Journal of Petroleum Science and Engineering 135 (2015) 323–335

Contents lists available at ScienceDirect

Journal of Petroleum Science and Engineering

journal homepage: www.elsevier.com/locate/petrol

Analytical solution for oil displacement by polymer slugs containing

salt in porous media
Anália S. de Paula, Adolfo P. Pires n
Universidade Estadual do Norte Fluminense, Rodovia Amaral Peixoto, km 163 27925-535 Macaé, RJ, Brazil

art ic l e i nf o a b s t r a c t

Article history: The displacement of oil by water containing dissolved chemical products in porous media is one of the
Received 30 January 2015 most used Enhanced Oil Recovery (EOR) techniques. The continuous injection of these displacing fluids is
Received in revised form very expensive; the alternative is the injection of finite volumes (slugs). Chemical methods of EOR in-
31 August 2015
volve mass transfer through adsorption and also changes transport properties. Mathematical modeling
Accepted 1 September 2015
Available online 12 September 2015
and experimental tests are important for the development of a project to increase oil recovery by EOR
methods. This paper presents the analytical solution for the injection of a water slug containing dissolved
Keywords: polymer and salt followed by water drive. In this model the salt does not adsorb nor modifies the water
Enhanced oil recovery fractional flow curve, but changes the adsorption isotherm. The solution is obtained by introducing a
Polymer flooding
potential function based on the conservation of the aqueous phase volume which splits the original
Conservation law
system into a thermodynamics auxiliary system and a transport equation. The thermodynamics auxiliary
Potential function
system depends only on the adsorption isotherm and allows predicting the distribution of components
between the different phases. The transport equation, also called lifting equation, is a function of the
relative permeabilities and viscosities of the phases. The solution of this problem is not self-similar, like
the continuous injection of dissolved chemical components in water, and interactions among waves
occur. A sensitivity analysis was performed to evaluate the effects of some parameters in the efficiency of
the technique. The solution allows determining the slug behavior, and shows the development of the
chromatographic cycle in porous media, illustrating the appearance and disappearance of the compo-
nents injected into the reservoir rock.
& 2015 Elsevier B.V. All rights reserved.

1. Introduction changes and mass transfer. The different methods of enhanced oil
recovery may lead to problems of multicomponent multiphase
Enhanced oil recovery is defined as oil recovery by injection of flow in porous media.
materials normally absent in the reservoir (Lake, 1989). The defi- The chemical methods of EOR include the injection of surfac-
nition of EOR is not associated with a specific production phase of tant, polymers and alkaline solutions, and usually don’t change the
the reservoir. The injection of these fluids can happen at any time chemical properties of oil. Polymer injection is the most wide-
during the productive life of an oil field and excludes water in- spread used technique among chemical methods of EOR. The main
jection and other methods of pressure maintenance. The methods objective of the polymer solution injection is to reduce the mo-
of enhanced oil recovery can be divided into three main cate- bility ratio by increasing the viscosity and decreasing the perme-
gories: chemical, solvents and thermal. Beyond the categories ability relative to water, consequently improving the sweep effi-
herein mentioned, there are other advanced recovery techniques ciency. Among others, the most used polymers are polysaccharides
that do not fit specifically into any of them, such as MEOR (in- and partially hydrolyzed polyacrylamides.
jection of microorganisms), electromagnetic methods and FOAM The injection of chemical solutions into reservoirs can lead to
(foam injection) (Lake, 1989). retention of molecules in the porous media, resulting in perme-
Enhanced Oil Recovery techniques are alternatives to get more ability reduction to the injected solution and decreasing its
production compared to primary and secondary recovery meth-
amount in solution. The adsorption occurs through the interaction
ods. When these EOR techniques are applied, thermodynamic
between the polymer molecule and the rock matrix and in most
phenomena take place in porous media: non-isothermal phase
cases it is the predominant retention mechanism. In this work, it
will be considered the amount of chemical component adsorbed
n
Corresponding author. reversibly. This phenomenon occurs in both static and dynamic

http://dx.doi.org/10.1016/j.petrol.2015.09.001
0920-4105/& 2015 Elsevier B.V. All rights reserved.
324 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335

Nomenclature X dimensionless distance

a concentration of adsorbed component Greek letters

A area
c chemical concentration in solution ϕ potential
f water fractional flow Ω slug volume
F density function of hyperbolic lifting equation
s dimensionless water saturation Subscripts
sw water saturation
t time
A condition at point A
T number of slug volumes injected
R reservoir
U flow function of hyperbolic lifting equation
x distance

