How can I estimate y+ for K-epsilon turbulence model in openFOAM?

[Mukut]

Good morning all!

As I told earlier about my simulation in a wind tunnel. I am expecting separation in the diffuser region, to get expected separation proper cell division is necessary. For standard k-e turbulence model I learned from the previous replies that it better to use y+>30. I have found these two sites to calculate y+

http://www.pointwise.com/yplus/

http://www.cfd-online.com/Tools/yplus.php

Here is my simulation domain:

Straight section height is 200mm which is 630mm long and a diffuser section with 20deg inclination having a length of 250mm.

In the above two y+ calculation site, there is a term boundary layer length and reference length, I am confused with these two term because both sites calculates the same but they mention these terms differently. Shall I use this length as 200mm? But in diffuser section these vertical height is a varying height, as it has 20deg slope in bottom wall.

BR

MuKuT

[ohbuchi]

Hi, Mukut

I think we should use consistent definition of Re and Cf that is suitable to target problem.

Your problem is internal flow, but distance from inlet is quite short. 

If inlet velocity profile is flat, we may treat it as boundary layer flow over the flat plate. 

If this assumption is appropriate, L should be length from inlet.

If the inlet flow profile is fully developed, we should treat it as internal flow, 

so the definition of L should be hydraulic equivalent diameter.

Ohbuchi

2014年12月24日水曜日 11時27分37秒 UTC+9 Mukut:

[Mukut]

Dear Ohbuchi and Nakagawa san


Good morning. I have extended the tunnel in diffuser section. Now straight section length is same as previous which is 630mm and diffuser section length is 770mm. So in total, WT length is 1.4m. Here is the screen shots of calculation y+ estimation:

So, near wall distance is 4mm. From origin, due to diffuser downward extreme end of diffuser is -280mm (770*tan20). In this location total height is 200+280=480mm. So 480/4=120 cells in y direction gives equal 4mm cell height. Now, if I use mesh grading higher than 1 in y direction, it gives gradual y cell height upwards......

Is my assumption okey??? Pls suggest me

BR

mukut

[Mukut]

Ux velocity contour for above mentioned conditions for 2000 iteration,