Numerical simulation of bedform effect on flow structure in shallow rivers

On way to understand the behavior of rivers is to study of flow structures and bedforms. Most rivers have rough beds, that are called bedforms. These shapes have different types depending on the hydraulic conditions that cause the resistance to flow, so sufficient knowledge of the hydraulic resistance of the flow is necessary to calculate discharge, depth, water velocity, flood prediction and sediment transport to reduce the damages that are caused by rivers. Despite years of research and experimentation on bedforms, there is still no adequate and accurate equation to predict the geometry of bedforms and their interaction with flow. The RANS turbulence model is not sufficiently accurate in detecting flow separation in high-altitude and high-angel lee side dunes, but it is more appropriate in other dunes. The LES turbulence model can be more appropriate in detecting flow separation in the vicinity of the bed. But in the layers away from the dune, it produces relatively severe turbulences. The DES turbulence model is more accurate in detecting the flow separation in large-scale dunes and has less execution time than the LES, but this model produces irregular vortices in smaller dunes near the bed. In this research, with the aim of investigating the effect of dune geometry on flow structure, numerical simulation of flow motion on dunes in open channel duct was investigated. In this regard, 29 simulations were performed to study the effect of the geometry of five types of dune with different angles and heights in different hydraulic conditions and different bed roughnesses with RANS and DES turbulence models. The STAR-CCM+  was used in order to simulate the numerical model in this research. This software provides highly realistic results by providing a seamless environment with high network production capability and extensive simulation tools, which helps professionals in the process of working with fluids.In this research, VOF method is used to calculate free surface area, and the solution method is Implicit Unsteady. In order to sensitize the numerical model to the number of computational cells as well as to optimize the runtime and output accuracy, several blocks of lattice dimensionality change in data points and dune locations were used. To select the appropriate networking dimensions, 8 models were run to optimize the time and accuracy of the model. To evaluate the accuracy of the simulations, the numerical model results were compared with those of previous researchers. The results of comparison of the numerical model and the experimental model showed that only about 9.5%, 15.5%, 14.5%, 9.4%, 12.2% and 7.4% were different. This error rate indicates good numerical model accuracy. Also the results of numerical model + STAR-CCM were compared with SSIIM numerical model. then, by using the dimensional analysis, effective factors on the interaction of dune geometry and flow structure were investigated and finally, a formula was developed to predict this interaction. The results of the evaluation of the obtained formula for investigating the effect of dune geometry on the hydraulic flow showed that the presented formula with error of 11.25% and 0.86 R2, due to the completely random nature of bedform formation, is very accurate.