Graph Network Surrogate Model for Subsurface Flow Optimization

Read original: arXiv:2312.08625 - Published 5/16/2024 by Haoyu Tang, Louis J. Durlofsky

Graph Network Surrogate Model for Subsurface Flow Optimization

Overview

This paper introduces a graph network surrogate model for optimizing subsurface flow problems.
The model uses a graph convolutional network to learn a compact representation of the subsurface flow field, which can then be used to efficiently optimize flow parameters.
The approach is demonstrated on several test cases, showing improved computational efficiency compared to traditional simulation-based optimization methods.

Plain English Explanation

The paper describes a new way to model and optimize subsurface flow problems, such as those encountered in groundwater management or oil and gas extraction. Traditionally, these problems are solved by running detailed computer simulations of the flow through the subsurface rock and soil. However, these simulations can be very computationally intensive, making optimization difficult.

To address this, the researchers developed a graph neural network model that can learn a compact representation of the subsurface flow field. This model takes the complex geometry of the subsurface as input and learns to predict the flow patterns efficiently. Once this surrogate model is trained, it can be used to quickly explore different flow parameter settings and find the optimal configuration, without having to run the full detailed simulation each time.

The key innovation is the use of a graph-based neural network architecture, which allows the model to naturally capture the complex, irregular structure of the subsurface. This graph-based approach is more flexible than traditional grid-based flow models, and can better represent the heterogeneous and anisotropic properties of real-world subsurface systems.

The researchers demonstrate that this graph network surrogate model can achieve significant computational speedups compared to traditional optimization methods, while still maintaining accuracy in predicting the flow. This could enable faster and more effective optimization of subsurface systems, with applications in areas like groundwater remediation, oil and gas production, and geothermal energy.

Technical Explanation

The paper presents a graph network surrogate modeling approach for optimizing subsurface flow problems. The key elements are:

Flow Equations: The model is based on the standard Darcy flow equations, which describe the pressure and flow field in a porous subsurface medium. These equations are discretized using a finite volume method on an unstructured mesh.
Graph Neural Network Architecture: The discretized flow domain is represented as a graph, where the nodes correspond to mesh elements and the edges represent connections between elements. A graph convolutional neural network is then used to learn a low-dimensional embedding of the flow field from this graph representation.
Optimization Procedure: The trained graph network surrogate model is used within an optimization loop to efficiently explore different flow parameter settings (e.g. permeability distribution) and identify the optimal configuration. This avoids the need to run the full flow simulation for each evaluation during the optimization.

The proposed approach is evaluated on several 2D and 3D subsurface flow test cases, including a heterogeneous permeability field and a fractured porous medium. The results demonstrate that the graph network surrogate model can achieve 1-2 orders of magnitude speedup in the optimization process compared to traditional simulation-based methods, while maintaining good accuracy in predicting the flow field.

Critical Analysis

The paper presents a promising approach for accelerating subsurface flow optimization, but there are a few caveats and areas for further research:

The current implementation is limited to single-phase, steady-state flow problems. Extending the graph network surrogate to handle more complex multiphase, transient flow scenarios would significantly broaden the applicability.
The training of the graph neural network model requires generating a large dataset of flow simulations, which can still be computationally expensive. Methods for more efficient data generation could help reduce this burden.
The paper does not explore the impact of the graph construction (e.g. node/edge features, connectivity) on the surrogate model performance. Further research is needed to understand the best ways to represent the subsurface geometry as a graph.
While the optimization results are promising, the paper does not provide a comprehensive comparison to other recent advances in subsurface flow optimization, such as multi-fidelity methods or differentiable simulators. A more thorough benchmarking would help contextualize the contributions.

Overall, the graph network surrogate modeling approach presented in this paper represents a valuable step forward in accelerating subsurface flow optimization. With further development and validation, this technique could have significant impact in fields like groundwater management, hydrocarbon recovery, and geothermal energy.

Conclusion

This paper introduces a novel graph network surrogate modeling approach for optimizing subsurface flow problems. The key innovation is the use of a graph convolutional neural network to learn a compact representation of the complex flow field, which can then be efficiently optimized without the need for repeated full-scale flow simulations.

The results demonstrate significant computational speedups compared to traditional optimization methods, while maintaining good accuracy in predicting the flow behavior. This could enable faster and more effective optimization of subsurface systems, with applications in areas like groundwater remediation, oil and gas production, and geothermal energy.

While the current implementation has some limitations, the graph-based surrogate modeling approach shows great promise. With further development and validation, this technique could become a powerful tool for optimizing the management of subsurface resources and systems.

This summary was produced with help from an AI and may contain inaccuracies - check out the links to read the original source documents!

