In recent years, low permeability reserves are taking a higher and higher proportion in China's proven petroleum reserves, and will be the main resource of increasing reserves and production. With the rapid growth of oil demand at home and abroad, exploration and development objects of oil and gas are increasingly complex, the grade of reserves is getting worse, and more and more low-permeability oilfields are put into development^[1,2]. At present, Wushi oilfield in Beibuwan Basin hasn’t been put into development, where of the proven oil reserves of nearly 100 million tons, more than half are in low-permeability sandy conglomerate fault-block reservoirs. Due to the limited natural energy of fault-block reservoir, water injection is needed to replenish energy. Moreover, the reservoirs in the oilfield have poor physical properties, low productivity, and strong water sensitivity, so the problem of "no injection and no recovery" occurs in the process of waterflooding development.

The ratio of fluid permeability to gas permeability of the low permeability sandy conglomerate reservoir in Wushi oilfield is 0.001-0.295, which are lower than other types of reservoirs in Beibuwan Basin^[3]. At the same time, according to the water-sensitive damage test results of a large number of cores, the reservoirs are generally greater than 50% in water-sensitive damage degree, belonging to the moderate to strong-strong water-sensitive ones. Due to the low content of clay minerals and expansive clay minerals in the low permeability sandy conglomerate reservoirs of Wushi oilfield, the water-sensitive damage mechanism of the reservoirs is not only conventional clay mineral expansion. Unclear understanding on reservoir water sensitive damage mechanism makes it difficult to select injection water source and hinders effective development of the oilfield. In addition, the reservoir water sensitive damage makes it difficult to evaluate the productivity of the low permeability sandy conglomerate reservoir correctly, and reservoir water sensitive damage caused by different salinities of injection water in oil displacement experiment cannot be evaluated, these issues need to be solved urgently.

In view of the above problems, a seepage capacity evaluation parameter of oil-water two phases for low permeability reservoirs was established based on the global mobility theory and large amounts of water-oil displacement experiment results^[4] to analyze the experimental effects of oil displacement by water with different salinities. By comparing the results of scanning electron microscopy, X diffraction of clay minerals, nonlinear seepage and nuclear magnetic resonance experiments, particle migration inhibition experiments and core compatibility experiments before and after oil displacement with water of different salinities, we tried to find out the mechanisms of water sensitive damage and enhancing water flooding effect of low permeability sandy conglomerate reservoirs in Wushi region of Beibuwan Basin. Meanwhile, based on the generalized Darcy formula considering the effective driving factor and mass conservation law, a production equation of the oil-water two phases flow well considering low-speed non-Darcy seepage and reservoir stress sensitivity is established to evaluate the effect of changes in reservoir physical properties and oil-water two-phase seepage capacity caused by water sensitive damage on productivity quantitatively and to select injection water source suitable for the low permeability sandy conglomerate reservoirs in Wushi oilfield.

The damage mechanisms of injected water to the reservoir mainly include clay mineral hydration and particle migration.

1.1.1. Clay mineral hydration

Clay minerals are ubiquitous in reservoirs, and when contacting with incompatible fluids injected into the reservoir, the clay minerals will hydrate, expand, disperse and migrate, thus blocking the throat, leading to water-sensitive damage in the reservoir^[5]. Clay mineral hydration includes internal surface hydration and osmotic hydration. The action mechanism of hydration is mainly as follows: polar water molecules are adsorbed by electric charge of clay particles and hydrogen and oxygen bonds formed by the hydrogen of oxyhydrogen or oxygen in clay mineral crystals with oxygen or hydrogen in water molecules to form hydration membrane of water molecules in directional arrangement around the clay particles; meanwhile, polar water molecules are also adsorbed by exchangeable positive ions on the surface of clay mineral crystals, to form hydration membrane of water molecules in directional arrangement at interlayers or on crystal surface of clay mineral^[6].

Internal surface hydration, also known as "crystal expansion" or "interlayer expansion", is mainly caused by exchangeable positive ions in interlayer of clay mineral absorbing polar water molecules and forming hydration membrane. When exchangeable positive ions are Mg²⁺, K⁺, H⁺, Ca²⁺, etc., the attraction between clay mineral chips is higher than the exchangeable positive ion of Na⁺, so the hydration membrane created is thinner with water molecules in regular directional arrangement. The expansive abilities of clay minerals in descending order are montmorillonite, mixed clay containing expansive layers, illite and kaolinite^[7].

