e
w
.
:
Societyof PetroleumEngineers ~
SPE 37001
Tubing Size Optimization
in Gas Depletion Drive Reservoirs
Ronald Gunawan, SPE, and George R. Dyer, SPE, Vito Indonesia
Copycght 19%
Soaety of Petroleum Engtneers, Inc
Typically, a new completion will start in the high pressure
system and move through the medium pressure system into the
low pressure system before depletion.
This paper was FC3PWedfor presentation at the 1996 SPE Asia PacfIc 011and Gas Conference
held In Adelalde, Austlalla, 2831 October 19S6
This paper was salec!ed for presentation by an SPE Prcqram Commttee following rewew of
mfonnakcmccwamed m an abstract submitted by the author(s) Contents of the paper have no!
been rv!awed by the SocIetY of Petroleum Engineers and are subeci to correctmn by the
author(s) The mated
as presented, dces not necaa$anly reflect any postmn of the SocIety of
Petroleum Engineers, Its cffcafs, or members Pawrs presented at SPE meebngs are sub)ect
to pubkalm revmw by Edkwal Commflaes of the SocIatY of Petroleum Engtne%rs Perm Issfon
to copy IS restricted to an abstract of not more than 300 words Illuslratlons maY not be copied
Tfw abstract should contain cnnspmous acknowledgment of where and by whom the paper IS
wsenled Mite Lix+nan, SPE, P O Box 833WS Rd?ardson, TX 7Soe33+33S U S A, fax 01214.952.9436
Tubing size selection is an important factor in completion design
that affects well performance and ultimate reserve recovery.
Producing a well at the maximum rate without affecting the
reserve recovery is desirable. Increasing the tubing size will
usually increase the rate, however, the question becomes how
does tubing size affect reserve recovery. A general tubing
selection guideline is required for optimum tubing size selection.
That guideline is also required for inventory control and
planning.
Abstract
Proper tubing size selection is essential to maximize economic
reserve recovery in depletion drive gas reservoirs. For wells
completed in multiple reservoirs with a wide range of reservoir
properties, tubing size selection can become quite complex.
This paper presents an approach used to determine optimum
htbing size using a model developed by applying Nodal Analysis
and Gas LoadUp Technology. A database consisting of 340 gas
wells was analyzed and used to confirm the validity of the
approach and develop a model. The model shows the relationship between Reservoir Abandonment Pressure (Pa), Permeability Thickness (kb), Tubing Size, and Flowing Wellhead Pressure.
The model shows that for high kh (Permeability Thickness)
completions, tubing size has only a minor effect on the reservoir
abandonment pressure. Actual field data confirms the models
predicted results.
Theoretical
Background
One of the critical factors in selecting tubing size is determining
when a well will load up and die. Determination of the factors
effecting this load up has been a matter of intense study over the
years.
The minimum flowrate required to unload a well
primarily depends on tubing size, liquid yields and flowing
tubing pressure. Two primary methods for determining when a
well begins to load up are Turner et als method using a physical
model and Nodal Analysis based on reservoir inflow performance and two phase flow correlations.
Turner Method. The Turner et al.s ) method for predicting gas
well load up is based on two physical models, liquid film
movement along the pipe walls and entrainment of liquid
droplets in a gas stream. Turner compared the two models with
observed field data and found that the LiquidDroplet Model was
superior. The LiquidDroplet Model is based on a freefalling
particle in a gas will reach a terminal velocity. This terminal
velocity depends on the particle size, shape, interracial tension,
liquid density, fluidmedium density and viscosity. In a gas well
if the upward velocity of the gas is less than the terminal
velocity, liquid will begin to accumulate and eventually the well
will load up and die. The Turner method predicts when liquids
can no longer be suspended in the gas as droplets. At this point
the well is at impending loadup.
Applying this terminal
velocity equation to wellbores and correcting to standard
conditions, the minimum gas flowrate, q,, for continuous
removal of liquids from a wellbore:
Refer to the result of the study, the tubing in 7 (seven) gas wells
were changed tlom 2 3/8 to 3 I/2. This project resulted in 50
MMCFD gas deliverability increase. The actual results agreed
closely with the predictions and demonstrated the accuracy of
the methods used.
Introduction
VICO Indonesia, the operator of SangaSanga Block in East
Kalimantan  Indonesia (Fig. 1) has four major gas fields that
produce an average of 1.6 BSCFD These fields consist of more
than 300 unique reservoirs that have permeability ranges from
less than 1 md to greater than 1,500 md. Most reservoirs are
depletion drive, Most of the 440 wells in the SangaSanga block
are dual completions. They range in depth from 5,000 to 14,000
ft. The SangaSanga block has three gathering system pressures,
high (950 psig), medium (375 psig), and low (150 psig).
