Round Chine Developable Surface Hull

Attached is a file of a sailboat hull shape with developable sides and bottoms, and constant radius “chines”. This hull shape was created as a demonstration of the method used, and is not intended to be used to build a sailboat. RC35-012.3dm (452.1 KB)

Note: This type of hull shape is sometimes used in metal boatbuilding because the sides and bottoms can be quickly plated using large sheets bent the framework. The round chines can be built using constant radius rolled sections. A similar approach is sometimes used with wood-epoxy construction.

Method used:

Design initial mid-ships section shape. It has straight segments for the upper and lower panels and an arc fillet connecting the straight segments. The “virtual chine” (intersection of extended straight segments) would be determined using this drawing. Additional sections shapes were also initially created for and aft.

Design longitudinal curves for sheer, virtual chine and keel. I used degree 3 curves. These curves were extended beyond the bow and stern.

Exact ruling lines for developable surfaces were created between adjacent cuves beyond the bow and stern using the method described in Developable surface - exact ruling lines from edge curves. Exact ruling lines were also created at intermediate locations. Longitudinal curves were copied and trimmed using the end ruling lines.

Developable surfaces for the side and bottom were created using -DevLoft (note hyphan version was used) based on the trimmed longitudinal curves. The intermediate ruling lines were used with the AddRuling option.

The side and bottom surfaces were simplified. Rebuild was used to simplify surface. Various numbers of control points in the u direction were tried to find the balance between number of control points and deviation from the original surfaces. Developability was checked using Curvature. The principal curves were visually checked to verify that one curve remained essentially straight as the cursor was moved over the surfaces. The numerical values of curvature was also checked at numerous locations on the surfaces to verify that one principal curvature remained essentially zero (10e-6 or smaller).

FilletSrf was used to create the round chine between the sides and bottom. The surfaces were not trimmed by the fillet.

The side, chine and bottom surfaces were trimmed at centerplane. The were also trimmed by at the stern. Then the side, chine and bottom surfaces were trimmed with each other.

Repeated the process again using modified longitudinal curves which improved the surface shape

Surfaces were mirrored.

Stem was rounded using VariableFilletSrf. Rounded transom surface was created using an arc which was extruded. Deck was created using the method described in Deck Method

1 Like

Hi David, I want to look at your hull when I am near a good internet source.
I looked at the Devsrf command and didn’t see much documentation on That command. The one dialog that comes up doesn’t seem to work. I know it’s not your problem, just a comment on that command.
I have used Devloft before ok. Does sweep 2 do the same thing as long as the cross section Curve is a straight line? Minus the ruling lines?
Anyway, it nice you want to share.—- Mark

With a straight line as the cross section Sweep2 will result in ruled surface but generally not a developable surface. All developable surfaces are ruled surfaces - BUT - Not all ruled surfaces are developable surfaces. Sweep2 can provide a (close to) developable surface With a sufficient number of developable surface ruling lines as cross sections. The more developable surface ruling lines the closer the results will be to an “exact” developable surface.

DevSrf compared to DevLoft deserves it’s own thread. In brief DevSrf when it works usually gives good results. However DevSrf is easily confused which happened with the curves in the example above. DevLoft is more reliable but suffers the serious flow of always continuing the resulting surface to the ends of the defining curves, even in the usual case of the corresponding ends of the defining curves not falling on the same ruling line. This causes the results of DevLoft to frequently not be accurate near the ends of the defining curves.