Professional Documents
Culture Documents
FINAL REPORT
OCTOBER 2006
2769-TTU-DOE-2098
MS 3111
LUBBOCK, TX 79409-3111
i
DISCLAIMER
United States government. Neither the United States Government nor any agency thereof,
nor any of their employees, makes any warranty, expressed or implied, or assumes any
information, apparatus, product, or process disclosed, or represents that its use would not
infringe privately owned rights. Reference herein to any specific commercial product,
United States Government or any agency thereof. The views and opinions of authors
expressed herein do not necessarily state or reflect those of the United States Government
ii
ABSTRACT
The minimum gas rate for unloading liquids from a gas well has been the
subject of much interest, especially in old gas producing fields with declining reservoir
pressures. For low-pressure stripper gas wells, liquid production accumulating in the
tubing is a pivotal factor that could lead to premature well abandonment and a huge
Some notable correlations that exist for predicting the critical rate required for
liquid unloading in gas wells include Turner et al., (1969), Coleman et al., (1991),
Nosseir et al. (1997), Li et al. (2001) and Veeken et al., (2003). However, these
correlations offer divergent views on the critical rates needed for liquid unloading, and
work on critical gas velocities required to keep liquid from accumulating in the tubing.
Also, during the course of the work, data points were collected using a flow test facility at
Texas Tech University. The critical gas rates were experimentally measured in order to
determine an improved correlation with specific application for low-pressure stripper gas
iii
TABLE OF CONTENTS
DISCLAIMER ....................... ii
ABSTRACT ...............iii
CHAPTER
I INTRODUCTION ........ 1
iv
2.2.1 Critical Velocity Definition .........10
v
4.3 Test Model Development ..... 35
REFERENCES ........ 42
APPENDICES ......... 44
EQUATIONS ..50
vi
LIST OF FIGURES
vii
2.9 Deformation of a Free-Falling Droplet .......... 16
viii
INTRODUCTION
Stripper gas wells are defined as the wells that produce 60 Mscfd or less at low
reservoir pressures. These marginal gas wells amount to nearly 261,000 wells in the
United States (U.S.) and account for about 7 percent of the total natural gas produced
on-shore in the nation, excluding Alaska1. The number of stripper gas wells in the U.S.
has steadily increased during the past ten years with a proportionate increase in
production. The average production from a U.S. stripper gas well in 2003 was 15.5
Mscfd.
Through normal reservoir depletion over time, all producing gas wells will
eventually become stripper gas wells. Figure 1.1 shows the trend of a typical gas well
decline curve. A depleted gas reservoir is a gas reservoir in its late production period with
declining gas production rate and pressure. It is characterized by low wellhead pressure
and probably having some liquid production towards the end of the life of the well.
Production Rate
(Mscf/d)
Time
1
Decreasing gas flow rate, as depicted in Figure 1.2, allows liquids to accumulate
within the wellbore, thereby imposing a backpressure on the formation that can reduce
production capacity and eventually cease gas production. Another way to express this is
the percentage of liquids occupying the tubing volume increases, which is the same as
holdup increasing. The term holdup is often used in multiphase flow modeling.
This process is termed liquid loading and is not peculiar with low-pressure gas
wells alone, but may exist in high-pressure gas wells with high liquid-gas ratio. Liquid
unloading is said to occur if the gas velocity has sufficient energy to lift the accumulated
liquid out of the wellbore. It is therefore important to maintain a minimum or critical gas
Several models have been proffered to help determine the critical velocity
needed for liquid unloading in gas wells. The most widely used is the correlation
2
They developed two models the continuous film and the entrained droplet movement
models and proved that the entrained droplet movement model, with a 20% upward
adjustment to match field data, is more adequate for predicting critical velocity. The field
data used to derive Turners model are from wells with wellhead pressures that are
greater than 800 psia. Another group of researchers, Coleman, et al.4 (subsequently
referred to as Colemans model), used the same entrained droplet movement model but at
wellhead pressures lower than 500 psia. They concluded that better results could be
obtained without a 20% upward adjustment to match field data. Other notable works on
this subject include the correlations obtained by Veeken, Bakker, and Verbeek5
(subsequently referred to as Veekens model), Li, Li, and Sun6 (subsequently referred to
as Lis model) and Nosseir et al.7 (subsequently referred to as Nosseirs model), with
different critical rate models and subsequent predictions, particularly at low wellhead
pressures. Particularly, Veekens model shows that a much higher rate than predicted by
Turners model is required at low pressures. It is noted that this was for offshore data and
some of the wells were deviated but in general the deviation seemed not to be a large
3
1.2 Statement of Problem
correlations at low wellhead pressures. The following objectives are met by this study:
2. Collect experimental data and experimentally determine the critical rate at pressures
3. Provide improved critical gas rate versus wellhead pressure correlations, specifically
for use with stripper gas wells at pressures lower than 50 psia at the wellhead.
Experimental data in this study were collected using compressed air (relative
density of 1.0) and water (with density of 62.4 lbm/ft3) at an average ambient temperature
of 64oF and atmospheric pressure of 13.2 psia (682 mmHg). The study is limited to
wellhead pressures of 50 psia or less and wells with liquid gas ratio in the range 40 195
bbl/MMscf. The liquid droplet movement model is adopted with the initial assumptions
of Turner et al.3.
