Dominant Controlling Factors and Displacement Performance in Heavy Oil Recovery Processes Authors: Han Cao, Bo Huang, Bo Jing, Xinsheng Xue Published: 2026 ABSTRACT During immiscible displacement in heavy oil reservoirs, the multiphase flow of oil–water or oil–steam systems is governed by the combined effects of viscous forces, capillary forces, and gravity. Identifying the dominant forces and clarifying their competitive relationships are essential for improving the oil recovery. On the basis of core-scale physical simulation experiments, a unified dimensionless framework for analyzing force competition was established, enabling dimensionless characterization and recovery prediction for both conventional heavy oil–water flooding and super heavy oil–steam flooding. By systematically analyzing residual oil variation under different injection rates, temperatures, permeabilities, steam qualities, and injection inclination angles, we quantitatively characterized the competitive interactions among viscous, capillary, and gravitational forces. The results indicate that conventional heavy oil–water flooding is predominantly controlled by capillary forces, and oil recovery exhibits a significant linear negative correlation with the modified capillary number. In contrast, super-heavy oil–steam flooding is mainly governed by gravity-driven flow and thermal viscosity reduction. The composite model developed based on the thermogravimetry number (N TG) demonstrates high predictive accuracy (R = 0.9692), providing a quantitative basis for parameter optimization in heavy oil reservoir development. Introduction The migration of oil, gas, and water in reservoir porous media is fundamentally governed by the combined action of multiple forces. Multiphase flow is controlled not only by pressure gradients and viscous forces but also by gravity and oil–water interfacial tension. Some of these forces act as driving mechanisms, whereas others serve as resistance. Therefore, immiscible displacement can be regarded as a dynamic balance among the forces acting on crude oil. Clarifying the dominant forces and their interactions during displacement is essential for improving oil recovery. Extensive studies have been conducted on the force behavior of droplets at the pore scale. Hinze first proposed the breakup criterion and the concept of a critical Weber number, indicating that droplet deformation and breakup are governed by dynamic pressure, viscous force, and surface tension; droplet breakup occurs when the Weber number exceeds a critical value. Madani et al. further identified three detachment modes of droplets under shear: (i) translation along a solid surface without significant deformation, (ii) deformation along the flow direction followed by movement, and (iii) deformation and movement followed by detachment from the surface. As droplet viscosity increases, the flow dynamics become more complex. Seevaratnam et al. observed droplet motion and possible separation at low viscosity, while deformation and partial detachment were reported at higher viscosity. The critical conditions for droplet movement and detachment were characterized using a combination of capillary number and Reynolds number. Lu et al. − investigated the deformation, migration behavior, and force balance of silicone oil droplets with different viscosities under shear flow and established a force model describing the balance between interfacial adhesion and hydrodynamic shear. Wang et al. studied ordinary heavy oil using microfluidic experiments, demonstrating that droplet morphology in emulsions is closely related to flow velocity and viscosity. At the macroscopic reservoir scale, residual oil saturation serves as a key parameter for evaluating displacement efficiency and predicting ultimate recovery. To quantify the relative magnitude of driving and resisting forces, dimensionless numbers, such as capillary number, Reynolds number, Bond number, and gravity number, are commonly introduced. Lenormand et al. , classified displacement patterns through numerical simulation and physical experiments and employed capillary number and viscosity ratio to represent the relative effects of capillary and viscous forces, identifying three regimes: viscous fingering, capillary fingering, and stable displacement. Subsequent modifications and extensions by Zheng, Zhang, and Chen improved the applicability and predictive capability of the Lenormand diagram. Consequently, the use of dimensionless parameters to characterize force competition and predict displacement patterns and recovery has become a fundamental theoretical approach in reservoir engineering. However, classical dimensionless theories were primarily developed for light oil systems, and their applicability to heavy oil reservoirs with extremely high oil–water viscosity ratios remains limited. Fingering refers to the phenomenon in which the displacement front penetrates the displaced phase in a finger-like or branched pattern. When a less viscous fluid displaces a more viscous one, the interface becomes unstable, and in heavy oil–water flooding, the high viscosity ratio significantly amplifies viscous fingering, thereby altering front stability and sweep efficiency. Abrams introduced a modified dimensionless number by incorporating the viscosity ratio into the definition of the capillary number, emphasizing its key role in determining sweep efficiency. Doorwar and Mohanty combined the capillary number with other dimensionless groups to propose a new parameter, N I, for predicting recovery in unstable immiscible displacement processes, such as heavy oil–water flooding. Jamaloei evaluated three types of capillary numbers for surfactant flooding in heavy oil reservoirs and reported that different flow conditions require different definitions, indicating that heavy oil systems exhibit more complex behavior than conventional reservoirs. However, existing capillary-number-based descriptions mainly focus on viscous-capillary competition under specific displacement conditions, while the influence of permeability, pore-scale flow restriction, and heavy oil boundary-layer effects on force competition has not been sufficiently considered. When a significant density contrast exists between the displacing and displaced phases, as in steam flooding, the competition between gravity and capillary forces may also strongly influence oil recovery. Rao et al. showed that the Lenormand diagram failed to accurately predict interfacial stability in gas-assisted gravity drainage (GAGD), largely because gravity effects were not properly accounted for in the physical models. In steam flooding, gravity segregation leads to phase separation between steam and condensate, making gravity a critical factor in the dynamic force balance of heavy oil reservoirs. Argüelles-Vivas et al. investigated steam-assisted gravity drainage (SAGD) through pore-scale experiments and clarified the gravity-driven displacement mechanisms and residual oil formation, demonstrating that the process is influenced by reservoir heterogeneity and wettability. Further experiments conducted in square capillaries revealed the variation of residual oil saturation under gravity-dominated two-phase and three-phase flow at different temperatures, and the interplay among viscous, capillary, and gravity forces was quantified under various trapping