Effect of free surface and froude number on the

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
__________________________________________________________________________________________
Volume: 02 Issue: 11 | Nov-2013, Available @ http://www.ijret.org 394
EFFECT OF FREE SURFACE AND FROUDE NUMBER ON THE
PROTECTION LENGTH AND TURBULENCE IN THE COMPOUND
TRANSITIONS THROUGH BRIDGES
Mahmoud.I.Attia1
, Abdullah Alhomidan2
1, 2
Civil Engineering Dept., Faculty of Engineering, Hail University, Saudia Arabia
m.elshewey@yahoo.com, aabdull@hotmail.com
Abstract
An experimental study was carried out to investigate the effect of free surface, Froude and Reynolds numbers on the protection length
and turbulence through compound transitions using Laser Doppler, in a rectangular channel. Measurements of turbulence are carried
out along the compound transition, at different contraction ratios, at different bed slopes. Vertical transition in the bed were also
changed. From the results, the protection length increases with the increase of the incoming Froude number, and increases with the
increase of the relative height, while it increases with the increase of the contraction ratio. The steep slope has a major effect on the
protection length. Free surface has a unique role in governing the turbulence in open channel flows. Within and downstream of the
transition, the turbulence occurs primarily in the wall region. Within the transition, flow tends towards the critical state (Fr
approaches close to 1) with a rise in the turbulence intensities in the free surface and the wall region which may be attributed to the
flow tending towards critical state. At small values of Froude number Fr ≤ 0.22, the effect of Froude number on turbulence intensities
and the free surface waves negligible.
Keywords: Turbulence intensities- Laser Doppler-Protection length-Free surface-Relative height-Contraction ratio-
Bottom slope-Froude number
---------------------------------------------------------------------***---------------------------------------------------------------------
1. INTRODUCTION
As the flow passes through a bridge, a channel transition in the
form of contraction and subsequent expansion is involved.
Flow in open channel transition being free surface flow, is
governed both by gravity and the viscous forces. Hence the
discussions are presented accounting both these effects in
terms of Froude and Reynolds numbers. Open channel
transitions have been studied extensively because of use in
water resources engineering and their efficacy in reducing the
energy loss in hydraulic structures. The information regarding
the turbulence characteristics in the transition structures is
somewhat scanty. It is well known that turbulent field in open
channel transition can be regarded as the superposition of
many eddies with different scales, and the fundamental
characteristics of turbulence are the intersection among these
structures. Paradoxically enough, the problem of separation of
the main stream of flow at open channel transitions or an
abrupt change of the boundary has attracted the attention of
investigators since the earliest times and yet it remains one of
the least understood and the most critical problems of fluid
dynamics today. In designing a channel transition, it is always
desirable to avoid excessive loss, to eliminate cross waves and
the turbulence, to ensure smooth streamlined flow, to
minimize standing waves, and to prevent the transition from
acting as a choke influencing upstream flow. Thus, one of the
purposes of this study of the turbulence in open channel
transitions is to gain insight about the role of the effect of
Froude and Reynolds numbers on the turbulence intensities in
open channel transitions. Therefore, precise and accurate
measurements of the mean and fluctuating flow quantities are
carried out to study the depthwise variation of the turbulence
intensities upstream, within and downstream of the transitions.
The turbulent flow models in open channel flows are
discussed by Chen and Jaw [2], Kim and Kline [7], Nezu and
Nakagawa [15] and Rodi [16]. Experimental investigation on
turbulent structure of back facing step flow have been reported
by several investigators such as Nakagawa [13]. Simulation of
turbulent flow at a Reynolds numbers, has been pointed by
Antonia [1]. Measurements of turbulence in open channel
flows have been reported by Nezu and Rodi [14], and
Kironoto [8], Nezu and Nakagawa [12].Turbulent boundary
layers have been reported by Clauser [3], and Kline and
Reynolds [9]. Turbulence and secondary flow over sediment
stripes has been studied by Mclelland and Ashworth [10]. The
present research involves measurements of the mean and
fluctuating flow quantities such as streamwise and vertical
mean velocity components u`/Uo and v`/Uo and streamwise
and vertical turbulence intensity components u`/Uo and v`/Uo,
and in a rectangular open channel with constant width. The

__________________________________________________________________________________________
measurements are carried out using a Laser Doppler
Velocimetry (LDV) a non-intrusive Fibre-optic state of the art
technique, in sudden and gradual channel transitions at
channel contraction rations Δ b/b 0.3 and 0.5. The main
objectives of the present study are, (a) To use the laser
technique (LDV), which includes the date acquisition system,
data processing to measure mean and fluctuating flow
characteristics at different locations of the channel transitions,
(b) To study the influence of free, Froude number (Fr) and
Reynolds number (Re) on the mean and fluctuating flow
quantities such as streamwise and vertical mean velocities
u`/Uo and v`/Uo, streamwise and vertical turbulence intensities
u`/Uo and v`/Uo upstream, within and downstream of the
channel transitions, (c) A comparative study of the depthwise
variation of the velocity components u`/Uo and v`/Uo (d) A
comparative of the depthwise variation of turbulence
intensities u`/Uo and v`/Uo at different cross sections along the
centerline of the channel. The experiments were conducted at
different values of Fr and Re viz. 0.240 , 0.090 and 0.5 105
,
and 0.15 105
respectively, generated by altering the depth of
flow, and (e) To study the effects of channel contraction,
bottom slope., approaching flow Froude number and vertical
transition on the relative protection length downstream of the
transition (sudden expansion).
