CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   The length of the first node (http://www.cfd-online.com/Forums/fluent/27506-length-first-node.html)

 Jennie February 18, 2000 15:37

The length of the first node

Hi! guys,

Could you tell me how to define the length of the first node to the wall, if the flow is laminar or turbulent?

Thanks.

Jennie

 Ahmed Hassaneen February 21, 2000 11:29

Re: The length of the first node

Hi Jennie, What you mean by the "length", I think you mean the length normal to the wall. If this is true, in the laminar flow problems this length is not important but in turbulent flow problems it is very important and it is a different story. Good luck =Ahmed=

 Volker Pawlik February 22, 2000 09:24

Re: The length of the first node

Hi Jenny,

I think your question is concerning the grid size of first cell adjacent to the wall?

Turbulent flows:

So for turbulent flow you should have a look to the definition of the y+=(density*u_tau*yp)/dyn.viscosity. u_tau is the so called friction velocity defined by sqrt(tau/density). For a pipe flow there is a relation-ship for u_tau/u=6.99(u_tau*R/kin. viscosity)^1/7 (R=pipe Radius, u=mean velocity) 1/7-Power law which can be generalized to a 1/n-law for different Re-numbers (e.g. see Blevins "Handbook of apllied fluid dynamics" for the relations between the friction coeff. lambda (or. s.t. called f) and n=1/sqrt(f), f=0.316/Re^0.25.

So solve for u_tau and put it into the def. of y+. Then you you get an equation for y+ as function of y which is the distance of the cell center (!!!) from the wall. Then you are able to estimate y+ even for non-pipe flows. Just exchange the Radius with the half of the hydr. diameter of your problem. It works fine. The y+-value which suits you fine depends of course on the turbulent model and wall model you want to use.

The y+ value is correct only for th full developed flow region.

Laminar flows:

have a look to the Fluent User's Guide p5-14 (fluent 5) or 5-23 (1996). For a fully developed pipe flow with Radius R choose: y/R<=0.1, for a flow between two plates choose y/H<=0.05 in order to resolve the flow profile correctly, that means that the stepwise linear approximation is not to far from reality. Unless the pressure loss due to shear stresses (or perpendicular momentum transport) will not be ok.

Good luck

 All times are GMT -4. The time now is 06:13.