conditions. Most previous studies have focused on gravity effects in SAGD, while systematic investigations of the coupled action of viscous, capillary, and gravity forces in conventional steam flooding remain limited. In particular, the combined effects of thermal viscosity reduction, steam quality variation, gravity segregation, and viscous instability have not been quantitatively integrated into a unified recovery prediction framework for steam flooding. Over the past decades, experimental, theoretical, and numerical studies have demonstrated that the competition between viscous and capillary forces determines the morphology of the displacement front and the extent of sweep, leading to the establishment of a dimensionless parameter-based analytical framework. However, most existing studies have been developed for conventional reservoirs, and the coupled effects of multiple forces under complex displacement conditions, such as heavy oil–water flooding and steam flooding, have not been fully characterized. A single dimensionless parameter is often insufficient to represent the dominant mechanisms in the different displacement systems. Therefore, further investigation of the evolution of dominant forces under various displacement conditions is required, building upon the existing theoretical foundations and supported by experimental evidence. In this study, physical core-scale experiments combined with dimensionless analysis were employed to systematically examine the interactions among viscous, capillary, and gravity forces in ordinary heavy oil–water flooding and super heavy oil–steam flooding. The force competition in the two displacement processes was characterized using different dimensionless parameters. The dominant forces and their critical ranges under different operating conditions were identified, which clarified the transition from capillary-dominated water flooding to gravity- and thermal-viscosity-reduction-dominated steam flooding. A recovery prediction model incorporating multiforce coupling effects was established, and a composite thermogravity number was further developed to quantitatively describe the coupled effects of gravity drainage, viscosity reduction, and capillary action in steam flooding. These results provide a more robust theoretical basis and practical guidance for understanding displacement mechanisms and evaluating development performance in heavy oil reservoirs. Experimental Section Materials The light oil and heavy oil used in this study were obtained from the D oilfield. The density of the heavy oil was measured in accordance with the petrochemical industry standard GB/T 13377-2010. Simulated formation water was used for all solution preparations and consisted of 1 wt % KCl. All reagents were of analytical grade and were purchased from Sinopharm Chemical Reagent Co., Ltd. The relevant fluid properties are summarized in Tables and . 1Parameters of Fluids Involved in the Water-Flooding Experiments fluid viscosity/mPa·sinterface related parameterstemperature/°Cheavy oillight oilformation waterIFT/mN·m–1 contact angle/deg251267222.9862.43260168 0.46739.718 Parameters of Fluids Involved in the Water-Flooding Experiments 2Parameters of Fluids Involved in the Steam-Flooding Experimentsenvironmental temperature/°C80100110120140crude oilIFT/mN·m–1 35.629.728.226.723.6viscosity/mPa·s53562389992288107density/kg·cm–3 969.6955.7951.6945.6933.6steam (234 °C)quality/%100807040 viscosity/mPa·s0.01180.03650.04620.0754 density/kg·cm–3 15.05176.42257.11499.16 Parameters of Fluids Involved in the Steam-Flooding Experiments Experimental Methods Water Flooding The experimental system consisted of a thermostatic oven, an ISCO high-precision displacement pump, pressure sensors, a data acquisition system, a high-temperature and high-pressure core holder, a piston, a 6-way valve, and high-pressure tubing. Unconsolidated sandstone artificial cores were used for the experiments. The cores were approximately 30 cm long and 2.5 cm in diameter. The core properties are given in Table . 3Parameters of Sandstone Cores in Core-Flooding Experiments displacement schemecore propertiescrude oil typetemperature/°Cinjection rate/mL·min–1 permeability/mDporosity/%light oil250.03, 0.05, 0.1, 0.3, 14204–463230–33heavy oil254354–462330–32901–94830–33604345–466030–33 Parameters of Sandstone Cores in Core-Flooding Experiments The experimental procedure was conducted as follows:1.Core permeability and porosity were measured. The core was evacuated for 2 h using a vacuum pump and then saturated with brine.2.The apparatus was assembled according to the experimental schematic (Figure ). Oil was injected at a rate of 0.1 mL·min–1 to establish the initial oil saturation until no further water was produced. The core was then aged for 24 h.3.The oven temperature was set to the target value and maintained for 2 h after stabilization.4.The displacement rate was set according to the experimental design, and water flooding was continued until a high water cut was reached. Pressure data were recorded throughout the experiment.5.The apparatus was cleaned in preparation for subsequent runs. 1.Core permeability and porosity were measured. The core was evacuated for 2 h using a vacuum pump and then saturated with brine. Core permeability and porosity were measured. The core was evacuated for 2 h using a vacuum pump and then saturated with brine. 2.The apparatus was assembled according to the experimental schematic (Figure ). Oil was injected at a rate of 0.1 mL·min–1 to establish the initial oil saturation until no further water was produced. The core was then aged for 24 h. The apparatus was assembled according to the experimental schematic (Figure ). Oil was injected at a rate of 0.1 mL·min–1 to establish the initial oil saturation until no further water was produced. The core was then aged for 24 h. 3.The oven temperature was set to the target value and maintained for 2 h after stabilization. The oven temperature was set to the target value and maintained for 2 h after stabilization. 4.The displacement rate was set according to the experimental design, and water flooding was continued until a high water cut was reached. Pressure data were recorded throughout the experiment. The displacement rate was set according to the experimental design, and water flooding was continued until a high water cut was reached. Pressure data were recorded throughout the experiment. 5.The apparatus was cleaned in preparation for subsequent runs. The apparatus was cleaned in preparation for subsequent runs. Schematic diagram of water-flooding tests. Steam Flooding The experimental system included a steam generator in addition to the equipment described above (thermostatic oven, ISCO pump, pressure sensors, data acquisition system, high-temperature and high-pressure core holder, piston, six-way valve, and high-pressure tubing). Unconsolidated sandstone artificial cores were also used for steam flooding. The cores were approximately 30 cm in length and 2.5 cm in diameter, with permeability ranging from 2800 to 3400 mD and a porosity of approximately 31%. Detailed core properties are provided in Table . 