Osmotic hydration refers to the hydration of the outer surface. It would cause clay minerals to hydrate, expand, disperse and migrate, thus blocking the throat. The reason is that when fluid with low salinity is injected into the reservoir, directional water membrane would be formed around the clay particles, which makes the electrostatic repulsion of the double electric layer increase and thus push clay mineral particles away from each other. The osmotic equilibrium state of semi-permeable membrane around the clay mineral particle is the main factor controlling the outer surface hydration of clay minerals^[7].

1.1.2. Particle migration

Particle migration mainly includes the invasion and retention of external particles and the release and capture of internal particles in the reservoir. External particles are mainly from injected water, while internal particles are mainly clay minerals and other particles such as quartz and feldspar. They release and migrate, thus blocking the pore throat, under the action of hydrodynamics or the combination of chemistry and hydrodynamics.

Al-Rasheedi et al.^[8] studied the quantitative law of non-Brownian motion of particles by flow and centrifugal experiments in the microscopic glass model and found that the critical velocity decreased with the increase of pH value and the decrease of salinity, but the effect of salinity was less significant than that of pH value, and pointed out that the critical velocity decreased with the increase of particle size. Leontaritis et al.^[9] found that there was a certain relationship between the critical velocity and the fluid properties. For example, under the same ionic strength for Berea sandstone, the critical velocity measured by KCl was about 5 times of that measured by NaCl and CaCl₂. Kartic et al.^[10] studied the water sensitivity phenomenon caused by the dispersed migration of kaolinite minerals in Berea sandstone by flow experiment, and found that the critical salinity was only related to the type of positive ion, but not related to the type and flow velocity of negative ion. The critical salinity of divalent positive ion is very small, while the critical salinity of monovalent positive ion decreases with the decrease of hydration ion radius. Omar et al.^[11] concluded through study that water-sensitive damage was related to the type and content of exchangeable positive ion in clay minerals, and based on the analysis of experimental data, they suggested that increasing exchangeable Mg²⁺ content and decreasing exchangeable Na⁺ and Ca²⁺ content could inhibit water-sensitive damage. Therefore, low pH value, high salinity and high K⁺ and Mg²⁺ concentrations are conducive to stabilize particles, inhibiting the migration of particles in the process of waterflooding development, and reducing water sensitive damage to reservoir.

For two-phase flow of oil and water in porous media, the flow capacity of each phase can be described by mobility. In order to describe the oil-water two-phase seepage capacity, Li and Horne^{[12,13,14,15]} established the Li-Horne model in 2006 and proposed the global mobility parameter. In 2019, Wang et al.^[4] demonstrated the reliability of the global mobility in characterizing oil-water two-phase seepage capacity with a large amount of displacement experimental data and production data. In the global mobility, the oil and water phases are considered as a whole, and the main parameters describing fluid flow in the porous medium are used to express global mobility as follows^{[12,13,14,15]} :

where \({{M}_{\text{w}}}^{*}={{K}_{\text{w}}}^{*}/{{\mu }_{\text{w}}}\), \({{M}_{\text{nw}}}^{*}={{K}_{\text{nw}}}^{*}/{{\mu }_{\text{nw}}}\).

According to the Li-Horne model, the global mobility can be calculated by using displacement experimental data^[4]. Based on the Li-Horne model, a new productivity index^[4,16] is proposed. The new productivity index does not only consider permeability and relative permeability of the reservoir rock, but also the fluid properties of the reservoir (oil and water viscosity), and relative storage capacity of the reservoir (1-S_wi) etc. The new productivity index is expressed as follows^[4,16]:

According to Li-Horne model and formula (2), the relationship between oil production and the new productivity index is obtained as follows^[4,16]:

where \({{b}_{0}}=A{{M}_{\text{e}}}^{*}\Delta \rho g\).

The generalized Darcy formula describing the low-velocity non-Darcy seepage law is as follows^[17]:

where \(\delta =1-G\left( {{r}_{\text{e}}}-{{r}_{\text{w}}} \right)/\Delta p\)

In equation (4), δ is the effective driving factor, and its physical meaning is the pressure ratio remaining for effective oil displacement after overcoming the resistance of low-velocity non-Darcy seepage in the reservoir.

