INTEGRATED RESERVOIR CHARACTERIZATION AND GEOLOGICAL UPSCALING FOR RESERVOIR FLOW SIMULATIONS OF THE SYCAMORE/MERAMEC AND HUNTON PLAYS IN OKLAHOMA

This study fully integrates multidisciplinary, multi-scalar subsurface and surface data for successful exploration and development programs in a fractured rock reservoir. Unconventional Sycamore/Meramec and conventional Hunton Carbonate plays in Oklahoma are the focus of this study. The following questions are addressed in this thesis: 1) what factors control natural fracture distributions, parameters, and their effect on fluid flow?, 2) what are the effects of geological upscaling on fluid flow simulations? and 3) what is the lithology and depositional environment of the Mississippian Sycamore/Meramec strata in the South Central Oklahoma Oil Province (SCOOP) area? Hydrocarbon production in naturally fractured reservoirs vary because some areas are more prone to fracturing than others. Also, some fracture parameters are more important than others. To address these issues, multiscale data was used to build a realistic fracture model for fluid flow simulations. As a result, a generic fracture model was developed to predict the lithology and structure of the rocks as two main factors controlling fracture distributions. Grain supported rock and/or curvature were found to be more prone to fracturing than mud supported rock and/or negative curvature. Also, fracture length was found to have a greater influence on production response than aperture. The implication of these findings helps optimize landing well locations. The upscaling process of geologic models can lead to losing fine-scale geological features, resulting in errors in production and reservoir performance predictions. To overcome this issue, the upscaling workflow was validated with history matching to find the optimal level of upscaling (OLU). OLU preserves the geological features and balance between simulation accuracy and simulation run time. As a result, horizontal upscaling larger than 100x150 ft results in increasing hydrocarbon production prediction errors. Logarithmic equations for different levels of upscaling were developed to define the production accuracy. Also, power law and lognormal relationships among grid cell size, computational simulation running time, and the number of processes were obtained. The implication of these findings can help predict the error in production if excessive upscaling is required. This workflow might be applied to other reservoirs to find optimal levels of upscaling. Additionally, many operators in the oil industry have been actively exploring the Mississippian Sycamore/Meramec strata in southern Oklahoma. The optimum drilling locations are not well known because the depositional environment, lithology, and reservoir quality, are still not well understood. To shed light on these issues, comprehensive quantitive and qualitative field, lab, and machine learning studies were conducted on two outcrops and a subsurface well. The lithofacies from the outcrop and subsurface are identified, outcrop-to-subsurface correlation were determined, and the depositional environment was interpreted. Sediment gravity flows was interpreted as the process of transport and deposition. I suggest that the bioturbated shale and/or the sandy siltstone of the Sycamore rock types can be potential target zones due to their reservoir quality, lithology, bed continuity, and brittleness. The implication of this specific study can be of direct benefit to the exploration and development programs of many companies in the Ardmore Basin of South Central Oklahoma