Available via license: CC BY 3.0
Content may be subject to copyright.
IOP Conference Series: Earth and Environmental Science
PAPER • OPEN ACCESS
Experimental study on flow structure in riffle-pool channels
To cite this article: Like Li et al 2018 IOP Conf. Ser.: Earth Environ. Sci. 199 052055
View the article online for updates and enhancements.
This content was downloaded from IP address 181.215.50.63 on 19/12/2018 at 17:12
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd
EPPCT 2018
IOP Conf. Series: Earth and Environmental Science 199 (2018) 052055 IOP Publishing
doi:10.1088/1755-1315/199/5/052055
1
Experimental study on flow structure in riffle-pool channels
Like Li, Niannian Fan* and Xingnian Liu
State Key Laboratory of Hydraulics and Mountain River Engineering, College of
Water Resource & Hydropower, Sichuan University, China
*Corresponding author: fannian7172@126.com
Abstract: Because of the riverbed evolution and geological tectonics, the river geometry has
changed obviously, forming a wide and narrow alternated channel .The area with shallow
water depth is easy to form riffles. The area with narrow deep water often forms deep pools.
Flow rate will increase or decrease. The hydraulic characteristics of riffle-pool with clear water
scouring under steady state were studied by using three-dimensional overlooking acoustic
Doppler velocimeter (ADV：Nortek-vectrino Ⅱ), a camera, a underwater high-speed camera
and a laser rangefinder . The results showed that with the change of channel geometry, the
water level of cross-section would change accordingly; logarithm law could well describe the
time-averaged velocity distribution near the wall. However, the bed morphology affected the
flow structure near the bed to a certain extent; because of the irregular cross-section and the
influence of the bed topography, the cross-section gradually changed. The Reynolds stress
varied significantly in the area where the cross-section decreased gradually, indicating that the
water body had strong shearing action. In the enlarged area, the Reynolds stress distributed
uniformly.
1. Introduction
Due to the development of the riverbed and the impact of geological structures, the geometry of the
river channel is obviously changed to form a wide and narrow river channel. The water depth in the
wide area is shallow, and it is easy to form a shoal. The water depth in the narrow part is deep, often
forming a deep pool, and the flow rate will increase or decrease with the depth. The slope of the
riffle-pool is generally less than 0.02, which is usually a straight or curved gravel riverbed river.
Usually, the distance between the deep pools is 5-7 times to the width of the river, but if there are a
large number of collapsed trees in the river, its distance will become smaller [1].
Since the riverbed gradient in the trailing edge area of the deep pool is larger than that of the water
surface, the water flow will spread in the vertical direction. Many scholars at home and abroad have
conducted extensive research on this wide and narrow river channel. Nelson et al. (2015) studied the
geomorphic dynamic response of gravel rivers with varying river widths to sediment transport changes
through a sink test. The change does not affect the location and topography of the shoal-deep pool.
The river channel responds to changes in sediment recharge by adjusting the gradient [2]. MacVicar
and Best (2013) generalized the bed surface morphology of smooth deep pools and shoals with smooth
boundaries using perturbation theory, pointing out that the shear velocity and Gaussian wake
parameters are not sensitive to the adjustment of river width, while the concentration of lateral flow is
The recovery of the main Reynolds stress can be described by two-stage propagation and a relaxation
response that varies with the width of the river [3]. Yan et al. (2011) analyzed the local head loss
characteristics of the river wide expansion section affected by water depth shrinkage based on
EPPCT 2018
IOP Conf. Series: Earth and Environmental Science 199 (2018) 052055 IOP Publishing
doi:10.1088/1755-1315/199/5/052055
2
laboratory tests [4]. Zhou et al. (2013) established a two-dimensional flow model of a wide and narrow
river channel in a mountainous area to simulate the characteristics of water flow motion [5]. Caamaño
et al. (2012) established a generalized model to explain the characteristics of the water flow structure
of the riffle-pool in the gravel river and how to respond to the change of external force through
detailed measurement and simulation of the three-dimensional water flow structure [6]. MacVicar and
Roy (2007) studied the distribution of the velocity and turbulence intensity of the riffle-pool forced by
the tree blocking channel, pointing out that the deceleration and acceleration flow generated by the
expansion and contraction in the vertical direction can explain many observations [7]. Papanicolaou
and Elhakeem (2007) combined with field measurements and experiments, pointed out that due to the
change of the section shape of the channel. The average shear stress does not approximate the
magnitude of the fluid shear stress [8]. Hoan et al. (2007) believe that the turbulence intensity and
Reynolds stress distribution deviate from the theoretical and experimental curves of uniform flow due
to the irregular geometry of the river [9]. Venditti et al., (2014) used the Acoustic Doppler Velocity
Profiler (ADCP) to measure the Fraser Canyon in Canada. When the water flowed into the canyon,
high-speed water flowed near the bed and low-speed water flowed on the surface, causing a reversal of
speed. The water then surges along the side walls of the canyon, creating a counter-rotation that causes
the main stream to deviate from the center of the river [10]. Yang et al., (2007) investigate the variation
of different resistance coefficients with the stage use symmetric compound channel with a large bed
roughness and summarize different representative methods for assessing the composite roughnesses in
compound channels [11].
