arXiv subjects

Martin J. Blunt

Publications and source records attributed to Martin J. Blunt.

16 recordsLinked to original sources

arXiv2026

Multiphase Reactive Transport in a Heterogeneous Flow Field: Channel Formation in Hydrocarbon-Bearing Carbonate Rock

Carbon capture, utilization and storage (CCUS) is a key approach for reducing anthropogenic CO2 emissions. Currently the vast majority of CO2 stored is injected into depleted hydrocarbon reservoirs. Reactive transport in hydrocarbon-bearing carbonate reservoirs is controlled not only by the balance between reactant delivery and surface reaction, but also by the heterogeneous pore-scale flow field created by pore structure and fluid distribution. Here, we used time-resolved micro-CT imaging, pore-network analysis, and direct numerical simulation to investigate channelized dissolution during injection of CO2-saturated brine into oil-bearing Ketton limestone at 0.5 mL/min. The system remained in a high Pe, low Da regime throughout the 180 min injection. Despite the advection-dominated flow regime, dissolution became strongly localized. Preferential flow pathways were already present originally because the heterogeneous pore structure and remaining-oil occupancy restricted brine flow to a subset of the connected pore space. Dissolution progressively amplified these pathways. The associated reduction in hydraulic resistance further focused flow, producing a continuous tortuous channel by 180 min. Meanwhile, the effective reaction rate was only 1.6*10^5 mol/m2s,, approximately one order of magnitude lower than the batch reaction rate, showing that rapid advective transport did not translate directly into rapid overall dissolution under multiphase conditions. These results demonstrate that channel formation arose from the coupling between strong reactant delivery and pore-scale flow-field heterogeneity, rather than from the bulk Pe Da regime alone. Accounting for flow-field heterogeneity is therefore important for predicting reactive transport and pore-structure evolution during CO2 injection into hydrocarbon-bearing carbonate reservoirs.

physics.flu-dyn↗

arXiv2026

Dynamic Evolution of Pore-scale Heterogeneity and Transport Conditions Control Mineral Dissolution Regimes

Mineral dissolution in porous media is classically partitioned into static regimes within the Pe-Da plane, but this framework fails to capture the dissolution behavior of structurally complex rocks. Using three-dimensional micro-continuum simulations on micro-CT images of three rock samples spanning a wide range of pore-space heterogeneity, we track the joint evolution of dissolution morphology, velocity distribution, and reaction rate. Our results reveal that initial flow heterogeneity controls accessibility of reactants, thereby controlling the dissolution regime,reshaping them as dynamic trajectories. Channeled dissolution emerges as a simultaneous reorganization of structure and flow, and the resulting permeability-porosity relationship cannot be captured by a single power-law. The effective power-law exponent increases with heterogeneity and changes over time, reaching a maximum of 9.8, 18.0, and 40.9 for the three samples. Consequently, the effective reaction rate falls one to three orders of magnitude below the uniform dissolution prediction, with the suppression scaling with flow heterogeneity due to mass transfer limitations in channeled dissolution.

physics.flu-dyn↗

arXiv2026

Intermittent two-phase flow in porous media: insights from pore-scale direct numerical simulation

Recent X-ray imaging experiments have revealed that multiphase flow through porous media involves transient fluctuations in local occupancy, even under fixed macroscopic steady-state conditions where capillary forces dominate at the pore scale. To examine how intermittency manifests at the pore scale we perform direct numerical finite volume simulations (DNS) of immiscible two-phase flow through a micro-CT-derived Bentheimer sandstone geometry at capillary numbers in the Darcy and intermittent flow regimes. We show that intermittent disconnection and reconnection are accompanied by strongly coupled local pressure redistribution and non-wetting phase flow. This behaviour contrasts with the Darcy flow regime, in which the phases remain predominantly in fixed pathways. Macroscopically the computed pressure-gradient-capillary-number relationship ($\nabla P$-Ca) recovers both the linear Darcy and the sub-linear intermittent scaling regimes consistent with previous experimental measurements. We show how an increase in intermittency leads to the transition from the linear to the sub-linear regime. Using topology-aware snap-off detection, we show that the spatial extent of intermittency increases with capillary number. Spectral, local-geometry, and network-connectivity analyses provide further evidence that the intermittent elements organise into connected conduits embedded within a stable backbone of fixed flow pathways: intermittency is a network-coupled rather than purely local process. This work characterises the pore-scale manifestation of intermittency as a periodic sequence of drainage and imbibition displacements triggered by local pressure fluctuations whose macroscopic consequence is to improve the overall mobility of the fluid phases.

