This paper documents the formation of natural fractures in the Horn River Group, a major Canadian shale gas play, and addresses relationships between natural-fracture development and rock-mechanical properties derived from cores and well logs. Most natural fractures in the Horn River Shale are narrow vertical fractures, sealed with carbonate minerals. In this study, the formation of observed fractures is primarily determined by a lithology type, mineral composition, and rock-mechanical properties at the timing of fracturing.
Brittleness is an important geomechanical property controlling the formation of fractures, because brittle shale is more easily fractured than ductile shale, and fractures in brittle shale tend to persist when the fracturing pressure is released. In this study, a hardness value measured by a commercial hardness tester is found to be a good proxy for the brittleness of shale layers. On the basis of a statistical analysis, the threshold values of both hardness and brittleness are estimated to predict the distribution of natural fractures, assuming that the mechanical properties of the host rock were relatively stable from at least the time at which fractures formed. Hardness values are shown to be more reliable than brittleness.
Kuznetsov, Oleksandr (Baker Hughes, A GE Company) | Agrawal, Devesh (Baker Hughes, A GE Company) | Suresh, Radhika (Baker Hughes, A GE Company) | Feng, Xianhua (Baker Hughes, A GE Company) | Behles, Jackie (Baker Hughes, A GE Company) | Khabashesku, Valery (Baker Hughes, A GE Company)
Oil-sand ore flotation is a primary method of bitumen recovery from mined Athabasca tar sands. In bitumen flotation, suspended biwettable ore fines, such as clays, tend to migrate to oil/water interfaces, creating a slime coating on liberated bitumen droplets. Slime coating significantly reduces the efficiency of the flotation process and overall oil recovery. Ultradispersed hydrophilic silica nanoparticles were found to stabilize biwettable ore fines in an aqueous phase by adsorbing onto fines surfaces, even at concentrations as low as 50 ppm. As a result, fine solids move away from oil/water interfaces, reducing the slime coating and increasing bitumen recovery during flotation of low-grade ore by more than 5%. The addition of nanoparticles has no negative effect on froth quality or oil, water, and solid separation in naphthenic and paraffinic froth-treatment processes. The study demonstrated that colloidal nanoparticles affect many stages of the bitumen-extraction process—from bitumen separation to clay-wetability alteration.
Nuclear-magnetic-resonance (NMR) measurements have been extensively used for determining porosity, pore-size distribution, and fluid saturation in porous media. However, internal gradients of the magnetic field generated by the presence of paramagnetic or diamagnetic centers such as shale or clay particles can significantly affect NMR response. Consequently, the resulting interpretation for pore-size distribution and porosity is also affected. In this paper, we quantify the effect of internal magnetic-field gradients and spatial distribution of matrix components such as clay minerals on NMR response using pore-scale NMR numerical simulations. We also quantify the influence of the aforementioned parameters on the NMR-based evaluation of porous media.
We used the finite-volume method to numerically solve the Bloch (1946) equations and simulated magnetization decay in porous media. We cross validated the reliability of numerical simulations using analytical solutions given for spherical pores in different diffusion regimes. The model was then used for simulation of NMR response in the pore-scale images of sandstone and carbonate rocks. We used Larmor frequency of 2 MHz, external magnetic-field gradient of 0.10 T/m, and half-echo spacing time of 0.5 ms for simulating NMR response in pore-scale images of sandstones and carbonates. We then developed synthetic cases using actual rock images covering a wide range of spatial distribution of clay minerals (i.e., paramagnetic centers) to quantify the sensitivity of NMR decay to internal magnetic-field gradients. We quantified the sensitivity of NMR response for distribution of clays as thin laminae in the rock and as thin layers on the surface of the grains.
The results showed that at low concentration (0.3 to 0.7%) of dispersed clay, there is a negligible effect of internal magnetic-field gradients on magnetization decay. At higher concentration of dispersed clay (5.1 to 7.3%), we observed a significant effect of internal magnetic-field gradients on magnetization decay. The presence of shale minerals can cause 53% variation in the location of the transverse-relaxation-time constant (T2) and up to 67% relative error in the assessment of dominant pore sizes. Shale laminations containing clay were found to produce an effect of up to 31.8% on T2 relaxation-time constant, which could cause a relative error of 50.0% in estimates of dominant pore size in the rock.
The outcomes of this paper demonstrated the effect of heterogeneous rock mineralogy on NMR response. The effect of internal magnetic-field gradients generated by shale and clay on NMR becomes relevant when shale and clay particles are close to the pore fluid and their magnetic field starts to affect the distribution of magnetic field in the pore space. The results reveal the importance of characterizing the distribution of shale and clay minerals before interpreting NMR response and can potentially improve conventional techniques of pore-network characterization (pore-size distribution and pore volume) in the presence of clay minerals where internal magnetic-field gradients are not negligible.
