| This thesis is an experimental study of turbulent flows, consisting of two parts. The first part is an investigation of the highly three-dimensional flow structure around a simplified vehicle model. The lasting high fuel cost has inspired resurgence in drag reduction research for vehicles, which calls for a thorough understanding of the vehicle wake. This investigation aims to revisit turbulent flow structure behind an Ahmed vehicle model. Two rear slant angles, i.e., α = 25° and 35°, of the model were examined, representing two distinct flow regimes. The measured velocity and vorticity data reconfirm the well documented major characteristics of the mean flow structure. However, both instantaneous and time-averaged PIV data unveil a number of important aspects of the flow structure, which have not been captured previously due to a variety of reasons. As such, a considerably modified flow structure model is proposed. The second part of the thesis is a detailed study of the extraordinarily complicated turbulence structure in a cylinder-on-cone cyclone. The turbulent flow within a cylinder-on-cone cyclone is highly three-dimensional and complicated; our knowledge of this flow has yet to be improved. Two separate efforts have been made. Firstly, measurements were performed at a fixed swirl number of 3 with the Reynolds number varying from 37,100 to 74,200. Two dominant frequencies of vortical structures were identified based on LDA-measured tangential and axial velocity spectra. One of them agreed quite well with previous reports, but the other was observed for the first time. One explanation was proposed. Secondly, special attention was turned to the swirl number effect on the flow. The swirl number examined was from 2.4 to 5.3, with the Reynolds number fixed at 7.4 x 104. Three regions of the flow have been identified after careful analysis of the data, which are referred to as the core, the outer and the wall-affected regions, respectively. It has been found that the flow, including its Strouhal numbers and radial distributions of the mean and fluctuating velocities, depends considerably on the swirl number. |
