[iilw1314]

hi everyone .you may help me if you are someone who knows gambit .
anyone can help me to mesh the zone for flow passing cylinder using the unstructured grid.

i choose the form 'pave ',but unfortunately,i cannot get a good mesh as choosing 'map'(structured grids).i really want to get concentric circle mesh for boundary layer of the cylinder.i should say the reason why that i did not use the operation of mesh boundary in gambit is that i find that using the transition patern "1:1" can cause strange mesh and i do not want use "4:2"or others for the boundary of the cylinder.anyone can tell me how to get good mesh for the flow passing cylinder useing ustructured grids?thx.

[ztdep]

Dear friends:
Multiblock stuructued grid system is strongly suggested for the problem. seperate the zone into several blocks, generate structured grid in each region.

[iilw1314]

Originally Posted by ztdep

Dear friends:
Multiblock stuructued grid system is strongly suggested for the problem. seperate the zone into several blocks, generate structured grid in each region.

yes,you are right ,but i have another problem.my case is 2D and i use another code to deal with the result *.neu from gambit.and that code require only one face exist.if i use multiblock ,there may be sevral faces .i know that fluent can do well with the 'sevral faces' but my code is not fluent.so i really need one face at last and i have tried many ways to get one face at the last but all failed,any suggestion?thx!

[ztdep]

Dear friends:
I often use the neutral file from the gambit. the face between every blocks will dispear when you output the mesh file. It is just a assistance to improve the mesh quanlity

[iilw1314]

Originally Posted by ztdep

Dear friends:
I often use the neutral file from the gambit. the face between every blocks will dispear when you output the mesh file. It is just a assistance to improve the mesh quanlity

thanks，now i think it's not that problems.my code can deal with the 'o' grids(concentric circle)badly.so the method of using multibolcks seems meaningless ,so i have to use the boundary setings in gambit to do it with one block and use the transition pattern '4:2"not "1:1".only in this way my code can get through.so strange!so boring!..e..

[tez]

Why don't you try structured grid in the boundary layer and then triangles outside the boundary layer. It should work fine, provided your solver allows that.