InterPore2024

Asia/Shanghai
Description

Register now

Join us for fascinating lectures, engage with fellow researchers from across the globe and discover cutting-edge exploration of porous media. 

Please consider making a donation to InterPore Foundation in any amount to enable the participation of your fellow researchers.

The Focus Theme of InterPore2024 will be “porous media and biology.” This includes occurrence and applications of biology in living organs, plants, and in medicine, as well as in soil, oil and gas reservoirs. It will include theories and (bio)physics, experiments, imaging, mathematics, and modelling.
 

Topics and applications

  • Transport phenomena    
  • Swelling and shrinking porous media    
  • Multiphysics-multiphase flow    
  • Reservoir engineering    
  • Soil Mechanics and Engineering    
  • Geothermal energy    
  • CO2 sequestration    
  • Constitutive modeling    
  • Wave propagation    
  • Energy Storage    
  • Biotechnology
  • Biofilms
  • Thin and nanoscale poromechanics
  • Fuel cells and batteries
  • Food
  • Paper and textiles
  • Filters, foams, membranes
  • Fibers and composites
  • Ceramics and constructions materials
  • Other porous media applications

 


Event Management

    • Plenary/Invited: Plenary Session I
      • 1
        Reservoir Simulator Development: The Past, Present and Future

        Reservoir simulators have been developed in the past 70 years. They have been widely used to predict, understand, and optimize complex physical processes in modeling and simulation of multiphase fluid flow in petroleum reservoirs. These simulators are important for understanding the fate and transport of chemical species and heat and maximizing the economic and environmental performance of exploration and production of fossil fuel energy.

        The development of reservoir simulators has been concentrated on conventional oil and gas reservoirs in the last century, and efficient black oil, compositional and thermal simulators have been successful in their application to the recovery of conventional oil and gas resources. As these conventional resources dwindle, the recovery of unconventional oil and gas (such as heavy oil, oil sands, tight and shale oil and gas, and coalbed methane) resources is now at the center stage. While the development of unconventional reservoir simulators has been focused on in this century, a lot of challenges still exist because of the significant differences between conventional and unconventional reservoirs in their multi-scale phenomena, fluid occurrence states, flow mechanisms, and production technologies.

        The speaker has engaged in the development of reservoir simulators for over 30 years. His group has developed parallel and intelligent simulators that can efficiently simulate complex fluid flow problems with giga (billion) grid block cells and reduce simulation time from days to seconds. For over ten years, his group has also incorporated artificial intelligence (AI) and quantum computing algorithms into these reservoir simulators. Fast and accurate simulators can increase energy production due to full utilization of available data and better understanding of the chemical and physical mechanisms involved, process designs and uncertainty analyses. In this plenary presentation, the speaker will give an overview on the development of conventional and unconventional reservoir simulators, the incorporation of parallel and AI algorithms into these simulators, and the quantum computing potential to solve reservoir simulation problems. The present status, existing challenges, and future prospects on reservoir simulators will be emphasized in this plenary presentation.

        Speaker: Zhangxing (John) Chen
    • Poster: Poster Session I
      • 2
        A molecular simulation study on adsorption and diffusion behaviors of hydrogen, methane and carbon dioxide

        Because underground hydrogen storage offers the potential for large-scale, long-term storage of hydrogen, understanding the adsorption and diffusion behaviors of hydrogen and cushion gas in the reservoir is critical to understanding the underlying mechanisms that control hydrogen storage and transport. Using molecular simulation methods, we investigated the adsorption and diffusion behaviors of hydrogen, methane and carbon dioxide in kaolinite slit pores (10 MPa and 303 K), respectively. The distribution characteristics, excess adsorption amounts, diffusion coefficients and gas-solid interaction energies of the three gases in the slit pores were analyzed. Near the pore wall surface, carbon dioxide formed a distinct double adsorption layer, methane formed a smaller second adsorption layer, and hydrogen formed a single adsorption layer. The order of excess adsorption amount is carbon dioxide > methane > hydrogen. The rank of diffusivity of gases under the same conditions is hydrogen > methane > carbon dioxide. The interactions between gases and pore walls are in the following order: carbon dioxide>hydrogen>methane. Van der Waals interactions dominate. However, hydrogen and carbon dioxide have significant coulombic interactions with the pore walls, while methane has negligible coulombic interactions with the pore walls. The minerology of the formation results in different charges on the pore surfaces, which has a significant effect on gas storage. This study provides better insights into the mechanisms of hydrogen and cushion gas storage, thus providing a theoretical basis for underground hydrogen storage site selection.

        Speaker: Zhenxiao Shang (China University of Petroleum (East China))
      • 3
        A novel CO2-responsive microgel for in-depth conformance control in CO2 enhance oil recovery (EOR)

        CO2-EOR (enhanced oil recovery) represents one of the most cost-effective methods for achieving carbon neutrality. However, CO2 flooding encounters serious preferential flow in porous mediums, which significantly reduce the economic benefits of CO2-EOR and exacerbate the risk of CO2 leakage. This study synthesized a tertiary amine-based, CO2-responsive microgel. The microgel possesses a particle size of ~7.50μm in neutral formation water and expands to 2-4 times upon contact with CO2, thereby offering advantages in deep profile modification and selective CO2 plugging. Core flooding experiments confirmed the microgel's effectiveness in deep profile modification and EOR in heterogeneous reservoirs.

        Speaker: Qihui Wu
      • 4
        A Robust Vapor-liquid-liquid Equilibrium Calculation Algorithm Considering Capillary Pressure and Critical Shift in Nanopores

        Porous media contains a huge amount of nanopores, and the behavior of confined fluid phases in nanopores will be affected by significant interfacial effects between the fluids and the walls. A large number of publications have recently investigated the influence of interfacial effects on the phase behavior of confined fluids in nanopores. Many influences such as adsorption, critical shift, capillary pressure have been discussed. However, most of these studies have focused only on vapor-liquid equilibrium (VLE), and few attempts have been made to generalize these influences to vapor-liquid-liquid equilibrium (VLLE) in nanopores. In this work, a robust and efficient algorithm for the calculation of VLLE is proposed. The algorithm can simultaneously consider the influences of capillary pressure and critical shift on VLLE in nanopores. In order to be able to accurately and efficiently calculate each of the possible phases, the algorithm adjusts some of the steps in the conventional VLE calculation and improves the solution methodology.

        The results of VLLE in this algorithm are obtained from three-phase flash calculations. It is well known that three-phase flash calculations are extremely dependent on initial guess and have poor convergence properties. In order to solve this problem, this work couples the successive substitution iteration with the Newton-Raphson iteration and builds a joint solver. The results show that this coupled joint solver not only improves the computational efficiency, but also enhances the stability of the flash calculations.

        The robustness of the present algorithm is verified by several computational examples, and influences of capillary pressure and critical shift on VLLE in nanopores are investigated. In this work, mixtures of hydrocarbons with water and carbon dioxide were calculated and several possible fluid distribution scenarios were considered. The computational results show that both capillary pressure and critical shift can change the phase distribution of the confined fluid in nanopores.

        1. The effect of capillary pressure on phase behavior is determined by the wettability between nanopores and the confined fluids. When the nanopores have a higher affinity for the vapor phase than the liquid phase, the generation of the bubble point will be advanced and the generation of the dew point will be lagged. If wettability is reversed, the effect of capillary pressure on phase behavior of the system is also reversed at the same time.
        2. Critical shift affects the phase behavior of confined fluids more than capillary pressure for a certain range of pore sizes. However, the critical shift is independent of the wettability. Regardless of the pore size, the effect of critical shift on phase behavior has the same trend.
        3. The effect of pore size on the behavior of confined fluids is nonlinear. In other words, there is an inflection point in the trend of pore size versus the phase behavior, and when the pore size is higher than this inflection point, the effect of capillary pressure or critical shift can be neglected when calculating the VLLE. Based on the findings of this work, we believe that this inflection point lies around 100 nm.
        Speaker: Binyao Xiao (China University of Petroleum (East China))
      • 5
        A Study on Stochastic Modeling of Channelized Reservoirs Based on Reinforcement Learning

        Channelized reservoirs, formed by ancient river sands, play a crucial role in oil and gas reservoirs. Their subsurface nature, however, poses significant challenges in direct visualization and characterization. This paper presents a novel approach to model these reservoirs by applying Q-learning, a machine learning technique. We redefine the reservoir modeling problem within a Q-learning framework, where well locations form the state space, and the reward mechanism differentiates between channel and non-channel wells. Our method utilizes the adaptive learning ability of Q-learning to predict the channel path, aiming to maximize the cumulative reward of identifying channel wells while minimizing traversal through non-channel wells. We extend our study by conducting multiple simulations and applying the results to porous media flow field simulation, which serves to test geological uncertainties. This not only provides a deeper understanding of channelized reservoirs but also showcases the potential of machine learning techniques in analyzing large datasets for geological exploration and characterization. Comparative results with traditional stochastic walk algorithms affirm the effectiveness and accuracy of our approach, offering new insights into the modeling of porous media.

        Speaker: Ms Xiufan Zhang (QUST)
      • 6
        Archaea community in gas hydrate-bearing sediments in the South China Sea

        The South China Sea, endowed with abundant natural gas hydrate resources, presents exceptional conditions for gas hydrate formation and exploration prospects. Gas hydrate-bearing sediments are characterized by methane saturation in pore waters, fostering rich and distinctive microbial ecosystems. These microorganisms play crucial roles in methane production, consumption, and global carbon cycling. This study focuses on 117 sediment samples from 11 sites across the Qiongdongnan Basin, Shenhu area, and Xisha Trough of the South China Sea , employing high-throughput MiSeq sequencing of the 16S rRNA gene to investigate the archaeal community structures and diversity. Our findings highlight significant microbial diversity variance across samples from the three geographic regions, with distinct differences noted between samples from the Qiongdongnan Basin and Shenhu area. The archaeal population is dominated by Halobacterota, Hadarchaeota, Lokiarchaeota, Euryarchaeota, and Woesearchaeota. Notably, methane-metabolizing taxa are prevalent, with a significant abundance of methanogenic archaea over anaerobic methane-oxidizing archaea (ANME). Additionally, the structure of methane-metabolizing groups varies significantly across the three regions, with ANME predominantly identified in the Qiongdongnan Basin. Methanogens show differing dominance in the Shenhu area compared to the Qiongdongnan Basin, and only a few methanogenic groups were observed in the Xisha Trough samples. This study provides the characteristics of archaeal community diversity within sediment cores from gas hydrate-bearing sediments in the South China Sea, contributing to our understanding of microbial group characteristics in these regions. Understanding these microbial populations and their functions is crucial for the comprehension of the biogeochemical processes involved in the formation of natural gas hydrates in the South China Sea.

        Speaker: 思炜 刘 (北京大学)
      • 7
        Attenuation Patterns of Low-Frequency Hydraulic Pulse Waves in Porous Media with Different Permeability

        Hydraulic pulsing is a widely used technology for cleaning, hydro-shredding, and soil improvement. In recent years, low-frequency hydraulic pulses have been used in oil and gas development, such as wellbore blockage removal and EOR. The effectiveness of this process depends on the attenuation pattern of the hydraulic pulse wave in the reservoir. In this paper, a numerical model of low-frequency hydrodynamic pulsed wave propagation in porous media is constructed and experimentally verified. The finite element method was used to solve the model and analyze the radial attenuation patterns of low-frequency hydraulic pulse waves in porous media with varying permeability. The results indicate that the rate of attenuation of the energy (pulse amplitude) of the hydraulic pulse wave with distance is significantly affected by the permeability of the porous medium. The rate of attenuation exhibits a pattern of deceleration, acceleration, and then deceleration again as distance increases. As the hydraulic pulse frequency is 0.1Hz, the pulse amplitude is 10MPa, and, the pulse amplitude attenuates faster along the path as the permeability decreases and is mainly concentrated within 0.5-2.5m around the wellbore. Specifically, the pulse amplitude attenuation to 20% corresponds to a propagation distance of 1.8 m in the 1 mD porous media, 2.7 m in the 200 mD porous media, and 6.8 m in the 1000 mD porous media. The pulse amplitude gradually decreases within 0.5 meters around the wellbore, which may be attributed to the impact of reflected waves. In a porous medium with a permeability of 100 mD, reducing the pulse frequency (0.1-30 Hz) can effectively decrease the interference between different wave levels during propagation and slow down the attenuation rate of pulse amplitude with distance. As distance increases, the effect of pulsation amplitude on attenuation decreases, and the degree of attenuation becomes similar for different amplitudes. For reservoirs with specific permeability, it is recommended to adjust the hydraulic pulse parameters to a lower frequency and higher amplitude to achieve a longer effective distance and better results.

        Speaker: Mr Kai Wang (China University of Petroleum (East China))
      • 8
        Brinkman double-layer model for flow at a free-porous interface

        The phenomenon of the Stokes–Darcy flow in coupled systems comprising a clear channel and a complex 3D porous medium is investigated through numerical and theoretical approaches. A quartet structure generation set (QSGS) method is used to generate random complex 3D porous structures imitating real structures in nature. Pore-scale flow simulations are performed using the Lattice Boltzmann method, enabling detailed analysis and characterization of the interfacial flow phenomena. Four key parameters with clear physical meanings are introduced to capture essential aspects of the flow dynamics quantitatively, revealing intriguing linear relationships with the square root of permeability – a fundamental characteristic length scale dominating the phenomenon. Several existing models are examined by these parameters. A Brinkman double-layer(BDL) model is proposed to address the limitations of existing models. Compared with several classic models, the present BDL model stands out due to its simplicity, accuracy, and robustness, providing a comprehensive understanding of the complex flow behavior in the coupled system.

        Speaker: Jinliang Kang (Tsinghua University)
      • 9
        Bypass flow of trapped droplet under seismic stimulations through pore double model analysis

        The study of trapping and releasing nonwetting droplets in porous materials has been extensively explored. However, dynamically characterizing residual nonwetting droplets under vibrations within porous media remains a challenging endeavor. Current theoretical models addressing seismic responses in two-phase flow primarily focus on single-channel geometries with fixed pressure differentials across inlet and outlet boundaries. In practical porous media, trapped droplets exist amidst flowing aqueous phases. External vibrations can induce significant pressure fluctuations due to surrounding flows, making the fixed pressure differential assumption invalid in a single-channel model. To overcome these constraints, this study delves into the micro-scale dynamics, aiming to surpass the limitations of the single-channel model. A theoretical framework involving a pore doublet, i.e., a sinusoidally constricted channel paralleled with a straight channel, is proposed to account for bypass flow effects. Initially, we analyze alterations in pressure differentials upstream and downstream of residual nonwetting droplets, considering flow dynamics during seismic excitation. We evaluate the impact of these pressure differential variations on predictive accuracy compared to the fixed pressure differential assumption in the single-channel model. Employing the pore doublet theoretical model, we examine how the permeability of parallel straight channels influences the dynamic response of residual nonwetting droplets. Furthermore, we compare predictive discrepancies between the single-channel and pore doublet models, integrating bypass flow, to determine critical acceleration amplitudes for releasing residual nonwetting droplets at different frequencies. Ultimately, we uncover the competitive interaction between the seismic response of residual nonwetting droplets at pore throats and bypass flow in surrounding pores. These research findings establish a robust theoretical foundation for comprehending seismic impacts on engineering, geological implications, and the potential incubation of geological disasters within the geological and environmental domains.

        Speaker: Wen Deng (Southeast University)
      • 10
        Density-Dependent Dynamics of Fines Retention and Pore Clogging in Rock Formations: A CFD-DEM Approach

        In the realm of geosciences, the phenomenon of fines migration and subsequent clogging in rock formations presents a complex challenge. This process can occur even when fines are smaller than a specific threshold size, known as the critical throat diameter. The dynamics of pore clogging involve interactions on multiple scales - ranging from the transport at the pore level to the mechanical and hydraulic behaviors at the colloid level, down to the electrochemical interplays at sub-colloid scales. Traditionally, the Colloid Filtration Theory (CFT) has been the go-to model, focusing on predicting how colloidal particles are retained under the assumption of clean bed conditions. This overlooks the significant impact of particle aggregation and clogging at the throat passages. While experimental measurements are ideal for assessing filtration efficiency, they fall short in directly examining the intricate movements and paths of particles at the pore level, especially due to the opaque nature of the media involved. Numerical models addressing the full scope of forces in pore clogging have been limited to two-dimensional simulations of rock structures. This research advances the field by employing a combined fines tracking method that integrates Computational Fluid Dynamics (CFD) with a Discrete Element Model (DEM). The approach is designed to predict the retention and clogging of fines, factoring in surface forces and the impact of gravity forces due to density variations between the fines and the saturating brine.

        To accurately represent the complex pore structures and simulate particle movement within the rocks, we use a three-dimensional X-ray computed microtomography image of Bentheimer sandstone. An innovative feature of our method is the use of a dynamically adaptive CFD mesh, which refines itself in areas dense with particles to better resolve the intricate fluid flow around them. We track particle trajectories and link them to specific pores and throats within the rock sample. Furthermore, we apply a method to compare the pair-wise trajectories of particles with different density, injected in a flux-weighted procedure through the inlet. This allows us to analyze how particle retention varies within the sandstone, considering the flow direction for fines with different densities. For heavier fines, we observe an interesting spatial variation in particle trapping, suggesting that gravity aids their movement in the direction of gravitational pull, but hinders it along the flow path. Our findings reveal that denser fines can move further along the flow path due to two key mechanisms: their ability to pass through gravitationally lower throats and the alterations in flow pathways caused by pore clogging.

        Speaker: Shitao Liu
      • 11
        Direct Pore-Scale Simulation of the Effect of Wettability Alteration by Low-Salinity on Oil Mobilization in 3D Natural Sandstone

        Wettability is an important factor controlling the pore-filling mechanism and displacement efficiency in the subsurface pore space. The trapped phase can be mobilized by wetting alteration, which is one of the main mechanisms of enhanced oil recovery technologies, such as surfactant flooding and low-salinity water flooding. Despite recent advances in the simulation of wetting alteration at the core scale or beyond, there are very few works that have modeled the wettability alteration at the pore scale, especially in three-dimensional (3D) micro-CT images, causing fluid displacement and retrapping mechanisms during wettability alteration are not well understood. With this objective, a wettability alteration model by low-salinity is developed and implemented in the open-source computational fluid dynamics software OpenFOAM (Open Source Field Operation and Manipulation), where both the Navier-Stokes equations for oil/water two-phase flow and the advection-diffusion equation for species transport are solved. The proposed model is validated against a published sinusoidal channel micromodel and then applied to 3D micro-CT images of sandstone to investigate the interplay between wettability alteration and pore structure. This study takes into account the effect of initial wettability, different degrees of wettability alteration, different time scales of wettability alteration, and different injection scenarios on oil trapping and the ultimate oil recovery factor. A larger degree of wettability alteration results in a higher oil recovery factor during tertiary low salinity water flooding. However, the oil recovery factor will first increase and then decrease with the increase of wettability alteration degree due to the snap-off effect during secondary low salinity waterflooding. In tertiary low salinity waterflooding, a lower wettability alteration time scale under the same degree of wetting alteration produces more oil. This study emphasizes the important interplay between wettability alteration, pore structure, and time scale during low salinity water flooding, and can explain some observations in recent micro CT experiments.

        Speaker: Haoyun Li (China University of Petroleum (East China))
      • 12
        Dynamic Effects on Solute Transport in an Unsaturated Soil

        The influence of flow regime and soil saturation on solute transport processes is significant, yet the associated effects have not been adequately addressed. To address this gap, we conducted three sets of solute transport experiments in a sandy soil, complemented by numerical modeling, under both steady-state and dynamic drainage conditions. The results from steady-state experiments revealed a non-monotonic relationship between dispersivity and saturation.Both classical advection-dispersion and dual-porosity (mobile-immobile) type transport equations were used to simulate the measurements. The fitted well defined dispersivity -saturation function was employed to the simulations of dynamic experiments. Taking into consideration the dynamic capillarity effect, our model accurately simulated solute transport processes and flow. Contrary to previous reports, our findings suggest that the flow regime does not significantly impact the dispersivity of solute transport.

        Speaker: Luwen Zhuang (Sun Yat-sen University)
      • 13
        Ensemble Variational Bayesian Uncertainty Quantification for High Dimensional Nonlinear Parameter Inversion of Darcy Flows in Porous Media

        The inversion and uncertainty quantification of parameters associated with governing PDEs are important in many scientific and engineering problems. For example, petroleum reservoirs are typically heterogeneous and uncertain due to the sparsity of hard data, and the uncertainty quantification of physical parameters associated with the governing PDEs of flows in porous media, given production history data, is a necessary step before reasonable forecasts can be made. Conventional history matching inversion methods are generally point-estimate, while uncertainty quantification using MCMC is computationally expensive. In the current study, an efficient ensemble variational Bayesian (EVB) uncertainty quantification method is developed for inverting high-dimensional parameters for the governing PDEs. Variational Bayes inference approximates the posterior using trial distributions such that the Kullback-Leibler divergence between the true posterior and the trial distribution can be minimised. In EVB, a reduced-order model is built using principle component analysis to enhance the convergence of small-size ensembles. The trial distribution is optimized simultaneously as the ensemble of realizations are updated by data assimilation. In particular, particle filtering is adopted for the nonlinear inverse problem under consideration. Two- and three-dimensional test cases of single- and two-phase Darcy flows in petroleum reservoirs are presented for validation.

        Speaker: zhao zhang (Shandong University)
      • 14
        Experimental and theoretical evidence for energy signal indicating flow regimes for two phase flow in porous media

        Many studies have been dedicated to examining flow regimes using two key parameters: viscous ratios and capillary numbers Ca. However, only a few studies have elucidated mechanisms that govern different flow regimes and how the work of displacement and surface aeras alter within porous media, as well as their influence on flow behavior is still unknown. In this study, we experimentally investigate the combined effect of wettability and flow rates on immiscible fluid-fluid displacement using high-resolution imaging in microfluidic flow cells with two different viscous ratios. We investigate morphology of oil cluster and displacement front and further calculate the relative change of energy conversion based on external work and surface energy. The morphology of invasion patterns in brine-silicone is sharper than that in brine-decane displacement, with the indication of larger ratios of length and width for fingers. The signature of the transition between the three regimes manifests itself in the efficiency of conversion of the external work to surface energy. Efficiency of conversion decreases with the increase of contact angles. With the increase of Ca, Efficiency of conversion reduces greatly to approximately zero. In high M displacement, efficiency of conversion is consistently higher than that in low M displacement. The signature of the transition between the three regimes (viscous flow, capillary dominated flow and capillary -stable displacement flow regime) manifests itself in the fluctuations of the external work and surface energy. We propose that it is possible to determine the nature of multiphase-flow displacement from the energy signal.

        Speaker: Shuangmei Zou
      • 15
        Experimental Validation of Pore-Scale Models for Gas Diffusion Layers in PEMFCs

        Pore-scale modeling developed over the past decades has become a powerful method to evaluate the effective transport properties of porous electrodes. Experimental verification for such a method is crucial to confirm the method's validity. In this study, experimental data of gas diffusion layer (GDL) are compared with results of pore-scale modeling. GDL microstructures are scanned and reconstructed by X-ray computed tomography. Explicit dynamic simulations based on the finite element method are performed on these reconstructed models to reveal the 3D displacement of the microstructure during compression. Over the deformed models, the effective diffusivity, thermal and electrical conductivities are then computed using a pore-scale model code. It is found that, as the compression ratio increases to 30%, the fiber displacement increases obviously with significant anisotropy, and the fibers gradually squeeze into nearby pores located in the adjacent layers inside GDL. The effective diffusivity and permeability decrease by about 15% and 35% respectively. The conductivity increases by 100% and 20% in the through-plane and in-plane direction respectively. The validated methods can support microstructure optimization and transport properties improvement for different types of porous electrodes.

        Speaker: Prof. Liusheng Xiao (Ningbo University)
      • 16
        Exploring the Relation Between Soil Salinity on Soil Organic Carbon Dynamics in Global Terrestrial Ecosystems

        Soil Organic Carbon (SOC) is a fundamental component of terrestrial ecosystems, connected to climate regulation, nutrient cycling, and soil health. The influence of soil salinity - referring to the concentration of soluble salts in the soil solution - on SOC content is acknowledged (1,2), but there is limited understanding regarding the precise direction and extent of SOC’s response to varying levels of soil salinity in real field conditions. This study explores the relationship between soil salinity and SOC content, necessary for understanding carbon sequestration processes, climate change mitigation, and terrestrial carbon stock stability. Using the SOC of 60,392 soil samples collected globally since 1950, we developed a statistical model (General Additive Model) and analyzed soil salinity’s relation with SOC dynamics while controlling the role of other environmental parameters. According to the results of the statistical analysis, we estimate that an increase in soil salinity from 1 to 5 dS m-1 would be correlated with a decrease in SOC, equivalent to dropping from 0.92 g kg-1 above the mean predicted SOC (31.77 g kg-1) to 6.34 g kg-1 below the mean predicted SOC (-700%), while considering other influencing environmental factors such as precipitation and temperature. Our results show the minor contribution of salinity to SOC while other factors such as climate, vegetation, and land management practices exert more substantial effects on SOC content. Key covariates in relation with SOC include soil nitrogen, anthropogenic phosphorous input, and soil pH. Additionally, our study estimates the effects of one standard deviation increases in soil salinity of the analyzed soil samples on topsoil (0 – 20 cm) SOC content, showing a 6.98% decrease in SOC. These findings highlight the importance of considering diverse factors in understanding SOC dynamics, providing insights into mitigating the impacts of soil salinity and climate change on terrestrial ecosystems.

        Speaker: Prof. Nima Shokri (Hamburg University of Technology)
      • 17
        Flow simulation of pore-scale deep shale gas under nano-confinement conditions

        Deep shale gas will become an important part of supporting the growth of China 's natural gas production. Compared with the middle shale and shallow shale, the properties and porous flow laws of deep shale gas are more complex. The nano-confinement effects such as adsorption and slippage cannot be ignored in the study of porous flow mechanism of deep shale gas. When the porous flow law of deep shale gas on the pore scale, the influence of nano-confinement effect on the porous flow law needs to be further clarified. In this work, a pore network model containing water-wet inorganic pore throats and gas-wet organic pore throats is established, which conforms to the pore structure characteristics of shale in actual depth. And the permeability of shale gas under different wettability, slippage, adsorption and surface diffusion is studied. Viscous flow, Knudsen diffusion, adsorption, slippage and surface diffusion are considered in organic pores, viscous flow and Knudsen diffusion are considered in inorganic pores. Different TOC contents are set to study the influence of nano-confinement effect on deep shale gas flow and the flow law of shale gas. The results show that the porous flow of deep shale gas is greatly affected by adsorption and slippage. The surface diffusion of adsorbed gas in organic pores provides more flux for the flow of deep shale gas. When the TOC content is high, the flow of shale gas is mainly controlled by organic pores.

        Speaker: Chaoyang Zhao (中国石油大学(华东))
      • 18
        Influence of non-stationarity within porous media sample on its flow properties

        The structure of the sample defines its physical properties. The homogenization based on different fields (e.g., pressure and velocity for permeability property) produces general property of the sample. For this property to be useful for continuum-scale modelling it has to be representative for the volume it will be substituted for in the next level model. This explains the importance of conventional REV concept. To be a REV, the structure has to be statistically homogeneous [1], but this is not necessarily achieved in real porous media samples such as rocks and soils [2]. This is where pore-scale modelling gets really handy – we can still perform homogenization and substitute the averaged property. In this contribution we discuss the influence of spatial non-stationarity on flow properties using full permeability tensor [3,4] as an example. To establish interrelationships, we create artificial porous media structures with different degrees of non-stationarity using stochastic reconstruction methodology [5,6].
        In this presentation we shall focus on:
        - Methodology to produce porous media structures with different degree of anisotropy;
        - Tensorial property assessment for such structures;
        - Applications for real homogenization and upscaling cases.
        The degree of stationarity of the stochastic reconstructions had a significant influence on the physical properties of the reconstructed binary structures—computed full permeability tensors showed different degree of anisotropy and off-diagonal terms values. The proposed approach to produce nonstationary structures from ensemble averaged set of correlation functions opens numerous ways to attack theoretical and practical problems with natural and artificial porous materials with statistically inhomogeneous structure. Moreover, it is possible to produce large scale inhomogeneous porous structures to parameterize, test and verify different upscaling schemes starting from pore-scale.

        Speaker: Dr Marina Karsanina (Schmidt Institute of Physics of the Earth of Russian Academy of Sciences, Moscow, Russia)
      • 19
        Investigation of fault damage zones from direct shear tests and implications for hydraulic fracturing process

        Hydraulic fracturing is an important stimulation technique for extracting resources from low-permeable formations. Apart from hydraulic fracturing (where pore pressure exceeds the minimum principal stress), hydraulic shearing (where pore pressure does not exceed the minimum principal stress) is an essential mechanism in forming the stimulated reservoir volume and controlling the ultimate stimulation results, especially for shale formations with layered beddings (Li et al., 2019). Shear failure in rocks will not only generate primary fractures but also cause stress alteration in the neighboring region around the primary fracture. Such stress alteration affects the rocks' stability and possibly induces more secondary fractures. This phenomenon is also known as fault damage zones, commonly observed in rock masses at different scales (Kim et al., 2004; Sui et al., 2019). In reality, it is almost impossible to directly observe the shear failure process in the subsurface. Therefore, in this work, we adopted an advanced dynamic direct shear testing device to break rocks (Qi et al., 2020) and a micrometer CT to observe the fault damage zones after the shear failure. Quantitative descriptions of the damage zone, such as fracture intensity, roughness, and connectivity, are summarized.
        In this research, an independently designed dynamic direct shear testing device (Qi et al., 2020) was utilized to conduct in-house direct shear tests on layered shale specimens with various layer angles (0°, 30°, 45°, and 60°). Subsequently, the sheared shale specimens were scanned with a micrometer CT, and 3D digital cores were reconstructed. Fine segmentation of micron-scale fractures in inhomogeneous shale was achieved using multiple processing algorithms. Considering different bedding structural surfaces, the physical properties and geometrical features of 3D micrometer-scale fractures in shale were quantitatively evaluated. Parameters such as fracture density, 3D shape factor, roughness, Euler number, and morphological filtering were employed to subdivide the damage zones along the main fracture surface.
        From the preliminary results, we divided the generated fractures into three categories based on their spatial distribution: the primary induced fracture, fractures in the connected damage zone, and isolated fractures. The range of damage zones varies significantly with different bedding angles. With the increase in bedding inclination, the density of fine fractures distributed along the primary shear fractures initially increases and then decreases, reaching the maximum in samples with 30°inclined layers. The roughness of primary fracture surfaces and secondary fractures in shale samples with different bedding angles displays significant anisotropy and asymmetry, with greater roughness observed at the microfractures and fractures in connected damaged zones. Morphological patterns of the three types of fractures are further discussed. An in-depth understanding of the fault damage zone provides valuable insights to evaluate and optimize the hydraulic fracturing process.

        Speaker: Ms Zifang Zhu (Institute of Geology an Geophysics Chinese Academy of Sciences)
      • 20
        Mechanism simulation on low salinity water flooding in high temperature sandstone reservoirs based on molecular simulation method

        Low salinity water flooding (LSWF) is novel technique which can be used to improve oil recovery for sandstone reservoirs. Although considerable experimental research has been conducted to identify the underlying pathways, there are a lot of debatable issues with the mechanics. On the basis of molecular simulation (MS) method, the models of rock, oil and brine in different salinity and ions compositions were constructed. The interactions among rock, oil and brine and the influence of brine salinity and concentrations of ions on the process of separating oil from sandstone surfaces were studied. The temperature range considered ranged from 298K to 373K. That altering of the wetting state of a sandstone induced the detachment of oil from the surface of the rock, even under elevated temperatures. The results will provide essential molecular state information for change in the wetting state of rock and increase in oil recovery.

        Speaker: Prof. Renyuan SUN (China University of Petroleum(East China))
      • 21
        Molecular dynamics simulation of ionic diffusion and mixing phenomena in polymer-enhanced low-salinity waterflooding

        Various studies have confirmed that water salinity and composition significantly impact the efficiency of the waterflooding process. Field-scale operation of low-salinity water injection has been proven to be a cost-effective enhanced oil recovery (EOR) method as well as compatible with environmental regulations. The success of this technique relies on the contact of low-salinity water with the rock surface to alter the wettability of the rock to more water-wetting conditions. However, the salinity of the injected water increases, as it contacts the resident high-salinity reservoir brine which then significantly impairs the efficiency of this technique. Under flowing conditions, two main mechanisms affect the mixing phenomena: hydrodynamic dispersion and molecular diffusion. Our preceding laboratory research has clearly shown that if a small amount of polymer, such as HPAM (partially Hydrolyzed Polyacrylamide) is added to the injection low-salinity water it can significantly reduce the mixing zone by suppressing the diffusion and dispersion phenomena and increase the integrity of the fluid-fluid front. To gain a more in-depth understanding of this process, molecular dynamics simulations were performed to investigate the system at an atomic scale. Polymer molecules were introduced into the low-salinity water and the high-salinity and low-salinity waters were made laterally in contact, under no-flow conditions, to start mixing. Sensitivity analysis was performed on the main factors affecting this phenomenon such as the presence of polymer molecules, the effect of polymer concentration, the salinity of low salinity water, and the salinity of resident brine (i.e., the salinity difference). The results indicate that the time for full mixing is controlled primarily by the effective diffusion coefficient. By adding polymer, the polymer strands and their chemical interaction with the brine ions would act as a diffusion barrier and reduce the diffusivity of low-salinity water, enhance the viscosity, and delay the ionic diffusion phenomenon; thereby reducing the growth rate of the mixing zone. In all cases the mixing zone grew linearly with the square root of time; indicative of Fickian diffusive mixing. Once the diffusivity is reduced, the salinity profile becomes sharper, leading to a more effective low-salinity water effect, and less volume of injecting EOR agent is required at large scales.

        Speaker: Prof. Hassan Mahani (Sharif University of Technology)
      • 22
        Molecular Simulation of the Effect of Imidazolium-Based Ionic Liquids on the Water/Toluene Interface

        Crude oil is a complex mixture of organic compounds which are conventionally categorized as saturates, aromatics, asphaltene and resins based on their polarity and solubility. Aromatics, in particular, comprise a large portion of many light, medium and heavy oils therefore it is important to understand their interfacial properties with water and surfactants in the context of enhanced oil recovery. In this work, we explore the effect of imidazolium-based ionic liquids on water/toluene interface using molecular dynamics (MD) simulation. Imidazolium ionic liquids are cationic organic compounds comprised of a heterocyclic aromatic head with a saturated aliphatic chain. The MD simulations were based on classical OPLS and SPC/E force fields with the leap-frog numerical integration scheme, 1.2 nm cut-off for Lennard-Jones potential and particle mesh Ewald summation for the electrostatic interactions. The interfacial tension (IFT) for pure water and toluene was found to be 36.09 mN/m at 300 K and 1 bar which is close to the experimental value of 35 mN/m. Adding imidazolium chloride at a surface coverage of two molecules for every square nanometer slightly reduced the IFT to 32.71 mN/m for the butyl chain. For Ionic liquids with octyl and dodecyl chains, the IFT of water/toluene has been significantly reduced by approximately a factor of 2 and 8, respectively. Calculations of the interaction energy between the two phases shows that imidazolium cations with longer chains interact more strongly with Toluene, thereby effectively reducing the IFT. Furthermore, density profiles across the axial-direction perpendicular to the interface shows that the ionic liquids with longer alkyl chain partition more toward the interface whereas the butyl imidazolium have a low surface excess and are not as surface active as the octyl and dodecyl imidazolium. Therefore, imidazolium interacts with toluene more via its alkyl chain and the association between the toluene and the imidazolium aromatic head is limited at the water/toluene interface.

        Speaker: Moataz Abu-Al-Saud (Saudi Aramco)
      • 23
        Numerical modelling of polymer support fluids permeating in sands

        Polymer fluids, a blend of polymers in water, offer a cost-effective and environmentally friendly solution for supporting deep underground excavations. However, being non-Newtonian fluids, their full potential in construction projects is hindered by a limited understanding of their behaviors. In this study, we will employ a combined approach of DEM-based and micro-CT imaging techniques to explore the pore-space topology in sands. Utilizing this data and considering the fluid-solid interaction, we will then develop a fully-resolved numerical model to comprehensively investigate the distribution pattern of strain and stress within the fluid phase, as well as drag forces on sand particles. Our numerical results will be validated against large-scale experimental observations, and provide insights for the development of upscaling modelling techniques.

        Speaker: Si Suo (Imperial College London)
      • 24
        Pore-scale Modeling of Dynamic CO2 Dissolution in Natural Porous Media with different Wettability

        In oil-gas-water three-phase systems, CO2 can be distributed either as a non-wetting phase, or as an intermediate-wetting phase. The morphology and distribu-tion of CO2 clusters under different wetting sequences are different, which has a complex influence on CO2 dissolution process. Based on phase distribution ob-tained from three-phase flow experiment, we constructed the physical models of initial CO2 phase distribution, subsequently simulated the CO2 dissolution pro-cess when CO2 is non-wetting phase and intermediate-wetting based on the VOF framework and CST method. The dynamic evolution of CO2 clusters and dis-solved CO2 distribution during dissolution process was tracked. The effect of wettability on CO2 dissolution trapping in three-phase systems was revealed. The characteristic parameters of CO2 dissolution process were also analyzed quantita-tively. Our results showed that CO2 clusters exhibited different dissolution states under different wetting conditions in three-phase systems. When CO2 serves as intermediate-wetting phase, the initial phase distribution is more dispersed, and the size of CO2 clusters is smaller, the CO2 saturation decreases more within the same time period, indicating that CO2 has a higher dissolution ability. The initial CO2 saturation determines the final CO2 concentration in the other phase. Disso-lution caused the originally connected large CO2 clusters to decompose into mul-tiple small clusters. When CO2 serves as intermediate -wetting phase, the mass of dissolved CO2 is higher, and thus the dissolution ability is higher.

        Speaker: Jinlei Wang
      • 25
        Reversing capillary trapping of nonaqueous fluid from dead-end structures by nanoparticle suspension and their self-adaptive control in complex porous media

        We report an anomalous capillary phenomenon that reverses typical capillary trapping via nanoparticle suspension and leads to a counterintuitive release of nonaqueous fluid from dead-end structures under weakly hydrophilic conditions. Fluid interfacial energy drives the trapped liquid out by hierarchical surfaces: the nanometric roughness formed by nanoparticle adsorption transfers the molecular-level adsorption film to hydrodynamic film by capillary condensation and maintains its robust connectivity, then the capillary pressure gradient in the dead-end micrometric structures drives trapped fluid motion out of the pore continuously. The developed mathematical models agree well with measured evolution dynamics of released fluid. This reversing capillary trapping phenomenon via nanoparticle suspension may be a general event in a random porous media and could dramatically increases displacement efficiency. Our findings have implications for manipulating capillary pressure gradient direction via nanoparticle suspensions to trap or release the trapped fluid from complex geometries, especially for site-specific delivery, self-cleaning, or self-recover systems.

        Speaker: Dr Wenhai Lei (KTH Royal Institute of Technology)
      • 26
        Stability, deformation and rupture of Janus oligomer enabled self-emulsifying water-in-oil microemulsion droplets

        Microemulsions exist widely in nature, daily life and industrial manufacturing processes, including petroleum production, food processing, drug delivery, new material fabrication, sewage treatment, etc. The mechanical properties of microemulsion droplets and a correlation to their molecular structures are of vital importance to those applications. Despite studies on their physicochemical determinants, there are lots of challenges of exploring the mechanical properties of microemulsions by experimental studies. Herein, atomistic modelling was utilized to study the stability, deformation, and rupture of Janus oligomer enabled water-in-oil microemulsion droplets, aiming at revealing their intrinsic relationship with Janus oligomer based surfactants and oil structures. The self-emulsifying process from a water, oil and surfactant mixture to a single microemulsion droplet was modulated by the amphiphilicity and structure of the surfactants. Four microemulsion systems with an interfacial thickness in the range of 7.4–17.3 Å were self-assembled to explore the effect of the surfactant on the droplet morphology. By applying counter forces on the water core and the surfactant shell, the mechanical stability of the microemulsion droplets was probed at different ambient temperatures. A strengthening response and a softening regime before and after a temperature-dependent peak force were identified followed by the final rupture. This work demonstrates a practical strategy to precisely tune the mechanical properties of a single microemulsion droplet, which can be applied in the formation, de-emulsification, and design of microemulsions in oil recovery and production, drug delivery and many other applications.

        Speaker: Yuequn Fu (University of Oslo)
      • 27
        The occurrence states of shale oil and its controlling factors in Yanchang Formation, Ordos Basin, China

        Oil mobility evaluation is the primary topic in shale oil development. The different occurrence states of shale oil, which closely relate to the pore structure and fluid properties tremendously affect the oil mobility in shale. As proved in previous studies, the higher the content of oil in free states, the better the oil mobility will be. In this study, the oil occurrence states and its influencing factors of the 7th member in Yanchang Formation (Chang7) in Ordos Basin, China were investigated by multiple experiments, including nuclear magnetic resonance (NMR), nitrogen gas adsorption (NGA), electronic scanning microscope (SEM), X-ray diffraction, and rock pyrolysis. The Chang7 samples were classified into four lithofacies based on mineralogy and TOC. With the increasing of the clay content, the four lithofacies are siliceous shale, OM-rich siliceous shale (TOC>5%), argillaceous shale, and OM-rich argillaceous shale (TOC>5%) (Figure 1). The NMR results indicate the fluids in Chang7 consist of structure water, free water, adsorbed oil, and free oil. The contents of adsorbed oil increase with TOC and clay percentages increase. However, the contents of free oil show negative relationships with TOC and clay percentages. Therefore, we speculate that the strong adsorbability in organic matter and clay minerals force the oil to be preserved as adsorbed oil in shale, which tremendously affect the oil occurrence states and mobility. Siliceous shale has the greatest content of free oil, and OM-rich siliceous shale has the greatest content of adsorbed oil. From the NGA results, mesopores have the domination in pore volume and specific surface area, especially in siliceous shale. The cumulative pore volumes decrease from siliceous shale to OM-rich argillaceous shale with the increase of clay contents. The NGA was also conducted on the samples after solvent extraction, which shows opposite changes of the volume in mesopores and macropores. Lots of mesopores were released by solvent extraction, particularly in siliceous shale which has the greatest increase in mesopores. Similarly, SEM images of the siliceous shale also show a large amount of mesopores in felsic grains. In the other three lithofacies, few pores were observed because of the tight compaction of clay and felsic minerals. Combing with the free oil content, the positive relationship between the free oil content and different percentage of mesopore volume indicates that the mesopores are the primary storage space of free oil (Figure 2). Compared with the lithofacies which have more clay content and TOC, the siliceous shale has more mesopores and low adsorbability in the pore system, which provides an optimum pore structure for the occurrence of free oil. To summarize, the free oil content shows an obvious preference in siliceous shale. The lithofacies defined by mineralogy compositions and TOC affect the oil occurrence states by pore structure. Parallelly, the pore structure was deeply affected by the grain size, spacial arrangement, and minerals (grain type). Therefore, the occurrence states of oil in shale were strongly affected by the pore structure through the lithofacies.

        Speaker: Chen Zhao (China University of Petroleum (East China))
      • 28
        Theoretical Foundation for Klinkenberg-corrected Permeability of Microporous Media in Pulse Decay Method

        We provide the theoretical foundation of directly adopting the Klinkenberg plot, the apparent permeabilities versus the reciprocal of the mean pressures, in Pulse Decay Method to eliminate the slippage effect, which is characterized by the Knudsen number and strengthens with the decrease of permeability, by asymptotic perturbation of Navier-Stokes equation in capillaric model in this work. Traditional late-time solution of Pulse Decay Method intrinsically cannot take the slippage effect into account, except for some tedious numerical attempts. We theoretically fill the gap of interpreting the experimental data got by the traditional late-time solution of Pulse Decay Method with consideration of the slippage effect. By considering the nature of the low value of permeability, asymptotic perturbation method is adopted to theoretically solve the governing equation. We show that the Klinkenberg plot can be safely used to interpret the experimental data of Pulse Decay Method for microporous media when the ratio between the pore volume of the microporous media and the upstream or downstream chamber is smaller than 0.1. This implies that when one chamber, upstream or downstream chamber, is totally sealed in experiment, the slippage effect cannot be eliminated for this kind of experimental apparatus. Our own experimental results, by comparing the intrinsic permeabilities got by Pulse Decay Method and steady-state method in different mean pressures for the same sample, verify the correctness of our derivation.

        Speaker: Tian Zhiguo
      • 29
        Understand Advection-Dispersion and their Relationship with the Scales of Heterogeneity through Lattice Boltzmann simulations

        The classical advection-dispersion equation has been a cornerstone in aquifer solute migration studies for decades. However, prevailing misunderstandings regarding advection-dispersion dynamics, their interplay with heterogeneity scales, the nature of ensemble averages, and their observational implications have sparked intense debate concerning the equation’s conceptual validity. Addressing these controversies is critical for demystifying phenomena such as macrodispersion, anomalous dispersion, and scale-dependent transport, as well as for evaluating contemporary models like the dual-domain dispersion model for solute movement in aquifers. This investigation employs the Lattice Boltzmann Method (LBM) for simulating solute transport within heterogeneous porous media. Our study delineates the evolution of the dispersion concept from molecular diffusion to encompass fluid dynamic effects caused by variations in scale-specific velocities and discusses the limitations of extending Fick's law for molecular-scale velocity variations to describe the effects of large-scale velocity variations.

        Speaker: Yanhui Dong
      • 30
        Unveiling Microbial Activity in Rock Pores: Tailored Sample Preparation and SEM-EDS Insights

        Studying the interactions among microbes within rock pores and their impact on mineral phases is essential for comprehending subsurface ecosystems and biogeochemical processes, particularly in the realm of subsurface energy operations. In this paper, we introduce an approach that merges customized sample preparation methods with traditional Scanning Electron Microscopy (SEM) featuring Energy-Dispersive X-ray Spectroscopy (EDS) to both visualize microbes in rock pores and analyze the alterations in mineral phases induced by microbial activity.
        Customizing the sample preparation process is crucial to preserving the delicate microbial structures within rock pores. We have devised a protocol that incorporates gentle fixation, embedding, and broad-ion-beam (BIB) polishing to preserve the in-situ arrangement of microbes while facilitating optimal SEM-EDS analysis. This method minimizes sample artifacts, enhancing the accurate representation of microbial structures. Through careful selection of imaging parameters and leveraging SEM imaging, we successfully visualized microbes in their natural habitat, allowing us to observe patterns of microbial colonization, biofilm formation, and interactions with mineral surfaces.
        The addition of EDS analysis complements SEM imaging by furnishing elemental composition data at micro- to nanoscale resolutions. Mapping mineral phases and identifying elemental changes induced by microbial activity provided valuable insights into biomineralization processes, dissolution, and precipitation events. Such information is crucial for understanding how microbial communities influence the mineralogical composition of rocks. Our findings highlight the diverse microbial communities residing in rock pores and their intricate impact on mineral phases. Distinct patterns of mineral alteration, including the formation of biominerals and the dissolution of specific phases, were observed. The integration of tailored sample preparation techniques with conventional SEM-EDS emerges as a straightforward yet potent tool for investigating microbial interactions in rock pores.

        Speaker: Mr Mingze Jiang (MaP - Microstructure and Pores GmbH)
      • 31
        Volatile Transport in Porous Lunar Regolith: Diffusion at Infinite Knudsen Number

        The lunar surface is covered with a layer of lunar regolith. Observational evidence[1-3] suggests that it may contain volatile substances such as water, methane, and helium-3 that could be utilized. Studying the diffusion behavior of volatiles in lunar regolith is of great significance for the exploration and exploitation of these extraterrestrial resources.

        Volatile in lunar regolith layer exist under extremely high vacuum conditions (~$10^{-9}Pa$)[4]. Under such extreme conditions, gas molecules undergo Knudsen diffusion, where the average free path is more than 10 orders of magnitude larger than the size of lunar regolith particles (Knudsen number $Kn > 10^{10}$). At this extreme (almost infinitely large) Knudsen number condition, gas molecules in porous lunar regolith rarely collide with each other, and the diffusion trajectories resemble chords (free paths) between solid surfaces which are determined solely by porous structure. Previous studies have measured tortuosity[5] or free path length distribution[6] to modify the diffusion coefficient, but the correlation between pore structure and the diffusion coefficient is still largely unexplored.

        In this study, we investigate the influence of pore structure on the diffusion of rarefied gases in porous media at infinitely large Knudsen number, based on a Monte Carlo program. Numerical experiments confirm that the linear relationship between the mean square displacement and time predicted by the Einstein equation still holds. However, the statistics of free path lengths shows clear bimodal-distribution even in homogeneous media, which is different from the unimodal-distribution as shown in porous media Fickian diffusion or in straight tube Knudsen diffusion. By statistically analyzing the molecular trajectories within the porous medium, we show that the bimodal distribution corresponds to the sizes of pore and the throat.

        According to the pore-throat bimodal distribution of free path length, we establish a bimodal random walk model to derive the diffusion coefficient from the pore and throat parameters. This analytical prediction successfully matches the numerical experiments with various structures. We further investigate the impact of adsorption and heterogeneity on the volatile transport in porous media at infinitely-high Knudsen number.

        Speaker: sunpeng zhou (College of Engineering, Peking University)
    • MS04: 1.1
      • 32
        Quantifying the Soil Swelling Potential by Soil Water Isotherm

        Swelling potential has long been used as a terminology to quantitatively describe the expansibility of soil. It encompasses multiple definitions such as the swelling pressure under specific strain constraints or the free swelling strain without confining stress. However, although these definitions are of great significance to engineering practice, they do not directly represent the physical nature of soil expansion. Resistance to mineral swelling occurs due to Van der Waals and Coulomb forces between crystal layers. The water adsorption ability within the interlayer space of expansive minerals can overcome this resistance by enabling water molecules to enter the interlayer space, resulting in macroscopic soil expansion. The interlayer hydration of expansive minerals serves as the intrinsic reason for soil swelling. Therefore, the authors propose utilizing the energy available for swelling of the interlayer space under a specific humidity as the soil swelling potential. To experimentally determine the proposed swelling potential, a framework based on the soil water isotherm (SWI), which establishes the constitutive relation between relative humidity and water content, was developed. SWI not only represents the energy state of soil water but also captures the interlayer water content change of expansive soil during wetting and drying through its hysteresis at low humidity. By leveraging the existing SWI model, which determines interlayer water content, and the method of calculating the water adsorption ability of soil (referred to as soil sorptive potential, SSP) using SWI, the authors can quantify the proposed soil swelling potential under any humidity. Several verifications were conducted to validate the proposed swelling potential. Firstly, the SWI of montmorillonite was generated using molecular simulation, combined with basal spacing measurements obtained from XRD tests under varying humidity. The interlayer water content and system energy change from the molecular simulation were analyzed to understand the volume change of mineral during wetting and drying. The energy change of soil-water system under humidity variations was used to verify the theoretical soundness of the established swelling potential framework. Secondly, the measured SWIs of different soil samples were utilized to calculate the swelling potential of these soils, thereby confirming the practical feasibility of the proposed framework. Finally, to facilitate comparison with existing indicators for identifying expansive soils, the energy used for crystal layer expansion during the water adsorption process, which can also be calculated using the proposed framework, was defined as the swelling potential index. This index exhibited superior performance in identifying expansive soils compared to other indicators of soil swelling ability. This study offers a novel perspective on the study of expansive soils and establishes a scientific basis for understanding the engineering behaviors of expansive soil under varying humidity environments.

        Speaker: Yijie Wang (The Hong Kong Polytechnic University)
      • 33
        Pore-scale network modelling of CO2–shale interaction with swelling effect

        Underground CO2 storage is a crucial effort to reach net-zero carbon emissions by 2050. By this, a large volume of gaseous CO2 is stored underground. Shale formations due to their proper storage capacity—composed of both bulk and adsorption types—can host CO2 for geo-sequestrations. Shale is also a common lithology of cap rocks over underground storage sites, which has the duty of sealing the storage site impeding upward migration of injected CO2. To understand dynamic behaviour of CO2 storage in shales, pore-scale insights are required to investigate physiochemical interactions between injected CO2 and host shale rock. Shale swelling defined as shale matrix deformation due to CO2 adsorption leading to fracture size change is to be studied in this research. In this study, we take an image-based approach to extract unstructured triple-porosity pore-network models (PNMs) based on volumetric synthetic images. At nanoscale, meso- and micro-pores of shale matrix together with fractures compose a network of pores for each shale sample. Various flow features of shale including gas rarefaction, sorption, and surface diffusion are included. It should be noted that this PNM benefits from considering gas sorption effect at pore-level, which is less included in previous shale PNM studies.

        By considering methane (CH4) as the host fluid within the pore network and considering injection of carbon dioxide (CO2) into shale samples, effect of swelling is studied. Such that, matrix deformation due to gas sorption imposes a size-reducing effect on fractures. This swelling effect is in competition with mechanical effective stress effect by which fractures are prone to size change as well. Thus, this study computes dynamic pore-scale fracture size change during the injection of CO2 into initially CH4-bearing shale samples. The results give insights into the fracture permeability change in two sets of low-density and high-density fractured shale samples. The results show how Darcy and Knudsen permeability values change over different CO2 injection pressures for both low-density and high-density fractured samples.

        Speaker: Mr Amin Taghavinejad (James Watt School of Engineering, University of Glasgow, Glasgow, United Kingdom)
      • 34
        Impact of Sand-Hydrogel Mixtures Swelling on Shearing Behaviour: An X-ray CT Study

        Hydrogels also known as superabsorbent polymers (SAPs) are crosslinked hydrophilic polymers characterized by a three-dimensional polymer network structure. These polymers are capable of absorbing water to thousands of times their own weight. Hydrogels can either be synthetic (polyacrylic acid) or natural such as biopolymers (xanthan gum). The use of hydrogels has been found to have increased utilization in diverse fields such as agriculture, environmental science, petroleum, and civil engineering. The key attractions in the use of hydrogels in these fields include the ability to absorb substantial amounts of water, selectively attracting and binding to pollutants and facilitating particle aggregation. Most previous studies have consistently demonstrated that during wetting swelling of hydrogel leads to restructuring of the soil particles and these studies claim that results of wetting affect the soil stiffness, increases erodibility, and reduces shear strength. However, we lack a more comprehensive understanding of the complex micromechanical processes that govern the behaviour of hydrogel-treated soils during wetting. This knowledge gap hinders our understanding of the interaction of hydrogels with soil and the effective use of hydrogels in soil remediation and geotechnical engineering applications. This study aims to unravel the swelling process of hydrogel in soil, how it leads to restructuring of soil particles and the resulting impact on the mechanical behaviour of hydrogel-treated soils. We employ a miniature triaxial cell connected to a humidifier and a peristaltic pump to supply humid air during saturation of the sand-hydrogel mixture. We utilize an X-ray Computed Tomography (CT) scanner to unveil the four-dimensional (3D + time) perspective of swelling-induced disturbance and its impact on the shearing behaviour. Preliminary results show that hydrogel swelling dramatically reduce sand-to-sand contacts, resulting in a much smaller peak strength. We further adopt image processing algorithms to denoise, segment phases (sand, hydrogel, and air), and label sand and hydrogel particles for performing quantitative analysis such as displacement and strain fields.

        KEY WORDS: Hydrogels; Swelling; Shearing behaviour; Soil restructuring; X-ray tomography.

        Speaker: Mr Mhlengi Masango (University College Dublin)
    • MS05: 1.1
      • 35
        Microcalorimetric Evaluation of Microbial Activity and Reaction Rate in Sand-packed Porous Media During Microbial-Induced Carbonate Precipitation For CO2 Leakage Remediation

        Carbon capture and storage (CCS) is a key technology to reach long-term climate goals that limit the temperature rise to 1.5 ◦C above pre-industrial levels. It consists in capturing CO2 from large industrial
        points and geological storage in underground formations, such as depleted oil and gas reservoirs, unminable coal beds, and deep saline aquifers [1]. The success of this technique depends on avoiding
        CO2 leakage to the surface through the complex subsurface geometric structures such as faults, fractures, and abandoned wells. Microbial induced calcite precipitation (MICP) is considered as a promising
        in-situ method for sealing subsurface leakage paths. This technique utilizes microbes to induce calcium carbonate precipitation, which effectively reduces the porosity and permeability of the porous media,
        thereby mitigating CO2 leakage risks [2].

