CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Software User Forums > SU2

convergence issue for transonic turbulent case

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   February 25, 2015, 12:38
Default convergence issue for transonic turbulent case
  #1
New Member
 
sam
Join Date: Sep 2012
Posts: 6
Rep Power: 12
aeroiitkgp is on a distinguished road
Hello,

I am new to SU2 and might ask very basic questions. I tried reading as much posts as possible, but didn't find what I am looking for.

I am trying to run a transonic case with SA model on an aircraft. The mesh I am using has a Y+ around 1. The problem runs very well when I choose first order for convective as well as turbulence terms. If any of the two I switch to second order, the problem diverges almost instantaneously.

I ran the case with first order for almost 11200 iterations. By this time the force and moment coefficients had converged well. But if I switch to second order for even the convective terms it diverges from the very first iteration.

The multigrid is also turned off. I tried turning on the multigrid without switching to second order, but that too diverged the solution after some 20 iterations.

I have attached the configuration file and the residual convergence plots for first order with and without multigrid. Please help me solve this issue.

Note: The angle of attack in the config file is 20, but the same thing happens for even very low angles of attack.

Other questions with some relation to the problem:

1. After I restart, the iteration number starts from 0 again and it overwrites earlier history.dat file. (For restart I had copied restart_flow.dat to solution_flow.dat). But I can see that the solution has indeed restarted as the coefficients value is from the last iteration.

2. The values of primary variables I get after non-dimesionalization is not 1.
Pressure is 0.9985....., velocity is 1.13, and so on. Please list down the equations used to derive other quantities from Mach Number, Temperature, angles and Reynolds Number if possible. What are the other possible input parameters, like, directly specifying the pressure, temperature and velocity components?
Attached Images
File Type: jpg residual-firstorder.jpg (35.5 KB, 29 views)
File Type: jpg residual-restart-firstorder-1200iter.jpg (29.0 KB, 20 views)
File Type: jpg residual-restart-firstorder-multigrid.jpg (33.0 KB, 21 views)
Attached Files
File Type: txt trialcase.cfg.txt (7.9 KB, 5 views)
aeroiitkgp is offline   Reply With Quote

Old   March 5, 2015, 20:37
Default
  #2
Super Moderator
 
Francisco Palacios
Join Date: Jan 2013
Location: Long Beach, CA
Posts: 404
Rep Power: 14
fpalacios is on a distinguished road
Hi, thanks for using SU2,

Have you tried to reduce the CFL number? sometimes it is also important to play with
LIMITER_COEFF= 0.3

I'm not sure what version of SU2 are you using but please note that we have improved the stability of the code in the latest SU2 releases.

Best Regards,

Francisco
SU2 lead developer
fpalacios is offline   Reply With Quote

Old   March 5, 2015, 21:31
Default
  #3
New Member
 
sam
Join Date: Sep 2012
Posts: 6
Rep Power: 12
aeroiitkgp is on a distinguished road
Thank you Francisco,

I am using the latest build of SU2.

I will try to change the limiter coeff as suggested by you. For CFL, how much should I start with? I am using an implicit formulation. Also should I use CFL ramping?
aeroiitkgp is offline   Reply With Quote

Old   March 9, 2015, 17:11
Default
  #4
New Member
 
Frederik
Join Date: Apr 2010
Posts: 18
Rep Power: 14
fchan is on a distinguished road
Hi,

I'm having similar issues with a second order simulation, although mine is unsteady and inviscid. It's subsonic (M=0.5), so I doubt it has anything to do with the turbulent solver.

Mine seems to work fine with first order (haven't tested it more than a couple of hundred time steps). However, I can't get more than 10 time steps done with 2nd order spatial. I have limited the number of iterations per time steps, so it eventually diverges. Additionnally, I'm not using MG nor CFL slopes (CFL=0.5 for now). I'm also using implicit time discretization.

@fpalacios: There are a couple of parameters for the slope limiter in SU2 and I'm having trouble identifying which parameters they relate to Venkatakrishnan's paper (AIAA 93-0880). Could you please elaborate on these:
  • REF_ELEM_LENGTH= 0.1 (default)
  • LIMITER_COEFF= 0.3 (default)
  • LIMITER_ITER= 999999 (default)
  • SHARP_EDGES_COEFF= 3.0 (default)
  • REF_SHARP_EDGES= 3.0 (default)
  • SENS_REMOVE_SHARP= NO (default)

Best regards,
Frederik
fchan is offline   Reply With Quote

Old   March 13, 2015, 01:29
Default Second Order Sucessful
  #5
New Member
 
sam
Join Date: Sep 2012
Posts: 6
Rep Power: 12
aeroiitkgp is on a distinguished road
Hi,

By just changing the limiter coeff to 0.3 from 10.0 earlier, I was able to run with second_order_limiter selected as discritization for both convective and turbulent terms. However, still multigrid does not work. All other settings are as shown in the previous configuration file. Kindly help.

It would be nice if we have all the possible parameters of configuration file available in one sample file. I am not able to find such a file.

Also the non-dimensionalization is not leading to unity for many variables. Please tell me how to solve this.

Last edited by aeroiitkgp; March 13, 2015 at 01:30. Reason: spelling mistakes
aeroiitkgp is offline   Reply With Quote

Old   May 12, 2015, 17:44
Default All configuration options
  #6
New Member
 
Santiago Padron
Join Date: May 2013
Posts: 17
Rep Power: 11
Santiago Padron is on a distinguished road
The file config_template.cfg in the main directory of the SU2 code contains all the possible parameters. You can find it here https://github.com/su2code/SU2/blob/...g_template.cfg
Also there is an option for different types of non-dimensionalization.
% Flow non-dimensionalization (DIMENSIONAL, FREESTREAM_PRESS_EQ_ONE,
% FREESTREAM_VEL_EQ_MACH, FREESTREAM_VEL_EQ_ONE)
REF_DIMENSIONALIZATION= DIMENSIONAL
Santiago Padron is offline   Reply With Quote

Reply

Tags
convergence, multigrid, restart, second order

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
pisoFoam turbulent case looks like laminar case inf.vish OpenFOAM Running, Solving & CFD 0 August 16, 2013 04:31
Free surface boudary conditions with SOLA-VOF Fan Main CFD Forum 10 September 9, 2006 13:24
Validation case for turbulent flow Ratan Main CFD Forum 0 October 4, 2005 04:03
Validation case for turbulent flow Ratan Main CFD Forum 0 October 4, 2005 04:02
Turbulent Flat Plate Validation Case Jonas Larsson Main CFD Forum 0 April 2, 2004 11:25


All times are GMT -4. The time now is 09:10.