conditions. The case of two-phase multicomponent flow, where the com-

The concentration of retained polymer, one of the key factors to ponents may adsorb and the composition affects the physical
determine the technical and economic viability of this technique properties of the phase (density, viscosity and interfacial tension,
varies with the characteristics of the adsorbate and rock. The re- etc.) have been solved based on two different approaches. One is
lationship between the amount of adsorbed product on the solid the theory of multicomponent chromatography, which describes
surface and the concentration of the solute in the aqueous phase is the behavior of a system formed by two phases, one stationary and
called adsorption isotherm. one movable, comprising n components where each component
Buckley and Leverett (1942) developed the solution for the affects the distribution of all others. An example is the separation
problem of oil displacement by immiscible water considering the of components by chromatography. Rhee et al. (1970) developed
theory of fractional flow and method of characteristics. The solu- the solution for adsorption and desorption of chemical compo-
tion may present more than one saturation at a point, and this nents in flow through a chromatographic bed. The Langmuir iso-
physical inconsistency introduces a discontinuity in the solution therm was used to model the mass transfer and the solution was
(shock), calculated by mass balance. Based on this solution, a obtained for both self-similar problems (continuous injection so-
graphical method for determining the recovery factor of oil was lutions), and also for the case where there is interaction between
developed (Welge, 1952). This technique was extended to include waves (adsorption followed by desorption and vice versa). The
the case of mass transfer and continuous change in mobility Langmuir isotherm can also be applied to model the adsorption in
(Claridge and Bondor, 1974). porous media, resulting in a problem similar to that of chroma-
There is a kind of physical systems that can be modeled by tography. The second approach is based on the flow of immiscible
hyperbolic systems of partial differential equations. In general, fluids in porous media considering two mobile phases, but does
problems involving conservation of mass, momentum and energy not include the distribution of the components between the
without dispersion effects lead to this type of mathematical phases. Helfferich (1981) extended the second approach to mul-
systems. ticomponent multiphase problems considering different models of
Analytical models for one-dimensional displacement of oil by phase equilibrium and fractional flow function. This theory was
water containing dissolved chemicals have been studied since the also applied to a three component system that simulates the in-
60s. Early work modeled continuous injection of surfactant in oil jection of surfactant in a two-phase environment (Hirasaki, 1981).
reservoirs through a 2
2 system of hyperbolic equations, using Isaacson and Temple (1986) analyzed the injection of a solution of
the Langmuir isotherm to relate the amounts adsorbed and dis- n polymers in porous media and solved the Cauchy–Riemann
solved (Fayers and Perrine, 1958). Patton et al. (1971) presented an problem for a system of hyperbolic conservation laws, using an s-
analytical solution to the one-dimensional problem of one che- shape fractional flow curve. Dahl et al. (1992) developed the self-
mical product injection, which provides a quick estimate of the similar solution of Riemann problems for a class of two-phase
amount of additional oil recovered. Pope (1980) included the ef- multicomponent displacement. It is based on the Riemann solu-
fects of the presence of chemicals in the water injection by tion for the corresponding one-phase problem.
changing the fractional flow curve to solve problems of enhanced The continuous injection of chemical solutions in porous media
oil recovery. is not economically feasible; an alternative is the injection of slugs.
The temperature of the injected fluid affects the performance of The solution of this initial and boundary value problem leads to
water injection, with or without additives, in oil reservoirs. In interactions between waves and it is not self-similar. Part of the
order to solve the problem, it is necessary to include the energy solution is identical to the case of continuous injection and it was
balance in the system of equations (Braginskaya and Entov, 1980). applied to three different adsorption isotherms (Bedrikovetsky,
In the 80s, detailed studies on oil displacement by water con- 1993).
taining chemicals were published. These papers developed several The introduction of a potential function associated with the
analytical solutions considering adsorption, the structure of these aqueous phase volume conservation can be used to solve systems
solutions was composed by a combination of discontinuities and of hyperbolic equations that arise in some problems of two-phase
rarefaction waves of saturation and concentration. The influence of flow in porous media. This potential is used to replace the time
saturation and concentration on the fractional flow curve for the variable and splits the original system into an auxiliary system
case of two phase displacement was analyzed (Alishaeva et al. with reduced size and a lifting equation. The unknowns of the
1982; Entov and Kerimov, 1986). An analytical solution when two auxiliary system are related only to the thermodynamic model, for
components are dissolved in water and one of the additives does the case of chemical injection they are defined by adsorption
not affect the mobility of the phases but can interact with the isotherms. The lifting equation depends on hydrodynamic prop-
active additive was developed (Entov and Zazovskii, 1982). erties such as viscosity and relative permeability (Pires et al.,
 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335 325

2006). saturation.
This methodology was subsequently applied to solve the pro- The dimensionless time and space variables are defined as a
t
blem of water injection containing a polymer in the presence of ∫0 u(y)dy
function of the number of slug volumes injected, T = and
salt, which does not adsorb nor alters the fractional flow curve φ(ΩB / AR )
x
(Boa, 2006), and the injection of n polymers (Cardoso et al., 2007). X= ΩB / AR
, where ΩB is the slug volume and AR is the reservoir
In both cases, it was considered the continuous injection of solu- cross-sectional area.
tions containing chemicals, i.e., the solutions were self-similar. It is considered that at the beginning of the injection there is no
Ribeiro and Pires (2008) developed analytical solutions for the polymer dissolved in the reservoir water. In the case of injection of
case of water slug injection containing one polymer considering a finite amount of water containing chemicals (slug), the initial
linear (Henry), convex (Langmuir) and concave isotherms and and boundary conditions for system (1) are:
convex fractional flow. It was also presented the solution for two
polymer slug injection where polymers adsorb according to Henry s w(X , 0) = s w (I ) X > 0
isotherm; slug injection of a polymer that adsorbs and a surfactant → →(I )
that does not adsorb and changes the residual oil saturation, and a c (X , 0) = c X>0
slug containing a polymer and a salt that does not change the
f (0, T ) = f (J ) = 1 T > 0
fractional flow function but the amount adsorbed. In all cases a
convex fractional flow function was used. Borazjani et al. (2014) ⎧ →(J )
→ ⎪c 0<T<1
applied the splitting procedure to solve the problem of polymer c (0, T ) = ⎨
⎪
⎩0 T > 1
slug injection considering Henry´s adsorption isotherm and sali- (2)
nity effects.
In this work we solve a similar problem (chemical slug injec- The initial condition is denoted (I) and the injection condition
tion containing one dissolved polymer in the presence of salt) (J). The polymer is injected at the reservoir inlet (X¼0) and after
considering a non-linear sorption isotherm using the splitting the slug injection (T ¼1) water drives the solution.
theory (Pires et al., 2006).
Next section shows the mathematical framework of the pro- 2.1. Splitting between thermodynamics and hydrodynamics
blem of oil displaced by water slugs containing dissolved chemi-
cals followed by the modeling of oil displaced by a polymer slug in In order to solve the system of Eq. (1) subject to initial and
the presence of a salt. A sensitivity analysis is performed to eval- boundary conditions (2) the following potential will be introduced
uate the influence of some parameters of the model on the results, (Pires et al., 2006):
and the last section brings the main conclusions. →
dϕ = f (s , c )dT − sdX (3)