2. THEORETICAL STUDY
The variables affecting the protection length (L) or the
distance defined by spacing between section [2] and section
[3] in Fig.(3), may be defined as: velocity of approach (vu),
channel bottom with (B), flow width within transition zone
(b), water depth of approach (yu), transition length (Lt),
channel bottom slope (So), gravity acceleration (g) and hump
height within the transition(h). The functional relationship of
the protection length downstream of sudden expansion could
be written as:
ƒ1 (y, , vu, yu, h , Δ b , Lt , So , L) = 0 (1)
Using the dimensional analysis principles (Pi theorem) based
on vu, yu and by repeating variables, the following
dimensionless relationship of the relative protection length
(L/b), could be expressed as:
(2)
Keeping in mind the properties of the dimensional analysis,
dividing L/yu by b/yu , to get L/b, Δb/yu by b/yu to get Δb/b,
h/yu by b/yu to obtain h/b and neglecting the effect of Lt as it
was kept constant during the course of the experimental work,
the following relationship is obtained.
(3)
In which Fu is the approaching flow Froude number, yu/b is
the relative approaching depth, h/b is the relative height of the
vertical transition and Δb/b is the contraction ratio.
3. EXPERIMENTAL SET UP AND
EXPERIMENTAL PROGRAM:
The experiments were carried out in a rectangular open
channel that is 8.0 m long, 0.3 m width and 0.5 m height with
glass wall 6mm thick and thick and a steel plate bed. Fig.1
shows layout of the test facility. The water is supplied from a
constant head overhead tank to the flume at a desired
discharge that is continuously monitored with an on-line
orifice meter. The discharges were measured using a pre-
calibrated orifice meter fitted in the feeding pipeline. And in-
line discharge control valve that is fitted into the main
supplying pipeline was used to regulate the flow rate. Depth
measurements were taken using a needle point gauge with a
reading accuracy of ±0.010 mm. The flume side walls are
made up of 6mm thick glass sheets. A tail gate is provided at
the downstream end of the flume to maintain a required water
depth of the channel flow. The water is finally collected in a
sump placed in the basement from where it is pumped back to
the overhead tank by a 15 hp pump. The experiments were
carried out using five different lateral contraction ratios, Δb/b,
of 0.1,0.2,0.3,0.4,0.5 and 0.6. Five different channel bottom
positive slopes, So, of 0.00, 0.005,0.01,0.015,0.02 and 0.025
were used to illustrate the effect of bottom slope So on the
protection length L/b due to local contraction. The slopes were
selected based on the flume facilities. For each combination of
lateral contraction ratio and bottom slope, five different flow
rates ranging from about 45 lit/sec to 15 lit/sec were use. The
upstream water depth was adjusted to produce a Froude
number Fu of approach ranging from 0.11 to 0.45. The flow
through the transition was always subcritical but it may
changed to supercritical state just at the end of the transition or
away from it, depending upon the incoming flow rate, the
applied flume bottom slope So and the contraction ratio Δb/b.
For each combination of contraction ratio and bottom slope,
the flow rate and the water surface profiles were measured.
The effect of the vertical sudden transition on the protection
length (L) was also studied, for a different lateral contraction
ratios Δb/b and a different bottom slope So. Five bed vertical
transitions, within the lateral contraction, were used with
relative heights of h/b =0.1,0.2,0.3,0.4,0.5 and 0.6 for each
particular value of h/b, five different discharges were
considered. Channel transitions were fabricated from
transparent prespex sheets. The construction at the inlet and
expansion at the outlet were, sudden.
With reference to the origin fixed at the channel bed and in the
centre of the channel transition, traverse of measuring volume
was run to obtain the profiles of both the mean velocity
components and root mean square (RMS) of the turbulence
intensities. The measuring points were closely spaced in the

__________________________________________________________________________________________
region of high velocity gradient. The Reynolds number based
on the free stream velocity 4 104
which ensured the turbulent
how for all the test conditions. Froude number of the free
stream flow Fr=0.24, ensured the free stream flow to be
subcritical. To obtain the vertical profiles the mean and
fluctuating flow quantities, the measurements were conducted
in the vertical plane at z/b=0 and z/b=0.25 at Q=40 1/s.
