Forced imbibition of a wetting liquid into a gas-filled tight rock is often expected to advance as a compact front because the liquid-to-gas viscosity ratio is favorable. This expectation can fail when narrow throats, pore-body/throat mismatch, capillary-valve pinning, and wall-associated wetting pathways reorganize the defending gas before it is displaced as a connected phase. We use a three-dimensional regularized color-gradient lattice Boltzmann model to examine these processes in a single, initially gas-saturated reconstructed tight-sandstone pore space. The simulations sample three capillary numbers, two or three Ohnesorge numbers depending on the capillary number, and two water-phase contact angles while keeping the water-to-gas viscosity ratio fixed at 26.11. The contact angle is measured through the aqueous phase, with θ = 20° representing strongly water-wet conditions and θ = 60° representing weakly water-wet conditions. For θ = 60°, terminal displacement efficiency changes little between the low and intermediate sampled capillary numbers and increases from approximately 0.660 to 0.736 at the highest sampled value. For θ = 20°, strong water-wetness is beneficial only after the bulk meniscus gains enough driving force to compete with precursor corner or wall flow; at a low capillary number, the same wetting affinity is associated with snap-off and premature gas isolation. The sampled Oh dependence is weaker than the Ca dependence and is consistent with conditional modulation of capillary-inertial damping and local interface relaxation; it is not interpreted as a new static entry criterion. Size-resolved and morphology-resolved statistics show contrasting terminal signatures: θ = 20° is associated with more large-pore gas and snap-off-consistent fragmentation, whereas θ = 60° is associated with more persistent small-pore gas and bypassing-consistent retention. Event-resolved phase-field sequences at one low-Ca condition directly show wall-first precursor advance, abrupt pore-body filling, gas-neck closure, persistent component splitting, and bypass-induced local entrapment. These events establish occurrence, not their frequency or dominance across parameter space. Because the study uses one pore-space realization and a sparse, non-factorial parameter matrix, the reported comparisons are restricted to the sampled conditions and do not define a continuous Ca–Oh–θ response surface or quantify structure-to-structure uncertainty.

Hydraulic fracturing in deep coalbed methane (CBM) reservoirs is commonly followed by coupled imbibition-displacement and flowback, yet the microscopic mechanisms of gas–water redistribution remain unclear because coal contains heterogeneous and multi-scale pore–fracture systems. This study employed an enhanced homogenized lattice Boltzmann method to construct a multiscale matrix–fracture model for deep coal, in which explicit cleat flow is coupled with homogenized matrix seepage. Gas–water evolution was simulated throughout the injection, shut-in, and flowback cycle to characterize exchange between fractures and coal matrix. Results show that wettability governs fluid invasion in the orthogonal cleat network and coal matrix. Hydrophilic conditions promote capillary imbibition and gas displacement but cause severe matrix water blocking during flowback. Strong hydrophobicity facilitates water removal from matrix; however, early gas breakthrough and bypassing disrupt water-phase continuity, leaving residual water trapped in dead-end cleats and thereby limiting the flowback recovery. During spontaneous imbibition, the matrix imbibition-displacement efficiency scales linearly with the square root of time, and two stages are identified from slope variations. Wettability effects weaken during late-stage water redistribution. Sensitivity analysis indicates that a weakly hydrophilic regime near neutral wettability provides a favorable compromise between imbibition sweep and flowback recovery. Prolonged shut-in yields diminishing marginal benefits because continued imbibition drives deeper matrix invasion and increases flowback resistance. For wettability modification during flowback, near-neutral wetting is recommended because it reduces capillary resistance while maintaining water-phase continuity. These findings clarify multiscale gas–water exchange mechanisms and provide a physical basis for optimizing shut-in and flowback strategies in deep CBM reservoirs.

