Fracture extension replications from commercially available software can provide granular hydraulic fracture properties resulting from material and energy balances in a grid-based environment. The results from purely physics-based platforms are largely accepted as the most accurate fracture replications available. However, there has been little success in integrating these physically governed fracture replications with a reservoir simulator for history matching and production prediction. Successfully doing so would allow more dependable and physically descriptive asset development sensitivities. This case study highlights a fruitful attempt in this context through commercially available packages.
This study demonstrates the results of this workflow for an asset in Midland County, Texas. MWD log corrected surfaces are used for correct horizons. Properties inherited by the Mechanical Earth Model (MEM) grid are calibrated to DFIT and drilling mechanics. Fracture mechanics are calibrated to DFIT, chemical oil/water tracers and pressure data from the stimulation operation. The reservoir grid inherits conductivity/geometry from simulation modeling and pressure variation away from the fracture face through power law integration. Petrophysics are mineralogy derived, facies biased and DFIT calibrated. Rel-perm is associated with facies distributions and serve as the primary iteration variable.
Post stimulation Instantaneous Shut In Pressure (ISIP) is replicated within 5% agreement of field acquired data using the MEM. This demonstrates accurate in-situ stress modeling, net fracture pressure modeling, and fluid leak-off modeling. Reservoir simulations performed simultaneously on the three Wolfcamp D wells studied are within 3% accuracy, including early time simulation while reservoir fluids and stimulation fluids were produced simultaneously. Modeled well interference is in good agreement with collected fluid/oil tracer data.
The novel workflow presented retains the values inherited from diagnostics and minerology calibrated petrophysical data, and fracture geometry from physics-based fracture replications. This is a paradigm shift from the usual industry practice where fracture geometries, conductivity and petrophysics are largely iterated on, often without physical context, for a successful history match. This distinction in the workflow provides a more physically consistent resource model that can be utilized for production prediction sensitivities with higher confidence, yielding optimal unconventional resource wine-rack development.
