A time accurate computational analysis of two-dimensional wakes
thesisposted on 16.12.2009, 10:41 by William Paul Bennett
The unsteady transport of large-scale coherent vortices can induce a redistribution in the stagnation temperature and pressure relative to the free stream flow. The time averaged result of this redistribution is the Eckert-Weise effect, by which a cooled region is defined along the wake centre. The time accurate characteristics of this mechanism for bluff body near wake flows are, however, sparsely documented at low transonic Mach numbers. For example, no available published research has, to date, studied the time resolved energy separation characteristics in the transonic near wake flow of a circular cylinder using a time accurate numerical model. A novel time accurate computational analysis is developed of the near wake energy separation characteristics downstream of a circular cylinder in a low transonic crossflow at high Reynolds number. This circular cylinder analysis is extended to a novel time accurate computational study of energy separation in an asymmetric turbine cascade wake at a low transonic exit Mach number. Energy separation is reported to primarily be a convective flow effect. A structured inviscid and turbulent test program examines the extent to which an inviscid model is able to predict energy separation. Results from this study indicate a good correlation of the time accurate and time mean flow statistics with published work. These results demonstrate that an inviscid model is able to capture the basic energy separation mechanism. However, inviscid models are shown to over-predict the stagnation temperature and pressure redistribution. The inviscid prediction suggests that air compressibility modifies the incompressible energy separation mechanism. Turbulence diffusion reduces the stagnation temperature and pressure extrema to demonstrate a better comparison with experimental data. A relationship between the energy separation and vortex strength is highlighted. This is shown in the turbine cascade prediction to be dependent on the boundary layer separation characteristics.