CO2 injection for enhanced oil recovery (EOR) and geological carbon sequestration (GCS) in shale reservoirs offers a promising win-win strategy. Yet, the mechanisms by which CO2 flooding and huff-and-puff (HNP) mobilize multicomponent shale oil in inorganic nanopores remain unclear, particularly regarding their respective contributions to oil recovery and CO2 storage. Here, we employ molecular dynamics (MD) simulations to compare CO2 flooding and HNP in calcite dead-end nanopores saturated with a multicomponent hydrocarbon mixture. We analyze interfacial and transport behaviors, extract localized molecular clusters to quantify self-diffusion coefficients, and examine the roles of flood and flowback pressures. Results show that flooding promotes greater CO2 penetration into nanopores but suppresses counter-current hydrocarbon migration, limiting swelling-driven recovery. In contrast, HNP enables CO2 diffusion during the huff stage and reverse pressure gradients during the puff stage, which enhance oil swelling and achieve higher recovery at the expense of lower storage efficiency. Elevated CO2 density increases the diffusion coefficients of all components, underscoring the critical roles of pressure gradients and molecular transport. By extracting localized molecular clusters, we further obtained pressure-dependent diffusion bounds for CO2 and hydrocarbons and quantify injection and flowback pressure windows that jointly optimize oil recovery and the CO2 storage efficiency. On this basis, we propose a well-count-agnostic and field translatable hybrid method that achieves 60.44% oil recovery and 47.18% CO2 storage efficiency. These findings provide molecular-scale guidance for full-process CCUS design in unconventional reservoirs.

Abstract Surfactant flooding is a promising strategy for subsurface fluid recovery and environmental remediation, wherein interfacial tension (IFT) reduction and wettability alteration are the two principal mechanisms governing residual phase mobilization. Yet attributing their interplay and relative contributions remains challenging because the two processes are intrinsically coupled. To address this, we develop a pore‐scale lattice Boltzmann model that concurrently incorporates surfactant transport, IFT reduction, adsorption‐induced wettability reversal, and tunable solubility contrast, providing a versatile platform to explore displacement physics. We then investigate the individual and combined effects of IFT reduction and wettability alteration under varying capillary numbers, viscosity ratios, and pore‐scale disorder. We find that both mechanisms improve displacement efficiency by stabilizing the displacement front: IFT reduction raises the local capillary number to suppress selective invasion, while wettability reversal promotes cooperative pore filling. Their synergy leads to more compact invasion patterns and reduced residual saturation. Morphologically, IFT reduction results in dispersed oil ganglia, whereas wettability reversal favors network‐like trapping. We observe a capillary number‐dependent shift in the dominant mechanism, controlled by surfactant diffusivity and adsorption loss: IFT reduction dominates at low capillary numbers, whereas wettability alteration becomes more influential at high capillary numbers and then levels off. Wettability reversal also shows superior robustness under high viscosity contrast and structure disorder by mitigating bypassing flow. These findings offer new mechanistic insight into the coupled physics of surfactant flooding and provide guidance for the rational design of injection strategies and chemical formulations in enhanced oil recovery and subsurface remediation.

Understanding and manipulating water intrusion at the nanoscale have increasingly attracted much interest in nature and technology. A velocity-dependent dynamic contact angle (DCA), as a boundary condition, serves a vital function in an advancing water front. However, an understanding of DCA mechanisms remains limited. Here, we develop a universal theory for DCA jointly produced by the friction dissipation near the contact line region and the viscous energy losses at the bulk meniscus, molecular movements overcoming these two modes of dissipation are characterized as slip velocity and bulk velocity in hydrodynamics, respectively. We find that contact angle significantly changes with velocity during water intrusion into nano-capillaries with different wettability, spanning from superhydrophilic to superhydrophobic. Our results demonstrate that the velocity dependence of DCA can be readily strengthened or suppressed by tuning confinement or wettability of nano-capillaries, which can be further used to manipulate nanoscale water flow at a more precise level.

