SPE 25880
Society of Petroleum engineers
Using Polymer Injectivity Tests To Estimate Fracture Porosity in Naturally Fractured Reservoirs
J.R. Gilman, S.B. Hinchman, and M.A. Svaldi, Marathon Oil CO.
SPE Members
Copyright 1993, Society of Petroleum Engineers, Inc.
This paper was prepared for presentation at the SPE Rocky Mountain Regional/Low Permeability Reservoirs Symposium held in Denver, CO, U.S.A., April 1214, 1993.
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. Permission to copy is restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment of where and by whom the paper is presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 750833836, U.S.A. Telex, 163245 SPEUT.
ABSTRACT
A new technique is presented to calculate fracture porosity in naturally fractured reservoirs using composite system analyses.
The concept outlined here is to inject a viscous solution (e.g. polymer) which will not invade the matrix rock, but remains in the fractures. Subsequent falloff tests can then be used to determine the bank radius of the polymer in the fractures and thus fracture porosity from material balance. In the early 1980's Marathon Oil Company performed a number of polymer injectivity tests in naturally fractured reservoirs. In several cases, apparent bank radius was much larger than anticipated. This observation lead to the technique proposed in this paper.
INTRODUCTION
A difficult problem in characterizing naturally fractured reservoirs is determining the fraction of the pore volume which can be assigned to the fracture network (i.e. fracture porosity). In tight formations, pressure transient testing can give a measure of the relative fracture storage capacity of the formation providing that wellbore storage is minimized. In order to estimate fracture storage capacity from pressure transient testing at least a portion of the
References and Illustrations at end of paper.
transition from fracture response to total system response must be seen on the pressure transient plot. However, in many naturally fractured reservoirs, the system may begin behaving like a homogeneous, singleporosity system in a very short time as given by the following equation 1:
t> 11377 (cj)ct)m J1
kmcr
(1 )
where (for cubical matrix blocks)
(2)
For moderate matrix permeability (>10 md) and high fracture intensity (cr > 0.1), the end of the transition period may be on the order of several seconds as calculated by Eq. 1 and thus may be masked by wellbore storage. Thus, in many highly fractured oil reservoirs, it is only possible to determine the total system effective permeability from a singlewell transient test and total porositycompressibility product from an interference test. In this paper, we propose the use of injectivity and falloff tests combined with composite system transient analysis to estimate the total porosity of the fracture network.
TECHNIQUE AND ASSUMPTIONS
The most basic application of this technique consists of injecting a fluid into a naturally fractured reservoir and then performing a falloff test. The injected fluid must have a mobility different than the insitu fluid and only occupy the fractures. Composite system pressure transient analysis is then used to estimate the bank radius
429
Using Polymer Injectlvlty Tests to Estimate
2 Fracture Porosity In Naturally Fractured Reservoirs
SPE025880
and material balance is used to estimate the fracture porosity. Throughout this paper we will assume that the injected fluid is a polymer solution and the insitu fluid is water.
As with all pressure transient analysis methods, we must make numerous simplifying assumptions about the system we are analyzing. This analysis assumes that the injected polymer solution behaves as a Newtonian fluid. OUr experience has shown that for many polymers in high salinity solutions, this is a reasonable assumption. As mentioned previously, the polymer solution is assumed not to invade the matrix or to be adsorbed on the fracture wall. The fracture geometry must be such that the flow in the fractures is essentially radial in nature during the test and the system behaves as a singlephase, homogeneous dualporosity reservoir of infinite extent. For the examples shown here, we assume no skin damage or wellbore storage, but these could be included in a straight forward manner. It is also required that the dualporosity transition period be over prior to the time at which the injected bank affects the transient response.
LIMITATIONS
