Proxy modeling pf reservoir hydrodynamics with graph neural networks

About the author:

Arseny W. Umanovskiy, Postgraduate Student, The Gubkin Russian State University of Oil and Gas; lynx.ff@gmail.com

Abstract:

The primary goal of hydrodynamic reservoir modeling is to predict the production wells performance, or more precisely, the dependence of this performance on the choice of the reservoir development plan. The achievement of this goal is hampered by the lack of accurate information about the properties of the reservoir. These properties have to be inferred from indirect data, first of all from the historical indicators of already functioning wells. This information is used to perform the adaptation of the numerical reservoir model or proxy models, which are less informative but have the advantage of speed and flexibility. The article proposes a reservoir proxy modeling method based on the use of a specific artificial neural network (ANN). The novel graph convolutional architecture of the ANN takes in the graph data describing the reservoir. The edges and vertices of the graph contain a spatial description of the reservoir along with the history of the well performance. Such architecture makes it possible to train the neural network for a whole class of situations instead of only one case. In accordance with the principles of the Physics-Informed Neural Networks (PINN), the task of the ANN is to derive a kind of formulation of a physical law guiding the system, rather than just a correlation between time series. The advantages of this approach over most ANN-based proxy models used today are, firstly, speed: adjustment to historical data and forecast output are made in seconds even for hundreds of wells; secondly, a certain degree of physical meaningfulness.

References:

1.       Ababkov А. V., Vasilyev V. M., Khisamutdinov N. I., Safiullin I. R., Shaislamov V. Sh. 2014. “Express method for assessing the degree of interaction between wells using
frequency analysis of data from the history of operation of injection and production wells”. Oilfield business, no. 7, pp. 10-13. [In Russian]

2.       Brilliant L. S., Dulkarnaev M. R., Danko M. Yu., Olisheva A. O., Nabiev D. Kh.,
Khutornaya A. I., Malkov I. N. 2022. “Oil production management based on neural network optimization of well operation modes at the site of pilot works of the YuV1 formation of the Vatyeganskoye field TPE ‘Povkhneftegaz’”. Georesources, no. 1 (24), pp. 3-15. [In Russian]

3.       Proskurin V. A., Khisamutdinov N. I., Antonov M. S., Sagitov D. K. 2013. “Methods for assessing the effectiveness of the formation of a waterflooding system at the site of the West-Ust-Balykskoye field”. Automation, telemechanization and communication in the oil industry, no. 6, pp. 36-38. [In Russian]

4.       Stepanov S. V., Sokolov S. V., Ruchkin A. A., Stepanov A. V., Knyazev A. V., Korytov A. V. 2018. “The problems of assessing the mutual influence of production and injection wells based on mathematical modeling”. Tyumen State University Herald. Physical and Mathematical Modeling. Oil, Gas, Energy, no. 3 (4), pp. 146-164.
[In Russian]

5.       Stepanov S. V., Beckman A. D., Ruchkin A. A., Pospelova T. A. 2022. “Maintenance of oil field development using CRM models”. IPC Express. [In Russian]

6.       Umanovskiy А. V. 2021. “Simulation method based on training data for two-phase flow in a heterogeneous porous medium”. Computer research and modeling, no. 4 (13), pp. 779-792. [In Russian]

7.       Umanovskiy А. V. 2020. “Adversarial convolutional neural networks as a heuristic model of the two-phase filtration process in a porous medium”. Computational mechanics of continuous media, no. 2 (13), pp. 231-241.

8.       Shevtsov N. O., Stepanov V. S., Pospelova T. A. 2020. “Study of predictive ability, numerical and analytical models on the example of well interference assessment”. Tyumen State University Herald. Physical and Mathematical Modeling. Oil, Gas, Energy, no. 3 (6), pp. 131-142. [In Russian]

9.       Cai S., Zhiping M. et al. 2022. “Physics-informed neural networks (PINNs) for fluid mechanics: A review”. Acta Mechanica Sinica. Pp. 1-12.

10.     Chakra N. C. C., Saraf D. N. 2016. “History matching of petroleum reservoirs employing adaptive genetic algorithm”. Journal of Petroleum Exploration and Production Technology, no. 4 (6), pp. 653-674.

11.     Chen X., Duan Y., Houthooft R. 2016. “InfoGAN: Interpretable representation learning by information maximizing generative adversarial nets”. Advances in Neural Information Processing Systems. Pp. 2180-2188.

12.     Evensen G. 2018. “Analysis of iterative ensemble smoothers for solving inverse problems”. Computational Geosciences, no. 3 (22), pp. 885-908.

13.     Gasmi C. F., Tchelepi H. 2021. “Physics informed deep learning for flow and transport in porous media”. ArXiv: 2104.02629.

14.     Gopa K., Yamov S., Naugolnov M. 2018. “Cognitive analytical system based on data-driven approach for mature reservoir management”. Society of Petroleum Engineers.

15.     Guo Zh., Reynolds A. C., Zhao H. 2018. “A physics-based data-driven model for history matching, prediction, and characterization of waterflooding performance”.
SPE Journal, vol. 23, no 2, pp. 367-395.

16.     Hamilton W., Ying Z., Leskovec J., Guyon I. et al. 2017. “Inductive representation learning on large graphs”. Curran Associates. Inc.

17.     Holl P., Thuerey N., Koltun V. 2020. “Learning to control PDEs with differentiable physics”. International Conference on Learning Representations.

18.     Illarionov E., Temirchev P., Voloskov D. et al. 2022. “End-to-end neural network approach to 3D reservoir simulation and adaptation”. Journal of Petroleum Science and Engineering, no. 208, pp. 109-332.

19.     Jaber A. K., Al-Jawad S. N., Alhuraishawy A. K. 2019. “A review of proxy modeling applications in numerical reservoir simulation”. Journal of Geosciences, vol. 12, no. 22, pp. 1-16.

20.     Jansen J. D., Fonseca R. M., Kahrobaei S. et al. 2014. “The egg model — a geological ensemble for reservoir simulation”. Data Journal, no. 2 (1), pp. 192-195.

21.     Karniadakis G. E., Perdikaris P., Wang S., Yang L. et al. 2021. “Physics-informed machine learning”. Nature Reviews Physics, no. 6 (3), pp. 422-440.

22.     Karumuri S., Tripathy R., Bilionis I. et al. 2020. “Simulator-free solution of high-dimensional stochastic elliptic partial differential equations using deep neural networks”. Journal of Computational Physics, no. 404, pp. 109-120.

23.     Kingma D. P., Ba J. L. 2015. “Adam: A method for stochastic optimization”. 3^rd International Conference on Learning Representations.

24.     Ng C. S. W. 2021. “Smart proxy modeling of a fractured reservoir model for production optimization: Implementation of metaheuristic algorithm and probabilistic application”. Natural Resources Research, no. 3 (30), pp. 2431-2462.

25.     Oliver D. S., Chen Y. 2011. “Recent progress on reservoir history matching:
A review”. Computational Geosciences, vol. 15, no. 1, pp. 185-221.

26.     Raissi M., Perdikaris P., Karniadakis G. E. 2019. “Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations”. Journal of Computational Physics, no. 378,
pp. 686-707.

27.     Wang H. 2022. “Mosaic flows: A transferable deep learning framework for solving PDEs on unseen domains”. Computer Methods in Applied Mechanics and Engineering,
no. 389, pp. 114-424.