Factors Affecting Waterflood Pattern

Performance and Selection

PAUL B. CRAWFORD
TEXAS PETROLEUM RESEARCH COMMITTEE
A&M COLLEGE Of TEXAS

Downloaded from http://onepetro.org/JPT/article-pdf/12/12/11/2237988/spe-1503-g.pdf by guest on 21 October 2024

MEMBER AIME
COLLEGE STATION, TEX.

Abstract In making these mathematical calculations, it was assumed

that the reservoir was uniform and homogeneous and the
A review is given of the results of early work showing
mobility ratio was one, steady-state conditions existed and
sweep efficiency and steady-state injection rates for various
gravitational and capillary effects could be neglected.
types of patterns which may be used for waterflooding
programs. This early work indicated that the sweep ef- In reviewing these assumptions, it would appear that
ficiency may be near 72 per cent for the five-spot pattern, these theoretical results may have limited application to
56 per cent for the direct, line-drive square pattern and water flooding in depleted reservoirs because it is well
//lay range from 45 to 90 per cent for the nine-spot pat- known that the mobility rate is not unity, that water
tern. The results were calculated on the basis that the 1110- flooding results in the creation of a flooded zone and an oil
bility ratio was one, the reservoir was uniform and homo- bank, oil is displaced in a transient potential field and
geneous, steady-state conditions existed, and gravity and gravitation and heterogeneities are important. In a discus-
capillary effects could be neglected. In actual flooding sion of flooding, Muskat points out that "while the steady-
operations it is found that all reservoirs are heterogeneous, state theory gives an instructive picture of the tendencies
steady-state conditions do not exist, the mobility ratio is for differential fluid movement over the different parts of a
not one, and gravity and capillary phenomena undoubtedly fluid-injection system and does have applicability to cycling
playa role. operations in condensate-producing reservoirs, it will hard-
ly represent even a semiquantitative treatment of the
Theoretical calculations show that rock heterogeneities motion of fluid injected for secondary recovery".'
can cause the sweep efficiency of both the five-spot and
direct, line-drive square patterns to vary from near zero
to 100 per cent. Actual laboratory data have been obtained Calculated Pattern Performance
to show that the oil recovery by water flooding in depleted Injectivity
reservoirs results in the production of substantially the
Fig. 1 shows the relative injection rates (i.e., the pattern
same quantity of oil for the five-spot, direct, line-drive
conductivity or pattern injectivity) for various well spacings
square pattern and nine-spot pattern for a wide range of
for three types of flooding patterns.'" The patterns include
initial fluid saturations. The experimentally observed sweep the direct line-drive square pattern, the five-spot pattern
efficiency of the five-spot pattern is shown to depend on and the seven-spot network. Note that, for well spacings
the initial connate-water saturation and oil saturation. near 330 ft, the injectivities are of the order of 5.0, 5.4
and 7.0 for the case of the line-drive, five-spot and seven-
Introduction spot networks, respectively. Increasing the well spacing to
660 ft decreases the injectivity by about 15 to 20 per cent,
It is believed quite likely that many of the decisions
as may be seen on the figure. It is desired to point out
affecting the selection of the pattern to be used in water
that the "y" axis represents a constant reflecting a com-
flooding are based largely on information and conclusions
bination of rock and fluid properties times the ratio
which were made many years ago. The purpose of this
QI6..P, which represents the barrels input per unit pres-
paper is to review this early work and include recent work
sure drop. From this study, it is apparent that the five-spot
showing various factors which affect the performance of
pattern is superior to the line-drive pattern, but the in-
pattern-type floods. In the mid 1930's Muskat4 and his
jectivity of the seven-spot is far superior to both.
associates presented the first analytical solutions showing
the sweep efficiency and injectivity for the five-spot, direct, Fig. 2 shows the injectivity vs well spacing ratios.' The
line-drive, staggered, seven-spot and nine-spot patterns. three curves shown are for the case of the direct line-drive
with a fixed at 660 ft, the direct line-drive with d fixed
Original manuscript received in Society of Petroleum Engineers
office May 10, 1960. Revised manuscript received Oct. 28, 1960. Paper at 660 ft and the staggered line-drive with a fixed at 660
presented at Fourth Biennial Secondary Recovery Symposium May 2'-3 ft. Note that, for the case in which d/ a is equal to 1.0,
1960, in Wichita Falls, Tex. ' ,
4References given at end of paper.
the conductivity of the patterns are substantially the same;
SPE 1503-G
DECEMBER, 1960 11
however, for values of dl a near two, it will be seen that adjacent input wells or the distance between adjacent out-
the direct line-drive with d fixed at 66D ft has greater in- put wells. The ordinate shows the sweep efficiency, or the
jectivity than either the direct drive with a fixed at 66D ft area swept at water breakthrough divided by the total
or the staggered drive with a fixed at 660 ft. The con- area of a unit pattern. Notice that, for the case in which
ductivity of the five-spot pattern is not shown but cor- dl a = 0.6, the sweep efficiency or area swept at water
responds to the injectivity of the staggered line-drive for breakthrough may be either 37 or 73 per cent depending
the special case when dja = 0.5. on whether a direct line-drive or staggered line-drive pat-
If a comparison of the injectivity had been made for the tern is operated. For the direct line-drive or square pat-
case in which the well density were the same, it would tern, dja is equal to one and has a sweep efficiency near
be found that the five-spot flood is superior to either the 56 per cent, compared to a sweep efficiency of near 77
line-drive or the seven-spot flood. This is true for any well per cent for the staggered line-drive pattern. For very
spacing; however, the differences are not large for very large values of dja, the sweep efficiencies may be as high
small well spacings. They do become fairly significant at as 90 per cent.
the larger well spacings in that the five-spot may be about
Five-Spot Pattern
10 per cent more conductive than the line-drive or seven-
spot. The five-spot pattern has a sweep efficiency of about
72 per cent. The sweep efficiency is substantially indepen-
Sweep Efficiencies dent of the well spacing or well diameter for most of the
Line-Drive and Staggered Patterns present waterflooding applications.
Fig. 3 shows the sweep efficiency of the direct and Nine-Spot Pattern

