FLUME TEST OF SUCTION-BLOWING METHOD FOR  SPIKE DAM CIRCULATION REPRESSION

Shao Xuejun, Chen Zhicong, Hu Huiwu and Ma Yongjun

Department of Hydraulic Engineering, Tsinghua University

Department of Hydraulic Engineering, Tsinghua University,

Beijing 100084, CHINA

Phone: (86) 10 62788543  Fax: (86) 10 62772463,

E-mail: shaoxj@mail.tsinghua.edu.cn

Abstract: This paper presents an experimental investigation into the use of boundary layer control method to reduce the size of a horizontal circulation behind a boundary protrusion, simulated by a spike dam. Experimental results show a more than 90% reduction of the circulation length when the “suction” and “blowing” method is used in flow about the spike dam. The circulation size is quite sensitive even for small increases in Q_S , when Q_S <0.01~0.02Q_F. After suction discharge Q_S exceeds a certain level, e.g., Q_S =0.02 Q_F , circulation size remains stable and does not show sensitivity to further increases in Q_S. Large Q_S can then produce a new circulation at the boundary opposite to the spike dam. From experimental results it can be seen that an optimal suction discharge exists which is large enough to reduce the circulation length to the required range (25~50cm in this study), and small enough compared with the flume discharge. For each of the four runs an optimal discharge is found out.  Based on theoretical results and dimensional analysis, the following non-dimensional parameter, L, is found to adequately represent the related hydraulic factors

L=

Keywords: boundary layer control, circulation, suction and blowing

1    INTRODUCTION

River morphology and flow pattern depends mainly on the features of river valley in a mountainous region. Deflection of flow by protruding mountain ridges into the water body often produce large scale horizontal circulations behind the obstruction, threatening navigation safety and leading to heavy local siltation in a sediment-laden river like the Yangtze. Diversion dike is a common engineering measure to regulate flow pattern and reduce the size of a circulation, but in a large reservoir like the Three-Gorges, the construction of a diversion dike can take a substantial amount of engineering work and increase cost.

Based on the understanding that circulation is the consequence of boundary layer separation, Professor Lin Bingnan pioneered the idea that circulation caused by valley features may be reduced by adopting boundary layer control measures, which was first described by L. Prandtl in 1904 and had since become an important method in the field of aeronautical engineering, and several reviews were made (e.g., Lachmann 1961, Chang 1976). 

Schlichting listed 6 methods of boundary layer control (Schlichting 1979, pp. 378-402). In the case of water flow in fluvial rivers, three of these methods may be used for boundary control purposes, i.e., blowing (acceleration of the boundary layer), suction (removal of decelerated fluid), and provision of suitable boundary shapes. As the last one is actually the diversion dike method, the present study examines only the suction and blowing methods.

2    EXPERIMENTAL SETUP AND TEST RUNS

The experiments are conducted in a slope-adjustable flume with a length of 13m and width of 50cm. The model spike dam has a dimension of 25´25cm and is fixed on one side of the flume wall in the test section. The downstream and lateral side of the model spike dam both have 4 openings on it, for blowing and suction purposes. Fig1. is schematic description of test setup.

Discharge through openings is controlled by a valve connected to each one, so that the blowing discharge equals that of the suction. In different test runs the flume discharge and the flow depth are varied. For each run, the length of circulation is controlled by adjusting the blow/suction discharge. Results of circulation length measurements in a test run are compared with each other, and the most effective discharge can be found out by analyzing the pattern of flow about the spike dam.

Measurements of velocity profile over flow depth are taken with a propeller velocimeter at section 1-1, 2-2 and 3-3 in each run, after stable flume flow is established. Flume discharge is evaluated from the velocity profile on several vertical lines at section 1-1. Flow pattern downstream of the spike dam is taken with PTV (particle tracking velocimeter) technique based on video recording of surface tracker motion and image processing method.

Fig. 1   Measurement section setup

Flow depths in test runs are H=10cm, 15cm and 18cm for different cases. For each depth, suction/blowing discharges have four variations by shutting down 0,1,2,3 openings on the lateral and downstream sides of the spike dam. In some cases half-opening is used to found out the optimal suction/blowing discharge. The results of test runs are summarized in Table 1.

Table 1    Test conditions and summary of results

Note: Case1 =no suction/blowing; case 2=one opening is used for suction/blowing; case 3=two openings used; case 4= three openings, case 5=only one half of the opening’s capacity is used.

3    EXPERIMENTAL RESULTS AND ANALYSIS

3.1    Effectiveness of the method

Base on experimental measurements, relationship between reduced circulation length L and suction/blowing discharge Q_S is established, as shown in Figure 2, in which Q_F is total discharge in the flume, L₀ is original circulation length without suction/blowing. The length reduction is most effective when Q_S is being increased from 0 to 0.01~0.02 Q_F, i.e., a small increase in Q_S/Q_F can cause a large drop in L/L₀. Further increase of Q_S after the point Q_S= 0.02Q_F does not repress the circulation much further, and the L/L₀ value is about 10~15%, almost independent of Q_S /Q_F .

