Lee, Jong-Hyun (Korea Institute of Construction Technology) | Kim, Jae-Jeong (Korea Institute of Construction Technology) | Yoon, Sang-Won (Korea Institute of Construction Technology) | Lee, Jung-Yub (Korea Institute of Construction Technology) | Koo, Ho-Bon (Korea Institute of Construction Technology)
In Korea, damage from the collapse of artificial slopes along national highways has significantly decreased due to the continuous survey and preparation of measures for the last 17 years, while casualties and property damage from the occurrence of debris flow at the valley parts of natural mountain area have rapidly increased. Accordingly, the necessity of the debris flow management for natural mountain area that is similar to the artificial slope management has been suggested at a national level.
In this study, specific sections vulnerable to debris flow damage were selected, and a complete enumeration survey was performed for the sections with debris flow hazards. Based on this, the characteristics of the sections with debris flow hazards and the current status of actions against debris flow were examined, and an efficient installation plan for a debris flow damage prevention method that is required in the future was suggested.
The results indicated that in the Route 56 section where the residential density is relatively higher between the two model survey sections, facilities for debris flow damage reduction were insufficient compared to those in the Route 6 section which is a mountain area. It is thought that several sites require urgent preparation of a facility for debris flow damage reduction. In addition, a numerical analysis showed that for a check dam installed as a debris flow damage prevention method, distributed installation of a number of small-scale check dam facilities within a valley part was more effective than single installation of a large-scale check dam at the lower part of a valley.
In Korea, more than 70% of the territory consists of mountain area. Thus, formation of slopes depending on road construction was inevitable, and there were many unstable slopes along the roads that had been constructed during the rapid economic growth in the past. Accordingly, in case of concentrated rainfall in the summer season where more than about 60% of the annual precipitation occurs, small and large rock falls and landslides have continuously occurred. However, for artificial slopes along national highways, nationwide surveys and proactive measures have been prepared at a national level for the last 17 years, and thus, casualties and property damage from the collapse of artificial slopes have significantly decreased compared to the past.
Rück, M. (German Research Centre for Geosciences) | Rahner, R. (German Research Centre for Geosciences) | Sone, H. (German Research Centre for Geosciences) | Dresen, G. (German Research Centre for Geosciences)
We studied the initiation and propagation of mode II fractures in granite and sandstone under confining pressure to investigate the controls on shear fracture propagation in rocks. An asymmetric loading set up was used to induce a fracture in cylindrical rock samples under confining pressure between 0-20 MPa. We achieved quasi-static fracture propagation with a refined AE feedback displacement control. This technique prolongs the fracturing process up to 42 hours, provides a higher AE resolution and thereby allowed the distinction of two different stages in shear fracture propagation. Granitic samples form vertical fractures in the strain strengthening stage that branch and stop propagating at peak stress. Simultaneously at peak stress a distinct diagonal fracture nucleates on the loaded side of the vertical fracture. During strain weakening we observed stable growth of this second diagonal fracture until the sample lost its integrity. On the other hand sandstone samples only form the diagonal fracture during the strain weakening stage. Analysis of AE source type and hypocenter as well as microstructural analysis indicate that porosity, either intrinsic (sandstone) or deformation inflicted (granite), primarily influences this fracture nucleation and propagation behavior for both rock types.
Fractures form as microscopic cracks coalesce into a planar structure, which can be captured by monitoring acoustic emissions (AEs). Macroscopically, rocks will fracture in the mode (I or II) corresponding to the mode of loading determined by the orientation of the fracture relative to the stress state. However at a microscopic scale, both modes of fracturing can be found in what may appear to be a pure mode of fracturing at the macroscopic scale [1, 2]. According to the acoustic emissions observed during rock fracturing experiments, shear-, tensile-, and compression-events all occur during macroscopic mode II fracture propagation . On a microscopic scale on the other hand, increased confining pressure suppresses the occurrence of mode I fractures and supports the occurrence of mode II fractures . These observations demonstrate the complexity of rock fracturing on a microscopic scale and thereby raised a controversy about the validity of macroscopic fracture modes. The existence of a fracture criterion for mode II is still debated in the literature [5, 6]. To study this, many authors used acoustic emissions to stabilize the fracture process [1, 7, 8, 9, 10]. By controlling the load in response to the intensity of the observed AEs, the rate of fracturing was varied and fracture propagation could be visualized in detail.