The reservoir stress-sensitivity can be characterized by the exponential relationship between permeability and effective stress as follows^[18,19]:

where \({{\alpha }_{\text{k}}}=\frac{1}{K}\frac{\partial K}{\partial p}\)

According to the mass conservation law, a quasi-pressure transformation equation (equation 6)^[20] has been introduced to establish a two-phase steady flow mathematical model considering both low-velocity non-Darcy seepage and reservoir stress sensitivity, as shown in equation (7).

where \({{\alpha }_{\text{k}}}^{*}=\frac{1}{K}\frac{\partial K}{\partial m\left( p \right)}\), \[{{q}_{\text{wo}}}\text{=}{{q}_{\text{w}}}/{{q}_{\text{o}}}\],

\[m\left( p \right)=\int_{{{p}_{\text{m}}}}^{p}{\left( \frac{{{\rho }_{\text{o}}}{{K}_{\text{ro}}}{{\delta }_{\text{o}}}}{{{\mu }_{\text{o}}}}+\frac{{{\rho }_{\text{w}}}{{K}_{\text{rw}}}{{\delta }_{\text{w}}}}{{{\mu }_{\text{w}}}} \right)\text{d}p}\]\[{{\xi }_{\text{wf}}}\text{=}\frac{\exp \left[ {{\alpha }_{\text{k}}}^{*}m\left( {{p}_{\text{wf}}} \right) \right]-1}{{{\alpha }_{\text{k}}}^{*}}\],

\[{{\xi }_{\text{init}}}\text{=}\frac{\exp \left[ {{\alpha }_{\text{k}}}^{*}m\left( a{{p}_{\text{init}}} \right) \right]-1}{{{\alpha }_{\text{k}}}^{*}}\]

Oil production equation of the oil-water two-phase flow well considering low-velocity non-Darcy fluid seepage and reservoir stress sensitivity is derived as follows:

\[{{q}_{\text{o}}}\text{=}\frac{0.543Kh\left( {{\xi }_{\text{init}}}-{{\xi }_{\text{wf}}} \right)}{\left( 1+{{\alpha }_{\text{k}}}^{*}{{\xi }_{\text{wf}}} \right){{B}_{\text{o}}}\ln \frac{{{r}_{\text{e}}}}{{{r}_{\text{w}}}}}\times \]

The effect of reservoir water-sensitive damage on productivity includes two aspects: (1) Water-sensitive damage makes the physical property of formations near the injection well and oil-water transition zone turn worse, thus the injection capacity and energy transfer rate of the injection well reduce. In order to ensure the injection-production balance, the injection pressure difference has to be constantly increased. But restricted by the reservoir fracture pressure, the injection rate and reservoir sweep speed must be limited. In this circumstance, formation energy cannot be replenished in time with the production of the oil well, so formation pressure and in turn productivity of oil well will drop constantly. Reservoir pressure level control coefficient (a) in the oil production equation is used to control the effect. (2) After water breaks through in the production well, water-sensitive damage will affect oil-water two-phase seepage capacity of the reservoir near the production well and in turn the productivity of the production well. The relative permeability of oil and water (K_ro and K_rw) in the oil production equation are used to control the effect.

The 9 core samples in Table 1 were taken from the same layer of an exploration well of Paleogene Eocene Liushagang Formation low-permeability sandy conglomerate reservoir of the Wushi region in Beibuwan Basin, and they are strongly representative in the whole Wushi region. Table 2 shows the parameters of injected water used in different displacement experiments. The crude oil used in the experiment was 4.0 mPa·s in viscosity and 0.8 g/cm³ in density. According to the velocity sensitivity experiment, the reasonable displacement velocity was 0.3 mL/min, to avoid the interference of velocity sensitivity.

Core samples 1-6 were used in experiments testing reservoir water sensitive damage mechanisms under normal temperature and pressure. The displacement experiments include two types, the salinity of injected water increasing and the salinity of injected water decreasing. In the former type of experiments, the formation water, and then water with salinity 2 times of the formation water and water with salinity three times of the formation water were injected into the same core in turn. In the latter type of experiments, the formation water, water with salinity 1/2 times of the formation water and distilled water were injected into the same core. After the displacement experiment of one type of injected water was done, the same core sample was directly re-saturated with oil to the initial irreducible water saturation and aged, and then the displacement experiment of the next type of injected water was done on this core sample. The displacement experiment was done at constant flow rate, and the cumulative injected water was 6-8 times of pore volume multiples, and the displacement time was about 2 h.

Core samples 7-9 were used in enhancing water flooding effect experiment under the formation temperature and pressure. The injected water was formation water, formation water+KCl and formation water+KCl+MgCl₂. The displacement experiment was done at constant flow rate. The cumulative injected water volume of these core samples were 100, 204 and 215 times of pore volume multiples respectively.

