Effect of microscopic pore–throat structure heterogeneity on waterflooding seepage characteristics of tight sandstone reservoirs

3.1 XRD

The mineralogical composition and content analysis of whole rocks and clay minerals was carried out using the Bruker D8 Advance X-ray diffractometer. Before the experiment, approximately 3–5 g of sample was ground to achieve a uniform powder with a particle size of 40 μm. The dried powder sample was then placed into a glass tube, mixed with distilled water, and thoroughly dispersed using ultrasound. The resulting suspension was poured onto a microscope slide for testing. Based on the parameters such as position, intensity, and width of diffraction peaks, the crystal structure, size, mineral composition, and relative mass content of the sample were determined by fitting with the characteristic spectra of standard minerals.

3.2 NMR and HPMI

The full-scale distribution of PTS was determined by the combination of NMR and HPMI experiments. The NMR tests were performed using the MARAN-2 core analyzer and the PC-102 centrifuge, the samples were saturated with a CaCl₂ solution of a total salinity of 50 g/L. Centrifugation at a force of 300 psi was used to obtain movable fluid saturation (MFS). After completion of NMR testing, the samples were dried and subjected to HPMI testing using the AutoPore IV 9520 mercury porosimeter. Based on the Washburn equation, the PTS characteristics were determined by converting the injection pressure of the HPMI into pore–throat radii.

3.3 T ₂ spectral transformation and fractal theory

Previous studies have suggested that the uncentrifuged T ₂ spectrum characterizes the connected pore space of saturated water, and the transverse relaxation time T ₂ is related to the size and shape of the pore space [11,12], following the relationship:

where n is the power index and m is a constant. Therefore, by determining the parameters m and n, the NMR T ₂ spectrum of saturated water can be transformed into the full PSD. Specific formula derivations and operational steps have been described in the literature [48,49].

This study employs the widely applied Brooks–Corey model to obtain the fractal dimension of the full PSD. The model assumes that the pore space of the rock consists of a series of capillary bundles [50,51], and the relationship between the cumulative volume percentage V(c) of pore–throat radii less than r and r can be expressed as [52]

where D is the fractal dimension, typically ranging between 2 and 3, with higher values indicating higher PTS heterogeneity, and C is a constant. As seen from equation (2), V(c) is positively linearly correlated with r in a double logarithmic coordinate system. If the full PSD of the core exhibits multi-segment linear relationships, it indicates that the sample has a multi-fractal characteristic. The total fractal dimension (D _t ) represents the degree of heterogeneity across the full PSD of the sample and serves as a porosity-weighted average of the fractal dimensions corresponding to different pore–throat types. It can be obtained using the following formula [53]:

3.4 Microscopic high-temperature and high-pressure waterflooding experiment

Thirteen flake core models were subjected to oil cleaning, drying, slicing, and polishing and then bonded between two glass plates. The length, width, and thickness of the sandstone model were 2.5, 2.5, and 0.065 cm, respectively (Figure 2a). The glass plates were heat-resistant (up to 150°C), pressure-resistant (up to 20.0 MPa), and explosion-proof tempered glass. The experimental temperature was controlled at approximately 65°C, with a maximum fluid displacement pressure of 16.0 MPa. The experimental water used was the same as for the NMR tests and was dyed with a small amount of methylene blue to make the formation water appear blue. A mixture of kerosene and vacuum pump oil in a certain ratio was used to prepare the simulated crude oil, with a viscosity of 1.45 mPa s at 65°C, and was dyed with a small amount of oil-soluble red to make the crude oil appear red, facilitating the observation of oil–water displacement characteristics.

Figure 2

Schematic diagram of the waterflooding experiment setup: (a) sandstone model diagram; (b) high-temperature and high-pressure reactors; and (c) experimental device, wherein (1) positive displacement pump; (2) confining pressure pump; (3) fluid tank; (4) stereomicroscopes; (5) high-temperature and high-pressure reactors; (6) computer; (7) metering device; and (8) hand pump.

