GYRE 2.0 beta Cep test

[ehsan]

Hi all,
I try to calculate adiabatic frequencies of the bcep.mesa model which ships for free with GYRE bundle; you can find it under the /test/ subdirectory.
To get an ordered mode output in n_pg, I use the following input inlist. However, I have not managed to end up with a reasonable grid set up.

Code: Select all

 &coeffs
    coeffs_type = 'EVOL'
    file = 'bcep.mesa'
    file_format = 'MESA'
    deriv_type = 'FINDIFF'  ! 'MONO'
 / ! end of coeff namelist

 &osc
    l = 1
    outer_bound_type = 'UNNO' !'DZIEM': this is a great option 
    variables_type = 'DZIEM'
    tag = ''
 / ! end of osc namelist

 &num
    n_iter_max = 500
    ivp_solver_type = 'MAGNUS_GL2'
    use_banded = .false.
 / ! end of num namelist

 &scan
    grid_type = 'LINEAR'
    freq_min = 1.0
    freq_max = 10
    n_freq = 100
    freq_units = 'NONE'
 / ! end of scan namelist

  &scan
    grid_type = 'INVERSE'
    freq_min = 0.03
    freq_max = 10
    n_freq = 20000
    freq_units = 'NONE'
 / ! end of scan namelist

 &shoot_grid
!    op_type = 'CREATE_CLONE'
 / ! end of 1st shoot_grid namelist

 &shoot_grid
    op_type = 'RESAMP_DISPERSION' !'RESAMP_THERMAL' 
    n = 10
    s = 0d0
    alpha_osc = 50
    alpha_exp = 10
    alpha_thm = 10
 / ! end of 1st shoot_grid namelist

 
  &recon_grid
    op_type = 'CREATE_CLONE'
    alpha_osc = 50
    alpha_exp = 10
 / ! end of recon_grid namelist

 &recon_grid
    op_type = 'RESAMP_DISPERSION' !'RESAMP_THERMAL' 
    n = 10
    s = 0d0
    alpha_osc = 50
    alpha_exp = 10
    alpha_thm = 10
 / ! end of recon_grid namelist

 &output
    summary_file = 'd2/ad-bcep.h5'
    mode_prefix = 'd2/ad-eig-'
    summary_item_list = 'l,n_pg,n_p,n_g,omega,freq,E,E_norm,W,M_star,R_star,L_star' 
    freq_units = 'HZ'
    !mode_item_list = 
!'M_star,R_star,L_star,l,n_p,n_g,n_pg,omega,freq,E,E_norm,W,beta,x,V,As,U,c_1,Gamma_1,nabla_ad,delta,xi_r,xi_h,Yt_1,Yt_2,phip,dphip_dx,delS,delL,delp,delrho,delT,dE_dx,dW_dx,
!prop_type,K,m,p,rho,T'
 / ! end of output namelist

I have to insist on saying that I have tried several variants of op_type, alpha_osc, alpha_exp, alpha_thm, n, ivp_solver_type, etc.
I attract your attention to the huge n_freq = 20000 that I use (which costs nearly 15 minutes), and still I do not get there.
Can anyone please help me with this?

I also would like to receive email notifications when there is a new post on any of the forum pages.
How do I set up my account? I have all options set to "Yes" in my account settings.

Cheers.
Ehsan.

[rhtownsend]

Hi all,
I try to calculate adiabatic frequencies of the bcep.mesa model which ships for free with GYRE bundle; you can find it under the /test/ subdirectory.
To get an ordered mode output in n_pg, I use the following input inlist. However, I have not managed to end up with a reasonable grid set up.

Code: Select all

 &coeffs
    coeffs_type = 'EVOL'
    file = 'bcep.mesa'
    file_format = 'MESA'
    deriv_type = 'FINDIFF'  ! 'MONO'
 / ! end of coeff namelist

 &osc
    l = 1
    outer_bound_type = 'UNNO' !'DZIEM': this is a great option 
    variables_type = 'DZIEM'
    tag = ''
 / ! end of osc namelist

 &num
    n_iter_max = 500
    ivp_solver_type = 'MAGNUS_GL2'
    use_banded = .false.
 / ! end of num namelist

 &scan
    grid_type = 'LINEAR'
    freq_min = 1.0
    freq_max = 10
    n_freq = 100
    freq_units = 'NONE'
 / ! end of scan namelist

  &scan
    grid_type = 'INVERSE'
    freq_min = 0.03
    freq_max = 10
    n_freq = 20000
    freq_units = 'NONE'
 / ! end of scan namelist

 &shoot_grid
!    op_type = 'CREATE_CLONE'
 / ! end of 1st shoot_grid namelist

 &shoot_grid
    op_type = 'RESAMP_DISPERSION' !'RESAMP_THERMAL' 
    n = 10
    s = 0d0
    alpha_osc = 50
    alpha_exp = 10
    alpha_thm = 10
 / ! end of 1st shoot_grid namelist

 
  &recon_grid
    op_type = 'CREATE_CLONE'
    alpha_osc = 50
    alpha_exp = 10
 / ! end of recon_grid namelist

