Kang Duan

Dehydration–rehydration reactions of hydrous minerals modify rock strength, fluid transport, and deformation in reactive geological environments. Although dehydration‐induced weakening is widely studied, the mechanical consequences of subsequent rehydration remain poorly constrained. We combine triaxial compression tests with microstructural and pore‐network analyses of original gypsum, fully dehydrated anhydrite, and rehydrated gypsum. Dehydration markedly reduces strength and suppresses pronounced serrated stress drops, whereas rehydration produces strength overcompensation and a more localized strain‐softening response. Rehydration does not simply restore the original fabric but produces a reconstructed microstructure containing newly formed gypsum crystals, locally infilled voids, and a reconfigured accessible pore‐throat network. Reaction‐induced stresses may contribute to this restructuring but were not measured directly. These results show that gypsum rehydration can create a mechanically distinct, overcompensated state rather than simply reverse dehydration.

Rainfall-induced groundwater fluctuations may alter pore pressure and drainage discharge around metro tunnels, but the hydraulic pathway from recharge to tunnel-scale response remains difficult to interpret because recharge filtering, saturated-zone propagation, bottom-boundary effects, and tunnel drainage dynamics act concurrently. This study develops a staged closed-form frequency-domain analytical model to separate and quantify these processes for a drained metro tunnel. Field groundwater-head observations and model-derived effective recharge are used to support the staged interpretation of recharge transmission and depth-dependent groundwater-head modification. A fitted shallow groundwater-head perturbation provides the upper-boundary forcing, which is propagated vertically through a fully saturated formation under constant-head and no-flow bottom boundary conditions to obtain the ambient head response at tunnel elevation. Rather than prescribing this tunnel-scale fluctuation directly, the model explicitly resolves its upstream saturated-zone propagation before coupling it to a local radial seepage solution around an equivalent circular tunnel with a finite Robin drainage boundary. Two-dimensional numerical seepage simulations provide a controlled benchmark for evaluating the analytical solution. The analytical responses closely reproduce the numerical benchmark in waveform, amplitude, and timing under both bottom boundary conditions. Sensitivity analysis shows that hydraulic diffusivity controls attenuation and signed phase difference during vertical propagation, while bottom-boundary effects are most evident in the intermediate-diffusivity range. The dimensionless drainage number governs the local pressure–discharge response regime: weak drainage preserves pore-pressure fluctuations while suppressing discharge variation, whereas strong drainage suppresses tunnel-boundary head fluctuations and produces a stronger discharge response. The model provides a compact analytical basis for distinguishing upstream aquifer filtering from local tunnel drainage effects in rainfall-induced hydraulic response around drained metro tunnels.

Equivalent drainage boundaries are computationally convenient, but they combine distinct internal processes into a single effective condition and cannot identify which process controls the hydraulic response: primary-lining recharge, circumferential geotextile conveyance, discrete drainage-point inflow, collector-pipe conveyance, outlet control, or inward leakage. This study develops a unified semi-analytical model for steady-state seepage transfer under a prescribed hydraulic head at the outer surface of the primary lining. Closed-form solutions are obtained for representative circumferential cells with ideal or finite drainage-point inlet capacity, and a transfer-matrix formulation is established for full-circumference systems with spatially non-uniform inlet capacities. The circumferential formulation is coupled with a one-dimensional collector-pipe model to resolve longitudinal head variation and outlet control. Numerical benchmarks agree closely, with a maximum absolute relative error of 3.2%, while independent physical-model data support the predicted pressure–discharge behavior and deterioration-induced hydraulic responses. Geotextile conveyance reduces circumferential head non-uniformity but provides limited system-wide pressure relief when drainage-point inlet capacity is severely degraded. Under representative severe-degradation conditions, increasing geotextile in-plane hydraulic conductivity by a factor of 50 lowers the mean geotextile head by only about 0.2%, whereas restoring drainage-point inlet capacities to their reference levels reduces the mean head by 42.5–59.4%. Increasing inlet capacity enhances collector-pipe inflow but may induce upstream backwater when longitudinal pipe conveyance is limited. A higher outlet head suppresses pipe discharge, whereas waterproof-board hydraulic conductivity primarily controls inward leakage. By resolving the internal controls hidden by equivalent boundaries, the model enables diagnosis of hydraulic bottlenecks and targeted maintenance assessment.

