CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   HVAC meshing with ICEM 4.cfx (http://www.cfd-online.com/Forums/cfx/20638-hvac-meshing-icem-4-cfx.html)

 Gearóid Lydon September 14, 2004 13:26

HVAC meshing with ICEM 4.cfx

Dear Users,

Does anyone have experience using Tetra ICEM 4.cfx to build a mesh for internal air flow in buildings?

I am working on a model of an operating theatre. I am not sure if the mesh I have constructed is appropriate. What parameters are important for an adequate model?

Tetra Settings:

Domain Size – 3m x 6m x 3m

Global Mesh Size: 0.16 m

Surface Mesh Size:

Walls: 0.16m – 0.08m

Inlet/Outlet: 0.065 – 0.025m

Prism Settings:

Growth Law: exp

Initial Height: 15mm

Height Ratio: 1.2

Number of Layers: 5

From experimental data the velocity range is from 0.1 to 0.4 m/s.

What parameters are important for an adequate model? Is the mesh density ok? Are the prism setting ok for the walls? What range of y+ values is required?

Best Regards

Gearóid Lydon

 Glenn Horrocks September 14, 2004 18:10

Re: HVAC meshing with ICEM 4.cfx

Hi,

You have to determine whether the mesh is adequate yourself by performing a mesh independance study. The mesh required varies job to job.

What type of model are you trying to do? Steady state or transient? RANS or DES/LES? These factors will also affect your required mesh.

Glenn Horrocks

 Gearóid Lydon September 15, 2004 05:11

Re: HVAC meshing with ICEM 4.cfx

Hi Glenn,

The model is Steady State using the k-w model.

For the mesh independence study, I should increase the mesh density until there are no changes in results?

Are the y+ values relevant?

If the model is changed to transient, will the mesh have to be less dense?

Best Regards

Gearóid Lydon

 Glenn Horrocks September 15, 2004 18:16

Re: HVAC meshing with ICEM 4.cfx

Hi,

The basic concept of a mesh independance check is you increase the mesh density until there are no changes in results.

The y+ values are relevant. If you use the default wall functions then there is no real y+ limit, just what gives you a mesh independant solution - however you will probably want y+ to be less than 1 if you want really good accuracy for seperations.

The required mesh is probably the same for transient and steady state, but computing resources usually mean you have to run a coarser mesh in transient as the computing time gets pretty long.

Glenn

 Gearóid Lydon September 16, 2004 06:40

Re: HVAC meshing with ICEM 4.cfx

Hi Glenn,