The real core microscopic model displacement experiment mainly consists of a high-temperature and high-pressure reactor (Figure 2b), constant-speed and constant-pressure pumps, a constant-temperature system, a stereomicroscope, a metering device, a computer, and an image acquisition system (Figure 2c). The visual high-temperature and high-pressure reactor with an upper visual window was used to hold the sandstone models, through which the seepage process could be observed and recorded with the stereomicroscope. The computer and metering device were used to record the pressure difference and stabilized flow rates between the two ends of the model, as well as to analyze and process the captured images. The constant-speed and constant-pressure pumps included a positive displacement pump and a confining pressure pump. The positive displacement pump was used to pressurize the displacement fluid, while the confining pressure pump simulated the overburden pressure of the rock and protected the sandstone model. The hand pressure pump was used to maintain system pressure, and the constant-temperature system heated the model.

4.1 Petrophysical properties and mineral composition characteristics

The wettability of 13 samples was determined to be neutral to weakly hydrophilic using the spontaneous imbibition method. The porosity of each sample was essentially between 6.45 and 11.78%, averaging 8.79% (Table 1). The permeability mainly ranged between 0.050 and 0.433 × 10⁻³ μm², averaging 0.199 × 10⁻³ μm², indicating poor petrophysical properties. The XRD test results showed that the mineral composition of each sample consisted mainly of feldspar, quartz, calcite, and clay minerals, corresponding to mass percentages of 42.75, 39.73, 3.58, and 13.61%, respectively, indicating moderate compositional maturity. Among the clay minerals, illite was the primary component, followed by chlorite and kaolinite, with a minimal amount of illite–smectite (I/S) mixed layer. Their average relative contents were 52.35, 23.67, 17.94, and 6.04%, respectively. The lower petrophysical properties and the higher clay mineral content suggest a strong cementation degree and complex PTS in the sandstone samples.

4.2 Pore and throat types

According to the results of casting thin sections and SEM, the pore types in each sample consist of dissolution pores, residual intergranular pores, and intercrystalline pores. These dissolution pores include feldspar dissolution pores (Figure 3a and c), lithic dissolution pores, and intergranular dissolution pores, corresponding to the average face ratio of 1.15, 0.08, and 0.05%, respectively. The dissolution pores have complex shapes, and some feldspar grains are completely dissolved to form moldic pores with larger pore sizes, which has a positive effect on improving reservoir properties. Residual intergranular pores (average face ratio of 1.04%) are primary pores remaining after compaction and cementation of the reservoir (Figure 3b). These pores have irregular shapes and large diameters, with lining chlorite and hair-like illite often occurring along the pore edges. Intercrystalline pores (average face ratio of 0.04%) are micro-pores formed by clay minerals filling intergranular pores and dissolution pores generated by reservoir cementation (Figure 3a and c), and the pore spaces are fine and complex. The throat types in each sample are primarily sheet-shaped, bent sheet-shaped (Figure 3b and d), and tubular. Sheet-shaped and bent sheet-shaped throats indicate that the rock particles have undergone strong compaction, compressing the throats into narrow elongated ribbons, making it difficult for them to connect with surrounding large pores. Tubular throats primarily consist of clay minerals filling pores, splitting them into irregularly sized, tube-shaped, or bundle-shaped intercrystalline pores, thereby reducing the connectivity between pores and throats (Figure 3e and f).

Figure 3

Pore and throat types of typical samples: (a) Feldspar dissolution pores and kaolinite intercrystalline pores, L54, 2184.94 m; (b) residual intergranular pores and bent sheet-shaped throats, H36, 2,074.37 m; (c) kaolinite filled the feldspar dissolved pores and formed intercrystalline pores, H13, 2673.59 m; (d) development of sheet-shaped and bent sheet-shaped throats, and the chlorite film attached to the surface of the particles, L71, 2331.25 m; (e) illite filled pores and formed intercrystalline pores and tubular throats, H37, 2088.72 m; (f) kaolinite filled pores and formed a tubular throat, H21, 2183.01 m.

