The quest to accurately model well performance hinges on the proper determination of permeability, porosity, saturation, and zone thickness. Porosity and zone thickness can be readily measured using wireline logs. Hydrocarbon saturation can also be estimated within reasonable accuracy using wireline log data. The essential ingredient for accurately predicting well performance, however, is permeability. Currently, the SPE literature includes over 875 papers addressing the determination of permeability. A brief review of this literature leaves one very confused as to what permeability is the right permeability to use for simulation and well performance prediction. We have absolute, infinite, air, brine, relative, effective, and Klinkenberg definitions for permeability as well as arithmetic, geometric, and harmonic averaging techniques. It is also possible to determine permeability through reservoir simulation by history matching. But the question still remains: What is the most reliable method to determine values of permeability for use as a predictive well-performance tool?
In stacked lenticular sands of the Piceance Basin, the problem of reliable determination of permeability is even more complex as demonstrated by Craig and Brown. 1 They used the Diagnostic Pump-In (DPI) tests as a quick testing method to determine pore pressure, permeability and closure stress.1 Building on the idea of using Diagnostic Pump-In test-derived permeability, this study was initiated using a database of 15 wells with 209 DPI-derived permeability measurements. This database was used to derive a wireline-log-based permeability calculation. The analysis of this data lead to recognition and characterization of three flow units based upon their log characteristics. Each flow unit had a unique porosity-to- permeability relationship. While there was poor correlation between log-derived permeability and DPI permeability on an individual sand basis, there was better agreement between the total permeability-feet (Kgh) derived using logs and the Kgh derived using the Diagnostic Pump-In test when summed for an individual frac stage. There is even better agreement between the two methods of deriving permeability when log-derived permeability were compared to the DPI permeability-ft for an entire well. The difference between log-based permeability and the DPI-based permeability appears to be the consequence of natural fracturing away from the borehole. Fifty percent (50%) of the DPI tests exhibited pressure dependant leakoff 2 that was interpreted as due to fractures away from the wellbore. Welllogs, however, would not measure such features.
Log-based permeability proposed here provides a "base case" value for permeability. This value can thus be used as the indicator of minimum flow capacity when evaluating the quality of a well and when planning the stimulation treatment. Zones where DPI tests indicate the presence of natural fractures should serve as an indicator to modify the standard hydraulically fracturing design for the particular stage. This indicator provides a novel approach to optimizing the size of the fracturing treatment.