Layer | Fill | Outline |
---|
Map layers
Theme | Visible | Selectable | Appearance | Zoom Range (now: 0) |
---|
Fill | Stroke |
---|---|
Collaborating Authors
Results
Summary We introduce a rapid, quantitative nuclear magnetic resonance imaging (NMRI) technique to resolve and measure multiple fluid phases in porous media. Liquids are resolved on the basis of their NMR spin-spin (T2) relaxation times, and their intensities are corrected via attenuation analysis. The spatially resolved and corrected NMRI intensities are normalized to yield fluid saturations. In-situ saturation measurements are presented for three immiscible (oil and water) displacements in the same Berea sandstone core. NMRI and effluent recovery methods were compared. T2 of the displacement fluids were observed to be sensitive to displacement conditions. Introduction Fluid-fluid displacements in porous media are important for subsurface waste remediation techniques and enhanced oil recovery methods. However, they are poorly understood processes, due to the complexity of natural porous media, fluid-solid and fluid-fluid interactions at the pore scale. Nondestructive experimental technologies are shedding new light upon fluid-fluid displacement processes. Nuclear magnetic resonance (NMR) methods are particularly powerful, yielding both petrophysical (porosities, permeabilities, and pore sizes) and transport (displacement saturations, diffusion and flow rate) information. NMR involves the manipulation and detection of the magnetic moment of certain atomic nuclei which are precessing in a static magnetic field. The response signal M is detected and processed for information. NMR imaging (NMRI) involves the spatial mapping of the nuclei by rendering their precession frequencies spatially dependent during manipulation or detection, typically by implementing applied magnetic field gradients (ga). NMR petrophysical measurement techniques were recently reviewed. In particular, NMR spin-lattice (T1) relaxation times measured for pore water have been related to the pore size distribution (PSD) and permeability. T1 and NMRI measurements can be combined to yield spatially-resolved petrophysical properties. T1 measurements have been used to study wettability effects upon fluid displacement processes in chalks and Berea sandstone. Spatially and temporally-resolved fluid distributions (saturations) are provided by NMRI of laboratory corefloods. NMRI methods for studying multiple-fluid systems must be capable of resolving the liquid phases. One straightforward approach is to remove the signal of a fluid by paramagnetic doping or isotopic (deuterium) substitution. Another approach is to resolve fluids based upon differences in their nonimaging NMR (spectral, relaxation, or diffusion) parameters. Another concern is that NMRI-derived fluid distributions are often rendered semiquantitative due to NMR signal attenuation. Signal attenuation arises due to NMR spin-spin (T2) relaxation and molecular diffusion attenuation processes, and varies with the rock and fluid properties and with the time-dependent saturation. The signal intensity cannot be reliably corrected by calibration since attenuation varies in both space and time. Additional signal loss is probable when the (nonimaging) NMR spectral linewidth is large compared with the radiofrequency (RF) pulse bandwidth. Finally, RF attenuation losses should be considered for samples with high electrical conductivity. Typical attenuation losses are 10- 80% of the original signal. Thus, quantitative NMRI techniques are needed for investigating fluid transport in porous media. Quantitative imaging techniques collect and implement the information to independently correct each image pixel intensity at each timestep Quantitative NMRI involves collecting several progressively-attenuated images. The image intensities are then fitted, pixel-by-pixel, to an appropriate attenuation model. The fitting procedure yields the corrected image intensities and the corresponding spatial attenuation rate constants (T2 times), which are related to the petrophysical and fluid properties of the sample. Quantitative NMRI was first applied to porous media by Rothwell and Vinegar who used a Carr-Purcell-Meiboom-Gill (CPMG) multiple-spin-echo technique to resolve hydrogen spin density and T2 in brine-saturated sandstone cores.
- North America > United States > West Virginia (0.46)
- North America > United States > Pennsylvania (0.46)
- North America > United States > Ohio (0.46)
- North America > United States > Kentucky (0.46)
Summary Temperature logs have been commonly used to evaluate fracture height by locating cool anomalies which indicate the locations where cool fracture fluids were injected. However, instead of cool anomalies, warm anomalies (called "warm noses") often occur on temperature logs run after fracture treatments. When interpreting fracture height from a temperature log, warm anomalies make it difficult to identify the top and bottom of the fracture. We believe that a plausible reason for warm anomalies is that the wellbore and the fracture are not coincident over the entire extent of the fracture; instead, away from the perforations, the fracture and the wellbore may be separated a finite distance that varies with depth. This paper investigates the effect of the existence and magnitude of the displacement between the wellbore and the fracture on well bore temperature behavior after fracturing. The results obtained explain the "warm noses" on shut-in temperature logs run after a fracture treatment and, more generally, provide an improved method for interpreting fracture height from temperature logs. A mathematical model has been developed to simulate the wellbore temperature after fracturing for cases where the wellbore and the fracture are not coincident for the entire extent of the fracture. The study shows that the temperature behavior strongly depends on the pattern and the magnitude of the displacement. When the fracture is perfectly connected with the well bore, the cool region on the log indicates the top and bottom of the fracture clearly. However, the cool region is much smaller than the fracture height if the wellbore deviates from the fracture at a constant angle away from the perforations - a situation that may occur in a well that is deviated slightly from vertical, for example. Furthermore, "warm noses" appear on the log if the well spirals in a helical trajectory because the spacing between the wellbore and the fracture will vary with depth for this geometry. Therefore, when evaluating a post-fracture temperature log, the possibility of a deviated wellbore-fracture system must be considered to avoid misinterpreting the fracture height. Longer shut-in times are shown to improve the fracture height interpretation. Introduction Currently, the most widely used method for determining the height of a hydraulic fracture is to run a temperature log on the shut-in well shortly after the fracturing treatment.
