[EinarKC]

Hello All,


Im currently meshing a 3D cone for hypersonic simulation. My topology is C-type with a symmetry plane. Im having trouble with bad elements in front of the cone, mainly where the sharp tip ends - all the way to farfield.



MESH WITHOUT FARFIELD




CLOSE UP OF NOSE




BLOCKING





As you can see the blocking is wierd, this is because I use the "Sponge" approach, Im not really sure what its called. It's pretending to wrap the fluid domain in a sponge like so:





Can I get some tips and techniques on how to mesh a cone with this kind of topology. My reasoning for using such topology is because I intend to simulate this mesh in a range of angle of attack

[Stel]

I didn't understand your blocking strategy (judging from the green blocks, I mean). I see a triangular shaped block which crosses the tip without edge splitting, it looks weird form me at a first glance.

But anyway:

Your geometry is axisymmetric, right? I would suggest you to mesh it starting from a 2D (Planar) block and then using 2D to 3D rotation. It is easier and it should work pretty well.

[EinarKC]

Originally Posted by Stel

I didn't understand your blocking strategy (judging from the green blocks, I mean). I see a triangular shaped block which crosses the tip without edge splitting, it looks weird form me at a first glance.

But anyway:

Your geometry is axisymmetric, right? I would suggest you to mesh it starting from a 2D (Planar) block and then using 2D to 3D rotation. It is easier and it should work pretty well.

Thanks for your input. I tried doing your method but I got stuck in the nose again. I cant get the ogrid quite right. The sharp point I find very troublesome.

[Stel]

Are you getting negative volumes or just bad quality elements around the nose?

[Stel]

Triangle shapes are really tricky for structured meshes, indeed. Maybe what you could try is a round nose (with a radius small enough to avoid undesirable effects on results) to assign an edge to it (using a squared block where now you have a triangle one). Still this could result in non ideal aspect ratios but probably better mesh quality from the inside. From the outside, those wedges are considered by default as bad quality elements by ICEM, but sometimes this does not necessarily mean that they will cause problems in the solution.

[EinarKC]

Originally Posted by Stel

Triangle shapes are really tricky for structured meshes, indeed. Maybe what you could try is a round nose (with a radius small enough to avoid undesirable effects on results) to assign an edge to it (using a squared block where now you have a triangle one). Still this could result in non ideal aspect ratios but probably better mesh quality from the inside. From the outside, those wedges are considered by default as bad quality elements by ICEM, but sometimes this does not necessarily mean that they will cause problems in the solution.

You were right. I manage to generate a manageable mesh. I had to truncate the tip a bit. In the end, my minimum angle is 11 deg and for 2x2x2 determinant its 0.45. I was trying to aim for atleast 18 deg angle but its hard. the lowest quality cells were at the tip as expected. Also, no errors were found during quality checks



star-ccm+ and fluent manage to read it without any problem. Thanks for the inputs