The correlation derived was not matched with field data however. All units are
4
PROJECT REVIEW
When a gas well begins to liquid load, there are several symptoms that can be
observed that will indicate the onset of this condition. Some of these symptoms are:
This is typified in Figure 2.1, which shows severe slugging recorded by the
differential pen on a gas chart as the slugs pass through the orifice, indicating the onset or
presence of liquid loading. The chart would only have smooth circles without the spikes
on it if no produced fluid slugs were present. Many modern data taking instrumentation
techniques no longer make use of the chart but instead collect data electronically and
present a line trace. However the idea that a differential measurement would become
5
Figure 2.1: Gas Chart Indicating Liquid Loading2
A typical gas well decline curve is smooth and gradual but sharp drops in the
decline curve usually indicate problems related to liquid loading in the tubing as shown in
Figure 2.2. Nevertheless, it is quite possible that erratic surface pressures occurring on the
tubing from other scenarios happening in the field could lead to sharp drop in decline
curve as observed for liquid loading. If liquid loading occurs, the rate usually will drop
6
off the original decline curve to a lower, less productive decline curve. It could stop
Actual with
(Mscf/d)
Loading
Time
When liquid loading occurs and liquids accumulate in the tubing, the tubing
flowing gradient becomes heavier due to liquids in the tubing with subsequent increase in
the difference between the tubing and casing pressures over time. This loading
7
(PCSG-PTBG), psia
Time
Figure 2.3: The Casing-Tubing Pressure Profile of a Loading Well
A well without liquid loading should show a smooth gas gradient in the tubing.
Liquid condensation dropout in the tubing nearer the surface of the well would normally
have a higher gradient than the rest of the tubing as shown in Figure 2.4.
8
2.2 Critical Velocity Models Based on Spherical Droplet
The study of the minimum critical rate required to remove liquid phase material
from a loading well date back to the 1940s. Vitter8 and Duggan9 proposed that wellhead
velocities observed in the field would be adequate for keeping wells unloaded. Jones10
and Dukler11 presented analytical treatments resulting in equations for calculating the
Turner, Hubbard, and Dukler3, after studying the earlier observations, proposed
two physical models for the removal of gas well liquids. The models are based on: (1) the
liquid film movement along the walls of the pipe and (2) the liquid droplets entrained in
the high velocity gas core. They used field data to validate each of the models and
concluded that the entrained droplet model could better predict the minimum rate
required to lift liquids from gas wells. This is because the film model does not provide a
clear definition between adequate and inadequate rates as satisfied by the entrained
droplet model when it is compared with field data. A flow rate is determined adequate if
the observed rate is higher than what the model predicts and inadequate if otherwise.
Again, the film model indicates that the minimum lift velocity depends upon the
gas-liquid ratio while no such dependence exists in the range of liquid production
associated with field data from most of the gas wells (1 - 130 bbl/MMscf).
9
The step by step derivation of the critical flow rate expressions, using the
entrained droplet model and assuming that wellhead conditions control the onset of
Vc , w = 5.304
(67 0.0031P ) 4 (Field Units)...... (1)
1
(0.0031P ) 2
1
Vc ,cond = 4.03
(45 0.0031P ) 4 (Field Units)...... (2)
1
(0.0031P ) 2
Using the same model but validating with field data of lower reservoir and
wellhead flowing pressures all below approximately 500 psia, Coleman et al.4 were
convinced that a better prediction could be achieved without a 20% upward adjustment to
Vc , w = 4.434
(67 0.0031P ) 4 (Field Units)...... (3)
1
(0.0031P ) 2
1
Vc ,cond = 3.369
(45 0.0031P ) 4 (Field Units)...... (4)
1
(0.0031P ) 2
falling body in a fluid medium under the influence of gravity alone. The critical velocity
10
quantity to guarantee removal or upward movement of the largest liquid droplet. This
terminal velocity is therefore a function of the size, shape and density of the particle and
Hinze12 ascertained that the balance of two pressures the velocity pressure and
the surface tension pressure, determines the maximum size a droplet may attain. The
surface tension of the liquid phase acts to draw the droplet into a spheroidal shape. Both
the Turner and Coleman equations are based on fixed droplet shape, a constant size and
assume a solid sphere. The Weber number was established experimentally from droplets
falling in air but not for conditions existing in gas wells although gas properties can be
introduced.
Coleman et al.4 explained that the two known sources of wellbore liquids are: (1)
the liquids condensed from gas owing to wellbore heat loss and (2) the free liquids
produced into wellbores with gas. Depending on the specific reservoir, both liquid
hydrocarbons and water may also be present. The denser water controls the onset of
loading. The amount of water condensed is shown to increase exponentially as the static
11
reservoir pressure declines thereby compounding the wells loading problem. This is
primarily due to the fact that at lower flows, the upper section of the tubing becomes
cooler and liquids drop out. In fact, lower pressures hold water vapor in the vapor state
in Figure 2.5. Coleman et al.4 also contends that the examination of other variables such
as temperature, gas gravity and interfacial tension shows a minor influence on the
12
2.2.4 Flow Pattern Recognition
Taitel et al.13 identified four distinct flow patterns bubble, slug, churn and
annular flow and evaluated the transition boundaries among them. The various flow
These flow patterns, together with the transitions, were later mapped in Figure 2.7 by
combining the works of Taitel et al.13, Barnea14, and Scott and Kouba15. This is referred
13
Figure 2.7: Flow Pattern Mapping with Transitions16
The flow regime map used by Duns and Ros17, intended for steady state flow is
(5)
1
4
Gas Velocity Number = v sG G ...........
g
(6)
Where:
qL
v sL = ....... (7)
A
14
qG
v sG = ...... (8)
A
gas field, the flow regime switches from annular mist flow to churning flow and the
lifting capacity of the flowing gas decreases dramatically. They described the critical
flow rate as the rate at which this switch in flow regime occurs. At the start of gas flow in
unsteady-state tests, liquid is often produced by slugging flow. Therefore, flow regimes,
specifically the transition from slug flow to annular-mist flow, are important in liquid
lifting.
15
2.3 Critical Velocity Model Based on Flat-Shaped Droplet
Li, Li, and Sun6, in 2002, proposed that the critical Weber Number (NWE) of 30
put forward by Hinze12 and used to evaluate critical velocities in both Turners and
Colemans models did not take the deformation of a free falling droplet into
stream, a pressure difference exists between the fore and aft portions of the droplet. The
droplet is deformed under the applied force and its shape changes from spherical to that
of a convex bean (approximated to a flat shape) with unequal sides as shown in Figure
2.9.
