Mechanical Engineering - Theses

Permanent URI for this collection

Search Results

Now showing 1 - 2 of 2
  • Item
    Thumbnail Image
    Effects of rotation on turbulent boundary layers and wakes
    Witt, Henrik T. ( 1986)
    This work describes experiments of two simple flow cases with imposed system rotation. The first flow is a boundary layer that develops on the flat vertical side-wall of the working section. A trip-wire causes transition, so that the flow in the boundary layer is fully turbulent. The opposite side-wall has been adjusted so that the pressure gradient is zero. The axis of rotation is parallel to the span of the vertical side-wall. The rotation affects the boundary layers; depending on the direction of rotation the changes can be seen as caused by a stabilisation or destabilisation of the flow. A number of mean flow and turbulence data measured with dynamically calibrated hot-wires are presented, both measured as ordinary boundary-layer profiles and, in the presence of 3-dimensionalities, as contour plots. Similarities between the stabilising/destabilising effects of rotation and density stratification, and also curvature, are pointed out. The second flow is the turbulent wake behind a circular cylinder. Each model used in the study is suspended in the middle of the free stream of the working section, with its axis parallel to the axis of rotation. Since the effects of rotation are found to be negligible close to the cylinder, the wake has been followed for several hundred diameters downstream. There the two initially symmetric sides of the wake have grown asymmetric. The observed effects of rotation are explained in terms of transport equations and alternatively by a qualitative model based on assumed structures underlying the turbulence and their behaviour when subject to rotation.
  • Item
    Thumbnail Image
    The structure of turbulent pipe flow
    HENBEST, SIMON ( 1983)
    A brief literature survey of the modelling of wall turbulence is presented. In addition, a recent development in the modelling of wall turbulence (Perry & Chong 1982) is annotated. The structure of fully-developed turbulent flow in a smooth-walled pipe and in a rough-walled pipe are investigated experimentally over the Reynolds number range of 75,000 to 200,000. Each pipe had a nominal length of 400 diameters. The turbulence intensity results were measured using dynamically calibrated normal and crossed-wires. A dynamic calibration mechanism was built that enabled the wires to be calibrated inside the test section of either pipe. The longitudinal and transverse broad-band turbulence intensity results and their spectral distributions, in the turbulent wall region of both the smooth-walled and rough-walled pipes, show encouraging support for Townsend’s (1976) attached-eddy hypothesis and Perry & Chong’s (1982) model of wall turbulence. Using the spectral similarity laws proposed, the asymptotic longitudinal broad-band turbulence intensity distribution in the turbulent and viscous wall regions of the smooth-walled pipe are predicted with reasonable accuracy. In the core region of both pipes, the longitudinal and transverse spectral results show support for the existence of an inertial-subrange and hence of “locally isotropic” small scale motions. In the smooth-walled pipe, Townsend’s (1956) Reynolds number similarity hypothesis is found to be valid in the asymptotic limit with Reynolds number, while in the rough-walled pipe, the extended version of Townsend’s Reynolds number similarity hypothesis (Perry & Abell, 1977) is not upheld, at least for the results presented in this thesis.