Вы находитесь на странице: 1из 6

Analyzing Injectivity of Polymer

Solutions With the Hall Plot


R.S. Buell, SPE, Chevron U.S.A.; H. Kazeml, SPE, Marathon Oil Co.; and F.H. Po.ttmann, SPE
Colorado School of Mines '
Summa..,. The Hall plot was originally used to analyze water-injection wells. This paper demonstrates that the Hall plot can also
be used to analyze injection of polymer solutions. In particular, it is possible to determine the in-situ and residual resistance factors
?f a polymer solution r o ~ the Hall plot. The analysis methods developed are used to examine two field injection tests and one hypothet-
Ical example. The analytical results are verified with a reservoir simulator.
Introduction
Polymer floods, micellarlpolymer floods, and injectivity- or
productivity-profIle-modification treatments are the most common
applications of polymer solutions. The interpretation of injection
pressures and rates associated with polymer solution injection is
important to the efficient application of the solutions. The Hall
plot
l
-
3
is a useful tool for evaluating performance of injection
wells.
The Hall plot was originally developed for single-phase, steady-
state, radial flow of Newtonian liquids. Since the advent of poly-
mer and micellar solutions for EOR, it has also been applied to
the injection of these solutions. Moffitt and Menzie
4
used the Hall
plot to evaluate injection of polymer solutions but did not verify
the validity of the Hall plot for this application. This paper verifies
the validity of the Hall plot for evaluating polymer solution injection.
Because of the complex nature of polymer solution flow through
porous media, exact analytical solutions are generally not possi-
ble. However, some relatively simple approximate analytical so-
lutions can be developed. To verify the analytical solutions for
polymer solution injection, a two-phase, radial, numerical reser-
voir simulator was developed.
2
The simulator is designed to con-
sider the more important phenomena and effects that occur when
polyacrylamide or polysaccharide polymer solutions are injected
into porous media. The simulator has the following characteristics:
slightly compressible flow, two-phase flow, non-Newtonian rheol-
ogy, adsorption/retention with permeability reduction, concentra-
tion effects, skin, and wellbore storage. It was used to history match
two field injectivity data sets.
Development of the Hall Plot
The Hall I plot was originally proposed to analyze the performance
of waterflood injection wells. Hall simply used Darcy's law for
single-phase, steady-state, Newtonian flow of a well centered in
a circular reservoir:
q ........................ (1)
14I.2B
w
l'w[ln(r
e
lr w) +s]
Hall integrated both sides with respect to time to obtain
Wi
Separating the integral of Eq. 2, Hall then rearranged to obtain
141.2B
w
l'w[ln(r elr w) +s]
Wi+i'Pedt . ........... (3)
kkrwh
The relation between surface and bottomhole pressures for steady-
state vertical flow is given by
pwf=ptrtlpf+pgD . ............................... (4)
Copyright 1990 Society of Petroleum Engineers
SPE Reservoir Engineering, February 1990
Hall substituted Eq. 4 into Eq. 3 to arrive at
141.2B
w
l'w[ln(r elr w) +s]
--------Wi +J(Pe + tlprpgD)dt.
kkrwh
.................................... (5)
Hall simply dropped the second term on the right side ofEq. 5 and
plotted the integral of wellhead pressures with respect to time vs.
cumulative injection, which came to be known as the "Hall plot."
By plotting in this format, Hall observed that if an injection well
was stimulated, the slope decreased, and if a well was damaged,
the slope increased. While Hall's conclusions regarding changes
in slope are valid, the second term on the right side of Eq. 5 is
often not negligible in comparison with the other terms and there-
fore usually cannot be dropped.
In industry applications, the Hall integrals i'Pifdt and i'Pwfdt fre-
quently are used. The slopes calculated from these integrals should
not be used for quantitative calculations unless a correction proce-
dure is applied. Fig. 1 is a Hall plot based on the data for Well
A, where the integral fPwfdt has been plotted vs. cumulative in-
jection. Several changes in slope can be seen on the plot, but there
has been no change in transmissibility or skin. The changes in slope
are caused by changes in rate, which occur because the integral
fPedt has been neglected. Fig. 2 is a Hall plot based on data for
Well C. The three most common forms of the Hall integral have
been plotted for the same data. For each integration method, the
slopes of the curves are quite different.
Injection data must be plotted in the form of Eq. 2 to make valid
quantitative calculations; i.e., cumulative injection should be plot-
ted vs. f(Pwf-Pe)dt. The slope of the Hall plot from Eq. 2 is then
given by
1412B
w
l'w[ln(r elr w) +s]
..................... (6)
kkrwh
Eq. 6 will not be appropriate when multiple fluid banks with sig-
nificantly different properties exist in the reservoir.
Advantage. and DI.advantage.
The Hall plot is a steady-state analysis method, whereas falloff tests,
injection tests, and type-curve analysis are transient methods. Tran-
sient pressure analysis methods determine the reservoir properties
at essentially one point in time. The Hall plot is a continuous mon-
itoring method; i.e., reservoir properties are measured over a period
of weeks and months. The Hall plot, therefore, can help identify
changes in injection characteristics that occur over an extended
period.
Hall's method has several advantages. Integrating the pressure
data with the Hall integral [f(Pwf-Pe)dt] has a smoothing effect
on the data. Data acquisition for the Hall plot is inexpensive be-
cause only the recording of cumulative injection and surface pres-
sures is required. Surface pressures must be converted to bottomhole
pressures (BHP's), correcting for hydrostatic head and friction loss-
41

350 bbl/d
-
100
'r
Iwfd!
/
I
I VV,
'
I I/V I I
r !
=1=== I
I _--+- I I
----!----- I I I I I
' I , ,
0 4000 6000 8000 10000 ie-ODD HODO 16000
CUMULATIVE INJECTION (bbl)
Fig. 1-Comparlson of Hall Integration methods, Well A,
P. = 1,000 psi.
es. Injection and falloff tests usually require running gauges on wire-
line to depth, which is an additional expense.
The greatest disadvantage of the Hall plot is that the skin, s, and
transmissibility, khl/Jo' are combined in the slope. It is possible to
determine one of these if the other is known, but the determination
of both skin and transmissibility is not possible with the Hall plot.
To use the Hall plot effectively, running falloff or injection tests
periodically is still necessary to determine the individual values of
transmissibility and skin.
Quantitative Analysis
Newtonian Fluids. In a mature waterflood, the transmissibility
usually will not change significantly with time; therefore, any change
in the slope of the Hall plot will be a result of skin effects. Assum-
ing no change in transmissibility, the new skin can be calculated
as follows for water injection:
kkrwh
s2 =sI - (mHI -mH2), .................... (7)
141.2B
w
/Jow
where subscript 1 denotes the old slope and skin and subscript 2
the new slope and skin. This relationship can be useful in recog-
nizing formation damage or fracturing.
When a waterflood begins, two-phase flow will exist in the near-
wellbore region. As the water moves away from the wellbore, water
and oil banks form if the oil saturation is large enough. A simpli-
fied method to analyze this situation is to apply Darcy's law in a
series manner. Because the oil displacement is governed by the
Buckley-Leverett
5
equation, the saturations and relative permea-
bilities are not constant within each fluid bank; however, for sim-
plicity, they can be assumed to be constant within each bank. The
accuracy of the results is not significantly compromised with this
assumption. The slope of the Hall plot for a water and oil bank
is given by
(water bank) (oil bank)
.................................... (8)
The interface between the oil and water banks is rbl. The inter-
face of the oil and water banks can be estimated with Eq. 9, which
results from the Buckley-Leverett equation in radial coordinates.
6
r1I = 5.615W
i
(a/
w
) +ra . ....................... (9)
q,7rh as
w
F
The quantity (a/wlaSw)F is the derivative of the fractional-flow
curve at the flood front. The water saturation and the derivative
of the fractional-flow curve at the flood front are determined with
42
.... &
I'w,d'
/
Go


y

Og

,/

//


a.
o
/'
.... g
....12

----

Co
a:
o
/
------
CI:il
w'"
1#
-------
I PHdt
.,.0

zg
-:il
V
:18
C:il
X"


... 10DO 2000 3000 '000 5000 6000 7000 8000
CUMULATIVE INJECTION (Barrels)
Fig. 2-Comparlson of Hall Integration methods, Well C.
Welge's7 method. As the oil bank is pushed away from the well-
bore, the water-bank term will dominate owing to the logarithmic
nature of Eq. 8.
Non-Newtonian Fluids. The analysis methods for non-Newtonian
fluids are similar to the methods developed in the previous section,
except permeability reduction must be considered. The apparent
viscosity of the non-Newtonian fluids is taken to be a constant within
each fluid bank. Eq. 10 is for an injection sequence of polymer
and then water. The reservoir is assumed to be initially oil-saturated.
Three fluid banks will be created: oil, polymer, and water.
[
/JowBw[ln(rb2Ir w) +s] /JopBw In(rbI lrb2)
mH=141.2 +--'------
hkakrw hkakrp
(water bank) (polymer bank)
+ /JooBo In(relrbl) I ............................... (10)
hkk
ro
j
(oil bank)
Eq. 10 can be rewritten with just one absolute permeability and
one aqueous-phase viscosity after the introduction of resistance fac-
tor, Rj' and residual resistance factor, Rrj' which are defined
below.
water mobility (kkrw)//Jow
R
j
= = ................ (11)
polymer mobility (kakrp)//Jop
absolute permeability before polymer k
and Rrj= .... (12)
absolute permeability after polymer ka
Because residual resistance factor and resistance factor are useful
in the evaluation of polymer performance, Eq. 10 has been rewrit-
ten with these definitions:
(water bank) (polymer bank)
+ /JooBo In(relrbl) I ............................... (13)
hkk
ro
j
(oil bank)
In Eqs. 10 and 13, apparent viscosity is assumed to be constant
through space; i.e., the non-Newtonian rheology is ignored. The
variation of apparent viscosity in space can be taken into account
by applying Darcy's law, with the definition of effective viscosity
used in the same series manner as used to develop Eq. 10. For the
simple case of a power-law fluid bank occupying the whole reser-
SPE Reservoir Engineering, February 1990
0
> : ~
II
~
:ag
e. ..
III
r Wat" I-
<Co
C : ~
./
III
... Polymer
0
0 1000.,.
:Z:o
:EN
)
0
Start of 1000 ppm polymer
~ o ,/
0
0
I
J
CD-
~
<>0 100 '00 '00 -laO 500 600 700 .00 900
TIME (days)
Fig. 3-Bottomhole Injection rate vs. time, Wen A.
voir, the Hall plot would be altered to S(PIIf-Pe)dt vs. Jqllndt. The
addition of the non-Newtonian rheology makes the equations more
difficult to handle for hand computations, particularly for multiple
fluid banks. We will demonstrate with Wells B and C that the ac-
tual change in apparent viscosity through space is relatively small
and can be approximated by a constant.
When the bank in contact with the wellbore has moved out a sub-
stantial distance, the other terms can often be dropped without sub-
stantial error, and the bank in contact with the wellbore can be
assumed to extend to the drainage radius. The individual evalua-
tion of the three terms in Eq. 13 can be used to determine whether
the bank in contact with the wellbore will dominate. The terms in
Eq. 10 and 13 can be rearranged to account for any injection se-
quence. The position of each fluid bank can be located with Eq.
9, or the fluid-bank position can be located by assuming piston dis-
placement for a miscible process.
For most field situations where a reservoir is under water, poly-
mer, or micellar/polymer flooding, the transient period usually does
not lastinore than a few days. During the transient flow period,
the slope of the Hall plot is changing, given constant fluid and reser-
voir properties. The transient period usually is not observed on the
Hall plot because pressures in the field are typically recorded only
once a day. Because the transient period typically lasts only a few
days in most polymer floods, it is difficult to observe the transient
effects in field data with such infrequent pressure readings.
Hall Piol Anal,sls for Well A
The data for Well A, a hypothetical example, are given in Refs.
2 and 8. The relative permeability data of Blair and Weinaug
9
were
used for two-phase flow. Fig. 3 illustrates the well's injection his-
tory. Water was injected for 51 days; then injection of I,OOO-ppm
polymer solution with CarreaulO,ll rheological properties was be-
gun and continued for the remainder of the injection. The func-
tional form of the Carreau rheological model is
ILp =1'00 +(1'0 -1'00)[1 +(A .y)2][(n-1)12], ............... (14)
where the coefficient n -1 = slope of the power-law region for a
Carreau fluid, 1'0 = viscosity at zero shear, and 1'00 =viscosity at
infinite shear. The intersection of the zero-shear region and the
power-law viscosity region occurs at 1IA.
CUMULATIVE INJECTION (bbl)
Fig. 4-Han plot, Wen A, comparison of single- and two-phase
flow.
To illustrate the effects oftwo-phase flow, two cases are shown.
In the first case, the reservoir began at irreducible water saturation
(two-phase flow). For the second case, the reservoir began at residu-
al oil saturation (ROS); therefore. this case degenerates to single-
phase flow. Fig. 4 compares the Hall plot for single- and two-phase
flow. The pressures for two-phase flow are higher than for single-
phase flow. The greater the oil viscosity, the more separation will
occur in Fig. 4. The simulator input of an absolute permeability
of 100 md and a relative permeability to water of 0.187 at ROS
results in 18.7 md. Applying Eq. 6 to the single-phase-flow curve
during water injection yields a water permeability of 18.7 md,
which, as expected, is identical to the simulator input. For two-
phase flow during water injection, Eq. 6 yields a water permeabil-
ity of 15.5 md because the oil bank is ignored. The permeability
to water of 18.7 md is closely matched if the oil bank is considered
in the slope calculation by use of Eq. 8.
At the end of polymer injection, the slope of the Hall plot is 58.1
(psi-D)/STB [1.34 (kPa' d)/stock-tank m
3
] for single-phase flow.
The curve for two-phase flow has a slope of63.1 (psi-D)/STB [1.45
(kPa' d)/stock-tank m
3
]. The difference between the two slopes is
8.6%. As time continues, the slope of the two-phase Hall plot curve
will approach the slope of the single-phase curve. As the oil bank
is pushed farther from the wellbore, less error will occur with the
single-phase-flow assumption. Where large mobility contrasts ex-
ist or the oil bank is close to the wellbore, Eq. 8 should be used.
Hlslory Malchlng and Anal,sls
WeD B. Well B is based on the data published by Milton et al.
12
The data come from an evaluation of a reservoir in the Big Horn
basin, WY, for polymer flooding. The Milton et al. data set was
selected because detailed reservoir data, daily pressures and rates,
and rheological data were provided. The injection pressures were
also limited to prevent reservoir fracturing or fracture parting. Two
falloff tests were run to determine skin and permeability. Data on
Well B can be found in Refs. 2, 8, and 12. The polymer injection
sequence consisted of water injection followed by a polyacrylamide
polymer solution. The polyacrylamide polymer solution was dis-
placed by water. When the performance of this well was history
matched, all the parameters supplied by Milton et al. except two
were used as input for the simulation runs. The rheological data
TABLE 1-HISTORY MATCH, APPARENT VISCOSITY AS A FUNCTION OF
INTERSTITIAL VELOCITY, WELL B
Concentration
Interstitial Velocity (ftlO)
(ppm) 0.01 0.10 1.00 10.00 100.0 1.000.0
--
0.0 0.70 0.70 0.70 0.70 0.70 0.70
250.0 3.50 3.50 2.30 1.40 0.70 0.70
500.0 5.00 5.00 3.00 1.80 0.70 0.70
1,000.0 10.00 10.00 5.50 2.00 0.70 0.70
2,000.0 45.00 45.00 20.00 6.00 0.70 0.70
SPE Reservoir Engineering, February 1990 43
CUMULATIVE INJECTION (barrels)
Fig. 5-Hall plot, rate-contrOlled history match, Well B.
and resistance factor estimates provided by Milton et ai. were used
as a first approximation and were then adjusted to obtain the best
possible history match. All pressure data were recorded at the sur-
face. For history-matching purposes, all surface pressures were cor-
rected to BHP with friction included.
The residual resistance factor was estimated to be 1.05 on the
basis of transient well testing. To obtain the best possible match
with the field data, reducing the polymer solution viscosity and in-
creasing the residual resistance factor was necessary. The best his-
tory match was obtained with the rheology given in Table I and
a residual resistance factor of 1.33.
The rate-controlled (Neumann) boundary condition and match-
ing pressures were used to obtain Figs. 5 and 6. The reservoir was
modeled with single-phase flow because Milton et ai. considered
the reservoir to be at ROS owing to extensive waterflooding. The
Hall plot generated from history matching can now be used to ap-
ply the analytical procedures developed. The Hall plot shown in
Fig. 5 has three distinct sections: water, polymer, and water injec-
tion. Applying Darcy's law in a series manner yields an equation
for each section. For this example, four unknowns will be assumed:
kkrw, Rrf, Rfl, and Rj2. All other parameters are assumed to be
known. Rfl is the resistance factor of the polymer bank while poly-
mer solution is being injected. Rj2 is the resistance factor of the
polymer bank after polymer injection has stopped and water injec-
tion has begun. Rfl and Rj2 usually will not be equal because of
shear thinning of the polymer solution and because adsorption/reten-
tion will reduce polymer concentrations as the polymer slug prop-
agates through the reservoir. The slope of the Hall plot for the first
water-injection period is given by Eq. 6. The slope of the Hall plot
for the polymer-injection period is given by
(polymer bank) (water bank)
................................... (15)
An oil bank is assumed not to form in Eq. 15 because of the exten-
sive waterflooding. The slope of the Hall plot for water injection
following polymer injection is given by
[
R,frwBw[ln(rb2Ir w) +s] Rj2JlwBw In(rbl /rb2)
mH3 = 141.2 +
hkkrw hkkrw
(water bank) (polymer bank)
+ J ............................. (16)
(water bank)
Eq. 8 is used to calculate a permeability to water of 90.9 md with
a skin of7.2. The permeability to water used by the reservoir simu-
lator is 91.0 md. With one unknown eliminated, there are now three
44
i

--0 10 20 30 40 50 60 70 80 90 100
TIME (days)
Fig. 6-BHP vs. time, rate-controlled history match, Well B.
unknowns and only two equations. To solve for all the unknowns,
another equation is necessary. The residual resistance factor can
be estimated by taking the ratio of the Hall plot slopes for water
injection before and after polymer injection. The contributions from
the banks farther away from the wellbore become smaller and
smaller as injection continues when Eq. 16 is used. As injection
proceeds, the Hall plot slope after polymer injection, mH3, will ap-
proach the value given by
14I.2R
rf
JlwBw[ln(r e1r w) +s]
mH3=
................. (17)
hkkrw
(water bank)
The latest straight-line portion of water injection following poly-
mer is used to estimate mH3, so the influence of the other banks
will be at a minimum. Eqs. 6 and 17 can be combi:1ed to solve
for Rrf. The residual resistance factor can then be estimated with
Rrf=m
H3
/mH!' .................................. (18)
The residual resistance factor is computed to be 1.42. The in-
situ residual resistance factor calculated in the simulator is 1.33.
Eqs. 15 and 16 can be used to find the two remaining unknowns,
Rfl and Rj2. Because this problem was simulated with one-phase
flow, all displacement processes are miscible. Piston-like displace-
ment, therefore, is assumed to occur. The location of the interface
between banks can be calculated by volumetric calculations account-
ing for polymer adsorption. Using Eq. 15 to calculate the resistance
factor at the end of the polymer-injection period results in a
resistance factor of 2.22 for the polymer bank. It can be seen that
when numerical values are substituted into Eq. 15, the water bank
is less important than the polymer bank, supporting the conclusion
that the bank in contact with the wellbore will dominate. The water
bank away from the wellbore can be assumed to be negligible, and
the single-fluid-bank assumption can be used. When this single-
fluid-bank assumption is used, the resistance factor is calculated
to be 2.06. The single-fluid-bank assumption will underestimate
the resistance factor because, in this case, the polymer bank is
assumed to extend to the drainage radius when it actually extends
only some fraction of the drainage radius. The simulator provides
pressures for each cell. The average resistance factor for the poly-
mer bank is calculated with the simulator results to be 2. 17.
The slope of the water injection following polymer is used to solve
Eq. 16. Substituting numerical values into Eq. 16 results in a cal-
culated resistance factor of 4.19 for the polymer bank away from
the wellbore. The average resistance factor of the polymer bank
is calculated with the simulator results to be 3.51. The resistance
factor calculated for the polymer bank away from the wellbore can
be significantly in error. The reason for the larger errors is that
the pressure drop caused by the polymer bank is small away from
the wellbore. A small error in the determination of the slope re-
sults in an increased error in the calculated resistance factor. Ta-
ble 2 compares the approximate analytical methods with the
simulator results.
SPE Reservoir Engineering, February 1990
Well C. Well C was used to evaluate the injectivity of micellar
and polymer solutions. The daily injection data consisted of mice 1-
lar solution injection followed by polymer solution injection. The
polymer solution was then displaced with water. The reservoir data,
fluid properties, daily injection history, and polymer parameters
are given in Refs. 2 and 8.
The injection pressures were controlled to prevent reservoir frac-
turing or fracture parting. A falloff test was run before micellar
solution injection began. The test indicated a water mobility of 27
md/cp [27 md/mPa' s] and a skin of -1.14. The skin and permea-
bility calculated from the falloff test were used in the simulator for
history matching. This reservoir had been waterflooded extensive-
ly before the injection testing. The reservoir was estimated to be
at ROS; therefore, the history match was done with only single-
phase flow.
The history match was conducted in the same manner as for Well
B. The best history match was obtained by adjusting rheology and
resistance factors. All other parameters were taken from available
data and assumed to be correct. The Carreau model was used to
approximate the rheology of the polymer and micellar solutions.
History matching was done with both the rate- and pressure-
controlled boundary conditions.
The Neumann (rate-controlled) boundary condition was used for
Figs. 7 and 8. Table 3 gives the rheology of the polymer solution
used to obtain the best match. The amount of permeability reduc-
tion was much larger than for Well B. The residual resistance fac-
tor used in the best match was 11. 1. The rate-controlled boundary
condition was used for cases with various levels of adsorption/reten-
tion. As with Well B, the results were found to be relatively insen-
sitive to the amount of adsorption/retention.
The Hall plot for Well C was analyzed in the same manner as
that for Well B. There were no initial water-injection data for this
well; however, the slope before polymer and micellar solution in-
jection can be calculated because the skin and transmissibility are
known. The Hall plot slope for water injection is calculated with
Eq. 6 to be 1.68 (psi-D)/STB [0.0387 (kPa' d)/stock-tank m
3
] from
the falloff testing data before polymer injection. The slope for the
late water-injection period is 16.50 (psi-D)/STB [0.380
(kPa'd)/stock-tank m
3
]. Eq. 18 can now be used to estimate a
residual resistance factor of 9.80, which is reasonably close to the
simulator value of 11. 10. Table 4 compares the analytical solutions
with the simulator results.
Eqs. 15 and 16 can now be used to calculate the average resistance
factor of the polymer/micellar solution banks. At the end of poly-
mer/micellar solution injection, the average resistance factor is cal-
culated with Eq. 15 to be 29.53. The average resistance factor can
also be approximated by assuming a single bank that extends to
the drainage radius, which results in an average resistance factor
of20.95. The average resistance factor calculated with the simula-
tor is 29.40.
After water injection, the polymer/micellar bank is between 48
and 95 ft [15 and 29 m]. The simulator results were used to calcu-
late an average resistance factor of 14.7. A resistance factor of 16.6
was calculated with Eq. 16.
The apparent viscosity for Well C is given as a function of radial
distance in Ref. 8. The relative change in the apparent viscosity
within the polymer bank is small. Small changes in apparent vis-
TABLE 2-COMPARISON OF ANALYTICAL METHOD
WITH SIMULATOR RESULTS, WELL B
Analytical Methods, Analytical Methods,
Parameter Multiple Banks Single Bank
kkrw 90.9 90.9
R" 1.42 1.42
R 11 2.22 2.06
Rf2 4.13
~
0
0
'"
~
V
:g
/'
