Copyright 2006, Society of Petroleum Engineers

This paper was prepared for presentation at the 2006 SPE Annual Technical Conference and
Exhibition held in San Antonio, Texas, U.S.A., 2427 September 2006.

This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than
300 words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract

Steady-state flow experimental data have been analyzed
for two commonly used polymers representing two generic
classes: polysaccharides (xanflood), and partially hydrolyzed
polyacrylamides (pusher-700) flowing inside bead packs and
Berea sandstone. Oscillatory flow measurements have been
used to compute the polymer solution's longest relaxation time
(
f
1
) which is referred to as the characteristic relaxation time
in this paper. Steady-state flow experimental data for the two
polymers combined with measured polymer viscous properties
have been converted to average shear stress-shear rate data
inside porous media. An average power-law exponent ( n ) is
therefore obtained for the polymer flow inside porous media.
Using
f
1
and n , rock permeability (k) and porosity () and
fluid flow velocity (u), a viscoelasticity number (N
v
) is
calculated, and found to strongly correlate with the pressure
gradient inside porous media. This correlation is the basis for
defining a viscoelastic model for polymer flow, analogous to
Darcy's law. The proposed model asserts a non-linear
relationship between fluid velocity and pressure gradient. It
accounts for polymer elasticity, and for pore geometry
changes due to molecular adsorption and mechanical
entrapment.

Introduction

Polymers are used in a variety of petroleum engineering
applications. In enhanced oil recovery, polymers are added to
water to increase its viscosity and decrease its relative
mobility. This decrease in water mobility improves the
volumetric sweep efficiency and the efficiency of the water
flood.
1,2
Even after decades of research, the mechanism of
polymer flow in porous media remains vaguely understood.
Darcy's law for fluid flow in porous media is valid for
fluids behaving as a continuum with constant viscosity, and
for constant pore geometry. The flow of polymers in porous
media deviates from these assumptions because: (i) polymers'
viscosity is shear rate dependent, (ii) polymer molecules'
length is comparable to the pore throat length, a fact that
enhances certain elastic phenomena, and (iii) molecular
adsorption and mechanical entrapment make the geometry of
the porous medium variable. Consequently, the use of Darcy's
law for modeling polymer flow in porous media has serious
flaws. The traditional approach to modeling polymer flow in
porous media has been to use Darcy's law with an effective
viscosity, at a fixed shear rate, substituted for the Newtonian
viscosity. This procedure corrects for the shear-rate
dependence of viscosity but does not account for the non-
continuum and elastic flow properties.
3,4

van Poollen and Jargon
5
, and Willhite and Uhl
6
showed by
a simple empirical model the non-linear relationship between
the pressure drop and the flow rate for non-Newtonian fluid
flow. This relation was expressed as

P = F
Q
m+1
A
m+1
L. (1)

In this notation, A is the cross-sectional area, Q is the
volumetric flow rate, m is a power-law constant. F is an
empirical constant that has been correlated to the rock
permeability, porosity and tortuosity. A similar, more explicit,
non-linear dependence can also be obtained by extending the
tube flow rate equation for a viscous power-law fluid to the
flow in porous media using the bundle of capillary tubes
theory. The resulting model is as follows:

Q =

1/ n
2
1/ n
r
3n+1
n
3n+1
n
f
p
dr
0

r
2
f
p
dr
0

A
m
1/ n






P
L






1/ n
. (2)

Here, m and n are the power-law constants of the polymer, f
p

is the pore size distribution function of the rock, and r is a
pore radius. This model clearly asserts two important findings
achieved by van Poollen and Jargon
5
, and by Willhite and
Uhl.
6
The first is that the flow rate indeed relates non-linearly
to the pressure drop. The second is that the constant of

SPE 102015
A Novel Model for Viscoelastic Fluid Flow in Porous Media
A.A. Garrocuh, SPE, and R.B. Gharbi, SPE, Kuwait U.
2 SPE 102015
proportionality depends on both the fluid properties (m and n)
and on the rock properties such as the pore size distribution
and the porosity.
The above models fall short, however, in accounting for
the elastic phenomena manifested by polymer flow in porous
media. Because the length of polymer molecules is
comparable to the length of pore throats and pore bodies in the
porous media, these molecules compress as they hit the pore
walls. Then, they stretch, or relax, to resume the flow under
the influence of the flow potential. The characteristic
relaxation time is sometimes comparable to the transient time
defined as the time these molecules take to exit from a pore
throat, or a pore body. One can expect major elastic effects
when the polymer molecule characteristic relaxation time is
comparable to its transient time. In many instances these
effects are felt as an additional pressure drop, or an additional
resistance to flow. A thorough summary on the subject is
given by Savins.
7
In this review paper, Savins
7
highlights an
interesting theoretical model expressed in a non-dimensional
form in terms of the friction factor (f), the Reynold's number
(Re) and the Deborah number (N
DEB
):

f Re = 1 + N
DEB
( )
2
[ ]
. (3)

