Group Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Title: P1.89 - Effects of Solid Particulates on Coherent Structures in a Horizontal Channel Flow
ALL VOLUMES CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102023/00475
 Material Information
Title: P1.89 - Effects of Solid Particulates on Coherent Structures in a Horizontal Channel Flow Particle-Laden Flows
Series Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Physical Description: Conference Papers
Creator: Li, J.
Liu, Z.
Wang, H.
Zheng, C.
Publisher: International Conference on Multiphase Flow (ICMF)
Publication Date: June 4, 2010
 Subjects
Subject: PIV
turbulence moderation
horizontal channel
coherent structures
 Notes
Abstract: Measurements of the turbulence modulation are made in the lower boundary layer of a fully developed horizontal channel flow with either 60 or 110 μm diameter polythene beads. A simultaneous two-phase PIV measurement technique is adapted to acquire the gas-particle turbulent statistics, and examine the modification of coherent structures in the near-wall region. The experiment is performed at shear Reynolds number of Reτ=427 based on the shear velocity uτ=0.445m/s and the half channel height h=0.015m; the corresponding Reynolds number of Reh=6380 based on the mean velocity Um=6.644m/s. The particle loadings conducted in the present experiment are very small, there are 0, 2.5×10-4, 1×10-3 and 5×10-3, the corresponding volume fractions are 0, 2.85×10-7, 1.14×10-6 and 5.7×10-6. It is found that the presence of particles distinctly modifies the gas turbulent motion, even though the mass loadings are only on the order of 10-4. The streamwise velocity fluctuations of gas phase decrease in the near-wall region and increase in the near-core region in appearance with particle loading increasing. In addition, the gas wall-normal velocity fluctuations and Reynolds shear stresses considerably increase in the vicinity of the wall and decrease in the outer region due to the presence of particles in all cases concerned. These phenomena can be explained using the mechanisms of the interactions between particles and the coherent structures through the two-point gas velocity fluctuation correlation, quadrant and Reynolds shear stress analysis. First of all, the results of the gas two-point correlations of both the streamwise and wall-normal velocity fluctuations show that the addition of particles result in weaker dominant coherent structures in the boundary layer, with shorter streamwise extent of the quasistreamwise structures. Moreover, it is also found that the frequencies of the Q2 (ejections) and Q4 (sweeps) events are both decreased by the particles from the quadrant analysis, together with remarkable reducing of the Reynolds shear stresses from both Q2 and Q4 events in all cases concerned. As a consequence, the particulate phase can significantly modify the overall carrier-phase turbulence through the effects on the turbulence structures in the boundary layer, and there is the fundamental mechanism for the turbulence modulation especially for the low mass loading case.
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: VID00475
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: P189-Li-ICMF2010.pdf

Full Text


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


Effects of Solid Particulates on Coherent Structures in a Horizontal Channel Flow


Jing Li, Zhaohui Liut, Hanfeng Wang and Chuguang Zheng

Huazhong University of Science and Technology, State Key Laboratory of Coal Combustion
Wuhan 430074, China
zliuiihust.edu.cn



Keywords: PIV, turbulence modulation, horizontal channel, coherent structures




Abstract

Measurements of the turbulence modulation are made in the lower boundary layer of a fully developed horizontal channel flow
with either 60 or 110 pLm diameter polythene beads. A simultaneous two-phase PIV measurement technique is adapted to
acquire the gas-particle turbulent statistics, and examine the modification of coherent structures in the near-wall region. The
experiment is performed at shear Reynolds number of Re,=427 based on the shear velocity u,=0.445m/s and the half channel
height h=0.015m: the corresponding Reynolds number of Reh=6380 based on the mean velocity Um=6.644m/s. The particle
loadings conducted in the present experiment are very small, there are 0, 2.5x10 4, 1x10" and 5x10-', the corresponding
volume fractions are 0, 2.85 x10 7, 1.14x10-6 and 5.7x10-6. It is found that the presence of particles distinctly modifies the gas
turbulent motion, even though the mass loadings are only on the order of 10 4. The streamwise velocity fluctuations of gas
phase decrease in the near-wall region and increase in the near-core region in appearance with particle loading increasing. In
addition, the gas wall-normal velocity fluctuations and Reynolds shear stresses considerably increase in the vicinity of the wall
and decrease in the outer region due to the presence of particles in all cases concerned. These phenomena can be explained
using the mechanisms of the interactions between particles and the coherent structures through the two-point gas velocity
fluctuation correlation, quadrant and Reynolds shear stress analysis. First of all, the results of the gas two-point correlations of
both the streamwise and wall-normal velocity fluctuations show that the addition of particles result in weaker dominant
coherent structures in the boundary layer, with shorter streamwise extent of the quasistreamwise structures. Moreover, it is also
found that the frequencies of the Q2 (ejections) and Q4 (sweeps) events are both decreased by the particles from the quadrant
analysis, together with remarkable reducing of the Reynolds shear stresses from both Q2 and Q4 events in all cases concerned.
As a consequence, the particulate phase can significantly modify the overall carrier-phase turbulence through the effects on the
turbulence structures in the boundary layer, and there is the fundamental mechanism for the turbulence modulation especially
for the low mass loading case.


