T Plus

sasanghomi · July 15, 2018, 01:58

Dear Friends,

I have a question about T+. I cannot find any option in CFX in order to monitor T+ values. I am wondering if CFX applies wall functions for heat transfer. We always talk about the acceptable values for Y+ but T+ values are not discussed.
Could anyone do me a favor and cast light on this ambiguity?

Best Regards

ghorrocks · July 15, 2018, 19:21

This is discussed in the CFX theory documentation, under wall treatments. Yes, the thermal equivalent of y+ is T+. But wall functions on thermal modelling are known to be problematic and are not as widely used as for the momentum equations. Many researchers don't use wall functions if they require accurate heat transfer results and resort to integration to the wall. Hence parameters like T+ are not widely used.

Opaque · July 16, 2018, 16:09

If you look into the theory documentation for "Near Wall Modelling" chapter (as Glenn suggested) you will find the formula used for Tplus,

Then, you can write an expression for Tplus as a function of Yplus, and the related properties.

From the formula, and your knowledge of the Prandtl number you can infer Tplus behavior for all cases. You will need a material with extreme variations of properties along a wall to have very different correlation behavior between Yplus for every node.

sasanghomi · July 21, 2018, 06:41

Thank you so much for your responses.
I did not get any conclusion that whether or not CFX uses heat transfer wall functions by default?
My second question is that when we use wall functions for velocity domain, does it necessarily mean that we are using heat transfer wall functions as well?
Also, I can remember that in reference books about turbulent flow it was a must to take care of the first node's position of mesh in boundary layer. In fact, the first node should not be positioned in the buffer layer (5<Yplus<30). However, I have not seen any recommendation like that in CFX software program. What is the secret behind that?

Could anyone help me with these questions please?

Best Regards

ghorrocks · July 21, 2018, 07:04

If your setup is using wall functions for the momentum equation it is using wall functions for heat transfer as well. So if your turbulence model uses wall functions by default it will use them for heat transfer as well.

The model used in wall functions only applies to the logarithmic layer, from about y+=11 and out. It does not apply in the sublayer. So if you apply wall functions in y+<11 you are applying them in a regime where they are not applicable. The scalable wall functions in CFX have removed this lower limit problem: https://sharcnet.ca/Software/Ansys/1.../i1313358.html

July 15, 2018, 01:58	T Plus	#1
sasanghomi Senior Member Sasan Ghomi Join Date: Sep 2012 Location: Denmark Posts: 292 Rep Power: 14	Dear Friends, I have a question about T+. I cannot find any option in CFX in order to monitor T+ values. I am wondering if CFX applies wall functions for heat transfer. We always talk about the acceptable values for Y+ but T+ values are not discussed. Could anyone do me a favor and cast light on this ambiguity? Best Regards

July 15, 2018, 19:21		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,728 Rep Power: 143	This is discussed in the CFX theory documentation, under wall treatments. Yes, the thermal equivalent of y+ is T+. But wall functions on thermal modelling are known to be problematic and are not as widely used as for the momentum equations. Many researchers don't use wall functions if they require accurate heat transfer results and resort to integration to the wall. Hence parameters like T+ are not widely used. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

July 21, 2018, 07:04		#5
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,728 Rep Power: 143	If your setup is using wall functions for the momentum equation it is using wall functions for heat transfer as well. So if your turbulence model uses wall functions by default it will use them for heat transfer as well. The model used in wall functions only applies to the logarithmic layer, from about y+=11 and out. It does not apply in the sublayer. So if you apply wall functions in y+<11 you are applying them in a regime where they are not applicable. The scalable wall functions in CFX have removed this lower limit problem: https://sharcnet.ca/Software/Ansys/1.../i1313358.html __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

July 16, 2018, 16:09		#3
Opaque Senior Member Join Date: Jun 2009 Posts: 1,815 Rep Power: 32	If you look into the theory documentation for "Near Wall Modelling" chapter (as Glenn suggested) you will find the formula used for Tplus, Then, you can write an expression for Tplus as a function of Yplus, and the related properties. From the formula, and your knowledge of the Prandtl number you can infer Tplus behavior for all cases. You will need a material with extreme variations of properties along a wall to have very different correlation behavior between Yplus for every node.

July 21, 2018, 06:41		#4
sasanghomi Senior Member Sasan Ghomi Join Date: Sep 2012 Location: Denmark Posts: 292 Rep Power: 14	Thank you so much for your responses. I did not get any conclusion that whether or not CFX uses heat transfer wall functions by default? My second question is that when we use wall functions for velocity domain, does it necessarily mean that we are using heat transfer wall functions as well? Also, I can remember that in reference books about turbulent flow it was a must to take care of the first node's position of mesh in boundary layer. In fact, the first node should not be positioned in the buffer layer (5<Yplus<30). However, I have not seen any recommendation like that in CFX software program. What is the secret behind that? Could anyone help me with these questions please? Best Regards