Microscopic mechanism of water flooding in tight reservoirs

Tight oil, whose strategic significance has become increasingly prominent due to its effective development in domestic China, is widely distributed (Jia et al., 2012; Zou et al., 2012; Wu et al., 2016; Zhang et al., 2016; 2018, Yang et al., 2013; Guo et al., 2016; Wu et al., 2015; and Dejam et al., 2018). Tight oil reservoirs have complex pore structures, poor physical characteristics (Yao et al., 2013; Bai et al., 2013; Dejam et al., 2017a; 2017b; Zhao et al., 2015; Ghanizadeh et al., 2015; and Yang et al., 2012), small pores in crude oil occurrence, and large fluid seepage resistance, leading to difficulty in establishing an effective displacement system (Li et al., 2016; 2006; Zhao et al., 2015; Ghanizadeh et al., 2015; Wang et al., 2015; 2009; Yang et al., 2014; 2007; 2009; Yu et al., 2012; Loucks et al., 2009; Law and Curtis, 2002; Mullen, 2010; Zhong et al., 2012; Yao et al., 2013; and Quan et al., 2011). A microseepage mechanism of oil and water is the theoretical basis for water flooding recovery in oilfields. Understanding the micromobility laws and driving mechanisms of water-flooding in tight reservoirs is of great significance for the rational and effective development of reservoirs. Some scholars have carried out a lot of research on the micromechanism of water-flooding in low-permeability reservoirs. Xueling et al. (2018) studied liquid flow in nano- or microsized circular tubes. Wei et al. (2014) analyzed influencing factors on water-oil displacement. Suleimanov et al. (2017) studied fluid properties near to the phase transition point. Datta et al. (2014) studied the fluid flow mechanism from a three-dimensional porous medium. De Paoli et al. (2016) and Chen and Yan (2015) studied the flow behavior of fluids in heterogeneous porous media. Qu et al. (2018) believed that the static imbibition recovery factor is closely related to the reciprocal of the irreducible number. The optimal value of the reciprocal of irreducible is about 1. Jianming et al. (2019) studied the impact of the energy supplement from the unreturned fluid during volume fracturing in horizontal wells on the development effect. Qu et al. (2017) explored the influence of fractures on the seepage characteristics of dense rock water flooding. Wei et al. (2016) studied the effect of reservoir properties on imbibition recovery. Gu et al. (2017) revealed the microscopic influence mechanism of the permeability of tight reservoirs on the efficiency of oil imbibition recovery. Pan et al. (2016) adopted the combination of nuclear magnetic resonance test and displacement test to study the displacement process of medium injection and crude oil under different injection conditions and clarified the producing rules of crude oil in different pores under water injection and CO₂ injection. Qualitative visual observation and theoretical analysis of the microscopic mechanism of water flooding in low permeability reservoirs (Huang 1999; Guo et al., 1990; and Wang and sun, 2010) with different wettability values have been done. However, there are few research studies on the microscopic recovery mechanism of water flooding in tight reservoirs with no report on the quantitative characterization of oil recovery in the microscopic mechanism, and it is necessary to explore the quantitative characterization of different recovery mechanisms in water flooding.

Targeting at tight cores with different wettability values, this study combined nuclear magnetic resonance technology with water flooding experiments under different pressures to establish a quantitative analysis approach for oil production from different microscopic actions and quantitatively analyze the microscopic mechanism of water flooding. The production mechanism of hydrophilic cores and neutral wetting cores under low displacement pressure is studied quantitatively (flooding, imbibition, and denudation), and the main mechanism is clarified. The mechanism of improving water displacement recovery of hydrophilic core and neutral wetting core after increasing the displacement pressure is studied quantitatively (production from small pore throats, boundary layers, and large pores controlled by small throats), and the similarities and differences between them are analyzed. Figure 1 shows the general sketch of the physical model.

The steps of this work are as follows: First, the experimental core and fluid data are presented. Then, cores are saturated with kerosene, and the T₂ spectrum is measured. Next, cores are redried and saturated with heavy water. Afterward, cores of saturated water are flooded by kerosene to establish irreducible water, and the T₂ spectrum is measured. Finally, the water flooding experiment under the pressure of 0.32 MPa and 5.12 MPa is carried out for each core, and the nuclear magnetic resonance T₂ spectrum is measured separately.

