Print this pageAAPG ANNUAL CONFERENCE AND EXHIBITION
Making the Next Giant Leap in Geosciences
April 10-13, 2011, Houston, Texas, USA
The Role of Natural and Hydraulic Fracturing and Matrix Rock Properties in Controlling Gas and Water Production from Tight Gas Sands in the Piceance Basin, Colorado, USA
(1) Upstream Technology, Marathon Oil Company, Houston, TX.
(2) Retired, Bella Vista, AR.
Efficient development of tight gas sands demands that several important technical challenges be addressed. Owing to the low matrix permeability of these reservoirs, hydraulic-fracture stimulation is required, and the interaction between induced and any natural fractures must be understood in the process of designing drilling patterns, well orientations and completion stages. Furthermore, the origin of produced water must be addressed.
This study represents the kind of integrated workflow that is needed to consider these issues. It covers an area in the Piceance Basin, where gas is produced from tight fluvial sandstones in the Williams Fork Formation. This play is widely considered as a basin-centered gas system.
The workflow comprises both static and dynamic elements. The static components include core- and log-based petrophysical characterization, sedimentological and stratigraphic analysis and geocellular modeling, while the dynamic elements comprise analysis of build-up and fracture injection tests and reservoir simulation.
Conventional log analysis, integrated with core data, provided reasonable estimates of matrix porosity and permeability, but interpretation of water saturation was hampered by lack of data on formation-water salinity. However, analysis of an NMR log provided evidence for free water in at least some of the reservoir zones. This finding does not support the view that the Piceance Basin represents a basin-centered gas system.
A geocellular model was constructed to represent reservoir architecture and the spatial distribution of matrix and fracture petrophysical properties. A novel object-modeling workflow was developed to ensure that the model represents realistic reservoir connectivity. Fracture properties were not constrained directly by observation in wells, but were derived using a crack-propagation model constrained by core measurements of geomechanical properties and extensional strain.
Finally, the static geocellular model was upscaled into a dual-porosity, dual-permeability simulation model. Assisted history matching was performed by allowing the most critical uncertain parameters, isolated by experimental design, to vary. Satisfactory match, involving water production both from matrix and from natural fractures, was found to depend most heavily on depth to free-water level in specific reservoir zones.
The study resulted in a set of history matched models that will be used to evaluate alternative development strategies.