Application of Bayesian Generative Adversarial Networks to Geological Facies Modeling

Feng, Runhai; Grana, Dario; Mukerji, Tapan; Mosegaard, Klaus

doi:10.1007/s11004-022-09994-w

Cited by 17 publications

(9 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Deep generative modeling aims to train a generative model whose samples

\tilde{x}\sim {p}_{\theta }\left(\tilde{x}\right)

from the distribution learned from the training data

\tilde{x}\sim {p}_{\text{data}}\left(\tilde{x}\right)

(Bond‐Taylor et al., 2022; Goodfellow et al., 2020). Deep generative modeling neural networks have numerous applications including porous media reconstruction (Q. Chen et al., 2022), reservoir characterization (R. Feng et al., 2022; Yang et al., 2022; Zhan et al., 2022), mineral evaluation (Jordão et al., 2022), subsurface flow and transport (Song et al., 2023; Zhong et al., 2019; Zhou et al., 2022), geophysical inversion (Lopez‐Alvis et al., 2021, 2022; Mosser et al., 2020), and remote sensing (Nesvold & Mukerji, 2021), among others. Generative adversarial network (GAN) is the one of the most commonly used neural network for characterizing hydrological structures and scenarios (Bao et al., 2022).…”

Section: Introductionmentioning

confidence: 99%

SA‐RelayGANs: A Novel Framework for the Characterization of Complex Hydrological Structures Based on GANs and Self‐Attention Mechanism

Cui,

Chen,

Liu

et al. 2024

Water Resources Research

View full text Add to dashboard Cite

Increased availability and quality of surface and subsurface observations provide the chance to improve the perception of hydrological phenomena. An important factor that hinders our understanding of hydrological structures is the characterization of heterogeneous patterns with continuous attributes inside the structures (e.g., porosity, permeability, fluid saturation, etc.). Unlike categorical attributes, continuous attributes convey more realistic characteristics but require more computational resources to characterize such complex earth systems. In this work, we propose a novel deep learning approach for the characterization of complex hydrological realism with continuous attributes based on generative adversarial networks (GANs) and self‐attention mechanism, named SA‐RelayGANs. To address the complexity of heterogeneous hydrological structures, we divide the modeling process into two stages: facies construction stage and property reconstruction stage. In the first stage, we employ an improved GAN with self‐attention mechanism to construct the heterogeneous structures while adhering to hard conditioning constraints. In the second stage, we utilize another GAN with an attribute enhancement term to reconstruct realizations based on the constructed structures and observations. SA‐RelayGANs can successfully predict the statistical distributions of heterogeneous structures with continuous attributes by using limited observations. This study highlights the effectiveness of using GANs to characterize the heterogeneous patterns of hydrological realism and the application over large geoscience fields.

show abstract

“…Deep generative modeling aims to train a generative model whose samples

\tilde{x}\sim {p}_{\theta }\left(\tilde{x}\right)

from the distribution learned from the training data

\tilde{x}\sim {p}_{\text{data}}\left(\tilde{x}\right)

Section: Introductionmentioning

confidence: 99%

SA‐RelayGANs: A Novel Framework for the Characterization of Complex Hydrological Structures Based on GANs and Self‐Attention Mechanism

Cui,

Chen,

Liu

et al. 2024

Water Resources Research

View full text Add to dashboard Cite

show abstract

“…In addition, integrated modeling procedures have been conducted to assist history matching in fractured reservoirs, with a history-matched model used to optimize well spacing for shale gas formations in China by including uncertainty in geological variables and economic feasibility 41 . In other recent studies, geological facies models have been constructed using self-attention generative adversarial networks 42 , 43 but these studies developed models based on synthetic reservoirs. A significant challenge for the simulation modeling of complex reservoirs is to effectively apply the resulting models to real field datasets.…”

Section: Introductionmentioning

confidence: 99%

Stochastic lithofacies and petrophysical property modeling for fast history matching in heterogeneous clastic reservoir applications