Steam-assisted gravity drainage (SAGD) has been extensively applied in thermal recovery from oil sands reservoirs in the Athabasca region of Northern Alberta, Canada. As the steam chambers associated with SAGD well pairs become mature, a form of abandonment is often applied that may include pressure maintenance in the depleted zone. Quantification of potential surface subsidence associated with SAGD abandonment becomes critical especially when the mature wells are in proximity to future developments. In addition,induced shear stresses should be estimated to fulfill well-integrity requirements. In the context of this case study, first, the development of a static geomechanical model (SGM) derived from a fine-tuned geomodel realization is discussed, which forms the basis for the iteratively coupled simulation model. The calibration work flow of the coupled reservoir/geomechanical simulation model to historical heave data is then reviewed and the effects of different parameters on calibration quality are investigated. Finally, the estimation of subsidence and the induced shear stresses in the nearby wells are discussed, and the magnitude of residual heave is quantified. The results of this study show that only a fraction (up to 38%) of surface heave is reversible (in form of subsidence) during the abandonment phase. Therefore, the magnitude of the surface subsidence and the associated shear stresses are small. The modeling study has also shown that a small magnitude of subsidence may be recorded even 10 years after abandonment. However, more than 50% of the surface subsidence is observed in the first 2 years after abandonment. Other important findings of this study include documenting the effects of thief-zone interaction and pseudoundrained loading as they relate to irreversibility of surface heave; documenting the effects of various geomechanical parameters on the quality of calibration against the historical heave data; observation of the relative effects of the isotropic unloading, thermal expansion, and shear dilatancy on the magnitude of heave; and quantification of incremental, yet small, shear stresses along the nearby horizontal wells.
Gogri, Maulin Pankaj (University of Oklahoma) | Rohleder, Joseph (University of Oklahoma) | Kabir, Shah (University of Houston) | Pranter, Matthew (University of Oklahoma) | Reza, Zulfiquar (University of Oklahoma)
Oklahoma has been at center stage of induced seismicity. Water-disposal activities have been associated with triggering the increasing number of seismic events. The objective of the study is to provide a simple diagnostics method and procedure for safe water-disposal operations. A comprehensive suite of scenarios and parameters has been analyzed that affect water disposal. On the basis of this study, prognosis will lead to safe water-disposal operation without the adverse effect.
A suite of reservoir models involving water injection helped understand disposal-well performance. The well operational limits correspond to disposal-zone fracture gradient. The modified Hall analysis is used to ascertain the point of departure from normal injection behavior. Limiting cumulative injected volumes are determined and investigated for various scenarios from simple to increasingly complex subsurface conditions. This investigation includes studying the effects of disposal-zone porosity, compartment size, conductivity, formation compressibility, heterogeneity, and natural fractures. In addition, we explored the effects of communication with overlying producing zone, communication through completion anomaly, seal integrity, and fluid complexities.
This study illuminates an overall understanding of disposal-well performance through various scenario analyses. A relationship between disposal-zone fracture gradient and limiting cumulative injection volume is established. For a fracture gradient of 0.7 psi/ft, this limiting pore-volume (PV) injection is less than 2%, which corresponds well with the conventional wisdom learned from carbon dioxide (CO2) injection-well performance. The relationship of disposal-zone compartment size, established with rate-transient analysis (RTA), with limiting cumulative injection volume is formulated. Analyses from the various statistical design of experiments (DoEs) reveal the important variables that may affect disposal-well performance. The disposal-well operation can be devised in real time withthe modified Hall analysis that can reveal the departure from normal injection-well behavior. Factors accentuating the departure from normal behavior include disposal-zone porosity, formation compressibility, and seal integrity. Situations in which pressure release through leaks or communication with an adjacent formation takes place will naturally accommodate a larger volume of disposal water. Also, we learned that limiting cumulative injection volume and not injection rate (assuming injection pressure gradient is less than the fracture gradient) triggers a departure from normal injection behavior.
Using a suite of numerical reservoir models and the reservoir-monitoring tools involving modified. Hall analysis and RTA led to a comprehensive understanding of disposal-well performance. This study found a relationship of fracture gradient with limiting cumulative injection volume, and identified key variables affecting the disposal-well behavior. These findings led to a smart and safe disposal-well monitoring scheme, which will help disposal-well management become more economical and environmentally friendly.
Correction to Eq. 14 on page 278 of SPE-183640-PA; this equation supersedes the equation in the original published paper.
Large volumes of oil are being produced by waterflooding heterogeneous reservoirs. Careful flood-pattern design, well placement, and control are required to maximize oil recovery by delaying water breakthrough and optimizing sweep efficiency. Models that analyze the waterflood’s performance and predict the production forecast, such as the Dykstra-Parsons (DP) method, are routinely used for this purpose.
The DP method estimates the vertical-flooding efficiency between conventional wells producing from noncommunicating layers. This method and its various modifications have had a significant impact on the development of the theoretical description of the waterflooding process. The DP method is still routinely used for waterflood-performance prediction and analysis, flood-pattern selection, and recovery-factor calculation.
