Liquid Holdup in Large-Diameter Horizontal Multiphase Pipelines
Liquid Holdup in Large-Diameter Horizontal Multiphase Pipelines
Liquid Holdup in Large-Diameter Horizontal Multiphase Pipelines
Maley
W. P. Jepson
NSF, I/URC Center for Mulliphase
Technology,
Department of Chemical Engineering,
Ohio University,
184 Stocker Center,
Athens,OH 45701-2929
Introduction
Slug flow is a common flow regime in long-distance multi~hase pipelines. When the slug front moves through the pipe,
It overruns the slower moving liquid film and accelerates the
film to the velocity of the slug. During this process a mixing
vorte:<-is creat~d, called the mixing zone. The mixing zone
entrams a conSIderable amount of gas which is released in the
form of pulses of bubbles. The pulses of bubbles are shot toward
~e bottom. of the pipe where they can impact and collapse.
FIgure 1 shows the different regions of the slug. The bubble
collapse is believed to increase. the corrosion rate and reduce
the efficiency of corrosion inhibitors in slug flow. Green et al.
( 1990) have determined that slug flow causes a higher corrosion
rate than wavy/annular. Slug flow can be eliminated if the
production is reduced or the gas flow rate is increased. If the
production is reduced, a stratified flow regime occurs, which
reduces corrosion rate, but reduces the amount of product. If
the gas velocity is increased by injecting gas into the pipeline,
an annular flow regime occurs. Corrosion rates are reduced, but
erosion rates can increase, especially where sand is present
within the mixture.
'
This project u~es a stationary slug to study the liquid holdup
and length of the mixing zone in a two-phase mixture. Jepson
(198~) has s~own that the slug characteristics in the mixing
zone m a statIonary slug are equivalent to moving slugs for the
same film Froude numbers. Therefore, this study used a stationary slug for the experiments. The stationary slug allowed the
slug characteristics within the mixing zone of the slug to be
studied at different distances. Four different liquids were used
in this study, water, Conoco LVT200, and two different singlephase mixtures of Britol and Conoco LVT200.
Literature Review
. Liquid Holdup.
Gregory et al. (1978) developed a correlation for liquid holdup within the slug. The experiments were
performed using a light refined oil and air in 2.S8-cm and 5.12cm pipes. To determine the liquid holdup in the slug, they used
a capaci.tance-type liquid volume fraction sensor. The following
correlatiOn was developed for liquid holdup, HL, within the
slug:
[Y!1...]
(1)
1.39
8,66
where
VM
where VSL is the superficialliquidvelocity and VSG is the superficial gas velo~ity. This model gives the average liquid holdup
across the entire slug body. The majority of the gas is entrained
within the mixing zone of the slug, therefore, the value obtained
?y using Equation 1 will give a higher liquid holdup value than
If only the average liquid holdup within the mixing zone was
studied.
.
Fershneider (1983) also studied liquid holdup in a 0.146-mdia pipeline using an optical probe. The pressure varied between
10 an.d 50 bars at ambient temperature. The. superficial liquid
velOCItyranged from 0 to 3 mis, and the superficial gas velocity
ranged from 0 to 7 mI s. The model developed by Fershneider
agreed with the data from Gregory et al. ( 1978). Andreussi and
Bendiksen (1989) used air and water to develop a correlation
for liquid holdup in horizontal and near horizontal pipelines.
They used 5-cm and 9-cm i.d. pipelines ranging from an inclination of- 3 to :to.S deg. The data that were collected were used
with Grego~ et al. (1978) and Fershneider (1983) to develop
the correlatiOn. The model used the pipe diameter, inclination,
and fluid properties to determine the liquid holdup. However,'
there were empirical coefficients that were not clearly defined.
Jepson and Kouba (1987) used stationary slugs to perform
liquid holdup experiments in a IS-cm pipeline using air and
water. They determined that the liquid holdup decreased linearly
as the film Froude number increased. The film Froude number
is defined in Eq. (6). Jepson and Taylor (1988) performed
experiments using air and water in a 30-cm pipeline. They
d~te~ned that the liquid holdup also was dependent upon the
pIpe dIameter for a gas velocity above 3 mls.