4Parameters of Sandstone Cores in Core-Flooding Experimentsdisplacement schemecore propertiesconversion temperature/°Cinjection rate/mL·min–1 steam quality/%injection angle/degpermeability/mDporosity/%80, 100, 110, 120, 140310002897–335430–321103100, 80, 70, 60, 40902832–320430–321, 3, 5, 1080903133–334530–34602893–314830–32452901–317431–33302893–314830–33 Parameters of Sandstone Cores in Core-Flooding Experiments The procedure was conducted as follows:1.Permeability and porosity were measured, and the core was evacuated for 2 h before being saturated with a 1 wt % KCl solution.2.The apparatus was assembled according to the experimental layout (Figure ). The oven temperature was set to the designated conversion temperature and maintained for 2 h.3.Oil saturation was established by injecting oil at a constant rate until no further water was produced. The flow rate was then gradually increased, and an additional 1–2 PV was injected to ensure stable initial oil saturation. The core was aged for 24 h.4.The displacement rate was set according to the experimental plan. The steam temperature was fixed at 234 °C, and the back pressure was adjusted to obtain different steam qualities.5.The core was positioned at a specified inclination angle. Once stable steam was generated, steam was injected for a predetermined pore volume. The experiment was terminated when the water cut reached 99%. Temperature at the end of the test, cumulative oil and fluid production, and pressure were recorded. Oil production was measured by the gravimetric method, and total liquid production was determined using a graduated cylinder.6.The system was cleaned before the next experimental run. 1.Permeability and porosity were measured, and the core was evacuated for 2 h before being saturated with a 1 wt % KCl solution. Permeability and porosity were measured, and the core was evacuated for 2 h before being saturated with a 1 wt % KCl solution. 2.The apparatus was assembled according to the experimental layout (Figure ). The oven temperature was set to the designated conversion temperature and maintained for 2 h. The apparatus was assembled according to the experimental layout (Figure ). The oven temperature was set to the designated conversion temperature and maintained for 2 h. 3.Oil saturation was established by injecting oil at a constant rate until no further water was produced. The flow rate was then gradually increased, and an additional 1–2 PV was injected to ensure stable initial oil saturation. The core was aged for 24 h. Oil saturation was established by injecting oil at a constant rate until no further water was produced. The flow rate was then gradually increased, and an additional 1–2 PV was injected to ensure stable initial oil saturation. The core was aged for 24 h. 4.The displacement rate was set according to the experimental plan. The steam temperature was fixed at 234 °C, and the back pressure was adjusted to obtain different steam qualities. The displacement rate was set according to the experimental plan. The steam temperature was fixed at 234 °C, and the back pressure was adjusted to obtain different steam qualities. 5.The core was positioned at a specified inclination angle. Once stable steam was generated, steam was injected for a predetermined pore volume. The experiment was terminated when the water cut reached 99%. Temperature at the end of the test, cumulative oil and fluid production, and pressure were recorded. Oil production was measured by the gravimetric method, and total liquid production was determined using a graduated cylinder. The core was positioned at a specified inclination angle. Once stable steam was generated, steam was injected for a predetermined pore volume. The experiment was terminated when the water cut reached 99%. Temperature at the end of the test, cumulative oil and fluid production, and pressure were recorded. Oil production was measured by the gravimetric method, and total liquid production was determined using a graduated cylinder. 6.The system was cleaned before the next experimental run. The system was cleaned before the next experimental run. Schematic diagram of steam-flooding tests. Results and Discussion Controlling Mechanical Parameters in Ordinary Heavy Oil–Water Flooding Water flooding is one of the most widely used recovery methods in reservoir development. However, because of the high viscosity and poor mobility of ordinary heavy oil, conventional water-flooding theories established for light oil systems cannot be directly applied. In this section, core-scale physical simulation experiments were conducted to compare the displacement behaviors of the light oil and ordinary heavy oil. A modified dimensionless capillary number was introduced to characterize the competition between viscous and capillary forces, and a recovery prediction model for ordinary heavy oil–water flooding was established. Interaction between Driving and Resisting Forces The ultimate oil recovery E can be expressed as the product of displacement efficiency and sweep efficiency: 1 E=ED·EV where E D is the displacement efficiency (%), and E V is the sweep efficiency (%). The value of E D is governed by the residual oil saturation within the swept region and is affected by the pore structure, microscopic heterogeneity, wettability, and interfacial tension. In contrast, E V reflects the unswept volume and is primarily controlled by the mobility ratio. Whether residual oil trapped in pore spaces can be mobilized depends on the balance between the driving and resisting forces. Oil can only be displaced when the driving force exceeds the capillary resistance. This balance is commonly described by the capillary number, defined as the ratio of the viscous force to the capillary force. The viscous force is related to fluid viscosity and flow velocity, whereas capillary pressure depends on interfacial tension, contact angle, and pore radius. The capillary number, therefore, provides a quantitative measure of the relative dominance of these forces and the displacement capability of the injected fluid. The classical capillary number (eq ) is generally considered a microscopic form and includes displacement velocity, viscosity, interfacial tension, and contact angle: 2 Ca=vμσcos⁡θ where v is the displacement velocity (m·s–1), μ is the apparent viscosity of the displacing phase (mPa·s), σ is the oil–water interfacial tension (mN·m–1), and θ is the contact angle. The classical form was developed for light oil systems in which viscous fingering is negligible and the viscosity ratio is close to unity, leading to a relatively stable displacement front. This assumption does not hold for heavy oil reservoirs. In heavy oil systems, the oil–water viscosity ratio can be as high as 50,000, resulting in strong viscous fingering and early water breakthrough. In addition, significant differences in effective flow area between phases make the conventional capillary number insufficient to describe heavy oil displacement behavior. To address these limitations, a modified capillary number incorporating macroscopic parameters such as permeability, porosity, and geometric scale is required to better represent the flow characteristics of heavy oil systems. More than 30 forms of capillary number have been proposed in the literature. In this study, the modified capillary number proposed by Arab et al., derived from 178 core-flooding experiments and applicable to viscosity ratios up to 15,000, was adopted (eq ) to quantify the interaction between viscous and capillary forces in heavy oil–water flooding. 