Experimental characterization of turbulance in steep rough streams
Torrential flows have so far been studied less than have river flows, due to the difficulty of faithfully reproducing them in the laboratory, and the few field measurements available. For this reason, this present research tries to shed some light on the turbulent description of this type of flows,...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis doctoral |
| Fecha de publicación: | 2019 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/128856 |
| Acceso en línea: | https://hdl.handle.net/2117/128856 https://dx.doi.org/10.5821/dissertation-2117-128856 |
| Access Level: | acceso abierto |
| Palabra clave: | Àrees temàtiques de la UPC::Enginyeria civil |
| Sumario: | Torrential flows have so far been studied less than have river flows, due to the difficulty of faithfully reproducing them in the laboratory, and the few field measurements available. For this reason, this present research tries to shed some light on the turbulent description of this type of flows, evaluating the effects posed by the energy slope and the roughness of the bed. With this study we have made a significant contribution to the knowledge of how turbulent properties develop along the profile of the depth of the flow and thus improve all the engineering aspects that affect the high mountain flows (<<D50/d), from the valuation of the energy losses produced in the rough layer until the beginning of movement of the particles. For this purpose, different flows have been characterized by the description of classic turbulent parameters (turbulent intensity, turbulent kinetic energy, Reynolds shear stress, dissipation, length scales, etc.). In addition, a detailed analysis of quadrants has been carried out due to the implications of bursting turbulence at the beginning of particle movement. Some of the results obtained from the analysis of quadrants have shown that the turbulence does not happen randomly, showing certain temporal and spatial patterns that can be of great help in the development of generation models of synthetic velocity series. The analysis of quadrants has been complemented with an analysis of the pulses, promoted for the temporary coherence observed in the quadrant sequences. One pulse is defined as the uninterrupted time that an event u'w' remains in the same quadrant. This analysis yields very interesting results since they show a certain proportionality in both quantity and duration of turbulent pulses, as if it were a condition to be fulfilled by all flows. Many authors have shown that the impulse applied (force x time application) is responsible for the initial movement of the particles rather than only the magnitude of the applied force. This idea has been the point of departure for the realization of a conceptual model to assess the rate of transport of sediment of a flow. This model is based on the consideration that Reynolds shear stress has the most important role in the beginning of the bed material’s movement. Therefore, the time in which the bed shear stress applies in the same direction defines the intensity of the impulses applied to a surface. This last hypothesis together with the concept of pulses and pulse sequence are the bases for a methodology for assessing the transport rate of a flow and a specific sediment. The conceptual model has been evaluated by comparison with the transport rates measured in the laboratory. To carry out this study, different laboratory experiments were performed in the Laboratory of Fluvial Morphodynamics II "The Cube" of the GITS-UPC research group. The experiments are divided into two groups; the first encompasses the RG tests (Rounded Gravel) and second group the CG tests (Crushed Gravel). The first group consists of a total of 10 experiments, all with the same bed roughness, formed with rounded gravel (D50=55-mm y σ=16-mm), but with 10 different discharges, which allows the evaluation of the effects of the flow rate over turbulent variables. In the second group of experiments (CG), crushed gravel (not rounded) is used to form the test bed. Up to three different diameters are used (D50= 17.8-mm, D50= 30.0-mm, and D50= 51.1-mm) that give rise to three different experiments. In addition, each of these materials is tested twice, the first testing the material simply poured and the second smoothing the surface as much as possible to reduce relative roughness, which makes a total of 6 experiments. In this second group the data collection is made once the start of the movement of the particles has been reached. The instantaneous velocity at each point of the profile was measured using an Acoustic Doppler Velocimeter (ADV) to obtain time series of instantaneous velocity in three dimensions (X, Y, Z). The measurements obtained by ADV in highly turbulent flows are linked to certain uncertainties such as the presence of spikes, low SNR signal and low correlation data. The low correlation favors a high percentage of data filtered through the most common filtering limits (COR <70), obtaining up to 50% of erroneous data in some areas near the bed. The replacement of the low correlation data subtracts turbulent energy to the series of velocity with the consequent modification of the turbulent parameters, distancing it from its natural turbulent behavior. In addition, the elimination of data of low correlation does not avoid the need to apply other filters to eliminate the spikes produced by aliasing. For this reason, an in-depth study has been carried out on the quality of the data, using analysis of quadrants and pulses in addition to other turbulent parameters, finding that low correlation data have a behavior similar to high correlation data. On the other hand, the need for a broader knowledge of the effects produced by the ADV configuration on the characterization of the turbulence studied throughout the thesis motivated an additional study. These analyses related to the reliability of the data have been included as an appendix since they involve all the concepts analyzed during the execution of the thesis together. Since the checks have been made recursively throughout the study, to include these works within the body of the thesis would lead to confusion. The reader will be indicated at certain points of the thesis to consult these appendices |
|---|