Group Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Title: 14.3.1 - PTV Measurements to Study Void Fraction Influence on the Liquid Phase Turbulence in Subcooled Boiling Flow
ALL VOLUMES CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102023/00347
 Material Information
Title: 14.3.1 - PTV Measurements to Study Void Fraction Influence on the Liquid Phase Turbulence in Subcooled Boiling Flow Experimental Methods for Multiphase Flows
Series Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Physical Description: Conference Papers
Creator: Estrada-Pérez, C.E.
Dominguez-Ontiveros, E.E.
Hassan, Y.A.
Publisher: International Conference on Multiphase Flow (ICMF)
Publication Date: June 4, 2010
 Subjects
Subject: experimental two-phase flow
PTV
void fraction measurements
 Notes
Abstract: Experiments were carried out to investigate turbulent subcooled boiling flow of Novec-2000 refrigerant through a vertical square channel with one heated wall. Channel dimensions were selected to be similar to those encountered on a Boiling Water Reactor (BWR) channel flow, with an hydraulic diameter of Dh = 8:2mm. Flow visualization techniques such as Particle Tracking Velocimetry (PTV) and Shadowgraphy were used to measure time-average axial and normal velocities, axial and normal turbulence intensities, and Reynolds Stresses. Results are reported for hydraulic Reynolds numbers at channel inlet of 4638, 14513 and 24188 for up to thirteen wall heat fluxes (q00) ranging from 0.0 to 64.0 kW=m2. This work is an attempt to enrich the database already collected on turbulent subcooled boiling flow, with the hope that it will be useful in turbulence modeling efforts.
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: VID00347
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: 1431-EstradaPerez-ICMF2010.pdf

Full Text



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


PTV Measurements to Study Void Fraction Influence on the Liquid Phase
Turbulence in Subcooled Boiling Flow


C. E. Estrada-P~rez E. E. Dominguez- Ontivero s* and

Y. A. Hassant

Department of Mechanical Engineering, Texas A&M University, College Station, TX 77843, USA
i- Department of Mechanical and Nuclear Engineering, Texas A&M University, College Station, TX 77843, USA
cperez ~ne.tamu.edu, elvisdom~tamu.edu and y-hassan~tamu.edu
Keywords: Experimental Two-Phase Flow, PTV, Void Fraction Measurements




Abstract

Experiments were carried out to investigate turbulent subcooled boiling flow of Novec-2000 refrigerant through a
vertical square channel with one heated wall. Channel dimensions were selected to be similar to those encountered
on a Boiling Water Reactor (BWR) channel flow, with an hydraulic diameter of Dh = 8.2mm. Flow visualization
techniques such as Particle Tracking Velocimetry (PTV) and Shadowgraphy were used to measure time-average
axial and normal velocities, axial and normal turbulence intensities, and Reynolds Stresses. Results are reported for
hydraulic Reynolds numbers at channel inlet of 4638, 14513 and 24188 for up to thirteen wall heat fluxes (q") ranging
from 0.0 to 64.0 kW/m2. This work is an attempt to enrich the database already collected on turbulent subcooled
boiling flow, with the hope that it will be useful in turbulence modeling efforts.


INTRODUCTION

Proper characterization of subcooled boiling flow is im-
portant in many applications. It is of exceptional signif-
icance in the development of empirical models for the
design of nuclear reactors, steam generators, and refrig-
eration systems. Most of these models are based on ex-
perimental studies that share the characteristics of utiliz-
ing point measurement probes with high temporal res-
olution, such as Hot Film Anemometry (HFA), Laser
Doppler Velocimetry (LDV), and Fiber Optic Probes
(FOP). The open literature contains many examples of
such works, but for brevity, we will refer to only a
few. Roy et al. Roy et al. (1986) pioneered on the
research of turbulent liquid flow through annular chan-
nels, heated and unheated. Using Refrigerant-113 and
HFA, they obtained mean axial velocity profiles and tur-
bulence intensities for various Reynolds numbers and
heat fluxes. They conclude that accurate velocity field
measurements in turbulent liquid flow by constant tem-
perature anemometry are difficult since generally only
low sensor overheats can be used. Lance and Bataille
Lance and Bataille (1991) used LDV measurements to-
gether with HFA to study the turbulence of the liquid in a
bubbly, grid-generated turbulent flow field. They found
that the turbulent kinetic energy greatly increases with


