Unrolling a loft / developable surface

rhino
unhandled

#1

just wondering what anyone thinks about this unrolled surface. as you can see the intersection curves become distorted when they are unrolled with the surface.

unrolled, loft - developable surface.3dm (392.8 KB)


(Pascal Golay) #2

Hello - a normal loft makes a cleaner surface here that does unroll with the curves. In V6, DevSrf also makes a developable surface - a little different.

-Pascal


(David Cockey) #3

DevLoft in V6. DevSrf is a Lab plug-in for V5.

Users should be aware that in many (most?) cases DevLoft will return a surface which is not exactly developable everywhere. The reason is it creates a surface which runs the full extent of the selected rails, including portions of each rail for which there is not a corresponding part which results in a surface without twist. DevSrf in V5 did not cover the full length of each rail, only the areas which would result in a surface with sufficiently small twist.


Equivalent to DevSrf
#4

I took a look at your model and made a video here:

Unrollsrf

I would be interested in how you made the original surface on the loft-dev layer as I could not replicate that workflow.

Also, I get a different result in V6 as that of V5 of Rhino in the initial unrollsrf command.

Kent


(David Cockey) #6

@anajob Did you use DevLoft to create your initial surface? If so the kinks in the intersection curves when unrolled may be due to DevLoft not working well with curves which have straight sections. Also the result of DevLoft in some situations may have multiknots which need to be removed.

Another potential problem for any developable surface algorithm is the chine curve has a small amount of reverse curvature near the transom which means there are not theoretical corresponding points on the keel curve.

unrolled, loft - developable surface DC Fix10V5.3dm (454.4 KB) Note I worked in V6 but saved as V5 file.

I moved the end control point up .016 on the chine curve to eliminate the reverse curvature.

Extended the chine curve so that theoretical ruling lines could run from the straight portion of the keel curve to the chine curve. Note the shape of the extension influences the shape of the resulting developable surface.

Split the keel curve where it became straight.

-DevLoft with the chine curve and the forward portion of the chine curve; added “exact” ruling line; used Untwist option several times.

RemoveMultiKnot on the surface.

Sweep1 to create a planar surface along the straight portion of keel.

MergeSrf for single surface.

Trim the surface at the transom.

I still need to explain how I determine exact ruling lines.


#7

david, i used loft / developable in rhino 5. if you pick the rails a little back from the tip at the bow, you get the corrupted surface that has a little over lap on the center line or keel curve. that is what makes the anomaly in the unrolled surface. if you select the rails a little closer to the tip at the bow, you get a slightly different surface that doesn’t have the over lap in it and it unrolls just fine.


(David Cockey) #8

Loft with the developable option can produce some strange results.which is probably why it was dropped from V6.