IP Library › Granted Patent US 12,241,338
Granted Patent B2
US 12,241,338 · App. 17/753,598 · Granted Mar 4, 2025

Geological analog recommendation workflow using representative embeddings

Inventors: Zikri Bayraktar (Wilmington, MA); Hedi Driss (Boston, MA); Marie Emeline Cecile LeFranc (Jar, NO)
Assignee: SCHLUMBERGER TECHNOLOGY CORPORATION
E21B41/00G01V20/00G06F40/279E21B2200/22
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,241,338
App. No.
17/753,598
Granted
Mar 4, 2025
Kind
B2
Abstract

Aspects of the present disclosure relate to a well analog recommendation system. The well analog recommendation system may generate numerical representations indicative of text-based descriptions within a well report and/or well log associated with a well. Further, the well analog recommendation system may generate a well analog output that may include one or more text-based characterizations associated with one or more additional wells that are determined based on the numerical representation. For example, the well analog recommendation system may compare the numerical representation of the well to one or more numerical representations associated with the one or more additional wells and output the one or more text-based characterizations when the numerical representations are approximately equal or above a threshold.

Claims (48)

1. A method, comprising:

receiving a text-based characterization of a well, wherein the text-based characterization of the well comprises one or more words describing a geological characterization of the well, a petrophysical characterization of the well, and at least one of a depth characterization of the well, a depositional environment characterization of the well, or a well measurement characterization of the well;

generating a vector representation of the text-based characterization of the well based on the geological characterization, the petrophysical characterization, and the at least one of the depth characterization, the depositional environment characterization, or the well measurement characterization;

identifying one or more analog wells using the vector representation;

prompting, via a user interface, a user to provide an input indicating whether to modify a characterization of the well with text associated with the one or more analog wells;

in response to receiving a user input that indicates to modify the characterization of the well based on the one or more analog wells, generating an updated characterization of the well based on the text associated with the one or more analog wells; and

outputting the updated characterization of the well.

2. The method of claim 1 , wherein the vector representation of the text-based characterization is generated using an embeddings process.

3. The method of claim 1 , wherein identifying comprises:

comparing the vector representation to one or more vector representations associated with a plurality of wells; and

wherein the identified one or more analog wells comprise a subset of the plurality of the wells with a similarity index above a threshold.

4. The method of claim 1 , wherein generating the vector representation comprises inputting the vector representation to a geological analog recommendation model configured to output the vector representation.

5. The method of claim 1 , where in the text-based characterization comprises a plurality of words that describe the geological characterization, the petrophysical characterization, or both.

6. The method of claim 1 , wherein the updated characterization of the well comprises an additional text-based characterization of the well.

7. The method of claim 1 , wherein the text-based characterization further comprises a geological structure characterization or a lithological characterization.

8. A method, comprising:

receiving a plurality of well logs and a plurality of well reports, or beth, associated with a plurality of wellsites, wherein the plurality of well reports comprise text-based characterizations that include words describing at least one of a depth characterization, a depositional environment characterization, or a well measurement characterization;

prompting, via a user interface, a user to provide user-defined criteria that indicate one or more keywords;

identifying one or more well features from the plurality of well logs and the plurality of well reports based in part on the one or more keywords and the words describing the at least one of the depth characterization, the depositional environment characterization, or the well measurement characterization;

determining a numerical representation of each well based on the identified one or more well features;

generating a geological analog recommendation model using the numerical representation of each well;

providing a portion of a well report associated with a wellsite separate from the plurality of wellsites into the geological analog recommendation model;

receiving, as an output of the geological analog recommendation model, an updated well report associated with the wellsite separate from the plurality of wellsites, wherein the updated well report was generated by modifying the portion of the well report based on at least one analog wellsite included in the plurality of wellsites; and

outputting, via a display, the updated well report.

9. The method of claim 8 , comprising identifying the one or more well features from the plurality of well reports, wherein identifying the one or more well features comprises pre-processing the plurality of well reports to remove words that do not describe the plurality of wellsites and to generate a pre-processed well report.

