Sensitivity study of low salinity water injection in Zichebashskoe Oilfield

Presently, low salinity waterflooding is considered one of the most promising and cost-effective EOR methods. Although the recovery mechanisms are still uncertain, decrease of residual oil saturation and alternation of rock wettability are considered to be main mechanisms of the incremental oil recovery. In addition, laboratory and mathematical studies conducted over recent years suggested that mobility control is also a possible mechanism for enhanced oil recovery during low salinity water injection. The mobility control effect is due to induced fines migration and consequent permeability reduction in water invaded areas. The laboratory studies show that the incremental recovery gained from low salinity fines-assisted water flooding strongly depends on end point relative phase permeability for formation and injection waters. Permeability reduction due to fines migration decreases injected water mobility and increases the reservoir sweep. In this study, 24 years of production data from Zichebashskoe field (Russia) including 7 years of low salinity waterflooding are used to study the effect of water relative permeability reduction during low salinity waterflooding on improved oil recovery. The results of 3D reservoir simulations show low incremental oil recovery by low salinity water injection mainly due to two reasons: first a significant amount of water produced before the water injection, that is, a significant mixture between formation and injected waters that decrease the effect of low salinity; and second already high sweep efficiency as a result of water injection into water zone. The sensitivity study shows that the incremental recovery increases for greater relative permeability reduction by low salinity water injection .However, with 20 times decrease of Krwor from formation water to injected water, the incremental oil recovery is still negligible (4%).


INTRODUCTION
Recent studies of low salinity waterflooding have largely focused on the effects of water compositions on wettability, capillary pressure, relative permeability and residual oil saturation (Yildiz and Morrow, 1996;Tang and Morrow, 1999;Zhang and Morrow, 2006;Jerauld et al., 2008;Pu et al., 2010;Takahashi and Kovcsek, 2010;Berg et al., 2010;Cense et al., 2011;Mahani et al., 2011).The aforementioned studies conclude the effects of low salinity waterflooding to be similar to that for chemical EOR methods (Lake, 1989;Bedrikovetsky, 1993).The detailed analysis of microscopic physics mechanisms of low salinity waterflooding is explained in the reviews by Morrow and Buckley (2011) and Sheng (2014).
Other studies by Bedrikovetsky et al. (2011) and Zeinijahromi et al. (2011Zeinijahromi et al. ( , 2013Zeinijahromi et al. ( , 2014) ) propose that the mobility control effect of low salinity water injection that involves mobilization and migration of natural reservoir fines and consequent permeability reduction results in improved recovery.Morrow and Buckley (2011) also suggest that the formation of lamellae and emulsions, stabilized by fines, their migration and straining may result in mobility control and deep reservoir flow diversion.Tang and Morrow (1999) and Fogden et al. (2011), suggest another mechanism of oil-wet and mixedwet fines detachment by advancing water-oil capillary menisci; the resulting straining may also decrease the water relative permeability and increase oil recovery.
These effects appear to be separate phenomena from the fines migration by low salinity water and plugging of water-filled pores, but may occur simultaneously.Hussain et al. (2013) conducted an experimental study to confirm the above effects of the water phase permeability reduction during formation-and low-salinity waterflooding in oil-saturated rock.It was concluded that the water-wet particles have been removed from the rock by moving low salinity water, resulting in decrease in relative permeability for water and increase in fractional flow for oil.The conclusions agree with the mechanisms proposed by Sarkar and Sharma (1990).Some low salinity core flood studies have reported the release of significant amounts of fines (Bernard, 1967;Tang and Morrow, 1999;Pu et al., 2010), while others showed no evidence of fines migration (Lager et al., 2008;Jerauld et al., 2008) even though additional oil was recovered.In order to separate these effects, the injections leading to fines lifting and permeability decline are called "the fines-assisted waterflooding" (Kruijsdijk et al., 2011) in the current work.The fines-assisted version of low salinity waterflooding is a mobility control EOR technology.The present paper only considers the effects of fines mobilization and capture to provide mobility control and does not consider changes to the residual oil saturation or relative permeability curves as a result of wettability alternation during injection of low salinity water.
The available literature on laboratory studies and mathematical modelling of low salinity waterflooding highly exceeds that on the field trials.Very limited information on low salinity waterflooding pilot tests have been published in the open literature.Several limited field applications show significant recovery of residual oil (Webb et al., 2004;McGuire et al., 2005;Seccombe et al., 2010).However, the North Sea pilot, where the screening criteria for low salinity waterflood have been met, did not exhibit an incremental recovery (Skrettingland et al., 2010).The lack of information on real field applications of smart waterflooding with alteration of injected water composition as compared with the formation water is a serious restriction for large scale application of the technology in the oil industry.
The current paper presents analysis of a low-salinity water injection in a field case, based on limited available production and injection data from Zachubashskoye field (Russia, Tatarstan).The result of 7 years pressure maintenance by low salinity water injection in Zichebashskoe field is presented and modeled using mathematical modeling of fines-assisted waterflooding (Zeinijahromi et al., 2014).
The results of mathematical modelling of low salinity waterflooding strongly depend on relative phase permeability for formation and injection waters, particularly on Sor and Krwor.Decrease of Sor corresponds to well-known effect of wettability alteration with the salinity decrease, that is, the effect is equivalent to that in chemical EOR.Decrease of Krwor reflects the permeability damage induced by the reservoir fines mobilized by the injected low salinity water, that is, the recovery enhancing mechanism is the same as that in mobility control EOR.In the present study, we concentrate on sweep enhancement, that is, the latter case.

