Meshing a truncated cone with high "area ratio"

Coreform Cubit Meshing
#1

Hi,

I am trying to mesh a truncated cone with high “area ratio” (meaning the top radius is much smaller than the base radius).
When I mesh it like a Partial Sphere (http://www.csimsoft.com/partialsphere.jsp), the very high width of the cells in the outer region at the cone base are nonsatisfying:

create frustum height 5e-1 radius 5e-1 top 2e-2
create Cylinder height 5e-1 radius 2e-2
webcut volume 1 with sheet extended from surface 4
delete Volume 2
webcut volume 1 with general plane yz noimprint nomerge
volume all size auto factor 2
imprint all
merge all
mesh vol all

What would be the best way for meshing this geometry with hex cells of similar size?
It would be great if somebody has an idea. Thanks a lot. :slight_smile:

Edit:
Additionally, it would be possible (or even desired) if the cells are getting smaller from the base towards the truncated tip (but of course not with the large ratio of base and top radius).

#2

How about an approach like this:

reset
create frustum height 0.5 radius 0.5 top 0.02
webcut volume all xplane
webcut volume all yplane
delete volume all except 1
composite create surface 14 13
volume 1 scheme tetprimitive
mesh volume 1
volume all copy reflect plane yplane
volume all copy reflect plane xplane
merge all

–Corey

#3

Thanks for your reply.
When I change the top-radius to 0 (i.e. a non-truncated cone), it looks really good - thanks. But unfortunately with the top 0.02, I get:

Any idea why?

#4

Looks like Trelis 14.0.5 is getting hung up on the vertices at the tip. Try compositing them away like so:

reset
create frustum height 0.5 radius 0.5 top 0.02
webcut volume all xplane
webcut volume all yplane
delete volume all except 1
composite create surface 14 13
composite create curve 3 21
composite create curve 16 17
volume 1 scheme tetprimitive
mesh volume 1
volume all copy reflect plane yplane
volume all copy reflect plane xplane
merge all

#5

Actually, I think the approach with a tetprimitive-meshed quarter of non-truncated cone is really the solution.
I am going to try cutting a cylinder with the size of the top and additionally four blocks along the axes. In this case there will be just four of the mentioned quarters left. I will add the journal if I succeed. Thanks again.