We present a research study aimed at testing of applicability of machine learning techniques for prediction of permeability of digitized rock samples. We prepare a training set containing 3D images of sandstone samples imaged with X-ray microtomography and corresponding permeability values simulated with Pore Network approach. We also use Minkowski functionals and Deep Learning-based descriptors of 3D images and 2D slices as input features for predictive model training and prediction. We compare predictive power of various feature sets and methods. The later include Gradient Boosting and various architectures of Deep Neural Networks (DNN). The results demonstrate applicability of machine learning for image-based permeability prediction and open a new area of Digital Rock research.
A multi-step steady-state (MSSS) method is proposed here for the measurement of low permeability. This new method can accurately and easily measure very low permeabilities of rock samples using a new setup, where the targeted rock sample and ordinary apparatus components are connected with a reference rock sample to form a series circuit. Any conventional rock sample with high permeability could be used as a reference rock sample such that the traditional steady-state measurement is feasible to accurately determine its permeability as a reference value in the MSSS method. The challenging measu…
Permeability is the key parameter for quantifying fluid flow in porous rocks. Knowledge of the spatial distribution of the connected pore space allows, in principle, to predict the permeability of a rock sample. However, limitations in feature resolution and approximations at microscopic scales have so far precluded systematic upscaling of permeability predictions. Here, we report fluid flow simulations in capillary network representations designed to overcome such limitations. Performed with an unprecedented level of accuracy in geometric approximation at microscale, the pore scale flow simul…
Permeability has a dominant influence on the flow properties of a natural fluid. Lattice Boltzmann simulator determines permeability from the nano and micropore network. The simulator holds millions of flow dynamics calculations with its accumulated errors and high consumption of computing power. To efficiently and consistently predict permeability, we propose a morphology decoder, a parallel and serial flow reconstruction of machine learning segmented heterogeneous Cretaceous texture from 3D micro computerized tomography and nuclear magnetic resonance images. For 3D vision, we introduce contr…
A multiporosity extension of classical double and triple porosity fractured rock flow models for slightly compressible fluids is presented. The multiporosity model is an adaptation of the multirate solute transport model of Haggerty and Gorelick (1995) to viscous flow in fractured rock reservoirs. It is a generalization of both pseudo-steady-state and transient interporosity flow double porosity models. The model includes a fracture continuum and an overlapping distribution of multiple rock matrix continua, whose fracture-matrix exchange coefficients are specified through a discrete probabilit…
Accurate modeling of gas relative permeability has practical applications in oil and gas exploration, production and recovery of unconventional reservoirs. In this study, we apply concepts from the effective-medium approximation (EMA) and universal power-law scaling from percolation theory. Although the EMA has been successfully used to estimate relative permeability in conventional porous media, to the best of our knowledge, its applications to unconventional reservoir rocks have not been addressed yet. The main objective of this study, therefore, is to evaluate the efficiency of EMA, in comb…
This paper discusses a continuum approach to track the evolution of permeability in granular rocks by accounting for the combined effect of porosity changes, grain breakage and cement bond damage. To account for such a broad range of microscopic processes under general loading paths, the Breakage Mechanics theory is used and the computed mechanical response is linked with the Kozeny equation, i.e. a permeability model able to evaluate the reduction of the hydraulic conductivity resulting from the simultaneous loss of porosity and growth of surface area. In particular, the evolution of the inte…
Effects of heterogeneity of mineral distribution and reaction rate on the rock dissolution process are investigated using a pore-scale reactive transport model based on the lattice Boltzmann method. Coupled fluid flow, species transport, chemical reaction and solid structure alternation due to dissolution are simulated. Effects of mineral distributions and chemical heterogeneity on the dissolution behaviors and evolutions of hydrologic properties are studied under different reactive transport conditions. Simulation results show that the coupling between advection, diffusion and reaction as wel…
