water

1. Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=tish20
ISH Journal of Hydraulic Engineering
ISSN: 0971-5010 (Print) 2164-3040 (Online) Journal homepage: https://www.tandfonline.com/loi/tish20
Study of depth-wise profiles of velocity and
turbulence parameters in the proximity of mid-
channel bar
Md. Amir Khan & Nayan Sharma
To cite this article: Md. Amir Khan & Nayan Sharma (2019): Study of depth-wise profiles of
velocity and turbulence parameters in the proximity of mid-channel bar, ISH Journal of Hydraulic
Engineering, DOI: 10.1080/09715010.2019.1570359
To link to this article: https://doi.org/10.1080/09715010.2019.1570359
Published online: 24 Jan 2019.
Submit your article to this journal
Article views: 1
View Crossmark data

2. Study of depth-wise profiles of velocity and turbulence parameters in the
proximity of mid-channel bar
Md. Amir Khan and Nayan Sharma
Department of Water Resources Development and Management, IIT Roorkee, Roorkee, India
ABSTRACT
The presence of braid bars increases the total flow resistance and the energy losses along the
boundary, thereby promoting the development of networks of hydraulically inefficient channels.
The mid-channel deposition is considered as one of the main reasons of initiation of braiding.
Extensive laboratory investigations have been carried out to study the turbulence flow hydraulics in
the vicinity of mid-channel bar. Present knowledge of braiding process and mainly mid-channel bar
formation is obtained from only few research works. The mechanism of mid-channel bar formation
is not properly understood. In this study, the turbulent flow structure in the vicinity of mid-channel
bar is investigated in detail. As could be seen from literature, hardly any work had been done on
this type of experimental setup. Thus, for validating the experimental results, the laboratory model
is developed in the Fluent Software. The Ansys Mesher is used for meshing the simulated model.
The Reynolds stress model is used for modelling purpose. The three-dimensional implicit steady
pressure-based solver is utilized for the computation. Discretization of pressure is done using the
Presto method. The momentum and Reynolds stress are discretized using the second-order upwind
scheme. The meshing of simulated mid-channel bar model is done using the Ansys Mesher. Velocity
measurements are taken with the help of Acoustic Doppler Velocimetry instrument. The effect of
mid-channel bar on the flow velocities profiles is investigated. The turbulent parameters profiles are
plotted for points in the vicinity of mid-channel bar. The changes occurred in the turbulent
parameters profiles due to the fluid and mid-channel bar interaction are studied.
ARTICLE HISTORY
Received 12 April 2018
Accepted 11 January 2019
KEYWORDS
Mid-channel bar; Reynolds
stress model; Acoustic
Doppler Velocimetry;
velocity profiles; turbulent
intensity
1. Introduction
Qualitative and geomorphic models have cited mid-channel
bar initiation and growth due to stalling of bed load around
the channel centreline as a primary process in braiding
(Ashmore 1982; Ashworth et al. 1992; Leopold and
Wolman 1957; Schumm and Khan 1972; Schumm and
Lichty 1963). Mid-channel bar formation, termed ‘central
bar deposition’ by Ferguson (1993) is still recognized as
a major cause of channel division in many braided rivers.
In spite of the significance of mid-channel bar, only a few
studies have been done on assessing the flow behaviour in
the vicinity mid-channel bar.
Present knowledge of braiding process and mainly
mid-channel bar formation is obtained from only few
research works (Akhtar et al. 2011; Bluck 1979; Ore
1964). The factors responsible for mid-channel bar devel-
opment are not yet completely known. Leopold and
Wolman (1957) reported that the mid-channel deposition
diverts the flow towards the banks which leads to the
erosion and positive feedback then accentuates bar devel-
opment and widening.
Some researches have been done in assessing the flow
structure in the vicinity of mid-channel bar (Detert et al.
2007; Nelson and Smith 1989), which are considered as the
precursors to braiding. However, these research works do
not analyse the turbulent flow dynamics and may be not apt
for the low sinuosity channels that are characteristics of
braided channels (Bridge 1993; Klaassen 1999; Lewin 1976).
Ashworth (1996) had carried out experiments on the
mid-channel bar development in a flume model. He had
presented only the direction and magnitude of surface velo-
city. Detailed analysis of velocity and turbulence parameters
in the vicinity of mid-channel bar is required for under-
standing the mechanism of mid-channel bar growth.
Only a few researches have been done on the flow beha-
viour changes observed due to mid-channel bar formation.
The experimental study on turbulent flow structure analysis
is necessary for understanding the flow behaviour in the
vicinity of mid-channel bar.
The study of flow characteristics in the vicinity of mid-
channel bar is vital for unravelling the braiding process.
Thus, it is imperative to undertake research into the turbu-
lent flow hydraulics in the vicinity of mid-channel bar
which has hardly been investigated with proper insight.
