(281e) On the Use of V-LES as An Alternative to LES for Industrial Turbulent Flows

The limits of the RANS approach for turbulent single phase and multi-phase flow have been reached. The competition between industrial key players in automotive and land-based and aerospace technologies in particular is promoting the development of new modeling strategies that put flow unsteadiness at the forefront of their predictive capabilities. This added to the constant growth of computer resources in terms of speed and mass storage helps replace the LES concept at the front stage of applied CFD. Be it as it may, the LES strategy is still too expensive for various types of flows, in particular those in which steady state conditions are difficult to achieve (e.g. strong non-homogeneity of turbulence). This is the reason why alternative approaches based on scale separation are indispensable; including the V-LES (Very Large Eddy Simulation), the filter-based technique and the so-called SAS variant. The idea is to combine RANS and LES for a specific flow, where the knowledge of the size of the most important scales is key. Here the fow is decomposed into resolved and subscale part, the latter being independent from the grid (in contrast to the SGS in LES), but is dependent on the flow (and thus the obstacles). Larger scales than this specific scale are directly resolved, meanig actually no model is included, but the sub-scale part is modelled, but with a more refined statistical turbulence model than the 0-equation Smagorinsky one, i.e. a k-epsilon or even a Reynolds stress model. The approaches cited above differ in the way the transtion between resolved and sub-scale turbulence is achieved. We discuss our latest results, in which we have conducted a systematic sensitivity analysis of the involved parameters. The V-LES is shown to be a viable alternative to LES for very high Re flow conditions.