Abstract
Numerical reservoir simulations based on finite differences can obtain approximate global solutions of the governing equations. However, this comes at the cost of substantial computational effort, especially in the component models, which makes the history matching tedious. To this end, we propose a deep learning-based surrogate model to implement reservoirs simulation of CO2-flooding efficiently. In this work, a multiple output convolutional long-short-term-memory (ConvLSTM) surrogate model for CO2-flooding is proposed, which can simultaneously output the field pattern of pressure and saturation over time. To train the model, 300 geologically-aware inhomogeneous reservoir “realizations” are created, and forward simulations of these realizations are performed to obtain time-series data on pressure and saturation fields. In addition, to make the model convergence better, this paper uses SSIM-L2, the combination of structural similarity (SSIM) loss and L2 loss, as the loss function to train the model. The training results of the surrogate model show that the proposed method can accurately predict the pressure and saturation variations in inhomogeneous reservoirs and under different fluid conditions. The average relative error of predicted pressure in the test set was 3.79%, and the average relative error of saturation was 2.12%. For computational efficiency, the improvement over the numerical simulation model is two orders of magnitude. In this work, a surrogate model for numerical simulation of CO2 flooding considering inhomogeneous is developed based on deep learning, which has the significant advantages of high accuracy and high performance. This has important implications for reservoir development adjustment and even geological understanding.
Copyright 2022, IFEDC Organizing Committee
This paper was prepared for presentation at the 2022 International Field Exploration and Develop-ment Conference in **’an, China, 16–18 November 2022.
This paper was selected for presentation by the IFEDC Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the IFEDC Technical Team and are subject to correction by the author(s). The material does not necessarily reflect any position of the IFEDC Technical Committee its members. Papers presented at the Conference are subject to publication review by Professional Team of IFEDC Technical Committee. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of IFEDC Organizing Committee is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of IFEDC. Contact email: paper@ifedc.org.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Chaki, S., Zagayevskiy, Y., Wong, T.: A Proxy Flow Modelling Workflow to Estimate Gridded Dynamic Properties and Well Production Rates by Deep Learning Algorithms: SPE/IATMI Asia Pacific Oil & Gas Conference and Exhibition. OnePetro (2021)
Liu, Y., Sun, W., Durlofsky, L.J.: A deep-learning-based geological parameterization for history matching complex models. Math. Geosci. 51(6), 725–766 (2019)
Zhang, K., Zhang, J., Ma, X., et al.: History matching of naturally fractured reservoirs using a deep sparse autoencoder. SPE J. 26(04), 1700–1721 (2021)
Tang, M., Liu, Y., Durlofsky, L.J.: Deep-learning-based surrogate flow modeling and geological parameterization for data assimilation in 3D subsurface flow. Comput. Methods Appl. Mech. Eng. 376, 113636 (2021)
Wang, Y., Liu, H., Zhou, Y.: Development of a deep learning-based model for the entire production process of steam-assisted gravity drainage (SAGD). Fuel 287, 119565 (2021)
Wang, Y., Lin, G.: Efficient deep learning techniques for multiphase flow simulation in heterogeneous porousc media. J. Comput. Phys. 401, 108968 (2020)
**, Z.L., Liu, Y., Durlofsky, L.J.: Deep-learning-based surrogate model for reservoir simulation with time-varying well controls. J. Petrol. Sci. Eng. 192, 107273 (2020)
**ao, C., Lin, H., Leeuwenburgh, O., et al.: Surrogate-assisted inversion for large-scale history matching: Comparative study between projection-based reduced-order modeling and deep neural network. J. Petrol. Sci. Eng. 208, 109287 (2022)
Zhong, Z., Sun, A.Y., Jeong, H.: Predicting CO2 plume migration in heterogeneous formations using conditional deep convolutional generative adversarial network. Water Resour. Res. 55(7), 5830–5851 (2019)
Tang, K., Liao, X., Zhao, X., et al.: A Numerical Simulation Study of CO2 and N2 as Cushion Gas in an Underground Gas Storage: SPE/IATMI Asia Pacific Oil & Gas Conference and Exhibition. OnePetro (2021)
Shi, X., Chen, Z., Wang, H., et al.:Convolutional LSTM network: A machine learning approach for precipitation nowcasting. Advances in neural information processing systems. 2015,28
Tang, M., Liu, Y., Durlofsky. L.J.: A deep-learning-based surrogate model for data assimilation in dynamic subsurface flow problems. J. Comput. Phys. 413, 109456 (2020)
Hore, A., Ziou, D.: Image quality metrics: PSNR vs. SSIM: In: 2010 20th International Conference on Pattern Recognition. IEEE, 20102366–2369
Mashanova, A., Oliver, T.H., Jansen, V.A.: Evidence for intermittency and a truncated power law from highly resolved aphid movement data. J. R. Soc. Interface 7(42), 199–208 (2010)
Tang, T., McDonough, J.M.: A theoretical model for the porosity–permeability relationship. Int. J. Heat Mass Transfer. 103, 984–996 (2016)
Zhong, Z., Sun, A.Y., Ren, B., Wang, Y.: A deep-learning-based approach for reservoir production forecast under uncertainty. SPE J. 26(03), 1314–1340 (2021)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Zhao, Yw. et al. (2023). A Surrogate Model of CO2 Flooding Reservoir Simulation Based on Deep Learning. In: Lin, J. (eds) Proceedings of the International Field Exploration and Development Conference 2022. IFEDC 2022. Springer Series in Geomechanics and Geoengineering. Springer, Singapore. https://doi.org/10.1007/978-981-99-1964-2_602
Download citation
DOI: https://doi.org/10.1007/978-981-99-1964-2_602
Publisher Name: Springer, Singapore
Print ISBN: 978-981-99-1963-5
Online ISBN: 978-981-99-1964-2
eBook Packages: EngineeringEngineering (R0)