Professional Documents
Culture Documents
Optimising Gas Lift
Optimising Gas Lift
lE
6200 lbrth Central Expressway
Dallas, Texas 75206
;;!&SPE
. .,
5150
Optimizing
Gas-Lift
Systems
By
J. David Redden, T. A. Glen Sherman, and Jack R. BlannS Exxon Production Research CO.,
Members SPE-AIME
American
institute
@Copyright 1974
of Mining, Metallurgical, and Petroleum
Engineers,
Inc.
This paper was prepared for the h9th Annual Fall Meeting of the Society of Petroleux
Engineers of AIME, to be held in Houston, Texas, Oct. 6-9, 197~+.Permission to cOPY ~.s
restricted to an abstract of not more than 300 words. Illustrations may rmt be copies.
The zbstract should cm?tain conspicuous acknowledg!nentof where and by whom the paper is
presented. Publication elsewhere after publication in the JOURNAL OF PETROLEUM TECHNOLOGY
or the SOCIETY OF PETROLEUM ENGINEERS JOU~TAL is usually granted upon request to the Editor
of the appr~priate journal provided agreement to give proper credit is made.
Discussion of this _oaDeris invited. Three copies ofany discussion should be sent to
meeting and, with the paper, may be c~nsidered for publication in one of the two SPE magazines.
ABSTRACT
SPE 5150
2.
SPE >1>[1
3.
4.
.Wu-u.,
submersible electric pump. Therefore, they
are not altered by the gas-lift optimization program. On the other hand, the continuoas flow gas-lift wells can be controlled by varying the injection gas rate
from the gas-lift manifolds.
To determine the optimum gas injection
rate for each well, the wells performance
as a function of its gas injection rate must
be determined. The computer program calculates a theoretical optimum based on the
last well test utilizing vertical flowing
pressure correlations. Various correlations
are available.~ A correlation based upon
well test data for Creoles Lake Maracaibo
operations was specially prepared for Creole
by Dr. Herald Winkler, Chairman of
Petroleum Engineering at Texas Tech in
Lubbock, Texas, This correlation was
determined to accurately represent Creoles
producing well behavior. Exxon USA chose
to use the more general 0rkiszewski2
correlation.
Figure 4 illustrates how flowing
pressure calculations are used by the program in determining a wells performance,
Well tests provide information on the flowing wellhead pressure, fluid production
rate, formation gas-liquid ratio, total
gas-liquid ratio, and the water-oil ratio.
This information, combined with the well
description a>d the fluid properties, permits calculation of a flowing bottom-hole
pressure, Knowing static bottom-hole
pressure, a productivity index can be calculated for the given producing conditions.
Assuming that this calculated productivity
index remains constant and the flowing
wellhead pressure, water-oil ratio, and
formation gas-liquid ratio also remain
constant over the operating range of
producing rates, performance curves similar
to the ones in Figures 1 and 2 can be construtted. These performance curves are the
mathematical models of continuous flow
gas-lift wells in the system.
Separator Stations: Separator station
capacity for handling oil, water and gas
may also prohibit a well from producing at
its unlimited optimum rate. Separator
stations have both continuous flow gas-lift
and other wells producing to them. To
determine the capacity of a station for gaslift wells, the production rate of the other
wells must be subtracted from the total
capacity of the station. If the capacity
of the separator station is insufficient
for the gas-lift wells to produce at their
unlimited rates, then these rates must be
reduced.
SPE 5150
, WST~S
I
Calculating the Optimum Gas Distribution
Using the performance curves which
describe the wells production characteristics, the program calculates for each
continuous flow gas-lift well the optimum
gas injection and fluid production rate;
i.e., the point on the curve in Figure 2
where the slope is equal to one. Thj.srate
assumes
no limitations on gas availability
or system production capacity.
However, a check is made to see if the
well should be receiving any gas-lift gas
at all. The routine in the program for
calculating gradient curves is limited to
total gas-liquid ratios less than 3000
CU ft/bbl. Therefore, if the input data
for a well specifies a total gas-liquid
ratio greater than 3000 cu ft/bbl, the
program will digress and calculate a
theoretical total ratio for the well based
upon a minimum flowing pressure gradient
attainable for the specified conditions.
Using this method, optimum injection gas
is then obtained by subtracting the
formation gas produced from the total gas
requirements determined as above. If the
calculated injection gas is negative
(i.e., formation gas-liquid ratio is
greater than the calculated total gasliquid ratio) the program assumes that the
well will flow naturally and no gas injection is required.
If the sum of the optimum gas requirements for all wells in the system exceeds
the system limitations, a priority ranking
of the wells is established and gas usage
is curtailed to save the most gas for the
least loss of oil. This is done by incrementally stepping down from the optimum
point on the revenue-expense curve
(Figure 2) for priority wells. That is,
a small decrease in production rate is
assumed (e.g., 10 B/D) and the performance curves of all wells are examined.
