SU2 v.7.0.8 cannot solve expansion region that was solved by using v.6.1.0

aeronautics · February 16, 2021, 08:42

Dear all,

I have been trying to solve a supersonic flow domain around an aircraft with SU2 v.7.0.8 that was solved before with SU2 v6.1.0 and results were validated. I have adapted the .cfg file from v6 to v7 without any change in solvers, multigrid etc.. However, the solution I have obtained by using v7 has yielded non-physical points at the downstream of the aft where Mach number goes up to 1800

. The region behind this non-physical region has been solved almost the same as in v6.

This is the non-physical solution I've got. The aircraft is attached to wind tunnel sting. I've rescaled the Mach interval as 1.6-1.8 in the image. Freestream Mach is 1.7.
wrong_solution.jpg

This is the mesh around the region. Aft of the aircraft (has relatively higher diameter) and sting (has relatively lower diameter) can be seen.
mesh.png

I'm appending the .cfg file.
config_fluid_v7.txt

What am I doing wrong? What could be the possible problem(s) and solution(s)? Any kind of help is appreciated. Thank you all in advance.

Huseyin Emre

pcg · February 16, 2021, 10:36

Some of the things that changed:
Implementation of Euler and Symmetry boundaries, to improve them for viscous flows.
How the CFL adaption works.
And I believe the reference length for limiters was fixed at one, so you might need to multiply the Venkat constant by the reference length manually.
I recommend using the adaptive CFL strategy, even if you don't go higher than 15 it can reduce CFL in more challenging regions and possibly improve stability.
With multigrid I never use damping factors higher than 0.75.

aeronautics · February 16, 2021, 11:40

Quote:

Originally Posted by pcg

Some of the things that changed:
Implementation of Euler and Symmetry boundaries, to improve them for viscous flows.
How the CFL adaption works.
And I believe the reference length for limiters was fixed at one, so you might need to multiply the Venkat constant by the reference length manually.
I recommend using the adaptive CFL strategy, even if you don't go higher than 15 it can reduce CFL in more challenging regions and possibly improve stability.
With multigrid I never use damping factors higher than 0.75.

Thank you Pedro for your answer! I will try your suggestions and then let you know results.

Edit:

I've used adaptive CFL, multigrid damping factors of 0.75 and different limiter coefficients. Despite residuals has decreased about 0.7, I couldn't avoid non-physical solutions.

aeronautics · February 18, 2021, 07:43

I've overcome the issue with non-physical points. I don't know exactly why yet I think the problem was aroused from Roe scheme. I changed convective scheme to AUSM, adjusted multigrid damping factors to 0.75 and activated adaptive CFL then I could be able to get an accurate result.

February 16, 2021, 08:42	SU2 v.7.0.8 cannot solve expansion region that was solved by using v.6.1.0	#1
aeronautics New Member Hüseyin Emre Join Date: Mar 2017 Location: Istanbul Posts: 8 Rep Power: 9	Dear all, I have been trying to solve a supersonic flow domain around an aircraft with SU2 v.7.0.8 that was solved before with SU2 v6.1.0 and results were validated. I have adapted the .cfg file from v6 to v7 without any change in solvers, multigrid etc.. However, the solution I have obtained by using v7 has yielded non-physical points at the downstream of the aft where Mach number goes up to 1800. The region behind this non-physical region has been solved almost the same as in v6. This is the non-physical solution I've got. The aircraft is attached to wind tunnel sting. I've rescaled the Mach interval as 1.6-1.8 in the image. Freestream Mach is 1.7. wrong_solution.jpg This is the mesh around the region. Aft of the aircraft (has relatively higher diameter) and sting (has relatively lower diameter) can be seen. mesh.png I'm appending the .cfg file. config_fluid_v7.txt What am I doing wrong? What could be the possible problem(s) and solution(s)? Any kind of help is appreciated. Thank you all in advance. Huseyin Emre

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Can I use fvOptions to couple a solid region and a fluid region?	titanchao	OpenFOAM Running, Solving & CFD	4	January 14, 2022 08:55
How to let the mesh motion solver just solve a small region near a moving boundary?	zhajingjing	OpenFOAM Running, Solving & CFD	9	April 28, 2016 05:15
Define velocity profile in fully develpoed region to solve energy equation	haghgoo_reza	OpenFOAM Programming & Development	0	September 22, 2013 21:19
Can SU2 solve the blast wave problem?	newmancfd	SU2	2	May 13, 2013 04:33
How can I solve at the interface region	K. Baker	FLUENT	0	April 3, 2007 04:39

February 16, 2021, 10:36		#2
pcg Senior Member Pedro Gomes Join Date: Dec 2017 Posts: 466 Rep Power: 13	Some of the things that changed: Implementation of Euler and Symmetry boundaries, to improve them for viscous flows. How the CFL adaption works. And I believe the reference length for limiters was fixed at one, so you might need to multiply the Venkat constant by the reference length manually. I recommend using the adaptive CFL strategy, even if you don't go higher than 15 it can reduce CFL in more challenging regions and possibly improve stability. With multigrid I never use damping factors higher than 0.75.

February 18, 2021, 07:43		#4
aeronautics New Member Hüseyin Emre Join Date: Mar 2017 Location: Istanbul Posts: 8 Rep Power: 9	I've overcome the issue with non-physical points. I don't know exactly why yet I think the problem was aroused from Roe scheme. I changed convective scheme to AUSM, adjusted multigrid damping factors to 0.75 and activated adaptive CFL then I could be able to get an accurate result.