CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   Best mesh for heating and ventilation cfd? (http://www.cfd-online.com/Forums/ansys-meshing/70635-best-mesh-heating-ventilation-cfd.html)

kingjewel1 December 1, 2009 07:13

Best mesh for heating and ventilation cfd?
 
Hi there,

I'm looking to use a structured cartesian mesh for a CFD simulation of a passive scalar in CFX. Can you recommend this type of hex meshing or are tets better?

Best regards

PSYMN December 13, 2009 00:12

Depends.
 
Hex mesh generally gives better quality, more efficient, more accurate, etc. particularly for certain physics like combustion or when you really want to capture the viscous effects due to high speed flow. But this comes at a price (difficulty to generate for some configurations.)

If your configuration is simple, like a blade row model or a bypass fan duct or something like that. You could probably generate the Structured Hexa mesh very easily, perhaps quicker than a tetra/prism mesh, so the decision is easy.

If it is more complex, like a water jacket in an engine block, or an underhood cooling model, the Hexa blocking might take you such a long time that you would rather just fill it with an nice tetra mesh and take your chances on the slightly reduced fidelity.

But note that multiblock (like ICEM CFD Hexa) is very different from Cartesian mesh. Cartesian mesh is always aligned (XYZ) and probably doesn't fit the geometry very well, which may make it much less accurate than Tetra/Prism (unless the geometry is a box). Mutli-block Hexa can fit the geometry and give you the high quality. It can be structure or unstructured, but that is just referring to how the data in the mesh file is organized (we can covert between the two).

Best regards,

Simon

PSYMN December 13, 2009 11:42

Bfcart...
 
One other thought...

We have a new mesh method called Body Fitted Cartesian. I have not tried it on HVAC yet (Usually use it for Bio-medical or EMAG, but it may be ideal. If this is a model you could share, I could try it out...

We are also working on a new CutCell Cartesian mesher and could use it as a test case for that also.

Simon


All times are GMT -4. The time now is 07:33.