Influence of Oil Viscosity on Alkaline Flooding for Enhanced Heavy Oil Recovery

Oil viscosity was studied as an important factor for alkaline flooding based on the mechanism of “water drops” flow. Alkaline flooding for two oil samples with different viscosities but similar acid numbers was compared. Besides, series flooding tests for the same oil sample were conducted at different temperatures and permeabilities. The results of flooding tests indicated that a high tertiary oil recovery could be achieved only in the low-permeability (approximately 500mD) sandpacks for the low-viscosity heavy oil (Zhuangxi, 390mPa⋅s); however, the high-viscosity heavy oil (Chenzhuang, 3450mPa⋅s) performed well in both the lowand medium-permeability (approximately 1000mD) sandpacks. In addition, the results of flooding tests for the same oil at different temperatures also indicated that the oil viscosity put a similar effect on alkaline flooding. Therefore, oil with a high-viscosity is favorable for alkaline flooding. The microscopic flooding test indicated that the water drops produced during alkaline flooding for oils with different viscosities differed significantly in their sizes, which might influence the flow behaviors and therefore the sweep efficiencies of alkaline fluids.This study provides an evidence for the feasibility of the development of high-viscosity heavy oil using alkaline flooding.


Introduction
Thermal methods are the primary ways to develop heavy oil reservoirs.However, severe heat losses reduce the effectiveness of thermal processes in deep or thin heavy oil reservoirs and increase the production cost.Inexpensive alkaline reagents can react with the organic acids in heavy oil and form massive amounts of surfactant in situ at the oil/water interface, by which the interface tension (IFT) can be reduced greatly [1][2][3][4][5].In recent years, a new mechanism proposed by Ding et al. [6] has attracted the attention of researchers.He described the interfacial reaction between the oil phase and the alkaline solution in detail and proposed two stages related to the reaction during the alkaline flooding process.The first stage was the occurrence of water columns that resulted from the penetration of the alkaline solution into the crude oil.The second stage was the division of these water columns into small discontinuous water droplets due to the nonuniform enrichment of surfactants at the oil/water interface.The viscous fingering effect was significantly reduced by the presence of water drops inside the oil phase.In contrast, the water in oil (W/O) emulsion was simply a byproduct of alkaline liquid penetration rather than the basic mechanism of enhanced oil recovery (EOR) during alkaline flooding.They also observed differences of water breakthrough and alkaline breakthrough.In water flooding, after water breaks through, several connected water channels are created diagonally and little oil can be recovered by continued water flooding.Consequently, most of the oil is bypassed, due to the viscous fingering caused by the adverse mobility ratio between the oil and the water.In alkaline flooding, after alkaline breaks through, it can be seen that relatively uniform degree oil saturation is distributed over the entire model.Oil is subsequently displaced in the form of the "water drop" with little viscous fingering.Therefore, it is the "water drop" mechanism that reduces the mobility of water phase and diverts the injected alkaline solution to the unswept region of micromodel to improve the sweep efficiency.
The mechanism of "water drops" flow allowed the successful interpretation to some phenomena in alkaline flooding.
Many factors influence the performance of alkaline flooding.Arhuoma et al. [7] investigated the influence of alkaline concentration on the alkaline flooding for Alberta heavy oil through sandpack flooding tests using NaOH as the alkaline reagent.Their results indicated that the recovery initially increased with the increase of alkaline concentration; afterwards, the variation became very slight or even decreased, which suggests that an optimum concentration exists for enhancing the oil recovery.Almalik et al. [8] have studied the influence of the alkaline type and the injection mode on oil recovery.They observed that NaOH performed better than Na 3 PO 4 ⋅12H 2 O and KOH under identical conditions, and the continuous alkaline flooding yielded a higher recovery than when slugs were used.Previous research has been primarily focused on these dynamic injection conditions.However, a few static reservoir conditions have been also studied in some reports.Chiwetelu et al. [9] and Trujillo [10] have investigated the influence of reservoir temperature on alkaline flooding, the results showed that about 12% more oil was recovered at 65 ∘ C than at 25 ∘ C, and the IFT values were significantly greater at higher temperatures which was considered as the main reason for the change regulation.
The static conditions of heavy oil reservoirs are complicated and different with each other, which can also influence the effectiveness of alkaline flooding significantly.Among these static conditions, oil viscosity may exhibit the largest variation range in heavy oil reservoirs.However, the previous study about this aspect is insufficient to satisfy the requirement to design a good EOR plan.Therefore, the influence of oil viscosity on alkaline flooding in heavy oil reservoirs is investigated here by sandpack flooding experiments.

Experimental
2.1.Fluids and Chemicals.Oil samples were collected from the heavy oil reservoirs of Zhuangxi, Chenzhuang, Xia-8 and Binnan in Shengli Oilfield; their viscosities were measured by a rotary viscometer, and the viscosity-temperature properties are shown in Figure 1.The acid-number of these four oil samples were measured by potentiometric titration method, and the results are listed in Table 1.All the aqueous phase was brine that contained 0.5 wt% NaCl.Chemicals used in this study, such as NaOH and NaCl, were all analytical-grade reagents supplied by Sinopharm.