Introduction

Gas-particle flow is common in nature, industrial and
environmental processes, as result of that the mechanism of
interaction between particles and gas has been investigated
extensively. The presence of particles with sufficient mass
loading can dramatically modify the gas flow, such as
increase or decrease turbulence intensity of carrier phase.
Because of the various parameters involved in determining
the turbulence modulation, the interactions between
particles and carrier phase are extremely complex. Although
much work has been done, the mechanisms of gas
turbulence modification due to the presence of particles
have not yet been clearly understood especially for the low
particle-loading case as result of the lack of information on
contributions of solid particles to the gas turbulence
structures. This paper will present an experimental
investigation in the lower boundary layer of a horizontal
gas-particle channel flow and provide the some direct
experimental data of the effects of particles on the coherent
structures to explain the turbulence modulation at very low
mass loadings.


From the previous work, the results show that small
particles, relative to the characteristic length scale of carrier
phase flow, generally tend to attenuate the turbulence
kinetic energy, while the large particles tend to augment it.
Gore and Crowe (1989) summarized results of former
researchers and identified the ratio of the particle diameter
to the turbulence integral length scale as the parameter that
can determine the regime of turbulence attenuation and
augmentation: if the ratio is greater than 0.1, particle
increase the turbulence intensity: contrariwise, when the
ratio is smaller than 0.1, particle suppress the turbulence
intensity. Furthermore, Hetsroni (1989) claimed that the
particle Reynolds number Rep is also a critical parameter:
particles with a small Rep trend to suppress the turbulence,
while particles with large Re, (>400) enhance the
turbulence due to the vortex shedding. Besides above
mentions, there are also many other factors related to the
turbulence modulation. For example, Fessler and Eaton
(1994) found that turbulence attenuation increases with the
mass loading and particle Stokes number, moreover they
(1999) also indicated that the flow regime strongly affect the
degree of the turbulence modulation. Sommerfeld and his






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

attentions will be paid to the lower boundary layer of the
channel. In the first parts of this work, the results of
turbulence modulation near core region of channel flow
have been acquired and it is shown that the presence of
particles could noticeably modified the carrier-phase
turbulent motion under very low mass loadings (5x10-4).
And then, the aim of present paper is to examine the
particle-induced modification of the near-wall turbulence,
especially for the modulation of the coherent stmectures due
to the presence of particles. The paper is organized as
follows: first of all, the preliminary measurements of clear
gas are performed and compare with the results of DNS
from Moser et al. (1999) to evaluate the validity of present
experiment: Secondly, the turbulence statistics of
particle-laden flow is measured and compared with that of
clear gas to examine turbulence modulation near the wall:
finally, the two-point correlations of velocity fluctuation of
both clear and particle-laden flow, as well as the quadrant
and Reynolds stress analysis, are investigated to examine
the modification of coherent structures, in the respect of the
dimension of quasi-streamwise structures and the
burst-frequency and -strength.

Nomenclature


coworkers have studied the effects of the particle-wall
collisions with roughness and inter-particle collisions on the
particle behaviour and gas turbulence modification through
the experiments and numerical simulations. Recently Sato
(2000) indicated that the inter-particle spacing was a more
critical parameter for turbulence modulation than the
particle diameter. Among these researches, although some
consensuses have been reached, many phenomena and
mechanisms of the turbulence modification by the particles
have not been fully explained and clearly understood due to
the four-way coupling in such high particle concentration.
For example, in the work of Tanibre et al. (1997), Kiger &
Pan (2002), Guo et al. (21I r 4, and Wu et al. (2006), although
the mass loadings, conducted from 0.06% to 4%, were
much lower in compared to the earlier literature, i.e. Gore &
Crowe (1989) and Hetsroni (1989), all results of these
experiments still showed that the presence of particles could
result in the obvious turbulence modulations. The
interactions between particles and turbulence structures
serve to interpret these interesting and surprising facts.
Turbulence coherent structures, such as low- or high- speed
streaks, quasi-streamwise vortices and ejection or sweep
associated with the structure, have been studied by many
investigators because they play key roles in the mechanism
of turbulence production and self-sustaining (Cantwell 1981,
Robinson 1991, Kasagi & Sumitani 1995, Adrian 2007).
Subsequently, as understanding improvement on turbulence
structures, the investigations of interactions between
particles and near wall turbulence stmecture also have been
performed more and more. Although numbers of studies
have been done, the majority of those were limited to the
modulation of the particle behaviour by the near-wall
turbulence stmecture (Brooke et al. 1992, Rouson and Eaton
2001, Marchioli and Soldati 2002, Kaffori et al. 1995 a & b-
Kiger & Pan 2002) and there is little information available
for the effect of particles on these stmectures. Rashidi et al.
(1990) were the first to conduct experiments aimed at
determining the turbulence structure modification by the
particles. They found that large particles (1100pLm) caused
an increase in the number of ejections, as well as turbulence
intensity and the Reynolds shear stress, while small particles
(120pLm) caused a decrease in the number of ejections.
Whereafter, Kaftori et al. (1998) also performed an
experimental investigation of particle-turbulence interaction
using LDA and flow visualization. It was found that the
particles increased the interval between bursts in
comparison to the particle-free flow. Recently, Dritselis and
Vlachos (2008) investigated the interaction of small heave
solid particles with turbulence coherent structure near the
wall of a vertical downward channel using direct number
simulation (DNS) and Lagrangian particle tracking. Their
results demonstrated that the presence of particles resulted
in a weaker mean coherent structure relative to that of the
particle-free flow. However, none of them have presented
and analyzed the comprehensive details of coherent
structures modifications due to the presence of particles,
with inclusion of the dimension of quasi-streamwise
structure, the frequency and the strength of bursts.
The present work is a following and extending part to the
work of Guo et al. (21 *4) and Wu et al. (2006), in which a
simultaneous two-phase PIV measurement technique,
including two-phase discrimination algorithm and PTV
algorithm, was developed and validated, and the special


