A CNN-based approach for upscaling multiphase flow in digital sandstones

Siavashi, Javad; Najafi, Atefeh; Ebadi, Mohammad; Sharifi, Mohammad

doi:10.1016/j.fuel.2021.122047

Cited by 26 publications

(11 citation statements)

References 80 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, artificial intelligence (AI) has been widely used in the oil industry due to its outstanding performance, including development effect evaluation [77][78][79], data segmentation [80][81][82], and flow simulation [83,84]. In this study, the SOM machine learning method [55,85] was used to automatically classify the residual oil pattern.…”

Section: 42mentioning

confidence: 99%

Effects of Wettability and Minerals on Residual Oil Distributions Based on Digital Rock and Machine Learning

Zhang

Lin

et al. 2022

Lithosphere

View full text Add to dashboard Cite

The wettability of mineral surfaces has significant impacts on transport mechanisms of two-phase flow, distribution characteristics of fluids, and the formation mechanisms of residual oil during water flooding. However, few studies have investigated such effects of mineral type and its surface wettability on rock properties in the literature. To unravel the dependence of hydrodynamics on wettability and minerals distribution, we designed a new experimental procedure that combined the multiphase flow experiments with a CT scan and QEMSCAN to obtain 3D digital models with multiple minerals and fluids. With the aid of QEMSCAN, six mineral components and two fluids in sandstones were segmented from the CT data based on the histogram threshold and watershed methods. Then, a mineral surface analysis algorithm was proposed to extract the mineral surface and classify its mineral categories. The in situ contact angle and pore occupancy were calculated to reveal the wettability variation of mineral surface and distribution characteristics of fluids. According to the shape features of the oil phase, the self-organizing map (SOM) method, one of the machine learning methods, was used to classify the residual oil into five types, namely, network, cluster, film, isolated, and droplet oil. The results indicate that each mineral’s contribution to the mineral surface is not proportional to its relative content. Feldspar, quartz, and clay are the main minerals in the studied sandstones and play a controlling role in the wettability variation. Different wettability samples show various characteristics of pore occupancy. The water flooding front of the weakly water-wet to intermediate-wet sample is uniform, and oil is effectively displaced in all pores with a long oil production period. The water-wet sample demonstrates severe fingering, with a high pore occupancy change rate in large pores and a short oil production period. The residual oil patterns gradually evolve from networks to clusters, isolated, and films due to the effects of snap-off and wettability inversion. This paper reveals the effects of wettability of mineral surface on the distribution characteristics and formation mechanisms of residual oil, which offers us an in-deep understanding of the impacts of wettability and minerals on multiphase flow and helps us make good schemes to improve oil recovery.

show abstract

Section: 42mentioning

confidence: 99%

Effects of Wettability and Minerals on Residual Oil Distributions Based on Digital Rock and Machine Learning

Zhang

Lin

et al. 2022

Lithosphere

View full text Add to dashboard Cite

show abstract

“…The difference between the pooling and convolution layers is that the former does not activate the function [27]. convolution and pooling layers can be repeatedly combined within the hidden layers as per actual application needs [28]. The fully connected layer is connected after the convolution layer and pooling layer.…”

Section: B Functional Modules Of the Traffic Iasmentioning

confidence: 99%

Designing Electronic Traffic Information Acquisition System Using Deep Learning and Internet of Things