An on-going project at the University of Arizona is using Kartchner Caverns in Benson, Arizona as a natural analog to study time-dependent rock failure with subcritical crack growth modeling. Various material properties of the Escabrosa limestone composing the caverns are required for input into the damage model. Central among these properties are the subcritical crack growth parameters n and A, which can be calculated from modes I, II, and III fracture toughness tests conducted at different loading rates. This paper presents the results of modes I, II, and III testing on Escabrosa limestone, providing the material properties necessary for the larger goal of modeling breakdown in Kartchner Caverns and applying the model to the long-term stability of rock excavations. Additionally, fracture test results are compared with a previous study by Tae Young Ko at the University of Arizona, which tested Coconino sandstone and determined that the subcritical crack growth parameters were consistent among modes. This study expands upon Ko’s work by adding the characterization of a second rock material in all three modes; preliminary results indicate that for Escabrosa limestone the subcritical crack growth parameters are not consistent among modes.
Time-dependent rock failure is an important aspect in the analysis of long-term rock stability for slopes, dam and bridge foundations, and underground storage facilities. Under short-term loading, a crack will propagate when the stress intensity factor at the crack tip exceeds a critical value: the fracture toughness. Subcritical crack growth is the propagation of a crack at values of the stress intensity factor smaller than the fracture toughness. This type of growth happens under long-term loading and is environmentally-assisted, occurring through mechanisms such as stress corrosion, diffusion, microplasticity, dissolution, and ion exchange [1, 2]. There are three modes of crack tip displacement in critical and subcritical crack growth: mode I (tension), mode II (in-plane shear), and mode III (out-of-plane shear).
We have used a hydraulic fracture model to quantify fault rupturing promoted by injection and migration of fluid into a fault, which contains in-plane high-conductivity segments and out-of-plane jogs or branches. The fluid under elevated pressure can promote extension-shear fracturing. The model provides numerical results on the opening and pressure variations with position and time. High-conductivity segments aid penetration of the pressurized fluids, but limit fluid pressure increases, so as to sustain a stable rupture growth mode. In the presence of varying normal stresses, stable growth cannot be maintained by pressure variations leading to the unstable growth in shear, and rapid fault movement events can be triggered as the fault re-establishes stable growth, radiating seismic energy and accompanied by local backward slip. These coupled seismic and aseismic faulting processes are applicable to faults with jogs and branches, which interact with the main fault to produce changes in the local stress states. The opening and slip along jogs and branches, either pre-existing or induced by fluid flow, will not only contribute to fluid storage due to their suction pumping action, but also produce changes in the downstream flow rates. The slip along them can produce associated opening along the main fault, but their opening can increase the compressive stress across the main fault which restricts its opening. The deformation transfer at junctions can complicate the fracture and flow responses.
Injection of liquid waste has been found to result in generation of seismic events and some of these may be large enough to be felt at the surface [1, 2]. The total volume injected and the maximum size of the seismic events generated have been found to be correlated . Microseismic monitoring of low-level induced seismicity generated by hydraulic fracturing is a technology applied in unconventional gas reservoirs [4- 6] for the purpose of mapping the extent of fracturing.
Considerable effort has been devoted to detecting and locating seismic events associated with fluid injection. However, little attention has been paid to understanding the source mechanisms of seismic events for a pressurized fault [7, 8]. Hydraulic fracturing in lowpermeability reservoirs is strongly affected by natural fractures within the targeted rock layers. Stimulation improves the connection of these fractures to one another and to the hydraulic fracture. Slip on these natural fractures can be activated by stress changes to generate low-magnitude seismic events. Meanwhile, shearing and shear-induced dilation on the natural fractures enhances conductivity and increases the stimulated volume. The aperture distribution along a fracture, including the initial and subsequent propagation paths, controls the conductivity and pressure distributions resulting from the injection. An integrated coupled hydraulic fracture model is applied in this paper to this problem to predict the relationship among stress, pressure, deformation and rupture growth.