Core sample 1 and core sample 2 were used to conduct oil displacement experiments by water of different salinities, and the relative permeability curves obtained are shown in Fig. 1.

The relative permeability curves variation from experiments with injected water of different salinities were quantitatively analyzed by using the global mobility and the new productivity index. Li-Horne model was used to fit the relationship between production rate (q_o) and reciprocal value of oil recovery (1/R) of the oil displacement experiments by water with different salinities. The fitting results were all good, with correlation coefficients all greater than 0.9. Fig. 2 shows the fitting results of core sample 1 when the injected water is formation water.

According to the fitting relationship between production rate and reciprocal value of oil recovery, the global mobility and productivity index corresponding to the relative permeability curve under the injected water of different salinities were calculated with the core and fluid parameters, the results are shown in Table 3. It can be seen that in the experiment of injected water salinity decreasing, the degree of water-free recovery, oil displacement efficiency, movable oil saturation, global mobility and productivity index decrease continuously with the decrease of injected water salinity. In the experiments of injected water salinity increasing, oil displacement efficiency and movable oil saturation gradually increase with the increase of injected water salinity. When the salinity of injected water is 2 times that of formation water, the degree of water-free recovery, global mobility and productivity index are the highest. The core evaluation parameters of water flooding experiment are productivity index and oil displacement efficiency. Productivity index is related to reservoir productivity and displacement efficiency is related to reservoir recovery. Comprehensive analysis shows that the water flooding effect is the best when the salinity of injected water is 2 times that of formation water.

In order to further study the water-sensitive damage mechanism of low-permeability sandy conglomerate reservoir, scanning electron microscope experiment, clay mineral X-ray diffraction experiment, non-linear seepage experiment and nuclear magnetic resonance experiment were conducted on specimens taken from core samples 3-6 to find out the changes of the core samples before and after water flooding. The scanning electron microscope experiment and the viscosity mineral X-ray diffraction experiment are to observe and analyze a part of the core, while the nonlinear seepage experiment and the nuclear magnetic resonance experiment are to analyze the whole core. The water-sensitive damage mechanism of low permeability sandy conglomerate reservoir in Wushi oilfields was comprehensively determined through comparative analysis of relevant experiments before and after water flooding.

2.3.1. Scanning electron microscope experiment

shows the scanning electron microscope experiment results of core samples 3-6 before water flooding. It can be seen that the space structure of the rock is good, with channels clearly visible and some pores blocked. For example, some pores and throats of core sample 4 are blocked by illite, and core sample 5 have a large number of quartz and feldspar and some illite crystals along crystal boundary. This means that there are clay minerals, quartz, feldspar and other particles in the pore throat of the original low-permeability sandy con-glomerate reservoir in Wushi region, and particle migration may happen during displacement, causing damage to the reservoir.

Fig. 4 shows the scanning electron microscope results before and after water flooding of different locations of core sample 3. The injected waters in the displacement process were formation water, water with salinity 1/2 times that of the formation water and distilled water in turn (the salinity of injected water decreases). Before water flooding, the sample had clear pore distribution, good rock space structure, fully developed pores and throats, good pore structure, complex throat connection, pores free of clog and exfoliate, clear boundary between quartz and feldspar with some illite attached. After water flooding, the rock clastic particles are poor in sorting, clay fills between particles, pores are few, some quartz enlarge secondarily, and illite, a small amount of calcite and alveolate illite and montmorillonite mixed layer between particles.

Fig. 5 shows the scanning electron microscope results of different locations of core sample 6 before and after water flooding. The injected waters in the displacement process were formation water, water with salinity 2 times that of the formation water and water with salinity 3 times that of the formation water in turn (with the salinity of injected water increasing). Before water flooding, the sample had clearly distributed pores, quartz and feldspar intricately distributed, clay minerals attached to the pores, general throat connection, and pore channels in large diameter of flake feldspar. After water flooding, the sample has poor sorting of clastic particles, less pores, some quartz secondary growth, part of the grain edge dissolved, irregular clay distributed between particles.

By comparing the scanning electron microscope images before and after water flooding, it is found that there were clay minerals and debris particles in the pore-throats of the core samples, and these particles expanded, dispersed and migrated, thus blocking the throats during the waterflooding. For the experiments with the injected water decreasing in salinity, strong hydration of clay minerals and particle migration occurred after water flooding, resulting in strong water-sensitive damage to the reservoir. For the experiments with injected waters increasing in salinity, water-sensitive damage occurred in the reservoir to a certain extent, but there was also a certain degree of dissolution. Based on the above analysis, the water-sensitive damage degree of reservoir can be reduced by increasing the salinity of injected water.

