|
Christopher Lerch, BHP Billiton Petroleum (Americas) Inc (United States)
Troy Thompson, DownUnder GeoSolutions (Australia)
Gillian Apps, BHP Billiton Petroleum (Americas) Inc (United States)
Ian Hayes, Murphy Oil Co (Malaysia)
Markus Leishman, BHP Billiton Petroleum (Americas) Inc (United States)
Michael Gardner, Montana State University (United States)
Dean Stoughton, BHP Billiton Petroleum (Americas) Inc (United States)
Michael Glinsky, BHP Billiton Petroleum (Americas) Inc (United States)
Christopher White, Louisiana State University (United States)
|
|
Datasets from hydrocarbon discoveries in deepwater stratigraphy in the Gulf of Mexico have served as the basis for the creation of a 3D synthetic stratigraphic and seismic model. Datasets included thick vertical sections of turbidite stratigraphy and good quality 3D seismic images. Creation of the model served as a quantitative way to apply systematic stratigraphic principles derived from outcrop and subsurface studies, and predict variation in local reservoir quality and properties. In addition this work was able to measure the ability of seismic data to detect such variations, which in this case were relatively subtle.
The clastic system at these discoveries encompasses 15-25 Ma and equates roughly to one second-order sequence stratigraphic cycle. This interval has been split into three third order cycles, each c.2 to 10 Ma. Seismic character and well log analysis of the turbidite system defined vertical and lateral variation in reservoir geometries (for example, axial fairway versus lobe versus lobe fringe patterns) and properties (such as bed thickness, net/gross, sorting, and permeability). A general stratigraphic framework and facies model from the Brushy Canyon Slope and Basin industry consortium was applied to the dataset to predict the shape and size of third through fifth order stratigraphic bodies, and the evolutionary variation of such bodies. This stratigraphic model, known as the "AIGR" model, characterizes the evolution of a stratigraphic system through an initial slope adjustment phase, followed by initiation, growth, and retreat phases of a turbidite system itself.
The basic and easily modifiable building block used in this synthetic model creation was a "channel-levee" pattern where the length and width of a central channel element, the length and width of flanking levee elements, and the size of the combined pattern was varied. A sheet element was a special case of this pattern. The automation of body number, placement, size, shape, and reservoir property assignment was facilitated by a driving parameter called the "retreat index". This index parameter was set to vary in a cyclic fashion, thereby driving systematic variation in all the other parameters.
The second to third order description of the model in three dimensions was constrained by the primary seismic markers mapped on the actual seismic data and penetrated by the well control through the turbidite system. Depth and time models were built in tandem, using appropriate sub-regional time-depth functions. Fourth and fifth-order body placement was done by allowing randomly generated body centers confined by a probability function defined by isochore thickness of the interval being populated.
Following creation of the synthetic model geometry and property variation a number of synthetic seismic volumes, derived volumes, and attributes were calculated. In addition many of the volumes were stratigraphically flattened to allow easier display of layer-based lateral variation.
|
