[FLASH-USERS] Problem with applyBCEdge
Klaus Weide
klaus at flash.uchicago.edu
Wed Dec 2 13:35:27 EST 2009
James,
Thanks for the additional info. I still don't have a good idea about what
might cause these artifacts, but here are some ideas.
It seems you suspect the looping indices in the
Grid_bcApplyToRegionSpecialized.F90 file, and maybe you are right, but if
there was something wring in the logic there, it should also in other
simulations. Is your setup doing anything that might be a bit unusual?
For example, are your cell not (logical) cubes of cells? What I mean is
NXB == NYB == NZB not true?
On Tue, 1 Dec 2009, James Guillochon wrote:
> REQUIRES Driver
> REQUIRES physics/Hydro
> REQUIRES physics/Eos/EosMain/Helmholtz
> REQUIRES Simulation/SimulationComposition
> REQUIRES Multispecies
> REQUIRES Grid/GridBoundaryConditions/OneRow
> REQUIRES physics/Gravity/GravityMain/Poisson/Multipole
The combination of the Multipole gravity with your WindTunnel-like BC
seems suspicious to me. Are you sure your setup makes physical sense?
What gravity boundary conditions (grav_boundary_type) are you using?
Whether you use "isolated" or "periodic", my understanding is that
multipole gravity is only meaningful if one assumes zero (or near-zero)
matter density close to the boundaries. Which contradicts having an inflow
boundary.
Can you test your setup (at least for a few steps) with gravity disabled,
to see whether you still have the same kind of artifacts?
Klaus
More information about the flash-users
mailing list