Professional Documents
Culture Documents
Experimental Thermal and Fluid Science
Experimental Thermal and Fluid Science
Erosion/Corrosion Research Center, The University of Tulsa, 800 South Tucker Drive, Tulsa, OK 74104, United States
Thermal Science Department, University of Los Andes, Mrida, Mrida 5101, Venezuela
c
Helmholtz-Zentrum Dresden-Rossendorf (HZDR), Dresden, Saxony 01328, Germany
b
a r t i c l e
i n f o
Article history:
Received 17 December 2013
Received in revised form 25 June 2014
Accepted 2 August 2014
Available online 12 August 2014
Keywords:
Wire-mesh sensor
Gasliquid ow
Elbow
Void fraction
Phase distribution
Flow visualization
a b s t r a c t
Fluid owing through pipelines often encounters ttings such as elbows. Although it is true that twophase ow patterns observed in elbows are qualitatively the same as those seen in straight pipes, the
presence of a pipe elbow can modify relative positions and local velocities of the two phases as they
are subjected to forces in addition to those encountered in a straight pipe. That redistribution can affect
pressure drop values, chemical inhibitor concentration and distribution to the top of the pipe, as well as
the erosion pattern occurring from solid particles such as sand that is entrained in oil and gas transportation pipelines. In this work, a wire-mesh sensor technique based on conductance measurements of void
fraction was applied to investigate two-phase pipe ow through a standard elbow. The horizontal ow
test section, consisting of a 76.2 mm ID, 18 m long pipe, was employed to generate stratied-wavy and
annular ow conditions. Two 16 16 wire-mesh conguration sensors were positioned either 0.9 m
upstream or 0.6 m downstream of the bend. The experiments were conducted at supercial liquid velocities equal to 0.03 m/s and 0.2 m/s and supercial gas velocities that ranged from 9 m/s to 34 m/s. The
effects of liquid viscosity on the measured parameters are also investigated using two different viscosities
of 1 and 10 cP. Stratiedslug transition, stratied wavy and annular ow patterns were observed visually in the clear section placed upstream of the wire-mesh sensors. Analysis of time series void fraction
data from the dual wire-mesh sensors allows the determination of mean void fraction, local time average
void fraction distribution, liquid phase distribution around the tube periphery, interfacial structure velocities, as well as probability density function characteristic signatures within the cross-section of pipe
before and after the elbow. The results indicate that the distribution of gas and liquid phases and interfacial velocities are signicantly altered even 20 diameters downstream of the elbow.
2014 Elsevier Inc. All rights reserved.
1. Introduction
Pipe ttings such as valves, bends, elbows and tees signicantly
affect multiphase ow distribution including generation of secondary ows, uctuations in void fractions and pressure losses and
velocities of gas and liquid phases. Elbows are often used in oil
and gas production systems, and they cause redistribution of gas
and liquid which can affect distribution of corrosion inhibitors
within and downstream of the bends. Elbows are also a location
Corresponding author.
E-mail addresses: rev87@utulsa.edu (R.E. Vieira), nrk301@utulsa.edu (N.R.
Kesana), brenton-mclaury@utulsa.edu (B.S. McLaury), siamack-shirazi@utulsa.edu
(S.A. Shirazi), ctorres@ula.ve (C.F. Torres), e.schleicher@hzdr.de (E. Schleicher),
u.hampel@hzdr.de (U. Hampel).
http://dx.doi.org/10.1016/j.expthermusci.2014.08.001
0894-1777/ 2014 Elsevier Inc. All rights reserved.
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
73
Nomenclature
A
cross-sectional area of the pipe (m2)
Br
systematic uncertainty or bias limit
Cxy, Cxx, Cyy covariance functions between x(t) and y(t)
c
coverage factor
D
pipe diameter (m)
E[]
expected value of []
L
pipe length (m)
N
number of samples, ()
n
time level ()
ID
internal diameter
Pr
precision limit
Rxy
StdDev
t
ag
rxy
qxy
s
lL
lx
ly
cross-correlation function
standard deviation
time (s)
void fraction ()
averaged product
correlation coefcient function
temporal lag obtained from cross-correlation (s)
liquid viscosity
mean value of the random process x(t)
mean values of the random process y(t)
74
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
Each wire-mesh sensor consists of two parallel wire layers perpendicular to the pipe axis. The wires are made of stainless steel
316L. The spacing between the layers is 1.5 mm; each layer is built
of sixteen 0.125 mm diameter parallel steel wires. The wires in the
receiver layer are perpendicular to those in the transmitter. Therefore, a rectangular mesh with 256 junctions is obtained; only 192
are considered to be located within the pipe cross section. The
operation principle of the wire-mesh sensor is based on the difference in the electrical conductivity of the two phases (water and
air). Each wire in the transmitter plane is periodically activated
by a multiplexer circuit by a short voltage pulse. An additional
multiplexer circuit is used to connect consecutively each one of
the receiver wires during the single pulse supplied to the transmitter wire. Each output pulse depends on the local instantaneous
conductivity at the crossing of the transmitter and the receiver
wires.
