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/)
-   -   [Other] structured or unstructured? (http://www.cfd-online.com/Forums/ansys-meshing/86056-structured-unstructured.html)

morteza08 March 13, 2011 17:00

structured or unstructured?
 
hi dear all
a question?
what is the different between structured and unstructured mesh in generating good results in fluent?
for example? for a 3d impingement jet, what is the different between these two types? i want to obtain yplus=1 and good distribution for nusselt number

alastormoody11 March 14, 2011 04:30

Hi,

the idea is that using a structured mesh you can have faces normal to the direction of flow or aligned to the direction of flow thereby reducing the error during flux calculations.

In unstructured mesh you have no control over the orientation of faces thereby leading to an increase in the spatial discretization error.

PSYMN March 14, 2011 16:30

Hexa can be Unstructured...
 
I agree that alastormoody already answered your question. He is correct in that hexas aligned with the walls will better capture your fluxes (the first principles of the code are formulated with these assumptions), hexas also fill your volumes more efficiently, can be biased to better capture boundary physics, etc.

but it is not quite right to think of unstructured as tetra and structured as hexa... Maybe this is just a semantics thing, but I think it helps if we all speak the same language.

"structured" is more about how the data is stored. (domains with ijk instead of nodes with XYZ). It can only be done with Hexa and results in conveniently small files that solvers can turn into matrices very easily. Many in house solvers and older versions of comercial solvers are "structured" solvers that require a structured mesh. Without getting too technical, these solvers can use the implied information in the grid to take shortcuts.

More modern commercial solvers tend to be unstructured. Unstructured mesh stores the data for all the elements and nodes individually. Each element number is cross referenced with an element type and the node numbers. Each node number gives a physical (XYZ) location. You can't tell from any particular node or element number which elements are surrounding it or where it is in the big scheme of things, so the solver must do more work to figure out the matrix.

ICEM CFD hexa can output hexa mesh as structured or unstructured mesh (for unstructured solvers like Fluent). Older versions of Fluent (such as V4) actually needed "structured" mesh.

enviro March 15, 2011 12:06

Hi dear all,

i have another question referring to hexa/tetra-meshing.

You mentioned that hexa-mesh...
>is better to capture your fluxes
>fill your volumes more efficiently
>is better capture boundary physics
>etc.

Is there any literature you could suggest to get to know the reasons of these points?
Or maybe a study about the difference of tetra/hexa-meshing?

I often read sentences like "hexa-meshing is more sufficient for CFD, than tetra, because it captures the physics in a better way" but i do miss the references for it.

Would be great if you could help me!

kind regards

chg December 12, 2011 04:57

Here is one paper I found: Comparison of Hexa-Structured and Hybrid-Unstructured Meshing Approaches for Numerical Prediction of the Flow Around Marine Propellers


All times are GMT -4. The time now is 10:11.