UnRoll or Develope a bent Tube

Hi All, Is it possible to reverse engineer a model of a bent tubular structure back to straight lengths of tube so I can export a file to my tube laser cutting people. I need to preserve the orientation of any intersecting end shapes. Ive attached a simple automotive roll cage for reference. Any help pointing me in the right direction would be much appreciated.Laser tube sample.3dm (616.0 KB)

Hello - not sure what you need exactly - the 2d center lines for each planar part?


Hi, The best way to describe it would be " UnBend a Tube" .Im trying to reverse engineer a model of a bent tube back into its unbent straight length but more importantly retain the orientation of any intersection profiles on each end. At the moment I am generating the end profiles using Pipe then Boolean difference and the projecting one of those profiles along the tube centre line using Orient along curve to the other end to determine the axial angular misalignment and using analize length on the tube centre line I can redraw the straight length and attach each end profile. Is there a better way. “Unbend Tube” would be good in the same way unRoll a surface is used for developing surfaces. Its so I can send a cut file to the Tube Laser cutting people.Is there a better way?

Hello - does the attached have what you are looking for? Flow from the centerline curve to a line of the same length.

Tube_maybe.3dm (178.9 KB)

Better would be: Explode the part - Flow the parts with PreserveStructure=Yes.


Hi Pascal, Im using Rhino 5 which wont open your newer version file. Are you able to save as an older version? Thanks for your help. If this problem can be cracked its a very powerful tool in the laser tube industry which is normally very expensive to achieve.

Hello - here it is as V5 - I exploded the tube, flowed (“PreserveStructure” = Yes) the two end bits from the red centerline to the green line, then added the new straight section in between. Somewhat guessing as to what is right here…

Tube_maybe.3dm (110.4 KB)


That looks correct but on that particular tube the two bends are in the same plane so there is no angular misalignment between the two end profiles. Would your technic work for this sample where the two bends are not in the same plane. In this file the two end profiles would end up being angularly misaligned by 20 Deg but I would like the technic to determine it. Thanks for your help.Laser tube sample 2.3dm (114.9 KB)

Hm - well, here’s what I get - not sure if it does you any good.
Laser tube sample 2_maybe.3dm (147.3 KB)


The length is spot on but unfortunately the axial misalignment is incorrect. Using Orient, On curve, I can slide one end profile to the other which does preserve its position then determine the angle. I was hoping for a quicker method. “UnBend Tube” function would be good because the tube laser machines are ahead of the CAD needed to exploit them at the moment.
Thanks for your effort Pascal.

Hello - it should be possible to fairly quickly Rotate or Rotate3D the end pieces on their axes to get them lined up correctly if you know what the angle should be.


If only it was that simple but its not on bent components with multiple bends in several planes.

Hello - I think there may be more we can do… in the case you sent, you know (correct?) what the angle is supposed to be when it’s all straightened out, correct? How are you determining that? 20 degrees is the angle off the horizontal of the upper straight bit, but what about the angle or rotation of the cuts in the tube axis? That is the sticking point, correct?


In the sample I sent I know its 20 degrees because its a very simple sample with both bends being 90 degrees with the second bend being twisted up by 20 degrees between the two bends, But its not always so simple with multiple bends and varying bend angles. I did that sample so it was easier to to verify the technique with the known answer. For a complex bent component I wouldn’t know if it was correct or not. If we extract an isocurve from the surface does that flow along the tube maintaining axial orientation.?

Yes you are correct that is the sticking point.

Here is another way to explain the problem. If you imagine a horizontal plain straight cut piece of round tube with a long line scribed on its surface set to top dead centre. Next you bend the tube with a series of complex bends, where does the line end up at the other end? I simple terms if you put one vertical bend of 180 degrees the line would end up at bottom dead centre.

If I reverse, so to speak, by hand, your example, by rotating the end parts in three planes, (20 degrees down, 90 degrees from top, 90 degrees from Front… I have the miters/cuts lining up exactly - no angle between the ‘peaks’ on those cuts… is that a dumb way to do it?


yes because you removed the “difficult bit” which is the 20 degrees ( in this case known) that is the point they should be 20 degrees out of alignment. What we need to achieve is an unbent model part so the laser tube company cut the end details then we bend the tube with respect to the start orientation so that when the part is produced it will fit into the predesigned structure of tubes with the end intersections true.

Thanks… I see. Not really, I am too dense to see how that rotation happens, but then, what I don’t know about bending tubing could fill volumes, as they say. But I am curious… Just to humor me - can you look at the attached? Is that in any way different from what you had (follow the edges) ?

@jdelooze - V5… Laser tube sample 2Question.3dm (151.8 KB)


I’m sure you’re far from dense, stick with it your a CAD guy but you already forgot I need a V5 file.

Are the individual bends planar without any twist, or are individual bends twisted?