FIELD DESCRIPTION
Figure 1 shows current saturation map of the Zichebashskoe field and wells' location.The Zichebashskoe field consists of two sandstone reservoirs, Tula and Bobrik (Figure 1a and b).The layers are isolated with no hydrodynamic interaction and are connected to an active aquifer.The Tula (upper) layer has higher horizontal connectivity and permeability compared to that for Bobrik layer.
Production from Zichebashskoe field started in 1989 followed by pressure maintenance with low salinity water injection since 2006.Water production curve on Figure 2 (green line) shows that a significant volume of water has been produced during the period 1989 to 2006, before start of low salinity water injection.The main injectors are located below water-oil contact to inject water in the water zone in order to provide pressure maintenance during 2006 to 2013.
Water injection into aquifers yields better sweep and displacement than that in the oil zone, since the displacement of oil by water occurs "by the plane surface" moving upwards from wateroil contact during the injection into aquifer (Lake, 1989, Bedrikovetsky, 1993).Slope of the peripheral zones near the initial water-oil contact also increases the recovery during bottom-up waterflooding, since gravity effect decelerates water and accelerates oil.These two simultaneous mechanisms can explain high displacement efficiency during injection in water-oil contact in Figure 1 shows oil saturation averaged over the production thickness.Water cut in production wells gradually decreases from the position of initial oil-water contact up to the central part of the anticlinal field.One can also see that oil saturation increases from peripheral areas, where the injectors are located, towards the central part of the field.
The main properties of fluids and rocks are given in Table 1.The initial pressure is above the bubble point pressure; hence there is no initial gas cap and primary energy for the production is provided by adjacent active aquifer.
Tables 2 and 3 show the formation and injected water compositions, respectively.Extremely high formation water salinity is defined by sodium chlorite concentration that highly exceeds those for other salts, while magnesium and calcium salts dominate in injected water.Therefore, intensive ion exchange and consequent fines release is expected to occur during the displacement of formation water by low salinity injected water (Khilar and Fogler, 1998;Bedrikovetsky et al., 2011).