2.3.2. Clay mineral X-ray diffraction experiment

Table 4 shows the clay mineral changes of core samples 3-6 before and after water flooding. It should be noted that the samples for clay mineral X-ray diffraction analysis before and after water flooding were taken from different positions of the same core. It can be seen from Table 4, the clay mineral content after water flooding is lower than that before water flooding, indicating that clay mineral particles migrated to some extent during water flooding. In the experiments with injected water decreasing in salinity (core samples 3 and 4), the content of illite and montmorillonite mixed layer increased after water flooding, while in the experiments with injected water increasing in salinity (core samples 5 and 6), the content of illite and montmorillonite mixed layer decreased or remained the same as before water flooding. The hydration of illite and montmorillonite mixed layer is very strong. Increasing the salinity of injected water can inhibit hydration of illite and montmorillonite mixed layer.

2.3.3. Non-linear seepage experiment

For core samples 3-6, non-linear seepage experiments were conducted with injected water of different salinities. The experimental data was fitted by Darcy formula, generalized Darcy formula considering effective driving factor and pseudo-starting pressure formula. It is found that the generalized Darcy formula considering effective driving factor has the best fitting effect (Fig. 6). According to formula (4), the starting pressure gradient (Fig. 7) and effective permeability (Fig. 8) of core samples 3-6 in displacement by water of different salinities were obtained by fitting. It is found that the starting pressure gradient decreases with the increase of injected water salinity, indicating that the effect of micro-pore throat and clay mineral on fluid flow can be reduced with the increase of injected water salinity. With the increase of injected water salinity, the effective permeability increases, indicating that the increase of injected water salinity can enhance the physical properties of the reservoir.

2.3.4. Nuclear magnetic resonance experiment

In order to further study the effect of injected water of different salinities on the microscopic pores of the reservoir, nuclear magnetic resonance experiments were carried out on core samples 3-6 under different displacement conditions, as shown in Fig. 9. Micro-pores are divided into macropores, mesopores and micropores according to regional laws. Among them, the T₂ value (transverse relaxation time) of macropores is greater than 86.40 ms, the T₂ value of mesopores is 11.57-86.40 ms, and the T₂ value of micropores is less than 11.57 ms. Micropores belong to pores with immovable oil, while macropores and mesopores belong to pores with movable oil. According to Fig. 9 and Table 5, when the salinity of injected water increases, the saturation of movable fluid increases. According to the change rule of pore, the proportion of macropores is the highest when the injected water is water with salinity 2 times that of the formation water. The effect of large pore throats, which are not the majority in low permeability reservoirs, on the seepage capacity should not be underestimated. The large pore throat plays a role similar to fractures^[16].

The experimental results of non-linear seepage show that the higher the salinity of injected water, the higher the effective permeability of reservoir will be. The results of oil displacement experiments by water show that the global mobility and productivity index are the highest when the injected water is water with salinity 2 times that of the formation water. According to the experimental results of nuclear magnetic resonance, the proportion of macropores is the highest when the injected water is water with salinity 2 times that of the formation water, which further proves that the global mobility and productivity index can better characterize the seepage capacity of the fluid in the low permeability reservoir.

The clay mineral hydration and particle migration work jointly to block throats in the reservoir, making macropores turn to mesopores and mesopores to micropores. With the decrease of injected water salinity, this phenomenon becomes more and more prominent. When the injected water was water with salinity 3 times that of the formation water, movable fluid saturation was the highest, but the macropores reduced in number when the injected water was water with salinity 2 times that of the formation water. Some of the macropores were transformed into mesopores. The reason is that the rest Ca²⁺ and Mg²⁺ were more in number when the water with salinity 3 times that of the formation water was used to neutralize the negative charge in clay minerals and acted as exchangeable positive ions, they reacted with CO₃^2- in the reservoir to form CaCO₃ and MgCO₃ precipitates, plugging the macropores.

In order to verify the understanding that K⁺ and Mg²⁺ in injected water can inhibit the migration of particles which would cause blocking of throats and water-sensitive damage to reservoir in the process of waterflooding, displacement experiments with injected water types being formation water, formation water+KCl and formation water+KCl+MgCl₂ were designed. The experimental cores are sample 7, sample 8 and sample 9 respectively. The experimental results are shown in Fig. 10.

