Sie sind auf Seite 1von 8

Slw- (s3

Horizontal Barriers for Controlling Water Coning


J. C, KARP*
JUNIOR MEMBER AME
D. K, LOWE
N. MARUSCW
MEMBER AIME

GULF RESEARCH & DEVE1OPMENT CO.


PITTSBURGH, PA.

THEORY AND DEFINITIONS

ABSTRACT
This paper considers, from an engineering viewpoint,
several factors involved in creating, designing and locating
horizontal barriers for corrtrolling warer coning. This is
an eflort to consolidate new concepts with previous information so that a reasonable selection of barrier materia[s,
dimensions and vertical position can be made,
Coning theories previously developed are, briefly reviewed and an eflort is made to reduce the results o~
coning-theory calculations to a point where routine calculations can be made with a desk calculator, It is e.tpected
that these simplified calculations will give usable predictions of the amounts of improvement attainable with barriers of various dimensions,
Apparatus and procedures used for testing the suitability of commercial cements are described and test results
are presented.

INTRODUCTION
Just about as soon as the phenomenon of water coning
was recognized as an oilfield problem, there were suggestions that the production of water coning could be
controlled or completely suppressed by means of horizontal barriers. It wm also suggested that natural barriers
such as shale streaks were helpful in restricting bottomwater production. The implication was that wells not
having the benefit of continuous shale streaks to suppress
bottom waters should, in some way, be supplied with an
artificial barrier.

A schematic drawing of a horizontal barrier in a reservoir is shown in Fig, 1, Two maximum stable water cones
are illustrated-one
for the wellbore diameter and the
other for a barrier. ,A maximum stable cone is one that
has attained its maximum volume at the critical pmiuction rate and is on the threshold of producing water. The
volume displaced by each water cone and the rise in
water table are indications of the amount of water-free
oil produced. Corrections for porosity, connate water and
residual oil, of course, must be applied.
The water-oil interface which defines a maximum
cone is called a free surface, It is a surface over
the pressure is constant. This free surface has the
of a limiting streamline at and below which there
movement or transport,

The shape of the water cone describes the pressure drop


in the reservoir. The effect of a completely impermeable
barrier on the cone shape is essentially the same as extending the wellbore out to the barrier radius.
The placement of a horizontal barrier requires that
a horizontal fracture be created using a single-point entry
technique. Two methods may be employed for filling this
fracture, One method consists of propping the fracture
before filling with cement. The other consists of fracturing
with fracture fluid, then immediately following with cement

WWh regard to the selection of barrier materials, there


itre many references which give the merits of liquid barriers., These liquids include surface-active agents, precipitates and emulsions.6 There is no real need to review
the merits of these materials again, This paper deals with
the use of solid barriers which could be made from various
cements. The two major restrictions placed on these barrier cements are that (1) they must be commercially
available in large quantities, and (2) they must be applicable by means of standard tools and equipment with
techniques that have been fully developed and are in
general use at the present time.

Orkhwd manuscript received in Society of Petroleum Engineers office


Aus. 7, 1961. Revised manuscript received May 4, 19S2. Paper presented
at S6th Annual Fall Meeting of SI?E, Oct. S.11, 1961, in DallRs,
Wow associated with Advanced Microelectronics

2Refesencee
given
JULY,

1962

at end of p~per.

Oo.