Flow patterns taken from four cases of suction/blowing discharges (Cases 1,2,4,5) in Run No.1 are shown in Figure 3, with a constant flow depth H=10cm. Figure 3 shows that, when there is no suction/blowing discharge (Case 1), the original length of circulation is L₀=235cm. In Case 2, when a single opening is used, the circulation length is reduced to L=35cm. In Case 5, when by valve control, only half the discharge capacity of an opening is used, the circulation is still sufficiently under control with L=50cm. It is obvious from Table 1 and Figures 2 and 3 that both the length and the intensity of the circulation have been reduced by acceleration of the boundary layer and removal of decelerated fluid, which confirms the effectiveness of this method in circulation size control.  

Fig. 2    Observed relationship between L/L₀ and Q_S/ Q_F

3.2    Optimal discharge

Flow patterns of Case 4 in Figure 3 indicates that strong suction/blowing discharge, while reducing the size of the circulation downstream of the spike dam, also causes a new circulation at the opposite boundary of the flow. An optimal suction discharge should therefore be found out, designated by Q_OP, which not only reduces the size of circulation caused directly by the existence of the spike dam, but also avoids causing a new circulation by the suction/blowing action itself. From Figure 3 it can be seen that the use of half-opening (Case 5) can produce such an ideal result, and Case 5 is the optimal suction/blowing discharge for this test run condition. The use of optimal suction discharge Q_OP is also important in engineering practice, as larger discharges of suction/blowing mean higher construction costs, which could make this measure inefficient and not economical to build.

Fig. 3   Flow pattern for various suction/blowing discharges

Since it is not the best solution to make the circulation downstream of the dike completely disappear, the optimal discharge is found out by adjusting the valve on the opening so that the circulation length behind the dam is about 1~2 times the dimension of the spike dam, i.e., L=25~50cm. This flow pattern is called the optimal flow pattern. The optimal discharge, Q_OP, is large enough to reduce the circulation length to the required range (25~50cm), and is small enough compared with the flume discharge. During each test run, optimal flow pattern is achieved for the flow condition by adjusting the suction/blowing discharge to a suitable level, and at the same time the 4 optimal discharges are obtained.

3.3    Optimal discharge and flow condition

A number of methods can be used to calculate boundary layers with suction (e.g., see Wuest 1961). Theoretical analysis shows that, for a flat plate with homogeneous suction at zero incidence, the velocity of suction just sufficient to prevent separation all along the wall, v₀, can be expressed by the following equation (Schlichting 1979, 383-393)

where dU/dx is the longitudinal gradient of cross-sectional averaged velocity, and n is kinetic viscosity. In the present study it is also found that the lateral velocity gradient, du/dy, at the Section 2 in Figure 1, is a good indicator as well for flow disturbance due to the existence of the spike dam, where u is the mean velocity over depth at each point of the cross-section. Experimental results in this study give good relationships between both Q_OP /Q_F ~ dU/dx and Q_OP /Q_F ~ du/dy, as shown in Figure 4. It can be seen that higher dU/dx or du/dy requires larger Q_OP .

(a)                                                      (b)

Fig. 4    Observed Q_OP /Q_F ~ dU/dx and Q_OP /Q_F ~ du/dy relationships

Based on theoretical results and dimensional analysis, the following non-dimensional parameter, L, is found to adequately represent the related hydraulic factors

L=

In which h is depth of flow and u_* is shear velocity. The ratio of optimal suction discharge over total flume discharge, Q_OP /Q_F, shows good relationship with L, as can be seen in Fig. 5.

Fig. 5    Observed Q_OP /Q_F ~ L relationship

4   CONCLUSIONS

In the present study, flume tests are conducted to investigate the possibility of using boundary control method to reduce the size of circulation downstream of a spike dam. The following conclusion can be drawn from experimental data and analysis of results:

(1) Experimental data shows an effective reduction of circulation size by using the suction/blowing method. The circulation size is quite sensitive even for small increases in Q_S , when Q_S <0.01~0.02Q_F.

(2) After suction discharge Q_S exceeds a certain level, e.g., Q_S =0.02 Q_F , circulation size remains stable and does not show sensitivity to further increases in Q_S. Large Q_S can then produce a new circulation at the boundary opposite to the spike dam.

(3) From experimental results it can be seen that an optimal suction discharge exists which is large enough to reduce the circulation length to the required range (25~50cm in this study), and small enough compared with the flume discharge. For each of the four runs an optimal discharge is found out.

(4) The optimal discharge is related to hydraulic actors of the flow, which is represented by L= in the present study.

(5) It should be stressed that the experimental results of Q_OF /Q_F » 0.02 cannot be directly applied in an realistic engineering situation as site conditions will be greatly different from a flume flow. Besides, for large rivers like the Yangtze, where discharge in river channel is often on the magnitude of 10000~100000 m³/s, the ratio of Q_OF /Q_F » 0.02 can mean huge capacity of the suction/blowing discharge and may not be economically feasible to build.

Acknowledgements

This study was supported by the Three-Gorges Development and Construction Corporation. The authors are thankful to Prof. Lin Bingnan at International Research and Training Centre on Erosion and Sedimentation (IRTCES) for his suggestions in conducting this study.

References

Chang, P.K., 1976, Control of flow separation, Hemisphere Publishing Corporation, Washington D.C.

Lachmann, G.V. (ed.), 1961, Boundary layer and flow control, I and II, Pergamon Press, London.

Schlichting, H., 1979, Boundary Layer Theory, 5th ed., McGraw-Hill, Inc., New York.

Wuest, W., 1961, Survey of calculation methods of laminar boundary layers with suction in incompressible flow, in: Boundary layer and flow control, ed. by G.V. Lachmann, pp.771-800, Pergamon Press, London.