Unable to do sweeping after neighbor volume thexed

[y.wayne]

Hi my friends,

I try to generate the hex mesh of a missile’s surround region. The surrounding region
is cut into 3 volumes at plane z=5 and z=-5.
I try to mesh the half-sphere structure in z<=-5 with thex and the -5<z<5 region with sweep.
However, when thex mesh are generated for z<=-5 sphere, sweeping for intermediate
volume failed. The error message does not show why it fails.

Since thex mesh produce low quality mesh, any advice on new meshing scheme is welcomed.
geo_df_2.sat (86.2 KB)
geo_df_3_scheme3.trelis (2.6 MB)

[y.wayne]

I have another question that, can I allow nodes on suface to redistribute while doing ConditionNumber smoothing for better mesh quality? It seems mesh on surface never changed when executing smooth.

[y.wayne]

When hex mesh are generated from right to left, how to mesh sphere structures with a hole inside?
I run into similar trouble before when surface mesh of a half sphere shell’s cross section is generated first. That’s quite difficult to deal with.

[karl]

Hi,

Welcome to the forum. In general I try to avoid the thex algorithm because it creates such poor quality elements. What are the requirements for your mesh. Does it have to be all hex elements?

The best way to get all hex elements would be to decompose the mesh into hex-meshable sections. Here is one solution to your problem.

reset
import acis "geo_df_2.sat" nofreesurfaces attributes_on  separate_bodies 
webcut volume 1 with plane xplane offset 0 
webcut volume 1 2 with plane yplane offset 0 
webcut volume all with cylinder radius .1 axis z 
 
webcut volume all sweep curve 118 119 120 121 122  zaxis angle 360 

composite create surface 23, 99, 100, 102, 103, 107
composite create surface 4, 65 to 67, 69, 70
composite create surface 24, 127 to 129, 139, 140
composite create surface 35, 111, 112, 119 to 121
composite create surface 22, 84, 89 to 92
composite create surface 5, 44 to 47, 49
composite create surface 6, 39 to 41, 48, 50
composite create surface 17, 59, 60, 68, 71, 72
 
volume 13  scheme Sweep source surface 271 target surface 270
volume 14  scheme Sweep source surface 269 target surface 267
volume 15  scheme Sweep source surface 266 target surface 272 
volume 16  scheme Sweep source surface 273 target surface 268

vol all size .25
vol all scheme auto

curve 126 127 int 2
Curve 221 interval 10
vol 1 vis on
Curve 284 248 171 interval 10
Curve 284 248 171 interval 10
mesh vol all
quality vol 2 3 4 8 9 11 7 10 12 14 15 16 scaled jacobian draw mesh

[karl]

Hi,

Welcome to the forum. In general I try to avoid the thex algorithm because it creates such poor quality elements. What are the requirements for your mesh. Does it have to be all hex elements?

The best way to get all hex elements would be to decompose the mesh into hex-meshable sections. Here is one solution to your problem.

reset
import acis "geo_df_2.sat" nofreesurfaces attributes_on  separate_bodies 
webcut volume 1 with plane xplane offset 0 
webcut volume 1 2 with plane yplane offset 0 
webcut volume all with cylinder radius .1 axis z 
 
webcut volume all sweep curve 118 119 120 121 122  zaxis angle 360 

composite create surface 23, 99, 100, 102, 103, 107
composite create surface 4, 65 to 67, 69, 70
composite create surface 24, 127 to 129, 139, 140
composite create surface 35, 111, 112, 119 to 121
composite create surface 22, 84, 89 to 92
composite create surface 5, 44 to 47, 49
composite create surface 6, 39 to 41, 48, 50
composite create surface 17, 59, 60, 68, 71, 72
 
volume 13  scheme Sweep source surface 271 target surface 270
volume 14  scheme Sweep source surface 269 target surface 267
volume 15  scheme Sweep source surface 266 target surface 272 
volume 16  scheme Sweep source surface 273 target surface 268

vol all size .25
vol all scheme auto

curve 126 127 int 2
Curve 221 interval 10
vol 1 vis on
Curve 284 248 171 interval 10
Curve 284 248 171 interval 10
mesh vol all
quality vol 2 3 4 8 9 11 7 10 12 14 15 16 scaled jacobian draw mesh

[y.wayne]

Thanks for your reply karl,

Yes I need a all-hex mesh. Your solution has much higher quality than the thex mesh I generated.
I should adapt yours. Thank you again.

Best,
Wayne

[y.wayne]

Hi karl,

I think you didn’t do ‘imprint & merge all’ before mesh, so mesh are not conneted.
It’s difficult to mesh with same scheme considering conformal mesh.

[y.wayne]

The mesh scheme still works with merged volumes, except low quality narrow elements in cylinders on head and tail. Thanks you again.