Feasibility Study of Solving Oil-Water Two-Phase Flow Equations Using Fourier Neural Operator

Rong Zhong

doi:10.54097/pc0qn189

Authors

Rong Zhong

DOI:

https://doi.org/10.54097/pc0qn189

Keywords:

Deep learning; Neural operators; FNO; Oil-water two-phase flow; Partial differential equations.

Abstract

In petroleum reservoir engineering, deep learning shows promising potential for solving partial differential equations (PDEs) in reservoir numerical simulations. However, traditional neural network-based methods for PDE solutions are often limited, typically requiring retraining for different equations. This study introduces the Fourier Neural Operator (FNO) approach, incorporating the fast Fourier transform mechanism to facilitate operator learning. We conduct a feasibility analysis of using FNO for solving two-dimensional oil-water two-phase flow PDEs. Through extensive literature review, we summarize the current state and limitations of deep learning in oil-water two-phase flow modeling, highlighting the application and advancements of neural operators for PDE solutions. This study demonstrates the potential of FNO to solve high-dimensional oil-water two-phase flow equations, offering valuable insights into intelligent predictions for complex subsurface reservoirs and presenting new methods and perspectives for further development in reservoir numerical modeling.

Downloads

Download data is not yet available.

References

[1] Han Jiangxia, Xue Liang, Liu Yuetian, et al. Solving Oil-Water Two-Phase Flow Equations Using Deep Neural Networks [C]//Beijing Energy Association, China University of Petroleum (Beijing), State Key Laboratory of Petroleum Resources and Prospecting, State Key Laboratory of Heavy Oil Processing, Baidu Intelligent Cloud. Proceedings of the 2022 China Oil and Gas Intelligent Technology Conference - The 5th Petroleum and Petrochemical Artificial Intelligence High-End Forum and the 8th Intelligent Digital Oilfield Open Forum, 2022:11. DOI: 10.26914/c.cnkihy.2022.035114.

[2] Cha Wenshu, Li Daolun, Shen Luhang, et al. A Review of Neural Network-Based Methods for Solving Partial Differential Equations [J]. Chinese Journal of Theoretical and Applied Mechanics, 2022, 54(03): 543-556.

[3] Long Z, Lu Y, Ma X, et al. Pde-net: Learning pdes from data[C]//International conference on machine learning. PMLR, 2018: 3208-3216.

[4] McCulloch W S, Pitts W. A logical calculus of the ideas immanent in nervous activity[J]. The bulletin of mathematical biophysics, 1943, 5: 115-133.

[5] Rumelhart D E, Hinton G E, Williams R J. Learning representations by back-propagating errors[J]. nature, 1986, 323(6088): 533-536.

[6] Baydin A G, Pearlmutter B A, Radul A A, et al. Automatic differentiation in machine learning: a survey[J]. Journal of Marchine Learning Research, 2018, 18: 1-43.

[7] Wornik K, Stinchcombe M, White H. Universal approximation of an unknown mapping and its derivatives using multilayer feedforward networks. Neural Networks, 1990, 3(5): 551-560

[8] Li X. Simultaneous approximations of multivariate functions and their derivatives by neural networks with one hidden layer. Neurocomputing, 1996, 12(4): 327-343

[9] Lagaris I E, Likas A, Fotiadis D I. Artificial neural networks for solving ordinary and partial differential equations[J]. IEEE transactions on neural networks, 1998, 9(5): 987-1000.

[10] Aarts L P, Van Der Veer P. Neural network method for solving partial differential equations[J]. Neural Processing Letters, 2001, 14: 261-271.

[11] Ramuhalli P, Udpa L, Udpa S S. Finite-element neural networks for solving differential equations[J]. IEEE transactions on neural networks, 2005, 16(6): 1381-1392.

[12] Zha W, Zhang W, Li D, et al. Convolution-based model-solving method for three-dimensional, unsteady, partial differential equations[J]. Neural Computation, 2022, 34(2): 518-540.

[13] Liu Z, Yang Y, Cai Q. Neural network as a function approximator and its application in solving differential equations[J]. Applied Mathematics and Mechanics, 2019, 40(2): 237-248.

[14] Han J, Jentzen A. Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations[J]. Communications in mathematics and statistics, 2017, 5(4): 349-380.

[15] Han J, Jentzen A, E W. Solving high-dimensional partial differential equations using deep learning[J]. Proceedings of the National Academy of Sciences, 2018, 115(34): 8505-8510.

[16] Sirignano J, Spiliopoulos K. DGM: A deep learning algorithm for solving partial differential equations[J]. Journal of computational physics, 2018, 375: 1339-1364.

[17] Kani J N, Elsheikh A H. Reduced-order modeling of subsurface multi-phase flow models using deep residual recurrent neural networks[J]. Transport in Porous Media, 2019, 126: 713-741.

[18] Raissi M, Perdikaris P, Karniadakis G E. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J]. Journal of Computational physics, 2019, 378: 686-707.

[19] Meng X, Li Z, Zhang D, et al. PPINN: Parareal physics-informed neural network for time-dependent PDEs[J]. Computer Methods in Applied Mechanics and Engineering, 2020, 370: 113250.

[20] Jagtap A D, Kawaguchi K, Karniadakis G E. Adaptive activation functions accelerate convergence in deep and physics-informed neural networks[J]. Journal of Computational Physics, 2020, 404: 109136.

[21] Fraces C G, Tchelepi H. Uncertainty Quantification for Transport in Porous Media Using Parameterized Physics Informed Neural Networks[C]//SPE Reservoir Simulation Conference?. SPE, 2023: D011S004R003.

[22] Lu L, Jin P, Karniadakis G E. Deeponet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators[J]. arXiv preprint arXiv:1910.03193, 2019.

[23] Nelsen N H, Stuart A M. The random feature model for input-output maps between banach spaces[J]. SIAM Journal on Scientific Computing, 2021, 43(5): A3212-A3243.

[24] Patel R G, Trask N A, Wood M A, et al. A physics-informed operator regression framework for extracting data-driven continuum models[J]. Computer Methods in Applied Mechanics and Engineering, 2021, 373: 113500.

[25] Hornik K, Stinchcombe M, White H. Multilayer feedforward networks are universal approximators[J]. Neural networks, 1989, 2(5): 359-366.

[26] Lu L, Jin P, Karniadakis G E. Deeponet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators[J]. arXiv preprint arXiv:1910.03193, 2019.

[27] Li Z, Kovachki N, Azizzadenesheli K, et al. Fourier neural operator for parametric partial differential equations[J]. arXiv preprint arXiv:2010.08895, 2020.

[28] Zhang K, Zuo Y, Zhao H, et al. Fourier neural operator for solving subsurface oil/water two-phase flow partial differential equation[J]. Spe Journal, 2022, 27(03): 1815-1830.

[29] Du Y. Neural operator for accelerating coronal magnetic field computations in bifrost MHD model[J]. 2024.

[30] Wen G, Li Z, Azizzadenesheli K, et al. U-FNO—An enhanced Fourier neural operator-based deep-learning model for multiphase flow[J]. Advances in Water Resources, 2022, 163: 104180.

[31] Lehmann F, Gatti F, Bertin M, et al. 3D elastic wave propagation with a factorized Fourier neural operator (F-FNO)[J]. Computer Methods in Applied Mechanics and Engineering, 2024, 420: 116718.

[32] Zhao X, Chen X, Gong Z, et al. RecFNO: A resolution-invariant flow and heat field reconstruction method from sparse observations via Fourier neural operator[J]. International Journal of Thermal Sciences, 2024, 195: 108619.