A great deal of literature has been devoted to the topic of composite system analysis. ^{3}  ^{1}^{7} In this paper we use the intersection time of two semilog straight lines assuming an equivalent singleporosity system to estimate the distance to the bank radius. ^{4} • ^{6} ,9,14 Other methods to estimate the bank radius are discussed in the literature including dualporosity system analysis. ^{1}^{6}  ^{1}^{7} The intersection point technique works very well for our examples provided that two semilog straight lines develop. In order to estimate the bank radius, total system compressibility must be known. This is most accurately determined from an interference test. We also need to have some reasonable estimate of fracture porosity in order to properly design the test. Because of wellbore storage, limited test time, and boundary effects, it is easy to miss the double slope behavior of a composite system if the test is not properly designed. The usefulness of our technique depends on the ability to estimate the bank radius from pressure transient analysis. To minimize uncertainty in the intersection time, it is necessary to have reliable estimates of the injected and insitu mobilities. The technique will work best when injected mobility is greater than insitu mobility to prevent dispersion of the injected bank. Also, 'as mentioned previously, the injected fluid cannot enter the matrix rock. This means that this test is well suited for viscous polymer injection in water injection wells. Imbibition and gravity forces will be minimal in these situations and therefore, there will be little potential for the injected polymer solution to enter the matrix.
TEST DESIGN CONSIDERATIONS
The injection time should be designed such that both straight lines can be seen after the wellbore storage period and prior to boundary effects. This requires a preliminary estimate of fracture porosity. For example, if we desire the intersection time of the two straight lines to be in the range of 1 to 10 hours, then Eqs. 3 and 4 (shown later) can be used to determine the bank radius and thus the amount of injected polymer required to move the bank to this radius. Multiple tests with injection banks less than and greater than estimated should be run to cover the range of possible lower and higher fracture porosities. A falloff test prior to polymer injection is useful for estimating the true mobility of the water bank. A falloff test after an extended period of polymer injection is important for estimating the mobility of the polymer bank. Knowing these values will help to determine where to draw the straight lines on the semilog pressure plots and thus more easily locate the intersection time and bank radius.
SIMULATED EXAMPLES AND ANALYSIS TECHNIQUE
Injectivity and falloffs tests using the test sequence given in Table 1 were simulated using a singlephase, dualporosity simulator with capabilities for polymer transport. Fifty exponentially distributed radial grid blocks were used in the simulator. The polymer in these examples is assumed to behave as a Newtonian fluid. Six cases were run using all combinations of fracture porosity and polymerwater mobility ratios given in Table 2. All other data remained constant as shown in Table 2. The bank radius was evaluated using the standard technique of intersection time of the late and early straight Iines. ^{4} ,6,9,14 Because these examples are computer generated, the data is smooth and the true straight lines are well known. The method to estimate bank mobility as described by Yeh and Agarwal,3 also works well for estimating bank radius as will be shown. The difficulty of estimating bank radius from actual field injectivitylfalloff tests is not discussed in this paper.
The falloff tests were analyzed using equivalent time as defined by Agarwal ^{2} :
Ate =
lin" At
J
tinj + At
•
(3)
Intersection time of the two semilog straight lines (with slopes mi and rna) was used to estimate bank radius as given by the following equation ^{4} _{:}
(4)
430
SPE025880
J. R. Gilman, S. B. Hinchman and M. A. Svaldl
3
where the subscripts "i" and "0" represent the inner and outer bank.
Fracture porosity is then estimated from this bank radius using material balance. We assume injected volume q B tinj only enters the fractures and is thus equal to the
volume
in
the
fractures
to
(i.e.• ell 1t (tir~)h/5.6146).
a
bank
of
~
cj)r =5.6146 q B tinj
1t(tir~)h
(5)
These equations show that the fracture porosity estimate is inversely proportional to the intersection time. Thus if intersection is underestimated by a factor of two, then porosity is overestimated by the same factor.
FigS. 1 and 2 show the loglog type curve and semilog plots for the three falloff tests (Table 1) using a mobility ratio of 5.5 and fracture porosity of 0.1%. All other data is listed in Table 2. The intersection times of the first and last straight lines are 0.173, 0.739, and 3.20 hours respectively. There is no true late time straight line portion for the final falloff as shown by the derivative curve in Fig. 1; however, by knowing the mobility of the water bank, a reasonable late time straight line and intersection time can be found. The intersection time values result in bank radii of 88, 181, and 377 ft and fracture porosities of 0.087, 0.102, and .099% as given by Eqs. 4 and 5. These values of fracture porosity are within 13% of the actual value of 0.1 %. Table 3 summarizes all the simulated tests and the calculated fracture porosities. The intersection times on this table were determined by performing a least squares fit on the two semilog straight line portions of each test and then intersection time is given as
.1tx = 10
.1P1 hr, i 
.1Pl hr, 0
(
ffioffii
)
•
(6)
If the derivative curve showed no straight line segment, then no intersection time is given in Table 3. Although not shown here, the largest errors occur when one of the semilog straight lines is of very short duration. Also, it appears that errors are generally greater for the smaller mobility ratio. This is because an error in slope leads to a larger error in intersection time as the mobility ratio becomes smaller.
