Immiscible Displacement of
Non-Newtonian Fluids in Communicating
Stratified Reservoirs
Noaman El-Khatib, SPE, Sudan U. of Science and Technology
Summary
The displacement of non-Newtonian power-law fluids in commu-
nicating stratified reservoirs with a log-normal permeability dis-
tribution is studied. Equations are derived for fractional oil recov-
ery, water cut, injectivity ratio, and pseudorelative permeability
functions, and the performance is compared with that for Newto-
nian fluids. Constant-injection-rate and constant-total-pressure-
drop cases are studied.
The effects of the following factors on performance are inves-
tigated: the flow-behavior indices, the apparent mobility ratio, the
Dykstra-Parsons variation coefficient, and the flow rate. It was
found that fractional oil recovery increases for nw>no and de-
creases for nw<no, as compared with Newtonian fluids. For the
same ratio of nw/no, oil recovery increases as the apparent mobility
ratio decreases. The effect of reservoir heterogeneity in decreasing
oil recovery is more apparent for the case of nw>no. Increasing the
total injection rate increases the recovery for nw>no, and the op-
posite is true for nw<no. It also was found that the fractional oil
recovery for the displacement at constant total pressure drop is
lower than that for the displacement at constant injection rate, with
the effect being more significant when nw<no.
Introduction
Many of the fluids injected into the reservoir in enhanced-oil-
recovery (EOR)/improved-oil-recovery (IOR) processes such as
polymer, surfactant, and alkaline solutions may be non-Newtonian;
in addition, some heavy oils exhibit non-Newtonian behavior.
Flow of non-Newtonian fluids in porous media has been stud-
ied mainly for single-phase flow. Savins (1969) presented a com-
prehensive review of the rheological behavior of non-Newtonian
fluids and their flow behavior through porous media. van Poollen
and Jargon (1969) presented a finite-difference solution for tran-
sient-pressure behavior, while Odeh and Yang (1979) derived an
approximate closed-form analytical solution of the problem.
Chakrabarty et al. (1993) presented Laplace-space solutions for
transient pressure in fractal reservoirs.
For multiphase flow of non-Newtonian fluids in porous media,
the problem was considered only for single-layer cases. Salman
et al. (1990) presented the modifications for the Buckley-Leverett
frontal-advance method and for the JBN relative permeability
method for non-Newtonian power-law fluid displacing a Newto-
nian fluid. Wu et al. (1992) studied the displacement of a Bingham
non-Newtonian fluid (oil) by a Newtonian fluid (water). Wu and
Pruess (1998) introduced a numerical finite-difference solution for
displacement of non-Newtonian fluids in linear systems and in a
five-spot pattern. Yi (2004) developed a Buckley-Leverett model
for displacement by a Newtonian fluid of a fracturing fluid having
a Herschel-Bulkley rheological behavior. An iterative procedure
was used to obtain a solution of the model.
The methods available in the literature to predict linear water-
flooding performance in stratified reservoirs are grouped into two
categories depending on the assumption of communication or no
communication between the different layers.
Copyright © 2006 Society of Petroleum Engineers
This paper (SPE 93394) was first presented at the 2005 SPE Middle East Oil and Gas Show
and Conference, Manama, Bahrain, 12–15 March, and revised for publication. Original
manuscript received for review 5 January 2005. Revised manuscript received 6 April 2006.
Paper peer approved 12 April 2006.
356
In the case of noncommunicating systems, no vertical cross-
flow is permitted between the adjacent layers. The Dykstra-
Parsons (1950) method is the basis for performance prediction in
noncommunicating stratified reservoirs.
A model for communicating stratified reservoirs was presented
by Hiatt (1958). This model assumes complete crossflow between
layers to keep the pressure gradient the same in all layers (vertical
equilibrium) at any distance. Warren and Cosgrove (1964) applied
the Hiatt model to a system with log-normal permeability distri-
bution and normal porosity distribution. El-Khatib (1999) pre-
sented a closed-form analytical solution for communicating sys-
tems with log-normal permeability distribution. Hearn (1971) used
the same Hiatt model to develop expressions for pseudorelative
permeabilities that can be used to reduce a 3D model to a 2D areal
model with average (pseudo) functions for the vertical direction.
To the best of the author’s knowledge, no analytical or numeri-
cal models are available for the displacement of non-Newtonian
fluids in multilayer stratified reservoirs. In this study, an analytical
model will be presented to study the performance of immiscible
non-Newtonian power-law fluids in stratified reservoirs. Although
the model is applicable for any permeability distribution, a strati-
fied system with a log-normal permeability distribution is studied
because its behavior is well documented in the literature (Warren
and Cosgrove 1964; El-Khatib 1999) for Newtonian fluids.
Theoretical Considerations
Rheological Model. Rheological models or equations of state for
fluids describe the dependence of the shear stress T on the shear
rate ␥˙. The most general rheological model is the Herschel-Bulkley
model (1926), which is expressed in the following form:
T = T0 + H␥˙ n, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1)
where H is the consistency index and n is the flow-behavior index.
Most of the familiar and commonly used rheological models can
be deduced from this model by assigning specific values for the
parameters T0 and n. For n⳱1, we get the equation for the Bing-
ham fluids, and for T0⳱0, we get the widely used power-law
model that is used in this study.
T = H␥˙ n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2)
For T0⳱0 and n⳱1, Eq. 1 reduces to the Newtonian fluids case,
with H being the viscosity ␮:
T = ␮␥˙ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (3)
Capillary-Tube Model. The equation for the flow of a power-law
fluid in a bundle of capillary tubes is given by the equation (Teeuw
and Hesselink 1980)
冉 冊 冉 冊
n+1
1
4n ␾ r n dp n
v= − . . . . . . . . . . . . . . . . . . . . . (4)
3n + 1 n+1 2n+1 1 dx
T 2n 2 n Hn
For a Newtonian fluid, n⳱1, Eq. 4 becomes
v=
␾ r2
8T ␮
−
dp
dx冉 冊
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (5)
Comparing with Darcy’s law, it follows that the permeability
for Newtonian-fluid flow is given by
August 2006 SPE Reservoir Evaluation & Engineering
 ␾ r2 • The relative permeability characteristics (Swi , Sor , k•rw , and
k= , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (6) k•ro) are the same for all layers.
T 8
• The porosity is assumed the same in all layers.
where T is the tortuosity of the porous medium. • Adsorption of fluids on the solid surface of the porous media
Using this expression, Eq. 4 can be expressed as is negligible, so effects such as porosity and permeability alteration
or inaccessible pore volume are not considered.

