Machine learning and seismic attributes for petroleum prospect generation and evaluation: An example from offshore Australia

Seismic attributes play pivotal role in studying and understanding subsurface formations. Attributes can be extracted from seismic data and from postseismic inversion data. The growing number of seismic and elastic attributes poses a challenge, making the full benefit from each attribute very difficult, if not impossible. Various approaches are routinely used to select the best attributes for specific purposes. Machine learning algorithms have demonstrated good capabilities in combining appropriate attributes to address reservoir characterization problems. We aimed to use seismic and elastic attributes to detect hydrocarbon-saturated reservoirs, source rock, and seal rocks in the Poseidon field, offshore Australia. A large number of attributes weer extracted from seismic data and from impedance data. Artificial neural networks were implemented to combine the extracted attributes and convert them into petrophysical properties, namely, resistivity volume, and gamma ray volume, from which shale probability volume, sand probability volume, effective porosity volume, and gas chimney probability cubes are produced. The cubes were deployed for a detailed analysis of the petroleum system in the area. The produced shale and resistivity cubes helped delineate the seal rock and source rock in the area. Next, the reservoir intervals were identified using the porosity, shale, and resistivity volumes. A pretrained convolutional neural network was trained using another carefully selected attribute set to detect subtle faults that hydrocarbons might migrated through from source rock to trap. The integration of all the extracted cubes contributed to finding new prospects in the area and assessing their geologic probability of success. Our approach stands out for its multiphysical attribute integration, ML and human expertise incorporation, possible applicability to other fields.