Full text loading...
This study introduces an integrated workflow combining Relative Geological Time model (RGT) and Temporal Convolutional Networks (TCN) in order to enhance the seismic interpretation and property modelling accuracy. The RGT provides a strong stratigraphic and geology consistent framework, while the TCN uses the RGT to predict elastic and petrophysical properties.
The RGT permits to decipher and interpret thin carbonates geomorphologies that are debated in the literature. When integrated within the TCN, the RGT improves the property prediction above 90%, outperforming the conventional methods.
Overall, this combined workflow represent a next step towards fully digital and stratigraphy guided models. This application does not only applies to hydrocarbon exploration, but also to broader subjects such as CCS, reservoir development and model optimization.