The potential (3) is associated with the conservation of water

2. Mathematical modeling of oil displacement by water con- volume; and it splits the original system (1) into a lifting equation
taining dissolved chemical components and an n x n auxiliary system:
∂ s ∂ 1
In this section it will be shown the mathematical modeling of ( w→ ) − ( → ) = 0
∂ϕ f (s , c ) ∂X f (s , c ) (4)
oil displacement by a water slug containing different chemical w

species. The following hypothesis will be considered:

→
∂ai(c ) ∂c
 Homogeneous and isotropic porous media; + i =0 i = 1... n
∂ϕ ∂X (5)
 Residual oil saturation does not depend on polymer →
concentration; The unknowns of system (4)–(5) are the vector c (X, ϕ) and
 One-dimensional two-phase incompressible flow; s(X, ϕ). Change of independent variables (X,T) to (X, ϕ) leads to the
 Absence of dispersive effects; following set of initial and boundary conditions:
 Negligible capillary and gravitational effects; Initial condition (ϕ = 0)
 Chemical components dissolved in the aqueous phase only;
 Water density independent of the concentration of chemical s w(X , 0) = s w (I ) X>0
additives. → → (I )
c (X , 0) = c X>0 (6)
We assume that the salt concentration is not high enough to Boundary condition (X = 0)
decrease the viscosity of the polymer solution and change the
fractional flow curve. f (0, ϕ) = f (J ) (ϕ) = 1 ϕ > 0
The displacement of oil by water containing dissolved chemi-
⎧
⎪→
(J )
→
cals which can be adsorbed by the rock is described by the fol- c (0, ϕ) = ⎨ c 0 < ϕ < 1
⎪
lowing system of hyperbolic equations: ⎩0 ϕ > 1 (7)
←
∂s w ∂f (s w , c ) The potential function will be used as an independent variable
+ = 0
∂T ∂X instead of time. The auxiliary system (5) depends purely on the
→ →
∂(cis w + ai(c )) ∂(cif (s w , c )) thermodynamics equilibrium governed by the adsorption iso-
+ = 0 i = 1... n therm. Eq. (4), from now on called lifting equation, depends on
∂T ∂X (1)
both thermodynamics and transport properties (phase relative
→
where ci is the concentration of the component i, ai(c ) is the ad- permeabilities and viscosities) and can be solved after the solution
sorption of i, f is the water fractional flow and sw is the water of system (5) is known. After that, the solution is mapped from the
326 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335

(X, ϕ) plane to (X , T ) domain.

3. Injection of water slug containing a polymer in the presence

of salt

This section will discuss the displacement of oil by an aqueous

solution slug containing one polymer with water drive. The salt
concentration influence on the polymer adsorption is considered.
Section 3.1 describes the mathematical modeling of this pro-
blem. It presents the adsorption isotherm, the mathematical sys-
tem that characterizes the physics of the flow and the initial-
boundary conditions corresponding to the slug displacement. The
splitting theory is applied to change the coordinates system from
(X , T ) to (X, ϕ).
The n × n auxiliary system is solved and the polymer and salt
concentrations are found, followed by the solution of the hydro-
dynamic part of the problem to obtain the water saturation. The
inversion mapping of the solution to time domain is presented in Fig. 1. Auxiliary problem solution.
Section 3.4.

3.1. Mathematical modeling of oil displacement by water slug con- ⎧ c = c (I ) = 0

⎪ 1 1
taining one polymer and salt ⎪ (I )
ϕ = 0⎨ c2 = c2
⎪U → ∞
In this work the Langmuir type adsorption isotherm is used to ⎪
model the polymer distribution between solid and liquid phases. ⎩F → ∞ (12)
When there is salt dissolved in the injection water the isotherm is
given by: ⎧ c = c (J )
⎪ 1 1
(β1 + β2c2)c1 ⎪ c = c (J )
a1(c1, c2) = X = 0⎨ 2 2
1 + β3c1 (8) ⎪ U (J ) = 1
⎪
where c1 and c2 are the concentrations of the component 1 ⎩ F (J ) = − 1 (13)
(polymer) and of component 2 (salt) respectively, and β1, β2 and β3 It is considered that the reservoir water salinity is greater than
are model constants determined from lab experiments. Polymer
the injected water salinity ( c2(I ) > c2(J )).
mass transfer between solid and liquid phases follows the Lang-
muir model of reversible adsorption. Salt is not adsorbed by the
rock and does not change the fractional flow function. 3.2. Auxiliary problem solution
The system of equations that describes the water volume
conservation and the mass conservation of the polymer and salt is The solution is described by the eigenvalues of the auxiliary
given by: system. The speeds of the characteristic curves are:

⎧ ∂s λI = 0
∂f (s w , c1)
⎪ w + =0 ∂a (c , c )
⎪ ∂T ∂X λII = 1 1 2
⎪ ∂(c s + a (c , c )) ∂(c1f (s w , c1)) ∂c1 (14)
⎨ 1w 1 1 2
+ =0
⎪ ∂T ∂X In this case λII > λI for any concentration. The solution for the
⎪ ∂(c s ) ∂(c2f (s w , c1))
⎪ 2 w + =0 concentration is built in the auxiliary physical plane, divided in the
⎩ ∂T ∂X (9) following regions (Fig. 1):
→ → → - Region I (ϕ = 0): c1 = 0 and c2 = c2(I );
Defining F (U , c ) = − s/f (s, c ) and U = 1/f (s, c ), defining a di-
- Region 1: c1 = 0 and c2 = c2(J ) ;
s w (X , ϕ ) − s w ( I )
mensionless water saturation, s = , and applying the - Region 2: c1 = c1(J ) and c2 = c2(J ) ;
s w (J ) − s w (I )
potential function in (9) we obtain the following auxiliary system - Region 3: c1 varying from c1(J ) to c1 = 0 and c2 = c2(J );
and lifting equation - Region 4: c1 = 0 and c2 = c2(J );
⎧ ∂a1(c1, c2) The rarefactions in region 3 (c-rarefactions) are calculated by:
∂c
⎪ + 1=0
⎪ ∂ϕ ∂X ⎧ 1 ⎫
⎨ ⎡ ⎤2
1 ⎪ ⎢ (β1 + β2c2 )X ⎥ ⎪
(J )
⎪ ∂c2 c1r (X , ϕ) = ⎨ − 1⎬
⎪
⎩ ∂X
=0 β3 ⎪ ⎢⎣ ϕ−1 ⎥⎦ ⎪
(10) ⎩ ⎭ (15)
→ In each region the velocity of the characteristics, which carry
∂F (U , c ) ∂U
+ =0 constant concentration values, is given by:
∂ϕ ∂X (11)
dϕ ∂a (c , c ) β + β2c2
Initial and boundary conditions that describe the chemical slug = 1 1 2 = 1 = η(c1, c2)
dX ∂c1 (1 + β3c1)2 (16)
injection in the presence of salt are depicted by:
 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335 327

The path of the shock ϕV1(X ) that separates region 1 and 2 is

given by:
ϕV (X ) = V1X (17)
1

Shock speed V1 is calculated through:

dϕ [a ] β + β2c2(J )
V1 = = 1 = 1 ,
dX [c1] (1 + β3c1(J )) (18)

valid for XoXA. Region 2 is limited by point A, whose coordinates

are determined by the intersection of shock V1 and the last char-
acteristic of this region:

(1 + β3c1(J ))2
XA =
β3c1(J )(β1 + β2c2(J )) (19)

(1 + β3c1(J ))
ϕA =
β3c1(J ) (20) Fig. 2. Polymer (c1) and salt (c2) concentration profiles during slug injection.

When X4XA, a new shock between the characteristics of re-

gion 1 and the family of rarefactions of region 3 takes place and its
speed is given by

dϕ [a ] (β + β2c2(J ))
V2 = = 1 = 1
dX [c1] 1 + β3c1r (21)

An expression for the shock V2 as a function of X and φ can be

obtained substituting c1r from Eq. (15) in (21):

dϕ (β1 + β2c2(J )) ϕ − 1
V2 = =
dX X (22)

Shock ϕV 2(X ) path is calculated by the integration of the Eq. (22)

and it was found to be:
2
ϕV 2(X ) = { β1 + β2c2(J ) ( X − )
XA + ϕA − 1 } +1 (23)

The general solution of the auxiliary system is divided into two

different regions by point A:

→ ⎧
⎪ c1(I )(X , ϕ ) , c2(I )(X , ϕ ) X ≤ XA
c (X , ϕ) = ⎨
⎩ c1(II )(X , ϕ), c2(I )(X , ϕ) X > XA
⎪
(24) Fig. 3. Polymer (c1) and salt (c2) concentration profiles during 1 o φo φA.

⎧ 0, 0 ≤ ϕ < ϕ (X )
⎪ V1
⎪ (J ) (J ) (J )
⎪ c1 , ϕV1(X ) < ϕ < η(c1 , c2 )X + 1
⎪
→
c1(I )(X , ϕ) = ⎨
⎪ c (X , ϕ), η(c (J ), c (J ))X + 1 ≤ ϕ < η(0, c (J ))X + 1
⎪ 1r 1 2 2
⎪
⎪ (J )
⎩ 0, η(0, c2 )X + 1 ≤ ϕ < ∞

⎧ c ( I ), ϕ = 0
→ ⎪ 2
c2(I )(X , ϕ) = ⎨
⎪ c ( J ), ϕ > 0
⎩ 2 (25)

→
c1(II )(X , ϕ) = { 0, 0 ≤ ϕ < ϕV 2(X ) (26)

The concentration profiles are shown in Figs. 2 to 4. During the

injection of the water slug, the polymer concentration is constant
and equal to the injection value. For the water drive period the
polymer concentration increases from the reservoir inlet (where
c1 = 0) to the injection concentration. For ϕ > ϕA we note that the
polymer concentration is smaller than the injection concentration Fig. 4. Polymer (c1) and salt (c2) concentration profiles when φ4φA.
328 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335

Fig. 6. Solution path in plane (F,U) for X o XA.

Fig. 5. Solution for U (X, ϕ) in the auxiliary plane. The solution UII (X, ϕ) is represented by:
For ϕ < η(0, c2(J ))X + 1