stable
which
shape
is no

.
OIL

ZONE
BARRIER

ORIGINAL

OIL - WATER

CONE

CONTACT-

WATER
ZONE
FIG.lHOaLZONTAL
BARRIER
Foa tiNTSOLLING
wATEaCONING.
78s

containing propping agents. The propping agents are used


to insure adequate barrier thickness.
BARRIER DESIGN
The essential elements which need to be considered for
the design of a barrier are the radius, thickness, vertical
position and permeability.
The only consideration for the vertical position of a
barrier is that it must be at a depth that will permit oil
production at a desired rate such as may be governed by
the daily allowable.
Methods for determining the optimum perforation interval are well kncwn and are in common use at present.
The barrier is placed immediately below this perforation
interval.
The barrier thickness is dependent on the width to
which fractures can be held open by propping agents
after the propping agents have been deformed or imbedded.
It has been reported that popping agents as large as % in.
in diameter or greater could be placed in fractures.
The optimum radius of is b:,rrier is a compromise
selection govern~d by the cost of placing a barrier of a
given radius and the economic advantages to be gained.
One criterion for measuring the effectiveness of a barrier
could be based on the amount of water-free oil obtained
before the barrier is by-passed and flooded. Another criterion could be the reduction in total amount of water
produced after the barrier is flooded and by-passed. The
third criterion would be the total amount of oil produced
before the well is abandoned because water-handling costs
are as great as the value of oil recovered.
To determine the optimum barrier radius, the concept
of the free surface was used. Several solutions of the freesurface problem have been presented and reviewed in the
literature. The solution of the free-surface problem
yields two equations of interest for the present discussion.
The equation
(/ =

+w(p,t
p ) (/1: D:)
.
...
_L__
,w., In

numbers at the ends of the dashed lines indicate, the


maximum water-free production rates.
The curves for barriers were made by substituting the
barrier radius (r~) for the wellbore ,radius(r. ) in Eq. 1.
The figure also lists the parameters which are assumed
in order to arrive at the theoretical wirves, The oil-zone
thickness, area, porosity, residual oil and connate water
determine the reservoir volume.
The Cumulative Oil Production vs Time plots are
very restrictive as they stand, However, a certain amount
of dimensional variation is possible on the graphs, This
means that certain parameters can be changed, with the
result that the scales on the plots are changed but the
shapes of the individual curves are not. The parameters
which aliaw for such changes are displacement efficiency,
porosity, permeability, water density, oil density and oil
viscosity, The quantities such as well spacing, depth of
penetration, barrier radius and pay thickness cannot be
varied without changin~ the shape of the curves,
A more useful curve is shown in Fig. 3, This graph
was obtained by dividing the cumulative water-free production obtained at a constant production rate for each
barrier radius by that production rate. This simple division
gives time to water breakthrough.
For the well conditions shown in Fig. 3, the wuter-free
production nearly doubles with a 5-ft radius barrier and
then requires nearly a 100-ft radius barrier to double the
water-f ree production again. Such satisfactory results arc
not available for all well conditions. For example, Fig. 4
shows that for another set of well conditions a barrier
-,,

. 80 ACRE
.25 FOR,
_ 0.5 DARCYS
Vlsc.

1.0 CPS

0.6 SP GR.
- 25 FT PENE.

20

(1)

()
<

~I,c

.,

gives the critical production rate tibove which water can


be produced.
The equation

.-.

. . ..-

In (r./t-)
(2)!,
z = /f -h: (w f)).
in (r,./r,. )
defines the shape of the maximum stable water cone in
a radial system.
Eq. 1 may be used to prepare the basic family of curves
shown in Fig. 2, These curves give the water-free cumulative production as a function of years to first water for
various barrier radii,
For the development of this graph, it is assumed that
the reservoir pressure is always sufficient to produce the
well at any required rate, It is further assumed that gas
in solution in the reservoir remains in solution and that
crude oil produced from the reservoir is replaced by bottom water. The effects of capillary pressure and relative
permeability are neglected.
The water table rises as oil is removed, and the production rate must be reduced to keep the water out of
the well. This is the behavior pictured in Fig. 2. The
. . .. . . . . ... . ... . ..
..

.- ... .. .

~Derivation of Eq, 2 1s given in

78.*

the Appendix,

15

t)

TIME
FIG. 2(;I:MLLATIW

20

25

30

IN YEARS

WATI;R-FRRE OIL AS A

Fi xcrtox OFh \lE.

I
~

5 r-------r---------

i~.q
yj

La
2

--~--l

---

250 BOPD

m3

80 ACRE
100 FT OIL ZONE
25 FT PENE>
0.5 DARCYS

~z

22.
WA
-J
3
=0
3
u

I
Vlsc,
1.0 CPS .,
50% EFF.
:
0.6 SP (33.
:
o

10
YEARS

FIG.

