Classic hull geometry

A correspondence with Reinhard Siegel on the hull of the 21 meter sailing yacht, project 08029.

The mentioned models can be found in H:\Library\Models\MultiSurf.

Dear Steven,

thanks again for sending the modeling problem. You are going to design a classic hull. In the attached file "Remarks on the modeling of classic hulls" please see some basic considerations. (The mentioned models were made some years ago; today the dragging of a point can be restricted, so all those transverse help surfaces are no longer necessary.)

To achieve a fair surface, regardless of its particular shape, I generally use the vertex line approach. Suppose, you have a nice boat and a child draws a net of longitudinal and transverse curves onto it. These curves can be very wavy, but the surface is still smooth. That means, bumpy parameter curves are NO indication for unfairness. One must check lines and surface curvatures.

But if longitudinal curves, that "plank" the surface, run in a harmonic fashion from stem to stern, then the surface is fair. And the easiest way to achieve this is to build in uniformity: use master curves (mcs) of the same type, use the same number of control points (cps) for all mcs, connect corresponding cps by C-spline guide curves (vertex lines) and make those vertex lines run in a harmonic fashion from bow to stern (like in a hard-chine hull). Then nice longitudinals (u-const parameter curves) and a fair surface is an automatic result. The labeling of the mcs will be similar, thus the distribution of curve points, which serve as supports for the "planking" longitudinals.

So far to the basic thoughts. Given your design problem I would tend to use either approach 5 or 3; the keel portion of the hull is rather small in relation to the whole body and the garboard not that tight.

So I did the following changes to start with approach 5:

0. Load your model. In order to keep my explanations simple, just the base hull entities were saved into a new file.

1. In that "hull" model, set "Decimal Places" to 4. The model lenght unit is meter, thus rounding is to 1/10 of a millimeter. Especially with C-spline curves and surfaces small changes can decide for example, whether an inflection occurs or not.

2. Set the v-division of the C-spline lofted surface to 8x8 since this surface is longer then wide. My rule of thumb is: if a curve, snake or surface extends in longitudinal or transverse direction more than LOA/2, it gets a division of 8x8 in this direction. Standard division is else 8x4.

During the design process I set the normally to 2. If the model is big or complex and response is slow, 1 is also fine. On export of data for building purpose the "Division Multiplier" is increased to 3, seldom to 4.

3. Created the C-spline curve vl1 connecting all control points at the top end of the hull mcs (master curve). Checked its curvature profile: the maximum in profile view is about midship; might have some more curvature in the forebody area - but this is astethics and not important here. Plan view: slight local maximum at mc3 (master curve nr. 3). No change.

4. Created C-spline curve vl8 connecting all control points at the bottom end of the mcs). Checked curvature; minor changes required to remove unwanted inflection points. No change. Not that important for the problem in question.

5. Created new mcs as B-spline curves type 3, using 8 cps. Why 8 cps? From experience; three cps will be identical at mc1 and mc2; if there are less cps, the distance between cps will be too large, thus little local influence on shape.

Of course, some mcs would not require so many cps. But if mcs will have different number of cps, one cannot run the C-spline guide curves through them, what is important for the fairing process.

6. Added vertex curves vl2 to vl7. Moved cps to achieve a) nice vertex curves, b) nice mcs with harmonic curvature profile and c) wanted hull shape.

When the proper number of mcs is established (not too few (= little local control in longitudinal direction), not to much (= risk for unfairness)) and the number of cps is found (not too few (= little local control in transverse direction) and not to many (= more effort to get smooth curvature)), there is just some fiddling with cps positions necessary, to finish the surface.

Very rarely I look at surface curvature; those vertex lines must run nicely, and the mcs must look good. I like very much to move points while in perspective few and looking along waterlines, for example. Such a view foreshortens the longitudinal distance and increases curvature, a very intuitive and effective tool for fairing.

See the attached hull_1.ms2.

One sentence to the bow mc; the rounding in the forefoot is part of the B-spline curve. Note, that the 2nd cp is a bead on the tangent line at the t=1 end of vertex curve vl8 --> tangency hardwired.