at any point in the reservoir, but the increasing profile is kept. In ⎧ U (I )(X , ϕ) ϕ = 0
⎪
Figs. 3 and 4 it is possible to see the appearance and disappearance ⎪ (1)
of the polymer, the chromatographic cycle. The salt concentration ⎪U 0 < ϕ < ϕV (X )
2
⎪
profile is composed by a zero speed jump to the injection ⎪
condition. ⎨ ∂F (1, c1 = c1r , c2 = c2(J ))
⎪ U3(X , ϕ) ϕV (X ) < ϕ ≤ X
⎪ 2 ∂U
3.3. Lifting equation solution ⎪
⎪ (J ) ∂F (1, c1 = c1r , c2 = c2(J ))
⎪ U =1 X≤ϕ<∞
⎩ ∂U (32)
In order to solve the hydrodynamic part of the problem it is
necessary to determine the lifting equation solution, noting that For ϕ > η(0, c2(J ))X +1
the salt does not alter the fractional flow curve:
⎧ ∂F (1, c1 = 0, c2 = c2(J ))
∂F (U , c1) ∂U ⎪ U4− η(0, c2(J ))X + 1 ≤ ϕ ≤ X
+ =0 ⎪ ∂U
∂ϕ ∂X (27) ⎨
⎪ ∂F (1, c1 = 0, c2 = c2(J ))
For regions where the concentration is constant, U is constant ⎪ U (J ) = 1 X≤ϕ<∞
⎩ ∂U (33)
on the characteristics. The speed of the characteristic curve is gi-
dϕ ∂F
ven by: dX = ∂U . The superscript indicates a constant state in the region and the
The solution of the lifting equation in the auxiliary plane can be subscript represents families of rarefactions. Thus U2 represents
seen in Fig. 5. In region 3 there is interaction between two families the rarefaction wave in region 2, U3 represents values of U in
of rarefactions, a U-rarefaction and a c-rarefaction. In this region, variable concentration region 3, and U4− is the rarefaction family
therefore, U changes on a characteristic wave. The rarefaction fa- which leads to the injection condition. Values of U before the
milies in region 3 begin at ( X *, ϕ*) and in region 4 at ( X **, ϕ**). shock are called U þ , and values after the shock U  .
The general solution for U (X, ϕ) may be divided into two parts:

U (X , ϕ) = {UI (X , ϕ), X ≤ XA , (28)

and the solution for UI (X, ϕ) is given by:

For ϕ < η(c1(J ), c2(J ))X + 1
⎧
⎪ U (I )(X , φ) φ = 0
⎨ (1)
⎪
⎩U 0 < φ < φV (X )
I (29)

For η(c1(J ), c2(J ))X + 1 < ϕ < η(0, c2(J ))X +1

⎧ ∂F (1, c1 = c1r , c2 = c2(J ))
⎪ U3(X , ϕ) ϕ (X ) < ϕ ≤ X
⎪ V2
∂U
⎨
⎪ ∂F (1, c1 = c1r , c2 = c2(J ))
⎪ U (J ) = 1 X≤ϕ<∞
⎩ ∂U (30)

For ϕ > η(0, c2(J ))X +1

⎧ ∂F (1, c1 = 0, c2 = c2(J ))
⎨ U4− η(0, c2(J ))X + 1 ≤ ϕ ≤ X
⎩
⎪
∂U (31) Fig. 7. Solution path in plane (F,U) for X 4XA.
 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335 329

The solution path in (F,U) plane for XoXA is shown in Fig. 6. It

begins with a jump from the initial conditional U(I) to the point U(1)
on the fractional flow curve F (U , c1 = 0, c2 = c2J ) with zero speed. At
this point there is a new jump with velocity V1 from U(1) to U(2) on
the curve F (U , c1 = c1(J ), c2 = c2J ). From this point a family of rar-
efaction waves links the fractional flow functions
F (U , c1 = c1(J ), c2 = c2J ) to F (U , c1 = 0, c2 = c2J ) up to the injection con-
dition. The structural formula is given by I → 1 → 2 − 3 − 4 − J ,
where the symbol ( →) indicates a shock and ( −) a rarefaction.
When X4 XA, the solution begins with a jump from the initial
condition U(I) to point U(1) on the curve F (U , c1 = 0, c2 = c2J ) with
zero speed. At this point there is a jump with speed V2 up to U3 on
the curve F (U , c1 = c1r , c2 = c2J ). From this point there is a rarefaction
wave where both U and c1 change continuously up to the frac-
tional flow function F (U , c1 = 0, c2 = c2J ) (Fig. 7). The solution path is
represented by I → 1 → 3 − 4 − J .
After the solution of the lifting equation, it is necessary to de-
termine the saturation and the fractional flow function solution in
the plane (X
φ). First the fractional flow is calculated from
f = 1/U (X , ϕ) and subsequently the saturation is obtained from
s (X , ϕ ) = − F (X , ϕ ) / U (X , ϕ ) .

Fig. 9. Solution path in plane (s
f) for the polymer slug injection with salt for
3.4. Inversion of the solution to time domain X oXA.

Once the solution of the auxiliary system and the lifting

equation were developed in the auxiliary space, the inversion to
time domain can be performed through the following expression,
based on the definition of the potential function:
dϕ s
dT = → + → dX
f (s(X , ϕ), c (X , ϕ)) f (s(X , ϕ), c (X , ϕ)) (34)

The eigenvalues of the auxiliary system are related to the ei-

genvalues of the original system through:
f
Λ=
s+λ (35)

and the speeds of the shocks of the two systems by:

f±
D=
s± + V (36)

The solution for concentration and saturation in plane X
T is

shown in Fig. 8.
The regions are the same as the ones in the auxiliary plane
(Fig. 1). A new region (region I), that corresponds to the initial

Fig. 10. Solution path in plane (s
f) for the polymer slug injection with salt for
X 4XA.

condition in auxiliary plane (φ ¼0) appears. Region I is bounded

T s(1)
by the straight line with slope = .
X f (1)
In region 1, saturation is constant and equal to s(1) . Up to point
A, a straight line separates this region and region 2:

f (1)
D1 = (1)
s + V1 (37)

After A the shock D2 separates regions 1 and 3. This shock is not

a straight line, and its path is determined by:

f − (s−, c1r ) f (1)

D2 = − = (1)
s + V2(X , ϕ) s + V2(X , ϕ) (38)

Fig. 8. Saturation solution in plane (X
T). The characteristics that carry constant saturation (regions 2 and
330 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335

Fig. 11. Polymer concentration profile (a), salt concentration profile (b) and saturation profile (c).

