CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   STAR-CCM+ (http://www.cfd-online.com/Forums/star-ccm/)
-   -   mesh quality? (http://www.cfd-online.com/Forums/star-ccm/93626-mesh-quality.html)

sieginc. October 20, 2011 21:39

mesh quality?
 
I'm new to using STAR-CCM and was wondering if there are some general guidelines to mesh quality? Take for instance my problem: I'm simulating the forced convection of an eletrical component (simplified as a 'block') in a duct using a baffle to help cool it. I am basically interested in knowing the surface temperatures on the block and the flow around it. There's a volume control around the block. Basically I just kept refining the mesh until my results did not seem to change. Also, is there a way to check y+ values once the mesh is created?

abdul099 October 20, 2011 22:31

There is no way to check Y+ only after meshing without solving. Y+ is not just a geometrical value but also influenced by the flow. So you need to solve at least until the solution will not change that much anymore. Then you can create a scalar scene and display y+ on all walls you are interested in.

General guidelines for mesh quality is quite simple: As high as possible. The better the mesh quality, the better the solution. When the preview of the "remove invalid cells" function finds more than just some single cells, it's always suspicious. It's also pretty often possible to create a section scene, display the mesh on the section plane and look at it near corners, edges, small features etc. When it looks bad, it is bad in the most cases.

But regarding your description, I assume, you are more interested in mesh sizes than mesh quality. Even a "perfect" mesh in terms of cell quality can be too coarse to obtain a good solution. So refining is quite important, but there is no general rule how far this should go. Some people say, as far as needed to get a mesh independent solution, but for most cases this is just impossible when one wants to get the solution still in this decade.
Anyway I think you're on a good way, as you're refining until the results seem not to change anymore.


All times are GMT -4. The time now is 12:00.