Follow @aimodelsfyi on 𝕏 →

Related Papers

Graph Network Surrogate Model for Subsurface Flow Optimization

Haoyu Tang, Louis J. Durlofsky

The optimization of well locations and controls is an important step in the design of subsurface flow operations such as oil production or geological CO2 storage. These optimization problems can be computationally expensive, however, as many potential candidate solutions must be evaluated. In this study, we propose a graph network surrogate model (GNSM) for optimizing well placement and controls. The GNSM transforms the flow model into a computational graph that involves an encoding-processing-decoding architecture. Separate networks are constructed to provide global predictions for the pressure and saturation state variables. Model performance is enhanced through the inclusion of the single-phase steady-state pressure solution as a feature. A multistage multistep strategy is used for training. The trained GNSM is applied to predict flow responses in a 2D unstructured model of a channelized reservoir. Results are presented for a large set of test cases, in which five injection wells and five production wells are placed randomly throughout the model, with a random control variable (bottom-hole pressure) assigned to each well. Median relative error in pressure and saturation for 300 such test cases is 1-2%. The ability of the trained GNSM to provide accurate predictions for a new (geologically similar) permeability realization is demonstrated. Finally, the trained GNSM is used to optimize well locations and controls with a differential evolution algorithm. GNSM-based optimization results are comparable to those from simulation-based optimization, with a runtime speedup of a factor of 36. Much larger speedups are expected if the method is used for robust optimization, in which each candidate solution is evaluated on multiple geological models.

5/16/2024

🧠

Graph neural network-based surrogate modelling for real-time hydraulic prediction of urban drainage networks

Zhiyu Zhang, Chenkaixiang Lu, Wenchong Tian, Zhenliang Liao, Zhiguo Yuan

Physics-based models are computationally time-consuming and infeasible for real-time scenarios of urban drainage networks, and a surrogate model is needed to accelerate the online predictive modelling. Fully-connected neural networks (NNs) are potential surrogate models, but may suffer from low interpretability and efficiency in fitting complex targets. Owing to the state-of-the-art modelling power of graph neural networks (GNNs) and their match with urban drainage networks in the graph structure, this work proposes a GNN-based surrogate of the flow routing model for the hydraulic prediction problem of drainage networks, which regards recent hydraulic states as initial conditions, and future runoff and control policy as boundary conditions. To incorporate hydraulic constraints and physical relationships into drainage modelling, physics-guided mechanisms are designed on top of the surrogate model to restrict the prediction variables with flow balance and flooding occurrence constraints. According to case results in a stormwater network, the GNN-based model is more cost-effective with better hydraulic prediction accuracy than the NN-based model after equal training epochs, and the designed mechanisms further limit prediction errors with interpretable domain knowledge. As the model structure adheres to the flow routing mechanisms and hydraulic constraints in urban drainage networks, it provides an interpretable and effective solution for data-driven surrogate modelling. Simultaneously, the surrogate model accelerates the predictive modelling of urban drainage networks for real-time use compared with the physics-based model.

8/2/2024

Accelerated training of deep learning surrogate models for surface displacement and flow, with application to MCMC-based history matching of CO2 storage operations

Yifu Han, Francois P. Hamon, Louis J. Durlofsky

Deep learning surrogate modeling shows great promise for subsurface flow applications, but the training demands can be substantial. Here we introduce a new surrogate modeling framework to predict CO2 saturation, pressure and surface displacement for use in the history matching of carbon storage operations. Rather than train using a large number of expensive coupled flow-geomechanics simulation runs, training here involves a large number of inexpensive flow-only simulations combined with a much smaller number of coupled runs. The flow-only runs use an effective rock compressibility, which is shown to provide accurate predictions for saturation and pressure for our system. A recurrent residual U-Net architecture is applied for the saturation and pressure surrogate models, while a new residual U-Net model is introduced to predict surface displacement. The surface displacement surrogate accepts, as inputs, geomodel quantities along with saturation and pressure surrogate predictions. Median relative error for a diverse test set is less than 4% for all variables. The surrogate models are incorporated into a hierarchical Markov chain Monte Carlo history matching workflow. Surrogate error is included using a new treatment involving the full model error covariance matrix. A high degree of prior uncertainty, with geomodels characterized by uncertain geological scenario parameters (metaparameters) and associated realizations, is considered. History matching results for a synthetic true model are generated using in-situ monitoring-well data only, surface displacement data only, and both data types. The enhanced uncertainty reduction achieved with both data types is quantified. Posterior saturation and surface displacement fields are shown to correspond well with the true solution.

8/21/2024

Graph Convolutional Networks for Simulating Multi-phase Flow and Transport in Porous Media

Jiamin Jiang, Bo Guo

Numerical simulation of multi-phase fluid dynamics in porous media is critical for many energy and environmental applications in Earth's subsurface. Data-driven surrogate modeling provides computationally inexpensive alternatives to high-fidelity numerical simulators. While the commonly used convolutional neural networks (CNNs) are powerful in approximating partial differential equation solutions, it remains challenging for CNNs to handle irregular and unstructured simulation meshes. However, simulation models for Earth's subsurface often involve unstructured meshes with complex mesh geometries, which limits the application of CNNs. To address this challenge, we construct surrogate models based on Graph Convolutional Networks (GCNs) to approximate the spatial-temporal solutions of multi-phase flow and transport processes in porous media. We propose a new GCN architecture suited to the hyperbolic character of the coupled PDE system, to better capture transport dynamics. Results of 2D heterogeneous test cases show that our surrogates predict the evolutions of pressure and saturation states with high accuracy, and the predicted rollouts remain stable for multiple timesteps. Moreover, the GCN-based models generalize well to irregular domain geometries and unstructured meshes that are unseen in the training dataset.

4/16/2024