Yeh and Agarwal ^{3} also proposed calculating fluid mobility versus radial distance from the following equations
At. = 162.6 q B
.1p'h
,
_{(}_{7}_{)}
(8)
(9)
where Ata is a volumetric average mobility from rw to radius r, and At would represent the true mobility at the same radius. For the assumptions used in this paper At and Ata should be the same to the bank radius. This was not the case in reference 3 where relative permeability was varying with r.
Fig. 3 shows the volumetric average mobility
versus radius (Eqs. 7 and 8) calculated for the simulated falloffs tests for a fracture porosity of 0.1% and mobility
ratio of 10. It can be seen that the deviation from the
low
mobility inner bank region approximately corresponds the the bank radius calculated from the intersection time method (Table 3) This type of plot gives an alternative
method to estimate bank radius and thus fracture porosity from Eq. 5.
FIELD EXAMPLES
Recently, Marathon performed several polymer gel injection tests in two naturally fractured reservoirs in order to reduce water flow from the fractures and increase oil flow from the matrix. For the seven wells analyzed using the techniques outlined in this report, two tests gave very unreasonable results (Le. fracture porosity was greater than the previously estimated total system porosity). One well, Well #7, showed composite system behavior at a large radius prior to gel injection. Thus, analysis of the composite behavior after gel injection was complicated by the pregel composite behavior. A second well, Well #6, had no pregel injection transient test (all the other wells did). Composite system analysis of this well also resulted in an estimate of fracture porosity which was larger than estimated total system porosity. This well possibly also had a composite behavior prior to gel injection which would affect the final analysis.
All the wells analyzed are listed in Table 4 while Figures 4 through 10 show the pre and postgel falloff test data. The permeabilities given on the figures are actually mobilities (kill). Derivative curves were also developed for the field data, but are not included for brevity. In most cases, a second straight line did not adequately develop during the postgel falloff; therefore, bank radius was estimated from the following equation rather than Eq. 410:
lb =0.0257 Y'Ai .1tx l /
(10)
431
Using Polymer Injectlvlty Tests to Estimate
4 Fracture Porosity In Naturally Fractured Reservoirs
SPE025880
where At _{x} 1 is the time at which the data begins to deviate from the straight line which represents the inner gel bank. This analysis assumes that the deviation is caused by transistion from the inner to the outer bank and not because of boundaries or nearby well interference. If there is no deviation from a straight line by the end of the test, then the final time from the test is used in Eq. 10. By combining Eqs. 5 and 10, a very simple relation for fracture porosity results:
(11)
where Vinj is the volume of polymer gel injection (prior to the postgel falloff test). Table _{4} shows the data used and results of the calculations. Also shown is the time (At _{x} ) at which early and late straight lines would have to intersect to give a porosity which agrees with that calculated from
At x1·
The slopes of the last time straight line (mo) given in Table 4 are not actual values from the tranisent tests. They were estimated based on the pregel injection permeabilities and the postgel injection rates. An accurate estimate of total system porositycompressibility is very important in the analysis of bank radius. The data for these examples were not based on interference testing and thus could be in significant error. Based on other considerations (eg. cores, logs, tracer tests, and simulations), the expected fracture porosity shoud be less than 1%. Only two wells (Well #3 and Well #4), gave fracture porosities in this range. One reason for the high calculated porosities could be because of a high estimate of the porosity compressibility product. The estimated porosity is directly proportional to compressibility (Eq. 11). Thus, if actual compressibility is lower by a factor of two, the estimated porosity will be reduced by the same factor.
SUMMARY AND CONCLUSIONS
The main conclusion of this study is that a new method is presented to calculate fracture porosity in naturally fractured reservoirs using composite system analysis. This technique appears to be well suited to polymer injection in water injection wells. A prior estimate of fracture porosity is important to properly design the test. Total system compressibility must be known and mobilities of the insitu and injected fluid should also be known. The injected fluid must remain in the fracture network and cannot enter the matrix rock. If this is not the case, then fracture porosity will be overestimated.
TABLE 1. SIMULATED INJECTIVITV TEST SEQUENCE
Time perlod,days
_{W}_{e}_{l}_{l} _{S}_{t}_{a}_{t}_{u}_{s}
00.25 
Inject 
0.251.25 
Falloff 
1.252.25 
Inject 
2.256.25 
Falloff 
6.2510.25 
Inject 
10.2526.25 
Falloff 
TABLE 2. DATA FOR EXAMPLES
_{q} 
750 STB/D 

