Group Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Title: 7.1.2 – Flow Regime Transition in Large Diameter Pipes
ALL VOLUMES CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102023/00172
 Material Information
Title: 7.1.2 – Flow Regime Transition in Large Diameter Pipes Bubbly Flows
Series Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Physical Description: Conference Papers
Creator: Peng, D.J.
Ahmad, A.
Hale, C>P>
Matar, O.K.
Hewitt, G.F.
Publisher: International Conference on Multiphase Flow (ICMF)
Publication Date: June 4, 2010
 Subjects
Subject: flow regime transition
bubble flow
slug flow
churn flow
void wave
large diameter pipe
 Notes
Abstract: There is increasing interest in multiphase flows in large diameter vertical pipes (typically with diameters greater that 100 mm) in the context of hydrocarbon production systems. There are strong indications that flows in such pipes differ greatly from those in smaller diameter pipes on which most of the prediction methodologies are based. In small diameter pipes, an important mechanism for the bubble flow to slug flow transition is the formation of void waves. As the gas velocity is further increased, the slug flow itself breaks down into churn flow by a process of flooding in the Taylor bubbles. In large diameter pipes, it appears that conventional slug flow does not occur; this is probably due to the fact that there is a size limit on spherical cap bubbles. This change in nature of the slug flow transition may be illuminated by comparing void wave propagation in small and large diameter pipes. Only when spherical cap bubbles can occupy the full pipe does the slug flow regime occur. Thus, in large diameter pipes, the bubble/slug and slug/churn transitions appear to be by-passed in favour of a direct transition from bubble to churn flow with increasing gas mass flux. Some investigators report a gradual transition from normal bubble flow to a regime they call churn-turbulent flow. This is greatly different to the churn flow being investigated in this study where there is a continuous path for the gas phase. This paper describes work aimed at developing a phenomenological understanding of the bubble/churn and churn/annular transition regions in large diameter pipes. Investigation of the liquid transport mechanisms has led to the definition of two new flow regime transition criteria, namely liquid upflow potential and minimum entrained fraction. To estimate the bubble-to-churn flow transition, the liquid upflow potential of a churn flow at the particular local set of gas and liquid flow rates is estimated. In churn flow, liquid upflow is achieved by the net upward flow in the film (bearing in mind that both upflow and downflow are occurring in the film, though the net value must be positive) and by droplet transport in the gas core. Once a gas velocity is reached where the liquid upflow potential is greater than the liquid input rate, then it is postulated that the transition to churn flow occurs. As the gas velocity is further increased, the flow rate of entrained drops in the gas core decreases to a minimum and then rises again. This minimum is observed to occur at a dimensionless gas velocity approximately equal to one and this serves as a possible criterion for the churn-to-annular flow transition. Axial view experiments and the existing adiabatic equilibrium data have been adapted to verify these concepts.
General Note: The International Conference on Multiphase Flow (ICMF) first was held in Tsukuba, Japan in 1991 and the second ICMF took place in Kyoto, Japan in 1995. During this conference, it was decided to establish an International Governing Board which oversees the major aspects of the conference and makes decisions about future conference locations. Due to the great importance of the field, it was furthermore decided to hold the conference every three years successively in Asia including Australia, Europe including Africa, Russia and the Near East and America. Hence, ICMF 1998 was held in Lyon, France, ICMF 2001 in New Orleans, USA, ICMF 2004 in Yokohama, Japan, and ICMF 2007 in Leipzig, Germany. ICMF-2010 is devoted to all aspects of Multiphase Flow. Researchers from all over the world gathered in order to introduce their recent advances in the field and thereby promote the exchange of new ideas, results and techniques. The conference is a key event in Multiphase Flow and supports the advancement of science in this very important field. The major research topics relevant for the conference are as follows: Bio-Fluid Dynamics; Boiling; Bubbly Flows; Cavitation; Colloidal and Suspension Dynamics; Collision, Agglomeration and Breakup; Computational Techniques for Multiphase Flows; Droplet Flows; Environmental and Geophysical Flows; Experimental Methods for Multiphase Flows; Fluidized and Circulating Fluidized Beds; Fluid Structure Interactions; Granular Media; Industrial Applications; Instabilities; Interfacial Flows; Micro and Nano-Scale Multiphase Flows; Microgravity in Two-Phase Flow; Multiphase Flows with Heat and Mass Transfer; Non-Newtonian Multiphase Flows; Particle-Laden Flows; Particle, Bubble and Drop Dynamics; Reactive Multiphase Flows
 Record Information
Bibliographic ID: UF00102023
Volume ID: VID00172
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: 712-Peng-ICMF2010.pdf

