Godzilla

Andrea
The performance will depend on wind conditions. This boat will not perform well in light wind compared with a well designed monohull. In stronger wind it will go better.

With a hull shape like this you do not need to build it with the bumps that Godzilla produces. Each hull could be made from 4 pieces. A keel, two sides and a deck.

It is easy to make it hard chine and then merge all the unused contour lines and points.

If you do this you can check with Michlet that the drag has changed little.

Rick W

 

Andrea
The above comment is meant for the hard chine hull shown in the Freeship image.

The rounded hull in the offset file will need a different method of construction.

Rick

 

Rick, Leo,
Thanks for repeating me things twice, I need it and appreciate.
I’ll be looking at longer things ( 4m), smaller displacement (0.1) m3 and 2 m/s velocity.
I will not go hard chine.
Andrea.

 

Thank you Rick for taking time to answer

Of course, the purpose is to consider one hull in the water with full displacement and appropriate crew position @ different speeds.

As 1-Michlet & Godzilla only consider underwater hull sections
2-Wave drag effect of underwater volume decrease with square of depth or something similar.
3- I saw somewhere than the "spray drag": the drag generated by above waterline hull'sections crossing choppy water is proportional to beam^3

So the basic idea is to achive an hull section which is slightly larger below the waterline than above , not only at the bow level like today's pierce-wave design, but also at the main hull section level.

I have a lot of input from crews and speedtests in different sea conditions, but the challenge is to put the Capricorn, Nacra Infusion & HB Tiger hull's section on Michlet.

But according to your comment, I think the simplest approach could be to run Godzilla with appropiate inputs and compare to datas available for calm water.

Thanks again Rick

Regards

Erwan

 

The hulls I generated with Godzilla so far have LCB = 0, I understand Godzilla measures LCB from the hull centre. All the hulls I have so far are symmetric in respect to the central section. I tried to constrain the max LCB to a negative value but it does not look to work, the hull isn’t optimized. Is it possible to have hulls with LCB<0 with Godzilla ?
Andrea

 

Hi Andrea,

When you tried to constrain the LCB, what exactly happened? Did Godzilla shut down immediately, or did it iterate really quickly without making any changes to the ship?

If one of the multitude of hulls generated by Godzilla violates any of your constraints, the program will discard that hull. So if your initial hull violates your constraints, so will just about anything spawned by that initial hull. Nothing meets the criteria, thus nothing is evaluated, thus nothing evolves.

Try changing the geometry of your initial hull to lie within the constraints you want to optimize for. Given this starting point, Godzilla should do what you're asking. Keep in mind that similar or related constraints can appear in several different sections of the input file; one can be stricter than its counterpart, but they are not allowed to be in direct conflict. This is the source of about two-thirds of my Michlet frustrations- simply missing a constraint conflict somewhere.

 

Thanks a lot Matt, I understand. No idea of what makes the LCB of the initial hull, I'm using a shape 7 factors, with all factors unconstrained.
I took from an example file the offset values below, no idea of what they means, could that be the problem ?
# Offsets
1,1.0,1.0,0.0
Andrea.

 

The manual included in the \docs\ directory of Michlet includes a description of the offset series.

The line "1,1.0,1.0,0.0" reads as follows:

"First line of offsets is nonzero, so read as mathematical hull series. First value is 1, so use Series1 (3-parameter). Second value f0 is 1.0, so waterlines are parabolic. Third value f1 is 1.0, so cross sectional shapes are parabolic. Fourth value f2 is 0.0, so profile view of keel is rectangular."

Note that 1,1.0,1.0,0.0 is the Michlet series code for a Wigley parabolic hull- a mathematically elegant form that makes a nice default setting.

0 is a rectangle, 0.5 an ellipse, 1.0 a parabola, and >1.0 can be used to make cusped shapes.

Series7 and Series8 work in a similar way, with parameters:
* f0: Forebody waterplane shape (min. 0.0, max. 3.0)
* f1: Forebody cross-section shape (min. 0.0, max. 3.0)
* f2: Forebody sideview shape (min. 0.0, max. 3.0)
* f3: Afterbody waterplane shape (min. 0.0, max. 3.0)
* f4: Afterbody sideview shape (min. 0.0, max. 3.0)
* f5: Forebody relative length (min. 0.0, max. 1.0)
* f6: Afterbody relative length (min. 0.0, max. 1.0)
* f7: Afterbody cut-off ratio (min. 0.0, max. 1.0) (Series8 only)

(parameters copied from Leo's Michlet manual.)

You've probably noticed now, Andrea, that your start point is a Wigley hull defined by Series1- which, by definition, is fore/aft symmetric. Try a Series7 starter hull based on the Wigley form, but with f5 and f6 adjusted so that your midships section is somewhere in your desired LCB range. It's not perfect but it doesn't have to be- all you need is something for it to start from, the starter hull never survives anyway.

 

Andrea
I ran your in.mlt file at post #18 and the hulls I get have LCB of 55mm. I have attached the hull.mlt file and the fsoff1.txt file so you can compare with your results. I expect the reason for the value being positive is the wave interaction between the hulls. This output resulted from about 30 seconds of optimisation but shape was quite stable by then.

Rick

 

Matt, Rick,
Thanks, I had not understood those numbers were the shape 1 factors, so that in my in file I had a 7 factors shape defined soon after the shape 1, .....

Talk you later
Andrea.

 

Thanks Matt, I have the LCB under control now !!

Rick, it looks to me Godzilla tend to keep the LCB as closed to the hull central section as the constrains allow it. I'm attaching my in file just in case you are interested.
Thanks again...

 

Here's one Godzilla hull run as a possible main hull on a riverine OC1. It produced the negative LCB and LCF without being forced. This is the hull result after 500k evals.

 

Hull displacement performance versus planing

Hello, I already posted this into the cathamaran area, probably not the right one, let me try here too.
I’m not in the field, I played a bit with Leo Lazauskas toys, thanks him, but I forgot to ask this.
Understand Michlet is a for displacement hulls, is there and what is the extent of the interest in using Michelet to study a 4 m long sail catamaran ? I mean, in such a case what is the interest of the displacement performance of the hull versus the planing one ? Michlet is only interesting for big things or not ?
Thanks for reading this.
Andrea

 

Michlet is not particularly sensitive to whether a hull is large or small; as long as all the parameters are valid and the hull is fairly slender for its length, it will work just as well for a 90 cm model as it will for a 200 m ship.

It will not, however, tell you much of interest about a planing hull. The version of Michlet that most of us have access to is not able to compute changes in trim and sinkage, nor can it determine the lift on the planing surface. If your hull is designed to plane, Michlet results for it will be of little to no use above the speed where the bow begins to rise.

 

Thanks marshmat.
Is it correct to say displacement performance is interesting for all hulls because before they eventually start planing they work by displacement ?
And I probably should investigate how important is the planing into a 4 m long sailing cathamaran.
Thanks again.
Andrea.