IFT Measurements.
The dynamic IFT values between heavy oil and alkaline solutions were measured at 50 ∘ C using an American Texas-500 spinning drop tensiometer according to the following: where  is the interfacial tension (mN/m),   is the density of the water phase (g/cm 3 ),   is the density of the oil phase (g/cm 3 ),  is the rotational velocity (rpm),  is the measured drop width (mm),  is the length of the oil drop (mm), and  is the refractive index of the water phase.

Microscopic Flooding Tests.
The glass-etched micromodel was used to investigate the displacement mechanisms of alkaline flooding.The micromodel was made by the two same glass plates which were impacted tightly, and the pore throats were etched on the interface by special corrosive substances.The procedure of the microscopic flooding test was described as follows: the micromodel was first saturated with brine after being vacuumed, and it was subsequently displaced by heavy oil until no more brine was produced.The alkaline solution was then injected at a constant flow rate of 0.003 mL/min.The flooding tests can be visualized using a video recorder and camera apparatus.

Sandpack Flooding Tests.
The sandpacks used in this study were 30 cm in length and 2.45 cm in diameter.They were wet-packed as follows: first, fresh quartz sand with 100-200 and 80-100 mesh sizes was blended at a weight ratio of 3 : 1.The sandpack was positioned vertically, and the sand was then added into the sandpack filled with brine water in several increments.In each step, the sand in the sandpack was shaken slightly after being added.The water surface was kept above the sand surface to avoid the intrusion of air.The sandpack displacement was conducted at 50 ∘ C using the following procedure: first, the sandpack was saturated with brine solution, and then the permeability was measured and the porosity was calculated.Afterward, the sandpack was subsequently saturated with the heavy oil until no more water was produced (the water cut was less than approximately 1 vol%).After the oil injection, water flooding was conducted until the oil cut was less than 1 vol%, and then a chemical slug of 0.3 pore volume (PV) was injected.The injection of the chemical slug was followed by an extended period of water flooding until the oil production became negligible (oil cut < 1 vol%).The injection rate of the brine solution and the chemical slug was set at 0.5 mL/min.The equipment for the sandpack flooding is shown in Figure 2.

Performance of Alkaline Flooding for Oils with Different
Viscosities.To investigate the influence of oil viscosity on the performance of alkaline flooding, alkaline flooding for two oil samples with different viscosities, and similar acid numbers were compared.Then, series flooding tests for the same oil sample were conducted at different temperatures.
Zhuangxi and Chenzhuang heavy oil were selected for their similar acid numbers and different viscosities, as shown in Figure 1 and Table 1.18 alkaline flooding tests (Runs 1 ∼ 18) were conducted in sandpacks with three levels of permeability: low-permeability (approximately 500 mD), medium-permeability (approximately 1000 mD), and highpermeability (approximately 2000 mD).Three alkaline concentrations were investigated at each permeability level.The parameters of the sandpacks, the chemical formulas, and the flood results are summarized in Table 2.
The alkaline flooding performances of these two oil samples were significantly different, as shown in Figures 3 and 4. The correspondingly dynamic recovery with the production is shown in Figure 5.For the Zhuangxi heavy oil with a relative low-viscosity, the incremental oil recovery was greater than 20% when the sandpack permeability was approximately 500 mD; however, it declined significantly as the permeability increased, and the tertiary oil recovery was only 5 ∼ 10% in the high-permeability sandpacks.For the Chenzhuang heavy oil, which exhibits a higher  viscosity, the tertiary oil recovery was as high as 20% at both low-and medium-permeabilities, but it also decreased when the permeability was approximately 2000 mD.These results demonstrate that oil with a higher viscosity exhibits a wider range of applicable permeability and that only in the lowpermeability sandpacks can the low-viscosity heavy oil such as Zhuangxi oil achieve a high tertiary oil recovery.Subsequently, temperature was changed to achieve different oil viscosities for the same oil sample.Xia-8 and Binnan heavy oil were selected for these tests, and their basic properties are also shown in Table 1.
The incremental oil recoveries of the sandpack flooding tests are shown in Figure 6.The results lead to the conclusion that the behaviors of tertiary oil recovery for the two oil samples are similar with the variation of oil viscosity.The incremental oil recovery reached 20% at 30 ∘ C and 50 ∘ C for the oil with high-viscosity, and it declined intensively at 70 ∘ C for the oil with low-viscosity.In Figure 7, the displacement pressure drop is plotted as a function of the injected volume.It shows that the peak pressure drop during alkaline flooding is higher than that during water flooding for the Xia-8 heavy oil, except for the oil at low-viscosity (70 ∘ C).It means that the plugging effect of water drops formed by injected alkaline solution is weakened for the low-viscosity oil sample.This result is consistent with the results of the sandpack flooding tests.Pictures of the effluent in these tests are shown in Figure 8.It can be seen that the interface of the produced fluid is clear, and the upper phase is the water drops inside the oil phase when the oil viscosity is high (30 ∘ C and 50 ∘ C).However, because the formation of water drops inside the oil phase cannot be implemented easily for the oil with low-viscosity (70 ∘ C), the alkaline solution injected into the sandpack mainly streams forward along the water channels and forms a muddy oil in water (O/W) emulsion, as shown in Figure 8(c).Consequently, heavy oil with a relative lowviscosity is unfavorable for alkaline flooding.