Full Text


Paper No 7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010



Flow Regime Transitions In Large Diameter Pipes


D.J. Peng*, M. Ahmadt, C.P Hale*, O.K. Matar*, and GF. Hewitt*


Department of Chemical Engineering, Imperial College London
T Department of Mechanical Engineering, Imperial College London
Imperial College London, South Kensington Campus, London, SW7 2AZ, United Kingdom
d.peng07@imperial.ac.uk


Keywords: flow regime transition, bubble flow, slug flow, chum flow, void wave, large diameter pipe





Abstract

There is increasing interest in multiphase flows in large diameter vertical pipes (typically with diameters greater that 100 mm) in the context
of hydrocarbon production systems. There are strong indications that flows in such pipes differ greatly from those in smaller diameter pipes
on which most of the prediction methodologies are based. In small diameter pipes, an important mechanism for the bubble flow to slug flow
transition is the formation of void waves. As the gas velocity is further increased, the slug flow itself breaks down into churn flow by a
process of flooding in the Taylor bubbles. In large diameter pipes, it appears that conventional slug flow does not occur; this is probably due
to the fact that there is a size limit on spherical cap bubbles. This change in nature of the slug flow transition may be illuminated by
comparing void wave propagation in small and large diameter pipes. Only when spherical cap bubbles can occupy the full pipe does the slug
flow regime occur. Thus, in large diameter pipes, the bubble/slug and slug/churn transitions appear to be by-passed in favour of a direct
transition from bubble to churn flow with increasing gas mass flux. Some investigators report a gradual transition from normal bubble flow
to a regime they call chur-turbulent flow. This is greatly different to the churn flow being investigated in this study where there is a
continuous path for the gas phase. This paper describes work aimed at developing a phenomenological understanding of the bubble/churn
and churn/annular transition regions in large diameter pipes. Investigation of the liquid transport mechanisms has led to the definition of two
new flow regime transition criteria, namely liquid upflow potential and minimum entrained fraction. To estimate the bubble-to-churn flow
transition, the liquid upflow potential of a chum flow at the particular local set of gas and liquid flow rates is estimated. In churn flow, liquid
upflow is achieved by the net upward flow in the film (bearing in mind that both upflow and downflow are occurring in the film, though the
net value must be positive) and by droplet transport in the gas core. Once a gas velocity is reached where the liquid upflow potential is
greater than the liquid input rate, then it is postulated that the transition to churn flow occurs. As the gas velocity is further increased, the
flow rate of entrained drops in the gas core decreases to a minimum and then rises again. This minimum is observed to occur at a
dimensionless gas velocity approximately equal to one and this serves as a possible criterion for the churn-to-annular flow transition. Axial
view experiments and the existing adiabatic equilibrium data have been adapted to verify these concepts.


Introduction (UG) is at approximately 1.0 (Hewitt and Wallis 1963).


Flow regimes in two-phase or multiphase flows are very
important in understanding and designing equipment and
establishing the safe operating limits of such equipment.
Thus, investigations of flow regimes and of the transitions
between such regimes have attracted a great deal of
attention from both academic and industrial research
groups. However, a major area of uncertainty

In two-phase flow in vertical round pipes, the flow regimes
are often classified as bubble, slug, chum and annular flow
(Hewitt 1982). Here, we consider first regime transitions in
smaller diameter pipes (typically less than 100 mm in
diameter). The bubble-to-slug flow regime transition is
commonly linked to the growth of void waves and usually
takes place at critical void fraction of around 0.25
(Biesheuvel and Gorissen 1990, Song et al 1995). A likely
interpretation of the slug-to-chum flow transition, is that it
is a consequence of flooding in the Taylor bubble in slug
flow (Jayanti and Hewitt 1992). The chum-to-annular flow
transition occurs when the dimensionless gas velocity


U, is given by :
U.1 1/2
U = Up 2[gdo(p, pP)12

where UG is the uperficial gas velocity, pG and p,
the gas and liquid densities, g the acceleration due to
gravity and do the tube diameter.