Downloaded from http://onepetro.org/JPT/article-pdf/12/12/11/2237988/spe-1503-g.pdf by guest on 21 October 2024

staggered line-drive patterns for various well spacings" The In 1948, Muskat reported on the sweep efficiency of
abscissa shows the ratio of the distance d to a. The dis- nine-spot flooding networks.' Muskat pointed out that the
stance d represents the distance between the line of injec- sweep efficiency depended largely on the ratio of flow at
tion weIls and the line of producing wells. For the direct two wells (shown as q3 and q2 in Fig. 4). The minimum
line-drive pattern, it is equal to the distance between the sweep efficiency was 50 per cent. If the rate of production
injection and producing wells. However, for the staggered Well No.3. When the flow ratio q,! q2 was about 3.0, the
line-drive pattern, d becomes equal to the distance between sweep efficiency was 50 per cent. If the rate of production
the line or rows of injection wells and the rows of pro- at Well No.3 was about 10.6 times that at Well No.2,
ducing wells. The value a represents the distance between
100,-----------------------------------------------,
l0r----------------------------------------------,

90
_--_--=======1
Seven-Spot Network
.-; Staggered lme Dnve __ ....
Ii
,-
80

70
Uniform and hom~us media
/ Direct lme Dnve
0' dlslonc~ between odJoct'nl
Input wells

------ ---------J
60
d' distance between the lme
of mjectlon wells and the
Ime of producing wells

Five - Spol Network '0

40
line Dnve - Square Pollern

Well SpOCin9 Ratio - d /0

Fig. I-Effect of well spacing on injectivity.' Fi,g. 3-Effect of well spacing ratio on sweep efficiency.'

5.6 120.--- -----------------------------------------,

5.2
" "' __ Staggered Dnve with 100
"" Q Filed at 660Ft

"
" ~ 80
r
"
15
Direct Onve with <3
.0 2 .0 3
d Fi)(ed at 660 Ft
~ 60

~-Q---1 f
r . "-
w
w
;0
40 ®q! .0 2
d "'
36 L DIRECT
~
,;
20 .0'
Direct Dnvewlth/
Q Fixed at 660 Ft. NINE SPOT
3.2

Fig. 4-Effect of producing ratios on sweep efficienc~·

Fig. 2-Effect of well spacing ratio on injectivity.s of nine-spot pattern!

