Wu Caiping, Gou Zhaoli, Guo Huimin and Song Lixuan

Yellow River Institute of Hydraulic Research, Zhengzhou, 450003, China

Abstract: With the data measured on a bed movable physical model for a pilot project, the coarse diversion at the Xiaobeiganliu reach, the paper represents of the research results of critical technical problems like location of diversion (entrance) sluice, the entrance/exit sluice operation, separation of coarse sediment from fine sediment over the weirs at the bends, layout of the grid-patterned dikes. By combining all the techniques together, we could both trap on the local floodplains the coarse sediment, which is a part of the sediment diverted from the Middle Yellow River, and discharging the fine sediment back to the river.

Key words: Yellow River, sediment trapping, physical model, warping

In 2004, Yellow River Conservancy Commission (YRCC) did an on-the-site experiment of coarse sediment diversion at the local Libotan floodplain at the Xiaobeiganliu reach, a lower part of the Middle Yellow River. The pilot project was intended to reduce the amount of the coarse sediment entering the lower Yellow River so as to find out if it can be taken as a new strategy for training the Yellow River. There were many critical technical problems involved in its planning, design, and operation of the on-the-site experiment, such as how to determine a proper location for the diversion (entrance) sluice, how to operate the sluices, how to separate coarse sediment from fine sediment over the weirs at the bends, how to determine the layout of the grid-patterned dikes, and how to operate exit sluice. Our institute had been assigned to do experimental investigation of the problems on a physical model before the on-the-site experimental was done.

The paper shows the research results of the critical problems. The investigation provided sound technical support for implementing the 2004’ coarse sediment trapping on the site and is of practical implications in implementing large-scaled projects of the kind in the future.

The pilot project of trapping coarse sediment on the Lianbotan was done with gravity division. The area for sediment trapping is planned to be 2.5km²，which can be used to trap coarse sediment of 0.232 billion m³. Six canals are to be made through the trapping area. The first-stage project was to construct the first canal next to the Middle Yellow River and diversion sluice, sediment transport canal, cofferdam and grid-patterned dikes and flow return sluice. The canal is averagely 640m wide, 8.6km long and 5.5 km² in area. Investigation indicates the inlet diversion sluice should be at a river training works at Xiaoshizui. The design flow conveyance of the sluice is 67m³/s. When flow rate in the local Yellow River is 500～4000 m³/s，the sluice can deliver flow of 50～108m³/s. The sediment transport canal is 20m wide at its bottom and 2600m long，with side slopes of 1:2.5. Its longitudinal slope is 4 and 5 ，has roughness of 0.0017. Two overflow weirs are made at the two bends of different radii, respectively. The trapping area is broken down to three sub-regions by a longitudinal dike and a lateral dike (Fig.1).  

Location Of The Pilot Project

 

To serve the planning and the design of the on-the-site experiment, investigation was done on a physical model to test the performance of the diversion sluice in drawing flow and sediment from the Middle Yellow River. The model is a schematized bend flume with a horizontal scale 1:100，and a vertical scale 1:20. The investigation focused on how to determine the location for diversion sluice, the diversion canal orientation relative to the direction of the main flow in the local Yellow River, per-unit-width flow discharge of the local Yellow River, ratios of the diverted per-unit-width discharge to the per-unit-width discharge in the river, effects of the suspended particle sizes and sediment concentration on the performance of the diversion sluice.

Analysis of the experimental data of the fluid field and sediment transport in the front of the diversion sluice and the local river, together with theoretical analysis, led to the findings that the diversion sluice would have much better performance in diverting flow and sediment, in terms of diversion guarantee ratio, amounts of the diverted flow and sediment and the particle sizes when the sluice is located at the concave side of the leading arm of a bend of the river channel than at the other places; and that it will be more convenient to implement layout of the structures and complete the construction if the diversion sluice is at the concave side than at the other places. So it is recommended that the diversion sluice should be at the concave side of the leading arm of a bend of the river channel, namely, at No.1 spur-dike of the Xiaoshizui Works. The experimental results show that the location is liable to receive the river flow and make better performances in diverting flow and sediment for three kinds of flow and sediment regimes.

One of the principles in operating the diversion sluice is to draw more sediment, especially more coarse sediment from the river. To realize the objective of trapping coarse sediment and discharging the fine sediment back to the river, proper operation of the diversion sluice in terms of timing the sluice lifting and the lifting height (openness) is necessary. Our institute performed an experimental investigation of the effect of river flow and sediment (including particle size and concentration) and other factors on diversion performance. It is found that compared with the other factors, river flow is of little influence on the diversion performance and can be neglected and be excluded in the scheme for the diversion sluice operation. That provides a quite large free room for formulation of the operation scheme.