Mechanism of Oil Viscosity on the Performance of Alkaline
Flooding.In conventional oil reservoirs, high-permeability and low oil viscosity are both favorable for development.However, the results of sandpack flooding tests indicate that this behavior is different for alkaline flooding in heavy oil reservoirs.Microscopic flooding tests were conducted to investigate the mechanism of this anomalous phenomenon, and the images intercepted from the flooding process are shown in Figure 9.It can be seen that there are a lot of water drops inside the oil phase existing in the models for both Chenzhuang and Zhuangxi heavy oils.These drops can plug the pore throats and improve the sweep efficiency for displacing phase.Compared to Chenzhuang heavy oil, the size of water drops formed in Zhuangxi heavy oil is obviously smaller, and numerous drops can exist in the same pore throat.These smaller water drops can effectively plug the small pores in the low-permeability cores, but this capacity decreases intensively in the high-permeability sandpacks.Therefore, good flooding performance could be achieved only in the low-permeability sandpacks.Figure 10 shows the pressure changes during the flooding process for the Zhuangxi heavy oil in the cores with different permeabilities.The displacement pressure increased significantly when the alkaline solution was injected into the low-permeability sandpacks; however, it changed little in the medium-and high-permeability sandpacks, which indicated that the water drops cannot plug the pores effectively.
In contrast, when the alkaline solution penetrated into the oil with a higher viscosity, that is, the Chenzhuang heavy oil shown in Figures 9(d), 9(e), and 9(f), the water drops were large, and even the throat could be occupied by only one drop.These water drops could plug the pores in both the lowand medium-permeability sandpacks and resulted in the high tertiary oil recoveries, as shown in Figure 4.In conclusion, the high-viscosity of heavy oil is favorable for the improvement of sweep efficiency in alkaline flooding.
The results of the sandpack flooding tests discussed in Section 3.1 indicated that the two oil samples could exhibit similar behaviors when these tests were conducted at different temperatures.However, the changing temperature may influence the dynamic water/oil IFT, which is very important for chemical flooding in conventional oil reservoirs.Therefore, dynamic IFT between the alkaline solutions and crude oils was measured.The results are shown in Figures 11 and 12, and little change is observed as the temperature varies.These results are inconsistent with the conclusions of Chiwetelu et al. [9] and Trujillo [10] as described in Section 1. Besides, it can be seen from Table 2 that the efficiency of alkaline flooding has little relation to the equilibrium IFT.So, we think many factors must be considered to design an alkaline flooding plan for heavy oil besides the IFT.As previously discussed, two stages occur in the waterdrop formation process.The first stage is the penetration of the alkaline solution into the heavy oil and the occurrence of water columns covered with oil film.This process is related to the fast reduction in the oil/water IFT, which is difficult to measure using drop-spinning method.The second stage is the formation of discontinuous water droplets from the breakup of the water columns, which is caused by the non-uniform enrichment of in situ surfactant [11].High temperature accelerates the reaction and diffusion rate between alkaline solution and heavy oil, which reduces the degree of non-uniform enrichment of in situ surfactant.The water columns then cannot be easily divided into small discontinuous water droplets.Therefore, the temperature militates the performance of alkaline flooding.

Conclusions
(1) The results of flooding tests indicated that a high tertiary oil recovery could be achieved only in the lowpermeability sandpacks for the low-viscosity heavy oil; however, the high-viscosity heavy oil performed well in both the low-and medium-permeability sandpacks.In addition, the results of flooding tests for the same oil at different temperatures also indicated that the oil viscosity put a similar effect on alkaline flooding, and oil with a high-viscosity was favorable for alkaline flooding.
(2) Microscopic flooding tests were conducted to elucidate the influence of oil viscosity on the alkaline flooding.It indicated that the water drops produced during alkaline flooding for oils with different viscosities differed significantly in their sizes, which might influence the flow behaviors and therefore the sweep efficiencies of alkaline fluids.

Figure 1 :
Figure 1: Viscosity-temperature curves of four types of heavy oil.

Figure 6 :
Figure 6: Influence of oil viscosity on the alkaline flooding efficiency.

Figure 7 :
Figure 7: Pressure changes during alkaline flooding for Xia-8 heavy oil with different viscosities.

Figure 8 :
Figure 8: Pictures of the fluid produced during the displacement tests of Xia-8 heavy oil with different viscosities.(Effluents during alkali injection and the extended water flooding were collected in the tubes from left to right in turn.The pictures show their appearances at the end of flooding, without demulsification.)

Table 1 :
Basic properties of four types of heavy oil.

Table 2 :
of sandpack flooding tests for different heavy oils.