Overview
This project focuses on using and enhancing components of NETL’s offshore risk modeling (ORM) platform to improve and conduct geohazard and subsurface uncertainty modeling.
The ORM is a multi-component platform built from simulating and predicting the behavior of engineered-natural systems from 2011 through 2016—incorporating lessons learned from previous deleterious events and spanning natural and anthropogenic offshore energy activities. The platform provides data, tools, and technologies to assist with evaluating potential risks and identifying possible technology gaps using science-based, data-driven assessments. The modules comprising the ORM support analysis of the subsurface, wellbore, and water column to evaluate relationships, trends, risks of offshore spills, and uncertainty.
Much of the previous work on the ORM platform was focused on developing data and tools that can be used to characterize offshore hazards, including geohazards of the subsurface. This project will enhance ORM’s Subsurface Trend Analysis™ (STA) methodology, which couples expert geologic knowledge, subsurface property datasets, and geostatistics to improve the characterization of the subsurface environment and its geohazards. These enhancements will incorporate additional datasets and refine the STA methodology to the assessment of geologic properties and subsurface uncertainty in new offshore areas at multiple scales.
Approach
The goal of this three-year project is to identify potential subsurface hazards and innovate new, advanced data computing and artificial intelligence/machine learning (AI/ML) methods to improve prediction of subsurface conditions. The project applies STA to improve prediction of advanced subsurface properties in the central Gulf of Mexico (GOM), including fracture and fault distributions and reservoir thickness. Additionally, the team is developing the STA software tool to ingest wireline and well log data that adapts subsurface property prediction and reduces uncertainty, thereby providing updated forecasts of geohazards based on the newest data. Strategic development to enhance the STA approach with 3D implementation and visualization will more accurately capture geohazards and constrain property predictions. This research will use publicly available data and models from earlier research phases to develop intelligent databases and AI/ML methods, and accelerate and validate analysis that addresses subsurface interaction challenges, such as characterization and mapping of geologic hazards, safe operations, environmental and community impacts, and infrastructure reliability.
In the first year, the project initiated development of the STA method into a AI/ML-informed, 2D software tool that acts as a virtual research assistant to streamline application and iteration of STA analyses. The goal for this development is to leverage the large scale, big data framework offered by the STA to inform local field- and reservoir-scale forecasts.
In the second year, the strategy for 3D STA application and visualization was designed and assessed for integration into the STA tool. The analysis of advanced subsurface properties in the central GOM evolved with the new 3D approach and visualization techniques.
Multidimensional and NLP analytics for the AI/ML 2D STA Tool were finalized, with the 2D tool released internally at the end of the year.
In the third year, the STA tool for 3D analytics will be released. Integration of 3D analytical and visualization logic into the tool has been initiated. Advanced subsurface property analysis in the northeastern GOM continued, with a subsurface structural complexity dataset in development via the use of the fuzzy logic tool, Spatially Integrated Multivariate Probabilistic Assessment Tool. Adaptive subsurface prediction using new field data for near-real time subsurface property prediction is in development.
The STA tool will be copyrighted and available for licensing and use either through open-source and/or commercial agreements. The analyses and results will be available via EDX.
Expected Outcome
- Development of a science-based, data-driven, 2D & 3D STA tools that can be used at various scales, from basin to wellbore, for subsurface exploration and real-time geohazard monitoring of sedimentary systems.
- Enhancement of the STA method towards an AI/ML-informed tool that can increase safety of operations, reduce environmental and community impact, and improve subsurface and geohazard predictions, thereby contributing to environmentally prudent offshore energy.
Research Products
AI/ML Integration for Accelerated Analysis and Forecast of Offshore Hazards
Mark-Moser, M., Wingo, P., Duran, R., Dyer, A., Zaengle, D., Suhag, A., Hoover, B., Pantaleone, S., Shay, J., Bauer, J., Rose, K. “AI/ML Integration for Accelerated Analysis and Forecast of Offshore Hazards”, AGU Fall Meeting 2021, Dec. 13-17, New Orleans, LA/Virtual. Session: EP027 -Proven AI/ML applications in the Earth Sciences
Improving prediction of subsurface properties using a geoscience informed, multi-technique, artificial intelligence approach
Rose, K., Mark-Moser, M., Suhag, A., and Bauer, J. 2021. Improving prediction of subsurface properties using a geoscience informed, multi-technique, artificial intelligence approach (Invited). AGU Fall Meeting 2021, Dec. 13-17, New Orleans, LA/Virtual. Session H33C – Application of Multimodal Physics-Informed Machine Learning/Deep Learning in Subsurface Flow and Transport Modeling.