20
TO

40

30
FIRST

50

WATER

3-PIXUODOF WATER-FREE
OIL PRODUCTION
PORVARIOIX
BARRIER RAMI.
JOURNALOF PETROLELIMTSWHXOLOGY

with a radius of nearly 50 ft is needed before water-free


oil can be obtained.
Another important consideration for evaluating the
effectiveness of barriers is the amount of water production
after the barrier is flooded as compared with the water
produced without a barrier.
The previous water-coning solutions were extended to
include the condition where water under the oil phase is
also moving. This is a first approximation to the solution
of the two-phase radial-fiow problem and was obtained
from first-order partial differential equations and the freesurface concept. * In addition to the assumptions used for
the maximum stable water-cone solution, it is assumed
that in the water phase underlying the oil there exists a
limiting streamline which extends from the bottom of the
water zone at the drainage radius to the bottom of the
well perforations (see Fig. 5). The water below this limiting streamline is stationary. It is also assumed that the
height to which the interface rises along the wellbore is
dependent on the rate, density and viscosities of the fluids
produced and that Darcys radial formula can be applied
to tbe wellbore.
The solution makes use of dimensionless parameters
and is in the following form.
F ~x%, il D u z!., z,.,,
ln~ .
r,)
= 2~hkgAp
v,
q11 %777
.,.
, .,.
. . . . . . . . (3)

) q+

-----*A condensed derivation of this solution is shown in the Appendix.


,,omprehensive derivation 1s available on request (Ref. 10).

100

20 BOPD

10 ACRE
1
r%;;
?kN~oNE
0.5 DARCYS

[00

50

.25 POR.
Vlsc.
Lo CPS
50~o
EFE
0.8 W? GR.

~
I
/=j
.~

The equation is much too complicated for analytical


purposes; therefore, tables for numerical values of F for
the various dimensionless parameters were compiled. From
q,, for
l-h
such tables it is possible to obtain the ratio X ~
P.
various rises in water table.
These values then can be plotted as shown in Fig. 6
to give the instantaneous per cent water as a function of
per cent of available oil produced. It is assumed that the
oil-production rate is held constant for the life of the well.
(Available oil is equal to the pore volume less the connate water and residual oil.)
This plot, which predicts the water-production rate for
all periods in the life of a well, is a basis for determining
the economic advantages cf barriers for specific areas.
For example, in areas where the cost of lifting and disposing of water is negligible, even large reductions of
water production may not offset the cost of a barrier. In
other areas, #here the cost of handling water is high,
the payout period could be very short. Another consideration is the prevention of premature well abandonment because of high water production.
The solutions of the coning problems as outlined herein
are so cumbersome and time-consuming that one might
be reluctant to use them. The diagram shown in Fig. 6
can be approximated with simplified calculations. With
the aid of Eq. 1, the rise in water table needed to produce
a maximum stable water cone for a specific production
rate can he computed (original oil-zone thickness minus
calculated h). The volume of oil represented by the drainage area and the rise in water table will be water-free
oil. With the aid of Eq. 2 the cone shape can be plotted.
This cone can be broken down into thin concentric cylinders from which the volume in the cone can be estimated.
The sum of the oil volumes represented by the rise in
water table and the cone will be the water-free oil. This
point may be plotted on a graph similar to Fig. 7.
A straight line drawn from the point of first water to
the terminal point on the graph (100 per cent water-cut
and 100 per cent oil produced) gives the diagram shown
in Fig. 7, which is an excellent approximation of Fig. 6.
TEST METHODS

l~
2

YEARS

TO FIRST

10

WATER

FIG.4PERNM OFWATER-FREE
OIL PHODLXTIOS
WITURmalms.

OiL

PERMEABILITY OF BARRIERMATERIALS
A completely impermeable barrier is desirable. The
cements available, however, do have measurable amounts
of permeability. Considering the large area and thin crosssection of barriers, it is clear that the permeability of the
barrier is important. As an aid to determining the amount

ZONE
40

L
INTERFACE

.
-\

WATER

Z(r)
.
ZONE

0.5

i2
~

40

a
w
n

20

DARCYS

RADIUS .5 A
z

----,

-u-

~LIMITING
WATER
STREAMLINE
X()
k

ZONE

25 FT. PENE

\l
t
z
01

ACRE

100 FT# Oil.

5-CONE

SHAPE AND LMITING STahtwmE


FOR SIMULTANEOUS PRODUCTIONOF WATER ANDOIL.