physics.flu-dyn↗

arXiv2026

Time-Resolved Pore-Scale Imaging of Multiphase Dissolution during CO2-Saturated Brine Injection into a Carbonate: Competition between Hydrocarbon Mobilisation and Swelling

We present time-resolved pore-scale experiments in which CO2-saturated brine was injected into a water-wet Ketton limestone sample containing residual hydrocarbon under reservoir conditions (8 MPa, 50 °C) and monitored by 4D X-ray microtomography. Equivalent pore-network models were extracted at each scan time to track pore geometry, topology, and fluid occupancy, while fluid-fluid and fluid-rock interfacial areas and the effective reaction rate were determined from segmented images. The dissolution rate is non-monotonic in time and proceeds through three regimes, consistent with a shifting balance between hydrocarbon swelling and ganglion mobilisation, which control advective access to reactive surfaces. In the initial advection-dominated regime, pore-throat widening leads to ganglia mobilisation and efficient acidic brine delivery to reactive surfaces. The second, dissolution-inhibited regime is marked by up to two orders of magnitude reduction in effective reaction rate. Pore-network analysis shows that swollen hydrocarbon ganglia persistently occupy the largest throats throughout this regime. This occupancy is associated with a reorganisation of the advective flow field into preferential flow paths and stagnant zones. We interpret the rate suppression as primarily reflecting a path-dependent loss of advective access to reactive surfaces, with subordinate contributions from localised H+ depletion near ganglia and reduced near-wall mass transfer in widened flow paths. The inhibited state persists until hydrocarbon is displaced from the largest throats, after which, in the third stage, advective access improves and rock dissolution accelerates. These results show that the effective dissolution rate in residual-hydrocarbon-bearing carbonate depends dynamically on the competition between hydrocarbon swelling and ganglion mobilisation, governing advective access to surfaces.

physics.flu-dyn↗

arXiv2026

Time-dependent pore-network modelling of Ostwald ripening in porous media

We present a time-dependent pore-network model that couples transient mass transfer in the aqueous phase, capillary pressure heterogeneity, and realistic pore-throat geometries to capture the dynamic evolution of gas clusters during Ostwald ripening in porous media. The model is applied to Bentheimer sandstone to study Ostwald ripening after imbibition to residual gas saturation. Both imbibition (shrinkage) and drainage (growth) events occur as the local capillary pressure in trapped gas clusters approaches equilibrium. The model tracks event statistics, capillary pressure equilibration, cluster volume distributions, and spatial saturation profiles over 48 hours. While the volume-weighted average capillary pressure is constant, there is a rapid initial decline in average number-weighted cluster pressure and a shift in cluster size distributions toward fewer, larger ganglia, consistent with pore-scale imaging studies. Pore and throat occupancy analysis reveal persistent gas trapping in larger pore spaces. Since growth is by drainage, the pore-scale configuration of fluid is different from that predicted by an equilibrium percolation-without-trapping model that only allows imbibition events. The model reproduces displacement and ganglion rearrangement during time-limited laboratory experiments, and can then provide predictions of trapped saturation, relative permeability and capillary pressure under field-scale conditions with application to hydrogen, natural gas and carbon dioxide storage in the subsurface.