4.3 HPMI results

Based on the morphology of the HPMI capillary pressure curves (Figure 4a), PSD curves (Figure 4b), and PTS parameters (Table 2), the PTS of the 13 samples were divided into four categories. From type I to type IV, the capillary pressure curves shift progressively upward, with the threshold pressure (P _t) increasing accordingly (averaging 0.43, 0.69, 1.18, and 4.23 MPa, respectively). Accordingly, the PSD curves exhibit single-peak distributions, with the peak of the pore–throat radius gradually shifting to the left. The average pore–throat radius (R _a) (averaging 0.30, 0.19, 0.12, and 0.02 μm, respectively) and the maximum pore–throat radius (R _max) (averaging 1.71, 1.06, 0.71, and 0.21 μm, respectively) progressively decreased. Additionally, the maximum mercury injection saturation (S _max) (averaging 85.18, 80.18, 75.86, and 58.64%, respectively) and the mercury withdrawal efficiency (W _e) (averaging 50.09, 43.20, 37.89, and 15.89%, respectively) decreased sequentially, indicating that the reservoir space and pore–throat connectivity gradually deteriorated between the four types of samples. However, there are significant differences in the sorting coefficient (S _P) and homogeneity coefficient (α) of the four types of PTS, indicating complex pore–throat configurations and strong PTS heterogeneity in tight sandstone samples. Types I and III have lower S _P (averaging 2.66 and 2.09, respectively) and higher α (averaging 0.23 and 0.27, respectively), indicating that the PSD has good sorting and concentration. The difference lies in that type I PTS developed larger pore–throat than type III PTS. Type II and IV PTSs have higher S _P (averaging 3.00 and 4.39, respectively) and lower α (averaging 0.17 and 0.13, respectively), indicating poor sorting and non-uniform PSD. Type II PTS simultaneously develops numerous large and small pore–throats with good pore–throat configuration and connectivity, resulting in higher W _e and petrophysical properties. In contrast, type IV PTS develops numerous nanoscale pore–throats with the strongest PTS heterogeneity and the worst petrophysical properties.

Figure 4

Characteristics of capillary pressure curves (a) and PSD curves (b) obtained by HPMI.

4.4 Full PSD and fractal characteristics

NMR testing offers a non-destructive method for the comprehensive characterization of pore–throat spaces. As shown from the NMR saturated water T ₂ spectrum (Figure 5a), type I, III, and IV samples exhibit single-peak distribution characteristics with peaks increasingly shifted to the left, while type II samples display a bimodal feature. This is consistent with the findings from HPMI, which suggests that type II samples have a greater development of large and small pore–throats. The PSD curves obtained from HPMI were used to convert the NMR-saturated water T ₂ spectrum, quantitatively characterizing the full PSD of tight sandstone samples. The results indicate that the full PSD for each sample is between 0.0001 and 10 μm (Figure 5b).

Figure 5

NMR saturated water T ₂ spectrum curves (a) and full PSD curves (b) of 13 samples.

By plotting the Lg V(c) and Lg r curves of each sample (Figure 6), it is found that the fractal characteristic curves of the full PSD exhibit a distinct multi-segment pattern rather than a straight line, with significant inflection points at pore–throat radii of 1, 0.1, and 0.01 μm. This indicates that the PTS of tight sandstone samples have multi-fractal characteristics. Based on these fractal inflection points, the pore–throat spaces of tight sandstone samples are divided into macropores (>1 μm), mesopores (0.1–1 μm), micropores (0.01–0.1 μm), and nanopores (<0.01 μm), and the corresponding fractal dimensions were defined as D ₁, D ₂, D ₃, and D ₄, respectively. The fractal dimensions for each pore–throat type across all samples exhibit the relationship D ₁ > D ₂ > D ₃ > D ₄, with corresponding average fractal dimensions of 2.8509, 2.8189, 2.5462, and 2.3143, respectively (Table 3). This indicates self-similarity within each pore–throat type, with different levels of complexity. Macropores and mesopores with larger pore–throat radii are more susceptible to the filling and transformation of clay minerals, resulting in stronger PTS heterogeneity, while micropores and nanopores with higher proportions and relatively smaller pore–throat radii exhibit better overall homogeneity and similarity, corresponding to smaller fractal dimensions. Furthermore, the D _t value for each sample ranges from 2.2373 to 2.8712, with an average value of 2.5630, indicating strong overall PTS heterogeneity. Moreover, the D _t value increases progressively from type I to type IV PTS, with average values of 2.3571, 2.4728, 2.6293, and 2.7709, respectively.

Figure 6

Fractal characteristic curves of typical samples H153 (a), H36 (b), L91 (c), and H19 (d) in the four types of PTS.

