In this work, Monotone Integrated LES (MILES) and conventional LES models have been used to investigate the Taylor-Green Vortex (TGV) problem. We have examined the behavior of MILES based on various different high-resolution schemes, including 2nd and 4th order FCT, CB-Godunov, and Lagrange Remap algorithms. Conventional LES approaches tested included Smagorinsky, Dynamic Smagorinsky models, One-Equation Eddy-Viscosity model (OEEVM), as well as a mixed model (MM) combining an OEEVM and a scale similarity model. Comparisons of TGV integral measures based on the present simulations are made with previously reported and new DNS data. The results show that all the MILES approaches employed here can provide stable and acceptable (in terms of accuracy) solutions without resorting to an explicit subgrid scale (SGS) model, as well as using relatively coarse grids. The results also show that the kinetic energy dissipation rate does depend on the details of the numerical scheme employed (and its particular associated implicit SGS model). Therefore, even though MILES provides a fairly robust computational framework for LES, there is plenty of room to achieve improvements on MILES performance based on better understanding of the specific dissipation and dispersion properties of the different high-resolution schemes. In particular, further such investigations are clearly warranted in order to gain better insights into the accuracy (and computational behavior in general) of MILES in relation to LES.
|Title of host publication||Collection of Technical Papers - 44th AIAA Aerospace Sciences Meeting|
|Number of pages||11|
|Publication status||Published - 1 Dec 2006|
|Event||44th AIAA Aerospace Sciences Meeting 2006 - Reno, NV, United States|
Duration: 9 Jan 2006 → 12 Jan 2006
|Conference||44th AIAA Aerospace Sciences Meeting 2006|
|Period||9/01/06 → 12/01/06|