Symmetry boundary conditions
Introduction
A particular set of boundary conditions can have an important effect on the results of plasma simulations [11], [3], [25], [26]. An important class of boundary conditions is that of energy conserving boundary conditions. Energy conserving boundary conditions are sometimes appropriate for laboratory plasmas [2]. Energy conserving boundary conditions are not ideal for all problems. For instance, in simulation of space plasmas, it is often desirable for energy to radiate out of the system simulating an open boundary [12], [14], [5]. But it is difficult to find boundary conditions that let all of the various wave modes escape out of the simulation region, and such boundary conditions are often much less stable than energy conserving boundary conditions. Typically when numerical codes “blow up” (that is, values of quantities get very large), it happens in such a way that the computational energy grows. Related to this, the most stable boundary conditions are typically energy conserving. For some applications, such as the study of Alfvén waves that are guided along the magnetic field, the boundary condition across the magnetic field should not matter, and an energy conserving boundary condition may be convenient. Finally, energy conservation is one of the best tests of a numerical code, and energy conserving boundary conditions make it easier to check that the total simulation energy is constant.
Here, we demonstrate a simple method to determine and implement energy conserving boundary conditions using symmetries about the boundaries. The boundary conditions appropriate for Cartesian coordinates can be used for curvilinear coordinates if the scale factors describing those coordinates are taken to be symmetric about the boundary. We illustrate our method for several different sets of plasma physics equations. While not all aspects of the boundary conditions we describe here are new, we believe that our clear explanation of how to implement them (which we have not found elsewhere) will be helpful to researchers using fluid or particle simulations.
For the sake of completeness, we have included in Section 5 a description of symmetry boundary conditions for reduced MHD. Much of this material can be found in Ref. [9]. Here we have a more complete description of the boundary conditions for all fields (Table 4) using a notation consistent with that used in the rest of this paper.
Section snippets
Symmetry basics
In Table 1, we list the symmetries used in this paper. We use (+) to indicate that a field value is symmetric relative to a particular boundary, and (−) to indicate that a field value is antisymmetric relative to that boundary. “Symmetric” here means that the field is even across the boundary; that is, , where x is the coordinate normal to the boundary at , and . In this case, the first derivative of f is zero at the boundary. Antisymmetry means that the field is odd
Energy conserving boundary conditions for MHD
The normalized MHD equations arewhere B is normalized to the background magnetic field is normalized to the background mass density is normalized to the background Alfvén speed , where is the permeability of free space, and p is normalized to .
The first step toward deriving energy conserving boundary conditions is to derive an energy equation. This is found by dotting (24) with v,
Energy conserving boundary conditions for a hybrid code
Normalized hybrid code equations for two species, particle protons (p) and fluid electrons (e) [27], are written aswhere m is the particle index, i is the grid index (possibly representing grid points on a multidimensional grid), the particle weight is the volume of grid cell i, and is the number of particles. The
Energy conserving boundary conditions for linear reduced MHD
Now we consider a simple set of linear reduced MHD (RMHD) equationswhere the electromagnetic fields areand the equations have been written in a form valid for curvilinear coordinates with , but assuming . The and subscripts indicate components respectively parallel and perpendicular to the background magnetic field . We define and , the parallel and perpendicular scale
Summary and discussion
We have shown how to implement energy conserving boundary conditions using exact symmetries at the boundaries. This method can be used even for a curvilinear coordinate system if the symmetries are exact. (When equations do not have exact symmetries, it is usually better to let the equilibrium quantities and scale factors be continuous at the boundary.) We demonstrated how to work out the symmetries of the various fields for MHD, a hybrid code, and for linear reduced MHD.
In Section 2, we
Acknowledgments
Work at Dartmouth was supported by NSF grants ATM-0503371 and ATM-0120950 (Center for Integrated Space Weather Modeling, CISM, funded by the NSF Science and Technology Centers Program) and NASA grant NNX08AI36G (Heliophysics Theory Program). RED especially thanks James F. Drake for introducing him to symmetry boundary conditions. We also thank John Lyon, Jan Hesthaven, and Alain Brizard for helpful discussions.
References (28)
- et al.
Geophys. Res. Lett.
(1994) - et al.
J. Comput. Phys.
(1981) - et al.
Phys. Plasmas
(2004) Mathematical Methods for Physicists
(1970)- et al.
Proc. Roy. Soc. London, Ser. A
(1958) - et al.
Plasma Physics via Computer Simulation
(1985) Phys. Plasmas
(2005)- et al.
Phys. Plasmas
(2006) - et al.
J. Geophys. Res.
(1998) - et al.
J. Geophys. Res.
(2002)
Phys. Plasmas
Phys. Plasmas
J. Geophys. Res.
Cited by (11)
Novel tri-viscous model to simulate pumping of flowable concrete through characterization of lubrication layer and plug zones
2022, Cement and Concrete CompositesCitation Excerpt :On the other hand, the boundary conditions for pressure were zero gradient in inlet and fixed value in outlet for both phases [63]. The no-slip and symmetry boundary conditions were applied for walls and center axis of the pipe, respectively [64,65] (Fig. 3b). The boundary conditions for α are similar to the velocity boundary conditions (i.e., fixed value at the inlet and zero-gradient in outlet).
Multigrid treatment of implicit continuum diffusion
2019, Computer Physics CommunicationsGDB: A global 3D two-fluid model of plasma turbulence and transport in the tokamak edge
2018, Computer Physics CommunicationsIntuition and Symmetries in Electromagnetism: Eigenstates of Four Antennas
2022, Applied Sciences (Switzerland)Symmetric Molecular Dynamics
2022, arXiv