Calibration of the sensor was performed before each experimental run. The output voltages at each junction in a pipe lled
with water and in an empty pipe were recorded. Following Prasser
et al. [22], the variation of the output voltage between these limiting values was assumed to vary linearly with the local instantaneous liquid holdup, thus enabling conversion of the recorded
voltages into void fractions, using the calibration coefcients
determined before the beginning of each experimental run. The
raw data were recorded as a sequence of voltages in each junction.
The voltages were further translated into the local instantaneous
void fraction. The resolution in the axial direction is determined
by the interface propagation velocity, as well as by the sampling
frequency. The instrument is capable of sampling at frequencies
up to 10 kHz for the whole cross section; in all the cases the maximum sampling frequency was used. The wire-mesh sensor thus
provides high-frequency data on the temporal variation of the local
void fraction at each junction of the mesh. Further details about the
operation of the WMS can be found in Prasser et al. [22].
The current studies were performed using the Erosion/Corrosion Research Center (E/CRC) large scale tower-boom loop facility
located at The University of Tulsa. The loop main components are
two diesel compressors, electrical compressor, slurry tank, test section, diaphragm pumps, and separator. The test section in this large
scale facility can be modied to accommodate different pipe diameters and can be oriented between 0 (horizontal) to 90 (vertical).
For these experiments, the pipe diameter was 76.2 mm (3 in.) and
18 m long.
Experiments were performed horizontally and near atmospheric pressure conditions. The uids used were air and either
water or water mixed with Carboxy Methyl Cellulose (CMC) to
increase the liquid viscosity. In order to measure changes on the
75
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
Pr c StdDevag
h
i
C xy E xt lx fyt s ly g
Z
i
1 Th
fxt lx gfyt s ly g dt
lim
T!1 T 0
where c is the coverage factor and StdDev(ag) is the standard deviation of the sample of N readings of the void fraction ag. For N P 10,
c = 2 is assumed. The value of the standard deviation is calculated
from N readings over the specied time interval of the void fraction
Boom
Gas
and
Liquid
N
1X
xi lx yi ly
N!1 N
i1
rxy Ex lx y ly lim
For any time delay s, the covariance function between x(t) and
y(t) is:
Rxy s lx ly
Rxy s lim
Liquid
76.2 mm ID Test Section
1
T!1 T
Test
17 mSections
xtyt sds
Gas
V11
V12
V21
V22
Flow
meter
250 gallon tank
Gas
V1
Collecting
Tank
Stirrer
Pump 1.1)
Rxy s lx ly
C xy s
q
qxy s p
C xx 0C yy 0
Rxx 0 l2x Ryy 0 l2y
Gas &
Liquid
Pump 1.2)
Diesel Compressors
Pump 2.1)
Pump 2.2)
76
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
Dual Wire
Mesh Sensor
Standard
Elbow
Flow
Direction
Table 1
Uncertainty of supercial liquid velocities.
VSL (m/s)
Error (%)
0.03
0.2
8.5
5
Table 2
Uncertainty of supercial gas velocities.
VSG (m/s)
Error (%)
9
18
27
30
32
34
1.8
1.9
2.0
2.1
2.2
2.2
Stratified
Wavy
10
VSL, (m/s)
SLUG
ANNULAR
0.1
STRATIFIED
SMOOTH
0.01
0.001
0.01
Test
(#)
VSG
(m/s)
VSL
(m/s)
lL
ag
()
Pr
()
Br
()
(cP)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
9
18
30
9
18
27
34
9
18
27
34
9
18
25
0.03
0.03
0.03
0.20
0.20
0.20
0.20
0.03
0.03
0.03
0.03
0.20
0.20
0.20
1
1
1
1
1
1
1
10
10
10
10
10
10
10
0.93
0.96
0.98
0.83
0.90
0.94
0.97
0.89
0.96
0.97
0.97
0.82
0.88
0.92
0.02
0.02
0.02
0.14
0.09
0.05
0.03
0.03
0.02
0.02
0.01
0.12
0.12
0.08
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.04
0.04
0.04
0.04
0.04
0.04
0.04
Stratied-wavy
Stratied-wavy
Annular
Stratiedslug transition
Stratied-wavy
Annular
Annular
Stratied-wavy
Stratied-wavy
Annular
Annular
Stratiedslug Transition
Stratiedslug transition
Annular
Annular
Dispersed-Bubble
DISPERSED BUBBLE
Table 3
Experimental test matrix, cross-sectional averaged void fraction measurements and
precision limit of the data; horizontal airwater or airwater & CMC (upstream
section).
0.1
STRATIFIED
WAVY
10
100
VSG, (m/s)
Fig. 5. Investigated area of ow conditions for stratied and annular ow
conditions; 1 cP water viscosity.
