A. Tominaga1, K. Ijima2 and Y. Nakano2
1Dept. of Civil Engineering, Nagoya Institute of Technology,
Gokiso-cho, Showa-ku, Nagoya, 466-8555, Japan
Tel & Fax: 81-52-735-5490, E-mail: tomi@suiko1.ace.nitech.ac.jp
2Graduate Student, Dept. of Civil Engineering, Nagoya Institute of Technology
Abstract:
To predict and control the sediment transport and
the water-quality by using submerged spur dikes, it is necessary to clarify
three-dimensional flow structures and instantaneous vortex structures around
spur-dike zones. In this study, relative height of the super dike to the water
depth is chosen as a design item and its effects on flow structures in spur-dike
zones are investigated by using PIV method. Two-dimensional PIV analyses of
vertical and horizontal planes reveal the time-averaged flow structures in
spur-dike zones with reasonable accuracy. The vertical vortices caused by the
top flow and transverse vortices caused by the side flow interact with each other and produce characteristic
three-dimensional vortex structures. The vertical profiles of the longitudinal
velocity indicate similarity with respect to the crest height of the spur dike.
The inverse-flow strength becomes larger at an inner part of the spur-dike zone
with an increase of the relative spur-dike height.
Keywords: submerged spur dike,3D flow structures, local flow, PIV
The spur dikes have been recognized as riparian structures that have functions of retarding velocity near the riverbank and deflecting the main flow from the levee, and as environmental structures that can provide plentiful ecological surroundings by creating diverse flow fields. In the former case, the less influence is expected in riverbed changes, but in the latter case, they might to be used to promote the deposit behind the spur dike and provide the scour holes. There were several researches dealing with the roughness and deflection effects of submerged spur dikes and the bed evolution (e.g. Gill (1972), Klingeman et al.(1984), Fukuoka et al.(1998)). In any case, it is an important problem to clarify flow structures around spur dikes by experiments (e.g. Rajaratnam and Nwachukwu (1983), Chen and Ikeda (1996)) and numerical simulations (Wenka et al.(1993), Liu et al. (1994), Mayerle et al. (1995)). The series of low-height spur dikes submerged by water in flood are frequently used in middle and lower reaches of alluvial rivers. If the area of river channel enveloped between two spur dikes is called as "spur-dike zones", the flow structures in these spur-dike zones still have not been entirely elucidated. In the case of submerged spur dike, it has considered that the vertical vortices caused by the side flow and transverse vortices caused by the top flow interact with each other and produce characteristic three-dimensional vortex structures. It is necessary to clarify three-dimensional flow structures and to grasp turbulent structures and organized vortices, in considering the sediment transport and the mass exchanges in spur-dike zones. In this study, relative height of spur dike to water depth is chosen as a design item and its effects on flow structures in spur-dike zones were investigated by using PIV method. Combination of two-dimensional PIV analyses of vertical and horizontal planes can catch the time-averaged three-dimensional flow structures in spur dike zones (Tominaga et al.(2000)). Since vortex structures in spur-dike zones is relatively stable in the present experimental conditions, only time-averaged flow structures and turbulent statistical quantities were examined.
The experiments were conducted in an 8m long and 0.3m wide rectangular flume. The slope of the flume i was set as 1/2000. The spur dike zone was set 4.5~4.8m downstream from channel entrance. Two continuous spur dikes were set along the left wall, as shown in Fig.1. The spur dike model is made of acrylic acid resin, which is 5.0cm long, 2cm deep, and its height d was changed as 2, 3, 4, 5, 6cm (CASE SD2, SD3, SD4, SD5, SD6). The height d was equal to the flow depth in the non-submerged case (CASE SD8). The interval length of spur dike S was 10cm. In al all cases, the discharge Q was 4100cm3/s and the water depth h was set as 8cm without the spur dikes by adjusting the downstream weir.
Fig.1 Arrangement of spur-dike model
Fig.2 Projection plane of laser light sheet
Fig.3 Velocity vectors on Y2 section
Fig.4 Vertical distributions of primary velocity
Fig.5 Velocity vectors on Y1 and Y3 sections
Nylon resin particle with 50micron diameter and
1.02 specific weight was used for visualizing the flow. The argon laser light
sheet with about 3mm thickness was projected on vertical (x–z) plane and horizontal (x–y)
plane. Fig.2 shows the positions of the laser light projection for vertical
planes; 5, 25, 45, 55, 70mm from the left wall and for horizontal planes; 5mm
above the riverbed, the half height of the dike, 5mm below the dike crest, 5mm
and 15mm above the dike crest. For the non-submerged spur-dike case, the
horizontal planes were 5mm above the riverbed, the half flow depth, and 5mm
below the free surface. The projection sections of each case were named Y1,Y2,…,Y5
from the left wall and Z1,Z2,…,Z5 from the riverbed. The visual images were
taken by using high-speed video camera (Photoron Fastcam-Rabbit-3). Once the
images were recorded in its memory, then they were saved in the hard disk of the
computer as “TIFF” files with 640
480 pixels. The velocity vectors were measured by using ‘‘VISIFLOW PIV’’
system software (AEA Technology), and analyzed by cross correlation method. The
analyzing area was about 20cm span from 4cm upstream of the first spur dike to
the rear edge of the second one. The resolution of the image analysis was 32
32 pixels and overlap was 50%. Time-averaged velocity vectors were obtained by
processing 1963 successive images in 16s with an interval of 1/120s.
