My F16-class pontoon design (comments/suggestions?)

[Guest625101138]

I compared your hull against the lowest drag hull for 10kts displacing 254kg. I am assuming this would be the case of reaching with one hull just lifting with 2 on board. The speed selection is not based on any experience with the F16. If the selected speed was lower the block coefficient would be smaller.

You would need to consider a few points of sailing and the expected speed to do a proper analysis but I find it useful to know what the lowest drag displacement hull can achieve as a starting point.

Rick W

[Ad Hoc]

"...What causes the better strong wind performance of flattened-v profiles.."

The hull, when heeled, acts like a lifting surface, owing to the 'V' at the keel. Obviously not a pure foil, but a "flat" lifting surface that generates, in some instances, lift, over that of a smooth 'U' shape which does not.

[CTMD]

Rick, any chance of showing us what you consider the optimised hull shape for 10 knots?

[Guest625101138]

Originally Posted by CTMD

Rick, any chance of showing us what you consider the optimised hull shape for 10 knots?

I attached the .igs file of the hull. This is the lowest drag hull. Only the underwater portion is relevant. The draft is 230mm. I have not considered any aspect above the waterline. The hull is straight out of Godzilla. I would normally do some fairing and then recheck the result because I do not like the bumps.

I have avoided the word optimum as there are a number of factors to consider in addition to drag. Turning ability being one that is obvious for something that races around buoys.

I have had great success with Godzilla so I take a lot of notice of what it produces. Getting the design speed right is an important factor of course.

Rick W

[CTMD]

sorry didn't notice the second attachment.

[Ad Hoc]

drag values are meaningless without:
LWL, Beam, Draft, Speed, as a minimum.
And what drag, total, Rf, Rw ??
A graph showing lines is just that a graph showing lines...without any real world reference it has no meaning

[Guest625101138]

Originally Posted by Rick Willoughby

I compared your hull against the lowest drag hull for 10kts displacing 254kg. I am assuming this would be the case of reaching with one hull just lifting with 2 on board. The speed selection is not based on any experience with the F16. If the selected speed was lower the block coefficient would be smaller.

You would need to consider a few points of sailing and the expected speed to do a proper analysis but I find it useful to know what the lowest drag displacement hull can achieve as a starting point.

Rick W

Note that I have edited this post.

In loading the .igs file I did not turn off symmetry so I had a double hull. Made the Michlet file export nonsense.

The proposed hull is almost identical to what Godzilla can produce so it is very good under this condition. Even slightly better than Godzilla under 5kts.

Now I am wondering how you achieved it? I have never been able to do better than Godzilla.

It would be interesting to compare in condition of equally loaded hulls but that will be another day.

Rick W

[kerinin]

The two hulls seem to have virtually identical performance; what difference there is I would say falls within the margin of error for the simulation algorithm.

As I said, the whole design is essentially some nudges and pulls on a spherical shape; it would appear that elliptical curvature does a good job in reducing pressure gradients on a narrow hull. The horizontal bow profile is the result of trial and error; it's essentially a degree-3 spline that started as an ellipse.

Another thing I just remembered reading in the Michlet documentation as I was thinking about how I might have accidentally outperformed the algorithm is that wave drag decreases with the square of depth. So just based on that, it would seem that burying the fore mass of the hull and using a 'raked' bow might serve to reduce wave drag? Does Godzilla produce any raked bow shapes? (I haven't opened the IGS file yet).

I'm sure someone's mentioned this, but I've been taking the Michlet results with a grain of salt; the drag numbers Michlet produces are fairly different that what Solidworks is producing in similar flow situations, so I've been concentrating mostly on the differences between hulls.

[kerinin]

The hull, when heeled, acts like a lifting surface, owing to the 'V' at the keel. Obviously not a pure foil, but a "flat" lifting surface that generates, in some instances, lift, over that of a smooth 'U' shape which does not.
Reply With Quote Multi-Quote This Message Quick reply to this message

I might start a thread (or look for one) about this, because I keep hearing about dynamic lift and it doesn't make much sense to me. I'm mostly confused by the role the water surface plays in generating this lift.

[Ad Hoc]

kerinin

It is somewhat analogous to that of a deep V planing hull when banked, see attached. Does this help?

[Guest625101138]

Originally Posted by kerinin

.......
I'm sure someone's mentioned this, but I've been taking the Michlet results with a grain of salt; the drag numbers Michlet produces are fairly different that what Solidworks is producing in similar flow situations, so I've been concentrating mostly on the differences between hulls.

