Catenary Request

Hi David, is there anyway to adjust the number of segments in the Catenary component?

Also, I believe that the component actually generates a parabola, which is the shape that a catenary takes under uniform horizontal load (like an arch under a wall) as opposed to the shape of a pure catenary under self weight.

Replies to This Discussion

Permalink Reply by Giulio Piacentino on May 4, 2012 at 3:43pm

Hi Eli,

the code is related to the one on this conversation. This is code for the catenary equation and not a parabola, but please feel free to have a look and chime in.

I think David purposefully chose to use 50 divisions, even though it's possible to sample the underlying function as needed.

I hope this helps,

- Giulio
_________________
giulio@mcneel.com

Permalink Reply by Mateusz Zwierzycki on May 4, 2012 at 3:49pm

are there any plans or future possibilities to draw "exact" curves with equations with rhino/gh ? since gh is used by architects, this would be really an advantage...

Permalink Reply by David Rutten on May 5, 2012 at 12:47am

Nurbs cannot be used to represent every possible class of equation. They are good at Conic section types of curves (lines, arcs, parabolas, hyperbolas) as those can be represented exactly by adjusting control-point weights. Other types of curve will most likely need to be fitted to within tolerance.

The bigger problem however is that in order to get the curve shape right, one needs to 'understand' the equation being solved. Sometimes equations result in tiny detail which is easily missed if you just sample at fixed intervals.

Sometimes the equation simply has too much (or even infinite) detail, Sin(x^-1) for example has infinitely many wiggles near zero.

We can probably provide a better equation evaluator, but a truly guaranteed exact one is impossible and a nearly guaranteed exact one is well beyond my abilities as a programmer.

David Rutten

david@mcneel.com

Poprad, Slovakia

Permalink Reply by Eli Meltzer on May 5, 2012 at 8:16pm

It just seems that the chain "links" are located at even intervals perpendicular to the gravity vector, as opposed to evenly along the length of the chain, which is the behavior of a catenary chain under uniform horizontal loading, as opposed to self weight.

Either way, regards the number of divisions, I would be more interested in being able to reduce the number, say to 10 or 7, or something manageable, and not so much to better approximate a true catenary curve. Picking a finite number of links would be useful for things like setting out the voussoirs of an arch, or the members of a truss chord.

RSS