the void fraction (a). They described two regimes: the
first one corresponds to low values of a, where hydrody-
namic interactions between bubbles are negligible, and
the second one to higher values, for which the bubbles
transfer a greater amount of kinetic energy to the liquid.
The Reynolds stress tensor shows that the quasi-isotropy
is not altered. Furthermore, their one-dimensional spec-
tra analysis showed that a large range of high frequencies
associated with the wakes of the bubbles and the classi-
cal -1/5 power law is progressively replaced by a -8/3
dependence. Wardana et al. Wardana et al. (1992) used
LDV and a resistance thermometer to study air velocity
and temperature statistics in a strongly heated turbulent
two-dimensional channel flow, with wall temperatures
up to 700 oC and a Fixed Reynolds number of 14000.
They did not find \ignilk~.llll changes by the wall heat-
ing in either temperature or velocity fields, however they
found a suppression of the turbulent intensities of ve-
locity fluctuations far from the heated wall. Velidandla
et al. Velidandla et al. (1996) used a two-component
LDV and a micro-thermocouple to measure velocities
and temperatures. They found buoyancy effects on the
time-mean velocity and turbulence fields, even at very
low values of Gr/Re Roy et al. Roy et al. (1997)
measured the velocity field in turbulent subcooled boil-







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


ing flow. They found that the turbulence was inhomo-
geneous, and anisotropic and the turbulent kinetic en-
ergy significantly higher than in single-phase at the same
velocity. They found a marked shift toward the inner
wall of the zero location of the axial Reynolds shear
stress and that the magnitude of the shear stresses in-
creased sharply close to the inner wall. Zarate et al.
Zarate et al. (1998) developed velocity and temperature
wall laws in a vertical concentric annular channel from
measurements in turbulent liquid flow, noting that when
buoyancy forces influence becomes larger i.e. a large
Gr/Re2 the velocity and temperature data do not fol-
low the respective wall laws. Roy et al. Roy et al.
(2002) summarizes their findings on turbulent subcooled
boiling flow, and their models and simulations based
on their experimental data. These experimental studies
represent a rich description of phenomenological events
occurring in nucleate subcooled boiling flows. How-
ever, there appears to be a scarcity of experimental stud-
ies that can capture instantaneous whole-field measure-
ments with a fast time response. Particle Tracking Ve-
locimetry may be used to overcome the limitations as-
sociated with point measurement techniques. PTV is a
whole-flow-field technique providing instantaneous ve-
locity vectors capable of high spatial and temporal res-
olution Riethmuler (2001). PTV is also an exceptional
tool for the analysis of boiling flow due to its ability to
differentiate between the gas and liquid phases and sub-
sequently deliver independent velocity fields associated
with each phase Estrada-Perez (2004). In this work, time
resolved PTV and shadowgraphy techniques are used to
obtain flow measurements in turbulent subcooled boiling
flow of refrigerant 3M HFE 7000 through a rectangular
channel. This work is an attempt to enrich the database
already collected on turbulent subcooled boiling flow,
with the hope that it will be useful in turbulence mod-
eling efforts.


Experimental Facility

To obtain PTV measurements, a test facility was de-
signed for low system pressure and subcooled boil-
ing flow of refrigerant 3M 7000. The refrigerant is
pumped through a vertical, rectangular channel made
of transparent polycarbonate, with 530 min length and
a cross-sectional area perpendicular to the flow of
8.7. J: .6 min". A thin heater with a length and width
of 175 m i and 7 m i, respectively, was attached to one
of the lateral walls of the channel, leaving an unheated
length from the inlet of 320 min. Figure 1 shows the
schematics and dimensions of the test section.
Measurements were obtained 470 min from the in-
let, ensuring a fully developed turbulent flow. At this
site, pictures were obtained using a high speed cam-


outlet


Theirnoope /


Ihin heater

:lear Channel Top


nei


Tan a ttn
Bo


Tronsporent
-oCha ate


Ilt


Figure 1: Test section schematics and dimensions.


era. The camera was synchronized with a high energy
laser that provided al in m thick sheet of light for illu-
mination. The camera frame rate was 2 .1101p .. -- -/s
with an exposure time of 2ps; each image acquired con-
sisted of 600r J800 pixels with a spatial resolution of
12.3~mpin/iel. Three different Reynolds numbers (Re)
were considered in this work: 4838, 14513 and 24188,
for each Re about 11 different heat fluxes (q") were used,
ranging from 0 to 56.3 kW/in". For all cases a constant
inlet temperature of 25.5 uC was maintained and the
heater wall average temperature and fluid outlet temper-
ature were also measured.


