Concept of oil production autonomization based on multi-agent LLM systems

UDK: 681.518:622.276.5

DOI: 10.24887/0028-2448-2025-12-63-69

Key words: digital transformation, integrated operations, large language models (LLM), multi-agent systems, production autonomization, digital twins, artificial intelligence

Authors: B.M. Latypov1 M.M. Khasanov2 E.V. Yudin2 N.S. Markov3 R.A. Bondorov1 N.A. Zyryanov4 1Ufa State Petroleum Technological University, RF, Ufa 2Gazprom Neft Companу Group, RF, Saint Petersburg 3NEDRA LLC, RF, Saint Petersburg 4Saint Petersburg State University, RF, Saint Petersburg

The article examines the evolution of digital transformation in the oil and gas industry – from the concept of Integrated Operations to multi-agent systems based on Large Language Models (LLMs). Unlike traditional machine learning systems, LLMs are able to understand context, work with unstructured data, interact with corporate knowledge bases, and explain the logic behind decisions. The paper analyzes both international and domestic experience in implementing Integrated Operations Centers, identifies the limitations of existing approaches, and substantiates the need to move toward a new level of production process autonomization. A mathematical formulation of the decision-making optimization problem is proposed, along with a structured system of autonomy maturity levels. Using a practical example of a geological and technical operation for pump frequency control, the architecture of a multi-agent LLM-based system is demonstrated, highlighting its advantages over traditional approaches. It is shown that the transition to multi-agent systems reduces decision-making time from hours to minutes, creating potential for significant reductions in production losses and improvements in operational efficiency. The paper identifies levels of maturity for autonomous systems, ranging from advisory tools to partially and fully autonomous control loops. Implementation of the concept involves a gradual transition between levels based on accumulated experience, increased trust in the system's decisions, and confirmation of their reliability under industrial operating conditions.

References

1. Rosendahl T., Hepsø V., Integrated operations in the oil and gas industry: Sustainability and Capability Development, Business Science Reference, 2013, 427 p.

2. Clark A., Larsen H.W., Nordtvedt J.-E. et al., Integrated operations centers – Planning and delivering improved operational performance through IOC initiatives,

SPE-196148-MS, 2019, DOI: https://doi.org/10.2118/196148-MS. – EDN: PAJVYA

3. Chai C.F., Van den Berg F., Engbers P., Sondak G., Smart fields-10 years of experience in intelligent energy and collaboration, SPE-167872-MS, 2014,

DOI: https://doi.org/10.2118/167872-MS

4. Dickens J., Latin D., Verra G. et al., The BP field of the future programme: The continuing mission to deliver value, SPE-128672-MS, 2010,

DOI: https://doi.org/10.2118/128672-MS

5. Lilleng T., Øyen M., Farestvedt U. et al., Integrated operations in statoil – From ambition to action, SPE-150418-MS, 2012, DOI: https://doi.org/10.2118/150418-MS

6. URL: https://www.rosneft.ru/press/news/item/195043/

7. URL: https://neftegaz.ru/news/dobycha/205440-v-gazpromneft-khantose-sozdan-tsentr-upravleniya-dobychey-is...

8. Hepsø V., Bergum O.J., Kristiansen E., Digital platform for the next generation IO: A prerequisite for the high north, SPE-127550-MS, 2010,

DOI: https://doi.org/10.2118/127550-MS

9. Da Silva J.S., Nogueira I.B.R., Martins M.A.F., Alves R.M.B., Tools, technologies and frameworks for digital twins in the oil and gas industry: An in-depth analysis, Sensors, 2024, V. 24, DOI: https://doi.org/10.3390/s24196457

10. Liu H., Ren Y., Li X. et al., Research status and application of artificial intelligence large models in the oil and gas industry, Petroleum Exploration and Development, 2024, V. 51, pp. 1049–1065, DOI: https://doi.org/10.1016/S1876-3804(24)60524-0

11. Xia Y., Cao J., Duan W. et al., Control industrial automation system with large language model agents, 2024, DOI: https://doi.org/10.48550/arXiv.2409.18009

12. Chen C., Zhao K., Leng J. et al., Integrating large language model and digital twins in the context of industry 5.0: Framework, challenges and opportunities, Robotics and Computer Integrated Manufacturing, 2025, V. 74, DOI: https://doi.org/10.1016/j.rcim.2024.102899

13. Latypov B.M., Yudin E.V., Bondorov R.A. et al., Development of a large language model for data extraction from unstructured text documents: a case study on production geophysical survey reports (In Russ.), Neftyanoe khozyaystvo = Oil Industry, 2025, no. 9, pp. 108–111, DOI: https://doi.org/10.24887/0028-2448-2025-9-108-111