This paper reports new developments on the complex variables boundary element approach for solving three-dimensional problems of cracks in elastic media. These developments include implementation of higher order polynomial approximations and more efficient analytical techniques for evaluation of integrals. The approach employs planar triangular boundary elements and is based on the integral representations written in a local coordinate system of an element. In-plane components of the fields involved in the representations are separated and arranged in certain complex combinations. The Cauchy- Pompeiu formula is used to reduce the integrals over the element to those over its contour and evaluate the latter integrals analytically. The system of linear algebraic equations to find the unknown boundary displacement discontinuities is set up via collocation. Several illustrative numerical examples involving a single (penny-shaped) crack and multiple (semi-cylindrical) cracks are presented.
Accurate three-dimensional modeling of fracture is of key importance for rock mechanics applications such as simulation of mining in faulty rock or computer simulation of hydraulic fracturing. In both applications, the Boundary Element Method (BEM) can serve as an efficient tool for realistic modeling of mechanical deformation of fractured rock. In previous papers [1, 2] we introduced a new approach for solving three-dimensional crack problems. The approach featured the following new elements:
• The use of complex variables to create various combinations of the fields, e.g. in-plane components of tractions, displacement discontinuities, as well as geometric parameters.
• The use of triangular elements and analytical integration over those elements.
• The use of the “limit after discretization” procedure, i.e. enforcing the boundary conditions after the discretization and analytical handling of the internal fields.
The approach was illustrated in  for a simple case of constant approximations of the unknowns. In , the method was further extended to include higher order approximations of unknowns. In addition, a new technique is employed to analytically evaluate the involved integrals. It is based on the reduction of the area boundary integrals over an element to those over its contour using complex integral representations. The technique allows for enrichment of the library of boundary elements by including those bounded by combination of straight lines and circular arcs.
The Wadden Sea is a shallow tidal sea in the north of the Netherlands where gas production is ongoing since 1986. Due to the sensitive nature of this area, gas extraction induced subsidence must remain within the “effective subsidence capacity” for the two tidal basins (Pinkegat and Zoutkamperlaag) affected. We present a probabilistic method to monitor the “effective subsidence capacity” and ensure that subsidence is below the long term (18.6 years) volumetric rate for relative sea level rise that can be accommodated by the tidal basins without environmental harm. The role of sedimentation volume rate, relative sea level rise and subsidence volume rate due to gas depletion are taken into account including their uncertainties. The probability of exceeding the acceptable subsidence limit for the period 2012 to 2050 is 2.8% for the tidal basin called Zoutkamperlaag and 1% for the tidal basin of Pinkegat for climate scenarios that fit the current relative sea level rise observations on the Dutch coast. The values are shown to be dominated by the effect of relative sea level rise, and not due to subsidence induced by gas depletion in the Wadden Sea. To current knowledge no harm is done to nature.
Subsidence caused by extraction of hydrocarbons is a sensitive issue in the Netherlands due to its proximity to sea level. The Wadden Sea in the north of the Netherlands is a shallow tidal sea behind a chain of coastal barrier islands. It has been inscribed on the UNESCO’s World Heritage list since 2009 because of its unique morphodynamic features and its wildlife. Also, it is one of the most notable nature conservation areas in the Netherlands protected under the European Birds and Habitats Directives. In the Wadden Sea, gas production is ongoing since 1986 from the Ameland gas field (Fig. 1). Since 1990 the subsidence caused by depletion of the Ameland gas field is compensated by sand suppletions under a dynamic ‘maintain the coast line’ preservation policy . In 2006 additional gas production commenced in the Nes, Moddergat, Lauwersoog and Vierhuizen fields (Fig. 1). Due to the sensitive nature of this tidal sea the gas fields in the area (Ameland Nes, Lauwersoog, Vierhuizen, Anjum, Ezumazijl, Mestlawier and Moddergat) are being produced within the so-called “effective subsidence capacity” .
