WIP: Allow specifying which components of q to output on each fgout grid #17
Conversation
* origin/main: minor change: move g=grav above loop on cells in rpt2 set g=grav in transverse solver
* origin/main: add momentum autostop for binary delete fgmax files (address this later) this ran and wrote a file with the right number of columns. set tau to zero for very low m. improve filpatch.f90 and filval.f90 to fix m0 for refinement. adding fixed grid routines to dclaw to accommodate meqn=7. define g as dave had on main define g if bednormal is not 1 d0s subtract sea level
Restored geoclaw_radial_slide/set*.py and included fgout specification in radial_slide/setrun.py for further testing. * origin/main: put output time back to 3. update radial slide setrun add more indicies to qinit fix index error from first pass first pass at moving to indicies. remove old move bowl slosh from examples to test starting to remove examples (except for radial slide)
case for adding Bouss terms. This version runs GeoClaw, not D-Claw. Note that with order=2 there is some oscillation in the leading shock, more pronounced with transverse_waves = 1 than with 2.
…ta, requires new fgout capabilities
|
|
||
| #list of common modules needed for dclaw codes | ||
| COMMON_MODULES += \ | ||
| $(DIGLIB)/digclaw_module.f90 \ | ||
| $(DIGLIB)/qinit_module.f90 \ | ||
| $(DIGLIB)/auxinit_module.f90 \ | ||
| $(DIGLIB)/fgout_module.f90 \ |
There was a problem hiding this comment.
@rjleveque , @dlgeorge I suggest we delete dclaw/src/2d/dig/fgout_module.f90.
We only have it it as a placeholder that is no longer needed b/c Randy updated the GEOLIB/fgout_module.f90.
Thoughts?
There was a problem hiding this comment.
seems best to remove it I guess.
There was a problem hiding this comment.
I think that's fine. The only difference in it was that it set
fg%num_vars(1) = 10 ! DIG different number
but this is no longer needed with the geoclaw updates.
|
@rjleveque I've tested this and it works well for me. My only thoughts are (1) that we should remove the two tests because they are not really of dclaw. Should these not go in the geoclaw/tests/geoclaw_radial_slide and geoclaw/tests/geoclaw_bouss_radial_slide? (2) that we remove the dclaw/src/2d/dig/fgout_module.f90 file. Let me know what you think (also tagging @dlgeorge). Other than these two issues, I think this is ready to merge. |
|
@rjleveque @kbarnhart . plots look great, and q_out_vars method seems good to me. PR seems good to me (?). |
|
@kbarnhart: regarding moving the examples to geoclaw, I put them here since I was using the SWE version to compare to what I got when running D-Claw with mass fraction 0, so that it was pure water. I don't think it really makes sense to include in GeoClaw since it's not a situation most GeoClaw users would be setting up, although I suppose it could be interesting as a radially symmetric overland flood example. At any rate I'm fine with removing them from the (The |
|
@rjleveque : @dlgeorge and I discussed and we decided that it made sense to keep them here, broadly for the reasons you list. I think keeping geoclaw_bouss_radial_slide makes sense (eventually) when dclaw-bouss exists. So I went ahead and merged it. Thanks for this PR! |
This requires clawpack/geoclaw#624
The main
Makefile.dclawnow usesfgout_module.f90fromgeoclawrather than a separate one for D-Claw.(I also added a couple directories
tests/geoclaw_radial_slideandtests/geoclaw_bouss_radial_slidefor testing geoclaw SWE and Bouss on this radial problem with water on the hill instead of dirt, perhaps useful to others, but these don't have fgout grids.)I updated the example in
examples/radial_slideto addto output h, hm, and eta at 100 fgout frames.
There's also a new
pyvista_plotting.pymodule to make a 3D plot using PyVista, with the surface colored by mass fraction (hm / h).A few frames shown below:
