Formal quality criteria for automatic correlation of well logs
https://doi.org/10.18303/2619-1563-2025-2-51
Abstract
This article proposes and compares formalized criteria for the quality of automated well correlation. Three indicators are considered: the error in predicting well log values based on cross-validation; the ratio of the average intra-group variance to the total data variance; and the standard deviation of the depths of a set of stratigraphic boundaries from a reference model. A very close statistical relationship is demonstrated for the first two criteria, but due to the comparative simplicity of calculation, the second is preferable. The criterion associated with assessing the deviation from the reference model (created by an expert) can be used in machine learning for practical tasks, but is of little use because it requires the preliminary construction of a reference model.
Keywords
automatic wellbore correlation,
correlation model ranking,
stratigraphic correlation,
correlation quality criteria
Supporting agencies: The study was carried out as part of government assignment to the Russian Academy of Sciences in basic research, Project FWZZ-2022-0009.
About the Authors
V. V. Lapkovsky
Trofimuk Institute of Petroleum Geology and Geophysics, SB RAS ; Novosibirsk State Technical University
Russian Federation
Russian Federation
Vladimir V. Lapkovsky
Koptyug Ave., 3, Novosibirsk, 630090
K. Marks Ave., 20, Novosibirsk, 630073
V. I. Sheludko
Novosibirsk State Technical University
Russian Federation
Russian Federation
Vasilisa I. Sheludko
K. Marks Ave., 20, Novosibirsk, 630073
References
1. Bakdi S., Kannan N., Masini S., Chennakrishnan B. Automated well correlation using machine learning and facial recognition techniques // Abu Dhabi International Petroleum Exhibition & Conference. 2020. doi:10.2118/203301-MS.
2. Caliński T., Harabasz J. A dendrite method for cluster analysis // Communications in Statistics. 1974. Vol. 3 (1). P. 1–27. doi:10.1080/03610927408827101.
3. Lapkovsky V.V., Istomin A.V., Kontorovich V.A., Berdov V.A. Correlation of well logs as a multidimensional optimization problem // Russian Geology and Geophysics. 2015. Vol. 56 (3). P. 487–492. doi:10.1016/j.rgg.2015.02.009.
4. Lapkovsky V.V., Moiseev S.A., Lunev B.V. Quality criteria for automatic correlation of well sections based on logging data // Russian Journal of Geophysical Technologies. 2019. No. 2. P. 12–22. (In Russ.). doi:10.18303/2619-1563-2019-2-12.
5. Latypov I.D., Markov A.V., Evgrafov N.A., Shagimardanova L.R. Development of approaches to automated correlation from well log data using machine learning // Exposition Oil Gas. 2024. No. 4. P. 47–51. (In Russ.). doi:10.24412/2076-6785-2024-4-47-51.
6. Romanenkova E., Rogulina A., Shakirov A., Stulov N., Zaytsev A., Ismailova L., Kovalev D., Katterbauer K., AlShehri A. Similarity learning for wells based on logging data // Journal of Petroleum Science and Engineering. 2022. Vol. 215 (Part B). 110690. doi:10.1016/j.petrol.2022.110690.
7. Shaibakov R.A. Using a neural network device for identifying the boundaries of geological objects (In Russ.) // Engineering Sciences: Traditions and Innovations: Proceedings of the First International Scientific Conference. Chelyabinsk, 2012. P. 8–11. https://moluch.ru/conf/tech/archive/6/1491.
Review
For citations:
Lapkovsky V.V., Sheludko V.I. Formal quality criteria for automatic correlation of well logs. Russian Journal of Geophysical Technologies. 2025;(2):51-59. (In Russ.) https://doi.org/10.18303/2619-1563-2025-2-51
Views:
48
Email this article 