Gopal (1994) studied the liquid holdup as a function of distance into the slug. He studied two phase slug flow by using
water and carbon dioxide mixture, and ARCOPAC90 and
carbon dioxide mixture in a 7.S-cm pipeline at 298 K and 0.101
MPa. Gopal determined that the liquid holdup can be modeled
by using a second order process dynamic system. The differential equation which describes this system is
2
where
d2y.
dY
is a time constant,
X(t)
e is the damping
(3)
ratio, t is time,
Mixing Zone
Gas Pocket
Liquid
Film
_00
VIC
D
0
0000
Fig. 1
E
0
x (t)
J. 10 em Plexiglus Section
M. Heater
N. Pump
LMZ
T=--
Regulator
(4)
Fig. 2
Experimental layout
F
rf
between 5 and 10. He found that when the film Froude number
increased, the length of the mixing zone increased linearly. The
correlation that he developed, with LMZ in meters is
LMZ = 0.13 Frf - 0.31
v, = )gh
Vlf
(6)
ef
(7)
Experimental Setup
Description of the Flow Loop.
The experiments were performed in the system shown in Fig. 2. The experimental layout
is 18-m long and is made from PVC and acrylic pipe. The liquid
is stored in a 1.3 m3 stainless steel tank, and using a 7A-kW
centrifugal pump, the liquid is pumped into a 7.6-cm i.d. PVC
pipeline. The flow rate of the liquid was controlled using a
pump variable speed drive. The liquid then flows through an
orifice plate where the pressure drop is measured with an Omega
differential pressure transducer with a range of 0 to 5 psi. The .
liquid then flows into a 10.I-cm i.d. acrylic pipeline where the
flow is forced under a gate, and a fast moving liquid film is
produced. Three different gate geometries were used in this
study. These had ratios of hiD = 0.28, hiD = 0.33, and hiD
= OAO, respectively.
Nitrogen is then introduced into the system. The nitrogen
flows from a 5,000 ft3 storage tank to a pressure regulator at a
pressure of 150 psi. The flow rate is controlled by a ~in. stainless steel needle valve and the gas is injected immediately after
the gate. This forms a gas pocket and a hydraulic jump, or slug,
is created. The gas flow is adjusted and the slug is moved to
the test section and held there. Measurements are then taken.
Nomenclature
A
Eo
Fr
R
LMZ
Re
X
= area of pipe
= Etovos no.
= Froude no.
= holdup
= length of mixing zone
= Reynolds no.
= step input
= response of second-order
system
1998
Subscripts.
M = mixture
SL = superficial liquid
SG = superficial gas
f = film
If = liquid film
t = translational
Transactions
of the ASME
low film Froude numbers, and :t 15 cm for the high film Froude
numbers.
The second method is based upon the liquid holdup results,
and uses the fact that the average void fraction becomes constant
once the end of the mixing zone is passed. Therefore, the measurements were taken up to and 30 cm past the visual length.
A more accurate end of the mixing zone was then determined
when the average liquid holdup became constant.
FLOW DIRECTION
Uquid
Holdup
Sample
Port
D. PressureTappiags
E. Shen Stress Port
Fig.3
Test section
The liquid/gas mixture then passes into the tank where the
liquid is recycled and the gas is vented to the atmosphere.
The test section, shown in Fig. 3, is 10.1 cm in diameter and
2 m long and is manufactured from acrylic pipe with a 0.64cm wall thickness. The measurements for this study were taken
within "the test section. The temperature of the flow was measured by using a type-K thermocouple which is connected to
an OMEGA DP3200- TC thermometer. The temperature was
held constant at 40C by using a Wiegand heater with a Wellman thermal system thermostat. The pressure was maintained
at 0.136 MPa within the test section by setting the back pressure
in the storage tank.
Liquid Holdup.
Each liquid holdup sample was withdrawn
from the slug by using a 316 stainless steel sampling tube. The
sample then flowed into a calibrated tube after simultaneously
opening the inlet and outlet valves to the calibrated tube. With
the slug held stationary within the test section, an isokinetic
sample was withdrawn at a known vertical distance into the
slug. At each film Froude number, samples of the flowing mixture, on the center line of the slug, were withdrawn at known
distances into the slug.
The number of axial sampling distances was determined by
the length of the mixing zone. The method of determining the
length of the mixing zone is described in the next section. For
each axial point, samples were taken at five different vertical
heights across the pipe along the centerline of the pipe. After
performing this experiment at five differential vertical heights
at each axial location, an average void fraction was then calculated over the cross section of the slug by the following equation:
LtXidA,
=----
(8)
Liquid Holdup.
The liquid holdup at a known distance into
the slug was measured at five different locations across the
cross section of the pipe. Figure 4 shows an example of the
local liquid holdup at each of the five locations for a film Froude
number of 7.4 for the 16.6 cP mixture. This figure shows that
at 7.1 cm from the top of the pipe the liquid holdup is at the
maximum value because the sampling probe has penetrated the
liquid film. This figure also shows how the liquid holdup values
from each of the locations range from 0.29 to 0.84 at a distance
8 cm into the slug and converge to 0.60 to 0.76 before the end
of the mixing zone.