Weak water-drive offshore reservoirs with complex pore architecture and strong permeability heterogeneity present major challenges, including rapid depletion of formation energy, low waterflood efficiency, and significant lateral and vertical variability in crude oil properties, all of which contribute to limited recovery. To support more effective field development, alternative strategies and a deeper understanding of pore-scale flow behavior are urgently needed. In this work, CT imaging and digital image processing were used to construct a digital rock model representative of the target reservoir. A pore-scale flow model was then developed, and the Volume of Fluid (VOF) method was applied to simulate and optimize waterflooding schemes aimed at boosting oil recovery. Optimization focused on adjusting injection rates, varying the oil-water viscosity ratio, and implementing a water-alternating-gas (WAG) process. Results show that, for equal injection volumes, higher injection rates cause early water breakthrough through high-permeability pathways, yielding slower gains in recovery. Lowering the oil-water viscosity ratio improves mobility control, suppresses viscous fingering, enlarges sweep volume, and enhances recovery. When CH4 becomes fully miscible, it dissolves into the crude oil, lowering viscosity and eliminating interfacial tension, thereby providing greater displacement efficiency than partially miscible injection. Following a switch from water to gas injection, residual oil saturation decreases and becomes more uniformly distributed, indicating that the combined action of water and gas significantly improves both sweep efficiency and microscopic displacement.

Gas reservoirs exhibit diverse water production mechanisms and variable manifestations, posing complex challenges to efficient gas recovery. While numerous water control technologies have been developed (covering wellbore liquid unloading, shutoff interventions, and reservoir-scale water invasion prevention), existing research predominantly addresses isolated techniques or provides parallel overviews, lacking a systematic classification and integrated evaluation. This review introduces a unified classification framework for water control technologies in gas reservoirs, grounded in the two primary water invasion types: edge-water drive and bottom-water drive. By integrating reservoir-scale invasion patterns with near-wellbore production behaviors, four distinct categories are defined: (1) deliquification, (2) water invasion prevention in gas zones, (3) wellbore shutoff and control, and (4) life-cycle coordinated water management. The framework delineates the functional roles and operational features of each category, and critically examines their technical limitations and implementation challenges. Through this multidimensional and systematic approach, the review enhances understanding of gas reservoir water control and addresses gaps in prior literature. It underscores the complexity of multi-scale, multi-mechanism systems, emphasizing that no single technology is sufficient for dynamic water production scenarios. A transition toward intelligent, refined, and data-driven strategies is essential. Future development should prioritize real-time monitoring, intelligent completions, and full life-cycle water management to replace empirical practices with science-based decision-making. This review provides both theoretical insight and practical guidance, offering value for future research and field deployment.

Gas displacement by water invasion in low-permeability gas-bearing rocks is commonly evaluated using gas saturation; saturation alone cannot identify whether the remaining gas is connected or isolated. The pore-scale controls on displacement efficiency remain unclear in three-dimensional pore structures, especially under initial-water conditions in low-permeability reservoirs and subsurface natural gas and hydrogen storage. Here, micro-computed tomography (micro-CT) experiments on tight cores are combined with three-dimensional lattice Boltzmann simulations on reconstructed digital rocks to link saturation decline with pore-size occupancy and residual-gas topology. Micro-CT images show that water invasion progressively converts a coarse-pore gas backbone into fragmented residual gas through limited early mobilization, selective removal from small-to-medium pore-throat regions, and late-stage fragmentation after water percolation. Simulations reveal that preexisting water is the leading control on displacement efficiency: increasing initial water saturation from 0 to 0.7 reduces efficiency by approximately 25%–56% across the digital rocks by weakening gas connectivity and promoting preferential flow along water-connected pathways. Capillary number governs capillary-threshold crossing, whereas Ohnesorge number mainly modulates local interface depinning and relaxation. Wettability and pore structure determine whether water invasion produces effective sweep, snapoff, or bypassing. Pore-size occupancy shows that small-pore gas responds early, while gas retained in medium-to-large pores controls sustained removal and late trapping. Topology analysis distinguishes droplet, columnar, multi-pore, and clustered gas, identifying whether remaining gas is isolated, partially connected, or preserved as bypassed bodies. This work provides a pore-scale topological basis for assessing gas mobility loss and trapping behavior in water-invaded reservoirs and related subsurface energy systems.

