Beruflich Dokumente
Kultur Dokumente
May
Improve
Condensate
A.S.
CuHick,
M.F.
Cohen,
SPE,
r.A
sensitivity
H.S.
and
Lu,
J.P.
and
Mobil
Watson,
E&P
SPE,
-.O
of the process
gas.
The
to several
process
iinchdkrz
variables,
mechanics
were
reservoir
confrmed%y
Technical
Jones,
SPE,
Consultant,
and
Center
process
may improve
sweep efficiency
and gas-condensate
recovThe study used extensive
numerical
stimulation
to inYeStiga@
the
Iavering.
ktbomto~
L.G.
Center
E&P
(WAG)
reservoirs.
--v-,
---
Technical
Mobil
.. .. .. .mmn.a
. .. . . .
trapped
Recovery
SPE,
ummary.
Gas=
uermeabtiw,
dis~ia~ements
relative
permeability,
in jaycred
caPflm
pressure,
ad
co;e.
Introduction
Many
WAG
process
to improve
cible
sweep
floods
ciency
Gas
lead
pointing
study,
verting
agent
yield,
swept
that
Water
because
pmce.vs
waters
practice,
because
and
traps
be avoided
and
water
of
the
present
simulation
by
ad
that
by
total
by
is un-
possl%lities
detailed
concerns
about
production.
volume
potential
slug
with
Gas
small
and
es-
re$ervok
discuss
the effects
performance.
effects
from
We
water
ing
Fluid
the proc-
of the simdatiom
process
VOU. Our
remfts
maintenance
from
the WAG
process.
represents
a 16CMcre
positional
simulator
state
@OS)
model
inverted
that
fluid
has
prototype
compae
gas injection
three
five-spot
incorporates
strata
pdtem.
The
re.serpressure-
with those
and
from
approximately
We use a ddly
a Peng-Robirmm
properties.
mcdels
simulator
was
described
layer
model,
The
is a layer-cake
meability
layers
resenmir
kh core
(??&
1).
and
permeabfily
strata
a geometric
average
permeability
layer
is
ratios
of=
moderate-permeability
model
data
moderate,
within
=8%
100:1
called
model
The
welI-log
(high,
permeability
and
model,
the
strata,
with
is a
proto~e
&at
were
three
different
developed
averaged
into
of the total
and
the three-permeability
ranges.
single
phase
sensitivity
density
tive
mobility
thickness
= 10: l.compared
respectively.
and
For
the
199S Society.5
Pe!mleutn Engineers
slug
distance]
y di-
balanced
water
at a reservoir
water
slug
water
cycles
a cumuki-
was
cvcle
=0.035%
WAG
Fluids
chosen.
injection)
atier
HCPV,
were
initial
and the
performed
representative
Table
dur-
gas were
of three
compares
in representative
and
gas-cycling
They
were
typical
the fluid
char-
projecfi.
three-component
butane:
and
The
synfietic
assumed
case,
fluids
to be fmt-
gas
(OHIP)
after
injection
is at
Relative
assumed
rela-
to be 75%
repotied
by
et al. 2
2 shows
12%,
was
permeability
Chierici
Fig.
was
by water
relative
limestone.
enters
tie
and
the rela-
kv/kH
was
layer
(i.e.,
use
stratum,
injection
the
WAG.
compared
sweep
reducing
it sweeps
that stratum.
Be.cau*
water
gas
cases
of gas
Almost
90%
with
=80%
is less
efficient
from
layers
required
287.
and
SW=PS
more
40%
continuous
ii the
witi
less
original
the
second
100-md
gas-swept
because
condensate
illustrates
the
the
significantly
l-red
WAG
fraction
for both
mecba-
preferentially
gas therefore
is @creased
injection,
for
gas break-
the dry
but recovered
=61%
gravig
wate<
wellbore,
total
with
fnitial
and
Injected
injected
displaces
curves
hydroca-
HCPV).
(mobility),
where
efficienq
4 compares
cases,
at 0.17
hjected
ratio,
and
injection,
compared
at the injection
mobility
At 1 HCPV
with
years
in that stratum.
10-md
the recovery
of origimd
(a 28 % increase.
viscosity
also crossflow
WAG
19.7
years
Y*S
efficiently
continuous
3 shows
is =78%
to tfds improvement.
100-md
gas
Fig.
recovery
after
19.1
=3.3
contribute
stratum
Re.mIts.
permeability,
ni.srns
both
were
Porosity
WAG
gas
in
were
gas saturation
used
cuwe.
rbon in place
achieved
viscosity
(i.e.,
to be 0.5:
Sknufator
Fw.
fluid
characteristics
happed
of an outcrop
permeability
HCPV
flow
saturations,
watec
saturation
injecdom
initial
endpoints
at 1 HCPV
ty.
fluid
permeabiliV,
segregation).
and 25%
saturation
(except
to relative
gravity
be representative
regions;
the simulation
determine
Base-Case
gas
copyright
those
base
The
Water
the low-
was
were
ratios
at trapped
has average
with
watsr
throughout
gas condensate
from
bigb-
of 76-BID
the first
.sa.ve). fn addition
and
tluee
by taking
The
base-
constant
fnjection
of kjecting
Fifteen
and dry
of a favorable
per-
(with
and reached
23 years.
days
Properties.
tive permeabili~
Configuration.
I-red
the
weWto-well
pattern
beeirmine
was 0.92.
propane,
ed to me top
Model
Each
with
through
of
well
after
(240
Rock
of ethane,
com-
equation
1 shows
con%ct miscible
because
the pressure. was maintained
above the gascnndemate
de~oint
pressure;
the hydmcabons
therefore
remdmd
to.
by contimmus
The
for
in general,
one cmss-sextional
operations
previously.
is to use synthetic,
mechanisms
a 23-ft-thick,
Fig.
HCPV
projects
assumed
propose
and
100~md
acres.
the sas.
gas condensate
tive
The
middlq
10-md
S-tl-thick,
23 years.
and
28%.
permeabilities..