        Complex bio-geochemical interactions considering rock-microbes-reactant solution are needed to get a broad assessment of MICP efficiency in geological porous media. However, this is not an easy task
        due to the complexity of the microbial activity and rock-forming minerals. In this study, we aim to understand the impact of particle size, specific surface area and pore volume on microbial activity and
        geochemical rates during MICP. An extremely sensitive microcalorimetry technique called Isothermal titration calorimetry (ITC) is used to assess the microbial activities and reaction rates within various
        water-saturated reservoir rocks inoculated with bacterial solutions [3]. In the ITC experiments, 100 mg of sandstone particles with different size was placed in a reaction vessel and 200μL of bacterial solution (sporosarcina pasteurii stains) was added to the rock particles [4]. The titration ampule containing the rock-bacterial solution was lowered stepwise into the calorimeter and equilibrated for 1 hour at 35 ◦C.
        Seven injections of 9.948 μL of the reactant solution (calcium chloride solution) were titrated with a time interval of 420 seconds into the slurry to determine the bio-geochemical reactions by monitoring heat
        changes. A quasi-2D sandpack (Fluidflower) was used to identify CO2 flow patterns after MICP treatment [5].

        This work shows that the bio-geochemical interactions are exothermic (thermodynamically favorable) and therefore proceed spontaneously. The reaction activity within sandstone is 10 to 18 times higher than in corresponding bulk solutions. This observation underscores the significance of available surface area in influencing both microbial colonization and the speed at which reactions occur in the MICP process.

        Speaker: Jacquelin Cobos (University of Bergen)
      • 36
        Predicting the tensile strength of sands treated via microbially induced carbonate precipitation (MICP)

        Microbially Induced Carbonate Precipitation MICP has been studied over the years as a promising bio-mediated alternative to enhance the mechanical performance of porous media. Multiple studies have investigated different MICP treatment techniques and their application to coarse-grained soils. Results of these works show the evolution of transmission properties (e.g., ultrasound waves and permeability) and the increase in strength measured by direct shear, Uniaxial Compressive Strength UCS, tensile and triaxial testing, among others. These analyses have enriched our understanding of the capabilities of MICP, yet there remains a lack of predictability regarding the soil strength that can be achieved with MICP treatments.

        This study uses an extensive dataset of triaxial results collected from the literature alongside new experimental data to propose a model that predicts the evolution of the tensile strengths in MICP-treated sands based on the calcium carbonate content achieved and the untreated soil index properties. In this study, we first analysed the influence of multiple soil characteristics on the final mechanical strength of MICP-treated sands using data from the literature. These preliminary analyses uncovered a gap in the data on the treatment of angular sands. To fill this gap, we conducted experiments using highly angular sand in which the MICP treatment strategy was varied to obtain multiple levels of cementation. After the MICP treatments were completed, the specimens were scanned via X-ray computed tomography before and after consolidated-drained triaxial tests, giving insight into the deformation behaviour during failure. The whole dataset was then used to assess the validity of an analytical model that builds on previous analyses to predict the tensile strength of MICP-treated sands. Our model results are remarkably consistent with all published datasets. Our research provides a robust framework for predicting the mechanical enhancement of MICP-treated sands, based on the mass of calcite precipitated and the untreated soil index properties. This is a critical step towards a more reliable use of bio-mediated soil enhancement techniques.

        Speaker: Gloria Castro (University of Strathclyde)
      • 37
        Pore-scale hydrodynamics influence the spatial evolution of preferential flow paths in porous media bioclogging system

        Biofilm is a universal form of microbial existence, which is formed by microbial cells and their extracellular polymers bonded to each other. It’s ubiquitous in rivers, human organs and drinking water distribution systems, where microorganisms attach to the surface of particles and cause bioclogging, which often results in negative impacts. In this paper, we developed a visualization experimental system, to realize the real-time dynamic and multi-scale observation of microbial growth under different pore structure, flow rate and nutrient concentration conditions. Visualization experimental results show that microbial growth was spatially obviously non-homogeneous due to the randomness of microbial attachment sites and preferential seepage of nutrients. In the early stage of the experiment, microorganisms mainly existed in the form of suspended cells, clusters and streams, and with the growth of microorganisms, clusters gradually coalesced to form individual biofilm clusters connecting the inlet to the outlet. In the late stage of the experiment, the biofilm formed a relatively fixed structure, and the nutrient solution mainly flowed along the preferential flow paths. Under constant flow conditions, the microbial growth led to the narrowing of the preferential flow paths, and the shear stress of the fluid would cause the preferential flow paths to become wider, the competition between microbial growth and fluid shear leads to the intermittent opening and closing of the preferential flow paths.

        Speaker: Ms Rui Peng (Wuhan University)
      • 38
        Exploiting induced carbonate precipitation to improve reservoir storage integrity and geothermal system efficiency

        Biomineralization, through microbially, thermally, or enzyme induced carbonate precipitation (MICP/TICP/EICP), is a cost-effective cementation process for changing porosity and permeability in the subsurface. This study aims to optimize compositional and injection parameters for biomineralization fluids, and to develop understanding of the interactions between geochemical reactions and fluid transport properties at the pore (micron) scale. Utilizing real-time in situ X-ray computed tomography (XCT), we compare traditional Microbially Induced Calcium Carbonate Precipitation (MICP) with novel thermally delayed (TICP) and Enzyme Induced Calcium Carbonate Precipitation (EICP) in a range of lithologies. This allows us to investigate the impact of mineralogy, grain size distribution, and temperature as well as the injection composition and strategy. We present quantitative analysis of crystal locations, the volume of carbonate and of individual crystals, and the effect of crystals on permeability and flow localisation over time. Coupled to measured changes in microstructural and macroscopic properties over repeated precipitation and dissolution cycles we present refined models of reactive transport for different injection strategies, and identify the optimal treatment strategy for different subsurface applications. This includes validation of the durability of precipitated calcite during dissolution phase simulating the behaviour of CO2-enriched brines.

        This work provides the underpinning understanding principles of crystal formation, growth and hydrodynamic feedback mechanisms necessary for accurate modelling of reservoir scale dynamic processes. However we also show how TICP and EICP strategies can improve performance of real-world Carbon Capture and Storage systems, driving more homogeneous, widely distributed and larger volumes of precipitated CaCO3 by maintaining permeability during treatment at higher degrees of cementation when compared to MICP. We also show how variable injection strategies allow improvement of other physical properties (e.g. mechanical strength) and enables the addition of highly conductive additives or phase change materials without reducing precipitation and flow. Using CaCO3 precipitation we observed a 470% increase in the thermal conductivity of unsaturated quartz sand after 9 cycles of MICP, and an 800% increase following addition of 5 wt% expanded natural graphite (ENG). Our findings also demonstrate the compatibility of integrating paraffin as a phase-change material within the porous structure of ENG prior to MICP/EICP treatment significantly increasing specific heat capacity. These new geomaterials have widespread implications for thermal energy storage, specialized geothermal grouts/backfill, shallower wells and reduced geothermal energy costs.

        The project's outcomes impact the commercialization of engineered biomineralization and its role in the subsurface energy transition.

        Speaker: Philip Salter (University of Strathclyde)
    • MS06-B: 1.1
      • 39
        Numerical modeling of the phase separation process driven by a porous membrane

        The separation of liquid and gas phases using porous media has been considered for various applications such as propellent management in aerospace, petroleum engineering, carbon storage and so on. However, such a process is usually difficult to model as the multiphase flow involving porous media usually spans several characteristic lengths, and the interface conditions between the free fluid and the porous media flow are relatively complex. A direct numerical simulation (DNS) using fully resolved Navier-Stokes equations is limited by the immense computational cost. In this work, a mixed-dimensional flow model was developed for the highly coupled free fluid and thin porous media flow, in which the flow process in the porous membrane was reduced to one dimension. Besides, the parameter transfer equations at the regional interface were evaluated, and the proposed model was validated against DNS results. The effects of pore structures, porosity, and flow direction on the two-phase transport process were studied in detail. The present study will provide guidance for the optimization of phase separation systems driven by thin porous media.

        Keywords: phase separation, multiphase flow, numerical modeling, coupled flow, porous media

        Acknowledgements
        This work is supported by the National Natural Science Foundation of China (Grant No. 52276013)

        Speaker: Mengyi Jiang (Shanghai Jiao Tong University)
      • 40
        The mathematical model and analysis of the nanoparticle-stabilized foam displacement

        This work proposes a mathematical model to study the foam displacement in porous media stabilized by nanoparticles [1]. We consider a simplification of the Stochastic Bubble Population balance model in local equilibrium, with nanoparticle dependence inspired by the experimental data from the literature. It consists of a non-strictly hyperbolic system of conservation laws, which is solved for the generic initial and injection conditions. We investigate the existence of a global solution as a sequence of waves following the Conservation Laws Theory and the procedure proposed in [2], where a similar problem was solved for a two-phase flow containing an active tracer (with linear adsorption). When the solution is composed of two or more waves, we present necessary and sufficient conditions to guarantee the compatibility of these wave sequences. The analytical solution for the nanoparticle-stabilized foam displacement in porous media allowed us to quantify the effect of nanoparticles on foam displacement, focusing on the breakthrough time and cumulative water production. In agreement with the literature, when only gas is injected, the breakthrough time and the water production increase with the nanoparticle concentration. Although, we also observe that the effect of nanoparticles is less pronounced for high nanoparticle concentration. Counterintuitively, adding nanoparticles changes the mathematical solution qualitatively, yielding a negligible effect on water production during gas-water co-injection for a certain parameter range. We discuss the most favorable conditions to observe the action of nanoparticles in laboratory experiments.

        Speaker: Tatiana Danelon de Assis
      • 41
        The Impact of System Softness on Haines Jumps in Porous Media

        Haines jump is an interfacial instability characterized by fluid redistribution and sudden pressure changes. It is a pore-scale phenomenon that occurs during displacement front movement and is widespread in multiphase flow processes in porous media. It is an essential physical process that affects fluid distribution, flow regimes, and displacement efficiency. Previous studies have shown that Haines jump is an instability phenomenon taking place in a soft system comprised of entrapped gas bubbles, deformable porous media, and interacting menisci. This study aims to further investigate the impact of system softness on Haines jumps in porous media.
        We conduct fluid displacement experiments in polydimethylsiloxane (PDMS) microfluidic chips and analyze the pressure signature and flow phenomena during Haine jumps. A syringe pump (Chemyx Fusion 200) maintains a constant injection rate, and a microscope (ZEISS) records the displacement process. A microfluidic pressure sensor (Fluigent) placed at the inlet measures the pressure change. The system softness is controlled by adjusting the volume of an air bubble entrapped near the inlet. The PDMS microfluidic chip is fabricated using the standard soft lithography technique. De-ionized water and ethanol are used as wetting fluids, and air is the nonwetting fluid. The impact of system softness is investigated in three conditions: single pore throat, pores in series, and pore network.
        Results show that system softness affects the position where Haines jumps occur at the pore throat and the distance of the interface jump. The interface jumps a longer distance when the entrapped air bubble volume increases. It also takes a longer time for the interface to reach the unstable point. An analytical model is proposed to explain the pressure change and interfacial jump at the pore throat and matches well with the experimental observation. In cases of pores in series, the pressure signature shows a “saw” shape where sudden pressure changes and interface jumps occur at pore throats. However, the interface can jump across more than one pore as the system softness increases. As the system softness affects the magnitude of Haines jumps, it also influences the snap-offs, entrapped fluid saturation, and displacement efficiency in pore networks.
        In summary, the influence of system softness on Haines jumps and its impact on fluid displacement are analyzed in detail in both single pores and pore networks. Our experiments provide insights into the pore-scale physics of Haines jumps and the impact of Haines jumps on multiphase flow in porous media.
        Keywords: Haines jump, porous media, microfluidics, instability, multiphase flow, interface

        Speaker: Zhonghao Sun (Xi'an Jiaotong University)
    • MS07: 1.1
      • 42
        Simulation of density-driven flow in heterogenous and fractured porous media

        The study of fractured porous media is an important and challenging problem in hydrogeology. One of the difficulties is that mathematical models have to account for heterogeneity introduced by fractures in hydrogeological media. Heterogeneity may strongly influence the physical processes taking place in these media. Moreover, the thickness of the fractures, which is usually negligible in comparison with the size of the whole domain, and the complicated geometry of fracture networks reduce essentially the efficiency of numerical methods. In order to overcome these difficulties, fractures are sometimes considered as objects of reduced dimensionality (surfaces in three dimensions), and the field equations are averaged along the fracture width.
        Fractures are assumed to be thin regions of space filled with a porous material whose properties differ from those of the porous medium enclosing them. The interfaces separating the fractures from the embedding medium are assumed to be ideal. We consider two approaches: 

        (i) the fractures have the same dimension, d, as the embedding medium i.e. they are d-dimensional; 

        (ii) the fractures are considered as (d-1)-dimensional manifolds, and the equations of density-driven flow are found by averaging the d-dimensional laws over the fracture width.
        We show that the second approach is a valid alternative to the first one. For this purpose, we perform numerical experiments using a finite-volume discretization for both approaches. The results obtained by the two methods are in good agreement with each other.
        We derive a criterion for the validity of the simplified representation. The criterion characterises the transition of a mainly parallel flow to a rotational flow, which can not be reasonably approximated using a d-1 dimensional representation. We further present a numerical algorithm using adaptive dimensional representation.

        Speaker: Gabriel Wittum (G-CSC, Goethe-Universität Frankfurt)
      • 43
        Multilevel Monte Carlo Method for Simulation of Propagation of Uncertainties in Fractured Porous Media

        Simulation of salinization of coastal aquifers plays an important role in prediction of availability of pure water resources. In these geological formations, fractures introduce strong heterogeneities and their influence on the groundwater flow and the transport of the salt is significantly (cf. [1]). However uncertain variations in hydrogeological parameters such as porosity, permeability, fracture aperture etc. may essentially reduce accuracy of the prediction of the transport phenomena. In this talk, we present an application of the multilevel Monte Carlo method for estimation of propagation of the uncertainty from the parameters of the fractured porous medium to the solution in the subsurface density-driven flow model represented by a system of non-linear PDEs. This research is a continuation of our recent work on the uncertainty quantification for this type of models (cf. e.g. [2]). We test this approach on a model problem with the random porosity field, recharge and fracture aperture that represent the limited knowledge of the data. Parallelization is applied to the Monte Carlo method. We present results of numerical experiments on the supercomputer Shaheen II.

        Speaker: Dr Dmitry Logashenko (King Abdullah University of Science and Technology)
      • 44
        A Hybrid-Dimensional Stokes--Brinkman--Darcy Model: Derivation, Analysis and Validation

        Coupled free-flow and porous-medium systems have received rising attention in recent years due to their broad applications in the environment, biology, and industry. A suitable coupling concept should be applied to characterize fluid behavior between the free flow and porous medium. However, the majority of coupling conditions are restricted to flows parallel to the fluid-porous interface.

        In this talk, we present a hybrid-dimensional model for coupled free-flow and porous-medium systems which is suitable for arbitrary flow directions. We consider a narrow transition region between these two flow systems that stores and transports mass, momentum, and energy. The proposed hybrid-dimensional model incorporates the Stokes equations in the free flow, the averaged Brinkman equations along the transition region, and Darcy's law in the porous medium. Appropriate transmission conditions are considered between the three regions. The well-posedness of the developed hybrid-dimensional model is proven. The model is validated against the pore-scale resolved simulations and compared with other coupling concepts. Numerical simulation results demonstrate the advantages of the proposed model in comparison to the coupling concepts available in the literature.

        Speaker: Linheng Ruan (Institute of Applied Analysis and Numerical Simulation, University of Stuttgart)
    • MS08: 1.1
      • 45
        Dispersion and Straining Behaviors of Non-Spherical Suspended Particles in Saturated Randomly Packed Beads: A Numerical and Theoretical Study

        Understanding the transport of particles in porous media, including dispersion and straining, plays a pivotal role in optimizing various engineering processes, such as drug delivery, wastewater treatment, and fracking proppants displacement. While prior numerical endeavors have significantly expanded our understanding of the microscopic behavior of particles within porous media, they have frequently overlooked the shape anisotropy of particles. When particles are non-spherical, such as pills, bacteria, and microplastics, the shape anisotropy of particles may determine their dispersion and straining behaviors in porous media, even in weakly heterogeneous environments like beadpacks.

        To bridge this knowledge gap and evaluate our hypothesis, we simulated the Lagrangian transport of 3-D non-spherical particles through a 3-D porous network generated by a randomly sedimented, saturated bead pack, employing a computational fluid dynamics-discrete element method (CFD-DEM) approach. To account for the particles’ asphericity as well as its impact on particle transport, we modeled the particles as superquadrics of varying asphericities and implemented a particle-fluid two-way coupling algorithm, where the fluid flow influences particles’ motion, and conversely, particles also affect the fluid streamlines.

        Our results suggest that, compared to spherical particles, highly aspherical particles tend to migrate along streamlines more readily, resulting in a higher mean dispersivity; such particles also tend to sweep a larger volume of the pore space, leading to a more uniform spatial distribution of retained particles. To support our numerical observations, we report a particle velocity probability distribution function that encompasses the impact of particles’ asphericity on their dispersion and straining behaviors. Said function compiles all numerical observations and distinguishes between the straining and dispersion characteristics. We also deliberate on the similarities and differences between this new function and the function applicable to spherical particles, as previously reported [1]. The presented function can be useful in designing particle topology to achieve specific velocity distributions or mean dispersivity of interest.

        Speaker: Mr Yaoming Chen (Southern University of Science and Technology, China University of Petroleum (East China))
      • 46
        Dispersion Control in Fractured Multi-Layer Porous Media System

        Multi-layered porous media are present in a variety of natural and engineered systems, and their structure can have a significant impact on flow and transport processes. This study proposes a hybrid analytical-numerical solution to examine the relationship between scalar dynamics and media properties in coupled systems comprising a two-dimensional free flow layer and a heterogeneous porous medium operating under fully developed laminar flow conditions. Perturbation and homogenization methods are used to obtain a set of one-dimensional upscaled equations for passive scalar transport. These equations are then used to develop a semi-analytical solution based on integral transforms, which enables the relationship between the properties of the porous system and scalar mixing and spreading to be determined. To validate the solution for the upscaled system, we compare the results with numerical findings for two-dimensional scalar transport. In addition, we analyze the influence of the multilayered system on macroscopic transport by examining the breakthrough curve, dispersion coefficient, and mixing of the scalar cloud. The results suggest that the semi-analytical solution can be used to optimize and determine the arrangement of porous media properties to achieve desired mixing objectives.

        Speaker: Bowen Ling (Stanford University)
      • 47
        Non-monotonic effect of compaction on dispersion coefficient of porous medium

        Numerous studies have shown a non-monotonic relationship between the dispersion coefficient and the degree of compaction of porous media [1, 2]. However, the mechanism responsible for the non-monotonic variations of the dispersion coefficient remains unclear, which brings difficulties and challenges for the regulation of the dispersion coefficient of porous media.
        By combining the discrete element method and the pore network model, we investigate the impact of compaction on the dispersion coefficient of the porous medium. The dispersion coefficient exhibits a non-monotonic dependence on the degree of compaction, which is distinguished by the presence of three distinct regimes in the slope of the dispersion coefficient to the pressure load. The non-monotonic variation of the dispersion coefficient is attributed to the disparate effect of compaction on dispersion mechanisms. Specifically, the porous medium becomes tightly packed with increasing pressure load, reducing the effect of molecular diffusion that primarily governs at small Péclet numbers. Simultaneously, the elevated pressure load reinforces the heterogeneity of the pore structure while reducing its connectivity, leading to enhanced disorder and elevated proportion of low-velocity regions within the porous media flow, further strengthening mechanical dispersion and hold-up dispersion, respectively, which dominate under high Péclet numbers. The competition between weakened molecular diffusion and enhanced hold-up dispersion and mechanical dispersion, together with the shift in the dominance of dispersion mechanisms across various Péclet numbers, results in multiple regimes in the slope of the dispersion coefficient to the pressure load. Our study provides unique insights into the structural design and modulation of the dispersion coefficient of porous materials.
        Keywords: dispersion; compaction; non-monotonic effect.
        Reference
        [1] E. Charlaix, J.P. Hulin, T.J. Plona, Experimental study of tracer dispersion in sintered glass porous materials of variable compaction, Physics of Fluids 30 (6) (1987).
        [2] C.T. Karin C.E. Östergren∗, Characterization of hydrodynamic dispersion in a chromatographic column under compression, Chem. Eng. J. (2000).
        [3] Y. Liu, W. Gong, Y. Zhao, X. Jin, M. Wang, A Pore-Throat Segmentation Method Based on Local Hydraulic Resistance Equivalence for Pore‐Network Modeling, Water Resour. Res. 58 (12) (2022) e2022WR033142.

        Speaker: Yang Liu
      • 48
        Probing Transport in Geologic Porous Materials by Fast X-ray Micro-Computed Tomography

        Understanding the time evolution of solute transport at the pore scale in geologic porous media is crucial in many subsurface engineering applications, including underground gas (e.g. H2 and CO2) storage. While transport processes in rocks can be investigated by dynamic 4D imaging, such as X-ray and neutron-based computed tomography, most of the observations so far have been limited to the continuum scale (image resolution approx. 1 mm and above). These observations have improved our understanding of non-Fickian transport in rocks [1]. Yet, the lack of spatial resolution of these methods precludes the unambiguous interpretation of the transport mechanisms at play, because the relevant mixing processes initiate at the pore scale. Recent advances involve the uses of fast X-ray micro-computed tomography (fast µCT), which allows direct micron-scale imaging of fluid transport at a time resolution on the order of tens of seconds.

        Here, we analyse a comprehensive data set of 4D imagery acquired by fast µCT available on the Digital Rocks Portal [2,3]. The dataset consists of dynamic images of solute transport during miscible displacement in three porous media (diameter: 6 mm, and lengths: 16 or 20 mm), namely sintered beadpack, Bentherimer sandstone, and Savonnières limestone. Tracer tests were performed at various Péclet numbers, covering the range Pe = 2 – 20. In each test, fast µCT scans were continuously acquired at the sample inlet with the field of view of 8.8 (H) x 8.8 (W) x 5.0 (L) mm, and at spatial and time resolutions of 14 µm and 12 s, respectively. We have analysed this rich dataset by applying the concept of residence time (RT) at different length-scales and by considering its spatial variability within the sample. Specifically, we computed RT curves for individual pore volume elements (PVE) in the sample (> 100 curves) and identified variations with both size and spatial location of the PVE. The strength of pore-scale heterogeneity is thus quantified by comparing the experimental RT with the theoretical counterparts. To quantify the degree of non-uniformity of the concentration field, we also examined the extent of mixing within individual PVE using various metrics, including the dilution index, the intensity of segregation and the spreading length.

        Speaker: Takeshi Kurotori (Imperial College London)
    • MS09: 1.1
      • 49
        Direct pore-scale modeling of foam flow through 3D rough fractures

        Foams are dispersions of gas bubbles within a liquid. They can be generated in porous and fractured media during co-injection of two fluids in the presence of surfactants that stabilize foam bubbles. Since foam viscosity is much higher than the constituent gas and liquid phases, they are used for diverting fluid to less permeable subsurface formations in applications such as enhanced oil recovery or carbon dioxide storage. Pore geometry, thermodynamic conditions, molecular structure and behavior of stabilizing agents such as surfactants or nanoparticles near fluid/fluid or fluid/solid interfaces are some important factors affecting stability and regeneration of foam in porous media.
        We present the first 2D and 3D results from a direct pore-scale modeling of foam in porous media. We use free-surface lattice Boltzmann method adapted for simulation in an imaged realistic fracture geometry. All geometries are available from Digital Rocks Portal, https://www.digitalrocksportal.org/. The model couples Navier-Stokes equation for fluid flow between bubbles and diffusion of dissolved gas within liquid into bubbles, and is adapted from LBfoam (https://github.com/mehdiataei/LBfoam) open solver that originally does not account for porous medium. To our knowledge, this is the first 3D model with the foam flow driven by pressure gradient in a fractured porous medium, gas diffusion through liquid phase and the interface changes as a result of both aforementioned mechanisms at pore scale. The bubble morphological variations we observe are caused by bubble coalescence, splitting and merging and we quantify them using morphological parameters (Minkowski functionals) in different liquid viscosity, pressure gradient, surface tension and temperature conditions.

        Speaker: Xuesong Ma
      • 50
        Numerical simulation of two-phase flows in digital core samples with underresolved porosity

        Micro-CT scans are widely used for rock models in Digital Rock Physics applications. However, up to one-half of the connected porosity of carbonates and shales is underresolved with micro-CT due to the small pore size. This underresolved pore space may still support multiphase fluid flow. To simulate two-phase flows in models where both large-scale and underresolved pores are present, we developed a numerical algorithm based on the combination of the phase-field model with two-phase filtration, which supports continuous phase transport in a multi-scale pore space.
        The fluid flow is simulated using the unified Navier-Stokes-Brinkman equation, which is well suited for the models where the absolute permeability is at the level of mcroDarcy, which is the case for the underresolved porosity of carbonates and shales. This equation is solved using the projection-based method. The phase transport in the resolved pores is governed by the Cahn-Hilliard equation of the phase field, which makes it simple to treat the complex geometry and topology of the pore space and the phase. Phase transport in the underresolved pores satisfies the two-phase filtration equation, accounting for the capillary pressure. The two models are coupled at the interface between the resolved and underresolved pores based on flux continuity. Additionally, the wetting-angle boundary condition is satisfied for the phase-field model.
        The designed algorithm and its GPU-based implementation are used to estimate the relative permeability and capillary pressure of the samples with underresolved porosity.
        The research was supported by RSCF grant no. 21-71-20003

        Speaker: Vadim Lisitsa (Institute of Mathematics SB RAS)
      • 51
        Volume of Fluid based study of the three phase dynamic contact line in wetting of the nanometric rough micro-channels

        The precise representation of molecular motion near the three-phase dynamic contact line remains a significant research challenge [1], with substantial practical implications [2]. We investigate the two-phase flow in a pressure driven micro channel (width ~ 1µm - 10µm) having a nanometric surface roughness. The two phases are separated by an interfacial layer with surface tension, that meets the moving pipe wall, hence, a three phase dynamic contact line is formed. Numerical simulations are conducted by solving the 2D two-phase Navier-Stokes equation using the Basilisk flow solver. The Volume-of-Fluid method is employed to capture the interface, and the surface tension force is computed using the Continuous Surface Force method. Additionally, curvature calculation is done using height functions. To address the influence of surface roughness, we develop a hybrid Volume-of-Fluid coupled embedded boundary solver. This hybrid solver enables the imposition of a contact angle on arbitrarily shaped solids. We explore scenarios where (a) surface roughness exhibits periodicity, (b) the surface is scratched or bumped with rough patches, and (c) surface heterogeneities are present. The study quantitatively demonstrates the emergence of stick-slip behavior in these scenarios allowing us to verify the thesis of Hocking [3] and Jansons [4]. Our findings serve as a prerequisite for full pore-scale Direct Numerical Simulation (DNS), ensuring a high-fidelity representation of dynamic wetting phenomena in porous media.

        [1] Lācis U, Pellegrino M, Sundin J, et al. Nanoscale sheared droplet: volume-of-fluid, phase-field and no-slip molecular dynamics. Journal of Fluid Mechanics. 2022;940:A10. doi:10.1017/jfm.2022.219

        [2] Liu C-Y, Vandre E, Carvalho MS, Kumar S. Dynamic wetting failure and hydrodynamic assist in curtain coating. Journal of Fluid Mechanics. 2016;808:290-315. doi:10.1017/jfm.2016.594

        [3] Hocking LM. A moving fluid interface on a rough surface. Journal of Fluid Mechanics. 1976;76(4):801-817. doi:10.1017/S0022112076000906

        [4] Jansons KM. Moving contact lines on a two-dimensional rough surface. Journal of Fluid Mechanics. 1985;154:1-28. doi:10.1017/S0022112085001392

        Speaker: Dr Tianyang Han (Sorbonne University)
    • MS13: 1.1
      • 52
        Anomalous phase transition behavior of dilute electrolyte solutions in nanoconfinement under cryogenic environment

        The impact of confined spaces on the phase transition of water or electrolyte solutions has garnered considerable interest due to their widespread occurrence in natural processes and technological applications. [1-3] Specifically, phenomena such as freezing, melting, and vapor condensation have been extensively studied. Recently, there has been growing attention towards the phase transition of aqueous solutions, as they represent situations closer to natural and realistic scenarios than pure water.[4-5]
        Notably, electrolyte solutions exhibit a slight variation from pure water, wherein an ice-salt phase separation occurs at low temperatures. The phase transition starts when ice or salt precipitates and continues until the eutectic point is reached. The eutectic point, being the lowest temperature at which a liquid solution can exist, remains constant regardless of the initial molality, and decreases with decreasing pore radius.[4] Anomalous behavior in confinement arises when the initial concentration is significantly diluted. In all studies involving a dilute solution in confinement, the thermal signal of ice and salt crystallization at the eutectic point was not observed, neither in bulk nor in the pore.[3,5]
        In this work, we conducted a systematic calorimeter measurement to analyze the influence of salt molality, pore filling extent, and pore size on the transition routine of CaCl2 and NaCl solution. The results indicate that in situations with fewer salts, such as smaller pores, dilute solutions, or lower filling degrees, only water freezing at the beginning is detected, and the eutectic transition is absent due to the lack of available free ions for crystallization in the pore center. Conversely, the eutectic transition could be well detected in the solution with a larger amount of salts. This may be attributed to the uneven distribution of cations and anions in pores[6]

        Reference
        1. T. Talreja-Muthrejia, K. Linnow, D. Enke, M. Steiger, Deliquescence of NaCl Confined in Nanoporous Silica, Langmuir, 38, 36 (2022) 10963-10974.
        2. X. Wang, G. Shi, S. Liang, J. Liu, D. Li, G. Fang, R. Liu, L. Yan, H. Fang, Unexpectedly high salt accumulation inside carbon nanotubes soaked in dilute salt solutions. Physical Review Letters, 121, 22 (2018), 226102.
        3. E. Jantsch, C. Weinberger, M. Tiemann, T. Koop, Phase transitions of ice in aqueous salt solutions within nanometer-sized pores. The Journal of Physical Chemistry C.123, 40 (2016) 24566-24574.
        4. J. Meissner, A, Prause, G. H. Findenegg, Secondary confinement of water observed in eutectic melting of aqueous salt systems in nanopores. The Journal of Physical Chemistry Letters. 7, 10 (2016), 1816-1820.
        5. M. Koniorczyk, D. Bednarska, Kinetics of water freezing from inorganic salt solution confined in mesopores. Thermochimica Acta, 682 (2019), 178434.
        6. M. Argyris, D. R. Cole, A. Striolo, Ion-specific effects under confinement: the role of interfacial water, ACS Nano, 4 (2010) 2035-2042.

        Speaker: Mr Shaoheng Wang (University of Hamburg)
      • 53
        Effect of partial saturation on acoustic properties of nano-porous media

        Nanoporous materials provide high surface area per unit mass and are capable of fluids adsorption. While the measurements of overall amount of fluid adsorbed by a nanoporous sample are straightforward, probing the spatial distribution of fluids is non-trivial. For macro-porous media the effect of partial saturation on acoustic properties is described by the theory of poroelasticity.

        To test applicability of poroelastic patchy saturation models to nano-porous materials, we consider ultrasonic measurements during adsorption and desorption of n-Hexane vapor on nanoporous Vycor glass (Page et al., 1995). As vapor pressure is increased from zero to the saturation pressure, the vapor is adsorbed on the pore walls, resulting in gradual increase of the liquid fraction. The reverse process occurs when pressure is decreased, but the ‘drying’ of the nanopoorus glass is heterogeneous, resulting in a very different velocity-saturation relationship.

        On adsorption, we model ultrasonic properties of partially saturated glass using Continuous Random Model (CRM) of Müller and Gurevich (2005), also known as the Dynamic equivalent medium approach (DEMA). In this model, the liquid fraction is considered a random function of position, controlled by the correlation length d (“patch size’), which may itself vary with saturation.

        As noted by Kobayashi and Mavko (2016), during imbibition, some significant portion of the liquid fraction should be uniform. In other words, if we consider the medium to be saturated with a binary mixture of two fluids, one of these fluids should be liquid, while the other should be a uniform mixture of liquid and vapor with liquid fraction SL0, which itself increases with the increasing overall liquid saturation SL. This is even more so for nanoporous media, where adsorption tends to produce rather uniform patterns. Our modelling shows that there is a strong coupling between the patch size and uniformly saturated fractions, which cannot be resolved with ultrasonic data only. However, this can be resolved using light scattering data (Ogawa and Nakamura, 2013). Very weak light scattering during adsorption shows that 99.3% of the increase of the saturation is uniform. Yet the saturation is not entirely uniform, as shown by the deviation of the longitudinal modulus from the unform saturation limit (as discussed in the next section).

        The desorption process results in macroscopic liquid patches, and cannot be modelled with CRM. We model this process with elastic finite element methods.

        Our calculations show that on adsorption the characteristic patch size is of the order of 100 pore diameters, while on desorption the patch size is comparable to the sample size. These results are supported by optical data for similar systems. Our analysis suggests that one can employ ultrasound to probe the uniformity of fluid spatial distribution in nanoporous materials.

        Speaker: Boris Gurevich (Curtin University, Australia)
      • 54
        Slip correction theory and transient solution of the pressure oscillation method

        The pressure oscillation method is a widely employed technique for measuring the permeability of time-varying and tight porous media. The previous analytical solution for permeability calculation neglects the unsteady-state condition of the slip effect, and the application of the Klinkenberg correction lacks theoretical support. Existing permeability calculations rely on the periodic part, and the utilization of the transient part needs further development. In addition, parameter regulation in experiments incurs trial-and-error costs, and the reasonable prediction of parameter setting is necessary. In this study, the analytical solution of the pressure oscillation process considering the slip boundary is derived based on the capillary model and perturbation expansion. The correspondence between the Klinkenberg correction relation and the Knudsen number is clarified, which provides a theoretical basis for applying Klinkenberg correction to the pressure oscillation method. A new data processing method is proposed for permeability calculation based on the transient solution, and the scope of application of the Klinkenberg correction for the new method is given. Experiments of the pressure oscillation method and pulse decay method are carried out to validate the theoretical model and data processing method. Through comparison of the permeability measurement results, the transient solution is consistent with the periodic solution, and the unification of the quasi-steady-state and unsteady-state methods under pressure oscillation conditions is achieved. In contrast to the pulse decay method, the pressure oscillation technique exhibits advantages in terms of measurement duration. Under conditions of higher permeability, a tenfold increase in measurement speed can be attained, while under lower permeability conditions, there is a minimum threefold improvement. Through the inverse solution process for permeability calculation, this study analyzed the main factors influencing measurements in the pressure oscillation method. The reason for the inaccuracy of porosity measurement is that porosity is extremely sensitive to the amplitude ratio and the phase difference, and the measurement error is magnified several times. The contour of the amplitude ratio response based on the dimensionless number is established to provide a reference for the selection of experimental parameters for practical engineering applications.

        Speaker: Mingbao Zhang (Tsinghua University)
      • 55
        Nanoporosity controls on the carbon storage and mineralization potential of basalts: insights from hydrothermal alteration at Newberry Volcano

        This study quantifies the pore structures and reactive flow capacity of basalt rocks, specifically a range of flow top (vesicular) and seal basalt samples from Newberry Volcano drill core (Oregon, USA). Dissolution and precipitation reactions in basalts and other mafic and ultramafic rocks (silicates rich in Mg, Ca, and Fe) are the foundation for carbon mineralization, in situ mining, and geologic hydrogen technologies due to high contents of reactive minerals (e.g., pyroxene and olivine) or critical minerals. However, despite their high reactivity, these rocks exhibit large variations in porosity and permeability. In fact, many of these rocks are nanoporous (<1um) or poorly connected, thereby challenging fluid access to critical and reactive mineral surfaces. The bulk volumes of qualifying basaltic and mafic/ultramafic rocks are vast, with the US’s Pacific Northwest alone estimated to have 10^5 Gt CO2 basalt storage/mineralization capacity, assuming pore space and reactive minerals are accessible to fluid flow. Towards addressing accessible pore space and reactive minerals surfaces, this work examines the (a) pore size distribution and connectivity and (b) accessible mineral surface area within Newberry volcano basalt samples.

        The Newberry Volcano basalts provide a natural laboratory for understanding multiphase and reactive flow potential of basalt pore structures. We collected 20 basaltic sandstone/volcaniclastic rocks and basalt/basaltic andesite rocks from USGS Newberry volcano drill core. The studied samples have been subjected to varied amounts of gases (CO2, H2S) and aqueous fluids in situ. The samples display varied degrees of hydrothermal alteration, where rock properties (porosity, permeability, lithology), temperature, and fluid composition dictate the extent of alteration. We characterize the macropores (vesicles) and nanopores (clays and matrix) of samples with N2 adsorption-desorption isotherms (BET surface area), TD-NMR T2, pycnometry, thin section and SEM/EDS, and microCT. The amount of Fe2O3, MgO, and CaO ranges from 18% (volcanic siltstone)-25% (flow-top basalt), confirming the reactivity of these Newberry volcano rocks. Direct numerical pore-scale simulations are used to study fluid flow capacity in image- and process-based domains for characteristic sample pore-scale features. We find distinct differences in pore structures among lithologies: for example, the volcaniclastic siltstones are dominated by a bimodal distribution, fresh/seal samples are dominated by a unimodal distribution, and multimodal distribution is significant in all hydrothermal altered basalts. Combining these distributions with imaging and modeling supports that nanoporosity is a driver of reactive flow capacity in basalts: Primarily nanoporous samples with extremely low permeability remain relatively “fresh” (unaltered) over geologic time. In originally more porous samples, aqueous fluids have altered primary minerals (plagioclase) into clays, quartz polymorphs, and carbonates which fill the pore systems, resulting in a secondary nanoporosity system. All studied hydrothermal altered basalts have similar pore size distribution and mainly contain slit-shaped pores (per BET analysis); the signal is likely dominated by clays. Overall, the natural CO2-fluid-rock system of the Newberry Volcano can be leveraged to understand anthropogenic CO2 movement in basalts. The alteration-nanoporosity-flow capacity feedback loop summarized in this work has implications for basalt storage capacity and seal performance for the aforementioned energy transition applications, especially CO2 storage and mineralization.

        Speaker: Shaina Kelly (Columbia University)
    • MS15: 1.1
      • 56
        Machine learning accelerated molecular simulation: Implications for oil and gas problems

        In the last few decades, deep learning (DL) has afforded solutions to macroscopic problems in petroleum engineering, but mechanistic problems at the microscale have not benefited from it. Mechanism studies have been the strong demands for the emerging projects, such as the gas storage and hydrate production, and for some problems encountered in the storage process, which are common found as the chemical interaction between injected gas and mineral, and the formation of hydrate. Emerging advances in DL technology enable solving molecular dynamics (MD) with quantum accuracy. The conventional quantum chemical method is computational expensive, whereas the classical MD method cannot guarantee high accuracy because of its empirical force field parameters. With the help of the DL force field, precision at the quantum chemistry level can be achieved in MD. Moreover, the DL force field promotes the computational speed compared with first-principles calculations. In this study, the basic knowledge of the molecular force field and deep neural network (DNN) is first introduced. Then, three representative open-source packages relevant to the DL force field are introduced. As the most common components in the development of oil and gas reservoirs, water and methane are studied from the aspects of computational efficiency and Chemical reaction respectively, providing the foundation of oil and gas researches. However, in the oil and gas problems, the complex molecular topo structures and various element types set a high challenge for the DL techniques in MD. Regarding the computational efficiency, it needs improvement via GPU and parallel accelerations to compete with classical MD. Even with such difficulties, the booming of this technique in the area of petroleum engineering can be predictable.

        Speaker: Dr Jie Liu (King Abdullah University of Science and Technology)
      • 57
        CO2 Leakage Detection using Optimized Deep Learning
        1. OBJECTIVE/SCOPE
          Geologic CO2 sequestration (GCS) has been considered as a promising engineering measure to reduce global greenhouse emission. Real-time monitoring of CO2 leakage is an essential aspect of large-scale GCS deployment. This work introduces a deep-learning-based algorithm using a hybrid neural network for detecting CO2 leakage based on bottom-hole pressure measurements.

        2. METHODS, PROCEDURES, PROCESS
          The hybrid neural network, called CNN-BiLSTM, leverages the strengths of convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM), in which CNN is used for spatial feature extraction and BiLSTM is applied for temporal dependency recognition. The CNN-BiLSTM enables us to build a spatial-temporal-based image-to-value regression model to learn the nonlinear mapping between high-dimensional input data (e.g., permeability, porosity, injection rate) and predicted bottom-hole pressure as output. The proposed workflow incorporates the generation of train-validation samples, the coupling process of training-validating, and the model evaluation. In this work, the diffusivity equation for pressure is solved within the CMG framework used to generate datasets under no-leakage conditions. A Bayesian optimization process is performed to optimize the network architecture, model hyperparameters, and the ratio of train to validation samples.

        3. RESULTS, OBSERVATIONS, CONCLUSIONS
          We test the CNN-BiLSTM performance on the bottom-hole pressure data collected from CO2 leakage simulations. Results show that the CNN-BiLSTM model can successfully detect CO2 leakage events by comparing the difference between the predicted (no leakage) and tested bottom-hole pressures. We further compare its superiority with CNN, LSTM, BiLSTM, and CNN-LSTM. Our proposed model achieves the highest accuracy with the same datasets. The CNN-BiLSTM outperforms other models owing to 1) its capacity to process image-based input, which could accurately capture input formation, especially cases with highly heterogeneous permeability; 2) its bidirectional ability to capture time-series dependency. Other models, like LSTM and BiLSTM, take value-based input, which is insufficient to describe the input information in highly heterogeneous cases. In contrast, CNN model suffers from capturing the temporal dependency features. Because of the bidirectional feature, CNN-BiLSTM shows higher accuracy, even 10% when applied to a small number of datasets, than the CNN-LSTM model.

        4. NOVEL/ADDITIVE INFORMATION
          We propose a hybrid neural network featuring Bayesian optimization for CO2 leakage detection. We demonstrate its applicability in terms of accuracy and robustness with other models. The proposed workflow can be implemented in commercial-scale GCS for real-time monitoring applications.

        Speaker: Dr Xupeng He (Saudi Aramco)
      • 58
        Prediction of Upscaled Permeability of Digital Rock Cores Using Machine Learning Techniques

        A workflow has been proposed to directly predict the upscaled absolute permeability of a rock core from CT images whose resolution is insufficient to directly calculate the pore-scale permeability. The workflow employed the deep learning technique with the raw CT image data of rocks and their corresponding permeability values, which were obtained through high-resolution flow simulation on these images.
        A binarized pore-geometry model was first cropped and reconstructed from the high-resolution micro CT images of the rock sample. Then the binarized model was divided into small subsamples of dimensions $100\times100\times100$, $200\times200\times200$, and $300\times300\times300$. To meet the demand of large datasets, the subsamples are allowed to be overlapped during the extraction. The permeability of these subsamples was calculated by the lattice Boltzmann flow solver.
        The calculated permeability and corresponding subsample’s low-resolution CT image pairs are then input as a dataset to train a neural network (Resnet34). Using the trained Resnet, the permeability map of an extended region in the rock core can be accurately predicted.
        Finally, the Darcy flow solver calculates the upscaled absolute permeability of the entire rock core. In this study, we explore the influence of the digital rock subsample core's dimensions and the quantity of training data on the prediction accuracy and upscaled absolute permeability. Three subsample dimensions, specifically $100\times100\times100$, $200\times200\times200$, and $300\times300\times300$, were evaluated.
        At a subsample dimension of $100\times100\times100$, the prediction accuracy was unstable, and good prediction accuracy was not obtained because of the presence of outliers within the subsample where only solid or only void were present. When the subsample dimension was $200\times200\times200$, the training accuracy was stable and acceptable prediction accuracy was attained when the subsample data size exceeding 20,000. Expanding the sample dimension to $300\times300\times300$, it was observed that 10,000 training data points were sufficient to achieve satisfactory prediction accuracy.
        The findings of our study emphasize that the selection of an appropriate subsample dimension for training data, identified as a Representative Elementary Volume (REV) for upscaling, plays a pivotal role in optimizing the trade-off between computational time and prediction accuracy.

        Speaker: Yaotian Guo
      • 59
        Large Scale Efficient 3D Domain Transfer for Digital Images of Porous Materials using Pseudo-3D Architectures

        Characterisation of the internal 3-dimensional (3D) structure of complex porous materials has been revolutionised with deep-learned image processing and segmentation, promising second-scale scan times with hour-scale quality, and beyond-human multi-label segmentation accuracy at a fraction of the time. However, these claims are currently only true for single-sample, single-domain cases using 2D networks on 3D data, or small 3D subdomains (<$10^{8}$ voxels) on 3D networks. These limitations are fundamental to domain mismatch between trained networks and inference inputs, dimensional blindness of 2D networks on 3D data causing z-axis misalignment (the coin-stack (CS) effect), and the incompatibility between memory inefficient 3D networks and large-scale 3D data. These interconnected issues that have prevented the true application of deep learning to 3D volume data ($10^{11}$ voxels, typical of synchrotron and nano/micro-CT imaging) are resolved in this paper. Herein, we introduce an unpaired semantically consistent pseudo-3D approach to domain transfer capable of inference on domains approaching the tera-scale. Several important domain transfer applications are exhibited and validated using pixel metrics and physical parameters, including the enhancement of the time resolution from hour-scale to minute- and second-scale of static and dynamic scans of geological rocks while maintaining the hour-scale image quality, accurate segmentation of out-of-domain nano/micro-CT images using a pretrained segmentation models of lithium-ion batteries and hydrogen fuel cells, and efficient large-scale 3D inference ($10^{11}$ voxels) on single GPU.

        Speaker: Kunning Tang (UNSW)
    • MS04: 1.2
      • 60
        Unified surface poromechanics theory capturing condensation-induced contraction of mesoporous materials

        The phenomenon of adsorption-induced deformation is prevalent in both natural materials such as wood and coal, as well as in engineered materials like cement, MOFs (Metal-Organic Frameworks), and porous polymers. As the partial pressure of adsorbate vapor rises, the strain isotherm of these materials can display intricate nonlinear and non-monotonic behaviors.

        Under low partial pressures, most porous materials undergo volumetric expansion. This expansion can be attributed to the reduction of surface stress and the subsequent relaxation of adsorption stress experienced by the solid skeleton—a phenomenon commonly known as the "Bangham effect." This effect is well-described by the surface poromechanics formulation proposed by Zhang (2018). For microporous materials, early adsorption can lead to a subtle shrinkage before the onset of swelling. This is linked to the development of negative disjoining pressures in nanopores, as explained by Eskandari-Ghadi and Zhang (2021).

        Despite these advances, the current surface poromechanics formulation is only for a single-phase pore fluid and therefore, does not apply to partially saturated porous media nor capture the dynamics of phase transition of the pore fluids. For this reason, it is unable to model the sudden contraction of mesoporous media at intermediate vapor pressure levels induced by capillary condensation. This contribution outlines our progress toward developing a unified surface poromechanics formulation that meets the following criteria:
        1. It takes into account the phase transition of pore fluid from vapor to liquid and the emergence of the liquid-vapor interface.
        2. It accurately reproduces the water retention characteristic curve unique to each porous system.
        3. It captures both the early Bangham expansion (without condensation) and the significant contraction resulting from condensation in a consistent manner.
        4. The theory's asymptotes at degrees of saturation equal to 0 and 1 align with the conventional poromechanics theory for single-phase pore fluid.

        REFERENCES:
        Eskandari-Ghadi, M., Zhang, Y., 2021. Mechanics of shrinkage-swelling transition of microporous materials at the initial stage of adsorption. International Journal of Solids and Structures 222, 111041.
        Zhang, Y., 2018. Mechanics of adsorption-deformation coupling in porous media. Journal of the Mechanics and Physics of Solids 114, 31-54.

        Speaker: Yida Zhang (University of Colorado Boulder)
      • 61
        Role of Substrate Roughness in Soil Desiccation Cracking

        Soil desiccation crack is ubiquitous in nature, yet the physics of its initiation and propagation remain under debate, as it involves complex interactions across multiple fields of mechanics, hydraulics, and thermals. Here, an experimental attempt is made to uncover the role of substrate roughness on the soil desiccation process. The substrate roughness is deliberately fabricated by 3D printing, whereas the thickness of sample and environmental humidity are controlled to eliminate the effect of large hydraulic gradient. Four types of soils with varying expansibilities were desiccated on substrates with varying roughness. It reveals that: (1) soil desiccation crack evolution can be conceived as a competing process between the shear failure of soil-substrate interface, i.e., slippage of interface, and the tensile failure of soil, i.e., crack initiation, in minimizing the total energy of drying soil; (2) substrate roughness alters the failure mode and shear strength of soil-substrate interface and its sensitivity to moisture, thereby it regulates the pattern of how soil crack propagates upon drying; (3) soil expansibility is recognized as a key factor governing the crack-initiation point in addition to the widely recognized air-entry, and flaws in soil are the sources for the 120° crack angle and bimodal crack angle distribution.

        Speaker: Mr Yuhan Yang (Hunan University)
      • 62
        Molecular Dynamics Simulations of Porous Illite Clay Surfaces and Particles

        Illite clay constitutes the main component of the Norwegian quick clay that is known for its tendency to transform rapidly from a solid to a liquid state under certain pressure.1 Currently the practical method for stabilizing quick clay involves the use of cement and lime, resulting in significant CO2 emissions.2 To explore more environmentally friendly stabilizers, nanoscale theoretical understanding of the mechanical forces between illite particles is essential. The interaction between clay particles depends on the thickness of electrical double layer (EDL) that can be controlled by the types and concentrations of salts.3 In quick clay, a higher salt content results in a thinner EDL and reduced repulsive force. Conversely, if the salt is leached out due to underground water, the EDL thickens, leading to a stronger repulsive force and worse stability. This research primarily focuses on non-equilibrium molecular dynamics simulations involving the direct contact between an illite particle and surface. It aims to elucidate the connections between this interaction and the variations in the EDL resulting from the addition of different salts, such as NaCl, KCl, CsCl, and CaCl2.

        Speaker: Ge Li (NTNU and PoreLab)
      • 63
        Dynamic soil structure imaging experiments and their digital model representation

        It is now well recognized that soil structure is dynamic and changes due to numerous reasons, most notably due to saturation changes. In or study, we have sampled 15 soil samples and performed a detailed X-ray microtomography (XCT) imaging study of the full wetting-drying curve [1]. By analyzing the XCT images, we revealed the dynamics of soil pore structure under slow water changes. In total, our analysis is based on 135 3D tomography scans (9 soil moisture points for each sample). We were able not only to visualize structural dynamics (which showed significant changes within the soil at ~10 µm – 3 mm pore sizes range) but also computed major classical morphological metrics. The analysis of these parameters and conceptual model of structural behavior revealed that after the wetting-drying cycle the studied soil degraded in general. This is contrary to the prevailing previous findings for mainly compacted soils where wetting-drying cycles led to structural improvements. We also found that classical metrics are not able to describe structural changes due to their low information content [2].
        Now, we need something reliable to describe all structural changes and create model to describe the changes we observe din the experiments. Such a descriptor to create a digital structural model has to fully describe both geometry and topology of the soil sample and possess high information content. We shall argue that a set of directional correlation functions [3] is enough for this purpose. Compared to classical metrics they 1) contain all classical metrics within them and can also be extended to include topological measures such as persistent diagrams [4]; 2) allow to describe anisotropic structures [5]; 3), have measurable information content [2]; 4) if needed, the information content can be augmented with higher number of functions in the set and with higher-order functions [6,7]; 5) allow to establish stationarity and representativity of the structure itself [8,9]; 6) can collect dynamic structural information from different scales [10] – the very aim of the research in this area.
        In this presentation we shall focus on:
        - Experimental studies of soil structure dynamics by imaging with XCT;
        - Description of dynamic soil structure without the use of classical approaches in the form of scalar characteristics;
        - Creation of digital model of soil structure dynamics with high information content vector descriptors in the form of correlation functions.
        In addition to experimental results and their interpretation, we discuss the major implications of our findings and outline a possibility to deepen our understanding of soil structure-function relationships, including dynamic hydraulic soil properties and 3D soil structure digital model based on correlation functions.

        Speaker: Dr Marina Karsanina (Schmidt Institute of Physics of the Earth of Russian Academy of Sciences, Moscow, Russia)
      • 64
        Impact of dynamic pore structure on local macroscopic parameters

        In continuum models for drying, macroscopic parameters are integral, relying on the microstructure of porous media. These parameters are determined through the volume averaging of state variables, often derived from simulations using fixed pore network models (PNM). While fixed PNM is a prevalent computational method, it typically assumes a static microstructure throughout the drying process, neglecting the dynamic changes inherent in drying deformable porous media. These changes significantly alter both the structural and transport properties of the porous medium. This study introduces an adaptive pore network model developed to capture the dynamic microstructure and mass transport kinetics of a model capillary porous medium under slow drying conditions. Through adaptive PNM simulations, key parameters, including local relative humidity and moisture transport coefficient, are derived. The findings reveal that, particularly in the later stages of drying, the non-local equilibrium effect becomes pronounced, evidenced by the deviation of local vapor pressure from equilibrium vapor pressure. Moreover, the moisture transport coefficient is primarily influenced by the liquid phase, leading to an extended transport region where the process dynamics are mainly governed by the presence and movement of liquid. This adaptive approach provides a detailed view of how microstructural changes interact with local macroscopic parameters during the drying of deformable porous media.
        Acknowledgement

        Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Project-ID 422037413 – TRR 287.

        Speaker: Jing Chen
      • 65
        Predictive modelling of liquid ingress into disintegrating pharmaceutical tablets

        In pharmaceutical science, the disintegration process refers to the mechanical breakup of an intact tablet into small fragments to increase the surface area of the drug substance in contact with the dissolution medium [1]. Within the pharmaceutical industry, the disintegration time, which is the time required to disintegrate a tablet until no palpable residues remain, has been employed as one of the critical quality attributes to ensure the bioavailability and efficacy of end products. Despite its significance and extensive studies on the process, a universally accepted and practically employed model to quantitatively describe the disintegration process remains elusive due to the intricate physiochemical interplay of the process, coupled with complex formulations and manufacturing conditions.

        At the microscale, determining the propagation of a capillary driven flow requires a solid understanding of the capillary structure including its deformation over time. However, it is practically impossible to capture the complexity in sufficient detail using established analytical methods for pharmaceutical tablets. We instead studied the propagation of the liquid front within the porous matrix, recognising the role of the advancing liquid as the initiator of all subsequent phenomena. Employing the terahertz pulsed imaging (TPI) technique coupled to an open immersion setup allowed us to precisely monitor the in-situ location of the liquid front whilst controlling the influence from the experimental setup on the liquid flow [2]. This approach successfully captured the liquid ingress profiles of complex formulation tablets and revealed two regimes of liquid propagation: 1) an initial rapid uptake regime and 2) a subsequent slower linear regime, which was rate-limiting in determining the disintegration time.

        Our results suggest that the linearity in transport results from the synchronised propagation of the liquid front and the erosion at the interface in touch with the dissolution medium [3]. Consequently, we employed this mechanism to develop a predictive model for the disintegration process of pharmaceutical tablets where each regime was modelled as a time-evolving porous medium in terms of swelling and erosion and its terminal structure transitioning inward. Based on TPI measurements, the associated fitting parameters were extracted to quantify the mass transport behaviour of the two regimes and compared across different tablets. This methodology and modelling offers insights into the disintegration process of pharmaceutical tablets and shows potential applicability in understanding disintegration for a range of swelling and eroding porous media.

        Speaker: Jongmin Lee (University of Cambridge)
    • MS05: 1.2
      • 66
        Co-transport of engineered nanoparticles and bacteria in soil

        Increasing production and wide application of engineered nanoparticles lead to their ultimate release into the environment, thereby contaminating water, air, and soil. Suspended bacteria are ubiquitous in the subsurface and can affect the transport of nanoparticles. This study investigates the fate and transport of zinc oxide nanoparticles (nZnO) in saturated soil in the presence of suspended E. coli through a coupled experimental and modelling approach. E. coli was found to enhance the transport of nZnO. However, E. coli transport was retarded in the presence of nZnO. The difference in the transport behaviour of nZnO and E. coli during the cotransport and individual transport studies is because of the competition between them in finding attachment sites on grain surfaces and also due to the formation of nZnO-E. coli heteroaggregates. The experimental results were successfully simulated using a model which accounted for nZnO and E. coli retention in soil, heteroaggregation kinetics, and heteroaggregate retention in soil.