In the water flooding experiments with water of different salinities, the salinity 2 times that of the formation water was 37 000 mg/L, which is basically the same as that of the formation water+KCl and formation water+KCl+MgCl₂ (Table 2). For core sample 7, when the injected water was formation water, the ratio of displacement permeability and initial permeability was 0.926-0.992, and 0.951 on average. When the injected water was water with salinity 2 times that of the formation water, the ratio of displacement permeability and initial permeability increased little. For core sample 8, when the injected water was formation water+KCl, the ratio of displacement permeability and initial permeability was 1.00-1.59, and 1.45 on average. For core sample 9, the injected water was formation water with both KCl and MgCl₂, the ratio of displacement permeability and initial permeability was 1.00-1.68, 1.56 on average. Therefore, increasing K⁺ and Mg²⁺ can effectively enhance the seepage capacity of the reservoir after eliminating the influence of salinity.

According to the whole rock data and clay mineral X-ray diffraction data, the low-permeability sandy conglomerate reservoir in Wushi region has a content of clay mineral of 4.6%-13.0%, on average 7.3%. The relative contents of illite and montmorillonite mixed layer, illite, kaolinite and chlorite are 22.4%, 46.4%, 27.4% and 3.8% respectively, and the interlayer ratio of illite and montmorillonite mixed layer is 10%-25%. According to a large number of water sensitivity experiments, the water sensitivity damage degree of the reservoirs is 43.5%-93.0% and generally greater than 50%, indicating the reservoirs are medium to strong-strong water-sensitive. Although not high in clay mineral content, the low-permeability sandy conglomerate reservoir in the Wushi region contains a certain amount of expansive clay minerals (such as illite and montmorillonite mixed layer). Besides,

according to the scanning electron microscope images before displacement, there is a certain amount of quartz, feldspar, clay minerals and other clastic particles in the reservoir. Based on water-sensitive damage mechanisms and experimental analysis results of the reservoir, it is concluded that the water-sensitive damage mechanisms of the low permeability sandy conglomerate reservoir in Wushi region include clay mineral hydration and particle migration.

According to the results of flooding experiments with water of different salinities and related matching experiments, increasing injection water salinity can lower water sensitivity damage to the reservoir and the mechanisms include: (1) Increasing the salinity of injection water can keep a semi-permeable membrane around the clay particles in osmotic balance state and reduce the osmotic hydration of clay minerals. (2) When the salinity of injection water is increased, positive ions in injection water also increase, which can adsorb and neutralize negative charges on the clay mineral surface and between crystal layers, and thus reduce the thickness of the clay mineral diffusion double layer and Zeta potential and inhibit the negative charge and adsorption of hydration positive ions, realizing the goal of preventing expansion and stabilizing clay minerals

In order to improve the effect of waterflooding development and reduce water sensitivity damage to the reservoir, particle migration inhibition experiments were carried out. The result shows that increasing K⁺ and Mg²⁺can effectively enhance reservoir seepage capability. The mechanisms include: (1) When exchangeable positive ions are Mg²⁺, K⁺, H⁺, and Ca²⁺, etc., the attraction between clay mineral chips increases than that of the exchangeable positive ion of Na+, so the hydration membrane created is thinner and water molecules are in regular directional arrangement. (2) The diameter of K⁺ is 0.26 nm, which matches the hexagonal space of 0.28 nm in diameter made by 6 oxygen atoms on the surface of clay mineral, making it easier to enter this space and not easy to release, effectively neutralizing the negative charge on the surface.

In addition, when the content of bivalent positive ions (especially Ca²⁺ and Mg²⁺) is high, the rest Ca²⁺ and Mg²⁺ are more after bivalent positive ions are used to neutralize the negative charge in clay minerals and act as positive ions, which would react with CO₃^2- in the reservoir to generate CaCO₃ and MgCO₃ precipitate, plugging the large pore throats. The small number of large pore throats in low permeability reservoirs cannot be underestimated, as they function similar to fractures. This is the reason why the core seepage capacity when the injected water is water with salinity 2 times that of the formation water is better than that when the injected water is water with salinity 3 times that of the formation water.

Based on the above analysis of the impact of water sensitivity damage on reservoir productivity, combined with the production equation of the oil-water two phases flow well considering low-speed non-Darcy seepage and reservoir stress sensitivity, the effects of reservoir pressure maintaining level and changes of oil and water relative permeabilities on reservoir productivity were quantitatively analyzed respectively.