However, it has to be noted that these flow regimes and
their transitions appear when pipe diameter is smaller than
100 mm, hereafter called small diameter pipes.
Conventional slug flow does not exist in pipes whose
diameter is above 100 mm, so-called large diameter pipes.
Thus, the flow regime transition criteria have to be altered
owing to the fact that different physical phenomena
dominate the transition mechanisms. Some researchers
have suggested that the flow regime transitions in large
diameter pipes are due to bubble size changes (Cheng et al.,
1998; Prasser et al., 2005), these changes being a






Paper No


consequence of bubble coalescence and breakup. In this
case, the flow regime changes from bubble flow to
churn-turbulent flow where large vortices and gas
entrainment exist between bubble clusters. There was no
evidence of liquid films with interfacial waves flowing
along the pipe wall. This type of flow regime is not being
investigated in the present study. However, it is likely to
be a precursor regime to the ultimate creation of chum flow
in large diameter pipes.

Here, chum flow is defined as a regime where a thick,
highly oscillatory liquid film exists on the pipe wall. The
film is swept by large flooding-type waves in between
which there exist regions of downwards film flow. The
nature of this regime was clearly demonstrated by Hewitt
et al (1985) using photochromic dye tracing. These
experiments strongly suggested that, in chum flow, the gas
core is continuous though the observed interactions
between the large waves and the gas core seemed likely to
be leading to entrainment of droplets from these waves into
the core.

Barbosa et al (2002) report measurements of liquid
entrained fraction in the chum/annular transition region.
Traditionally, entrained fraction in annular flow has been
measured indirectly by sucking off the liquid film and
estimating the entrained flow rate as the difference between
the film flow rate and the total liquid flow rate. This
technique cannot be applied in churn flow because of the
fluctuation of film flow direction (as observed by Hewitt et
al ,1985). Thus, Barbosa et al (2002) used a sampling
probe to measure the local droplet mass flux and integrated
these measurements to obtain entrained fraction.

In a parallel paper (also presented at ICMF7) Ahmad et al
(2010) present a correlation for entrainment in chum flow
(based on the Barbosa et al data) and also present the
results of a study using axial view photography which
demonstrate unequivocally that the gas core is continuous
in chum flow (an implicit assumption in the Barbosa et al
experiments).

In order to develop new transition criteria in large diameter
pipes, the mechanisms of upward liquid transport in chum
flow needs to be understood, at least approximately.
Essentially, there are two mechanisms for upward liquid
transport, namely net upwards transport in the liquid film
and net upwards transport of liquid drops. One may
define (for chum flow) a Liquid Upflow Potential (LUP)
which is the liquid rate which can be carried upwards at a
given gas flow rate. One can argue that if the liquid flow
rate is less than the LUP, then upwards transport of the
liquid is possible in the chum flow regime; if it is not, then
the transport mechanism has to be one associated with
bubble flow or slug flow.

In what follows, we first discuss the transitions in bubble
flow, and specifically the effect of tube diameter on void
wave growth. Results in the literature are consistent with
the idea that void wave growth in bubble flow occurs only
with small tube. Next, we summarize and present further
analysis of the Barbosa et al chum flow entrainment data in
terms of the LUP concept annular flow, A closely related


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

matter is that of "liquid loading" in natural gas wells and
this is discussed in the next section, It is shown that
upwards liquid transport is possible at much lower gas flow
rates than those commonly assumed. The nature of the
liquid loading phenomenon may differ from that postulated
previously. Finally, some conclusions are drawn an some
possible directions for further work indicated.

