[FLASH-USERS] Checkboard of Magnetic Monopoles

[Aaron Froese]

Hi Dongwook,

Thanks for responding so quickly.  I hope I can answer your questions to your satisfaction.
 
> (1) What is the grid geometry you're using? I understand that you're
> interested in working on cylindrical implementation at some point and I
> wonder if you are seeing these monopoles on your own cylindrical
> implementation or not. Is this any non-Cartesian implementation?

I am using Cartesian geometry.  (But I am still interested in cylindrical geometry.)
I initially encountered this bug in my spheromak compression problem, but have
recreated it in the OrszagTang problem supplied with flash3.

> (2) What is the dimension of the problem? Is it 2D?

I am running the OrszagTang problem in 2D.  The setup calls are:
./setup magnetoHD/OrszagTang -auto -2d +usm -objdir=OrszagTang
./setup magnetoHD/OrszagTang-Cool -auto -2d +usm -objdir=OrszagTang-Cool
In my first email, I attached plots showing the bug in both 2D and 3D versions
of my problem.

> Have you tried running the same OrszagTang problem WITHOUT including your
> cool module?

DivB looks random and does not exceed 1e-12 when running the OrszagTang problem
exactly as it is distributed with flash3 with the flash.par file included in
the directory.

> In case you still see the monopoles without it, did you need to implement
> any other custom files other than the cool unit?

The only change I make to cause the monopoles is to add that Cool.F90 file to
the directory and include the Radloss module.  I am attaching both simulation
directories.

> In general, the monopole issue you're seeing is mostly caused by the wrong
> initial condition or the boundary conditions of the cell
> interface-centered magnetic fields (facevars in the code).
> Can you please check if your initial divB is small too?

DivB is initially zero for both cases, with and without cooling.  I have
outputted divB at every time step and it is not suddenly appearing at step 1.
It grows slowly and takes about 14 steps to form a clear checker board
pattern when using a cooling rate of 1e-6.

> Do you have any special boundary conditions?

No, I have tried periodic and outflow, but there is no change,
except where the checker board effect wraps at the edges.

>  In general, if you can tell me more about your simulation along with your
>  log file and flash.par, it will be helpful to understand the issue better.

I have included the flash.par and log files in the attachment.  The parameter
file has not been changed from the default found in the OrszagTang directory.

> However, if you are calling ENER_VAR OUTSIDE the hydro unit, for example
> in your Cool.F90, then it only includes SPECIFIC eint and ekin.

I simplified the cooling in the Radloss module to eliminate possible
sources of error, but otherwise the expressions for ENER_VAR remain
as found in the distributed version.

If you need any other information, please let me know.

Thanks,
Aaron
-------------- next part --------------
A non-text attachment was scrubbed...
Name: magnetoHD.tgz
Type: application/x-compressed-tar
Size: 32412 bytes
Desc: magnetoHD.tgz
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20120418/8f653fc9/attachment-0001.bin>