Using Physics Informed Generative Adversarial Networks to Model 3D porous media

Zihan Ren,Sanjay Srinivasan

2024-09-18

Abstract:Micro-CT scanning of rocks significantly enhances our understanding of pore-scale physics in porous media. With advancements in pore-scale simulation methods, such as pore network models, it is now possible to accurately simulate multiphase flow properties, including relative permeability, from CT-scanned rock samples. However, the limited number of CT-scanned samples and the challenge of connecting pore-scale networks to field-scale rock properties often make it difficult to use pore-scale simulated properties in realistic field-scale reservoir simulations. Deep learning approaches to create synthetic 3D rock structures allow us to simulate variations in CT rock structures, which can then be used to compute representative rock properties and flow functions. However, most current deep learning methods for 3D rock structure synthesis don't consider rock properties derived from well observations, lacking a direct link between pore-scale structures and field-scale data. We present a method to construct 3D rock structures constrained to observed rock properties using generative adversarial networks (GANs) with conditioning accomplished through a gradual Gaussian deformation process. We begin by pre-training a Wasserstein GAN to reconstruct 3D rock structures. Subsequently, we use a pore network model simulator to compute rock properties. The latent vectors for image generation in GAN are progressively altered using the Gaussian deformation approach to produce 3D rock structures constrained by well-derived conditioning data. This GAN and Gaussian deformation approach enables high-resolution synthetic image generation and reproduces user-defined rock properties such as porosity, permeability, and pore size distribution. Our research provides a novel way to link GAN-generated models to field-derived quantities.

Image and Video Processing

What problem does this paper attempt to address?

The problem this paper attempts to address is how to generate 3D core microstructures that are consistent with actual rock properties, in order to overcome the challenges in existing methods due to the lack of sufficient CT scan samples and the difficulty in linking pore networks to rock properties at the formation scale. Specifically, the paper proposes a method based on Generative Adversarial Networks (GAN) combined with physical information, constructing 3D rock structures controlled by observed rock properties through a progressive Gaussian deformation process. This method can generate high-resolution synthetic images and reproduce user-defined rock properties such as porosity, permeability, and pore size distribution. This provides a new approach to connect GAN-generated models with field observation data, which is significant for designing a systematic machine learning workflow that interpolates subsurface properties by combining pore-scale and formation-scale data. Additionally, this study represents an important step towards upscaling important multiphase physical properties (such as relative permeability or capillary pressure) from the pore scale to the formation scale.

Using Physics Informed Generative Adversarial Networks to Model 3D porous media

3D-Aware Image Synthesis Via Learning Structural and Textural Representations

Stochastic reconstruction of 3D porous media from 2D images using generative adversarial networks

Reconstruction of three-dimensional porous media using generative adversarial neural networks

Reconstruction of 3D Digital Rocks with Controllable Porosity Using CVAE-GAN

3D Reconstruction of Digital Rock Guided by Petrophysical Parameters with Deep Learning

An End-to-end Approach to Predict Physical Properties of Heterogeneous Porous Media: Coupling Deep Learning and Physics-Based Features

New 2D to 3D Reconstruction of Heterogeneous Porous Media via Deep Generative Adversarial Networks (GANs)

Reconstruction of three-dimension digital rock guided by prior information with a combination of InfoGAN and style-based GAN

Multi-scale Reconstruction of Porous Media from Low-Resolution Core Images Using Conditional Generative Adversarial Networks

3D-porous-GAN: a high-performance 3D GAN for digital core reconstruction from a single 3D image

Reconstruction of large-scale anisotropic 3D digital rocks from 2D shale images using generative adversarial network

Digital rock reconstruction with user-defined properties using conditional generative adversarial networks

Digital Rock Reconstruction Enhanced by a Novel GAN-based 2D-3D Image Fusion Framework

Multiscale fusion of digital rock images based on deep generative adversarial networks

Constrained Transformer-Based Porous Media Generation to Spatial Distribution of Rock Properties

Enhancing digital rock image resolution with a GAN constrained by prior and perceptual information

Rock Reconstruction with Deep Generative Network

Pore-scale modeling of multiphase flow in porous media using a conditional generative adversarial network (cGAN)

RockGPT: reconstructing three-dimensional digital rocks from single two-dimensional slice with deep learning

Sliced Wasserstein Distance-Guided Three-Dimensional Porous Media Reconstruction Based on Cycle-Consistent Adversarial Network and Few-Shot Learning