冉 冊冉 冊 冉 冊
n+1
n−1 1 As in the conventional prediction methods for stratified reser-
4n ␾ 2n k 2n dp n
v= − . . . . . . . . . . . . . . . . . . . . . . (7) voirs, the layers are ordered in a decreasing order of absolute
3n + 1 2 1 dx horizontal permeability, with each layer i having a thickness ⌬hi
Hn and permeability ki. In the horizontal direction between injection
and production faces, the system is divided into N+1 zones sepa-
Darcy-Law Analogy. Different approaches were used to trans- rated by the displacement fronts. Zone 0 is at the production end,
form Eq. 7 into a form similar to Darcy’s law by introducing an while Zone N is at the injection end, with N being the number of
apparent-viscosity term. Bird et al. (1960) used the form layers. All layers in Zone 0 (zero zones flooded) are at initial

vn =
k
␮eff 冉 冊
−
dp
dx
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (8)
conditions, while all layers in Zone N (N layers flooded) are at the
residual oil saturation. At Zone j (j layers flooded), Layers 1
through j are at residual oil saturation, while Layers j+1 through N
It is clear that the effective viscosity as defined in this equation are at initial conditions (see Fig. 1).
will not be a function of the rheological-model parameters H and n
only. It also will depend on the porosity and permeability of the rock. Fractional-Flow Formula. At the time of water breakthrough in
Other investigators such as Gogarty et al. (1972) and Cannella the jth layer, the fraction of water flowing at the outlet boundary
et al. (1988) suggested the form (water cut) fwj, as derived in Appendix A, is given by

v=
k
− 冉 冊
dp
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (9) fwj =
␥Cwj
. . . . . . . . . . . . . . . . (14)

冉 冊
␮ap dx fwj Qt ␮wap
nw
−1
no
As can be seen from Eq. 7, the apparent viscosity in this case ␥Cwj + 共Cot − Coj兲 •
Wkrw Cwj
also will be dependent on the pressure gradient dp/dx in addition to
the rock and fluid properties. It should be noted that Eq. 14 is implicit in fwj. It may be written
In this work, we separate the effect of rock properties and in the following form:

冉 冊
pressure gradient from the definition of the apparent fluid viscosity nw
and introduce the following definition for ␮ap, which depends only 共Cot − Coj兲 Qt ␮wap no
− 1 nw
fwj + f no = 1. . . . . . . . . . . . . . . . . . (15)
on the rheological-model parameters H and n. ␥Cwj •
Wkrw Cwj wj

␮ap = 2 2n
n−1

冉 冊
3n + 1
4n
1
H n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (10) Eq. 15 can be solved iteratively for fwj. This can be performed for
j=1 . . . N−1. It is clear that fwo⳱0 and fwN⳱1.
With this definition, the Darcy-law analogy for power-law flu-
Fractional Oil Recovery. The equations for oil recovery will be
ids may be written as
the same as those for Newtonian fluids except for the expression

冉 冊
1 for the fractional-flow formula.
k dp n
v=␣ − , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (11) Once the function fwj is evaluated, the equations for communi-
␮ap dx cating stratified reservoirs as given in the literature can be used
where the non-Newtonian flow coefficient ␣ is defined as (Hiatt 1958; Warren and Cosgrove 1964; El-Khatib 1999):

冉冊
1−n
k 2n
␣= . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (12)
␾
For multiphase flow, we may write

冉 冊
1
kkri dp ni
vi = ␣i − , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (13)
␮iap dx
where kri is the relative permeability to the specified phase i (oil,
gas, or water).
The definition of the apparent viscosity in this formulation is
different from those of Bird et al. (1960), Gogarty et al. (1972),
and Cannella et al. (1988) in that only the fluid parameters H and
n are used in the definition. The rock parameters k and ␾ and the
pressure gradient dp/dx are not used. The effect of the rock prop-
erties on the flow is introduced in the non-Newtonian-flow coef-
ficient ␣. Although Eq. 7 is the same in all models, this formula-
tion is more convenient in handling two-phase-flow problems.

Development of the Model Equations

The following assumptions are made:
• The system is linear, horizontal, and of constant thickness.
• The flow is isothermal, and rock and fluids are incompressible.
• The initial fluid distribution is uniform with irreducible
water saturation.
• The displacement is pistonlike, with only residual oil behind
and initial conditions ahead of the displacement front. Fig. 1—Stratified system showing zones and layers before wa-
• Capillary and gravity forces are negligible. ter breakthrough in the first layer.

August 2006 SPE Reservoir Evaluation & Engineering 357

 The fractional oil recovery at time of breakthrough in the jth Curves for oil and water pseudorelative permeabilities can be used
layer is given by in reservoir simulation, collapsing the vertical direction into a
single block.
Rj = hDj + 共1 − fwj兲␶j , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (16) The pseudofractional-flow curve fw can be calculated from
where the dimensionless time ␶ is the injected volume relative to pseudorelative permeabilities by use of the relation
the ultimate oil recovery, and the fractional oil recovery is also Qwj 1
relative to the ultimate oil recovery so that at j=N, R⳱1. fwj = = . . . . . . . . . . . (30)

冉 冊
The dimensionless time at breakthrough in layer j is given by Qwj + Qoj ␣o k̃ro ␮wap dp
1 1
−
no nw
1+ −
1 ␣w k̃rw ␮oap dx
␶j = , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (17)
⌬fwj
Injectivity Variation. As the displacement proceeds and more of
⌬hDj
the displacing fluid enters into the formation, either the injection
where rate, the total pressure drop, or both will change. The variation is
expressed in terms of the injectivity ratio Ir or the resistivity ratio
⌬fwj = fwj − fw共j−1兲, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (18) ␤, which is the inverse of the injectivity ratio and is defined as
with fw0⳱0. 1 共⌬pt /Q not 兲j
The dimensionless thickness ⌬hDj and the cumulative dimen- ␤j = = . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (31)
sionless thickness hDj are defined as Ir 共⌬pt /Q no
t 兲 in