Spherical liquid droplets have a smaller efficient area and need a higher
terminal velocity and critical rate to lift them to the surface. However, flat-shaped
16
droplets have a more efficient area and are easier to be carried to the wellhead. The full
derivation of Lis critical rate model is given in Appendix C and is summarized below:
1 1
vc = 2.5
4
( )
l g
4
1
(S.I Units)........... (9)
( ) g
2
Apvc
qc = 2.5 10 5 (S.I Units)......... (10)
zT
The drag coefficient for Lis Model at Reynolds Number 104 to 105 is approximately 1.0
Nosseir et al.7 focused their studies on the impact of flow regimes and changes
in flow conditions on gas well loading. They assumed the droplet model and analyzed the
downward gravitational force and the upward drag force on the model. The representative
drag coefficient (Cd) equations are defined under laminar, transition and turbulent flow
regimes, which in turn determine the expression of the drag force and hence critical
velocity equations.
dominance of turbulent flow regime in a certain range. They showed that most of
Turners data points fall in the highly turbulent region where Reynolds number (Re)
exceeds the value of 200,000. This region is termed the highly turbulent flow regime
with Re extending from 200,000 to 106 with a Cd of 0.2. They again confirmed that most
17
of Colemans data points fell in the Re range of 1,000 to 200,000 with its corresponding
drag coefficient Cd of 0.44 for spherical droplets. They argued that this is the reason why
Turners initial equation has a good match for Colemans data without adjustment. A
presentation of the critical velocity equations for the Nosseirs model is given in
Appendix D.
Nosseirs model gave analytical critical velocity equations for the transition
flow regime and the highly turbulent flow regime with Re of 2 105 to 106. They
suggested that transition regime might occur in low pressure and flow rate systems and
14.6 0.35 ( l g )
0.21
Again, the critical velocity equation for highly turbulent flow regime is given as:
21.3 0.25 ( l g )
0.25
Veeken, Bakker, and Verbeek5 devised a ratio term called Turner Ratio (TR),
which is the ratio of the actual flow rate and Turners flow rate required for liquid
performance as:
18
0.172
P
TR = 3.77 A 0.5 ... (13)
14.5
Where:
0.1678 Pwf Pr
A= ... (14)
qg
A plot of the Turner Ratio term and best-fit combination of inflow parameter
A with wellhead pressure is shown in Figure 2.10. The correlation below showed that at
lower wellhead pressures, the required critical rate is greater than what the Turners
19
Figure 2.10: Turner Ratio vs. Product of A and FTHP5
The Veekens correlation data included some deviated wells and some casing
flow. They also reported that critical flow for deviated wells is about the same for vertical
critical flow rate using the bottom-hole static pressure. The Veekens model very well
contradicts Colemans correlation at low wellhead pressures and a plot showing the
comparison between the Turners, Colemans and Veekens models are shown in Figure
2.11.
20
Figure 2.11: Comparison of Turners, Colemans and Veekens Predictions5
diameter, and specific gravity on minimum gas flow rate using the Turners model. They
made use of water and condensate as liquids, with specific gas gravity ranging from 0.55
to 0.7, flow diameters in the range 2.0 to 5.0 inches, pressure ranging from 250 to 2,000
selected temperatures, at constant specific gravity and flow diameter and deduced that
21
minimum flowrate increases as the pressure increases at a fixed temperature. This is
Again, they showed the relationship between minimum flowrate and pressure at
selected flow diameters with constant specific gravity and temperature. Their relationship
is depicted in Figure 2.13 and showed that at each pressure, the minimum flowrate
increases as the diameter of flow increases. This can be explained by the corresponding
22
Figure 2.13: Flow Diameter Effect on Minimum Gas Flowrate19
function of specific gravity of gas with constant flow diameter and temperature was also
investigated. They showed that for wellhead pressures up to 1,500 psia, lower gas gravity
values yield a higher minimum gas flowrate required to unload the well. At wellhead
pressure above 1,500 psia, the difference in gas gravity values has less profound effect on
the minimum gas flowrate. This is shown in Figure 2.14. The decrease in the difference
factor involved in the gas flowrate equation. Ilobi and Ikoku20 also hinted that momentum
23
could be a vital factor in the ability of gas to carry liquid droplets up in the wellbore.
They also suggested that increased pressure, which results in increased gas gravity, was
METHODOLOGY
b) Gas tank.
24
c) Compressor.
e) Pressure gauge.
f) Temperature gauge.
g) Pressure regulator.
j) Mercury barometer.
k) Stop watch.
25
10
7 5
6
4
3
8 1
26
The liquid flow meter is read in percentages and multiplied by a constant factor to get an
equivalent flowrate in gallons per minute. Similarly, the gas flow rate was measured by
the time required for 100 cubic feet of gas to pass through the orifice.
a) Set the pressure regulator at the top of the tubing to a predetermined pressure
b) Set the liquid flow rate at a minimum predetermined rate and allow water to
c) Set the flow rate of air and check to see if water is moving up.
d) Run for 10 minutes and shut both water and air supply off.
f) If the water level is less than 18 inches, this implies that the prevailing gas rate is
more than what is required for critical rate. Reduce the gas rate and re-run steps (c) to
(e). Otherwise, if the water level is higher than 18 inches, this implies that the
prevailing gas rate is less than what is required for critical rate. Increase the gas rate
g) Continue to do this until the water level is constant before and after the test.
h) Record the gas pressure, liquid flow rate, temperature and tubing pressure data.
27
i) Re-run steps (b) to (g) with a higher liquid rate and a higher tubing pressure and
record data at which water level in the tubing is same before and after the test.