4. LASER MEASURING TECHNIQUE
Laser Doppler Velocimetry (LDV). A 5 Watt Argon-ion Laser
with two Laser beams; one blue (488 nm) and one green
(514.5nm), were focused at a measuring point from one side
of the channel through on optical lens. Two burst spectrum
Analyzers(BSA)were used to evaluate the Doppler
frequencies. Subsequent computer analysis consisted of
velocity bias averaging and outlier rejection. The number of
samples taken at every point was 4000 bursts. This
corresponds to a simple averaging time of about 80-100
seconds. The data rate was about 50-60 per second. Before
acquiring the data, the LDV signal was checked for its regular
Doppler burst that correspond to a particle passing through the
measuring volume. The measurements were taken at different
cross sections upstream, within, and downstream of the
transition. Figure 3 shows the location grid of the measuring
stations for all the channel transitions.
5. RESULTS AND DISCUSSION
The relationship between the relative protection length L/b
and the approaching flow Froude number Fu is shown in Fig.4,
for different bed slope So at particular contraction ratios Δb/b
of 0.1,0.2,0.3,0.4,0.5 and 0.6, respectively. It can be noticed
that the protection length L/b increases as the bottom slope So
increases at particular Fu and constant value of contraction.
Also, L/b increases as the Froude number Fu increases, at
particular slope and fixed value of contraction ratio.
Comparing figures, it is obvious that as the contraction ratio
Δb/b increases from Δb/b=0.1 to 0.5,L/b increases,
considering particular bottom slopes. This observation is valid
for all tested bottom slope and all tested range of Fu. These
observations are attributed to the increase in the velocity of
flow as the width of the channel is reduced from b to (b-Δb)
leading to longer L/b at the small contraction. Since the
incoming flow is subcritical and the incoming unit discharge is
smaller than the unit discharge through the transition zone, the
depth of flow is lowered resulting in higher velocity because
the pressure head is converted into velocity head. This higher
velocity of flow that exits from the transition zone forces the
jump to form away from the end of the transition leading to
longer protection length L/b compared to the cases of larger
contraction ratios Δb/b. Similarly, the resulting high velocity
due to the steeper bottom slope So causes the jump to form
away from the end of the transition zone yielding longer
protection length compared to flatter slopes.
The relationships between L/b and the approaching Froude
number Fu are presented in Fig.5 for different contraction
ratios Δb/b and particular slopes of
So=0.00,0.005,0.01,0.015,0.02 and 0.025 respectively.
Inspection of this figure indicates that the trend of variation of
L/b with Fu is similar for different slopes and different
contraction ratios Δb/b. It can be noticed that at fixed bottom
slope So and Δb/b, the relative protection L/b increases as Fu
increases. It can be observed that the relative length of
protection is longer for the bigger contraction ratio at
particular Fu when the bottom slope is kept constant.
Moreover, the present research investigated the effect of a
vertical transition formed within the lateral constriction zone
on L/b for different Froude number Fu and different bottom
slope So and fixed contraction ratio Δb/b=30%, as mid values
of the tested ranges. The relationship between the relative
protection length L/b and the relative vertical transition h/b for
different Froude number Fu at different bed slopes So is
presented in Fig.6. It can be easily observed that at fixed
bottom slope and Froude number, the relative protection
length L/b increases as h/b increases. Also, L/b increases as
the relative vertical height h/b increases at the same Froude
number Fu. This may be attributed to additional increase in the
velocity head due to the transformation of the pressure head to
velocity head when the flow meets the vertical transition. This
can be simply explained as follows: since the approaching
flow is subcritical and the incoming energy is higher than the
energy at a downstream section, consequently, according to
the specific energy principle, the depth of flow is lowered
resulting in higher velocity which is turn leads to a longer
projection length.
Root mean square (RMS) values of streamwise and vertical
components of turbulence intensities u` and v` are made non-
dimensional with respect to the steamwise mean free stream
velocity Uo. Figs.8,9 and 10 depict the variation of non
dimensional streamwise and vertical components of
turbulence intensities u`/Uo and v`/Uo as a function of relative
channel depth y/yo in sudden channel transitions for
contraction ratios Δb/b of 0.3 and 0.5 at different locations
upstream, within and downstream the transitions along the
length for different discharges (15-40)Lit/sec. Reynolds and
Froude numbers are respectively 0.4 105
and 0.24 for free
stream flow. The trend of variation of streamwise and vertical
turbulence intensities u`/Uo and v`/Uo are similar in all the
cases of channel transition. The trend of u`/Uo and v`/Uo
upstream, within and downstream in all the cases of transition
have higher values close to the bed, following a gradual fall in
the wall region (wall effect) defined by y/yo≤0.2, reaching
minima in the middle region (core region) defined by
0.2<y/yo≤0.6. Turbulence intensities u`/Uo and v`/Uo rise
gradually and then rapidly in the upper region (free surface
region) defined by y/yo>0.6 reaching the maximum at the free
surface. Generally, the minimum streamwise and vertical

__________________________________________________________________________________________
turbulence intensities u`/Uo and v`/Uo always lie in the core
region at the same location of the profiles of channel
transition. The maximum turbulence intensities u`/Uo and
v`/Uo close to the bed or at the free surface depending on the
location of the profile station at the same location of the of the
profiles of transition. Also, as a comprehensive observation, it
is noted that the streamwise turbulence u`/Uo is always
stronger compared to the vertical v`/Uo. In the wall region
defined by y/yo≤0.2, the turbulence intensities u`/Uo and v`/Uo
have substantially large magnitude closer to the wall (wall
effect). With the increasing distance from the boundary, both
the turbulence intensities decrease in wall region tending
towards a minimum in the intermediate region (core region)
defined by 0.2≤y/yo≤0.6, where the location of the minimum
value of the turbulence consistently corresponds to that of
maximum streamwise mean velocity ū/Uo where the velocity
gradient is locally zero, reaching a higher value in the free
surface.