        Keywords: ZnO nanoparticles, bacteria, cotransport, modeling, soil

        Speaker: Rima Manik (Indian Institute of Technology, Hyderabad)
      • 67
        Pore-Scale Modeling MICP Process and Investigation of the Effect of Pore Structures on Calcite Distribution

        Abstract: The microbially induced calcite precipitation (MICP) technique holds promising applications in groundwater remediation, gas storage, soil improvement, and rock fracture sealing. In this study, a two-dimensional pore-scale numerical model is developed to simulate the coupled flow, reactive mass transport processes, and precipitation processes in MICP. In the presented model, the lattice Boltzmann method (LBM) and finite element method (FEM) are employed to solve the incompressible Navier-Stokes equations and the advection-diffusion-reaction (ADR) equation, respectively. The presented model considers the processes of bacterial transport and attachment, ureolysis, and the bacterial and calcite detachment induced by the shear effects of the flow. With the presented model, multiple field profiles including the flow field, concentration field, and the calcite distribution can be obtained in the pore space. The presented numerical model is validated based on the experimental data from the literature. To investigate the effect of heterogeneous pore structures on calcite distribution, different scenarios are carried out. The simulation results demonstrate that the pore structures with large pore throats result in more calcite accumulation. For heterogeneous pore structures with upper or lower distribution, the calcite distribution is dominated by the flow direction. Moreover, the quantitative results are presented in the evolution curves of calcite proportion, and the trend of evolution curves in different areas can reflect the homogeneity of the calcite distribution. The distribution of the evolution curves in various areas is aggregated, indicating a uniform calcite distribution.

        Speaker: Dianlei Feng
      • 68
        Bioclogging during underground hydrogen storage: Assessing impact of biofilm formation on hydrogen injection and recovery.

        In response to the urgent global concerns regarding climate change, there is a critical need for the evaluation and implementation of reliable renewable energy solutions. To bridge the energy demand-supply gap, immediate research into effective energy storage methods is imperative. The use of green hydrogen, generated via renewable electricity's electrolysis, is receiving increased attention, owing to its low volumetric calorific value (3 kWh/m3) and high mass energy density (33.3 kWh/kg). Hydrogen gas as an energy carrier can be stored in large amounts in subsurface reservoirs, such as salt caverns, saline aquifers, and depleted hydrocarbon reservoirs. However, even at minimal aqueous concentrations, hydrogen serves as an attractive electron donor for subsurface microorganisms, including methanogens, sulphate-reducers, homoacetogenic bacteria and iron (III)-reducers. Microbial growth in porous media leads to biofilm formation, narrowing the rock pores and causing potential bioclogging. Therefore, assessing these microbial effects in underground hydrogen storage is crucial to estimate risks associated with gas injectivity, loss, and recovery in large-scale operations.
        This study investigated hydrogen consumption by two sulphate-reducing microorganisms in a pressurized microfluidic chip at 10-100 bar and 37 ℃, mimicking conditions akin to shallow gas reservoirs. The microbial cells congregated at the interface of the aqueous phase and hydrogen gas, utilizing the hydrogen to form biofilms. However, bioclogging resulting from biofilm formation was observed when utilizing Lactate as the carbon source, while no such clogging was observed with hydrogen gas. Under the microscope, the comparison between biofilm formations using Lactate and hydrogen gas revealed notable differences. The Lactate-formed biofilm appeared denser and tightly packed, whereas the hydrogen gas-formed biofilm displayed a crystal-like structure. Further analysis with Raman spectroscopy uncovered disparities in the protein structures within these biofilms.After one week of cultivation in a hydrogen-rich environment, the biofilm appeared to detach from the pore network following the second hydrogen injection drainage. Our hypothesis proposes that this reduction in the biofilm might be due to a shift in bacterial behavior, potentially transitioning from a biofilm-mode to a planktonic-mode state in an environment abundant in hydrogen. Biofilm formation and its adhesion to solid surfaces directly impact wettability, notably increasing hydrophobicity. This change in contact angles leads to a decrease in capillary entry pressure during hydrogen injection, thereby improving drainage efficiency.
        These findings emphasize the substantial impact of biofilm dynamics and wettability changes on the efficiency of hydrogen injection and recovery processes in subsurface reservoirs. The results offer critical experimental evidence concerning the microbial risks linked to underground hydrogen storage, thereby contributing to the validation of the upscaled model within this context.

        Speaker: Na LIU (University of Bergen)
      • 69
        Impacts of viscous fingering on bio-methanation risks during underground hydrogen storage

        Hydrogen (H2) can be used as an effective energy vector and is a key element in the energy transition [1]. To accommodate the significant volume of H2 involved in the future energy mix, subsurface porous media, such as saline aquifers and depleted hydrocarbon reservoirs, is increasingly promoted as a viable option for underground H2 storage [2]. However, the reliability of this form of storage is not yet proven. One of the concerns is the impacts of microbial activity on the storage performance of H2 [3]. This is because H2 is a superb electron donor and can trigger a variety of microbial metabolism [4]. For example, H2 may initiate the bio-methanation process when carbon dioxide (CO2) is used as cushion gas in the subsurface environment. This process may lead to H2 loss and the contamination of the back produced gas. On the other hand, H2 has very low viscosity and thus is susceptible to the development of viscous fingering, when being injected to displace a more viscous fluid [5]. In this numerical work, we investigated the joint impacts of bio-methanation and viscous instability on the storage and recovery performance of H2. We have performed a range of 2D vertical cross-sectional models with a very fine cell size (0.1 m) to capture the viscous fingering in detail. It has been found that the viscous instability can expand the total size of the mixing zone and thus promote H2 consumption by methanogenesis. Since the process leads to the reduction in total gas volume, the primary purpose of cushion gas injection, which is to prevent water breakthrough, can be compromised. As a comparison, a gravity-dominated operational strategy is designed to isolate and thus ascertain the role of viscous instability on the bio-methanation process. Although gravity can drive the segregation between H2 and CO2, permeability heterogeneities lead to flow dispersions and gas mixing. However, the total mixing zone is much reduced and thus the methanogenesis is suppressed. The results of this work can be used to improve the numerical simulations associated with H2 storage in subsurface porous media, including both hydrodynamic and microbiological processes. This study should also provide useful insights and definitions of “target properties” (e.g. acceptable rate of methanogenesis) for experimentalists and industry engineers involved in screening projects for subsurface H2 storage.

        Speaker: Gang Wang (Heriot-Watt University)
      • 70
        Field-scale mathematical modelling and simulations of biofilm effects in hydrogen storage

        Underground hydrogen storage (UHS) allows for large-scale energy retention using depleted hydrocarbon reservoirs, saline aquifers, and salt caverns. Biofilms, defined as an aggregate of microbes enclosed in a matrix of extracellular polymeric substance (EPS), are sophisticated systems where different biological, chemical, and physical processes occurs such as growth, erosion, sloughing, attachment, formation of metabolites, etc. While in some applications we can benefit from biofilms (e.g., food industry, water quality), biofilms can also become an obstacle, especially for UHS regarding injectivity and hydrogen loss. Numerical simulations can assist on a better understanding of the interactions between biofilms and hydrogen in cyclic operations involving injection, storage, and withdrawal periods.

        The aim of this work is to develop and implement a mathematical model to perform field-scale UHS simulations including biofilm processes. We can find in literature comprehensive multi-component bio-reactive models for UHS (e.g., Hagemann et al. (2016)). Since field-scale simulations require running the model on large spatial and temporal scales, then simplified models are suitable to deal with the heavy computational burden. Still, the simplified model must capture the key processes and quantities. Here, the main mechanisms related to microbial activity are the consumption of hydrogen by the biofilm, porosity reduction due to the development of the biofilm, and biofilm detachment because of higher flow velocities. To this end, the fluid is modelled as a two-phase (liquid and gas), two-component (water and hydrogen) system, while the biofilm is modelled as a solid phase attached to the rock, which grows due to hydrogen consumption and suffers erosion due to the flow.

        The mathematical model is implemented in the industry-standard simulator Open Porous Media (OPM) Flow (Rasmussen et al., 2019). The existing hydrogen module implementation is extended to include biofilms, which allows for flexibility to account or neglect the biofilm effects on the simulations. For example, in Strobel et al. (2019) the authors presented a history matching study where microbial activity was identified in the field during hydrogen injection, leading to a successful match after adjusting the initial biofilm density.

        We apply the model to assess the hydrogen loss under different injection strategies and microbial parameters. The complexity in the geological models is increased from core samples to layered heterogeneous field-scale reservoirs. We use the pyopmnearwell software (Landa-Marbán and von Schultzendorff, 2023), an open-source framework for creating the required input files by OPM (e.g., corner-point grids, tables for the saturation functions, injection schedules) via configuration files, to perform the simulations, which allows for reproducibility of the results and further studies (e.g., history matching, optimization).

        Speaker: Dr David Landa-Marbán (NORCE Energy)
    • MS06-B: 1.2
      • 71
        A gel front liquid system with delayed properties for pore-type cracks

        With the deepening of oil and gas resources exploration and development, drilling wells will encounter various problems, especially the leakage problem, which will lead to a substantial increase in drilling costs. At the same time, the uncertainty of the leakage channel in the process of reservoir leakage not only has a great impact on the drilling efficiency but also causes a large amount of drilling fluid leakage and other problems. At this stage, cement slurry plugging material is the most commonly used plugging material to deal with cracks, holes, and other malignant leakage, but the disadvantage of cement slurry is easy to be dilute with formation water mixing, so it is difficult to stay in the leakage layer in the near-well zone to the point that it can't form a dense blocking layer, so the effect of the cement plugging agent isn't particularly ideal. This paper uses anionic polyacrylamide HPAM and organic chromium crosslinker, a stabilizer to form a temperature and PH-sensitive gel system, the gel system can be added to the pre-positioning liquid to delay the gel formation time, and the system before and after the addition of the pre-positioning liquid, respectively, the performance of the analysis and evaluation, and ultimately determine the gel system. It was found that within a certain range, the release rate of Cr3+ from the organochromium crosslinker in this system accelerated with the increase in temperature, and the release rate of Cr3+ was slower when the pH was 5-7. Since the release rate of Cr3+ determines the speed of hydroxyl bridge reaction with HPAM, the purpose of delayed gel formation in a weak acidic environment can be achieved by controlling these two variables, so that the predisposed liquid system added into the gel can be controlled and sealed after arriving at the target layer, obtaining good plugging effect and laying the foundation for the future cement cementing process.

        Speaker: Doudou Wang (中国石油大学华东)
      • 72
        Experimental Investigation and Molecular Dynamics Analysis of the Fluid-Fluid Interactions between Binary Surfactant Systems for EOR

        This work aims to explore the properties and interactions between binary surfactant systems due to their ability to form mixed micelles with lower interfacial tension (IFT). The focus is on determining the synergistic or antagonistic behaviors of these systems for effective application in enhanced oil recovery (EOR) in carbonate oil fields. Our study employed a methodology comprising experimental analysis, mathematical modeling, and molecular dynamics simulations. In the experimental study, we examined eight individual surfactants and six binary surfactant systems at various ratios to determine their critical micelle concentrations (CMCs), using reservoir oil and performing experiments at reservoir conditions. Then, Rubingh's Regular Solution Theory (RST) was applied to evaluate interactions within the binary surfactant mixtures. Finally, using molecular dynamics simulations, we characterized the microscopic interactions to comprehend how hydrophilic and hydrophobic parts of the surfactants interact with surrounding media, and how they self-assemble into aggregates such as micelles or bilayers. The key findings of our work showed that the occurrence of synergism or antagonism in lowering the CMC of binary surfactant mixtures depend on both the concentration of the individual surfactant and the type of surfactant used. Nevertheless, we noted a prevalent synergistic phenomenon in all binary surfactant systems, notably influenced by the concentration of the non-ionic surfactant. Increased concentrations of non-ionic surfactants notably enhanced synergistic interactions, fostering lowered CMC values when combined with anionic, cationic, and zwitterionic surfactants. On the other hand, an excessive concentration of cationic surfactants demonstrated relatively 'weak' synergistic effects, attributed to their comparatively smaller hydrophobic tail. Moreover, the formation of mixed micelles in binary surfactant systems led to a more negative free energy of micellization, thereby achieving synergistic effects between surfactants and resulting in lower CMC values. This emphasizes the crucial role of surfactant concentration in achieving synergistic outcomes within mixed systems. Generally, binary surfactant systems demonstrated lower CMC values compared to single surfactants, suggesting the potential for their use at lower concentrations to achieve desired interfacial and recovery outcomes, thereby reducing operational costs.

        Speaker: Ayomikun Bello (Skolkovo Institute of Science and Technoloy)
      • 73
        Wettability acoustic probing in granular porous media

        Surface wettability determines the property of a solid surface in contact with a fluid. It plays a major role in reservoir engineering applications involving fluid transport phenomena. For example, it is critical to hydrocarbon recovery where wettability not only influences the sweep efficiency, but also influences the remaining oil distribution. Since the wettability tends to affect the mechanical properties of rocks, it provides a potential means to remotely monitor the movement of spilled nonaqueous phase liquids in the subsurface via acoustic monitoring. However, there is a lack of direct wettability measurement conditions in underground applications. As widely utilized methods in geophysical exploration, borehole sonic logging and seismic surveys could be used as a potential far-field probing tool for wettability changes. Understanding the acoustic response to wettability changes becomes crucial for this purpose.
        To generate a favorable experimental condition for wettability acoustic monitoring, we establish an ultrasonic experimental basis for observing changes between hydrophilic and weakly hydrophilic conditions in (partially) saturated granular porous media. Glass beads are employed to construct granular porous media with fully interconnected pores and are chemical-treated to alternate the wettability condition. Particular attention is paid to ensure the uniform distribution of water across bead packings at different saturation levels, aiming to diminish the impact of patchy water distribution.
        The P- and S-wave velocities and attenuations are measured with increasing saturation for the bead packings before and after chemical treatment, respectively. The results illustrate that the chemical treatment increases contact angles and improves the water–bead coupling, leading to higher velocity and lower attenuation of coherent waves. The Gassmann-Wood-Walton model predicts the behavior of coherent waves under different wettability conditions assuming a change in the coordination number. Once reaching a critical saturation, incoherent high-frequency waves are developed with higher propagating velocities. The treatment reduces the amplitude of these incoherent waves to some degree, probably due to the improvement of water–bead coupling. We interpret the observed incoherent waves in terms of the wettability-dependent ability for water to bridge neighboring beads. The velocity of incoherent waves is exceeding the Gassmann-Wood prediction at partial saturation but is close to the fully water-saturated condition. This is suggestive of a propagation path of incoherent pulses resembling the fully water-saturated condition. In our interpretation, this is because the presence of a critical amount of water creates a favorable and wettability-dependent condition to form liquid bridges connecting neighboring grains.

        Speaker: Dr Yangpu Chen (China University of Petroleum (East China))
      • 74
        Contact angle on rough curved surfaces and its implications in porous media

        Equilibrium contact angle depends on both the chemistry of the two fluids and solid base, and the microstructure on the solid surface. Actual surface of the pore wall in porous media is typically rough and curved, which has not been well-considered in related applications. This work uses a free interfacial energy minimization approach to theoretically derive the equilibrium contact angle on two specific surface structures on flat surfaces and extends the derivation considering the surface curvatures in porous media. Results reveal the equilibrium contact angle is not dependent on the curvature of spherical surfaces, and we further prove that this conclusion applies to any point along the apparent common line at solid surfaces with any arbitrary curvature. The fundamental physics is the local mechanical balance of a composite contact among three interfacial tensions. Furthermore, the contacting mode can shift from non-wetting to wetting when the pressure difference between two fluids exceeds the entry pressure of the microstructures, which should be considered in relative dynamic scenarios such as rain droplet impact and fluid displacement in porous media. Note these conclusions are from pure theoretical analysis based on idealistic assumptions and real circumstance may deviate from these assumptions.

        Speaker: Lei Liu (Westlake University)
      • 75
        Enhanced CO2 Storage in Saline Aquifer by Electric Field Considering Formation Wettability

        Geological CO2 storage involves injecting captured CO2 into various geological formations, in which saline aquifers have the largest storage potential around the world. In the context of carbon neutrality, one of the key issues is to store CO2 as much as possible on the premise of formation stability. In this paper, we find that external electric field can enhance CO2 storage in saline aquifers. The different mechanisms of CO2 storage enhancement in hydrophobic and hydrophilic formations are revealed by molecular dynamics simulations. The following conclusions can be drawn. (1) In order to consider formation wettabilities, a carbon-based pore wall, a hydroxylated quartz pore wall and a calcite wall are constructed. CO2 tends to accumulate more readily on carbon-based wall, while H2O exhibits a higher tendency to accumulate near hydroxylated quartz wall and carbonate wall, indicating different wettabilities of the three walls. (2) On a short time-scale, CO2 is stored in adsorbed and dissolved states in saline aquifers of hydrophobic and hydrophilic formations. (3) In the absence of an electric field, the dissolved CO2 accounts for 42.50% in the hydrophobic formations; the adsorbed CO2 accounts for 12.4% in the hydrophilic quartz formations. When an external electric field is applied vertically to the wall, the proportion of dissolved CO2 in the hydrophobic formations increases to 63.23%; the proportion of adsorbed CO2 in the hydrophilic quartz formations increases to 21.76%. However, the external electric field has negligible effects in the hydrophilic carbonate formations. (4) The orientation of H2O molecules and the hydrogen bonds are further analyzed to reveal the different enhancement mechanism. In the hydrophobic formations, the external electric field induces oriented H2O molecules, leading to their preferential accumulation near the wall, rendering the initially hydrophobic wall hydrophilic, thereby reducing the available space for adsorbed CO2 and promoting its dissolution in the H2O phase. In the hydrophilic quartz formations, the external electric field drives H2O away from the surface, concurrently reducing the number of hydrogen bonds formed between H2O and the hydrophilic wall by approximately 24.22%. This reduction diminishes the hydrophilicity of the wall. In the hydrophilic carbonate formations, there are no hydrogen bonds between H2O molecules and the wall. Therefore, the electric field has negligible effects on the wettability of the wall and CO2 storage. This study proposes a novel technique to enhance CO2 storage in saline aquifers of different wettabilities by electric field. The molecular perspective revealing the enhancement mechanism is expected to provide theoretical guidance in the future practical application.
        Keywords: Geological CO2 storage; electric field; molecular dynamics simulation;

        Speaker: Liangyu Zhao (Chengdu University of Technology)
      • 76
        Experimental Investigation of Illite Clay in Norwegian Quick Clay for Sustainable Ground Stabilization

        Quick clay is a young geological clay formation and has been formed during the last glacial ice age. Due to the isomorphic substitutions in these phyllo-silicates, their flat surfaces are net negatively charged (1). The Na+ cations which come from the main salt in the marine environment (35g/L) is attracted by the negatively charged surface and neutralizes it. As a result, the particles have a thin diffuse double layer (DDL), leading to flocculation in a “House- of-cards” structure (2,3,4). Due to the isostatic rebound above sea level, fresh water can infuse into the marine deposits, causing salts to leach out. With the loss of cations, the Coulomb forces between the clay platelets increase and overcome the attractive van der Waals forces, and the structure becomes extremely sensitive (5,4). When this structure is disturbed, the clay formation turns into a low-viscosity fluid, and a catastrophic landslide occurs.

        To stabilize quick clay, a deep mixing technology with lime and cement is used. Although this technology has been improved over the past decades, it still has a high carbon footprint.
        To provide a more sustainable stabilization, our project “Sustainable Stable Ground” investigates the quick clay formation from an atomic scale and upscale to geological scale.

        Here, our focus lies in examining the interaction between clay particles through experimental investigation of illite suspensions under different conditions, such as different ionic strength and adding various additives. For our experiments we are using the Tiller-Flotten quick clay from the area around Trondheim, Norway.
        Our process involves purifying the clay from the natural ground and characterize the physical- chemical properties using diverse methods such as Scanning Electron Microscopy (SEM) and Elemental Analysis (EDS). With the objective of identifying potentially sustainable stabilizers, we create different suspensions with diverse (sustainable) additives, allowing for a long-term observation of the clay-rich ground.

        Speaker: Rene Tammen (NTNU Porelab)
    • MS08: 1.2
      • 77
        Investigating the limits of averaging: a numerical case study employing diffusion-reaction in porous media

        Effective descriptions are often utilized to describe mass transfer phenomena in porous media, i.e. in heterogeneous catalysis, filtering or subsurface transport. Besides more than a century of research, the a priori determination of the relevant effective transport parameters has shown to be elusive and is still subject of research. A major challenge is the appropriate mathematical upscaling of the intricate influence of pore-scale phenomena on the Darcy-scale behavior for realistic morphologies. Typically, such upscaling procedures incorporate convenient estimates to determine the significance of the pore scale transport phenomena with respect to the Darcy scale. Often, such estimates are based on the pore space geometry, macroscopic transport properties and external process parameters. A popular example there is the assumption, that the pore side length scale is significantly smaller than the representative dimension of the porous medium.
        To gain insight into the applicability of commonly used averaged descriptions with respect to the coarseness of the pore space, direct numerical simulations of diffusion with first order surface reaction in a resolved pore space were employed. There, a 3D resolved model was developed, based on the finite volume approach utilizing a second order implicit immersed boundary method to accommodate the representation of the pore space. The developing transient species profile was monitored and compared with the analytically derived profiles to the complementary averaged problem.
        To investigate the limitations of the averaging approach, the numerical model was applied to a variety of model porous media with varying characteristic length scales and particle Thiele moduli. The generated insights concerning the emergence of Darcy scale behavior from pore scale phenomena will be presented and implications discussed.

        Speaker: David Rieder (TU Eindhoven)
      • 78
        Reactive Solute Transport in Heterogeneous Porous Media

        CO2 geological storage in saline aquifer is significantly influenced by the reactive solute transport in the fracture media. However, the governing factors and coupling mechanisms of solute transport within the fracture at different periods under seepage-chemical coupling remain unknown. In this investigation, reactive solute transport experiment on sandstone fracture was conducted to determine the main mineral reactions. The fracture reactive solute transport analysis model taking mineral precipitation into account is proposed, and the fracture reactive solute transport characteristics under varying Pe number and Da number conditions are analyzed. The results of the dimensionless parameter analysis revealed that the Pe number was the governing factor of the solute transport process in the short period between injection and peak concentration of CO2, and that the time to peak solute concentration decreased with increasing Pe number. Under long periods of injection conditions, the solute transport process is governed by the Da number. Mineral precipitation at a high Da number obstructs the fracture, resulting in a progressive decrease in the variation rate of Ca2+ concentration during the late stage and the bypassing flow phenomenon of flow lines and species transport pathways.

        Speaker: Mengzi Ren (Northeastern University)
      • 79
        Fingering Instability During Mixing‐Driven Precipitation Flow: Experiments and Simulations

        Reactive flows in porous media that results in precipitation of solids are ubiquitous in a wide range of applications. Laboratory studies focusing on microscopic changes of the porous media have elucidated the complexity of the precipitation patterns due to the highly nonlinear coupling between advection, diffusion, reaction, and the intrinsic heterogeneity of the pore geometry and mineralogy. Here, we study the displacement of aqueous solutions of calcium chloride by sodium carbonate in a Hele–Shaw cell where the two fluids react, upon mixing, to form calcium carbonate precipitates. We examine the case of equal reactant concentrations in detail via high-resolution imaging, which reveals a variety of precipitation patterns at different injection rates and reactant concentrations. We find that reaction along the moving fluid–fluid interface forms a precipitation band in the form of particle suspensions, whose width and particle concentration are controlled by the injection rate. This injection rate dependent behavior arises due to particle–particle agglomeration in the precipitation band. Higher injection rates generate larger particles and lower suspension mobility, resulting in miscible viscous fingering at the precipitation band. Critically, fingering has important control over the growth of precipitation amount in time, which is diffusive with time in the absence of fingering but is linear with time in the presence of fingering. Furthermore, we show that the precipitates uniformly deposit onto the top and bottom surfaces of the Hele–Shaw cell as a thin particle layer at low injection rate, but they form large particle islands at high injection rates. We develop a novel reaction-diffusion-convection model that not only captures the phenomenology of the precipitation and deposition process, but also the scaling of the temporal precipitation amount.

        Speaker: Dr Benzhong Zhao (McMaster University)
      • 80
        Pore-scale digital twin of sorption thermal energy storage in packed bed reactor using a machine-learning assisted dual-network model

        The adsorption thermal energy storage system is widely utilized for low-grade heat storage and recovery due to its environmentally friendly and efficient characteristics. In this work, we utilized a machine-learning assisted dual-network model to construct an upscaling model from micro-kinetics to reactor in order to simulate an adsorption heat release process involving heat and mass transport on a meter-scale packed bed reactor. The simulation results were compared with experimental measurements and analytical models to demonstrate the accuracy of the model in predicting temperature and concentration distribution within the system. Subsequently, we explored the impact of different boundary conditions on the internal state parameters during the adsorption heat storage process, offering valuable insights for the design of adsorption heat storage systems.

        Speaker: Mingliang Qu (Zhejiang University)
      • 81
        Quantification of crystal surface reactivity using positron emission tomography (PET) techniques

        Predicting the migration behavior of dissolved contaminants in the pore space of rock and soil is crucial for assessing the feasibility of remediation and long term waste storage strategies.
        Positron emission tomography (PET) using conservative radiotracers is an established and reliable method for investigating advective flow and diffusive flux in porous geomaterials and for validating transport models [1, 2]. However, solute transport is often significantly influenced by sorption effects. Reliable data concerning these effects are crucial for analyzing remediation processes as well as predicting desired immobilization in waste storage applications.
        To understand and quantify the effects of solute-mineral surface interactions, analyses beyond breakthrough curve measurements are essential. PET techniques offer unique capabilities by providing in-situ tracer propagation and concentration data with high temporal and spatial resolution, surpassing traditional flow and lysimeter experiments.
        For many materials, it is desirable to quantify both reactivity and hydrodynamic flow. The simultaneous quantification of both effects requires the use of a dual tracer system. In this presentation, we discuss the possibilities of utilizing a tracer pair consisting of $^{18}$F as a reactive tracer and $^{76}$Br as its conservative counterpart. This allows the prediction of spatially resolved surface reactivities as well as the evaluation of advective flow. Using different sandy sediments as model systems, we demonstrate the quantifiability of localized sorption effects as low as 10 pmol/mm³.

        Speaker: Jann Schöngart (HZDR, FS Leipzig)
    • MS09: 1.2
      • 82
        Volume of Fluid based study of the three phase dynamic contact line on rough surfaces relevant for Underground Hydrogen Storage

        Large-scale (TWh) renewable energy storage is crucial to achieve a net-zero green world. To accomplish this, renewable energy can be converted into hydrogen (H$_2$) and stored in large-scale volumes in giant subsurface geological reservoirs. The feasibility of underground hydrogen storage in porous reservoirs highly depends on the flow and transport behaviour of hydrogen during subsequent injection and withdrawal cycles in the reservoir, which is governed by complex pore-scale processes [1-3].

        Recently, several experimental studies [4-8] have taken place, which allow for the characterization of hydrogen transport properties in the subsurface. However, some discrepancies were found in contact angle characterization results using different measurement techniques and solid surfaces. The roughness of the solid surface is a possible explanation for this [9]. To date, no study has investigated the impact of roughness on the characterization of H$_2$-brine flow.

        To help shed new light on the characterization of this crucial property, the Basilisk flow solver is used to conduct numerical simulations by solving the 2D two-phase Navier-Stokes equation. The H$_2$-brine interface is captured using the Volume-of-Fluid method, and the Continuous Surface Force method is employed to compute surface tension forces. The calculation of the curvature is done using height functions. To investigate the influence of surface roughness on H$_2$-brine flow in sandstone channels, a hybrid Volume-of-fluid coupled embedded boundary solver is used. Within this solver, an intrinsic contact angle can be imposed on solids with diverse shapes, facilitating the replication of a rough sandstone surface. Dynamic contact line motion and apparent contact angles can be analysed.

        By comprehending the influence of surface roughness on the contact line motion, we will gain insight into the reported experimental measurements and assess the appropriateness of using specific data as input for larger-scale simulations.

        Speaker: Willemijn van Rooijen (TU Delft)
      • 83
        Simulation of boundaries and parameters variations of natural gas hydrate in thermofluidic dissolution based on multi-field coupling under pore-scale modeling

        When natural gas hydrates are heated and dissolved, the boundaries of fluid-solid will apparently be changed, and average permeability and equivalent thermal conductivity change in coupling. We designed two different microstructures, grain‐coating type and pore-filling type, based on two common storage modes. The model size is 50 μm × 22 μm, and the solid particles are 2.0 μm × 2.0 μm. Then we use the Dual Distribution Function model (DDF) of Lattice Boltzmann method to simulate the processes of hydrate dissolution and heat convection based on coupling thermal-flowing-mechanic-chemical (TFMC).
        To reduce the impact of nonlinear conditions on calculation results, it is necessary to partition the simulation area along the flowing direction and calculate the data of each partition. The grain-coating type has an initial hydrate saturation of 43.6% and its permeability, as determined by Darcy's law, increases from 0.43D to 2.91D. Similarly, the pore-filling type model has an initial hydrate saturation of 36.6% and its permeability increases from 0.43D to 2.64D. The relationship between permeability K and hydrate saturation Sh is linear, as the simulation area was divided into four equal parts. However, the relationship between thermal conductivity (λ) and hydrate saturation (Sh) in both models is non-linear, which calculated by the convective heat transfer formulation. As the hydrate saturation decreases, the equivalent thermal conductivity firstly increases exponentially and then linearly. The initial thermal conductivity λ of the two models is about 1.47, and the final λ of the grain-coating type is about 8.47, and the final λ of the pore-filling type is about 11.63. Both models split at around 2/3 of initial saturation. The thermal conductivity equivalent, λ, is exponential with saturation from the starting point to the cut-off point, but becomes linear when saturation is less than the cut-off point. The exponential approximation is due to the high proportion of hydrates and the gradual weakening of the thermal diffusion rate compared to the convective heat transfer rate.
        To investigate the multi-field coupling effect of TFMC, we analysed the flow rate (seepage field), the initial hydrate saturation Sh (solid field) and the activation energy ΔE or phase change potential ΔH (chemical field). Changes in velocity will not affect the linear relationship of permeability, but will significantly increase thermal conductivity and shorten the nonlinear section. The changes of initial hydrate saturation Sh will not affect the linear law of permeability, but it is necessary to increase the number of calculated partitions to make the results more linear. The demarcation point of equivalent thermal conductivity is kept on about 2/3, only the final results are changed. The activation energy change (ΔE) and phase change potential (ΔH) indicate different types of hydrates. The permeability or equivalent thermal conductivity do not change significantly.
        In summary, the DDF-LBM can be used to simulate the unsteady convective heat transfer process of hydrate dissolution. According the above analysis, more accurate parameters can be provided for thermal flow mining natural gas hydrate under seafloor, which considering the multi-field coupling conditions of TFMC.

        Speaker: Zhengyi Li (China University of Petroleum (East China); University of Science and Technology Beijing)
      • 84
        Identification and assessment of three-phase boundaries in porous electrodes of solid-oxide electrolysis cells based on a 3D microstructure model

        Hydrogen energy is regarded as a promising energy carrier due to its high energy density and non-pollution. Solid oxide electrolysis cells (SOECs) have been studied extensively as a promising way for massive hydrogen production from renewable but unstable energy sources. The electrode microstructure of SOECs has a significant influence on their electrochemical performance. To better understand the relationship between their microstructures and electrochemical performance, the quantification of key microstructural parameters such as the three-phase boundary (TPB) density and phase connectivity are required. The density and activity of the TPB sites are crucial in determining the electrochemical performance of SOEC electrodes. Therefore, it is also important to find the quantitative relationships between the active TPB density and electrochemical performance. Many efforts in microstructural analyses of SOEC electrodes via focused ion beam-scanning electron microscopy (FIB-SEM) have provided a great opportunity to link the microstructural properties to the electrode performance and the active TPB density is usually evaluated from geometrical models. However, due to the influence of conductivity and mass transfer of the gas phase, not all TPB identified from the geometric models are electrochemically active. Also, it is hard to quantitively calculate the active TPB density linking to the charge transfer. Here, we developed a comprehensive method to distinguish the active and inactive TPB which combined the image-based method and finite-element (FE)-based method. A 3D pore-scale model based on the real three phases and TPB lines was built to capture mass transfer, electron/ion transfer, and electrochemical reaction processes. The TPB sites were assessed with their contributions to the total current of the electrodes. By comparing the active TPB densities calculated from geometrical models (image-based method) and that of the developed model in this study (combined image and FE method), the proposed model was more accurate in predicting electrode electrochemical performance. This provides an effective way to reduce experimental costs and time but also deepens our understanding of the microstructure of porous SOEC electrodes.

        Speaker: Yuzhu Chen (South University of Science and Technology)
      • 85
        Digital Porous Material Analysis with Multiscale REV

        Digital Porous Media Analysis (DPMA) is the process of using imaging and simulation techniques (e.g., x-ray computed microtomography, Pore-Network Modelling (PNM) and Direct Numerical Simulations (DNS)) to investigate the properties of porous materials. With DPMA, physical processes are investigated in the real structure of a sample and effective porous media properties (e.g., permeability and capillary pressure) are estimated. These properties can then be used for upscaling, provided that the sample investigated is a Representative Elementary Volume (REV) of the full domain of interest. However, most porous materials relevant to the energy transition are multiscale, and thus have pore structures (e.g., microporosity, channels, vugs) spanning several orders of magnitude in size. An REV of such materials cannot be fully characterised by a single image, which would either be too small to be an REV or lack the resolution to accurately resolve the pore/solid interfaces. In this work, we define a multiscale REV of a porous material as an image that can be segmented into resolved and unresolved parts, and for which, for each unresolved voxel, the properties can be associated with a higher resolution image that is an REV of the underlying structure inside that voxel, the combination of which provides an REV of the full domain of interest. Our multi-scale REV workflow is demonstrated on several examples, including microporous carbonate rocks and 3D printed hierarchical foams, and employs multiscale simulation techniques (e.g., Darcy-Brinkman-Stokes models, multiscale PNM) first to confirm a multi-scale REV and then to simulate reactive transport and multiphase flow processes while estimating properties such as permeability, dispersivity and capillary pressure.

        Speaker: Julien Maes (Heriot-Watt University)
      • 86
        Digital rock reconstruction considering high stress

        Abstract
        As the development of medium and shallow oil and gas reservoirs progresses into the mid-to-late stages, the focus of petroleum exploration is shifting towards deep and ultra-deep reservoirs. Deep oil and gas reservoirs are defined as those with burial depths exceeding 4500 m, while ultra-deep reservoirs refer to those buried beyond 6000 m. These reservoirs exhibit characteristics of high stress, significantly impacting the pore structure of reservoir rocks and, consequently, influencing microscopic flow of oil and gas. Digital rocks serve as a crucial platform for simulating flow at the pore scale. However, existing methods for reconstructing digital rocks fail to account for the effects of high stress. In this study, we propose a novel method for reconstructing digital rock cores considering high stress based on the discrete element method (DEM). The first step involves transforming scanned results obtained under room temperature and pressure conditions into a DEM model. We employ the watershed algorithm to segment CT scan images, utilize spherical harmonic functions to represent particle contours, and transform them into clump particles in PFC3D. Subsequently, a DEM model is established with porosity matching that of the actual rock. The accuracy of the model is evaluated using two-point correlation and linear path correlation functions. The second step involves setting micro-mechanical for the contact constitutive model between particles, applying stress simulation calculations, and converting the results into voxel data. The third step analyzes the geometric and topological structure of pores under different stress combinations, along with the evolution of permeability. The feasibility of the proposed digital rock core reconstruction method is validated through the analysis of Bentheim sandstone as a case study.
        Keywords: digital rock reconstruction; the discrete element method; high stress; pore structure
        Acknowledgments
        This project was jointly supported by the National Natural Science Foundation of China (Grant No. 52034010), Sinopec Science And Technology Entry Program (No. P21072-1).

        Speaker: Chunqi Wang
    • MS13: 1.2
      • 87
        Pore aperture regulated surface adsorption and mass transfer of hydrocarbon and CO2 in organic nanopores

        Shale oil, widely distributed in organic (i.e., kerogen) nanopores, is playing an ever-increasing role in addressing the global energy crisis, but is faced with challenges of low recovery efficiency due to well-developed nanopores. It is believed that the pore size distribution of kerogen is in the range of several angstroms to tens of nanometers (AAPG bulletin 96 (6): 1099-1119). In such a narrow pore space, oil molecules are dominated by adsorbed phase, which is hard to recover relying on pressure drop (International Journal of Coal Geology 147 (2015): 9-24). CO2 huff-and-huff is identified as a promising method to enhance oil recovery while achieving CO2 sequestration. Clarifying the adsorption and extraction behaviors of hydrocarbons in kerogen nanopores is crucial for accurately predicting oil recovery and revealing CO2 enhanced oil recovery (CO2-EOR) mechanisms.
        In this work, we adopted molecular dynamics (MD) simulation to study the static spatial distribution and dynamic mass transfer of hydrocarbon and CO2 in slit-shaped kerogen nanopores by carefully designing a series of pore apertures. It shows that the adsorption and extraction behaviors of oil molecules are closely related to pore aperture. Interestingly, we found that the surface adsorption of oil molecules demonstrates a non-monotonic trend of rising after falling as the pore size decreases. Specifically, when the pore size is reduced to a certain value, oil molecules exhibit a pseudo-double layer adsorption state, in which the surface adsorption peaks of oil molecules are significantly weakened. On the other hand, although the reduction of pore width adversely affects the extraction speed of oil during CO2 soaking, the recovery efficiency presents a jump for the oil at pseudo-double layer adsorption state. Meanwhile, the surface adsorption of CO2 is also greatly enhanced, which leads to the highest CO2/Oil ratio in the nanopores. Collectively, our work provides fresh and important insights into hydrocarbon occurrence state and CO2-EOR mechanisms in organic-rich shale reservoirs and builds up a good foundation for accurate predictions of oil recovery.

        Speaker: mingshan zhang (东北大学)
      • 88
        Molecular Simulation of Competitive Adsorption of H2S-Containing CO2 and CH4 in Organic and Inorganic Shale Nanopores

        Objectives/Scope:
        Geological storage of acidic gases can reduce atmospheric emissions of CO2 and H2S, thus serving as a critical part of low-carbon energy systems. Depleted shale gas reservoirs are good storage candidates owing to their intrinsic gas storage capacity. However, shales exhibit complex structural characteristics and abundant micro- and nano-scale pores, where gases primarily adsorb. Consequently, comprehending gas adsorption mechanisms in shale nanopores is imperative for shale gas geological storage. In this study, we simulated CO2, H2S and CH4 adsorption in organic/inorganic shale nanopores under various pressures using grand canonical Monte Carlo (GCMC) simulations. We calculated their adsorption quantities and selectivity coefficients under assorted conditions and analyzed competitive adsorption. This work elucidates shale gas adsorption from a molecular perspective and lends insights into storage assessments.
        Methods/Procedures/Process:
        We constructed molecular models of shale nanopores with diverse structures, encompassing hydroxylated quartz nanopores and graphitic organic nanopores. We examined CO2, H2S and CH4 adsorption in these pores under various pressure conditions (373.15K). The molar ratio of CO2, H2S and CH4 was fixed at 4:1:5. CLAYff and Steele 10-4-3 potential models described quartz and graphite correspondingly, while the TraPPe and Nath force field models represented CO2/CH4 and H2S individually. Component densities and chemical potentials were computed based on gas composition. CO2, H2S and CH4 insertions, translations and rotations were conducted over 20 million MC cycles, equalizing pore gas chemical potentials to calculated values. The first 12 million cycles achieved chemical potential equilibrium, and the subsequent 8 million cycles collected gas distribution data. The Langmuir-Freundlich adsorption model fitted the isotherms. Competitive gas adsorption was examined by quantifying adsorption quantities and selectivity coefficients.
        Results/Observations/Conclusions:
        Gas adsorption capacities were superior in organic versus inorganic pores across all gases examined. Graphite ostensibly furnished additional adsorption sites, whereas hydrogen sulfide and carbon dioxide occupied further sites in quartz nanopores. Therein, CH4 adsorption selectivity was markedly inferior to CO2 and H2S. Excess CH4 adsorption approached 0 as pressure rose, however excess CH4 adsorption in organic pores increased with pressure. H2S displayed the maximal adsorption selectivity coefficient universally, imprinting the strongest competitive adsorption capacity, which intensified in organic nanopores. Under certain quartz pore conditions, CO2 adsorption selectivity closely approximated H2S and exceeded CH4 substantially, unveiling comparable CO2 and H2S competitive adsorption capacities in inorganic pores given natural gas presence in reservoirs.
        Applications/Significance/Novelty:
        This study elucidates the competitive adsorption between acidic gases and natural gas in depleted shale gas reservoirs from a molecular vantage, thus facilitating augmented comprehension of CO2/H2S adsorption mechanisms in natural gas-laden formations while steering acidic gas geological storage in shale.

        Speaker: Jingkai Cui (Chengdu University of Technology)
      • 89
        Microscopic mechanism of CO2 huff-n-puff promoting shale oil mobilization in organic/inorganic nanopores

        CO2 huff-n-puff is a potential promising approach for enhanced recovery and sequestration of CO2 in shale reservoirs. It is of great practical significance to understand the CO2 huff-n-puff mechanism from a microscopic point of view. Here, we investigate three stages of CO2 huff-n-puff promoting shale oil mobilization from organic-inorganic nanopores by molecular dynamics simulation. We show that during the adsorption process of shale oil, due to the presence of active molecules, the adsorption density and strength of shale oil on kaolinite wall are higher than kerogen, but the influence range of shale oil is smaller than kerogen. In the CO2 soaking stage, although CO2 has a desorption effect on shale oil near both sides of the wall, stripping shale oil near the inorganic surface was more effective than the kerogen surface. In addition, due to the presence of hydroxyl on the surface, when CO2 is slightly away from the equilibrium position on the surface of kaolinite, the attraction between CO2 and kaolinite will become repulsive force under the action of electrostatic force. In the CO2 puff stage, compared with the ideal model of 0 pressure, when the CO2 puff pressure is 10MPa, CO2 can effectively dissociate the "bullet head" structure of the medium component blocking the pore exit through the synergistic effect of miscible phase, viscosity reduction and swelling. Increase overall shale oil recovery by more than 37%. This work first investigates the CO2 huff-n-puff mobilization of shale oil from multiple stages, and effectively reveal the promoting effects of CO2 on different components of shale oil in each stage of huff-n-puff.

        Speaker: Huaisen Song (China University of Petroleum (East China))
      • 90
        Investigation of Fluid Flow Mechanism Considering Multi-Component Fluids, Nanopore Roughness, and Nanopore Flexibility

        Understanding the occurrence and flow mechanisms of shale oil in nanopores, as well as the impact mechanisms of fluids on solid deformation, is crucial for advancing our comprehension of fluid behavior in porous media. Prior neglect of factors such as the multi-component characteristics of shale oil, the properties of real shale nanopore walls, and nanopore flexibility has resulted in insufficient knowledge regarding the occurrence and flow mechanisms of shale oil in nanopores. In this study, molecular dynamics simulations were employed to extensively investigate the occurrence and flow mechanisms of fluids in graphene, hydroxylated quartz, rough kerogen rigid nanoslits, and flexible nanotubes. The following conclusions were drawn: (1) The occurrence patterns of multicomponent shale oil in organic and inorganic nanopores were revealed. The adsorption characteristics of shale oil are related to the pore wall elements. Components containing oxygen, nitrogen, and aromatic hydrocarbons tend to adsorb more readily on quartz surfaces, while sulfur-containing components also tend to adsorb on the kerogen surface due to interactions with sulfur elements in kerogen. (2) Real shale oil flows in real nanopores without slippage. Through simulations comparing the flow of single/multicomponent shale oil in smooth/rough nanopores, we found that slip phenomena occur only under ideal conditions (single-component oil, smooth surface). The slip does not occur in realistic shale oil flow, offering theoretical support for setting slip length in pore-scale simulations. (3) In quartz, kerogen, and graphene nanoslits, an increase in pore pressure was observed with the elevation of pressure gradients. In rigid graphene nanoslit, fluid flow induces an elevation in nanoslit pressure, with a critical pressure gradient of 1 MPa/nm. Below this threshold, pore pressure exhibits minimal variation; above it, a significant increase is observed. Higher pressure gradients lead to an increase in kinetic energy in the direction perpendicular to the wall, indicating an escalation in collision intensity between the fluid and the wall, as well as among fluid particles, resulting in a rise in pore pressure. Increased pressure gradients reduce the interaction energy between the fluid and the wall, signifying that fluid molecules are propelled further from the wall upon collision, underscoring the gradual intensification of fluid-wall collisions. (4) The intricate relationship between pressure gradient, nanopore stress, and nanopore strain was revealed. Under static conditions, the transition of a rigid and smooth nanopore to a flexible one can result in an increase in surface roughness, which leads to a reduction in the density of the adsorption layer. The pore width decreases and the pore pressure increased slightly. Fluid flow induces an increase in pore pressure and width. Simulations of fluid flow in rigid nanoslits with coupled pore width and rock compressibility, as well as in flexible nanotubes, indicate that an increase in pressure gradient leads to pore expansion. This study revealed the intricate interactions of shale oil in nanopores, offering theoretical support for understanding its flow in porous media and contributing to the efficient extraction of shale oil from unconventional reservoirs.

        Speaker: Dr Tianhao Li (China University of Petroleum (East China))
      • 91
        Molecular dynamics investigation of water-gas two phase flow in rough clay nanopores

        The gas-liquid two-phase flow in rough nanopores plays a crucial role in shale gas extraction. To deeply understand the flow mechanisms, molecular dynamics simulation (MDS) method is often employed to simulate fluid flow in nanoscale channels. However, current researches on two-phase flow at the nanoscale have mainly focused on smooth channels. In addition, accurate simulation of the rock wall material is challenging. This work aims to investigate the mechanisms of gas-liquid two-phase flow in rough shale nanopores by using MDS.
        A new rough surface model is constructed by illite crystals, and the water-gas two-phase flow is simulated in it which can achieve a more accurate characterization of flow phenomena in microscale shale reservoirs. To better represent the actual formation conditions, rough nanopores are constructed by adding roughness elements to smooth wall surfaces. Methane and water molecules are introduced into the pore models. The flow process is simulated using the EF-NEMD method. Based on these, the effects of rough particles, different arrangements of rough particles, and varying relative roughness on two-phase flow are investigated. The simulation results reveal that rough particles have a significant impact on gas-water two-phase flow. Statistical analysis is used to quantify the density, velocity distributions and boundary conditions in two-phase flow.
        Simulations performed under different roughness conditions demonstrate: the presence of rough particles leads to three adsorption layers of water molecules near the pore walls; it also induces a phenomenon similar to macroscopic Jamin effect during two-phase flow, which severely affecting the flow velocity. Another important observation is that compared to smooth channels, the presence of rough particles significantly increases the boundary slip length i.e. the thickness of the immobile water layer. The different arrangements of rough particles will generate different negative slip lengths. Furthermore, with the roughness decrease the influence of rough channels on gas phase flow is negligible in hydrophilic channels. The aforementioned findings provide valuable insights into the gas-liquid two-phase flow behavior in rough nanopores, which is crucial for understanding and optimizing the transport and mass transfer processes in nanoscale systems.
        This work simulates water-gas two-phase flow in rough nanopores constructed by illite crystals, which has not been previously explored. The major contribution of this work lies in analysing the impact of roughness elements on two-phase flow through MDS. It provides a basis for the development of subsequent mathematical models. Simulating the actual shale reservoir can guide the optimization of production measures. These findings provide insights into the intricate dynamics of gas-liquid two-phase flow in rough channels and contribute to a better understanding of fluid transport in porous media with real-world applications, such as shale reservoirs.

        Speaker: Ying Bi (University of Science and Technology of China)
      • 92
        Smart Water Flooding: Experimental study and Molecular Simulation

        In the background of the strong oil wettability and low production by water flooding in carbonate reservoirs, low salinity water containing sulfate ions and nanoparticles can significantly change the surface wettability of carbonate rocks and thus increase the sweeping area, however, the absorption and desorption mechanisms of the oil film in the carbonate rock surface remain unclear. In this work, These problems is addressed in the framework of molecular dynamics simulation (Material Studio software) and experiments. The results were showed that sodium sulfate solution could accelerate the rate from oil-wet to water-wet and the interaction of oil molecules, water molecules, and SO42- ions at molecular scale was explained. The results of the simulations show that many water molecules travel down the water channel under the influence of several powerful forces, including the electrostatic force, the van der Waals force and hydrogen bond, crowding out the oil molecules on the calcite's surface and causing the oil film to separate.

        At the same time, a hybridization technique of combining low salinity water and nanofluids was performed by using experiments such as contact angle measurement, core displacement, and NMR (Nuclear Magnetic Resonance), and the effects of different salinity water and the nanofluids concentrations on wettability alteration and enhanced oil recovery were revealed. The parameters of wettability changes and contact angle were measured at different nanofluid solutions with high/low salinity water. The experimental results revealed that the test with KCl-1+NF outperformed other compositions. As for the new method of hybridization technique, the insights presented in this study provide a good reference for further research in this area. In a word, these investigations can guide the practical application of low salinity water flooding in carbonate reservoirs.

        Speaker: Nannan Liu (常州大学)
    • MS15: 1.2
      • 93
        Deep learning-assisted technology transition in natural hydrogen development

        Although the global energy sector is shifting from the fossil-based energy systems to the renewable energy resources, the conventional energy development techniques has received increasing attentions with the mature development and the recharge by AI. With the help of AI techniques, drawing lessons from thousands of years of traditional energy development in the technology transition into the next-generation energy is an effective approach to accelerate energy transition and avoid repeated research causing unnecessary wasting. In this talk, we will introduce an iterative flash calculation scheme and a deep learning algorithm using a thermodynamics-informed neural network (TINN) to perform accurate, robust, and fast phase equilibrium calculations for realistic fluid mixtures of natural hydrogen. The development of natural hydrogen is an emerging topic in the current energy transition trend. The production process involves compositional multiphase flow via subsurface porous media. This makes studying the compositional phase equilibrium behavior essential for reliable reservoir simulation and prediction. The application of TINN architecture can accelerate the calculations for nearly 20 times. The effect of capillarity on phase equilibrium states is demonstrated. Based on simulation results, suggestions for the natural hydrogen industry chain are provided to control the possible phase transitions under certain environmental conditions that may be observed in the natural hydrogen reservoirs, storage and transportation facilities. The extremely low critical temperature of hydrogen challenges the robustness of flash calculations but facilitates the separation of impurities by liquefying certain undesired species. Moreover, phase transitions under control can be an effective approach for carbon dioxide capture and sequestration with optimized operating conditions over the phase equilibrium analysis.

        Speaker: tao zhang (King Abdullah University of Science and Technology)
      • 94
        Multiparameter Inversion of Reservoirs Based on Deep Learning

        Reservoir parameter inversion is an important technique in oil and gas exploration and development that can estimate the reservoir physical properties, such as skin factor and permeability, using observed data, such as well test data and production data. In this paper, we propose a physical accelerated neural network with multiple residual blocks (PRNN-Acc) for multiple parameter inversion of the seepage equation with a source term and a sink term. PRNN-Acc is based on the idea of physical residual neural network (PRNN), which uses deep neural networks to approximate the solution and parameter spaces of partial differential equations. PRNN-Acc adds multiple residual blocks to enhance the expression ability and flexibility of the network and avoid gradient explosion or degeneration phenomena. In addition, the input of PRNN-Acc is multiplied by three adaptive parameters, which can adjust the network training process according to the characteristics of the data and loss function and improve the accuracy and stability of the inversion. We use bottomhole pressure (BHP) data before and after shut-in as labels to invert multiple parameters for homogeneous and heterogeneous reservoirs. In this paper, three numerical experiments are designed. For homogeneous and heterogeneous reservoirs, the inversion results of this method are up to 36 times more accurate than those of PRNN. It is fully proven that the inversion effect of this method is better than that of PRNN.

        Speaker: Qian Wang
      • 95
        A Vision Transformer for Size-Agnostic Modelling of Two-Phase Drainage in Complex Porous Media Considering Wettability, Interfacial Tension, and Resolution

        The estimation of pore-scale multiphase flow fields in complex geometries using deep learning has proven challenging. This is partly because researchers have historically focused on model architecture and data quality, while the volume and variety of data may have been inadequate to capture the intricacies of multiphase flow. In this work, we introduce a novel deep learning methodology to predict phase distributions within realistic porous rocks during two-phase capillary-dominated drainage. We use Computerised Tomography (CT) images and incorporate pressure gradient, resolution, wettability (contact angle), and interfacial tension as inputs without relying on complicated expert-crafted features.
        To create ground-truth datasets, we extract subsamples from CT scans of both synthetic and real rocks, including sandstones and carbonates. Primary drainage is then simulated in these sub-images by an in-house Pore Morphology-based Simulator (PMS), yielding millions of fluid occupancy instances. To maintain both pixel-wise accuracy and physical fluid connectivity, we devise a Higher-Dimensional Vision Transformer (HD-ViT). We train the model on phase distributions where the wetting phase is drained from pores solely based on their sizes, regardless of their relation to other pores and the inlet, allowing the network to focus on subtle details such as generating valid fluid-fluid interfaces. Fluid continuity is then enforced as a post-processing step by removing patches of the invading phase that are not connected to any desired inlet(s). This approach facilitates efficient inference for images of varying sizes and resolutions with any inlet-outlet setup. After training on a massive dataset of images and rock-fluid data, the model achieves outstanding results with a testing F1 score and saturation correlation coefficient above 0.95.
        We confirm the model’s validity by demonstrating consistently high performance on larger images of unseen sandstone and carbonate rocks through an effective patch-and-stitch strategy. The model maintains accuracy across a wide range of scales, from microns to centimetres, within the range of properties used in this study. Such scalability enables distributed computing, facilitating the processing of extremely large images. Therefore, the reported methodology can be considered a solution to the computational constraints encountered for large images. Interestingly, the HD-ViT proves even faster than the PMS, itself considered one of the most efficient simulators of drainage. This underlines the immense potential of models trained at scale, like ours, to be fine-tuned for computationally intensive simulations using smaller datasets, where the speed advantage becomes increasingly significant.
        Our final model introduces multiple innovative aspects. Firstly, setting it apart from similar models, ours incorporates all factors influencing capillary drainage as inputs, offering a comprehensive approach. Secondly, the model is trained on a dataset of unprecedented size and diversity, comprising millions of highly heterogeneous and realistic images. Thirdly, by avoiding complex feature engineering, we ensure an end-to-end, easy-to-use model. Fourthly, we implement a simple and effective strategy to enforce phase connectivity in fluid distributions and to also allow for size-agnostic predictions on any inlet-outlet configuration. As such, the HD-ViT is a multiscale, practical, and efficient model for pore-scale drainage.

        Speaker: Ping Wu (School of Science, Southwest Petroleum University, Chengdu 610500, China)
      • 96
        A neural network model with physics constraints for simulating CO2 storage in deep saline aquifers during and after injection

        The exploration of CO2 capture and storage has become a crucial element in strategies aimed at mitigating climate change, where deep saline aquifers are of particular interest due to their extensive storage capacity and widespread availability. The complexities involved in effectively monitoring and simulating CO2 behavior within these geological formations present significant challenges. To address these challenges, our research introduces a specialized neural network, designed to simulate and monitor CO2 storage in deep saline aquifers during both the injection and post-injection phases. This neural network represents an integration of physics-based principles and advanced deep learning techniques. This integration facilitates the modeling of CO2's complex movement and distribution under diverse conditions. The network's ability to process and analyze spatial data, coupled with key geological characteristics, significantly enhances the accuracy of its simulations. This aspect is crucial for understanding the heterogeneous nature of subsurface systems and the dynamic behavior of CO2. The architecture of this network, encompassing various computational layers and physics-informed constraints, is designed to ensure comprehensive and precise modeling of CO2 storage processes. This approach aims to contribute to the existing body of knowledge in the field of carbon capture and storage, offering a new perspective in the simulation and understanding of CO2 behavior in subsurface systems.