⌬hDj = ⌬hj Ⲑ ht . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (19) The expressions for the resistivity ratio ␤ are derived in Ap-
pendix B and given below.
j
Before water breakthrough in the first layer,
and hDj = 兺 ⌬h . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (20)
冉 冊
Di ,
␮wap

冤 冥
i=1 nw

兺冉 冊 冉 冊
N
where ht is the total thickness of the system. fw1 Wk•rw fwj nw ⌬fwj

冉 冊
␤=1−␶ − Q nw−no ⌬
⌬hD1 t
␮oap no
j=1 Cwj ⌬hDj
Pseudorelative Permeability Functions. For non-Newtonian-
fluid flow in a homogeneous system having a uniform permeabil- Wk•roCot
ity equal to the average permeability of the stratified system k and . . . . . . . . . . . . . . . . . . . . . . . . . . (32)
total thickness ht, Eq. 13 expressed for the water phase can be
written as At time of water breakthrough in layer j,

冉 冊
1
冉 冊 ␮wap nw

兺冉 冊 冉 冊
khtWk̃rwj dp nw N
Qwj = ␣w − , . . . . . . . . . . . . . . . . . . . . . . . . . . (21)
•
Wkrw fwi nw ⌬fwi

冉 冊
␮wap dx ␤j = ␶j Q nw−no ⌬ . . . . . . . . (33)
t
␮oap no
i=j Cwi ⌬hDi
where the average (pseudo) non-Newtonian-flow factor ␣w is de-
fined as Wk•roCot
j It is to be noticed from Eqs. 32, 33, and 14 that for the constant-
兺␣
i=1
wiki⌬hi
injection-rate case, the calculation of the injectivity (resistivity)
ratio is straightforward. In this case, the value of fwj needs to be
␣w = j
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (22) calculated only once for each j (j⳱1, 2, . . . N). The resistivity ratio
兺 k ⌬h
i=1
i i
is linear with the dimensionless time ␶ before water breakthrough
in the first layer and linear with ␶ between times of water break-
through in the successive layers. On the other hand, for the case of
Comparing with Eq. A-2 of the stratified system, we get constant total pressure drop ⌬pt, the total injection rate Qt will vary
Cwj with time. In this case, as shown by Eq. 14, the fractional flow fw
k̃rwj = krw
•
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (23) will be time dependent (i.e., it will depend on the location of the
Cwt displacement front in the different layers). As realized from Eqs.
Similarly for the oil phase using Eq. A-4, 32 and 33, the terms inside the summation involving fwj and ⌬fwj
will change. At any given value of the dimensionless time ␶, there
Cot − Coj will be a specific fractional-flow curve. The same also applies for
k̃roj = k•ro , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (24) times after water breakthrough in the first layer. An iterative pro-
Cot
cedure must be used to estimate both Qt and fwj simultaneously at
where k̃rwj and k̃roj are the pseudorelative permeabilities for water different dimensionless times ␶ before water breakthrough in the
and oil, respectively. first layer and at the dimensionless times of water breakthrough in
The dimensionless pseudowater saturation at the time of water the successive layers (j⳱1, 2, . . . , N). At the time of water break-
breakthrough in layer j is given by through in Layer j, only values of fwi for i=j, j+1, . . . . . . , N−1
need to be calculated (fwN⳱1). It is clear from the above discus-
SDj = hDj . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (25)
sion that for the case of constant total pressure drop, the resistivity
The dimensionless saturation SDj is defined as ratio is not linear with the dimensionless time ␶, neither before
water breakthrough in the first layer nor between times of water
SDj = 共S̃w − Swi兲/⌬Sw, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (26) breakthrough in the successive layers.
S̃w = Swi + ⌬Sw · hDj , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (27) Computational Procedure
⌬Sw = 1 − Swi − Sor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (28) The system parameters that are needed to perform the computa-
tions include the number of layers N; the values of absolute hori-
Unlike the case of pseudorelative permeabilities for Newtonian fluids, zontal permeability ki and thickness ⌬hi for each layer; the poros-
it can be seen from Eqs. 23 and 24 that for non-Newtonian fluids, ity ␾; the rheological-model parameters Ho, no, Hw, and nw; and
the relative permeability endpoint characteristics Swi, Sor, k•rw, and
k̃rw k̃ro
+ ⫽ 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (29) k•ro. The displacement is specified either at constant injection rate
•
krw k•ro Qt or at constant total pressure drop ⌬pt.

358 August 2006 SPE Reservoir Evaluation & Engineering

 The layers are arranged in order of decreasing permeability, Results and Discussion
and the following terms are calculated for each layer: The developed method was applied to a hypothetical stratified
• Oil and water non-Newtonian-flow coefficients ␣oj and ␣wj reservoir of 20 layers with the permeability generated from a log-
are determined by use of Eq. 12. normal distribution with VDP⳱0.5. The porosity and endpoint
• For dimensionless thickness ⌬hDj and dimensionless cumu- relative permeabilities are assumed the same for all layers.
lative thickness hDj, Eqs. 19 and 20 are used. The log-normal distribution of permeability is given by
• For non-Newtonian formation capacity Cwj and Coj, Eqs. A-3
and A-5 are used.
• For pseudorelative permeabilities k̃rwj and k̃roj, Eqs. 23 and
24 are used; for dimensionless pseudowater saturation SDj, Eq. 25
P共k兲 = 0.5 + 0.5 erf 冋公 册
ln共k/km兲
2␴k
, . . . . . . . . . . . . . . . . . . . . . . . (38)

is used.
where P(k) is taken as the relative cumulative thickness (fraction
These values are used both for cases of constant injection rate
of the total thickness) with permeability less than k.
and for cases of constant total pressure drop.
Noting that P(kj)⳱1−hDj, Eq. 38 may be rearranged as

关公2␴ erf 兴
Constant Injection Rate Qt. The following procedure is performed: −1
Calculate fwj for each layer (j⳱1, 2, . . . , N−1) from Eq. 15 k = km exp k 共1 − 2hDj兲 . . . . . . . . . . . . . . . . . . . . . (39)
using the following Newton-Raphson procedure:
To investigate the effect of the different parameters that affect