Figs.8 and 9 depict the variation of u`/Uo and v`/Uo in the
transition for different Fr and Re viz. 0.240,0.090; and
0.5 105
and 0.5 105
, for free stream respectively at
contraction ratio Δb/b=0.3, comparison of u`/Uo and v`/Uo
upstream, within and downstream the transition, indicates that
an increase in the Froude number and Reynolds number,
decreases the turbulence intensities u`/Uo and v`/Uo upstream
and within the transition slightly with a slight increase in the
turbulence intensities downstream of the transition in the
expansion zone at the same flow conditions. This slight
decrease in the turbulence intensities u`/Uo and v`/Uo with
increasing Fr and Re occur in the wall region and the free
surface region and slight corresponding increase occurs in the
downstream region. However, in the core region no noticeable
change in the turbulence levels upstream, within and
downstream region of the transition occurs. This increase or
decrease in Fr and Re is not associated with the drastic change
in the nature of turbulence intensity profiles maintaining the
similarity of turbulence profiles with Fr and Re in the
transition. This observation is consistent with the observation
reported by Nezu and Nakagawa [14].
Figure 10 depict the turbulence intensity profiles along the
depthwise and channel axis, for contraction ratio Δb/b=0.5. In
the upstream region of the transition, the Froude and Reynolds
numbers are 0.23 and 4 105
on an average respectively,
turbulence is relatively lower compared to the turbulence
within and downstream regions of the transition. Upstream of
the entrance to the transition, surface waves were relatively
mild. Upstream the transition, since the incoming flow is
subcritical and turbulent as Froude and Reynolds numbers are
0.24 and 4 105
on average respectively, for x/b<-1,
turbulence is relatively lower upstream of the transition. At the
location of x/b=-4, z/b=0; x/b=-3, z/b=0.25; x/b=-3, z/b=0;
and x/b=-1, z/b=0 at Reynolds number of 4 105
and the range
of Froude number 0.218<Fr<0.229, clearly, the trend of
variation in streamwise and vertical turbulence intensities
u`/Uo and v`/Uo are similar in all cases. In all the cases bath
u`/Uo and v`/Uo, the maximum turbulence intensities occur
close to the bed, following a gradual fall in the wall region at
y/yo<0.2, reaching up to the free surface. It is also noted that
the values of u`/Uo and v`/Uo are approximately constant in
the core region. At x/b=-1, z/b=0, that is the centre just before
the entrance of transition, the trend in variation curve of u`/Uo
and v`/Uo is parallel. It has been observed during this
experimentation, that surfaces waves play am important role in
the turbulence production. Before the entrance to the
transition, surface waves were relatively mild. Hence far away
from the entrance, at x/b=-4 for instance, the turbulence
intensities in the core region and at the surface are relatively
low. The turbulence intensities are almost the same up to x/b=-
2 for Δb/b=0.5. As the flow approaches the entrance, there is a
gradual increase in the turbulence intensities u`/Uo and v`/Uo
in the core region as well as at the surface, with a subsequent
fail in water depth within the transition. The minimum depth
of water occurs approximately at the centre. The turbulence
intensities at the free surface reach a large value of about
u`/Uo≈30% and v`/Uo≈23% at the centre compared about 10%
and 9% of the corresponding values at the free water surface is
primarily due to oblique waves that are generated within the
transition. The vicinity of the upstream corner of the transition
a likely location separation at x/b=-1, z/b=0.25, turbulence
levels are substantially higher compared to the one at centre
z/b=0 at the location. The trend u`/Uo and v`/Uo are similar in
the nature. At the entrance of the transition, the production of
turbulence is higher compared to previous upstream sections.
This increase in both turbulence intensities u`/Uo and v`/Uo
occurs throughout the depth of flow. The transverse
component of deflected velocity vector at the corner, interact
with the longitudinal velocity vector at the centerline at the
entrance, may be instrumental in higher turbulence production
at x/b=-1, x/b=0.