In this study, eight tight oil reservoir cores from the Ordos Basin were selected for water flooding experiments under low displacement pressure and high displacement pressure, and nuclear magnetic resonance detection was performed on each state of the cores. The porosity of the eight cores ranges from 5.4% to 12.5% with an average of 9.9%, and the gas measured permeability range is (0.035–0.21) mD with an average of 0.12 mD (see Table I). The experimental water is standard brine with a mineralization degree of 50 g/l (NaCl: 22.5 g/l; KCl: 22.5 g/l; CaCl₂: 3.8 g/l; and MgCl₂: 1.2 g/l. Heavy water does not contain H core, and so no nuclear magnetic signal is generated during nuclear magnetic resonance testing). The experimental oil is aviation kerosene with similar properties to the crude oil of the reservoir, with the physical parameters such as viscosity being basically the same as those of the actual crude oil, 2.65 cP at 25 °C. The gas medium used in the experiment is nitrogen.

The application of nuclear magnetic resonance in the petroleum industry is to make full use of the relaxation of NMR to detect and analyze the occurrence state and properties of oil and water in rocks. The T₂ relaxation time of fluids in rocks can be expressed by the following formula:

In formula (1), $1T2S$ is the relaxation contribution of fluid from the surface of rock particles, $1T2B$ is the relaxation contribution of bulk fluid itself, and $1T2D$ is the relaxation contribution from self-diffusion of fluid molecules. When the static magnetic field is uniform, $1T2D$ is close to zero and negligible. The relaxation of bulk fluids is much weaker than that of rock surface fluids, which can be neglected in petroleum nuclear magnetic research and application. For surface fluid relaxation in the presence of a single channel, the following expression can be used:

The rock pore is composed of different sizes of pores. Each size of pore has its own characteristic relaxation time T_2i under different occurrence conditions. Therefore, there are many exponential decay processes in rock. The total relaxation is the superposition of these relaxations,

In formula (3), A_i is the proportion of component i and T_2i is the relaxation time of component i, which is related to the specific surface S/V or pore size of the rock. When the fluid is in close contact with the solid surface in large pores or fluid in small pores, the nuclear magnetic transverse relaxation time of the fluid is small. Conversely, when the fluid is not in close contact with the pore surface in large pores, the nuclear magnetic transverse relaxation time of the fluid is large.

Experimental steps are as follows: ① core marking, oil washing (Dean Stark extraction), drying and weighing dry weight, and measuring core length and diameter; ② achieving gas measurement porosity and gas measured permeability (steady state gas-perm; the gas medium is nitrogen); ③ vacuum pressurized and saturated with kerosene, calculating kerosene porosity (using the core weight difference between saturated kerosene and dry weight to calculate core pore volume and using core length and diameter to calculate core volume); ④ measuring the nuclear magnetic resonance T₂ spectrum under saturated oil state (using the Reccore-04 core NMR analyzer for T₂ spectrum detection); ⑤ core redrying; ⑥ vacuum pressurized and saturated with heavy water at a mineralization of 50 000 mg/l, calculating the porosity (water measured porosity) by using the difference between the wet weight and the dry weight of the core; ⑦ loading the core into the displacement stream, selecting an appropriate displacement pressure, flooding the core of the saturated water by kerosene, and establishing the state of the saturated oil of the core under irreducible water state (the displacement experiment is completed by using the SL-2012 nonlinear test system; the displacement multiple is about 10 PV, measuring the displaced water volume and the core weight); ⑧ measuring the T₂ spectrum of NMR under the condition of the saturated oil under irreducible water state; and ⑨ based on core parameters including physical properties, each core is subject to the water flooding experiment under the pressure of 0.32 MPa and 5.12 MPa, separately, and each displacement pressure is driven until no more oil is produced (the displacement volume is about 5 PV). Measure the amount of produced oil, measure the core weights, and test the NMR T₂ spectrum, separately.

Figure 2 shows the NMR T₂ spectra of No. 5 and No. 8 specimen under the conditions of saturated oil, saturated oil under irreducible water, and 0.32 MPa water flooding. The wettability of No. 5 was hydrophilic (wetting angle 28.5°, Fig. 3), and the wettability of No. 8 was neutral (wetting angle 96.6°). It can be seen from this figure that there are significant discrepancies in the T₂ spectra of the different states of the hydrophilic and neutral cores. The right peaks of No. 8 saturated oil state and the saturated oil under irreducible water state T₂ spectrum (corresponding to the macropores) are basically coincident, indicating that the crude oil in partial large pores of the neutral cores has been fully filled, while partial large pores of the hydrophilic cores have not been fully filled. After 0.32 MPa displacement, the right peak of No. 8 sample decreased greatly and the decrease in gradient of the left peak was lower, indicating that the producing degree of the crude oil in partial large pores was relatively high, while the left and right peaks of No. 5 sample decreased by an equivalent amplitude. According to the principle of nuclear magnetic resonance, small relaxation time corresponds to the fluid in small throats or at the surface of the large pore throats of the cores. Further comparative analysis shows that the produced crude oil in the pore space below 10 ms in No. 8 specimen was little (corresponding to Fig. 2, right panel), indicating that the crude oil in small throats or on the surface of large pore throats in No. 8 specimen was rarely produced. Since the displacement pressure is only 0.32 MPa, which is relatively low, the absorption at the surface of the pores in the tight reservoir is very strong, suggesting that such displacement pressure cannot drive out the oil on the surface of the large pores of the cores; on the other hand, the tight reservoir is featured by mixed wettability, with some of the pore throats showing hydrophilic characteristic and some of the pore throats showing lipophilic characteristic. For the lipophilic small throats, the small displacement pressure cannot overcome the capillary resistance and drive out the oil in the small throat, and so the oil in the small throats can only be produced by the imbibition of the hydrophilic capillary portion; therefore, the mechanism of the crude oil in the pore space below 10 ms in No. 8 specimen is imbibition.