冉 冊
nw the performances, 19 different combinations of the non-Newtonian
共Cot − Coj兲 Qt ␮wap no
−1 nw
flow-behavior indices nw and no, the Dykstra-Parsons variation
g共fwj兲 = fwj + fwj no − 1 = 0. . . . . . (34)
␥Cwj •
Wkrw Cwj coefficient VDP, the total flow rate Qt, and the consistency indices
Hw and Ho (as shown in Table 1) are considered. Values of the

冉 冊
nw apparent mobility ratio for the different cases as given by Eq. A-8
nw 共Cot − Coj兲 Qt ␮wap no
−1 nw
−1
g⬘共fwj兲 = 1 + fwj wjno . . . . . . . . (35) are calculated and listed in Table 1. The first case representing the
no ␥Cwj •
Wkrw Cwj Newtonian fluids is used for comparison.

g共f wj
k
兲 Effect of Non-Newtonian Flow-Behavior Indices. The basic
k+1
f wj = f wj
k
− . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (36) equations of the model are Eq. 14 for the water cut fw and Eq. 16
g⬘共f wj 兲
k
for the fractional oil recovery R, with the formation capacity for
water and oil given by Eqs. A-3 and A-5, respectively. All these
The iteration is continued until Eq. 34 is satisfied within a speci- equations include the ratio nw/no explicitly except Eq. 16, where
fied tolerance. Inspection of Eq. 35 shows that g⬘ is always posi- the ratio is included implicitly in the terms fw and ␶. It is clear from
tive (>1). This guarantees the convergence of the Newton-Raphson these equations that for nw/no⳱1 (i.e., for Newtonian fluids or
iteration scheme because the function g does not have extreme non-Newtonian fluids with the value of the flow-behavior indices
points. Furthermore, the procedure has a quadratic convergence. for the displacing and displaced fluids), the model equations re-
Before Water Breakthrough in the First Layer. During this duce to those for the case of displacement by Newtonian fluids.
time, R=␶ and fw⳱0. The following calculations are performed to This case is well documented in the literature by Hiatt (1958),
obtain the total pressure drop ⌬pt. Hearn (1971), Warren and Cosgrove (1964), and El-Khatib (1999).
For X1⳱0.1, 0.2, . . . . . . . , 1.0, calculate These conditions are represented by Case 1 (Table 1) and are
• The dimensionless time ␶ from Eq. B-6. included for comparison.
• The total pressure drop ⌬pt from Eq. B-7. To study the effect of the flow-behavior indices, Cases 1
After Water Breakthrough in the First Layer. At times of through 5 are considered. The reference Case 1 is for Newtonian
water breakthrough in the successive layers j (j⳱1, 2, . . . , N), fluids (nw/no⳱1). Cases 2 and 3 are for a Newtonian fluid (nw⳱1)
• The values of fwj, SDj, k̃rwj, and k̃roj are already calculated. displacing a non-Newtonian fluid (no⳱0.8, 1.2), while Cases 4
• The dimensionless time of breakthrough in layer j is esti-
mated from Eq. 17.
• The fractional oil recovery Rj is obtained from Eq. 16.
• The total pressure drop ⌬pt is calculated from Eq. B-9.

Constant Total Pressure Drop ⌬pt. The same procedure as with

constant injection rate is followed, with the following modifications:
• The value from the previous time is used as an initial guess
for Qt at the start of the new time. At the first time, we choose Qtin
calculated from Eq. B-10.
• The values of fwj are calculated for j⳱1, 2, . . . , N–1 before
water breakthrough in the first layer and for j, j+1, . . . , N–1 at
times of water breakthrough in layer j using the iterative procedure
outlined by Eqs. 34 through 36 using the assumed value of Qt.
After all values of fwj are obtained, Qt is calculated from Eq. B-7
before water breakthrough in the first layer and from Eq. B-9 at the
time of water breakthrough in layer j. Eq. B-9 is explicit in Qt and
can be solved directly. An iteration procedure is needed to solve
Eq. B-7 for Qt. Using a Newton-Raphson procedure, Eq. B-7 can
be written in the following form with y=Qtno:
nw
g共y兲 = ay + by no − ⌬pt = 0. . . . . . . . . . . . . . . . . . . . . . . . . . . . (37)

Inspection of Eq. 37 indicates that it does not have any maxima or

minima (g⬘⫽0), and, hence, the iteration procedure converges qua-
dratically. After solving for y, the value of Qt is obtained (⳱y1/no).
A computer program was written to compute the performance
of the stratified system for a given set of model parameters.

August 2006 SPE Reservoir Evaluation & Engineering 359

Fig. 2—Effect of flow-behavior indices on performance (R vs. ␶). Fig. 3—Effect of flow-behavior indices on performance (fw vs. R).

and 5 are for a non-Newtonian fluid (nw⳱0.8, 1.2) displacing a It is to be noted that at the time of breakthrough in the last layer,
Newtonian fluid (no⳱1). the pressure-drop ratio ⌬Pfin/⌬Pin is not equal to the inverse of the
Fig. 2 shows the fractional oil recovery R as a function of modified mobility ratio, as would be the case for Newtonian fluids.
dimensionless time ␶ for the five cases. For the case of a Newto- As can be derived from Eqs. B-9 and B-10, this ratio is given by

冉 冊
nian fluid displacing a non-Newtonian fluid, it is seen that the oil
Qt ␮wap nw
recovery is higher for no⳱0.8 and lower for no⳱1.2 as compared
with the Newtonian-fluids case. For the case of a non-Newtonian ⌬pt fin •
Wkrw Cwt

fluid (water) displacing a Newtonian fluid (oil), the oil recovery is
higher for nw⳱1.2 and lower for nw⳱0.8 as compared with the
Newtonian-fluids case. It is clear that the recovery is high when
nw>no and low when nw<no, as compared with the Newtonian-
fluids case (nw/no⳱1).
Fig. 3 shows the water cut fw vs. the fractional oil recovery R.
The results indicate delayed water breakthrough with lower water
cut for nw>no and earlier water breakthrough with higher water cut
for nw<no as compared with the Newtonian-fluids case nw/no⳱1.
This behavior can be explained by investigating Eq. 15, which can
be written as
冉 冊
= . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (41)
⌬pt in Qt ␮oap no
Wk•roCot
This ratio is equal to 1/␥ only if nw/no⳱1.
Effect of Apparent Mobility Ratio. The apparent mobility ratio ␥
as given by Eq. A-8 is calculated for all cases and listed in Table 1.
Fig. 7 shows the performance in terms of water cut fw vs.
fractional oil recovery R for Cases 1, 6, 7, 16, and 17. As stated
previously, the fractional recovery increases and the water cut
decreases for nw>no. It may appear that the ratio nw/no controls the

