CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   How to decide the number of grid lines in boundary layer (http://www.cfd-online.com/Forums/main/112027-how-decide-number-grid-lines-boundary-layer.html)

Anna Tian January 20, 2013 17:21

How to decide the number of grid lines in boundary layer
 
Hi,

May I ask how to decide how many grid lines(at the direction proportional to the wall) do I need(to resolve boundary layer well)?

Of course grid independence study is always good practice. But I always need a first guess of this number to have my first grids before the grid independence study. Thank you very much.

cfdnewbie January 21, 2013 03:52

That depends on the order of your scheme and on the type of flow you are considering. Impossible to answer in general....

Anna Tian January 24, 2013 17:08

Quote:

Originally Posted by cfdnewbie (Post 403031)
That depends on the order of your scheme and on the type of flow you are considering. Impossible to answer in general....

Could you please recommend any paper about this?

agd January 24, 2013 22:25

If you are performing a RANS or URANS simulation the standard approach is to settle on the distance of the first point off the wall based on an acceptable y+ value, and then keep any grid stretching ratio below 1.2. The resulting simulation should always be reviewed after the fact to ensure that the actual value of y+ is still acceptable and the boundary layer is well-resolved.

venkateshaero January 25, 2013 05:04

After running problem in solver. you will get actual value of Y+ . before running also we can calculate by Y+ estimatation calculater in CFD online(Rough). based on calculation you can find first cell distance.

Anna Tian January 25, 2013 17:20

Quote:

Originally Posted by venkateshaero (Post 403959)
After running problem in solver. you will get actual value of Y+ . before running also we can calculate by Y+ estimatation calculater in CFD online(Rough). based on calculation you can find first cell distance.

Thanks. May I ask another question?

If I use laminar model(no model), do I need to have Y+ smaller or equal to 1?

Thank you.

venkateshaero January 26, 2013 08:48

please post your problem brief. whether its Heat transfer or only fluid flow problem. I will send some data based on that. that data will answer your question. .

Anna Tian January 26, 2013 10:42

Quote:

Originally Posted by venkateshaero (Post 404177)
please post your problem brief. whether its Heat transfer or only fluid flow problem. I will send some data based on that. that data will answer your question. .

It's a very simple case. Just laminar flow go through a rectangular pipe. But there's a 90 degree turning in the middle. There's nothing about heat transfer. Do I need to consider Y+ for laminar flow?

Just wondering: If it's for heat transfer, Y+ will need to be below 1, right?

venkateshaero January 27, 2013 05:38

Based Turbulence models Y+ limitation is there for each models(like K-omega ,K-epsilon). if u want caputure flow parameter along vertical. u have make fine mesh near wall.For Y+ consideration i will refer and answer your question later .

immortality June 7, 2013 18:04

Hi
I have a question like Anna.
Anna did you find an answer to your question about importance of y+ in laminar cases?
and is it important that flow be compressible or incompressible?
thanks.

Aeronautics El. K. June 8, 2013 09:04

Quote:

Originally Posted by Anna Tian (Post 402979)
Hi,

May I ask how to decide how many grid lines(at the direction proportional to the wall) do I need(to resolve boundary layer well)?

Of course grid independence study is always good practice. But I always need a first guess of this number to have my first grids before the grid independence study. Thank you very much.

There are formulas for the estimation of the boundary layer thickness whether that is laminar or turbulent. Use them.

Anna Tian June 13, 2013 05:10

Quote:

Originally Posted by immortality (Post 432712)
Hi
I have a question like Anna.
Anna did you find an answer to your question about importance of y+ in laminar cases?
and is it important that flow be compressible or incompressible?
thanks.

I think usually the boundary layer thickness doesn't strongly depend on whether the flow is compressible or incompressible. Y+ can be defined for laminar flow but doesn't make too much sense. We can use Blasius theory to estimate the laminar boundary thickness then put 11 points to the boundary layer. This could be a good starting point. Checking and grid independence study need to be done afterwards.


All times are GMT -4. The time now is 08:05.