Robin/Inhomogeneous Neumann BC for EB

[ThomasHowarth]

Myself and some colleagues were looking to add in surface chemistry in Pele. This requires either Robin or inhomogeneous Neumann boundary conditions. For cases without EB, we can already work on using the existing LinOpBCTypes; however, from what I can tell, these boundary conditions are currently unavailable for EB. Do you have any suggestions on making a start here, and how much effort would be required to implement such a feature?

[WeiqunZhang]

I assume Pele here means PeleLMeX, because LinOpBCType is mentioned. There are several types of linear systems in PeLeMeX, MAC projection, nodal projection, tensor solver, and diffusion. Could you be more specific about what you need for these solvers?

@drummerdoc might have more comments.

[drummerdoc]

Indeed this is LMeX. I believe that the MAC, nodal and tensor solvers don't need to be involved in this work. This is likely only the CC ABeC stuff.

[WeiqunZhang]

Okay. This is about MLEBABecLap, which currently supports Dirichlet and homogeneous Neumann on EB. It should not be too hard to support inhomogeneous Neumann, but I am not sure about Robin BC.

[ankithadas]

Just thinking out loud, could we encapsulate all the EB operations (fluxes, normalization, etc.) within a dedicated Stencil class? This EB stencil object would store standard stencils for common flux operations, while also supporting user-defined EB stencils and EB-specific functionality.