CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   What if coase mesh at the boundary? (https://www.cfd-online.com/Forums/main/8114-what-if-coase-mesh-boundary.html)

GER September 20, 2004 19:42

What if coase mesh at the boundary?
 
For a turbulence flow, if I am not interested in flow next to wall, can I use coarse mesh at the bounday (in normal direction of wall)?

Also, from model aspect, do I have to use fine grid if the model does not use a wall function?

Thanks a lot for your valuable input.

Ger


ag September 21, 2004 10:03

Re: What if coase mesh at the boundary?
 
1. No - behavior at and near the wall is probably the major influence on the flow.

2. Yes - see answer to 1. You need to have a y+ value of about 1 if you do not have wall functions. Otherwise you are wasting CPU cycles.

amol palekar September 22, 2004 13:16

Re: What if coase mesh at the boundary?
 
If you are not interested in flow near the wall, then why dont you change the boundary conditions at the wall to extrapolation. This way your computational domain will also be small. That should work for flow away from wall and which is not affected by the wall. -amol

ag September 22, 2004 17:35

Re: What if coase mesh at the boundary?
 
The original poster states that his interest is in turbulent flow. Ignoring the wall (a major source of turbulence) seems rather odd to me.

amol palekar September 22, 2004 19:02

Re: What if coase mesh at the boundary?
 
I think that may depend on problem at hand. if we have a jet type of configuration which is not wall bounded then i guess effect of wall may no matter. so may be if GER tells abt the particular problem we may know more. amol

GER September 23, 2004 20:31

What if only interested in large scales
 
I meant that we are interested in large scale motion of turbulence flows in which wall effects cannot be ignored.

What's your idea?

Ger


amol palekar September 24, 2004 00:57

Re: What if only interested in large scales
 
In that case I guess you would need fine enough grid to resolve the trubulence as ag had mentioned. -amol


All times are GMT -4. The time now is 01:57.