Solving coupled steady and unsteady equation in OpenFOAM

jd87 · July 18, 2020, 18:26

Hello,

I have two equations, the momentum (velocity and pressure) equation which is steady in my case and a convection-diffusion (transport) equation which is unsteady and drives the flow.

Physically, the momentum spreads much faster and becomes steady before there is any effect from the unsteady transport equation. Thus, I want to couple them. I had few thoughts:
1) To use simpleFoam to solve the steady part and then go for the unsteady inside a time loop. However, simpleFoam itself uses a dummy time loop and there goes the issue.
2) Use pimpleFoam to run the momentum equation until steady and using its results solve the transport equation. So, I have made a custom solver for the 2nd case. But how to tell the pimpleFoam momentum equation (the pimple.loop() function) to only exit pimple.loop() after the momentum equation has reached steady state and then go for the unsteady transport equation?

My overall idea is to make the momentum steady and use it to run unsteady transport and loop over these two (moment and transport) until a convergence is obtained for a particular time point. Then I move on to the next time point and do the same. Is there any better way to do that besides point 2?

klausb · July 21, 2020, 12:22

Hmm...

read this page carefully: https://openfoamwiki.net/index.php/O...hm_in_OpenFOAM

You can run pimpleFoam in PIMPLE and PISO mode. PIMPLE uses SIMPLE in each timestep. Practically, do you mean, you want to switch between PIMPLE and PISO mode in each timestep?

Klaus

jd87 · July 22, 2020, 12:36

Thanks for your reply. Yes, I have already looked in the PIMPLE ago. PIMPLE uses a dummy timestep to reach a steady state (that uses SIMPLE), however as it runs that dummy time step to obtain a steady state value, the equation which is actually unsteady also changes its value. So I have two equations one steady momentum (U,p variables) and one transport (say, T as variable) equation. ideally, the term ddt(T) in the unsteady transport equation is broken as (T - T_oldTime)/dt. Now, in changing the dummy time step, the value of T_oldTime will also change while I want it to be constant and want it to only change when the momentum equation (U,p variables) converge along with T for a particular time step. So, the algo goes like:

find converged value for U, p and T equations for a particular incremental dt where U and p have reached their steady state then move on to change the time step dt.

Thanks again.

July 18, 2020, 18:26	Solving coupled steady and unsteady equation in OpenFOAM	#1
jd87 New Member Jayabrata Dhar Join Date: Nov 2018 Posts: 17 Rep Power: 8	Hello, I have two equations, the momentum (velocity and pressure) equation which is steady in my case and a convection-diffusion (transport) equation which is unsteady and drives the flow. Physically, the momentum spreads much faster and becomes steady before there is any effect from the unsteady transport equation. Thus, I want to couple them. I had few thoughts: 1) To use simpleFoam to solve the steady part and then go for the unsteady inside a time loop. However, simpleFoam itself uses a dummy time loop and there goes the issue. 2) Use pimpleFoam to run the momentum equation until steady and using its results solve the transport equation. So, I have made a custom solver for the 2nd case. But how to tell the pimpleFoam momentum equation (the pimple.loop() function) to only exit pimple.loop() after the momentum equation has reached steady state and then go for the unsteady transport equation? My overall idea is to make the momentum steady and use it to run unsteady transport and loop over these two (moment and transport) until a convergence is obtained for a particular time point. Then I move on to the next time point and do the same. Is there any better way to do that besides point 2? neelarun likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
iterations process for steady solver and time dependence for unsteady	serguei	Main CFD Forum	22	May 10, 2018 11:58
Simulating an unsteady phenomenon with a steady model	JochemGrietens	Main CFD Forum	2	April 13, 2016 12:51
Steady and Unsteady nature	Natali_Beginner	FLUENT	15	April 21, 2013 13:43
steady or unsteady? (in dpm)	winnie	FLUENT	1	April 28, 2003 12:30
steady or unsteady equation?	Harry Qiu	FLUENT	0	April 2, 2001 07:35

July 21, 2020, 12:22		#2
klausb Senior Member Klaus Join Date: Mar 2009 Posts: 281 Rep Power: 22	Hmm... read this page carefully: https://openfoamwiki.net/index.php/O...hm_in_OpenFOAM You can run pimpleFoam in PIMPLE and PISO mode. PIMPLE uses SIMPLE in each timestep. Practically, do you mean, you want to switch between PIMPLE and PISO mode in each timestep? Klaus

July 22, 2020, 12:36		#3
jd87 New Member Jayabrata Dhar Join Date: Nov 2018 Posts: 17 Rep Power: 8	Thanks for your reply. Yes, I have already looked in the PIMPLE ago. PIMPLE uses a dummy timestep to reach a steady state (that uses SIMPLE), however as it runs that dummy time step to obtain a steady state value, the equation which is actually unsteady also changes its value. So I have two equations one steady momentum (U,p variables) and one transport (say, T as variable) equation. ideally, the term ddt(T) in the unsteady transport equation is broken as (T - T_oldTime)/dt. Now, in changing the dummy time step, the value of T_oldTime will also change while I want it to be constant and want it to only change when the momentum equation (U,p variables) converge along with T for a particular time step. So, the algo goes like: find converged value for U, p and T equations for a particular incremental dt where U and p have reached their steady state then move on to change the time step dt. Thanks again.