Within the transition, the Reynolds and Froude numbers
increased to 6.8 104
and 0.7 on average respectively. As the
flow approaches critical state as indicated by Fr reaching close
to unity, it tends to be unstable due to higher wave
disturbance, although the flow in transition is not exactly
critical at Fr=0.6. In addition, oblique surface waves are seen
during experimentation within the transition and few
centimeters, about 30-35 cms downstream of the exit of the
transition. The combined influence of these surface waves in
the subcritical flow and the oblique waves due to constriction
could have enhanced, the turbulence intensities in the free
surface and wall region at the centre x/b=0, and beyond up to
exit section at x/b=1. This surface wave disturbance effect
appears predominant from the centre and at all subsequent
section downstream. At the center itself at x/b=0, however, in
the core region turbulence is minimum. This may be attributed
to the high linear velocity generated due to constriction and
reduced depth. Along the centerline defined by x/b=0, z/b=0;

__________________________________________________________________________________________
x/b=1,z/b=0; and x/b=2, z/b=0, the trend variation of both
u`/Uo and v`/Uo is quite similar as stated earlier, viz.,
turbulence decreases from large magnitude in wall region
reaching minimum in the core region and again rises to a large
value in the free surface region. Within the core region
turbulence decreases slightly as one approach the centre. At
the centre itself the turbulence in the core region attains a
minimum value. Continuing from the centre downstream up to
the exit of transition, turbulence profile show a distinct
change. Compared to the turbulence at the centre, the
turbulence intensities increase in the magnitude in the core
region. The turbulence variation, both in the wall region and
the free surface region at the centre and up to exit section of
the transition, at x/b=0 and 1, are specifically noteworthy.
Turbulence levels are enhanced significantly in the free
surface and the wall regions compared to the core region. The
streamwise component u`/Uo is much larger at the surface in
comparison with v`/Uo at the centre and exit.
Downstream of the transition in the expansions zones,
Froude’s number increases slightly after inlet of expansion
zone and then decreases slightly, Reynolds number remains
almost same at 4 104
. The conditions of the flow at the inlet
of the expansions zones cause unidirectional distortion of the
fluid elements which may be expected to produce high
nonhomogeneous and anisotropic turbulence downstream of
the transition. Under the action of dynamic process, the
turbulence is produced to some degree all over the field. A
high level of turbulence intensities u`/Uo and v`/Uo in all
expansion zone occurs compared to the upstream sections as
seen in the Figure 10. Specially the maximum turbulence
intensities u`/Uo and v`/Uo can be seen to occur either close to
the free surface of close to the bed. Even in his region,
turbulence intensities u`/Uo and v`/Uo in core region are
relatively low. It can be 2 to 4 times as high compared to
minimum value in the wall or free surface region. The nature
of variation in turbulence intensities u`/Uo and v`/Uo at the
entry of expansion zone and subsequent sections downstream
is somewhat distinct compared to the turbulence profiles
before and within the transition. Herein, in the core region
turbulence intensities profiles u`/Uo and v`/Uo do not exhibit
the tendency towards constancy unlike in the transition or
before the transition. Generally, after reaching the minimum
turbulence intensities u`/Uo and v`/Uo as the distance increases
from the wall, the turbulence tends to increase consistently till
the free surface is reached. Turbulence intensities are
particularly largest at u`/Uo=52%,v`/Uo=35% and
u`/Uo=63%,v`/Uo=40% at x/b=2, z/b=0 and x/b=2, z/b=0.25
closer to the wall region and free surface region respectively.
Similarly as stated earlier both the turbulence intensities u`/Uo
and v`/Uo are large at all the sections investigated downstream
of the transition in the wall region and the free surface region.
The general trend in variation of depthwise turbulence is
similar in the downstream zone up to x/b=5 observed in this
work. The turbulence intensities u`/Uo and v`/Uo are, however,
observed to be of the same order at the wall and free surface
along the centerline located at entrance to the expansion zone
x/b=1, farthest section at x/b=5, z/b=0 and x/b=5 and
z/b=0.25. However, at an intermediate location x/b=2, these
values do differ significantly. The changes in turbulence
intensities u`/Uo and v`/Uo both in the wall region and the free
surface region, and noteworthy at and after the centre of
transition x/b=0 are specifically significant. The increase in
the turbulence in the downstream may be attributed to the
probable separation occurrence at the exit corner and the
occurrence of diverging flow, associated with the oblique
surface waves.
Summing up it may be concluded that the depthwise
turbulence characteristics upstream, within and downstream of
the transition differ as revealed in the profiles shown. The
vertical side walls may be affect the depthwise turbulence
away from the transition upstream of it. However, at the
entrance of the transition, at the corner, turbulence has a large
magnitude particularly near the bed, but not at the centerline.
Just downstream of the entrance to the transition, turbulence is
significant throughout the depth for the probable role of the
lateral component of velocity as explained earlier.