69-t%tbick,
an
and a 1,870-ft
to the gas
ratio
top;
bottom.
five-spot
consisted
with
the
(22x5)
WA,G
gas breaktfuough.
of
ad-
injection
different
water
injectivity
injeztion.
on WAG
adverse
to
production
only
compositional
system
displacement
core
water
so that
of a fully
layered
reserves
Simulation
to study
equal
at the
in the
gas injection
acteristics
water invasion,
and reducin a WAG prOc=s,
wat~
me used.
water
parameters
losing
condensate
volume
results
laborato~
in a two-layer
of
the process
each
of a synthetic
process
gas
into a gas-ccmden-
= 160
of = 1.22
gaslwater
through
mobility,
of the reservoir
tive
akematelv
channels
strata
is not injected
not
water
after
and
and present
ess
gas,
total
is restored
reservoir
dress
dry
by designing
a small
sentially
We
as a di-
cb~ek,
favorable
pat
tipped
gas candensae,
killing wells with
ing injectivity.
h conuast
to a wate.fflood,
slugs
gas sweep
high-pemneability
[110 cefls
of
has a horizontal,
at the
an inverted
Condmmus
of
simu-
by acting
gas to lower-permeab~ky
the lower
because
,improves
stratum
stratum
model
gas-cycling
reservoir
fo!lows
costs
recovery
entering
high-penneabfity
mension)
dkap-
In this computer
model
straw
low-permeability
representing
reservoirs
efficiency,
compression
increases
of the
eftl
in a gas-
case
sweep
strata.
The 2D cross-secdonal
moderate-permeabiliV
gas.
conventional
can
sweep
process
sweeping
~s-
to improve
low
the WAG
dry
applied
ad
in gas-condensate
high
by preferentially
and
by
sate
and
Wd
reported.
breakthrough,
gas coridensate
imbibition
been
pressure
we show
preferentially
h
gas
injected
sweeping
not
recove~.
&vetiing
proposed
gas fi.fiscibl~
of WAG
in high-permeability
and ultimate
and
has
condensate
gas channeling
lation
Use
to maintain
to early
been
pressure-maintenance
reservoir
injected
have
of tijected
in oil reservoirs.
in a gascycling,
condemsatc
can
applications
efficiency
WAG.
of gmwigas
than
condensate.
remaining
at 1
superior
sweep
is swept
by dry
gas injection.
layer
because
TABLE
PRODLWON
lNJEOT[ON
I-MODEL,
iA60Rii0RY
CONDENSATE-RESERVOIRS
wEU
WELL
FLUID
PROPERTIES
Laboratory
and
Fluid
Resewoir
Simulator
Model
Properties
ABC
--
..:.
viscosity
#
6.0
.3.8
5,7
5.5
13.5
10,0
16.3
16.5
H*OUJM
i%omk
Density
p
H,O/p
2.1
2.2
2.1
3.4
3.3
5.0
pressures
for simulator
RI
pH,O@Ig
pus
Fig.
1 Three-permeability
layer
the
average
tibkout
model.
Because
affect
of gas override.
Gas channeling
tum is predominant
Although
botb
tum
from
water
the
in the lower
tum
Fig.
that
the
movement
high
fn addition
continuous
sequence,
dfierent
Tab[e
for
a 3D
permeabilities
tions.
The
puted
by normalizing
the
relative
case,
Fig.
gas
recovery
ities
were
duction
water
in Fig.
6 shows
same
capill~
used
for
the simulation
the same
with
tbmugh
to, retain
The
results
and
without
somewhat
the water..
remaim
low
until
Fig.
very
from
case
for
model.
there
h?wever,
of the
flood.
key
for
@e other
relative
permeabili~
potentially
would
result
tr~ds
low-
wO~d
model
showed
gas
injection
2D model,
a 23 % increase
pared
with
that
nOt
same
layering
ceifs.
The
model
fiv,~spot
pattern
iniection
and
1,870
production
of gas
that
75%
the
with
WAG
ery
while
is diverted
9 com-
cQmhad
the
and 330
160-acre
The
as those
and
gas
strata.
because
the
used
for
only
override.
Recovery
ratiO
which
Shmdation
bas
a mobfitv
a ccmrsen.ingstratum
decreasing
38%
ORE
Gas
does
with
However,
straw.
to 54%
incremental
became
Th?refOre,
be-
effectively
OHfP
and increased
any
displacement
shows
injected
crossflow
by 42%
at the
depth)
at 1 HCPV
not
improves
I-red
dis-
Of twO shOuld
of
gas channeling
is fmmwater
is about
of a miscible
100%.4
permeability
stratum.
sweep
water,
because
mobility
the high-permeabtity
of reduced
10-md
stratum
that
with
areal. sweep
dw-gas
WiOI mO~Jity
Sequence.
recovers
lower
that
same
good
areal
approach
(i.e.,
model
of severe
from
~je~~
and producer.
the
The
The
pattern
should
sequence
gas cycling
achieved
condensate.
to 85%,3
=0.1,
that
from
by the
3D model
of an inverted
injector
displacements.
on a five-spot
of the
tie
with
shown
in the y dimension
one-eighti
and WAG
mobility
top
is no cap!lary
The
were
Re-
the bas~
improvement
to that
at 1 Hcpv
ft between
conditions
grids.
from
stu~y.
gas injection
of
recove~
simik
3 cells
simulated
with
simula-
by subdividm~
m those
gas injection.
as the 2D model,
of vertical-
six e+i-tiicknms
in rwOveV
of continuous
of tive
identical
A 3D
into
Pr*
into
almost
number
we conducted
instead
continuous
water
pro-
.gids
vertically
were
of the
effect
prediction,
vertical
Coarsening-Upward
As expected,
the
over
cause
Fz.
test
override
Model.
the
as those
water
nine
the model
permea.bil-
the
with
hysteresis
but, the
WAG
ratio
60!~u0m
the same
and
pmperdes
made
3D
reach
for
To
on gas
10-md-strata
top
sults
upw~d
fOr tie
pressnre.
when
end
using
mm
placement
strata.