The test facility is set up in such a manner that there will always be a pressure
drop across the choke, as shown in Figure 3.1, when gas passes through the flowline. Gas
qsc = 23.088Cv
(p2
1 )
p22
..... (15)
gT1
Where:
T1 = absolute temperature, oR
The coefficient of flow (Cv) is a formula used to determine a valves flow under
various conditions and most valve manufacturers have flowcharts that display the Cv
value at different turns for calculating gas flow rates. The Cv value of the valve used in
28
Figure 3.2: MC 10 Flow Curves21
29
DISCUSSION OF RESULTS
stripper gas wells, the data gathered from the laboratory experiment are analyzed and
compared with previous correlations, using a value of 62.4 lbm/ft3 for water density and
The initial data, shown as measured critical rate data in Appendix A, were first
smoothed with moving average method. The smoothed data were then used to develop
the critical rate versus pressure equation by the least-square technique. The best fit for
1000
Lab. Measurement
Smoothed Data
Expon. (Least Square Prediction)
Critical Gas Flow Rate (Mscf/d)
R2 = 0.35
100
10
15 20 25 30 35 40 45 50
FTP (psia)
30
4.1 Flow Regime and Mapping
The critical rate data obtained in this study are shown to be in the Reynolds
number (Re) range of 36,500 and 153,000, defining turbulent flow during the course of
the experiment. Thus, we can predict the flow pattern of each test according to the
Ansari16 model by mapping on a superficial liquid velocity versus superficial gas velocity
It could be observed that all test data fall outside the transition zone (D1)
required to prevent the entrained liquid droplets from falling back into the gas stream.
The test velocity data also approach the Barnea transition zone D2, which includes the
effect of film thickness and instability of the liquid film that may result in downward
flow of the film at low liquid rates. Yamamoto and Christiansen18 also gave the opinion
that the transition from slug flow to annular-mist flow is important in liquid lifting.
31
10
Barnea
B
Superficial Liquid Velocity (m/s)
Transition
U considering film
B thickness
1 A
B N
L SLUG OR CHURN N
Y U
0.1
L
A
Our Data R
0.01
D1 D2
0.001
0.1 1 10 100
Superficial Gas Velocity (m/s)
Having determined that most of the test data collected in this study fall in the
turbulent flow category, it is necessary to compare the collected critical rate data vis--vis
other models. Table D.2 in Appendix D compared our critical rate data (by least square
Turner, Coleman and Li. The various results from different models are depicted in Figure
4.3.
32
200
o
Turner's Model@ 64 F
170
o
Critical Gas Flow Rate (Mscf/d)
Coleman's Model@ 64 F
140
o
Our Model@ 64 F
110
80 o
Li's Model@ 64 F
50
20
14 22 30 38 46 54 62
FTP (psia)
determining critical rates in wells with flowing tubing wellhead pressures of 800 psia or
higher, Colemans model is judged to be better suited for flowing tubing well pressures
of 500 psia or less. A careful analysis of Colemans data at pressures below 110 psia
reveals data inconsistency in the pressure ranges of 40-50 psia and 50-85 psia, with
critical rates decreasing as wellhead pressures increases. This is shown in Figure 4.4.
33
Colemans data marked as Ragged Trace are data collected from test well or gas charts
at the onset of liquid loading. Colemans prediction model may also have predicted
higher critical rates at low pressures since the model includes tubing wellhead pressures
720
Coleman's Model
Critical Gas Flow Rate (Mscf/d)
520
420
320
220
Our Model
120
20
20 30 40 50 60 70 80 90 100 110
FTP (psia)
Lis model, on the other hand, brings a divergent but important point of view in
showing that it is possible for the shape of entrained droplet movement model to change
from spherical to another shape because of pressure difference across the fore and aft
34
4.3 Test Model Development
stripper gas wells with our experimental data using the Turner et al.3s entrained droplet
movement model with well articulated assumptions about the shape of the entrained
droplet. Starting with Turners model and assumptions stated in Appendix B and
Equation B3:
( l g )d
Vc = (V g Vd ) = 6.55 ... (17)
Cd g
g c
The initial assumption is a spherical droplet with diameter (d ) = 30 . Equation 17
gVc2
then becomes:
30(0.00006852 ) g c ( l g )
Vc = 6.55 ... (18)
C d g2Vc2
1.68
1
2
( l g )2
1
Vc = . (19)
C d Vc g
1 1
1.2978 ( ) 4
l g
4
Vc = 1 1
(20)
(C ) ( )
d
4
g
2
1
( )
l g
4
Vc = 1
..... (21)
g2
Where:
3.612
= 1
.. (22)
Cd 4
35
The expression in Equation 21 was used to derive an average term for using
gas rate data obtained by the least square technique and shown in Table A.1 of Appendix
A. is found for each data point and the average value over the range considered is
3.612
= 1
= 3.391 ...... (23)
Cd 4
The Cd value of 1.3 perfectly matches the drag coefficient for a disk or cylinder for
turbulent flow with Reynolds number of range 1000 200,000. This is shown in Figure
4.5.
36
Figure 4.5: Drag Coefficients for Spheres, Disks and Cylinders22
Thus, results obtained and modeled here could be represented fully by the
expressions:
1
( l g ) 4
Vc = 3.391 1
... (25)
g2
3060 PAtVc
q= ...... (26)
(T + 460) z
The relationship between the critical rate and critical velocity by is shown in Figure 4.6.
120 35
100
90 25
80
20
70
60 15
10 15 20 25 30 35 40 45 50
FTP, psia
37
1. Liquid droplets entrained in gas wells at very low pressures could approximate either
Turners data indicated that the LGR does not influence the critical lift velocity
in the observed LGR range from 1-130 bbl/MMscf while the extent of LGR effect on
A plot of LGR and critical rates using our test model reveals that high LGR
corresponds to low critical rates. This is shown in Figure 4.7 and suggests a
wells.