Free surface has a unique role in governing the turbulence in
the open channel flows. Hence the effect of the free surface is
emphasized in this section. As already presented, Figs.8,9 and
10 depict the variation of turbulence intensities u`/Uo and
v`/Uo in the channel transition at different contraction ratios, Δ
b/b of 0.3 and 0.5, at different discharges of 15 and 40 Lit/sec
at upstream, within and downstream locations. The free
surface influences to a larger extent outer region. The
influence of increased turbulence in the outer layer may also
be transferred in the lower core region. At upstream of the
transition, the most noticeable effect is the dampening
streamwise and vertical components of turbulence fluctuations
u`/Uo and v`/Uo near the free surface. Associated with this
effect is a reduction in eddy near the free surface, as compared
to the wall region(wall effect)at y/yo≤0.2. Clearly, upstream of
the transition, the free surface suppress the movement of
eddies and the surface waves. Also, the most noticeable
feature is that, Froude numbers of less than of equal 0.22, the
surface waves do not occur at the free surface. The effect of
surface waves on the turbulence becomes significant when the
Froude number exceeds about 0.22 as can be seen in the
figures. This influence drawn from the Laser technique
measurements is the same as stated by Nezu and Rodi (1986).
If both the upper and lower boundaries were fixed, one would
expect symmetric distribution of u`/Uo and v`/Uo. The non-
symmetric turbulence distribution along the vertical with
respect to the centre of the depth of channel, with larger
turbulence intensities u`/Uo and v`/Uo at the channel bed and
top of the surface, are not necessarily equal. It is concluded
that, upstream of the transition, turbulence fluctuation u`/Uo
and v`/Uo are slightly decreased at the free surface as

__________________________________________________________________________________________
compared to the wall region. Within the transition, the flow
approaching the critical condition from the tranquil state,
Froude number close to 1.0, the turbulence intensities u`/Uo
and v`/Uo near the free surface are higher as compared to the
core region. The turbulence intensity u`/Uo at the free surface
is always larger than v`/Uo although both increase as the flow
proceeds downstream. It may be explained by noting that, as
the flow approaches near critical condition, the oblique waves
under chocking state were formed within the transition. This
leads to high values of velocity gradients and surface
disturbances leading to the greater turbulence production.
Downstream of all the transition, the probable separation and
wave generation coupled with the oblique surface waves
transmitted to about 30-40 cms, generates nonhomogeneous
and anisotropic turbulence. This can be clearly noted, for
instance, at Δ b/b=0.5 as shown Fig.10, the maximum
turbulence intensities u`/Uo and v`/Uo occur at the free surface;
u`/Uo at the free surface being twice as large as at the wall at
x/b=2, z/b=0.25 and also at x/b=1 and z/b=0.25.
Fig.11 depicts the profiles of streamwise mean velocity
distribution ū/Uo of channel transition along the depth for
discharge 40 Lit/sec at different locations upstream, within
and downstream of the transition at free stream Reynolds and
Froude numbers of 0.4 105
and 0.240 respectively. The main
observations are as follow. The streamwise mean velocity
profile along the longitudinal direction at the centerline of the
transition indicates general increase from upstream zone up to
the entrance of expansion zone of the transition. Subsequently,
the magnitude decreases gradually downstream. At the corner
locations of the transitions, reversed flow and flow separation
could be observed downstream of the transition as show in
Fig.11 at x/b=1 at z/b=0.25 as can be seen by the shape of the
velocity profile and was observed by dye injection. Reversed
flow was also observed at the beginning of the transition. With
the increase in the contraction Δb/b, the flow reversal also
increased. On the other hand the maximum velocity occurs
close to the end of the transition at the centerline z/b=0, x/b=1
and 2, below the free surface in the downstream zone.
Velocity distribution along the depth shows a curvilinear
nature, with greater distance downstream.
Fig.12 depicts the profile of vertical mean velocity /Uo
profiles is marked changes in the magnitude along the depth.
In ū/Uo profile, changes were relatively smooth, with smaller
number of points of contraflexure in wall, core and the free
surface regions. Secondly, /Uo, assumes both positive and
negative values at the same section and hence the zero value
also at some intermediate locations. This nature of variation,
viz., change from positive to negative magnitude and vice-
versa occurred almost at all the sections. However, in the
profile of ū/Uo such a variation occurred near the separation
zone due to reversal of flow. Thirdly, the zero magnitude of
vertical velocity component /Uo occurs at more than one
point at several locations. This observations is somewhat more
intriguing as one may not expect more than one location at
which /Uo could be zero. Fourthly, the magnitude /Uo of
increases steadily from the upstream section from relatively a
small value (albeit predominantly negative in the farthest
section of this investigation) with a gradual increase reaching
maximum positive and negative values at the centre of the
transition, x/b=0 and z/b=0. At exact entrance location,
velocity /Uo is negative all along the vertical height. At
x/b=1, z/b=0.25 just outside the exit section of transition,
velocity /Uo is positive and a fairly of large magnitude.
Subsequently, although /Uo fluctuates as one moves
downstream of the transition, the magnitude gradually
decreases reaching relatively small value again at the farthest
downstream of x/b=5.