Summary With a distinct element code, we analyzed a compartmentalized North Sea oil reservoir's mechanical behavior during fluid pressure drawdown. Special emphasis was given to the affect of compartment pressure reduction on the activation of existing faults, the change of the fault-sealing properties as a result of fracturing of the intact rock mass, and the determination of the existing faults' behavior during fluid pressure drawdown with respect to shear displacement and changes in the fault/fracture conductivities. The simulations showed that several segments of the existing faults in the vicinity of the drawdown region could be activated. Some faults with shear stresses below peak showed large shear displacements and were therefore considered to be activated. Shear fracturing of the intact rock was found to be possible in several zones around the drawdown region. Introduction The field is a fault-bounded sandstone reservoir located in the Central graben. During production, the fluid pressure in the 4600-m deep reservoir could be reduced by 40 to 50 MPa. The reservoir is most likely subdivided into several compartments, and depletion is expected to lead to high pore-pressure gradients across sealing faults. Simulation of the rock mechanical aspects of the fault behavior during depletion had the following aims:To determine whether the existing faults could be activated during fluid pressure draw down, assuming that the existing faults are healed before drawdown. To determine whether the intact rock mass will fracture during fluid pressure drawdown. To determine the nonlinear behavior of the existing faults during fluid pressure drawdown; in particular, whether the faults could experience significant shear deformations and changes in the fault/fracture conductivity. Analyses were performed by use of the 2D universal distinct element code (UDEC), with the Barton-Bandis nonlinear fracture behavior model, NGI-BB (UDEC-BB).
- North America > United States (1.00)
- Europe > Norway > North Sea > Central North Sea (0.34)
- Europe > United Kingdom > North Sea > Central North Sea (0.24)
- Geology > Geological Subdiscipline > Geomechanics (1.00)
- Geology > Rock Type > Sedimentary Rock > Clastic Rock > Sandstone (0.54)
- Geology > Geological Subdiscipline > Economic Geology > Petroleum Geology (0.34)
- Oceania > Australia > Western Australia > Ashmore Cartier Territory > Timor Sea > Bonaparte Basin > Londonderry High > Vulcan Basin > Eclipse Field (0.99)
- Oceania > Australia > Western Australia > Ashmore Cartier Territory > Timor Sea > Bonaparte Basin > Bonaparte Basin > Vulcan Basin > Eclipse Field (0.99)
- North America > United States > California > Los Angeles Basin > Wilmington Field (0.89)
- (2 more...)
- Well Completion > Hydraulic Fracturing (1.00)
- Reservoir Description and Dynamics > Reservoir Characterization > Reservoir geomechanics (1.00)
- Reservoir Description and Dynamics > Reservoir Characterization > Faults and fracture characterization (1.00)
- Reservoir Description and Dynamics > Reservoir Characterization > Exploration, development, structural geology (1.00)
Abstract This paper investigates the mobilization and recovery of waterflood residualoil by immiscible gas flooding in 2D glass micromodels under strongly oil-wetand water-wet conditions. Two oil/water/gas systems, one displacing a positive and the other a negative spreading coefficient (oil over water), were used. The results clearly show the importance of fluid-solid and fluid-fluid interactions on tertiary oil recovery. Oil recovery was highest for the oil-wet displacements where the oil formed thick oil wetting films. Recovery was lowest for considerably lower for the positive spreading system under strongly water-wet conditions where the oil formed continuous but thinner spreading films. Recovery was lowest for the negative spreading system under strongly water-wet conditions where oil films were not formed. Introduction The recovery of waterflood residual oil by immiscible gas flooding has been the subject of a number of recent studies. Although these studies have demonstrated that high oil recoveries are possible under conditions where gravity drainage is important, little is presently understood of the basic mechanisms responsible for the mobilization of waterflood residual oil, or of the factors which influence the effectiveness of the pore scale processes responsible for oil recovery - particularly the roles played by wettability and the spreading coefficient. An understanding of these processes is necessary both for the development of improved gas flooding applications and for predicting phase mobilities undertertiary gas flooding conditions. In a previous study of the pore scale mechanisms responsible for the recovery of waterflood residual oil in glass micromodels under strongly water wet conditions, Oren et al. have shown that oil is displaced by a double-drainage mechanism in which a gas-oil displacement(first drainage event) is always associated with a corresponding oil-water displacement (second drainage event). The second drainage event leads to there connection of residual oil blobs and to water production at the model outlet. This mechanism is responsible for the frequently reported observations that high oil recovery in laboratory tertiary core floods is associated with highwater recovery and that gas breakthrough usually occurs shortly after oil breakthrough. Further, Oren et al. show that recovery efficiency in water-wet displacements depends on the sign of the oil-water spreading coefficient for the gas-oil-water system. Oil recovery for a positive spreading system was demonstrated to be significantly higher than that for the corresponding negative spreading system. The higher recovery for the positive spreading system was shown to result from flow through thin but continuous oil films which play an important role in mobilising and reconnecting waterflood residualoil. The absence of these films for the negative spreading system resulted in lower oil recovery.