Table 4
Experimental test matrix, cross-sectional averaged void fraction measurements and
precision limit of the data; horizontal airwater or airwater & CMC (downstream
section).
Test
(#)
VSG
(m/s)
VSL
(m/s)
lL
ag
()
Pr
()
Br
()
(cP)
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
9
18
27
36
9
18
27
36
9
18
27
36
9
18
27
36
0.03
0.03
0.03
0.03
0.20
0.20
0.20
0.20
0.03
0.03
0.03
0.03
0.20
0.20
0.20
0.20
1
1
1
1
1
1
1
1
10
10
10
10
10
10
10
10
0.94
0.97
0.98
0.99
0.85
0.91
0.96
0.98
0.91
0.95
0.96
0.97
0.82
0.90
0.94
0.97
0.02
0.02
0.02
0.02
0.15
0.07
0.04
0.02
0.03
0.03
0.02
0.01
0.14
0.08
0.04
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
Stratied-wavy
Stratied-wavy
Annular
Annular
Stratied-wavy
Annular
Annular
Annular
Stratiedslug Transition
Stratied-wavy
Annular
Annular
Stratiedslug Transition
Stratied-wavy
Annular
Annular
the pipe can be higher than half the diameter of the pipe. The wave
instability is the most notable feature of the stratied to slug transition. Fig. 8 shows that in the downstream section more liquid is
in the perimeter of the cross-sectional area of the pipe and waves
are smaller as compared with the upstream conditions. As can be
observed for the stratied ow (Fig. 9) at higher gas ow rates,
the holdup is very small. Fig. 9 shows that the average liquid layer
is thinner than the average liquid layer observed in Fig. 7a and c.
The liquid lm thickness decreases with increasing supercial
gas velocities for a xed supercial liquid velocity of 0.2 m/s.
Fig. 10 shows the downstream conditions where the liquid layer
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
1.2
Upstream
Downstream
Rxy()
0.8
0.6
0.4
0.2
0
-0.2
-10
-5
10
Time, (s)
Fig. 6. Cross-correlation sample result between two planes of the dual wire mesh
sensor; stratiedslug transition (VSG = 9 m/s, VSL = 0.2 m/s and, lL = 1 cP).
in the bottom part of the pipe tends to climb the inner and outer
pipe walls in an oscillating pattern.
4.2. Cross-sectionally averaged void fraction
Figs. 1113 show plots of instantaneous cross-section void fraction averages obtained from the output of the WMS placed
upstream and downstream of the elbow for stratiedslug, stratied wavy and annular ows, respectively. The observed stratied
slug transition pattern in Fig. 11 is characterized by the intermittent appearance of pseudo-slugs or highly-aerated slugs which
bridge the entire pipe section. Previous investigations [10,30] also
have identied this phenomenon by the occurrence of signicant
pressure uctuations in the stratiedslug transition. Similar uctuations are observed in the upstream section of the pipe as
observed in the downstream section of the elbow.
Fig. 12 shows that the mean void fraction in the upstream location for stratied ow (ag = 0.90) is smaller than for annular ow
(ag = 0.97) which is shown in Fig. 13. The periodic structures
obtained from the cross-sectionally averaged void fraction time
series show similar structures in the upstream and downstream
locations for stratied and annular ows. In both cases, the
77
(a) t = 10 s
(b) t = 20 s
(c) t = 30 s
(d) t = 40 s
(e) t = 50 s
(f) t = 60 s
Fig. 7. Sequence of frames at 10 s intervals showing the location of the gas and liquid (in blue) for stratiedslug transition in the upstream section of the elbow (VSG = 9 m/s,
VSL = 0.2 m/s and, lL = 1 cP). (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)
78
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
(a) t = 10 s
(b) t = 20 s
(c) t = 30 s
(d) t = 40 s
(e) t = 50 s
(f) t = 60 s
Fig. 8. Sequence of frames at 10 s intervals showing the location of the gas and liquid (in blue) for stratiedslug transition in the downstream section of the elbow
(VSG = 9 m/s, VSL = 0.2 m/s and, lL = 1 cP). (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)
(a) t = 10 s
(b) t = 20 s
(c) t = 30 s
(d) t = 40 s
(e) t = 50 s
(f) t = 60 s
Fig. 9. Sequence of frames at 10 s intervals showing the location of the gas and liquid (in blue) for stratied-wavy ow in the upstream section of the elbow (VSG = 18 m/s,
VSL = 0.2 m/s and, lL = 1 cP). (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)
interfacial shear stress which can oppose gravity and transport the
liquid upwards.