4.5 Fluid mobility

Table 2 shows the movable fluid parameters derived from NMR for 13 samples, with the MFS for each sample ranging from 25.26 to 62.16%, averaging 45.77%. This indicates overall low fluid mobility in tight sandstone samples, with most of the pore–throat spaces occupied by bound fluids. The MFS decreases sequentially for the four types of PTS, with average values of 57.24, 50.09, 44.96, and 31.05%, respectively. Furthermore, the proportion, MFS, and permeability contribution values of different pore–throat types were statistically analyzed to quantitatively assess the fluid mobility of tight sandstone samples (Table 3 and Figure 7). The results show that micropores have the highest fluid mobility among the 13 samples (average MFS of 18.53%), primarily due to their large reservoir space (averaging 40.75%) and their significant contribution to permeability (averaging 0.0182 × 10⁻³ μm²), which facilitates the occurrence and seepage of mobile fluids. Mesopores and nanopores show relatively high fluid mobility (average MFS of 13.72 and 12.87%, respectively), with similar proportions of reservoir space (averaging 26.70 and 31.43%, respectively), but mesopores have significantly higher permeability contribution values than nanopores (averaging 0.1528 × 10⁻³ and 0.0003 × 10⁻³ μm², respectively). Macropores develop only in type I PTS, with the largest pore–throat radii and permeability contribution values (averaging 0.1543 × 10⁻³ μm²), but their development is limited (average proportion of 1.12%) and reduces pore–throat interconnectivity, resulting in the lowest MFS (averaging 2.82%).

Figure 7

Proportion of different pore–throat types and distribution of MFS in each sample.

From the perspective of PTS classification, type I PTS have higher development of mesopores (average proportion of 52.25%) and greater seepage capability (average permeability contribution value of 0.1729 × 10⁻³ μm²), resulting in higher fluid mobility for mesopores (average MFS of 29.95%). In types II and III PTS, macropores are absent and the proportion of mesopores decreases (average proportions of 30.92 and 21.47%, respectively), resulting in mainly movable fluid being present in micropores (average MFS of 20.92 and 23.97%, respectively). In type IV PTS, movable fluids are mainly present in nanopores (average MFS of 18.46%). Although nanopores are more developed (average proportion of 60.08%), their smaller pore–throat radii result in lower connectivity and percolation ability, resulting in poor overall fluid mobility.

4.6 Microscopic waterflooding characteristics

The waterflooding results for each sample showed that the initial oil saturation (S _io) ranged from 40.36 to 71.36%, averaging 54.44%, while the average E _w was 39.07% (Table 2), indicating that most of the crude oil remained undisplaced and trapped in the pore–throat spaces. Significant differences in oil–water seepage paths and residual oil distribution characteristics were observed among different PTSs (Figure 8). From type I to type IV PTS, the waterflooding patterns with decreasing core permeability and deterioration of PTS successively showed uniform, network, finger-network, and finger displacement, and the morphology of residual oil gradually transitioned from scattered membranous, angular, isolated, and patchy shapes to continuous cluster and sheet shapes [38,54].

Figure 8

Oil–water seepage paths and residual oil distribution characteristics corresponding to type I PTS (a–d), type II PTS (e–h), type III PTS (i–l), and type IV PTS (m–p).

In type I PTS, mesopores and macropores account for a higher proportion, corresponding to the samples with the best petrophysical properties and the highest S _io (averaging 67.06%), the waterflooding patterns primarily exhibiting uniform displacement. In the early stages of waterflooding, the formation water preferentially enters the larger pores with less resistance, forming multiple seepage paths, and the forward movement rate of formation water along each seepage path remains stable. The waterflooding front remains substantially parallel at different times, achieving uniform waterflooding coverage and high E _w (averaging 51.30%) (Figure 8a–c). After waterflooding, the residual oil primarily exists as dispersed isolated, angular, patchy shapes, and a small amount of cluster shapes, trapped in the pore blind ends and the grain edge of the core (Figure 8d).

In type II PTS, mesopores and micropores account for a higher proportion, with a relatively high S _io (averaging 57.27%) and a relatively uniform distribution, primarily characterized by network oil–water seepage paths. During the initial waterflooding phase, injected water enters through larger channels at the inlet end and forms multiple injection lines that advance irregularly in a branch-like manner. The waterflooding front advances in a network-like pattern, with significant local bypassing. As the displacement pressure and injected water volume increase, the injected water gradually penetrates smaller pores, thereby widening the seepage paths and increasing the waterflooding coverage and E _w (averaging 43.47%) (Figure 8e–g). After waterflooding, the residual oil types are mainly membranous, angular, patchy, and cluster shapes (Figure 8h).

