It is very difficult to see and understand a panel with twist in a serious of section cuts.
However, generally you should see a gentle continuous curve, not something with a quick curve in an isolated section of the cut.

It is hard, especially since cross sections with straight lines have the highest amount of twist along ruling lines, i.e., the cross sections that look best are mathematically less developable (if you use twist as a measure).

I believe you are not generating a simple twist to a panel, rather something which has a localized bend.
This is probably what Bill Hamm identifies as a "chine".

The reason this geometry will not work in the real world is what Nick said. The actual material cannot be bent in this manner and stay in this shape with only supporting it at the edges. The stiffness across the panel inherent in the material will prevent the quick localized curve.

The localized bend is a result of the slant in the ruling lines and actually goes away if twist is introduced. In my original post I described this as a bulge or soft chine. From my original post:

Despite being developable, I do not believe the shape that is described by these ruling lines to be buildable because the curvature in the direction perpendicular to the ruling lines is very high. Furthermore, plotting cross sections through the ruling lines shows that this shape has a pronounced bulge or "soft" chine below the hard chine described by the Bezier curve outside the center of the kayak.

I think this is the same thing Bill Hamm referred to as a chine, and it is the problem I have been trying to overcome.

You are going to have to check the analysis sometime.

I have checked the analysis many times. Unless I keep making the same mistake, it is correct. The code produces correct results for test cases that have a known solution, so I am fairly confident that it produces the correct ruling lines at this point, but I will check again.

Will you do it at full size or at a scale size?
The cost is relative. What does it cost for the 0.8mm ply?

It is about $60 for the 0.4mm and about $30 for the 0.8mm for a 4'x4' sheet, plus shipping from Florida. I guess that size would make a few models.

I was planning to make a scale model first, but not of the hull I showed in the picture, that was just for illustrating the problem.

Are you using a 3-D modeler to generate the actual surfaces you used to make the cuts? If so you should be able to generate a view of the panel, then find the cuts which show a straight line across the panel, if it is truly a develop-able panel.

I used a very simple code to compute the intersections of ruling lines with the cut planes and output the points to a text file which I then plotted using gnuplot.

I have generated 3D models from the data and it shows the straight lines you mention. However, having a straight line in the surface is only a necessary condition for developability, not a sufficient one. The surface normal must also stay constant along the ruling lines. In discrete models that use boundary triangulations, even this condition is fulfilled in the case of non-developable surfaces because the discrete models have creases.