Properties of a sketcher for lofting..? 1

[TurbulentFluid]

Hi I'm trying to make a 3D geometry of a gulf for meshing purposes; I'd like to loft it bottom->top and I did the profiles but "loft" won't select them. They're all closed curves in different sketchers and different planes, containing only the visible geometry. What might be the problem? Thax a lot!

 

[Azrael]

First check the sketch with the sketch analysis tool and if you don´t find anything try using the output feature in the sketcher.

 

[TurbulentFluid]

Thanx... nope, it didn't work... I imported the stuff from a.dxf file, then I tried the .igs format but loft tool doesn't want to select them at all (but it does select if I want to define them as leaders). They're very complex, but non-intersecting polylines, though - is it even possible to loft such complex geometry...?

 

[solid7]

Polylines are not good lofting geometry, by definition. The resultant of a polyline loft would be a multifaceted surface, and if your linecount doesn't match from one curve to the next, there is no solution available in loft. It's like trying to smooth blend a P-line with 3 segments to another with 5. (try it and tell us what you get)

My suggestion would be for you to extract and explode the P-lines, and then make splines (from points) at the line intersections. When you have smooth geometry, you'll have a much better idea of what you challenges are going to be. (or maybe you'll get what you want with no trouble)




**************
Check out CATBlog!

 

[Configurator]

^^ Good post.

* for you.

I'll disagree on one point. Splines aren't the best choice because they have reversals.

What he should do, IMO, is fit in conic sections.

Lee

http://www.lmco.com

​

 

[solid7]

You are correct - splines are not the best choice... But it was only my intention to use them as a starting point, since I can't see the geometry. (something better than a polyline)

Thanks for the star. I feel honored to receive one from you.




**************
Check out CATBlog!

 

[TurbulentFluid]

OK, after a looong and exhaustive work, which also included Orbix wars in I-DEAS, I have found an answer to this situation, and will post it here in case someone else has such a problem to solve (modelling&meshing of a complex geometry such as a gulf for CFD purposes):

1. the geometry of the gulf was re-created by SMS (triangulated) and the points (just points, not the geometry) were exported into a .dxf file;
2. the .dxf file was exported to .igs via mechanical desktop
3. the .igs was dumped into I-DEAS and made into a domain by Fit Surface to Points and extrude options;
3. I-deas file was exported into .igs, imported into Catia and made into a solid (I-deas won't swollow making a solid out of such a complex geometry; the surface was fitted to over 56 000 points).
4. Catia file was modified into the desired domain, .igs-ed, and returned to I-DEAS
5. meshing was executed in ideas by partitioning the domain as you would for mapped meshing, which you can't do because of the complexity of the geometry. Use "sweep mesh" instead (you'll find it under options->menus->generate)...
6. Export to unv and FINALLY: FLUENT....