Fractured Reservoir Modeling in a Foreland-Basin Anticline: from Outcrop to Flow Simulation

Neil Hurley¹, Chris Zahm², Ali Raba’a³, Alper Karadavut^4
1Colorado School of Mines, Golden, CO
²ConocoPhillips, Houston, TX
³Saudi Aramco, Dhahran, Saudi Arabia
⁴Turkiye Petrolleri A. O. (TPAO), Ankara, Turkey

Little Sand Draw field, Bighorn Basin, Wyoming, is an anticlinal reservoir that produces oil and gas from fractured and faulted dolomites of the Phosphoria Formation (Permian). This anticline, which has a steeply dipping, near-vertical forelimb and a gently dipping backlimb, is similar to foreland-basin anticlines in many parts of the world. Most production is from the Ervay Member of the Phosphoria Formation, a dolomitic unit that averages 100 ft (30.5 m) in thickness, with approximately 65 ft (19.8 m) of net pay. Several factors make this small field an outstanding laboratory to study geologic controls on hydrocarbon production: 1) 14 of the 32 wells in the field have been cored, 2) excellent production and pressure data are available, 3) quality logs exist, and 4) there is very little commingled production.

The purpose of this project is to combine geological and engineering data to build a flow simulation that honors observed directional permeabilities in this faulted and fractured reservoir. Anisotropic directional permeabilities, roughly perpendicular to the fold axis, are known from pressure-interference tests at Little Sand Draw field. Flow units have been defined using cumulative storage capacity (phi-h) and flow capacity (k-h) plots. In wells with no core data, neural-network software has been used to predict permeability. Fault intensities and orientations have been quantified at Thermopolis anticline, an analogous outcrop that lies only 6 mi (10 km) to the south. 

This study constructed 3-dimensional geological and fluid-flow models of the Little Sand Draw field. We built the geological model using EarthVision software. We built the full-field, 3-D fluid-flow model using the ECLIPSE 100 black-oil simulator. The study generated three fluid-flow models: (1) an unfaulted-unfractured model, (2) a faulted-unfractured model, and (3) a faulted-fractured model. History matching showed that the unfaulted-unfractured model could not produce enough total fluids to match historical data, and reservoir pressures were too low. The faulted-unfractured model improved the history match, but could not match production data. The faulted-fractured model does the best job of matching the observed past performance at Little Sand Draw field.