Beruflich Dokumente
Kultur Dokumente
Texas
SPE
=R
1531
I5206
THIS IS A PREPRINT --- SUBJECT TO CORRECTION
Gravity
Effects
in Miscible
Displacement
By
M. C. Miller, Jr. Member AIME, Atlantic Richfield Co., DQlas, Tex.
0 Copyright
Ameriesn
Institute of Mining,
1966
Metallurgical and Petroleum
Engineers,
Inc.
This paper ?~asprepared for the hlst Annual Fall Meeting of the Society of Petroleum Engineers
of AIME, to be held in Dallas, Tex., Oct. 2-5, 1966. Permission to copy is restricted to an abstract
of not more than 300 words. illustrations m~
copied. The abs;;act should contain conspicuous acknowledgment of where and by whom the ~aper is m?esented. Publication elsewhere after
publication in the JOURNAL OF PETROLEUM TECfiO~OGY or-the SOCIETY OF F!ETROLEUMENGINEERS JOURNAL is
usually granted upon request to the Editor of the appropriate journal provided agreement to give
proper credit is made.
Discussion of this paper is invited. Three copies of any discussion should be sent to the
Society of Petroleum Engineers office. Such discussion may be presented at the above meeting and,
with the paper, may be considered for publication in one of the two SPE magazines.
ABSTRA13T
Segregation of hydrocarbons and water can
severely reduce the ultimate vertical sweepout
efficiency of mlsclble displacement processes
when a mobile water phase is present in all or
part of the reservoir. Computer studies have
shown that the decreased miscible sweep resulting from segregated flow can result in uneconomic operations in some fields in wh~ch economic
operations were predicted for non-segregated
flow .
Gravity override is enhanced as a result
or segregation, and once the reservoir fluids
segregate, the high rel&tive permeability of th(
lighter fluids at the top of the reservoir
increases the gravity override with the result
that oil irrthe lower portion of the reservoi$
is not displaced miscibly.
The shape andrnovement of the oil bank
formed.during secondary and tertiary projects
can be paramount to a successful field operation. Results from our computer model study
show that an oil bank formed ahead of an overriding miscible fluid is forced down below the
solvent tongue and moved relatively slowly
towara the protiucingwell. The oil bank i.s
slowly recovered immiscible at an extremely low
rate by the water injected along with the gas.
Thus at the economic limit of the fielcloperation, a substantial.amount of the oil displacea
References and illustrations at end of paper.
GRAVITY EFFECTS IN M~
process, a propane or rich-gas slw which iS
miscible with the reservoir oil is driven by a
gas which is misctble with the slug but not
miscible with the oil. lower pressures [1,200
psi to 3,000WI are usedand a sufficiently
large slug must be employed In order to prevent
premature loss of miscibility by diffusion and
dispersion of the slug material.g Injection of
water along with the gas has been beneficial,
and oil recoveries in the 50 to 70 per cent
range can be obtained.3 Iackof complete
contact between the reservoir oT1 and miscible
fluid at the economic limit of a field operation prevents oil recoveries of 100 per cent.
Interpretation and prediction of Lhe
reservoir performance for miscible processes
are difficult as a result of the effects of
viscous fingering1~2~11~12~16 [dynamic instability of high mobllf.tyfluids displacing
lower mobility fluids], loss of miscibility
from diffusion and dispersion of the slug
materia19?16 heterogeneities and
gravity.1)2)13>16 me effects of these factors
and their interaction plays a major role in the
areal and vertical sweepout performance and
thus, ultimately, the economic 0<1 recovery.
methodox
Unfortunately, no laboratory
mathematical procedure is available which can
be used to analyzeall the forces and effects
in a complex reservoir system. In spite of
their limitations, physical [laboratory]and
mathematical [computer] models have given
significant insight int6 oil recovery mechanism
which allow us to predict reservoir behavior.
Habermann12 found that if a gas-driven miscible
slug is less than 10 per cent of the oil in
place, the slug will deteriorate and the displacement will revert to an ordinary immiscible
gas drive. Craig et al.1 found gravity can
severely reduce the vertical sweepout performance in frontal drives in displacement studies
which were predominantly controlled by gravity
[single override finger]. On the other hand,
Blackwell et al.16 tests were finger controlled
[many random fingers] and observed recoveries
were considerably higher. Crane et al.15
investigated displacements In the transitional
range between finger-and gravity-controlled
displacements, and postulated that a functional
relationship must exist between the ratio of
viscous to gravity forces, ratio of reservoir
length to helght~ and mobility ratio. We agre<
with Crane et al.15 fora homogeneous reservoi~
but the square root of the ratio of vertical tc
horizontal permeability times the length to
kv
height ratio,
J
~H
, as publtshed by
E31LEDISPLACEMENT
SpE-~5%1
-, /-
L.i .
il-Jq>J.
