hollow waterlines, rowboat

Daiquiri,

Holding both displacement and wetted surface area (WAS) constant for two versions of a practical hull, one with hollow waterlines and one hull without hollow waterlines, is not straight forward.

We can describe the problem with a thought experiment of a simplified hull.
Consider only the forward portion of a simple hull. A case where the forward waterline is a triangle, with straight lines. A slight concave, or hollow waterlines, will increase surface area and decrease displacement. A slight convex, or full waterlines, will increase surface area and increase displacement. We can adjust to an equal WSA, but we will need to change displacement with another parameter.

Consider the case of a full-lines hull and a hollow-lines hull with equal surface areas, with equal L, B, T.

If we decrease Bwl in the full-lines hull to reduce displacement, we decrease WSA.
If we increase Bwl in the hollow-lines hull to increase displacement, we increase WSA.
If we decrease Tc in the full-lines hull to reduce displacement, we decrease WSA.
If we increase Tc in the hollow-lines hull to increase displacement, we increase WSA.
Same will apply for Lwl.

We are left with the option to change the shape of the keel line, or canoe bottom curve. If we start with a full keel curve, a convex shape, we will have relations as above.

None of these parameters are complementary, which would allow the adjustment of the two parameters to result in changing displacement and re-adjustment to the original WAS.

There is an option however.
If we start the thought experiment with a straight line keel curve, we find a complementary relation in going from a straight keel to a concave keel curve in the full-lines hull, which will reduce displacement and increase WSA.

Looking at a simplified hull, we would need to move to a concave shape in the bow keel curve, along with adjusting any one of the above mentioned parameters, to create both a full-lines hull and a hollow-lines hull with the same displacement and WAS. It would work for the sake of a michlet experiment, but I don’t think it would represent very many practical hulls.

Did I over look something?

 

This is exactly what i did with hull 3 in comparison with hull 1 and yes, it was about can michlet show differences or not. hull 3 and 1 share the same LWL, BWL, CP, displacement, wetted surface and max draft. Can you open the delftship files? If not, i will post the hydrostatics of both hulls.

wo

 

Wolle, I cannot open your .mlt file. The Michlet windows shows for a fraction of a second and then shuts down.

 

Delftship produces mlt files for a previous version.
The attached should work with versions 9.20 and 9.30.

 

Yes, i uploaded the wrong file, where these two lines are missing, sorry.

here both hulls as .mlt, both should work now, and hydrostatics from delftship as ziped .html.

wo

 

I was not fast enough,

please dont use this file, or use it but correct the LOA to the value LWL is in delftship before usage, its in a line after the offset lines, value must be corrected to 5.671. Else you will get wrong results.
(See earlier posts)
Or use the last files i posted.

wo

 

You can avoid hollow lines in Series 7 by using waterplane shape factors
between 0.0 (rectangular) and 1.0 (parabolic). Using factors greater than 1.0
will produce cusped waterlines.
The small hollows at the ends you can see on-screen are graphical artefacts -
they are not hollow in hydrodynamic calculations.

 

Yes I know. We went over this some time ago. Short version was that I was getting better predicted drag when the waterplane shape factor was > 1, and the hollows I was seeing were not artifacts.

 

Pardon my senior member moment

 

No worries. I know you have to explain your program to so many people that they must all become a blur.

If anyone is really bored/interested, there's something they could try with Michlet/Godzilla.

With recreational rowboats, it seems that if you allow Series 7 hulls the algorithm will always spit out "optimal" hulls that have hollow waterlines. As far as I can tell, the reasons it does this are as follows.

First is that most recreational rowboats will be required to have a certain amount of positive metacentric height, which means the waterline beam will be far greater than a scull. This in turn means that Godzilla will want to trade off a lower prismatic against greater length to minimise total resistance. The lower prismatic saves some wetted surface, so the length can be increased, meaning the total resistance goes down a bit.

Michlet also likes to minimise rocker for these sorts of hulls, which means it generally wants to reduce prismatic by reducing the waterplane coefficient.

This leads directly to the next reason, which I think is purely a result of how the Michlet/Godzilla algorithm is coded. From what I can tell, Michlet doesn't have the ability to produce non-hollow waterlines if you want to reduce the waterplane coefficient. It seems to work by applying an exponent to a base ellipse, or possibly a sine curve, to produce the waterline shape. This will result in hollow waterlines (I'm glossing over the mathematical details a bit here, but anyone who has messed around with exponents and ellipses will get the idea).

Note that I can see why the algorithm for waterlines was coded this way, since it does give unlimited scope for waterplane coefficient.

Now for the "if you're really bored" bit. What would be possible is to manually create a series of hulls which used a basic exponential curve, of generic form y=kxⁿ, to generate the waterline shapes. If you wrote a routine to generate the offsets it shouldn't be too tedious.

The reason for doing this is that such waterlines shapes will give you the ability to generate a lower prismatic without having hollows in the waterlines, since such exponential curves are always convex. No, I don't need to be told about the possible reflex curvature when some such curves pass through the mathematical origin, since that wont be relevant to generating hull offsets.

Anyway, this would give a way of comparing hollow waterlines generated by Godzilla for Series 7 hulls, with non-hollow waterlines on hulls that have the same length, displacement, midship section and prismatic coefficient. That might give some more insight into how bad or otherwise the hollows are.

ETA: For those who aren't mathematically inclined, do note though that this method of generating waterline shapes is likely to get bonkers for waterplane coefficients under 0.55. The reason is that using basic exponential curves for the waterlines will tend towards a diamond shape at a waterplane coefficient of 0.5 (in which case the exponent would be 1).

 

This kind of undertaking is out of my skope in the case for this boat, but in avoiding hollowness i did something similar, i didnt use a formula, i used a bezier-curve with only two inner controll points at oposite distances from midship (lower distance means here more like a diamond shape) and at the same distance from centerline as a master waterline. I used a standart station (derived from hull ps2_1), wich is scaled in width by this curve and in height by another, then i introduced a derivated section with more width, but only above the waterline (to get more space on board toward the ends), and finaly a curve to control the smoth change from standart station at amidship to this extended station at bow and stern.

Well, it took some time, but finaly i managed to generate the hullform with the cad-software i am familiar with (because i use it everyday) and import it in delftship (import curves). This was not straightforward and others like rhino are much easyer in doing these things, but i liked to do the programming too . A small problem ist that delftship imports the geometry not correct if i set the control intervalls lower, so the result is now a mess of points with no way to make some further adjustments inside delftship.

I am happy with the result, resistance is reduced, the hollowness is almost gone and stability is sufficiant for my porpuse. With the information given by NoEyeDeer, the boat should be hard to reach 6 kts single rowed, maybe only for a very short period, but with 2 things look better. But to reach 4.5 with 30N in cruising is a nice imagination.

I have to adjust things for optimised rowing position, freeboard (25cm) seems a bit low, but i know, there is another thread(s) talking about this - i have to read it again.

wo

 

IMHO, freeboard is fine for sheltered waters. I would expect such a boat to be very hard to push to 6 knots at design displacement with only one person rowing, but your DWL gives enough volume for two people anyway.

One thing: with a forefoot that deep I think you will have tracking problems.I think some rocker for'd would be advisable.

ETA: By the way, you'll find the hull a lot easier to fair with fewer stations and control curves. You shouldn't need that many to get the shape you want.

 

I did not any fairing inside delftship. Because i did not get the shape i want (without hollows) with few control points (and curves) with hull ps2_1, i computed this new hull outside and imported it then into delftship, to get the hydrodynamics. If i choose a lower number of control intervall during import, the hollow lines reapear. I mentioned that in the previous post.

wo