In a low oil-price environment it is definitely challenging to conduct much data collection, particularly along horizontal section which in most cases only gamma ray log is acquired. As horizontal drilling activities have grown the areal heterogeneity in the unconventional play is revealed to be far more complex than what was initially thought. In order to address this complexity a 3D model is necessary to guide landing the horizontal well, designing the completion and determining the appropriate well spacing. Fortunately, 3D models can be built from the thousands of commingled vertical wells that exist in many development areas in the Permian Basin that provide valuable information, ranging from cores and logs, to pressure and production. In some areas, 3D seismic data exists to provide inter-well information. In this paper, we present an integrated workflow which utilizes multi-scale data from multiple sources that has been successfully applied to our unconventional reservoir factory-model development in the Permian Basin.
The proposed workflow adapts a top-down concept and fits factory-model batch development: progressing the model from a big 3D model to pad-scale and single well models. A fine-grid field-scale earth model covering tens of square miles including structure, reservoir properties and geomechanical properties is built based on existing vertical wells, newly drilled horizontal wells and seismic interpretation if available. A pad-scale model can be modified from the field model based on local information and used to evaluate multi-well interference, landing strategy and well spacing. A single well model with the smallest scale can be cut from the pad-scale model for landing and completion design. Hydraulic fracture with a discrete naturally fractured network is modeled subsequently and directly converted to unstructured reservoir simulation grid fitting reservoir simulator. The workflow forms an iterative process to update field-, pad- and well-models, as existing wells deplete and new wells are put on production through time. A case using more than 1,000 vertical wells from the Midland Basin will be presented, clearly demonstrating an approach to effectively leveraging all existing data and improving factory-model deployment.
In-house uncertainty analysis package is linked to both hydraulic fracture modeling and reservoir simulation. The package handles a variety of key parameters for hydraulic fracturing and flow performance, such as completions design, discrete fracture network (DFN) characterization and generation, unpropped hydraulic fracture properties, fracture compaction, and matrix permeability. The uncertainty analysis helps understand the relative impacts of different parameters and drive specific data collection.