JULY, 1962

10

20

PERCENT

40
AVAILABLE

so

60
CUMULATIVE

OIL

100

PRODUCTION

FIG. 6-PER
CENTWATER PRODUCTIONAS A FUNCTION
OF PER
CENT AVAILABLE OIL PRODUCED(BY RICOROUS CALCULATIONS).
78s

of permeability that can be tolerated, the following equationvias derived, . .

a hand-operated piston pump, A dead-weight tester modified to act as a pressure-relief valve was used to maintain
the pressure at a constant value,
The cement to be injected into the fracture was mixed
in a high-speed blender in accordance with API standards
or the recommendations of the supplier. The cement,
contained in a sump, was delivered to the wellbore end
of the fracture system through M-in, steel tubing by
means of the hand pump.
A differential-pressure gauge connected between the
lines leading to the inlet and discharge ends of the test
block gave the pressure drop through the fracture system,
Two graduated cylinders were used to measure the
amount of cement injected-one
cylinder was at the inlet
of the injection pump, the other at the discharge of the
dead-weight tester. After the cement was injected, the
pressure and temperature were maintained for a period of
several hours to alIow the cement to attain an initial set.
After the cement cured in oil for a period of at least
72 hours, l-in. in diameter cores were cut from the block,
These cores were then used to determine the permeability
of the barrier by means of a Hassler cell assembly. The
permeating fluid was separated from the pressure source
by mercury to insure against gases going into solution in
the fluids,
It was found that for one of the cements, called Resin
Cement A, tne resin phase filtered into the fracture faces
under very law pressures. There was concern that under
high pressure the resin would penetrate deep enough into
the oil zone to damage the well. To check this possibility,
tests were conducted in a Hassler cell by placing enough

%=(+)(5)+
r+-~n(rbr)l
.

.?

.0.

(4) :3

The condition under which this equation applies is that


the oil production be high enough to maintain a water
cone under the entire barrier without producing any
water around it, The water-oil ratio is influenced by the
ratio of barrier to formation permeabilities. Higher barrier permeabllities can be tolerated inhigher-permeability
formations.
TEST EQUIPMENT AND PROCEDURES
It ivas the purpose of the icsting procedure to determine
the permeability of cements after they had been subjected
to conditions similar to those expected to be encountered
in a reservoir.
Tesi Specimen

Cements were injected into a 50 section of a 5-ft radius


fracture system. The fracture system consisted of two picshaped blocks of Berea srmdstone separated by either
shims or propping material. The sides and ends were
sealed with epoxy resin-impregnated cloth. Enough layers
of cloth were used to keep the blocks together uhder injection pressures of 50 psi above simulated reservoir pressures. The top, bottom and inside surf aces were not
treated. To subject the cements to the same type of
fluid-loss conditions that would be encountered in a reservoir, a saturation history simulating reservoir conditions
was imposed on the blocks. The procedure consisted of
saturating the blocks with water, and then displacing the
water to a low water saturation with oil.
Cor&tant-Temperature

OIL
CEMENT
SLURRY

Pressure Vessel

The fracture assembly inside a constant-temperature


pressure vessel is shown in Fig. 8, This device consists
essentially of a section of .8-in. steel pipe with pipeline
closures on each end. The pressure vessel was designed
to withstand pressures up to 1,500 psi, The copper coil
inside was a heat exchanger. An electric heater in a sump
outside the pressure system was the heat source. A centrifugal pump circulated the hot oil and was turned on
and off by a thermo regulator.
The pressure in the vessel was increased by means of
-.-.-.

II

FIG.6-CONSTANT-TEMPERATURE
PRESWR~
VESSEL.

*The derivation of Eq. 4 is given in the Amendix,

100

- 40 ACRE
100 FT OIL ZONE
25

FT. PENE.

0.5

DARCYS

RADIUS .5

.25 POR. VISC. 1.0CPS 100io EFE .6 W? GR. 10

20

PERCENT
Fxc. 7P?R

786

40
AVAILABLE

CENTWATE.ASA

60
CUMULATIVE

60
OIL

100

PRODUCTION

FUNCTION OF PER CENT AVAILABLX


OIL PRODUCED(SIMPLIFIED METNOD).

i2&:,, 16(J: , .200.


340
500
:. ., :
... ....
.. .. . . . ,..-..;PRES5XJRE PSIG
FCC.9PENETRATION OF RSSIN A.

