Physics-informed machine learning with differentiable programming for heterogeneous underground reservoir pressure management
Abstract
Abstract Avoiding over-pressurization in subsurface reservoirs is critical for applications like CO $$$$_2$$$$ sequestration and wastewater injection. Managing the pressures by controlling injection/extraction are challenging because of complex heterogeneity in the subsurface. The heterogeneity typically requires high-fidelity physics-based models to make predictions on CO $$$$_2$$$$ fate. Furthermore, characterizing the heterogeneity accurately is fraught with parametric uncertainty. Accounting for both, heterogeneity and uncertainty, makes this a computationally-intensive problem challenging for current reservoir simulators. To tackle this, we use differentiable programming with a full-physics model and machine learning to determine the fluid extraction rates that prevent over-pressurization at critical reservoir locations. We use DPFEHM framework, which has trustworthy physics based on the standard two-point flux finite volume discretization and is also automatically differentiable like machine learning models. Our physics-informed machine learning framework uses convolutional neural networks to learn an appropriate extraction rate based on the permeability field. We also perform a hyperparameter search to improve the model’s accuracy. Training and testing scenarios are executed to evaluate the feasibility of using physics-informed machine learning to manage reservoir pressures. We constructed and tested a sufficiently accurate simulator that is 400 000 times faster than the underlying physics-based simulator, allowing for near real-time analysis and robust uncertainty quantification.
- Authors:
- Publication Date:
- Research Org.:
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
- Sponsoring Org.:
- USDOE National Nuclear Security Administration (NNSA); USDOE Laboratory Directed Research and Development (LDRD) Program; USDOE Office of Science (SC)
- OSTI Identifier:
- 1897291
- Alternate Identifier(s):
- OSTI ID: 1909554
- Report Number(s):
- LA-UR-22-24347
Journal ID: ISSN 2045-2322; 18734; PII: 22832
- Grant/Contract Number:
- 20200575ECR; 89233218CNA000001
- Resource Type:
- Published Article
- Journal Name:
- Scientific Reports
- Additional Journal Information:
- Journal Name: Scientific Reports Journal Volume: 12 Journal Issue: 1; Journal ID: ISSN 2045-2322
- Publisher:
- Nature Publishing Group
- Country of Publication:
- United Kingdom
- Language:
- English
- Subject:
- 54 ENVIRONMENTAL SCIENCES; carbon capture and storage; computational science; hydrology
Citation Formats
Pachalieva, Aleksandra, O’Malley, Daniel, Harp, Dylan Robert, and Viswanathan, Hari. Physics-informed machine learning with differentiable programming for heterogeneous underground reservoir pressure management. United Kingdom: N. p., 2022.
Web. doi:10.1038/s41598-022-22832-7.
Pachalieva, Aleksandra, O’Malley, Daniel, Harp, Dylan Robert, & Viswanathan, Hari. Physics-informed machine learning with differentiable programming for heterogeneous underground reservoir pressure management. United Kingdom. https://doi.org/10.1038/s41598-022-22832-7
Pachalieva, Aleksandra, O’Malley, Daniel, Harp, Dylan Robert, and Viswanathan, Hari. Fri .
"Physics-informed machine learning with differentiable programming for heterogeneous underground reservoir pressure management". United Kingdom. https://doi.org/10.1038/s41598-022-22832-7.
@article{osti_1897291,
title = {Physics-informed machine learning with differentiable programming for heterogeneous underground reservoir pressure management},
author = {Pachalieva, Aleksandra and O’Malley, Daniel and Harp, Dylan Robert and Viswanathan, Hari},
abstractNote = {Abstract Avoiding over-pressurization in subsurface reservoirs is critical for applications like CO $$_2$$ 2 sequestration and wastewater injection. Managing the pressures by controlling injection/extraction are challenging because of complex heterogeneity in the subsurface. The heterogeneity typically requires high-fidelity physics-based models to make predictions on CO $$_2$$ 2 fate. Furthermore, characterizing the heterogeneity accurately is fraught with parametric uncertainty. Accounting for both, heterogeneity and uncertainty, makes this a computationally-intensive problem challenging for current reservoir simulators. To tackle this, we use differentiable programming with a full-physics model and machine learning to determine the fluid extraction rates that prevent over-pressurization at critical reservoir locations. We use DPFEHM framework, which has trustworthy physics based on the standard two-point flux finite volume discretization and is also automatically differentiable like machine learning models. Our physics-informed machine learning framework uses convolutional neural networks to learn an appropriate extraction rate based on the permeability field. We also perform a hyperparameter search to improve the model’s accuracy. Training and testing scenarios are executed to evaluate the feasibility of using physics-informed machine learning to manage reservoir pressures. We constructed and tested a sufficiently accurate simulator that is 400 000 times faster than the underlying physics-based simulator, allowing for near real-time analysis and robust uncertainty quantification.},
doi = {10.1038/s41598-022-22832-7},
journal = {Scientific Reports},
number = 1,
volume = 12,
place = {United Kingdom},
year = {Fri Nov 04 00:00:00 EDT 2022},
month = {Fri Nov 04 00:00:00 EDT 2022}
}
https://doi.org/10.1038/s41598-022-22832-7
Works referenced in this record:
The relation between the lowering of the Piezometric surface and the rate and duration of discharge of a well using ground-water storage
journal, January 1935
- Theis, Charles V.