Periodic recharge can generate groundwater-head fluctuations that propagate through saturated strata and are further modified by localized drained underground openings. Quantifying this coupled far-field propagation and near-opening transformation remains challenging. This study develops analytical transfer functions to describe recharge-driven groundwater-head propagation and the associated local discharge response near drained underground openings under alternative bottom boundary conditions. A two-stage frequency-domain framework is established. In the first stage, an effective upper-boundary head signal is transmitted to opening depth through one-dimensional vertical groundwater diffusion. In the second stage, the ambient head at depth is mapped to local head, pore-pressure, and discharge responses using a closed-form radial solution with an equivalent Robin drainage boundary. The analytical solutions are benchmarked against two-dimensional seepage simulations over a range of hydraulic conductivity and specific storage. The results show that bottom boundary conditions mainly shift the overall response gain, whereas hydraulic diffusivity governs attenuation and phase-difference trends. After diffusivity-based normalization, response patterns collapse over broad parameter ranges, and phase-difference contrasts between bottom boundary conditions diminish in the high-diffusivity regime. The drainage conductance of the opening further controls the conversion of groundwater-head fluctuations into local discharge response. The proposed framework provides a computationally efficient hydrological tool for interpreting recharge-signal propagation, boundary effects, and drainage response in saturated groundwater systems.

Permeability in gypsum is jointly modified by chemical alteration during static dissolution and mechanical confinement by effective stress, yet their coupled controls remain poorly constrained. This study measures the steady state permeability of gypsum rock after long term static dissolution under multiple effective stress levels and interprets the response using NMR, MIP, SEM, and aqueous ion chemistry. The results show that permeability increases significantly with static dissolution duration while porosity changes remain minor, indicating that improved connectivity and preferential modification of bottleneck throats dominate rather than bulk pore volume growth. Stress dependent results further show that permeability sensitivity to effective stress weakens as dissolution proceeds, consistent with the progressive dominance of connected flow channels and a reduced contribution from weakly connected, easily compacted pathways. A time dependent empirical formulation with a baseline term is established to relate permeability to static dissolution duration and effective stress. Multiscale evidence indicates that grain boundary and micro throat dissolution promotes microcrack initiation and linkage, whereas precipitation, recrystallization, and fines retention locally constrict or clog throats. Under higher effective stress, preferential closure of fine conduits delays and limits permeability gains. This parametric framework supports forecasting permeability trends relevant to fluid transport and long-term performance in soluble rock settings.

Stable fluid displacement in porous media under confining stress underpins processes from gas extraction to geological storage. While the onset of hydraulic fracturing is well characterized, the confinement-controlled transition to stable infiltration—governed by evolving permeability and matrix stiffness—remains poorly resolved. Laboratory studies rarely combine precise stress control with pore-scale visualization of invasion patterns, and many numerical approaches neglect boundary compliance (fixed vs movable confinement) and grain-scale force chains that are critical for predicting regime transitions. Here we use a fully coupled hydromechanical pore-network–discrete element method model to capture two-way feedbacks among pore-pressure buildup, permeability evolution, and granular mechanics under fixed and movable confinement boundaries. Increasing confining stress suppresses grain deformation and shifts displacement from fracturing- to infiltration-dominated regimes. Decreasing the viscosity ratio weakens viscous forcing and further favors infiltration, whereas boundary mobility and stress anisotropy promote fracture growth and directional channeling. We introduce a composite, morphology-based classification factor with an empirical transition band (0.19–0.33) that, together with dimensionless velocity and force ratios, delineates nonlinear transition boundaries across variations in fluid viscosity, inflow rate, matrix stiffness, and permeability. Energy analysis shows that confinement and boundary mobility jointly control displacement efficiency and reservoir integrity and that fixed-boundary Hele–Shaw idealizations can underestimate displacement efficiency and fracture propensity under low confinement. The framework links pore-scale mechanisms to field-relevant efficiency–damage tradeoffs, providing guidance for injection strategies that improve displacement while minimizing reservoir damage.