Investigation of turbulent mixing layer flow in a vertical water channel by particle image velocimetry (PIV)

Abstract

Turbulent mixing layer flow in a vertical water channel was experimentally investigated by particle image velocimetry (PIV). The mixing layer is produced by a specially designed insert plate placed in the channel with a low-and high-speed side velocity ratio of 0.25. The Reynolds number based on the velocity difference of two streams and the spanwise vorticity thickness at the place where the mixing layer start merging ranges from 2184 to 14 672. The results show that there are large coherent vortex structures near the centreline of the mixing layer. Both instantaneous kinetic energy and spanwise vorticity always concentrate at the location where the coherent structures connect or meet each other. The normalised dimensionless Reynolds stresses and average spanwise vorticity show self-similar, respectively, under different Reynolds numbers at the same cross-section in the down streamwise direction. Every component of Reynolds stresses increases but the vorticity decreases with the downstream distance. For all Reynolds number, the peak values of mean vorticity in the streamwise direction appear the same decay speed. The splitter plane wake causes a negative peak of the mean vorticity where the mixing layer merges. The negative peak values of vorticity increase with the Reynolds number. The dimensionless negative peak values decrease exponentially with Reynolds number and reach a constant when the Reynolds number is large enough.Le débit de couche de mélange turbulent dans une conduite d'eau verticale aété analysé de façon expérimentale par vélocimétrie par images de particules (PIV). La couche de mélange est produiteà l'aide d'une plaque d'insertion de conception particulière placée dans le conduit avec un rapport de vitesse latérale basse etélevée de 0,25. Le nombre de Reynolds fondé sur la différence de vitesse de deux courants et l'épaisseur du tourbillon dans le sens de l'envergureà l'endroit où la couche de mélange commenceà fusionner se situe entre 2,184∼14,672. Les résultats indiquent qu'il existe de grandes nappes de tourbillons cohérentes près de la ligne centrale de la couche de mélange.À la fois l'énergie cinétique instantanée et le tourbillon dans le sens de l'envergure se concentrent toujoursà l'endroit où les nappes cohérentes se rencontrent. Les tensions de Reynolds normalisées adimensionnelles et le tourbillon dans le sens de l'envergure moyen démontrent qu'ils sont autosimilaires respectivement sous différents nombres de Reynoldsà la même section transversale en aval. Chaqueélément des tensions de Reynolds s'accroît, mais le tourbillon diminue avec la distance en aval. Pour tout nombre de Reynolds, les valeurs de crête du tourbillon moyen dans la direction du flux semblentêtre de la même vitesse de désintégration. Le sillage plan de séparation cause une crête négative de tourbillon moyen lorsque la couche de mélange est fusionnée. Les valeurs de crête négatives du tourbillon augmentent avec le nombre de Reynolds. Les valeurs de crête négatives adimensionnelles diminuent de façon exponentielle avec ...

Vertical mixing layer development

“…In most studies of turbulent mixing layers [5,14,31], the results are non-dimensionalised or normalised, typically using an average velocity and the initial roller diameter or splitter plate boundary layer thickness. This is because in the conditions tested by others, there has been a constant stream of regular rollers.…”

“…For example, Azim and Islam [7] test at similar velocity ratios to those present in a current and wave testing tank, but use air as a medium as well as having significant differences in splitter plate geometry and free stream Reynolds numbers. An example of a water flume is given by Guo et al [5], who investigate flows with far higher velocity differences than those tested here. The high sensitivity of turbulent mixing layers to physical conditions and geometry, along with there being no previous studies with similar conditions to those of interest here mean that it is prudent to perform a new experiment.…”

“…Turbulent mixing layers are a significant source of aero-acoustic noise and understanding them is of importance to aerospace design [4]. Mixing layers are also present in chemical plants and are critical to increase mixing performance [5,6]. This mixing situation also lends itself to pure fluid dynamic study and model validation due to the absence of bounding walls and rapid but measurable growth of turbulent structures [7].…”

“…PIV is capable of measuring velocities over a large area instantaneously and therefore aiding the investigation of turbulent mixing layers. Guo et al [5] used a 2D PIV system to produce instantaneous vector and contour velocity plots in which a series of rollers can be seen. Along with the visualisation of the rollers, Guo et al [5] use PIV to provide an efficient method of obtaining distributions of Reynolds stresses and other flow quantities.…”

“…Guo et al [5] highlighted that surface waves affected the development of the mixing layer. In these experiments the free surface was lifted slightly (<2 mm) near the inlet but no significant surface waves were observed.…”