In order to study turbulences variation downstream of the
compound transition covering the entire cross section and
longitudinal direction, the turbulence intensities u`/Uo and
v`/Uo were measured at three transverse locations z/b=0.0,015
and 0.30, and several longitudinal x/b locations, in a
horizontal plane y/yo=0.5 for three contraction ratios of
Δb/b=0.5 and 0.3, for Q=45 lit/s. Figs.13 and 14 depict these
variation of turbulence intensities u`/Uo and v`/Uo. The trend
of variation in both streamwise and vertical turbulence
intensities is quite similar, the maxima of both the intensities
also occurring at the same location of x/b. A slight shift in x/b
for such a maxima was sometimes noticed. The turbulence
rises rapidly to reach maxima in the of 3<x/b<4.5. In the
region beyond x/b>4.5, turbulence decreases smoothly and
continuously at slower rate for all the cases. For z/b=0 and
0.15, the trend is reverse at 3<x/b<4.5 in all the cases. It is
noted that the steamwise turbulence u`/Uo is stronger
compared to the vertical turbulence v`/Uo . In the inlet region
of the expansion (transition outlet), the turbulence increases as
the side wall is approached as noted earlier. For z/b of 0.3 the
turbulence levels are higher along the length of the channel.
The nature of variance of turbulence in all the cases clearly
indicate that the flow is nonisotropic and nonhomogeneous at
all locations. As well as later, isotropy of turbulence has been
observed at some locations and depths. Although the nature of
variation in turbulence u`/Uo and v`/Uo are similar in all the
cases, the difference in magnitude of the variation in
turbulence for Δb/b=0.5 and 0.3, and also at the location
z/b=0.3 is significant.
Figs.15 and 16 show the variation of turbulence intensities
u`/Uo and v`/Uo at y/yo of 0.5 downstream of compound
transition in the expansion zones beyond x/b=0 at three
spanwise locations z/b of 0 and 0.25 for Q=45 lit/sec at
different Δb/b of 0.5, 0.4 and 0.3. Both the u`/Uo and v`/Uo
increased rapidly and almost simultaneously. Rate of increase
of turbulence is greater than the rate of decrease after reaching
the maximum, in all the cases. A slight gradual increase in x/b
at which the u`/Uo and v`/Uo are maximum, is observed as the
contraction ratio Δb/b decreases from 0.5 to 0.3, in all the

__________________________________________________________________________________________
cases of experimentation. The maximum turbulence occurs for
the maximum Δb/b ratio, viz 0.5 in the present investigation
and the least of the maximum turbulence occurs for minimum
Δb/b of 0.30. Similarly, the largest maximum turbulence
intensities u`/Uo and v`/Uo occur at spanwise location z/b=0.25
and smallest of the maximum turbulence intensity occur at
spanwise location z/b=0. At z/b=0 i.e. at the centerline
turbulence intensities u`/Uo and v`/Uo assume intermediate
values. Turbulence intensities at the wall itself will be nil.
Thus the turbulence intensities in transverse direction decrease
form the maximum approximately at z/b=0.25 and fall again
to zero at the wall. The trend is the same in all the cases of
experimental observation.
CONCLUSIONS
The conclusions arising out of this study can be summarizes as
follows:
The relative protection length L/b increases with the increase
of the incoming Froude number Fu at particular contraction
ratio Δb/b and fixed bottom slope So. Also, the L/b increase
with the increase of the bottom slope. The flatter the bottom
slope So is, the shorter the relative protection length, at
particular lateral contraction ratio and certain value of
upstream Froude number. The L/b increases with the increase
of lateral contraction ratio Δb/b. The smaller the contraction
ratio, the smaller the relative protection length, at particular
bottom slope. At particular bottom slope and lateral
contraction ratio, the relative protection length L/b as the
vertical transition ratio h/b increases for particular upstream
Froude number.
Free surface has a unique role in governing the turbulence in
the open channel flow. The free surface influences to a larger
extent outer region. The influence of increased turbulence in
the outer layer may also be transferred in the lower core
region. At upstream of the transitions, the most noticeable
effect is the dampening of turbulence intensities near the free
surface, Associate with this effect is a reduction in eddy
viscosity near the free surface, as compared to the wall region
(wall effect), also the free surface suppresses the movement of
eddies and the surface waves in the upstream of the transition.
An increase in the Froude and Reynolds numbers, decreases
the streamwise ad vertical turbulence intensities upstream
slightly, followed by slight increase in the turbulence
intensities downstream of the transition in the expansion zone
at the same flow condition. This effect in the turbulence
occurs primarily in the wall region and the free surface region.