Han, Yanhui (Shell International Exploration and Production Company) | Tallin, Andrew G. (Shell International Exploration and Production Company) | Wong, George K. (Shell International Exploration and Production Company)
In this paper, the integrity of sand screen is evaluated for various depletions using fluid-mechanical coupling analysis. The interaction between the screen and reservoir is captured by rock-structure interaction model, based on laboratory tests that measured the mechanical properties of ceramic proppant, the formation sandstone and sand screen. The frictional hardening properties of the ceramic proppant and sandstone core plugs are measured by performing constant mean-stress tests in rock mechanics laboratory. The volumetric hardening curve of sandstone is measured in isotropic compression tests (with unloading excursions). The elasto-plastic mechanical properties of sand screen are determined by calibrating the laboratory crushing test data. For a given maximum depletion, this model can be used to select appropriate sand screens; conversely, for a given sand screen, this model can be used to estimate the depletion where failure is likely.
The reduction of formation pore pressure induced by fluid production in water and hydrocarbon reservoirs could result in many serious consequences, including subsidence of earth surface, loss of overburden integrity, reduction of porosity and permeability of formation, damage of well construction and production parts such as casing and tubing strings (Schutjens et al., 2004, 2008).
Open-hole standalone screens and gravel packs are often-used sand control methods in unconsolidated sandstone. As reservoir pressure depletes, the effective stresses in the formation increase. The depletion-induced increase in effective stresses are further magnified by the near wellbore stress concentration. Because the strength of unconsolidated sandstone is low and the gravel pack provides only a limited support to the wellbore wall, plastic deformation near the well seems inevitable in many wells. As a result, the forces acting on the sand screen increase. When formation loads exceed the loading capacity of the sand screen, the integrity of the screen is lost resulting in sand control failure and lost production.
We simulated the Defense Threat Reduction Agency (DTRA)-sponsored Jointed Limestone Test (JOLT) using the Abaqus fully coupled Euler-Lagrange computation scheme. JOLT was a small-scale experiment consisting of over 37,000 individual cubes of limestone surrounding simulated tunnels. This test bed was loaded with ground shock generated by a spherical CompB charge, thereby providing data for the study of computer codes used to compute the response of buried tunnels to explosive loading. We modeled the cubes as continuum blocks with contact, making for a dis-continuum simulation using a hybrid continuum code including Eulerian modeling of the explosive region and Lagrangian modeling of the remainder of the test bed. This paper has been approved for unlimited release under LA-UR-15-21382.
The Defense Threat Reduction Agency (DTRA) has funded studies of tunnel vulnerability to blast effects for decades. The studies have included experiments of various complexity and scale as well as numerical analysis. Several projects have combined computational studies with experiments to validate predictive algorithms. Numerical approaches have evolved based not only on prior knowledge but also due to advances in computing hardware and software capabilities. The focus of this paper is on a recent three-dimensional (3-D) simulation effort, but we will also provide some background material by describing an older twodimensional (2-D) computational effort that laid some of the groundwork for this later work
2. MIGHTY NORTH
The MIGHTY NORTH event was a laboratory-scale high explosive test performed in the mid-90s to investigate various approaches to numerical simulation of the response of tunnels to explosive-induced ground shock [1, 2]. The test bed (Figure 1) consisted of 1.2-m long, 5-cm square cross-section bars of limestone stacked in an imbricate pattern to simulate a layered, jointed geologic setting. The “block of bricks” was 2 m square in section and 2 m long. The center of the block included a 0.4-m diameter aluminum-lined “tunnel.”
