Non-convergence of transonic RAE2822 airfoil

LaurentvdB · June 25, 2015, 04:32

Hi all,

I think I have a basic question, but somehow I cannot find the problem.

It is as follows. I am running the latest RAE2822 testcase of SU2 (v.4) and the default configuration does converge in about half an hour on my machine. Now when I (only!) change the Mach number to .766 in stead of the default value of .729, SU2 keeps running without converging anymore. The cl and cd do not tend to converge to a fixed value after the maximum amount of iterations have been done. Increasing the amount of iterations and turning on adaption of the CFL-number does not help.

I thought the problem could be that the mesh is adapted to the specific shockwave location of the default Mach number, but after a visualization of the mesh that does not seem to be the case. Currently, I think that I need to scale another parameter, but I cannot come up with which one.

Does anybody know what I do wrong in this case?

Thanks! Laurent

Santiago Padron · July 14, 2015, 16:46

The first thing to try will be turning down the CFL number, then you can also try modifying aspects of the convective numerical method.

LaurentvdB · July 15, 2015, 03:05

I already have an idea what the "error" might be: the mesh. I already lowered the CFL number and changed almost all parameters of the numerical method, but in the situation of mach .766 it turns out that the grid is too course too converge. Choosing a finer mesh resolves the problem.

Thanks for your reply!

June 25, 2015, 04:32	Non-convergence of transonic RAE2822 airfoil	#1
LaurentvdB New Member Laurent van den Bos Join Date: Jun 2015 Posts: 4 Rep Power: 10	Hi all, I think I have a basic question, but somehow I cannot find the problem. It is as follows. I am running the latest RAE2822 testcase of SU2 (v.4) and the default configuration does converge in about half an hour on my machine. Now when I (only!) change the Mach number to .766 in stead of the default value of .729, SU2 keeps running without converging anymore. The cl and cd do not tend to converge to a fixed value after the maximum amount of iterations have been done. Increasing the amount of iterations and turning on adaption of the CFL-number does not help. I thought the problem could be that the mesh is adapted to the specific shockwave location of the default Mach number, but after a visualization of the mesh that does not seem to be the case. Currently, I think that I need to scale another parameter, but I cannot come up with which one. Does anybody know what I do wrong in this case? Thanks! Laurent

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Convergence Problem icoFoam steady flow over an airfoil	Lucas	OpenFOAM Running, Solving & CFD	6	July 12, 2018 15:06
convergence issue for transonic turbulent case	aeroiitkgp	SU2	5	May 12, 2015 16:44
Airfoil analysis convergence problem	mayursindhav	Fluent Multiphase	0	August 27, 2014 09:47
Transonic Aeroelastic CFD code for an Airfoil	Nicholas Amuyedo	Main CFD Forum	8	August 8, 2007 08:25
Simulation of transonic flow over NACA0012 airfoil	MSc Student	Siemens	2	August 9, 2006 13:49

July 14, 2015, 16:46		#2
Santiago Padron New Member Santiago Padron Join Date: May 2013 Posts: 17 Rep Power: 12	The first thing to try will be turning down the CFL number, then you can also try modifying aspects of the convective numerical method.

July 15, 2015, 03:05		#3
LaurentvdB New Member Laurent van den Bos Join Date: Jun 2015 Posts: 4 Rep Power: 10	I already have an idea what the "error" might be: the mesh. I already lowered the CFL number and changed almost all parameters of the numerical method, but in the situation of mach .766 it turns out that the grid is too course too converge. Choosing a finer mesh resolves the problem. Thanks for your reply!