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

Hou, Zhiyu; Cao, Danping; Ji, Siqi; Cui, Rongang; Liu, Qiang

doi:10.1016/j.cageo.2021.104939

Cited by 19 publications

(4 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On the other hand, while using binary images, the MAPE is lower: 2.19%, 3.04%, and 6.08% for predicting porosity, pore surface area, and throat area, respectively (Figure ). This is because the model generated with grayscale images is more sensitive to noise and defects, resulting in a higher prediction error than the model developed with binary images . Accurate rock image segmentation is essential for extracting the precise properties of the rock sample.…”

Section: Discussionmentioning

confidence: 99%

“…This is because the model generated with grayscale images is more sensitive to noise and defects, resulting in a higher prediction error than the model developed with binary images. 66 Accurate rock image segmentation is essential for extracting the precise properties of the rock sample. This study applied two different segmentation methods, the U-Net model and the Otsu thresholding method, to extract the properties.…”

Section: Discussionmentioning

confidence: 99%

“…The model presented in this work for predicting the morphological descriptors of porous media still has room for improvement. Methods like the generative adversarial network (GAN) 66 and hybrid stochastic deep-learning (HSDL) 67 can be applied to increase the image resolution and thus increase the prediction accuracy. In addition, the hyperparameters of the CNN model can also be tuned for better model performance.…”

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

Machine Learning Assisted Prediction of Porosity and Related Properties Using Digital Rock Images

Khan,

Khanal

2024

ACS Omega

View full text Add to dashboard Cite

Accurately estimating reservoir rock properties is paramount for modeling the storage and flow of fluids (hydrocarbon, carbon dioxide, and groundwater) in porous media. However, existing laboratory techniques to measure rock properties are usually time-consuming, expensive, and computationally intensive. This work proposes an efficient workflow that uses the machine learning algorithm, based on the convolutional neural network (CNN) framework, to predict rock properties from microcomputed tomography (micro-CT) X-ray images. The workflow involves data preprocessing, label extraction, training, and prediction using the segmented images of the rock to predict porosity, throat area, and pore surface area, which are essential for pore-scale modeling. The model was trained and validated on the Bentheimer sandstone, which was then used to predict properties of other sandstones (Castlegate and Leopard) with different pore structures and flow properties. The model yielded a good prediction for the throat and pore surface area but a significant error for porosity. Subsequently, a new complex model was trained and validated using diverse images from Bentheimer and an additional rock Castlegate, which was then used to predict the properties of Leopard sandstone. The new model improved the prediction of each property, resulting in mean absolute percentage error (MAPE) values of 2.19%, 3.04%, and 6.08% for porosity, pore surface area, and throat area with the binary images, respectively. In addition, we present a novel data-driven method using a simple regression model to predict the absolute permeability of a digital rock sample using the pore network parameters as predictors. The extreme gradient boost (XGBoost), which performed the best among several machine algorithms, was trained and validated using digital rock images from Bentheimer and Castlegate sandstone. The generated model was then used to predict the absolute permeability of the Leopard sandstone with an R 2 of 0.813, which was a significant improvement over the model generated solely by using either the Bentheimer or the Castlegate sandstone images. Furthermore, our analysis showed that the tortuosity had the most significant effect on the absolute permeability prediction of the rock sample. This study showed that we can reliably predict the morphological properties of porous media using computationally efficient models generated from digital rock images, which can be used to build a regression model to predict the crucial petrophysical properties needed to model the flow of fluids in porous media.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Machine Learning Assisted Prediction of Porosity and Related Properties Using Digital Rock Images

Khan,

Khanal

2024

ACS Omega

View full text Add to dashboard Cite

show abstract

“…In recent years, the petroleum industry has extensively employed artificial intelligence ( AI ) technology, owing to its outstanding predictive performance. This technology has been utilized for various tasks including reservoir history matching, image segmentation, − flow simulation and parameter prediction, − and super-resolution modeling. − AI algorithms have impressive nonlinear data prediction capabilities, enabling the prediction of oil phase utilization efficiency based on pore structure characteristics. This study combines XGBoost and Shapley additive explanation ( SHAP ) technology to establish a regression model between pore structure parameters and RCPOC and evaluates each parameter’s contribution to the absolute change of oil occupancy.…”

Section: Introductionmentioning

confidence: 99%

Flow Patterns and Pore Structure Effects on Residual Oil during Water and CO₂ Flooding: In Situ CT Scanning

Zhang,

Lin,

Ren

2023

Energy Fuels

View full text Add to dashboard Cite

Carbon dioxide (CO 2 ) enhanced oil recovery (EOR) is an important technology to achieve carbon neutrality by sequestering CO 2 underground while simultaneously recovering crude oil. Reservoir pore structure is a key factor influencing CO 2 EOR. In this study, we utilized advanced online in situ CT scanning and digital rock techniques to obtain, for the first time, evolution profiles of the finger area during water flooding and CO 2 flooding processes, quantitatively assessing the differences in fluid patterns. Additionally, we first introduced an innovative approach using advanced machine learning techniques, especially XGBoost and SHAP, to construct a predictive model of the relative change of oil phase occupancy (RCPOC) based on pore structure parameters and evaluated the importance of each pore structure parameter. Importantly, our results revealed that CO 2 can significantly increase the sweep efficiency area while substantially reducing residual oil saturation, in stark contrast to the relatively uniform water front observed during water flooding. Furthermore, we elucidated the critical role of capillary forces, demonstrating that water flooding primarily extracts trapped oil from small pores, while CO 2 flooding effectively extracts oil from larger pores. During CO 2 flooding, there is a positive correlation between coordination number, mean throat radius (MeanTR), and mean throat length (MeanTL) and the change in oil occupancy, whereas their influence during water flooding is limited. In summary, this study contributes to the understanding of flow patterns and pore structure effects on residual oil during water flooding and the CO 2 flooding processes. It also provides a novel approach based on pore structure parameters to predict RCPOC and assess the importance of influencing factors, thereby expanding our research perspective on this issue.

show abstract

Intelligent rock fracture identification based on image semantic segmentation: methodology and application

Pan

Lin

et al. 2023

Environ Earth Sci

View full text Add to dashboard Cite

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

Cited by 19 publications

References 18 publications

Machine Learning Assisted Prediction of Porosity and Related Properties Using Digital Rock Images

Machine Learning Assisted Prediction of Porosity and Related Properties Using Digital Rock Images

Flow Patterns and Pore Structure Effects on Residual Oil during Water and CO₂ Flooding: In Situ CT Scanning

Intelligent rock fracture identification based on image semantic segmentation: methodology and application

Contact Info

Product

Resources

About

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

Cited by 19 publications

References 18 publications

Machine Learning Assisted Prediction of Porosity and Related Properties Using Digital Rock Images

Machine Learning Assisted Prediction of Porosity and Related Properties Using Digital Rock Images

Flow Patterns and Pore Structure Effects on Residual Oil during Water and CO2 Flooding: In Situ CT Scanning

Intelligent rock fracture identification based on image semantic segmentation: methodology and application

Contact Info

Product

Resources

About

Flow Patterns and Pore Structure Effects on Residual Oil during Water and CO₂ Flooding: In Situ CT Scanning