Table 1 In the region sI (X , T ) the solution is given by:

Geometric and production parameters used for sensitivity analysis. (η(c1(J ), c2(J )) + s2)X + 1
For T < ,
f (s2, c1(J ), c2(J ))
Length 1000 m
⎧ (I ) X
Height 50 m ⎪ s (X , T ) = 0 0 < T < f
Width 200 m ⎪ s
Porosity 0.15 ⎪
Flow Rate 5000 m3/d ⎪ X X
⎪ s (1) f ≤ T <
⎪ D1
⎪ s
⎨
4) in auxiliary plane are also straight lines in this plane. ⎪ s (X , T ) X1 ≤ T ≤ X
For all rarefaction families the speed of the characteristics for ⎪ 2 D1 (J ) (J )
∂f (1, c1 = c1 , c2 = c2 )
⎪ ∂s
constant saturation can be calculated from (34): ⎪
⎪ X
X ∂f (s , c1, c2) f f ⎪ s (J ) = 1 (J ) (J )
≤T<∞
= = ∂F (U, c1, c2)
= ϕ ⎪ ∂f (1, c1 = c1 , c2 )
T ∂s s+ s+ ⎩ (42)
∂U X (39) ∂s

In region 3 where both saturation and concentration change we (η(c1(J ), c2(J )))X + 1 (η(0, c2(J )) + s3)X + 1
have For <T<
f (s2, c1(J ), c2(J )) f (s3, 0, c2(J ))

ϕ* 1 ϕ 1 ⎧ X X
T3 = ∫0 f (s2, c1(J ), c2(J ))
dϕ + ∫ϕ* f (s(X , ϕ), c1(X , ϕ), c2(J ))
dϕ ⎪ s3(X , T ) ≤T≤ (J )
⎪ D 2 ∂f ( 1, c 1 = c1r, c2 = c2 )
X* X ⎪ ∂s
s2 s(X , ϕ) ⎨
+ ∫0 f (s2, c1(J ), c2(J ))
dX + ∫X* f (s(X , ϕ), c1(X , ϕ), c2(J ))
dX
⎪ (J ) X
(40) ⎪s = 1 ≤T<∞
(J )
⎪ ∂f (1, c1 = c1r, c2 = c2 )
The saturation solution can be divided into two regions: ⎩ ∂s (43)

⎧ sI (X , T ), X ≤ XA
s(X , T ) = ⎨ (η(0, c2(J )) + s3)X + 1
For T > ,
⎩ sII (X , T ), X > XA (41) f (s3, 0, c2(J ))
 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335 331

Fig. 12. Polymer concentration profile (a), salt concentration profile (b) and saturation profile (c) for the case of 0.3 pore volumes slug size.

⎧ (η(0, c2(J )) + s3)X + 1 ⎧ (η(0, c2(J )) + s3)X + 1

⎪ s4−(X , T ) X ⎪ s4−(X , T ) X
≤T≤ ≤T≤
⎪ f (s3, 0, c2(J )) (J )
∂f (1, c1 = 0, c2 = c2 ) ⎪ f (s3, 0, c2(J )) (J )
∂f (1, c1 = 0, c2 = c2 )
⎪ ∂s ⎪ ∂s
⎨ ⎨
⎪ (J ) X ⎪ (J ) X
⎪s = 1 (J )
≤T<∞ ⎪s = 1 (J )
≤T<∞
⎪ ∂f (1, c1 = 0, c2 = c2 ) ⎪ ∂f (1, c1 = 0, c2 = c2 )
⎩ ∂s (44) ⎩ ∂s (46)

The polymer and salt concentration solution in the physical

The solution for the region sII (X , T ) is given by:
plane is given by:
(η(0, c2(J )) + s3)X + 1
For T < ,
f (s3, 0, c2(J )) ⎧ c1(I )(X , T ), c2(I )(X , T ) X ≤ XA
→ ⎪
c (X , T ) = ⎨
⎧ (I ) X ⎪
⎪ s (X , T ) = 0 0<T< f ⎩ c1(II )(X , T ), c2(II )(X , T ) X > XA (47)
⎪ s
⎪
⎪ X X ⎧ X
⎪ s (1) ≤T < ⎪0 0 ≤ T <
⎪ f D2 ⎪ D1
⎪ s
⎪
⎨ ⎪ (J ) X (η(c1(J ), c2(J )) + s2)X + 1
⎪ s (X , T ) X X ⎪ c1 <T≤
≤T≤
⎪ 3 D2 (J )
∂f (1, c1 = c1r, c2 = c2 ) ⎪ D1 f (s2, c1(J ), c2(J ))
⎪ ∂s ⎪
⎪ ⎪
⎪ (J ) (η(c1(J ), c2(J )) + s2)X + 1
X c1(I )(X , T ) = ⎨ c1r (X , T ) ≤T
⎪s = 1 (J )
≤T<∞ ⎪ f (s2, c1(J ), c2(J ))
⎪ ∂f (1, c1 = c1r, c2 )
⎪
⎩ ∂s (45) ⎪ (η(0, c2(J )) + s 3)X + 1
⎪≤
⎪ f (s3, 0, c2(J ))
(η(0, c2(J )) + s3)X + 1
For T > , ⎪
f (s3, 0, c2(J )) ⎪ (η(0, c2(J )) + s 3)X + 1
⎪0 ≤T<∞
⎩ f (s3, 0, c2(J )) (48)
332 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335

Fig. 13. Polymer concentration profile (a), salt concentration profile (b) and saturation profile (c) for the case of 0.5 pore volumes slug size.