Nomenclature


pipe diameter (m)
deposition rate (kgm-2-1)
entrained fraction (-)
mass flow rate (kgm-2S-1)
superficial velocity (ms-1)
(wave amplitude)d-1
dimensionless velocity (-)
quality (-)
axial direction (m)


Greek letters
8 liquid film thickness (m)
p density (kgm-3)
C surface tension (Nm 1)
T shear stress (nm-2)

Subsripts
G gas
L liquid
LE liquid entrainment
LF liquid film
LUP liquid upflow potential
M maximum

Void Wave Propagation and the bubble-slug
transition

The transition from bubble flow to slug flow is commonly
linked to the coalescence of bubbles into a size and shape
characteristic of Taylor bubbles which occupy nearly the
whole cross section of the pipe. It seems unlikely that such
coalescence will occur in bubble-to-bubble collisions in
free-flowing bubble flow. The contact time between bubbles
will not necessarily be long enough in such free flows to
facilitate coalescence. However, if void waves are formed,
then the bubbles in such waves are maintained in closer
proximity for longer and coalescence leading to to bubble
growth and ultimately the formation of Taylor bubbles and
hence slug flow may occur (Biesheuvel and Gorissen 1990). ,
Song et al 1995).have proposed a statistical factor ( the
Spatial Attenuation Factor SAF) which is obtained from
analysis of void fluctuations. SAF is defined as the
minimum of the imaginary part of the complex wave
number and represents the most unstable factor dominating
wave growth. If SAF is positive, then void waves will
attenuate; if SAF is negative, the void waves will grow. SAF
results obtained by Song et al for air-water upflow in a 0.025
m pipe are shown in Figure 1. The transition from bubble
flow to slug flow corresponds closely to the void fraction at
which SAF becomes negative. The transition is seen to be
affected by the bubble size; for larger bubbles, the transition
occurs at a lower void fraction.






7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010


(Cheng et al., 1998) carried out experiments using air-water
flow in a 150 mm pipe. These experiments suggested
strongly that the bubble/slug flow transition did not take
place in this large diameter channel.

Some important further tests on the effect of channel
diameter on flow pattern transition are reported by
Omebere-Iyari et al (2007). In these tests, use was made of a
0.189 m diameter and 50 m tall riser facility in the Tiller
Loop in Trondheim, Norway. The fluids used were nitrogen
and naptha with a gas density range of 23-103 kg/m3 and a
liquid density of 700 kg/m3; the experiments were carried
out at a pressure 20 bar. The range of superficial liquid
velocity (0.004-4.0 m/s) and superficial gas velocity
(0.1-15.0 m/s) were such as to cover the bubble and slug
flow regimes and the transition between them in smaller
diameter pipes. Measurements of temporal and spatial void
fraction distributions reported by Omebere-Iyari et al (2007)
clearly show that slug flow as normally envisaged does not
occur in these large diameter pipes. This is probably because
the spherical cap bubbles formed by coalescence of smaller
bubbles are themselves unstable and their growth beyond a
diameter of around 5 cm is inhibited by the shedding of
bubbles from around the "skirt" of the spherical cap bubble.

The data from the Omebere-Iyari et al (2007) experiments
were kindly made available to us by Professor B. J.
Azzopardi of Nottingham University and were analysed
using the procedure developed by by Song et al (1995). This
led to the calculation of SAF values for this data and the
results are shown in Figure 2.


- First order exponential decay fitting


009


..


02 04 06 08
Void Fraction


First order exponential decayfitting





.. ..... .. .i n 1


02 04 06
Void Fraction


08 10


i (m) JLO.t2 nr






Db |
p 10 2D a S
vO troam (


(b)JL.0-18 nV






a

to 2 3 40
voWm Naeon %)


4, Dr 1 .T
S 10 o20 30 40 0 0 10 o 0 30 0
Void ftWcUn II Vowi radtiln M)
Figure 1: Instability of void wave propagation related to
the bubble-to-slug flow transition in 0.025 m pipe (Song et
al., 1995).

As will be seen from Figure 2, the values of SAF for the
Omebere-Iyari et al (2007) data are small compared to the
Song et al data (Figure 1) for the smaller diameter pipe but
the values are positive over the full range of void fractions
covered. This implies that void waves will not grow under
the conditions investigated and this is consistent with the
non-existence of slug flow as observed from the void
distribution measurements.