10. The method of claim 8 , comprising determining the numerical representation of each well using a neural network.

11. The method of claim 8 , wherein the text-based characterization further comprises one or more of a geological structure characterization or a lithological characterization.

12. The method of claim 8 , comprising:

inputting a search query into the geological analog recommendation model, wherein the search query comprises a text-based characterization associated with a second wellsite separate from the plurality of wellsites; and

receiving, as an output of the geological analog recommendation model, a well analog output indicating a similarity between the second wellsite and at least one wellsite of the plurality of wellsites.

13. The method of claim 12 , wherein the well analog output comprises a well similarity index.

14. The method of claim 12 , wherein the well analog output comprises an additional text-based characterization.

15. The method of claim 8 , wherein the numerical representation comprises a vector representation of the identified one or more well features; and the method comprising generating the numerical representation using an embeddings process.

16. A system, comprising

a display;

one or more processors coupled to the display; and

one or more memory devices storing machine-executable instructions that, when executed by the one or more processors, cause the one or more processors to:

receive a text-based characterization of a well, wherein the text-based characterization of the well comprises one or more words describing a geological characterization of the well, a petrophysical characterization of the well, and at least one of a depth characterization of the well, a depositional environmental characterization of the well, or a well measurement characterization of the well;

generate a vector representation of the text-based characterization of the well based on the geological characterization, the petrophysical characterization, and the at least one of the depth characterization, the depositional environmental characterization, or the well measurement characterization;

identify an analog well using the vector representation;

prompt a user to provide an input indicating whether to modify the text-based characterization of the well with text included in a text-based characterization of the analog well;

in response to receiving a user input that indicates to modify the text-based characterization of the well with the text included in the text-based characterization of the analog well, modify the text-based characterization of the well with the text included in the text-based characterization of the analog well;

generate a well analog output comprising a well report based on the modified text-based characterization of the well; and

display the well analog output using the display.

17. The system of claim 16 , wherein the well report comprises a drilling report, a geological report, a completion report, or any combination thereof.

18. The system of claim 16 , wherein the instructions, when executed by the one or more processors, cause the one or more processors to generate the vector representation based on numerically encoded expert characterizations from well reports and/or encoded log measurement embeddings representing unique geological formation trends.