The well which will result in the maximum
savings in gas-lift gas (i.e., has the
flattest performance curve) is adjusted
LO the lower rate. The process is repeated each time examining all of the
wells, including the one just adjusted,
until the total gas injection requirement
falls within system limitations. This
represents optimum use of the available
gas. Appendix A illustrates the application of the gas-lift optimization procedure for a hypothetical ten-well field.
A simplified outline of Gas-Lift Optimization Logic is given in Appendix B.
FIELD RESULTS
Two major problems have caused the program to have only limited success to date.
These are (1) the requirement for good well
data, and (2) the computer time required to
perform the calculations for a large system.
However, the program has been successfully
used in Creoles Lake Maracaibo operations
in Venezuela in a group of thirzy wells.
By implementing the recommendations of the
program, Creole was able to increase oil
production by 2050 B/D with an increased
gas injection rate of only 1920 Mscf/D.
Creole is presently in the process of updating their computer files which store well
test data and well completion data. When
this data storage is complete, Creole plans
to use the program on a routine basis. They
have also made steps to reduce the computer
time required by the program by eliminating
portions of the program which were
unnecessary for their application. Another
step made to improve the computer time
required was to break the total field into
several smaller areas of supervisory
responsibility. This made a considerable
reduction in the size of the model re+uir.,d
to describe the field.
Exxon Company, USA has used the program
on a very limited basis but realized a
significant improvement in the wells which
were tested. Six wells were selected from
approximately 150 wells in an East Texas
field near Houston. The programs recommendations were implemented and a net
increase of 180 barrels of oil per day was
observed for an increase of only 145 Mscf/D
of gas-lift gas. The program has been made
available in three ~ther East Texas fields
but presently is not in use. Exxons plans
are to simplify the output and t en to use
9
the procedure on a larger scale.
CONCLUSIONS
Based on the observed field results,
the following conclusions are made.
1.
2.
3.
CDW
.-
<1 %f)
...4.
T. n. Rlmn17N
.
------
=
=
=
=
=
=
=
=
=
=
=
=
tubing diameter
depth
formation depth
operating gas-lift valve depth
producing water-oil ratio
productivity index
pressure
producing wellhead pressure
producing bottom-hole pressure
static bottom-hole pressure
production rate
oil density
ACKNOWLEDGMENTS
We express our appreciation to
Mr. Gordon Bystrom and Sr. Juan Bless of
Creole for their assistance in the development of the gas-lift optimization computer
program used in Venezuela$ and to
Mr. Terry Mayhill for his engineering
assistance in implementing a similar
computer program in Exxon Company, USAs
operations.
REFERENCES
.
1. Simmons, W. E.: Optimizing
Continuous Flow Gas Lift Wells, PetPolewn
Engineer (August, 1972) 46-48, and
(September, 1972) 68-72.
2. Orkiszewski, J.: Predicting
Two-Phase Pressure Drops in Vertical
Tubing, paper SPE 1546 presented d SPE
41st Annual Fall Meeting, Dallas, October
2-5, 1966.
3, Mayhill, T. D.: Simplified
Method for Gas-Lift Well Problem Identification and Diagnosis, paper SPE 5151
presented at SPE 49th Annual Fall Meeting,
Houston, October 6-9, 1974.
APPENDIX A
Example Optimization Problem
The results of utilizing the gas-lift
optimization calculations are illustrated
here for an example ten-well field. Figure 5
is a model of the example field containing
one compressor plant, one separator station,
and ten gas-lift wells, During normal opersting conditions, the compressor plant, which
has several individual compressors, has a
(WTTMT7TNC
/?AS T.T171SYST17MQ
...--...
.-...
----
----------
5.
Simplified Outline
of Gas-Lift Optimization Logic
Retrieve well test data, well descrip~~ons, and field producing sYstem
description from data cards and computer
data storage files.
b.
c.
b.
c.
2,
3* Productivity index.
J .~
Pws - Pwf
****,*. ,...
2,
(1)
D.
Sum the optimum gas injection requirements for all continuous-flow gas lift
wells supplied by a compressor plant.
2,
Q .**,..,**.*...*
(2)
a.
3. Calculate the gas injection rate
required at the new rate using a
--Az
6.
1.
4.
APPENDIX B
B.
A.
~Pl? %1<
u7n.
7
I
u,
@nm
nw m-mx7
LU.IWIJX4AV
3.
4*
E.
b.
c.
d,
e.
4
f.
F.
g.
b..
c1 En
JJ. JU
2.
If any compressor plant has insufficient gas available for the continuoua-flow gas-lift wells to operate
at their unlimited optimum rate,
calculate a reduced gaa injection
requirement in the following manner.
a,
urn
a.
b.
3.
4.
For each compressor plant, c3lculate the amount of gas being produced to it from continuous-flow
gas-lift wells (gas-lift gas plus
formation gas).