Summary Coreflood experiments in naturally heterogeneous sandstone outcrop cores were conducted and simulated. Displacements included waterfloods and polymerfloods using tracers for the oil and/or water during both single-phase flow and two-phase flow. Those involving the displacement of viscous polymer solution were unstable. The corn were characterized by polymer solution were unstable. The corn were characterized by air permeability measurements on each square centimeter of each face and by C.T. scans of cross-sections of the core spaced one Centimeter apart along the length of the core. Fine-grid simulations were then made using these characterization data as input. The agreement between the experiments and simulations is good. Simulations using coarser grid physical descriptions were then made using effective properties. Introduction The results given in this paper summarize our efforts to use numerical simulation to analyze our own experimental data on naturally heterogeneous sandstone cores called Antolini sandstone. Antolini sandstone is an eolian outcrop in northern Arizona. A series of displacements were conducted in two samples of this outcrop. Experiment Al is one of several corefloods done by Wreath to study polymerflooding in heterogeneous cores. Experiment Ahtolini-4 is one of several corefloods done by Ganapathy to study the effects of heterogeneity on both miscible and immiscible displacements. In both cases, the cores were characterized by a combination of air permeability measurements on each square centimeter of each face of rectangular slabs of sandstone, C.T. scans of cross-sections spaced one centimeter apart along the length of the cores ad by multiple water and oil tracers. Fine-grid simulations were then made using this characterization data ad comparisons made with the pressure, tracer and production data from these corefloods. The chemical flooding simulator developed at The University of Texas called UTCHEM was used for these simulations. An important factor that distinguishes this work from other results reported in the literature is the use of an outcrop sandstone rather than either artificially heterogeneous cores or small samples of reservoir cores. We present an integrated approach that includes experiments using an air minipermeameter, C.T. scanning, multiple tracer experiments in both water and oil, pressure data, oil recovery data and both stable and unstable polymer displacements in large samples of heterogeneous outcrop sandstone combined with both fine-mesh and coarse-mesh simulations. The use of this particular outcrop sandstone for these and other similar experiments has a number of advantages over the use of reservoir cores because we can readily obtain large samples and as may samples as needed with variable degrees of heterogeneity. It also has advantages over artificially heterogeneous permeable media because as a practical matter samples of consolidated sandstone with large and practical matter samples of consolidated sandstone with large and variable heterogeneity on a variety of scales are difficult and expensive to make and may still not minic nature in some ways we may desire. For example, in addition to the macroscopic flow characteristics that are included in this paper, results of how the residual oil saturation differs between waterflooding and polymerflooding can be found in Wreath.
Summary Dispersion and convection scaling can be used to model miscible and immiscible flows, respectively, but these scaling techniques fail in multidimensional heterogeneous permeability distributions. The length dependence observed for dispersion induced by correlated heterogeneity in miscible displacement processes precludes the definition of an effective dispersion coefficient. Effective relative permeabilities can be defined that will reproduce the results of the cross-sectional simulations from which they were derived, but complications arise when they are used in areal models. Introduction The study of reservoir heterogeneity and its influence on oil recovery has received more attention recently. The issues that arise in modeling flow inheterogeneous reservoirs can be placed in four categories.Geologic quantification: quantitative specification of fluid flow properties includes statistical measurements of variation and spatial correlation. Data integration: data from different sources and measured at different scales am combined at different degrees of reliability, resolution, and spatial sampling into a reservoir model consistent with all the available data and their spatial structures. Numerical rocks: simulations of geologic property distributions at arbitrary resolutions that reproduce measured data and statistical correlation measurements. Scale-up: definition of effective flow properties for coarse-grid, finite-difference simulations or streamtube scaling methods for field-scale performance predictions. This paper focuses on scale-up by reviewing the scaling behavior of miscible and immiscible displacements to determine the validity of defining effective flow properties to represent the effects of permeability heterogeneity on frontal displacements. We show that heterogeneity effects on miscible processes cannot be reduced to an effective dispersion coefficient or be modeled with an appropriate level of numerical dispersion. Heterogeneity effects on immiscible displacements can be captured in pseudoreladve-penneability functions, but their use entails many complications and restrictions that have not been discussed previously. These restrictions, which include maintaining the same cell previously. These restrictions, which include maintaining the same cell size and stream wise sequence and using the same initial saturation distribution as the cross section from which they were derived, will be shown. First, it is necessary to develop scaling laws for 1D displacement processes. These scaling laws form the basis for scaling the equivalent processes. These scaling laws form the basis for scaling the equivalent 1D form of the heterogeneous cross-sectional results. Scaling Laws In Multiphase Displacement The 1D transport equation for multicomponent, multiphase flow in a porousmedium is (1) (2) (3) (4) (5) (6) (7) Fi = fractional flux of Component i, which depends on the relative permeabilities and viscosities of Phase j, which, in