| In this thesis the longstanding need for a dynamically adaptive Large Eddy Simulation (LES) method has been addressed. Current LES methodologies rely on, at best, a zonal grid adaptation strategy to attempt to minimize computational cost in resolving large eddies in complex turbulent flow simulations. While an improvement over regular grids, these methodologies fail to resolve the high wave number components of the spatially intermittent coherent eddies that typify turbulent flows, thus not resolving valuable physical information. At the same time the flow is over resolved in regions between the intermittent coherent eddies. The Stochastic Coherent Adaptive Large Eddy Simulation (SCALES) methodology addresses the shortcomings of LES by using a dynamic grid adaptation strategy that resolves the most energetic coherent structures in a turbulent flow field. This new methodology inherits from Coherent Vortex Simulation (CVS) the ability to dynamically resolve and "track" the most energetic part of the coherent eddies in a turbulent flow field, while using a field compression similar to LES, which could be considerably higher than with CVS. Unlike CVS, which is able to recover low order statistics with no subgrid scale stress model, the effect of the unresolved subgrid scale stresses must be modeled in SCALES. The required SGS stress modeling is addressed in the form of a dynamic SGS stress eddy viscosity model based on an alternative Germano identity using a wavelet thresholding test filter. Using this new dynamic model SCALES is shown to moderately outperform LES in the unit test problem of decaying isotropic turbulence, yet the real strength of SCALES will be seen in complex real world flows, where dynamic grid adaptation will completely eliminate the onerous overhead of current grid generation methods. |
