The flow dynamics of a rectangular cavity with a thick incoming boundary layer at low Mach numbers is investigated by Large Eddy Simulation (LES) and parallel CFD, as a simplified model of automobile bodywork recesses. The cavity inflow is generated by means of synthetic stochastic time-dependent methods in a precursor simulation, in order to identify and analyze quantitatively the streaks in the spatially developing boundary layer approaching the cavity. In the cavity flow model, no self-sustained oscillation is found, due to the high value of the boundary layer thickness. The influence of the approaching boundary layer turbulent scales on the cavity instabilities is examined. The intermittent cavity flow behaviour is related to the injection and ejection of vortex structures across the cavity opening and downstream edge. The space and time resolution of the LES enables to identify the flow dynamics of vortical instabilities and of the three-dimensional structures in the cavity shear layer. Cavity noise sources are identified by correlation and spectral analysis. In the upstream region of the cavity, the streaks break down into smaller and less coherent structures, as shown by the reduction of the integral length scale. In the rearmost region of the cavity, a spanwise negative velocity correlation is interpreted as a dipole-type noise source, which is likely to reduce the radiated noise level with respect to a two-dimensional cavity flow. The velocity spectra show broadband amplification of modes related to the dominant scales in the cavity, as opposed to the selective mode amplification of cavities with a thin boundary layer inflow. A novelmultivariate non-dimensional analysis of the CFD parameters is presented, that explicits the modelling process for a cavity flow test case. This is used for estimating the simulation cost and the spatial and temporal resolution trade-off in the cavity flow simulation.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:574126 |
| Date | January 2013 |
| Creators | Monti, Manuele |
| Contributors | Rona, Aldo; Gao, Shian |
| Publisher | University of Leicester |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://hdl.handle.net/2381/28023 |
Page generated in 0.0025 seconds