MATHEMATICAL MODELING OF FINES-ASSISTED LOW-SALINITY WATERFLOODING
Let us discuss a system of two-phase flow in porous media with varying water salinity resulting in the fine particles lifting.Following Muecke (1979), it is assumed that the water-wet particles are transported by the water phase.The detached forces mobilize water-wet fines that have been water-wet originally or, according to Berg et al. (2010) and Cense et al. (2011), became water-wet after the arrival of low salinity water; the mobilization occurs if the detaching torque of drag and lifting forces exceeds the attaching torque of electrostatic and gravity forces.It is assumed that the detached fines are inert, that is, they are intact and keep their integrity during detachment.The effects of clay swelling are assumed to be negligible.For simplicity, it is assumed that the volumetric concentrations of attached and retained particles are negligibly small compared to the porous space, that is, the retention of fine particles does not affect the porosity.It is also assumed that the initial salt concentration is the critical salt concentration for the reservoir fines,  a0 = cr ( i ), that is, the reservoir fines start leaving the rock surface with the decrease in salt concentration starting from = i .We also assume that the dissipation effects of diffusion and capillary pressure are negligibly smaller than those of fines straining.Alteration of water salinity affects the attached concentration stronger than the velocity alteration; therefore, the velocity dependency of the maximum concentration of attached fines is neglected.The permeability damage by fines straining is significantly higher than that by attachment.Other assumptions include constant temperature, incompressibility of water and oil, constant water and oil viscosities.
Volumetric balance of the overall flux of incompressible water and oil is: Volumetric balance for incompressible water is (Lake, 1989)   , 0 where the fractional flow function accounts for the reduction of relative phase permeability for water according to Equation 6: and U  is a three dimensional vector of the overall water-oil flux: The mass balance of suspended, attached and retained particles is: Here it is assumed that no fine particle attachment occurs in the reservoir during the injection of water without fines.Particle detachment occurs during injection of low salinity water into oilfield, where the attached fines with maximum concentration are in contact with water with continuously decreasing salinity, where drag and lifting forces are determined by interstitial velocity of water (Yuan and Shapiro, 2011).Equation ( 4) means an instant particle release governed by the torque balance.
Size exclusion capture of mobilized fine particles in small pores is described by the equation of linear kinetics (Bedrikovetsky, 2008) Here, the straining rate is proportional to water flux f(s)U since the mobilized fine particles are transported by the water phase.
The mass balance of salt in the aqueous phase assumes low salt concentration not affecting the aqueous phase density  w : The modified Darcy's law for two-phase flow accounting for permeability damage to water is: Finally, the system of governing equations for two-phase oil-water flow with fines mobilization, due to the decrease in water salinity and consequent reduction of relative permeability for water, consists of 7 equations: i) Volumetric balance for incompressible flux of carrier water and oil, ii) Volumetric balance for incompressible water, iii) Mass balance for suspended, attached and strained particles, iv) The maximum concentration of attached fine particles as a function of interstitial water velocity, salinity and saturation, v) Size exclusion retention rate, vi) Advective mass transfer of salt in porous space with retained fines and vii) Modified Darcy's law accounting for permeability reduction due to fines straining.
This system determines 7 unknowns  a , s, p, c, ,  s and U.
The initial conditions corresponding to injection of low salinity water into oil bearing formation include initial water saturation and initial concentrations of salt and of attached particles, zero values of suspended and strained fines.Boundary conditions on the injection wells include rate, unit fractional flow for water, salt concentration and zero concentration of suspended fines.Well bottomhole pressure is a boundary condition at the production wells.
The above system describes fines assisted waterflooding for all length scales, from core to a reservoir.Zeinijahromi et al. (2013) provides detailed derivation.

Mechanism for improved sweep efficiency due to fines migration
The observations, that fines migration can cause permeability decline because of changes in water composition and are sufficient to warrant the consideration of the effects of induced fines migration on waterflooding.During a waterflood, the rapid breakthrough of water can be a significant problem, leading to high water cut at producing wells and lower volumetric sweep efficiency.The problem is particularly pronounced for a mobility ratio significantly greater than unity or where the variation of permeability across the reservoir is significant.Fines release, due to the alteration of the chemistry of the injected water, and the consequent decrease in permeability, may be able to provide the mobility control and hence the ability to improve waterflood performance (Lemon et al., 2011;Zeinijahromi et al., 2011).Since the mobilization of fines by changing the chemistry of the injected water can only take place in the water-swept zone, only the effective permeability to water of the reservoir is decreased, reducing the mobility ratio.However, the main disadvantage of mobility control is that, for a given injection rate, the induced formation damage results in an increased injection pressure.
In displacement of oil by water in a heterogeneous reservoir, water propagates preferentially in highly permeable zones, with slow displacement of the oil in low permeability zones.A further displacement front in the low permeability zone occurs after water breakthrough in highly permeable zones and the creation of an injector/producer channel filled by high mobility water.The formation damage induced by mobilized fines in the swept zone homogenizes the permeability distribution across the reservoir and diverts the injected water into unswept areas.Hence, the induced formation damage causes the delayed breakthrough period and improved sweep efficiency for a given volume of injected water.

HISTORY MATCHING
Injection of low salinity water results in detachment of reservoir fines, their migration and straining in thin pore throats.The mathematical model for fines assisted water flooding is similar to those of mobility control EOR.Hence, in this study, the mathematical model for low salinity waterflood with changing relative phase permeability and accounting for fines mobilization and consequent permeability reduction in water swept areas, is used (Zeinijahromi et al., 2014).The system of equations for fines assisted waterflood is mapped on the system of equations for polymer flooding, allowing low salinity water injection with fines migration to be modeled using chemical option of black-oil model.Reservoir simulation software Tempest (Roxar, 2014) is used for modelling of low salinity and "normal (formation)" waterflooding in this study.The tracer option in Tempest is equivalent to polymer option without adsorption, where relative permeability can be made dependent on tracer (salt) concentration.The tuning parameters are pseudo (at the reservoir scale) phase permeability for oil and formation water, and the reduction factor to obtain the phase permeability for low salinity water from the phase permeability for formation water.
The Corey parameters are obtained by tuning the curves of cumulative oil and water production.The form of tuned pseudo relative permeability is shown in Table 4 for Tula and Bobrik layer.The Corey powers smaller than unity determine the convex forms of pseudo phase permeability, which is typical for those as obtained at the reservoir scale.
The result of history matching is presented in Figure 2 and exhibits a good match between the field history and the modelling data after the history matching.This model is later used to simulate normal (formation) water and low salinity water injection scenarios in Zichebashskoe field.