38
250
200
LGR, Bbls/MMscf
150
100
50
0
60 70 80 90 100 110 120 130
Critical Rates, Mscf/d
39
CONCLUSIONS AND RECOMMENDATIONS
5.1 Conclusions
The following conclusions were made based on the analysis of our experimental
data used to determine critical rates at pressures between 15 to 50 psia using air and
water:
1. Wellhead conditions are often used to describe the minimum gas velocity required to
keep a well unloaded. However, for deeper wells, most expressions show that more
rate is needed to be above critical at the bottom of the well than at the surface. In tests
carried out here, the tube used was very short compared to a deep well with higher
2. Critical flow rates at low pressures fall under the turbulent flow regime with
3. Critical rate predictions based on Turners model are very much higher than actual
minimum rates required for liquid unloading. Most of Turners data are in the highly
4. The minimum rates required for liquid unloading of stripper gas wells at low
pressures are not sufficiently modeled by either the Colemans or Lis models
because of the incorrect assumption about the shape of the entrained liquid.
40
5. A cylindrical disk with thickness of about two-third of its diameter and drag
coefficient of 1.3 best describes the shape of an entrained liquid droplet flowing under
6. The critical velocity required for unloading a low-pressure stripper gas well, based on
the data collected and modeled in this study, could be described by equations:
1
( l g ) 4 3060 PAtVc
Vc = 3.391 1
and q = .
g2 (T + 460) z
5.2 Recommendations
1. The correlation obtained from this study could be refined by matching with field data
41
REFERENCES
2. Lea, J.F., Nickens, H.V., and Wells, M.: Gas Well De-Liquification, first edition,
Elsevier Press, Cambridge, MA (2003).
3. Turner, R.G., Hubbard, M.G., and Dukler, A.E.: Analysis and Prediction of
Minimum Flow Rate for the Continuous Removal of Liquids from Gas Wells, J. Pet.
Tech. (Nov.1969) 1475-1482.
4. Coleman, S.B., Clay, H.B., McCurdy, D.G., and Lee Norris, H. III: A New Look at
Predicting Gas-Well Load Up, J. Pet. Tech. (March 1991) 329-333.
5. Veeken, K., Bakker, E., and Verbeek, P.: Evaluating Liquid Loading Field Data and
Remedial Measures, presented at the 2003 Gas Well De-Watering Forum, Denver,
CO, March 3-4.
6. Li, M., Li, S.L., and Sun, L.T.: New View on Continuous-Removal Liquids from
Gas Wells, paper SPE 75455 presented at the 2001 SPE Permian Basin Oil and Gas
Recovery Conference, Midland, TX, May 15-16.
7. Nosseir, M.A., Darwich, T.A., Sayyouh, M.H., and El Sallaly, M.: A New Approach
for Accurate Prediction of Loading in Gas Wells Under Different Flowing
Conditions, paper SPE 37408 presented at the 1997 SPE Production Operations
Symposium, Oklahoma City, OK, March 9-11.
8. Vitter, A.L.: Back Pressure Tests on Gas-Condensate Wells, API Drilling and
Production Practice (1942) 79.
9. Duggan, J.O.: Estimating Flow Rates Required to Keep Gas Wells Unloaded, J. Pet.
Tech. (December 1961) 1173-1176.
10. Jones, P.J.: Petroleum Production, Reinhold Publishing Corp., New York, NY (1947)
3, 100.
11. Dukler, A.E.: Design of Multi-Phase Flow Systems, paper presented at the API
Southwestern District Meeting, Albuquerque, NM (March 1961).
42
12. Hinze, J.O.: Fundamentals of the Hydrodynamic Mechanism of Splitting in
Dispersion Processes, AIChE J. (Sept. 1955), 1, No.3, 289-295.
13. Taitel, Y., Barnea, D., and Dukler, A.E.: Modelling Flow Pattern Transitions for
Steady Upward Gas-Liquid Flow in Vertical Tubes, AIChE J. (1980) 26, 345.
14. Barnea, D.: A Unified Model for Predicting Flow-Pattern Transition for the Whole
range of Pipe Inclinations, Intl. J. Multiphase Flow (1987) 13, 1.
15. Scott, S.L. and Kouba, G.E.: Advances in Slug Flow Characterization for Horizontal
and Slightly Inclined Pipelines, paper SPE 20628 presented at the 1990 SPE Annual
Technical Conference and Exhibition, New Orleans, LA, Sept. 23-26.
16. Ansari, A.M., Sylvester, N.D., Sarica, C., Shoham, O., and Brill, J.P.: A
Comprehensive Mechanistic Model for Upward Two-Phase Flow in Wellbores,
paper SPE 20630 presented at the 1990 SPE Annual Technical Conference and
Exhibition, New Orleans, LA, Sept. 23-26.
17. Duns, H., Jr. and Ros, N.C.J.: Vertical Flow of Gas and Liquid Mixtures in Wells,
Proceedings of the 6th World Petroleum Congress (1963) 451-465.
18. Yamamoto, H. and Christiansen, R.L.: Enhancing Liquid Lift from Low Pressure
Gas Reservoirs, paper SPE 55625 prepared for presentation at the 1999 SPE Rocky
Mountain Regional Meeting, Gillette, WY, May 15-18.
19. Bizanti, M.S. and Moonesan, A.: How to Determine Minimum Flowrate for Liquids
Removal, World Oil, (Sept. 1989) 71-73.
20. Ilobi, M.I. and Ikoku, C.U.: Minimum Gas Flow Rate for Continuous Liquid
Removal in Gas Wells, paper SPE 10170 presented at the 1981 SPE of AIME
Annual Fall Technical Conference and Exhibition, San Antonio, TX, Oct. 5-7.
