«Oilfield engineering»

Application of machine learning methods when constructing a volumetric lithology model

UDC: 536+550.3
DOI: -

Authors:

¹ Ufa University of Science and Technology, Ufa, Russia

Keywords: machine learning, petrophysical properties prediction, acoustic logging, density logging, well logging interpretation, linear regression method, volumetric lithology model

Annotation:

Building a volumetric lithology model of rocks is an important task, especially in deposits with complex geological structure, which are characterized by multicomponent mineral composition. Currently, the solution of this problem is based on the application of the numerical inversion algorithm of well logging data. The successful building of a volumetric lithology model requires availability of complete geological and geophysical information. At the same time, the reliability of the model depends on the quantity and quality of the initial data. Nowadays a limited set of well logging methods is often performed in wells for various reasons. In this regard, the question of completeness of the well logging data set becomes relevant. The problem can be solved by recovering (predicting) missing data using different machine learning methods. Most often it concerns such properties as density and travel time. The article discusses the possibility of predicting the volumetric lithology model by using machine learning algorithms.

Bibliography:

1. Primenenie metodov mashinnogo obucheniya v obrabotke dannykh geofizicheskikh issledovaniy skvazhin otlozheniy vikulovskoy svity / V.I. Sakhanyuk, E.V. Novikov, A.M. Sharifullin [i dr.] // Georesursy. – 2022. – T. 24, № 2. – S. 230–238. – DOI: 10.18599/grs.2022.2.21

2. Potekhin D.V., Galkin S.V. Primenenie tekhnologii mashinnogo obucheniya pri modelirovanii raspredeleniya litotipov na permokarbonovoy zalezhi nefti Usinskogo mestorozhdeniya // Zapiski gornogo instituta. – 2023. – T. 259. – S. 41–51. – DOI: 10.33285/2073-9028-2023-3(312)-41-51

3. Gafurov D.O. Metodika kompleksnoy interpretatsii geologo-geofizicheskikh dannykh obuchaemymi neyronnymi setyami na primere Talakanskogo neftegazokondensatnogo mestorozhdeniya: avtoref. dis. … kand. geol.-miner. nauk: 25.00.16. – Tomsk, 2013. – 22 s.

4. Malygin I.V. Metody mashinnogo obucheniya v geofizicheskikh zadachakh s defitsitom dannykh: avtoref. dis. … kand. tekhn. nauk: 25.00.10. – M., 2021. – 25 s.

5. Lin W.Ch., Tsai Ch.F., Zhong J.R. Deep learning for missing value imputation of continuous data and the effect of data discretization // Knowledge-Based Systems. – 2022. – Vol. 239, Issue 10. – DOI: 10.1016/j.knosys.2021.108079

6. Prediction of permeability from well logs using a new hybrid machine learning algorithm / M. Matinkia, R. Hashami, M. Mehrad [et al.] // Petroleum. – 2022. – Vol. 9, Issue 1. – DOI: 10.1016/j.petlm.2022.03.003

7. Prediction of Water Saturation from Well Log Data by Machine Learning Algorithms: Boosting and Super Learner / F. Moghaddam, M. Ostadhassan, M.A. Sadri [et al.] // J. of Marine Science and Engineering. – 2021. – Vol. 9, Issue 6. – P. 666. – DOI: 10.3390/jmse9060666

8. Vakhitova G.R., Kasimov A.M. Avtomaticheskoe vosstanovlenie karotazhnykh diagramm razlichnymi metodami approksimatsii // Geonauki: vremya peremen, vremya perspektiv: sb. st. 10-y yubileynoy nauch.-prakt. konf. – SPb., 2023. – S. 467–470.

9. Fedoseev A.A., Glinskikh V.N., Kazanenkov V.A. Otnositel’noe soderzhanie porodoobrazuyushchikh komponentov i osnovnye litologicheskie tipy porod bazhenovskoy svity i ee stratigraficheskikh analogov po dannym geofizicheskikh issledovaniy skvazhin i kerna // Neftegazovaya geologiya. Teoriya i praktika. – 2018. – T. 13, № 2. – DOI: 10.17353/2070-5379/15_2018