        Speaker: Mengjie Zhao
      • 97
        2D to 3D deep learning reconstruction of CO2 electroconversion Gas Diffusion Electrode : a validation study

        Direct 3D imaging of natural or synthetic porosity below ~1μm in diameter often requires the application of Focused Ion Beam Scanning Electron Microscopy (FIB-SEM). This technique has several limitations: high cost and time demands, instrument availability, complex sample preparation and low field-of-view (FoV), restricting its suitability for operational industrial studies. SliceGAN, a generative adversarial network algorithm developed by Kench and Cooper (2021) reconstructs 3D pore space based on a 2D image input (Fig. 1). In conjunction with 2D SEM imaging, SliceGAN has the potential to generate representative porous media images considerably faster and cheaper than FIB-SEM. However, the original work (Kench and Cooper, 2021) did not validate the methodology against 3D images to check whether the synthetic reconstructed media replicate the original porous materials porosity, permeability, and pore size distributions.
        The microporous layer (MPL) of the Gas Diffusion Electrode (Fig. 2), used for the electroconversion of CO2 into ethylene, methanol and other products (Weekes et al., 2018), was selected for the validation of the pore space reconstruction algorithm. Fully resolved images of the MPL porosity are necessary for subsequent modelling of electrode physical properties: permeability, diffusivity and conductivity (McLaughin et al., 2022). For this study, four FIB-SEM validation volumes from two scans at 5 and 10nm resolution and FoV varying from 51 to 94μm3 were analysed. Raw data was segmented, porosity calculated, and permeability of each volume modelled using a steady state single-phase direct numerical simulation in OpenFOAM (Fig. 3). Permeability variation from 4.04e-16 to 8.03e-16 m2 between different volumes is attributed to an insufficient FoV of individual scans, leading to unrepresentative models. A large image aspect ratio, inherent to FIB-SEM scanning, also led to variation of permeability in three principal orientations in the absence of any noticeable material anisotropy.
        The open-source network architecture of SliceGAN was adapted to accommodate varying training data crop sizes. Two orthogonal SEM images at 1.7nm resolution and FoV of 315 μm2 were used to prepare training datasets. Ten different studies, combining different crop sizes and input image resolutions, were trained to assess the effect of the 2D training image FoV on the quality of the reconstructed 3D volume. For each study, inference volumes with a maximum FoV of 163μm3, limited by RAM capacity, (Fig. 4) were generated, and their porosities and permeabilities calculated (Fig. 5). While properties vary between studies and between inferences within each study, an overall good match of porosity, permeability, and pore size distributions with FIB-SEM data was observed, with further research to be done on multiphase flow properties of the artificial volumes.
        This study shows the verification of the 2D – 3D reconstruction algorithm, which, while not creating an exact copy of the real pore space, is able to replicate properties necessary for larger-scale physics modelling. It is also capable of generating significantly larger volumes when compared to FIB-SEM validation data. This, together with the fact that the network is trained on 2D data with a large FoV, addresses the representability issues observed during the FIB-SEM image analysis.

        Speaker: Georgy Borisochev (CHLOE)
      • 98
        Deep Learning enhanced multiscale rock typing for digital core modeling

        Currently, there is rapid development in the approaches for constructing and utilizing digital cores. Digital Rock Physics (DRP) methods allow for quick and non-destructive acquisition of rock properties. The process of digital rock physics involves two primary stages: model construction and simulation of physical processes on the created models.
        For heterogeneous reservoir rocks, the usage of DRP is not as straight forward as for high porosity sandstones. This is due to the inherent trade-off between the spatial resolution of data and the representativeness of the model size. The primary goal of this study was to establish a technique for upscaling digital core models from micro to macro scale, enabling the computation of rock properties while accounting for heterogeneities of various scales.
        The upscaling procedure involved searching for correlations between tomography data of different resolutions and transforming low-resolution tomography into a multiclass model according to the found correlation. The approach of using convolutional neural networks for high-resolution tomography data was considered as the optimal algorithm for transforming low-resolution tomography into a multiclass model. The output of the neural network was an upscaled model of lower resolution than the original tomography. Each element (voxel) of the upscaled model belonged to one of several digital types of rock, whose generalized characteristics were determined based on the analysis of high-resolution tomography data.
        To validate the upscaling technique we constructed a digital model of complex carbonate reservoirs based on data from multiscale microtomography. A multiclass model concept has been created and experimented with, enabling the computation of flows in pore spaces of various scales. By incorporating diverse pore space structures as supplementary classes in the multiscale model, it is possible to preserve a substantial physical size of the model while enhancing its level of intricacy.

        Speaker: Denis Orlov (Skolkovo Institute of Science and Technology)
    • MS21: 1.2
      • 99
        Wave-mediated diffusion model for semi-sealed systems: effective diffusion coefficient and experimental validation

        It is widely known that the motion of a gas at a low Mach number can be approximated as an incompressible flow at the leading order in a small Mach number expansion of the full solution (Anderson 1995; Panton 2013). However, it has been shown recently that such an incompressible approximation becomes invalid for a semi-sealed system with no inlets and no boundary movements. Studies based on linearized compressible Navier-Stokes equations for such a system made of small capillary tubes have revealed some counter intuitive flow characteristics such as no-slip flow with a slip-like mass flow rate (Chen and Shen 2018a, b; Shen and Chen 2019a, b; Shen and Chen 2020). In this work, we extend these works to a semi-sealed porous system which has applications to microfluidics and primary production from a tight gas reservoir.

        Based on the compressible N-S equations and the theory of Klainerman and Majda (1982) for low Mach number flow, Jin and Chen (2019) has shown that at the pore-scale, the flow of the gas obeys a damped wave equation. Applying multi-scale analysis and volume averaging upscaling to the pore scale equation, Jin and Chen (2019) obtained a self-diffusion equation at the macroscopic scale in the limit of infinitesimal pore size. To account for small but not infinitesimally small pores, the effective macroscopic diffusion coefficient must be modified to consider the effect of wave-mediated diffusion. We first perform pore-scale numerical simulations of drainage flow from a porous plug using the damped wave equation. The mass flow rate from this simulation is then matched to the one computed from the macroscopic diffusion equation with an adjustable diffusion coefficient. The diffusion coefficient that provides such a match in the mass flow rate is then the effective diffusion coefficient.

        A large number of pore scale simulations are performed for various pore structures. We study slow viscous drainage flow of a viscous compressible gas from a semi-sealed porous plug to a large vessel. The semi-sealed porous plug has a length of , height of ; and the vessel has a height of and extends to infinity downstream. We then use the homogenized medium shown in Figure 1 and the macroscopic diffusion equation with various effective diffusion coefficients to compute mass production rate from the plug. From the simulation results, it is found that the larger the porous plug length and its porosity, the larger the effective diffusion coefficient. The larger the expansion ratio, the smaller the effective diffusion coefficient. An empirical correlation of the effective diffusion coefficient is then established for applications to larger physical size. The proposed wave-mediated effective diffusion model as well as Darcy’s law are both used to perform historic match with the data from laboratory experiments. The comparison shows that the wave-mediated effective diffusion model provides good agreement with experiments whilst Darcy’s law severely underestimates the flow rate.

        The proposed wave-mediated diffusion model is promising for applications for primary production from tight gas reservoirs. Testing of this model with field data is currently underway.

        Speaker: Shiming Wei (China University of Petroleum, Beijing)
      • 100
        Non-linear growth of fingers during two-phase flow in porous media

        Invasion of a fluid in a porous medium filled with another fluid immiscible to the injected one produces a wide variety of displacement patterns depending on the fluids’ viscosity contrast and the capillary number of the flow. In the case of a low-viscosity fluid advancing through a high-viscosity immiscible fluid, a viscous instability occurs, leading to viscous fingers which have long been assumed to exhibit a linear Laplacian growth behavior. This means that the interface velocities of the advancing fronts depend linearly on the local pressure gradient. This (linear) Laplacian growth behavior is also observed for viscously-unstable fingers observed in continuum Hele-Shaw cells by Saffman and Taylor [1], as well as for diffusion limited aggregates (DLA). However, an experimental study of drainage in a porous Hele-Shaw cell around 20 years ago [2] demonstrated that drainage fingers in porous media can also exhibit non-linearity in the growth in a certain regime. Recently we further investigated this configuration of drainage displacement with a dynamic pore-network model [3] and measured the local growth rate of the fingers as a function of the local pressure drop. We showed that there exists a regime where the two quantities relate nonlinearly with a power-law, which then crosses over to a linear Laplacian growth regime at higher capillary numbers [4]. The origin of this nonlinearity is the disorder in the capillary barriers at the pores [2, 4], and the pore-size distribution, through the distribution of capillary thresholds, controls this non-linearity. In our recent study [4], the pores size distribution was uniform. Here in this talk we will present our further investigations on the sensitivity of these results on the distribution of pore sizes, and how the exponent related to the fingers’ growth law depends on the functional form of that distribution.

        Speaker: Dr Santanu Sinha (PoreLab, Department of Physics, University of Oslo, N-0316 Oslo, Norway)
      • 101
        Modeling Non-Newtonian Polymer Flooding in Heterogeneous Carbonate Rock: An Experimental and Simulation Investigation

        Polymer flooding is a widely used chemical Enhanced Oil Recovery (EOR) technique in carbonate reservoirs that can decrease water-oil mobility ratio and thereby enhance sweep efficiency. However, the accuracy of simulating polymer flooding in porous media relies on integrated characterization on polymer properties, especially rheological behavior of polymer. The objective of this study is to accurately model and predict shear-thickening and shear-thinning behavior of polymer injection of heterogeneous porous media by incorporating Special Core Analysis (SCAL), bulk rheology test, injectivity test and coreflooding experiments.
        This study started with integrated fluid and reservoir rock characterization. Injectivity tests and coreflooding experiments were then conducted, including sea water flooding and polymer flooding in low-permeability (20 mD) and high-permeability (200 mD) outcrops. The pre-constructed polymer coreflooding simulation model was history matched with experimental results and uncertain parameters were calibrated by optimizing key indicators reflecting polymer non-Newtonian behavior in porous media. The calibrated model was then used to model the polymer in-situ rheology and EOR performance in a heterogeneous core sample combining high-permeability and low-permeability layers. Nuclear Magnetic Resonance (NMR) technology was used before and after polymer flooding to confirm the pore size distribution affected by polymer injection.
        Following the rock characterization study, the base simulation model M1 was incorporated with data on porosity, end-point water and oil permeabilities, and fluid viscosity. After polymer rheology and adsorption studies, the model was upgraded to M2, which displayed increased accuracy in polymer in-situ rheology and integrity affected by adsorption. Model M2 was utilized to history match base water and polymer oil displacement efficiency experiments, and the initial match degree was evaluated. Uncertain parameters, including apparent viscosity, Inaccessible pore volume, and relative permeability curve, were then adjusted by optimizing an objective function that included pressure drop, water breakthrough time, and cumulative oil production. After multiple iterations of history match, significant improvements in accuracy were observed in the calibrated model, which was then utilized to forecast polymer coreflooding performance in the heterogeneous carbonate core sample. However, the initial match was insufficient due to the complex crossflow of polymer solution between the high permeability layer and low permeability layer. Adjustments to the vertical permeability resulted in the final model M3, which achieved a high history match degree with experimental results.
        In this investigation, A polymer that has both shear-thickening and shear-thinning features was modeled and calibrated step by step based on experimental results. The outcome was the creation of highly precise simulation models with the capability of providing forecasts for polymer in-situ rheological behavior, viscous fingering phenomenon, and EOR performance in heterogeneous carbonate rock. This methodology is instrumental in advancing our understanding of the mechanisms behind the non-Newtonian flow behavior of polymers in Darcy-scale porous media through integrated experimental and simulation investigations.

        Speaker: Mr Chuangchuang Qi (Khalifa University of Science and Technology)
      • 102
        Incorporating Pore Size Distribution into Dynamic Permeability Modelling for Porous Media

        Previous dynamic permeability models often relied on simplified and assumed pore-scale parameters such as average pore radius, potentially leading to inaccuracies. This study introduces a novel approach that directly incorporates measured pore size distributions, addressing these limitations and providing a more realistic representation of fluid flow in porous media. Key findings include:

        • Pore size distribution's primary impact on dynamic permeability:
          The model demonstrates that pore size and its distribution exert a
          first-order effect on dynamic permeability.
        • Integration of measurable pore size data: The model utilizes pore size distributions obtained from techniques like MICP or NMR, enabling direct
          incorporation of measured data for enhanced model accuracy and
          applicability.
        • Incorporation of slip boundary conditions for
          wettability effects: The model accounts for wettability's influence
          on dynamic permeability, providing a more comprehensive understanding
          of fluid flow behaviour in porous media with varying wettability
          characteristics.

        Overall, this study presents a practical dynamic permeability model that overcomes limitations of existing approaches by incorporating pore size distribution and wettability effects. This model holds significant potential for improved characterization and understanding of fluid flow in diverse porous media applications, potentially leading to advancements in various fields.

        Speaker: Dr Jimmy Xuekai Li (The University of Queensland)
      • 103
        Universal scaling law of bubble dissolution in porous media

        Dissolution of bubbles in porous media affects many important geological and engineering processes, such as CO$_{2}$ sequestration, hydrogen storage, fuel cell water management and Li-battery electrolyte filling. However, although dissolution of continuous gas phase has been investigated extensively, there is still no universal theory for static bubble dissolution kinetics in porous media. We thus conduct experimental and theoretical investigation on the universal dissolution theory for bubbles in porous media.
        We found that porous structure does not only change the effective diffusivity, but also regulate the characteristic mass transfer distance and surface area. In open space, the mass transfer area (S) of the bubble is the surface area of the bubble itself, and the characteristic mass transfer distance is proportional to the bubble radius R 1. These result in V~t scaling in 2-d and V$^{2/3}$~t scaling in 3-d in quasi-steady state, where V is the dissolved volume. However, in porous media, both its mass transfer area and its characteristic mass transfer distance are re-regulated by the porous structure. The concentration in the pore body is almost uniform, while the concentration gradient is concentrated along the throat (Fig.1a). Consequently, the characteristic mass transfer distance is determined by throat length (L$_{t}$), and effective mass transfer area is determined by the throat cross section area (A$_{t}$) and available number of throats (n).
        For a bubble in a single pore, n is a constant, resulting in a linear V~t scaling. For a large bubble that occupies many pores, n depends on the shape of the bubble. If the bubble is completely liner, n is proportional to the volume of the bubble V, so we can deduce lnV~t. While if the bubble is a bulk, n$^{2}$ is proportional to V, leading to a V$^{1/2}$~t scaling. These conclusions have been well verified by experiments. (Fig.1b).
        In addition ,we investigate the dissolution of CO$_{2}$ bubble in porous media under gravity field, where the Rayleigh-Darcy convection can be induced in the liquid phase as dissolved CO$_{2}$ increases water density. In case that the mass transfer is dominated by convection, the bubble dissolution rate becomes proportional to the flux of fresh water through the bubble. For an isolated CO$_{2}$ bubble in an infinitely-large porous medium, the flux is determined by the number of pores it occupies perpendicular to the gravitational field n_vert. Therefore, for a CO$_{2}$ bubble in a single pore (Fig.1c), the dissolution is in constant rate (V~t), that is experimentally examined.
        In summary, we establish the scaling laws for bubble dissolution in porous media. We show that bubble dissolution kinetics in porous media becomes very different from that in bulk, not only in the value of diffusivity, but (more significantly) also in the volume-time scaling, because porous structure regulates the mass transfer length and area.

        (a) The concentration field of dissolved bubbles in a single pore (b)Data on the volume of bubble dissolution over time under different conditions(c)The concentration field in which a bubble dissolves in a gravitational field

        Speaker: Yuehongjiang Yu
    • Poster: Poster Session II
      • 104
        A fully implicit single-phase multi-component phase transition pore network model based on automatic differentiation and GPU acceleration

        The complex phase behavior of hydrocarbon mixtures is encountered in miscible flooding in the oil-saturated reservoir and liquid dropout in gas-condensate reservoir. In pore-network models, phase equilibrium calculations (Michelsen, 1982) have been coupled with convective-diffusion equations to evaluate the influence of hydrocarbon phase behaviors to flow and transport at given hydrocarbon mole compositions and temperature, as reported by Chen et al. (2020) and Santos, M.P.P.C. et al. (2020) where a fully implicit Euler method is implemented to get the set of non-linear algebraic equations at each time step solved by Newton’s method.
        The number of convective-diffusion equations varies with pseudo-components number, which brings great inconvenience to the analytical programming of Jacobian matrix. In this work, the reverse derivation technique based on the chain derivation rule (Baydin et al., 2017) is applied to evaluate Jacobian matrix using a open source C++ library named fadbad++ and the program is comparable in time to the analytical programming of Jacobian matrix.
        In addition, the GPU parallelization is promising to the numerical studies of multi-component flow and transport with computationally intensive nature. In our test, a graphic card RTX 3060 with 12GB memory together with CUDA library is used to speed up the steady-state and transient simulation process. The program is compared with the one running on a CPU i5-12400F with the Eigen mathematical library. The comparison results show that the acceleration rate can be about 5 times in the single-phase dynamic simulation with a pore network of 8000 pore elements, and about 10 times in the single-phase steady-state simulation with 2 million pore elements.
        The coupled model is used for gas-condensate reservoir simulation to investigate the effect of droplets on gas production processes. In future work, a more efficient coupling way will be explored (Collins et al., 1992), and the existing model will be advanced to two-phase flow simulation (An et al., 2023).

        Speaker: JIanqi Rong (Chongqing University)
      • 105
        A pore-scale lattice Boltzmann model for solute transport coupled with heterogeneous surface reactions and mineral dissolution

        In this paper, we propose a pore-scale lattice Boltzmann model to treat heterogeneous surface reactions coupled with mineral dissolution. The primary innovation lies in the transformation of surface reactions, originally treated as boundary conditions, into volume source terms through dimensionality augmentation within the framework of sharp liquid-solid interfaces. This significantly simplifies the implementation, particularly for reactions occurring in porous media with intricate geometric structures. Several benchmark tests were performed to validate the accuracy of this model, including a reaction-diffusion problem in a rectangular domain, a two-dimensional reaction and dissolution of a circular grain, as well as a three-dimensional calcite crystal dissolution in a micro-channel. All the obtained simulation results agree well with the reference solutions. In addition, a dissolution problem in a three-dimensional porous meida bulit with the sandpack is then investigated. Cases with different Peclet numbers (Pe) and Damkohler numbers (Da) were simulated, and five dissolution modes were obtained, which were finally summarized in a diagram of Pe and Da.

        Speaker: Ju Long (King Abdullah University of Science and Technology)
      • 106
        A study on the CO2 displacement behavior at nanoscale considering rough surface

        CO2 displacement is considered as a potential method to enhance shale oil recovery. CO2 can reduce the viscosity and surface tension of crude oil, making it possible to recover crude oil in the nanopores. At the same time, the CO2 can also be partially stored underground, reducing the carbon footprint of the hydrocarbon extraction process. Therefore, understanding the CO2 displacement in nanometer pores of shale is critical for developing effective CO2 injection techniques. In this work, we applied direct numerical simulation to study the effect of rough surface on CO2 displacement in nanometer pores of shale. By quantifying the CO2 displacement in rough nanochannels, we aim to understand how surface roughness and morphology affect the displacement process. After considering the influence of slip effect, the CO2 displacement process in three channel models was studied (single channel, pore throat structure, nanoporous media). We found that in a single channel, the rough surface leads to the reduction of CO2 displacement paths, slowing down the displacement rate. In addition, the pinch-point effect of the rough nanochannel prevents the smooth progression of the interface contact line. The Periodic fluctuations at the interface further hinder CO2 displacement. The smoother the convex and convex transition of rough surface, the smaller the resistance effect of the pinch-point effect. In the pore throat structure model, the rough surface makes it easier for residual oil to remain in the pore. We also simulated CO2 displacement in rough nanoporous media and found that rough surfaces lead to a substantial reduction in CO2 displacement efficiency. Our simulation results show that the surface roughness of shale nanometer pore has nonnegligible effect on CO2 displacement.

        Speaker: Keli Ding (China University of Petroleum (East China))
      • 107
        Adsorption Swelling and Anisotropic Characteristics of CO2 in Shale

        Abstract. Geologic sequestration of carbon dioxide (CO2) is one of the most significant technologies to combat climate change at present. Nevertheless, the CO2 injected into shale reservoirs can expand to affect the permeability and strength of the reservoirs, affecting the efficiency of injection and the safety of storage. In this work, the strain behavior of He (1300 psi) andCO2 (850 psi) on shale samples at constant hydrostatic pressure was investigated using a self-developed high temperature and high pressure gas adsorption and expansion apparatus measuring temperatures at 308 K. The results indicate that adsorption expansion of CO2 exists in shale samples. With increasing pressure, the swelling rate increases and then decreases, and the adsorption-induced swelling strain of shale shows a Langmuir-like relationship with pressure. The adsorptive deformation of shale is anisotropic, with deformation perpendicular to the direction of the laminae being greater than that parallel to the plane of the laminae. The asynchronous response of adsorptive swelling and mechanical compression produced by CO2 gas can lead to crack expansion in rocks and rock fracture. The amount of swelling is dependent on the CO2 concentration, and the swelling of shale is mainly determined by the partial pressures of the component gases.

        Speaker: Shuangshuang Lin (Institute of Rock and Soil Mechanics, the Chinese Academy of Science)
      • 108
        Anti-hydrate Surface Design for Utilization in CO2 Sequestration Processes

        The undesired CO2 hydrates formation in wellbores or pipelines often poses a significant risk to production safety. CO2 hydrates can develop during numerous processes, such as CO2 injection for geological or saltwater storage, production wells in CO2 flooding for enhanced oil recovery, and CO2 pipeline transportation. These hydrates form when gas-water two-phase conditions are met within the hydrate stability zone. Particularly during the transition of supercritical CO2 into the gas phase, the decrease in temperature promotes the hydrate formation. Once these hydrates accumulate on a large scale and form blockages, the bulk hydrates significantly compromise the safety and efficacy of CO2 storage or hydrocarbon production. To address the challenges posed by hydrates, an effective alternative solution is to develop a new generation of passive anti-hydrate surfaces that can prevent hydrates from accumulating over time. These new surfaces can incorporate smart properties such as anti-hydrate nucleation on the surface initially and possess low hydrate adhesion strength if hydrate deposition inevitably occurs. Therefore, it is necessary to understand the fundamental interactions between CO2 hydrates and solid surfaces. To achieve this, the current study employs systematic atomistic modeling and large-scale molecular dynamics (MD) simulations to explore the underlying mechanisms and key factors influencing hydrate adhesion. The results indicate that the gas concentration in the vicinity of solid surfaces plays a crucial role in determining the structures of the hydrates intermediate layer formed on those surfaces. By increasing the gas content near solid surfaces, it becomes possible to weaken CO2 hydrate adhesion, enabling the automatic detachment of hydrates under the influence of shear flow. With a better understanding of these mechanisms, it is conceivable to develop more effective anti-hydrate strategies and enhance the safety and efficiency of CO2 utilization process.

        Speaker: Rui Ma (University of Regina)
      • 109
        Assessing the Representativeness and Precision of Three-Dimensional Digital Rock Modeling: A Case Study on Tight Sandstone

        Tight sandstones are characterized by low porosity and permeability, high clay content. Measuring the rock physics properties under low water saturation conditions using the displacement method poses significant challenges. Digital rock physics (DRP) has been emerged as a valueble method for studying of rock physics of unconventional reservoir. It should be noted that the resolution of X-ray Computer Tomography (CT) scans and sample size can impose mutual restrictions. In order to enhance the applicability of rock physics numerical simulation results, it is crucial to adequately assess the representativeness and accuracy of three-dimensional digital rocks. In this study,3 sandstone samples with porosities of 17.0%,10.8%, and 8.4%, and permeabilities of 339.7,13.2, and 0.94 mD, were selected to construct digital rocks. Seven sub-samples with diameters of 25.4,9,7,5,3,2, and 1mm were prepared for each sample. We utilized X-ray CT scanning to generate three-dimensional grayscale images of the samples, with resolutions ranging from 13.5μm to 1.1μm. These images were then segmented into five components- pores, clay, feldspathic, potassium feldspar, and high-density minerals- using a machine learning image segmentation algorithm. The volume content of the principal minerals in the multi-mineral component digital rocks was calculated and compared with the XRD measurement to assess the representativeness of the three-dimensional digital rocks with different size. The porosities of the digital rocks were determined and compared with the porosity measured in lab. This comparative analysis was conducted to evaluate the precision of the digital rocks. The outcomes of three-dimensional digital rock modeling for tight sandstones reveal that three-dimensional grayscale image acquire via CT scanning for the sample with the diameter of 25.4 mm exhibits difficulty in distinguishing between pore spaces and primary mineral types. By considering the composition of randomly distributed high-density minerals as a metric for assessing representativeness, it was found that the variability of this mineral component increases when the sample diameter is less than 5 mm. This suggests that samples smaller than this size may not adequately capture the macroscopic physical properties. As the sample size decreases, the porosity identified in the digital rock increases. However, it consistently remains lower than the experimentally measured porosity, even in the highest resolution 1 mm sample. When accounting for micropores that are smaller than the scanning resolution of CT, and incorporating them into the multi-mineral digital rocks, the computed porosities agree well with those measured in lab.

        Keywords:Tight Sandstone, Machine Learning, Digital Rock Physics, Multi-mineral 3D Modeling, X-ray CT Scanning.

        Speaker: Mr Fei Xian (China University of Petroleum East China)
      • 110
        Changes in the acoustic signature of tight sandstone during spontaneous imbibition process

        Abstract: Spontaneous imbibition is a process in which porous media spontaneously inhales wetting liquid driven by capillary force, which is an effective means to enhance oil recovery in tight reservoirs. At present, the observation methods of spontaneous imbibition mainly include nuclear magnetic resonance method and computer tomography method. In this paper, the method of ultrasonic testing is used to link the change of seismic attributes of rocks with the distribution of fluids, indirectly observe the spontaneous imbibition process of rocks, and explore the imbibition law of tight sandstone. The high-pressure mercury intrusion method and low-temperature nitrogen adsorption method were carried out on two kinds of tight sandstone. The pore structure parameters of the rock were calculated, and the complexity of the pore structure was quantitatively described according to the fractal characterization method. In addition, the ultrasonic test of the imbibition process of the two rocks was carried out, and the flow of the fluid was observed by the change of the velocity and amplitude of the elastic wave, and the influence of the pore structure of the rock on the imbibition was analyzed. The results show that the pore structure of tight core is complex, mainly micron pores. The initial rate of imbibition is faster, and the rate gradually slows down with the increase of imbibition height. When the fluid front reaches the vicinity of the sensor, the velocity and amplitude of the ultrasonic wave are strongly affected.

        Keywords: Tight sandstone, spontaneous imbibition, ultrasonic monitoring

        Speaker: Mr Fangzhou Zhao (China University of Petroleum (Beijing))
      • 111
        Consideration of the effect of interlayer spatial distribution on the mechanical behaviour of porous media

        In the process of forming porous media by deposition of particles, due to factors such as deposition rate, gravity sorting effect and fragmentation of coarse particles, different structures of interlayers are usually formed, and the form of distribution of interlayers has an important effect on the stress distribution, structural strength and deformation properties of the porous media. However, existing structural parameters such as voidness, coordination number and friction angle cannot fully characterise the effect of interlayers on the structural stability and contact anisotropy of porous media under stress. In carrying out the research on the structural stability of porous media interlayer, the discrete element method is used to simulate the mechanical behaviour of particles in the process of compaction under triaxial stress servo, to reveal the mechanism of the influence of different interlayer parameters on the overall structure of the porous media, and to analyse the influence of the thickness and quantity of the interlayer on the mechanical behaviour and structural deformation of porous media according to the structural parameters of the formation of the porous media and the parameters of the interlayer particles. Based on the structural stability and stress anisotropy of porous media, the mechanical properties of the intercalation were found to have a particularly significant effect on the macroscopic strength and structural stability of porous media, taking into account the inter-particle contact force, contact direction, and peak stress correlation.

        Speaker: Mingxin Liu (China University of Petroleum(East China))
      • 112
        Covalent Organic Frameworks Supported Highly Active Fe-N-C Catalyst Boosting Oxygen Reduction in Direct Formate Fuel Cell

        Direct liquid fuel cells have become an ideal power source for rapidly emerging miniaturized and portable electronic products due to their advantages of cleanliness, environmental friendly, high efficiency, safety, long endurance, and fast "charging". However, at present, the cathode catalysts for oxygen reduction reaction (ORR) of such fuel cell are still mainly platinum or platinum-group noble metals, which leads to the high cost. In addition, these precious catalysts may be poisoned and inactivated during the operation of the fuel cell, seriously affecting the output performance and stability.
        In recent year, covalent organic frameworks (COFs) have emerged as a potential materials for energy storage and electrochemistry conversion due to their high porosity, atomically precise structures and designable topological architectures. Thus, COFs material was synthesized in this study to serve as the support for FePc by facilely sovolthermal process, forming highly active Fe-N-C catalyst to boost ORR in direct formate fuel cell (Fig. 1a). Different load of FePc into COFs (Fig. 1b) was investigated and one could see that the most active COFs-supported catalyst (FePc1@COFs5) exhibited higher onset potential of 0.929 V (vs. RHE) and half wave potential of 0.862 V (vs. RHE) than that of commercial Pt/C (0.928 and 0.845 V (vs. RHE)) (Fig. 1c). Further, the direct formate fuel cell with FePc1@COFs5-coated cathode also archived higher power density and limiting current than that with Pt/C catalyst (Fig. 1d). The facile synthesis process and high performance of COFs-supported catalyst broaden the development for COFs application in electrochemistry energy conversion.

        Fig. 1 (Seen in attachment) (a) Schematic illustration of membrane-free direct formate fuel cell with Pd-deposited Ti mesh anode and COFs-supported catalyst-coated air cathode, (b) Fe content in different samples, (c) Linear sweep voltammetry scanning for different samples in 0.1 M KOH by scanning rate of 10 mV/s, (d) power density curves of membrane-free direct formate fuel cells with 2 mg/cm2 Pt/C-coated cathode and 2 mg/cm2 COFs-supported catalyst-coated cathode.

        Speaker: Mr Linghan Lan (Guilin University of Electronic Technology)
      • 113
        Determination of the type of free gas transport in shale gas formations based on Knudsen number from molecular perspectives

        The type of free gas transport in shale gas formations includes viscous flow, slip flow, and Knudsen diffusion. These three types of transport are categorized based on Knudsen number (Kn), which is defined as the ratio between the mean free path (MFP) of gas and the pore width. The MFP of gas in nanopores is usually estimated based on the ideal gas model. However, the gas in the nanopores is not evenly distributed due to the interactions between gas and walls, and thus the gas in nanopores cannot be viewed as ideal gas, meaning the real value of Kn may deviate from the value obtained by ideal gas model. In this study, we calculated the Kn of methane (CH4) in nanopores by molecular dynamics simulations. The values of MFP in nanopores were obtained based on the trajectories of CH4. We investigated the proportions of viscous collision, slip collision and Knudsen collision, which determine the type of gas transport. By analyzing the proportions of forementioned three types of collision for different values of Kn in nanopores, a real criterion for determining the type of free gas transport was established. Results show that, at 353.15 K with the pressure lower than 50 MPa, the value of Kn of CH4 is smaller than 0.1 in the pore with the width less than 5 nm. The major type of CH4 flow is the viscous flow when Kn < 0.07, and the slip flow should be considered when Kn > 0.07. The Knudsen diffusion cannot be ignored when Kn > 0.08. The results obtained in this study are crucial for correctly determining the type of gas transport in shale formations.

        Speaker: Mr Xinyi Zhao (China University of Petroleum (East China))
      • 114
        Do capillary and film water have equal matric suction or not in simple geometries ?

        The pore water retained in the unsaturated soil includes film water stagnant on the solid surface and capillary water in corners or pores, which can be morphologically quantified by film thickness and radius of meniscus curvature, respectively. The current procedure and theory of soil water retention (SWR) measurement assume that capillary water and film water are connected in the medium saturation range and have equal suction values when they coexist especially in the clayey and loamy soils. However, the results of theoretical derivation and numerical simulations from this work reveal that the film thickness remains constant with the varying radius of meniscus curvature, not following the classical relationship. Moreover, the adsorptive suction is much higher than the capillary suction and the adsorptive interaction on solid boundary affects the shape of SWR curve significantly. These findings not only help to establish the SWR function with accurate physical meanings in the field of soil physics and hydrology, but also have important implications for understanding the general relationship between film thickness and meniscus curvature as well as measuring the disjoining pressure isotherm in colloid chemistry and interface science.

        Speaker: Dr Zi Li (School of Engineering, Westlake University)
      • 115
        Droplet motion in flexible channels: Effects of opening angle and wettability

        Passive and directional droplet transport has gained significant interest due to their potential applications, e.g., self-cleaning surfaces and atmospheric water harvesting. One novel mechanism, known as bendotaxis, involves droplets spontaneously deforming an elastic channel via capillary pressure, thereby inducing droplet motion. However, current studies have primarily focused on parallel channels, neglecting the potential influence of channel geometry on droplet motion and transport efficiency. This study aims to investigate the combined effects of channel opening angle, structural flexibility, and surface wettability on droplet motion dynamics. We employ a comprehensive approach, combining macroscopic-scale experiments, numerical simulations, and a simplified mathematical model to explore different transport modes and their associated timescales. The current study offers insights into directional droplet transport phenomena, leading to potential technological advancements in various fields.

        Speaker: Haiyi Zhong
      • 116
        Effect of catalyst particle size distribution in the catalytic layer on the performance of water electrolysis in proton exchange membrane pore scale simulation

        In this study, the effect of catalyst particle size on the performance of proton exchange membrane water electrolyzer (PEMWE) was studied by using the Lattice Boltzmann Method (LBM). The results show that compared with the catalyst particle distribution, the catalyst particle size is the main factor affecting the performance of the catalytic layer of the anode. Homogenized catalyst particles with smaller particle size can effectively increase the specific surface area of catalyst particles and increase the ECSA of the catalytic layer, thereby improving the electrochemical reaction performance of the catalytic layer. the electrochemical performance of the catalytic layer can be effectively improved by using catalyst particles with smaller particle size and more uniform distribution in the preparation of the catalytic layer (the average local reaction current is 0.13A/cm2 at 3V)
        This study investigates the effect of catalyst particle sizes on the performance of a PEMWE using the lattice Boltzmann method (LBM). The findings reveal that the size of catalyst particles plays a crucial role in determining the performance of the anode CL, surpassing the influence of catalyst particle size distribution. Utilizing smaller and more uniform catalyst particles enhances the specific surface area and electrochemical reaction performance of the CL.
        Key words: PEMWE,LBM,Catalyst layer,Gaussian distribution

        Speaker: Jiaxin He (ChongQing University)
      • 117
        Efficient solution strategies for a generalized coupled fluid-porous problem

        Coupled fluid-porous systems appear routinely in environmental, biological, and industrial applications. The flow interaction between the free fluid and the porous medium is strongly interface driven and can be described by the sharp interface or the transition region concept. Classical interface conditions based on the Beavers--Joseph approach are valid only for unidirectional flows parallel or perpendicular to the fluid-porous interface.

        In this work, we present a coupling concept which is suitable for arbitrary flow directions. We consider a narrow transition region between two flow domains and derive a hybrid-dimensional Stokes--Brinkman--Darcy model (Ruan & Rybak, FVCA, 2023). The transition zone resolves the storage and transfer of mass, momentum, and energy and can be regarded as a complex interface. We validate the proposed coupling concept numerically against the pore-scale resolved simulations. To solve the coupled problem efficiently, we develop and investigate several preconditioners.

        Speaker: Linheng Ruan (Institute of Applied Analysis and Numerical Simulation, University of Stuttgart)
      • 118
        Extensive pore modelling (XPM) – a coherent framework for multiscale pore network modelling

        It is not uncommon for porous media to span multiple scales of heterogeneity. Geological formations are examples of such complex systems that may act as natural aquifers, hydrocarbon reservoirs or greenhouse gas sequestration units. Application of conventional single scale modelling approaches is not sufficient for representative prediction of flow in such heterogenous permeable media. Instead, a method that marries features of different heterogeneity scales needs to be established and validated.

        Three-dimensional digital images of pore spaces are the foundation for numerical pore scale modelling. Depending on the image resolution and the underlying pore structure, voxel data may not be exclusively binary (void or solid), but rather a collection of grey values that indicate under-resolved porous regions. Traditional pore networks have already demonstrated their efficiency and accuracy when modelling single scale macroscopic properties where the porosity is fully resolved. However, rigorous capture of under-resolved heterogeneity remains a difficult task for this class of models.

        In our work, we aim to address this shortcoming by introducing an additional set of entities referred to as Darcy nodes that complement existing pore network macro nodes and throats. Physically, the Darcy nodes correspond to under-resolved regions that are characterized by its porosity and permeability. The proposed novelty is the more systematic consistency and flexibility of the Darcy nodes allocation and integration into the existing pore network modelling workflow in comparison to the previously published methods of microlinks (Bultreys et al., 2015) or very large stochastic explicit networks (Jiang et al., 2013). We established a methodology that unites laminar and Darcy flow mechanisms as well as their transitional behaviour, similarly as it is done in the multiscale Darcy-Brinkman formulation, as shown in Figure 1. The accuracy and robustness of our model, as implemented in the XPM (extensive pore modelling) simulator, is confirmed by a comparison with more physically complicated direct numerical simulation modelling results. Finally, our development is open source that is freely and readily available to the wider audience.

        Speaker: Julien Maes (Heriot-Watt University)
      • 119
        InPore: Image-based and GPU-Accelerated Volumetric Lattice Boltzmann Method for Pore-Scale Porous-media Flows with Applications

        Recently developed image-based computational fluid dynamics (ICFD) techniques have revolutionized the study of pore-scale porous media flows (PSPMFs) by allowing for simulations within realistic porous structures extracted directly from images. Pore-scale fluid dynamics delve into the fundamental physics governing flow, transport, reaction, adsorption, and deformation within heterogeneous porous materials, marking a significant leap towards establishing heterogeneous porous media flow as a standard analytical tool. The applications of this advancement are diverse and far-reaching, encompassing scenarios such as tracking chemical contaminant propagation in underground reservoirs, understanding ink permeation dynamics, modelling sedimentation processes, optimizing hazardous waste storage, and predicting fluid flow behaviours in oil reservoirs and biological tissues. Traditionally, porous media flow was approached through temporally and spatially averaged models, relying on phenomenological and empirically derived equations such as Darcy's law. However, these conventional methods often fell short in capturing the intrinsic complexity of porous media due to the lack of suitable research tools. In this context, we present InPore, a groundbreaking computational platform that employs a kinetic-based volumetric lattice Boltzmann method to solve PSPMFs within image-derived porous structures. InPore stands out for its integrated modelling approach, seamlessly combining image extraction and fluid dynamics simulation, thereby eliminating the need for additional grid or mesh generation steps and simplifying data transfer across software packages. Furthermore, InPore leverages state-of-the-art GPU (Graphic Processing Units) parallel computing technology to enable rapid and localized computations, facilitating high-fidelity simulations. During our presentation, we will showcase InPore's capabilities through application studies and discuss its integration with supplemental mechanisms such as mass/heat transfer, interfacial dynamics, and chemical reactions. These enhancements aim to broaden InPore's functionality for tackling real-world porous media flows, thereby advancing our understanding of intricate phenomena within porous materials.

        Speaker: Prof. Huidan Yu (Purdue University)
      • 120
        Investigating Hydrogen Storage in Pore Media of Saline Aquifers: A Numerical Study on Wettability and Pore Structure Impact

        Hydrogen, as a promising clean energy source, holds significant potential for energy transition and the efficient utilization of clean energy. However, large-scale hydrogen storage poses a limitation to its large-scale utilization and further development. Saline aquifers, characterized by favorable pore space and temperature-pressure conditions, are considered promising candidates for large-scale hydrogen storage. Therefore, our study focuses on investigating the flow of hydrogen in porous sandstone media during the initial injection and extraction process.
        Utilizing the volume of fluid method, we conducted direct numerical simulations of this process, scrutinizing the impact of wettability, capillary number, and pore structure on hydrogen flow, storage capacity, and loss rate. The result reveals that hydrogen flow in underground porous media is predominantly governed by capillary forces, with hydrogen primarily stored in larger pores and channels. Increasing hydrogen wettability enhances reservoir storage capacity but concurrently results in elevated residual hydrogen after the extraction process. Regardless of reservoir wettability, hydrogen losses during the initial injection and extraction process are significant. Reservoirs characterized by larger pore and throat radii exhibit higher effective hydrogen storage capacity. Additionally, reservoirs featuring higher coordination numbers and enhanced connectivity contribute to greater hydrogen storage capacity and improved recovery rates.

        Speaker: Jiawei Li
      • 121
        Is it safe to continue relying on traditional porosity-permeability relationships?

        The precipitation of secondary phases in porous media carries profound implications for the functionality and efficiency of diverse natural and engineered systems. This encompasses applications ranging from subsurface CO2 storage sites, geothermal systems, deep geological disposal repositories, tunnels, oil and gas reservoirs, to the treatment of contaminated groundwater. These precipitation processes alter the structure of porous media, reduce pore space, influence hydrodynamics, and even modify reaction rates by reshaping reactive surfaces. As a result, it becomes crucial to thoroughly investigate the hydrodynamic consequences of mineral precipitation in porous geometries. However, the prevailing practice of assessing the impact of precipitation reactions on flow and transport relies on simplistic permeability-porosity relationships. Commonly employed empirical, experimental, or theoretical models such as Kozeny-Carman, Verma-Pruess, and power law are favored for their convenience and simplicity. These models find widespread application in commercial or open-access simulators for diverse geo-energy and geo-environmental purposes. Nevertheless, our previous research has revealed that relying solely on such porosity-permeability relations introduces significant uncertainty. To address this knowledge gap and mitigate the associated uncertainty, we propose a hierarchical statistical approach to upscale the porosity-permeability relationship from the microscale to the macroscale. Our approach acknowledges the complexity of permeability-porosity evolution while still leveraging practical and readily available formulations. Simulations of the mineral precipitation process in diverse homogeneous and heterogeneous settings were conducted, and a power-law formulation for the porosity-permeability relation was fitted, resulting in a distribution of power-law parameters for each setting. This resulted in a lognormal probability distribution function (PDF) for all the cases. By incorporating this PDF into continuum scale simulations, a fit-for-purpose porosity-permeability relation is established, linking the microscopic dynamics of probabilistic nucleation and growth in porous media with the macroscopic application domain. For most objectives in reactive transport modeling, a three-step scheme adequately captures the pore-scale physics and dynamics, ensuring the representation of these properties at the application scale.

        Speaker: Mohammad Masoudi (University of Oslo, Department of Geosciences)
      • 122
        Machine-Learning-Based Robust Optimization of Brine Extraction Well Placement in CCS Projects Using Fast Marching Method

        In carbon sequestration projects, ensuring the safe management of reservoir pressure is essential for long-term security. The injection of CO2 can lead to pressure build-up, risking safety issues like caprock damage, induced seismicity, and potential leaks. While brine extraction offers a practical solution to mitigate those safety issues, it is crucial to optimize the location of the brine extraction well, especially in heterogeneous reservoirs.

        Optimizing the brine extraction well location is computationally expensive, requiring numerous simulation runs to identify the most effective configuration. To perform robust optimization, which employs multiple reservoir models representing reservoir uncertainties, the computational complexity further increases.

        In this study, we propose a machine-learning-based surrogate model that accurately predicts the effectiveness of the input well location with low computational cost. The proposed model incorporates the fast-marching method (FMM) to calculate the hydraulic connectivity and convolutional neural network (CNN) to extract the features of the connectivity map and predict the net present value (NPV). NPV is used as an objective function that represents the effectiveness of a brine extraction well. We applied this model to a CO2 injection site in the Pohang basin, Korea. Our model showed strong predictive performance, significantly reducing the computational costs by utilizing only 5% of the total location candidates.

        Speaker: Hyunjee Yoon
      • 123
        Nanomechanical properties of Janus nanoparticle-stabilized Pickering emulsion in confined nanochannels

        The crucial role of the interaction between Pickering emulsions and confined nanochannels in their industrial applications is well acknowledged. However, there is a limited understanding of how the modulation of deformation stability and rupture limits of Pickering emulsions occurs when they come into contact with solid walls, particularly in relation to the influence of solid particle shells. This study employs molecular dynamics (MD) simulations to elucidate the nanomechanical properties of Pickering emulsions stabilized by Janus nanoparticles (JNP) in confined channels. For the first time, a comprehensive predictive model is developed to characterize the contact behavior of Pickering emulsions with surfaces exhibiting distinct wettability. The contact stress experienced by an emulsion is found to be dependent on factors such as the equivalent elastic modulus of the emulsion, geometric deformation function, and the influence of the JNP shell along with its interactions. Additionally, it is observed that hydrophobic surfaces induce the rupture of Pickering emulsions under compression. The delay in rupture is achieved by increasing the surface coverage (ϕ) of JNP. Notably, when ϕ reaches a critical value, the JNP shell can assume an ordered quasi-solid structure, leading to a significant enhancement in emulsion stability. These findings have practical implications for the design and screening of specific Pickering emulsions, especially in applications such as enhanced oil recovery, drug or food delivery, and cosmetic ingredient absorption, where the management of deformation and rupture on solid surfaces is crucial.

        Speaker: Dr yuanhao chang (university of regina)
      • 124
        Numerical Simulation of the Microbial Induced Calcite Precipitation (MICP) Process in Darcy-scale and Pore-scale

        Microbial Induced Calcite Precipitation (MICP) technique is a "green" bio-grouting method developed in recent years, which has been applied in many engineering fields. The MICP technique has attracted extensive attention due to its high reinforcement strength and environment-friendly properties. However, MICP reinforcement often faces the problem of non-uniformity precipitation, which happens in different spatial scales and is one of the bottleneck problems restricting the further development of this technology. In this study, the uniformity of calcium carbonate precipitation in the MICP process has been studied numerically in both the Darcy-scale and the pore-scale. The influence of grouting injection strategy, non-uniform distribution of soil material properties, and pore characteristics on the uniformity of MICP reinforcement have been preliminarily investigated.

        Speaker: yajie chu (Tongji University)
      • 125
        Numerical simulation of yttrium oxide grain sintering

        Yttrium oxide is a promising and poorly studied material for the field of catalysis. It can be used as a support in catalytic processes such as carbon dioxide reforming of methane and $CO_2$ methanation. Predicting changes in the texture of $Y_2O_3$ during temperature treatment is an important material science and a computational task.

        In this study, we applied a phase-field approach to obtain an accurate mathematical description of $Y_2O_3$ sintering over a wide temperature range. The general principle of the phase-field method is to describe physical quantities by a set of continuous fields that take constant values in specific regions and smoothly change in the interfaces between these regions. In the case of sintering, such areas are the individual grains of the material. The interface of the microstructure has a finite width along which the sintering materials move. The Allen-Cahn and Cahn-Hilliard equation system is used to describe changes in order parameters and mass density distribution.

        To verify the mathematical model, yttrium oxide sintering experiments were carried out and data on the textural and structural properties of $Y_2O_3$ were obtained. The developed model makes it possible to calculate the decrease in the specific surface area and pore volume of yttrium oxide for pores ranging from 3 to 70 nm and determine the growth rate of $Y_2O_3$ crystallites during sintering. The model allowed us to determine that stepwise heating from 600° C to 900° C and then 1200° C decreases the specific surface area of yttrium oxide from 54 $m^2/g$ to 15 $m^2/g$ and then to 5 $m^2/g$, respectively.

        It should be noted that the obtained experimental micrographs of the cross sections of yttrium oxide samples are in visual accordance with the model images. The approach used in work can be used to predict the evolution of the textural properties of porous materials (catalysts, sorbents, ceramics) under high-temperature conditions.

        The study was carried out with funding provided by Russian Science Foundation grant number 21-71-20003

        Speaker: Dmitry Prokhorov (Sobolev Institute of Mathematics SB RAS)
      • 126
        Optimizing Battery State Estimation: Overcoming Computational Challenges with Hybrid Models

        To ensure the safety and reliability of batteries it is critical to accurately estimate the internal state of the battery which is crucial in Battery Management Systems (BMSs). It is crucial to have methods which, aside from yielding accurate predictions, can be applied for real time estimations. However, the advanced BMSs generating accurate results are computationally intensive and time-consuming, limiting their direct application in real-time estimation. To overcome the computational demand Deep Neural Networks (DNNs) have been applied. However, to have highly accurate models, DNNs with more complex architectures should be applied. The complexity of their architecture will hinder their efficiency for online state estimation algorithms. To tackle the goal of having highly accurate predictions while being computationally efficient, we propose a BiLSTM model as a state estimator, with its hyperparameters automatically optimized using a Bayesian Optimization (BO) framework. We show that leveraging Bayesian inference enables the use of a highly accurate state estimator with a less complex DNN architecture, ensuring computational efficiency.

        Speaker: Dr Serveh Kamrava (Colorado School of Mines)
      • 127
        Relative permeability curve prediction directly from 3D digital rocks based on AI approaches

        The relative permeability curve is one of the key features to evaluate the flow property of a porous medium, which is important in many subsurface engineering problems such as underground energy storage and recovery. Recently, rapid developments in the technology of artificial intelligence (AI) have offered new views to revisit the acquisition of relative permeabilities. Here, we present our systematic work on the developments of AI models for the predictions of relative permeability curves directly from 3D digital rock images. The training and testing data are generated from pore-network simulations and core-flood experiments. It avoids the use of indirect geometrical parameters as inputs in previous AI methods. It is able to cover 3D digital rocks with variable sizes and further equiped to have the upscaling capability. The results show that the AI models have high prediction accuracies over 95%, with scale information being the most important physics feature accounting for 51%, and the upscaling prediction of relative permeability curve is in good agreement with macroscopic experiment data. The new framework is also flexible and can be easily extended for the prediction of other rock physical properties according to practical demands.

        Speaker: Jingwei Zhu
      • 128
        Remote hydraulic fracturing at weak interfaces

        As the most effective reservoir stimulation technique, hydraulic fracturing has been applied since the 1950s. At the same time, hydraulic fracturing can induce seismicity or result in the loss of containment of subsurface fluids due to the high injection pressure applied during its operation, leading some projects to eventual shut-down. To mitigate such adverse impacts, an alternative approach known as hydro-shearing has been promoted for some enhanced geothermal system projects, wherein the injection pressure is kept at a low level, aiming to stimulate pre-existing networks of fractures by shearing. However, the practical effectiveness of hydro shearing is yet to be proven. In this talk, we propose another alternative stimulation approach using a low-viscosity fluid. We numerically demonstrate that with low-viscosity fluid injection, we can fracture discontinuous interfaces such as grain boundaries or natural fractures without initiating fractures at the injection point. Our results indicate the possibility of engineering reservoir stimulation operations without applying high injection pressure.

        Speaker: Tao You (Montanuniversität Leoben)
      • 129
        RepoTREND: Software Tools for Robust Safety Analysis of Radioactive Waste Repositories

        RepoTREND [1], [2] is a novel simulator that has been designed to emulate the processes that occur within a radioactive waste repository in a variety of geological formations. It provides robust functionality to simulate the release and migration of contaminants from the near-field through the geosphere to the biosphere, while estimating their radiological impact on human health and the environment.

        Designed with modularity in mind, RepoTREND consists of computational modules tailored to simulate the processes within each subsystem of a repository. The inherent heterogeneity of typical repository models poses significant challenges. In addition to fundamental processes such as two-phase contaminant transport, numerous specific effects (such as container corrosion or rock convergence) have to be considered in the simulation.

        The structure of RepoTREND has been designed to meet a number of challenges. These include the flexible selection of models for different regions, the seamless combination of models during simulations and the easy integration of new models and effects. The RepoTREND code is a framework for the solution of a general nonlinear system of equations. Different physics are implemented as models in library form.

        Each model is defined by specific equations of state and routines for handling relevant effects, organised in libraries of equations and effects. This structure simplifies the integration of new equations and effects, and allows different models to be assigned to different grid blocks. The coupling of physical models is managed implicitly. This facilitates the solution of linear couplings between variables across grid blocks within the same matrix system.

        This conceptual approach ensures ease of implementation for new effects. It also maintains flexibility, transparency and reusability as the code is extended and refined.

        References:
        [1] Reiche, T.: RepoTREND Das Programmpaket zur integrierten Langzeit-sicherheitsanalyse von Endlagersystemen, GRS-413
        [2] https://www.grs.de/en/news/projects/repotrend-repository-safety-analysis

        Speaker: Tatiana Reiche
      • 130
        Rock-Fluid Interaction Mechanisms between Binary Surfactants Systems for Enhanced Oil Recovery in a Carbonate Formation

        Carbonate rocks exhibit a complex surface charge, making it challenging to generalize the use of a single surfactant type. Hence, the utilization of binary surfactant mixtures is proposed as a more efficient alternative. This work focuses on static adsorption, wettability alteration, and spontaneous imbibition tests to gain comprehensive insights into the underlying fluid-rock interactions in carbonate formations. The objective is to propose more effective solutions for enhanced oil recovery in carbonate formations. Our study centered on binary surfactant systems and their interactions with carbonate rock. We conducted several laboratory experiments, including static adsorption tests on eight different surfactant systems. This aimed to compare their adsorption behaviors against individual surfactants, with the aim of studying their synergistic interactions. Additionally, wettability and spontaneous imbibition tests were conducted under the reservoir conditions of a producing oil field to understand the primary mechanisms and synergistic effects of binary surfactant systems in enhancing oil recovery from carbonate formations. Our results showed a significant influence of the nonionic surfactant leading a considerable reduction in adsorption values of 53% and 28% in its anionic-nonionic and cationic-nonionic mixtures, respectively. The efficient synergism between binary surfactant systems to reduce surfactant adsorption in carbonate rocks was also confirmed in the physicochemical evaluations with a reduction in both zeta potential and pH values when compared to their individual surfactants. Furthermore, spontaneous imbibition results showed that binary surfactant mixtures exhibit maximum synergism, particularly when they system is composed of of zwitterionic and non-ionic surfactants. This surfactant blend resulted in the highest recovery factor of nearly 60%, signifying significant improvement in oil recovery from carbonate formations. According to the analysis of contact angle, the binary surfactant systems did not significantly change wettability. However, this can be beneficial because it implies that the surfactant molecules are not adsorbed to the rock surface within the porous medium. Instead, they are utilized to their maximum potential within the porous medium. The findings presented in this work demonstrate that careful screening, selection, and combination of binary surfactants can effectively reduce surfactant adsorption, maintain rock wettability, and substantially lower interfacial tension in carbonate rock, ultimately aiming to enhance oil recovery. This approach paves the way for the development of innovative surfactant blends that ensure the economic viability of EOR projects, suitability for CO2 foam sequestration projects, and broad applicability in carbonate formations.

        Speaker: Ayomikun Bello (Skolkovo Institute of Science and Technoloy)
      • 131
        Simulation of the Microscopic Three-Phase Flow Process in CO2 Miscible Flooding at the Pore Scale

        CO2 flooding after water flooding can effectively improve the recovery efficiency of low-permeability reservoirs. At present, the seepage law of CO2 flooding after water flooding is generally determined through indoor core experiments and macroscopic numerical simulation methods, and simulations of the seepage process at the microscopic pore scale are lacking. Among the existing microscopic numerical simulation methods, two-phase flow simulation is generally the main focus, and multiphase flow simulation under the conditions of three-phase coexistence of oil, gas, and water is lacking. In view of the above problems, this paper conducts a microscopic numerical simulation of the CO2 flooding seepage process after water flooding based on a two-dimensional heterogeneous pore model of circular media and studies the effects of interface tension and injection velocity on the three-phase seepage process, gas breakthrough time, and gas recovery degree during the multistage miscible process. The research shows that when the interfacial tension between CO2 and oil is high, CO2 pushes water and oil forward in a piston-like manner and penetrates the water layer to contact the oil, which ultimately causes the continuous water phase to separate from the gas phase and form the main flow line of the continuous gas phase. With decreasing interfacial tension between CO2 and oil, i.e., closer to the miscible state, the gas diffuses into the water after injection and accumulates at the water-oil interface, the crude oil is displaced toward the production end, and the flow speed of CO2 is faster than that of the water phase. The lower the interfacial tension is, the shorter the gas breakthrough time at the outlet after CO2 injection. Before the miscible state of CO2 and oil, the lower the interfacial tension is, the earlier the gas channeling time, and the lower the recovery degree. After mixing, a turning point occurs. An increase in the injection velocity will advance the gas breakthrough time and gas channeling time at the outlet, leading to an increase in the gas recovery degree. This study has reference and guiding significance for understanding the three-phase flow characteristics of oil, water, and gas during CO2 miscible flooding in mines.