Experimental Results

On the analysis of single and two-phase flows, is nec-
essary to determine whether or not the flow is two di-
mensional and fully developed at the measurement area.
To characterize the flow, liquid velocity measurements
using PTV technique were performed at different posi-
tions along the channel. The camera, mirrors and lenses
arrangement on movable positioners allowed to change
with minimum effort the measurement area along the



































.000 .500 1.000
ylh

Figure 3: Axial mean velocity 17 and axial turbulence
intensity u' profiles at different positions
along the channel.


measurement area P56 (at 455 min from the inlet) was
selected as a suitable measurement area for the heated
wall experiments. This measurement area is located at
the mid-point between the fifth and sixth thermocouples,
therefore the average wall temperature can be readily es-
timated. Also, flow development from lower measure-
ment areas is known and can be used as inlet boundary
conditions for computer simulations.


Single and Two-Phase PTV Experiments

Figure 4 shows the experimental images obtained from
the PTV experiment at a Re = 14513. Figure 4(a) repre-
sents the unheated single-phase flow images, where the
flow seedings are easily identified from the black back-
ground. The heated single-phase experimental images
are similar and are not shown for brevity. Figure 4(b)
represents the boiling flow images at a heated condition
of q" 56.9 kW/in". A bubble layer is shown on the
left side of the channel where boiling is occurring. From
these images bubbles can be discriminated from the flow
seedings due to differences in size, gray scale value and
distribution.
Figure 5 shows instantaneous liquid velocity fields
obtained from the PTV analysis of the experimental im-
ages at a Re = 14513. Figure 5(a) shows the veloc-
ity field from the unheated single-phase flow experi-
ments. Figure 5(b) shows the liquid velocity field ob-
tained from the boiling flow experiment. Vector gaps


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


`. Flow out

a
5 IC 9
12
312"
21) 12"




~low in


~x(mm)
455
*425
395
S365
335
305
275





P .''


f~- -- :


Pulsed Laser




Mirror High Speed
Mirror Camera




Halogen Lanp Channel


1.5


0.15










o o


Figure 2: Visualization system arrangement.


Table 1: Test sections names and positions
Name Distance from the inlet
P56 455 mm
P45 425 mm
P34 395 mm
P23 365 mm
P12 335 mm


osE


length of the channel (see Figure 2). Seven different
measurement areas along the channel were selected. The
distances from the measurement area to the channel in-
let are shown in Table 1. For each measurement area,
pictures were obtained using the high speed camera. A
Reynolds number (Re) of 14513 was considered in this
experiment, and a constant inlet temperature of 25.5 av
was maintained.
Figure 6 shows the mean axial velocity (17) and mean
axial turbulence intensity (u') for different measurement
area positions, both normalized by the mean centerline
axial velocity (Uc). The distance from the wall is nor-
malized by half of the channel height (h). The mean ve-
locity profiles along the test section agree quite well with
those of Wardana et al. (1992). The turbulence intensity
profiles showed \ignilk~.llll differences depending on the
measurement area position. These differences may be
the result of having a small ratio of channel width and
height (w/H 1 ), which means that the flow is not en-
tirely two-dimensional. This geometry was chosen to
simulate the small ratios of w/H typically found in the
coolant flow channels of boiling water nuclear reactors
(BWR).
In a fully developed flow of a constant-property fluid,
the mean normal velocity V should be zero. We mea-
sure nonzero values of this velocity but were only about
0.7% of the mean axial velocity component in the core
region of the flow. The normal velocity profiles are not
shown for these experiments. From the previous results,







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


(a) (b)


Figure 4: PTV experimental images (a) unheated
single-phase flow (b) boiling flow with q" =
56.1) k~lly".


are found at positions fully occupied by bubbles, con-
firming that only the liquid velocity is being obtained,
it is also noticeable that the vector magnitude difference
in between the two cases is larger in regions close to the
heated wall (left part of the channel).