3 NCa=(μwvwσcos⁡θφ)0.26(μwμo)0.5(KLD)0.18 where μw and μo are the apparent viscosities of water and oil (mPa·s), respectively; v w is the water-phase displacement velocity (m·s–1); σ is the oil–water interfacial tension (mN·m–1); θ is the oil–water contact angle; φ is the porosity (%); K is the permeability (μm2); L is the core length (m); and D is the core diameter (m). During steam flooding, the spatial distribution of oil, steam, and water is governed by multiple, coupled physical processes. Low-viscosity steam or water tends to induce viscous fingering and alter flow patterns, gravity segregation promotes vertical migration, and capillary pressure in water-wet porous media drives lateral redistribution. The relative importance of these mechanisms depends on flow rate, density difference, mobility ratio, permeability, and capillary pressure characteristics. To quantitatively evaluate the competition among capillary, viscous, and gravity forces, dimensionless numbers were introduced to characterize their interactions under temperature gradient conditions and to reveal the variation of residual oil with key influencing factors. Effect of Flow Velocity on Residual Oil Saturation In porous media, the driving force acting on oil droplets is the pressure difference across the droplet, which, according to Darcy’s law, can be expressed as the product of apparent viscosity and flow velocity. Therefore, flow velocity plays a critical role in determining residual oil saturation. In general, an increase in injection rate enhances the viscous force. If recovery increases with velocity, viscous force is dominant; conversely, if lower velocities result in higher recovery, the incremental oil production can be attributed to capillary-driven imbibition. Previous studies have also indicated that the effect of injection rate depends strongly on oil viscosity. Figure presents breakthrough recovery, postbreakthrough recovery, and total recovery at different flow rates for light and heavy oils. As shown in Figure b, higher injection rates led to higher recovery for light oil, reaching 69.83% at 1 mL·min–1, indicating that viscous force dominates the displacement process. In contrast, Figure a shows that for heavy oil, total recovery decreased with increasing flow rate. The maximum recovery (58.96%) was obtained at 0.03 mL·min–1, suggesting that capillary force plays the dominant role. A similar trend was observed for breakthrough recovery: it increased with flow rate for light oil but decreased for heavy oil. Although higher injection rates generated greater pressure in heavy oil systems, enhanced viscous fingering and shortened breakthrough time resulted in earlier water production and reduced recovery. Under high-rate injection, the injected water preferentially flows through high-permeability channels and bypasses the surrounding oil, leading to an unstable displacement front and severe viscous fingering. At the pore scale, the large viscosity contrast between water and heavy oil weakens the ability of the displacing phase to uniformly mobilize trapped oil, while at the macroscopic scale, flow instability reduces volumetric sweep efficiency. Consequently, although the pressure gradient increases at higher flow rates, the displacement efficiency decreases because a significant portion of the heavy oil remains unswept. After water breakthrough, a preferential water channel formed, and further production was defined as postbreakthrough recovery. For light oil, postbreakthrough recovery remained nearly constant (∼31%) regardless of flow rate. In contrast, for heavy oil, postbreakthrough recovery increased markedly from 20.84 to 53.39% under low-rate injection, contributing significantly to total recovery. Oil recovery of (a) heavy oil and (b) thin oil at different flow rates. Figure illustrates the relationship between residual oil saturation and the capillary number. For heavy oil, residual oil saturation increased with an increasing capillary number, whereas the opposite trend was observed for light oil. For light oil, capillary force dominated at N Ca < 0.0008, while viscous force became dominant at higher NC a, leading to reduced residual oil. For heavy oil, capillary force-controlled displacement at N Ca < 0.0001; as N Ca increased beyond this range, viscous fingering intensified, and residual oil saturation increased significantly. These observations indicate that increasing the flow rate benefits light oil recovery by strengthening viscous displacement, whereas in heavy oil systems, excessive viscous force aggravates fingering and channeling. Improved recovery is achieved under low-rate conditions that enhance capillary action and suppress instability. Relationship curve between residual oil saturation and capillary number. Effect of Temperature on Residual Oil Saturation Temperature affects not only fluid viscosity and mobility but also interfacial tension and contact angle, thereby altering capillary resistance at the oil–water interface. Water flooding experiments were conducted at 25 and 60 °C to quantify these effects (Figure ). Although heavy oil viscosity was significantly reduced at 60 °C, total recovery decreased with increasing flow rate at both temperatures, confirming that capillary force remains dominant. At all flow rates, total recovery at 60 °C exceeded that at 25 °C. For example, at 60 °C, reducing the flow rate increased recovery from 44.78 to 71.36%. However, the sensitivity to flow rate was stronger at 25 °C, indicating that capillary effects are more pronounced at a lower temperature. This difference was mainly reflected in postbreakthrough recovery. At 25 °C, postbreakthrough recovery varied significantly with flow rate, reaching 53.39% at a low rate, whereas the variation at 60 °C was smaller. The breakthrough recovery exhibited the opposite trend. At higher temperatures, it was more sensitive to the flow rate. As the flow rate increased from 0.03 to 1 mL·min–1, the breakthrough recovery decreased from 30.16 to 20.25% at 60 °C, while it decreased from 5.57 to 2.21% at 25 °C. The data indicate that increasing temperature reduces viscosity contrast and pressure gradients, leading to a more stable displacement front. Oil recovery of heavy oil at different flow rates at 60 °C. Figure shows the relationship between residual oil saturation and capillary number at different temperatures. Residual oil saturation was more sensitive to capillary number at 60 °C, with a shorter plateau region. At 25 °C, residual saturation remained approximately 31% within 0.00008 < N Ca < 0.00012, indicating capillary-dominated flow. When N Ca > 0.0012, the difference between temperatures diminished, suggesting that fingering instability weakened the temperature effect. The reduction in contact angle (32–18°) and interfacial tension (62.4–39.7 mN·m–1) with increasing temperature further influenced capillary behavior and residual oil saturation. Relationship curve between residual oil saturation and capillary number at different temperatures. Effect of Permeability on Residual Oil Saturation When permeability is fixed, the influence of viscous force can be evaluated more directly by varying operating conditions. However, once permeability changes, both viscous force and capillary pressure are affected simultaneously. A decrease in permeability increases the pressure gradient at a given injection rate, while a smaller effective pore