However, in the core region no noticeable change in the
turbulence levels upstream, within and downstream regions of
the transition occurs. Within the transition, flow tends towards
the critical state (Fr approaches close to 1) with a rise in the
turbulence intensities u`/Uo and v`/Uo in the free surface and
the wall region which may be attributed to the flow tending
towards critical state. At small values of Froude number
Fr≤0.22, the effect of Fr on turbulence intensities u`/Uo and
v`/Uo and the free surface wave is negligible. As a
comprehensive observation, it is noted that, the stream wise
turbulence u`/Uo is always greater to the vertical turbulence
v`/Uo. The turbulence intensities u`/Uo and v`/Uo are higher
nearer the bed in the wall region defined by y/yo≤0.2 due to
wall effect and the free surface region defined by y/y≥0.6 due
to free surface effect. In the intermediate core region defined
by 0.2<y/yo<0.6, minimum turbulence intensities u`/Uo and
v`/Uo occur. With the increasing channel contraction, the
variation of turbulence intensities u`/Uo and v`/Uo increase and
become pronounced within and downstream of the transition,
changing rapidly in the wall, core and the free surface regions
NOMENCLATURE
b Channel width,
Δb Channel contraction,
Fu Upstream Froude number,
Re Reynolds number,
L Protection length, spacing between(2)and(3),
Lt Transition length (length of contracted reach)
Uo Streamwise mean free steam, velocity averaged over the
cross section.
u` Streamwise turbulence intensity component in x-direction
(RMS)
Streamwise mean velocity component in x- direction
(RMS) Root mean square
v` Vertical turbulence intensity component in y- direction
(RMS).
Vertical mean velocity component in y- direction.
x Longitudinal axis along channel length
y Transverse axis along channel height
yo Free stream water depth
yu Upstream water depth
z Lateral axis along channel width
h Height of vertical transition
So Bottom slope.
Q Flow discharge
REFERENCES:
[1] Antonia, R.A., and Rajagoplala, S. (1982), “Reynolds
Number Dependence of the Structure of a Turbulent
Boundary Layer”, Prof. J. Fluid Mechanics, 121: Pp.
121-140.
[2] Chen, C., and Jaw, S. (1997), “Fundamentals of
Turbulence Modeling”, Teylor & Francis, London.
[3] Clauser, F. (1956), “The Turbulent Boundary Layer”,
Adv. Appl Mech., 4, 1-51.
[4] Dancey, C.H. (1990), “Measurement of Second order
Turbulence statistics in an Axial Flow Compressor via
3-Component LDA”, AIAA J., 90, 2017.
[5] Ead, S., Rayaratnam, N., Katopodis, C., Ade, F. (2000),
“Turbulent Open Channel Flow in Circular corrugated
Culverts”, J. Hydraul. Eng., 126 (10), 750 – 757.

__________________________________________________________________________________________
[6] Hitoshi, S., Akiyama, M., Mitsunobu, F.O., and Sato,
T. (1995), “Numerical Analysis of Developing
Turbulent Flow in a Square sectioned 90 degree bend”.
Trans. Jpn. Soc. Mech. Eng., Ser. B, 61 (587), 2393-
2400.
[7] Kim, J., Kline, S., Johnston, J.P. (1980), “Investigation
of a Reattaching Turbulent Shear Layer; Flow over a
Backward Faing Step”. J. Fluids Eng., 102, 302.
[8] Kironoto, B.A., and Graf, W. H. (1994), “Turbulence
Characteristics in Rough Uniform Open Channel
Flow”, Proc., Inst. Civ. Eng., Waters, Maritime
Energy., 98.
[9] Kline, S. J., Reynolds, W.C., Schraub, F.A., and
Runstadler, P.W. (1967), “The Structure of Turbulent
Boundary Layer. “J. Fluid Mech., 30, 741-773.
[10] MeLelland, S., Asworth, P., Best, J., and Livesey, J.
(1999), “Turbulence and Secondary Flow over
Sediment Stripes in Weakly Bimodal Bed Matevial”, J.
Secondary Flows, 125 (5), 463-473.
[11] Mehta, P.R. (1981), “Separated Flow Through Large
Sudden Expansions”, J. Hydraaul. Div., Am. Soc. Civ.
Eng., 107 (4), 451-460.
[12] Nezu, J., and Nakagawa, H. (1993), “Turbulence in
Open Channel Flows”, A. A. Balkema, Rotterdam, The
Netherlands.
[13] Nakagawa, H., and Nezu, I.,(1987)”Experimental
Investigation on Turbulent Structure of Back facing
Step Flow in an Open Channel”, J. Hydraulic Research,
IAHR, 25, Pp. 67-88.
[14] Nezu, I., and Rodi, W.,(1993)“Open Channel Flow
Measurements with a Laser Doppler Velocimetry”, J.
Hydraulic Engg., ASCE, 112, Pp. 335-355, 1986.
[15] Nezu, I., and Nakagawa, H., “Turbulence in Open
Channel Flows”, IAHR-Mongraph, A.A. Balkema
Publishers, Old Post Road, Brookfield, VT 05036,USA.
[16] Rodi, W.,(1994)“Turbulence Models and their
Application in Hydraulics”, IAHR Monograph, A.A.
Balkema Publishers, Old Post Road, Bookfiels, VT
05036, USA.

__________________________________________________________________________________________

Effect of free surface and froude number on the

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Effect of free surface and froude number on the (20)

More from eSAT Publishing House (20)

Recently uploaded (20)

Effect of free surface and froude number on the