625
~~

JOURNALOF PETROLEUMTECHNOLOGY

cement on top of a 3.5-in. long core to more than saturate


the entire pore space, then applying high pressure on the
cement. The titer-cake build-up was such that 625-psi
dtierential pressure was required to penetrate 2.28 in.
of core, Typical results of these tests are illustrated in
Fig. 9. No appreciable damage to the oil zone is anticipated because of this penetration.
TEST RESULTS AND OBSERVATIONS
Table 1 summarizes the results of tests made with four
commercially available cements, The costs of the various
cements tested are listed in Table 2.
DISCUSSION OF RESULTS
BARRIER CEMENTS
The cements reported herein are commercially avaifable
and small volumes can be placed either with a dump bailer
or through tubing using tubing plugs. For larger volumes,
methods and equipment applicable for conventional Wment-squeeze techniques may be used.

is obtained. In one Lest, the main body of the cement was


flushed through, leaving only the damaged faces to act
as a barrier. With the irdet tube open to atmospheric
pressure and a hydrostatic pressure of 50 psi applied to
the outside of the entire block, the rate of seepage was
such that only 0.3 BWP!3 would be produced through
a 5-ft radius barrier. This damaging characteristic of Resin
Cement A constitutes a fail-safe feature in case of accidental overflush cm backwash of the main body of cement.
Another possibility exists. By damaging the fracture
faces near the wellbore and extending the fracture out
to a large radius, it may be possible to obtain the advantages of a fracture and at the same time control either,
or both, water coning and gas coning.
Resin Cement B

Resin Cement B is a plastic cement composed of gypsum


cement and a powdered resin, There was no apparent
penetration into Berea sandstone at a 3,000-psi differential
pressure. The flow behavior was found to be satisfactory.
Latex Cement

Resin Cement

Resin Cement A is a blend of water-soluble resin,


Portland cement, water and a set control additive. After
this cement is set, it will not dissolve in water, The flow
behavior was found to be satisfactory. The slurry tills
.a 5-ft radius open fracture (1/32-in. wide) or a fracture
propped with a monolayer of 8- to 10-mesh walnut shells
at pressures less than 50 psi above the surrounding pressures. A filter cake forms on the faces of the fracture and
limits the resin-phase penetration evenly along the faces
of the fracture (see Fig. 10). At 625-psi differential pressure, the resin phase penetrates 2.28 in. into Berea sand stone, It was not possible to fill a closely packed multilayer sand-propped fracture (20-40 mesh) out to a radius
of 5 ft at 50 psi with Resin A or any other cement.
With Berea sandstone, the depth of penetration of the
resin phase into the fracture faces was %s in. for pres-.
sures under 25 psi. This penetratku decreases the permeability of the sssndstone to a point where a good barrier
..
lPEf!MEASl
TA81E

L17Y AND

WATER-FLOW

CAPACITV

OF

VARIOUS

Cement
Resin

..

Resin A
Spearhead

Resin A
,Penefrotion

Portland
only

%sin

0.00009

0.0020

().00022

The neat cement used during these tests was Portland


Class A, cement, Considerable difficulties were encountered whale trying to fill a %,-in. fracture. The fluid
loss was so great that the filter cake bridged the gap and
a 5-ft radius fracture could not always be filled. Fig. 11.
shows the pressure build-up as the fracture bridged. It
is realized that in the field much higher, pressure and flow
rates are possible than would be practical in these tests.
Cement in oil wells can be forced into the formation by
either increasing the fracture width or producing new
fractures. However, because of the fluid loss and bridging,
it is doubtful if channeling in the barrier could be prevented.

AF1

Resin A and Portland Cemenr

0.00025

0.000173
0.00194
0.0400

0.0160
0.0180

0.3/00

Portland

Cement

In an effort to find a way to reduce the cost of matwiids,


tests were made with a limited amount of Resin A cement
used as a spearhead, then following with neat cement.
No bridging difficulties were encountered. The permeability
and flow capacity of the combination harrier were slightly

SID/ftz/psi
.-. .

0.000s

0.0023

Latax

Portland

BARRIERS

W;:;.yv
Permeobllity
-(red)

Latex is a blend of Portland cement, water and a


synthetic latex. There was no penetration into the pore
spaces at 50 psi. The nature of this blend is such that
penetration is not expected at higher pressures. The fl,ow
behavior was satisfactory.

