Scalar field blows up (only) after restarting the simulation

backscatter · September 15, 2018, 00:32

Hi,

I have a modified form of pimpleFoam where I have added a source/sink term in it (TEqn) to make the temperature field periodic in periodic flows. So far so good. I initialize the T field with a uniform value everywhere and start the simulation. It runs and after few flow-passes, I can obtain a fully developed T-profile. No blowing up at all.

Now when I stop the simulation and try to re-run it, the T field blows up at places where the grid is too fine.

Please note that if I didn't stop the original simulation and had let in run, it won't crash. The trouble is just with stopping and rerunning the simulation with the developed T-profile.
Any ideas how to tackle this issue?
Thanks,

pete20r2 · September 15, 2018, 09:12

Hi Spockner,
Can you share more information, the case and the solver would be ideal if you can share them.
Otherwise, have you got an under-relaxation term on your sink/source? When restarting a case there is often no cached fields (grad(U)) so the initial residuals/ continuity error can be larger in the first few steps than when the same job stopped. This may be causing your active sink/source to overreact.

backscatter · September 16, 2018, 02:24

Quote:

Originally Posted by pete20r2

Hi Spockner,
Can you share more information, the case and the solver would be ideal if you can share them.
Otherwise, have you got an under-relaxation term on your sink/source? When restarting a case there is often no cached fields (grad(U)) so the initial residuals/ continuity error can be larger in the first few steps than when the same job stopped. This may be causing your active sink/source to overreact.

Thanks Peter,

I am attaching the TEqn for you to refer to my source/sink term. I am not able to access my case right now, but to let you know I did not use any under-relaxation for my T field so far. I will try it tomorrow to see if that helps. Please take a look attached TEqn...

pete20r2 · September 16, 2018, 07:56

For good measure, it is common to use a limitTemperature constraint in fvOptions with some limits a little outside of what you're expecting. This can help get past the first few iterations without crashing.

September 15, 2018, 00:32	Scalar field blows up (only) after restarting the simulation	#1
backscatter Member Anonymous Join Date: Aug 2016 Posts: 75 Rep Power: 10	Hi, I have a modified form of pimpleFoam where I have added a source/sink term in it (TEqn) to make the temperature field periodic in periodic flows. So far so good. I initialize the T field with a uniform value everywhere and start the simulation. It runs and after few flow-passes, I can obtain a fully developed T-profile. No blowing up at all. Now when I stop the simulation and try to re-run it, the T field blows up at places where the grid is too fine. Please note that if I didn't stop the original simulation and had let in run, it won't crash. The trouble is just with stopping and rerunning the simulation with the developed T-profile. Any ideas how to tackle this issue? Thanks,

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Howto use scalarCodedSource in fvOptions	hanness	OpenFOAM Running, Solving & CFD	61	November 18, 2022 13:11
whats the cause of error?	immortality	OpenFOAM Running, Solving & CFD	13	March 24, 2021 08:15
Steady simulation does not iterate in Time=1	agustinvo	OpenFOAM Running, Solving & CFD	3	November 19, 2015 05:57
Passive scalar in "frozen" field	mrcsjnssn	FLUENT	3	October 5, 2015 10:41
restarting paused transient simulation using reactingFoam	JMDag2004	OpenFOAM Running, Solving & CFD	1	August 10, 2015 11:15

September 15, 2018, 09:12		#2
pete20r2 Senior Member Peter Baskovich Join Date: Jul 2014 Posts: 127 Rep Power: 12	Hi Spockner, Can you share more information, the case and the solver would be ideal if you can share them. Otherwise, have you got an under-relaxation term on your sink/source? When restarting a case there is often no cached fields (grad(U)) so the initial residuals/ continuity error can be larger in the first few steps than when the same job stopped. This may be causing your active sink/source to overreact.

September 16, 2018, 07:56		#4
pete20r2 Senior Member Peter Baskovich Join Date: Jul 2014 Posts: 127 Rep Power: 12	For good measure, it is common to use a limitTemperature constraint in fvOptions with some limits a little outside of what you're expecting. This can help get past the first few iterations without crashing.