diameter (Cpm)
channel half height (m)
wall Reynolds number
friction velocity (m s )
dimensionless distance from the wall
mean velocity (m s )
streamwise r~m.s. velocity (m s )
wall-normal r.m.s. velocity (m s )
bulk Stokes number
wall Stokes number


Greek letters
p density (kg m )
x, kinematic viscosity (m s ')
Mass loading ratio (x100%)

Subsripts
f fluid
P particle
cl center line


Experiments and procedure

Experimental Facility
The experiment was performed in a horizontal channel with
a height (2h), width and length of 3cmx30cmx600cm
respectively, as show in Fig. 1. The air flow was driven by a
frequency-controlled centrifugal blower, the tracers were
introduced through smoke chamber by means of producing
steam of paraffmn oil and the polythene particles (used as
solid phase and called particles later) were added in
upstream of the development section. Downstream of the
test section, the particles were removed from the flow by
the cyclone separator and reused: while the tracers were
cycled through the tracers recycle pipe. In the recycle pipe
end, flow passes through honeycombs to break up the large




































Table 1: PIV parameters
Parameter value

PIV image resolution (~l~iIll 12800 1024
PIV spatial resolution (pmll'p~itel 7.785
Frame interval (pLs) 18
Sampling rate (pairs/s) 3.75
Interrogation area (pixel ) 32 x16*

*(50% windows overlap in the x direction)


j II


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

two-dimensional flow is created in the center plane of test
section.
The measurement instrument was the PIV system assembled
by the TSI Inc. The measurement was taken in
streamwise-normal (x-v) plane (the center plane in the
spanwise direction) and the plane was illuminated by
aligned pulse laser sheet from two Nd: YAG lasers. A
PIVCAM 13-8 CCD, synchronized with the lasers, captured
pairs of images and transferred them to the computer as TIF
files. The imaging parameters were listed in Table 1.


eddies and to obtain a turbulent uniform flow. Afterwards
flow enters into a development channel whose length is 200
times channel height in comparison to 120 times in the
Kiger and Pan (2002). After passing through the
development section, according to available literature
(Alfredsson and Johansson, 1984: Eckelmann, 1974), the
flow could be considered to reach a fully developed state in
the position of the test section. Meanwhile, because the ratio
of the channel width to the height is 10 greater than 7.0, so
the side wall effects can be neglected and then a


Figure 1: Schematic diagram of experiment facility: 1 Blower, 2 Cyclone, 3 Stabilization tank, 4 Test section, 5 Development
section, 6 Conditioning and contraction section, 7 Tracer generator, 8 Tracer recycle pipe, 9 Particle feeder


Table 2: Channel cc


conditions and particle properties
1.1754
15.623 x106
8 8445


Fluid density, p, (kg xm ')
Kinematic viscosity, u (n? xs-')
Fluid centerline velocity, 11; (mxs')
Fluid mean velocity, L, (mxs')
Fluid friction velocity, u, (mxs')
Fluid response time, r, -0 u/ (ms)
Channel half height, h (m)
Mass loading ratio, # (x100%/)
Particle density, p, (kg m3)
Mean diameter, d, (pLm)
Nondimensional, d,' -dy, a
Size range (min~- max) (pLm)
Location slip velocity, max(u,-up) (mxs')
Particle Reynolds number, Rep
Viscous response time, r,=p,cf 18p (ms)
Corrected particle response time, r, (ms)
Corrected Particle setting velocity, V, (mxs')
Bulk Stokes number, Sth = zp U1 h
Wall Stokes number. St,, = r, u, o