Due to the turbulent flow, the kinetic transfer of the bed, and the obvious feedback relationship
between the bed surface micro-topography [12], Carbonneau and Bergeron (2000) pointed out through
experiments that the bed load will change the dissipation rate of turbulent kinetic energy, and thus
affect the near-bed area. Flow rate distribution [13]. In the past, the experimental study on the wide and
narrow phase water channel was mostly fixed bed surface [3-5], and it was rarely concerned about the
water flow structure under the condition of clear water scouring. This paper aims to study the water
flow under the steady state of wide and narrow rivers through the sink test. The study of flow structure
can improve the understanding of the type of river in the shoal-deep pool.
2. Experimental materials and settings
The test was conducted at the University of British Columbia, Canada, with a test tank length of 18 m
and a trough-wide gradient sink with a width of 0.8 m and a narrowest point of only 0.38 m. The initial
slope of the sink is 0.015 and the flow rate is 50 L/s. The test is divided into six stages: clear water
flushing, stable sanding, clear water flushing, increasing water flow, increasing water flow again and
adding fine particles. At the beginning of the test, the sand was first laid flat at the bottom of the tank.
The particle size distribution of the sand before and after the test is shown in Figure 1. The water is
then flushed and the airfoil is mounted on the connecting rod of the three-dimensional overhead
acoustic Doppler velocimetry and the underwater camera to reduce wake. During the test, there was no
sand in the upper reaches, and the sediment transported by the water flow originated from the bed
surface. After 78 hours of clean water flushing, the sediment transport rate at the exit section has been
very small, dropping to 0.075 g/s, falling below the maximum sediment transport rate of 7.55 g/s (one
hour average sediment transport rate). The bottom of the sink has also been shaped into a certain shape
of the bed. It is washed out of the deep pool at the narrow position of the river, and is stacked into a
shallow beach at the widened position of the river, as shown in Figure 2. At this time, the
cross-sectional flow velocity was measured using ADV, and the surface morphology information was
collected using a camera and a laser topographic scanner. This paper mainly studies the water flow
structure when the bed surface is in a stable state after 78 hours of water washing.
EPPCT 2018
IOP Conf. Series: Earth and Environmental Science 199 (2018) 052055 IOP Publishing
doi:10.1088/1755-1315/199/5/052055
3
110
0
20
40
60
80
100
Percentage(%)
Grain size (mm)
before the experiment
after the experiment
Fig. 1 Grain size distribution
Fig. 2 Bed topography
The test tank has three wide and narrow river sections. In order to avoid the influence of the inlet
and outlet water flow, the middle section of the tank is selected as the study area. A total of 17
measuring sections are set in this area, the 1 # section is set at x= -5.6 m, the 17 # section is set at x=
-10 m, and the specific arrangement of the measuring section is shown in Fig. 3. According to the
change of the geometry of the water tank, measuring sections with unequal spacing and different
numbers of measuring vertical lines are set, and each vertical line has different number of measuring
points due to the difference of water depth. The arrangement of specific measuring points is shown by
the dashed line in Figure 3. The single short line segment in each broken line represents a measuring
section in the longitudinal direction to explore the water flow structure of the wide and narrow phase
channel in detail. The instrument measures the instantaneous velocity of the non-uniform flow on the
lateral boom overlooking the ADV. The instrument measures the flow of water at 4～7.4 cm below the
bottom of the probe. The sampling frequency and time are set to 50 Hz and over 40 seconds
respectively. The number of sampling points in each measurement period is greater than 2000, which
has a high signal-to-noise ratio and correlation. Despikng of ADV data is performed by Derek G.
Goring [14].