radius enhances capillary pressure. In the modified capillary number, this behavior is reflected as a lower N Ca, indicating relatively stronger capillary effects. To evaluate the permeability effect, two artificial cores with permeabilities of 1000 and 5000 mD were tested (Figure ). For both cases, total recovery decreased with increasing flow rate. At low injection rates, the total recovery of the low-permeability core was only approximately 3 percentage points lower than that of the high-permeability core, confirming that heavy oil–water flooding is primarily capillary-driven. At the same flow rate, breakthrough recovery was higher in the high-permeability core under relatively high-rate conditions. For example, at 0.5 mL·min–1, breakthrough recovery in the 5000 mD core reached 3.75%, which was 1.43% higher than that of the 1000 mD core. This indicates that lower capillary resistance in larger pores allows viscous force to contribute during the early stage of displacement, although its effect diminishes as flooding proceeds. Despite a significantly higher pressure gradient in the low-permeability core (with a pressure difference exceeding 900 kPa), breakthrough recovery was not necessarily improved. After the breakthrough, high postbreakthrough recovery was observed at low injection rates for both cores. The low-permeability core exhibited slightly better performance due to stronger capillary pressure, further confirming that reducing flow velocity to suppress viscous effects and enhance capillary driving is essential for heavy oil recovery. Oil recovery of heavy oil at different flow rates in a 1000 mD core. Figure shows that when N Ca < 0.00007, residual oil saturation remained within 30–33% and was weakly dependent on permeability. As N Ca increased, residual oil saturation rose more rapidly in the low-permeability core, indicating an earlier onset of instability. When N Ca > 0.00007, the viscous force became dominant and enhanced fingering in the low-permeability system. In contrast, the high-permeability core maintained stable, capillary-dominated displacement within N Ca < 0.00011. The data indicate that lower-permeability media are more prone to interfacial instability under identical injection conditions, leading to earlier viscous domination and intensified channeling. Relationship curve between residual oil saturation and capillary number at different permeabilities. Recovery Prediction Method The results demonstrate that the capillary number significantly influences oil recovery in ordinary heavy oil–water flooding (Figure ). Different recovery correlations were obtained for different permeability levels, indicating that boundary-layer effects become pronounced when the permeability is lower than 1000 mD, leading to distinct flow behaviors. , In low-permeability porous media, the pore throat radius decreases significantly, causing the adsorption boundary layer of heavy oil on the pore surface to occupy a larger proportion of the effective flow channels. As a result, the effective seepage area is reduced, and the nonlinear flow characteristics become more pronounced. Under these conditions, a larger pressure gradient is required to mobilize the heavy oil, and the displacement process becomes more sensitive to the viscous instability and microscopic flow heterogeneity. In contrast, high-permeability cores possess larger pore channels and weaker boundary-layer influence, resulting in relatively stable flow behavior and a different recovery–capillary number relationship. Overall, recovery decreases with an increasing capillary number, suggesting that a stronger viscous force reduces displacement efficiency. Therefore, the injection rate should be minimized in heavy oil–water flooding to suppress viscous fingering and enhance capillary-driven displacement. Fitting a relationship curve between capillary number and oil recovery of ordinary heavy oil–water flooding. A linear regression between calculated capillary number and recovery was performed, yielding eq with correlation coefficients greater than 0.94, indicating good predictive performance. It should be noted that this relationship is derived from laboratory experiments, and field applications require appropriate correction according to specific reservoir conditions. 4 E={−292885NCa+89.486,R2=0.9452,high permeability−461205NCa+85.873,R2=0.9861,low permeability In eq , E represents the ultimate oil recovery factor (%), and N Ca is the modified capillary number defined in eq . Controlling Mechanical Parameters in Super Heavy Oil–Steam Flooding During steam flooding, the spatial distribution of oil, steam, and water is governed by multiple, coupled physical processes. Low-viscosity steam or water tends to induce viscous fingering and alter flow patterns, gravity segregation promotes vertical migration, and capillary pressure in water-wet porous media drives lateral redistribution. The relative importance of these mechanisms depends on the flow rate, density difference, mobility ratio, permeability, and capillary pressure characteristics. To quantitatively evaluate the competition among capillary, viscous, and gravity forces, dimensionless numbers were introduced to characterize their interactions under temperature gradient conditions and to reveal the variation of residual oil with key influencing factors. Interaction between Driving and Resisting Forces in Steam Flooding Steam flooding of heavy oil represents a complex multiphase flow system. At the pore scale, steam, condensed water, crude oil, emulsions, and organic deposits may coexist. Field performance is influenced by temperature gradients, crude oil properties, and thermodynamic behavior. In this process, recovery is governed by competition among viscous, capillary, and gravity forces. Capillary-driven displacement is controlled by wettability, fluid properties, pore geometry, and solid–fluid interactions, whereas gravity-driven flow is primarily determined by the density contrast and interfacial effects. , Although gravity is often neglected in oil–water systems, it becomes a dominant factor in oil–steam and oil–steam–water systems due to significant density differences. Buoyancy causes injected steam to migrate preferentially toward the reservoir top. Sweep efficiency is further influenced by gravity effects, well configuration, and permeability heterogeneity. Nonuniform front propagation resulting from density contrast and reservoir heterogeneity reduces volumetric sweep and overall recovery. High-temperature emulsification behavior can further influence viscosity reduction and residual oil mobilization during extra-heavy oil recovery processes. Therefore, the coupled effects of viscous, capillary, and gravity forces must be considered simultaneously to realistically describe the steam-flooding process. The effective driving force in steam flooding originates from thermal energy, particularly the latent heat of wet saturated steam, which reduces oil viscosity and sustains steam chamber growth. When heat loss exceeds latent heat supply, steam chamber expansion is restricted. The chamber volume is also pressure-dependent: at a constant injection rate, a lower reservoir pressure leads to higher steam quality and specific volume, resulting in a larger steam zone. Thus, the driving force cannot be regulated solely by adjusting the injection rate; instead, steam