cond-mat.soft↗

arXiv2026

SAMamba3D: adapting Segment Anything for generalizable 3D segmentation of multiphase pore-scale images

Reliable segmentation of multiphase pore-scale X-ray images of rocks is necessary to quantify fluid saturation, connectivity, and interfacial geometry. However, current 3D segmentation methods are typically dataset-specific, requiring retraining or extensive fine-tuning whenever rock type, fluid pattern, scanner, or acquisition conditions change. Foundation models such as the Segment Anything Model (SAM) provide strong 2D boundary priors, but they are not directly applicable to 3D data. We present SAMamba3D, a parameter-efficient framework that adapts a largely frozen SAM encoder to generalizable 3D pore-scale segmentation by coupling it with Mamba-based volumetric context modeling and progressive cross-scale feature interaction. For sandstone and carbonate datasets, with different fluids, wettability, and scanning conditions, SAMamba3D matches or outperforms current 3D baselines while reducing the need for case-specific retraining. The resulting segmented images preserve physically meaningful descriptors, including fluid saturation, connectivity, and interface morphology, enabling more reliable and rapid analysis of large 3D multiphase images.

cs.CV↗

arXiv2026

Learning Pore-scale Multiphase Flow from 4D Velocimetry

Multiphase flow in porous media underpins subsurface energy and environmental technologies, including geological CO$_2$ storage and underground hydrogen storage, yet pore-scale dynamics in realistic three-dimensional materials remain difficult to characterize and predict. Here we introduce a multimodal learning framework that infers multiphase pore-scale flow directly from time-resolved four-dimensional (4D) micro-velocimetry measurements. The model couples a graph network simulator for Lagrangian tracer-particle motion with a 3D U-Net for voxelized interface evolution. The imaged pore geometry serves as a boundary constraint to the flow velocity and the multiphase interface predictions, which are coupled and updated iteratively at each time step. Trained autoregressively on experimental sequences in capillary-dominated conditions ($Ca\approx10^{-6}$), the learned surrogate captures transient, nonlocal flow perturbations and abrupt interface rearrangements (Haines jumps) over rollouts spanning seconds of physical time, while reducing hour-to-day--scale direct numerical simulations to seconds of inference. By providing rapid, experimentally informed predictions, the framework opens a route to ''digital experiments'' to replicate pore-scale physics observed in multiphase flow experiments, offering an efficient tool for exploring injection conditions and pore-geometry effects relevant to subsurface carbon and hydrogen storage.

cs.LG↗

arXiv2025

Pore-Scale Dynamics of Multiphase Reactive Transport in Water-Wet Carbonates under CO2-Acidified Brine Injection: Dissolution Patterns and Reaction Rates

Depleted carbonate hydrocarbon reservoirs are promising sites for geological CO2 storage, yet the presence of residual hydrocarbons introduces complex pore-scale interactions that influence the dynamics of solid dissolution. This study reveals how residual oil affects dissolution patterns and effective reaction rates during CO2-acidified brine injection into Ketton limestone under reservoir conditions. We combine time-resolved X-ray microtomography (micro-CT), core-flooding experiments, and direct numerical simulations to assess the impact of pore space heterogeneity, oil distribution and injection rate. We find that the coupling between pore structure, residual oil saturation and oil displacement control flow heterogeneity, reactive surface accessibility, dissolution patterns and the reaction rates. At low injection rate, dissolution by channel widening is enhanced by oil displacement. This mechanism is especially important when dissolution is suppressed by heterogeneity in the pore space and the residual oil. At high injection rates, a more uniform dissolution occurs and can be enhanced by re-mobilisation of oil blocking brine flow. Effective reaction rates in two-phase flow are found to be lower than in the the equivalent single-phase case and up to two orders of magnitude lower than the batch rates due to persistent transport limitations. These findings offer mechanistic insights into multiphase reactive transport in carbonates and highlight the need for accurate understanding of the impact of the hydrocarbon phase on reaction to improve predictions of CO2 storage efficiency.