Deborah number is a measure of polymer elasticity. This
number is expressed as a function of polymer concentration,
molecular weight, viscosity at zero shear rate, the fluid
superficial velocity and a characteristic grain size diameter of
the rock.
7
Based on a perturbation analysis of flow of a spring-
type of viscoelastic fluid characterized by a single relaxation
time, Savins
7
showed that, for a porous medium, the purely
viscous pressure drop must be multiplied by a factor of
[1+(N
DEB
)
2
] to obtain the pressure drop in the presence of
elastic effects. The parameter is a constant. A great number
of published data, however, do not agree with the rate of
increase in pressure loss predicted by this model.
7

The objective of this research is to develop an empirical
model for modeling polymer flow in porous media. The model
accounts for both viscous and elastic effects. Steady-state
experiments for polymer flow in porous media are the means
for defining the non-linear dependence between the fluid
velocity and the pressure gradient for horizontal flow. This
dependence is expressed in terms of fundamental rock and
fluid properties.


Experiemental Equipment and Procedures

Data used in this analysis are extracted from three
categories of experiments: measurement of polymer solution
properties, measurements of petrograhic parameters for porous
media, and flow experiments of the polymer solutions through
porous media.

Preparation of Polymer Solutions. Two types of water-
soluble polymers were used to study the viscoelasticity effects
on flow behavior in porous media. These are xanthan gum
(Kelkos xanflood, hereafter referred to as xanflood), and
partially hydrolyzed polyacrylamide (Dow's pusher-700,
hereafter referred to as pusher-700). The polymers all came in
powder form, and aqueous solutions were prepared in different
ways depending upon the polymer type. Pusher-700 polymer
solutions were made directly at the desired polymer and salt
concentrations with a magnetic stirrer. The rate of addition of
polymer was slow enough for the particles to disperse in water
without lump formation, but not so slow that the solution
thickens before all the solid was added. Eight to twenty-four
hours were generally required for complete hydration. The
xanflood solution was prepared by first dissolving sodium
bicarbonate and an enzyme in distilled water. Then the
xanthan gum powder was added and the mixture sheared in a
high speed blender for 10 minutes. This provided a
concentrated solution containing 1% by weight (1 wt %)
which was then blended with brine containing a biocide to
obtain the desired polymer and sodium chloride
concentrations. The final composition was 1000 ppm xanthan
gum, 10000 ppm NaCl, 100 ppm NaHCO
3
, 4 ppm enzyme
(Alcalase P1.5), and 120 ppm biocide (Dowicide B).

Rheological Measurements. The bulk fluid property
characterized is the viscosity as a function of shear rate
ranging from approximately 100000 sec
-1
at the injector
wellbore to as low as 0.01 sec
-1
at some distance into the
reservoir. This study included most of this range. The
Contraves L-S 30 rheometer was used with various bob-and-
cup arrangements for measuring shear stress-shear rate
profiles.
8
The instrument is valuable since ultra-low shear
stress measurements are possible for shear rates as small as
0.0174 sec
-1
. The viscoelasticity was measured using an
oscillatory tube flow system in which the frequency of
oscillation and the amplitude are independently controlled.
The apparatus and theory of measurement have been described
by Thurston and Pope.
9
In this method, the shear stress at the
tube wall is determined from the pressure gradient while the
shear rate at the tube wall is determined from the volume flow.
The sinusoidal electrical signals representing the instantaneous
shear stress and shear rate are analyzed to obtain the
magnitudes and phase difference between the fundamental
components, free of harmonics and noise.

Porous Media Description. Two types of permeable media
were used. These are glass-beads packed-columns
(experiments A, B, C, and D) and Berea sandstone cores
(experiments E and F). Glass beads provided a uniform
unconsolidated pack that was free of clays. Two different sizes
of glass beads were used in this study, namely 100-110 m
beads (for experiments A, B, and C) and 250-300 m (for
experiment D). The glass beads were packed in a 9/16 inch
inner-diameter stainless steel pipe that was one-foot long. The
column was composed of a piece of stainless steel pipe and
two end caps. A detailed description of the glass-beads
columns used to generate data for experiments A, B, C and D
is given by Yuan.
10
Multiple pressure taps were installed on
these columns to observe the fluid entrance and exit end-
effects. These pressure taps were also useful for monitoring
the pressure distribution to determine if a fully developed flow
was present inside the packed column, or if any plugging
occurred in the column.