6.644
0.445 (particle-free)
7.889x10-2
0.015
0.025, 0.1, 0.5
1030
60 110
1.71 3.13
53~74 97~125
0.996~1.162 1.204~1.355
3.83~4.46 8.48~9.54
11.218 37.705
7.905~8.147 22.096~22.833
0.077~0.08 0.217~0.224
4.66~4.80 13.03~13.46
100.2~103.3 280.1~289.4


Flow conditions and particles
In order to quantify the gas turbulence modification under
different particle sizes and mass loadings, the input power
of the blower was kept as a constant in all experiments, i.e.
constant rotate speed, which was different from those in
Kulick et al. (1994). In their experiments, the rotate speeds
were adjusted to promise a constant velocity at the
center-plane of channel.


All the experimental results are presented in a
non-dimensional form by the wall variables. In order to
improve the experimental validity, a precise evaluation of
the friction velocity is necessary. In present experiments, we
adopted a method of iterative procedure proposed by
Hagiwara (2002) and the friction velocity (uz) in the clear
gas flow was estimated to be 0.445m/s. The wall Reynolds
number Rez=uzxh/v=427, where h is the half height of






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

particles in all cases concerned, once the particle flux
appeared to have reached a steady state within the channel.
The captured two-phase images contain both particle and
tracer images. In order to get the velocities of the
solid-phase and gas-phase respectively, the solid particle
images were separated from the tracer images by a phase
discrimination algorithm, following those used by Guo et al.
(2002). Firstly, each digital image was filtered to reduce
high-frequency spatial noise and background noise was
removed with a local threshold. And then, a brightness
threshold and a size threshold could be determined through
analyzing the distributions of the size and brightness of
images and used to separate the tracer and particle images.
This separation method originates from the work of
Khalitov et al. (2002) and was improved by Guo et al.
(2002). It had been proved to be able to separate the two
phase images correctly and efficiently.
After the separation, the tracer-only images were filtered
again to eliminate the halos left by the particle images and
analyzed with the commercial software INSIGHT to
compute gas-phase velocity fields; while solid-phase images
was analyzed with a PTV algorithm introduced in the work
of Guo et al. (2Is1-1,1. In the PTV algorithm, the particle
central position was determined with a brightness-weighted
method, also described in the work of Guo et al. (2002). The
relaxation method introduced by Ohmi et al. (2000) was
adopted and simplified to track particles, since the particle
concentration was particularly dilute and there were few
particles in each searching area.


channel and v is the kinematic viscosity of fluid. The
Reynolds number base on the bulk mean velocity Um
(=6.644m/s) was 6417. The density of particles was
1030kg/m3 and two different sizes, 60pLm and 110pLm, were
used for dispersible phase. For each size, there are three
different particle mass loading ratios, including 2.5x10-4,
1x10-3 and 5x10-3, volume fractions from 2.85x10-7 to
5.7x10-6 correspondingly. The experimental conditions for
both clear-gas and particle-laden acquisitions are
summarized in Table 2.
The slip velocities used in the particle Reynolds number are
obtained by measuring the local relative velocity, and the
same procedure was adopted by Tanibre et al. (1997).
Correspondingly, the present value of both relative velocity
and particle Reynolds number are therefore more accurate
than those obtained from the mean slip velocity. Although
the particle Reynolds number is relative small varied from
3.83 to 9.54 in all cases concerned, the wake may also arise
in the downstream of particle because of the large velocity
gradient near the wall. Hereinafter, this surprising
phenomenon will be discussed in details and proved by
experimental data.
The methods adopted to calculate the properties of both
carrier phase and particles in present experiment are
identical with those used by Kiger and Pan (2002).

Image processing techniques
A series of 6000 image pairs of the two-phase flow field was
acquired, which promised there are sufficiency samples of


- DNS R=395
-*PIV R=440





(a)


0.2

0.0
1.0

0.8

0.6

0.4

0.2

0.0


100


100


y =yu /v y =yu /v
Figure 2: Comparison of single-phase PIV results with DNS simulations of Moser et al (1999). (a) Mean streamwise velocity,
(b) Reynolds shear stress, (c) streamwise turbulent stress component, (d) wall-normal turbulent stress component


Results and Discussion

Comparison of single-phase PIV results with DNS
For the single-phase conditions, 2000 image pairs were
acquired with a Model 630047 PIVCAM 13-8 CCD (1280
pixels x 1024 pixels) and processed with the INSIGHT


3.3TM. A comparison of the single-phase flow conditions
was made with the DNS simulations of Moser et al. (1999),
which are shown in figure 2. The DNS and PIV were
performed at a Reynolds number of 395 and 427, based on
the friction velocity and channel half-height, respectively.
The results for the streamwise mean flow are



























