2.
OPTIMIZING G/
8
a.
b.
3.
4.
TABLE 1
(ZNERAL WELL DATA
35 API
2.992 in.
5500 ft (subsurface)
5000 ft (subsurface)
1400 psig
100 psig
0.8
$1.00/bbl
$0.25/Mcf
Oil Density
Tubing Inside Diameter
Depth of Producing Interval
Depth of operating Gas-Lift Valve
Static Bottom-Hole Pressure
Producing Wellhead Pressure
Gas Specific Gravity (air = 1.0)
Value of Oil
Cost of Gas-Lift Gas
TPBLE 2
EXAMJLEWELLT EST
Total
Liquid
J!l!Q-
Water
Oil
!UL!U!L
1000
1000
1000
700
600
500
500
500
500
500
1000
750
500
700
600
500
300
250
500
.500
6800
5600
25:
500
0
0
20:
250
0
0
1200
DATA
Formation
Gas-Oil
Ratio
CU
ft/bbl
Gas-Lift
Gas
_!!Q&-
500
500
500
500
500
500
500
500
550
600
200
200
200
200
200
200
200
200
200
200
2000
Calculated
ProcJ:J:jvity
B/D/psi
1.12
1.28
1.64
0.72
0.61
0.51
0.66
0.73
0.50
0.49
TABLE 3
OPTIMIZATION WITHOUT SYSTEM LIMITATIONS
.
Wel1
No.
lx
~-Lift
Total
Liquid
-w!_
Oil
Iw._
1013
1049
1098
695
608
530
527
532
519
506
1013
787
549
695
608
530
316
266
519
506
258
312
316
176
216
262
248
245
247
216
13
49
98
-5
3:
27
32
19
6
3:
16
16
19
6
7077
5789
2496
277
189
Gas
~
;;
49
-5
496
TABLE 4
OPTIMIZATION WITH COMPRESSOR CAPACITY LIMITATIONS
(2000 Mcf/O AVAILABLE)
Wel1
No,
1
2
3
4
5
6
;
1:
983
978
1048
695
608
520
497
512
509
486
983
734
524
695
608
520
298
256
509
486
160
160
240
176
216
237
197
214
222
168
-17
-22
48
-5
-17
-16
24
-5
2:
7;
2;
-2
6
-1:
-1:
6836
5613
1990
36
13
-10
TABLE5
OPTIMIZATIONWITH COWRESSORCAPACITY LIMITATIONS
(1000 Mcf/D AVAILP8LE)
Calculated Optimum Rate
Total
Gas-Lift
Liquid
Oil
Gas
J!L-
!w_
913
959
968
655
498
430
447
442
419
426
913
719
484
655
498
430
268
221
419
426
6157
5033
-!!!!m-
.JiL!L_
1
148
157
62
58
-87
-41
-32
-45
-102
-70
-53
-58
-81
-74
-87
-31
-16
-45
-102
-70
-32
-29
-81
-74
-193
-73
-19
-88
-139
-119
-52
-43
-138
-142
994
0643
-567
-1006
12;
181
112
TA8LE 6
OPTIMIZATION WITH SEPARATOR CAPACITY LIMITATIONS
(4000 BO/D MAXIMUM)
Wel1
No.
903
509
324
515
418
340
184
126
329
346
4737
3994
456
-97
-241
-176
-185
-182
-160
-116
-124
-171
-154
-200
-179
-113
-189
-170
-156
-97
-89
-170
-181
-2063
-1606
-1544
--
---
F
I
I
I
I
I
I
I
LIMITED
.---- --
I
l!
MCF/D
GAS-LIFT
$/DAY
GAS
INPUT
GAS
COMPRESSION
COST
Pd,,
GIVEN:
d,
FWo
P.
Pt[
~
------+----1
D
Rsx!QD
valve
I
Df
\\\
u&r&+
}/00
O.OO0.l
I
L
..
HIGHPRESS.
GAS
---
LOW?RESS.
GAS
---
GAS,OIL&
--- OIL*
WATER
OCOMPRESSORPLAN1
o GAs
LIFT
A GAS
A
INJECTION
J-J
I-----J
WELL
sE.pARA,oR~~A,,o~
?f&WJ%%U
WATER
. . ..
x$.o~
WELL
GL MANIFOLD
Pwf
Ws
= pw~.pwf
FLARE
=O
SEPARATOR
~1
5700
STATION
CAPACITY
BO/D
SEE TABLES 1
AND2 FOR WELL
DATA
WELL NO,
GAS LIFT
MANIFOLD
t1
LOW PRESSURE
COMPRESSOR
GAS
PLANT
b
5600 MCF/O
v
SALES, FUEL
REINFECTION,
ETC.
3600 MCFID
Fig.
5 -
Example
gas 1ift
system model.
CAPACITY