Han

2022

IEEE Access

View full text Add to dashboard Cite

In implementing the Intelligent Traffic Monitoring System (ITMS), timely and effective access to road traffic information is an essential link. It requires an effective traffic Information Acquisition System (IAS) to collect real-time data and transmit the collected information to the background for processing. Therefore, this paper studies on-road vehicle information recognition based on Deep Learning (DL). Firstly, a framework of traffic IAS is proposed. Then, an improved MT-GooGleNet model based on Convolutional Neural Network (CNN) is proposed to locate and recognize vehicles in traffic images. Finally, the performance of the model is analyzed by simulation. The experimental results of vehicle position recognition show that the classification accuracy of Multi-Task (MT)-GooGleNet after fine-tuning is 99.5%. Compared with other models, the MT-GooGleNet model proposed is the best in vehicle position recognition, and its positioning accuracy is very high. The results of vehicle identification show that after data enhancement and pre-training, the testing set accuracy of the MT-GooGleNet model is 79.96%. The results show that the model's accuracy has been dramatically improved after processing. The research provides a reference for establishing IAS in the future INDEX TERMS deep learning; Internet of Things; signal acquisition; system design; vehicle identification and positioning

show abstract

“…However, acquiring 3D micro‐structural data is currently expensive both experimentally and computationally. In addition, using 3D data volume as input, although CNN has good accuracy and relatively low computational cost compared to numerical methods, CNN still requires high computational costs during training (Siavashi et al., 2022). Meanwhile, there is a contradiction that the resolution needs to be sacrificed to expand the scanning field of view (Najafi et al., 2021; Siavashi et al., 2022).…”

Section: Introductionmentioning

confidence: 99%

“…In addition, using 3D data volume as input, although CNN has good accuracy and relatively low computational cost compared to numerical methods, CNN still requires high computational costs during training (Siavashi et al., 2022). Meanwhile, there is a contradiction that the resolution needs to be sacrificed to expand the scanning field of view (Najafi et al., 2021; Siavashi et al., 2022). In this regard, 2D slices are cheaper and more available experimentally, compare with 3D structured data volumes.…”

Section: Introductionmentioning

confidence: 99%

Predicting 3D Physical Properties From a Single 2D Slice Based on Convolutional Neural Networks: 2D‐Slice‐To‐3D‐Properties for Porous Rocks

Chen,

Yang,

et al. 2023

Water Resources Research

View full text Add to dashboard Cite

The 3D physical properties of porous rocks directly determine the subsurface flow and modelling. However, predicting a wide range of 3D physical rocks remains a formidable challenge and requires a large amount of input. Reliable microstructure‐property correlations can accurately predict 3D physical properties, avoiding time‐consuming experimental testing. Here, we propose a new dimensionless CNN‐based method to find inherent correlations for accurately predicting 3D properties of porous rocks using only a single 2D slice or a series of 2D slices. Training and testing were conducted on 9 properties of 16,409 semi‐realistic 3D porous rocks. Eight properties, except for the formation factor, were predicted with an acceptable correlation coefficient of 0.92. Furthermore, the modeling results indicate that the network architecture of the model has a significant impact on its performance. The classical CNN Model 4 achieved an average validating loss of 1.0×10‐3. Interval 2D slice sampling can significantly reduce the computational cost in training model. The developed model was verified and tested by five independent porous samples. Additionally, the proposed method accurately predicted permeability in different directions using only 2D slices trained in one direction, for porous rocks where anisotropy is not significant. This work demonstrates the ability to predict various 3D properties rapidly and accurately using only a single 2D slice from semi‐realistic 3D volumes. It provides insights into the conversion of 2D slices to 3D properties for porous media in a wide range of applications, including property estimation from 2D images without structural construction and simulation in 3D.This article is protected by copyright. All rights reserved.

show abstract

A CNN-based approach for upscaling multiphase flow in digital sandstones

Cited by 26 publications

References 80 publications

Effects of Wettability and Minerals on Residual Oil Distributions Based on Digital Rock and Machine Learning

Effects of Wettability and Minerals on Residual Oil Distributions Based on Digital Rock and Machine Learning

Designing Electronic Traffic Information Acquisition System Using Deep Learning and Internet of Things

Predicting 3D Physical Properties From a Single 2D Slice Based on Convolutional Neural Networks: 2D‐Slice‐To‐3D‐Properties for Porous Rocks

Contact Info

Product

Resources

About