Fully explicit time accurate simulation

nilesh · May 23, 2014, 13:14

Hi,
I want to simulate flow over NACA 0018 using MUSCL (without slope limiter) and "explicit runge-kutta" combination (preferably RK4).
I wish to compute from t=0 to t=50 without using any kind of implicit methods. However the concept of dual time stepping is confusing to me. Dual time stepping uses euler implicit to take the problem to a steady state and then another 3 time level method to proceed to the next time step.
I am into calculating "time-accurate" solutions in which I do not use any turbulence model and explicitly compute upto an estimated non-dimensional time until the flow has fully developed. However I am unable to do so in SU2 as the config file settings confusing me. Please provide me the config file settings and solution writing settings.
Start time t=0
End time t=50
time step delt=10^-5
solution write interval: After 1000 steps.

Thanks.

nilesh · May 25, 2014, 08:09

time integration=rk explicit
internal iterations=0
multigrid level=0

But the results aren't good. Drag prediction is 100 times higher.

demanosalvas · June 5, 2014, 15:20

Nilesh,

If you are interested on learning more about the details of the dual time stepping, you can take a look at "Time dependent calculations using multigrid, with applications to unsteady flows past airfoils and wings" by Prof. Antony Jameson.

With regards of the time integration schemes, there is the option to use dual time stepping, explicit euler or RK methods. All you need to do is to make sure that the method that you are planning to use is selected in your config file.

To improve the accuracy of your solution, you need to look at your mesh and make sure that it is refined in the places where you expect the highest gradients.

Hope this helps,

David

May 23, 2014, 13:14	Fully explicit time accurate simulation	#1
nilesh New Member nilesh Join Date: Mar 2014 Location: Kanpur / Mumbai, India Posts: 27 Rep Power: 12	Hi, I want to simulate flow over NACA 0018 using MUSCL (without slope limiter) and "explicit runge-kutta" combination (preferably RK4). I wish to compute from t=0 to t=50 without using any kind of implicit methods. However the concept of dual time stepping is confusing to me. Dual time stepping uses euler implicit to take the problem to a steady state and then another 3 time level method to proceed to the next time step. I am into calculating "time-accurate" solutions in which I do not use any turbulence model and explicitly compute upto an estimated non-dimensional time until the flow has fully developed. However I am unable to do so in SU2 as the config file settings confusing me. Please provide me the config file settings and solution writing settings. Start time t=0 End time t=50 time step delt=10^-5 solution write interval: After 1000 steps. Thanks.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Multiple floating objects	CKH	OpenFOAM Running, Solving & CFD	14	February 20, 2019 09:08
time step directories naming issue	Andrea_85	OpenFOAM	3	April 3, 2014 08:38
How to write k and epsilon before the abnormal end	xiuying	OpenFOAM Running, Solving & CFD	8	August 27, 2013 15:33
dynamic Mesh is faster than MRF????	sharonyue	OpenFOAM Running, Solving & CFD	14	August 26, 2013 07:47
pisoFoam with k-epsilon turb blows up - Some questions	Heroic	OpenFOAM Running, Solving & CFD	26	December 17, 2012 03:34

May 25, 2014, 08:09		#2
nilesh New Member nilesh Join Date: Mar 2014 Location: Kanpur / Mumbai, India Posts: 27 Rep Power: 12	time integration=rk explicit internal iterations=0 multigrid level=0 But the results aren't good. Drag prediction is 100 times higher.

June 5, 2014, 15:20		#3
demanosalvas New Member David Manosalvas-Kjono Join Date: Feb 2014 Posts: 25 Rep Power: 12	Nilesh, If you are interested on learning more about the details of the dual time stepping, you can take a look at "Time dependent calculations using multigrid, with applications to unsteady flows past airfoils and wings" by Prof. Antony Jameson. With regards of the time integration schemes, there is the option to use dual time stepping, explicit euler or RK methods. All you need to do is to make sure that the method that you are planning to use is selected in your config file. To improve the accuracy of your solution, you need to look at your mesh and make sure that it is refined in the places where you expect the highest gradients. Hope this helps, David