Fig. 18 shows that for annular ows, the presence of the liquid
lm in the upper part of the tube is not clearly detected by the sensor. Several mechanisms have been proposed to explain the presence of very thin lms in the upper part of the tube in horizontal
annular ows, the more important and credible among these are
secondary ow mechanism [31] entrainment and re-deposition
mechanism [32], wave spreading mechanism [33] and pumping
action mechanism [34]. Figs. 19 and 20 show that if the liquid
79
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
(a) t = 10 s
(b) t = 20 s
(c) t = 30 s
(d) t = 40 s
(e) t = 50 s
(f) t = 60 s
Fig. 10. Sequence of frames at 10 s intervals showing the location of the gas and liquid (in blue) for stratied-wavy ow in the downstream section of the elbow (VSG = 18 m/s,
VSL = 0.2 m/s and, lL = 1 cP). (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)
0.8
Void fraction
Void fraction
0.8
0.6
0.4
0.6
0.4
Upstream
Downstream
0.2
0.2
Upstream
Downstream
10
15
20
Time, (s)
0
10
15
20
Time, (s)
Fig. 11. Representative cross-sectionally averaged time series of void fraction for
stratiedslug transition in the upstream and downstream section of the elbow
(VSG = 9 m/s, VSL = 0.2 m/s and, lL = 1 cP).
Fig. 13. Representative cross-sectionally averaged time series of void fraction for
annular ow in the upstream (VSG = 34 m/s), and downstream (VSG = 36 m/s) section
of the elbow (VSL = 0.03 m/s and, lL = 1 cP).
1.00
0.8
Void fraction
0.90
0.6
0.4
Upstream
Downstream
0.2
0
10
15
20
Time, (s)
Fig. 12. Representative cross-sectionally averaged time series of void fraction for
stratied-wavy in the upstream and downstream section of the elbow (VSG = 18
m/s, VSL = 0.2 m/s and, lL = 1 cP).
0.80
0.70
Upstream; Vsl=0.03 m/s
Downstream; Vsl=0.03 m/s
Upstream; Vsl=0.2 m/s
Downstream; Vsl=0.2 m/s
0.60
0.50
0.40
10
15
20
25
30
35
40
80
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
always equal to one, as conrmed by the standard deviation prole. As the liquid viscosity is increased, the standard deviation in
the bottom layers decreases because the liquid lm is traveling
at lower velocities, with fewer bubbles entrained. The analysis of
the standard deviation highlights that for VSG = 9 m/s and
VSL = 0.67 m/s with 10 cP, the presence of highly unstable liquid
lm can be veried by an increase of the standard deviation in
the core region.
1.00
0.90
0.80
0.70
Upstream; Vsl=0.03 m/s
Downstream; Vsl=0.03 m/s
Upstream; Vsl=0.2 m/s
Downstream; Vsl=0.2 m/s
0.60
0.50
4.4. Effect of the elbow on phase distribution close to the upper part of
the inner pipe wall
0.40
10
15
20
25
30
35
40
Voidm
g(%)
95
90
80
80
60
30
20
5
10
95
90
80
60
30
20
5
95
95
90
80
70
60
50
40
30
20
10
5
Voidm
g(%)
95
95
95
95
80
90
70
50
70
30
90
70
80
30
40
20
5
90
80
70
60
50
40
30
20
10
5
Fig. 16. Cross-section local average void fraction for stratiedslug transition (VSG = 9 m/s, VSL = 0.2 m/s and, lL = 1 cP).
95
60
95
95
40
70
20
80
70
40
20
10
80
Voidm
g(%)
Voidm
g(%)
90
80
70
60
50
40
30
20
10
5
90
80
70
60
50
40
30
20
10
5
95
95
90
80
90
60
95
80
40
50 4 0
10
Fig. 17. Cross-section local average void fraction for stratied wavy ow (VSG = 18 m/s, VSL = 0.2 m/s and, lL = 1 cP).
81
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
95
90
80
95
90
90
80
70 80
70
40 30
95
Voidm
g(%)
Voidm
g(%)
90
80
70
60
50
40
30
20
10
5
90
80
70
60
50
40
30
20
10
5
95
90
70
95
95
90
70
Fig. 18. Cross-section local average void fraction for annular ow conditions (VSL = 0.2 m/s and, lL = 1 cP).
90
70
50
20
95
95
95
90
80
90
50
90
70
50
20
5
60
30
Voidm
g(%)
95
95
70
95
80
60
20
5
95
90
80
70
60
50
40
30
20
10
5
95
95
90
50
20
Voidm
g(%)
95
80
60
30
90
80
70
60
50
40
30
20
10
5
Fig. 19. Cross-section local average void fraction for stratiedslug transition (VSG = 9 m/s, VSL = 0.2 m/s and, lL = 10 cP).
Voidm
g(%)
95
95
90
50
90
70
70
20
20
5
95
50
90
80
70
60
50
40
30
20
10
5
20
95
95
95
80
90
90
60
30
60
95
80
60
Voidm
g(%)
90
80
70
60
50
40
30
20
10
5
30
Fig. 20. Cross-section local average void fraction for stratied-wavy ow (VSG = 18 m/s, VSL = 0.2 m/s and, lL = 10 cP).