The present study addresses the variation of turbulent flow
structure that occurs due to the fluid bar interaction. Although
the static mid-channel bar is not precisely an actual represen-
tation of the natural prototype bar, the study for fixed bar will
undeniably improve the understanding of turbulent structure
in the vicinity of mid-channel bar. It will give an insight to the
comprehension of the phenomena which are vital for under-
standing the intrinsic fluvial processes of fluid bar interaction.
The deviations of mean velocity and turbulence due to the
presence of mid-channel bar need a special attention. The
present research is basically a small but important step for
enhancement in discerning the turbulent flow structure in the
vicinity of mid-channel bar. The main motive of this study is
to analyse the changes that occur in turbulent flow due to the
fluid bar interaction and to better understand the mechanics of
flow which are responsible for sediment entrainment and
CONTACT Md. Amir Khan amirmdamu@gmail.com
ISH JOURNAL OF HYDRAULIC ENGINEERING
https://doi.org/10.1080/09715010.2019.1570359
© 2019 Indian Society for Hydraulics

3. deposition in the vicinity of mid-channel bar. It is desirable to
study the mechanics of turbulence from the data collected
using Acoustic Doppler velocimetry (ADV) at different
locations.
As could be seen from literature, hardly any work had
been done on this type of experimental setup. Thus, for
validating the experimental results, the laboratory model is
developed in the Fluent Software. The Reynolds stress
model is used for modelling purpose.
2. Experimental setup and methodology
The experiments were conducted at River Engineering
Laboratory, Department of Water Resources Development
and Management, Indian Institute of Technology Roorkee,
India. The experiments were conducted in a flume of
2.6-m wide, 1-m deep and 10-m long (Figure 1). Experiments
were carried out at water flow rate of 0.3 m3
/sec. The tailgate
was used to maintain the desired flow depth.
The details of experiments are given in Table 1. The
velocities are measured at 24 different points for both
experimental conditions (Figure 1). The measuring points
are shown by the dot (Figure 1). The effect of fluid
bar interaction is more at location in the close vicinity of
mid-channel bar. The points at which effect of fluid and
mid-channel bar interaction is prominent are named from
a to l (Figure 1). The present study is mainly focused on
these named points.
The bed levels at these measuring points were taken
before starting and after completion of the experiments by
using the point gauge. The differences between the initial
bed level and the final bed level after the completion of
experiments are calculated, wherein negative values of dif-
ference represent the scouring and positive values represent
the deposition at those Points (Table 2).
The depth of flow is kept constant for both experimental
runs using the tail gate. The bed slope is kept constant at
0.005 for all experimental runs. Velocity is measured with
the help of ADV. The ADV is designed to record
instantaneous velocity components at a single point with
a relatively high frequency. For each point, the velocity is
measured at 10 different vertical distances from the bed.
The ADV is used to measure the three-dimensional
velocity components of flow. Measurements are performed
by measuring the velocity of particles in a remote sampling
volume based upon the Doppler shift effect. In most of
these cases, ADV is the technique of choice, because it is
relatively low in cost, can record at a relatively high fre-
quency up to 100 Hz. The velocity measurements are taken
at a frequency of 25 Hz in our experiment; this value is
taken on the basis of fact that the high frequency of mea-
surement may induce spike in velocity measurement
(Voulgaris and Trowbridge 1998). The sample volume is
a cylinder of diameter 4.0 mm and height of 5.6 mm.
A fairly recent near-bed turbulence measurement by
Radice et al. (2009) showed that a typical size of sweep
event contributing to bed load motion has a representative
length of 0.8 cm given the duration of 0.05 s and celerity of
Figure 1. Showing the sketch of the mid-channel bar model and measuring points (All dimensions in meters).
Table 1. Showing the details of experiments.
Experiment code
Discharge
m3
/sec
Bar size
(l  b  hbÞ cm Condition
Depth (h)
(cm)
No bar condition 0.3 No bar No bar 0.30
bar condition 0.3 100 by 70 by 10 Presence
of bar
0.30
Table 2. Showing the scouring and deposition patterns at different points in
the vicinity of mid-channel bar for bar condition.
Points Scouring/deposition in cm (bar condition)
a −4.56
b −5.48
c −3.73
d −4.63
e −0.88
f −1.29
g 0.17
h 0.36
i 0.41
j 0.74
k 0.86
l 0.98
2 M.A. KHAN AND N. SHARMA

4. 16 cm/s. This length is of similar order of magnitude as that
of a typical ADV sampling volume.
The technique employed in ADV is superior to the other
conventional methods, since the actual sampling volume is
located at a lower depth (0.05 m below the probe in the
present case) than the probe and hence is less disturbed.
ADV has been recommended as a capable instrument for
characterizing near-bed flow, particularly in the first 10 mm
above the bed (Finelli et al. 1999). The measurements of
ADV are affected by the noise contributions. Considering
the importance of noise, several methodologies have been
proposed to detect and remove noise from velocity signals
(Nikora and Goring 2000). Phase space threshold method is
the most widely accepted technique due to its non-
parametric characteristics (Cea et al. 2007; Mori et al.
