Direct Multi‐Modal Inversion of Geophysical Logs Using Deep Learning
Abstract Geosteering of wells requires fast interpretation of geophysical logs which is a non‐unique inverse problem. Current work presents a proof‐of‐concept approach to multi‐modal probabilistic inversion of logs using a single evaluation of an artificial deep neural network (DNN). A mixture densi...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
American Geophysical Union (AGU)
2022-09-01
|
Series: | Earth and Space Science |
Subjects: | |
Online Access: | https://doi.org/10.1029/2021EA002186 |
_version_ | 1811208377955516416 |
---|---|
author | Sergey Alyaev Ahmed H. Elsheikh |
author_facet | Sergey Alyaev Ahmed H. Elsheikh |
author_sort | Sergey Alyaev |
collection | DOAJ |
description | Abstract Geosteering of wells requires fast interpretation of geophysical logs which is a non‐unique inverse problem. Current work presents a proof‐of‐concept approach to multi‐modal probabilistic inversion of logs using a single evaluation of an artificial deep neural network (DNN). A mixture density DNN (MDN) is trained using the ”multiple‐trajectory‐prediction” loss functions, which avoids mode collapse typical for traditional MDNs, and allows multi‐modal prediction ahead of data. The proposed approach is verified on the real‐time stratigraphic inversion of gamma‐ray logs. The multi‐modal predictor outputs several likely inverse solutions/predictions, providing more accurate and realistic solutions compared to a deterministic regression using a DNN. For these likely stratigraphic curves, the model simultaneously predicts their probabilities, which are implicitly learned from the training geological data. The stratigraphy predictions and their probabilities obtained in milliseconds from the MDN can enable better real‐time decisions under geological uncertainties. |
first_indexed | 2024-04-12T04:20:13Z |
format | Article |
id | doaj.art-88fabc735dad4c2fa588b6d39184bffd |
institution | Directory Open Access Journal |
issn | 2333-5084 |
language | English |
last_indexed | 2024-04-12T04:20:13Z |
publishDate | 2022-09-01 |
publisher | American Geophysical Union (AGU) |
record_format | Article |
series | Earth and Space Science |
spelling | doaj.art-88fabc735dad4c2fa588b6d39184bffd2022-12-22T03:48:16ZengAmerican Geophysical Union (AGU)Earth and Space Science2333-50842022-09-0199n/an/a10.1029/2021EA002186Direct Multi‐Modal Inversion of Geophysical Logs Using Deep LearningSergey Alyaev0Ahmed H. Elsheikh1NORCE Norwegian Research Centre Bergen NorwayHeriot‐Watt University Edinburgh UKAbstract Geosteering of wells requires fast interpretation of geophysical logs which is a non‐unique inverse problem. Current work presents a proof‐of‐concept approach to multi‐modal probabilistic inversion of logs using a single evaluation of an artificial deep neural network (DNN). A mixture density DNN (MDN) is trained using the ”multiple‐trajectory‐prediction” loss functions, which avoids mode collapse typical for traditional MDNs, and allows multi‐modal prediction ahead of data. The proposed approach is verified on the real‐time stratigraphic inversion of gamma‐ray logs. The multi‐modal predictor outputs several likely inverse solutions/predictions, providing more accurate and realistic solutions compared to a deterministic regression using a DNN. For these likely stratigraphic curves, the model simultaneously predicts their probabilities, which are implicitly learned from the training geological data. The stratigraphy predictions and their probabilities obtained in milliseconds from the MDN can enable better real‐time decisions under geological uncertainties.https://doi.org/10.1029/2021EA002186geophysical inversionmulti‐modal inversiondeep neural networkmixture density networkwell‐log interpretationstratigraphic‐based geosteering |
spellingShingle | Sergey Alyaev Ahmed H. Elsheikh Direct Multi‐Modal Inversion of Geophysical Logs Using Deep Learning Earth and Space Science geophysical inversion multi‐modal inversion deep neural network mixture density network well‐log interpretation stratigraphic‐based geosteering |
title | Direct Multi‐Modal Inversion of Geophysical Logs Using Deep Learning |
title_full | Direct Multi‐Modal Inversion of Geophysical Logs Using Deep Learning |
title_fullStr | Direct Multi‐Modal Inversion of Geophysical Logs Using Deep Learning |
title_full_unstemmed | Direct Multi‐Modal Inversion of Geophysical Logs Using Deep Learning |
title_short | Direct Multi‐Modal Inversion of Geophysical Logs Using Deep Learning |
title_sort | direct multi modal inversion of geophysical logs using deep learning |
topic | geophysical inversion multi‐modal inversion deep neural network mixture density network well‐log interpretation stratigraphic‐based geosteering |
url | https://doi.org/10.1029/2021EA002186 |
work_keys_str_mv | AT sergeyalyaev directmultimodalinversionofgeophysicallogsusingdeeplearning AT ahmedhelsheikh directmultimodalinversionofgeophysicallogsusingdeeplearning |