In type III PTS, micropores account for a higher proportion, the pore–throat radius and connectivity decrease, and PTS heterogeneity increases. The S _io is lower (averaging 50.61%) and discontinuously distributed, primarily characterized by finger-network displacement. During the initial waterflooding phase, injected water extends forward in a finger-like manner along larger pore–throats. As displacement proceeds, the injected water gradually enters smaller pores and spreads outward along the original paths to form tree branch-like paths. The waterflooding network gradually becomes smaller and denser, and the waterflooding coverage expands, but the E _w remains low (averaging 37.85%) (Figure 8i–k). After waterflooding, the residual oil types include membranous, angular, cluster, and sheet shapes (Figure 8l).

In type IV PTS, nanopores account for a higher proportion, with the lowest pore–throat radius and connectivity, higher clay mineral content, higher cementation degree, and PTS heterogeneity, leading to the undersaturation of some pores with crude oil. S _io and E _w are the lowest (averaging 44.11 and 25.05%, respectively) and unevenly distributed. During waterflooding, there are relatively high-permeability channels, and injected water advances along one or more channels in a finger-like manner, with significant bypasses occurring, resulting in large areas of residual oil remaining in the pore–throat space (Figure 8m–o). After waterflooding, the residual oil types are mainly in cluster and sheet shapes (Figure 8p).

5.1 Effects of petrophysical properties on MFS and E _w

The correlation between permeability and MFS and E _w is stronger than that between porosity and these two (Figure 9a and b). This is primarily because both MFS and E _w reflect fluid mobility within the reservoir, which is closely related to pore–throat connectivity and seepage capacity. High-porosity samples merely have larger storage spaces and weaker control over fluid mobility in the reservoir. Higher permeability in tight sandstone samples promotes fluid penetration into pores, where movable fluids in pore–throats are less affected by capillary resistance in finer throats and adsorption by hydrophilic minerals, resulting in higher MFS. For samples with better petrophysical properties, injected water preferentially enters larger pore–throats, increasing the number and width of seepage paths and expanding the waterflooding coverage, resulting in a significant increase in E _w. Interestingly, the MFS of each sample is higher than the E _w and has a high correlation with petrophysical properties. This is because only a single-phase flow of the formation water in the pore–throat space is simulated by NMR during centrifugation, resulting in MFS that reflects the theoretical or maximum movable degree of the fluids in tight sandstone samples. E _w, on the other hand, is the relative percolation result of oil–water two-phase fluids in the pore–throats, influenced by reservoir PTS, fluid properties, solid–liquid interfacial interactions, etc., and accurately evaluates the recoverable degree of tight oil [55,56]. Furthermore, the reservoir quality index (RQI) of each sample was found to have a good correlation with MFS and E _w (Figure 9c), indicating that fluid mobility in tight sandstone samples is influenced by both porosity and permeability, with higher pore–throat space and larger pore–throat development both facilitating fluid seepage. There is also a good positive correlation between permeability and MFS for different pore–throat types in each sample (Figure 9d). Fewer macropores develop, so they are not included in the analysis. Mesopores and micropores contribute to higher permeability, which corresponds to higher MFS, indicating that the development of larger pore–throats favors the occurrence and seepage of movable fluids. Nanopores have smaller pore–throat radii, resulting in lower fluid seepage capability, but their high pore–throat proportion and homogeneity allow some movable fluids to occur.

Figure 9

Correlation between MFS and E _w with porosity (a), permeability (b), and RQI (c), and the correlation between MFS and permeability in different pore–throat types (d).

