Cylindrical domain

[nilsvu]

• Add spherical shells
• Add time-dependent maps
• Log and 1/r distributions

[geoffrey4444]

01/20/26: Nils Vu has been attempting the cylindrical domain runs, so has Alex Carpenter, ... Iago working on log maps for cubic domain

[geoffrey4444]

@markscheel : shouldn't be hard to remove the layer that turns the spherical shells / cylindrical shells and instead put in hollow cylinders or filled cylinders... cylindrical filtering is applied to pairs, but it's just the exponential filter

@macedo22 : has started trying to evolve with the Cylindrical domain... it imports the initial domain, can't create the target domain...

[nilsdeppe]

03/03/2026:
Alex Macedo: Able to evolve a little bit now. Takes time steps, looks like domain is expanding, need to understand why.
Geoffrey: Were the bugs in input files or code?
Alex M: In the code, e.g. grid anchors were on the z-axis instead of x-axis.
Mark: Why didn't unit test problem?
Geoffrey: Alex M could look into that. This is great. For domain expanding depends on the expansion map. Also, gauge dynamics can cause the domain to expand a bit near the BHs.
Nils D: The outer boundary should never expand.
Geoffrey: Yep, Alex M can look into this.

[macedo22]

To clarify on my end, by "domain expanding" I mean the radius of the apparent horizon seems to be increasing as opposed to the outer boundary. Going to do sanity checks on my part and if those check out, then look at the time dependent maps.

The specific error that suggested an expanding domain took this form and was consistent with seeing the apparent horizon radius appearing to grow:

Wall time: 08:23:50
Node: 5 Proc: 63
void evolution::dg::Actions::detail::apply_boundary_condition_on_face(const gsl::not_null<db::DataBox<DbTagsList> *>, const BoundaryCorrection &, const BoundaryCondition &, const Direction<Dim> &, const Variables<tmpl::list<EvolvedVariablesTags...>> &, const Variables<db::wrap_tags_in<Tags::Flux, ty [...] intGamma1, gh::Tags::ConstraintGamma2, gr::Tags::Lapse<DataVector>, gr::Tags::Shift<DataVector, 3>>] in /home/fs01/spec1182/spectre/src/Evolution/DiscontinuousGalerkin/Actions/BoundaryConditionsImpl.hpp:410

We found the radial mesh velocity points in the direction of the outward normal, i.e. we possibly have an expanding domain. Its unclear if proper boundary conditions are imposed in this case.

In element:[B76,(L2I2,L2I3,L2I3)]
In direction: +2
############ ERROR ############

[nilsdeppe]

This is the outer boundary, though. It shouldn't expand even when the BHs are moving apart, or am I misremembering @markscheel ?

[markscheel]

The expansion map should have two FunctionOfTimes associated with it:

1. The FoT for the bulk (which can be increasing if the holes move apart)
2. The FoT for the outer boundary (which should be slowly drifting inward)

The expansion map uses both of those FoTs to construct a smooth map that agrees with (1) for most of the interior domain, (2) at the outer boundary, and some superposition of (1) and (2) in the wave zone.

[macedo22]

Apologies, I had the blocks wrong and misunderstood the output. Thanks @nilsdeppe