SPE 102015 3
Berea core samples were used to generate data for experiments
E, and F. Berea provided a complex form of porous medium
since it is a consolidated rock with some clay material present.
A detailed description for the Berea core holder setup and
procedures used to generate experiments E, and F is provided
by Jones
8
and Wreath.
11
Two-foot long Berea core samples
were cut with a two-square-inch cross-sectional area. Nylon
screening material of 500 m opening size was placed
between the end piece and the core face to insure an adequate
initial fluid contact area which was uniform for the entire core
face. Fittings were positioned on the top of the core, 6 inches
from each core face, providing two pressure taps 12 inches
apart. Tables 1 and 2 give a summary of the petrographic
properties of each porous medium and the corresponding
polymer solution used.

Pumping System and Flooding Procedures. A constant
flow-rate feed-system was employed to conduct the flow
experiments through the packed columns and the Berea core
samples. The flow system consisted of a Zenith pump used
for controlling the flow rate, a solution reservoir, a packed
column with uniform glass beads or Berea as the porous
medium, a system of pressure transducer-carrier demodulator
and a chart recorder for pressure drop measurements, as well
as a fraction collector to collect the effluent samples
automatically. All the solutions were filtered and degassed
before they were introduced into the solution reservoir.
For all porous media samples, an initial brine flood was
conducted before any polymer solutions were pushed into the
medium to determine the absolute permeability. Brine with
1% NaCl concentrations by weight was made and filtered
through 0.45 m Millipore filter paper with the application of
the Fann Filter Press under 40 psi pressure. It was then fed
into the solution reservoir after degassing. After all the air
bubbles in the tubing and fitting were purged out, the flow
experiment started. Pressure drops were measured from
pressure taps using pressure transducers. Three pressure drop
values were read for each flow rate. The volumetric flow rate
was determined after steady flow was established, which could
be indicated by the stabilized pressure drop values.
After the initial brine flood, polymer solutions were
filtered through 1.2 m Millipore filter paper under 40 psi.
Then they were pushed through the porous medium to study
the viscoelastic nature of the solutions. The porous medium
which has been flooded with 1% NaCl brine solution was then
flushed with polymer solution at a medium flow rate. The
volumetric flow rate was arbitrarily selected to be 1.7
cm
3
/min. This flush had been conducted until a steady
pressure drop was established and the viscosity of the effluent
was measured to be identical to the injected polymer solution
viscosity. This ensured that the packed column was indeed
fully saturated with the polymer solution. Polymer flooding
was then conducted from low rates to high rates while
measuring the pressure drops and flow rates. The effluents
were collected and viscosities were measured to see whether
shear degradation occurred during the flow. The time span for
each experiment was kept to a minimum to maintain the
freshness of the polymer solutions. After the polymer flood
had finished, the brine solution was run again to obtain the
flushed permeability. The brine solution was run at a relatively
high flow rate (25 cm
3
/min) until about 25 to 30 pore volumes
were injected and a steady pressure drop had been established.

Characterizing Polymer Viscous Effects. The experimental
data of flow rate versus pressure drop was converted to shear
stress versus shear rate data inside the porous medium. The
apparent viscosity of polymer solution inside the porous
medium was calculated using Darcy's law as

qL
P A k
f
app

= . (4)

Here k
f
is the permeability of the flushed porous medium. It
was measured after the brine flood following the polymer
flood. The porous medium shear rate was calculated using the
Blacke-Kozeny model
8,10
as

f
1 n
n
p
k 8
u 4
n 4
1 n 3

+
=
. (5)

The exponent n is the power-law parameter determined from
the viscous flow measurements obtained using the LS-30
rheometer by plotting the shear stress-shear rate profile. The
parameter n is the slope from this log-log plot. The
corresponding shear stress inside the given porous medium
was therefore calculated using

p app p
=
. (6)

