Detailed geomodelling within high-resolution, 3D seismic data is a time-consuming and arduous process. However, recent advances in deep learning practices are accelerating the speed at which geologic features can be mapped. While most geoscientific deep learning applications have focused on mapping features such as faults and salt, we propose a novel, interactive deep learning methodology that enables the interpreter to characterize a petroleum system by labeling and training networks on associated elements proven by exploration well data. This study uses available data from the complex Central Graben Basin within the North Sea, which contains many producing fields.
The F3 seismic survey contains several seismic representations of petroleum system elements such as migration chimneys and dry gas shows. Dry gas migrates vertically through overlying strata and along faults. Results from well-trained deep learning networks can accurately map various petroleum elements of the basin, which is traditionally very challenging and time-consuming.
These results were obtained in a fraction of the time compared to traditional interpretation workflows and enables geoscientists to better characterize regional trends while also making observations at the petroleum system scale.