2007). Therefore, this method is used for removing the
signal noise using the WinADV software.
The feasible method to determine the threshold
Correlation value (COR) is Shapiro–Wilks test statistic for
normality of distribution. The Shapiro–Wilks test done in
Khan et al. (2016) shows that the COR decline at 60.
Therefore, the cutoff value of COR is taken as 60. The
velocity data having COR value lesser than the 60 and
SNR value greater than the 15 is removed using the
WinADV software. Error associated with ADV data due
to spikes is removed by the method of (Nikora and
Goring 2000). This method is based on the concept of
phase space plot where the points outside ellipsoid are
defined as spikes. After the detection of spikes, they are
replaced by interpolated values. WinADV software is used
for filtering and removing the spikes from ADV data using
the method of (Nikora and Goring 2000).
Bed material is classified on the basis of grain size dis-
tribution. From the grain size distribution curve, represen-
tative bed size (d50) is calculated. Sand of d50 = 0.3 mm is
used as the bed material in the present study. The experi-
ments are performed in clear water condition.
For both experimental runs, the bed shear stress is kept
less than the critical shear stress τc for sediment movement.
As mentioned above, the bed shear stress is less than the
critical shear stress. Therefore, the scouring occurs only due
to the interaction of fluid and mid-channel bar.
In Table 1, l and b represent the length of major axis and
minor axis of elliptical mid-channel bar, respectively, and hb
represent the height of mid-channel bar.
The velocity fluctuations are defined as a variation of
temporal velocities. Algebraically these fluctuations in long-
itudinal, transverse and vertical directions are defined as
u;
¼ u À "u, v;
¼ v À "v, w;
¼ w À "w respectively. Here, u, v
and w are the instantaneous sample velocities in the long-
itudinal, transverse and vertical directions, respectively. "u,
"v and "w are the temporal mean velocities in the longitudi-
nal, transverse and vertical directions, respectively. The
temporal mean velocities are defined (Equations (1)–(3)).
"u ¼
1
n
Xi¼N
i¼1
ui (1)
"v ¼
1
n
Xi¼N
i¼1
vi (2)
-9
-6
-3
0
3
-6
-3
0
3
Q-3
Q-2Q-1
u'
Q-4
w'
Figure 2. Example of the fluctuating components of longitudinal and vertical velocity.
Figure 3. Showing the mesh of three-dimensional mid-channel bar model
developed in Fluent Software.
ISH JOURNAL OF HYDRAULIC ENGINEERING 3

5. "w ¼
1
n
Xi¼N
i¼1
wi (3)
where n is the total number of velocity samples. ui, vi and wi
are the magnitude of velocity in longitudinal, transverse and
vertical directions, respectively, for ith velocity sample.
Figure 2 shows the sample velocity fluctuations of long-
itudinal and vertical velocity with respect to the mean
value. Figure 2 shows the representative velocity fluctua-
tions and their corresponding quadrant. For example Q-1
represent the first quadrant, the longitudinal velocity fluc-
tuation u;
and vertical velocity fluctuation v;
are positive for
the first quadrant. In this same manner, the other quadrant
can be explained with the help of Figure 2.
3. Validation of experimental results using
the fluent software
As could be seen from literature, hardly any work had been
done on this type of experimental setup. Therefore, for
validating the experimental results, the commercial CFD
code Fluent is used. Steady-state simulations are carried
out using the Reynolds stress model of Fluent Software on
the basis of the study by Sarkar and Ratha (2014). The
Reynolds stress model (RSM) is one of the most elaborate
turbulence model inbuilt in the CFD code Fluent. In RSM,
the isotropic eddy-viscosity hypothesis is neglected. The
RSM closes the Reynolds-Averaged Navier-Stokes (RANS)
equations by solving transport equations for the Reynolds
stresses together with an equation for the dissipation rate.
This indicates that five more transport equations are
required for 2-D flows and seven additional transport equa-
tions are required for 3-D closure of RANS equations. The
RSM takes into accounts the effects of rotation, swirl, abrupt
change in strain rate and streamline curvature in a more
effective manner as compared to the one-equation and two-
equation models. It has a greater accuracy for complex flow
(Cambon et al. 1992; Durbin 1993).
Reynolds Stress equation models rely on the Reynolds
Stress Transport equation. The equation for the transport of
kinematic Reynolds stress.
Rij ¼ hu0
iu0
ji ¼ Àτij=ρ is
DRij
Dt
¼ Dij þ Pij þ Åij þ Ωij À εij (4)
In Equation (4), Rate of change of Rij + Transport of Rij by
convection = Transport of Rij by diffusion + Rate of pro-
duction of Rij + Transport of Rij due to turbulent pressure-
strain interactions + Transport of Rij due to rotation + Rate
of dissipation of Rij.