.-...
TABLE

2AOST

OF

BARRIER

MATERIALS

cost

Material
Resin

S Cement

Resin A Cement
[LOW Tompsroture,

200

901

Next

300

WI

above

506

sal

.40

Neat

Class

Siurtv

API

VOlume-1.lS

JULY, 1962

56.90/sal slum

.
.

$3.75/s01

slurry

$3.50/Bal

slurry

$3.25/wIl $Iwrf

$1 .10/001

$1 .65/sack

$0.19/s01

,$,

Latex Cement
Slwry Vehm+l
Cemenf,

SO to 120 F)

First
All

,..

fP/5ack
A

fF/sack

S1.w

,.,

,-

.
slurv

FIG. l&PENETRATION

OF RSSIN PH.KS.EOF RESIN X

CEMENT.
787

v. == linear velocity at a specific radius

I
~

100

u)
U3

80

~:

60

&w
&
1I&l
-f
z

::2-

z ,.

p
~.

. NO APPARENT

FLOW OF CEMENT40 - AFTER 25 PSI


20 I
I
0
150
100
0
50

P.
p

p.

i%, 11PRESSURE

P.
~

200

VOLUME CEMENT DISPLACED


CUBIC CENTIMETERS
BARRIER

height of oil-water interface


=height ofoil-water interface atdraitmge radius
=height of oil-water interface at wellbore
= viscosity
= viscosity of oii
= viscosity of water
= density
= density of oil
= density of water

= height of limiting water streamline

Z =
Z..

1/8 THICK

BLXLD-lJP FOR NEAT CEMEXI,

higher than that obtained with RLSin A alone. When considering the cost of resin cement ($3.25/gal slurry) against
the cost of Portland cement ($0.20/gal slurry), a considerable amount of savings can be realized.
CONCLUSIONS
[t should not be inferred from these discussions that
harriers can always be placed and that these will be
effective in solving all water problems for all reservoir
conditions. The question of whether or not a horizontal
fracture can be produced in a given formation is of fundamental importance and should be resolved first. Reservoirs containing high iensity or high-viscosity crude oils
or having .very low pwmeabilities or a small oil-zone
thickness may prove to be poor candidates for barrier
treatment. It is believed that the methods outlined in this
paper will permit reasonably sound decisions regarding
the design of barriers and will permit reasonable predictions of advantages to be gained by piacing a barrier in
a specific reservoir.

ACKNOWLEDGE

ENT

The authors are indebted to P. L. Terwiliiger who supervised the work covered herein, They are aiso indebted
to J. E. Warren for the mathematical assistance and
advice given during the development of the analytical
soiutions, Finaily, the authors acknowledge the work of
R. A. Neison who assisted with the laboratory tests.
itEFflRENCES
Oil WeRTreatment,U.S. Patent No. 2,832,416
(April 29, 1958).
Piety, R. G.: Oil Well Prwluctiol~ U.S. Patent No. 2,886,108
(May 12, 1959),
Michaek, Aian S., Bergman,R. I. and Brown:
CoT!..;tY-. .m
Coning byInjecting Reverse-Wetting.4dditi@, P@. En&.
(Aug., 1957) B45,B-52.
I-foward,George C: PIugging Forma,ims with Hot SOIU.
tions, U. S. Patent No. 2,779,415[Jan, 29, 1957).
Kravcilenko,1, I. and Kartsev,E. V.: Field r~stson Isolatir!:

1.

Aflen, J. C,:

2.
.3.

4.
5.

of Bottom Water in Fields of Bashkiria by the UFII Method ,


fie~.tyunocKhoz (Oct., 1960) 38, No. 10, 20 (Translated by
Associated Technical Services, Inr., P. O. Iiox 271, East
Orange, N. J.)

6. Huitt, J. L,: Hydrarrlic Fracturing

with the Single-Point


f3ntry Ieclurique,Jour. Pet, Tcrh. [March, 1960) XII, No.
3, 11,
McGraw.
7. Muskat, M,: Physicvd Principles o~ oil Production,
Hill Book Co., Inc., N.Y. (1949) 226.

8.