The river sediment factor imposes an apparent influence on the diversion performance. When the river carries high sediment-concentrated flow, there are high ratios of the diverted sediment load to the river sediment load. Similarly, it is true for sediment particle sizes. So it is recommended that when high-concentrated river flow with coarse sediment particles comes, the diversion sluice should be lifted.

Analysis of the experimental data and the hydrological data measured at the Longmen gauging station led to the finding that the local Yellow River carries quite low sediment concentrated flow, which is averagely under 10 kg/m³, in the non-flood seasons. So diversion should be done, not in the non-flood seasons, but in the flood seasons.

With the investigation, we recommended that the diversion operation at the floodplain of Libotan should be performed when river flow is greater than 500m³/s with sediment concentration greater than 50kg/m³ with mean particle size of the suspended greater than 0.026 mm，and the sediment load with the particle  size greater than 0.05 mm more than 20％ of the total suspended load..

Overflow weirs at bends are the facilities to sort out the fine sediment from the coarse sediment in the sediment-loaded flow drawn from the Yellow River. Circular flow at the bends had found many applications in riverine engineering, port and navigation channel improvement, but had not found any application in the Yellow River before completion of the experimental investigation. The findings made in the investigation led to the understandings as follows.

The sediment-transport canal has two bends where the overflow weirs are planted. The upstream bend has its radius four times the canal width while the downstream one has its radius 2.5 times the canal width. Monitoring of flow and sediment factors at the bends were conducted. Fig.2 shows vertical distribution of measured concentration at the bends, which is averaged over the whole diversion period. From Fig.2 we can see that due to effect of the circular flow, lateral sediment transport is in non-equilibrium, which can be embodied by the fact that the concentration at the convex side is higher than at the concave side. Vertical distributions at the convex and the concave sides show that concentration at lower layers is higher than at upper layers.

Fig.2 shows another feature, which is that average concentration is somewhat higher at the lower bend than that at the upper bend due to the fact that a part of water and sediment has been diverted over the upper overflow weir to the water exit canal.

Analysis of the particle grading data shows that the suspended particles are finer at the concave side than at the convex side. At the upstream bend, D₅₀ is 0.018 mm at the concave side with 60% of the particles being smaller than 0.025mm while at the convex side D₅₀ is 0.024mm with 51％ of the particles being smaller than 0.025m.

It is true for the situation at the downstream bend. At the downstream bend, D₅₀ is 0.02mm at the concave side with 58% of particles being smaller than 0.025mm while at the convex side, D₅₀ is 0.026mm with 48% of particles being smaller than 0.025mm.

In addition, it is apparently seen that the suspended particles are somewhat coarser at the upstream bend than at the downstream bend.

Fig.3 shows that the ratios of the concentration measured at the branches downstream of the overflow weirs to the concentration upstream of the weirs. Although the data points seem

scattered more or less, they are still robust in indicating a fact that the mean concentration at the branches is about 10% lower than the concentration at the canal upstream of the weirs.

Similarly, the representative sizes of the suspended particles at the branches are about 10% smaller than those at the canal upstream of the weirs.

The objective of the on-the-site pilot project is to trap coarse sediment particles and discharge fine particles. By “trapping coarse sediment” we mean that the particles bigger than 0.05mm in diameter will be detained on the low floodplains, and by “discharging the fine” we mean that the fine particles in the flow will be sent back to the Yellow River channel. To realize the objective, we need to operate the sluices properly, to separate the sediment by using bends and weirs, and make a proper layout of the dikes over the floodplains (Fig.1).

We performed an experimental investigation of the layout on the model with horizontal scale being 1:120 and vertical scale 1:20.

The input data was the typical hydrographs of sediment and flow measured in 1998, totally 37 days on the model run, representing 2 years and 5 days in the reality. The mean daily flow rate is 63 m³/s～101m³/s while the mean daily concentration is 51 kg/m³～328 kg/m³ . The fed suspended sediment grading is determined by following both the model scale law and the measured sediment grading (D₅₀=0.025mm measured).

The objective of trapping the coarse and discharging the fine is realized by operating the sluices properly. The objective in the sub-region (1) in Fig.1 is realized by the proper operation of the height of the sand bags at the outlet at the lateral dike so as to control the out-going sediment concentration and sediment grading while the objectives in the sub-regions (2) and (3) are realized by the operating the heights of the horizontal beams of the exit sluice.

Due to the limitation of survey time with the Optical-Electronic Grading Meters, there is no way to control the index of the particle sizes. The operation of the exit sluice beams should be such that the ratios of the sediment discharging of the trapping area are 30％ ～ 50％. The operation of the exit beams and the sand bags at the outlet on the physical model is shown in Fig.4.

