Section: New Results

Stochastic Downscaling Method

Participant : Antoine Rousseau.

In collaboration with TOSCA (Inria Sophia-Antipolis), LMD (Ecole Polytechnique) and CETE (Clermont-Ferrand), we investigate a new method for the numerical simulation of the wind at small scales. In this work, we consider a new approach for the downscaling in CFD, The local model that we propose is inspired from S.B. Pope's previous works on turbulence. We investigated a new numerical simulation method for the downscaling in CFD, with a strong orientation in applications to meteorology, particularly for the simulation of wind at small scales. The local model that we propose consists in modeling the fundamental equations of fluid motion by a stochastic Lagrangian model describing the behaviour of a fluid particle. Because of the both Lagrangian and stochastic nature of our model, it is discretized thanks to an interacting particle system, combining a time Euler scheme for stochastic differential equations and a Monte–Carlo approximation method. This model called SDM (Stochastic Downscaling Method) is adapted from previous works introduced by S.B. Pope [84] (see http://sdm.gforge.inria.fr/Accueil/index.en.php ).

This year, we worked on the comparison of the SDM model (endowed with a physical geostrophic forcing and a wall log law) with simulations obtained with a LES method (Méso-NH code) for the atmospheric boundary layer (from 0 to 750 meters in the vertical direction), in the neutral case. This work allowed to deeply understand the contribution of each elements of the Lagrangian model in terms of the turbulence production and dissipation, we analyzed the returns of various closure parametrization approaches, including viscosity turbulent approach. We also investigated anisotropic effect, with the introduction of GLM model in SDM (see [84] ), in particular the isotropic relaxation case. We gave our conclusions as a part of the final report for ADEME [58] . A paper is in preparation.