CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   unstructured grid wall cells algorithm (https://www.cfd-online.com/Forums/main/159114-unstructured-grid-wall-cells-algorithm.html)

researcher September 9, 2015 05:42

unstructured grid wall cells algorithm
 
How do we find whether a cell is located at the wall. How would the algorithm do it ?

Thanks in advance


shah

mprinkey September 9, 2015 10:53

Quote:

Originally Posted by researcher (Post 563081)
How do we find whether a cell is located at the wall. How would the algorithm do it ?

Thanks in advance

shah

If you are talking about an unstructured code like FLUENT or OpenFOAM, to touch all of the cells adjointing a wall boundary, you loop over the faces in the wall thread/patch and look up the cell connected to each face (F_C0() in FLUENT, etc).

researcher September 10, 2015 08:09

Thank you for your reply.
In fact I am programing it using fortran. Plus I am using unstructured mesh. I am curious how would a triangular cell with only one of its vertices located on the wall effect the algorithm.
Thanks

mprinkey September 10, 2015 09:00

Quote:

Originally Posted by researcher (Post 563081)
How do we find whether a cell is located at the wall. How would the algorithm do it ?

Thanks in advance


shah

Quote:

Originally Posted by researcher (Post 563287)
Thank you for your reply.
In fact I am programing it using fortran. Plus I am using unstructured mesh. I am curious how would a triangular cell with only one of its vertices located on the wall effect the algorithm.
Thanks

It won't. If you are using FVM, the entire scheme is based on surface integrals over each cell. A single vertex has no area to integrate over, so its contribution is zero. Only finite area faces will contribute to the flux balances that you write for each cell.

researcher September 10, 2015 11:06

Quote:

Originally Posted by mprinkey (Post 563300)
It won't. If you are using FVM, the entire scheme is based on surface integrals over each cell. A single vertex has no area to integrate over, so its contribution is zero. Only finite area faces will contribute to the flux balances that you write for each cell.

Haha, of course. Thank you. Your reply made both my work and the day.


All times are GMT -4. The time now is 09:02.