8 compares
cases,
tion
twice
com-
curve
and relative
earlier
grids
of favorable
satura-
were
adjusted
two
W=
capillary
fq both
the
pressure
are essentially
for its
pressure
other
saturations
Model
duection
Both
recovery
the residual
).
wa.i assigned
stratum
the capilhy
the
A mapped-gas
slightly,
total
layers
mcdel
residual
was
caMs.
stratum
the crews
where
layer
individual-stratum
The
breaks
duction
the
four
appropriate
stratum.
curves
pressure
lists
a liited
with
were
assumed.
in the 2D
Each
saturations
and WAG
a coarsening-
pressure
permeabtity
7 compqes
a model
an
results.
Fig.
model,
sequence,
and
and a capilkmy
3
rela-
variations
runs
base-case
recoiev
the
we compared
model
a fine-gfid
saturations.
10-md
residual
model,
several
Model.
Table
@e.
water
(No
properties
ihi
er
Free-Grid
stra-
from
we used
cases.
WAG
with
change.)
produced.
low
(the four-permeability
the
permeability
fi awaY
sweep,
a coarsening-downward
summdzes
resid~d
stra-
l-red
results,
hysteresis
occurs
the
has been
to the base-case
gas
case,
permeability.
340
no water
of effective
IndiniduaWfralum-Property
a relative
is SW
model,
vertical-crossflow
with
l-red
displacement
from
in
viscosiv.
individual-stratum-propew
upward
water
and flat
is the result
and
Cases.
with
stratum
nms
pressure.
use of indi~iduaf-stratum
the
sensitivi~
in WAG.
I-red
the bottom
recovery
waterfront
injection
permeability,
Sensititi@
the
into
A stable
consequently,
at 1 HCPV
WAG
bypasses
downward
stratum.
stra-
significantly.
5 shows
Its slow
tive
100-md
stratunu
high-permeabfity
completely
crossflow
is improved
producer
tie
cases.
through
reservoir
capilfay
2,0
4.3
recovwwmy
sweePs
,.-.?.
1NJSC70N
0.8
0.6
WLATNE
04
=RMwlmv
0,2
E.
00
0,2
0.4
WATER
0.6
0,8
SATURATION
Pig.
2Base.case
relative
permeability
curve.
Fig.
3Three-permeability
model
layer
recoveries.
.,
SPE Reservoir
EI&er&
Augu~
1993
CONTNUOUS
Y/m
PRODWER
1NJEC70R
QA.S
m%
,80,
M
Fig.
4Original
tbe
t-red
er,
improving
gas
condensate
stratum
bottom
in addition
overall
to diverting
vertical
25%
WAG
compared
and
gas
for WAG).
case
100-md
case.
in-
At
100-md
witi
faster
front
with
is not invaded
depth,
continuous
this case
in the bottom
has
340
at
recovery
for
between
moves
water
lay-
stratum
and increases
difference
the
100-md
increasing
sweep
compaed
stratum
the
(61 % recoveV
A noticeable
1 HCPV,
the
moved
ft away
efficiently
to only
by gas or water
case.
fverticul
between
WAG.
stmta,
restricted
at 1 HCPV
stratum
Penneabil@).
reservoir
With
OHIP
into
the
mm.
Consequently,
bas
crossflow,
To prevent
five
cycles
was
continued
.abnost
water
(i.e.,
mntinuous
O.O4%
of HCPV),
cantly
when
The
process.
WAG
as
WAG
also
projects.
yam,
an
is recovered,
167
z 0.60
plus
pressore
The
Stone
The amount
OtiY
was
that water
46%
ery
relative
HCPV
Gas
for
after
injection
19.4
years,
compared
prnducdon
that yields
was
critical
10%
that
higher
with
reduced
from
of that
(approximately
movement
slowed
during
through
injected
the
the 10@md
in the
operation
and to
Laboratory
fommnce
gas
effects
relative
,cecovery
uldmate
gas recov-
after
reported
psia.
0.30.
condensate-liquid
h addition,
of the blowdown
in a layered
ed a core of two
a 10-red
12-in. -long
air permeability,
a 0.3-md
core
to that
WAG
by Hawes
demonstrated
air permeabili~,
that
in the simulations.
hemicylindricd
267.
porosity
11.3%
TABLE
outcrop
Texas
porosity
2SIMULATION
CASE
(gas
et al. 6
per:
construct-
ti,estone
cream
kdiana
WAG
We
slabs
ltiestone
and
limestone.
The
SENSITIVITIES
<2%
Recovery
(Oh
OH[p)
Recovery
WAG
Recovery
(oh)
(o/a OHIP)
case
at 1 HCPV
Baze
case
61
78
2s
Coarsening
upward
3s
54
42
Coarsening
downward
62
78
25
crossflow,
= 0.1
atSw
3D, one.eighth
spot pattern
Four-permeability
layer model
80
Incremental
WAG
Gas
smtum
injection
100
*D
was
hysteresis
w25 analogous
kw=l
40
done,
WAG
PRE2SURE
for tbr=-phase
saturation
improved
The
YO1% at 5,981
Coreflood
floods
kv/kH
P.:
it the surface.
=40
to account
an EOS
dewpoint
included).
Limited
of
signifi-
injected.
CAPILLARY
used
of
we used
=6,170-psi
Continuous
of gas channeling
65%
with
at the end
Case
was
15 years
produced
significantly
15
resemoir
or
effect,
dropout
WAG
stopped.
improved
cycling
gas for
bas~case
of blowdowi
fluid
condensate
permeability
not
the
= 175 bbllMMscf
liquid
SI models
Laboratory
injection
injected).
injection
for a reservoir
shows
in older
we cycled
stra-
recovered.
water
were
using
25 years
25 % over b[owdown
was
WAG,
of blowdown.
a mzximum
10
l-red
l-red
0.2
Fig.
with
location.