12 .JOURN,\L OF PETROLEIJM TECHN()LO(;Y

the sweep efficiency of the nine-spot reached a maximum 100~-------------------------------,
of 78 per cent. When this producing ratio exceeded 30,
it was equivalent to shutting-in Well No.2, and the sweep
efficiency approached that of the five-spot, or 72 per cent.
90
Field Calculations
Tables 1 and 2 show comparisons of field results for
various modes of operation. The tables were originally
presented by Muskat and are based on the theoretical ~ 80
operating conditions. I
~
o
Z
Balancing Injection Rates ILl
(3
In water flooding five-spot patterns, it seems important iL: 70
~
to balance the injection rate so that premature break- ILl
through does not occur from an injection well long before a..
ILl
water reaches the well from an adjacent input well. Fig. 5 ILl
shows the effect of unequal injection rates on the sweep ~ 60
efficiency of the five-spot pattern. Note that, as the injec-
tion ratio q,/ q, varies from zero to 1.0, the sweep efficiency
varies from 45 to 72 per cent.' The low value of 45 per

Downloaded from http://onepetro.org/JPT/article-pdf/12/12/11/2237988/spe-1503-g.pdf by guest on 21 October 2024

cent results as the stream lines cross the normal boundaries
when q, is reduced.
Effect oj Mobility Ratio 40L-____ ____
~ _____L_ _ _ _- L____
~ ~

Dyes, Caudle and Erickson presented the results of o 0.20 0.40 0.60 0.80 1.00
an extensive study showing the effect of mobility ratio on INJECTION RATIO -ql/q2
the swept area and fractional water cut for the five-spot,
Fig. 5-Effect of unequal injection rates on sweep
direct, line-drive square pattern and the staggered line- efficiency of five-spot pattern.
drive patterns: In performing the work there was no gas
saturation and there was 100 per cent liquid fill-up when
the water was first injected. me ability in the east-west direction differs from that in
the north-south direction, the sweep efficiency may vary
Fig. 6 shows a comparison of the swept area for the from near zero to 72 per cent for the five-spot pattern.
five-spot and direct, line-drive square patterns: The abscis- This is shown in Fig. 7. The sweep efficiencies are restric-
sa shows the mobility of the fluids in the unswept region ted to the orientations shown, as a 45° rotation could
divided by the mobility in the waterflooded region. At result in a near-zero sweep for the line-drive and 100
breakthrough, the five-spot is definitely superior to the line- per cent sweep for the five-spot pattern.
drive. After injecting 2.0 pore volumes of water, the swept
areas were substantially the same for mobility ratios of
0.4 and greater.* Actual WaterOood Results

Heterogeneous Reservoirs Five-Spot and Direct Line-Drive Patterns

Reservoir heterogeneities can have an appreciable effect A laboratory study has been made to compare the oil
on the sweep efficiency of flooding patterns. If the per- recoveries of the five-spot, direct, line-drive square pattern
and the nine-spot pattern. The model was 10 X 10 X Y2-
in. thick. An unconsolidated sand was used. The initial
*Since this work was presented, AIME symbols and nomenclature connate water, oil saturation and oil viscosity were varied.
have inverted th" definition of mobility ratio as first used by Dyes.
Caudle and Erickson. Fig. 8 shows a comparison of the oil recovered by water
flooding the five-spot and line-drive square pattern. The
TABLE l-THE RELATIVE EFFICIENCIES OF VARIOUS FLOODING NETWORKS initial water saturation was 25 per cent. The oil-water vis-
Relative
Production
cosity ratio was 1.6. For the figure presented, the initial
Rate (Con. oil-in-place was 35, 50 and 65 per cent of the pore space.
ductivity
For Equal Flood Note that after injection of 2.0 pore volumes of water the
Well/Acre Efficiency
Type of Network Densities) (per ce".!!.
oil recoveries were about the same, independent of pat-
Staggered Line-drive tern, when the initial oil saturation was 50 per cent or less.
d/o = 1.50, a = 539 It, d = 808 ft 0.383 80.0 For the higher oil saturation, fill-up was obtained more
Five-spot Array, ds = 660 ft 0.433 72.3
Regular line-drive, dlo = 1.50; quickly and the five-spot was definitely superior for the
a = 539 ft; d= 808 ft 0.383 70.6
conditions studied.
Seven-spot Array, d7 = 709 ft 0.378 74.0

TABLE 2-COMPARISON OF VARIOUS FLOODING NETWORKS FOR A SPECIFIC FLOODING PROGRAMS

Total area of tracts, 1000 acreSi total number of wells, 400; total pressure differential, 1,000 fb; sand thickness, 40 fti sand permeability, 0.01 darey; sand porosity,
, 10 per cent; effective oil-water viscosity, 10 cpo

Rate
of Time
Fluid to Total
Input- Input- Input Reach Water-free
No. No. Input Output or Output Flood Oil
Input Output Spacing Spacing Output Wells Efficiency Prod.
Wells Wells (ltl _(_It_I__ (B/DI (yrsl (per centl (10· bbll
Staggered line-drive, d/o
Direct Line-drive, d =0
1.50 200
200
200
200
270
330
426
330
3776
4094
~.