U'
10 100
J' =yut v


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

processing with high resolution, which was adopted in
present experiment, is able to provide adequately accurate
results of gas turbulence statistic. The identical method has
been also adopted and demonstrated by the Liu et al (1991).
They indicated that the precision of the PIV with sufficient
resolution approached that of LDA. In the present
experiment, each measurement area has dimensions of (ax,
Ay)=(7.1x3.55) in wall units, corresponding (ax,
Ay)=(7.5x2.25) for the Liu et al (1991). In the next section,
interest will be focus on the modification of the turbulence
statistics due the presence of particles.


indistinguishable from the DNS, with less than a 1%
deviation for y'>8. The streamwise fluctuating velocity
agrees to within 8% for y'>17, and within 5% for
20 is within 13% for 1540. The
Reynolds shear stress variation is also within 15% for y'>10,
and within 5% for 30 it is believed that highly reliable measurements of the
higher-order moments are available in the outer log layer,
with a 10-15% variability for regions close the wall
(10 From the comparisons of figure 2, it is obvious that the PIV


-Fluid, No particles
-a- Fluid, 110pni, g=0.02500
Fluid, 110pni, =0.100
-r- Fluid, 110pni, g=0.500


- Fluid, No particles
-m- Fluid, 60pni, g=0.02500
Fluid, 60pni, =0.100
-r- Fluid, 60pni, g=0.500


J' =yut v


J' =yut v


Figure 3: Streamwise mean velocity profiles for both particle-free and particle-laden cases.


10 100
J' =J'tt /


Figure 4: Gas-phase streamwise velocity fluctuation for both particle-free and particle-laden cases.


-Fluid, No particles
- - Fluid, 60pni, g=0.0250 0
Fluid, 60pni, 9=0.10 0
-r- Fluid, 60pni, g=0.50 0


"----- -Fluid, No particles
-- Fluid, 110pni, g=0.0250
2 Fluid, 110pni, g=0.10
-r- Fluid, 110pni, g=0.50


Figure 5: Gas-phase wall-normal velocity fluctuation for both particle-free and particle-laden cases.






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


-i




-Fluid, No particles
-m- Fluid, 110pni, g=0.02500
Fluid, 110pni, =0.10
-r- Fluid, 110pni, g=0.50


0.8







0.2


0.0


0.8





7 0.

0.2


0.0


J' =yu/v


J' =J'u v


Figure 6: Gas-phase shear Reynolds stress for both particle-free and particle-laden cases.


Statistics of two-phase turbulent property
The gas-phase mean streamwise velocities are plotted in
wall scaling coordinates for both particle-free and
particle-laden cases in figure 3. It is found that the mean
velocities increase in the near-wall region and slightly
decrease in the near-core region with increasing mass
loadings, which show the drag effects of particles on the gas.
As a result of that, the profile of the gas mean streamwise
velocity for the particle-laden case becomes flatter than that
of the particle-free case. Furthermore, the results also show
that the thickness of the viscous sublayer was decreased by
the present of particles along with the gas streamwise
velocity gradient increased in this region.
Figure 4 shows the gas streamwise velocity fluctuations for
both particle-free and particle-laden cases. The presence of
particles causes the decrease in the near-wall region, while
the increase in the outer region. This interesting
phenomenon was also observed by the some former
investigators[], and it is similar to the results of the fixed
roughness elements. Careful insight into the profiles, it is
also found that the peak values of the streamwise velocity
fluctuations adjacent to the wall increase due to the presence
of particles, while the peak positions also shift to the wall.
These phenomena are consistent with the modification of
the viscous sublayer mentioned above: The increase of
velocity gradient in the viscous sublayer causes the
streamwise velocity fluctuation enhanced in the vicinity of
the wall, while the decrease of the thickness of the viscous
sublayer induces the peak position of the streamwise
velocity fluctuation to approach the wall. As a result of
those, the apparent trends of the overall streamwise velocity
fluctuations represent the decrease in the near-wall region
and the slight increase in the near-core region.
The profiles of the gas wall-normal velocity fluctuations are
plotted in figure 5. The addition of particles causes the gas
wall-normal velocity fluctuations increase with increasing
mass loading in the vicinity of the wall, while noticeably
decrease in the near-core region, especially for the largest
mass loading (5x10-3). The augmentations of the
wall-normal turbulence intensities near the wall may be
caused by the particle-saltation and the particle-wake. In the
vicinity of the lower wall, the saltation of the particle is very
intensive and the particulate concentration is also relative
high, thus the gas wall-normal turbulence intensities can be
increased by the particles through the drag force. In addition,