The volume of crude oil produced from the pore space below 10 ms in No. 5 specimen is relatively high (corresponding to Fig. 2, left panel). Due to the hydrophilicity of No. 5 specimen, the small displacement pressure can also drive out the oil on the surface of the large pores through denudation action. Therefore, the crude oil in the pore space of No. 5 specimen below 10 ms is the result of the combined mechanism of denudation and imbibition. Figure 2 shows that the T₂ spectrum amplitude of the irreducible water under saturated oil state below 10 ms in No. 5 specimen is relatively large, indicating that the effect of oil recovery through denudation is relatively strong.

Figure 4 shows the NMR T₂ spectra of No. 4 and No. 7 under conditions of saturated oil state, saturated oil under irreducible water state, and states after flooding at different displacement pressures. The wettability of No. 4 was hydrophilic (wetting angle 21.6°, Fig. 3), and the wettability of No. 7 was neutral (wetting angle 88.9°). It can be seen from this figure that compared with the T₂ spectrum after 0.32 MPa displacement, the T₂ spectrum variations after the 5.12 MPa displacement of both the hydrophilic and neutral cores are significantly different. The relatively larger relaxation time part of the two cores has a certain degree of reduction, indicating that there is a certain volume of residual oil in the large pores that has been driven out; the residual oil of this type can be divided into two categories: one is the crude oil in large pores controlled by small throats (as this part of throats is very small and 0.32 MPa is not enough to start the displacement, the crude oil in which cannot be produced until the displacement pressure value has increased to a certain extent) and the other type is the oil droplets controlled by large pore throats (due to the Jiamin effect, 0.32 MPa is not enough to overcome the seepage resistance caused by the Jiamin effect, resulting in the residual oil being distributed in the large pores after 0.32 MPa displacement until the displacement pressure is increased to a certain extent before being produced). The relatively small relaxation time part of the two cores also has a certain degree of reduction, but the neutral wetting cores have a greater reduction. For the studied tight cores in this essay, the displacement at 0.32 MPa for 5 PV takes a long time (all over 10 days) and the core static imbibition experiment shows that the core imbibition mainly occurs in the initial stage of imbibition, and the oil recovery from imbibition becomes little after 72 h. Therefore, the oil recovery effect of imbibition has been fully utilized in the 0.32 MPa displacement experiment. When the displacement pressure is increased from 0.32 MPa to 5.12 MPa, the effect of imbibition recovery is very small and the improvement of oil recovery depends mainly on other functions. After increasing the displacement pressure, some part of the oil controlled in the small pore throats can be produced; on the other hand, part of the boundary layer oil film in the neutral wetting cores became thinner, which increases the oil recovery ratio (the influence of this mechanism is minimal as the portion of the hydrophilic core boundary layer oil is little). After increasing the displacement pressure, the common point for the oil recovery mechanism through water flooding of both the hydrophilic and the neutral wetting cores reachedthat both types started the oil controlled by the small throats and the residual oil droplets trapped in the large pores; the discrepancy is that the oil recovery ratio of the neutral wetting cores improved due to the thinning of the oil film in the boundary layer, while this kind of oil recovery mechanism influences the hydrophilic cores little.

