[FLASH-USERS] Status/capabilities of the 3D MHD module

[Anshu Dubey]

 Greetings,
>
> It would also be a good idea to make a very specific
> announcement to the effect in the manual.
>
> Is there any chance that a MHD solver for the low corona
> will become available in FLASH?

That is not likely to happen anytime soon. However, if you want to
add one, we'll be happy to give you any FLASH specific help you need.

>
> Ken Dere
>
> Timur Linde wrote:
>> Hi Mark,
>>
>> The answer to your question depends on what part of the heliosphere you
>> want to simulate. If you are thinking about lower corona and the
>> transition region with all physics included (heating, low-beta
>> conditions, etc), no the solver in version in 2.5 is not a good one to
>> use. The current solver is written mostly for low-beta, compressible
>> plasmas and it would take too much effort to adapt it to lower corona
>> conditions. As far as I know most people use special purpose codes to
>> study that region. Btw, BATS-R-US is no exception. It is true that the
>> origin of MHD  solvers in BATS-R-US and FLASH are similar but the
>> low-beta version of the MHD solver in BATS-R-US is very different from
>> the one
>> originally described in Powell et al.
>>
>> If you need to study outer heliosphere and the solar wind, the solver
>> should be OK. I have a global heliosphere problem setup that is working
>> with both the Roe and HLLE solvers (the HLLE solver is working but it is
>> not incorporated into version 3.0).
>>
>> Best,
>> Timur
>>
>>
>> Greetings,
>>
>> I'm looking for advice on the 3D MHD module (cartesian geometry).
>> My colleagues and I are thinking about using this as a basis for a
>> heliosphere model and we're trying to decide whether or not this is
>> wise.
>>
>>  From talking with people who have used FLASH and from reading the users
>> list archives I see that people have had some problems with the MHD
>> module.  Some have reported problems with div B conservation which was
>> attributed to a change in the calling sequence for the Poisson solver.
>> Others have reported that the code hangs under certain circumstances
>> which exhibit large gradients.  It was suggested that the HLLE module
>> might be more stable in these circumstances.
>>
>> So, my first question is whether these two problems have been addressed
>> in the most recent version of FLASH.  I downloaded FLASH 2.5 in March.
>> Has the 3D cartesian MHD module been incorporated into version 3.0 yet?
>> Has the Poisson call been fixed and has the HLLE module been fully
>> incorporated into versions 2.5 and/or 3.0?
>>
>> My second question is with regard to Timor's comment on the users list
>> that FLASH is not well suited for low-beta flows.  Timor (or anyone
>> else) can you elaborate on this?  According to the FLASH manual, the
>> default MHD solver is based on the Roe-type Riemann solver described by
>> Powell et al (1999).  As I understand it, this solver was developed with
>> the low-beta heliosphere in mind and forms the basis of the BATS-R-US
>> code used for heliospheric simulations at the University of Michigan.  I
>> can appreciate that the HLLE solver may be more robust but I don't see
>> why either would be intrinsically ill suited for low-beta flows.
>> Also, is there a written description somewhere of the HLLE
>> method/algorithm?
>>
>> Is a heliosphere (low-beta) model based on FLASH doomed to fail without
>> a substantial re-write of the MHD module?
>>
>> Any insight is appreciated.
>>
>> - Mark Miesch
>>
>
> --
> _______________________
> Ken Dere
> Research Professor of Solar Physics
> MS 6A2
> Department of Computational and Data Sciences
> College of Science
> George Mason University
> 4400 University Drive
> Fairfax VA 22030
> 703-993-4555
> 703-993-1993 (FAX)
> kdere at gmu.edu
>