Godzilla

Hi, im toying with Michlet and Godzilla but i dont seem to get the latter to work. It works with the example files obviously but when i create my own ini file it just pops up for a fraction of a second and closes. I gave a look at the out file and it says "bad min. beam".
My hull beam at the waterline, since thats the only portion of hull in the offsets, its about 0.45and a few pennies.
Ship Constrictions show a min width of 0.01 and a max width of 1000, i suppose those dont matter since i have to imput the values in Hull Constrictions. I started with a min of 0.45 and a max of 0.6 since the original hull is a tad more than 0.45 and i dont really want less than that and for sure not more than 60cm but didnt work and no matter what kind of value above or below 0.45 i put as a minimum beam it doesnt seem to be a good value, i even put the min a 0.01 and max at 1000 and still nada. What im missing?
There for sure will be other errors that i dont know about but since im stuck with this i will have the surprise later... and probably post again.

Oh btw, godzilla evolves just one hull from another single one or its possible to have two or more parents hull that it can use to produce a final single evolved "best" specimen? And if it can start with multiple parents how you fix the ini file for that, the examples are only for multihulls (and thats probably because godzilla can use just one parent hull i suppose)

 

If you post your in.mlt file for Godzilla I will work out what the conflict is.

The program just searches out the optimum hull for the specidied conditions and the constraints set. The main constraints are usually:
1. Total displacement.
2. WL length range. (Can be just a value)
3. WL beam range. (Can also be a set value)
4. If it is a catamaran then the separation range of the hulls can be set.
5. You can set stability limit.
6. You can limit hull shape range such as creating a hard chine, amount of rocker etc.
7. You can limit the range of a wide variety of hull paramaters.

You normally optimise for a single speed or a couple of speeds. I usually optimise for my target cruising speed. I normally start with a set displacement and design speed to get a totally unconstrained hull form. This is a good starting point as you can then establish what other constraints cost in required power.

Rick W.

 

thanks Rick, i attached the file.

Anyone knows about some software able to evolve one optimum from a couple hulls as parents? (or maybe im asking something that hasnt been done yet?)

 

The attached should now run.

You had an inconsistency between the displacement on the starting hull and constraint displacements. The Starting hull must be within the soecified range. I have set all values to 0.09. You could set all values to 0.099 if that is what you want.

Also the hull shape function value was set at 0. The data has 7 shape factors so this is what I set it to. Also the last two factors were set at 0.75. Unless you want to constrain shape these should be set at 1. I have made this change.

Let me know when you get it going. If you tell me what you are trying to do I could help you set constraints and parameters.

Your question about morphing between two hulls seems inconsistent with Godzilla. It produces the optimum hull for the constraints you set so why would you want something that is sub-optimal. Just change the constraints to suit.

Rick W.

 

well i guess the displacement was the biggest problem because fixing that in other models made it work, i wish it would have just came out witha displacement error in first place then.
the hull function shape was set at 0 because i read in the manual that if you set the shape parameters at something else will be those governing the hull generated shape rather than the offset of the actuall initial hull, or i got it wrong?, and of course i forgot to put a single dummy value in the datas after, who knows hom many other times i will forget that.
With the last values set at 1 though the hull has a rectangular waterline with tapered ends... but anyway since i took those off..

What im trying to do is to learn how to use it for now, to actually understand how useful can be for a couple builds or so and to compare the results with the changes i make, just by eye, looking at the waterlines in a cad.
I use for now the function 13, to minimize the power, it seems the one that would give a more useful and complete result to me.

The idea of starting from two hulls was in first place because i would find it fun, then because i think it would be interesting to help with a choice in between different designs, to see towards which of the parent hulls the "evolved" generations would tend or if they would take a different direction, and at the same time have an optimum resulting hull. Knowing all the compromises etc i use optimum very loosely.
And yes i understand, now, that thats not what godzilla is for, i was in doubt if it could or not.
Its not an easy software to manage and navigate though, and i have the feeling that until i will get exactly what each command line means and does i wont get much out of it. Although i might try to put together an Excel spreadsheet to compilate the ini file in a faster and easier way (that in my case can be read as dumbproof)

 

A ysefule idea is to have an in.mlt file for a catamaran and a monohull that work and keep them intact. When you get a file that doea not work compare it with the one that does.