11.8
ao:o-
57.0
~
1.77
five-spot 200 200 467 330 4305 14.3 72.3 2.25
Seven-spot 267 133 355 355 3765 16.7 74.0 2.30

DECEMBER, 1960 13
N'ine-Spot-Five-Spot Patterns the oil saturation was slightly higher initally), the nine-
A two-stage modified nine-spot injection program was spot performance was slightly superior to the performance
studied by Smith and Nobles using an electrical network: of the five-spot. However, the differences at water-oil
In their two-stage technique, the streamlines and isopoten- ratios greater than 50:1 were found to be very small.
tials were constant for a single stage; but, ~n the second Similar-type studies were made with other operating
stage of the injection program, a separate field of stream- schedules for the nine-spot pattern and for various oil-
lines and isopotentials existed. By use of a two-stage nine- water viscosity ratios; but, in general, it was found that
spot displacement method, Smith and Nobles found that there was very little effect of scheduling technique on the
a sweep efficiency of 90 per cent could be obtained. This oil recovery (i.e., the five-spot performance was very
high sweep efficiency resulted when in the initial stage the similar to that of the nine-spot, practically independent of
center well was an injection well and in the final stage the method in which the nine-spot was operated). This was
a side well was the sole injection well. not true in the early oil recoveries where the nine-spot
From a review of the literature, it is apparent that the pattern was superior to the five-spot, but for high water-
sweep efficiency of the nine-spot pattern may range from oil ratios there appeared to be very little difference in the
near 45 per cent to 90 per cent depending on the produc- recovery.
ing-injection scheduling techniques. In view of the possible
limited application of steady-state results to depleted water Sweep Efficiencies
floods, it appeared desirable that actual laboratory flood- In performing the laboratory waterflooding tests, it was
ing programs be conducted in porous media to evaluate found that the sweep efficiency or swept area when water
several nine-spot injection-producing schedules to supple- broke through depended on the initial oil, water and gas

Downloaded from http://onepetro.org/JPT/article-pdf/12/12/11/2237988/spe-1503-g.pdf by guest on 21 October 2024

ment the theoretical work and provide additional infor- saturation: If the initial oil saturation remained constant
mation on the relative efficiency of the five-spot compared and only the quantity of connate water and free-gas satu-
with the nine-spot pattern in a transient displacement pro- ration varied, it was found that the sweep efficiency in-
gram. creased as the connate water decreased.
A study was made to simulate and compare the flooding If the connate water was maintained constant and the
efficiencies in the five-spot and the nine-spot pattern.! In initial oil and free-gas saturations varied, it was found
one particular method of operation, the center well was the that the sweep efficiency increased as the initial oil satu-
injection well and Well No.2 was produced until water ration decreased. These results appear proper inasmuch
breakthrough; then, the well was shut in and the No. 2 as radial flow is maintained for a greater period of time for
well used was a sole injection well. Results of this par- lower initial liquid saturations; and, during radial displace-
ticular case are shown in Fig. 9. This particular flood ment, the sweep efficiency is independent of the flooding
was conducted for the case in which the oil-water viscosity well pattern.
ratio was 1.6, the initial connate-water saturation ranged
from 8 to 10 per cent and the initial oil saturation ranged Discussion
from 35 to 63 per cent, as shown on the figure. The cumu-
lative oil produced in terms of per cent pore volume is At the present time very little information has been
shown as the ordinate. presented in the literature on the quantitative effects of
The abscissa shows the water injection in terms of per reservoir heterogeneities, gravitational phenomena, capillary
cent pore volume of the pattern. The solid lines on the
curve show the performance for the nine-spot curve pat- 80
tern. The dashed lines on the curve show the performance
or oil recovery for the five-spot pattern. In reviewing this
figure, it will be noted that, for initial oil saturations rang- 70
ing from 35 to near 50 per cent, there was very little
difference in the cumulative oil produced whether the nine- >-
u
spot or the five-spot pattern was used; however, for the z 60
w
case in which there was greater fill-up initially (i.e., where ~
lJ..
lJ..
W 50
1oor-------------~----------------~----------_,
Il.
2.0 Pore Volumes W
Water ~ /. W
90 J/\/ ~ 40

