Periodic Immersed Boundaries

[danieljvickers]

There is currently no concept of periodicity for immersed boundary generation when using periodic boundary conditions. This means that moving immersed boundaries that cross the boundary will have errors, as only cells on the side of the grid with the centroid will be included in the calculations of the boundary markers.

A PR for this needs to resolve this issue. Ideally a test case of a moving immersed boundary crossing the period boundary diagonally (as that will test the periodicity wrapping around two dimensions at once) should be added.

[sbryngelson]

Hi @danieljvickers can you square away your 'fix' (i think i merged it, maybe?) w/ #1043 from Conrad? What's the difference for the periodic part? anything?

[danieljvickers]

I haven't added periodicity to the IB's yet. But I can look at his PR and make sure it makes sense.

[sbryngelson]

You might cherry-pick that part and create a new PR for it... to isolate it from the rest of the rather large PR

[danieljvickers]

Yeah, I left comments. It is fine how it is just to get the feature merged, but it will probably need a second PR to extend the periodicity to all patch types and reduce code size a bit.

[sbryngelson]

well the PR from Conrad is quite large so it might take a bit to merge

[danieljvickers]

Ok, are you asking me to work on just the periodic IB part of it and then help that get merged in with Conrad?

[sbryngelson]

depends on how quickly we want that feature merged it, really. If you want it merged in, you can cherry-pick it out (or merge in your own version).

[danieljvickers]

I am still working on 2-way coupling right now. It was next on my list after that. So I was fine not stepping in to look at it until after that.

[sbryngelson]

sounds good!