Fig. 3 Test measurement section layout (m)
EPPCT 2018
IOP Conf. Series: Earth and Environmental Science 199 (2018) 052055 IOP Publishing
doi:10.1088/1755-1315/199/5/052055
4
3. Results and analysis
3.1Water depth changes along the river
Figure 4 shows the distribution of water depth along the course. As the geometry of the channel
changes, the water level changes relatively small. Due to the narrowness of the channel, the deep
water was washed out, and the water depth increased significantly. After that, the river was widened
and piled up into shoals, and the water depth began to gradually decrease. It can be seen from the
figure that the minimum water depth does not correspond to the widest part of the section, but occurs
at the position of x= -6.3 m (4 #section); the maximum depth of water does not correspond to the
narrowest part of the section, but occurs at x = -7.9 m (9 #section) position, 9 # section is also the
lowest point of the bed, combined with Figure 2 can be found that this position is the most severe.
-10-9-8-7-6-5
0.10
0.15
0.20
0.25
0.30
0.35
Narrowest section
z(m)
x(m)
water level
bed elevation
water depth
Widest section
flow
Fig. 4 Water depth changes along the river
3.2 Vertical distribution of time averaged velocity
During the flow of water in the tank, the flow velocity distribution is not uniformly affected by the
geometry of the channel and the morphology of the bed. Generally, the flow velocity in the lateral
direction increases from the boundary of the sink to the intermediate position, and increases in depth
along with the increase in depth. We used the ADV flow meter to measure the water flow rate from the
diverging to the tapered section at different depths from the riverbed.
According to the previous study of the time-averaged velocity distribution of uniform flow and
non-uniform flow [15-20], in order to study the turbulence characteristics, the velocity distribution is
fitted to the turbulent region following the logarithmic distribution (h/z≤0.2 ). The outlet section of
the experimental tank can be approximated as a uniform flow. Since the frictional velocity u* of the
non-uniform flow is very difficult to determine, in order to simplify the calculation, the frictional flow
velocity u* of the outlet section is assumed to be 0.046 m/ s. The longitudinal velocity of all sections is
normalized by using the frictional velocity u* as a general parameter. The instantaneous velocity of
each point measured by the ADV is arithmetically averaged to obtain the average flow velocity value
of the point on the vertical line. The calculation formula is:
*
u
uu
(1)
*
yu
yv
(2)
In the formula, u: the average flow velocity along the vertical direction,
: the kinematic viscosity
of the water, y: the distance from the measurement point to the bed surface. See Figure 5 below for the
change in flow velocity from the gradually expanding section to the tapered section.
EPPCT 2018
IOP Conf. Series: Earth and Environmental Science 199 (2018) 052055 IOP Publishing
doi:10.1088/1755-1315/199/5/052055
5
100 1000
0
5
10
15
20
25
y=0 m
y=0.05 m
y=0.10 m
y=0.15m
u
+
y
+
(a) x=-7.4 m
100 1000
0
5
10
15
20
25
y=0 m
y=0.05 m
y=0.10 m
y=0.15m
u
+
y
+
(b) x=-7.7 m
100 1000
0
5
10
15
20
y=0 m
y=0.05 m
y=0.10 m
y=0.15m
u
+
y
+
(c) x=-8.3 m
100 1000
0
5
10
15
20
y=0 m
y=0.05 m
y=0.10 m
y=0.15m
u
+
y
+
(d) x=-8.5 m
Fig. 5 Time-averaged velocity distribution
It is obvious from the figure that all the velocity profiles are basically in line with the logarithmic
distribution, which is consistent with the previous research results of non-uniform flow [21].
3.3 Reynolds stress
The Reynolds stress reflects the momentum transfer caused by the pulsation of the water flow.
According to the pulsation speeds u, v and w of the water flow in the x, y and z directions, the
Reynolds stress of the XOY and XOZ planes is calculated by the pulsation correlation method, and the
Reynolds stress is obtained. The stress is normalized by u * 2, and the distribution of the dimensionless
Reynolds stress along the vertical direction can be obtained. The component of the dimensionless
Reynolds stress has anisotropy in the vertical and lateral directions, and the distribution from the
divergent section to the tapered section along the groove width is shown in Figure 6.