Figure 1 shows the shear stress-shear rate profile for a typical
xanflood solution flowing inside a bead pack. The xanflood
solution appears to fit reasonably well the power-law model
given by

n
p p
m = , (7)

implying insignificant elastic effects. The exponent n is an
average power-law constant inside the porous medium. The
molecules of xanflood are described as relatively stiff in the
literature.
12
Their flow behavior is governed by their shear
viscosity. The results obtained in this study confirm these
observations. The pusher-700 solutions showed some
deviation from the viscous power-law model given by Eq. 7.
Figure 2 shows the shear stress-shear rate profile for
experiment B which is run on a similar sand pack to that of
experiment A, but with pusher-700 solution instead. The
shear-rate magnitude increased by approximately two orders
of magnitude, as compared with values for experiment A,
because of the shear-thickening behavior of pusher-700. The
deviation of the shear stress-shear rate profile from a straight
line in a log-log plot confirms the literature observations that
classify pusher-700 as a viscoelastic polymer.
9,12


4 SPE 102015
Characterizing Polymer Elastic Effects. The viscoelasticity
was measured using an oscillatory tube flow system in which
the frequency of oscillation and the amplitude were
independently controlled. The complex coefficient of viscosity

*
was obtained from the complex ratio of shear stress to the
shear rate
,

E v
*
i

= =

, (8)

where
v
is the energy dissipative viscous component and
E

is the non-dissipative elastic energy storage component.
9
Both

v
and
E
are functions of the angular frequency and the
shear rate
at which the measurement is made. The
experimental data were fitted to the classic Rouse model.
13
In
this model, the polymer molecule is represented as a flexible
chain of N beads connected by N-1 elastic springs. When
immersed in a solvent fluid, such a model molecule has N-1
relaxation times which are related to the storage and loss
moduli as follows:

v
=
s
+ RT

f
j
1 +
f
j

2
j=1
N1

, (9)
and

E
= RT

f
j
2
1 +
f
j

2
j=1
N1

. (10)

The longest relaxation time,
f
1
, was used for the correlation
with flow data through cores and packs. By fitting the
predictions of the model for
v
and
E
to the experimental
data, an estimate for
f
1
was obtained. This characteristic
relaxation time was used to calculate the Deborah number
defined as
7

N
DEB
=

f
1
u
D
p
eff

, (11)

where, D
p
eff
is the effective porous medium diameter
8,10
and
is given by

D
p
eff
=
150 1
( )
2
k

3
. (12)

Selecting a Viscoelasticity Number. For two similar sand
packs one flooded with xanflood (experiment A) and the other
flooded with pusher-700 (experiment B), the Deborah
number-flow rate profile appears to be identical (Figure 3).
This is not surprising since a slight difference in the
characteristic relaxation time for these two polymer solutions
does not cause a significant difference in N
DEB
, according to
Eq. 11. The surprising element though is the fact that xanthan
material which is described in the literature
12
as long rod-like
molecule with little flexibility has a relaxation time
comparable to pusher-700 which is a partially hydrolyzed
polyacrylamide made of flexible and elastic chain-structured
molecules. This might be explained by the sensitivity of
pusher-700 to solution salinity and hardness. Salinity causes
the molecular chain to collapse which results in a much
smaller random coil structure with a reduced relaxation time.
A fully extended partially hydrolyzed polyacrylamide
molecule in a good solvent would exist as a filament of 7-25
A

in diameter and exceeds 10 m in length. The molecular

length of xanthan is estimated to be 2-10 m. The diameter of
a single-chain xanthan backbone is 20 A

.
12
In a saline
solution (1wt %) the length and width of the xanthan molecule
make up for its elasticity as compared to the narrow partially
hydrolyzed molecule whose length has been reduced. This is
confirmed in Figure 4 where a comparison is made between
the Deborah number of two pusher-700 solutions of different
salinities (1 wt % and 0.1wt % NaCl). The lower salinity
solution appears to be more elastic than the higher salinity
solution. This result indicates that the Deborah number may
not be an adequate parameter for characterizing the
viscoelasticity of polymer flow in porous media since an
elastic polymer (pusher-700) appears to behave similarly to a
viscous polymer (xanflood) in porous media because of
pusher-700 sensitivity to solution salinity. This suggests the
introduction of a group number (N
v
) which we refer to as the
viscoelasticity number in this manuscript, and it is given by
14


N
v
=
k

f
1
u
n 1
. (13)

The viscoelasticity number becomes dimensionless when n
takes a value of 2, and carries units of velocity when n equals
1. It also increases with increasing n values as long as the
fluid velocity (u) is maintained less than 1 m/sec which is
always the case for polymer flow in porous media.
Figure 5 shows a plot of the viscoelasticity number versus
flow rate for two similar packs (experiments A and B) but for
two displacing polymers (pusher-700 for experiment B, and
xanflood for experiment A). Here the distinction between the
two polymers is crystal clear both in magnitude and in profile.
The shear-thickening pusher-700 appears to have a negative
slope unlike the shear-thinning xanflood, despite the solution
high salinity. The viscoelasticity of the pusher-700 is clearly
indicated by the magnitude of N
v
at any flow rate compared to
xanflood.