Now let us consider approach 3 - a canoe body hull, keel attached as a separate surface.

1. The hull is a typical C-spline lofted surface design, BCurve mcs, with vertex lines to coordinate cp positions.

2. The appendage is a B-spline lofted surface; 4 longitudinal mcs. Mc1 and mc2 are snakes on the canoe body hull.

This is model hull_2.ms2.

One will have to trim the canoe body surface along the mc1 snake, in order to remove the overlapping portions.

The advantage of approach 3 is, that for the appendage a variety of constructions can be used: an FLoftSurf with trailing edge mc, leading edge and thickness mc, or a BLoftSurf spanning a series of longitudinal FCurve mcs and two snakes on the hull to form a fillet and so on. One can use the minimum of mcs for the hull, and add a complex appendage. I used this separation for modeling a Dragon hull, that fits closely to the measurement rules, or a Lacustre or a "30er Schärenkreuzer" hull.

When you work on the appendage, you can be sure, that your canoe body will not be effected, whatever you do down there.

The attached examples are meant to illustrate my explanations. Certainly they will need some tuning here and there, but it is the principle that is of importance here.

Also, there is a 3da file attached, that holds your hull mcs. It is used for comparison purpose.

I will quit now to bring this matter on its way to you. I do hope, there is something helpful for you.

If you have any further question or if there is any other way I can be of assistance to you, please let me know.

With best regards,

Reinhard

Construction of a geometric model of a classic hull

I know about 4 approaches, how to model a longkeel hull. Each has its advantages and disadvantages.

1) Antiope.ms2 - the stern mc aft is a PolyCurve consisting of BCurve b1 and BCurve b2. b1 runs along the transom edges down to cl, b2 follows the stern overhang and keel trailing edge. Two curves are used here; a single one would require closely spaced control points at the knuckle (with strong influence on the surface mesh).

Advantage: direct definition of stern overhang and trailing edge. Disadvantage: the slope discontinuity at the end of b1 sometimes affects the station shape; fiddling around with the end t-value of the PolyCurve can help. Little local control of the stern overhang.

2) Pl-3.ms2: same story. Addional disadvantage are the "non-planar" mcs; indirect relation between mc shape and station shape.

3) Cutter.ms2: there is a surface for the canoe body, a surface for the keel, and a fillet, that creates the transition between canoe body and keel.

Advantage: independent definition of hull and appendage shape Disadvantage: SubSurfs or TrimSurfs required to remove overlapping surface portions.

4) Longkeel-x.ms2: One single surface; all mcs except the stem mc are in transverse planes. Note, that there are curves (vertex curves) running through corresponding control points (CCurve v1...v6). Also, control points are fastened to help surfaces. This allows the positioning of control points in perspective view (I love to look along waterlines, buttocks in perspective view and move this or that point to see the effect). If those vertex curves are fair, the surface is fair too.

Advantage: direct section definition by mcs in transverse planes; vertex curves possible (if no. of control points is the same for all mcs).

Disadvantage: no knuckle possible in fairbody profile; some boats show a sharp corner at trailing edge and bottom of the keel. Closely spaced mcs required at the end of the keel and at the begin of stern overhang. If the hull is a CLoftSurf, you will need more mcs than with a BLoftSurf, but the CLoftSurf is advantageous if you retrofit an existing design (you can put mcs where you have templates, and fit BCurves to the templates . This drawback can be reduced somewhat by "dependent" control points. For example, in model Longkeel-x.ms2, the top 4 control points of both mastercurves MC4a and MC5a are IntBeads on the vertex lines - so you just have to position the last control point, all the others are au- tomatically generated (of course, you cannot use them to shape MC4a or MC5a).

5) Longkeel-xyz.ms2: the surface at the stern is extended towards the negative side of the centerplane.

Advantage: discontinous fairbody curve possible at the end of the keel + advantages of Longkeel-x.ms2

Disadvantage: unusual extension of stern mcs beyond centerplane. Indirect definition of stern profile and keel trailing edge.

My favorable approach are 3) and 4). Minimum setup required with approach 1), but there is not much local control of the aftbody and the overhang portion.