the particle diameter is larger than the characteristic length
scale of gas flow adjacent to the wall and the income flow
for the particle is also inhomogeneous with high velocity
gradient, thus the weak of the particles and the vortex
shedding may arise in the downstream of particles at
relative low particle-Reynolds-number and then increase the
wall-normal turbulence intensities. The attenuations of
wall-normal velocity fluctuations in the near-core region are
attributed to the suppressions of coherent structures by the
particles which will be discussed in details hereinafter.
The gas shear Reynolds stresses, as shown in figure 5,
exhibit the similar trends to those of the wall-normal
velocity fluctuations: in the region y <10, for the both size
classes, the Reynolds stress of gas (with particles) is
enhanced in comparison with the clear-gas result: whereas
in the region y >10, the Reynolds stress of gas is reduced by
the present of particles. The identical observation was
obtained in the Righetti & Romano work which was
conducted in the relative high mass loading.
From the results of figure 3 to 6, regarding the comparison
between the two size classes (60pLm and 110pLm), it is found
that the turbulence modulations with smaller particles
(60pLm) are more obvious than those with larger particles
(110pLm), especially for the lowest mass loading (2.5x10 )~.
This phenomenon may be attributed to the particle number
density: the number density is larger for the smaller particle
relative to the larger particle case at the identical mass
loading. And then, it also indicates that the particle number
density may be more critical than the particle diameter
under the low mass loading, when the differences between
two size classes are not very large such as 60pLm and 110pLm
in present experiment.
In order to interpret the mechanism of gas turbulence
modulation with particles, the effects of particles on the
turbulence structures in boundary layer will be discussed in
details from two-point correlation, quadrant analysis and
Reynolds stress.

Statistics of two-point correlation
The statistics of two-point correlations for streamwise and
wall-normal velocity fluctuation components, R,,, & R,,,. can
reveal the information about the scale of the gas turbulence
structures, such as streaks and quasistreamwise vortices.
The spatial two-point correlation R,,, & R,,. is defined as






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


1~

1


Two Point Correlation at Y =30
-Clear Fluid
- -Fluid, 60pLm, g=0.025%
Fluid, 60pLm, g=0.1%
~--A- Fluid, 60pLm, g=0.5%






50 100 150 200
Y -yu/v


0 50 100 150 200
X' xu /v


250 300


100 1
Y -yu /v


0 50 100 150 200 250 300
X =xu /v


Figure 7: Two-point correlations of streamwise and wall-normal velocity component in the X-Y plane, for the clear gas and
carrier gas with 60pLm particles


1 ~:


100 150
Y -yu /v


50 100 150 200
X -xu /v


250 300


100 150 200 0 50 100 150 200 250 300
Y -vu /v x -xu /v


Figure 8: Two-point correlations of streamwise and wall-normal velocity component in the X-Y plane, for the clear gas and
carrier gas with 110pLm particles






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

correspondingly, the particulate motions in the wall-normal
direction were remarkably intensified by the gravity. This
fact induced the intensive sedimentation and saltation of the
particles near the lower wall. ( Tanibre et al., 1997) In other
words, the coarse-wall effects and the crossing-trajectory
effects due to the addition of particles were both augmented
by the gravity in the horizontal channel, and thus the gas
coherent structures were suppressed with shorter streamwise
extent of the quasistreamwise structures.

Quadrant analysis
In the quadrant analysis, the frequencies of occurrence of
turbulence burst along the channel height are examined for
both clear-gas and particle-laden cases. The four quadrant
events adopted in the present experiments are defined as:
Quadrant 1 (Q1) represents (u'>0, v'>0) events: Quadrant 2
(Q2) represents (u'<0, v'>0) events, called ejections;
Quadrant 3 (Q3) represents (u'<0, v'<0) events: Quadrant 3
(Q3) represents (u'>0, v'<0) events, called sweeps. The
percentum of 4 quadrant events is defined as:


u, '(x,, ) xI u,(x)
Ru,, (x) = (1)

Where the bar indicates the temporal averaging.
Figure 7 & Figure 8 show the results of the two-point
correlations of the gas velocity fluctuations at wall-normal
location of y =30 in the X-Y plane. It is clearly found that
the presence of particles obviously reduces the correlations
with increasing particle loading, for both two size classes.
These results indicate the presence of particles suppresses
the turbulent structures through the cutting of
quasistreamwise structures in the streamwise direction.
As distinct from the present results, the numerical study of
Dritselis and Vlachos indicated that the addition of particles
elongated the quasistreamwise vortices in the streamwise
direction along with the mean structure decreased. This
distinction may be due to the difference of the flow regime
between their numerical simulations and our experiments.
Dritselis and Vlachos performed DNS in a vertical
downward channel. In contrast to that, the present
experiment was conducted in a horizontal channel:

0.22 Q1 Events




0.20 -




0 .16 - ... . ... .
10 100




0.30- Clrui Q3 Events








0.15


,,I.. Q2 E"vent~ s:ejetons
10 100


10 100 10


y~=yu_/o


Figure 9: Quadrant analysis of gas: percentum of 4 quadrant events.
S(A10nber (ie,actran -.9 )
Percentage(ieuactranf~t 9)= sample gladrant = (1, 2, 3, 4) ,

sample r=1


where sample represent the number of image pairs acquired
in one series.
The percentum of 4 quadrant events as a function of
wall-normal coordinate is plotted in the Figure 9. It is found
that the presence of particles reduces the percentum of Q4
events, and enhances the percentum of Q1, Q2and Q3
events. Furthermore, regarding comparison between Q2


events and Q4 events, it is also found that the degree of the
Q4 events suppression is much larger than that of Q2 events
augmentation under identical mass loading. And then, the
amount effects of both Q2 and Q4 events are reduced by
addition of particles.
Because the Q2 events (ejections) and the Q4 events
(sweeps) are associated with turbulence burst process, so the






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


region 1080) the
differences between clear-gas and particle-laden flows are
relative small: this results are coincident with the fact that
the coherent structures are dominant in the buffer region
( Jeong et al. 1997).


2.0


1.5


1.0

aReynolds stress from Q2 Event
0.5 -ClearFluid
--Fluid with 110pm~ 0.025%
ejections Fluid with 110pm~ 0.1%
-A- Fluid with 110pm~ 0.5%
0.0
10 100

y~yu/o




1.5


1.0


Reynolds stress from Q4 Event
0.5 C -ClearFluid
--Fluid with 1101rm_ 0.025%
sweeps Fluid with 1101rm_ 0.1%
-A~- Fluid with 110pmn_ 0.5%
0.0
10 100

y~yu/o


decrease of the amount percentum of Q2 and Q4 events
means that the frequencies of bursts are reduced. In other
words, the presence of particles causes a decrease in the
number of turbulence burst cycles and suppresses the
coherent structures. In the Figure 9, it is also shows that the
obvious modifications of 4 quadrant events occur in the


Reynolds stress from Q2 Event
-ClearFluid
-m- Fluid with 60pm~ 0.025%
ejections F"l"idwth 60pm 0.1%
-A~- Fluid with 60pm~ 0.5%


y+yur/o


hi
L
1
s
Y


y yu~/v


1


Reynolds stress from Q1 Event
-ClearFluid
-m- Fluid with 60pm~ 0.025%
Fluid with 60pm~ 0.1%
-A Fluid with 60pm~ 0.5%


Reynolds stress from Q1 Event
-ClearFluid
-m- Fluid with 110pm~ 0.025%
Fluid with 110rm~ 0.1%
-A- Fluid with 110pm~ 0.5%


y Ju,/o


y yu~/o


y Ju,/o


y yu~/o









Figure 10: Intensity of shear R
Shear Reynolds Stress
Among the 4 quadrant events, the Q2 events (ejections) and
Q4 events (sweeps) play important roles in the turbulence
production and self-sustaining mechanism of turbulence:
In the Q2 ejection events, low-speed fluid adjacent to the
wall is ejected by the passing of several streamwise vortices.
Reynolds shear stress product take a high value because of
negative fluctuating velocity in the streamwise direction and
positive fluctuating velocity in the wall-normal direction
during this event. On the other hand, high-speed fluid is
transported in the wall-ward direction by the streamwise
vortices during the Q4 sweep events. This also gives high
values of the Reynolds shear stress product. The Q2 and Q4
events product contribute mainly to the Reynolds shear
stress, which plays an important role for the production of
turbulent kinetic energy.
In order to gain a deeper insight into the intensity
modification of coherent structures due to the presence of
particles, an analysis of the shear Reynolds stress from each
quadrant events is performed as plotted in Figure 10. It is
clearly shown that the addition of particles causes the
Reynolds stresses from Q2 and Q4 events increase in the
near-wall region (y <15) and decrease in the outer region
(y >15), meanwhile the Reynolds stress from both Q1 and
Q3 events are remarkably reduced. Moreover, the degree of
the Reynolds stress modification increases with the mass
loading.
These results can be explained as follows: in the viscous
sublayer (y <15), as a result of high streamwise velocity
gradient, the wake may arise in the downstream of particles
and then the shear Reynolds stress (with particles) is
augmented in comparison with the clear-gas flow; whereas,
in the outer region (y >15) or buffer region where the
coherent structures are dominant, the presence of particles
suppress the turbulence bursts and reduce the intensity of
coherent structures, so the shear Reynolds stresses also
decrease correspondingly.


Conclusions

The modulations of the turbulence properties and structures
at relatively low mass loadings were both investigated using
the simultaneous two-phase PIV measurement technique.
The results show that the effects of particles on the coherent
structures are the important mechanisms for the turbulence
modulations, especially for the case of very low mass
loading. The particulate phase can significantly modify the
overall gas turbulence through the influence on the coherent
structures in the boundary layer:
(1) The gas coherent structures were suppressed with
shorter streamwise extent of the quasistreamwise
structures.
(2) The presence of particles reduces the percentum of Q4
events, and enhances the percentum of Q1, Q2and Q3
events. Furthermore, the amount effects of both Q2 and
Q4 events are reduced by addition of particles.
(3) In the near-wall region, the wake causes the shear
Reynolds stress increase; in the outer region, the
suppression of turbulence bursts and the reduction of
coherent structures intensity cause the shear Reynolds
stress decrease.