5.2 Effects of PTS on MFS and E _w

R _a and R _max, which represent the pore–throat size, are significantly positively correlated with MFS and E _w (Figure 10a and b). This indicates that larger pore–throat radii are more conducive to fluid seepage. Conversely, finer pore–throats have a stronger binding effect on movable fluids, causing greater capillary resistance, resulting in more frequent waterflooding bypassing and lower E _w. P _t exhibits a good negative correlation with MFS and E _w (Figure 10c). This is because higher displacement pressures indicate poorer petrophysical properties of the sample, with a higher proportion of smaller pore–throats that are unfavorable for movable fluid flow. Additionally, injected water initially requires a higher starting pressure to enter the pore–throats, and once it breaks through smaller pore–throats, the displacement pressure decreases, allowing water to preferentially enter larger pore–throats for displacement, which easily leads to finger displacement and lower E _w. S _max and W _e have a good positive correlation with MFS and E _w (Figure 10d and e), indicating that higher reservoir space and pore–throat connectivity are beneficial to the occurrence and flow of reservoir fluids. Since the mercury withdrawal process is similar to the water-oil displacement process, both involve the non-wetting phase (mercury and injected water) displacing the wetting phase (air and crude oil) under high pressure, many scholars approximate the W _e as the E _w [57,58]. The difference lies in the mercury withdrawal process, in which the pressure drops rapidly, causing mercury in smaller pore–throats to be preferentially expelled and isolated, disrupting the continuity of the mercury and causing more mercury trapped in larger pore–throats. In contrast, during the water-oil displacement process, the injection pressure gradually increases, allowing a more continuous flow of the non-wetting phase, resulting in a generally higher E _w than W _e across samples. Therefore, in the actual waterflooding development of the reservoir, it is advisable to avoid sudden pressure changes and maintain gentle water injections, which is conducive to improving the recovery rate of tight oil.

Figure 10

Correlation between MFS and E _w with R _a (a), R _max (b), P _t (c), S _max (d), W _e (e), S _p (f), α (g), and D _t (h), and the correlation between MFS and fractal dimensions for different pore–throat types (i).

S _p and α of the pore–throats in tight sandstone samples have a relatively weak impact on MFS and E _w (Figure 10f and g). This is because samples with high S _p and low α indicate significant PSD variation. Although a higher proportion of larger pore–throats can reduce the displacement pressure differentials, the injected water tends to advance along high-permeability paths formed by larger pore–throats, leading to bypass flow phenomena that make it difficult to displace oil from certain pore–throats. Conversely, samples with smaller S _p have a more uniform distribution, although exhibiting they generally have finer pore–throats. Under larger displacement pressure differentials, once the injected water breaks through, it can advance uniformly, and the displacement process often manifests as uniform or network displacement, with a larger sweep range and a better displacement effect. The D _t of various PTS shows a good negative correlation with MFS and E _w (Figure 10h), indicating that stronger PTS heterogeneity is detrimental to the flow of reservoir fluids. Furthermore, there is a good negative correlation between MFS and fractal dimensions for different pore–throat types (Figure 10i). Micropores serve as primary sites for movable fluids and possess high seepage capacity, the development and complexity of micropores has a significant impact on fluid mobility. Thus, for tight sandstone reservoirs, low permeability and fine pore–throats are not necessarily problematic; rather, PTS heterogeneity emerges as the principal factor governing fluid mobility and E _w, highlighting the importance of integrated pore–throat characterization in reservoir engineering.

5.3 Effects of clay mineral composition and content on MFS and E _w

The effects of different clay mineral compositions, occurrences, and their relative contents on the occurrence and seepage of fluids in reservoirs vary. The total clay minerals content has a good negative correlation with MFS and E _w (Figure 11a). The presence of clay minerals tends to increase the specific surface area and roughness of the pore–throats, enhancing the interaction forces between the pore walls and the crude oil, which leads to elevated water injection pressures and difficulty in separating the crude oil, thereby reducing E _w. High relative content of illite often occurs in the form of curved sheets or filamentous structures that fill into larger pore–throats, easily reducing the pore space or blocking the throats, making it difficult for fluids to flow and form trapped fluids. Additionally, during waterflooding, illite easily swells in the presence of water, blocking pore–throat spaces and reducing pore–throat connectivity and seepage capacity. Therefore, the relative content of illite shows a good negative correlation with MFS and E _w (Figure 11b). Chlorite has dual effects on reservoir fluid mobility and reservoir space size. Early cemented chlorite films can effectively preserve pores, while late-formed chlorite liners fill pore spaces, resulting in less significant correlations between chlorite and MFS and E _w (Figure 11c). Kaolinite is often associated with feldspar dissolution pores and commonly fills pore spaces in hexagonal thin plates, increasing PTS heterogeneity. Under water injections, kaolinite can migrate easily causing severe water sensitivity. However, given its relatively low content, it has little impact on MFS and E _w (Figure 11d). In the process of waterflooding development, it is crucial to fully consider the clay mineral sensitivity and design waterflooding schemes tailored to different types of clay minerals.

Figure 11

Correlation between MFS and E _w with total clay mineral content (a), illite relative content (b), chlorite relative content (c), and kaolinite relative content (d).