367
RONALDGUNAWANand GEORGER DYER
2
3,06
9= =
where
Vg
generally accepted as the best correlation for wet gas wells. The
Gray correlation was developed from data on 108 wet gas cases,
mostly medium to high rate gas wells. The minimum wellhead
flowing pressure used to develop the correlation was 224 psig.
This correlation describes the pressure drop accurately over a
wide range of tubing sizes and water cuts in highpressure
wells4). However, in wells with low well head flowing pressure
and low rates, the Gray correlation presents a problem. Field
results show that the Gray correlation under predicts the actual
flowing bottom hole pressure4~.
(1)
Tz
= 5.62
(67
0.0031
(0,0031
P)
(2)
P)*
Determination
of Reservoir Abandonment
Pressure. Reservoir abandonment pressure can be estimated from Nodal
Analysis as the reservoir pressure where the inflow and tubing
performance curves will no longer intersect. At this point, Nodal
Analysis shows that the wellbore will loadup and the well will
die.
Turner et al. showed that usually wellhead conditions controlled
the onset of liquid loadup. The work of Turner shows that
liquid/gas ratios between I and 130 Bbl/MMSCF did not
influence the minimum lift velocity. The validity of the Turner
method has been tested for both high and low flowing wellhead
pressures (Fig. 2). In his original work, Turner suggested that a
safety factor of 20A be used in determining critical velocity.
Recent work by Coleman et al.z) however suggests that for high
flowing wellhead pressures (>500 psig FTHP) the 20A safety
factor should be used but for low flowing wellhead pressures
(<500 psigFTHp) no safety factor is required. Fig. 3 shows the
critical flowrate for 2 3/8, 2 7/8, 3 1/2, and 4 1/2 tubing at
various flowing wellhead pressures using Eq. 1.
Work by Coleman et aL2) shows that the critical rate
liquiddropletmodel
(Turner Method) accurately predicts
loadup for low flow rate and low wellhead flow pressure wells.
If the Turner critical rate is used as the criteria for determining
reservoir abandonment pressure then the required reservoir
pressure is significantly higher than that based on Nodal
Analysis. A comparison of the two methods is shown in Fig.4.
Since the Gray correlation under predicts the actual flowing
bottom hole pressure, it is not surprising that Nodal Analysis
using the Gray correlation also predicts a flowrate less than the
Turner critical rate at reservoir abandonment. However, if the
Turner rate is used as the rate at which the well will load up and
die then the Nodal Analysis can be used to predict the reservoir
abandonment pressure accurately.
Liquid load up can also be determined by
Nodal Analysis.
Nodal Analysis. Nodal Analysis predicts well performance by a
process that involves the intersection of the reservoir inflow
performance curve and tubing outflow performance curve. The
results of the integration can be plotted and analyzed. Fig. 4
shows an example of such a plot.
Inflow performance. The inflow performance relationship
(IPR) describes the relationship between gas production rate and
flowing bottomhole pressure. Inflow performance is controlled
by reservoir rock and fluid properties, including nearwellbore
effects, heterogeneities and reservoir pressure. A number of
expressions are available to describe the inflow performance of
a gas well. For pseudo steady state gas flow the inflow performance curve can be constructed using]):
0.000703
klr (P
~.
T p z [In (0.472
Determination
+ S + Dq]
Tubing
Size
The independence the critical flowrate to the amount of liquid
production is because the well is producing in the (annular) mist
flow regime, Therefore, the produced liquid is in individual
droplets and not in slugs or a continuous phase, The size of the
droplets is determined by the surface tension between the liquid
and the gas and not the amount of liquid being produced. If the
velocity in the tubing is high enough to lift the droplets to
surface, liquid will not accumulate in the tubing and well will
continue to flow.
(3)
Tubing Performance.
A tubing performance curve is a
prediction of the flowing bottomhole pressure at various gas
rates at a constant flowing wellhead pressure. A tubing performance curve is a function of fluid propefiies, tubing size, depth
and temperature. The tubing performance curves can be
developed using vertical single or twophase flow correlations.
The Gray correlation
of the Optimum
The Turner Model shows that for wells producing less than i 30
Bbl/MMSCF of liquid, load up does not depend on how much
liquid the well makes but is a function of the production rate. In
other words a well that makes 50 Bbl/MMscf is not any more
likely to load up than a well that makes 5 Bbl/MMscf provided
both wells produce above the Turner critical flowrate.