rate, steam quality, and pressure must be optimized in combination. A larger steam chamber generally corresponds to improved sweep efficiency and higher thermal efficiency. As steam propagates from the injector to the producer, distinct zones are formed, including the steam zone, condensed hot water zone, heated oil zone, and original oil zone. Because of the low density of steam, gravity segregation leads to steam override after breakthrough, while condensed water advances along the lower section of the reservoir, continuously heating colder oil. The viscosity reduction enables oil production under gravity and steam/hot water displacement. Consequently, after steam breakthrough, the temperature and sustained driving capacity of the hot water zone become critical to the ultimate recovery. To quantify the relative magnitudes of capillary, gravity, and viscous forces, the capillary number N C, Bond number N B, and gravity number N G were employed to describe force competition during steam flooding. , Equations – define these parameters. The capillary number represents the ratio of viscous to capillary force; the Bond number characterizes the balance between gravity and capillary forces; and the gravity number combines the effects of viscous and gravity forces. − 5 NC=μovgσog 6 NB=Δρg(kϕ)σog 7 NG=Δρg(kϕ)μovg where μo is the oil viscosity (mPa·s), v g is the steam seepage velocity (m·s–1), σog is the oil–steam interfacial tension (mN·m–1), Δρ is the density difference between steam and oil (kg·m–3), g is the gravitational acceleration (m·s–2), k is the absolute permeability (mD), and φ is the porosity (%). Effect of Preheating Temperature on Oil Recovery In super-heavy oil reservoirs, cyclic steam stimulation is commonly applied to preheat the formation, reduce reservoir pressure, and improve steam utilization before transition to steam flooding. Determining a suitable conversion timing is therefore critical to overall economic performance. Previous studies suggest that two conditions should be satisfied before converting to steam flooding: (1) the interwell thermal communication temperature should reach or exceed the oil conversion temperature and (2) a sufficiently large steam chamber should be formed, with a lateral sweep radius of approximately one-third of the well spacing. Based on these considerations, this section focuses on identifying the minimum interwell thermal communication temperature. Before the steam breakthrough, displacement is dominated by steam drive, and viscous forces prevail, while gravity effects can be relatively neglected. To isolate the influence of preheating temperature, the core was placed horizontally so that only viscous and capillary forces were considered. The oven temperature was adjusted to simulate different interwell communication temperatures after cyclic steaming. Steam-flooding experiments were conducted at five temperature levels ranging from 80 to 140 °C. Figure shows that when the preheating temperature was 80 °C, which was below the anomalous temperature point, recovery remained low. Although the injected steam heated the oil, the initial formation temperature was insufficient to compensate for heat loss, and oil viscosity could not be reduced to the flowing threshold. When the temperature increased to 110 °C, the largest increment in recovery was observed, rising from 37.25 to 42.67% with a 10 °C increase. At temperatures above 110 °C, the incremental improvement diminished to approximately 1% per 10 °C. Steam condensation during transport reduces specific volume and heat injection rate, thereby weakening steam chamber propagation. While higher preheating temperature promotes viscosity reduction, excessive heating increases steam consumption and cost and may cause premature steam breakthrough. Considering technical and economic factors, 110 °C was selected as the optimal preheating temperature for subsequent steam-flooding experiments. Oil recovery at different conversion temperatures. Effect of Steam Quality on Residual Oil Saturation After a steam breakthrough in super-heavy oil reservoirs, the viscous force decreases significantly, and production enters an “erosion” stage. This behavior results from the coupled effects of viscous, capillary, and gravity forces, among which gravity segregation plays a key role. To emphasize gravity effects, vertical core orientation was adopted so that the directions of viscous, capillary, and gravity forces were aligned. Different steam qualities were tested to determine an appropriate operating range. The viscosity and density of wet saturated steam varied with steam quality and were calculated using eqs and . 8 μws=Xs·μs+(1−Xs)μw 9 ρws=Xs·ρs+(1−Xs)ρw where μws is the viscosity of wet saturated steam (mPa·s), μs is the viscosity of dry steam (mPa·s), μw is the viscosity of saturated water (mPa·s), and X s is the steam quality (%). The corresponding densities are denoted by ρws, ρs, and ρw (kg·m–3) for wet saturated steam, dry steam, and saturated water, respectively. The experimental results (Figure ) indicate that recovery generally decreased as steam quality decreased. When steam quality was reduced to 80%, recovery declined slightly from 52.25 to 51.45% (≈1%). However, further reduction to 40% caused recovery to drop sharply to 32.69%. This trend is attributed to gravity effects under vertical conditions. Lower steam quality increases wet steam density and reduces the density contrast with oil, thereby strengthening gravity influence and partially compensating for reduced heat transfer within a moderate range. When steam quality becomes excessively low, viscosity reduction is insufficient, interfacial instability intensifies, and pronounced fingering occurs, ultimately reducing sweep efficiency and oil recovery. Oil recovery at different steam drynesses. Effect of Steam Injection Inclination on Residual Oil Saturation During steam injection, the flow behavior is governed by the combined effects of gravity, capillary, and viscous forces. In this section, the gravity effect was represented by varying the injection inclination angle, while the viscous force was adjusted through the injection rate. Unlike conventional gases, increasing the steam injection rate raises the pressure and apparent driving force but may simultaneously reduce steam quality, thereby weakening effective displacement. Therefore, the heat injection rate must be considered concurrently. The heat injection rates of the steam zone and hot water zone are expressed in eqs and : 10 Qsi=1000qsXsLv 11 Qwi=1000qs(Hs−Hwr) where Q si is the steam heat injection rate (kJ/day), Q wi is the hot water heat injection rate (kJ/day), q s is the steam injection rate (m3/day), X s is the steam quality (fraction), L v is the latent heat of vaporization (kJ·kg–1), H s is the steam enthalpy at temperature T s (kJ·kg–1), and H wr is the water enthalpy at initial reservoir temperature T r (kJ·kg–1). Equation indicates that the heat injection rate depends on both steam rate and steam quality, while steam quality is controlled by temperature and pressure. Excessively low injection rates reduce the heat transfer efficiency, whereas overly high rates increase the pressure and decrease the steam quality, both of which weaken effective heating. These results imply the existence of an optimal injection rate that balances efficient heating with a stable