physics.flu-dyn↗

arXiv2024

Percolation without trapping: how Ostwald ripening during two-phase displacement in porous media alters capillary pressure and relative permeability

Conventional measurements of two-phase flow in porous media often use completely immiscible fluids, or are performed over time-scales of days to weeks. If applied to the study of gas storage and recovery, these measurements do not properly account for Ostwald ripening, significantly over-estimating the amount of trapping and hysteresis. When there is transport of dissolved species in the aqueous phase, local capillary equilibrium is achieved: this may take weeks to months on the centimetre-sized samples on which measurements are performed. However, in most subsurface applications where the two phases reside for many years, equilibrium can be achieved. We demonstrate that in this case, two-phase displacement in porous media needs to be modelled as percolation without trapping. A pore network model is used to quantify how to convert measurements made ignoring Ostwald ripening to correct trapped saturation, capillary pressure and relative permeability to account for this effect. We show that conventional measurements over-estimate the amount of capillary trapping by 20-25\%.

physics.flu-dyn↗

arXiv2019

DeepFlow: History Matching in the Space of Deep Generative Models

The calibration of a reservoir model with observed transient data of fluid pressures and rates is a key task in obtaining a predictive model of the flow and transport behaviour of the earth's subsurface. The model calibration task, commonly referred to as "history matching", can be formalised as an ill-posed inverse problem where we aim to find the underlying spatial distribution of petrophysical properties that explain the observed dynamic data. We use a generative adversarial network pretrained on geostatistical object-based models to represent the distribution of rock properties for a synthetic model of a hydrocarbon reservoir. The dynamic behaviour of the reservoir fluids is modelled using a transient two-phase incompressible Darcy formulation. We invert for the underlying reservoir properties by first modeling property distributions using the pre-trained generative model then using the adjoint equations of the forward problem to perform gradient descent on the latent variables that control the output of the generative model. In addition to the dynamic observation data, we include well rock-type constraints by introducing an additional objective function. Our contribution shows that for a synthetic test case, we are able to obtain solutions to the inverse problem by optimising in the latent variable space of a deep generative model, given a set of transient observations of a non-linear forward problem.

cs.LG↗

arXiv2018

Multimodal Functions as Flow Signatures in Complex Porous Media

This study refutes the premise that the distribution of flow speeds in complex porous media can be described by a simple function such as a normal or exponential variation. In many complex porous media, including those relevant for subsurface storage and recovery applications, a separation of scales exists between larger inter-granular pores and micro-porosity inside the grains, leading to different flow signatures that need to be described by multimodal functions with distinct flow field characteristics. We demonstrate this finding by devising a novel methodology to simulate fluid flow in carbonate rock based on a representation of the pore space obtained by differential X-ray imaging. The permeability assignment in the micro-porous space is estimated from the pore size inferred from mercury injection porosimetry. Model predictions considering micro-porosity agree well with experimentally measured porosities and permeabilities. The micro-porosity can contribute significantly to the overall permeability, particularly in the more heterogeneous media.

physics.flu-dyn↗

arXiv2018

Stochastic seismic waveform inversion using generative adversarial networks as a geological prior

We present an application of deep generative models in the context of partial-differential equation (PDE) constrained inverse problems. We combine a generative adversarial network (GAN) representing an a priori model that creates subsurface geological structures and their petrophysical properties, with the numerical solution of the PDE governing the propagation of acoustic waves within the earth's interior. We perform Bayesian inversion using an approximate Metropolis-adjusted Langevin algorithm (MALA) to sample from the posterior given seismic observations. Gradients with respect to the model parameters governing the forward problem are obtained by solving the adjoint of the acoustic wave equation. Gradients of the mismatch with respect to the latent variables are obtained by leveraging the differentiable nature of the deep neural network used to represent the generative model. We show that approximate MALA sampling allows efficient Bayesian inversion of model parameters obtained from a prior represented by a deep generative model, obtaining a diverse set of realizations that reflect the observed seismic response.

