[Users] evolving flat boundary conditions

[Roland Haas]

Hello Ryn,

> I am trying to do a cosmology simulation project using Cactus. For this
> project, I'm evolving an expanding homogeneous fluid in a box (FLRW
> spacetime), and I would like to use flat boundary conditions. I've attached
> the par file I am currently using. However, when I run this, the boundary
> cells do not seem to update at each time step to match the current value of
> the grid cells on the edge, as seen in this plot of the second time step
> output:


The parameter admbase_boundary_condition is only acted on if you set
the evolution method to "static" (see schedule.ccl for routine
ADMBase_Boundaries which is the only one using the parameter).

Since you have a BSSN evolution thorn enabled it will compute the ADM
variables from the BSSN ones. So to set a certain boundary condition on
the ADM variables you need to set McLachlan's options. Eg

ML_BSSN::rhs_boundary_condition     = "scalar" # set RHS to 0 in boundary

ML_BSSN::ML_log_confac_bound = "flat"
ML_BSSN::ML_metric_bound     = "flat"
ML_BSSN::ML_Gamma_bound      = "flat"
ML_BSSN::ML_trace_curv_bound = "flat"
ML_BSSN::ML_curv_bound       = "flat"
ML_BSSN::ML_lapse_bound      = "flat"
ML_BSSN::ML_dtlapse_bound    = "flat"
ML_BSSN::ML_shift_bound      = "flat"
ML_BSSN::ML_dtshift_bound    = "flat"

Though this may or may not actually make things crash with bad values
("flat" being an unusual boundary condition for the metric).

Yours,
Roland

-- 
My email is as private as my paper mail. I therefore support encrypting
and signing email messages. Get my PGP key from http://pgp.mit.edu .