NOMENCLATURE
D =depthof

F
~
h
k

wellbore penetration

= function
= acceleration.of gravity
=oil-zone thickness
= permeability
k. =barrier permeability
k, =formation permeability
P = pressure
P. = pressure atthe drainage radius
p. = pressure attheweiibore
q = volume flowrate
go = vohtme flowrate of oil
910= VOiUmeflowrate of water
r = radius
radius
~r. = drainage
r. = welibore radius
t, = barrier thickness
U= original thickness of water zone underlying. the
oil
v = linear v.!ocity
788

9.

10. I.m;e, D: K.: Two.lbase Flow-, Unpthlishcd

File Note
KR.51 (.$pril IO, 1%1) Gulf Research & Dewlopment Co.,
Pittsburgh 30, Pa.
APPENDIX

PRODUCTION

OF WATER THROUGH

BARRIER

When an impermeable barrier is used toprevent water


coning, it is important to know how much water wiii
penetrate the barrier and be produced with the oii. If the
problem is restricted only to the case where the water cone
stays below the barrier, then the greatest rate of water
production occurs when thecone isatits maximum stabie
position.
To solve this probiem, it is necessary to know the
velocity of the oil as it flows through the formation just
above the barrier. A simpie approximation of tilis veiocity
is the average velocity. As a function of radial distance,
the average radial velocity is
v

=v(r)=~=_A2p -,. . . . .

(A-1 )

JOURNAL OF PETROLEUMTECHNOLOGY

Since the quantities above are functions of r only, the partial derivative can be replaced by the total derivative. Then
integrating between the limits r,. and r, the pressure in
the formation above the barrier can be obtained. i.e.,

