Layer | Fill | Outline |
---|
Map layers
Theme | Visible | Selectable | Appearance | Zoom Range (now: 0) |
---|
Fill | Stroke |
---|---|
Collaborating Authors
DC Rocks 2001, The 38th U.S. Symposium on Rock Mechanics (USRMS)
Hydraulic fracture containment in layered media, experiment and computer simulation
Chudnovsky, A. (The University of lllinois at Chicago) | Fan, J. (The University of lllinois at Chicago) | Shulkin, Y. (The University of lllinois at Chicago) | Dudley, J.W. (Shell International Exploration & Production, Inc.) | Nichols, W.B. (Shell International Exploration & Production, Inc.) | Wong, G.K. (Shell International Exploration & Production, Inc.)
ABSTRACT: The paper describes a new model and computer simulation technique for hydraulic fracture in layered media, and its application to laboratory experimental data. Field and laboratory observations show that an initially circular crack (or notch) in a layered media can often turn into a highly elongated fracture. Such fracture containment is often assumed, rather than modeled directly. In the method presented here the simulation of the phenomenon is based on a model of elliptical crack propagation in a 3D layered poro-elastic body. The global Griffith-type fracture criterion and an anisotropic fracture toughness parameter (specific fracture energy of rock) are employed. A computer code based on the proposed model is then used to describe a laboratory hydraulic fracture experiment. The large block hydraulic fracture experiment in a layered material (diatomite) is described. The computer simulations with appropriately selected characteristics of the model give results in reasonable agreement with the experimental measurements. Possible applications of the model are discussed. IINTRODUCTION The present work is concerned with the primary features of hydraulic fracture processes in stratified rock materials. As field and laboratory experiments show, an initially small radial crack (or notch) can often grow significantly more in one direction that results in noticeably non-circular (an oval) geometry. The reason for such crack growth behavior may be fracture anisotropy of the layered medium. There are three apparent causes of a non-circular fracture growth: 1) a non-uniform traction on the crack face, 2) an anisotropy of fracture toughness, and 3) heterogeneity of the fracture toughness resulting from layered structure of the rock. This paper describes a new model for hydraulic fracture in layered media, and its application to a laboratory hydraulic fracture test. The model is developed tirst, followed by a description of the experiment. Computer simulation results of the test are then described, and conclusions drawn.
ABSTRACT: An injection experiment was conducted to investigate the pressure domain within which hydromechanical coupling significantly influences the hydrologic behavior of a granite rock mass. The resulting data base is used for testing a numerical model dedicated to the analysis of such hydro-mechanical interactions. These measurements were performed in an open hole section, isolated from shallower zones by a packer set at a depth of 275 m and extending down to 840m. They consisted in a series of flow meter injection tests, at increasing injection rates. Field results showed that conductive fractures form a dynamic and interdependent network, that individual fracture zones could not be adequately modeled as independent systems, that new fluid intakes zones appeared when pore pressure exceeded the minimum principal stress magnitude in that well, and that pore pressures much larger than this minimum stress could be further supported by the circulated fractures. These characteristics raise the question of the influence of the morphology of the natural fracture network in a rock mass under anisotropic stress conditions on the effects of hydromechanical couplings. INTRODUCTION Characterizing and modeling the hydromechanical behavior of a natural fractured rock mass still remains a challenging problem in rock mechanics. Multiple disciplines that were often used independently are now being integrated and we may expect significant improvements of our understanding. Flow through single fractures of varying apertures is generally investigated (see the review by Zimmerman & Bodvarsson, 1996) as a function of aperture distribution parameters and contact area. The effect of normal stress to explain the deviation from the well known cubic law at large stress levels, at the sample scale or at larger scale as reported by Raven & Gale (1985) is also widely discussed. Numerous empirical models have been proposed for the normal closure behavior of joints (Bandis, 1983, Brown & Scholz, 1985). The relationships between the fracture stiffness, as a link between the hydraulic and seismic properties of a single natural fracture are also reported (Pyrak-Nolte, 1992). Unger & Mase (1993) propose a theoretical model where the aperture distribution is determined during the closure of two random elastic surfaces. During closure, asperities that come into contact deform elastically and deformations are transmitted through the infinite halfspace to neighboring asperities. Capasso & a1.