21. Cameron Willis In-line Chokes Models MC10 & MC20 Technical Manual.
22. Lapple and Shepherd: Ind. Eng. Chem. (1940) 32, 605.
43
APPENDIX A
44
Table A.1 continued
Tubing Measured Least Square Measured Gas Lab. Measured Superficial Vel
Data Pressure Critical Rate Critical Rate Liq. Rate Density Temp. LGR VsG VsL Re
3 o
(psia) Qg (Mscf/d) Qg (Mscf/d) (BPD) (lbm/ft ) ( F) (bbl/MMscf) (m/s) (m/s)
28 25.37 68.33 82.81 8.67 0.1314 78 126.94 11.10 0.0079 57,795
29 25.66 74.06 83.17 2.54 0.1329 60 34.26 12.04 0.0023 46,980
30 25.66 88.16 83.17 7.61 0.1329 62 86.33 14.33 0.0069 66,527
31 25.66 95.40 83.17 12.69 0.1329 64 132.97 15.50 0.0116 81,963
32 25.66 101.50 83.17 8.67 0.1330 74 85.46 16.49 0.0079 76,391
33 25.87 98.97 83.43 17.35 0.1340 60 175.29 16.08 0.0158 94,174
34 26.13 115.01 83.76 2.54 0.1354 58 22.06 18.69 0.0023 69,647
35 26.16 127.22 83.80 13.01 0.1356 68 102.28 20.67 0.0119 100,587
36 26.18 140.12 83.82 2.54 0.1356 66 18.11 22.77 0.0023 83,546
37 26.18 127.84 83.82 5.71 0.1356 61 44.65 20.77 0.0052 84,200
38 26.37 136.00 84.06 26.02 0.1366 60 191.35 22.10 0.0237 134,093
39 26.43 75.32 84.14 5.07 0.1370 59 67.37 12.24 0.0046 53,573
40 26.66 85.41 84.43 17.35 0.1381 68 203.13 13.88 0.0158 86,352
41 26.83 95.05 84.65 7.61 0.1390 58 80.08 15.44 0.0069 70,368
42 27.03 109.18 84.90 10.15 0.1401 58 92.95 17.74 0.0093 84,028
43 27.16 119.86 85.07 21.69 0.1407 66 180.93 19.48 0.0198 115,505
44 27.33 131.39 85.29 12.69 0.1416 58 96.55 21.35 0.0116 102,187
45 29.31 135.61 87.85 8.67 0.1519 61 63.97 22.04 0.0079 95,386
46 29.61 140.30 88.25 17.35 0.1534 59 123.65 22.80 0.0158 117,641
47 30.18 75.22 89.01 3.81 0.1564 80 50.60 12.22 0.0035 50,582
48 30.18 83.71 89.01 7.61 0.1564 76 90.93 13.60 0.0069 64,037
49 30.31 79.94 89.18 26.02 0.1570 57 325.53 12.99 0.0237 100,871
50 30.81 91.69 89.85 34.70 0.1596 56 378.42 14.90 0.0317 125,152
51 31.65 95.43 90.99 3.81 0.1640 66 39.88 15.51 0.0035 61,791
52 31.65 90.32 90.99 7.61 0.1640 66 84.27 14.68 0.0069 67,733
53 32.63 100.86 92.34 2.54 0.1691 68 25.15 16.39 0.0023 61,819
54 32.65 114.41 92.36 3.81 0.1692 74 33.26 18.59 0.0035 72,304
55 32.65 90.90 92.36 7.61 0.1692 73 83.74 14.77 0.0069 68,054
56 33.13 117.84 93.04 7.61 0.1717 68 64.59 19.15 0.0069 83,056
57 33.13 129.48 93.04 12.69 0.1717 64 97.97 21.04 0.0116 101,119
58 34.69 99.55 95.24 3.81 0.1797 81 38.23 16.18 0.0035 64,071
59 35.00 125.90 95.69 3.81 0.1814 42 30.23 20.46 0.0035 78,667
45
Table A.1 continued
Tubing Measured Least Square Measured Gas Lab. Measured Superficial Vel
Data Pressure Critical Rate Critical Rate Liq. Rate Density Temp. LGR VsG VsL Re
3 o
(psia) Qg (Mscf/d) Qg (Mscf/d) (BPD) (lbm/ft ) ( F) (bbl/MMscf) (m/s) (m/s)
60 35.00 139.47 95.69 7.61 0.1814 41 54.57 22.66 0.0069 95,078
61 35.19 140.35 95.95 7.61 0.1823 82 54.23 22.81 0.0069 95,562
62 38.66 60.46 101.09 3.81 0.2003 79 62.94 9.82 0.0035 42,386
63 38.66 66.54 101.09 7.61 0.2003 78 114.39 10.81 0.0069 54,420
64 39.31 93.61 102.07 7.61 0.2037 49 81.31 15.21 0.0069 69,566
65 40.31 109.64 103.61 2.54 0.2088 50 23.14 17.82 0.0023 66,678
66 40.31 92.08 103.61 12.69 0.2088 52 137.76 14.96 0.0116 80,088
67 40.68 106.54 104.19 3.81 0.2108 54 35.72 17.31 0.0035 67,946
68 40.68 107.49 104.19 7.61 0.2108 55 70.81 17.47 0.0069 77,303
69 40.86 119.77 104.47 8.67 0.2117 52 72.42 19.46 0.0079 86,577
70 41.16 53.19 104.95 3.81 0.2133 80 71.55 8.64 0.0035 38,342
71 41.76 62.51 105.89 17.35 0.2164 52 277.54 10.16 0.0158 72,860
72 43.19 92.20 108.19 3.81 0.2238 76 41.27 14.98 0.0035 60,002
73 43.19 103.96 108.19 7.61 0.2238 74 73.21 16.89 0.0069 75,336
74 45.13 86.17 111.37 2.54 0.2338 76 29.44 14.00 0.0023 53,683
75 45.13 99.28 111.37 5.07 0.2338 79 51.11 16.13 0.0046 66,882
76 45.13 75.50 111.37 7.61 0.2338 77 100.82 12.27 0.0069 59,446
77 45.13 97.52 111.37 10.15 0.2338 75 104.07 15.85 0.0093 77,497
78 45.13 86.22 111.37 12.69 0.2338 74 147.14 14.01 0.0116 76,761
79 47.19 110.83 114.88 3.81 0.2445 71 34.34 18.01 0.0035 70,323
80 47.19 113.27 114.88 7.61 0.2445 67 67.20 18.41 0.0069 80,516
81 48.31 119.03 116.82 2.54 0.2503 56 21.32 19.34 0.0023 71,872
82 48.31 135.62 116.82 5.07 0.2503 54 37.41 22.04 0.0046 87,033
83 48.31 112.56 116.82 7.61 0.2503 52 67.62 18.29 0.0069 80,123
84 48.31 132.72 116.82 10.15 0.2503 52 76.46 21.57 0.0093 97,165
