Introduction and a Question

Hi there,

I've been lurking on these forums for a while now but now I have a question so I thought I'd better register and Introduce myself before asking.

I'm a software professional in Adelaide, South Australia and I just like messing about in boats.

I'm building a John Welshford design, Navigator, in my shed at the moment and have become quite interested in small boat design while doing so. I'd like to design my next boat myself and I'd like to design a different sail rig for the Navgator when I've finished the Hull.

I've obtained and read Skene's book and am currently going through the Larsson/Eliasson book where they mention the use of "Copenhagen ships curves" to draw body plans.

My question is how would one go about drawing out the shape of these ships curves. Is there a specification for the maths algorithm's that define the shapes, or is it something that has evolved over the years and is a closely held secret by the makers? I cannot find any documented history on the development of these instruments on the Net. I have seen sets that were obviously hand made, but they may just have been traced out from an existing set.

So, who made the first set and how did they do it?

Many thanks

Andy

 

Hello, and welcome to the forum!!! Sorry that I can Answer your question...

 

Those set of curves are named generically. Each manufacturer has a sumilar but somewhat different set of curves. Many of us prefer to use battens. They can be bent to any shape without the restrictions of an algorithm or a curve manufacturer. I think that if whatever designing program has curves that look pleasing to you, then it's OK. I always loft lines full size, because it is impossible to see them well in a small screen. For example, half an inch difference in a sheer line forty feet long, makes an ugly boat beautiful. Those lines where quite difficult to master. The idea behind them is that they do not follow an algorithm. In other words they are not regular curves. The drafstman would use parts of them and splice them to obtain whatever compound curve he was looking for. I find a set of battens and ten or twelve lead pigs are much easier and faster to use.

 

In terms of algorithms, the battens and lead ducks are drafting splines and many CAD programs (eg HULLFORM) use cubic splines for producing longitudinal fair curves, with control pionts replacing the ducks. Transverse sections with complex shapes are often modelled by sequences of bezier curves, which probably approximates the splicing together of parts of hand drawn curves using ship's curves as templates.

-Charles Magnan

 

Welcome,
There are a large number of methods available in various algorithms. the two most powerful are probably the B-Spline and NURBS systems. There are various peices of code on the web for these.

My only warning is to watch out for the different definitions of the NURBS curve/surface (if you want to write it yourself). The control point weighting should be applied outside the basis function, not inside it.

I am currently writing the basis of an open-source Hull Design package. It will contain the basic curve tools you're after. If you want to develop the code drop me an e-mail. If you e-mail me I can let you have a copy of the B-spline or Nurbs algorithms (in Fortran 90 or C), there is no great secret involved in the code itself.

Good luck,

Tim B.

 

For Copenhagen shipcurves, see MacNougthon, www.macnaughtongroup.com,
http://www.macnaughtongroup.com/copenhagen_ships_curves.htm
http://www.macnaughtongroup.com/yds_drafting_tools_order_form.htm
For small boats it's probably better to use flexible battens to draw the curves.
For simple and inexpensive design programs, see Prolines and NavalDesigner.

 

I think I use the Gonzo style of curves. Straight grained piece of wood. Place it in tension to give a close approximate curve. Then lay another piece of wood along the bowstring base. A couple of lines at right angles, fairs in any curves as required. It is right, when it looks right.

 

method of drawing curves

AndyC,
"My question is how would one go about drawing out the shape of these ships curves."

"Ships curves" can be drawn by creating a series of circles, whose radius vary, where the lines decribing these circles' circumfrence are tangent to one another AND the radius change between any two adjacent arc segments results in arc segments that are "fair". Fair being the non-mathmatically defined flow of curves, arcs and lines which don't have a point of deflection or descernable transition which is historically called 'hogged'. Fair was described above as when it looks right.

The author and programmer of the New Wave System's application Prosurf had a good paper on the topic at his site. http://www.newavesys.com/news.htm
Steven Hollister discusses the various aspects of nurbs and b-splines regarding software and fairing.