Hydraulic conductance is crucial for determining fluid transport capacity in porous media during energy recovery and storage. Inspired by the quasi-steady-state assumption of seepage flows, we developed a numerical model coupling creeping flow, momentum transfer, and interfacial slip, enabling precise modeling of two-phase hydraulic conductance under various mechanisms. We generated tens of thousands of data samples and constructed three machine learning models for hydraulic conductance prediction. Validation results demonstrate that these models outperform traditional empirical models in accuracy. Viscous coupling enhances the wetting phase conductance more significantly than the non-wetting phase, with its effect controlled by fluid and geometric configuration. A dimensionless slip length exceeding 0.01 elevates wetting phase conductance but has limited impact on the non-wetting phase. Non-wetting phase slip boosts the hydraulic conductance of both phases at a Knudsen number exceeding 0.01. The wetting film profoundly reshapes the velocity profiles and amplifies viscous coupling effect, with wetting film flow contributing as much as 60 similar to 100% to the total wetting phase flow. Wetting film also strengthens the impact of slip on both phases' conductance. These findings highlight the intricate mechanisms controlling gas-liquid hydraulic conductance, especially the powerful control of wetting film, which is often underestimated in current pore-scale to core-scale analyses.

Pronounced wall-molecule interactions arise at nanoscale, profoundly varying the CH4-CO2 mixture thermodynamics as well as adsorption behavior, apparently different from that at bulk space where intermolecular interactions dominate. Inherently, wall-molecule interaction strength is crucial for controlling nanoconfined CH4-CO2 behavior, which however remains a knowledge gap as the most research focus on the mixture behavior within specific nanopore wall types to date. In this work, calibrating the attraction term resulting from the wall-molecule interactions, a modified EoS (Equation of State) for nanoconfined fluids is developed. Then, on the basis of fundamental equilibrium of chemical potential, a model characterizing the nanoconfined CH4-CO2 behavior is proposed by incorporating the modified EoS, multiple-component SLD (Simplified local density method) framework, and the adsorption mixture thickness, which can be efficiently solved by well-designed iteration procedures. Results indicate that: (a) CH4 fraction evidently reduces while approaching the nanopore wall, both small pore size and strong wall interactions contribute to its magnitude; (b) High CO2 adsorption selectivity over CH4 can be achieved under low pressure-temperature condition, especially with the bulk CO2 fraction of 0.5; (c) A large variation in CO2 selectivity, reaching up to 123.8%, can be induced by wall energy parameter, which is much greater than the pore size effect. The research is able to provide a clear understanding of molecule-wall interaction strength on CH4-CO2 behavior in nanoconfinement.

In oil and gas field development, accurate prediction of the near-wellbore stress distribution is important for ensuring wellbore integrity, particularly in tight shale reservoirs where the medium can be approximated as nearly homogeneous. The large stress gradient and multi-physics coupling of the near-wellbore stress field are key factors limiting high-precision prediction. Physics-informed neural networks (PINNs) allow integration of governing physical laws into network training processes. However, conventional PINNs cannot accurately capture local stress concentration features and lack the ability to generalize across different parameter settings. This study uses the near-wellbore stress concentration problem as an example to improve the prediction accuracy of PINNs through the incorporation of additional physical constraints and modifications to the network architecture. Furthermore, the method is extended to a physics-informed Deep Operator Network (PI-DeepONet) framework with enhanced generalization capability. The results show that, after introducing the proposed constraints, the stress field exhibits strict biaxial symmetry. The angle-adaptive residual module decreases the near-wellbore stress error from about 40–45% to less than 5%. Based on this method, the near-wellbore stress field under different in situ stress combinations can be predicted instantaneously without retraining.

Hydrogen has attracted much public attention amid global energy transitions and renewable energy utilization. Underground hydrogen storage (UHS), offering large-scale storage capacity at low cost, plays a key role in enhancing hydrogen production and utilization efficiency. This review systematically examines UHS from micro to macro perspectives. It begins with comparing the properties of hydrogen, CO2 and CH4, highlighting key differences between hydrogen and other gases stored in underground media. Low molecular weight of hydrogen increases leakage risks, presenting unique storage challenges. The review then outlines common storage structures and existing global UHS sites. Five primary storage mechanisms are identified, including structural trapping, residual trapping, dissolution trapping, mineral trapping and adsorption. Among these, structural trapping emerges as the dominant mechanism. Site screening methods and potential sites in China are also discussed. Using established evaluation criteria, the review examines multiscale influencing factors, such as solid-fluid interaction, geochemical reactions, mechanical analysis, multiphase flow, storage capacity, and injection and withdrawal schemes, across micro, mesoscopic, and macro levels. These factors interact dynamically. It was observed that geochemical reactions may alter rock geo-mechanical properties, potentially compromising storage integrity. Injection and withdrawal schemes must account for complex fluid-solid interactions. Despite advancements, unresolved challenges persist due to the complexity of underground systems. Accordingly, the review concludes with targeted recommendations to address research gaps, providing a theoretical foundation for optimizing underground systems.