2Pw f*)
rc/rw)
SPE 37001
Model
Development,
As shown in Eq.3, reservoir inflow
performance is strongly affected by permeability thickness (kb)
and skin factor, Our analysis shows that for gas rates below 25
is widely used in wet gas wells and is
368
TUBING SIZE OPTIMIZATION
SPE 37001
.
IN GAS DEPLETION DRIVE RESERVOIRS
MMscfd, normal rate of VICOS wells, the effect of NonDarcy
coefficient is not significant. The other variables T, p and z do
not vary significantly and can be considered constants.
Tubing performance is a function of the produced
fluid
properties, tubing size, depth and temperature. Liquid yields,
tubing size and depth have significant impact on the tubing
performance. By determining the skin factor, the liquid yield
and the depth, the reservoir abandonment pressure for each
tubing size can be developed as a function of permeability
thickness by applying Nodal Analysis and Gas Loadup Technology.
The effect of the water yield, well depth, and skin factor on
reservoir abandonment pressure with 3 1/2 inch tubing is
presented in the following discussion. Other tubing sizes had
similar results. The effect of skin factor on reservoir abandonment pressure is presented in Fig. 5. As expected, Fig, 5 shows
that a higher skin factor increases the reservoir abandonment
pressure. However, as the permeability thickness increases, the
difference becomes smaller. For example in a 2,000 mdft well,
the difference in reservoir abandonment pressure for skin factor
of 25 and 5 is 190 psi while in a 6,000 mdft well the difference
is only 35 psi.
The effect of well depth on reservoir abandonment pressure is
presented in Fig. 6. A deeper wel I has a higher reservoir
abandonment pressure, but the effect in general is minor. For
example, for a 500 mdft well the difference in reservoir
abandonment pressure between a 5,000 ft and a 13,000 ft
completion is only 150 psi. This difference decreases with
increasing permeability thickness, in a 6,000 mdft completion
the reservoir abandonment pressure difference is only 100 psi.
The effect of water yield on reservoir abandonment pressure is
presented in Fig. 7. Increasing water yield will affect the
reservoir abandonment pressure because the flowing gradient
will increase as the yield increases, however, the effect is minor.
For example in a 2,000 mdft completion the difference in
reservoir abandonment
pressure for water yield of 3.6
Bbl/MMscf and 150 Bbl/MMscf is only 110 psi,
The key point from the above sensitivity analysis is that a
difference in skin factor, liquid yield, or well depth will increase
or decrease the reservoir abandonment pressure, but, the general
relationship between reservoir abandonment pressure, permeability thickness (kb) and tubing size does not change.
Rate. The results are plotted in Figs. 8,9, and 10.
By comparing Fig. 8,9, and 10, although the absolute value of
the reservoir abandonment pressure changes with the changes in
the wellhead flowing pressure the difference caused by different
sizes of tubing remains minor in high permeability thickness
completions.
The results from the model are compared with actual field data
in Fig. 11. A flowing wellhead pressure of 375 psig is used
because not enough data is yet available on the new 150 psig
gathering system. The actual field data agrees quite well with
the theoretical model. For VICOS application, only 2 3/8 and
3 1/2 tubing can be used. Based on Figs. 8, 9, 10, and 11, 3
1/2 tubing should be used in wells with a permeability thickness
greater than 2,000 mdft.
Refer to the above results, 7 (seven) Nilam gas wells with 2 3/8
tubing in the short string or single string were selected for a
tubing changeout (2 3/8 to 3 I/2). Each well has production
rates greater than 3 MMCFD in high pressure system (950 psig).
The Tubing ChangeOut technique is limited to the short string
or single string completions since they do not require killing the
completion during the workover ( a plug is installed in the tubing
below the packer).
The developed model was used in the candidate selection
process to make sure that reserve recovery is not significantly
affected by changing out the existing 2 3/8 tubing with 3 1/2(
tubing. Nodal Analysis was used to calculate the production
increase by changing the tubing. The method used in the Nodal
Analysis was the same as used in the model, From the seven
selected wells, a total of 50 MMCFD was gained compare to 52
MMCFD predicted. Fig. 12 shows an example of a prediction
and the actual results of one of the tubing changeout well.
Conclusions
Tubing size does not have significant effect on reservoir
abandonment pressure in high permeability thickness depletion
drive reservoirs. Above 2,000 mdft, the reservoir abandonment
pressure for 2 3/8, 2 7/8, 3 1/2, and 4 1/2 tubing is not
significantly different.
