A fully-implicit finite-volume method for multi-fluid reactive and collisional magnetized plasmas on unstructured meshes

Research output: Contribution to journalArticlepeer-review

Abstract

We present a Finite Volume scheme for solving Maxwell's equations coupled to magnetized multi-fluid plasma equations for reactive and collisional partially ionized flows on unstructured meshes. The inclusion of the displacement current allows for studying electromagnetic wave propagation in a plasma as well as charge separation effects beyond the standard magnetohydrodynamics (MHD) description, however, it leads to a very stiff system with characteristic velocities ranging from the speed of sound of the fluids up to the speed of light. In order to control the fulfillment of the elliptical constraints of the Maxwell's equations, we use the hyperbolic divergence cleaning method. In this paper, we extend the latter method applying the CIR scheme with scaled numerical diffusion in order to balance those terms with the Maxwell flux vectors. For the fluids, we generalize the AUSM+-up to multiple fluids of different species within the plasma. The fully implicit second-order method is first verified on the Hartmann flow (including comparison with its analytical solution), two ideal MHD cases with strong shocks, namely, Orszag-Tang and the MHD rotor, then validated on a much more challenging case, representing a two-fluid magnetic reconnection under solar chromospheric conditions. For the latter case, a comparison with pioneering results available in literature is provided.

Original languageEnglish
Pages (from-to)252-276
Number of pages25
JournalJournal of Computational Physics
Volume318
DOIs
Publication statusPublished - 1 Aug 2016
Externally publishedYes

Keywords

  • Finite volume method
  • Magnetohydrodynamics (MHD)
  • Multi-fluid
  • Plasma

Fingerprint

Dive into the research topics of 'A fully-implicit finite-volume method for multi-fluid reactive and collisional magnetized plasmas on unstructured meshes'. Together they form a unique fingerprint.

Cite this