Blowdown
to blowdown
for liquid<ondensate-dropout
at 175P
has
water-front
Early
injeqted,
by either
description
fluid
to
blowdown
10 years
To account
model
model
an alternative
HCPV
folfowed
WAG
by about
preferen-
the
im.prOvement
water
of water
injection
profde
after
water
HCPV
a 54%
Akbougb
bbl
indicadng
water
injection
flows
into
OHIP
stopped
HCPV
At
water
419S
only
be
To compare
model,
layered
Alternative
may
or
with
54%
the bottom
goes
through
we
at =0.175
only
little
only
production,
gas injection.
WAG,
and
with
for 23 yeizrs.
71 % OHIP
completely,
with
is restrict-
recovers
recove~tlom
breaks
.Wa=I
cros$ow
a dramatic
gas injection
stratum
injected,
verdcal
improves
Lhtle
however,
100-md
been
ff
recovery
injection.
is obtiinet
tially
water
5Three-permeaMhw
permeability;
Crossflow
ed
vertical
injection
the producer,
in this
1 HCPV
recovery.
Whh
stratum,
ft from
and
permeability
improves
with
200
gas from
sweep
Sequence.
of the model
at 1 HCPV,
at
W.
Coarsening-Doynwmd
the
remaining
At
HCPV
46
71
54
61
SI
33
61
75
23
45
72
60
of five-
Iniected.
,,
D
k
TABLE
3LAYER
RESIDUAL
.D~
o
0,2
04
WATER
(4=
1294)
6.Capiilary
pressure
curve
for
lo-red
7
!.,
0.s
SATURATION
.-..1
Individual-1
Fig.
SATURATIONS
.20
rock
layer.
ayer-property
3
model
,,,
...
,..
1
,.
84
Fig.
7-Gas-condensate
10
10WAG
as
10
m.
W.*
......$:
f=)
mGEkLwER
k L4YER
03
,.
40
alternative.
CORE
cE.4aAcrEmar1c3
k L.u%R
30
a blowdown
recovew.
BIG+
1.
p,:,
Fin.
BWIWKWN
WAG
1-+1
1:
CORE
FiOLDEX
m
FIUSXJRE
TRANSDUCER
GAS
GAsWAma
:
SER411ATOR
C~x
C3
C4
Fig.
8-Cumulative
water
Fig.
production.
11 Coreflood
apparatus
schi!kailc.
. ... . .. .
.,
m
!4
II
m
0,s
WAG
,,7
0.,
WAG
RECOVERY,
MK.lloll
m?
-?*
0.5
17
c.3N.Nu.0%
m
,.,
--
.x2
,,03
,.10
Fig.
9Resewoir
slabs
were
placed
The
10-md
layer
in a rubber
end cross
filled
eDoxv
with
coretio~
Fluids
sate,
used
sleeve
29%
section.
to form
of the
The
comparison.
a 2-in,
PV.
interface
inlet.
Fig.
were
between
Fig.
5%
mately
the
same
was
brine,
propane
% butaoe
ratios
mix
at 1,500
as those
for
fuat-contact
the
psia
11 ~so
and
simulation
miscible.
The
lioear
velocity
of = 2 tVD,
velocity
depended
on the
degree
of channeling
rate
12-WAG
ayered
was
is a schematic
effect
and found
gas
to represent
at a nominal
210
the layera
to represent
ducted
the injection
~g.
core
11 is a schemmi[
in the
vs.
cora,
Texas
that
dry
gas.
150F
were
fluid
(Table
coreflood
although
conden-
1); gas
were
acmd
or bypassing.
gas floods
Fluid
approxicon-
velocities
connate
water
relative
permeability
conducted
and
on
achieved
and
28%
essentially
the same
bre~ough
Endpoint
for trapped
gas
for
HCPV
saturations
and
measured
measured
from
90 + % recovery
at 0,65
10 ftlD.
were
saturation
recovery
bottom.
= 0.1
for
the
the
fndkma
for linear
of 25 % for
endpoint
Texaa
cream.
liiestone
water
The
was
29.5%.
Initial
fluid
We
layer
of 0.05,2,
trapped-gas
eontinuOus_gaS-injectiOn
Io-md
cream
at 1.2 HCPV
and a 95 % ethane/5
and-viscosity
tested
-diameter
-.
pmwm
e:
app&atus.
density
displacement
pressure
contains
of the core
of the
.4.
.WfAnr
+.-
QAS
Iilfing
25
PV
fluid
samrafiom
the core
of mineral
with
in the
brine,
oil.
Water
then
layered
flushing
saturations
SPE
Reservoir
core
out
after
were
established
the brine
the
Engineering,
01
by
with
up to
flush
were
Angnst
WY3
m
5+
.
WA,
,.7
WAG
,0
*.17.
,,0
OIUGUWL
69
FLum
IN
,,
RECOVEW,
mA&lOh
0..
&
EFFmFNT
CONTINUOUS
GAS
,..
40
,,
,.,
--
10
,.,
~J
.40..50.3112
Fig.
0.0
13Effluent-concentration
Io-md
layer
=27
% of total
core
until
on
Soltrolw
solvent
folowed
floods:
a continuous
a WAG
flood
witi
the
cycles
weter
by
continuous
the
layer
HCPV
Fluids
iayeied
Fig.
Core-
produced
oil was
from
14WAG
We
conducted
four
ex-
the
layer
on
10-md
the 10 md-layer
on the top,
The
with
at a constant
layer
on
0.s
HCRV
i.
PRODUCE.
injection
iecove~fror
top.
bottom,
a WAG
floods
alternating
produced
on.the
and
WAG
0.6
continuous-gas.
10-md
with
with
layer
,N,WTION
0..
vs.
core,
.3As
the
flushed
gasflood
gas slug
were
not
propane.
with
10-rnd
was
mineral
on the bottom.
of a 0.05
slug.