displacement. Under these conditions, heat loss is minimized, viscous fingering and steam override are suppressed, and the steam chamber volume is maximized. Effect of Gravity Number Figure a shows that residual oil saturation exhibits a consistent trend with gravity number under different inclination angles and that a critical value exists. Taking an inclination angle of 30° as an example, when N G < 0.05, residual oil saturation increased with increasing gravity number; when N G > 0.05, residual oil decreased gradually. A larger inclination angle intensified the adverse gravity effect after steam breakthrough. These results indicate that at lower injection rates (corresponding to larger N G), gravity effects become more pronounced, but excessively low rates impair heat transfer and viscosity reduction. When the injection rate exceeds the optimal range, the viscous carrying capacity after breakthrough reduces residual oil. However, larger inclination angles promote gravity segregation and steam override, amplifying postbreakthrough resistance and diminishing sweep efficiency. Relationship curve between residual oil saturation with (a) capillary number and (b) gravity number at different steam injection inclination angles. Effect of Capillary Number As shown in Figure b, residual oil saturation first decreased and then increased with increasing capillary number. At an inclination angle of 90°, when (N C < 0.00003), residual oil decreased slightly, indicating that moderate viscous force contributed positively to oil mobilization. When N C increased further, residual oil saturation rose markedly. This behavior suggests that at low capillary number, viscous force assists displacement, whereas excessive viscous force destabilizes the oil–steam interface, induces steam channeling and preferential pathways, and ultimately reduces sweep efficiency. Although higher viscous force enhances heat transfer and viscosity reduction to some extent, the intensified instability outweighs the thermal benefit, leading to increased residual oil saturation. As shown in Figure , the capillary number, bond number, and gravity number all influence oil recovery during super-heavy oil–steam flooding. However, regression analysis using any single dimensionless number yields poor correlation with recovery (eq ), indicating that viscous, capillary, and gravity forces are strongly coupled and cannot be represented independently. 12 E={−200563NC+50.738,R2=0.798311.146ln⁡NB+187.66,R2=0.1417.593ln⁡NG+62.672,R2=0.708 Fitting relationship curve among (a) capillary number, (b) bond number, and (c) gravity number with oil recovery of super heavy oil–steam flooding. To establish a more representative predictive relationship, composite dimensionless parameters such as trapping number and gravity drainage number have been proposed to account for multiforce interactions. In this study, several combinations were evaluated through regression analysis. Among them, the low-temperature oxidation number proposed by Chen et al. (eq ) showed the strongest correlation with experimental recovery data. 13 NLTO=NBA1μRA2NCA3 where N LTO is the low-temperature oxidation number (dimensionless), μR is the viscosity ratio before and after low-temperature oxidation (dimensionless), and A 1, A 2, and A 3 are the correction coefficients (dimensionless). On the basis of retaining the original mathematical form of the low-temperature oxidation parameter, the experimentally determined bond number and capillary number were introduced as adjusting factors, and the parameter was modified through coupling correction so that the combined effects of gravity, thermal viscosity reduction, and capillary action could be simultaneously represented. The revised composite parameter was thus defined as the thermogravity number (N TG), intended to characterize the integrated controlling mechanisms governing steam flooding. As shown in Figure , regression fitting and sensitivity analysis determined the correction coefficients A 1, A 2, and A 3 to be 0.26, 0.17, and 0.0001, respectively. The resulting relationship between oil recovery and N TG is given by eq , with a correlation coefficient of 0.9692, indicating high predictive accuracy at the laboratory scale. It should be noted that practical field applications require further calibration according to specific reservoir conditions. 14 E=35.008ln⁡NTG+180.4,R2=0.9692 Fitting a relationship curve between capillary number and oil recovery of super heavy oil–steam flooding. The magnitudes of the correction coefficients reflect the relative importance of Bond number N B, viscosity ratio μR, and capillary number N C. The dominant contribution of N B confirms that steam flooding is primarily gravity-driven. The second-largest contribution of μR highlights the critical role of viscosity reduction. The minimal weight of N C (0.0001) indicates that capillary force plays a minor role in steam flooding. From an engineering perspective, gravity effects can be enhanced by moderately reducing steam quality, while viscosity reduction can be strengthened by appropriately increasing injection rate and conversion temperature to optimize steam-flooding performance. Conclusions By varying displacement conditions to obtain recovery under different dimensionless numbers and analyzing the interaction among viscous, capillary, and gravity forces, the following conclusions are drawn:(1).Ordinary heavy oil–water flooding is predominantly capillary-driven, while viscous force exerts an overall negative effect on recovery. When the capillary number is less than 10–4, displacement remains relatively stable and residual oil saturation is maintained at 30–33%. As the capillary number increases, viscous fingering intensifies, sweep efficiency declines, and residual oil rises significantly.(2).Injection rate is the key parameter controlling heavy oil–water flooding performance. Lower injection rates suppress viscous fingering and enhance capillary imbibition, significantly improving postbreakthrough recovery. Increasing the temperature enhances overall recovery but does not alter the capillary-dominated mechanism. Although a higher temperature reduces oil viscosity and modifies interfacial properties, capillary force remains the principal driving force under low capillary number conditions.(3).A significant linear negative correlation exists between recovery and the modified capillary number in ordinary heavy oil–water flooding. The linear prediction models established under different permeability conditions exhibit correlation coefficients greater than 0.94, indicating that the modified capillary number effectively characterizes the competition between viscous and capillary forces and provides a theoretical basis for the recovery prediction and parameter optimization.(4).Super heavy oil–steam flooding exhibits optimal operational windows. The optimal conversion temperature is approximately 110 °C. A moderate reduction in steam quality (around 80%) enhances gravity contribution while maintaining effective heat transfer. Steam injection rate must be balanced to achieve sufficient viscosity reduction without inducing severe flow instability; excessively high or low rates are detrimental to recovery.