(1999) use a similar approach using an elasto-plastic behavior for the contact zones. Both approaches lead to a non linear dependency between normal stress, ratio of contact area and mean closure. They provide at each stress levels the parameters required by Zimmerman and Bodvarsson (1996) formalism to derive the equivalent hydraulic conductivity of a rough fracture. An application at the field scale of coupled models using such non linear stress-closure relationships is illustrated by Rutqvist (1995) on single fractures isolated in vertical bore holes by two impermeable packers. The purpose was to determine their normal stiffness from well tests, namely pulse tests and hydraulic jacking tests. However the tensile situation could not properly be reproduced in these simulations without starting to adjust additional parameters, like the fracture size, the fluid pressure at the fracture tip or the Young's modulus of the surrounding rock blocks. Coupling between flow along a given fracture and stresses across this fracture under site specific conditions is therefore dependent upon the global mechanical behavior of the host rock and the hydraulic conditions at the fracture boundaries or at intersections with other fractures, that will control the ov
- Geology > Geological Subdiscipline > Geomechanics (1.00)
- Geology > Rock Type > Igneous Rock > Granite (0.61)
ABSTRACT: Hydraulic fracturing is used at Moonee Colliery to induce caving as part of the routine operation of this longwall mine. Measurements undertaken to successfully introduce hydraulic fracturing to Moonee and pressure records routinely obtained from each treatment provide a unique opportunity to develop and test a new model of hydraulic fracture growth near a free surface. This paper presents the results of the comparison for several fracture treatments, demonstrating that the model is able to match the treatment data. INTRODUCTION Moonee Colliery is owned and operated by Coal Operations Australia Limited and the mine is located just south of Newcastle, NSW at Catherine Hill Bay. Mining extracts 3.2 m of the Great Northern seam, leaving, on average, 1.8 m of roof coal and claystone above the seam. This weak roof coal sequence typically caves immediately behind the supports leaving the 30 to 35 m thick conglomerate section bridging the 100 m wide longwall panel. Hydraulic fracturing is used at Moonee Colliery to grow fractures in the roof rock behind the face. The fractures formed are horizontal and parallel to the base of the conglomerate. Their growth produces caving of the massive conglomerate roof strata at an interval designed to avoid natural caving events. A new model that accounts for the strong interaction of the hydraulic fracture with the base of the roof conglomerate (a free surface) has been developed. The development and application of the model are the subject of this paper. down both toward the rib line and the face. Further back from the face, the arch maintains an essentially constant shape with a maximum height or 15 m above the base of the conglomerate at the centerline of the panel (see the profile in Fig. 1 labeled "old panel"). Figure 1 contains calculated contours of the vertical stress magnitude on vertical sections across and along the longwall panel.
- Oceania > Australia (0.48)
- North America > United States (0.47)
- Geology > Rock Type > Sedimentary Rock > Organic-Rich Rock > Coal (1.00)
- Geology > Geological Subdiscipline > Geomechanics (1.00)
ABSTRACT: This paper describes a numerical model to simulate the propagation of a plane-strain (KGD) hydraulic fracture in an elastic, impermeable medium with zero toughness. The fracture is driven by injection of an incompressible fluid with power-law rheology. The numerical model, which is formulated in terms of a moving coordinates system, is based on the displacement discontinuity method and on an explicit finite difference scheme. The accuracy of the algorithm is validated against the available self-similar solution for a Newtonian fluid. INTRODUCTION Hydraulic fracturing (HF) is a technique widely used to enhance the flow of oil or natural gas from the reservoir formations towards the extracting wells. The uncertainty about the in situ conditions, the complexity of the mechanisms taking place and the difficulty in obtaining precise measurements of the fracture geometry, make necessary the use of idealized models (e.g., the KGD or "plane-strain" model, the "pennyshaped" or radial model and the PKN model) for studying this process. Even with these simplified models, the mathematical formulation for the propagation of hydraulic fractures is given by a relatively complicated system of integral and non-linear differential equations. Some analytical solutions of these mathematical models are already available (Spence & Sharp, 1985; Savitski & Detournay, 1999; Garagash, 2000; Savitski, 2000; Adachi, 2001; Adachi et al., 2001). However, these solutions are constructed on the basis of various restrictive assumptions (e.g., constant injection rate; very small or very large material toughness, Newtonian theology for the fracturing fluid, no fluid leakoff, etc.). In order to extend the applicability of these models, it is necessary to release some of these assumptions and consequently, the solution of the governing equations demands the use of numerical techniques. Commonly, the numerical solution of non-linear problems is restricted to the use of implicit schemes. Explicit schemes are not often applied in this type of problems due to the difficulty in reaching numerical stability, even though the latter are simpler to implement. In this paper, we introduce an explicit finite difference scheme with a moving mesh that can be used to simulate the propagation of a KGD hydraulic fracture. This scheme is shown to be numerically stable, accurate and "flexible", in the sense that additional features (e.g., fluid leak-off, poroelastic effects, etc.) can be easily incorporated into the model (Detournay et al., 1990).