The
gasflood
with
10-md
from
0,,
-wA~wA~R-
F8A~10NmTAL,
layered
water
~je.ction.
the top,
PRODUCED
history
addkional
WAG
HCPV
bottom.
perimental
tion
0.,
TUTAL
FR.4CTION
floud,
COwmws
flood
used
a 0.05
iajecHCPV
1,500-psi
back-
s.iK
pressure
Experimental
Rusufts.
recuvered
46%
With
more
OHIP
the 10-md
than
layer
on the bottom,
continuous
gas
L_12221
,
WAG
injection
recov-
U,4R60SI
ered
produced).
Fig.
12 shows
propane
recoveries
XNMXtON
vs.
total
Fig.
fluid
produced
13 shows
tion decreases
with
mix)
WAG
began
and
.-s
dramatically
WAG
have
core
layers.
diverted
gas
and
history;
continuous.
gw
WAG
shows
flood.
after
was
stopped
The
floud
bre~ough
along
Fig.
cOmP@f
that
tie joint
the intervals
Water
0.34
was
HCPV
at 0:54
water
HCPV
moduced
when
produced
water
first
had been
produced,
=0.1
was
the
itieckd.
or after
1 HCPV
HCPV
total
Water
eight
water
during
totaf
pro-
~j@On
Whb
the
10-md
the recoveries,
drops
The
duction
kick
prod&ion
cycles.
of the 0.39
HCPV
in the effluent
did
water
tutd
=0.65
begin
iqiected.
water
layer
there
After
1.2 HCPV
produced
total
when
of Experiments.
the same
bottom.
produced.
enceontbe.
interface
flood.
pending
Whh
cases
on whether
the 10-md
effectively
was
between
diverts
slightly
is on the
if tie
10-md
The
between
layer
on the bottom,
along
the
is on the
recuvery
0.3-md
equal
layer.
infln-
basis
models,
floods,
or tie
higher
Wkh
layer
processes,
then
and
the 0.3-md
lay-
itdoes
than
is on
is completely
top,
layer,
volume
layer
top.
swept
recovery
the
the
of crossflow
when
fills more
10-red
10-md
and
of our
siimdation
we can estimate
fluid
voir.
in both
improvement
when
the
10-md
respectively.
results
the WAG
from
five
process
different
reservoir
sensitivity
to reservoir
properties.
Pohasity
In our
cogmmnly
studies,
a lower
target
tie
&ata
for
WAG
correlates
thepomsi~of
porusity
Pemneabifity
detop.
WAG
contrast,
higher
because
voir
cremental
The
both
severe
can
contrast,
of
the interface.
bottom
is much
10The
SWWP
WAG
later
WAG
and
enters
because
Sensitivities
tluencm
recoveries
subsequent
gas channeled
top
the
propane.
is a potential
almost
was
the
layer
bottum
core
to permeability
a.lllayers
in the low-penneabitity
would
have
been
in a reser-
is tie
same.
somlum,
If we
the recOv-
lower.
,when
than
the water
the
the low-permeability
initially
as well
occurs
when
and
from
Porosity.
water
at 1.2 HCPV
earlier
lighter
there
difference
the
layer
even
that
was
recovery
the layers
dine.
indicated
is a significant
=58%
layer
that
layered
or
production,
HCPV
floods
10-md
ethaneis
gas-breaktbmugh
gas floods
thematrixinboth
There
had
because
high-permeabilig
in the two
the
breakthrough
wasontop
continuous-gas
whether
Bodugastloods
Etime
mdlayer
The
recoveries
layer
continuous-gas
had used
Dikussi?n
because
O.3-md
layer
ery
or tie
tie
OD the
pro-
injected.
ally
water
layer
breakthrough
er is on the bottom
Process
gasflood
produced,
of the 0.63
however,
the low-permeability
Water
from
top.
in the
WAG
HCPV
on
by
is a propane
in the
0.79
history
layer
ontbetup,
sweeps
tbe
14 compares
Cuncentmtion
HCPV
until
had been
recuve~
Fig.
whereas
at
not
improved
produced).
propane
breakthrough,
beginning
HCPV
HCPV
WAG
15 compzes
propane
after
Water
=0.40
on the top,
(at 1.2 HCPV
and Fig.
rapidly
0.16
layer
gas injecti&
effluents.
Io-md
If we assume
WAG
irijekd:
2g % over
15Eff[uent-concentration
flood,
tbe gas
between
injected
at 0.51
l.1
PRODUCED
produc-
inje~On
indicates
by channeling
ECPV
floods.
propane
imbibition.
12 also
water
with
The rapid
cases.
been
the WAG
duced,
the continuous
concentmtion
propane
pmducdon.
DW gas (i. e., the ethaneibutane
to be produced
at = 0.15 HCPV
in both the continuous
may
Fig.
for
the effluent
TOTAL
is less
gas
cont+wt
strongly
higher
permeability
with
recove~
of irjected
effectively
witl
WAG
can be expected
gaS in a highly
by water
channeling
stmdOed
injection.
and
thus
in-
With
resersmaller
a Iower
WAG
in-
a marked
effect
recovery.
direction
of permeabilky
the continuous-gas
meabfity
Generafly,
incremental
channeling
be reduced
there
Permeability
Contrast.
performance.
water
to swq
lower
10-md
stratum,
water
strata.
layering
and WAG
When
undersides
gas severdy
the Iiighest
und
has
processes.
diverts
~ltb
overrides
Per-
and water
tends
permeabfity
gas
on
the. &he$t
to higher
is ih the lowest
layers.
211
Relative
Water
meabdity
Pemtwability.
endpoint,
:CY. when
We
used
a
instead of 0.1 at residual
was
slightly
23 years
higlier
was
more
at the producer
gravity
had
a small
of water
effect
relative
OD WAG
perIecOv-
relative
permeabtity
endpoint
of 1.0
gas saturation,
WAG incremental
recovery
at 1 HCPV
through
curves
ordy
water
the same.
exhibits
bility
The magnitude
has
km,
injection,
With
crossflow,
km,
The
effect
recovery
water
has higher
earlier.
little
but tddmate
larger
is more
injectivigf,
shape
100-md
Gas
Saturation.
of residud
or trapwd-gas
been
tht
Iower
WAG
effect
in 0.24-darcy
with
saturation
our basmase
result
because
front
in the more-permeable
recovery
would
have
by water,
28%,
a higher
have been
strata
been
depends
WAG
from
watertlooded
S,g h~
recovery
behind
and waterflood
fdgher
If the
S,!.