Fig.11 shows the relationship between productivity of different layers and water content when the bottom hole flowing pressure is 30% of the original formation pressure. The productivity of the production well decreases with the decrease of reservoir pressure maintaining level. Therefore, when the reservoir pressure is kept higher for higher productivity of the low permeability sandy conglomerate reservoir.

According to the relative permeability curve of core sample 1, the oil production equation was used to calculate the relationship between productivity and water content with injected water of different salinities when the bottom hole flowing pressure is 30% and the reservoir pressure is maintained at 100%, the results are shown in Fig. 12. It can be seen that reservoir water sensitivity damage affects the reservoir productivity through affecting the oil-water two-phase seepage capacity of the reservoir near the production well.

Water sensitive damage of reservoir affects reservoir productivity through affecting reservoir pressure maintaining level and oil and water two-phase seepage capacity of the reservoir near the production well after breakthrough. Therefore, selecting suitable injection water source and controlling injection water quality strictly are very important for ensuring continuous replenishment of formation energy and keeping reservoir fluid seepage capacity unaffected during the development of low permeability sandy conglomerate reservoirs.

Seawater, nanofiltration seawater or shallow formation water is often selected as injection water source for offshore oilfields. All these waters are fairly high in salinity. Moreover, the low permeability sandy conglomerate reservoir in Wushi region is not high in content of expansive clay minerals. Therefore, reservoir water sensitivity damage is mainly caused by osmotic hydration and particle migration during the development of Wushi oilfields. Adding K⁺ and Mg²⁺ into the injected water can effectively inhibit the osmotic hydration of clay minerals and reduce reservoir damage caused by particle migration.

Available injection water sources for the low-permeability sandy conglomerate reservoir in Wushi oilfield include seawater, nanofiltration seawater and the formation water of shallow Paleogene Oligocene Weizhou Formation, with a salinity of 33 583 mg/L, 23 593 mg/L and 48 631 mg/L, respectively. Fig.13 shows the results of core dynamic compatibility experiments with the three types of injected water. The cores used in the experiments were representative core samples of the low-permeability sandy conglomerate reservoir in Wushi region. The shallow formation water of Weizhou Formation has the highest K⁺ content, and the proportion of bivalent positive ions (such as Mg²⁺ and Ca²⁺) in the total amount of positive ions (8.6%) between the nanofiltration seawater and seawater with the highest salinity. Therefore, the shallow formation water in Weizhou Formation has the best dynamic compatibility with the core. In contrast, the seawater has the worst dynamic compatibility with the core, the main reason is that bivalent positive ions in the seawater account for 13.1% of total positive ions, the Ca²⁺ and Mg²⁺ left after neutralizing the negative charge in clay minerals and acting as exchanging positive ions are large in quantity, which would react with CO₃^2- in the reservoir to generate CaCO₃ and MgCO₃ precipitates, plugging large pore throats and making reservoir physical properties turn worse. Moreover, K⁺ content is not high in the seawater. The nanofiltration seawater has a proportion of bivalent positive ions of only 0.4%, and K⁺ content similar to that of seawater. Although CaCO₃ and MgCO₃ precipitates are not produced during injection of the nanofiltration seawater, with no positive ions (such as K⁺ and Mg²⁺) that can improve water flooding effect, this water has compatibility with the core between that of shallow formation water of Weizhou Formation and the seawater. In conclusion, the shallow formation water of Weizhou Formation is most suitable for the low permeability conglomerate reservoir in Wushi region.

The global mobility theory was used to evaluate the experimental results of flooding with water of different salinities. The results show that the seepage capacity of reservoir is the strongest when the injected water is water with salinity 2 times that of formation water. The results of scanning electron microscopy, X diffraction of clay minerals, nonlinear seepage and nuclear magnetic resonance experiment and particle migration inhibition experiment before and after flooding with water of different salinities were compared and analyzed, which show that the water-sensitive damage mechanisms of the reservoirs in Wushi region include hydration of clay minerals and particle migration. By increasing the content of cations (especially K⁺ and Mg²⁺) in the injected water, the seepage capatity of the reservoir can be effectively enhanced.

