document preview

Abstract

A fundamental problem facing the petroleum industry is to effectively use the large amounts and diverse types of data that are collected to define and exploit stratigraphic and structural compartments that contain undrained hydrocarbons. The key is to build data-driven, deterministic geological interpretations to intelligently target infill wells. This approach is fundamentally different from that used in geostatistically-driven approaches which interpolate the sparse data support points without maximizing the value of the data that has been collected or geological knowledge.

We describe a workstation tool and interpretation method for that allows one to combine, in 3D, well-based interpretation and quantitative analog information from fields or outcrops to make testable predictions about the location of geological bodies that are prospective infill drilling locations. This tool combines (1) 3D visualization in a common viewing environment of diverse data that are viewed at true scale (e.g., 3D surface seismic; vertical seismic profiles, reservoir simulation results, conventional wireline and borehole imaging logs, core photographs); (2) a well-based interpretation environment; and (3) an archive of digital 3D geological analog shapes and textures that one can use to relate textures seen in image logs or core images to those observed in analog data, i.e., other fields or outcrops. These shape and texture analogs can then be used to place geological bodies that can be rescaled and oriented in 3D. Comparison with seismic data may then support or refute these interpretations.

We have applied this tool to the interpretation of the Atokan Bend Conglomerate in the Boonsville Field in north Texas, a mixed siliciclastic-carbonate succession containing deltaic, estuarine, and fluvial valley-fill sandstone reservoirs. The key to exploitation of this field is identifying sandstone bodies within the lowstand, incised valleys, which are commonly less than 500 meters in width, and 20 meters in thickness. By combining core, image log, and seismic information in the 3D visualization tool, we have recognized Bend Conglomerate reservoir sandstone bodies and further, we have been able to define them in inter-well space using sandbody shape analogs from the tool's digital archive.

The key advantages of this approach are that we preserve information about the interpretation process and multiple hypotheses; we see all data at the appropriate scale; and we view the implications of the deterministic geological interpretations within the same data volume as our measured data. This provides a means of capturing and applying geological knowledge of analog formations, as well as tracking the steps in the interpretation process.