_{B} 
1.00 RB/STB 

h 
50 ft 

_{J}_{l}_{w} 
0.82cp 

_{P}_{i} 
100 psia 

_{+}_{m} 
0.15 

^{.}^{,} 
0.05, 0.1, and 0.5% 

_{k}_{f} 
1000 md 

k _{m} 
1 rnd 

(J 
0.75 1/sq. ft 

_{r}_{w} 
0.30 ft 
5000 ft 

_{r}_{e} 

_{C}_{w} 
3.1 e 6 psi 1 

_{c}_{m} 
1.0e 5 
psi 1 

_{C}_{f} 
2.0e 
5 
psi 1 
_{M}_{o}_{b}_{i}_{l}_{i}_{t}_{y} _{r}_{a}_{t}_{i}_{o}
2.45, 5.5, and 10.0
TABLE 3. RESULTS OF COMPOSITE SYSTEM ANALYSIS FOR SIMULATED FALLOFF TESTS
_{I}_{n}_{p}_{u}_{t} _{D}_{a}_{t}_{a}
_{+}_{t}
0.0005
o.ooos
o.oOOS
0.0005
_{0}_{.}_{0}_{0}_{0}_{5}
0.0005
ADlA.i
_{2}_{.}_{4}_{5}
2.45
2.45
_{5}_{.}_{5}
_{5}_{.}_{5}
_{5}_{.}_{5}
_{r}
_{1}_{1}_{6}
259
530
116
_{2}_{5}_{9}
530
0.0005
0.0005
0.0005
0.001
0.001
0.001
0.001
_{0}_{.}_{0}_{0}_{1}
0.001
0.001
0.001
0.001
0.005
0.005
_{0}_{.}_{0}_{0}_{5}
0.005
0.005
_{0}_{.}_{0}_{0}_{5}
0.005
_{0}_{.}_{0}_{0}_{5}
_{0}_{.}_{0}_{0}_{5}
_{1}_{0}
_{1}_{0}
_{1}_{0}
2.45
_{2}_{.}_{4}_{5}
2.45
_{5}_{.}_{5}
_{5}_{.}_{5}
5.5
10
10
10
_{2}_{.}_{4}_{5}
_{2}_{.}_{4}_{5}
_{2}_{.}_{4}_{5}
_{5}_{.}_{5}
_{5}_{.}_{5}
_{5}_{.}_{5}
10
10
_{1}_{0}
_{1}_{1}_{6}
_{2}_{5}_{9}
_{5}_{3}_{0}
_{8}_{2}
_{1}_{8}_{3}
375
_{8}_{2}
_{1}_{8}_{3}
375
_{8}_{2}
_{1}_{8}_{3}
375
_{3}_{7}
_{8}_{2}
_{1}_{6}_{8}
_{3}_{7}
_{8}_{2}
_{1}_{6}_{8}
37
_{8}_{2}
_{1}_{6}_{8}
0.485
2.42
10.1
0.0857
0.426
1.791
0.151
_{0}_{.}_{7}_{5}_{4}
3.17
0.243
1.21
5.09
0.0182
0.0895
_{0}_{.}_{3}_{7}_{6}
0.0322
0.158
_{0}_{.}_{6}_{6}_{4}
0.0518
_{0}_{.}_{2}_{5}_{4}
_{1}_{.}_{0}_{7}
0.304
1.409
0.473
2.194
0.123
0.433
1.768
0.173
0.739
0.261
un
0.130
0.398
0.188
0.684
0.254
_{1}_{.}_{1}_{2}_{7}
calculated Results
115.9
247.4
0.000413
0.000601
0.000499
0.000548
114.0
242.3
101.2
183.2
3n.6
88.5
178.4
85.3
178.0
0.000515
0.000571
0.000654
0.000998
0.000987
0.000855
0.001053
0.000922
0.001057
100.8
173.0
90.2
169.7
82.2
172.56
0.003300
0.004704
0.004121
0.004886
0.004956
0.004727
% error
17.42%
20.14%
0.14%
9.53%
3.06%
14.18%
34.61%
0.21%
1.29"10
14.49%
5.33%
7.80%
5.73"10
33.99%
5.920/0
17.59%
2.27"10
0.88%
5.46%
intersection time could not be calculated because one straight line segment did not appear on pressure transient.
432
SPE025880
J. R. Gllman~ S. B. Hinchman and M. A. Svaldl
_{5}
TABLE 4. COMPOSITE SYSTEM ANALYSIS FOR FIELD FALLOFF TESTS
WEll:
Well
#1
Well
#2
Well
#3
Well
#4
Well
#5
Well
_{#}_{6}
Well
_{#}_{7}
_..._..._....._
..__....._
mj
351.1
784.2
74.7
7.8
178.8
59.5
_{2}_{8}_{7}_{.}_{2}
m _{o}
_{A}_{o}
_{'}_{C}_{d}
r _{w}
_{A}_{t} _{x} _{1}
_{r}_{b}
1032
6.6
22.5
228.2
5.8
19.3
8.7
38.8
332.9
1.7
17.4
4.1
444.1
36.3
109.1
2033.3
371.8
_{1}_{5}_{7}_{7}_{.}_{0}
_{3}_{8}_{.}_{4}
_{2}_{.}_{4}
_{1}_{6}_{.}_{7}
7.5E06 7.2EoEI
1.3E·04
0.326
0.229
0.292
15
20
93.49
101.08
120
153.41
1.5E04
1.7E04
1.7E04
_{1}_{.}_{l}_{E}_{}_{0}_{4}
0.385
0.354
0.354
0.292
110
_{9}_{0}