The reservoir water sensitivity damage affects the reservoir productivity through the drop of formation pressure by lowering injection capacity of injection well and energy transfer speed and lowering oil-water two-phase seepage capacity of the reservoir near the production well after water breakthrough in the production well. Based on the generalized Darcy formula considering the effective driving factor and mass conservation law, the production equation of the oil- water two-phase flow well considering low-speed non-Darcy seepage and reservoir stress sensitivity has been established, which can be used to quantitatively analyze the effect of reservoir water-sensitive damage on productivity. Based on the results of core dynamic compatibility experiments of seawater, nanofiltration seawater and the shallow formation water of Weizhou Formation, and the analysis results of water sensitive damage and enhanced water flooding mechanisms, the shallow formation water of Weizhou Formation is the injected water source most suitable for the low-permeability sandy conglomerate reservoir in Wushi region.

Nomenclature

a—control factor of reservoir pressure maintaining level;

A—cross-section area of the core, m²;

b₀—coefficient associated with gravity, 10^-12 m³/s;

B_o, B_w—oil and water volume coefficients, m³/m³;

g—acceleration of gravity, m/s²;

G—starting pressure gradient, MPa/m;

h—effective thickness of reservoir, m;

I_p—productivity index, 10^-3 μm²/(mPa·s);

K—effective permeability under current reservoir pressure, 10^–3 μm²;

K_abs—absolute permeability, 10^-3 μm²;

K_init—effective permeability under initial formation pressure, 10^-3 μm²;

K_nw^*—effective permeability of the nonwetting phase at 1-S_wf, 10^-3 μm²;

K_r—relative permeability, f;

K_ro, K_rw—relative permeability of oil and water, f;

K_w^*—effective permeability of the water or the wetting phase at S_wf, 10^-3 μm²;

L—core length, m;

m(ap_init)—pseudo-pressure of oil-water two phases at formation pressure, 10¹² kg/(m³·s);

m(p_wf)—pseudo-pressure of oil-water two phases at well bottom flowing pressure, 10¹² kg/(m³·s);

m(r)—pseudo-pressure of oil-water two phases at r, 10¹² kg/(m³·s);

M_e^*—global mobility of the two phases at S_wf, 10^-3 μm²/(mPa·s);

M_nw^*—mobility of the nonwetting phase at 1-S_wf, 10^-3 μm²/(mPa·s);

M_w^*—mobility of the water or the wetting phase at S_wf, 10^-3 μm²/(mPa·s);

p—current reservoir pressure, MPa;

\(\nabla p\)—pressure gradient, MPa/m;

Δp—differential pressure of production, MPa;

p_c^*—capillary pressure at S_wf, Pa

p_init—initial formation pressure , MPa;

p_m—atmospheric pressure, MPa;

p_wf—bottom hole flowing pressure, MPa;

q_o, q_w—oil production and water production, 10^-3 m³/s;

q_oL—oil production of displacement experiment under at S_wf, 10^-12 m³/s;

r—distance from the wellbore, m;

r_e—drainage radius, m;

r_w—wellbore radius, m;

R—recovery of core flooding experiment, f;

S_or—residual oil saturation, f;

S_wf—water (or the wetting phase) saturation at a certain moment, f;

S_wi—initial water (or the wetting phase) saturation, f;

ν—fluid velocity vector, 10^-6 m/s;

α_k—permeability variation coefficient, MPa^-1, indicating reservoir stress sensitivity;

α_k^*—apparent permeability variation coefficient, 10^-12 (m³·s)/kg;

δ—effective driving factor;

δ_o, δ_w—effective driving factors of oil phase and water phase;

μ—fluid viscosity, mPa·s;

μ_nw—viscosity of the nonwetting phase, mPa·s;

μ_o—viscosity of oil, mPa·s;

μ_w—viscosity of water (or the wetting phase), mPa·s;

ξ(r)—pseudo-pressure transformation parameter of oil-water two-phase seepage at r, 10¹² kg/(m³·s);

ξ(r_e)—pseudo-pressure transformation parameter of oil-water two-phase seepage at r_e, 10¹² kg/(m³·s);

ξ(r_w)—pseudo-pressure transformation parameter of oil-water two-phase seepage at r_w, 10¹² kg/(m³·s);

ξ_init—pseudo-pressure transformation parameter of oil-water two-phase seepage at formation pressure, 10¹² kg/(m³·s);

ξ_wf—pseudo-pressure transformation parameter of oil-water two-phase seepage at well bottom flowing pressure, 10¹² kg/(m³·s);

ρ_o, ρ_w—density of oil and water, kg/m³;

Δρ—density difference between the wetting and nonwetting phases, kg/m³.