冉 冊
nw
Qt ␮wap no

共Cot − Coj兲 •
Wkrw Cwj nw

冉 冊
fwj + no = 1. . . . . . . . . . . . . . . . . . . (40)
f wj
␥Cwj Qt ␮wap
•
Wkrw Cwj

It is clear that the coefficient of the second term on the left side of
the equation increases with the increase of nw/no. Hence, fw must
decrease to keep the sum of the two terms on the left side of the
equation constant at the value of unity. The decrease in fw is a
favorable indication of reservoir performance.
Fig. 4 is a plot of oil and water pseudorelative permeability
curves. The results indicate that the water pseudorelative perme-
ability is influenced only by nw and not by no, while the oil pseu-
dorelative permeability is influenced only by no and not by nw. The
water pseudorelative permeability decreases as nw increases, while
the oil pseudorelative permeability increases as no decreases. This
will result in a decrease in fractional flow for increasing nw/no, as
can be realized from Eq. 30. This behavior can be seen in Fig. 5,
which shows the fractional-flow curves vs. the dimensionless
(pseudo) water saturation.
Fig. 6 is a plot of the total-pressure-drop ratio ⌬P/⌬Pin. It can
be seen that the ratio drops from unity at the start of displacement
to constant values at the time of breakthrough in the last layer. The
ratio is, however, higher for nw>no and lower for nw<no as com- Fig. 4—Effect of flow-behavior indices on pseudorelative per-
pared to the Newtonian-fluids case. meability curves.

360 August 2006 SPE Reservoir Evaluation & Engineering

 Fig. 6—Effect of flow-behavior indices on pressure-drop ratio.
Fig. 5—Effect of flow-behavior indices on fractional-flow
curves.
decreases as polymer concentration increases. The increase in the
performance. Three different cases with the same value of the ratio consistency index H, however, is much more than the decrease in
nw/no⳱1.5 but with different individual values for each nw and no the flow-behavior index. This behavior causes a decrease in the
(1.2/0.8, 1.35/0.9, and 1.8/1.2) were investigated. As shown in apparent mobility ratio and, hence, an increase in the fractional oil
Fig. 7, the performance of the three cases is quite different. The recovery with the increase of polymer concentrations for both sets
three cases have largely different apparent mobility ratios (3.59, of polymers. The consistency index H for the PEO polymers is
2.89, and 1.87, respectively). It is clear that the recovery increases much lower than that for the XAN polymer at all concentrations.
and the water cut decreases as the apparent mobility ratio de- This would result in a higher apparent mobility ratio and lower
creases. The results of Cases 1, 6, and 7 indicate that the effect of fractional oil recovery for the PEO polymers as compared with the
the apparent mobility ratio is secondary compared to the effect of XAN polymers. However, because of the high values of the flow-
the flow-behavior indices. Although the apparent mobility ratio for behavior index, the fractional oil recovery for PEO polymers does
the 0.8/1.2 case is 1.99 compared to the value of 2.5 for the 1/1 not differ too much from that for the XAN polymers. These results
case, the performance deteriorates because nw<no. On the other are in agreement with the results of the developed model.
hand, the performance improves for the 1.2/0.8 case over that of
the 1/1 case despite the increase in the apparent mobility ratio from Effect of Reservoir Heterogeneity. The heterogeneity of the res-
2.5 to 3.59. This is also attributed to nw being greater than no ervoir is described by means of the standard deviation of the per-
despite the higher apparent mobility ratio. meability distribution ␴k or the Dykstra-Parsons coefficient of varia-
It can be seen from Eq. 15 or Eq. 40 that at the same value of tion VDP, which is related to ␴k by the relation ␴k=ln [1/(1 – VDP)].
nw/no, a decrease in the value of ␥ will cause the coefficient of the A value of VDP⳱0 represents a homogeneous reservoir (constant
second term on the left side of the equation to increase, thus k), while a value of 1 represents a totally heterogeneous reservoir.
resulting in a decrease in fw to keep the sum of the two terms on
the left side of the equation constant at a value of unity.
As indicated by Eq. A-8 and Eq. 10, the value of the apparent
mobility ratio is determined by the flow-behavior indices nw and no
and the consistency indices Hw and Ho. The above cases investi-
gated the effect of apparent mobility ratio controlled by the non-
Newtonian flow-behavior indices. The consistency indices were
taken as 1 and 5 for water and oil, respectively. For polymer
solutions, the flow-behavior index as reported by many investiga-
tors (Teeuw and Hesselink 1980; Cannella et al. 1988) is usually
less than 1 (usually 0.3 to 0.7 and decreasing with increased poly-
mer concentration). This will result in unfavorable displacement
conditions (nw<no). The consistency factor for polymer solutions,
however, is much higher than that of water (10 to 500 and increas-
ing with increased polymer concentration), which will result in a
lower apparent mobility ratio and favorable displacement conditions.
Gleasure and Phillips (1990) reported that some synthetic polymer
solutions such as polyethylene oxide (PEO) and partially hydrolyzed
polyacracrylamides (PHPA) have a shear thickening behavior in core
flow (nw>1). This, combined with the high value of the consistency
index, will result in very favorable displacement conditions.
The results reported by Gleasure (1990) are in agreement with
the results of the developed model. Two sets of polymers (XAN
and PEO) at polymer concentrations of 500, 1,500, and 2,500 ppm
were used to displace oil in unconsolidated porous media. The
consistency index H increases while the flow-behavior index nw Fig. 7—Effect of apparent mobility ratio on performance (fw vs. R).

August 2006 SPE Reservoir Evaluation & Engineering 361

 Fig. 9—Effect of total injection rate on performance (R vs. ␶).
Fig. 8—Effect of reservoir heterogeneity on performance (fw vs. R).

