Mesh size and equivalent sand grain rougthness

kordou · January 27, 2011, 05:52

Hi!

I am trying to model a windtunnel experiment and i have to apply a wall roughness on the ground equal to 1.53m.

In cfx for rough wall you need to specify the equivalent sand grain rougthness which is about 30*1.53. (according to this http://www.mi.uni-hamburg.de/fileadm...5-2007-www.pdf).

In the cfx manual there is the a notation about the cell size of the first cell. it says that the center of the first cell must not be smaller than the sand grain roughness. So for a wind tunnel of hight 200m (full scale) i have to have a fisrst cell with size 22.95m which is extremely hight.

any recomentation about how create this mesh?

P.S. the geometry is an empty simple rectangular box
P.S2. the ground material on the wind tunnel have roughness 0.0045m (full scale) of 0.0002 m (wind tunnel scale)

Thank you in advanced

PSYMN · January 29, 2011, 21:11

What mesher? What kind of mesh?

You need boundary layers, but there are many ways to get them.

kordou · February 14, 2011, 06:28

I was not thinking about mesher. I was thinking as a general problem. Let's say ICEM with Hexahedral mesh

as for the boundary layer can you please tell me how can I get them?

PSYMN · February 14, 2011, 11:01

In ICEM CFD, you would start with a single block, and create an OGRID. You would put an OGRID Face at the inlet and outlet so your inflation would be parallel to the side walls, but pass thru the inlet and outlet.

Then you would set the Ogrid edge distribution with the Edge params so that your spacing 1 = what ever first cell height you wanted and your growth ratio was what you wanted. You could also set the growth law to geometric1 or exponential1 or what ever you wanted.

Beyond that starts to get specific again to your problem, but if necessary, you can adjust the total thickness of your OGrid, have different initial heights on different edges, etc.

RobC56 · February 15, 2011, 07:16

The value you quote is what is expected. See Blocken et al. Atmospheric Environment 41 (2007) 238 - 252 for full discussion about this how to do ABL's in CFD.

Regards,

January 27, 2011, 05:52	Mesh size and equivalent sand grain rougthness	#1
kordou New Member Join Date: Mar 2009 Posts: 19 Rep Power: 17	Hi! I am trying to model a windtunnel experiment and i have to apply a wall roughness on the ground equal to 1.53m. In cfx for rough wall you need to specify the equivalent sand grain rougthness which is about 30*1.53. (according to this http://www.mi.uni-hamburg.de/fileadm...5-2007-www.pdf). In the cfx manual there is the a notation about the cell size of the first cell. it says that the center of the first cell must not be smaller than the sand grain roughness. So for a wind tunnel of hight 200m (full scale) i have to have a fisrst cell with size 22.95m which is extremely hight. any recomentation about how create this mesh? P.S. the geometry is an empty simple rectangular box P.S2. the ground material on the wind tunnel have roughness 0.0045m (full scale) of 0.0002 m (wind tunnel scale) Thank you in advanced

February 15, 2011, 07:16		#5
RobC56 New Member Rob Connell Join Date: Feb 2011 Location: Christchurch, New Zealand Posts: 1 Rep Power: 0	The value you quote is what is expected. See Blocken et al. Atmospheric Environment 41 (2007) 238 - 252 for full discussion about this how to do ABL's in CFD. Regards, __________________ RobC56

January 29, 2011, 21:11		#2
PSYMN Senior Member Simon Pereira Join Date: Mar 2009 Location: Ann Arbor, MI Posts: 2,663 Blog Entries: 1 Rep Power: 47	What mesher? What kind of mesh? You need boundary layers, but there are many ways to get them.

February 14, 2011, 06:28		#3
kordou New Member Join Date: Mar 2009 Posts: 19 Rep Power: 17	I was not thinking about mesher. I was thinking as a general problem. Let's say ICEM with Hexahedral mesh as for the boundary layer can you please tell me how can I get them?

February 14, 2011, 11:01		#4
PSYMN Senior Member Simon Pereira Join Date: Mar 2009 Location: Ann Arbor, MI Posts: 2,663 Blog Entries: 1 Rep Power: 47	In ICEM CFD, you would start with a single block, and create an OGRID. You would put an OGRID Face at the inlet and outlet so your inflation would be parallel to the side walls, but pass thru the inlet and outlet. Then you would set the Ogrid edge distribution with the Edge params so that your spacing 1 = what ever first cell height you wanted and your growth ratio was what you wanted. You could also set the growth law to geometric1 or exponential1 or what ever you wanted. Beyond that starts to get specific again to your problem, but if necessary, you can adjust the total thickness of your OGrid, have different initial heights on different edges, etc.