Graph Convolutional Networks for Simulating Multi-phase Flow and Transport in Porous Media
2307.04449
0
0
Abstract
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.
Create account to get full access
Overview
- This paper explores the use of Graph Convolutional Networks (GCNs) for simulating multi-phase flow and transport in porous media, which is a complex process with important applications in fields like oil and gas, groundwater management, and geological carbon sequestration.
- The researchers developed a GCN-based model that can efficiently capture the complex spatial and temporal dynamics of multiphase flow, outperforming traditional numerical simulation methods in terms of both accuracy and computational efficiency.
- The proposed approach has the potential to significantly accelerate the design, optimization, and decision-making processes in various porous media applications, making it a valuable tool for researchers and practitioners in this field.
Plain English Explanation
Porous media, such as soil, rocks, and sediments, are found in many natural and engineered systems, and understanding the flow and transport of fluids (like oil, water, and gas) through these media is crucial for applications ranging from oil and gas extraction to groundwater management and geological carbon storage.
The flow and transport processes in porous media can be very complex, involving interactions between multiple fluid phases, heterogeneous material properties, and intricate geometries. Traditional numerical simulation methods, while accurate, can be computationally intensive and may struggle to capture all the nuances of these complex systems.
The researchers in this paper propose using a type of machine learning model called a Graph Convolutional Network (GCN) to simulate multi-phase flow and transport in porous media. GCNs are well-suited for this task because they can effectively capture the spatial and temporal relationships between different parts of the porous medium, allowing them to model the complex flow and transport dynamics more efficiently than traditional methods.
The key idea is to represent the porous medium as a graph, where the nodes represent different regions or "control volumes" within the medium, and the edges represent the connections between these regions. The GCN model can then learn to predict the flow and transport behavior within this graph-based representation, using the relationships between the different parts of the porous medium to make accurate predictions.
By leveraging the power of GCNs, the researchers were able to develop a simulation model that is both accurate and computationally efficient, potentially enabling faster design, optimization, and decision-making in a wide range of porous media applications. This work builds on previous research in using machine learning for porous media simulations and represents an important step forward in the field.
Technical Explanation
The researchers in this paper developed a Graph Convolutional Network (GCN) model for simulating multi-phase flow and transport in porous media. The key elements of their approach are:
-
Graph Representation: The porous medium is represented as a graph, where the nodes correspond to different "control volumes" or regions within the medium, and the edges represent the connections between these regions. This graph-based representation allows the model to effectively capture the complex spatial relationships within the porous medium.
-
GCN Architecture: The researchers designed a custom GCN architecture to predict the flow and transport behavior within the porous medium graph. The GCN model takes in the graph structure and the initial conditions (e.g., fluid saturations, pressures) as inputs, and outputs the predicted flow and transport dynamics over time.
-
Training and Validation: The GCN model was trained on a large dataset of porous media simulations generated using traditional numerical methods. The researchers used this dataset to optimize the model's parameters and evaluate its performance on a held-out test set.
-
Benchmarking: The researchers compared the accuracy and computational efficiency of their GCN-based simulation model against traditional numerical methods, demonstrating significant improvements in both aspects. The GCN model was able to achieve higher accuracy while being much faster to run, making it a promising alternative to existing simulation techniques.
The key insights from this work are that GCNs can effectively capture the complex spatial and temporal dynamics of multi-phase flow and transport in porous media, and that leveraging this machine learning approach can lead to significant improvements in simulation performance compared to traditional numerical methods. This research opens up new possibilities for using advanced data-driven modeling techniques to accelerate the design, optimization, and decision-making processes in a wide range of porous media applications.
Critical Analysis
The researchers have presented a compelling approach for using Graph Convolutional Networks to simulate multi-phase flow and transport in porous media, demonstrating impressive results in terms of both accuracy and computational efficiency. However, there are a few potential limitations and areas for further research that could be explored:
-
Generalization and Transferability: The performance of the GCN model was evaluated on a specific set of porous media simulations used for training and testing. It would be important to further investigate the model's ability to generalize to a broader range of porous media systems, with varying geometries, material properties, and flow conditions.
-
Model Interpretability: As with many machine learning models, the internal workings of the GCN model may be difficult to interpret, which could be a concern for applications where transparency and explainability are important, such as in regulatory or decision-making contexts. Investigating methods to improve the interpretability of the GCN-based simulations could be a valuable direction for future research.
-
Coupling with Physical Models: While the GCN model demonstrated superior performance compared to traditional numerical methods, it is still an approximate, data-driven model of the underlying physical processes. Exploring ways to combine the GCN approach with physical models could potentially lead to even more robust and reliable simulations of multi-phase flow and transport in porous media.