P=(2+M:3
~~~~
_
()
q,, ,n

/1,,

~,

727rD

.,

Then, from a continuity standpoint and neglecting compressibility, the flow rate of fluids through a vertical cylindrical surface, concentric with the wellbore at any radius,
will be the same as the flow rate into the wellbore, The
flow of each fluid becomes as follows.

A-2)
(A-3 )

r,.
At the end of the barrier, the pressure in the static
water phase just below the barrier is the same as the
pressure in the moving oil phase just above the barrier.
The pressure in the water pha~ is the same at all points
immediately under the barrier; this pressure is

$=:$D1n
()T.
rti
,.

..

f,,
~p==~~-ifi-in

[A-5)

(A.6)

k, 2zD
()
To find the total production of water through the barrier, it was necessary to assume the barrier to be a flat
disc of radius r, and thickness I,. In addition, the barrier
was assumed to have a uniform permeability k~.

From Darcys law, the velocity of water through the


harrier is
k,, ?P
v:=-.-=-

k,, Ap

p,. az I!.w/,
The incremental water production is

( A-7)

( A-8)

&

Then integrating between the limits of r, and r,,, q. is

,.

SIMULTANEOUS PRODUCTION
WATE~ AND OIL

ar

= p,x (z - z,. ) + p.g(zo z),


p. p. = p.gz - p,m,o + Apgz

X(r)

==
ar

Apg:

. . .. - 94.

2zr,r:(/l_

2wk(Z

(A-IS)

dz+= 27r .k. (Z X) !??,!


J.L,,.

z)

-X)

= 8Z(r)

,.

(A-16)

(A-17)

(A-lfl)

-/3

(A-19)

h= Y(r.)

Di-U
Z(r,r)

(A-n)

(A-12)

- D) Z(rp) t UZ(r,. )
Z(r,,) Z(rm) o

1
q,c/.Lw -
q,,p,, . .
Ap&Z$
, e ( A-20)
2zkr [ Z -x
h-z
I
Since r and Z only are involved, it is possible to set
dr

(l

=-

The partial derivatives in Eqs. 17 and 18 cm now he


substituted in Eq. 14;

dz
az
= ~,

OF

The partial derivative of Eq. A-12, with respect to r, becomes

ar

where the constants 8 and /3 were determined to be

and the following differential equation is ob-

tained where the expression X(r) from Eq. A-18 has been
included.
1
q,.!.,.,,
.
.. ..._ q+,
2irkr(li-8)Z+#
11-Z
[

=I

Apg$.

(A-21)

Eq. A-21 in integral form is


r.

dr
.
.. .
2xkApg H r
rw

Z(r.

or

apn

Assuming that the shape of the limiting water streamline is established in tbe same manner as the oil-water
interface and is similar to the interface, then

(A-1O)

The simultaneous: production of water and oil is a


problem in two-phase flow. Fig. 5 illustrated the flow behavior of the two fluids. For the following approximate
steady-state solution, it is assumed that a static pressure
relationship holds between all points at tbe same radial
distance. The pressure relationship between the three
points shown may be expressed as follows.
p(r,z,a) - P(r,zo) = P!. p

apw

qt./.&
.-

~P
.

q,, .,,.,~)(:);zf<.+.=$~]
.

and

8
(A-9)

and

k,, Lip
= 2~r dr
p.!, tl>

dq,. = ( 2~r) (v:dr)

% H)
x
. . .

ap.
s

.)

~-

Rearranging, Eqs. A-15 and A-16 become

In+

Irr:

k
-- q. = 27r

The pressure drop across the barrier is


z~&

(A-4 )

wr.ere r~ is barrier radius.

.\p+p=.

Z(rtc)

[k-z] [(l+8)z +/3]


J -.
1.Y!41- qd -[%rwr+ 9.2-(I + 8)1 = \z .
. . . . . . . . .!.
. . . (A-22)

To make the results more usable, the following dimensionless quantities have been introduced.

(A-13)

Only the horizontal components of velocity as defined


by Darcys law are considered,
v. = JULY,

1962

k
~%and~
=
p.. ar
r

apt,

p,. &

,.,

(A-14)

The result of the integration of Eq. A-22 then is


789

*.,

J+H+E

Y= 0.999:::;

J+H+ E- Y+a
Rw( E-Y)
[-aRw
(EY)][J+H
+E-Y-(JE+JfY)]
[J+
H+
EYl[J+
H+
EY-l-aRw(E
{

,nY-J-aRm(l

-Y)

E+ H-aRw(l
-E)
(J+ H+ EY) S;aRw(E-Y)(JE+HY)
[J+
H+ EY][J+H+E
Y+aR.

(.E-

Y)l

Y) + % (~

- f?)

=27&Pln

Y)]

%?
27rr,h( 1 Y)

~w

kap

.,
P.

= %rrwh( Y - J)

-kap
..
~Lm

iw

q+

q.p.

27WWWZ(
1 - Y) - 2wrWk/z(Y - J)
+J

aRm+l. ..

A-24)

-.

JA-25)

This approximation of Y cannot be used in Eq, A-23


as it stands since the value for Y makes the natural log
term indeterminate. However, by adding a small increment
to Y, it can be used. The revised expression used for Yis

790

THE SHAPE OF A CONE

To predict tbe shape of the maximum stable cone in a


radial system, it is necessary to know h, D, r. and r-.
= O.

Eqs, A-13 and A-15 can be combined to give


82 . , . . . (A-27)
= 27w~ (h - z) APg%
P-.
Only (r) and (Z) are involved in this relation, so the
partial derivative can be replaced by the total derivative
q.

q. = ~kg(pe po) (hz D)


poln(r,/rw)
.
Substituting for q.
1 (h* D) dr
(fZ-z)dz=
2 In (r,/rW) 7
.
re and r. are lked and D and h are known,

(A-29)

(A-30)

(A-31 )

(A-32)

This expression reduces to


y_aR.

(A-26)

But q. has already been listed as Eq. 1 of the report.

ar

If the pressure gradienta in the two zones are assumed


equal at the wellbore, then
ap

be determined independently.

and

-5-

For a radial system with no water production, ~

This is a case of one equation and two unkncwns. In


Eq. A-23, it is desired to assign va!.ues to E (the fractional height of rise of the water table at the drainage
radhrs) and obtain R., the water-oil ratio and to plot
these values as shown in Fig. 8.

v. =

()

= F(cvR, J; H, E, Y) (A-23)

Y must tit

When Eq. A-26 ia used, the number of variables involved in the value of F is reduced from six to five. Thus,
Fin Eq. A-23 is now F(aRW, J, H, E),
COMPUTING

(E-

-1-0.4)01.

(h-

Z)dZ=~-

:
[

;-2hZ+(hZ-D2)Z=h

=;.

h2(h_Dm)
lnrdr
(A-33)
m.
4
The free surface can be plotted by substituting various
**
values of r, mnging from r = r. to r = r..

JOURNAL

OF

PETROLEUM

TECIZNOiOGY

Das könnte Ihnen auch gefallen