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0 5 10 15 20 25
Relativedepth(z/h)
Longitudional Velocity (cm/sec)
(a)
For bar (Exp) No bar (Exp) For bar (Fluent) No bar( Fluent)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0 5 10 15 20 25
Relativedepth(z/h)
Longitudional Velocity (cm/sec)
(b)
For bar (Exp) No bar (exp) For bar (Fluent) No bar (Fluent)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0 5 10 15 20 25 30
Relativedepth(z/h)
Longitudional Velocity (cm/sec)
(c)
For bar (Exp) No bar (Exp) For bar (Fluent) No bar (Fluent)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0 2 4 6 8 10 12 14 16 18Relativedepth(z/h)
Longitudional Velocity (cm/sec)
(d)
For bar (Exp) No bar (Exp) For bar (Fluent) No bar (Fluent)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0 2 4 6 8 10 12 14 16 18
Relativedepth(z/h)
Longitudional Velocity (cm/sec)
(e)
For bar (Exp) No bar (Exp) For bar (Fluent) For bar (Fluent)
Figure 4. (a), (b), (c), (d) and (e) show the depth-wise variation of longitudinal velocity for ‘b’, ‘d’, ‘f’, ‘h’ and ‘l’ points, respectively (Experimental and simulation).
4 M.A. KHAN AND N. SHARMA

6. At the top surface above the air, zero normal velocity and
zero normal gradients of all variables are applied by defin-
ing a symmetric boundary condition. The upstream inlet is
placed at sufficient distance from the mid-channel bar to
ensure that the flow becomes fully developed. The no-slip
boundary condition is specified to set the velocity to be zero
at the solid boundaries. The roughness height of the bottom
is specified according to the D50 of the bed material for
each simulated experiment. The side boundaries of the
domain are assumed to be smooth. Defining the side
boundaries as walls or symmetry conditions has been
found to have no effect on the flow field around the pier.
The mid-channel bar model is simulated using the Ansys
workbench. The meshing of simulated mid-channel bar
model is done using the Ansys Mesher. The view of the
mesh is shown in Figure 3. Face sizing method is used for
meshing the area in the vicinity of the mid-channel bar. The
element size for face sizing is kept at 0.001 mm for a region
in the vicinity of mid-channel bar. For the remaining por-
tion of the model, the element size is kept at 0.01 mm. The
low value of element size in the vicinity of mid-channel bar
is taken in order to study the velocity and turbulence para-
meters precisely at a region close to the mid-channel bar.
The three-dimensional implicit steady pressure-based
solver is utilized for the computation. Air and water are
considered as two immiscible fluids. The implicit volume
of fluid model and channel boundary conditions are used
for solving the momentum equation and track the volume
fraction of each of the fluids over the computational
domain. The velocity boundary condition is applied at
the inlet, whereas velocity outlet and velocity inlet bound-
ary conditions are utilized for the outlet. The turbulence
parameters are computed using the SIMPLE pressure-
velocity coupling. Discretization of pressure is done
using the Presto method. The momentum and Reynolds
stress are discretized using the second-order upwind
scheme. The convergence criteria for the residual of all
the parameters are kept at 1 Â 10À8
. The number of
iterations was kept at 1 Â 106
. The Initialization of solu-
tion was done using the standard initialization toolbox in
Fluent software. The comparison of experimental data
with the results obtained from the Fluent simulation will
be discussed in the next two sections.
4. Velocity distribution
The depth-wise profile of velocity distribution helps in under-
standing the behaviour of flow in the vicinity of mid-channel
bar. The longitudinal and vertical velocities are plotted against
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0 0.5 1 1.5 2 2.5 3
Relativedepth(z/h)
Vertical velocity (cm/sec)
(a)
For bar (Exp) No bar (Exp) For bar Fluent No bar (Fluent)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0 0.5 1 1.5 2 2.5
Relativedepth(z/h)
Vertical velocity (cm/sec)
(b)
For bar (Exp) No bar (Exp) For bar (Fluent) No bar (Fluent)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0 0.5 1 1.5 2 2.5
Relativedepth(z/h)
Vertical velocity (cm/sec)
(c)
For bar (Exp) No bar (Exp) For bar (Fluent) No bar ( Fluent)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
-2 -1 0 1 2 3
Relativedepth(z/h)
Vertical velocity (cm/sec)
(d)
For bar (Exp)
No bar (Exp)
For bar (Fluent)
No bar (Fluent)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
-1 -0.5 0 0.5 1
Relativedepth(z/h)
Vertical velocity (cm/sec)
(e)
For bar (Exp)
No bar (Exp)
For bar (Fluent)
No bar (Fluent)
Figure 5. (a), (b), (c), (d) and (e) show the depth-wise variation of vertical velocity for ‘b’, ‘d’, ‘f’, ‘h’ and ‘l’ points, respectively (Experimental and simulation).
ISH JOURNAL OF HYDRAULIC ENGINEERING 5

7. the relative depth of flow (z/h). Here, z is the vertical distance
from the bed, and h is the depth of flow.