-
Computational Resources and Scalability: The computational efficiency of the GCN model is a significant advantage, but the training and deployment of such models may still require substantial computational resources, especially for large-scale, high-resolution porous media systems. Investigating ways to further optimize the model architecture and training process could be important for ensuring the scalability of this approach.
Overall, this research represents an important step forward in the use of advanced machine learning techniques for simulating complex physical processes in porous media. By continuing to build on this work and addressing the potential limitations, the researchers and the broader community can further unlock the potential of GCNs and other data-driven modeling approaches to accelerate progress in this critical field.
Conclusion
This paper presents a novel approach for simulating multi-phase flow and transport in porous media using Graph Convolutional Networks (GCNs). The researchers developed a GCN-based model that can effectively capture the complex spatial and temporal dynamics of these processes, outperforming traditional numerical simulation methods in terms of both accuracy and computational efficiency.
The ability to simulate multi-phase flow and transport in porous media accurately and efficiently has important implications for a wide range of applications, including oil and gas extraction, groundwater management, and geological carbon sequestration. By leveraging the power of machine learning and GCNs, this research opens up new possibilities for accelerating the design, optimization, and decision-making processes in these domains, potentially leading to significant advances in the field.
While the results presented in this paper are promising, there are still opportunities for further research and development, such as investigating the model's generalization and transferability, improving its interpretability, and exploring ways to integrate it with physical models. Nonetheless, this work represents an important step forward in the use of advanced data-driven modeling techniques for simulating complex physical processes in porous media, and it is likely to inspire further innovations and advancements in this critical area of research.
This summary was produced with help from an AI and may contain inaccuracies - check out the links to read the original source documents!
Related Papers
🧠
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
0
0
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.
4/17/2024
Knowledge-Based Convolutional Neural Network for the Simulation and Prediction of Two-Phase Darcy Flows
Zakaria Elabid, Daniel Busby, Abdenour Hadid
0
0
Physics-informed neural networks (PINNs) have gained significant prominence as a powerful tool in the field of scientific computing and simulations. Their ability to seamlessly integrate physical principles into deep learning architectures has revolutionized the approaches to solving complex problems in physics and engineering. However, a persistent challenge faced by mainstream PINNs lies in their handling of discontinuous input data, leading to inaccuracies in predictions. This study addresses these challenges by incorporating the discretized forms of the governing equations into the PINN framework. We propose to combine the power of neural networks with the dynamics imposed by the discretized differential equations. By discretizing the governing equations, the PINN learns to account for the discontinuities and accurately capture the underlying relationships between inputs and outputs, improving the accuracy compared to traditional interpolation techniques. Moreover, by leveraging the power of neural networks, the computational cost associated with numerical simulations is substantially reduced. We evaluate our model on a large-scale dataset for the prediction of pressure and saturation fields demonstrating high accuracies compared to non-physically aware models.
4/5/2024
Graph Neural Networks for Emulation of Finite-Element Ice Dynamics in Greenland and Antarctic Ice Sheets
Younghyun Koo, Maryam Rahnemoonfar
0
0
Although numerical models provide accurate solutions for ice sheet dynamics based on physics laws, they accompany intensified computational demands to solve partial differential equations. In recent years, convolutional neural networks (CNNs) have been widely used as statistical emulators for those numerical models. However, since CNNs operate on regular grids, they cannot represent the refined meshes and computational efficiency of finite-element numerical models. Therefore, instead of CNNs, this study adopts an equivariant graph convolutional network (EGCN) as an emulator for the ice sheet dynamics modeling. EGCN reproduces ice thickness and velocity changes in the Helheim Glacier, Greenland, and Pine Island Glacier, Antarctica, with 260 times and 44 times faster computation time, respectively. Compared to the traditional CNN and graph convolutional network, EGCN shows outstanding accuracy in thickness prediction near fast ice streams by preserving the equivariance to the translation and rotation of graphs.
6/27/2024
Advection Augmented Convolutional Neural Networks
Niloufar Zakariaei, Siddharth Rout, Eldad Haber, Moshe Eliasof
0
0
Many problems in physical sciences are characterized by the prediction of space-time sequences. Such problems range from weather prediction to the analysis of disease propagation and video prediction. Modern techniques for the solution of these problems typically combine Convolution Neural Networks (CNN) architecture with a time prediction mechanism. However, oftentimes, such approaches underperform in the long-range propagation of information and lack explainability. In this work, we introduce a physically inspired architecture for the solution of such problems. Namely, we propose to augment CNNs with advection by designing a novel semi-Lagrangian push operator. We show that the proposed operator allows for the non-local transformation of information compared with standard convolutional kernels. We then complement it with Reaction and Diffusion neural components to form a network that mimics the Reaction-Advection-Diffusion equation, in high dimensions. We demonstrate the effectiveness of our network on a number of spatio-temporal datasets that show their merit.
6/28/2024