Conditioning generative adversarial networks on nonlinear data for subsurface flow model calibration and uncertainty quantification

Razak, Syamil Mohd; Jafarpour, Behnam

doi:10.1007/s10596-021-10112-8

Cited by 13 publications

(3 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…GANs are a class of unsupervised machine learning methods which can learn to generate new formatted data with the same statistics as the training set. Motivated by successful applications of GANs for modeling channelized structures for reservoir-simulation workflows [16,17,18,19,20], we use a GAN for efficient earth modeling.…”

Section: Earth Modeling Using Ganmentioning

confidence: 99%

Verification of a real-time ensemble-based method for updating earth model based on GAN

Fossum¹,

Alyaev²,

Tveranger³

et al. 2022

Preprint

View full text Add to dashboard Cite

The complexity of geomodelling workflows is a limiting factor for quantifying and updating uncertainty in real-time during drilling. We propose Generative Adversarial Networks (GANs) for parametrization and generation of geomodels, combined with Ensemble Randomized Maximum Likelihood (EnRML) for rapid updating of subsurface uncertainty. This real-time ensemble method combined with a highly non-linear model arising from neural-network modeling sequences might produce inaccurate and/or biased posterior solutions. This paper illustrates the predictive ability of EnRML on several examples where we assimilate local extra-deep electromagnetic logs. Statistical verification with MCMC confirms that the proposed workflow can produce reliable results required for geosteering wells.

show abstract

Section: Earth Modeling Using Ganmentioning

confidence: 99%

Verification of a real-time ensemble-based method for updating earth model based on GAN

Fossum¹,

Alyaev²,

Tveranger³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…K. Zhang et al (2022) presented a GAN-based geological modeling approach with a multi-source information fusion strategy. In addition to the characterization of subsurface structures, Razak and Jafarpour (2022) presented two approaches for generating subsurface flow models by using conditional GAN (Mirza & Osindero, 2014). Moreover, some studies have demonstrated that employing an ensemble-based strategy can enhance the performance of hydrogeological structures generated by GAN-based approaches (Abadpour et al, 2018;Fossum et al, 2022).…”

mentioning

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

“…Maybe the first work which applied generative adversarial networks to the generation of geological facies was performed by Chan and Elsheikh (2017), with an extended version presented later (CHAN; ELSHEIKH, 2019a). This work FENG et al, 2022;RAZAK;JAFARPOUR, 2022;ZHANG et al, 2022).…”

Section: Parameterizationmentioning

confidence: 90%

Conjunto suavizado iterativo com localização adaptativa e parametrização com aprendizado profundo em reservatórios carbonáticos.

Ranazzi

View full text Add to dashboard Cite

I am deeply grateful to my advisor, Prof. Marcio Sampaio, for his continuous support, guidance, and unwavering commitment to my academic and professional development. His expertise, invaluable insights, and willingness to share his knowledge have significantly influenced the outcome of this research. I would like to extend my sincere thanks to the Energy Production Innovation Center (EPIC), FAPESP and Equinor for their generous financial support throughout the course of this research. Their investment in this project has enabled me to pursue in-depth investigations and achieve meaningful results. I am grateful to the UNISIM group for providing the simulation models that were crucial for conducting this study. Their contribution has greatly enhanced the accuracy and reliability of the research outcomes. Additionally, I would like to acknowledge the Computer Modelling Group (CMG) for providing the essential licenses of the reservoir flow simulator. Lastly, I want to express my gratitude to the Polytechnic School of the University of São Paulo (USP) and the Laboratory of Petroleum Reservoir Simulation and Management (LASG) for providing the necessary infrastructure and resources that facilitated the progress of this research. Their commitment to academic excellence and dedication to fostering a conducive research environment have been invaluable. "Explore the world. Nearly everything is really interesting if you go into it deeply enough."

show abstract

Conditioning generative adversarial networks on nonlinear data for subsurface flow model calibration and uncertainty quantification

Cited by 13 publications

References 61 publications

Verification of a real-time ensemble-based method for updating earth model based on GAN

Verification of a real-time ensemble-based method for updating earth model based on GAN

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

Conjunto suavizado iterativo com localização adaptativa e parametrização com aprendizado profundo em reservatórios carbonáticos.

Contact Info

Product

Resources

About