Figure 2 : Results for SAF obtained by analysis of transient
void fraction results obtained by Omebere-Iyari et al (2007)
for nitrogen/naptha flow in a 0.189 m pipe plotted in (a)
small scale and (b) same scale as Song et al. (1995) used.

Thus in large diameter pipes, the non-existence of slug
flow means that the Taylor bubble flooding mechanism
(Jayanti and Hewitt, 1992) cannot be used to describe the
transition to chum flow and an alternative approach has to
be taken. Clearly, at at high enough gas velocity, annular
flow will exist even in large diameter pipes. It is not
unreasonable to assume that churn flow (with upwardly
transported large waves and falling films between the
waves) will also occur in large diameter pipes as a precursor
to annular flow. The crucial issue is to estimate the liquid
flow rate which can be carried upwards by a chum flow -
the "liquid upflow potential" (LUP).

Liquid transport in churn flow.

Though the indications from previous experiments such as
those of Hewitt et al (1985) gave strong indications of the
nature of the near-wall region in chum flow (i.e. the
existence of a continuous liquid film at the wall with
waves and flow reversals between the waves), these
experiments did not give a clear indication of the nature
of the gas core and of the entrained liquid within it. The
gas core can be conveniently visualised using axial view
photography. Figure 3 shows axial view photographs
obtained on the Imperial College LOTUS facility for
upwards air-water flow in a 32 mm diameter vertical tube.


Paper No






7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010


Further details are given in the parallel paper in this
conference (Ahmad et al, 2010). Figure 4 shows results
obtained (also using the LOTUS facility) by Barbosa et al


Z 30
0
I-
C
U-
0
La 20
z
L-
z
uJ

C3
-J


Figure 3: LOTUS axial view experiments undertaken at
superficial liquid velocity Jf = 0.165 m/s in (a) Chum Flow
(UG = 0.528), (b) Chum-to-Annular Flow transition (UG =
0.968) and (c) Annular Flow (UG = 1.043).


0---- -\--I
0,40 0.80 1.20 1,50 2,00
DIMENSIONLESS GAS VELOCITY
Figure 4: Trends of liquid entrained fraction in vertical
flows at (a) P = 2.0 bar (b) P = 3.6 bar (Barbosa et al.
2002).

(2002) for entrained liquid fraction in the region spannng
chum and annular flow. The pictures shown in Figure 3
demonstrate unequivocally that there is a clear gas core in
chum flow. Figures 3 and 4 show that the extent of gas
entrainment decreases throughout the chum flow region,
reaches a minimum at the transition to annular flow (i.e at a
dimensionless gas velocity U0 of around unity) and

increases again in the annular flow region ( U > 1).


Entrained fraction in churn flow is an important boundary
condition for the annular flow region and (in the parallel
paper at this conference) Ahmad et al (2010) report a
correlation for entrainment rate in chum flow which can be
applied to both fully developed chum flow and to the
developing flows which can be found in evaporating
systems. Figure 5 shows predictions of fully developed
entrained fraction compared to the data of Barbosa et al
(2002). As will be seen, there is good agreement between
the correlation and the data for fully developed chum flow.

The ability to predict one component of the liquid upflow
potential (i.e. the entrained drop component) still leaves the
problem of predicting the other component, namely the
amount of liquid which can be carried in the liquid film.
Figure 6 shows a plot of liquid film flow rate as a function
of inlet liquid flow rate for one of the Barbosa et al (2002)
data sets similarr results are obtained for the other data sets).
Another way of plotting this data is to plot the fraction of
the liquid flowing in the film as a function of
dimensionless gas velocity a plot of this type is given in
Figure 7. As will be seen from this latter plot, the film flow
fraction remains high for the whole of the Barbosa data set
except for the region adjacent to the condition of the
slug/chum transition. In that region the film fraction falls
consistent with the rise in entrained fraction as seen in
Figure 4. A possible reason for this rise is the existence of
residual liquid in the core arising from the slug/chum
transition.


Paper No






7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010


Liquid mass flow rate = 23 kg/m2/s


--- Correlation in this work
--- Barbosa et al. 2001 data


Churn-to-Annular transition


02 04 06 08 10 12 1
Dimensionless Air Flow Rate, U.