⎧ (I ) X (time when shocks D1, D2 and the last rarefaction of region 2 cross)
⎪ c2 0 < T < f and time T2 is greater. In these figures we notice the decrease of
⎪
⎪ s the polymer slug size with time caused by the presence of salt,
c2(I )(X , T ) = ⎨
⎪ (J ) X which will disappear when shock D2 finds the last rarefaction
⎪ c2 T > f region 3.
⎪
⎩ s (49)

⎧ X 4. Sensitivity analysis
⎪0 0≤T<
⎪ D2
⎪ In order to evaluate the effect of some parameters on the so-
⎪
⎪ X (η(0, c2(J )) + s 3)X + 1 lution, a sensitivity analysis was performed. The geometry of the
c1(II )(X , T ) = ⎨ c1r (X , T ) <T≤
reservoir and the flow rate chosen are presented in Table 1.
⎪ D2 f (s3, 0, c2(J ))
⎪ The polymer solution viscosity is given by the following ex-
⎪ (η(0, c2(J )) + s 3)X + 1 pression (John et al., 2005):
⎪0 ≤T<∞
⎪
⎩ f (s , 0, c (J ))
3 2 (50) μP = μw [1 + (81.0*c1 + 2500.0*c12 + 2700.0*c13)c2 SP ] (52)

⎧ (I ) X where mp is the polymer viscosity and mw is the viscosity of pure

⎪ c2 0 < T < f water. In all cases oil viscosity was set equal to 6 cp and water
⎪
⎪ s viscosity equal to 0.5 cp. We did not consider the salt effect on
c2(II )(X , T ) = ⎨
⎪ (J ) water viscosity so we set the polymer correlation parameter SP
X
⎪ c2 T > f equal to zero.
⎪
⎩ s (51) The water and oil phases relative permeabilities were calcu-
lated using the following power law:
Figs. 9 and 10 show the path of the solution in the plane (f
s).
The superscript (n) means a constant saturation in region n, and k rw = 0.8*s 3
the subscript (n) corresponds to the rarefaction wave saturation in k ro = 0.8*(1 − s )3 (53)
region n of plane (X
T).
The polymer and salt concentration and saturation profiles are For all cases studied, the initial polymer concentration was
shown in Fig. 11a, b and c, respectively. Time T1 is smaller than TA zero, and the initial reservoir salt concentration was 2 meq/ml.
 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335 333

Fig. 14. Polymer concentration profile (a), salt concentration profile (b) and saturation profile (c) when β3 is 10.

Case 1. : Different slug volumes order to evaluate the effect of the adsorption constants in the re-
sults. The results are presented in Fig. 14a, b and c and will be
First we compare the results for two different slug sizes,
compared with Fig. 13.
0.3 and 0.5 pore volumes. In both slugs, the polymer concentration
In this analysis, even the self-similar solutions are different (113
was 0.05 wt.%, the salt concentration was 0.17 meq/ml and the
days). When β is smaller, there is more adsorption and the front of
adsorption coefficients were β1 ¼1.5, β2 ¼ 0.5 and β3 ¼ 100 (John
the polymer slug is located at approximately 480 m, while for the
et al., 2005).
case of β set to 100, its position is beyond 600 m. This happens
Fig. 12a, b and c shows the polymer concentration, the salt
because in the case where there is more adsorption, the polymer
concentration and the saturation profiles after 113 and 3100 days
concentration front “travels” with smaller velocity. Although the
when the 0.3 pore volumes slug is injected. Similar results when
maximum concentration for the smaller β is bigger, the amount of
slug size is 0.5 pore volumes are presented in Fig. 13a, b and c.
polymer in solution in the reservoir is smaller (area below the
At 113 days both solutions are self-similar. As time increases,
concentration curve).
the maximum concentration of the bigger slug is greater. For
A similar evaluation can be made for the sweep efficiency of
0.5 pore volumes injected the concentration peak is 0.015 wt%,
the oil as previously (different slug sizes). For the case when more
while for the case of 0.3 pore volumes injected, it is 0.012 wt%
polymer is adsorbed, the oil slug behind the saturation front does
(Figs. 12a and 13a). In other words, there is less polymer in porous
not appear and the water average saturation is smaller, leading to
media.
a less efficient recovery.
Another effect of the slug size may be seen in the saturation
profile (Figs. 12c and 13c). For the case of 0.5 pore volumes in- Case 3. : Different salt concentrations in the slug
jected, the average water saturation is bigger, and an oil bank
develops behind the saturation shock. This oil bank does not ap- The last case evaluated is the effect of the salt concentration in
pear when the polymer slug is smaller, and the saturation profile the injected slug. The parameters of Case 1 for 0.5 pore volume
resembles the self-similar profile. It is clear that for these para- slug size are compared with the case when the salt concentration
meters set the oil recovery is anticipated by the increase of the in the injection fluid is changed to 1.9 meq/ml. Fig. 15a, b and c
slug size. present the results for the higher salt concentration.
Again, the self-similar solution is different (113 days). It is
Case 2. : Different adsorption parameters possible to see that the polymer adsorption increases due to the
In this case the slug size was set to 0.5 pore volumes, the initial higher salinity, and the polymer slug “travels” with a smaller ve-
and injection concentrations of polymer and salt were the same as locity in the porous media. It is interesting to note that the poly-
the previous ones, but the β3 parameter was changed to 10, in mer concentrations profiles are much more similar in this case
334 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335

Fig. 15. Polymer concentration profile (a), salt concentration profile (b) and saturation profile (c) when injected salt concentration is 1.9 meq/ml.