SENSITIVITY STUDY OF WATER RELATIVE PERMEABILITY REDUCTION
The mobility control effect of low salinity water injection is due to The modeling results show that the incremental recovery as a result of fines migration is highly dependent on degree of damage to water relative permeability.The purpose of this section is to study the effect of the degree of water permeability reduction on recovered oil.This is done by back calculating the recovery from the reservoir if formation water had been injected and comparing it with different low salinity water injection scenarios.Experimental results from different studies showed that during the injection of formation water, no ionic exchange or fines migration due to alteration of electrostatic occur.Therefore, we define formation water injection as a basic waterflood option, which is referred to as "normal" waterflooding.
Four cases of low salinity water injection are modeled and compared with normal waterflooding using tracer option of Tempest reservoir simulator.All reservoir, well and injection parameters are similar for four scenarios except the degree of damage caused by fines migration.Low salinity water injection is modeled for a case with negligible damage (3 folds decrease of water relative permeability as compared with normal waterflood), a normal damage case (6 folds decrease) and two extreme damage cases (10 and 20 folds decrease) (Figure 3).The pseudo relative permeability Kr depends on saturation and salinity  (or tracer concentration in Tempest C).Following the coreflood study by Hussain et al. (2013), it is assumed that pseudo relative permeability for oil, Krowi, residual oil saturation, Sor and power for oil, n o are independent of salinity or tracer concentration.The value of end point relative permeability Krwor for injected salinity (=0 or in Tempest: C=C max ) is decreased 3, 6, 10 and 20 times if compared with that for formation water ( 0 =1 or in Tempest: C=0).
In all cases, the well pressure is maintained below the fracturing pressure to avoid unrealistic prediction.In order to study the production improvement from the field, cumulative oil and water production are compared with normal water flooding (formation water injection) as the base case.

RESULTS AND DISCUSSION
Figure 4 shows the comparison for recovery factor and cumulative oil and water production between the normal and low salinity waterflooding.One can see that fines assisted water flooding resulted in production improvement in Zichebashskoe field by increasing the oil production and decreasing produced water.However, the results show that the improved recovery benefits are highly sensitive to the degree of reduction of water relative permeability during low salinity water injection (Figure 4a, b, c and d).
Figure 3a shows that with the three times decrease of water relative permeability due to low salinity, the incremental recovery is insignificant (~0.7%) and a reduction of produced water volumes is also small.The gained incremental recovery increases if the relative permeability of water decreases 6 times during low salinity water injection (Figure 3b); however it is much smaller than the 11% reported in Zeinijahromi et al. (2014) for a homogeneous 5 spot pattern.It can be explained by significant amount of water that has been produced before the start of low salinity water injection in Zichebashskoe field, that is, the injected water displaces the oil under already high water saturation.Another reason is injection of low salinity water into the aquifer in Zichebashskoe field.Oil is directly displaced by high salinity formation water and injected water lags significantly behind.The above are the main reasons why the incremental recovery factor with low salinity waterflooding is not significant.
The model for low salinity waterflood only accounts for fines migration and consequent decrease of relative permeability for water, that is, the effects of wettability change and residual oil saturation decreasing are ignored.Accounting for decrease in relative permeability for water and decrease in oil residual can bring additional incremental recovery if compared with the normal waterflooding.
Figure 4c and d present modeling results for severe water relative permeability reduction.It can be observed that the higher damage caused by fines migration results in more oil production improvement (Table 5).The cumulative volume of produced water also decreases with increase in damage.Since the mobilization of fines by changing the salinity of the injected water only takes place in the water-swept zones, only the effective permeability to water is decreased that reduces the water mobility.The formation damage induced by mobilized fines in the swept zone tends to homogenize the permeability distribution across the reservoir and diverts the injected water into unswept areas.Hence, higher induced formation damage causes more homogenized water front and improved sweep efficiency for a given volume of injected water in the case under study.It must be mentioned that the incremental recovery caused by

