y+ concepts

howhs · August 15, 2013, 12:25

Hi people,

I would like to clear up my concepts for y+ in cfd. It is commonly said that for laminar flow, it is alright for us to use a y+<10 since there is no buffer layer and viscous sublayer in a laminar flow. However, when it comes to turbulent flow, buffer layer and viscous sublayer are both presence and we have to input the correct y+ value according to the turbulence model that we have chosen. Essentially, if we use a high-Re turbulence model without any enhance wall treatment, it is crucial to choose a y+ value that is above 30 since we want to avoid any unbounded errors from the wall function. For low-Re model such as SST, it is alright if we have a lower y+ number since this model is able to capture the wall function phenomenon.

Do correct me if there is any conceptual misunderstanding in the above explanation on y+. I am actually currently working on natural convection, thus, since there is no velocity, can I say that this is going to be a laminar flow and y+ can be chosen to be less than 10?

Thank you for your assistance!

flotus1 · August 15, 2013, 12:36

The low-Re formulations of the wall functions are only able to capture the flow near the wall correctly if the values of Yplus are in the order of 1 or below.

In natural convection processes, the velocity is not zero.
Especially for heat transfer simulations like this, it is crucial to have the boundary layer well-resolved. Accordingly, the Yplus-values in such a simulation should be in the order of 1 or below.
You cannot say generally that the flow regime will be laminar.

howhs · August 15, 2013, 12:57

Thank you foltus1 for the rapid reply.

Can I ask if there is any reason why the y+ has to be below the order of 1 for the low-Re formulation? Does that mean that the model is able to capture the laminar sublayer region?

From your explanation for natural convection, does that mean that we should not use standard k epsilon since it is not able to capture the natural convection phenomenon? SST model would be more appropriate for this case. Anyway, do you have any papers which explain the need to use y+ below the order of 1 for natural convection?

Thank you for your time.

August 15, 2013, 12:25	y+ concepts	#1
howhs New Member Join Date: Jul 2013 Posts: 18 Rep Power: 13	Hi people, I would like to clear up my concepts for y+ in cfd. It is commonly said that for laminar flow, it is alright for us to use a y+<10 since there is no buffer layer and viscous sublayer in a laminar flow. However, when it comes to turbulent flow, buffer layer and viscous sublayer are both presence and we have to input the correct y+ value according to the turbulence model that we have chosen. Essentially, if we use a high-Re turbulence model without any enhance wall treatment, it is crucial to choose a y+ value that is above 30 since we want to avoid any unbounded errors from the wall function. For low-Re model such as SST, it is alright if we have a lower y+ number since this model is able to capture the wall function phenomenon. Do correct me if there is any conceptual misunderstanding in the above explanation on y+. I am actually currently working on natural convection, thus, since there is no velocity, can I say that this is going to be a laminar flow and y+ can be chosen to be less than 10? Thank you for your assistance!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Basic turbulence concepts, vortex stretching energy cascade concepts help	irtiza	Main CFD Forum	7	August 4, 2013 17:37
How to distinguish between concepts of Temperature	Twiti	Main CFD Forum	3	April 19, 2005 05:40
fundamental concepts of turbulence	sarah_ron	Main CFD Forum	2	January 20, 2005 10:56
basic concepts	Max Wang	Main CFD Forum	0	June 29, 2004 11:15
Turbulent flow reference/statistical concepts	Tom Brown	Main CFD Forum	2	January 3, 2001 05:19

August 15, 2013, 12:36		#2
flotus1 Super Moderator Alex Join Date: Jun 2012 Location: Germany Posts: 3,427 Rep Power: 49	The low-Re formulations of the wall functions are only able to capture the flow near the wall correctly if the values of Yplus are in the order of 1 or below. In natural convection processes, the velocity is not zero. Especially for heat transfer simulations like this, it is crucial to have the boundary layer well-resolved. Accordingly, the Yplus-values in such a simulation should be in the order of 1 or below. You cannot say generally that the flow regime will be laminar.

August 15, 2013, 12:57		#3
howhs New Member Join Date: Jul 2013 Posts: 18 Rep Power: 13	Thank you foltus1 for the rapid reply. Can I ask if there is any reason why the y+ has to be below the order of 1 for the low-Re formulation? Does that mean that the model is able to capture the laminar sublayer region? From your explanation for natural convection, does that mean that we should not use standard k epsilon since it is not able to capture the natural convection phenomenon? SST model would be more appropriate for this case. Anyway, do you have any papers which explain the need to use y+ below the order of 1 for natural convection? Thank you for your time.