- Transactions, American Geophysical Union, Vol. 16, Issue 2
Decision support analysis for safety control in complex project environments based on Bayesian Networks
journal, September 2013
- Zhang, Limao; Wu, Xianguo; Ding, Lieyun
- Expert Systems with Applications, Vol. 40, Issue 11
From Fluid Flow to Coupled Processes in Fractured Rock: Recent Advances and New Frontiers
journal, February 2022
- Viswanathan, H. S.; Ajo‐Franklin, J.; Birkholzer, J. T.
- Reviews of Geophysics, Vol. 60, Issue 1
A robust deep learning workflow to predict multiphase flow behavior during geological C O 2 sequestration injection and Post-Injection periods
journal, April 2022
- Yan, Bicheng; Chen, Bailian; Robert Harp, Dylan
- Journal of Hydrology, Vol. 607
Measuring the efficiency of decision making units
journal, November 1978
- Charnes, A.; Cooper, W. W.; Rhodes, E.
- European Journal of Operational Research, Vol. 2, Issue 6
Machine learning based co-optimization of carbon dioxide sequestration and oil recovery in CO2-EOR project
journal, July 2020
- You, Junyu; Ampomah, William; Sun, Qian
- Journal of Cleaner Production, Vol. 260
Development and application of a machine learning based multi-objective optimization workflow for CO2-EOR projects
journal, March 2020
- You, Junyu; Ampomah, William; Sun, Qian
- Fuel, Vol. 264
Workflow using sparse vintage data for building a first geological and reservoir model for CO2 geological storage in deep saline aquifer. A case study in the St. Lawrence Platform, Canada
journal, June 2012
- Claprood, Maxime; Gloaguen, Erwan; Giroux, Bernard
- Greenhouse Gases: Science and Technology, Vol. 2, Issue 4
Earthquake hypocenters and focal mechanisms in central Oklahoma reveal a complex system of reactivated subsurface strike-slip faulting: Earthquake Source Parameters in Oklahoma
journal, April 2015
- McNamara, D. E.; Benz, H. M.; Herrmann, R. B.
- Geophysical Research Letters, Vol. 42, Issue 8
Earthquakes in Switzerland and surrounding regions during 2006
journal, December 2007
- Baer, Manfred; Deichmann, Nicolas; Braunmiller, Jochen
- Swiss Journal of Geosciences, Vol. 100, Issue 3
Microseismic imaging of a geothermal reservoir stimulation
journal, July 2008
- Dyer, B. C.; Schanz, U.; Ladner, F.
- The Leading Edge, Vol. 27, Issue 7
Development of robust pressure management strategies for geologic CO2 sequestration
journal, September 2017
- Harp, Dylan R.; Stauffer, Philip H.; O’Malley, Daniel
- International Journal of Greenhouse Gas Control, Vol. 64
An intelligent decision support system for management of petroleum-contaminated sites
journal, April 2001
- Geng, Liqiang; Chen, Zhi; Chan, Christine W.
- Expert Systems with Applications, Vol. 20, Issue 3
Machine learning and shallow groundwater chemistry to identify geothermal prospects in the Great Basin, USA
journal, September 2022
- Ahmmed, Bulbul; Vesselinov, Velimir V.
- Renewable Energy, Vol. 197
Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
journal, February 2019
- Raissi, M.; Perdikaris, P.; Karniadakis, G. E.