Liquid mass flow rate = 105 kg/m/s


-- Correlation in this work
-*- Barbosa et al. 2001 data


Churn-to-Annular transition

02 04 06 08 10 12 1
Dimensionless Air Flow Rate, U.


I Avg UG*=/ /2 / 99 bar
---Avg UG*=/ 5/ / 99 bar
Y--Avg UG*=0 6 96 bar
-- --Avg UG*=0 74_ 96 bar
I---Avg UG*=0 1_ 9/ bar
S---Avg UG*= /2 356 bar
I-Avg UG*= 5 3 56 bar
Avg UG*= 43 56 bar
-^*A g U *=1 1_ 56 bar
-Avg UG*=0 3/ 502 bar
-Avg UG*=0 79 5 02 bar
-Avg UG*=0 955 02 bar
TA g UG *=1 295 02 bar
02 04 06 08 10 12 14 16
U,


Figure 7: Fraction of the liquid flow in the film as a
function of the dimensionless gas velocity

The main lesson from the results shown in Figures 6. and 7
is that, once the churn flow region has been entered, then it
is possible to carry upwards any liquid which enters the
tube. Thus, in this region, the LUP is equal to the total
liquid flow. Here, the crucial phenomenon is theflooding
condition. If the situation is such that flooding is predicted
to have occurred, then chum flow will occur. The most
commonly used correlations for flooding are those of Wallis
(1961) and Pushkina and Sorokin (1969). The Wallis
correlation has the form:


(UG)1/2 +(U ,1/2


where C is a constant of the order of unity and UL then
is the dimensionless liquid velocity deined as:


U, = ULp[gdo(PL,


PG )1/2


Figure 5: Comparison of entrained fraction predicted
from the correlation with data of Barbosa et al (2002).
Agreement of the work proposed in this study against the
experiments.


Avg. Pressure 3.56 bar
360
320- --Avg UG=0 61
-4--Avg UG*=0 74
280- -A-Avg UG= 91
--Avg UG*=1 12 ."
240- ---Avg UG=1 51
200-
160-
120-
80
40-

0 40 80 120 160 200 240 280 320 360
Inlet Liquid Flow Rate (kg/m 2s)


Figure 6: Plot of film flowrate as a function of inlet liquid
flow rate for the Barbosa et al (2002) data set at 3.56 bar.


The Pushkina and Sorokin correlation suggests that flooding
will occur for gas velocities for which the Kutateladse
number K is greater than the following:


K = Up2[g (pL


P )] 1/4


The latter correlatio shows no effect of liquid flow rate and
diameter and is conventionally used for tubes of larger
diameter.

In order to compare the results for various transiions, a plot
was prepared in terms of the flow pattern map of Hewitt nad
Roberts (1969) as shown in Figure 8. The following are
shown on this map: a series of lines have been calculated
and placed on the flow pattern map:



(1) The original lines and flow pattern limits.

(2) The Barbosa et al (2002) data. This lies, as would be
expected in the churn/annular tradition region in the map.


Paper No






Paper No






100000



10000



1000


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010


Bubble flow with
developing structure


p,(U,)2, kg/m/s2


Figure 8: Comparisons with flow pattern map of Hewitt and Roberts (1969)


(3) Lines calculed from the Wallis correlation for
atmspeheric pressure air-water flow in pipes of various
diameters and each line assymtotes to the value
corresponding to UG = 1.
(4)A line for K=3.2. Where the gas velocity condition is
above this line, then the Puskina and Sorokin correlation
should be used as an indicator of the onset of ceum flow. .



Liquid Loading

In gas-condensate wells producing, a matter of
considerable concern is that of "liquid loading" where
liquid collection at the bottom of the well leads to a
pressure which is too high to allow continued production.
The most common method of determining the conditions at
which this would occur is that developed by Turner et al
(1969) who suggested that, to carry liquid out of the well, a
gas velocity greater than the terminal velocity Ur of the
largest drop would be required. In SI nits, the Turner et all
criterion can be written as follows:

005

U, = 5.463 05( p9 050
PG

Clearly, in the experiments of Barbosa et al (2002), liquid
is being carried out of the test section by both liquid
entrainment and by net transport in the liquid film. How do


the gas velocities to achieve this transport compare with
those predicted by the criterion of Turner et al?
Comparisons of the Barbsa data with the Turner et al
criterion are shown in Figure 9.


