Heat transfer and vortex shedding effects in single- and multi-scale grid-generated turbulence

Abstract

In this thesis the potential of single- and multi-scale turbulence-generating grids in improving the heat transfer from a circular cylinder and from a flat plate is experimentally explored. Hot-wire and heat transfer measurements downstream of low-blockage grids with different geometries were performed in a wind tunnel. Particular emphasis is given to the impact of the vortex shedding from the bars of the grids both on the turbulent flow field and on the heat transfer. The effects of vortex shedding suppression in grid-generated turbulence can be summarised in: (i) a decrease of turbulence intensity Tu in the grid's production region; (ii) a downstream shift of the peak of turbulence intensity; (iii) an attenuation of the streamwise growth of the integral length scale L_u together with a reduction of the ratio between L_u and the Taylor length scale; and (iv) an increase of the skewness and the flatness of the streamwise turbulent fluctuations in the turbulence production region. A fractal square grid (FSG17) exhibits vortex shedding suppression properties in analogy to fractal plates and fractal trailing edges. It is discovered that an intense vortex shedding reduces the heat transfer in the laminar boundary layer region of a cylinder placed in the grid's turbulence production region. For this reason, if the shedding is prominent, the heat transfer coefficients in the production region are lower than those measured, for the same Tu, in the turbulence decay region where the vortex shedding signature tends to vanish. The heat transfer around the cylinder is enhanced more and is more persistent downstream by using FSG17 or a single square grid (SSG) in place of a regular square-mesh grid (RG60) with higher blockage. Grid FSG17 has the practical advantage of allowing high heat transfer together with a less energetic vortex shedding from the grid. For the case of a flat plate, it is found that a new class of multi-scale inhomogeneous grids (MIGs), which produce a gradient of mean velocity in the direction normal to the plate, can be optimised for heat transfer augmentation. These grids can allow a near-wall high mean velocity and high turbulence intensity, thus leading to sustained heat transfer enhancement with downstream distance.