Torpedo boat stern in Rhino3D -- how to edit?

From the national archives I have some drawings for the USS Ericsson, a steam torpedo boat of 1895. I will try to make a 3D model from these in Rhino, but I am having trouble getting a smooth and fair hull, particularly at the stern.

The hull is essentially double ended, and the stern tapers sharply to form a "torpedo boat stern". Viewed from a certain favorable angle it looks okay, but from any other angle, particularly from below, the surface flaws are evident.

The underlying curves are fair and have been rebuilt or refit. I have created the surface via multiple construction paths and this seems to be about the best one. Everything I have attempted involves editing curves, however, not the surface itself.

Could I edit this surface in some way to make it relax into the shape suggested by the underlying frame of lines and curves?

Thank you for your insights and suggestions

Michael

 

You might want to try building it quite differently. If you put a 'Point' at the most forward part of the bow bulb and loft through some sections it will give a different result. You may need to combine this technique with the technique you have shown.
It is always a bit hard to be absolutely specific because there are at least 3 or 4 ways of approaching this kind of modelling. My gut feel on this one, would be to use a loft through sections which is mostly pretty good.

If you go to the Rhino tutorial of the Dinosaur or T Rex or whatever it is you will see the value of a 'Point' being the end of a Loft surface. It should then give a nice smooth rounded bulb. If say you were modelling a 2.4 Meter Class boat, you would model two hull shapes and intersect and/or blend to get the kind of shapes they have.

You are correct in using the Curves to get the Surfaces. Rhino rewards carfully constructed Curves so check Degree and number of nodes/control points. The simpler the better. Much easier to edit curves first rather than surfaces. Your curves look OK but it does not look like a longditudinal lofting or sweeping surface using Section curves that have created the surface shown.

 

Thank you for your suggestions. I will look for the dinosaur tutorial.

Most of the several strategies I have tried so far involve a Surface from a Network of Curves. The idea is to create a half-hull model using the shear line; the keel line; and a few (3 or 4) sections. The sections and shear line are degree 5 curves rebuilt with 6 control points. In this particular hull, owing to the compound curves of the stern, NetworkSrf was used once for the forward part of the hull and twice for the stern, which was cut up into two distinct surfacing problems.

A surface from a Network of Curves works nicely if you have a cut-off transom, but these curvaceous torpedo boat sterns were initially designed for US Navy torpedo boats by the sailing yacht designer Nathanael Herreshoff. He was coaxing high speed from a displacement hull. He achieved a beautiful line but it is tricky to surface in CAD.


Some progress. The simplest NetworkSrf approach, which is to run a complete half hull as a single surface, gives a pretty good result, except for a huge puffy unexpected laugh line at the stern.

I watched a tutorial on Rhino surface editing, and decided to try to get rid of the laugh line. In Rhino you can light up rows of control points using Sel-U or Sel-V. I used Sel-V in this instance to select every other row of control points underlying the laugh line bulge -- and then deleted all three rows. The unedited port side and edited starboard side results are shown side by side, both as wireframes and as renders.

To restore the keel line after this edit, it was necessary to fiddle with two control points on the center plane, using the Right view. Many remaining problems including a lack of tangency at the half hull junctures.

 

TBH Surface from Curve Network is probably the last command I would use to get a fair sweet shape. Mostly the surfaces are too complex. Much better to use Loft or Sweep 2 rails to get the best result. You might need to to use full section curves - and it matters how you select them because of he Normals.
In the worst case you should be able to get 95% accurate and then using the tangency control in the Surface from Curve Network tweak out the last bit.

I would be tempted to try the whole hull not a half hull on this shape. Hence joining the section curves (each half). I would not use any buttock or waterline curves except to find how close they are to your resulting surface. If you put a plane through at LWL and use Curve from Intersection of Surfaces that would give you the static waterline from your shape. You may use multiple planes to get other waterline/buttock lines and diagonals. This allows you to check these lines against your original lines. There may well be a discrepancy, 3D CAD will be more accurate than the original pencil drawings so a bit of leeway will be required.

The reason Loft will work better as a whole hull is because the Point will be resolved better and more evenly. You may find the best option is to free the sheerline a little. If you cut the 'hull' surface with another curved surface it will give you a sheerline. It doesn't actually matter if the original surface is closed unless you intend to get it out of a GRP mould ie you need draft taper. My inclination would be extend the section curves up to a horizontal plane bar the torpedo ends. Loft the surface, then cut later with another surface to get the sheerline. The sheerline is a minor part of the design ie it needs to be low enough to let you paddle but high enough to give wave protection. The exact height being defined by the last mm is not helping you get a sweet surface. A few mm of float which you would get by modelling with Loft or Sweep 2 Rails will not be detrimental to the design. You may even discover something about the shape.

 

Below is a really crude attempt at getting a smooth torpedo stern shape using the loft command and a Point at (in this case) both ends. The surface was later trimmed to another to give the sheerline.

 

mcg, What information do you have on the shape of the hull: a full set of lines, offsets, an outline of the hull? What level of accuracy is needed; does it just need to "look right"?

 

SukiSolo, the hull looks smooth and fair. I will try your method, thank you for suggesting it.

David Cockey, altogether I have 9 drawings from the national archives -- kind of a grab bag selection from 1895. All the drawings are now in the form of TIFF files.

For the hull I do not have waterlines or buttocks, unfortunately. I do have all the sections, plus a clear side view marked with section numbers. In addition there are several top views.

The rest is technical detail of interest to modelers, e.g., a deck crane, steam engines, et cetera. From a CAD point of view this stuff seems straightforward. The hull is not.

A number of photos appear in this excellent book:

http://www.amazon.com/Building-Mosq...qid=1382561187&sr=8-5&keywords=mosquito fleet

From a comparison of the drawings and photos, I would judge the hulls were modified after the drawings I have were made. The drawings show a much sleeker torpedo boat stern than the photos.

The main objective is to improve my Rhino. I think it should be possible to get a TLAR hull from these resources, but so far I am getting curious ripples, distortions and near-creases in the surface, chiefly at the extremes of the stern and the bow.

I will next try a 2-rail from point to point, as suggested above, and see what results.

Thank you for your insights. Michael