The slug front is constantly changing shape due to the high
turbulence, therefore, it was difficult to obtain accurate data any
closer than 15 cm into the slug front. However, at low film
Froude numbers the slug front is not as turbulent as at high film
Froude numbers, and a few data points at 8 cm into the slug
were obtained." The slug was then moved down the pipe to
another known distance and the experiment was repeated until
the end of the mixing zone was reached. After obtaining the
experimental data, the average liquid holdup was calculated by
using a weighted average across the cross section of the pipe.
A curve fit was performed through the data to show that the
liquid holdup at zero distance into the slug is equal to the
nondimensional area of the slug. Video images taken by Gopal
et al. (1995) show that the liquid holdup in the slug front is
equal to that of the liquid film.
The first fluid that was studied was 100 percent water with
a nondimensional liquid gate and film height of 0.33 for a range
of film Froude numbers from 4.8 to 18. The nondimensional
area of the liquid film, the ratio of the area of the liquid film
to the cross-sectional area of the pipe, was 0.29. Since, the
liquid holdup is defined as the area of liquid divided by the
area of the pipe, the nondimensional area is equal to the liquid
holdup of the liquid film before the slug. Therefore, the liquid
holdup at zero distance into the slug should be 0.29 for this
fluid. The void fraction was plotted against the distance into
the slug, as shown in Fig. 5, for a film Froude number of 8.6.
The mixing zone length was determined to be 65 cm, and the
liquid holdup at zero distance into the slug is 0.31, which is
1.00
0.80
'"=
:ll
0.60
'0
.;
.,.
;:;
~
...
:!l
.4l
0.40
..
:4l
I
+
EndOf
Mixing Zone
Distance from
top of pipe
+
f',
O.5em
2.1em
3.8em
S.4em
7.lem
0.20
0.00
0.00
0.25
0.50
Distance
0.75
1.00
1.25
1.50
Fig. 4 Local holdup versus distance into the slug for a film Froude
number 7.4 for 16.6 cP mixture
SEPTEMBER
1.00
1.00
0.80
0.80
".""0=
0.60
0.60
;:;
0.40
:>::
."
.":;
or
End of
MixingZone
End of
Mixing Zone
0.40
0.20
0.20
0.00
0.00
0.25
0.50
0.75
Distance
1.00
1.25
1.50
0.25
0.50
0.75
Distance
Fig. 5 Liquid holdup versus distance into the slug for film Froude number 8.6 for 100 percent water and a nondimensionalliquid
film area of
0.29
1.00
1.00
1.25
1.50
Fig. 7 Liquid holdup versus distance into the slug for film Froude number 13.8 for 100 percent water and a nondimensionalliquid
film area of
0.37
1.00
0.80
.""0"-=
0.60
.'
T'
:>::
."
'.
or
;:;
..L
0.40
End Of
Mixing Zone
..L
0.20
0.00
0.00
0.25
0.50
0.75
Distance
1.25
1.00
1.50
Fig. 8 Liquid holdup versus distance into the slug for film Froude number 18.6 for 100 percent water and a nondimensional liquid film area of
0.23
0.60
..
in=
0 0.23
0.50
,:
.~
B
-=
=
:E"-=
0.29
0.37
~ ~
0.40
~
+
0.30
+
0
0
."==
'.
C'
;:;
0.20
0.80
0.10
10
0.60
"'TT=
:=
0.40
~ '.
-l
Film Fronde
12
14
16
18
20
Number
Fig. 9 Liquid holdup at 0 cm into slug versus film Froude number for
varying ratios of aLFIA for 100 percent water
End of
..L
Mixing Zone
0.20
0.00
0.00
0.25
0.50
0.75
1.00
1.25
1.50
Fig. 6 Liquid holdup versus distance into the slug for film Froude number 16.4 for 100 percent water and a nondimensionalliquid
film area of
0.29
of the ASME
1.00
1.00
0.80
0.80
Q.
."=
c;
0.60
! .!
;:;
or
End
End or
Mixing Zone
.".;
C'
0.40
0.60
:ll
0.40
Mixing. Zone
0.20
0.20
0.00
0.00
0.00
0.25
0.50
0.75
Distance
1.25
1.00
I
I
0.25
0.75
0.50
Distance
Fig. 10 Liquid holdup versus distance into the slug for film Froudenumber 5.7 for 100 percent LVT and a nondimensionalliquid
film area of 0.31
0.00
1.50
1.00
1.25
1.50
Fig. 12 Liquid holdup versus distance into the slug for a film Froude
number of 5.8 for 10.9 cP mixture and a nondimensionalliquid
film area
of 0.31
1.00
0.80
Q.
