Problem generating contact file with three refined grids

[marcmestres]

I am trying to generate the contact file for a setup with three telescoping grids. Grid A fully contains grid B, which in turn fully contains grid C, similarly to the ADE nesting in the Lake_jersey test case (which I have managed to replicate). Both nested grids have been generated using the coarse2fine routine, and the refine factors are 5 (grid A -> grid B) and 3 (grid B -> grid C); the boundaries of the different grids do not touch. Following a step-by-step approach, I have successfully created contact files for A+B nesting, and B+C nesting, and both configurations have run smoothly. Now, when I try to create the full A+B+C contact file I run into trouble. For this configuration I believe I should have 4 contact regions (A->B, B->A, B->C and C->B), but the contact routine gives me 6, including regions A->C and C->A, and ends up failing to generate the file.
I have looked into the grid_connections matlab routine and I see that, for some reason, at least one of the edge boundary vectors defining the connection between grids A and C is not zero, so it assumes that the A+C contact region is valid, but I cannot discover why. Any help will be truly appreciated.
Thanks.

[wilkin]

I just took a look at my 3-level nested setup and I agree you should have

Code: Select all

netcdf contactfile_doppio_pioneer_array_v3 {
dimensions:
        Ngrids = 3 ;
        Ncontact = 4 ;

In my case I have a factor of 3 refinement at both levels. Would it be trouble to test whether this is a bug with refinement factors - i.e. could you re-run coarse2fine with factors of 3 in both just to see whether the contact file processes correctly. It would rule out (or in) a certain class of possible bugs.

[marcmestres]

Thanks John. I have followed your suggestion and repeated the nesting using a factor of 3 for both refinements, but the problem remains. I can bypass it by manually invalidating the contact regions between the finest and coarsest meshes in the grid_connections routine, but it still bothers me since it is likely to be due to some silly mistake on my side.

[Vera]

hello~
I want to know that if I have two nested grids,can I set different N(number of vertical layers)for them?
Thanks in advance.

[arango]

Hi Marc and John,

I removed the restriction of having the noncoincident grid points in the indirect connectivity between grids for telescoping applications. See the following trac ticket for more information.

Please update. Your particular set-up should work now.

[marcmestres]

Hernan and John

Thank you very much. Will try it as soon as possible.

Regards