." 0
, 0
., 0
0; ....
/v
Q.
:g
Simulatot Match
: ~
/
..
Q.
Field Cat ... 0
:I g
.s ~ .
If'
:=; Mic_Uar SO'U
L
:::l g
I/) 0
-I V
"'. N
7-polymer t---.- -- ---- -Wat&r- 1---
Simulator
91.0
1.33
2.17
3.51
v
f - - - ~
0
0 1000 ZOOO 3000 4000 5000 6000 7000 BODO
CUMMULATIVE INJECTION (barrels,
Fig. 7-Hall plot, rate-controlled history match, Well C.
40 50 60
TIME (days,
Fig. 8-BHP vs. time, rate-controlled history match, Well C.
TABLE 3-HISTORY MATCH, APPARENT VISCOSITY AS A FUNCTION OF
INTERSTITIAL VELOCITY, WELL C
Concentration
Interstitial Velocity (ft/O)
(ppm) 0.01 0.10 1.00 10.00 100.0 1,000.0
--
0.0 1.00 1.00 1.00 1.00 1.00 1.00
2,445.0 3.00 2.85 2.01 1.55 1.39 1.33
2,800.0 5.00 4.72 3.07 2.19 1.87 1.76
3,100.0 6.99 6.59 4.19 2.91 2.45 2.29
3,430.0 8.49 7.83 4.30 2.83 2.44 2.34
5,000.0 19.98 18.70 12.55 10.51 10.10 10.02
'Savins 13 shear-rate/velocity relation used.
SPE Reservoir Engineering, February 1990
45
TABLE 4-COMPARISON OF ANALYTICAL METHOD
WITH SIMULATOR RESULTS, WELL C
Analytical Methods, Analytical Methods,
Parameter Multiple Banks Single Bank
Rtf 9.80 9.80
R 11 29.53 20.95
R'2 16.63
Simulator
11.10
29.40
14.73
cosity within the polymer bank also occurred with Well B. Accord-
ing to the history matching of Wells B and C, the non-Newtonian
flow effect is relatively small. Apparent viscosity stays relatively
constant within the polymer bank.
Conclusions
Using a numerical reservoir simulator,
2
we have demonstrated that
quantitative analysis can be performed on the Hall plot when non-
Newtonian solutions are injected. The best method for analyzing
the Hall plot would be to use a reservoir simulator like that devel-
oped for this study. For the practicing engineer, however, a simu-
lator may not be available. Therefore, two approximate analysis
methods for the Hall plot have been developed to estimate permea-
bilities, resistance factors, and residual resistance factors.
1. The Hall plot can be used to estimate the performance charac-
teristics of injected polymer and micellar/polymer solutions.
2. The multibank analysis method will yield more accurate an-
swers than the single-bank method. When the fluid bank in contact
with the wellbore has moved out a substantial distance, the single-
fluid-bank analysis method can be used with acceptable accuracy.
3. The non-Newtonian rheology effect is small because the change
in the in-situ apparent viscosity of the polymer solutions through
space is relatively small.
4. The amount of permeability reduction has a significant effect
on the Hall plot and simulator results.
5. The transient flow period has little effect on the Hall plot be-
cause, in most field situations, the transient period rarely lasts more
than a few days. Because most Hall plot data are recorded daily,
it is usually not possible to observe the transient flow period on
the Hall plot.
Nomenclature
B = FVF, dimensionless
D = true vertical hole depth, ft [m]
fw = fractional flow of water, dimensionless
g = gravity constant, 32.2 ft/sec
2
[9.81 m/s2]
h = formation thickness, ft [m]
k = absolute permeability, md
ka = absolute permeability after polymer, md
k
ro
= relative permeability to oil, dimensionless
krp = relative permeability to polymer, dimensionless
krw = relative permeability to water, dimensionless
mH = Hall plot slope, (psia-D)/STB [(kPa 'd)/stock-tank m
3
]
n = Carreau and power-law-fluid slope parameter,
dimensionless
Pe = pressure at external drainage radius, psia [kPa]
