Flow resistance and sediment transport in compound channels =&GV\ ju
Yang Kejun 2e YkWHi
ABSTRACT &>{L"{
I9 m
A compound channel comprising the main channel and floodplains exists in natural rivers widely, especially in alluvial streams. It differs from a single channel in adjusting flood, cutting flood peak, transporting sediment etc. When water in the main channel flows in an out-of-bank condition and on to the adjoining floodplain, owing to abrupt change of the shape of cross section and heterogeneous boundary roughness, there are a bank of vertical vortices along the vertical interface between the main channel and its floodplain, which will result in complex variations in flow structure, flow resistance, sediment transport and fluvial processes. The study of the complex behavior of flow movement and sediment transport in compound channels is profoundly important for the basic theory of hydraulics, mechanics of sediment transport and river dynamics. The research results benefit to flood control, channel training, floodplain exploitation etc. The thesis will systematacially investigates flow resistance, mechanism of momentum transfer, conveyance capacity in non-vegetated compound channels, flow structure in a compound channels with vegetated floodplains, flow resistance and sediment transport in a self-formed one. The main contents of the thesis are as follows: g' E^@1{
1. Flow resistance in compound channels "H@AT$Ny(
(1) To study resistance coefficients in compound channels. This thesis analyzes and discusses the effect of cross-sectional shape on Manning’s and Darcy-Weisbach resistance coefficients. By analyzing the experimental data from Science and Engineering Research Council Flood Channel Facility (SERC-FCF), the relationships between overall, zonal, local resistance coefficients and a wide range of geometries and different roughness between the main channels and its associated floodplains have been established. Moreover, the reason why the conventional methods can not assess the conveyance capacity of compound channels is analyzed. The reason is that single channel method doesn’t consider the fact the composite roughness varies with flow depth and cross-sectional division method ignores the extra resistance produced due to the momentum transfer between the main channel and floodplains, in assessing the conveyance capacity in compound channels. According to the experimental results of SERC-FCF, it is shown that the overall Darcy-Weisbach coefficient for a compound channel is the function of Reynolds number, but the function relationship is different from that for a single channel. ']x]X ,
(2) To compare and analyze the method for predicting composite roughness in compound channels. This thesis systematically sums up all kinds of different representative methods for predicting composite roughness. According to the hydraulic parameter required and whether the momentum transfer is considered or not, they can be simply classified into several groups. A vast number of experimental data and field data for compound channels are applied to check the validity of the mentioned methods. Meanwhile, the effect of the division type of cross section on the computation of composite roughness is analyzed. Any method for predicting composite roughness will result in error. Among them, Lotter method is closely related to the division type of cross section, while Einstein-Banks method is not related to that. By comparing and analyzing the above methods, it is pointed out the methods are not fit to assess the composite roughness in compound channels. The reasons why the methods result in errors are analyzed. AWcP OU
2. Mechanism of momentum transfer in compound channels ppA8c 6
(1) To undertake the analysis of kinetic energy loss intensity in compound channels. The intense momentum transfer on the vertical interface between the main channel and floodplains in a compound channel, makes its conveyance capacity decrease. To reflect the kinetic energy loss intensity in a compound channel, two new concepts, transverse kinetic energy correction coefficient (TKECC) and kinetic energy loss rate (KELR) are put forward for steady, uniform and turbulent flow in the thesis. By the analysis of the mechanism of kinetic energy loss in compound channels, a conclusion is drawn that TKECC is larger than 1 and KELR is larger than 0 in compound channels. By analyzing the experimental data from SERC-FCF, it is found that TKECC and KELR are both related to shapes of cross section. Kinetic energy loss becomes weaker with main channel side slope factor increasing and becomes stronger with the ratio of main channel and floodplain widths increasing. Kinetic energy loss in a symmetric compound channel is stronger than that in an asymmetric one. For all the shapes of cross section, kinetic energy loss increases with the relative depth increasing. After it reaches the largest value, it decreases with the relative depth increasing, and the compound channel ultimately shows a characteristic of single channels. P6rL;_~e
(2) To investigate momentum transfer coefficient in compound channels. Momentum transfer coefficient plays an important role in computing conveyance, apparent shear stress on the vertical interface between the main channel and floodplain, mean boundary shear stress on floodplains or in the main channel respectively. In this thesis, beginning with Boussinesq’s assumption and combining the characteristics of velocity distribution in the interacting region, the expression of momentum transfer coefficient is derived theoretically; on the basis of force balance, the expression of vertical apparent shear stress is obtained; applying experimental data from SERC-FCF, the variation of momentum transfer coefficient with relative depth and the ratio of floodplain and main channel width, is analyzed; basing on the momentum transfer coefficient relationship obtained and applying Liu Peiqing method, the conveyance capacity in compound channel is calculated. The computed results show the momentum transfer coefficient relationship obtained is viable. St<\qC
3. Conveyance capacity in compound channels #D8)r s.9