&apped
et al.s
sis,
Densities
those
used
recovery
Viscosities.
in the base
because
fwtively.
wilf
and
water
Viscosity
block
cosity
the
ratio
conversely,
case,
we
will
ratios
expect
strata.
compensate
intluence
would
result
a low
viscosi~
incremental
would
ef-
of dry
WAG
condensate
models
ance
water
to inject
the
and
trast
between
to benefit
gas;
segregation,
incremental
contrast
strata,
between
Pvs,
a WAG
sitive
ratio
of
low
ratio
would
de-
vertical
permeability
con-
1:1
appeam
1:1
gas
for
Discussion
ventional
and
showed
that
tical
dering
gas
affect
results7
from
cotinned
this.
show
that
layer,
tending
ability
layer
Sorbic
gas
frontal
in singfe-phase
in proportion
served
to kh.
in our
in all strata
the
displacing
simukuiorq
i.e.,
and
water
Soucemarianadin
et al 9 reported
practice
waker
water
Iock
reduce
blowdown
needed
et al.
stage.
to remobilize
and
Fishlock
aPPmxkatelY
PV or more
with
Chierici
et il.
thre.$holdfor
trapped
pressure
s2
costs,
improve
releasing
reservoir
and lowered
injectivity.
injection
recovety
after
Our
and
with
the WAG
gas
blowdown
is occupied
water
samration
study
low,
space
in conto trapPed
injectivity
process.
is Iarger
after
by water.
S =
or-
SiW
eadier
L
L2
sm+mtion,
fraction
irreducible
residual
y = viscosity,
p =
density,
.$ =
porosity,
tit
m/L3
fraction
work
gas remobtition
to
a front
during
WAG.
hysteresis
that
indicated
when
gas
=
horizontal
~ti~
injected
~ = re$idu~
ob-
moved
&=
whereas
rm.mvoir
fluid
V = vertic2J
w = water
strata.
from
laboratov
a WAG
after
there
water
process,
occur
it
tie
increase
Both
blowdown
mm
was
waterflood
no
injection
gas
th~
Mobfl
J.A.
Bamhart
R&D
COIP.
for
for permission
conducting
the
to publish
laboratory
this
work
experiments.
References
is
Hawes
after
were reached
with 2
they did not study gas
the
We
and
et al.
during
saturation
blowdown.
their
Acknowledgments
Fkh-
to ffawes
would
a large
in ga$)water.
et al. 6 and
According
gas during
maintained
layers
those
mobifity,
fhe hysteresis
tie
conducted
g=
immis-
the core
results
Subscripts
nnit-mobility
similar
by Hawes
with
fiat
per-
media.
exists
showed
et al.
swept
arc
true residual
gas saturations
of water
injection.
However,
injection
212
work,
They
less tijected
could
of reserves
water
reservoir
thickness,
of different
two-phase
sitqifai
repmted
hysteresis
et al.s
witi
WAG
loss
production
more
is sen-
gas saturation,
liquids.
primarfly
&-y-gas
permeability,
to crossflow
For
developed
gas recove~
Fishfock
because
trapped
condensate
of potential
after
and layer
core.
displacement
porous
curves
et al. lo If this
would
and
studies
permeabiliw
total
gmviy
performance
a gas-condensate
production,
restored
perform-
compression
more
into
ffiat water
condensate;
WAG
Therefore,
by reducing
because
in the highest-permeabili~
in multilayered
Oux simulation
relative
gas
to gas mobility,
WAG
more
irjecti&.
recovering
on water
stable,
results
of its favorable
preferentially
layers
fluid
er al.s
and
heterogeneous
sortie
because
relative
fluid properties,
is not injected
in the lower-perme-
between
miscible
in
displacement,
to superior
ga.s-
contbmous
g:
ver-
confirm
leads
with
contributing
simhicmtly
the mo-
higher-permeabtity
advancement
crossflow
displacements
miscible
mobiIity
recovery
dkplace-
stratified
displacements.
et al. 8 repoti
meabilities
cible
contrast
to an adjacent
aids
studies
Our gasflcods
ftom
that
when
Our
out of
as long
work%e.?
imbibition
Permeability
layer
in miscible
Early
systems.
an unfavorable
ta reduce
inecbadsms
recove~.
displacement
si@tic2ntly.
the Iow-permeability
identified
capilfary
in stratified
the process
that
from
in ah immiscible
is favorable
displacing
to higher
resulting
efficiency
ratio
studies
lead
crossflow
sweep
bility
laboratory
and
hystere
recovery
28 % to 54%
layering,
however,
WAG
in WAG
Fish-
gas
coreflood
recovery
permeability.
substantially
sale,
shows,
is Pre-
factors
continuous
is essentially
simulation
inchrded
gas
and
and
of trapped
from
permeabili~-
5. Ultimate
Our
that
laboratory
with
ranged
recovers
gas saturation,
pefmeabflity
1.?s than
gas than
and
watermobility
to reservoir
4. Water
ferred.
opemte
et al
liited.
gas-condensate
and relative
economics
needed
water
or large
of much
of
vertical
of approximately
crossflow
a WAG
amount
nomtalfy
amount
relatively
restricted
layers,
optimum
We
minhmiin
With
with
The
strata pertneabihies,
layer
the
recovery.
recovery;
Ratios.
on resemoir
to inject
good
and
individual
process
to achieve
Sizes
depends
permeabiliv,
sign
SIug
process.
investigated.
favorable
are
3. WAG
Gas
was
WAG
in Hawes
ultimate
experi-
no ~eshold
water-invaded
percolation
limit
compmd
dominant
vis-
and
and
not
uncon-
hysteresis
fn our
coks,
simulations
to simulations
improves
reservoirs
2. The
recovery
Water
core.
partially
than
that
did
Improvement
lation
water
water-to-gas
reduce
their
volume
WAG
irjection.
more
diversion
lower
concluded
regions
1. &cordmg
than
injected
A ~gh
effective
ratio
WAG
effectively
strata.
in more
larger
how
high-permeahilby
dfierences
higher
in 1.28-darcy
smaller
the constant-pressure
much
From
4- to 13-darcY
a much
was oidy
the hystere-
Conclusions
gas-condensate
the low-permeability
At density
during
was
et al.
with
low-pemneabtity
the reservoir
as injected-water
would
than
that
in lower-permeability
the water
ments,
Fluid
core
packs.
