Boundary Conditions

[ax3l]

We need a huge bunch of useful boundary conditions:

• PML (perfectly matching layer)
• thermal (foils)
• reflecting add reflective particle boundary conditions #3806
• open space (Lindman)
• ...
• periodic
• absorbing (field right now exp. decaying, ) -> remove particles from memory (done), maybe record them Output: Particles That Leave the Box #517 and think about Transition Radiation for leaving particles

Maybe we should differentiate between particle and field boundary conditions!

Some refs:

• Lars D. Ludeking, Andrew J. Woods. Well Matched Electromagnetic Boundary in FDTD-PIC for Charged Particle Penetration, The Open Plasma Physics Journal 3, 53-59, 2010
• PML: Taflove, Hagness "Comp. Electrodyn - the FDTD Method 3rd ed.", ch. 7
• PML: Jean-Pierre Berenger. A Perfectly Matched Layer for the Absorption of Electromagnetic Waves Journal of Computational Physics 114, 185-200, 1993

[ax3l]

in the moment where fields or particles are manipulated, the "lost" or "added" energy due to the boundaries should be calculated and written out.

[PrometheusPi]

@ax3l I agree. Otherwise our energy conservation tests only work for periodic boundary conditions.

[bussmann]

++1

[dgholstein]

I presume the code will not handle arbitrary conducting boundaries such as one would find on a magnetron, is that correct? That it would do RF simulations as would HFSS or CST?

[ax3l]

welcome to our issue tracker @dgholstein !

Conducting boundary conditions can be, e.g., modelled with PML conditions (which just need to be implemented and have been so in many PIC codes before).

Please be aware that the PIC method as implemented in this code is a FDTD PM method. HFSS is using FEM methods and CST is providing a whole framework of codes with arbitrary methods and tracers.

[n01r]

Ping @sbastrakov.
A very old issue - I already ticked the box for PML.
Is there anything else from the list that is done or planned for the next months?

[sbastrakov]

I am not aware if we plan to implement some more soon, but here is my view on the BCs in the list:

• thermal: for particles can be useful to have, although @ax3l pointed out in an offline discussion (quite a while ago) that it also has some limitations
• reflecting: this one should be easy to do, not sure how useful though
• Lindman open space: should be not needed since we now have PML

Some time ago I was also thinking of moving BCs to runtime parameters, we should do this sometime, should not be that hard.

[ax3l]

reflecting: this one should be easy to do, not sure how useful though

very useful addition if you like to that that, e.g. to simulate shock scenarios :)
(one downstream plasma -> reflected at boundary -> nice counter-prop shock physics & pretty image ;) )

[ervw12]

It is quite some time ago. But are there any plans of implementing reflecting boundaries soon? We want to simulate some shock scenarios already mentioned by @ax3l, where this BC would be very useful.

[PrometheusPi]

@ervw12 There is currently a pull request open, that will implement reflecting boundary conditions for particles. See here: #3806 It will be merged within a few days.

[steindev]

With reflecting and thermal particle boundaries implemented in #3806 and #3858, respectively, as well as PML field boundaries in #2950 (as an 'equivalent' to the Lindman bc in the description), I take the liberty to close this long standing issue.
Well done @lennertsprenger and @sbastrakov!