Reference : Hydrodynamics and bed stability at smooth-to-rough transitions. Experiments based on ...
Dissertations and theses : Doctoral thesis
Engineering, computing & technology : Civil engineering
Hydrodynamics and bed stability at smooth-to-rough transitions. Experiments based on acoustic flow measurements
Duma, Diana mailto [Université de Liège > Département ArGEnCo > Hydraulics in Environmental and Civil Engineering >]
Université de Liège, ​Liège, ​​Belgique
Docteur en Sciences de l'Ingénieur
ix, 210 + 6
Dewals, Benjamin mailto
Nguyen, Frédéric mailto
Erpicum, Sébastien mailto
Pirotton, Michel mailto
Bung, Daniel B. mailto
Peltier, Yann mailto
[en] As reflected in the scale of large contemporary hydropower schemes, flowing water may carry huge amounts of energy. If not mastered properly, this energy may cause impressive channel erosion and local scour. Therefore, one lasting concern in hydraulic engineering has been the stability of structures, riverbeds and riverbanks against flow erosion.
Two main limitations of current understanding of riverbed and riverbank stability are highlighted: the first one is directly linked to the evaluation of the flow action, while the second one relates to the conceptual framework in which stone stability is evaluated.
Since the standard approaches use the mean bed shear stress to quantify the flow forces, they may only apply under uniform flow conditions, for which the ratio of turbulence intensity to the bed shear stress remains almost constant and the influence of turbulence is therefore implicitly incorporated. For non-uniform flow, correction factors have been conventionally applied to account for the turbulence fluctuations; but this approach does not reproduce the physical influence of turbulence higher in the water column and can only be used as a rule of thumb.
A new approach was initiated recently in literature. Instead of using the standard Shields parameter, it quantifies the flow forces by means of a new set of parameters which combine explicitly the velocity and turbulence distributions over a certain water depth above the riverbed, while remaining reasonably accessible for engineering applications. Next, this quantity is related to a mobility parameter, which describes the bed damage. This new approach requires additional high quality experimental data to confirm its validity for a wider range of non-uniform flow.
In this research, the focus was set on a single canonical configuration, namely a smooth-to-rough transition. We take a preventive perspective by focusing on the flow and bed stability conditions before a scour hole starts to develop and we are interested in characterising the flow conditions to ensure bed stability, i.e. prevent the dislodging of bed material downstream of the structure.
As a second specific objective of the thesis, we aimed at evaluating the feasibility of using acoustic techniques to properly estimate the new bed stability parameters proposed in literature in the last decade and draw conclusions on stone mobility at smooth-to-rough transitions.
The experimental tests were conducted in two laboratory flumes, a horizontal bottom flume (6 m long and 15 cm wide) and a tilting flume (up to 4% slope, 20 m long and 50 cm wide), in which we measured the flow velocity at a 100 Hz frequency, using two different acoustic instruments: an Ultrasound Velocity Profiler (UVP) and a 3D Acoustic Doppler Velocity Profiler (ADVP). The measurements were done immediately downstream of the transition from smooth-to-rough. Two types of tests were undertaken. In the first type, the stones were glued with silicone on the flume bottom and velocity measurements were performed without stone motion. This enables measurements of the flow characteristics both below and above the threshold for inception of sediment motion, without perturbations induced by stone displacements. In the second type, the stones in the measurement area were laid on the bottom of the flume (i.e. not glued) and the number of entrained stones was recorded. The tests were conducted by varying (i) the grain size of the bed material (8 mm, 15 mm and 30 mm), (ii) the flow velocity (between 0.74 m/s and 1.16 m/s), (iii) the sediments density (1410 kg/m³, 1690 kg/m³ and 2650 kg/m³) and (iv) the configuration (quasi-uniform vs. smooth-to-rough transition). Indeed, for the purpose of comparison, a uniform rough bed (quasi-uniform configuration) was considered also, by replacing the smooth part of the bed with similar sediments as in the measurement area. After a specific signal processing, flow variables were derived from the measurements and were next exploited to evaluate several bed stability parameters, with the aim of correlating them with bed damage data obtained from dedicated experiments.
The results showed no correlation between the bed damage and the bed shear stress or between the bed damage and flow turbulent kinetic energy. In contrast, when the flow action is described by both mean velocity and turbulence characteristics, a relative good correlation with the bed damage was observed. Nonetheless, these correlations appear only per subset of points corresponding to the same grain size and/or same material density or flume geometry.
With a total of 45 hydraulic configurations for which the flow characteristics were measured and 66 (times four repetitions) tests in which the bed damage was observed, the present research ends up with a unique dataset, which may prove useful in various research such as for the validation of 2D-vertical and 3D turbulent flow and morphodynamic simulations.
Another key outcome of the present doctoral research is a systematic comparison between measurements conducted with the UVP and the ADVP. The ADVP is deemed generally more accurate and reliable. Nonetheless, both instruments remain somehow complementary. We showed that, under certain conditions, the considered instruments have the potential to contribute to the assessment of bed stability in the considered configurations.

There is no file associated with this reference.

Bookmark and Share SFX Query

All documents in ORBi are protected by a user license.