Considering the fact that sediment grading can be better controlled on the site than on the model, we expect that the performance in trapping the coarse sediment and discharging the fine sediment could be better on the site.

 

The experiment ran for 37 days on the model and made a deposition of 11.29 million m³ in volume (in prototype). Fig.5 shows the longitudinal profiles in the trapping sub-regions (1) and (3) while Fig.6 shows the longitudinal profile in the sub-region (2). The figures show that the trapped sediment is distributed in strips and the thickness of the deposition increases downstream. The average deposition thickness in the upper parts of the trapping sub-regions is about 2~2.5m and the surface slope is 5.3  while in the lower parts the average thickness is 2.5~3.5 m and the surface slope 3.2 .

The area upstream of the milestone km 6.2 shows apparent existence of a channel and there is 0-1 m difference in elevation of the channel bottom and the floodplains while the area downstream of km 6.2 shows no apparent existence of the kinds due to the backwater effect made by the ever-rising beams of the exit sluice or the ever-rising sand bags. The average slopes in the upper and the lower parts are about 4.4 。

   

After completion of the trapping, sediment samples were taken from the trapping area. Fig.6 show the longitudinal distributions of mean diameters of the deposited particles. They show that the particles become finer and finer downstream in the channel or on the floodplains due to sediment sorting-out effect, and that the particles in the channel are coarser than on the floodplains.

The deposited particles upstream of the milestone km 5.2 in the sub-region (2) and the milestone km 6.2 in the sub-regions(1) and (3) are quite coarse, about 0.030~0.040mm in diameter, while those downstream of the milestones are quite fine, about 0.018~0.030 mm due to the backwater effect by the exit sluice or the sand bags at the outlet of the trapping area.

Fig.6  Longitudinal distributions of the mean particles in sub-regions

Fig.7 shows the particle size grading curves taken longitudinally in the channel in the sub-reginon(1), (2) and (3) . It can be seen that, 40％~25％ of the deposited sediment particles in the areas upstream of the milestone km 6.2 in the sub-regions (1) and (3) and the milestone km 5.2 in the sub-region (2) are bigger than 0.050 mm while 10％～25％ of the deposited in the lower parts are bigger than 0.050 mm.

Base on the data measured on the model, there is 24.90 million m^3，of sediment drawn from the Yellow River channel, 5 million m^3，of which are coarse sand with their diameters being bigger than 0.050mm. There is 11.29 million m^3，of sediment deposited in the trapping area. 60% of the deposited is in the areas upstream of km 6.2 in the sub-regions (1) and (3) and km 5.2 in the sub-region (2), with 35% of it consisting of coarse sand (bigger than 0.05mm in diameter) while the rest 40% is in the areas downstream with 17% of it consisting of the coarse sand. The rate of trapping all the fractions of sediment in the whole trapping area is 45% and the rate of trapping the coarse sediment fraction is 63%. That means that 2/3 of the coarse sand that is diverted into the trapping area is successfully trapped there. The rate of discharging all the fractions of sediment at the exit sluice is 55% and the rate of discharging the fine fraction (D<0.025mm) is 68%. That means that 2/3 of fine sediment that is diverted to the trapping area is discharged back to the Yellow River channel.

The experimental results show that the design can realize the objective of trapping the coarse and discharging the fine. The on-the-site experiment improved the performance by controlling higher discharging standard.

Most of the key technical problems that were involved in the coarse sediment trapping project at the Libotan floodplain were solved on a physical model. The experimental results provided sound technical support for completing the on-the-site project. Selection of proper location of the diversion sluice is the guarantee for the successful implementation of the on-the-site project. Proper operation of the diversion sluice and exit sluices are the important measures. Overflow weirs at the bends help to separate the coarse sediment from the fine sediment. By combining all the techniques together, we could both trap on the local floodplains the coarse sediment, which is a part of the sediment diverted from the Middle Yellow River, and discharging the fine sediment back to the river. With the support of the research done on the physical model, the on-the-site pilot project successfully realized its objective of trapping the coarse sediment and discharging the fine sediment back.

Zhao Yean, Zhou Wen Hao, etc, Fundamental Law of Fluvial Process in the Lower Yellow River, Yellow River Press, 1998.

Chen Ji Wei, Xu Ming Quan, etc, Training the Yellow River and Developing the Water Resources in the Yellow River Basin, Yellow River Press, Dec.1998.

Physical Model Experimental Investigation of Trapping Coarse Sediment in the Liantan Floodplain at the Middle Yellow River (Research Report), Yellow River Institute of Hydraulic Research, May 2004.

Analysis of the Performance of the on-site Experiment of Trapping the Coarse Sediment in the Liantan Floodplain at the Middle Yellow River (Research Report), Yellow River Institute of Hydraulic Research, 2004.3