        Speaker: Mr Jing Li (China University of Petroleum (East China))
      • 132
        Simulation study of hydrogen storage in a depleted gas reservoir: Microbiological influences in porous media

        Storing hydrogen in depleted gas reservoirs is a viable method for balancing seasonal energy demand fluctuations. However, these reservoirs harbor a diverse population of microorganisms. H2 are considered one of the most important electron donors for subsurface microbial respiration. Under high salinity, high temperature, and high pressure conditions, microbial reactions such as methane generation, sulfate reduction, and acetate production are most common [1]. These reactions result in hydrogen loss, gas acidification, pore plugging by metabolic biofilms, and alteration of the hydrogen-brine-rock three-phase interface properties due to the generation of organic acids [2]. Currently, there is very little research on the impact of microorganisms in depleted gas reservoirs.
        This study was conducted on a specific depleted gas reservoir, utilizing the CMG-STARS to simulate the impact of microorganisms on hydrogen storage. Firstly, the diffusive distribution of solid-phase microorganisms (biofilms) on porous media was designed, and Fick's law was employed to characterize the concentration-driven microbial transport process. Subsequently, based on reaction conditions, four reactions were designed to generate CH4 (PH>7), H2S, acetic acid (PH<7), and microbial growth. The microbial population within the community was considered to control the rates of hydrogen uptake and microbial growth. The shedding of biofilms was influenced by the number of microorganisms and shear rate. Multiple sets of relative permeability curves were designed to match changes in wetting angle caused by acetic acid generation. Finally, the injection pressure was limited by reservoir fracture pressure and capillary forces causing leakage to the overlying formation. Seasonal hydrogen storage was conducted over four cycles, with a cycle consisting of 6 months of injection and 6 months of production.
        The simulation results revealed the presence of high microbial saturation zones in near-wellbore region and higher parts of structure. The generated CH4 and H2S account for a maximum of 1.4% and 0.1% of the injected hydrogen volume, respectively, and accumulated below the H2 layer. The loss of hydrogen gas was highest at 5% in first cycle and decreased to a minimum of 0.6% in third cycle. As cycling period increased, the purity of hydrogen in produced gas became higher. Throughout the entire process, the effective porosity of the gas reservoir decreased by a value ranging from 0.1% to 0.5%, while the pH remained relatively unchanged.

        Speaker: Zanfu Xiong (China University of Petroleum (East China))
      • 133
        The Crushing Characters of Quartz Sand Based on a New Experimental Image Processing Methods

        Hydraulic fracturing has gradually increased as an important means to enhance production in deep low permeability reservoirs. Whether proppant can maintain long-term high conductivity in fractures has become a hotspot. Quartz sand has been widely used due to its affordability and easy preparation. However, the strength of quartz sand is low. Under high closure stress, quartz sand is crushed, producing fine particles that reduce fracture width and permeability of sand piles, resulting in a rapid decrease in fracture conductivity. Therefore, it is crucial to study the fragmentation law of proppants under high closure stress. For the study of the fragmentation law of quartz sand, a lot of related experiments have been conducted both domestically and internationally, and quantitative methods such as screening method and laser particle size analysis have also been formed, but there are certain limitations. In this study, a new image processing-based quantitative method is developed to determine the compression proppant crushing rate. Compared with conventional screening methods to verify accuracy, this method can more quickly and efficiently quantify the proppant crushing rate. Subsequently, a crushing experiment of quartz sand was conducted using the proposed method, and the proppant crushing rate under different conditions was calculated. The influence of factors such as sand spreading concentration, particle size combination, and sand placement method on the proppant crushing rate was analyzed. The results of this analysis were consistent with previous studies, confirming the applicability of the proposed method. This research provides a theoretical foundation for hydraulic fracturing and optimization of sand placement in order to maintain long-term high conductivity in fractures.

        Speaker: liansong Wu (Southwest Petroleum University)
      • 134
        The Mechanism and Quantification of Threshold Pressure for Oil Flow in Silica Nanochannel by Molecular Simulation

        Shale oil is an abundant unconventional resource in the world. However, due to the highly heterogeneous of shale reservoirs, the shale oil is difficult to flow in the porous media, resulting in the uncertainty efficiency of industrial exploration for shale oil. Threshold pressure gradient is the key property to characterize the mobility of shale oil. Usually, shale rock has abundance of pores in nanometer, and the structure is more tight than other reservoirs. It is challenging to study threshold pressure gradient for shale oil by traditional theoretical and laboratory methods in low-permeability reservoirs. In view of this, our work will quantify the threshold pressure gradient and reveal the mechanism of threshold pressure for fluids flow at nanoscale by molecular simulation. In our work, the threshold pressure of oil flow in silica nanochannel was quantified firstly by “pressure - velocity” method. The “pressure-velocity” shows the oil flow would be divided into three stages, via vibrate stage, initial stage and flow stage, showing non-Darcy flow. After that, the size of silica nanopore with 2nm, 4nm, 6nm, 8nm and 10nm, was constructed to study the size effect of threshold pressure. The calculated threshold pressure indicates that the critical pressure for oil to flow in nanopores would exponential increased with the decrease of height. When the height of nanochannel increases to 10nm, the threshold pressure was very small. Also, the height of silica nanochannel would influence the fluids properties, such as density, viscosity, etc. Due to the decrease of nanochannel, the proportion of adsorbed oil molecules in nanochannel would increase, thus influencing the viscosity and critical pressure to flow. Meanwhile, with the decreased nanochannel, the interaction between shale oil and nanochannel would increase, and more pressure are needed to make the shale oil flow. There is a highly correlation between threshold pressure and interaction energy. Intrinsically the critical pressure of oil is the energy needed to disturb or force the adsorbed oil film flow in nanochannel. Furthermore, the threshold pressure for shale oil is also related with rock types, such as quartzite, carbonate and kaolinite, etc, especially for oil in very small nanochannel. Our work provides a new insight into the shale oil flow in porous media, which is meaning for exploration of shale oil resources.

        Speaker: BingBing Liu (China University of Petroleum (East China))
      • 135
        Two-Phase Flow Displacement Morphologies in Cohesive Granular Media

        Multiphase flow in granular materials is intricate and subject to pattern formation resulting from the interplay between hydrodynamic and mechanical forces. While considerable effort has been devoted to studying systems with cohesionless grains, our understanding of the two-phase flow behavior through the cohesive counterpart held together by intergranular bonds is limited. Herein, we study the novel coupling between viscously unstable fluid-fluid displacement and bonded-grain deformation in a synthetic cohesive granular pack. We experimentally inject a low-viscosity fluid into a monolayer of cohesion-tuneable bonded glass beads that are initially saturated with a more-viscous fluid, with injection capillary numbers and cohesion levels varying among experiments. We map out a first-ever phase diagram showing displacement patterns transitioning from deformation-dominated fracturing with bond breakage to infiltration-dominated viscous fingering without grain motion as cohesion increases. Strikingly, we find that peak injection pressure exhibits a non-monotonic trend with increasing cohesiveness. The injection pressure reaches maximum when fracturing is favored against infiltration while begins to drop due to the regime transition irrespective of the increasing cohesion. Furthermore, we characterize the onset of fracturing via dimensional analysis, effectively capturing the transition based on the balance between viscous and cohesive forces. Our findings shed light on the multiphase flow behavior within cohesive materials which is fundamental in various subsurface technologies such as carbon geostorage, oil/gas recovery, and groundwater remediation.

        Speaker: Mr Feihu KE (The University of Hong Kong)
      • 136
        Upscaled model for steady slip flow fluid structure coupling in shale system

        Shale is a highly heterogeneous porous material rich in organic matter. Injecting fluid into a porous material can expand the pore space, distorting the solid skeleton. The detailed flow and mechanics of this solid deformation has not yet been systematically investigated. This work reports on modelling steady liquid flow in shale system, considering the slip effect and fluid-structure wall deformation. A microscale (pore level) fluid structure interaction (FSI) problem is formulated in terms of incompressible Newtonian fluid and a linearized elastic solid. The slip effect is adopted means of a Navier-type boundary condition. Combining different mechanical properties of organic and inorganic matter, an asymptotic solution to the FSI problem is derived for a certain geometry. A nonlinear Darcy-type upscaled equation for the averaged pressure is obtained, as well as introducing an apparent permeability dependent on interface position and slip coefficient. Based on the obtained results, relevant results for more general situations are obtained through extended analysis. The accuracy of the result is assessed by comparisons with numerical simulations. Our results may be useful for a better understanding of shale oil rocks at the micrometer scale, studying the large squeezing deformation of carbonaceous shale in practical situation or studying the deformation of other porous media.

        Speaker: Dongyan Fan (China University of Petroleum (East China))
      • 137
        Water Impact on Adsorbed Oil Detachment from Mineral Surfaces by Supercritical CO2

        Geochemical reactions are crucial for in-situ CO2 mineralization underground associated with CO2-enhanced oil recovery (CO2-EOR) in a hydrocarbon reservoir. However, the presence of formation water and adsorbed oil on rocks generates physical barriers to CO2’s access to mineral surfaces, which may yield impedance to CO2 mineral trapping that has yet to be accounted for. In this study, we mimic the dynamic oil detachment process using molecular dynamic (MD) simulations and analyzed the influence of an adsorbed oil film on supercritical CO2 (scCO2) diffusion towards the mineral surface in the presence and absence of a water phase. CO2-oil-water-rock reaction experiments are performed to substantiate the simulated data. Our results demonstrate a negative impact of water on oil film detachment by scCO2, which may give rise to a substantial delay in mineral reactions or even impede their occurrence and is unfavorable for mineralized CO2 storage underground. Carbonated water, regardless of whether it is saturated, showcases the same inhibitory effect on the miscibility of scCO2 and oil, thereby restraining oil film detachment and the contact of CO2 with the rock surface.

        Speaker: Ms Rui Gao (China University of Petroleum (Beijing)))
      • 138
        Wave Velocity Dispersion and Attenuation in Partially Saturated Porous Media

        The understanding of the seismic signature of the partially saturated formation is critical to seismic monitoring in the hydrogen geo-storage, CO2 geo-sequestration and geophysical survey and exploration of oil and gas reservoir. The main objective of this study is to model the wave propagation in partially saturated rocks containing two immiscible fluids (i.e., gas-water), with a comparative case study on hydrogen (H2), methane (CH4), nitrogen (N2) and carbon dioxide (CO2) bearing rocks. The sonic velocities and the attenuations are influenced by several parameters, which interact in a complex pattern, particularly when the rock is saturated with multiple fluids. We developed a rock physics model that considers the effects of patchy saturation, wettability, effective pressure, and relative permeability. By examining wave propagation in each fluid-saturated case against water saturation, we improve our understanding of changes in sonic velocity and attenuation during the water saturation varies. This provides valuable insights for seismic and sonic monitoring during the injection and extraction of gas in the reservoir formation.

        Speaker: Dr Jimmy Xuekai Li (The University of Queensland)
    • MS01: 1.3
      • 139
        Solar energy storage in saline aquifers: Insights from coupled hydro-thermo-mechanical modeling

        The utilization of saline aquifer for solar energy storage is recognized as a promising solution to address the spatial and temporal mismatch between energy demand and supply. This approach holds significant potential for future renewable energy storage and conversion. Thermal energy storage in saline aquifer can effectively transform intermittent solar energy into stable geothermal energy at high temperatures. In this study, we examine a previously proposed solar energy storage and conversion system. This system entails the initial conversion of solar energy into heat through parabolic troughs, followed by the storage of thermal energy in a saline aquifer facilitated by high-temperature hot water circulation. Currently, the impact of poroelasticity and thermal stress induced by high-temperature hot water injection on injectivity and heat storage efficiency remains unclear. In this study, three-dimensional porosity and permeability fields for typical saline aquifers are generated by geostatistical modelling. The circulation of high-temperature hot water in the aquifer by doublet vertical well system is explored through coupled hydro-thermo-mechanical modeling. We analyze the effects of poroelasticity and thermal stress on the injectivity of hot water. The influences of in-situ stress and permeability heterogeneity on spatial and temporal evolution of hot water zone are then explored. In addition, the efficiency of solar energy storage in various heterogeneous saline aquifer is evaluated and compared. Considering the hydro-thermo-mechanical coupling effects, saline aquifers conducive to solar energy storage are identified. This study enhances our understanding of the mechanisms involved in solar energy storage in saline aquifers, providing crucial insights for its practical implementation.

        Speaker: Mr Kunpeng Zhong (Chengdu University of Technology)
      • 140
        Towards an open-source digital twin for subsurface geothermal systems: a proof-of-concept study for a doublet system

        Geothermal energy can provide clean and sustainable baseload energy for heating, cooling, and power. The injected working fluid undergoes flow and heat transfer within the surrounding porous rocks during the geothermal reservoir extraction. The geological complexity and lack of data require digital technology like computer simulation to assist in the optimization and decision-making of operation strategies. A digital twin denotes a virtual representation of a physical product, process or facility, and is used to understand and predict the physical counterpart’s performance. A digital twin for geothermal production can help to mitigate operational risks, reduce maintenance costs, extend reservoir longevity, and enhance overall sustainability of a geothermal resource.
        We propose a workflow for an open-source digital twin for geothermal energy that contains the following elements: a) Well logs and seismic data are utilized to design multiple reservoir models that capture possible geological scenarios using the Rapid Reservoir Modeling (RRM) software. RRM is a sketch-based modelling software that allows users to rapidly sketch geologically consistent models in 3D. b) Possible property distributions will be assigned to geological domains to capture uncertainty in the petrophysical data. c) The Delft Advanced Research Terra Simulator (DARTS) is combined with machine learning techniques to create proxy models that enable fast simulations. d) As new production and monitoring data becomes available, data assimilation techniques like Ensemble Smoother with Multiple Data Assimilation (ESMDA) are applied to update property distributions for each scenario. This iterative process of data assimilation will help users constrain geological and production uncertainties, both of which are key to optimizing operational strategies.
        We demonstrate the digital twin framework using a proof-of-concept study of a low-enthalpy geothermal system located in a channelized fluvial reservoir. Heat is produced from a geothermal doublet. The geological scenarios were designed using RRM. These models consider key uncertainties, such as Net to Gross, sinuosity of the channels, paleo flow direction, and the distribution of porosity and permeability within the geological domains. One of the RRM models was chosen to be the “truth case” for which synthetic production data (well temperatures and pressures) and dynamic data along the wells were simulated using DARTS. Each individual RRM model adheres to the well constraints “observed” for the truth model. DARTS was applied to rapidly predict production performance for each scenario, and these data can subsequently serve as the training set to obtain the proxy model. Both machine learning and ESMDA will be performed to reduce the difference between the prediction and observation of truth case to update reservoir properties.
        The outcomes of this proof-of-concept study demonstrate the feasibility of the digital twin framework for geothermal systems. A broader range of monitoring data in the reservoir and transient data will be included in the future to enhance the performance of the digital twin in geothermal energy applications.

        Speaker: Dr Guofeng Song (Delft University of Technology)
      • 141
        Insight on the stability of gas hydrate in montmorillonite slits by molecular dynamics simulations

        The extent of interactions between clay surfaces and water molecules and their impact on hydrate stability in clay reservoirs have been a source of debate. This uncertainty arises from the inherent challenges associated with the nanoscale temporal and spatial detection of bound water molecule distribution characteristics. This study employs molecular dynamics simulations to investigate the stability of methane hydrates in montmorillonite slits at various temperatures, focusing on the surface influence scale, bound water molecule distribution characteristics, and binding strength. The results show that hydrates in close proximity to the clay surface exhibit lower stability and are more prone to decomposition. The hydrophilic nature of the surface leads to water molecule aggregation at the interface, driving methane molecules away during decomposition. Additionally, compared to the charged tetrahedral layer surface of montmorillonite, the quasi-liquid layer on the neutral tetrahedral layer surface is thinner, with semicage structures persisting within the vacancies of the Si-O rings. The analysis suggests that variations in the range of surface influence and binding strength can be primarily attributed to intermolecular Coulomb interactions and charge redistribution at the interface. These research findings offer valuable molecular insights into the microscopic characteristics and behavior of hydrates within clay slits.

        Speaker: Dr Jie Chen (China University of Petroleum (East China))
      • 142
        Microfluidic Study of Formation, Dissociation, and Dissolution Dynamics of Gas Hydrates in Porous Media

        Gas hydrates are crystalline solids in which guest molecules are trapped within cages formed by water molecules at high pressure and low temperature. These solids have important applications in natural gas hydrate exploration, CO2 or H2 storage, water desalination, gas separation, and gas/oil transportation. Natural methane (CH4) hydrates are abundant in the seabed sediments and are potential sources for future energy harvesting[1]. On the other hand, carbon dioxide (CO2) hydrates are promising forms of CO2 sequestration due to the large storage capacity. The phase transition of hydrates and the transport behaviors of the relevant gas and liquid phases in porous media are crucial to CH4 production and CO2 storage using hydrates. Therefore, many studies have been conducted on investigating the dynamics of formation, and dissociation of CH4 and CO2 hydrates in porous media using sand or glass bead packs with the help of in-situ imaging methods such as X-ray synchrotron tomography and magnetic resonance imaging (MRI) technology. In this work, we developed a low-temperature and high-pressure microfluidic system for gas hydrate study, allowing for in-situ imaging of the phase transition of hydrates under realistic reservoir conditions of deep seabeds. We studied the formation, dissociation, and dissolution mechanisms of CH4 and CO2 hydrates in both pore scale and chip scale. The hydrates were generated in pure water at 10 MPa and 5 oC subcooling temperature. The dissociation of these two hydrates was induced by decreasing pressure and increasing temperature, respectively. During hydrate formation, we observed the nucleation and propagation of hydrates from the gas-liquid interfaces into the bulk gas, showing various morphologies at the pore scale. The growing kinetics were calculated by analyzing the optical images obtained by a high-resolution camera. Further, we successfully captured the crustal fingering of CH4 gas encased by CH4 hydrates due to the local pressure gradient [2]. From the chip scale, the location of hydrate formation and its propagation in the porous media is stochastic. We found that the induction time for hydrate formation is also stochastic, and the nucleation of hydrate should be triggered by external stimuli such as flow and pressure[3]. During hydrate dissociation, the hydrates remained stable until the pressure or temperature exceeded equilibrium. Then, a drastic transition of hydrates into gases occurs, which results in the fast displacement of gas with liquid in the porous media. In addition, the reformation of hydrates was observed during hydrate dissociation. Finally, we studied the dissolution of CH4 hydrate in undersaturated water and revealed the formation, dissolution, exsolution, and reformation mechanisms of gas hydrates in porous media.

        Speaker: Wei Yu (King Fahd University of Petroleum and Minerals)
    • MS03: 1.3
      • 143
        An Efficient Numerical simulation of Reactive Flow in Fractured Vuggy Carbonate Reservoirs Considering Hydro-Mechanical coupling effects

        Fractured vuggy carbonate reservoirs are one of the most important reserves in the world, which hold great importance for increasing reserves and production. However, fractured vuggy reservoir has greatly different reservoir space and flow patterns challenging low recovery. Three types of reservoir space, including matrix pores, fractures, and vugs, coexist with strong heterogeneity, and the spatial distribution scale varies from millimeter to the meter. Acidizing is a vital stimulation technique to boost production in deep fractured vuggy carbonate reservoirs since it can effectively enhance the connectivity of fractures and vugs. The real-time dynamic alterations in the volume of matrix pores, fractures, and vugs during acidification, coupled with changes in reservoir in-situ stress, signifies a multi-field coupled problem. Currently, research on hydrological-mechanical coupling processes throughout reactive flow in porous media is restricted to single-pore and fracture models, with little consideration given to the influence of pore, fracture, and vug deformation on reactive flow. This paper puts forth a set of mathematical models and numerical simulation techniques for analyzing reactive flow in fractured vuggy carbonate reservoirs while accounting for hydro-mechanical coupling effects. Validation of the model and method is achieved through a numerical example. The results show that fractures and vugs are leading in acid flow through the medium during the acidification of fractured vuggy media. Under stress conditions, fracture closure exhibits the most substantial impact on acid flow in the fracture, followed by vug deformation. Acid fluid preferentially flows via dominant channels connected by fractures and vugs, dissolving the rock.

        Speaker: Kang Liu (中国石油大学(华东))
      • 144
        Characteristics of CO2 Dissolution in Fractured Saline Aquifers

        Dissolution of CO2 in brine, one of the key mechanisms for securely storing CO2 in the subsurface, involves diffusion of CO2 into the brine and subsequent buoyancy-driven convective migration. The stable stratification along the CO2-brine interface, predominated by diffusion, stimulates the density-driven convection, resulting in an enhanced dissolution rate. In fractured porous media, the complex geometry increases the uncertainty of CO2-plume migration. Predicting the characteristics of CO2 dissolution into the resident brine in fractured saline aquifers is important to understand the potential for long-term storage. In this work, a discrete fracture-matrix model (DFM), where fractures are explicitly characterized in the model with individual grid cells, is adopted to describe the geometry of the fractured saline aquifer. We first carry out a sensitivity study to obtain a reasonable resolution of grid discretization which can capture both the fast convective flow and converged dissolution rate with fractures. Based on the selected resolution, the properties of fractures, e.g., the permeability and aperture of fractures, are investigated to confirm their impacts on density-driven convection within an individual-fracture model. In addition, an aquifer containing a complicated fracture network that includes highly intersected/dead-end fractures is used to highlight the effects of fractures on the interactions between gravity currents and convective dissolution. Our simulation results demonstrate that due to a low porosity/permeability matrix, CO2 dissolution is driven by diffusive/convection transport along the CO2-brine interface, while the density-driven convection is very weak, i.e., relatively long onset time with a small Rayleigh number. The directions of fractures play a critical role in the convective behavior of CO2-enriched brine. The fracture network enhances CO2 dissolution compared to the case of an aquifer containing isolated fractures, i.e., no connections among fractures. These estimates of the dissolution rate with fractures show that the geometry of fractures plays an important role in enhancing storage security.

        Speaker: Dr Xiaocong Lyu (China University of Petroleum (Beijing))
      • 145
        A new model for predicting conductivity under nonlinear fracture closure and proppant crushing grading curve evolution

        To solve the problem of rapid decline in conductivity of sand filled fractures in deep shale, based on the mechanical process of compression and deformation of proppant pile, the constitutive equation of deformation and fracture of proppant pile was established, and the transfer matrix of gradation curve after fracture of proppant pile was established based on fractal theory. Combined with KC equation, the equivalent particle size and permeability evolution equation of sand pile under multi-particle fracture were deduced. Finally, a new model for predicting the conductivity of proppant embedment deformation and fracture is obtained. The new model is compared with the results of fracture width deformation experiment, sand pile permeability experiment and multi-particle size combination conductivity experiment, which proves the correctness of the model. The results show that the change of net closing pressure will change the fracture width and permeability, and thus change the fracture conductivity. The dominant factor of the change of conductivity is the net closing pressure, and the greater the net closing pressure, the smaller the conductivity. The main factor affecting the change rate of fracture width is the apparent elastic modulus of proppant pile. The larger the apparent elastic modulus is, the larger the fracture width is. The main factor affecting the change rate of fracture permeability is the fracture degree of proppant pile. The larger the fracture degree is, the smaller the permeability is. The larger the combination ratio of large particle size, the higher the conductivity. Therefore, the development of proppants with high apparent elastic modulus and low degree of breakage is of great importance to improve the conductivity. At the same time, this model also improves the theoretical guidance for proppant particle size combination selection, which is helpful to the optimization design of field construction.

        Speaker: liansong Wu (西南石油大学)
      • 146
        Impact of corner-bridge flow on capillary pressure curve

        The capillary pressure curve is essential for predicting multiphase flow processes in geological systems. At low saturations, wetting films form and become important, but how wetting films control this curve remains inadequately understood. In this study, we combine microfluidic experiments with pore-network modeling to investigate the impact of corner-bridge flow on the capillary pressure curve in porous media. Using a CMOS camera and a confocal laser scanning microscopy, we directly observe the corner-bridge flow under quasi-static drainage displacement, revealing that corner-bridge flow serves as an additional flow path to drain trapped water. Consequently, the capillary pressure curve shifts towards lower saturations, resulting in a reduced water residual saturation. We establish a theoretical criterion for the occurrence of corner-bridge flow and develop a pore-network model to simulate quasi-static drainage, taking into account this additional flow path. Pore-network modeling results agree well with our experimental observation. On this basis, we employ our pore-network model to systematically analyze the impact of corner-bridge flow on capillary pressure curve across varying porosity, pore-scale disorder, and system size. Results indicate that the impact of corner-bridge flow becomes more pronounced as porosity decreases and shape factor increases. Our findings demonstrate that the maximum decrease of water residual saturation is 0.19 when porosity is at its minimum, and the shape factor is at its maximum. This work bridges the gap between the pore-scale mechanism and capillary pressure behavior and has significant implications for estimating the amount of extractable water and the CO2 storage capacity.

        Speaker: Dr Tian Lan (Wuhan university)
    • MS06-B: 1.3
      • 147
        Interface Evolution During Pore Water Evaporation in Micromodels

        The interface between liquid and vapor phases within porous media plays a pivotal role in enhanced vapor diffusion and water evaporation. The liquid-vapor interfaces can be classified into internal interfaces (pertaining to vapor diffusion) and external interfaces (associated with phase change) based on their distinct mechanisms. However, the intricate geometric and topological complexities within these interfaces pose challenges in discerning between their internal and external manifestations, hindering a comprehensive understanding of heat and mass transfer mechanisms within soil pores. In this study, a meticulously engineered Hele-Shaw cell integrated with patterned micropillars offers an innovative approach for comparing interface evolution in pores with diverse patterns. A sophisticated image-processing analysis was employed to accurately compute the evaporation rate of the pore water from the micromodels. Equivalent lengths were obtained to determine interface areas at distinct time intervals. Comparison of the continuous recession of the liquid-vapor interface in a stable micropillar pattern to the air-entry of the internal interface and pinned external interface in an unstable micropillar pattern provides an approach to quantitatively separate the internal and external interfacial evaporation rates. Furthermore, employing image processing during evaporation enables the calculation of both global and local interface curvatures. Consequently, a correlation between the characteristic curvature radius and the averaged interfacial evaporation rate was established, consistent with prior experimental findings documented in the literature. The distinction between internal and external evaporation rates offers a fresh perspective, shedding light on the mechanisms that drive enhanced evaporation and diffusion within porous media.

        Speaker: Yi Dong
      • 148
        Evaporation in porous media with salt precipitation

        Visualization experiments are performed to disclose the salt precipitation and gas-liquid displacement in microfluidic pore networks during evaporation. Two forms of salt precipitation are revealed: aggregated polycrystalline structures and large bulk crystals. It is found that gas bubbles can be formed because of imbibition of liquid into aggregated polycrystalline structures. The length of a corner liquid film can affect the direction of growth of the aggregated polycrystalline structures connected to the corner liquid film. Discontinuous corner liquid films can be transformed to continuous ones when they are touched by growing aggregated polycrystalline structures. The "sleeping" aggregated polycrystalline structures at the open surface of a microfluidic pore network, i.e. efflorescence, can grow again if they are touched by growing aggregated polycrystalline structures inside the microfluidic pore network, i.e. subflorescence. Because of efflorescence, the evaporation rate from a microfluidic pore network can increase first and then decrease. In addtion, the distribution of the precipitated salts also depends on the thermal gradient along the microfluid pore networks.

        Speaker: Dr Rui Wu (Shanghai Jiao Tong University)
      • 149
        Interfacial tension reduction mechanism by nanoparticles at heavy oil-carbonized water interface from molecular dynamics approaches

        Nanoparticle-enhanced carbonated water (NP-enhanced CW) is a novel and promising injection agent for coupled enhanced heavy oil (HO) recovery and CO2 storage. The main objective of this study is to investigate the interfacial tension reduction mechanism by nanoparticles (NPs) at HO-CW interface from molecular dynamics approaches. The influences of NPs on the Interfacial tension of HO-CW systems under reservoir conditions were studied. In addition, the influences of NP type, NP concentration, CO2 concentration, pressure, and temperature were investigated and the enhanced oil recovery mechanisms of nanoparticle-enhanced CW were also discussed. The results revealed that the ability of the five NPs to reduce the IFT of the HO-CW systems was as follows: SiO2 NPs > Al2O3 NPs > TiO2 NPs > Fe NPs > CuO NPs. The interfacial tension value of the HO-CW system decreased by 39.69 % due to the presence of SiO2 NPs. An optimal NP concentration existed to decrease the interfacial tension of the HO-CW systems. The interfacial tension values of the HO-CW systems in the presence of SiO2 NPs decreased with increasing the CO2 concentration, temperature, and pressure. This study is helpful in deeply understanding the microscopic mechanisms of NPs affecting the interfacial tension of HO-CW systems during the injection processes of NP-enhanced CW. The results also provide a valuable reference for the injection of the NP-enhanced CW as a new technology to enhance HO recovery and underground storage of CO2.

        Speaker: Xiaofei Sun (China University of Petroleum (East China))
      • 150
        Thermodynamics and Morphology of Ganglia in 2D Heterogeneous Porous Media

        Ganglia (bubbles, or droplets) are widespread in porous media of various industrial applications. Thermodynamic properties of a ganglion, including its volume ($V$), surface free energy ($F$), and capillary pressure ($P_{c}$), play pivotal roles in determining its transport and reactive performance. Although these properties in homogeneous porous media have been recently resolved [1, 2], quantitatively description of ganglia in heterogeneous media remains a challenge [3-5].

        In this study, we develop a pore-scale algorithm for determining the morphologies and thermodynamic properties of hydrostatic ganglia in heterogeneous porous media (a 2D pillar array, as illustrated in Figure 1a). Notably, we reveal novel ganglion morphologies: the fluid-fluid interface can emerge between non-adjacent solid particles that do not share a pore unit (referred to as the “cross-pore interface”), although it has long been assumed that a pore is a basic unit of fluid and interface behaviors in porous media [2]. The presence of cross-pore interfaces is strongly associated with the pore-throat ratio: a smaller pore-throat ratio (wider throat) leads to a greater number of metastable morphologies. Interestingly, these novel cross-pore interfaces can also be found in homogeneous media.

        We track cycles of quasi-static growth and shrinkage of a ganglion (Figure 1b) and resolve the corresponding thermodynamic properties’ evolution (Figure 1c&d). During growth, the ganglion invades pore by pore, with only one major length scale (the throat) controlling $P_{c}$. In contrast, during shrinkage, the boundary of the ganglion in different pores contracts cooperatively, exhibiting multiple scales of $P_{c}$ during different stages of ganglion shrinkage. In addition, although the $F-V$ correlations of both growing and shrinking ganglia are statistically linear, the surface free energy ($F$) of a shrinking ganglion is, in most cases, higher than that of a growing ganglion at the same $V$.

        This work provides insights for investigating quasi-static degassing, ganglia dissolution, and ripening processes, as well as to analyze the thermodynamic stability of dispersed fluid clusters in heterogeneous porous media. In addition, we call for attention that the term “pore” may not always be a valid basic representative unit during the description of fluid and interface behaviors in porous media, as demonstrated by the presence of cross-pore interfaces.

        (a) Ganglion with cross-pore interfaces, in pore-scale experiments (left) and our model (right). (b) Snapshots of ganglion growth and shrinkage in heterogeneous porous media. (c&d) Evolution of capillary pressure (Pc) and surface free energy (F) with ganglion volume (V) during ganglion growth-shrinkage cycle in heterogeneous porous media.

        Speaker: Chuanxi Wang (Peking University)
    • MS08: 1.3
      • 151
        Dissolution patterns and permeability evolution in dissolving fracture under mechanical deformation

        Fractured rock is widely present in the crust of the Earth and provides main permeable pathways. Mineral dissolution due to reactive fluid flows would enlarge the fracture aperture, and lead to different dissolution patterns and increase the permeability. However, normal stresses would cause mechanical deformation of the fracture and pressure dissolution of contacting asperities, which can further lead to fracture closure and reduced permeability. Here, we systematically study fracture dissolution processes at different normal stresses to reveal the conditions under which fracture permeability increases or decreases. First, we develop a computational model incorporating mechanical deformation, chemical reaction at the free-fracture surfaces and pressure dissolution at contacting asperities, subsequently validating it through experiments. Comparison to existing experiment demonstrate the ability of the computational model to simulate fracture dissolution under normal stress $\sigma$. Then we use the computational model to simulate more than 300 fracture dissolution processes with a wide range of Peclet number $Pe$, second Damkohler number $Da_Ⅱ$, normal stress $\sigma$ and fracture length $L$. We elucidate the underlying mechanisms of different dissolution modes and their permeability evolution. We establish theoretical predictions for transitions of dissolution patterns: $Da_{eff}$ (effective Damkohler number) predicts the transition from wormhole to uniform dissolution; $\Lambda^{-1}$ (thickness ratio of reaction front) predicts the transition from compact to wormhole dissolution. We further develop theoretical predictions for the increase or decrease in fracture permeability under normal stress. There are two conditions under which decrease in fracture permeability occurs: (a) 1/$Da_{eff}$>2 and $a_{eff}$>1; (b) $\Lambda^{-1}$<32 and $a_{eff}$>1, where $a_{eff}$ is effective activity of solid. In all other cases, fracture permeability will increase. This work improves our understanding of fracture dissolution under mechanical deformation and pressure dissolution and is important for many subsurface engineering applications.

        Speaker: Kai Li
      • 152
        A mechanistic investigation of oscillatory zoning using reactive transport modeling

        Solid solutions are widely studied because their formation is ubiquitous in natural and anthropogenic systems. Co-precipitation in rock matrix can result in oscillatory zonation phenomena with solid solutions exhibiting compositional variations (e.g., plagioclase). The principle of co-precipitation of sulphate solid solution is relevant for wastewater treatment of produced waters from hydraulic fracturing and oil/gas extraction, for removing contaminant in uranium mines etc. For nuclear waste disposal, the formation of solid solutions is considered as an important retention mechanism for 226Ra. Despite the widespread occurrence of solid solutions and well-established thermodynamic models, their formation in rock matrix and the effects of transport and kinetics are poorly understood. Previous microfluidic experiments of diffusion-controlled precipitation showed patterns of oscillatory zoning of solid solution crystals of (Ba,Sr)SO4 [1]. In this study, reactive transport modeling is performed to provide a mechanistic understanding of the oscillatory zoning behavior. A micro-continuum approach based reactive transport model that considers probabilistic nucleation was used to simulate the precipitation of (Ba,Sr)SO4 solid solutions following the experimental geometry and setup [2]. It enabled us to compare the contributions of physical-chemical processes that include species-specific diffusion at the solid-fluid interface, solubilities, nucleation kinetics and crystal growth. The models have highlighted that reaction kinetics, rather than transport, are more important in shaping the oscillatory zoning phenomena.

        1. Poonoosamy, J., et al., A lab-on-a-chip approach integrating in-situ characterization and reactive transport modelling diagnostics to unravel (Ba,Sr)SO4 oscillatory zoning. Scientific Reports, 2021. 11(1): p. 23678.
        2. Deng, H., J. Poonoosamy, and S. Molins, A reactive transport modeling perspective on the dynamics of interface-coupled dissolution-precipitation. Applied Geochemistry, 2022. 137: p. 105207.
        Speaker: Dr Hang Deng (Peking University)
      • 153
        Quantifying dissolution dynamics in porous media using a spatial flow focusing profile

        Dissolution processes in porous media produce a diverse range of patterns, depending on the relations between flow, diffusion, and reaction rates. Determining the dissolution regimes is critical for controlling contaminant migration, preventing CO2 leakage during geological carbon sequestration, or assessing the long-term stability of hydraulic structures. In all of these cases, the emergence of highly efficient flow paths within rock can dramatically alter its transport properties. To distinguish between these regimes, we analyze a spatial flow focusing profile based on the focusing index developed by Jang (2011). We segment the medium into cross sections along the flow direction and calculate the flow focusing index for each of them. Consequently, we obtain a profile which is a function of distance from the inlet. Through the analysis of this profile and its temporal changes, we observe various types of evolution, such as a front of increased flow focusing progressing from the inlet in the wormholing regime or a decrease in focusing along the entire medium in the uniform regime.

        We employ this measure in numerical simulations of a dissolving porous medium using a pore-network model. In this approach, we treat the porous medium as a system of interconnected pipes (Budek, 2012) with the diameter of each segment increasing in proportion to the local reactant consumption. By modifying flow and reaction rates in the simulations, we obtain a phase space diagram displaying a variety of dissolution patterns, which we characterize using our quantitative flow focusing measure. Through this analysis, we identify the elusive channeling regime (Menke, 2023). To generalize our findings, we investigate the impact of network heterogeneity on the emerging flow patterns by varying its magnitude and spatial correlation. The findings elucidate the key parameters that determine the dissolution regimes.

        Speaker: Tomasz Szawełło (University of Warsaw)
      • 154
        Dynamics of contaminant flow through porous media containing biochar adsorbers

        Biochar is used as an additive in green roof soil substrates to aid in the regulation of fertilizer storage and dispersal, preventing unwanted runoff of the chemicals. The evolution of contaminant transport and adsorption by biochar added to a packed bed is analyzed using experiments and simulations. Experiment 1 is used to determine the equilibrium capacity and adsorption rate of two types of biochar when immersed in a methylene blue solution. Experiment 2 determines the breakthrough curves of a packed bed of glass beads with randomly interspersed biochar as a methylene blue solution is circulated. Simulations are run using the properties extracted from experiment 1 and the results are compared with experiment 2. An analytical model is proposed and utilized to mimic the behavior of biochar reaching equilibrium, unable to remove additional solute. Monodisperse beds are superior in the removal of solute but removal efficiency is heavily related to the surface area of the reactive particles and the rate at which they become unable to remove additional solute. The cases using the analytical model display a tight distribution of particle surface concentration at times after the solution front passing, indicating full immersion in the solution and therefore maximum removal efficiency. In comparison, the cases with constant reactivity display a much wider distribution of surface concentrations, indicating uneven exposure. The polydisperse beds create more channeling effects which reduce reactive particle efficiency and lead to higher breakthrough concentration profiles. Comparison between experiments and simulations show good agreement with breakthrough curves.

        Speaker: Dr Kaj Pettersson (Chalmers University of Technology)
    • MS09: 1.3
      • 155
        Multiscale Generalized Network Modeling of Carbonates with Sub-Resolution Porosity

        With advances in digital rock physics, pore-scale numerical methods have been developed to estimate various petrophysical parameters based on 3D micro-CT images. However, currently pore-scale models mostly rely on segmented dry scan of the rock sample for network extraction, and the resulting network consists only of resolved pores and throats. For complex rocks such as carbonates that encompass multiple length scales, capturing pores at every scale is often not possible due to the size-resolution trade-off. In this study, we develop a multiscale generalized network model (GNM) by including sub-resolution porosity as another throat type, called micro-links, and modify the flow model by including flow through micro-links. In single-scale GNM, resolved throats are the main pore elements in the network and are divided into corners by certain discretization levels. GNM has several benefits, such as realistic representation of the pore space with the effect of throats expanding from throat center to neighboring pore centers and detailed corner description (Raeini et al. 2017). Moreover, GNM improves the physical accuracy of model predictions by formulating the 3D interfacial curvature between two phases not only in the axial plane but also in the sagittal plane (Raeini et al. 2018; Giudici et al. 2023). We employ differential imaging of brine and dry scans to characterize connectivity and quantify unresolved porosity. We obtain a porosity map containing all voxels with their sub-resolution porosities. Using the dilation algorithm developed by Foroughi et al. (2023), each microporous voxel is labeled according to its two closest pores, and then microporous voxels with the same closest pores are classified as a micro-link. Since we consider micro-links as continuous Darcy-like porous media, we use classical empirical relationships to describe flow in micro-links. We first tested our multiscale model with highly permeable Ketton limestone. We tuned our model to mimic the reported Ketton mercury injection capillary pressure (MICP) data, which exhibited a bimodal throat size distribution, one peak at larger pores is attributed to interparticle resolved pores large enough to be captured at micro-CT voxel size, and the peak at small pore size is for intraparticle micropores. And in between, there is an intermediate interval covering unresolved macropores, which are larger than micropores but are under resolution. To achieve a good match with the MICP curve, we determine the critical micro-link porosity as the boundary between different porosity regions and evaluate their saturation exponent and grain diameter values separately. After calibrating the model with measured permeability, formation factor, and MICP data, we generate a bimodal capillary pressure curve for oil drainage into initially water-wet system. Even in the early stages of drainage, we observe an increase in the relative permeability of water, contributed by unresolved porosity. In summary, our approach shows significant promise in addressing sub-resolution porosity in a less computationally costly manner using micro-links. We aim to extend our approach to other complex multiscale systems, including fuel cells, membranes, and batteries.

        Speaker: Branko Bijeljic (Imperial College)
      • 156
        The interplay between temperature evolution, species distribution, and microstructure dynamic in a calcining porous particle

        Reacting particle systems play a crucial role in various industrial applications, with limestone calcination serving as a prime example. In this process, applying high-energy input to calcium carbonate (CaCO$_3$) particles results in the production of active lime solid (calcium oxide CaO) and the by-product carbon dioxide (CO$_2$) gas. The legal obligation to mitigate CO$_2$ emissions has a notable impact on production costs, emphasizing the need for a thorough understanding of the calcination process. Such an understanding can enhance conversion process efficiency and enable the achievement of desired CaO structures, crucial for high efficiency in CO$_2$ adsorption. This study introduces a pore network model to explore the interplay between intraparticle heat and mass transfer, pore structure changes during chemical reactions, and their interactions with the surrounding fluid-solid environment in a single particle. The pore space and solid skeleton of the particle are approximated as regular-lattice networks, incorporating cylindrical pores and volume-less nodes. Local reaction rates are determined based on the effective specific surface area and local CO$_2$ pressure. Thermal energy is supplied through hot gas flow convection at the network's surface. Solid element dimensions and corresponding pore structures are updated during each time step, enabling the tracking of temperature evolution, local conversion, and void space structure changes within the particle. Simulation results reveal that when the bulk gas has initially low CO$_2$ pressure, calcination extends beyond the particle surface and occurs within the particle. As a result, the released CO$_2$ becomes trapped due to internal mass transfer resistance, impeding further calcination. On the contrary, exposure to high initial CO2 pressure exclusively promotes local calcination reactions at the particle's surface. The initial bulk CO$_2$ pressure also has a notable impact on the final structure of CaO. This intricate interplay in calcination, as demonstrated by the simulations, provides valuable insights for a better understanding and optimization of industrial processes, applicable not only to calcination but also to other heterogenous reactive systems.

        Acknowledgement

        Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Project-ID 422037413 – TRR 287.

        Speaker: Xiang Lu
      • 157
        Quantification of geometric and flow characteristics for CO2 storage at pore-scale using a DC-GAN based digital experiment approach

        The trapping efficiency of CO2 storage in porous subsurface is influenced by various geometric and flow characteristics. Conducting experimental studies on reservoir structure characteristic parameters and actual storage efficiency consumes a significant amount of resources, making it difficult to analyze the uncertainty of parameters through a large number of experiments. In this work, a deep convolutional generative adversarial network (DC-GAN) was employed to generate 1000 sets of images that are visually indistinguishable by using tomographic images of Bentheimer sandstone as the training data. This is followed by performing image analysis and pore network modelling to obtain geometric (e.g. Minkowski functionals) and flow (absolute and relative permeability, capillary pressure, saturation, and trapping efficiency) properties. With maximum capillary of 7.0 KPa, we found that the trapping efficiency ranged from 32% to 40%. We then explored the uncertainty of all geometric and flow characteristics to determine the minimum number of digital experiments to reproduce the same statics. This work proposes a strategy for coupling deep learning method and pore network models to conduct a large number of digital experiments on complex porous media. This can be used to correct experimental errors obtained through traditional experimental methods and guide the design of geological CO2 storage systems.

        Speaker: Ms Yifan Zhang (Zhejiang University)
      • 158
        The numerical simulation of two-phase flow in multi-mineral shale digital rock cores

        Shale oil and gas resources are widely distributed and have abundant reserves in China, with broad development prospects and potential. Due to the inherent characteristics of shale oil, such as the large number of nanometer-sized pores with complex pore structures, significant fluid-wall effects, complex mineral compositions, abundant organic matter, and complex and diverse wettability, the flow law of multiphase flow under shale reservoir conditions differs significantly from conventional reservoirs. Therefore, it is essential to characterize the flow of multiphase flow under shale pore scales, considering TOC content, complex wettability, and adsorption conditions. In this work, a new method for simulating the pore scale of shale oil-water two-phase flow based on mixed multi-mineral phase digital cores is proposed. This method is based on identifying the pore wall surfaces of each mineral and considering the corresponding adsorption and wettability conditions. Firstly, the multi-mineral phase shale digital core is reconstructed from the two-dimensional scanning electron microscope image of the shale sample, and the pore space of the corresponding minerals is divided by grids. Secondly, based on the Navier-Stokes equation, considering the TOC content, complex wettability conditions, and adsorption layer, the VOF method is used to simulate the shale oil-water two-phase flow process at the pore scale. Finally, the influence of TOC content, complex wettability, and adsorption on the shale oil-water two-phase flow is analyzed. The results have shown that the effects of TOC content, complex wettability, and adsorption on shale oil-water two-phase flow cannot be ignored. As the TOC content increases, the contact area between the organic pore walls and oil and water also expands. The flow of the water phase is hindered by the oil-wet organic pore walls, which significantly reduces the movable degree of oil. The movable degree of oil is also influenced by the organic matter distribution. In addition, the change in wettability of shale pore walls can also affect the flow of oil and water phases in shale. With the stronger oil-wetness of shale pore walls, especially organic pore walls, the movable degree of oil decreases. Due to the narrow pore radius and strong fluid-wall interaction, the adsorption phenomenon is significant, causing oil to remain on the pore surface, thus, greatly reducing the movable degree of oil. Our findings are critical for enhancing the efficiency of shale oil recovery, carbon dioxide geological sequestration, and other related areas.

        Speaker: Guangyuan wei (China University of Petroleum(East China))
    • MS13: 1.3
      • 159
        Wetting behaviors and oil occurrence status of shale reservoirs

        A contrast-variation technique in small-angle neutron scattering (CV-SANS) is employed to investigate the interfacial chemistry of shale oil reservoirs using reagents that span a range of polarities, including water, toluene, and dimethyl methanamide. Through five different experimental strategies, the work demonstrates a modification of shale wettability, ranging from enhancement, weakening, to reversal. This study also presents an innovative approach for characterizing the status of oil occurrence at the nanometer scale, and new insights into the scattering vector-particle size (q-r) relationship in polydisperse systems. The unique CV-SANS technique shows that shale with low contents of total organic carbon, clays, and porosity typically indicates better oil mobility, associated with larger particle scales. Meanwhile, the results indicate that an increase in pore scale does not necessarily accompany the rise in radius of gyration when mass density spatial variation occurs in the system. Collectively, this work establishes a direct correlation between size r in real space and q in reciprocal space and decodes the interfacial wettability traits in nanopores of shale oil reservoirs.

        Speaker: Dr Tao Zhang (Department of Earth and Environmental Sciences, University of Texas at Arlington, Arlington, Texas 76019, United States.)
      • 160
        Experimental study of gas flow and relative permeability in low-porosity media using LF-NMR

        The key characteristics of complex oil and gas reservoirs include high heterogeneity in the porous space, ultra-low permeability resulting from the nanoscale dimensions of pores, and the presence of solid insoluble organic compounds in the rock. These factors complicate the application of existing physical and mathematical flow models with sufficient accuracy. This causes challenges in accurately predicting producible hydrocarbon reserves and impedes the development of unconventional formations.
        To enhance the characterization of the gas flow in low-porosity and low-permeability media, a detailed study of mass transfer is essential. This involves obtaining the relative permeability for multiphase systems that is typically determined from laboratory filtration experiments using rock samples. However, assumptions made during relative permeability determination, including methodological ones, often prove invalid when describing mass transfer in low-permeability porous media, resulting in quality of the field hydrodynamic model and increase development risks.
        The objective of this study is to develop an enhanced laboratory-based method for determining gas-water and oil-water relative permeability in medium- and low-permeability reservoirs using NMR relaxometry under high-pressure conditions. NMR relaxometry enables the quick and accurate determination of sample saturation in tight rocks without additional contrasting during the core flooding experiment. The study considers surface phenomena in liquid-gas, liquid-solid, liquid-liquid, and gas-solid systems, determining their impact on mass transfer during filtration in rock samples. Experimental study was conducted using pressure-pulse decay porosimetry for obtaining reservoir properties, low-field NMR relaxometry (2 MHz) on 1-inch core plugs and X-ray computed microtomography. Probes of degassed oil samples, deuterium oxide, carbon dioxide and methane were used as fluids in experimental modeling of the relative permeability.
        As result, the method for determining the relative permeability in two-phase systems using NMR relaxometry was proposed. The current study also reports the results of fluid adsorption on crushed rock samples containing clays and organic matter using high-pressure NMR experiments. The analysis of results include the three-dimensional pore-network model based on experimental work. Validation of the model was conducted using values of reservoir properties (porosity, permeability, and pore size) determined by computed microtomography, as well as permeability values from the obtained relative permeability curves. In addition, sensitivity analysis of the developed pore-network model highlights the influence of pore connectivity, pore size distribution, gas phase diffusion on the absolute and relative gas permeability.

        Speaker: Dr Aliya Mukhametdinova (Skolkovo Institute of Science and Technology)
      • 161
        A comprehensive study on shale pyrolysis dynamics by real-time in-situ imaging technology

        Shale oil and gas primarily exist in nanoscale pore-fracture networks. Despite of the large resources of the oil shale and low-medium mature shale, limited removable hydrocarbon and extremely low permeability due to limited pores restrict in the development of those unconventional resources. Therefore, different pyrolysis technologies, such as in situ conversion pyrolysis, superheated steam, nitrogen, electrofrac and etc., were emerged to accomplish the recovery. With the pyrolysis and maturing process, the kerogen was transformed into oil and gas, and more fractures and pores were generated, which increases both the permeability and hydrocarbon in the shale.
        In order to study this dynamic process, a real-time in-situ imaging via environmental scanning electron microscope was applied to characterize and analyze the nano to micro scale changes of the shale quantitatively. Afterwards, Energy Dispersive Spectrometer (EDS), Rock Evaluation, and Thermal Gravity Analysis-Fourier Transform Infrared Spectroscopy (TGA-FTIR) were conducted for the physical and chemical alternation of the shale components and expulsed fluid. The real-time in situ SEM showed that 1) nano-fractures started to appear below 100 °C; 2) inorganic nano-fracture width demonstrated a non-monotonous relationship with temperature; and 3) kerogens amount decreased monotonously as temperature increased, especially during 400-500oC. TGA-FTIR indicated 4 pyrolysis stages with different characteristic changes, in which main expulsed products were CO2, H2O, and light hydrocarbons C1-C5. SEM images along with EDS characterized the inorganic components and their changes after pyrolysis.
        These findings will promote fundamental understanding of oil shale pyrolysis dynamics at nanoscale and provide key guidance on oil shale extraction at reservoir scale.

        Speaker: Xia Yin (Petroleum Exploration and Production Research Institute, Sinopec)
      • 162
        Occurrence characteristics and quantitative evaluation of micro-nano pore shale oil: A case study of Lianggaoshan Formation shale strata in northeast Sichuan, China

        The reservoir space of shale series is mainly composed of micro and nano pores. It is significance to carry out the occurrence characteristics and quantitative evaluation of shale oil with micro- nano pores for the study of shale oil rich and integrated reservoirs. In 2020, Well Ping’ an 1, located in the northeastern region of Sichuan Basin, achieved a high yield of 100 tons in Jurassic Lianggaoshan Formation shale, realizing a major breakthrough in shale oil and gas exploration of Lianggaoshan Formation in Sichuan Basin and showing a broad exploration prospect of shale oil in this area [1]. However, the rapid sedimentary facies change, complex and diverse lithofacies, large maturity distribution range (Ro: 0.9%-1.9%), and multiple oil types of the shale strata in the Lianggaoshan Formation lead to unclear occurrence characteristics and content of micro-nano pore shale oil, which seriously restricts the optimization of favorable zones in this area [2]. Therefore, taking the medium-high mature shale strata of Lianggaoshan Formation in northeast Sichuan area of China as an example, this study carried out experimental analysis of conventional core and preserved core samples in the study area, such as scanning electron microscopy, low-temperature nitrogen adsorption, high-pressure mercury injection, nuclear magnetic resonance, multi-step pyrolysis, and original oil and gas chromatography. The results show that oil film/carbon slag is widely developed in the micro-nano pores of shale, and the lower limit of free oil pore size is ~3nm. The occurrence (oil) volume of organic-rich layered clay shale and organic-low layered felsic shale is the largest. TOC, clay content and maturity are the main control factors of oil adsorption in Lianggaoshan shale. Combined with light hydrocarbon recovery and heavy hydrocarbon correction, the total oil content calculation model of the original formation in the study area is determined. The total oil content of the organic-rich layered clay shale and organic-low layered felsic shale is higher than that of other rock facies, which can be used as a favorable oil reservoir for research.

        Speaker: Xuefeng Bai (China University of Petroleum (East China))
    • MS15: 1.3
      • 163
        Application of Diffusion Models to Generate Multiphase Fluid Pore-Scale Images

        Generative Adversarial Networks (GANs) have been a typical example of how machine learning has been successfully applied, using three-dimensional images as training datasets, to generate realizations of the pore space, as well as to produce super-resolution images. We further this work with a new generative model: diffusion models (DMs), to generate images of both the pore space and two fluid phases within the pore space, using experimental high-resolution three-dimensional X-ray images of the pore space and fluids at different fractional flows as training datasets. We demonstrate that using DMs, we can generate images for a range of saturations and compare the quality of these realizations against experimental data in terms of Minkowski functionals: saturation, interfacial area, mean curvature, and connectivity (Euler characteristic), as well as contact angle. DMs are a very promising algorithm type for the study of multiphase flow in porous media, with effectiveness comparable to, if not surpassing, GANs. We discuss the use of this methodology to complement pore-scale displacement and imaging experiments, to generate images of arbitrary size and for a wide saturation range. These images provide a basis for further analysis and pore-scale modeling, including the prediction of averaged multiphase flow properties, such as capillary pressure and relative permeability.

        Speaker: Linqi Zhu (Imperial College London/ Chinese Academy of Sciences)
      • 164
        Predicting ultimate hydrogen production and residual volume during cyclic underground hydrogen storage in porous media using machine learning

        In the context of climate change mitigation, underground/subsurface hydrogen storage (UHS) is regarded as a solution that could help tackle the imbalance in renewable energy supply. Excess energy can be stored as molecular hydrogen (H2) and re-used when it is needed. To enable large-scale storage in underground geologic formations, reservoir simulation of cyclic loading scenarios will be used to optimize the storage operations. Hydrogen storage in geological reservoirs involves many physical phenomena related to reservoir dynamics, trapping mechanisms, and potential reactions with minerals and bacteria. Simulating different scenarios of fast H2 injection and production while considering those physical constraints and optimized economic and operational parameters generates loads of data and therefore requires high computational power. The nature of UHS operations and the underlying storage reservoir physics make variations in the generated data sets extensive. Predictive tools like machine learning (ML) that are data dependent can to some extent fill the knowledge gap while simultaneously making the operations more viable. It is therefore interesting to develop tools that can predict parameters associated with fast cyclic operations while minimizing the computational cost. Such methods could help optimize storage operations and reduce operational costs.
        The work summarized in this abstract attempt to showcase how machine-learned models trained with data generated from simulated UHS systems in porous media can be used to predict ultimate hydrogen production. The same approach is applied to predict H2 amounts that remain trapped in the reservoir due to physics-related parameters. The OPM flow reservoir simulator is used to build models encompassing physical and dynamic parameters to generate cyclic field data which are then used to train time series neural network (NN) models. In the presented work, the reliability and accuracy of the model are ensured through hyper-parameter tuning and cross-validation analysis on a windowed time-series NN. The results obtained in the study show the relevance of machine learning (ML) methods in predicting ultimate H2 production and residual H2 amount in geological reservoirs. The trained models captured the data trends with mean squared error (MSE) and mean absolute error (MAE), commonly termed as loss functions, from training and validation steps used as accuracy metrics. In one of the reservoir-scale cases, the machine-learned training and predictions in a physics-oriented approach reduced the computational time by about 6773% in comparison with simulation runs by OPM flow on a 4-layer reservoir model. The accuracy metrics and predictions are even much better on simulation data obtained with a one-layer model having a horizontal well along its top and based on a complex cyclic schedule. By showing how machine-learned models can capture some of the complex physical uncertainties associated with underground/subsurface hydrogen storage, our research aims to bring a technical contribution to the development of this technology. Field-scale simulated production and injection data are used to train the models. The paper therefore presents the methodology followed, results from the machine learning methods, and the outlook for future tasks.

