CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Fluent UDF and Scheme Programming (http://www.cfd-online.com/Forums/fluent-udf/)
-   -   How to define the turbulent k near the wall ? (http://www.cfd-online.com/Forums/fluent-udf/116985-how-define-turbulent-k-near-wall.html)

joy2000 April 30, 2013 02:27

How to define the turbulent k near the wall ?
 
hello everyone, :)

I wish to define the solution of the turbulent kinematic energy solution near the wall, in Fluent.

My case is standard k-e with standard wall function.

From the ansys help we can see that the k-equation is solved in the whole domain including near the wall, while the epsilon adjacent to the wall is calculated from the near-wall equ with is derived from the local equilibrium hypothesis.

After the solution I found that the production of k (Gk) is zero everywhere on the wall, which is very different from the Gk equ based on the local equilibrium hypothesis.

I wish to give a definition of the k in order to force the local equilibrium near the wall. Could anybody give some advice how i can do this?

Many thanks!! :)

blackmask April 30, 2013 02:58

There is no fluctuation on the wall ( the velocity is always zero) so that the turbulent kinetic energy k vanishes on the wall.

joy2000 May 13, 2013 22:54

1 Attachment(s)
Quote:

Originally Posted by blackmask (Post 424076)
There is no fluctuation on the wall ( the velocity is always zero) so that the turbulent kinetic energy k vanishes on the wall.

From the ansys help we can see the k-equation would be solved in the whole domain including the near-wall cells while the epsilon would be given by functions near the wall. The boundary condition for the k-eq is the zero normal partial differential: dk / dn =0, on the wall. (d here also means "partial differential" for that I can't type the "round" symble :)) This is a Neumann boundary.


I saw in many literatures they specified the value of k on the wall, which is, to give a Diriclet boundary for the equation. For example, some can specify the k as u*^2/sqrt(Cmu), which is, to force the local equilibrium on the wall. Or some can give a zero k on the wall.

I checked the results of the Fluent on an external flow around a cylinder with k-e and standard wall function. When I compared the computed k to the "u*^2/sqrt(Cmu)" , I saw in the wake region the computed value differed a lot. I wonder whether this would be a key point to cause the differences between my results and those in the literatures.

So I wish to specify the k on the wall, in the Fluent, eg. though a UDF~~

Could anybody shed some light~~:)

Attachment 21716


All times are GMT -4. The time now is 19:14.