Using elements of machine learning to solve the inverse problem of reconstructing the hydraulic conductivity field for a filtration problem

Tyumen State University Herald. Physical and Mathematical Modeling. Oil, Gas, Energy


Release:

2022. Vol. 8. № 2 (30)

Title: 
Using elements of machine learning to solve the inverse problem of reconstructing the hydraulic conductivity field for a filtration problem


For citation: Kosyakov V. P., Legostaev D. Yu. 2022. “Using elements of machine learning to solve the inverse problem of reconstructing the hydraulic conductivity field for a filtration problem”. Tyumen State University Herald. Physical and Mathematical Modeling. Oil, Gas, Energy, vol. 8, no. 2 (30), pp. 129-149. DOI: 10.21684/2411-7978-2022-8-2-129-149

About the authors:

Vitaly P. Kosyakov, Cand. Sci. (Phys.-Math.), Senior Researcher, Tyumen Branch of the Khristianovich Institute of Theoretical and Applied Mechanics of the Siberian Branch of the Russian Academy of Sciences; Associate Professor, Department of Oil and Gas Flow Metering, University of Tyumen; eLibrary AuthorID, Web of Science ResearcherID; lik.24@yandex.ru; ORCID: 0000-0002-2297-408X

Dmitry Yu. Legostaev, Junior Researcher, Tyumen Branch of the Khristianovich Institute of Theoretical and Applied Mechanics of the Siberian Branch of the Russian Academy of Sciences; Senior Lecturer, Department of Applied and Technical Physics, University of Tyumen; eLibrary AuthorID, legostaevdy@yandex.ru; ORCID: 0000-0001-6371-7031

Abstract:

In the modern world, machine learning methods are widely used. In the oil industry, there is also a noticeable trend to use these methods in the context of digitalization and intellectualization of the entire production process.

The present work is devoted to the development of a technique for solving the inverse problem of restoring the permeability field of an oil reservoir with the combined use of machine learning elements and a filtration model. A computational algorithm has been implemented, which implies close mutual integration of the filtration part and the machine learning block, the results of which are used to parameterize the physically meaningful model. A network of radial basis functions is used as a machine learning model. The proposed solution search procedure includes the numerical solution of the direct and adjoint problems for the filtration model. Solving the adjoint problem allows one to apply gradient optimization methods widely used in machine learning methods.

The paper presents the results of a numerical experiment. On the example of a symmetrical two-dimensional development element, a solution was obtained for the problem of restoring the permeability field for a set of zonal-heterogeneous oil reservoirs. For the reconstructed fields, the characteristic sizes of inhomogeneities coincide with the initial ones with sufficient accuracy. The fundamental possibility of a qualitative restoration of the porosity-permeability characteristics of the interwell space is shown, which is impossible when using classical interpolation methods without involving additional data. The paper studies the influence of the choice of the type of control parameter on the behavior of the objective function and its derivative, which affects the process of solving the inverse problem. As a result of the study, the use of hydrodynamic resistance as an adaptable parameter in solving the inverse problem is proposed.

References:

  1. Zakirov I. S. 2006. Development of the theory and practice of oil field development. Moscow; Izhevsk: Institute of Computer Technologies. 356 p. [In Russian]

  2. Kosyakov V. P., Legostaev D. Yu., Musakaev E. N. 2021. “The problem of the combined use of filtration theory and machine learning elements for solving the inverse problem of restoring the hydraulic conductivity of an oil field”. Tyumen State University Herald. Physical and Mathematical Modeling. Oil, Gas, Energy, vol. 7, no. 2 (26), pp. 113-129. DOI: 10.21684/2411-7978-2021-7-2-113-129 [In Russian]

  3. Basniev K. S., Dmitriev N. M., Kanevskaya R. D., Maksimov V. M. 2006. Underground hydromechanics. Moscow; Izhevsk: Institute of Computer Research. 488 p. [In Russian]

  4. Illarionov E., Temirchev P., Voloskov D., Kostoev R., Simonov M., Pissarenko D., Orlov D., Koroteev D. 2022. “End-to-end neural network approach to 3D reservoir simulation and adaptation”. Journal of Petroleum Science and Engineering, vol. 208, art. 109332. DOI: 10.1016/j.petrol.2021.109332

  5. Innes M. 2018. “Flux: Elegant machine learning with Julia”. Journal of Open Source Software, vol. 3, no. 25, art. 602. DOI: 10.21105/joss.00602

  6. Kosyakov V. P. 2020. “Structural and parametric identification of an aquifer model for an oil reservoir”. Lobachevskii Journal of Mathematics, vol. 41, no. 7, pp. 1242-1247. DOI: 10.1134/S1995080220070239

  7. Musakaev E. N., Rodionov S. P., Legostaev D. Yu., Kosyakov V. P. 2019. “Parameter identification for sector filtration model of an oil reservoir with complex structure”. AIP Conference Proceedings, vol. 2125, art. 030113. DOI: 10.1063/1.5117495

  8. Otchere D. A., Arbi Ganat T. O., Gholami R., Ridha S. 2021. “Application of supervised machine learning paradigms in the prediction of petroleum reservoir properties: Comparative analysis of ANN and SVM models”. Journal of Petroleum Science and Engineering, vol. 200, art. 108182.

  9. Tahmasebi P., Kamrava S., Bai T., Sahimi M. 2020. “Machine learning in geo- and environmental sciences: from small to large scale”. Advances in water resources, vol. 142, art. 103619. DOI: 10.1016/j.advwatres.2020.103619

  10. Yanji Wang, Hangyu Li, Jianchun Xu, Shuyang Liu, Xiaopu Wang. 2022. “Machine learning assisted relative permeability upscaling for uncertainty quantification”. Energy, vol. 245, art. 123284. DOI: 10.1016/j.energy.2022.123284