Figure 4 shows the depth-wise distribution of longitudinal
velocity for five points. At points ‘b’, ‘d’ and ‘f’, the long-
itudinal velocity for bar condition is greater than the corre-
sponding value for no bar condition. This indicates that the
acceleration of longitudinal flow velocity occurred at these
points due to the presence of mid-channel bar.
For points ‘h’ and ‘l’, the longitudinal velocity for no bar
condition is greater than the corresponding values for bar
conditions. This indicates that the decrease in longitudinal
flow velocity occurred at these points due to the interaction
of fluid and mid-channel bar.
Figure 5 displays the depth-wise distribution of vertical
velocity for five points. For Points ‘b’, ‘d’ and ‘f’, the vertical
flow velocity is positive. The vertical velocity at points ‘b’, ‘d’
and ‘f’ for bar condition is greater than corresponding values
for no bar condition. This indicates that the lifting of flow at
these locations due to interaction of fluid and mid-channel bar.
The vertical velocity is negative for ‘h’ and ‘l’ points for
bar condition. For ‘h’ and ‘l’ points, the vertical velocity
profile is more scattered. Table 2 shows that the scouring
observed at points ‘b’, ‘d’ and ‘f’ and deposition occurred at
points ‘h’ and ‘l’.
The high longitudinal velocity and positive vertical velo-
city causes scouring at ‘b’, ‘d’ and ‘f’ points. Evidently, the
low longitudinal velocity and negative vertical velocity leads
to the deposition at ‘h’ and ‘l’ points.
The experimental velocity profiles are plotted along with
the simulated profiles obtained from the Reynolds stress
modelling. The experimental profiles are lying very close
to the simulation profiles (Figures 4 and 5). This indicates
that the experimental results are reasonably validated by the
Reynolds stress modelling.
5. Depth-wise distribution of turbulent parameters
The depth-wise distribution of turbulent parameters helps
in analysing the turbulence structure generated by the fluid
and mid-channel bar interaction. The depth-wise distribu-
tions of turbulent parameters are studied in detail in this
section. The depth-wise profiles of turbulent parameters are
analysed for all measuring points. But, for sake of clarity,
the depth-wise of turbulent parameters for only five points
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 2 4 6 8 10 12 14
RelativeDepth(z/h)
Normalized Longitudional Turbulent Intensity
(a)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 2 4 6 8 10 12 14 16
RelativeDepth(z/h)
Normalized Longitudional Turbulent Intensity
(b)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 2 4 6 8 10 12 14
RelativeDepth(z/h)
Normalized Longitudional Turbulent Intensity
(c)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1 2 3 4 5 6 7
RelativeDepth(z/h)
Normalized Longitudional Turbulent Intensity
(d)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1 2 3 4 5 6 7 8
RelativeDepth(z/h)
Normalized Longitudional Turbulent Intensity
(e)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
Figure 6. (a), (b), (c), (d) and (e) Shows the depth-wise variation of normalized longitudinal turbulent intensity for ‘b’, ‘d’, ‘f’, ‘h’ and ‘l’ points, respectively
(Experimental and simulation).
6 M.A. KHAN AND N. SHARMA

8. are displayed. These five points grossly represent the turbu-
lent structure of all measuring points.
5.1. Turbulence intensity
The root mean square (RMS) values of turbulence fluctuat-
ing velocity components (u0
; v0
and w0
) are normalized by
the shear velocity and presented in the form of Tiu ¼
ffiffiffiffi
u;2
p
uÃ
(in the longitudinal direction), Tiv ¼
ffiffiffiffi
v;2
p
uà (in the trans-
verse direction) and Tiw ¼
ffiffiffiffiffi
w;2
p
uà (in the vertical direction).
These normalized turbulent intensities are plotted against
the relative depth of flow (z/h).
Here ; uà is the shear velocity of approach flow. u0ð Þ2
,
v0ð Þ2
and w0ð Þ2
are the RMS value of fluctuating velocity in the
longitudinal, transverse and vertical direction, respectively.
Figure 6 shows the depth-wise profiles of normalized long-
itudinal turbulent intensity plotted for five points located in
the vicinity of mid-channel bar. The normalized longitudinal
turbulent intensity (TiuÞ values for bar condition is much
greater than the corresponding values for no bar condition.
For bar condition, the magnitude of Tiu for points ‘b’, ‘d’ and
‘f’ have significant greater magnitude as compared to the Tiu
values for ‘h’ and ‘l’ points.
The higher values of Tiu are observed for points located
near the upstream end of mid-channel bar (‘b’, ‘d’ and ‘f’) and
lower values of Tiu are observed for points’ (h’ and ‘l’) that are
located near the downstream end of mid-channel bar.
Figures 7 and 8 show the depth-wise profiles of Tiv and
Tiw respectively plotted for five points located in the vici-
nity of mid-channel bar. Similar to the plot of Tiu, the
depth-wise plot of Tiv and Tiw indicates that the magnitude
of these turbulent intensities is greater for bar condition as
compared to the no bar condition.
