CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Phoenics (
-   -   y+ value (

Dipak November 7, 2000 03:30

y+ value
Dear all, I am solving flow through staggered wall plate channel using k-e model with wall function approach. I have to take the near wall grid point at such a distance so that y+ lies between 30-130(log-law region). I have calculated that distance assuming friction factor as i find dy=f(hydraulic diameter,Re,friction factor,y+). I have tested this for flow through circular pipe and got the y+ value close to the one i took for calculating near wall grid point distance(i.e. dy). But in case of flow through staggered wall channel i find that dy is about half of the total distance in y-direction. And here also the y+ value is close to the taken one. Now my doubt, is the result correct as the flow field for half of the domain solved analytically(wall function approach)?? Is there any other way to get the the near wall grid point distance??

Please help me and make me understand.


Mike Malin November 9, 2000 13:31

Re: y+ value
It is impossible to answer this question without knowing the precise geometry and the actual formulae used. It is clear that the method would work for pipe flow which is one of the flows from which data were used to develop empirical wall functions. Does staggered channel mean baffles normal to the flow are positioned in a staggered fashion at various intervals down the channel? Or are the baffles laid in in line with the flow at staggered positions inside the channel and along it's length? The critical question is whether the formulae used are valid for this type of flow.

Dipak November 10, 2000 01:24

Re: y+ value
Thanks for your comments. Plates are arranged at staggered positions along the channel length(i.e flow direction). It's a simplified model of rectangular offset srtip-fin used in plate-fin heat exchanger.

As we know, (a) y+=friction_velocity*(dy/2)/enul (b) friction_velocity=sqrt(wall_shear_stress/density) (c) skin_friction_factor=4*wall_shear_stress/(0.5*density*mean_vel**2)

So from (a),(b) and (c) we get, dy=funtion_of(Reynolds_no,hydraulic_diam,skin_fric tion_factor,y+)

Now my question is: if dy comes the half of the distance in normal to the flow direction(y-direction) then will be my solutions correct??

Thanks Dipak

Mike Malin November 13, 2000 11:10

Re: y+ value
The solutions are probably not correct, but I would need to see what you are doing in detail before making a firm conclusion. The derivation of the formula is not a problem, but rather what length and velocity scale are used in the Reynolds number. It seems to me that if you have a main channel which is effectively sub-divided in smaller channels by the insertion of in-line plates, then you have to be careful in the analysis you are using for estimating the near-wall mesh size. I think is impossible to say anymore without knowing what you are doing in detail, i.e. the exact geometry and the main flow parameters and boundary conditions.

All times are GMT -4. The time now is 18:44.