- Journal of Computational Physics, Vol. 378
Predicting solubility of CO2 in brine by advanced machine learning systems: Application to carbon capture and sequestration
journal, October 2019
- Menad, Nait Amar; Hemmati-Sarapardeh, Abdolhossein; Varamesh, Amir
- Journal of CO2 Utilization, Vol. 33
A new machine learning ensemble model for class imbalance problem of screening enhanced oil recovery methods
journal, March 2021
- Pirizadeh, Mohsen; Alemohammad, Nafiseh; Manthouri, Mohammad
- Journal of Petroleum Science and Engineering, Vol. 198
Adversarial uncertainty quantification in physics-informed neural networks
journal, October 2019
- Yang, Yibo; Perdikaris, Paris
- Journal of Computational Physics, Vol. 394
A physics-constrained deep learning model for simulating multiphase flow in 3D heterogeneous porous media
journal, April 2022
- Yan, Bicheng; Harp, Dylan Robert; Chen, Bailian
- Fuel, Vol. 313
Combining brine extraction, desalination, and residual-brine reinjection with CO2 storage in saline formations: Implications for pressure management, capacity, and risk mitigation
journal, January 2011
- Buscheck, Thomas A.; Sun, Yunwei; Hao, Yue
- Energy Procedia, Vol. 4
Machine learning to discover mineral trapping signatures due to CO2 injection
journal, July 2021
- Ahmmed, Bulbul; Karra, Satish; Vesselinov, Velimir V.
- International Journal of Greenhouse Gas Control, Vol. 109
Normal or abnormal? Machine learning for the leakage detection in carbon sequestration projects using pressure field data
journal, December 2020
- Sinha, Saurabh; de Lima, Rafael Pires; Lin, Youzuo
- International Journal of Greenhouse Gas Control, Vol. 103
Maximizing the Value of Pressure Monitoring Data from CO2 Sequestration Projects
journal, January 2013
- Mishra, Srikanta; Kelley, Mark; Zeller, Evan
- Energy Procedia, Vol. 37
Backpropagation Applied to Handwritten Zip Code Recognition
journal, December 1989
- LeCun, Y.; Boser, B.; Denker, J. S.
- Neural Computation, Vol. 1, Issue 4
GaS_GeoT: A computer program for an effective use of newly improved gas geothermometers in predicting reliable geothermal reservoir temperatures
journal, January 2021
- Acevedo-Anicasio, A.; Santoyo, E.; Pérez-Zárate, D.
- Geothermal Energy, Vol. 9, Issue 1
Machine learning for creation of generalized lumped parameter tank models of low temperature geothermal reservoir systems
journal, November 2017
- Li, Y.; Júlíusson, E.; Pálsson, H.
- Geothermics, Vol. 70
Data‐Driven Design of Wave‐Propagation Models for Shale‐Oil Reservoirs Based on Machine Learning
journal, December 2021
- Xiong, Fansheng; Ba, Jing; Gei, Davide
- Journal of Geophysical Research: Solid Earth, Vol. 126, Issue 12
A gradient-based deep neural network model for simulating multiphase flow in porous media
journal, August 2022
- Yan, Bicheng; Harp, Dylan Robert; Chen, Bailian
- Journal of Computational Physics, Vol. 463
A geomechanical approach to casing collapse prediction in oil and gas wells aided by machine learning
journal, January 2021
- Mohamadian, Nima; Ghorbani, Hamzeh; Wood, David A.
- Journal of Petroleum Science and Engineering, Vol. 196
Julia: A Fresh Approach to Numerical Computing
journal, January 2017
- Bezanson, Jeff; Edelman, Alan; Karpinski, Stefan
- SIAM Review, Vol. 59, Issue 1
A machine learning framework for rapid forecasting and history matching in unconventional reservoirs
journal, November 2021
- Srinivasan, Shriram; O’Malley, Daniel; Mudunuru, Maruti K.
- Scientific Reports, Vol. 11, Issue 1
Application of machine learning and artificial intelligence in oil and gas industry
journal, December 2021
- Sircar, Anirbid; Yadav, Kriti; Rayavarapu, Kamakshi
- Petroleum Research, Vol. 6, Issue 4
Earthquakes Induced by the Stimulation of an Enhanced Geothermal System below Basel (Switzerland)
journal, September 2009
- Deichmann, N.; Giardini, D.