turn, depend permeabilities and viscosities of Phase j, which, in turn, depend only on the local fluid compositions, Ci. Di = the dispersive flux of Component i, which depends on the dispersion coefficient of Component i in Phase j, Kij. This equation includes the simplifying assumptions that the fluids and rock are incompressible and isothermal and that the fluids flowing through the mediumare in local thermodynamic equilibrium and do not go through chemical reactions. In Eq. 1, the first term represents the local accumulation of Component i, the second term represents the net flux out of a local volume element byconvective fluid motion, and the right side represents the net flux into a volume element by dispersion in the phases. The scaling behavior of displacement fronts depends on the phases. The scaling behavior of displacement fronts depends on the forms and relative magnitudes of the flux terms, which contain the transport properties, ki and Ki. If we consider displacements in asemi-infinite medium, (i.e., the downstream boundary is far enough away that the boundary condition does not influence local displacement behavior), then the only length scale is the mean displacement distance after a time to, L0=vt0. The relative magnitudes of the two flux terms then are given by the Peclet number, defined as (8) When diffusion is negligible compared with mechanical mixing in the microscopic pore space, =, where = microscopic mixing length and (9) Measurements of or from outcrop sandstones show values typically greater than 0.5 cm, while measurements of in carbonates are more variable and typically are in the 0.1- to 2-cm range. For typical interwell distances(greater than 200 m), Np greater than 10,000. In simulations with finite-difference models, the truncation error effects introduce a numerical dispersion that, for small timesteps, scales as =, where N=number of gridblocks along the flow directions. This means that in most flow simulations, numerical dispersion effects will be larger than properly scaled physical dispersion. In defining effective flow properties to represent permeability heterogeneity effects, it must be understood that they can reproduce only the kinds of scaling behavior produced by the flux terms in Eq. 1.
Summary Permeability heterogeneity in porous media is evaluated with standard-pressure and production-history measurements from miscible displacements. Variations in the production history with mobility ratio and deviations of the pressure history from an idealized response provide complementary information on permeability heterogeneity. These heterogeneity tests have been permeability heterogeneity. These heterogeneity tests have been developed to assist in core sample selection for unsteady-state relative permeability tests and other transient displacement tests. Introduction Many laboratory tests used to characterize the ability of porous materials to transmit fluids involve transient displacement. During any displacement test, at least one recognizable feature, such as saturation or phase composition, is propagated through the porous medium in the direction of flow. The unsteady-state porous medium in the direction of flow. The unsteady-state relative permeability test is an example of a displacement test in which the saturation and effective viscosities are transient properties. A variety of EOR methods commonly are investigated properties. A variety of EOR methods commonly are investigated in the laboratory by use of displacement tests with transientsaturations and phase compositions. Displacement tests usually rely on the assumption of uniform permeability for measurement interpretation. Unsteady-state relative permeability tests on heterogeneous samples have been shown to violate the assumptions required by explicit interpretation methods, which lead to errors in the relative permeability analysis. Heterogeneity will be classified as transverse or longitudinal to facilitate the discussions in this paper. Transverse heterogeneity refers to permeability variations overcross sections orthogonal to the flow direction. Longitudinal heterogeneity refers to permeability variations that are strictly a function of position permeability variations that are strictly a function of position along the flow axis. Although heterogeneous cores generally will not fit into either classification, the miscible-displacement tests to be discussed have features that differentiate these two heterogeneity modes. Permeability Heterogeneity Characterization Permeability Heterogeneity Characterization Permeability heterogeneity characterization is not routine because Permeability heterogeneity characterization is not routine because of limited available experimental methods and difficulties in interpreting heterogeneity results. The most frequently used heterogeneity test has been the unit-mobility-ratio miscible-displacement test. The mixing of miscible fluids in 1D flow through homogeneous porous media is known to follow the dispersion equation (1) The hydrodynamic dispersion coefficient, K, is used to model the bulkeffective mixing rate for miscible fluids flowing through porous media. porousmedia. Capacitance effects in the production history from this type of test can be used to detect heterogeneity. Capacitance production behavior refers to along period of low-level production of the displaced phase after an early breakthrough of the injected phase. The 1D dispersion model (Eq. 1) cannot match production histories with capacitance effects, which generally have been modeled by including mass transfer between a flowing and nonflowing fraction of pore space. Although the unit-mobility-ratio miscible displacement determines heterogeneity when capacitance occurs, permeability heterogeneity does not always produce capacitance permeability heterogeneity does not always produce capacitance effects. Transverse heterogeneity may enhance longitudinal dispersion without demonstrating capacitance effects or any other deviations of the production history from a dispersion model. Longitudinal permeability heterogeneity will not produce a characteristic response from a unit-mobility-ratio miscible displacement because the production history is sensitive only to variations in porosity and dispersivity. Therefore, it is not always possible to identify permeability heterogeneity from a