Exploring Appropriate CFD Model and Impact Scale for Non-submerged Spur Dikes

Spur dike is a common hydraulic structure, which is widely used in hydraulic engineering such as channel regulation, flood prevention and river diversion for preserving the desired water depth, deflecting the main current and protecting river banks. In actual operations, spur dikes exert influence on river system usually in the form of groups, and the flow field around spur dike group is complicated and not well understood yet [1-2]. Compared with experimental investigation, numerical simulation can provide more details with reasonable cost. However, it still depends on observed data as well as appropriate choice of turbulence models and boundary conditions [3-4]. In this study, firstly, the flow field around non-submerged spur dike groups which were eight spur dikes in staggered arrangement was selected as the research objects for understanding the performance of the CFD models of spur dike group flow based on flume experiments [5-6]. A set of flume experiment was conducted in laboratory, based on its observed data the numerical simulations were carried out through Finite Volume Method (FVM) and three turbulence models-standard k-ε model, Reynolds Stress Model (RSM) and Large Eddy Simulation (LES). In each numerical model, the free surface boundary respectively employed two methods-rigid-lid assumption and VOF (volume of fluid) model. The comparisons between computation results from CFD and observed data from flume experiment showed that the three turbulence models can simulate the three-dimensional flow around spur dike group in a certain extent. Specifically, the simulation effect of standard k-ε model and RSM has no significant difference under either rigid lid assumption or VOF method, both of them can preferably simulate flow field of mainstream area, while the simulation performance of recirculation zone near the spur dikes is poorer relatively; in contrast, the simulation effect of LES is best especially in the region of recirculation. For computation time, k-ε model is shortest while LES is longest, and the gap between the two is huge. Therefore, to rapidly evaluate the field characteristics of mainstream, standard k-ε model with rigid lid assumption is recommended; while to achieve the fine simulation of spur dike field, especially the flow pattern in recirculation zone, LES with VOF method is appropriate. On the basis of the previous, numerical models based on standard k-ε model, finite volume method and rigid lid assumption were built to classify the impact scale of non-submerged spur dikes on river systems. The conceptions of spacing threshold of non-submerged double …