The depth-wise plot of Tiv and Tiw also indicates that
the magnitude of these intensities is greater for points
located near the upstream end of mid-channel bar (‘b’, ‘d’
and ‘f’ points).
The above discussion indicates that the fluid and mid-
channel bar interaction create the zone of high turbulence
region. The high turbulence region is located near the
upstream end of the mid-channel bar.
5.2. Turbulent kinetic energy (TKE)
In fluid dynamics, theTKE is the mean kinetic energy per
unit mass associated with eddies in turbulent flow.
Physically, the TKE is characterized by measured RMS
value of velocity fluctuations (Afzal et al. 2009; Kozioł
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.5 1 1.5 2 2.5
RelativeDepth(z/h)
Normalized Transverse Turbulent Intensity
(a)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.5 1 1.5 2 2.5
RelativeDepth(z/h)
Normalized Transverse Turbulent Intensity
(b)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.5 1 1.5 2
RelativeDepth(z/h)
Normalized Transverse Turbulent Intensity
(c)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.2 0.4 0.6 0.8 1 1.2
RelativeDepth(z/h)
Normalized Transverse Turbulent Intensity
(d)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.2 0.4 0.6 0.8 1 1.2
RelativeDepth(z/h)
Normalized Transverse Turbulent Intensity
(e)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
Figure 7. (a), (b), (c), (d) and (e) show the depth-wise variation of normalized transverse turbulent intensity for ‘b’, ‘d’, ‘f’, ‘h’ and ‘l’ points, respectively
(Experimental and simulation).
ISH JOURNAL OF HYDRAULIC ENGINEERING 7

9. 2015; Lien and D’Asaro 2006). The study of TKE is neces-
sary for thoroughly understanding the turbulent structure
of flow (Afzal et al. 2009).
The TKE is given by Equation (5).
TKE ¼
1
2
u0ð Þ2
þ v0ð Þ2
þ w0ð Þ2
n o
(5)
The TKE is normalizsed by the square of average shear
velocity uÃ
ð Þ of incoming flow in the uniform width portion
of model.
Normalized turbulent kinetic energy (NTKEÞ is given
by NTKE ¼ TKE
uÃ
2
The depth-wise distribution of NTKE is plotted for five
points (Figure 9).
Figure 9 indicates that the value of NTKE is maximum in
the near-bed region (z/h <0.15). The values of NTKE for bar
condition are much greater as compared to corresponding
values for no bar condition. It was also observed that the
high value of NTKE is observed at points located near the
upstream end of mid-channel bar (‘b’, ‘d’ and ‘f’). This
indicates that the fluid and mid-channel bar interaction is
present greatly only for a region located near the upstream
end of the mid-channel bar. Presence of mid-channel bar
causes the separation of a high NTKE zone that is located
near the upstream end of mid-channel bar from the low
NTKE zone that is located near the downstream end of the
mid-channel bar.
The high value of NTKE at ‘b’, ‘d’ and ‘f’ points for bar
condition plays an active role in extracting the sediment
from the bed which leads to the scouring at these locations.
The depth-wise profiles of turbulent parameters com-
puted from the experiments are compared with the simula-
tion profiles obtained using the Reynolds stress modelling
(Figures 6–9). The value and pattern of simulation results
are almost similar to the experimental results (Figures 6–9).
This indicates that the experimental results are in agreement
with the results obtained from the Reynolds stress
modelling.
6. Conclusions
The present research is basically a small but important step
for enhancement in discerning the turbulent flow structure
in the vicinity of mid-channel bar. The main motive of this
study is to analyse the changes that occur in turbulent flow
due to the fluid bar interaction. In the present research, the
laboratory investigations have been conducted to study in
depth the turbulent flow hydraulics in the vicinity of mid-
channel bar of an alluvial stream. From the literature sur-
vey, it could be found that the turbulent flow structure in
the vicinity of a mid-channel bar is not much researched in
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.5 1 1.5
RelativeDepth(z/h)
Normalized Vertical Turbulent Intensity
(a)
For bar (Exp)
No bar (Exp)
No bar (Fluent)
For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.5 1 1.5 2 2.5 3 3.5 4
RelativeDepth(z/h)
Normalized Vertical Turbulent Intensity
(b)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.5 1 1.5 2 2.5
RelativeDepth(z/h)
Normalized Vertical Turbulent Intensity
(c)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.5 1 1.5 2RelativeDepth(z/h)
Normalized Vertical Turbulent Intensity
(d)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.5 1 1.5 2 2.5
RelativeDepth(z/h)
Normalized Vertical Turbulent Intensity
(e)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
Figure 8. (a), (b), (c), (d) and (e) show the depth-wise variation of normalized vertical turbulent intensity for ‘b’, ‘d’, ‘f’, ‘h’ and ‘l’ points, respectively
(Experimental and simulation).