Modeling Polymer Flow. As shown in Figure 6, a log-log
plot of the flow rate term ( Qm
1/ n
/ A ) versus the pressure
gradient term P / L ( )
1/n
, for pusher-700 solution flooded in
a bead-pack, did not result in a straight line. This suggests that
the pusher-700 floods did not follow the flow rate-pressure
drop relationship indicated by the capillary tube model given
by Eq. 2. This is not surprising since the capillary-tube model
was derived for viscous but not elastic polymers. On the other
SPE 102015 5
hand, xanflood in both bead packs and Berea appeared to
mimic the capillary-tube model reasonably well (Figures 7 and
8). One expects xanflood to follow the capillary-tube model
since the polymer has insignificant elasticity, and since the
bundle-of-tubes model is a close approximation for bead
packs, and these facts are in line with the original assumptions
made in deriving the capillary-tube model given by Eq. 2. The
results of xanflood flow in Berea (Figure 8) make a little bit of
surprise though since Berea sandstones pore geometry is far
more complex than the simple geometry assumed in
developing the capillary-tube model. These results may
suggest that the flow of a viscous polymer with insignificant
elasticity inside bead packs, or Berea sandstones adheres to
the following non-linear relationship between flow rate and
pressure gradient:

Q
A
m
1/ n
=
P
L




1/ n

. (14)

Here and are two empirical constants that depend on the
porous medium geometry and the polymer rheological
properties. Both and are obtained from a log-log plot of the
flow rate term versus the pressure gradient term. A statistical
summary of the values of these constants, and the
corresponding regression coefficients for each experiment are
given in Table 3.

Data from the six experiments investigated appeared to fit
very well in a log-log plot of viscoelasticity number versus
pressure gradient. Examples of these responses for both Berea
and bead packs flooded with xanflood and pusher-700 are
displayed in Figures 9 through 12. A viscoelatic model is
therefore suggested as

N
v
=
P
L

. (15)

In this notation is the slope of a log-log plot of N
v
versus
pressure gradient. The empirical constant is the intercept at
a value of (P/L) equal to one from the same log-log plot. By
using the definition of the group number N
v
(Eq. 13), the
above equation can be rearranged, for n values different from
one, to give the following relationship between Darcy's
velocity (u) and the pressure gradient (P/L):

u =
k

f
1






1
n 1
P
L

1n
. (16)

The model presented by Eq. 16 asserts the non-linear
relationship between fluid velocity and pressure gradient. It
also captures both viscous and elastic effects of the polymer
solutions, since it includes the characteristic relaxation time
(
f
1
) and the power-law constant ( n ) that reflects the porous
medium relationship between shear stress and shear rate.
Adsorption and mechanical entrapment effects on the pore
structure are implicitly accounted for through the use of the
power-law constant n . This is true since the flushed
permeability k
f
is used to calculate the porous medium shear
rate (Eq. 5). This data is ultimately used to generate n from a
plot of shear stress versus shear rate inside the porous
medium. The model also accounts for the effects of the
original microscopic structure and geometry of the porous
medium through rock permeability (k) and porosity (). As
shown in Table 4, it is interesting to note that the pressure
gradient exponent ( / (1 n )) of the proposed viscoelastic
model (Eq. 16) is approximately equal to the pressure gradient
exponent (1/ n ) of the capillary tube model (Eq. 2). This
implies that the term ( / (1 n )) of Eq. 16 makes up for
the viscous part of the pressure drop needed for the polymer to
flow. The model presented by Eq. 16 can, therefore, be
rearranged to give an expression for the fluid velocity in terms
of n instead of , as follows:

n
1
1 n
1
1
f
L
P k
u

. (17)


A statistical summary of the and values and the
corresponding regression coefficients for each experiment are
given in Table 3. The intercept appears to strongly correlate
with n . The correlation coefficient between and n is
approximately 0.86. The deterministic relationship between
and the remaining fluid and porous media parameters is yet to
be determined, though.

