A hybrid discontinuous Galerkin method for tokamak edge plasma simulations in global realistic geometry

Progressing toward more accurate and more efficient numerical codes for the simulation of transport and turbulence in the edge plasma of tokamaks, we propose in this work a new hybrid discontinuous Galerkin solver. Based on 2D advection-diffusion conservation equations for the ion density and the particle flux in the direction parallel to the magnetic field, the code simulates plasma transport in the poloidal section of tokamaks, including the open field lines of the Scrape-off Layer (SOL) and the closed field lines of the core region. The spatial discretization is based on a high-order hybrid DG scheme on unstructured meshes, which provides an arbitrary high-order accuracy while reducing considerably the number of coupled degrees of freedom with a local condensation process. A discontinuity sensor is employed to identify critical elements and regularize the solution with the introduction of artificial diffusion. Based on a finite-element discretization, not constrained by a flux-aligned mesh, the code is able to describe plasma facing components of any complex shape using Bohm boundary conditions and to simulate the plasma in versatile magnetic equilibria, possibly extended up to the center. Numerical tests using a manufactured solution show appropriate convergence orders when varying independently the number of elements or the degree of interpolation. Validation is performed by benchmarking the code with the well-referenced edge transport code SOLEDGE2D (Bufferand et al., 2013, 2015 [1,2]) in the WEST geometry. Final numerical experiments show the capacity of the code to deal with low-diffusion solutions.

Giorgio Giorgiani, H. Bufferand, G. Ciraolo, P. Ghendrih, Frédéric Schwander, et al.. A hybrid discontinuous Galerkin method for tokamak edge plasma simulations in global realistic geometry. Journal of Computational Physics, Elsevier, 2018, 374, pp.515-532. ⟨10.1016/j.jcp.2018.07.028⟩. ⟨hal-01946999⟩

Journal: Journal of Computational Physics

Date de publication: 01-12-2018

  • Giorgio Giorgiani
  • H. Bufferand
  • G. Ciraolo
  • P. Ghendrih
  • Frédéric Schwander
  • Eric Serre
  • P. Tamain

Digital object identifier (doi): http://dx.doi.org/10.1016/j.jcp.2018.07.028

x >