8 M.A. KHAN AND N. SHARMA

10. the past. Therefore, for validating the experimental results,
the commercial CFD code Fluent is used. Notably,
Ashworth (1996) observed that the mid-channel bar forma-
tion is mainly responsible for the initiation of braiding
process. The study of flow characteristics in the vicinity of
mid-channel bar is vital for unravelling the braiding pro-
cess. Thus, it is imperative to undertake research into the
turbulent flow hydraulics in the vicinity of mid-channel bar
which has hardly been investigated with proper insight.
(1) The experimental velocity profiles are plotted along
with the simulated profiles obtained from the
Reynolds stress modelling. The experimental profiles
are lying very close to the simulation profiles. This
indicates that the experimental results are reasonably
validated by the Reynolds stress modelling.
(2) The depth-wise profiles of turbulent parameters com-
puted from the experiments are compared with the
simulation profiles obtained using the Reynolds stress
modelling. The value and pattern of simulation results
are almost similar to the experimental results. This indi-
cates that the experimental results are in agreement with
the results obtained from the Reynolds stress modelling.
(3) The results indicate that the acceleration of long-
itudinal flow velocity is occurred at ‘b’, ‘d’ and ‘f’
points due to the presence of mid-channel bar. For
points ‘h’ and ‘l’, the decrease in longitudinal flow
velocity is occurred due to the interaction of fluid
and mid-channel bar.
(4) The high values of longitudinal velocity and positive
values of vertical velocity cause scouring at ‘b’, ‘d’ and
‘f’ points. Evidently, the low values of longitudinal
velocity and negative values of vertical velocity at ‘h’
and ‘l’ points lead to the deposition at these points.
(5) The normalized turbulent intensities values for bar
condition are much greater than the corresponding
values for no bar condition. The higher values of
normalized turbulent intensities are observed for
points located near the upstream end of mid-
channel bar (‘b’, ‘d’ and ‘f’) and lower values of
these turbulent intensities are observed for points’
(h’ and ‘l’) that are located near the downstream
end of mid-channel bar. The discussion indicates
that the fluid and mid-channel bar interaction cre-
ate the zone of high turbulence region. The high
turbulence region is located near the upstream end
of the mid-channel bar.
(6) The value of NTKE is maximum in the near-bed
region (z/h <0.15). The high value of NTKE is
observed at the points located near the upstream
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 2 4 6 8 10 12 14 16
RelativeDepth(z/h)
Normalized Turbulent Kinetic Energy
(a)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 5 10 15 20 25 30 35
RelativeDepth(z/h)
Normalized Turbulent Kinetic Energy
(b)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 2 4 6 8 10 12 14 16
RelativeDepth(z/h)
Normalized Turbulent Kinetic Energy
(c)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1 2 3 4 5 6 7 8
RelativeDepth(z/h)
Normalized Turbulent Kinetic Energy
(d)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 2 4 6 8 10 12
RelativeDepth(z/h)
Normalized Turbulent Kinetic Energy
(e)
For bar (Exp) No bar (Exp) No bar (Fluent) For bar (Fluent)
Figure 9. (a), (b), (c), (d) and (e) show the depth-wise variation of normalized turbulent kinetic energy for ‘b’, ‘d’, ‘f’, ‘h’ and ‘l’ points, respectively (Experimental
and simulation).
ISH JOURNAL OF HYDRAULIC ENGINEERING 9

11. end of mid-channel bar. Presence of bar causes the
separation of a high NTKE zone that is located near
the upstream end of mid-channel bar from the low
NTKE zone that is located near the downstream end
of the mid-channel bar.
(7) The high value of NTKE at ‘b’, ‘d’ and ‘f’ points for
bar condition plays an active role in extracting the
sediment from the bed which leads to the scouring at
these locations.
Disclosure statement
No potential conflict of interest was reported by the authors.
References
Afzal, B., Faruque, M.A., and Balachandar, R. (2009). “Effect of
Reynolds number, near-wall perturbation and turbulence on
smooth open-channel flows.” J. Hydraul. Res., 47(1), 66–81.
doi:10.3826/jhr.2009.3113
Akhtar, M., Sharma, N., and Ojha, C. (2011). “Braiding process and
bank erosion in the Brahmaputra river.” Int. J. Sediment Res., 26(4),
431–444. doi:10.1016/S1001-6279(12)60003-1
Ashmore, P.E. (1982). “Laboratory modelling of gravel braided stream
morphology.” Earth Surf. Process. Landf., 7(3), 201–225. doi:10.1002/
esp.3290070301
Ashworth, P., Ferguson, R., and Powell, M. (1992). “Bedload transport
and sorting in braided channels.” Dynamics of gravel-bed rivers,
P. Billi, et al., ed., John Wiley, New York, 497–515
Ashworth, P.J. (1996). “Mid-channel bar growth and its relationship to
local flow strength and direction.” Earth Surf. Process. Landf., 21(2),
103–123. doi:10.1002/(ISSN)1096-9837
Bluck, B. (1979). “Structure of coarse grained braided stream
alluvium.” Earth Environ. Sci. Trans R Soc. Edinb., 70(10–12),
181–221. doi:10.1017/S0080456800012795
Bridge, J.S. (1993). “The interaction between channel geometry, water
flow, sediment transport and deposition in braided rivers.” Geol.
