A Dataset of 3D Structural and Simulated Transport Properties of Complex Porous Media

Abstract

Physical processes that occur within porous materials have wide-ranging applications including - but not limited to - carbon sequestration, battery technology, membranes, oil and gas, geothermal energy, nuclear waste disposal, water resource management. The equations that describe these physical processes have been studied extensively; however, approximating them numerically requires immense computational resources due to the complex behavior that arises from the geometrically-intricate solid boundary conditions in porous materials. Here, we introduce a new dataset of unprecedented scale and breadth, DRP-372: a catalog of 3D geometries, simulation results, and structural properties of samples hosted on the Digital Rocks Portal. The dataset includes 1736 flow and electrical simulation results on 217 samples, which required more than 500 core years of computation. This data can be used for many purposes, such as constructing empirical models, validating new simulation codes, and developing machine learning algorithms that closely match the extensive purely-physical simulation. This article offers a detailed description of the contents of the dataset including the data collection, simulation schemes, and data validation.

Learning a general model of single phase flow in complex 3D porous media

“…In this work, we have implemented the following enhancements to the previous model: • Input features: to capture simulations encompassing diverse thermodynamic conditions, we introduce a spatially-varying 3D normalized MFP in our model. This feature, which arises from the Peng-Robinson equation, accounts for the reduction of the local kinetic viscosity caused by the confinement effects at the nanoscale (the detailed explanation can be found in [35]) 2 . This input feature captures the resolution of the various samples, which is crucial when working with a heterogeneous dataset of geological materials.…”

“…For a thorough review of the dataset, we refer the reader to [35]. The DRP-372 dataset includes more than 1200 flow simulations from a wide variety of samples hosted in the Digital Rocks Portal [36].…”

“…All the data used to train and test our model is described in [35] and it is openly-accessible at www.digitalrocksportal.org/projects/372.…”

“…For the first time, we used hundreds of computationally large flow simulations in synthetic samples to train our model, including a wide range of pore sizes, rock types, and boundary conditions. The comprehensive dataset used for all the training and testing data are publicly available [35]. To accommodate the larger model and dataset, we likewise applied a number of computer science tools to ameliorate the increased computational cost, which we detail.…”

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

“…From DRP, micro‐CT non‐destructively images the internal 3D microstructure of porous rocks (J. Liu & Regenauer‐Lieb, 2021; Ma et al., 2021). Currently, 3D digital rocks can be constructed through commercial or open‐source codes (Prifling et al., 2021; Sahimi & Tahmasebi, 2021; Santos et al., 2022). Then, 3D properties can be calculated based on the 3D digital model through many numerical methods (Andrä et al., 2013; Sahimi & Tahmasebi, 2021), such as lattice Boltzmann method, finite volume method (LVM), pore network model (PNM) for permeability, formation factor, Young's modulus (J. Liu et al., 2016).…”