82
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
95
95
Voidm
g(%)
90
80
70
60
50
40
30
20
10
5
90
80
70
60
50
40
30
20
10
5
95
95
80
Voidm
g(%)
95
95
80
95 70
30
30
30
95
70
30
Fig. 21. Cross-section local average void fraction for annular ow (VSL = 0.03 m/s and, lL = 10 cP).
15
20
30
25
30
20
15
15
20
25
10
5
30
15
20
StdDev
g(%)
Voidst
10
15
20
25
30
30
25
20
15
10
5
1
30
30
20
10
10
10
30
25
20
15
10
5
1
15
10
10
StdDev
g(%)
Voidst
25
10
30
10
20
10
30
Fig. 22. Standard deviation of local average void fraction for stratiedslug transition (VSG = 9 m/s, VSL = 0.2 m/s and, lL = 1 cP).
5
10
20
20
10
20
15
25
1
10
20
30
15
20
10
20
30
25
20
15
10
5
1
30
30
30
20
10 5
15
25
30
30
25
20
15
10
5
1
StdDev
g(%)
Voidst
15
StdDev
g(%)
Voidst
30
20
20
Fig. 23. Standard deviation of local average void fraction for stratied-wavy ow (VSG = 18 m/s, VSL = 0.2 m/s and, lL = 1 cP).
and liquid viscosity. It can be concluded from the void fraction contour plots and time series that in the downstream section the presence of continuous liquid lm along the periphery of the tube is
more frequent than the upstream section of the elbow.
4.5. Reconstructed 3D images of stratiedslug transition and
stratied wavy ows
The iso-surface plots of the air phase in the two-phase ow
structures are obtained by xing the threshold of void fraction to
the individual air and liquid phases. In the current study, the
threshold values for the gas and liquid phases were xed to 0.9
and 0.1. Fig. 31 shows the structure of the waves and the slugs
formed at liquid viscosities of 1 cP and 10 cP. It can be observed
that with an increase of liquid viscosity from 1 cP to 10 cP, there
is a signicant change in the wave characteristics in the upstream
section. Therefore, a pseudo slug structure can be clearly observed
at 10 cP liquid viscosities. Fig. 31d shows that the downstream section of the pseudo-slug structure is even more aerated. This pseudo
slug structure may be described as liquid slug structures that lack
sufcient quantities of liquid to form a traditional slug. In stratied
ows, when a wave keeps growing, its amplitude will eventually
83
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
1
10
20
20
20
5
15
30
25
20
15
10
5
1
30
25
20
15
10
5
1
StdDev
g(%)
Voidst
StdDev
g(%)
Voidst
10
10
20
15
25
5
15
30
Fig. 24. Standard deviation of local average void fraction for annular ow (VSG = 0.03 m/s and, lL = 1 cP).
5
10
10
10
15
15
25
30
20
15
20
30
25
20
25
10
StdDev
g(%)
Voidst
10
30
25
20
15
10
5
1
StdDev
g(%)
Voidst
20
15
15
10
25
15
20
20
30
30
25
30
25
20
15
10
5
1
30
20
15
30
25
25 30
15
5
Fig. 25. Standard deviation of local average void fraction for stratiedslug transition (VSG = 9 m/s, VSL = 0.2 m/s and, lL = 10 cP).
StdDev
g(%)
Voidst
10
20
30
10
20
15
25
30
15
25
10
15
20
30
15
15
5
10
20
30
25
20
15
10
5
1
30
30
15
StdDev
g(%)
Voidst
30
25
20
15
10
5
1
20
30
10
20
Fig. 26. Standard deviation of local average void fraction for stratied-wavy ow (VSG = 18 m/s, VSL = 0.2 m/s and, lL = 10 cP).
be so large that the top of the wave hits the top of the pipe. When
the system is operating at relatively low pressures and at relatively
low gas and liquid ow rates, very long slugs with sizes reaching
several pipe diameters may form. These experiments clearly show
that the high viscosity liquids tend to promote slug ow.
Fig. 32 shows the structure of stratied-wavy ow at liquid viscosities of 1 cP and 10 cP. It can be observed that with an increase
of liquid viscosity from 1 cP to 10 cP, there is no signicant change
in the stratied gasliquid structure in the upstream section.
84
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
StdDev
g(%)
Voidst
StdDev
g(%)
Voidst
30
25
20
15
10
5
1
1
1
5
1
15 0
10
30
25
20
15
10
5
1
5
10
10
20
1
5
10
5
15
10
15 5
2
Fig. 27. Standard deviation of local average void fraction for annular ow (VSL = 0.03 m/s and, lL = 10 cP).
Void fraction
0.95
0.9
0.85
0.8
0.75
Upstream
Downstream
0.7
0.65
0
10
20
30
40
50
60
Time, (s)
Fig. 30. Time series of averaged void fraction around the top of the pipe for
stratied-wavy ow (VSG = 18 m/s, VSL = 0.2 m/s; lL = 1 cP).