."c;=
0.60
:ll
.".;
C'
;:;
End or
MixingZone
0.40
0.20
0.00
0.00
0.25
0.50
Distance
0.75
1.00
1.25
1.50
Fig. 13 Liquid holdup versus distance into the slug for film Froude number 12.5 for 10.9 cP mixture and a nondimensional area of 0.31
1.00
0.80
Q.
:!!
0.60
:=.".;
C'
;:;
End or
Mixing Zone
0.40
,I
1.00
0.20
0.80
0.00
0.00
."=
.L
End or
Mixing Zone
0.00
0.00
0.25
0.50
Distance
0.75
1.00
1.25
1.50
Fig. 11 Liquid holdup versus distance into the slug for a film Froude
number of 17.5 for 100 percent LVT and a nondimensional liquid film
area of 0.31
.
!.
1.00
1.25
1.50
0.40
0.20
0.75
Fig.14
Liquid holdup versus distance into the slug for film Froude number 7.4 for 16.6 cP mixture and a nondimensionalliquid
film area of 0.36
.".;
;:;
0.50
0.60
~
C'
0.25
Distance
Q.
Technology
1.00
:.50
0.80
1.20
:!i!
0.60
".
..
=.;
"I:l
;:;
~..
""
0.90
.~
End Of
Mixing Zone
:;
..;;,"
0.40
0.60
:J
0.20
0.30
0.00
0.00
0.25
0.00
0.50
0.75
Distance
1.00
1.25
1.50
..
..
.E"
0.50
+
a
0040
0.36
/}.
;;
/}.
0.30
/}.
0.31
/}.LVT
"I:l
.;
-;:;
".
14
16
20
18
""=
:!i!
Fig. 18
12
Number
0.60
10
Film Froude
Fig. 15 Liquid holdup versus distance into the slug for film Froude number 9.5 for 16.6 cP mixture and a nondimensionalliquid
film area of 0.36
1i'i
oS
+
a
0.20
10.9 cP Mixture
16.6 cP Mixture
0.10
4
10
12
14
16
18
20
(9)
0.18
Fig. 16 Liquid holdup at 0 cm into the slug versus film Froude numbers
for varying ratios of aLE! A
1.00
0.80
""=
"I:l
0.60
"I:l
.;
".
+
+
++
+
++
"t- +
-q.++'+
;:;
"
0.40
~""
+
+
+
0.20
+
+
t+
0.00
0.00
0.20
Fig.17
0.60
0.40
Experimental
Average
0.80
1.00
Liquid Holdup
Froude number of 9.5 are shown in Fig. 15. The liquid holdup
was 0.51 at 15 cm and increased to 0.69 at 60 cm into the slug.
The length of the mixing zone was 70 cm and the liquid holdup
at zero distance into was 0.38.
Figure 16 shows the liquid holdup at zero distance into the
slug for each film Froude number for the three different oils
studied. This plot shows that the extrapolated intercept for the
liquid holdup versus distance into the slug closely resembles
the nondimensional area before the slug for each of the oil based
slugs studied. Figure 17 compares the experimental results to
the results obtained from Gopal' s model. This figure shows that
190 I Vol. 120, SEPTEMBER
1998
Experimental
Transactions
length
of the ASME
2.40
o
o
2.10
~~
g
..
1.80
1.50
=
';<
1.20
~
Q
.c
0.90
:l=
0.60
DISCUSSION
0.30
G. E. Koubal
0.00
0.00
0.20
0.40
Experimental
0.80
0.60
Length
1.00
1.20
1.40
Fig. 20 Model's and Gopal's length of the mixing zone versus experimental length of mixing zone
Conclusions
Slug flow characteristics were studied using water, LVT, a
1O.9-cP and a 16.6-cP single-phase mixture of Britol and LVT
with nitrogen as the gas phase. The experiments were conducted
in a 10.I-cm pipeline at a constant temperature of 40C and a
constant pressure of 0.136 MPa. The nondimensionalliquid film
height before the slug ranged from 0.28 to 0040. A stationary
slug was used in all experiments.
The results from this study show that as the film Froude
number increases the amount of gas entrained in the slug at any
given distance also increases. When the liquid film height before
the slug increases, so does the liquid holdup at any given distance into the slug. The data also suggest that the higher the
viscosity of the fluid, the less gas the slug will entrain at the
same film Froude number, and the liquid holdup at the slug
front is equal to the non dimensional area of the liquid film
height before the slug.