_{4}_{6}_{3}_{.}_{7}_{7}
_{1}_{1}_{2}_{.}_{6}_{6}
_{6}
_{3}_{0}
_{5}_{0}_{.}_{4}_{3}
_{2}_{0}_{.}_{7}_{9}
_{V}_{i}_{n}_{j}
h
At _{x}
4390
6690
35
38
0.0256
0.0308
28
37
15000
70
0.0163
180
15000
15000
_{1}_{5}_{0}_{0}_{0}
_{8}_{5}_{0}_{0}
_{1}_{1}_{0}
_{4}_{0}
_{4}_{0}
_{1}_{0}_{0}
0.0011
0.0528
0.2636
0.3515
190
131
8.3
47
NOMENCLATURE
B
c
h
k
L
m
=
=
=
=
=
p
q
r
t
tinj
Vinj
dP
dP1hr
=
=
=
=
_{=}
_{=}
_{=}
_{=}
dP'
dt
dte
dtX
dtx1
A
=
=
=
formation volume factor, RB/STB
compressibility, psi ^{1} reservoir thickness, ft permeability, md
matrix block size, ft
slope of semilog pressure transient plot, psi/cycle pressure, psi flow rate, STB/Day radius, ft time, hours injection time, hours volume injected, STB pressure difference, psi pressure difference at 1 hour, psi
d
t·
enva Ive 0
.
f
pressure difference,
.
d Ap
, psi
d2nAt
time since start of falloff, hours equivalent falloff time, hours time at intersection point, hours time for deviation from first straight line, days mobility, kill, md/cp
A'
_{I}_{l}
_{c}_{r}
_{c}_{j}_{l}
_{=}
_{=}
_{=}
derivative of mobility, dA, mdlcpft
dr
viscosity, cp
matrix shape factor, ft ^{2} porosity, fraction
Subscripts
1hr
_{a}
_{b}
_{e}
f
_{=}
inj
_{=}
_{0} _{=}
_{t}
w
one hour average bank equivalent, external fracture
inner bank, initial injection
matrix
outer bank total wellbore, water
intersection
ACKNOWLEDGEMENTS
We thank Marathon Oil Company for permission to prepare this paper. We would also like to thank J. M. Paneitz and M. J. Stover for providing the data and analysis of the field examples.
REFERENCES
1.
Satman, A. : "Solution of Heat and FluidFlow Problems in Naturally Fractured Reservoirs, Part 2:
FluidFlow Problems", SPE Prod. Eng. 467473, (Nov.
1988).
2. Agarwal, R. G.: "A New Method to Account for Producing Time Effects When Drawdown Type Curves Are Used to Analyze Pressure Buildup and Other Test Data," SPE 9289 presented at SPE Annual Technical Conference, Dallas, TX, (Sept. 2124, 1980).
3. Yeh, NS., Agarwal, R. G.: "Pressure Transient Analysis of Injection Wells in Reservoirs With Multiple Fluid Banks," SPE 19775 presented at the 64th SPE Annual Technical Conference, San Antonio, TX (OCt. 811, 1989).
4. Brown, L. P. :"Pressure Transient Behavior of the Composite Reservoir," SPE 14316 presented at the 60th SPE Annual Technical Conference, Las Vegas, NV (Sept. 2225, 1985).
5. Hazebrook, P., Rainbow, H. and Matthews, C. S.: "Pressure FallOff in Water Injection Wells," Trans. AIME (1958) 213, 250260.
433
Using Polymer Injectlvlty Tests to Estimate
6 Fracture Porosity In Naturally Fractured Reservoirs
SPE025880
6. Hurst, W.: "Interference Between Oil Fields," Trans. AIME (1960) 219, 175190.
7. Carter, R. D.: "Pressure Behavior of a Limited Circular Composite Reservoir," Soc. Pet. Eng. J., (Dec. 1966),
328334. 