        Speaker: Raymond Mushabe (University of Bergen)
      • 165
        Efficient 3D Digital Rock Detail Reconstruction and Quality Enhancement with Super-Resolution Transformer

        High-quality digital rocks are essential for high-precision pore-scale modeling. However, limited by the imaging hardware, meeting the requirements of high resolution (HR) and a wide field of view (FOV) simultaneously is challenging. In this study, we propose a novel Efficient Attention Super-Resolution Transformer (EAST) to boost digital rock quality, which reconstructs HR details from low resolution 3D images with wide FOV. To address the specific characteristics of digital rock tasks, EAST employs a hybrid loss function to recover sharp pore edges and combat noise. Furthermore, we utilize data augmentation techniques to improve model generalization. The hyperparameters of EAST are optimized to trade off speed and super-resolution quality. Through quantitative evaluations and qualitative visualizations, we validate the superior reliability of EAST in terms of recovering sharp edges and eliminating noise. Compared to the efficient convolutional neural network-based model RCAN, EAST achieves higher performance with only 22% of the parameters due to hybrid efficient attention mechanism. Finally, we verify the physical accuracy of the EAST reconstruction results by direct flow simulation method. The results demonstrate that EAST significantly reduces the relative error of single-phase permeability from 35% in Tricubic interpolation to 8%. Moreover, EAST is 185% faster than RCAN, which implies that EAST could process a digital rock with $10^{9}$ voxels in 4.3 hours, thereby generating an impressive 6.4×$10^{10}$ voxels.

        Keywords: digital rock physics, 3D image super-resolution, deep learning, Transformer

        Speaker: Mr Zhihao Xing (China University of Petroleum (East China))
      • 166
        Integrating deterministic geological model with multimodal machine learning to predict shale productivity

        Over the past ten years, diverse machine learning techniques have been extensively employed in forecasting output for non-traditional reservoirs. Nevertheless, these techniques primarily utilized discrete point data obtained from field databases, such as well drilling, completion, monitoring, experiments, and production data of horizontal wells. However, this data fails to capture the spatial heterogeneity of reservoir properties, which ultimately undermines the reliability of shale gas production. This study proposes a multimodal machine learning approach that utilizes a geological model restricted by well-logging and 3D seismic data. The deterministic geological model is constructed by utilizing the high vertical resolution of well logs and the planar resolution of reflection 3D seismic attributes. Subsequently, the detailed and precise data regarding geological properties such as porosity, permeability, gas saturation, TOC, brittleness, thickness, etc. in the vicinity of horizontal wells are acquired using the aforementioned geological model. These data are then combined with traditional tabular datasets to accurately represent the heterogeneity of the reservoir. Subsequently, a multimodal model is created that combines a convolutional neural network (CNN) module and an artificial neural network (ANN) module. The CNN module is designed to handle high-level information from the visual dataset, while the ANN module is used to evaluate the typical tabular datasets. A fusion module integrated and processed input from both modalities. The results demonstrate that the multimodal approach attained a coefficient of determination (R2) of 0.845 for the 12-month shale gas production prediction, which is greater than the R2 value of 0.721 obtained using simply the ANN model. In addition, this approach based on the multimodal of analysis can elucidate the varying levels of shale gas production between two horizontal wells that have similar average reservoir attributes. This is achieved by taking into account the lateral heterogeneity of the producing formations that the two horizontal wells have penetrated. Hence, the exceptional predictive accuracy of the multimodal machine learning technique offers valuable insights into forecasting shale gas production. This approach may be utilized to guide the selection of optimal locations for new horizontal wells and facilitate the efficient exploitation of shale gas reservoirs.

        Speaker: Gang Hui (China University of Petroleum Beijing)
    • Poster: Poster Session III
      • 167
        A multiscale simulation method for aerosol transport in a mouth-to-lobar bronchi model

        Prediction of aerosol deposition in the respiratory tract has become a major focus for inhaled drug delivery and air pollution prevention. Computational fluid-particle dynamics (CFPD) provides the most accurate local prediction results, but the computational cost is unbearable for the CFPD simulation of the whole respiratory tract. This challenge arises due to the multiscale nature of the respiratory tract and numerous bronchioles [1]. A common practice is to truncate the bronchial tree and simulate on the truncated model using CFPD. In this study, a novel boundary condition for the truncated respiratory tract model based on multiscale simulation is proposed, named extended-bronchus-network (EBN) boundary condition. The truncated model is extended to the terminal bronchial and the air flow in the extended part is simulated using local hydraulic resistance equivalence pore network model (PNM) [2]. The pressure and flow rate at the outlet of truncated model is consistent with PNM, which provides the outlet boundary condition for CFPD of truncated model. A comparison against EBN with the widely used uniform pressure outlet boundary condition [3] is made. It reveals that EBN boundary condition in this study is more physiologically and closely to the clinical data. The maximum relative disparity of nano-micro aerosol penetration fraction of the right middle lobe and right lower lobe between these two methods is 93% and the maximum relative disparity of aerosol deposition fraction within the trachea-lobar bronchi is 30%. EBN boundary condition is implemented for the simulation of nano-micro particles transport in the mouth-to-lobar bronchi (MLB) model at the inspiration volume rate of 15, 60, 90 L/min, respectively. Result shows that for particles equal to or less than 1 μm in size, over 90% penetrate deeper into the pulmonary lobes, with inspiration volume rate and particle size having minimal impact on penetration fraction. However, micro particles more probably deposit in the MLB with larger inspiration volume rate. Notably, when particles larger than 6 μm are inhaled at 15 L/min or particles larger than 3 μm are inhaled at 60 L/min, over 40% of them deposit in the MLB. Particle deposition hotspots forming reason is qualitatively analyzed. This work provides a reference for the optimization of drug delivery, targeted therapy, the prevention and control of pollutants. It also lays a foundation for the simulation of aerosol transport in whole lung.

        Speaker: Han Xiao (Tsinghua univeristy)
      • 168
        A Novel Approach for Advancing Lithology Classification Through Machine Learning and Deep Learning Models

        Manual interpretation of geophysical logging data can be a tedious and time consuming task in the case of the non-linear behavior of well-logging signals. In this study, we introduced three novel algorithms including GrowNet, Deep-Insight and blender in the classification of rock facies. To compare the performance of these models, we used algorithms such as XGBoost, Random Forest and Support Vector Machine. The data employed is from the South and North Viking Graben, comprising twelve lithological rock facies. Deep-insight was used to convert tabular data into images and these generated images were employed as inputs for a convolutional neural network. It demonstrated better performance in lithology classification compared to traditional models such as Decision Tree and Logistic Regression. The GrowNet and blender models for lithology classification successfully increased the penalty score and accuracy compared to the FORCE2020 competition. This study highlights the value of a hybrid approach, integrating the SMOTE and NearMiss algorithms in order to balance the data. Addressing missing data is crucial for dependable analysis; employing regression models, rather than simplistic techniques such as mean imputation, enhances accuracy. Additionally, knowledge-based feature augmentation techniques are selectively applied based on the availability of relevant features, thereby enhancing the effectiveness of the overall model learning process. To more efficiently evaluate and compare the performance of the models in a multi-class classification, we introduced the class prediction error plot instead of using confusion matrix.

        Speaker: Dr Seyed Mojtaba Hosseini-Nasab (School of Chemical, Oil and Gas Engineering)
      • 169
        Acoustic Properties of Hydrate-Bearing Porous Media Based on Electrical-Mechanical-Acoustic Multi-physics-Field Coupling Model

        There are technical difficulties in accurately controlling and evaluating the micro-distribution mode and saturation of hydrate in physical simulation experiments. Limitations exist in the experimental technologies for investigating acoustic properties of hydrate-bearing sediments and establishing interpretation models of reservoir parameters. The acoustic properties of hydrate-bearing sediments are influenced by hydrate saturation and micro-distribution modes, skeleton particle arrangement and shape significantly. Currently, there is a lack of research work on the influence mechanisms of skeleton particle arrangement and particle shape. Three-dimensional numerical models were established for hydrate-bearing porous media based on digital rock physics technology. For the three kinds of hydrate micro-distribution modes (suspension, contact and cementation), finite-element models were established individually based on the method of electrical-mechanical-acoustic multi-physics-field coupling. The effects of micro-distribution mode and hydrate saturation on sound velocity and attenuation of porous media were examined. The results of sound velocity from the numerical and theoretical models were compared. The influences of skeletal particle arrangement modes and shapes on the sound velocity and attenuation characteristics of sediments under different hydrate micro-distribution modes and saturation conditions were explored, and the mechanisms were discussed. It was demonstrated that: (1) when the hydrate saturation is low, the volumetric proportion of quartz sand particles in the diamond-arrangement model is higher than that in the cubic-arrangement model, thus the sound velocity of the diamond-arrangement model is higher; as the hydrate saturation increases, the difference in the volumetric proportion of hydrates between the two models increases and the volumetric proportion of hydrates in the cubic-arrangement model is higher, consequently the sound velocity growth rate in the diamond-arrangement model is lower; (2) the porosity of the diamond-arrangement model is smaller than that of the cubic-arrangement model, and the energy attenuation during the propagation of sound waves is lower; (3) compared with the spherical-particle model, the elliptical-particle model contains more pores with smaller aspect ratios, resulting in a smaller bulk modulus and lower sound velocity; (4) the ellipsoidal-particle model contains more and smaller pores, which results in lower wave-energy loss than that of the spherical particle model. This study may provide a theoretical support for the data interpretation of seismic exploration and sonic logging for natural gas hydrate reservoirs.

        Speaker: Lanchang Xing (China University of Petroleum (East China))
      • 170
        Advancing Underground Hydrogen Storage: Insights from Molecular Simulations of Wettability and Interfacial Tension

        The intermittency in energy supply and demand from renewable resources, which is often caused by seasonal variations, necessitates the development of long-term energy storage solutions. One promising approach is Underground Hydrogen Storage (UHS), a scheme in which hydrogen is strategically stored in subsurface formations to maintain market equilibrium. Depleted oil and gas reservoirs and saline aquifers present an attractive option for UHS. The appeal of these reservoirs lies in their existing infrastructure and the presence of subsurface traps, which are essential for sealing reservoir fluids. To fully leverage these reservoirs for UHS, it’s crucial to understand the wettability and interfacial tension (IFT) of the gas, brine, and rock systems. These properties play a significant role in determining capillary pressure, fluid migration, and drainage in porous media. We used molecular dynamics simulations to study the impact of temperature (300 and 323 K) and pressure (5, 10, 15, and 20 MPa) on the IFT between hydrogen and formation brine. We also investigated the effects of two carrier gases, methane and carbon dioxide, on the system. Our findings revealed that the IFT of pure hydrogen/brine did not change significantly (less than 1 mN/m) with pressure. However, the addition of methane and carbon dioxide to the system lowered the IFT, with carbon dioxide having a more pronounced impact. In addition to IFT, we studied the contact angle between a gas phase (comprising hydrogen and its mixtures with methane and carbon dioxide), the brine and rock (calcite and silica) at a pressure of 20 MPa and a temperature of 300 K. With the exception of the H2-CO2/brine/silica, all systems exhibited complete water-wetting with a zero-contact angle. The exception can be attributed to the complete protonation of silanol groups, which is caused by the dissolution and hydration of CO2 to form carbonic acid, resulting in a low pH of 3. This research contributes to our understanding of UHS and provides valuable insights that could aid in the optimization of energy storage strategies. It underscores the importance of considering the effects of temperature, pressure, and carrier gases on the IFT and wettability of the system, which are critical for the efficient storage and retrieval of hydrogen in subsurface formations.

        Speaker: Moataz Abu-Al-Saud (Saudi Aramco)
      • 171
        Applications of pore network modelling in predicting the permeability in hydrate-bearing sediment

        Permeability plays a pivotal role in governing the fluid flow within hydrate-bearing sediment (HBS) and significantly influences the efficiency of natural gas production from hydrate reservoirs. However, the measurement of HBS permeability is challenging due to the complexities of maintaining phase equilibrium conditions during testing. This study focused on the sandy hydrate-bearing sediments and intended to elucidate the evolution of absolute and relative permeability as a function of hydrate saturation by the mean of pore network modelling. In the developed model, the hydrate formation process in the porous media is simulated incorporating two key sub-processes: hydrate nucleation and hydrate growth. We integrated various theories from hydrate kinetics, including the random nucleation theory, interface growth theory, Ostwald-Ripening effect, and pore water activity theory, to control the hydrate formation process. For modeling fluid flow within the pore networks, we utilized the conductivity calculation method. The constructed pore network model was employed to analyze permeability variations within different pore networks. Simulation results from a series of regular networks demonstrated that the distribution of formed hydrate and the permeability of HBS are influenced by factors such as model dimension, hydrate nucleation fraction, and hydrate growth type. Further simulations based on CT images showed the changes of absolute permeability and gas-water two-phase permeability during hydrate formation within different sandy sediments. As a result of these simulations, we provided parameter ranges suitable for the application of the Masuda model in predicting absolute permeability in sandy hydrate reservoirs, as well as the Brooks-Corey model and van Genuchten model for predicting two-phase gas-water permeability. This study is hoped to bring new insights into the field of micro-scale seepage research within hydrate reservoirs.

        Speaker: Dr Yongchao Zhang (Qingdao Institute of Marine Geology)
      • 172
        Assessment of CO2 Storage Capacities in Saline Aquifers Using Material Balance Equations

        Estimation the capacities for CO2 storage plays a pivotal role in carbon capture and storage (CCS) projects. The material balance equation (MBE) approach, commonly applied in this domain, provides direct estimations of CO2 storage potential. Traditional MBE techniques, however, often compute the original fluid in-place volume via volumetric assessments without subsequent validation, posing challenges to estimation accuracy. Recognizing the impact of precise original fluid in-place volume calculations reflective of the reservoir's pore volume on storage capacity, our research introduces an enhanced MBE methodology. We evaluated the effectiveness of our enhanced MBE method through comparisons with a commercial reservoir simulator, applying it to both a constructed synthetic aquifer scenario and the Sleipner field model. The application to the synthetic aquifer scenario resulted in a notably precise CO2 storage capacity estimate, with a relative error of just 2.085%, based on data from a single year of injection. This high level of accuracy was similarly observed in the Sleipner field application, where the method achieved a relative error of 1.6873%. Our developed MBE method further demonstrates its reliability in accurately forecasting CO2 storage capacities across a range of conditions, including variations in grid sizes, rock properties, injection rates, operational strategies, and geological heterogeneity.

        Speaker: Sangkeon Park (Seoul National University)
      • 173
        Comparisons between a dual-pore-network model and a hybrid pore-network-continuum model for predicting permeability and formation factor of multiscale carbonate digital rocks

        Many underground rocks have been found to possess complex multiscale porous structures with bimodal/multimodal pore size distributions, such as carbonate rocks, tight sandstones, and shales (Bultreys et al., 2016; Shan Wang et al., 2022; Nijat Hakimov et al., 2022). Flow and transport in these rocks play an important role in many subsurface applications. In addition to in-situ core experiments, several pore-scale numerical models have been developed to simulate flow and transport in multiscale porous structures, including dual-pore-network, micro-continuum and pore-network-continuum models (Francisco J. Carrillo et al., 2020; Zhang et al., 2021; Tom Bultreys et al., 2015). Compared to micro-continuum and pore-network-continuum models, a dual-pore-network model is computationally efficient and can be used to the full core analysis. However, the effect of smearing heterogeneity of microporous domains (i.e., sub-resolution domains) on numerical predictions needs to be studied.
        In this work, absolute permeability and formation factor of Estaillades carbonate rocks are modelled by both a dual-pore-network model and a hybrid pore-network-continuum model. We show the key difference between the dual-pore-network model and pore-network-continuum model for treating microporous domains. By comparing numerical predictions of the two models, the influence of microporous heterogeneity on seepage characteristics of carbonate rocks is extensively explored. Moreover, the dual-pore-network model is used to study the influence of image resolution on the prediction of permeability and formation factor. As reducing the resolution of the original image, it is observed that more and more resolved pores are identified as microporosity, while the modelled permeability decreases.

        Speaker: Mr Xingyuan Zhao (chongqing university)
      • 174
        Constructing Three-Dimensional Digital Rock of Continental Shale with Multi-Mineral Components Using Machine Learning Segmentation Algorithms

        Continental shales are characterized by their highly developed laminations and a high clay content, which pose significant challenges in terms of sample preparation and fluid saturation process for traditional rock physics experiments. Digital rock physics (DRP) has been emerged as an alternative method for unconventional reservoir. The establishment of a high-precision three-dimensional digital rock is crucial to ensure the accuracy of numerical simulations for determining rock physics properties. While continental shale reservoirs exhibit numerous nanopores, indistinct clay particle boundaries, and small fractures in bedding planes, which pose challenges to the segmentation of two or three dimensional grayscale images.X-ray Computer Tomography (CT), Scanning Electron Microscope Mineral Quantitative Evaluation (QEMSCAN), and Multi-spectral Automated Petrographic System (MAPS) tests are performed sequentially on continental shale samples. The scanning resolutions for these tests are 1.35μm, 1μm, and 10nm, respectively. Initially, the grayscale ranges for different mineral components were identified by combining QEMSCAN with CT scans images. Afterwards, a machine learning image segmentation algorithm was employed to partition the CT scan grayscale images into five components: pores, organic matter, clay minerals, feldspathic minerals, carbonate minerals, and pyrite. Subsequently, the same machine learning segmentation algorithm was applied to the two-dimensional MAPS images of the shale sample. This was done to identify pore spaces that were smaller than the CT resolution present in the carbonate minerals, organic matter, and clay minerals, and to calculate the surface porosity.The segmentation results of the X-ray CT scan images indicate that the machine learning segmentation algorithm improve the accuracy in identifying the boundaries of the matrix and pores compared to traditional grayscale-based segmentation method. The machine learning-based image segmentation algorithms can also accurately identify unidirectionally extended microcracks. The contents of the main mineral components calculated from digital rocks agree well with those measured by X-ray Diffraction (XRD). However, the porosity identified in CT images was considerably lower than helium porosity of the samples because only large intergranular pores and fractures can be resolved by CT. There are lots of sub-resolution pores in continental pores as illustrated in MAPS images compared to marine shale gas reservoirs, the organic matter pores in continental shale oil reservoirs are not well developed, and the microporosity is about 5%. The intercrystalline pores in clay minerals are the main reservoir space for continental shale oil, with a microporosity of 10%. The total porosities of the multi-mineral component digital rocks were calculated by considering the volume fractions of the main minerals and their corresponding microporosities. The porosity values obtained from the digital rocks exhibit excellent correlation with those derived from laboratory measurements. The three-dimensional digital rock model serves as a precise representation of the pore structure, enabling quantitative analysis of the microstructure and numerical simulation of the physical properties of continental shale.

        Key words: Digital Rock Physics, Continental Shale, Machine Learning Image Segmentation, Pore Structure Characterization, X-ray Computer Tomography (CT) scans Imaging

        Speaker: Ms Min Li (China University of Petroleum (East China))
      • 175
        Coupled studies of oil compositions and storage spaces in the Kongdian Shale Formation, Bohai Bay Basin, Eastern China

        The shale in the Kongdian Formation, Bohai Bay Basin is one of the leading shale oil targets in China and attracted attention from both industry and academics since the 2010s. Shale oil is known to distribute in the $\mu$m to nm pores in the shale formations. The oil compositions and pore spaces of shale formations strongly influence the shale oil production behavior for an efficient development. The four components of shale oil (saturates, aromatics, resins, and asphaltene) and pore space of shale have been studied separately in many previous researches. However, the combined distribution of different shale oil compositions in the pore spaces has rarely been studied. To investigate the storage space of different compositions of shale oil, t, this study combined organic geochemical and petrophysical methods for core samples of Kongdian Formation. Extracts were collected after the 1st, 5th, 10th, and 30th day of solvent extraction, then analyzed by bitumen extraction and separation and GC-FID to determine the composition changes vs. time. Shales were collected at the same time interval to investigate the pore structure changes vs. time by mercury intrusion porosimetry and nitrogen adsorption. Results show that as the solvent extraction time gets longer, (1) the main composition of extracts changes from saturates to resins, and saturates gradually become heavier, (2) the porosity of shale gradually increases, and pore space first mainly increases in the $\mu$m-nm pore, then mainly increase in nm pores. In summary, this study is important to understand the distribution of different shale oil compositions and provide fundamentally important knowledge in reservoir evaluation.

        Speaker: Weixing Yan (China University of Petroleum (East China))
      • 176
        Development of multiphase flow simulation method in DEM under a movable-grain condition

        Understanding how hydrodynamics of multiphase flow couple porous media deformation is essential to ensure successful engineering practices such as geological carbon sequestration. However, existing hydro-mechanical coupled models face significant challenges in reliably and efficiently capturing fluid-grain displacement patterns. In response, we present a novel two-way coupled hydro-mechanical discrete-element method model that manages fast fluid transport and considers the synchronising pore deformation. This model, which employs the implicit finite volume approach to solve pore pressure under a remarkable timestep, predicts fluid-fluid and fluid-grain interactions unconditionally stable. We design a pressure-volume iteration scheme to balance injection-induced pressure changes with pore structure rearrangements dynamically. Additionally, we incorporate adaptive flow front advancement criteria to enhance the capture of interface motion, particularly in complex flow scenarios where fluid migration surpasses the frontline pores or is impeded by capillary effects. The robustness and reliability of our model, validated against Darcy flow theory and experimental observations from Hele-Shaw tests, demonstrate its capability in accurately analysing multiphase fluid migration and dynamic fluid-grain interactions in porous media. We are confident in considering this model a powerful tool to illustrate the micro-mechanisms of multiphase flow in deformable porous media.

        Speaker: Mr Quanwei Dai (PhD Candidate)
      • 177
        Direct numerical simulation of CH4 - CO2 mixture flow in nanoporous media

        CO2 injection into shale is believed beneficial for both enhanced gas recovery and CO2 storage. The confined space and strong solid-molecule interactions in nanoporous media lead to different occurrence states of CH4 and CO2, causing the flow of CH4 - CO2 mixture in shale to deviate from predictions of continuum models. In this study, we employed a modified pseudo-potential based lattice Boltzmann (LB) model to study gas mixtures in nanoporous media. The mixed equation of state is used to calculate interaction force between gas mixtures. The solid-molecule interaction force is determined by comparing density profiles from LB simulation and molecular dynamics. The proposed method can model the flow of CH4-CO2 mixture in complex topological nanopores with various surface properties. Our results demonstrate that the Langmuir model and BET theory cannot accurately describe the adsorption isotherms of CH4 and CO2 in nanoporous media. The transport capacity of CH4-CO2 mixture in nanoporous media is found affected by surface properties. In organic nanoporous media, CO2 molecules tend to accumulate near the pore surface, hindering their flow compared to CH4 molecules. In contrast, inorganic nanoporous media facilitate the flow of both CH4 and CO2 molecules. We propose a modified apparent permeability model to describe the flow capacity of a CH4-CO2 mixture in nanoporous media.

        Speaker: Mr Chenyue Xie (China University of Geosiences (Wuhan))
      • 178
        Dynamic X-ray computed microtomography imaging of multiphase flow in porous media using deep learning

        Dynamic imaging of multiphase flow in porous media using X-ray microcomputed tomography (micro-CT) has been a technique exclusive to synchrotron-based systems. With the emergence of deep learning, however, the lower X-ray flux from a standard micro-CT system, and thus lower signal and higher noise under dynamic imaging conditions can be compensated for by use of convolutional neural networks with a priori knowledge of the imaged domain and the noise signature.

        In this work, a cycle consistent generative adversarial network (CycleGAN) based on the principle of unpaired image-to-image translation is utilized for transforming noisy micro-CT images to clean images. The two main objectives of this study are to assess the levels of noise that would be prevalent during dynamic imaging, and to design a DL network to denoise these images. To obtain the relevant data, fast and slow scans were performed at set saturation levels in the sample. The examined two-phase flow system consisted of air and water (the latter doped with potassium iodide – KI). The sample in consideration was a Bentheimer sandstone sample and the experiment was conducted with a custom-built benchtop micro-CT system located at Oregon State University. To obtain ground truth (GT) images for the training of the CycleGAN model, a high-quality dry scan of the sample was acquired before the KI doped water was injected. Once the fluid was injected, subsequent fast scans were acquired, and finally, a high-quality multiphase scan was captured.

        The results from denoising micro-CT scans indicate that the proposed workflow is robust and capable of improving the quality of images with good accuracy and ease of implementation. The fastest scan was conducted at 1min43s while the high-quality scans were acquired at 1hr24mins. Three fast scans with varying scan times of 1min43s, 2min45s, and 3min26s were tested. It was observed that when subjected to the CycleGAN network, the denoised images of 1min43s were adding features (often called hallucinations) in the generated results indicating that the images were too noisy as a starting point. On the other hand, 2min45s and 3min26s scans showed promising results. The accuracy of denoising was then validated by pixel-wise accuracy of the segmented denoised images.

        In multiphase flow imaging, it is often not practical to acquire paired images for denoising. With the implementation of cycleGAN, the proposed research not only enhances image quality, but also indicates that the acquisition time can be decreased by more than 25 times from hours to minutes (as low as 2min45s) for dynamic imaging in standard benchtop systems.

        Speaker: Ryan Armstrong (University of New South Wales)
      • 179
        Effect of elevated-temperature on mechanics and microstructure of basalt fibre-modified cementitious composites

        Cementitious materials, known for their brittle nature, are often vulnerable to thermal degradation in deep underground and hydrothermal environments. Basalt fiber (BF), an inorganic silicate additive used in cement, has garnered significant attention due to its outstanding mechanical and thermal resistance properties. However, key experimental data are scarce on the role of BF in cement exposure at elevated-temperature conditions, and a lack of understanding of the key mechanisms of reinforcement processes. In this study, we investigate the mechanical behavior of cementitious composites with varying BF content under both ambient and elevated temperature conditions (up to 200°C) to determine the optimal dosage. Microstructural analysis and phase composition assessments are conducted to reveal the mechanisms of reinforcement, encompassing the state of the cement matrix, BF itself, BF-matrix interaction, and phase evolution. The results indicated that the addition of 0.05-0.1 wt.% BF to cement can significantly enhance its mechanical strength and crack resistance, with flexural strength improving by up to 60% after exposure to 200°C for 6 days. This enhancement is primarily attributed to multiple energy-consumption processes, including the bridging effect, breaking effect, pulling-out effect, crack deflection effect and adhesive effect. Notably, the adhesion between BF and the cement matrix improves after exposure to 200°C, leading to the formation of "network-like and granular hydration products" on the surface of BF. While needle-like hydration products are commonly observed under ambient conditions, contributing to a slight increase in the surface friction of BF, the accelerated growth of hydration products at higher temperature conditions emerges as the predominant factor enhancing mechanical strength and improving ductility of cement composites at elevated temperatures. The insights from this study provide promising prospects for the application of BF-modified cement composites in elevated-temperature environments, including deep geothermal wells, CO2 storage wells, and various other deep geo-energy applications.

        Speaker: HAO HAN (University of Manchester)
      • 180
        Effect of pore size of electrospun membrane on quality and ion separation of nanofiltration membrane

        In this research, the effect of the pore size of the electrospun membrane in the preparation of a three-layer thin film nanofiber composite membrane (TFNC) was investigated. Due to its special properties, such as high porosity and the ability to produce pore sizes ranging from tens of nanometers to several micrometers, along with different mechanical properties, it finds wide applications in various fields, including medicine and health (i.e., tissue engineering, drug delivery, protective clothing, and biosensors), environment (air and water filtration membranes), energy (solar cell, battery fuel) and makes the use of electrospun membranes highly promising in separation technology.
        The three-layer membrane comprised a first layer of mesh-shaped polyester and a middle layer of a substrate consisting of hydrophobic polysulfone with a concentration of 20% by weight. The middle layer was produced by electrospinning with varying pore sizes. The third layer was a polyamide layer formed through interfacial polymerization between piperazine monomers (2wt.%) and trimesoyl chloride monomers (0.2wt.%). The polyamide layer and polysulfone fibers were characterized using infrared spectroscopy (FTIR), scanning electron microscope (SEM), bubble point, and MgSO4 divalent ion separation.
        Based on the FTIR test, peaks of 1618 and 2990 were observed, indicating the presence of the polyamide layer and polysulfone substrate, respectively. The electrospinning was conducted under constant conditions, including a voltage of 17 kV, a needle-to-collector distance of 120 mm, and a variable polymer injection rate set at 2, 1.2, 0.8, and 0.5 ml/h. The diameter of the fibers was measured using SEM images (0.11 ± 1.25, 0.45 ± 0.9, 0.37 ± 0.58, and 0.12 ± 0.3 micrometers), and the pore sizes of each substrate were measured as 9.3, 7.1, 3.5, and 1.1 microns by bubble point. The MgSO4 salt separation test was conducted on membranes with various pore sizes and fiber diameters after the coating process. In this experiment, the separation percentage for MgSO4 divalent salt was measured as 0%, 23%, 51%, and 83%, respectively. The separation of MgSO4 ions increased with the reduction of the pore size.
        Nanofiltration is a relatively recent separation process that has found widespread applications in the chemical and environmental industries due to its lower energy consumption and higher flux. In this study, we investigated the effect of the pore size of the electrospun layer. It was observed that the average diameter of the electrospun membrane fibers has a direct relationship with the pore size. As the diameter of the fibers decreases, the pore space also becomes smaller. Subsequently, the layer uniformity of polyamide is enhanced on the electrospun membrane, leading to a higher separation rate of bivalent ions.

        Speaker: Masoud Riazi (School of Mining and Geosciences, Nazarbayev University, Kabanbay Batyr 53, Astana, Kazakhstan, 010000)
      • 181
        Evaluating the performance of asphalt mixture with additives to withstand salt erosion and freeze-thaw cycles

        Asphalt pavement is widely used in road construction due to its smoothness, wear resistance, and ease of maintenance, making it the most commonly chosen material. However, asphalt concrete exposed to the natural environment is susceptible to various external factors, resulting in different types and degrees of damage. This greatly shortens its service life and reduces the durability of asphalt pavement. Especially in coastal regions with seasonal freezing, salt spray, tides, and rainfall can all contribute to the penetration of salt into asphalt pavement through porous media. The inner salt will accumulate and dissolve, weakening the cohesiveness of the asphalt-aggregate interface and causing diseases such as spalling and pitting. In addition, the seasonal freeze-thaw cycle causes early damage to the asphalt mixture. Admixture is an effective means of improving the performance of porous asphalt mixtures. Efforts done by previous researchers have shown that ferrocyanide can inhibit salt crystallization and reduce salt erosion damage of porous materials. Basalt fiber and anti-stripping agent can improve the service performance of asphalt mixture. However, the effects of three additives on salt erosion and freeze-thaw coupling environments have not been studied, and their specific improvement effects are unclear.
        This paper focuses on the damage resistance of asphalt mixture with crystallization inhibitor, basalt fiber and anti-spalling agent under salt erosion and freeze-thaw, respectively. The splitting tensile strength of asphalt mixtures with additives was tested after 15 cycles of salt erosion and freeze-thaw. We used X-ray computed tomography to analyze the initial internal structure after 0, 7, and 20 freeze-thaw cycles, as well as salt erosion. The effects of salt erosion and freeze-thaw cycles on asphalt mixtures with additives were evaluated through the analysis of changes in strength and internal structure. Based on the changes in internal structure, the performance enhancement of porous asphalt mixtures with additives was assessed using grey relation analysis and analytic hierarchy process analysis.
        Results showed that the damage caused by salt spray erosion on the mechanical properties of asphalt mixture is greater than that of salt solution erosion. Compared to the AC-13 asphalt mixture, the SMA-13 asphalt mixture has better mechanical properties against the effects of freeze-thaw cycles and salt erosion. The performance improvement of asphalt mixture with a crystallization inhibitor is better than that of an anti-stripping agent and a basalt fiber asphalt mixture, respectively. The three types of additives only retard the evolution of internal structure and do not change the damaged formation of internal structure inside asphalt mixture under freeze-thaw cycles and salt erosion. The internal damage index of asphalt mixture with crystallization inhibitor is the minimum, indicating good resistance to freeze-thaw cycles and salt erosion. The combination of SMA-13 gradation and crystallization inhibitor was found to be suitable for designing asphalt mixtures in coastal seasonal frozen regions. The research results will help improve the level of construction and maintenance of asphalt pavement in coastal seasonal freezing areas, extend its service life, and save on maintenance costs.

        Speaker: Weidong Ji
      • 182
        Experimental and theoretical study of unsaturated flow in fractured media

        Unsaturated flow in fractured media is an important process with relevance to a large number of industrial and environmental application. In this work, we report recent experimental and theoretical investigations on unsaturated flow in single fractures, fracture intersections, and fracture networks. Wo focus on how small-scale flow physics influences the spatial and temporal characteristics of unsaturated flow in discrete fracture networks. We propose theoretical models for predicting water splitting at fracture intersections and for predicting water breakthrough time in an unsaturated fracture network. We validate these models by comparing with experimental observations. We show that the breakthrough time in a fracture network decreases with the increase of initial saturation. We also find that avalanche infiltration mode, i.e., sudden release of a large amount of water from the network, emerges spontaneously in the network, and is modulated by the local splitting behavior. We further show that the power spectral density of the water saturation time series in the network follows a power law with an exponent of −2 for all simulations with different structural parameters and local flow rules, suggesting a universal self-organized criticality behavior for unsaturated flow in fractured rocks.

        Speaker: Zhibing Yang (Wuhan University)
      • 183
        Experimental evaluation of dynamic seepage in tight/shale reservoirs under the coupling of matrix fractures based on NMR

        Meng Du1,2,3, Shuyi Lu4, Zhengming Yang*1,2,3, Weifeng Lyu1,2,3, Xinliang Chen2,3, Xiang Qi 3, Pengwei Fang 1,3, Zhuoying Dou 1,3
        (1. University of Chinese Academy of Sciences, Beijing 100049, China;2. Institute of Porous Flow & Fluid Mechanics, Chinese Academy of Sciences, Langfang 065007, China; 3. Research Institute of Petroleum Exploration & Development, PetroChina, Beijing 100083, China; 4. Beijing Normal University, Beijing 100875; 5. State Key Laboratory of Enhanced Oil Recovery, Beijing 100083, China)

        The imbibition and displacement between fractures and matrix have a significant effect on the development of tight/shale reservoirs, a combination of dynamic displacement and imbibition online physical simulation method was established by integrating nuclear magnetic resonance (NMR) and CT scanning. Through real-time dynamic monitoring of multiphase flow and migration behavior of crude oil in each stage of dynamic imbibition, the development effect of dynamic imbibition and the micro-production mechanism of pore throats with different sizes of tight/shale oil were quantitatively studied. The effects of displacement pressure, permeability, and fractures on the dynamic imbibition effect and pore crude oil production were analyzed. On this basis, the dynamic seepage process of fracking-soaking-backflow-production integration was simulated, which reveals the dynamic production characteristics of different development stages and their contribution to enhancing oil recovery (EOR). The results show that the dynamic imbibition process of tight/shale oil water flooding can be divided into three stages: strong displacement and weak imbibition stage of rapid production of large pores and fractures under displacement action; weak displacement and strong imbibition stage of slow production of small pores and fractures under counter-current imbibition action and dynamic equilibrium stage of weak displacement and weak imbibition. The greater the displacement pressure, the lower the degree of imbibition recovery and the stronger the contribution of displacement, but it is easy to produce water channeling, leading to an early breakthrough, as a result, the recovery increases and then decreases. The higher the permeability and the better the pore throat connectivity, the greater the degree of both imbibition and displacement recovery, and the shorter the percolation equilibrium time and the greater the recovery. Fractures can effectively increase the imbibition contact area between the matrix and water, reduce the resistance of oil and water seepage, and increase the rate of matrix oil release and total recovery. There are differences in dynamic production characteristics and the degree of contribution to recovery at different development stages. Conducting a soaking program after fracturing is beneficial for fully utilizing the effects of fluid imbibition, displacement, and energy storage; also, the key to EOR is to effectively utilize the carrying effect of the backflow fluid and the displacement during the production stage. This study provides theoretical support for the efficient development of tight/shale oil.

        Speaker: Dr Meng Du
      • 184
        IMPACT OF DUAL POROSITY SYSTEMS ON FLOW IN HEAP LEACHING USING MICRO COMPUTED TOMOGRAPHY IMAGING

        The global progress with the energy transition from fossil fuels to renewable energy has boosted the demand for metals. Copper, amongst all metal resources, is regarded as an essential raw material in producing the conduit to reduce the energy needed during electricity production. Most of the low-grade copper is recovered from heap leaching, which is a well-established hydrometallurgical method.
        However, a low recovery remains a significant challenge of this method. This poor leaching performance is caused by the variation of the leaching kinetics driven by not only the leaching fluid transport at the particle scale but also the mass transport at the grain scale. Previous studies have implemented a collection of numerical and experimental methods to investigate the leaching performance under a variety of scales and leaching conditions but rarely investigated the process using a multiscale approach.
        Herein, we implemented a fundametal study the flow behaviour in a dual porosity system, which is by visualising the fluid transport and measuring the liquid contents under the particle and grain scale. A series of leaching experiments is conducted using water with packed glass beads and chalcopyrite beads. We image the column in regular periods under Micro Computed Tomography (Micro-CT) imaging over several days of the leaching period. We label and track mineral grains from high-resolution images and monitor the alternation of the mineral size and porous media structure. Our study shows the impact of dual porosity systems on flow patterns and the performance of leaching on mineral grain within different particles. Our results highlight the importance of multiscale simulation for the design and optimisation of heap leaching.
        Keywords: Heap leaching, dual-porosity, Micro-CT, column test

        Speaker: Quan Zheng (Mr)
      • 185
        Inferring electrochemical performance and parameters of Li-ion batteries based on deep operator networks

        Li-ion battery is a complex physicochemical system that generally takes observable current and terminal voltage as input and output, while leaving some unobservable quantities, e.g., Li-ion concentration, for serving as internal variables (states) of the system. On-line estimation for the unobservable states plays a key role in battery management system since they reflect battery safety and degradation conditions. Several kinds of models that map from current to voltage have been established for state estimation, such as accurate but inefficient physics-based models, and efficient but sometimes inaccurate equivalent circuit and black-box models. To realize accuracy and efficiency simultaneously in battery modeling, we propose to build a data-driven surrogate for a battery system while incorporating the underlying physics as constraints. In this work, we innovatively treat the functional mapping from current curve to terminal voltage as a composite of operators, which is approximated by the powerful deep operator network (DeepONet). Its learning capability is firstly verified through a predictive test for Li-ion concentration at two electrodes. In this experiment, the physics-informed DeepONet is found to be more robust than the purely data-driven DeepONet, especially in temporal extrapolation scenarios. A composite surrogate is then constructed for mapping current curve and solid diffusivity to terminal voltage with three operator networks, in which two parallel physics-informed DeepONets are firstly used to predict Li-ion concentration at two electrodes, and then based on their surface values, a DeepONet is built to give terminal voltage predictions. Since the surrogate is differentiable anywhere, it is endowed with the ability to learn from data directly, which was validated by using terminal voltage measurements to estimate input parameters. The proposed surrogate built upon operator networks possesses great potential to be applied in on-board scenarios, since it integrates efficiency and accuracy by incorporating underlying physics, and also leaves an interface for model refinement through a totally differentiable model structure.

        Speaker: Qiang Zheng
      • 186
        Integration of Digital Core and Molecular Simulation for Research on Reservoir Mechanical Properties

        The mechanical parameters of reservoir rocks play a crucial role in the evaluation of unconventional reservoirs. However, due to time and cost constraints, conducting rock mechanics experiments on all formations is not feasible. Furthermore, conventional testing methods may not accurately characterize and test rocks with complex micro-structures, such as tight sandstones and shales, where pore-throat connectivity is intricate. Therefore, this study proposes a multi-scale coupled simulation approach for reservoir rock mechanical parameters. Firstly, digital core models were constructed through CT scans of four core samples obtained from the Shengli oilfield. The analysis of thin section observations, combined with the machine learning approach of random forest, was used to classify mineral phases in the cores, such as Feldspar, Quartz, Mica, and Calcite. These mineral phases were well distinguished within the core. Subsequently, crystal lattice models of these minerals were established at the molecular scale, and mechanical parameter simulations based on molecular modeling were performed. The finite difference method was employed to calculate the stress under various strain states, resulting in the Young's modulus and Poisson's ratio for each mineral component. At the core scale, finite element simulation of uniaxial compression tests was conducted using the mechanical parameters obtained from molecular simulations, to calculate the Young's modulus and Poisson's ratio of the entire core. The model established by this method accurately reflects the fine structure and heterogeneity of the rocks, thus improving the simulation accuracy and faithfully representing the mechanical behavior of rocks in experiments. Comparison with actual uniaxial compression test results indicates an average Young's modulus error of 2.44 GPa and an average Poisson's ratio error of 0.036, demonstrating that this simulation method effectively predicts rock mechanical properties, reduces the time and cost required for experiments, and retains the majority of the accuracy of the results obtained.

        Speaker: Mr Yifan Yin (s22020114@s.upc.edu.cn)
      • 187
        Investigation of single particle crushing characteristics considering non-spherical shape based on DEM

        The evaluation of fracture construction processes hinges on the critical factors of conductivity and the validity period of artificial fractures. It is imperative not to overlook the conductivity damage resulting from the crushing of proppant particles. Being a specialized geotechnical granular material, proppant particles undergo fragmentation when the applied compressive load surpasses their bearing capacity. The smallest unit in breakage behavior, the single particle, serves as the foundation for pertinent mechanical research.
        Quartz-sand proppant particles pose a challenge for classical theories due to their complex structures and irregular shapes. Existing studies often neglect the intricate structural shape of the proppant and the complex stress environment in the reservoir, opting instead for investigations based on regularly shaped particles. This study delves into the crushing behavior of irregularly shaped quartz sand single particles, determining the critical conditions of single-particle breakage through a combined approach of numerical simulation and experiments. A discrete element model (DEM) for the crushing of quartz sand single-particles under closure pressure and confining pressure was established. The analysis of crushing characteristics involves examining the dynamic distribution of the crushing belt and stress-strain curves. Additionally, the primary controlling factor influencing the crushing behavior of single particles is investigated.
        The results reveal a unimodal distribution in the stress-strain curve for proppant particles. Smaller particle sizes correspond to higher effective peak values of fragmentation and smaller strains. Specifically, under consistent parameter conditions, 40–70 mesh quartz sand particles exhibit stress peaks over twice as high as those of 8–16 mesh particles. Larger particle sizes harbor a greater number of internal natural cracks and defects, thereby diminishing the particles' bearing capacity. The influence of particle sphericity on crushing patterns and crack locations is governed by the mode of contact between particles and the wall surface. Particles with high roundness are more likely to be broken into two parts. Concurrently, as particle irregularity intensifies, secondary cracks emanate within the particles prior to complete fragmentation, resulting in a stress-strain curve exhibiting a multi-peak distribution. For a consistent 8–16 mesh of quartz sand particles, the peak crushing force under conditions of high roundness surpasses that under conditions of low roundness by approximately 15 N. Notably, the crushing strain is significantly greater, nearly 10%, for small-sized particles compared to large-sized particles under conditions of low sphericity, whereas this difference diminishes to 1.8% under conditions of high sphericity. The experimental approach was used to validate the simulation results, immensely contributing to providing theoretical underpinnings for the application and optimization of proppants.

        Speaker: Dr Xiangyu Wang (School of Petroleum Engineering, China University of Petroleum (East China))
      • 188
        Microscopic damage rules of water flooding in ultra-low permeability reservoir: an experimental study based on the combination of microfluidic and low-field NMR technology

        Water flooding is one of the important ways of oilfield development in China. However, due to the poor-compatibility between injected water and reservoir rocks and fluids, free particles such as inorganic scale and rock clay often block the pores and throats under the carry-over effect of water seepage, which leads to engineering problems such as water injection pressure increasing, underinjection of injection-wells and decreased oil production capability of production-wells. Conventional core experiments and numerical simulations were used to study the damage induced by water injection. Since the black-box nature of the core, researchers cannot observe the fluid flow rules in porous media, nor can they know the specific location and degree of core damage. Although the numerical simulation method can simulate the scale formation and the occurrence of scale particles, it ignores the real situation of scale particles being carried and moved by fluid, and can not obtain the real rules of water flooding damage to the reservoir.
        Considering the shortcomings of previous studies, we combined low-field NMR and microfluidic techniques to explore the damage rules of injected water to the core of an ultra-low permeability reservoir in northwest of China. Firstly, low-field NMR online displacement technology was used to detect the signal intensity changes of the core along the axial direction, as well as the pores of various sizes during the water flooding process, and we obtained the injury rules of the core at different locations during the water flooding process. Further, we used μ-CT to scan the core slice and obtain its pore distribution information, and etched it on the glass chip for displacement experiment, and observed the damage rules of the scale particles generated by the combination of injected water and formation water on the porous media in real time.
        It is found that free particles are selective to reservoir damage. Due to the low-resistance flow characteristic of fluid, free particles will be carried by fluid to the micro-fractures and large pores, causing greater damage. This kind of damage has mobile characteristics, free particles will be carried to the center and tail end of the core, and increase the blockage these parts over time. The microfluidic experiment further confirmed the rules obtained by low-field NMR method, and the accumulation of scale particles in the middle and rear of the chip was also found in the microfluidic displacement experiment. The particles size of these areas continues to increase, which increases the damage to the middle and rear of the reservoir, blocks the flow channel of the production wells, and seriously reduces the swept efficiency of the displacement phase.

        Speaker: Dr Yiping Wen (Institute of Rock and Soil Mechanics, Chinese Academy of Sciences)
      • 189
        Microscopic mechanism investigation of counter-current imbibition in tight reservoirs using the Lattice Boltzmann method

        Tight oil reservoirs are typically developed through hydraulic fracturing to create a network of fractures, with counter-current imbibition between fractures and matrix playing a crucial role. However, during the counter-current imbibition process, as water displaces oil, it increases water saturation and leads to water blocking phenomena, resulting in reduced oil relative permeability and heightened flow resistance in the oil phase. The oil relative permeability is pivotal in determining the matrix's ability to elastically expel oil, rendering counter-current imbibition with both advantages and disadvantages. To quantitatively characterize the damage caused by counter-current imbibition in tight reservoirs, we employed the lattice Boltzmann method to simulate counter-current spontaneous imbibition in Jimsar tight rocks. We also introduced a novel method for assessing reservoir damage: the water-to-solid transformation method.

        During counter-current imbibition, the water phase occupies the flow space of the oil phase, acting as a solid, consequently diminishing the oil's flow capacity. Given the challenge of calculating the relative permeability of two phases during counter-current imbibition, we transformed the water phase into a solid phase to construct a new digital core. This transformation allowed for the computation of unidirectional permeability, which was then used to characterize the reduction in oil relative permeability during counter-current imbibition. The study elucidates that the reduction in oil relative permeability in the digital core of Jimsar during counter-current imbibition is limited, ultimately establishing a relatively stable oil flow channel. This method facilitates the evaluation of the effectiveness of counter-current imbibition in reservoirs, enabling the implementation of appropriate improvement measures. Furthermore, a comparison between counter-current and co-current imbibition indicates that the water phase in counter-current imbibition only affects the oil phase near the fracture, without reaching the interior of the matrix. Nevertheless, the imbibition front of counter-current imbibition remains relatively stable, without significant fingering.

        Speaker: Prof. Shuyang Liu (China University of Petroleum (East China))
      • 190
        Multi-scale Pore Structure Characteristics of Deep Marine Shale and Its Controlling on Gas Transport Mode: Silurian Longmaxi Formation in Southern Sichuan, China

        The attributes of shale pore structure govern the accumulation, presence, and dissipation of gas. Deep marine shale undergo intricate geological evolution, with pore development at the nanoscale. Consequently, quantifying the impact of deep shale pore structure on gas is challenging. In this paper, the microscopic storage space structure of deep shale is quantified, and the correlation between pore structure and mode of gas transport is established. The study focused on the Silurian Longmaxi Formation shale, utilizing techniques such as SEM, CO2 and N2 adsorption, HPMI, and the Frenkel-Halsey-Hill method approach to quantify the development characteristics and controlling factors of pores at multiscale. Based on the pore structure, fractal, and molecular dynamics of methane, the numerical correlation linking pore structure and gas transportation mechanisms was established. The results indicate that the deep marine shale formations are predominantly composed of ORSS and ORMS. The evolution of pores is impacted by the TOC content and mineral composition. Based on the governing function of pore structure in gas transport mechanisms, the pores can be classified into three categories: ultramicro adsorption pore, nano-diffusion pore, and micro-nano flow-diffusion pore. Correspondingly, there are five types of gas transport modes: surface adsorption diffusion, Knudsen diffusion, Fick diffusion, slip flow, and continuously flow. These diverse modes collectively form a complex gas transport network. Deep shale exhibits a greater abundance of micropores and mesopores compared to shallow shale. The contribution of micro-fractures to shallow shale gas transport is crucial, and the contribution to deep shale decreases significantly. In conclusion, the favorable exploration lies in targeting ORSS formations with low D1 (fractal dimension) and high D2 (fractal dimension), as well as ORMS segments with high D1 and high D2. Specifically, within hydrocarbon-rich basins located below 3500 m, it is recommended to search for overpressure regions with weak structural deformation. These areas hold potential for successful gas exploration. This research establishes a basis for the exploration, development, and geological principles of deep shale gas.

        Speaker: Shijie He (China University of Petroleum-Beijing)
      • 191
        Multiphase Flow Through Rough Porous Layers in Proton-Exchange Membrane Fuel Cells (PEMFCs)

        Polymer electrolyte membrane fuel cells (PEMFCs) have emerged as ideal energy-conversion devices for hydrogen energy applications. The performance of PEMFCs is significantly affected by the accumulation and transport of water inside porous components and flow channels. Here, we focus on investigating the role of surface roughness on the fluid transport and droplet impact behaviours in the porous components of PEMFC. We start by examining the fluid transport characteristics at the interfacial region of microporous layer (MPL) and catalyst layer (CL), considering the effects of compression stress, porosity, and wettability. Liquid and gas permeabilities are also investigated to assess water drainage and fuel supply efficiency with different compression conditions. Then, surface roughness effects of gas diffusion layer (GDL) on liquid droplet removal inside a flow channel are investigated. We simulated the complete process of droplet removal in flow channel, including emergence, growth, detachment, and removal. We also identified different regimes of detachment modes based on the droplet breakup location and detachment ratio. Finally, we experimentally examined the liquid droplet impact dynamics on rough surfaces with various topological parameters. We observed different modes of droplet spreading and bouncing behaviour, and droplet impact outcome transition from bouncing to no bouncing is identified. To quantify the influence of surface wetting area on the bouncing-wetting transition, we proposed a modified Weber number that incorporates a combined effect of droplet kinetic energy, surface energy, and adhesion force. We found that the droplet impact outcomes in the transition regime can be accurately described by a single curve as a function of the modified Weber number. The results can provide valuable insights into selecting appropriate parameters of diffusion media for optimising water management and fuel supply during fuel cell operation processes.

        Speaker: Yixiang Gan (The University of Sydney)
      • 192
        Neural Operator Predictions of Electrical Properties in Porous Media

        The characterization of electrical transport properties through porous media is commonly used for reservoir formation evaluation, groundwater management, mineral exploration, and carbon capture and storage (CCUS) monitoring. We discuss image-based characterization and prediction of a range of transport properties including, but not limited to, flow/momentum transport, mass transport as well as electrical properties, often in the presence of multiple fluid phases. We then present novel data science methods and their impacts on prediction speed and accuracy. The data used in this work is openly available as part of a benchmark dataset that includes complete geometric characterization and simulations hosted on Digital Rocks Portal1.

        Deep learning (DL) has emerged to be a powerful tool for modeling these transport properties while reducing computation time by several orders of magnitude. Previously, we have applied convolutional neural networks (CNNs) to predict the electric potential field given a segmented x-ray microtomographic image. We found that the set of input features used to predict momentum and mass transport properties did not translate well to electrical transport and had to be redesigned. We further found that because classical DL networks (including CNNs) typically provide a mapping between Euclidean spaces (e.g. image vector-to-scalar quantity, image vector-to-image vector, etc.), they effectively only learning a single instance of the governing equation. If the discretization of the partial differential equation (PDE) changes, the model needs to be retrained or can otherwise lead to issues when applying the model to new domains. Here, we present new work on overcoming that limitation by training neural operators.

        Neural operators are a new subclass of scientific machine learning in which data-driven models learn mappings between function spaces, thus learning entire families of PDEs. Here, we apply a Fourier neural operator network to predict the electric potential through a 3D porous medium and contrast them to classical convolutional neural networks (CNNs). The results show improved accuracy over previously trained CNNs, particularly when introducing new types of porous media. Further, we recognize that training a neural operator network typically requires large amounts of high-fidelity data, and despite the existence of open data portals, such as Digital Rocks Portal, this is nevertheless a bottleneck. To address this, we explore the development of a multiscale neural operator network. We expect that the initial time-intensive training will be compensated by its ability to generalize to new parameterizations and other transport problems in porous media.

        Speaker: Bernard Chang (The University of Texas at Austin)
      • 193
        Novel Learning-based Pattern-Data-Driven Forecast Approach for Predicting Future Well Responses

        Accurately predicting future well response is crucial for making investment decisions in developing subsurface reservoir resources. Future well responses have been forecasted using history-matching methods, but traditional history-matching methods often incur high computational costs in calibration steps and have difficulties in maintaining geological constraints. Recently, data-driven forecast methods like data-space inversion (DSI) and learning-based data-driven forecast approach (LDFA) have been introduced to mitigate the computation cost and geological constraint issues of history-matching methods. However, DSI and LDFA have extrapolation, conditioning, and prediction variance issues. In this study, we propose two simple approaches, a learning-based pattern-data-driven forecast approach (LPFA) and an ensemble conditioning step (ECS), to resolve the issues associated with DSI and LDFA. For the extrapolation issue, LPFA provides accurate predictions by scaling prior data using observed data, even when the observed data are outside the prior data range. ECS resolves the conditioning and variance issues of LDFA and LPFA by selecting the predictions of ensemble learning that honor the observed data. The prediction performances of DSI, LDFA, LPFA, and ECS are compared using two well-known benchmark models (Brugge and Olympus models). All the forecast approaches showed reliable performance in predicting future well responses, achieving an average median relative error of 2~3% under a 2% noise level of the observed data. However, LPFA was the only approach that provided the most accurate predictions for future well responses when the prior data did not contain data close to the observed data. ECS improved the prediction performances of LDFA and LPFA in both the Brugge and Olympus models as it selects only the predictions from multiple learning models honoring the observed data.

        Speaker: Yeongju Kim
      • 194
        Oscillation Method for Measuring Gas Storage in MCM-41

        The total storage of gas in a mesoporous material MCM-41 with adsorption/capillary condensation was measured using a new oscillation-based method. With an improved setup and procedure, the accuracy of the measured isotherm was significantly improved. Experiments were conducted using both condensable (propane and carbon dioxide) and non-condensable (argon and methane) gases. The results show that this method can be used to measure not only the total storage of gas but also the excess due to adsorption/capillary condensation with considerable accuracy. In experiments conducted with propane, the occurrence of capillary condensation, which increased the total storage of gas, was well depicted.

        Speaker: Xiaolong Yin (Eastern Institute of Technology, Ningbo)
      • 195
        Pore-scale and Reservoir-scale Investigations on H2 Trapping: Impact of Temperature and Salinity

        Underground H2 storage is arguably considered one of the promising techniques to achieve net-zero emissions goals. The storage of H2 in geological formations is influenced by a complex function of the physicochemical, petrophysical, and geo-mechanical characteristics of the H2/brine/rock system. This results in the existence of different trapping mechanisms (e.g., residual and dissolution trapping), which will lead to the loss of H2 within the formation. Therefore, it is important to understand the trapping of H2 at different scales to provide a better understanding of the H2 withdrawal efficiency.

        To this end, contact angles and interfacial tensions of the H2/brine/sandstone at different temperatures and salinities were collected for the pore network modeling to investigate the pore-scale H2 trapping behaviors. The obtained results were then fitted using the Land trapping model. Subsequently, these trapping behaviors and different H2/brine properties, including density, viscosity, and H2 solubility under different conditions, were then considered in the field-scale simulations. One single injection and production well and four annual injection-withdrawal cycles were considered. The results indicate that a higher temperature leads to less residual trapping in both pore-scale and reservoir scale. The higher temperature and higher salinity conditions are more favorable for H2 production (i.e., a higher H2 recovery factor). In addition, the H2 dissolution trapping is also influenced by the temperature and salinity, which contribute to a maximum of ~5% H2 loss. Furthermore, the H2 plume migrations are also influenced by different temperature/salinity conditions.