Advanced well completions (AWCs) control the fluid flow at the reservoir sandface. They have become a proven, widely used technology (particularly in waterflooded reservoirs) for modifying a production or injection well’s inflow/outflow rate profile along the well. In addition to this, new AWC designs that react to water breakthrough have recently become available. Incorporating a description of the AWC performance into the waterflood-analysis models will allow fast forecasting of the production profile and oil recovery, as well as help in optimizing the AWC configuration and control at the well-design stage.
This work extends the DP method for rapid prediction of the waterflood’s performance to AWC wells. It provides a simple means of estimating the additional, long-term value derived by controlling zonal flow rate using AWCs or any other means (e.g., well workover). The accuracy of the extended DP method’s prediction has been verified by comparison against the results from a numerical reservoir simulator. Several examples illustrate the extended DP model’s performance and value. The method’s limitations and possible future extensions are also discussed.
The presented model is a simple, transparent approach to evaluating the impact on the waterflood’s oil recovery efficiency (RE) of various well-completion and control options. It can be implemented as an analytical model or as a fast simulator. This model is also the missing link between the various AWC design methods available today and the AWCs’ long-term value evaluation.
AlKaaoud, Hassan (KOC) | Singh, Bharat (Kuwait Oil Company) | Marston, David (Golder Associates Ltd.) | McQuaid, James (Suncor Energy Inc) | Devon, John (Devon Mining Geology Consulting LLC) | Preene, Martin (Preene Groundwater Consulting) | Hornbrook, John (DeGolyer and MacNaughton Inc) | Pope, Gary (University of Texas at Austin)
Surface mining of hydrocarbon deposits is not a new technique, but its application has been mainly limited to the excavation of oil sands in the Athabasca Basin of Alberta, Canada, where the mining method has proved to be commercially successful, although in a narrow set of geological and environmental conditions. This paper discusses the scope for a broader application of a surface-mining approach and builds on the results of a conceptual study that examined the possibility of surface mining the viscous crude oil of the Ratqa Lower Fars (RQLF) reservoir in northern Kuwait. The study findings indicate that a large rate of crude oil might be profitably and sustainably produced for many decades through a surface-mining approach.
Giant reservoirs such as Lula (Santos Oil Basin, Brazil) and Ghawar (Saudi Arabia) have high-permeability intervals, known as super-k zones, associated with thin layers. Modeling these small-scale flow features in large-scale simulation models is difficult. Current methods are limited by high computational costs or simplifications that mismatch the representation of these features in simulation gridblocks. This work has two purposes: present an upscaling work flow to integrate highly laminated or interbedded reservoirs with thin, highly permeable layers in reservoir simulations through a combination of an explicit modeling of super-k layers using the Parsons (1966) formula and dual-medium flow models, and compare this method with two conventional upscaling approaches that are available in commercial software.
We use the benchmark model UNISIM-II-R (Correia et al. 2015a), a fine single-porosity grid dependent on field information from the Brazilian presalt and Ghawar oil fields, as the reference solution to compare the upscaling matching between the three methods. We compare oil recovery factor (ORF), water cut (WC), average reservoir pressure (RP), water front, and the time consumption for simulation. Our proposed Parson’s dual-medium (PDP) methodology achieved better upscaling matches with the reference model and had minimal time consumption compared with the representation of super-k layers through an implicit matrix modeling by single-porosity flow models (IMP) and through the explicit representation of super-k zones in the fracture system of dual-medium flow models (DFNDP).
Abbaszadeh, Maghsood (Innovative Petrotech Solutions) | Varavei, Abdoljalil (Innovative Petrotech Solutions) | Rodriguez-de la Garza, Fernando (Pemex E&P) | Villavicencio, Antonio Enrique (Pemex E&P) | Lopez Salinas, Jose (Rice University) | Puerto, Maura C. (Rice University) | Hirasaki, George (Rice University) | Miller, Clarence A. (Rice University)
An integrated methodology is presented for the development of a comprehensive empirical foam model based on tailored laboratory tests and representative numerical simulations that encompass processes of foam generation, coalescence, and shear thinning along with rheological characteristics and associated flow regimes. Steady-state and unsteady-state laboratory experiments of foam floods in a vertical column of sandpack with and without oil at different surfactant concentrations and at varied gas/surfactant-solution injection rates are designed, conducted, and analyzed. The logic and basis of these experiments are provided. Test results from experiments in the presence of oil provide information on the oil-induced foam/lamella coalescence functions. Unsteady-state experiments capture foam-generation and foam-dry-out phenomena, whereas steady-state experiments capture the effects of foam quality, foam velocity, and surfactant concentration. Process-based numerical simulations of these experiments are combined with basic governing analytical relationships of foam flow to provide a methodology for a comprehensive empirical foam model and to uniquely define the model parameters to preserve consistency with simulations of foam-flow processes. A procedure is presented to fully model the effect of surfactant concentration on foam strength and to quantify all concentration-function parameters, and, in particular, epsurf.