流　体　力　学　セ　ミ　ナ　ー

日時:  ３月１３日（月） １０：３０ から １２：００

場所:  京大 理学研究科５号館(物理学教室) ４１３号室

講師： Professor Mikhael Gorokhovski
       Laboratoire de Mécanique des Fluides et d’Acoustique
       École Centrale de Lyon      

講演題目： Stochastic subgrid acceleration models in LES of two-phase flows. 

講演要旨：
It was recently recognized that the turbulence at the high Reynolds number is characterized by quick peaks in the velocity signal, and the corresponding velocity gradients may attain intermittently very large values. This is a manifestation of intense and long-lived flow structures at small spatial scales.  However, in practical computations of highly turbulent flows, the smallest scales are filtered, and consequently, the effects of intermittency are mainly neglected. In this talk we propose to discuss an approach to account the sub-filtered effects of intermittency in the stochastic way. The idea is to force the filtered Navier-Stokes equations by the stochastic acceleration term with prescribed properties known from the experimental and DNS statistics of the Lagrangian acceleration. The stochastic model for this acceleration contains the lognormal process for its norm (with long-range correlations), and the model for the acceleration direction (with short range correlations). The latter represents the Ornstein-Uhlenbeck process on the unit sphere with relaxation to the locally resolved vorticity vector. Both models – one for the norm and another for the direction of acceleration - depend on the local Reynolds number. The model is referred to as LES-SSAM (Stochastic Subgrid Acceleration Model). In this talk, the fluid mechanics from LES-SSAM is assessed by DNS and is considered as a background turbulent flow in the following situations: 
- The box turbulence laden by heavy particles; DNS/LES-SSAM. The assessment of 
single and two time statistics of the particle acceleration and velocity illustrated clearly the advantage of the proposed flow model in predictions of the non-Gaussian statistics and autocorrelations of heavy particles. 
- The homogeneous shear flow laden by heavy particles; DNS/LES-SSAM. Here we will show that the volume averaged acceleration norm of particle exhibits a pseudo-cyclic behavior, which is a signature of the periodic action of largest confined vortexes against the mean shear. Another interesting discussion will concern the preferential orientation of the particle acceleration given by mean shear-induced vortical structures. These both principal findings may have relation to the turbulence-induced intermittent structures in clouds.
- Turbulent incompressible flows with interface. Here the possibility of combination of LES-SSAM with the Volume of Fluid (VoF) approach will be discussed and illustrated for primary atomization process.