Restart run with different AOA

hyunko · October 26, 2019, 22:53

Hi everyone.

I am using SU2 to simulate flight vehicle with different altitude, mach, aoa, aoss.
Now I have completed the simulation for one particular case. Such as aoa=0, aoss=0. (case0)
I want to use case0 as an initial condition for next run.

I am trying the solution of case0 as a restart file, and change cfg file with different aoa.

Then, Does SU2 reflect the boundary conditions correctly as I want?

Thanks in advance.

aero_amit · October 27, 2019, 04:49

It is likely to work and you may save some computational efforts. You already may be doing this - copy restart_flow.dat to solution_flow.dat than use
RESTART_SOL= YES in the cfg file. Change AOA and SIDESLIP_ANGLE and run it.

Solution will start changing from the boundaries to interior over the iterations.

Share the advantage you obtain by using a different aoa initial solution.

pcg · October 27, 2019, 06:06

Hi Hyun,

It depends on the version, I think there is a bug in v6.2 that causes the angle of attack to be read from the restart file.
I just tested with the current develop branch (soon to be v7.0) and that issue seems to have been solved.
If you are unsure whether the solver is taking your inputs correctly just compare the residuals on the last iteration before the restart file is created and on the first iteration of your new run, the residuals should be very different.

Cheers,
Pedro

aero_amit · October 27, 2019, 06:07

Quote:

Originally Posted by pcg

Hi Hyun,

It depends on the version, I think there is a bug in v6.2 that causes the angle of attack to be read from the restart file.
I just tested with the current develop branch (soon to be v7.0) and that issue seems to have been solved.
If you are unsure whether the solver is taking your inputs correctly just compare the residuals on the last iteration before the restart file is created and on the first iteration of your new run, the residuals should be very different.

Cheers,
Pedro

I was not aware of this bug. Great.

hyunko · October 27, 2019, 06:14

Quote:

Originally Posted by pcg

Hi Hyun,

It depends on the version, I think there is a bug in v6.2 that causes the angle of attack to be read from the restart file.
I just tested with the current develop branch (soon to be v7.0) and that issue seems to have been solved.
If you are unsure whether the solver is taking your inputs correctly just compare the residuals on the last iteration before the restart file is created and on the first iteration of your new run, the residuals should be very different.

Cheers,
Pedro

Thank you for your answer.

I am using 6.2 and think some bug on restart run with different Bc.

I will wait until v7.0 and run the simulation without restart run.

Thanks

hyunko · November 4, 2019, 12:06

Following input statement may solve this problem in v6.2.

DISCARD_INFILES= YES

Additionally, plz read this. Will help.

Modification of a cfg parameter in restart

I’m not tested yet. Back after testing.

October 26, 2019, 22:53	Restart run with different AOA	#1
hyunko Member Hyun Ko Join Date: Jun 2009 Posts: 32 Rep Power: 16	Hi everyone. I am using SU2 to simulate flight vehicle with different altitude, mach, aoa, aoss. Now I have completed the simulation for one particular case. Such as aoa=0, aoss=0. (case0) I want to use case0 as an initial condition for next run. I am trying the solution of case0 as a restart file, and change cfg file with different aoa. Then, Does SU2 reflect the boundary conditions correctly as I want? Thanks in advance.

October 27, 2019, 04:49		#2
aero_amit Member Amit Join Date: May 2013 Posts: 85 Rep Power: 12	It is likely to work and you may save some computational efforts. You already may be doing this - copy restart_flow.dat to solution_flow.dat than use RESTART_SOL= YES in the cfg file. Change AOA and SIDESLIP_ANGLE and run it. Solution will start changing from the boundaries to interior over the iterations. Share the advantage you obtain by using a different aoa initial solution. Last edited by aero_amit; October 27, 2019 at 11:35.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
cannot run naca0012 case using JST	paulocaveman	SU2	0	December 26, 2016 01:08
Restart to adjoint for Unsteady Turbulence Cylinder flow	yweng3	SU2	0	October 7, 2016 14:50
Restart run of dsmcFoam from nonzero time	mechy	OpenFOAM Running, Solving & CFD	2	April 28, 2016 02:24
Restart FSI run on ANSYS CFX	lingdeer	CFX	3	October 10, 2011 09:47
Restart Residual	diegon	OpenFOAM Running, Solving & CFD	0	December 6, 2006 08:34

October 27, 2019, 06:06		#3
pcg Senior Member Pedro Gomes Join Date: Dec 2017 Posts: 465 Rep Power: 13	Hi Hyun, It depends on the version, I think there is a bug in v6.2 that causes the angle of attack to be read from the restart file. I just tested with the current develop branch (soon to be v7.0) and that issue seems to have been solved. If you are unsure whether the solver is taking your inputs correctly just compare the residuals on the last iteration before the restart file is created and on the first iteration of your new run, the residuals should be very different. Cheers, Pedro

November 4, 2019, 12:06		#6
hyunko Member Hyun Ko Join Date: Jun 2009 Posts: 32 Rep Power: 16	Following input statement may solve this problem in v6.2. DISCARD_INFILES= YES Additionally, plz read this. Will help. Modification of a cfg parameter in restart I’m not tested yet. Back after testing.