Flow in two-stage channels with the main channel skewed to the flood plain direction
The work of this thesis is an experimental study of flow in a compound or two-stage channel, where the flow in the main channel is skewed slightly to the flood plain flow direction. The purpose of the study is to investigate flow mechanisms and flow behaviour in a situation which is more complex than a straight/parallel compound channel and less complex than meandering compound channel. In other words, an element of `lq cross over flow is introduced, where the flood plain flow crosses over a skewed main channel beneath. This element is absent in study of a straight and parallel channel/flood plain system. The thesis begins by introducing the subject of two-stage channels and river flooding in Chapter 1, followed by a literature survey in Chapter 2. The literature survey attempts to draw together various strands of two-stage channel research including stage-discharge assessment, lateral turbulent shear, bend behaviour for inbank and overbank situations, and flow behaviour in the cross-over region. Chapter 3 describes the experimental rigs used during the experiments including the main flume for the skewed compound channel flow tests, and also a smaller straight flume with a slot of variable length in the main bed, designed to simulate certain aspects of cross-over flow. This chapter emphasises the instrumentation used to measure accurate point velocities and accurate streamline angles throughout the flow field. Both these measurements were crucial in determining transverse velocity components. Chapter 4 describes flow resistance and stage-discharge data for the skewed compound flume with both smooth and rough flood plains, and also the skewed main channel operating in isolation from adjacent flood plains. An analysis of drag coefficients for the vertical roughening rods for the flood plain is also presented. Chapter 5 presents the experimental data of longitudinal velocities throughout the flow field and transverse velocities within the skewed main channel, for the case of both smooth and rough flood plains. Observations are made on the asymmetry of the longitudinal velocity profile, the magnitude of transverse velocity components, as well as variations in the downstream direction. Chapter 6 concentrates on flow measurements in a smaller flume with a slot in the channel bed, representing an idealised version of flow crossing-over the skewed main channel. Measurements are presented of variations in water surface level, energy losses at the expansion and contraction region as well as velocity measurements at various sections throughout the slot region in an attempt to produce a universal velocity profile. Finally Chapter 7, attempts to draw together experimental evidence in Chapter 4, 5, and 6, to draw some conclusions about the behaviour of two-stage channels. This is done by putting forward ten ideas on the behaviour of two-stage channels involving stage-discharge assessment for the smooth boundary case, stage-discharge assessment for the rough boundary case, flow resistance concepts for two-stage channels, deviation of the flood plain streamlines in the cross-over region, entrainment of flood plain flow down into the main channel flow, the magnitude and distribution of secondary circulation within the skewed main channel (and how this is influenced by boundary roughness, aspect ratio and relative flow depth), the magnitude and distribution of the depth-averaged longitudinal velocity (including comparison of left and right flood plain velocities, position of maximum velocity filament, width of lateral shear layers on either flood plain), variations of flow behaviour in the downstream direction, application of the result to two-stage channel design and finally some speculations on the longer term behaviour of two-stage channels from the viewpoint of scouring and deposition of bed load and suspended sediment. The thesis shows very clearly that flow is a skewed compound channel, even at a small skew angle of 5.84o, is in many ways dissimilar to flow in the straight/parallel case. This is highlighted in the magnitude of secondary cells in the skewed main channel, deviation of flood plain streamlines, greater horizontal shear and less lateral shear between main channel and flood plain and asymmetry of the depth-averaged velocity. This thesis also reveals a very limited application of energy losses at a slot in the channel bed to a skewed cross-over flow situation.