You have to be logically correct with the data. For example a starting hull must be in the range of the optimised hull. Also you do not need to use hull offsets to make a starting hull. You can just use the parameter function. I usually start with:
1,1,1,0
This will generate a hull. It is likely the end result will be the same as if you have a very complex starting hull.

Once you get familiar with the data it is very easy to use. You would spend a lot of time making a spreadsheet that ensures the data set you enter is logically correct.

Rick W.

 

well. i dont find it much useful if its not for optimizing one of my designs so id have to start with the offsets.
And, lol, the way i see it, its that im spending lots of time trying to understand everything correctly and fixing up when (often) i forget changes here and there through the file, luckily im a lot more proficient in Excel than in Michlet.

The first trials though werent satisfactory at all, beside the hull having been optimized through a stretching of the waterline and a vertical bow (to mantain displacement since it got thinner?), the worst was that once imported the offsets in Freeship the "stations", or markers whatever, were ... squiggly, waterlines were fair but the station lines werent, so the offsets were unuseful.
Im still working out how to import the whole hull not just the underwater portion, the fact is that id be supposed to keep the displacement for the whole hull like it was immersed then? how im going to have it keeping the right displacement to the waterline through the optimization? would godzilla get it automatically when i trim the hull to lift up to the supposed right draft?
I have the impression that results would be obtained best by telling godzilla to find the best hull for a certain lenght beam, draft and displacement starting from scratch rather than refining an exhistent one. Maybe keeping playing with it ill get how it has to be done right.

 

In a practical sense there is only ONE optimum for the constraints you have set so it does not matter what you start with, the end hull will always be the same. You can alter the final shape by the constraints you set and the limits of the hull shape functions. Godzilla should be your starting point once you have an idea of displacement and the speed or power level.

Godzilla does nothing with the hull above the waterline so it does not matter what is there it just gets thrown out. You only get the underwater part. When you import it into Freeship you have to check that it has the correct number of stations and waterlines. These are the values you nominate in the in.mlt file. You also have to set the length, beam and draft in FreeShip from the hull.mlt file. Once you do this the hull will be reasonably fair. There may be bumps that will reduce wave making but typically fairing them out will not reduce performance much.

Once you fair out the hull you should export to Michlet and check performance. I have sometimes been able to get a slightly better result than Godzilla by smoothing things out.
Rick W.

 

I should stress that it is best to start with the absolute minimum constraints and full range of the shape functions. I usually only set displacement.

The end result is the lowest drag hull you could make for that displacement and design speed.

After you have this you then start applying constraints. For example of you need to limit length for storage or transport then you could set your limit. The new hull can be compared with the original to see what this constraint will cost. It only takes a few seconds to test each constraint so you can do a lrge number of tests in a few hours. It is much faster than building models and doing tank tests.

Rick W.

 

I was thinking that putting offsets for the whole hull would have saved me time in re-attaching the "dry" part later on, and i noticed that the file created from freeship keeps as hull leght the total hull lenght not the lenght at the waterline and the underwater part was stretched to that lenght too, something else i have to remember to change unless i import the whole hull. But if you say it gets thrown out does that means that in the out file the, hopefully unchanged, top offsets wouldnt even be written?

I wish there was a little more in-depth manual, this not to be taken as complaint, i mean, the software is a very special thing and its put out for free, its a pity that there isnt more people playing with it and helping developing it even more. It just would save a lot of trial and errors.

See when i think about "optimizing" i think about a final touch not starting point, to refine a design that already has some characteristics i want and i dont want to lose, im sure that in that the error is mine since i probably set the constraints wrong (i.e. the rocker, when i was looking for less power it happened that the bow got pulled down becoming vertical, the beam reduced and the rocker disappeared.) If i start from the optimum and change stuff then i could change it so much that wouldnt be useful to have an optimum hull at all.
Back to learning how to use it i guess.