A new and fast Matlab algorithm for predicting absolute permeability is presented. The developed tool relies on measuring the connectivity of pores in a given three-dimensional (3D) micro-CT rock image. An index of pore connectivity is introduced. After a calibration step, the developed index is used to estimate permeability in a variety of rocks with challenging pore structures (e.g. complex carbonate formations). The developed algorithm was tested on sandstone and carbonate rock samples. It offers large computational and memory savings when compared with algorithms based on the Lattice Boltz…
Shear deformation of partially molten rock in laboratory experiments causes the emergence of melt-enriched sheets (bands in cross-section) that are aligned at about 15-20 degrees to the shear plane. Deformation and deviatoric stress also cause the coherent alignment of pores at the grain scale. This leads to a melt-preferred orientation that may, in turn, give rise to an anisotropic permeability. Here we develop a simple, general model of anisotropic permeability in partially molten rocks. We use linearised analysis and nonlinear numerical solutions to investigate its behaviour under simple-sh…
Each grid block in a 3D geological model requires a rock type that represents all physical and chemical properties of that block. The properties that classify rock types are lithology, permeability, and capillary pressure. Scientists and engineers determined these properties using conventional laboratory measurements, which embedded destructive methods to the sample or altered some of its properties (i.e., wettability, permeability, and porosity) because the measurements process includes sample crushing, fluid flow, or fluid saturation. Lately, Digital Rock Physics (DRT) has emerged to quantif…
Stony Brook University（石溪大学）；Halicioglu Data Science Institute, University of California, San Diego（加州大学圣地亚哥分校哈利西奥格鲁数据科学研究所）；Columbia University（哥伦比亚大学）
We present a SE(3)-equivariant graph neural network (GNN) approach that directly predicting the formation factor and effective permeability from micro-CT images. FFT solvers are established to compute both the formation factor and effective permeability, while the topology and geometry of the pore space are represented by a persistence-based Morse graph. Together, they constitute the database for training, validating, and testing the neural networks. While the graph and Euclidean convolutional approaches both employ neural networks to generate low-dimensional latent space to represent the feat…
Apparent liquid permeability (ALP) in ultra-confined permeable media is primarily governed by the pore confinement and fluid-rock interactions. A new ALP model is required to predict the interactive effect of the above two on the flow in mixed-wet, heterogeneous nanoporous media. This study derives an ALP model and integrates the compiled results from molecular dynamics (MD) simulations, scanning electron microscopy, atomic force microscopy, and mercury injection capillary pressure. The ALP model assumes viscous forces, capillary forces, and liquid slippage in tortuous, rough pore throats. Pre…
The interconnectivity of the porous space is an important characteristic in the study of porous media and their transport properties. Hence we propose a way to quantify it and relate it with the intrinsic permeability of rocks. We propose a measure of connectivity based on geometric and topological information of pore-throat network, which are models built from microtomographic images, and we obtain an analytical method to compute that property. The method is expanded to handle rocks that present a higher degree of heterogenity in the porous space, which characterization requires images from d…
The presence of mesoscopic heterogeneities in fluid-saturated porous rocks can produce measurable seismoelectric signals due to wave-induced fluid flow between regions of differing compressibility. The dependence of these signals on the petrophysical and structural characteristics of the probed rock mass remains largely unexplored. In this work, we derive an analytical solution to describe the seismoelectric response of a rock sample, containing a horizontal layer at its center, that is subjected to an oscillatory compressibility test. We then adapt this general solution to compute the seismoe…
The permeability of a pore structure is typically described by stochastic representations of its geometrical attributes. Database-driven numerical solvers for large model domains can only accurately predict large-scale flow behaviour when they incorporate upscaled descriptions of that structure. The upscaling is particularly challenging for rocks with multimodal porosity structures such as carbonates, where several different types of structures are interacting. It is the connectivity both within and between these different structures that controls the porosity-permeability relationship at the …
Machine learning enhanced data assimilation framework for multiscale carbonate rock characterization
