A neural network approach to modeling oil production rates

UDK: 622.276.5.001

DOI: 10.24887/0028-2448-2026-4-80-83

Key words: predictive analytics, neural networks, multilayer perceptron, ranking, production rate forecasting

Authors: V.A. Markin (Surgutneftegas PJSC, RF, Surgut); L.V. Markina (Oil and Gas Production Department Fedorovskneft, Surgutneftegas PJSC, RF, Surgut); V.R. Bayramov (Surgutneftegas PJSC, RF, Surgut); M.Yu. Lobanok (Surgutneftegas PJSC, RF, Surgut)

This paper addresses the development of a neural network-based model for predictive assessment of potentially productive areas and effective exploration of remaining oil reserves. The object of study is a subsurface area comprising three oil fields with a reserve depletion rate of approximately 80 %. Additionally, the subject area has experienced a significant decline in annual oil production over an extended period. Input data include reservoir characteristics from more than 2000 wells producing exclusively from a single formation. The uniqueness of the approach lies in incorporating temporal dynamics (cumulative well operation time) together with static parameters (net pay thickness, net sand ratio, porosity, oil and gas saturation), enabling the modeling of oil production rates with graphical data representation. The optimal network configuration was identified as a multilayer perceptron with six hidden layer nodes. Predictive analysis was performed for 14 target areas identified for potential geological and technical operations. Forecast results for production rates and performance characteristics of future typical wells were obtained. The trained neural network models were preserved and can be applied to any field area using only a few input variables. Thus, the objective of predictive assessment and identification of productive zones was achieved, area ranking was performed, the most high-risk zones were identified, and the effectiveness of integrating neural network modeling into geological and technical operations planning was confirmed.

References

1. Boyd D., Crawford K., Six provocations for Big Data, SSRN Electronic Journal, 2011, V. 123, No. 1, DOI: http://doi.org/10.2139/ssrn.1926431

2. Siegel E., Liftoff: The basics of predictive model deployment, Predictive Analytics World, 2021, URL: https://www.predictiveanalyticsworld.com/blog/liftoff-the-basics-of-predictive-model-deployment/

3. Markin V.A., Markina L.V., Bayramov V.R. et al., Data Mining methods as a decision support system under conditions of data limitation (In Russ.), Neftyanoye khozyaystvo = Oil Industry, 2024, No. 5, pp. 138–142, DOI: https://doi.org/10.24887/0028-2448-2024-5-138-142

4. Markin V.A., Markina L.V., Bayramov V.R. et al., Intellectual analysis as a method of knowledge discovery in field development (In Russ.), Neftyanoye khozyaystvo = Oil Industry, 2025, No. 5, pp. 132–136, DOI: https://doi.org/10.24887/0028-2448-2025-5-132-136

5. Nisbet R., Miner G., Yale K., Handbook of
statistical analysis and data mining applications, Oxford, UK: Elsevier Inc.,
2018