In wells with completions below 2,000 mdti the determination
of the proper tubing size is more complicated and it appears that
smaller tubing may be economically justified.
Nomenclature
Application
of the Model,
A database of340
VICO
wells was
analyzed to determine average parameters for the model,
Table I shows data used for the model. Using this data, a model
was generated for the three different pressure systems (950 psig,
375 psig, and 150 psig) and four different tubing sizes (2 3/8,
2 7/8, 3 1/2, and 4 1/2) with varying net permeability
thickness to determine reservoir pressure at the Turner Critical
A
D
h
k
P
P,
Pwf
=
=
=
=
=
=
=
flow area of the conduit, Lz, ft~
NonDarcy flow coefficient, t/L3, (MMscf/D)l
net pay, L, tl
effective gas permeability, L*, md
flowing tubing head pressure, m/Ltz, psi
average reservoir pressure, m/Lt2, psi
flowing bottomhole pressure, m/Lt2, psi
RONALD GUNAWAN and GEORGE R. DYER
~
%
rW
r,
=
=
=
=
=
=
=
=
n
z
P
gas flowrate, L3/t, MMscf/D
critical gas flowrate, L]/t, MMscf/D
welIbore radius, L,tl
drainage radius, L,ft
skin factor, dimensionless
temperature, T,OR
terminal velocity ofaparticle, L/t, ft/sec
gasdeviation factor, dimensionless
gasviscosity, m/Lt, cp
Acknowledgments
We acknowledge the management of VICO Indonesia and
Pertamina BPPKA for their permission to publish this paper.
Special thanks to Mr. Rick J. Louden who initiated and support
this study and to Mr. Edison Napitupulu for his support in
preparing this paper.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10
11
12.
13.
Turner, R. G., Hubbard, M. G,, Dukler, A. E. : Analysis
and Prediction of Minimum Flow Rate for Continous
Removal of Liquids from Gas Wells, JPT (Nov. 1969)
1475.
Coleman, S. B., Clay, H. B., McCurdy, D. G., and Norris
111,H. L. : A New Look at Predicting GasWell LoadUp,
JP7( March 1991) 329.
Brown, K. E : The Technology of Artificial LiR Metods Vol. 4, PennWell Publishing Company, Tulsa ( 1984) 249.
Oudeman, P. : Improved Prediction of WetGasWell
Performance; SPE PE (Aug. 1990)212.
Coleman, S. B., Clay, H. B., McCurdy, D. G., and Norris
111,H. L. : Understanding GasWell LoadUp Behaviour7
JPT( March 1991) 334.
Coleman, S. B., Clay, H. B., McCurdy, D. G., and Norris
111, H. L : The BlowdownLimit Model,.JPT ( March
1991)339.
Coleman, S. B., Clay, H. B., McCurdy, D, G., and Norris
III, H. L. : Applying GasWell LoadUp Technology,
JPT( March 199 I) 344.
Greene, W. R. : Analyzing the Performance of Gas Wells,
~PT(July 1983)1378.
Ikoku, Chi U. : Natural Gas Reservoir Engineering, John
Wiley & Sons, Tulsa (1984)141.
Libson, T. N., and Henry, J. R: Case Histories : Identification of and Remedial Action for Liquid Loading in Gas
Wells  Intermediate Shelf Gas Play: .Wf( Apr. 1980)685.
Moltz, A. K. :Predicting Gas Well LoadUp Using Nodal
presented at the 1992
System Analysis, pape; SPE
Annual Technical Conference and Exhibition, Washington,
Oct. 47.
Brown, K. E,, and Lea, J. F. : Nodal System Analysis of
Oil and Gas Wells,.lPT( Oct. 1985)175.
Upchurch, E. R. : Expanding the Range for Predicting
Critical Flow Rates of Gas Wells Producing from Normally
Pressured Waterdrive Reservoirs, JPT ( Aug. 1989)3 12.
370
SPE 37001
SPE 3700!
TUBING SIZE OPTIMIZATION
TABLE
Completion
Condensate
Condensate
IN GAS DEPLETION DRIVE RESERVOIRS
1 AVERAGE
Depth
Yield
Gravity
Gas Specific
Gravity
DATA FOR MODEL
.
(ft)
( BbllMMscf
9166
8.2
dimensionless
45
dimensionless
Water
Yield
( Bbl/MMscf
0.7
3.6
Water
Gravity
dimensionless
1.01
dimensionless
=
=
0.51
2000
0.05
Skin
Factor
Wellbore
Drainage
Radius
Radiua
Non Darcy
Fig.
