Racing Canoe

Alright, so I've got Godzilla optimizing my hull according to the constraints I've specified in the in.mlt file. How do I receive my output once I'm done. I open the fsoff5 file and it reads "# blank". Maybe I'm exiting the program incorrectly because when I hit the escape button from the main window the program reads "press y to exit, any other to continue". If I hit y nothing happens, if I hit any other key, nothing happens. I just press escape again and it brings me back to the main interface. So then I hit the X at the top right of the window to close it. Am I doing something wrong. Is there a save or output command?

 

Corey
After you have let it optimise, you have to do the resistance analysis to get it it to record the information - see point 12 in my earlier post.

It should exit the program when you press 'y'. If it is not doing this then I do not know what is wrong. If you have to kill the window by pressing the X then there is a problem.

Godzilla uses the same resistance calculation as Michlet it is just an optimising routine on top of Michlet. If you can get Michlet data then you should be getting the same data for your optimised hull.

Once you have an offset text file you can import this into Freeship. I usually set the number of waterlines and stations to smaller values than the in.mlt files I sent you. You need to ensure Freeship has the right number of both when you import it. Also you need to set the hull dimensions in Freeship. You get this data from the hull.mlt file in the Godzilla folder.

I am interested to see how you go. If you cannot get it going then send me your in.mlt file and I will check that it is working OK. The ones I posted were OK.

 

Here's my in.mlt file. Everything seems to work fine until I try to exit Godzilla. I press y and nothing happens. What am I doing wrong? Rick, could you send me the mlt file for the opt_850 canoe you did?

 

Corey
Your in.mlt file ran without any alteration. I have produced the attached files.

One thing I can suggest is to restart your computer if you have forced the exit of Godzilla.

Just to recap - I expect you have seen the hull shape altering when Godzilla is running. From that point I hit any key to stop the optimisation, then <shift> "R", then "t". The drag curves are displayed. From that point you have to hit <Esc> three times to get to the exit question; then "y". It should just shut down. I use microsoft XP.

I have attached the offset, hull and speed files.

You have not set a stability constraint so the resulting boat is quite narrow and will be tippy.

You should be able to import the offset file into Freeship, set the hull dimensions and display the below water sections. Make sure you use the same number of stations and waterlines as your in.mlt when you import. I typically work on these lines to smooth out bumps and generally use fewer stations and waterlines to make this less time consuming. (I have not checked to see where the bumps are generated but once I have a hull I think can be built I recheck drag with Michlet. It might sound like a lot of fiddling but a few hours at the computer can save days with building.

 

How do I optimize the hull for 850 lbs in michlet? I attempted this by converting weight to displaced volume:

I assumed 850 lbs * .454 kg/lb / 999.1 m^3/kg = 0.386 m^3 displacement

this is what I specified in the in.mlt file as my displacement

After optimizing I take the resulting offset file into freeship and calculate the hydrostatics. These display the resulting displacement in lbs = about 650 lbs at the displacement volume of 0.386 m^3 I specified. How does that make sense?

I'm accounting for the fact that a ton in freeship is 2240 lbs.

 

The 0.386Cu.m is correct.

What length, beam and draft values did you set when you imported the hull offset file into Freeship? Did you do these conversions correctly?

It is probably simpler to just work with metric units if you are moving between Freeship and Michlet. You find that 1mm is a nice base unit as it is the sort of tolerance you aim to achieve with a hull.

What was the problem with getting Godzilla to close down?

 

Getting Godzilla to close down? I'm not sure, I just know that I don't have the problem if I use the computers at school. As I set up my in.mlt file I find myself sticking with metric units more and more in freeship.

I'm going to post my in.mlt file and resulting .fbm file upon importation into Freeship. Could you tell me why my canoe looks like a block? And how do I get Michlet to produce the above water portion of the hull? I tried the sinkage method, probably incorrectly as you'll see in the in.mlt file, and still all I get is up to the draft I specified. Slowly but surely I believe I'm reaching hull design independence. I'm sorry I've been such a burden, you probably wish you never got involved Rick.

 

Corey
Responding to your questions provides a brief distraction from my current mind numbing workload.

Have you got the fsoff1.txt file that Godzilla produced. This is what you should be importing as a surface into Freeship.

I have never looked to see if Michlet provides any surfaces above the waterline. Once I get the underwater section into Freeship I just do a vertical extrusion on the entire outer edge.

 

Corey
I am assuming that your lack of posts means you are getting interesting results.

Have you arrived at a nice hull shape yet?

Rick W.