To investigate the effect of reservoir heterogeneity, permeabil- Newtonian fluids, the fw/Sw curve remains the same during the
ity distributions were obtained from a log-normal distribution with displacement process.
VDP values of 0.25, 0.5, and 0.75 and nw/no values of 0.8/1, 1/1, Fig. 11 shows the fractional oil recovery vs. the dimensionless
and 1/0.8 (Cases 1, 2, 4, and 10 through 15). The results for these time, and Fig. 12 shows the water cut vs. the fractional oil recovery
cases are shown in Fig. 8 in terms of water cut fw vs. fractional oil for displacement at constant total pressure drop in comparison with
recovery R. As can be expected, heterogeneity (higher VDP) tends those at a constant injection rate. The figures show a decrease in
to have a negative effect on the performance. For any pair of fluids fractional oil recovery and an increase in water cut for displace-
with a fixed value of nw/no, the value of fw increases and the value ment at constant total pressure drop. This effect is more noticeable
of R decreases at higher values of VDP. This effect is more notice- when nw<no than when nw>no. This behavior may be explained by
able, however, when nw/no>1. the large increase in the injection rate for nw<no as the displace-
It also can be seen that for any value of VDP , water cut fw ment proceeds, as seen in Fig. 13.
decreases and the fractional oil recovery R increases for nw/no>1
and vice versa, as was noticed before. This effect, however, is Conclusions
more apparent in the less-heterogeneous reservoirs (VDP⳱0.25) 1. A mathematical model is developed for power-law non-
than in the more-heterogeneous reservoirs (VDP⳱0.75). Newtonian-fluid displacement in communicating stratified res-
ervoirs. Equations are derived for fractional oil recovery, water
Effect of Total Injection Rate. To investigate the effect of the cut, injectivity ratio, and pseudorelative permeabilities. Cases of
total injection rate on the performance, six cases were considered. displacement at constant injection rate and at constant total pres-
First, Cases 8, 2, and 9 (with values for Qt of 50, 100, and 200,
respectively, and an nw/no value of 0.8/1), and then Cases 18, 4,
and 19 (with values for Qt of 50, 100, and 200, respectively, and
an nw/no value of 1/0.8) were investigated. All other parameters are
the same for all cases. Fig. 9 shows the fractional oil recovery R as
a function of dimensionless time ␶ for the three rates. It is clear that
the increase in the total injection rate results in an increase in the
fractional oil recovery R when nw>no and a decrease in the frac-
tional oil recovery when nw<no. Again, this behavior can be ex-
plained by Eq. 15. The total injection rate Qt is raised to the power
of nw/no−1. For nw/no>1, this power is positive, so the coefficient
of the second term on the left side of the equation increases as Qt
increases. This causes fw to decrease to keep the sum at a value of
1. On the other hand, for nw/no<1, Qt is raised to a negative power,
and the coefficient of the second term decreases as Qt increases,
resulting in an increase of fw.

Displacement at Constant Pressure Drop. The procedure out-

lined before for injection at constant total pressure drop was ap-
plied for nw/no values of 0.8/1.2 and 1.2/0.8. Fractional-flow
curves were calculated at values of dimensionless distance of the
displacement front in the first (most-permeable) layer X1 of 0.2,
0.4, 0.6, 0.8, and 1 and at the times of water breakthrough in the
successive 20 layers of the stratified system. Fig. 10 shows the
fractional-flow curves as a function of the pseudo (average) water
saturation. It is seen that fw increases as X1 increases and as water
breaks through the successive layers. This behavior is more effec- Fig. 10—Fractional-flow curves for constant-pressure-drop
tive for the nw/no case of 0.8/1.2 than for the 1.2/0.8 case. For displacement.

362 August 2006 SPE Reservoir Evaluation & Engineering

Fig. 11—Performance (R vs. ␶) for constant-pressure-drop
displacement.
sure drop are considered. An iterative procedure is applied to
solve for the implicit equations obtained.
2. The performance is controlled mainly by the relative values of
the rheological-model indices nw and n0. The performance im-
proves (higher recovery and lower water cut) over the Newto-
nian case for nw>no, and the opposite occurs for nw<no. For the
same nw/no ratio, the performance is controlled mainly by the
apparent mobility ratio. The recovery increases and the water
cut decreases as the apparent mobility ratio decreases.
3. The recovery increases and the water cut decreases as the total
injection rate is increased for nw>no; the opposite is true for nw<no.
4. Increased reservoir heterogeneity (high VDP) results in lower oil
recovery and higher water cut. This behavior becomes more
apparent for nw>no.
5. Fractional oil recovery is lower and water cut is higher for
displacement at constant pressure drop as compared to those at
constant injection rate, with other parameters being the same.
This effect is more noticeable for nw<no.
Fig. 13—Flow-rate ratio for constant-pressure-drop displacement.
August 2006 SPE Reservoir Evaluation & Engineering
Fig. 12—Performance (fw vs. R) for constant-pressure-drop
displacement.
6. The displacement at constant pressure drop does not have a
single fw/Sw curve. The values are time dependent, and fw in-
creases as displacement progresses. This effect is more notice-
able for nw<no.
7. The best performance may be obtained if the displacing fluid is
shear thickening (nw>1) and the displaced fluid is Newtonian or
shear thinning (noⱕ1).
Nomenclature
C ⳱ dimensionless formation capacity
fw ⳱ water cut, dimensionless
hD ⳱ fraction of total thickness
ht ⳱ total formation thickness, ft (m)
H ⳱ consistency index in power-law model
Ir ⳱ injectivity ratio, dimensionless
k ⳱ absolute horizontal permeability, md (␮m2)
km ⳱ mean of log-normal permeability distribution, md (␮m2)
k•ro ⳱ oil relative permeability at irreducible water saturation,
dimensionless
k̃ro ⳱ pseudorelative permeability for oil, dimensionless
k•rw ⳱ water relative permeability at residual oil saturation,
dimensionless
k̃rw ⳱ pseudorelative permeability for water, dimensionless
L ⳱ total length of flow system, ft (m)
n ⳱ flow-behavior index in power-law model, dimensionless
N ⳱ total number of layers
Q ⳱ flow rate, B/D (m3/s)
r ⳱ radius of capillary tube, ft (m)
R ⳱ fractional oil recovery, dimensionless
SD ⳱ dimensionless water saturation
Sor ⳱ residual oil saturation, fraction
Sw ⳱ water saturation, fraction
S̃w ⳱ pseudowater saturation, fraction
Swi ⳱ initial water saturation, fraction
T ⳱ tortuosity, dimensionless
v ⳱ average velocity, ft/D (m/s)
VDP ⳱ Dykstra-Parsons coefficient, dimensionless
W ⳱ width of layers, ft (m)
X ⳱ dimensionless distance, dimensionless
␣ ⳱ non-Newtonian-flow parameter
␤ ⳱ resistivity ratio, dimensionless
363
 ⌬P ⳱ pressure drop, psi (kPa) van Poollen, H.K. and Jargon, J.R. 1969. Steady-State and Unsteady-State
⌬Sw ⳱ displaceable oil saturation, fraction Flow of Non-Newtonian Fluids Through Porous Media. SPEJ 9 (1):
␾ ⳱ porosity, fraction 80–88; Trans., AIME, 246.
␥ ⳱ apparent-mobility ratio, dimensionless Warren, J.E. and Cosgrove, J.J. 1964. Prediction of Waterflood Behavior in
a Stratified System. SPEJ 4 (2): 149–157; Trans., AIME, 231. SPE-
␥˙ ⳱ shear rate, s−1
581-PA.
␭ ⳱ mobility
Wu, Y.-S. and Pruess, K. 1998. A Numerical Method for Simulating Non-
␮ ⳱ apparent viscosity, cp (Pa·s) Newtonian Fluid Flow and Displacement in Porous Media. Adv. in
␴k ⳱ standard deviation of log-normal permeability Water Res. 21 (5): 351–362.
␶ ⳱ dimensionless time Wu, Y.-S., Pruess, K., and Witherspoon, P.A. 1992. Flow and Displace-
T ⳱ shear stress ment of Bingham Non-Newtonian Fluids in Porous Media. SPERE 7
(4): 369–376. SPE-20051-PA.
Subscripts Yi, X. 2004. Model for Displacement of Herschel-Bulkley Non-Newtonian
ap ⳱ apparent Fluid by Newtonian Fluid in Porous Media and Its Application in
D ⳱ dimensionless Fracturing Fluid Cleanup. Paper SPE 86491 presented at the SPE In-
eff ⳱ effective ternational Symposium and Exhibition on Formation Damage Control,
Lafayette, Louisiana, 18–20 February.
fin ⳱ final
in ⳱ initial
Appendix A—Derivation of the Fractional-Flow
m ⳱ mean
Formula for Non-Newtonian Fluids
o ⳱ oil
For Layer i in Zone j (i⳱1, 2, . . . , j), from Eq. 13:
r ⳱ relative

