[snailstb]

I have to create two kinds of meshes in my model. The first one is hex, the second is tet, and pyramid meshes for the transition bewteen two. we can use the "merge meshes method" tool to complete. Th best way is that we make sure the surface meshes of interface between two regions in distribution and scale are the same.

Just take one simple model as an example shown in fig.0 and fig.00. Hex meshes will be created in the inner region, tet meshes will be created in the out region. I finished the hex meshes in the inner region. In order to better control meshes, surface meshes of out region are created firstly, and want to use T-GRID method (i do not know much about the volume meshes method, i puzzle why ICEMCFD can not create mixed meshes using tri and quad surface meshes directly instead of using "mesh merges

specifically, after finishing the inner hex meshes, i want to create surface 1(in fig.2) meshes based on the mesh points of surface 2(quad meshes).But unfortunately, it does not work, seen in fig.3,4 and 5. It is no use even respect line elements selected. why? how to solve this problem in ICEMCFD? everyone whoever can help me is grateful greatly.thank you all.

[snailstb]

fig.4, fig5 are been as follows:

[snailstb]

how to create triangle meshes of one surface based on the points of other surface meshes before volume meshes creation?

no matter how i do it, i can not deal with it. It made me puzzled and confusing.

I can do it easily when i use gambit, i mean the mesh points consistency. firstly, you can create quad meshes and tri meshes on the surface of volume, then create Tet/Hybid elements based on TGrid method. it means tet and pyramid meshes are in the volume meshes. very simple and easy.

but what about ICEMCFD? Even so many methods for surface meshes and volume meshes creation, i can not find a way to solve it. point consistency(common points in surface common curves) is my great care, it is also a good way to drop down the numerical error.

who can tell me how to solve it in ICEMCFD.I also believe it is not a problem and just a piece of cake in ICEMCFD.

I spend a lot of time in studying ICEMCFD. I find that the advantage comparing with the gambit is the hex technology. However, the unstucture mesh technology disappointed me a little. maybe i am not proficient about it.

anybody in anytime tell me something about unstructure mesh technology and merits is grateful.

[PSYMN]

Sorry this is so late, I often miss CFD Online posts because there are just too many of them... I just jumped down and chose one old email at random for this evening.

You can't just use the Gambit methods to mesh in ICEM CFD... the tools are very different. ICEM CFD can do what you want, but you need to use the ICEM CFD way of doing things or you will keep being frustrated...

I see you are trying to match the surface tri mesh to the Hexa mesh... This is not necessary for ICEM CFD.

First, do you have a volume hexa mesh? I will assume that you do. I will also assume that there is a unique interface part between the Hexa and tetra regions...

Yes, you want the tetra size to be "similar" to the hexa size, but it can easily be off by a factor of three and it certainly doesn't need to be point matched. The only requirements are

1) that they share an interface part completely
2) that both the hexa and tetra side perimeters be projected to that interface part.

For the hexa side, you can make sure of 2) by associating the edges to the curves.

To get your mesh sizing similar, just set the correct mesh size on the interface surface so that your tetra mesh will come out the right size.

To mesh the tetra portion, put a material point in that section and use the Octree Tetra or DeLaunay tetra mesher. The idea is to generate a volume mesh and not worry about individual surface meshes, they will take care of themselves.

As for the question of generating tetra and Hexa at the same time, you can do that, but that is another story.