Fig. 28. Selected area for averaged void fraction time series around the top of the
pipe (in blue). (For interpretation of the references to color in this gure legend, the
reader is referred to the web version of this article.)
Void fraction
0.8
0.6
0.4
0.2
0
Upstream
Downstream
10
20
30
40
50
60
Time, (s)
Fig. 29. Time series of averaged void fraction around the top of the pipe for
stratiedslug transition (VSG = 9 m/s, VSL = 0.2 m/s and, lL = 1 cP).
representative experimental conditions in the test matrix. The gures are organized horizontally when the supercial gas velocity is
constant and vertically when the liquid supercial velocity is constant. As can be seen in Fig. 33, there are typical shapes or signatures as the supercial velocities change. A narrow single peak
for void fractions between 0.9 and 0.95 is typical of annular ow.
For the second type, a single peak at high void fraction with a small
tail down to lower void fractions is recognized as the signature of
stratied-wavy ow. As the gas velocity is increased, the lower
void fraction peak moves to higher values and the peaks appear
at void fractions of 0.97, which represent annular ows. As the
liquid velocity is increased, all dominant void fraction peak frequencies diminish. For lower gas velocities, a double peak is usually expected for slug ow. The observed peaks in the data are
not bimodal, these slugs are highly aerated not yet fully developed
and do not present the characteristic frequency for slug ow. This
is not surprising as the amount of liquid present in the pipe is too
small for slug ow to exist and as a result, a characteristic single
peak signature of stratied ow is still observed.
One interesting observation is that, for a constant VSG, the
observed differences between the maximum PDF values upstream
and downstream vary for both liquid velocities of 0.03 and 0.2 m/s.
For the lower liquid velocity, as the gas velocity increases from
9 m/s to 27 m/s, the differences between the maximum values
decrease; on the other hand, for higher liquid velocity of 0.2 m/s,
a slight increase of this difference is observed. A following broad
tail present in the PDF of the upstream section does not exist in
the PDF results for data collected after the elbow for supercial
gas and liquid velocities of 27 m/s and 0.2 m/s, respectively. The
secondary peak disappears and a single peak is observed at a maximum void fraction value.
4.7. Structure velocity using cross-correlation techniques
The velocities of the periodic structures, such as highly unstable
liquid lms and disturbance waves can be obtained by computing
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
85
(a) Upstream; L = 1 cP
(b) Dowstream; L = 1 cP
(c) Upstream; L = 10 cP
(d) Dowstream; L = 10 cP
Fig. 31. Inuence of liquid viscosity on the structure of liquid waves and slugs (VSG = 9 m/s, VSL = 0.2 m/s; lL = 1 cP).
(a) Upstream; L = 1 cP
(b) Dowstream; L = 1 cP
(c) Upstream; L = 10 cP
(d) Dowstream; L = 10 cP
Fig. 32. Inuence of liquid viscosity on the structure of stratied wavy-ow (VSG = 18 m/s, VSL = 0.2 m/s).
86
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
100
100
Upstream
Downstream
60
40
20
40
0
0.6
0.7
0.8
0.9
0.6
0.7
0.8
0.9
Void fraction
Void fraction
100
100
Upstream
Downstream
Upstream
Downstream
80
PDF, %
80
PDF, %
60
20
60
40
20
60
40
20
0
0.6
0.7
0.8
0.9
0.6
0.7
0.8
0.9
Void fraction
Void fraction
100
100
Upstream
Downstream
Upstream
80
PDF, %
80
PDF, %
Upstream
Downstream
80
PDF, %
PDF, %
80
60
40
20
Downstream
60
40
20
0
0.6
0.7
0.8
0.9
0.6
0.7
0.8
0.9
Void fraction
Void fraction
compared with the structure velocity for 1 cP liquid for the mixture
velocities considered.
4.0
Upstream; Vsl=0.03 m/s, 1 cP
Downstream; Vsl=0.03 m/s, 1 cP
Upstream; Vsl=0.03 m/s, 10 cP
Downstream; Vsl=0.03 m/s, 10 cP
3.0
5. Conclusions
2.0
1.0
0.0
5
10
15
20
25
30
35
40
A dual wire-mesh sensor was used to detect the local instantaneous cross-sectional distribution of the phases in gasliquid stratied and annular ows in the upstream and downstream locations
of a standard elbow. Data were obtained for a wide range of ow
rates in the horizontal orientation. Detailed information of the void
fraction distribution and interfacial structures are obtained. The
signals of the sensor, that are proportional to the liquid conductivity, are processed to obtain void fraction time series, cross-section
average time series, local void fraction distributions and structure
velocities. The effects of the supercial liquid and gas velocities
and liquid viscosity on the average and local void fraction have
been experimentally investigated. The analysis included times series plots of void fraction, contour plots of void fraction distribution
and local standard deviation, 3D image reconstruction, PDF of void
fraction and average void fraction for various ow regimes considered. These results shed additional light on the hydrodynamics of
horizontal gasliquid ows, in particular, regarding the
R.E. Vieira et al. / Experimental Thermal and Fluid Science 59 (2014) 7287
distribution of void fraction in the upstream and downstream section of a standard elbow.