than in the previous one, even though the effects of changing the saturations discontinuities (s-shocks) and regions of continuously
adsorption parameter and the salt concentration on the adsorption changing saturations (s-rarefactions). A detailed derivation is ad-
are the same (adsorption increases). dressed and a sensitivity analysis was performed considering dif-
In this case an oil slug develops behind the water saturation ferent parameters and slug sizes. From the sensitivity analysis it is
shock for both salt concentrations. Again, the average water sa- possible to note that the smaller adsorption leads to more
turation is greater when there is less adsorption (the case of homogeneous profiles, greater polymer concentrations and more
smaller salt concentration in the polymer slug), indicating an effective oil displacement.
improved oil recovery. It was shown that the presence of salt can affect the efficiency
of oil recovery. The polymer concentration profile presents the
development of the chromatographic cycle in porous media, il-
5. Conclusions lustrating the appearance and disappearance of the components
injected into the reservoir rock.
In this paper, the one-dimensional two-phase flow that models
the slug injection of a polymer solution in the presence of salt is
presented. It was considered that the salt does not change the References
fractional flow curve, but influences the amount of polymer
adsorbed. Alishaeva, O.M., Entov, V.M., Zazovskii, A.F., 1982. Structures of the conjugate sa-
The introduction of a potential function based on the water turation and concentration discontinuities in the displacement of oil by a so-
lution of an active material. J. Appl. Mech. Tech. Phys. 23 (5), 675–682.
volume conservation splits the original system (water volume and Bedrikovetsky, P.G., 1993. Mathematical Theory of Oil and Gas Recovery: With
chemical components mass conservation) into an auxiliary system, applications to ex-USSR Oil and Gas Fields. Kluwer Academic, Dordrecht, p. 575.
related to thermodynamics and one hyperbolic equation, a func- Boa. P. M. F. Análise da Injeção de Água com Polímeros em Reservatórios na Pre-
sença de Sal. 2006. 78 f. Dissertação (Mestrado em Engenharia de Reservatório
tion of the transport properties and of the solution of the auxiliary e Exploração de Petróleo)-Universidade Estadual do Norte Fluminense. Rio de
system. This new independent variable allows the solution of the Janeiro. 2006.
non-self-similar problem that arises in this case of variable Borazjani, S., Bedrikovetsky, P., Farajzadeh, R., 2014. Exact Solution for Non-Self-
Similar Wave-Interaction Problem during Two-Phase Four-Component Flow in
boundary condition.
Porous Media. Abstr. Appl. Anal. 2014, 13.
It was presented the solution for the water saturation and Braginskaya, G.S., Entov, V.M., 1980. Nonisothermal displacement of oil by a solu-
polymer and salt concentrations. For the Langmuir type polymer tion of an active additive. Fluid Dyn. 15 (6), 873–880.
adsorption isotherm the solution is composed by concentration Buckley, S.E., Leverett, M.C., 1942. Mechanism of fluid displacement in sands. Trans.
AIME 146 (1), 107–116 (SPE 942107-G).
discontinuities (c-jumps), regions where the concentration chan- Cardoso, C.B., Silva, R.C.A., Pires, A.P., 2007. The Role of Adsorption Isotherms on
ges continuously (c-rarefactions) and these families interact with Chemical-Flooding Oil Recovery SPE Annual Technical Conference and
 A.S. de Paula, A.P. Pires / Journal of Petroleum Science and Engineering 135 (2015) 323–335 335

Exhibition. (SPE 109642), California, USA. Conservation Laws. J. Differ. Equ. 65, 250–268.
Claridge, E.L., Bondor, P.L., 1974. A graphical method for calculating linear dis- John, A., Han, C., Delshad, M., Pope, G.A., Sepehrnoori, K.A., 2005. New Generation
placement with mass transfer and continuously changing mobilities. SPE J. 14 Chemical-Flooding Simulator. SPE J. 10 (2), 206–216 (SPE 89436).
(6), 609–618 (SPE 4673-A). Lake, L.W., 1989. Enhanced Oil Recovery. Prentice Hall Inc., Upper Saddle River, p.
Dahl, O., Johansen, T., Tveito, A., 1992. Multicomponent cromatography in a two 550.
phase environment. SIAM J. Appl. Math. 52 (1), 65–104. Patton, J.T., Coats, K.H., Colegrove, G.T., 1971. Prediction of polymer flood perfor-
Entov, V.M., Kerimov, Z.A., 1986. Displacement of oil by an active solution with a mance. SPE J 11 (1), 72–84 (SPE 2546-PA).
nonmonotonic effect on the flow distribution function. Fluid Dyn 21 (1), 64–70. Pires, A.P., Bedrikovetsky, P.G., Shapiro, A.A., 2006. A splitting technique for ana-
Entov, V.M., Zazovskii, A.F., 1982. Displacement of oil by a solution of an active and lytical modelling of two-phase multicomponent flow in porous media. J. Pet.
a passive additive. Fluid Dyn. 17 (6), 876–884. Sci. Eng. 51, 54–67.
Fayers, F.J., Perrine, R.L., 1958. Mathematical description of detergent flooding in oil Pope, G.A., 1980. The application of fractional flow theory to enhanced oil recovery.
reservoirs. Fall Meeting of the Society of Petroleum Engineers of AIME, Hous- SPE J. 20 (3), 191–205 (SPE 7660-PA).
ton, USA (SPE 1132-G). Rhee, H.K., Aris, R., Amundson, N.R., 1970. On the Theory of Multicomponent
Helfferich, F.G., 1981. Theory of multicomponent, multiphase displacement in Chromatography. Phil. Trans. Royal Soc. Ser. A 267, 419–455.
porous media. SPE J. 21 (1), 51–62 (SPE 8372-PA). Ribeiro, P.M., Pires, A.P., 2008. The displacement of Oil by Polymer Slugs Con-
Hirasaki, G.J., 1981. Application of the theory of multicomponent, multiphase dis- sidering Adsorption Effects. SPE Annual Technical Conference and Exhibition,
placement to three-component, two-phase surfactant flooding. SPE J. 21 (2), Denver, USA (SPE 115272).
191–204 (SPE 8373-PA). Welge, H.J., 1952. A simplified method for computing oil recovery by gas or water
Isaacson, E., Temple, B., 1986. Analysis of a Singular Hyperbolic System of drive. J. Pet. Technol. 4 (4), 91–98 (SPE 124-G).