Diagnose dG bug - interface-hugger

[lsawade]

There is a wave traveling tangentially to this diagonal interface. I have seen errors up to 5e-1.

This also holds for Ivy-like bathymetry, but I have not seen errors that large in this case.

Tasks

• re-examine flat case
  • if it appears, implement periodic bdries in gmshlayerbuilder. If they vanish, it is likely an interaction between stacey and interface (unlikely, though).
  • otherwise, it's an error that probably appears in most cells on the interface, so the debugger is probably the go-to at that point.

Acceptance Criteria

• traces are consistently error < 5e-2 for ivy bathy, < 1e-2 for non-curved

[lsawade]

Test Cases

• Keep results for flat bottom with non conforming

• Diagonal results

• Sinusoidal

  • Rough bathymetry
  • smoothed bathymetry
  • different resolutions. Use your fraction 7/3 but do factor of it, 1x, 2x, 3x, number of elements and show that the error is reduced.

Once you can show that those cases (better, finer bathymetry) reduce the error.

• Ivy's problem with the same setup as the sinusoidal, as a real world example.

[lsawade]

Todo list is now #1434 #1435 #1436