Velocity and sediment transport in a physical scale model of a longitudinal training dam
High and low discharge extremes in the Dutch river systems are a continuous concern of the Dutch state authority for infrastructure and environment (Rijkswaterstaat). To manage these extremes in discharge, new river training structures are investigated to secure the hinterland from flooding, and to guarantee sufficient depth for navigation. One of these new structures is a longitudinal training dam (Vermeulen et al., in press). A physical scale model with a mobile bed is built in the Kraijenhoff van de Leur Laboratory for Water and Sediment Dynamics at Wageningen UR to investigate the morphological impact of the construction. The ultimate aim of this study is to establish the extent to which the physical scale model can be used to predict velocities and sediment transport rates in the prototype situation of the River Waal. Particular attention is given to the methods of flow measuring techniques under laboratory conditions. Velocity in the scale model is measured using an Acoustic Doppler Velocimeter (ADV), which operates using the Doppler shift of a transmitted acoustic signal. This signal is reflected by fine suspended particles, which may be insufficiently present in tap water. Seeding of extra particles is needed to increase the velocity signal quality, decreasing the number of spikes related to the absence of scatterers. Spike detection and removal is performed following the phase space threshold techniques introduced by Goring and Nikora (2002), which were discussed by Wahl (2003). The influence of seeding rates on mean velocities and turbulence stresses is investigated in detail. Results show a large difference between experiments with or without seeding, but the seeding concentration has only a small influence on mean flow velocity and turbulence quantities. Once some of the seeding material is in the system, additional seeding only marginally affects the averaging time necessary to reach a stable mean velocity or covariance value. The physical scale model predicts flow velocity in the prototype satisfactorily, despite irregularities in the first half of the flume. Contraction of flow lines and the concomitant bed response are conform expectations, and similar to prototype values. To monitor sediment transport rates, a bedform tracking method was tested and compared with a more direct way of measuring sediment transport, by diverting the return flux of suspended sediment to a purpose-built measurement reservoir. Results of alternative the measurements were nearly the same. Sediment distributions are found to be in accordance with the sediment transport distributions of Hamamori (1962), and the Shields parameters are found to be conform prototype values.