 

As I said earlier an optimum is an optimum. It is not your idea of an optimum. If you want to constrain the shape in some way then you set that constraint and Godzilla should produce the most efficient underwater body for those constraints. However there is a lot more to a boat than the underwater body. Godzilla cannot anticipate what you want to do above the waterline.

For example if you want it wider then set the minimum beam to the value you want.

If you want to have rocker you have to limit the range of two particular hull functions. Look up the manual to find out which ones or simply experiment. Like I said an optimising run takes about 30 seconds so you can experiment with all of the 7 shape functions in a few minutes. Make sure the shape function seed you set is within the limits you set. You can make notes in the in.mlt file to help guide you - just put the # before the note.

I import Michlet output date into Excel and I usually always set the same plot range from 1m/s to 6m/s and 51 intevals. This gives speed at even interval of 0.1m/s. As I have said before, it is best to set up an in.mlt file for a catamaran and a file for a monohull that you can use as starter files that you know work and understand. Just change the name to say incat.mlt or inmono.mlt and leave them in the Godzilla folder.

There is a manual - Have you read every page? This is something Leo strongly recommends.

Rick W.

 

To make the most of Godzilla certainly entails reading its entire manual as Rick suggests above. The manual is not horrifically long and is quite well written.

Trial and error is part of the process of learning any piece of software. When you are dealing with Michlet and Godzilla, and getting frustrated, keep in mind a few things:

- Leo's philosophy of code writing is to make the algorithms as incredibly efficient as possible. When you go to run an enormous simulation, this is a Very Good Thing. It comes with a tradeoff- when you are trying to use all the computing power for the math, there's not much point in wasting power on a fancy graphical interface. Thus the program feels very old-school and cumbersome for someone used to Windows, at least until you learn its interface. Most highly optimized, math-heavy engineering and analysis codes share this concept.

- Part of that efficiency is that the program is not loaded up with automatic ways to fix your mistakes. If it can't make sense of your input, it will stop running and tell you to fix it. This is more difficult in the Godzilla side of things, because the different sections of your input file must not conflict with each other. If when simulating a monohull, you tell it the hull beam cannot exceed 3.0 m and the ship's overall beam cannot exceed 4.0 m, this is fine- it'll go with 3.0 m. But then say your initial hull works out to 3.2 m. It can't figure out what you want in this case. Your constraints must agree with each other. Virtually every time I have crashed Godzilla, I have traced the problem to conflicting constraints and/or initial conditions.

Godzilla will find the optimum hull for the goal and constraints you give it. If you're distorting and scaling a given hull form, then you have already done part of the work yourself and it will try to optimize that shape as much as possible, but will not necessarily find the global optimum. To find something close to a global optimum for your constraints, you need to turn it loose with a Series 7 or 8 hull and freedom to play with all parameters.

Remember with both input and output files that Michlet will only consider the hull from the waterline down. Both input and output offsets will always be this way.

 

Godzilla quit as soon as I run it

Rick W.,

Could I ask you to look at my "in" file too ?, godzilla quit as soon as I run it.
Thanks a lot.
Andrea.

 

As it says in the "out.mlt" file, the error is because your design speed is out of range.

Also, you left off the value of the special hull constraint in the last line.

The attached file should work. I set the design speed to 3.0 m/sec.

I'm not sure that your constraints will allow much optimisation to occur. You will have to manage these yourself.

Good luck,
Leo.

 

Andrea
Like Leo pointed out it is best not to overly constrain the boat initially. I also use the 7 shape function rather than the 3 you are using for the optimisation.

The process I use is to start with a set displacement and the design speed. Everything else is unconstrained so limits set very wide. This gives me an indication of what is possible. I then set various constraints to see what each costs in terms of extra power.

If you provide an idea of what you want to do I might be able to help you get there.

Rick W