[hilmialf]

Hello.


As far as I understand, MUSCL is used for higher order reconstruction and thus might somehow spurious oscillations in the region where discontinuity exists. To overcome this, one might add slope limiter to include a mechanism such that it switch to lower order scheme in the discontinuities.



My question is, I somehow found a strange configuration in one of the testcases as is attached below. I am now not really sure about my understanding thus could please someone explain a bit about it?


Here it is written for flow numerical method that it turns MUSCL_FLOW on, but SLOPE_LIMITER_FLOW is set to NONE. Does that mean the following lines specifying the VENKAT_LIMITER_COEFF become useless?


Moreover, for turbulent numerical method, it does not use MUSCL_TURB yet it somehow put SLOPE_LIMITER_TURB. Does this slope limiter affect then?


I hope someone can correct my understanding.


link to the cfg file: https://github.com/su2code/SU2/blob/...ST_RAE2822.cfg




Best regards,
Hilmi

[aero_amit]

JST scheme is a central scheme. Here two dissipation terms (2nd order and 4th order) are used. 4th order dissipation is meant for suppressing the high frequency oscillations. Near to strong gradients like shock 2nd order dissipation becomes active and suppresses the overshoots (you can read the details online)
I think you don't need a limiter here unlike other second order cases. JST is a central scheme (second order accurate supplemented by dissipation terms with appropriate switches).

Regarding second part when MUSCL (MUSCL_TURB=No) is off, it won't use limiter as shown in your. cfg file.

Best
Amit

[hilmialf]

Dear Amit,


Thank you for your reply. It is really helpful.

If I may, I would like to also ask if you know which kind of entropy fixing is used for ROE solver in SU2. Since as far as I understand there are a lot of algorithms proposed by different researcher to fix the entropy for roe solver.


Kind regards,
Hilmi

[aero_amit]

It uses Mavriplis entropy correction. You can check the numeric_direct_mean.cpp

[Shenren_CN]

Originally Posted by aero_amit

It uses Mavriplis entropy correction. You can check the numeric_direct_mean.cpp

Hi Amit,

Is there a publicly available reference on the so-called Mavriplis entropy correction?

Regards,
Shenren