What is it about?
In this work, deep learning methods, particularly Convolutional neural networks were used to predict reservoir properties (porosity and volume of clay) from remote sensing data (seismic). Convolutional neural networks proved powerful in learning patterns in training data and making reasonable predictions in test datasets for both synthetic case and a real field example from Western Australia.
Featured Image
Photo by Markus Winkler on Unsplash
Why is it important?
Predicting reservoir properties from remote sensing (seismic) data is crucial for exploration and development. It helps in the identification of areas that need to be drilled and is a prime input in volume calculations of reserves in a field that in turn impacts economic decisions.
Perspectives
This is a novel application of deep learning methods in finding reservoir properties from seismic data. The learnings from this work can help revolutionize the oil and gas workflows used in decision making. The impact of this work also can be generalized for geophysical inversion problems.
Vishal Das
Stanford University
Read the Original
This page is a summary of: Petrophysical properties prediction from prestack seismic data using convolutional neural networks, Geophysics, August 2020, Society of Exploration Geophysicists,
DOI: 10.1190/geo2019-0650.1.
You can read the full text:
Contributors
The following have contributed to this page