I have validated the Michlet result on a variety of slender hulls for froude number up to 0.7.

With the most accurate power measurement we have used I was getting within 2% of design estimate against actual from pedals to boat speed. Leo even adjusted the viscosity limits so we could get values to suit the cold lake water testing:
http://www.youtube.com/watch?v=0yk7tkX9mOY

I am impressed with what Godzilla produces and you have done well to get virtually identical results in a first attempt.

Rick W

[kerinin]

Rick:
I just took a look at your IGS file; that's very interesting, I wonder if that design is optimal due to some quirks of the algorithm. I'm a brick-and-mortar architect in my day job, and one of the things we've been discussion around the office is the possibility of using genetic algorithms as a design tool for buildings; giving the computer a set of spaces and adjacencies and allowing the computer to produce a set of spatial arrangements that optimize whatever constraints we set - it's called 'generative design'.

Looking at your hull shape and mine, it reminds me of something that often comes up when dealing with algorithms like GODZILLA; local minima. Gradient descent algorithms such as what's probably used for GODZILLA are very susceptible to 'valleys' in the optimization curve that are surrounded by less optimal values but which aren't the 'true' minima for the curve as a whole.

I would be curious to see what GODZILLA produces if you use my hull as a starting point; would it converge on your optimal design (in which case your hull probably isn't a local minima) or would it converge on something different?

This is getting interesting...

[Ad Hoc]

"..Leo even adjusted the viscosity limits.."

I don't understand this comment, care to explain?

When performing resistance calculations one cannot calculate without knowing the Reynolds number, this requires the value of the viscosity of the fluid which is also related to the temperature of the fluid. It is the first thing one checks when doing tank testing....so i don't understand what is being suddenly adjusted?

[bad dog]

Gee you blokes have been busy! One day at work and look at all the feedback! Some of these posts have some really good info and are beyond what I can usefully remember from my days enrolled in Naval Architecture at UNSW back in 1974, which was tug boats and trawlers - so I'll contain my comments to 'user experience' rather than 'theoretician' (for fear of making a goose of myself).

Re raked bows: strongly raked bows seem to shed all but plastic bags quite readily (why are bags still legal?). Plumb bows do not: seagrass builds up on mine all the time. I reckon the steeply reverse raked bow that drives hard with the bow 100mm below waterline have a terrific wave drag advantage - once again this just seems intuitive. Hull length and Froude numbers go into the mix somewhere, but I still wonder if there is a theoretical explanation?

Re vee in hulls: as I understand it, and corroborated at least in part by Ad Hoc above, is about planing area. Yes I know cats like F16s and As don't truly plane, but flat sections generate more lift than curved sections. Ron Given's designs have relied on this for years (Paper Tiger et al), and most current As have a definite shallow vee flattening of the aft sections. The benefit of the boats that have this (compared to those that don't) is evident especially off the wind.

Re rocker: interesting comments on the LR2 and LR3, which are longer hulls with less crew weight. F16 two-up must do a bit more work than an A with 75kg crew! Light air upwind in my humble experience calls for more rocker, crew weight way forward. This is where the non-vee (and perhaps heavier) hulls can match it with any others. Turn the corner at the top mark and everything changes - the flatter vee hulls just streak away, which the LR2 experience backs up. So maybe this decision is informed by the local conditions you are expecting to sail in most: heavy/light air, choppy/smooth water, (and maybe plastic bag strewn/pristine! ;-)

[Guest625101138]

Originally Posted by kerinin

........
Looking at your hull shape and mine, it reminds me of something that often comes up when dealing with algorithms like GODZILLA; local minima. Gradient descent algorithms such as what's probably used for GODZILLA are very susceptible to 'valleys' in the optimization curve that are surrounded by less optimal values but which aren't the 'true' minima for the curve as a whole.

I would be curious to see what GODZILLA produces if you use my hull as a starting point; would it converge on your optimal design (in which case your hull probably isn't a local minima) or would it converge on something different?

This is getting interesting...

Leo advises that there are two lowest drag hulls but I have never seen this. It does not matter what hull you start with it will arrive at the same form each time. Sometimes it can take many attempts to get going so I usually start with a hull that I know will be nothing like the optimum so it kicks off in a few seconds.

I normally use the 7 function hull form but I have the 18 function series as well and this will produce quite convoluted hulls similar to things like the Ward Optimum Symmetrical Ship form depending on what objective function is set.

The bumps on the hull give the lowest drag but fairing them out does not add much drag.

Rick W