        Speaker: Mr Haiyang Zhang (Chemical and Petroleum Engineering Department, Khalifa University)
      • 196
        Pore-scale Modeling and Numerical Simulation for Viscoelastic Emulsion Flow

        In this study, a body force driven two-phase fluid flow in porous media is numerically simulated by weakly compressible smoothed particle hydrodynamics, a Lagrangian mesh free particle method. The dispersed phase consists of viscoelastic emulsive droplets and is assumed to obey Oldroyd-B rheological model. The background phase is a Newtonian fluid. The interfacial tension between two phases and the wettability between fluids and solid boundaries are evaluated by pairwise force model. Different flow velocity, viscosity ratio and Weissenberg number are investigated. All simulations are implemented on GPU to achieve high efficiency.

        Speaker: Haoran Cheng (Qingdao University of Science & Technology)
      • 197
        Promoting Ultra-High-Density Nanoparticles Exsolution in Layered Perovskite Ferrites via a Facile Cobalt Doping Method: A High-Performance, Stable Anode for Direct Ethane Solid Oxide Fuel Cells

        Nanoparticles anchored on the perovskite surface have gained considerable attention for their wide-ranging applications in heterogeneous catalysis and energy conversion due to their robust and integrated structural configuration. Herein, we employ controlled Co doping to effectively enhance the nanoparticle exsolution process in layered perovskite ferrites materials. CoFe alloy nanoparticles with ultra-high-density are exsolved on the (PrBa)0.95(Fe0.8Co0.1Nb0.1)2O5+δ (PBFCN0.1) surface under reducing atmosphere, providing significant amounts of reaction sites and good durability for hydrocarbon catalysis. The morphology evolution measurements reveal a significant transform in CoFe alloy nanoparticles at around 600 °C, transforming from larger to ultra-densely decorated smaller nanoparticles. A single cell with PBFCN0.1 anode exhibits high performance in wet ethane atmosphere (e.g., a typical peak power density of 455 mW cm−2 at 800 °C), which is significantly improved by 36%-70% compared to the sample without Co doping. This discovery emphasizes how temperature influences alloy nanoparticles exsolution within doped layered perovskite ferrites materials, paving the way for more targeted material-specific research and broadening the spectrum of practical applications.

        Speaker: Mr Rubao Zhao (Shenzhen University)
      • 198
        Reactivity of porous media under continuous injection

        High flow rates within reactive porous media occur both in industrial applications and in natural media, e.g. in permeable soil substrates subjected to extreme weather events [1,2]. The reactive surfaces of the soil grains interact with the transported species, determining the overall porous media reactivity and capacity of retaining nutrient and contaminants. We show via pore-scale lattice-Boltzmann simulations in a packed bed column that, under the conditions of uniform injection, the uniformity of porous surface reactions is determined by the mixing of the low concentration wakes forming behind the reactive soil grains. Scaling arguments are proposed to extract analytical models for the probability distributions of concentrations at the reactive surfaces.

        Speaker: Kaj Pettersson
      • 199
        Research on the occurrence states of microscopic remaining oil in ultra-low permeability reservoirs

        The Chang 6 reservoir represented by Wangyao Area in Ansai Oilfield is the earliest developed reservoir in Changqing Oilfield. After nearly 40 years of exploration and development, the comprehensive water cut has reached 72.8%, and the degree of geological reserves recovery is only about 17.8%. The reservoir has entered the middle and late stage of development, and the reserve-production ratio has decreased year by year, and the problems of production decline and low recovery degree have become more and more serious. The effect of conventional water flood adjustment and cyclic water flooding are not obvious. So it is necessary to use polymer injection, phlogisticated air injection and carbon dioxide injection to enhanced oil recovery. However, due to the great differences in the micro-pore structure of the reservoir, the location and shape of the micro-remaining oil are different, and the corresponding tapping methods for different types of micro-remaining oil are also significantly different. Therefore, it is necessary to carry out systematic water drive experimental research on the ultra-low permeability reservoir whose micro-remaining oil occurrence characteristics and production mechanism are not completely clear at present. In this study, the ultra-low permeability lithology reservoir of Chang 6 formation in Ansai Oilfield was taken as the research target area. By using the dynamic displacement scanning technology to obtain two-dimensional image information of micro-plunger core in different water flooding stages, and reconstruct the occurrence states of micro-remaining oil in three-dimensional pore space, and the types of micro-remaining oil were divided and quantitatively calculated according to its formation mechanism and three-dimensional structural parameters. Furthermore, the types and dynamic changes of micro-remaining oil in different water drive stages are analyzed, and the potential of micro-remaining oil in different development stages and corresponding utilization methods are defined, in order to provide theoretical guidance for the smooth implementation of water drive and tertiary oil recovery in ultra-low permeability reservoir.

        Speaker: Dr Peng Chi (China University of Petroleum (East China))
      • 200
        Reshaping the Imaging Landscape: AI-Supercharged Swin Transformer for Unprecedented Detail

        Super-resolution imaging, a transformative technique spanning various scientific disciplines, holds the potential to revolutionize our understanding of complex porous structures within the realm of porous media and modeling. Traditional imaging approaches often struggle to capture the intricate details of porous media's intricate structures. To overcome this limitation, our research employs advanced AI-driven super-resolution methods, aiming to transcend inherent resolution constraints. Our goal is to bridge the gap between large-scale imaging methods, which excel at capturing macroscopic features, and small-scale techniques known for their detailed focus. By developing techniques to reconstruct high-resolution representations from lower-resolution inputs, our study promises a profound characterization of porous media's internal architecture, critical for applications such as filtration, oil recovery, and groundwater flow.

        Our research journey embarked with a meticulous exploration of AI-based super-resolution techniques. Initially, we employed the ESRGAN model with the rrdb net as the generator. While it exhibited commendable performance, the model's large size posed practical challenges. Subsequently, we transitioned to the SwinIR model, which delivered results characterized by their remarkable smoothness and sharpness. Building on this progress, we leveraged GAN-assisted Swin transformers, an approach that not only yielded exceptional outcomes but also presented a significantly reduced model size. This transition not only streamlined GPU memory usage during training but also accelerated the training process.

        Our methodology included in-depth examinations of various loss functions, highlighting the significance of a hybrid approach that combines GAN loss and pixel loss. This novel approach proved instrumental in effectively training the model and enhancing the quality of super-resolution results. Throughout our research, we meticulously studied the individual impact of each loss function and established relevant metrics, providing a robust foundation for our study's methodology.

        Beyond the advancements in super-resolution techniques, our research culminated in the development of a model capable of achieving super-resolution beyond current technological limitations. This breakthrough allowed us to push the boundaries of resolution in porous media imaging, and we validated our findings on new datasets with resolutions of 1μm, 4μm, and 16μm.

        The results were nothing short of groundbreaking. We observed that our AI-driven super-resolution approach consistently outperformed conventional methods, producing images with unprecedented clarity and detail. Notably, our model achieved remarkable results even at resolutions previously considered unattainable. This newfound capability opens doors to a wealth of possibilities in the fields of porous media analysis, offering insights that were once hidden from view.

        In conclusion, our research demonstrates the immense potential of AI-driven super-resolution in revolutionizing porous media imaging. By unveiling hidden details and pushing the boundaries of resolution, our findings hold promise for a wide range of applications, from enhancing filtration processes to optimizing oil recovery and improving our understanding of groundwater flow dynamics.

        Speaker: Mr Yang Meng
      • 201
        Study on the Distribution Patterns and Resistivity Characteristics of THF Hydrates in Sandstone Sediments

        Investigating the distribution patterns of hydrates in the pore spaces of sedimentary rocks and their resistivity characteristics is crucial for understanding and predicting the storage and extraction of natural gas hydrates. However, CT and resistivity joint measurements indicate that while the distribution of Tetrahydrofuran (THF) hydrates in sandstone sediments is similar to that in conventional oil and gas reservoirs, the resistivity measurements do not fully conform to the Archie model. This discrepancy arises because the formation of THF hydrates is a dynamic process which involves fluctuations in hydrate saturation, formation water salinity, temperature, and spatial distribution, making it challenging to accurately characterize these variables using traditional resistivity measurements. To address this issue and gain a deeper understanding of the complex petrophysical relationships in THF hydrate sediments, this study established digital rock based on the hydrate distribution from in-situ CT scans by segmenting the three-dimensional grayscale images of the sediments. The spatial distribution of hydrates within the sediment pores was then simulated using the Diffusion-Limited Cluster Aggregation (DLCA) model. Subsequently, the resistivity of sediments at various hydrate saturations is determined using the finite element method. The results show that at low hydrate saturations, hydrates exist in a dispersed form within the pores, while at high saturations, they are distributed in cementing or encapsulating forms, enhancing the structural integrity of the sediments. Furthermore, analysis of the aggregate morphology, including changes in gyration radius and fractal dimension, effectively reflects the evolution of the structure. The combined results from experimentation and simulation demonstrate that varying hydrate saturations significantly influence the distribution forms and their impact on resistivity: at low saturations, hydrates have a minor effect on resistivity, whereas at high saturations, extensive hydrate formation and distribution patterns lead to a rapid increase in resistivity. Overall, this study offers novel insights and methodologies for understanding the petrophysical characteristics.

        Keywords:THF hydrate formation, Resistivity measurements, Digital rock, DLCA model, Finite element method.

        Speaker: Mr Zizeng Li (China University of Petroleum East China)
      • 202
        Unraveling Heat Transfer Routes in Unsaturated Soils

        Thermal conductivity is a key parameter dominating the heat transfer process in a number of engineering applications such as geothermal exploitation, borehole thermal energy storage and carbon dioxide geo-sequestration. Since soil consists of multiple phases including solid particles, gas and/or water, not only the physical properties of these phases but also the routes that they form affect the heat transfer and reflect on the value of thermal conductivity. However, the heat transfer routes have not been properly identified in existing studies, which hinders the fundamental understanding of the physical mechanisms for thermal behaviour and accurate prediction of thermal conductivity in unsaturated soils.

        This work aims to identify the different types of heat transfer routes and study the contribution of each type to the effective thermal conductivity of unsaturated soils. Based on computed tomography (CT) images of five glass bead assemblies varying in degree of saturation, a series of image processing algorithms were applied to the images and identified six types of heat transfer routes: (1) particles in contact surrounded by air, (2) particles in contact surrounded by both air and water, (3) particles in contact surrounded by water, (4) separate particles connected by air, (5) separate particles connected by both water and air and (6) separate particles connected by water, respectively. In addition, the amount of each type was quantified. After reconstructing digital samples from CT images and computing their effective thermal conductivity, the contribution of each heat transfer route to the effective thermal conductivity of unsaturated soils was analysed. Results show that the contribution of types 3 and 6 increases while types 1 and 4 decrease during the increase of saturation. In addition, type 2 and 5 reach their maximum contribution when the degree of saturation is at 0.2.

        Speaker: Wenbin Fei (College of Civil Engineering, Hunan University, Changsha 410082, China)
    • MS01: 2.1
      • 203
        Evaluating the impact of Hysteresis and Heterogeneity on Hydrogen Storage Performance in Saline Aquifers

        The imperative to achieve net zero carbon emissions by 2050, aligned with global efforts to limit temperature rise, emphasizes the urgent shift to low-carbon energy sources. Hydrogen is identified as a key player in global decarbonisation however, concerns about the efficiency of hydrogen storage accompany its expanding production. This study investigates underground geological hydrogen storage in saline aquifers, emphasizing the impact of reservoir heterogeneity, flow function hysteresis, and injection/production flow rates on storage efficiency. The results show that hysteresis plays a crucial role in affecting storage efficiency, leading to significant entrapment and a lower recovery factor in initial production cycles. Reservoir heterogeneity leads to non-uniform gas movement in heterogeneous systems and as a result, the efficiency of hydrogen storage is greatly compromised. In addition, the optimal selection of production flow rates presents a challenge in balancing hydrogen recovery and water management. The study highlights the need for customized approaches, emphasizing the importance of aligning flow rates with specific reservoir characteristics for efficient large-scale hydrogen storage.
        A set of simulations was conducted to investigate the influence of relative permeability hysteresis, injection/production rates, and reservoir heterogeneity (permeability) on large-scale hydrogen storage performance. The spatial continuity was established as omni-directional in the heterogeneous model and the simulation was conducted in a two-dimensional cartesian system (Figure 1). Both injectors and producers were controlled by the flow rate at reservoir conditions (0.015 PV/day), assuming the fracturing pressure would not be exceeded. The model's end simulated a large aquifer as a spill point, ensuring sufficient volume for input or output to maintain well flow rates in the gas injection and production cycles. The simulation was designed to terminate as soon as the gas reaches the spill point.
        Both the homogeneous and omni-directional models exhibited similarities and differences in their hydrogen injection and production behaviour. The homogeneous system, with a consistent injection rate of 6.01305 m3/day, showed higher primary production recovery (72.56%) compared to the omni-directional system (56.60%), where high permeability in specific blocks led to rapid hydrogen advancement and more hydrogen trapping during production cycles. The second injection cycle demonstrated comparable trends in recovery factors for both systems, with the heterogeneous system achieving a slightly higher final recovery of secondary production (97.1%) compared to the omni-directional system (94%). In both systems, by decreasing the gas injection and production rates the final recovery of primary and secondary production was lowered.
        Lowering the production flow rate consistently reduces the gas-water ratio in both models due to the decreased mobility difference between hydrogen and water. In the heterogeneous model, since water was displaced unevenly throughout the system (viscous fingering), the waterfront progressed faster towards the well during the production cycle. As shown in Figure 2, the waterfront eventually reaches the production well, causing a significant reduction in the gas-water ratio, particularly in the second production cycle since a smaller amount of hydrogen was injected during the second injection cycle. This dynamic transition holds substantial implications for storage operation efficiency and becomes a pivotal factor in operational decision-making.

        Speaker: Mr Abdolali Mosallanezhad (Researcher)
      • 204
        Numerical simulation of depleted and cushion gases impacts on hydrogen storage in a depleted gas reservoir

        Depleted gas reservoirs have large storage capacity, pre-proved containment security, in-place depleted gas, and well-established surface infrastructures, thus are viewed as the most feasible hydrogen storage space. However, the impacts of depletion time, volume ratio ($VR_{H_{2}:CH_{4}/CO_{2}}$) of hydrogen ($H_2$), depleted gas (mainly $CH_4$) and cushion gas (e.g., $CO_2$), and injection/withdrawal mode on hydrogen storage performance have not been systematically studied. Therefore, we examined these impacts using a numerical simulation method. The results demonstrate that: 1) As the $VR_{H_{2}:CH_{4}}$ decreases from 100% : 0 to 50% : 50%, both $H_2$ withdrawal factor ($W_{F-H_2}$) and purity ($W_{P-H_2}$) firstly increase and then decrease; during the $1^{st}$ withdrawal cycle, the highest $W_{F-H_2}$ is 42% and the smallest $W_{P-H_2}$ is 51%, both of which occur at the $VR_{H_{2}:CH_{4}}$ = 60% : 40%. 2) In case of $CO_2$ as cushion gas, $W_{F-H_2}$ and $W_{P-H_2}$ are decrease as the $VR_{H_{2}:CO_{2}}$ increases from 50% : 50% to 25% : 75%; during the $1^{st}$ withdrawal cycle, the highest $W_{F-H_2}$ is 38% and the smallest $W_{P-H_2}$ is 50%, both of which occur at the $VR_{H_{2}:CO_{2}}$ = 25% : 75%. 3) A smaller $H_2$ withdrawal rate ($W_{R-H_2}$) results in a lower $W_{F-H_2}$, but a higher $W_{P-H_2}$, e.g., at the $VR_{H_{2}:CH_{4}}$ = 60% : 40%, $W_{F-H_2}$ and $W_{P-H_2}$ are 42% and 51% at $W_{R-H_2}$ = 100×$10^4$ S$m^3$/day, respectively, while they become 31% and 64%, respectively, at $W_{R-H_2}$ = 50×$10^4$ S$m^3$/day. These simulation results indicate that cushion gas injection is beneficial to reducing hydrogen loss, and depleted gas can be used as cushion gas. Depletion time and the ratio of hydrogen, depleted and cushion gas have significant influence on hydrogen storage performance. These insights provide important guidance for industrial hydrogen storage in depleted gas reservoirs.

        Speaker: Yawen Yang (University of Science and Technology Beijing)
      • 205
        Pore Storage for Green Hydrogen: A Sensitivity Analysis of Geological Parameters at Ketzin Anticline (Germany)

        The increasing demand for green hydrogen in Europe, particularly in Germany, requires extensive storage capacities to compensate the imbalance between production and consumption. Although salt caverns are known as safe and reliable medium for large scale hydrogen storage, however, these are geographically limited. Therefore, porous reservoirs - such as saline aquifers and depleted hydrocarbon fields - are suggested as alternative solutions.

        This study investigates the preliminary feasibility of underground hydrogen storage in Ketzin, Brandenburg, Germany. The Ketzin site was already used for the storage of town gas, natural gas and CO2 as well in the past decades. The potential hydrogen storage unit is the Stuttgart formation, which exhibits fluvial facies characterized by channel and floodplain elements. In this study, we consider a homogenous model for a sensitivity analysis, where porosity, permeability, capillary pressure, relative permeability, hydrodynamic dispersivity as well as salinity of the brine are investigated within their known ranges. The effect of geological parameters is analysed with respect to storage performance such as gas injectivity and productivity as well as sustainability derived from long-term injectivity and productivity.

        The results from the sensitivity study reveal that changes in capillary pressure and permeability exhibit the most significant influence on productivity, with variations of up to 150% observed in both positive and negative directions. The sustainability, on the other hand, is significantly negatively influenced by low porosity, low permeability as well as high dispersivity by up to 80 %. However, the results of the high salinity of the water solution have the most positive influence on this index.

        The analysis not only holds significant importance in advancing fundamental knowledge on the storage of hydrogen in porous reservoirs and its influencing parameters. It also provides a solid starting point for the systematic evaluation of the feasibility of a prospective regional hydrogen demonstrator. Future work will be carried out with a more complex model including the lithological heterogeneity and focus on the optimization of the operating parameters like the duration of cushion gas injection and storage cycles as well as the injectivity and productivity.

        Speaker: Lea Döpp (GFZ German Research Centre of Geosciences)
      • 206
        Performance Study of Underground Hydrogen Storage in a Saline Aquifer for a Prospective Hydrogen Pore Storage Site in Northeast Germany

        To address seasonal fluctuations in supply and demand for renewable energy, hydrogen (H$_2$) can be produced using excess electricity and temporarily stored in geological formations [1]. Due to their large volumes, widespread occurrence and distribution in sedimentary basins, saline aquifers have significant potential for underground hydrogen storage (UHS). However, large-scale UHS of pure H$_2$ in the porous subsurface has not been demonstrated yet. The Helmholtz research project GEOZeit focuses on preparatory research for the construction of a hydrogen pore storage demonstrator in a saline aquifer.

        The precursor research contains numerical reservoir simulations with the reservoir software CMG GEM. It targets to assess the capability of UHS operations at the Triassic Stuttgart anticlinal formation near Ketzin, Germany. This formation is lithologically heterogeneous, consisting of mudstone and siltstone, and the reservoir sandstone varies in reservoir properties and thicknesses [2]. In the recent past, a large-scale CO$_2$ storage research project was successfully realised at the flank of the anticline [3]. Now, the top of this structure is explored to serve as a structural trap for storing H$_2$. However, seismic surveys revealed the presence of a fault zone at the top [2,4], indicating possible migration pathways for the gas. To study fluid flow across the fault system, different fault leakage scenarios are carried out by adjusting fault transmissibility to represent sealing or leaky faults.
        To access areas situated at an increased distance from the fault zone, we are exploring the option of horizontal directionally drilled (HDD) wells to bypass the fault zone. Although vertical drilling is acknowledged as a cost-effective method, HDD excels in exploring a wider expanse of the reservoir. Given that the performance of a storage operation is strongly dependent on the well location, orientation and integrity, the comparative gas injection and withdrawal performance of a vertical versus a horizontal well layout will be presented.

        For all evaluated scenarios, crucial metrics are applied to assess the quality and effectiveness of the storage operation, such as gas purity, sweep efficiency, and cyclic efficiency. The findings from the numerical studies on UHS, encompassing both general considerations and site-specific analyses at the Ketzin site, will play a crucial role in preparing and developing a prospective hydrogen demonstrator and evaluating its feasibility.

        Speaker: Dr Anna-Maria Eckel (GFZ German Research Centre for Geosciences)
    • MS03: 2.1
      • 207
        Effects of cross-scale fracture surface roughness in crystalline host rocks on hydrodynamics

        Granite is considered a suitable host rock for a deep geological repository for radioactive waste. Since fractures are the main flow pathways for solute transport in this material, accurate and efficient calculation of solute transport and retention phenomena is essential for predictions related to the safety case of the repository. A key issue is the effect of cross-scale surface topography and roughness on hydrodynamics such as fluid channeling and residence time. In this study, we use a fracture geometry model based on µ-CT data and apply a finite element method to reveal the influence of fracture geometry on solute transport behavior. Due to the heterogeneity of fracture shape and aperture width distribution, it is difficult to describe the fracture geometry and morphology by a single variable. In addition, the surface roughness of fracture walls exhibits cross-scale variability due to heterogeneous material composition, which hinders the application of simplifying self-affine geometry descriptions. Instead, investigating of the role of cross-scale surface roughness in solute transport modeling is a promising approach. We investigated the sensitivity of the roughness effect by systematic modification across scales using µ-CT data of granite fractures. By comparing 2.5D vs. 3D transport model results, the role of long wavelength surface constituents and fracture bending can be investigated. The solute transport modeling was performed using the finite element code COMSOL Multiphysics. We discuss the quantitative effect of long wavelength surface building blocks on the tailing of the breakthrough curves and a weakening of the Fickian behavior. The tracer concentration fields in the 2.5D models show a high sensitivity to spatial heterogeneity. The solute transport in larger half-pores is overestimated compared to 3D models. The differences between 2.5D and 3D models due to small-scale surface roughness are considerably smaller. Nevertheless, the effect of surface roughness wavelengths on the BTC tailing behavior is not simply monotonic, which is an important effect to consider when implementing roughness parameters in transport modeling. Finally, we discuss the potential application of using power spectral density (PSD) curves as a means of assessing changes in roughness on fracture surfaces. PSD curves provide a cross-scale quantification of surface topographies. We propose the implementation of PSD curves in transport models to increase their predictive capability for contaminant migration in fractures.

        Speaker: Mr Wenyu Zhou (HZDR, FS Leipzig)
      • 208
        Numerical simulation of transport mechanisms for cyclic high-speed injection and production in fractured-vuggy underground gas storage

        Underground gas storage (UGS) exhibits various transport mechanisms due to their multi-cycle injection and production, often overlooked in numerical simulations. In fractured-vuggy UGS, certain mechanisms may have a stronger effect. L underground gas storage is crucial to China's first fractured-vuggy UGS group. A simulation study was conducted to investigate the effect of three transport mechanisms—stress sensitivity, relative permeability hysteresis, and high-speed non-Darcy effect of the fractured-vuggy UGS L during high-speed injection and production. Based on the geological model of the L UGS, the history matchings were separately conducted with and without considering transport mechanism to ensure model accuracy and elucidate the significance of transport mechanisms (as shown in Fig. 1). The multi-component fluid characterizations were implemented separately using the Peng-Robinson equation of state to perform the compositional simulation.

        Our study found that stress sensitivity resulted in a 3.31% reduction in storage capacity and a 6.07% decrease in working gas volume. Relative permeability hysteresis led to a 9.05% decline in storage capacity and a 4.09% decrease in working gas volume. The high-speed non-Darcy effect only caused a 0.16% reduction in storage capacity but led to a 4.19% decrease in working gas volume. With increased injection and production cycles, the storage capacity steadily rises. After 25 cycles, there was a 3.97% increase in storage capacity. Stress sensitivity increased the capacity increment to 4.66%, while relative permeability hysteresis and high-speed non-Darcy effect raised the increment to 6.05% and 4.13%, respectively. The greater the impact of the transport mechanisms on storage capacity and working gas volume, the more significant the increase in storage capacity. However, this increase in capacity is attributed to the rise in cushion gas volume and does not reflect an increase in working gas volume. The coupling of stress sensitivity and relative permeability hysteresis resulted in a reduction of 6.51% in storage capacity and a decrease of 11.65% in working gas volume. The coupling of these two mechanisms reduced the loss in storage capacity but amplified the decline in working gas volume. We analyzed six effects resulting from the coupling of two mechanisms, as illustrated in Fig. 2. The coupling of the three mechanisms resulted in a reduction of 6.53% in storage capacity and a decrease of 13.44% in working gas volume (as shown in Fig. 3). Coupling with the high-speed non-Darcy effect had no extra coupling effect on storage capacity but led to a further decline in working gas volume. The mutual influence relationships among the three mechanisms are depicted in Fig. 4.

        This study presents the utilization of compositional simulation to investigate the coupled effect of stress sensitivity, relative permeability hysteresis, and high-speed non-Darcy effect, as well as the coupled effects of stress sensitivity with relative permeability hysteresis on the operation of fractured-vuggy UGS. It comprehensively quantifies the varying degrees of influence exerted by different transport mechanisms on the operation of UGS. The study offers guidance for optimizing operational strategies for the UGS L and other similar UGS converted from fractured-vuggy reservoirs.

        Speaker: Dr Ye Tian (Southwest Petroleum University)
      • 209
        A New Numerical Well-Test Model Using an Analytically Modified Embedded Discrete Fracture Model

        The technology of multi-stage, multi-well pad fracturing is an effective way to increase the stimulated volume and recoverable reserves in shale reservoirs. During the fracturing treatments, there are common phenomena of well interferences from the multi-well pad. However, there still lacks an effective tool to analyze the parent-child interactions and to evaluate the fracture parameters quantificationally. To narrow this gap, a numerical pad-well model is developed for pressure transient analysis in fractured horizontal wells with secondary fractures and well interferences, based on a discrete fracture model (DFM) and unstructured PEBI grid system.

        Using methods of automatic differentiation and Newton iteration, the proposed model is more efficient for computations and interpretations of well testing curves. Its accuracy and practicality have been demonstrated by model verifications and field applications. The results show that the flow regime of interference effects caused by parent-child interactions are more obvious, with a larger child-well production, a smaller well spacing, and a larger hydraulic-fracture angle. The well interferences are also stronger when the child well has more secondary fractures, longer secondary fractures, and higher fracture conductivity, as the pressure drop caused by child well will propagate more quickly. Once the complex fracture networks have developed within the multi-well pad, the interactions between parent and child well will be weaker with the increase in area and conductivity of fracture networks. By comparison, the pressure transient behaviors of Parent well are remarkably affected by Child-well production rate, well spacing as well as connectivity degree. However, the angle, length, number, and conductivity of secondary fracture have weaker impacts on the pressure transient behaviors of Parent well. The field application shows that the single-well testing model without considering well interference cannot match with field data at the late stage. In this case, the estimation errors will occur. With considering the well interferences, the well testing data are interpreted and the fracture parameters are evaluated successfully. This work provides a meaningful way to understand the pressure transient behaviors and to evaluate the fracture parameters of multi-stage, multi-well pads.

        Speaker: Prof. Zhiming Chen (China University of Petroleum at Beijing)
      • 210
        Generalized framework for flow in fractured subsurface formations

        While isolated fractures are difficult to fully characterize in subsurface formations, they serve as highly conductive long-range flow conduits and thus may have a strong influence on flow and transport. Recently, we have proposed a new model for flow in fractured formations that provides predictions of the expected flow field. Unlike existing methods, this model accounts for the non-local effect of these flow conduits using kernel functions that appear in an integro-differential flow balance equation. In the present work, kernel functions predicting mean flow rates and pressure profiles for a variety of fracture shapes are presented. Furthermore, discrete fracture length distributions are incorporated leading to a formalism that can account for mixtures of different fracture families. A series of numerical experiments are presented with the results being successfully compared to expensive fine-scale reference simulations.

        Speaker: Daniel Stalder
    • MS08: 2.1
      • 211
        Salinity-induced melting of underlaying permafrost

        The rise of sea levels and the expansion of plateau salt lakes are among major consequences of the ongoing climate change[1]. When saline water overlays above permafrost (ice in porous soil), ice may melt because salinity reduces the melting/freezing point. Permafrost melting may alter the mechanical properties of the soil and affect the safety of coastal structures[1], and even may induce the release of underground methane gas into the atmosphere[2]. Therefore, studying the kinetics of salinity-induced melting of underlaying permafrost is of great environmental significance.

        We conducted visualized experiments to study the kinetics of permafrost melting induced by overlying saline water. Water in bead-pack is first frozen to mimic permafrost and then is immersed under excessive saline water at -5°C. Glass bead diameter varies from 0.1mm to 0.5mm, and salt concentration varies from 10wt% to 25wt%. Melting front (ice-water interface) in porous media can be visually identified (Figure 1(a)) and recorded by camera. As the dilute saline water at the melting front is of lower density than the overlying saline water, Rayleigh-Darcy convection is induced in the porous medium[3,4], so we use Rayleigh number (ratio of gravitational-induced flux over diffusion), Ra, to characterize the mass transfer in liquid-saturated porous layer[5].

        Surprisingly, we found two distinct melting patterns: 1) when Ra is high, the melting front is flat and moves down stably; 2) when Ra is low, “fingers” emerge and develop at melting front. This seems to be different from the previous research results that greater Ra implies higher instability[6–8].

        We theoretically show that the melting pattern is a result of interplay between local circumflux shaped by the melting front perturbation and the global Rayleigh-Darcy convection (Figure 1(b)). When a perturbation emerges at the front, density contrast induces local convection from the trough to the peak, that further enlarge the perturbation (Figure 1(c)). This local convection is proportional to Ra. Meanwhile, global Rayleigh-Darcy convection enhances lateral mixing which compress the development of the perturbation. This lateral mixing is proportional to Ra^(1.5~2). As a result, when Ra is low, melting is dominated by the local circumflux and fingers grow; when Ra is high, strong lateral mixing homogenizes concentrations along the solid-liquid interface and results in flat melting front. Numerical simulations further support the above theory and match the experiments well. When the melting front is flat, the melting rate can be predicted by classical Rayleigh-Darcy convection theory. However, when a fingering melting front forms, the melting rate is one order of magnitude slower than classical theory prediction. Moreover, fingering melting front implies penetration of permafrost layer before melting all ice, that may induce unexpected groundwater pollution and subsurface methane release.

        Speaker: Yumin Wang
      • 212
        How does access to continuous brine sources in saline aquifers enhance salt precipitation dynamics during geological CO2 storage?

        Meeting the long-term expectations of Carbon Capture and Storage (CCS) technology hinges on injecting massive volumes of CO2 annually into deep saline aquifers. These aquifers, due to their storage capacity and proximity to emission sources, are prime candidates for CO2 sequestration. The near-wellbore environment experiences significant thermo-hydro-mechanical-chemical (THMC) perturbations, necessitating a comprehensive understanding through experimental tests and simulations to maximize the safety and cost-efficiency of CO2 storage. The injection of large volumes (on a million-ton scale) of supercritical CO2 into the geological formations causes evaporation of formation water near wellbores and precipitation of salt crystals inside the porous medium. CO2-induced salt precipitation can substantially threaten sequestration in saline aquifers. While existing works primarily focus on predicting salt location and amounts, our study delves into the physics, growth dynamics, and behavior of the fluid-solid interface near the evaporation/precipitation front. We present a series of experiments, including microfluidic, hele-shaw, and sandbox, along with pore-scale reactive transport modeling using the Lattice Boltzmann Method (LBM), providing fresh insights into brine evaporation and salt growth dynamics. Our research challenges the current understanding, revealing a common shortcoming in many experimental studies—failure to facilitate access and replicate in-situ continuous brine sources. This shortcoming significantly alters the dynamics of salt nucleation and growth in porous reservoir rocks, where the availability and continuity of solute through water film movement control geometric alterations. To address this issue, we also designed surface mineral precipitation tests and large-scale sandbox experiments to investigate salt precipitation and growth under two scales and regimes in porous geometries. The laboratory results indicated massive salt accumulation close to the injection port and underlined the effect of solute availability and continuity on the intensifying severity of salt accumulation. The research outcome highlights the interplay of complex processes (some of which are not yet fully characterized) crucial in investigating salt precipitation induced by million-tons-scale CO2 injection. The observed characteristics call for further in-depth investigation because, in the context of subsurface CO2 storage, we need to redefine how we see injectivity impairment due to salt precipitation.

        Speaker: Dr Mohammad Nooraiepour (Environmental Geosciences Section, Department of Geosciences, University of Oslo, Oslo, Norway)
      • 213
        Experimental and Numerical Study of Carbon Dioxide Geological Storage in Coal – A Comparative Analysis with the application of Positron Emission Tomography Imaging.

        With the global energy mix predominantly fossil fuel based1, carbon dioxide (CO2) capture, utilisation and geological storage (CCUS) is a key tool in reducing anthropogenic CO2 emissions. One of the main challenges facing CCUS in coal seams is the loss of injectivity due to CO2 coal swelling2–7. This work aims to improve the understanding of CO2 transport mechanisms in coal by applying in-situ Positron Emission Tomography (PET) imaging to obtain direct images of CO2 flow in coal, allowing for better CO2 geosequestration techniques.

        This work presents a comparative history matching analysis between a one-dimensional Advection Diffusion Equation (ADE) model and experimental data obtained from in-situ PET imaging during core flooding. Traditional core flooding methods usually rely on assumptions that flow within samples is piston like, possibly leading to inaccuracies during modelling8. The use of PET imaging provides an actual representation of gas flow behaviour and serves as a ground truth point of reference during history matching. This comparative analysis focuses on determining how the diffusion coefficient of CO2 in coal changes vis-à-vis coal properties such as initial adsorbate molecules, coal swelling and their effects on gas (particularly CO2) diffusion within coal samples.

        [11C]CO2 was utilised as the PET radiotracer during core flooding experiments to directly image carbon dioxide (CO2) diffusion dynamics and mechanisms within coal samples. A 1-D ADE model in MATLAB was then history matched to experimental data for the purpose of obtaining the diffusion coefficient that best represented what was observed. Additionally, X-Ray μCT imaging technology was utilised to obtain high resolution (30μm) images of the core samples. Machine learning algorithms were then applied to these CT images as a method of digital image segmentation to obtain a good estimate of sample porosity, further improving the accuracy of the 1-D ADE model. In-situ PET scans allow for a dynamic observation of gas flow during the core flooding experiment as well as a source on which gas diffusion effects can be confirmed through the use of history matching with a 1-D ADE model in MATLAB and subsequently back calculating the diffusion coefficient of best fit from the ADE model that accounts for key core flooding parameters such as coal porosity, gas flow rate and type of injected gas.

        The results show that stable diffusion coefficients arise when samples are dry and an inert gas (He) is used as the carrier gas. In the cases of competitive adsorption between methane (CH4) and CO2 in samples that were CH4 saturated show a decreasing diffusion coefficient. Diffusion coefficients in CO2 saturated samples were in the order of 100 times lower than in samples that were not exposed to CO2 prior to injection. This indicates that coal swelling has a significant impact on the ability of gas to effectively diffuse in the coal matrix. These findings further develop the body of knowledge surrounding CO2 geosequestration in unmineable coal seams and contributes to the optimisation of CCUS processes during the transition to low-carbon and renewable energy sources.

        Speaker: Aaron Uthaia Kumaran (University of New South Wales)
      • 214
        Cotransport of clay and microplastics in saturated porous media

        The widespread use of plastics for various applications lead to their inevitable release into the environment. The disintegrated microplastics particles may ultimately find their way into the subsurface, thereby contaminating soil and groundwater. Hence, it is essential to understand the transport behaviour of microplastics in soil to protect drinking water wells from contamination. The presence of natural colloids such as clays in the subsurface are known to alter the transport behaviour of several contaminants. This study aims to understand the cotransport of clays and microplastics in saturated soil through column experiments and mathematical modeling. Experimental results showed enhanced transport of microplastics and retarded transport of clays during their cotransport as compared to their individual transport. This contrasting transport behaviour of clays and microplastics during their cotransport may be due to the competition between them in finding deposition sites on grain surfaces and also due to the formation of clay-microplastics heteroaggregates which may have different surface properties than individual clay and microplastics particles. The experimental results were successfully simulated using mathematical model which accounted for clay and microplastics retention in soil, heteroaggregation kinetics, and heteroaggregate retention in soil.

        Speaker: Mahima Horta (Indian Institute of Technology Hyderabad, India)
    • MS09: 2.1
      • 215
        Micro-Continuum Modeling of Mineral Nucleation and Precipitation at Pore-Scale

        Mineral nucleation and precipitation commonly occur in nature and plays an important role in many energy-related applications with reactive flow, especially, at pore-scale. For instance, minerals nucleate and precipitate as scale in the pore structure in unconventional reservoirs and significantly reduce the permeability of the porous media. This phenomenon could lead to a rapid decrease in production and cause significant financial loss. The need to predict the dynamic properties of such systems has resulted in questions about the fundamental mechanisms of reactive flow as well as mineral nucleation and precipitation in pores. Additionally, there is still a discrepancy between laboratory molecular scale findings and large-scale observations. To address this discrepancy, modeling methods at the pore scale started gaining interest recently due to the capability of capturing reactive and nonreactive species transport, effects of pore topology, and interface chemical reaction within the same approach, which typically is difficult to observe directly in experiments.
        For some solutions, especially high saturation index solution, the nucleation process could potentially play an important role in the precipitation due to either heterogeneous or homogenous nucleation, which was largely overlooked in most previous numerical models for mineral precipitation. In this study, we coupled the micro-continuum simulation approach based on Darcy-Brinkman-Stokes (DBS) equation with the classic nucleation theory (CNT) to study the stochastic nucleation process in reactive flow. A range of different parameters were studied to understand their impact on the nucleation process and precipitation. It was discovered that such a nucleation process was affected by the Damköhler number and Peclet number as well as other effects. As the precipitation reaction on the crystal surface enhances, the total amount of nucleus formed on the substrate decreases due to the depletion of species in the vicinity of the substrate. In general, the competition between flow/transport of species and precipitation consumption governs the behavior of phase change procss and produces different scenarios. The results of this study are expected to shed light on the mechanism of liquid-solid interaction within porous medium.

        Speaker: Fengchang Yang (Institute of Mechanics, Chinese Academy of Sciences)
      • 216
        Investigate the effect of pore heterogeneity on elastic wave velocity evolution under mineral dissolution process

        Mineral dissolution is a common phenomenon in many subsurface geo-systems, such as carbon sequestration, wastewater disposal and oil and gas recovery. Dissolution can change the topology of porous rock, which affects the rock’s geophysical parameters, such as permeability and elastic wave velocity. We numerically investigate the relationships between the evolutions of P-wave and S-wave velocities and permeability induced by mineral dissolution under different pore heterogeneities. We use a linear Boolean model to represent sedimentary rocks with various pore heterogeneities. We reproduce three typical dissolution patterns: compact, uniform and wormhole, by adjusting the Péclet and Damköhler numbers. For these numerical simulations, we use the lattice Boltzmann method to compute the velocity and concentration fields, and the finite element method to compute the strain fields. Our results indicate that the evolution trends of both P-wave and S-wave velocities are similar in all simulations. When the initial pore heterogeneity is fixed, the uniform dissolution pattern cases show a faster decrease of elastic wave velocity as the dissolution progresses; when the dissolution pattern is fixed, the more heterogeneous rock shows a faster decrease of elastic wave velocity. The findings have important implications for subsurface engineering applications involving pore network and fluid path evolutions caused by mineral dissolution.

        Speaker: Yutian Zhang
      • 217
        A Benchmark Study of Pore-scale Multiphase Flow in Pore-doublet: The Impacts of Hydrodynamics on Mineral Dissolution Reaction Rate

        Abstract: Multiphase flow in porous media is ubiquitous in soils, oil and gas reservoirs, geologic carbon storage and hydrogen storage systems, and batteries. In recent years, direct observation using microfluidic experiments and pore-scale numerical modeling have become increasingly important tools for studying pore-scale fluid dynamics. To further examine the precision of the experiment and evaluate the performance of numerical models in order to expand beyond experimental conditions, it is necessary to develop proper benchmark experiments.
          In this work, a benchmark study is developed for gas-water two-phase flow through a pore-doublet geometry. Microfluidic experiments for a range of capillary numbers and fluid properties were performed and characterized using optical microscopy techniques. Subsequently, the experiments were numerically simulated using the interFoam and interFlow solvers in OpenFOAM, and the phase-field and level-set methods in COMSOL. The comparison enables us to systematically examine the impacts of modeling decisions (e.g., mesh resolution, model dimensionality) under a range of flow rates and wettability conditions. Finally, mineral dissolution was numerically simulated using CrunchFOAM, a solver based on OpenFOAM and coupled with the CrunchTope geochemical framework, to evaluate the subsequent impacts on geochemical reactions in two-phase systems.

        Speaker: Xin Wang (Peking University)
      • 218
        Pore-scale multiphase reactive transport and CO2 mineralization capacity in vesicular basalts

        The geological storage of CO2 has emerged as a critical pathway for decarbonization, where in situ carbon mineralization in mafic/ultramafic rocks such as basalts is considered the most stable form of CO2 storage. in-situ CO2 mineralization pilot projects in basaltic formations include the Wallula project in Columbia River basalt and the Carbfix project in Icelandic basalt. Multiphase flow governs the invasion and distribution of native brine, carbonated water, and injected supercritical CO2 and will determine the accessibility and carbonation capacity of reactive mineral pore surfaces during geochemical processes. As such, what is the mix of injectate or injection scheme that optimizes tons of anthropogenic CO2 injected (storage) and mineralization capacity (security) for different formations?
        We leverage pore-scale, multiphase computational fluid dynamics (CFD) models, enhanced by experimentally- and theoretically-informed reactive transport relationships and mineral-fluid wettability values, to assess the complex interplay between mineral hydrophilicity, capillary trapping, thin films, dissolution, precipitant nucleation, and mineralization. We simulate various injection schemes, including supercritical (dry) CO2 invading in-situ brine and water-alternating-gas (WAG) injection, within several representative vesicular basalt samples (including one fresh basalt sample, one from the Carbfix site, and two from different flow-top zones in the Wallula site). The pore-scale models are informed by petrographic data of pore morphology (e.g., thin section, SEM, micro-CT), physical-chemical mineralization behavior (coupled with PHREEQC), and routine core analysis data. The models are tuned with different boundary conditions and initial conditions to represent the basalt units in different locations in the reservoir under the selected injection schemes. For each sample, we quantify crucial dynamic relationships for geologic storage and mineralization, including porosity-permeability, accessible reactive mineral surface area, brine-CO2 capillary pressure-saturation (Pc-Sw), and relative permeability (Kr-Sw) relationships. These relationships are explored as a function of native basalt groundwater composition, mineral-specific surface area, and the sequence of pore-scale alteration processes.
        The aforementioned dynamic pore-scale relationships are integrated with fluid characterization and core-scale measurements, including hydraulic tests, helium pyconometry, and NMR measurements. Results indicate a strong correlation between the location of precipitated nodules and the menisci of CO2 bubbles under steady state, depicting a vital role of multiphase flow in understanding geochemical processes.
        Ongoing efforts involve extrapolating pore-scale functional relationships to gridblock-scale reactive transport reservoir models (e.g., STOMP and MRST) to refine predictions of invasion depth, carbon storage and mineralization capacity, with the consideration of evolving accessible reactive surface area on a larger scale.

        Speaker: Tianxiao Shen (Columbia University)
    • MS13: 2.1
      • 219
        Bound water transport by diffusion in wood-revealed by Nuclear Magnetic Resonance

        Wood is extensively applied in various fields such as construction, tooling, sculpture, boat building. The water content within wood plays a crucial role in influencing its performance across different contexts. For example, a large portion of water must be removed from wet or green (fresh cut) wood to mitigate further dimensional variations under varying humidity conditions. In this context, the transport of bound water (absorbed between cellulose microfibrils, up to 30% of the dry mass, and at the origin of swelling) plays a fundamental role. However, measuring these transport properties is challenging as this bound water is contained in nanopore inclusions. Moreover, it was shown that during standard imbibition there is a strong coupling between bound water and free water [1-2].
        Here, for the first time, we develop experimental conditions allowing to prevent most free water (in vessels or fibers) imbibition in hardwood (oak and poplar). We then follow the progression of bound water by NMR relaxometry. This allows to determine in a straightforward way the diffusion coefficient of bound water. The results reveal that the transport diffusion coefficient of bound water in hardwood is rather large, typically in the order of 10-9 m2/s. More precisely, the diffusion in poplar occurs at a faster rate compared to oak samples. Additionally, we show that the fastest rates of diffusion are observed in the longitudinal direction, followed by the radial and the tangential directions. This research underscores the mechanisms and complexity of bound and free water transfer in bio-based materials and provides an insight into the processing and protection of wood.

        Speaker: Ms Luoyi Yan (Laboratoire Navier, Ecole des Ponts ParisTech,CNRS)
      • 220
        Coarse-grained modeling of fluid transport in swelling porous media

        Fluids can induce solid adsorption and swelling in porous materials when they infiltrate pores. For example, alkaline liquids created after water mixing with alkali metals can react with minerals like quartz or feldspar and form a new substance called alkali calcium silicate hydrate. Fluid flow facilitates the damage of concrete as this new substance can swell upon adsorbing water and crack the concrete. In carbon sequestration, CO2 injectivity can significantly decrease with time as gas adsorption induces swelling of the rock matrix and reduces pore spaces for the flow pathway. However, the coupling between fluid flow and solid deformation remains challenging to be captured by numerical models.
        Coarse-grained molecular dynamics (CGMD) bridges nano- and micro-scales by mapping a group of atoms/molecules into a single coarse-grained (CG) particle. Compared with all-atomic molecular dynamics (MD), it overcomes the difficulty of simulating multiphase flow with multiphysics in complex pore networks. This study introduces a novel CGMD model to achieve the coupling between fluid transport and solid deformation at the microscale. This model accurately simulates the interactions between fluids and solids, and between fluids and solids themselves. The solid comprises bead-spring chain networks considering bonding and non-bonding interactions and reproduces a broad range of Young's moduli and swelling ratios. The fluid is modeled by dissipative particle dynamics (DPD) and calibrated against density and viscosity at different pressures.
        The proposed CGMD model has been adopted to study fluid transport through deformable and non-deformable nanochannels of varying sizes (35.4 nm~123.9 nm) and a simplified nanoporous medium composed of spherical solids. The results are analyzed using the Hagen-Poiseuille equation and the Kozeny-Carmen equation for validation. The effect of swelling on reducing fluid permeability is justified, and a relationship is established between fluid permeability and solid swelling. This study provides a straightforward new approach to modeling fluid transport in swelling porous media at the microscale within the framework of CGMD, with potential applications in concrete design and energy storage technologies.

        Speaker: Jian Wu (The University of Sydney)
      • 221
        Confinement-guided self-assembly of ionic superdiscs

        Discotic ionic liquid crystals (DILCs) consist of self-assembled superdiscs of cations and anions that spontaneously stack in linear columns with high one-dimensional ionic and electronic charge mobility, making them prominent model systems for functional soft matter. Compared to classical non-ionic discotic liquid crystals (DLCs), many novel liquid crystalline structures with a unique combination of electronic and ionic conductivity have been reported, which are of interest for separation membranes, artificial ion/proton conducting membranes and optoelectronics. Unfortunately, a homogeneous alignment of the DILCs on the macroscale is often not achievable, which significantly limits the applicability of DILCs. Infiltration into nanoporous solid scaffolds can overcome this drawback. However, little is nknown about the structures of DILCs in nanoscale confinement. Here, we present temperature-dependent high-resolution optical birefringence measurement and 3D reciprocal space mapping based on synchrotron-based X-ray scattering to investigate the thermotropic phase behavior of dopamine-based ionic liquid crystals confined in cylindrical channels of 180 nm diameter in macroscopic anodic aluminum oxide (AAO) membranes. As a function of the membranes hydrophilicity and thus the molecular anchoring to the pore walls (edge-on or face-on) and the variation of the hydrophilic-hydrophobic balance between the aromatic cores and the alkyl side chain motifs of the superdiscs, we find a particularly rich phase behavior, which is not present in the bulk state. It is governed by a complex interplay of liquid crystalline elastic energies (bending and splay deformations), polar interactions and pure geometric confinement, and includes textural transitions between radial and axial alignment of the columns with respect to the long nanochannel axis. Furthermore, confinement-induced continuous order formation is observed in contrast to discontinuous first-order phase transitions, which can be quantitatively described by Landau-de Gennes free energy models for liquid crystalline order transitions in confinement. Our observations suggest that the infiltration of DILCs into nanoporous solids allows tailoring their nanoscale texture and thus their electrical and optical functionalities over an even wider range than in the bulk state, in a homogeneous manner on the centimeter scale as controlled by the monolithic nanoporous scaffolds.

        Speaker: Mr Zhuoqing Li
      • 222
        Fractal nanopore structure of anthracite and CO2 adsorption-induced alteration: A synchrotron radiation SAXS study

        Abstract:
        Coal is a porous medium material with a highly developed pore network inside. Nanopores dominate gas adsorption and transport behavior in geological reservoirs. The fractal nanopore structure of anthracite from the Qinshui Basin were characterized using synchrotron radiation small angle X-ray scattering (SAXS). Based on the fractal theory of SAXS, the fractal characteristics of nanopores were obtained by analyzing the scattering data. The results indicate that the nanopores at 10~70 nm exhibit surface fractal characteristics, with irregular self-similar surfaces. The fractal nanopore structure of different sizes can be obtained by dividing the logarithmic curve into different regions. The pores at 10~20 nm, 20~30 nm, and 30~40 nm exhibit surface fractal characteristics, with the greatest contribution to the surface fractal characteristics of the overall pores (10~70 nm). However, the pores at 40~50 nm, 50~60 nm, and 60~70 nm exhibit pore fractal characteristics, reflecting the self-similar pore structure of the nanopores. Compared to the initial state, the surface fractal dimension of pores at 10~70 nm with CO2 adsorption gradually decreases. There is a negative correlation between surface fractal characteristics and adsorption pressure. From the initial state to 3 MPa adsorption, the fractal dimensions of pores at 10~20 nm, 30~40 nm, and 60~70 nm decreased by 5.597%, 2.397%, and 8.214%, respectively. CO2 adsorption weakens the fractal characteristics most significantly. The fractal dimensions of pores at 20~30 nm and 50~60 nm exhibit fluctuations under different adsorption pressures. CO2 adsorption has a relatively small impact on the fractal nanopore structure. Specifically, the pores at 40~50 nm with CO2 adsorption have the maximum fractal dimension (up to 2.97) and remain constant. It is difficult to alter the self-similarity of pore structure between 40~50 nm for CO2 adsorption.
        Key words: Coal; Nanopore; CO2 adsorption; SAXS; Fractal

        Speaker: Prof. Yixin Zhao (China University of Mining and Technology (Beijing))
    • MS15: 2.1
      • 223
        Stable Diffusion in Digital Rock Analysis: Applications, Challenges, and Future Prospects

        Digital rock analysis has shown promise in visualizing geological microstructures and elucidating transport mechanisms in subsurface rocks, particularly in unconventional reservoirs such as tight sandstone and shale. Accurate image reconstruction techniques, which provide valuable insights into the pore network, grain distribution and connectivity, are essential to capture the intricate features and heterogeneity present in digital rock samples.

        Stable diffusion (SD), a new hotspot in the field of artificial intelligence-generated content (AIGC), holds promising potential for the production of high-quality digital rock images. The SD is a deep learning model based on diffusion techniques, and has revolutionized the field of computer vision by generating highly realistic images from textual prompts, since its first release in 2022. While it is already being used in fields such as illustration, game design and electronic-commerce, its application in the digital core field is still in its early stages.

        In this study, we examine the primary applications of SD in the field of digital rock analysis. Specifically, we explore its potential in enhancing image resolution, improving image quality through denoising and deblurring techniques, segmenting images into multiple regions, filling in missing sections, extending images in any direction using outpainting, and reconstructing 3D digital rocks based 2D images. Furthermore, this research highlights certain limitations of existing pre-trained models such as WebUI, Midjourney, and DALL-E. These limitations come from the fact that their databases do not encompass digital rock images obtained from scanning electron microscopes (SEM) or computed tomography (CT). Therefore, it is imperative to fine-tune the existing models or develop new ones specifically tailored to the realm of digital rock analysis, which deserves further attention and investigation.

        Speaker: Yutian Ma (China University of Petroleum (Beijing) Karamay Campus)
      • 224
        Optimizing Underground Hydrogen Storage through Surrogate Modeling: A CNN-LSTM-Attention Network Approach

        Underground hydrogen storage (UHS) presents a viable solution for storing excess energy in suitable geological sites, ensuring a stable and scalable energy supply [1]. While extensive experience exists in underground natural gas storage [2], the significant differences in the properties of hydrogen pose unique challenges [3]. To deepen insights into the hydrogen recovery in UHS projects, conducting reservoir simulation and optimization to pinpoint optimal operating parameters becomes essential. However, this process is typically time-consuming. The integration of a surrogate model proves invaluable in expediting the optimization process, addressing the significant time constraints associated with traditional methods.

        We develop a base UHS simulation model with a 3D heterogeneous depleted natural gas reservoir featuring an anticline structure. The model integrates various physics, encompassing compositional fluid flow, hydrogen methanation reaction, gravity segregation, hysteresis, and capillary effects. The cycling schedule starts with an initial phase of cushion gas injection and idle periods, followed by five distinct hydrogen injection-idle-production cycles spanning five consecutive years. Notably, injection rates and bottomhole pressure (BHP) of the production well vary across these cycles. The base model incorporates diverse cushion gas types and layers of perforation. Upon parameterizing these decision variables and employing the Latin-Hypercube method for sampling, we generate a comprehensive database comprising approximately 1000 simulation cases, executed in parallel. To predict cumulative productions of hydrogen and other components, we train a surrogate model utilizing a CNN-LSTM-Attention network, leveraging the NVIDIA RTX A6000. The CNN component transforms 3D heterogeneous permeability and porosity fields into 1D datasets. This well-tailored surrogate model seamlessly integrates into the optimization workflow based on the stochastic simplex approximate gradient (StoSAG) [4] method. The primary optimization objective is to maximize hydrogen recovery while concurrently minimizing losses attributed to micro-bio reactions within a predefined timeframe.

        Due to gravitational segregation in the base model, hydrogen, cushion gas, methane, and water exhibit a vertical distribution from top to bottom. Additionally, we note a progressive enhancement in hydrogen recovery efficiency with consistent injection rates during production. Our numerical experiments highlight nitrogen's superior effectiveness as a cushion gas for augmenting hydrogen recovery compared to carbon dioxide and identify a specific percentage of micro-bio-induced hydrogen loss. Regarding the performance of the surrogate model, the R2 scores for both training and testing datasets mostly exceed 0.95, affirming its robustness and feasibility. To demonstrate the acceleration achieved through the proxy model in optimization, we compare CPU times between the reservoir simulation and surrogate models. The former averages 210-300 seconds per case, while the latter ranges from 0.01 to 0.1 seconds. This translates to a remarkable speedup of approximately 1000 times compared to optimization conducted solely with reservoir simulation, all while maintaining equivalent accuracy.

        This research introduces a comprehensive framework designed for reservoir simulation and optimization in UHS, integrating a CNN-LSTM-Attention network and StoSAG. Important mechanisms, including compositional flow, cushion gas dynamics, and micro-bio reactions, are thoroughly incorporated in the UHS simulation. This framework serves as an important guideline, offering crucial insights into accelerating the optimization of the UHS process and related projects.