Optimizing Prediction of Reservoir Properties with Artificial Intelligence, Big Data, and the Subsurface Trend Analysis Method
Rose, K., Suhag, A., Mark-Moser, M., and Wingo, P., “Optimizing Prediction of Reservoir Properties with Artificial Intelligence, Big Data, and the Subsurface Trend Analysis Method,” invited talk, accepted at the 2020 Machine Learning in Oil & Gas Virtual Conference, November 9–11, 2020.
Forecasting 3D Structural Complexity with AI/ML method: Mississippi Canyon, Gulf of Mexico
Pantaleone, S., Mark Moser, M., Bean, A., Walker, S., Rose, K., 2021, “Forecasting 3D Structural Complexity with AI/ML method: Mississippi Canyon, Gulf of Mexico”. AAPG/SEG IMAGE conference, Denver, Colorado, September 26, 2021 October 1, 2021.
Enhancing Knowledge Discovery of Unstructured Data to Support Context in Subsurface Modeling Predictions
Hoover, B., Mark Moser, M., Wingo, P., Suhag, A., Rose, K., 2021, “Enhancing Knowledge Discovery of Unstructured Data to Support Context in Subsurface Modeling Predictions,” AAPG/SEG IMAGE conference, Denver, Colorado, September 26, 2021 October 1,2021.
Geohazards and Subsurface Uncertainty Smart Modeling
Mark-Moser, M., “Geohazards and Subsurface Uncertainty Smart Modeling,” Carbon Management and Oil and Gas Research Project Review Meeting Aug. 26, 2021. https://edx.netl.doe.gov/offshore/wp-content/uploads/2021/12/Geohazards-and-Subsurface-Uncertainty-Smart-Modeling-082621.pdf
Ocean and Geohazard Analysis
Duran, R., Mark-Moser, M., Wingo, P., Dyer, A., Zaengle, D., Pantaleone, S., Hoover, B., Bauer, J., and Rose, K., “Ocean and Geohazard Analysis,” Carbon Management and Oil and Gas Research Project Review Meeting Aug. 26, 2021. https://edx.netl.doe.gov/offshore/wp-content/uploads/2021/12/Ocean-and-Geohazard-Analysis_08262021.pdf
Subsurface Trend Analysis Domains for the Northern Gulf of Mexico
Mark-Moser, M., Miller, R., Rose, K., and Bauer, J., “Subsurface Trend Analysis Domains for the Northern Gulf of Mexico,” (2020), https://edx.netl.doe.gov/dataset/subsurface-trend-analysis-domains-for-the-northern-gulf-of-mexico, DOI: 10.18141/1606228.
A Systematic, Science-Driven Approach for Predicting Subsurface Properties
Rose, K., Bauer, J.R., and Mark-Moser, M., “A Systematic, Science-Driven Approach for Predicting Subsurface Properties,” Interpretation, 8:1 (2020), 167–181, https://doi.org/10.1190/INT-2019-0019.1.
A Systematic, Science-Driven Approach for Predicting Subsurface Properties
Rose, K., Bauer, J., and Mark-Moser, M., “A Systematic, Science-Driven Approach for Predicting Subsurface Properties,” Interpretation, 8 (2020), T167–T181, https://library.seg.org/doi/10.1190/INT-2019-0019.1.
Developing a Virtual Subsurface Data Framework: Transforming DOE’s EDX Data Lake Using ML/NLP
Rose, K., Rowan, C., Sabbatino, M., Baker, V., Bauer, J., Creason, C.G., Jones, T.J., Justman, D., Romeo, L., Suhag, A., Yeates, D., and Walker, S., “Developing a Virtual Subsurface Data Framework: Transforming DOE’s EDX Data Lake Using ML/NLP”, AGU, 12/2019.
https://agu.confex.com/agu/fm19/meetingapp.cgi/Paper/596761
A Novel Fuzzy Logic Geospatial Approach to Characterize and Predict Geologic Structural Complexity
Justman, D., Creason, C.G., Rose, K., and Bauer, J.R., “A Novel Fuzzy Logic Geospatial Approach to Characterize and Predict Geologic Structural Complexity”, AGU, 12/2019.
