Abstract:Abstract The permeability of a pore structure is typically described by stochastic representations of its geometrical attributes (e.g. pore-size distribution, porosity, coordination number). Database-driven numerical solvers for large model domains can only accurately predict large-scale flow behavior 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 type of structures (e.g. micro-porosity, cavities, fractures) are interacting. It is the connectivity both within and between these fundamentally different structures that ultimately controls the porosity–permeability relationship at the larger length scales. Recent advances in machine learning techniques combined with both numerical modelling and informed structural analysis have allowed us to probe the relationship between structure and permeability much more deeply. We have used this integrated approach to tackle the challenge of upscaling multimodal and multiscale porous media. We present a novel method for upscaling multimodal porosity–permeability relationships using machine learning based multivariate structural regression. A micro-CT image of Estaillades limestone was divided into small 60 3 and 120 3 sub-volumes and permeability was computed using the Darcy–Brinkman–Stokes (DBS) model. The microporosity–porosity–permeability relationship from Menke et al. (Earth Arxiv, https://doi.org/10.31223/osf.io/ubg6p , 2019) was used to assign permeability values to the cells containing microporosity. Structural attributes (porosity, phase connectivity, volume fraction, etc.) of each sub-volume were extracted using image analysis tools and then regressed against the solved DBS permeability using an Extra-Trees regression model to derive an upscaled porosity–permeability relationship. Ten test cases of 360 3 voxels were then modeled using Darcy-scale flow with this machine learning predicted upscaled porosity–permeability relationship and benchmarked against full DBS simulations, a numerically upscaled Darcy flow model, and a Kozeny–Carman model. All numerical simulations were performed using GeoChemFoam, our in-house open source pore-scale simulator based on OpenFOAM. We found good agreement between the full DBS simulations and both the numerical and machine learning upscaled models, with the machine learning model being 80 times less computationally expensive. The Kozeny–Carman model was a poor predictor of upscaled permeability in all cases.

Upscaling an Extended Heterogeneous Stefan Problem from the Pore-Scale to the Darcy Scale in Permafrost

Hierarchical Multiscale Modeling of Fluid-Saturated Soils

Parallel Hierarchical Multiscale Modelling of Hydro-Mechanical Problems for Saturated Granular Soils

Upscaling the porosity–permeability relationship of a microporous carbonate for Darcy-scale flow with machine learning

A heterogeneous multiscale MPFA method for single-phase flows in porous media with inertial effects

Non-local modelling of freezing and thawing of unsaturated soils

Permeability up-scaling using Haar Wavelets

Transient and Transition Factors in Modeling Permafrost Thaw and Groundwater Flow

Pore-to-meter Scale Modeling of Heat and Mass Transport Applied to Thermal Energy Storage: How Local Thermal and Velocity Fluctuations Affect Average Thermal Dispersivity

Unified analysis of algorithms for equilibrium, non-equilibrium, and hysteresis models of phase transition in permafrost

Meshfree multiscale method for the infiltration problem in permafrost

Upscaling and Effective Behavior for Two-Phase Porous-Medium Flow Using a Diffuse Interface Model

Thermo‐hydro‐mechanical coupled material point method for modeling freezing and thawing of porous media

The Method of Finite Averages: A rigorous upscaling methodology for heterogeneous porous media

Convection in the active layer speeds up permafrost thaw in coarse-grained soils

Improved Stefan Equation Correction Factors to Accommodate Sensible Heat Storage during Soil Freezing or Thawing

Non-equilibrium modeling hysteresis of water freezing: Ice thawing in partially saturated porous building materials

Impact of lateral groundwater flow on hydrothermal conditions of the active layer in a high arctic hillslope setting

Homogenization of Two-Phase Flow in Porous Media From Pore to Darcy Scale: A Phase-Field Approach

Large Eddy Simulations of the Interaction Between the Atmospheric Boundary Layer and Degrading Arctic Permafrost