Experimental study on the influence of the geometry of shallow reservoirs on flow patterns and sedimentation by suspended sediments

The worst enemy of sustainable use of reservoirs is sedimentation. Often the main silting process is the result of settling down of suspended sediments. In shallow reservoirs the flow pattern and the sediment deposition processes are strongly influenced by the reservoir geometry. The trap efficiency of a shallow reservoir depends on the characteristics of the inflowing sediments and the retention time of the water in the reservoir, which in turn are controlled by the reservoir geometry. With the purpose of controlling the sedimentation in shallow reservoirs, the effects of the geometry on flow pattern and deposition processes were investigated with systematic physical experiments and numerical simulations. This allowed identifying ideal off-stream reservoir geometries, which can minimize or maximize the settlement of suspended sediments. The objective was also to gain deeper insight into the physical processes of sedimentation in shallow reservoirs governed by suspended sediments. The systematic experimental investigations were carried out in a 6m long, and 4m wide and 0.3m deep shallow basin. The influence of the shallow reservoir geometry was studied for the first time by varying the width, the length, and the expansion angle of the basin in the experiments for clear water and with suspended sediment. In total 11 different reservoirs geometries and 4 water depths were analyzed. During the tests several parameters were measured, as 2D surface velocities, 3D velocity profiles, thickness of deposited sediments, and sediment concentration at the inflow and outflow. Crushed walnut shells with a median grain size, d50, of 50 µm, and a density of 1500 kg/m3 were used to simulate the suspended sediments. After having reached a stable flow pattern with the clear water, velocity measurements were performed. In a second phase, the evolution of the flow and deposition patterns under suspended sediment inflow were investigated. Tests were carried out with durations from 1.5 hours up to 18 hours, in order to follow the morphological evolution. In order to investigate the efficiency of flushing, flushing operations at normal water level as well as with drawdown were examined after the sedimentation tests. Numerical simulations of the laboratory basin were performed and compared with experimental results. The purpose was to assess the sensitivity of the results on different flow and sediment parameters and different turbulence closure schemes. The experimental investigation of the flow and sediment behavior in axi-symmetric geometries with different shapes provides further information on the evolution of the flow pattern and the sediment deposition. Beside the expansion ratio and form ratio of the basin the flow regime was classified by the geometry shape factor SK and inlet Froude number Frin. The geometry shape factor, defined as a function of wetted perimeter, total reservoir surface area, aspect ratio and expansion density ratio, was used to compare and analyze the experimental results obtained from the different investigated basin geometries. The clear water experiments investigations revealed under what geometrical conditions the flow changes from symmetrical to asymmetrical behavior. The length of the basin has a strong influence on the change of the flow field. The water depth also significantly affects on the type of jet and vortex structure forming in the basin. On the other, hand the sediment deposition pattern was strongly influenced by the jet type and the flow behavior changed with increasing deposits. The prediction of sediment deposition is linked to the prediction of flow behavior. Furthermore it is very sensitive to the basin geometry and the boundary conditions of inflow and outflow. Some tests were performed until the sediment released at the outlet was equal to the sediments entering at inlet into the basin that means a quasi equilibrium was reached. Flushing at normal water level allows only a relatively small part of the deposited sediment to be evacuated. As expected, drawdown flushing is much more effective and a significant amount of sediment deposits could be flushed out of the basin. Regarding the flow pattern in shallow basins empirical relationships for the estimation of the reattachment length of gyres and the normalized residence time as a function of the geometry shape factor, SK, were established. Also empirical equations for the prediction of the jet flow type and velocity magnitude depending on the basin geometry could be formed. Finally empirical equations for the prediction of sedimentation index, silting ratio, trap efficiency and relative deposition thickness, as well as normalized residence time and flushing efficiency could be found. The numerical simulations revealed that the observed asymmetry in flow and deposition patterns can be explained by the sensitivity of the flow regarding geometry and boundary conditions. The influence of the length and width of the basin on the flow pattern can be predicted in good agreement with experiments by these simulations. Some recommendations are given for the design procedure of a new shallow reservoir in view of minimizing the sedimentation due to suspended sediment. The deposited sediment volume can be efficiently minimized by an optimal designed reservoir geometry.