Al-Mudhafar,

Vo Thanh,

Wood

et al. 2024

Sci Rep

View full text Add to dashboard Cite

For complex and multi-layered clastic oil reservoir formations, modeling lithofacies and petrophysical parameters is essential for reservoir characterization, history matching, and uncertainty quantification. This study introduces a real oilfield case study that conducted high-resolution geostatistical modeling of 3D lithofacies and petrophysical properties for rapid and reliable history matching of the Luhais oil reservoir in southern Iraq. For capturing the reservoir's tidal depositional setting using data collected from 47 wells, the lithofacies distribution (sand, shaly sand, and shale) of a 3D geomodel was constructed using sequential indicator simulation (SISIM). Based on the lithofacies modeling results, 50 sets of porosity and permeability distributions were generated using sequential Gaussian simulation (SGSIM) to provide insight into the spatial geological uncertainty and stochastic history matching. For each rock type, distinct variograms were created in the 0° azimuth direction, representing the shoreface line. The standard deviation between every pair of spatial realizations justified the number of variograms employed. An upscaled version of the geomodel, incorporating the lithofacies, permeability, and porosity, was used to construct a reservoir-flow model capable of providing rapid, accurate, and reliable production history matching, including well and field production rates.

show abstract

“…Subsurface modeling and forecasting are essential components for optimum decision-making for hydrocarbon exploration and field development [1], hydrogeological resources exploration and efficient management [2,3], optimal injection and production well locations [4], and safe and reliable CO 2 injection and long-term sequestration [5]. For all of these field scale models, the common simulated spatial feature is discrete facies (or rock type), as it represents the vast majority of variance in spatial features like porosity and permeability that impact fluid flow [6,7]. Over smaller scales, e.g., individual rock pores, we are interested in directly reconstructing the pore spaces' geometry [8,9].…”

Section: Introductionmentioning

confidence: 99%

Minimum Acceptance Criteria for Subsurface Scenario-based Uncertainty Models from Single Image Generative Adversarial Networks (SinGAN)

Liu,

Salazar,

et al. 2024

Preprint

View full text Add to dashboard Cite

Evaluating and checking subsurface models is essential before their use to support optimum subsurface development decision making. Conventional geostatistical modeling workflows (e.g., two-point variogram-based geostatistics and multiple-point statistics) may fail to reproduce complex realistic geological patterns (e.g., channels), or be constrained by the limited training images and computational cost. Deep learning, specifically generative adversarial network (GAN), has been applied for subsurface modeling due to its ability to reproduce spatial and geological patterns, but may fail to reproduce commonly observed nonstationary subsurface patterns and often rely on many training images with the inability to explore realizations around specific geological scenarios. We propose an enhanced model checking workflow demonstrated by evaluating the performance of single image GAN (SinGAN)-based 2D image realizations for the case of channelized subsurface reservoirs to support robust uncertainty around geological scenarios. The SinGAN is able to generate nonstationary realizations from a single training image. Our minimum acceptance criteria expand on the work of Leuangthong, Boisvert, and others tailored to the nonstationary, single training image approach of SinGAN by evaluating the facies proportion, spatial continuity, and multiple-point statistics through histogram, semivariogram, and n-point histogram, along with evaluating the nonstationarity reproduction through multiple distribution checks ranging from local scale pixel distribution to multiscale local distribution. Additionally, our workflow incorporates reduced-dimensionality analysis through self-attention, providing a flexible approach for deep learning-based enhanced model realization to single training image comparison. With our proposed workflows, the robust application of SinGAN is possible to explore uncertainty around geological scenarios.

show abstract

Application of Bayesian Generative Adversarial Networks to Geological Facies Modeling

Cited by 17 publications

References 43 publications

SA‐RelayGANs: A Novel Framework for the Characterization of Complex Hydrological Structures Based on GANs and Self‐Attention Mechanism

SA‐RelayGANs: A Novel Framework for the Characterization of Complex Hydrological Structures Based on GANs and Self‐Attention Mechanism

Stochastic lithofacies and petrophysical property modeling for fast history matching in heterogeneous clastic reservoir applications

Minimum Acceptance Criteria for Subsurface Scenario-based Uncertainty Models from Single Image Generative Adversarial Networks (SinGAN)

Contact Info

Product

Resources

About