19. The system of claim 16 , wherein the well analog output comprises an additional text-based characterization.

20. The system of claim 16 , wherein the text-based characterization further comprises a geological structure characterization or a lithological characterization.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Mar 10, 2022
From: BAYRAKTAR, ZIKRI; DRISS, HEDI; LE FRANC, MARIE EMELINE CECILE
To: SCHLUMBERGER TECHNOLOGY CORPORATION
Reel/Frame 059224/0145 →
Continuity (2)
Provisional Application 62897640 · Sep 9, 2019
Related Publication 20220341292A1 · Oct 27, 2022
References Cited (28)
US 20040257240A1 · Chen et al. · 2004 [cited by applicant]
US 20130341093A1 · Jardine · 2013 [cited by examiner]
US 20140116776A1 · Marx · 2014 [cited by examiner]
US 20170075872A1 · Tong · 2017 [cited by examiner]
US 20180232342A1 · Tong · 2018 [cited by examiner]
US 20180240021A1 · Al-Harbi et al. · 2018 [cited by applicant]
US 20190286934A1 · Lai · 2019 [cited by examiner]
US 20210334433A1 · Taieb · 2021 [cited by examiner]
WO 2018067131A1 · 2018 [cited by applicant]
T. Mikolov, K. Chen, G. Corrado, J. Dean, “Efficient Estimation of Word Representations in Vector Space,” Sep. 7, 2013 (12 pages). [cited by applicant]
T. Mikolov, I. Sutskever, K. Chen, G. Corrado, J. Dean, “Distributed Representations of Words and Phrases and their Compositionality,” NIPS13 Proceedings of the 26th International Conference on Neural Information Proces… [cited by applicant]
J. Pennington, R. Socher, and C. D. Manning. 2014. “GloVe: Global Vectors for Word Representation”. (12 pages). [cited by applicant]
P. Bojanowski, E. Grave, A. Joulin, T. Mikolov, “Enriching Word Vectors with Subword Information”, Transactions of the Association for Computational Linguistics, vol. 5, 2017, p. 135-146. [cited by applicant]
M. E. Peters, M. Neumann, M. lyyer, M. Gardner, C. Clark, K. Lee, L. Zettlemoyer, Deep contextualized word representations, Proccedings of NAACL, 2018, New Orleans, Louisiana, Jun. 1-6, 2018, pp. 2227-2237. [cited by applicant]
J. Howard, S. Ruder. “Universal Language Model Fine-tuning for Text Classification,” May 23, 2018 (12 pages). [cited by applicant]
A. Radford, K. Narasimhan, T. Salimans, and I. Sutskever. Improving language understanding with unsupervised learning. Technical report, OpenAI. Jun. 11, 2018 (17 pages). [cited by applicant]
J.Devlin, M. Chang, K. Lee and K. Toutanova, BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding, Proceedings of NAACL-HLT 2019, Minneapolis, Minnesota, Jun. 2-Jun. 7, 2019, pp. 4171-4418. [cited by applicant]
M. Grbovic, H. Cheng, “Real-time Personalization using Embeddings for Search Ranking at Airbnb,”, Applied Data Science Track paper, KDD 2018, Aug. 19-23, 2018, London, United Kingdom (10 pages). [cited by applicant]
J. Wang, P. Huang, H. Zuan, Z. Zhang, B. Zhao, D. L. Lee, “Billion-scale commodity embedding for e-commerce recommendation in Alibaba,” May 24, 2018 (10 pages). [cited by applicant]
Ramzi Karam, Using Word2vec for Music Recommendations, https://towardsdatascience.com/using-word2vec-for-music-recommendations-bb9649ac2484, Dec. 7, 2017 (17 pages). [cited by applicant]
D. Cer, Y. Yang, S. Kong, N. Hua, N. Limtiaco, R. St. John, N. Constant, M. G.-Cespedes, S. Yuan, C. Tar, Y-H. Sung, B. Strope, R. Kurzweil, “Universal Sentence Encoder”, Apr. 12, 2018 (7 pages). [cited by applicant]
Nitish Srivastrava, Elman Mansimov, Ruslan Salakhutdinov, “Unsupervised Learning of Video Representations using LSTMs”, Proceedings of the 32nd International Conference on Machine Learning, Lille, France, 2015 (10 pages… [cited by applicant]
Naveen Sai Madiraju, Seid M. Sadat, Dimitry Fisher Homa Karimabadi, “Deep temporal clustering: Fully unsupervised earning of time-domain features”, Feb. 4, 2018 (11 pages). [cited by applicant]
Mark Palatucci, Dean Pomerleau, Geoffrey Hinton, Tom M. Mitchell, “Zero-Shot Learning with Semantic Output Codes” https://www.cs.cmu.edu/afs/cs/project/theo-73/www/papers/zero-shot-learning.pdf, 2009 (9 pages). [cited by applicant]
Richard Socher, Milind Ganjoo, Hamsa Sridhar, Osbert Bastani, Christopher D. Manning, Andrew Y. Ng, “Zero-Shot earning Through Cross-Modal Transfer”, 2013 (10 pages). [cited by applicant]
International Search Report and Written Opinion issued in PCT Application PCT/US2020/049963, dated Nov. 30, 2020 (12 pages). [cited by applicant]
Ian Goodfellow and Yoshua Bengio and Aaron Courville, “Deep Learning”, https://www.deeplearningbook.org/, Chapter 15—Representation Learning, pp. 524-554, (48 pages). [cited by applicant]
International Preliminary Report on Patentability issued in PCT Application PCT/US2020/049963, dated Mar. 17, 2022 (9 pages). [cited by applicant]
Cited By (1)
US 12,608,412