EPPCT 2018
IOP Conf. Series: Earth and Environmental Science 199 (2018) 052055 IOP Publishing
doi:10.1088/1755-1315/199/5/052055
6
-5 -4 -3 -2 -1 0 1 2 3 4 5
0.02
0.04
0.06
0.08
0.10
0.12
0.14
-5 -4 -3 -2 -1 0 1 2 3 4 5
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
-5 -4 -3 -2 -1 0 1 2 3 4 5
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
-5 -4 -3 -2 -1 0 1 2 3 4 5
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
-5 -4 -3 -2 -1 0 1 2 3 4 5
0.06
0.08
0.10
0.12
0.14
0.16
0.18
-5 -4 -3 -2 -1 0 1 2 3 4 5
0.04
0.06
0.08
0.10
0.12
0.14
0.16
-20246810
0.08
0.10
0.12
0.14
0.16
0.18
0.20
-2024681012
0.06
0.08
0.10
0.12
0.14
0.16
0.18
-5 -4 -3 -2 -1 0 1 2 3 4 5
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
0 1020304050
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
-2 0 2 4 6 8 10 12
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
-5 -4 -3 -2 -1 0 1 2 3 4 5
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
z/h
(a
1
) y=0.05 m (a
2
) y=0.10 m
xy
/
xz
/
z/h
(a) x=-7.7 m
(a
3
) y=0.15 m
z/h
z/h
z/h
z/h
z/h
z/h
z/h
z/h
(d) x=-8.5 m
(d
3
) y=0.15 m
(d
2
) y=0.10 m
(d
1
) y=0.05 m
(c
3
) y=0.15 m
(c
2
) y=0.10 m
(c
1
) y=0.05 m
(c) x=-8.3 m
(b) x=-7.9 m
(b
3
) y=0.15 m
(b
2
) y=0.10 m
(b
1
) y=0.05 m
z/h
z/h
Fig. 6 Typical section Reynolds shear stress distribution
Overall, the Reynolds stress has a higher value near the bed surface, indicating that turbulence is
due to wall shear. The difference of shear stress in vertical and horizontal directions is obvious. With
the increase of water depth, the influence of bed surface is gradually weakened. The difference
between
xy and
xz gradually becomes smaller. In the tapered section, about 0.12 h water depth,
Renault The stresses
xy and
xz are hardly affected by the morphology of the bed surface and tend to
EPPCT 2018
IOP Conf. Series: Earth and Environmental Science 199 (2018) 052055 IOP Publishing
doi:10.1088/1755-1315/199/5/052055
7
be parallel along the vertical direction. In the divergent section, the influence of the bed surface shape
extends only to the depth of 0.08 h.
The value of
xy is relatively stable, and the fluctuation range is relatively small. The section x=
-7.7 m and x= -8.5 m are relatively far from the narrowest position of the water tank, and are relatively
less affected by the geometry of the water tank. Nearby, and the section x= -7.9 m and x= -8.3 m, the
channel geometry changes significantly,
xy deviates from the value of 0, and fluctuates roughly
between 1 and -1.
The fluctuation range of
xz is relatively large.
xz is affected by the riverbed topography on the near
riverbed surface, and the Reynolds stress fluctuation is abnormally significant, especially for the area
where the section is gradually reduced. It can even reach several times the area where the section is
gradually enlarged (section x= -7.7 m and x= -7.9 m), indicating that the shearing effect of water in
this area is very strong.
4. Conclusion
This paper focuses on the water flow structure under the steady state of water flushing between wide
and narrow rivers, and draws the following conclusions:
1. The water level is less affected by the geometry of the river channel, and there is no obvious
change. The minimum water depth of the tank corresponds to the lowest point of the bed surface,
indicating that the location is the most severe.
2. The time-averaged velocity distribution of the vertical direction of different water depths of
each section is studied. It is found that the logarithm law is not only applicable to the near-wall area of
the open channel uniform flow, but also can better describe the turbulent flow area of the open channel
under the condition of wide and narrow river channel clear water scouring. The distribution of average
speed. The average flow velocity near the bed surface does not conform to the logarithmic distribution,
and the shape of the bed surface affects the water flow structure near the riverbed surface to some
extent.
3. The Reynolds stress fluctuation of the near wall surface is large, and the water body has
strong shearing effect. With the increase of water depth, the Reynolds stress is less affected by the
shape of the bed surface, and the difference between
xz and
xy is smaller. Due to the influence of the
channel geometry, the
xy and
xz are almost unaffected by the shape of the bed surface in the tapered
section, which is almost independent of the water depth of the bed. The distribution along the vertical
line tends to two parallel lines, while in the divergent section; the influence only extends to the depth
of 0.08 h.
References
[1] Montgomery, D. R., & Buffington, J. M. (1997). Channel-reach morphology in mountain
drainage basins. Geological Society of America Bulletin, 109(5), 596-611.