1.
L
geometry, ~
u .
kv
kh
F
2.
kr~
mobility ratio, IvI = _
Bd
k ro
/
40
/lo
~
viscous to capillary pressure
gradients,
+V /Lo
we
v .@w
= - +
= gravitational constant
k; = specific permeability in horizontal
direction
kv = specific permeability in vertical
direction
= relative permeability to oil
k
k; = relative permeability to the displacing fluid
M= mobflity ratio
q = fluid.injection rate
q
v= total linear pore velocity, v =
~
A cross-sectional area
L= length of system
= thickness of system
$ = density difference between the injected and displaced fluids
c= interracial tension between the injetted and displaced fluids
e= contact eagle
Go = oil viscosity
Pd = displacing fluid viscosity
0= porosity
In the mathematical model, it is not
necessary to use scaling groups to predict the
behavior for a given field test. However, the
use of dimensionless scaling groups Is useful
In correlating the results of several field
tests. The resulting correlation can be used
to estimate the flow behavior for an unknown
field and is therefore useful in screening
potential miscible flood prospects.
The effects of geometry and the ratio of
VISCOUS to gravitypressure gradients were found
to be paramount in the flow behavior as will be
shown later.
a%
n-i-
q=
ash=
= -+
[3] ~rcys
8SW
,
at
. [11
[2]
k krW
o c [31
V*W,
. .
P~
[h] Darcys law for the hydrocarbon phase
k r(hc)
q==.
%hc
7*
~hc
[4]
[51
~hc=%c+
~hcgh,
. . . . . . . [6]
=So+ss+sg=(co
hc=l-%v
+c~+cg)shc,
[8]
. .....
.[7]
~ h~ = CO(IO+
CsPs+
CgPg,
. .
.,[81
l/4
(~)
~hc
=Co
(b
1%
+c~(
1/4
)
IJs
--+
w=-
where:
TWO-DIMENSIONALMAT
HEMATICAL
[11avolumetric balance[continuity
equation] for the water phase
Po
L,LLJJLJULI
+-c
g@40[
GRAVITY
EFFECTS IN MISCIBLE
DISPIACEMEXI!T
.-- .- -,-..
---
-. . .
pc(sw)=Phc
-Pa,.....
,...[10
where
Subscripts
c = capillary
g = gas
hc = hydrocarbon phase
o . oil
rw = relative to water
ro = relative to oil
s = solvent
w = water
A finite-difference, alternatingdlrection, leap-frog procedure, similar to the
procedure outlined by Couglas, Peaceman and
Rachford5 and discussed in detail in the
doctoral dissertations of Nielsen6 and Goddin,7
was employed to solve the above equatio~s.
This method was modified to include the flow
behavior of three contiguous miscible flulds
[oil, solvent snd gas] In the hydrocarbon phase
The effects of diffusion and dispersion between
the fluids in the hydrocarbon phase were
Ignored, leading to a sharp interface between
the solvent and oil and the gas and solvent.
the denstty
Eqs. 8 and 9 were used to calculate
and viscosity, respectively, of the hydrocarbon
phase.
Although the mathematical model is restricted to,two-climensfonalflow systems,
the effects of gravlty~ capillary pressure,
stratification, flow rate, slug size, reservoir
length and thickness, reduced vertical permeability, and reservoir fluid pr~pertles can be
investigated.
The special
case of no vertical crossflow
[zero vertical permeability] was treated sepaa
rately ana analyzed for an itiealizedsituation
in which there are no relative permeability or
capilla~ pressure effects. These calculations
+
-%
v?aoviscous-to-gravitygradient ratio,
. The resulting
J~g
AP
product is a viscobs-to-gravity-pressure-drop
@v~L
ratio,
predicted using the two-dimensional, mathematical model aescribea above. Since the
physical properties of the reservoir are not
precisely known, [e.g., the ratio of vertical
to horizontal permeability may be as high as
1.0 or as low as 0.003] several different case:
were,run for each reservoir. The velocity
[throughput rate] was varied for proposed fi.elf
tests to determine the injection rates require<
to prevent gravity override and segregation.
The reservoir flutd properties are summarized
in Table 1 and the twenty cases investigated
listed in Table 2.
