RSM without wall function

michal.s · March 30, 2018, 03:32

Hey there!

I am trying to compare simulation performed with k-w SST with RSM.
My mesh prepared for k-w has very fine mesh with boudary layer (y+~1).
As far I know (and read), RSM uses wall function.
Is it possible to simulate flow without using any wall function in order to calculate heat transfer between the wall and the fluid properly?
Otherwise, mesh should have y+>30 right?

LuckyTran · March 30, 2018, 12:22

You can use RSM for y+~1 meshes.

Quote:

Originally Posted by michal.s

Is it possible to simulate flow without using any wall function in order to calculate heat transfer between the wall and the fluid properly?

No, never because you always need boundary conditions for the transported turbulence variables (whether they be k, epsilon, reynolds stresses, whatever). You need some type of "wall function" unless you are doing DNS. The issue is too many people misunderstand what is a wall function because the meaning of wall function has always meant many different things.

michal.s · March 30, 2018, 16:40

Ok, I'll keep it in my mind

I'll explain my problem.
I've used very fine mesh with y+<1 near the wall that I'm cooling by the water jet. After changing turbulence model to RSM, my total heat transfer rate decreased over two orders of magnitude.
I've prepared coarse mesh (y+ about 15), heat transfer rate was more realistic with this one.
I was wondering how to simulate the flow with RSM using old (finer) mesh.
As far I understood,RSM Stress Omega can be used with my low y+.
Anybody could say sth about it?

Many thanks,
Michal

LuckyTran · April 1, 2018, 18:43

Two orders of magnitude has to be a joke right? That means there is an insignificant amount of heat transfer for the RSM case.

Does the flow make sense? Do your sanity checks and make sure you didn't goof some parameter. If y+ ~15 gave a result that was two orders of magnitude different, then it looks like you need to do a mesh dependence study.

I have compared r-k-e, kwSST, RSM (k-e based) and RSM (omega based) on very fine meshes in pipe flows and I would say the results should be +/- 25% or 50% on a bad day. Although, I would say the omega based RSM model is a bit finicky and I personally prefer the k-e based approach.

michal.s · April 1, 2018, 18:52

Firsty, I run steady-state simulation to get my flow converged. Surface temperature is fixed - after XXX iterations everything is steady, my total heat flux is about 250W (resonable for my case).
Then I switch to transient and turn my surface cooling UDF. It works perfect with k-omega. My UDF reads surface temperature, heat flux and calculates new temperature after the timestep.
Here is funny thing. Heat flux is now very low, but why? I dont get it.
UDF works great with k-omega so it's ok.

My mesh was checked with k-omega. I hoped that it would be also OK for RSM

March 30, 2018, 03:32	RSM without wall function	#1
michal.s Member Michal Join Date: May 2017 Posts: 35 Rep Power: 8	Hey there! I am trying to compare simulation performed with k-w SST with RSM. My mesh prepared for k-w has very fine mesh with boudary layer (y+~1). As far I know (and read), RSM uses wall function. Is it possible to simulate flow without using any wall function in order to calculate heat transfer between the wall and the fluid properly? Otherwise, mesh should have y+>30 right?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[Other] mesh airfoil NACA0012	anand_30	OpenFOAM Meshing & Mesh Conversion	13	March 7, 2022 17:22
[mesh manipulation] RefineMesh Error and Foam warning	jiahui_93	OpenFOAM Meshing & Mesh Conversion	4	March 3, 2018 11:32
What's the problem with turbulence models near the wall region?	Jaydi_21	Main CFD Forum	6	July 7, 2017 02:39
[swak4Foam] installation problem with version 0.2.3	Claudio87	OpenFOAM Community Contributions	9	May 8, 2013 10:20
Use of enhanced wall function in the RSM viscous model for a 3D sector	a.hasan	FLUENT	0	March 27, 2012 07:40

March 30, 2018, 16:40		#3
michal.s Member Michal Join Date: May 2017 Posts: 35 Rep Power: 8	Ok, I'll keep it in my mind I'll explain my problem. I've used very fine mesh with y+<1 near the wall that I'm cooling by the water jet. After changing turbulence model to RSM, my total heat transfer rate decreased over two orders of magnitude. I've prepared coarse mesh (y+ about 15), heat transfer rate was more realistic with this one. I was wondering how to simulate the flow with RSM using old (finer) mesh. As far I understood,RSM Stress Omega can be used with my low y+. Anybody could say sth about it? Many thanks, Michal

April 1, 2018, 18:43		#4
LuckyTran Senior Member Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,675 Rep Power: 66	Two orders of magnitude has to be a joke right? That means there is an insignificant amount of heat transfer for the RSM case. Does the flow make sense? Do your sanity checks and make sure you didn't goof some parameter. If y+ ~15 gave a result that was two orders of magnitude different, then it looks like you need to do a mesh dependence study. I have compared r-k-e, kwSST, RSM (k-e based) and RSM (omega based) on very fine meshes in pipe flows and I would say the results should be +/- 25% or 50% on a bad day. Although, I would say the omega based RSM model is a bit finicky and I personally prefer the k-e based approach.

April 1, 2018, 18:52		#5
michal.s Member Michal Join Date: May 2017 Posts: 35 Rep Power: 8	Firsty, I run steady-state simulation to get my flow converged. Surface temperature is fixed - after XXX iterations everything is steady, my total heat flux is about 250W (resonable for my case). Then I switch to transient and turn my surface cooling UDF. It works perfect with k-omega. My UDF reads surface temperature, heat flux and calculates new temperature after the timestep. Here is funny thing. Heat flux is now very low, but why? I dont get it. UDF works great with k-omega so it's ok. My mesh was checked with k-omega. I hoped that it would be also OK for RSM