Abstract: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.

Intelligent Digital Rock Physics Assisting Quantitative Seismic Interpretation

High-precision Digital Rock Construction and Electrical Property Upscaling in Tight Sandstone

An Intelligent Method for Reconstructing Large-Size Digital Rocks by Joining Multi-Dimension Information

A Method to Construct High-Precision Complex Pore Digital Rock

Rock Physical Modeling of Tight Sandstones Based on Digital Rocks and Reservoir Porosity Prediction From Seismic Data

A Rapid Reconstruction Method of 3D Digital Rock with Strong Pore Connectivity

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

Introduction to Digital Rock Physics and Predictive Rock Properties of Reservoir Sandstone

Multi-task learning for digital rock segmentation and characteristic parameters computation

Direct Prediction of Relative Permeability Curve from 3D Digital Rock Images Based on Deep Learning Approaches

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

Digital Rock Physics of Sandstone Based on Micro-Ct Technology

Process-Based Reconstruction of Digital Rock Based on Discrete Element Method Considering Thermal-Mechanical Coupling Effectand Actual Particle Shape

A New Multi-Scale Method to Evaluate the Porosity and MICP Curve for Digital Rock of Complex Reservoir

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

Relative permeability curve prediction from digital rocks with variable sizes using deep learning

Controllable image expansion of rock castings based on deep learning

Analysis of reconstructed multisource and multiscale 3-D digital rocks based on the cycle-consistent generative adversarial network method

Predictive Soft Computing Methods for Building Digital Rock Models Verified by Positron Emission Tomography Experiments

Application of improved and efficient image repair algorithm in rock damage experimental research

Investigating the causes of permeability anisotropy in heterogeneous conglomeratic sandstone using multiscale digital rock