5.4 Effect of water injection volume multiple on E _w

Statistics were conducted on the E _w of each sample at different water injection volume multiples (Figure 12). The results showed that when the water injection volume multiple was less than 1 pore volume (PV) (water-free period), the seepage channels and sweep area of the injected water for each sample expanded rapidly, increasing E _w. However, when the water injection volume multiple reached 1 PV, the increase in E _w was small for each sample. This was because, with the increase in water saturation or water cut in the pore–throat space, the relative permeability of the water phase increased and became more mobile, and oil–water seepage channels had already formed. The continuous increase in water injection pressure and volume made it difficult to displace the oil trapped in smaller pore–throats. The type IV samples had the worst PTS and petrophysical properties with strong PTS heterogeneity, yet their E _w increased slowly at a high water injection volume multiple. This was due to the overall small pore–throat radius of these samples, with a high proportion of nanopores. Under higher displacement pressure, injected water gradually entered the nanopores and displaced the oil phase. However, due to the limited reservoir space of the nanopores and lower S _io, the overall increase in E _w was small. Additionally, the E _w of type III sample L91 was significantly higher than that of type II samples H36 and H37. Although samples H36 and H37 had better petrophysical properties and PTS with developed mesopores and micropores, their S _p and D _t were both higher than those of sample L91, indicating stronger PTS heterogeneity. Increasing the content of larger pore–throats resulted in a reduced displacement pressure differential, causing injected water to mostly advance along the high-permeability channels and form continuous cluster and sheet shape residual oil. Therefore, it is recommended to maintain a stable injection pressure and injection volume multiple of less than 2 PV in the process of waterflooding in tight sandstone reservoirs, which is conducive to improving the water injection utilization rate and protecting the environment.

Figure 12

E _w of each sample under different injection volume multiples.

5.5 Analysis of residual oil formation reasons

In the actual water injection development process of tight oil reservoirs, the morphology and formation mechanisms of residual oil are extremely complex. The formation of a specific residual oil morphology in pore–throats is typically controlled by several mechanisms. For instance, clustered residual oil is often controlled by a combination of injected water bypass flow and low displacement pressure. Therefore, this paper only analyzes and summarizes the common morphologies and formation mechanisms of residual oil, which mainly include the following five categories: (1) non-piston displacement within pore–throats. Due to the differences in the wettability of rock particle surfaces and their heterogeneity, as well as the unbalanced driving forces and capillary pressures that the injected water experiences during seepage, the velocity of the injected water moving through the pore channels is uneven. Water tends to flow faster in the center or edges of the pore channels, resulting in the formation of membranous and isolated residual oil (Figure 13a and b). (2) No seepage outlet or blind pores. The sedimentation and diagenesis of the rock particles in the reservoir create linear or closer contacts between local rock particles, creating blind pores. This configuration makes it difficult for the oil in the pores to be separated by the displacing agents, leading to the formation of angular and patchy residual oil (Figure 13c and d). (3) Blockage due to small pore–throats. Continuous oil droplets in the pore–throats may experience blockage when passing through very small pore–throats due to an imbalance between driving force and resistance, which easily leads to the occurrence of the Jamin effect and the formation of isolated residual oil (Figure 13b). (4) Low driving force or low displacement pressure differential within the pore–throats. When the connectivity of pore–throats is poor or the displacement direction of the injected water is perpendicular to the pore channels, the pressure differential at both ends of the pore channel may be small or even zero, making it difficult to displace the oil contained therein, resulting in the formation of patchy residual oil. Furthermore, as the displacement radius increases, the driving force of the injected water in the pore–throats near the outlet becomes very low, preventing further displacement and leading to the formation of cluster and sheet residual oil (Figure 13e and f). (5) Water coning and bypassing of injected water. Due to the overall small pore–throat radius and severe local cementation of the tight sandstone samples, the injected water is likely to seep along the channels of larger pore–throats, often advancing in a finger displacement, which easily results in the formation of cluster and sheet residual oil (Figure 13e and f).

Figure 13

Schematic diagram of distribution and formation mechanisms of microscopic residual oil: (a) membranous shape residual oil; (b) isolated shape residual oil; (c) angular shape residual oil; (d) patchy shape residual oil; (e) cluster shape residual oil; and (f) sheet shape residual oil.