[FLASH-USERS] 3T MHD with multi species EOS tables - negative internal energy

[Mario Manuel]

Hello Everyone,

I have been trying to run an MHD simulation of a magnetized, laser-generated plasma. The original, non magnetized simulations, were based off of the LaserSlab example and I was able to run that fine. I then created an MHD simulation based on the FieldLoop example, but incorporating all of the necessary modules for the 3T and laser packages. I have a running simulation and the initial outputs look fine, but the simulation only runs 10-15 time steps before crashing.

When I run in the full MHD mode, whether I use eos_tab or eos_gam, the error that eventually crashes it is:

 DRIVER_ABORT: [hy_uhd_ragelike] Negative 3T internal energy, CHECK LOG

however, before every step I get the overflow error from multiple processors:
…
 Overflow! Mesh PE          24 , gettingFrom           1 , recvCount          60 , max particles       20000
 Overflow! Mesh PE          25 , gettingFrom           1 , recvCount          12 , max particles       20000
 Overflow! Mesh PE          24 , gettingFrom           2 , recvCount          69 , max particles       20000
…
but this does not seem to crash the run, it just seems odd (computationally unhealthy?)

I have tried changing the resolution of the simulation and set my cfl=0.2, but the negative internal energy error still comes up. I increased ed_maxRayCount to 40000 to see if I could get rid of the Overflow! error and all that changes is “… max particle   40000”. 

I have tried a different polystyrene table (both tables from the original LaserSlab example), and it still crashes with a ’Negative 3T internal energy' error. 

I changed eosModeInit from “dens_temp_gather” to “dens_ie_gather” and now the internal energy error is gone, but a different EOS error pops up ending in:

DRIVER_ABORT: [Eos] Error: too many iteration in Newton-Raphson

I have increased eos_maxNewton=2000 and set eos_tolerance=1.0e-07, but this doesn’t seem to fix anything.

I tried the ForcedHydroLimit=.true. scenario and I still end up with the same error behavior.


—> With the same geometric setup and EOS tables, the previous simulation using the +uhd3t solver seems to run okay for the most part. However, I have not gotten any version of the simulation to run with the +usm3t solver even when I use the ForcedHydroLimit. 


Any and all suggestions are much appreciated. :)

Cheers,
Mario Manuel






______________________________________
                                                                                                                                
Dr. Mario Manuel
Einstein Postdoctoral Fellow, Smithsonian Astrophysical Observatory
mmanuel at umich.edu
617-899-8381

University of Michigan
Space Research Building
2455 Hayward Street
Ann Arbor, MI 48109
Office: 734-764-4585
______________________________________

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20150403/d037d132/attachment.htm>