A Mach-Sensitive Splitting Approach for Euler-like Systems - Archive ouverte HAL
Article Dans Une Revue ESAIM: Mathematical Modelling and Numerical Analysis Année : 2018

A Mach-Sensitive Splitting Approach for Euler-like Systems

Résumé

Herein, a Mach-sensitive fractional step approach is proposed for Euler-like systems. The key idea is to introduce a time-dependent splitting which dynamically decouples convection from acoustic phenomenon following the fluctuations of the flow Mach number. By doing so, one seeks to maintain the accuracy of the computed solution for all Mach number regimes. Indeed, when the Mach number takes high values, a time-explicit resolution of the overall Euler-like system is entirely performed in one of the present splitting step. On the contrary, in the low-Mach number case, convection is totally separated from the acoustic waves production. Then, by performing an appropriate low-Mach correction on the acoustic step of the splitting, the numerical diffusion can be significantly reduced. A study made on both convective and acoustic subsystems of the present approach has revealed some key properties as hyperbolicity and positivity of the density and internal energy in the case of an ideal gas thermodynamics. The one-dimensional results made on a wide range of Mach numbers using an ideal and a stiffened gas thermodynamics show that the present approach is as accurate and CPU-consuming as a state of the art Lagrange-Projection-type method.
Fichier principal
Vignette du fichier
Mach_Sensitive_HAL.pdf (1015.85 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01466827 , version 1 (13-02-2017)
hal-01466827 , version 2 (20-12-2017)

Identifiants

Citer

David Iampietro, Frédéric Daude, Pascal Galon, Jean-Marc Hérard. A Mach-Sensitive Splitting Approach for Euler-like Systems. ESAIM: Mathematical Modelling and Numerical Analysis, 2018, 52 (1), pp.207-253. ⟨10.1051/m2an/2017063⟩. ⟨hal-01466827v2⟩
519 Consultations
340 Téléchargements

Altmetric

Partager

More