Folds
Oil Incremental RF 20 folds decrease 0.038 10 folds decrease 0.019 6 folds decrease 0.011 3 folds decrease 0.007 induced fines migration also depends on reservoir heterogeneity; thus there is an optimum permeability reduction above which no extra incremental recovery can be obtained (Zeinijahromi and Bedrikovetsky, 2014).
Experimental studies show that the degree of permeability reduction with injection of low salinity water is a function of the difference between formation and injection water salinity as well as type, size and concentration of initial reservoir fines (Khilar and Fogler, 1998;Bedrikovetsky et al., 2011;Zeinijahromi and Bedrikovetsky, 2013).A great difference between salinity of formation and injected water and also high concentration of movable fines can result in large release of fines from reservoir rock and consequently sever permeability damage.However, in high permeable reservoir with small volume of in-situ fines, change of injected water salinity may not result in a significant damage to rock permeability to be used as mobility control method.To further validate the method, the amount of coreflood studies with double-salinity waterflood must be enhanced.Low salinity water injected in Zichebashskoe field aimed to maintain reservoir pressure above the bubble point pressure (1.3 to 2.1 MPa).The pressure maintenance above the bubble point pressure by the waterflooding yields the commingled flow of oil with water that has significantly higher viscosity than the associated gas.In addition, oil viscosity increases during gas liberation during pressure depletion; thus, pressure maintenance during water injection is applied to avoid decrease of the oil flux due to gas evaporation and flow (Dake, 1998).Despite salinity of the injected water does not affect the pressure maintenance; the main shortcoming of waterflooding is water breakthrough to producers and cutting-off oil flux by lower viscosity water.Low salinity water can highly affect water flux and oil production.Decrease of the contact angle between oil and water results in decrease of residual oil saturation and improving water and oil relative permeability yielding some decrease in water flux, increase in the oil rates and the recovery factor.Moreover, lifting of the reservoir fines during low salinity waterflooding and the consequent fines migration result in significant decrease in relative permeability for water (Zeinijahromi et al., 2014;Hussain et al., 2013).Zeinijahromi et. al. 19 It should be mentioned that the current study does not emphasize the fines migration as the primary mechanism for low salinity effects in the field under investigation.The present paper concludes low efficiency of tertiary finesassisted low-salinity waterflooding.In the case where the reservoir rock contains a lot of kaolinite, tertiary lowsalinity water injection into water zone yields some minimum incremental oil recovery.
The mathematical model for low salinity water injection with increase of the rock wettability by water accounts for reduction in residual oil saturation, some increase in relative permeability for oil and some decrease in water relative permeability at the core scale.Despite ion exchange is the essential part of the mathematical model and residual oil alteration is triggered by change of the multi-component vector of ion concentrations, the above effects of chemical EOR is the main mechanism of incremental recovery due to salinity reduction (Pires et al., 2006).The mathematical model for so-called finesassisted low-salinity waterflood accounts for permeability damage to water due to migration of mobilized fines (Zeinijahromi et al., 2013).So, relative permeability for water decreases at the reservoir scale and the model for fines migration representing the mobility-control EOR also feature the variation of relative permeability.Finally, the salinity-dependence of relative permeability is the main EOR mechanism in the mathematical modeling of low salinity waterflooding.
The coreflood data are unavailable for the presented study.The forms of relative permeability as extracted from the coreflood would reveal whether chemical or mobility-control EOR effects dominate.In the present study, we concentrate on decrease in relative permeability for water.

Conclusions
Oil and water production data for low salinity waterflooding in Zichebashskoe oilfield are history matched by the fines-assisted-waterflood model (tracer model in Tempest) with high accuracy.The modeling show that injection of low salinity water results in improved oil production and reduction in produced water volume.However, the improved recovery with fines assisted waterflooding depends on degree of the permeability damage during low salinity water injection.The higher the water relative permeability reduction due to fines release, the greater is the incremental recovery; however, obtained improved recovery is limited by heterogeneity of the reservoir.
Low salinity water injection under the conditions of Zichebashskoe field results in negligible incremental recovery and small decrease in the produced water if compared with the waterflooding by formation water.The phenomenon is explained by high flooding of the reservoir before commencement of low salinity water injection, by high salinity aquifer water.Another explanation is that low salinity water injection into aquifer causes lower incremental recovery than that with the injection into oil-zone due to usual high sweep efficiency of water injection into aquifers.

Figure 3 .
Figure 3.Relative permeability for water and oil for the case studies with the sensitivity analysis curves.

Figure 4 .
Figure 4. Comparison between formation and low salinity waterflood for 30 years prediction a) 3 time decrease of Krwor b) 6 times decrease c) 10 times decrease d) 20 times decrease.

Table 1 .
Properties of rocks and fluids in Bastrykskoye field.

Table 2 .
Composition of formation water in Zichebashskoe field.

Table 3 .
Composition of fresh lake water injected in Zichebashskoe field.