冉 冊
t ⳱ total 1
ki⌬hiWkrw
•
dp
w ⳱ water qwi = ␣wi −
nw
, . . . . . . . . . . . . . . . . . . . . . . . (A-1)
␮wap dx
Superscripts where k•rw and k•ro are the water and oil relative permeabilities at
_
⳱ average the endpoints.The total flow rate of water in Zone j is

冉 冊
1
References j •

兺
Wkrw dp nw
Qwj = qwi = − Cwj , . . . . . . . . . . . . . . . . . . . (A-2)
Bird, R.B., Stewart, W.E., and Lightfoot, E.N. 1960. Transport Phenom-
i=1 ␮wap dx
ena, 206. New York City: John Wiley & Sons Inc.
Cannella, W.J., Huh, C., and Seright, R.S. 1988. Prediction of Xanthan where
Rheology in Porous Media. Paper SPE 18089 presented at the SPE
j
Annual Technical Conference and Exhibition, Houston, 2–5 October.
Chakrabarty, C., Tortike, W.S., and Farouq Ali, S.M. 1993. Complexities
in the Analysis of Pressure-Transient Response for Non-Newtonian
Cwj = 兺␣
i=1
wi ki ⌬hi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-3)

Power-Law Fluid Flow in Fractal Reservoirs. Paper SPE 26910 pre- Similarly for the oil phase,
sented at the SPE Eastern Regional Meeting, Pittsburgh, Pennsylvania,

冉 冊
2–4 November. N 1
Wk•ro
兺
dp no
Dykstra, H. and Parsons, R.L. 1950. The Prediction of Oil Recovery by Qoj = qoi = − 共Cot − Coj兲, . . . . . . . . . . . . . . (A-4)
Waterflooding. In Secondary Recovery of Oil in the United States, i=j+1 ␮oap dx
second edition, 160–174. Washington, DC: API.
El-Khatib, N. 1999. Waterflooding Performance of Communicating Strati- where
fied Reservoirs With Log-Normal Permeability Distribution. SPEREE j
2 (6): 542–549. SPE-59071-PA.
Gleasure, R.W. 1990. An Experimental Study of Non-Newtonian Polymer
Coj = 兺␣
i=1
oi ki⌬hi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-5)
Rheology Effects on Oil Recovery and Injectivity. SPERE 5 (4): 481–
486. SPE-17648-PA. and
Gogarty, W.B., Levy, G.L., and Fox, V.G. 1972. Viscoelastic Effects in
Polymer Flow Through Porous Media. Paper SPE 4025 presented at the
Cot = CoN. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-6)
SPE Annual Meeting, San Antonio, Texas, 8–11 October. From Eqs. A-2 and A-4, we obtain
Hearn, C.L. 1971. Simulation of Stratified Waterflooding by Pseudo Rela-
tive Permeability Curves. JPT 23 (7): 805–813. SPE-2929-PA. Qwj ␥Cwj
fwj = = , . . . . . . . (A-7)

冉 冊
Herschel, W. and Bulkley, R. 1926. Consistency Measurements on Rubber- Qwj + Qoj dp
1 1
−
Benzene Solutions. Koll. Zeit. 39: 291. ␥Cwj + 共Cot − Coj兲 −
no nw

Hiatt, N.W. 1958. Injected-Fluid Coverage of Multi-Well Reservoirs With dx

Permeability Stratification. Drill. and Prod. Prac. 165: 165–194.
Washington, DC: API. where ␥ is the apparent-mobility ratio defined as
Odeh, A.S. and Yang, H.T. 1979. Flow of Non-Newtonian Power-Law 0
krw ␮oap
Fluids Through Porous Media. SPEJ 19 (3): 155–163. SPE-7150-PA. ␥= . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-8)
Salman, M., Baghdikian, S.Y., Handy, L.L., and Yortsos, Y.C. 1990. k0ro ␮wap
Modification of Buckley-Leverett and JBN Methods for Power-Law
Fluids. Paper SPE 20279 available from SPE, Richardson, Texas. From Eq. A-2, we can write