Wellbore zonal isolation is particularly important for subsurface storage of CO2 , where well integrity must be ensured for very long time spans. In this study, three dimensional discrete element models of wellbore systems have been used to simulate failure and damage of wellbore cement and surrounding rock. The models allow simulation of wellbore failure and damage in wellbore systems with different geometries and with perfect (idealized) or imperfect plug-casing-cement and cement sheath-rock interfaces. The aim is to determine critical stress conditions for mechanical wellbore failure and associated damage. Comparison of model simulations with conventional geomechanical analysis is used to determine limits in injection pressures. If these limits are exceeded, wellbore failure and upward fluid migration through wellbore cement may occur due to alignment of fractures and the formation of connected fracture networks. Imperfect cementation or a cement sheath that is not evenly distributed around the wellbore enhances local damage or failure. Although the analysis relies on upper bounds on the changes in stresses around wellbores caused by an increase in pressure, it shows that loss of well integrity during injection of CO2 or other fluids may occur, in particular by axial loading due to reservoir compaction.
Wells that penetrate subsurface CO2 storage complexes may act as potential pathways for leakage of CO2 to the surface [e.g., 1-3] . For storage of CO2 to be meaningful, well integrity has to be ensured for large injection volumes and over much longer time scales (100’s to 1000’s of years) than in the case of hydrocarbon production (10’s of years of depletion). Ensuring well integrity not only during injection period but also in the post-operational period over such long timescales represents a challenge for Carbon Capture and Storage (CCS) projects that is not encountered before in the E&P industry. The long term-ability of wells to inhibit CO2 migration has been identified as a significant potential risk for the long-term security of geological storage facilities [e.g., 2]. The risk of leakage through “old” abandoned wells requires particular attention. Regulations for well abandonment were less comprehensive than they are today. In addition, future use of abandoned reservoirs for CO2 storage was not taken into account when old wells were completed and abandoned. The risks of leakage through abandoned wells may be a showstopper for a CCS project as it was in the case of CO2 storage in the depleted De Lier field in the North Sea, offshore of the Netherlands . The project was discontinued after a feasibility assessment of CCS found that there was a high potential for leakage through old wells. Repair of these wells was uneconomic and in some cases, technically impossible.
This paper employs the Particle Flow Code in 2 dimensions (PFC2D) to simulate the shear behaviors of rock joints under direct shear tests. A series of rock joints and regular joints with various saw-tooth angles are investigated. The shear resistance - shear displacement relationship, shear strength parameters, crack propagation, and failure modes are observed and investigated. Based on the numerical simulation results and application of the regular joint models, the simulation results compared well with the shear strength models of Patton  and Zhang  for higher and lower normal stresses but only compared well with the model of Zhang  for critical normal stresses. For observation of crack propagation, the initial micro-cracks appeared near the saw-tooth interfaces at 60% ~ 70% of the peak stress state before the peak. Soon, micro-cracks began to propagate into a macro-crack until the stress state reached the peak. The simulated failure modes are compare well with Zhang . This paper further established the chart of the distribution of failure modes based on the saw-tooth angle and the normal stress conditions
Rocks usually accompany inherent discontinuities, including joints, cleavage, and foliation, which were occasioned by tectonic stresses and pressure release. For engineering activities of tunnel excavation and slope stability, a rock joint is one of key components that could lead to tunnel collapse or slope failures during construction. Several factors could influence the mechanical behaviors of a rock joint, i.e., roughness, persistence, wall strength, aperture, and spacing, of which roughness is one of the most important factors for rock sliding.
The natural rock joint geometry never follows a specific regular pattern. Hence, most of the shear strength criteria of a rock joint have been developed by empirical analysis [1-2]. Several scholars simplified the rock joint geometry as a saw-tooth shape and established analytical solutions by observing the results of experimental tests [3-4]. Although shear strength criteria can be developed by experimental tests, the sliding failure process is still difficult to observe. Numerical simulation is one of approaches that can help to understand the failure mechanism, and the use of the discrete element method (DEM) is a magnificent demonstration of this capability [5-13]. To understand the relationship between the joint failure process and its shear strength, a fundamental investigation is required (a simplified joint geometry (saw-tooth), a basic sliding test (direct shear test, DST), and a 2 dimensional numerical analysis were considered). Hence, this paper used the Particle Flow Code in 2 dimensions (PFC2D, a DEM program) to simulate the mechanical behaviors of saw-tooth rock joints under the direct shear test. The shear resistance - shear displacement relation, the shear strength parameters, the crack propagation, and the failure modes will be discussed in this paper.