A predicted performance is classifieiias
gravity controlled if a single finger [gravity
tongue] is predicted as shown in Fig. lA. How.
ever, if the fingers initiated in the higher
permeability layers [alternate layer-permeabilities were increased 6 per cent to Initiat<
fingers] continued to grow as shown in Fig. X3:
the displacement is designated as ffnger controlled. If the fingers grow, but the fingers
at the top of the reservoir grow considerably
faster than the fingers at the bottom when a
lighter fluid is injected, the displacement is
designated as finger-gravity controlled.
,
The predictea performance of the 20 cases
investigated are presented in Fig. 2. These
results were correlated with the geometry ,ratic
F
%L
k~
M.
.-.
-,/-
f!.
.
MTT,T,F!R
.--
5
.
ratio,
ini~ial
EWE-1531
2$7/
6. Nielsen, R. L.:
2. If aQ override
formed, the oil bank ts
centierof the reservoir
immiscible by the water
gas.
gravity tongue is
pushed down into the
and slowly displaced
injectecialong with the
J--
REFERENCES
I
1.
M. C. MILLER
E-1531
..
APPENDIX--GRAVITY OVERRIDEIN TWODIMENSIONAL SYSTEMS WITH NO CROSSFLOW
L
and theabove-
ratio [dimensionless]
khgA~
cited mobll%ty ratios,
where:
g =
AP=
H=
v=
PO
gravitational constant
density difference
reservoir
thickness
fluid velocity
fluid viscosity of the in-place
fluid
L= reservotr length
kh = reservoir horizontal permeability.
=(3 f(yo)
- (4f(yo) +
+
wh.e
re:
Sf(yl)
2f(Yl)
6(vP))(~)z
-6(Vp))(~)
f(Y~)
= position of interface at a
given value of dimensionless
hetght.
TABLE
TABLE I
CASE
RESS2RVOIRAND
FLUID
PROPERTIES
- CASE
STUDmS
Permeability
Ratio
Field
Reservoir
Length.
ft.
Reservoir
Thickness,
Reservoir
Porosity,
1034
ft.
Reservoir
Horizontal
Permeability
(Oil at
conoate water) md
15
%
22.0
Field
2660
Field
2660
32.2
84
22,8
7.8
102
22.4
STUDIES
Field
.
Field
E
Case
Field
kJkh
330
3020
0.1
25
57
Geometry
Ratio
YE+
Viscous
to Granty
40.7
5.3
0.03
22.4
9.7
0.03
22.4
48.4
56.7
10
0.03
22.4
97.0
20.7
10.3
0.02
18.3
59.5
Gradient
0.969
0.992
0.992
1.000
1.060
0.015
16.2
13.5
0.594
0.603
0.603
0.727
0.547
0.01
12.9
16.8
Solvent
Gravity,
gin/ccO.111
0.254
0.254
0.504
0.019
0.003
-1.1
30.7
LeanGas Gravity,
0.130
0.130
0.082
0.03
14.3
5.8
0.27
0.355
0.355
0.92
0.29
10
0.03
14.3
11.5
0.19
0.248
0.248
3.65
0.13
11
0.01
8.3
9.8
12
0.03
59.0
5.6
10.6
9.2
9,2
33.2
5.4
13
0.03
59.0
11.2
14
0.03
59.0
44.8
15
0.1
4.2
12.3
16
0.1
4.2
24.6
17
0.01
53.0
0.77
18
0.1
16.8
0.24
19
0.1
lb. s
2.42
20
0.01
5.3
0.77
Water
Gravity,
gmlcc
OilGravity,
gmfcc
Water
Viscosity,
Oil Viscosity,
gmlcc
cp
cp
Q 100
Solvent-Lean
Gas
Viscosity Ratio p~lfig
1.2
1.4
1.4
Effective
Tension,
20
20
20
Interracial
dynes/ cm
7.8
20
--
--
20
Ratio
A,
CASE
2:
GF?AVITY
CONTROLLED
A
SOLVENT
d
+
z
0
i=
0
a
0.2
0.4
0.6
DIMENS1ONLESS
B. CASE
16:
FINGER
0.8
1.0
LENGTH, f
CONTROLLED
A
0,2
0.4
0.6
DIMENSIONLESS
0,8
1.0
LENGTH, f
Fig. 1 - Predicted
Hydrocarbon
Ilistribut,ion Showing
Gravity
and
Finger
loo~
Controlled
I 1 1111
A-FIELD
[11(1
II
07
O-FIELD
-FIELD
V-FIELD
B-FIELD
Displacements
I 1[111
12
#
, #$
*I
,~~:
At
Oe
10 .
