[FLASH-USERS] hydrostatic models in flash

[Ji Suoqing]

Hi Dongwook,

> I would like to understand more on your
> statement "2.5d (and hence, the unsplit solvers)". Can you explain this to
> me more?

I guess Phil mentioned this because the split PPM solver does not support 2.5-D with rotation (v_phi is set to zero by the code), consequently he uses the unsplit solver.

To Phil,

If you still have problem in unsplit solver, you could try split solver again: to add a centrifugal term caused by rotation by yourself, which enables you to simulate a rotational case.

Best wishes,
Suoqing

On Jan 3, 2013, at 4:40 PM, dongwook at flash.uchicago.edu wrote:

> Hi Phil,
> 
>> Hi everyone,
>> 
>> I'm working with the new flash 4 in 2.5 d simulating a white dwarf at
>> sits in hydrostatic balance; the white dwarf has significant rotation
>> and hence the 2.5 d.  However, I am finding that when I put in a
>> hydrostatic white dwarf, the white dwarf initially oscillates a little,
>> but soon these oscillations grow to be very violent over a few hundred
>> seconds.  Obviously there is a source of energy that is being tapped.
>> 
>> Looking at Zingale et al (2002, ApJ 143, 539), it seems that one thing
>> might be that these oscillations arise from the imperfect cancellation
>> between pressure and gravity.  What he did was to subtract out \rho g
>> from pressure gradient before calculating the state.  This was
>> implemented in the PPM code and can be turned on with the
>> ppm_modifystates = .true. flag.
>> 
>> As I am using 2.5 d (and hence, the unsplit solvers), I don't have this
>> same luxury.
> 
> You can still check your simulations runs as you expected by using the
> split PPM solver for the testing purpose if needed. But in order to
> understand your concern correctly, I would like to understand more on your
> statement "2.5d (and hence, the unsplit solvers)". Can you explain this to
> me more?
> 
>> However, as I am reading FLASH, it seems that this
>> cancellation may be built in.  In particular, for the unsplit solvers,
> 
> Just to understand correctly: You say "unsplit solvers". And I am guessing
> you are using the unsplit "hydro" solver, rather than the MHD one,
> correct?
> 
>> it seem that gravity is used to calculate the various fluxes.  But I am
>> not sure.  Is there some combination of use_gravHalfUpdate,
>> use_gravConsv, or use_GravPotUpdate in the unsplit solver that would
>> have the same effect as ppm_modifystates = .true. in the split PPM solver?
> 
> Such implementation of ppm_modifistates=.true. in the split PPM solver has
> not been implemented in the unsplit (hydro & MHD) solvers, but it can
> certainly be done easily.
> 
> As for those gravity switches, the choices you want to use are the
> following (although they are not giving you the same effect as
> ppm_modifystates=.true.):
> 
> use_gravHalfUpdate = .true.
> use_gravConsv = .false.
> use_gravPotUpdate = .false.
> 
> If you wish you can send me your log file and flash.par, I will have a
> look at them.
> 
> Thanks,
> Dongwook
> 
>> Cheers,
>> Phil
>> 
>> 
> 
> =========================================
> Dongwook Lee, Ph.D., Research Scientist
> The Flash Center for Computational Science
> The University of Chicago
> 5747 S. Ellis Ave., Room 319
> Chicago, IL 60637
> (773) 834-6830