 

Well, I had been using michlet. However, the best thing I could come up with was almost identical to the opt_850 you came up with. So I took a new approach and built University of Madison Wisconsin's National's winning hull in Freeship. After looking at the numbers, it's no surprise they won. And you will be surprised to see that their KM value = .32 meters = 12.6 inches. So now I'm trying to see how I'm going to improve on this hull - and that's not going to be easy. They say in their design report that they used michlet to design their hull. I have not been able to get michlet to design a hard-chined hull - which is the only way I see capable of attaining high KM values and low midship areas. Check out the Freeship fbm file of their hull I attached below. I like their canoe also because most of the bottom is completely flat which makes for a really low draft and easy maneuverability. Yet, the chines allow for a high Cb allowing for good tracking. Finally, I've decided that my canoe design will incorporate the flat back as I figure the water will eddie off the hull at about that width anyhow. So you might as well do the flat back to keep eddies from forming along the sides of the hull.

 

Corey
Makes sense to study a proven design.

I ran the numbers. It is about 0.3kts slower fully loaded and 0.1kts faster with 2-up than Opt-850 and substantially lower KM.

If you have a lot of weight low down in the boat then this will offset crew weight so you can tolerate lower KMt. The tolerance also depends on the crew skill level. Obviously 0.3m is tolerable but don't underestimate the effort to get weight low and the required skill of the crew.

You can constrain the shape when you optimise with Godzilla by limiting the range of the hull shape factors. The X-section is controlled by the second factor. If you make the limit small it will harden the chine if this is what you want. You need to make sure the starting seed is between the limits.

If they used the same design tool then the result will be similar. Just a matter of working out what your crew can operate in. A hard chine should help with roll damping as well. You can also do things above the waterline to increase righting moment is it rolls. Note the increasing beam on the Wisconsin hull. You can use the Cross Curves function in Freeship to compare righting moment.

 

Rick,
I recently developed a pretty decent design in Michlet and then upon importing it into freeship I flattened the stern so I could make it more streamlined. I'll list off the features:

Fully Loaded (850 lbs)
.39 KM (last years winners - .32)
draft of .16 m and .1 m bow rocker ought to make for excellent turning
decent block coefficient of .53 for good tracking (Last years winners - .59)
Midship area of 0.94 square meters (Last years winners - .97)

My aim now is to get a higher block coefficient to improve the tracking. Aside from that, I'd say it's a pretty decent boat. It's actually slightly faster than the opt_850. Does more stern rocker necessarily improve turning? I started this whole project thinking the prismatic coefficient and the length/beam ratio were the two most important issues. However, that doesn't seem to be the case here as they're pretty sad for this boat. I suppose it's because we've got a lot of weight in a short hull?

One last question - I'm considering taking this project into FLUENT for further analysis - a program which may take me several months to learn how to use. In your experience, do you think Michlet is accurate enough that I shouldn't waste my time?

 

Corey
I only looked at the data for Looking_Good fully loaded but it certainly looks good in this condition. You have slightly better stability fully loaded and much better stability lightly loaded than the Wisconsin boat. It should be a very stable canoe 2-up. If the CofG of the hull is low it should be more than tollerable 4-up.

Michlet calculates viscous drag empirically using a tried and true method. You can do this with a few equations yourself but the answer will be similar. The wave drag is based on a numerical method that I think all other CFD programs use. Other methods should arrive at similar results unless they can cater for more complexity such as sinkage.

I get good correlation with long slender boats operating with little sinkage or change in trim. I feel the numbers are within my measurement accuracy - within 5%. The results look reasonable when you push a hull a bit faster and you start to get trim changes but I doubt that Michlet would mean much at planing speed. My rule-of-thumb for planing is that hull drag is 1/8th the total weight.

Michlet can produce wave patch pictures if you want them. I have done maybe 20 comparisons between the Michlet image and real wakes from boats of different types and these always compare well. There is a wave image attached. This is for a cat and the actual was very close in terms of pattern and heights looked OK as best as could be guaged from the camera position.

In terms of allocating time I think you will get better value out of getting the hull true and smooth with weight down low rather than checking the Michlet result.

There are many other aspects of the boat AND crew that will make it perform better than trying to get another 1 or 2% from improved hull design. It takes some time on the water to get the best result. For example good paddling technique could make a difference of 10 to 20%.

 

What's the best way to constrain more fore rocker in Godzilla? Also, I keep getting wavy profile lines on the bow - for which the best solution seems to be fewer station lines although that hasn't eliminated the problem. When I go to fair them out in Freeship, I ruin the minimized resistance.

 

Corey
To answer your question here. I did mention the shape parameter controls that can be set in the in.mlt file. You will find then in the manual but here they are:

Series 7
This hull series takes the following seven shape parameters; minimum and maximum values for each are given in brackets.
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)

You can set the limits at any value in the range just make sure the seed is within your new limits.

I always start with the no constriants other than my design displacement so I get an idea of what each constraint costs in performance.

The only think I can suggest for bumps is to make sure the number of waterlines and stations you nominate in Freeship when importing a Michlet offset file matches the Michlet setting you used.