ABSTRACT: We carried out hydraulic fracturing tests in hollow cylinders of Tablerock sandstone (porosity: 26%) subjected to vertical, and horizontal far-field stresses, and initial pore pressure. Pressuretime records reveal that hydraulic fracture initiation (breakdown) occurs before peak pressure is reached in the first pressure cycle. In a series of tests in which รณh, รณv, and Po were kept constant throughout, breakdown pressure Pc increased significantly with wellbore pressurization rate, but appeared to asymptotically tend to an upper and a lower bound. These bounds correspond to fast and slow rates, respectively, as expected by Detournay-Cheng criterion. However, tests conducted at a preset pressurization rate in unjacketed specimens (รณh = Po) revealed that (PcPo) increased with confining/pore pressure, contrary to the constant (PcPo) predicted by the Detournay-Cheng criterion. We modified the criterion by replacing the Terzaghi effective stress (รณeff = รณ Po) with a general effective stress (รณeff = รณ รขPo, where 0 รข<1), and determined ,8 from the unjacketed test results. Several series of hydrofracturing tests in jacketed specimens reinforced the applicability of the modified Detournay-Cheng criterion to Tablerock sandstone in terms of correctly estimating the relationship between the unknown far-field stress and the typically known test parameters: breakdown pressure, initial pore pressure, tensile strength and pressurization rate.
- Geology > Geological Subdiscipline > Geomechanics (1.00)
- Geology > Rock Type > Sedimentary Rock > Clastic Rock > Sandstone (0.86)
ABSTRACT: Longitudinal and transverse fractures were observed in two laboratory hydraulic fracturing experiments under similar loading conditions. In-depth data analyseshowed that the difference might be due to the secondary stress field induced by the packers. A three-dimensional, non-linear module from a comn?rcial numerical simulator (ABAQUS) was used to model the stress field in the tested blocks. It is found that when the packer is working properly, it transfers tensile stress concentration from the packer edge to the central section of the sealed borehole and forms a longitudinal stress concentration band. This band induces a longitudinal fracture initiated from the wall of the sealed section. On the other hand, if the packer malfunctions, high tensile stress concentrations will be induced at its edges. As a result, circular tensile stress concentration bands form which eventually initiate transverse fractures. INTRODUCTION Packers are widely used in both field hydraulic fracturing (I-IF) operations (Brown et al. 2000) and laboratory experiments (Guo et al. 1993, Morita et al. 1996, Wilson et al. 1999). The primary function of packers is to seal the pressurized section from the rest of the borehole. There have been many investigations on the influence of packers. While most of them were concemed with field applications, some involved topics related to laboratory work. yon Schoenfeldt & Fairhurst (1969) were the first to mention that the stress field in the borehole would be influenced by the type of packer, though no quantitative results were given in their paper. Using a 2-D FEM simulator, Roegiers et al. (1973) investigated the distribution of longitudinal and circumferencial stresses in the borehole near a packer. Both influences of the packer rigidity and the steel mandrel length were studied. Ong (1994) investigated the function of different packers on laboratory tests of inclined boreholes and developed an epoxy to backup the packers located in the borehole, limiting their inherent deformation. In a recent experimental study, longitudinal and transverse fractures were observed in two different laboratory tests, which were run under the same conditions except for the packer length. The purpose of those HF experiments was to investigate the fracture initiation and propagation in asymmetrical stress field situations. The one with a longer packer showed normal fracturing behavior, which means the fracture initiated on the borehole wall in the sealed section and propagated in the direction of higher stresses (Scott et al. 2000). On the other hand, the one with a shorter packer showed transverse fracturing behavior, that is, the fracture initiated at the end of the packer and propagated in the direction perpendicular to the borehole. Due to the complexity of the applied stress, the previously mentioned 2-D FEM modeling could not fully explain the results. In order to understand these phenomena, 3-D FEM numerical modeling using ABAQUS was carried out.
