CO<sub>2</sub> Storage Monitoring via Time-Lapse Full Waveform Inversion with Automatic Differentiation
In the field of CO<sub>2</sub> capture utilization and storage (CCUS), recent advancements in active-source monitoring have significantly enhanced the capabilities of time-lapse acoustical imaging, facilitating continuous capture of detailed physical parameter images from acoustic signal...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2024-01-01
|
Series: | Nanomaterials |
Subjects: | |
Online Access: | https://www.mdpi.com/2079-4991/14/2/138 |
_version_ | 1797342853657001984 |
---|---|
author | Jixin Yang Pengliang Yu Suran Wang Zheng Sun |
author_facet | Jixin Yang Pengliang Yu Suran Wang Zheng Sun |
author_sort | Jixin Yang |
collection | DOAJ |
description | In the field of CO<sub>2</sub> capture utilization and storage (CCUS), recent advancements in active-source monitoring have significantly enhanced the capabilities of time-lapse acoustical imaging, facilitating continuous capture of detailed physical parameter images from acoustic signals. Central to these advancements is time-lapse full waveform inversion (TLFWI), which is increasingly recognized for its ability to extract high-resolution images from active-source datasets. However, conventional TLFWI methodologies, which are reliant on gradient optimization, face a significant challenge due to the need for complex, explicit formulation of the physical model gradient relative to the misfit function between observed and predicted data over time. Addressing this limitation, our study introduces automatic differentiation (AD) into the TLFWI process, utilizing deep learning frameworks such as PyTorch to automate gradient calculation using the chain rule. This novel approach, AD-TLFWI, not only streamlines the inversion of time-lapse images for CO<sub>2</sub> monitoring but also tackles the issue of local minima commonly encountered in deep learning optimizers. The effectiveness of AD-TLFWI was validated using a realistic model from the Frio-II CO<sub>2</sub> injection site, where it successfully produced high-resolution images that demonstrate significant changes in velocity due to CO<sub>2</sub> injection. This advancement in TLFWI methodology, underpinned by the integration of AD, represents a pivotal development in active-source monitoring systems, enhancing information extraction capabilities and providing potential solutions to complex multiphysics monitoring challenges. |
first_indexed | 2024-03-08T10:39:12Z |
format | Article |
id | doaj.art-3f6ce387d49a480ba2f61fe8567ef394 |
institution | Directory Open Access Journal |
issn | 2079-4991 |
language | English |
last_indexed | 2024-03-08T10:39:12Z |
publishDate | 2024-01-01 |
publisher | MDPI AG |
record_format | Article |
series | Nanomaterials |
spelling | doaj.art-3f6ce387d49a480ba2f61fe8567ef3942024-01-26T17:58:01ZengMDPI AGNanomaterials2079-49912024-01-0114213810.3390/nano14020138CO<sub>2</sub> Storage Monitoring via Time-Lapse Full Waveform Inversion with Automatic DifferentiationJixin Yang0Pengliang Yu1Suran Wang2Zheng Sun3Department of Geoscience, Pennsylvania State University, University Park, PA 16802, USADepartment of Geoscience, Pennsylvania State University, University Park, PA 16802, USACNOOC Research Institute Co., Ltd., Beijing 100028, ChinaCUMT-UCASAL Joint Research Center for Biomining and Soil Ecological Restoration, State Key Laboratory of Coal Resources and Safe Mining, China University of Mining and Technology, Xuzhou 221116, ChinaIn the field of CO<sub>2</sub> capture utilization and storage (CCUS), recent advancements in active-source monitoring have significantly enhanced the capabilities of time-lapse acoustical imaging, facilitating continuous capture of detailed physical parameter images from acoustic signals. Central to these advancements is time-lapse full waveform inversion (TLFWI), which is increasingly recognized for its ability to extract high-resolution images from active-source datasets. However, conventional TLFWI methodologies, which are reliant on gradient optimization, face a significant challenge due to the need for complex, explicit formulation of the physical model gradient relative to the misfit function between observed and predicted data over time. Addressing this limitation, our study introduces automatic differentiation (AD) into the TLFWI process, utilizing deep learning frameworks such as PyTorch to automate gradient calculation using the chain rule. This novel approach, AD-TLFWI, not only streamlines the inversion of time-lapse images for CO<sub>2</sub> monitoring but also tackles the issue of local minima commonly encountered in deep learning optimizers. The effectiveness of AD-TLFWI was validated using a realistic model from the Frio-II CO<sub>2</sub> injection site, where it successfully produced high-resolution images that demonstrate significant changes in velocity due to CO<sub>2</sub> injection. This advancement in TLFWI methodology, underpinned by the integration of AD, represents a pivotal development in active-source monitoring systems, enhancing information extraction capabilities and providing potential solutions to complex multiphysics monitoring challenges.https://www.mdpi.com/2079-4991/14/2/138automatic differentiationCO<sub>2</sub> capture utilization and storagetime-lapse monitoringfull waveform inversiondeep learning tool |
spellingShingle | Jixin Yang Pengliang Yu Suran Wang Zheng Sun CO<sub>2</sub> Storage Monitoring via Time-Lapse Full Waveform Inversion with Automatic Differentiation Nanomaterials automatic differentiation CO<sub>2</sub> capture utilization and storage time-lapse monitoring full waveform inversion deep learning tool |
title | CO<sub>2</sub> Storage Monitoring via Time-Lapse Full Waveform Inversion with Automatic Differentiation |
title_full | CO<sub>2</sub> Storage Monitoring via Time-Lapse Full Waveform Inversion with Automatic Differentiation |
title_fullStr | CO<sub>2</sub> Storage Monitoring via Time-Lapse Full Waveform Inversion with Automatic Differentiation |
title_full_unstemmed | CO<sub>2</sub> Storage Monitoring via Time-Lapse Full Waveform Inversion with Automatic Differentiation |
title_short | CO<sub>2</sub> Storage Monitoring via Time-Lapse Full Waveform Inversion with Automatic Differentiation |
title_sort | co sub 2 sub storage monitoring via time lapse full waveform inversion with automatic differentiation |
topic | automatic differentiation CO<sub>2</sub> capture utilization and storage time-lapse monitoring full waveform inversion deep learning tool |
url | https://www.mdpi.com/2079-4991/14/2/138 |
work_keys_str_mv | AT jixinyang cosub2substoragemonitoringviatimelapsefullwaveforminversionwithautomaticdifferentiation AT pengliangyu cosub2substoragemonitoringviatimelapsefullwaveforminversionwithautomaticdifferentiation AT suranwang cosub2substoragemonitoringviatimelapsefullwaveforminversionwithautomaticdifferentiation AT zhengsun cosub2substoragemonitoringviatimelapsefullwaveforminversionwithautomaticdifferentiation |