Effect of bed roughness on the hydrodynamic properties of dam-break waves

Abstract

A dam-break wave can result in considerable damages and casualties. Events that can lead to such waves include dike and dam breaches, storm surges and impulse waves. Dutch history is familiar with a number of coastal dike breaches but also river dike breaches have occurred, resulting in dam-break waves reaching far inland. Beyond the Netherlands, propagating walls of water could occur when dams of water reservoirs collapse. More specific knowledge of the hydrodynamic behaviour of dam-break waves is necessary in the battle against water.
The impact of a dam-break wave can be reduced through a better understanding of the currently poorly understood link between bed roughness and wave hydrodynamic properties. The objective of the present study is therefore to study experimentally the effect of bed roughness on the hydrodynamic properties of dam-break waves, on both dry and wet beds. The approach was to generate dam-break waves through a lift-gate and a reservoir with a depth d0 = 0.4 m and to measure the water levels over time in the downstream channel. 7 different bed roughness configurations were made (range of ks = 0.6 to 48 mm) in addition to smooth bed reference test. Dry bed tests were performed, 5 wet bed tests with an initially still water level (h0) (range of h0/d0 = 0 to 0.125) were performed. The analysis focused on wavefront celerity, inundation depths and roller length which are used to characterise the hydrodynamic properties.
Visual observations showed substantial differences in wavefront behaviour between surges and bores as well as between smooth and rough beds. Validation of the generated waves was done by successfully comparing the reconstructed water levels over the flume length with a new approach.
Results showed that an increase in bed roughness resulted in a decrease in wavefront celerity for surges and that the initially still water level can work as a lubricant for bores. The initial still water level is dominant over the bed roughness when h0 >0.5ks.
The maximum water level increased with increasing bed roughness for surges, while bores showed an opposite behaviour. Increasing the initially still water level reduced the relative maximum water level for rough bed configurations.
In the present study, only some test configurations revealed the presence of a plateau height. If there was one, an increase in bed roughness showed a decrease in plateau height for bores with h0/d0 = 0.125. Three sections in the h0/d0 range with different plateau height behaviours were noticed.
The roller lengths were determined with the wavefront celerity and the plateau height. No trivial relations were found for the influence of bed roughness or initial still water level.
Overall, this project revealed some interesting behaviours that need to be further investigated in the future. It is recommended to conduct a full quantitative analysis once precise values of the bed roughness configurations are determined. Nevertheless, these results present some preliminary, yet interesting, findings that will contribute to a better understanding of the behaviour of dam-break waves over rough bed.