85 49.19 124.93 118.38 3.81 0.2549 67 30.46 20.30 0.0035 78,135
46
Table A.2: Critical Flow Rates Comparison
Tubing Measured New Model Scalar Turners Colemans Lis
Data Pressure Critical Rate Critical Rate Multiplier Critical Rate Critical Rate Critical Rate
(psia) Qg (Mscf/d) Qg (Mscf/d) () Qg (Mscf/d) Qg (Mscf/d) Qg (Mscf/d)
1 14.62 55.06 64.17 3.725 99.55 82.96 38.96
2 14.82 55.06 64.61 3.710 100.23 83.52 39.22
3 15.01 55.06 65.02 3.697 100.87 84.06 39.48
4 15.01 55.06 65.02 3.697 100.87 84.06 39.48
5 15.01 55.06 65.02 3.697 100.87 84.06 39.48
6 15.02 55.06 65.04 3.697 100.90 84.08 39.49
7 15.11 55.06 65.24 3.691 101.21 84.34 39.61
8 15.11 55.06 65.24 3.691 101.21 84.34 39.61
9 15.11 55.06 65.24 3.691 101.21 84.34 39.61
10 15.18 55.06 65.39 3.686 101.44 84.53 39.70
11 15.18 55.06 65.39 3.686 101.44 84.53 39.70
12 15.22 55.06 65.47 3.683 101.57 84.64 39.75
13 15.42 55.06 65.90 3.671 102.24 85.19 40.01
14 15.42 55.06 65.90 3.671 102.24 85.19 40.01
15 16.22 55.06 67.59 3.622 104.85 87.37 41.03
16 17.02 55.06 69.23 3.579 107.40 89.50 42.03
17 20.11 55.06 75.25 3.449 116.74 97.28 45.68
18 20.21 55.06 75.43 3.445 117.03 97.52 45.80
19 20.21 55.06 75.43 3.445 117.03 97.52 45.80
20 21.11 55.06 77.09 3.417 119.60 99.66 46.81
21 21.81 55.06 78.36 3.397 121.57 101.30 47.57
22 22.31 55.06 79.25 3.384 122.95 102.45 48.12
23 22.81 55.06 80.13 3.372 124.32 103.60 48.65
24 23.07 55.06 80.59 3.366 125.02 104.18 48.93
25 23.77 55.06 81.80 3.352 126.90 105.75 49.66
26 24.07 55.06 82.31 3.346 127.70 106.41 49.97
27 24.77 55.06 83.50 3.333 129.54 107.95 50.69
28 25.37 55.06 84.50 3.323 131.10 109.24 51.30
29 25.66 55.06 84.98 3.319 131.84 109.86 51.59
30 25.66 55.06 84.98 3.319 131.84 109.86 51.59
31 25.66 55.06 84.98 3.319 131.84 109.86 51.59
32 25.66 55.06 84.99 3.319 131.85 109.87 51.60
47
Table A.2 continued
48
Table A.2 continued
49
APPENDIX B
VELOCITY EQUATIONS
Turner et al.3 analyzed two physical models to determine the minimum flow rate
sufficient to remove liquids (condensed hydrocarbon gas and/or water) from gas wells.
These are the liquid film movement along the walls of the pipe and liquid droplets
entrained in the high velocity gas core, as depicted in Figure B.1. The analysis of the
Drag
Gravity
A droplet is subjected to gravity force (FG) and drag force (FD). The downward gravity
3
g
FG = ( l g )d ....B1
gc 6
50
The upward drag force is as follow:
1
g C d Ad (V g Vd ) .....B2
2
FD =
2g c
( l g )d
Vt = (V g Vd ) = 6.55 ..B3
Cd g
Hinze12 showed that the droplet diameter (d) can be expressed in terms of the
Vt 2 g d
N WE = ..B4
g c
Hinze posited that the largest droplet occurs at NWE in the value range 20-30 and at NWE
greater than 30, the droplet will shatter. Hence the expression for the largest droplet
g c
d = 30 .B5
g Vt 2
Reynolds number ranging from 1,000 to 200,000 (Newtons Law region), as shown in
Figure 4.4. Having the interfacial tension ( ) unit changed from pound force per foot
51
The gas density expression is:
PM P( M a g ) g P
g = = 2.715 ....B7
zRT zRT zT
P
g = 2.715 0.6 = 0.0031P ..B8
(460 + 120) 0.9
Typical values of density and interfacial tension for both water and gas condensate are:
w = 67 lbm/ft3
condensate = 45 lbm/ft3
w = 60 dyne/cm
condensate = 20 dyne/cm
freely falling body in a fluid medium under the influence of gravity alone and substituting
the typical values of density and interfacial tension together with Equation B8 into
Equation B6, the Turner et al.3 and the Coleman et al.4 equations are expressed as:
Vc = Vt (max) .....B9
Vc , w = 5.304
(67 0.0031P ) 4 ....B10
1
(0.0031P ) 2
1
Vc ,cond = 4.03
(45 0.0031P ) 4 ...B11
1
(0.0031P ) 2
52
Coleman et al.4 equations (without any adjustment to fit data):
1
Vc , w = 4.434
(67 0.0031P ) 4 ....B12
1
(0.0031P ) 2
1
Vc ,cond = 3.369
(45 0.0031P ) 4 .....B13
1
(0.0031P ) 2
The critical velocity is easily translated to critical rate using the following expression:
3.06 PAtVc
q= .......... B14
(T + 460) z
53
APPENDIX C
Li, Li, and Sun6 argued that a pressure difference exists between the fore and aft
portions of a free falling droplet. The droplet is thus deformed, under the applied forces,
from a sphere to a convex bean (or a flat-shaped droplet). This is depicted in Figure C.1.