 &recon_grid
    op_type = 'RESAMP_DISPERSION' !'RESAMP_THERMAL' 
    n = 10
    s = 0d0
    alpha_osc = 50
    alpha_exp = 10
    alpha_thm = 10
 / ! end of recon_grid namelist

 &output
    summary_file = 'd2/ad-bcep.h5'
    mode_prefix = 'd2/ad-eig-'
    summary_item_list = 'l,n_pg,n_p,n_g,omega,freq,E,E_norm,W,M_star,R_star,L_star' 
    freq_units = 'HZ'
    !mode_item_list = 
!'M_star,R_star,L_star,l,n_p,n_g,n_pg,omega,freq,E,E_norm,W,beta,x,V,As,U,c_1,Gamma_1,nabla_ad,delta,xi_r,xi_h,Yt_1,Yt_2,phip,dphip_dx,delS,delL,delp,delrho,delT,dE_dx,dW_dx,
!prop_type,K,m,p,rho,T'
 / ! end of output namelist

I have to insist on saying that I have tried several variants of op_type, alpha_osc, alpha_exp, alpha_thm, n, ivp_solver_type, etc.
I attract your attention to the huge n_freq = 20000 that I use (which costs nearly 15 minutes), and still I do not get there.
Can anyone please help me with this?

I also would like to receive email notifications when there is a new post on any of the forum pages.
How do I set up my account? I have all options set to "Yes" in my account settings.

Cheers.
Ehsan.

Hi Ehsan --

Thanks for the bug report. I can confirm this problem with the latest development version of GYRE. This is indeed very strange behavior, and can be triggered using quite-reasonable parameters. I'm looking into it right now, and I'll let you know as soon as I find something.

One comment: in your input file, you don't need to specify the parameters which take default values. So, for instance, all of the tag='' lines can be removed.

cheers,

Rich

[ehsan]

Thanks Rich,
I will be impatient to know about a fix.
How would I also get a notification when anyone posts anything on the forum, so that I would catch up? Is it possible?

Best regards.
Ehsan.

[rhtownsend]

Thanks Rich,
I will be impatient to know about a fix.
How would I also get a notification when anyone posts anything on the forum, so that I would catch up? Is it possible?

Best regards.
Ehsan.

Go to 'user control panel' -> 'overview' -> 'manage subscriptions'

cheers,

Rich

[rhtownsend]

Ehsan --

Is this one of the stars that doesn't appear to be in hydrostatic equilibrium? I've been encountering difficulty with non-consecutive n for RGB models which turn out to not be quite in HSE. I know you've been looking at the HSE issue in other models -- can you check this one?

cheers,

Rich

[ehsan]

Hi there,
So, I quote myself:

I try to calculate adiabatic frequencies of the bcep.mesa model which ships for free with GYRE bundle; you can find it under the /test/ subdirectory.

This is the bcep.mesa model that you have calculated and comes with the GYRE in the test directory.
So, it could be the same HSE issue. If you refer to Bill's message, he also reproduced my poor case, so maybe a sort of global. However, we checked that in the MESA school, and it was Ok at that time. So, I still do not know maybe some MESA version change is involved.
Anyway, Bill convinced me to use lnPgas as a variable for the EOS solver.

I also have to say that I made some calculations with red giants and had no issue. Maybe this is more on the massive star side than the low-mass stars.
Honestly, I am not going to dig further in it. I just want to switch to lnPgas.

Cheers.
Ehsan.

[rhtownsend]

Hi there,
So, I quote myself:

I try to calculate adiabatic frequencies of the bcep.mesa model which ships for free with GYRE bundle; you can find it under the /test/ subdirectory.

This is the bcep.mesa model that you have calculated and comes with the GYRE in the test directory.
So, it could be the same HSE issue. If you refer to Bill's message, he also reproduced my poor case, so maybe a sort of global. However, we checked that in the MESA school, and it was Ok at that time. So, I still do not know maybe some MESA version change is involved.
Anyway, Bill convinced me to use lnPgas as a variable for the EOS solver.

I also have to say that I made some calculations with red giants and had no issue. Maybe this is more on the massive star side than the low-mass stars.
Honestly, I am not going to dig further in it. I just want to switch to lnPgas.

Cheers.
Ehsan.

Hi Ehsan --

It looks like there are some issues with the Brunt-Vaisala frequency of this model (big spikes), which in turn cause the problems you've been encountering with the mode radial order. If you're interested in pursuing this further, then we should look into fixing the B-V --- perhaps by turning on some diffusive mixing.

cheers,

Rich

[ehsan]

Thanks Rich for this.
I would suggest to show me a MESA inlist, and possibly some diagrams like BV profile, and we re-attempt it.
However, I am not sure if in a real star there would be no sharp variation in BV, specifically in the borders between convective and radiative regions.
So, removing the spike in BV might help GYRE do its job more smoothly, but whether or not that's a correct way to go needs a very careful examination with specific stars.
This thing I am utterly interested in.

Best.
Ehsan.