Top-level heading

An all-speed scheme for isentropic two phase flows

Categoria
Seminari di Modellistica Differenziale Numerica
Data e ora inizio evento
Data e ora fine evento
Aula
Sala di Consiglio
Sede

Dipartimento di Matematica Guido Castelnuovo, Università Sapienza Roma

Aula esterna
on-line su ZOOM
Speaker

Andrea Thomann, Johannes-Gutenberg Universität Mainz

We are interested in the numerical simulation of liquid-gas mixtures, where the sound speed of the liquid phase is consistently faster than the one of the gas phase. If in addition, the material wave is significantly slower than the individual acoustic waves, the system can exhibit three different scales of wave speeds. In these regimes, which are characterized by small, potentially different phase Mach numbers, using an explicit scheme requires a time step that scales with the smallest appearing Mach number. Moreover, the main interest often lies on a sharp resolution of slow dynamics which would allow for a much larger time step. Therefore, we use implicit-explicit (IMEX) time integrators where fast waves are treated implicitly leading to a CFL condition which is restricted only by the local flow velocity. In this talk, we present an all-speed finite volume scheme for isentropic two-phase flows based on a symmetric hyperbolic thermodynamically compatible model given developed by Toro and Romenski (2004). Since the flow regimes can range from compressible for gases to almost incompressible for some liquids, the asymptotic preserving (AP) property together with the correct numerical viscosity are essential. Since the flow regime of the considered two-phase flow model is characterized by two potentially distinct phase Mach numbers, different singular Mach number limits can be obtained which depend on the constitution of the mixture. The AP property of our IMEX scheme is obtained by using a reference solution approach. The consistency with single phase flow, accuracy and the approximation of material waves in different Mach number regimes is illustrated in numerical simulations.