 fractured reservoirs. Results showed that the effect of α and scale on the keff became more significant as the contrast between matrix and fracture permeabilities increased. We also fit the percolation‐based effective‐medium approximation (P‐EMA) to the simulations and optimized its two parameters critical fracture density and scaling exponent. Results exhibited that both P‐EMA model parameters were scale‐dependent. Through linear regression analysis, we found that the critical fracture density and scaling exponent were highly correlated to other matrix‐fracture system properties and proposed two regression‐based models evaluated using a new six sets of simulations. Comparing the estimated keff values with the simulated ones demonstrated the reliability and predictability of the P‐EMA. The matrix‐fracture systems studied here were finite in size. We also showed that one may extend results to infinitely large reservoirs using the P‐EMA framework.
Key Points
Scale‐dependent effective permeability in matrix‐fracture systems is investigated using numerical simulations and theoretical modeling
Critical fracture density and scaling exponent are linked to reservoir properties and its size, and results are extended to large systems
Theoretical estimations of effective permeability in six unseen fractured reservoirs agree with numerical simulations well