0 g
GRAVITY
J?
C~NTROLLED++
E
M
z
o
kH,kv=HORIZONTAL
AND
kro,km
PERMEABILITIES
1.0
v=
TOTAL
+=
POROSITY,
Af2
w
to
~RELATIVE
FLUID
=DENSITY
[,,
I
VELOCITY,
MILLIDARCIES
ANO *ATER
Fmt/Ooy
SLUG,
OF
BETWEEN
MOBILE
OIL
ANO
WATER,
PLACE
FLUID
AND
1111111
1 I
111111
3644
m
{
2 - Predioted
Linear
Reservoir
Performance
Flow
System:
(Two
i
1111-11
Vertical
Dimensional
Computer
1000
~rO
_._+~},N,TML
Ap
.
Fig.
1 I I
100
10
Sweep-out
INJECTED
CENTIPOISE
IN
Gm /Qc
,,
111111 I I II11111
0,
OIL
DIFFERENCE
VISCOSITY
05!01
VISCOUS
PORE
TO
~.
FRACTION
MISCIBLE
#Lo,~wm
PERMEABILITY,
VERTICAL
co ROLLED
Is
Vv
[
I IllJj
;ER
15
&+*
:4
6/0A
lo4&7/
020
>
I I 1[11
Model)
DIMENSIONLESS
LENGTH, +
DIMENSIONLESS
(C) AT
END
OF
FIFTH
LEAN
LENGTH , +
GAS -WATER
OIMENS1ONLESS
(D) AT
END OF EIGHTH
LENGTH,
CYCLE
(0.25
LEAN
PV)
LEAN
c3AS
PV)
I
SOLVENT
ii
**
z
4
OIL
40
BANK
F
o
a
-a
o
c
0.2
0.4
0.:
DIMENSIONLESS
(E) AT
END OF ELEVENTH
LEAN
3 - Predicted
LO
LENGTH, f
DIMENSIONLESS
Fig.
0.8
LENGTH, +
Hydrocarbon
Case 9
Distribution
Pv)
g
b
0.4 -
2
L
~
0.2 -
PREDICTED
1-
SOLVENT B.T.
OIL B,T.
00
0.4
0.6
- Oil
0.8
PORE
Fig.
EXTRAPOLATED
___
----3
---
1.0
1.2
1.4
1,6
VOLUME
Trom
(A)
Flow
Field
CASE
1.8
THRUPUT
Fractional
Stream
B,
in Producing
Case
1, $Hmo.l
-1
SOLVENT
4
-0
-o
0
0.2
0.4
DIMENSIONLESS
(B)
0.6
0.8
1,0
n
n
LENGTH,*
CASE 7, ~=0.01
DIMENSIONLESS LEN6TH,f
(C)
CASE 8,&&90.003
-1
-s
-g
Ii
z
q
-o
&
0.2
0.4
DIMENSIONLESS
Fig.
at
5 - Predicted
end
of
Solvent
0.6
o.s
I.o
LENQTH,+
Hydrocarbon
Slug
----
(O.1OPV)
Distribution
- Field
of
#p
PORE
0.2
c!?
VOLUME
0.4
~osition
Front
vs
of Leading
and Trailing
PV Thruput
in Linear
with
no Crossflow
??1:
THRUPUT
0.6
0.8
Edges
Reservoirs
x+
.
w
:
Lo
LLJ
g
:
~ 0.8
UJ
*S (*)(*)
0.8
1.0
i+ (*)(+)
*=I
a
w
-1
#.1
x~A
$=3
*.3
0.6
L
0
0.6 -
f-lo
o
~ 0.4
g =100
1-
o-l
o
a
0
a
(l)
0.2
z
g
0
0.2
PORE
A-1
- For
(Up
0.4
VOLUME
0.6
0.8
5
THRUPUT
Mobility
Ratio
to Breakthrough)
~ 0.2U-J
w
J
30
65
z
w
s
=
a
0.2
PORE
A-2
0.1
0.4
VOLUME
0.6
08
E
THRUPUT
Ratio
- For Mobility
(Up to Breakthrough)
= 1
1.
XA
X*P
PORE
0.2
ElJii
0.8
VOLUME
THRUPUT
0.4
0.6
0.6
0.4
0.6
0.8
1.0
#,[
#=3
0.6 -
g=,o
04 - &loo
0.2
o
=
P~RE
A-3
VOLUME
THRUPUT
= 10
0,2
PORE
A-4
VOLUME
THRUPUT
= 100