Bixel, 
C. 
and 
Van 
Poolen, H. K.: "Pressure 
Drawdown and Buildup in the Presence of Radial Discontinuities," Soc. Pet. Eng. J., (Sept. 1967), 301
309.
9. Ramey, H. J., Jr.: "Approximate Solutions for Unsteady State Liquid Flow in Composite Reservoirs," _{J}_{.} _{C}_{d}_{n}_{.} _{P}_{e}_{t}_{.} _{T}_{e}_{c}_{h}_{.} (Jan.March 1970),3237.
10. Merrill, L. S., Jr., Kazemi, H. and Gogarty W. B.:
"Pressure Falloff Analysis in Reservoir with Fluid Banks," J. Pet. Tech. (July 1974), 809818.
11. Sosa, A., Raghavan, R. and Limon, T. J.: "Effect of Relative Permeability and Mobility Ratio on Pressure FallOff Tests," J. Pet. Tech. (June 1981),11251135.
12. Raghavan, R.: "Well Test Analysis for Multiphase Flow," SPE 14098, presented at the International Meeting on Petroleum Engineering, ,Beijing, China, (March 1720, 1986).
13. Abbaszadeh, M. and Kamal, M. M.: "Pressure Transient Testing of Water Injection Wells," _{S}_{P}_{E} Reservoir Engineering, (February 1989), 115124.
14. Rowan, G. and Clegg, M. W.: "An Approximate Method for Transient Radial Flow, " Soc. Pet. Eng. J., (Sept., 1962) 225.
15. Satman, A., Eggenschwiler, M. and Ramey, H.J., Jr.: "Interpretation of Injection Well Pressure Transient Data in Thermal Oil Recovery," SPE 8908 presented at the 50th Annual SPE California Regional Meeting, Los Angeles, CA (Apr. 911,1980).
16. Satman, A. : "PressureTransient Analysis of a Composite Naturally Fractured Reservoir", SPE Form. Ev., (June 1991) 169175.
17. Kikani, J. and Walkup, G. W. Jr.: "Analysis of PressureTransient Tests for Composite Naturally Fractured Reservoirs", SPE Form. Ev., (June 1991)
176182.
434
Figure 4.
Well #1 Pre and PostGel Falloff Tests
L .
.
_
435
436
437