冉 冊
Savins, J.G. 1969. Non-Newtonian Flow Through Porous Media. Ind. Eng. • 1
Chem. 61 (10): 18–47. Wkrw dp nw
Qwj = fwjQt = − Cwj , . . . . . . . . . . . . . . . . . . . . . (A-9)
Stiles, W.E. 1949. Use of Permeability Distribution in Water Flood Cal- ␮wap dx
culation. Trans., AIME 186: 9–13.
Teeuw, D. and Hesselink, F.T. 1980. Power-Law Flow and Hydrodynamic from which we get

冉 冊冉 冊
Behaviour of Polymer Solutions in Porous Media. Paper SPE 8982
dp fwjQt ␮wap nw
presented at the SPE Oilfield and Geothermal Chemistry Symposium, − = •
. . . . . . . . . . . . . . . . . . . . . . . . . . (A-10)
Stanford, California, 28–30 May. dx Wkrw Cwj

364 August 2006 SPE Reservoir Evaluation & Engineering

Substituting in Eq. A-7, At the time of water breakthrough in Layer j (j⳱1, 2, . . . , N),
from Eqs. B-1 through B-3 we obtain
␥Cwj
fwj = , . . . . . . . . . . . . . (A-11)

冉 冊 冉 冊 兺冉 冊 冉 冊
nw
Qt ␮wap
N
fwjQt␮wap −1 nw fwi nw ⌬fwi
⌬pt = L␶j ⌬ , . . . . . . . . . . (B-9)
no
␥Cwj + 共Cot − Coj兲 •
Wkrw Cwj
•
Wkrw i=j Cwi ⌬hDi
which is the fractional-flow formula for non-Newtonian fluids. It where ␶j is given by Eq. 17.
is clear that when nw/no⳱1, Eq. A-11 reduces to the familiar Stiles Initially, from Eq. B-7 at ␶⳱0,
formula (1949) for communicating stratified reservoirs.
Appendix B—Derivation of the
Injectivity-Variation Formula
⌬pt in = L 冉 Qt in ␮oap
Wk•roCot
冊 no
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-10)

As the displacement proceeds and more of the displacing fluid

enters into the formation, either the injection rate, the total pressure Dividing Eqs. B-7 and B-9 by Eq. B-10 and rearranging, we
drop, or both will change. obtain expressions for the inverse injectivity ratio 1/Ir or the re-
From Eq. A-10, we get, for j⳱1, . . . , N, sistivity ratio ␤.

冉 冊
Before water breakthrough in the first layer,
fwjQt␮wap nw
⌬pj = •
L⌬Xj , . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-1) 1 ⌬pt /Qno
Wkrw Cwj t
␤= =
Ir 共⌬pt /Qno
t 兲 in

冉 冊
where
␮wap

冤 冥
nw
⌬Xj = Xj − Xj+1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-2)
兺冉 冊 冉 冊
N
fw1 •
Wkrw fwj nw ⌬fwj

冉 冊
⌬fwj =1−␶ − Qnw−no ⌬ ,
and Xj = ␶ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-3) ⌬hD1 t
␮oap no
j=1 Cwj ⌬hDj
⌬hDj
Wk•roCot
Before water breakthrough in the first layer, for Zone 0, . . . . . . . . . . . . . . . . . . . . . . . (B-11)
X0⳱1, and

⌬p0 = 冉 Qt␮oap
冊 no
L共1 − X1兲. . . . . . . . . . . . . . . . . . . . . . . . . (B-4)
and at the time of water breakthrough in Layer j,
Wk•roCot
冉 冊
Qtj␮wap nw

冉 冊 冉 冊
From Eqs. B-1 and B-4, we obtain N
•
⌬fwi

冉 冊兺
Wkrw fwi nw
␤j = ␶j ⌬ . . . . . . . . . . . (B-12)
冉 Qt␮oap
冊 冉 冊 兺冉 冊
Qt␮wap nw N Qt in ␮oap ⌬hDi
no nw no Cwi
fwj i=j
⌬pt = L共1 − X1兲 + L ⌬Xj .
Wk•roCot •
Wkrw j=1 Cwj Wk•roCot
. . . . . . . . . . . . . . . . . . . . . . . . . (B-5)
Substituting for the values of Xj from Eq. B-3 and noting that
SI Metric Conversion Factors
⌬fw1 fw1
X1 = ␶ =␶ , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-6) bbl × 1.589 873 E–01 ⳱ m3
⌬hD1 ⌬hD1 cp × 1.0* E–03 ⳱ Pa·s
we obtain ft × 3.048* E–01 ⳱ m

冉 冊冉 冊
Qt ␮oap no fw1 psi × 6.894 757 E+00 ⳱ kPa
⌬pt = L 1−␶
Wk•roCot ⌬hD1 *Conversion factor is exact.

冉 冊 兺冉 冊 冉 冊
Qt␮wap
N
nw fwj nw ⌬fwj
+ L␶ ⌬ . . . . . . . . . . . (B-7) Noaman El-Khatib is a professor of petroleum engineering at
•
Wkrw j=1 Cwj ⌬hDj
Sudan U. for Science and Technology in Khartoum, Sudan.
Eq. B-7 gives the change in total pressure drop ⌬pt for constant e-mail: nkhatibafm@gmail.com. Previously, he was a professor
total injection rate Qt up to the time of water breakthrough in the at King Saud U. in Riyadh, Saudi Arabia, where he had been on
first layer ␶1, where the faculty for 25 years. He also taught at Cairo U. and the U.
of Pittsburgh. His research interests include well-test analysis,
1 1 reservoir simulation, enhanced oil recovery, reservoir charac-
␶1 = = . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-8) terization, and geostatistics. El-Khatib holds a BS degree from
⌬fw1 fw1 Cairo U. and MS and PhD degrees from the U. of Pittsburgh, all
⌬hD1 ⌬hD1 in petroleum engineering.

August 2006 SPE Reservoir Evaluation & Engineering 365