Adding arbitrary (up to 50 of each) active and scalar fields #408
Conversation
After resolving conflicts and attempting to move code appropriately this compiles but has not been tested.
…sing segfaults. Note that the additional scalar field is currently mandatory.
Fix an error in bcs for svar and allow initial conditions to be specified (only through generic input). Temporarily add a test for debugging purposes.
…tial Conditions. Fixes segfault found using intel compiler.
…ons of all the parameters. c(12) has now been taken for the buoyancy coefficient.
…ve or passive fields. A limited number of output quantity codes are implemented currently. For active fields these begin at 10000 and increment by 200 for each subsequent field. For passive fields these begin at 20000 and also increment by 200. main_input parameters are either indicated by chi_a or chi_p (active or passive). A test is included to demonstrate the functionality on 2 active and passive fields, comparing to temperature, which does the exact same thing.
cianwilson
requested review from
feathern and
gassmoeller
and removed request for
gassmoeller
post_processing/reference_tools.py
Outdated
| @@ -12,6 +12,10 @@ class equation_coefficients: | |||
| f_dict = {'density':1, 'buoy':2, 'nu':3, 'temperature':4, 'kappa':5, 'heating':6, | |||
| 'eta':7, 'd_ln_rho':8, 'd2_ln_rho':9, 'd_ln_T':10, 'd_ln_nu':11, 'd_ln_kappa':12, | |||
| 'd_ln_eta':13, 'ds_dr':14} | |||
| c_chi_a_dict = {'diff_fact':1, 'source_fact':2, 'buoy_fact':3} | |||
There was a problem hiding this comment.
This can probably be ignored for now. It was my attempt at starting to deal with custom reference states but it is not complete and not necessarily the right way to do it.
There was a problem hiding this comment.
Changes should be removed before merge (if not fixing custom reference states first)
| @@ -62,11 +62,11 @@ Subroutine Initialize_Benchmark_Equations | |||
| Integer :: neq, nvar,lp, l, nlinks | |||
| Integer, Allocatable :: eq_links(:), var_links(:) | |||
| If (magnetism) Then | |||
| neq = 6 | |||
| nvar = 6 | |||
| neq = 6 + n_active_scalars + n_passive_scalars | |||
There was a problem hiding this comment.
These changes are redundant but may indicate to future developers that these fields need to be taken into account.
…in PDE coefficients rather than in the declaration of the implicit terms. This is intentionally done in only 3 out of the 4 definitions. For custom reference state (not implemented yet) the user will have to get the sign correct themselves. This should ultimately be reflected in the documentation.
|
Not ready to merge yet... still need to add room for extra ra_functions and ra_constants with custom reference states! |
| @@ -0,0 +1,72 @@ | |||
| &problemsize_namelist | |||
There was a problem hiding this comment.
Committed by mistake?
| @@ -0,0 +1 @@ | |||
| ../c2001_case0_input_pscalar | |||
There was a problem hiding this comment.
Whole directory committed by mistake?
post_processing/reference_tools.py
Outdated
| @@ -12,6 +12,10 @@ class equation_coefficients: | |||
| f_dict = {'density':1, 'buoy':2, 'nu':3, 'temperature':4, 'kappa':5, 'heating':6, | |||
| 'eta':7, 'd_ln_rho':8, 'd2_ln_rho':9, 'd_ln_T':10, 'd_ln_nu':11, 'd_ln_kappa':12, | |||
| 'd_ln_eta':13, 'ds_dr':14} | |||
| c_chi_a_dict = {'diff_fact':1, 'source_fact':2, 'buoy_fact':3} | |||
There was a problem hiding this comment.
Changes should be removed before merge (if not fixing custom reference states first)
src/Physics/Fields.F90
Outdated
| n_variables = 6 | ||
| n_equations = 6 + n_active_scalars + n_passive_scalars | ||
| n_variables = 6 + n_active_scalars + n_passive_scalars | ||
| if (n_active_scalars>0) then |
There was a problem hiding this comment.
Can you change the ">" to ".gt." -- for FORTRAN's sake.
src/Physics/Sphere_Linear_Terms.F90
Outdated
| @@ -207,6 +235,11 @@ Subroutine Load_Linear_Coefficients() | |||
| amp = -ref%Buoyancy_Coeff | |||
| Call add_implicit_term(peq, tvar, 0, amp,lp, static = .true.) ! Gravity --- Need LHS_Only Flag | |||
|
|
|||
| do i = 1, n_active_scalars | |||
| amp = ref%chi_buoyancy_coeff(i,:) | |||
There was a problem hiding this comment.
The negative in the T term above is due to the decomposition into stream functions. Any negative signs associated with assumptions concerning light vs. heavy elements of Chi should be handled inside the buoyancy coefficient itself. So, here, need to remove negative sign and make appropriate adjustments in PDE_Coefficients.F90.
No description provided.