To address the insufficient consideration of osmotic pressure in existing studies on water huff-and-puff in tight oil reservoirs, this paper introduces both osmotic and capillary pressures into two-phase flow theory and develops a full-cycle model for low-salinity water injection. The effects of salinity, injection rate, and soaking time on pressure, saturation, and salt concentration are analyzed. Results show that lower salinity enhances osmotic drive and improves recovery. Specifically, 1000 mg/L yields the highest recovery (10.94%) and an osmotic gradient of ~3.9 MPa. Capillary pressure acts synergistically with osmotic pressure during the soaking stage, promoting spontaneous imbibition and fluid redistribution. An optimal injection rate exists: excessive rates reduce swept volume, while very low rates prolong the cycle. Quantitatively, 80 m³/d is preferred (recovery 10.94%) over 40 m³/d (11.13%) as it halves the injection time with comparable recovery. There is also an optimal soaking time beyond which incremental gain diminishes. Recovery rises from 9.32% at 20 d to 10.94% at 40 d, with only 0.28% and 0.20% gains afterward; thus, 40 d is optimal. The model captures osmotic-driven behavior and reveals the secondary displacement effect during soaking, supporting parameter optimization.

In the development of carbonate reservoirs via water flooding huff-n-puff, the osmotic pressure effect is frequently overlooked, and existing models inadequately quantify the matrix imbibition and oil expulsion driven by salinity gradients. To address this issue, this study establishes a coupled oil–water two-phase huff-n-puff flow model for carbonate reservoirs that incorporates the interplay between salt concentration and osmotic pressure, which, for the first time, fully couples the van ’t Hoff osmotic pressure equation with solute transport equations for fractured-vuggy carbonate huff-n-puff, filling the gap that prior tight/shale reservoir low-salinity flow models fail to adapt to cyclic injection-soaking production regimes of carbonates. Based on the IMPES (implicit pressure–explicit saturation) numerical simulation method, an equivalent single-nucleus model is adopted to characterize the fractured-vuggy reservoir architecture. The model integrates the osmotic pressure formula, solute transport equation, and two-phase seepage governing equations, enabling a systematic analysis of the mechanisms by which osmotic pressure affects the multi-stage seepage process and the influence of key parameters on development performance. Quantitative simulation reveals three core laws controlled by salinity-induced osmosis: first, osmotic pressure drives water molecules to spontaneously migrate from the high-permeability fracture inner core toward the tight matrix pores, thereby modifying the water saturation distribution, expanding the water sweep region, and smoothing the saturation gradient between the inner and outer cores, which effectively mitigates water channeling in fractured reservoirs. Under the base case (injected water salinity = 1000 mg/L, inner-core permeability = 1000 mD, shut-in time = 80 d), the oil recovery factor with osmotic pressure considered reaches 13.46%, representing a 3.50% increment over the case without osmotic pressure. The recovery factor decreases monotonically with increasing injected water salinity, while it increases with longer shut-in time and higher inner-core permeability, both exhibiting pronounced diminishing marginal returns; the optimal shut-in time is approximately 80 d under the simulated conditions. This work delivers a fully coupled numerical tool and quantitative evaluation standard for osmotic imbibition mechanisms in fractured-vuggy carbonates. The quantified recovery increment and optimal soaking window established herein can directly guide field parameter optimization of injection water salinity, shut-in cycle and fracture reconstruction scale, balancing oil increment revenue and water treatment/well shutdown operation costs for on-site low-salinity huff-n-puff design.

Water retention after imbibition and the recovery response to low-salinity injection suggest that capillarity alone does not describe all fluid exchange in tight formations. How osmotic forces influence successive injection, soaking, and production stages in a fractured reservoir therefore requires further examination. We constructed a simulator that combines aqueous salt transport with two-phase flow and represents hydraulic fractures by assigning equivalent properties to the intersected grid cells. A comparison with an independent simulator was used to assess the computed saturation response. Sensitivity cases varied in brine salinity, shut-in duration, injection rate, branch position, and network geometry; pressure, aqueous saturation, salt distribution, and produced oil were examined together. With 4800 m³ injected at 80 m³/d and a subsequent 40 d soak, the 1000 mg/L case produced 13,741.24 m³ of oil and recovered 17.28% of the initial oil volume. The corresponding values at 50,000 mg/L were 9507.86 m³ and 11.95%. Thus, reducing salinity added 4233.38 m³ of oil and 5.33 percentage points of recovery, equivalent to a 44.60% increase relative to the reported high-salinity recovery factor. Extending shut-in beyond 40 d provided progressively smaller benefits. A faster injection schedule reduced the operating time at a fixed water volume but also favored early water return; 80 m³/d provided a useful compromise within the tested range. Placing branches close to the well raised recovery to 22.35%, versus 17.28% for the unbranched fracture. This improvement was 5.07 percentage points, or 29.34% relative to the baseline. In the network cases, matrix blocks enclosed by fractures supplied most of the recovered oil, while more isolated rock retained substantial oil. Taken together, the simulations link the benefit of dilute injection water to fracture access and the time available for matrix exchange. For the assumptions and cases examined, a 1000 mg/L brine, an 80 m³/d injection schedule, a 40 d soak, and branches close to the well provide a favorable operating combination.