unit-mobility-ratio miscible-displacement test. The use of favorable-mobility-ratio miscible displacements provides a comparison test to the unit-mobility miscible provides a comparison test to the unit-mobility miscible displacement to aid in the diagnosis of heterogeneity. Mobility ratio values less than 1 are favorable because of good displacement efficiency. Giordano et al. studied favorable-mobility-ratio miscible displacements in aheterogeneous porous medium by use of numerical simulation of miscible displacement in a square, 2D random permeability field. Results from these calculations show sharpening of the breakthrough curve at the lower mobility ratios compared with the unit-mobility-ratio test. For the particular heterogeneity field used in these simulations, no differences were detected between the 0.1 mobility-ratio displacement and the 0.01 mobility-ratio displacement. The permeability field contains random permeability values drawn from a normal distribution where the permeability values drawn from a normal distribution where the ratio of the standard deviation to the mean is 0.5. The permeability is uniform over squares with sides of length 0.1 permeability is uniform over squares with sides of length 0.1 times the total length of the porous medium. Experiments that Sorbie et al. performed on a Clashach quarrysandstone give further evidence of reduced dispersion at lower mobility ratios. A favorable-mobility-ratio miscible slug test (M = 0.33) resulted in a dispersivity (for the leading edge) close to one-half the dispersivity found for a unit-mobility-ratio test. However, further decreases in the mobility ratio (M = 0.125) did not influence dispersivity. These results indicate that stable, miscible flow is sensitive to a certain range of the mobility ratio. Comparing a miscible displacement at unit mobility ratio with a miscible displacement at one or more lower mobility ratios provides a means to determine whether permeability heterogeneity enhances dispersion. The average effective dispersion for equal mobility fluids in a heterogeneous porous medium is a result of pore-level mechanisms (which create dispersion in a homogeneous porous medium) plus macroscopic deviations from 1D flow caused by permeability heterogeneity. A favorable mobility ratio can suppress deviations from 1D flow, thereby reducing effective dispersion. The sensitivity of the dispersivity to the mobility ratio has theoretical implications with regard to modeling favorable-mobility-ratio miscible displacement. The mobility ratio as used in this paper is strictly the endpoint mobility ratio. The sensitivity of the dispersion coefficient to the endpoint mobility ratio might suggest that the dispersion coefficient should be a function of the local mobility ratio (i.e., a function of the concentration). However, experimental results here and elsewhere indicate that the linear dispersion model (Eq. 1) is sufficient to match the production behavior. SPEFE P. 112
Summary Nuclear-magnetic-resonance (NMR) imaging was used to examine pore structures and liquid flow paths in rocks nondestructively. Various pore types were recognized in a spectrum of limestones, including growth-framework, moldic, vuggy, between- particle, and within-particle porosities. Pore connectivity was particle, and within-particle porosities. Pore connectivity was examined by fluid displacement and flow-weighted and 3D imaging. Introduction NMR signals from hydrogen in liquid molecules can be used t obtain information about the location, displacement, and flow o liquids in poresystems of rocks. The direct proportionality of the NMR pore systems of rocks. The direct proportionality of the NMR signal frequency to the magnetic fieldstrength is the basis of NMR imaging. The application of a magnetic fieldgradient across a liquid-saturated rock results in a range of NMR frequencies along the direction of the gradient. A series of gradients properly oriented in3D space gives rise to a spatially dependent distribution of NMR frequencies that can be processed by computer to yield an image of the pore system of the rock. Liquid flow during the measurement can modify the NMR signal, allowing flow-rate imaging, even during steady-state flow. Several characteristics make NMR imaging useful for examining geologic core samples. Because it is nondestructive, valuable material need not be destroyed and such multiple experiments as repetitive fluid flow through the core may be conducted with a single sample. In addition, the solid rock matrix does not interfere with measurement of fluids in the pores. Such fluid properties as now rate, diffusion coefficient, and chemical properties as now rate, diffusion coefficient, and chemical composition can be measured (e.g., oil can be distinguished from water). Finally, 3D NMR imaging 1.2 can measure 3D pore structures accurately. In contrast, methods that build a 3D structure from 2Dslices are subject to errors from nonuniform, overlapping, or missing slices, The only other technique available that directly provides 3D pore structures of a rock without these problems is 3D provides 3D pore structures of a rock without these problems is 3D X-ray microtomography. A potential limitation of NMR imaging is the difficulty in obtaining signals proportional to porosity when iron or clay is present in samples, causing the spin-spin relaxation time(T2) of present in samples, causing the spin-spin relaxation time (T2) of the saturating fluid to be short. This commonly is worse for sandstones than for limestones. Using fast switching and strong gradients in the spin-warp imaging methods or using the projection reconstruction method should overcome this limitation. The earliest geologic applications of NMR imaging were profiling water distribution in rocks and imaging bedding planes profiling water distribution in rocks and imaging bedding planes and clay seams in sandstones. Fluid displacements, flow rate, and relaxation time of geologic rock cores have been imaged. Of particular interest is the ability to distinguish oil from water in rocks. Recently, NMR imaging has been used for pore structures in rocks. This enables investigation of processes at the pore level that ultimately control the processes at the pore level that ultimately control the macroscopically observed phenomena. In this report, we first show several examples in which various types of pores in limestones are recognized in NMR images of water-saturated samples. Pore types are classified as the reservoir potential of a rock is described and this classification can help indicate the difficulty of producing oil from the rock. Next, three methods of using NMR imaging to examine pore connectivity are demonstrated. Connectivity is an aspect of pore structure that strongly influences fluid flow through reservoir rock; it is difficult to study by other means. Experimental Details The instrument used is a Bruker MSL-200 NMR spectrometer with mini-imaging and micro-imaging accessories. The capabilities of the mini-imaging equipment and the imaging parameters are a 4.7-T (200-MHZ) magnetic field, 150-mm magnet bore, 60-mT/m magnetic field gradient strength, 6.5- to 17.0-ms echo time, 80-s hard 180' pulse length. 