4D Variational Data Assimilation with Large Eddy Simulation - Mécanique Accéder directement au contenu
Communication Dans Un Congrès Année : 2019

4D Variational Data Assimilation with Large Eddy Simulation

Assimilation 4DVAR avec simulation des grandes échelles

Résumé

Data assimilation as a tool for the study of fluid mechanics has attracted a lot of research attention over the past few decades with a lot of novel techniques being developed in the field. However, most techniques are restricted to 2D studies, or to reduced order implementations due to the high computational requirements in terms of time and memory needed for performing data assimilation of high dimensional fluid mechanical systems. An interesting approach to redress this is to apply the concepts of turbulence modeling via large eddy models to the dynamical system within data assimilation. This is the focus of this study, wherein an apt large eddy model is used to perform variational data assimilation of a transitional wake flow around a circular cylinder. By introducing turbulence modeling into the dynamical model of the data assimilation approach, the computational requirements are significantly reduced facilitating higher Reynolds number assimilation studies. The capability of the variational assimilation algorithm is analyzed with a synthetic data-set. The assimilation algorithm is shown to significantly improve the initial condition on the a-priori estimate provided while also predicting an improved time-evolution of the initial state.

Mots clés

Fichier non déposé

Dates et versions

hal-02610087 , version 1 (16-05-2020)

Identifiants

Citer

Pranay Chandramouli, Etienne Mémin, Dominique Heitz. 4D Variational Data Assimilation with Large Eddy Simulation. AIAA AVIATION Forum, Jun 2019, Dallas, United States. pp.1-13. ⟨hal-02610087⟩
18 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More