ХЕДДЕР-АНГЛ-новый.jpg

Neftyanaya Provintsiya 

No.3(27),2021

Using statistical machine learning methods to optimize well operation

A.V. Nasybullin, R.R. Baiburov

DOI: https://doi.org/10.25689/NP.2021.3.84-94

PP.84-94

Download article

Adobe_PDF_Icon.png
 
 

Abstract

Machine learning finds its way into a wide variety of fields of science and technology. The essential condition for its use is the availability of digital factual material. Over the long history of the operation of oil fields, a significant database has been accumulated related to the development and applied methods of well stimulation.

The paper discusses the statistical methods of machine learning for the analysis of operational parameters at the producing oil wells of the Sotnikovskoye field. In particular, based on the production wells fund, the values of the target parameters are calculated by choosing a set of factors (the nominal number of oscillations of the pumping unit per minute, the nominal stroke length of the stuffing box rod), which make it possible to optimize the operation of the well, namely, to achieve the highest pump flow rate.

Key words:

well mode optimization, well performance forecasting, machine learning, neural networks, decision tree regression, MLPRegressor multilayer perceptron, LightGBM gradient boosting

References

  1. I.T Mishchenko Raschety pri dobyche nefti i gaza [Calculation of parameters during oil and gas production]. Moscow, Neft i Gaz, 2008, 296 p. (in Russian)

  2. R.S. Andriasov, I.T. Mishchenko, A.I. Petrov Spravochnoe rukovodstvo po proektirovaniyu razrabotki i ekspluatacii neftyanyh mestorozhdenij. Dobycha nefti. [Reference guide for reservoir and production engineering. Oil production]. Moscow, OOO TID Alyans Publ., 2005, 455 p (in Russian)

  3. O.V. Denisov Sovershenstvovanie processov monitoringa i regulirovaniya razrabotki neftyanyh mestorozhdenij na osnove statisticheskih, optimizacionnyh i nejrosetevyh algoritmov [Improving oil-field management processes based on statistical, optimization and neural network algorithms].  Abstract of PhD thesis. Bugulma, 2019, 25 p (in Russian)

  4. S. Ruschka Python and machine learning. Translated from English. Moscow, Press Publ., 2017, 418 p

Authors

A.V. Nasybullin, Dr.Sc, Professor, Head of the Department for Development and Operation of Oil and Gas Fields, Almetyevsk State Oil Institute, Almetyevsk

2, Lenin st., Almetyevsk, 423450, Russian Federation

E-mail: arsval@bk.ru

 

R.R. Baiburov, TatNIPIneft Institute–PJSC TATNEFT

32, Musa Jalil st., Bugulma, 423236, Russian Federation

E-mail: robert145xb@gmail.com

For citation:

A.V. Nasybullin, R.R. Baiburov Ispol'zovanie statisticheskih metodov mashinnogo obuchenija dlja optimizacii jekspluatacii skvazhin [Using statistical machine learning methods to optimize well operation]. Neftyanaya Provintsiya, No. 3(27), 2021. pp. 84-94. DOI https://doi.org/10.25689/NP.2021.3.84-94 (in Russian)

 
 
 

   © ​A.V. Nasybullin, R.R. Baiburov, 2021
       This is an open access article under the Creative Commons Attribution 4.0 License (https://creativecommons.org/licenses/by/4.0/)