Morphological modelling with suspended sediment transport in river bends

Depth-averaged morphological computations using Delft3D tend to produce ‘strange’, results for river bends when the sediment transport formulas of Van Rijn are used. Even pools in inner bend regions and bars in outer bend regions can result from the model. The objective of the present thesis is to identify the cause of the problem and propose potential solutions. A short analysis shows that the ‘strange’ predictions are not connected to the Van Rijn’s formulations, but to suspended sediment transport modelling in general. An axi-symmetric solution (viz. the equilibrium solution of the transverse slope in an infinitely long bend with constant width and curvature) is derived for the case where suspended sediment transport is included, based on a formulation for the bed load transport by Olesen (1987). It is indicated that the limitations of the model including suspended sediment transport, along with bad modelling practices, can produce unrealistic results. Delft3D implicitly models the influence of secondary flow on the suspended sediment transport with the help of a mixing coefficient. Numerical experiments show that this is not possible in a river bend context, unless calibrating the parameters in such a way that they would lose their physical meaning. In order to overcome this problem, it is concluded that the influence of secondary flow on the suspended sediment has to be modelled explicitly. A suspended sediment transport model is designed, which includes the influence of secondary flow due to the variation in vertical direction of the profiles of secondary flow and suspended sediment. The depth averaged expressions including the dispersion of suspended sediment by the main and the secondary flow are derived by using analytical similarity profiles of sediment concentration (0th order Rouse profile) and secondary flow (Kalkwijk & Booij, 1986) for closing the extra terms. These expressions are implemented in Delft3D and the resulting code is used to carry out numerical simulatoons for hypothetical and real test cases including flume experiments. The analysis is confirmed by the simulations and the results of the improved depth-averaged code are much more realistic (from a mathematical point of view) as they show, especially for the hypothetical test cases, good agreement with the results from 3D computations (using also Delft3D). On the other hand, the applicability of the model to cases with large curvatures is more problematic, unless a more accurate expression for the secondary flow is used. It is concluded that incorporating the influence of secondary flow on the direction of suspended sediment along with the option of using a distribution function with variable shape is highly beneficial for morphological modelling in river bends. Another alternative is indicated to be the appropriate calibration of the morphological model. This is a less viable option, as it could even lead on disconnecting parameters from their physical meaning.Coastal EngineeringHydraulic EngineeringCivil Engineering and Geoscience