(5).Super heavy oil–steam flooding is primarily governed by gravity drainage and viscosity reduction, whereas capillary force plays a minor role. When N G < 0.05, gravity acts as a favorable driving force; beyond this threshold, gravity segregation and steam override become pronounced. Residual oil shows a “decrease–increase” trend with capillary number; excessive viscous force promotes steam channeling and interfacial instability, thereby reducing sweep efficiency.(6).A single dimensionless number is insufficient to predict steam-flooding recovery accurately. A composite model based on the thermogravity number (N TG) achieved high fitting accuracy (R 2 = 0.9692). The dominant contribution of the bond number (N B), followed by the viscosity ratio (μR), and the minimal weight of capillary number (N C) confirm that steam-flooding recovery is controlled mainly by coupled gravity drainage and thermal viscosity reduction. The proposed model effectively captures the integrated effect of multiple forces. Future studies may further conduct multiphysics simulations and pore-scale investigations, and verify the applicability of the NTG model under field-scale reservoir conditions. This will provide more reliable guidance for the prediction and optimization of steam-flooding performance. (1).Ordinary heavy oil–water flooding is predominantly capillary-driven, while viscous force exerts an overall negative effect on recovery. When the capillary number is less than 10–4, displacement remains relatively stable and residual oil saturation is maintained at 30–33%. As the capillary number increases, viscous fingering intensifies, sweep efficiency declines, and residual oil rises significantly. Ordinary heavy oil–water flooding is predominantly capillary-driven, while viscous force exerts an overall negative effect on recovery. When the capillary number is less than 10–4, displacement remains relatively stable and residual oil saturation is maintained at 30–33%. As the capillary number increases, viscous fingering intensifies, sweep efficiency declines, and residual oil rises significantly. (2).Injection rate is the key parameter controlling heavy oil–water flooding performance. Lower injection rates suppress viscous fingering and enhance capillary imbibition, significantly improving postbreakthrough recovery. Increasing the temperature enhances overall recovery but does not alter the capillary-dominated mechanism. Although a higher temperature reduces oil viscosity and modifies interfacial properties, capillary force remains the principal driving force under low capillary number conditions. Injection rate is the key parameter controlling heavy oil–water flooding performance. Lower injection rates suppress viscous fingering and enhance capillary imbibition, significantly improving postbreakthrough recovery. Increasing the temperature enhances overall recovery but does not alter the capillary-dominated mechanism. Although a higher temperature reduces oil viscosity and modifies interfacial properties, capillary force remains the principal driving force under low capillary number conditions. (3).A significant linear negative correlation exists between recovery and the modified capillary number in ordinary heavy oil–water flooding. The linear prediction models established under different permeability conditions exhibit correlation coefficients greater than 0.94, indicating that the modified capillary number effectively characterizes the competition between viscous and capillary forces and provides a theoretical basis for the recovery prediction and parameter optimization. A significant linear negative correlation exists between recovery and the modified capillary number in ordinary heavy oil–water flooding. The linear prediction models established under different permeability conditions exhibit correlation coefficients greater than 0.94, indicating that the modified capillary number effectively characterizes the competition between viscous and capillary forces and provides a theoretical basis for the recovery prediction and parameter optimization. (4).Super heavy oil–steam flooding exhibits optimal operational windows. The optimal conversion temperature is approximately 110 °C. A moderate reduction in steam quality (around 80%) enhances gravity contribution while maintaining effective heat transfer. Steam injection rate must be balanced to achieve sufficient viscosity reduction without inducing severe flow instability; excessively high or low rates are detrimental to recovery. Super heavy oil–steam flooding exhibits optimal operational windows. The optimal conversion temperature is approximately 110 °C. A moderate reduction in steam quality (around 80%) enhances gravity contribution while maintaining effective heat transfer. Steam injection rate must be balanced to achieve sufficient viscosity reduction without inducing severe flow instability; excessively high or low rates are detrimental to recovery. (5).Super heavy oil–steam flooding is primarily governed by gravity drainage and viscosity reduction, whereas capillary force plays a minor role. When N G < 0.05, gravity acts as a favorable driving force; beyond this threshold, gravity segregation and steam override become pronounced. Residual oil shows a “decrease–increase” trend with capillary number; excessive viscous force promotes steam channeling and interfacial instability, thereby reducing sweep efficiency. Super heavy oil–steam flooding is primarily governed by gravity drainage and viscosity reduction, whereas capillary force plays a minor role. When N G < 0.05, gravity acts as a favorable driving force; beyond this threshold, gravity segregation and steam override become pronounced. Residual oil shows a “decrease–increase” trend with capillary number; excessive viscous force promotes steam channeling and interfacial instability, thereby reducing sweep efficiency. (6).A single dimensionless number is insufficient to predict steam-flooding recovery accurately. A composite model based on the thermogravity number (N TG) achieved high fitting accuracy (R 2 = 0.9692). The dominant contribution of the bond number (N B), followed by the viscosity ratio (μR), and the minimal weight of capillary number (N C) confirm that steam-flooding recovery is controlled mainly by coupled gravity drainage and thermal viscosity reduction. The proposed model effectively captures the integrated effect of multiple forces. Future studies may further conduct multiphysics simulations and pore-scale investigations, and verify the applicability of the NTG model under field-scale reservoir conditions. This will provide more reliable guidance for the prediction and optimization of steam-flooding performance. A single dimensionless number is insufficient to predict steam-flooding recovery accurately. A composite model based on the thermogravity number (N TG) achieved high fitting accuracy (R 2 = 0.9692). The dominant contribution of the bond number (N B), followed by the viscosity ratio (μR), and the minimal weight of capillary number (N C) confirm that steam-flooding recovery is controlled mainly by coupled gravity drainage and thermal viscosity reduction. The proposed model effectively captures the integrated effect of multiple forces. Future studies may further conduct multiphysics simulations and pore-scale investigations, and verify the applicability of the NTG model under field-scale reservoir conditions. This will provide more reliable guidance for the prediction and optimization of steam-flooding performance.