The printable version is no longer supported and may have rendering errors. Please update your browser bookmarks and please use the default browser print function instead.

Cylinder-wall junction flow

Semi-confined Flows

Underlying Flow Regime 3-35

Abstract

Cylinder-wall junction flow is an example for a flow around bluff bodies mounted on a flat plate. Such flows appear in various technical applications, e.g. wing-body junctions, or environmental flows, e.g. flow around a bridge pier. The flow dynamics in such situations are often dominated by a so-called horseshoe vortex system establishing in front of the obstacle and wrapping around it. It has been demonstrated that this system undergoes a complicated dynamics including bi-modal velocity distributions and large levels of turbulent fluctuations around the horseshoe vortex center. The horseshoe vortex is linked to strongly enhanced wall shear stresses in front and around the cylinder which e.g. can give rise to scouring of bridge piers.

This study follows a dual approach using a numerical as well as an experimental approach. We applied a highly resolved large eddy simulation as well as a particle image velocity experiment to the flow around a cylinder mounted vertically on a flat rigid plate. Both data sets refer to the same set-up, but were acquired independently and are made accessible in files at the end of the Evaluation section of this document. The data for the inflow condition can be found at the end of the Test Case Studies section of this document, too.

The results presented show the time-averaged streamlines visualizing the main flow structure. The distribution of the c-shaped turbulent kinetic energy and its budget terms such as mean convection, production, diffusive transport and dissipation are shown as well. Furthermore, selected profiles of the velocity components and the Reynolds stresses, the pressure coefficient and the friction coefficient are presented.

In general, the numerical and experimental results do agree with each other. However, slight deviations are visible such as for the time-averaged position of the vortex system. To ease the comparison of the velocity profiles at the same positions relative to the vortex system, we introduced an adjusted horizontal coordinate. The data presented below were evaluated at distinct horizontal positions in this adjusted coordinate system and are thus located at the same positions relative to the vortex center in both numerical and experimental datasets.