The results of water flooding of 8 tight cores were compared according to wettability (Figs. 5–7). The relatively large pores in saturated oil under irreducible water state of cores with a different wettability have a relatively high oil content (Fig. 5). The neutral and weak hydrophilic cores are higher than the hydrophilic ones, indicating that the crude oil in the large pores of the neutral and weak hydrophilic cores has been fully injected, while the filling of the crude oil in the large pores of the hydrophilic cores is not completely full. After 0.32 MPa displacement, the recovery ratios of the crude oil in the larger pores of cores with a different wettability are higher; the hydrophilic and weak hydrophilic cores have higher production in the pore space below 10 ms (6 cores range from 7.5% to 17.7% with an average of 11.8%), while the producing degree of such pores in the neutral wetted cores is relatively low (2 cores range from 0.5% to 3.8% with an average of 2.1%) (Fig. 6, Table I); this indicates that the hydrophilic and weakly hydrophilic cores have stronger denudation and imbibition under low displacement pressure, while the neutral wetting cores have weaker denudation and imbibition under low displacement pressure. From Fig. 6, it can be further seen that the relative recovery of oil below 10 ms in water-wet cores can reach 41.8%, while that of neutral cores is only about 7.5%. This indicates that water-wet reservoirs more easily play the role of imbibition in oil recovery. After 5.12 MPa displacement, the crude oil recovery ratio of pores in different wettability cores in different T₂ intervals all increased to some extent; the recovery ratios of crude oil in the pore space below 10 ms in neutral wetting cores are significantly higher than that in hydrophilic and weakly hydrophilic cores. It shows that after increasing the displacement pressure, part of the boundary layer oil film has been activated in the neutral wetting cores, which has increased the oil recovery.

The microscopic mechanisms of water flooding in 8 tight cores have been quantitatively calculated (Fig. 8, Table I). The oil recovery mechanism of hydrophilic cores mainly includes flooding, imbibition, and denudation, producing oil controlled by small pore throats, and by large pores controlled by small pore throats. The percentage of produced oil is 14.7%, 11.8%, 2.2%, and 4.3%, respectively, with flooding, imbibition, and denudation being dominant. The oil recovery mechanism of neutral wetting cores, mainly including flooding, imbibition, and denudation, starts producing oil controlled by small pore throats and boundary layers, and large pores controlled by small pore throats. The recovery ratios are 24.5%, 2.1%, 5.8%, and 3.3%, respectively, with flooding production being dominant.

In order to improve the development effect of tight reservoirs in the Ordos Basin of China, field tests, such as synchronous injection and production of horizontal wells, asynchronous injection and production of horizontal wells, and asynchronous injection and production of staggered fracture distribution, were carried out in the oilfield. The energy supplement mode of cyclic water injection and wheel-injection production in the mixed injection-production pattern of directional wells and horizontal wells is attempted. In typical tight oil areas of China, tests of imbibition production and supplementary energy by fracturing fluid in “artificial reservoirs” have also been carried out. The purpose of these technologies is to increase formation energy, reduce displacement distance, and give full play to displacement and imbibition. This is consistent with the technical thinking and development mechanism proposed in this study. Moreover, the method of this study can quantitatively give oil production of each mechanism and provide support for development decision-making of oilfields.

Nevertheless, there are some limitations in this study. In this study, core experiments cannot take into account the impact of large natural fractures in the actual formation. In addition, the wettability of actual reservoir is more complex, and this study can only carry out experiments on specific wettability cores.

Combining water flooding experiments under different pressures with nuclear magnetic resonance, a quantitative analysis approach for oil recovery from different microscopic actions has been established and the microscopic mechanisms of water flooding in tight oil cores have been quantitatively studied. The production mechanism of hydrophilic cores and neutral wetting cores under low displacement pressure is analyzed quantitatively, and the main mechanism is clarified: The oil recovery mechanism of hydrophilic cores mainly includes flooding, imbibition, denudation, and other mechanisms (the oil recovery ratio in hydrophilic cores with flooding and imbibition and denudation was 14.7% and 11.8%, respectively, which are the main mechanisms of water flooding); the mechanism of neutral wetting cores recovery mainly includes flooding and imbibition (the oil recovery ratio in neutral wetting cores with flooding is 24.5%, which is the main oil recovery mechanism of water flooding). The mechanism of improving water displacement recovery of hydrophilic core and neutral wetting core after increasing the displacement pressure is analyzed quantitatively, and the similarities and differences between them are analyzed. Both the hydrophilic cores and the neutral wetting cores have a certain increase in oil recovery; the common point of the two producing mechanisms is that both of them activated the oil controlled small throat control and the residual oil droplets trapped in the large pores; the discrepancy is that the oil film of the neutral wetting core boundary layer became thinner and the oil recovery degree was improved, while these kinds of the oil recovery mechanisms have less effect on the hydrophilic cores (the oil produced by the hydrophilic and neutral wetting cores increased by 6.4% and 9.1%, respectively, and the neutral wetting cores was slightly higher, indicating that the increased displacement pressure might convert some crude oil in the boundary layer into producing volume).

We gratefully acknowledge financial support from the National Science and Technology Major Project (Grant Nos. 2017ZX05013-001, 2017ZX05069003, and 2017ZX05049005-004) and the Ministry of Science and Technology of PetroChina (Grant Nos. 2017E-1514 and 2018E-11-05).