Temperature fall under physical value

HorstvanGrass · August 25, 2021, 04:39

Hi,
I have a multiregion simulation which a Fluid starting around 293K. It is a classical buoyancy problem nothing to serious. However during the calculation process in some cells the temperature falls under 293, which is not physical in my understanding, as I'm heating and there is no serious expansion.

I doubt this may due to some convergence issues. With running time there are less cells, which temperature falls under 293. It is not much, but it can affect the solvers stability when I do not limit it.

Does someone encountered similar behavior before? Do you know which part of the energy equation is responsible for this? For example, falsely calculated alpha , transferring to much heat to the wall?

Thx in advance

agustinvo · August 25, 2021, 05:13

Hi,

could you add some info, like the solver you're using, numerical schemes... The reason of your low T values might be there.

Once I was running unsteady simulations in buoyant flows and I got higher temperatures than the prescribes ones when moving from 1st order to 2nd order schemes.

You could try to limit the temperature field using a fvOption.

Regards

HorstvanGrass · August 25, 2021, 05:29

Quote:

Originally Posted by agustinvo

Hi,

could you add some info, like the solver you're using, numerical schemes... The reason of your low T values might be there.

Once I was running unsteady simulations in buoyant flows and I got higher temperatures than the prescribes ones when moving from 1st order to 2nd order schemes.

You could try to limit the temperature field using a fvOption.

Regards

Hi thanks for the response,
I'm using only second order numerical schemes as Gauss linear and Crank Nickolson with 0.9 and also cellMDLimited leastSquares 1.0.

Also already using fvOption for stabilizing.

Therefore I'm a bit confused.

HorstvanGrass · August 26, 2021, 04:21

Hi,
could manage to restrain the influence factor.

It's the thermal diffusion which cause the temperature drop:
fvm::laplacian(turbulence.alphaEff(), he)

Using other laplacian scheme seems to help like: Gauss linear orthogonal but doesn't get rid of the problem. The mesh is also very smooth, therefore that does not seem to be a source.

August 25, 2021, 04:39	Temperature fall under physical value	#1
HorstvanGrass Member Horst van Gras Join Date: Oct 2018 Posts: 45 Rep Power: 8	Hi, I have a multiregion simulation which a Fluid starting around 293K. It is a classical buoyancy problem nothing to serious. However during the calculation process in some cells the temperature falls under 293, which is not physical in my understanding, as I'm heating and there is no serious expansion. I doubt this may due to some convergence issues. With running time there are less cells, which temperature falls under 293. It is not much, but it can affect the solvers stability when I do not limit it. Does someone encountered similar behavior before? Do you know which part of the energy equation is responsible for this? For example, falsely calculated alpha , transferring to much heat to the wall? Thx in advance

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Shadow Wall and temperature	norger	FLUENT	10	September 28, 2019 12:43
Outlet temperature to temperature boundary condition of velocity inlet. (UDF)	dillon	Fluent UDF and Scheme Programming	3	March 25, 2018 12:06
unexpected constant Temperature on a clip surface	Sungki	OpenFOAM Running, Solving & CFD	0	August 4, 2015 05:50
Difficulty In Setting Boundary Conditions	Moinul Haque	CFX	4	November 25, 2014 18:30
Inlet won't apply UDF and has temperature at 0K!	tccruise	Fluent UDF and Scheme Programming	2	September 14, 2012 07:08

August 25, 2021, 05:13		#2
agustinvo Senior Member Agustín Villa Join Date: Apr 2013 Location: Alcorcón Posts: 314 Rep Power: 15	Hi, could you add some info, like the solver you're using, numerical schemes... The reason of your low T values might be there. Once I was running unsteady simulations in buoyant flows and I got higher temperatures than the prescribes ones when moving from 1st order to 2nd order schemes. You could try to limit the temperature field using a fvOption. Regards

August 26, 2021, 04:21		#4
HorstvanGrass Member Horst van Gras Join Date: Oct 2018 Posts: 45 Rep Power: 8	Hi, could manage to restrain the influence factor. It's the thermal diffusion which cause the temperature drop: fvm::laplacian(turbulence.alphaEff(), he) Using other laplacian scheme seems to help like: Gauss linear orthogonal but doesn't get rid of the problem. The mesh is also very smooth, therefore that does not seem to be a source.