esolution of the boundary layer

Bananenflanke · December 21, 2020, 08:16

Hi,

I have a question regarding the resolution of the boundary layer. In the CFX Solver Modeling Guide (4.2.7.2.1), it says that the boundary layer must be resolved with at least 10 nodes when using wall functions for the turbulence models to work properly.

I wonder if this only applies to fully developed boundary layers? If it also applies to developing boundary layers, at what point should there be 10 nodes in the boundary layer? Maybe after the transition from laminar to turbulent flow?
For example, in a pipe flow with a constant velocity profile at the inlet, there is no boundary layer directly at the inlet and shortly after the inlet it is only very small. In this case, it would be very impractical to impossible to have 10 nodes in this area of the boundary layer. How do you have to handle something like this?

Best regards and stay healthy!

karachun · December 21, 2020, 14:34

Region near an inlet (in pipe flow) should be considered as a region with nonphysical flow (if you use constant velocity and turbulence variables). Therefore you do not need to accurately resolve BL in this part. Use the same inflation as in the fully developed part. Anyway you do not use results from the inlet region.

ghorrocks · December 21, 2020, 16:50

Sorry Alexander, your answer is wrong. The BL in the developing part definitely is physical - it does develop as the flow develop. So you cannot say it is nonphysical. And you definitely can model it if that is important to your model - and if you need accurate results from the inlet region that can be done.

If you need accurate results in the developing region then you need a mesh fine enough to resolve it, and a turbulence model suitable to model the physics. The 10 nodes in the BL guide is just a guide, and as you point out it gets a bit hard to interpret what that means in the inlet region. The best thing to do is to do a mesh sensitivity study and see what mesh is required to get accurate results in your case.

EDIT: rereading Alexander's comment, he says that the inlet is non-physical if you use constant velocity and turbulence conditions. While this is correct, surely the best way to fix this problem is to move the inlet boundary condition further upstream from the region of interest such that it does have an accurate BL development.

Bananenflanke · December 22, 2020, 04:30

Thank you very much for your help! Doing a mesh study will probably be the best thing to do.

December 21, 2020, 08:16	esolution of the boundary layer	#1
Bananenflanke New Member Jannis Join Date: Jul 2020 Posts: 20 Rep Power: 6	Hi, I have a question regarding the resolution of the boundary layer. In the CFX Solver Modeling Guide (4.2.7.2.1), it says that the boundary layer must be resolved with at least 10 nodes when using wall functions for the turbulence models to work properly. I wonder if this only applies to fully developed boundary layers? If it also applies to developing boundary layers, at what point should there be 10 nodes in the boundary layer? Maybe after the transition from laminar to turbulent flow? For example, in a pipe flow with a constant velocity profile at the inlet, there is no boundary layer directly at the inlet and shortly after the inlet it is only very small. In this case, it would be very impractical to impossible to have 10 nodes in this area of the boundary layer. How do you have to handle something like this? Best regards and stay healthy!

December 21, 2020, 16:50		#3
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,933 Rep Power: 145	Sorry Alexander, your answer is wrong. The BL in the developing part definitely is physical - it does develop as the flow develop. So you cannot say it is nonphysical. And you definitely can model it if that is important to your model - and if you need accurate results from the inlet region that can be done. If you need accurate results in the developing region then you need a mesh fine enough to resolve it, and a turbulence model suitable to model the physics. The 10 nodes in the BL guide is just a guide, and as you point out it gets a bit hard to interpret what that means in the inlet region. The best thing to do is to do a mesh sensitivity study and see what mesh is required to get accurate results in your case. EDIT: rereading Alexander's comment, he says that the inlet is non-physical if you use constant velocity and turbulence conditions. While this is correct, surely the best way to fix this problem is to move the inlet boundary condition further upstream from the region of interest such that it does have an accurate BL development. aero_head likes this. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum. Last edited by ghorrocks; December 22, 2020 at 05:26.

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
Question about adaptive timestepping	Guille1811	CFX	25	November 12, 2017 17:38
Centrifugal fan-reverse flow in outlet lesds to a mass in flow field	xiexing	CFX	3	March 29, 2017 10:00
Any formula for approximating the boundary layer thickness around a cylinder?	bestniaz	Main CFD Forum	0	October 24, 2015 02:00
[ICEM] 3D Dynamic Mesh - Boundary layer mesh issues	nathanricks	ANSYS Meshing & Geometry	0	September 23, 2015 05:14
Low Mixing time Problem	Mavier	CFX	5	April 29, 2013 00:00

December 21, 2020, 14:34		#2
karachun Senior Member Join Date: Nov 2015 Posts: 247 Rep Power: 12	Region near an inlet (in pipe flow) should be considered as a region with nonphysical flow (if you use constant velocity and turbulence variables). Therefore you do not need to accurately resolve BL in this part. Use the same inflation as in the fully developed part. Anyway you do not use results from the inlet region.

December 22, 2020, 04:30		#4
Bananenflanke New Member Jannis Join Date: Jul 2020 Posts: 20 Rep Power: 6	Thank you very much for your help! Doing a mesh study will probably be the best thing to do.