Designing without curvature

I don't like to speak for him, but Richard Woods uses developable designs in some of his cats.

The lines are quite good imo.

I am not sure about his magic, however. And I believe he may be offline now.

ps-the pink extruded is for a built in cooler of sorts...

 

You are correct, though with a little care the lofted surface can be held fairly close to a conic section.

To achieve true developability in SolidWorks I might suggest:
1. Sketch a profile at some waterline. As my previous comment a minimally constrained spline will give a fair curve.
2. Extrude a surface from this. Use draft angle to set the flare.
3. Extrude other surfaces from the centreline for sheer & the bottom of the hull.
4. Trim the surfaces or generate a solid from the intersection of them.
5. Mirror to give a complete hull

Solidworks features are in italics. Additional surfaces can be generated from sketches on arbitrary planes (e.g. at the turn of the bilge for a double chine hull).

 

This is the same sequence I use when designing a hard chine hull; just that no computer program is involved. But, I am sure, using a computer allows many more options.

 

For an experienced designer, which I'm not, I think the main advantage with something like DelftShip is he automated hydrostatistics. Otherwise I imagine some may feel less constrained with paper and their imagination.

One feature I wish DelftShip (at least the free version) had that Blue Peter Marine's Hullform has is the ability to specify roll, the (approximate) displacement, and then find how the hull settles in the water giving an approximate yaw and pitch for that amount of heeling. It isn't automated, and since the calculations don't hold the displacement exactly as specified (pretty close though), it isn't precise, but for trying to figure out what a sailboats hull does while heeling Hullform 9, which is free, has some nice features.

 

Thanks for the feedback, tlouth7. This is basically what I have been doing, but I want more advanced shapes. For instance, I don't want my panels constrained to be conic. More specifically, I don't want all of the straight lines that exist on the plates to meet at one point.

 

You would probably combine different cones or cones and cylinders so that not all the straight lines converge at the same point.
Having a graphic idea of what you intend to do could be useful to give you better advice.

 

Surely you can use parallel lines by putting the focus of your cone at infinity, if nothing else.

 

And how do you know that "ordinary" conic/multi-conic development won't be good enough ? You want something that looks like a compound curvature, but isn't ? I am getting confused again !

 

All developable surfaces are ruled surfaces.
- but -
Not all ruled surfaces are developable surfaces.

All conical and cylindrical surfaces are developable surfaces.
- but -
Not all developable surfaces are conical or cylindrical surfaces.

 

Sounds awfully tricky doesn't it, but so far as I know, through any point on a developed surface, a straight line can be drawn, that touches the surface along its whole length, and at any and all points along that line, lines drawn "normal" (perpendicular) to the surface at those points, will form a flat plane. Not that that helps much !

 

Correct. Equivalent is that through any point on a developable surface there will be a straight line where a plane will be tangent to the surface along the line.

Added: If the surface is developable and not planar then through any point on the surface there will be on only one straight line on the surface which passes through that point (unless the point is an isolated apex singularity or a crease). If there are multiple straight lines then the surface is twisted and not developable.

 

I developed a method for determining exact ruling lines of the developable surface between two edge curves. The method is described as implemented using Rhino in this thread on the Rhino forum: Developable surface - exact ruling lines from edge curves https://discourse.mcneel.com/t/developable-surface-exact-ruling-lines-from-edge-curves/73928 I'll try to find time to create a more general post here but it may be January before I have time.

 

I think people get a little misled in thinking of conical developed surfaces being like a witch's hat, but there is no limitation of having to be based on upright, circular cones.

 

The "conical surfaces" I referred to my post above are general conical surface which can be produced by sweep a straight line from a fixed point along an arbitrary curve. Similarly the "cylindrical surfaces" I referred to can be produced by sweeping line which remains parallel to it's initial orientation about an arbitrary curve. Even with those generalizations not all developable surfaces are conical surfaces or developable surfaces.

 

Sheet metal work manuals probably offer a good introduction to the principles of developable surfaces.