Physics-informed machine learning for fault-leakage reduced-order modeling

Journal Article · 19 March 2023 · International Journal of Greenhouse Gas Control

DOI: https://doi.org/10.1016/j.ijggc.2023.103873 · OSTI ID:1963647

Meguerdijian, Saro  ^[1]; Pawar, Rajesh J.  ^[2]; Chen, Bailian  ^[2]; Gable, Carl Walter  ^[2]; Miller, Terry Ann  ^[2]; Jha, Birendra  ^[3]

---

1. Univ. of Southern California, Los Angeles, CA (United States); Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
2. Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
3. Univ. of Southern California, Los Angeles, CA (United States)

Geologic carbon storage (GCS) is a promising technology for mitigating CO₂ emissions. The overall success of GCS depends on safe operations that are informed by risk assessment and have proper mitigation plans in place. Performing quantitative probabilistic risk assessment for a GCS site using traditional reservoir simulators can be challenging due to the high computational costs. To overcome this challenge, the US Department of Energy’s National Risk Assessment Partnership (NRAP) project has developed an integrated assessment modeling approach that utilizes computationally efficient reduced-order models (ROM) for simulating various parts of a GCS storage site to quantify uncertainty. Here, in this study, we develop a reduced-order model for fault leakage risk assessment. We use a deep learning approach to build the reduced-order model. We perform a sensitivity analysis and find that the deep learning model yields high accuracy with a much smaller computational cost than full-physics simulation. We also evaluate the performance of the model in scenarios where simulations are not possible to run, providing analysis not previously performed in fault-leakage ROM analyses. Based on a sensitivity analysis of the model, we suggest a simplified conceptual model for fault leakage and site monitoring.

• Research Organization: Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
• Sponsoring Organization: USDOE Office of Fossil Energy (FE), Clean Coal and Carbon Management
• Grant/Contract Number: 89233218CNA000001
• OSTI ID: 1963647
• Report Number(s): LA-UR-22-28020
• Journal Information: International Journal of Greenhouse Gas Control, Journal Name: International Journal of Greenhouse Gas Control Vol. 125; ISSN 1750-5836
• Publisher: ElsevierCopyright Statement
• Country of Publication: United States
• Language: English