16-
14-
12-
S10-
S08-

CD 06-
04-
02-


0 V



vg
> Avg
a
0 Avg
SAvg


S 2 4 6 8 10 12
U,


14 16


Figure 9: Comparison of the Barbosa et
with the criterion of Turner et al (1969)


al (2002) data


As will be seen from Figure 9, liquid transport is possible
at velocities much lower than those predicted by the Turner
et al criterion. This criterion takes no account of the film
transport possible in chum flow. However, there must be
some explanation of the deterioration in performance at
velocities below the Turner et al value. Perhaps this
explanation may lie in the excursion in pressure gradient


F






Paper No


which occurs at the onset of chum flow. This is illustrated
by some results obtained by Owen (1986) as illustrated in
Figure 10.





Buble D.032m

0.7 -

0.68





DA4
Ir -

C O.K -I
2g '
W L8 Si
o -2 A fk


- I I 1 1 1 1 1 1 1 1 1.. .
0 D. 4 0 a i 1.0 1.2 IA IJ
Dimensionless gas velocity Uc
At the chum flow transition, an excursion in pressure
gradient occurs (mainly associated with interfacial friction)
and significantly higher gas velocities may be required to
give acceptable pressure losses and, hence, flows, in the
well.

Conclusions

The results presented here show that significant differences
may occur in the nature of flow pattern transitions in small
diameter and large diameter pipes. The analysis of data for
bubble flow in large diameter pipes shows that void wave
growth does not seem to be occurring, consistent with the
observed absence of Taylor-type bubbles. For chum flow,
the combined mechanisms of liquid film and entrained
drop transport seemed sufficient to carry the liquid upwards
over a wide range of flow rates. Thus, calculation of the
conditions for flooding seems to be an appropriate bass for
estimating the onset of chum flow. For large diameter pipes,
the chosen correlation to predict flooding would be that of
Pushkina and Sorokin(1969). Comparison of the results (i.e.
those of Barbosa et al, 2002) analysed here with the
criterion of Turner et al (1969) indicated that liquid upflow
would occur at velocities well below those ideated by this
criterion. This discrepancy may be related to the excursion
in pressure gradient occurring at the onset of chum flow,
but this needs to be investigated further.


Acknowledgements

This work has been undertaken within the Joint Project on


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

Transient Multiphase Flows and Flow Assurance. The
Author(s) wish to acknowledge the contributions made to
this project by the UK Engineering and Physical Sciences
Research Council (EPSRC) and the following: Advantica;
BP Exploration; CD-adapco; Chevron; ConocoPhillips;
ENI; ExxonMobil; FEESA; IFP; Institutt for
Energiteknikk; PDVSA (INTEVEP); Petrobras;
PETRONAS; Scandpower PT; Shell; SINTEF;
StatoilHydro and TOTAL. The Author(s) wish to express
their sincere gratitude for this support.

References

Ahmad, M., Peng, D.J., Hale, C.P, Walker, S.P and Hewitt,
GH., "Droplet Entrainment in Chum Flow" 7th
International Conference on Multiphase Flow, ICMF 2010,
Tampa, FL USA, May 30-June 4 (2010)

Batchelor, G.K., A new theory of the instability of a
uniform fluidized bed, J. Fluid Mech. Vol. 193, pp. 75-110
(1988)

Barbosa, J.R., Govan, A.H., Hewitt, G.F., Visualisation and
modeling studies of chum flow in a vertical pipe, Int. J.
Multiphase Flow, Vol. 27, pp. 2105-2127 (2001)

Barbosa, J.R., Hewitt, G.F., Konig, G., Richardson, S.M.,
Liquid entrainment, droplet concentration and pressure
gradient at the onset of annular flow in a vertical pipe, Int.
J. Multiphase Flow, Vol. 28, pp. 943-961 (2002)

Biesheuvel A. & Gorissen W.C.M., Void fraction
disturbances in a uniform bubbly fluid, Int. J. Multiphase
Flow, Vol. 16, pp. 211-231 (1990)