- Seismological Research Letters, Vol. 80, Issue 5
Machine learning and multi-agent systems in oil and gas industry applications: A survey
journal, November 2019
- Hanga, Khadijah M.; Kovalchuk, Yevgeniya
- Computer Science Review, Vol. 34
Application of machine learning techniques for selecting the most suitable enhanced oil recovery method; challenges and opportunities
journal, October 2021
- Cheraghi, Yasaman; Kord, Shahin; Mashayekhizadeh, Vahid
- Journal of Petroleum Science and Engineering, Vol. 205
Machine learning reveals cyclic changes in seismic source spectra in Geysers geothermal field
journal, May 2018
- Holtzman, Benjamin K.; Paté, Arthur; Paisley, John
- Science Advances, Vol. 4, Issue 5
Induced seismicity associated with Enhanced Geothermal Systems
journal, June 2007
- Majer, Ernest L.; Baria, Roy; Stark, Mitch
- Geothermics, Vol. 36, Issue 3
Predicting the Geothermal Heat Flux in Greenland: A Machine Learning Approach
journal, December 2017
- Rezvanbehbahani, Soroush; Stearns, Leigh A.; Kadivar, Amir
- Geophysical Research Letters, Vol. 44, Issue 24
Machine learning in oil and gas; a SWOT analysis approach
journal, May 2019
- Hajizadeh, Yasin
- Journal of Petroleum Science and Engineering, Vol. 176
High-rate injection is associated with the increase in U.S. mid-continent seismicity
journal, June 2015
- Weingarten, M.; Ge, S.; Godt, J. W.
- Science, Vol. 348, Issue 6241
Prediction of Reservoir Temperatures Using Hydrogeochemical Data, Western Anatolia Geothermal Systems (Turkey): A Machine Learning Approach
journal, November 2019
- Tut Haklidir, Fusun S.; Haklidir, Mehmet
- Natural Resources Research, Vol. 29, Issue 4
Knowledge-based machine learning techniques for accurate prediction of CO2 storage performance in underground saline aquifers
journal, May 2022
- Vo Thanh, Hung; Yasin, Qamar; Al-Mudhafar, Watheq J.
- Applied Energy, Vol. 314
Characterization of CO2 storage and enhanced oil recovery in residual oil zones
journal, September 2019
- Chen, Bailian; Pawar, Rajesh J.
- Energy, Vol. 183
Optimal well placement and brine extraction for pressure management during CO2 sequestration
journal, November 2015
- Cihan, Abdullah; Birkholzer, Jens T.; Bianchi, Marco
- International Journal of Greenhouse Gas Control, Vol. 42
Sharp increase in central Oklahoma seismicity since 2008 induced by massive wastewater injection
journal, July 2014
- Keranen, K. M.; Weingarten, M.; Abers, G. A.
- Science, Vol. 345, Issue 6195
Flux: Elegant machine learning with Julia
journal, May 2018
- Innes, Mike
- Journal of Open Source Software, Vol. 3, Issue 25
Inferring CO2 saturation from synthetic surface seismic and downhole monitoring data using machine learning for leakage detection at CO2 sequestration sites
journal, September 2020
- Wang, Zan; Dilmore, Robert M.; Harbert, William
- International Journal of Greenhouse Gas Control, Vol. 100
PPINN: Parareal physics-informed neural network for time-dependent PDEs
journal, October 2020
- Meng, Xuhui; Li, Zhen; Zhang, Dongkun
- Computer Methods in Applied Mechanics and Engineering, Vol. 370
Prediction of CO2 diffusivity in brine using white-box machine learning
journal, July 2020
- Amar, Menad Nait; Jahanbani Ghahfarokhi, Ashkan
- Journal of Petroleum Science and Engineering, Vol. 190
Earthquakes in Switzerland and surrounding regions during 2007
journal, December 2008
- Deichmann, Nicolas; Baer, Manfred; Clinton, John
- Swiss Journal of Geosciences, Vol. 101, Issue 3
Geologic CO2 sequestration monitoring design: A machine learning and uncertainty quantification based approach
journal, September 2018
- Chen, Bailian; Harp, Dylan R.; Lin, Youzuo
- Applied Energy, Vol. 225
On the feasibility of using physics-informed machine learning for underground reservoir pressure management
journal, September 2021
- Harp, Dylan Robert; O’Malley, Dan; Yan, Bicheng
- Expert Systems with Applications, Vol. 178
Use of machine learning and data analytics to increase drilling efficiency for nearby wells
journal, April 2017
- Hegde, Chiranth; Gray, K. E.
- Journal of Natural Gas Science and Engineering, Vol. 40