Fig.3 shows the velocity vectors in the cases of SD2, SD4, SD6 and SD8 at y=25mm (Y2 section). All the cases of submerged spur dike contain the large-scale transverse vortices with their centers located at about 10~15% below the crest of each spur dike. The width and the strength of inverse flow becomes larger at the lower part of the spur dike zone with an increase of the spur-dike height. The longitudinal center of the transverse vortices moves downstream with an increase of the spur-dike height. Fig.4 shows the vertical distribution of the primary velocity on the vertical section Y2 at x=6cm; the middle point of the spur-dike interval. The vertical velocity distributions show almost similar profiles irrespective of the relative spur-dike height, except for the non-submerged case (SD8). The zero-cross points move upward with an increase of the spur-dike height and the profiles are shifted up in parallel with each other. Therefore, it is recognized that the inverse velocity becomes larger and its vertical distribution range becomes wider with an increase of the spur-dike height. Thus, the maximum inverse velocity is proportional to the spur-dike height. The velocity in the upper part of the spur-dike zones does not vary so much with respect to the spur-dike height. In the case of non-submerged spur dike (SD8), the velocity in the spur-dike zone is obviously smaller than that in the submerged cases. In front of the first spur dike, the impinging flow is separated into upflow over the crest and downflow to the bed and the latter forms the small transverse vortex. These front transverse vortices become larger in scale with an increase of the relative height. In the non-submerged case, SD8, the transverse vortex is not recognized in the spur-dike zone. Only the inverse flow is observed and this inverse velocity becomes weak behind the first spur dike and in front of the second one. This means that there is an inflow through the front of the second dike and an outflow through the back of the first dike. Therefore, it is considered that the vertical vortex entirely controls the flow in the non-submerged spur-dike zone.
Fig.5 shows the velocity vectors on vertical sections of y=5mm (Y1); near the left wall, and of y=45mm (Y3); near the edge of the spur dike, in the representative case, SD4. On the section of y=5mm, the inverse flow near the riverbed are also recognized, but the downflow in front of the second spur dike becomes small. This suggests the existence of the transverse flow toward the left wall along the front of the second spur dike. Accompanying the significant increase of the upward velocity in front of the first spur dike, the upflow behind the first spur dike becomes remarkable. On the section of y=45mm, the transverse vortex structure becomes ambiguous. The downflow in front of the second spur dike is only outstanding.
Fig.6 shows the velocity vectors on the horizontal section
at the half height of the spur dike (Z2 section). On these horizontal sections,
the large-scale vertical vortices are recognized over the whole spur-dike zones
in all cases. In the case of d=2cm
(SD2), however, this vertical vortex becomes weak in strength. In the cases
that the dike height d is greater
than 3, the skewed inverse velocity appears. The center of the vertical vortex
is located on about (x, y)=(5cm, 4cm), irrespective of the relative height of
the spur dike. At the edge of the first spur dike, the flow is deflected toward
the main stream and a small-scale separation bubble is recognized. It was
verified that the main-flow deflection effect becomes larger with an increase
of the relative spur-dike height. Fig.7 shows transverse distribution of the
primary velocity on the horizontal section Z2 at x=6cm; the middle point of the
spur-dike interval. The velocity on the horizontal sections does not show clear
changes like the velocity on the vertical sections, and each case show similar
distribution. The velocity becomes larger in the main flow area (Y
7.0cm) with an increase of the spur-dike height. This is attributed to an
increase of the area sheltered by the spur dike.
Fig.6 Velocity vectors on Z2 section
Fig.7 Transverse distributions of primary velocity
Fig.8
Velocity vectors on Z1 and Z3 sections
Fig.8 shows the velocity vectors on the horizontal section of z=5mm (Z1); near the riverbed, and of z=35mm (Z3); just below the crest of the spur dike. On the section of z=5mm, the vertical vortex structure seems almost similar to that on Z2 section shown in Fig.6, but the the strength of the vertical vortex becomes larger. On the section of z=35mm, the vertical vortex structure loses its shape and the velocity near the left wall becomes very small. The outflow along the back of the first spur dike stands out. These structures suggest the existence of the boiling flow from the riverbed near the left wall.