1
(ft)
Coefficient
East
(ft)
( d/MMscf
Kalimantan
371
Oil and
Gas
18.5
Fields
RONALD GUNAWAN and GEORGE R. DYER
<.
A
.
.. .
..,*
.
,.
%
,
.
.,.
/
..*
m
.
SPE 37001
.:
..
0,1
Calculated Mlnimu;
Fig. 2 Critical
10
Flowrate [MMacfd)
Rate Database
..
Iu
Gas
gravity
= 0.7
a
7
6
5
4
3
1
1
250
Soo
750
1000
FTHP
Fig. 3 Critical
Flowrate
372
12s0
1S00
17S0
2000
2250
2S00
(psig)
for different
tubing sizes
TUBING SIZE OPTIMIZATION
.
SPE 37oO1
IN GAS DEPLETION DRIVE RESERVOIRS
500
L
Turners
critical
rate
<
original
interpretation
1.. . ~.,
1501
0.0
200
400
x.,
.
600
. ........
1000
800
GAS RATE (Mscf/d)
Fig. 4
I&x
Determination
of Reservoir
Abandonment
Pressure
.0
a
4C0
2C0 .
lax
Zom
4CJ20
woo
6JXQ
7000
Eml
k h ( mdft)
Fig. 5
Effect
Case
of Skin Factor on Abandonment
: 31/2 tbg, Low Pressure
Gas
373
Pressure
System
(150
psig)
9m
RONALD GUNAWAN and GEORGE R. DYER
SPE 37001
%00
...........EEEEl
...............
1400
lzoo
~
woo
2
n
no
800
600
400
200
Oi
moo
2000
3000
4000
6000
5000
7000
8000
9000
k h ( mdft)
Fig. 6
Effect
Case
of Well Depth on Abandonment
: 31/2 tbg, Low Pressure
Gas
Pressure
System
(150
psig)
600
!400
g
w
n
1300
.

W@w YIe1d=lWBBUMM8d
800
n
CL
600

. . 

. . . . . . . . . . . . . . . . . . . . . . . . 
400
200
0
0
woo
2000
3000
4000
5000
6000
7000
8000
k h (mdft)
Fig. 7
Effect
Case
of Water Yield on Abandonment
: 31/2 tbg, Low Pressure
Gas
374
Pressure
System
(150
psig)
9000
TUBING SIZE OPTIMIZATION
SPE 37oo1
IN GAS DEPLETION DRIVE RESERVOIRS
23/8
u
Zm
(199S ID)
27/8 (2 447 ID)
312
..
(2.992 ID)
 4K? (3.9W
Iq
a
!03,

k h ( mdft)
Fig. 8
Abandonment
Pressure
Case : Low Pressure
2m
for different
tubing
sizes
Gas System (150 psig)
2?W(l
SQYID)
2.7/8 (244 flD)
&
V2(29WlD}

4 VZ(39WID)
El
txo.
tire.
o
0
an)
Uml
am
k h ( mdft)
Fig. 9 Abandonment
Pressure for different tubing sizes
Case : Medium Pressure Gas System (375 psig)
375
L?mJ
RONALD GUNAWAN and GEORGE R. DYER
10
SPE 37001
4000
23/8
I
I
I
3%)0
3000
\:
(1.99S ID)

2.7/8 (244~
3.

K? (2.992
.4
ID)
ID)
W(3958ID)
2530
2000
ao
2000
4000
6000
8000
Qooo
moo
k h ( mdft)
Fig.
10
Abandonment
Case : High
MP Gas System
Pressure
for different
tubing sizes
Pressure
Gas System
(950 psig)
(375
pslg FTHP)
P
c9ia3v2,
&
ActW4Daa3v2
Mualc91a238
msl
mm
4(J3I
em
12col
KoM
k h ( mdft)
Fig. 11 Abandonment
Theoretical
Pressure
vs Actual
376
for23/8
Field Data
and
3112
tbg
TUBING SIZE OPTIMIZATION
SPE 37001
IN GAS DEPLETION DRIVE RESERVOIRS
11
..
Boo
9000
4 Eao
8000
4000
moo
3520
6000
5000
4000
\
\
two
3000
%
2000
moo
XIO
moo .
0,
3
Yea r
R,te@cur
Fig. 12
(m!
tb, ng
Well
Rale
L43
Deliverability
l/~Tub,
ng
Forecast
377
Aclu,lf
late,
2 3/8

. 
WHPd,
vs 3 1/2
urren!tbo
tubing
flHPrJf3
(Well
1{2 TLw
A)