Institute of GeoEnergy Engineering, Heriot-Watt University（赫瑞-瓦特大学地质能源工程研究所）；Department of Geoscience & Engineering, Technical University of Delft（代尔夫特理工大学地球科学与工程系）；PETRONAS Research Sdn Bhd（马来西亚国家石油公司研究有限公司）
Direct pore scale simulations of two-fluid flow on digital rock images provide a promising tool to understand the role of surface wetting phenomena on flow and transport in geologic reservoirs. We present computational protocols that mimic conventional special core analysis laboratory (SCAL) experiments, which are implemented within the open source LBPM software package. Protocols are described to simulate unsteady displacement, steady-state flow at fixed saturation, and to mimic centrifuge experiments. These methods can be used to infer relative permeability and capillary curves, and otherwis…
Depleted oil fields in the Gulf of Suez (Egypt) can serve as geothermal reservoirs for power generation using a CO2-Plume Geothermal (CPG) system, while geologically sequestering CO2. This entails the injection of a substantial amount of CO2 into the highly permeable brine-saturated Nubian Sandstone. Numerical models of two-phase flow processes are indispensable for predicting the CO2-plume migration at a representative geological scale. Such models require reliable constitutive relationships, including relative permeability and capillary pressure curves. In this study, quasi-static pore-netwo…
Recovery of oil from oil shales and the natural primary migration of hydrocarbons are closely related processes that have received renewed interests in recent years because of the ever tightening supply of conventional hydrocarbons and the growing production of hydrocarbons from low permeability tight rocks. Quantitative models for conversion of kerogen into oil and gas and the timing of hydrocarbon generation have been well documented. However, lack of consensus about the kinetics of hydrocarbon formation in source rocks, expulsion timing and how the resulting hydrocarbons escape from or are …
In the context of a deep geological repository of high-level radioactive wastes, the French National Radioactive Waste Management Agency (Andra) has conducted an extensive characterization of the Callovo-Oxfordian argillaceous rock and surrounding formations in the Eastern Paris Basin. As part of this project, an accurate 3D seismic derived geological model is needed. The paper shows the procedure used for building the 3D seismic constrained geological model in depth by combining time-to-depth conversion of seismic horizons, consistent seismic velocity model and elastic impedance in time. It a…
Salt-Rock Creep Deformation Forecasting Using Deep Neural Networks and Analytical Models for Subsurface Energy Storage Applications
IIT (ISM) Dhanbad（印度理工学院（印度矿业学院）丹巴德分校）；Sorbonne University, Abu Dhabi（索邦大学阿布扎比校区）；Sorbonne Center for Artificial Intelligence, Sorbonne University（索邦大学人工智能中心）；CSIRO Energy（澳大利亚联邦科学与工业研究组织能源研究所）
Clays and claystones are considered in some countries (including Belgium, France and Switzerland) as a potential host rock for high activity long lived radioactive waste disposal at great depth. One of the aspects to deal with in performance assessment is related to the effects on the host rock of the temperature elevation due to the placement of exothermic wastes. The potential effects of the thermal impact on the excavated damaged zone in the close field are another important issue that was the goal of the TIMODAZ European research project. In this paper, some principles of waste disposal in…
Stochastic image reconstruction is a key part of modern digital rock physics and materials analysis that aims to create numerous representative samples of material micro-structures for upscaling, numerical computation of effective properties and uncertainty quantification. We present a method of three-dimensional stochastic image reconstruction based on generative adversarial neural networks (GANs). GANs represent a framework of unsupervised learning methods that require no a priori inference of the probability distribution associated with the training data. Using a fully convolutional neural …
ICSA, The University of Edinburgh（爱丁堡大学ICSA）；EPCC, The University of Edinburgh（爱丁堡大学EPCC）
Drilling boreholes for gas and oil extraction is an expensive process and profitability strongly depends on characteristics of the subsurface. As profitability is a key success factor, companies in the industry utilise well logs to explore the subsurface beforehand. These well logs contain various characteristics of the rock around the borehole, which allow petrophysicists to determine the expected amount of contained hydrocarbon. However, these logs are often incomplete and, as a consequence, the subsequent analyses cannot exploit the full potential of the well logs. In this paper we demonstr…
A Computationally Efficient Hybrid Neural Network Architecture for Porous Media: Integrating Convolutional and Graph Neural Networks for Improved Property Predictions
