3-D Modeling, Upscaling, and Simulation of the Fullerton Clear Fork Unit, Andrews County, West Texas

Fred P. Wang, F. Jerry Lucia, and Stephen C. Ruppel
Bureau of Economic Geology, University of Texas at Austin, TX

Discovered in 1942, Fullerton Clear Fork field in Andrews County, West Texas, produces 42o API crude from lower Clear Fork and Wichita carbonates. The reservoir consists of inter- to intratidal dolomite and limestone. The field was unitized in 1953, and waterflood began in 1960. Production peaked in 1986 at 15,000 bopd and declined sharply to 6,000 bopd in 2001. Cumulative oil production from the field reached 289 million barrels as of July 2003.

A 2,000-acre study area containing the highest density of cores and best suites of wireline logs was chosen for detailed, integrated geologic, petrophysical, and engineering characterization. Thirty-nine rock-fabric flow layers were defined on the basis of high-frequency cycles identified from core and wireline-log studies. These flow layers are used to provide a geologically constrained framework for 3-D modeling. First, a fine-scale geological model containing 3.2 million cells (140 × 90 × 256) was constructed for the area, and porosity, permeability, and water saturation were mapped through the 3-D space. Although phih is higher in Wichita, kh is greater in the lower Clear Fork because of differing rock fabrics. The geological model was then scaled up to a coarse reservoir model containing 130,000 cells (73 × 48 × 39) for reservoir simulation. Scale-up of oil and water relative permeabilities was conducted during simulations. 

The simulation study was divided into two phases: (1) sensitivity analysis and (2) history matching. From the sensitivity study we could rank importance of reservoir parameters affecting production performance. Because fractures and breccias are common, negative skin factors (or effective well-bore radii) were used to simulate hydraulic fractures, and permeability values were modified to simulate karst breccias. Through history matching, optimal fluid and rock properties could be determined. Because most high-permeability stakes were averaged out by up-scaling, both oil and water relative permeabilities used in simulations were higher than those measured in the lab.

Simulation results provide an important glimpse into reservoir response in West Texas Clear Fork carbonate reservoirs. The techniques used to construct both model and results can serve as a basis for improved modeling and simulation of other shallow-water carbonate-platform reservoirs.