[FLASH-USERS] Fwd: Supersonic Turbulence using the PPM method

[Tristan]

Hi, Christoph:
	Thank you for your advice, I tried your advice to reduce the cfl and set nriem =1000, it works better than before.But the artificial viscosity of PPM is a little large. And the larger mach number couldn’t be achieved yet.  I heard that the artificial viscosity of un_split method is little. But I can only run  StirFromFile at 3 mach number with the un_split method. If I improved the st_energy, the program broke down and output “Nan”
I list some features I used in the un_split method:

```
order		= 2      # Interpolation order (first/second/third/fifth order)
slopeLimiter    = "limited"   # Slope limiters (minmod, mc, vanLeer, hybrid, limited)
LimitedSlopeBeta= 1.     # Slope parameter for the "limited" slope by Toro
charLimiting	= .true. # Characteristic limiting vs. Primitive limiting

use_avisc	= .false. # use artificial viscosity (originally for PPM)
#cvisc		= 0.1     # coefficient for artificial viscosity
use_flattening	= .true. # use flattening (dissipative) (originally for PPM)
use_steepening	= .false. # use contact steepening (originally for PPM)
use_upwindTVD	= .false. # use upwind biased TVD slope for PPM (need nguard=6)

#	II. RIEMANN SOLVERS:
RiemannSolver	= "HLLC"       # Roe, HLL, HLLC, LLF, Marquina
entropy         = .false.     # Entropy fix for the Roe solver

#	III. STRONG SHOCK HANDELING SCHEME:
shockDetect	= .true.     # Shock Detect for numerical stability
shocklowerCFL   = .true.
```

I also attached the flash.par and the output file for you to check my work.



Our purpose is to run the code at 10 mach number. I wonder if you changed the Solver yourself to make the code stable or just use solver in the flash4. If the latter, could you please give some advice about the features and the solver you used?

If everyone have some suggestion about the high mach number using the un_split method in the Stir unit. Please connect me!
Thank you all!
Best wishes!


> 下面是被转发的邮件：
> 
> 发件人: Christoph Federrath <christoph.federrath at gmail.com>
> 主题: 回复：[FLASH-USERS] Supersonic Turbulence using the PPM method
> 日期: 2021年6月2日 GMT+8 下午5:47:12
> 收件人: Tristan <juwj at mail2.sysu.edu.cn>
> 抄送: flash-users at flash.uchicago.edu
> 
> 
> Hi Tristan,
> 
> for PPM, you can try nriem = 1000, which helped sometimes (in the past). Also reducing cfl helps. However, there is a limit to those tweaks. I would suggest to try the unsplit solvers. Dongwook Lee implemented a number of features to make them more stable, such as shock detection and automatic reduction of the cfl. We are using a different solver altogether (Waagan et al. 2011, JCP, based on work by Bouchut et al., which is not (yet) in the public flash repository though.
> 
> Best regards,
> Christoph
> 
>> On 2 Jun 2021, at 19:21, Tristan <juwj at mail2.sysu.edu.cn> wrote:
>> 
>> Hello,everyone:
>> 	I am working on a simulation	with high Mach number turbulence using the StirFromFile unit, I using the forcing_generator.F90 to generate the Driving force seed with the Mach at 12(by changed the st_energy). But I did’t know how to choose a suitable Solver for the Stir unit. 
>> 	I tried the split PPM solver, and the two parameters to the flash.par:
>> ```
>> 	use_steepening = .false. # use contact steepening (originally for PPM) 
>> 	hybrid_riemann = .true.
>> ```
>> My program broke down quickly, and the error output is:
>> 
>> Terminating execution.
>> DRIVER_ABORT: Nonconvergence in subroutine rieman
>> application called MPI_Abort(MPI_COMM_WORLD, 1) - process 46
>> 
>> I also attached the flash.par and the output file for you to check my work. If someone have a better Solver choice in Supersonic (Mach above 12) turbulence simulation. Please help me! 
>> Thanks
>> Best wishes!
>> 
>> <flash.par><slurm-4341014.out>
> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20210608/2db530d5/attachment-0002.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: flash.par
Type: application/octet-stream
Size: 4207 bytes
Desc: not available
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20210608/2db530d5/attachment-0001.obj>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20210608/2db530d5/attachment-0003.htm>