Conclusions

A steady-state flow experimental data set of two polymer
solutions (xanflood and pusher-700) flowing in porous media
(Berea sandstone and bead packs) has been analyzed. The
analysis led to the development of an empirical viscoelastic
model for polymer flow in porous media analogous to Darcy's
law. A non-linear relationship between fluid velocity and
pressure gradient is asserted though. The proposed model
accounts for polymer elasticity by using the longest relaxation
time (
f
1
), and accounts for polymer viscous properties by
using an average porous medium power-law constant ( n ). It
also accounts for porous medium structure by using porosity
and permeability. The changing pore structure is accounted for
in estimating ( n ) since the in-situ shear rate, used to generate
n , depends on the after-flush permeability (k
f
).
Analysis of similar data, for a variety of polymers flowing
inside heterogeneous rocks, is essential for confirming the
empirical relationships between parameters , , the fluid
rheological properties, and the petrographic properties. This
confirmation may then generalize our results to other types of
porous media.
The xanflood flow in both bead packs and Berea
sandstone is found to mimic a non-linear relationship between
flow rate and pressure gradient similar to that predicted by the
6 SPE 102015
capillary-tube model. This implies the development of a
second empirical model that may work only for flow of
viscous polymers with insignificant elasticity.
The study shows that the Deborah number may not be an
adequate parameter to characterize viscoelastic flow of
polymers in porous media. A dimensional number N
v
, called
the viscoelasticity number, is found to be more adequate since
it clearly distinguishes viscous from viscoelastic flow.
Acknowledgments
The authors would like to thank Kuwait Foundation for the
Advancement of Sciences (KFAS) for its financial support of
project KFAS-94-8-2.


Nomenclature

A cross sectional area
a
o
regression constant for Eq. 17 given in the
Appendix
a
1
regression constant for Eq. 17 given in the
Appendix
a
2
regression constant for Eq. 17 given in the
Appendix
a
3
regression constant for Eq. 17 given in the
Appendix
a
4
regression constant for Eq. 17 given in the
Appendix
D
p
eff
effective porous medium diameter
P pressure drop
F empirical constant
f friction factor
f
p
pore size distribution
k rock permeability
k
f
flushed permeability
L length of core sample
m fluid power-law constant from rheological
measurements
m intercept from a log-log plot of shear stress
versus shear rate of the rock
n fluid power-law constant from rheological
measurements
N number of beads of a flexible chain that
models a polymer molecule
n average power-law constant inside porous
medium
N
v
viscoelasticity number
N
DEB
Deborah number
r pore radius
R
e
Reynold's number
r

2
coefficient of determination for regression
line given by and
r

2
coefficient of determination for regression
line given by and
Q volumetric flow rate
u Darcy velocity
Qm
1/ n
/ A flow rate term
P / L ( )
1/n
pressure gradient term


Greek Letters

intercept from a log-log plot of N
v
versus
pressure gradient
slope from a log-log plot of N
v
versus
pressure gradient
slope from a log-log plot of flow rate term
versus pressure gradient term
rock porosity

complex shear rate

p

average porous medium shear rate

intercept from a log-log plot of flow rate
term versus pressure gradient term
Newtonian fluid viscosity

