RealizableKE & nutkAtmRoughWallFunction

Sylvain · October 28, 2025, 11:19

Good afternoon,

We found a bug when using the RealizableKE turbulence model together with the nutkAtmRoughWallFunction.

With a very simple test case of a parallelepiped domain with an ABL profile at the inlet, such as in the TurbineSitting tutorial, one can see that the velocity profile is maintained throughout the domain.

When switching to realizable KE, the profile changes throughout the domain, with the velocity decreasing too quickly near the ground.
A slice at z=10m from the ground shows it well.

This problem was not happening in OpenFOAM 10. It is present in both OpenFOAM 12 and OpenFOAM 13.

If anyone has experienced such a problem, any insight would be appreciated.

Best regards

Tobermory · November 10, 2025, 07:34

There are a number of possible reasons for the boundary layer developing over the length of the domain, none of which are "bugs in the turbulence model".

In fact, a constant profile is an unnatural state for a zero pressure gradient boundary layer - like any boundary layer, it wants to grow in height with downstream distance. However, this evolution will generally take a number of boundary layer heights of downstream distance to occur, so is unlikely to be the cause here.

Things for you to check: do your inlet boundary profiles match the wall roughness? What boundary condition have you applied at the top of the domain and is this consistent with the height of your domain and height of your boundary layer? This "simple" test case is very sensitive to the boundary conditions that you have applied.

Sylvain · November 12, 2025, 06:12

Well, there was actually a bug, but it was solved

https://bugs.openfoam.org/view.php?id=4165

I didn't check out, my bad

You have to add "nutMaxCoeff 1e8" in "constant/momentumTransport"

It was not present in the turbineSitting tutorial in the early releases of OF12, and i didn't check back the tutorial in OF13, my bad again...

It was due the addition of the boundEpsilon function in the kEpsilon model

Thank you very much Henry Weller for helping us on this subject

Once you add "nutMaxCoeff 1e8" in "constant/momentumTransport" it works perfectly fine

Hope the post might be helpfull anyway

Tobermory · November 12, 2025, 06:25

Aaah - I had forgotten about that - I was the one who actually put the bug request in to Henry to get that fixed in the tutorial!

October 28, 2025, 11:19	RealizableKE & nutkAtmRoughWallFunction	#1
Sylvain Member Sylvain Aguinaga Join Date: Feb 2010 Posts: 44 Rep Power: 17	Good afternoon, We found a bug when using the RealizableKE turbulence model together with the nutkAtmRoughWallFunction. With a very simple test case of a parallelepiped domain with an ABL profile at the inlet, such as in the TurbineSitting tutorial, one can see that the velocity profile is maintained throughout the domain. When switching to realizable KE, the profile changes throughout the domain, with the velocity decreasing too quickly near the ground. A slice at z=10m from the ground shows it well. This problem was not happening in OpenFOAM 10. It is present in both OpenFOAM 12 and OpenFOAM 13. If anyone has experienced such a problem, any insight would be appreciated. Best regards

November 12, 2025, 06:12		#3
Sylvain Member Sylvain Aguinaga Join Date: Feb 2010 Posts: 44 Rep Power: 17	Well, there was actually a bug, but it was solved https://bugs.openfoam.org/view.php?id=4165 I didn't check out, my bad You have to add "nutMaxCoeff 1e8" in "constant/momentumTransport" It was not present in the turbineSitting tutorial in the early releases of OF12, and i didn't check back the tutorial in OF13, my bad again... It was due the addition of the boundEpsilon function in the kEpsilon model Thank you very much Henry Weller for helping us on this subject Once you add "nutMaxCoeff 1e8" in "constant/momentumTransport" it works perfectly fine Hope the post might be helpfull anyway dlahaye likes this.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Issue compiling solver for OF 2.4	klausb	OpenFOAM Programming & Development	8	August 1, 2018 20:15
Variation between kEpsilon and realizableKE	Thangam	OpenFOAM Running, Solving & CFD	1	January 5, 2014 05:29
realizableKE hanging and then exploding	vaina74	OpenFOAM Running, Solving & CFD	4	January 10, 2013 09:54
RealizableKE	cfdengineering	OpenFOAM Running, Solving & CFD	9	March 31, 2011 04:45

November 10, 2025, 07:34		#2
Tobermory Senior Member Join Date: Apr 2020 Location: UK Posts: 869 Rep Power: 19	There are a number of possible reasons for the boundary layer developing over the length of the domain, none of which are "bugs in the turbulence model". In fact, a constant profile is an unnatural state for a zero pressure gradient boundary layer - like any boundary layer, it wants to grow in height with downstream distance. However, this evolution will generally take a number of boundary layer heights of downstream distance to occur, so is unlikely to be the cause here. Things for you to check: do your inlet boundary profiles match the wall roughness? What boundary condition have you applied at the top of the domain and is this consistent with the height of your domain and height of your boundary layer? This "simple" test case is very sensitive to the boundary conditions that you have applied.

November 12, 2025, 06:25		#4
Tobermory Senior Member Join Date: Apr 2020 Location: UK Posts: 869 Rep Power: 19	Aaah - I had forgotten about that - I was the one who actually put the bug request in to Henry to get that fixed in the tutorial!