Interfacial effects and intermolecular forces dominate in shale nanopores， and the classical theory of continuum medium mechanics has limitations or even failures， thus becoming a key issue limiting the development of shale oil and gas storage and flow theories， as well as mathematical simulations. In order to address this issue， the interaction between alkane molecules in shale oil and graphene walls in shale nanopores and the adsorption/retention characteristics of multi-component alkanes in shale oil were studied. By using molecular dynamics simulation， single-component and multi-component systems of shale oil were constructed to analyze the adsorption behavior and retention phenomena of different alkane molecules in the pores， as well as the effects of van der Waals forces， pore sizes， and other factors on the strength of the interactions and the characteristics of adsorbed layers. The simulation results show that the interaction strength between alkane molecules and graphene walls increases nonlinearly with the increase in the number of carbon atoms in alkane molecules due to the irregularity of molecular arrangement and nonlinear van der Waals forces. In the single-component system of shale oil， the number of adsorption layers of alkane molecules increases with the increase in the interaction strength between alkane molecules and graphene walls， but the adsorption strength gradually decreases with the increase in the number of adsorption layers； in the multi-component system of shale oil， there are obvious component differentiation and heavy hydrocarbon retention phenomenon.

Corner film flow governed by dimensionality and wettability profoundly impacts displacement patterns, yet its effect on relative permeability remains unclear. We use a multiple-relaxation-time color-gradient lattice Boltzmann model with geometric wetting boundaries to examine, under identical pore structures, how capillary number, wettability, and viscosity ratio modulate the influence of wetting films on relative permeabilities. We find that in 3D (three-dimensional) capillary bridges link corner films into secondary pathways that increase wetting connectivity and strengthen interfacial viscous coupling. Increasing capillary number raises the relative permeabilities of both phases in 2D (two-dimensional) and 3D. At high capillary number or high wetting saturation, bulk flow dominates and the difference in wetting permeability shrinks, whereas the non-wetting permeability remains higher in 3D than in 2D due to stronger viscous coupling. Decreasing the contact angle enhances viscous coupling and weakens non-wetting-solid interactions, thereby increasing the non-wetting relative permeability and widening the 3D over 2D gap. The wetting-phase response to wettability depends on the presence of films and saturation. Increasing the viscosity ratio markedly elevates the non-wetting relative permeability, with the larger interfacial area in 3D further amplifying viscous coupling. The wetting-phase relative permeability decreases with viscosity ratio, yet at very high viscosity ratios the initial phase configuration becomes influential. This study clarifies the dimensional mechanisms governing relative permeability, especially the controlling role of wetting films, which is critical for judging when multiphase-flow physics can be simplified.

Forced imbibition dynamics are critical for enhancing recovery rates in reservoirs, as efficient fluid displacement directly impacts resource extraction. This study employs the Zou-He velocity boundary condition and a modified convective boundary condition (mCBC) within the multi-component Shan-Chen lattice Boltzmann method (LBM), thus facilitating unobstructed flow of multi-component fluids at the outlet while maintaining constant outlet pressure. The model's validity was established through outflow tests of immiscible droplets in channels. Its accuracy was further confirmed by comparing results from forced imbibition dynamics tests in dual-wetted pores with theoretical predictions. A symmetrical porous medium was constructed using the stacking method, achieving dual wettability by fixing the contact angle in the upper region and varying it in the lower region. We performed 20 simulation sets under unfavorable viscosity ratios and varying capillary numbers, focusing on overall displacement efficiency before and after breakthrough, and employing energy balance equations to evaluate the dominant forces. Results reveal that capillary forces predominantly dictate forced imbibition dynamics in low capillary number scenarios. In strongly wetted regions, the invading fluid fully occupies pore spaces; however, in weakly wetted regions, displacement efficiency significantly declines as wettability approaches neutrality, even nearing zero. As capillary numbers increase, viscous forces become more prominent, controlling dynamics and leading to fingering and trapping of defending fluids. In weakly wetted areas, the increasing influence of viscous forces enhances fluid displacement, resulting in significant improvements in overall displacement efficiency compared to conditions dominated by capillary forces.