Cheng, H., Hills, J.H., Azzorparrdi, B.J., A study of the
bubble-to-slug transition in vertical gas-liquid flow in
columns of different diameter, Int. J. Multiphase Flow, Vol.
24, pp. 431-452 (1998)

Govan, A.H., Modelling of vertical annular and dispersed
two-phase flows, PhD Thesis, Imperial College London
(1990)

Hewitt, G.F. & Wallis, G.B., Flooding and associated
phenomena in falling film in a vertical tube, In Proc. of
Multiphase Flow Symp., pp. 62-74, Philadelphia, PA
(1963).

Hewitt, G. F., and Whalley, PB. Advanced optical
instrumentation methods. Int. J. Multiphase Flow, Vol. 6,
No. 12 p136-156 (1980)

Hewitt, G.F., Handbook of Multiphase Flow System,
Hemisphere Publishing Corporation (1982)

Hewitt, G.F., Martin, C.J., Wilkes, N.S., Experimental and
modelling studies of annular flow in the region between
flow reversal and the pressure drop minimum,
PhysicoChemical Hydrodynamics, Vol. 6, pp. 69-86 (1985)

Ishii, M. & Hibiki, T., Thermo-Fluid Dynamics of
Two-Phase Flow, Springer, Printed in the United States of






Paper No 7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

America (2006)

Jayanti, S. & Hewitt, GF., Prediction of the slug-to-chum
flow transition in vertical two-phase flow, Int. J.
Multiphase Flow, Vol. 18, pp 847-860 (1992)

Koizumi, Y, Enari, R., Ohtake, H., Characteristics for a
liquid film falling down along a vertical wall, Journal of
Heat Transfer, Vol. 131, pp. 082901-1-082901-9 (2009)

Ombebere-Iyri, N. K.. Azzopardi, B. J. and Ladam, Y.
Two-phase flow patterns in large diameter vertical pipes at
high pressure. AIChE Journal, Vol. 53, pp. 2493 -2504
(2007)

Park, J. W., Lahey, R. T. Jr., Drew, D. A., The measurement
of void waves in bubbly two-phase flow, Proc. 6th Int. MTG.
on Nuclear Reactor Thermal-Hydraulics, Grenoble, 5-8
Oct., pp. 655-665 (1993).

Prasser, H.M., Beyer, M., Bottger, A., Carl, H., Lucas, D.,
Schaffrath, A., Schutz, P, Weiss, F.P, Zschau, J., Influence
of the pipe diameter on the structure of the gas-liquid
interface in a vertical two-phase pipe flow, Nucl. Technol.,
Vol. 152, pp. 3-22 (2005)

Pushkina O.L. and Sorokin, Y. L. Breakdown of liquid
film motion in vertical tubes. Heat Transfer Soviet
Research, Vol.1, pp. 56-64, 1969

Shearer, C.J. and Davidson, J.F., The investigation of a
standing wave due to gas blowing upwards over a liquid
film; its relation to flooding in wetted-wall columns, J.
Fluid Mech., Vol. 22, pp. 321-325 (1965)

Song, C.H., No, H.C., Chung, M.K., Investigation of
bubble flow developments and its transition based on the
instability of void fraction waves, Int. J. Multiphase Flow,
Vol. 21, pp. 381-404 (1995)

Turner, R. G, Hubbard, M. G. and Dukler, A. E., Analysis
and prediction of minimum flow rate for continuous
removal of liquids from gas wells. J. Petroleum Tecnology,
November 1969, pp.1475-1482.

Wallis, G. B. Flooding velocities for air and water in
vertical tubes. UKAEAReport AEEW-R123 (1961)





Paper No 7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010




100000-


10000- Annular Wispy annular

UG*=1 in 150 mm pipe
_UG*=1 in 100 mm pipe
1000- UG*=1 in 67 mm pipe
UG*=1 in 32 mm pipe

10 -0--- Kutateladze no. =3.2
| 100 +__-
E + Churn
.. .. ... --7

_S" '+' Slug Bubble flow with
S10, + Barbosa data developing structure


L(UL)2, kg/m/s2




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.7 - mvs