Deep learning for sustainable subsurface data analysis
Contact person: Anita Torabi
Keywords: Machine learning, 1D and 3D deep learning, Automatic seismic interpretation, Subsurface data characterization
Research group: Sedimentary Basin Group
Department of Geosciences (host)
Understanding geological heterogeneities that result from geological structures such as faults and fractures are important research areas for many applications in subsurface studies including CO2 storage underground, and geothermal energy management.
Deep Neural Networks (DNN) has a wide range of applications in geoscience, including in the reflection seismic interpretation related tasks (Fig. 1, an efficient U-Net from Bonke, et al., 2023). The size of data used for training in the interpretation of subsurface data (e.g. seismic data) is not as large as in the other disciplines, where DNN algorithms were originally developed. In addition, the architecture of most of networks are designed for prediction or classification of problems that are not directly relevant to subsurface data.
The objective of the research is to modify the architecture of Deep Neural Networks to fit classification of geological structures and subsurface data properties. This would enhance subsurface data imaging and characterization of geological features such as faults and fractures. The position will be hosted in the department of Geosciences in collaboration with informatics, and mathematics departments at the University of Oslo.
Examples of methodological research topics:
- Deep Neural Network selection and architecture customization.
- Revisiting training parameters, model parameters, optimization, and loss functions in DNN.
- Integration of geological parameters into DNN training parameters.
Examples of relevant application from natural sciences or technology:
- CO2 storage underground
- Sustainability
- Green transition
- Geothermal reservoirs
External partner:
- The Norwegian Computing Center (NR)
