Basic loft issue for ship hulls

Hi Guys,
I’m really ashamed to ask for this since it is obviously a very basic issue which has been answered countless times, but I just cannot get the bow of my ship to loft kindly with the rest of the stations.
The issus is the curvature of the bow line, which is in the XZ plane, while all other stations have an axial curvature.
When lofting, I can’t seem to find a way to properly change the direction of the curvature at the bow edge of the surface. Therefore, there is some kind of singular point where rhino tries to change the curvature direction…
The surface looks good but I cannot add thickness to it without getting errors in the direction of offset at that location.

I mean, everything is quite logical and I understand why rhino actually does that. But I tried A LOT of things to work on the surface and fix this but I can’t find anything proper. I used the hull tutorial of rhino as much as I could without success so far !
It seems to me that the most sensical thing would be to be able to change the direction of the curvature at the edge, but I couldn’t do it…

Any help would be appreciated !

Nicolas

hard to judge from a screenshot - can you post the file ?
my guess:
you have a problem like this at the lower corner:
if the 3 CVs at the corner (left) form a nice triangle - everything is fine
if the 3 CVs form 180 degree - already not clear (middle) but seams like rhino can handle it
if the 3 CVs form a corner that is more then 180 degree - the normal is pointing to opposite side


_dir command showing normals
green surface = offest

EDIT
check out this older topic:

Well, actually lofting for a ship hull may not be the best way to go. Looks like you have a very “dense” surface to deal with which is hard to edit. You cold create a basic surface with as minimum CV’s as possible - say degree 5, 6 CV surface- and push-pull CV’s to get a decent/fair surface. You can create a loft surface of course, but probably you’ll need , again, curves with as little CV’s as possible with all same CV count - degree5 6CV curve to start with. And create a loose loft to edit easily.

But if you’re happy with what you already have, you can create a big surface on Y plane, like a couple of millimeters away from the hull’s longitudinal symettry axis, and trim the hull. That way, you can get rid of that problematic point and create a better offset (or blend surface with the mirrored hull). I believe this will work, but if not, maybe you can share the file and someone will check it and suggest something different ?

this might be interesting:

and this:

Thank you very much for your kind
la_belle_test_ecport.3dm (256.3 KB)
replies, I will look into each of them.

for further information, please find attached the file. I lofted the surface and displayed the curvature.

I must add that this my aim is to design this od ship at 1/36 scale and CNC mill most of the parts.

I agree that in my previous attempt, I trimmed the hull in the Y axis and ‘repaired’ the surface. But the issues with the offseet propagates a few millimeters along the X axis. Thus, It remains an issue.

Being of the utmost accuracy is not a problem for small ship modelling but I fear that if I do not resolve properly this issue, giving thickness to this hull will be a nightmare

Looking into various answers, it seems that what I’m dealing with a a ‘degerate surface’ where u and v share the same edge of a surface at some point. And this is a bad thing.

I can easily mirror your surface and add some surfaces from edge curves to make a watertight solid, which should be good enough for milling the hull shape from a solid piece of material. That would be the most simple solution, and maybe good enough for your problem. If you want to shell the hull either you have to repair the naked edges, or build better surfaces.
labelle-jgh1.3dm (238.8 KB)

I do not intend to mill only the external surface but also the internal for each frame.
So I’m willing to go a little further for accuracy
In the end, here is what I got with my previous attempt:
It was quite nice but I did not loft the insides of the frames but rather added 6 mm to the thickness of the surface. Therefore the accuracy of those frame was entirely unsufficient for my purpose.

That is why I intend to loft both surfaces and link them as a closed polysurface before using boolean operations with rectangular frames along the axial axis to get my final frames. But It was my idea that to get properly closed polysurfaces both lofts had to be as clean as possible.

EDIT: I tried one of the provided method ( ‘trimmed surface’) to get what I want using a ‘fake’ bow out of axis to loft a bit further than needed then trim the surface along the Y axis. It is not that bad so far.


Yes, I fiddled a bit more with it, and it should do the trick.
At least, with some more use of the SoftEditSrf function, the bow should be round enough while keeping the surface itself clean enough for the next steps. I’ll keep you updated.

Allright,
I found a solution, albeit it lacks elegance…
Took me a while.

I used a first untrimmed external surface with the bow prolongated in the X dan Y direction such as in the previous picture (method C of the ‘bow tutorial’ on this forum). This surface had a very low nuber of control point. I faired it as much as I could.

Then, I used a variable distance offset to create the internal surface of the frames. for instance, the upper end of the frames ranges from 4.5 to 6 mm thickness. the lower end up to 9 mm. I had to tweak it a lot to get something rather accurate and rather fair (which is less important for this area that won’t be as visible as the outside of the model).

Third, I connected the two surface with planar surfaces. I joined all 6 surfaces.

And Voilà ! A closed polysurface. The last step was to Wirecut the solid.

It is surprisingly accurate, with around 1mm max error as compared to the blueprints from the monography of this old ship. Considering the material will be wood, It should be ok for my purpose.

Thanks for your help. I hope this method will help someone.

Beginner question, but: is there a way to rebuild the solid to get less edges?

1 Like