Gravity segregation is a critical phenomenon in thick condensate gas reservoirs, significantly influencing fluid composition and phase behavior. Reservoir-scale numerical simulation, serving as an indispensable technical approach in modern petroleum engineering, provides both quantitative data support and theoretical frameworks for development strategy optimization. However, the impact of gravity segregation on the distribution of initial fluid compositions is often overlooked in conventional numerical simulations due to data limitations or underestimated importance. This oversight leads to systematic deviations between simulated reservoir performance and actual field observations, ultimately compromising the efficient development of reservoirs. This study analyzed PVT data from reservoir fluid samples at different depths to determine the initial fluid composition distribution. Two models were developed: one incorporating gravity segregation and another neglecting it, to evaluate their performance during gas injection. Key findings include: (i) Gravity segregation alters the initial fluid composition, creating lighter components near the reservoir top and heavier ones at the bottom, resulting in distinct phase behaviors and production dynamics. (ii) The model accounting for gravity segregation aligns better with historical production data, while the model neglecting it underestimates oil production rates by about 9% and overestimates oil recovery by 2–5% during gas injection, due to inaccurate fluid composition assumptions. (iii) The model without gravity segregation also underestimates differences in oil recovery between injection–production strategies, such as top versus bottom injection. This study highlights the critical role of gravity segregation in reservoir simulation and provides valuable insights for optimizing the development of condensate gas reservoirs with complex fluid distributions. The findings reveal that accounting for gravity segregation in reservoir simulation models through proper initialization of fluid distribution leads to improved simulation accuracy, thereby enabling more precise development strategy design.

Longitudinal dispersion coefficient is a key parameter governing solute transport in porous media, with significant implications for various industrial processes. However, the impact of microfractures on the longitudinal dispersion coefficient remains insufficiently understood. In this study, pore-scale direct numerical simulations are performed to analyze solute transport in microfractured porous media during unstable miscible displacement. Spatiotemporal concentration profiles were fitted to the analytical solution of the convection–dispersion equation to quantify the longitudinal dispersion coefficient across different microfracture configurations. The results indicate that the longitudinal dispersion coefficient is highly sensitive to microfracture characteristics. Specifically, an increased projection length of microfractures in the flow direction and a reduced lateral projection length enhance longitudinal dispersion at the outlet. When Peclet number ≥1, the longitudinal dispersion coefficient follows a three-stage variation pattern along the flow direction, with microfracture connectivity and orientation dominating its scale sensitivity. Furthermore, both diffusion-dominated and mixed advective-diffusion regimes are observed. In diffusion-dominated regimes, significant channeling alters the applicability of traditional scaling laws, with the relationship between longitudinal dispersion coefficient and porosity holding only when the Peclet number is below 0.07. These results provide a comprehensive scale-up framework for CO2 miscible flooding in unconventional reservoirs and CO2 storage in saline aquifers, offering valuable insights for the numerical modeling of heterogeneous reservoir development.

The inverse problem of capillary imbibition involves determining the capillary geometry from imbibition kinematics. Solving this inverse problem is essential for nondestructive characterization of the internal geometry of nanoporous materials, as well as for the design of microfluidic and porous media channels. Previous research works have shown that this problem may have multiple solutions, often requiring time-varying meniscus position in both forward and reverse directions to ensure accuracy. This study introduces a machine learning approach to address the inverse problem. A training dataset is constructed using numerical simulation data, and an optimized neural network model is used to capture the nonlinear relationship between imbibition kinematics in two directions and the capillary radius. The approach is validated through two examples using numerical simulations and experimental data. In both cases, the capillary radius predicted by the neural network model closely matches the true capillary radius. Compared to previous iterative and analytical methods, this approach is computationally efficient, faster, less dependent on data quality and quantity, and exhibits strong robustness.