Closed WPringle closed 5 months ago
Keep in mind, this approach is set up to work for a single domain. You'll need to create a nested function inside mesh1d to correctly handle the multiscale case.
Also keep in mind the pseudo deltaT for the 1d mesh generation problem should be pretty small or otherwise it goes numerically unstable.
Moving this along a bit.
meshgen
object not when building it. The logical behind that design is so that users can just call meshgen
to develop the fixed points and edges but not build the mesh yet. Note you can plot those constraints with ...
mshopts = meshgen('ef',fh,'bou',gdat);
mshopts.plot()
The only changes the user needs to make are adding the high_fidelity
numerical flag 1 to the geodata
class constructor they want to employ the automatic generation of point and edge constraints.
gdat = geodata('shp',coastline,'bbox',bbox,'h0',min_el,'high_fidelity',1);
Constraints are displayed during meshing as blue points.
TestJBAY.m not passing (Example_5b_JBAY_w_weirs.m). On further inspection appears weirs are not appearing correctly: Projection branch result zoom-in:
This branch result zoom-in:
Can work out what changed wrt to the weir stuff?
I have restored a mesh improvement strategy that improves the element qualities nearby the weirs. This will be auto deactivated when using the high fidelity mode. The JBAY test produced a difference of 6% in total number of nodes from the test value but the more important part is the weirs were correctly meshed.
@krober10nd It looks good to me, can you do the approval and merging and new version release?
Looks like there was a formatting issue with
meshgen.m
and it shows up as 1000 lines changed...