ABSTRACT: Research conducted by the National Institute for Occupational Safety and Health (NIOSH) has shown that knowledge of the direction and magnitude of the in situ horizontal stress can be critical to the success of many underground mining operations. One techniquemployed by NIOSH for determining in situ stress has been hydraulic fracturing from boreholes drilled in mine. However, in some cases the combined rock strength and stress magnitude have been too high to allow successfid stress determination using available hydraulic fracturing equipment. NIOSH developed high pressure straddle and impression packers capable of operating at pressures up to 103 MPa in a 51 mm diameter borehole. The high pressure hydraulic fracturing system was used to measure horizontal and vertical stresses in a western Pennsylvania limestone mine with a high incidence of stress related roof failures. Breakdown pressures ranged from 43.9 to 95.7 MPa and in two cases the rock could not be broken down even at the maximum system pressure, 103 MPa. The estimated ranges for the maximum horizontal principal stresses were 27.2 to 65.9 MPa. The measured vertical stress was 3.8 MPa. The calculated K ratio was betweeh 7 and 17. The stress directions, determined from the crack directions in roof and floor holes, were N76รธE and N66รธE, respectively. Because of the high breakdown pressures required to initiate the fractures, neither of the horizontal stress measurements could have been made with any other existing hydraulic fracturing equipment. The author believes the packers represent a significant improvement in the state of the art of in situ stress measurement under near mine conditions. INTRODUCTION Research conducted by the National Institute for Occupational Safety and Health (NIOSH) over the last decade has shown that horizontal stress can have a significant effect upon the success or failure of underground mines (Molinda et al. 1992, Mark & Mucho 1994). Recent studies by NIOSH have documented cases in both coal (Dolinar et al. 2000) and limestone (Iarmacchione et al. 1998) mines where understanding and mitigating the effects of horizontal stress have been critical to the safety and continued operation of mines. Research (Agapito et al. 1980, Mark & Mucho 1994) has also shown that horizontal stresses in the range of 10 to 30 MPa, at mining depths from a few tens of meters to well below 1000 m, once characterized as "high" are actually a predictable consequence of plate tectonics (Zoback & Zoback, 1989). An important part of NIOSH's recent research efforts has been the development of methods to identify the conditions under which horizontal stress adversely affects mining operations, and making the mining industry aware of both the effects of horizontal stress and the means of mitigating these effects. NIOSH and others have used several techniques to measure horizontal stress and assess its potential for causing ground control problems, including the "stress mapping" technique (Mucho & Mark 1994), direct measurement of strain, generally through overcoring and the technique of interest in this paper, the hydraulic fracturing method. This technique is sufficiently well documented to be standardized under ASTM Designation D 4645-87. Although ASTM D 4645-87 is applicable to any depth, the standard is primarily concerned with stress measurements conducted in vertical boreholes drilled from the surface, rather than from underground mine sites, from which most of NIOSH's stress testing has been conducted. Most of the equipment used for conducting hydraulic fracturing stress measurements has been developed as an outgrowth of the oilfield technique of hydraulic fracturing to increase oil and gas production and is of
- Geology > Geological Subdiscipline > Geomechanics (1.00)
- Geology > Rock Type > Sedimentary Rock > Carbonate Rock > Limestone (0.47)
- Government > Regional Government > North America Government > United States Government (1.00)
- Energy > Oil & Gas > Upstream (1.00)