Heat flux influence on the liquid-phase
behavior

In the following section, a brief discussion of the sub-
cooled boiling liquid phase velocity field measurements
is presented Figure 6(a) shows the profiles of mean
liquid axial velocity U for a Re 4638 with wall
heat fluxes ranging from 0 to 64 k~llyi. Wall heating
brought \ignilk~.lill changes in the distribution. First, the
mean liquid axial velocity in regions close to the wall
increased accompanied also with a decrease in the ve-
locity for regions far from the wall. Second, there was a
marked shift of the maximum liquid axial velocity loca-
tion toward the wall. For wall heat fluxes ranging from
3.1) to !).0 k~llyi the increase of velocity close to the
wall shifted the maximum towards a common position
located at about y = 2.5 m i. These heat fluxes shared
a common maximum velocity magnitude of about 0.2
in/s. Further increase of the heat flux forced the max-
imum velocity magnitude to increase and to shift closer

'Measurements were not made beyond y = 7.2 mm. The magni-
fication selected to obtain an optimal particle image size for the
PTV analysis, did not covered the whole width of the channel. We
focused on the heated wall, since most of the changes are in this
region


Figure 5: Instantaneous liquid velocity fields obtained
from experimental images for (a) unheated
single-phase flow, (b) boiling flow with q'



to the wall. This behavior is observed up to a heat flux of
42.3 k~llyi. The velocity profiles for these heat fluxes
collapsed towards a common position (y = 2.5 min).
These trends changed for the highest heat flux cases
(from 48.7 to 64.0 kW/in"). For these cases a new
behavior is found where the maximum velocity now
starts shifting towards the center of the channel, and it
does not increase much with the increase of heat flux,
reaching a maximum velocity of about 0.34 n/.s. Fur-
thermore, it is clear that the common collapsing point
found for the lower heat fluxes (y = 2.5 min) is not
present for the higher heat flux cases. Instead a shifting
of this point towards the center of the channel is noticed.
Figure 6(b) shows the profiles of mean liquid axial ve-
locity U for Re 14513 with wall heat flux ranging
from 0 to 64 kW/in". With this Reynolds number, the
influence of heat flux over the axial velocity follows the
same general trends as the ones observed for the lower
Reynolds case (Re 4638). However, additional fea-
tures should be pointed out such as: the increase of the
velocity for regions close to the wall is smaller, and the
velocity reduction effect for regions far from the wall is
still present, but with a smaller intensity. Furthermore,
the unheated single-phase profile, together with those of
low heat flux (0 to 16.0 kW/in"), showed no \ignilk~.slll
differences in between them, only small discrepancies

2Liquid velocity measurements close to the wall for the higher heat
fluxes are subject to an increased error, because of the vector sam-
pling reduction due to a thicker bubble layer








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


shows that the single-phase profile, together with those
of low heat flux (0 to 9 kW/m2), showed no \ignilk~.lllI
differences from the law of the wall, and only small dis-
crepancies for points close to the heater wall are noted.


for points close to the heater wall are noted. For medium
heat fluxes (18.3 to 35.9 kW/ ~~ ), there is an increase
of the axial velocity, persistent from y 0 to about
y 3.5 mm. For the heat fluxes (42.2 to 56.3 kW/m2),
the increase of axial velocity follows the same behavior
of the medium heat flux cases; however, the maximum
velocity location is shifted from the center of the chan-
nel toward the heated wall. When reaching the heat flux
of 64.0 kW/m2, the general behavior trend does not ap-
ply anymore. Now the maximum velocity location is
shifted to the right and a reduction of velocity magni-
tude is observed for points close to the wall. This trend
is similar to the one found for the higher heat flux cases
at Re 4638, with the difference that now the maxi-
mum velocity is about 0.54 m/s. Figure 6(c) shows the
profiles of mean liquid axial velocity U for Re 24188
with wall heat flux ranging from 0 to 64 kW/m2. It is
clear that at this Reynolds number, the heat flux influ-
ence seemed to be minimal. The increase of velocity
for regions close to the wall as a function of wall heat-
ing is small and velocity reduction for regions far from
the wall is not noticeable. Small changes start appear-
ing only up to a wall heat flux of 56.6 kW/m2, where
a small increase of velocity is observed for points in the
region from y = 1 to about 2.5 mm.