0.1-mm resolution in the image plane and 0.3-mmresolution normal to that plane (slice thickness), and pixel dimensions of 128ร 128 to 512 ร 512. The microimages were obtained at the same field, but with200-mT/m gradient strength, 2. - to 3.4- ms echo time, 7- s hard 180 pulse length, and 0.05-mm isotropic 3D resolution. The resolution, data acquisition time, and other experimental parameters can be varied to fit the requirements of the investigation. Most rocks examined in this study are limestones cut into25- mm-diameter cylinders 25 to 38 mm long. The microimages were of 4.5 x4.5-mm samples. All the 2D images presented here are slices orthogonal to the cylinder axis and therefore are round in cross section. The terms 2D and 3Dimaging refer to generating and collecting NMR signals. not to the signal display. The 2D spin-warp imaging method, usually with a hard 180' pulse, was used to obtain all the 2D images. The 3D spin-warp imaging method was used to obtain the 3D images. Two methods are available to display 3D arrays of imagedata (sometimes called volume image data), regardless of the generation method(from 3D imaging or from a 3D reconstruction of stacked 2D images). In one method, computed slices through a 3D array are displayed the same way as 2Ddata. It is possible to "step through" a sequence of slices generated from a 3D array of image data in movie like fashion. In addition, 3D arrays of data can be displayed in a rendering that resembles a view of a pore cast with the added flexibility of deleting selecting pores. The method used here to display 3D NMR image data is similar to that described by Herman et al. A cube is drawn at the coordinates of each volume element that contains more than 50%water. Cube faces that would be hidden from the viewer if the cube were opaque are not drawn. Visible faces are displayed with an intensity as if they reflected light diffusely from a point source oriented about 45 from the direction of the viewer and as if they emitted light about 25 % as intense as the point source. Views for selected rotation angles of the sample are point source. Views for selected rotation angles of the sample are generated. Various water saturations, lighting and viewing angles, and light intensities are possible. A computer program was written to define a porous network by identifying adjacent voxels containing more than 50% water. This method allows further examination of pore structures by enabling selecting parts of a poresystem to be displayed. Advanced display programs can show pore system to be displayed. Advanced display programs can show part of the pore structure in a translucent, opaque, or transparent part of the pore structure in a translucent, opaque, or transparent appearance. NMR Imaging of Various Pore Types In Limestones Pores in rocks are classified according to their sizes, shapes, and Pores in rocks are classified according to their sizes, shapes, and modes of formation. They often are classified while the geologic description of the area is generated and can help indicate the difficulty of producing oil from the rock. Choquette and Pray's classification of pore types and Dunham's classification of limestones as modified by Jordan were used. Several examples of NMR imaging of pore types found in coarse-grained limestones from subsurface reservoirs and surface outcrops is discussed below. The quality of the NMR images allows various pore types to be recognized in the images. SPEFE P. 123
- North America > United States (0.68)
- Asia > Middle East > Jordan (0.25)
Summary We present visual cell and coreflood data pertaining to the influence ofwater saturation on miscible gasflood residual oil (water blocking). In visualcell studies, we equilibrate miscible gases and oils at reservoir conditionswith an intervening water phase. We find that the solvent-induced deposition ofseparated phase. We find that the solvent-induced deposition of separatedsolids causes initially water-wet quartz surfaces to become oil-wet. A numberof miscible coreflood displacements at high water saturation are considered, and the extent of water blocking is determined for Berea sandstone andreservoir cores. Displacements of fluid systems with wettability alteration andthose of refined oil systems without wettability alteration are reported. Wefind that water-blocking measurements with pure hydrocarbons overestimatemiscible flood Sor regardless of the porous medium (Berea or reservoirsandstone or a reservoir porous medium (Berea or reservoir sandstone or areservoir carbonate) and the flood (CO2 or hydrocarbon miscible). We furtherfind that water blocking is significantly more severe in Berea sandstone thanin reservoir rock. We assert that water-blocking measurements with refined oilsystems in Berea sandstone greatly overestimate water blocking in reservoirs, partially because refined oils cannot reproduce miscible-flood-inducedwettability alteration. Introduction The large viscosity contrast between injected miscible gas and inplacereservoir fluids leads to unstable displacements. The simplest method ofcontrolling this poor mobility ratio is to inject the miscible gas alternatelywith water - i.e., water-alternating-gas (WAG) injection. The WAG processimproves the volumetric sweep efficiency of the miscible flood but results in awater saturation that is high and mobile. Literature data show that, inwater-wet porous media, a high water saturation decreases the efficiency of amiscible flood by shielding remaining oil from injected