ABSTRACT: Flow in fractured porous media is investigated by a direct, exact and complete numerical solution of the flow equations for arbitrary distributions of permeabilities in the porous matrix and in the fracture network. For single phase flow, the macroscopic permeability of the network has been systematically computed; the main parameters are the fracture density, the ratio between the characteristic fracture and porous medium permeabilifies. Some results are presented and discussed; they show the importance of the percolafion threshold of the fracture network and possibly of the porous matrix. The same work is underway for twophase flow. The unsteady non-linear equations are solved in the same general setting. The method of solution is presented, together with illustrative simulation results. INTRODUCTION Consider a set of permeable fractures embodied in a permeable porous solid matrix. When a fluid is flowing through such a medium, the fractures and the porous matrix interact. Historically, this complex situation was first addressed by Barenblatt & Zheltoy 0960) and Barenblatt et al. 0960); though these papers motivated many further works, progress in this area has been slow and so far no complete solution of this problem for any fracture network and for any permeability distribution in the porous matrix has been described to the best of our knowledge in the literature. Warren & Root (1963) modeled a fractured porous rock as an idealized system made up of identical rectangular porous parallelepipeds separated by an orthogonal network of fractures. Flow is assumed to take place in the fracture network which is fed by the porous blocks. This sort of model was further developed by Odeh (1965). All these equations have been thoroughly studied in the literature with many different boundary conditions (van Golf-Racht 1982, Chen 1989, Chen 1990 and Pinder 1993). Other possible approaches have also been presented by Adler & Thovert 0999). The most important is probably the multiscale analysis of flow through fractured porous media, initiated by Aifantis (1980) and extended by Arbogast (1990), Levy 0988, 1990) and Panfilov (1990,1994). The purpose of this paper is to briefly present the methodology and the first results obtained in the determination of the single-phase permeability of fractured porous media. It is a significant extension of our previous paper on the permeability of fractured media Koudina et al. (1998). In addition, a further extension to two-phase flow is presented, with a set of illustrative results.
ABSTRACT: We consider the use of a modified diffusion-limited aggregation.(DLA) process to model immiscible viscous displacements in rough-walled fractures. We include the effects of local aperture field variability to channelize flow within a fluid phase and capillarity to smooth the displacement front. Model behavior is illustrated over a limited set of simulations in both a Hele-Shaw cell, and a variable aperture field that is correlated well below the simulation scale. INTRODUCTION The issue of immisicble displacement of one fluid by another rises in the context of enhanced oil recovery, remediation of nonaqueous phase liquids, CO2 sequestration, and other applied geologic problems. In discrete fractures, accessibility will combine with competition between capillary, gravity, and viscous forces to control the displacement process. Here we focus on behavior in a single horizontal fracture at the sub-meter scale for conditions where viscous forces cannot be ignored. Immiscible displacement processes can (and have been) modeled using a number of different approaches that may be broadly categorized as continuum or discrete. In continuum approaches, one either: 1) solves the Navier-Stokes equations for each fluid within complex geometry, and separated by a moving boundary where kinetic and dynamic boundary conditions must be met; or, 2) describes the system as a two-phase continuum with relative permeability and pressure-saturation relations defined for both fluids at every point. In the discrete approaches, a set of local 'rules' are applied to model pore-scale behavior; broad categories of rules include: 1) Particle interaction models (PIM) such as cellular automata, lattice-Boltzman, or lattice-gas where large quantities of particles simultaneously move and interact according to a list of rules; 2) Invasion percolation (IP) where individual apertures along the fluid-fluid interface are filled according to local accessibility and capillary criteria; and, 3) Diffusion-limited aggregation (DLA) in which random particles moving through the displaced fluid are used to identify locations on the interface for advancement. Both the Navier-Stokes and PIM approaches are currently limited by the difficulty of defining the fluid-fluid interface, and computational requirements at our scale of interest. Two-phase continuum approaches are computationally tractable, but do not correctly include the physics of displacement processes; i.e. the fluid phases intermingle where they should be separate. IP and DLA have potential at the scale of interest, both are growth models that encapsulate the critical unit process (aperture filling), and iterate that process to yield large scale behavior. The basic IP process models