dc.description.abstract | Sediment bypass tunnels (SBTs) belong to the active sediment management strategies of combating reservoir sedimentation, which route high-speed sediment-laden flows around reservoirs and primarily experience supercritical narrow channel flow conditions that characterize secondary currents and redistributed primary velocity, turbulence properties, and bed shear stress, which intensified by the presence of in-plan bend. These alternations in the flow characteristics influence the sediment transport, invert abrasion, and hence, the design of SBTs. Although recent laboratory-scaled experiments in straight channels produced two-dimensional flow properties, individual sediment transport, and hydro-abrasion estimate models, three-dimensional investigations are scarce, and therefore, a supplementary and cost-efficient design alternative can be the computational fluid dynamics (CFD) simulation. Furthermore, field investigations at SBT bends demonstrated deeper invert abrasion near the inner wall than toward the outer wall. However, a detailed investigation of such flow characteristics and sediment transport was missing. Under these circumstances, research was initiated at NTNU to bridge the gap between experimental and numerical studies of high-speed flows while emphasizing some of the challenges and phenomena associated with SBT hydraulics.
This research focuses on (a) developing open-source CFD models to simulate the turbulence-driven secondary currents and their interactions with the flow characteristics in straight narrow channels conveying supercritical flows and subsequently applying this model to investigate the effect of cross-sectional geometries found at SBTs and to compare open channel flows with pressurized flows or duct flows, which are sometimes observed at SBTs and (b) experimental investigation of the effect of a channel bend on the flow properties, bed shear stress, and bed load transport those are linked to the abrasion patterns observed at SBT bends and producing experimental data useful in further development of numerical models. Here, computationally economical Reynolds stress models were utilized to compute steady-state results of high-speed flows after modifying the existing model codes in OpenFOAM, as the alternate transient options are still computationally very demanding. Furthermore, a state-of-the-art particle tracking velocimetry system in conjunction with the Shake-The-Box particle tracking algorithm was employed to measure the flow characteristics in a channel bend model of the Solis SBT, Switzerland, and high-speed cameras were utilized to observe the bed load movements.
The modified Reynolds stress models compute promising results of turbulence-driven secondary currents, distributions of longitudinal velocity and turbulence parameters, and bed shear stress undulation and its averaged value for uniform supercritical flows. Channel aspect ratio influences the secondary current formations and their influence on the flow characteristics and bed shear stress. Intermediate vortex develops with decreasing aspect ratio and is fully developed for aspect ratios ≤ 1.05. The bottom vortex guides the near-bed redistribution of flow properties and bed shear stress undulation. Additionally, altering the sidewall curvature influences the bottom vortex and developing intermediate vortex, bottom corner zone of low-momentum fluids, and bed shear stress distribution. A horseshoe cross-section with plane invert is the suitable design choice over the archway and circular geometries, as it helps in shrinking the corner zone of low-momentum fluids while providing a wider invert than a circular cross-section without imposing a stress concentration. Besides, the flow properties in the bottom half flow depth and bed shear stress distribution of open channel flows are comparable to that of duct flows with the same aspect ratios, particularly for a square cross-section due to similar bottom corner configurations and comparable secondary currents. However, duct flows experience greater boundary shear stress than open channel flows. In addition, the experimental investigation indicates that the free surface undulates significantly across and along a channel bend caused by the cross-waves, the curvature-induced secondary currents and their impacts develop over a stretch in the bend, and the bed load movements in the channel bend is not a function of the bed shear stress magnitude but rather of the curvature-induced secondary currents which push and keep the sediments near the inner wall. As a result, deeper invert abrasions are formed toward that wall at SBT bends and further downstream.
This study contributes to the available knowledge of three-dimensional narrow channel flows and the interactions between hydraulic and hydrodynamic phenomena associated with high-speed sediment-laden flows, particularly for aspect ratios ≤ 2. The produced open-source CFD models are useful in computing complex three-dimensional supercritical flow characteristics in uniform narrow channels. In addition, the obtained experimental findings enlighten the hydrodynamics behind the abrasion patterns observed at SBT bends and generate data useful in validating numerical models in the future. Further advancements in computational resources in the years ahead are expected to make the alternate and more accurate transient CFD techniques feasible for supercritical flows in straight and curved channels. | en_US |