The experimental study highlights the details of gas and liquid
distributions for stratied and annular gasliquid conditions inside
the pipe:
In general, it is found, within the considered operating conditions, that the effect of the 90 horizontal bend on the time
average void fraction was to slightly increase the cross-sectional time averaged void fraction value.
The sequence of instantaneous cross-sectional void fraction
showed that wave instability is the most notable feature of
the stratied to slug transition in both upstream and downstream locations of the elbow.
The contour plots of cross-sectional averaged void fraction
show that, irrespective of liquid viscosity, the cross-sectional
distribution of the void fraction changes after the elbow. The
analysis of the void fraction time series around the top of the
pipe highlights that for stratied ows the liquid tends to climb
the inner wall of the pipe for the measurements after the elbow.
These results help to gain a better understanding of how the
liquid phase wets the inner wall at the top of the pipe and
assure continuous wetting of corrosion inhibitors.
Three-dimensional structures have been identied for liquid
viscosities of 1 cP and 10 cP in stratiedslug transition and
stratied-wavy ows. Results show that high viscosity liquids
tend to promote slug ow. The pseudo slugs structures found
in the downstream section of the elbow are more aerated than
the structures identied in the upstream section.
Probability density function was used to discriminate non-stable
stratied ows from stratied wavy ow, as evident in Fig. 33.
The PDF for non-stable stratiedslug transition shows a at
region without a peak which is different from previous works
found in literature. Examination of the probability density function plots of the cross-sectional averaged time series of void fraction leads to the conclusion that these different trends are linked
to hydrodynamic changes in ow patterns and transition zones.
On the other hand, within the downstream pipe ow of horizontal 90 elbows, the same unstable stratied, stratied-wavy and
annular ow patterns are observed.
The contours of the standard deviation of the void fraction signal reect the stochastic behavior of the interface and the
amplitude of the interface oscillations.
The velocities of the periodic structures were obtained by computing the cross-correlation of average cross-sectional void
fraction time series from both meshes. The velocities were plotted against the mixture velocity. Positive slopes were obtained
for liquid viscosities of 1 cP and 10 cP in the downstream section. Almost a constant trend was obtained when the liquid viscosity was increased to 10 cP for the same ow conditions in the
upstream section.
References
[1] [1] M.I. Cohen, An investigation of pressure drop in a two-phase twocomponent ow in bends, M.Sc. Thesis, MIT, 1957.
[2] J.R. Castillo, Study of two-phase ow in pipe bends, M.Sc. Thesis MIT, 1957.
[3] S.A. Shirazi, J. Shadley, B. McLaury, E. Rybicki, A procedure to predict solid
particle erosion in elbows and tees, J. Pressure Vessel Technol. 117 (1995) 4552.
[4] B.S. McLaury, J. Wang, S.A. Shirazi, J. Shadley, E. Rybicki, Solid particle erosion
in long radius elbows and straight pipes, in: SPE Annual Technical Conference
and Exhibition, San Antonio, Texas. SPE38842 (1997) pp. 977986.
87
[5] B.S. McLaury, S. Shirazi, E. Rybicki, Sand erosion in multiphase ow for slug and
annular ow regimes, CORROSION/2010 San Antonio, Texas, paper No.10377
(2010).
[6] N. Kesana, S. Grubb, B.S. McLaury, S.A. Shirazi, Ultrasonic measurement of
multiphase ow erosion patterns in a standard elbow, ASME J. Energy Resour.
Technol. 1353 (2013) 032905.
[7] D.C. Silverman, Rotating cylinder electrode geometry relationships for
prediction of velocity-sensitive corrosion, Corrosion 44 (1) (1988) 4249.
[8] J. Postlethwaite, E. Tinker, M. Hawrylak, Erosioncorrosion in slurry pipelines,
Corrosion 30 (1974) 285290.
[9] J. Villarreal, D. Laverde, C. Fuentes, Carbonsteel corrosion in multiphase slug
ow and CO2, Corros. Sci. 48 (2006) 23632379.
[10] K. Usama, Long liquid slugs in stratied gas/liquid ow in horizontal and
slightly inclined pipes. Ph.D. Thesis, Delft University, Netherlands, 2009.
[11] G. Gardner, P. Neller, Phase distributions ow of an airwater mixture round
bends and past obstructions, Proc. Inst. Mech. Eng. 184 (1969) 93101.
[12] K. Usui, S. Aoki, A. Inoue, Flow behavior and pressure drop of two-phase ow
through C-shaped bend in a vertical plane-I: upward ow, J. Nucl. Sci. Technol.
17 (1980) 875887.