Hawes
compared
relatively
permeability
indicated
gas is low
et al. 6 r~ported
observed
studi.i-s,
on the amount
cores
layered,
were
et af. 10 also
to mobtie
Hawes
ments
the
recoveries.
recovery
sands.
fn these
permea-
Ffihlock
needed
consolidated
solidated
lock
Trapped
depletion.
sis in saturation
effects
and breaks
of the relative
on predicted
after
mObiie,
during
phase.
saturation
was continued
1.
Harper,
I.L.
et
al.:
Composid.oml
Sitdator
for
Performing
l-age
13714
and
4s3.
SPE Resemoir
Engineering,
August
1993
3.
Smlkup
son,
4.
Craig,
(1983)
.%ies,
St...
6.
Hawes,
Data,
R.I.
et
Condensate
8.
Yokoyama,
Y.
al the
tion,
Antonio,
San
K. S.,
menrs
in
paper
SPE
ference
9.
Mediz
tion,
10,
R. S.,
and
Dallas,
19S7
Sept.
Sweep
stable
SPE
and
CapMary
Pressure
pap
SPE
and
Recovery
SPE
on
10109
Exbibi.
O1-Displacemmt
E.x@-
Mechanisms,
Annual
Technical
Culllck
Lu
A.S.
in Multilayered
Unstable
Flow,
Technical
Porous
paper
C.mfcrence
SPE
and
Culllck
ing
16955
ing
%+erimental
Studies
on the Waledlced
SPERE
by Blowdown,
Residual
@lay
in
Metric
Conversion
bbl
ff
ft3
4.046873
E01
ha
1.589873
E01
m3
3.C4S*
E01
2.831685
E02
~F-32)/l.8
m3
OC
Cohen
hysicsl
chemistry
ngimser
>ined
in
2.54*
E+OO
cm
ombustion
md
9.869233
E04
pm2
ons.
psi
6.894757
E+OO
kPa
in.
1980
bandonment.
Woversio
factor
original
mnuscrlpt
sql,
at
SPE
3,1992.
the
i 989
t,
exacl
received
PaFaracmPted
SPE
G=
for
Technology
fcr
rev$ew
Pukllcathm
SYnwas(um
June
Cct
7,1989.
5,1992
held
Ravised
Pa$ar(SPE
in
Dallas
mmmrlpt
191
J...
7-9.
14)
ultsnt,
received
M
both
retired
. >ined
all
and
from
Transient
in
is
and
an
U.
was
He
holds
Tech.
Photo
unavailable.
and
PhD
in 1992.
and
degrees
Jones
wss
a PhD
degree
biographical
and
petroleum
studies.
1970-71.
in applied
information
He
en-
engineering,
of
the
during
and
a coninterests
in petroleum
in chemical
MEPTEC
meas-
Jones,
principal
a member
development
applica-
steamflood
LG.
His
He
in-situ
rate
sand-control
during
at
on
well
civil
in
staff
MEPTEC.
and
degree
chairman
adviser
at
and
in
a BS
in chemdegrse
research
degrees
holds
Develop-
is a senior
quality
CommOtee
simulation
f.u
horizontal
Technical
engineering
reservoir
1980.
are
and
Stste
Testing
1986-89
hen
MS
Ohio
He
a PhD
U. of Wyoming.
MEPTEC
in 1959.
U. and
modeling,
MS
the
Shell
Group
and
flow
Cullick
degree
in steam
and-completion
Mobil
gineering
are
fe-
and
for
U. H.S.
gas
charac-
BS
conducted
projects
holds
fmm
from
re well-testing
presented
State
flowline
Lu
ngineering,
.smm
reservoir
holdsa
Tulane
hss
studies
and
current
assessment,
Recovwy
and
steam
He
Ohio
steam
current
urements,
from
Thermal
In
and
His
istry
born
the
Mobil
Co.
on
enriched-gss
His
worked
ment
join-
worked
behavior,
are
risk
Since
has
scaleup/averaging.
previously
acre
phase
interests
attribute
Factors
he
projeots.
terization,
SI
Technical
in Dallas.
1981,
fluid
search
engineer-
E.SP
fiooding,
production
387
associate
Mobil
miscible
fiooding,
1988)
at
(MEPTEC)
Mobil
C02
Its Pmducdon
is an
adviser
Center
Exhibi-
27-30.
ad
Jones
Con-
27-30.
Efficiency
Annual
Gas-
20-22.
Conference
of
19S7
Media;
Technical
Analysis
.,
..-=
~
LLiL
and
53.
Wateflmdhg
of
Porous
at the
etaf.:
et aL:
T.P.
Saturation
4,
Oct.
T.C.:
Mono-
PemfeabiliN
No.
of
Effects
Rowe,
Core
Wlween
Sept.
The
Annual
5-7.
presented
at the
FMhxk,
Gas
SPE
Oct.
A.
Dallas,
.%Jdies
in StadfIed
Exhibition,
Conuast
prewnted
12,
London,
L.:
Lake,
Wat,
Soucemarianadin,
I@hard-
50.
1981
16706
SPE,
Wate@mding,
Relative
Tech.,
Feasibility
Heterogeneous
and
Pet.