app
apparent viscosity of polymer solution
inside rock

complex coefficient of viscosity

E
the non-dissipative elastic energy storage
component of

v
the energy dissipative viscous component of

s
steady-flow viscosity
complex shear stress

p
average shear stress in porous medium

f
relaxation time

f
j
longest relaxation time
angular frequency
a constant


References

1. Lake, L.W. Enhanced Oil Recovery; Prentice Hall:
Englewood Cliffs, New Jersey, 1989.
2. Xian-Qing, H.; Wei-Ying, W.; Ying, X. The Viscoelastic
Behavior of HPAM Solutions in Porous Media and its
Effects on Displacement Efficiency. SPE International
Meeting on Petroleum Engineering, Beijing, November
1995; Paper 30013.
3. Pruess, W. K.; Witherspoon, P.A. Displacement of a
Newtonian Fluid by a Non-Newtonian Fluid in a Porous
Medium. Transport in Porous Medium. 1991, 6, 115.
4. Flew, S.; Sellin, R.H.J. Non-Newtonian Flow in Porous
Media - A Laboratory Study of Polyacrylamide Solutions.
Journal of Non-Newtonian Fluid Mechanics. 1993, 47,
169.
5. van Poollen, H.K.; Jargon, J.R. Steady state and unsteady
state flow of non-Newtonian fluids through porous media.
In AIME Transactions: Dallas, TX, 1969; Vol. 246, p 80.
6. Willhite, G.P.; Uhl, J.T. Correlation of the mobility of
biopolymer with polymer concentration and rock
properties in sandstone. Polymeric Science and
Engineering. 1986, 55, 577.
SPE 102015 7
7. Savins, J.G. Non-Newtonian Flow Through Porous Media.
Industrial and Engineering Chemistry. 1969, 61, 18.
8. Jones, K. The Rheology of Viscoelastic Fluids for Oil
Recovery: A Study of Micellar/Polymer Fow Over a Wide
Range of Shear Rates in Porous Media. M.S. Thesis, The
University of Texas at Austin, Austin, TX, 1981.
9. Thurston, G.B.; Pope, G. Shear Rate Dependence of the
Viscoelasticity of Polymer Solutions. Journal of Non-
Newtonian Fluid Mechanics. 1981, 9, 69.
10.Yuan, M. A Rheological Study of Polymer and
Microemulsion in Porous Media. M.S. Thesis, The
University of Texas at Austin, Austin, TX, 1981.
11. Wreath, D. A Study of Polymerflooding and Residual Oil
Saturation. M.S. Thesis, The University of Texas at
Austin, Austin, TX, 1989.
12. Sorbie, K. Polymer-Improved Oil Recovery; Blackie and
Son Ltd: London, 1991.
13. Heemskerk, J.; Rosmalen, R.J.; Holtslag, R.J.; Teeuw, D.
Quantification of Viscoelastic Effects of Polyacrylamide
Solutions. SPE/DOE Fourth Symposium on Enhanced Oil
Recovery, Tulsa, OK, April 1984; Paper 12652.
14.Wreath, D., Pope, G.; Sepehrnoori, K. Dependence of
Polymer Apparent Viscosity on the Permeable Media and
Flow Conditions. In Situ. 1990, 14, 263.



































































8 SPE 102015
Table 1. Listing of all Experiments Analyzed

Experiment Solution Pack
A 1000 ppm xanflood in
1 wt % NaCl brine
100-110 m beads
B 1000 ppm pusher-700 in
1 wt % NaCl brine
100-110 m beads
C 1000 ppm pusher-700 in
0.1 wt % NaCl brine
100-110 m beads
D 1000 ppm pusher-700 in
1 wt % NaCl brine
250-300 m beads
E 1000 ppm xanflood in
1wt % NaCl brine
Berea sandstone
F 1000 ppm pusher-700 in
0.1 wt % NaCl brine
Berea sandstone


Table 2. A Petrographic Summary of the Experimental Data Analyzed

Experiment Porosity (%) Permeability (md)
A 39.0 4000
B 39.0 4200
C 39.3 3600
D 37.4 37900
E 19.7 312
F 22.0 530


Table 3. Statistical Summary of Regression Parameters

Experiment r

2
r

2

A 5.130 E-6 0.4514 0.9947 1.1367E-3 0.9955 0.9953
B 7.261 E-3 -0.2954 0.9810 6.3600E-2 0.9836 0.8437
C 3.005 E-4 -0.0523 0.9715 3.1410E-1 0.9915 0.8906
D 5.171 E-3 -0.1975 0.9831 3.0892E-3 0.9880 0.8524
E 3.790 E-6 0.5725 0.9963 3.2970E-6 1.561 0.9989
F 4.442 E-4 -0.01023 1.0000 2.7072E-6 1.033 1.0000



Table 4. Rheological Data Summary of all Experiments Analyzed

Experiment
n m 1000 /(1- n ) 1/ n
A 0.6902 45.326 1.4570 1.4488
B 1.3512 1.7260 0.8411 0.7400
C 1.0558 19.992 0.9408 0.9471
D 1.2495 2.9221 0.7916 0.8003
E 0.6400 31.190 1.5903 1.5625
F 1.0100 36.600 1.0230 0.99009
SPE 102015 9

FIGURE CAPTIONS


0.1
1
10
100
1000
1 10 100 1000 10000 100000
SHEAR RATE (1/sec)
S
H
E
A
R

S
T
R
E
S
S

(
P
a
)

Figure 1: Shear stress versus shear rate profile for experiment A (100-110 m beads-pack flooded with 1000 ppm xanflood in 1wt %
ppm NaCl brine solution.