(1) To compare and analyze the method for predicting the conveyance capacity in compound channels. For a compound channel, when water in a main channel flows in an overbank manner and inundates its floodplains, if the conveyance capacity is directly calculated by Manning equation, it results in a deal of error. All kind of methods for predicting discharge are systemically summarized in this thesis and are applied to compute the cross-sectional discharge and the distribution of discharge in the main channel and floodplains, respectively. By comparing with different series of experimental data from SERC-FCF, it is found that the discharge error is large by cross-sectional vertical division method (VDM), single channel method (SCM) and equivalent velocity division method (EVDM). Otherwise, the calculation accuracy is high by the other methods. According to the order from high accuracy to low, the methods are，in turn， channel coherence method (COHM), Shiono and Knight method (SKM), inclined division method type 1 (IDMT1), momentum transfer method (MTM), Xie Hanxiang method (XHXM), superposing cross section method (SCSM), Tong Hanyi Method. In the meantime, the methods with high-calculation accuracy are used to calculate the discharge distribution. It is found that for certain given cross section, it is difficult to determine which method has the highest discharge distribution’s accuracy, but for all the series, as a whole, the highest is SKM, the next is IDMT1 and THYM, and the rest is COHM，MTM，XHXM and SCSM. In addition, the thesis analyses the effect of secondary flow on conveyance capacity, discharge distribution and flow resistance. The results show that it has a very small influence on conveyance capacity, but it does distinct influences on discharge distribution, lateral distributions of depth mean velocity and boundary shear stress, the mean velocities and boundary shear stresses of main channel and floodplains, respectively. By analyzing the merit and demerit of the above mentioned methods, it is pointed out that COHM is a good method in calculating discharge. lom4z\ 6
(2) To establish the system dynamics model of conveyance capacity in compound channels. By analyzing a vast number of experimental data from SERC-FCF, it is found that the ratio of the Darcy-Weisbach resistance coefficients between the floodplain and main channel decreases with the relative depth increasing. That in the main channel varies with the relative depth and follows the paraboloidal distribution. On the basis of the resistance coefficient relationships obtained, applying system dynamics method, the thesis establishes the system dynamics model of conveyance capacity when water flows in an out-of-bank manner and onto the adjoining floodplain. The model gives the relationship between stage and discharge. There is a good agreement with modelled and experimental values. The absolute values of the relative errors are very little. up6LO7drW/
4. Flow patterns in compound channels with vegetated floodplains cvt2P}ma#
The thesis experimentally studies the velocity distribution of flows over different types of vegetations such as arbors, shrubs and grass. For vegetations on the floodplain, the thesis chooses plastic grass, duck feathers and plastic straws as model grass, model shrubs and model arbors, respectively. ADV is used to measure the local flow velocities for the cases of different types of vegetation on the floodplain, discharges and flume slopes. All measured streamwise velocities follow the logarithmic distribution for the case of non-vegetated floodplain, and obey S-shaped distribution for vegetated floodplain. The S-shaped distribution divides the flow into three regions. The range of every region is related to flow depth, lateral location and vegetation type. For different vegetation, the S-shaped distribution is different. In the meantime, it is found that the velocity in the main channel increases and that on floodplain decreases after the floodplain is vegetated. The increasing or decreasing degree is related to vegetation type. Furthermore, the influences of flume slope on local velocity distribution for different types of vegetation are distinct. For all cases, the fluctuating velocity follows approximately a normal distribution. The time-averaged velocity is related to the sampling duration. After the floodplain is vegetated, the lateral exchange of momentum becomes intensive. On the whole, the momentum transferrs from the main channel to its associated floodplain. In addition, the effects of different types of vegetation on turbulence intensity are distinct. The turbulence intensity increases after the floodplain is vegetated. The lateral and vertical ones are proximately equal and follow an S-shaped distribution. After the floodplain is vegetated, the distribution of Reynolds stress distinctly differs from that for non-vegetated case. Meanwhile, for different types of vegetation, the distributions are largely different. In the main channel side-slope zone, the Reynolds stresses become particularly complex. The effect of vegetation makes both tyx and tzx negative as a whole, differing from that of the non-vegetated-floodplain compound case. lM]7 @A
5. Flow resistance and sediment transport in a self-formed compound channels B}.ia_&DLR
The thesis experimentally studies flow resistance and sediment transport for inbank and overbank flows in a self-formed channel with complex cross section, including the variation of energy loss and resistance coefficients along the channel, the streamwise and lateral variations of median diameter of bed material after armoring is completed, the variation of bed load discharge with time during the armoring process. By analyzing the experimental data, it is found that water surface slope is largely different from energy slope for alluvial channels. Energy loss varies along the channel. The variation trends of resistance coefficient and energy slope are similar, as a whole, corresponding to the variation of representative diameter of bed material. Under the action of clear water, the river becomes wide quickly at the begging of the experiment. Then, the rate of river width adjustment becomes smaller. For the given experiment, retreat rate of river bank becomes small along the channel for the same discharge. Comparing with the original cross section, the main channel deposits while its two sides erode. The erosion or deposition areas decrease, corresponding to the decrease of mean velocity of cross section along the channel. It is surprised that the erosion areas of the two sides of the main channel are not equal although the original cross section is symmetric and the original size distribution is the same. Applying the concept of Shannon’s entropy, the equation of bed load rate is derived during the armoring process. Whether the bed laod rate increases or decreases with time, the structure and form of the equations are both the same. According to the characteristics, the general form of stochastic variable, X, is obtained. By theoretic analysis, the relationship between the parameter, ki, and the shape of Curve is established. If the curve is concave, ki is less than zero; otherwise, ki larger than zero. For inbank and overbank flows, the median diameter of bed load increases with time. Finally, it will approach to certain constant. The variation trend may be modeled by the expression of X. For the type of channel, during the course of armoring process, not only sorting along the channel but also across the channel exists. $
a7^3
gnt[l 0m
~ nIZ g5
Key words: Compound channels; Flow resistance; Momentum transfer; Conveyance capacity; Vegetation; Sediment transport