Conference,
and
of
(1971)3,
Three-Phase
C&.
Displacements
presented
Sorbic,
al.:
Pcmleurn
Immiscible
Series,
Aspects
TX
of
.?.
Reservoirs,
European
7.
En@neenng
Richardson,
,Estimation
Oil
Monogmph
44-46.
SPE,
H. L.:
Residual
Di@ac?ment,
8;
Reserwir
F. F.:
graph
5.
Miscible
F.I.:
TX
Pressure
1983-85
Martin
and
has
and
F.
worked
applications
mechanics
for
J.P.
Co.
since
from
Watson
Cd
FRmC~ANDIOG
Geometric
Average
Permeability
(md~
Res. C
w
44
112
Permeability
Strata
!iigh Perm,
z of
_TStal
Thicknes
~v,
B
&
9
8
Permeability
Ratio to
COSMOS
Lo
e rm Layer
MQ!M
&p
e
8
80
132
COSMOS
100
Strata
Moderate
Strata
Perm.
9.7
69
69
69
13
11
10
0.55
0.85
22
23
23
(<30 md)
Low Perm.
Strata
6.913
(<2 md)
continuous
Gas
Caae
Baae Caae
Regressive
Layering
VAG
VAGIncremental
MCQYSKx~
61
70
28%
38
54
422
62
78
25%
46
71
54%
61
81
33Z
61
75
23Z
Layering
Krw
Cross-Flow
1 at Srg
All
recoveries
at 1 HCPVinjected.
Recovery
,,.
!.
SeE 19114
.
.
INJECTOR
I
~~RMEABIL\Ty
Ml
1
PRODUCER
A
4351+.,
12
THICKNESS
feet
15
44
0.55
20
4.8
24
4.8
23
~=o
FIGURE t
feetl
FOUR-PERMEABILITY
0.6 EcovERy
0.5 FRACTION OHIP
0.4 0.3 0.2 0.1 -
0.0
0.2
ox
0.4
0.8
to
1,2
HCPV INJECTED
FIGURE 2
WAG
INJECTOR
STRATA MODEL
CONTINUOUS GAS
PRODUCER INJECTOR
HIG
L
PRODUCER
81-100%
~lq
11-5f!~
51-80%
IJ
0-10%
FIGURE 3: INJECTEO+AS
CONCENTRATION AT 1 HCPV INJECTED FOR THE
FOUR-PERMEABILITY STRATA MODEL
6s7
..
INJECTION
WELL
PERMEABILITY
-i4-
34.5
10 MD
34.5
10 MD
100 MD
11.5
1 MD
11.5
1 MC
-. 1870
z-
feet
FIGURE 4: THREE-PERMEABILITY
STRATA MODEL
1.0
1 HCPV INJECTION
0.9
0.8
0.7
WAG
.0.6
RECOVERY
0*5
FRACTION OHIP
0.4
1 HCPV
INITIAL WAG
WATER INJECTION
CONTINUOUS
NJECTION
GAS
0.3
0.2
0.1
/
0.0
0.0
_,
4.0
8,0
12.0
16.0
20.0
24.0
TIME (YEARS)
FIGURE 5: RECOVERIES FROM THE THREE-PERMEABILITY
STRATA MODEL
.,-
CONTINUOUS
0-
GAS
1911
WAG
10%
11-
50%
INJECTOR
PRODUCER
T
3600
DAYS
5400 DAYS
Tr
00 DAYS
HIGH
PERM
LAYER
m
m
5s9
2400
t
2200
F
2000 PRESSURE,
PSIA
1800 -
1600 -
ti
1400 -
1200
o
12
16
20
24
TIME (YEARS)
FIGURE8:
100
90
80 .
~~
70
CONDENSATE
RECOVERY
(% OCIP)
1:1 WAG
BLOWDOWN
5183 psia
60
50
40
~1
GAS CYCLING
1156 psia
30
.BLOWDGWN
20
10
0
o
10
20
30
40
TIME (YEARS)
FIGURE9:
CASE
:.. ..
... .,...,.
CORE
HIGH k LAYER
26%
11.3%
28%
29.8%
.i!iii!f
C2
/ C4
AP TRANSDUCER
RIN
c~
LOW k LAYER
0.3
10
.
a
0:4
L
I
1.5
1
u
GASI WATER
SEPARATOR
0.9
t
0.8 WAG
0,7 0.6 RECORERY,
FRACTION 0.5 OHIP
0.4 -
CONTINUOUS
0.3 -1 -n-
GAS
m-u
0.2 -
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Ml
100
90 80 70 60 MOLE
lX$t4:;
EFFLUENT
50 40 CONTINUOUS
GAS
30 20 10 0
0.0
I
0.2
[
0.4
I
0.6
1
0.s
1
1.0
1
1.2
0.9
0.8 I
WAG
0.7 -
0,6 RECOVERY,
FRACTION
OHIP
.~
0.5 -
CONTINUOUS GAS
0.4 0.3 0.2 -
1-11.-0
0.1
0.0 ~-0.0
0.2
0.4
H
MH*H
WAG WATER
INJECTION
0.6
+-l
0.8
1.0
1.2
100
90 -
so70 60 MOLE %
ORIGINAL
FLUID IN
EFFLUENT
50 40 -
CONTINUOUS GAS
30 20 10 0
0.0
1
0.2
t
1
I
1
0.4
0.8
1.0
0.6
FRACTION TOTAL HCPV PRODUCED
I
1.2
SPE
=-Lf~~m#m-
SPE 19114
WAG Process Promises Improved Recovery From
Gas Condensate Reservoirs: Part n-Simulation
Sensitivity Studies and Lab Corefloods
by A.S.
Cullick,
H-S.
Lu, and
J,P,
Watson,
Mobil
R&D Corp.
SPE Members
ERRATA
CORRECTED
FI IIRE
WAG
CONTINUOUS GAS
m
m-
I$?2zl11- 5%
FIGURE 6: ORIGINAL GAS