https://agu.confex.com/agu/fm19/meetingapp.cgi/Paper/600615
Moving Data “Rocks” Out of Hard Places: Adapting and Innovating Data Science Tools to Improve Geoscience Analytics
Yeates, D., Walker, S., Fillingham, J., Sabbatino, M., Suhag, A., Rose, K., Mark-Moser, M., Creason, C.G., and Baker, V., “Moving Data “Rocks” Out of Hard Places: Adapting and Innovating Data Science Tools to Improve Geoscience Analytics”.
https://agu.confex.com/agu/fm19/meetingapp.cgi/Paper/608555
Putting Data to Work: Transforming Disparate Open-Source Data for Engineered-Natural Systems and Models
Creason, C.G., Romeo, L., Bauer, J., Rose, K., Rowan, C., and Sabbatino, M., “Putting Data to Work: Transforming Disparate Open-Source Data for Engineered-Natural Systems and Models”.
https://agu.confex.com/agu/fm19/meetingapp.cgi/Paper/600618
Integrating Artificial Intelligence with the Subsurface Trend Analysis to Predict Porous Media Properties
Mark-Moser, M., Rose, K., Bauer, J., Wingo, P., and Suhag, A., “Integrating Artificial Intelligence with the Subsurface Trend Analysis to Predict Porous Media Properties”, AGU, 12/2019.
https://agu.confex.com/agu/fm19/meetingapp.cgi/Paper/609860
Back to the Future: Rescue, Curation, and Transformation of a Corpus of Carbon Storage Data
Sabbatino, M., Baker, V., Bauer, J., Creason, C., Romeo, L., Rose, K., Rowan, C., and Zoch, G., “Back to the Future: Rescue, Curation, and Transformation of a Corpus of Carbon Storage Data”, AGU, 12/2019.
https://agu.confex.com/agu/fm19/meetingapp.cgi/Paper/600855
Subsurface Trend Analysis: A Methodical Framework for Artificial Intelligence Subsurface Property Prediction
Rose, K., Mark-Moser, M., and Suhag, A., “Subsurface Trend Analysis: A Methodical Framework for Artificial Intelligence Subsurface Property Prediction,” Machine Learning for Unconventional Resources (MLUR), University of Houston, Houston, TX, November 18, 2019.
Hydrate‐Filled Fracture Formation at Keathley Canyon 151, Gulf of Mexico and Implications for Non‐Vent Sites
Oti, E.A., Cook, A.E., Welch, S.A., Sheets, J.M., Crandall, D., Rose, K., and Daigle, H.,
“Hydrate‐Filled Fracture Formation at Keathley Canyon 151, Gulf of Mexico and Implications for Non‐Vent Sites,” Geochemistry, Geophysics, Geosystems, 20 (2019)
https://doi.org/10.1029/2019GC008637
A Systematic, Science-Driven Approach for Predicting Subsurface Properties
Rose, K., Bauer, J., and Mark-Moser, M.
Interpretation, INT-2019-0019. In review.
https://www.researchgate.net/publication/334277526_Steady_Flow_of_a_Cement_Slurry/citation/download
Explore research products that are related to this project.
*Image Source: NETL
The STA workflow integrates geologic information with subsurface datasets to produce science-based, data-driven predictions for subsurface properties in areas with scarce data.
*Image Source: NETL
Twenty-one distinct geologic domains derived from the STA workflow. Each contains a ternary diagram characterizing the influence on the domain by the three geologic end-members: structure, lithology, and secondary alteration. Colored points represent different chronozones analyzed for end-member influence within the domain.
Screenshots of the 2D STA Tool software user interface, showing analysis of geologic domains in the northern Gulf of Mexico and document classification for geologic domain formation using Natural Language Processing.
Application of the Spatially Integrated Multivariate Probabilistic Assessment tool to the Mississippi Canyon, Gulf of Mexico, to evaluate the advanced property of fault and fracture distribution. This analysis integrates with the Subsurface Trend Analysis method to provide constraint of structural geologic domains for analysis of subsurface geohazards.
Contacts
Kelly Rose
Offshore Portfolio Lead
Co-Principal Investigator
MacKenzie Mark-Moser
Research Geologist
Co-Principal Investigator
Philip Reppert
Associate Director
Geological & Environmental Systems
Roy Long
Offshore Portfolio Technical Manager
Effective Resource Development
Alexandra Hakala
Senior Fellow (Detail)
Geological & Environmental Systems