Equating the gravity of a liquid droplet to the buoyancy and drag force, we obtain (in S.I
Units):
1
l gV = g gV + g v 2 sC d ........C1
2
Where:
54
l = liquid density, kg/m3
v = velocity, m/s
Cd = drag coefficient
As the droplet falls relative to the gas stream velocity, the pressure difference
existing between the fore and aft portions, by Bernoulli, is expressed as:
10 3 g
p = v 2 .....C2
2
The condition for the balance of pressure and interfacial tension forces is:
p
3 sh + s = 0 ....C3
10
For a constant volume of the droplet, the following conditions must be satisfied:
V = sh = constant...C4
Where:
55
From Equation C4, we have:
s V sh s
= 2 = 2 = ...C6
h h h h
By combining Equations C5 and C6, we calculate the thickness of the droplet as:
10 3
h= .C7
p
2
h= ..C8
gv2
g v 2V
s= ....C9
2
4( l g )g
vc = v = 4 ....C10
g2 C d
For a typical field condition, the particles Reynolds number ranges from 104 to
105 and the drag coefficient (Cd) is close to 1.0 for the shape considered. Hence, we have:
1 1
vc = 2.5
4
( )
l g
4
1
......C11
( ) g
2
Apvc
qc = 2.5 10 5 .....C12
zT
The density of gas is determined by the expression:
gp
g = 3.4844 ....C13
zT
56
With T measured in degree Kelvin.
For a gas gravity ( g ) of 0.6 and within the pressure range (4,500 23,000 kPa
or 653 3,350 psia) of Lis field data, the average z is 0.85. The typical wellhead
temperature is given as 322oK. Substituting these data into Equation C13, we obtain:
g = 0.00764 p ...C14
Taking water density ( l ) of 1074 kg/m3, a typical value of interfacial tension between
water and gas ( w ) of 0.06 N/m and z value of 0.85 into Equations C11 and C12
respectively, we obtain:
1
1.2373(1074 0.00764 p ) 4
vc = ....C15
(0.00764 p )12
1
At p(1074 0.00764 p ) 4
q c = 0.113 1
10 4 .......................................................C16
(0.00764 p ) 2
Where:
At = area of tubing, m2
57
APPENDIX D
Nosseir et al.7 assumed the droplet model and analyzed the downward
gravitational force and the upward drag force on the model. The gravitational force,
d 3
Fg = ( l g )g c ...D1
6
The drag force pushing the droplet upwards is expressed by the term:
C d lVc2 d 2
Fd = .....D2
2 4
carried out. For laminar flow regime, Stokes, in 1851, introduced his equation for
24 g
Cd = .......D3
dVc l
Combining Equations D1 and D4, the terminal velocity equation reduces to Stokes Law:
58
g c d 2 ( l g )
Vc = ...D5
18 g
Stokes Law applies better to small impure fluid particles, which is probably the case in
gas wells.
0.625
C d = 30(Re ) ...D6
The turbulent flow regime extends from a Reynolds number (Re) of about 1,000
to 200,000 where the drag coefficient is fairly constant (Figure 4.4) at approximate Cd of
0.44. The terminal velocity equation can be described by one form of Newtons Law:
0.5
( l g )
Vc = 1.74d 0.5 g c .......D8
g
dominance of turbulent flow regime in a certain range, which determines the shape of the
drag coefficient and subsequently the critical velocity equation. They showed that most
of Turners data points fall in the highly turbulent region where Re exceeds the value of
59
200,000. This region is termed the highly turbulent flow regime with Re extending from
200,000 to 106 with an average drag coefficient of 0.2. This is shown in Figure D.1.
Nosseir et al.7 again confirmed that most of Colemans data points fall in the Re range of
1000 to 200,000 and its corresponding Cd of 0.44. They argued that this is the reason why
Turners initial equation has a good match for Colemans data without adjustment.
Nosseir et al.7 went on to develop two analytical models, employing the hard,
smooth, spherical droplet of liquid, for the transition and highly turbulent flow regimes.
Equation D7:
60
1.18
30g
0.2 2 c
0.72
Vc g ( l g )
Vc = 0.45
* gc
...... (D9)
g
g
g
0.72
g c ( l g )
* 1.18 / 1g.18Vc2.36 / ( g / g ) ........ (D10)
0.45
Vc = K *
g
Where:
0.21
1 l g
* 0.35 / g0.35 / ( g / g )
0.134
Vc3.36 = K 3.36 * ...... (D12)
g
14.6 0.35 ( l g )
0.21
Vc = ....... (D13)
g0.134 g0.426
Equation D13 is the Nosseir et al.7 critical velocity equation for flow under transition
flow regime.
The derivation of critical velocity equation for the highly turbulent flow regime
follows equally. Starting from Equations D1 and D2, and substituting for maximum
454.67( l g ) 0.5
0.5
Vc = ........ (D14)
gVc
61
21.3 0.25 ( l g )
0.25
Vc = ...... (D15)
g0.5
62