A translation scheme must be established to link core properties, such as rock type and permeability, from core intervals to non-cored intervals. Such a translation determines accurate and consistent permeability predictions across the field. This study focuses on an Early - Middle Devonian sand-dominated, tidal-estuarine depositional sequence. The formation is characterized by permeability variations caused by rapid facies changes, modified by a highly variable diagenetic overprint. Reservoir production is characterized by relatively thin limited intervals with prolific flow; and other intervals where the flow is low, despite permeability data, which suggest better productivity. These properties reflect a high uncertainty scatter in the permeability distribution. Existing permeability modeling techniques often fail to capture the actual range and scatter of permeability present in the formation: At the low porosity range the modeled permeability overestimates the potential flow and underestimates the performance of the prolific intervals.
This paper presents an approach to improve the permeability prediction by reducing the permeability uncertainty scatter using electrofacies groupings; then, by including the remaining permeability uncertainty scatter in the modeled result. Furthermore, recently acquired and more accurate core permeability measurements at the micro-Darcy range have improved the definition of the low permeability range for this reservoir through the higher resolution measurement.
The results of this work were an improvement in the permeability predictive model. In particular, the calibration of the low permeability intervals was better defined through an improved stress correction. At the high permeability end the identification of productive intervals was improved by the reservoir quality clustering approach combined with the core permeability statistics.
This approach results in improved reservoir characterization with better predictive capacity through better modeling of the dynamic permeability range.