Application of machine learning methods for well logging data interpretation
UDC: 550.832
DOI: 10.33285/2413-5011-2022-9(369)-48-54
Authors:
GULIN ALEXANDER B.1,
KERIMOV ABDUL-GAPUR H.
2
1 LUKOIL-Engineering Limited KogalymNIPIneft Branch Office in Tyumen, Tyumen, Russia
2 North Caucasus Federal University, Stavropol, Russia
Keywords: machine learning, outliers, data normalization, neural networks, decision trees
Annotation:
The study describing the use of machine learning (ML) in order to interpret GIS data for Cretaceous sediments has been carried out. An analysis was carried out to identify the best ML method, which is planned to be used in future for the express interpretation of GIS data. The paper analyzed the collection and preparation of input data, which is one of the main tasks in the application of machine learning algorithms. This topic is relevant due to the fact that the data volume of logging diagrams continues to grow rapidly as recording technologies and methods improve. The introduction of machine learning methods in the geological sector will help speeding up working processes with large datasets without worsening the quality. In future, it is planned to refine algorithms for interpreting not only chalk deposits, but also the entire section of a well, which will allow, together with the expert’s conclusion, obtaining a more accurate interpretation result.
Bibliography:
1. Aleksandrov D.V. Sozdanie edinoy postoyannodeystvuyushchey seysmogeologicheskoy modeli yurskikh i achimovskikh otlozheniy Tevlinsko-Russkinskogo mestorozhdeniya na osnove pereobrabotki, pereinterpretatsii i obobshcheniya seysmicheskikh materialov 3D: otchet. – Tyumen’: Filial OOO "LUKOYL–Inzhiniring" "KogalymNIPIneft’" v g. Tyumeni, 2020. – 170 s.
2. Shier D.E. Well log normalization: Methods and guidelines // Petrophysics, 2004. – № 45(3). – Pp. 268–280.
3. Devis Dzh.S. Statisticheskiy analiz dannykh v geologii. Kn. 1. – M., 1990. – S. 109–110.
4. Elektronnyy uchebnik po statistike. [Elektronnyy resurs]. – M.: Statsoft, Inc., 2012. – URL: https://habr.com/ru
Back to issue