Based upon the 16sec time-averaged vertical and transverse velocity, mass exchange was examined on the boundary of the crest and the side of the spur-dike zone. Fig.9 and Fig.10 show the longitudinal distribution of the vertical velocity W along the boundary of the crest on Y2 section and of the transverse velocity V along the side boundary of the spur-dike zone on Z2, respectively. On both the crest and side boundary of the spur-dike zones, the inflow occurs in front of the second spur dike and outflow occurs behind the first spur dike. The outflow and inflow velocity tend to become larger in proportion to the relative spur-dike height. However, when the relative spur-dike height becomes more than half of the water depth, there occurs an extremely increase of the inflow discharge from the front of the second spur dike and the outflow discharge from the back of the first spur dike. These phenomena are considered to affect the riverbed evolution and the water quality exchange.
Fig.11 shows the
Reynolds stress
at y=25mm section (Y2). The negative maximum
value appears near the crest of the first spur dike. The negative value of
means that the upward fluctuation
of v causes the acceleration of u and the downward one brings about the
deceleration of u. Whereas in the
spur-dike zones, there are long slender area with positive large value 5mm above
the crest level. This indicates the turbulent structure controlled by the
vertical velocity gradient and the magnitude in SD4 is the largest among all
cases. Fig.12 show the velocity vectors and Reynolds stress
at the half height section of the
spur dike (Z2). There are areas with negative large value at the edge of the
first spur dike. These negative areas expand with an increase of the relative
spur-dike height. This is explained by the same mechanism as the vertical
section. The long slender positive areas are observed from the edge behind the
first spur dike to downstream. These areas are moved to the main flow region
with enlarging their values as increasing the relative spur-dike height.
Effects of relative spur-dike height on flow structure in submerged spur-dike zone are experimentally investigated by using PIV method. First, by using the present PIV system, the high-speed drifted flow at the edge of first spur dike and vortex structures with inverse flow in spur-dike zone could be caught qualitatively and quantitatively. With an increase of the relative spur-dike height, the center position of transverse vortex is shifted up and the inverse flow area below the center is enlarged and the inverse velocity itself increases linearly. In submerged spur dikes, the vertical vortices caused by the side flow and the transverse vortices caused by the top flow are composed. The regions with the same flow direction are strengthened and the regions with the opposite flow direction are weakened. As a result, it is considered that the vortex has an axis normal to the diagonal plane connecting the crest of the first spur dike to the bottom of the second one. It is notable feature of submerged spur dikes that the vortex structures are strengthened and the mass exchange is more activated as compared with the non-submerged spur dikes in a spur-dike zone.
Fig.9 Longitudinal distributions of verticalvelocity W along the crest boundary
Fig.10 Longitudinal distributions of transverse velocity V along the side boundary
Fig.11 Contours of
Reynolds stress
in Y2 section
Fig.12 Contours
of Reynolds stress
in Z2 section
Chen, F. and Ikeda, S. (1996): Experimental study on horizontal separetion eddies in open-channel flow with groins, Annual Journal of Hydraulic Engineering, JSCE, vol.40, pp.787-792 (in Japanese).
Fukuoka, S., Nishimura, T., Okanobu, M. and Kawaguchi, H. (1998): Flow and bed topography around groins installed in a straight channel, Annual Journal of Hydraulic Engineering, JSCE, vol.42, pp.997-1002 (in Japanese).
Gill, M. A. (1972): Erosion of sand beds around spur dikes, J. Hydraulics Division, ASCE, vol.98(9), pp.1587-1602.
Klingeman, P.C., Kehe, S.M. and Owusu, Y.A. (1984): Streambank erosion and channel scour manipulation using rockfill dikes and gabions, WRRI Report for Project No.373909, Oregon State Univ., pp. 1-169.
Liu, J., Tominaga, A. and Nagao, M. (1994): Numerical simulation of the flow around the spur dikes with certain configuration and angles with bank, J. Hydroscience and Hydraulic Engineering, vol.12(2), pp.85-100.
Mayerle, Y., Toro, F.M. and Wang, S.S.Y. (1995): Verification of a three-dimensional numerical model simulation of the flow in the vicinity of spur dikes, Journal of Hydraulic Research, vol.33(2), pp.2150-2156.
Rajaratnum, N. and Nwachukwu, B.A. (1983): Flow near groin-like structures, J. Hydraulic Engineering, ASCE, vol.109(3), pp.463-480.
Wenka, T., Rodi, W. and Nestmann, F. (1993): Depth-averaged calculation of flow in river reaches with flood control and sediment regulation structures, Proc. of 25th IAHR Congress, Tokyo, A-2-1, pp.41-48.
Tominaga, A., Nakano, Y. and Fujita, T. (2000): Visualization analysis of flow structures in submerged spur dikes, Annual Journal of Hydraulic Engineering, JSCE, vol.44, pp.1041-1046 (in Japanese).