.. 80 _ _ Direct Une Drive

LlNEnRIVE
r-=-(J'-'---,
."' ___ Five Spot
I I
II:

."'
.. 70
I-
II <kNEII
I
~ I k I
II: 60
L--o-_w
"'
~
50

40~~------~----~~--------~----~~------_;40
o 3.5
0.1 0.4
MOBILITY IN UNSWEPT

Fig. 6-Comparison of swept areas in direct-line-drive Fig. 7-Effect of directional perllleability on

and five-spot patterns: sweep efficiency.

14 JOURNAL OF PETROLEUM TECHNOLOGY

 OOI.---------------------------------------~ w GOr----------------------~
~ J.1 1il- "I.6 - - - - Performonce of Five -Spot Schedule -53
0 w
1;) <5 __________ 5 " 63.".
~ > 50 - 0
::> w ; _~==4==:=:===-c'----sw" 9.65%
...Jso ~
~ I-
UJ ~ 40
a:
oQ.
!!40
I /
/'
",,-- -- _ - - - - - - - - - ---65

65 ~
8 30
o
UJ / "o;5
<> I
1530 / '"-'a. 20
______~" 35.0~

o
a: / o Sw:"IO.O't
Q. w
/ >
...J
6
I
I
~
S 10
... 20
2: I Oil Saturatioy "
::>

~ I (% P.V.l "
:5 ! ? - - Direct Line Drive
3S0

~IO
<> - - - Five Spot
Fig. 9-Comparison of nine-spot and five-spot
performances.

280

Downloaded from http://onepetro.org/JPT/article-pdf/12/12/11/2237988/spe-1503-g.pdf by guest on 21 October 2024

2. Muskat M.: "The Theory of Nine-Spot Flooding Networks",
Prod. Monthly (March, 1948) 12, 14.
Fig. 8--Comparison between five-spot and
direct-line-drive square performances. 3. Smith, H. W. and Nobles, M. A.: "Two Stage Modified Nine-
Spot Flooding Procedure", Prod. Monthly (Dec., 1958) 33.
4. Muskat, M.: Physical Priniciples of Oil Production, McGraw-
effects, wettability, interfacial properties, etc., on oil re- Hill Book Co., Inc., N. Y. (1949).
covered by water flooding. In actual laboratory water 5. Dyes, A. B., Caudle, B. H. and Brickson, R. A.: "Oil Produc-
flooding in stratified sands, it has been found that the esti- tion After Breakthrough-as Influenced by Mobility Ratio",
mation of the water rate in layered systems using recom- Trans., AIME (1954) 201, 8l.
mended techniques may range from 10 to 3,000 per cent 6. Talash, A. W., Pruitt, G. T. and Crawford, Paul B.: "Labora-
of the actually observed flooding rates. It is readily ap- tory Performance of Five-Spot Floods", Report No. AM 59-17,
parent that considerable work is yet to be done before TPRC (Oct. 24, 1959) unpublished.
the detailed results to be expected in flooding actual reser- 7. Muskat, M.: Flow of Homogeneous Fluids, McGraw-Hill Book
voirs can be predicted with a reasonable degree of confi- Co., Inc., N. Y. (1937).
dence in many cases. 8. Muskat, M. and Wyckoff, R. D.: "A Theretical Anaylsis of
Water-Flooding Networks", Trans., AIME (1933) 107, 62.
References
1. Cotman, N. T., Still, G. R. and Crawford, Paul B.: "Labora·
***
EDITOR'S NOTE: A PICTURE AND BIOGRAPHICAL SKETCH
tory Comparison of Oil Recovery in Five·Spot and Nine·Spot OF PAUL B. CRAWFORD WERE PUBLISHED IN THE OCT, 1960
Waterflood Patterns", Report No. AM 59·21, TPRC (Jan. 4,
1960) unpublished. ISSUE OF JOURNAL OF PETROLEUM TECHNOLOGY.

DECEMBER, 196<1 IS