What is it about?
This article discusses using multiple cutting-edge techniques to better understand and map fracture systems deep underground. These fractures can act as pathways or barriers for oil and gas flow, so identifying them accurately is crucial for locating productive reservoirs and optimizing resource extraction. The researchers combined deep learning algorithms to detect faults on seismic data, analysis of seismic wave velocities to find fractured zones, and monitoring of ambient micro-earthquakes caused by rock movements along fractures. By integrating the strengths of each method, they could map both large faults as well as smaller, hard-to-detect fractures that conventional methods often miss.
Featured Image
Why is it important?
This multi-method approach is innovative and important because conventional techniques frequently struggle to fully characterize complex subsurface fracture networks, which can lead to missed opportunities or inefficiencies in oil and gas operations. By harnessing the power of methods like deep learning fault detection, seismic anisotropy analysis, and passive seismic monitoring, the researchers were able to gain unprecedented insights into the fracture systems - identifying their locations, orientations, densities, and activities in much greater detail than possible before. This work demonstrates how combining multiple state-of-the-art geophysical analyses, made feasible by access to a uniquely comprehensive dataset, can overcome the limitations of any single technique to provide a more complete and accurate picture of the subsurface. Such integrated high-resolution fracture mapping has the potential to reduce costs and environmental impacts by optimizing well placement and production.
Perspectives
Read the Original
This page is a summary of: Integration of deep learning fault segmentation, HTI analysis, and ambient microseismic methods to enhance fracture prediction in the Crisol Anticline, Colombia, Interpretation, February 2024, Society of Exploration Geophysicists,
DOI: 10.1190/int-2023-0046.1.
You can read the full text:
Contributors
The following have contributed to this page