        Speaker: Mr Zhilei Han (KAUST)
      • 225
        Feasibility study of the inversion method for non-uniform hydrate saturation distribution based on ensemble Kalman filter algorithm

        Field tests and laboratory experiments indicate that the spatial distribution of hydrate saturation in hydrate reservoirs is non-uniform. This non-uniform distribution significantly impacts the reservoir’s temperature changes, and gas and water production rates during reservoir development. Currently, the primary methods for determining hydrate saturation distribution in porous media are nuclear magnetic resonance (NMR) and computed tomography (CT) scanning. However, these methods have limitations such as small detection ranges, high costs, and the necessity of interrupting experiments. During depressurization exploitation of hydrate reservoirs, abundant data on gas and water production, as well as temperature and pressure monitoring, are available. These highly reliable observational data vary with changes in hydrate saturation distribution, providing the possibility of using inversion methods to determine this distribution.
        This study first conducts secondary development of the Tough+Hydrate simulator. Energy and mass conservation equations are separately constructed for the matrix and high-conductivity channels after reservoir stimulation. The transfer of mass and heat in the matrix and high-conductivity channels was characterized using the discrete fracture method. A numerical simulation method for reservoir stimulation assisted depressurization development of hydrate reservoirs was established and implicitly solved. Then, by combining the ensemble Kalman filter algorithm with the simulator, the inversion method of the hydrate saturation distribution was built and then validated using three cases: core scale depressurization development, hydraulic fracturing assisted depressurization development, and radial well stimulation assisted depressurization development. The impact of the number of observation points on the inversion results also was investigated. Finally, based on
        the observation data of Masuda’s classic experiment, the inversion method was used to obtain the distribution of hydrate saturation in the core of Masuda’s experiment successfully.
        Research results indicate that the established inversion method continuously assimilates observational data in the ensemble. Hydrate saturation distributions obtained through inversion in the three cases tend to approach the preset distributions, demonstrating the reliability of the inversion method. The quantity of observational data has a certain influence on inversion results; more observational data lead to the assimilation of more information, resulting in hydrate saturation distributions closer to the actual values. Inversion results based on Masuda's experimental data reveal a strong non-uniformity in hydrate saturation distribution within the core, with a relatively high hydrate saturation zone in the central region and lower hydrate saturation at the inlet and outlet ends.

        Speaker: Yongge Liu (China University of Petroleum (East China))
      • 226
        A Comprehensive Approach to In-Situ Stress Estimation in Subsurface Energy Structures using Numerical Simulation and Machine Learning

        The objective of this research is to establish a consistent relationship between nonlinear numerical simulations and the obtained results for use in inverse analysis. We simulate the shape of breakouts, taking into account inelastic deformation of high-porosity limestone, using developed finite element methods under various in-situ conditions. Subsequently, the dataset is employed to train four machine learning algorithms, as well as white-box algorithms, in order to determine the relationship between in-situ stress and breakout shape.
        This study employs a two-phase approach through inverse analysis to determine in-situ stress. In the initial phase, we utilize nonlinear elastoplastic finite element modeling to generate a dataset. This dataset serves as the training data for a machine learning (ML) algorithm designed to establish a predictive correlation between in-situ stress and borehole breakout measurements. In the second phase, the trained ML algorithm is applied to estimate the equivalent in-situ stress based on provided borehole breakout measurements. To investigate in-situ stress from borehole breakouts and construct robust correlations, we employ a combination of four black-box algorithms and three white-box algorithms.
        A numerical simulation has been performed to determine the geometry of borehole breakouts under various in situ stress levels and taking into account plastic deformations. The breakout cross-section's non-circular shape can be modeled using an elastoplastic model that was created using the finite element approach. This shape fluctuates as the breakout develops until it stabilizes. The depth of the breakouts rises until a stable state, just like in earlier models based on the elastic assumptions. The width of the breakouts, however, does not change as the breakouts develop. The growth of the breakout is stopped by taking into account inelastic deformations, which also gives the chance to model the V-shaped type breakouts seen in both field and laboratory data. According to laboratory research, disregarding plastic deformations in very porous and weak rocks results in an incorrect understanding of the relationship between in situ stress and rock failure state.
        To determine the correlation between in situ stress and breakout shape, four machine learning techniques and three whitebox algorithms have been applied to the data set generated from numerical tests. To calculate the in situ stress from breakout shapes, trained algorithms were put through an inverse analysis. The XGBoost and GP algorithms mean square error (RMSE) of 0.419541, 0.9977and a determination coefficient (R2) of 0.99565 and 0.97564 outperform others in terms of accuracy and suitability.
        The novelty of the proposed approach lies in its consideration of inelastic deformation for estimating in-situ stresses, which is a crucial factor in the failure of high-porosity and unconsolidated rocks. Additionally, it involves establishing a relationship for estimating in-situ stresses through a combination of machine learning and numerical simulation.

        Speaker: Dr Erfan Mohammadian
    • MS19: 2.1
      • 227
        Electrical response during drying and imbibition of mesoporous materials.

        Upon the contact of the conductive mesoporous material with an aqueous electrolyte solution, ions adsorb on its surface, spontaneously forming an electrical double layer. In this case, due to the absence of an applied external potential difference, while the total charge of the system is zero, there is already accumulated a local charge at the interface. The number of adsorbed ions is determined by the chemical composition of the material, the bonds on the surface and, due to the spontaneity of this process, to a large extent by the surface area. Therefore, mesoporous materials with high specific surface area and porosity become the most favorable objects for research. However, when a second electrode with a different surface chemistry is introduced into the circuit, a potential difference occurs. This leads to spontaneous charge redistribution between the electrodes and rearrangement of ions at the interface. Current relaxation and potential difference evolution are the key characteristics of this process. Changes in the imbibition parameters as well as decreasing of wetted surface area during drying affect these electrical responses. Here we investigate the nature of these electrochemical processes and their correlation with fluid dynamics using gravimetric mass uptake measurements in combination with Zero Resistance Amperometry and other Open circuit methods.

        Speaker: Mrs Mariia Liseanskaia
      • 228
        Main controlling factors and pore structure of low resistivity shale

        Abstract:The phenomenon of abnormal low resistivity values in shale is widespread at the base of the Longmaxi Formation in the southern Sichuan region,and the gas content between different low resistivity shale gas wells have obvious differences.In order to explore the controlling factors and the differences in nanoscale pore structures of different resistivity shale, the shale of the Longmaxi Formation in the Changning area of the Sichuan Basin was taken as the research object.Firstly,according to the characteristics of electric logging curves and the production capacity,the shale wells of Longmaxi Formation in Changning area were divided into ultra-low resistivity wells(Rt<1Ω·m),low resistivity wells(1Ω·m<Rt<20Ω·m),and normal resistivity wells (Rt>20Ω·m).Secondly,the effects of organic matter,conductive minerals and pore fluids on the resistivity of shale were analyzed through the rock electrical experimental,and the main controlling factors of shale resistivity were clarified.Finally, the reservoir space of the shale reservoirs of the Longmaxi Formation in the study area was characterized by argon ion polishing scanning electron microscopy, low-field nuclear magnetic resonance,carbon dioxide and nitrogen adsorption.The results show that the ultra-low resistivity wells are mainly affected by the graphitization of organic matter,which leads to the exponential decrease of shale resistivity,while the low resistivity wells are mainly affected by the high water saturation,and the shale resistivity decreases relatively little.There are great differences in the microscopic pore structure of shale reservoirs with different resistivity.The ultra-low resistivity wells have the lowest porosity (the mean is 3.38%),and the worst inter-pore connectivity and openness (the mean hysteresis coefficient is 0.21).There was little difference between the porosity of low resistivity wells(The mean is 6.22%) and normal resistivity wells (the mean is 6.14%),and the normal resistivity wells (the mean hysteresis coefficient is 0.13) were better than those of low resistivity wells (the mean hysteresis coefficient is 0.15).

        Speaker: Yijiang Leng (Southwest Petroleum University)
      • 229
        Elastic indication of fluid patch clustering in partially saturated porous media: critical saturation model

        Knowledge related to the relations between elastic observations and fluid state cannot be overemphasized in porous media, which is of fundamental concern for the extraction of hydrocarbons, the monitoring of carbon dioxide geological sequestration, and the underground hydrogen storage. As known from previous theoretical and experimental experience, the P-wave velocity-saturation relation for partially saturated porous media is a concretization of the effect of fluid patch evolution on elastic properties. However, the P-wave velocities are not only governed by the overall saturation, but also depend on the fluid patches and their size. The patch size variation as saturation changes is commonly ignored in modelling investigations, even though it is natural to assume that fluid patches will form larger as saturation progresses and that percolating clusters will form at some critical saturation levels. To capture the evolution of patch size with saturation implied in the velocity-saturation relations, we are inspired by percolation theory. By incorporating the connectivity of water-filled patches in the continuous random medium model, we develop a critical saturation model. We apply this critical saturation model to examine recently reported experimental measurements, specifically analyzing the patch size changes. For measurements of drainage or imbibition processes in four sandstone samples, we indeed find a clear indication of growing patch size with water saturation. The predictions of the critical saturation model are in reasonable agreement with elastic observations. Our approach enhances the interpretation accuracy of the velocity-saturation relations and lays the foundation for a profound understanding of the effects of fluid clustering on elastic properties in partially saturated porous media.

        Speaker: Dr Qiang Liu (China University of Petroleum (East China)-Curtin University)
      • 230
        Analyzing uncertainties of the instability of the anode /electrolyte interface in solid state batteries

        Modern batteries must meet stringent performance standards to qualify for use in technological solutions that seek to address current global environmental challenges. Such batteries should exhibit high energy densities, fast charging, and long cycle lives while maintaining a high degree of safety. Solid-state batteries (SSBs) exploit high-capacity anode materials such as Lithium or Sodium metal and are expected to deliver high standards that meet the stringent needs of long-range electric vehicles and large-scale renewable energy storage. However, stability in these devices presents important challenges. The interface anode/electrolyte interface is home to structural imperfections that lead to heterogeneous stripping and plating during cell cycling, significantly reducing cell capacity and compromising cell safety. Although numerous studies have attempted to shed light on the root causes of inhomogeneous electrochemical processes at metals anodes in SSBs, the detailed atomistic processes that lead to ubiquitous dendrites growth in metal anodes are not well elucidated. Critically lacking is the detailed understanding of the thermodynamic driving forces that lead to such degradation at the atomistic level.
        We analyze the forward propagation of the imperfection parameters that are susceptible to highly defeat the stability of the anode/electrolyte interface. The imperfections in inputs are parametrized as random variable and Monte Carlo method and sensitivity analysis approaches allow a better understanding of Lithium plating and stripping behaviors.

        Speaker: Ben Mansour Dia (CPG, KFUPM)
    • MS01: 2.2
      • 231
        Realistic evaluation of prototypical porous materials for carbon capture

        Carbon capture, utilization, and storage (CCUS) is an attractive approach to help decarbonization from point sources, like energy supply and other industries, as well as for pulling CO2 out of the atmosphere (i.e., direct air capture, DAC). Among several approaches at differing technology readiness levels, solid sorbents are promising as they generally combine high uptakes and selectivity with milder regeneration energies.

        Adsorption screening and testing of promising materials are often performed using pure component or point uptake experiments, which only give information about adsorption capacity and ideal selectivity. At realistic process conditions, competitors such as moisture and temperature have a large effect on the uptake of CO2, wherein the presence of water could either increase CO2 capacity, compete for the same adsorption sites, or even induce material collapse. The kinetics on the other hand is another important factor for an effective separation.

        Figure 1 shows that apparent CO2 uptake decreases by 5% RH in Zeolite 13X. Figure 2 presents the details of the sorption kinetics of both components highlighting replacement effects.

        In this work, several porous materials including zeolites, MOFs, and functionalized resins are screened in realistic conditions for CO2 capture using advanced dynamic gravimetric sorption and breakthrough methods. The tests were conducted under varied conditions, e.g., different CO2 concentrations and relative humidity. The results showed that humidity is the key factor affecting the CO2 capture efficiency. This study provides a reference for screening the effective sorbents for carbon capture.

        Speaker: Meishan Guo (Surface Measurement Systems Ltd.)
      • 232
        Effect of reactive impurities in CO2 gas storage in carbonate reservoirs

        Storage of carbon in the form of compressed CO2 in the subsurface represents a potentially viable and cost-effective way to reduce emission of heat-trapping CO2 to the atmosphere. The feasibility of a CO2 storage scheme depends on many factors, including CO2-induced corrosion and scale; availability of inexpensive CO2 sources, available pipeline, pipeline integrity, temperature and pressure conditions, caprock integrity, biological activity in the subsurface, injectivity, mineral trapping and interactions with the rock surface amongst others. All these factors are potentially affected by the presence of impurities in the CO2 supply. The chemical composition of the CO2 stream will depend on the fuel sources and capture methods, and CO2 with impurities is much more widely available in sufficient quantities for transport to offshore facilities as capture processes and transport generally lead to some content of impurity. The effects of major impurities (SO2, N2 and O2) on phase behavior as well as corrosion in pipelines are quite well understood and widely reported in literature. Some studies have addressed the geochemical effects of impurities on the matrix in the well, mainly for shales and sandstone reservoirs. However, the geochemical effects of long-term storage of impurity-containing CO2 are not well known, particularly for carbonates, including chalk. The theoretical and experimental predictions of the interaction energies for complexes of CO2 with relevant impurities (aminoethanol, ethylene glycol, methanol, ethanol, water, H2S, NH3, CO, NOx and SO2) suggest that the interactions with CO2 vary and some of the species interact strongly even in small quantities compared to small gaseous impurities like N2, Argon and O2. Injection testing of these impurities in Danish North Sea Chalk cores are conducted in core flood experiments combined with chemical analysis of the effluent fluids and chalk surfaces to investigate the alterations caused by impurities.

        Speaker: Karen Feilberg (Danish Hydrocarbon Research and Technology Centre)
      • 233
        Effect of dissolution and heterogeneity on supercritical CO2 invasion in porous media: an experimental study using X-ray micro-computed tomographic imaging

        In this study, with the utilization of quasi-dynamic X-ray micro-computed tomographic (MCT) imaging, pore-scale fluid configurations were tracked for CO$_{2}$ injected into two different brine-saturated Bentheimer sandstone cores under conditions relevant to geologic carbon sequestration. CO$_{2}$ injection was performed at low capillary number (Ca = 10$^{-9}$) into cores saturated with live- and dead-brine, consecutively. Two cores with different pore space characteristics were used to investigate the impact of heterogeneity on the resultant fluid configurations. We also interrogated possible wettability alteration during CO$_{2}$ injection based on the obtained MCT images. We find that invasion patterns continue to evolve long after breakthrough, with distinct and gradual saturation changes occurring after decades of pore volumes injected. For one core, the invasion patterns for both live- and dead-brine conditions eventually converge after 16.5 pore volumes; for the second core, the patterns are distinct under the different injection conditions for up to 30.1 pore volumes. The presence of pore-scale heterogeneities in the cores has a strong influence on the ultimate CO$_{2}$ distribution under the different conditions. It is expected that results from this study will contribute to better understanding of the pore-scale invasion of CO$_{2}$ and ultimately, the field-scale application of geologic carbon sequestration.

        Speaker: Ruotong Huang
      • 234
        Pore-Scale Dynamics in Carbonate Reservoirs: Understanding Heterogeneity's Influence on CO2 Storage in Indiana Limestone

        Subsurface carbon dioxide (CO2) storage is one of the most critical strategies in combatting climate change. One of the principal challenges encountered by the Carbon Capture and Storage (CCS) industry is the accurate understanding, representation, and upscaling of fluid flow dynamics within targeted reservoir formations. This problem is rather complex in carbonate formations due to their varying spatial heterogeneities and complex pore structures. In our experiment, we assess the impact of microporosity, heterogeneity and connectivity on saturation changes, and trapping in Indiana limestone samples.
        We image Indiana limestone core samples using a high-resolution μCT scanner, with a resolution of 4.9 µm. Through two cycles of drainage and imbibition, we flooded the core with two different flow rates, to understand the influence of heterogeneity on the mobility of both wetting and non-wetting phases within the porous media. During the two cycles, the pore-scale capillary number was kept well within the capillary flow regime (10-7 - 10^-8). Our study highlights noticeable differences in saturation, non-wetting connective path, and dynamics of pore-filling between the two flooding cycles. Additionally, we show the redistribution of the non-wetting phase across the pore space when increasing the non-wetting phase flow rate. Furthermore, we investigate intermittent flow observed during imaging manifesting as artefacts within the reconstructed 3-D volume. This exploration aims to elucidate the origins and implications of intermittency, providing valuable insights into its impact on imaging quality and interpretation of pore-scale fluid dynamics.

        Speaker: Nihal Darraj (Imperial College London)
    • MS03: 2.2
      • 235
        Study on Foam Flow Behavior in Fractured-Vuggy Systems

        For deep fractured-vuggy carbonate reservoirs, foam flooding is an effective oil recovery method. However, the connectivity and anisotropy of the fractured-vuggy network affect the plugging performance of foam and the ability to adjust the displacement profile. Therefore, it is necessary to conduct a comprehensive investigation on the migration characteristics of foam, in order to provide guidance for the oilfield application of foam flooding.
        The fractured-vuggy system exhibits heterogeneity and strong diversion capabilities. When developing a model that can represent reservoirs with fractured-vuggy formations, it is challenging to simultaneously satisfy the characteristics of multiple experiments with a single model. The flow behavior of foam in fractured-vuggy system is a crucial factor that needs to be observed, so it is necessary to appropriately relax the requirements for simulating reservoir temperature and pressure conditions. Based on the combination relationships of fractures, wall effects, and fluid properties, a multi-dimensional and multi-scale fractured-vuggy model was developed. This model, combined with the selected foam system, was used to study the evolution of foam structure, flow characteristics, gas-liquid distribution patterns, and oil displacement properties within the fractured-vuggy model. The study summarized the dynamic and static matching relationships between fractured-vuggy dimensions and foam, investigated the improvement effects of foam on shielding fractured-vuggy flow, and comprehensively analyzed the changes in the foam displacement front and the different distribution characteristics of gas and liquid in fractures under the influence of various factors. The study clarified the foam displacement characteristics corresponding to different production scenarios.
        The experimental results show that, due to limitations in the channel dimensions, there are differences in the quantity and shape of foam distribution within fractured-vuggy formations after injection. Significant variations also exist in the evolution patterns during the static stable stage of foam. The shielding effect of foam displacement between fractures is dynamically adjusted. This is because high-quality stable foam gradually "plugs" dominant fractures, increasing the flow resistance for subsequent foam in the dominant fractures. Consequently, some foam is still able to divert towards the inferior fractures.

        Speaker: Dr Zhengxiao Xu (Changzhou University)
      • 236
        Experimental study on the effect of supercritical CO2 and acid alternative injection mode on the acid-etching behavior and conductivity of fracture in carbonate rocks

        In order to enhance acid penetration depth and fracture conductivity, acid fracturing techniques involving the alternating injection of non-reactive fluids (fracturing fluids) or weakly reactive fluids (self-generated acid) with acid are considered a pivotal enhanced oil recovery rate in carbonate reservoirs. In recent years, the CO2-enhanced acid fracturing technique has gained prominence in the Middle East. This method adopts a mixed injection mode of CO2 and acid liquid in the wellbore, featuring the advantageous effects of retarding acid-rock reaction rates, improving fracture conductivity, and conserving water. However, its application in deep wells is limited by the high friction associated with the mixture of CO2 and acid. Supercritical CO2 and acid alternating injection, conducted under the conditions of conventional acid fracturing with established dominant fracture channels, involves supercritical CO2 injection to reduce the flow resistance of CO2 into the reservoir, showcasing potential for application in deep wells. Nevertheless, the impact of supercritical CO2 and acid alternating injection on hydraulic fractures and conductivity has not been reported.

        This paper utilizes a self-developed supercritical CO2 acid etched fracture conductivity simulation device and employs downhole cores to conduct experiments on hydraulic fracture acid etching and conductivity under two scenarios: alternating injection of weakly reactive fluid (self-generated acid) with gelled acid and alternating injection of supercritical CO2 with gelled acid.

        Research results indicate that supercritical CO2 and gelled acid alternating etching exhibits more pronounced elevation variations on the fracture surface, demonstrating a stronger and more dispersed non-uniform etching compared to the self-generated acid/gelled acid alternating injection mode. In terms of fracture conductivity, CO2/gelled acid alternating injection exhibits higher initial fracture conductivity, and under medium to high closure pressures (≤52MPa), self-generated acid/gelled acid alternating injection demonstrates higher conductivity. However, at high closure pressures (>52MPa), the supercritical CO2/gelled acid alternating injection mode can create higher fracture conductivity.

        This study experimentally explores the impact of supercritical CO2 and acid alternating injection on fracture conductivity, demonstrating that for deep carbonate reservoirs with high closure stress, adopting this mode has the effect of reducing the usage of non-reactive or weakly reactive fluids and enhancing fracture conductivity under high closure pressure.

        Speaker: Ke Xu (Southwest Petroleum University)
      • 237
        Bubble growth and induced flow characteristics in porous media under heating conditions

        The release of trapped bubbles from dead-end porous media filled with nonvolatile liquid holds extensive applications in gas‒liquid reactors, CO2-assisted srteam flooding, ceramic sintering, and droplet microcarriers. Since traditional pressure-driven flow fails to induce bubble transport in dead-end pores, this study investigates the potential use of heating to control the release of bubbles from dead-end porous media. This study addresses the issue of bubble retention in dead-end porous media and designs various microfluidic chips with different porous structures, including upward sparse and downward dense, upward dense and downward sparse, and isotropic porous media. The porosity of the sparse and dense regions is 0.7355 and 0.8718, respectively. Using CO2 as the gas and dimethyl silicone oil as the liquid, a self-built Micro-PIV visualization experimental system is employed to investigate the influence of porous media pore structures on the growth and release of bubbles, as well as the induced flow field patterns under heating conditions. The results show that an increase in temperature leads to the transfer of dissolved gas to the bubble, resulting in an increase in bubble pressure, which serves as the driving force for the bubble to pass through the pore throat. Under heating conditions, changes occur in the surface tension of the gas-liquid interface and the viscosity of the liquid phase, while capillary pressure is a key factor for the bubble to pass through the channel. Blockage occurs when the bubble interface capillary pressure is less than the threshold pressure and breakthrough happens when it is greater. The structure of porous media with dense upper and sparse lower regions impedes bubbles from entering low porosity zones, reducing the coalescence probability. Conversely, the structure with sparse upper and dense lower regions increases the coalescence probability, facilitating bubble expulsion. When bubbles seal pores, their continuous growth causes the liquid velocity of corner film flow to rise. During bubble release, the surrounding liquid accelerates to fill the original space, causing a simultaneous transition that induces vortices along the microcolumn walls.

        Speaker: Zhi Feng (China Jiliang University)
    • MS06-A: 2.2
      • 238
        Aging of liquid foam in porous media

        One of the key stages in materials recycling is their crushing into finer elements, i.e., granular material or powder to be sorted and re-used. Those crushed granular materials can be mixed and reshaped using binders which will make it possible to reproduce solid objects with useful purposes. Mainly, the major practical difficulty in the implementation of these granular assemblies, whether recycled or not, is the dispersion of the binder at the contacts between the grains in order to produce binder bridges that will ensure the cohesion of the whole.
        Complex liquid foam (liquid foam loaded with a binding component) represents a first-choice low carbon binder precursor to be pushed through the voids offered by a packing made with such grains, to give shape to the whole and to confer significant mechanical strength. This strength is expected to depend on the microstructure of the confined foam, the latter being controlled by the bubble-to-pore size ratio "r" [1]. However, as the liquid foam undergoes the so-called coarsening mechanism, which consists in the exchange of gas between the different bubbles, the size ratio increases as function of time.
        Here, we study the coarsening of liquid foam confined into the porosity of granular packings. During these experiments the liquid fraction is maintained uniform in the system by appropriated rotation of the samples in order to counteract the effects of gravity (see Figure 1). We show that coarsening is faster whenever we increase the initial confinement parameter r0 at constant liquid saturation. We find the result known from the literature for which the bubbles eventually reach pore size, which marks the end of coarsening, but we also highlighted that before stopping, there exists a regime of self-accelerated coarsening. The main deviations with respect to the coarsening of unconfined foams will be presented.

        Speaker: Mr Ali Salamé (Laboratoire Navier, Univ Gustave Eiffel, Ecole des Ponts, CNRS-UMR8205)
      • 239
        Trapping criteria for three-dimensional periodic liquid particles in micropillar scaffolds

        Liquid particles within three-dimensional periodic scaffolds play a crucial role in various natural and engineering applications, for example, cellular arrays composing living tissue, 3-D materials that mimic tissue with an unprecedented level of control, and innovative liquid-infused materials designed for carbon capture. Although it has been known that fluid interfacial energy during multiphase displacement can drive the emergence of 3-D periodic liquid particles in micropillar scaffolds, the underlying microscale physics and the macroscopic formations of liquid particles in scaffolds remain unclear. Here we establish trapping criteria for the formation of liquid particles in micropillar scaffolds based on the evolution of fluid-fluid interfacial curvature during multiphase displacement, considering four key parameters (pillar size α, contact angle θ, capillary number Ca, and viscosity ratio M). The fundamental trapping criteria are theoretically derived in α-θ space under a viscous stable state and main meniscus-dominated flow, where critical α distinguishes trapping mode and complete displacement, and critical θ further distinguishes between diamond-like and spherical particles in the trapping mode. However, the critical pillar size α for trapping mode or completed displacement can be further affected by viscous instability under lower viscosity ratio M and inter-pillar interface-dominated flow under lower capillary number Ca. These criteria are validated through numerical CFD simulations and confirmed by microfluidic experiments. These results indicate the conditions for trapping 3-D periodic liquid particles in micropillar scaffolds, offering insights that extend and clarify prior literature. The derived criteria provide valuable guidelines for the design of micropillar scaffolds and for the emergence of 3-D periodic liquid particles in micropillar scaffolds under various conditions by controlling multiphase displacement processes.

        Speaker: Dr Wenhai Lei (KTH Royal Institute of Technology)
      • 240
        Spontaneous Symmetry Breaking during Dispersed Fluid Flow through Porous Media

        Dispersed fluid systems (foam, emulsion, bubbly liquid, etc.) involves in many key geophysical/geochemical, environmental, and engineering processes[1-3]. However, regardless of many pore-scale and channel-scale approaches[4, 5], predicting dispersed fluid flow behavior in porous media is still a major challenge. Here we conduct experimental and theoretical investigation, trying to rationalize the long-puzzling gap between single-channel scale and porous media scale models for dispersed fluid flow.
        We first conduct experiments in doublet microfluidic model (Fig. 1a). We fix $Ca_d$ (dispersed fluid capillary number) and observe the flow state at varying $Ca$ (total capillary number). At low Ca, very significant difference for blob fluxes between two parallel channels are identified even when the two channels are highly symmetric (Fig. 1b). After careful quantitative analysis, we realize that minor fabrication error cannot rationalize this symmetry breaking.
        We thus hypothesize that this asymmetric flow of dispersed fluid is originated from a new symmetry breaking mechanism. We thus conduct stability analysis in a doublet system, that introduce an infinitesimal perturbation to assess its potential for self-amplification. Analysis shows that such symmetry breaking can emerge, if there is a negative correlation between pressure drop and total flow rate at a constant dispersed fluid flux ($\left.\frac{\partial F}{\partial Ca}\right|_{Ca_d} < 0$) within a specific channel. The existence of $\left.\frac{\partial F}{\partial Ca}\right|_{Ca_d} < 0$ correlation is successfully validated: theoretically, by classic Bretherton's correlation for non-viscous gas slug flow in a uniform cylindrical tube [6] (Fig. 1c); and experimentally, by microfluidic experiment along a single channel of sequential pore-throat structure (Fig. 1d).
        This breaking of flow path symmetry, if emerges in porous media, may result in preferential flow even in homogeneous media. We conduct a demonstrative experiment in a homogeneous porous medium to validate the above inference. Surprisingly, we do observe significant non-uniform flow at steady state, as shown in Fig. 1e and Fig. 1f. Preferential paths carry almost all dispersed fluid flux, while blobs in other paths flow only occasionally and slowly. Noticeably, the dispersed fluid saturation is negatively correlated with its flux, which is contradictory against classic relative permeability idea, but can be well explained by the abovementioned spontaneous symmetry breaking.
        This discovery of spontaneous symmetry breaking of dispersed fluid in porous media may bring new insight into the understanding and modeling of complex fluid behaviors in disordered geometry.
        (a) The design of the doublet-channel model, single-channel model and homogeneous porous media model. (b) The experimental results of the cumulative number of droplets in doublet system for Cad = 1.5×10-5. Hollow symbols denote channel 1 results, while solid symbols represent channel 2 results. (c) Schematic diagram of F-Ca curve in a single channel for Bretherton model. (d) The experimental results of F-Ca curve along a single channel with Cad = 3.0×10-5 (in the main plot, and colors mark different regimes), and with Cad = 2.5×10-4 and 5.0×10-4 (in the inserted plotting). (e) Time-average dispersed fluid saturation map of the whole micromodel. (f) Time-average blob-passing frequency field map of the whole micromodel.

        Speaker: Dr Jie Qi (Peking University)
      • 241
        A three-layer Hele-Shaw problem driven by a sink

        We investigate a sink-driven three-layer flow in a radial Hele-Shaw cell performing numerical simulations. The three fluids are of different viscosities with one fluid occupying an annulus-like domain, forming two interfaces with the other two fluids. Using a boundary integral method and a semi-implicit time stepping scheme, we alleviate the numerical stiffness in updating the interfaces and achieve spectral accuracy in space. The interaction between the two interfaces introduces novel dynamics leading to rich pattern formation phenomena, manifested by two typical events: either one of the two interfaces reaches the sink faster than the other (cusp-like morphology) or they touch each other (interface merging). In particular, the inner interface can be wrapped by the other to have both scenarios. We find that multiple parameters contribute to the dynamics including the width of annular region, the location of the sink, and the mobilities of the fluids.

        Speaker: Meng Zhao
    • MS07: 2.2
      • 242
        A robust two-level overlapping preconditioner for Darcy flow in high-contrast porous media

        In this talk, we present a two-level overlapping domain decomposition preconditioner for solving linear algebraic systems obtained from simulating Darcy flow in high-contrast media. Our preconditioner starts at a mixed finite element method for discretizing the partial differential equation by Darcy?s law with the no-flux boundary condition and is then followed by a velocity elimination technique to yield a linear algebraic system with only unknowns of pressure. Then, our main objective is to design a robust and efficient domain decomposition preconditioner for this system, which is accomplished by engineering a multiscale coarse space that is capable of characterizing high-contrast features of the permeability field. A generalized eigenvalue problem is solved in each non-overlapping coarse element in a communication- free manner to form the global solver, which are accompanied by local solvers originated from additive Schwarz methods but with a non-Galerkin discretization to derive the two-level preconditioner. We provide a rigorous analysis indicating that the condition number of the preconditioned system could be bounded above with several assumptions. Extensive numerical experiments with various types of three-dimensional high-contrast models are exhibited. In particular, we study the robustness against the contrast of the media as well as the influences of numbers of eigenfunctions, oversampling sizes, and subdomain partitions on the efficiency of the proposed preconditioner. Besides, strong and weak scalability performances are also examined. The work is partially supported by the Hong Kong RGC General Research Fund (Projects: 14305222 and 14304021).

        Speaker: Prof. Eric Chung (Chinese University of Hong Kong)
      • 243
        Exploration of robust and fast L-splitting schemes for nonlinear double degenerate equations

        Nonlinear advection diffusion equations model diverse physical phenomena.
        Some examples include flow through porous media (as found in subsurface and reactive flows),biological processes, Stefan problem and permafrost models; and many others. In this work, we investigate numerical methods for nonlinear parabolic equations that show doubly degeneracy, i.e. for example the diffusion coefficient of the equation is allowed to vanish (degenerate diffusion) at zero concentration which leads to hyperbolic equation with free-boundaries as observed in nature, and become singular at full concentration leading to an elliptic problem. First we propose a semi-implicit (Backward Euler) time discretization. Implicit time stepping methods are popular due to their stability, allowing to avoid severe restrictions on the time step. This leads to nonlinear, time-discrete elliptic equations, for which linear iterative schemes are needed for approximating the solution, which combines the features of the Newton method and the L-scheme, i.e., a modified L-scheme. The linearization scheme is shown to be globally convergent (even for double degenerate cases). Moreover, it is linearly converging in the non-degenerate case accelerate with a small time-step. Numerical results will present which revealed that it is robust and stable when compared to the standard linearization schemes.

        Speaker: Ayesha Javed (Hasselt University)
      • 244
        Stable unfitted finite element method for poroelasticity with weak discontinuity

        The traditional finite element method requires that the mesh must match with various discontinuous, which can significantly increase the difficulty of preprocessing for hydrodynamic coupling problems with complex boundaries and material interfaces. In such case, the finite element method using unfitted mesh is obviously more advantageous, however, this method also has certain problems, for example, irregular mesh cutting may lead to ill-conditioned coefficient matrix to appear, which in turn affects the accuracy and stability of the algorithm. The ghost penalty technique was proposed to overcome the ill-conditioning issue. Recently, an unfitted finite element was proposed for two-field poroelasticity problem, where stabilization terms based on the ghost penalty were developed. Material interfaces are even more difficult to deal with than the boundary as they require careful treatment of the weak discontinuity conditions as well as the mesh cutting stabilization. In this paper, we formulate an unfitted finite element for the poroelastic problem with both material interfaces and complex boundaries. A weak formulation based on the Nitsche’s method was developed. Ghost penalty stabilization terms are designed for both sides of the elements intersected by the material interface. The performance of the proposed methodology is tested by several benchmark and practical hydraulic problems of complicated rock-soil mixtures. The numerical results demonstrate optimal convergence rates and low-level condition numbers independent of the mesh cutting.

        Speaker: Prof. Zhijun Liu (Lanzhou Univ)
      • 245
        Multiscale Extended Finite Element Method for the Simulation of Fractured Geological Formations

        In the prevailing context of the 21st century, characterized by a predominant reliance on oil and gas, or in the promising future where green energy shapes a human society committed to net-zero emissions, the role of underground fractured formations in energy production (geothermal) and storage remains pivotal and irreplaceable. In the past decade, hazardous consequences of failing to predict the geomechanics behaviors of fractured formations has led to a pronounced focus on developing simulation strategies that are both accurate and efficient for subsurface fractured formations.
        As a widely used simulation method in fracture mechanics, the extended finite element method (XFEM) provides a precise approach to simulate deformation and fractures propagation within highly fractured media. It is also a convenient strategy as it allows for the use of structured grids. However, the expensive computational cost of using classical XFEM in the simulation of fracture networks makes this method not immediately suitable in the geoscientific community.
        To resolve this challenge, a multiscale extended finite element method (MS-XFEM) is proposed to provide a novel approach to simulate the highly fractured subsurface formations accurately and efficiently. The deformation and fractures propagation are both simulated by interpolating the solutions from a larger yet sparser coarse grid to the original fine-scale grid. This interpolation process requires the construction of the basis functions matrix. The novelty of this work is to involve the fractures into basis functions only, thus the coarse scale system is constructed based on a standard finite element method. More importantly, this construction of basis functions is fully algebraic and can be updated locally and adaptively for the simulation of propagating fractures. This method has been implemented and tested to prove its efficiency and accuracy. All test results prove the good qualities of solutions computed from MS-XFEM when compared to fine scale XFEM solutions. Basis functions are constructed successfully with the algebraic method since they capture all different types of discontinuities. These tests reveal the potential of MS-XFEM in simulating real-world subsurface fractured formations.

        Speaker: Fanxiang Xu
    • MS10: 2.2
      • 246
        Pore structure evolution of low-permeability sandstone under acid treatment: a Micro-CT investigation

        The use of acid for permeability enhancement has gained popularity in mining and oil industries to enhance the recovery rate of low-permeability formations. This study employed static acid permeability enhancement tests, flooding acid permeability enhancement tests, and micro-CT scanning to investigate the mechanism of permeability enhancement and changes in pore structure during acid treatment. The extent of reaction between low-permeability sandstone samples and four different acids was evaluated by static tests, with hydrochloric and formic acids demonstrating good performance in dissolving filling minerals. Acid flooding experiments were conducted under reservoir conditions with a constant flow rate, and decreases in pressure difference between flow inlet and outlet were observed for most experiments, indicating an increase in permeability. The pressure difference was lower for hydrochloric acid compared to formic acid at the end of flooding, with permeability increases of 283% and 120%, respectively. Micro-CT scanning before and after acid permeability enhancement tests revealed changes in pores, pore throats, and coordination numbers using Avizo software. Based on micro-CT results, acid treatment led to an increase in the number of interconnected pores, pore throats, and their equivalent radii, resulting in higher permeability. The improved permeability was primarily due to the dissolution of dolomite, as identified by SEM-EDS and ICP.

        Speaker: Prof. Liwei Zhang (Institute of Rock and Soil Mechanics, Chinese Academy of Sciences)
      • 247
        Void deformation and connecting visualization in asphalt mixture under dynamic water pressure

        Water transport under dynamic vehicle load is the primary causation of asphalt pavement water damage. The dynamic water breaks through the inner voids and destroys the micro-structure of asphalt mixture, and consequently degrades asphalt pavement durability. Understanding the microstructure evolution and void connecting during dynamic water load contributes to the water damage mechanism of asphalt pavement.
        This study developed a water seepage device for asphalt mixture and used pulse water pressure to simulate the dynamic water load caused by tire crimping. The pulse sinusoidal water pressure with a frequency of 10Hz and a range from 0 to 0.7 MPa was served. X-ray CT scanning was performed on the dry asphalt mixture and also in-situ seepage asphalt mixture after 3, 8, 15, and 20h water load. A 3D digital void model was developed to analyze the void structure evolution. The translation, volume changing, and connecting in void structure were recognized and analyzed. The 3D water-activated void and water passageway were reconstructed.
        The result explained the variation of dynamic flow rate curves by analyzing the microstructure evolution during dynamic water pressure load. The deformation characteristics of voids were addressed and its contribution to water extension was analyzed. The 3D visible water passageway and saturation showed the dynamic water transport process and gave direct evidence of the macro seepage behavior variation.
        This study proposes a method to quantify the void deformation in asphalt mixture and explains the macro seepage behavior from the micro aspect. It contributes to understanding the water transport in asphalt pavement and improving its durability.

        Speaker: Hao Shi
      • 248
        Experimental study on optimization of acidizing acidizing fluid in heterogeneous oolitic limestone reservoir

        The oolitic limestone reservoir of Qianjiang Formation in Jianghan Oilfield has the characteristics of shallow burial, thin layer, developed upper and lower water layers and strong heterogeneity, so it is difficult to be reformed on site. In order to optimize the transformation process and the optimal process parameters suitable for the reservoir, the core dissolution experiment and acid displacement experiment of two different acid systems of conventional acid and retarded acid were carried out in this paper. Combined with the analysis of CT scanning results, the action law of oolitic limestone and different acid systems, the quantitative characterization of permeability change at different injection rates and the CT three-dimensional imaging of acid etching pore characteristics were clarified; at the same time, according to the results of acid rock reaction kinetics experiment and acid etching conductivity experiment, the effective distance of acid rock reaction in different acid system is predicted theoretically, and the best acid system and injection parameters are optimized. The results showed that the optimum concentrations of hydrochloric acid and retarded acid were 20 % and 10 %, respectively. The displacement rate has no obvious difference in the increase of permeability after core acidification, but it has a great influence on the wormhole structure formed by acid etching. The increase of displacement rate will form wormholes with better connectivity and less damage to rock skeleton and physical properties; by calculating the experimental parameters of acid-rock reaction kinetics, it is found that the effective distance of retarded acid is 10 m longer than that of conventional acid under the same injection rate of large displacement acid. Finally, according to the characteristics of the reservoir, a large displacement injection of ' 10 % retarded acid + 0.3 % corrosion inhibitor ' acid system can effectively increase the acid action distance and enhance the transformation effect.

        Speaker: Yanying Qu
      • 249
        Flow field tomography identifies and quantifies pore opening and clogging in sandstones

        Fluid-rock interactions drive changes in porosity and permeability. This has important consequences for the flow field development in the complex porous material and thus controls the evolution of reactive transport processes. Important applications are in the vast field of reservoir rock alteration, e.g. by coupled dissolution and precipitation processes. While dissolution processes can cause local increases in pore space and permeability, they can also lead to pore throat blockage, which can cause formation damage due to precipitation reactions and particle retention in pore throats. Although these mechanisms are understood in principle, the direct changes in the flow field they cause are difficult to observe directly. Using positron emission tomography (PET), we show how flow field heterogeneities are quantitatively affected by the coupling of dissolution reactions and pore throat blockage by particles in a long-term experiment.

        Specifically, we performed a dissolution experiment focusing on calcite cement in sandstones. While dissolution is responsible for a local increase in pore space, mobilized iron oxide and sheet silicate colloids are trapped and cause a local decrease in permeability. Direct comparison of sequences of PET-derived flow field data reveals a pattern of flow field modification during this experiment. PET thus becomes a key analytical tool to localize and quantify pore-scale flow field changes, in addition to recent advances focused on the identification of flow channeling effects of advective flow [1]and on the heterogeneity of diffusive flux in low permeability rocks [2].

        1. Pingel, J. L.; Kulenkampff, J.; Jara-Heredia, D.; Stoll, M.; Zhou, W.; Fischer, C.; Schäfer, T., In-situ flow visualization with Geo-Positron-Emission-Tomography in a granite fracture from Soultz-sous-Forêts, France. Geothermics 2023, 111, 102705.

        2. Bollermann, T.; Yuan, T.; Kulenkampff, J.; Stumpf, T.; Fischer, C., Pore network and solute flux pattern analysis towards improved predictability of diffusive transport in argillaceous host rocks. Chemical Geology 2022, 606, 120997.

        Speaker: Cornelius Fischer (Helmholtz-Zentrum Dresden-Rossendorf)
    • MS13: 2.2
      • 250
        Quantifying oil- and water-wettable pore networks of the lacustrine- and marine-sourced shale oil reservoirs

        Microscopic pore structure (both geometry and connectivity) characteristics control fluid flow and hydrocarbon movement in shale oil reservoirs. Considering the uniquely wide spectrum of pore sizes (nm to sub-mm), microscale mixed wettability, as well as the interplay of pore structure and wettability in organic-rich shale oil reservoirs, this work presents various approaches to quantifying the oil- and water-wettable pore networks for several important tight oil formations in China and USA with different depositions, as well as a range of maturation and mineral compositions. The approaches include the utility of different wetting fluids (deionized water or API brine, n-decane and/or toluene, isopropyl alcohol or tetrahydrofuran or dimethylformamide), fluid pycnometry, fluid immersion porosimetry after vacuum saturation, mercury intrusion porosimetry, nuclear magnetic resonance, and field emission-scanning electron microscopy. In particular, (ultra-) small angle neutron & X-ray scattering techniques, (U)SANS & (U)SAXS, are used to quantify the total (both edge-accessible and isolated) porosity and characterize pore size distribution in a pore length size from 1 nm to 10 m; in addition, the employment of contrast matching technique of (U)SANS enables the discrimination of accessible (open) pores and inaccessible (closed) pores to a particular liquid fluid. For example, our results show that the marine-sourced Bakken samples in USA have a relatively high total porosity (8.87-12.95%) with no more than 30% of the pores are accessible from sample surface, and are not preferentially wet by oil or water.

        Speaker: Mr Cunjian Zhang
      • 251
        Theory of electrolyte solutions in a slit charged pore: Effects of structural interactions and specific adsorption of ions

        The study of electrolyte solutions in confined geometries is crucial for developing energy storage devices and water purification systems. One challenge in this field is accurate modeling ion behavior while considering the interplay between ion-specific effects and electrostatic interactions. Although self-consistent field theory enables the simulation of ionic fluids, it overlooks several effects, such as short-range correlations of the ions. Recent research [1] incorporates into the grand thermodynamic potential of ionic fluid the short-range correlations of the ions and derives the mechanical equilibrium condition using the Noether's theorem formalism. In particular, the authors have derived the total stress tensor of ionic fluid taking into account electrostatic and steric interactions alongside the structural effects (short-range correlations).

        In this work, we use the model developed by Blossey et al. [2], which takes into account both structural and steric interactions between ions. The structural interactions are described through a bilinear form of the gradients of the local ionic concentrations, while the steric interactions are modeled using the lattice gas approach.This framework allows for a phenomenological description of the molecular properties of ions, such as steric interactions due to their non-spherical shape, changes in configuration, and the influence of the solvent. Additionally, we investigate the specific interactions between ions and pore surfaces by incorporating external attractive forces.

        Our main interest lies in analyzing how ionic concentration profiles and disjoining pressure are influenced by variations in pore size. Starting from the local mechanical equilibrium condition, we derive a general formula for the disjoining pressure.
        Our findings [3] indicate that taking into account the structural interactions between ions leads to a pronounce minimum in the disjoining pressure curves at small pore widths. This minimum is attributed to the formation of electrical double layers on the electrically charged surfaces of the pores.In addition, our results indicate that the attractive interactions between ions and the pore walls contribute to the formation of this minimum and shift it to smaller pore sizes. These theoretical findings have practical implications for researchers in the field of electrochemical engineering for supercapacitors, particularly in applications involving porous electrodes filled with concentrated electrolytes and room temperature ionic liquids.

        [1] Brandyshev P. E., Budkov Y. A. Noether’s second theorem and covariant field theory of mechanical stresses in inhomogeneous ionic liquids. The Journal of chemical physics. – 2023.– Т. 158. – No. 17.
        [2] Blossey R., Maggs A. C., Podgornik R.Structural interactions in ionic liquids linked to higher-order Poisson-Boltzmann equations. Physical Review E. – 2017
        [3] Victoria A. Vasileva, Daria A. Mazur, Yury A. Budkov. Theory of electrolyte solutions in a slit charged pore: Effects of structural interactions and specific adsorption of ions. Journal of Chemical Physics. 2023. Vol. 159. No. 2. Article 024709

        Speaker: Daria Mazur (Laboratory of Computational Physics, HSE University,)
      • 252
        Using fractal theory to study the influence of movable oil on the pore structure of different types of shale: A case study of the Fengcheng Formation shale in Well X of Mahu Sag, Junggar Basin, China

        In order to investigate the influence of movable oil on the pore structure of various shale types, this study systematically selected 19 shale samples from Well X in the Mahu Sag of the Junggar Basin. Initially, X-ray diffraction (XRD) analysis was conducted to classify the shale samples. Subsequently, the geochemical properties and pore structures of the samples, both pre and post oil extraction, were comparatively analyzed through Total Organic Carbon (TOC) content measurement, rock pyrolysis, and nitrogen adsorption experiments. Additionally, fractal theory was employed to quantitatively describe the impact of movable oil on the pore structure of different shale types.
        The findings reveal that siliceous shale exhibits a higher content of movable oil compared to calcareous shale. Following oil extraction, there was a notable increase in both specific surface area and pore volume across all shale samples, with a more pronounced variation observed in the pore structure of siliceous shale as opposed to calcareous shale. Calcareous shale predominantly displays H2-H3 type hysteresis loops, indicative of ink-bottle-shaped pores, suggesting a relatively uniform pore structure. Conversely, siliceous shale exhibits a diverse range of hysteresis loops, reflecting its complex pore structure. The fractal dimension of calcareous shale samples appears primarily influenced by pore structure, exhibiting no significant correlation with TOC content before or after oil extraction. Conversely, the change in fractal dimension of siliceous shale samples demonstrates no clear correlation with either TOC content or pore structure, suggesting that variations in fractal dimension may arise from the combined effects of TOC and pore structure.

        Speaker: Ms Hong Zhang (Peking University)
    • MS15: 2.2
      • 253
        A deep learning enabled massive parallel simulator for porous media flow

        Due to the complex composition of oil and gas resources, reservoir engineers usually switch between different mathematical models when describing the properties of petroleum reservoirs. In addition to the commonly used black oil model, various compositional models have been proposed. Some EOR techniques, such as polymer flooding, must be simulated based on the framework of compositional models. Some other applications of porous media flow, such as CO2 sequestration, groundwater contamination, and geothermal resource development, can also be simulated using compositional models. But the compositional models tend to be associated with more complex PDEs, more variables, and higher computational costs. In this talk, we will discuss a general-purpose compositional framework and our efforts in developing its solution methods, including discretizations, nonlinear solvers, linear solvers, parallelization and AI capabilities. Furthermore, we will introduce an open-source software project for simulating multi-component multi-phase porous media flow.

        Speaker: Prof. Chensong Zhang (Academy of Mathematics and Systems Science)
      • 254
        Application of Machine Learning and Deep Learning Methods in Reservoir Development

        Machine learning (ML) has revolutionized various aspects of underground seepage, geological modeling, reservoir numerical simulation, production optimization, and big data analysis in the oil and gas industry. In particular, when it comes to reservoir development, ML methods, e.g., deep learning (DL) and intelligent computing, have proven to be superior to traditional methods in terms of effectiveness and efficiency. Our study focuses on the application of cutting-edge ML methods to real-time reservoir optimization problems. The research includes reservoir history matching, well placement optimization, production optimization, wellbore fault diagnosis, big data analysis, and so on. Through extensive research and experimentation, we have observed that ML-based methods, especially DL methods, not only enhance the performance of traditional techniques but also significantly reduce the computational effort. They can quickly give reliable prediction results for variables of interest almost within seconds. ML-based methods can also accurately predict the performance of stimulation measures in oilfields, where an ML-based model is obtained with only available data, instead of expert knowledge in oilfields utilized in traditional patterns. These findings demonstrate the immense potential of ML methods in improving the performance of traditional techniques, providing valuable insights for practical oilfield management and development.

        Speaker: Prof. Kai Zhang (China University of Petroleum (East China);Qingdao University of Technology)
      • 255
        3D Pore Segmentation and Pore-Scale Simulation by Deep Learning

        Objectives
        This study aims to optimize the characterization and prediction of permeability and relative permeability in porous media through a multi-faceted approach. The primary objectives include achieving accurate 3D reconstruction of rock core images, implementing advanced deep learning models for segmentation, and addressing computational challenges associated with the Lattice Boltzmann Method (LBM) for permeability calculations. Furthermore, the goal is to enhance scalability by training models on relatively small representative elementary volumes (REVs).
        Methodology
        In this work, we perform the data acquisition of rock core images through CT scanning, followed by the 3D reconstruction of multiple 2D slices. We develop deep learning-based segmentation models to enhance pore segmentation, including 3D UNet, Attention-3D UNet, and 3D UNet-transformer. To resolve the computational bottleneck of LBM for large REV, we develop 3D-CNN models to take 3D segmented pore images offered by the segmentation network as input and learn to predict labels of permeability and relative permeability curves generated by LBM solver at relatively small REVs. Once trained, these 3D-CNN models can predict rock properties on a larger REV without computational limitations.
        Results
        The 3D reconstruction using the 2D slices yields highly accurate representations of rock core images. Subsequent segmentation employing 3D UNet, Attention-3D UNet, and 3D UNet-transformer identify the most accurate and effective porous network, providing valuable insights into the rock structure. The segmentation model achieves scalability by efficiently predicting permeability for larger REVs by addressing the computational challenges associated with LBM. The permeability and relative permeability values determined through up-scaled predictions of the trained model closely align with LBM data, affirming the reliability and utility of the integrated approach. This research contributes a comprehensive and efficient workflow for permeability prediction in porous media, combining advanced imaging, deep learning segmentation, and permeability prediction.
        Additive Information
        The deep learning-based workflow developed in this work can efficiently solve the scalability limitation of physics-based solvers, such as the Lattice Boltzmann Method, as it can accelerate the computation and extend to predicting permeability in larger REVs, which is about scalability. This dual advantage of efficiency and scalability represents a notable breakthrough, highlighting the efficiency and effectiveness of our methodology in overcoming the limitations occurring in digital rock.

        Speaker: Haotian Li
      • 256
        A multi-well deep learning model considering geological and engineering parameters for the long-term forecasting of shale gas production

        The production forecasting of shale gas wells is an important research topic in the natural gas industry. The underground pore structure is extremely complicated after hydraulic fracturing. Conventionally, researchers try to construct forecasting models via theoretical or experimental approaches. However, both theoretical and experimental approaches are faced with difficulties. Firstly, many parameters required during the theoretical computations cannot be timely or accurately measured, which limits the usage of theoretical models. Secondly, experiments can hardly match the real scale since a typical shale gas well has several thousand meters in length and depth. As a result, machine learning has become a popular modelling approach in the production forecasting of shale gas wells under complex pore structures induced by hydraulic fracturing.

        With the large-scale development of shale gas blocks, an increasing number of production records as well as geological and well completion data is becoming available. This enables the construction of shale gas production forecasting models using dynamic time series data as well as static geological and well completion data. The former represents the historical production features of a well while the latter reflects the underground pore structure. Nevertheless, most existing methods do not incorporate static parameters into the modelling process. Moreover, most existing methods apply a single-well modelling scheme where a model specifically trained for a target well is constructed. This scheme has two major flaws when it comes to field application. Firstly, it overlooks useful information from other existing wells in the same production block. Secondly, it requires a long piece of initial data for training and therefore cannot be used on newly developed wells. The problems faced with single-object methods have also been addressed in many other fields where the use of a long piece of initial data is impractical [1].

        Considering the problems in existing shale gas production forecasting models, we propose a multi-well long-term forecasting model with static parameters. The proposed model is built on recurrent and fully connected neural networks and is trained on both production and static data from multiple existing wells. Since the training process does not need a long piece of production history from the target well, the proposed model can be used on a newly opened or re-opened well after a short initial period (for example, 14 days). Then, the long-term forecasting can be carried out iteratively using previously forecasted values. Therefore, the proposed model can achieve timely forecasting of shale gas production with both time-series data and pore structure after fracturing taken into consideration. The accuracy and usefulness of machine-learning-based shale gas production forecasting models are thus improved.

        Speaker: Yilun Dong (University of Science and Technology of China)
    • MS17: 2.2
      • 257
        Advanced 4D Imaging of Shales at Micro- to Nano-scale: Investigating the time-lapse evolution Under Subsurface Thermal, Hydrological, Mechanical, and Chemical Conditions

        Mudstones and shales are the source of unconventional oil and gas reservoirs, as well as the primary control on the sealing efficiency for subsurface storage applications (e.g. energy, CO2, H2, waste). However, their fine-grained and heterogeneous nature makes their full characterisation highly challenging. Here we demonstrate the multi-scale and dynamic imaging approaches that can help meet these challenges and discuss limitations and future opportunities.
        Microstructures can be characterised at scales from sub- nm (<1 nm) to over 1 m, using multi-scale and multi-model imaging approaches [1-3], including X-ray tomography, Focused Ion Beam Scanning Electron Microscope and Transmission electron microscopy tomography. The majority of pores in mudstones/shales range from 0.2 nm to 3µm, and we have documented 4 major types with 3 distinct size distributions [4]. Based on the REV analysis, pore sizes, types and distribution can be upscaled via three stages from sub-nm to cm-scale[5]. The permeability is pressure dependant, ranging from 1.0×10-17 to 1.0×10-22m2 [3, 5]. CO2 adsorption is 3-7 times higher than CH4 and over 10 times higher than H2 [6]. Image based modelling has demonstrated that the non-Darcy effects (e.g., slip flow and Knudsen diffusion). Adsorption/desorption and surface diffusion takes major controls over time after injection [7].
        Dynamic imaging of mudstones/shale has provided the opportunity to characterise the thermo-hydro-mechanical-chemical (THMC) properties and the coupling mechanism in mudstones/shales to investigate the sealing ability under realistic reservoir conditions. These include high temperature (from less than 10 °C up to 1000 °C) [8], high pressure (e.g. confining pressure, indentation, torsion, deformation and fractures; up to 65 MPa) [9], fluids (e.g. diffusion, adsorption, flowing through, multi-phase flow) [10] and complex chemistry environment (brine and drilling fluids) [11]. Based on the dynamic behaviours observation and quantification, It can be concluded that mudstone/shales with horizontally thin-layered laminations, few fractures and less reactive minerals may act as the best caprocks.
        Whilst the above has led to an improved understanding of shale/mudstone microstructure under static and dynamic conditions, significant challenges still remain regarding representivity and up-scaling, experimental analysis at subsurface-realistic temperatures, pressure and chemistry, accurate estimations of the long-term behaviours and the proper monitoring techniques.
        1. Taylor, K. G., & Ma, L. 2021. GeoExpro, Vol. 18, No. 1 - 2021. 2. Ma, L., et al., 2017. International Journal of Coal Geology, 180, 100-112. 3. Ma, L. et al., 2016. . Marine and Petroleum Geology, 72, 193-205. 4. Ma, L., et., al 2019 . Scientific reports, 8(1), 1-14. 5. Ma, L., et al., 2019. Energy, 181, 1285-1297.6. Ma, L. et al. 2021. Energy & Environmental Science, 14(8), 4481-4498. 7. Guo, B., et al., 2018. Advances in Water Resources, 122, 70-84. 8. Wang, K., et al., 2021. International Journal of Coal Geology, 244, 103-116. 9. Figueroa Pilz, F., et al., 2017. Journal of Geophysical Research: Solid Earth, 122(4), 2553-25, 11. Godinho, J. R., et al., 2019). Minerals, 9(8), 480.

        Speaker: Lin Ma (University of Manchester)