Liquid turbulence statistics

Since most of the changes in the liquid behavior due
to wall heating are observed close to the heated wall,
it is common to use wall coordinates. In the wall co-
ordinate system, a characteristic velocity is needed to
obtain non-dimensional variables. This characteristic
velocity was chosen to be the friction velocity (u* -
[ ~l1/p] The friction velocity can be estimated by
plotting the dimensional total stress profile of the un-
heated single phase case and taking a best fit of the
near-wall total stress to determine 7. Following the
previous procedure the friction velocities obtained for
Re = II.; 14513, and24188 are respectively u*=
0.012, 0.027, and 0.040 m/s. The overall influence of
wall heating over the axial velocity was explained in the
previous section, while some remarks concerning the ap-
plicability of the wall law are discussed. All results pre-
sented on this section are normalized by the correspond-
ing friction velocity for each Reynolds number. Figures
7, 8, and 9 show the axial velocity profile at various val-
ues of wall heating for Re I 11 >, 14513, and 24188.
From Figure 7, it is clear the dramatic influence of heat
flux for the low Reynolds number case. Only the un-
heated single-phase flow case (q" 0) has a fairly good
approximation to the law of the wall. For this case small
discrepancies of points below y+ 10 exist due to the
error induced by the heater wall reflections. Figure 8


Fi :-ii13


. *


0, 30


..


-


~020

0 00


0 000 0 002 0 D4
y(m)


0 0065


Re=14513







g.
d


_


0.20 -


0.00 _La


ill


0.000 0.002 0.04


Re=24188


1 20




,0 8


0.40


O 000 0 002 0 DD4 O000
y(m)


Figure 6: Mean axial liquid velocity profile for q" =
0.0 O, 3.9 , 16.0 V, 20.3 r>, 22.3 35.9 <1,
42.3 4, 48.7 O, 56.6 #, 64.0 O [kW, -







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


file even more and an inverted peak is observed. The lo-
cation of this inverted peak (minimum value) will shift
towards the center of the channel with wall heat flux in-
crements. The same trends are found for the medium
Reynolds number (Figure 17), where the profile reduc-
tion is more clear for points up to yt 380. For po-
sitions larger than this value, a smaller decrease is seen.
The zero shifting and peak inversion is also present in
this Reynolds number. For the highest Reynolds number
the effect of the wall heat flux starts to be noticeable only
at high q" values. There are not significant differences in
the profiles up to a heat flux value of 48.7 k~lm nt. For
this Reynolds number, the zero value shifting and peak
inversion were not observed.


Discussion

In this section, some of the mechanisms that govern
the fluid behavior are presented and discussed. First,
the heated single-phase behavior is explored to serve
as a basis to explain the further more complex mech-
anisms present when boiling appears. Previous works
discussed the effects of buoyancy forces on the single-
phase heated velocity fields (no boiling involved). Ac-
cording to Petukhov et al.Petukhov and Poyakov (1988)
,buoyancy affects the flow field in a heated channel via
two mechanisms. Firstly, the buoyancy force acts on
the entire flow because of the non-homogeneous fluid
density distribution, or the so-called external t Ite -t. The
second effect arises from the fluctuating liquid density in
the gravity field. This direct effect of buoyancy on tur-
bulence is termed the structural t:&*<-7l~. The transport of
momentum and thermal energy are influenced by a com-
plicated interaction between the structural and external
effects of buoyancy. In turbulent mixed convection in a
vertical channel, the structural effect appears first, influ-
encing the axial velocity in points away from the heated
wall. The external effect would not be \igni lk~.lll during
phases of low Gr/Re2. This explains the essentially un-
changed mean axial velocity near the heated inner wall
and a discernible effect farther from the wall (see Figure
6(c)). By increasing the wall heat flux, the external ef-
fects are expected to become more \ignlilk~.slll leading to
the well-known free convection effect of a fuller mean
axial velocity profile near the heated wall. The turbu-
lent kinetic energy in the proximity of the heated wall
is not influenced by the structural effects, at least for
low Gr/Re2 For points far from the heated wall, the
fluctuating buoyancy force contribution is expected to be
negative, and the production term which is positive, de-
creases in magnitude compared to the isothermal flow.