Back to laying out your own templates. The series of circles involved isn't constant between all drafting tool vendors; but with some paper you can verify the design. Trace any ship's curve, french curve or other fixed curve drafting tool onto paper and you can spend as many hours as you'd like discovering where these centers are, as well as the radii.

Some of the edges are parabolic or eliptical but even these can be drawn by graduating radii of adjoining arc segments and 'fairing' the lines at the intersections.

My guess is that scale, accepted historic vessel sections and lines as well as the practice of different groups of shipwrights influenced the most often used curves which were then made into hardwood tools to speed the batten process in smaller physical size work.

Today these drafting tools most likely conform in dimension to imperial and metric scale drawings that will fit on standard paper sizes.

Cheers,
kmorin

 

The pocket calculator and a B-spline sounds like less work Kmorin.

Go on, write a B-spline curve routine then plot the answers.

Tim B.

 

It is less work. That is how most boats are designed today and they are mostly ugly.

 

AndyC

Welcome

Forget Gonzo's serbic comments. The design comes from the designer not from the design method.

Forget the traditional archaic methods, they were slow and innacurate, the methods you refer to given in Skene's are no longer practical.

The computer can draw a fair curve. The output from the computer is absolute (no errors to be corrected at lofting). The design process is considerably easier and faster. All the critical data, displacement, area centroids, floatation etc are absolute and error free.

For example for a 60 footer you can change your lines, generate a hull get a curve of areas, COB, and Displacement. 4 or 5 times in one morning. Honing in on to the desired characteristics. The paper and batten approach is a very poor option given todays computer tools and no current pro. designer worth his salt would contemplate hand drawing other than a concept sketch.

The computer has revolutionised the design process. Embrace it. You don't need any fancy computer either, a 500Mhz 14inch laptop is more than adequate for the job.

Software wise you can email me for any recommendations or help. (Just click on my user Name on this post)

 

easier is 'better'

Tim B, I was responding to the question posed in the post, not advocating the hand creation of a set of drafting templates! AndyC asked about the theory of layout of ships curves and I was answering (well trying to).

The thread got a bit off the path to an answer to AndyC's question about the history, pattern or conventions that are implied but not stated by the presence in the designer's tools of these curves. He seemed to be interested in the history of the device so I was trying to relate what little I knew of them.

I use Prosurf, Rhino, Maxsurf (student edition) Sketchup and ACAD, an E sized scanner and all of our CAD stations have D or E sized plotters. I don't use templates unless I get forced by correspondence courses (Westlawn wants to be sure you can use a pencil) and find board drafting tedious. I've even given up paper for lots of sketch work and use the Cintiq18 from Wacom running Alias; I'm an 'easy-way' advocate but none of that would answer the question of how to layout or where have the ships' curves come from.

Cheers,

 

Serbic?
As I said before, it is less work;particularly if you have no experience. Computer programs give fair curves. However, they are mostly ugly. Also, the programer limits your options. I prefer to control my designs.

 

Gonzo

You are talking about hull generation software, which is a quick and poor way of producing a hull to someone else’s design ideas. I agree.

However some modern PC based CAD packages can completely replicate the older hand drawn environment without error, with more control and with analysis of the curve itself.

After generating the hull from your own curves you have far more powerful downstream tools available.

I too had a prejudiced view of PC based CAD for hull design, when we moved to Autocad some years back (since the industry demands numerical output now) the curves were generated in a spreadsheet. Now the modern Nurbs based CAD eats hull design and is a joy to use.

 

There are a couple of reasons I don't like to design aesthetically in a computer. The screen is too small to show the lines. The pixels are too big to show small imperfections. I like pencils (don't laugh). For calculations, displacement, CG,CF, etc. a computer program is way superior. I can use a planimeter and do it by hand, but it is extremely boring and time consuming.