Twisted Developable Surfaces

Nor did I expect there would be. That's why I originally asked about guidelines, not universal answers. Let me give you an example of the kind of thing I mean. When you form a 90-degree bend near the edge of a piece of sheet metal there's no fixed minimum of lip height that can be simply or universally applied to guarantee that you'll be able to successfully form the bend. The material, the tooling, and the thickness all factor into it, and perhaps other things. However, a well-known guideline for engineers who design sheet metal components is that a bend height of at least 8 times the material thickness works for most of the common materials, in most of the common thicknesses, on most tooling. I thought that, after centuries of forming developable surfaces, that sort of tribal knowledge might exist in this area, and that somebody here might know it.

 

The material itself will tell you what it is happy with, you cannot compare sheet metal to thin ply, the bending radii limitations are not comparable. Be guided by what shape the material will naturally follow with the least "persuasion", especially with a form of construction that is supposed to be easy and uncomplicated

 

Nope.
Love the optimism though.
Page 2 and nobody's come forward yet...
Try a smaller build maybe to get skills up.
Or, better yet, go with a proven design that meets all your SOR in their prioritized entirety.

 

DC summed it up well.

Too many variables

 

What is your actual goal in building a kayak?

Proof of some theory or getting a specific shape?

What is wrong with the shapes available from current designs?

Can you identify a benefit to your desired shape?

A discussion that is based on the definations of various surfacing methods is usually worthless.

What cad package are you talking about?

 

I have a similar scenario. I have a panel on a boat I'm hoping to build that is mainly cylindrical but will have a little bit of double curvature. I'm planning on running a series of rough tests on some of the leftover plywood (after I cut out most of the other panels) to see how much it can curve in both directions and how difficult it is to force the panel into shape. I suspect that my panel will work just fine but the tests will confirm that before I do anything permanent. If the plywood ends up being much more difficult to bend in two directions, I will at least know that I need to work out some other solution. I also plan to test how tight of a curve the plywood can handle. It would be interesting to see how much a piece of plywood can be twisted before it pops into a curve.

 

There are tables created by plywood manufacturers that show the practical bending radii for ply of various thicknesses. Developable means basically this, that at any point on the surface, a straight line can be drawn across it, that touches the surface at all points, and that the lines "normal" (perpendicular) to the surface, at all points along that line, will be in the same plane.

 

I do know all of that. It's still good practice to do some testing when things are less familiar or when information (in the case of the OP's question) is not cataloged anywhere. In my case a handful of rough tests will be easy enough to do in my garage (to the accuracy I want) that there isn't really a reason not to.

 

This occurs in the hyperbolic paraboloid, which is a ruled surface but is not developable. Perhaps it is the only case contrary to that rule.

 

The twist can also create very concave (or convex) shapes too - while being developable. Ross Lillistone Fleet (or Flint) here.

 

The visible line (2D) is concave but I think the surface of the bottom plank (3D) is not. In my opinion a surface is concave only if its inner area in any direction is "deeper" than its outline. In the pictured shape you will find one direction for every point in it which provides a straight line to the outline. It is more like the hyperbolic paraboloid Tansl mentioned.

An plane rectangular plank if twisted, will form geometrical a double helix. So its longitudinal outer lines must be longer (tensile strain) than its neutral middle line (compressive stress). Regarding this the twisted surface to me is no longer developable.

I was building hulls with a twisted bottom plank and the designer told me the plywood should bear it. That was correct for most of the lenght but not near the stem where the hardest twist was located. A friend testet it and got a crack in the plank.
https://uploads.tapatalk-cdn.com/20170823/d0c6e8411d24beffa9e8d6bce8b91db0.jpg

After cutting it to strips the dry fitting shows contact in the middle line and wedge shaped small gaps towards the outer lines.
https://uploads.tapatalk-cdn.com/20170823/cd62b3dbec078996e6cd8ddecf4a3733.jpg

When I built it, I cut the plank in strips (beveled edges) and used a tiny plane to make the middle of the strips a bit less wide (compared with their ends) to fit without gaps.
https://www.boote-forum.de/attachment.fphp?attachmentid=767168&d=1504035017

https://www.boote-forum.de/attachment.fphp?attachmentid=767169&d=1504035023

 

Designing and building boat hull using large sheets of plywood in non-developable shapes is frequently called "tortured plywood". A previous thread on the topic: tortured ply theory https://www.boatdesign.net/threads/tortured-ply-theory.37934/