3The results of this work showed differently, even for low values of
GlesR esmalldbut s gnificanh differences of the turbulence inten-


For higher heat fluxes the discrepancies were larger.
For the highest Reynolds number, (Figure 9) most of
the heat flux cases considered followed the law of the
wall closely. It is clear that the heat flux influence is
dampened in the whole profile except for points below
yt 30. Figures 10, 11, and 12 show the axial turbu-
lence intensity u' at various values of wall heating for
He = II.. 14513, and24188. For the low Reynolds
number (Figure 10) the heat flux influence is large, even
for low heat flux cases. Two trends are observed: first,
from q" 3.95 to 9.02 k~lm nz, there is a decrease
in the turbulence intensities with respect to the isother-
mal case and the second trend is observed for the higher
heat fluxes where an increase of the wall heat flux will
increase drastically the turbulence intensity. The axial
turbulence intensity profiles for the medium Reynolds
number (Re 14513) are shown in Figure 11. Two
trends are observed here as well, but in this case the
first trend will persist for more heat flux cases, rang-
ing from q" 3.0 to 22.3 k~lm nt, where the profiles
are seen to be below or really close to the isothermal
case. The starting point for the second trend is shown
for wall heat fluxes larger than 35.9 k~lm n, but a dra-
matic increase of the turbulence intensity profile is ob-
served after reaching a wall heat flux of 48.7 k~lm nt.
Figure 12 shows the turbulence intensity profiles for
the highest Reynolds (Re = 24188). Here the wall
heat flux influence is dampened and only a small incre-
mental effect over the turbulence intensity profile is ob-
served. No profile is observed to be below the isother-
mal case. Figures 13, 14, and 15 show the normal tur-
bulence intensity profile (v') at various values of wall
heating for Re 4638, 14513, and 24188. These fig-
ures show similar trends than those found for the ax-
ial turbulence intensities (u'). The two trends discussed
previously are observed in Figure 13. The first heat
flux case q" 3.95 showed a decrease on the profile
compared to the isothermal case, while the rest of the
heat flux cases presented a drastic increase on turbu-
lence intensity due to the wall heating. Similar behav-
ior is found for the medium and high Reynolds number
however, by increasing the Reynolds number value, the
wall heating influence is reduced. The Reynolds stresses
u'v' profiles are shown in Figures 16, 17, and 18, for
Re 4638, 14513, and 24188. The wall heat flux
brought \ignilk~.lill changes on the u'v' profile. In gen-
eral a profile decrease tendency is found as a result of a
wall heat flux increment as well as a marked shift of the
zero location toward the heated wall can be observed.
For the low Reynolds number case (Figure 16), these
changes are large. At the beginning of heating from
q" 3.95 to 12.23 k~lm nt a peak reduction and shift
of the zero Reynolds stress location towards the wall is
found. Further increase of the heat flux reduced the pro-







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


nents were obtained, but also instantaneous 2D veloc-
ity fields were measured with high temporal and spatial
resolution. New and detailed information was obtained;
specifically for the cases of low Reynolds number and
high wall heat fluxes. The influence of buoyancy and
bubble interaction on the axial direction reached a max-
imum. Funther increase on the heat flux showed that the
influence of buoyancy and bubble interaction extended
normally to the heated wall. Both dimensional and non-
dimensional data are presented with the hope that they
will be useful in turbulence modeling efforts.

References

C.E. Estrada-Perez. Analysis, comparison and modifi-
cation of various particle image velocimetry (PIV) algo-
rithms. Master thesis, Texas A&M University, USA, Tx,
December 2004.

M. Lance and J. Bataille. Turbulence in the liquid phase
of a uniform bubbly air-water flow. Journal of Fluid
Mechanics, 222:95-118, 1991.

B. S. Petukhov and A.F. Poyakov. Heat Transfer in
Turbulent Mixed Convection. Hemisphere, New York
(1988), 1988'

M. Riethmuler. Particle Anage Velodinetry and Associ-
ated Techniques. Von Karman Institute, 2001.

R.P. Roy, V.S. Krishnan, and A. Raman. Measurements
in turbulent liquid flow through a vertical concentric an-
nular channel. Journal of Heat Transfer, 108:216-218,
1986.

R.P. Roy, V. Velidandla, and S.P. Kalra. Velocity field in
turbulent subcooled boiling flow. Journal of Heat Trans-
fer, 119:754-766, 1997.

