Experimental investigation of floodplain vegetation density effect on flow hydraulic ‎in divergent compound channels

The velocity difference in the main channel with higher velocity and floodplain with lower velocity creates a strong shear layer in their junction, causing the production of additional turbulence structures, especially large-scale vertical vortices in this interface. In addition, because of turbulence anisotropy in the bottom and wall of the channel, secondary currents occur around the longitudinal axis and in a spiral shape. On the other hand, in most cases, due to the existence of vegetation on floodplains, investigation of the flow mechanism is far more complicated. There are usually three methods to explain the flow field and shear stress with the existence of vegetation on floodplains: 1) field measurements, 2) hydraulic models, and 3) analytical and numerical models. In natural rivers, since the flow cross-section changes along the river and the cross-section shape changes from prismatic to non-prismatic, with these conditions causing more mass and momentum exchange from the floodplain to the main channel and vice versa. this study has explored the effects of divergence angle and vegetation density on the flow structures in a non-prismatic compound channel. The experiments of this study were performed in an asymmetric compound channel made of Plexiglas with a length of 12 m, width of 0.6 m with a bed slope (So) of 0.8810-3. In order to model the vegetation on the floodplain, rigid cylindrical plastic rods with a diameter of (D) 10 mm were used. The spacing ratio (Sr = ly / D) for the three vegetation densities will be equal to 5, 7.5, and 10. Three divergence angles () equal to 3.8, 5.7, and 11.3 ͦ were created on the floodplain. Due to the formation of non-uniform flow in non-prismatic sections, the relative depths (Dr=yf/H) of 0.15, 0.25, 0.35, and 0.45 were set in the middle of the divergence region for all experiments. The longitudinal, transverse, and vertical components of the instantaneous flow velocity were measured by a 3D Vectrino profiler velocimeter at three sections: entrance, middle, and end of the divergence region. Using the transverse distribution of depth-averaged velocity, contribution of each section to the conveyance capacity was calculated. Due to the interaction between the rods, the flow structures are very different from the behavior of a single rod; thus, this should be considered in calculating the drag coefficient of an element set on the floodplain. For determine f, the Keulegan (1938) equation for smooth surfaces was modified. Jafari et al. (2011) proposed the an equation to calculate Strouhal number in a row arrangement. Because of outbreak the Kelvin–Helmholtz instability due to the existence of vegetation on the floodplain, in the interface between the main channel and the floodplain, coherent vortices and intense momentum exchange were formed from the main channel to the floodplain. Since the flow momentum prepared a shear layer around the vegetation stems, which causes inflection points in the velocity profile, which is consistent with Sanjou et al. (2010), Mulahasan et al. (2017), and Ahmad et al. (2020) results. At Sr = 7.5, the distance of the elements well forms Von Kàrmàn vortex streets and increases the flow resistance. At all relative depths, increasing vegetation density has reduced the Ufp / Umc ratio. The discharge rate through floodplain with vegetation has reduced by an average of 58.6 and 69.3% compared to non-prismatic channel without vegetation in the middle and end of the divergence reach, respectively. The results indicate that with increasing Dr, zonal roughness coefficient in the floodplain has increased nonlinearly and is linear in the main channel. This result is consistent with the Musleh and Cruise (2006) research. Drag coefficient has decreased nonlinearly with increasing the rod Reynolds number. In addition, it can be found that the drag coefficient caused by floodplain vegetation is directly related to the vegetation density. The results show that with increasing the vegetation density from Sr = 10 to Sr = 5 on the floodplain in the middle and end of the divergence, the bed shear stress has decreased by 44.2 and 54.6%, respectively. The vortex frequency is a linear function of Rerod and the increasing rate of vortex frequency versus Rerod in the middle of the divergence is higher than the end. In the zone close to the vertical interface between the main channel and the floodplain, the secondary currents have suddenly reached their maximum and minimum values. The results showed that with emergent vegetation, Kelvin-Helmholtz instability caused the generation of primitive Von Kàrmàn vortex streets in downstream of the elements. The existence of vegetation in the floodplain caused a sharp reduction in the bed shear stress in this region and increased it in the main channel. As the vegetation density increased, so did the drag coefficient and flow friction factor significantly. The flow passing through the vegetation was controlled by coherent vortices whose maximum size was in the interface between the main channel and the floodplain.