1
10
100
1000
10000
100000
1000000
1 10 100 1000 10000
SHEAR RATE (1/sec)
S
H
E
A
R

S
T
R
E
S
S

(
P
a
)


Figure 2. Shear stress versus shear rate profile for experiment B (100-110 m beads-pack flooded with 1000 ppm pusher-700 in 1 wt
% NaCl brine solution).
10 SPE 102015

0.0001
0.001
0.01
0.1
1
10
0.0001 0.001 0.01 0.1 1 10
FLOW RATE (cm
3
/sec)
D
E
B
O
R
A
H

N
U
M
B
E
R
Experiment B
Experiment A

Figure 3. Comparing Deborah number for experiments A and B (100-110 m bead packs flooded with 1000 ppm xanflood in 1 wt %
NaCl brine, and 1000 ppm pusher-700 in 1 wt % NaCl brine solution, respectively).







0.001
0.01
0.1
1
10
0.0001 0.001 0.01 0.1 1
FLOW RATE (cm
3
/sec)
D
E
B
O
R
A
H

N
U
M
B
E
R
Experiment B
Experiment C


Figure 4. Comparing Deborah number for experiments B and C (100-110 m bead packs flooded with pusher-700 in 1 wt % NaCl,
and with pusher-700 in 0.1 wt % NaCl brine solution, respectively).
SPE 102015 11
0.000001
0.00001
0.0001
0.001
0.01
0.1
1
0.0001 0.001 0.01 0.1 1 10
FLOW RATE (cm3/sec)
Nv
Experiment B
Experiment A

Figure 5. Viscoelasticity number versus flow rate for experiments A and B (100-110 m bead packs flooded with xanflood and
pusher-700 in 1 wt % NaCl brine solutions, respectively).





1.0E-05
1.0E-04
1.0E-03
1.0E-02
1.0E-01
0.001 0.01 0.1 1 10 100
PRESSURE GRADIENT TERM
F
L
O
W

R
A
T
E

T
E
R
M
n / 1
n / 1
L
P
versus m
A
Q

Figure 6. Flow rate term versus pressure gradient term for experiment D (250-300 m beads-pack flooded with 1000 ppm pusher-700
in 1 wt % NaCl brine solution).
12 SPE 102015
1.0E-06
1.0E-05
1.0E-04
1.0E-03
1.0E-02
1.0E-01
1.0E+00
0.001 0.01 0.1 1 10 100
PRESSURE GRADIENT TERM
F
L
O
W

R
A
T
E

T
E
R
M
n / 1
n / 1
L
P
versus m
A
Q

Figure 7. Flow rate term versus pressure gradient term for experiment A (100-110 m beads-pack flooded with 1000 ppm xanflood in
1 wt % NaCl brine solution) .





1.00E-07
1.00E-06
1.00E-05
1.00E-04
1.00E-03
0.1 1 10 100
PRESSURE GRADIENT TERM
F
L
O
W

R
A
T
E

T
E
R
M
n / 1
n / 1
L
P
versus m
A
Q

Figure 8. Flow rate term versus pressure gradient term for experiment E (Berea sandstone flooded with 1000 ppm xanflood in 1 wt
% NaCl brine).
SPE 102015 13

N
v
= 5.13E-06(P/L)
0.4514
1.0E-06
1.0E-05
1.0E-04
0.1 1 10 100 1000
PRESSURE GRADIENT (psi/ft)
N
v

Figure 9. Viscoelasticity number versus pressure gradient for experiment A (100-110 m beads-pack flooded with 1000 ppm xanflood
in 1 wt % NaCl brine solution).






Nv = 3.79E-06(P/L)
0.5725
1.0E-06
1.0E-05
0.1 1 10
PRESSURE GRADIENT (psi/ft)
Nv


Figure 10. Viscoelasticity number versus pressure gradient for experiment E (Berea sandstone flooded with 1000 ppm xanflood in 1
wt % NaCl brine solution).
14 SPE 102015

N
v
=3.005E-4(P/L)
-0.0523
0.0001
0.001
0.1 1 10 100 1000
PRESSURE GRADIENT (psi/ft)
Nv

Figure 11. Viscoelasticity number versus pressure gradient for experiment C (100-110 m beads pack flooded with 1000 ppm pusher-
700 in 0.1 wt % NaCl brine solution).





Nv = 4.442E-04(P/L)
-0.01023
0.0001
0.001
0.1 1 10
PRESSURE GRADIENT (psi/ft)
Nv


Figure 12. Viscoelasticity number versus pressure gradient for experiment F (Berea sandstone flooded with 1000 ppm pusher-700 in
0.1 wt % NaCl brine solution).