Friday, 15 January 2016

Parallel, block-adaptive MHD simulations for solar coronal dynamics.


The SAC and SMAUG codes are based on the Versatile Advection Code we heard from one of the authors of the code about MPI-AMRVAC. This code is an advance on VAC,  the parallel scaling of the code exhibits weak scaling up to 30000 processors allowing to exploit modern peta-scale infrastructure. In particular the code has been developed to allow adaptive refinements of the computational mesh.

 The discussion focussed on solar physical applications modeled by the magnetohydrodynamic module of MPI-AMRVAC. The spatial discretizations available cover standard shock capturing finite volume algorithms, there are extensions to conservative high order finite difference schemes,  employing many flavors of limited reconstruction strategies. Multi-step explicit time stepping includes strong stability preserving high order Runge-Kutta steppers to obtain stable evolutions in multidimensional applications realizing up to fourth order accuracy in space and time.

 There was a discussion of the strategy for the AMR code. For a hypothetical grid arrangement exploiting 4 × 3 grid blocks at level l = 1, the left panel shows the global grid indices, while the right panel gives the tree representation with boolean variables indicating grid ‘leafs’.

Solar physics applications target the formation of flux rope topologies through
boundary driven shearing of magnetic arcades, following the in situ
condensation of prominences in radiatively controlled evolutions of arcades and
flux ropes, and the enigmatic phenomenon of coronal rain, where small-scale
condensations repeatedly form and rain down in thermodynamically structured
magnetic arcades.

The example above shows the density distribution at t = 3.01 × 105 yr after a supernova explosion. The entire domain is shown. Typical features such as the bow shock, the disturbed cloud with Richtmyer–Meshkov features on the front side, and a low-density region Rayleigh–Taylor instability behind the cloud are shown. The right hand image shows a zoomed-in look at the dust density distribution of species two in the cloud region. The dust can be seen to be tightly coupled to the dynamics of the cloud.

References

 `MPI-AMRVAC for solar and astrophysics', O. Porth, C. Xia, T. Hendrix, S.P. Moschou, & R. Keppens, 2014, ApJS 214, 4 (26pp) Full paper, doi:10.1088/0067-0049/214/1/4 )

`Parallel, grid-adaptive approaches for relativistic hydro and magnetohydrodynamics', R. Keppens, Z. Meliani, A.J. van Marle, P. Delmont, A. Vlasis, & B. van der Holst, 2012, JCP 231, 718-744. Full paper, doi:10.1016/j.jcp.2011.01.020. Accepted for special topical issue, with R. Keppens as Associate Editor. See also the Editorial Preface: Computational Plasma Physics.

 https://gitlab.com/groups/mpi-amrvac
http://homes.esat.kuleuven.be/~keppens/

 `Three-dimensional prominence-hosting magnetic configurations: creating a helical magnetic flux rope', C. Xia, R. Keppens, & Y. Guo, 2014, ApJ 780, 130 (11pp) Full paper, doi:10.1088/0004-637X/780/2/130

`Simulating the in situ condensation process of solar prominences', C. Xia, R. Keppens, P. Antolin, & O. Porth, 2014, ApJ Letters 792, L38 (6pp) Full paper, doi:10.1088/2041-8205/792/2/L38

https://perswww.kuleuven.be/~u0016541/Pub.html
https://perswww.kuleuven.be/~u0016541/Cont.html
https://perswww.kuleuven.be/~u0016541/Teach.html