The length of the mixing zone was estimated for each film
Froude number by using a visual means of determining the
length of the turbulent part of the slug. The length was then
determined by studying the liquid holdup data. The end of the
mixing zone was determined to be the point where the average
liquid holdup became constant. The length of the mixing zone
was then plotted against the film Froude number and a linear
relationship was observed. There was no effect on the length
of the mixing zone when the viscosity of the liquid inside the
slug increased, for a viscosity range from 1 to 16.6 cPo An
empirical correlation for the length of the mixing zone was
developed.
References
Andreussi, P., and Bendiksen. K., 1989, "An Investigation of Void Fraction
in Liquid Slugs for Horizontal and Inclined Gas-Liquid Pipe Flow," International
Journal of Multiphase Flow, Vol. 6, p. 937.
Andreussi, P., Bendiksen. K., and Nydal, O. J., 1993, "Void Distribution in
Slug Flow," International Journal of Multiphase Flow, Vol. 5, p. 817.
Dukler, A. E., and Hubbard, M. G., 1975, "A Model for Gas-Liquid Slug Flow
in Horizontal and Near Horizontal Tubes,"
Industrial Engineering
Chemistry
Fundamentals.
Vol. 14, NO.4, p. 337.
Ferschneider, G., 1983, Revue de L'lnstitut Francais du Petrole, Vol. 38, No.
2, p. 153.
Gopal, M., 1994, "Visualization
and Mathematical
Modelling of Horizontal
Multiphase Slug Flow" dissertation, Ohio University, Aug.
Gopal, M., and Jepson, W. P., 1997, "The Study of Dynamic Slug Flow Characteristics Using Digital Image Analysis,"
ASME JOURNALOF ENERGY RESOURCES
AND TECHNOLOGY.
Gopal, M., Kaul, A., and Jepson, W. P., 1995, "Mechanisms
Contributing to
Enhanced Corrosion in Horizontal Three-Phase Oil, Water, and Gas Slug Flow,"
NACE Paper No. 105.
Green, A. S., Johnson, B. V . and Choi, H., 1990, "Flow-Related
Corrosion in
Large-Diameter
Multiphase Flowlines,"
SPE 20685, p. 677.
Gregory, G. A., Nicholson, M. K .. and Aziz, K., 1978, "Correlation
of the
Liquid Volume Fraction in the Slug for Horizontal Gas-Liquid Slug Flow,"
International Journal of Multiphase Flow, Vol. 4, p. 33.
Journal
of Energy Resources
Technology
2
3
Redistribution
Region.
The magnitude of the large-scale
turbulence has declined dramatically by the end of the entrainment region and buoyancy begins to dominate the motion of
large bubbles, forcing the redistribution of gas. In the redistribution region, the large gas bubbles migrate upward. The size of
bubbles that can be entrained by the decaying turbulence stead-
1 Chevron
Petroleum Technology
vard, La Habra, CA 90631-6374
Company.
SEPTEMBER
Beach Boule-
Gas redistribution
Slug tail
Slug body
region
Fig. 21 Regions of mixing zone in liquid slug. Two counterrotating
gas redistributes as turbulence decreases.
eddies entrain gas and liquid at the slug front. Further back
1.4
1.2
Model
o Maley Correlation
0.8
0.6
+O.2m
0.4
0.2
-O.2m
where
his
v,
Vm
Vb
= 1.53 [
J'4
(T~~P
(12)
h,,1vb*(v,
- vm)
(13)
- Vm)IVb
(14)
LmID;:;,,: 1.5
hi,lD*(v,
his
Fr,iii)
(15)
0
0
0.2
0.4
0.6
1.2
0.8
1.4
Vb
Results
The mechanistic model of the MJ mixing length can be represented by any of the equivalent forms given by Eqs. (13), ( 14),
or (15). Predictions of mixing length from the mechanistic
model and the MJ correlation are compared against MJ data in
1998
Conclusions
Maley and Jepson's investigation into holdup in slug flow
yields insight into the mixing region. A simple mechanistic
model has been developed that is consistent with observations
and measurements of the mixing region.
Additional References
Dulder, A. E., and Hubbard, M. G .. 1975, Industrial Engineering
Fundamentals, Vol. 14, NO.4, pp. 337-347.
Harmathy, T. Z., 1960, AIChE Journal, Vol. 6, p. 281.
Transactions
& Chemistry
of the ASME