Stable solution in version 7.0.0, diverging in 7.2.0

Dónal · January 19, 2022, 10:13

I am running a compressible (M=2.5) simulation in SU2 version 7.0.0. In 7.0.0 I am able to get a stable solution running for 1000s of iterations and residuals approach 10^-3. When I try running the simulation in version 7.2.0 it diverges after a small number of iterations. The number of iterations varies with CFL, less than 10 with a CFL of 5 and around 60 with a CFL of 0.05.

I suspect the issue is related to the mesh quality. The model has some sharp edges where small angles and high volume ratios are difficult to avoid. I'm not sure why this would be an issue with the most recent version and not a previous one. I have tried different slope limiters and solvers but the solution still diverges.
Looking at the solution after this short number of iterations, it appears to have just populated the mesh with the freestream values and nothing else.

Any ideas on why this may be happening, or version differences that might be related to this would be greatly appreciated

Medium722.txt

pcg · January 19, 2022, 19:06

At some point we fixed a bug that was freezing the solution in places where the rate of change of the solution was too high.

It was a kind of deadlock between turbulence and mean flow logic that caused the automatic relaxation factor to never increase, and so if it went to zero at any point...

This was probably saving your case, I would try running 1st order first to get rid of initial transients.

January 19, 2022, 10:13	Stable solution in version 7.0.0, diverging in 7.2.0	#1
Dónal New Member Dónal McCaughey Join Date: Sep 2021 Posts: 9 Rep Power: 4	I am running a compressible (M=2.5) simulation in SU2 version 7.0.0. In 7.0.0 I am able to get a stable solution running for 1000s of iterations and residuals approach 10^-3. When I try running the simulation in version 7.2.0 it diverges after a small number of iterations. The number of iterations varies with CFL, less than 10 with a CFL of 5 and around 60 with a CFL of 0.05. I suspect the issue is related to the mesh quality. The model has some sharp edges where small angles and high volume ratios are difficult to avoid. I'm not sure why this would be an issue with the most recent version and not a previous one. I have tried different slope limiters and solvers but the solution still diverges. Looking at the solution after this short number of iterations, it appears to have just populated the mesh with the freestream values and nothing else. Any ideas on why this may be happening, or version differences that might be related to this would be greatly appreciated Medium722.txt

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Solution diverging after updating from 7.0.0 to 7.1.1	Dónal	SU2	1	September 8, 2021 05:07
NACA0012 validation with different OF version	Ivangzp	OpenFOAM Running, Solving & CFD	0	February 17, 2021 11:34
Different version of Fluent, different solution!	Kim yongchul	FLUENT	2	January 14, 2021 00:05
IcoFoam parallel woes	msrinath80	OpenFOAM Running, Solving & CFD	9	July 22, 2007 02:58
HELP! Diverging Solution for Incompressible Prob.	Bob	FLUENT	3	April 16, 2006 19:33

January 19, 2022, 19:06		#2
pcg Senior Member Pedro Gomes Join Date: Dec 2017 Posts: 465 Rep Power: 13	At some point we fixed a bug that was freezing the solution in places where the rate of change of the solution was too high. It was a kind of deadlock between turbulence and mean flow logic that caused the automatic relaxation factor to never increase, and so if it went to zero at any point... This was probably saving your case, I would try running 1st order first to get rid of initial transients.