Graduation of Yannick Steenman

16 november 2020 13:30 t/m 14:30 - Locatie: TBA - Door: Webredactie

Numerical modelling of shallow jet flows

  • Professor of graduation:   Dr. ir. R.J. (Robert Jan) Labeur
  • Supervisors of graduation: Dr. ir. B.C. (Bram) van Prooijen (TU Delft),  dr. ir. S. (Sierd) de Vries (TU Delft), dr. ir. Y.B. (Yorick) Broekema (Deltares), ir. Y. (Ype) Attema (Svasek Hydraulics)

This work aims to gain information regarding numerical representations of shallow jet flows in both laboratory and field applications. The primary focus lies in the differences between 2DH (two-dimensional horizontal) numerical models and 3D models and therewith the effect of the three-dimensionality of the flow. Numerical models were set-up for both a laboratory experiment investigating a shallow jet over a longitudinal slope and for the field application at Waterdunen. It was found that in a controlled environment, proper reproduction of the dissipative processes such as wall friction are crucial for the reproduction of the asymmetric pattern of the flow. As a result, recirculating velocities and the shape of the recirculation zones were modelled incorrectly. Furthermore, due to the assumed vertical log-profile of the velocity, horizontal streamline contractions were shown to solely be determined by variations in the water-depth due to adaptations of the log-profile. In the laboratory data, the vertical velocity profile was shown to deviate from the standard log-profile, indicating an inaccuracy in the 2DH model. In 3D models, the easier parameterization of wall friction allows for a proper representation of the flow asymmetry. In both the 3D model and the laboratory observations it was shown the streamline contraction persisted after the changes in water-depth were passed. It was speculated the horizontal streamline contraction was directly correlated to the vertical velocity profile, meaning a proper representation of the vertical velocity profile would yield an accurate reproduction of the streamline contraction. 

When considering a field scenario, the differences between 2DH and 3D models were less pronounced. The asymmetry of the flow was reproduced nearly equally by a 2DH and 3D model. It was speculated the better reproduction of the asymmetric pattern was the result of the more complex geometry, topography and inflow boundary condition. Whereas in a controlled environment the asymmetry is the result of subtle instabilities, in the field these instabilities are much more pronounced. As a result of the better reproduction of the asymmetry, other hydrodynamic processes were better reproduced as well. However, similarly to the laboratory models, the horizontal streamline contraction was shown to solely be a function of the water depth changes rather than the changes in the vertical velocity profile. 

It was concluded both a 2DH and a 3D numerical model are capable of reproducing most hydrodynamic processes in shallow jet flows. However, if an asymmetric flow pattern can be expected, a 3D model might be required to properly simulate the asymmetry. Furthermore, it can be expected the 2DH model simulated an incorrect vertical velocity profile on longitudinal topographic changes and downstream thereof.