solvent, a phenomenonsometimes called water blocking. Data also show that, in an oil- or mixed-wetmedium essentially all oil is accessible to the flood, even at very high watersaturations. The wettability of a porous medium could be altered during amiscible flood. Wettability alteration has been implicated as a contributor toa low residual oil for a high WAG flood in an originally water-wet porousmedium. When crude oils are mixed with solvents, suspended asphaltenes can bedestabilized. These highly polar molecules are surface-active and could adhereto the pore walls of a porous medium and change its wettability. pore walls ofa porous medium and change its wettability. For this study, we look at whathappens if a miscible flood causes the wettability to change from water- tooil-wet, focusing on (1) the effect that high water saturation has on themiscible flood recovery, (2) the extent to which water blocking is diminished, and (3) whether the initial or final wettability determines the extent of waterblocking. Our goal is to determine how severe water blocking affects oilrecovery in the field. Experimental Approach Our approach to the water-blocking problem is experimental. We used acapillary tube visual cell (CTVC) to study pore-level fluid contacts atreservoir conditions. Our objectives were to observe oil swelling and tomonitor the oil/water contact (OWC) angle. For some of the fluid systemsstudied in the CTVC, we performed coreflood displacements, again at reservoirconditions. In addition, we measured the extent of water blocking in reservoircores during refined oil displacements. Visual Cell Studies. The CTVC is a simple model of the pore-level contactsthat occur between injected gas and waterflood residual oil in water-wet porousmedia. In a water-wet medium, waterflood residual oil exists as isolated dropscompletely surrounded by water. Injected solvent does not contact oil directlybut initially is separated from the oil by a water film. Diffusion of solventthrough the water film is the only route of oil/solvent contact. In the CTVC, we experimentally reproduce this water-mediated oil/solvent contact in aquartz, thick-walled capillary tube (0.7 to 1.5 mm ID) (Fig. 1). We introduce abubble of solvent into the brine-filled tube and then inject oil into the tubeto obtain an oil droplet 1 cm from the near end of the solvent bubble. Thesephases are held at test pressure and reservoir temperature for a period rangingfrom 3 days to 1 month. We studied live and stock-tank oils and contacted oilwith both CO2 and propane. During a test, we monitor oil-phase volume, solidformation, and oil/water wettability. We determine phase volume by measuringphase length because the tube diameter is constant. The system is phase lengthbecause the tube diameter is constant. The system is valved so that we can movethe droplets carefully back and forth within the tube. We use these motions todetermine whether any material is adhering to the quartz tube and to monitorthe wettability of the tube where the oil has contacted it. Wettability ischaracterized with a contact angle measurement. The meniscus shape, however, isdistorted by the tube, which acts as a cylindrical lens. The distortion isparticularly severe away from the tube axis and interferes with directobservation of the contact angle. To measure the contact angle, we use thegeometrical relationship between the meniscus height, the tube radius, and thecontact angle, assuming that the meniscus is a spherical cap (Fig. 2). Then, the contact angle, c, is determined by measuring the meniscus height, h, and byuse of the following equation: (1) Because the interfaces can be moved with great control, both receding andadvancing contact angle measurements can be made. All contact angles reportedwere measured through the water phase, so angles less than 90 degrees arewater-wet and angles phase, so angles less than 90 degrees are water-wet andangles greater than 90 degrees are oil-wet. In addition to measurements of thecontact angle, qualitative wettability data are determined easily from themeniscus curvature. Displacement Experiments. To enable us to understand the effect ofwettability alteration on laboratory and field displacements, we performedseveral coreflood displacements. We used composite cores and reservoir fluidsto perform miscible WAG floods at a range of WAG ratios. To improve confidencein our understanding of the effect of wettability alteration on displacementefficiency, we also performed displacements with refined oils. Refined oilscannot alter wettability as substantially as crude oils because they lack thehigh-molecular-weight polar species principally responsible for surfaceactivity. The refined oil displacements are steady-state measurements where wedetermine the saturation of oil trapped by a WAG flood with afirst-contact-miscible solvent. We use n-tridecane (n-C13) as the oil to bedisplaced and n-tetradecane (n-C14) as the solvent. SPEFE P. 167
- North America > United States > West Virginia (0.66)
- North America > United States > Pennsylvania (0.66)
- North America > United States > Ohio (0.66)
- North America > United States > Kentucky (0.66)
Summary This paper evaluates the accuracy of log-inject-log (LIL) tests, analyzesthe fluid-displacement behavior around the wellbore, discusses the transientbehavior of pulsed-neutron-capture (PNC) logs before steady state is reached, and establishes the conditions under which LIL tests work best. This papershows that both completion and reservoir properties affect the accuracy of LILtests. Introduction Accurate methods for measuring reservoir oil saturations (ROS's) areessential for determining remaining reserves. Many methods currently are usedto evaluate oil saturations. These methods include drilling and coring a newwell, openhole logs, cased-hole logs, single-well tracer tests (SWTT's), and LIL tests. LIL tests generally are recommended as one of the best methods forestimating oil saturation in existing wellbores. The LIL technique evolved fromconventional cased-hole-logging methods for estimating oil saturations. Thecased-hole PNC log eliminates the casing and cement response and measures onlythe formation capture cross section. The formation capture cross sectiondepends on three unknowns: rock capture cross section, porosity, and oilsaturation. The major uncertainty that remains in