Soc. Spec. Publ., 75(1), 13–71. doi:10.1144/GSL.SP.1993.075.01.02
Cambon, C., Jacquin, L., and Lubrano, J. (1992). “Toward a new
Reynolds stress model for rotating turbulent flows.” Phys. Fluids
A. Fluid Dyn., 4(4), 812–824. doi:10.1063/1.858298
Cea, L., Puertas, J., and Pena, L. (2007). “Velocity measurements on
highly turbulent free surface flow using ADV.” Exp. Fluids, 42(3),
333–348. doi:10.1007/s00348-006-0237-3
Detert, M., Weitbrecht, V., and Jirka, G. (2007). “Simultaneous velo-
city and pressure measurements using PIV and multi layer pressure
sensor arrays in gravel bed flows.” HMEM, Lake Placid.
Durbin, P. (1993). “A Reynolds stress model for near-wall turbulence.”
J. Fluid Mech., 249, 465–498. doi:10.1017/S0022112093001259
Ferguson, R. (1993). “Understanding braiding processes in gravel-bed
rivers: Progress and unsolved problems.” Geol. Soc. Spec. Publ., 75
(1), 74–87. doi:10.1144/GSL.SP.1993.075.01.03
Finelli, C.M., Hart, D.D., and Fonseca, D.M. (1999). “Evaluating the
spatial resolution of an acoustic doppler velocimeter and the con-
sequences for measuring near-bed flows.” Limnol. Oceanogr., 44(7),
1793–1801. doi:10.4319/lo.1999.44.7.1793
Khan, M.A., Sharma, N., and Singhal, G.D. (2016). “Experimental
study on bursting events around a bar in physical model of
a braided channel.” ISH J. Hydraul. Eng., 3, 1–8.
Klaassen, G. (1999). “Flow structure and transport of sand-grade
suspended sediment around an evolving braid bar, Jamuna River,
Bangladesh.”
Kozioł, A. (2015). “Scales of turbulent eddies in a compound
channel.” Acta Geophys., 63(2), 514–532. doi:10.2478/s11600-
014-0247-0
Leopold, L., and Wolman, M. (1957). “River channel patterns.” Fluv
Geom: Geom Crit Conc , Geological Survey, United States
Department of Interior, Washington, USA, Vol, 3
Lewin, J. (1976). “Initiation of bed forms and meanders in
coarse-grained sediment.” Geol. Soc. Am. Bull., 87(2), 281–285.
doi:10.1130/0016-7606(1976)87<281:IOBFAM>2.0.CO;2
Lien, R.-C., and D’Asaro, E.A. (2006). “Measurement of turbulent
kinetic energy dissipation rate with a Lagrangian float.” J. Atmos.
Ocean. Technol., 23(7), 964–976. doi:10.1175/JTECH1890.1
Mori, N., Suzuki, T., and Kakuno, S. (2007). “Experimental study of air
bubbles and turbulence characteristics in the surf zone.” J. Geophys.
Res. Oceans, 112(C5). doi:10.1029/2006JC003647
Nelson, J.M., and Smith, J.D. (1989). “Flow in meandering channels
with natural topography.” River Meandering, 4, 69–102.
Nikora, V., and Goring, D. (2000). “Flow turbulence over fixed and
weakly mobile gravel beds.” J. Hydraul. Eng., 126(9), 679–690.
doi:10.1061/(ASCE)0733-9429(2000)126:9(679)
Ore, H.T. (1964). “Some criteria for recognition of braided stream
deposits.” Rocky Mt. Geol., 3(1), 1–14.
Radice, A., Ballio, F., and Nikora, V. (2009). “On statistical properties
of bed load sediment concentration.” Water Resour. Res., 45(6).
doi:10.1029/2008WR007192
Sarkar, A., and Ratha, D. (2014). “Flow around submerged structures
subjected to shallow submergence over plane bed.” J. Fluids. Struct.,
44, 166–181. doi:10.1016/j.jfluidstructs.2013.10.002
Schumm, S., and Khan, H. (1972). “Experimental study of channel
patterns.” Geol. Soc. Am. Bull., 83(6), 1755–1770. doi:10.1130/0016-
7606(1972)83[1755:ESOCP]2.0.CO;2
Schumm, S.A., and Lichty, R.W. (1963). Channel widening and
flood-plain construction along Cimarron River in southwestern
Kansas, Geological Survey, United States Department of the
Interior, Washington USA.
Voulgaris, G., and Trowbridge, J.H. (1998). “Evaluation of the acoustic
doppler velocimeter (ADV) for turbulence measurements*.”
J. Atmos. Ocean. Technol., 15(1), 272–289. doi:10.1175/1520-0426-
(1998)015<0272:EOTADV>2.0.CO;2
10 M.A. KHAN AND N. SHARMA