What is it about?
We trained shallow and deep machine learning algorithms with well dynamic data and post-stack seismic multi-attributes to predict carbonate reservoir productivity during early exploration phase in E&P projects.
Featured Image
Why is it important?
The seismic amplitude response is not unique. Furthermore, reliable seismic inversion data is not available during early exploration phases. On the other hand, post-stack seismic attributes can be derived from highly available fullstack amplitude volumes and be used as inputs for machine learning algorithms for prediction of inportant reservoir properties obtained during drill stem test such as flow capacity and productivity index.
Perspectives
Our method has demonstrated that it is possible to predict reservoir productivity using post stack seismic attributes with performance of uo to 85% using random forest regression. Multi-layer perceptron regression yielded to a 75% at a 30x higher computational cost. It has the power to de-risk the reservoir presence and deliverability during the early phase of exploration projects and, thus, adding economical value to them.
Marcus Maas
Petroleo Brasileiro S/A
Read the Original
This page is a summary of: Unraveling the hidden relationships between seismic multiattributes, well-dynamic data, and Brazilian pre-salt carbonate reservoirs productivity: A shallow versus deep machine-learning approach, Interpretation, September 2024, Society of Exploration Geophysicists,
DOI: 10.1190/int-2023-0113.1.
You can read the full text:
Contributors
The following have contributed to this page
