The Lower Cretaceous Thamama Group is one of the proven and productive hydrocarbons bearing intervals within Abu Dhabi area. The geological setting of Thamama Group is complex due to the nature of its structural evolution in a strike-slip regime with high-angle fault and subtle vertical displacement. Additionally, the stratigraphy of Thamama Group consists of multiple stacks of Carbonate reservoirs which are separated by various Anhydrite layers. Some of the carbonate layer cannot be interpreted properly due to data resolution and quality of the seismic reflector. Due to the geological complexities, using the existing 3D Seismic Data to interpret faults and horizons within due time frame for critical decision making can be extremely challenging. Vertical displacement of the fault often cannot be easily recognized from seismic section and auto-tracking horizon interpretation doesn't produce a good result due to lack of reflector continuity. In the specific area of this study, the main geological challenge was to interpret those N45W subtle faults developed at the central region of the seismic volume. To address this challenge, we first applied one of the existing machine-learning models proposed by Jiang and Norlund (2020) to pre-train a multi-channel convolutional neural network model with a set of synthetic seismic volumes (over 200 different cases), that resembles multiple geological scenarios with similar structural characteristics to those observed in the study area. We utilize a new point-based method that leverages a network analysis technique to automatically extract fault surfaces from fault imaging volumes. For horizon interpretation we use a new Assisted Horizon Interpretation workflow we are developing. This approach is good for identifying and tracking larger, more continuous surfaces throughout a seismic volume. This results in many output horizons that fill the entire seismic cube, we refer to this as “Dense Horizon” extraction. This portion of the project uses a deterministic approach for identifying & tracking patches throughout the entire seismic volume & then uses several automated and manual tools to join those patches to generate complete surfaces. The results achieved with this study helped to improve the quality of faults and horizon generation in complex geological area with improved efficiency and accuracy.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.