[13] K. Usui, S. Aoki, A. Inoue, Flow behavior and phase distributions in two-phase
ow around inverted U-bend, J. Nucl. Sci. Technol. 20 (1983) 915928.
[14] S. Das, M. Biswas, A. Mitra, Non-Newtonian liquid ow in bend, Chem. Eng. J.
45 (1991) 165171.
[15] T. Bandyopadhyay, T. Banerjee, S. Das, Gas-non-Newtonian liquid ow through
elbows, Chem. Eng. Commun. 182 (2000) 2133.
[16] A. Ribeiro, T. Bott, D. Jepson, The inuence of a bend on drop sizes in horizontal
annular two-phase ow, Int. J. Multiphase Flow 27 (2001) 721728.
[17] P. James, B. Azzopardi, D. Graham, C. Sudlow, The effect of a bend on droplet
size distribution in two-phase ow, in: 7th International Conference on
Multiphase Flow in Industrial Plants, Bologna, 1315 September 2000.
[18] A. Azzi, L. Friedel, R. Kibboua, B. Shannak, Reproductive accuracy of two-phase
ow pressure loss correlations for vertical 90 degree bends, Forsch.
Ingenieurwesen 67 (2002) 109116.
[19] P.L. Spedding, E. Benard, Gasliquid two phase ow through a vertical 90
elbow bend, Exp. Therm. Fluid Sci. 31 (2007) 761769.
[20] M. Abdulkadir, D. Zhao, S. Sharaf, L. Abdulkareem, I. Lowndes, B. Azzopardi,
Interrogating the effect of 90 degree bends on airsilicone oil ows using
advanced instrumentation, Chem. Eng. Sci. 66 (2011) 24532467.
[21] N. Kesana, Erosion in multiphase pseudo-slug ow with emphasis on sand
sampling and pseudo-slug characteristics, Ph.D. thesis, The University of Tulsa,
Oklahoma, USA, 2013.
[22] H.M. Prasser, A. Bttger, J. Zschau, A new electrode-mesh tomograph for gas/
liquid ows, Flow Meas. Instrum. 9 (1998) 111119.
[23] J.C. Rodriguez, Effects of liquid viscosity and sand size on erosion in slug
multiphase ow, MSc. thesis, The University of Tulsa, Oklahoma, USA, 2008.
[24] R. Vieira, N. Kesana, B.S. McLaury, S. Shirazi, Sand erosion in multiphase ow
for low-liquid loading and annular conditions, in: Proceedings of the
International Mechanical Engineering Congress-American Society of
Mechanical Engineers, 2012, Houston, Texas, 812 November.
[25] R. Vieira, N. Kesana, B. McLaury, S. Shirazi, E. Schleicher, U. Hampel,
Experimental investigation of horizontal gasliquid stratied and annular
ow using wire mesh sensor, in: Proceedings of the ASME Fluids Engineering
Division Summer Meeting, 2013, paper no. FEDSM2013-16117, Incline Village,
Nevada, July 711.
[26] D. Barnea, A unied model for predicting ow pattern transitions for the whole
range of pipe inclinations, Int. J. Multiphase Flow 13 (1987) 112.
[27] F. Stern, M. Muste, M. Beninati, W. Eichinger, Summary of Experimental
Uncertainty Assessment Methodology. Iowa Institute of Hydraulic Research
Technical Report No. 406, College of Engineering, The University of Iowa, 1999.
[28] J. Bendat, A. Piersol, Random Data: Analysis and Measurement Procedures,
fourth ed., Wiley, New Jersey, 2010.
[29] S. Orfanidis, Optimum Signal Processing: An Introduction, second ed.,
McGraw-Hill, New York, 2007.
[30] P.Y. Lin, T.J. Hanratty, Detection of slug ow from pressure measurements, Int.
J. Multiphase Flow 13 (1987) 1321.
[31] J.E. Laurinat, T.J. Hanratty, W.P. Jepson, Film thickness distribution for gas
liquid annular ow in a horizontal pipe, PhysicoChem. Hydrodynam. 6 (1985)
179195.
[32] T.W.F. Russell, D.E. Lamb, Flow mechanism of two-phase annular ow, Can. J.
Chem. Eng. 43 (1965) 237245.
[33] D. Butterworth, Airwater annular ow in a horizontal tube, Prog. Heat Mass
Transfer 6 (1972) 235251.
[34] T. Fukano, A. Ousaka, Prediction of circumferential distribution of lm
thickness in horizontal and near-horizontal gasliquid annular ow, Int. J.
Multiphase Flow 15 (1989) 403420.
[35] P. Attwood, K. Van Gelder, C. Charnley, CO2 Corrosion in wet gas systems.
CORROSION/1996 San Antonio, Texas, 32, 1996.
[36] R. Nyborg, A. Dugstad, Top of line corrosion and water condensation in wet gas
pipelines. CORROSION/2007 San Antonio, Texas, paper No. 07555, 2007.