porosity, and oil saturation. The major uncertainty that remains in conventional PNC logging is the reservoirrock capture cross section, which can vary over short distances. In addition, certain minerals, such as clays, can have an inordinately large effect on therock matrix capture cross section. The best method for getting around thisdifficulty is to conduct LIL tests in which a baseline log is run first, theresident water is replaced completely by water of a different salinity, and thewell is relogged. By subtracting the two log responses, we eliminate thecontribution from the rock matrix, which allows the fluid properties to bemeasured unequivocally. The ROS in an LIL test can be calculated from (1) The LIL technique has some distinct advantages over otheroil-saturation-measurement methods: the technique is fast and the test can becompleted within I to 2 days, rock mineralogy does not affect the results, andthe technique gives a foot-by-foot distribution of oil saturation within azone. However, the porosity distribution within the zone must be knownaccurately, and porosity distribution within the zone must be known accurately, and the resident brine must be replaced completely by the injected brine. Themotivation for performing an in-depth analysis of fluid displacement in LILtests was that oil saturations measured in some cases by LIL tests were notconsistent with other oil saturation measurements. Two major concerns were thatthe oil saturations measured by the LIL test were consistently higher thanthose measured by other tests and that LIL tests run in wellbores completed inmultiple zones gave comparable results in some zones and in only parts of wholezones in other cases. Even if a single zone was isolated before performing thetest, the LIL again gave reliable results in only part of the whole zone. Incomplete displacement of the resident brine by the injected brine is believedto cause these problems. Incomplete displacement of the resident brine by theinjected brine can lead to erroneously high estimates of ROS. In case ofincomplete displacement, the term (-) in Eq. 1 is lower than the correctvalue, and the ROS is overestimated. In essence, the fraction of resident brinenot displaced by the injected brine is taken to be an additional hydrocarbonfraction. Reedy suggested the most commonly used model for dealing with theincomplete displacement of resident water by injected water. In addition to thedisplacement of the resident brine by the injected brine, his method involvesone more displacement step in which the injected brine is displaced by theresident brine and the well is relogged. The major assumption in the Reedymodel is that the fluid-displacement efficiencies and the correspondinglog-response efficiencies in the two injection steps are equal if the fluidquantities injected at each step are equal. Although the Reedy model is a goodstarting point for understanding fluid displacement, it has two importantshortcomings. First, fluid flow around perforations is complex, and theinjection of equal volumes does not necessity imply equal displacementefficiencies. Second, the PNC log response can vary significantly betweenhigh-salinity brine displacing low-salinity brine, and conversely, betweenlow-salinity brine displacing high-salinity brine. Both factors must beaccounted for when interpreting LIL tests. The major objectives of this paperare to investigate the fluid-flow behavior around perforations in theformation, to quantify the fluid-displacement efficiency as brine salinitychanges, to examine the transient behavior of the PNC logs before steady stateis reached, and to provide recommendations for designing and executing future LIL tests. This paper presents a general viewpoint for extracting all usefulinformation from logs run during an LIL, especially those run during thetransient period when log responses are changing. This paper analyzes LIL testsrun in the South Cuyama field and uses these logs as benchmarks for developingand illustrating underlying concepts in LIL testing. LIL Model The fluid-flow model and the PNC log-response model comprise the model forthe LIL process. This paper concentrates on the fluid-flow model andincorporates the PNC log-response models developed by well logging companies. The wellbore and near-wellbore geometries are so complex that simplifiedgeometries and numerical solution methods are necessary to get meaningfulresults. Fluid-Flow Model. Fluid-Flow Numerical Model. The trend in numericalsimulation of fluid flow around perforations has been to use increasinglysophisticated finite-element models, which have been shown to solve the flowequations around perforations very accurately, especially for turbulent gasflows. Improvements for laminar liquid flows are marginal, and trendsdocumented in earlier finite-difference and finite-element studies remainreasonably accurate. Finite-element models were not used in this study becauseof their increased complexity, especially in modeling multicomponent fluidflow. A three-phase, compositional, finite-difference simulator was used, whichincorporated longitudinal and transverse dispersion, convection, and 3Dgeometry. Fluid-Flow Model Parameters. Parameters that influence near- wellborefluid flow are perforation shape, crushed zones, shot density, and reservoirheterogeneities. The perforation shape is characterized by the perforationdiameter and length. Usually, shot density and perforation diameter are knownwith a high degree of certainty, but the perforation length and the shape ofthe crushed zone are difficult to ascertain for each perforation. The 3Dgeometry of the perforations is approximated by a 2D Cartesian model (Fig. 1). The perforation is modeled as a thin slit that extends in the z dimension, andthe reservoir also extends in the z direction. Thus, the simulation area shownin Fig. 1 represents the basic symmetry element for fluid flow. SPEFE P. 157
- Geology > Rock Type > Sedimentary Rock > Clastic Rock (0.95)
- Geology > Geological Subdiscipline (0.66)
- North America > United States > California > Santa Barbara Basin > South Cuyama Field (0.99)
- North America > United States > California > Sacramento Basin > 4 Formation (0.99)
- North America > United States > California > Sacramento Basin > 3 Formation (0.99)
- North America > United States > California > Sacramento Basin > 2 Formation (0.99)