Date of Graduation
2016
Document Type
Dissertation
Degree Type
PhD
College
Eberly College of Arts and Sciences
Department
Geology and Geography
Committee Chair
Thomas H Wilson
Committee Co-Chair
Alan Brown
Committee Member
Timothy Carr
Committee Member
Shahab Mohaghegh
Committee Member
Jaime Toro
Abstract
The Tensleep oil reservoir at Teapot Dome, Wyoming, USA, is a naturally fractured tight sandstone reservoir that has been considered for carbon-dioxide enhanced oil recovery (CO2-EOR) and sequestration. CO2-EOR analysis requires a thorough understanding of the Tensleep fracture network. Wireline image logs from the field suggest that the reservoir fracture network is dominated by early formed structural hinge oblique fractures with interconnectivity enhanced by hinge parallel and hinge perpendicular fracture sets. Available post stack 3D seismic data are used to generate a seismic fracture intensity attribute for the reservoir fracture network. The resulting seismic fracture intensity is qualitatively correlated to the field production history. Wells located on hinge-oblique discontinuities are more productive than other wells in the field. We use Oda's method to upscale the fracture permeabilities in the discrete fracture network for use in a dual porosity fluid flow simulator. We analytically show that Oda's method is sensitive to the grid orientation relative to fracture set strike. Results show that the calculated permeability tensors have maximum geometric mean for the non-zero permeability components (kxx,kyy,kzz,kxy) when the dominant fracture set cuts diagonally through the grid cell at 45° relative to the grid cell principal directions (i,j). The geometric mean of the permeability tensor components falls to a minimum when the dominant fracture set is parallel to either grid wall (i or j principal directions). The latter case has off-diagonal permeability terms close to zero. We oriented the Tensleep reservoir grid to N72°W to minimize the off-diagonal permeability terms. The seismic fracture intensity attribute is then used to generate a realization of the reservoir fracture network. Subsequently, fracture properties are upscaled to the reservoir grid scale for a fully compositional flow simulation. We implemented a PVT analysis using CO2 swelling test results to build an 8 component equation of the state. A fully compositional flow simulation is conducted to acquire a history match between model production and production history. The history matching process reveals that high fracture permeabilities enhance water conning around the producers and decreases the oil production. Moreover, increasing apertures in the model DFN will result in higher oil production from the field. Thus, aperture and vertical permeabilities are adjusted for the model DFN to approximate the production history. We analyzed two CO2-EOR cases with different injection patterns. One has the injectors parallel to the main fracture set and the second one has injectors perpendicular to the main fracture set. Results show that the former model has higher oil recovery with later CO2 breakthrough than the second model. The dominant fracture set (N76°W) affects the CO2-EOR sweep efficiency in the Tensleep reservoir. We show that CO2 breakthrough is inevitable in both cases. The fault transmissibility multipliers are also assumed; they are uncertain parameters that could influence CO2-EOR. The model with completely impermeable faults yields a lower CO2-EOR sweep efficiency compared to the case for which all faults are fully permeable.
Recommended Citation
Kavousi Ghahfarokhi, Payam, "Reservoir Characterization and Flow Simulation for CO 2-EOR in the Tensleep Formation Using Discrete Fracture Networks, Teapot Dome, Wyoming" (2016). Graduate Theses, Dissertations, and Problem Reports. 5947.
https://researchrepository.wvu.edu/etd/5947