No abstract

Field studies and numerical simulations suggest that solute transport in fractured rock occurs primarily along preferred pathways or channels that involve only a small percentage of the available fracture network. By examining relationships between path length and fluid velocity in a series of discrete network simulations, we conclude that mass tends to travel longer distances in fractures in which the fluid velocity is higher, although the longest paths traveled do not correspond to the highest fluid velocities within the network. These observations suggest that a velocity optimization process, operating at fracture intersections, causes mass channeling at the network scale. In networks composed of fractures of varying aperture, this mechanism explains the increasingly tortuose paths traveled as mass moves selectively through the larger apertures.

This paper describes the development of a discrete feature network (DFN)model for the South Oregon Basin field in the Big Horn Basin of Wyoming. This DFN model is being developed to support well placement and gel treatment in order to recover previously bypassed oil in this highly heterogeneous, structurally controlled carbonate reservoir. The DFN model developed in this paper is based on data from outcrop studies, fracture and lithologic data from wells and lineament maps from 3D seismic. The model is calibrated against tracer test results. Introduction The South Oregon Basin field in the Big Horn Basin of Wyoming, currently operated by Marathon Oil Co., was discovered in 1912 (Fig. 1). Thisfield has produced over 83 million bbl of oil from the Permo-TriassicPhosphoria Formation, a thin marine limestone. This unit has moderate matrix permeability, but the significant structural deformation to which the reservoir units have been subjected has produced good large-scale fracture connection. However, local smaller-scale fracture connections appear to be variable and often results in poor fracture permeability. The field has been under waterflood since the 1960's. Because of the significant heterogeneity in reservoir permeability, current production suffers from very high >95%) water cuts, while the oil saturation remains high (upto 80%). Several options are being considered for improving recovery from this reservoir that require better knowledge of the fracture system. These options include improved targeting of water injection for waterflooding, optimal horizontal drilling to better connect and link together lower recovery portions of the reservoir, and the selective reduction in water cycling through improved gel conformance treatment design and placement. As part of a DOE-funded study, the discrete fracture network (DFN) technique is being applied to optimize well placement and gel treatment in order to recover previously bypassed oil. As the basis for the fracture study, four independent data sets were analyzed: outcrops at Wind River Canyon and Zeisman Dome, fracture and lithologic data from field wells, lineament maps from 3D seismic, and tracer tests. Individually, each data set confirms that fractures dominate the fluidflow in the reservoir rocks. Collectively, these data sets were used to construct a consistent, calibrated DFN reservoir model. The DFN approach is well-suited for these goals, since it creates realistic3D models of the fracture "plumbing" in the reservoir, can be calibrated against a wide variety of production tests and data, and can be used to carryout numerical simulations of flow and transport. DFN models also integrate data from different scales, including individual wellbore-scale data that is treated stochastically, and reservoir-scale, deterministic fault models derived from seismic and outcrop data. The flow parameters of the DFN model were calibrated using breakthrough times and concentrations from tracer tests. The final DFN model, conditioned to the structural geology, lithology, production, and tracertest data from this complex field, quantifies fracture intensity, surface area, volume, permeability, and connectivity. This calibrated DFN model is being used to support optimization of well placements and gel treatment. DFN Model Implementation The DFN model implemented for the South Oregon Basin is illustrated inFig. 2. This model was derived by integration of structural geological and hydraulic data as described below. The structural information is synthesized to obtain parameters for thespatial model,orientation distribution,size, andintensity of the natural fractures.

This report describes the results made in fulfillment of contract DE-FG26-00BC15190, "3-D Reservoir and Stochastic Fracture Network Modeling for Enhanced Oil Recovery, Circle Ridge Phosphoria/Tensleep Reservoir, Wind River Reservation, Arapaho and Shoshone Tribes, Wyoming". The goal of this project is to improve the recovery of oil from the Tensleep and Phosphoria Formations in Circle Ridge Oilfield, located on the Wind River Reservation in Wyoming, through an innovative integration of matrix characterization, structural reconstruction, and the characterization of the fracturing in the reservoir through the use of discrete fracture network models.Fields in which natural fractures dominate reservoir permeability, such as the Circle Ridge Field, often experience sub-optimal recovery when recovery processes are designed and implemented that do not take advantage of the fracture systems. For example, a conventional waterflood in a main structural block of the Field was implemented and later suspended due to unattractive results. It is estimated that somewhere less than 20% of the OOIP in the Circle Ridge Field have been recovered after more than 50 years' production.Marathon Oil Company identified the Circle Ridge Field as an attractive candidate for several advanced IOR processes that explicitly take advantage of the natural fracture system.These processes require knowledge of the distribution of matrix porosity, permeability and oil saturations; and understanding of where fracturing is likely to be welldeveloped or poorly developed; how the fracturing may compartmentalize the reservoir; and how smaller, relatively untested subthrust fault blocks may be connected to the main overthrust block.For this reason, the project focused on improving knowledge of the matrix properties, the fault block architecture and to develop a model that could be used to predict fracture intensity, orientation and fluid flow/connectivity properties.Knowledge of matrix properties was greatly extended by calibrating wireline logs from 113 wells with incomplete or older-vintage logging suites to wells with a full suite of modern logs. The model for the fault block architecture was derived by 3D palinspastic reconstruction. This involved field work to construct three new cross-sections at key areas in the Field; creation of horizon and fault surface maps from well penetrations and tops; and numerical modeling to derive the geometry, chronology, fault movement and folding history of the Field through a 3D restoration of the reservoir units to their original undeformed state. The methodology for predicting fracture intensity and orientation variations throughout the Field was accomplished by gathering outcrop and subsurface image log fracture data, and comparing it to the strain field produced by the various folding and faulting events determined through the 3D palinspastic reconstruction. It was found that the strains produced during the initial folding of the Tensleep and Phosphoria Formations corresponded well without both the orien...

No abstract