lp:~fluidity-core/fluidity/ice-melt-mesh-movement

Branch merges

Propose for merging

No branches dependent on this one.

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

4218. By Stephan Kramer on 2013-05-18

Mesh movement according to ice melt.

To use choose ice_melt as mesh movement scheme and add a prescribed field
"ThermalConductivity" with the conductivity values for water and ice in the
corresponding regions. As the ice melt depends on the normal temperature
gradient at the interface you'll need enough resolution to resolve the
diffusive boundary layer.

4217. By Stephan Kramer on 2013-05-13

Reconstruction of internal boundary after adaptivity.

Seems to got happily through several adapts now. Not tested in parallel yet.

4216. By Stephan Kramer on 2013-05-08

Make add_faces() accept internal boundaries without element ownership information.

If no element ownership information is provided it is assumed that each internal facet is only represented once in the provided surface mesh (sndgln). In add_faces that facet will be duplicated and also its surface ids will be copied. This means after the add_faces() call the surface_element_count() may be higher than the number of provided surface elements.

Note that this commit also adds some more sanity checks and improved error messages. If something stops working at this commit it may be that your mesh input was already broken and it's only now being picked up. Passes short tests at this point. New add_faces() functionality not actually tested yet.

4215. By Stephan Kramer on 2013-05-08

Don't shortcut get_coplanar_ids() if stotel==0.

This breaks in parallel because merging should be collective.

4214. By Stephan Kramer on 2013-05-08

Properly fix get_coplanar_ids with internal boundaries.

Turns out my previous fix was wrong; Just removing duplicate facets from the surface mesh is not sufficient to be able to call make_eelsit on it, because the surface mesh will contain "splits" (e.g. where the internal and external boundary meet) where there will still be surface elements which have multiple neihbouring surface elements that can be reached through the same edge (facet of the surface element).
Therefore, giving up on using eelist for the surface mesh. Changed the forward front algorithm to only use a nelist, considering all surface elements that are connected via just a single node for inclusion the current coplane.

4213. By Stephan Kramer on 2013-05-07

Fix coplanar ids calculation for internal boundaries.

This fixes https://bugs.launchpad.net/fluidity/+bug/1088860 (hopefully) which prevented make_eelist from giving
an error when elements were adjacent to more than one elements through the same facet. Fixing this means we can
no longer call make_eelist on the surface_mesh if there is an internal boundary. Instead in the coplanar calculation
we now construct a temporary copy of the surface mesh with one of each pair of duplicate facets on the internal boundary
removed.

4212. By Stephan Kramer on 2013-05-07

Also pass internal boundary on to mba2d.

This is a first step to fix prescribed_region + adaptivity. Next step is to reconstruct the internal boundary after the adapt.

4211. By Jonathan Bull on 2013-04-29

fixing another typo in a Helmholtz test.

4210. By Christian Jacobs on 2013-04-27

Also update the BFS 2D example so diamond_validation passes: Include discretisation_options for the Density field used in the KEpsilon equation_type.

4209. By Christian Jacobs on 2013-04-27

Bug fix following Cian's commit: Include discretisation_options for the Density field used in the KEpsilon equation_type.