R.P. Roy, S. Kang, J.A. Zarate, and A. Laporta. Tur-
bulent subcooled boiling flow -experiments and simula-
tioins. Journal of Heat Transfer, 124:73-93, 2002.

V. Velidandla, S. Putta, and R.P. Roy. Turbulent velocity
field in isothermal and heated liquid flow through a ver-
tical annular channel. international Journal of Heat and
Mass Transfer, 39:3333-3346, 1996.

I.N.G. Wardana, T Ueda, and M. Mizomoto. Structure
of turbulent two-simensional channel flow with strongly
heated wal. Experiments in Fluids, 13:17-25, 1992.

J.A. Zarate, M. Capizzani, and R.P. Roy. Velocity and
temperature wall laws in a vertical concentric annular
channel. international Jorunal of Heat and Mass Trans-
fer, 41:287-292, 1998.


Thus, suppression of turbulence is expected for regions
far from the heated wall (see the low wall heat flux cases
in Figures 10 to 15). In this work, the turbulence sup-
pression in regions farther from the wall, was found in
all cases, except for the higher Reynolds number case
(Re = 24188). For this case a consistent increase
with wall heat flux was found. We now briefly discuss
some of the physical implications of the subcooled boil-
ing measurements. From two-phase flow analysis sev-
eral conclusions have been obtained from other works.
Lance et al. Lance and Bataille (1991) suggested that
two-phase flow turbulence is the result of nonlinear in-
teraction between wall turbulence and bubble-induced
pseudo-turbulence, the latter being perturbations due to
random stirring of the liquid by the bubbles and defor-
mation of their surface. It has been conjectured that
these perturbations are proportional to the local vapor
fraction and the square of the vapor bubble velocity rel-
ative to the liquid, and additionally that they contribute
directly to the normal Reynolds stresses only. Through
dynamic interactions, they contribute to the Reynolds
shear stresses. In this work, the axial Reynolds shear
stress magnitude increased sharply near the inner wall,
where the vapor fraction was high, as was the normal
gradient of liquid mean axial velocity. The high vapor
fraction served to diminish the production rate, whereas
the high Reynolds shear stress and mean strain rate aug-
mented the production rate in the wall vicinity. It is im-
portant to note that one feature was found in this work
that was not measured or explained in previous stud-
ies. As observed by other researchers, at high values of
Gr/Re the maximum axial velocity was augmented
and shifted towards the heated wall. The consistency
of this behavior stopped when reaching a critical value
of Gr/Re At this value, the maximum axial veloc-
ity shifting towards the heated wall changed towards the
center of the channel. The maximum axial velocity in-
creased no longer with increments of the heat flux. This
behavior can be clearly seen in Figure 6(a). For high heat
fluxes (high Gr/Re2), buoyancy and bubble interaction
influence over the axial velocity in the axial direction,
reached a maximum. Once this point is surpassed, buoy-
ancy and bubble interaction influence start extending in
the normal direction towards the center of the channel.


Conclusions

Using PTV, liquid velocity fields of a turbulent sub-
cooled boiling flow in a rectangular channel were suc-
cessfully obtained. The present results agree with sim-
ilar studies that used point measurement probes [War-
dana et al. (1992), Roy et al. (1997), Roy et al. (2002)].
However, the present study provides additional informa-
tion; not only averaged profiles of the velocity compo-









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


Re=4638


Re=24188


P
P


20 -



15 -



10 -



5-



0-


kW-m


kW- m'

-q"= 0 0
-q"= 3 9
Sq"=16.0
q"=20.3
q"=22.3
q"=35.9
q"=42.3
q"=48.7
q"=56.6
aq"=64.0
SLaw of the wall


Sq"= 0 00
q"= 3 95
q"= 9 02
q"=12 23
Sq"=16 07
Sq"=18 31
Sq"=20 30
q"=22 31
Sq"=35 97
q"=42.30
q"=48.77
Sq"=56.68
q"=64 00
SLaw of the wall


111 1 il


100 101 102 103 104




Figure 9: Mean axial velocity profile at various values
of wall heating, normalized with friction ve-
locity u* = 0.04 m/s for Re = 24188


100 101 102 103 104




Figure 7: Mean axial velocity profile at various values
of wall heating, normalized with friction ve-
locity u* = 0.012 m/s for Re = 4638