[2] Nelson, P. A., Brew, A. K., Morgan, J. A. (2015). Morphodynamic response of a variable‐width
channel to changes in sediment supply. Water Resources Research, 51(7), 5717-5734.
[3] MacVicar, B., & Best, J. (2013). A flume experiment on the effect of channel width on the
perturbation and recovery of flow in straight pools and riffles with smooth
boundaries. Journal of Geophysical Research: Earth Surface, 118(3), 1850-1863.
[4] Yan, X. F., Yi. Z. J., Liu, T. H. et al.. (2011). Flow structure and characteristics of local head loss
in transition channel. Journal of Yangtze River Scientific Research Institute.
[5] Zhou, S. F, Yi, Z. J, Yan, X. F. et al. (2013). Two-dimensional numerical simulation of flows in
wide and narrow alternated channels in mountainous areas. Advances in Science and
Technology of Water Resources, 33(01):22-26.
[6] Caamaño, D., Goodwin, P., Buffington, J. M. (2012). Flow structure through pool‐riffle
sequences and a conceptual model for their sustainability in gravel‐bed rivers. River research
and applications, 28(3), 377-389.
[7] MacVicar, B. J., & Roy, A. G. (2007). Hydrodynamics of a forced riffle pool in a gravel bed river:
EPPCT 2018
IOP Conf. Series: Earth and Environmental Science 199 (2018) 052055 IOP Publishing
doi:10.1088/1755-1315/199/5/052055
8
1. Mean velocity and turbulence intensity. Water Resources Research, 43(12).
[8] Papanicolaou, A. N., & Elhakeem, M. (2007). Turbulence Characteristics in a Gradual Channel
Transition. In World Environmental and Water Resources Congress 2007: Restoring Our
Natural Habitat (pp. 1-6).
[9] Hoan, N. T., Booij, R., Stive, M. J. et al. (2007). Decelerating open-channel flow in a gradual
expansion. In Asian and Pacific Coasts Conference, September 21-24, 2007, Nanjing, China.
APAC.
[10] Venditti, J. G., Rennie, C. D., Bomhof, J. et al. (2014). Flow in bedrock
canyons. Nature, 513(7519), 534.
[11] Yang, K. J., Cao, S. Y., Liu, X. N. (2007). Flow resistance and its prediction methods in
compound channels. Acta Mechanica Sinica, 23(1), 23-31.
[12] Naden, P. S. (1988). Models of sediment transport in natural streams. Modelling
Geomorphological Systems. John Wiley and Sons New York. 1988. p 217-258, 13 fig, 6 tab,
86 ref.
[13] Carbonneau, P. E., & Bergeron, N. E. (2000). The effect of bedload transport on mean and
turbulent flow properties. Geomorphology, 35(3), 267-278.
[14] Goring, D. G., & Nikora, V. I. (2002). Despiking acoustic doppler velocimeter data. Journal of
Hydraulic Engineering, 128(128), 117-126.
[15] Wang, S. Y., Zhou, S. F., Zhao, X. E. (2013). Experimental study on the flow characteristics at
local diverging-converging sections in mountain lotus root shape channel. Journal of Sichuan
University, 45, 51-54.
[16] Liu, C. J, Li, D. X, Wang X. K. (2005). Experimental study on friction velocity and velocity
profile of open channel flow. Journal of Hydraulic Engineering, 36(8), 0950-0955.
[17] Afzalimehr, H., & Anctil, F. (1999). Velocity distribution and shear velocity behaviour of
decelerating flo. Canadian Journal of Civil Engineering,26(26), 468-475.
[18] Balachandar, R., Hagel, K., & Blakely, D. (2002). Velocity distribution in decelerating flow over
rough surfaces. Canadian Journal of Civil Engineering, 29(2), 211-221.
[19] Wang, X. K, Yi Z. J., Yan X. F. et al. (2015). Experimental study of the flow structure of
decelerating and accelerating flows under a gradually varying flume. Journal of
Hydrodynamics. B27(3), 340-349.
[20] Liu, C., Wright, N., Liu, X., Yang, K. (2014). An analytical model for lateral depth-averaged
velocity distributions along a meander in curved compound channels. Advances in Water
Resources, 74, 26-43.
[21] Wang, J. J., & Dong, Z. N. (1996). Open-channel turbulent flow over non-uniform gravel
beds. Applied Scientific Research, 56(4), 243-254.