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

leynolds stress from each quadrant.
(4) The presence of particles causes the gas wall-normal
turbulence intensity and shear Reynolds stress
decreased in the near-core region through the
suppression of the coherent structures and increased in
the vicinity of the wall by the wake of the particle.
(5) The increase of streamwise velocity gradient in the
viscous sublayer and the decrease of the viscous
sublayer thickness cause the peak-values of the
streamwise velocity fluctuations increase along with
their positions shift to the wall, that is why the
streamwise intensity decrease in the near-wall region
and slightly increase in the near-core region in
appearance.


Acknowledgements

The study was partially supported by the National Natural
Science Foundation of China (50276021, 50576027), and
Program for New Century Excellent Talents in University,
Ministry of Education, China (NCET-04-0708)

References

Smith, C.R. Walker, J.D.A. Haidari, A.H. & Sobrun, U. On
the dynamics of near-wall turbulence. Phil. Trans. R. Soc.
Lond., Vol. A 336, 131-175 (1991)

Christensen, K.T. & Adrian, R.J. Statistical evidence of
hairpin vortex packets in wall turbulence. J. Fluid Mech.,
Vol. 431, 433-443 (2001)

Rashidi, M. Hetsroni, G. & Banerjee, S. Particle-turbulence
interaction in a boundary layer. Int. J. Multiphase Flow, Vol.
16, No. 6, 935-949 (1990)

Kaftori, D. Hetsroni, G. & Banerjee, S. Particle behavior in
the turbulent boundary layer I: Motion, deposition, and
entrainment, Phys. Fluids, Vol. 7, No. 5, 1095-1106 (1995)

Kaftori, D. Hetsroni, G. & Banerjee, S. Particle behavior in
the turbulent boundary layer II: Velocity and distribution
profiles, Phys. Fluids, Vol. 7, No. 5, 1107-1211 (1995)

Kaftori, D. Hetsroni, G. & Banerjee, S. The effect of
particles on wall turbulence, Int. J. Multiphase Flow, Vol. 24,
No. 3, 359-386 (1998)

Kiger, K.T. & Pan, C. Suspension and turbulence
modification effects of solid particulates on a horizontal
turbulent channel flow, J. Turbulence, Jot. 3, No. 019,
(2002)

Yoshimichi Hagiwara, Toru Murata, Mitsuru Tanaka &
Takumi Fukawa, Turbulence modification by the clusters of
settling particles in turbulent water flow in a horizontal duct,
Power Technology, Vol. 125, 158-167 (2002)

Wu, Y. Wang, H. Liu, Z. Li, J. Zhang, L. & Zheng, C.
Experimental investigation on turbulence modification in a
horizontal channel flow at relatively low mass loading. Acta
Mech Sinica, Vol. 22, 99-108 (2006)






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


Kulick, JD. Fessler, JR. & Eaton, JK. Particle response
and turbulence modification in fully developed channel flow.
J. Fluid Mech, Vol. 277, 109-134 (1994)

Guo, F. Wang, H. Liu, Z. & Zhang, L. PIV measurement
technique of dilute two-phase flow (in Chinese). Annual
Symposium on Engineering Thermophysics, Dalian,
353-359 (2002)

Khalitov, DA. & Longmire, EK. Simultaneous two-phase
PIV by two-parameter phase discrimination. Exp. Fluids,
Vol. 32, 252-268 (2002)

Guo, F. Wang, H. Liu, Z. & Zheng, C. Experimental
investigation on turbulence modulation in a horizontal
channel flow using PTV (in Chinese). Journal of
Engineering Thermophysics, Vol. 25, 622-624 (21 r 14)

Ohmi, K. & Li H. Particle-tracking velocimetry with new
algorithms. Meas. Sci. Techno, Vol. 11, 603-616 (2000)

Moser, R.D. Kim, J. & Mansour, N.N. Direct number
simulation of turbulent channel flow up to Re,= 590. Phys.
Fluids, Vol. 11, 943-945 (1999)

Tanibre, A. OesterlC, B. & Monnier, JC. On the behavior of
solid particles in a horizontal boundary layer with
turbulence and saltation effects. Exp. Fluids, Vol. 23,
263-271 (1997)

Dritselis, C. D. and Vlachos, N. S.. Numerical study of
educed coherent structures in near-wall region of a
particle-laden channel flow. Phys. Fluids 20, 055103 (2008)

Jeong, J. Hussain, F. Schoppa, W. and Kim, J. Coherent
structures near the wall in a turbulent channel flow. J. Fluid
Mech. 332. 185-214 (19971)




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