Dear All,

When using the standard k-epsilon then must Y+ for the cells at the wall be > 11, Otherwise it is possible to use the low Reynold models where y+ should be =1 or Y+=2..3 depending on model.

1) Is that also true when heat transfer is involved? As I recall from many years ago, the thermal boundary layer behave different from the turbulent boundary layer.

2) I have been told from CFX when using the differential stress model then should Y+>11. But I want to resolve the boundary layer meaning I want to calculate a case where I have y+ =< 1 and using the differential stress in ordrer resolve the non-isotropic behaviour in the boundary layer. (uu' NOT EQUAL TO vv' NOT EQUAL TO ww') in a boundary layer.

Is there any one out there that has made a correction to CFX, that can handle my problem ?

Regards Jan

Hi Jan,

This is more speculation than an answer, but here goes anyway...

You can turn the logarithmic wall functions off with the command:
:>MODEL DATA
:>WALL TREATMENTS - WALL PROFILE 'LINEAR'

Now, using the linear wall functions, you should be able to take equations all the way to about y+=1 or so. You will have to be careful, as the equations start to get pretty unstable near the wall.

I am not sure whether this approach will work at all, it is just an idea.

Regards, Glenn

Hi Glenn,

Excellent idea, so obvious that it must be correct �. I can see in the manual that the linear wall profile is used for the low Reynolds number model (y+=1). Maybe you then can correct my on this one, I thought that very close to the wall, that there would be a laminar sublayer (y+<1 ??) should the near wall cell then not be treated as a 'laminar' cell and using the quadratic wall model ??

Regards Jan

Damn! Some times it helps to read the manual, to get more information. I have checked the manual, it is a correct approach to adjust the wall model, and when then sublayer is properly resolved it is correct to use the laminar wall model.

BUT!! The linear and quadratic model can not be used with high Reynolds number models (see app. C and wall treatments pages p. 187 in cfx4.3 manual).

So I can not use this approach

Jan

I have discussed this problem with user support. Only the standard wall model is available in connection with the differential stress model => y+ >>11

A two-layer model for the differtial stress will be available in CFX?.? year 2???.

Is there really nobody that have implemented a two -layer model for the differential stress in CFX4 that will share the knowledge.

Regards Jan