capillary dominated horizontal displacements on a random aperture field. However, IP has recently been modified to include gravity and the dynamic calculation of interfacial curvature at the fluid-fluid interface within a fracture (e.g. Glass, 1993, Glass et al., 1998). The latter, termed MIP, yields simulated behavior that matches experimental measurements quite well. In addition, there have also been recent attempts to further include viscous forces in IP or MIP (Xu et al., 1998, Ewing & Berkowitz, 1998) with excellent correspondence to experiments in heterogeneous porous media where capillary, gravity, and viscous forces all played a role (Glass et al., in press). Here we focus on a DLA based approach for simulating fluid-fluid displacement. The DLA process begins with a cluster that consists of a single occupied grid block at the center of a large regular lattice (Witten & Sander, 1981). A particle introduced at a random location far from the cluster walks randomly on the lattice u
ABSTRACT: The study of hydromechanics has identified that surface roughness has an impact on the flow characteristics of single and two-phase fitfids. Technical developments in the field of two-phase flow are of great importance for improving the understanding of underground inundation and gas outbursts, in order to reduce the risks to personnel. The paper describes recent advances in the understanding of two-phase (airwater) stratified flow. A new constitutive model is presented, based upon an extension of Darcy's Law and using the concept of relative permeability. The proposed model is verified by experimental results using 'state of the art' Two Phase High Pressure Triaxial Apparatus (TPHPTA). This study presents the results of laboratory testing that will enable the development of a relationship between roughness (Joint Roughness Coefficient, JRC) and the flow rate for steady state conditions. INTRODUCTION Fracture roughness in the form of the Joint Roughness Coefficient (JRC) is acknowledged to have a fundamental impact on the hydromechanical properties of discontinuous media. Previous studies (Barton et al, 1985) have used a series of standard roughness profiles that enable the estimation of fracture hydraulic conductivity. This relationship has also been examined in terms of gas flow (Schrauf & Evans, 1986). With the onset of research into twophase flow, several different approaches have been adopted. A number of studies have considered applied numerical techniques, typically using fractures generated with various mathematical models (Rasmussen, 1991). Laboratory studies have also been carried out using artificial fi'actures (Fourar et al, 1993). The nature of two-phase flow is of practical interest to civil and mining engineering projects especially with regard to the storage of waste in fraetured rock mass and minerals extraction in the mining and petroleum industries. The aim of the current research program is to study the effect of roughness on fracture aperture and two-phase flow behaviour for both natural and induced fractures, extending over a range of JRC values. A further objective is to evaluate the effect of capillary pressure, phase interference and fracture roughness on the relative permeability and twophase flow behaviour. In underground rock mass, excavation of multiple openings causestress redistribution and associated fitfid flow through existing and newly created discontintfities. In the Australian coal industry, the risks from gas outburst and groundwater inundation are still only partially understood, and damage to underground eqtfipment and fatalities occur too frequently. The study of two-phase flow characteristics provides a more thorough understanding of nearfield pore pressure variations associated with the redistribution of stresses. Phase interference and the 'blocking-off' of pockets of gas can lead to 1ocalised pressure concentrations that can result in outburst. It is hoped that from the better understanding of fracture characteristics obtained in the laboratory, the field behaviour can be more accurately predicted so that the risks from outburst and inundation can be controlled more effectively.
- Geology > Geological Subdiscipline > Geomechanics (1.00)
- Geology > Rock Type > Sedimentary Rock > Organic-Rich Rock > Coal (0.74)
- Well Completion > Hydraulic Fracturing (1.00)
- Reservoir Description and Dynamics > Reservoir Fluid Dynamics > Multiphase flow (1.00)
- Reservoir Description and Dynamics > Reservoir Fluid Dynamics > Flow in porous media (1.00)
- Reservoir Description and Dynamics > Reservoir Characterization > Faults and fracture characterization (1.00)