Ptf = surface tubing injection pressure, psia [kPa]
Pwf = bottomhole injection pressure, psia [kPa]
t::..pf = pressure loss caused by friction, psi [kPa]
q = rate, BID [m
3
/d]
rbl = Bank I radius, ft [m]
rb2 = Bank 2 radius, ft [m]
re = external drainage radius, ft [m]
r w = wellbore radius, ft [m]
R
f
= resistance factor, dimensionless
Rrf = residual resistance factor, dimensionless
s = skin, dimensionless
S = saturation, dimensionless
46
t = time, days
ilt = change in time, days
Wi = cumulative injection, bbl
"y = shear rate, seconds-\
A = Carreau rheological parameter, seconds
/./, = viscosity, cp [mPa's]
/'/'e = effective viscosity, cp [mPa's]
/'/'0 = viscosity at zero shear, cp [mPa's]
/./,00 = viscosity at infinite shear, cp [mPa 's]
p = fluid density, Ibm/ft3 [kg/m3]
cf> = porosity, dimensionless
Subscripts
F = flood front
o = oil
P = polymer
w = water
wf = injection pressure at r w
Acknowledgments
We thank the Colorado School of Mines Petroleum Engineering
Dept., Chevron U.S.A., and Marathon Oil Co. for their support
in presenting and preparing this paper.
References
1. Hall, H.N.: "How To Analyze Waterflood Injection Well Perform-
ance," World Oil (Oct. 1963) 128-30.
2. Buell, R.S.: "Analyzing Injectivity of Non-Newtonian Fluids: An Ap-
plication of the Hall Plot," MS thesis, Colorado School of Mines, Gold-
en, CO (1986).
3. DeMarco, M.: "Simplified Method PintJoints Injection Well Problems,"
World Oil (1968) 95-100.
4. Moffitt, P.D. and Menzie, D.E.: "Well Injection Tests of Non-
Newtonian Fluids, " paper SPE 7177 presented at the 1978 SPE Rocky
Mountain Regional Meeting, Cody, WY, May 17-19.
5. Buckley, S.E. and Leverett, M.C.: "Mechanism of Fluid Displace-
ment in Sands," Trans., AIME (1942) 146, 107-16.
6. Collins, R.E.: Flow of Fluids Through Porous Marerials, Petroleum
Publishing Co., Tulsa, OK (1961) 149.
7. Welge, H.J.: "Simplified Method for Computing Oil Recoveries by
Gas or Water Drive," Trans., AIME (1952) 91, 195-98.
8. Buell, R.S., Kazemi, H., and Poettmann, F .H.: "Analyzing Injectivi-
ty of Polymer Solutions with the Hall Plot, " paper SPE 16963 presented
at the 1987 SPE Annual Technical Conference and Exhibition, Dallas,
Sept. 27-30.
9. Blair, P.M. and Weinaug, C.P.: "Solution of Two-Phase Flow Prob-
lems Using Implicit Difference Equations," SPEl (Dec. 1969) 417-24;
Trans., AIME, 246.
10. Carreau, J.P.: "Rheological Equations from Molecular Network The-
ories," PhD dissertation, U. of Wisconsin, Madison, WI (1968).
11. Vogel, P. and Pusch, G.: "Some Aspects of the Injectivity of Non-
Newtonian Fluids in Porous Media," Enhanced Oil Recovery, F. John
Fayers (ed.), Elsevier Science Publishers, New York City (1981)
179-96.
12. Milton, H.W. Jr., Argabright, P.A., and Gogarty, W.B.: "EOR Pros-
pect Evaluation Using Field-Manufactured Polymer," paper SPE 11720
presented at the 1983 SPE California Regional Meeting, Ventura. March
23-25.
13. Savins, J.G.: "Non-Newtonian Flow Through Porous Media," Ind.
& Eng. Chern. (Oct. 1969) 61, No. 10, 18-47.
51 Metric Conversion Factors
bbl x 1.589 873 E-Ol
ft x 3.048* E-Ol
psi x 6.894 757 E+OO
'Conversion factor is exact. SPERE
Original SPE manuscript received for review Sept. 27. 1987. Paper accepted for publica-
tion Nov. 2, 1989. Revised manuscript received June 22, 1989. Paper (SPE 16963) first
presented at the 1987 SPE Annual Technical Conference and Exhibition held in Dallas,
Sept. 27-30.
SPE Reservoir Engineering, February 1990

Вам также может понравиться