physics.geo-ph↗

arXiv2018

Conditioning of three-dimensional generative adversarial networks for pore and reservoir-scale models

Geostatistical modeling of petrophysical properties is a key step in modern integrated oil and gas reservoir studies. Recently, generative adversarial networks (GAN) have been shown to be a successful method for generating unconditional simulations of pore- and reservoir-scale models. This contribution leverages the differentiable nature of neural networks to extend GANs to the conditional simulation of three-dimensional pore- and reservoir-scale models. Based on the previous work of Yeh et al. (2016), we use a content loss to constrain to the conditioning data and a perceptual loss obtained from the evaluation of the GAN discriminator network. The technique is tested on the generation of three-dimensional micro-CT images of a Ketton limestone constrained by two-dimensional cross-sections, and on the simulation of the Maules Creek alluvial aquifer constrained by one-dimensional sections. Our results show that GANs represent a powerful method for sampling conditioned pore and reservoir samples for stochastic reservoir evaluation workflows.

stat.ML↗

arXiv2017

Stochastic reconstruction of an oolitic limestone by generative adversarial networks

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 network allows fast sampling of large volumetric images.We apply a GAN based workflow of network training and image generation to an oolitic Ketton limestone micro-CT dataset. Minkowski functionals, effective permeability as well as velocity distributions of simulated flow within the acquired images are compared with the synthetic reconstructions generated by the deep neural network. While our results show that GANs allow a fast and accurate reconstruction of the evaluated image dataset, we address a number of open questions and challenges involved in the evaluation of generative network-based methods.

cs.CV↗

arXiv2017

Generalized network modeling of capillary-dominated two-phase flow

We present a generalized network model for simulating capillary-dominated two-phase flow through porous media at the pore scale. Three-dimensional images of the pore space are discretized using a generalized network -- described in a companion paper (https://doi.org/10.1103/PhysRevE.96.013312) -- that comprises pores that are divided into smaller elements called half-throats and subsequently into corners. Half-throats define the connectivity of the network at the coarsest level, connecting each pore to half-throats of its neighboring pores from their narrower ends, while corners define the connectivity of pore crevices. The corners are discretized at different levels for accurate calculation of entry pressures, fluid volumes and flow conductivities that are obtained using direct simulation of flow on the underlying image. This paper discusses the two-phase flow model that is used to compute the averaged flow properties of the generalized network, including relative permeability and capillary pressure. We validate the model using direct finite-volume two-phase flow simulations on synthetic geometries, and then present a comparison of the model predictions with a conventional pore-network model and experimental measurements of relative permeability in the literature.

physics.comp-ph↗

arXiv2017

Reconstruction of three-dimensional porous media using generative adversarial neural networks

To evaluate the variability of multi-phase flow properties of porous media at the pore scale, it is necessary to acquire a number of representative samples of the void-solid structure. While modern x-ray computer tomography has made it possible to extract three-dimensional images of the pore space, assessment of the variability in the inherent material properties is often experimentally not feasible. We present a novel method to reconstruct the solid-void structure of porous media by applying a generative neural network that allows an implicit description of the probability distribution represented by three-dimensional image datasets. We show, by using an adversarial learning approach for neural networks, that this method of unsupervised learning is able to generate representative samples of porous media that honor their statistics. We successfully compare measures of pore morphology, such as the Euler characteristic, two-point statistics and directional single-phase permeability of synthetic realizations with the calculated properties of a bead pack, Berea sandstone, and Ketton limestone. Results show that GANs can be used to reconstruct high-resolution three-dimensional images of porous media at different scales that are representative of the morphology of the images used to train the neural network. The fully convolutional nature of the trained neural network allows the generation of large samples while maintaining computational efficiency. Compared to classical stochastic methods of image reconstruction, the implicit representation of the learned data distribution can be stored and reused to generate multiple realizations of the pore structure very rapidly.

cs.CV↗