Centerline 2

[forfun]

Does anyone know how to get the centerline of a 3d fillet?
I have a HD2 with HA1 and GSO.

Thanks in advance
forfun

 

[KooKoo]

I had to find the centerline for a 3D tube (dumb solid) and I used the Spine option. In some lengths of the tube I had to set the Start Point to get the correct result.
Maybe someone has a better suggestion.

 

[Kapitan]

In the picture the background model has the (Red) fillet whose centreline is required.

The foreground model is the construction one to do it.

[ul]
[li] Analise the original Fillet Radius.[/li]
[li] Undo the model to the point where the fillet was to be made.[/li]
[li] Offset the two original surfaces that were filleted, by the Fillet Radius.[/li]
[li] The (White)intersection curve of the two (Transparent) offsets is the Fillet Centreline.[/li]
[li] If the fillet had a Law driven variable radius, then use the same Law to create the surface offsets.[/li]






[/ul]

 

[solid7]

Without destroying any geometry, you can also construct perpendicular surfaces at each of the fillet boundaries, and intersect them, as shown.






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

 

[Kapitan]

V5, like other systems, allows various ways of doing things and naturally some will be better, or more accurate, than others.

In my first picture the Backtrack method was deliberately used to reduce the risk of error creeping in.
The Red Fillet was derived from the two surfaces, so to find the centreline using them is the most accurate way. To work from a derived element is analogous to photocopying a photocopy, there is bound to be some degradation.

In this picture two fences have been created on the trimmed boundaries of the original surfaces and normal to them. It is best to hide the fillet and it's boundaries first. Going back to the source elements is the better way.
Of course, the fillet may be a Blend or something else with a spine, so that the fillet is rather more complex and requires a different approach.

I seem to remember, years ago, previous versions of Catia had a function in Curve 1 or Curve 2, called Bi-Tangent and Tri-Tangent. These would compute the curves created by rolling an imaginary ball along the valley of two surfaces, the output was the contact curves and the locus of the centre of the rolling ball.

But, whatever way it is done, an understanding of the geometry and what is really happening is invaluable. It's too easy to create something that looks right but may be technically, rubbish.

 

[solid7]

But, whatever way it is done, an understanding of the geometry and what is really happening is invaluable. It's too easy to create something that looks right but may be technically, rubbish.

The general assumption is, and should always be, "garbage in, garbage out."

I failed to clarify that, but I don't think that the analogy of, "To work from a derived element is analogous to photocopying a photocopy, there is bound to be some degradation" is entirely true. It CAN be true, but certainly is not always, or even most of the time.

To find out if the fillet is workable without destroying the geometry, one would only need to do a preliminary surfaic analysis. It is also assumed that anyone working in surfacing, asking specific questions, must be familiar with the basic tools.

Your mileage may vary, but I've found that quite frequently one way OR the other is THE ONLY way. Sometimes, even what seems like a simple fillet, won't produce a quality result using the method you described first, which is why I added the extra thought. Not to outdo you by any means - your example was excellent - just to add "plan B," if you will.






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