Re=14513


Re=4638


10










-6




4




2


kW- m


kW- m-2

q"= 0.00
q"= 3.95
q"= 9.02
q"=12.23
q"=16.07
q"=18 31
q"=20 30
q"=22 31
q"=35 97
q"=42 30
q"=48 77
q"=56 68
q"=64 00


Sq"= 0 0
9= 3.9
Sq"= 9.0
Sq"=16.0
q"=18.3
q"=20.3
q"=22.3
q"= 35.9
q"=42.3
Sq"=48.7
Sq"=56.6
q"=64.0
SLaw of the wall


z "




m-
v*
***. $


100 101 102 103 104




Figure 10l: Mean axial fluctuating velocity at various
values of wall heating, normalized with fric-
tion velocity u* 0.012 m/s for Re
4638


100 101 10
y+


103 10'


Figure 8: Mean axial velocity profile at various values
of wall heating, normalized with friction ve-
locity u* 0.027 m/s for Re 14513









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






I Re=4638


Re=14513


4.00


i 1


kW- m_


kW- m-2

q"= 0.00
q"= 3.95
q"= 9.02
q"=12.23
q"=16.07
q"=18 31
q"=20 30
q"=22.31
q"=35.97
q"=42.30
q"=48 77
q"56 M8


q"= 0.0
q"= 3 9
Sq"= 9 0
q"=16 0
q"=18.3
Sq"=20.3
q"=22 3
n q=35 9
q"=42 3
o q"=48 7
Sq"=56 6
Sq"=64 0


3.00






2.00


o, t

a~~


100 101 102 10"




Figure 11: Mean axial fluctuating velocity at various
values of wall heating, normalized with fric-
tion velocity u* 0.027m/s for Re
14513


10' 102 103 104




Figure 13: Mean normal fluctuating velocity at various
values of wall heating, normalized with fric-
tion velocity u* 0.012m/s for Re
4638






4 Re= 14513


Re=24188


kW- m2

- q"= 0.0
-q"= 3.9
q"= 9.0
q"=16 0
Sq"=18 3
Sq"=20 3
q"=22 3
Sq"=35 9
Sq"=42 3
1 q"=48 7
* q"=56 6
Sq"=64 0


kW- m-,

Sq"= 0 0
q"= 3 9
q"=16 0
Sq"=20 3
q"=22 3
Sq"=35 9
Sq"=42 3
o q"=48 7
Sq"=56 6
q"=640


L'


Y o~
o 1

,


10' 102 103 10'


100 10' 102
Y+


Figure 12: Mean axial fluctuating velocity at various
values of wall heating, normalized with fric-
tion velocity u* = 0.04 m/s for Re =
24188


Figure 14: Mean normal fluctuating velocity at various
values of wall heating, normalized with fric-
tion velocity u* = 0.027 m/s for Re =
14513


5.


"
v . 1
..,-**









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


Re= 24188


Re= 14513


2.00 +


kW- m-2

- q"= 0.0
q"= 3 9
q"=16 0
q"=20 3





Sq"=64.0


kW- m-2

q"= 0.0
q"= 3.9
Sq"= 9.0



q"=42 3
Sq"=48.7



Sq"=56.6
Sq"=64.0


1.50 -






1.00-


1 1o


10' 102 103 104


0 200 400 600 800


Figure 15: Mean normal fluctuating velocity at various
values of wall heating, normalized with fric-
tion velocity u* = 0.04 m/s for Re=
24188


Figure 17: Reynolds stresses, normalized with friction
velocity u* = 0.027 m/s for Re = 14513


Re= 24188








1.: ";

1.


Re= 4638


kW- m-2


q"= 3.9
q"=16.0
q"=20.3
q"=22 3
Sq"=35 9
Sq"=42.3
Sq"=48.7
* q"=56.6
q"=6409


kW- m-2

Sq"= 0 00
q"= 3 95
q"= 9 02
q"= 1223
Sq"=16 07
q"=18 31
q"=20 30
q"=22 31
q"=35 97
Sq"=42 30
a q"=48.77
* q"=56.68
q"=64.00


/ . i~-


rh 8


o~ ?


0 150 300 450


Figure 18: Reynolds stresses, normalized with friction
velocity u* = 0.04 m/s for Re = 24188


Figure 16: Reynolds stresses, normalized with friction
velocity u* = 0.012 m/s for Re = 4638


I,-,




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