Re: [FLASH-USERS] Re: MHD hangs

Hi,
Did anyone have a solution on how to use a different MHD solver?
In the user guide it says that I just have to 'require hydro/mhd/hlle' but
this module is not included my version (2.5) of FLASH. Couldn't find it
anywhere else either.

Thanks in advance,
Andreas

-- 
Andreas Ekenbäck                      andreas@irf.se
Ph.D. Student                         http://andreas.irf.se
Swedish Institute of Space Physics    http://www.irf.se
Tel direct: +46 (0)980 79113
On Wed, 10 Aug 2005, Matthias Vigelius wrote:
> Hi !
>
> Thanks for your answer!
>
>>> 2) if my guess is right, that this is a problem of steep gradients or if
>>> this might be something else?
>> 
>> You are correct. The only place where MHD can hang is in mhd_interpolate.
>> The routine tries to minimize gradients when they are too steep and when
>> conditions are particularly severe an infinite loop may occur. This
>> "feature" has been around for a while, and I will fix it in the code.
>> I will send a piece of code to you with the fix in the next couple of
>> days.
>
> That would be great. I kind of fixed with an artificial break condition (a 
> counter) and some useful output information, but this is just a makeshift.
>
>> Note, however, that this typically happens when the code is about to blow
>> up anyway. Just fixing gradients in this case may not be enough. Try also
>> to switch the solver form the default one to HLLE. It is supplied with the
>> latest  release of the code. Also, I would suggest to increase resolution,
>> because dipole fields features strong field curvature.
>
> How can I use the rieman_hlle solver with the mhd module? Including 
> hydro/explicit/split/ppm doesn't work, they are mutually exclusive.
>
>> I would bet that your setup contains quite strong fields. What is plasma
>> beta in your simulation? Dipole field is actually a very difficult
>> configuration to represent in a general purpose code. People write
>> specific code to take care of this configuration. The basic trick is to
>> represent the dipole field analytically and solve for perturbation around
>> this field. The current version of FLASH does not have this method
>> implemented.
>
> The problem does even arise without magnetic fields. Is it possible, that it 
> is due to my contact discontinuity?
>
> Thanks again for your help!
>
> Matthias
>