Surfaces of revolution from arbitrary

[HelterSkelter]

quilts?

Lets assume I have an arbitrarily shaped quilt located some distance away from a CSYS. The quilt is orientated roughly normally to the CSYS. You can imagine that if you spun this arbitrary quilt around, say, the Y axis of the CSYS a surface of revolution would be synthesised. Something vaguely like this (no surface of revolution shown):

http://img50.imageshack.us/img50/5061/surface3bi.jpg

Lets also assume the the quilt is not of a shape ameanable to calculating the surface of revolution from first principles i.e. lets say the quilt is built from a cloud of arbitrary points.

Is there an elegant modeling technique in proe that would enable me to create such a surface of revolution from this quilt about an arbitrary axis, such as the Y axis??

So far all I can come up with is a brute force method. Something like:
-Intersect a plane normal to CSYS Y with quilt to create an intersection curve.
-Scan along intersection curve to find point furtherest from CSYS Y (i.e. largest radius). Record this point's XYZ coordinates.
-Shift plane normal to CSYS Y down a notch, and repeat all above steps ad infinitum.

Not elegant at all and not even something I think you could do in proe...


Any suggestions???

Edited by: HelterSkelter

 

[ThreeSigma]

Helter,


This is not an exact method, but perhaps it would help. Try creating an arbitrary spline curve in a plane that contains the axis about which you want to revolve (Y in this case?). Now revolve this spline about the axis to create a surface, and make this surface a different color than your original quilt. You should now be able to see those locations where the quilt protrudes through your new surface. Iteratively adjust your spline curve until your original quilt is as close to your new surface as you desire. Maybe this will get you close enough.


Since you can't rotate the quilt to generate its swept volume, you insteadrotate an adjustable fence to try to circumscribe it.

 

[jeff4136]

I'm not real sure but think I might look at doing something like sectioning; planes normal to axis, some aribitrary number of 'em (a dozen?). For each section I'd create two circles, each tangent to the inner and outer most extremities of the section curve. (Might be more of an "eyeball" excercise than anything.) Once that's done create splines thru circle interesections with a radial plane.

Curiosity question; Why?